Joao Das Couves Vol4

title: Flotation Science and Engineering
author: Matis, K. A.
publisher: CRC Press
isbn10 | asin: 0824792645
print isbn13: 9780824792640
ebook isbn13: 9780585358840
language: English
subject Separation (Technology) , Flotation.
publication date: 1995
lcc: TP156.S45F46 1995eb
ddc: 660/.2842
subject: Separation (Technology) , Flotation.
Page i
Flotation Science and Engineering

Edited by
K. A. Matis
Aristotle University
Thessaloniki, Greece
Page ii
Library of Congress Cataloging-in-Publication Data
Flotation science and engineering / edited by K. A. Matis.
p. cm.
Includes bibliographical references and index.
ISBN 0-8247-9264-5 (acid-free paper)
1. Separation (Technology) 2. Flotation. I. Matis, K. A. (Kostas A.)
TP156.S45F46 1995
660'.2842dc20 94-22246
CIP
The publisher offers discounts on this book when ordered in bulk
quantities. For more information, write to Special
Sales/Professional Marketing at the address below.
This book is printed on acid-free paper.
Copyright © 1995 by Marcel Dekker, Inc. All Rights Reserved.
Neither this book nor any part may be reproduced or transmitted in
any form or by any means, electronic or mechanical, including
photocopying, microfilming, and recording, or by any information
storage and retrieval system, without permission in writing from
the publisher.
Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 10016
Current printing (last digit):
10 9 8 7 6 5 4 3 2 1
PRINTED IN THE UNITED STATES OF AMERICA
Page iii
In memory of
Prof. E. Voyatzakis
Page v
Preface
Flotation is known as an effective unit operation in mineral
processing. In the last decades new horizons have opened for this
process in chemical technology, particularly in its use as a selective
separation technique. This latter field has not been covered
extensively in the literature on flotation. Research on the process,
mainly at the laboratory stage, has indicated novel applications, for
instance, in the recovery of valuable metals from dilute aqueous
solutions. Much work, however, remains to be done, and hopefully,
sometime in the near future, the results of these applications will be
evident. Many recent advances in flotation technology have come
about through the application of new theories (i.e., adsorption
mechanisms, electrochemical control), novel flotation reagents,
various alternative flotation techniques (rather than the
conventional ones), fine particle processing, and so on. Most of
these unique features are reviewed in the present volume.
Since 1983, when a similar book on flotation was published by
Marcel Dekker, Inc. (1), a lot of progress has been made in both the
science and engineering of flotation. Problems were overcome by
innovations that met the various challenges, mostly economic,
facing the industry worldwide.
Other books published during this time were either proceedings
volumes or concerned specifically with ores and industrial mineral
applications. However, research in flotation is certainly ongoing in
an effort to improve production. Scientists and engineers are trying
to enhance flotation recov-
Page vi
ery and grade, reduce running costs of operation, and achieve
greater overall plant efficiency. Therefore, this new book will
provide up-to-date coverage of the latest developments in flotation
by known experts in their fields.
The book is divided into five parts. In Part I, the fundamentals of
flotation are briefly introduced and stress is given to technology.
Part II focuses on the mineral-processing origins of flotation, which
is a completely different area from chemical engineering. This
information will provide better understanding of the separation
process. Developments in flotation machines and units constitute
another area of application (Part III), and are useful for design
purposes.
It may be true that mineral processing has recently been at a
"crossroads" facing many problems. Thus Part IV will describe the
other successful applications of flotation, particularly water and
wastewater engineering, which are most interesting in today's
world of environmental (and energy) constraints. Current and
future prospects are stressed in Part V.
This book will be a valuable resource for scientists, technologists,
and graduate students in mining, sanitary and chemical
engineering, and chemistry. Thanks to the cooperation of people
from different backgrounds, both academic and industrial, flotation
has stopped being an "art" and is based on solid foundations. New
directions can be found in research and development.
I consider it an honor that some of the best names in the field have
contributed to this volume: I would like to express my gratitude to
them for their participation. I also wish to acknowledge that a great
part of this book was written by our research team in the
Laboratory of General and Inorganic Chemical Technology. And
finally, thanks to the publisher for providing me with this
opportunity.
K. A. MATIS
Reference
1. A. N. Clarke and D. J. Wilson, eds. Foam Flotation Theory and
Applications. New York: Marcel Dekker, Inc., 1983.
Page vii
Contents
Preface v
Contributors xi
Part I
Fundamentals
1 1
An Overview of the Process
K. A. Matis and A. I. Zouboulis
2 45
The Role of Surface Tension
Josef Freund and Bohuslav Dobias *
3 63
The Role of Bubble/Particle Size
Part II
Mineral Processing
4 89
Oxide and Silicate Flotation
Maurice C. Fuerstenau
Page viii
5 127
Flotation of Iron Sulfide Minerals: Electrokinetic Aspects
K. A. Kydros and K. A. Matis
6 157
Electrochemistry of Sulfide Minerals
Lordwell Kampamba Witika and Bohuslav Dobias *
7 179
Complexation at the Metal SulfideWater Interface
Lars Rönngren, Staffan Sjöberg, Zhong-xi Sun, and Willis
Forsling
8 207
Salt-Type Minerals
Bohuslav Dobias
Part III
Flotation UnitsCells
9 261
Developments in Flotation Equipment
Risto Lindsberg
10 273
Residence Time Distribution in Flotation Cells
Paul Mavros
11 291
Column Flotation
James A. Finch, A. Uribe-Salas, and Manqiu Xu
12 331
The Jameson Cell
Geoffrey M. Evans, Bruce W. Atkinson, and Graeme J.
Jameson
Part IV
Wastewater Engineering
13 365
Removal of Organics from Water by Nonfoaming Flotation
Kalliat T. Valsaraj
14 385
Electrolytic Flotation: An Unconventional Technique
15 415
Flocculation in SolidLiquid Separation Processes
Brij M. Moudgil and Sanjay Behl
16 441
Application of the Factorial Design to Flotation
Experiments
N. K. Lazaridis, George A. Stalidis, and K. A. Matis
Page ix
Part V
New Processes
17 455
Biomodification of Mineral Surfaces and Flotation
M. K. Yelloji Rao and P. Somasundaran
18 473
The Use of Nitrogen in Flotation
K. A. Kydros, K. A. Matis, and P. K. Spathis
19 493
Dissolved-Air Flotation of Fine Particles of Salt-Type
Minerals
G. P. Gallios, K. A. Matis, and Th. N. Balabanidis
20 517
Metal Ion Flotation in Hydrometallurgy: The Case of
Germanium Recovery
A. I. Zouboulis and K. A. Matis
Index 551
Page xi
Contributors
Bruce W. Atkinson Department of Chemical Engineering,
University of Newcastle, Newcastle, New South Wales, Australia
Th. N. Balabanidis Department of Chemistry, Aristotle University,
Sanjay Behl* Mineral Resources Research Center, University of
Florida, Gainesville, Florida
Bohuslav Dobias * Department of Surface Chemistry and Mineral
Processing, University of Regensburg, Regensburg, Germany
Geoffrey M. Evans Department of Chemical Engineering,
James A. Finch Department of Mining and Metallurgical
Engineering, McGill University, Montreal, Quebec, Canada
Willis Forsling Department of Inorganic Chemistry, Luleå
University of Technology, Luleå, Sweden
*Present affiliation: Technical Center, Engelhard Corporation,
Gordon, Georgia
Page xii
Josef Freund Department of Surface Chemistry and Mineral
Processing, University of Regensburg, Regensburg, Germany
Maurice C. Fuerstenau Department of Chemical and Metallurgical
Engineering, University of Nevada, Reno, Nevada
G. P. Gallios Department of Chemistry, Aristotle University,
Graeme J. Jameson Department of Chemical Engineering,
K. A. Kydros Department of Chemistry, Aristotle University,
N. K. Lazaridis Department of Chemistry, Aristotle University,
Risto Lindsberg Outomec Oy, Espoo, Finland
K. A. Matis Department of Chemistry, Aristotle University,
Paul Mavros Department of Chemistry, Aristotle University,
Brij M. Moudgil Department of Materials Science and
Engineering, University of Florida, Gainesville, Florida
M. K. Yelloji Rao Department of Chemical Engineering, Materials
Science, and Mineral Engineering, Henry Krumb School of Mines,
Columbia University, New York, New York
Lars Rönngren Department of Inorganic Chemistry, University of
Umeå, Umeå, Sweden
Staffan Sjöberg Department of Inorganic Chemistry, University of
Umeå, Umeå, Sweden
P. Somasundaran Department of Chemical Engineering, Materials
Science, and Mineral Engineering, Henry Krumb School of Mines,
Columbia University, New York, New York
Page xiii
P. K. Spathis Department of Chemistry, Aristotle University,
George A. Stalidis Department of Chemistry, Aristotle University,
Zhong-xi Sun Department of Inorganic Chemistry, Luleå
University of Technology, Luleå, Sweden
A. Uribe-Salas Department of Mining and Metallurgical
Engineering, McGill University, Montreal, Quebec, Canada
Kalliat T. Valsaraj Department of Chemical Engineering, Louisiana
State University, Baton Rouge, Louisiana
Lordwell Kampamba Witika Department of Surface Chemistry and
Mineral Processing, University of Regensburg, Regensburg,
Germany
Manqiu Xu J. Roy Gordon Research Laboratory, INCO Limited,
Mississauga, Ontario, Canada
A. I. Zouboulis Department of Chemistry, Aristotle University,
Page 1
1
An Overview of the Process
Aristotle University, Thessaloniki, Greece
I
Introduction
The importance of the froth flotation process to the economy of the
whole industrial world is considered to be enormous. Rough
estimates of the quantity of crushed ore that is treated by flotation
is about 9 × 109 tons annually, and the proportion of base metals
won nowadays by this process is about 95%. Certainly, without
flotation many familiar metals and inorganic raw materials would
be exceedingly scarce and costly, because the high-grade ores,
which could be processed by simple physical and mechanical
methods, have long since been used up (see, for instance,
Kitchener, Ref. 1).
Flotation originated initially from the mineral-processing field,
since for many years various particulate solids (besides minerals)
have been extracted from water by using this effective separation
method. These flotation applications include mainly the treatment
of waste water but also bacteria, coal, clays, corn, resins, proteins,
fats, rubber, dyes, glass, plastics, fruit juices, cane sugar, etc.
For many of these applications, flotation is best carried out by the
dissolved-air generation method, rather than the dispersed-air
method used for minerals. Alternatively, electrolytic flotation or
some other flotation technique is feasible. The distinguishing
feature is that where extremely
Page 2
small (or even colloidal) particles have to be floated, micro-bubbles
are required. Furthermore, what is floated is often an aggregated
colloid rather than a dispersed one; consequently, high shear rates
must be avoided in order to obviate destruction of the aggregates.
In principle, the adsorption of surface active solutes at the air/water
interface is brought about by the same forces that cause flotation of
mineral particles. Actually, there is no particle size from a 1-mm
grain of coal down to a 1-nm molecule of alcohol that cannot be
extracted by means of bubbles, provided that it has the right
balance of hydrophilic-to-hydrophobic character. Only the
mechanism of capture changes radically with the size. Colloidal
particles are no exception.
Many factors actually influence the flotation process and among
them the principal ones seem to be pulp aeration, bubble
mineralization, agitation intensity, residence time of bubbles in
pulp, pulp density, processing of the froth layer, bubble size,
particle size, specific gravity of particle, shape of particle, presence
of slimes, hydration of the surface, and adsorbed layer of flotation
reagents. Because many of these previously mentioned factors are
interrelated, the whole flotation process becomes quite difficult to
model.
Thin liquid films play a significant role as basic structural elements
of many disperse systems, as well as in processes of major
industrial importance, such as flotation (see Pugh and Maney, Ref.
2). The stability of thin liquid films from aqueous surfactant
solutions is of primary interest for froth generation and also in
order to obtain better contact conditions between air bubbles and
mineral particles (3).
The formation of the froth and its subsequent removal are
considered as the final two steps of the flotation process.
Phenomena that occur in the froth phase are known to affect
significantly the results of the whole process (4). The phenomena
that take place in the froth probably have little effect on the yield of
mineral values, because there is little evidence to suggest that there
is a substantial return of values from the froth to the pulp by
rupture and disengagement from the bubbles in the froth. However,
the froth effects have a significant effect on the grade, i.e., on the
mass fraction of values in the concentrate. This is because the
practical measures that can be taken in order to increase the mass
of values recovered from a cell, such as increasing the residence
time of the pulp, lead inevitably to an increase of the pulp quantity
entrained into the froth and recovered together with the
concentrate. Since the values normally represent only a small mass
fraction of the gangue present in the feed, a disproportionate
quantity of gangue ends up in the concentrate as flotation time
increases.
A strong effect of froth height above the level of the pulp in the cell
was found. Increasing the airflow rate increased the wetness of the
froth. An-
Page 3
other interesting conclusion was that if in a conventional flotation
cell the froth residence time could be controlled so that only the top
layers of froth were skimmed off, the grade of the concentrate
would be much improved. The improvement is clearly due to the
drainage of water and gangue into the lower layers and back into
the pulp. Even further improvement might be expected by
''washing" the pulp with a spray of clean water or possibly of water
containing a froth stabilizer (see, for instance, Jameson, Ref. 1).
The froth, because it has an increased free surface energy (as a
result of the extensive interfacial surface between liquid and air),
appears thermodynamically unstable. The foaming is a dynamic
balance, as new bubbles are added continuously from the solution,
while others disappear from the top of it. In steady-state conditions
the volumes of added and subtracted air are equal (5). The average
thickness of the liquid hymenion (film) between the bubbles is 1
µm. The quality of the foam is affected by a large number of
variables (like temperature and agitation), but it depends mainly on
the chemical composition of the hymenion, i.e., on the type and
concentration of surface active reagents that are present in the
flotation pulp (6).
The stability of the froth was described by bubble coalescence and
breakage parameters. The bubble coalescence is mainly due to the
bigger pressure, which exists inside the smaller bubbles, while the
breakage is mainly dedicated to lowering the thickness of the
hymenion due to the force of gravity. This force is also responsible
for the return of the liquid, trapped between the bubbles, back to
the pulp, a process known as drainage. When more than one
surface active substance exists in the solution, it is possible for it to
be fractionated in the froth layer, due to different adsorption rates, a
flotation method termed bubble fractionation.
There are two basic mechanisms by which the solid particles can
be recovered into the froth phase, i.e., by attachment or by
entrainment. Small amounts of gangue material recovered by
entrainment can seriously decrease the selectivity of the flotation
process. This entrainment problem places a limit on the
effectiveness of the conventional equipment when treating fine
particles. When the particles are larger than about 30 µm, the
gravitational forces become sufficient to cause the entrained
particles to drain back into the pulp from the froth phase. The rate
of entrainment was established by recognizing that the nonselective
recovery of fines is directly proportional to the amount of water
reporting in the froth product. It is possible, in this area of
controlling the phenomena occurring in the froth, that the greatest
advances will be made in the future design of flotation machines,
especially in order to handle the flotation of fine particles (7).
From a macroscopic point of view, the examination of the kinetics
of the flotation process for waste water treatment applications has
attracted the researcher's interest for many years (8). Table 1 gives
a summary of
Page 4
Table 1 A Selection of Flotation Kinetics Models.
Model R = R* (1 - e-kt) Classical first-order model, where
1a R is the recovery at time t and R*
is the ultimate recovery for an
infinite time
Model First-order model with rectan-
2 gular distribution of float-
abilities
Model Fully mixed reactor model
5
Model Gas/solid adsorption model
6
Model Second-order model with rect-
9 angular distribution of float-
abilities
Model R = (1 - z) (1 - e-kt) Three-parameter model, fast (k)
10 + z(1 - e-k*t) and slow (k*) floating components
Model Three-parameter model (gamma
13 distribution)
aThe author's numbering has been kept.
Source: After Ek, Ref. 2.
several mineral flotation kinetic models, known from the literature

(see Ek, Ref. 2). Kinetic models were used also in order to analyze
industrial flotation circuits (9). The cell flotation mixing and design
was reviewed recently (Mavros, Ref. 2). There is little prospect of
theoretically predicting the performance of a flotation cell from
direct principles; it has been argued that flotation cells are invented,
not designed (Kitchener, Ref. 1).
Significant progress has also been made in the development of
automatic control systems for flotation circuits, since the first
availability of on-stream analysis data, which took place
approximately 20 years ago. A selection of case studies
demonstrated different approaches (see, e.g., McKee, Ref. 2).
It was elsewhere suggested that the so-called parameter L of the
flotometric equation can be used to characterize the flotation
properties of sulfides, taking values between 0.02 and 0.14. The
flotometric equation has the form:
Page 5
Table 2 Particle Size Classification.
Super particle less than 0.2 µm in
colloids diameter
Colloids particle less than 1 µm
Ultrafines particle less than 5 µm
Very fines particle less than 20 µm
Fines particle less than 100 µm
Intermediatesparticle less than 500 µm
Coarse particle more than 500 µm
Source: Ref. 12.
where a50 is the maximum diameter of the particle that can be

successfully floated, while r is the density of the solids to be
floated in water (10).
Many other aspects of the flotation process, referring mainly to the
separation of particulate matter such as hydrodynamics, were also
examined. The applications of various flotation methods and
techniques were recently reviewed (11). Because classification
between the different particle sizes is needed at this stage, that by
Sivamohan (Table 2) is adopted (12).
II
Mineral-Processing Applications
The key to floatability (or nonfloatability) of minerals and
therefore to their separation by flotation is whether the solid is
hydrophobic or perhaps hydrophilic. The vast majority of minerals
are hydrophilic in varying degrees and need treatment to render
them floatable. The degree of hydrophobicity of the various
minerals can be controlled by direct surface-chemical action of
flotation reagents, which could be collectors, frothers, activators, or
depressants. Any mineral can be rendered hydrophobic by
adsorption of the appropriate surfactant, whereby the polar groups
on the mineral's surface are eliminated and covered by nonpolar
groups. This principle has two consequences:
1. Weakly polar surfaces need only a few nonpolar groups attached
to render them floatable.
2. The larger (or longer) the nonpolar part of a surfactant molecule,
the stronger the hydrophobicity it will impart when adsorbed.
Therefore, the basic question is, What are the exposed atoms, ions,
or molecular groups on the surfaces of various minerals, as they
exist in an aqueous and aerated slurry? The last phrase of this
question is a reminder that what matters for this research is not
necessarily what might be observed by a surface physicist
examining a freshly cleaned crystal in a high vacuum
Page 6
with some modern spectroscopic technique. The problem is more
complicated.
The following main mineral types are commonly distinguished
according to their surface properties:
1. Elemental minerals (e.g., gold, sulfur, graphite, diamond). These
are natural hydrophobic minerals; hence, they are easily floatable.
2. Sulfide minerals (e.g., galena, PbS; sphalerite, ZnS; chalcosite,
Cu2S; pyrite, FeS2; chalcopyrite, CuFeS2; and molybdenite, MoS2).
It is known that sulfide ores are among the most suitable minerals
for enrichment through flotation, due to their excellent flotation
properties. The principal collectors used for this category are the
thio compounds (like alkyl xanthates, dithiophosphates, and
dithiocarbamates), which are strongly chemisorbed by slightly
oxidized sulfides. Therefore, the uptake of the reagent appears to be
irreversible, and the consumption of xanthate is progressive and
depends on access of oxygen, formation and accumulation of
byproducts, etc.
It is noted that one of the most important characteristics of sulfide
minerals is their instability in atmospheric conditions and also in the
presence of water and oxygen. They are transformed slowly into the
stable oxidized minerals (e.g., PbSO4 and ZnSO4). Considerable
changes on the particle's surface occur during this oxidation, which
may be represented for a sulfide (denoted as MeS) by the simple
reaction:
It is established that the interaction between sulfide minerals and

thio collectors takes place by a corrosion-type, mixed potential
mechanism, involving simultaneous electrochemical processes (13).
Perhaps pyrite flotation and its subsequent separation from raw coal
constitutes another interesting application for obvious environmental
reasons. As mechanization increased the amount of fine coal
produced, froth flotation offers an economic method for recovering
this useful coal from the rest of mineral fines (gangues). (See also
Chap. 5.)
Generally, thio collectors exhibit a high level of chemical reactivity
with heavy metal cations, existing either at mineral surfaces or in
solution. Dixanthogen can also be formed after the reaction of
xanthate anions with copper cations, as for instance happens in the
following reaction (14):
Cuprous xanthate and dixanthogen precipitate out from the solution.

This was taken into account in our ion flotation studies of copper
from dilute aqueous solutions, using xanthate as collector (15).
The reaction between sphalerite and cupric ion may be a good
example
Page 7
of its activation, in order to enhance its separation from pyrite (iron
sulfide), when applying flotation. A pertinent example is the
mixture of zinc sulfate with an alkali, which is a common
depressant for flotation of sphalerite and its subsequent separation
from galena. Cyanide ion, another widely applied depressant for
pyrite and sphalerite, is used for their separations from galena,
although it is a harmful toxic pollutant if released. In general, the
surfaces of sulfide minerals are considered neither strongly
hydrophobic nor hydrophilic. The native and induced floatabilities
of sulfide minerals were also investigated by flotation tests in the
absence and presence of EDTA, in order to affect their surface
properties and hence to alter their floatabilities (16).
A comparison was attempted between sulfide and nonsulfide
flotation showing the difference in the actual flotation response of
chalcocite (with xanthate) and hematite (with alkyl sulfonate) (17).
The inability of sulfide minerals to form hydrogen bonds is
considered to be the principal reason why they, unlike the oxygen-
containing minerals, are not strongly hydrophilic.
3. Oxide minerals. These can be either simple oxides (e.g.,
hematite, Fe2O3; and rutile, TiO2) or mixed oxides (e.g., ilmenite,
FeTiO3; and chromite, FeCr2O4). There are two main conditions to
be met for a surfactant to operate as a flotation collector for oxides:
a. It needs to have about 10 or more carbon atoms in its
hydrophobic chain.
b. Its polar group must carry a charge of opposite sign to that
borne by the oxide, i.e., an electrostatic mechanism of
adsorption of ionic surfactants is followed.
Amphoteric oxides adsorb cationic surfactants above the point of
zero charge and anionic below. The adsorption is reversible and
gives a characteristic form of isotherm, which is related to
micellization and has been named as formation of hemimicelles.
The latter is accompanied by charge neutralization and then
reversal of charge. The surface is most hydrophobic in the
concentration region around the point of charge reversal, where the
adsorption amounts to about one monolayer. An excess of
surfactant reduces the contact angle, eventually to zero.
4. Silicates (e.g., mica and feldspar). As these minerals account for
the bulk of the rocks in the earth's crust, they commonly form the
unwanted gangue from which the more valuable species have to be
separated by flotation. These minerals accept ionic surfactants in
much the same way as oxides. As the silicate network is generally
negatively charged, amine-type collectors are commonly suitable.
The main difference is that the exposed surfaces are, in effect,
mixtures of basic and acidic oxides, and therefore
Page 8
their surface potentials depend on the conditions of leaching.
Separations of these types of minerals are mainly empirical rather
than based on theory (see also Chap. 4).
5. Salt-type minerals (e.g., magnesite, fluorite, apatite, baryte, and
scheelite). These are composed of sparingly soluble ionic salts. A
considerable amount of research has been devoted to the
mechanism of collector adsorption on minerals. Part of the
obtained results fits the electrostatic model, i.e., anionic surfactants
are adsorbed onto the positively charged surfaces and cationic, onto
the negatively charged ones. However, the most commonly applied
collectors for these minerals are the fatty acid soaps (such as
sodium oleate). For these collectors there is a good deal of
evidence in favor of the deposition of a sparingly soluble metal
soap.
The form of the product seems very unlikely to conform with the
monolayer model. There is no doubt that given a chance calcium
minerals, for example, would go on reacting with sodium oleate
almost indefinitely; this is not a reversible physical adsorption. It
was shown that adsorption isotherms for such systems are almost
certainly abstraction-by-precipitation curves (18).
Figure 1 presents the remarkable influence of the surfactant
concentra-
Figure 1
Effect of pH on froth flotation of dolomite for various anionic collector
(sodium oleate) concentrations.
Courtesy of IMM, London (19).
Page 9
tion on dolomite (Mg,Ca)(CO3)2 flotation recovery, in relation to
the solution pH (19). Salt-type minerals have recently attracted
much attention. It is proved that the first molecules of the oleate
adsorb on the surface of the calcium minerals (see Rao and
Forssberg, Ref. 2), and they are not attached by a simple
precipitation mechanism of calcium oleate at the surface as was
previously thought. Furthermore, a monolayer coverage was
proposed for calcite but a bilayer formation, for the cases of
fluorite, apatite, and scheelite. Depending on the surface potential
and on its magnitude at basic pH values, sodium oleate is either
chemisorbed on the calcium atoms of the surface or mono-
coordinated through the counter-charged sodium and calcium ions
for the monolayer filling. Diffused reflectance IR spectra were used
in order to obtain more information for the adsorption isotherms.
The mechanisms of adsorptionabstraction or retention of
surfactants on heterogeneous surfaces was extensively reviewed
(see Cases and Villieras, Ref. 2), taking into account the following
main points:
1. The importance of lateral bonds which, in general, favor the
formation of aggregates through a cooperative process on surfaces.
2. The intensity of normal (electrostatic) adsorbateadsorbent bonds,
which in turn favor:
a. the formation of two-dimensional aggregates in the monolayer
range and the formation of a bilayer at higher concentrations
when the interaction with the solid surface is strong.
b. the formation of micelles, when the temperature is lower than
the Krafft point and the interaction with the solid surface is
weak.
3. The importance of heterogeneity of the energetic surface, which
controls the size of the two-dimensional aggregates.
4. The possible association of ionic surfactants with metal cations
in the solution, which leads to their bulk or surface precipitation.
Certainly, the adsorption of surfactants from an aqueous solution is
a phenomenon of dominant importance in different applications,
ranging from ore flotation to water or waste water treatment (20).
It is natural to speculate that, in the presence of a mixture of
particles of different sizes, densities, and hydrophobicities, there
could be considerable interaction between one particle and another.
Nevertheless, that species interaction is negligible in a
conventional stirred cell (see Jameson, Ref. 1). This result is of
considerable importance because it increases the confidence one
can place on the results of tests on single species, which are used so
often in the laboratory for practical reasons.
However, this conclusion applies strictly to phenomena within the
pulp. When the particles enter the froth, there is considerable
potential for species interaction especially if the particle sizes are
radically different. An interac-
Page 10
tion between the particles was reported for the magnesite/dolomite
system during Hallimond tube flotation experiments (21). This may be
due to
1. the higher concentrations of calcium and magnesium ions existing in
the flotation dispersion
2. the possible consumption of oleate anions by these cations taking into
consideration the much stronger affinity, which exists between oleate
and calcium or magnesium ions.
Elsewhere, it was suggested that when the pH value of the dispersion
was increased, a rapid decrease in the solubility of the calcium
carbonate component of dolomite is dominated (19). Figures 2 and 3
show that the use of modifying reagents (of inorganic or organic nature)
can affect significantly the selectivity between these minerals, by
changing the respective flotation response (22). Certainly, the presence
of chemical reagents also influences the probability of adhesion.
The current knowledge of how depressants, such as starch and dextrin,
act or why they often improve or lower the selectivity of collector
action is rather inadequate. Of course they "work" by being adsorbed on
the minerals and by providing a hydrophilic coating in opposition to the
collector's
Figure 2
Influence of 50 mg L-1 sodium hexametaphosphate on flotation of magnesite and
dolomite by 10 mg L-1 sodium oleate.
Courtesy ofáMarcel Dekker, Inc. (22).
Page 11
Figure 3
Effect of carboxymethyl cellulose on dolomite flotation by 20 mg L-1 sodium
oleate with 100 mg L-1 cellulose and without modifier.
Courtesy of Marcel Dekker, Inc. (22).
hydrophobicity. But the mechanism of attachment is not perfectly
understood, so the prediction of their application's behavior is often
not possible. Yet recipes based on trial and error are in successful
everyday use.
Fuerstenau (23) reviewed and summarized several factors that
control the selective adsorption of both inorganic and organic ions at
the mineral/water interface. In particular, the role of the electrified
double layer in the adsorption of surface active substances on oxides
in aqueous media was stressed. The importance of interfacial
processes to mineral chemistry is illustrated through the application
of specific adsorption phenomena to the understanding of the
physical chemistry of flotation systems.
It has to be recognized that the present understanding of the surface
chemistry is rather inadequate for a science-based flotation
technology. For basic flotation research it is usual to obtain "pure"
samples of minerals for study. It should never be forgotten that this
strategy is often not a necessity. There is no such thing as the perfect
mineral, only endless varieties. Real minerals, as they occur in ores
to be processed, depart from the ideal in many respects, inasmuch as
they are more or less chemically impure and the mineral grains are
physically imperfect.
Page 12
III
The Problem of Fines
The problem of fine mineral particles becomes significant with the
increasing demand for minerals and the continuously declining
grade of available ores. The presence of very fine particles in
mineral enrichment methods is a difficult handling problem,
particularly because they have small mass and relatively high
surface area. The factor of small mass may lead to the following
phenomena:
1. low particle momentum
2. heterocoagulation
3. particle entrainment in concentrates (e.g., in froth)
4. low probability of collision with a bubble
5. difficulty in overcoming the energy barrier between
particleparticle and particlebubble
The factor of high surface area may lead to
1. higher dissolution rate in water
2. adsorption of larger quantities of chemical reagents
3. rigidity of froth
4. high pulp viscosity
5. undesirable coating of the valuable particles with ultrafine slimes
(gangues)
Nevertheless, a number of examples do exist where ores have been
successfully floated at finer (than usual) sizes. Such systems
include galena/xanthate at a particle size range of 120 µm;
fluorspar/oleic acid at 637 µm; molybdenite/xanthate at a less than
44 µm; leadzinc ore/xanthate at less than 10 µm; hematite/octyl-
hydroxamate at a less than 15 µm; oxidized leadzinc ores/oxime at
a less than 12.3 µm; scheelite ore/oleate at 1 µm, etc. (24).
Elsewhere, a mathematical model was derived for fine coal particle
beneficiation by flotation, its basis being the transport of solids
relative to the flux of bubble surface in the concentrate stream (25).
The background of fine particle flotation was reviewed (26), and a
descriptive illustration of the many alternative methods was given
(see Fig. 4), i.e., froth flotation (a); oil extraction and/or spherical
agglomeration (b,df); carrier separation (c); selective
flocculation/hydrophobic association (d); aggregate flotation (de);
ultra flotation (ch,cj,cji); and oil flotation (bi,dfi). Certain of these
separation methods, based on surface activity and wettability have
been discussed already; others may be future trends of this area.
Various other flotation methods and techniques (like precipitate or
adsorbing colloid flotation, and electrolytic flotation) contribute to
the solution for the difficult separation problems of fine particles.
Page 13
Figure 4
Schematic representation of surface-based concentration methods.
(From Ref. 26.)
Although the froth flotation process, invented at the end of the
previous century, is still the best available technology for
separating mineral fines, its efficiency deteriorates rapidly with
decreasing particle size below approximately 10 µm. The first
major subprocess during flotation, i.e., the particlebubble collision,
is nonselective in the sense that both hydrophobic and hydrophilic
particles have the same probability for collision with the bubble. At
constant bubble size the collision rate decreases, as the particle size
is decreasing (27). Another important factor that was also noted
was the unfavorable hydrodynamic conditions surrounding the
bubbles generated in conventional flotation machines. It has been
shown that when bubbles are too large, the fine particles follow the
streamlines around the bubbles, due to their small inertial force and
do not collide with the bubbles. The consequence of this low
probability of collision would be a low flotation rate and, hence,
poor recovery. It was suggested, however, that the probability of
bubbleparticle collision can be improved greatly when using
smaller bubbles.
Another reason for the low recovery of fine particles is related to
the low rate of flotation. The flotation rate is related rather to the
bubble number than to the bubble diameter, for bubbles smaller
than 0.1 mm. This means that for a given flow rate, bubbles must
be as small as possible.
Page 14
On the other hand, if the flotation time is increased, then the
recovery of fine particles increases also, but in a practical case the
quantity of gangue entrained and removed in the froth would
increase as well, giving rise to a reduction in the grade, i.e., the
proportion of valuable material in the concentrate.
As a conclusion, a possible way to improve the flotation of fine
particles would be the use of smaller bubbles. It is worth pointing
out that if the bubble size is reduced proportionally to the particle
size, then the rate of collision remains practically unchanged,
which possibly indicates that flotation using finer bubbles would
recover effectively finer particles (11).
Based on this observation, two nonconventional flotation
techniques, i.e., dissolved-air and electrolytic flotation, were
suggested for mineral processing of fines (28). Both of them,
already known from effluent treatment, are characterized, among
others, by the production and utilization of finer bubbles than the
conventional dispersed-air flotation technique. A large volume of
experimental work in this area has its origin in the application of
flotation in water purification and in waste water treatment.
However, any alternative processing method, which will be applied
for fine particle treatment, has to meet at the same time, energy,
environmental, and economic constraints (see Chap. 19).
The bubbles in conventional flotation machines are typically 0.52
mm in diameter. For bubbles in free rise, at a constant gas rate, the
rate of flotation should vary inversely as db3, so there would be a
real advantage to using bubbles in the range of 50100 µm in
diameter. Such bubbles are produced in dissolved-air flotation. It
has been found that bubbles 80 µm in diameter can be produced,
with air at a saturation pressure of 485 kPa (29).
Figure 5 presents some results obtained for the removal of fine
mineral particles from dilute aqueous solutions using dissolved-air
flotation and tetrasodium pyrophosphate as a modifier (30). These
results are showing also that selectivity, here of dolomite from
magnesite, may be possible applying this technique in the presence
of a modifier, like the tetrasodium pyrophosphate is. A dissolved-
air flotation microcell for floatability tests with particulate systems
has also been evaluated and tested for pyrite fines (31).
Figure 6 shows a laboratory comparison between electrolytic and
dispersed-air flotation (using a porous diffuser) of the same system
(magnesite/dolomite). The effect of pH value was examined (32).
The mineral particles are at the subsieve size range with diameters
lower than 45 µm; however, they present an appreciable difference
in its specific surface areas: 3.89 m2 g-1 of magnesite compared to
only 0.52 m2 g-1 of dolomite. This observation can partly explain
the apparent differentiation between the
Page 15
Figure 5
Effect of tetrasodium pyrophosphate on recovery of fine
carbonates by dissolved-air flotation; sodium oleate 40 mg
L-1, pH 7.1.
Courtesy of Elsevier, Amsterdam (30).
efficiencies of the two techniques, when applied to the magnesite
enrichment.
The problem with dissolved-air flotation, at least where mineral
processing is concerned, is that the volume of gas produced (in the
form of bubbles) is far lower than with mechanical systems. In a
mechanical cell, the air used is typically 1 m3air/(m3 cell)/min or
with a residence time of 10 min, the airflow rate is 10 m3/m3 pulp.
In dissolved-air flotation, the air supply rate is 810 g air/m3 raw
water, or about 0.007 m3 air/m3 water, which means three orders of
magnitude smaller than the previous case (29). Therefore, it seems
to be desirable to be able to produce large numbers of small
bubbles by mechanical means, but so far this appears to be beyond
the present technology in large scale.
Page 16
Figure 6
Comparison of electryolytic flotation with dispersed-
air flotation: (a) dolomite fines and (b) magnesite fines (20 mg L-1 commercial fatty acid collector).
(From Ref. 32.)
Page 17
For the flotation of ores of the finest particle ranges (having 100%
lower than 10 µm diameter), a free jet flotation cell has been
developed, which uses a combination of the conventional
dispersed-air with an induced-air (dissolved) device. This device
causes air to be sucked in the cell due to the under pressure caused
by the slurry jet movement (33).
The literature concerned with methods based on modifications of
surface phenomena was recently reviewed (12). Some of the basic
approach structures to the important problem of very fine particle
recovery during mineral processing have been reported. For
instance, the recognition that a better approach to this problem can
be made through modifications to well-established, surface-based
methods than by developing entirely new processes. A typical
example of the former is the hydrophobic association. Gravity or
magnetic concentration methods also proved to be effective for the
treatment of fine particles.
It would seem desirable to try to control flotation systems
involving fine particles, so that the charge on the particles and the
bubbles could be opposite in sign. However, this would be difficult
to achieve in practice, because the collector ions that are usually
applied to render the particles hydrophobic adsorb similarly on the
surface of the bubbles. Hence, there may be scope for charge
modification via control of pH or ionic concentration.
The importance of charge effects in flotation of fine particles was
first demonstrated by Collins and Jameson (34). They floated fine
polystyrene latex particles with cetyl-trimethyl-ammonium
bromide as collector, modifying the double-layer thickness with
sodium sulfate, in order to vary the charge on the particle and the
bubble. By changing the concentration of sulfate ions by 100-fold,
the charge on both the particles and the bubbles was reduced from
about 65 to 40 mV, and the rate of flotation increased by an order
of magnitude.
One of the reasons for creation of mineral fines is obviously
grinding. However, opportunities do exist to make appropriate use
of flocculants and dispersants, introduced specifically into the
grinding circuit or else-where (i.e., in the preliminary treatment
stages), in order to modify the pulp properties and thereby to
achieve enhanced grinding, mainly in the ultrafine size range. In
addition, aggregation can be retarded by using a number of other
techniques. For instance, it can be minimized by removing the
finest particle fraction continuously, by using closed circuit
grinding. Also, by cooling the grinding machines, agglomeration
can be minimized, due to rising temperatures.
Selective coagulation of colloidal particles also may be possible,
even predicted by employing a proper chemical control of the
surface potential between the different chemical
constituents/species, which are present in
Page 18
the solution. The successful and economic use of polymers as
agents for flocculating colloidal dispersions, such as those
encountered in mineral processing, depends on the careful control
of the many variables that affect the system. The most important of
these variables from a practical point of view are the polymer
concentration, the conditions of agitation, and the molecular weight
of the polymer. The important finding that a flocculating system is
thermodynamically metastable is the key concept in understanding
the role of agitation on flocculation (see, for instance, Gregory,
Ref. 2).
This was demonstrated for the case of hematite fines, where it was
found that the presence of polyacrylic acid improved significantly
the flotation of hematite at pH 4.1 and 0.001 M NaCl with sodium
dodecylsulfonate (35). The polymer dosage must be carefully
controlled, as excess polymer results in low flotation yields, due to
surface saturation by the hydrophilic polymer. Ultrafine scheelite
was also floated selectively after a preliminary flocculation step
(36).
The role of long chain xanthates, containing 2, 6, 8, and 10 carbon
atoms in a normal chain, has been examined for the flotation of
fine chalcocite grains, having diameter lower than 10 µm.
Aggregating effectiveness increased with the number of carbon
atoms in xanthate molecules. Flotation rate, recovery, and
selectivity were improved by using octyl-or decyl-xanthate (37).
In other related laboratory experiments for systems of well-defined
binary mixtures, the most important examined parameters were the
ionic radii, the ionic strength, the Hamaker constant, and the
surface potential of the particles. Several examples occurring in
clays, mineral slimes, and natural water systems were also
summarized (38).
Another example of selective coagulation or flocculation may be
that developed by the U.S. Bureau of Mines, which was used for
oxidized taconites as iron resources (39). Before the separation by
flotation of the silicate gangue took place, the preconditioning of
the ore by starch and its subsequent desliming proved to be very
important during the coagulation process.
Possibilities for selective flocculation of chalcopyrite and
pentlandite using polyacrylamide and polyethylene oxide have
been examined. The selectivity was found to depend on polymer
type and concentration and the presence of dissolved species from
each mineral. The dissolved species are found to interact with
minerals to be treated and reagents, resulting in nonselective
flocculation. The interference by dissolved species can be
eliminated by using a complexing agent, such as diphenyl-
guanidine that interacts specifically with copper minerals (40).
Certain (peculiar) experimental difficulties of these systems were
reported, which make the interpretation of the obtained results
rather diffi-
Page 19
cult (41). The chemistry of these systems is characterized by the
variety of phases, which can be formed during the course of the
process (ions, colloids, precipitates) and the difficulty to reach
chemical equilibrium (steady-state conditions). Usually, flotation is
quite a fast process. Additionally, the solubility of these minerals is
affecting primarily the pH value of the dispersion, as has been
shown already (22).
In the xanthate flotation of galena, it was found that the depression
due to slime coatings is less significant than in salt-type
mineral/oleate systems, and this appeared to be because the specific
gravity of slime coating is generally lower for the salt-type
minerals under equivalent conditions in comparison with the
sulfide minerals (42). Slime coating occurs when there is no
potential energy barrier to heterocoagulation and is dependent on
the z-potential and on sizes of the mineral and of the slime. In the
case of sulfide minerals, z-potential can vary greatly, depending
mainly on the conditioning method (43).
It was pointed out also that it would be undesirable to float near the
point of zero charge, because this could lead to coagulation of
values and gangue, hence lowering the separation. Experimental
results showed that the flotation rate was not severely affected
when the particle charge was low and it was suggested that a
charge of 2030 mV would be a practicable minimum that would
give reasonable flotation rates but prevent flocculation.
A variation of the previous method may be the so-called shear
flocculation, aiming at recovering fine particles of a valuable
mineral from a suspension containing equally fine particles of
many other minerals. Shear flocculation is an effect whereby fine
particles (0.510 µm in diameter) suspended in an aqueous
surfactant solution are made to aggregate by applying a shear field
of sufficient magnitude. Such behavior differs from electrolytic
coagulation or polymeric flocculation in that the aggregates
produced do not break up under turbulent conditions. The produced
aggregates are hydrophobic and thus suitable for direct recovery by
froth flotation, while other agents used to induce flocculation, such
as electrolytes or polymers, usually form hydrophilic flocs (44).
Among others, this already has been applied for scheelite
separation and recovery (45).
Perhaps another solution to the fine particle recovery problem may
be the application of column flotation, instead of the conventional
flotation machines (46, 47). Flotation columns were designed to
prevent hydraulic entrainment of fine particles into the froth layer,
a fact that is known to decrease the grade of concentrate in
conventional mechanical flotation cells by maintaining a net
downward fluid flow of water through the froth. Therefore, a
flotation column consists of two principal zones, the collection or
bubbling zone and the cleaning or froth zone (48). The use of tall
col-
Page 20
umns and small air bubbles in combination with countercurrent
wash water addition can greatly improve the product quality as has
been proved with a variety of fine particles, including coal and
kaolin (49) (see also Chap. 11).
IV
Water and Waste Water Treatment
For many decades, flotation has been applied successfully for
waste water treatment, as in the oil and chemical industry, for
solids and nutrient removal, but also for sludge thickening after
biological treatment, etc. (see Zouboulis et al., Ref. 2). The main
parameters affecting these applications of the flotation process
were recently analyzed (11). A comparison has also been attempted
between dissolved-air, dispersed-air, and electrolytic air techniques
applied for waste water treatment (50).
In mineral flotation the particles are quite large, typically having
diameters of 50 µm or more. Therefore, bubbles have to be large,
usually of the order of 1 mm, while the solid content in the pulp is
normally high, of the order of 25% by weight and a high selectivity
between solids is generally wanted. In effluent treatment, by
contrast, the particles are typically small (less than 20 µm in
diameter), close to neutral buoyancy, and present in very dilute
concentrations, often as low as 50 mg L-1 or 50 × 10-4% w/v.
Dissolved-air flotation is perhaps the dominant flotation technique
in water or waste water treatment. The application of dissolved-air
flotation to the mineral fine beneficiation, although it was observed
to be an efficient separation technique, showed that it cannot treat
dispersions of high density. Furthermore, high recycle ratios (of the
order of 50%) were required, being rather unrealistic (see also
Chap. 19).
Various methods of air dissolution into water at elevated pressures
have been investigated in order to examine the saturator
performance in dissolved-air (pressure) flotation. The method of
spraying water over proprietary packing media proved to be most
satisfactory. The main factors affecting the design of dissolved-air
flotation systems have been examined, and an evaluation of
interrelationships between process variables and their optimization
for design has been suggested (51). Different types of sewage
suspensions have been studied in batch and continuous flow
dissolved-air flotation units in order to assess their behavior during
flotation (52). Laboratory (batch) and full-scale (continuous)
dissolved-air flotation performance has been compared, and the
relationships between design and imposed parameters have been
studied.
Nevertheless, advances have been investigated applying this
technique in the form of a rotating mass transfer apparatus (used
already as rotary de-aerator), named ''Higee" (high g), where
centrifugal forces are used (53). It is known that the bubble
diameter is inversely proportional to the cubic
Page 21
root of the g force. In this technique the g force is replaced by the
centrifugal acceleration (wr2), which takes higher values than g,
thus lowering the produced bubble diameters. On the other hand,
the smaller the bubble size, the greater the produced interfacial area
available for mass transfer.
The latter technique has a number of special features such as small
size, low weight, short contact time, low liquid hold-up, being
insensitive to motion or orientation, rapidly attaining steady-state,
nonfoaming apparatus, quickly emptied, and easily cleaned. The
operating features, which are among the greatest considerations,
are those related to mass transfer, throughput capacity, and power
consumption.
The flotation process is suitable even at the colloidal particle size
range, as was examined in the case of lead precipitate flotation
(54). Lead hydroxide, previously precipitated from a dilute aqueous
solution by pH modification, was floated quantitatively (see Fig.
7). A comparison was attempted in the same figure with theoretical
solubility data of Pb(OH)2. Precipitate flotation can also work
selectively for separating metal cations from dilute solutions. An
experimental design (factorial) using a statistical approach
Figure 7
Dispersed-air flotation of lead (1.53 × 10-4 M), precipitated as hydroxide
by dodecylamine (2 × 10-4 M). Effect of pH value on removal and
comparison with optimum precipitation area of Pb(OH)2 (dotted line).
Flotation time 300 s, gas flow rate 3.3 cm3 s-1, ethanol 0.5%.
Courtesy of Gordon and Breach Science Publishers (54).
Page 22
has been proved successful in modeling the precipitate flotation
system of CuS/ZnS (see Chap. 16).
In order to explain the mechanism of the flotation process when it
is applied to dilute aqueous solutions, more information is required
on the different chemical species, which exist in the solution.
Figure 8 presents the ionic equilibria diagram of dilute copper
solutions. Thermodynamic equilibrium data obtained from the
literature were used in order to design this diagram with the help of
a computer program (55). The knowledge of the available metal
ions and their charges in solution can also help to foresee the
efficiency of a separation process.
The high separation efficiency of flotation has also been tested for
speciation studies, i.e., for the selective separation of metal ions
that belong to the same chemical element, but exist in different
valency states. For example, using ion or precipitate (dispersed-air)
flotation, it was found possible to separate quantitatively either
Cr(VI) or Cr(III) from an aqueous solution containing small
concentrations (5 mg L-1) from both of these species (see Fig. 9)
(56).
In most water or waste water applications, a total flotation of the
suspended solids is generally wanted. Dissolved-air flotation was
suggested to
Figure 8
Copper ionic equilibria diagram (for initial concentration 0.8 mM).
(From Ref. 55.)
Page 23
Figure 9
Influence of pH on the precipitate flotation/separation of Cr(III)/Cr(VI) mixture
(5 mg L-1), using sodium dodecylsulfate (1 × 10-4 M) as surfactant.
Courtesy of Kluwer Academic (21).
be a very effective solid/liquid separation process used for water or
waste water treatment as an alternative to sedimentation, in order to
lower the initial solid concentration before granular filtration, offering
certain advantages (see Hahn, Ref. 1). This enables filters to be
operated more effectively and to produce at a moderate cost high-
quality filtered water, which was suitable after disinfection for potable
supply.
During the treatment of surface waters, supplementary addition of
surfactants was found to be unnecessary, as the presence of natural
organic substances (humic or fulvic) even in low concentrations, lower
the surface tension, creating a sufficient foam layer (see Zabel, Ref. 2).
In Figure 10, long-term experiments (48 h duration) were presented
for comparison of flotation with sedimentation, using a pilot-scale
induced-air flotation apparatus (a variation of dissolved-air flotation)
for the physicochemical removal of phosphates from surface waters.
Better results in terms of turbidity were obtained with flotation (57).
Flocculation generally precedes dissolved-air flotation in this
application. The role of different coagulants (and coagulant aids) and
especially, of polyelectrolytes on colloid stability was examined. The
main factors
Page 24
Figure 10
Comparison of induced-
air flotation with sedimentation for phosphates removal in
turbidity terms.
Courtesy of Kluwer Academic (2).
influencing the flocculation rate (hence, the process kinetics) were
outlined, and the importance of particle size was emphasized. In
essence, the outcome of flocculation, being actually an increase of
particle size, would give similar results with the aforementioned
reasoning, when the bubble size was reduced, i.e., keeping the
probability of collision high (see Chap. 2).
In Figure 11 a comparison is presented between the phosphate
removal efficiencies of three different coagulants (FeClSO4,
Al(III), and PACl (polyaluminum chloride)). Similar results in
terms of total phosphate removal were obtained using FeClSO4 or
PACl, while lower results were found using Al(III) (57).
The problem of fines disposal and utilization, which has to be faced
by most of the mineral-processing plants (58) in relation to the
removal of residual flotation reagents (like fatty acids) found in
flotation circuit waters (59), led to the study of adsorption (and
hence removal) of these reagents on fine carbonate (magnesite and
dolomite) minerals (18). Different flotation techniques (dispersed-
air, dissolved-air, and electrolytic) were tested for the subsequent
solid/liquid separation (60). For better rationalization of water
resources use and to limit pollution as far as possible, the method
of recy-
Page 25
cling residual waters constitutes, nowadays, the ideal solution for
the majority of industrial sectors.
Figure 12 presents some other quite similar work using a technical
zeolite (NaY), which is a known cation exchanger, having size of
the order of 3.5 µm (61). The solid/liquid separation was
undertaken in the laboratory by using foam flotation, and
promising results were obtained. The zeolite had been previously
used for zinc ion exchange and removal from aqueous solutions.
Zeolites are crystalline, hydrated alumino-silicates, having a very
high internal surface area in the form of micropores of fixed
geometry.
Ideally, fine particles should be produced only for the purpose of
complete liberation of the valuable constituents (as in the case of
mixed sulfides) from gangue minerals (such as calcite and quartz).
Unfortunately, in some mineral-processing operations, as in the
case of phosphates, it is possible that almost 30% of the original
mineral is finally produced as fines that cannot be further processed
or enriched; hence, they are stockpiled in the mine area, creating a
lot of environmental problems. However, the fine size range is
well-suited to processes like adsorption, where the increased
surface area generally improves the process (and removal)
efficiency.
Figure 11
Comparison of three different coagulant concentrations for removal of total
phosphates using induced-air flotation.
(From Ref. 57.)
Page 26
Figure 12
Ion exchange removal of zinc ions by zeolite in tap water followed by dispersed-
air
flotation recovery of the exchanged form of the zeolite, using as surfactant
dodecylamime (in 0.5% ethanol solution) at two different initial concentrations:
5 × 10-5 M and 1 × 10-5 M; pH 7 to 7.5.
The removal of copper ions from dilute aqueous solutions by the
addition of mineral (pyrite) fine particles was investigated followed by
dissolved-air flotation, in order to separate the solids from the liquid
(55). Pyrite constitutes a residual or even more a solid industrial waste
by-product, and its further utilization for waste water treatment was
proposed with this in mind. Selected results are presented in Figure 13.
The transfer of mineral-processing know-how to environmental
technology may be a further advantage for both parts.
Another incentive for this work came from the so-called carrier
flotation (ultraflotation or "piggy-back" flotation), which generally uses
an auxiliary mineral (for instance, calcite) for the separation with
flotation of useful fine particles from the rest solids, as for example in
the case of anatase fines, which were separated from kaolin. This
technique was recently reviewed (45).
Electrokinetic measurements are often carried out in these
investigations in order to explain the experimental findings (see Fig.
14). In the case of copper ions removal by pyrite fines, a reversal of z-
potential was shown
Page 27
around the neutral or the slightly alkaline pH range, which was
possibly due to pyrite surface coverage by hydrolyzed copper
cationic species (see also Fig. 8). Hence, a surface-induced
hydrolysis reaction/precipitation mechanism was the more realistic
proposal to explain the experimental data.
Ion flotation certainly constitutes another promising technique (11).
A kinetic model was studied for a full-scale application of
molybdenum and tungsten ion flotation, in order to specify the
necessary number of cells in the flotation machine, resembling
somehow the design with concurrent liquid-liquid extraction process
by using immiscible solvents. This model was also used in order to
optimize the methods of technological computation, regarding the
physical parameters of the technique (62).
It has been determined that the interactions between (ionwater) and
(cationsurface active anion) govern the ion flotation selectivity when
sublate was not formed in the bulk solution. The flotation selectivity
order of anions was correlated with the experimentally determined
viscosity coefficients, whereas the flotation selectivity order of
cations was correlated with their ionic potential values. On the other
hand, if sublate precipitation
Figure 13
Effect of added pyrite concentration on copper removal ([Cu] 50 mg L-1); pH
influence.
Page 28
Figure 14
Zeta-
potential measurements of pyrite (iron sulfide) as a function of pH in the
presence of copper ions (suspension of 1 g L-1 pyrite).
Courtesy of Marcel Dekker Inc. (55).
takes place in the bulk solution, the surface charge of sublate was
the key factor influencing the floatability of an ion (see Walkowiak,
Ref. 2).
The selectivity in foam separation processes can be also influenced
by controlling the charge of the extracting solute, as for instance in
the case of metal chlorocomplexes or by regulating the pH and the
other flotation conditions (time, concentration, and type of
surfactant, etc.) (63). The effect of particle sizes, formed by
interactions between surfactant and the metal ions, has also been
studied by examining the influences of mixing time, temperature,
and surfactant chain length (64).
The separation of copper from zinc cations in laboratory
experiments was reported, applying continuous flow precipitate
flotation of the respective insoluble metal sulfides (see Fig. 15)
(65). A specific advantage of the precipitate flotation method over
that of ion flotation is the substantial lower surfactant requirement
by the first one.
Precipitate flotation using sulfides can be distinguished from the
other precipitate flotation methods by the low solubility of the
produced metal sulfides, better sludge thickening, and reduced
leaching of the removed
Page 29
toxic metal, hence, safer environmental disposal (66). Another difference from the
precipitate flotation using hydroxides, especially important for highly acidic wastes, is
the added capital and operating costs of the treatment system, since the precipitation of
hydroxides usually must be conducted at highly alkaline pH values. The nonexistence
of a unique pH value for the precipitation of all metal hydroxides is noted (being the
usual applied removal method of toxic metals), i.e., a certain pH value where all metal
hydroxides are to be found at their minimum solubility. The recovery and reuse of
applied surfactant could certainly highly improve the economics of the aforementioned
flotation methods.
Figure 15
Selective separation of the mixture Cu/Zn, using continuous flow dispersed-
air flotation (airflow rate
100 cm3 min-1). Concentrations of metals Zn 250 mg L-1, Cu 50 mg L-1, dodecylamine 5 mg L-1,
cetylpyridinium chloride 20 mg L-1, ethanol 0.025%, pH 1.7.
Page 30
V
Challenges and Conclusions
One of the most important characteristics and advantages of ion or
precipitate flotation (in the broad sense) is that it can work
selectively, contributing in this way to the recycling of the
recovered metal values (67). Numerous studies have appeared in
the literature, recently summarized by Matis and Mavros (11).
Figure 16 presents, as an example, the case of copper/arsenic(V)
separation. Phase separation was accomplished by applying the
dissolved-air flotation technique, while potassium ethyl xanthate
was used as surfactant and precipitant (15). Ferric sulfate was used
afterward as the collector/adsorbent for arsenic(V) removal
applying the adsorbing colloid flotation method.
Figure 16
Selective flotation of copper from arsenic ions by potassium
ethyl-xanthate (10% excess of stoichiometric); concentrations
of Cu and As(V) equal to 50 mg L-1.
Page 31
Figure 17
Effect of initial lead content on separation of Pb/Ge system
using dispersed-air flotation; concentration of Ge: 1 × 10-4
M, pH: 10.010.3, surfactant to ion (Pb) ratio ø: 0.5.
By-products are valuable for recycling, as for instance in the paper
industry for the removal of ink and the purification of pulp, in
order to be recycled (see Dobias et al., Ref. 2). Apparently,
flotation has been successfully used for a long time in Scandinavia
for recovering wood pulp from the effluent of paper mills, serving
the dual function of saving cellulose (raw material) and minimizing
pollution.
Other examples of selective separation are those between
germanium and arsenic(III) (68), as well as between germanium
and lead (see Fig. 17), leaving germanium (a valuable metal) in
solution (69). Dodecylamine in ethanolic solution was the
surfactant used, as in the case of lead hydroxide (shown in Fig. 7)
and a mechanism of precipitate flotation of lead as hydroxide was
also followed.
Page 32
The so-called field of adsorptive bubble separation processes has
been greatly widened by the use of specific reagents, which render
the desired solute surface active. For example, proper complexing
agents, which will react with trace metals and carry them to an
interface, can be found, as in the case of ion flotation of
germanium, which was realized after the activation of this element
through the formation of a complex between germanium and
pyrogallol (70). Similarly, a weakly surface active electrolyte can
be combined with one of opposite sign to form a highly surface
active (and maybe insoluble) complex, a process classified as ion
flotation (71). This method is also related to the use of chelating
agents (like sodium ethylene-diamine-tetra-acetate, sodium
diethyldithiocarbamate, or citric acid), which form complexes with
metal ions in solution. The subsequent separation of the complexes
from the bulk liquid by foaming has been studied recently as a
method for removal of trace toxic metals (such as cadmium, zinc,
and lead) from waste waters (72).
Selective flotation of silver, uranium, and gold from their dilute
aqueous mixtures has been examined elsewhere. In fact, several
examples concerning the combination of hydrometallurgy and
flotation have been proved successful for the selective recovery of
metals from their hydrometallurgical liquors (73,74).
Foaming and nonfoaming adsorptive bubble separations (bubble
fractionation and solvent sublation) have a significant potential for
the removal not only of several toxic metals but also of
hydrophobic organic compounds of environmental significance
from aqueous solutions (like pentachloro-phenolate and crystal
violet chloride). These methods have been found to present certain
advantages over other more conventional removal/separation
methods (see Chap. 13). For instance, solvent sublation can achieve
a higher degree of removal than solvent extraction, can remove
both volatile and nonvolatile compounds unlike air stripping, and
has less sensitivity to axial dispersion than bubble fractionation.
The utilization of flotation for trace analysis dates from the second
half of this century. The late Sebba (75) first reported the use of ion
flotation in chemical analysis. This idea as well as the even more
sensitive technique of flotation of co-precipitated ions, has
meanwhile been introduced to elemental analysis of sea or surface
water samples as a preconcentration technique. Flotation in
combination with usual determination techniques permits an
increase in the detection sensitivity. A number of comprehensive
reviews appeared in the literature covering this interesting
application of flotation (see, for example, Ref. 76).
Attention is also given to the biochemically enhanced flotation,
although the application of flotation to the separation of the
biological systems was systematically proposed several years ago
by Gaudin (77). Nowadays, mi-
Page 33
crobial ''biosorption" of metals from aqueous solutions presents a
method of increasing importance. Subsequent separation of metal-
loaded microbial cells from aqueous solutions is an important step.
Flotation of microbial cells is based on their preferential
accumulation at the gas-liquid interface and has been suggested as
a viable alternative method to other more convenient bio-separation
methods, like centrifugation (78). The surface active cells having
hydrophobic surfaces exhibit excess concentration at the interface
over that in the bulk. The use of bioconditioning prior to a flotation
process, in order to help the removal of pyrite from coal, was also
described, and the effect of bacterial adaptation on the activity of
the microorganisms was demonstrated (79).
Among the significant innovations presented during a recent
Advanced Study Institute on flotation technology held in Kallithea,
Greece, was the use of vibro-acoustic techniques in flotation. It was
suggested that it is possible to facilitate the attachment of mineral
particles to air bubbles by means of acoustical vibrations to form
bubbles with appropriate dimensions and to increase the flotation
rate, recovery, selectivity, etc. (see Stoev et al., Ref. 2). The role of
bubble size certainly requires more attention. The mineral surface
complexation was also examined by contacting potentiometric
titrations of mineral suspensions in order to predict the adsorption
behavior of these surfaces with the flotation reagents and hence the
different floatabilities of minerals (Sun and Forsling, Ref. 2).
Amines, which comprise an important category of flotation
surfactants, have found extensive application mainly for nonsulfide
metallic and silicate ore processing. They also were recently
proposed for pyrite beneficiation (80). This could be of interest in
the case of pyrite separation from ores, in which the remaining
nonsulfides containing increased concentrations of gold, silver,
etc., are valuable. It was suggested that the xanthate reagents,
already widely applicable for pyrite flotation, have a deleterious
effect on the subsequent cyanidation method used for gold
recovery. Hence, the cationic flotation of pyrite could be a solution.
Another technique, applied quite extensively in flotation research,
is the use of electrokinetic measurements in order to explain the
flotation mechanism and to predict, as well as perhaps to monitor,
this separation process. A typical example is shown in Figure 18
(80); these results were obtained for aged and freshly ground
pyrite. It is also worth pointing out that only a few researchers have
previously investigated the low acidic pH range, as the first point
of z-potential reversal was noticed at a pH value of only 2.5. The
different behavior of the two samples is due mainly to its different
surface oxidation, and these results stand in good agreement with
the literature. The addition of ferric ions in a pyrite dispersion
actually gave the same electrokinetic results with the moderate
atmospheric oxidation of
Page 34
Figure 18
(a) Zeta-potential of pyrite as a function of the pH of solution.
(b) Presence of ferric ions.

Courtesy of Academic Press (80).
Page 35
the mineral surface. The different characteristics of surface
coatings, which form when the sulfide minerals react with aqueous
solutions in the presence (and absence) of flotation reagents were
discussed elsewhere (81). (See also Chap. 5.)
A rather new modification, termed gamma flotation, relies on
controlling the interfacial tension of the solution. The minerals
desired to float should satisfy the condition gc < g1v, while the
minerals not desired to float should satisfy the condition gc > g1v.
Here, g1v is the liquidvapor interfacial tension and gc is the critical
surface tension of wetting of a solid. In this way, the separation of
pyrite, chalcocite, and chalcopyrite was demonstrated (82).
An important innovation was suggested by Outokumpu, called
skim-air flotation or flash flotation. In conventional flotation, the
mill discharge is treated in a unit flotation cell to separate easily
floatable minerals. This technique has been further developed in
order to treat the recirculating load in the grinding classification
circuit, i.e., the cyclone underflow material before further grinding,
instead of the mill discharge. Its benefits include increased
recovery by reduced overgrinding, lower dewatering costs, and
reduced flotation volume requirements (83).
Electrochemistry can certainly offer a lot in the field of flotation
research (see Chap. 6). Eh measurements have been carried out for
a long time, giving useful information, for instance concerning the
several chemical species that are responsible for different flotation
behavior. The so-called Pourbaix diagram or Eh/pH (see Fig. 19)
shows what products are possible or thermodynamically stable but
gives no information about the mechanism or the rate at which the
reactions occur.
In recent years modern electrochemical methods have been used to
investigate the mechanisms of oxidation of sulfide minerals and
their reactions with flotation reagents (13). Understanding the
possible interactions between the flotation reagents can lead to a
substantial decrease in recent consumption. Using thermodynamic
analysis data, calculations are possible for various temperatures
leading to the mapping of the flotation circuit (see Rao et al., Ref.
2). Polarograms have been used by Tolun (84) to elucidate the
mechanism of a galenaoxygenxanthate flotation system (see Fig.
20). Voltammograms were also used, superimposed upon the
flotation response. Redox control in sulfide flotation is now
possible, and it has already been applied for operational control of
a flotation circuit (85).
Another interesting aspect is the application of various gases in
order to differentiate the efficiency of mineral separation obtained
through flotation (see Chap. 18) or to increase the purification
efficiency when flotation is applied in waste water treatment.
Figure 21 presents some results on electrolytic flotation of oil
emulsions, showing different behavior between the
Page 36
Figure 19
Stability relations of pyrite and ferric hydroxide at 25°C and 1 atm total pressure, at
an activity of dissolved sulfur of 3 × 10-4 M.
(From Ref. 14.)
used gases. A membrane was applied to separate the two horizontal
electrodes (cathode and anode), thus allowing the use of a specific gas
(hydrogen or oxygen) each time (86). The flotation of 0.10.2-µm diameter
MnO2 particles with small hydrogen bubbles in the absence of collector
was also described. Flotation progressed normally, although very slowly
(87).
In the aforementioned processes the bubbles needed for flotation were led
through or induced in a solution to be treated by either mechanical
(injection of air/dispersed-air or dissolved-air flotation) or electrolytical
methods (electrolytic flotation). Another efficient method for bubble
generation is possibly the treatment of the solution by chemical flotation.
In this process bubbles can be generated, e.g., by the spontaneous
decomposition of H2O2 in an alkaline solution. The method has been
tested for the treatment of lye waste waters giving very good results (88).
The combination
Page 37
Figure 20
Polarograms at a galena electrode.
Courtesy of M. Nijhoff (84).
Figure 21
Experiments with the conventional electrolytic cell and with separated anode.
(From Ref. 86.)
Page 38
of flotation with the activated sludge process resulted in the "deep
shaft" technology, which has been applied with high efficiency for
the treatment of high-strength industrial waste waters (89).
Collectorless flotation is one more possibility having apparent
advantages (90). The collectorless application of electrolytic
flotation for salt-type mineral beneficiation was already reported
(91). Another example (see Fig. 22) presents some recent work on
pyrite fine removal from dilute aqueous dispersions in the presence
of copper ions, using dissolved-air flotation (92). The application
of dissolved-air precipitate flotation in the absence of collector and
frother has also been investigated for the purification of waste
waters containing metal ions (93).
The influence of the pulp potential for altering the flotation
efficiencies has finally been examined. The pulp potential may be
controlled by the application of proper chemical methods, i.e.,
through the activity of oxygen in the system or by adding suitable
oxidizing or reducing agents (the latter, in order to make the
potential more cathodic). As an example, hydrazinium sulfate was
applied on a bulk pyrite concentrate (see Fig.23) (94).
A further conclusion was that, with the application of suitable
flotation reagents, an industrial flotation product may be processed
either for more
Figure 22
Dissolved-air flotation of pyrite (0.5 g L-1) both in the presence (10 mg L-
1) of
surfactants and not; effect of pH.
Page 3
Figure 23
Flotation in the presence of hydrazinium sulfate. Effect of modifier concentration. Conditioning: 300 s with modifier
900 s with KEX. Flotation: 300 s, T: total solid recovery (%), A: Arsenopyrite recovery (%).
Courtesy of Pergamon Press (94).
time to achieve increased selectivity for environmental reasons (for example, to enhance the separation
of arsenic), or for technological reasons [to recover some contained values, like gold (95)]. Flotation
seems to be an important separation method for this purpose; it is a unique process in the sense that it
combines three separate phases, gas/liquid/solid (96).
Due to its advantages, like high separation rates and good clarification, flotation competes satisfactorily
with the more conventional processes in this area, like sedimentation. It is expected that in the near
future new horizons for flotation will appear, broadening the applications of this interesting field.
References
1. K. Ives, ed., The Scientific Basis of Flotation, M. Nijhoff, The Hague. (1984).
2. P. Mavros and K. A. Matis, eds., Innovations in Flotation Technology, Kluwer Academic, Dordrecht
(The Netherlands). (1992).
3. R. B. Grieves and R. N. Kyle, Models for interactions between ionic surfac-
Page 40
tants and non-surface active ions in foam fractionation
processes, Sep. Sci., Technol., 17: 465 (1982).
4. M. Szatkowski, Factors influencing behaviour of flotation froth,
Trans. IMM, 96: C115 (1987).
5. Y. V. Nguyen and C. R. Phillips, Prediction of equilibria and
mass transfer rates in continuous foam fractionation of phosphate,
Sep. Sci., 8(1): 11 (1973).
6. M. C. Fuerstenau, J. D. Miller, and M. C. Kuhn, Chemistry of
Flotation, SME/AIME, New York. (1985).
7. K. A. Matis, G. P. Gallios, and K. A. Kydros, Separation of fines
by flotation techniques, Sep. Technol., 3: 76 (1993).
8. L. D. Skrylev and K. B. Armanov, Kinetics of ion flotation, J.
Appl. Chem. USSR (Engl. transl. of Zh. Prikl Khim.), 46: 819
(1973).
9. E. C. Dowling, R. R. Klimpeland, and F. F. Aplan, Use of kinetic
models to analyze industrial flotation circuits, Proceedings, The
Reihards Schuhmann Inter. Symp. on Innovative Technology and
Reactor Design in Extraction Metallurgy (D. R. Gaskell, J. P.
Hoyer, J. E. Hoffmann, and P. J. Mackey, eds.), AIME, New York,
p. 533 (1987).
10. J. Lekki and J. Drzymala, Flotometric analysis of the
collectorless flotation of sulphide minerals, Colloid. Surf., 44: 179
(1990).
11. K. A. Matis and P. Mavros, i) Part I. Recovery of metals by ion
flotation from dilute aqueous solutions, ii) Part II. Foam/froth
flotation. Removal of particulate matter, Sep. Purif. Methods, 20: 1
& 163 (1991).
12. R. Sivamohan, The problem of recovering very fine particles in
mineral processingA review, Int. J. Miner. Process., 20: 247
(1990).
13. R. Woods, Flotation of sulfide minerals, Reagents in Mineral
Technology (P. Somasundaran and B. M. Moudgil, eds.), Marcel
Dekker, New York, p. 39. (1988).
14. M. C. Fuerstenau, M. C. Kuhn, and P. A. Elgillani, The role of
dixanthogen in xanthate flotation of pyrite, Trans. AIME, 241: 148
(1968).
15. G. A. Stalidis, K. A. Matis, and N. K. Lazaridis, Selective
separation of copper, zinc and arsenic(V) from solution by flotation
techniques, Sep. Sci. Technol., 24: 97 (1989).
16. X. Wang and E. Forssberg, EDTA-induced flotation of sulfide
minerals, J. Colloid. Interface. Sci., 140: 217 (1990).
17. T. W. Healy and W. J. Trahar, Challenges in sulfide flotation
fundamentals, Challenges in Mineral Processing (K. V. S. Sastry
and M. C. Fuerstenau, eds.), SME/AIME, Littleton, p. 3. (1989).
18. K. A. Matis, G. P. Gallios, and A. I. Zouboulis, Adsorption of
fatty acids on fine mineral particles as a purification method of
industrial waste waters. Toxicol. Environ. Chem., 2021:425 (1989).
19. K. A. Matis, G. P. Gallios, G. A. Stalidis, and C. T. Hollick,
Flotation of magnesite and dolomite by fatty acids, Trans. IMM,
98: C99 (1989).
20. B. Dobias, Surfactant adsorption on minerals related to
flotation. Structure and BondingNew Developments (M. J. Clark et
al., eds.), vol. 56, Springer Verlag, Berlin, p. 91. (1984).
Page 41
21. K. A. Matis and G. P. Gallios, Anionic flotation of magnesium
carbonates by modifiers, Int. J. Miner. Process., 25: 261 (1989).
22. G. P. Gallios and K. A. Matis, Floatability of magnesium
carbonates by sodium oleate in the presence of modifiers, Sep. Sci.
Technol., 24: 129 (1989).
23. D. W. Fuerstenau, Interfacial processes in mineral/water
systems, J. Pure Appl. Chem., 24: 135 (1970).
24. R. P. King, Flotation of fine particles, Principles of Flotation
(R. P. King, ed.), S. Afr. IMM, Johanessburg, p. 215. (1982).
25. S. A. Flynn and E. T. Woodburn, Development of a froth model
for fine particle beneficiation by flotation, Trans. IMM, 96: C191
(1987).
26. D. W. Fuerstenau, Fine particle flotation, Fine Particles
Processing (P. Somasundaran, ed.), SME/AIME, New York, p. 669.
(1980).
27. K. V. S. Sastry, Flotation of mineral fines, Beneficiation of
Mineral FinesProblems and Research Needs (P. Somasundaran and
N. Arbiter, eds.), AIME, Ann Arbor, p. 283. (1979).
28. K. A. Matis and G. P. Gallios, Dissolved-air and electrolytic
flotation, Mineral Processing at a Crossroads (B. A. Wills and R.
W. Barley, eds.), M. Nijhoff, Dordrecht, p. 37. (1986).
29. A. J. Rees, D. J. Rodman, and T. F. Zabel, Dissolved air
flotation for solid/liquid separation, J. Sep. Process. Technol., 1: 19
(1980).
30. K. A. Matis, Th. N. Balabanidis, and G. P. Gallios, Processing
of magnesium carbonate fines by dissolved-air flotation, Colloid.
Surf., 29: 191 (1988).
31. K. A. Matis, P. Mavros, and K. A. Kydros, A dissolved-air
flotation micro-cell for floatability tests with particulate systems,
Sep. Technol., 1: 255 (1991).
32. K. A. Matis and G. P. Gallios, "The problem of fine particles
and its solution by flotation techniques," Proceedings 2nd Natl.
Conf. Environ. Sci. Technol. (T. Lekkas, ed.), Univ. Aegean,
Mytilene, p. 518. (1991) (in Greek).
33. T. Atmaca and W. Simonis, Free jet flotation of oxide and
sulphide ores in finest particles ranges, Aufber. Tech., 2, 88 (1989).
34. G. L. Collins and G. J. Jameson, Double-layer effects in the
flotation of fine particles, Chem. Eng. Sci., 32: 239 (1977).
35. J. E. Gebhardt and D. W. Fuerstenau, Flotation behavior of
hematite fines flocculated with polyacrylic acid, Min. Metall.
Proc., Aug., 164 (1986).
36. P. T. L. Koh and L. J. Warren, Flotation of flocs of ultrafine
scheelite, Trans. IMM, 86: C94 (1977).
37. W. Janucz, A. Monies, A. Komosa, and J. Szczypa, "Flotation
of Fine Chalcosite Grains by Means of Long-Chain Xanthates,"
Proceedings XVIth Inter. Minr. Process. Congr. (K. S. E.
Forssberg, ed.), Elsevier, Amsterdam, p. 1155. (1988).
38. R. J. Pugh, Selective coagulation of colloidal mineral particles,
Colloid Chemistry in Mineral Processing (J. S. Laskowski and J.
Ralston, eds.), Elsevier, Amsterdam, p. 243 (1992).
39. A. F. Colombo, Selective flocculationflotation of oxidized
taconites, Theory, Practice and Process Principles for Physical
Separations (M. P. Freeman and J. A. Fitzpatrick, eds.), Eng.
Found., New York, p. 221. (1981).
Page 42
40. S. Acar and P. Somasundaran, Flocculation of sulfides and the
role of a complexing agent in it, Int. J. Miner. Process., 27: 111
(1989).
41. N. P. Finkelstein, Review of interactions in flotation of
sparingly soluble calcium minerals with anionic collectors, Trans.
IMM, 98: C157 (1989).
42. P. G. Parsonage, Effect of slime and colloidal particles on the
flotation of galena, Flotation of Sulfide Minerals (K. S. E.
Forssberg, ed.), Elsevier, Amsterdam, p. 111. (1984).
43. P. Somasundaran, Role of surface chemistry of fine sulfides in
their flotation, Complex Sulfide Ores (M. J. Jones, ed.), IMM,
London, p. 118. (1980).
44. R. Sivamohan and J. M. Cases, Dependence of shear-
flocculation on surface coverage and zeta potential, Int. J. Miner.
Process., 28: 161 (1990).
45. T. V. Subrahmanyam and K. S. E. Forssberg, Fine particle
processing: shear-flocculation and carrier flotationA review, Int. J.
Miner. Process., 30: 265 (1990).
46. P. Mavros, N. K. Lazaridis, and K. A. Matis, A study and
modelling of liquidphase mixing in a flotation column, Int. J.
Miner. Process., 26: 1 (1989).
47. P. Mavros, N. K. Lazaridis, K. A. Matis, and G. A. Stalidis, A
study of flotation column performance for the recovery of fine
particles, Sep. Sci. Technol., 25: 155 (1990).
48. K. V. S. Sastry (ed.), Column Flotation '88, Proceedings of
S.M.E. Annual Meeting, AIME, New York (1988).
49. R. H. Yoon, G. T. Adel, G. H. Luttrell, M. J. Mankosa, and A.
T. Weber, Microbubble flotation of fine particles, Interfacial
Phenomena in Biotechnology and Materials Processing (Y. A.
Attia, M. Moudgil, and S. Chander, eds.), Elsevier, Amsterdam, p.
363. (1988).
50. A. I. Zouboulis and K. A. Matis, Ionenflotation von chrom,
Fortschritte in Theorie und Praxis der Aufbereitungstechnik (K.
Uhlig, ed.), FIA, Freiberg (Germany), p. 418. (1989).
51. J. Bratby and G. v.R. Marais, Flotation, SolidLiquid Separation
Equipment Scale-up (D. B. Purchas, ed.), Uplands Press, Croydon
(UK), p. 155. (1977).
52. R. Gehr and J. G. Henry, Assessing flotation behaviour of
different types of sewage suspensions, Prog. Water. Technol., 12: 1
(1980).
53. K. Repanas, A study of bubble generation and hydrodynamics
in dissolved-air flotation, Ph. D. Thesis, Dept. Chem. and Process
Eng., Univ. Newcastle-upon-Tyne (1992).
54. A. I. Zouboulis, K. A. Matis, and G. A. Stalidis, Parameters
influencing flotation in the removal of metal ions, Int. J. Environ.
Studies (Sect. B: Environ. Sci. Technol.), 35: 183 (1990).
55. A. I. Zouboulis, K. A. Kydros, and K. A. Matis, Adsorbing
flotation of copper hydroxo-precipitates by pyrite fines, Sep. Sci.
Technol., 27: 2143 (1992).
56. A. I. Zouboulis and L. Goetz, Ion flotation as a tool for
speciation studies: selective separation in the Cr(III)/Cr(VI)
system, Toxic. Environ. Chem., 3132: 539 (1991).
57. A. I. Zouboulis and A. Grohmann, Removal of phosphates
from surface waters applying induced-air flotation, Nutrient
Removal from Wastewaters (N. Holan, ed.), Technomic, Lancaster
(USA), preprint. (1993).
Page 43
58. K. C. Dean, L. J. Froisland, and M. B. Shirts, Utilization and
stabilization of mineral wastes, Bulletin/Bureau of Mines, No. 688.
(1986).
59. T. P. McNulty and R. B. Coleman, Impact of environmental
regulation on mineral processing and hydrometallurgical plants,
Challenges in Mineral Processing (K. V. S. Sastry and M. C.
Fuerstenau, eds.), SME/AIME, Littleton, p. 638. (1989).
60. G. P. Gallios and K. A. Matis, Fatty acids removal from
effluent on mineral fines, Environ. Technol., 11: 811 (1990).
61. A. I. Zouboulis, D. Zamboulis, and K. A. Matis, Foam flotation
of zeolitesApplication on zinc ion removal, Sep. Sci. Technol., 26:
355 (1991).
62. A. M. Gol'man, ''Principles of theory and industrial application
of ion flotation", Proceedings XVI Inter. Miner. Process. Congr. (K.
S. E. Fossberg, ed.), Elsevier, Amsterdam, p. 1701. (1988).
63. W. A. Charewicz and R. B. Grieves, Separation of the
oxyanions of Re(VII), Mo(VI), Cr(VI), W(VI) and V(V), from a
multicomponent solution at pH 6 by foam fractionation. Anal. Lett.,
7(4): 233 (1974).
64. R. D. Grieves, D. Bhattacharyya, and J. K. Ghosal, Surfactant-
colligend particle size effects on ion flotation: Influences of mixing
time, temperature and surfactant chain length, Colloid. Polymer
Sci., 254: 507 (1976).
65. G. A. Stalidis, N. K. Lazaridis, K. A. Matis, and I. N.
Papadoyannis, Continuous precipitate flotation of CuSZnS, Sep.
Sci. Technol., 24: 1033 (1989).
66. G. A. Stalidis, K. A. Matis, and N. K. Lazaridis, A statistical
approach to precipitate flotation of CuSZnS, Int. J. Miner. Process.,
24: 203 (1988).
67. G. A. Stalidis, K. A. Matis, and A. I. Zouboulis, Flotation
techniques for the separation of trace pollutants, Chim. Chron.
(New Ser.), 15: 133 (1986).
68. K. A. Matis, I. N. Papadoyannis, and A. I. Zouboulis,
Separation of germanium and arsenic from solutions by flotation,
Int. J. Miner. Process., 21: 83 (1987).
69. A. I. Zouboulis, K. A. Matis, and N. K. Lazaridis, Separation of
germanium from lead in solutions by flotation, Sep. Sci. Technol.,
25: 463 (1990).
70. K. A. Matis, G. A. Stalidis, and A. I. Zoumboulis, Flotation of
germanium from dilute solutions, Sep. Sci. Technol., 23(4&5): 347
(1988).
71. A. I. Zouboulis and K. A. Matis, Ion flotation in environmental
technology, Chemosphere, 16(2&3): 623 (1987).
72. T. E. Carlesson and M. Moussavi, Chelation and foam
separation of metal ions from solution, Sep. Sci. Technol.,
23(10&11):1093 (1988).
73. E. H. De Carlo and D. M. Thomas, Removal of arsenic from
geothermal fluids by adsorptive bubble flotation with colloidal
ferric hydroxide, Environ. Sci. Technol., 19: 538 (1985).
74. H. Yamasita and K. Kato K., Recovery of individual metals
from galliumarsenicindium semiconductor wastes by flotation,
Anal. Sci., 6: 783 (1990).
75. F. Sebba, Ion Flotation, Elsevier, New York. (1961).
76. M. Caballero, R. Cela, J. A. Perez-Bustamante, Analytical
applications of some flotation techniquesA review, Talanta, 37(3):
275 (1990).
77. A. M. Gaudin, Flotation of microorganisms, Froth Flotation,
50th Anniversary Volume (D. W. Fuerstenau, ed.), SME/AIME,
New York, p. 658. (1962).
Page 44
78. S. Parthasarathy, T. R. Das, and R. Kumar, Foam separation of
microbial cells, Biotechnol Bioeng., 32: 174 (1988).
79. Y. A. Attia, Feasibility of selective biomodification of pyrite
floatability in coal desulfurization by froth flotation, Resources.,
Conserv. Recycl., 3: 169 (1990).
80. K. A. Kydros, K. A. Matis, and G. A. Stalidis, Cationic
flotation of pyrites, J. Colloid. Interface Sci., 155: 409 (1993).
81. S. Chander, Electrochemistry of sulfide flotation: Growth
characteristics of surface coatings and their properties, with special
reference to chalcopyrite and pyrite, Int. J. Miner. Process., 33: 121
(1991).
82. B. Yarar and J. Alvarez, "Separation of sulfide minerals by the
Gamma Flotation method," Proceedings XVIth Int. Miner. Process.
Congr. (K. S. E. Forssberg, ed.), Elsevier, Amsterdam, p. 547
(1988).
83. Anon., Flash flotation, Int. Min., May, 45 (1986).
84. R. Tolun, Chemistry of sulfide mineral flotation, Mineral
Processing Design (B. Yarar and Z. M. Dogan, eds.), M. Nijhoff,
Dordrecht, p. 37. (1987).
85. D. W. Fuerstenau, "Flotation science and engineeringAdvances
and challenges", Proceedings XVIth Int. Miner. Process. Congr. (K.
S. E. Forssberg, ed.), Elsevier, Amsterdam, p. 63 (1988).
86. K. A. Matis, Electrolytic flotation in industrial effluent
treatment, Chim. Chron. (New Ser.), 11: 159(1982).
87. Y. Fukui and S. Yuu, Measurement of the charge on small gas
bubbles, AlChE J., 28: 866 (1982).
88. T. Bechtold, D. Gmeiner, E. Burtscher, I. Boesch, and O.
Bobleter, Flotation of particles suspended in lye by the
decomposition of hydrogen peroxide, Sep. Sci. Technol., 24(5&6):
441 (1989).
89. T. Gallo and D. S. Sandford, The application of deep shaft
technology to the treatment of high strength industrial waste
waters, AlChE Symp. Ser., Water1979, 76(197): 288 (1980).
90. R. A. Hayes and J. Ralston, The collectorless flotation and
separation of sulphide minerals by Eh control, Int. J. Miner.
Process., 23: 55 (1988).
91. K. A. Matis, G. P. Gallios, and G. A. Stalidis, Applicability of
electrolytic flotation for the recovery of carbonate fines, Chim.
Chron. (NewSer.), 17: 29 (1988).
92. K. A. Matis, A. I. Zouboulis, and K. A. Kydros, Removal of
copper ions during flotation of pyrite fines, Fresenius Environ.
Bull., 1: 450 (1992).
93. D. Miskovic, E. Karlovic, and B. Dalmacija, The investigation
of application of dissolved air precipitate flotation in the absence of
collector and frother for the purification of waste water containing
metal ions, Chemistry for Protection of the Environment (L.
Pawlowski, A. J. Verdier, and W. J. Lacy, eds.), Elsevier,
Amsterdam, p. 245. (1984).
94. K. A. Matis, K. A. Kydros, and G. P. Gallios, Processing a bulk
pyrite concentrate by flotation reagents, Miner. Eng., 5(3&5): 331
(1992).
95. A. I. Zouboulis, K. A. Kydros, and K. A. Matis, Recovery of
gold from thiourea solutions by flotation, Hydrometallurgy, 34: 79
(1993).
96. A. I. Zouboulis and K. A. Matis, Separation of heavy metals
from effluents by flotation, Pretreatment in Chemical Water and
Wastewater Treatment (H. H. Hahn and R. Klute, eds.), Springer-
Verlag, Berlin, p. 159. (1988).
Page 45
2
The Role of Surface Tension
Josef Freund and Bohuslav Dobias *
University of Regensburg, Regensburg, Germany
Flotation requires the formation of a stabile bubbleparticle
aggregate, which enables the particle to be carried out from the
pulp. The stability of the aggregate depends on the interfacial free
energies, or the corresponding interfacial tensions, of the
solid/liquid, solid/vapor, and liquid/vapor interfaces, which are
involved in the attachment process. The tendency of the particle to
replace its solid/liquid by the solid/vapor interface, as it happens
during bubbleparticle attachment, is commonly termed
hydrophobicity. Surface chemistry has developed a number of
methods, such as the adsorption of collectors or depressants, to
change the hydrophobicity of the solid for the purpose of selective
flotation. As the bubble surface also is involved in the attachment
process, the liquid/vapor surface tension, commonly termed only
surface tension, plays a predominant role. In recent years,
liquid/vapor surface tension control is applied to regulate flotation.
As the attachment process is very similar to the dewetting of a
solid, some useful concepts, formulated previously for the wetting
process, are now applied to bubbleparticle attachment. The concept
of the critical surface tension of wetting proves to be important for
selective flotation.
I
Interfacial Free Energy and Bubble
Particle Stability
For energy calculations, the attachment process can be subdivided
into two steps (1). In the first step, the bubble surface is partly
adjusted to the shape
Page 46
of the solid surface (Fig. 1). At equilibrium, the bubble takes the
shape of a sphere in which it exhibits a minimum surface. The
deformation of this shape causes an extension of surface. The
reversible work required for this extension represents the energy
contribution W(deformation), which depends on the bubble's surface
free energy and on the shape of the attachment area. In the second
step, the adjusted bubble attaches to the solid surface. A new
solid/vapor interface is formed in this process, while the solid/liquid
and the liquid/vapor interfaces diminish. The reversible work
involved in this step is composed of the contributions of the specific
interfacial free energies gsv, gsl, and glv, respectively. The work of
the integral process (related to a unit area) can then be formulated as:
Figure 1
Schematic representation of
bubble attachment to a
solid surface.
Page 47
The bubbleparticle aggregate is stabile, if the energy gain of this
process is high enough, not to be overcome by detachment forces
acting under actual flotation conditions. Thus, for flotation to occur,
the following relationship must be satisfied:
As W(deformation) + W(detachment) ³ 0, it follows that the

necessary condition for flotation to occur is
This is the reverse condition for the spontaneous spreading of a

liquid on a given solid, forming a wetting film, i.e., gsv - gsl - glv > 0
(2,3). When the particle attaches to the bubble, the liquid film
between bubble and particle ruptures, forming the new solid/vapor
interface. This is the reverse of the wetting process, i.e., the particle
surface is "dewetted."
At the three-phase contact of the bubbleparticle aggregate, as well as
for a liquid drop resting on a solid surface, there exists a defined
contact angle q between the liquid/vapor and the solid/liquid
interface. According to Young's equation, the contact angle can be
calculated from the mechanical equilibrium of the acting interfacial
tensions:
Using this equation, the necessary condition for flotation (or

nonwetting) to occur [Eq.(3)] can be formulated as
The stability of the bubbleparticle aggregates increases as the
negative work of attachment increases. At a fixed surface tension,
this correlates with an increasing contact angle, as can be seen from
Eq.(5). At a zero contact angle, i.e., complete wetting, flotation is not
possible. If the measurements are carried out in pure water, the
contact angle can be used as a measure of the hydrophobicity of the
solid surface. In principle, flotation is possible in every liquid that
forms a contact angle greater than zero with the given solid. In this
case, the tendency to dewet is not termed hydrophobicity but
lyophobicity.
To examine the effect of surface tension on bubbleparticle stability, it
is useful to know its influence on the contact angle. Zisman (4)
carried out very extensive contact angle measurements with a variety
of solids and liquids. His investigations showed that the contact
angle always decreases with surface tension (Fig. 2). If nonpolar
solids and liquids are used, the results are linear plots of cos q vs.
glv. The surface tension for which cos q
Page 48
Figure 2
Schematic representation of Zisman's concept of the critical
surface tension of wetting and of gc determination by the
flotation method.
= 1 was denoted the critical surface tension of wetting (gc) of the
given solid. The gc value proved to be independent of the liquids
used and can therefore be applied to characterize the wettability of
the given solid.
Zisman used the gc value to distinguish between high- and low-
energy surfaces. Surfaces that are completely wetted by water (gc³
72 mN/m) are denoted as high-energy surfaces. Most metals and
minerals exhibit high-energy surfaces. Hydrocarbon surfaces are
low-energetic (gc between 20 and 35 mN/m). gc depends on
surface constitution. Fluorinated hydrocarbons show that lower,
chlorated, and nitrated hydrocarbons exhibit higher gc values. A
number of gc values depending on surface constitution were
presented by Shafrin and Zisman (5).
Those liquids, whose surface tensions are equal to or less than the
critical surface tension, wet the solid completely. Series of pure
organic liquids (e.g., n-alkanes), aqueous solutions of short chain
alcohols (methanol, ethanol), or surfactant solutions can be applied
to surface tension control. The surface tensions of the commonly
applied liquids are in the range between 18 (n-hexane) and 72
(water) mN/m at 25 °C. The surface tension of n-alkanes decreases
with alkyl chainlength. Using aqueous alcohol solutions, the
surface tension can be regulated by the water/alcohol ratio between
that of the pure alcohol (22 mN/m for methanol and ethanol) and
Page 49
that of the pure water (6). As the solid is completely wetted at gc, it
is also obvious that flotation is not possible in solutions of equal or
lower surface tension.
II
The Critical Surface Tension of Flotation
Since the beginning of the 1970s, the concept of critical surface
tension has been extended to flotation (79). The methods applied
for determining gc values are as follows:
Test tube experiments. The sample is placed in a series of test tubes
containing liquids of progressively decreasing surface tension and
shaken for a standard time. The critical surface tension lies
between the surface tension of the liquid, in which the sample just
adheres to the liquid/vapor interface, and the liquid, in which it just
remains in the bulk (10).
Film flotation tests. The sample is placed on a series of liquids of
lower densities with progressively decreasing surface tension. The
critical surface tension lies between the surface tension of the
liquid, in which the sample just floats, and the liquid, in which it
just sinks (7,1113).
Bubble pick-up tests. The adhesion of particles from a particle bed
to a captive bubble is investigated in liquids of decreasing surface
tension. The pick-up/no pick-up boundary determines the critical
surface tension (14).
Flotation tests. These tests are carried out as small-scale frothless
flotation tests on narrow size fractions of samples in a series of
liquids of decreasing surface tension. Yarar and Kaoma (15)
applied this type of flotation in a PartridgeSmith cell for the
flotation of molybdenite, elemental sulfur, oil shale, and teflon in a
series of aqueous methanol solutions. They determined the critical
surface tension of flotation by extrapolating the %-recovery vs.
surface tension plot to 0% recovery. By comparing the floatability
results with contact angle measurements carried out on flat
specimens of these solids, they were able to demonstrate that the
critical surface tension of floatability (gcf) was equal to the critical
surface tension of wetting (gc). This demonstration proved that
under these conditions (i.e., small particles, sufficient
bubbleparticle contact time, and low turbulence) floatability
depended solely on the nonwetting condition.
Commonly applied for all these experiments are aqueous solutions
of methanol or ethanol. Surface tensions slightly higher than that of
water can be established by the addition of inorganic salts, such as
NaCl or
Page 50
Na2SO4. Yarar (16) relates the higher floatability of naturally
hydrophobic solids, such as freshly ground coal, in highly
concentrated salt solutions to the increased surface tension.
The relationship between the critical surface tension of wetting and
the critical surface tension of flotation was examined theoretically by
Hornsby and Leja (17). gcf depends on the minimum contact angle
necessary for flotation under the existing conditions. Under static
conditions, when detachment forces have to be taken into account, it
may vary from <1° to 90°. A higher contact angle is required if the
particle weight increases. Dynamic conditions also can contribute to
the value of the minimum contact angle, as the minimum contact
time required for successful bubbleparticle attachment (induction
time) increases with decreasing contact angle. At the minimum
contact angle, the induction time exceeds the existing average
contact time and flotation ceases. From all these conditions, it
follows that
Flotation using liquid/vapor surface tension control, also known as

gamma flotation, can be applied to the selective separation of two
inherently hydrophobic solids if they differ sufficiently in their gcf
values. The surface tension of the liquid used has to be adjusted
between the critical surface tensions of the respective samples:
Some authors specify the floatability of heterogeneous mineral

samples by a gcf range. The lower limit of this range determines the
surface tension where no particle floats, whereas at the upper limit
all particles float (1719).
Investigations using lv-surface tension control have been carried out
on a number of naturally hydrophobic solids, such as sulfides and
coal (13,20) or talc and stibnite (21). The critical surface tensions of
hydrophobic solids are strongly dependent on their pretreatment.
Surface oxidation causes an increase of the gc values. Surface
coating, like the smearing of the soft, hydrophobic mineral on the
gangue mineral or the embedding of hydrophilic particles into the
surfaces of the hydrophobic ones, causes changes in the wettability
and floatability properties. In general, the gc values of heterogeneous
samples decrease with the increase of the hydrophobic component.
The decrease of gc of sulfides by Na2S treatment was reported by
Yarar and Kaoma (22). The same authors (15) examined the
influence of SiO2 coating in molybdenite flotation. The dependence
of gc on the oil content of oil shales was evaluated by Yarar and
Hemphill (23). Kelebek
Page 51
(24) found an increase of gc with the polar nature of molybdenite
and sulfur, which he estimated from octane contact angle
measurements.
A general treatment of the effects of particle composition and
interparticle contamination on gc is dealt with in Yarar (16). A
review of the critical surface tensions of minerals, with and without
pretreatment, was presented most recently and very comprehensively
by Ozcan (25). The lowest gc values reported were 26.5 mN/m
(chalcopyrite), 25.0 mN/m (pyrite), and 24.0 mN/m (chalocite).
III
Collector Adsorption and Critical Surface Tension
Most minerals are hydrophilic and cannot be floated because their
critical surface tensions of floatability are higher or equal to the
surface tension of water. For them flotation collectors (e.g.,
surfactants) have to be applied. An example of the effect of collector
adsorption on the critical surface tension of floatability is presented
in Figure 3 (26). It shows the floatability of fluorite particles,
covered with an adsorption layer of sodium laurate, in aqueous
methanol solutions of decreasing surface tension. The critical surface
tension of floatability decreases with increasing adsorption density
of the surfactant. At the highest adsorption density, the gc value
decreases to a lower limit of 22 mN/m, which can be related best to a
surface covered completely with methyl groups as given in a
surfactant monolayer (5,27).
By means of the collector adsorption, the mineral surface is
completely or partly covered with a hydrophobic adsorption layer. At
these sites the solid surface changes from a high-energy into a low-
energy surface. The integral energy of the surface, which is
composed of the energies of both the covered and the uncovered
sites, diminishes with increasing collector coverage. With that, when
a statistical surface coverage is assumed, the critical surface tension
also diminishes. At a monolayer coverage the solid surface exhibits
solely the hydrocarbon groups of the collector and therefore equals
the surface of an organic, low-energy solid. Likewise, the critical
surface tension changes with increasing collector coverage from that
of the uncovered, high-energy solid to that of the low-energy
collector-monolayer:
Kelebek et al. (28) found that the gc values of galena, precovered by

the adsorption of alkyl-xanthates, decreased with increasing alkyl
chainlength from 30 mN/m (methyl) to 23 mN/m (dodecyl). This
seems to be due to a higher surface coverage of low-energy alkyl
groups caused by the increase
Page 52
Figure 3
Floatability of fluorite, coated with increasing amounts of collector
(No. 1 to 4), as a function of lv-surface tension.
(From Ref. 26.)
in chainlength. Further results of gc determination of minerals
precoated with surfactants are presented in the review by Ozcan
(25). The gc values of nearly all reported minerals could be reduced
to lower than 30 mN/m by surfactant adsorption.
Some authors examined the surface, free energy of minerals related
to flotation in more detail. Bilinski * and Wójcik (2931), Chibowski
and Holysz (32,33), and Janczuk* et al. (34) determined the
dispersion and nondispersion (polar) components of the surface free
energy separately. Applying the combining laws (35)
they determined the dispersion and polar components of the solid
surface (gsd and gsp) from wetting processes (contact angle, gas-
adsorption), using liquids of known energy components. They found
that the high surface free energy of hydrophilic minerals mainly
depends on the polar compo-
Page 53
nent. With increasing surfactant adsorption, this component
decreases drastically and disappears when the surfactant adsorption
reaches a statistical monolayer. The surface free energies determined
for these monolayers were in the range between 20 and 30 mN/m
and are in accordance with the critical surface tensions found for
minerals coated in this manner. The results of the floatability
measurements correlated with the energy calculations, i.e., flotation
occurred when gsv - gsl - glv < 0. Floatability increased as the
surface free energy of the solid decreased by surfactant adsorption.
Very recent publications (36,37) discuss the surface free energy
components on the basis of the Van Oss' surface thermodynamic
approach (38). According to this, the surface free energy components
can be separated into apolar (gLW, LW = Lifshitz-van der Waals),
electron donor (g-) and electron acceptor (g+) components:
This approach provides an extended understanding of the surface

polarity, taking into account structural (hydration) forces.
IV
Dynamic Surface Tension and Flotation
Floatability measurements, carried out in homogeneous liquids, such
as solutions of water and a short-chain alcohol or a series of organic
liquids of decreasing surface tension, show a good correlation with
the corresponding static contact angle measurements. If surfactants
are used for surface tension control, however, some discrepancies
between the floatability and contact angle measurements have been
reported. These referred to a good flotation response even in the
cases that the static contact angles were zero. It was stated that time-
dependent (dynamic) phenomena may be involved. Surface tension
control using surfactants acts in a different way from the other
methods mentioned. Using surfactants, the tensions of the
liquid/vapor and liquid/solid interfaces are reduced by the oriented
adsorption of the surfactant at these interfaces. In this manner, the
water surface is substituted by an adsorption layer of lower
interfacial energy (4).
Surfactant adsorption is a time-dependent process. Thus, for static
measurements, enough time for establishing the adsorption
equilibrium must be guaranteed. It is obvious that in a flotation cell
equilibrium is not achieved. While the equilibrium at the solid/liquid
interface may be established by conditioning the mineral in the pulp,
this is not possible for the
Page 54
bubble. It remains in the cell for just a few seconds after its
generation, which may not be enough for establishing the
adsorption equilibrium.
Smith and Lai (39) demonstrated that the contact angle of a gas
bubble, attached to a silica surface in an alkaline dodecylamine
solution, decreased from 80° to 0°, as the bubble aged to 150 s.
Finch and Smith (8) examined the dynamic surface tension of
alkaline dodecylamine solutions related to flotation. Using the
bubble pressure technique, they found that the typical time range
for establishing equilibrium at the bubble exceeds 100 s. There was
a lack of correlation between the static contact angle and the
floatability results for magnetite. The floatability reached its
maximum in a concentration range where the static contact angle
decreased to zero. They referred this to the high surface tension of
the freshly generated bubble, which was still above the critical
surface tension of wetting. In a later publication (14), they
supported this by captive bubble experiments on glass, hematite,
and teflon. They demonstrated that bubble attachment to a flat
specimen of the solid, which was preconditioned in the surfactant
solution, decreased to zero as the bubble age increased from <1 to
300 s. There was, however, a good correlation between bubble
pick-up and the dynamic surface tension of the bubble generated in
the surfactant solution.
Other kinds of investigations, which give evidence of the role of
dynamic surface tension, are those dealing with the dependence of
floatability on the Critical Micelle Concentration (CMC) of the
surfactant. These are strongly related to those discussed previously,
because the reduction of interfacial tension by surfactant adsorption
increases sharply in the region just before the CMC. Above the
CMC no significant changes in the thermodynamic properties of
the system are observed under equilibrium conditions, and
therefore changes of the floatability in this region cannot be related
to equilibrium processes.
In many cases a complete cessation of flotation has been observed
whenever the surfactant concentrations reached their CMCs.
Watson and Manser (40) observed a correlation between the CMCs
of dodecylamine, as affected by changes in the pH of the solutions,
and the cessation of floatability for various silicates. This
suppression is due to the surfactant adsorption at the solid/liquid
and the liquid/vapor interfaces, as mentioned previously. On the
other hand, there are also results that show good floatability above
the CMC and a cessation at concentrations considerably higher.
Floatability tests of Dobias * (41) using alkyl sulfates as collectors
for barite flotation showed good floatability at the respective
CMCs and a cessation slightly above them. Freund and Dobiàs (42)
found a cessation of the floatability of quartz in solutions of
cetylpyridinium chloride in the concentration range between 10-3
and 10-2 mol/L, even when the CMC was shifted to much lower
concentrations by the addition of sodium chloride (Fig. 4).
Floatabil-
Page 55
Figure 4
Effect of sodium chloride addition on the CMC and the floatability of quartz in solutions of cetylpyridinium
chloride [NaClconc./mol L-1; 0 (No. 1), 10-2 (No. 2), 10-1 (No. 3)].
(From Ref. 42.)
ity tests of aluminum oxide in solutions of sodium dodecyl sulfate and sodium octadecyl sulfate
exhibited a cessation in the concentration range between 10-2 and 10-1 mol/L, although in the
case of sodium octadecyl sulfate the respective CMC was much lower (Fig. 5).
At the CMC, a mineral particle is probably covered by a surfactant double layer. While in the
primary absorbed layer, the surfactant molecules are oriented in a way that a hydrophobic
surface is formed; the second, inverse oriented surfactant layer again renders the particle
hydrophilic. Bubbleparticle attachment requires the removal of this adsorption layer, so that the
primary adsorbed, hydrophobic layer comes into contact with
Page 56
Figure 5
Floatability of aluminum oxide as a function of sodium dodecyl sulfate (SDS) and sodium octadecyl sulfate
(SOS) concentration.
(From Ref. 42.)
the gas phase of the bubble (43) (Fig. 6). This may be possible if the bubble itself is not
strongly armored by a surfactant adsorption layer. The probability of this process occurring
depends on the degree of bubble armoring, which is a function of bubble age. Dynamic surface
tension measurements, using the maximum bubble pressure technique, demonstrated that, at
small bubble age above the CMC, changes in the surface tension occur, depending on the
bubble age (Fig. 7). This relates to a still incomplete surfactant adsorption which may permit
bubble attachment at the CMC. The decrease of dynamic surface tension above the CMC
correlates well with the cessation of flotation in the concentration range above it.
From the energetic point of view [Eq. (3)], flotation is principally possi-
Page 57
Figure 6
Schematic representation of bubble attachment to a
solid covered by a surfactant double layer.
ble, if the exerted surface tension of the bubble is greater than the
adhesion tension gsv - gsl. At equilibrium, it is possible to show that
and the adhesion tension can be expressed by the critical surface

tension of wetting:
As the particle is not affected by dynamic adsorption processes

during flotation, the adhesion tension can be represented by the
respective gc, so that attachment is energetically favored [Eq. (3)] if
or
This demonstrates that bubbleparticle attachment is in principle

possible if the dynamic surface tension of the bubble exceeds the
critical surface
Page 58
Figure 7
Dynamic surface tension as a function of cetylpyridinium chloride concentration (No. 1: equilibrium, No. 2:
bubble age = 1 s, No. 3: bubble age = 0.1 s).
(From Ref. 42.)
tension of wetting. Even if the particle is covered by a second, inversely oriented surfactant
adsorption layer, the removal of this layer may be energetically favorable in the attachment
process. As mentioned previously, this layer can be considered as part of the liquid, forming the
interface to the solid. As gc is independent of the kind of liquid applied for surface tension
control, it is not affected by the formation of this surfactant layer. It depends only on the
properties of the solid surface, either uncovered or covered with a primary adsorbed,
hydrophobic surfactant layer. The energy gain in the dewetting process decreases with
decreasing dynamic surface tension. This is in accordance with the observed floatability
changes above the CMC (Figs. 4 and 7). With an increase of gc, this energy gain also decreases,
and suppression of floatability becomes more probable. At relatively high gc values, the
floatability results may therefore coincide with the
Page 59
results of the static measurements, as gc approaches the exerted
surface tension.
References
1. J. Leja and G. W. Poling, ''On the Interpretation of Contact
Angle," Proceedings of the Fifth International Mineral Processing
Congress, IMM, London, pp. 325332 (1960).
2. W. F. Cooper and W. H. Nuttall, The theory of wetting, and the
determination of wetting power of dipping and spraying fluids
containing a soap basis, J. Agr. Sci., 7: 219 (1915).
3. W. D. Harkins, The Physical Chemistry of Surface Films,
Reinhold, New York. (1952).
4. W. A. Zisman, Relation of the equilibrium contact angle to liquid
and solid constitution, 20 Years of Colloid and Surface Chemistry,
The Kendal Award Addresses (K. J. Mysele, C. M. Samour, J. H.
Hillister, eds.), ACS, Washington, D.C., p. 109. (1973).
5. E. G. Shafrin and W. A. Zisman, Constitutive relations in the
wetting of low energy surfaces and the theory of the retraction
method of preparing monolayers, J. Phys. Chem., 64: 519 (1960).
6. J. Leja, Surface Chemistry of Froth Flotation, Plenum Press,
New York. (1982).
7. J. P. Mutchler, J. Menkart, and A. M. Schwartz, Estimation of
the critical surface tension of fibers, Advan. Chem. Ser., 86: 7
(1969).
8. J. A. Finch and G. W. Smith, Dynamic surface tension of
alkaline dodecylamine acetate solutions in oxide flotation, Trans.
IMM, 81: C213 (1972).
9. B. K. Parekh and F. F. Aplan, Proc. Chemical Inst. of Canada,
Flocculation and Dispersion Symp., Division of Macromolekular
Science, Protective Coatings, Toronto, pp. 2135 (1974).
10. D. T. Hornsby and J. Leja, Critical surface tension and the
selective separation of inherently hydrophobic solids, Colloids
Surf., 1(34): 425 (1980).
11. S. Garshva, S. Contreras, and J. Goldfarb, Hydrophobic
characterisation of powders: Some criteria and experimental
evidence, Colloid Polym. Sci., 256(3): 241 (1978).
12. D. W. Fuerstenau and M. C. Williams, Characterisation of the
lyophobicity of particles by film flotation, Colloids Surf., 22(1): 87
(1987).
13. D. W. Fuerstenau, J. Diao, and J. S. Hanson, Estimation of the
distribution of surface sites and contact angles on coal particles
from film flotation data, Energy Fuels, 4(1): 34 (1990).
14. J. A. Finch and G. W. Smith, Bubble solid attachment as a
function of bubble surface tension, Can. Metall. Q., 14(1): 47
(1975).
15. B. Yarar and J. Kaoma, Estimation of the critical surface
tension of wetting of hydrophobic solids by flotation, Colloids
Surf., 11: 429 (1984).
16. B. Yarar, "Gamma Flotation: A New Approach to Flotation,
Using LiquidVapour Surface Tension Control," Developments in
Mineral Processing, Vol.
Page 60
9 (D. W. Fuerstenau, ed.), Proceedings of the 2nd Latin
American Congress on Froth Flotation, Conceptión, Chile, pp.
4164 (1985).
17. D. T. Hornsby and J. Leja, Critical surface tension of
floatability, Colloids Surf., 7(4): 339 (1983).
18. S. Kelebek and G. W. Smith, Selective flotation of inherently
hydrophobic minerals by controlling the air/solution interfacial
tension, Int. J. Miner. Process., 14(4): 275 (1985).
19. A. Marmur, W. Chen, and G. Zografi, Characterisation of
particle wettability by the measurement of floatability, J. Colloid
Interface Sci., 113(1): 114 (1986).
20. S. Kelebek, G. W. Smith, J. A. Finch, and S. Yörük, Critical
surface tension of wetting and flotation separation of hydrophobic
solids, Sep. Sci. Technol., 22(6): 1527 (1987).
21. S. Kelebek, Wetting behavior, polar characteristics and flotation
of inherently hydrophobic minerals, Trans. IMM, 96: C103 (1987).
22. B. Yarar and J. Kaoma, Critical surface tension of wetting of
sulfide minerals, Trans. Am. Inst. Min., Metall., Pet. Eng., Soc.
Min. Eng. AIME, 276: 1875 (1984).
23. B. Yarar and G. P. Hemphill, "Use of Critical Surface Tension
of Wetting in the Benefication of Oil Shales: Gamma Flotation,"
Proc. 17th Oil Shale Symp., pp. 433440 (1984).
24. S. Keiebek, Critical surface tension of wetting and of
floatability of molybdenite and sulfur, J. Colloid Interface Sci.,
124(2): 504 (1988).
25. O. Ozcan, Classification of minerals according to their critical
surface tension of wetting values, Int. J. Miner. Process., 34: 191
(1992).
26. J. Freund and B. Dobias *, Characterisation of adsorption
layers on fluorite particles with gamma flotation, Miner. Eng., 5(7):
851 (1992).
27. B. K. Parekh and F. F. Aplan, "The Critical Surface Tension of
Wetting of Minerals Coated with Collectors," 15. Congr. Int.
Mineralurgie, Vol 2, St. Etienne, France, pp. 315 (1985).
28. S. Kelebek, J. A. Finch, S. Yörük, and G. W. Smith, Wettability
and floatability of galenaxanthate system as a function of solution
surface tension, Colloids Surf., 20: 89 (1986).
29. B. Bilinski* and W. Wójcik, A gas adsorption study of the
influence of a flotation collector on the polar/nonpolar surface
properties of barite, Colloids Surf., 25: 183 (1987).
30. B. Bilinski and W. Wójcik, Gas-adsorption studies on
correlations between the floatability of minerals and the work of
water adhesion to their surfaces, Colloids Surf., 30: 275 (1988).
31. B. Bilinski and W. Wójcik, An estimation of surface free
energy of galena, Colloids Surf., 36(1): 77 (1989).
32. E. Chibowski and L. Holysz, Correlation of surface free energy
changes and floatability of quartz, J. Colloid Interface Sci., 112(1):
15 (1986).
33. E. Chibowski and L. Holysz, Adsorptiondesorption equilibria,
surface free energy and flotation activity changes in the
quartz/dodecylamine chloride-water system, Colloids Surf., 41(12):
61 (1989).
Page 61
34. B. Janczuk *, E. Chibowski, T. Bialopiotrowicz, and F.
González-Caballero, Surface free energy of celestite and its
flotation activity, Colloids Surf., 35(1): 41 (1989).
35. F. M. Fowkes, Attractive forces at interfaces, Ind. Eng. Chem.,
56(12): 40 (1964).
36. J. Skvarla*, The role of interfacial interactions in the
particle/bubble attachment, J. Disp. Sci. Technol., 12(3&4): 349
(1991).
37. L. Holysz and E. Chibowski, Surface free energy components
and floatability of barite precovered with sodium dodecyl sulfate,
Langmuir, 8(1): 303 (1992).
38. C. J. Van Oss, M. K. Chaudhury, and R. J. Good, Interfacial
Lifshitzvan der Waals and polar interactions in macroscopic
systems, Chem. Rev., 88: 927 (1988).
39. R. W. Smith and R. W. M. Lai, On the relationship between
contact angle and flotation behaviour, Trans. AIME, 235: 413
(1966).
40. D. Watson and R. M. Manser, Some factors affecting the
limiting conditions in cationic flotation of silicates, Trans. IMM,
77: C57 (1968).
41. B. Dobias*, Adsorption, electrokinetic and flotation properties
of minerals above the critical micelle concentration, Phenomena in
Mixed Surfactant Systems, Am. Chem. Soc., Washington D.C., pp.
216224 (1986).
42. J. Freund and B. Dobiás, "The Influence of Surface Tension on
the Stability of BubbleParticle Aggregates in Mineral Flotation,"
Proceedings of the International Conference on Environment and
Mineral Processing, Part II, Ostrava, Czechoslovakia, pp. 1724
(1992).
43. H. Schubert, Zum Aufbau und zur Struktur von
hydrophobierenden Tensidfilmen an Festkörperoberflächen,
Tenside-Detergents, 6: 297 (1971).
Page 63
3
The Role of Bubble/Particle Size
I
Bubbles:
Some Theoretical Considerations
For many years there have been numerous studies on bubble
formation, where an orifice (in the simplest case) has been used as
gas distributor. Bubble formation could take place in different
regimes depending on the physical properties of the liquid, orifice
size, volume of gas chamber underneath the orifice, and gas flow-
rate. Several models have been proposed in order to predict the size
of produced bubbles. All these models (one-, two-, or three-stage
models) are based on the same form of balance to describe the
growth of the bubble, which takes place in discrete phases (1).
Bubble growth takes place either at constant flow conditions, at
constant pressure conditions, or in the intermediate conditions.
Bubble formation at constant flow conditions means that the gas
flow rate passing into the bubbles remains constant during bubble
formation. This can be obtained by using a small gas chamber
underneath the orifice and a large pressure drop across the orifice.
Growth at constant pressure conditions arises from the presence of
a large gas chamber volume, such that the pressure drop across the
orifice is small; hence, the chamber pressure remains constant.
Finally, when flow rate and pressure drop are time dependent, the
bubble growth is said to be in the intermediate conditions (2).
Page 64
An example of a one-stage model, developed for prediction of the
growth of a bubble in an inviscid liquid when the surface tension is
neglected, gives the bubble volume (V) by the following equation
(3):
where C has a constant value of 1.378, g is the gravitational

acceleration, and Q is the volumetric flow rate. If the fluid is very
viscous, so that the Reynolds number associated with the growing
bubble is small, then the previous equation becomes
where µ is the viscosity of the liquid, which gives good results at

low flow rates (3).
A two-stage model has been developed in order to predict the
detachment diameter of a bubble from a nozzle (having diameter dh)
(4):
1. In the first stage, called expansion, the bubble remains attached to
the nozzle up to the point when the resultant of the lifting forces just
exceeds the resultant of the restraining.
2. Further increase in bubble volume causes imbalance of forces,
detachment, and the bubble lifts, while still being connected to the
nozzle through a neck. The neck contracts and the bubble finally
detaches itself from the nozzle.
At small flow rates, when the orifice diameter (d0) is small, the
bubble volume at detachment is determined by a balance between
the surface tension g and gravity, which finally gives the following
expression:
where the subscripts l and g refer to the liquid and gas, respectively.
This rather simple equation is not particularly accurate. Corrections
can be made using a force balance equation, solved to predict the
diameter of the produced bubble by taking into account the
buoyancy, the pressure forces, and the force due to gas momentumall
of which favor the detachment of the bubble. These forces were
equilibrated by the surface tension, the drag, and the inertia forces.
The Weber number (We), known from liquid mixing theory, was also
introduced incorporating the surface tension of the liquid phase:
Page 65
where S, L, and T are constants given by the following expressions:
where V is the bubble volume at the end of the detachment stage.

The first constant S takes into consideration the surface tension force
reduced by the force due to gas momentum. The second constant L is
due to viscous drag. The third constant T combines the drag force
and the force due to bubble acceleration. Since the drag force is
basically due to liquid inertia, the last constant is the resultant of all
inertia forces.
For low or moderate gas pressure, the following approximations can
be made:
Bubble sizes produced by shear and turbulence in a bubble column

have also been studied (5), as well as the size reduction of bubbles
by an orifice mixer (6).
The concept of drift flux analysis was introduced in order to relate
phase flow rates, gas hold-up, and physical properties. From this
theory, an estimation of the terminal bubble rise velocity can be
obtained which, in turn, can be used to calculate an estimation of the
bubble diameter (7). It was further suggested that in large columns
the actual mean bubble diameter can be 4080% higher than the
predicted values using the ideal bubbly model (8).
The hindered settling equations were adapted also, in order to predict
the mean bubble size in a swarm, giving finally the following
expression (9):
where ubs is the slip velocity between bubbles and slurry (or fluid),
eL is the fractional liquid hold-up being a function of Reynolds
number, and the subscript b refers to the bubble. Bubbles and rigid
spheres in water have virtually equivalent drag coefficients up to a
Reynolds number of approxi-
Page 66
mately 500. From a more practical point of view, bubbles formed at
orifices slow down when they leave the orifice and are then
"caught up" by the one emerging from behind. The main problem
here appears to be the prevention of coalescence between the newly
formed bubbles.
The hydrodynamic performance of very fine gas required for
dissolved-air (pressure) flotation, which are produced by dissolving
air in water under pressure and subsequent flushing to normal
pressure through a nozzle, has been studied. The results clarify the
effects of dissolved pressure, liquid flow rate, and nozzle geometry,
and empirical equations for the number of generated bubbles were
obtained (10).
Several methods for experimentally measuring the size of the
bubbles have been proposed. In mechanical flotation machines,
which have been in industrial use for many decades, gas bubbles
are generated by shear at the impeller edges. It is generally
accepted that they produce effectively and consistently rather small
bubbles, having diameters less than 1000 µm. Photographic
techniques are the most common approach for bubble diameter
measurements (7). However, these techniques are generally tedious
and quite difficult to apply in actual flotation pulps.
A number of other methods also have been developed, including
electroresistivity measurements (11), optical sensors (12), gas hold-
up combined with pressure measurements (8), and calculations
using empirical or semi-empirical correlations (7,9). The bubble
size determination in static foams is also possible by a continuous
direct method (13). Certainly, the bubble size is influenced by a
number of chemical and physical variables like viscosity, density,
ionic strength, and frother or other reagent concentrations (12).
Recently, bubble size measurements in dissolved-air flotation were
carried out in tap water at different operational conditions of liquid
flow rate, gas flow rate, or pressure. The performance of two types
of gas absorbers, being a conventional packed column and a
"Higee" absorber, was compared (14). Photographic measurements
of bubble diameters in electrolytic flotation experiments were also
performed (15).
The size of bubbles has been shown to affect the performance of
flotation processes greatly. Smaller bubbles than those encountered
in conventional froth flotation can be generated by dissolved-air
flotation or by electrolytic flotation. Although the bubbles
produced during dissolved-air flotation are generally small, having
a size on the order of 100 µm, even smaller bubbles are generally
considered advantageous, when flotation is applied for water or
waste water treatment (16). Studies of the nucleation and growth of
microbubbles during dissolved-air flotation have shown that energy
input must be imparted to a liquid supersaturated with air before
bubbles will form (10).
Page 67
The bubble size distribution also has been found to influence the
enrichment and collapse in foam during the separation of a non-
ionic surfactant. When the inlet bubble size distribution is narrow,
the enrichment factor is found to be almost the same as that for
equal size bubbles of mean size. As the inlet bubble size
distribution becomes broader, the enrichment factor decreases
because of the coalescence of the small bubbles, particularly at
lower superficial gas velocities, lower viscosities, smaller mean
bubble sizes, and lower inlet concentrations. For very broad inlet
bubble size distributions, a rapid coalescence of bubbles is found to
occur, resulting in the collapse of the foam bed (17).
After the formation of bubbles, they tend to coalesce. Coalescence
is the phenomenon where bubbles grow in diameter by feeding on
small bubbles. Generally, bubbles coalesce when two bubbles
collide and form a single bubble. There are three main factors that
favor coalescence:
1. The presence of bubbles having nonuniform diameters, where
the larger bubbles gobble up the smaller by collision.
2. Violent agitation tends to coalesce bubbles.
3. The presence of a large surface area in contact with the
bubbleliquid surface. The large surfaces serve as deposition points
for the microbubbles, where coalescence can readily take place
(18).
Also, as the bubbles move toward the surface of the solution, they
collide with each other; therefore, it is quite possible for
coalescence to occur. In this way, the surface areas per unit volume
are decreasing, and, consequently, they tend to rise faster as a result
of increased buoyancy. Both of these phenomena can be inhibited
with the addition of surfactants, because they
1. are adsorbed on the bubble surfaces during the flotation process,
having their polar groups oriented toward the liquid. The
surfactants are hydrated and eventually, the bubbles are surrounded
with a hydrated layer. This layer stabilizes the surface of the
bubbles by forming a wet hymenion and thus prevents adhesion
and reduces the coalescence potential by protecting the bubbles
from rupturing.
2. allow better adhesion of the bubbles to the particles, ions, and
solution constituents, thus permitting their effective transfer to the
surface of the solution (19).
Another consequence of surfactant adsorption is that, if bubbles are
introduced into the base of a column containing a solution of
surfactant, they will carry adsorbed material with them to the top of
the column and concentrate it perhaps in the foam, and, hence, the
liquid becomes depleted in the surfactant. Gas bubbles, of course,
are no exception to the general behavior
Page 68
of interfaces, but when bubbles are rising though a surfactant
solution, adsorption is not a steady state. Surfactant molecules are
brought to the interface by convection and diffusion, and the local
surface concentration G is therefore related to the Reynolds
number and the Schmidt number (Sc = v/D), or more generally
through the Peclet number (Pe = Re · Sc), where v is the kinematic
viscosity and D is the molecular diffusivity.
It has been reported that an increase in concentration of the
foaming reagent reduces the size of the bubbles to a certain extent.
As their concentration increases, there is a continuous removal and
assembly of surface active substances on the surface of the liquid,
and the bubble size finally reaches a steady state (20).
By enhancement of gas bubbles in dilute surfactant solution, it is
possible to form emulsions of gas in water, which if the bubbles are
small enough, have an appreciable (if limited) lifetime. These gas
emulsions, named microfoams, possess some interesting and
unusual properties and, in addition to many potential industrial
applications (e.g., for waste water treatment), provide a promising
new method for studying bubbles and the gasliquid interface (21).
The presence of inorganic salts, although not yielding any
persistent foam, may also help avoid the coalescence of bubbles
through a dynamic surface elasticity (see, for instance, Kitchener,
Ref. 22).
It is well known that colloidal particles carry a net charge as the
result of adsorbed surfactants (or collectors), and since charged
surfactant ions adsorb similarly at the gasliquid interface, it would
be reasonable to assume that bubbles would also carry a net charge.
In this case, it would follow that the sign and magnitude of the
charge on both bubbles and particles would have an important
bearing on the rate at which coalescence between the two could
occur, i.e., on the flotation rate. Such effects should be observable
not only for colloidal particles, but also for many particlebubble
combinations, where the electrostatic force of repulsion or
attraction was significant when compared with other forces arising
in the system, like the buoyancy, inertia, or viscous drag forces are.
The importance of bubble charge in its effect on the rate of
flotation of small particles was first reported by Collins and
Jameson (23). The charge was varied by addition of sodium sulfate,
which by conventional double-layer theory was collapsing the
double layer progressively, as the concentration was increased. In
these experiments, it was also found that the charge on the bubbles
and on the particles was about the same, ranging from
approximately 40 to 60 mV.
These results led to an important conclusion. In a single species
pulp, e.g., an emulsion of oil in water, it could be possible to
maximize the flotation rate by controlling the pH value of the
emulsion, for example, in
Page 69
order to give essentially zero charge on the particle and perhaps on
the bubbles. In this case the flotation rate would be maximum for
the following two reasons:
1. The electrostatic repulsion between bubbles and particles would
be a minimum.
2. The conditions would also lead to coagulation of the oil droplets
in order to form larger aggregates whose flotation rate would be
enhanced (24).
However, with a pulp containing mixed species, it would be
undesirable to try to create conditions of zero charge for the
material to be floated, because conditions may then favor
heterocoagulation of mineral values and gangue slimes. In view of
the observed effect of the particle/bubble charge on flotation rate, it
would probably be desirable to try to maintain the particle charge at
about 20 mV, thereby preventing coagulation, but without a drastic
reduction in the flotation rate.
The conditions where the particles and bubbles had opposite signs
were able to be controlled. An unexplicable but interesting
observation was that the flotation time in batch operation increased
the zeta-potential of the particles remarkably from -10 to +35 mV
within 30 min (25). The charge on the bubbles was also determined
as a function of surfactant concentration and of bubble size, but the
obtained results concerning the z-potential of the bubbles were
reliable only for very small bubbles (26).
An experimental technique has been developed in order to study
the electrokinetic behavior of gas bubbles in aqueous electrolyte
solutions. It was found that bubbles are negatively charged in NaCl
solutions, whereas under certain conditions, they become positively
charged in MgCl2 solutions. The observed charge reversal in the
pH range of 9 to 11 is attributed to the precipitation of Mg(OH)2(s)
at the gasliquid interface; while the reduction in the magnitude of
the z-potential of the bubble in the acidic pH range is considered to
be due to the specific adsorption of Mg2+ (27).
Gas bubbles generally are reported to be negatively charged in pure
water and in solutions of inorganic electrolytes (28). Therefore, the
anions are gathered basically in the air side of the interfacial
surface, while the cations are located toward the liquid side. In
general, the sign of the bubble charge is determined by the polar
head of the surfactant, when an ionic surfactant is used to produce
the bubbles. When using hydrolyzable ionic surfactants, however,
the bubble charge is significantly affected by the hydrolysis
products. In the presence of non-ionic surfactants, bubbles can be
charged either positively or negatively, depending on the pH, and
the isoelectric points appear to be related to the oxygen-to-carbon
ratio of the surfactant molecule. With ionic surfactants, an increase
in concentration results in an in-
Page 70
crease in bubble charge. With non-ionic surfactants, the z-
potentials change little in the concentration range usually studied.
The negative changes observed with air bubbles and oil droplets in
the absence of surfactants can be explained by the differences in
the hydration energies of H+ and OH- ions (29).
A convenient experimental method was suggested elsewhere for
measuring of z-potentials generating on the bubbles suspended in
aqueous surfactant solutions and created during dissolved-air
flotation. These measurements would provide predictions for the
floatabilities in the dissolved-air flotation technique (30).
Ascending to the surface (top of the cell) through the solution, the
bubbles increase in size because of the decrease of hydrostatic
pressure, so the rising velocity is also increased. The presence of
surfactants has a drastic effect on the rising bubble velocities,
decreasing them gradually as the surfactant concentration is
increased (31). The bubble rising velocity depends also on the
temperature of the solution. During dissolved-air flotation
experiments bubble rise rates at 80°C were three times faster than
at 20°C, due to the reduced viscosity of the water (32). Because of
the positive superficial potential that exists in every interfacial
surface, bubbles eventually tend to take a spherical shape, if they
have sizes below 1.5 mm, regardless of the flotation technique
applied for their generation.
II
The BubbleParticle Collision
Regarding the physicochemical aspect of the bubbleparticle
collision process, two models are considered today as the most
important (33):
1. The electrostatic model. According to this model, the energy for
the coupling of the particle with the bubble is due to electrostatic
attraction between the particle and an interfacial surface carrying
opposite charge.
2. The hydrophobic model. According to this model, the particles
become hydrophobic after the adsorption of the surfactant, and this
allows their adhesion with the bubbles.
There are three mechanisms mainly by which bubbles can assist the
suspended solids in raising them to the surface:
1. Adhesion of a gas bubble to a suspended or solid phase.
2. Trapping of a gas bubble in a floc structure, as the gas bubble
rises.
3. The adsorption of gas bubbles in a floc structure, as the floc
structure is formed (34).
Page 71
The maximum possible contact in the first mechanism may be
established by dissolving the gas into the stream that contains the
suspended phase. This mechanism is due to surface tension, and the
solid phase acts as a plane for bubble nuclei formation, when the
gas precipitates out of the solution.
The second mechanism depends upon a floc structure that is
sufficiently irregular in structure formation that a rising bubble
may be trapped in the surface irregularities. The larger the floc
diameter, the greater the probability that a rising gas bubble shall
trap itself and either not miss the floc structure or flow around the
floc structure, because of flow lines established around the rising
gas bubbles.
The third mechanism includes the formation of a floc around the
gas bubbles. This permits the formation of nuclei on the surface of
the suspended phase as well as allows the trapping of the rising
bubbles in the surface irregularities of the floc in addition to
adsorbing gas bubbles in the floc formation.
Mineral particle attachment to air bubbles is the basis of the froth
flotation process. By all the aforementioned mechanisms, a three-
phase system (gls) is created, in which the following three
interfacial tensions exert governing effects:
1. the gassolid tension (g/s)
2. the solidliquid tension (s/l)
3. the gasliquid tension (g/l).
These tensions determine together the contact angle concept
introduced by the Young equation at the point of conjunction of the
three phases. The larger the contact angle, the greater the tendency
of the bubbles to adhere with the particles. In Table 1 the influence
of several flotation variables can be seen on the speed of
particlebubble adhesion (35).
Table 1 Influence of Different Flotation Variables on the
Speed of Adhesion Between Particles and Bubbles
Size of the The bigger the size of the particle, the
particle slower the adhesion takes place.
Size of the The bigger the size of the bubble, the
bubble quicker the adhesion takes place.
Surfactants Dramatically improve the adhesion.
Depressing Dramatically decrease or hinder the
agents adhesion.
Bubbleparticle This must be in the range of 5 ms.
contact time
Source: Ref. 35.
Page 72
Figure 1
Correlation of contact angle and flotation recovery of magnesite against pH at
sodium oleate concentration of 30 mg L-1.
Courtesy of IMM, London (36).
Contact angle measurements were used to serve as a measure of the
hydrophobicity of mineral surfaces. A finite contact angle means that
the work of adhesion of water to the solid is less than the work of
cohesion of water to water. That principle leads to the question of the
nature of the intermolecular forces that bind water molecules to one
another or to solid surfaces. Figure 1 shows some selected
experimental results, where the contact angle measurements and the
flotation recoveries had the same curve shape, when plotted versus
the pH values of the solution (36).
Nevertheless, these measurements are usually carried out on flat and
pure mineral surfaces, obviously different in size, shape, surface
area, etc., than the particles to be floated. Therefore, the above
correlation seldom works in industrial practice. Arguments and
criticisms in this area have been summarized (37).
Fundamentally, when flotation is viewed as a stochastic event, the
overall probability P of particle collection by a bubble (i.e., the
flotation probability) can be represented by using the following
product of three separate probability terms (48):
Page 73
where Pc is the probability of collision between the particle and a
bubble, Pa is the probability of adhesion after collision, and Pd is
the probability of maintaining the adhesion (or detachment).
The first probability depends mainly on the hydrodynamic
characteristics of the flotation cell. For industrial flotation cells
with a dense pulp containing a mixture of mineral species, all with
their own ranges of particle size and the whole subjected to violent
turbulence, it becomes very difficult to describe quantitatively
these terms. However, three different regions can be distinguished:
1. With large particles and large bubbles, collisions will depend
greatly on inertial forces.
2. With small particles and medium size bubbles, particlebubble
collisions will occur by the sedimentating particles encountering
rising bubbles.
3. Extremely small particles will be carried along the streamlines
and will not make a close encounter with a bubble, unless they
cross the streamlines by executing Brownian motion (diffusion).
Theoretical models have been derived, especially for regions 2 and
3, but these are highly idealized, namely they are dealing with very
dilute suspensions of monodispersed spherical particles and large,
monodispersed spherical bubbles. The hydrodynamic theory of
collision between these model particles and bubbles has been
developed (39).
The second probability depends on the superficial forces and the
third on the collision with other particles and on the turbulence of
the flotation cell. The basic mechanism of flotation is the adhesion
between the particle and the bubble; hence, this stage defines the
kinetic characteristics of the flotation process (40).
The adhesion is a result either of the collision between the rising
bubbles with the particles or of the formation of a new bubble on
the surface of a particle, a phenomenon that is relevant to
nucleation. Other ways (as it was shown) include the trapping of
the rising bubbles from the particles, which is due to particle
flocculation and, finally, the confinement of the bubble in an
aggregate as it forms. Certainly, the probability of adhesion
between the particle and the bubble after the collision is greatly
influenced (enhanced) by the addition of surface active substances.
It is possible that all three of the aforementioned mechanisms take
place during the flotation process. It is noted that for a completely
hydrophobic particle, Pa is equal to 1, while for fine particles
(having diameters less than 100 µm) Pd is equal to 0; hence,
detachment is not considered possible, as far as adhesion is
concerned. It is often assumed for fine particles that they have no
inertia and thus follow the streamlines around the bubbles, when
Page 74
the bubbles have diameters lower than 1200 µm. In this case, they
are assumed to behave as nondeformable spheres (41). The flotation
probability is also influenced by a number of other parameters, like
the shape of particles or the adsorbed layer of chemical reagents,
etc., as well as from the multiple interrelation of all the previous
phenomena, so the whole process becomes extremely complicated.
A selectivity index (SI), taking values between 0 and 1.0, was
defined (42) using the probability of adhesion after collision as
follows:
where the superscripts s and w refer to strongly or weakly floating

particles, respectively. The larger the value of SI, the higher the
efficiency of the flotation separation will be. It was suggested that
the selectivity index remains high, until the particle size is reduced to
approximately 40 µm. Below this size limit the SI decreases very
rapidly, with decreasing particle size and eventually passes through
zero for an infinitely small particle (see Fig. 2). However, as the
particle size is reduced, the population of middling particles is also
reduced.
According to hydrodynamics, the particles will follow the
streamlines around the front of the bubble, except insofar as they
cross these lines by sedimentation. Assuming further that the
streamlines come closest to the bubble at its equator, a grazing
streamline was defined as one that passes through the distance of
particle radius at the equator rp and the particle radius r0 at an
infinite distance from the bubble. It then becomes clear that among
the particles located in the path of a bubble, only those originally
situated within the limiting (critical) radius r0, of the central vertical
axis of the bubble, can actually collide with it. Those outside this
area will sweep past the bubble, without having an opportunity for
contact with it. This critical radius depends simply on calculable
hydrodynamic factors, namely, on the rate of rise of the bubble and
the rate of settling of the particle. Those particles distant further than
r0 from the axis of the system will pass around the bubble without
touching it, while those within r0 may bounce off or slide around.
Capture is not necessarily instantaneous, but the period of contact
between particle and bubble is very shortonly a few milliseconds at
most.
The polar coordinate system is often used to describe the trajectory
of a particle moving past a bubble in streamline flow (see Fig. 3),
and the NavierStokes equations were applied to solve numerically
the relative motion (43). The trajectory of a particle is considered to
be determined by the streamline that passes through its center. For
apparent reasons, this becomes less valid as the ratio of the particle
to the bubble diameter increases.
Page 75
Figure 2
Effect of particle size on the selectivity index for bubbles of
100 and 1000 µm in diameter.
Courtesy of Gordon and Breach Science Publishers (42).
Whether or not a colliding particle sticks to a bubble is a question
of which surface forces come into play, when the aqueous film
between the two has thinned sufficiently. It has been shown that if a
solid is hydrophobic (defined as forming a finite contact angle),
very thin water films spontaneously rupture and shrink into a series
of droplets, leaving dry areas between them. These films must be
subject to a negative disjoining pressure. The origin of this pressure
is possibly devoted to electrostatic attraction where the solid/liquid
and gas/liquid interfaces carry electrical double layers of opposite
sign (44). If a positive disjoining pressure develops, thinning will
be retarded. If negative, it will be accelerated.
Consequently, when a bubble is pressed against a hydrophobic
solid under water, if the particle is to be attached to the surface of
the bubble and subsequently collected in the froth phase, the
intervening (disjoining)
Page 76
Figure 3
(a) Polar coordinate system describing the trajectory of a
particle moving past a bubble in streamline flow.
(b) Polar coordinate system used in determining the critical
angle of incidence and probability of adhesion.
(From Ref. 42.)
Page 77
water layer at first thins down, under the influence of hydrostatic
forces until a certain critical thickness is reached. At this point
surface forces of attraction between the two interfaces become
dominant. Inversely, with hyrophilic solids the repulsive forces get
stronger as the film thins.
This critical thickness should be understood as a nominal, average
thickness related to the dimensions of the bubble and the roughness
of the solid, rather than the theoretical thickness at which the
disjoining pressure becomes negative. Values of the critical thickness
for collapse of wetting films at the order of 30 nm have been
determined, in the case of very dilute solutions of a cationic
surfactant on quartz (44).
Then the film rapidly ruptures, in order to establish a three-phase
contact forming a contact angle, and the meniscus runs until the
equilibrium contact angle is reached. The sudden rupture of the
wetting film probably comes about by local fluctuations of the film.
Model experiments that have been carried out on wetting films
between small captive bubbles and smooth polished plates have
shown that hydrophilic solids (such as quartz) retain an elastic water
layer that repels the particle from the bubble, whereas hydrophobic
solids showed a sudden collapse of the aqueous film at a certain
critical thickness.
The probability of particlebubble collision (or collection efficiency)
is defined as the fraction of particles in the path of the bubble that
actually collide with the bubble. Therefore, it is determined by the
ratio Pc of the area A0 (inscribed by the limiting critical diameter d0)
to the area Ab (inscribed by the bubble diameter db), i.e.,
The denominator of the equation should actually be (db + dp), but,
for density reasons in order to have flotation, the basic assumption is
that the bubble diameter is well over that of the particle, i.e., db is
much greater than dp in Eq. (13). As a result, the aggregate formed
by their collision may float (42). The value of d0 is generally
unknown and must be calculated from the mathematical description
of the grazing streamline.
The particle behavior during the examination of the hydrodynamics
of particle collection by bubbles was found to be characterized by
two dimensionless groups (43):
Page 78
where ut is the bubble terminal rise velocity and the subscript p
refers to particle; the other terms were defined already.
In case of fine particles where K < 0.1 and d < 20 µm, the collection
efficiency was rather independent of K but strongly dependent on G,
i.e., for fine particles the inertia effects were rather small in
comparison with the viscous effects. In this regime the collection
efficiency was found to be
Because of its importance on a wide range of disciplines, the study

of fluid flow around a spherical obstruction (such as a raindrop or a
bubble) is a classical hydrodynamical problem with extensive
literature. Table 2 presents some of the obtained equations for the
calculation of Pc in the various flow regimes, following the general
expression (42)
where a and n obtain the respective values shown in the table. A

similar equation was found elsewhere (45), where the constants a
and n were functions of the particle-to-fluid density ratio. They were
dealing with bubbles less than 100 µm and particles less than 20 µm
in diameter.
The importance of this relation is twofold:
1. It suggests a string dependence of Pc on the particle diameter,
which can help explain the difficulty so often observed in the
flotation of fine particles.
2. It also suggests a similar strong dependence of Pc on the gas
bubble diameter.
At a constant gas flow rate, the number of bubbles will vary
inversely as , while the collection efficiency for single bubbles
varies as . The total rate of removal will vary as the product of Pc
and the number of bubbles
Table 2 Values of a and n Used in the Determination
of Probability of Collision Under Different Flow
Conditions
Flow Regime a n Reynolds Number
Stokes 3/2 Re ® 0
Intermediate 2 0 < Re < 300
Potential 3 1 Re ®¥
Source: From Ref. 42, Eq. (17).
Page 79
Table 3 Calculated Collision Efficiencies for Fine
Particles
Authors Collision Efficiencies Proportional
toa
Sutherland (38) ru
Flint & Howarth r2u (K + r2u)
(43)
Woodburn (46) exp (-Î'/r2u)
Reay & Ratcliff r2u
(45)
au is the velocity of the particle relative to the bubble,
and K and Î' are constants.
Source: Ref. 46.
produced and the area swept by each bubble. Hence, it would be

predicted that the rate constant k should vary as:
and obviously k will increase markedly as the bubble radius

increases. Actually, a limited amount of data for bubble size suggests
that for small bubbles in the range of 40 to 80 µm.
Equation (18) has been derived for collection by bubbles rising in a
stationary fluid; it predicts that the rate constant should vary as ,
while the experimental data for such systems suggest that the
exponent should be approximately 1.5 (41). It was indicated that in
flotation cells with conventional impellers, the rate constant varies
linearly with dp. Results of several analyses concerning the
bubbleparticle collision efficiencies for fine particles (calculated
theoretically) were presented by King in Table 3 (46).
Bubbles having sizes of the order of 0.2 mm behave as solid spheres
in pure water. In this case the Stokes equation can be applied for
Reynolds numbers lower than 1 and assuming unhindered movement
of an airsolid agglomerate:
where ut is the terminal rise velocity of the gassolid agglomerate and

da is the diameter of the agglomerate.
Several approximate expressions for the probability of collision have
been developed in recent years, for application especially in the
intermediate Reynolds number range, recognizing the difficulty in
using the Stokes or the potential flow regimes for analyzing flotation
processes, particularly in column flotation (47). It is noted, for the
case of low Reynolds numbers (which is of interest also in effluent
treatment), that Gaudin obtained several years ago an identical form
of equation (48):
Page 80
Sutherland found (also several years ago) a relatively simple

expression, but there was considerable scatter in the experimental
data (38):
It is noted also that the Reynolds number for bubbles is given by the
expression:
where ub is the bubble rise velocity (considering a single bubble)

and µ is the dynamic viscosity of the liquid. The rising velocity
depends on the size of the bubbles and also on the presence or not of
surface active substances in the solution. Smaller bubbles move with
smaller rising velocities than the bigger ones, due to the lower
buoyancy forces. The surface active substances not only decelerate
the rising bubbles but also reduce the adhesion between them due to
the selective adsorption on the surface of the bubbles. This property
is valid only in the case when all the bubbles have similar sizes;
otherwise, the bigger bubble will absorb the smaller one.
Of course, there is certain oversimplification in this theory. Since the
early 1940s, research of Volkova (Ref. 35) on the mechanism of
bubble mineralizationmade by taking microphotographs of the real
flotation process of calcite by oleic acidshowed clearly that it was
often possible to observe more than one particle in the surface of the
same bubble, as well as some particles floating by a group of several
bubbles, depending always on the size of particles and bubbles (35).
A certain conclusion from the aforementioned observations may be
that a specific role is played by the particle/bubble size (49). This is
clearly shown in Figure 4, where with decreasing the bubble
diameters the possibility of collision is increased, reaching almost
10%see Yoon et al. (50). A narrowly sized sample of very
hydrophobic particles was used in these experiments, and quite good
agreement between the predicted values of Pc and those determined
experimentally was observed.
Nevertheless, another conclusion extracted from Figure 2 could be
that the influence of bubble size is rather slight on the probability
attachment; hence, selectivity is being influenced mainly by the
particle sizes and deteriorates with fine particles. This fact might
explain the rather mediocre experimental results obtained with
artificial mixtures of magnesite and dolomite, presented as Table 4.
The grade of magnesite, following electrolytic flotation (i.e., using
bubbles of smaller diameters than in the usual dispersed-air flotation
technique) was improved in comparison with the initial concentra-
Page 81
Figure 4
Probability of collection as a function of bubble size for various particle
collision models.
Courtesy of SME/AIME (50).
Table 4 Selective Electrolytic Flotation Experiments of Mixtures of
Pure Ores with Modifiers
A. Use of Sodium Concentration of collector 40 mg L-1,
Fluorosilicate: modifier
520 mg L-1, pH 11.5
Magnesite 80%-Dolomite RM = 98.4% GM = 84.2% calc.: 90%
20%
60 40 100 67.6 78
40 60 100 48.1 61
B. Sodium Concentration of collector 20 mg L-1,
Hexametaphosphate: modifier
80 mg L-1, pH 8.0
Magnesite 80%-Dolomite RM = 74.9% GM = 85.8% calc.: 89%
20%
60 40 81.7 70.4 76
40 60 82.3 47.6 58
C. Carboxymethyl Concentration of collector 40 mg L-1,
Cellulose modifier
Reverse Flotation: 100 mg L-1, pH 9.5
Magnesite 80%-Dolomite Concentrate RD = 70.1% GD = 23.3%
20%
Tails RM = 42.3 GM = 85
50 50 Concentrate RD = 70.2 GD = 58.7
Tails RM = 50.6 GM = 62.9
Source: Ref. 51.
Page 82
tion, if it is compared with the calculated grade from the tests with
the individual minerals (51). The carbonate particles tested were at
the subsieve size range.
Only a small part of the particles, which collide with bubbles, would
actually become attached by them. Otherwise, no separation would
be possible by flotation. Quantitatively, only those particles that are
hydrophobic enough attach themselves to the bubble surface.
Therefore, it becomes necessary to evaluate also this probability (Pa)
in addition to the previously defined Pc term [Eqs. (20) and (21)].
For the Stokes regime flow conditions (where Re is below 0.1), it
was found that the Pa term can be described, by using the following
expression (52):
where ti is the induction time, i.e., the minimum time that is required
to thin and, subsequently, to rupture the disjoining film between the
bubble and the particle. It is worth noting that while the concept of
contact angle offers a thermodynamic criterion for bubbleparticle
adhesion, the induction time can provide kinetic information
regarding the drainage of the disjoining film between a bubble and
particle. Measurements of this parameter for coal particle flotation
(1100 ms) were reported in the literature (53).
As soon as a particle collides with a bubble, it begins to slide over
the surface of the bubble and resides on it for a finite period of time,
up to the bubble equator. This is generally referred to as the sliding
(or contact) time (Fig. 3). The magnitude of the sliding time is
determined by the velocity of the liquid, which is flowing behind the
rising bubble, and by the size of the bubble. Therefore, the sliding
velocity of a particle is calculated by the tangential velocity of the
streamline at a distance of one particle radius from the bubble
surface.
The induction time can be subdivided into two separate steps:
1. the time required to thin up to a critical thickness
2. the time required for the thin film to rupture spontaneously, in
order to form a contact angle
In cases where bubbleparticle attachment is thermodynamically
possible, adhesion occurs only when the sliding (contact) time is
longer than the induction time. It has been shown that the time it
takes for the thin film to be spontaneously ruptured at the critical
thickness is of the order of 10-9 s. This value is determined by the
thermodynamics of the system. The experimentally determined
induction times reported in the literature are of the order of 10-2 s or
greater. Thus, induction time refers essentially to the
Page 83
time required for the film thinning and drainage process, which is
determined by the hydrodynamics of the system.
Schulze (44) has determined that the time of collision was no greater
than 4 ms, i.e., attachment took 14 ms to occur. The sliding time
without attachment occurring was 3050 ms. Accordingly, in these
experiments the induction time was not constant but varied with the
angle ü from the axis at which contact with the bubble was first
made, i.e., the impingement angle.
The following expression for the induction time was found to give
values about one-half of those experimentally observed (see
Jameson, Ref. 22):
The induction time is certainly valuable in that it gives a physical

explanation of the resistance that must be overcome before bubble
and particle can coalesce. Nevertheless, it has rather limited validity,
because for any particular particlebubble combination there is no
single induction time, as the behavior of the couples will be strongly
influenced by the relative positions of the two bodies before the
capture process begins. The trajectory taken by the same particle will
obviously vary according to its initial position relative to the bubble
in the flow field in question, so the induction time will also vary.
Measurements of induction time were shown to be useful in studying
the effect of coal oxidation on flotation and in determining the
optimum flotation conditions in terms of pH and reagents addition.
In addition, these measurements conducted as a function of
temperature gave the activation energy for bubbleparticle adhesion,
which may be useful in providing information on the role of flotation
reagents (42).
The Stokes number St, which represents a ratio of inertia-to-drag
forces, is useful for determining the tendency of a particle to adjust
itself to changes of fluid flow. The Stokes number is given as a
function of the Reynolds number and of the bubble diameter by the
following equation:
where d is the diameter and r is the density. The tendency of a

particle to follow a streamline around the bubble depends on both
bubble and particle sizes. It can be readily shown that for db of 1 mm
and rp of 2.6 g · cm-3, the inertial forces need not be considered for
particles smaller than approximately 50 µm in diameter. Particles
having large values of St will deviate from the streamlines, while
those having small values of St will follow the streamlines closely
(42).
Page 84
The influence of the roughness of particles also seems to be
important when determining adhesion during collision as does the
influence of the shape of particles when their size is decreasing. It
often has been taken as obvious that during extensive mineral
grinding the produced fines would be like spheres. However, this is
not always the case. Certain ores (for example, natural hematite)
can exhibit increased angularity, when broken down to colloidal
size range, because of the dispersal of fibrous clay minerals (54). In
this case, an angular particle may be expected to cause rupture of
the disjoining liquid film between the bubble and the particle much
more easily than a spherical particle.
In conclusion, although much attention has been paid to the particle
size, the role of bubble size was somehow comparatively
disregarded in the literature according to our knowledge and
certainly deserves more attention. The main explanation for the
former is perhaps the fact that there are practical reasons involved
in the bubble size change. However, various bubble generation
methods (most of them unconventional) utilize fine bubbles.
It is believed that the control of both particle and bubble size is of
paramount interest in affecting the process. The scope of this paper
was particularly limited to fine flotation in minerals and/or
chemical engineering.
References
1. A. K. Khurana and R. Kumar, Studies in bubble formationIII,
Chem. Eng. Sci., 24: 1711 (1969).
2. R. D. Lanauze and I. J. Harris, A note on gas bubble formation,
Chem. Eng. Sci., 29: 1663 (1974).
3. J. F. Davidson and B. O. G. Schuler, (i) Bubble formation at an
orifice in a viscous liquid, (ii) Bubble formation at an orifice in an
inviscid liquid, Trans. Inst. Chem. Eng., 38: (i) 144, (ii) 335 (1960).
4. E. S. Gaddis and A. Vogelpohl, Bubble formation in quiescent
liquids under constant flow conditions, Chem. Eng. Sci., 41: 97
(1986).
5. D. A. Lewis and J. F. Davidson, Bubble sizes produced by shear
and turbulence in a bubble column, Chem. Eng. Sci., 38: 161
(1983).
6. H. Unno and I. Inoe, Size reduction of bubbles by orifice mixer,
Chem. Eng. Sci., 35: 1571 (1980).
7. G. S. Dobby, J. B. Yianatos, and J. A. Finch, Estimation of
bubble diameter in flotation columns from drift flux analysis, Can.
Metall. Q., 27: 85 (1988).
8. J. B. Yianatos and A. R. Levy, ''Estimation of Gas Hold-Up,
Diameter and Apparent Density of Mineralized Bubbles in
Industrial Flotation Columns," Proceedings Int. Colloq.
Developments on Froth Flotation, Vol. 2, Cape Town (1989),
preprint.
Page 85
9. J. B. Yianatos, J. A. Finch, G. S. Dobby, and M. Xu, Bubble size
estimation in a bubble swarm, J. Coll. Interf. Sci., 126: 37 (1988).
10. T. Takahashi, I. Miyahara, and H. Mochizuki, Fundamental
studies of bubble formation in dissolved-air pressure flotation, J.
Chem. Eng. Jpn., 12(4): 275 (1979).
11. A. Vasumishi, M. Fujkuma, and K. Muroyama, Measurement
of behaviour of gas bubbles and gas hold-up in a slurry bubble
column by a dual electroresistivity probe method, J. Chem. Eng.
Jpn., 19: 444 (1986).
12. C. T. O'Connor, E. W. Randall, and C. M. Goodall,
Measurement of the effects of physical and chemical variables on
bubble size, Int. J. Miner. Process., 28: 139 (1990).
13. A. Selecki and R. Wariak, The continuous direct method of
bubble size determination in static foams, J. Colloid Interface Sci.,
102(2): 557 (1984).
14. K. Repanas, A Study of Bubble Generation and
Hydrodynamics in Dissolved-Air Flotation, Ph. D. Thesis, Dept.
Chem. and Process Eng., Univ. Newcastle-upon-Tyne (1992).
15. K. A. Matis, Bubble measurements in electrolytic flotation,
Chim. Chron. (New Ser.), 9: 71 (1980).
16. J. A. Kitchener, "Flotation", Proceedings of Seminar,
Technologically Important Aspects of Interface Science, Imper.
Coll. Sci. and Tech., London, preprint (1983).
17. G. Narsimhan and E. Ruckenstein, Effect of bubble size
distribution on the enrichment and collapse in foams, Langmuir, 2:
494 (1986).
18. R. Jackson, The formation and coalescence of bubbles and
drops in liquids, Chem. Eng., 107 (1964).
19. E. V. Krieg, C. J. King, and H. H. Sephton, Determination of
surface equilibria and mass transfer rates in a concurrent flow
bubble column, AIChE J., 150: 46 (1975).
New York. (1982).
21. F. Sebba, MicrofoamsAn unexploited colloid system, J. Colloid
Interface Sci., 35(4): 643 (1971).
22. K. Ives (ed.), The Scientific Basis of Flotation, M. Nijhoff, The
Hague. (1984).
23. G. L. Collins and G. J. Jameson, Double-layer effects in the
flotation of fine particles, Chem. Eng. Sci., 32: 239 (1977).
24. Y. Sato, Y. Murakami, T. Hirose, H. Yamamoto, and Y. Urgun,
Removal of emulsified oil particles by dissolved air flotation, J.
Chem. Eng. Jpn., 12: 454 (1979).
25. Y. Fukui and S. Yuu, Measurement of the charge on small gas
bubbles, AIChE J., 28: 866 (1982).
26. G. L. Collins, M. Motarjemi, and G. J. Jameson, A method for
measuring the charge on small bubbles, J. Colloid Interface Sci.,
63: 69 (1978).
27. C. Li and P. Somasundaran, Reversal of bubble charge in
multivalent inorganic salt solutionsEffect of magnesium, J. Colloid
Interface Sci., 146(1): 215 (1991).
28. S. Usui and H. Sasaki, Zeta-potential measurements of bubbles
in aqueous solutions, J. Colloid Interface Sci., 65: 36 (1978).
Page 86
29. R. H. Yoon and J. L. Yordan, Zeta-potential measurements on
microbubble generated using various surfactants, J. Colloid
Interface Sci., 113(2): 430 (1986).
30. K. Kubota, S. Hayashi, and M. Inaoka, A convenient
experimental method for measurement of zeta-potentials generating
on the bubbles suspended in aqueous surfactant solutions, J.
Colloid Interface Sci., 95(2): 362 (1983).
31. J. F. Anfruns and J. A. Kitchener, Rate of capture of small
particles in flotation, Trans. IMMSect. C, 89:9 (1977).
32. W. T. Shannon and D. H. Buisson, Dissolved-air flotation in hot
water, Water Res., 14: 759 (1980).
33. A. N. Clark and D. J. Wilson, Foam FlotationTheory and
Applications, Marcel Dekker, New York. (1983).
34. K. A. Matis and P. Mavros, (i) Part I. Recovery of metals by ion
flotation from dilute aqueous solutions, (ii) Part II. Foam/froth
flotation. Removal of particulate matter, Sep. Purif. Methods, 20: 1
and 163 (1991).
35. V. I. Klassen and V. A. Mokrushov, An Introduction to the
Theory of Flotation, Butterworths, London. (1963).
Flotation of magnesite and dolomite by fatty acids, Trans.
IMMSect. C, 98: 99 (1989).
37. I. Wark, Contact angles and flotation research, Principles of
Mineral FlotationThe Wark Symposium (M. H. Jones and J. T.
Woodcock, eds.), Austr. IMM, Parkville, p. 1. (1984).
38. K. L. Sutherland, Kinetics of the flotation process, J. Phys.
Chem., 52: 394 (1948).
39. B. V. Derjaguin, S. S. Dukhin, and N. N. Rulyov, Kinetic
energy of flotation of small particles, Surf. Colloid Sci., 13: 71
(1984).
40. K. Muhle, H. J. Schulze, H. Stechelmesser, and A. Ivanauskas,
Study of the basic steps in flotation and flocculation, Miner.
Process. Extr. Metall., 91: 112 (1982).
41. W. J. Trahar, A rational interpretation of the role of particle size
in flotation, Int. J. Miner. Process., 8: 289 (1981).
42. R. H. Yoon and G. H. Luttrell, The effect of bubble size on fine
particle flotation, Miner. Process. Extr. Metall. Rev., 5: 101 (1989).
43. L. R. Flint and W. L. Howarth, The collision efficiency of small
particles with spherical air bubbles, Chem. Eng. Sci., 26: 1155
(1971).
44. H. J. Schulze, Physico-chemical Elementary Processes in
Flotation, Elsevier, Amsterdam. (1984).
45. D. Reay and G. A. Ratcliff, Removal of fine particles from
water by dispersed-air flotation. Effects of bubble size and particle
size on collection efficiency, Can. J. Chem. Eng., 51: 178 (1973).
46. R. P. King, Flotation of fine particles, Principles of Flotation
(R. P. King, ed.), S. Afr. IMM, Johanessburg, p. 215. (1982).
47. M. E. Weber and D. Paddock, Interceptional and gravitational
collision efficiencies for single collectors at intermediate Reynolds
numbers, J. Colloid Interface Sci., 94: 328 (1983).
48. A. M. Gaudin, Flotation (2nd ed.), McGraw-Hill, New York.
(1957).
flotation. Mineral
Page 87
Processing at a Crossroads (B. A. Wills and R. W. Barley, eds.),
M. Nijhoff, Dordrecht, p. 37. (1986).
50. R. H. Yoon, G. H. Luttrell, and G. T. Adel, Advances in fine
particle flotation, Challenges in Mineral Processing (K. V. S.
Sastry and M. C. Fuerstenau, eds.), SME/AIME, Littleton, p. 487.
(1989).
Chron. (New Ser.), 17: 29 (1988).
52. G. S. Dobby and J. A. Finch, Particle size dependence in
flotation derived from a fundamental model of the capture process,
53. J. L. Yordan and R. H. Yoon, Induction time measurements for
a coal flotation system, Interfacial Phenomena in Biotechnology
and Materials Processing (Y. A. Attia, B. M. Moudgil, and S.
Chander, eds.), Elsevier, Amsterdam, p. 333 (1988).
their flotation, Complex Sulfide Ores (M. J. Jones, ed.), IMM,
London, p. 118 (1980).
Page 89
4
Oxide and Silicate Flotation
Maurice C. Fuerstenau
University of Nevada, Reno, Nevada
I
Origin of Surface Charge
Surface charge can be generated by a number of means, namely,
specific chemical interaction, preferential dissolution of surface
ions, and lattice substitution. In these systems, specific chemical
interaction is, by far, the most important mechanism involved in
surface charge generation. In the case of quartz, following fracture,
H+ and OH- adsorb on anionic and cationic sites, respectively (see
Fig. 1). Surface silicic acid or silanol, SiO H, dissociates to release
hydrogen ion to solution, leaving the surface with a negatively
charged silicate group. If the hydrogen ion activity is increased, the
reaction is reversed, and at high hydrogen activity, the surface
becomes positively charged. Hydrogen ion is said to be potential-
determining because the surface charge and surface potential are
determined by its activity in the bulk phase. At some intermediate
hydrogen ion activity, the surface is uncharged, and this particular
activity presents the point-of-zero-charge (pzc) of the solid. The
pzc of quartz, for example, occurs at pH 1.8. The pzc for any given
oxide or silicate is dependent on the nature of the hydrolysis
reaction for that mineral or its tendency to act as a weak acid. The
specific value of the pzc is determined by the magnitude of the
dissociation constant for surface acidic groups. Points-of-zero
charge
Page 90
Figure 1
Schematic representation of surface charge development
on quartz (1,2).
that arise from this phenomenon are given for various oxide and
silicate minerals in Table 1.
In the development of surface charge, the solid surface acquires a
potential with respect to the solution. The surface charge is
compensated for by an equal charge distribution in the aqueous
phase. The charge in solution together with the charge on the solid
surface is referred to as the electrical double layer, a schematic
representation of which is given in Figure 2. Shown are potential-
determining ions at the surface, a layer of adsorbed counter ions
and counter ions arranged diffusely in the solution surrounding the
solid. Potential-determining ions are those ions that establish the
surface charge. These ions include ions of which the solid is
composed; hydrogen and hydroxyl ions; collector ions that form
insoluble metal-collector salts with ions comprising the mineral
surface; and ions capable of forming complex ions with the surface
species.
Page 91
Table 1 Point-of-Zero-Charge (pzc) of Several Oxide
and Silicate Minerals.a
Mineral pzc Reference
Augite 2.7 Fuerstenau, M. C. et al. (3)
Bentonite <3.0 Iwasaki et al. (4)
Beryl 3.1 Smith and Trevedi (5)
Biotite 0.4 Cases (6)
Cassiterite 4.5 Johansen and Buchanan (7)
Chromite 7.0 Palmer et al. (8)
Chrysocolla 2.0 Palmer et al. (8)
Corundum 9.4 Fuerstenau (9)
Cummingtonite 5.2 Iwasaki et al. (10)
Cuprite 9.5 Parks (11)
Diopside 2.8 Fuerstenau (3)
Garnet 4.4 Cases (6)
Goethite 6.7 Iwasaki (12)
Hematite 6.7 Iwasaki et al. (12)
Kaolinite 3.4 Iwasaki et al. (4)
Magnetite 6.5 Iwasaki et al. (4)
Pyrolusite 7.4 Fuerstenau (13)
Quartz 1.8 Fuerstenau (14)
Rhodonite 2.8 Palmer et al. (8)
Rutile 6.7 Purcell and Sun (15)
Tourmaline 4.0 Rice (16)
Zircon 5.8 Cases (6)
aThese are typical results. Sources of oxide, trace
impurities, etc., cause variations in values.
Counter ions are those ions that have no special affinity for the
surface and are adsorbed by electrostatic attraction. Examples are
collector anions at low concentration adsorbed on a positively
charged surface.
Certain ions possess special affinity for the surface but are not
chemisorbed, and these are termed specifically adsorbed ions.
Examples include polyvalent metal ions in a pH region in which
the metal ions hydrolyze to hydroxy complexes and also longer-
chained homologs of collectors at moderate to high concentration.
At certain concentrations the hydrocarbon chains of the surfactant
ions and/or molecules associate and micellization occurs. See
Figure 3. The concentration at which this phenomenon occurs is
the critical micelle concentration and is referred to as the CMC.
When this phenomenon occurs at the mineral surface, these
aggregates are termed hemimicelles.
Page 92
Figure 2
Schematic representation of electrical
double layer (17).
II
Collectors
A
Cationic
The cationic collectors of importance are the amines. Important
features of these reagents are hydrocarbon chain length, ionization
constant, solubility, and critical micelle concentration. The amines
may be primary, secondary, tertiary, or quaternary (Table 2).
Primary amines are the most extensively used amines in industry.
This reagent ionizes in aqueous solution by protonation. In the case
of dodecylamine (19,20),
Figure 3
Schematic representation of an
anionic micelle.
Page 93
Table 2 Cationic Collectors.
Collector Formula Ion
Primary amine salt RNH3Cl
Secondary amine salt RR'NH2Cl
Tertiary amine salt R(R')2NHClR(R')2NH+
Quaternary ammonium salt R(R')3NCl R(R')3N+
Source: Ref. 18.
In saturated systems,
A speciation diagram for 1 × 10-4 mol/L dodecylamine is presented

in Figure 4.
Solubilities of the undissociated molecule and CMC for selected
primary amines have been established (21).
Figure 4
Logarithmic concentration diagram for 1 × 10-4 mol/L total
dodecylamine (18).
Page 94
B
Anionic
The collectors that fall into this category are the carboxylates (fatty
acids), sulfonates, alkyl sulfates, and certain chelating agents. The
structural formulas of these reagents are given in Table 3.
Fatty acids dissociate into negatively charged ions, RCOO-. As the
hydrocarbon chain length is increased, the solubility of the
undissociated molecule is decreased dramatically. A speciation
diagram for 3 × 10-5 mol/L oleate has been presented by
Somasundaran and Ananthapadmanabhan (23) (see Fig.5). It is
seen that:
1. The maximum activity of the ionomolecular species RCOOH ·
RCOO- occurs at pH 7.8.
2. The activities of the ionic oleate (RCOO-) and the dimer
(RCOO-)2-2 increase as the pH is increased up to pH 7.8 and then
remain constant at higher pH values.
Table 3 Structural Formulas of Sodium Salts of
Various Anionic Collectors.
Collector Structural Formula
Carboxylate
Sulfonate
Alkyl sulfate
Hydroxamate
R represents the hydrocarbon chain

Source: Ref. 22.
Page 95
Figure 5
Species distribution diagram for an aqueous oleate solution (23).
3. Neutral oleic acid (RCOOH) starts to precipitate from this
aqueous solution at pH values less than pH 7.8.
Critical micelle concentrations of various carboxylates, sulfonates,
and alkyl sulfates have also been established (24).
The solubilities of polyvalent metal-collector salts are a function of
the inorganic component of the collector molecule, the
hydrocarbon chain length, and the metal ion. The solubility
products of various metal-collector salts have been established
(25).
III
Adsorption
Depending on concentration and hydrocarbon chain length,
collector species may be present on the surface as individual ions,
as aggregates of ions, termed hemimicelles, or as precipitates.
These data are demonstrated well by that presented by Wakamatsu
and Fuerstenau (26), deBruyn (20), Laskowski (27) an Fuerstenau
et al. (28). As shown in Figure 6, at relatively low concentrations
of dodecyl sulfonate, i.e., less than about 5 × 10-5 molar, individual
sulfonate ions adsorb as counter ions at the solidliquid interface by
electrostatic attraction. This phenomenon occurs in Region I and is
shown schematically in Figure 7(a). As the concentration of
dodecyl sulfonate is increased, the adsorption density of surfactant
becomes sufficiently high that interaction between hydrocarbon
chains of the sulfonate ions occurs through van der Waals forces.
This phenomenon results in
Page 96
Figure 6
Adsorption density and zeta-potential of alumina as a function of
dodecyl sulfonate concentration at pH 7.2 and 10-3 M ionic strength (26).
hemimicelle formation and occurs at concentrations designated as
Region II and Region III. This solidliquid interface under these
conditions is depicted schematically in Figure 7(b). The break
between Region II and Region III occurs at about one-tenth
monolayer coverage.
The adsorption density of sulfonate ions adsorbed in Region I may
be calculated by the following relationship (29):
where Gi is the adsorption density of species i (mol/cm2); ri is the
radius of
Page 97
the adsorbed counter ion (cm); Ci is the concentration in the bulk
solution (mol/cm3); n is valence of the ion; z is the zeta-potential
(volt); R is the gas constant (cal/mol-deg); and F is the Faraday
constant (cal/volt-equiv).
When species are adsorbed specifically at the surface, a force in
addition to the electrostatic force is involved, and Eq. (91) takes on
the following form (30):
where f is the specific adsorption potential. In the case of

hemimicelle formation, van der Waals forces assume importance
and are equal to 0.62 kcal per mole per CH2 group. This is the free
energy decrease that the system experiences when the hydrocarbon
chains are removed from water.
With reference to the data presented in Figure 6, f is zero for
concentrations less than about 5 × 10-5 molar sulfonate. At the pzr,
which is the concentration at which the sign of z is reversed and, in
this case, is about 3 × 10-4 molar, one-tenth monolayer coverage is
present. Under these conditions, f will equal 0.62 kcal/mol CH2
group.
Adsorption data in Figure 6 show that hemimicelles form at the
alumina surface when the bulk equilibrium concentration of
dodecyl sulfonate is approximately 5 × 10-5 mol/L. For
hemimicelle formation to occur under
Figure 7
Schematic representation of the
electrical double layer in the
presence of surface active
organic compounds: (a) adsorption
as single ions at low collector
concentrations
(b) hemimicelle formation at

higher concentrations
(c) co-adsorption of
collector ions and neutral
molecules (17).
Page 98
these conditions, the concentration of dodecyl sulfonate ions in the
Stern plane would have to be about 100 to 200 times greater than
that in the bulk solution. In view of the potentials involved [Eq. (1)
and Fig. 6], such a concentration is realizable.
A
Precipitate Adsorption
The solubility of long-chain amines is limited. For example, the
solubility of dodecylamine is 2 × 10-5 mol/L (20). The pH values at
which RNH2(s) is precipitated at various values of pH is given in
Figure 8. As anticipated, the surface of the amine precipitate is
electrically charged. The isoelectric point (i.e.p.) of dodecylamine
at a concentration of 1 × 10-4 mol/L is pH 10.7 (31). Below the
i.e.p., the precipitate is positively charged, and above this pH, it
will be negatively charged. The formation and adsorption of the
amine precipitate on a quartz surface is well demonstrated in
Figure 9. The rapid rise in zeta-potential corresponds to the
precipitation of dodecylamine and its adsorption on the quartz
surface.
Figure 8
Thermodynamic equilibrium diagram showing the CMC
(30 °C) and the lines of critical pH of
precipitation and solubility limit at 25 °C (31).
Page 99
Figure 9
Zeta-potential of quartz at 4 × 10-5 M
dodecylamine acetate as a function of pH and
dodecylamine precipitate alone (32, 33).
1
Hydroxy Complexes/Metal Hydroxides
The role of metal ion hydrolysis is substantial in oxide and silicate
flotation. In this view, it will be well to review these concepts.
When ferric sulfate, for example, is added to water, both ferric and
sulfate ions will be hydrated. Due to its small size and high charge,
the extent of hydration of ferric ion is substantial. Six waters are
coordinated with ferric ion arranged in octahedral fashion.
The attraction of ferric ion for the electrons of oxygen of a
coordinated water molecule is relatively great with the result that
one of the oxygenhydrogen bonds in the water molecule is
weakened. When the hydrogen ion activity in solution is decreased,
the hydrogen ion dissociates from the water molecule to
compensate for this reduced activity. When this occurs, the first
hydroxy complex of ferric ion, FeOH(H2O)5+2, is formed. Further
decrease in the hydrogen ion activity in solution results in
additional hydrolysis of ferric ion. The complete hydrolysis scheme
of ferric ion, excluding ferrite and the dimer, is as follows:
Page 100
A speciation diagram for 1 × 10-4 mol/L ferric is presented as a function of pH
in Figure 10. Note that FeOH+2 becomes significant in concentration at about
pH 2.5.
As the size of the metal ion is increased and its electrical charge is reduced, the
pH at which hydrolysis occurs is increased.
The adsorption potential of hydroxy species on solid surfaces is large. In fact,
these species adsorb on positively charged substrates (14). In the
quartzcalcium system, limited adsorption of hydrated calcium ion is observed
above the pzc (pH 1.8) until about pH 11 above which pH Ca+2 species are
adsorbed extensively (Fig. 11).
Comparison of these data with the speciation diagram for 1 × 10-3
Figure 10
Logarithmic concentration diagram for 1 × 10-4 M Fe3+. Equilibrium data from Ref. 34.
Page 101
Figure 11
Adsorption of calcium species on quartz as a function of pH from
solutions 100 ppm in Ca2+ (35).
mol/L CaCl2 in Figure 12 shows that CaOH+ becomes significant

in concentration at about pH 11.
These species also change the sign of the zeta-potential of oxides
and silicates (see Figs. 13 and 14). The pH at which the reversal of
sign from negative to positive occurs is the same pH that PbOH+
becomes significant in concentration.
The metal hydroxide also assumes a significant role in these
systems. This is shown clearly in the talcchromium system (Fig.
15). First, the zeta-potential can be noted to change sign from
negative to positive at the pH at which CrOH+2 becomes
significant in concentration. The pH at which precipitation of
Cr(OH)3(s) occurs in this system has been noted, and the excellent
correlation of the zeta-potentials of talc in the presence of 1 × 10-3
mol/L CrCl3 with those of Cr(OH)3 precipitated externally can be
noted.
The mechanism of hydroxy complex adsorption has been
postulated to occur through the entropy increase that the system
experiences when water, adsorbed on the solid surface, replaces
coordinated water molecules from the hydration sheath of the
hydroxy complex (36). Once adsorbed, further
Page 102
Figure 12
Logarithmic concentration diagram for 1 × 10-3 M Ca2+.
Equilibrium data from Ref. 34.
hydrolysis of the adsorbed hydroxy complex to metal hydroxide
can occur with increased hydroxyl ion activity.
IV
Adsorption
A
Flotation by Physical Adsorption
The adsorption of many collectors occurs by electrostatic
interaction with oxide and silicate surfaces. When these collectors
are used, knowledge of the pzc's of the minerals in question must
be known.
Three mineral systemsthe hematite, alumina, and quartz
systemshave received considerable attention. The dependence of
hematite flotation
Page 103
on electrostatic phenomena when amine, sulfonate, and sulfate
collectors are used is shown clearly in Figure 16. The point-of-
zero-charge for hematite was measured at pH 6.7. Below the pzc
the surface is positively charged; negatively charged dodecyl
sulfonate ions are adsorbed in this region, and complete flotation is
effected. Above the pzc the surface is negatively charged, and
sulfonate ions are repelled from the surface. On the other hand,
dodecylammonium ions, which are positively charged, are
adsorbed on the negatively charged surface.
Similar phenomena can be noted when corundum (Al2O3) is
floated with dodecylamine and dodecyl sulfate (Fig. 17). The pzc
of alumina is pH 9.1.
In the case of quartz, the pzc is pH 1.8. Anionic collectors, such as
sulfonates and alkyl sulfates, are not adsorbed in sufficient amount
below the pzc to result in flotation, since these species must
compete with anions present (from pH adjustment) in
concentrations greater than 1 × 10-2 mol/L. Above the pzc,
aminium ions are adsorbed. Similar results have been reported by
Cases for other silicates (6).
As discussed previously, the surfaces of oxide and silicate minerals
are extensively hydrated due principally to hydrogen bonding of
water molecules to the surface. As a result, the presence of
hemimicelles of collector appears to be necessary to render these
surfaces sufficiently hydrophobic
Figure 13
Zeta-potential of quartz as a function of pH in the presence of
1 × 10-4 M PbCl2 (14).
Page 104
for flotation to occur. With reference to Figures 18 and 19, the
concentrations at which the zetapotential of quartz changes sign
from negative to positive and the concentrations at which a rapid
rise in flotation recovery of quartz occurs provide evidence of this
phenomenon. Verification of this premise has been provided by
Fuerstenau et al. (28), who have shown that when the logarithm of
collector concentration required for a rapid rise in flotation
recovery is plotted as a function of the number of carbon atoms in
the hydrocarbon chain, a straight line with a slope that corresponds
to a specific adsorption potential f of -1.1 kT is obtained, which is
indicative of hemimicelle formation. This value of -1.1 kT
corresponds to -0.62 kcal/mol-CH2 group, the free energy decrease
associated with hydrocarbon chain removal from solution.
Similar to the case of dodecyl sulfonate and dodecyl sulfate
flotation of
Figure 14
Logarithmic concentration diagram for 1 × 10-4 M Pb2+.
(Equilibrium data from Ref. 37.)
Page 105
Figure 15
Zeta-
potential of talc as a function of pH in the absence and presence of Cr(III) together with the
zeta-potential of precipitated Cr(OH)3 (38).
hematite, the adsorption of relatively short-chain carboxylates also occurs by
electrostatic interaction with oxide and silicate minerals. For example, when
chromite with a pzc of pH 7 is floated with octanoic and decanoic acids at
concentration levels of 10-2 molar, flotation is possible only below the pzc (41).
When this same mineral is floated with 1 × 10-4 molar laurate, flotation also
occurs only below the pzc. However, with a laurate addition of 1 × 10-3 mol,
chemisorption of the collector occurs, since flotation above the pzc is possible
(see Fig. 20).
B
Modulation of Flotation
1
Effect of Basicity
Primary, secondary, and tertiary amines are weak bases, and the concentrations
of the dissolved species that are present in solution are related to the pH of the
solution. As shown in the logarithmic concentration diagram presented in Figure
5, dodecylamine is present predominantly as the aminium ion RNH3+ below
about pH 10. The neutral species RNH2(aq) becomes significant in concentration
above this same value in pH.
The correlation between contact angle, adsorption density, zeta-potential, and
flotation recovery for the dodecylaminequartz system is
Page 106
Figure 16
Flotation recovery of geothite as a function of pH with 1 × 10-3 M
additions of collector compared with surface properties of the same
mineral (12).
given in Figure 21. The pH of maximum recovery, contact angle,
and zeta-potential corresponds to the pH at which precipitation and
adsorption of dodecylamine occurs (see Figs. 9 and 10). The
depression observed above about pH 12 is probably due to the
negative charge on the amine precipitate.
2
Activation with Fluoride
Manser (42) has obtained flotation data with minerals contained in
the five groups of silicates,
1. orthosilicates (andalusite, beryl, tourmaline)
2. pyroxenes (augite, diopside, spodumene)
Page 107
Figure 17
Flotation recovery of corundum as a function of pH 4 × 10-5 M
additions of collector (39).
3. amphiboles (hornblende, tremolite, actinolite)
4. sheet silicates (muscovite, biotite, chlorite)
5. framework silicates (quartz, feldspar, nepheline)
It was found that orthosilicates are sensitive to fluoride addition,
whereas the pyroxenes and amphiboles are scarcely affected by
fluoride. Sheet silicates are activated by fluoride, while framework
silicates are activated to a lesser extent with the exception of
quartz.
Quartz and feldspar have been floated by Smith (43) with
dodecylamine in the absence and presence of sodium fluoride.
These results are presented in Figure 22. As shown, maximum
contact angle on microcline and, hence, maximum flotation are
obtained at approximately pH 2 in the presence of 10-2 mol
fluoride.
A number of conflicting theories have been presented as to the
possible role of fluoride ion under these conditions (4446). Smith
and co-workers have suggested the following (46,47):
1. HF attack of the surface silicic acid to form SiF6-2
2. Fluosilicate ion adsorption on aluminum sites:
Page 108
Figure 18
Zeta-potential of quartz as a function of concentration of ammonium acetate
and aminium acetates with different hydrocarbon chain length at neutral pH (40).
3. Adsorption of aminium ions on the alumino-fluosilicate sites:
3
Depression with Inorganic Ions
Onoda and Fuerstenau (48) and Somasundaran (49) have studied in
detail the depressing effect that inorganic cations exhibit in cationic
flotation systems. The flotation of quartz with 1.5 × 10-4 mol/L
dodecylamine as a function of potassium nitrate concentration and pH
is presented in Figure 23. As the pH is reduced, the surface charge is
reduced, and the concentration of K+ needed to depress the system is
reduced. Under these conditions K+ is able to compete effectively with
RNH3+ for the surface.
C
Flotation by Chemisorption
Chemisorption of high molecular weight collectors on oxides and
silicates occurs under conditions where metal ions comprising the
mineral have hy-
Page 109
Figure 19
The effect of hydrocarbon chain length on the relative flotation response of
quartz in the presence of aminium acetate solutions at neutral pH (28).
Figure 20
Flotation recovery of chromite as a function of pH in the presence
of various additions of lauric acid (8).
Page 110
Figure 21
Correlation of adsorption density, contact angle, and
zeta-potential with flotation of quartz with 4 × 10-5 M
dodecylammonium acetate additions (32).
drolyzed to hydroxy complexes and metal hydroxides have formed.
When pyrolusite is floated with oleate as collector, for example,
two maxima in flotation response are present (Fig. 24). The pzc of
this mineral is pH 7.4. From electrophoretic (13) and activation
(22) work in other systems, the flotation response observed at pH
8.5 is intimately involved with the species, MnOH+ and Mn(OH)2.
This may be seen by referring to Figure 25, which shows the zeta-
potential of rhodonite (manganous silicate) in the absence and
presence of manganous sulfate, and also to Figure 26, the
speciation diagram of manganous ion. As can be noted, zeta-
reversal and flotation occur in the pH range in which MnOH+ is
present maximally and in which Mn(OH)2(s) is formed. Further,
this is the same pH range in which
Page 111
rhodonite responds to flotation with octyl hydroxamate (8) and
activation of quartz occurs in the presence of manganous salts (Fig.
25).
Pyrolusite is manganic oxide, and formation of the hydroxy
complex of manganous ion could occur only after manganic
reduction by reaction with water. In practice dissolution of
pyrolusite has been enhanced by the addition of sulfur dioxide to
ore systems (50).
Flotation observed at pH 4 can probably be attributed to physical
adsorption of an oleate species. Unfortunately, pyrolusite is opaque
to light in the infrared region, and identification of species present
on the surface cannot be made.
The role that hydroxy complexes and metal hydroxides assumes in
chemisorption in anionic flotation systems is also clearly illustrated
in the chromite and augite systems. The theoretical composition of
chromite is FeO · Cr2O3; Fe3+, Al3+, Fe2+ and Mg2+ are common
impurities. As would
Figure 22
Contact angle on microcline and quartz as a function
of pH with 4 × 10-5 M dodecylamine in theabsence
and presence of fluoride (43).
Page 112
Figure 23
Flotation recovery of quartz as a function of KNO3 concentration
and pH with 1.5 × 10-4 M dodecylamine (49).
Figure 24
Flotation recovery of pyrolusite as a function of pH with 1 × 10-4 M
oleate (13).
Page 113
Figure 25
Zeta potential of rhodonite as a function of pH in the absence and
presence of Mn2+ (8).
be expected, composition exhibits a direct effect on the electrical
properties of this mineral's surface. In fact, pzc's ranging from pH
4.4 to 9.6 have been measured for chromites of various
compositions (41). The composition of the chromite whose
flotation response is given in Figures 20 and 27 is 41.72% Cr+13,
7.05% Fe2+, 3.73% Fe3+, 8.94% Mg2+, 4.21% Al3+, 0.11% Ca2+,
0.31% Si4+, and 0.07% Ti4+.
The response in the vicinity of pH 11 is attributed to
MgOH+/Mg(OH)2 and that in the vicinity of pH 8 is attributed, in
part, to FeOH+/Fe(OH)2. These are the pH values at which zeta-
reversal occurs and those at which hydroxy complexes are present
in significant concentration at which hydroxides form (8) (see Figs.
28 and 29).
Dixit and Biswas (53) have shown that non-iron-bearing minerals
are also floated at approximately pH 8 as shown in Figure 30.
Furthermore, other investigators (14,54) have shown that hematite
is floated maximally at this same pH with oleic acid. Apparently,
phenomena in addition to hydroxy complex/metal hydroxide
formation are involved in the flotation response observed at pH 8.
For example, Kulkarni and Somasundaran (55) have shown for
hematite flotation with oleate that collector adsorption at the
air/water interface and subsequent lowering of surface tension may
also be involved in the maximal flotation observed at this pH. The
dynamic surface tension in such systems seems to be related to acid
soap formation.
Page 114
The flotation of chromite observed at about pH 3.5 can probably be
attributed to physical adsorption of an oleate species, either oleate
ion or colloidal oleic acid. Since this chromite contains Al3+, one
would suspect that hydroxy complexes of aluminum are
responsible for flotation at this pH. This is apparently not the case,
however. Similarly, the hydroxy complexes of chromium are
apparently also not involved in chromite flotation. The pH range in
which CrOH2+ is most abundant is the pH range in which
depression is observed (41). Aluminum and chromium are
coordinated octahedrally with oxygen, whereas the divalent cations
are coordinated tetrahedrally with oxygen. As a result, the divalent
cations would be expected to dissolve much more readily than the
trivalent cations.
Figure 26
Logarithmic concentration diagram for 1 × 10-4 M Mn2+.
Equilibrium data from Ref. 34.
Page 115
Figure 27
Flotation recovery of chromite as a function of pH and oleate
concentration (8).
Responses similar to these were also observed by Sobieraj and
Laskowski (56).
With augite, the response observed at pH 3 can probably be
attributed to physical adsorption of an oleate species (see Fig. 31).
Since the pzc of this mineral is pH 2.7, the response of this mineral
is much more limited than that observed for chromite and
pyrolusite under the same conditions. The pzc's of these samples of
chromite and pyrolusite are pH 7.0 and pH 7.4, respectively. The
responses observed at around pH 8 may be attributed to
FeOH+/Fe(OH)2 and that from pH 10 to pH 12 to
MgOH+/Mg(OH)2, and CaOH+/Ca(OH)2 (22). The distribution
diagrams for calcium, ferrous ion, and magnesium are presented in
Figures 12, 28, and 29.
Chrysocolla responds to flotation with octyl hydroxamate in the pH
range in which CuOH+ is present in significant concentration and
Cu(OH)2(s) is formed (57). Optimum flotation occurs at
approximately pH 6.3, which is also the pH at which the zeta-
potential becomes positive in the presence of cupric salts (see Figs.
32 and 33).
As opposed to pyrolusite, chrysocolla responds well to infrared
absorption analysis (8). Major absorption bands of cupric octyl
hydroxamate occur at 925, 1095, 1380, 1450, and 1535 cm-1.
Identical absorption bands at 1380, 1450, and 1535 cm-1 are noted
on chrysocolla after contact with 1
Page 116
× 10-3 mol octyl hydroxamate at pH 6.3. In fact, the multilayers of
cupric hydroxamate are visibly apparent on the surface. The
chrysocolla changes color from its natural blue-green to a vivid
green color during the conditioning step prior to flotation.
Multilayers of metal hydroxamate have been observed in other
systems, notably the hematitehydroxamate system (14). In the case
of pyrolusite, multilayer adsorption of octyl hydroxamate was
observed, but whether manganous hydroxamate was present as
multilayers was not established (58).
D
Quartz Activation
In the case of quartz, since its solubility is very limited, and since
the only cation comprising the mineral is silicon, anionic flotation
is obtained only
Figure 28
Logarithmic concentration diagram for 1 × 10-4 M Mg2+.
(Equilibrium data from Ref. 34.)
Page 117
Figure 29
Logarithmic concentration diagram for 1 × 10-4 M Fe2+.
(Equilibrium data from Refs. 51 and 52.)
after metal ions are added to or are contained in the system in a pH
range in which in which hydrolysis to first hydroxy complexes and
metal hydroxides occurs. Quartz activation can be achieved with
most anionic collectors, including xanthate, providing certain
conditions are satisfied (22).
In the case of sulfonate flotation, edges outlining minimum values
of pH at which flotation is obtained in the presence of 1 × 10-4
mol/L sulfonate and 1 × 10-4 mol/L of various metal ions are
presented in Figure 34. Comparison of these data with the
distribution diagrams of these metal ions (Figs. 10, 12, 14, 26, and
28) should be made. The range of pH in which flotation is obtained
in the presence of some of these activators is shown in Figure 35.
The absence of flotation below the lower flotation edge is
attributed to an insufficient amount of hydroxy complex/metal
hydroxide required for flotation. The depression observed above
the maximum flotation edge is attributed to the negative surface
charge of the metal
Page 118
Figure 30
Effect of pH on flotation recovery of the beach sand
minerals with sodium oleate (53).
hydroxide precipitate and to the stability of the metal hydroxide
relative to the metal-collector salt. The pH ranges in which
flotation is effected in the presence of a number of metal salts are
listed in Table 4. These ranges are those in which flotation recovery
greater than 90% is obtained. Regions of flotation response of
quartz similar to those in Table 4 have been presented by Kraeber
and Boppel (59).
Similar behavior has been found for carboxylates. Flotation
response in the presence of various additions of calcium chloride
and lauric acid at pH 11.5 is presented in Figure 36. Arrows
indicating the activity of laurate at which calcium laurate
precipitates in each of the systems is shown. As can be noted,
flotation occurs only after precipitation of calcium laurate has
occurred. Similar results with oleate as collector have been
reported (61).
With shorter-chain homologs of fatty acids, e.g., nonanoic acid,
complete flotation of quartz is effected prior to bulk precipitation
of the calcium salt, but high concentrations of collector are
required (60). At these
Page 119
Figure 31
Flotation recovery of augite as a function of pH and oleate
concentration (8).
Figure 32
Flotation recovery of chrysocolla as a function of pH with various
concentrations of octyl hydroxamate (57).
Page 120
Figure 33
Zeta-potential of chrysocolla as a function of pH in the absence and
presence of Cu2+ (8).
Figure 34
Minimum flotation edges of quartz as a function of pH. Conditions:
1 × 10-4 M sulfonate; 1 × 10-4 M metal ion (22).
Page 121
Figure 35
Flotation recovery of quartz as a function of pH. Conditions:
1 × 10-4 M sulfonate; 1 × 10-4 M metal ion (22).
concentrations it seems likely that hemimicelles would have
formed at the interface.
The critical nature of the addition of collector relative to that of
activator has been shown by Gaudin and Rizo-Patron (62), Cooke
and Digre (63), Fuerstenau and Atak (64), and Fuerstenau and
Elgillani (65). In those sys-
Table 4 Ranges in pH in Which
90% Recovery of Quartz Is
Obtained in the Presence of
Various Activators.a
Metal Salt pH Range
Fe(III) 2.9 3.8
Al 3.8 8.4
Pb 6.512.0
Mn(II) 8.5 9.4
Mg 10.911.7
Ca 12.0 and greater
aConditions: 1 × 10-4 M
sulfonate, 1 × 10-4 M metal
salt.
Source: Ref. 22.
Page 122
Figure 36
Flotation recovery of quartz as a function of lauric acid and calcium
chloride additions at pH 11.5 (60).
tems in which precipitation of metal-collector has occurred at
constant collector addition, increasing the metal ion addition by an
order of magnitude reduces the minimum pH at which flotation is
possible by one unit (64).
On the other hand, at constant metal ion addition, increasing the
collector addition reduces the pH at which flotation is possible until
a critical level of addition is reached, at which point higher values
of pH are required for flotation.
As these facts indicate, the phenomena occurring in these systems
are complex, and they are not well understood at the present time.
Further study is needed to delineate the mechanisms involved.
Acknowledgments
The author wishes to acknowledge the permission granted by the
Society of Mining Engineers of AIME to publish material from the
book, Chemistry
Page 123
of Flotation, written by M. C. Fuerstenau, J. D. Miller and M. C.
Kuhn and to publish selected figures and tables from the books,
Flotation, A. M. Gaudin Memorial Volume, edited by M. C.
Fuerstenau and Challenges in Mineral Processing, edited by K. V.
S. Sastry and M. C. Fuerstenau. The permission to publish a figure
from the International Journal of Mineral Processing is also
gratefully acknowledged.
References
1. J. A. Yopps and D. W. Fuerstenau, The zero point of charge of
alpha-alumina, J. Colloid Sci., 19: 61 (1964).
2. A. C. Zettlemoyer, Hydrophobic surfaces, J. Colloid Interface
Sci., 28: 143 (1968).
3. M. C. Fuerstenau, B. R. Palmer, and G. B. Guttierez,
Mechanisms of flotation of selected iron-bearing silicates, Trans.
AIME, 262: 234 (1977).
4. (a) I. Iwasaki, S. R. B. Cooke, and Y. S. Kim, Some surface
properties and flotation characteristics of magnetite, Trans. AIME,
223: 113 (1962).
(b)I. Iwasaki, S. R. B. Cooke, D. H. Harraway, and H. S. Choi,
Iron wash ore slimesSome mineralogical and flotation
characteristics, Trans. AIME, 223: 97 (1962).
5. R. W. Smith and N. Trivedi, Variation of point of zero charge of
oxide minerals as a function of ageing time in water, Trans. AIME,
256: 69 (1974).
6. J. M. Cases, On the normal interaction between adsorbed species
and adsorbing surface, Trans. AIME, 247: 123 (1970).
7. P. G. Johansen and A. S. Buchanan, An application of the
microelectrophoresis method to the study of the surface properties
of insoluble oxides, Aust. J. Chem., 10: 398 (1957).
8. (a) B. R. Palmer, G. B. Gutierrez, and M. C. Fuerstenau,
Mechanisms involved in the flotation of oxides and silicates with
anionic collectors: Part 1, Trans. AIME, 258: 257 (1975).
(b) B. R. Palmer, M. C. Fuerstenau, and F. F. Aplan, Part 2,
Trans. AIME, 258: 261 (1975).
9. D. W. Fuerstenau and J. H. Modi, Streaming potential of
corundum in aqueous organic electrolyte solutions, J. Electrochem.
Soc., 106: 336 (1959).
10. I. Iwasaki, S. R. B. Cooke, and H. S. Choi, Flotation of
cummingtonite, Trans. AIME, 220: 394 (1961).
11. G. A. Parks, The isoelectric points of solid oxides, solid
hydroxides and aqueous hydroxo complex systems, Chem. Rev.,
65: 177 (1965).
12. I. Iwasaki, S. R. B. Cooke, and A. F. Colombo, Flotation
characteristics of goethite, U.S. Bur. Mines, RI 5593 (1960).
13. M. C. Fuerstenau and D. A. Rice, Flotation characteristics of
pyrolusite, Trans. AIME, 241: 453 (1968).
14. (a) M. C. Fuerstenau, R. W. Harper, and J. D. Miller,
Hydroxamate vs. fatty acid flotation of iron oxide, Trans. AIME,
247: 69 (1960).
Page 124
(b) M. C. Fuerstenau, D. A. Elgillani, and J. D. Miller,
Adsorption mechanisms in nonmetallic activation systems,
Trans. AIME, 247: 11 (1970).
15. G. Purcell and S. G. Sun, Significance of double bonds in fatty
acid flotation, Trans. AIME, 226: 6 (1963).
16. D A. Rice, Some Mechanisms of Collector Adsorption on
Tournaline as Related to Other Silicates, Doctor of Science Thesis,
Colorado School of Mines (1968).
17. F. F. Aplan and D. W. Fuerstenau, Principles of nonmetallic
mineral flotation, Froth Flotation (D. W. Fuerstenau, ed.), AIME,
New York, pp. 170, 173. (1962).
18. R. W. Smith and A. Akhtar, Cationic flotation of oxides and
silicates, Flotation (M. C. Fuerstenau, ed.), AIME, New York, p.
87. (1976).
19. A. W. Ralston, Fatty Acids and Their Derivatives, Wiley, New
York, p. 671. (1948).
20. P. L. deBruyn, Flotation of quartz by cationic collectors, Trans.
AIME, 202: 291 (1955).
21. W. D. Harkins, The Physical Chemistry of Surface Films,
Reinhold, New York. (1952).
22. M. C. Fuerstenau and B. R. Palmer, Anionic flotation of oxides
and silicates, Flotation (M. C. Fuerstenau, ed.), AIME, New York,
p. 148. (1976).
23. P. Somasundaran and K. P. Ananthapadmanabhan, Solution
chemistry of surfactants and the role of it in adsorption and froth
flotation of mineral-water systems, Solution Chemistry of
Surfactants (K. L. Mittal, ed.), Plenum, New York, p. 777. (1979).
24. P. Mukerjee and K. J. Mysels, Critical micelle concentrations of
aqueous surfactant systems. Natl. Bur. Standards, NSRDS-NBS 36
(1971).
25. C. DuRietz, Chemisorption of collectors in surface chemistry,
Proceedings XI Intl. Min. Proces. Congr., Calgary, p. 376. (1975).
26. T. Wakamatsu and D. W. Fuerstenau, Effect of alkyl sulfonates
on the wettability of alumina, Trans. AIME, 254: 123 (1973).
27. J. S. Laskowski, The colloid chemistry and flotation properties
of primary aliphatic amines. Challenges in Mineral Processing (K.
V. S. Sastry and M. C. Fuerstenau, ed.), SME, Littleton, p. 15.
(1989).
28. D. W. Fuerstanau, T. W. Healy, and P. Somasundaran, The role
of hydrocarbon chain of alkyl collectors in flotation, Trans. AIME,
229: 321 (1964).
29. H. R. Kruyt, ed., Colloid Science, Vol. 1, Elsevier, Amsterdam.
(1952).
30. D. Stigter and J. Th. G. Overbeek, Proceedings 2nd Intl. Congr.
Surface Activity, Butterworths, London, p. 311. (1958).
31. J. S. Laskowski, R. M. Vurdela, and Q. Liu, The colloid
chemistry of weak-electrolyte collector flotation, Proceedings XVI
Intl. Min. Proces. Congr. (E. Forssberg, ed.), Elsevier, Amsterdam,
p. 703. (1988).
32. D. W. Fuerstenau, Correlation of contact angles, adsorption
density, zeta potentials and flotation rate, Trans. AIME, 208: 1365
(1957).
33. J. S. Laskowski, Intl. Conf. Copper 87Mineral Processing and
Process Control, Vol. 2 (A. L. Mular, G. Gonzalez, and C.
Barahona, eds.), Univ. Chile, p. 137 (1988).
Page 125
34. J. N. Butler, Ionic Equilibrium, Addison-Wesley, Reading, MA,
pp. 287 and 364. (1964).
35. S. W. Clark and S. R. B. Cooke, Adsorption of calcium,
magnesium and sodium ion by quartz, Trans. AIME, 241: 334
(1968).
36. R. O. James and T. W. Healy, Adsorption of hydrolyzable metal
ions at the oxide-water interface, J. Colloid Interface Sci., 40: 53
(1972).
37. W. M. Latimer, Oxidation Potentials, Prentice-Hall, Englewood
Cliffs, N.J., pp. 128, 152, 169. (1952).
38. M. C. Fuerstenau, A. Valdivieso, and D. W. Fuerstenau, Role of
hydrolyzed cations in the natural hydrophobicity of talc. Int. J.
Miner. Process., 23: 161 (1988).
39. H. J. Modi and D. W. Fuerstenau, Flotation of corundum: An
electrochemical interpretation. Trans. AIME, 217: 381 (1960).
40. D. W. Fuerstenau, Streaming potential studies on quartz in
solutions of aminium acetates in relation to the formation of
hemimicelles at the quartz-solution interface, J.Phys. Chem., 60:
981 (1956).
41. B. R. Palmer, Adsorption of Surfactants on Chromite and
Chrysocolla, Ph.D. Thesis, Univ. Utah (1972).
42. R. M. Manser, Handbook of Silicate Flotation, Warren Spring
Laboratory, Stevenage, UK. (1975).
43. R. W. Smith, Effect of fluoride addition on contact angle in the
system microcline-dodecylamine solution-nitrogen, Proceedings S.
Dakota Acad. Sci., vol. 42, p. 60. (1968).
44. R. W. Smith, Coadsorption of dodecylamine ion and molecule
on quartz, Trans. AIME, 226: 427 (1963).
45. L. J. Warren and J. A. Kitchener, Role of fluoride in the
flotation of feldspar: adsorption on quartz, corundum and
potassium feldspar, Trans. IMM, 81: C137 (1972).
46. R. W. Smith and T. J. Smolik, Infrared and X-ray diffraction
study of the activation of beryl and feldspars by fluorides in
cationic collector systems, Trans. AIME, 232: 196 (1965).
47. D. B. Sullin and R. W. Smith, Hallimond tube investigation of
fluoride activation of beryl and feldspars in cationic collector
systems, Trans. IMM, 75: C333 (1966).
48. G. V. Onoda and D. W. Fuerstenau, Amine flotation of quartz in
the presence of inorganic electrolutes, Proceedings VII Intl. Min.
Proces. Congr., New York, p. 301. (1964).
49. P. Somasundaran, Cationic depression of amino flotation of
quartz, Trans. AIME, 256: 64 (1974).
50. S. J. McCarroll, Upgrading manganese ores. Min. Eng., 6: 289
(1959).
51. D. D. Wagman, Natl. Bur. Standards, Technical Note 270-4, p.
74. (1969).
52. D. Langmuir, The Gibbs free energies of substances in the
system, U.S. Geological Survey, Professional Paper 650-B, p. 181.
(1969)
53. S. G. Dixit and A. K. Biswas, pH-dependence of the flotation
and adsorption properties of some beach sand minerals, Trans.
AIME, 244: 173 (1969).
54. A. S. Peck, L. H. Raby, and M. E. Wadsworth, An infrared
study of the
Page 126
flotation of hematite with oleic acid and sodium oleate, Trans.
AIME, 235: 301 (1966).
55. R. D. Kulkarni and P. Somasundaran, Kinetics of oleate
adsorption at the liquid/air interface and its role in hematite
flotation, AlChE Symp. Ser. Vol. 71, p. 124 (1975).
56. S. Sobieraj and J. Laskowski, Flotation of chromite: Early
research and recent trends, and surface properties of spinel
minerals, Trans. IMM, 82: C207 (1973).
57. H. D. Peterson, M. C. Fuerstenau, R. S. Rickard, and J. D.
Miller, Chrysocolla flotation by the formation of insoluble surface
chelates, Trans. AIME, 232: 389 (1965).
58. R. Natarjan and D. W. Fuerstenau, Int. J. Miner. Process., 11:
139 (1983).
59. L. Kraeber and A. Boppel, Metall. Erz., 31: 412 (1934).
60. M. C. Fuerstenau and W. F. Cummins, The role of basic
aqueous complexes in anionic flotation of quartz, Trans. AIME,
238: 196 (1967).
61. J. B Clemmer, B. H. Clemmons, C. Rampacek, M. F., Williams,
and R. J. Stacey, Beneficiation of iron ores by flotation, U.S.
Bur.Mines, RI 3799 (1945).
62. A. M. Gaudin and A Rizo-Patron, The mechanism of activation
in flotation, Trans. AIME, 153: 462 (1942).
63. S. R. B. Cooke and M. Digre, Studies on the activation of
quartz with calcium ion, Trans. AIME, 184: 299 (1949).
64. (a) M. C. Fuerstenau and S. Atak, Lead activation in sulfonate
flotation of quartz, Trans. AIME, 232: 24 (1965).
(b) M. C. Fuerstenau, R. E. Pray, and J. D. Miller, Metal ion
activations in xanthate flotation of quartz, Trans. AIME, 232:
359 (1965).
65. M. C. Fuerstenau and D. A. Elgillani, Calcium activation in
sulfonate and oleate flotation of quartz, Trans. AIME, 235: 405
(1966).
Page 127
5
Flotation of Iron Sulfide Minerals:
Electrokinetic Aspects
K. A. Kydros and K. A. Matis
I
Introduction
The main source of a great variety of metals universally is the
mixed sulfide ores, found in North America, Australia, ex-USSR,
Asian countries, South Africa, and Scandinavia; the countries that
are members of the European Community have no large resources.
Their beneficiation is based traditionally on the flotation process, a
unit operation taking advantage of mineral surface properties.
Actually, the sulfide minerals are often the main example of
flotation application. Table 1 presents a brief historical background
(1).
Another incentive of the present work is environmental protection
to prevent acid rain, since pyrite flotation is significant in coal
preparation (coal has been used in power plants); sulfur is present
in coal in three forms, i.e., sulfate, organic, and pyritic. The first
content is usually very low, while the organic sulfur is ultimately
associated with the coal. Basically, only pyritic sulfur may be
removed (2). Perhaps due to environmental constraints, this is of
major importance nowadays.
Pyrite was long used in sulfuric acid production, before being
abandonedat least, in the populated and developed countries,
particularly from the air pollution point of viewand replaced by
pure sulfur, coming from mineral resources or oil desulfurization
plants. For obvious economic
Page 128
Table 1 Significant Steps in Sulfide Flotation
Development.
Year Scientist Innovation
1913Bradford Use of copper sulfate for
sphalerite activation
1922Sheridan Introduction of potassium
cyanide to depress pyrite and
sphalerite
1925Keller Invention of the collecting action
of xanthates
1929Gaudin Regulation of pH value of pulp
reasons, pyrite is one of the comparatively little studied sulfide

minerals. Nevertheless, it is often associated with gold or uranium,
which makes its processing attractive. It is also associated with
other iron sulfides like arsenopyrite (FeAsS)being an unwanted
admixtureand pyrrhotite, Fe(1-n)S. A certain relation between
arsenic and gold concentrations in bulk iron sulfide concentrates
often exists (see, for instance, Ref. 3), indicating that gold is
preferably associated with arsenopyrite. Precious metal behavior
and distribution are of major importance in the economic
performance of selective flotation.
The heart of a flotation circuit is the cells. There are many types
and designs (about 40), distinguished by the method of mixing and
air bubble generation; the feed enters through a feed box and
passes from one compartment to the next, through appropriate
openings. It is noted that several cells, containing a number of
compartments, are joined together to form a bank of cells, in which
a particular flotation operation is accomplished, such as roughing,
scavenging, and cleaning (4).
Figure 1 presents a typical scheme of a complex sulfide ore-
processing plant (without a separate copper concentrate). However,
often the situation is more complicated, as shown in Figure 2,
where a fine-grained ore is processed (5). Steps have been taken to
study alternatives whereby some or all nonferrous metals (Pb, Zn,
Cu) could be recovered selectively from pyrite. The predominant
collectors in sulfide mineral flotation are the thio or sulfydryl
reagents.
A
Principles of Sulfide Flotation
It is known that sulfide minerals are of semiconducting nature.
Thus they present strong electrode properties that are directly
connected with their ability to adsorb sulfydryl collectors (6). The
usual cathodic half-reaction results in the production of hydroxyl
ions, while some of the typical anodic
Page 129
Figure 1
Complex sulfide ore processing: a typical flowsheet.
Page 130
Figure 2
Flotation flowsheetproduction of three concentrates and a bulk concentrate.
(From Ref. 5.)
products are adsorbed alkyl xanthate (denoted as X) and
dixanthogen and its metal precipitate (see Fig. 3).
These anodic reactions are combined with the cathodic one by
electron exchange through the semiconductive lattice of sulfide.
Such interactionswhich are the most basiccontribute (to the greater
extent) to the hydrophobicity of sulfide minerals. Nevertheless, it
should be noted that the adsorption of xanthate collector may occur
by pure physical adsorption, as well as by ion exchange
mechanisms of xanthate ions on sulfide surface, with ions existing
already adsorbed on the surface, like HS-, OH-, and SxOyn- (7).
Every chemical reaction of charge transfer corresponds to a certain
value of electrochemical potential, which is related to the Gibbs
free energy change and the number of electrons exchanged. It is
apparent that the electrochemical potential immediately affects the
function of flotation in its fundamental stage of collector
adsorption.
Consideration also should be given to the fact that, from the
thermody-
Page 131
namic point of view, sulfides are unstable systems and tend to be transformed to their
respective oxides (more stable compounds) (8). The electrochemical potential influences
the chemical nature of sulfide surfaces, which may affect indirectly the collector
adsorption. So, on oxidized surfaces the electron transfer reactions are largely hindered, as
the oxide layer negatively affects the electron transfer (9).
The electrochemical interactions between electrodic surfaces, called galvanic, are due to
the different electrochemical activities of the conductive solids (6). In other words, the
combination of a cathodic solid surface (which acts as the cathodei.e., attracts electrons)
and an anodic surface (which offers electrons) forms a galvanic cell.
Figure 3
Schematic representation of collectorsulfide mineral (denoted as MeS): interactions of charge transfer.
Page 132
These interactions significantly influence flotation operation,
because there is a mixing of solid particles with varying
electrochemical activity (as the sulfides are) and of metals originated
from the corrosion of grinding media (the latter being steel balls and
rods). A table (Ref. 10) was presented with the values of rest
potential of sulfide minerals referenced to standard hydrogen
electrode, at pH 4. Sulfides with greater rest potential are more
cathodic and are considered as less electrochemically active. For
example, pyrite (rest potential 0.66 V vs. SHE) is a more inert
sulfide, while arsenopyrite, is more active (rest potential 0.39 V vs.
SHE).
Figure 4(a) shows the results of electrolytic contact of two sulfides.
The hydroxides produced on the cathodic mineral result in the
depression of its flotation. Elemental sulfur formed on the anodic
mineral increases its hydrophobicity, so that it may sometimes lead
to collectorless flotation. Similar actions [see Fig. 4(b)] are
developed between minerals and grinding media. The latter are often
iron alloys, which are more anodic than sulfides and hence seriously
hinder flotation, as is the case for pyrrhotite (10). In summary, the
galvanic actions are responsible for the phenomenon of flotation by
sulfydryl collectors, and this is illustrated schematically in Figure 5.
Figure 4
Galvanic interactions: (a) mineral/mineral
(b) mineral/grinding media.

Page 133
Figure 5
Effect of electrochemical actions on sulfide mineral flotation.
B
Electrical Properties of Interfaces
Flotation is a process in which three phases take place. Hence, the
properties of interfaces, which are developed between them, are of
major importance for a fundamental understanding of what is
happening and whyi.e., the mechanism of the process. When a
mineral particle is introduced in a certain solution, then at the
solid/liquid interface an electrical charge is developed due to the
potential-determining ions (those that are in an adsorbed state on
the solid); the origin of the latter is explained by a series of
reactions (11).
In this way, the interactions of water molecules and their
dissociation products with various species of mineral surface may
lead in the formation of other charged species and the creation of a
surface charge. These interactions happen with most minerals, but
also there are other ways for a surface charge to be generated, such
as the specific adsorption of surface active (or not) ions from the
solution, the selective dissolution of certain surface ions from the
mineral, and the substitution of ions in the crystal lattice. In order
to accomplish electroneutrality, a number of ions (the counter ions)
from the solution are adsorbed by electrostatic forces on the
interface and an excess
Page 134
charge is developed nearbybeing in a diffused state toward the
solution and having equal but opposite value to the surface charge
(see Fig. 6).
The surface potential y0 (i.e., the potential difference between the
mineral surface and the solution) depends on the adsorption density
of the cations and anions that determine the potential (12). The
activity of the potential-determining ions in the solution, where the
density of surface charge is zero, gives the point-of-zero-charge
(pzc). Often these ions are the H+ and OH-, and the pzc is
expressed by some value of the pH of solution.
Figure 6
Schematic representation of the electrical double layer and
the potential drop across the double layer.
Page 135
Table 2 Techniques of z-Potnetial Measurement.
Name Principle
ElectrophoresisDealing with mobility of a charged colloidal
particle suspended in an electrolyte, in which a
potential gradient is set up
Streaming Concerned with the magnitude of induced electric
potential potential, generated by a flow of liquid in a single
capillary, or in a system of capillaries (i.e., a
porous plug of compressed solid particles)
Electroosmosis Denoting a movement of liquid in a capillary, or
in a system of capillaries, under an applied
potential
Sedimentation Determining the magnitude of the potential
potential developed by a dispersion of particles settling
under the influence of gravity
Generally, the surface potential is not directly measurable. So, the

electrical behavior of mineral dispersions (pulps) is usually made
known by a more easily measured parameter, the zeta-potential.
Employing a simplified double-layer model, it may be accepted
that the potential difference developed between a hypothetical level
known as ''Stern plan" ("b" in Fig. 6), which represents the
minimum distance of the hydrated counter ions from the surface,
and the solution gives the zeta-potential value (13). The latter is
measured by four basic methods, as described in Table 2 (4). In
flotation research, electrophoresis and streaming potential are
generally applied; the former is applied in the present study.
II
Application of Electrokinetic Studies
The investigation of zeta-potential changes and, in general, of the
electrokinetic behavior of a system can give much information on
the ways the different existing ions may function. The basic
applications in flotation are classified in the
1. study of the method of collector adsorption,
2. selection of the appropriate flotation reagents, and
3. examination of activation mechanisms.
The interaction mechanism between collector and a certain mineral
was clarified (14) for galena (PbS) and pyrite in the presence of the
same concentration of a xanthate (O-alkyl-dithiocarbonate) salt. A
shift to more negative values of zeta-potential of galena was
observed; the mineral ad-
Page 136
sorbs a negative-collecting species, being the xanthate ion. The zeta-
potential of pyrite was practically unchanged, due to its surface
coverage by a nonpolar species; the latter was proved to be
dixanthogen (discussed later).
The identification of the surface charge can be a useful advisor as to
the suitable flotation reagent. For example, goethite presented a
positive zeta-potential in an acidic environment, while at pH 7 it was
reversed to negative values (15). So, it floated satisfactorily by
anionic collectors (like sodium alkyl sulfate and sulfonate) at acidic
media; in alkaline solutions, the cationic collectors were appropriate,
and the anionic collectors were inactive.
For the case of quartz (16), it was found that the different polyvalent
cations activate its anionic flotation (by sodium alky sulfonate).
Activator is a compound, usually ionic, that intensifies the
adsorption of collector on the mineral surface. This activation was
maximum at the pH value where the cationic hydroxy complexes
were formed.
Often electrokinetic data (as electrophoretic mobility) are correlated
and give similar curves, with other flotation parameters, such as
adsorption density, contact angle, settling rate, or floatability. The
system of alumina/dodecylsulfonate is well known (17). A sharp
increase in the slope of the adsorption isotherm along with those of
other parameters at an adsorption density of 10-12 mol cm-2 was
clearly noticed.
An extensive review of metal sulfides was presented (18), which at
that time (1976), raised some interesting points: little attention was
given to experimental details such as adequate analysis of the
sulfide, conditioning effects, sample equilibration, prevention
measures that control oxidation. However, it concluded that the zeta-
potential can be an excellent tool for observation and description of
surface reactions at the mineralwater interface.
III
Pyrite Flotation Behavior
Gaudin, since 1956, identified adsorbed dixanthogen on the pyrite
surface (1). It has been proved since then by extensive research that
dixanthogen aids in pyrite flotation (19). Fuerstenau et al. (20), by a
combination of spectroscopic and electrokinetic indications with
flotation data, explained the pyrite flotation by potassium ethyl
xanthate due to dixanthogen and proposed the following mechanism:
At the same time, Majima and Takeda (21) by electrochemical

techniques supported the mechanism:
Page 137
It may be noted that in Eq. (2), the electrodic nature of the sulfide
surface was considered. The aforementioned views were certified by
electrochemical studies of Usul and Tolun (22).
Analyzing the electrochemical reaction, we obtain
These half reactions are connected by electron passage and exchange

through the sulfide lattice (14), and this is generally the accepted
mechanism today. However, there are indications that, except for
dixanthogen, other species also contribute in pyrite flotation, like
ferric xanthate, ferric hydroxy-xanthate, xanthate ion, and elemental
sulfur (23,24).
From the reaction of xanthate oxidation, it is clear that the dissolved
oxygen of pulp plays a significant role in the process. The influence
of oxygen was described by Leja (4). Recently, work has been
conducted even in the substitution of air in flotation circuits by
nitrogen (25). The selective flotation of pyrite from its mixture (an
industrial product) with arsenopyrite, by the application of nitrogen
as the flotation gas, was published by Kydros et al. (26).
The semiconducting nature of sulfides and the strong trend of thio-
type collectors to be oxidized offer possible explanations for the
observed interactions between them. From the 1950s the view that
the presence of oxygen is mandatory for sulfide flotation by
sulfydryl collectors has been argued (see, for instance, Ref. 27).
Also, electrochemical techniques, like voltammetry, were introduced
in flotation chemistry research (28). Plaksin and Shafeyef (29)
correlated the electrophysical properties of sulfides with the ability
to reduce oxygen and, hence, adsorb xanthate compounds.
According to them, the sulfide surfaces are composed of a mosaic of
cathodic and anodic regions. For each sulfide in dispersion, there is a
certain oxygen concentration over which the mineral is transformed
superficially to a p-type semiconductor.
Allison et al. (30) correlated the mixed potentials of electrodes, from
various minerals in xanthate solutions, with the equilibrium potential
of the couple xanthate/dixanthogen and analyzed the active
collective species in each case. Those sulfides having mixed
potential greater than the equilibrium potential of the
xanthate/dixanthogen couple, form dixanthogen. Trahar (31)
examined floatability in relation to redox potential. An excellent
review of the field was published by Rao et al. (6).
Hayes and Ralston (32) presented a potentiostatic Hallimond tube,
being a development of the earlier one by Chander and Fuerstenau
(33); they
Page 138
investigated flotation without collectors and only the control of
redox potential. Labonté and Finch (34) as well as Rand and
Woods (35) examined the area from the instrumental aspect and
proposed different electrodes for potential measurements in
industrial conditions. Potentiostatic flotation was also examined by
Panayotov (36).
Lime is a strong depressing agent for pyrite (20). First, its addition
in the pulp results in an increase of pH value. Complete flotation
depression is successful at pH 10. It was supposed (37) that the
calcium ion is attracted by electrostatic forces from the negatively
charged surface of pyrite. This suggestion, however, did not take
into consideration that pyrite surface is negative from a pH value of
7 and above. The fact that lime influence is maximized in a pH
value of 10 indicates that calcium ion adsorption takes part through
specific interactions following the known mechanism of interfacial
precipitation (13) of its hydrolysis products (formed in pH's over
10).
Another well-known depressant is cyanide ion, which is used
extensively in practice. The presence of potassium cyanide in
flotation pulp leads to a complete pyrite depression at a pH of
around 8. It was suggested that the compound Fe4[Fe(CN)6] is
responsible for its depression (38). Nevertheless, due to
environmental concern, there are many attempts recently at
noncyanide technology.
In noncyanide technology, mention will be given to
biomodification of mineral surfaces (39), a really promising
method. Applications include the beneficiation of sulfide minerals
from complex ore body by selective or bulk flotation, removal of
pyritic sulfur from coal, and extraction of precious metals
encapsulated in a sulfide matrix.
Finally, the removal of particulate matter by foam or froth flotation
was reviewed recently (40). Process parameters (particle and
bubble size, surfactants), available techniques (dissolved-air,
electrolytic, column flotation), and design aspects (operation,
hydrodynamics, kinetics, and modeling) were examined. A small
chapter on mineral processing was also added, based on earlier
experience with magnesium carbonate ores.
IV
Experimental Details
Microelectrophoresis
A microelectrophoresis cell consists of a glass tube either
cylindrical or with a square vertical section (flat cell), equipped
with an electrode in each end. An optical system (microscope or
video monitor through a camera) for microscopic observations is
usually accommodated near this tube. The flat cells are generally
preferred in the studies of solids having high settling rates, like the
mineral particles, while the cylindrical cell is selected to examine
the electrokinetic properties of emulsions (13).
Page 139
When a potential is applied to the cell's edges, the particles are
moving toward the direction of the electrode with the reverse
charge (electrophoresis). Except for this motion, other
electrokinetic effects appear due to electro-osmosis. The inner glass
surface of the cell is charged in the presence of liquid. Due to the
electric field, a motion of the liquid appears near the glass wall,
which (as the cell is closed) leads to a liquid recirculation showing
maximum velocity in the tube center (see Fig. 7). So, a parabolic
distribution of liquid velocity is revealed, and the particle acquires
an apparent velocity greater than the real, electrophoretic one that
is observed only in the stationary levels. There, the electro-osmotic
flow and the liquid recirculation are counteracted.
The measurement at the stationary levels is based on the Komagata
equation (12), giving their place:
where s is the distance of the plane from the cell wall, d its depth,
and l its width (see Fig. 7). From the measurement of the particle
mean velocity for a certain potential difference, its electrophoretic
mobility can be deduced. Twenty-five measurements were
conducted for each run, and the electrode's polarity was reversed
between each run; the distance covered and the potential difference
value were adjusted in order to have measurable times on the
Figure 7
(a) Velocity distribution in the flat cell
(b) stationary levels.

(From Ref. 13.)
Page 140
order of 1015 s with accuracy of two decimal digits (19).
Electrophoretic mobility values are often converted to zeta-
potential by the known Smoluchowski equation. The latter is easily
applied to aqueous dispersions of mineral particles (14). The sign
of zeta-potential is taken from the particle motion direction,
according to the electrode polarity.
In this work, the zeta-potential measurements were carried out with
an apparatus of Rank Brothers, UK (Mark II); the principle is
presented in Figure 8. To avoid surface oxidation, before each test
the mineral sample was ground in an agate mortar. The
conditioning was similar to that of the flotation experiments, in
order to get comparable results. An amount of 0.05 g of sulfide
fines was weighed and mixed in a 50 cm3 reagent (modifier,
regulator, collector, etc.) solution. The dispersion was then left to
rest for 120 s, and a part of the supernatant liquid was taken in the
flat cell, being inserted in the water bath (at 298 K) where it
remained for another 120 s. The electrodes used in the apparatus
were from platinized platinum to avoid polarization problems. The
applied potential was up to 100 V, and the current varied from a
few hundred µA to some mA, depending on the ionic strength of
the solution in each case.
Figure 8
Microelectrophoresis apparatus: 1, electrodes; 2,
microscope; 3, micrometric screw; 4, light source;
5, flat cell; 6, water bath; 7, water recycle.
Page 141
V
Measurements of z-Potential and Discussion
A
Influence of pH
It is known that the regulation of the solution pH (in the pulp) is
one of the most common methods of flotation control, particularly
in the sulfide minerals (see also Fig. 1). In sulfide minerals, this
parameter plays a rather complex role. So, in the alkaline
environment, the flotation of pyrite is depressed when xanthate
salts are used as collectors (1,3,37). The critical value, i.e., the
boundary value over which flotation is not possible, was found for
the system pyriteoxygenethyl xanthate to be 10.5, with 25-ppm
collector. A first explanation of this phenomenon was given by
Fuerstenau et al. (20), who reported the absence of dixanthogen in
alkaline pH values. It was attributed to a shift of equilibrium of
reaction 2 toward the left, due to the excess of hydroxyls.
According to Woods (29), the oxidation of pyrite at pH values over
11.4 happens in lower potentials than the xanthate oxidation. In this
way, the action of dissolved oxygen is limited to the corrosion of
pyrite rather than to the oxidation reaction 2. Thermodynamic
analysis also showed that at high pHs, the surface of pyrite is
covered by Fe(OH)3, which makes the aforementioned reaction
difficult. Similar observations were published for arsenopyrite (41).
The optimum flotation recovery of pyrite by potassium ethyl
xanthate, in relatively low additions, was recorded to be succeeded
at pH ranges of about 3 and 8. On the contrary, in the neutral region
(pH 5.57.5), a significant depression appeared (20). Later,
Fuerstenau and Mishra (23), based on electrokinetic data, assumed
that pyrite depression around the pH region of zero-point-of-
charge, at pH value of 6.9, was due to the fact that the xanthate ion
was not possible to be attracted on its surface by electrostatic
forces.
Today, it is well known that the sulfide minerals are unstable to
oxidation. The theory accepts the existence of an inner
hydrophobic layer, being reached in sulfur, which is covered by a
layer of metal hydroxy-oxides that makes the mineral hydrophilic.
This layer might be enough to prevent the collector adsorption
(42).
The zeta-potential of pyrite from two different areas (from Mexico
and Greece) was recently presented for comparison (26). In the
first case, a reversal of zeta-potential was noticed in the neutral pH
region (43,44). This reversal was apparently due to the presence of
metal ions from surface oxidation.
However, the sample from Chalkidiki, Greece remained
unoxidized, in these conditions, which may be attributed to the
varying microstructures of
Page 142
the mineral samples (19). The presence of admixtures (impurities)
differentiated the energy of crystal lattice and, consequently, the
mechanism of the oxidation of sulfides, which follows different
stages (37). It seems that the sample origin and preparation assume
important roles.
Indeed, during the absence of oxygen, it was found by Fuerstenau
et al. for the pyrite from Naica, Mexico that the zeta-potential
reversal was not appearing any more (43,44). On the contrary, the
Greek pyrite also had a characteristic zeta-potential reversal when
ferric ions were added in the pulp.
This film of iron hydroxides, which was traced on pyrite,
negatively influenced flotation by xanthates (20) and also by
amines. When the dissolved oxygen is driven out from the flotation
liquor, nitrogen leads to the preservation of surface; nevertheless, it
does not maintain discrete flotation properties in all the cases (26).
A fundamental investigation of flotation processing of pyrite and
arsenopyrite by a cationic collector in the laboratory was lately
reported (45) based mainly on electrokinetic as well as floatability
experiments; electrokinetic experiments were applied on an
auriferous pyrite industrial concentrate.
At first glance, certainly the former will not be considered as
original (see, for example, Refs. 20,23,38,46). However, from
careful reading, a lot of contradiction was realized regarding pyrite
behavior.
B
Cationic Collectors
Amines are a category of flotation collectors that find extensive
application, mainly for nonsulfide metallic and silicate ore
processing (47). They are well known for their great collective
action in flotation, but, due to the lack of selectivity particularly in
nonacidic medium compared to sulfydryl collectors, they have not
found much use on complex sulfide beneficiation. Nevertheless,
the applicability of amines to sulfides has been discussed for a long
time now (1). This could be of particular interest, for instance, in
the case of pyrite removal from ores in which nonsulfides are
valuable.
Limited work, according to our knowledge, exists on the cationic
flotation of pyrite (45). In South Africa (48), it was suggested that
the xanthate reagents have a very deleterious effect on the
cyanidation of gold; accordingly, the flotation of pyrite was usually
conducted after cyanidationcyanide is inherently a depressant for
pyrite (38). Perhaps another application of amine flotation of pyrite
would solve this problem. Cetyl-trimethyl-ammonium bromide was
the quaternary amine applied as cationic collector in the present
(denoted as CTMAB).
The zeta-potential of pyrite in the presence of iron and amine
(conditioning 5 and 15 min, respectively) is presented in Figure 9.
No substantial
Page 143
Figure 9
Zeta-
potential measurements of freshly ground pyrite modified by ferric ions
(800 ppb Fe3+) in the presence of CTMAB.
Reprinted with permission from Academic Press (45).
difference was observed, compared with the behavior when no
amine was added (45). The positive charge of the mineral, in this
case, possibly inhibits a significant adsorption of amine. Similar
results, shown as Figure 10, were noticed with the oxidized pyrite.
Generally, the action of iron on pyrite has some consequences with
respect to that of specifically adsorbed hydrolyzable polyvalent
cations, studied mainly on oxide and silicate ores (49). This
phenomenon was excellently reviewed by Hunder (13). According
to that analysis, the point of zeta-potential reversal (zpr) at the pH
of 2.5 represents the pzc (point of zero charge) or the isoelectric
point of pyrite. While the pzr at pH 67 is shifted at higher pH
values with the increased concentration of iron, the limit is the pzc
of the metallic hydroxide; indeed, that of trivalent iron hydroxide is
at a pH around 7.
Figures 11 and 12 present the respective studies on arsenopyrite
freshly ground and aged. This time, the known reversal of zeta-
potential was seen in the presence of amine, which has the same
characteristics as pyrite in the presence of ferric ions.
Page 144
Figure 10
As in Figure 9, but with aged samples of pyrite.
(Adapted from Ref. 45.)
A possible mechanism of collector adsorption that was proposed
from these investigations has been an ion exchange on the
arsenopyrite surface between the chemisorbed iron and the cation
of amine collector. The free ion, which is produced from the
substitution by the amine cation on the mineral surface, is
hydrolyzed and, hence, reverses the zeta-potential of the mineral.
Similar mechanisms of adsorption of cationic surfactants, mainly
during flotation in an acidic environment, were described by
Gaudin (1).
These measurements of zeta-potential characteristically showed
that the aging of pyrites led to a mineral surface alteration, which
simplifies their selective separation (45). This was also met in the
literature (50).
C
Other Anionic Collectors
The collective action of anionic collectors (except the sulfydryl,
which was already mentioned), such as the soaps, on sulfides is
significant. Their use on the removal of sulfide minerals from ores,
in which other minerals are of economic interest, was reported (1).
However, the classical anionic collectors act, to a certain extent, by
chemisorption as they form with various metallic ions soaps that
are difficult to dissolve (51). So, sodium dodecyl-sulfonate was
applied, which is adsorbed by electrostatic attrac-
Page 145
Figure 11
Zeta-
potential measurements of freshly ground arsenopyrite modified by ferric
ions (800 ppb Fe3+) in the presence of CTMAB.
Reprinted with permission from Academic Press (45).
tions. Also, they have been previously used for the beneficiation of
oxides and salt-type minerals (14).
Electrokinetic tests of pyrite (and arsenopyrite) with and without
any collector was conducted (52). When sulfonate was added, a
shift toward more negative values of zeta-potential was observed
both in acidic and alkaline medium, which indicated an intense
adsorption of the collector on the mineral surface. An advantage of
these nonsulfydryl collectors is that they can promote a reverse
flotation process in the system pyrite/arsenopyrite, i.e., floating
arsenopyrite (19), which is a more reasonable solution. Floating
arsenopyrite is important as arsenopyrite usually exists in lower
concentration than pyrite. It is worth noting that pyrite flotation
concentrates having arsenopyrite, as low as 12% are not accepted
as raw materials in sulfuric acid production.
Such interactions (as the aforementioned) have already been
reported (53) and are related to the adsorption of sulfonates on
kaolinite surface. They were attributed to an ion exchange
mechanism, involving foreign cations added with the pH regulator
(Ca2+, Na+).
Page 146
D
Influence of Hydrous Copper Sulfate
It is generally known that copper sulfate pentahydrate activates
pyrite flotation by these collectors, a property that finds industrial
application from the 1950s (1,37). However, little is published on
the chemistry of activation; this field was reviewed by O'Connor et
al. (54). Possible mechanisms seem to be the following:
1. Complexation of cyanide ions, which were added in preceding
stages by the divalent copper ion.
2. Adsorption of Cu2+ ions on pyrite surface and help in the
oxidation of the thio collector.
The possibility of arsenopyrite separation by flotation from pyrite
in the presence of copper sulfate was also suggested by
Glembotsky et al. (37).
The zeta-potential of pyrite as a function of pH in the presence of
increasing amounts of copper ions is shown in Figure 13. A pzr can
be observed around the neutral and slightly alkaline pH range. This
reversal is due to pyrite surface coverage by hydrolyzed copper
cationic species. A mechanism of interfacial precipitation of the
metal hydroxide was proposed
Figure 12
As in Figure 9, but with aged samples of pyrite.
Page 147
Figure 13
Zeta-
potential measurements of pyrite (0.1 g/L) in the presence of copper ions.
Reprinted with permission from Marcel Dekker, Inc. (55).
(55). The latter may occur at a pH lower than that required for
precipitation in the absence of pyrite.
The trend of copper sulfate to be adsorbed and activate the flotation
of pyrite is so great that it was applied for the removal of copper
ions, as hydroxy precipitates, from dilute aqueous solutions after
their adsorption on fine pyrite particles. The next separation stage
was dissolved-air flotation, and the whole process was termed
adsorbing flotation. In this way, a further utilization of pyrite was
investigated and suggested in environmental technology.
E
Influence of Ammonium Salts
The role of ammonium salts in pyrites flotation was described
mainly by Russian scientists, like Glembotsky (37) and Klassen and
Mokrousov (56). According to them, NH4+ was selectively
depressing pyrite flotation in presence of lime, but it had no such
action on arsenopyrite. This made possible their separation.
The application of magnesia mixture (mixture of magnesium
chloride, ammonium chloride, and ammonia in aqueous solution)
was first suggested
Page 148
as modifier during flotation following depression by AsI3 (57). The
authors assumed that the hydrophilic component that depressed
arsenopyrite was the MgNH4·AsO4·H2O. The flotation processing
of a bulk auriferous pyrite concentrate by magnesia mixture has
been presented (3).
Figure 14 presents the electrokinetic part of this work on pure
minerals; the range of good separation was at pH values between 9
and 10. Arsenopyrite was shown to be more positive than pyrite in
a pH range of 7 to 11 approximately. Arsenopyrite had a pzc at pH
10, where complete depression was noticed, while pyrite had a pzc
at pH 10.5.
F
Oxidative Reagents
The depressing action of oxidants, like potassium permanganate
and dichromate, has been pointed out even from the 1930s (58).
Later, the crystal structure of iron sulfides was related with their
susceptibility to spontaneous oxidation (37). This theory also led to
the development of a separation method of pyrite/arsenopyrite from
a bulk concentrate by pyrolusite addition.
Nevertheless, the assumption that pyrite was more susceptible to
oxida-
Figure 14
Zeta-potential of iron sulfides in presence of (4 ml/L) magnesia mixture;
conditioning 10 min.
Page 149
tion than arsenopyrite soon fell to pieces, since electrochemical
studies proved that pyrite is the most noble sulfide mineral (4,15)
(i.e., the most refractory in oxidation). Particularly excellent work
was conducted at the University of British Columbia (9) where they
found that the oxidation of arsenopyrite happens in lower potentials
than that of pyrite. In this way, by using common oxidative agents
(such as KMnO4, H2O2, NaOCl) the selective oxidation of
arsenopyrite in a bulk concentrate was accomplished.
Similar conclusions related to the use of permanganate were reached
elsewhere (59,60), and methods were published to clean a pyrite
concentrate from arsenic. The measurement of zeta-potential of
pyrite and arsenopyrite, in the presence of KMnO4, was also
reported (52).
When the modifier was added, it resulted in a slight modification of
the zeta-potential profile of pyrite, but the mineral surface remained
essentially negatively charged. On the other hand, arsenopyrite
appeared to be positively charged in the acidic region (pH < 4); its
pzr was at a pH of around 4.
The products of permanganate reduction in this pH range are
expected to be hydrous manganese oxides. It seemed possible that
the heterogeneous redox process between the aqueous permanganate
and the sulfides (mostly arsenopyrite) led to a surface deposition of
hydrated Mn(IV) species, which were responsible for the zeta-
potential reversal. Indeed, the pzc of precipitated manganese oxide
(MnO2) was reported to be 4.2 (4).
When the sulfonate collector was added in the mineral dispersion,
with KMnO4 acting as a modifier, pyrite was shown to be slightly
affected, but it remained negatively charged. The zeta-potential of
arsenopyrite, however, shifted toward more negative values in the
pH range of interest for separation, around 4, indicating a strong
collector adsorption (52).
G
Reducing Modifiers
Since dixanthogen has been seen as the more significant collector
species in pyrite conventional flotation, it has been apparent that
every chemical reagent applied as modifier, which prevents the
formation, will act as a flotation depressant of the minerals. For the
half-element xanthate-dixanthogen, we have Eq. (3):
According to Eq. (6), every agent more reducing than xanthate will
not permit its oxidation acting as a flotation depressant. Such
reagents are sodium sulfide and the various sulfites. For instance, for
the couple SO32- - SO42- we have the reaction
Page 150
Hence, sodium sulfite is more reducing than the xanthate ion. At pH
values over 6.5, an amount of the order of 10-2 M Na2SO3 was
enough to depress pyrite flotation completely (14). The
chemisorption of SO32- on pyrite was mentioned in electrokinetic
studies. Hoyak and Raghavan suggested the formation of the
hydrophilic compound Fe2(SO4)2·Fe(OH)3 (61).
The use of sulfoxy-reagents in the modification of mixed sulfides,
in fact, has been practiced since 1919 (8). Sodium metabisulfite (or
disulfite) Na2S2O5 constitutes a less common agent with rather
weak reducing properties. In this case, it was found that the
selectivity was lowered after extensive mixing of the pulp. So, no
conditioning and a short flotation time were applied (3). This
modifier performed well in an acidic environment, and good results
were obtained, both in terms of recovery and grade of the product.
Figure 15 presents the electrokinetic measurements.
Hydrazinium sulfate [H2NNH3]HSO4 is another reducing agent
with different formulai.e., not producing sulfoxy compounds in the
solution. It has the further advantage that its oxidation gives gas
nitrogen, which does not modify the surface properties (19). The
electrokinetic measurements were presented earlier (see Fig. 16). It
was found that this modifier shifted
Figure 15
Zeta-
potential of pyrite in presence (and not) of metabisulfite; conditioning 5 min.
Page 151
Figure 16
Zeta-potential of pyrite in presence (and not) of hydrazinium sulfate;
conditioning 10 min.
the zeta-potential of pyrite (and arsenopyrite) to more negative
values in the pH range investigated. This behavior was explained
by the possible ''cleaning" of surfaces, i.e., a depression of the
formation of superficial oxidized species. The level of dissolved
oxygen concentration is a critical parameter in differential flotation
(62).
This reagent was further observed to depress arsenopyrite over all
the pH range, giving especially clean concentrates at values over 8.
The selective modifying action of a reducing and nonsulfoxy
reagent, like hydrazinium sulfate, was a strong indication that this
modification was mainly due to their reducing properties and not to
a possible selective sorption of modifying species on mineral
surfaces. With the addition of chemicals consuming oxygen (like
the different reducing agents), it is expected that a pyrite would
float selectively from arsenopyrite, as also proved by Hallimond
tube tests.
In conclusion, electrochemistry is the basis to explain in-depth the
mechanisms of pyrites flotation and even to control the process.
Moreover, the electrokinetic measurements may provide useful
indications on flotation, which are usually followed by floatability
experiments.
Page 152
References
1. A. M. Gaudin, Flotation, 2nd ed., McGraw-Hill, New York.
(1957).
2. F. F. Aplan, Coal flotation, FlotationA. M. Gaudin Memorial
Volume (M. C. Fuerstenau, ed.), SME/AIME, New York, p. 1235.
(1976).
pyrite concentrate by flotation reagents, Miner. Eng., 5(35): 391
(1991).
New York. (1982).
5. G. Barbery, Complex sulphide ores: Processing options, Mineral
Processing at a Crossroads (B. A. Wills and R. W. Barley, eds.),
Martinus Nijhoff, Dordrecht, p. 157. (1986).
6. S. R. Rao, G. Labonté, and J. A. Finch, Electrochemistry in the
plant, Innovations in Flotation Technology (P. Mavros and K. A.
Matis, eds.), Kluwer Academic, Dordrecht, p. 57. (1992).
7. G. W. Poling, Reactions between thiol reagents and sulphide
minerals, FlotationA. M. Gaudin Memorial Volume (M. C.
Fuerstenau, ed.), SME/AIME, New York, p. 334. (1976).
8. I. Bayraktar, U. A. Ipekoglu, and R. Tolun, Features and
flotation of complex Cu-Pb-Zn sulphides, Innovations in Flotation
Technology (P. Mavros and K. A. Matis, eds.), Kluwer Academic,
Dordrecht, p. 307. (1992).
9. G. W. Poling and M. J. V. Beattie, Selective depression in
complex sulphide flotation, Principles of Mineral FlotationThe
Wark Symposium (M. H. Jones and J. T. Woodcock, eds.), Austral.
IMM, Melbourne, p. 137. (1984).
10. D. Kocabag and M. R. Smith, The effect of grinding media and
galvanic interactions upon the flotation of sulphide minerals,
Complex SulfidesProcessing of Ores, Concentrates and by
Products (A. D. Zunkel, R. S. Boorman, A. E. Morris, and R. J.
Weseley, eds.), AIME and Can. IMM, San Diego, p. 53. (1985).
11. D. W. Fuerstenau and S. Raghavan, Some aspects of the
thermodynamics of flotation, FlotationA. M. Gaudin Memorial
(1976).
12. D. J. Shaw, Electrophoresis, Academic Press, London. (1969).
13. R. J. Hunder, Zeta Potential in Colloid Science. Principles and
Applications. Academic Press, New York. (1981).
14. M. C. Fuerstenau, J. D. Miller, and M. C. Kuhn, Chemistry of
Flotation, AIME, New York. (1985).
15. F. F. Aplan and D. W. Fuerstenau, Principles of nonmetallic
mineral flotation, Froth Flotation, 50th Anniversary Volume (D. W.
16. M. C. Fuerstenau and B. R. Palmer, Anionic flotation of oxides
and silicates, FlotationA. M. Gaudin Memorial Volume (M. C.
17. P. Somasundaran, Interfacial chemistry of particulate flotation,
Advances in Interfacial Phenomena of Particulate/Solution/Gas
Systems; Applications to
Page 153
Flotation Research (P. Somasundaran and R. B. Grieves, eds.),
AIChE, New York, p. l. (1975).
18. T. W. Healy and M. S. Moignard, A review of electrokinetic
studies of metal sulphides, FlotationA. M. Gaudin Memorial
(1976).
19. K. A. Kydros, A Study of Floatability of Pyrite, Arsenopyrite
and their Concentrates, Ph.D. Thesis, Department of Chemistry,
Aristotle University, Thessaloniki (1992).
20. M. C. Fuerstenau, M. C. Kuhn and D. A. Elgillani, The role of
(1968).
21. H. Majima and M. Takeda, Electrochemical studies of the
xanthatedixanthogen system on pyrite, Trans. AIME, 241: 431
(1968).
22. A. H. Usul and R. Tolun, Electrochemical study of the
pyriteoxygenxanthate system, Int. J. Miner. Process., 1: 135
(1974).
23. D. W. Fuerstenau and R. K. Mishra, On the mechanism of
pyrite flotation with xanthate collectors, Complex Sulfide Ores (M.
Jones, ed.), IMM, London, p. 271. (1980).
24. J. K. Critchley, Reaction of ethylxanthate at an iron electrode,
Trans. IMM, 90: C42 (1980).
25. C. J. Martin, J. A. Finch, and S. R. Rao, Complex sulphide ore
processing with pyrite flotation by nitrogen, Int. J. Miner. Process.,
26: 95 (1989).
26. K. A. Kydros, K. A. Matis, and P. K. Spathis, The use of
nitrogen in flotation, present book.
27. R. Woods, Electrochemistry of sulphide flotation, as in
reference 2, p. 298.
28. S. M. Ahmed, Electrochemical studies of metal sulfides in
relation to flotation, Recent Developments in Separation Science,
Vol. 5 (N. N. Li, ed.), CRC, Florida, p. 95. (1979).
29. I. N. Plaksin and R. Sh. Shafeyef, Influence of surface
properties of sulphide minerals on adsorption of flotation reagents,
Trans. IMM, 72: 715 (1963).
30. S. A. Allison, L. A. Goold, M. J. Nicol, and A. Granville, A
determination of the products of reaction between various sulfide
minerals and aqueous xanthate solution and a correlation of the
products with electrode rest potentials, Met. Trans., 3: 2613 (1972).
31. W. J. Trahar, The influence of pulp potential in sulphide
flotation, Principles of Mineral FlotationThe Wark Symposium (M.
H. Jones and J. T. Woodcock, eds.), Austral. IMM, Melbourne, p.
117.
separation of sulphide minerals by Eh control, Int. J. Miner.
Process., 23: 55 (1988).
33. S. Chander and D. W. Fuerstenau, Effect of potential on the
flotation and wetting behaviour of chalcocite and copper, Trans.
AIME, 258: 248 (1975).
34. G. Labonté and J. A. Finch, Measurement of electrochemical
potentials in flotation systems, CIM Bullentin, 81: 920 (1988).
35. D. A. J. Rand and R. Woods, Eh measurements in sulphide
mineral slurries, Int. J. Miner. Process., 13: 29 (1984).
36. V. Panayotov, Design of a non-cyanide technology for floating
lead-zinc ores,
Page 154
"Non-Ferrous MetallurgyPresent and Future," 1st Europ. Metals
Conf. Proceeds., GDMB & IMM, Brussels, (1991), preprint.
37. V. A. Glembotsky, V. I. Klassen, and I. N. Plaksin, Flotation,
Primary Sources, New York. (1972).
38. D. A. Elgillani and M. C. Fuerstenau, Mechanisms involved in
cyanide depression of pyrite, Trans. AIME, 241: 437 (1968).
39. M. K. Yelloji Rao and P. Somasundaran, Biomodification of
mineral surfaces and flotation, present book.
40. K. A. Matis and P. Mavros, Foam/froth flotation. II. Removal
of particulate matter, Sep. Purif. Methods, 20: 163 (1991).
41. M. J. V. Beattie and G. W. Poling, A study of the surface
oxidation of arsenopyrite using cyclic voltametry, Int. J. Miner.
Process., 20: 87 (1987).
42. A. N. Buckley, I. C. Hamilton, and R. Woods, Investigation of
the surface oxidation of sulphide minerals by linear potential sweep
voltametry and x-ray photoelectron spectroscopy, Flotation of
Sulfide Minerals (K. S. E. Forssberg, ed.), Elsevier, Amsterdam, p.
41. (1985).
43. M. C. Fuerstenau, C. A. Natalie, and R. M. Rowe, Xanthate
adsorption on selected sulphides in the virtual absence and
presence of oxygen, part 1, Int. J. Miner. Process., 29: 89 (1990).
44. M. C. Fuerstenau, M. Misra, and B. R. Palmer, Xanthate
presence of oxygen, part 2, Int. J. Miner. Process., 29: 111 (1990).
45. K. A. Kydros, K. A. Matis, and G. A. Stalidis, Cationic
flotation of pyrites, J. Colloid Interface Sci., 155: 409 (1993).
46. X. Wang and K. S. E. Forssberg, EDTA-induced flotation of
sulfide minerals, J. Colloid Interface Sci., 140: 217 (1990).
47. L. Lindstrom, Flotation of Ore Minerals with Amines as
Collector Reagents, Acta Polyt. Scand., Stockholm, p. 53. (1967).
48. K. G. Ashurst and N. P. Finkelstein, The flotation of pyrite
from siliceous gangue by cationic collectors. A literature survey, S.
Africa Inst. Metal., RR 316, Transvaal (1968).
49. M. C. Fuerstenau, A. Lopez-Valdivieso, and D. W. Fuerstenau,
Role of hydrolysed cations in the natural hydrophobicity of talc,
50. C. T. O'Connor, D. J. Bradshaw, and A. E. Upton, The use of
dithiophosphates and dithiocarbamates for the flotation of
arsenopyrite, Miner. Eng., 3(5): 447 (1990).
51. C. DuRietz, "Fatty Acids in Flotation," Trans. 4th Int. Min.
Dress. Congr., Stockholm, p. 417 (1958).
52. K. A. Kydros, P. Mavros, and K. A. Matis, "Flotation of an
Auriferous Arsenical Pyrite Concentrate by a Non-thio Anionic
Collector," 4th Intl. Min. Proces. Symp. Proceeds., Antalya, p. 345
(1992).
53. J. E. Poirier and J. M. Cases, Sur l' origine et la nature de l'
interaction normale adsorbat-adsorbant dans les systemes a
interactions faibles, Solid-Liquid Interactions in Porous Media (J.
M. Cases, ed.), Technip., Paris, p. 429. (1984).
Page 155
54. C. T. O'Connor, C. Botha, and M. J. Walls, The role of copper
sulphate in pyrite flotation, Miner. Eng., 1(3): 203 (1988).
55. A. I. Zouboulis, K. A. Kydros, and K. A. Matis, Adsorbing
flotation of copper hydroxo-precipitates by pyrite fines, Sep. Sci.
Technol., 27: 2143 (1992).
56. V. I. Klassen and V. A. Mokrousov, An Introduction to the
Theory of Flotation, Butterworths, London (1963).
57. A. Abeidu and A. Almadhy, Magnesia mixture as a regulator in
the separation of pyrite from chalcopyrite and arsenopyrite, Int. J.
Miner. Process., 6: 285 (1980).
58. B. Ball and R. S. Rickard, The chemistry of pyrite flotation and
depression, FlotationA. M. Gaudin Memorial Volume (M. C.
59. D. Draskic, M. Gifing, and J. Pavlica, "Possibility of Arsenic
Content Reduction in Pyrite Concentrate by Flotation
Concentration Process," XIVth Int. Min. Proces. Cong., CIM,
Toronto, (1982), preprint.
60. L. Guongming and Z. Hongen, The chemical principles of
flotation, activation and depression of arsenopyrite, Processing of
Complex Ores (G. S. Dobby and S. R. Rao, eds.), Pergamon Press,
New York, p. 61. (1989).
61. M. E. Hoyak and S. Raghavan, Interaction of aqueous sodium
sulphite with pyrite and sphalerite, Trans. IMM, 96: C173 (1987).
62. A. A. Beyzani and L. P. Kitschen, Selective flotation of sulfidic
complex ores with special reference to the interaction of specific
surface, redox potential and oxygen content, XVI Int. Min. Proces.
Cong. Proceeds. (K. S. E. Forssberg, ed.), Elsevier, Amsterdam, p.
565. (1988).
Page 157
6
Electrochemistry of Sulfide Minerals
Lordwell Kampamba Witika and Bohuslav Dobias *
I
Introduction
A
General Considerations
Electrochemistry is defined as the physical chemistry of ions in
solutions and the phenomena that occur at the solidliquid interface.
Bockris and Reddy (1) viewed electrochemistry as being
subdivided into two fundamental domains, ionics and electrodics,
where the former was restricted to the behavior of ions in solutions
with examples including many familiar redox systems and the later
representing the study of charged interfaces and the conditions
governing charge transfer reactions across them. Since most
reactions occurring in electro/hydrometallurgical processes and
indeed mineral-processing systems involve these kinds of
reactions, it is pertinent to attach more emphasis on the elucidation
of the electrochemical nature of these reactions, and more
especially to sulfide minerals.
Several investigators have presented information on the
electrochemical behavior of sulfide and oxide minerals, the major
sources of day-to-day widely used metallic and nonmetallic
elements. Sulfide minerals have received the most attention as they
are recognized valuable sources of nonferrous metals and because
of their electronic properties, which are such that corrosion and
galvanic couples are readily established in their aqueous sys-
Page 158
tems. The electronic conductivity of most of these minerals enables
them to act as sources or sinks for electrons and hence enables them
to participate in coupled charge transfer processes. This phenomenon
of chemical reactions/interactions on the surface of minerals
occurring through an electrochemical mechanism was first proposed
by Salamy and Nixon (2).
1
Most Common Reaction Types in Sulfide Flotation
Several kinds of electrochemical reactions involving sulfide
minerals, dissolved oxygen, and flotation reagents may occur
depending on various thermodynamic conditions prevailing. The
most common ones are:
1. Cathodic reduction of oxygen, which may proceed through an
intermediate reaction involving H2O2
to give the overall reaction
2. Anodic oxidation of collector through an adsorption phase:
to give an overall reaction

or
3. Sulfide mineral oxidation and formation of polysulfides:
4. Decomposition of metal/reagent compound by oxidation:
This decomposition is similar to that of collector to a nonactive form

through chemical or electrochemical reactions.
5. Decomposition of MX due to depressant action to form a sulfide-
type mineral surface with a decrease in pH:
Page 159
6. Formation of depressant compounds with metallic species such as

in pyrite-cyanide system:
Equation (1), the most important cathodic reaction, suggests that the
presence of oxygen in sulfide systems is a prerequisite for flotation
to occur as was earlier postulated by Plaskin and Bessonov (3). In
his early works, Gaudin (4) explained the requirement for oxygen in
terms of this species being a component in chemical reactions
between sulfide surfaces and collector species.
II
Oxidation of Sulfide Minerals
Interaction between the constituents of sulfide minerals in flotation
systems depends on the prior or simultaneous oxidation of these
minerals. The occurrence of various reactions and their kinetics is
strongly influenced by the potential difference across the
mineralsolution interface (5). Thus, the floatability of sulfide
minerals, excluding those that are believed to be naturally
hydrophobic, is greatly influenced by the degree of oxidation of the
mineral surface. Unfortunately, it is not possible to control the
potential of each of the particles by such methods as are used in
electrochemical research for realistic flotation. The presence of
oxygen is generally sufficient to satisfy this condition, although
oxidants other than oxygen may be used to induce an oxidizing
environment (6,7). The surface oxidation of sulfide minerals has
been studied by examination of leaching and solution analysis
procedures, investigation of electrical characteristics of particle in
suspension, electrochemical measurements on sulfide electrodes,
surface spectroscopic analysis of sulfide surfaces under high vacuum
in dry state, and fairly recently a combination of in situ techniques.
Of late, much attention has been focused on the electrochemical
behavior of sulfide minerals in neutral and alkaline solutions, which
generally are the most relevant conditions in practical sulfide
flotation.
Various investigators have used different electrochemical techniques
to further their understanding of the electrochemical behavior of
different sulfide mineral electrodes in solutions of different
compositions. Linear Potential Sweep Voltammetry (LPSV) is
probably the most extensively used technique and has successfully
been used to characterize pyrite (8,9), galena (10,11), chalcopyrite
(12), chalcocite (13), pyrrotite (9), violalite and pentlandite (14), and
carrollite (15) in detailed studies.
It has generally been postulated that the initial oxidation of simple
sulfide minerals in acidic solutions corresponds to a reaction of the
type (5):
Page 160
with its equivalent in neutral or alkaline solution as:
Further oxidation may lead to the production of oxy-sulfur species:
Indeed, this generalization may also be used to describe the

oxidation of much more complex sulfide minerals. Oxidation leading
to sulfurlike products is usually considered to be critical in rendering
some sulfide minerals such as chalcopyrite (12,16), galena (11,17),
and sphalerite (17) naturally floatable (self-induced floatability).
However, oxidation leading to further oxidation products deteriorates
this self-induced floatability property. This means that the nature of
the oxidation products is very critical, especially when they are
soluble. Under these conditions, they will be expected to dissolve in
the solutions under well-stirred conditions like those in conventional
flotation cells and will not be available for reactions when the
system is subjected to a reducing environment. However, the
insoluble ones such as those expected in Eq. (12) will readily be
available for reduction to reverse the forward reaction. The
appearance of metal species during reduction in the former case will
adversely affect the subsequent mineralcollector interactions.
Electrochemical methods, however, cannot be used alone to identify
reaction products and whether they are insoluble (in which case they
remain on the mineral surface) or soluble (diffuse in solution). This
requires the use of other specific surface analysis techniques, which
will be covered in a subsequent section.
III
Electrochemical Potentials and Their Significance
Several notations for electrochemical potentials have been used
among which electrochemical potentials (E), oxidation/reduction
potential (ORP, redox, Eh) rest potential or open circuit potential,
thermodynamic or reversible potential (E), and standard potential
(E0) are the most widely used, and all refer to chemical reactions
involving electron transfer having a definite associated potential.
From a critical point of view, these potentials indicate different
situations. Oxidation/reduction and electrochemical potentials are
general terms. Reversible, standard, and rest ("open circuit")
potentials refer to potentials measured when the system is believed
to be at equilibrium at known state of conditions of ionic activity,
pressure, and temperature. These can be related to thermodynamic
properties of reactions
Page 161
occurring in the system and at unit ionic activity, P = 1 atm, T =
25°C, the measured potential is the standard potential, and Eh
implies that the measured/calculated potential is converted to the
Standard Hydrogen Electrode (SHE).
In mineral chemistry, it is important to appreciate other frequently
used terms such as solution, mixed, galvanic, and pulp potentials.
These potentials, unlike the first set of potentials, refer to a
nondynamic situation. The equilibrium state is never fulfilled in
actual flotation, and the potentials measured indicate that a steady
value is a result of various reactions proceeding at constant rates.
Thus, the so-called mixed potential arises when the potential of the
system/mineral electrode is determined from more than one redox
reaction (i.e., complex system), whilst the galvanic potential refers
to potential between conducting solid phases such as metals and
semiconductors. The pulp potential is perhaps the widely accepted
description of the measured potential in mineral slurries, which is
in most cases a compromise between the galvanic potentials of the
minerals and the solution potential. However, in order for all
measured potentials to be meaningful values that can be
interpreted, it is very important to relate them to some standard or
reference electrodes such as gold, platinum, silversilver chloride,
saturated calomel, or SHE (18,19). Probably the most effective
would be platinum, although gold electrodes have proved to be
more effective elsewhere (20). In actual mineral pulp, it may
appear that the electrode made from the mineral being concentrated
may be the most appropriate to measure Eh. However, due to
oxidation surface layers, which may form on the mineral surface,
the electrode may cease to be effective.
A
Principles of EhpH Diagrams for Sulfide Mineral Flotation
In well-defined systems, the measured rest "open circuit" potentials
(Eh), as they are vaguely referred to, may provide valuable
information on the identity of the reactions occurring at the
electrode/solution interface when interpreted with caution. One
approach is to compare the measured potential with the equilibrium
potentials for various processes derived from various
thermodynamic data. It is usual to express these equilibrium
potentials for a system in the form of an EhpH diagram as derived
by Pourbaix (21) for metal systems and adapted to minerals by
Garrels and Christ (22).
The use of these diagrams has proved to be a convenient and useful
method to present free-energy data in relation to
hydrometallurgical processes and mineral processing. The use of
EhpH diagrams, with only stable phases to be directly applied to
mineral processing, has always been questioned because of the
relatively short time scales, which makes kinetic limi-
Page 162
tations favor the formation of metastable phases. Only in a few cases
have the problems of metastable phases been addressed in the
construction of EhpH diagrams in relation to processing of sulfide
minerals (2329). In the majority of those cases, computer-based
methods are used that permit mass balance to be attained.
One of the most widely accepted programs is the so-called
SOLGASWATER computer program based on the free-energy
minimization technique (30). This program is believed to be more
flexible than those based on mass balance techniques such as those
reported before by Basilio et al. (24) and Pritzker and Yoon (28,29)
and is believed to handle up to 80 species with a maximum of 48
solid phases. These calculations and investigations led Woods et al.
(31) to establish that the stability regions of various copper sulfides
depended on the copper-to-sulfur ratios. Figure 1 shows such a
Figure 1
Eh-
pH diagram of chalcocite(Cu2S) considering metastable copper sulfide and not
containing sulfur-oxy anions (31).
Page 163
Eh-pH diagram for chalcocite as prepared by Woods et al. (31)
considering metastable copper sulfides and not considering sulfur-
oxy anions. This figure is consistent with the behavior of chalcocite
with its voltammetric behavior reported before (24).
The establishment of similar diagrams for other sulfide minerals
promises to better the understanding of most electrochemical
observations in these mineral systems. These diagrams are
sometimes drawn as pe-pH where pe = -log (electron activity) is
hypothetically used to measure the relative tendency of a solution to
accept or transfer electrons. High positive pe values indicate a
relatively strong tendency for oxidation, whilst high negative pe
values mean a strong tendency for reduction. The pe value is related
to the redox potential Eh vs. SHE in one of the works of Wang and
Forrsberg (32) through the following equation:
It should be appreciated that these diagrams are much the same,

differing only by a factor.
The use of equilibrium potentials to predict electrochemical
reactions or the composition of a mineral on the basis of EhpH
diagrams may be a useful tool. However, the main weaknesses of
this method are: first, the calculated equilibrium potentials are
derived from thermodynamic properties of bulk phases under
conditions where multilayers may be possible to form, a case that is
not applicable to flotation; and second, many reactions involving
sulfides are irreversible (26).
Thus, a process may occur not because it has a more favorable
thermodynamic potential, but due to its faster kinetics. This is
particularly important for a complex reaction for which the potential
has not been established experimentally. It is therefore not logical to
draw firm conclusions on a reaction on the basis of measured
potentials. Allison et al. (33) and Goold and Finkelstein (34) were
probably the first to realize this and correlated the measured
potentials with products of reaction extracted from the mineral
surface. Although the method suffered from various interpretation
problems depending on sample preparation (3537), the procedure
demonstrated the need to make such comparisons.
IV
Collectorless Flotation
Whether a sulfide mineral surface is naturally hydrophobic is still
open to discussion. It was generally accepted that, apart from
minerals such as molybdenite, orpiment, realgar, and stibnite, pure
and clean sulfide minerals are hydrophilic (38). The property of such
natural floatability is attained when the spreading coefficient for
water on the mineral surface is negative.
Page 164
However, Miller (39) argued that sulfide minerals are
thermodynamically unstable and as such are easily oxidized to
sulfurlike surface products first proposed by Wark (40) and later by
many other researchers and believed to be responsible for
collectorless flotation of sulfide minerals. However, this theory has
since been refined in light of recent electrochemical and surface
characterization studies (41,42) which have identified metal-
deficient sulfide layers. As early as 1932, Gaudin (43) proposed
that native floatability of sulfide minerals may be due to the
resulting hydrophobic surface layers formed after rapture of van
der Waals forces. However, after assessing the various arguments
advanced by different researchers (4446), it is probably logical to
conclude that the reasons for the collectorless floatability of sulfide
minerals may largely depend on the mineral type and may vary
from mineral to mineral. Other researchers have demonstrated that
the preparation conditions and methods also play a very important
role in this natural floatability, whilst other documented evidence
points to the role of surface roughness and defects (4750). From
published works, it is possible to rank minerals in the approximate
order of ease of collectorless flotation. Table 1 lists the order of
floatability of some minerals and their rest potentials in water at pH
4 (18). Although this list is always looked at with suspicion, it,
however, appears that the more cathodic or reducing the rest
potentials and the less easily oxidizable the sulfide mineral is, the
more readily it floats without a collector (51). Figure 2 illustrates
an example of collectorless flotation of chalcopyrite as a function
of Eh when it is mixed with quartz (16). It shows that the mineral
floated at oxidizing Eh values but
Table 1 Rest Potentials of Sulfide
Mineral in Water at pH 4 Arranged
in an Approximate Order of Their
Easy Collectorless Floatability.a
Rest Potential
Mineral vs. SHE
Molybdenite, MoS2 0.11
Stibnite, Sb2S3 0.12
Argentine, AgS 0.28
Galena, PbS 0.40
Bornite, Cu5FeS4 0.42
Covellite, CuS 0.45
Sphalerite, ZnS 0.46
Chalcopyrite, CuFeS2 0.56
Marcasite, (Zn,Fe)S2 0.63
Pyrite, FeS2 0.66
aFrom Ref. 50.
Page 165
Figure 2
General relation between applied potential and recovery of chalcopyrite (16).
not in reducing atmosphere. The presence of pyrite enhances the
oxidation of other sulfides but may deteriorate their floatability for
which a well-understood galvanic mechanism is responsible (52).
The Eh and pH effect to promote collectorless flotation is an
excellent invention from Guy and Trahar (7). Through a good
control of Eh and pH, the collectorless floatability of different
minerals can be appreciably improved and then can be used as a
preconcentration stage. The function of collectors such as
dithiocarbonates would seem to increase the range (in terms of Eh
and pH) under which a hydrophobic layer is formed on the surface
(53). The major finding of research directed toward investigations
pertaining to good sulfide flotation without collectors is that
galvanic coupling between sulfide minerals of different
electrochemical properties can adversely influence the collector-
induced flotation separation of these minerals (54). These were the
observations of Nakazawa and Iwasaki (55,56) and the extended
works on galvanic coupling between sulfide minerals and the
grinding media (55,57). The recent work of Hayes and Ralston (58)
has also confirmed the early findings of Guy and Trahar (16) for the
chalcopyrite/galena that mineral selectivity is reduced when the
minerals are ground
Page 166
together in an oxidizing environment but is restored when ground
in a reducing environment. Gebhart and Richardson (59) have
reported similar results for a chalcocite/pyrite system.
V
Cathodic Reduction of Oxygen at Sulfide Surfaces
The reduction of oxygen is one of the most important reactions in
various processes involving sulfide minerals. It plays a major role
in the weathering of sulfides in nature (60,61), in oxidation during
mining, storage, and transportation (62), and in a number of
leaching systems (63).
The electrocatalytic activity for oxygen reduction on various
sulfide minerals depends on the type of sulfide mineral, with
Linnaeite (Co3S4) being the most electroactiveslightly above pyrite
(64). The general notion is that the reduction proceeds through an
intermediate hydrogen peroxide stage, which is believed (65) to
play an important role in the flotation of these minerals at least in
the formation of perxanthate (66,67), a compound that is believed
not to have collecting properties and that should be very much
avoided to reduce the consumption of useful xanthate.
VI
Various Electrochemical Techniques
A summary of the various useful electrochemical techniques in
sulfide mineral research has been outlined by Chander (68). They
include the potentio-dynamic techniques such as Linear Potential
Sweep Voltammetry, Small Amplitude Cyclic Voltammetry
(SACV) (69), and Cyclic Voltammetry (CV), which have been
extensively used to study sulfide mineral flotation. These
techniques are particularly useful in conducting preliminary
investigations of the reduction/oxidation reactions. The principle
involved is the application of a potential to a working electrode
(mineral) linearly through a potentiostat with a ramp generator. The
potential varies linearly with time in single sweep voltammetry and
as a triangular wave in CV. The current flow is monitored as a
function of time and is often plotted against the potential to obtain
the so-called voltammogram; a typical example is shown in Fig. 3
(11). The potential at which a current peak is registered can be used
to predict a possible chemical reaction, and the magnitude of the
peak is an indication of the surface concentration of the species
undergoing a chemical reaction. From the shape of the peak and the
dependence of both the peak potential and the peak current on the
sweep rate, details about the reaction mechanisms can be
elucidated (11). In order to identify the mechanism due to
voltammetric behavior, the following parameters are important to
determine:
Page 167
Figure 3
Voltammograms for chalcopyrite in pH 11 buffer solution.
Linear sweeps at 4 mV are taken from rest to one limit
followed by 1 1/2 cycles. Initial scan positive going in A
and negative going in B.
(From Ref. 12.)
1. the potential at which each electrode reaction commences,
2. the charge passed in these processes
3. the effect of pH on the potential at which various reactions take
place
4. the effect of rotating the electrode to disperse soluble reaction
products
When cyclic voltammetry is used to study a system for the first
time, qualitative tests are initially done and then followed by
semiquantitative and finally quantitative analyses to determine the
kinetic parameters. This is usually done by using a wide range of
sweep rates and various values of initial and final potentials. It is
common to obtain several peaks, and by a careful scrutiny of the
appearance and disappearance of the peaks within different
potential ranges and sweep rates, and through the comparison of
the first and subsequent cycles, it is possible to relate the process
with the peaks. From these observations, it is viable to identify
adsorption, diffusion, and couple homogeneous chemical reactions.
Information can also be
Page 168
obtained on the reversible nature of an electron transfer process,
the number of electrons transferred in the rate-determining step,
and the degree of coverage or the amount of product formed.
The electrochemical characteristics as studied by CV or LPSV can
be done in a conventional two/three electrode cell, which may or
may not have a counter electrode through which current flows, a
reference electrode, and a working electrode, which is the material
to be studied. The most common practice in sulfide research is to
fabricate a working electrode from the purest grains of the mineral
itself using the procedure of spark erosion to cut the electrode as a
mineral block/core similar to that reported before (71) to enable
polishing of the surface. However, care should be exercised not to
induce an electrical resistance that is too high due to various
connections. In some cases mercury has been recommended as the
conducting media between the mineral and potentiostat.
However, in all cases, it is important to realize that the design of
working electrodes depends on the problem to be studied. They can
vary from small spheres and discs to metal foils or single crystals
of semiconductors such as sulfides or even powdered beds (as the
popular fluid/packed bed electrodes are called), which is probably a
much more practical method in relation to electrochemical flotation
research (67,72).
Fairly recently intermittent galvanostatic polarization (IGP) was
used by various researchers in investigating further the
electrochemical nature of sulfide minerals (73). In this method, the
galvanostatic current is repeatedly interrupted at equal short
intervals to separate the capacitive and faradaic contributions.
However, the main drawback of this method is to relate the current
and potential to the actual electrochemical process occurring. Small
amplitude triangular potential technique was used by Hu et al. (74)
to reevaluate the diethyldithiocarbonate-galena flotation system in
which the interfacial capacitance and resistance were measured and
related to collector coverage in which a good agreement was
obtained with mathematical calculations.
The other probably not very widely used technique is AC
impedance spectroscopy, which was recently used by Chander and
Briceno (73,7678) to study the properties of surface films. In this
technique, impedance of the electrode/solution interface is
measured as a function of angular frequency and is drawn as a bode
plot. It is possible to characterize the stages where uniform and
nonuniform oxidation layers may form using this information.
VII
Surface Characterization Techniques
Various methods for surface characterization are all basically
microanalytic techniques. These could be classified in three distinct
categories depending
Page 169
on the source of surface excitation elements, which give a signal
interpreted in terms of energy units. The techniques based on ion
beams include SIMS and ISS whose signals come from secondary
and primary ions, respectively. ESCA (UPS/XPS), LAMMA, and
MOLE (LAMAN) use photoelectrons, ions, and raman photons,
respectively. EPMA employs X-rays; and AES, TEM, and
cathodoluminescence uses electrons and photons, respectively. A
detailed summary of the various features of these techniques are
found in a review by Barbery (53). In most cases, these techniques
have been used either on their own or together with other
techniques. XPS, like CV, is the most favored combination
technique, and it has been used by a number of researchers. The
details about the operation of XPS and the wide applicability of the
technique can be found in an excellent review by Cecile (79). The
energy spectra of X-ray stimulated electrons are composed of photo
peaks, which are characteristic of atomic orbital energy, which
allows a qualitative analysis of surfaces. However, most of these
techniques are not in general useful for in-situ characterization of
surfaces of particles in solutions. In a recent publication,
Mielczarski and Yoon (80) used infrared reflectionabsorption
spectroscopy (IRAS) and infrared attenuated total reflection (ATR)
for in situ measurements to characterize collector orientation on the
mineral surface. The outstanding questions of the nature of sulfide
surface on which collector adsorption takes place or which role is
played by oxidized surface on collector adsorption seem to be
handled well by the newly developed techniques of FTIR and
DRIFT spectroscopic techniques. FTIR is based on the Michelson
interferometers and has many advantages over the old grating
instruments (81). The main advantage of DRIFT is its ability to
detect very low concentrations of various compounds. The need to
accurately define surfaces of sulfide minerals has, indeed, been
revolutionized with the coming of these various apparatus, which
may be used in combination with other electrochemical techniques
as described in the works of Leppinen et al. (82). In this work the
combination of ATR, FTIR, and cyclic voltammetry led to a better
understanding of the surface xanthate compounds on chalcocite,
chalcopyrite, pyrite, and galena. The choice of the technique is
determined by various things such as required information, the
nature of the specimens to be used, and the flexibility of the
technique as well as the cost of the auxilliary and accessory
components.
VIII
Sulfide Flotation Using Gases
The use of gases other than air in mineral-processing plants was
probably first reported by Lin and Metzer (83) and later by
Podobnik and Shirley (84). Nitrogen, however, has won favor, due
to its abundance and chemical inactivity. Today, nitrogen is used in
chalcopyritemolybdenite separation
Page 170
(85,86). The role of nitrogen is here believed to either suppress the
consumption of some specific reagents (e.g., hydrogen sulfide) or
depress other various chemically coupled reactions (87).
The presence of nitrogen in flotation pulps would be expected to
reduce the pulp oxygen content, which in turn decreases the pulp
potential. Both effects should ideally play a decisive role in the
flotation of sulfide minerals. The flotation of pyrite in complex ore
has been shown to decrease considerably when floated with
nitrogen (88). However, this is true only when nitrogen is
introduced throughout the grinding stage to flotation or else when it
is applied only at the flotation stageits floatability is reactivated. It
is believed that this is due to the weakening of the electron activity
on pyrite when coupled to other sulfide minerals so that the
preferential oxidation of air (oxygen) to form hydroxyl ions, which
depresses pyrite, does not take place. It is felt amongst many
researchers that, probably with the introduction of columns in the
mineral industry, this trend to nitrogen flotation of sulfide minerals
will win popularity.
Sulfur dioxide has been known for a long time to be used in
differential flotation of sphalerite at Broken Hill Mines (89). Today,
its use is even broader, and in certain cases it is used in
combination with depressing agents to achieve the required goals
(9092). The Japanese are on the forefront in the use of sulfoxy
gases, where they are extensively used depending on the
complexity of the ore (93,94). The role of sulfur dioxide in the
depression of sulfide minerals like sphalerite and its pulp
modification role is still an area of heated debate. Various
mechanisms (9597) have been postulated such as:
1. Formation of surface deposits such as zinc sulfite or calcium
sulfite (when quick lime is used as a pH modulator).
2. Formation of insoluble heavy metal sulfite salts and subsequent
reduction in heavy metal ions concentration in pulp, which
prevents unintentional activation.
3. Xanthate decomposition and its reduced adsorption.
Although the work of Peres (95) showed that the depression action
of SO2 is not appreciable for chalcopyrite and other copper-bearing
minerals (98), the actual depression mechanism is still unclear.
However, what appears to be certain is that SO2, like N2, plays a
decisive role in the flotation of sulfide minerals probably through
an electrochemical route. It is important to note that the presence of
both SO2 and N2 may result in a totally different pulp environment,
which may adversely change the floatability of various minerals. In
a particular case, the addition of SO2 resulted in reduced frothing;
in other situations frothing was nonexistent. Furthermore, it may be
necessary to change collectors from, say, xanthates, which easily
decom-
Page 171
pose to dithiophosphines. The use of modifiers such as cyanide,
sodium sulfide, and zinc/copper sulfate also may not be required
favoring the economics of the process and a better environmental
working atmosphere (99).
IX
Discussion
The dependence of the floatability of most sulfide minerals on the
electrochemical conditions as established from laboratory and pilot
plant tests goes to support the contention that it may be possible to
control most sulfide flotation electrochemically. The use of
electrochemical principles in industrial plants has been successfully
applied in various plants in Finland (100,101), Australia, and
Canada. There are various independent factors that have to be
borne in mind when carrying out such an ambitious step. The
technology requires the use of industrial scale units for the control
of pulp potential and will be determined by, among other factors,
economic merits. Apart from a new blend of Process Engineers,
there will be a need to adopt units as electrochemical reactors,
which may have to work on the principle of fluid/packed bed
electrodes. Industrial and safe potentiostats may not be easy to
handle and control for existing steel-based structure mineral-
processing plants. The problem of coupled pH and Ep control will
have to be addressed seriously as the modifier may have both
effects. The use of nitrogen/oxygen/sulfur dioxide will have
occupational health and safety hazards, and it will be necessary to
recycle gases. On the other hand, the use of various modifiers may
not be necessary. Furthermore, it might be apparent that different
sections of the mineral-processing plants may require the use of
different gases of different purity, which may contribute to a rise in
operational and investment costs. Integrated Mine/Mill/
Smelter/Refinery complexes will in this respect provide the best
setup as opposed to separate plants. This principle is already being
tried in some plants in Canada where by-product nitrogen from the
smelter is used in the mineral-processing plants.
X
Summary
The complex nature of reactions in sulfide mineral pulps is outlined
and the various electrochemical techniques for elucidating these
reaction mechanisms are identified. The role of both pH and Eh and
the use of Pourbaix diagrams and their limitations is discussed and
related to the collectorless floatability of sulfide minerals in
general. Different surface characterization techniques and their
advantages over others in sulfide mineralcollector systems are
summarized. Finally, the use of various gases other than air
Page 172
has been discussed in terms of both the advantages and
disadvantages of such a replacement.
References
1. J. O'M. Bockris and A. K. N. Reddy, Modern Electrochemistry,
Vol. 1, Plenum, New York, p. 3 (1973).
2. S. G. Salamy and J. C. Nixon. The application of
electrochemical methods to flotation research, Recent
Developments in Mineral Dressing, Int. Min. Metallurgy, London,
pp. 503516. (1953).
3. I. N. Plaskin and S. V. Bessonov, ''Role of Gases in Flotation
Reactions," Proceedings of the 2nd Int. Congress on Surface
Activity, Vol. 3, pp. 361367 (1957).
4. A. M. Gaudin, Flotation. 2nd ed., McGraw Hill, New York,
Chap. 9. (1957).
5. P. J. Guy and W. J. Trahar, The effects of oxidation and mineral
interaction of sulfide flotation, Developments in Mineral
Processing, Vol. 6 (K. S. E. Forrsberg, ed.), Elsevier, Amsterdam.
(1985).
6. G. W. Heyes and W. J. Trahar, Oxidation-reduction effects in the
flotation of chalcocite and cuprite. Inter. J. Miner. Process., 6:
229252. (1979).
7. P. J. Guy and W. J. Trahar, The influence of grinding and
flotation environments on the laboratory batch flotation of galena,
Int. J. Miner. Process., 12: 1538.(1984).
8. N. D. Janetski, S. I. Woodburn, and R. Woods, An
electrochemical investigation of pyrite flotation and depression,
Int. J. Miner. Process., 4: 227239. (1977).
9. I. C. Hamilton and R. Woods, An investigation of surface
oxidation of pyrite and pyrrhotite by sweep voltammetry,
Electroanal. Chem., 118: 327343. (1981).
10. P. E. Richardson and E. E. Maust, Jr., Surface stoichiometry of
galena in aqueous electrolytes and it's effect on xanthate
interaction, Flotation A. M. Gaudin Memorial Vol. 1 (M. C.
Fuerstenau, ed.), Am. Inst. Min. Metall. Eng., pp. 364392. (1976).
11. J. R. Gardner and R. Woods, A study of the surface oxidation of
galena using cyclic voltammetry, J. Electro. Anal. Chem., 100:
447459. (1979).
12. J. R. Gardner and R. Woods, An electrochemical investigation
of the natural floatability of chalcopyrite, Int. J. Miner. Process., 6:
116. (1979).
13. G. W. Walkers, J. V. Stout III, and P. E. Richardson,
Electrochemical flotation of sulfides: Reactions of chalcocite in
aqueous solution, Int. J. Miner. Process., 12: 5572. (1984).
14. T. Warner, "Electrochemical studies of violalite and pentlandite
by cylic voltammetry, chronopotentiometry and
chronoaperometry," Ph.D. Thesis, University of Leeds, 1988,
unpublished.
15. L. K. Witika and W. C. Lombe, Electrochemical behaviour of
carrollite (CuCo2S4) in aqueous buffer solutions, Trans. Int. Min.
Metall. Sec. C: Extra. Metall., Min. Process., C26C32,
JanuaryApril. (1989).
Page 173
16. G. W. Heyes and W. J. Trahar, The natural floatability of
chalcopyrite, Int. J. Miner. Process., 4: 317344. (1977).
17. R. K. Clifford, "Mechanisms of flotation of natural sphalerite
with xanthate." Ph.D. Thesis, University of Utah, Salt Lake City,
1972.
18. J. Ralston, Eh and its consequences in sulfide mineral flotation,
Min. Eng., 4 (711): 859878. (1991).
19. N. W. Johnson et al., Oxidation-reduction effect in galena
flotation: Observation on PbZnFe sulfide separation, Trans. Inst.
Min. Metall., 91, C32C37. (1982).
20. D. A. J. Rand and R. Woods, Eh measurements in sulfide
mineral slurries, Int. J. Miner. Process., 13: 2942. (1984).
21. M. N. J. Pourbaix, Atlas of Electrochemical Equilibria,
Gauthiers-Vilars, Paris. (1963).
22. R. M. Garrels and C. L. Christ, Solutions, Minerals and
Equilibria, Harper and Row, New York. (1965).
23. K. S. E. Forrsberg, B. Antti, and B. I. Palsson, Computer
assisted calculations of thermodynamic equilibria in the
chalcopyrite-ethyl xanthate system, Reagents in the Minerals
Industry, IMM, London, pp. 251264. (1984).
24. C. Basilio, M. D. Pritzker, and R. H. Yoon, "Thermodynamics
Electrochemistry of the ChalcocitePotassium Ethyl Xanthate
System," 114th Annual Meeting, New York, Reprint No. 8586.
(1985).
25. E. Peters, Leaching of sulfides, Advances in Mineral
Processing (P. Somasundaran, Ed.), SME/AIME, Littleton, CO, pp.
445562. (1986).
26. E. Peters, "Electrochemical Mechanism for Decomposition of
Sulfide Minerals," Proceedings of the International Symposium on
Electrochemistry of Minerals and Metal Processing, Vol. 10 (P. E.
Richardson, S. Srinivasan, and R. Woods, eds.), The
Electrochemical Society, Pennington, NJ, 343361. (1984).
27. M. D. Pritzker, R. H. Yoon, C. I. Basilio, and W. Z. Choi,
Solution and flotation chemistry of sulfide minerals, Can. Metall.
Q., 24: 2738. (1985).
28. M. D. Pritzker and R. H. Yoon, Thermodynamic calculations
on sulfide flotation systems. I. Galena-ethyl xanthate system in the
absence of metastable species, Int. J. Miner. Process., 12: 95125.
(1984).
29. M. D. Pritzker and R. H. Yoon, "Thermodynamic Calculations
and Electrochemical Studies on the GalenaEthyl Xanthate System,"
(P. E. Richardson, S. Srinivasan, and R. Woods, eds.). Proceedings
of the International Symposium on Electrochemistry in Mineral
and Metal Processing, Vol. 10, The Electrochemical Society, pp.
2653. (1984).
30. G. A. Eriksson, An algorithm for the calculation of aqueous
multicomponents, multiphase equilibria, Anal. Chim. Acta, 112:
375383. (1979).
31. R. Woods, R. H. Yoon and C. A. Young, Eh-pH diagrams for
stable and metastable phase in the coppersulfurwater system, Int. J.
Miner. Process., 20: 109120.(1987).
32. X. H. Wang and K. S. E. Forrsberg, Mechanism of pyrite
flotation with xanthates, Int. J. Miner. Process., 33: 275290.
(1991).
33. S. A. Allison, L. A. Goold, M. J. Nicol, and A. Granville, A
determination of the products of reactions between various sulfide
minerals and aqueous
Page 174
xanthate solutions and a correlation of the potential, Metall.
Trans., 3: 26132618. (1972).
34. L. A. Goold and N. P. Finkelstein, The reactions of sulfide
minerals with thiol compounds, National Institute for Metallurgy,
Republic of South Africa Report No. 1439 (1972).
35. R. C. Greener, An infrared investigation of xanthate adsorption
on lead sulfide, J. Phys. Chem., 66: 877886. (1962).
36. A. M. Gaudin and W. D. Wilkinson, Surface action of some
sulfur bearing organic compounds on some finely ground sulfide
minerals, J. Phys. Chem., 37: 833845.(1933).
37. A. S. Allison and N. P. Finkelstein, Study of the products of
reaction between galena and aqueous xanthate solutions, Trans.
Inst. Min. Metall., 10: C223239.(1972).
38. J. Rogers, Froth Flotation Fiftieth Anniversary Volume (D. W.
Fuerstenau, ed.), AIME, pp. 139169. (1962).
39. J. D. Miller, "The Significance of Electrochemistry in the
Analysis of Mineral Processing Phenomena," 7th Australian
Electrochemistry Conference, Sydney, Australia, February 1419,
1988.
40. I. W. Wark, Principles of Flotation, Australasian Institute of
Mining and Metallurgy, Melbourne, p. 125. (1938).
41. R. Woods, A. N. Buckley, and I. C. Hamilton, Investigation of
the surface oxidation of sulfide minerals by linear potential sweep
voltammetry and X-ray photoelectron spectroscopy, Flotation of
Sulfide Minerals, Development in Mineral Processing, Vol. 6 (K. S.
E. Forrsberg, ed.), Elsevier, Amsterdam, pp. 4160. (1985).
Technology (P. Somasundaran and B. M. Moudgil, eds.), Marcel
Dekker, New York, pp. 3978. (1988).
43. A. M. Gaudin, Flotation, McGraw-Hill, New York. (1932).
44. W. J. Trahar, The influence of pulp potential in mineral
flotation, Principles of Flotation: A Wark Symposium (J. T.
Woodcock and M. H. Jones, eds.), AUIMM, pp. 117135. (1983).
45. N. P. Finkelstein, S. A. Allison, V. M. Lovell, and B. V.
Steward, Advances in inter facial Phenomena of
paticulate/solution/gas System, (P. Somasundaran, and R. B.
Grieves, eds.), AIME, pp. 165175. (1975).
46. M. C. Fuerstenau and B. J. Sabasky, Int. J. Miner. Process., 8:
79. (1981).
47. W. J. Trahar, A laboratory study of the influence of sodium
sulfide and oxygen on the collectorless flotation of chalcopyrite,
Int. J. Miner. Process., 11: 57. (1983).
48. P. J. Harris and R. Richter, The influence of surface defect
properties on the activation and natural floatability of sphalerite,
Flotation of Sulfide Minerals, Developments in Mineral Processing
(E. K. S. Forrsberg, ed.), Elsevier, Amsterdam, pp. 141158. (1985).
49. J. F. Oliver, C. Huh and S. G. Mason, An experimental study of
some effects of solid surface roughness on wetting, Colloids Surf.,
1: 79. (1980).
Page 175
50. R. H. Dettre and R. E. Johnson. Adv. Chem. Ser., 43. (1964).
51. H. Majima, How oxidation effect selective flotation of complex
sulfide ores, Can. Metall. Q., 8: 269273. (1969).
52. D. A. J. Rand, Oxygen reduction on sulfides: Part II
Comparison of activities of various copper, iron, lead and nickel
mineral electrodes, J. Electroanal. Chem., 83: 1932. (1977).
53. G. Barbery, Recent progress in the flotation hydrometallurgy of
sulfide ores, Flotation of Sulfide Minerals, Developments in
Mineral Processing (K. S. E. Forrsberg, ed.), Elsevier, Amsterdam,
pp. 140). (1985).
54. M. K. Yelloji Rao and K. A. Natarajan, Effect of
electrochemical interactions among sulfide minerals and grinding
media on the flotation of sphalerite and galena, Int. J. Miner.
Process., 29: 175194. (1990).
55. H. Nakazawa and I. Iwasaki, Galvanic contact between nickel
arsenide and pyrrhotite and it's effect on notation. Int. J. Miner.
Process., 18: 203215. (1986).
56. H. Nakazawa and I. Iwasaki, Effect of pyrite-pyrrhotite contact
on their floatability, Miner. Metall. Process., 2: 206211. (1985).
57. K. Adams and I. Iwasaki, Pyrrhotite-grinding media interaction
and it's effect on the floatability of different applied potentials,
Miner. Metall. Process., 1: 8187. (1984).
separation of sulfide minerals by Eh, control. Int. J. Miner.
Process., 23: 5584. (1988).
59. J. E. Gebhardt and P. E. Richardson, Differential flotation of a
chalcocite-pyrite bed by electrochemical control, Min. Metall.
Process., 4: 140145. (1987).
60. M. Sato, Oxidation of sulfide ore bodies: Oxidation
mechanisms of sulfide minerals at 25°C, Econ. Geol. Survey Bull.,
55: 12021231. (1960).
61. H. Tributsh and H. Gerischer, The oxidation and self heating of
metal sulfide as an electrochemical corrosion phenomena, J. Appl.
Chem. Biotechnol., 26: 747761. (1976).
62. E. H. Nickel, J. R. Ross, and M. R. Thornber, The supergene
alteration of pyrrhotite-pentlandite ore at Kambala, Western
Australia, Econ. Geol., 69: 93107. (1974).
63. E. Peters and H. Majima, Electrochemistry of sulfide
dissolution in hydrometallurgical systems, Proceedings of the 8th
Mineral Processing Congress, Leningrad, Vol. II, Inst. Mekhanobr,
Leningrad, pp. 517. (1969).
64. S. M. Ahmed, Electrochemical activity of galena, pyrite and
cobalt sulfide for the oxygen reduction in relation to xanthate
adsorption and flotation, Int. J. Miner. Process., 5: 163174. (1978).
65. T. Biegler, D. A. J. Rand, and R. Woods, Oxygen reduction on
sulfide minerals, Trends in Electrochemistry (J. O'M. Bockris, D.
A. J. Rand, and B. J. Welch, eds.), Plenum Press, New York, pp.
291302. (1977).
66. M. H. Jones and J. T. Woodcock, PerxanthateA new factor in
the theory and practice of flotation, Int. J. Miner. Process., 5:
285296. (1978).
67. P. E. Richardson, J. V. Stout III, C. L. Proctor, and G. W.
Walker, Electrochemical flotation of sulfide: Chalcocite-ethyl
xanthate interaction. Int. J. Miner. Process., 12: 7393. (1984).
Page 176
68. S. Chander, Electrochemistry of sulfide mineral flotation,
Miner. Metall. Process., 104. (August 1988).
69. J. R. Ross, J. P. Celis, and A. S. Sudarsono, Electrochemical
control of chalcocite and covellitexanthate flotation, Int. Miner.
Process., 29: 1730. (1990).
70. R. Greef, R. Peat, L. M. Peter, D. Pletcher, and J. Robinson,
Instrumental Methods in Electrochemistry, John Wiley, New York,
pp. 178228. (1985).
71. L. K. Witika and B. Dobias, "Rest Potential and Cyclic
Voltammetry Measurements in a CarrolliteXanthate System,"
Proceedings of the Internation Conference on Environment and
Mineral Processing (P. Fecko *, ed.), Ostrava. (1992).
72. G. W. Walker, J. V. Stout III, and P. E. Richardson,
Electrochemical flotation of sulfides: Reactions of chalcocite in
aqueous solution, Int. J. Miner. Process., 12: 5572. (1984).
73. S. Chander and A. Briceno, "The Nature of Surface Films on
Pyrite and Their Role in Flotation," 117th SME Annual Meeting,
January 2528, Phoenix, AZ (1988).
74. Q. Hu, D. Wang, and B. Li, An electrochemical investigation of
the diethyldithiocarbamategalena flotation system, Int. J. Miner.
Process., 34: 289305. (1992).
75. A. Granville, N. P. Finkelstein, and S. A. Allison, Review of
reactions in the flotation system galenaxanthateoxygen, Trans. Inst.
Min. Metall., Sec. C: Extra. Metall. Miner. Process., 81: C1C30.
(1972).
76. S. Chander and J. E. Gebhardt, Application of electrochemical
techniques in sulfide mineral flotation. Challenges in Mineral
Processing (K. V. S. Sastry and M. C. Fuerstenau, eds.), SME,
Littleton, CO, pp. 90108. (1989).
77. R. Woods and P. E. Richardson, The flotation of sulfide
minerals: electrochemical aspect. Advances in Mineral Processing
(P. Somasundaran, ed.), SME, Littleton, CO, pp. 137153. (1985).
78. R. Woods, Electrochemistry of sulfide flotation, Flotation, A.
M. Gaudin Memorial Volume (M. C. Fuerstenau, ed.), AIME, New
York, pp. 298333. (1976).
79. J. L. Cecile, Application of XPS in the study of sulfide mineral
flotationA review, Flotation of Sulfide Minerals, Developments in
Mineral Processing Vol. 6 (K. S. E. Forrsberg, ed.), Elsevier,
Amsterdam, pp. 6180. (1985).
80. J. A. Mielczarski and R. H. Yoon, "Spectroscopic Analysis of
Collector Adsorption," Proceedings of Engineering Foundation
Conference on Flocculation and Dewatering, January 1015, Palm
Coast, Florida (1988).
81. P. F. MacMillan and A. M. Hofmeister, Infrared and Raman
spectroscopy, Reviews in Mineralogy, Vol. 18 (F. Hawthorne, ed.),
pp. 98159. (1988).
82. J. O. Leppinen, C. I. Basilio, and R. H. Yoon,
"Spectroelectrochemical Study of Ethyl Xanthate Adsorption on
Sulfide Minerals," Proceedings of 173rd Meeting of the
Electrochemical Society, Atlanta, Georgia, May 1520 (1988).
83. I. J. Lin and A. Metzer, Effect of dissolved hydrocarbon gases
on froth flotation, AlChE Symp. Ser.74(150): 7580. (1975).
Page 177
84. D. M. Podobnik and J. F. Shirley, Molybdenite recovery at
Cuajone, Min. Eng., 14731476. (1982).
85. M. A. Redfearn, "The Role of Nitrogen in the Flotation of By-
Product Molybdenite at Gibraltar Mines," SME-AIME Annual
Meeting, Atlanta, Georgia, Preprints No. 83-64 (1983).
86. K. Y. Onstott and P. L. Person, Society of Mining Engineers,
Preprints No. 84-65 (1984).
87. C. J. Martin, S. R. Rao, J. A. Finch, and M. Leroux, Complex
sulfide ore processing with pyrite flotation with nitrogen. Int. J.
Miner. Process., 26: 95110, (1989).
88. I. Sandoval-Caballero, M. Leroux, R. S. Rao, and J. A. Finch,
Testing nitrogen flotation of pyrite in mini-cells, Min. Eng., 3:
9373. (1990).
89. L. Bradford, "The Bradford Sulfur Dioxide Process," Austr.
Pat. No. 11253, (1913).
90. G. O. Deshler and T. R. Herndon, New techniques in
copperlead separation, Eng. Min. J. 172(5): 67. (1946).
91. R. W. Mathews and D. Raicevic, Sulfur dioxide as a
replacement for sodium sulfide in sulfide flotation, CIMM Bull.:
April 26. (1960).
92. F. W. McQuiston, "Flotation of Complex CopperLeadZinc
Ores," Progress in Mineral Dressing, Trans. Inst. Miner. Dressing
Congress, pp. 511 (1957).
93. T. Yonezawa, World Mining, 20(4): 58. (1967).
94. T. Yonezawa, "General Survey on Processing Techniques of
'Kuroko' (Black ore)," Joint Meeting of Min. Metall. Inst. Japan.
AIME, Print No. T111d2 (1972).
95. A. C. E. Peres, "The interaction between xanthate and sulfur
dioxide in flotation of nickelcopper sulfide ores," Ph.D. Thesis,
University of British Columbia, 1981, unpublished report.
96. G. R. Hyde, "Flotation Research of Avondale Research
CenterAn Introduction," Bureau of Mines information circular,
8818, p. 13 (1980).
97. S. K. De Kok, "Differential flotation of copper-zinc at Priska
Copper Mines," J. S. Afr. Inst. Min. Metall., 72: 305. (1972).
98. M. Misra, J. D. Miller, and Q. Y. Sony, The effect of sulfur
dioxide in the flotation of sphalerite and chalcopyrite, Flotation of
Sulfide Minerals, Developments in Mineral Processing, Vol. 6 (E.
S. K. Forrsberg, ed.), pp. 175194. (1985).
99. P. G. Broman, J. Hultqvist, and U. Marklund, "Experience from
the Use of SO2 to increase the Selectivity in Complex Sulfide Ore
Flotation," Technical Report from BOLIDEN. Presented at the
Workshop on Flotation of Sulfide Minerals, Stockholm, Sweden,
June 1984.
100. S. Heimala, M. Himmi, S. Jounela, and M. Saari, "The
Potential Controlled Flotation (OK-PCF) Method in Outokumpu
Nickel Concentrate Production," Proceedings of the 25th Annual
Conference of Metallurgists, Aug. 1720, Toronto (1986).
101. S. Heimala, S. Jounela, S. Ranapuska, and M. Saari, "New
Potential Controlled Flotation Methods Developed by Outokumpu
Oy," IVth International Mineral Processing Congress, III, Cannes,
France, pp. 8898 (1985).
Page 179
7
Complexation at the Metal SulfideWater Interface
Lars Rönngren and Staffan Sjöberg
University of Umeå, Umeå, Sweden
Zhong-xi Sun and Willis Forsling
Luleå University of Technology, Luleå, Sweden
I
Introduction
Sulfide minerals constitute an integral part of virtually all types of
rocks in the outer crust of the earth. This is a result of their
formation in magmatic, metamorphic, and sedimentary processes.
Under local conditions, the fraction of these minerals may rise to
almost 100%, and the resulting, so-called ores have been the
primary source of certain metals to mankind for centuries. Some of
these metals have even been named after places where early ores
were found, e.g., the metal copper is named from the island
Cyprus.
In nature, as well as in waste-water treatment and in modern
mineral separation processes, interactions between mineral surfaces
and water play a most important role. Thus chemical interactions
with sulfide mineral surfaces are likely to have a major impact on
the transport and deposition of heavy metals.
Sulfide surfaces have also been recognized to be very good
adsorbents for metal ions in reducing environments (1,2). Low
levels of mercury and lead in anoxic fluids are believed to be due
to their scavenging by sulfide particles (3). Furthermore, Brown (4)
and Brown et al. (5) have shown that iron sulfide minerals are very
good adsorbents for mercury(II)-complexes. They therefore
suggested the possibility of using iron sulfides to remove mercury
from waste waters. Above all, the abundance of metal sulfides in
Page 180
anoxic environments has led to increased interest in studying their
interracial chemistry.
Also in the mineral processing of sulfides, surface
complexation/precipitation reactions play an extremely important
role. Surface modifications of the sulfide minerals are a prerequisite
for efficient and selective flotation. It is, for instance, well known
that xanthate can be used as a collector; however, some naturally
occurring minerals like zinc sulfide are rendered floatable to
xanthate collectors by the adsorption of more active metal ions (e.g.,
Cu2+ , Pb2+ , Ag+) on the mineral surface. Furthermore, reagents
like HS- together with high pH have been found to be effective
depressants in the flotation of sulfide minerals (68).
In spite of all these very significant interactions between sulfide
minerals and the aqueous phase, it is true to say that basic
knowledge about the physical and chemical nature of the metal
sulfide/water interface still is relatively limited. The aim of the
present work is to contribute to the understanding of this interface.
To this end, it has been logical to use the much better characterized
metal oxidewater interface as a good starting point.
A
Surface Reactions at the Metal Oxide/Metal SulfideWater Interface
Adsorption at hydrous inorganic solids, in terms of surface
complexation models, has so far mainly dealt with metal oxides,
principally those of Si, Al, Fe, and Mn. It has been found that these
hydrous metal oxide surfaces can be described according to Figure 1.
The resulting surface hydroxyl groups have been found to represent
complex-forming donor groups which may coordinate metal ions or
metal complexes from the bulk solution. Similarly, the metal ion at
the oxide
Figure 1
Hydration of a metal oxide surface and formation of surface hydroxyl groups.
From Ref. 9.
Page 181
Figure 2
Hydration of a metal sulfide surface and formation of two different
surface groups ºSMeOH and ºMeSH.
surface may be treated as a Lewis acid; i.e., the OH group may be
replaced by other coordinating ligands. Several types of surface
species may be formed in this manner, including binary and ternary
complexes, as well as surface chelates of metal ions and ligands
(1012).
It seems reasonable to assume that similar complexation may take
place at the surface of hydrous metal sulfides. If so, one can expect a
hydrating mechanism as illustrated in Figure 2.
It has generally been found that hydrous oxide surfaces exhibit an
amphoteric character due to the formation of ºMeOH2+ or ºMeO-
groups. Amphoteric properties of the hydrated MeS surface can, by
analogy, be ascribed to the presence of ºSMeOH and ºMeSH binding
sites, presumably according to reactions 14 (13).
Due to the higher electronegativity and smaller size of O compared
to S, ºSMeOH will probably behave as a stronger base than ºMeSH.
Hence the protonation of the surface is ascribed to reaction (1), while
the deprotonation is ascribed to reaction (4).
Desorbed metal ions (13,14) as well as metal deficient surfaces
(15,16) have been reported to result at low pH. This makes it
reasonable to assume that an ion exchange process between protons
and metal ions is occurring on the metal sulfide surface.
Page 182
Theoretically it seems possible that sulfide ions might be exchanged
for hydroxide ions at the surface:
Adsorption of metal ions can occur due to either an ion exchange

process on the metal binding site (ºSMeOH) or surface complexation
at either the sulfide binding site (ºMeSH) or at the oxygen on the
metal binding site.
According to the literature, reaction (7a) occurs mainly when M

forms a metal sulfide that is less soluble than MeS(s) (14,17). If M
forms a metal sulfide that is more soluble than MeS(s), reaction (8)
or (9) is claimed to occur (1820). In Ref. 14 it was found that the
adsorption of metal ions is strongly pH-dependent even when
reaction (7a) occurred, a behavior explained by an ion exchange
process between the released metal ions and protons on the surface:
Inorganic and organic ligands have also been reported to form

complexes with the metal and sulfide binding sites (21) according to
reactions (10) and (11).
These ligands may also form complexes/solid phases with metal ions
released through ion exchange processes at low pH. This implies that
the relative amount of surface groups on a metal sulfide surface
might depend on an excess of surface-bound metal or sulfide ions.
On a stoichiometric surface, the concentrations of MeSH and
SMeOH should be the same.
However, since the surface is a small part of the total amount of
metal sulfide, a slight excess of metal or sulfide ions will produce a
nonstoichiometric surface. It should also be mentioned that oxidation
of the sulfide ions on the surface probably leads to a surface with an
excess of metal ions. This means that it is very important to have
nonoxidized samples to ensure surface stoichiometry.
B
Solubility Considerations
The majority of heavy metal sulfides are known to be scarcely
soluble in water (22). It is therefore possible to study their surface
properties over a wide pH-range.
Page 183
As can be seen from Figure 3, the solubility of metal sulfides shows
a considerable pH-dependence. At low pHs the solubility increases
due to a formation of HS- and H2S. At high pHs the solubility
increases due to a hydrolysis of the metal ion (mainly Zn(OH)2 and
Zn(OH)3- for ZnS and PbOH+ Pb(OH)2(aq), and Pb(OH)3- for PbS).
However, in the range 3 £ -log[H+] £ 11, the concentrations of metal
and/or sulfide ions in solution, due to dissolution, are negligible
provided that neither oxidation processes occur nor large excesses of
metal or sulfide ions are present.
At alkaline pH and with excess of sulfide ions, some metal sulfides
are dissolved due to the formation of thio complexes:
However, the only thio complex of Zn2+ and Pb2+ found in the
literature (23) is Zn(OH)(HS), but due to difficulties in separating
the aqueous complex from colloidal ZnS(s), the equilibrium constant
for this complex can be considered as very uncertain (24). By using
the recently estimated formation constant for this complex given by
Dyrssen and Wedborg (24), it was found that the possible influence
from it could be neglected.
Dissolution of metal sulfides can also occur due to oxidation of the
sulfide ion (25), which then results in the formation of many
different products such as polysulfides, elemental sulfur, thiosulfate,
sulfite, or sulfate according to reactions (13)(17).
As can be seen, reactions (13)(15) are highly pH-dependent. It
should also be mentioned that the oxidation of sulfides involves very
complex reactions and that many different products can occur (15,
26).
C
Aim of the Present Work
The aim of the present work has been to characterize the acid/base
and complexing properties of hydrous zinc and lead sulfide surfaces.
The interpretation of experimental data is made in terms of a surface
complexation model. This will make it possible to give a qualitative
and quantitative description of surface speciation as a function of
parameters such as pH and pS. The complexation studies have been
performed with special refer-
Page 184
Figure 3
The solubility of (a) ZnS (wurtzite)
(b) PbS (galena) as a function of -log[H+].

Page 185
ence to the ligands HS- and ethyl xanthate, all of vital importance
for the understanding of selective flotation of sulfide minerals.
The results are expected not only to find applications in the
flotation process of sphalerite and galena but also to throw some
light upon the significance of the interfacial chemistry of sulfide
particles in anoxic aquatic environments.
II
Experimental
A
Sample Preparations
1
Zinc Sulfide
Two different types of zinc sulfide samples used in this study: (a) a
synthetic ZnS (KEBO) and (b) a sample of well-ground sphalerite
mineral supplied by Boliden Minerals.
2
Lead Sulfide
Two different samples were used in the lead sulfide studies as well:
(a) a synthetic PbS prepared in the titration vessel by
stoichiometrically titrating a Pb2+ solution with a freshly calibrated
S2- solution. It was found that an error of more than 4% in the
determined concentrations would cause a surface that consisted of
only Pb2+ or S2-. (b) A sample of well-ground galena mineral
(from Wards Natural Science Establishment, Rochester, New
York).
The precipitated samples were rinsed several times under an inert
atmosphere, using pure ionic medium (0.1 M NaNO3 or 0.1M
NaClO4). After rinsing, a known amount of strong acid was added,
and the precipitate was allowed to ''age" overnight (12 h) while
stirring.
B
Sample Characteristics
The synthetic ZnS was found to consist of a mixture of sphalerite
and wurtzite. A foreign acid with pKa» 4.8 was found on its
surfaces. This acid, amounting to 0.02 × 10-3 mole g-1, was
corrected for in the calculations. The surface areas of the different
samples were measured by the BET method (27) and found to be
0.7 m2 g-1 (sphalerite) and 6.4 m2 g-1 (synthetic ZnS). In the
experiments the solid concentrations were 3040 g dm-3 and 60 g
dm-3 for the synthetic ZnS and the mineral, respectively.
The precipitated PbS was found to be microcrystalline as deduced
from x-ray powder diffraction and showed the same spectrum as
galena. The specific surface area was found to be 21 ± 2 m2 g-1
(precipitated PbS) (28) and 0.63 m2 g-1 (galena) using the BET
method. The solid concentration
Page 186
of the precipitated PbS was about 1012 g dm-3, whereas the
mineral concentration was 40 g dm-3 in the suspensions.
C
Apparatus
The potentiometric titrations were performed with an automatic
system for precise emf titrations designed and constructed by
Ginstrup (29). The cell arrangement, which was immersed in an oil
thermostat at 25.00 ± 0.05 °C, was similar to that described by
Forsling et al. (30). All the experiments were conducted in a room
thermostated at 25.0 °C. Ag/AgCl electrodes, prepared according
to Brown (31), were used as reference electrodes. For the [H+]
measurements, a glass electrode (Ingold type 201-NS) was
employed. The glass electrode was calibrated against a hydrogen
electrode within the range 3 £ -log[H+] £ 10.5. The S2- and Pb2+
concentrations in solution were monitored using an Ag/Ag2S
electrode (Orion 94-16) and a PbS/Ag2S electrode (Orion 94-82),
respectively. The total concentrations of soluble zinc and sulfur
were determined using an ARL 3410 ICP-AES spectrometer.
Calibrations were made with respect to aqueous standard solutions.
An x-ray powder diffractometer, Rigaku (D/MAX II A), was used
to characterize different solid samples. A Perkin-Elmer 1760x
FTIR-spectrometer with a TGS detector was used to record the
infrared spectra. The internal reflection method was used with an
ATR reflection accessory and a vacuum tight measurement cell
with a germanium reflection element according to Leppinen (28).
D
Special Precautions
Special precautions were undertaken to avoid oxidation of the
samples. Fine particles of the minerals galena and sphalerite were
prepared by grinding selected block crystals for about 8 h in an
agate mortar in an oxygen-free nitrogen box. The samples were
stored under an argon atmosphere. The argon gas was bubbled
through a vanadium (III, IV)- or a Na2S solution in order to
remove traces of oxygen.
E
Ionic Media
In order to keep the activity factors of reacting species as constant
as possible (32), ionic media were employed as background
electrolytes in the measurements. The media chosen were 0.1 M
NaNO3 or 0.1 M NaClO4. NaNO3 was the primary choice, while
NaClO4 was used to check for a possible anion dependence in the
measurements. Since no differences were observed, it was
concluded that oxidation of metal sulfides by NO3- in acidic
solutions (33) could be neglected.
Page 187
III
Methods
A
Potentiometric Titrations
During the titrations, performed at 25.0 °C, -log[H+, log[S2-], and
log[Pb2+] were monitored using appropriate electrodes. The
measurements were performed in a cell schematically represented as:
where ME denotes a glass electrode, or ion selective electrodes

(Pb2+ , S2-). The emf (E in mV) for the glass electrode is related to
log[H+] according to relation (18):
where E0 is an apparatus constant determined within each titration

and Ej is the liquid junction potential given by expression (19) for
0.1 M ionic strength (34).
Kw in this equation denotes the ionic product of water and has the
value 1.660 × 10-14 M2 at 0.1 M ionic strength.
The emf of the sulfide electrode can be expressed according to
equation (20):
Due to the controversy in the value of the constant for the reaction
HS-Û S2- + H+ (35,36), absolute values for the concentration of
sulfide ions in solution are uncertain. However, in the range 9 < -
log[H+] < 12, HS- is the dominating species, implying that equation
(20) can be recalculated to yield equation (21) (37)
where E·0S and KS were determined in separate experiments, in

which KS was found to be » 29.5 mV M-1. The emf of the lead
electrode can be expressed as for the sulfide electrode according to
equation (22):
where E0Pb and KPb were separately determined in experiments with

lead ions added to an ionic medium at -log[H+] » 5, and KPb was
found to be » 28.8 mV M-1.
Page 188
1
Proton Binding Capacity
After having washed the various MeS suspensions several times, a
known amount of acid was added to obtain -log[H+] » 3, and the
suspensions were allowed to equilibrate overnight. The excess of
protons added (Hex) was determined within the range 3 < -log[H+]
< 4. From a Gran plot (38), it became possible to perform an
accurate determination of Hex, which facilitated calculation of the
total concentration of protons reacted (HS). Furthermore, an in situ
calibration of the glass electrode was obtained at the same time.
2
Equilibration
As a criterion for equilibrium, a drift in the measured E values of »
0.6 mV/h was accepted. This required equilibration times in the
acid/base systems of 610 h for titrations with base; at -log[H+] £ 4
about 15 min was sufficient for equilibration, while in titrations
with acid at least 24 h were needed to fulfill this criterion. In the
titrations in the sulfide systems the equilibrium times were 68 h.
B
Solubility Studies
By titrating the supernatant of a suspension treated at -log[H+] » 3,
a buffer range with -log[H+] > 6.5 was observed. However, no
buffering capacity could be found by acidifying the supernatant of
a suspension treated at -log[H+] » 8. These observations made it
necessary to analyze the bulk solution with respect to zinc or lead
and sulfur to interpret possible dissolution reactions. In separate
batch experiments, the total concentrations of aqueous Me(II),
sulfur, and H+ were determined. After rinsing, known amounts of
acid (or base to acidified suspensions) were added to cover the
range 3 < -log[H+ < 8. These experiments were performed using
15-ml Pyrex glass tubes, which were deaerated with argon gas
before acid or base was added. The suspensions were allowed to
equilibrate under continuous stirring for 2448 h. This was achieved
by means of an end-rotating test tube holder operating at about 30
rpm. The suspensions were then centrifuged and the supernatant
solutions were analyzed.
C
FTIR-FTRAMAN Measurements
1
Diffuse Reflectance
A sample of 0.5% conditioned sulfide mixed with KBr (99.5%) is
placed in a small container, where source radiation strikes it and is
diffusely reflected in various directions. The reflected radiation is
collected and measured by the spectrometer. The spectrum is
measured against a reference spectrum of pure powdered KBr.
Page 189
2
KBr Discs
A sample of 1% conditioned dry sulfide mixed with KBr (99%) is
placed in a special device and compressed into a disc at high
pressure. The transparent disc obtained is put into the spectrometer
and run.
3
ATR
The measurements were performed with an ATR-technique
described by Leppinen (28). The wet sample collected after
conditioning over 12 h with known -log[H+].
4
FT-Raman Measurement
A sample of conditioned sulfide was pressed into a sample holder
and excited with 600 mW of unpolarized, intensity stabilized (0.1%
rms) 1064-nm radiation from a Spectron SL 301 series ND : YAG
laser, and the scatted light was collected with 180 backscattering
geometry optics. The InGaAs detector and integral preamplifier were
cooled to 77 K (liquid nitrogen) as this yields a fourfold gain in
signal-to-noise performance. The interference filters used to reject
light at the excitation wavelength allowed the collection of scattered
light with a Raman shift greater than 200 cm-1. The mirror drive
speed was 0.1 cm/s and 100 scan were accumulated at 1 cm-1
resolution.
D
UV-Visible Spectrophotometric Measurements
In the H+ - PbS-ethylxanthate system, the concentration of free
ethylxanthate was measured by a UV-visible spectrophotometer at a
wavelength of 301 nm.
IV
Data Treatment
A
Equilibria
The surface equilibrium reactions in the system H+ºSMe - Me2+ - L
(charge omitted), where ºSMe represents a stoichiometric surface at
pHZPC and Me2+ indicates metal ions released through the ion
exchange process, can be expressed according to:
The aquatic complexes and solid phases used in the calculations are
shown in Table 1.
Page 190
Table 1 Composition Matrix for the Aquatic
Species and Solid Phases Used in the
Calculations.
Species log b pq rstu
H+ 0.0 10 0000
Zn2+ 0.0 01 0000
Pb2+ 0.0 00 1000
HS- 0.0 00 0100
H2CO3* 0.0 00 0010
X- 0.0 00 0001
H2S 7.02 10 0100
S2- -13.9 -10 0100
HCO3- -6.16 -10 0010
CO32- -16.16 -20 0010
OH- -13.78 -10 0000
HX 1.54 10 0001
ZnOH+ -9.15 -11 0000
Zn(OH)2 -17.10 -21 0000
Zn(OH)3- -28.39 -31 0000
Zn(OH)42- -40.71 -41 0000
Zn2OH3+ -8.89 -12 0000
Zn2(OH)62- -57.53 -62 0000
ZnHCO3+ -5.86 -11 0010
ZnCO3(aq) -12.00 -21 0010
ZnO(s) -11.34 -21 0000
ZnCO3(s) -6.10 -21 0000
Zn5(OH)6(CO3)2(s) -44.11 -105 0020
PbOH+ -7.86 -10 1000
Pb(OH)2 -17.27 -20 1000
Pb(HO)3- -28.02 -30 1000
Pb2OH3+ -6.16 -10 2000
Pb3(OH)42+ -23.95 -40 3000
Pb4(OH)44+ -20.30 -40 4000
Pb6(OH)84+ -43.29 -80 6000
PbCO3(aq) -10.76 -20 1010
Pb(CO3)22- -23.46 -40 1020
PbO(s) -12.97 -20 1000
PbCO3(s) -4.00 -20 1010
Pb3(OH)2(CO3)2(s) -16.10 -60 3020
Pb10(OH)8(CO3)6(s) -88.20 -200 10060
PbX2(s) 16.70 00 1002
Pb(OH)X(s) 2.04 -10 1001
Page 191
The conditions for the total concentration and the law of mass action
applied on reactions (23)(26) will give equations (27)(30) where H,
B, C, and D are total concentrations of H+, SMe, Me2+ , and L,
respectively, and h, b, c, and d are the corresponding free
concentrations of these components.
B
The Constant Capacitance Model
For metal oxide systems, it has been found that the formation of a
charged surface introduces a complicating factor when evaluating
the equilibrium constants describing these reactions. A correction
with regard to the coulombic energy appearing on these surfaces has
to be undertaken. For this purpose a number of electrostatic models
have been developed (39), and since it was reasonable to assume that
the same complication occurred also in the metal sulfide systems,
one of these models, the so-called constant capacitance model, was
applied to the experimental data. With this model the conditional
constants were corrected to obtain the corresponding intrinsic
constants, i.e., the constants valid for an uncharged surface:
y in this equation represents the potential at the surface. According

to the constant capacitance model, y is related to the surface charge s
by the expression:
where K is the specific capacitance (C V-1 m-2). The surface charge

(C m-2) is defined according to
Page 192
where F is the Faraday constant, a is concentration of solid (g · dm-
3), S is the specific surface area (m2 · g-1), and Ts is the surface
charge (mole dm-3) defined as follows:
A combination of equations (32) and (33) then gives
For the calculations the least-squares error program FITEQL (40)

version 2.0 has been used. It was, however, found that the expected
charge dependence in the equilibrium constants were of negligible
importance.
C
Graphical Representation Data
In order to visualize the experimental data in an accurate way, Z(-
log[H+])-curves were calculated. In the systems, H+ - MeS(s) and
H+ - MeS(s) - HS-, Z was defined as the average number of protons
reacted per surface group, or expressed in equation terms:
By plotting the Z values, a visual representation of the data is

obtained as shown in Figure 4. From the slope of the Z-curve the
relative buffering capacity can be shown. A steeper slope means a
higher buffering capacity.
D
Computer Programs
Evaluation of the equilibrium model was performed using one of the
computer programs LETAGROPVRID (41) (version ETITR),
FITEQL (40), or SOLGASWATER (42).
In the LETAGROPVRID program the error sum of squares for the
total concentration of H+ ((Hcalc - Hexp)2) was minimized.
In the FITEQL program the weighted error sum of squares was
minimized in the same way as in LETAGROPVRID. In the FITEQL
calculations, the possible charge dependence of the different
formation constants was tested by employing the constant
capacitance model. It was also possible to minimize many different
weighted error sum of squares in the calculations, e.g., (Hcalc -
Hexp)2and (Dcalc - Dexp)2.
The SOLGASWATER program is basically a modelling program,
but since neither LETAGROPVRID nor FITEQL can handle
simultaneous formation of bulk precipitation during the experiments,
the SOLGASWATER program was used for the evaluation of some
of the experimental data.
Page 193
Figure 4
ZB(-log[H+])-
curve for acid/base titrations on synthetic ZnS. The symbols represent experimental
data points,while the line was calculated using the model given in Table 2.
(From Ref. 47.)
The constants were optimized according to a "pit-mapping" procedure
described by Dyrssen et al. (43).
The distribution and predominance area diagrams were calculated with
the SOLGASWATER program.
V
Results and Discussion
A
The Systems H+-ZnS(s) and H+-PbS(s)
The acid/base properties of stoichiometric hydrous zinc- and lead sulfide
surfaces have been studied (4651). In the ZB-curve (Fig. 4) it is shown
that at -log[H+] £ 5 two protons have reacted per surface group(ZB = 2.0).
Two main buffer regions can be identified, one at 5 £ -log[H+] £ 7.5 and
one at -log[H+] ³ 9. These buffer regions were found on both zinc and
lead sulfide surfaces.
The ICP-measurements (Fig. 5) showed that there was a desorption of
Zn2+ as protons were adsorbed. The low concentration of total sulfur in
solution indicated that these zinc ions were not the result of a zinc sulfide
dissolution. It was also found that [H+]ads/[Zn2+]des» 2 provided -
log[H+] £ 5.
Page 194
Figure 5
The total bulk concentration of Zn2+( ), sulfur (
), and protons reacted (D) in mM as
a function of -log[H+] (synthetic ZnS).
(From Ref. 47.)
Evaluation of this data clearly showed that the buffer capacity in the
region 5 £ -log[H+] £ 7.5 was due to an ion exchange process in
which metal ions on the surface were exchanged for protons
according to reaction (37) and to a protonation of the surface,
reaction (38).
The buffer capacity at -log[H+] ³ 9 was assumed to originate from a

deprotonation reaction at the surface.
The results from the calculations (c.f., Table 2) clearly show that
hydrous zinc and lead sulfide surfaces have very similar acid/base
and ion exchange properties. The values of pHZPC correlate well to
the solubility minima of ZnS and PbS, which are at -log[H+] » 8.5
(c.f., Fig. 3). A problem with these purely stoichiometric equilibrium
reactions is that they seem to result in chemically unlikely
compositions. Reasonable microscopic surface reactions can,
however, be obtained if it is assumed that the surface consists of
alternating ºSMeOH and ºMeSH groups. With such an assumption
the equilibrium model presented in Table 2 can be visualized as
follows:
Page 195
In this model it is assumed that the SH-group at the surface is

deprotonated at high -log [H+] and that the OH-group is protonated
at -log[H+] £ 7.5. This assumption is based upon the fact that the
OH-group has a higher basicity than the HS-group. According to the
distribution diagrams in Figure 6, the concentration of free zinc and
lead ions in solution correlates well with the model. As can be seen
from Figure 6, a stoichiometric metal sulfide surface at -log[H+] £ 5
basically consists of SH-groups, while the metal ions are found in
solution due to the ion exchange process. In very acidic solutions,
this surface probably reacts with H+ ions to produce H2S(g), thereby
initiating a dissolution process. It is also likely that the SH-rich
surface could easily be oxidized (i.e., favorable conditions are
created to yield surface oxidation products like polysulfides and/or
elemental sulfur).
B
The Systems H+-ZnS(s)-HS- and H+-PbS(s)-HS-
In these systems the complexation between sulfide ions and the
metal sulfide surfaces was studied, and it was found that different
reactions occurred depending on at what -log[H+] the sulfide ions
were added.
When sulfide ions were added at -log [H+] » 4, a precipitation of
MeS(s) involving the added sulfide ions and the metal ions in
solution occurred. It was found that the amount of sulfide added was
quantitatively precipitated provided [HS-]tot/[Me2+]tot£ 1.3. This
high ratio indicates not only that a precipitation of MeS (ratio = 1)
but also that an additional sulfide species may be adsorbed at
precipitate surfaces. (To some extent this high ratio may be
explained by loss of H2S(g) in these slightly acidic solutions;
however, no smell of H2S was noted during the experiments.) At
Table 2 Results of FITEQL Calculations on Acid/Base
Properties of Hydrous Zinc and Lead Sulfide.
log log
b2,1,-1 ± b1,1,0 ± log -
3s 3s b-1,1,0 ± 3s log[H+]ZPC
Synthetic 9.59 ± 6.91 ± -10.28 ± 8.60
ZnS 0.03 0.03 0.10
Sphalerite 9.65 ± 7.14 ± -10.29 ± 8.72
0.03 0.03 0.10
Synthetic 9.48 ± 7.11 ± -10.0 ± 0.09 8.56
PbS 0.03 0.04
Galena 10.21 ± 7.15 ± -10.2 ± 0.09 8.68
0.02 0.05
Page 196
Figure 6
Distribution diagrams showing the surface speciation of the acid/base system of (a) ZnS
and (b) PbS. The symbols denote experimentally found fractions (a) Zn2+ and
(b) Pb2+ in the bulk solution, which corresponds to a (ºSH)2.
(From Refs. 47 and 48.)
Page 197
higher ratios (> 1.3), a sharp increase in -log[H+] and log[HS-] indicated completed
precipitation adsorption processes.
Since the total concentration of Me(II) in these solutions is given by the ion exchange
process, the solubility product for the precipitate formed, log K (Me2+ + HS-Û MeS(s)
+ H+), could be evaluated. The SOLGASWATER calculations performed using data
with [HS-]tot/[Zn2+]tot < 1 resulted in a "best" fit with log K = 10.6 ± 0.7 for ZnS and
with log K = 15.6 ± 0.4 for PbS. The buffering capacity at 5 £ -log[H+] £ 7.5 previously
found on stoichiometric metal sulfide surfaces disappeared when the metal ions in
solution were precipitated as metal sulfide.
When adding sodium sulfide at 8 £ -log[H+] £ 9, the sulfide ions adsorbed on the metal
ions at the surface and formed surface complexes. It was found that the sulfide ions
were bound almost stoichiometrically to the surface as long as [HS-]tot/[ºSMe]tot < 1.
When the total amount of sulfide ions added exceeded the surface concentration ºSMe,
the excess of sulfide ions was found as free hydrogen sulfide ions in solution. Acid/base
titrations of sulfide-covered metal sulfides ([HS-]tot/[ºSMe]tot» 1) were also performed.
The ZB-curve (Fig. 7) exhibited a plateau at ZB = 1, corresponding to the formation of a
surface complex ºSHMeSH (i.e., H2S
Figure 7
B
Z (-log[H+])-curve for a sphalerite surface covered with sulfide ions. The filled and open symbols
represent H+ and OH- additions, respectively. The full curve represents the model given in Table 3,
while the dotted line represents the model for a stoichiometric surface.
(From Ref. 51.)
Page 198
bound to the stoichiometric surface). At neutral to alkaline pH, this
complex starts to deprotonate and the complexes ºSMeSH- and
ºSMeS2- (corresponding to ZB = 0 and ZB = -1) are successively
formed.
It is also clearly shown by the ZB-curves (Fig. 7) that the buffer
capacity occurring in the range 5 £ -log[H+] £ 7.5 for a
stoichiometric surface disappears when the surface is covered with
sulfide ions. This implies that the corresponding reaction depends on
reactive metal ions at the metal sulfide surface. It was found, not
illustrated, that the metalproton exchange process occurring on a
stoichiometric surface at low -log[H+] was more or less inhibited by
the sulfide complexation. By applying the previously suggested
model of two surface groups (ºSMeOH and ºMeSH) on the metal
sulfide surface, all these observations can be conceptualized by
relation (41).
In the FITEQL calculations on these systems, the ''best" fit was

found for a model including the reactions:
The resulting equilibrium constants from these calculations are given

in Table 3.
A predominance area diagram showing the major forms of surface
complexes as a function of -log[H+] and log[HS-]tot for the H+ -
ZnS(Sphaler-
Table 3 Results from the FITEQL Calculations on the H-
MeS(s)-HS System.
log log log
b1,1,0,1 ± b0,1,0,1 ± b-1,1,0,1 ±
3a 3a 3a pKa1pKa2
Synthetic 21.00 ± 12.92 ± 3.33 ± 0.04 8.1 9.6
ZnS 0.10 0.04
Sphalerite 16.72 ± 7.44 ± 0.13 -2.91 ± 0.12 9.3 10.4
0.11
Synthetic 23.18 ± 15.71 ± 5.22 ± 0.07 7.5 10.5
PbS 0.13 0.02
Galena 22.9 ± 0.6 16.04 ± 5.55 ± 0.11 6.8 10.5
0.05
Page 199
Figure 8
Predominance area diagram showing the surface speciation for sphalerite as a function of log[HS-]tot
and -log[H+]. [ºSZn]tot = 1.0 mM.
(From Ref. 51.)
ite) -HS- system is given in Figure 8. In this diagram it can be seen that when [HS-]tot >
0.5[ºSMe]tot, sulfide-containing surface species are dominating.
As shown by the results in Table 3, lead sulfide forms stronger surface complexes with
sulfide ions than zinc sulfide. This is in good agreement with the corresponding solubility
products, and indicates that the stabilities of the surface sulfide complexes are probably
related to the metal atom. It can also be seen from Table 3 that pKa1 is lower for a PbS
surface than for a ZnS surface, which means that this pKa-value probably is related to the
metal ion.
However, the pKa-values for the reaction:
are quite similar (c.f., Table 3). The similarities with the left reaction in relation (40)
occurring on stoichiometric surfaces at -log[H+] > 9 indicates that these reactions also are
deprotonations of surface sulfide binding sites.
Page 200
As is apparent in Table 3, the differences in the b-values between
sphalerite and synthetic ZnS are unexpectedly high. It was shown by
powder diffraction studies that the synthetic zinc sulfide consisted of
a mixture of sphalerite and wurtzite. Since it is known that wurtzite
has a lower solubility than sphalerite, it is reasonable to believe that
the sulfide ions are bound to the wurtzite sites on the synthetic ZnS
surface.
C
The System H+-PbS(s)-Ethylxanthate
As a conventional collector in sulfide mineral flotation,
ethylxanthate complexation at the synthetic PbS surface was studied
(21, 50). A simple reaction model was confirmed by
spectrophotometric determinations of residual concentrations of
ethylxanthate. The previously suggested surface chemical reaction
model (50) is modified according to the following scheme:
Due to the acidbase properties of the sulfide mineral surface, the

ethylxanthate adsorption is certainly pH-dependent. In alkaline
condition, competition takes place due to ion exchange between the
hydroxide ions at the surface and ethylxanthate ions in the bulk
solution or vice versa. Therefore ethyl xanthate adsorption is more
favorable in neutral or acidic pH ranges. Bearing in mind the
considerable metal ion concentration in acidic solution caused by ion
exchange between protons and metal ions at the surface, metal
xanthate precipitates may form due to their low solubility product. A
distribution diagram showing the distribution of ethylxanthate at a
lead sulfide surface is shown in Figure 9.
D
FTIR-FTRAMAN Measurements
The experiments performed by Gärd et al. confirmed surface
identities of S-H and ZnOH groups (52). It was observed that the
surface SH groups enhances as pH decreases, and surface ZnOH
sites appears with increasing pH in ZnS suspension. The SH bonding
around 2500 cm-1 and 14751375 cm-1 were identified as SH and
ZnOH bonding, respectively. The adsorption band at 2560 cm-1 due
to SH streching was confirmed by FTRAMAN measurement.
Furthermore, the surface oxidation products can be observed at
around 10001100 cm-1. The FTIR and FTRAMAN measurements
are in good agreement with our surface complexation models.
Page 201
Figure 9
Distribution diagram for the H+-PbS(s)-ethylxanthate system.
(From Ref. 21.)
VI
Concluding Remarks
The acid/base titrations on hydrous zinc and lead sulfide surfaces
have shown that an ion exchange reaction between metal and
hydrogen ions occur at the particle surface. At -log[H+] £ 5 the
concentration of metal ions found in solution corresponded to half
the quantity of protons adsorbed from the point of zero proton
condition (-log[H+] » 8.5). It was shown that the buffer region at 5 £
-log[H+] £ 7.5 was caused by this ion exchange reaction. However,
the calculations also showed that the ion exchange process occurred
in two stages where the first step was a protonation of the surface
and the second was the ion exchange between protons and metal
ions. A second buffering area was found at -log[H+] ³ 9, this one
caused by a deprotonation reaction on the surface.
The reactions between added sulfide ions and a metal sulfide surface
were found to be dependent on the -log[H+] at which the addition
was made. At -log[H+] » 4 a bulk precipitate of metal sulfide was
formed between added sulfide and metal ions released by the ion
exchange process.
At -log[H+] » 89 the metal ions were bound to the surface, and the
sulfide ions added were bound as surface complexes. It was found
that the
Page 202
complexation was very strong provided [HS-]tot/[ºSMe]tot < 1. At
this ratio practically all sulfide ions were bound to the surface but,
from that point on, added sulfide ions were found in solution as
free hydrogen sulfide. The acid/base properties of a metal sulfide
surface covered with sulfide ions were found to be completely
different from the acid/base properties of a stoichiometric metal
sulfide surface. The buffer region at -log[H+] ³ 8 was enhanced
while the buffer region at 5 £ -log[H+] £ 7.5 due to the ion
exchange reaction had disappeared. A model with the surface
complexes ºSHMeSH, ºSMeSH-, and ºSMeS2- provided a good
explanation of the experimental data.
The reactions between a lead sulfide surface and ethylxanthate
were studied, and it was found that two surface complexes
corresponding to ºSPbXH and ºSPbX- were formed.
VII
Practical Aspects
Through the present work it has been shown that the properties of a
hydrous metal sulfide surface are highly dependent on the presence
of sulfide or metal ions on the surface. Accordingly, a metal sulfide
probably shows different flotation qualities depending on which
ions are present in excess on the surface. The results presented on
sulfide complexation to metal sulfide surfaces might provide a
qualitative and quantitative explanation to the depression of galena
by sulfide ions. This depression is based on the relative adsorption
of hydrophobic species like the xanthate ion and hydrophilic
species like the sulfide ion. Since sulfide ions form much stronger
complexes with the metal sulfide surface, it is conceivable that
these ions can prevent xanthate from being adsorbed and that this
will result in a hydrophilic, nonfloatable galena. Since air (p(CO2)
» 0.0003 atm) is also bubbled through the flotation pulp during the
flotation process, it is also probable that carbonate ions present in
the pulp compete with xanthate ions and prevent xanthate
adsorption. However, due to kinetic aspects (i.e., dissolution
kinetics of carbon dioxide), it is questionable whether true
equilibrium conditions actually occur in the flotation process.
It must also be realized that the presence of air in the flotation
process probably causes a partial oxidation of the sulfide surfaces
as well as an oxidation of the xanthate ion. This implies that an
adequate description of the flotation process cannot be given until
studies of oxidized metal sulfide surfaces and of complexation with
oxidized sulfur have been performed.
The ion exchange process, in combination with the surface
oxidation process, might be a possible explanation for part of the
leaching of metal ions from mine waste deposits. In this case
protons can be exchanged for metal ions, and the resulting sulfide
groups on the surface can be oxidized
Page 203
due to oxygen in the water. These two processes might thus lead to
a dissolution of metal sulfide.
Due to the ion exchange process it is also probable that metal ions
leached from metal sulfide sediments in reduced bottom areas of
lakes when the pH in these lakes is lowered due to acid pollution.
Acknowledgments
Financial support provided by the Swedish Board for Technical
Development (STU) is gratefully acknowledged. We thank
Professor Paul W. Schindler, University of Bern, Switzerland, for
many valuable discussions and suggestions during this work.
References
1. R. O. James and G. A. Parks, Am. Inst. Chem. Eng. Symp. Series,
150(71): 157174. (1975).
2. J. W. Morse, F. J. Millero, J. C. Cornwall, and D. Rickard, Earth
Sci. Rev., 24: 142. (1987).
3. D. Dyrssen, P. Hall, C. Haraldsson, A. Iverfeldt, and S.
Westerlund, Complexation of Trace Metals in Natural Waters C. J.
Kramer and C. J. Duinker, Eds. Martinus Nijhoff/Dr. W. Junk
Publishers, The Hague, pp. 239245. (1984).
4. J. R. Brown, "Ion adsorption by mineral surfaces," Ph.D. Thesis,
University of Western Ontario, 1978.
5. J. R. Brown, G. W. Bancroft, W. S. Fyfe, and R. E. N. MacLean,
Environ. Sci. Technol., 13: 11421144. (1979).
6. I. W. Wark and A. B. Cox, Trans. AIME, 134: 725. (1939).
7. G. A. Last and M. A. Cook, J. Phys. Chem., 56: 637642. (1952).
New York, p. 613. (1982).
9. J. C. Westall, Geochemical Processes at Mineral Surfaces (J. A.
Davis and K. F. Hayes, eds.), American Chemical Society,
Washington, pp. 5478. (1986).
10. L. Lövgren, S. Sjöberg, and P. W. Schindler, Geochim.
Cosmochim. Acta, 54: 13011306. (1990).
11. H. Hohl and W. Stumm, J. Colloid Interface Sci., 55: 281288.
(1976).
12. P. W. Schindler, B. Fürst, R. Dick, and P. U. Wolf, J. Colloid
Interface Sci., 55: 469475. (1976).
13. S. W. Park and C. P. Huang, J. Colloid Interface Sci., 117:
431441. (1987).
14. S. W. Park and C. P. Huang, Wat. Res., 23: 15271534. (1989).
15. J. R. Gardner and R. Woods, J. Electroanal. Chem., 100:
447459. (1979).
16. A. N. Buckley, R. Woods, and H. J. Wouterloods,
"Electrochemistry in Mineral and Metal Processing," Conference
Proceedings, Pennington, N.J., (1988).
Page 204
17. M. M. Hyland, G. E. Jean, and G. M. Bancroft, Geochim.
Cosmochim. Acta, 54: 19571967. (1990).
18. S. W. Park, and C. P. Huang, J. Colloid Interface Sci., 128:
245257. (1989).
19. R. O. James and G. A. Parks, AIChE Symposium Series,
150(71): 157164. (1975).
20. M. S. Moignard, R. O. James, and T. W. Healy, Aust. J. Chem.,
30: 733740. (1977).
21. Z. X. Sun and W. Forsling, to be published.
22. A. Ringbom, "Solubilities of Sulfides," Report to Anal. Sect.
IUPAC, July (1953).
23. R. M. Smith and A. E. Martell, Critical Stability Constants,
Volumes 4, 5, and 6, Plenum Press, New York. (1976, 1982, 1989).
24. D. Dyrssen and M. Wedborg, Water, Air and Soil Pollution, 56:
507519. (1991).
25. Y. H. Hsieh and C. P. Huang, J. Colloid Interface Sci., 131:
537549. (1989).
26. P. Eadington and A. P. Prosser, Inst. Min. Met. Trans., 78:
C7482. (1969).
27. S. Brunauer, P. H. Emmett, and E. Teller, J. Am. Chem. Soc.,
60: 309319. (1938).
28. J. Leppinen, "On the interaction between thiol collector ions
and lead sulfide surface," Doctoral Thesis, University of Turku,
Turku, Finland, 1986.
29. O. Ginstrup, Chem. Instrum., 4: 141155. (1973).
30. W. Forsling, S. Hietanen, and L. G. Sillén, Acta Chem. Scand.,
8: 901909. (1952).
31. A. S. Brown, J. Am. Chem. Soc., 56: 646647. (1934).
32. G. Biedermann and L. G. Sillén, Ark. Kemi, 5: 425440. (1952).
33. C. S. G. Phillips and R. J. P. Williams, Inorganic Chemistry,
Vol. I, Clarendon Press, Oxford. (1965).
34. S. Sjöberg, Y. Hägglund, A. Nordin, and N. Ingri, Mar. Chem.,
13: 3544. (1983).
35. R. J. Myers, J. Chem. Educ., 63: 687690. (1986).
36. S. Licht and J. Manessen, J. Electrochem. Soc., 134: 918921.
(1987).
37. I. Toth, P. Solymosi, and Z. Szabo, Talanta, 35(10): 783789.
(1988).
38. G. Gran, Acta Chem. Scand., 4: 559577. (1950).
39. P. W. Schindler and W. Stumm, Aquatic Surface Chemistry (W.
Stumm, ed.) John Wiley, New York, pp. 83110. (1986).
40. J. Westall, Report 82-01 and 82-02 Chemistry Department,
Oregon State University, Corvallis (1982).
41. N. Ingri and L. G. Sillén, Ark. Kemi, 23: 97121. (1964).
42. G. Eriksson, Anal. Chim. Acta, 112: 375383. (1979).
43. D. Dyrssen, N. Ingri, and L. G. Sillén, Acta Chem. Scand., 15:
694697. (1961).
44. R. L. Wilfong and E. E. Maust, Jr., Infrared Spectra of
Compounds of Importance in Lead Sulfide Flotation, United States
Department of the Interior, Bureau of Mines (T. V. Falkie,
Director). (1974).
Page 205
45. P. W. Atkins, Physical Chemistry, 2nd ed., Oxford University
Press, Oxford, Great Britain. (1982).
46. Z. Sun, W. Forsling, L. Rönngren, and S. Sjöberg, Surface
reactions in aqueous metal sulfide systems. 1. Fundamental surface
properties of hydrous PbS and ZnS., Int. J. Miner. Proc., 33: 8393.
(1991).
47. L. Rönngren, S. Sjöberg, Z. Sun, W. Forsling, and P. W.
Schindler, Surface reactions in aqueous metal sulfide systems. 2.
Ion exchange and acid/base reactions at the ZnS-H2O interface. J.
Colloid Interface Sci., 145: 396404. (1991).
48. Z. Sun, W. Forsling, L. Rönngren, S. Sjöberg, and P. W.
Schindler, Surface reaction in aqueous metal sulfide systems. 3. Ion
exchange and acid/base properties of hydrous lead sulfide, Colloids
Surf., 59: 243254. (1991).
49. Z. Sun, W. Forsling, and S. Sjöberg, Surface reaction in
aqueous metal sulfide systems. 4. The correlation between some
fundamental surface properties in galena Flotation System, XVII
IMPC Preprints, Dresden. (1991).
50. Z. Sun and W. Forsling, Mineral surface complexation in
flotation, Innovation in Flotation Technology (P. Mavros and K. A.
Matis, eds.), Kluwer Academic, pp. 263281. (1992).
51. L. Rönngren, S. Sjöberg, Z. Sun, and W. Forsling, Surface
reaction in aqueous metal sulfide systems. 5. Surface complexation
of sulfide ions, J. Colloid Interface Sci., 162: 227235. (1994).
52. R. Gärd, Z. Sun, and W. Forsling, FT-IR and FT-raman studies
of colloidal ZnS. 1. Identity of S-H and Zn-OH bonds at the
ZnS/water interface, in manuscript.
Page 207
8
Salt-Type Minerals
Bohuslav Dobias *
I
Introduction
The selective flotation of salt-type minerals is a rather complicated
process especially due to the ionic composition of the solution and
the influence of lattice ions on changes of surface properties of all
mineral components present in the suspension. A number of papers
(summarized in Ref. 1) reporting on the study of phenomena
occurring in the three-component system of lattice
ioncollectormineral has been published. Based on the knowledge of
reactions taking place in the flotation pulp, these studies aim to find
optimum conditions for a selective separation of individual salt-
type minerals from one another but also of other useful minerals
present in the pulp.
This chapter summarizes the most recent information on the
surface chemistry of salt-type minerals and the most important
phenomena with respect to collector adsorption and flotation of
these minerals.
II
Characteristics of the MineralWater Interface
The surface properties of most minerals are generally influenced by
the chemical composition and the structure of the mineralwater
interface. The
Page 208
knowledge of these characteristics is required to understand the
surface chemistry of solids. From the thermodynamic point of
view, adsorption of collectors on a mineral surface is very
complicated. This is caused by the complex structure and ionic
composition of the mineralwater interface formed during the
grinding of the mineral in the water.
Wetting of a mineral with water generates a surface charge and a
region of electrical inhomogeneity at the solidwater interface, as a
result of the interaction of water molecules with the crystal surface.
A surplus of positive or negative charge, which is typical for the
crystalline structure of a mineral surface, is compensated by a
region of counter ions (H+, OH-, lattice ions, etc.) of the opposite
sign from the bulk phase. Thus an electrical double layer (EDL)
results at the mineralwater interface.
The formation of a charge on a mineral surface wetted with pure
water is controlled by two processes:
1. adsorption of water molecules (dipols), H+ or OH- ions on the
solid surface followed by a dissociation of the hydrolytic product
formed
2. preferential release of a certain lattice ion species from the solid
phase as a result of different lattice energies and readsorption of
hydrolytic lattice products
The first mechanism is characteristic of simple and complex metal
oxides and a number of silicates, whereas the other mechanism is
typical of sparingly soluble salts such as CaF2, BaSO4, and
CaCO3. These mechanisms represent two extreme cases of a
surface charge formation, because, in most minerals, both
mechanisms proceed simultaneously in proportions depending on
the chemical composition and crystalline structure.
Ions causing electrical charge of a mineral surface are called
potential-determining ions (PDI). The characteristic of these PDIs
that controls the surface reactions is the most important parameter
for judging adsorption mechanisms (see, e.g., Refs. 26). In this
connection, the pH value of the isoelectric point (IP) is a
characteristic material quantity. However, the IP obtained from
electrokinetic measurements represents the sum of all interactions
occurring at the mineralwater interface, e.g., H+/OH- adsorption,
distribution of dissolved lattice ions, or hydrolytic reactions of
H+/OH- with the dissolved lattice ions at the interface. Therefore, it
is not possible to interpret the surface reactions using only IP data,
without further information of the H+, OH-, and other ion species
at the mineralwater interface in relation to the final value of the IP.
This hypothesis agrees very well with measurements done on a
series of minerals in which the adsorption of the H+/OH- was
investigated with respect to the IP (7,8).
Page 209
A
Solubility and Electrokinetic Behavior
The presence of polyvalent lattice ions in dispersion systems
containing salt-type minerals very often leads to complicated
interactions with the mineral surface and with the dissolved
components of the bulk phase. The transition of lattice ions from
the mineral surface into the aqueous phase (suspension) is
controlled first of all by the pH value. Solubility investigation of
some salt-type minerals chosen for this purpose, which depends on
the equilibrium pH value (g) at 20 °C, have shown (Fig. 1) that the
lattice ion concentration strongly increases under a pH value
typical for each mineral. In fluorite, this is at about pH 3 because of
the formation of HF whilst for scheelite and fluoroapatite pH is
over 6.5 because of the formation of tungsten acids and dihydrogen
phosphate, and in the case of calcite
Figure 1
Lattice ion concentration in saturated mineral suspensions Mineral: fluorite;
calcite; scheelite; fluoroapatite. Anion: 1fluoride; 3total carbonate;
5wolframate; 7total phosphate; 8fluoride. Cation: 2Ca2+;
4Ca2+; 6Ca2+; 9Ca2+.
Page 210
pH is over 10 because of the formation of hydrogen carbonate. For
fluorite, the concentration ratio of dissolved lattice F- and Ca2+
ions is 2 to 1 and corresponds to the ratio of the ions in the
crystalline lattice.
On the other hand, for calcite, the sum of the carbonate and
hydrocarbonate ion concentrations (total carbonate concentration)
increases with decreases in the pH value. The relationships in the
electric double layer change in opposing directions. This means
that while the proton and hydroxyl ion adsorption plays an
important role for the z-potential* of fluorite, the z-potential course
of calcite is characterized by a preferential dissolution of the
carbonate at a reduced pH and by a formation of hydrogen
carbonate on the surface (Fig. 2).
Figure 2
Zeta potential (z) of minerals after 24 h as
related to the pH value. 1fluorite;
2apatite; 3calcite; 4scheelite.
*z-potential means zeta-potential.
Page 211
In the scheelite, the calcium concentration was higher in the entire
pH range than that of wolframite showing a marked minimum at
the neutral value. This minimum corresponds to a maximum of the
z-potential. The negative z-potential of scheelite in the entire pH
range results from the wolframite surplus on the surface. A
reduction of the pH leads to a further preferential Ca dissolution
that causes a shift toward negative z-potential values. In an alkaline
medium, the Ca concentration remains almost constant; on the
other hand, the wolframite concentration increases. In agreement
with data in the literature (9), this points to a wolframitehydroxyl
ion exchange on the surface.
In fluorapatite in the acidic and neutral pH range, the ratio of lattice
ion concentration corresponds to that of the crystal. At an alkaline
pH value, the fluoride concentration increases rapidly indicating an
exchange of F- for OH- ions.
B
Exchange Enthalpy of the H+/OH- Adsorption
Calorimetric measurements with sparingly soluble salt-type
minerals (Fig. 3) showed a relatively good agreement between
equiadsorption points obtained from calorimetric measurements
(EAPcal) and values obtained from adsorption measurements
(EAPads) (11), as opposed to pH values at which the IP was
reached (Table 1). When compared to oxides (10), the H+exchange
heats of the salt-type minerals, except those of barite and fluorite,
have very high values, although the H+/OH- adsorption densities of
salt-type minerals are 100 to 1000 times lower than those of the
oxides. The heats of exchange are therefore not only caused by H+
adsorption but are also due to a strong dissolution of the salt lattice
ions and to the hydration of the free lattice ions. Measuring mineral
solubility independent of the solution pH, Schulz and Dobias * (7)
found a more pronounced dissolution of lattice ions of calcite,
apatite, and fluorite below pH 10, 6.5, and 4, respectively. The
same was true for scheelite below pH 6.5 (12).
On the other hand, sparingly soluble barite showed neither a
dissolution heat nor a significant H+ adsorption heat, so that the
EAPcal could not be determined.
Fluorite differs from all the other minerals in the acidic range by its
endothermic heat. This difference was caused by the behavior of
the fluorite ion toward the proton. While no measurable association
heat of dissolved ions with H+/OH- occurred in the case of the
above-mentioned minerals, H+ ions reacted with dissolved fluorite
ions endothermically. Beside the exchange heats, Figure 3 also
shows the heat of association of the following reactions for fluorite:
Page 212
Figure 3
Exchange enthalpy (DH) of H+/OH- on salt-
type minerals as dependent on pH at constant
ionic strength (1 × 10-3 M NaCl). Fluorite ( ) DHass + DHads + DHdiss; (
)DHass.
Because of the endothermic association reaction of H+ with

dissolved F-, it can be expected that the H+ adsorption on the fluorite
surface also proceeds endothermically like a H+/F- association and
that the overall exchange enthalpy is positive.
Page 213
Table 1 EAP and IP of Salt-Type Minerals.
Ca5F
Mineral CaCO3 CaF2 (PO4)3 CaWO3 BaSO4
EAPcal 10.0 ± 6.50 ± 7.20 ± 0.3 7.60 ± -
0.1 1.0 0.2
EAPads 10.3 ± 6.95 ± 7.85 ± 8.20 ± 8.60 ± 0.1
0.1 0.1 0.05 0.1
IP 10.45 10.2 4.05 2 5.3
Electric charge
in KAP near zero + - - -
III
Interaction of Lattice Ions with Anionic Collectors
The flotation of sparingly soluble salts of earth alkaline elements is
usually carried out in the neutral and alkaline pH range. As
collectors, alkyl sulfates, alkane sulfonates, alkyl
benzenesolfonates, and alkyl ammonium salts have also been used
in the technical scale beside carboxylates (13). The selective
flotation from salt mixtures proves to be more difficult than their
separation from sulfides, oxides, or silicates.
Due to dissolution of salts, the concentration of lattice ions in the
flotation pulp increases, which can result in alterations of the
surface of all minerals present in the pulp as well as in a
precipitation of the anionic collector used. Therefore, the mineral
components often float at another concentration range than the
same minerals in a monomineral dispersion. This has been
described, for example, in a mixture of calcite and fluorite (14,15).
A
Precipitation Reactions
Problems of precipitation reactions of anionic collectors with Ca2+
will be treated by using some examples.
When the Na dodecyl sulfonate (Na-DDSO) concentration is
increased in a 1 × 10-4 mol L-1 CaCl2, solution (Fig. 4), a turbidity
increases at a surfactant concentration of 5 × 10-4 mol L-1. With
increasing surfactant concentration, the precipitation and thus the
turbidity are enhanced, and the z-potential of the precipitate
becomes more negative.
The exact starting point of precipitation at different concentration
ratios is shown in the precipitation diagram in Figure 5. The slope
of the lines is not -2.00 as originally expected, but -2.30. Because
of this it is interesting to calculate the activity coefficients for Ca2+
and the surfactant ion.
For ionic strength it is possible to show that:
Page 214
Figure 4
Turbidity units and zeta potential (z in mV)
of a Na-dodecyl sulfonate solution with 1 ×
10-4 M CaCl2 at pH 10 as dependent on the
surfactant equilibrium concentration.
cCa + csurf are read on the precipitation diagram, cOH is 10-4 mol L-
1 corresponding to pH 10, cCl is equal to double the CaCl2
concentration used, and cNa is the sum of the surfactant
concentration used and 10-4 mol L-1.
For the uppermost point of the curve (c·surf = 2.5 × 10-4, c·Ca = 10-3
mol L-1), the ionic strength values are 3.06 × 10-3 mol L-1.
According to equations (5), (6), and (9), those of activity coefficients
fCa and fsurf are 0.790 and 0.944, respectively.
Page 215
Figure 5
Precipitation diagram of the system
CaCl2 Na-dodecan sulfonate at pH 10. Total
Ca concentration: 1 - 1 × 10-3 M, 2 - 1 × 10-4 M.
For the lowest point of the curve (c·surf = 2 × 10-3 and c·Ca = 10-4
mol L-4), the value of I = 2.12 × 10-3 mol L-1 and fca and fsurf are
0.819 and 0.952, respectively.
As opposed to the solubility product, the slope of the precipitation
border curve does not substantially change when the activities are
used.
In Figures 6 and 7 the turbidity and the z-potential of some Ca
alkylsulfate solutions and corresponding precipitation diagrams are
shown. In all four surfactants, the turbidity increases at pH 10 with
increasing surfactant concentration, i.e., with increasing
precipitation. Sodium hexadecylsulfate (Na-HDS), like Na-DDSO,
has a Krafft point greater than 20 °C. Therefore, in a solution
saturated at 20 °C, no micelles exist. With Na tetradecyl sulfate
(Na-TDS) and Na dodecyl sulfate (Na-DDS), the turbidity
decreases over the CMC, as could be expected according to data in
the literature. The measurements were not performed with Na-DS
up to the CMC, because the amounts needed for flotation would
have been too high. The maximum turbidity and z-potential of the
precipitation occurs at the CMC of the alkyl sulfates.
Page 216
Figure 6
Turbidity units and zeta potential of a solution with 1 ×
10-4 M CaCl2 at pH 10 as dependent on the surfactant
initial concentration. Surfactant: Na-hexadecyl sulfate;
Na-tetradecyl sulfate; Na-dodecyl sulfate; Na-decyl
sulfate. Turbidity: 1; 2; 3; 4. Zeta potential: 5; 6; 7; 8.
A comparative measurement with Na-DDS at pH 4 showed a
precipitation diagram identical to that at pH 10. The z-potential of
the precipitation was somewhat less negative.
Precipitation diagrams with CaCl2 and Na-TDS or Na-DDS show
(Figure 7) lines with slopes of -2.46 and -2.25, respectively, under
the CMC of the surfactants. Above the CMC the solubility of the
Ca-surfactant salts increases, as expected.
With Na-DS at a Ca concentration of 10-4, no analytically
detectable precipitation was found up to surfactant concentrations
of 10-2 mol L-1, although some solutions were slightly turbid. With
other alkyl sulfates, the
Page 217
Figure 7
Precipitation diagram of the system CaCl2 Na alkyl sulfate at pH 10. Surfactants:
1Na-dodecyl sulfate; 2Na-tetradecyl sulfate.
turbidity also occurred before the precipitation of the anionic surfactant
started due to salting-out of traces of long-chain alcohols that are almost
always present as impurities in alkyl sulfates.
The composition of the precipitated Ca-surfactant salt probably plays a
certain role in the deviation from the ideal slope in case of Na-DDSO
and alkyl sulfates. During the precipitation of Na-surfactant salt, co-
precipitate and surfactant ions can adsorb on the surface. This was
demonstrated by the negative z-potential at pH 4. In the literature, slopes
somewhat smaller than -2.0 are therefore reported. The z-potential of the
precipitation gets more negative with an increase in surfactant
concentration, i.e., with increasing surfactant adsorption. Above the
CMC the monomer concentration does not increase further, as opposed
to free Na ions, which leads to a reduction of the z-potential.
The turbidity and the z-potential of a Ca-laurate solution show a plateau
at a surfactant concentration of about 3 × 10-4 M. Otherwise the nature
of the curves is similar to that of other surfactants (9).
B
Adsorption and Abstraction Isotherms
Adsorption isotherms of surfactants at the mineralwater interface
describe the relation between the adsorbed surfactant amount and its
equilibrium
Page 218
concentration in the solution. Thus they furnish important knowledge for
drawing up the adsorption mechanism and for the interpretation of
flotation results. This is valid under the presumption that the surfactant
molecules adsorb only on the mineral surface.
The adsorption of anionic surfactants at the mineral interface is very often
accompanied by their own precipitation with Ca ions. From the reactions
with fluorite, for example,
it follows that
Figures 811 show the lattice ion concentrations and the reduction of the
Na-DDSO concentration due to adsorption and precipitation in the
suspensions as reported elsewhere (16).
The amount of precipitated surfactant (CT pre) can be calculated from the
Figure 8
Lattice ion concentrations (c) and reduction of the surfactant concentration (cT) of a
fluorite suspension as a function of Na-
dodecyl sulfonate equilibrium concentration cT.
The dashed line shows the Ca-surfactant precipitation boundary.
Page 219
Figure 9
Lattice ion concentrations and decrease in surfactant concentration (c, mol L-1) in the
calcite suspension versus the Na-dodecyl sulfonate concentration (cT, mol L-1).
The dashed line shows the Ca-dodecyl sulfonate precipitation boundary.
data for the lattice ion concentrations in the aqueous supernatant (C·ca and
C°F) and in the supernatant with surfactant (Cca and CF) as reported before
16) for fluorite:
for calcite:
for scheelite:
Table 2 gives the lattice ion concentrations, and , for the minerals
at pH 10. In systems with little surface, the ratio of the concentrations
should resemble the stoichiometry of the chemical formula. In the four
systems studied, the differences from this ratio arise from the predominant
Page 220
Figure 10
Lattice ion concentrations and decrease in surfactant concentration (c, mol L-1)
in the scheelite suspension versus the Na-dodecyl sulfonate concentrations
(cT, mol L-1). The dashed line shows the Ca-dodecyl sulfonate precipitation
boundary.
exchange of the lattice anion in the mineral surface with dissolved
hydroxide ions in the solution and perhaps from minute impurities in
the minerals. From comparison with the literature, the high calcium
concentration in the scheelite suspension and the resulting solubility
product is assumed to come from a rather soluble calcium salt in the
scheelite (9), while the low lattice ion concentrations of the fluorite
arise from the extremely low dissolution velocity (18,19).
For the four minerals, the composition of the supernatant, together
with data in the literature (10), indicates a calcium-rich surface with
adsorbed hydroxide ions, resulting in a negative zeta-potential at pH
10.
Adsorption tests showed more than 90 and 98% of the given
equilibrium surfactant adsorption density after 10 min and 2 h,
respectively. The equi-
Page 221
Figure 11
Lattice ion concentrations and reduction of the surfactant concentration
c, and adsorption density G of a fluoroapatite suspension as a function
of Na-dodecyl sulfonate equilibrium concentration cT. The dashed line
shows the Ca-surfactant precipitation boundary.
librium time of 24 h was chosen to achieve lattice ion
concentrations near equilibrium.
In the system fluoroapatite/Na dodecyl sulfonate, neither a
significant adsorption nor precipitation occurred. The lattice ion
concentrations remain constant at all surfactant concentrations.
In general, one can say that the lattice ion concentrations remain
con-
Table 2 Lattice lon Concentration is Suspensions with 25 g-L-1 Mineral
pH 10.
Mineral Fluorite Calcite Scheelite Fluoroapatite
Chemical CaF2 CaCO3 CaWO4 Ca5F(PO4)3
formula
Lattice Ca2+:1.1 × Ca2+:8.2 × Ca2+:3.2 × Ca2+:3.3 ×
10-4 10-5 10-4 10-5
Ion concn. F-:4.8 × 10-4 CO32-:1.6 × WO42-:5.0 × PO43-:5.5 ×
(mol · L-1) 10-4 10-5 10-5
F-:8.3 × 10-5
Page 222
stant at a low surfactant concentration, and the abstraction densities
of surfactants correspond to real adsorption densities, because no
precipitation occurs. At higher surfactant concentrations
precipitation starts, the calcium concentration is reduced, and the
anion concentration increases owing to the dissolution of the
mineral. The solubility product for Ca-dodecyl sulfonate was found
to be 2.9 × 10-11 mol L-3. The anionic surfactant does not undergo
exchange with lattice anions on the mineral surface at a low
surfactant concentration.
After attaining the solubility product of Ca-dodecyl sulfonate, the
ceq of the surfactant corresponds to lower values, which would be in
accordance with the amount of the surfactant bound at the mineral
interface, because the equilibrium surfactant concentration is
influenced by its precipitation with Ca ions. The relations ceq/G
obtained thus can be called abstraction isotherms. From the
experimentally obtained CT values, it is possible to calculate the
adsorption densities of the surfactant (9). As an example, reactions
occurring on adsorption of an anionic surfactant on fluorite surface
can possibly be of the form:
with
The abstraction density Gabs can be calculated from the diminution
of surfactant concentration cT:
where V is suspension volume, A is specific mineral surface, and m

is mineral amount weighed.
The adsorption density Gads may be defined as:
Page 223
Figure 12 shows the abstraction (equation 21) and adsorption
(equation 23) densities of Na-DDSO on fluorite. The surfactant
precipitation starts at an adsorption density of approximately 2.3
µmol m-2.
From the data of the area needed for one surfactant molecule in a
single adsorption layer (20), it follows that the theoretical monolayer
cover for Na-DDSO is reached at about 6.3 µmol m-2. The area
needed for this surfactant is mainly controlled by the hydrocarbon
chain.
The high abstraction values possibly can be explained by surfactant
precipitation. The calculated maximum adsorption density of 14
µmol m-2 suggests the formation of a double layer at high surfactant
concentrations.
A possible adsorption owing to an exchange of the surfactant for F
ions according to the equations (24) and (25) leads to too high a
value of the precipitated surfactant amount:
The resulting error would not be too large, because this ion exchange
should materialize only during formation of the first adsorption
layer. The maximum adsorption density is, therefore, calculated too
low with a maximum error given by the difference of the values at
the monolayer cover and the
Figure 12
Abstraction density G of Na-dodecyl sulfonate
on fluorite as a function of the surfactant
equilibrium concentration. The dashed line
shows the adsorption density.
Page 224
beginning of precipitation, i.e., approximately 3 µmol m-2, which
brings about no substantial deviation in the results.
The stoichiometric equations derived can distinguish between a true
surfactant adsorption and surfactant precipitation, but not between
precipitation in the solution and precipitation on the mineral surface.
Figures 1214 show comparable measurements with Na-DDSO on
calcite, scheelite, and fluoroapatite. In the calcite and scheelite
suspensions, the concentration course can be explained in the same
way as in fluorite. The equations for calculating the precipitated
surfactant amounts and the adsorption densities of calcite and
scheelite were derived like those for fluorite. Because of different
stoichiometry of the crystalline lattice one gets
and neglecting the ionic exchange of the surfactant for hydrogen

carbonate and wolframite ions:
Figure 13
Abstraction density G of Na-dodecyl sulfonate
on calcite as a function of the surfactant
equilibrium concentration. The dashed line
Page 225
Figure 14
Variation of abstraction density (G mol m-2)
on scheelite with Na-dodecyl sulfonate
concentration c (mol L-1). The dashed line
The adsorption density on calcite (Fig. 13) does not further increase
after the surfactant starts to precipitate at 2 µmol m-2. In scheelite
(Fig. 14) the precipitation starts fairly soon because of the high Ca2+
concentration. The maximum adsorption density at 2 × 10-3 M Na-
DDSO was approximately 7 µmol m-2, thus corresponding to the
monolayer covering.
During adsorption of Na-DDSO on fluoroapatite, the lattice ion
concentrations remain constant (Fig. 11), which means that no
surfactant precipitation occurs, although it can be expected
according to the precipitation diagram. This is in agreement with the
formation of soluble calcium phosphate complexes that have been
described by Changgen and Yongxin (21).
The adsorption densities on apatite are much lower throughout the
entire range measured than on all other minerals.
Page 226
Figure 15
Floatability (A) of fluorite (1) and calcite (2) as a function of the concentration of Ca-
dodecyl
sulfonate precipitate (mol L-1).
In regard to flotation of suspensions that correspond to the adsorption

experiments concerning both the lattice ion and the surfactant
concentrations, a good agreement existed between the mineral floatability
and the surfactant adsorption isotherms.
High lattice ion concentrations, especially of polyvalent ions, reduce the
surfactant adsorption because of their own adsorption. In addition,
phosphates form soluble complexes with Ca2+ ions and reduce surfactant
precipitation in this way.
To study the influence of precipitation on the mineral surface, flotation
experiments were carried out with fluorite and calcite. As shown in Figure
15, the mineral floatability does not change in either system with increasing
precipitation product concentration. The deposition of precipitation
products at the mineralwater interface was not demonstrated in the system
studied. This probably happens as a result of the strongly negative z-
potentials of the precipitate and the mineral after the surfactant adsorption.
C
Effect of the Polar Group and the CMC of the Collector
Investigations of the solubility of Ca-surfactant salts have shown that for
the same chain length (C12-) the lattice ion concentrations of minerals
strongly depend on the chemical character of the polar group.
Page 227
Figure 16 shows the lattice ion and surfactant concentrations in a fluorite suspension with
dodecyl sulfate. At lower surfactant concentrations, no precipitation and no F- ion
exchange occur, like with NaDDSO (Fig. 8). In the precipitation range, there is a shift of
the precipitation bounds to a higher solubility of the Ca-surfactant salt because of an
influence of the ionic strength. Beyond the CMC, Ca ions are taken up on the micelle
surface leading to a greater solution of fluorite.
Taking the deposition of Ca ions on micelles into consideration, the following equation
results:
where cCasoln is free Ca ion concentration in the solution and cCamic is concentration of
Ca bound on the micelle. For the total Ca concentration, it holds that
Figure 16
Lattice ion concentrations and reduction of the surfactant concentration c of a fluorite suspension as a
function of Na-dodecyl sulfate equilibrium concentration Ct. The dashed line shows the Ca-
surfactant
precipitation boundary.
Page 228
Figure 17
Lattice ion concentrations and reduction of the surfactant concentration c of a fluorite suspension as
a function of Na laurate equilibrium concentration cT.
From this it follows that
Equation (31) is identical to equation (32). With the measuring methods used, the
precipitated surfactant amount can also be calculated in micellar solutions according to
equation (33) and the adsorption density according to equation (23) (8).
Figure 17 shows results of the adsorption measurements of Na laurate (Na-L) on

fluorite. Because of the low solubility product of Ca laurate (Ca-L), the surfactant
precipitation starts at very low concentrations. The
Page 229
course of the Ca concentration as dependent on the surfactant
equilibrium concentration furnishes a solubility product for Ca-L of
about 1 × 10-13 mol L-1.
With increasing surfactant concentration, the amount of
precipitated Ca laurate decreases. From about 10-3 mol L-1 on,
filtrates of the suspension remained slightly turbid, even when
micropore filters were used and centrifuged. In this range the Ca
concentration filtrates increase with elevating surfactant
concentration: fluoride concentration and the abstracted surfactant
amount decrease. The increase in Ca concentration does not result
from micelle formation, since in that case the fluoride
concentration should also be expected to increase. Most probably,
the finest precipitate particles were not separated but rather
measured together. This is in agreement with the ability of
carboxylates to form a very low aggregate number under the CMC.
The minimum of the calcium concentration curve corresponds well
to the plateau of the turbidity determination in Figure 18. From this
point on,
Figure 18
Turbidity units and z-potential of a Ca laurate solution
with 1 × 10-4 CaCl2 (mol L-1) at pH 10 as a function
of the surfactant equilibrium concentration cT.
Page 230
a change in the precipitation behavior can be expected. This was also
observed by Young and Matijevic (22) who found ranges with
different physical properties in the precipitation diagram of the
system Ca nitrate/Na-L. It is better to consider a co-precipitation of
Na-L here, because the Na concentration is much higher than that of
Ca. For the reaction
it follows by analogy with equations (18)(33) that
The precipitated surfactant amount calculated according to equation

(33) is therefore too small when co-precipitation occurs. For x = 1
the adsorption density of the surfactant from 10-3 mol L-1 on would
be almost constant and would correspond to a double layer as in the
case of an adsorption of other surfactants on fluorite. This is a useful
approach since at a surfactant concentration of approximately 3 ×
10-4 mol L-1 Ca laurate precipitates, and the suspension can be well
filtered. An adsorption density will be reached that suggests the
formation of a double layer according to equation (33).
D
Abstraction and Exchange Enthalpy
The adsorption of an ionogenic surfactant on the surface of salt-type
minerals is characterized by a manifold adsorption mechanism and is
often combined with a parallel occurring precipitation reaction. A
typical example is the adsorption of Na oleate (Na-OL) on the
fluorite or calcite surface (Fig. 19). The experimental conditions
were that the surface of both minerals offered for adsorption and the
pH value (10) were the same. In this work an almost identical
abstraction at low concentrations was obtained; at intermediate and
high oleate concentrations, the abstraction isotherm had a linear
course with calcite, while it tends to saturation with fluorite. A linear
increase in the abstraction enthalpy of calcite that stretches far
beyond a concentration of 2 × 10-3 mol L-1 indicates that beside the
Na-OL adsorption on the surface, Ca2+ ions are released from the
mineral surface and react with the added oleate under formation of a
Ca-OL precipitate:
The precipitate remains partly in the bulk phase, and partly bound to
the adsorption layer. The calculation of the molar abstraction
enthalpy results in a mean value of Hm = -29.4 kJ/mol. This value
agrees very well with the Ca-OL formation enthalpy ( -28 kJ/mol)
found by von Rybinski and Schwunger at 40 °C (23).
Page 231
Figure 19
Abstraction exchange enthalpies (DH) and abstraction densities (A) of Na oleate on
calcite and fluorite as a function of the initial concentration (pH 10 + 0.05) at a constant
ionic strength of NaCl (1 × 10-3 mol/L).
A separation of the abstraction in adsorption and precipitation in calorimetric

measurements is not possible experimentally. Experiments carried out with
the filtrate to determine the precipitated portion cannot be compared with
experiments performed with mineral dispersions because the amount of
dissolved Ca2+ ions is different. The Ca2+ ion amount in filtrate is accurately
added while in the dispersion new Ca2+ are supplied steadily on oleate
addition from the calcite surface so that more Ca-OL particles precipitate.
A different situation exists in fluorite. Here the dissolution of Ca2+ is reduced
by the adsorption of OL- on the mineral surface, and the abstraction tends to a
maximum. The OL- adsorption combined with a co-adsorption of the Ca-OL
precipitate results in a more firm and strong hydrophobic adsorption layer
than in the case of calcite (23) and enables a very good floatability even at a
very low OL concentration. In fluorite the molar abstraction enthalpy DHm
decreases with increasing OL- concentration (Table 3). This means that the
adsorption, which took place through Ca-OL complex formation at the
beginning (DHm = -26.6 kJ/mol), is
Page 232
Table 3 Molar Abstraction Enthalpy DHm of Oleate on
Fluorite a Different Initial Oleate Concentration C0.
C0 (x10-3 mol · L-1) 0.5 1 2 3 4
- DHm (kJ · mol-1) 26.5 18.8 11.7 8.2 8.0
controlled by weaker forces (such as hydrophobic interactions) at

higher concentrations.
Adsorption densities of surfactants obtained from the corrected
abstraction isotherms can be used to calculate the exchange enthalpy
using the Graham's equation (25,27):
where r is the radius of ions adsorbed in the Stern layer, c is the ion
concentration in the solution, and Gs is the adsorption energy of one
ion in the Stern layer.
Table 4 shows the surfactant concentrations of the system's
surfactant/mineral investigated (9, 28) and the calculated adsorption
energies at an adsorption density of 1 µmol m-2.
Ionic radius of the polar group radius of 2.5 × 10-10 m was
employed at temperature 20°C.
For the same surfactant chain length, the rank order of adsorption
energy diminution is carboxylate > sulfonate > sulfate. This
corresponds to the order of solubility increase of the Ca-surfactant
salts.
The energy difference of -7.8 kJ mol-1 between the adsorption of
Na-DS and Na-TDS results from the association energy of two
additional CH2 groups of Na-TDS. As a maximum association
energy value of one CH2 group, -1.39RT is given (28), i.e., 3.39
kJ/mol at 20 °C. Since the experi-
Table 4 Concentration in the Solution (CT) and
Adsorption Energy (Gs) of Some Surfactants at an
Adsorption Density in µmol · mol-2.
Mineral Surfactant CT(mol · L-1) Gs(kJ · mol-1)
Fluorite Na-DDSO 3 × 10-4 -21.4
Fluorite Na-DDS 1 × 10-3 -18.5
Fluorite Na-TDS 4 × 10-5 -26.3
Fluorite Na-L 6 × 10-5 -25.4
Calcite Na-DDSO 2 × 10-4 -22.4
Calcite Na-DDS 7.5 × 10-4 -19.2
Scheelite Na-DDSO 3.5 × 10-4 -21.0
Fluoroapatite Na-DDSO 2 × 10-3 -16.8
Page 233
mental value is somewhat higher, an almost complete association
can be presumed. This can also be expected in other surfactants
used because of a similar structure.
The whole association energy of hydrocarbon chains is higher for
all systems than the adsorption energy. The adsorption starts on
energetically favorable sites. Higher adsorption densities are
achieved only owing to the association of hydrocarbon chains of
the surfactants. The interaction energies of polar groups with the
surface are not high; this might be the reason that no measurable
exchange was found of lattice ions for surfactant ions in any
system. The main operating adsorption mechanism is thus the
exchange of OH- ions for surfactant ions and a further adsorption
due to an association accompanied by a surface negativity. In
fluorite, many exchangeable OH- ions are present on the surface;
therefore, a double layer forms after a monolayer is complete. In
other minerals, a higher negative charge of the mineral surface
limits the maximum adsorption density.
The start of a precipitation can be shifted toward higher adsorption
density values by choosing surfactants with a better solubility of
their Ca salts. The differences are, however, small, and it can be
assumed that the same interactions are responsible for adsorption
and for precipitation: the interactions of a polar group with Ca2+
ions and the surfactant association.
IV
Adsorption Mechanisms of Anionic Collectors
As shown earlier, binding of an anionic collector on the surface of
salt-type minerals is a very complicated process, which is
accompanied by a number of antagonistic reactions. This chapter
attempts to describe schematically the adsorption mechanism from
the viewpoint of different physicochemical and chemical criteria
based on studies performed in previous investigations and on the
data from the literature.
A
Polar Head of the Collector Is Charged opposite to the Mineral
Surface
The system Na-OL-fluorite is the most often discussed example of
the adsorption mechanism of an anionic surfactant on a salt-type
mineral with a positive sign of the z-potential under natural
conditions of its aqueous dispersion concerning the pH.
In spite of the unequivocal mechanism of the z-potential
dependence of fluorite on the Na-OL concentration showing a sign
change of the z-potential with increasing the Na-OL concentration,
the adsorption process
Page 234
Figure 20
Zeta potential of fluorite as a function of initial oleate concentration at pH 10 ± 0.1
is very complicated (Fig. 20), especially due to the formation of a
precipitated phase resulting from a reaction of Ca2+ with the oleate ion
and is often a matter of controversy. For example, according to Rao et al.
(30), the adsorption mechanism of Na-OL on synthetic fluorite is
discussed from the viewpoint of adsorption measurements that are
supplemented with data on the z-potential and FTIR-spectroscopy of the
surface compounds. These studies show that the chemisorption of Na-OL
on surface Ca species followed by the physical adsorption of
hydrocarbon chains (bilayer formation) precede the precipitation of Ca-
OL in bulk phase. With an increase in the solid/liquid ratio, the bulk
precipitation of Ca-OL after bilayer formation was found to be
diminished, and at a high solid concentration the isotherm was found to
level off with bilayer formation. As the solid concentration increases, the
Ca concentration in the bulk phase decreases due of its adsorption on the
heads of the hydrophilic oleate layer (bilayer), and thus the extent of bulk
precipitation decreases. The adsorption density for a monolayer coverage
corresponds to the condensed state of an alkyl chain with a cross-
sectional area of 0.33 nm2 (liquid-crystal state). Comparison of the
theoretical chemical equilibria with the experimental data shows that the
infinite increase in the slope of the isotherms is due to the precipitation
Page 235
of Ca oleate. The solubility product of this three-dimensional phase
is found to be in the pK range 14.414.5 and agrees with most of the
values quoted in the literature. An application of the two-
dimensional condensation model is considered, and no evidence for
the surface precipitation process is observed.
Sivamohau et al. (31) characterized the oleate adsorption on
fluorite by four regions (Fig. 21). At low oleate concentrations
(region I), the oleate ions adsorb on the surface through Na+ ions
on monocoordinated Ca2+ ions, and the adsorbed layer is filled
with hemimicelles. At higher concentration (region II) and at pH
9.45, e.g., a surface Ca-oleate precipitate region appears with a
substantially higher adsorption density, reaching a mixed bilayer
made of Na-OL, Ca-OL, and carboxylic acid in a solid state. In
these regions, adsorbances of all the bands investigated
(hydrocarbon chains and carboxylate groups) were linearly
correlated with surface coverage. In regions III and IV, a change in
adsorbance (sharp or monotonic decrease according to the pH)
indicates a change in the state of the adsorbed layer.
Keller et al. (32) showed on the basis of FTIR-spectroscopy that, at
low equilibrium oleate concentration, the adsorbed species are
chemisorbed at
Figure 21
Adsorption isochrone of Na-OL on fluorite (1), z-
potential (2), and turbidity (3) as a function of
residual NaCl concentration for pH 9.45 and a temperature of 20°C.
Page 236
the fluorite surface and that at higher oleate concentrations a
surface precipitate of Ca-OL is formed. From changes in the
asymmetric -CH2 stretching band location, it was concluded that
oleate initially adsorbed on fluorite at room temperature with its
alkyl chain in a predominantly flexible gauche conformation
unassociated with the nearest neighbors. Further loading involves
increasing van der Waal's interactions between adjacent adsorbed
oleate species, causing the alkyl chains to become more rigid as a
result of increased proportions of traps conformers. Upon heating
from room temperature, the -CH2 stretching frequencies of the
chemisorbed oleate shift to higher values, thus indicating melting,
which increases the number of gauche conformers along the alkyl
chain. This thermotropic behavior resembles the transition of a gel
or coagel to a micellelike state. By way of comparison, surface-
precipitated oleate remains in a predominantly all-trans
conformation at both 25 and 40 °C (Fig. 22).
Based on the adsorption data, contact angle measurements, and
induction time in the system fluorite/oleate, Hu et al. (33) conclude
that binding the surfactant occurs in three regions: (1) a chemical
plateau region, (2) a surface Ca-oleate precipitation with a
significantly higher adsorption density reaching an adsorption
maximum, and (3) a region that exhibits a decrease in adsorption
density at higher equilibrium oleate concentration. They thought
that adsorption seemed to involve a chemisorbed layer at low
equilibrium oleate concentrations, which can be distinguished from
a surface Ca-OL precipitate at equilibrium concentrations
exceeding 5 × 10-6 M.
In another paper characterizing oleate adsorption on fluorite on the
basis of FTIR-spectroscopy, Hu et al. (34) suggest that a chemical
interaction occurs between adsorbed oleyl groups at higher
temperatures and/or oxygen potential and is responsible for the
increased hydrophobicity. Specifically, the intensity of the =CH
bond of the adsorbed oleate diminishes at high temperature and/or
high oxygen potential. Such a phenomenon was not observed in
control tests with the Ca-oleate soap. For extensive reaction at the
fluorite surface, the presence of a new surface species has been
identified. These results indicate that the fluorite surface catalyzes
an oxidation and/or polymerization reaction between adsorbed
oleyl groups, which results in a more hydrophobic structure at the
fluorite surface.
Apart from the problem of Na-OL adsorption, a number of papers
have been published dealing with the adsorption mechanisms of
other anionic collectors on the surface of salt-type minerals. The
adsorption of Na dodecyl sulfate, Na dodecyl sulfonate, Na-
dodecylbenzene sulfonate, and Na tridecylate was studied by
Dobias * (3, 35). Most of the adsorption isotherms have a double
S-shape characterizing a polylayer adsorption of the surfactant,
e.g., the adsorptions of Na-DDS and Na-tridecylate are
characterized by a plateau and a maximum on the adsorption
isotherm (Fig. 23). The
Page 237
Figure 22
Conformation of chemisorbed oleate species
at the surface of fluorite for different levels
of adsorption density revealing the transition
from the gauche state at submonolayer
coverage to the more rigid trans state at near
molecular coverage. The gauche bonds are
designated by asterisks.
plateaus on the adsorption isotherm correspond to the completion
of the first adsorption layer like results from the calculated values
of the square area requirements (Table 5). For CaF2 these values
are somewhat higher in both adsorbates than the values of the
compressed monolayer films on the phase interface Na-DDS
solutionair (0.204 nm2) (36) and Na tridecylateair (0.209 nm2)
(37). The deviations of the square area requirement depend first of
all on the number of free adsorption sites related to the composition
of the phase interface (38) and the degree of dissociation of the
surfactant, as shown in an earlier study (39).
Page 238
Figure 23
Adsorption isotherms of Na-DDS (1) and Na-TD
on fluorite at 25°C and pH of 6.10 ± 0.25 (1) and
of 6.45 ± 0.45 (2). G-number of molecules
adsorbed per cm2.
Table 5 Surface Area of Na-Dodecyl Sulfate and Na-
Tridecylate Molecule Adsorbed on Fluorite and Barite.
CaF2 BaSO4
Surface Surface
Ads Molecule area Molecule area
Absorbate layer number (nm2) number (nm2)
Na-dodecyl 1 3.465 × 1014 0.2886 5.010 × 1014 0.1996
sulfate 2 6.447 × 1014 0.1551 8.772 × 1014 0.1139
Na-tridecy- 1 2.135 × 1014 0.4682 4.375 × 1014 0.2286
late 2 8.428 × 1014 0.1186 9.964 × 1014 0.1003
Page 239
The maxima of the adsorption isotherms correspond, according to
the values of the square area requirements in Table 5, to the
formation of a second adsorption layer and thus a bilayer film.
Because of a reverse orientation of the surfactant molecules, the
second adsorption layer has an associative hydrophilic (polar)
character. A number of measurements performed showed that in
the region of the second (hydrophilic) layer formation the fluorite
floatability decreased (44). The surfactant adsorption was
accompanied by precipitation reactions of Ca2+. Similarly like
during Na-OL adsorption on fluorite (Fig. 20), the z-potential value
changes of fluorite corresponds to the concentration of other
anionic surfactants (Fig. 24). The plateaus on CaF2 adsorption
isotherms are at those Na-OL concentration regions at which the
first adsorption layer occurs.
Tamamushi and Tamaki (40) and Shergold (41) also postulate a
bilayer adsorption mechanism in the CaF2/NaDS system. A recent
paper of Mielczarski et al. (42) showed that adsorption of Na-DS
on CaF2 surface using internal reflection spectroscopy in the IR
range was a surfactant anion that chemisorbs on the mineral
surface.
Figure 24
Zeta potential (z in mV) of CaF2 as a function of Na-DDS
(1) and Na-tridecylate (2) concentrations at 25°C,
c - mol L-1, pH value 6.10 ± 0.25.
Page 240
Figure 25
Adsorption isotherm of Na-DDS on CaF2 at different pH values. c - mol L-1;
G - number of molecules in cm-2, pH values - 2.8 (1), 4.1 (2), 5.2 (3), 6.4 (4), 8.3
(5), 0.1 (6), 11.5 (7).
A good floatability of CaF2 was attained at a concentration of NaDS, at
which probably the close-packed monolayer of chemisorbed dodecyl
sulfate anions is formed (about 1 × 10-3 mol L-3). An abrupt decrease in
floatability was observed above the concentration of Na-DDS at which
the precipitation of Ca-DDS occurs. It is very likely that Ca-DDS is
partly precipitated at the surface of the CaF2 particles. Experimental
results suggest that the concentration of Ca2+ ions in the solution and
the presence of other ions (e.g., Ba2+) forming sparingly soluble
precipitates with DDS- ions play an important role in the CaF2-Na-DDS
system.
The extent to which pH influences the surfactant adsorption is
illustrated in Figure 25.
As can be seen from this relation, a plateau forms on all adsorption
isotherms that slightly shift to lower equilibrium concentrations with
decreasing the pH.
If a plot of the pH against the square area requirement F of the adsorbed
Page 241
surfactant molecule is made a linear relation is obtained in the
region pH < 10 (Fig. 26). The deviation from linearity at pH > 10 is
related to a greater hydroxylation of the fluorite surface (41). From
the linearity for pH < 10, it can be concluded that the H+
concentration determines the surfactant adsorption density in the
monomolecular layer (up to a plateau). However, because a
polycomponent system is involved in which the H+ concentration
strongly influences the CaF2 solubility simultaneously, as shown
by Bahr (43) and Shergold (41), this effect of H+ ions on the
surfactant adsorption is not direct. According to Bahr, at pH < 3 the
formation of HF2- ions and at pH < 4 the increased CaF2 solubility
must be taken into consideration. This CaF2 solubility decreases in
NaOH solutions up to pH < 12.
According to the results of Shergold, mainly Ca2+ and H+ ions and
molecular HF occur in the system CaF2/H2O at pH < 3.25. At pH >
3.25, however, only Ca2+ and F- ions exist in the system according
to his data. The CaF3- and CaF+ concentrations are much lower
than those of the above-mentioned ion species. In the absence of
Na-DDS, precipitation
Figure 26
Surface area (F in nm2) of Na-DDS molecules
on fluorite as dependent on the pH value at 25°C.
Page 242
Figure 27
Graphic representation of the adsorption of anionic surfactant at the salt-
type mineral/water interface. The sign of the electric
charge of the mineral is opposite to that of polar heads of the surfactant.
of the Ca salt occurs in equilibrium solutions (surfactant-CaF2) at a
surfactant concentration of 3.7 × 10-3 mol L-1 and at pH 4.
It can be assumed that the H+ and OH- are mainly responsible for the
exchange of the surfactant for OH- and for its association with H+
ions.
The above-mentioned new information on adsorption and binding of
anionic substances agrees well with the mechanisms proposed
elsewhere (e.g., Refs. 27 and 35). All considerations on anionic
surfactant adsorption on the surface of salt-type minerals with a
positive charge sign of the mineralwater phase interface are
summarized in the following scheme (Fig. 27). In this respect, it is
again necessary to recall the relation on Figure 15, according to
which a negatively charged macroscopic Ca-surfactant precipitate
has no effect on an increase in mineral hydrophobicity because in the
presence of a precipitate, an equilibrium surfactant concentration
corresponds to an adsorption density at which the fluorite surface is
already charged negatively (cf., Fig. 24).
Page 243
B
Mineral Surface with the Same Charge Sign as the Collector Head
A typical example of adsorption of an ionogenic collector on a
mineral surface with the same charge sign is the adsorption of an
anionic surfactant on barite. If we consider the adsorption
mechanism as an ionic exchange in the electric bilayer, a sufficient
adsorption of the anionic collector should not occur because of the
negatively charged mineral surface, and no flotation should be
expected (27,45). However, the z-potential trend on barite in
dependence on Na-DDS or Na-TD concentration shows that the
anions of the collector do adsorb on the mineral surface (the
negative z-potential value increases) (Fig. 28). The adsorption
densities of both collectors are much higher than in the case of
fluorite (Fig. 24). On the adsorption isotherms two plateaus occur
as in the case of CaF2 (Fig. 29). Therefore, it is possible to
postulate a formation of the first hydrophobic and the second
hydrophilic adsorption layer. A proof of the formation of a second
amphi-
Figure 28
Zeta potential (z in mV) of BaSO4 as a function of
Na-DDS (1) and Na-tridecylate (2) concentration at
25°C. c - mol L-1, pH value 5.45 ± 0.45.
Page 244
Figure 29
Adsorption isotherms of Na-DDS (1) and Na-TD
on fluorite at 25°C and pH value of 6.10 ± 0.25
(1) and of 6.45 ± 0.45 (2). G number of
molecules adsorbed per cm2.
pathically oriented layer was brought, for example, by Cap (36)

who used the contact angle measurement technique in a system of
BaSO4/Na-DDS. The formation of a second layer takes place in the
area of the CMC values. During anionic collector adsorption up to
the formation of the first hydrophobic layer, one can assume
binding of anions on Ba2+ ions occurring in the solid part of the
electric double layer (the inner and outer Helmholtz layer), which
can be regarded as part of the crystalline lattice of the mineral. At
its higher concentrations or with extending the adsorption time, the
existence of chemical binding (chemisorption) can be supposed of
Ba2+ ions directly in the crystalline lattice or an exchange of SO42-
for DDS- and TD-. Formation of chemical bonds between lattice
ions and surfactant anions was confirmed by several authors
(46,47). The formation of a second adsorption layer takes place to
the detriment of the association binding between hydrocarbon
residues as is clear from the shear z-potential increase of the barite
surface with increasing collector concentration. It can be assumed
that the Ba-surfactant compound has a microscopic character in the
Page 245
electric double layer but a macroscopic nature in the bulk phase. The
higher adsorption density of the anionic collector on the barite
surface in comparison to that of fluorite can be explained by the
presence of a Ba-surfactant precipitate on the barite/water phase
interface. A possible adsorption mechanism of the system discussed
is shown schematically in Figure 30 whereby the author proceeds
from earlier assumptions (34,35) supplemented by new knowledge.
C
Adsorption in the Presence of Complexing Agents
A very comprehensive study on the effect of complexing agents on
the adsorption mechanism of collectors on a salt-type mineral
surface was published by Baldauf (48). To explain the effect of the
inhibitors, he proceeds from their adsorption on the mineral surface
that can be proved even for some hydrophilizing compounds. A later
comprehensive study comes from Orthgiess (49) who studied the
mechanism of interaction of complexing agents with collectors, with
the mineral surface, and with lattice ions in the
Figure 30
Graphic representation of the adsorption of anionic surfactant at the salt-
type mineral/water interface. The sign of the electric
charge of the mineral is the same as this of the surfactant polar head.
Page 246
solution, by adsorption, electrokinetic, and flotation measurements,
as well as by determinations of lattice ion solubility. Calcite,
fluorite, and barite were used as minerals; the collectors were ionic
surfactants, and the complexing agents were compounds containing
O or N as possible donor atoms. Beside the solution of practical
problems, the author concentrated on the explanation of basic
mechanisms of the depression effect of these substances.
Results of his investigations (49) can be summarized as follows:
1. Complexing agents in the flotation pulp generally influence:
a. the adsorption density of the collector on the mineral surface,
b. the solution of the mineral through complexing lattice ions in
the solution,
c. the structure of the mineral/aqueous solution phase interface
and of the electric double layer, and
d. the state of the fluid/gaseous phase interface.
2. Some of the complexing agents tested act as selective inhibitors
on mineral flotation with anionic surfactants.
3. When using cationic surfactants as collectors, an activating
effect of complexing agents is often seen on mineral floatability.
4. Structurally closely related reagents sometimes show a strongly
different effect on flotation (e.g., ethylene-diaminetetraacetic acid,
EDTA, or trans-1,2-diaminecyclo-hexantetraacetic acid, CDTA,
with fluorite).
5. In spite of their strong effect on floatability, the chelating agents
do not adsorb on the mineral surface at all, or very slightly
(contrary to the often assumed mechanism of their inhibitory
effectBaldauf).
6. The effect of complexing agents on flotation is not uniform: Five
different mechanisms could be proved in the systems investigated:
a. The complexing agent (not a chelating agent) adsorbs on the
mineral beside the surfactant and causes a certain hydrophilicity
of the particle surface through its polar groups, which leads to
alterations of the floatability (p-hydroxy-benzaldoxime,
HBOfluorite).
b. The complexing agent (chelating agent) complexes mineral
cations in the interface or in the bulk phase as well (i.e., directly
or indirectly through surplus solubility), thus stripping the
mineral surface of cations and causing the surface negative.
With the anionic collector Na-DBS this leads to a strong
decrease in adsorption density of the surfactant and thus to an
inhibitory effect of the complexing agent, without its significant
adsorption on the mineral.
Page 247
c. Complexing of metal ions by the chelating agent in the
mineralsolution interface leads to a decrease in free cations.
Using Na-OL as collector reduces the formation of metal-oleate
associates and drives back the surfactant adsorption in the sense
of surface precipitation, which results in a markedly lower
floatability.
d. The induced negativity of mineral surface described under b
favors the surfactant adsorption of cationic flotation collectors,
which in many cases activates mineral floatability. Moreover, an
associate of an (anionic) complexing agent and a (cationic)
surfactant can act as an especially interface active collector.
e. The chelating agents CDTA and 8-hydroxyquinoline (8-HC),
both containing hydrophilic and hydrophobic ring structures,
selectively activate fluorite flotation (in spite of driving back the
collector adsorption by CDTA according to the mechanism
described under b and c). The reason is that both complexing
agents are effective collectors themselves, especially for fluorite.
Two characteristic examples for the adsorption mechanism of
anionic collectors on the salt-type mineral/water interface are
shown schematically in Figures 31 and 32.
D
Adsorption in the Presence of Ion Exchangers and Non-Ionic
Surfactants
The adsorption of silicates and precipitation of anionic collectors
by polyvalent lattice cations often causes problems on selective
separation of salt-type minerals from gangue material during
flotation. Inorganic or organic complexing agents may also be
added to the flotation pulp to prevent undesirable activation and
precipitation effects of multivalent cations. As was shown in
Chapter 7, a number of complexing agents have an effect on
removing lattice cations from the mineral surface and this, on the
other hand, causes an increase in the concentration of these ions in
the flotation pulp. Two possibilities are presented here on how to
decrease the concentration of lattice ions in a flotation dispersion to
the concentrations at which these unwanted reactions do not take
place or occur only on a very limited scale.
1
Use of Ion Exchangers
Because of these effects, the flotation process control becomes
more complex. Insoluble inorganic ion exchangers, such as zeolite
A, do not have these undesirable properties. Polyvalent lattice
cations are being exchanged for Na+ by these compounds (50,51)
in their pore structure and are therefore removed from aqueous
solutions. Because of the insolubility of zeo-
Page 248
Figure 31
Graphic representation of the mechanism of action of complexing agents. Case 1:
The complexing agent adsorbs on the mineral surface and causes a reduced
hydrophobicity through its polar groups or depression of surfactant adsorption.
lites, only small interactions occur with the individual mineral
components in the flotation pulp.
The simplest example of an activation as a consequence of limited
solubility of salt-type minerals is the influence of Ca2+ ions in the
system fluorite/quartz during CaF2 flotation using an anionic collector
(50). For this reason, the flotation of quartz in saturated solutions of
various minerals was investigated as a function of the amount of zeolite
A added. The addition of zeolite A reduced significantly the recovery of
activated quartz because it removes Ca2+ from the solution and inhibits
the collector adsorption over active bridges. The effect of zeolite A is
related to the collector used and the type of mineral. The most
significant effect was observed with Na hexadecyl sulfate (Na-HDS) as
collector and fluorite as mineral. The maximum flotation recovery of
quartz in saturated solutions of scheelite at low zeolite concentrations
results from a change in the pH value by the
Page 249
Figure 32
Graphic representation of the mechanism of action of complexing agents. Case
2: The complexing agents releases cations from the mineral surface leading to
a greater negativity of the zeta potential and a reduction of anionic collector
adsorption.
addition of zeolite A, the aluminum silicate having a certain alkalinity
because of its process of manufacture. At high concentrations of
zeolite A, the pH value may vary from 9.5 to 10.5.
The advantages of zeolite A in flotation processes are not confined to
the deactivation of quartz and silicatesthe dosage of anionic collectors
may also be reduced by precipitation. Anionic collectors form
insoluble precipitates with Ca2+ ions (see Sec. III). At high Ca2+ ion
concentrations almost the whole of the collector precipitates. The
presence of quartz causes a further reduction in the Na-HDS solution
concentration by the adsorption of collector at quartz surface
activated by Ca2+ ions. Collector consumption in flotation with
anionic surfactants is strongly increased by these effects, the major
portion of the collector forming insoluble Ca salts and being unable to
adsorb on the mineral surface.
Figure 33 shows the influence of zeolite A on the precipitation of Na-
Page 250
Figure 33
Influence of an addition of zeolite A (1) and Na silicate (2) on solution
concentration of Na-
HDS at a constant concentration of CaCl2 (c = 3.5 × 10-4
mol L-1) (initial solution concentration of Na-HDS co 82 × 10-6 mol L-1.
Solution concentration of Na-HDS in the presence of 3.5 × 10-4 mol CaCl2
L-1 c1 8 × 10-6 mol L -1; T = 25°C).
HDS with Ca2+ ions. In the absence of zeolite A, the concentration
of dissolved collector decreases to a value of c1 by the addition of
3.5 × 10-4 mol Ca2+ ions/L. When the solution is equilibrated with
different amounts of zeolite A prior to the addition of collector, the
precipitation of insoluble Ca-HDS may be prevented almost
completely with sufficient dosage of zeolite A. If, instead of zeolite
A, equal amounts of Na silicate are added, the effectiveness is only
minor. It may be expected from the results obtained that, in ore
flotation with Ca2+ ions in the pulp, zeolite A would reduce the
consumption of anionic collector.
This effect was verified by flotation experiments of the model
system apatitecalcite with an anionic collector and on industrial
scale in a flotation plant when obtaining apatite (51).
Page 251
2
Use of Non-Ionic Surfactants
The co-adsorption of non-ionic surfactants in the adsorption layer
of ionogenic collectors bound to a mineral surface leads to an
increase in the mineral hydrophobicity and thus to its better
floatability (50,52). If we decrease the concentration of an ionic
collector under its critical concentration at which no flotation of the
mineral occurs and replace it with a nonionogenic collector, the
mineral will float. If the ionogenic collector concentration is
reduced under a value that is smaller than the solubility product of
the corresponding salt of the polyvalent cation and the surfactant,
the occurrence of precipitation reactions in the flotation pulp can be
influenced.
Figure 34 shows the adsorption isotherms of hexadecyl sulfate on
pure scheelite in the absence and presence of nonyl phenol
pentaglycol ether (NP 5). In the absence of NP 5, the isotherm rises
very steeply at the beginning and then levels out to a plateau at low
equilibrium concentrations. This behavior corresponds to the
commonly observed adsorption characteristic of ionic surfactants
on solid matter (53). If hexadecyl sulfate is adsorbed
Figure 34
Adsorbed amounts of Na-HDS on scheelite as a function of
the solution concentration of hexadecyl sulfate in the
absence (×) or presence ( ) of NP 5. T = 40°C,
cHS/cNP 5 = 1 : 1.
Page 252
Figure 35
Enthalpy of displacement DHdis on scheelite as a function of the adsorbed amounts
G of HDS (×) and for a mixture of HDS and NP 5 at a ratio of 1 : 1 ( ). T = 40°C.
from a 1 : 1 mixture with NP 5, the amounts adsorbed are significantly
lower, but the isotherm shape is maintained. The influence of the non-
ionic surfactants becomes even more marked when calcite is used as
adsorbent instead of scheelite.
The theoretical area requirement of a Na-HDS molecule with vertical
arrangement in a fully packed monolayer can be used for the
interpretation of the adsorption isotherms obtained. Starting with a value
of 0.20 nm2, the adsorbed amount is 8 × 10-6 mol m-2 for a complete
monolayer and 16 × 10-6 mol m-2 for a double layer. As Figure 34 shows,
the adsorbed amounts measured are clearly above these values and also
coincide with earlier investigations (34). An exception is the adsorption of
hexadecyl sulfate on scheelite in the presence of NP 5.
Microcalorimetric methods used on pure Na-OL and Na-HDS have
already shown that Ca surfactant salts are mainly formed at higher
solution concentrations in the mineral slurry (22). This is why the
corresponding measurements were also carried out for the mixtures of
hexadecyl sulfate and NP 5. In Figure 35, the heat occurring during
adsorption has been plotted against the adsorbed amounts of scheelite
from Figure 35. The heat effect is called enthalpy of displacement DHdis
and results from the following equation:
Page 253
where Qm is the heat effect measured, Qdil is the dilution heat
during adsorption, ms is the mass of adsorbent, and as is the
specific surface of adsorbent.
The term enthalpy of displacement has been chosen because the
heat effect includes the surfactant adsorption, the desorption of
water molecules from the adsorption surface and the precipitation
of Ca-surfactant salts.
According to Figure 35, three is a linear correlation between the
heat effect and the amount of surfactant that is abstracted from
solution at high surfactant concentration. The differential enthalpy
of displacement is therefore independent of the surface coverage
above a certain concentration. In the presence of the non-ionic
surfactant, the curve is analogous but the enthalpies of
displacement obtained are markedly lower. Basically similar effects
can be observed with the adsorption on calcite (52).
To distinguish between the adsorption and association of surfactant
molecules and the precipitation reaction, it is useful to determine
the reaction heat of the surfactant anions with the Ca ions in the
absence of minerals. The results of these tests are presented in
Figure 36 for different additions of NP 5. In the absence of non-
ionic surfactant, the reaction enthalpy between hexadecyl sulfate
anions and Ca ions is DHR = -50 kJ mol-1. Because of the low
solubility product of long-chain alkyl sulfates (53), almost the
entire amount of surfactant is precipitated. As the NP 5 content
increases, the reaction enthalpy undergoes a linear decrease. This
indicates that the precipitation of Ca-surfactant salts can be
suppressed by the addi-
Figure 36
Reaction heat Q of the precipitation of Ca-HDS from
Na-HDS and Ca ions as a function of the NP 5
concentration. cHS = 2.23 × 10-3 mol L-1, cCa =
2.6 × 10-3 mol L-1, T = 40°C.
Page 254
tion of the non-ionic surfactant. The measurements thus also
coincide with results obtained with mixtures of alkylbenzene
sulfonate and non-ionic surfactants in the washing process (54).
When comparing the results from Figure 36 with the slopes of the
curves for scheelite and calcite with high adsorption, a high degree
of coincidence of results, within the limits of the accuracy of the
measurements, becomes apparent in the absence or presence of the
minerals (Table 6). This means that the formation of the Ca-
surfactant salts is strongly superimposed on the adsorption process
proper on the mineral surfaces with higher surfactant
concentrations. The formation of the Ca-surfactant salts is
influenced by the solubility of the minerals so that the effects are
greater with the calcite, due to its higher degree of solubility. By
the addition of the non-ionic surfactants, the differential enthalpy
of displacement decreases, i.e., the precipitation of Ca-surfactant
salts is reduced. However, the precipitation cannot be completely
prevented by the non-ionic surfactants. On the other hand, with a
very low surfactant concentration, the enthalpy of displacement is
determined by direct interaction between the surfactant anions and
the mineral surface. A possible mechanism is the formation of
mixed micelles between anionic and nonionic surfactants leading to
an increased solubility of the collector-cation salt (55).
The surface precipitation is of great importance for practical
flotation (35,56). Since complex mineral mixtures have to be
separated in practical work, the ion content of the flotation pulp is
determined by the minerals with the maximum solubility. In
particular, ores with a high calcite content may cause major
problems. The separation of valuable minerals from calcite is
therefore especially important.
Table 6 Comparison of the Formation Enthalpy of
Calcium Hexadecyl Sulfate from Calcium Ions and
Sodium Hexadecyl Sulfate (HS) According to Fig. 36
with the Differential Enthalpy of Displacement
DHdis/dG in the Linear Range of the Curves for
Scheelite (Fig. 35) and Calcite [52] in the Absence or
Presence of NP 5; CHS = 800 mg · 1-1, CNP S = 800
mg · mol-1.
DHR DHdis/dG
Mineral Surfactants (kJ md mol-1) (kJ md mol-1)
- HS -51 -
Scheelite HS - -43
Calcite HS - -57
- HS/NP 5 -28 -
Scheelite HS/NP 5 - -28
Calcite HS/NP 5 - -28
Page 255
Table 7 Flotation of a Scheelite ore with Hexadecyl Sulfate, Alkyl
Sulfosuccinate and Mixtures with Nonyl Phenol Pentagylcol Ether (NP 5)
in the Microflotation Cell. R-recovery, pH 8.59.0, flotation time 2 min.
Concentration grade (%)
Collector Co-collector Rtot Rwo3
(g · t-1) (g · t-1) (%) (%) WO3 CaO SiO2 Al2O3 Fe2O3
- - - - 0.4 6.8 59.5 12.1 7.0
HS 500 - 1.3 3.2 1.0 10.7 50.6 13.8 9.4
HS 330 NP 5 170 4.3 10.5 1.0 9.8 51.4 13.9 8.9
HS 250 NP 5 250 4.8 12.0 1.0 9.5 50.9 13.1 9.0
HS 170 NP 5 330 4.6 11.5 1.0 9.2 51.7 13.7 9.2
AS 500 - 3.9 52 5.3 19.1 36.2 11.6 8.6
AS 250 NP 5 250 8.7 39 1.8 11.5 47.7 12.5 8.2
AS 100 NP 5 100 3.8 55 5.9 15.8 39.6 10.2 9.5
Page 256
This is why the effect of the collector mixtures tested was checked
by flotation tests on a natural scheelite ore with calcite as a gangue
mineral (Table 7). The tests were carried out in a microflotation
cell in order to initially exclude the influence of the non-ionic
surfactants on the foaming behavior. The results show that the
hexadecyl sulfate has very poor collector properties. At moderate
selectivity, the recovery is very low after a short flotation period of
2 min. By adding the non-ionic surfactant, the recovery undergoes
a strong increase with identical selectivity. More favorable results
were obtained with alkyl sulfosuccinate as a collector with its
higher selectivity and recovery. Here, the collector concentration
can be clearly reduced with a slightly increased selectivity and
recovery. This results in an optimum ratio between anionic
collector and non-ionic surfactant, which presumably depends on
the Ca ion content of the flotation pulp and has to be optimized for
different types of ores. The collector mixtures thus show a positive
effect in flotation, and this was confirmed by tests with a laboratory
flotation cell.
References
1. N. P. Finkelstein, Review of interactions in flotation of sparing
soluble calcium minerals with anionic collectors, Trans. IMM 98:
157. (1989).
2. B. Dobias, Electrokinetische Interpretation der Adsorption
ionogener Tenside an Festkörperoberflächen, Teil I. Tenside
Deterg.: 322. (1972).
3. B. Dobias, Electrokinetische Interpretation der Adsorption
ionogener Tenside an Festkörperoberflächen. Teil II.
Potentialbestimmende Gitterionenanionische Tenside, Tenside
Deterg.: 131. (1976).
4. D. W. Fuerstenau and S. Raghavan, Some aspects of the
thermodynamics of flotation, Flotation (M. C. Fuerstenau, ed.),
Vol. 1, Am. Inst. Min. Metall. Petrol. Eng., New York, p. 21.
(1976).
5. S. H. Hanna and P. Somasunduran, Flotation of salt type
minerals, Flotation (M. C. Fuerstenau, ed.), Vol. 1 Am. Inst. Min.
Metall. Petrol. Eng., New York, p. 197. (1976).
6. G. S. Basetti, M. Tschapek, and A. K. Helmy, Calorimetric
determination of the point of zero charge, J. Electroanal. Chem. 36:
507. (1972).
7. P. Schulz and B. Dobias, ''Effect of Lattice Ions in the Selective
Flotation of Salt-Type Minerals," Proceedings XV Int. Min.
Process. Congr., Tome II, Flotation and Hydrometallurgy, Cannes,
p. 16. (1985).
8. B. Dobias and K. A. Wierer, "Electrokinetic and adsorption
properties of mineral particles in solid/liquid separation," Preprints
of the 1st World Congr. Particle Technology, Part IV, p. 221. Partec
Nürnberg (1986).
9. J. Oberndorfer, "Adsorption of Anionic Surfactants on Ca Salts,"
PhD Thesis, Univ. Regensburg (1989).
10. M. R. Atademir, J. A. Kitchener, and H. L. Shergold, The
surface chemistry and flotation of scheelite. Part I, J. Colloid
Interface Sci., 71(3): 466. (1979).
11. K. A. Wierer and B. Dobias, Exchange enthalpies of H+ and
OH- ion adsorp-
Page 257
tion on minerals with different characters of potential
determining ions, J. Colloid Interface Sci., 122: 171. (1988).
12. T. Wakamatsu and Sh. Mukai, AIChE Symp. Ser. 71, 150: 81.
(1975).
13. H. Schubert, Aufbereitung fester mineralischer Rohstoffe, Bd.
II, Veb. Dt. Verlag Grundstoffindustrie, Leipzig (1977).
14. C. Tanneberger and H. Baldauf, Zur Rolle der lösungsphase bei
der Flotation schwerlöslicher Salzminerale, Neu Bergbautecknik,
13(5): 271. (1983).
15. H. Baldauf, Neu Bergbautecknik 16(6): 228. (1986).
16. J. Oberndorfer and B. Dobias, Adsorption mechanisms of
anionic surfactants on sparingly soluble minerals, Colloid Surf., 41:
69. (1989).
17. J. Oberndorfer and B. Dobias, Interaction of surfactants with
polyvalent cations in solution and at the mineralwater interface,
Progr. Colloid Polym. Sci., 76: 286. (1988).
18. J. D. Miller, M. E. Wadsworth, M. Mistra, and J. S. Hu,
Flotation chemistry of fluorite/oleate system, Principle of Mineral
Flotation. The Wark Symposium, Trans. Austr. Inst. Min. Metall.
Symp. Ser. 40: 31. (1984).
19. C. N. Litsakes and P. Ney, Forstchr. Mineral., 63: 165. (1985).
20. B. Dobias, Einige physiko-chemische Eigenschaften von
Sammlern für nichtsulfidische Minerale, Bergakademie, 3: 162.
(1965).
21. L. Changgen and L. Yongxin, Selective flotation of scheelite
from calcium minerals with sodium oleate as a collector and
phosphates as modifiers, Int. J. Miner. Process., 10: 219. (1983).
22. S. L. Young and J. Matijevic, Precipitation phenomena of
heavy metal soaps in aqueous solutions. III. Metal laurates, J.
Colloid Interface Sci., 61: 278. (1977).
23. W. von Rybinski and M. Schwunger, Kalorimetrische
Untersuchungen zur Adsorption von Tensiden an Mineralen, Ber.
Bunsenges. Phys. Chem., 88: 1148. (1984).
24. M. A. Lovell, L. A. Goold, and N. P. Finkelstein, Infrared
studies of the adsorption of oleate species on calcium fluorite, Int.
J. Miner. Process., 1: 183. (1974).
25. O. Stern, Zur Theorie der elektrolytischen Doppelschicht, Z.
Electrochem., 30: 508. (1924).
26. D. C. Grahame, The electrical double layer and the theory of
electrocapillary, Chem. Rev., 41: 441. (1947).
27. J. M. W. Mackenzie, Min. Sci. Eng., 25: (1971).
28. B. Dobias *, Dispersion stability in mineral processing.
Coagulation and Flocculation (B. Dobiás, ed.), Marcel Dekker,
New York, p. 539. (1993).
29. S. I. Chou and J. H. Bae, Surface precipitation and
redissolution in brine, J. Colloid Interface Sci., 96: 192. (1983).
30. K. H. Rao, J. M. Cases, and K. S. E. Forrsberg, Mechanism of
oleate interaction on salt type minerals. V. Adsorption and
precipitation reactions in relation to the solid/liquid ratio in the
synthetic fluoritesodium oleate system, J. Colloid Interface Sci.,
145: 330. (1991).
31. R. Sivamohan, P. de Donato, and J. M. Cases, Adsorption of
oleate species at the fluorite aqueous solution interface, Langmuir,
6: 637. (1990).
32. J. J. Kellar, C. A. Young, K. Knutson, and J. D. Miller,
Thermotropic phase
Page 258
transition of adsorbed oleate species at a fluorite surface by in
situ FTIR-IRS spectroscopy, J. Colloid Interface Sci., 144: 381.
(1991).
33. J. S. Hu, M. Misra, and J. D. Miller, Effect of temperature and
oxygen on oleate adsorption by fluorite, J. Miner. Process., 18: 57.
(1986).
34. J. S. Hu, M. Misra, and J. D. Miller, Characterization of
adsorbed oleate species at the fluorite surface by FTIR
spectroscopy, J. Miner. Process., 18: 73. (1986).
35. B. Dobias, Flotierbarkeit und elektrokinetischen Eigenschaften
von CaF2 und BaSO4, Erzmetall., 21: 275. (1968).
36. M. Cap, Rudy, 13: 1. (1965).
37. B. Dobias, unpublished results.
38. D. F. Sears and J. H. Schulman, Influence of water structures
on the surface pressure, surface potential and area of soap
monlayers of lithium, sodium, potassium and calcium, J. Phys.
Chem., 68: 3529. (1964).
39. B. Dobias, Beitrag zur Theorie der Flotation von
nichtsulfischen Mineralen, Freiberger Forsch.,-H., Ser. A 335: 7.
(1965).
40. B. Tamamushi and K. Tamakai, Trans. Faraday Soc., 55: 1007.
(1959).
41. H. L. Shergold, Infrared study of adsorption of sodium dodecyl
sulphate by calcium fluoride(Fluorite), Trans. Inst. Min. Metall.
Sec., C81: 148. (1972).
42. J. Mielczarski, P. Nowak, and J. W. Strojek, Correlation
between the adsorption of sodium dodecyl sulphate on calcium
fluorite and it's flotation, Int. J. Miner. Process., 11: 303. (1983).
43. A. Bahr, Ph.D. Thesis, Univ. Clausthal-Zellerfeld, (1966).
44. B. Dobias, Phenomena in Mixed Surfactant Systems, (J. F.
Scamehorn, ed.), Am. Chem. Soc., Washington, D.C., p. 216.
(1986).
45. B. Dobias, Surfactant adsorption on minerals related to
flotation, Structure and Bonding Series, Vol. 56, Springer Verlag,
Berlin, p. 92. (1984).
46. K. I. Marinakis and H. L. Shergold, The mechanisms of fatty
acid adsorption in the presence of fluorite, calcite and barite, Int. J.
Miner. Process., 14: 161. (1985).
47. Prasad and D. W. Fuerstenau, The adsorption of hydroxamate
on semisoluble salts. I. Adsorption on barite, calcite and
bastnaesite, Colloids Surfaces, 8: 103. (1983).
48. H. Baldauf, Freiberger Forsch.-H, Ser. A: 619. (1980).
49. E. Orthoiess, "Komplexbildner als Reglersysteme bei
Mineralflotation," Ph.D Thesis, Univ. Regenburg, 1991.
50. P. Schulz, "Einfluss der Gitterionen auf die Adsorption- und
Flotationeigenschaften schwerlöslicher Salze," Ph.D Thesis, Univ.
Regensburg, 1985.
51. W. von Rybinski, M. J. Schwunger, P. Schultz, and B. Dobias,
"Zeolite for increased selectivity and recovery in flotation
processes," Proc. Reagents in the Mineral Industry., Rome, p. 197.
(1984).
52. W. von Rybinski, M. J. Schwunger, and B. Dobias, Surfactant
mixtures as collectors in flotation, Colloids Surfaces, 26: 291.
(1987).
53. E. H. Lucassen-Reynders, Anionic surfactants, Surfact. Sci.
Ser., Vol. 11, Marcel Dekker, New York. (1981).
Page 259
54. D. L. Smith, K. L. Mateson, and M. F Cox, J. Am. Oil. Chem.
Soc., 62: 1396. (1985).
55. W. von Rybinski and M. J. Schwunger, Langmuir Lett., 2: 639.
(1986).
56. P. K. Anasnthapamanabhan and P. Somasundaran, Surface
precipitation of inorganics and its role in adsorption and flotation,
Colloids Surfaces, 13: 151. (1985).
Page 261
9
Developments in Flotation Equipment
Risto Lindsberg
Outomec Oy, Espoo, Finland
I
Introduction
The mining industry is struggling with rising costs, decreasing ore
grades, and low metal prices. By using flotation in the grinding
circuit, bigger units as roughers and scavengers, and carefully
selected machines in the cleaning flotation, it is possible to
overcome these problems.
Flotation in the grinding circuit has become well accepted in a
short time, even though the old "unit-cell" concept is still in the
minds of many metallurgists.
Outokumpu has developed the old-fashioned unit-cell concept
further, so that instead of the total mill discharge, only the
recirculating load (cyclone sands) is treated in a special coarse
flotation cell (Skim-Air). This cell is installed under the cyclone in
such a way that the tailings flow by gravity to the secondary
grinding without any extra pumping (see Fig. 1). The concentrate
from the coarse flotation cell can be combined with the final
concentrate without cleaning, when metal sulfide ores are
discussed. With reverse flotation (of industrial minerals), froth
product is naturally final tailings.
Within flotation the large cells having a unit volume of 30 m3 or
more have become generally accepted. There is no doubt about the
savings in energy consumption, maintenance costs, and good
availability. Some types
Page 262
Figure 1
Schematic drawing showing Skim-Air
flotation machine in grinding circuits.
are also efficient in handling coarse material and can be easily
restarted after a power failure. Increased unit size makes it possible
to use fewer parallel process lines, thus leading to the improved
circumstances for onstream automatic control. The reasons for the
success of the big cells are obvious, but relatively little attention
has been paid to the new opportunities in designing flotation
circuits consisting of large cells. The engineering seems to follow
more or less conventional paths that have been used in connection
with small cells. Among the factors having an influence on the
"correct size" flotation cell circuit design are the following:
the need of effective volume is similar to the small cell
the selectivity of big cells is often better than that of small ones
(thick froth layer, etc.)
the possibilities to control product quality by adjusting air and/or
pulp level are better than with smaller cells (fewer control
points)
Cleaning is an obvious part of conventional flotation circuit,
although it is possible to produce part of the final grade concentrate
in one stage (e.g., Flash Flotation with the Skim-Air or the first
cells of the conventional roughers). However, cleaning flotation
has to be done for the rest of the rougher concentrate. There are
also many types of column cells available now in addition to
conventional cells of several sizes. Some words about large OK
cells as cleaners will be said later, since it seems to be beneficial to
use large cells (of same size) in the whole process in some cases.
The newest invention in cleaning flotation area is the HG (high
grade) flotation machine, which will be discussed in detail later.
Page 263
II
Flotation in Grinding Circuit
It is well known among metallurgists that commonly used
hydrocyclone has a tendency to separate mineral particles
according to their specific gravity rather than size. As an example,
the separation limit is shown in Table 1. Thus, quite often, the
cyclone underflow has a higher content of valuable minerals than
fresh ore. This is because heavy minerals stay in the circulating
load until they are ground fine enough to be taken into the cyclone
overflow. By that time part of heavy minerals is ground to slime,
and silicate middlings are still too coarse for the conventional
flotation.
Eight years ago the first Skim-Air flotation machine was installed
at the Hammaslahti concentrator in Finland. Results from this test
installation and from many other permanent installations later on
are very promising:
Avoidance of overgrinding, because of floatable mineral
particles in cyclone sands are selectively captured before
secondary grinding. Mixed or too coarse particles do not float
within the short retention time and thus they will be reground.
The load in the actual flotation circuit decreases, less volume is
needed, or feed rate can be increased, if flotation has been the
bottleneck.
The final concentrate is coarser and thus easier to dewater.
The overall slime losses are reduced.
This is especially true with heavy and soft minerals.
Since then the Flash Flotation method with the Skim-Air flotation
machines has been used in several other grinding circuits all over
the world. Results similar to the original Hammaslahti installation
have been achieved, and some new features have been found,
especially improved precious metal recovery. Low-grade gold and
silver ores have been very suitable for the Flash Flotation process.
Also platinum, lead, copper, and zinc ores seem to suit the process,
as well as industrial minerals.
The suitability of the Skim-Air machine can be tested in the
laboratory scale simply producing a standard series of flotation
tests so that it is possible for very coarse material to settle on the
bottom of the cell. This simulates the short circuiting of oversize
material in plant scale Skim-Air.
Table 1 Separation Limits.
Specific Separation
gravity limit
(kg/m3) (µm)
Silicates 2.6 300
Sulfides 5.0 60
Gold 19 20
Page 264
Generally the SK concentrate is final grade material; except when
precious metals are present, the grade is higher than from
conventional flotation.
Recovery improvement is in many cases less than can be detected
(e.g., 1%), but also 2 to 6% total recovery improvements have been
measured. With precious metals, the figure can be up to 20%.
Reasons for the better recovery are briefly:
reduced overgrinding and
steady feed grade to the conventional flotation.
All installations reported the coarser combined final concentrate,
which further resulted in easier dewatering. Generally, the moisture
of concentrate after filtration was 1 or 2% lower, when the SK unit
was operating. This means substantial savings in thermal drying if
it is used. Energy savings in some cases were achieved because of
the reduced flotation volume. These savings are at the level of 2
kWh/t of ore.
One large Skim-Air installation and completely new scope of
flotation is in Brazil where rougher flotation for gold ore is done by
eight SK-240 machines from the cyclone underflows. Scavenging
and cleaning are done in the cells of "correct sizes" (16 and 8 m3).
III
Rougher-Scavenger Flotation
As mentioned earlier, rougher flotation can be done (and
beneficially) in many cases already in the grinding circuit from
cyclone underflow, which is preconcentrated feed from flotation
without harmful slimes. However, normal rougher-scavenger
flotation is done in small or large conventional cells. Since large
cells have been discussed many times before during the last
decade, only some new features will be brought up.
During the last few years with increasing cell sizes, some
installations have re-encountered an old problem: short circuiting.
If total residence time is short (e.g., 13 min), then 6 or 7 shafts are
not enough to prevent short circuiting if large (e.g., 40 m3) cells are
used. In a case like this, 12 shafts are recommended in one
rougher-scavenger line to prevent short circuiting.
Figure 2 shows some measurements from operating installations
(squares) with large cells (30 to 50 m3) and a theoretical curve, if
the flotation cell mechanism were an ideal mixer. It can be seen
that in some cases (black squares) short circuiting is rather a strong
phenomenon and can be helped only by increasing residence time.
Thus it has to be remembered that not only a long line of cells will
prevent short circuiting. It can be done
Page 265
Figure 2
Residence time in one cell (Tr) versus number of cells (units) in relation to short
circuiting.
also with large cells, residence time allowing. For example, four shafts
are enough if total residence time is 30 min.
Another interesting phenomenon can be seen in Table 2, which shows
the average retention times of various materials in the 2 × 38 m3 unit
with 22 min total residence time. The retention time of all minerals is
longer than that of water; a heavy, nonfloatable mineral seems to be
especially slow.
Another benefit of large cells with respect to their size is the ease with
which they are installed and used and the good metallurgy that is
obtained. Outomec has recently developed a new size and type of large
cells: OK-50
Table 2 Average Residence Times of Different
Materials Tested.
Material Tailings Second cell
(min) concentrate
(min)
Water 22 21
Glass, fine (-0.04 mm) 36 25
Glass, coarse (+0.15 mm) 36 21
Baryte, fine (0.04 mm) 37 42
Baryte, coarse (+0.15 53 53
mm)
Pyrite, fine (-0.04 mm) 32 22
Pyrite, coarse (+0.15 mm) 26 36
Page 266
and OK-100 (effective volume 50 m3 or 1750 ft3 and 100 m3 or
3500 ft3, respectively). These cells have inside launders so that
they are part of cell walls and box walls. Special attention in
designing and testing these OK cells has been paid to:
easy installation and froth removal
minimum space requirement and layout work
piping and pumping
The basis for OK-50 and OK-100 cell is the famous OK rotorstator
combination and the U-shaped tank. Due to the modular
construction, the installation of the cell is relatively easy, even with
the huge tank, 100 m3.
Froth removal is arranged so that the inside launders will collect
the froth and discharge it on either side of the cell. Leaving out the
outside launders will substantially save space required for a certain
flotation capacity.
Also froth removal on either side of the cell without any extra
piping will give new opportunities for the layout and will offer
savings in installation costs. Since froth discharge is so high up
from the floor level, it is possible to feed smaller cleaner cells
without pumping, even if they are on the same floor.
It has been noticed that not only good suspension and air dispersion
capabilities give the best metallurgy. Froth removal is as important
an aspect as the others for flotation cell operation. A small but
rather interesting detail is that the froth removal is arranged also for
the froth from the junction box. The junction box froth has been
killed normally by water sprays, and for some minerals, it is
difficult to float again the once-floated particle in the same
flotation stage.
All this combined with excellent metallurgy offers again new
opportunities for better overall economy in the flotation plant.
Rotor speed is one variable in the flotation cell that has not been
optimized yet completely. There are indications that optimum
speed can be found when taking into account:
metallurgical results
power and air consumption
wear rate and lifetime of wear parts
maintenance costs, reagent consumption, etc.
Also the optimum varies with different ores and different types of
the flotation equipment. In a few cases, where plant scale
optimization for rotor speed has been done, variation has been ±
15%.
Page 267
IV
Cleaning Flotation
Although the flotation results achieved with the existing selection
of machinery have been satisfactory, there has been a desire to
simplify and intensify the complex cleaning circuits often plagued
by big circulating loads, small old cells, lots of piping and pumping
with high operational and maintenance costs, etc.
To fulfill these needs Outokumpu's metallurgists set up a
comprehensive development project targeting a totally new type of
flotation machine having the following properties:
high selectivity
high concentration ratio
high-grade concentrate with high recovery without any need for
extra scavengers
simple construction, installation, and operation
reasonable space requirements
As a result of the project, a prototype of an HG machine was born
and installed at Pyhäsalmi mine in 1988.
A
HG Flotation Machine
Figure 3 shows the construction and operation principles of the HG
flotation machine. The bottom part of it, including the tank and the
rotorstator mechanism, is basically similar to the corresponding
part of a standard OK machine. The most distinctive difference
from a conventional OK machine is the upper part, where the rising
froth is forced to pass a tapered channel before falling into the
launders. The amount of flotation air is higher than in the
conventional flotation. The aim of this is to create conditions
favorable to difficult flotation tasks, e.g., fine-particle flotation. It
is successfully performed in an HG cell where an abundance of
finely dispersed air prevails. The cell walls on two sides of the
channel are inclined so that the area through which the concentrate
leaves the channel is considerably smaller than the surface area of
the pulp.
The froth layer is remarkably thicker than in the standard flotation
machine. The froth is squeezed in the channel and washed by water
sprayed into the froth layer by special nozzles, which distribute the
wash water evenly and gently throughout the rising froth without
collapsing it.
As a result of this, a clean, high-grade concentrate is recovered
while the gangue minerals unintentionally lifted by the bubbles are
washed away.
Page 268
Figure 3
Construction and operation principles of
the HG flotation machine.
B
Testing and Operating Experience
1
Tara Mines, Ireland
The tests at Tara Mines can be divided into three parts:
1. Comparison of the plant's first zinc cleaners (6 shafts) with an
OK-0.5-2RM-HG (2 shafts, 0.5 m3 each) and a column. The
average results from several tests are shown in Table 3. The
residence time of the HG cell was one-third of the plant residence
time. The column
Table 3 Comparison of Cleaners at
Tara Mines.
Plant HG Column
Feed, % Zn 29 30 30
Concentrate, % Zn 47 52 49
Recovery of Zn 97 93 93
Page 269
Table 4 Comparison of Cleaners at
Tara Mines.
Plant HG Column
Feed, % Zn 4550 50 45
Concentrate, % Zn 55 54 54
Recovery of Zn 8896 97 90
concentrate was not acceptable as final concentrate, whereas the

HG cell produced it in one step.
2. Comparison of the plant's second and third cleaners with OK-
0.5-2RM-HG and column resulted in the averages shown in Table
4. The results the HG cell produced in one step were similar to
conventional cells, but the residence time was only one-fourth of
the conventional one. The recovery against the column was clearly
better.
3. The upgrading of zinc concentrate after leaching and the
removal of calcium sulfate by OK-0.5-2RM-HG gave the average
results in Table 5. The requirements set for the concentrate were
easily fulfilled.
2
Pyhäsalmi Mine, Finland
After the trial installation, the OK-3-2U-HG has been operating
continuously in the copper cleaning circuit parallel to the
conventional three-stage cleaning. It has upgraded the feed material
(rougher concentrate) from 12% Cu to 26% Cu in a single stage
with a recovery of 90%. The machine has produced 70% of the
total amount of copper concentrate. Compared to the standard OK
machines, both the grade and the recovery have been better in the
HG machine when using the same retention times. Lowering the
results of the HG to the same level with the OK machines increases
the capacity by 200%.
3
Guemassa, Morocco
OK-0.5-2RM-HG tested in various stages of the process. The
results are shown in Table 6. Although the feed material to copper
cleaning circuit is very fine (90% 80 µm), the cleaning can be
successfully performed in three stages instead of the existing five
by using the HG cells. Also the num-
Table 5 Results After
Upgrade.
Feed 54% Zn
Concentrate 55% Zn
Upgrade over 1%
Recovery 99.8%
Page 270
Table 6 Test Results from Guemassa,
Morocco.
1st Pb Combined 4th
Cleaning and
5th Pb
Cleaning
Feed, % Pb 29 63
Concentrate, % 47 71
Pb
Recovery, % 68 94
1st Cu Combined 4th
Cleaning and
5th Cu
Cleaning
Feed, % Cu 9 21
Concentrate, % 20 29
Cu
Recovery, % 78 95
ber of lead cleaning stages can be drastically reduced by adopting

the HG cells.
V
Conclusion
The tests and the full-scale industrial operation have proven the HG
machine to be an effective tool in cleaning circuits and a viable
alternative for column flotation. The column cell has during the
recent years gained a reputation as a high-grade concentrate
producer, but attention has to be paid to the fact that, almost
without exception, the column recovery is low and the tailings need
further purification. Consequently, many plants have been forced to
install afterwards mechanical flotation machine lines to refloat the
column tailings.
Substituting HG cells for old columns simultaneously gives two
benefits: the complicated column cell is replaced by a cost-
efficient, easy-to-use unit, producing a high-grade concentrate and
a disposable tailing, and the need of excess purifiers disappears.
Considering new plants, adopting HG machines enables the
construction of simple and efficient cleaning circuits. Its
dimensions are similar to the conventional flotation machines;
therefore, it does not need extra height to the building, which
typically has to be 1215 m high when installing columns.
VI
Summary
Outomec's Flash Flotation method with Skim-Air flotation
machines has proven that there are still ways to improve
metallurgical results and economic recovery in most flotation
processes. In the existing Skim-Air instal-
Page 271
lations one of the most important economic benefits has been the
improved precious metal recovery. This has been achieved by
reducing the overgrinding of the heavy metal particles, which have
a tendency to stay in the circulating load until they are ground
extremely fine and consequently lost in slimes fraction. The
tendency to separate valuable minerals as early as possible seems
to be the correct way. Flotation in the grinding circuit is one step in
that direction.
Outomec OK-50 and OK-100 flotation cells are additions to the
product line cover all cell sizes from laboratory and pilot plant cells
up to industrial plants. Thus it is possible nowadays to use the
''correct size" cell concept to achieve maximum economic recovery,
at a concentrator of any size.
The latest development, the HG cell for cleaning and difficult
separation processes, has been a success because other types of
flotation machines cannot do these difficult tasks.
Page 273
10
Residence Time Distribution in Flotation Cells
Paul Mavros
Residence time distributions are measured in flotation cells to
establish the gas, liquid, and solid-particle flow patterns in them;
the results are used to identify regions of flow maldistribution in
them and the effects of geometrical, operational, and
physicochemical parameters. Residence time distributions are
analyzed according to either the axial dispersion model or the
mixed zones-in-series one; the merits and problems of both
approaches and the work related to flotation cells and columns is
discussed.
I
Introduction
Flotation cells were initially constructed as mechanically agitated
vessels, with a rotor for dispersing gas and solid particles and
sometimes a stator for enhancing the dispersion of air into fine
bubbles; the cells were practically perfectly mixed. New cell
designs, however, have been either established, such as the
flotation column (1) or are appearing, e.g., the increasingly large
mechanical cells (2), the hydrocyclones (3), the flotation reactors
with separate mixing and disengagement regions (4,5), and the
Jameson column (6,7), among others.
Mixing of the various phases is one of the crucial parameters, since
it affects both particlebubble collection and detachment process; a
high degree of mixing is hence necessary for bringing the solid
particles into
Page 274
contact with the bubbles, but it is otherwise deleterious to the cell
performance, if it happens in the disengagement zone. It is
therefore desirable, when designing or optimizing the flotation cell,
to have as much information as possible about the effects of
various geometrical, physicochemical, and operational parameters
on the mixing process. This information is obtained by performing
residence time distribution (RTD) measurements for the liquid, gas,
and solid phase. (See Note Added in Proof on p. 286).
In this work, the fundamental principles, the problems encountered,
and the work related to the RTD measurements and their
interpretation are discussed.
II
Residence Time Distributions in Mechanically Agitated Cells
It has always been assumed that mechanical flotation cells are well
mixed, especially in the vicinity of the rotorstator region; the fact
that part of the vessel is not really mixed was recognized by
Dinsdale and Berube (8), who determined by RTD measurements
that about 15% of the volume of a 19-m3 Maxwell cell could be
considered as "dead," i.e., where little pulp circulated.
Harris and co-workers studied extensively the mixing behavior of
single flotation cells (911) and used a simple model, combining a
mixed region with recycle, to represent their results. Later, the
amount of the "effective" volume was correlated to several
parameters, e.g., pulp density, flow rate, and agitation speed among
others (12).
Transient experiments with step changes of feed flow rates
disclosed an unexpected oscillatory behavior (13), showing that
another two-zone model, proposed by Arbiter and Harris (14), may
not be totally appropriate. This may be more important in the
currently observed trend of constructing large cells, with volumes
up to 85 m3 (2).
A model comprising three zones, one perfectly mixed for the
impeller region and two, partially mixed zones for the froth and the
intermediatedisengagementregion may be more realistic in
representing the flow pattern inside the agitated flotation cell. Its
main disadvantage, however, would lie in the fact that the more
complex a model becomes, in terms of parameters, the more
sophisticated experimentation is necessary. The benefit from such
an approach would probably be counteracted by the degree of
uncertainty introduced by the variation of key parameters, like the
feed grade or the aeration rate.
Residence time distribution studies have also been performed in
circuits comprising banks of flotation cells, mainly in order to
study the flow pattern between adjacent cells, e.g., the work of
Frew (1517). An intense internal backmixing flow was determined,
with the magnitude of the ratio
Page 275
Figure 1
Experimental residence time distribution obtained on a bank of five
2.7-m3 mechanical flotation cells
(redrawn from Ref. 18 with permission);
comparison with theoretical RTD for the same number of perfectly
mixed zones-in-series.
of backmixing vs. average net tailing flow ranging from 4 : 1 in the
rougherscavenger bank to 11 : 1 in the recleaner bank. The
magnitude of the backmixing flow varied from cell to cell and
depended also upon the aeration rate, the addition of surfactants,
and the volume of the cell.
The measurement of residence time distributions in industrial cell
banks was also used to characterize the design of particular cells.
An example will be given from data published in the literature (18).
The test was performed on a bank of five cells of a particular
manufacturereach cell having a volume of 2.7 m3and the
experimental RTD profile for this bank was found to be skewed to
the left, compared to the theoretical profile expected for the same
number of identical cells (Fig. 1). This indicated a badly designed
cell, since this RTD shift is characteristic of short-circuiting or by-
passing, i.e., of systems where flow remains inside the control
volume for less time than is expected. It is interesting to note that
the manufacture of this particular cell was discontinued by the
manufacturers, because of its poor performance (18).
III
Residence Time Distributions in Flotation Columns
Flotation columns are increasingly becoming the standard
equipment for the processing and beneficiation of fine particle
mineral ores (19).
Page 276
The mixing process affects all three phases present in a column;
most attention has been paid to the liquid phase. Residence time
distribution studies for the gas phase in flotation columns are, to
the best of the author's knowledge, nonexistent; work has been
done, so far, only for stirred reactors and bubble columns.
For the solid particles, there is some evidence that the residence
time distribution for solids is faster than the liquid one, but tending
to come close to the liquid-phase RTD as the particle size decreases
(20). The solids axial dispersion coefficient, which has been used
to characterize the extent of mixing, has been reported as being
either equal to the liquid-phase one (1,21) or smaller (22,23) (for
particles having mean diameters less than 150 µm).
Liquid-phase residence time distribution measurements have been
published by several workers (20,21,2432) and have been recently
reviewed (3335); the main parameters (column dimensions, type of
tracer, method of introduction and measurement, and analysis
technique) are summarized in Table 1. The results obtained so far
show that the mixing of the liquid phase increases with the increase
of the column diameter. This is illustrated in Figure 2 with results
obtained in columns with the same height but with varying
diameters (from 2.5 to 11.0 cm) (36); as the diameter increases, the
peak of the RTD becomes less sharp and at the largest diameter it
displays a behavior corresponding to a highly mixed vessel. The
mixing process is also affected (adversely) by an increase in bubble
diameter (29) or pulp density.
Results obtained in industrial-scale as well as laboratory-scale
columns indicate that the aspect ratio of the column, i.e., the
dispersion height : diameter ratio, may also affect the mixing
process; the results obtained in the largest laboratory-scale column
(dc = 11.0 cm), with an aspect ratio of 150 : 11 or 13.6 : 1 (Fig. 2),
are interestingly similar to results obtained in an industrial-scale,
91.4-cm (36-in.) square column (25), which had an aspect ratio of
9.7 (calculation based on the circular column diameter equivalent
to the square one, dc = 103.1 cm).
The aforementioned experimental results in flotation columns with
internal diameters ranging from 4.0 to 11.0 cm (36) indicate also
that the mixing is affected more by the superficial liquid velocity
than the (superficial) gas velocity; however, more work, mainly in
larger columns, is required to establish the scale-up rules necessary
for the design of flotation columns.
IV
Interpretation of Residence Time Distributions
There are essentially two ways of analyzing the residence time
distribution results: (a) If it is assumed that the mixing of the
system is intense, approaching the condition of perfect mixing,
then the "zones-in-series" model
Page 277
Table 1 RTD Measurements in Flotation Columns.
Column
dimensions
(dc/ht, mm) Phasea Tracer Conditionsb Modelc Ref.
82/4570 L KCl O-C ADM 21
80/800 L KCl C-C NZS 24
516/12000d S,L Dyee/MnO2 O-C ADM 25
1031/12000d S,L Dye/MnO2 O-C ADM 25
102/4000 L KCl O-C ADM 26
102/4000 L KCl C-C ADM 27
54/2200 S 89Au191 C-C ADM, NZS 28
54/2200 L NaCl O-C ADM, NZS 29
192/1994 L NaCl C-C ADM, NZS 31
2500/13000 L LiCl ?-? ? 32
25/1500 L KCl C-C ADM, NZS 36
54/2300 S 89Au191 C-C ADM, NZS 47
150/10000 L Li? ?-C ADM 48
3250 + FC/
11450f L Li? ?-C ? 48
aS, Solids, L, Liquids.
bTracer introduced through wash water.
cADM, Axial Dispersion Model; NZS, N Zones-in-Series.
dColumn height not specified.
eFluorescein.
fA froth crowder was installed at the top of the column.
be used; (b) if, on the other hand, it is considered that very little
mixing occurs in the cell, then the axial dispersion model may be
considered.
These two approaches will be presented here, and their respective
problems and benefits will be discussed.
A
The Axial Dispersion Model
The axial dispersion model postulates that mixing is due to
departures from ideal plug flow, caused by disturbances in the flow
front. It is strictly valid for cases where it is expected that very little
mixing occurs, e.g., in a column. In that sense, its use for
representing flow in mechanically agitated cells would not be
appropriate.
Page 278
Figure 2
Experimental residence time distributions obtained in various laboratory-
scale flotation columns.
Comparison with results obtained in a 91.4-cm (36-in.), square, industrial-
scale column.
From Ref. 25.
An equation similar to the Fickian diffusion may be written for the
concentration C of some species:
where D is the axial dispersion coefficient characterizing the

"macroscopic" diffusion, x is the linear dimension (i.e., the distance
along which the concentration varies), and D is the axial dispersion
coefficient, corresponding to a macroscopic diffusion coefficient.
The equation is usually rewritten in dimensionless form:
where Nd is the dimensionless dispersion number (Nd = D/uL),

which was used by Levenspiel (37) to characterize mixing in
chemical reactors (its inverse, the Peclet number, is often used
instead of Nd); L is the vessel length; z is the dimensionless linear
dimension; q = t/t, where t is the mean residence time in the vessel:
Page 279
(V being the cell volume and Q the volumetric flow rate entering the
cell).
The integration of equation (2), in order to obtain the concentration
profile with respect to time for a given system, requires boundary
and initial conditions. The initial condition depends upon the type of
experiment performed or simulated. For example, when a step input
is performed, then C(x, t = 0) = C1 and C(x = 0, t > 0) = C2, whereas
if an impulse disturbance is introduced into the system, then C(x, t =
0) = C1, C(x = 0, t = 0) = C2, and C(x = 0, t > 0) = C1 (back to the
initial state).
The establishment of the correct boundary conditions is related to the
experimental technique envisaged (38,39). When the tracer
introduction or detection is done inside the cell, then the
corresponding boundary condition is considered as being of the
"open" type. If the injection/detection point is outside the cell (done
through the feed/tailings/concentrate line), then the corresponding
boundary condition is of the "closed" type.
When the openopen boundary conditions are considered, the end
effects caused by restrictions in flow are absent. However, the
residence time distribution E(q) may not be defined easily anymore
since, in principle, a fluid particle may travel upstream and hence
reenter the system more than once (the same also holds for particles
going by the detection point). In such cases, involving either one or
both boundary conditions of the open type, analytical expressions are
available for the concentration of the tracer at the exit C(q), which is
not necessarily the residence time distribution E(q) (40).
In the case of the closedclosed boundary conditions, the following
equation is obtained after translation of equation (2) into the Laplace
domain:
where q = (1 + 4s/Pe)1/2. Its inversion, in order to obtain the time-

domain theoretical residence time distribution for a given value of
Pe, has not been yet achieved. Only pseudoanalytical equations have
been proposed, for a Dirac delta (impulse) tracer input (40) or a step
input (41), both relying on an infinite series requiring the roots ln of
the following transcendental equation:
The Dirac delta equation has been shown to diverge for Peclet values
as low as 7 (42). The profile generated by differentiating the
"washout" profile yielded by the latter, on the other hand, diverges
for Pe > 10.
Page 280
Fortunately, it is possible to determine the Pe number corresponding
to a particular experiment by calculating the mean residence time t
and the variance s2 of the distribution; these are related to Pe by the
following equation:
Thus, it is possible to use equation (6) to estimate the Peclet or the

dispersion number for a particular experimental condition; however,
this technique has a disadvantage. Since it relies on the variance s2
of the experimentally determined distribution, long or short tails will
affect the accuracy of the determination. For this reason, it has been
suggested to determine the Peclet or the dispersion number by
matching the experimental to theoretically calculated profiles and
determining the Peclet number for the optimum match through the
sums of squared deviations. But this again necessitates the use of
some profile-generating equation, and this is possible only for
certain Pe ranges. An alternative method, using a numerical
integration of equation (2) has been proposed to circumvent this
problem (43).
As the value of Pe increases, the profile produced by the various
equations tend to become identical, even if the wrong boundary
conditions are used (44); for Pe values above 16 (39) or 20 (44), it
has been proposed that the openopen equation (45)
may be used without serious loss of accuracy. For the intermediate

region (10 < Pe < 16), it has been suggested (42) that the openopen
equation (7) should be used first to gain a first estimate of Pe (Peoo)
and then correct it to find the value corresponding to the
closedclosed boundary conditions (Pecc) by using the following
relation proposed by Ityokumbul et al. (45):
B
The Mixed Zone Model
The alternative approach is to consider that the cell is either a single
perfectly mixed region or that it behaves like a series of perfectly
mixed zones connected in series, like the banks of mechanically
agitated cells. The theoretical RTD profile is given for any number N
of cells in-series by the following equation (37):
Page 281
Once the mean t and variance s2 of an experimental residence time
distribution are determined, the value of N may be calculated:
It is obvious that in most cases the experimentally determined value

of N will be some real number, whereas equation (9) was initially
derived only for integers; however, this has been generalized (46) to
allow for real values of N:
where G(N) is the gamma function:
Therefore, in order to determine the value of N corresponding to a

particular experiment, it is possible to calculate it either through
equation (10)which suffers from the same problems related to long
or short tails, as in equation (6)or by generating theoretical profiles
using equations (11) and (12) and determining the value of N
yielding the optimum match.
One of the advantages of this approach is that a variety of models
may be constructed using some combination of perfectly mixed
and/or plug flow zones. Theoretical E(q) profiles for various such
models have been published (see, for example, Refs. 12 and 37).
These models are relatively simple in that the number of parameters
involved is usually small, yet they are able to represent quite a
variety of residence time distribution corresponding to widely
different mixing conditions.
Such a simple model, for example, consists of a series of perfectly
mixed zones connected in series, but with "backflow" allowed
between adjacent cells, i.e., pulp flows from each cell to both the
previous and the subsequent cells. This is a two-parameter model,
involving the number of mixed zones-in-series N and the ratio of
backflow vs. net incoming pulp flow l. This is quite realistic in case
of banks of mechanically agitated cells and has been used by Frew
and co-workers (1517) to determine the amount of internal
circulation. It also has been used in a similar fashion to determine
the internal flow in a flotation column (24), revealing the intense
recirculating flow pattern established inside the flotation columns.
The facility of this model will be demonstrated by analyzing the
aforementioned RTD results obtained in two industrial-scale square
columns; the smaller one had a square side of 45.7 cm ("18 in.,"
which corresponds to an equivalent circular-column diameter of 51.6
cm), while the larger one had a square side of 91.4 cm ("36 in.,"
having an equivalent circular-column
Page 282
diameter of 103.2 cm). Both have an approximate height of 10 m
(25). Residence time distributions for each column are presented in
Figure 3, together with several theoretical profiles for a series
comprising up to five perfectly mixed zones. It is obvious that
although the contents of both columns seem highly mixed, the
simple mixed zones-in-series model is not adequate to represent
their internal state.
The backflow model was used to model these data. A search for
optimum matching was performed for both the number of zones N
and the backflow parameter l, by calculating the sum of squared
deviations of the predicted from the experimental value. The results
of the calculations are illustrated in Figure 4 for the first column
(18-in.) and in Figure 5 for the second one (36-in).
For the smaller column, if the number of zones N = 5 is taken as
the starting point, then a degree of backflow of l = 1.3 is required
to achieve a reasonable representation of the experimental data.
This is equivalent to saying that inside the column more than twice
the amount of liquid circulates, as compared to the incoming flow.
However, an even better match is obtained with a larger number of
zones (N = 10) and a higher degree of backflow (l = 2.5). The
increased backflow is due to the fact that the larger the number of
zones-in-series, the less mixed is a given vessel; hence, an increase
in the backflow is necessary to represent the mixing state of the
column.
The larger column (36-in.; Fig. 5) seems even more mixed, with
fewer zones required to model the experimental data. Here, again, a
degree of backflow is required to obtain a good fit. Surprisingly,
the best fit is obtained with only three zones-in-series, compared to
ten or even more for the smaller column. The enhancement of the
mixing process inside the column, due to its enlargement, is thus
illustrated even by the model parameters required to represent the
experimental data.
In principle, several pairs of (N,l) will yield acceptable matching
profiles to a given set of experimental data. This is illustrated by
the 2-D optimization of the (N,l) values, presented in Figure 4
(top), where it seems that for N > 8 and l > 2, any combination of
parameter values will give an acceptable match. A method, based
on the estimation of the pulp velocity circulating inside the column,
has been proposed (49) to obtain an independent estimate of the
backflow parameter, so that the number of zones is the sole
parameter to be determined by fitting. Alternatively, it has been
suggested that inside bubble columns (hence, probably inside
flotation columns, too) a flow structure is established, with a
number of zones corresponding to the aspect ratio of the column
(50)thus leaving l as the sole parameter to be fittedalthough this
may be questionable in the large columns, with internal diameters
up to 3 m, that tend to be constructed
Page 283
Figure 3
Experimental residence time distributions obtained in industrial-scale
columns.
(Data redrawn from Ref. 25 with permission.)
Comparison with theoretical perfectly mixed zones-in-series profiles
(number of zones in series N = 1,2,3, and 5).
Page 284
Figure 4
Two-
dimensional search for minimum sum of squared deviations of predicted from experimental residence
time distributions; 18-in. column.
From Ref. 25.
Fit of theoretical with experimental residence time distributions for various zone number (N) and backflow
parameter (l) pairs.
Page 285
Figure 5
Fit of theoretical with experimental residence time distributions for various zone numbers (N)backflow
parameter (l) pairs for the 36-in. column.
From Ref. 25.
lately, where a more irregular flow pattern may be conceived. More work will be required
to establish whether one of these parameters may be independently set, possibly by relating
them to the metallurgical performance of the column or measurements of fluid velocities
inside the column, among others. The degree of uncertainty associated with the
experimental determination of the residence time distribution in a given column for a
particular set of operating conditions should also be taken into account.
V
Conclusions
Residence time distributions are practically the only tool available to the engineer for
determining regions inside a flotation cell that do not correspond to the flow pattern sought
to be established, as well as for the study of the way these flow patterns are affected by the
various parameters involved in flotation, without having recourse to other, rather
sophisticated experimental techniques, like Laser Doppler Anemometry, or perhaps
unpractical ones, like hot-wire anemometry.
The axial dispersion model, with the dispersion number as its sole parameter, is more
appropriate to use in cases with little mixing; flotation columns results have been
extensively analyzed following this model. However, the lack of a precise method for the
determination of the dispersion number and the need to resort to numerical tricks is a
disadvantage.
Page 286
The perfectly mixed zones-in-series model, on the other hand, is
capable of dealing with all kinds of flow patterns, through the
construction of models incorporating plug flow as well as mixed
regions. One additional advantage of this modeling approach is that
in the case of complex ore systems, involving fast-and slow-
floating minerals, as well as gangue, the determination of the type
and number of zones corresponding to the optimum representation
of the internal flow may then be used to examine possible
operation improvements, considering the whole system under its
mass transfer aspect, as has already been done in chemical reactors
involving gasliquid reactions with complex kinetics.
A final comment about the experimental method used for the
determination of the residence time distribution. From the two
techniques available, the step change is much easier; however, the
region of feasible profiles, which is enclosed between the RTD for
plug flow and the RTD for a single perfectly mixed region, is very
narrow. An impulse is rather more difficult in conducting but yields
much more informative profiles, since an erroneous profile, due to
a poor experimental procedure, is immediately obvious (usually, its
peak lying on the right side of the q = 1 lines). At the same time,
the plug flow profile is completely different from the perfectly
mixed one and allows a much wider variety of flow conditions to
be visualized.
Note Added in Proof

Yianatos and co-workers have started investigating the gas-phase
residence time distribution in an industrial-scale flotation column
(51).
References
1. J. A. Finch and G. S. Dobby, Column Flotation, Pergamon Press,
Oxford. (1990).
2. S. T. Hall, Mineral and coal processing, Mining Annual Review
(1992).
3. D. R. Spears and C. E. Jordan, ''A Novel Rapid Flotation System
Based on the Air-Injected Hydrocyclone," SME/AIME Annual
Meeting, Denver, 2528 Feb., preprint no. 9184 (1991).
4. R. Imhof, Pneumatic flotationA modern alternative, S. Afr.
Mining World, 8: 38. (1989).
5. G. D. Hood and C. E. Jordan, "Improved Flotation Kinetics by
Intense In-Line Agitation," SME/AIME Annual Meeting, Denver,
2528 Feb., preprint no. 91142 (1991).
6. G. J. Jameson, A new concept in flotation column design,
Column Flotation '88 (K. V. S. Sastry, Ed.), SME/AIME, Littleton,
CO, p. 281. (1988).
7. G. J. Jameson, "The Jameson ColumnA High-Intensity Flotation
Cell,"
Page 287
Proceedings Column Flotation Workshop, TUNRA, Adelaide
2728 July, pp. I1I14 (1989).
8. J. D. Dinsdale and Y. Berube, A characterization of
hydrodynamics in a 700 cu. ft. Maxwell flotation cell, Can. Metall.
Q., 11: 507. (1972).
9. C. C. Harris, A. Chakravarti, and S. N. Degaleesan, A recycle
flow flotation machine model, Int. J. Miner. Process., 2: 39. (1975).
10. C. C. Harris, A recycle flow flotation machine model:
Response model to parameter changes, Int. J. Miner. Process., 3: 9.
(1976).
11. C. C. Harris and A. Cuadros-Paz, Species interaction in
flotation: A laboratory scale semi batch study, Int. J. Miner.
Process., 5: 267. (1978).
12. S. P. Mehrotra and A. K. Saxena, Effects of process variables
on the residence time distribution of a solid in a continuously
operated flotation cell, Int. J. Miner. Process., 10: 255. (1983).
13. M. Greaves and B. W. Allan, Steady-state and dynamic
characteristics of floatation in a single cell, Trans. Inst. Chem.
Eng., 52: 136. (1974).
14. N. Arbiter and C. C. Harris, Froth Flotation (D. W. Fuerstenau,
ed.), AIME, New York, p. 215. (1962).
15. J. A. Frew, Back mixing in industrial flotation banks. 1. Effect
upon the determination of flotation rate coefficients, Int. J. Miner.
Process., 13: 239. (1984).
16. J. A. Frew and K. J. Davey, Back mixing in industrial flotation
banks. 2. Effect upon metallurgical performance. Int. J. Miner.
Process., 15: 239. (1985).
17. S. G. Wilson and J. A. Frew, Effect of operating variables upon
liquid back-mixing in a laboratory flotation bank, Int. J. Miner.
Process., 16: 281. (1986).
18. B. J. Arnold and F. F. Aplan, "A Practical View of Rate and
Residence Time in Industrial Coal Froth Flotation Circuits,"
Proceedings 5th Annual Intern. Pittsburgh Coal Conf., 1216 Sept.,
pp. 328341 (1988).
19. K. A. Matis and P. Mavros, Foam/froth flotation. Part II.
Removal of particulate matter, Sep. Purif. Methods, 20: 163.
(1991).
20. J. B. Yianatos and L. G. Bergh, "Radioactive Tracer Technique
Applied to Industrial Flotation Columns," Proceedings III Intern.
Miner. Proces. Symposium (G. Onal, ed.), Istanbul, 1113 Sept., pp.
537544 (1990).
21. R. G. Rice, A. D. Oliver, J. P. Newman, and R. J. Wiles,
Reduced dispersion using baffles in column flotation, Powder
Technol., 10: 201. (1974).
22. M. Xu and J. A. Finch, Solids mixing in the collection zone of
flotation columns, Miner. Eng., 5: 1029. (1992).
23. P. T. Mills, J. B. Yianatos, and C. T. O'Connor, The mixing
characteristics of solid and liquid phases in a flotation column,
Miner. Eng., 5: 1195. (1992).
24. P. Mavros, N. K. Lazaridis, and K. A. Matis, A study and
modelling of liquid-phase mixing in a flotation column, Int. J.
Miner. Process., 26: 1. (1989).
25. G. S. Dobby and J. A. Finch, Mixing characteristics of
industrial flotation columns, Chem. Eng. Sci., 40: 1061. (1985).
26. M. Xu and J. A. Finch, The axial dispersion model in flotation
column studies, Miner. Eng., 4: 553. (1991).
Page 288
27. M. Xu and J. A. Finch, Estimating vessel dispersion number in
flotation columns, Column '91, Vol. 2 (G. E. Agar, B. J. Huls, and
D. B. Hyma, eds.), CIM, Sudbury, Ontario, p. 437. (1991).
28. C. M. Goodall and C. T. O'Connor, "Residence Time
Distribution of the Solid and Liquid Phases in a Laboratory
Column Flotation Cell," Proceedings Int. Coll. Developments in
Froth Flotation, Gordon's Bay, Cape Town, vol 2. (1989).
29. P. J. T. Mills and C. T. O'Connor, The modelling of liquid and
solids mixing in a flotation column Miner. Eng., 3: 567. (1990).
30. J. B. Yianatos and L. G. Bergh, Column '91, Vol. 1 (G. E. Agar,
B. J. Huls and D. B. Hyma, Eds.), CIM, Sudbury, Ontario, p. 221.
(1991).
31. M. H. Moys, J. Engelbrecht, and N. Terblanche, The design of
baffles to reduce axial mixing in flotation columns, Column '91,
Vol. 1 (G. E. Agar, B. J. Huls and D. B. Hyma, eds.), CIM,
Sudbury, Ontario, p. 275. (1991).
32. R. Espinoza-Gomez, N. W. Johnson, J. D. Pease, and P. D.
Munro, The commissioning of the first flotation columns at Mount
Isa Mines Ltd, Processing of Complex Ores (G. S. Dobby and S. R.
Rao, eds.), Pergamon Press, Oxford, p. 293. (1989).
33. A. R. Laplante, J. Yianatos, and J. A. Finch, On the mixing
characteristics of the collection zone in flotation columns, Column
Flotation '88 (K. V. S. Sastry, Ed.), SME/AIME, Littleton, CO, p.
69 (1988).
34. J. A. Finch and G. S. Dobby, Column flotation: A selected
review. Part I, Int. J. Miner. Process., 33: 343. (1991).
35. P. Mavros, Mixing in flotation columns. Part I. Axial dispersion
modelling, Miner. Eng, 6: 465. (1993).
36. P. Mavros and A. Daniilidou, Mixing in flotation columns. Part
II. Liquid-phase residence time distributions studies, Miner. Eng.,
6: 707. (1993).
37. O. Levenspiel, Chemical Reaction Engineering, 2nd ed., Wiley,
New York, p. 273. (1972).
38. P. Mavros, Mixing and hydrodynamics in flotation cells,
Innovations in Flotation Technology (P. Mavros and K. A. Matis,
eds.), Kluwer, Dordrecht, p. 211. (1992).
39. E. B. Nauman and B. A. Buffham, Mixing in Continuous Flow
Systems, Wiley, New York, p. 101. (1983).
40. K. R. Westerterp, W. P. M. van Swaaij, and A. A. C. M.
Beenackers, Chemical Reactor Design and Operations, 2nd ed.,
Wiley, Chichester, p. 187. (1987).
41. H. Brenner, Diffusion model of longitudinal mixing in beds of
finite length, Chem. Eng. Sci., 17: 229. (1962).
42. P. Mavros, Validity and limitations of the closed-vessel
analytical solution to the axial dispersion model, Miner. Eng., 5:
1053. (1992).
43. M. Xu, J. A. Finch, and A. R. Laplante, Numerical solution to
axial dispersion model in flotation column studies. Can. Metall. Q.
30: 71. (1991).
44. A. Seidel-Morgenstern, Analysis of boundary conditions in the
axial dispersion model by application of numerical Laplace
inversion, Chem. Eng. Sci., 46: 2567. (1991).
Page 289
45. M. T. Ityokumbul, N. Kosaric, and W. Bulani, Parameter
estimation with simplified boundary conditions, Chem. Eng. Sci.,
43: 2457. (1988).
46. B. A. Buffham and L. G. Gibilaro, A generalization of the
tanks-in-series mixing model, Am. Inst. Chem. Eng. J., 14: 805.
(1968).
47. C. M. Goodall and C. T. O'Connor. RTD studies in a flotation
column. Part 1: The modelling of RTD in a laboratory column
flotation cell, Int. J. Miner. Process., 31: 97. (1991).
48. B. Alford, "The Application and Scale-Up of Flotation
Columns at Paddington," Proceedings Column Flotation Workshop,
TUNRA, Adelaide 2728 July, pp. E1E25 (1989).
49. P. Mavros, Mixing in flotation columns. III. Internal
circulation, Miner. Eng., 6: 1167. (1993).
50. J. B. Joshi and M. M. Sharma, A circulation cell model for
bubble columns, Trans Inst. Chem. Eng. 57: 244. (1979).
51. J. B. Yianatos, L. G. Bergh, O. U. Duran, F. J. Diaz, and N. M.
Heresi, "Measurement of Residence Time Distribution of the Gas
Phase in Flotation Columns," Proceedings Miner. Eng., Cape
Town, S. Afr., 2527 August, 17 pp. (1993).
Page 291
11
Column Flotation
James A. Finch and A. Uribe-Salas
McGill University, Montreal, Quebec, Canada
Manqiu Xu
J. Roy Gordon Research Laboratory, INCO Limited,
Mississauga, Ontario, Canada
I
Introduction
A schematic of the design of flotation columns, which is the
subject of this chapter, is shown in Figure 1(a) with some typical
flow rates of the various streams (given as volumetric rates per unit
cross-sectional area of column J). Three characteristic features are
the use of a sparger to generate bubbles near the base, a counter-
current slurry/bubble flow in the collection zone, and a deep froth
zone (0.52 m) coupled with the use of wash water to induce a
"cleaning action." This design was first patented in Canada in the
early 1960s and is sometimes known as the Canadian or
conventional column. It took some 20 years before column
flotation was seriously considered, but since 1980 applications
have become worldwide. Numerous studies have shown that the
column is capable of an improved grade-recovery relationship
compared to mechanical cells (usually when the column is operated
with a net downflow of waterthe difference being supplied by the
wash watercalled a positive bias), hence the widespread interest.
Much of the experience with columns up to 1989 is available in the
proceedings of an international symposium held in Phoenix,
Arizona in 1988 (1) and in a book (2) published in 1990. Since then
there has been a second international symposium held at Sudbury,
Ontario in 1991 (3), at least three workshops (46) and special
sessions at conferences (e.g., SME, 1992) and a book in Russian
(7) part of which is apparently about to
Page 292
Figure 1(a)
Schematic of Canadian or conventional column (with typical nomenclature
and flow rates in cm/s).
Page 293
Figure 1(b)
Example of cell design used to perform conductivity measurements in multiphase systems.
From Ref. 18.
become available in English. (There are interesting developments occurring in
Russia in parallel to those in the West.) Applications continue to expand. In
Canada, for example, units have been installed recently at Placer Dome's
Cambell concentrator (8) and Westmin Resources' Myra Falls Operations (9).
Columns have grown in size; for example, Freeport McMoran operates 13
columns each 3 m in diameter by 15 m tall for Cu cleaning at their mine site in
Irian Jaya. And circuits entirely composed of columns are in operation, for
example at Peak Gold, Cobar, Australia for Cu, Pb, and Zn flotation. Alternative
flotation devices are being tested against the column, e.g., the Jameson and
Outokumpu HG cells (10,11). The Jameson cell is now widely used in Australia
(12), and the first installation in Canada was in 1992 at the Kidd Creek Division
of Falconbridge Ltd.
To try to keep abreast of what remains a fast-changing field, two reviews have
been published by one of the authors (JAF) with G. Dobby (13,14). This chapter
is written as the third in the series. Since its completion (Oct. 1992)
developments have laid the ground for the fourth review: new spargers and wash
water delivery systems, "virtual" columns, on-line gas holdup measurement in
slurries, the role of solids on gas holdup, dynamic modeling, neural networks for
improved level detection accuracy, and at least six novel designs of flotation
device just to mention a few to whet the appetite. This review concentrates on
the properties of the collection and froth zones drawn from fundamental and
operational studies.
Page 294
II
Collection Zone
A
Gas Holdup
1
Measurement in GasSlurry (Flotation) Systems
Gas holdup Îg is a possible control variable in flotation. On-line
measurement in two-phase (e.g., gaswater) dispersions using
pressure or conductivity-based techniques is well established (15).
Measurement in gasslurry systems requires knowledge of the
appropriate slurry property (i.e., slurry density rls or conductivity
kls) and, in the case of conductivity, the appropriate model.
Uribe-Salas et al. (16) have evaluated the conductivity-based
technique. Using a grid electrode cell (Fig. 1) to give the desired
characteristics of uninterrupted flow and a near-uniform electric field
(as implicit in most models of the conductivity of dispersions), they
found Maxwell's model (17) was suitable in the form
where
The practical problem is one of making independent, simultaneous

measurements of klsg and kls. Some possibilities in flotation
columns have been considered (16).
The most direct demonstration of the technique, however, has been
in the Jameson cell (18). Using measurements in the downcomer (to
give klsg) and in the feed line (kls), acceptable estimates of holdup
were found as compared with direct measurements (obtained by
isolating the section of downcomer enclosed by the conductivity
cell) (Fig. 2). Tests in plant have confirmed the potential of this
method for on-line gas holdup estimation (19).
Apart from gas holdup estimation, conductivity signals also give
advance warning of disturbances in operation (20): Figure 3, for
example, shows conductivity is more sensitive than pressure to slug
formation in the Jameson cell downcomer.
2
Axial Profiles
Flotation columns, being up to 10 m or more in height, exert a
hydrostatic pressure drop of about 1 atm. This creates an axial
distribution (or profile) of gas holdup in the collection zone. [A
profile also exists in the froth zone, but this is due to coalescence and
water drainage rather than pressure changes (21).]
Page 295
Figure 2
Gas holdup from conductivity measurements and Maxwell's model (equation 1)
against that obtained from direct measurements for slurryair system.
Figure 3
Variation in conductivity or pressure inside the downcomer of a Jameson cell when a slug is formed.
Page 296
Theoretical examination suggests the holdup will decrease with
depth (2). Uribe-Salas et al. (22) verified this experimentally from
pressure gradient measurements in two-phase gaswater dispersions
(Fig. 4). As a general observation, gas holdup is halved over a depth
of 10 m.
Predicting this variation is not straightforward. In Figure 5, three
models are presented. The first assumes gas holdup is simply
proportional to bubble volume (which is inversely proportional to
pressure), i.e.,
where the subscript o refers to the top of the column. The second is
the drift flux approach (2), which has the approximate analytical
form
And, the third is one of two models from Zhou and Egiebor (23)
(rewritten to refer to distance H from the top of the column, i.e., Ho
= 0),
Figure 4
Gas holdup as a function of height in the collection
zone of a 0.91m × 13.5 m column (Strathcona
concentrator, Falconbridge, Ltd.) at three
superficial gas rates in batch water with frother
concentration ~ 20 ppm Dowfroth 250.
From Ref. 22.
Page 297
Figure 5
Comparison of three models with data in Figure 4 for
Jg = 2.03 cm/s: (a) equation 2; (b) equation 3;
(c) equation 4 (with Ubs = 0.1057 m/s; b = 0.00022).
where Ubs is Jgo/Îgo (i.e., the bubble swarm velocity, assumed
constant along collection zone) and b is obtained by data-fitting
and is linked to a reduced bubble buoyancy due to adsorbed frother.
The exponential form [Eq. (4)] clearly corresponds best to the data;
however, the physical meaning of the value of b (= 0.00022 in Fig.
5) is unclear as it implies a vanishingly small buoyancy force.
The data are for gaswater dispersions, but similar profiles are
expected in a gasslurry system. (Measurement here is more
difficult as the appropriate slurry parameter must be measured as a
function of height also.) The profile may need to be included in
refinements to scale-up routines, where gas holdup is usually
considered constant with height (e.g., Ref. 2).
Axial variation in gas holdup will need to be considered in order to
compare reported relationships of gas holdup vs. gas rate for
industrial-scale columns. To illustrate, if gas is defined on an
atmospheric pressure basis, i.e., at the top of the column Jgo (the
common practice), then different relationships are found at
different heights [Fig. 6(a)]: Defining gas rate on a local basis tends
to give a common relationship [Fig. 6(b)].
One question is whether different flow regimes exist at different
heights. This has been observed under some circumstances in
bubble column reactors (24). Figure 6(b) shows that for this
column operation the flow regime was essentially unchanged with
height.
Page 298
Figure 6
Gas holdup as a function of gas rate in a 0.91 m × 13.5 m column at three
locations (top: H = 1.75 m; middle: H = 6.89 m; bottom: H = 11.23 m): (a) gas rate
defined at top of column (i.e., Jgo); (b) gas defined locally (i.e., allowing for
measured pressure).
Data such as in Figure 6 provide a test of the assumption in the model of
Zhou and Egiebor, that the swarm velocity (Jg/Îg) in bubbly flow is a
constant independent of Jg and H. The data are in general support (Fig.
7). The constancy of Jg/Îg with Jg is a common observation and is
usually taken as characteristic of bubbly flow. Constancy with height,
however, does not seem to have been widely tested.
3
Radial Profiles
Variation in axial gas holdup is driven largely by pressure changes.
Radial variations, on the other hand, have their origin in the bubble
generation systemthe type and arrangement of sparger(s), the uniformity
of gas injection (across the column diameter), gas rate, and frother
concentration. Uniform radial gas holdup would seem the desirable
situation and is implicit in most discussions, an example being that of
Figure 6.
At least three profile shapes are known (Fig. 8): parabolic, saddle-
shaped, and W-shaped. To illustrate, Figure 9(a) shows the saddle-shaped
profile obtained with one centrally located Cominco-type sparger. As a
generality, the profiles seem to shift from W to saddle to parabolic as gas
rate is increased (25); however, factors such as bubble size and column
dimensions complicate generalizations.
The presumption in Figure 8 is that gas injection is symmetrical. In
Page 299
Figure 7
Jg/Îg (or swarm velocity) as a function of Jg from data in Figure 6 (Îg, Jg are the local values): the apparent
decrease in Jg/Îg with Jg is not considered significant, although it is predicted by drift flux analysis
(e.g., see Ref.69).
Figure 8
Three known radial gas holdup profile shapes: parabolic (27); saddle-shaped (26); W-shaped (25).
Page 300
Figure 9
Profiles in a 0.91 m × 13.5 m flotation column 10 m above a Cominco-
type sparger: (a) centrally
located and (b) off-
center. (Gas holdup was measured using a multicell conductivity probe
the probe position relative to the sparger is shown.)
(Adapted from Ref. 25).
practice this may not be the case, especially with multisparger
systems at varying stages of deterioration. An extreme situation is
shown in Figure 9(b): The one off-center sparger clearly produced
an asymmetric profile, with Îg highest above the sparger position,
although otherwise it appears to have distributed the gas quite
successfully.
Figure 9 gives some appreciation of what uniform means. A profile
will exist, but the experience to date in flotation columns is that the
variation is small. For spargers in both locations, the maximum
variation in gas holdup is only about 4% (absolute), which was
typical (25).
The ability of the off-center sparger to distribute the gas may be of
some practical consequence. For example, if the column was
baffled, this ability
Page 301
would be lostindeed it would induce circulatory flows between the
baffled chambers (28). The ability to distribute is attributed here to
the partially horizontal flow of the powerful gas jet given by this
type of sparger; the same may not hold, therefore, for porous media
spargers.
Knowledge of the gas holdup profiles (radial and axial) is a step
toward understanding the gas and slurry flow patterns existing in
the column and their dependence on design and operating
parameters and, ultimately, their effect on metallurgy. Measuring
the profiles does offer a possible diagnostic tool to detect
maldistribution of gas and to compare sparging systems.
B
Mixing
1
Liquid RTD
There has been considerable debate over the appropriate model for
the liquid RTD (e.g., see Ref. 13). For a wide range of conditions,
the axial dispersion model is adequate, provided the appropriate
boundary conditions and data-fitting routines are used (29,30). An
attractive feature of this model is that only a single fitting
parameter is required, the vessel dispersion number Nd (or its
inverse, the Peclet number), which in turn can be related to design
and operating variables. [The other single parameter model, the N
perfect tanks-in-series model does not fit, although more complex
compartment models can (31).]
In most studies, closedclosed boundaries are appropriate to the
experimental conditions. A quasianalytical solution is available for
this case but has been found difficult to apply, failing to converge
properly for Nd < 0.14, for example (32). Xu and Finch (29) and
Mills and O'Connor (30) recommend a numerical solution with a
least-squares fit to the experimental data. It is certainly desirable
that the same solution/fitting procedure be used in order to compare
Nd from different sources. Since this is not yet the case, some
caution is required when reviewing the available correlations.
Table la lists NdL correlations that have been proposed based on
column flotation work. Figure 10 compares the Nd estimates as a
function of dc for each correlation. For dc < 0.75 m the models
converge, but the extrapolations for larger (i.e., practical) columns
diverge significantly. This reflects the increasing insensitivity of
the RTD to Nd as mixing increases. For Nd > 1 perfect mixing can
be assumed.
A variable not included in Table la is verticality. Small tilts from
the vertical visibly disturb the flow pattern in laboratory columns
(31,110), which has been included in some dispersion coefficient
correlations (e.g., Ref. 110). The impact on industrial-size columns,
while considered (2), has not been evaluated: A significant effect
over an angle of tilt likely to be missed by normal column erection
procedures seems unlikely. In erecting
Page 302
Table 1 Mixing Correlations.
a. Correlations for Liquid Mixing Based on Flotation Column Testwork
Comments; Nd
Investigators Correlations (in SI units) estimation
Finch and dc < 1.0 m; gasliq-
Dobby (2) uidsolids; open
vessel, weighted
moments fit
Laplante et al. dc < 1.8 m; gasliq-
(33) uidsolids; closed
vessel, moments
matching
Mankosa et dc < 1.0 m; F-curve;
al. (34) closed vessel, least-
squares fit
Xu and Finch dc < 2.5 m; percent
(31) solids and vertical-
ity, little effect;
closed vessel (nu-
merical solution),
least-squares fit
Alford (35) dc < 1.2 m; gas rate,
percent solids little
effect
b. Correlations Between Liquid and Solid Dispersion Coefficient.
Investigators Correlations Comments
Rice et al. EL = Ep dp£ 150 µm
(36)
Finch and
Dobby (2)
Xu and Finch EL = 0.8 Ep dp£ 150 µm
(37),
Mills et al.
(100)
Hidaka et al. EL = 1.1Ep dp > 2.2 mm
(38)
Note: NdL = EL/(uiHc) and ui = JL/(1 - Îg).
laboratory columns, adjusting the unit till the disturbances

disappeared has been used to judge verticality, but more stringent
tests are probably desirable.
While there is not much to be gained practically from further
refinements to the Nd correlations, in fundamental terms gas hold
up and/or bubble size should be included. Attempts to include Îg
(31,34) are made difficult by
Page 303
Figure 10
Comparison of Nd vs. dc for the various models in
Table 1a. (Data for models in inset.)
data as in Figure 6. According to Rice and Geary (39), bubble size
may be the key length parameter for predicting tall bubble column
dynamics.
2
Solids RTD
The mixing behavior of the solids is ultimately more relevant than
that of the liquid. The available solids RTD data have been
successfully fitted with the same axial dispersion model used for
liquid RTDs (37,100) (e.g., Fig. 11). References to the model not
being suitable (40,41) appear to be the result, in part at least, of
inappropriate Nd estimation routines.
Ndp as a function of particle size (Fig. 12) shows the expected
trend, the magnitude of particle mixing tends toward that of the
liquid as particle size decreases. If, rather, the dispersion coefficient
Ep is considered, the trend suggests the opposite, that mixing
increases as particle size increases and even exceeds that of the
liquid (Fig. 13). Mills et al. (100) suggest this means the particles
interact with the bubbles more than the liquid does. An
Page 304
Figure 11
Liquid and solid RTD and axial dispersion model fit
(data from Ref. 41):
(a) -75 + 38 µm solids, (b) -150 + 75 µm solids.
Figure 12
Ratio of solid to liquid vessel dispersion number vs. average
particle size.
(From Ref. 37.)
Page 305
Figure 13
Ratio of solid to liquid dispersion coefficient vs. average particle size.
(From Ref. 37.)
alternative explanation is that the analogy with Fickian diffusion,
which the model uses, is not really sufficient to describe the mixing
induced by bubbles (40). Lumping all the various phenomena into
a single coefficient E is only a rough simplification of reality (101).
From a practical standpoint the authors suggest Nd is the better
measure when referring to ''degree of mixing."
To estimate Ep requires the particle settling velocity. For particles
<100150 µm the hindered settling correlations determined in two
phase solidliquid systems appear adequate (37,42). The reason
given was equivalence between solids concentration inside and
outside the bubble wake region at these particle sizes (42).
Tsutsumi et al. (102) suggest the equivalence holds provided
dp(rsrl) < 220. For a typical density difference of 2 g/cm3, this
gives dp < 110 µm which is in good agreement. For coarser
particles one of the settling equations given by Ityokumbul (40)
may be relevant.
Table 1b gives some reported relationships between EL and Ep. For
most practical purposes (e.g., column scale-up), the assumption
that EL = Ep is adequate; the 25% under estimation of Ndp that this
may entail (100) does not lead to significant overestimation of
recovery at typical Nd values (> 1).
Deckwer and Schumpe (101) in a recent review on bubble column
reactors concluded that "in contrast to the many investigations on
liquid mixing
Page 306
its effect on performance is only minor"; the same may be said in
the case of flotation columns and be extended to include solids
mixing in most situations. They also noted that, on the other hand,
gas phase mixing has been largely ignored but that in large
diameter columns it may be important; work on this topic in
column flotation will no doubt occur.
3
Baffling
One application of the mixing correlations is to predict the effect of
baffling: the correlations suggest that reducing dc (or dc/Hc) by
vertical baffling reduces Nd. This assumes that an even distribution
of slurry, wash water and gas is achieved in each baffled section.
Failure to do so is recognized as a limitation (2) and has been
elegantly demonstrated on the laboratory scale by Moys et al. (28).
They found baffling gave no systematic decrease in mixing but in
fact tended to increase it. Unequal flows between the sections was
shown to be the cause. Maldistribution of gas, in particular, was
examined. It was shown that water flow could become reversed in
the section with excess air causing recirculation between the
chambers, a phenomenon deliberately exploited in some designs of
bubble column (15). The limited data on baffling in full-size units
available to us does suggest caution is required. One example is in
Figure 14, which shows that the liquid RTD was essentially
unchanged upon baffling. Substituting the usual static baffles with
dynamic (i.e., mobile) ones has been claimed to reduce mixing
(105).
Moys et al. (28) made two recommendations to try to realize the
benefits of baffling: (1) raise the top of the baffle into the froth to
effectively seal each chamber and prevent recirculation and (2)
raise the bottom of the baffle sufficiently above the spargers to take
advantage of any self-compensating redistribution of gas that may
occur. The penalties with the first recommendation are: the level in
each chamber is different and distribution of feed to each section
individually is required (rather than the common expedient of one
centrally located feeder). In applications requiring high recoveries
(e.g., roughing), these recommendations should be considered.
4
Height
The effect of height on mixing can also be addressed with the
correlations in Table 1. Their use, however, in predicting
metallurgy is valid only if the system is operating below its
carrying capacity. Further, the impact of mixing on metallurgy is
only significant at high recoveries (e.g., > 80%). Respecting these
constraints, Yianatos et al. (44) showed theoretically that Hc/dc» 10
was a reasonable compromise, apparently confirming the ratio in
the original designs (45). This criteria was relaxed in subsequent
modeling work (31).
Page 307
Figure 14
The liquid RTD for two parallel 1.8 × 11.6 m columns
at INCO's matte separation plant one baffled into four
sections. The residence times were tail 1, 55.3 min;
tail 2, 58.7 min.
(Raw data from Ref. 43.)
One method of testing the effect of height in an industrial-scale
column is to change the location of the spargers (Fig. 15) (46,106).
Ounpuu and Tremblay (46), found a consistent 60% recovery while
increasing Hc/dc from 2.6 to 6.2, which suggested the system
rapidly attained carrying capacity. Maksimov et al. (106) found the
recovery increased but passed through a maximum at Hc/dc» 5; the
decrease above 5 was attributed to bubbles which became too
loaded and were carried to the tailings. The froth may play a role in
these observations also.
Reaching carrying capacity, as Ounpuu and Tremblay note, is not
unexpected for a system with the high mineral floatability and feed
grades typical of a cleaning section. There is clearly no advantage
to increasing the height above a Hmax at which the bubbles become
fully loaded. A number of approaches can be taken to predict Hmax
(40,47) [including modifying the commonly used scale-up model
(2)]. The difficulty is one of estimating the appropriate
parameter(s) to quantify floatability and the fact that it and other
key variables such as mineral feed rates change in practice, i.e.,
Hmax moves up and down. It was natural, therefore, to be
conservative in the early designs, following the presumption that
overdesign would incur fewer problems than underdesign, which is
generally the case (but see Ref. 95).
Page 308
Figure 15
Recovery as a function of column collection zone height to diameter ratio.
(Data from Ref. 46 for Zn, dc = 1.2 m, and Ref. 106 for Cu, dc = 1 m, square.)
The success of short conventional columns does remove some
incentive to the development of alternative, more compact versions.
For cleaning, the preferred design may be short columns in series (to
provide the carrying capacity). Such an arrangement clearly
approaches that of a bank of cells. The obvious question may be why
not add wash water (a key feature of columns) into banks of
mechanical cells? Indeed this has been done at a number of plants
(48,108). Mechanical cells tailored to include wash water addition
are now available and they have been tested against columns (10,11).
In roughing and scavenging, where high recovery demands more
attention be paid to mixing and mineral feed grades are low, tall
columns may remain the preferred geometry.
What will occur is increasing use of columns in circuits, including
other flotation machines (49). Any variations in response between
different flotation devices, e.g., to such factors as particle size (50)
and degree of oxidation (51), could then be profitably exploited. For
example, compared to columns mechanical cells generally give
higher recoveries at equivalent selectivity with coarse particles but
lower recoveries and much poorer selectivity at fine sizes (50,96,97).
A judicious combination of both may improve overall performance
and give some flexibility in circuit control (50). (The alternative of
introducing agitation into a column to improve coarse parti-
Page 309
cle recovery (97) loses this flexibility. To date, full-scale trials of
agitated columns (98) have failed to attract much interest.)
The concept, and the most successful application of columns, is to
treat slow floating fine particles where entrainment recovery limits
separation; the farther the application strays from this the less the
metallurgical incentive will likely be to employ columns.
C
Sparging
Sparger life and ease of maintenance largely determine the choice
of device rather than such considerations as control over bubble
size or metallurgical benefits, which are not always readily evident
(14,52,53). Porous media (e.g., filter cloth) continue to be
developed (49) and used successfully (e.g., 52,106). Limited life of
porous spargers, for example due to plugging with solids, has
prompted wide use of gas jetting techniques (USBM or Cominco-
type spargers). Initial problems of hole wear with these devices
have been solved with ceramic inserts but plugging can still occur,
sometimes from particulates in the sparger water. The advantage of
on-line removal with these spargers has not always been realized
because of precipitate buildup on the sparger making its subsequent
withdrawal from the column difficult.
These systems produce bubbles inside the column. External
generation offers the possibility of reliable on-line maintenance and
for this reason is of increasing interest. The systems that create
bubbles by shear between gas and slurry (as opposed to water) are
particularly attractive (Fig. 16).
The aerator of Yoon et al. (54) is now installed in at least three
locations: Cyprus Sierrita, Thiele Kaolin (both retrofits), and ANR
Coal Co. (Roxana, Kentucky). Compared with internal sparging
systems, higher capacities at reduced gas rates and the ability to
change on-line have been demonstrated (56).
Operational advantages alone will ensure an expanding use of
external aerators. But even more compelling in the future may be
the claim that these aerators act as a "reactor for mineral particles
and air bubbles" (57). The process of bubble creation by shear in
these aerators may be conducive to simultaneous particle
collection. Understanding and exploitation of this mechanism could
usher in a new era of flotation machines. The mixing section in the
downcomer of the Jameson cell (99) may be providing a similar
environment for rapid particle collection.
D
Bubble Velocity
1
Single Bubbles
The relationship between terminal velocity and size is used in
routines for estimation of bubble diameter (58). Several
relationships have been pro-
Page 310
Figure 16
Aerators using gasslurry contacting.
(a) Cross-sectional view of Bahr et al.
(55) device. (Note that slurry is
usually feed.)
(b) Aerator design of Yoon et al. (54).

(Note that slurry is usually column contents.)
posed (e.g., see Ref. 59), some of which try to account for the
effect of contaminants (e.g., frothers in this case): For the same size
bubble different frothers and dosages change the velocity.
Zhou et al. (60,61) have developed a model that attributes
contamination to the effect of frother on the properties of the liquid
film carried by a rising bubble. Mathematically, this film is
modeled by adopting the concept of virtual liquid mass and
attributing effects to changes in the thickness of the film. A single-
fitted parameter, called a contamination factor (related to frother
type and concentration), was required. The fit can certainly be
improved with this model compared to previous, commonly used
ones (Fig. 17).
The work points out the need for a fundamental interpretation of
the role of frothers on bubble motion. Adopting a concept
originally developed
Page 311
Figure 17
Changes in bubble rise velocities with the addition
of 1.5 cm3 per 100 L Dowfroth 250 and MIBC.
[Solid lines represent fit of model of Zhou et al. (61);
the dashed line is the Schiller and Naumann (62)
equation.]
for the case of an accelerating bubble, e.g., upon formation (see Ref.
63), is questionable (although an acceleration term could be
associated with velocity changes upon bubble growth as it rises). In
practical terms the contamination effect becomes less for the usual
range of interest, db < 1.0 mm (Fig. 17). For example, in an
extension of the model to estimate bubble size in swarms (64), no
improvement over the previous models is evident (Table 2).
Predicting bubble velocity in a slurry is one objective of this basic
work. From limited data (59) the slurry may be considered as
homogeneous if db/dp > 30. For many flotation systems, therefore,
this condition may not be met. In flotation the effect of attached
particles on reducing bubble velocity also has to be considered
(58,59); treating the problem as one of increased gas density, which
has been studied in other systems (101), may be an approach.
2
Bubble Swarms
Single bubble motion is of interest here primarily as an element in
understanding the motion of bubble swarms.
By definition the swarm velocity Ubs is
Page 312
Table 2 Comparison of Bubble Size Estimation with
Measured Values.a
Bubble diameter
(mm)
Estimationc
Frotherb Jg Jl Îg Mea-
(ppm) (cm/s) (cm/s) (%) sured 1 2 3
5 1.0 0.91 9.5 1.20 1.06 1.11 1.11
10 1.0 0.85 12.9 0.86 0.88 0.87 0.87
20 1.0 0.85 15.5 0.69 0.81 0.77 0.77
10 2.1 0.30 15.7 1.51 1.34 1.40 1.40
15 0.5 1.00 12.3 0.62 0.62 0.55 0.55
15 0.8 1.00 17.0 0.67 0.68 0.64 0.64
From Ref. 58.
a Data selected to cover full range in measured db.
b Dowfroth 250.
c 1. Zhou et al. (64); 2. Yianatos et al. (58); 3. Xu and
Finch (67)a simplified version of 2.
(i.e., the average gas velocity). Nicklin (65) proposed that (in
counter-current flow)
where Ubo is the buoyancy velocity that can be determined by

switching the gas off and measuring the velocity of the bubble front
thus created. Shen et al. (66) have developed a conductivity
technique for automatically recording bubble front velocity.
Results for batch water (JL = 0, the case analyzed by Nicklin) show
that the front velocity depends on what new gas rate (Jg2) is set, or
the ratio Jg2/Jg1 (Fig. 18).
The value of Ubo (i.e., at Jg2 = 0) is the maximum front velocity.
This seems at odds with equation (6) which says, with Jg > 0, the
front velocity should be greater than Ubo. Figure 19 compares three
estimates of Ubs, two using equation (5) (one with the local Jg, the
other with Jgo) and one using equation (6) with the Ubo measured
from Figure 18. Agreement between the equations is seen only if Jgo
(atm) is used in equation (5). In view of the discussion of Figure 6,
this does not seem logical. Further, if as a measure of Ubs the front
velocity at Jg2 = Jg1 is taken (which seems reasonable on casual
inspection), then none of the estimates correspond. At the moment,
therefore, the concepts expressed in equations (5) and (6) appear to
need further investigation.
Even in the absence of a fundamental understanding, data such as in
Figure 18 may help in control. For instance, gas rate is frequently
manipulated to control metallurgy, but this also alters the level. To
stabilize the
Page 313
Figure 18
Bubble front velocity upon step change in gas rate from
Jg1 to Jg2
(unpublished work, S. Gang, McGill University, 1992).
Figure 19
Comparison of various possible swarm velocities.
Page 314
level, a feed forward control algorithm that considers the effect of
gas rate on front velocity may have advantages over the current
feed back loops.
III
Froth Zone
A
Bias and Alternative Parameters
The important role of wash water in enhancing grade continues to
be reconfirmed (52,106,109). The control of wash water rate in the
absence of direct grade measurement, however, remains a
challenge.
Common practice is to control either to a bias or a displacement
wash ratio (Table 3). There are drawbacks with both parameters
[which are related (2)]: they require water mass flow
measurements, and they are not necessarily directly related to
hydrophilic fine particle recovery, which ultimately is what wash
water is intended to suppress. Much evidence points to an optimum
bias of about zero which severely tests any control system.
Inferring bias from other measurements has been suggested (68)
but would seem to introduce even more uncertainty (69). Perhaps,
therefore, bias should be substituted by another parameter: some
possible ones are given in Table 3.
The most attractive metallurgically is feed water recovery as it is
related
Table 3 Bias and Alternative Measures.
Parameter Advantage(s) Disadvantage(s)
Commonly used
Bias (difference or Requires mass
ratio between water flows;
in tails and feed) not directly
related to grade
Displacement wash Reflects action of Requires mass
ratio (ratio of wash wash water (to flows
water to water in replace slurry water
overflow) in froth)
Feed water recovery Directly related to Requires mass
fine hydrophilic flows and a
particle recovery tracer
(and hence to grade)
Fraction of feed Does not require Not directly
water in overflow mass flows related to grade;
(concentrate) water requires a tracer
(or entrainment)
Profiles (e.g., Do not require mass Interpretations
conductivity, flows incomplete
temperature,
pressure)
Adapted from Ref. 71.
Page 315
to hydrophilic particle recovery (95) (although the form of the
relationship remains an area of research (e.g., Ref. 70). Predicting
water recovery from empirical correlations with bias and feed rate
may be feasible (14,95) but this does not avoid the uncertainty
associated with the bias measurement and a general correlation
seems unlikely (71). Consequently direct measurement of feed water
recovery seems preferable. Measurement on-line requires a
detectable difference in some quality of the various input waters
(referred to as a ''tracer" in Table 3), e.g., temperature (95), and still
requires mass flow measurement.
More attractive practically is "entrainment" as it requires no flow
measurements although still requires a tracer. For example, from the
temperature of the overflow To, wash water Tw and feed Tf
entrainment a is given by
(this assumes no heat exchange with the environment, in particular

with the gas phase).
Entrainment is being explored. Figure 20 shows a general
relationship with the traditional bias difference (or rate) parameter:
the scatter suggests experimental uncertainty (much of which may be
in the bias) and probably an effect of secondary variables, for
instance, froth height. Moys and Finch (72), in one example, showed
grade and entrainment were related (Fig. 21).
Figure 20
Feed water entrainment vs. bias rate.
(Data from Refs. 35 and 71, and Espinosa-
Gomez (private communication, 1992).
Page 316
Last, profiles potentially carry the most information, but
interpretation is far from complete (7174). Pressure profiles are
closest to implementation; Figure 22 shows their use in detecting
loss of interface in an operating column (75).
Profiling is a valuable diagnostic tool. Uniform wash water
distribution is important in increasing recovery (76); wash water
distribution can be readily checked by temperature profiling. If
there is uncertainty over froth depth, profiling may resolve the
issue (77,78).
B
Carrying Capacity
That a maximum in the overflow solids rate exists and may be
capacity limiting in columns (or any flotation device) is well
understood, although its origin (79) and how best to determine it
(80) are still uncertain. Measures that have been used include: mass
per unit volume of gas (Cg), rate per unit of lip length (lip length
capacity, CL, or rate per unit area (Ca). The latter measure is often
called carrying capacity (although it is better termed area capacity).
CL and Ca are related through geometry, and both in turn are
related to Cg through the gas rate, Jg (2); Cg is usually a stronger
function of gas rate than either of the other two and is, therefore,
less useful for developing correlations.
Figure 21
Correlation between grade of Pb (the desired component)
in the concentrate and feed water entrainment at
Cominco's Polaris concentrator.
(From Ref. 72.)
Page 317
Figure 22
Example of pressure profiling; two pressure sensors in
collection zone to measure bulk density and one sensor in
froth to measure bulk density. Note that a change in operating
conditions (e.g., increase in gas rate) has caused the two
zone densities to approach one another; when they become equal
this means interface is lost.
(See also Ref. 69.)
For preliminary predictive purposes, an empirical model to predict
Ca was developed (13)
Carrying capacity is observed to decrease with dc above » 1 m

(76,79); Amelunxen (79) proposed the following to allow for this
where the constants k1 and n depended on the mineral being treated.

Since CL and Ca are related by
where k2 depends on cell geometry (e.g., k2 = p/4 if column is
circular with no internal launders), then a suggested modification of
equation (8) for dc > 1 m is
with dc in meters, and n ranges from 0.2 to 0.7 (79). A similar

correlation probably applies to Jameson cells also (109).
Table 4a reviews some recent carrying capacity estimations: the
predictions [equation (11) with n = 0.3] are reasonable. Table 4b
introduces a note of caution, however. Although these data are not
specified as carrying
Page 318
Table 4 Carrying Capacities.
a. Carrying Capacitya Estimates and Predicted Values.
Carrying capacity
(g/min cm2)
Predicted
Duty d80 (µm) dc (m) Measured Ca Ca* Source
Cu Clnr 38 1.2 7.2 7.6 6.7 (46)
Zn Clnr 28 1.2 5.6 5.9 5.2 (46)
Zn Clnr 25 0.05 4.4 5.0 - (80)
Zn Clnrb 17 2.4 1.35 3.4 1.8 (52)
Cu Clnrc 26 0.4 5.3 5.2 - (109)
b. Carrying Ratesa Compared to Predicted.
Carrying rate
(g/min cm2)
Predicted
Duty d80 (µm) dc (m) Measured Ca Ca* Source
Cu Clnr 38 2.13 5.3 8.2 4.8 (9)
Zn Clnr 62 1.68 4.5 12.7 8.8 (52)
(Boliden)
Zn Clnr 109 0.91 5.3 22.4 - (94)
Source: Data from Ref. 13.
a Average given if range quoted.
b Apirsa concentrator, one (of five) referred to
specifically at carrying capacity (see Table 4b).
c Jameson cell; Ca at bias rate = 0 taken.
capacity, there is reason to believe they are given the duty. Clearly
the measured data fall well below the predicted in some cases. One
reason may be the coarser d80 compared with those in Table 4a;
perhaps, the original constraint of d80 < 44 µm (2) is still relevant
for sulfides.
From operations, the following observations emerge: is of the
order to expect; is rarely exceeded significantly; and carrying
capacity is rarely greater than 8 g/min cm2.
Since failure to meet the anticipated carrying capacity can have
severe repercussions (e.g., see Ref. 52) optimistic predictions are
best avoided. The growing trend to internal launders is in response
to the limits imposed by overflow removal rates.
Page 3
C
Cleaning Action
From water-tracer studies Yianatos et al. (81) inferred that a significant rejection of fine hydrophilic
particles (or cleaning) took place at the collection/froth zone interface. Yianatos and Bergh (41)
subsequently validated this using radioactively tagged hydrophilic particles (Fig. 23). From direct
measurement of solids concentration in the interstitial slurry of the froth, Falatsu and Dobby (87) sugge
entrained particles are rejected in the first 30 cm.
Does this mean an interface is essential? In small diameter columns (say, dc < 15 cm), upgrading can
occur between the feed port and the interface (82). This, in fact, has led to periodic reference to a third,
washing, zone. In large diameter columns this upgrading is not apparent (e.g., Ref. 46), because the
higher degree of mixing renders the washing zone part of the collection zone (2,33). Nevertheless, it
remains an interesting question as to how much upgrading could be achieved with wash water in the
absence of a froth. The successful cleaning achieved by adding wash water below the froth (83) is also
pertinent to this point.
Figure 23
Evidence of (a) rejection of hydrophilic (gangue) particles (or cleaning) at the interface from radioactive tracer study
Yianatos and Bergh (41), and (b) comparison with earlier water-tracer study (81).
Page 320
In another approach to the question, Schultz et al. (84) found
metallurgy improved by feeding directly into the froth (Fig. 24).
Feeding into the froth, in effect, may also be what occurs in the
packed column where an interface does not seem to exist (85).
While firm conclusions cannot be reached, it can be stated that an
interface contributes significantly to cleaning but that it may not be a
necessary condition.
D
Recovery and Selectivity
The factors likely to affect froth zone recovery Rf can be introduced
using a simple plug flow model
where Hf/Jg is equivalent to bubble retention time in the froth. The

complement (1Rf) is sometimes called dropback (2).
1
Detachment Rate Constant
If there are differences in kd (i.e., there is selectivity) between
minerals of different hydrophobicity, then a change in grade with
froth depth (i.e., grade profiles) should exist along the froth.
Figure 24
Effect on metallurgy (oil grade and recovery for an oil shale) of
location of feed relative to interface. Note the improved performance
when feeding into froth in tests 3 and 4.
Page 321
The available profiles, however, seem largely flat (2,35,46,86,87). There
could be some problem in interpreting the profiles because of the
presence of entrained particles; Falatsu and Dobby (87), however,
isolated the solids attached to the bubbles but again showed flat profiles
(Fig. 25). Selectivity does appear to exist in conventional (draining)
froths (88).
Based on these observations, it is usual to adopt a single value of Rf for
all minerals for scale-up calculations; laboratory work tended to support
this observation suggesting that froth height was not a factor and
attributing Rf to the impact of the interface (91). More recent work on
plant columns, however, does not support this, showing that bubbles
shed particles as they rise through the froth (87).
2
Froth Depth and Gas Rate
Froth depth can be shown to influence metallurgy (89), but it is not
generally incorporated in control strategies (e.g., Ref. 90). On the other
hand, gas rate is perhaps the major operating variable in column
performance. As gas rate increases, the general trend is for grade to
decrease and recovery to pass through a maximum. The trend in grade
follows from the effect of gas rate on entrainment, and the trend in
recovery can be predicted from a
Figure 25
Grade of Zn for particles on the bubbles and in the slurry between the bubbles as a
function of height in the froth. Data taken on the 2.4-
m diameter column at Cominco's
Sullivan concentrator by Falatsu and Dobby (87).
Page 322
fundamental model of the collection zone; however, the froth zone
may also be a factor in this trend in recovery.
Wilson and Stratton-Crawley (92) show froth zone recovery
increased with gas rate (Fig. 26) and this appeared to dictate gas
requirements. This may be related to the need to force heavy froths
over the lip. Amelunxen (93), for example, experimented with
additional air just below the froth to try to overcome "a lack of
carry-over force required to transport concentrate over the lip."
Some versions of the Flotaire column include facility to add air
part-way up the column (103).
The role of gas rate on froth zone behavior is difficult to isolate
from its overall effect on column performance. Nevertheless its
possible impact on Rf should be considered, particularly under
loaded froth conditions. Gas rate will severely degrade metallurgy
if raised above a critical value, for example that which causes loss
of interface (69) (see also Fig. 22).
IV
Conclusions
The last 10 years have seen significant progress in the application
of flotation columns worldwide. While this may be considered a
triumph of "know how" rather than "know why" (104), the authors
believe real progress in the latter was achieved, which has formed
part of this review. Indeed, the ability to interpret the
hydrodynamic and metallurgical response of flota-
Figure 26
Froth zone recovery as a function of gas rate in 0.038 test column: Rf
estimated using a model (92).
Page 323
tion columns to design and operational variables was probably key
to convincing industry to test the concept outside the laboratory.
The flood of research articles over the last 2 years shows a topic
reaching maturity. Column flotation is now a subset of the field of
bubble columns (15).
What of the future? Inevitably research to understand fundamental
phenomenabubble velocity, bubbleparticle collision, etc.will
intensify. The industry, however, is already showing signs of
moving onto new devices. Many of these owe their existence to the
conventional column either as direct derivatives or simply because
a climate for change was created by the success of columns. In
accord with some other current workers (107), the authors see
future machines engineered to give intense particle collection in
one part (the reactor) followed by a quiescent water-washed section
in which bubble slurry separation occurs (the separator).
Independent optimization of the two sections should be possible:
research into this will be rewarding to both researchers and
industry.
Acknowledgments
Column flotation research at McGill is sponsored by a consortium
(Falconbridge, INCO, Teck, Cominco, CANMET) through MITEC
and the Natural Sciences and Engineering Research Council of
Canada through funding under the Research Partnerships program.
This support is gratefully acknowledged.
References
1. K. V. S. Sastry, Ed., Column '88, Proceedings of an International
Conference on Column Flotation, Phoenix, Ariz. Jan. 2528 (1988).
2. J. A. Finch and G. S. Dobby, Column Flotation, Pergamon,
Oxford. (1990).
3. G. A. Agar, B. J. Huls, and D. B. Hyma, Eds., Column '91,
Proceedings of an International Conference on Column Flotation,
Sudbury, Ontario, Canada, June 26 (1991).
4. TUNRA/AMIRA, Column Flotation Workshop, Proceedings,
Adelaide, Australia, July 2728, 1989.
5. J. A. Finch and G. S. Dobby, Column Flotation, Camborne
School of Mines, Camborne, U.K., July, 2122, 1989.
6. J. A. Finch and G. S. Dobby, Column Flotation, Singapore, Feb.
1617, 1991.
7. Y. B. Rubenstein, V. I. Meli-Gaikazyan, and S. B. Leonev, Froth
Separation and Column Flotation, Nedra, Moscow, 1989.
8. K. Norris, "Process Improvements at Placer Dome Campbell
Mine," Proceedings of 24th Annual Meeting of the Canadian
Mineral Processors, Ottawa, Canada, Jan. 2123, paper 22 (1992).
Page 324
9. M. A. Mular and C. Veloo, "Circuit Modifications at Westmin
Resources Myra Falls Operations," Proceedings of 24th Annual
Meeting of the Canadian Mineral Processors, Ottawa, Canada, Jan.
2123, paper 26 (1992).
10. W. W. Ulan, D. Green, and G. A. Kosick, "In-Plant Testing of
the Outokumpu High Grade Flotation Cell," Column, '91,
Proceedings of an International Conference on Column Flotation
(G. A. Agar, B. J. Huls, and D. B. Hyma, eds.), Sudbury, Ontario,
Canada, pp. 689702 (1991).
11. I. Farr, Evaluation of alternative flotation technologies for Cu
cleaning & Cu/Ni separation at INCO's Thompson Mill, Miner.
Eng., 5(10/12): 1169. (1992).
12. G. Jameson and E. V. Manlapig, "Applications of the Jameson
Flotation Cell," Column '91, Proceedings of an International
Conference on Column Flotation (G. A. Agar, B. J. Huls, and D. B.
Hyma, eds.), Sudbury, Ontario, Canada, June 26, pp. 673688
(1991).
13. J. A. Finch and G. S. Dobby, Column flotation: A selected
review. Part 1, Int. J. Miner. Process., 33: 343. (1991).
14. G. S. Dobby and J. A. Finch, Column flotation: A selected
review. Part 2, Miner. Eng., 4(7/11): 911. (1991).
15. L.-S. Fan, Gas-Liquid-Solid Engineering, London,
Butterworths. (1989).
16. A. Uribe-Salas, C. O. Gomez, and J. A. Finch, Gas holdup
measurement in flotation columns using electrical conductivity,
Can. Metall. Q., 30(4): 201 (1991).
17. J. C. Maxwell, A Treatise of Electricity and Magnetism, 3rd ed.,
Vol. 1, Part 11. (1892).
18. M. M. Marchese, A. Uribe-Salas, and J. A. Finch,
Measurement of gas holdup in a three-phase concurrent downflow
column, Chem. Eng. Sci., 47(13/14): 3475. (1992).
19. A. Summers, M. Xu, and J. A. Finch, "Jameson Cell testwork,"
Report to Falconbridge Ltd. (1992).
"Hydrodynamics of a Downflow Column," Proceedings of XVIII
International Mineral Processing Conference, Sydney, Australia,
May 2328 (1993, in press).
21. J. B. Yianatos, J. A. Finch, and A. R. Laplante, Holdup profile
and bubble size distribution of column flotation froths, Can.
Metall. Q., 25(1): 23. (1986).
22. A. Uribe-Salas, C. O. Gomez, and J. A. Finch, "Axial Gas
Holdup Profiles in Flotation Columns," Proceedings of 42nd
Canadian Chemical Engineering Conference, Toronto, Ontario,
Canada, Oct. 1821 (1992, in press).
23. Z. A. Zhou and N. O. Egiebor, Prediction of axial gas holdup
profiles in flotation columns, Miner. Eng., submitted.
24. W. Deckwer, U. Graeser, H. Langemann, and Y. Serpemen,
Zones of different mixing in the liquid phase of bubble columns,
Chem. Eng. Sci., 28: 1223. (1973).
25. M. Xu and J. A. Finch, Measurement of radial gas holdup
profiles in flotation columns, Int. J. Miner. Process., in press.
Page 325
26. A. Serizawa, I. Kataoka, and I. Michyoshi, Turbulence
structure of air-water bubbly flow, Int. J. Multiphase Flow, 2: 221.
(1975).
27. W. Freedman and J. B. Davidson, Holdup and liquid circulation
in bubble columns, Trans. Inst. Chem. Eng., 47: 251. (1969).
28. M. H. Moys, J. A. Engelbrecht, and N. Terblanche, ''Design of
Baffles to Reduce Axial Mixing in Flotation Columns," Column
'91, Proceedings of an International Conference on Column
Flotation (G. A. Agar, B. J. Huls, and D. B. Hyma, eds.), Sudbury,
Ontario, Canada, June 26, pp. 275288 (1991).
29. M. Xu and J. A. Finch, Axial dispersion model in flotation
column studies, Miner. Eng., 4(5/6): 553, (1991).
30. P. J. T. Mills and C. T. O'Connor, Use of axial dispersion model
to describe mixing in a flotation column, Miner. Eng., 5(8): 939.
(1992).
31. M. Xu and J. A. Finch, "Estimating Vessel Dispersion Number
in Flotation Columns," Column '91, Proceedings of an International
Hyma, eds.), Sudbury, Ontario, Canada, June 26, pp. 437454
(1991).
32. P. Mavros, Validity and limitations of the closed-vessel
analytical solution to the axial dispersion model, Miner. Eng., 5(9):
1053. (1992).
33. A. R. Laplante, J. B. Yianatos, and J. A. Finch, "Mixing
Characteristics of the Collection Zone in Flotation Columns,"
Column '88, Proceedings of an International Conference on
Column Flotation (K. V. S. Sastry, ed.), Phoenix, Ariz., Jan. 2528,
pp. 6980 (1988).
34. M. J. Mankosa, G. H. Luttrell, G. T. Adel, and R. H. Yoon, A
study of axial mixing in column flotation, Int. J. Miner. Process.,
35(1/2): 51. (1992).
35. R. A. Alford, "Modelling and design of flotation column
circuits," Ph.D. Thesis, Univ. of Queensland, Australia, 1991.
36. R. G. Rice, A. D. Oliver, J. P. Newman, and R. J. Wiles,
Reduced dispersion using baffles in column flotation, Powder
Technol., 10: 201. (1974).
37. M. Xu and J. A. Finch, Solids mixing in the collection zone of
flotation columns, Miner. Eng., 5(9): 1029. (1992).
38. N. Hidaka, M. Onitani, T. Matsumoto, and S. Morooka, "Axial
Mixing and Segregation of Multicomponent Coarse Particles
Fluidized by Concurrent GasLiquid Flow," Proceedings of the First
International Conference on Gas Liquid and GasLiquidSolid
Reactor Engineering, Columbus, Ohio, Sept. 1316 (1992, in press).
39. R. G. Rice and N. W. Geary, Prediction of liquid circulation in
viscous bubble columns, AIChE. J., 36: 1339. (Sept. 1990).
40. M. T. Ityokumbul, A new modelling approach to flotation
column design, Miner. Eng., 5(6): 685. (1992).
41. J. B. Yianatos and L. G. Bergh, "RTD Studies in an Industrial
Flotation Column: Use of Radioactive Tracer Technique," Column
Flotation, Sudbury, Ontario, Canada, June 26, pp. 221234 (1991).
42. J. B. Yianatos, J. A. Finch, and A. R. Laplante, Apparent
hindered settling in gasliquidsolid countercurrent column, Int. J.
Miner. Process., 18: 155. (1986).
Page 326
43. J. Miszczak, private communication, 1987.
44. J. B. Yianatos, R. Espinosa-Gomez, J. A. Finch, A. R. Laplante,
and G. S. Dobby, Effect of height on column flotation performance,
Miner. Metall. Process., 4(1): 11. (1988).
45. D. A. Wheeler, "Column FlotationThe Original Column," Froth
Flotation, Proceedings of 2nd Latin American Congress on Froth
Flotation (S. H. Castro and V. Alvarez, eds.), Chile, Aug. 1923, pp.
1740 (1985).
46. M. Ounpuu and R. Tremblay, "Investigation into the Effect of
Column Height on the 1200mm Diameter Column at Matagami,"
Column '91, Proceedings of an International Conference on
Column Flotation (G. A. Agar, B. J. Huls, and D. B. Hyma, eds.),
Sudbury, Ontario, Canada, June 26, pp. 303316 (1991).
47. Z. Zhou and N. O. Egiebor, Particle-bubble collision in
flotation columns, Can. Metall. Q., submitted.
48. M. Kaya, A. R. Laplante, C. Tan, B. J. Huls, A. Putz, P.
Lacombe, and R. Kemp, "Plant Applications of Froth Washing in
Mechanical Flotation Cells," Proceedings of 22nd Annual Meeting
of the Canadian Mineral Processors, Ottawa, Canada, Jan. 1618,
pp. 160198 (1990).
49. G. S. Dobby, Column Flotation Seminar, McGill University,
May 5, 1992.
50. B. J. Huls, D. B. Hyma, G. S. Dobby, and C. D. Lachance,
"Scale-Up and Installation of Flotation Columns at Falconbridge
Ltd.," Column '91, Proceedings of an International Conference on
Column Flotation (G. A. Agar, B. J. Huls, and D. B. Hyma, eds.),
Sudbury, Ontario, Canada, June 26, pp. 199211 (1991).
51. R. Lahoapanit, A. R. Laplante, and A. Boulet, "Column
Flotation Testwork at Les Mines Camchib, Inc., Chibougamau,
Quebec," Column '91, Proceedings of an International Conference
on Column Flotation (G. A. Agar, B. J. Huls, and D. B. Hyma,
eds.), Sudbury, Ontario, Canada, June 26, pp. 181198 (1991).
52. N. J. Brolin and L. Soderberg, "Experience from Installations
of Column Flotation in Sweden," Column '91, Proceedings of an
International Conference on Column Flotation (G. A. Agar, B. J.
Huls, and D. B. Hyma, eds.), Sudbury, Ontario, Canada, June 26,
pp. 573584 (1991).
53. A. J. Burger, "Performance of a Flotation Column in Recovery
of Gold and Pyrite from Backfill Material," Column '91,
Canada, June 26, pp. 595608 (1991).
54. R. H. Yoon, G. H. Luttrell, G. T. Adel, and M. J. Mankosa,
Application of MicrocelTM column flotation to fine coal cleaning,
Coal Processing (B. K. Parekh, ed.), in press.
55. A. Bahr, "Application and Sizing of a New Pneumatic Flotation
Cell," Proceedings of XVth International Mineral Processing
Congress, Cannes, France, June 29, pp. 314326 (1985).
56. M. J. Mankosa, private communication, 1992.
57. L. Changgen and A. Bahr, Flotation of copper ore in a
pneumatic flotation cell, Miner. Metall. Process., 9(1): 7 (1992).
Page 327
58. J. B. Yianatos, J. A. Finch, G. S. Dobby, and M. Xu, Bubble
size estimation in a bubble swarm, J. Colloid Interface Sci., 126(1):
37 (1988).
59. L. S. Fan and K. Tsuchiya, Bubble Wake Dynamics in Liquids
and LiquidSolid Suspensions, Butterworth, Boston, p. 48. (1990).
60. Z. Zhou, L. R. Plitt, and N. O. Egiebor, "On Bubble Size
Determination in a Flotation Column," Column '91, Proceedings of
an International Conference on Column Flotation (G. A. Agar, B. J.
pp. 249262 (1991).
61. Z. A. Zhou, N. O. Egiebor, and L. R. Plitt, Frother effects on
single bubble motion in a water column, Can. Metall. Q., 31(1): 11.
(1992).
62. L. Schiller and A. Naumann, Z. Vgl. Ing., 77: 318. (1933).
63. N. W. Geary and R. G. Rice, Bubble size prediction for rigid
and flexible spargers, AIChE. J., 37:161. (Feb. 1991).
64. Z. A. Zhou, N. O. Egiebor, and L. R. Plitt, Frother effects on
bubble size estimation in a flotation column, Miner. Eng., 6(1):
5567. (1993).
65. J. D. Nicklin, Two-phase bubble flow, Chem. Eng. Sci., 17:
693. (1962).
66. G. Shen, A. Uribe-Salas, N. Ahmed, and J. A. Finch,
"Measurement of Dynamic Gas Holdup in a Bubble Column:
Computer Applications in the Mineral Industry," Proceedings of
Second Canadian Conference (R. Poulin, R. C. T. Pakalnis, and A.
L. Mular, eds.) Vancouver, Canada, Sept. 1518, pp. 287298 (1991).
67. M. Xu and J. A. Finch, Simplification of bubble size estimation
in a bubble swarm, J. Colloid Interface Sci., 140(1): 298 (1990).
68. R. Pal and J. H. Masliyah, Flow characteristics of a flotation
column, Can. Metall. Q., 29(2): 97. (1990).
69. M. Xu, J. A. Finch, and A. Uribe-Salas, Maximum gas and
bubble surface rates in flotation columns, Int. J. Miner. Process.,
32: 233. (1991).
70. V. M. Kirjavainen, Mathematical model for the entrainment of
hydrophilic particles in froth flotation, Int. J. Miner. Process.,
35(1/2): 1. (1992).
71. A. Uribe-Salas, C. O. Gomez, and J. A. Finch, "Bias Detection
in Flotation Columns," Column '91, Proceedings of an International
Hyma, eds.), Sudbury, Ontario, Canada, June 26, pp. 391408.
(1991).
72. M. H. Moys and J. A. Finch, Use of temperature measurements
in the analysis and control of flotation columns, Can. Metall. Q.,
30(3): 131. (1991).
73. M. J. Fairweather and J. R. Egan, "Froth Phase Mixing by
Multi-Point Temperature Analysis in an Industrial Flotation
Column," Column '91, Proceedings of an International Conference
74. L. G. Bergh and J. B. Yianatos, "Advances in Flotation Column
Dynamics and Measurements," Column '91, Proceedings of an
International Conference on Column Flotation (G. A. Agar, B. J.
pp. 409422 (1991).
75. G. A. Kosick, G. S. Dobby, and P. D. Young, "ColumnEX: A
Powerful and Affordable Control System for Column Flotation,"
Column '91, Proceedings
Page 328
of an International Conference on Column Flotation (G. A.
Agar, B. J. Huls, and D. B. Hyma, eds.), Sudbury, Ontario,
Canada, June 26, pp. 359374 (1991).
76. R. Espinosa-Gomez and N. W. Johnson, "Technical
Experiences with Conventional Columns at Mount Isa Mines
Limited," Column '91, Proceedings of an International Conference
77. C. O. Gomez, A. Uribe-Salas, J. A. Finch, and B. J. Huls, A
calibration probe for level detection in flotations, C.I.M. Bull.,
83(935): 118 (1990).
78. A. Uribe-Salas, M. Leroux, C. O. Gomez, and J. A. Finch, "A
Conductivity Technique for Level Detection in Flotation Cells,"
Copper 91, Proceedings of the Copper 91-Cobre 91 International
Symposium: Mineral Processing and Process Control, Vol 11 (G. S.
Dobby, S. A. Argyropoulos, and S. R. Rao, eds.), Ottawa, Canada,
Aug. 1821, pp. 261276 (1991).
79. R. Ameluxen, "Lip Loading Considerations in Flotations
Columns," Column '91, Proceedings of an International Conference
80. R. del Villar, H. Soto, and P. Lacombe, "Column Flotation
Circuit Design at Les Mines Selbaie," Column '91, Proceedings of
an International Conference on Column Flotation (G. A. Agar, B. J.
pp. 149164 (1991).
81. J. B. Yianatos, J. A. Finch, and A. R. Laplante, Cleaning action
in column flotation froths, Trans. Inst. Min. Metall. (Sec. C), 96:
199 (1987).
82. R. A. Ynchausti, J. D. McKay, and D. G. Foot, Jr., "Column
Flotation ParametersTheir Effects," Column '88, Proceedings of an
International Conference on Column Flotation (K. V. S. Sastry,
ed.), Phoenix, AZ, Jan, 2528, pp. 157172 (1988).
83. M. Kaya and A. R. Laplante, Froth washing in mechanical
flotation cells, Proceedings 21st Annual Meeting of the Canadian
Mineral Processors, Ottawa, Canada, Jan. 1719, pp. 497540 (1989).
84. C. W. Schultz, R. K. Mehta, and J. B. Bates, The flotation
column as a froth separator, Min. Eng., December, pp. 14491450
(1991).
85. D. C. Yang, "Technical Advantages of Packed Flotation
Columns," Column '91, Proceedings of an International Conference
86. R. L. Amelunxen, R. Llerena, P. Dunstan, and B. Huls, Column
Flotation (K. V. S. Sastry, ed.), Phoenix, AZ., Jan. 2528, pp.
149156 (1988).
87. M. Falatsu and G. S. Dobby, Froth performance in commercial
sized flotation columns, Miner. Eng., 5(10/12): 1207. (1992).
88. G. Tarjan, Mineral Processing 2, Akademiai Kiado, Budapest,
p. 262. (1986).
89. B. J. Huls, C. D. Lachance, and G. S. Dobby, "Gas Rate and
Froth Depth Effects on Performance of a CuNi Separation Column:
Processing of Com-
Page 329
plex Ores," Proceedings of an International Symposium (G. S.
Dobby, and S. R. Rao, eds.), Halifax, Nova Scotia, Canada, Aug.
2024, Pergamon Press, New York, pp. 311324 (1989).
90. R. Espinosa-Gomez, N. W. Johnson, J. D. Pease, and P. D.
Munroe, "Commissioning of the First Flotations at Mount Isa
Mines Limited: Processing of Complex Ores," Proceedings of an
International Symposium (G. S. Dobby, and S. R. Rao, eds.),
Halifax, Nova Scotia, Canada, Aug. 2024, Pergamon Press, New
York, pp. 311324 (1989).
91. M. Falatsu and G. S. Dobby, "Direct Measurement of Froth
Zone Performance in a Laboratory Flotation Column: Processing of
Complex Ores," Proceedings of an International Symposium (G. S.
Dobby, and R. S. Rao, eds.), Halifax, Nova Scotia, Canada, Aug.
2024, Pergamon Press, New York, pp. 335348 (1989).
92. S. W. Wilson and R. Stratton-Crawley, "Design of Production
Columns Using a First Order Kinetic Model," Column '91,
Canada, June 26, pp. 165180 (1991).
93. R. L. Ameluxen, "Mechanics of Operation of Column Flotation
Machines," Proceedings of 17th Annual Meeting of the Canadian
Mineral Processors, Ottawa, Canada, Jan. 2224, pp. 1331 (1985).
94. C. Webber, B. Bharadwaj, and K. S. Moon, "Zinc Cleaning
Using Column Flotation at Ruttan Operation HBM&S," Column
Flotation (G. A. Agar, B. J. Huls, and D. B. Hyma, eds.), Sudbury,
Ontario, Canada, June 26, pp. 4562 (1991).
95. A. Maachar and G. S. Dobby, Measurement of feed water
recovery and entrainment solids recovery in flotation columns,
Can. Metall. Q., 31(3): 167. (1992).
96. W. Petruk and R. Lastra, "Evaluating the Recovery of Liberated
and Unliberated Chalcopyrite by Flotation Columns in a Copper
Cleaner Circuit," presented at 31st Ann. Conf. of Metallurgists,
The Metallurgical Soc. of CIM, Edmonton, Canada, Aug. 2327
(1992).
97. M. C. Harris, J-P. Franzidas, C. T. O'Connor, and P. Stonestreet,
An evaluation of the role of particle size in the flotation of coal
using different cell technologies, Miner. Eng., 5(10/12): 1225.
(1992).
98. J. C. Schneider and G. Van Weert, "Design and Operation of
the Hydrochem Flotation Column," Column '88, Proceedings of an
International Conference on Column Flotation (K. V. S. Sastry,
ed.), Phoenix, AZ., Jan. 2528, pp. 287292 (1988).
99. G. M. Evans, "A study of a plunging jet bubble column," Ph.D.
Thesis, University of Newcastle, Australia, 1990.
100. P. T. Mills, J. B. Yianatos, and C. T. O'Connor, The mixing
characteristics of solid and liquid phases in a flotation column.
Miner. Eng., 5(10/12): 1195. (1992).
101. W-D. Deckwer and A. Schumpe, "Improved Tools for Bubble
Column Reactor Design and Scale-Up," keynote paper, The First
Int. Conf. on GasLiquid and GasLiquidSolid Reactor Eng.,
Columbus, Sept. 1316 (1992).
Page 330
102. A. Tsutsumi, T. Charinpanitkul, and K. Yoshida, Prediction of
solid wake concentration profiles in three-phase fluidized beds,
Chem. Eng. Sci., 47(13/14): 3411. (1992).
103. D. E. Zipperian and U. Svensson, "Plant Practice of the
FLOTAIRE Column Flotation Machine for Metallic, Nonmetallic
and Coal Flotation," Column '88, Proceedings of an International
Conference on Column Flotation (K. V. S. Sastry, ed.), Phoenix,
AZ, Jan. 2528, pp. 4354 (1988).
104. M. T. Ityokumbul, A mass transfer approach to flotation
column design, Chem. Eng. Sci. 47(13/14): 3605. (1992).
105. S. B. Leonov, G. B. Rubinstein, and S. B. Polonsky, "The
Designing, Servicing and Modelling of Flotation Columns," Proc.
XVII Int. Min. Proc. Congress, Dresden, FRG, Sept. 2328, Vol. III,
pp. 313323 (1991).
106. I. I. Maksimov, A. D. Borkin, and M. F. Emelyanov, "The Use
of Column Flotation Machines for Cleaning Operations in
Concentrating Non-Ferrous Ores," Proc. XVII Int. Miner. Process.
Congress, Dresden, FRG., Sept. 2328, Vol. II, pp. 273281 (1991).
107. C. E. Jordan and F. J. Susko, Rapid flotation using a modified
bubble-injected hydrocyclone and a shallow-depth froth separator
for improved flotation kinetics, Miner. Eng., 5(10/12): 1225.
(1992).
108. M. Kaya and A. R. Laplante, "Froth Washing Technology in
Mechanical Flotation Machines," Proc. Int. Miner. Process.
Congress, Dresden, FRG, Sept. 2328, Vol. II, pp. 203216 (1991).
109. B. W. Atkinson, P. T. Griffin, G. J. Jameson, and R. Espinosa-
Gomez, "Jameson Cell Testwork on Copper Streams in the Copper
Concentrator of Mount Isa Mines Ltd.," Proc. XVIII Int. Miner.
Process. Cong., Sydney, Australia, May 2328 (1993).
110. R. G. Rice, D. T. Barbe, and N. W. Geary, Correlation of
nonverticality and entrance effects in bubble columns, AIChE J,
36(9): 1421. (1990).
Page 331
12
The Jameson Cell
Geoffrey M. Evans, Bruce W. Atkinson, and Graeme J. Jameson
I
Introduction
The Jameson cell is a type of flotation column in which the air and
the pulp are brought together in a vertical tube. The air and pulp are
dispersed into a dense foam of fine bubbles, which creates a
favorable environment for particlebubble collision and subsequent
collection of hydrophobic particles (1). The bubbly mixture then
discharges into the cell proper, which is essentially a
disengagement chamber in which the bubbles carrying the floatable
particles separate from the pulp. The disengagement zone behaves
in a similar fashion to the froth zone in a conventional flotation
column, and as in the column, it is possible to apply clean
washwater to the froth to remove unwanted gangue particles.
The main advantages of the Jameson cell relate to the rapid
collection of particles in the downcomer, leading to a compact
space-efficient device, and to the ability of the cell to operate with
self-aspiration, thus obviating the need for compressors or blowers.
Since its introduction at the commercial scale in 1989, the cell has
been applied to a variety of ores, and full-sized cells are now
operating on streams involving lead, zinc, copper, and nickel
sulfides, as well as coal and industrial minerals. It has been applied
as well to the removal of fine organic droplets from solvent
extraction liquors.
In this chapter, a brief overview of antecedents to the Jameson cell
is given. The general principles of operation are discussed, together
with a
Page 332
description of the various phenomena that take place in the
downcomer. The effect of the operational variables on cell
performance follows, including some data from operating cells.
II
Development
The cell was devised in the period 19851986, when the inventor
was undertaking a collaborative research project at Mt. Isa Mines
Limited, Mt. Isa, Queensland. This is a large mine, accustomed to
using the best modern practices in the production of copper, lead,
and zinc concentrates. A number of large flotation columns were
being designed and introduced at the time in the copper and
lead/zinc circuits. Although it was clear that the columns would
offer significant advantages over the conventional mechanical cells
in terms of the ability to wash the froth, it was also clear that the
collection process was very slow, necessitating large residence
times and consequently the large columns, which have become a
familiar feature of this technology. Accordingly, alternative
methods of bringing the particles and the air into contact were
investigated. Following initial work in the laboratory at Newcastle
University, prototypes were tested and further developed in the
concentrator at Mt. Isa, resulting in a production-scale design. A
number of Jameson cells were installed in the lead circuit and in
the nearby Hilton concentrator then under construction. Further
cells were installed in a coal operation at Newlands, Queensland,
and on a copper stream at the Peko Warrego mine, Tennant Creek,
Northern Territory, Australia.
The use of columns as gasliquid contactors, as well as the use of
plunging jets to entrain air, has been previously tried in a number
of designs, dating from the time of the first introduction of flotation
in mineral processing. Perhaps the first to appear in the patent
literature is the column described by Norris (2) in 1907, in which
air appears to be both dissolved and entrained into a stream of pulp,
which is then introduced to the bottom of a flotation column
complete with froth crowder. Various devices for entraining air into
a pulp by means of a plunging jet are described by Taggart (3) as
"cascade machines." None of those early machines was capable of
control of the air and bubble sizes, and it is apparent that they were
designed more by guesswork than from basic principles of fluid
mechanics. A review of alternative flotation machines has recently
been given by Jameson (4).
In the design of the Jameson cell, it has been possible to take
advantage of the many advances that have been made in the
dynamics of single bubbles and of two-phase bubbly flows. Thus
the jet entry conditions are such as to make micro-bubbles of
controlled size, so as to create a large surface
Page 333
area and hence to maximize the collection process in the
downcomer. The downcomer is sized so that the bubbles created by
the plunging jet are carried downward, but at the same time the
downward pulp velocity is sufficiently low to allow bubbles to rise
against the flow and create a mixture of high gas content. The
downcomer diameter and the pulp flow must also be matched in
such a way as to ensure that any large slugs that form are carried
downward. In the froth zone, the rising velocity Jg of the gas must
be such as to allow the froth to drain properly so that the gangue is
not carried forward into the concentrate. At the same time, the Jg
must be sufficiently high to minimize the drop-back of values and
consequent loss of recovery. When washwater is used, the froth
conditions must be such that again, the improvements in grade are
not linked with reduction in recovery.
III
Principles of Operation
A
Introduction
In this section the different hydrodynamic regions occurring within
the Jameson cell are discussed, with emphasis placed on the
interactions between regions that determine the overall flotation
performance. Before identifying the different hydrodynamic
regions, however, it is useful to define the basic geometry of a
Jameson cell and also to describe qualitatively what happens
during startup and operation of a typical unit. As illustrated in
Figure 1, the simplest geometry consists of a riser section, which is
often referred to as the cell, and a single downcomeralthough some
of the larger units incorporate up to 30 separate downcomers in
combination with a common riser section.
The downcomer consists of a vertical tube that is sealed at the top
except for a regulated air inlet and a vertical nozzle through which
the slurry feed is introduced. The base of the downcomer is located
below the pulp level inside the riser. On startup, the air inlet at the
top of the downcomer is closed, and the pulp feed is pumped
through the nozzle. The air in the downcomer is entrained into the
pulp, which forms the seal in the bottom of the cell; consequently,
the pulp is drawn upward from the cell into the downcomer. The
pulp level reaches the tip of the nozzle quite quickly, and as a
consequence of the hydrostatic suction developed by the head of
this pulp, the pressure in the head of the downcomer is less than
atmospheric. When the inlet is opened, air is drawn into the
headspace at the top of the downcomer where it is entrained into
the downcomer contents by the plunging jet. The entrained air is
broken up into fine bubbles that are quickly dispersed into the pulp
and carried downward by the bulk fluid motion.
Page 334
Figure 1
Layout of a single-downcomer Jameson cell.
The three-phase mixture passes from the base of the downcomer
into the cell proper, which has a much greater cross-sectional area
than the downcomer. Consequently, the downward superficial
velocity of the mixture is reduced, allowing the particle-laden
bubbles to disengage from the liquid, rise to the surface, and form a
layer of froth. The froth then drains before overflowing into a
collection launder, while the liquid phase and unrecovered particles
leave through a valve at the base of the cell.
B
Hydrodynamic Regions
The different hydrodynamic regions that constitute the Jameson
cell are shown in Figure 2. The free jet, submerged jet, mixing
zone, and pipe flow zone occur within the downcomer, while the
disengagement zone, which includes the froth layer, occupies the
entire riser volume.
Page 335
Figure 2
The various active zones in the Jameson cell.
1
Free Jet
The free jet is defined as the stream of liquid (pulp) feed between
the tip of the nozzle and the horizontal free surface inside the
downcomer. Once the jet leaves the nozzle, its diameter is
determined by the relaxation of the velocity profile inside the jet
and also by the interaction of the free surface of the jet with the
surrounding atmosphere (5). The velocity relaxation inside the jet
occurs as a result of the change from the velocity profile, when the
pulp is inside the nozzle, to a plug flow velocity profile once the jet
has left the confines of the nozzle. Further downstream the velocity
profile changes from plug flow to parabolic flow due to the
interaction of the outer boundary layer of the jet with the
surrounding atmosphere. This interaction slows the velocity of the
jet at its free surface resulting in an expansion in the free jet
diameter.
Associated with the velocity relaxation within the jet once it leaves
the nozzle is a lateral movement that creates undulations on the jet
free surface (6). As shown in Figure 3, the undulations increase in
magnitude with
Page 336
Figure 3
The mechanisms of air transport by a free jet.
increasing free jet length thereby increasing the effective diameter
of the jet. The irregular nature of the jet surface created by the
undulations is often called the surface roughness, and jets that
exhibit this characteristic are called rough jets. Jets that display no
surface roughnessand are generated using specially designed
nozzlesare called smooth jets.
The surface roughness S is defined in Figure 3 as the difference
between the nozzle diameter DN and the effective diameter of the
jet Dj(z), which is a function of the length of the free jet Lj(z). The
surface roughness is strongly influenced by the upstream
conditions of the jet delivery system. Disturbances such as swirl
generated by a pump, vibration from a motor, or irregularities on
the inside surface of the pipework are all amplified once the pulp
passes through the nozzle, resulting in an increase in the surface
roughness.
Also illustrated in Figure 3 is the transport of air by the free jet as it
passes through the headspace at the top of the downcomer. First, a
thin annular boundary layer of air QB(z) is carried along adjacent to
the jet free surface. Second, a quantity of air QT is trapped within
the effective diame-
Page 337
ter of the jet. Therefore, the total amount of air transported by the
free jet QF can be written as
Smooth jets that exhibit no surface roughness have no trapped air

component. Hence, for two jets with the same velocity and pulp flow
rate, a rough jet transports a greater volume of air than a smooth jet.
Usually this results in a greater air/feed ratio QG/QL into the
downcomerwhere QG and QL are the air and liquid volumetric feed
rates into the downcomer, respectivelywhich leads to higher air void
fractions and interfacial areas. Therefore, it is desirable to operate
the Jameson cell using rough jets, a situation that is not difficult to
reach in practice since the jets normally have rough surfaces created
by the disturbances generated within the delivery system.
2
Plunging Jet
The plunging jet is defined as the region where the free jet impacts
with the horizontal free surface at the top of the downcomer,
resulting in air entrainment. At the point of impact, a depression is
formed in the horizontal free surface, which is often referred to by
McCarthy (6) as the induction trumpet. A half-sectioned profile of
the induction trumpet is illustrated in Figure 4, which shows the free
surface at the top of the downcomer being drawn downward by the
momentum of the free jet. The induction trumpet has a wide opening
at the top, which tapers down to a thin annular film adjacent to the
effective boundary of the free jet. The fluted entrance channels the
moving air boundary layer into the induction trumpet, and a portion
of it Qf enters the thin film, while the remaining air inside the
boundary layer QU travels radially out along the horizontal free
surface and returns to the headspace (7).
The induction trumpet periodically collapses due to instabilities
generated on its free surface, resulting in entrainment of the annular
film. Combining this quantity with the air trapped within the
effective jet diameter at the point of impact, the total entrainment
rate of the plunging jet QE is
It should be noted that the entrainment capacity of the plunging jet

given by Equation (2) is not equivalent to the air feed rate into the
downcomer. The reason is that air is being continuously circulated
within the mixing zone, and a portion QH returns to the headspace,
from which it is reentrained by the plunging jet. The air recirculated
from the mixing zone effectively reduces the amount of new air QG
that can enter the downcomer.
Page 338
Figure 4
Components of air flow into plunging jet.
3
Mixing Zone
The mixing zone is defined as the volume occupied by (1) the fluid
inside the submerged jet immediately below the plunge point that
expands to occupy the cross-sectional area of the downcomer and
(2) the body of fluid recirculating between the submerged jet
boundary and the column walloften referred to as the recirculating
eddy. The high velocity gradients between the submerged jet and
the recirculating eddy result in high energy dissipation rates within
the mixing zone, which are responsible for the breakup of the air
once it is entrained by the plunging jet. The entrained air is broken
into fine bubbles before being transported downward into the pipe
flow zone by the bulk fluid motion.
The size of the bubbles produced inside the mixing zone is related
to the forces acting on the bubble surface. In low viscosity fluids,
like those normally encountered in flotation circuits, the bubbles
are deformed by forces arising from liquid velocity fluctuations
acting over distances of the order
Page 339
of the bubble diameter. The restoring force resisting the deformation
of the bubble is due to surface tension acting at the airliquid
interface. The ratio of these forces is known as the Weber number We
and is defined as
where is the average value of the squares of the velocity

differences acting over a distance of the order of the bubble diameter
d, and rL and s are the liquid (pulp) density and surface tension,
respectively. If the bubbles are small compared to the turbulent
macroscale but large compared to the microscale, the velocity
fluctuations, following Hinze (8), can be related to the energy
dissipation rate per unit volume by the relationship
where C1» 2 according to Batchelor (9).

For a given bubble size and at low Weber numbers, the deformation
forces are dominated by the restoring forces, and the bubble is
stable. However, as We increases, the shape of the bubble becomes
increasingly distorted. Eventually a critical Weber number Wec is
reached where the bubble becomes unstable and breaks up into
smaller bubbles. The maximum bubble size dm corresponding to the
critical Weber number is
Equation (5) has been used by Evans et al. (10) to predict the
maximum stable bubble diameter inside the mixing zone, using
estimates of the critical Weber number and the energy dissipation
rate per unit volume.
The volume of the mixing zone is determined by the point of contact
of the expanding jet and the wall of the enclosing downcomer. By
considering the radial flow of momentum from the incoming jet to
the recirculating eddies in the mixing zone, Evans (11) derived the
following expression for the half angle b subtended by the expanding
jet at the virtual origin:
where h is a jet energy transfer efficiency found by experiment to be

0.089, rMZ is the density of the two-phase mixture in the mixing
zone, Rc and Rj
Page 340
are, respectively, the downcomer and jet radius, P1 is the pressure in
the headspace at the top of the downcomer, Uj is the jet velocity, and
Ue,max is the maximum return velocity in the eddy, which has been
found by experiment to be 0.085 Uj.
The volume of the mixing zone is then assumed to be the volume
enclosed by the cylinder formed by the downcomer of internal
diameter Dd and the length LMZ of the mixing zone.
Evans et al. (10) have compared the sizes of bubbles predicted by
Equation (5), with measurements of the sizes of bubbles formed in
an air water system, and found agreement within ±20%. Two
downcomers of diameters 44 mm and 74 mm were used, and the jet
velocity varied from 7.8 to 11.53 m/s. The bubble sizes were
generally in the range 200400 µm.
4
Pipe Flow Zone
The pipe flow zone is the region below the mixing zone inside the
downcomer. The flow here resembles the downward flow in a
vertical pipe. At low air flow rates, bubbly flow exists in which
small discrete bubbles of different diameters move downward with
the bulk liquid flow but not generally with the same velocity as the
liquid phase. If the air flow rate is increased, the discrete bubbles
coalesce and alternating air and liquid regions, or slugs, form inside
the downcomer. The air slugs or Dumitrescu bubbles (12) have
hemispherical caps that occupy nearly the entire cross-sectional area
of the downcomer. The rise velocity Ur of these bubbles is a function
of the downcomer diameter Dd and g, the acceleration due to
gravity:
At air flow rates when the Dumitrescu bubbles are initially formed,
the drag and viscous forces are sufficiently great to give the bubbles
a net downward velocity, and they are carried out through the base of
the downcomer. However, if the air flow rate is further increased, the
Dumitrescu bubbles become elongated, and the shearing forces
present in the liquid make them unstable. This results in a chaotic
mixture of air and liquid pockets known as churn-turbulent flow.
In terms of flow characteristics for the pipe flow zone, bubbly flow
is the most desirable because it produces the greatest collection area
for a given air/feed ratio, and it also provides for stable Jameson cell
operation. With churn-turbulent and slug flow, the surface area per
unit of gas volume is very low, and the collection efficiency is
significantly reduced. Moreover, there is a much greater tendency for
larger bubbles, with rise velocities greater than the net downward
motion, to return to the top of the down-
Page 341
comer and be reentrained by the jet, reducing the amount of new
air that can be introduced, and hence reduces the effective air/feed
ratio.
5
Disengagement Zone
When the bubbly mixture reaches the bottom of the downcomer, it
passes out into a vessel of larger cross-section, and the bubbles
then disengage from the main pulp flow. For efficient action, the
base of the downcomer is below the liquid level in the cell proper,
and a hydraulic seal is maintained. Because of its low effective
density, the bubbly flow initially hugs the outer wall of the
downcomer, but as it rises it tends to entrain slower-moving fluid
and then spreads laterally. The bubbles rise relative to the liquid
and pass from the pulp layer into the base of the froth layer. The
processes taking place in the froth are very important in the
determination of the overall grade and recovery. If the gas
superficial velocity Jg is too small or the froth depth too large, the
bubbles coalesce and the froth degrades, leading to squeezing of
the least hydrophobic particles and loss of recovery. If on the other
hand the air rate is too high and the froth depth too shallow, the
recovery may be high but at a low grade due to entrainment of
gangue. Consequently, the cell must be operated to give optimum
grade and recovery by manipulating the air rate and froth depth. If
the air rate is too high altogether, the cell may ''flood." This
phenomenon occurs when the upward gas velocity exceeds the rate
at which the liquid can drain back into the pulp layer and is
evidenced in practice by the production of a very wet froth. In fact,
when the cell floods, there is no distinction between the froth and
pulp phases, and the interface that normally exists between them is
lost. The phenomenon of flooding has been investigated in several
recent papers. Pal and Masliyah (13) have given an empirical
equation for the critical flooding velocity, and Langberg and
Jameson (14) have given a more basic description, relating the flow
in the froth and pulp phases to the flow of liquid in a packed bed,
where in the present case the packing consists of the bubbles. Xu et
al. (15) have given a comprehensive account of the various flow
regimes and methods for calculating the limiting flow rates.
Washwater can be used with the Jameson cell as with other
flotation columns, when high-grade products are required. In some
cases it is possible to achieve the grades required by proper design
of the cell, regarding the typical Jg allowed for, and by controlling
the froth depth. However, the use of washwater adds another
control variable of great power, whose proper use can lead virtually
to total flushing out of the entrained gangue. The use of washwater
does not unduly complicate the analysis of the flow presented in
Refs. 1315, and the same principles apply. The washwater flow is
sometimes used to control liquid level as well as to wash the froth,
Page 342
but this practice seems to have nothing to recommend it and should
be avoided as it adds further complexity, especially when the water
flows needed for level control purposes are such as to be in conflict
with optimum metallurgical conditions in the froth, for best grade
and recovery.
The liquid layer in the cell forms a complex system of liquid and
air recirculation patterns of varying turbulence and void fraction.
The cell design is based on downcomer flows and downcomer
placement to optimize this system to produce best grade and
recovery. There is no limit on its volume, provided the net
downward velocity of the pulp JL is sufficiently low to avoid the
entrainment of bubbles in the underflow. When the froth and
disengagement zones have the same cross-sectional area, the two
important velocities are the rate of rise of the bubbles in the pulp,
and the rate of drainage of liquid in the froth. The former is usually
greater than the latter, so that a column sized to give the correct Jg
will also give the correct JL, and bubble entrainment in the
downward flow will not be a problem.
C
Interactions and Operating Stability
The interactions taking place between regions inside the Jameson
cell and the various recycle paths taken by the air are illustrated by
the flowchart shown in Figure 5. The performance of the cell is
determined principally by the collection rate of the bubbles inside
the downcomer and to a lesser extent by the separation of the
product from the tailings stream in the riser. In general, for a given
cell cross-sectional area, the collection rate is increased by
increasing the air/feed ratio into the downcomer. However, an
increase in this ratio can also lead to instabilities and in some cases
to the complete collapse of the system. The operating stability of
the Jameson cell as a function of air feed rate is shown in Figure 6.
At low rates of air induction, the level inside the downcomer is
sustained just below the level of the nozzle, indicating that the
plunging jet can effectively entrain all of the air in the headspace.
The air is dispersed into very fine bubbles, which are carried
downward by the net fluid movement, resulting in minimal
recirculation of air back into the headspace from the mixing zone.
In general, this is the ideal mode of operation as it provides (1) the
most stable operating condition, (2) very small bubbles, and (3) the
longest residence time inside the downcomer, thereby maximizing
the bubbleparticle contact time. The only disadvantage with this
operating mode is the relatively smaller air/feed ratio inside the
downcomer, which may lead to lower rates of production of
interfacial area. (Recent unpublished work has shown, however,
that in some practical cases a reduction in the ratio is compensated
directly by the reduction in bubble size and hence an increase in the
surface area of interface produced by a given volume of
Page 343
Figure 5
Interaction paths in the Jameson cell.
air. Thus a reduction of air rate, which gives much more stable
downcomer operation, does not necessarily lead to a reduction in
recovery, providing the Jg is maintained constant.)
If the air rate is increased, a point is reached where the jet can no
longer entrain all of the air being introduced into the headspace,
and the froth level in the downcomer starts to drop. Fortuitously, it
appears that some type of compensatory phenomenon is at work.
This phenomenon has the effect that the rate of entrainment of air
by the plunging jet increases up to a point, as the length of free jet
increases. The phenomenon is possibly linked to the increase in the
effective jet diameter with increase in the free jet length. Whatever
the reason, a new equilibrium height is reached inside the
downcomer, marking the length where the jet can effectively
entrain all of the entering air. In a sense the jet is self-regulating in
that the length of the free jet will increase to accommodate the
increase in the amount of air added to the downcomer (up to a
point).
An increase in the free jet length also leads to (1) a reduction in the
froth height inside the downcomer, resulting in a reduction in the
residence time
Page 344
Figure 6
Variation of downcomer operation with airflow rate.
of the bubbles, and (2) an increase in mean bubble size generated
within the mixing zone (11), resulting in a decrease in the amount
of interfacial area generated per volume of air inside the
downcomer.
More important than the potential loss in recovery rate is the
possible effect of increasing the air feed rate on the overall stability
of the system. At very high air rates, bubble coalescence results in
the formation of large gas slugs in the pipe flow zone, which
restrict the flow of air through the downcomer. This leads to an
increase in the pressure in the headspace causing a drop in the froth
level and, in the extreme case, the complete collapse of the system.
D
Conditions in the Downcomer
A number of investigators have studied the void fraction and gas
holdup in the downcomer. Sanchez-Pino and Moys (16) measured
the hydrostatic
Page 345
pressure between two points in a vertical downcomer and used a
drift-flux method to analyze the data. The void fraction or gas
holdup varied from 20% at an air/feed ratio of 0.1 to 55% at a ratio
of 0.9. Marchese et al. (17,18) used a conductometric technique to
measure the void fraction in the downcomer and compared their
results with those from a total isolating method in which the
contents of the downcomer are trapped between two valves in-line,
which are closed simultaneously.
The values they obtained can be contrasted with those found in
conventional columns that are typically of order 15%. The high
void fractions explain the extremely rapid kinetics found in the
downcomer; the values suggest that the bubbles are in fact
approximating the close-packed spherical limit and that it would be
unrealistic to expect void fractions in excess of 55 to 60%. Void
fractions above 60% have been observed, but the most likely
explanation is that the contents of the downcomer then consist of a
dense foam of average void fraction 55 to 60%, embedded in which
is a small number of relatively large slugs or Dumitrescu bubbles,
which would contribute virtually nothing to the collection
processes in the froth.
IV
Jameson Cell Operating Parameters
Some general operating characteristics of the Jameson cell are now
discussed.
A
Froth Depth
The froth phase produced in a Jameson cell can be controlled in the
same manner as in conventional columns. Shallow froth depths
(less than 200 mm) are used where high recovery is necessary and
grade is of secondary importance, while deeper froths (up to 1000
mm) are employed to obtain maximum concentrate grade. Shallow
froths result in significant entrainment of very fine (less than 10
µm) gangue mineral components that accompany the pulp phase.
Use of deeper froths results in significant drainage of hydrophilic
gangue producing a higher grade concentrate and a higher
percentage of solids in the concentrate. Under some circumstances,
addition of counter-current froth washwater assists froth mobility,
where otherwise excessive froth drainage would result in an
immobile froth.
B
Superficial Gas Velocity
The superficial gas velocity Jg (cm/s) is the upward superficial
velocity of air in a flotation cell. In the Jameson cell, the Jg is
calculated by dividing the downcomer air rate (cm3/s) by the cross-
sectional area (cm2) of the riser part of the cell. The cell is
normally circular or rectangular in section, and
Page 346
the appropriate cross-sectional area is simply the area normal to the
direction of the flow of the froth, excluding the area occupied by
the downcomer(s). It is conveniently expressed in units of cm/s
because values typically range from 0.5 to 4 cm/s in practice.
For a stream that is not carrying-capacity limited (see Sec. IV.H),
the recove ry and concentrate carrying rate (g/min/cm2) tend to
increase with increasing Jg, as in conventional columns (see Refs.
19 and 20). For a given stream and frother concentration, a
maximum air rate (subsequently Jg,max) is reached, above which
froth flooding occurs, resulting in the loss of frothpulp interface, a
very wet froth, and total loss of selectivity. In flooding, the entire
cell fills with froth as the only stable phase, and there is no pulp
phase.
Some data of interest are shown in Figure 7 for a lead/zinc circuit.
Data were obtained using a Jameson cell in parallel with a
conventional column. In each case, it is seen that there is a strong
correlation between the production rate of solids and the superficial
gas velocity Jg. For a given air rate, the Jameson cell gives greater
production rates than the column, presumably a consequence of the
smaller bubbles generated in this device.
Figure 7
Variation of concentrate production rate with superficial air velocity Jg,
for a lead/zinc flotation, with a conventional column and a Jameson cell
in parallel. The spread in data is due to variations in froth depth. The
numbers in the legend refer to various test runs.
Page 347
The operating Jg used in the sizing of the Jameson cell depends
strongly on the application, and on the residual reagent
concentrations from any upstream processes. Generally speaking,
low values (Jg = 0.4 to 0.8 cm/s) are employed in cleaning
applications, and high Jg's (1.0 to 2.0 cm/s) are employed in
roughing or scavenging applications. The reasons for these choices
are as follows. In cleaning operations, a high proportion of the feed
reports to the concentrate, and the froth loading tends to be high.
Consequently, the bubbles are well coated with particles, which
tend to stabilize the froth. The drainage rate of the interstitial liquid
in the froth is retarded by the relatively high concentration of
particles, so it is necessary to design for lower values of Jg to allow
time for the gangue to drain from the froth to obtain the required
high grade. In roughing applications, however, only a small
fraction of the feed reports to the concentrate, and the froths
formed tend to be less stable as a consequence. Also, gangue
entrainment is not such a serious problem, because it can be coped
with in the downstream cleaning circuit. In order to maintain a
stable froth, it is therefore usual to design a Jameson cell for a
roughing application with a higher Jg than in the cleaners.
In some circumstances, high residual concentration of reagents in
the feed necessitates the use of low Jg's to avoid froth flooding.
Although frother concentration is of primary importance to bubble
size and hence the advent of froth flooding, circumstances have
risen where collector (xanthate) and frother (methyl isobutyl
carbinol, MIBC) interaction has been observed (21). In such a case,
the frother dose should be decreased if the collector dose is
increased, and vice versa. Too high a frother or collector
concentration can lead to froth flooding, while too low a dose can
lead to loss of froth stability.
Particle size can also have an influence on the maximum Jg, due to
its effect on froth stability through bubble bridging. Small particles
(less than 100 µm) are easily collected at low gas rates (Jg less than
1.0 cm/s), while recovery of coarser particles (greater than 100 µm)
may be assisted by higher rates (Jg greater than 1.0 cm/s).
C
Bubble Size
Bubble sizing determinations for full-scale operating cells and test
cells at the University of Newcastle show that the Jameson cell
produces an arithmetic mean bubble diameter of the order of
300600 µm, while the Sauter (volume-to-surface) mean diameter
dvs is of order 360950 µm (21). These sizings compare very
favorably with conventional columns where the Sauter mean
bubble size is typically 2 to 3 mm.
Figure 8 shows bubble size versus air/feed ratio for a 300-mm
diameter
Page 348
Figure 8
Measured values of the Sauter mean bubble diameter in exit stream
from a Jameson downcomer as a function of the air/feed ratio, in
flotation of a copper retreatment stream. Results for two separate
determinations are shown.
Jameson pilot cell treating a fine copper retreatment stream (d80 =
30 µm). These data confirm visual observations that bubble size
decreases with decreasing air/feed ratio.
Chatiar (22) carried out experiments in the laboratory using air and
water with MIBC as frother and found that as the concentration
increased, the mean bubble size was reduced, up to a concentration
of 30 ppm V/V, beyond which there was little change.
D
Air/Feed Ratio
Jameson cells generally operate with a volumetric air/feed ratio of
0.30.9. Experiences with large (23 m diameter) Jameson cells
indicate that operation at a low air/feed ratio does not appear to
detract from metallurgical performance providing the superficial
gas velocity Jg is maintained above a certain minimum value.
Operation at lower air/feed ratios has a stabilizing effect producing
a more uniform and finer bubble size. A significant advantage of
operation at lower air/feed ratios is that lower concentrations of
frother are required.
Maintenance of metallurgical performance at low air/feed ratios
can be explained by considering the flux of interfacial area Sb
(interfacial area/s) per unit of column cross-sectional area (s-1) as
defined by Finch and Dobby (19):
Page 349
where d is the bubble diameter. The superficial bubble surface rate

(and hence the collection capability) can be maintained with reduced
superficial gas rate providing the bubble size decreases accordingly.
The data in Figure 9 were calculated using information from Figure
8 using Equation (8) and show the superficial bubble surface rate as
a function of the air/feed ratio for a fixed cell cross-sectional area. It
is seen that the superficial bubble surface rate remains almost
constant at 55 s-1 approximately, despite two-fold changes in the
air/feed ratio.
E
Washwater Ratio
The washwater ratio is defined as the ratio of the washwater addition
rate to the flow rate of water in the concentrate:
where QWW is the washwater flow rate and QWC is the rate of flow
of water in the concentrate. The washwater ratio is a relative
measure of the amount of washwater applied.
Figure 9
Flux of interfacial area for bubbles generated by downcomer and
passing into the froth phase at the air/feed ratios shown in Figure 8.
The interfacial flow rate Sb is calculated from Equation (8). The drop
in airflow rate at low values of the ratio is mainly compensated by a
corresponding decrease in the mean bubble size.
Page 350
Another common method of describing the washwater addition is in
terms of the bias, that is, the absolute excess of the washwater
applied over the quantity of water being recovered in the
concentrate, expressed as a superficial velocity Jb (cm/s):
where AC is the cross-sectional area of the column. If no washwater

is used, the washwater ratio is zero, and the bias is negative. When
Jb = 0, W = 1; positive bias corresponds to washwater ratios greater
than unity.
Although the bias does give an indication of the absolute amount of
washwater being added, its use can be misleading because it does not
take into account the wide variation in the absolute values of the rate
of water entrainment in the concentrate. It is preferable to use the
washwater ratio, which is a relative figure.
Consider, for example, a rougher application in which the mass flow
of recovered solids is only a small fraction of the feed. In such a
case, the flow of concentrate will be low and a correspondingly low
volume of washwater will be needed to replace the water being
carried out of the cell in the concentrate. Thus good froth washing
may be achieved with a relatively low positive bias, say 0.005 cm/s.
On the other hand, in a cleaning operation, where the mass flow of
concentrate is high, a much higher superficial flow rate of washwater
would be needed to replace the water in the concentrate, since there
is much more water being removed in the froth.
Figure 10 shows a comparison of measured washwater ratios and
bias rates for a range of operating conditions for a Jameson cell
operating on a zinc rougher application. Although the bias covers a
relatively narrow range, the washwater ratio has much greater
variation. To an operator, the difference in bias between 0.02 and
0.04 cm/s may seem negligible, but it can be seen that in some
circumstances, such changes in Jb may lead to a threefold increase in
the washwater ratiosupplying in effect a 200% overload of
washwater.
Although it might be thought that a cell should be operated with W >
1, there is evidence to show that the optimum operating value should
be in the range 0.5 to 1. Figures 11 and 12, for example, show the
effect of washwater ratio on recovery and grade, respectively, of a
fine copper retreatment stream (d80 = 30 µm). As the washwater
ratio is increased, the recovery decreases significantly up to W » 1
and then decreases more gradually for W > 1. The grade increases
progressively up to a washwater ratio of 1.0, but no further grade
improvement is observed at washwater ratios above this value. These
data, together with other Jameson cell operating
Page 351
Figure 10
Plot of superficial washwater bias Jb against the washwater ratio
calculated for the same zinc rougher flotation runs. Note the large
variation in washwater ratio for relatively small changes in the
absolute bias.
data, confirm that the best washwater operating point is at a
washwater ratio of slightly less than unity, where recovery and
grade are both high. Higher washwater addition simply results in
decreased recovery without compensating improvements in grade.
Precise on-line washwater bias measurement is highly desirable in
order to maximize recovery at required grade.
F
Reagents
In streams tested to date, MIBC, long-chain alcohols, polyglycol
propylenes, and polyglycol ethers have been used as the frothers in
the feed to the Jameson cell, usually in the range 5 to 25 ppm. In
cleaning applications, generally no frother addition is required due
to the residual concentration from the roughing stage. In some
cases, excessive frother from the upstream stage can lead to a
reduction in maximum superficial gas rate Jg, which can be applied
in the Jameson cell. Froth flooding is initiated at lower air rates due
to the finer bubble size generated by an excess frother
concentration (see Ref. 19, p. 22).
Page 352
Figure 11
Effect of washwater ratio on recovery in the flotation of a copper
retreatment stream. The apparent scatter here and in Figure 12 is due to
differences in flotation conditions between individual runs, particularly
in the froth depth and air rate. Note that the progressive decline in
recovery as washwater is increased.
Figure 12
Grade vs. washwater ratio in flotation of a fine copper retreatment
stream. The grade increases markedly until the washwater ratio
reaches a value of 1 approximately and thereafter remains constant.
Page 353
G
Downcomer Void Fraction
Experimental studies by Evans (11) and Marchese et al. (17,18) have
shown that the void fraction (gas holdup) in a Jameson downcomer
is generally in the range 50 to 60% by volume.
The downcomer void fraction can be measured directly, either from
the isolating technique or from conductivity measurements. It can
also be calculated using a downcomer momentum balance.
The gas void fraction Î in the dense foam inside the downcomer is
defined as
where VG and VL are the volumes occupied by the gas and liquid
(pulp) phases, respectively. In the isolating method, the contents of a
section of the downcomer are trapped between two valves that are
closed simultaneously, thus allowing direct measurement of VG and
VL.
With the conductometric method, use is made of classical potential
theory to relate the measured conductivity to the voidage. Thus
Marchese et al. (17) applied the relationship of Maxwell (23) and
found
where kL and kLG are the measured conductivities of the liquid

(pulp) and the liquidgas mixture, respectively.
The void fraction can also be estimated from measurements of the
pressure in the headspace in the downcomer. In the absence of
dynamic effect, Î can be calculated from simple hydrostatic pressure
considerations. Thus a pressure balance over the downcomer
contents yields
where P1 and P2 are the pressures in the headspace and in the

discharge end of the downcomer, respectively; h is the height of the
dense foam in the downcomer; and rm is the average density of the
two-phase mixture inside the downcomer. When the gas density is
negligible compared with the liquid density, it can be shown that Î is
given by
Thus from Equation (14) we have

Page 354
Dynamic effects can be included by application of the momentum
principle to the froth zone. The sum of the forces acting on the froth
is equal to the gain in momentum flow rate between inlet and outlet,
so
where Ad is the area of cross-section of the downcomer and QG and

QL are the gas and liquid flow rates, respectively. Ignoring the term
in the gas density, we find:
where Aj is the cross-sectional area of the jet. It is seen by

comparison with equation (15) that a dynamic correction term has
been introduced into the expression for the void fraction. The
correction is usually significant, contributing 25 to 50% of the
calculated overall void fraction, depending on conditions.
H
Carrying Capacity
The carrying capacity Ca is the limiting or maximum concentrate
production rate per unit of area of cell cross-section, usually
expressed as concentrate solids rate (g/min) per unit of available cell
cross-sectional area (cm2). The carrying capacity is strongly
influenced by the gas flow rate and the size of the bubbles in the
froth discharging at the overflow lip, since these two factors
influence the rate of surface area discharge into the launder. The
mass of the hydrophobic particles that can be carried by the froth
varies directly with the surface area. The other important factor is the
size of the particles, because when a layer of particles is adhering to
a gasliquid interface, the mass of particles per unit of interfacial area
varies directly as the mean particle size.
Carrying capacity limitations generally apply where high-grade
streams are treated, as in cleaning applications, where the bubbles
can easily become fully loaded with particles. The recovery is then
determined by the carrying capacity rather than by the flotation
kinetics.
The expected carrying capacity of conventional columns has been
given by Espinosa-Gomez et al. (24) as
where d80 is the size at which 80% by mass of the concentrate

passes, expressed in µm, and rp is the density of the particles
(g/cm3).
In conventional columns, the carrying capacity depends on column
diameter. The parameter a has been found to be 0.068 in small
columns (as reported by Espinosa-Gomez et al. (24) for a 50-mm
column). For larger
Page 355
columns, up to 1.0 m diameter, Finch and Dobby (19) give a =
0.05; and for columns greater than 2.0 m diameter, a = 0.035 (25).
Figure 13 shows carrying capacities measured by Atkinson et al.
(21) as a function of the washwater ratio in a Jameson cell treating
coarse chalcopyrite (autogenous mill discharge, d80 = 400 µm). It
is seen that the maximum carrying rate is strongly influenced by
washwater addition.
These data show that the carrying capacity for a Jameson cell must
be defined at a particular washwater ratio. For maximum
concentrate grade and recovery, it is generally necessary to aim for
a washwater ratio of 1.0 or just slightly lower. Thus it would be
appropriate to define the carrying capacity at a washwater ratio of
1.0. In all Jameson cell applications, concentrate carrying
capacities have been shown to be at least as high as the values
predicted for conventional columns using Equation (18).
V
Applications
A
General
Table 1 shows a summary of sites employing Jameson cells in
mineral or coal flotation applications. In most applications, the
Jameson cell is capable of providing a final grade concentrate in a
single stage, providing the
Figure 13
Concentrate production rate as function of washwater ratio for coarse
chalcopyrite flotation, showing a strong reducing trend as washwater
ratio is increased.
Page 356
Table 1 Summary of Jameson Cell Sites (January
1993).
Company Location Project
Mount Isa Mount Isa, Lead/Zinc
Mines Australia slimes
Mount Isa Mount Isa, SX-EW
Mines Australia
Mount Isa Hilton, Australia Low-grade
Mines middlings
Mount Isa Hilton, Australia Lead cleaning
Mines
Peko Mines Warrego, Australia Copper cleaning
Amalg Moonta, Australia Copper cleaning
Syndicate
Mamut Copper Malaysia Copper cleaning
Newlands Coal Qeensland, Fine coal
Australia
Amalg Spargoville, Nickel cleaning
Syndicate Australia
Amalg Spargoville, Nickel roughing
Syndicate Australia
Kidd Creek Timmens, Canada Copper/Zinc
Mount Isa Mount Isa, Fine zinc
Mines Australia cleaning
Western Mining Olympic Dam, SX-EW
Corp Australia
Phelps Dodge Morenci, USA SX-EW
Cons Murchison South Africa Antimony/gold
Matos Blancos Chile Copper cleaning
C.C.P. Collinsville, Fine coal
Australia
Sth Atlantic Oracle Ridge, USACopper
Ventures roughing
White Mining Nth Goonyella, Fine coal
Australia
BHP Australia Blackwater, Fine coal
Coal Australia
Girilambone NSW, Australia SX-EW
values are sufficiently liberated and washwater is applied (where

necessary). The Jameson cell is efficient for the recovery of the full
range of particle sizes experienced in flotation feed streams.
Examples of extremes include the recovery of coal slimes (d80 =
25 µm) and very fine sphalerite/galena particles (d80 = 10 µm) at
the lower end of the particle size spectrum and, at the coarse end of
the range, coal particles up to 1000 µm and chalcopyrite particles
up to 400 µm.
Details of the metallurgical performance of operating Jameson cells
have been given in a number of papers. Kennedy (26) described
some of the early full-scale applications, in the lead/zinc
concentrators at Mt. Isa, Queensland, and Hilton, Queensland; at
the copper concentrator at the Peko Warrego mine, Tennant Creek,
Northern Territory; and at the Newlands Coal operation in the
Bowen Basin, Queensland, all in Australia.
The commissioning and operation of the Peko Warrego copper
flotation cells has been described by Jameson et al. (27). The
original concentrator produced a concentrate of 23% copper from a
chalcopyrite ore containing pyrite, magnetite, hematite, and quartz,
with traces of covellite and bis-
Page 357
muthinite. The reason for the Jameson cell installation was to
increase the concentrate grade from 23 to 28% or better, in order to
reduce transport costs. Two 1.4-m circular cells were installed and
commissioned in April 1990, taking a feed of up to 40 tph dry
solids at 35% solids in the slurry. The required grade increase,
obtained in the original pilot-plant tests, has been proven in plant
operation.
A description of the Jameson cell installations at the Mt. Isa and
Hilton mines has been given by Jameson and Manlapig (28) and
Clayton et al. (29).
The flotation of fine coal slimes at the Newlands Coal operation
has been described by Jameson et al. (30). The washery has an
output of 1200 tph of coal, which is washed mainly in Batac jigs.
The water stream from the jigs is allowed to settle and is returned
for reuse. A side stream of 1250 m3/h is taken off to prevent
accumulation of slimes in the circuit, and previously this bleed
stream had been sent to a tailings dam. However, the solids in the
stream were in the range 3 to 8%, representing a considerable loss
in ultrafine coal, around 200,000 tpa. Accordingly it was decided to
recover the very fine coal, whose d50 is 22 µm, by flotation. Tests
with mechanical cells showed that it would not be possible to
recover the coal at the required ash level (12% maximum) except at
unacceptably low yields, but with the Jameson cell it was found
that the required ash level could readily be achieved, without the
use of washwater. With a single stage of operation, the combustible
recovery was 70%, and with two stages the overall recovery was
90% or better. A full-scale installation was commissioned in early
1990 and has operated ever since. Large installations are scheduled
to start in 1993 at two other coal washeries in the Bowen Basin, at
the BHP Australia Coal Blackwater mine and the North Goonyella
joint venture.
B
Rougher Applications
The Jameson cell has been successfully employed in roughing
applications and can produce a concentrate of final grade straight
from run-of-mine ore. In such cases, it is possible to upgrade the
capacity of a mill that has been operating with conventional
flotation machines. The Jameson cell is then used as a scalper to
remove a significant proportion of the feed, removing the load
from the rest of the existing circuit, which can then be used to
recover and clean the tails from the Jameson cell.
An interesting set of data has been obtained when a Jameson cell
was set up to treat an autogenous mill discharge stream, on a
chalcopyrite ore. Figure 14 shows the overall grade-recovery
performance of the Jameson cell for single stage treatment of this
stream. The average recovery was
Page 358
Figure 14
Grade-recovery data for single-stage Jameson cell flotation of discharge
stream from autogenous mill discharge, chalcopyrite flotation circuit. Feed
size d80 = 400 µm.
50%, at a concentrate grade of 27% Cu, from a feed grade of 3%
Cu. The size analyses of the feed and the corresponding
concentrate are shown in Figures 15 and 16, respectively. The d80
of the feed is of the order of 400 µm, while the d80 of the
concentrate is 50 µm; most of the floatable particles are in the < 20
µm fraction.
These data demonstrate the wide-ranging capability of the Jameson
cell as an excellent flotation machine, capable of treating all
particle sizes typically encountered in mineral processing.
VI
Control Requirements
The principal control requirements for a Jameson cell are
1. froth depth
2. washwater ratio
Froth depth control is essential for concentrate grade stability. It
can be readily achieved by some type of direct or indirect level
sensor (e.g., bubble pressure probe, diaphragm pressure sensor,
ultrasonic sensor, float level indicator) or by a simple gravity
overflow. In most applications the froth depth is between 300 and
800 mm, although for high-recovery operations where grade is not
so important, froth depths as low as 50 mm are used. Where
washwater is used, some measure of control is desirable in order to
maximize recovery while maintaining required concentrate grade.
There is
Page 359
Figure 15
Cumulative size distribution of chalcopyrite flotation feed,
autogenous mill discharge.
Figure 16
Cumulative size distribution, concentrate from flotation of
autogenous mill discharge; d80 » 50 µm.
Page 360
some latitude in that a high grade is usually achieved for all
washwater ratios greater than 1, but recovery may suffer if too
much washwater is used.
Aeration of pulp occurs in the Jameson cell in the downcomer. In
order to cope with feed changes, it is customary to design the cell
for a maximum expected flow rate. Variations in feed can be dealt
with by a number of strategies, including a variable-speed drive on
the feed pump, controlling make-up water additions in the feed
sump, and recycling of tails or an intermediate stream to maintain a
constant flow. The cell design has been developed so that isolation
of downcomers in conjunction with suitably positioned baffle
plates allows fractions of the cell to be operated without affecting
Jg or other design fundamentals. The solution chosen depends on
the circumstances pertaining in the plant.
VII
Amenability Testing
Amenability testing is the term used to cover pilot plant testing for
the purpose of assessing a technology for a given application. Table
2 provides a matrix covering the ranges of variables that should be
tested in order to evaluate the capabilities of a Jameson cell fully.
In particular applications, some of the conditions may be omitted
or given lower importance, e.g., mineral cleaning applications
would always require washwater, whereas roughing applications or
coal flotation may not. Each test should be repeated at least three
times to ensure that experimental outliers are identified.
Table 2 Jameson Cell Amenability Testing Program
for a Given Stream.
Washwater ratio Froth depth (mm) Air/feed ratio
Test 0 0.5 1.0 250 500 750 0.4 0.6 0.8
1 * * *
2 * * *
3 * * *
4 * * *
5 * * *
6 * * *
7 * * *
8 * * *
9 * * *
10 * * *
11 * * *
Page 361
As described earlier, the washwater ratio is an extremely important
variable, particularly in mineral slurry applications. For pilot-plant
trials, washwater requirements can easily be estimated by
collecting and weighing a timed concentrate sample. From
knowledge of the approximate concentrate percent solids, the
concentrate water rate can be found, which then allows setting of
washwater addition to a nominated washwater ratio. In the absence
of measured solids contents, it can be assumed as a good
approximation that the flow rate of water in the concentrate
(L/min) is the same numerically as the flow rate of froth
concentrate (kg/min). The procedure should be repeated after
initially setting up washwater addition, as the addition of
washwater itself affects concentrate recovery rate.
Acknowledgments
A worldwide exclusive license has been granted to MIM Holdings,
Ltd., 400 Ann Street, Brisbane 4000, Australia, for the use and sale
of the Jameson cell for metallurgical purposes. We are grateful to
MIM Holdings, Ltd for their assistance in the preparation of this
article. We also acknowledge the assistance of J. A. Finch, M. H.
Moys, G. Harbort, and M. Marchese. One of us (BWA)
acknowledges the support of an Australian Post-graduate Research
Award, in conjunction with MIM Holdings, Ltd. and The
University of Newcastle Research Associates Limited (TUNRA).
References
1. G. J. Jameson, A new concept in flotation machine design, Min.
Met. Process., Feb., 4447. (1988).
2. D. H. Norris, Apparatus for separating the metallic particles of
ores from the rocky constituents thereof, US Patent 873586, 10
December (1907).
3. A. F. Taggart, Handbook of Mineral Processing, Wiley, New
York, pp. 1261. (1927).
4. G. J. Jameson, ''Flotation Cell Development," Proc. Ann. Conf.,
Aus. Inst. Min. Metall., pp. 2531 (1992).
5. M. J. McCarthy and N. A. Molloy, Review of stability of liquid
jets and the influence of nozzle design, Chem. Eng. J., 7: 120.
(1974).
6. M. J. McCarthy, "Entrainment by plunging jets," Ph.D. Thesis,
University of Newcastle, Australia, 1972.
7. G. M. Evans and G. J. Jameson, "Prediction of the Gas Film
Entrainment Rate for a Plunging Liquid Jet Reactor," AIChE
Spring National Meeting, Houston, paper 34G (1991).
8. J. O. Hinze, Fundamentals of the hydrodynamic mechanism of
splitting in dispersion processes, AIChE J., 1: 289295. (1955).
9. G. K. Batchelor, Proc. Cambridge Philos. Soc., 41: 359. (1951).
Page 362
10. G. M. Evans, G. J. Jameson, and B. W. Atkinson, Prediction of
the bubble size generated by a plunging liquid jet bubble column,
Chem. Eng. Sci., 47: 32653272. (1992).
11. G. M. Evans, "A study of a plunging jet bubble column," Ph.D.
Thesis, University of Newcastle, Australia, 1990.
12. D. T. Dumitrescu, Flow past an air-bubble in a vertical pipe, Z.
Angew. Math. Mech., 23: 139149. (1943).
13. R. Pal and J. H. Masliyah, Flow characteristics of a flotation
column, Can. Metall. Q., 29(2): 97103. (1990).
14. D. A. Langberg and G. J. Jameson, The coexistence of the froth
and liquid phases in a flotation column, Chem. Eng. Sci., 47:
43454355. (1992).
15. M. Xu, J. A. Finch, and A. Uribe-Salas, Maximum gas and
bubble surface rates in flotation columns. Int. J. Miner. Process.,
32: 233250. (1991).
16. S. E. Sanchez-Pino and M. H. Moys, "Characterisation of Co-
Current Downwards Flotation Columns," Column '91, Proceedings
of an International Conference on Column Flotation (G. E. Agar,
B. J. Huls, and D. B. Hyma, Eds.), Sudbury, Ontario, July 1991,
Vol. 1, 341355 (1991).
Measurement of gas holdup in a three-phase concurrent downflow
column, Chem. Eng. Sci., 47: 34753482. (1992).
"Hydrodynamics of a Downflow Column," to be presented at
XVIII International Mineral Processing Congress, Australasian
Institute of Mining and Metallurgy, Sydney, Australia, 2328 May
(1993).
19. J. A. Finch and G. S. Dobby, Column Flotation, Pergamon
Press, Oxford. (1990).
20. J. A. Finch and G. S. Dobby, Column flotation: a selected
review. Part I, Int. J. Miner. Process., 33: 343354. (1991).
21. B. W. Atkinson, P. T. Griffin, G. J. Jameson, and R. Espinosa-
Gomez, "Jameson Cell Test Work on Copper Streams in the
Copper Concentrator of Mount Isa Mines Limited," to be presented
at XVIII International Mineral Processing Congress, Australasian
Institute of Mining and Metallurgy, Sydney, Australia, 2328 May
(1993).
22. R. Chatiar, "Report on Test Work on the Operating Parameters
of the Jameson Cell," Internal Report, Department of Chemical
Engineering, University of Newcastle, February (1992).
23. J. C. Maxwell, A Treatise on Electricity and Magnetism, 3rd
ed., Vol. I, Part II, pp. 435449, Oxford University Press, London
(1892).
24. R. Espinosa-Gomez, J. A. Finch, J. B. Yianatos, and G. S.
Dobby, Flotation column carrying capacity: particle size and
density effects, Miner. Eng., 1(1): 7779. (1988).
25. R. Espinosa-Gomez and N. W. Johnson, "Technical experiences
with conventional columns at Mount Isa Mines Limited," Column
Flotation (G. E. Agar, B. J. Huls, and D. B. Hyma, Eds.), Sudbury,
Ontario, July 1991, pp. 511524 (1991).
26. A. Kennedy, The Jameson flotation cell, Min. Mag. October
(1990).
Page 363
27. G. J. Jameson, G. Harbort, and N. Riches, "Development and
Application of the Jameson Cell," Proc. 4th Mill Operators
Conference, Burnie, Tasmania, 1014 March 1991; Melbourne: The
Australasian Institute of Mining and Metallurgy (1991).
28. G. J. Jameson and E. V. Manlapig, "Applications of the
Jameson Cell," Column '91, Proceedings of an International
Conference on Column Flotation (G. E. Agar, B. J. Huls, and D. B.
Hyma, Eds.), Sudbury, Ontario, July 1991, pp. 673687 (1991).
29. R. Clayton, G. J. Jameson, and E. V. Manlapig, The
development and application of the Jameson cell, Miner. Eng.,
4(711): 925933. (1991).
30. G. J. Jameson, M. Goffinet, and D. Hughes, "Operating
Experiences with Jameson Cells at Newlands Coal Pty Ltd,
Queensland," Proc. 5th Australian Coal Preparation Conference,
Australian Coal Preparation Society, Newcastle, NSW, Australia,
1217 May (1991).
Page 365
13
Removal of Organics from Water by Nonfoaming
Flotation
Kalliat T. Valsaraj
Louisiana State University, Baton Rouge, Louisiana
I
Introduction
A large number of organic compounds that are of significance as
water pollutants have low aqueous solubilities and are slightly
volatile or nonvolatile in nature. They are mostly nonpolar, neutral
hydrophobic compounds. Removal of these compounds at dilute
concentrations from the aqueous phase is a challenging problem in
separation science. Use of activated carbon or macroreticular resin
columns for removing these compounds involves large capital and
operating costs, which often cannot be justified by the value of the
reclaimed water. Due to the hydrophobic nature of the compounds
and their airwater interfacial behavior, flotation on air bubbles can
be an alternative technology to remove them from the aqueous
phase. The simplicity of the equipment (viz., a bubble column) and
the ease of operation make flotation processes attractive for the
removal of refractory organic compounds. Dissolved or induced air
flotation is a common waste water treatment process for the
removal of suspended solids, oils, greases, fibers, and other low-
density solids as well as for the thicening of activated sludge.
Flotation processes are classified as foaming and nonfoaming
separations. The first category requires the presence of surfactants
to produce a relatively stable foam that acts as the carrier fluid for
the removal of floated particles; the second, on the other hand, does
not. This chapter is concerned
Page 366
exclusively with the nonfoaming separations that are composed of
three techniques, namely, bubble fractionation, solvent sublation,
and predispersed solvent extraction. All three involve transport of
solute from the aqueous phase on tiny air bubbles (see Fig. 1).
Nonfoaming techniques are especially attractive for waste water
treatment since they involve neither the formation of large
quantities of foam nor the need to collapse and dispose of the wet
foam.
Over the past decade my group has been involved in the
investigation of one of these processes, namely, solvent sublation,
for the removal of a variety of hydrophobic compounds from waste
water. I will summarize in this chapter the various mechanisms of
solute transport in solvent sublation, the advantages of solvent
sublation over a conventional extraction process, recent work on
the application of sublation in semibatch and continuous
countercurrent modes for water treatment, and a brief review of
predispersed solvent extraction.
II
Advantages of Solvent Sublation
Bubble fractionation is a process involving only two phases,
namely, aqueous and air (1). Both solvent sublation and
predispersed solvent extraction, on the other hand, involve three
phases, namely, aqueous, air, and an immiscible organic solvent
(2,3). Figure 1 is a comparison of the three nonfoaming flotation
processes in the semibatch mode.
Solvent sublation involves the transport of solutes by air bubbles
from the aqueous phase to an overlying immiscible organic solvent.
Volatile material will be transported within air bubbles and
nonvolatile materials on the surface of air bubbles. As the bubbles
transit the organic solvent layer, the materials carried will be either
solubilized in the solvent or deposited on the solventair interface
when the bubble breaks through the solvent surface. A semibatch
sublation with the air phase in a one-pass mode of contact with the
aqueous and organic solvent phases is shown in Figure 1(b). The
process can be conducted in two other modes as wellair and
aqueous phases in continuous countercurrent contact but with a
stagnant organic solvent layer, or with all three phases in
continuous contact (4).
Solvent sublation enjoys several advantages over conventional
solvent extraction. Conventional extraction suffers from a few
drawbacks that make it less useful for waste water treatment. These
are: (a) the need for a mixersettler stage; (b) the need to comminute
the phases; (c) the formation of emulsions in the aqueous phase
that are difficult to break; (d) extensive dissolution of the organic
solvent within the pregnant aqueous solution; and (e) a large
volume of the organic solvent to the aqueous pregnant solution. In
order to explain why solvent sublation is more useful than
Page 367
Figure 1
Schematic of single-
stage, semibatch bubble (a) fractionation, (b) solvent sublation, and (c) predispersed solvent
extraction processes.
Page 368
conventional extraction, I shall enumerate the mechanisms of
solute transport between the aqueous and organic solvent phases.
III
Mechanisms of Solute Transport in Solvent Sublation
Figure 2 is an illustration of the modes of solute transport at the
aqueousorganic solvent interface. There are three important
transport mechanisms:
1. a unidirectional transport of solute by air bubbles
2. a molecular diffusion mass transfer, driven solely by the solute
activity gradient between the phases such as in solvent extraction
3. transport associated with the water droplets carried up into the
organic solvent by the air bubbles.
The first mechanism is dependent on the specific surface area of
the bubbles, the airflow rate, the mass transfer coefficient for the
solute from the aqueous to the air bubbles, and the equilibrium
relationship for the solute between the airwater interface provided
by air bubbles and the water phase. The second mechanism is a
competing one that is the primary transport pathway in
conventional solvent extraction. It depends on the solute
Figure 2
Mechanisms of solute transfer between the aqueous and
organic solvent phases in solvent sublation.
Page 369
activity gradient between the phases, the mass transfer coefficient for
solute from the aqueous to organic solvent phase and the area of
contact between the organic solvent and aqueous phases (i.e., the
cross-sectional area of the column). It is a molecular diffusion driven
process and is slow in comparison to the first mechanism. The third
mechanism involves solute transport with the water droplets taken
into the organic solvent by the air bubbles. The water so transported
returns as droplets that are depleted of solute. The amount of water
transported is so small that this mechanism of solute transport is
negligible in comparison to the other two mechanisms. Conventional
extraction equilibrium between the water and organic solvent exists
only in the case of the third mechanism. Since the first two transport
mechanisms are rate-controlled, the process of solvent sublation is a
rate-controlled process.
In order to illustrate the competing effects of the dominant transport
mechanisms, let us consider a nonvolatile solute of low aqueous
solubility and high hydrophobicity (e.g., neutral pentachlorophenol)
subjected to solvent sublation in a semibatch mode as shown in
Figure 1(b). Let us assume that the air bubble in the aqueous phase is
in equilibrium with the liquid surrounding it. Large aqueous phase
mass transfer coefficients for the solute will ensure that equilibrium
is established in the aqueous phase. The following analysis is
therefore for a semibatch, single-stage, equilibrium-controlled
operation.
The rate of change of solute in the aqueous phase is given by
Equation (1) (5):
where Cw is the solute concentration in the aqueous phase (moles
cm-3), Co is the solute concentration in the organic solvent layer
(moles cm-3), k1 is the mass transfer rate coefficient for solute
transport between the organic and aqueous phases (cm s-1), rc is the
radius of the sublation column (cm), Kmw is the distribution
coefficient for solute between the organic solvent and aqueous
phases (moles cm-3/moles cm-3), Vw is the total volume of the
aqueous phase in the column (cm3), and r is the average radius of an
air bubble (cm). Gm and C1/2 are constants in the Langmuir
isotherm, which gives the equilibrium concentration of the solute at
the airwater interface provided by the air bubbles. The Langmuir
equation is given by the following relationship
Page 370
where Gm is the maximum adsorbed concentration (moles cm2) and
C1/2 is the aqueous solute concentration when G is Gm/2 and has
units of moles cm-3.
Since the process is semibatch, we have from mass balance
considerations
where is the initial solute concentration in the aqueous phase and

Vo is the volume of the organic solvent layer.
At steady state the left-hand side of Equation (1) will be zero. Then
using Equation (3) to eliminate Co from Equation (1), we have the
following quadratic in Cw:
where
and
In order to obtain a physically realistic value of concentration, we

choose only the positive root of the quadratic and hence the aqueous
phase solute concentration at steady state is given by
and the solute concentration in the organic solvent phase at steady
state is
The partition constant for the solute between the organic solvent and
aqueous phases in semibatch sublation at steady state is given by
. An enhancement factor can be defined as given by
.
Let us now explore the effects of various parameters on the
enhancement factor E. For this we choose a set of standard
parameters: column radius rc = 5 cm, phase ratio, Vw/Vo = 100, air
flow rate G = 1 cm3 s-1, average air bubble radius in the aqueous
phase r = 0.05 cm, mass transfer coefficient between the phases k1 =
0.001 cm s-1, Langmuir isotherm constant Gm =
Page 37
3.3 × 10-10 moles cm-2, C1/2 = 2 × 10-7 moles cm-3, and the initial solute concentration in the aqueous
phase moles cm-3. The solute parameters chosen are for neutral pentachlorophenol
(PCP) and the immiscible organic solvent chosen is mineral oil. As has been described elsewhere, there
is considerable uncertainty in the value of C1/2 and hence it is an adjustable parameter (6). Apart from
C1/2, the three other significant parameters in the model sare G, k1, and r. Let us choose a value of Kmw
for the solute between mineral oil and aqueous phase to be 480.
Figure 3 shows the effects of airflow rate and bubble radius upon the enhancement factor. With an
increase in G from 1 to 10 cm-3 s-1, the value of E increases from 1 to 13.5. However, with increasing
air bubble radius the enhancement falls off, since larger bubble radius means decreased airwater
interfacial area per unit volume of air. In actual practice, air bubble
Figure 3
Effects of air flow rate and air bubble radius upon the enhancement factor in solvent sublation (semibatch operation)
Page 372
radius increases with G and hence large air flow rates afford only minimal
enhancement in sublation (5,7). Thus the primary limitation to solvent sublation on
a large scale at present is the necessity to use low air flow rates so as to maintain the
small size of air bubbles. The presence of large air bubbles is also detrimental in that
it increases the axial circulation of the aqueous phase that will destroy the
concentration gradient in the aqueous phase (8). These observations are also
important in other nonfoaming separations.
With increasing k1, the mixing and back diffusion of the solute from the organic
solvent layer to the aqueous phase become more significant. This tends to bring the
two liquid phases to equilibrium faster. At high values of k1, the maximum removal
obtained is therefore that of conventional solvent extraction. Figure 4 shows that
with increasing k1, the enhancement factor tends toward one. Increasing the air flow
rate will help maintain the two
Figure 4
Effects of k1 and C1/2 upon the enhancement factor in solvent sublation (semibatch operation).
Page 373
liquid phases under nonequilibrium conditions and hence larger
enhancements can still be achieved with increasing k1. The
enhancements in the removal of several hydrophobic species by
sublation as compared to conventional extraction has been
demonstrated experimentally by several investigators (5,9,10).
Figure 4 shows that increasing C1/2 also decreases the
enhancement factor. C1/2 is related to the binding energy for a
hydrophobic species at the airwater interface of the air bubble.
With decreasing C1/2 the binding energy increases (6), thereby
increasing the airwater interfacial concentration and the capability
of air bubbles to carry materials into the organic solvent.
Apart from the above characteristics of solvent sublation, there is
one other important aspect that deserves mention. In solvent
sublation, it has been observed that the degree of redissolution of
the organic solvent in the aqueous phase is smaller than that due to
conventional solvent extraction (11,12). Table 1 lists the residual
solvent concentrations in the aqueous phase after a single-stage,
semibatch solvent sublation at steady state and a single-stage
solvent extraction using the same volumes of aqueous and organic
solvent.
IV
Applications of Solvent Sublation
A large number of neutral hydrophobic compounds (volatile,
semivolatile, and nonvolatile) and a number of ionizable organics
complexed with surfactant counter ions have been removed by
solvent sublation on a semibatch scale. A number of different
solvents have been used. These solvents satisfy certain criteria,
viz., they are immiscible with water, have large affinities for the
solutes, have low interfacial tensions with water, and are
nonvolatile. For the treatment of waste water and drinking water,
the solvent should also be nontoxic. Table 2 is a list of compounds
removed by solvent sublation along with the type of solvents used.
Table 1 Comparisons of Residual Solvent
Concentrations in the Aqueous Phase After Sublation
and Extraction.
Residual solvent
concentration in
moles L-1
Air flow rate (for
sublation) in mL
Solvent min-1 SublationExtraction
Methyl 65 0.135 0.194
isobutyl
ketone (12)
Amyl acetate 30 0.0056 0.0095
(11)
Page 374
Table 2 List of Organic Compounds Removed by
Solvent Sublation on Semibatch Scale.
Compound Solvent Reference
Volatile hydrophobic organics:
1,1,1-trichloroethane octanol 13
chloroform mineral oil 14
1,2-dichlorobenzene octanol, 6,16
mineral oil
toluene octanol 32
1,2-dichlorobenzene octanol 32
Semivolatile hydrophobic organics:
1,2,4-trichlorobenzene mineral oil 16
1,2,4-trichlorobenzene lauryl alcohol 16
Aroclor 1254 (a PCB mineral oil 6
mixture)
p,p'-DDT mineral oil 16
lindane mineral oil 6
endrin mineral oil 6
aldrin mineral oil 17
naphthalene mineral oil 18
naphthalene decyl alcohol 7
phenanthrene mineral oil 18
biphenyl mineral oil 19
hexachlorobutadiene mineral oil 15
Nonvolatile hydrophobic organics:
2,4,6-trichloroanisole mineral oil 34
phenol amyl acetate 11
chlorophenols isopropanol 20
pentachlorophenol mineral oil 21
pentachlorophenol decyl alcohol 5
indene mineral oil 17
magenta dye mineral oil 23
direct red dye mineral oil 24
bis(2-ethyl hexyl) mineral oil 25
phthalate
phenol and dissolved methyl 12
organics isobutyl
ketone
HTMAB (surfactants) 2-octanol 9,26
dibutyl phthalate mineral oil 25
trichlorophenol mineral oil 7
trichlorophenol decyl alcohol 7
Nonvolatile ionsurfactant complexes: 6
di octanol
nitrophenolate/HTMAB
picrate/HTMAB octanol 6
chlorophenolate/HTMABisopropanol 20
chlorophenolate/HTMABdecyl alcohol 5
methylene blue/SDS octanol 27
rhodamine B/SDS octanol 28
methyl orange/HTMAB octanol 28
metal ions + surfactants octanol, 22,2933
benzene
Page 375
Only a few studies on the use of solvent sublation as a continuous
process have been reported. Air and aqueous phases in continuous
counter-current mode with a stagnant organic solvent layer has been
reported from my laboratory (7,35,36).
A mathematical model where both the aqueous and organic solvent
phases are in continuous counter-current contact with air can be
obtained by extending the model for semibatch sublation described
earlier. The operation will be assumed to be single stage, equilibrium
controlled, and steady state in nature. The aqueous phase will be
assumed to be a well-mixed stage connected in series with the thin
solvent layer, which will also be considered well mixed. Equations
can be derived for the solute concentration in both the aqueous phase
and the organic solvent layer. The equation for the steady-state
concentrations of a nonvolatile solute in the aqueous phase is given
by
where L is the aqueous feed rate, Cinfl is the solute concentration in

the aqueous feed, and Cw is the solute concentration in the aqueous
raffinate at steady state. In order to simplify our discussion further
and avoid dealing with a quadratic in Cw, we shall assume that the
concentration of solute in the aqueous phase is such that we are
working in the linear region of the Langmuir adsorption isotherm
[Eq. (2)]. In other words, we have the condition that in
Equation (10) and the first term on the lefthand side simplifies to -
G(3/r)(Gm/C1/2) Cw. This is a valid approximation in most cases,
since the solute concentration in the aqueous phase is generally very
small.
At steady state we have the following overall mass balance
relationship
where Lo is the organic solvent flow rate and Co is the solute

concentration in the extract organic layer at steady state.
The degree of separation (separation factor) at steady state in either
counter-current sublation or extraction is given by the ratio of solute
concentration in the organic solvent extract (Co) to the solute
concentration in the aqueous raffinate (Cw). Using Equation (11) in
(10), we have for the degree of separation in continuous counter-
current sublation, Kss = Co/Cw, the following equation:
Page 376
Figure 5 is an illustration of the variation of Kss with k1 using
Equation (12). The parameters used are the same as those for the
semibatch conditions described earlier. The aqueous feed rate was
chosen to be 1 cm3 s-1. The volumetric rate of mineral oil was chosen
to be 0.05 cm3 s-1. If k1 is zero, the diffusional mass transfer between
the two liquid phases is negligible. The value of the degree of
separation is then dependent on that due to air bubbles and is directly
proportional to (G/Lo)(3/r)(Gm/C1/2). With a very large k1, the value
of Kss approaches Kmw very rapidly. An approximation to k1 would be
the ratio D/d, where D is the diffusion constant for the solute in the
aqueous phase and d is the thickness of the diffusion boundary layer
between the aqueous and organic solvent phases. A typical D value
would be of the order of 10-6 cm2 s-1 and a typical d value would be
of the order of 10-3 cm. Hence a typical k1 value would be about 10-3
cm s-1. As shown in Figure 5, for such a k1 value one can exceed the
Kmw value by increasing the value of G and decreasing r so that the
term (G)(3/r)(Gm/C1/2) is quite large in comparison to . The
influence of k1 is large when separation by air bubbles is small, i.e.,
extraction mechanism makes a bigger contribu-
Figure 5
Effects of air flow rate and k1 upon the separation factor in continuous counter-
current
solvent sublation.
Page 377
tion when the solute transport by air bubbles is small. If air flow is
turned off, the only mechanism of removal is that of counter-current
solvent extraction from the aqueous to the organic solvent phase and
the degree of separation Kext = Co/Cw is given by
The ratio of the degree of separation in sublation to that in solvent

extraction is defined as an enhancement factor E for steady-state,
continuous counter-current operation. This is given by Equation
(14).
Equation (14) is the same expression for E in semibatch operation

described earlier if only the linear portion of the Langmuir
adsorption isotherm is retained in Equations (1) and (2). From
Equation (14) we conclude that when the separation by air bubbles is
large and the contribution from molecular diffusion is small, large
enhancements in steady-state removal by continuous counter-current
sublation are possible. Separation by air bubbles will be significant if
air flow rate G is large and the isotherm parameter C1/2 and air
bubble radius r are small. However, with increasing G, the air bubble
radius r will also increase, and hence there is an optimum set of
values that will give the maximum removal as demonstrated by
Valsaraj et al. for counter-lakcurrent sublation (36). According to
Equation (14), the parameters that determine the enhancement factor
are air flow rate, air bubble radius, adsorption isotherm parameters,
column radius, and molecular diffusion mass transfer coefficient
between the aqueous and organic solvent phases.
If the aqueous phase is contained in a tall, narrow column, it will
have to be modeled as a series of well-mixed stages. If we consider
N stages in the aqueous phase, the separation factor for sublation will
be given by Equation (15).
When compared with the separation factor for a single-stage process

given by Equation (12), we observe that Equation (15) is larger by a
factor of
Page 378
Table 3 List of Organic Compounds
Removed by Counter-Current Solvent
Sublation.
Compound Solvent Reference
2,4,6- mineral 34
Trichloroanisole oil
2,4,6- mineral 36
Trichlorophenol oil
Pentachlorophenolmineral 36
oil
Pentachlorophenoldecyl 38
alcohol
Naphthalene mineral 36,39
oil
PCP + HTMAB decyl 38
alcohol
{1 + [(G/L)(3/r)(Gm/C1/2)] + [G/L)(3/r)(Gm/C1/2)]2 + . . . + [(G/L)

(3/r) (Gm/C1/2)] (N-1)}. Thus the separation factor in sublation can
be improved considerably if more than one stage is made possible
in the aqueous phase. This can be accomplished by reducing the
axial dispersion in the aqueous phase by suitable baffling of a tall,
narrow flotation column.
Both the semibatch and continuous counter-current models
described previously are for steady-state situations with respect to
solute removal from the aqueous phase. More importantly, they
assume that the rate of mass transfer of solute to the air bubble in
they aqueous phase is large so that the air bubble is in equilibrium
with the aqueous phase surrounding it. This may not be true if the
height of the aqueous layer is small and the average bubble
velocity is large so that the air bubble has a short residence time in
the aqueous phase (4,6,8,37).
Several compounds have been removed by continuous solvent
sublation in which the air phase was in a single-pass mode of
contact with both the aqueous and organic solvent phases. The
aqueous phase was in counter-current with air, but the organic
solvent phase was stagnant. Both ionic species complexed with
surfactant counter ions and neutral hydrophobic species have been
successfully removed from the aqueous phase (Table 3). The
removal efficiencies were crucially dependent on the air bubble
radius, and hence only low aqueous and air loading rates were
possible.
Based on preliminary investigations of the feasibility of solvent
sublation, it has become clear that significant limitations exist for
scale-up of the process for treatment of waste waters (1,7).
Specifically, the need to produce and maintain micrometer size air
bubbles in the aqueous phase, the low air and water loading rates
that are required for a successful sublation, and the need to reduce
the axial dispersion in the column are to be addressed on a pilot
scale column.
Page 379
V
Predispersed Solvent Extraction
Another flotation/extraction separation process that is a variation of
solvent sublation was suggested by Sebba recently (41). A
schematic of the process in a semibatch mode is shown in Figure
1(c). It involves precomminuting solvents into polyaphrons, which
are used to scavenge solutes from the aqueous phase. The
polyaphrons are then buoyed up to the aqueous surface by
microbubbles called colloidal gas aphrons (CGAs). Flotation using
CGAs is necessary because the polyaphrons by themselves rise
through the water phase very slowly. Upon reaching the water
surface the polyaphrons coalesce leaving a thin layer of organic
solvent loaded with the solute. Both the formation of CGAs and
polyaphrons require small amounts of surfactant. They are both
generated with minimum effort and are stable for extended periods
of time. More details on the generation, structure, and properties of
both CGAs and polyaphrons are described in an excellent
monograph (3) and a project report by Sebba (40).
Just as with solvent sublation, this process also avoids an energy
wasteful mixersettler stage that is required of a conventional
extraction process. Formation of emulsions in the aqueous phase is
avoided and the volume of the organic solvent required to treat the
pregnant solution is very small, on the order of one to a thousand or
even lower. Because of the large surface area and rapid equilibrium
for the polyaphron with its surrounding aqueous phase, solute mass
transfer to the solvent phase is large. Extraction will be virtually
complete using a very small volume of the organic solvent to
pregnant solution. For example, Sebba (41) demonstrated that an
oil soluble dye, waxoline blue at 2 parts per billion concentration
can be removed from 2 L of water into 1 ml of kerosene
polyaphron within a minute of contact time.
Significant differences exist between the theory of conventional
solvent extraction and predispersed solvent extraction. Both
processes are dependent on a large partition constant for the target
solute between the organic solvent and aqueous phases. However,
in a conventional extraction apparatus, the transfer from the
aqueous to the solvent phase that occurs in the mixersettler stage
ceases when equilibrium is established between the phases. In
predispersed solvent extraction, on the other hand, equilibrium
exists only between each aphron and the solution immediately
surrounding it, just at the water surface where the aphron is
effectively removed from contact. Thus there is a procession of
aphrons moving through the aqueous column that is progressively
depleted of the solute, each aphron contacting a small volume of
aqueous solution that is lower in solute concentration than that
contacted by its predecessor. Sebba defined a depletion volume
Page 380
that refers to the volume of the pregnant solution with which the
aphron is in equilibrium before leaving the aqueous phase (41).
Based on simple equilibrium theory for the semibatch process shown
in Figure 1(c), the following equation was derived for the moles of
solute left in the pregnant solution after extraction by n aphrons (41):
where a is the moles of solute in the depletion volume VD before

extraction, Kmw is the equilibrium partition constant for the solute
between the solvent and aqueous phase, and v is the volume of a
single polyaphron. If f is the ratio of the total volume of organic
solvent to the total volume of the pregnant aqueous solution, then n
is given by fVD/v. Thus the value of n is dependent on the ratio f and
can be quite large if the size of aphrons is also small. The value of v
is typically on the order of 1 µm3 and VD is typically of the order of
500 µm3. If Kmw is very large, then the value of xn/a can be quite
small with an n value as small as 100 or so. Thus it is the value of n
that makes predispersed solvent extraction so efficient.
Both CGAs and polyaphrons require the presence of surfactants for
their generation. Hence when the aphrons rise to the surface they
tend to foam slightly. However, this can be avoided and the
formation of foam eliminated by a judicious choice of surfactants for
CGA and polyaphron and by taking advantage of the contrafoam
phenomena that Sebba has discussed (41). Thus, for example, a
polyaphron made using an anionic surfactant requires a CGA with a
cationic surfactant.
Metal ions such as copper, uranium, and chromate ions have been
removed by single-pass extraction using polyaphrons to
concentrations that were virtually undetectable in the aqueous
raffinate. Several applications of polyaphrons were listed by Sebba
(41).
There is an important difference between predispersed solvent
extraction and solvent sublation. In sublation the only mixing and
contact between the aqueous and organic solvent phases occur at the
top of the aqueous phase. In predispersed solvent extraction,
however, the organic solvent in the form of polyaphrons contacts the
entire aqueous phase. Therefore, chances of dissolution of organic
solvent in the pregnant aqueous solution will be greater in
predispersed solvent extraction. Because of the use of surfactants for
the generation of CGAs and polyaphrons, residual surfactants in the
aqueous phase are also likely in predispersed solvent extraction. As a
result, after removal of the solute-laden solvent, a subsequent
flotation will be necessary to clear up the aqueous phase of both
residual solvent and surfactants.
Only one reported use of predispersed solvent extraction for the
removal
Page 381
of organics from water is known. Michelsen et al. (42) reported a
feasibility study for the removal of 1,2-dichlorobenzene from waste
water. The process was a single-pass extraction using polyaphrons
made of decane as the solvent and conducted in a flotation cone
apparatus. They observed that predispersed solvent extraction was
about 5 to 10 times more effective in removing 1,2-
dichlorobenzene from water at less than 1/500 solvent/feed ratio
than with identical tests using a similar quantity of straight solvent.
Scale-up tests were recommended for accurately assessing the
applicability of predispersed solvent extraction for waste water
treatment operations.
VI
Conclusions
Nonfoaming flotation/extraction processes such as solvent
sublation and predispersed solvent extraction have significant
advantages over a conventional solvent extraction process.
Specifically, they do not require the use of an expensive
mixersettler stage as do conventional extraction operations.
Separation factors can exceed those achieved by a conventional
solvent extraction process. Moreover, the amount of residual
extracting solvent in solvent sublation has been observed to be
much smaller than that in extraction. Flotation/extraction processes
are limited in their effectiveness only due to the low air and
aqueous loading rates that are necessary to maintain the small size
of air bubbles so as to maximize the bubblewater interfacial area
for adsorption. Further studies to avoid this limitation are necessary
before these processes can enjoy large-scale applicability.
Acknowledgments
I thank the U.S. National Science Foundation for financial support
of my research in adsorptive bubble separations and the LSU/EPA
Hazardous Waste Research Center for providing the facilities.
References
1. K. T. Valsaraj, G. J. Thoma, L. J. Thibodeaux, and D. J. Wilson,
Nonfoaming adsorptive bubble separation processes, Sep. Technol.,
1: 234 (1991).
2. R. Lemlich (Ed.), Adsorptive Bubble Separation Techniques,
Academic Press, New York (1972).
3. F. Sebba, Foams and Biliquid FoamsAphrons, Wiley, Chichester,
U.K. (1987).
4. K. T. Valsaraj and L. J. Thibodeaux, Studies in batch and
continuous solvent sublation. II. Continuous countercurrent solvent
sublation of neutral and ionic species from aqueous solutions. Sep.
Sci. Technol., 26: 367 (1991).
5. K. T. Valsaraj and L. J. Thibodeaux, Studies in batch and
continuous solvent
Page 382
sublation. I. A complete model and mechanisms of sublation of
neutral and ionic species from aqueous solutions, Sep. Sci.
Technol., 26: 37 (1991).
6. K. T. Valsaraj and D. J. Wilson, Removal of refractory organics
by aeration. IV. Solvent sublation of chlorinated organics and
nitrophenols, Colloid Surf., 8: 203 (1983).
7. K. T. Valsaraj, X. Y. Lu, and L. J. Thibodeaux, Continuous
countercurrent solvent sublation for the removal of hydrophobic
organics from aqueous solutions, Water Res., 25: 1061 (1991).
8. A. N. Clarke and D. J. Wilson, Foam Flotation: Theory and
Applications, Marcel Dekker, New York (1983).
9. I. Sheiham and T. A. Pinfold, Mechanism of sublate removal
studied by solvent sublation. II., Sep. Sci., 5: 619 (1970).
10. B. L. Karger, Solvent sublation, Adsorptive Bubble Separation
Techniques (R. Lemlich, Ed.), Academic Press, New York (1972).
11. R. B. Grieves, W. Charewicz, and S. M. Brien, Separation of
phenol from dilute alkaline aqueous solutions by solvent sublation,
solvent extraction and foam fractionation, Anal. Chim. Acta, 73:
293 (1974).
12. B. T. Nolan and W. F. McTernan, Application of solvent
sublation for the simultaneous removal of emulsified coal tar and
dissolved organics, Chem. Eng. Comm., 63: 1 (1988).
13. T. Lionel, D. J. Wilson, and D. E. Pearson, Removal of
refractory organics by aeration. I. Methylchloroform, Sep. Sci.
Technol., 16: 907 (1981).
14. K. T. Valsaraj and L. J. Thibodeaux, Diffused aeration and
solvent sublation for the removal of volatile organics from aqueous
solutions, ACS Preprint Abs. (Div. Env. Chem.), 27: 220 (1987).
15. K. Y. Shih, W. D. Han, and S. D. Huang, Solvent sublation of
hexachlorobutadiene and 2,4,6-trichlorophenol, Sep. Sci. Technol.,
25: 477 (1990).
16. K. T. Valsaraj, J. L. Porter, E. K. Liljenfeldt, and C. Springer,
Solvent sublation for the removal of hydrophobic organic
compounds from aqueous solutions, Water Res., 20: 1161 (1986).
17. L. K. Foltz, K. N. Carter, and D. J. Wilson, Removal of
refractory organics by aeration. VII. Solvent sublation of indene
and aldrin, Sep. Sci. Technol., 21: 57 (1986).
18. S. D. Huang, K. T. Valsaraj, and D. J. Wilson, Removal of
refractory organics by aeration. V. Solvent sublation of naphthalene
and phenanthrene, Sep. Sci. Technol., 18: 941 (1983).
19. W. K. Wang and S. D. Huang, Solvent sublation and adsorptive
flotation/sublation of diphenyl, Sep. Sci. Technol., 23: 375 (1988).
20. M. Caballero, R. Cela, and J. A. Perez-Bustamante, Solvent
sublation of some priority pollutants (phenols), Anal. Lett., 21: 63
(1988).
21. K. T. Valsaraj and C. Springer, Removal of traces of
pentachlorophenol from aqueous acidic solutions by solvent
extraction and solvent sublation. Sep. Sci. Technol., 21: 789 (1986).
22. A. D. Jones and C. Robinson, Solvent extraction and adsorptive
bubble separation of metal ions from aqueous solutions. II.
Adsorptive bubble separation
Page 383
of Ni(II) using carboxylic acids and their salts as collectors, J.
Inorg. Nucl. Chem., 36: 1871 (1974).
23. G. L. Shen and S. D. Huang, Solvent sublation and adsorbing
colloid flotation of magenta, Sep. Sci. Technol., 22: 2253 (1987).
24. M. H. Cheng and S. D. Huang, Solvent sublation and adsorbing
colloid flotation of direct red, J. Colloid Interface Sci., 126: 346
(1988).
25. K. Tamamushi and D. J. Wilson, Removal of refractory
organics by aeration. VI. Solvent sublation of alkyl phthalates, Sep.
Sci. Technol., 19: 1013 (1985).
26. H. H. Tabak and R. L. Bunch, Measurement of non-ionic
surfactants in aqueous environments, Proc. Purdue Ind. Waste
Conf., 36: 88 (1981).
27. J. L. Womack, J. C. Lichter, and D. J. Wilson, Removal of
refractory organics by aeration. II. Solvent sublation of methylene
blue and methyl orange, Sep. Sci. Technol., 17: 897 (1982).
28. B. L. Karger, A. B. Caragay, and S. B. Lee, Solvent sublation
and extraction of methyl orange and rhodamine B, Sep. Sci., 2: 39
(1967).
29. R. Cervera, R. Cela, and J. A. Perez-Bustamante, Analytical
solvent sublation of metallic dithizonates. Part I. Solvent sublation
of copper, Analyst, 107: 1425 (1982).
30. P. E. Spargo and T. A. Pinfold, Mechanism sublate removal
studied by solvent sublation II., Sep. Sci., 5: 619 (1970).
31. Z. Szeglowski, M. B-Jankowska, and J. Mikulski, Frothless ion
flotation of some rare earth elements, Nukleonika, 18: 299 (1973).
32. H. S. Shin and R. W. Coughlin, Removal of organic compounds
from water by solvent sublation, J. Colloid Interface Sci., 138: 105
(1990).
33. D. Huiru, Y. Shouzhen, and L. Jialai, Solvent sublation-
spectrophotometric determination of trace (Ni(II) in drinking water,
Anal. Sci., 5: 601 (1989).
34. K. T. Valsaraj and C. Springer, Solvent sublation for the
removal of a taste and odor causing organic from aqueous solution,
J. Indian Chem. Soc., 65: 369 (1988).
35. K. T. Valsaraj and L. J. Thibodeaux, Nonfoaming adsorptive
bubble processes for wastewater treatment, Preprint No: 92-43 for
the Annual Meeting of the Society for Mining, Metallurgy and
Exploration, Inc., Phoenix, AZ, February 2427 (1992).
36. K. T. Valsaraj, X. Y. Lu, and L. J. Thibodeaux, Continuous non-
foaming adsorptive bubble separation process for the removal of
hydrophobic organic compounds from the aqueous phase, ACS
Symp. Series., 509: 116 (1992).
37. J. M. Smith, Chemical Engineering Kinetics, 2nd ed., McGraw
Hill, New York (1970).
38. K. T. Valsaraj, X. Y. Lu, and L. J. Thibodeaux, Studies in batch
and continuous solvent sublation. III. Solubility of
pentachlorophenol in alcoholwater mixtures and its effects on
solvent sublation, Sep. Sci. Technol., 26: 977 (1991).
39. D. J. Wilson and D. E. Pearson, Solvent Sublation of Organic
Contaminants for Water Reclamation, Report OWRT/RU-84/5,
U.S. Dept. of Interior, Washington, D.C. (1984).
Page 384
40. F. Sebba, Investigations of the Modes of Contaminant Capture
in CGA Foams, Report OWRT/RU-82/10, U.S. Dept. of Interior,
Washington, D.C. (1981).
41. F. Sebba, Sep. Purif. Methods, 14: 127 (1986).
42. D. L. Michelsen, K. W. Ruettimann, K. R. Hunter, and F.
Sebba, Feasibility study on the use of predispersed solvent
extraction/flotation techniques for the removal of organics from
wastewaters, Chem. Eng. Comm., 48: 155 (1986).
Page 385
14
Electrolytic Flotation:
An Unconventional Technique
I
Introduction
Electrolytic flotation or electroflotation is a rather unconventional
flotation technique, consisting of electrolysis of the aqueous part of
a dispersion, followed by flotation of the contained suspended
particles, colloids, or ions with the help of generated fine bubbles
of hydrogen and oxygen, which will form a foam ''blanket" at the
surface of the flotation cell. Therefore, it is believed that the term
electrolytic flotation is a better description of the process than the
more common electroflotation. This process is generally envisaged
as a system involving a tank in which two electrodes (mainly
metalliclike Fe and Albut also graphite) are being placed
horizontally at the bottom creating, during the passage of direct
electric current, a stream of tiny bubbles, which result in the
flotation process. It is also possible, when sodium chloride is used
as electrolyte (or, more practically, sea water), that chlorine
evolution at certain current densities occurs. This separation
method was applied to different areas ranging from water and
waste water treatment to mineral processing.
When an effluent is brought between two electrodes, of which one
is the positive anode and the other is the negative cathode, and
electricity is supplied to the electrodes, an electric field is built up
between them through the use of the suspension conductivity (1).
Even without any other addition of chemical reagents, a
preliminary coagulation occurs within the particulate
Page 386
matter of the effluent, which results in the grouping of the negative
and positive particles together. Furthermore, with the
decomposition of water into its constituents, free atomic oxygen is
evoluted in the diffusion layer of the anode. It is carried by
convection into the waste water and immediately combines with
the present organic and inorganic constituents, resulting in its
oxidation. In a similar manner, there is also a transformation
brought about by the electrically excited hydrogen, resulting in
reduction of the contained constituents.
The first scientist who observed that electrolysis could also be used
for sewage treatment was Webster in 1887 (2). He erected a plant
that treated some of London's sewage. In 1946, Ryvkin et al.
applied electrolytic flotation for mineral beneficiation purposes (3).
The mineral slurry (pulp) was being mixed in the flotation cell by
the use of an ordinary agitator. The distinguishing characteristic of
the used chamber (cell) was attributed to the fact that a set of
electrodes was placed inside it, which allowed electrolytic
preparation of the slurry to be carried out directly in the cell,
simultaneously with flotation.
In technological enrichment processes, the interaction forces of an
electric field and the features of electric charge movement have
often been used (4). Generally, this process is termed
electroseparation and takes place when electricity acts directly on
the treated material without necessarily converting it to other forms
of energy in order to enhance or accomplish separation. Although
electroseparation methods proved to be very effective in some
cases (like electrodialysis), in other cases (like the separation of
fine ore fractions, in particular, for hydrophobic minerals with low
electrical conductivity) it seems to be quite unsuitable (5). These
difficulties stimulated the investigators to search for other ways of
using electricity in order to improve the technical and economic
results for efficient mineral beneficiation. In this way, electrolytic
flotation has been developed mainly in what was formerly the
USSR (6).
The methods, which have been used for the enhancement of
flotation as an ore beneficiation process, are directed primarily
toward the improved preliminary preparation of the mineral grain
surfaces and secondarily toward the creation of suitable conditions
for the attachment of selected mineral particles onto the gas
bubbles (mainly air, but also nitrogen). In the first case, the
presence of an optimum reaction process seems to be basic. The
solution of the second problem is based on the creation of proper
conditions for perfect aeration of the mineral slurry, i.e., the
number and the size of bubbles, the possible use of surface active
agents, etc. To a certain degree, electrolytic flotation provided
solutions for both problems. With its help and due to the
application of hydrolysis of the surface-determining ions, the
preparation of grain surfaces from the minerals to be
Page 387
floated was realized in order to achieve a better adherence of them
to the bubble surface. More suitable conditions were created also
for the mineralization of air bubbles in the flotation cell.
Electrolytic flotation has also been applied to water and waste
water treatment (7,8). More specifically, it has been applied for the
reduction of suspended solids (9) and the removal of COD (10),
phosphates (11), and dissolved toxic metals (like Cu, Zn, Ni, and
Cd) from plating operations (12), as well as for sludge thickening
(13). Another application is for oily and refinery effluents,
particularly when the treatment unit is required to break and
separate oil/water emulsions and not free oil droplets (14).
The electrolytic flotation process has (to some extent) already been
reviewed (15,16). In the first review, more attention was given to
the application of the process for fine mineral particle
beneficiation, and results have been presented for calcite and
quartz. Calcite has been examined (for comparison) by applying
also dissolved-air flotation, a dominant technique applied mainly to
waste water treatment. The problem of processing fine mineral
particles constitutes an immense challenge, as the application of
conventional methods is generally uneconomical.
II
Literature Review
Historically, one of the first electrolytic flotation applications was
the purification of radioactive waste waters, containing traces of
90Sr and 137Cs (17). This technique was examined also as an
attractive alternative to dissolved-air flotation in water treatment
technology (1820). The effect of the most significant operational
parameters of electrolytic flotation upon bubble size characteristics
and bubble concentration was assessed quantitatively (18). Usually
in water treatment a solid/liquid separation process is applied in
order to lower the present suspended solid concentration, before the
application of granular filtration. This enables filters to be operated
more effectively, lowering simultaneously the cost of producing a
filtered water of suitable quality, which can be used after
disinfection for potable supply (21). Generally, flotation requires
smaller tanks and produces better treated water quality, compared
to sedimentation (22).
The collection mechanism of colloidal particles (polystyrene
lattices of 0.6 µm in diameter) in electrolytic flotation was studied,
both experimentally and theoretically (23), as an extension of an
earlier work (24). The electrolytic flotation of fine dispersions has
also been studied (25).
Electrolytic flotation was applied at bench-scale batch experiments
for the removal of hexavalent chromium anions from dilute
aqueous solutions (26). Very good results were obtained, and in
some cases removals of the
Page 388
order of 100% were noticed. The optimum conditions are
summarized in Table 1.
The technique was also applied for the selective separation and
recovery of ions from multicomponent systems (containing Ni, Co,
Cu, Zn, Ti, and Mo), over a wide pH range, with the addition of
ferrocyanide, oxyquinoline, hydroxide, and carbonate precipitants
(27). The removal of other toxic metals (mainly, Mo) from acid
mine tailing waste waters (28) or from chemical industry waste
waters (mainly Cd, Cr(III), and CN-) have also appeared in the
literature (29).
After 2 years of pilot plant experimentation, an integrated closed-
loop waste/process water treatment was devised, having as its main
components an electrolytic flotation and a reverse osmosis system
(30). This approach had to take into account the new treated
effluent standards, as well as the process water requirements of an
airline maintenance basement. First results of the full-scale system
operation were presented and discussed.
A simultaneous electrolytic flotation and disinfection sewage
treatment was reported elsewhere, after chemical coagulation and
flocculation and in the presence of chloride ions at various
concentrations (31). The anodic oxidation of chloride ions gives
hypochlorite ions, which could be used for the subsequent
disinfection of the treated sewage. The results, obtained from a
continuous small pilot plant, showed that solid/liquid separation
was improved over static clarification for at least 2 h.
The electrolytic flotation of a palm oil mill effluent was also
described using a lead dioxide-coated titanium anode as first-stage
treatment, with an aerobic or physicochemical process as a
secondary treatment step (32). The anode was found to be
corrosion-resistant under the applied conditions.
Table 1 Electrolytic Flotation of Cr(VI)-Containing
Solutions: Optimum Conditions.
Initial Cr(VI) content: 10 mg L-1
Horizontal electrodes made from stainless steel
Current density: 200 A m-2 Voltage: 18 V
Flotation time: 900 s Frother (optional): 0.4% ethanol
Agitation (accomplished with a mechanical stirrer): 300
rpm
pH value: 39, i.e., rather independent
Conductivity: 709000 µmhos
Temperature: 2050°C
Surfactants: cetyl-terimethyl-ammonium bromide 2 × 10-4
M (aqueous solution) or dodecylamine 4 × 10-4 M
(ethanolic solution, 0.5%)
No negative effect was noticed when ionic strength was
increased
Page 389
About 40% of the initial COD and 86% of suspended solids could
be removed. Waste waters from a meat industry were also
successfully treated with electrolytic flotation (33).
In another paper, an electrolytic cell was evaluated for the
treatment of oily waste water and an automatic control system was
installed (34). In this process, the oil emulsion was electrolytically
demulsified, the iron anode was dissolved producing Fe(III) ions,
and the subsequent application of dissolved-air flotation gave an
oil-rich floating sludge containing Fe(OH)3 and an oil-free water.
The use of electrolytic flotation for oily waste water treatment has
also been studied in pilot or in full-scale plants (35,36).
In mineral processing, electrolytic flotation was tried for the
separation of carbonate (magnesite and dolomite) fines. A novel
electrolytic flotation microcell of Hallimond type was reported
(37). The separation of this group of minerals from each other is
extremely difficult due to similar flotation responses. For this
reason, and in order to obtain selectivity, various modifiers were
tested, both of inorganic and organic nature. However, problems
were noticed with the flotation of artificial mixtures attributed
mainly to hydrolysis and to interactions of the dissolved cations.
One of the interesting conclusions of this paper was the observation
that electrolytic flotation could also operate collectorless, a
conclusion that is in agreement with other literature findings (38).
This technique also was applied at laboratory scale for the
enrichment of fine particles of enargite (63% of which have had
diameters below 8.8 µm), which is a mixed copper-arsenic sulfide
(39). The results were compared to those obtained in a standard
Hallimond tube, but using nitrogen for bubble production and not
(as usual) air. The microcell, specially constructed for these
experiments, was a modified Hallimond tube to which platinum
electrodes and an electrolyte receiver were adapted in order to
obtain an electrolytically generated H2-O2 mixture. The
disadvantage of this system is its operation in rather high current
densities, which are needed for optimum recoveries.
Another interesting bench-scale apparatus was used for
chalcopyrite fines, of average size below 4.42 µm (40). Flotation
tests were carried out with hydrogen or oxygen separately, as a
membrane diaphragm was dividing the two horizontal electrodes
and an electrolyte was flowing to sweep away the gas of the lower
electrode, preventing it from entering the flotation zone. From this
paper, however, values of applied current density are not available;
they cannot even be deducted (as the electrode area was not
reported). On the other hand, the applied voltage (4.59.0 V) was
quite low, hence acceptable.
Recently, the electrolytic flotation of quartz fines was published
(41). The influence of changing the diameters of hydrogen or
oxygen bubbles,
Page 390
when they are used separately, on fine recovery was studied. A
change in the electrode surface geometry (plates or mesh), and
subsequently in current density, alters the bubble diameters, as well
as bubble flux, thereby affecting the metallurgical results obtained.
III
The ProcessSelected Results and Discussion
There is reasonable evidence that the capital cost of flotation
installations is lower than that of sedimentation plants of the same
capacity. Overflow rates of 24 mm s-1 obtainable with flotation are
compared with 0.31.3 mm s-1 rates of sedimentation. These high
treatment rates suggest significant initial construction capital
savings. A space savings of around 80% in comparison with
settlement facilities of similar throughput was reported (42).
Certainly, the introduction of gas bubbles in flotation cells leads to
a higher power consumption than for solids settling.
Initial work on electrolytic flotation was conducted using a dilute
aqueous solution of an emulsion paint (white color), as the
laboratory simulation of an effluent liquid, because such a mixture
was quite difficult to clarify by conventional means (8).
Additionally, this system had the further advantage of simple
quantitative analysis by a spectrophotometric technique in the
visible range of the spectrum.
Typical batch results are shown in Figure 1, where three different
current densities and two initial concentrations of 1000 and 2000
mg L-1 were examined. These results were produced using a
cylindrical flotation cell, made from Perspex and having a 170-cm2
surface area, 1-m height, and 17-L volume. The cell was equipped
with two horizontal, circular, stainless steel electrodes, having a
diameter almost equal to that of the column. Flotation retention
time was found to decrease by increasing current density or
decreasing the initial paint concentration. It is worth noting here
that the higher current densities, although they lead to higher
achievable removal rates, simultaneously introduce increased
operating costs, so it becomes a matter of best optimization (43).
Optimum conditions in continuous mode are summarized in Table
2. Furthermore, in an economic study of electrolytic flotation, the
relative costs were calculated and compared with those of
dispersed-air flotation (blowing air through a porous diffuser). This
comparison is presented as Table 3.
The effect of the initial pH value on electrolytic flotation is quite
critical, as in any other flotation technique. The final (treated) paint
concentrations of the effluent and the relative retention times have
been found to depend strongly on the pH, correlating well with the
corresponding values. In this system, flotation removal rates were
very high within a narrow pH range
Page 391
Figure 1
Electrolytic flotation of a paint dispersion.
and very low outside it. However, in the case of Cr(VI) removal,
electrolytic flotation was found to be efficiently applicable over a
wide pH range of 39 (see Fig. 2).
Also, experiments with a bipolar cell were reported. A bipolar
arrangement is characterized by the fact that only the terminal
electrodes are connected to the source of electric current, while the
rest of the electrodes conduct each other through the solution
conductivity (7).
Table 2 Electorlytic Flotation as a Solid/Liquid
Separation Process: Operation DataContinuous Mode.
Flow rate (maximum): 5 Effluent flow: 4.56 m3 per
cm3 s-1 day
Feed concentration: 1500 Tank capacity: 0.17 m3
mg L-1
Suspended solids removal: Interelectrode gap: 3.2 mm
(minimum) 90% Power consumption: 3.6 MJ
m-3
Retention time: 55 min Current density: 300 A m-2
No surfactants/chemical Current intensity: 51 A
additives
Concrete or steel tank Voltage: 21.3 V
Transformer three-phase, Horizontal set of electrodes
oil-immersed (primary (platinum expanded mesh as
specifications 2400/4160 V, anode, stainless steel 304
secondary 120/240 V) sheet as cathode)
Page 392
Table 3 Economic Comparison of Two Flotation
Techniques (Same Application).
Electrolytic flotation Dispersed-air flotation
Operational costs: 0.034 Operational costs: 0.024
Pounds (per 4 m3) Pounds (per 4 m3)
Additional capital costs Additional capital costs
(Pounds) (Pounds)
electrodes 20.28 diffuser 4.00
transformer 547.8 blower 7,358
Note: UK currency figures to be updated (from 1975).
In Figure 3 some other results are presented, when different initial

concentrations of sodium chloride or sea water were used. These
may be compared with the tests of dispersed-air flotation (see Fig.
5).
The principle of electrolyzing sea water, i.e., using sea water in
order to influence the conductivity of the dispersion during
electrolytic flotation, is
Figure 2
Influence of the solution pH to Cr(VI) removal with and without the
presence of CTMA-Br 2 × 10-4 M. The other conditions were Cr(VI)
concentration of 10 mg L-1, Fe electrodes, flotation time of 15 min,
current density of 200 A m-2, and an external magnetic stirring speed
of 125135 rpm.
(Reprinted with permission from Ref. 26.)
Page 393
Figure 3
Electrolytic flotation of a paint dispersion at 200 A m-2.
quite interesting (7). During electrolysis, alkalinity is created at the

cathode (separate electrodes were used), and phosphorus is
precipitated together with flocculent magnesium hydroxide. It is
noted that phosphate supply is a limiting factor in algae growth.
Increasing the conductivity of the effluent by salt addition
decreased the necessary voltage (hence, the power input). This
power savings (calculated as being about 22.7 pence), due to
increased conductivity, can easily cover the purchasing price of
sodium chloride for the 0.1 % addition (being 0.82 pence, for the
same throughput). The running cost of electrolytic flotation is
largely a function of the conductivity of the effluent. Of course,
there is the possibility of using sea water in combination with the
effluent for water or waste water treatment plants that are located
near the coast. It was observed that the use of 5.5% sea water had
approximately similar effects with a 0.1% NaCl concentration,
while the presence of 33.5% sea water gave equal paint removals
with the presence of 1% NaCl. When increasing the conductivity,
attention should be paid to anode corrosion, mainly by the chloride
ion. The data from experiments with 0.1 % NaCl gave a voltage of
around 22 V for an interelectrode gap of 2.54 cm. For a gap of 0.32
cm, a voltage of only 2.8 V was found to be sufficient.
Page 394
It was also found (Fig. 4) that, in the presence of tap water (instead
of the deionized water commonly used in laboratory
experimentation), the solution conductivity and the Cr(VI) final
removal increased substantially during the Cr(VI) removal
experiments. However, the presence of other cations and anions
(Ca2+, Mg2+, HCO3-) was increased in the meantime (26).
Further results of electrolytic flotation with the paint dispersion in
the presence of various electrolytes (NaCl, NaOH, KOH) are given
in Figure 5. This figure was scheduled to be compared later with
the dissolved-air flotation experiments, where a similar volume of
gas was applied. However, as was shown afterward, there was a
difference in bubble size diameters, which presents difficulties in
this comparison. For example, one bubble with a 2-mm diameter
contains about the same amount of air as 64,000 bubbles with 50-
µm diameters.
Another laboratory-prepared sample, which was an emulsified
oilwater dispersion, has been tested (6). A machine cooling oil
(Revol oil) was used. The oil contains an emulsifier, and when it
was diluted in water, it gave a very stable white emulsion. In this
way, the effectiveness of the electrolytic flotation as a liquid/liquid
separation method, and mainly its applicability
Figure 4
Effect of initial CTMA-Br concentrations on Cr(VI) removal. Distilled
water (conductivity around 80 µmhos) and tap water (conductivity
600 µmhos) were compared. The pH value was kept constant at 8,
while the other parameters were kept the same as in Figure 2.
Page 395
Figure 5
Electrolytic flotation of a paint dispersion, for comparison with the dissolved-
air flotation tests
(see Table 3).
on emulsions was examined. In this case, much lower initial
concentrations were suitable (on the order of 100 mg L-1) for
satisfactory flotation removal. It is also worth mentioning that in
refinery waste water treatment the American Petroleum Institute
gravity separator is often used as the first separation step, its primary
function being to separate free oil from the refinery waste waters.
However, the respective design manuals cannot predict the effluent
quality after the treatment, but can provide only a separator size for
specified gravity and flow rate. Such a unit is not capable of
breaking or separating emulsions, nor is it specified for these
purposes.
Generalized approaches that previously were acceptable are no
longer adequate because of more stringent regulations that demand
more efficient gravity-type oil/water separators, in addition to further
waste water treatment downstream. It is exactly in this field that the
flotation process was expected to operate more satisfactorily, i.e., as
a posttreatment step after the preliminary gravity separator
application.
Another important parameter investigated in flotation experiments,
thus in electrolytic flotation also, is the presence of a surfactant and
its suitable concentration. Figure 2 shows the results obtained with a
common cationic surfactant (cetyl-trimethyl-ammonium-bromide),
applied for the removal of Cr(VI) ions. Nevertheless, in the
electrolytic flotation experiments with paint or oil emulsions, no
surfactant was used.
Page 396
The two electrodes of a horizontal set are often separated, for
instance by a cation exchange membrane, as only one of the
previously mentioned gases is usually necessary (7). It was noticed
that only one gas (hydrogen) activated the process, when the
emulsified oil/water emulsion was studied, whilst the other acted as
a depressor or merely contributed in the transfer of the coagulated
matter to the flotation cell surface. In the meantime, when the
separation between the gases was applied, an electricity savings
was noticed at laboratory scale. These advances could be achieved
on an industrial scale by changing the surface area of the working
electrodes, i.e., using electrodes having the shapes of a plate and a
rod.
A
Electrochemistry in the Mineral-Processing Plant
It is now well established that sulfide minerals interact with water
and with flotation surfactants largely by charge transfer
mechanisms and that every reaction of this type corresponds to a
certain electrochemical potential (44). This suggests the possibility
of controlling flotation selectivity by manipulating the potential of
the pulp, for example, by chemical methods, i.e., through the
activity of oxygen or with the addition of suitable oxidizing or
reducing agents.
The pulp potential certainly can be varied by an external power
source. This was done earlier in a modified Hallimond tube, where
the mineral particles (chalcocite) were in electrical contact with a
gold electrode (45). Recently, a laboratory flotation cell (Denver)
was modified, and when an electric potential was applied, the
flotation recovery of covellite was enhanced (46). It is noted that
the measurement of the electric current of solution under the
application of an external electric potential is known as cyclic
voltammetry. Voltammographs of sulfide mineral systems have
been used already in flotation studies (44). Eh-pH mapping of a
flotation circuit may provide additional information on the
operating regime and the relevant predominant species involved in
the process.
Applying the electric potential externally could also minimize the
application of costly reagents (such as copper sulfate) and could
have an ecological importance by excluding other highly toxic
reagents like sodium cyanide (47). Panayotov has investigated the
industrial implementation of the non-cyanide technology in depth,
while considerable increase in the recovery of noble elements (like
silver) was also reported.
In general, the electric field gradient between the electrodes aids in
the flocculation of suspended solids. Furthermore, when an electric
current is applied to a dispersion, the oxidationreduction potential
changes depending on the type of electrode treatment. Also,
nascent hydrogen and oxygen
Page 397
are highly active, and as they attack the solid surfaces, a change in
the flotation characteristics of these can be realized. Any one or
more of these effects could be operated during electrolytic
flotation. These have been described with the general term
electrochemical effect of the process, which is not found in other
flotation techniques (1). Nevertheless, there should exist a certain
differentiation between the previously mentioned electrochemical
effects and the application of gas bubbles as the transport medium
in the flotation process.
B
Electrolytic Flotation in Mineral Processing
Mineral processing has been approaching a crossroads recently.
The demand for mineral supply is continuing, while the grade of
available ores diminishes. In order to optimize the recovery of fine
mineral particles, and thus avoid loss of value, considerable interest
is growing in developing new processes and improving old ones. A
part of improving old processes is the attempt to try and test
electrolytic flotation.
Fines are produced as a result of liberation and, in other cases, as
unavoidable by-products during mining, grinding, or classification.
There are many examples of ores that consist mostly of fines (15).
Flotation, of course, is perhaps one of the most important unit
operations in mineral processing. But the application of
conventional separation methods, such as froth flotation, for fine
beneficiation and enrichment, is considered generally
uneconomical.
An application of electrolytic flotation to the recovery of dolomite
and magnesite fines was reported (37,48). As shown in Figures 6
and 7, where the effects of usual examined parameters on the
performance technique were examined (i.e., pH influence and
flotation time), the solution pH was observed not to affect flotation
recoveries in a wide range. However, the most interesting
conclusion was that the process was working collectorless.
C
Dispersed-Air Flotation
A comparison between the three main flotation techniques for
bubble generation was presented, i.e., (a) dispersed-air flotation;
(b) dissolved-air flotation; and (c) electrolytic flotation, when all
these methods were applied for Cr(VI) anion removal (49).
Elsewhere, the three technologies were also compared, giving
advantages and disadvantages in industrial waste water treatment
(50). In summary, each plant adopts a method that best suits its
specific needs. In certain applications, such as removal of algae
from aeration lagoons, the effect of electrolytic current on the algae
coagulation properties was said to be especially helpful (9).
Air flotation uses mainly air bubbles (or, seldom, nitrogen), while
elec-
Page 398
Figure 6
Effect of electrolysis time on flotation recoveries of dolomite and magnesite, without any collector.
Current density 100 A m-2, pulp density 11.1 g L-1, NaOH 3.7 × 10-3 N.
trolytic flotation releases hydrogen and oxygen bubbles. One of the first obvious
questions is whether the chemical difference is important. Another question is whether
the size of the bubbles and their ascent rate differ and whether this differentiation is
important. Answers to these questions are being examined. Dispersed-air flotation,
which resembles electrolytic flotation in that atmospheric pressures are being generally
used, has been known for a long time now, and studies of this process, termed froth
flotation, exist both for waste water treatment and (mainly) for mineral flotation.
A literature survey has shown that there is a wide application of a great variety of
chemical reagents/aids in the dispersed-air flotation experiments, such as surfactants,
frothers, and modifiers (51). As any interference from chemicals was not wanted, only
the use of air bubbles (different gas flow rates) has been tested in these experiments of
the paint dispersion treatment, as shown in Figure 8. The main scope of this work was
to obtain data using the same flotation cell as in electrolytic flotation and to keep
unchanged the type of treated solution. These data could be used in order to relate and
compare the two processes. It is easily observed that by increasing the applied gas flow
rate, the necessary flotation time for obtaining high paint removal is greatly decreased.
Page 399
Figure 7
Influence of solution pH on electrolytic flotation of magnesitedolomite.
Flotation time 600 s, current density 100 A m-2, pulp density 11.1 g L-1,
Acintol FA-1 (fatty acid mixture) 20 mg L-1.
The porosity of the diffuser, made in this case from sintered
carborundum, was 150250 µm, while the air coming from a blower
had a pressure of about 10.8 kN m-2. When a second diffuser with
coarser pores was tested, it was found less effective. In another
experimental series using continuous flow and treating the same
paint dispersion as in the previously mentioned electrolytic
flotation experiments, a concentration decrease from 1645 down to
85 mg L-1 was obtained, which corresponds to a reduction of
94.8%. The foam layer, formed at the surface of the dispersion
after the phase separation, was observed to be in a less dry
condition than in electrolytic flotation, having 25- to 50-mm
thickness, but very stable. Foam heights over 1 m were noticed
with higher air rates.
On other occasions, the bubbles produced by dispersed-air flotation
were found just to mix the waste water system and not to treat it.
When this process was tried on the emulsified oilwater dispersion,
as well as on another sample from a refinery's interceptor, it was
found ineffective without a previous flocculation step. Perhaps a
possible cause for these results was that, under the dispersed-air
flotation conditions, bubbles were found to be insufficiently finely
divided. As the air flow rate was quite high, the
Page 400
Figure 8
Dispersed-
air flotation experiments of the same paint dispersion for comparison with electrolytic
flotation (Fig. 2), at three different air flow rates.
bubbles were producing a high degree of turbulence in the suspension,
which could be a disadvantage of this process.
Bubble measurements, done on a pilot plant scale, showed a linear increase
of bubble size from 1.7 to 2.0 mm with air flow rates from 6 to 15 L min-1,
respectively. A diffuser of the same type and porosity, having 172-mm
diameter, was used in this case. The increase of bubble size with depth at a
given flow rate was found to be less than in the previous case (52).
D
Dissolved-Air Flotation
The basic principles of dissolved-air flotation, which nowadays is the
dominant flotation technique in water and waste water treatment, were
introduced elsewhere (15). It is well known that the solubility of a gas (like
air) is mainly a function of its pressure and temperature, but also (to a lesser
extent) of the presence of dissolved solids. The following experiments,
which are presented in Table 3, had only the aim of comparison with
electrolytic flotation (see also Fig. 5).
The constructed setup for the following experimental series consisted of a
feed tank, a flotation tank, and a pressurized tank; the flotation chamber had
the same dimensions as before. The pressurization tank (0.4 m in
Page 401
height and 0.27 m in diameter) was filled with water (for the
laboratory experiments) from a port on the top, where also a purge
valve and a manometer were fitted. The compressed-air inlet was
connected to a pipe near the bottom of the tank, which had small
perforations (on the top side), and it was used as a distributor.
There was also a bypass in the air inlet, connected to the top of the
tank, in order to keep the pressure constant when emptying the
pressurized tank. Around one-third of the volume of the tank was
filled with a packing, composed of glass spheres, in order to
increase the air/water contact. Finally, a 12.7-mm needle valve, at
the bottom of the flotation tank, was used for outlet of the
pressurized air/water mixture, causing a pressure release and thus
creating the tiny precipitating bubbles.
The usual experimental practice after the pressurization tank was
filled was to leave it for stabilization/solubilization of air in water
for about 1 h, having the air inlet on, while the purge valve was
slightly open; therefore, the pressure was kept at 290304 kN m-2.
That was the available pressure from the compressed-air line, and it
is considered near the low limit pressure, which is usually applied
in dissolved-air flotation applications (350550 kN m-2). A special
device was also constructed in order to measure the air volume
released from the saturation tank (52). Theoretically, it was found
that the precipitated-out air (due to pressure release) was about
72% of the calculated volume. Curves for the latter can be found in
the literature (e.g., Ref. 53). From the research experience gained
in the laboratory, it can be suggested that the pressure-reducing
device was an important parameter. The tendency for bubble
coalescence should be minimized, because this may prevent their
attachment to the suspended solids.
The procedure followed can be described as a semibatch process
and had the characteristic of a free rising surface, a phenomenon
not occurring in continuous flow. In practice, the saturator feed is
continuously blended with the influent waste water. A tangential
inlet with concentric piping for good mixing is often used (53). In
this chapter an interesting application of dissolved-air flotation to
activated sludge thickening was presented. Sometimes, also,
nozzles (or fine openings) are used at the pressure release system in
order to produce finer bubbles. However, these may be blocked by
solids, decreasing the air/water contact.
It was observed that in some tests only partial separation was
achieved. Quite often no floated layer was formed due to
turbulence in the vessel. Meanwhile, settling of the particles was
apparent. Therefore, a detergent surfactant was used to help the
formation of a more stable froth layer. Frothers or more generally
surface active agents reduce the surface tension, and are frequently
used in froth flotation. The velocity of the rising surface, termed
rise rate, was approximately 1 mm s-1.
In these experiments a number of different coagulants/flocculants
were
Page 402
Table 4 Dissolved-Air Flotation Results.
Total Ratio air-
C(initial) volume water C(final)
No. (mg L-1) Additive (L) effluent (mg L-1) Re%
1 1000 1% 10 2/1 330 67
NaCl
2 1000 0.1% 15 2/1 203 79.7
KOH
3 1000 0.1% 10 7/3 235 76.5
NaOH
4 2000 0.1% 10 1/1 305 84.8
NaOH
5 1000 0.1% 15 2/1 32 96.8
KOH
6 1000 0.1% 15 2/1 90 91
NaOH
subsequently used, such as ferric chloride, alum, and a cationic

polyelectrolyte. However, it was not the aim of this work to study
specifically the flocculation phenomena taking place. Nevertheless,
in some cases, flocculation was so effective that there was not a
real need for application of flotation as the next separation stage of
treatment. The emulsified oil droplets, especially after the emulsion
broke down, were easily settled out and separated from the rest of
the waste water. Due to their density, which was lower than that of
the water, settling was antagonistic with flotation in this case.
It could be concluded that the results in Table 4 were quite
promising. For instance, with an initial concentration of paint of
1000 mg L-1, a total volume of 15 L (after the addition of 0.1%
NaOH or KOH and of a frother), and a ratio of air-water/effluent
2:1, the final concentrations were, respectively, 90 and 32 mg L-1.
These concentration values have been analytically detected, whilst
further dilution from the air-charged water has not been taken into
account. However, it is noted that the applied recycle ratio is rather
unrealistic for a real industrial case, usually being in this case on
the order of only 1020%.
It was initially thought that a comparison could be feasible between
dissolved-air and electrolytic flotation. However, a difference in the
respective bubble size was noticed, as shown in Figure 9. The data
obtained were plotted on normal probability paper, in rather good
agreement with the Gaussian distribution (see also Ref. 18).
Average sizes of as small as 5 µm and as large as 200 µm were
reported in different papers in the literature dealing with
electrolytic flotation (6). Surprisingly, in our case the bubble size
decreased with increased current densities, and at the same time the
number of bubbles increased (8). The direct measurement method
used here has been described already elsewhere (54). Electrolytic
flotation bubbles (both hydrogen and oxygen) had a mean diameter
of around 50 µm, while those precipitated out from dissolved-air
Page 403
Figure 9
Bubble photographic measurements in flotation: electrolytic (100 A m-2) and dissolved-
air flotation.
flotation had a mean diameter of 287 µm. The latter was not in agreement
with the literature (55), probably due to coalescence of bubbles after their exit
from the needle valve.
Many investigators have also proposed that the flotation separation of fine
particles will be improved by applying finer bubbles. The idea of a critical
diameter of a bubble, which is just buoyant when fully covered with mineral
particles, also has been introduced (56). If fine bubbles will be introduced into
a very dilute slurry, as this actually happens in dissolved-air flotation, these
bubbles will become covered rapidly by particles, in spite of the high initial
collision efficiency between fine bubbles and fine particles, simply because of
the deficiency of particles to be attached. This situation will allow bubbles to
coalesce more rapidly because of the high concentration of bubbles and the
long residence time. As a result, the bubbles will grow in order to exceed the
critical size and the effect of the use of fine bubbles will be diminished
significantly by excessive coalescence.
E
Design Aspects
The knowledge of the rate at which suspended matter will be
separated/floated in a flotation unit is very important. Also, an operator
should be able to identify the main variables that control this removal rate
(57). The
Page 404
Figure 10
First-order kinetics of electrolytic flotation.
understanding of the process can lead to a substantial improvement
in the electrolytic flotation design (58). A macroscopic approach
was followed in the kinetics study. This usually consists of the
analysis of suspended matter concentrations as a function of time.
It is pointed out also that the conditions governing the entire period
of the batch process were not constant.
Figure 10 presents some of the results of the kinetics study in a
semilogarithmic plot. Three different current densities were applied
(100, 200, and 300 A m-2), while the initial paint concentrations
were 500 or 1000 mg L-1. Photographs for some of the electrode
types used in these experiments were given recently (59). The use
of flotation cells with consumable anodes is also of interest, since
their slow dissolution provides a continuing injection of metal
hydrolyzable cations into the system [usually Fe(III) or Al(III)],
which then subsequently aids flocculation (see also Ref. 11).
It was observed (Fig. 10) that for most of the experimental time,
except perhaps the initial period, the first-order equation (as known
from chemical reaction kinetics) could satisfactorily describe the
process, i.e., ln(C/C0) = -kt, where k denotes the flotation rate
coefficient, C is the concentration at time t = t, and C0 at time t = 0.
Flotation kinetics were previously discussed (60).
In batch tests the ratio dC/dt is a measure of the flotation rate and
will be equal to the rate of clarification (in effluent treatment) if the
amount of removed constituents (e.g., suspended solids or oil) that
returns to the
Page 405
solution from the floated layer is negligible. In mineral processing
attention should be paid to the fact that flotation is applied mostly
as a selective separation process and that only certain mineral
particles (either useful or gangue) have to be transferred toward the
slurry surface.
If the same equation is applied to continuous flow experiments,
where at the steady state dC/dt = 0, the aforementioned equation, as
it stands and is used in Figure 10, breaks down. Batchwise tests can
be used to gain preliminary information on the flotation
characteristics of a system, while continuous flow experiments may
give the final design data.
Another observation is that a deviation was apparent at the start of
batch electrolytic flotation experiments, where a retardation
appeared. This fact was expected, however, to be overcome during
continuous flow. This period, which was explained as the time
needed for destabilization of the treated dispersion, was affected by
the current density and the added chemical or other flocculent
concentrations. As is shown rather clearly in Figure 11, the
application of Zetag 94, a common cationic polyelectrolyte (kindly
supplied by Allied Colloids) reduced the necessary retention time,
producing also cleaner effluent.
These experiments took place at a larger scale. The scale-up was on
the order 1:21.8, where the relative calculations were based on the
different
Figure 11
Electrolytic flotation experiments on a larger scale. Use of a
cationic polyelectrolyte (10 mg L-1) at current density 100 A m-2.
Page 406
tank surfaces. In the larger scale the same kinetic pattern was
followed, although with increased retention times (61).
The principles of scale-up in chemical technology are generally
rather well established. This knowledge would not only benefit
operation and prediction, but also electrolytic flotation plant design
placing it on a more concrete basis, and thus reducing the excessive
reliance on costly trial-and-error development of new and
improved units (1).
Experiments also were carried out scaling up the tank depth
(height). In Figure 12 some of the results for oil/water emulsion
treatment are presented, using electrolytic flotation. The flotation
rate was calculated from the respective chemical kinetics equation
as a function of retention time, enabling the flotation rate to be
presented as shown in Figure 10. It was observed that the flotation
rate reaches a maximum, which increases with higher feed
concentrations. A constant maximum rate of solid removal was
reached at any given current density (8). There was, however, a
limited concentration of the dispersion system to be treated in the
flotation cell, above which the unit was not actually in operation,
i.e., the generated bubbles were only helping to mix the liquid.
An investigation of hydrodynamics was then undertaken, as the
efficiency is expected to be a function of the residence time
distribution (62). The stimulusresponse technique was used in this
experimentation. The
Figure 12
Variations of the electrolytic flotation rate in oil/water
emulsion treatment. Batch tests at current density 100 A m-2.
Page 407
tracer input signal was an electrolyte and had the form of a step
function. Assuming that the flow regime was composed of various
flow types, a theoretical multiparameter model was applied, and
the experimental data of electrolytic flotation were fitted
accordingly (63).
A flotation unit is fundamentally a form of reaction vessel, with a
separation occurring instead of a chemical reaction. Consequently,
a detailed knowledge of the flow characteristics involved in a
particular tank may be expected to comprise an important feature
of its design (60). The advantages of different operation modes in
continuous flow have also been discussed. It was found that the
plug flow region was about the same (around 12.4%) for both
operation modes, counter-current and concurrentarranged by
varying the inlet port of the feed near the top of the tank or at the
bottom near the electrodes, respectively. As might be expected,
differences were noticed in the mixed and dead-water regions,
while no bypass was found. In the counter-current flow, there was
more backmix flow (around 87.7%), while in the concurrent
operation, a dead water region appeared (around 6.7%) (62).
A simple equation was subsequently used to describe the process in
continuous operation, based on the assumption of a first-order
kinetics equation (2). Under steady-state conditions, a mass balance
was computed around the electrolytic flotation vessel, which gave
promising results in comparison with the experimental data (Figs.
13 and 14).
The mass of effluent per unit of time (calculated as the product of
flow rate by the concentration at the respective sampling point) is
going into the tank, minus the mass per unit of time going out of
the tank constituted the left-hand side of the equation applied. The
product of the flotation rate coefficient and the concentration in the
tank, in the applied operating conditions, gave the flotation rate for
a differential volume. Hence, it was multiplied by the tank volume
(51).
The inherent assumption in the mass balance equation was that the
concentration was independent of the distance, a principle that
holds for a perfectly mixed vessel. A regression analysis of these
equations for the two different dispersion applications gave
significant results. Therefore, within the range investigated, these
assumptions were quite well justified.
The significance of this work is that the effluent concentration from
the electrolytic flotation unit could be foreseen with reasonable
approximation as a function of the feed concentration entering the
unit, if the other operating parameters (i.e., flow rate, current
density) are kept stable. Finally, if the outlet concentration is
expressed or known (for instance, the limit imposed by the Water
Control Authorities), it is then possible for the optimum flow rate
to be predicted by calculation. Therefore, the applied model
represented a simple and reasonable tool in flotation treatment.
Page 408
Figure 13
Comparison of the proposed model in continuous mode. Electrolytic flotation rate against
effluent concentration
(at current density 100 A m-2).
Figure 14
Experimental versus calculated concentrations for the paint dispersions
from the mass balance equation.
Page 409
IV
Conclusions
Electrolytic flotation offers certain advantages over other flotation
techniques of bubble generation (dispersed-air, dissolved-air).
Among others, the following points are perhaps the most
important.
1. The electrode grids can be arranged to provide good coverage of
the whole surface area of the flotation tank, avoiding any bypass.
2. The electric field gradient between electrodes aids flocculation
and flotation, even without the presence of any surfactant.
3. Gas production (evolution), flotation time, and the other
operating conditions can be checked quickly and are relatively
easily controlled.
4. The equipment is reliable and safe in operation.
5. Ionic strength of the effluent solution (slurry) to be treated is not
a crucial parameter, as in the other flotation techniques.
6. By-product recovery is a potentially profitable aspect of
electrolytic flotation treatment, so proteins, fats, oil, organics, but
also toxic or valuable metals could be recovered from waste
streams, especially when no other chemical reagents (like
surfactants) are added to help the pollutant separation.
References
1. K. A. Matis, Electrolytic flotation in industrial effluent
treatment, Chim. Chron. (New Ser.), 11: 159 (1982).
2. K. A. Matis, ''A study of electrolytic flotation and its application
to the treatment of industrial liquid wastes," Ph.D. Thesis,
University of Newcastle-upon-Tyne (1977).
3. K. A. Matis, Electroflotation and its application to the treatment
of industrial effluents, Tech. Chron. (Chem. Eng.), 4: 19. (1977).
(In Greek with extended English summary.)
4. N. C. Lockhart, Electrical separationsa colloid perspective, Filtr.
Sep., 15: 211 (1978).
5. A. A. Mamakov and A. I. Avvakumov, Flotation of minerals by
electrochemical reactions, Elektron. Obrab. Mater. (translated as
Appl. Electr. Phenomena), 5: 54 (1968).
6. K. A. Matis and J. R. Backhurst, Laboratory studies of
electrolytic flotation as a separation technique, SolidLiquid
Separation (J. Gregory, Ed.), Ellis Horwood, London (1984), p. 29.
7. K. A. Matis, Treatment of industrial liquid wastes by
electroflotation, Water Poll. Control, 79: 136 (1980).
8. J. R. Backhurst and K. A. Matis, Electrolytic flotation in effluent
treatment, J. Chem. Technol. Biotechnol., 31: 431 (1981).
9. E. Sandbank, G. Shelef, and A. M. Wachs, Improved
electroflotation for the
Page 410
removal of suspended solids from algal pond effluents, Water
Res., 8: 587 (1974).
10. M. Palak, J. Sikora, and J. Zielinski, Investigations on
combined electrocoagulation and electroflotation application to
removal of phosphates from waste water, Environ. Protection Eng.,
8: 15 (1982).
11. P. A. Treille and L. C. Willners, "A New Process for Treating
Industrial Waste Waters: Electro-Flocculation," Proc. Inter.
Scandinavian Congr. Chem. Eng., Copenhagen (1974), p. 23.
12. C. P. C. Poon, "Removal of Copper, Nickel, Zinc, Cadmium
and Cyanide from Plating Waste Water by Electroflotation,"
Proceedings Conference Management and Control of Heavy
Metals in the Environment, Commission EEC, London (1979), p.
572.
13. Chemical Engineering, Electrolysis-bed bubbles clarify water,
Chem. Eng., 75: 82 (1968).
14. L. M. Balmer and A. W. Foulds, Separating oil from oil-in-
water emulsions by electrocoagulation/electroflotation, Filtr. Sep.,
23: 366 (1986).
W. Barley, Eds.), Martinus Nijhoff, Dordrecht (1986), p. 37.
16. R. Mallikajunan and S. Venkatachalam, ElectroflotationA
review, Electrochemistry in Mineral and Metal Processing,
Electrochem. Soc., Pennington (1984), p. 233.
17. V. P. Shredov and M. F. Yakushev, The use of electrophoresis,
electrocoagulation and electroflotation for the purification of
radioactive wastes, Sov. J. Radiochem. (transl. of Radiokhimiya),
12: 840 (1970).
18. J. Mayer, L. Zhang, and H. H. Hahn, Liquidsolid separation by
electroflotation: An attractive alternative to dissolved-air flotation,
Chemical, Water and Wastewater Treatment (H. H. Hahn and R.
Klute, Eds.), Springer-Verlag, Berlin (1990), p. 151.
19. C. Camilleri, Electroflottation et flotation a l' air dissous, Ind.
Miner., Supp. Tech., 67: 25 (1985).
20. E. R. Ramirez, D. L. Johnson, and A. Clemens, "Direct
Comparison in Physico-Chemical Treatment of Packing House
Waste Water Between Dissolved-Air and Electroflotation,"
Proceedings 31st Industrial Waste Conference, Ann Arbor, MI
(1976), p. 563.
21. T. F. Zabel, Flotation in water treatment, Innovations in
Flotation Technology (P. Mavros and K. A. Matis, Eds.), Kluwer
Academic, Dordrecht (1992), p. 431.
22. H. H. Hahn, Waste water treatment, The Scientific Basis of
Flotation (K. J. Ives, Ed.), Martinus Nijhoff, The Hague (1984), p.
379.
23. Y. Fukui and S. Yuu, Removal of colloidal particles in
electroflotation, AIChE J., 31: 201 (1985).
24. Y. Fukui and S. Yuu, Collection of submicron particles in
electroflotation, Chem. Eng. Sci., 35: 1097 (1980).
25. B. Dobias, Koagulation und Flotation von feinen Dispersionen,
Tenside Detergents, 17: 68 (1980).
Page 411
26. A. I. Zouboulis and K. A. Matis, Electrolytic flotation of
chromium from dilute solutions, Environ. Technol. Lett., 10: 601
(1989).
27. V. P. Nebera, V. I. Zelentsov, and K. A. Kiselev,
Electroflotation of ions from multicomponent systems, Fine
Particles Processing (P. Somasundaran, Ed.), SME/AIME, New
York (1980), p. 886.
28. E. R. Ramirez, "Electroflotation of Group VI-A Heavy Metals
from Mining Tailings Waste Water," Proceedings 35th Industrial
Waste Conference, Purdue University (1980), p. 242.
29. Mitsui, Mining and Smelting Co., Elimination of heavy metals
in waste water by electrolytic flotation, Tech. Inf. Bull., Japan
(1989).
30. H. P. Roth and P. V. Ferguson, An integrated industrial waste
water treatment system using electroflotation and reverse osmosis,
Desalination, 23: 49 (1977).
31. P. Costaz, J. Miquel, and M. Reinbold, Simultaneous
electroflotation and disinfection of sewage, Water Res., 17: 3
(1983).
32. C. C. Ho and C. Y. Chan, The application of lead dioxide-
coated titanium anode in the electroflotation of palm oil mill
effluent, Water Res., 20: 1523 (1986).
33. S. E. Hrudey, The management of waste water from the meat
and poultry products industry, Survey in Industrial Wastewater
Treatment, Vol. I: Food and Allied Industries (D. Barnes, C. F.
Forster, and S. E. Hrudey, Eds.), Pittman, Boston (1984), p. 169.
34. F. G. Oblinger, M. Weintraub, E. J. Blais, and R. L. Gealer,
Automation of an electrolytic cell for the treatment of oily waste
water, Environ. Progress, 5: 283 (1986).
35. G. R. Kramer, A. Buyers, and B. Brownlee, "Electrolytic
Treatment of Oily Waste Water," Proceedings 34th Industrial Waste
Conference, Purdue University (1980), p. 673.
36. E. Baer, Einsatz der Electroflotation bei der Behandlung von
Abwassern mit emulgierten und dispergierten Inhaltsstoffen,
Handbuch der Wasserversorgungsund Abwassertechnik, 2.
Ausgabe, Vulkan-Verlag, Essen (1990), p. 488.
Chron. (New Ser.), 17: 29 (1988).
38. C. Manohar, V. K. Kelkar, and J. V. Yakhmi, Electroflotation of
colloids without surfactants, J. Colloid Interface Sci., 89: 54
(1982).
39. H. Ledesma and M. Guzman, Electroflotation of enargite, Froth
FlotationDevelopments in Mineral Processing (S. H. Castro and J.
Alvarez, Eds.), Elsevier, Amsterdam (1988), p. 355.
40. G. B. Raju and P. R. Khangaonkar, Electroflotation of
chalcopyrite fines, Int. J. Miner. Process., 9: 133 (1982).
41. D. R. Ketkar, R. Mallikarjuman, and S. Venkatachalam,
Electroflotation of quartz fines, Int. J. Miner. Process., 31: 127
(1991).
42. Institute of Chemical Engineers, Application of Chemical
Engineering to the Treatment of Sewage and Industrial Liquid
Effluents, Symp. Ser. 41, York, UK(1975).
Page 412
43. K. A. Matis, "A study of flotation techniques in effluent
treatment," M.Sc. Thesis, University of Newcastle-upon-Tyne
(1975).
44. S. R. Rao, G. Labonte, and J. A. Finch, Electrochemistry in the
Matis, Eds.), Kluwer Academic, Dordrecht (1992), p. 57.
45. S. Chander and D. W. Fuerstenau, Effect of potential on the
flotation and wetting behavior of chalcocite and copper, Trans.
SME/AIME, 258: 284 (1975).
46. J. R. Roos, J. P. Celis, and A. S. Sudarsono, Electrochemical
control of chalcocite and covelliteXanthate flotation, Int. J. Miner.
Process., 29: 17 (1990).
47. V. Panayotov, "Design of a Non-Cyanide Technology for
Floating Lead-Zinc Ores: Energy Prerequisites, Implementation,
Results," Proceedings 1st European Metals Conference, IMM,
Brussels, preprint (1991).
48. K. A. Matis, G. A. Stalidis, and G. P. Gallios, Beneficiation of
fine mineral particles by electrolytic flotation, Tech. Chron. (Chem.
Eng.), 5: 35 (1985). (Greek with extended English summary.)
49. A. I. Zouboulis and K. A. Matis, Ionenflotation von Chrom,
Fortschritte in Theorie und Praxis der Aufbereitungstechnik (K.
Uhlig, Ed.), FIA, Freiberg (1989), p. 418.
50. E. R. Ramirez, "Comparative Physicochemical Study of Waste
Water Treatment by Electrolytic, Dispersed-Air and Dissolved-Air
Flotation Technologies," Proceedings 34th Industrial Waste
51. K. A. Matis and P. Mavros, Foam/froth flotation: Part II.
Removal of particulate matter, Sep. Purif. Methods, 20: 163 (1991).
52. K. A. Matis, Flotation and its application in industrial waste
treatment, Chim. Chron. (Gen. Ed.), 46: 27 (1981). (In Greek:
Chem. Abstr., 95: 208948j.)
53. J. Bratby and G. v. R. Marais, Flotation, Solid/Liquid
Separation Equipment Scale-up (D. B. Purchas, Ed.), Uplands
Press, Croydon (1977), p. 155.
54. K. A. Matis, Bubble measurements in electrolytic flotation,
55. Vrablik, E. R., "Fundamental Principles of Dissolved-Air
Flotation of Industrial Wastes," Proceedings 14th Industrial Waste
56. M. Szatkowski and W. L. Freyberger, Kinetics of flotation with
fine bubbles, Trans. IMM, 94: C61 (1985).
57. F. Barrett, ElectroflotationDevelopment and application, Water
Poll. Control, 59 (1975).
58. Y. Fukui and S. Yun, Development of apparatus for electrolytic
flotation, Chem. Eng. Sci., 39: 939 (1984).
59. K. A. Matis, Electrolytic flotation, Innovations in Flotation
Technology (P. Mavros and K. A. Matis, Eds.), Kluwer Academic,
Dordrecht (1992), p. 301.
60. G. P. Gallios and K. A. Matis, "Flotation Applied in Waste
Water Engineering," Proceedings 2nd International Symposium,
MESAEP, Heraklion, Creta (1983), p. 427.
Page 413
61. G. P. Gallios and K. A. Matis, "Flotation Techniques in
Environmental Technology," Proceedings 7th National Chemical
Conference, Yiannina (1982), p. 90. (In Greek.)
62. K. A. Matis, Nonideal flow and its application to flotation,
63. O. Levenspiel, Chemical Reaction Engineering, Wiley, New
York (1972), p. 253.
Page 415
15
Flocculation in SolidLiquid Separation Processes
Brij M. Moudgil and Sanjay Behl*
University of Florida, Gainesville, Florida
I
Introduction
Solidliquid separation is an important unit operation in the mining
and chemical industries. The increase in the awareness of the
effects of industrial wastes on the environment has further
enhanced the need for improving the solidliquid separation
efficiency in different process industries. In most cases where such
separation is desired for fine particles, conventional separation
techniques such as settling, forced settling, and filtration are not
effective. Flocculation of the suspension followed by separation by
a conventional method has been proposed as an alternative strategy.
Flocculation essentially involves aggregating fine particles using
synthetic polymeric materials. This is of course different from the
aggregation with salts and organic molecules such as starches and
gums, commonly termed coagulation (although there is no
universal agreement on this terminology).
Flocculation as a process for solidliquid separation is used in
different applications and the characteristics of the floes
(aggregates) required in each case are unique. The mechanisms
involved in flocculation are not understood well enough at present
to obtain flocs of desired properties. The major aim of most of the
current research in flocculation is twofold: first to understand the
mechanisms involved during the aggregation process and
*Present affiliation: Technical Center, Engelhard Corporation,
Gordon, Georgia
Page 416
Table 1 Nature of the Flocculant.
Ionic nature Examples
of flocculant
Non-ionic Polyvinyl alcohol, polyethylene oxide,
polyacrylamide
Anionic Anionic polyacrylamide, polyacrylic acid
(high pH), sodium polystyrene sulfonate
Cationic Polyethyleneimine, polydialldimethyl-
ammonium chloride
second to deal with correlation of the flocculant properties with the

process efficiency. This chapter deals with the basic concepts of
flocculation with the aim of presenting guidelines to design and use
flocculants for solidliquid separations efficiently.
II
Nature of the Flocculants
Early examples of polymeric flocculants were based on natural
products such as gelatin, alginates, and starches. Many of these are
still used in applications such as clarification of beer and wine.
However, most of the flocculants used in the mineral and chemical
industries these days are synthetic in nature. These are broadly
classified in three major categories. (See Table 1.)
Flocculants are usually linear polymers, which may be ionic or
non-ionic in nature. Some branched polymers such as starches also
act as effective flocculants. The most important characteristics of a
flocculant are its monomeric unit, molecular weight, and charge
density. Unfortunately, most of the polymer-induced flocculation
work in the past was done with poorly characterized materials of
commercial origin and therefore meaningful interpretation of the
data is difficult. For example, the quoted molecular weights (which
may range from a few thousand to several million) used in a
process are perhaps average values where the higher molecular
weight fraction of the polymer acts as a flocculant and the lower
molecular weight fraction as a dispersant. This may have been the
reason for several discrepancies observed in the past from the
theoretically expected flocculation trends.
III
Separation Processes and Floc Properties
The aim of a solidliquid separation process may be to obtain either
a clear supernatant (e.g., potable water) or a dense mass of solids
(such as in dewatering of clays). In either case, flocculation is a
method of aggregating
Page 417
the suspended fine particles into larger agglomerates followed by
another step such as sedimentation, filtration, or floc flotation to
separate the flocculated material from the suspension. Floc
properties desired in a given application are governed more by the
separation step. For instance, in filtration, strong, porous, and less
dense flocs are desired, whereas in sedimentation dense, large flocs
with minimum porosity are required. In cases where consolidation
of the solids is the overall objective (e.g., in dewatering of clays),
density of the flocs would be of major importance since there is a
direct correlation between floc densities and sludge (sediment)
densities (1). A summary of desired floc characteristics for specific
applications is presented in Table 2. Appropriate flocculant
selection, therefore, should take into account the solidliquid
separation method to be employed. It is to be noted that
measurements of size and density of the flocs are difficult to
classify or measure independently.
A
Criteria for Evaluating Flocculants
Different parameters such as turbidity, amount of flocculated mass,
settling rate, sediment volume, filtration rate, and floc size and size
distribution are used for evaluating the efficiency (both rate and
degree of flocculation) and the optimum dosage of a given
flocculant. Each of these parameters measures a different
characteristic of the floc, and tests must be designed to be
consistent with the intended flocculant application. For instance, in
a study of the Na-kaolinite and polyacrylamide system (2), the
optimum polymer concentration for the suspension was determined
to be 25 mg/kg by settling rate, while the sediment volume
measurement yielded a value above 30 mg/kg (see Fig. 1). These
results clearly indicate the danger of reaching conclusions with
regard to the optimum dosage based on only one measurement. The
turbidimetric method for monitoring flocculation has been used by
a number of researchers (35). An approach utilizing optical
parameters such as transmittance for evaluating flocculation using
activity diagrams and contour diagrams has also been proposed (6).
These techniques, however, are not sensitive to the structure of the
sedimented flocs, an important
Table 2 Floc Characteristics for Specific
Applications.
Method of Desired floc properties
separation
Filtration Porous, strong, permeable flocs
Sedimentation Dense, strong, large, regular
shape
Centrifugation Strong, dense, large flocs
Floc flotation Low-density, strong, narrow size
distribution
Page 418
Figure 1
Adsorption and flocculation response as a function of polyacrylamide dosage.
(From Ref. 2.)
aspect in a number of applications. La Mer and Healy (7) have
pointed out that filtration rate measurements characterize flocculated
systems better since they are influenced by both floc properties and
unflocculated fines.
B
Floc Characterization
Appropriate techniques must be used to measure the characteristics
of flocs so that meaningful correlations can be established between
the experimental parameters being examined and the results obtained
(8). Since flocs are not homogeneous with respect to size or
structure, only averages or distributions are measured. Measurement
of floc size has been attempted with the aid of photography (910),
optical micrographs, or SEM micrographs of freeze-dried flocs on
electron microscope grids (13). More recently, Koh et al. (12) have
reported measuring floc size distribution of scheelite using a Leitz
Photosedimentometer, a method first proposed by Harner and
Musgrave (13). This method has the advantage of not breaking the
flocs during
Page 419
analysis. Moreover, it incorporates the porosity of the flocs in the
measured size, obviating the need for independent porosity
measurements. Floc densities have been estimated using direct
measurements of the size and free settling velocities (1). Floc
structure has been studied by stereological analysis of SEM and
TEM micrographs of freeze-etched and freeze-fractured flocs
(11,14). Fractal dimension, a number that relates the density of the
floc with the size has also been used effectively to assess the floc
characteristics (15).
IV
Factors Affecting Floc Properties
The nature of flocs obtained using a flocculant depends on many
factors including the nature of the solid (e.g., surface chemistry,
surface energy, size, size distribution, shape, and density), the
liquid (e.g., viscosity and dielectric constant), the suspension (e.g.,
solid loading, pH, ionic strength, and temperature), and the
flocculant (e.g., chemical nature of backbone and side chains,
molecular weight, molecular weight distribution, and charge
density). Extensive research has been conducted on the stability of
colloids and modes of their aggregation using inorganic salts and
polymer molecules. However, most of the studies are related to
specific systems and do not postulate general principles for
selecting an appropriate flocculant for a specific application. Few
systematic studies correlating floc structure and different variables
have been reported in the literature. Even when the major factors
can be controlled, flocs of desired qualities may not always be
produced.
Under given conditions, one of the major factors that influence the
floc structure is the mechanism by which the flocculation takes
place. It should be noted that synthesis of a polymer to follow a
particular flocculation mechanism would allow one to control the
ultimate properties of the flocs. The major features of these
mechanisms with specific reference to polymer properties are
discussed next.
A
Polymer Properties
1
Flocculation Mechanisms
Bridging
The bridging mechanism is associated with polymers of very high
molecular weights wherein the loops and tails of the
macromolecules can form links between particles (see Fig. 2)
(16,17). In this mechanism, the electrostatic charge repulsion
between the particles is overcome since the polymeric chain can
span over the electrical double layers of the approaching particles.
The polymer itself may be charged or uncharged, but its molecular
weight has the major influence on the flocculation process. The
Page 420
Figure 2
Schematic of bridging flocculation.
higher the molecular weight, the more efficient is the polymer in
flocculating a given suspension. It must be realized that the effect
of molecular weight cannot be compensated by dosage. In other
words, using a high dosage of a low molecular weight polymer
would not yield the same results as low dosage of a high molecular
weight polymer. This was clearly established in a study of a
polyacrylamide-type polymer and silica system (18)
Figure 3
Optimum polymer concentration for various
samples of varying molecular weight.
(From Ref. 18. Used with permission of Society
for Mining, Metallurgy, and Exploration, Inc.)
Page 421
illustrated in Figure 3. A mathematical explanation of this
observation is provided by Behl (19). It must be pointed out that
the maximum flocculation takes place at half the surface coverage.
However, for every molecular weight, the total polymer required
for half surface coverage is different. In fact, in certain very high
molecular weight flocculants, half the surface coverage might not
be physically achievable in the small time frames involved in the
flocculation experiments.
Charge Patch Neutralization
The floes formed in this case are similar to those formed with
inorganic coagulants. Ionic polymers of low to moderate molecular
weight but high charge density have been reported to form flocs by
the ''charge-neutralization" mechanisms (20,21). In this case the
charged polymer, visualized to be much smaller than the surface
area of the particle, adsorbs on the charged particle and forms a
patch, not only neutralizing locally the charge on the particle but
inducing a charge of the opposite nature due to its own high charge
density (see Fig. 4). This results in a particle with patches of the
opposite charge. Upon collision of particles, charge interaction
between oppositely charged patch-bare surface leads to floc
formation. For flocculation to occur by this mechanism, the
polymer and the particles must be charged. Polymers with charges
opposite to those on the particle can be expected to be more
strongly adsorbed than others and to lead to higher settling rates
and superior supernatant clarities (18).
Network Model
It is reported that certain polymers, polyethylene oxide in
particular, form a network of cross-linked polymer chains. This
could occur by ions or polymers through electrostatic interactions
or hydrogen bonding. The dispersed material is enmeshed in the
polymer network formed (see Fig. 5). Not much is known about the
structure of the flocs formed by this mechanism, although it is
reported that in some cases the
Figure 4
Schematic of charge-patch neutralization mechanism of
flocculation.
Page 422
Figure 5
Schematic of network flocculation.
flocs formed are not very strong (22). In phosphatic clay
flocculation with PEO, however, flocs could be compressed to
release water without disintegrating the flocculated mass,
suggesting formation of high-strength aggregates (23). Another
interesting variation of network flocculation is pelleting
flocculation. In this case, hydrous oxide precipitates of aluminum
and iron salts in association with small amounts of high molecular
weight anionic polymer under controlled hydrodynamic conditions
produce very strong flocs (24). In another study of flocculation of
palygorskite with polyethylene oxide, it was found that the flocs
formed were "stringy," and a "network" formation of the polymer
was reported to be partly responsible for this observation (14).
2
Mode of Polymer Addition and Dosage
The amount and mode of polymer addition are important factors
governing the structure of the aggregates formed. It is well known
that an optimum polymer dosage exists for flocculation and that
less than the optimum dosage would not yield complete
flocculation, whereas excess polymer addition would lead to steric
stabilization. The method of separating flocs from the suspension
also influences the polymer dosage to be used. Preliminary
evidence from a study by Healy (25) indicates that filtration is
more sensitive to degree (completeness) of flocculation than
settling; therefore, usually higher polymer dosages are required for
filtration than settling to ensure efficient solidliquid separation.
Polymer dosages recommended by commercial flocculant
manufacturers are usually higher for filtration than for thickening,
lending further support to this observation.
Page 423
Besides the amount of polymer added, flocculation is also affected
by the mode of polymer addition. For example, in a study of a
palygorskitePEO system (14), the optimum flocculation was shown
to occur at a lower dosage when polymer was added in stages than
when added all at once. It can be seen from Figure 6 that the same
extent of flocculation occurs at about half the dosage (1.25 kg/t) for
multistage addition (polymer added in 0.25 kg/t increments every
15 s) when compared to single-stage addition (2.5 kg/t). This
increase in efficiency is attributed to a more efficient uptake of
PEO molecules by the solid and a more uniform distribution of the
polymer in solution. Similar results were obtained by La Mer, et al.
(26).
The rate of polymer addition also affects the extent of flocculation.
Hogg et al. (1) showed that the faster the rate of polymer addition,
the larger the size of flocs obtained (see Fig. 7). However, the flocs
obtained with a high rate of addition are relatively weaker and
break down more easily than those formed with a lower rate of
addition. Too high a rate of polymer addition would lead to
insufficient mixing in the suspension, leading to poorer
flocculation.
Figure 6
Sediment volumes of flocculated palygorskite
suspension after 24 h.
(From Ref. 14.)
Page 424
Figure 7
Settling rate as a function of the rate of polymer additions at fixed polymer dosage.
(From Ref. 1. Used with permission of Society for Mining, Metallurgy, and Exploration, Inc.)
B
Solution Properties
Selection of a polymer for a given application should also take into account the
suspension conditions of the system. A brief discussion of the various suspension
parameters follows.
1
The pH
An important and at times controllable variable is suspension pH. Since H+ and
OH- act as potential-determining ions for most particles, the magnitude and nature
of the charges on them may be manipulated by varying the pH. If bridging
flocculation is desired, it is advantageous to reduce the charges on the particles
(close to isoelectric point) so that they may approach each other closely, promoting
bridging by the loops and tails of the adsorbed polymers. However, as noted
earlier, if the molecular weight of
Page 425
the polymer is high enough, it can span over the electrical double
layers of the two particles, leading to their aggregation. In the case
of charge-patch mechanism, higher charges on the particle would
increase polymer adsorption and flocculation. If an ionic polymer
is used, the effect of pH on the polymer must also be considered.
Higher charge densities on the polymer (e.g., by hydrolysis) would
increase repulsion between polymer segments, leading to extended
conformation favoring bridging. When the charge on the polymer
is of the same nature as that on the particle, an optimum exists for
the charge density on the polymer (27). In such cases, although the
polymer chain is extended at a high charge density favoring
bridging, polymer adsorption on the particles is decreased due to
electrostatic repulsion.
2
Ionic Strength
Ionic strength of the suspension is another important parameter
since it affects thickness of the electrical double layer associated
with the particles. Additionally, ionic competition could affect
polymer adsorption on a given substrate. In both charge-patch and
bridging mechanisms, higher ionic strengths would be desirable
since compression of electrical double layers would allow close
approach of particles. Again, the effect of the electrolyte on the
polymer must not be ignored because it may cause coiling of
charged polymer molecules by shielding intersegment repulsion,
leading to less efficient adsorption and bridging.
3
Solvency
The role of conformation of the polymer in flocculation is not clear
and has been examined only indirectly. Besides pH and ionic
strength, conformation is also affected by the quality of the solvent,
usually water in industrial applications and the temperature of the
suspension. Stuart et al. (28) showed that adsorption of polymer on
a substrate can be altered by the quality of solvents. The adsorption
of polyvinyl pyrrolidone on silica, for example, could be varied
from about 0.7 to 0 mg/m2 by changing the ratio of water and N-
methyl pyrrolidone in the mixture used as the solvent.
C
Solid Properties
Physical properties (such as size, size distribution, and solids
loading) and surface chemical characteristics of the solid surface
play an important role in flocculation.
1
Effect of Particle Size and Size Distribution
Coarse particles (e.g., more than 100 µm) are believed not to
flocculate for three reasons: (1) rate of collisions is very small and
kinetically the process is not favored; (2) polymer molecules are
not capable of attaching to the
Page 426
surface during collisions; and (3) the floc breakage is high in the shear field induced and the
"flocs" break as soon as they are formed. In general, it is believed that flocculation is best
observed for fine particles, typically less than 50 µm. A systematic study of particle size effect
on flocculation was conducted recently and, contrary to expectations, coarse particles that
normally do not agglomerate (e.g., greater than 100 µm) were found to do so if the flocculant of
appropriate molecular weight (high) was employed (29). In order to flocculate coarse particles,
the binding force (due to bridging) needs to exceed the breaking force (due to the turbulence) in
the system to form a strong floc. This implies that floc formation would not be observed below a
certain molecular weight. In other words, there exists a critical molecular weight for
flocculation to occur. It can be seen from Figure 8 that for a given particle size there is indeed a
critical molecular weight of
Figure 8
Effect of particle size and molecular weight of the flocculant on flocculation of apatite with polyacrylic acid.
(From Ref. 29.)
Page 427
Figure 9
Change in size distribution during flocculation of bauxite
ore using a hydrolyzed polyacrylamide flocculant.
(From Ref 30. Used with permission of
Elsevier Science Publishers.)
the polymer above which it acts as a flocculant. The molecular
weight at which the flocculation starts increases with the increase
in the particle size. This trend was also observed in dolomite-
poly(ethylene oxide) systems where flocculation started at about 1
million molecular weight for -37 µm (-400 mesh) particles but the
-53 +44 µm (-270 +325 mesh) particles did not flocculate with the
same polymer. The different flocculation optimum observed with
450,000, 1 million, and 4 million PAA may be attributed to the
molecular weight distribution of the polymer employed.
The lower limit of particle size for flocculation in the above system
was observed to be about 2 µm below which no significant flocs
were observed. The value of the lower limit, however, would be
different for different polymer molecules considering that the same
particles were flocculated very well with a non-ionic
polyacrylamide flocculant. Similar observations for fine particle
flocculation were made by Yu and Attia (30) in a
bauxitehydrolyzed polyacrylamide system. It can be seen from
Figure 9 that there was a marked increase in large particle sizes.
However, no appreciable flocculation of fine particles was
observed indicating flocculation of only coarser particles.
Page 428
2
Effect of Solids Loading
Mechanism of Flocculation
It should be noted that besides the nature of the flocculant itself,
suspension parameters such as solids loading also can affect the
mechanism of flocculation. Mangravite et al. (16) showed that
flocculation of kaolin clay by cationic polyacrylamide occurred by
charge neutralization up to a clay phase ratio of about 10-3. Beyond
this solid loading, molecular weight was found to be an important
factor, signifying that flocculation was occurring by a bridging
mechanism. It was proposed that high molecular weight polymers
with charges opposite to those of the particles in suspension
represent a convergence in bridging and charge neutralization
mechanisms and are efficient in dewatering and thickening of
waste water and industrial organic sludges.
Effect on Flocculation Efficiency
Besides the mechanism of flocculation, the extent and efficiency of
flocculation is directly dependent on the solid loading of the
suspension (31,32). Large numbers of particles promote
interparticle collisions either by Brownian motion and/or external
agitation leading to increased flocculation. However, the effect of
solid loading is not pronounced at high pulp densities (450%) (18).
In the case of raw water treatment, which contains high
concentrations of microscopic organisms, flocculant aids like
pulverized limestone and activated silica are used to increase the
rate of water treatment and obtain strong, rapidly settling flocs. In
fact, floc strength can be increased as much as six times by adding
activated silica (32). Increasing the pulp density also raises the
resistance of flocs to agitation. When flocs are broken due to
agitation, the detached polymer segments can rebridge floc
fragments or reabsorb on the same surface. Large numbers of flocs
would increase the interfragment collisions, leading to high
stability. While high pulp densities promote flocculation, their
effect on floc settling rates must be considered carefully. Under
low-pulp-density conditions, sedimentation of individual flocs
occurs, and hence the settling rate of the suspension is higher. At
higher pulp densities, on the other hand, the flocs may form a
network structure, regarding the sedimentation rate (34). The solid
loading at which the three-dimensional structure starts forming is
called the gel point.
3
Effect of Agitation Intensity
While the chemical nature of the flocculant and suspension
conditions are important, physical parameters encountered during
the flocculation process, such as the mode of polymer addition,
agitation intensity, and time of agitation, also play a role in the
structure and properties of the flocs produced. It has been shown
that the normally accepted method of instantaneous flocculant
addition followed by prolonged agitation does not result
Page 429
in larger flocs (1). Instead, continuous flocculant addition during
stirring was demonstrated to give optimum floc properties. It is
also postulated that agitation can lead to floc densification (35,36).
A limited effort has also been made in correlating agitation
intensity and velocity gradients in a given environment with the
size and structure of flocs produced (37).
The effect of turbulence on the floc formation also was studied for
apatitepolyacrylic acid (molecular weight 450,000) system (29)
(see Fig. 10). The polymer molecular weight was selected so that
the effect of turbulence was not masked by strong interaction of the
polymer with the surface. At faster agitation or higher energy input,
less flocculation is expected since floc breakage would be larger.
This, however, was found not to be totally true. At lower impeller
speeds (300 rpm), less flocculation of the particles may be
attributed to the difficulty in keeping them in suspension. Of the
particles that flocculate, the impeller speed has a significant effect
on flocculation of coarser particles but not of fines. This implies
that the flocs formed with fine particles using PAA are stronger
with distinctive capabilities to handle high shear values, or
indirectly the strength of the flocs formed with fine particles is very
high.
4
Hardness of the Solids
Another interesting observation was also made about the
grindability of a material and its flocculation behavior. It was found
that hard materials had relatively less tendency to flocculate as
compared to the softer materials. This behavior was explained by
considering fracture and wetting characteristics of the materials.
Figure 10
Effect of stirring speed on flocculation of coarse particles.
(From Ref. 29.)
Page 430
V
Flocculation Kinetics
The nature of a polymeric flocculant also affects the flocculation
rate, an important consideration in industrial applications. The
molecular weight of the polymer, especially in the case of the
bridging mechanism, plays an important role in this respect. The
adsorbed polymer is postulated to increase the effective collision
diameter of the particles, leading to increased flocculation with
higher molecular weight polymers due to an increase in the
interparticle collisions (38). However, Gregory (20) argues that the
increase in the effective diameter of the particles is compensated by
a lower diffusion rate. Gregory's work with negatively charged
polystyrene and cationic polymers has demonstrated that there is no
appreciable difference in rates of flocculation with polymers having
molecular weights from 50,000 to 106. It seems that the charge-patch
neutralization mechanism of flocculation, the one operative in this
case, is not very sensitive to molecular weight. However, in a later
review on effects of polymers on colloid stability, Gregory (39,40)
has observed that higher polymer molecular weight increases the
flocculation rate. It appears that the role of molecular weight in
flocculation kinetics, especially in the case of charge-patch
neutralization mechanism, needs to be further investigated. Further,
the physical parameters, such as the rate at which polymer is mixed
in the suspension, can also affect the kinetics of flocculation,
particularly when polymer adsorption rather than particleparticle
collisions is the rate-determining step (40).
VI
Mathematical and Molecular Description
Polymer-induced flocculation may be considered to consist of only
the binary collisions between the particles/clusters in a manner
consistent with well-known orthokinetic flocculation proposed by
Smoluchowski (42). The rate expression for flocculation is a slightly
modified form of the expression suggested by Smoluchowski and is
given by
where Fij is the rate of aggregation of species i with j, K is the

flocculation rate constant, ni, nj are the number of interacting species
per unit volume, and Eij is the collision efficiency factor.
In the development of the model for flocculation, fine particles are
considered to be dispersed and the polymer adsorption to be
irreversible, as is generally the case for adsorption of polymers and
macromolecules on solid surfaces (43). This assumption is valid for
the time scale involved in the flocculation process. As observed by
Wagberg and Lindstrom (44), the
Page 431
polymer adsorption is fast since very rapid flocculation is observed
with high molecular weight polymers. Also, this assumption is valid
in case of flocculation in turbulent medium and for particles not
exhibiting large porosities. It can, therefore, be assumed that the
polymer adsorption is extremely fast and particleparticle collision is
the rate-controlling step. Further, it is assumed that the particles are
large enough to neglect the Brownian diffusivities.
A number of expressions have been suggested for the collision
efficiency Eij in the literature by LaMer et al. (26), Hogg (45), and
Moudgil and Vasudevan (46). The latter expression is general in
nature. In fact, in the limiting case, Moudgil's expression given in
Equation (2) reduces to the corrected version of the one suggested
earlier (26).
where Fi, Fj are the fraction of sites on i and j that are active for
polymer adsorption, qi, qj are the fractional surface coverage (of
active sites) at a given polymer dosage, FR is the ratio of ''active
sites" on the active component to "active sites" on the inert
component = Fi/Fj, and FR > 1 (by definition).
Equation (2) is based on the number of active sites on various
substrates, assuming that the conformation of the adsorbed polymer
on the different surfaces is the same (47,48). It has been shown that
there exists a critical collision factor above which flocculation is
observed (48). This theory was further improved to incorporate the
differences in the conformation of the adsorbed polymer. F for a
material is estimated from the saturation adsorption data with respect
to a reference material, the entire surface of which is considered to
be active. On the other hand, q is the ratio of the nonequilibrium and
saturation adsorption values for a given material.
Flocculation is often directly correlated with the polymer adsorption
on solid particles. Behl (19) has recently postulated that flocculation
is observed only above a critical value of F.
The model has been extended to take into account the presence of
low molecular weight fraction of the polymer in the given flocculant.
The lower molecular weight fraction (LMWF) acts as a dispersant,
while the higher molecular weight fraction has flocculating
capabilities. Flocculation is first observed for polymers whose
molecular weight corresponds to the critical F for flocculation. Since
the adsorption of the smaller molecular weight fraction is much
faster than larger molecular weights, therefore, to develop an
expression for estimation of the collision efficiency for the particles
in the presence of the LMWF polymer, it is assumed that the lower
molecular weight polymer is introduced in the system prior to the
injection of the flocculating component. In other words, the higher
molecular weight
Page 432
fraction would encounter a partially covered surface. The site on the
surface can be defined as was done earlier by Hogg (45) and
Moudgil et al. (47) to be the area occupied by the adsorbed polymer
molecule. Floc formation would be observed if a coated site collides
with the "active" bare site or vice versa. Any collision of a
flocculant-coated site with a LMWF-coated site or vice versa will
not yield flocculation. The implicit assumption of La Mer's model,
that orientation of the particles, as well as the adsorbed polymer
conformation remains unchanged as the particles approach one
another, are maintained in this model.
A model was developed on the same lines as that of the site blocking
parameters done earlier (19,50). In the presence of the lower
molecular weight fraction, the collision efficiency factor is given by
the following expression:
where is the collision efficiency in the presence of the LMWF.

and are the active fraction and coverage of the particle surface for
the lower molecular weight fraction. It is clear that, in the absence of
LMWF, Equation (3) reduces to the expression proposed by Moudgil
et al. (4648).
In particular cases where all the particles have the same polymer
coverage and the same fraction of the active sites, for example, in a
single mineral suspension, Equation (3) reduces to
which can be written as

where E is the collision efficiency for single mineral particles.
Further, if all the surface sites are considered to be active, then
Equation (5) reduces to
and if no LMWF was present (i.e., ), Equation (6) reduces to

the expression suggested by Hogg as the corrected formulation of the
La Mer model. The methods for determining adsorption parameters
(F, q, , ) have been discussed in detail elsewhere (19,50).
To quantify the role of LMWF in flocculation, it is clear from
Equation (6) that collision efficiency would be lower in the presence
of LMWF. This decrease would, of course, depend on the product of
and (termed as LMWF adsorption parameters). Physically the
former term signifies the molecular weight of LMWF used while the
latter denotes the dosage employed. A higher F would be expected
for a larger molecular weight of
Page 433
LMWF molecule, which clearly should have higher site-blocking
capabilities. Since the fraction of active surface for a flocculant is much
larger than that for the LMWF, it would not be possible to depress
flocculation of the active component (F = 1).
It is clear from Figure 11 that for material with F = 1 the surface coverage
qmax for maximum E would decrease from 0.5 for no LMWF to about
0.25 if the LMWF adsorption parameter is 0.5. However, if F = 0.5, then
on increasing the LMWF adsorption parameter the qmax reduces from 0.5
to zero, implying that LMWF would strongly depress flocculation if the
material itself exhibits a low tendency toward floc formation. This can be
observed from the decrease in the absolute value of E for cases with F less
than one (see Fig. 12). Moudgil et al. (47) and Hogg (45) have shown that
the collision efficiency factor is directly related to the amount of
flocculation. Therefore, flocculation is difficult to suppress in cases with F
= 1 as the LMWF adsorption parameter value of 1.0 is not practical.
However, in cases where F is 0.5, which implies using a weaker flocculant
molecule, depression of flocculation is easily achievable with LMWF.
A
Coverage for Maximum Flocculation
The dosage required to obtain maximum flocculation in single minerals
can be calculated from the expression proposed earlier (26,45,48). It can
be easily shown that maximum flocculation would be observed at half
surface
Figure 11
Effect of lower molecular weight fraction adsorption parameter on surface coverage
for maximum collision efficiency.
(From Ref. 50.)
Page 434
Figure 12
Effect of lower molecular weight fraction adsorption parameter on the maximum
collision efficiency factor.
(From Ref. 51.)
coverage. This, however, was not found to be true in a number of
systems studied earlier (18,26,43,45,49).
In the presence of a LMWF, the surface coverage corresponding to
maximum flocculation can be obtained by simply maximizing q in
Equation (4) and is given by the expression:
The maximum value of collision efficiency can be obtained by

substituting qmax in Equation (4):
Equation (8) reduces to the limiting cases of LaMer's model and Hogg's
expression (qmax = 1/2) when the LMWF is absent and the entire
surface is considered active. It can be seen from Figure 13(A), that in
the absence of the LMWF, maximum E is obtained at half the surface
coverage. However in the presence of the LMWF, depending on the
adsorption parameters , the maximum E would be obtained at a
value lower than 0.5. For example,
Page 435
Figure 13
Effect of effective surface coverage on the collision efficiency factor.
(From Ref. 50.)
at (Fig. 13D) maximum flocculation or collision

efficiency factor (value of 0.125) would be obtained at a surface
coverage of 0.25.
It is clear that if the flocculant employed consisted of a distribution
of molecular weights, which normally is true in commercial
flocculants, then, depending on the amount of various fractions in
the polymer, optimum flocculation may or may not be achieved at
half surface coverage.
VII
Concluding Remarks
An important criterion to bear in mind while using polymers as
flocculants for a particular application is the presence of ions in the
system. In the case of certain industrial wastes, and often in the
case of mineral suspensions, multivalent cations and anions are
present, which could affect the charged polymers. For instance,
polyacrylate can form insoluble calcium salts with calcium ions
while acrylate-polyacrylamide copolymers can lose effectiveness in
the presence of divalent cations (17).
It is interesting to note that most of the work reported in the
literature
Page 436
has been carried out with linear homopolymers or random
copolymers. Only limited efforts have been directed at
investigating bloc copolymers (51). Graft copolymers or branched
homopolymers could increase the possibility of bridging due to
more "tentacles" reaching out to the surface of suspended particles.
The foregoing discussion indicates that selecting a polymer as a
flocculant is a complex problem involving the optimization of a
number of parameters. Unfortunately, theory has not developed to
the extent where floc properties and polymer specifications can be
correlated directly so that specific polymers can be synthesized for
individual applications. Empirical testing is still the best method to
select the most effective flocculant for a specific application,
although the broad principles enumerated in the foregoing
discussion may be utilized to narrow the list of polymers to be
examined. In addition to the properties of the polymer as a
flocculant, its toxicity effects also need to be considered so that it is
environmentally acceptable.
Acknowledgments
The authors acknowledge the financial support of this work by the
Minerals Resources Research Center, University of Florida.
References
1. R. Hogg, R. C. Klimpel, and D. T. Ray, "Agglomerate structure
in flocculated suspensions and its effects on sedimentation and
dewatering," SME Annual Meeting, New Orleans, Louisiana,
Preprint Number 8696 (1986).
2. P. Somasundaran and R. Ramachandaran, Mineral solution
equilibria and the role of it on polymer flocculation, Flocculation,
Sedimentation and Consolidation (B. M. Moudgil, and P.
Somasundaran, Eds.), Engineering Foundation, New York, p. 23
(1985).
3. F. B. Birkner and J. J. Morgan, Polymer flocculation kinetics of
dilute colloidal suspension, J. Am. Water Works Assoc., 60: 175
(1968).
4. A. P. Black and S. A. Hannah, Measurement of low turbidities, J.
Am. Water Works Assoc., 57: 901 (1957).
5. G. W. Holmes, Determination of suspended solids in sewage by
light adsorption, Sewage Works J., 7: 642 (1935).
6. S. L. Daniels, The utility of optical parameters in evaluation of
processes of flocculation and sedimentation, Chem. Eng. Prog.,
Symp. Ser., 65(97): 171 (1969).
7. V. K. La Mer and T. W. Healy, The role of filtration in
investigating flocculation and redispersion of colloidal dispersions,
J. Phys. Chem., 67: 2417 (1963).
8. B. M. Moudgil and S. Behl, Characterization techniques for
agglomerate
Page 437
structures, Characterization Techniques for Solid-Solution
Interface in Ceramic Systems, American Ceramic Society,
Westerville, Ohio (1992).
9. C. Dirican, "The structure and growth of aggregates in
flocculation," M.Sc. Thesis, Pennsylvania State University (1981).
10. B. M. Moudgil, "Aggregation of Phosphatic Clay," Proc.
Consolidation and Dewatering of Fine Particles, U.S. Bureau of
Mines, Tuscaloosa, AL (1982).
11. R. F. Stewart and D. Sutton, (i) Control of structure in
particulate solids suspensions, SolidLiquid Separation (J. Gregory,
Ed.) Horwood, Chichester (1984) p. 111.
(ii) Structure of flocculated suspensions, Chem. Ind., 21 May,
(1984).
12. P. T. L. Koh, J. R. G. Andrews, and P. H. T. Uhlherr, Floc size
distribution of scheelite treated by shear flocculation. Int. J. Miner.
Process., 17: 46 (1986).
13. H. R. Harner and J. R. Musgrave, A photoelectric
sedimentation method for particle size determination in the
subsieve range, ASTM Spec. Tech. Pub. 234 (1959), p. 172.
14. B. Hoghooghi, "Flocculation of palygorskite suspensions and
the orientation of particles in flocs," M.Sc. Thesis, University of
Florida, Gainesville (1985).
15. B. M. Moudgil and T. V. Vasudevan, Characterization of flocs,
Flocculation and Dewatering, (B. M. Moudgil and B. J. Scheiner,
Eds.), Engineering Foundation, New York (1988) p. 167.
16. F. J. Mangravite, Jr., C. R. Leitz, and P. E. Galick, Organic
polymeric flocculants: Effect of charge density molecular weight
and particle concentrations, Flocculation, Sedimentation and
Consolidation (B. M. Moudgil and P. Somasundaran, Eds.),
Engineering Foundation, New York (1985), p. 139.
17. J. A. Kitchener, Flocculation in mineral processing, The
Scientific Basis of Flocculation (K. J. Ives, Ed.), NATO Advanced
Study Institute, Sijthoff and Noordhoff, Alphen dan dey Rijn, The
Netherlands, pp. 283328 (1978).
18. W. F. Linke and R. B. Booth, Physical chemical aspects of
flocculation by polymers, Trans. Am. Inst. Min. Metall. Eng., 217:
364 (1959).
19. S. Behl, "Controlled polymer adsorption for enhanced
selectivity in flocculation," Ph.D. Dissertation, University of
Florida (1993).
20. J. Gregory, Rates of flocculation of latex particles by cationic
polymers, J. Colloid Interface Sci., 42(2): 448 (1973).
21. F. Mabire, R. Audebert, and C. Quivoron, Flocculation
properties of some water-soluble cationic copolymers toward silica
suspensions: A semiquantitative interpretation of the role of
molecular weight and cationicity through a 'patchwork' model, J.
Colloid Interface Sci., 97(1): 120 (1984).
22. T. Lindstrom and G. Glad-Nordmark, Network flocculation and
fractionation of latex particles by means of a polyethylene oxide
phenol formaldehyde resin complex, J. Colloid Interface Sci.,
97(1): 62 (1984).
23. D. A. Stanley and B. J. Scheiner, Mechanically induced
dewatering of ion exchange attapulgite flocculated by a high
molecular weight polymer, Colloid Surfaces, 14: 151 (1985).
24. J. Gregory, The action of polymeric flocculants, Floccuolation,
Sedimentation and Consolidation (B. M. Moudgil and P.
Somasundaran, Eds.), Engineering Foundation, New York, p. 125
(1985).
Page 438
25. T. W. Healy, Polymer flocculation of colloidal dispersion,
Water Soluble Polymers (F. W. Stone and J. J. Stratta, Eds.), High
Polymers Monograph Series, Interscience, New York (1970).
26. V. K. La Mer, R. H. Smellie, Jr., and Pui-Kum Lee,
Flocculation, subsidence and filtration of phosphate slimes, J.
Colloid Interface Sci., 12: 230 (1957).
27. A. S. Michaels, Aggregation of suspensions by
polyelectrolytes, Ind. Eng. Chem., 46: 1485 (1954).
28. M. A. C. Stuart, J. M. H. N. Scheutjens, and G. J. Fleer, The
role of the solvent in polymer adsorption: Displacement and
solvency effect, Polymer Adsorption and Dispersion Stability (E.
D. Goddard and B. Vincent, Eds.), American Chemical Society,
Washington, D.C. (1984), p. 53.
29. S. Behl, B. M. Moudgil, and T. S. Prakash, Effect of particle
size on flocculation, J. Colloid Interface Sci., 158: 511 (1993).
30. S. Yu and Y. A. Attia, Review of selective flocculation in
mineral fines, Flocculation in Biotechnology and Separation
Systems (Y. A. Attia, Ed.), Elsevier, Amsterdam (1987), p. 601.
31. S. L. Daniels, Flocculation, Encyclopedia of Chemical
Processing and Design, Marcel Dekker, New York, (1985), p. 184.
32. Y. C. Cheng, "The effect of surface hydration on the adsorption
and flocculation of a model silica suspension using polyethylene
oxide," M.Sc. Thesis, University of Florida, Gainesville (1985).
33. H. E. Hudson, Jr., Flocculation and flocculation aids, J. Am.
Water Works Assoc., 49: 242 (1957).
34. R. Buscall, R. F. Stewart, and D. Sutton, Structure in colloid
suspensions and effects on solidliquid separation, Filtr. Sep.,
May/June, 183 (1984).
35. R. C. Klimpel, The structure of agglomerates in flocculated
suspensions, M.Sc. Thesis, Pennsylvania State University (1984).
36. R. C. Klimpel and R. Hogg, Effects of Flocculation conditions
on agglomerate structure, J. Colloid Interface Sci., 113(1): 121
(1986).
37. D. T. Tomi and D. F. Bagster, The behavior of aggregates in
stirred vessels, Trans. I. Chem. E., 56: 1 (1978).
38. W. E. Walles, Role of flocculant molecular weight in the
coagulation of suspension, J. Colloid Interface Sci., 27(4): 797
(1968).
39. J. Gregory, The effect of cationic polymers on the colloidal
stability of latex particles, J. Colloid Interface Sci., 55(1): 35
(1976).
40. J. Gregory, Effects of polymers on colloid stability, The
Scientific Basis of Flocculation (K. J. Ives, Ed.), NATO Advanced
Study Institute, Sijthoff and Noordhoff, Alphen aan dey Rijn (The
Netherlands), p. 101 (1978).
41. A. L. Wingstein and R. A. Stratton, Polymer adsorption and
particle flocculation in turbulent flow, Polymers Adsorption and
Dispersion Stability (E. D. Goddard and B. Vincent, Eds.), ACS
Symposium Series 240 (1984), p. 429.
42. M. V. Smoluchowski, Versuch einer mathematischen theorie
koagulations kinetic kolloider lösungen, Z. Phys. Chem., 92: 129
(1917).
43. J. M. Singer, F. C. A. Vekemans, J. W. Th Lichtenbelt, F. Th.
Hesselink, and P. H. Wersema, Kinetics of flocculation of latex
particles by human gamma gloubin, J. Colloid Interface Sci., 45(3):
608 (1973).
Page 439
44. L. Wagberg and T. Lindstrom, Kinetics of polymer-induced
flocculation of cellulosic fibers in turbulent flow, Colloid Surfaces,
27: 29 (1987).
45. R. Hogg, Collision efficiency factors for polymer flocculation,
J. Colloid Interface Sci., 102(1): 232 (1984).
46. B. M. Moudgil and T. V. Vasudevan, Role of active sites in
selective flocculation, J. Colloid Interface Sci., 127(1): 239 (1989).
47. B. M. Moudgil, B. D. Shah, and H. S. Soto, Collision efficiency
factors in polymer flocculation of fine particles, J. Colloid
Interface Sci., 119(2): 466 (1987).
48. B. M. Moudgil and S. Behl, A model of the selective
flocculation process, J. Colloid Interface Sci., 146(1): 1 (1992).
49. P. Somasundaran, Y. H. Chia, and R. Gorelik, Adsorption of
polyacrylamide on kaolinite and its flocculation and stabilization,
Polymer Adsorption and Dispersion Stability (E. D. Goddard and
B. Vincent, Eds.), American Chemical Society, Washington, D.C.
(1984), p. 393.
50. S. Behl, B. M. Moudgil, and T. S. Prakash, Control of active
sites in selective flocculationPart I: A mathematical model, J.
Colloid Interface Sci., 161: 414 (1993),
51. J. V. Dawkins and G. Taylor, Surface layer thickness in non-
aqueous polymer dispersions stabilized by AB copolymers of
polystyrene and poly (dimethyl siloxane), J. Chem. Soc. Faraday I,
76: 1263 (1980).
Page 441
16
Experiments
N. K. Lazaridis, George A. Stalidis, and K. A. Matis
I
Introduction
The investigation of processes involving a large number of
parameters is often laborious and tedious, leading to a large
experimental program that yields results not easily quantifiable. It
would be very helpful if it were possible to determine in a swift
manner which of the parameters are really important for the
process, using the minimum number of experiments, and then
proceed with a detailed investigation of these only. Statistical
methods may prove extremely useful in such cases, mainly in
determining which of the parameters warrant further study and
which of them are of relatively little importance.
The factorial design of experiments is one such technique of a
priori experimental program design; parameter values are changed
often simultaneously but in a coherent way, and this offers the
advantage of obtaining quantitative information not only about the
effect of the main variables of the process, but also of their mutual
interactions (1,2). The main benefit of the factorial design is
savings in time, experiments, and labor for the experimenter.
Flotation is such a process, involving many parameters, such as the
pH of solution or dispersion, the concentration of various species
(e.g., the material to be floated, the other materials present in the
solution or dispersion, and the chemicals added for the
enhancement of the process like the collector and frother), the
operating conditions (gas, pulp, and wash water
Page 442
flowrates). The optimum conditions for the treatment of a particular
system are usually established with a trial-and-error procedure. The
factorial design has been applied in some cases, e.g., to the
precipitation of iron as jarosite (3) and goethite (3,4).
In the following chapters the basic steps of the factorial design are
presented; its use is demonstrated by the application of the technique
to the investigation of a case of dissolved-air flotation of ions [the
differential separation technique for copper from zinc ions (5)], a
process widely used in water and waste water treatment.
II
The Complete Factorial Design
Let us assume that a physicochemical process (e.g., flotation) is
being studied. Such a process may involve several independent
variables (e.g., pH, concentrations, and flow rates) as well as several
dependent ones (e.g., recoveries and selectivity). The purpose of the
investigation is to determine which set of values of the independent
variables yields an optimum value for one (or possibly all) of the
dependent ones.
The optimization of the process (i.e., the determination of the
maximum or minimum) may be conducted in the usual way: One of
the independent variables is changed while the others are kept
constant. Thus its influence on one of the dependent variables is
determined. Therefore, for a system involving M independent
variables, if each procedure is repeated K times, then the number of
necessary experiments will be
which obviously results in a large number of experiments, even for
relatively simple systems. One further disadvantage of this method is
that it does not yield any information about the possible interaction
between the various independent variables.
The alternative way is to use the factorial design to plan the required
experiments. For this, it is not even necessary to know the process in
depth; it is only enough to decide a priori which variables are going
to be the independent ones to be investigated. The factorial design
proceeds by minimizing the number of necessary experiments; this is
succeeded mainly with the simultaneous alteration of the
independent variables in an orderly and predetermined way. The
factorial design was initially proposed by Fisher (6) and further
developed by Box and Wilson (7,8).
If there are n independent variables xi (x1, x2, . . ., xn) involved in the
processwhich in the factorial design are called factorsand one or
more dependent variables yi (y1, y2, . . ., ym)termed responses or
parametersthe aim of the design is to express the phenomenon with
a second-order response function of the following type:
Page 443
where y is the parameter to be optimized (e.g., recovery), is its

estimated value, b0 is a constant term, bi, is the linear effect of the
ith factor, bij and bijk are the interaction effects of the factors, and
bii is the squared effect of the ith factor. In this way, every possible
interaction between factors and their effect on the process is taken
into account. The purpose of the factorial design is to determine the
minimum number of experiments required to obtain Equation (2).
There are several kinds of factorial design, depending upon various
details: the full (or complete) factorial design and the incomplete
factorial design; these are further classified into the orthogonal or
the rotable factorial design. In this work, the complete orthogonal
technique will be presented.
In performing an experimental plan according to the factorial
design, a preliminary elaboration stage is followed. Here
information for the system under study is gathered to define the
factors and parameters. The former should be uncorrelated, in order
to be true independent variables. From these factors, only the
important, or basic ones, are kept; the others are discarded and do
not take part in the following calculations.
The next stage, termed insertion stage, involves the planning and
performance of the necessary experiments. Initially, the first-order
experiments are done. For this, two extreme levelsa high one and a
low oneare chosen for the n factors. Alternatively, the basic level
(or center) and the interval unit (step) may be defined. The upper
level is denoted by +1; the lower, by -1; and the center by 0 (Table
1). Each experiment corresponds
Table 1 Program Schedule of Factorial
Experiment.
Factors
x1 x2 x3 . . . xn
Definition of xi
Units of xi
Basic level or 0
Interval Dxi
High level or + 1
Low level or - 1
Page 444
Table 2 Complete Factorial Design of the First Order,
N = 22.
Parameter
Experiment no. x0 x1 x2 x1x2
1 +1+1+1+1
2 +1+1-1 -1
3 +1-1 +1-1
4 +1-1 -1 +1
Properties of the scheduled plane
1.
2.
3.
Property 3 means that the vector x is orthogonal; hence,
the whole schedule is also called orthogonal.
to a combination of factor levels; the various combinations constitute

the schedule of the full factorial experiment. The total number of
experiments is
(It is usual, however, to repeat each experiment at least twice.) The

response function obtained by the statistical processing of the results
yields a response function of the following type:
Table 2 presents an example of a schedule for the case of two
factors, x1 and x2 (n = 2, or 22 schedule); space has been allowed for
a replication of each experiment (yielding two values of y, y1, and
y2, as well as the mean ). An additional column for x0 is also
allocated; it is used to estimate the coefficient b0. Once the
experiments are performed, the interaction coefficients are calculated
from the relation:
Page 445
Following this calculation, the equation of coalteration can be

written as follows:
The values of the coefficients bi, may now be used to determine

which of the factors are significant for the process (always
considering that these conclusions are strictly true for the range of
factor values covered by the upper and lower levels). For this, the
variance of and of are calculated using the following
equations:
and
where r is the number of replications of the experiments (usually r =

2). The Student t criterion (t*) is then found in tables (9,10) for a
given probability level (or confidence level), and the degree of
freedom is calculated by the following equation: f = N(r - 1). Finally,
the following test is performed: Only those bi coefficients are
retained for which
In this way, a response function containing only the significant
factors is obtained. This, however, has to pass another test, to check
whether this function describes satisfactorily the process. For this,
the Fisher test is applied. The experimental Fisher criterion is
calculated, using the following equations:
and
where fR is the degree of freedom of the model, fR = N - (k* + 1)

with k* being the number of significant interaction coefficients bi,
omitting b0.
Page 446
The value of Fexp is then compared with the theoretical one (found
as before in statistical tables); the following relationship should hold
for the model to be satisfactory:
The (absolute) value of the coefficients is an indication of their

influence: The higher the value, the more important the
corresponding factor is. The elimination of some factors from the
function may be due to various reasons, e.g., that the factor is not
basic or that the scheduled range is narrow.
The next stage involves the optimization, in which the optimum
function is sought by adding second-order factor coefficients. Since
the number of experiments required to calculate the values of bi for
Equation (2) is quite large, the second-order factorial design is based
on the results already obtained with the first-order one, adding some
additional experiments. The following function is thus determined:
where
If N1 = 2n is the number of experiments required for the first-order

factorial experiment, then the total number of experiments needed
for a second-order factorial experiment is given by the following
equation:
The extra experiments correspond to new points on the schedule,
located one at the center of the plane and the others at a distance a
on the axes. The distance a can be calculated from the following
equation:
However, the new vector xi is not orthogonal anymore. Therefore,

new variables are introduced so that all the properties of the
scheduled plan are valid again. The new variable is defined by
the following equation:
Table 3 illustrates the schedule for the second-order factorial

experiment. The following response function is obtained first:
Page 447
Table 3 Second-Order Factorial
Experiment of the Type 2.2
Run no. x1 x2 x1x2
1 -1 -1 +1 +1 +1 1/3 1/3
2 +1 -1 -1 +1 +1 1/3 1/3
3 -1 +1 -1 +1 +1 1/3 1/3
4 +1 +1 +1 +1 +1 1/3 1/3
5 -1 0 0 +1 0 1/3 -2/3
6 +1 0 0 +1 0 1/3 -2/3
7 0 -1 0 0 +1 -2/3 1/3
8 0 +1 0 0 +1 -2/3 1/3
9 0 0 0 0 0 -2/3 -2/3
and then Equation (13) is determined (the various equations required

for the calculation of the coefficients bi and the variances are given
in Ref. 11). The statistical significance of the coefficients and the
validity of the model are again checked with the Student t and the
Fisher criteria.
III
The factorial design technique will be illustrated below with its
application to a complex physicochemical process, precipitate
flotation.
Leaching solutions or effluents from metal-producing plants are
often acidic aqueous solutions of various heavy and toxic metals,
like copper and zinc. Their removal before disposal or water reuse is
one of the most important effluent and/or waste water treating
processes; if the concentration of the metal species is sufficiently
high, then it is also possible to achieve some degree of recycling,
too.
Precipitation of these heavy metals is the main removal method, in
the form either of hydroxides (usually by lime addition) or sulfides
(with addition of Na2S); the latter is slightly advantageous compared
to the former (11). The removal of the precipitate is done by various
techniques, like sedimentation or filtration.
Flotation is an alternative separation method, offering high removal
rates and a better effluent quality (12,13). Flotation of sulfides as
minerals (14) or as precipitates (15,16) has been thoroughly
investigated.
Page 448
The precipitation of dissolved metals as sulfides in aqueous solutions
occurs by the following chemical reaction:
where M denotes the metal; the reactivity of sulfides (S2-, HS-) with
heavy metal ions is high (15). A drawback of this method is that
colloidal or very fine particles are precipitated (17), which generally
settle poorly; therefore, this further justifies the selection of flotation
as the solidliquid separation method.
A binary system, comprising copper and zinc ions, was chosen for
the application of the factorial design; the purpose of the work was
to investigate the feasibility of selective separation of copper from
zinc ions by precipitate flotation as sulfides, expecting copper to be
removed into the foam layer, while zinc remained in solution.
Dissolved-air flotation, a technique that generates fine air bubbles
(with bubble diameters less than 120 µm), was applied; experiments
were performed at low pH values, using dodecylamine as a collector
(11). Residual copper and zinc concentrations were determined by
atomic absorption spectrophotometry, while dodecylamine was
determined spectrophotometrically. The batch tests were carried out
in duplicate.
A
First-Order Experiment
In the preparatory stage, four of the parameters considered as being
the most important were selected: the Cu2+ concentration, the
solution pH, and
Table 4 First-Order Factorial Design N = 24.
Factors
x1 x2 x3 x4
Definition of xi [Cu2+] [RNH2] pH [Na2S]
Units of xi ppm ppm - % w/v
Basic level or 0 150 150 2.0 ± 0.1 14.0
Interval Dxi 100 100 0.5 ± 0.1 10.5
High level or + 1 250 250 2.5 ± 0.1 24.5
Low level or - 1 50 50 1.5 ± 0.1 3.5
From Ref. 11.
Page 449
Table 5 Treatment Combinations of 24 First-Order
Factorial Experiment.
Cu removal (%) Amine removal (%)
Run no. x1 x2 x3 x4 (i) (ii) (i) (ii)
1 +1+1+1+1 99.22 99.90 31.80 26.96
2 +1 -1 -1 +1 99.11 96.59 17.49 15.29
3 -1 +1 -1 +1 91.19 95.27 30.04 31.80
4 -1 -1 +1+1 99.80 95.70 7.59 3.19
5 +1+1 -1 -1 15.72 12.60 28.72 26.96
6 +1 -1 +1 -1 16.56 13.60 3.19 7.59
7 -1 +1+1 -1 80.44 78.30 18.60 12.00
8 -1 -1 -1 -1 72.81 69.46 3.19 1.00
9 +1+1+1 -1 12.11 15.58 12.00 18.60
10 +1 -1 -1 -1 13.11 14.60 4.29 1.00
11 -1 +1 -1 -1 72.83 74.25 34.00 29.60
12 -1 -1 +1 -1 77.22 76.50 5.39 9.79
13 +1+1 -1 +1 98.81 98.05 34.00 28.28
14 +1 -1 +1+1 98.57 99.67 40.59 40.59
15 -1 +1+1+1 99.80 97.36 26.96 19.04
16 -1 -1 -1 +1 87.62 85.94 5.39 11.99
From Ref. 11.
the concentrations of Na2S and dodecylamine added. The

percentage of copper ion and dodecylamine removal was defined
as the two responses. Table 4 illustrates the schedule for the first-
order 24 factorial experiment, including the factors and their basic
and interval values. The Zn2+ concentration was kept constant
throughout these experiments.
The experiments were carried out according to the schedule in
duplicate; residual copper and dodecylamine concentrations were
determined, and the results for copper and dodecylamine removal
(%) in the foam were calculated. The relative levels of the factors
and the results are shown in Table 5 (11). The coefficients bi of
Equation (4) were calculated, and their significance was
determined using the Student t test at a 95% confidence level, with
t(16,0.95) = 1.746.
For the Cu2+ recovery in the foam, the results indicated that the
linear term x2 could be ignored; the interactions x1x2, x3x4, x2x3,
x2x4, x1x2x3, x1x2x4, x1x3x4, x2x3x4, and x1x2x3x4 were found to
have only a minor contribution and could also be ignored. Hence,
the surfactant concentration was found to be insignificant (in the
range selected for the experiments).
Page 450
The following equation was obtained:
or
where a negative sign means an adverse effect of the factor(s). Due

to the high recoveries obtained, further optimization was not deemed
necessary for this response (RCu).
For the amine removal, it was found that the terms x3, x1x2x3, and
x2x3x4 could be ignored. The removal of amine (RRNH2) was thus
found to be independent of pH (again, for the selected range of
factor values). The model obtained had the following form:
where x1 = [Cu2+], x2 = [RNH2], x3 = [pH], and x4 = [Na2S].

The statistical significance of both models [Equations (19) and (20)]
was tested for a 95% confidence level by applying the Fisher
criterion [Equations (11) and (12)]; the theoretical Fisher criteria
were for the first model
and for the second one
In both cases, the required condition (Fexp < Ftheor) was obeyed.
Hence, the models were considered as satisfactory.
B
Second-Order Experiment
A second series of experiments was also conducted with the same
ionic system, according to the factorial design, in order to illustrate
the second-order scheme. Three variables were chosen as factors: the
initial Cu2+ (x1) and Zn2+ (x5) concentrations and initial mixing
speed RPM (x6). The percentage of Cu2+ removal (RCu) and Zn2+
remaining in solution (TZn) were defined as responses.
For this series, the solution pH was kept at an average value of 2;
this is
Page 451
more realistic, since in practice fine pH adjustment is rarely
achieved. The required Na2S added for the ion precipitation was a
30% excess of the stoichiometric amount needed for Cu2+ each time.
The schedule is presented in Table 6.
A first-order 23 factorial experiment was conducted first; the model
obtained was checked and found to be statistically significant, as in
example A above, and the second-order 23 factorial experiment was
then planned. Seven additional points were found to be necessary,
and these were added to the initial first-order eight-point schedule;
the distance of the axial points from the center was calculated as a =
1.215. The experimental conditions and the results obtained for Cu2+
recovery (RCu) and Zn2+ remaining in solution (TZn) are presented
in Table 7; this is divided in two parts, the upper part being the
nucleus for the first-order factorial design and the lower part being
the additional points.
The results were used to obtain a response equation similar to
Equation (13), e.g., for the percentage of Cu2+ removal (RCu). The
statistical analysis for the coefficients bi using a value of the Student
t of t(15,0.99) = 2.60 for a 99% confidence level showed that all
terms were significant for Cu2+ recovery:
where, again, x1 = [Cu2+], x5, = [Zn2+], and x6 = [RPM]. The model

was found to be significant according to the Fisher test (Ftheor =
F(5,15,0.99) = 4.56).
It is also possible to determine the absolute optimum of a response.
For
Table 6 Second-Order Factorial Design.
Factors
x1 x5 x6
Definition of xi [Cu2+][Zn2+] stirring
speed
Units of xi ppm ppm rpm
Basic level or 0 150 150 200
Interval Dxi 100 100 150
High level or 250 250 350
+1
Low level or 50 50 50
-1
From Ref. 11.
Page 452
Table 7 Treatment Combinations of 23 Second-Order
Factorial Experiment.
Cu removal Zn remaining
(%) (%)
Run
No. x1 x5 x6 (i) (ii) (i) (ii)
1 +1 +1 +1 99.80 99.85 89.39 86.90
2 +1 -1 -1 98.60 98.57 89.22 89.22
3 -1 +1 -1 95.57 94.61 89.39 86.90
4 -1 -1 +1 98.50 97.67 89.22 89.22
5 +1 +1 -1 99.06 99.22 84.77 80.50
6 +1 -1 +1 99.81 99.74 87.79 89.22
7 -1 +1 +1 99.01 98.38 86.90 82.28
8 -1 -1 -1 91.05 91.82 89.22 85.66
9 +1.2150 0 99.85 99.85 84.54 84.15
10 -1.215 0 0 95.21 95.71 84.54 85.72
11 0 +1.2150 99.62 99.62 80.16 81.03
12 0 -1.215 0 99.09 99.46 87.74 87.33
13 0 0 +1.215 99.74 99.46 84.54 85.72
14 0 0 -1.215 95.07 94.94 87.29 88.47
15 0 0 0 99.50 99.62 84.54 86.51
From Ref. 11.
example, Equation (21) may be differentiated with respect to the

variables involved and set equal to 0:
from which optimum flotation conditions (i.e., factor values yielding

the maximum percentage of RCu) are determined: [Cu2+] = 237
mg/L, [Zn2+] = 158.2 mg/L, and stirring rate = 248.5 RPM.
It should be noted that, although the determination of values for
Cu2+, Zn2+, and RPM yielding the maximum of Cu2+ removal from
solution is numerically correct, in actual practice (i.e., waste water or
effluent treatment), ion concentrations are random variables. Hence,
the differentiation is meaningful only for those variables that may
actually be controlled, like the stirring rate in the present case.
IV
Conclusions
Flotation is a complex process, influenced by a multitude of
parameters (e.g., the pH of the solution, the concentration of the
various species in-
Page 453
volved, and the time and intensity of agitation provided at the
initial stage, among others). Some of them may prove to be of
major or minor importance, depending upon the case, as has been
demonstrated in the previous examples.
A complete investigation of the process, given such a large number
of parameters, would require a considerable amount of resources
and time and would lead to an extremely large number of
numerical results; the interpretation of such results is always
strenuous and often shows that some of the experiments were
performed in vain. In contrast, the application of the factorial
design provides a concise experimentation scheme and yields in a
reasonably rapid way to achieve quantitative results, indicating
possibly which of the parameters warrant further experimentation
and which are of relatively little importance to the process. It is
also possible, in some cases, to determine the optimum conditions
for a particular process.
It should be emphasized, however, that there are two limitations to
the application of the factorial design: (a) that it does not help in
the investigation of the kinetics of the underlying processes and (b)
that the conclusions are strictly applicable in the given range of
factor values.
References
1. N. L. Johnson and F. C. Leone, Statistics and Experimental
Design in Engineering and the Physical Science, Vol. 2, Wiley,
New York (1977).
2. D. L. Massart, A. Dijkstra and L. Kaufman, Evaluation and
Optimization of Laboratory Methods and Analytical Procedures,
Vol. 1, Elsevier, Amsterdam (1979).
3. A. R. Burkin, Use of statistical methods of experimental design
in optimizing hydrometallurgical processes, Hydrometallurgical
Process Fundamentals (R. G. Bautista, Ed.), Plenum, New York
(1984), p. 529.
4. S. Agatzini, ''A statistical approach to the precipitation of iron as
goethite," Ph.D. Thesis, Royal Sc. Mines, Imperial Coll., London
(1981).
5. N. K. Lazaridis, "Selective removal of copper, zinc and arsenic
ions from aqueous solutions by flotation techniques," Ph.D. Thesis,
Chemistry, Aristotle University, Thessaloniki (1991).
6. R. A. Fisher, The Design of Experiments, 6th ed., Oliver and
Boyd, London, (1951).
7. G. E. P. Box and K. B. Wilson, On the experimental attainment
of optimum conditions, J. R. Stat. Soc. (B) 13: 1 (1951).
8. G. E. P. Box, Evolutionary operation, Appl. Stat., 4: 3 (1957).
9. G. E. P. Box, W. G. Hunter, and J. S. Hunter, Statistics for
Experimenters, Wiley, New York (1978).
10. C. Chatfield, Statistics for Technology, Chapman and Hall,
London (1975).
11. G. A. Stalidis, K. A. Matis, and N. K. Lazaridis, A statistical
approach to precipitate flotation of CuS/ZnS, Int. J. Miner.
Process., 24: 203 (1988).
Page 454
12. T. A. Pinfold, Precipitate flotation, Adsorptive Bubble
Separation Techniques (R. Lemlich, Ed.), Academic Press, New
York (1972), p. 75.
13. K. A. Matis and P. Mavros, Recovery of metals by ion flotation
from dilute aqueous solutions, Sep. Purif. Methods, 20: 1 (1991).
flotation of complex CuPbZn sulfides, Innovations in Flotation
15. D. Bhattacharyya, A. B. Jumawan, Jr., and R. B. Grieves,
Separation of toxic heavy metals by sulfide precipitation, Sep. Sci.
Technol., 14: 441 (1979).
16. C. P. Beitelshees, C. J. King, and H. H. Sephton, Precipitate
flotation for removal of copper from dilute aqueous solutions.
Recent Developments in Separation Science (N. N. Li, Ed.), Vol. 5,
CRC, Boca Raton, FL (1979), p. 43.
17. N. K. Lazaridis, K. A. Matis, G. A. Stalidis, and P. Mavros,
Dissolved-air flotation of metal ions, Sep. Sci. Technol., 27:
17431758 (1992).
Page 455
17
Biomodification of Mineral Surfaces and Flotation
M. K. Yelloji Rao and P. Somasundaran
Henry Krumb School of Mines, Columbia University, New York,
New York
I
Introduction
Association of microorganisms with the formation and
solubilization of mineral deposits since geological times is well
known. Mining operations have long benefited from the activities
of such naturally occurring microbes, especially from the ability of
some bacteria to leach metals from insoluble ores. In spite of the
fact that biooxidation of minerals has been occurring in nature for
centuries, it was believed until recently to be a purely chemical
process mediated by water and dissolved oxygen. The contribution
of the bacteria to metal leaching was recognized in 1947 when
Colmer and Hinkle (1) identified a bacterium, now called
Thiobacillus ferrooxidans, from the acid drainage of bituminous
coal mines. The presence of bacteria in the leach waters of Rio
Tinto mines was first confirmed in 1963 (2). Today, many countries
throughout the world are adapting bioleaching processes to recover
metals from a wide variety of ores. The principal metals recovered
microbiologically on an industrial scale include copper, uranium,
and gold, although other metals could also be recovered. Apart
from Thiobacillus ferrooxidans, innumerable bacterial species are
known to be associated with leaching operations. All these
organisms in general are acidophilic chemolithotrophs.
Several review articles and symposium proceedings summarize
various aspects of bacterial leaching and provide excellent
background information
Page 456
(310). Apart from leaching, these microorganisms could also be
used with advantage for mineral beneficiation. Bacterial leaching
generally requires longer periods for dissolving the mineral
substrate, while biomodification of surfaces in principle should
require only a few minutes. Applications of biomodification of
mineral surfaces include beneficiation of sulfide minerals from
complex ore body by selective or bulk flotation, removal of pyritic
sulfur from coal and the extraction of precious metals, encapsulated
in sulfide matrix, by bacterial pretreatment of sulfide concentrate
followed by flotation for enrichment of precious metals in sulfide
concentrate. Reagents of biological origin such as secreted
metabolites and bacterial debris could also be used in the flotation
of nonsulfide and fluorspar minerals. The purpose of this chapter is
to provide relevant recent information on bioflotation and to
describe emerging microbial-based technologies for mineral
beneficiation.
In this chapter, a critical discussion of the available literature is
made with special reference to the following aspects:
1. bacteria of relevance to mineral beneficiation
2. influence of bacteria and their products on flotation
3. factors that influence biomodification process
II
Bacteria of Relevance to Mineral Beneficiation
Thiobacillus ferrooxidans is the most widely studied bacterium and
is currently the major leaching microorganism of economic
importance. It is a chemolithotroph, obtaining its energy for
metabolism through the oxidation of ferrous iron and reduced-
valence inorganic sulfur compounds. This acid-generating
microorganism is intimately associated with sulfide mineralization
in mine waters. Among the other species widely involved in
leaching operations Thiobacillus thiooxidans, Leptospirullum
ferrooxidans, and Sulfolobus acidocaldarious are important. T.
ferrooxidans can directly oxidize sulfide minerals through prior
bacterial attachment or indirectly through ferric sulfate generated
as a metabolical product (1114). Galvanic interactions could also
come into play if more than one sulfide mineral is present. The role
of T. ferrooxidans in such a case is to remove passive elemental
sulfur by oxidizing it to sulfate (12,1517). Thiobacillus thiooxidans
is a similar chemolithotroph capable of oxidizing sulfur and
reduced-valence sulfur compounds, but not ferrous iron (6).
Leptospirallum ferrooxidans species is a mesophile while
Sulfolobus acidocaldarious species are thermophilic organisms
capable of optimum activity at higher temperatures on the order of
6080°C (18). The organism Sulfolobus is mixo-
Page 457
trophic, capable of using both inorganic substrate (ferrous and
reduced sulfur) and simple organic compounds (glucose and yeast
extract) as energy sources. Sulfate-reducing bacteria (SRB),
desulfovibrio, is a specific group of anaerobic bacteria capable of
using the oxygen of sulfates, reducing them to hydrogen sulfide,
for performing anaerobic respiration (19).
III
Influence of Bacteria and Their Products on Mineral Flotation
A
Biomodification of Sulfide Mineral Surfaces
Solojenken (20,21) reported for the first time the use of
microorganisms of the type SRB, microbe fat, and biomass in the
flotation of several sulfide and nonsulfide minerals. The SRB was
found to depress the flotation of both chalcopyrite and sphalerite
but not those of molybdenite and galena (Fig. 1) (20). Studies with
different sulfide concentrates show that SRB can desorb the
xanthogenate coatings making them lose their flotation activity. In
the case of bulk concentrate containing both sphalerite and galena,
although control experiments do not show any selectivity in their
separation,
Figure 1
Effect of conditioning with sulfate-reducing bacteria on sulfide flotation.
(Previously published in Proceedings of ICCM-79, BARC, India. Reprinted
with permission from Indian Institute of Metals.)
Page 458
treatment with SRB yielded about 95% recovery of galena while
sphalerite recovery under these conditions was only 4.5%.
Yelloji Rao et al. (22,23) have reported the effect of bacterial
conditioning with T. ferrooxidans on the floatability of sulfide
minerals. Figure 2 shows the effect of bacterial conditioning on
sphalerite recovery under different flotation conditions (23). While
pretreatment with sulfuric acid solution at pH 2 without any bacteria
itself improved sphalerite flotation significantly (with and without
flotation reagents), conditioning at the same pH with T. ferrooxidans
(108 cells/mL) further improved the floatability. However, bacterial
treatment did not show any effect when flotation was carried out after
conditioning with both activator and collector. On the other hand,
when the cell dosage was increased to 109 cells/mL, the floatability
of sphalerite was reduced drastically (Fig. 3), even when flotation
was carried out after conditioning with flotation reagents (23). In the
case of galena also, natural floatability was enhanced appreciably
upon pretreatment with sulfuric acid solution (Fig. 4) (23). However,
when T. ferrooxidans (108 cells/mL) was also used for conditioning,
such enhancement of natural floatability was not observed. The
floatability of collector-treated
Figure 2
Effect of bacterial conditioning on the floatability of sphalerite at 108 cells/mL
concentration of T. ferrooxidans at 1% pulp density.
(Previously published in Minerals and Metallurgical Processing, Vol. 9, 1992.
Reprinted with permission from Society of Mining Engineers, Inc.)
Page 459
Figure 3
Effect of bacterial conditioning on the floatability of sphalerite at 109 cells/mL
galena was reduced by biopretreatment, and increase of cell dosage to
109 cells/mL further depressed the flotation drastically (Fig. 5) (23).
During the conditioning with sulfuric acid solution, dissolution or
surface oxidation of the mineral is possible. The dissolution reactions
for sphalerite and galena can be represented as follows:
Elemental sulfur thus generated on the mineral surfaces is
hydrophobic and hence can increase the natural floatability of both
sphalerite and galena. Thiobacillus ferrooxidans is known to oxidize
such elemental sulfur to form sulfate (14). At the acidic pH of 2,
while the zinc sulfate formed on sphalerite surface is soluble, lead
sulfate species formed on galena is insoluble. Presence of oxidized
insoluble products on the sulfide mineral surface is known to interfere
with the action of the collector (24). Hence, the flotation recovery of
galena was significantly decreased after biopretreatment. At the high
cell dosage of 109 cells/mL, floatability is proposed to be governed
mainly by the enhanced attachment of the bacteria. Recovery of both
sphal-
Page 460
Figure 4
Effect of bacterial conditioning on the floatability of galena at 108 cells/mL
erite and galena were thus affected markedly by the microbes and
their products (22,23).
Lyalikova and Lyubavina (25) have discussed the possibility of using
T. ferrooxidans to separate antimony and mercury sulfides by
flotation. Bacterial conditioning for an hour and a half produced no
change with respect to cinnabar recovery (89%) while antimonite
recovery decreased from 89% to 6.2% leading to almost complete
separation. It is suggested that T. ferrooxidans can oxidize the surface
of antimonite crystals leading to depression of the floatability while
cinnabar remains unaffected with no change in its floatability.
B
Biomodification of Nonferrous and Nonsulfide Minerals
Solojenken (20) has demonstrated the use of reagents of biological
origin such as microbe fat and biomass as flotation reagents in the
case of nonsulfide and fluorspar ores of different deposits. This
investigation showed the good selectivity of microbe fat as a
collector in the flotation of fluorspar.
Page 461
Figure 5
Effect of bacterial conditioning on the floatability of
galena at 109 cells/mL concentration of T. ferrooxidans
at 20% pulp density.
(Previously published in Minerals and Metallurgical
Processing, Vol. 9, 1992. Reprinted with permission
from Society of Mining Engineers, Inc.)
With 10 mg/L of microbe fat, fluorite was selectively floated while
the associated minerals, calcite and barite, floated little, and quartz
floated practically none. The optimum flotation of fluorite with
oleic acid, a conventional collector for nonsulfide ores, is obtained
in the pH range of 710 while with microbe fat is in the range 410.
The expanded pH range is considered to be due to the fact that
microbe fat contains a number of saturated and unsaturated fatty
acids, with the former fixed more rapidly to the fluorite surface.
Infrared spectra for collector and microbe fat interactions with
minerals were identical. The use of microbe fat (in 1:1 ratio of
oleic acid) yields practically the same quality scheelite concentrate
with increased WO3 recovery and reduced cost.
The use of biomass as a flotation agent has been demonstrated also
for celestine and associated minerals, calcite, barite, and quartz
(20). Flotation recoveries of these minerals using biomass are
shown in Figure 6 as a function of pH. Both calcite and barite were
depressed with 10 mg/L of biomass while about 20% of celestine
could be selectively floated with it. With an increase of biomass
concentration to 20 mg/L, the selectivity could be increased to
about 80% around pH 10. With about 5075 mg/L, all the
Page 462
Figure 6
Effect of biomass on the floatability of celestine
and associated minerals (1,2 3: 20 µg/L; 4: 5 µg/L;
5: 10 µg/L).
(Previously published in Proceedings of ICCM-79,
BARC, India. Reprinted with permission from Indian
Institute of Metals.)
associated minerals were practically depressed. The depressing
action of biomass was further compared with dextrine, a
conventional depressant: 300 g/ton of dextrine yielded a 96.2%
CaF2 with 84.9% recovery to be compared with 96.3% of CaF2
with a recovery of 86.5% obtained with 50 g/ton of biomass. The
ability of biomass macromolecules to hydrate in aqueous solution
and to more selectively adsorb on the gangue minerals made it
possible to use them as depressors in nonsulfide ore flotation.
C
Removal of Pyritic Sulfur from Coal
Townsley et al. (26) have reported the effect of bacterial
conditioning with T. ferrooxidans on the suppression of pyritic
sulfur as a part of the coal-cleaning process. Advanced coal-
cleaning processes are necessary to treat pyritic sulfur coals in an
environmentally acceptable and cost-effective manner. The effect
of conditioning pyrite with bacterial suspension in direct bacterial
liquor and membrane-filtered liquor with and without bacteria at
pH 2.0 is shown in Figure 7 (26). The natural floatability of 84.5%
decreased to about 7.7% upon conditioning with bacteria
suspended at pH 2 for 2.5 min. Conditioning with membrane-
filtered bacterial liquor gave 32.5% recovery, while with direct
bacterial liquor was about 17.4%. Best
Page 463
Figure 7
Comparison of effect of different pretreatments on subsequent pyrite flotation.
(Previously published in Biotechnology and Bioengineering, Vol. 30, 1987.
Reprinted with permission from John Wiley & Sons, Inc.)
suppression was obtained with membrane-filtered liquor
supplemented with bacteria wherein the recovery was only 4%. This
suggests the importance of bacteria and the associated medium in
altering the floatability of pyrite. Attia and Elzeky (27) have shown
that at natural pH, neither the nutrient medium without bacteria nor
the bacterial suspension produced any depression of coal (Fig. 8). On
the other hand, although nutrient medium alone did not affect pyrite
flotation, the bacterial conditioning was found to affect the flotation,
and such an effect was severe when old culture (refers to the duration
of culture previously grown on pyrite substrate) was used.
Pyrite suppression due to bacterial conditioning has been suggested
to be due to different reasons. Townsley et al. (26) have suggested it
to be due to changes in surface charge in response to either
adsorption of intact bacteria, bacterial metabolites, or bacterial
debris. On the other hand, Harada and
Page 464
Figure 8
Flotation of coal and pyrite at natural pH after
preconditioning with nutrient medium and T.
ferrooxidans. (a) Coal with aerated nutrient medium
[O] and with T. ferrooxidans [·]. (b) Pyrite with
aerated nutrient medium. (c) Pyrite with 2-week old
T. ferrooxidans. (d) Pyrite with 6-week old
T. ferrooxidans.
(Previously published in Coal Science and
Technology, Vol. 9, 1985. Reprinted with
permission from Elsevier Science Publishers B.V.)
Kuniyoshi (28) have attributed pyrite depression to bacterial
oxidation, leading to formation of jarosite or a jarosite-like
insoluble sulfate (hydrophilic film) on the pyrite surface. Attia and
Elzeky (27) have suggested that the bacteria could adsorb on the
mineral surface and grow. Also, T. ferrooxidans are capable of
producing polymeric surface active substances (mainly
polysaccharides and lipids) that can adsorb on the pyrite surface.
Growth of bacteria coupled with adsorption of cell-excreted
compounds can be expected to make the mineral surface more
wettable thereby affecting the floatability. The bioadsorption
process is believed to be rapid enough to be completed in a matter
of a few minutes. In fact, in the case of pyrite, 90% of the
inoculated cells attached to the surface within 2 min of
conditioning (29). Earlier work on pyrite flotation has shown that a
short conditioning for 2.5 min with T. ferrooxidans can reduce the
natural floatability (26). It is unlikely that within such a short
period enough bacterial metabolites are produced to affect the
floatability to any measurable extent. The only other possibility for
flotation depression under these conditions is the bacterial
attachment onto the minerals, which will result in a hydrophilic
surface. Therefore, unlike in bacterial leaching processes, the
formation of bacterial metabolites and their attachment to minerals
during short
Page 465
periods have to be adequately considered in any study of
biomodification processes.
Dogan et al. (30) have reported that bacterial conditioning with T.
ferrooxidans followed by flotation not only removed pyritic sulfur
but also resulted in a coal with lower ash content than sulfur
removal by bacterial leaching alone. However, the reasons for the
removal of ash content as well as the removal of pyrite due to
bacterial conditioning were not presented.
IV
Factors That Influence Biomodification Process
An understanding of the effect of various parameters would help in
developing appropriate biomodification of the mineral surfaces for
selective flotation or suppression. The factors that can be controlled
prior to or during conditioning are listed in Table 1, and their
possible effects on the biomodification process are discussed.
A
Particle Size
Use of microorganisms in mineral beneficiation through
biomodification of the surface involves bacterial conditioning
followed by flotation. In cases where biomodification process is
governed by surface oxidation to only a few molecular layers to
change its surface characteristics, it is beneficial to avoid smaller
particles as it would take longer periods for modifying sufficient
surface to bring out the desired changes. Also, higher cell
concentration may be necessary to achieve the desired effect. On
the other hand, with coarser particles, although the cell
concentration and the conditioning time required is relatively less,
there could be practical difficulties in carrying out subsequent
flotation.
Table 1 Factors To Be Controlled for
Better Biomodification Process.
Mineral 1. particle size
2. pulp density
Bacteria1. inoculum (cell) concentration
2. preadaptation to
a. mineral substrate
b. medium pH
Medium1. effect of pH
2. presence and absence of
nutrients
3. composition
Page 466
B
Pulp Density
Pulp density would be an important controlling factor where
biomodification process is governed by either bacterial attachment
or surface oxidation by bacterial activity. For the latter case, at low
pulp densities, the modification rate is limited by the external
surface area of the minerals, while at high pulp densities, the
important parameters are transfer and solubility of gaseous
nutrients necessary for bacterial activity. For a process governed by
bacterial attachment, the controlling parameter is always the ratio
of available surface area to the cell concentration. Studies on the
pyritic sulfur removal rate from coal by the thermophilic
microorganism S. acidocaldarious showed an increase in the
removal rate of up to 15% and then a decrease with further increase
in the pulp density (Fig. 9) (31). This has been attributed to particle
agglomeration at relatively high pulp densities leading to a
reduction in external surface area. The influence of bacterial
concentration as well as the absence of gaseous nutrients may also
be the reasons for such a decrease in the removal rate.
C
Inoculum Concentration
Regardless of the mechanism of microbial action during
biomodification processes, cell concentration and their activity will
play a dominant role in bringing out the desired changes. A change
in cell concentration alone could change the floatability of the
mineral to a larger extent. The flotation
Figure 9
Diagram showing pyritic sulfur removal as a function of coal pulp
density.
(Previously published in Biotechnology and Bioengineering, Vol.
27, 1985. Reprinted with permission from John Wiley & Sons, Inc.)
Page 467
recovery of sphalerite (Table 2) after biotreatment with T.
ferrooxidans, although almost unaffected for an initial cell
concentration in the range of 102 to 108 cells/mL, was found to be
decreased with further increase in initial cell concentration to 109
cells/mL (23). On the other hand, the floatability of galena was
found to decrease gradually with increase in initial cell
concentration from 102 cells/mL to 109 cells/mL during
biotreatment. As explained earlier for sphalerite, while at low cell
concentrations, biomodification is governed by bacterial-induced
oxidation; at high concentrations, it is determined by the adsorption
of the bacteria on the minerals. This study is significant for the
differential flotation of sphalerite and galena from a complex
sulfide as well as for bioflotation of the sulfide matrix associated
with precious metals. In the case of pyritic sulfur removal also,
flotation suppression was better at higher cell concentrations (26).
Importance of optimum cell concentration is further shown in
Figure 10 for the removal of pyritic sulfur from coal by the
thermophilic organism S. acidocaldarious (31). The removal rate
increased initially with increasing cell concentrations up to 2 × 108
cells/mL and decreased with any further increase in the cell
concentration. At very high cell concentrations, heavy foaming was
observed, and this may reduce the transfer of gaseous nutrients, O2
and CO2, hindering the desired bacterial activity (20).
Since bacteria could grow on mineral substrates alone (23) during
prolonged conditioning, even in the absence of nutrient medium,
care should be taken to control the cell population whenever longer
duration of conditioning is involved. This may not be a critical
factor in biomodification processes since time involved in initial
conditioning is limited to a few minutes.
Table 2 Effect of Initial Cell Concentration During
Biopretreatment on the Floatability of Sphalerite and
Galena (20% pulp density; 24 h of biopretreatment).
Percent recovery
Initial cell Sphalerite (10-6 M Galena (10-5
concentration/mL CuSo4 + 10-5 M NaIX) M NaIX)
0 97.0 92.1
102 94.6 84.0
104 96.3 78.2
106 95.9 75.9
108 94.2 48.3
109 52.9 22.5
Previously published in Minerals and Metallurgical
Processing, Vol. 9, 1992. Reprinted with permission
from Society of Mining Engineers, Inc.
Page 468
Figure 10
Diagram showing the effect of initial cell concentration
on pyritic sulfur removal.
(Previously published in Biotechnology and
Bioengineering, Vol. 27, 1985. Reprinted with
permission from John Wiley & Sons, Inc.)
D
Preadaptation of Bacteria
The major question that arises in using microbial processes is with
respect to longer conditioning time. Preadaptation of bacteria to
mineral substrate can reduce the duration of bacterial conditioning
prior to flotation. Figure 11 shows the effect of bacterial adaptation
on the bacterial leaching of sphalerite (32). Although the extraction
of zinc does increase with increased duration of bioleaching, for a
given leaching time extraction also increases with the bacterial
adaptation period. The growth and activity of bacteria, in general,
follow a lag-log-stationary phase. Due to preadaptation, the lag
phase will be either reduced or diminished depending on the
duration, and the bacteria would begin their biomodification
process immediately. Adaptation of T. ferrooxidans to various
sulfide mineral substrates is a well-known phenomenon (3234).
For the biomodification process wherein chemical oxidation is also
involved, a higher acidic pH is advantageous for faster rate. The
optimum growth and microbial activity for T. ferrooxidans is at pH
2.3 in the temperature range of 2035°C. Preadaptation of these
bacteria to acidic media could improve tolerance to lower pH
values of 0.5 to 1.0. Modified strains
Page 469
Figure 11
Diagram showing the effect of the bacterial adaptation period and bioleaching time
on the sphalerite bioleaching.
(Previously published in The Chemical Engineering Journal, Vol. 44, 1990.
Reprinted with permission from Elsevier Sequois S.A.)
of T. ferrooxidans, which can withstand very acidic solutions (pH <1),
have been developed (35).
E
Medium pH and Composition
As biomodification process is a two-stage process involving initial
conditioning followed by flotation, the pH maintained has to be suitable
for both optimum bacterial activity and flotation. In many cases, these
two pH conditions may be different, and hence either pH has to be altered
after initial conditioning or the bacteria has to be preadapted to possess
more or less the same activity at the flotation pH.
Effect of nutrient medium on the flotation of minerals should be
understood in order to determine whether the cells have to be separated
prior to subjecting them for the biomodification purpose. Nutrients are
necessary for the growth and activity of bacteria. It is reported that in the
case of pyrite flotation, nutrient medium alone does not have any effect
on flotation (27) while membrane-filtered bacterial liquor can reduce
flotation markedly (26). Also membrane-filtered liquor supplemented
with bacteria
Page 470
brings out better suppression of pyrite than bacterial suspension at
pH 2.0 (26).
V
Summary
Biomineral technology involving surface modification of minerals
by bacterial conditioning prior to flotation could be successfully
used in mineral beneficiation. Applications of biomodification of
mineral surfaces include beneficiation of sulfide minerals from
complex ore body by bulk or selective flotation, removal of pyritic
sulfur from coal, and the enrichment of precious metals
encapsulated in sulfide matrix. In the case of flotation of nonsulfide
and nonferrous ore deposits, reagents of biological origin such as
microbe fat and biomass act as collector and depressant. These
reagents not only provide better selectivity and recovery but also
are less expensive than conventional reagents. Mechanisms for
surface modification can involve either adsorption of microbes and
their products or surface oxidation to a few molecular layers to
change its surface characteristics. Change in bacterial concentration
alone plays a dual role in altering the floatability of minerals.
Presence and absence of nutrients, necessary for bacterial growth
and activity, show different effects on the flotation of minerals.
Preadaptation of desired bacteria to mineral substrates enhances the
rate of biomodification, thereby decreasing the period of
conditioning required prior to flotation. This would be an important
factor with respect to use of microorganisms in mineral
beneficiation. Adaptation of bacteria to medium pH and
composition could also be useful for obtaining maximum
selectivity in flotation. Most significant development in this area
will come from development of microbes that have functional
groups that can selectively interact with minerals. It should also be
possible to develop microbes with fimbriae with selective
recognition for mineral surface species when the precise role of the
fimbriae in attachment and adhesion to surfaces is understood.
Acknowledgments
The authors acknowledge Unilever Research, USA for partial
support of this work.
References
1. A. R. Colmer and M. E. Hinkle, The role of microorganisms in
acid mine drainage, A preliminary report, Science, 106: 253256
(1947).
2. W. E. Razell and P. C. Trusell, Isolation and properties of iron
oxidizing Thiobacillus ferrooxidans, J. Bact., 85: 595603 (1963).
3. C. L. Brierly, Bacterial Leaching, CRC Crit. Rev. Microbiol., 6:
207262 (1978).
Page 471
4. D. Woods and D. E. Rawlings, Bacterial leaching and biomining,
A Revolution in Biotechnology (Jean L. Marx, Ed.) Cambridge
University Press, New York, pp. 8193 (1989).
5. A. E. Torma, Leaching of metals, Biotechnology, A
Comprehensive Treatise in Eight Volumes (H. J. Rehm and G.
Reed, Eds.), VCH, Federal Republic of Germany, 6B:367399
(1988).
6. S. R. Hutchins, M. S. Davidson, J. A. Brierly, and C. L. Brierly,
Microorganisms in reclamation of metals, Ann. Rev. Microbiol., 40:
311316 (1986).
7. J. O. Gregory and M. K. Robert, Microbiological metal
transformations, Biotechnol. Progr., 2: 115 (1986).
8. L. E. Murr, Theory and practice of copper sulfide leaching in
dumps and in-situ, Min. Sci. Eng., 12: 121189 (1980).
9. K. A. Natarajan, I. Iwasaki and K. J. Reid, Some aspects of
microbe mineral interactions of interest to Duluth Gabbro copper,
nickel sulfides. Recent Progress in Biohydrometallurgy, (G. Rossi
and A. E. Torma, Eds.), Associazono Mineraria Sarda, Italy (1983),
pp. 169184.
10. K. A. Natarajan, ''Biohydrometallurgy of Gold and copper,"
Proc. Research and Practice in Hydrometallurgy, Udaipur,
February, pp. 3.1.119. (1990).
11. V. K. Berry and L. E. Murr, Direct observations of bacteria and
quantitative studies of their catalytic role in the leaching of low
grade copper bearing waste, Metallurgical Applications of
Bacterial Leaching and Related Microbiological Phenomena (L. E.
Murr, A. E. Torma, and A. Brierly, Eds.), Academic Press, New
York (1978), pp. 103136.
12. V. K. Berry, L. E. Murr and J. B. Hiskey, Galvanic inter-actions
between chalcopyrite and pyrite during bacterial leaching of low
grade waste, Hydrometallurgy, 3: 309326 (1978).
13. A. E. Torma, Biohydrometallurgy as an emerging technology,
Biotechnology and Bioengineering Symposium (1986), 16: 4963.
14. L. C. Bryner, J. V. Beck, D. B. Davis, and D. G. Wilson,
Microorganisms in leaching mineral sulfides, Ind. Eng. Chem., 46:
25872592 (1954).
15. K. A. Natarajan and I. Iwasaki, Microbiological Effects on
Metallurgical Processes, (J. A. Clum and L. A. Haas, Eds.), AIME,
New York (1985), pp. 113.
16. N. Jyothi, "The role of galvanic interactions in the bioleaching
of mixed sulfides," M.Sc. (Engg.) Thesis, IISc., Bangalore (1988).
17. B. E. Purkiss, Corrosion in industrial situations by mixed
floras, Microbial Aspects of Metallurgy (J. D. A. Miller, Ed.),
Elsevier, New York, pp. 107128. (1970).
18. T. D. Brock, K. M. Brock, R. T. Belly, and R. L. Weiss,
Sulfolobus: A new genus of sulfur-oxidizing bacteria living at low
pH and high temperature, Arch. Mikrobiol., 84: 5468 (1972).
19. J. R. Postgate and L. L. Campbell, Classification of
Desulfovibrio species, the nonspolrulating sulfate-reducing
bacteria, Bact. Rev., 30: 732738. (1966).
20. P. M. Solojenken, "Floatability and Leaching of Low Grade
Ores with Reagents of Biological Origin," International Conference
on Advances in Chemical Metallurgy, BARC, Bombay, India
(1979), pp. 37/1-37/13.
Page 472
21. P. M. Solojenken, L. L. Lyubavina, V. K. Larin, L. D.
Bergelson, and E. V. Dyatlovitskaya, A new collector in nonsulfide
ore flotation, Bulletin Nonferrous Metall., 16: 21 (1976).
22. M. K. Yelloji Rao, K. A. Natarajan, and P. Somasundaran,
"Effect of Bacterial Conditioning of Sphalerite and Galena with
Thiobacillus ferrooxidans on Their Floatability," Engineering
Foundation Conference Mineral Bioprocessing, (R. W. Smith and
M. Misra, Eds.), TMS Publication (1992), pp. 105120.
23. M. K. Yelloji Rao, K. A. Natarajan, and P. Somasundaran,
Effect of Biotreatment with Thiobacillus ferrooxidans on the
floatability of sphalerite and galena, Miner. Metall. Process.
(1992), 9:95100.
24. I. W. Wark, Principles of Flotation, Aust. Inst. Min. Metall.,
Melbourne (1938), pp. 163181.
25. N. N. Lyalikova and L. L. Lyubavina, On the possibility of
using a culture of Thiobacillus ferrooxidans to separate antimony
and mercury sulfides during flotation, Fundamentals and
Biohydrometallurgy (R. W. Lawrence, R. M. R. Branion, and H. G.
Ebner, Eds.), Elsevier, New York (1986), pp. 403406.
26. C. C. Townsley, A. S. Atkins, and A. J. Davis, Suppression of
pyritic sulfur during flotation tests using the bacterium Thiobacillus
ferrooxidans, Biotechnol. Bioeng., 30: 18 (1987).
27. Y. A. Attia and M. A. Elzeky, Biosurface modification in the
separation of pyrite from coal by froth flotation, Coal Science and
Technology, Vol. 9 (Y. A. Attia, Ed.), Elsevier, New York (1985),
pp. 673682.
28. T. Harada and N. Kuniyoshi, Effects of bacterial oxidation on
the floatability of pyrite, J. Min. Metall. Inst. Jpn. (1985), 101:
719724.
29. R. M. Bagdigian and A. S. Myerson, The adsorption of
Thiobacillus ferrooxidans on coal surfaces, Biotechnol. Bioeng.
(1986), 28: 467479.
30. M. Z. Dogan, G. Ozbayoglu, C. Hicyilmaz, M. Sarikaya, and
G. Ozcengiz, "Bacterial Leaching Versus Bacterial Conditioning
and Flotation in Desulfurization of Coal," XV International
Mineral Processing Congress, Cannes, June 29, 1985, pp. 304313.
31. F. Kargi and J. M. Robonson, Biological removal of pyritic
sulfur from coal by the thermophilic organism Sulfolobus
acidocaldarious, Biotechnol. Bioeng., 27: 4149. (1985).
32. Y. A. Attia and M. A. El Zeky, Bioleaching of non-ferrous
sulfides with adapted thiophilic bacteria, Chem. Eng. J., 44:
B31B40. (1990).
33. Y. A. Attia, J. Litchfield, and L. Vaaler, Application of
Biotechnology in the Recovery of Gold, Microbiological Effect on
Metallurgical Processes (J. A. Clum and L. A. Haas, Eds.), Met.
Soc., AIME, Warrendale, Penn., pp. 111120. (1985).
34. C. Livesey-Goldblatt, P. Norman, and D. R. Livesey-Goldblatt,
Gold Recovery from Arsenopyrite/Pyrite Ore by Bacterial
Leaching and Cyanidation, Recent Progress in Biohydrometallurgy
(G. Rossi and A. E. Torma, Eds.), Assoc. Mineraria Sarda, Italy,
pp. 627641 (1983).
35. A. Bruynesteyn, R. P. Hackl, and F. Wright, The
BIOTANKLEACH Process Gold 100, Proc. Int. Conf. Gold.,
Extra. Met. of Gold, Vol. 2, SAIMM, Johannesberg, p. 353 (1986).
Page 473
18
The Use of Nitrogen in Flotation
K. A. Kydros, K. A. Matis, and P. K. Spathis
I
Introduction
When one tries to define flotation, it is true that reference is made
to, among other things, the generation of gas bubbles in the
dispersion; in other words, the start is a pulp feed to the cell and the
result will be the creation of a froth of a certain height, bearing the
concentrate (i.e., the particulate matter that has floated by the
bubbles). It is the former stage that constitutes the scope of this
chapter. Certainly, the usual practice in flotation is the application
of air as the gas phase. A typical example, from the mineral-
processing point of view, is sulfide ore.
Nevertheless, reference is also required to the flotation collectors,
which are the most important class of chemical reagents used in the
process, and particularly to the way they adsorb on minerals.
According to one of the classifications available for collectors, one
of the categories (suitable for sulfides) forms the thio compounds,
usually containing the sulfhydryl or mercapto group; common
representatives are the so-called xanthates.
A simple and efficient method to investigate the surface reactions,
such as interfacial oxidation and collector or modifier adsorption in
a certain mineral, is the study of z-potential and its changes (1).
The determination of the electrokinetic behavior of a system gives
useful information on how the different ions operate. The surface
charge can also advise us in the proper selection of flotation
reagents.
Page 474
Contact angle measurements are perhaps the first indications of
flotation applied. The magnitude of contact angle is usually
correlated with the flotation efficiency (2). Nevertheless, a lot of
exceptions exist, arising from the experimental procedure applied,
during the determination of contact anglessince some of the
interactions, between various ionic and solid components of a real
flotation pulp, are ignored. Both z-potential and contact angle
measurements have been applied in the following.
It is now established that the interaction between sulfide minerals
and thio collectors takes place by a corrosion type, mixed potential
mechanism involving simultaneous electrochemical processes.
Sulfides, being electronic conductors, can act as a source of electrons
supporting electrode reactions at the surface. Such charge transfer
reactions are involved with the sulfhydryl collector adsorption, as
will be analyzed next (3).
A
Electrochemical Principles
The electrochemical interactions between xanthates and sulfides
were reviewed (4). The common cathodic half-reaction can be
represented as follows:
If MeS denotes the sulfide mineral and KEtX the applied potassium
ethylxanthate, as collector, some of the typical anodic reactions
follow:
1. chemisorption of the thio-ion
2. oxidation of the collector ion to dixanthogen
3. reaction with the sulfide giving an oxy-sulfur ion
or even sulfur
Reactions (4) and (5) can also occur in separate stages.

Each charge transfer reaction corresponds to a definite
electrochemical potential E related to the Gibbs free energy of the
process,
where F is the Faraday constant and z is the number of electrons

transferred per unit mole. An anodic reaction can occur at a mineral
surface, if a
Page 475
potential exists at which both this process and the cathodic
reduction of oxygen proceed at finite and equal rates; such a
potential is called a mixed potential (4).
The electrochemical aspects of a multielectrode system, which are
relevant to flotation, were also discussed (5), and the electrical
resistivities of pyrite and arsenopyrite, for instance, were presented.
They are 0.011000 and 0.0055 Ohm cm for pyrite ore and mineral,
respectively, and 0.110 and 0.03 Ohm cm for arsenopyrite ore and
mineral, respectively. The lower the resistances of the electrodes,
the higher the degree by which the multielectrode system is
determined by the polarization characteristics of the electrodes.
With a complex sulfide ore, the overall result of reactions depends
upon the properties of all the minerals and ions present in the pulp.
The aforementioned interactions contribute extensively to the
sulfhydryl-collector induced hydrophobicity of the sulfide
minerals. However, adsorption of such collectors also may take
place by ion exchange mechanisms of the xanthate ions on mineral
surface, with ions (SH-, OH-, SxOyn-) already abstracted by the
surface, as well as plain physical adsorption (6).
From the thermodynamic point of view, sulfide minerals are
unstable systems and tend to be transformed to more stable,
oxidized minerals. Therefore, the redox potential influences the
chemical state of their surface, which indirectly affects the
collector abstraction. So, in excessively oxidized surfaces the
electrode reactions are largely depressed, as the layer of oxides
eliminates the electron transfer (7).
1
Galvanic Interactions
The electrochemical interactions between the electrode surfaces are
often called galvanic. They are due to the different electrochemical
activity that the conducting solids present. The combination of a
cathodic surface (i.e., electron acceptor) and an anodic one
(electron donor) results in the creation of a galvanic cell. The
existence of a convenient oxidant, such as dissolved oxygen,
enhances the creation of a galvanic current, serving as the ultimate
electron acceptor. Such interactions can be hindered either by
lowering the activity of oxygen in water or by increasing the pH of
solution. The electrochemical activity of the minerals is
characterized by means of their rest potential (i.e., the potential
reached spontaneously by the mineral electrode in a given aqueous
solution). The mineral with the higher rest potential acts as a
cathode, while that with the lower rest potential is the anode.
During the flotation process, the galvanic interactions have a
significant influence as there is a mixing of solids with varying
electrochemical activity, such as sulfide minerals and metals
originated from the corrosion of grinding media (being iron alloys).
The latter are more anodic than the sulfides and, consequently,
affect the selectivity of the process.
Page 476
Figure 1 presents the mechanism of galvanic interactions and the
possible reactions during the electrolytic contact of two sulfides or
with grinding media (steel balls, rods, etc.). The hydroxyls that are
produced on the cathodic mineral result in the depression of
flotation, while the elemental sulfur that forms on the anodic sulfide
can increase its hydrophobicity and even lead sometimes to
collectorless flotation.
The rest potentials of sulfide minerals at pH 4 were reported (5).
Among all of them, pyrite has the highest rest potential of 0.66 V
versus SHE (standard hydrogen electrode). This mineral may be
regarded as the least electrochemically active or the most cathodic
one. On the contrary, arsenopyrite has a rest potential 0.39 V,
presenting high reactivity. Certainly, electrons flow from a less
cathodic mineral to a more cathodic one; pyrite thus serves as an
electron acceptor. The less cathodic mineral loses electrons; for
example, sulfide is oxidized to sulfur.
Generally, the galvanic interactions modify the mineral surface and
affect the ability of minerals to electrocatalyze the charge transfer
reactions. Therefore, they interfere with the flotation of sulfides by
sulfhydryl collectors.
Figure 1
Electrochemical models for galvanic interactions: (a) mineral/mineral
(b) mineral/grinding medium.

(From Ref. 5.)
Page 477
B
Application of Various Gases
The specific chemical effect of various gases during the process
has been studied in early works. This is an important topic for
research in flotation of minerals, and a number of studies were
devoted to the selective action of gases.
Oxygen, acting as an electron acceptor, contributes significantly in
the flotation of sulfides by sulfhydryl collectors, as already
mentioned. The topic was sufficiently reviewed (8). The presence
of nitrogen was found to shift the rest potential of pyrite electrode
to more cathodic values (9), whereas oxygen was found to maintain
the mixed potential of the system pyritexanthate to more anodic
values, favorable for the electrooxidation of xanthate to
dixanthogen. As a result, bubble contact could be obtained (10).
In a recent paper (11) from a workshop, the variations in the
oxidizing conditions of the pulp were studied as were their effects
on flotation for three types of natural sulfide ore (i.e., a complex
pyrite ore, a sphaleritegalena ore, and a chalcopyrite ore). Different
gas mixtures ranging from pure nitrogen to pure oxygen were used
to control the oxidizing conditions. The use of an oxygen-rich
flotation gas was proposed to be a suitable method for changing the
conditions and improving the metallurgical results.
In industrial froth flotation, the use of gases other than air has been
reported from time to time (12). Of all the gases, nitrogen would
appear to hold the greatest practical promise because of its ready
availability, as a by-product (from oxygen plants at the smelter
site), being also cheap to produce. Since it is chemically inert, it is
unlikely to be consumed by side reactions. It had been used in
chalcopyritemolybdenite separation, and a novel sphaleritepyrite
process was demonstrated on a complex ore from Canada.
Experiments were conducted also in pilot plant scale, both in
conventional cells and columns (13). Flotation of pyrite was
facilitated significantly when nitrogen was used as the flotation
gas. This selective activation of pyrite was due to the elimination of
the galvanic interactions, and an electrochemical mechanism was
proposed to explain the phenomenon.
Finally, preflotation aeration of sulfide pulps was practiced in
Canada for many years to depress pyrite and pyrrhotite (resulting
from excessive surface oxidation) and promote chalcopyrite and
galena (14). Investigations into the effects of aeration led to the
finding that the rate of oxygen consumption of a pulp was a
function of the susceptibility of mineral surfaces to oxidation. The
oxygen demand declines with time if sufficient aeration is
provided.
Page 478
Attention was also paid to electrolytic flotation (15). In the latter
technique, finely dispersed bubbles of hydrogen and oxygen are
generated as a result of electrolysis of the aqueous part of the
liquid. When sodium chloride is used as electrolyte, chlorine
evolution from the anode is also possible at certain current
densities.
Plaksin stated that gases play an important chemical role in
flotation (16). The behavior of minerals in particular depended on
the concentration of oxygen in solution. The effect of oxygen was
not limited to sulfide minerals but also entered into the behavior of
nonsulfides. The crystal structure of minerals was noticed to be a
major factor in their response to variations in oxygen level.
Elsewhere (17), the role of carbon dioxide in a calcite flotation
system was examined. This application was improved at low
sodium oleate concentration by the use of carbon dioxide instead of
air. Electrophoretic mobility studies were also carried out to obtain
an insight into how this gas served better as an activator, due to
increased electrostatic attraction of the collector anion.
The co-adsorption of paraffinic gases in the solidliquidgas system
and its effect on flotation recovery was finally investigated (18).
The latter increased when using ethane or butane for flotation of
quartz with dodecyl-ammonium chloride. The adsorption of these
dissolved hydrocarbon gases and the interfacial phenomena
associated with it were interpreted in terms of effective chain
length of collector.
Gases, particularly oxygen, contribute significantly in
collectormineral interactions as well as in systems other than
sulfide/sulfhydryl collector. The case was extensively reviewed by
Klassen and Mokrousov (19), as much of the work in the area is
due to scientists from the former USSR. On the influence of
dissolved gases in flotation, mention was given to:
1. change of the pH of the medium
2. adsorption of gases on the surfaces of mineral grains
3. oxidation of mineral surface
4. precipitation of gases from solution (occurring also in dissolved-
air flotation)
5. dehydration of the surface
6. change in adsorbed reagents
7. change in ionic content of pulp.
These phenomena are interrelated and a general scheme was
proposed (19).
Figure 2 shows some results of Plaksin and Chaplygina (19). It was
found that the flotation of many nonsulfide minerals was possible
to be improved by using oxygen as an activator and nitrogen as a
depressant.
Page 479
Figure 2
The effect of preliminary interaction with gases in the flotation of pure quartz by sodium
oleateactivation by copper sulfate.
(Courtesy of Butterworths, Oxford.)
Quartz and calcite were observed to have improved recoveries to a
considerable extent by oxygen.
Generally, it was suggested that oxygen activates the anionic flotation of
sulfides, oxides, or salt-type minerals, whereas nitrogen depresses it (19) (see
also Fig. 2). Although the role of oxygen and nitrogen in sulfide flotation is
well defined nowadays, this is not the case for the soap flotation of salt-type
and oxide minerals.
The activation of fluorite in oleate flotation under relatively high temperature
conditions has been reported. The effect of increased partial pressure of
oxygen was found to lead to similar activation effectiveness (20). Fluorite,
barite, and calcite exhibit some semiconducting behavior (21). Further
oxidation in the grinding or the conditioning stages was found to induce
surface semiconductivity of p-type and generally to reduce the Fermi level of
these minerals. This has a result of enhancing their surface properties as
electron acceptors, which aids the chemisorption of anionic collectors (22).
The presence of nitrogen, obviously, is related to the maintenance of the Fermi
level at high values. However, the collector speciation in oxidative conditions
should also be taken into account, especially in the case of nonconducting
minerals.
Page 480
II
Experimental
A bulk auriferous pyrite concentrate from the Olympias plant in
Chalkidiki (Greece) was the sample used in the investigation of
flotation, which was carried out in a modified Hallimond tube in the
Laboratory. Due to the interactions involved in the flotation pulp, it
was preferred during this floatability examination to have actual
mixture of minerals than pure samples. In this way, the selective
separation of pyrite/arsenopyrite by flotation was investigated.
The x-ray diffraction analysis showed that the concentrate contained
mainly pyrite, arsenopyrite, and some quartz. By chemical analysis,
it was found that it assayed about 43.7% Fe, 7.5% As, and 4.5%
insoluble matter and 21 g/ton Au. A positive relationship between
the gold and arsenic contents in the studied ore has already been
reported (23). Therefore, the separation of arsenic would be of
technological interest.
As flotation collector, potassiumethylxanthate was used at a
concentration of 25 mg/L. Both conditioning and flotation time were
600 s. The pH of the solution (220 cm3) was altered by sulfuric acid
or sodium hydroxide aqueous solutions. In tests involving nitrogen
conditioning, the wanted amount of distilled water was
deoxygenated by nitrogen bubbling for 600 s, prior to the addition of
the ore and reagents for conditioning. As flotation gas, either air or
nitrogen was applied (as stated). The results from the floatability
experiments and after the chemical analysis were calculated and
expressed as separate recoveries of pyrite and arsenopyrite.
Pure sulfide minerals, hand-selected from the mine area, were used
in the z-potential measurements conducted in a Rank Brothers Mark
II micro-electrophoretic apparatus. An arsenopyrite electrode was
also conducted from a pure monocrystal, and the redox potential
(Eh) of the flotation liquor was measured in response to a saturate
calomel electrode (SCE). The measured potentials were converted to
the SHE scale. The calculation of reversible potential of the xanthate
dixanthogen couple for the examined initial xanthate concentration
was done considering the following reaction (20):
As dixanthogen is practically insoluble in water, its activity was

assumed to be unity.
III
Although nitrogen does not interact either electrochemically or
chemically with some of the constituents of the flotation pulp, it acts
significantly influencing the behavior of sulfides during their
flotation (also nonsul-
Page 481
fides). This activity may be attributed to the ousting of dissolved
oxygen from the flotation liquor, which is successful when a
current of nitrogen is passed through it for a sufficient extent of
time. The following consequences can be considered as
responsible:
1. changes in the hydration state of mineral surfaces
2. effect of dissolved gases on the surface chemistry of minerals
3. use of nitrogen to maintain cathodic redox potentials
4. influence of nitrogen on galvanic interactions.
The adsorption of a gas on a surface is possible by physical
adsorption. The presence of a thin, gaseous film on a mineral
surface decreases its wetting characteristics and alters significantly
the further adsorption of flotation reagents. On the other hand, such
a gas film eases appreciably the attachment of a bubble on the
mineral (19).
Wetting results showed that the hydrophobicity of sulfide surfaces
was increased by the interaction of some gases, as well as that of
oxides and salt-type minerals. The effect of oxygen was more
intense compared to that of other gases tested. For the case of
sulfide minerals (examined in the present), the influence of oxygen
is particularly important.
The latter may be due to the partial oxidation of sulfides to
products like S or MeSx. Under certain conditions the sulfur reacts
with sulfide to form polysulfide species, or metal-deficient sulfides.
Any of these systems cause the mineral surface to become
hydrophobic to some degree and leads sometimes to collectorless
flotation (4). This should be distinguished from inherent
floatability exhibited by a few sulfides, which have layered
structure, such as molybdenite.
The interaction of a mineral like chalcopyrite with oxygen led to a
contact angle of the order of 85° (19), which is quite a large value.
The effect of nitrogen, under similar conditions, gave a contact
angle of only 55°. This example can clearly show that the gas used
influences, to a great extent, driving to activation (or not) of the
minerals.
Today, it is well known that sulfide minerals are unstable during
oxidation (24). The theory accepts the existence of an inner
hydrophobic layer, which is accessible in sulfur and covered by a
layer of metal hydroxy-oxides that makes the mineral hydrophilic
(25). This layer might be enough to prevent the adsorption of
collector (26).
In Figure 3 the z-potential of pyrite from two different areas is
presented for comparison. In the case of pyrite coming from Naica
(Mexico), a reversal of z-potential is noticed in the neutral pH
region (27,28). This reversal is apparently due to the presence of
metal ions from surface oxidation. To the contrary, the sample from
Chalkidiki (Greece) remained unoxidized, in these conditions,
which may be attributed to the varying microstructure of
Page 482
Figure 3
Zeta-potential measurements of pyrites against the pH of
solution: (a) sample from Mexico.
(Courtesy of Elsevier, Amsterdam.)
(b) sample from Greece.

mineral samples. The presence of admixtures (impurities)
differentiates the energy of crystal lattice and, consequently, the
mechanism of oxidation of sulfides, which follows different stages
(29). It seems that the sample origin and preparation assume
important roles.
Indeed, under conditions of absence of oxygen, Fuerstenau et al.
found that for the Mexican pyrite the z-potential reversal was not
appearing any more, as shown in Figure 4(a) (27,28). On the
contrary, the Greek pyrite
Page 483
Figure 4
The same comparison as in Figure 3: (a) the absence of
oxygen.
(Courtesy of Elsevier, Amsterdam.)
(b) The presence of 0.8 ppm Fe3+.
also had a characteristic z-potential reversal when ferric ions were

added to the pulp [see Fig. 4(b)].
This film of iron hydroxides, which was traced on pyrite,
influenced negatively flotation both by xanthates (30) and by
amines. When the dissolved oxygen is driven out from the flotation
liquor, nitrogen leads to the preservation of sulfide surface,
however, without assuring in all the cases discrete flotation
properties.
Dissolved oxygen (DO) is the most common oxidant, being an
unavoid-
Page 484
Figure 5
Contact angle measurements of pyrite and arsenopyrite as a function of pH (15 mg/L KEtX).
(Courtesy of Pergamon Press, Oxford.)
able component of all the flotation systems. It was reported (31) that during
grinding of a sulfide ore the level of DO was lowered from near saturation in the
feed to the rod mill (at around 8 mg/L of dissolved O2) to 0.10.5 mg/L in the mill
discharge. It was increased to around 4 mg/L in the classifier overflow (feed to
the flotation circuit) and, finally, reached the near-saturation point after 0.6 to 1.2
ks of conditioning or aeration in the cells. These measurements indicate that
oxygen dissolved in water is rapidly consumed during grinding whenever a large
extent of fresh sulfidic surface is produced. Oxidation of sulfides and of abraded
iron is primarily responsible for the consumption of oxygen (8).
The level of concentration of DO is a crucial parameter for selective flotation
(31). Because one of the constituents that affects the oxidation-reduction potential
is DO, it is clear that the alteration of its concentration can lead to changes of
potential that directly influence selective separation. The effect of DO on
adsorption of xanthate on pyrite was examined (19) as was the effect of redox
potential on pyrite floatability (4). Nevertheless, it was commented that the use of
Eh for control in commercial plants did not appear too promising (32), because
the difference in potential (corresponding to a few percent difference in flotation
efficiency) could well be quite small to be accurately differentiated from
background drifts.
In Figure 5 some contact angle measurements on pure pyrite and arsenopyrite (as
flat surfaces in the usual way) are given (24). In the alkaline
Page 485
region (pH 8.510.5) a significant differentiation was noticed from these monomineral
experiments, showing that a separation could be possible. No such indication, however,
was obvious from the flotation tests without the further application of modifiers.
Figure 6 shows the floatability of the concentrate as a function of pH; where the
recoveries are given separately calculated for pyrite and arsenopyrite, while the potential
of solution was measured under the same conditions with the arsenopyrite electrode. The
two minerals were found to be strongly depressed at pH values where the redox potential
was decreasing under the reversible potential of the couple xanthate/dixanthogen.
In Moscow Plaksin investigated the influence of oxygen (and oxidants) on selective
flotation of pyrite and arsenopyrite as a function of their crystal structure (16). It was
stressed that the two minerals had very similar flotation properties, although they differ
substantially in structure, as shown in Figure 7. The position of atoms of sulfur in pyrite
on the faces and edges of the lattice units makes them more accessible for the joining of
oxygen; whereas in arsenopyrite, the sulfur atoms are so dispersed that they are less
accessible to oxygen. The peculiarities of the crystal structure of pyrite and arsenopyrite
revealed the different influence of oxidation upon them. Elsewhere (29), it was further
reported that in this way it becomes clear why a low oxygen content in the aqueous
solution (e.g., 2.5 mg/L) was sufficient for 93% of the pyrite to pass into the froth during
flotation with xanthates, while arsenopyrite could hardly be floated at all.
Figure 6
Flotation experiments with the industrial concentrate; application of air: recoveries and Eh versus the
pH of solution.
Page 486
Figure 7
Crystal structures of: (a) pyrite, FeS2, and (b) arsenopyrite, FeAsS.
Nowadays it is generally accepted that pyrite is the most ''noble"
sulfide mineral (i.e., the less susceptible to oxidation) in contrast to
arsenopyrite, which is more active (26). Nevertheless, it was
further reported that arsenopyrite floats well by xanthates in more
intense oxidative conditions than pyrite.
This fact could be attributed to the different electrocatalytic activity
of the two minerals in oxygen reduction [Equation (1)]. Pyrite
presents a behavior similar to noble metals in oxygen reduction,
and it is the strongest electrocatalyst among the other sulfide
minerals, following the series (33):
This means that arsenopyrite demands more anodic potentials (i.e.,

more intense oxidative conditions) to balance the anodic oxidation
of xanthate to dixanthogen [see Equation (3)] in comparison to
pyrite. The oxygen consumption of an ore is usually determined
from measurements in the flotation cell with a dissolved oxygen
probe in process pulps. Nevertheless, it should be noted that, for
the system pyrite/ethyl xanthate, dixanthogen is not the sole species
responsible for the flotation behavior of the mineral (34,35). This
was the outcome of selected electrochemical studiesflotation and
infrared spectroscopic techniquescorrelated with thermodynamic
calculations.
Flotation has been also conducted, under experimental conditions
similar to those previously given, with the application of nitrogen
both in the conditioning and the flotation stages, as presented in
Figure 8. A decrease of Eh values is observed with the
corresponding depression of the two minerals. Also, an increased
selectivity is noticed, compared to Figure 6. Arsenopyrite generally
floats at higher potential values, and the results are in agreement
with the literature (12).
Hence, this suggests the possibility of controlling flotation
selectivity by
Page 487
manipulating the potential of the pulp. The pulp potential can be varied either by
chemical methods or by an external power source (4).
Chemical methods control the potential either through the activity of oxygen in
the pulp or by adding suitable oxidizing agents (to make the potential more
anodic) and reducing agents (to make the potential more cathodic). The activity of
oxygen in the pulp can be varied by the oxygen content of the flotation
gasnitrogen enrichment to decrease the activity and oxygen enrichment to raise it.
The practice of adding modifying agents (activators/depressants) to control
selective flotation was largely established prior to any electrochemical
interpretation of their role. It is true that several potential-modifying agents (for
instance, sulfide salts) are also pH regulators. In certain cases, both parameters are
critical and should be controlled.
The chemical nature of the added agents should also be taken into account. The
advantage of the application of nitrogen-enriched gas to maintain cathodic
potentials is that nitrogen is inactive from the chemical point of view and its
presence does not cause unpredictable interactions with the constituents of the
pulp. Its sole action is the purge of dissolved oxygen. Of course, in the presence of
nitrogen, the achievable potentials are not cathodic enough in several applications
of mineral processing (i.e., galena/sphalerite separation applying low cyanide
technology). In such cases, real reducing gases or salts such as SO2, Na2S, or
Na2SO3 should be used.
Figure 8
As in Figure 6, but with nitrogen bubbling instead of air, in conditioning and flotation stages.
Page 488
Another major consequence of the presence of nitrogen in a
flotation pulp is the depression of the galvanic interactions between
conductive solids with different electrochemical activities. This is
achieved by the elimination of oxygen, being a necessary electron
acceptor for such interactions to occur. In Figure 9, some results
concerning the further processing of the examined concentrate
applying nitrogen only in the flotation state are given. The
respective results achieved with the application of nitrogen both in
the conditioning and flotation stages are also given for comparison.
In the former case, arsenopyrite was floating the same in contrast to
pyrite, which was activated slightly by nitrogen flotation. Coming
back to the wetting results (Fig. 5), it may be assumed that the
reason for the nonexistent correlation between the conclusions
drawn from contact angle and floatability experiments is the
galvanic interaction, which of course cannot be carried out in the
case of wetting experiments. During galvanic interactions, pyrite
acts as the cathode (as it is the most cathodic sulfide), whereas
arsenopyrite acts as the anode. This results in the formation of a
hydroxide outer layer that depresses the flotation of pyrite. The
presence of nitrogen in the conditioning stage certainly eliminates
galvanic interactions but also prevents collector adsorption onto
pyrite surface, due to the absence of sufficient amounts of oxygen.
By the application of nitrogen only in the flotation stage, improved
separation for the system was achieved.
Figure 9
Effect of nitrogen to the pyrite concentrate flotation, when the
former applied: (i) only in the flotation stage and (ii) both in
conditioning and flotation.
Page 489
IV
Summary and Conclusions
Dissolved gases present in flotation pulps have a serious influence
on the process. The presence of oxygen, being adequate for the
flotation of sulfide minerals with anionic sulfhydryl collectors, is
well studied nowadays. Oxygen activation and nitrogen depression
were also observed during the soap flotation of salt-type, (some)
oxide, and sulfide minerals, without a clear interpretation being
obtained.
Electrochemical potentials measured in the pulp may be considered
to be indicators of the readiness of sulfide minerals to float, since
every charge transfer reaction, which is responsible for collector
adsorption, corresponds to an electrochemical potential. The results
of many laboratory investigations (but rather limited plant work) in
relation to the electrochemical interpretation of the flotation
process makes the technology promising for improving plant
performance. The pulp potential can be varied easily by various
chemical methods, including nitrogen bubbling for sufficient time.
By the application of nitrogen in the flotation stage, both the
adsorption of xanthate collector on pyrite and, in the meantime, the
interruption of the galvanic interactions was assured; the latter are
enhanced from the dissolved oxygen and may depress some
selective separations. It was observed finally that these interactions
were largely increased during the pulp aeration stage for the system
pyrite/arsenopyrite, which was examined as an example.
References
1. K. A. Kydros, "A study of floatability of pyrite, arsenopyrite and
their concentrates," Ph.D. Thesis, Aristotle University,
Thessaloniki (1992).
Technology (P. Somasundaran and B. M. Moudgil, Eds.), Marcel
Dekker, New York (1988), p. 39.
3. J. M. Cases and F. Villieras, The mechanisms of collector
adsorptionabstraction (ionic and non-ionic surfactants) on
heterogeneous surfaces, Innovations in Flotation Technology (P.
Mavros and K. A. Matis, Eds.), Kluwer Academic, Dordrecht
(1992), p. 25.
4. S. R. Rao, G. Labonte, and J. A. Finch, Electrochemistry in the
Matis, Eds.), Kluwer Academic, Dordrecht (1992), p. 57.
5. D. Kocabag and M. R. Smith, The effect of grinding media and
galvanic interactions upon the flotation of sulfide minerals,
Complex SulfidesProcessing of Ores, Concentrates and by
Weseley, Eds.), AIME & Can. IMM, San Diego (1985), p. 53.
6. G. W. Poling, Reactions between thiol reagents and sulfide
minerals, Flota-
Page 490
tionA. M. Gaudin Memorial Volume (M. C. Fuerstenau, Ed.),
AIME, New York (1976), p. 334.
7. M. J. V. Beattie and G. W. Poling, A study of the surface
oxidation of arsenopyrite using cyclic voltametry, Int. J. Miner.
Process., 20: 87 (1987).
New York (1982), p. 619.
9. A. H. Usul and R. Tolun, Electrochemical study of the
pyriteoxygenxanthate system, Int. J. Miner. Process., 1: 135
(1974).
10. H. Majima and M. Takeda, Electrochemical studies of the
xanthatedixanthogen system on pyrite, Trans. AIME, 241: 431
(1968).
11. G. Berglund, Pulp chemistry in sulfide mineral flotation, Int. J.
Miner. Process., 33: 21 (1991).
12. C. J. Martin, S. R. Rao, J. A. Finch, and M. Leroux, Complex
sulfide ore processing with pyrite flotation by nitrogen, Int. J.
Miner. Process., 26: 95 (1989).
13. C. J. Martin, J. A. Finch, and S. R. Rao, Pilot flotationColumn
tests of nitrogen flotation of pyrite in pyrite/sphalerite separation;
Trans. IMM, 99: C1154 (1990).
14. K. V. Konigsmann, Flotation techniques for complex ores,
Complex SulfidesProcessing of Ores, Concentrates and By
Weseley, Eds.), AIME & Can. IMM, San Diego (1985), p. 5.
15. K. A. Matis, "A study of electrolytic flotation and its
application to the treatment of industrial liquid wastes," Ph.D.
Thesis, University of Newcastle-upon-Tyne, UK (1977).
16. I. N. Plaksin, Interaction of minerals with gases and reagents in
flotation, Trans. AIME, 214: 319 (1959).
17. V. Y. S. Kumar, N. Mohan, and A. K. Biswas, Fundamental
studies on the role of carbon dioxide in a calcite flotation system,
Trans. AIME, 250: 182 (1971).
18. I. J. Lin and A. Metzer, Effects of dissolved hydrocarbon gases
on froth flotation, Advances in Interfacial Phenomena of
Particulate Solution/Gas Systems; Applications to Flotation
Research (P. Somasundaran and R. B. Grieves, Eds.), AIChE, New
York (1975), p. 75.
Theory of Flotation, Butterworths, London (1963), p. 207.
20. M. C. Fuerstenau, J. D. Miller and M. C. Kuhn, Chemistry of
Flotation, AIME, New York. (1985).
21. M. Carta, R. Ciccu, C. Delfa, G. Ferrara, M. Ghiani, and P.
Massacci, "Improvement in Electric Separation and Flotation by
Modification of Energy Levels in Surface Layers," 10th Intl. Min.
Proces. Congr., IMM, London, paper 3 (1973).
22. I. N. Plaksin and R. Sh. Shafeev, Influence of surface properties
of sulfide minerals on adsorption of flotation reagents, Trans. IMM,
72: 715. (1963).
pyrite concentrate by flotation reagents, Miner. Eng., 5(35): 331.
(1992).
Page 491
flotation of complex CuPbZn sulphides, Innovations in Flotation
25. A. N. Buckley, I. C. Hamilton, and R. Woods, Investigation of
the surface oxidation of sulphide minerals by linear potential sweep
voltammetry and X-ray photoelectron spectroscopy, Flotation of
Sulphide Minerals (K. S. E. Forssberg, Ed.), Elsevier, Amsterdam
(1985), p. 41.
26. G. W. Poling and M. J. V. Beattie, Selective depression in
complex sulphide flotation. Principles of Mineral FlotationThe
Wark Symposium (M. H. Jones and J. T. Woodcock, Eds.),
Australasian IMM, Melbourne (1984), p. 137.
27. M. C. Fuerstenau, C. A. Natalie, and R. M. Rowe, Xanthate
presence of oxygenPart 1, Int. J. Miner. Process., 29: 89 (1990).
28. M. C. Fuerstenau, M. Mishra, and B. R. Palmer, Xanthate
presence of oxygenPart 2, Int. J. Miner. Process., 29: 111 (1990).
29. V. A. Glembotskii, V. I. Klassen, and I. N. Plaksin, Flotation,
Primary Sources, New York (1972).
30. M. C. Fuerstenau, M. C. Kuhn, and D. A. Elgillani, The role of
(1968).
31. A. N. Beyzani and L. P. Kitshen, "Selective Flotation of
Sulfidic Complex Ores with Special Reference to the Interaction of
Specific Surface, Redox Potential and Oxygen Content," Proc. XVI
Int. Miner. Process. Congr. (K. S. E. Forssberg, Ed.), Elsevier,
Amsterdam (1988), p. 307.
32. R. Woods, Electrochemistry of sulfide flotation, FlotationA. M.
Gaudin Memorial Volume (M. C. Fuerstenau, Ed.), AIME, New
York (1976), p. 298.
33. D. A. J. Rand, Oxygen reduction on sulfide minerals. Part III.
Comparison of activities of various copper, iron, lead and nickel
mineral electrodes, J. Electroanal. Chem., 83: 19 (1977).
34. D. W. Fuerstenau and R. K. Mishra, On the mechanism of
pyrite flotation with xanthate collectors, Complex Sulfide Ores (M.
J. Jones, Ed.), IMM, London (1980), p. 271.
35. X.-H. Wang and K. S. E. Forssberg, Mechanisms of pyrite
flotation with xanthates, Int. J. Miner. Process., 33: 275 (1991).
Page 493
19
Dissolved-Air Flotation of Fine Particles of Salt-
Type Minerals
G. P. Gallios, K. A. Matis, and Th. N. Balabanidis
I
Basic Principles
Dissolved-air (or pressure) flotation uses bubbles of air generated
under controlled conditions by pressure release, from a super-
saturated solution of air in water. The whole technique is based on
Henry's law, which gives the solubility of air in water: the mass of
a slightly soluble gas that dissolves in a definite mass of a liquid (at
a given temperature) is, very nearly, directly proportional to the
partial pressure of that gas (1).
Vacuum may also be used to decrease the pressure after first
saturating water at atmospheric pressure; however, the latter
technique (so-called vacuum flotation) is avoided due to scale-up
difficulties.
There are generally three ways of pressurization, which are
represented schematically in Figure 1:
1. Total pressurization of the feed (around 300 kPa)
2. Partial pressurization of the feed (around 500 kPa)
3. Partial pressurization of the effluent (at the previous case) and
recycle
The last method is preferred to avoid problems in the saturator
from feed suspended particles or microbial growth. If full
saturation is achieved, the quantity of air dissolved in the pressure-
saturated water is proportional ideally to the pressure in the
saturator. In laboratory work, one of the aims
Page 494
Figure 1
Schematic representation of the various
ways of pressurization.
was to measure the air volume produced from the saturation tank
and compare it with the theoretically calculated air volume, so a
suitable apparatus was built, as shown in Figure 2. The apparatus
consists of two interconnected glass vessels. Water saturated with
air at a relatively high pressure (usually 400500 kPa) is allowed to
flow through a pressure-reducing device (A) and then into the
container, while at the same time water is displaced from the
second vessel having valve C slightly opened, so that the pressure
remains constant. The volume of water displaced from the second
vessel is equal to the volume of water-plus-air added to the
container, i.e., the difference of the two measured water volumes in
the cylinders gives the released air volume.
Much research has been carried out on the saturator design (2).
Possible methods of air dissolution (Fig. 3) include among others
packed columns and sparged air or pump suction air injection
systems. Bratby and Marais (3) proposed a packed column system,
which was markedly superior in
Page 495
Figure 2
Experimental setup for dissolved-air volume measurement.
performance; air was introduced at the top of the saturator and the
material was Raschig rings. To monitor the process and its
performance it is usually not only a matter of the mass of air
precipitated per volume of saturator feed. The bubble size
generated is also of interest, consisting of an important parameter
in effluent treatment and fines processing.
For instance, the specific gravity of flocs produced in water and
waste treatment is often similar to that of water; therefore, only
very small air bubbles (being of 20120 µm) are required to float the
flocs to the surface. The presence of a frother or surfactant in the
pulp lowers the surface tension, hence the bubbles' size. Many
methods, among them several estimation techniques, particularly in
flotation columns, have been used for bubble diameter
measurement, which are generally needed for design purposes. The
latter are certainly less tedious and applicable to slurries, but are
rather approximate. The most common approach is some
photographic technique. Figure 4 shows a suitable arrangement
applied to this sort of size range.
A large number of dissolved-air flotation plants, treating potable
water and liquid wastes, for example, from oil refineries (4), are
currently operating in Europe. The majority of them are utilizing a
circular flotation cell design; such a unit is presented in Figure 5.
The process is used generally as a secondary waste water
treatment, the primary usually being gravity
Page 496
Figure 3
Designs of saturation tanks: (a) static designs
(b) dynamic designs.

sedimentation in large clarifiers operating under low-velocity
gradients. After the primary treatment, waste water passes into a
flocculation tank, into which various chemical reagents for
coagulation are dosed in proportion to the flow rate and
concentration.
From the flocculation tank, the waste water containing the flocs
flows under quiescent conditions into the flotation unit. Usually,
below the inlet point the pressurized recycle water is introduced
into the cell through a pressure-reducing valve, followed by a pipe
fitted with nozzles; effort was devoted to evaluate the proper inlet
design (1). Downstream of the valve the pressure is virtually
atmospheric, and air commences to precipitate out of the solution
as the solubility is decreased. This air/water mixture is blended
Page 497
Figure 4
Experimental setup for photographic measurement of bubble size in a flotation vessel.
with the waste water stream. If attachment of the precipitating air bubbles to
the suspended matter is possiblefor instance, if the particles are
hydrophobicthe agglomerates due to gravity then rise to the top of the
flotation chamber and accumulate as a float. The floated layer (including
solid particles, oil droplets, etc.) is removed to off-take points by a certain
scrapping mechanism, while the clarified liquid is generally suitable for
final discharge.
II
Fine Mineral Particles
An investigation of salt-type mineral flotation may be interesting both from
the practical and theoretical point of view. The best example is perhaps
Page 498
Figure 5
Dissolved-air flotation unit.
apatites, which constitute the greater amount of a raw material
produced by flotation techniques (in the United States and other
countries). The selective separation of this group of minerals is
difficult, due to similar behavior of them toward flotation and
problems with the increased solubility, for instance, in the pH
control (5). Often modifiers are applied for this reason (6). But the
main difficulties arise with crypto-crystalline-type deposits (e.g.,
magnesite, which exists in Greece) due to the creation of a large
amount of fines. This was the initiative of the present work.
Among the salt-type minerals, magnesite MgCO3 and dolomite
(Ca,Mg)(CO3)2 (called magnesium carbonates) are comparatively
less studied and have been investigated mainly by Predali (7),
Brandao (8), and Gallios (9). Dolomite is often a gangue mineral of
apatites.
Ideally, one would like fine particles to be produced only for the
purpose of liberation of valuable minerals. However, millions of
tons are produced each year mainly from size reduction operations
and disposed of (i.e., in the Florida phosphate industry) after
classification in the mine area, their processing being inefficient or
uneconomical. Parsonage (10) recognized that in the salt-type
mineral/oleate systems and due to density reasons, the depression
of slime coatings was markedly increased. Additionally, because
many current ore reserves have lower grade and increasingly
complex mineralogy (i.e., sulfides), technology has to find
solutions for the processing of fines. Such flotation techniques as
dissolved-air flotation were suggested (11) as worth development
particularly for the beneficiation of sulfide fines.
Page 499
To the authors' knowledge, rather little work has been published on
fines processing by this technique. Vacuum flotation (similar in
principle to dissolved-air) was at first reported for the processing of
fluoritequartz and baritequartz mixtures (12). Dissolved-air
flotation was applied, in the separation of fine coal from coal
washery waste water (13) and also of cassiteritequartz mixtures
(14).
Flotation of minerals is usually performed in agitated cells. This
may prove inconvenient when treating fine particles, especially
when entrainment of gangue or loss of valuable mineral is not
desired. Also, the solubility problem is enhanced with the finer
range. The significance of particle size and its consequences in
flotation was thoroughly examined by Somasundaran (15) and
Fuerstenau (16). Although the maximum flotation rate is generally
achieved at the intermediate size range, a number of examples do
exist where ores have been successfully floated at fine sizes.
Klassen and Mokrousov (12), examining the tenacity of mineral
particle attachment to bubbles, mentioned the old work of Volkova
(since 1940). It seems that she was the first scientist who studied
the role of particle/bubble size in flotation. Reviewing the
literature, Szatkowski and Freyberger (17) found that the flotation
rate constant was proportional to the particle diameter to a power
of 12, but inversely proportional to the bubble diameter to a power
of 2.53. Ahmed and Jameson (18) also concluded that the flotation
rate of fine particles was accelerated considerably by using very
small bubbles, of less than 100 µm.
Conventional froth flotation cells, nevertheless, generally produce
air bubbles in the size range of 0.6 to 1.0 mm. Matis and Gallios
(1) reviewed two unconventional flotation techniques for mineral
processingdissolvedair and electrolytic flotation, which is
characterized by the production of fine gas bubbles and quiescent
conditionsas possible candidates to mineral processing of fines;
both have their origin in effluent treatment. The first experimental
results on the former technique, which sounds economically more
attractive, were already published (19).
III
Experimental
The pure minerals used in the batchwise experiments were
magnesite from the Yerakini area (Chalkidiki) and dolomite from
the Neraida area (Kozani), both in Northern Greece. It is noted that
the latter, however, was crystalline. Crypto-crystalline deposits of
magnesite exist in many places all over the world, but in rather
limited amounts for utilization and with a low content in valuable
mineral. Among the major magnesite-producing countries is
Greece.
The chemical analysis of the minerals (which were identified by x-
ray
Page 500
diffraction analysis) is presented in Table 1. The specific surface
was 3.89 m2 g-1 for the magnesite and 0.5147 m2 g-1 for the
dolomite. A subsieve particle size range has been selected for this
work, namely, -45 µm, prepared in the laboratory after crushing
and wet grinding. The very fine particles of -15 µm were discarded
by washing and decantation. The proposed size classification by
Sivamohan and Forssberg (20) is accepted and followed in the
present, giving as fines the particles less than 100 µm in size.
The main volume of experiments in this paper was carried out by
dissolved-air flotation (DAF). In this process, usually a part of the
cleaned effluent from the flotation unit is saturated with air and
then mixed with the feed (to be processed), hence recycled. So, the
recycle ratio, expressed in percent, is traditionally usedcoming
from the chemical reactor theoryand meaning the volume of fluid
returned to the reactor entrance over the volume leaving the
system.
The dissolved-air flotation bench-scale procedure was described
previously (19) as was the modified type of Hallimond tube (6).
Conventional froth flotation tests were conducted in a 4-L Denver
laboratory flotation cell. Conditioning time was 600 s, mixing at
about 850 rpm and air flow rate of 83.3 cm3 s-1.
The anionic collector applied in the flotation experiments was a
commercial fatty acid, Acintol FA-1 tall oil, kindly supplied by
Arizona Chemical Co. (USA). As this was not a pure reagent, the
ppm expression of concentration was adopted.
The modifying reagents used were the following: fluosilisic acid,
carboxymethyl cellulose (as the sodium salt), sodium silicate,
tetrasodium pyrophosphate, and sodium hexametaphosphate. The
pH of the solution (in deionized water) was adjusted with sulfuric
acid or sodium hydroxide solutions. The pH value recorded
hereafter is that before flotation (and after conditioning). Finally,
the experimental conditions applied with the modifi-
Table 1 Chemical Analysis of Minerals
Used.
Dolomite (%) Magnesite (%)
MgO 22.81 47.77
CaO 29.84 0.35
R2O3a 0.18 0.12
SiO2 0.25 0.25
LOIb 46.92 51.51
aMixed oxides.
bLoss on ignition.
Page 501
ers were already established (21) (6) and are not within the scope
of this chapter.
IV
A
The New Technique
Although there are certain similarities between the fields of
conventional mineral beneficiation and effluent treatment, as far as
flotation is concerned, several differences, which are presented in
Table 2, should be kept in mind. It is true that a large volume of
experimental work on the role of particle/bubble size has its origin
in the application of flotation in water and waste water purification
(1). In mineral-processing practice, it is difficult to reduce the
bubble size without affecting other variables in the flotation
system. Perhaps it is for this reason that the published studies
concerning the effect of bubble size on the industrial flotation of
fine ore pulps are rather limited.
Initial work involved experimentation with the conventional
laboratory mechanical cell (Denver) and comparison with
dissolved-air flotation. Figure 6 presents some of the results in this
series. The difference in scale was 1:4 initially (before the entering
of recycle), but the pulp density could be kept the same (3% is this
case). As the particle size is reduced, the solid surface to be
conditioned is increased; this reflects to optimum pulp density.
However, from preliminary work, a rather low pulp density was
found to be necessary, in the case of magnesite fines, for the
conventional froth flotation cell (around 5%, compared to 25%
usually used in the coarser sizes). As has been confirmed
experimentally, the classical cell was working rather as a froth
pump with more than 5% solids (magnesite) and various collector
concentrations.
Table 2 Differences of Flotation from the Point of
View of Fine Processing.
Conventional Effluent
mineral processing treatment
Concentration 1540% ppm range
(pulp density)
Particle size 25500 µm colloidal
Mixing intense, continuous unnecessary
Separation selective S/L
Economy attractive unattractive
Page 502
Figure 6
Effect of collector concentration (noted as Ccol) on
flotation recovery of (a) dolomite
(b) magnesite: comparison of DAF, with 20%

recycle, conventional froth flotation, 3% pulp
density, and pH 10.0.
Page 503
The present experiments were carried out without any modifiers
added, except for pH control. Previous work (9) has indicated that
for a simple binary mixture satisfactory separation of magnesite
from dolomite can be effected by reverse flotation, using only
marginal amounts of sodium oleate collector and sodium hydroxide
as pH modifier (22).
Dolomite floated better with the conventional cell, having a 90%
recovery with 400600 ppm collector. In dissolved-air flotation
(with 20% recycle), the maximum recovery was around 60%.
Increasing the collector concentration above 40 ppm decreased the
recovery to 20%. A further increase of collector (up to 200 ppm,
not shown in Fig. 6) in dissolved-air flotation was noticed to
worsen the recoveries (about 15%). Magnesite recoveries were
lower (3040%) with both cells. Nevertheless, an increase of recycle
to 50%, presented in Figure 7, improved dolomite recoveries in
dissolved-air flotation (about 80%). Magnesite recovery increased
to 45% with dissolved-air flotation, and it was a little better than
with the conventional cell. Approximately ten times more collector
consumption was seen than was necessary for the conventional
cell, showing increased operational costs.
The aforementioned comparison justified the following work on
dissolved-air flotation. As shown in Figure 8, a 5% (in solids)
density was observed to give reasonable results in dissolved-air
flotation, mainly for dolomite (80% recovery), with an initial cell
content of 1000 cm3 (benchscale). The 8% pulp density worsened
the flotation recovery (Fig. 8) of both minerals. A 50% recycle was
applied here for bubble generation, as used earlier (19).
This was considered rather unrealistic, from the experience in
effluent treatment (1) and was compared to a 20% recycle, which is
presented in Figure 9; the latter was, however, reducing the
recoveries, because the decreased air volume introduced in the pulp
was not enough to carry all the fine particles to the surface. At a
300-ppm collector concentration and 50% recycle (not shown in
the figure), dolomite gave approximately the same recovery, while
magnesite floated about 70% (from 64% with 200 ppm). In these
experiments, pyrophosphate was also present, being found to affect
mainly dolomite, but at pH greater that 9 (21). At pH 9.5 also a
reduction in collector adsorption was noticed (23). Nevertheless, at
pH 7.1 (as shown in Fig. 9) the modifier did not affect dolomite
and slightly activated magnesite.
A close examination under the microscope revealed numerous
fines, adhered to larger particles of magnesite, which were
themselves pitted with many cavities. Inspection of the dolomite,
on the other hand, showed, as expected, that the dolomite was free
from cavities and fines. A photograph was recently presented (24).
Page 504
Figure 7
Comparison of (a) dolomite and(b) magnesite
(as in Fig. 6), but 50% recycle for
dissolved-air flotation and both with 5% pulp density.
Page 505
Figure 8
Dissolved-air flotation of (a) dolomite
(b) magnesite (at 50% recycle and pH 10.0); effect

of collector concentration on flotation recoveries
at two different pulp densities: 5 and 8%.
Page 506
Figure 9
(b) magnesite (3% pulp density, pH 7.1, 600 ppm

pyrophosphate); study of recycle at 20 and 50%.
Page 507
Figure 10
(b) magnesite by a 10-ppm collector, with and without

100-ppm sodium silicate; effect of pH at 50% recycle
and pulp density 3%.
Page 508
Figure 11
(b) magnesite with and without 20-ppm sodium

hexametaphosphate (SHMP); effect of collector
concentration at 3% pulp density, 20% recycle,
and pH 10.0.
Page 509
Figure 10 shows the experiments in the presence of sodium silicate,
a usual modifier; this undergoes a number of hydrolysis reactions.
Under the applied conditions, the modifier did not affect flotation.
Dolomite recoveries were quite good (around 80%), while
magnesite's was only approximately 40%, in the whole pH range
investigated. In contrast, the same collector concentration of 10
ppm, in Hallimond tube tests by sodium oleate, but at a coarser size
gave a distinct minimum (of 20%) for both minerals at pH 89, with
greater recoveries mainly toward the acidic region (9).
The effect of hexametaphosphate (20 ppm) on the dissolved-air
flotation of minerals is shown in Figure 11. Dolomite (which had a
smaller specific surface) was observed to appear decreased
recoveries in increased collector concentrations (from about 50 to
100 ppm) and without any modifier. This was possibly due to the
formation of a second hydrophilic layer on particles.
Hexametaphosphate is known to adsorb first on a salt-type mineral
surface. From z-potential measurements it has been found that the
modifier increases the negative charge of dolomite (21). For a
coarser size, hexametaphosphate was found to depress mainly
dolomite. However, in this case, the hexametaphosphate stabilized
the recovery of dolomite (to about 50%) with increased collector
concentrations. It did not affect the recovery of magnesite.
Today many flotation techniques, other than the conventional, have
been investigated and proposed as capable for mineral fines
processing, because of the losses of mineral and metal values in the
fine size range; among them are the carrier flotation, oil flotation,
flocculation/flotation, flotation at high temperature, and electrolytic
flotation. It is believed that promising results were obtained in the
present by dissolved-air flotation. As the two studied minerals have
been found to appear different floatabilities under specified
conditions, this technique possibly could be used for selective
flotation.
B
Comparison of Flotation Techniques
It was proposed (18) that serious consideration should be given to
devising separate flotation circuits, designed specifically to handle
the fine size ranges. Something similar (i.e., processing at close
size ranges) is applied at other gravity separation methods, for
example, the shaking tables.
In laboratory work on calcite, a conventional bench-scale flotation
cell was compared to the modified Hallimond tube, being (as
known) a classical microcell; quite comparable results were
obtained (24). The latter is generally a fast apparatus needing small
samples, of around 2 g, for experimentation, and it operates with a
low pulp density. Useful conclusions on the floatability of mainly
pure minerals can be extracted. However, a definite
Page 510
Figure 12
Comparison of dissolved-air flotation (20% recycle)
with dispersed-air flotation (Hallimond
tube)30-ppm collector and 100-ppm
carboxymethyl cellulosefor (a) dolomite
(b) magnesite.
Page 511
Figure 13
Comparison (as in Fig. 12) of (a) dolomite
(b) magnesite in the presence of 100-ppm fluosilisic

acid; the collector was 50 ppm in dissolved-air
flotation and 30 ppm in the Hallimond.
Page 512
Figure 14
Comparison (as in Fig. 12) of (a) dolomite
(b) magnesite in the presence of pyrophosphate,

but 50% recycle in dissolved-air flotation with pH
7.1 and a 40-ppm collector.
Page 513
conclusion on the required collector concentration is not easy, and
carefully controlled conditions are also necessary. For instance, the
solubility of salt-type minerals might be a problem, i.e., in pH
control at long conditioning times.
In the present, dissolved-air Rotation was compared to Hallimond
tube (220 cm3 content) experiments. In this microcell special care
was taken to prevent mechanical transport and entrainment of
ultrafines in the froth layer; the vertical part had a height of 265
mm. Figure 12 shows the results in the presence of carboxymethyl
cellulose. It is noted that the dissolved-air flotation cell had a pulp
density of 3%, while the Hallimond had a pulp density of 1.8%.
This means that the latter had more collector per unit of solids.
Both minerals floated better with the Hallimond microcell, but the
recoveries were generally low. At pH values over 10, dolomite was
floating at 60% (better recovery). At alkaline pH, carboxymethyl
cellulose is known to be completely ionized and possesses a high
negative charge.
A drop of flotation recoveries in the presence of carboxymethyl
cellulose was also reported (19). This was explained by the
reduction found in adsorption on dolomite. In the meantime, the z-
potential of magnesite was unchanged (21).
Another comparison of dissolved-air flotation with Hallimond, in
the presence of fluosilisic acid as a modifier, is given in Figure 13.
There was a different amount of collector used this time in the two
cells to equilibrate the difference in pulp densities. Again, both
minerals floated better with Hallimond, but the maximum
recoveries were also low. With sodium fluorosilicate a considerable
depression of dolomite was found (6).
In contrast, dissolved-air flotation was observed to operate much
better than Hallimond for dolomite with a 40-ppm collector
(around 80% recovery) in the presence of pyrophosphate, as shown
in Figure 14. It is noted that a 50% recycle was used, in this case,
for dissolved-air flotation. A drop in magnesite recoveries was also
noticed. This modifier was seen to affect markedly the z-potential
of salt-type minerals and so the flotation behavior showed
selectivity (19).
Another point is the low recovery of magnesite in the conventional
cell (unless an inverse flotation is envisaged). The disadvantages of
the conventional cell in fine flotation (Figs. 6 and 7) can be
explained by the observed production of very fine particles during
mixing, mainly with the cryptocrystalline mineral. It is known that
most mechanically agitated cells produce high turbulence in the
region of the impeller. The cells should be agitated at the minimum
speed needed for complete suspension of the mineral particles
(valuable and gangue), to possibly obtain the best separation.
From this comparison of flotation techniques, it can be said that
each
Page 514
flotation system must be studied individually and that no general
conclusions should be drawn, as it seemed to depend on the system
but mainly on the conditions applied. Nevertheless, useful
observations were made for the operation of different flotation
cells, noting the difference in the collector amounts required.
V
Conclusions
Generally, it can be observed that:
1. Dissolved-air flotation, being a technique for producing fine
bubbles (<120 µm) and dominant already in effluent treatment,
may be suitable for the processing of particles at the subsieve size
range.
2. The two minerals investigated, although of the same type (salt-
type minerals), did not behave similarly as far as flotation is
concerned. Dolomite (here, mineralogically crystalline) was the
more amenable to processing, giving high flotation recoveries at
this size range. The presence of modifying reagents could
significantly depress its floatability, as expected.
3. The existence of very fine particles presents difficulties in
flotation. These can be created even during the conditioning (or
mixing), particularly in the case of crypto-crystalline minerals
(magnesite in the present) being friable. Due to salt-type mineral
solubility, this constitutes a serious problem for flotation, if
selectivity is required.
4. Optimum conditions for dissolved-air flotation were found to be
3% solids, pH 7.1, 50% recycle, 200-ppm collector, 600-ppm
pyrophosphate. Recoveries were dolomite = 90% and magnesite =
63% (see Fig. 9).
Acknowledgments
Experimental collaboration of chemistry students Th. Nikopoulos
and G. Thymiatis is gratefully acknowledged.
References
W. Barley, Eds.), Nijhoff, Dordrecht (1986), pp. 3770.
2. Th. F. Zabel, Flotation in water treatment, Innovations in
Flotation Technology (P. Mavros and K. A. Matis, Eds.), Kluwer
Academic, Dordrecht (1992), pp.431454.
Page 515
3. J. Bratby and G. v. R. Marais, Flotation, Solid/Liquid Separation
Equipment Scale Up (P. B. Purchas, Ed.), Uplands Press, Croydon
(1977), pp. 155198.
4. K. A. Matis and P. Mavros, Recovery of metals by ion flotation
from dilute aqueous solutions, Sep. Purif. Methods, 20: 148 (1991).
5. H. S. Hanna and P. Somasundaran, Flotation of salt-type
minerals, FlotationA. M. Gaudin Memorial Volume (M. C.
Fuerstenau, Ed.), AIME, New York (1976), pp. 197272.
6. K. A. Matis and G. P. Gallios, Anionic flotation of magnesium
carbonates by modifiers, Int. J. Miner. Process., 25: 261274 (1989).
7. J. J. Predali, Adsorption des Sels d' Acides Gras a l' Interface
Solide-Liquide et Flottation des Carbonates, Sciences de la Terre,
Memoire 27 (1973).
8. P. R. G. Brandao, ''Adsorption of oleate on magnesite and its
influence on flotation," Ph.D. Thesis, University of British
Columbia, 1982.
9. G. P. Gallios, "A study of the behavior of magnesium carbonate
minerals during their flotation," Ph.D. Thesis, University of
Thessaloniki, 1987.
10. P. G. Parsonage, Effect of slime and colloidal particles on the
flotation of galena, Flotation of Sulfide Minerals (E. Forssberg,
Ed.), Elsevier, Amsterdam (1985), pp. 111139.
their flotation, Complex Sulfide Ores (M. J. Jones, Ed.), IMM,
London (1980), pp. 118127.
Theory of Flotation, Butterworths, London (1963), pp. 129, 412.
13. J. Shimoiizaka and I. Matsuoka, "Applicability of Air-
Dissolved Flotation for Separation of Ultrafine Mineral Particles,"
Proceedings 14th Int. Miner. Process. Congr., Toronto, preprint
(1982).
14. R. J. Gochin and J. A. Solari, Dissolved-air flotation for
recovery of fine cassiterite, Trans. IMM, 92: C52C58 (1983).
15. P. Somasundaran, An overview of the ultrafine problem,
Mineral Processing at a Crossroads (B. A. Wills and R. W. Barley,
Eds.), Nijhoff, Dordrecht (1986), pp. 136.
16. D. W. Fuerstenau, Fine particle flotation, Fine Particles
Processing (P. Somasundaran, Ed.), AIME, New York, pp. 669705
(1980).
17. M. Szatkowski and W. L. Freyberger, Kinetics of flotation with
fine bubbles, Trans. IMM, 94: C61C70 (1985).
18. N. Ahmed and G. J. Jameson, The effect of bubble size on the
rate of flotation of fine particles. Int. J. Miner. Process., 14: 195215
(1985).
19. K. A. Matis, Th.N. Balabanidis, and G. P. Gallios, Processing
of magnesium carbonate fines by dissolved-air flotation. Colloid
Surf., 29: 191203 (1988).
20. R. Sivamohan and E. Forssberg, Recovery of heavy minerals
from slimes, Int. J. Miner. Process., 15: 297314 (1985).
21. G. P. Gallios and K. A. Matis, Floatability of magnesium
carbonates by sodium oleate in the presence of modifiers, Sep. Sci.
Technol., 24: 129143 (1989).
Flotation of magnesite and dolomite by fatty acids, Trans. IMM,
98: C99C105 (1989).
Page 516
23. K. A. Matis, G. P. Gallios, and A. I. Zouboulis, Adsorption of
fatty acids on fine mineral particles as a purification method of
industrial wastewaters, Toxicol. Environ. Chem., 20: 425435
(1989).
24. G. P. Gallios and K. A. Matis, Flotation of salt-type minerals,
Innovations in Flotation Technology (P. Mavros and K. A. Matis,
Eds.), Kluwer Academic, Dordrecht (1992), pp. 357382.
Page 517
20
Metal Ion Flotation in Hydrometallurgy:
The Case of Germanium Recovery
A. I. Zouboulis and K. A. Matis
I
Introduction
Pyrometallurgy is admittedly the ancient art for metal production.
People learned thousands of years ago how to build furnaces and to
use fire in order to melt ores and to produce metals. Use of water or
aqueous solutions for replacing the dry, high-temperature methods
of ore processing came much later. Modern hydrometallurgy, in
fact, can be traced back only to the end of the 19th century. Its
discovery can be attributed to the invention of two novel (for that
time) processes, i.e., the cyanidation process for gold recovery and
the Bayer process for alumina production. Another major
breakthrough came during the 1940s, with the introduction of new
leaching agents for uranium ores and the combined use of ion
exchange and solvent extraction in order to recover this metal
quantitatively (1).
In general, hydrometallurgy involves two distinct steps:
1. The selective dissolution of metal values from a mineral (raw
material), a process known as leaching
2. The selective separation and recovery of these metal values,
commonly existing in small concentrations, from the bulk solution
In most cases, the leach solution often must be purified previously
and properly concentrated before the second step can be realized
(2).
Germanium, on the other hand, rarely forms minerals of its own,
which
Page 518
contain this element in concentrations over 10,000 g tn-1 (3,4).
Germanium is a scattered element occurring in traces in other ores,
mainly mixed sulfides especially adjoining with sphalerite. The
beneficiation of these sulfide minerals takes place applying mainly
flotation, as, for example, in Tsumeb plant of South Africa (5),
where germanium is finally separated selectively as a concentrate.
More recently, flotation has been applied also for the recovery of
germanium from Fargo Oil's Lang Bay property (6). Germanium
can also be found, but in trace amounts, in coal and oil (7,8).
Today, major reserves of germanium occur in sphalerite, and
germanium is usually obtained after zinc electrolysis. But in
industrialized countries the efforts to find new germanium sources
have been intensified, like the recovery of germanium from coal
flue dust (9). Studies are oriented also toward the more efficient, in
economical terms, separation of germanium from raw materials, so
that the various requirements are satisfied, i.e., for the electronics
industry. The use of germanium in fiber optics and infrared lenses
increased dramatically in the late 1970s. These became the
dominant applications for germanium, increasing the significance
of this element (10).
Several million tons of pulverized fuel ash (fly ash) are produced
every year from coal-burning power plants and collected mainly by
using electrostatic precipitators. Actually, fly ash is an industrial
solid waste, being in some cases almost 20% of the initial coal
weight and may represent serious environmental problems
(collection, disposal, etc.) (11). In order to eliminate these
problems, several applications of fly ash have been proposed (12).
Among them, well studied and already widely applied, is the use of
fly ash as a pozzolan replacement in the Portland cement
production industry. This cement, after fly ash addition, is also
considered to present increased resistance to corrosion, and
appreciable work has been conducted and published in this area
(13). Research also has been carried out on the possibilities of fly
ash use as a primary resource material, aiming to its recycle
through metal recovery (14,15)or its use in environmental
engineering as a cheap solid adsorbent material (1618).
Coal fly ash can be considered as an already mined oresize reduced
and readily available for engineering applicationsin substantial
quantities and rather consistent composition, depending on the
origin and type of burned coal. Only the total (if possible) recovery
of all the mineral constituents from fly ashi.e., aluminum, iron,
calcium, silicon, sulfur, sodium, and titanium, but also (from the
trace-contained values) uranium, germanium, molybdenum, etc.is
envisioned as economically feasible and will have a positive
impact on the ash disposal problem. For instance, recovery of
aluminum alone from fly ash, a procedure extensively investigated,
is not antagonistic with the traditional production of the metal from
bauxites. The different available processes (mainly chemical),
which can be applied
Page 519
in this field were recently reviewed (19). These are generally
distinguished into four categories:
1. Hydrometallurgical leaching
2. Sinteringleaching
3. Gas/liquid reaction
4. Direct reduction
II
Germanium:
Occurrence, Analysis, and Aqueous Chemistry
For many years it has been well known that many coals contain
germanium (2022). In this way, after the burning of coal in thermal
power stations, the produced ash (fly or bottom) would be enriched
in germanium, provided that it is not lost with the waste gases
(23,24). The vertical distribution of germanium in the layers of a
coal deposit was found to vary depending on the nearby existing
mineral gangues. Furthermore, germanium was noticed to have the
tendency to be concentrated near the edges of the coal layers (3).
Certainly, the petrographic coal distribution in germanium seems to
be of little importance in evaluating the possibilities of its recovery.
This is the reason why it is not technologically feasible to select the
most appropriate type, or even the specific part of coal that
contains the higher germanium concentration that it is fed for
burning in the plant. Hence, attention should be given mainly to
methods that are capable of concentrating germanium (which exists
mainly in the form of the respective oxide) in the waste solid
products of coal burning. This is due to the very low initial
germanium concentrations in the initial raw materials, being
usually on the order of few g tn-1 only. The form of germanium,
with which it initially exists, was considered to be insignificant.
When the relation between the particle sizes of the collected fly ash
and its respective germanium composition was examined in coal-
fired power plants, it was realized that in the -20-µm particle size
fraction germanium was concentrated almost quantitatively.
Possibly, this was due to the fact that the finer particles present
higher surface areas and remain for longer retention times in the
flue gases, particularly in units equipped with electrostatic filters
(3). In this way, some preliminary classification of fly ash might be
performed.
The different ways of coal burning were also studied and mass
balances were calculated around the combustion unit. It was shown
that concentration of germanium in fly ash was inversely
proportional to the amount of fly ash produced, which also depends
on the specific type of furnace used
Page 520
(a total of 13 were examined). In samples from dust bins the
germanium content was reported to be on the order of 1560 g tn-1.
The enrichment ratio from coal to the fly ash was found to be about
7 (3).
In other papers, different techniques of burning and gasification of
coal samples in the laboratory were investigated, showing that
depending on the conditions (oxidation, reduction, etc.) germanium
may either pass in the gaseous products or could not leave the
combustion chamber, thus remaining with the solid residue (slag)
(25,26). Main parameters influencing the volatility of the chemical
compounds that germanium forms in coal are temperature and
quantity of oxygen supplied (as air) (27).
Basic possible equations describing these chemical reactions are
given in Table 1. In the system germanium/oxygen, two stable
oxides are forming, of which monoxide is a very volatile
compound. The gases that leave the combustion zone undergo a
stagewise cooling, thus preventing the escape of germanium with
the waste gases. It is worth noting here that the usual industrial
conditions for coal burning are over 1270 K (8). Table 2 shows the
representative chemical analysis of three different Greek fly ash
samples used for the present studyin order to get an idea of its main
constituents.
Germanium dioxide can be combined with the minerals usually
encountered in coals, which in the high burning temperatures are
transformed mainly to the respective oxide forms, like CaO, MgO,
Al2O3, FeO, Fe2O3 and SiO2, giving various mixed products, as
shown by a thermographic
Table 1 Reactions of Germanium in Coal-Burning
Conditions
Type Note
2 GeS2(s)® 2 Possible up to the melting point of GeS2
GeS(g) + S2(g) (being 1098 K).
2 GeS(g)® 2 Further decomposition is thermodynam-
Ge(1) + S2(g) ically of low possibility even at high
temperatures.
GeS(g) + 2 High rate reaction, at 770 K.
O2(g)®
GeO2(g) +
SO2(g)
GeO2(g)® Low probabilityeasier in presence of solid
GeO(g) + 1/2 C, H2, or CO (at around 1200 K).
O2(g)
GeO(g) + The analysis of the last two equations
CO(g)® Ge(s) showed that in high temperatures the
+ CO2(g) equilibrium turns to the left side.
2 GeO(g) +
C(s)® 2 Ge(s)
+ CO2(g)
From Ref. 3.
Page 521
Table 2 Chemical Analysis of Greek Fly Ash
Samples.a
Constituent Kardia Kardia Megalopolis
(%) Unit 1 Unit 2
Loss on 3.15 3.00 1.87
ignition
SiO2 29.88 28.49 49.48
Al2O3 12.66 12.72 18.25
Fe2O3 4.77 4.65 8.50
TiO2 0.26 0.24 0.90
CaO 36.07 35.80 14.98
MgO 3.98 5.15 3.15
SO3 5.67 5.46 1.06
Na2O + K2O 1.01 0.93 1.06
aProvided by Public Electr. Corp.
analysis followed by x-ray identification of the created chemical

compounds (28,29). The formation of germanium salts between
magnesium and calcium (from their oxide or carbonate
compounds) and germanium dioxide, happens even in the
relatively low temperatures of 10701170 K. The oxides of
aluminum, iron, and silicon do not react with germanium dioxide at
temperatures below 1270 K. The formation of solid solutions of
germanium dioxide in silicon dioxide occur only at 1570 K. In
other words, the burning temperature and content of oxides in coals
play a significant role for the distribution of germanium between
fly ash, waste gases, and slag.
The chemical analysis of germanium, except from the application
of different modifications of the atomic absorption
spectrophotometry methods (3037), also includes one of the most
sensitive colorimetric methods for the determination of small
amounts of this element, using phenylfluorone (2,3,7-trihydroxy-9-
phenol-6-fluorone) in an acid solution made from sulfuric or,
better, hydrochloric acid (3841). This reagent forms a red-coloredd
complex with tetravalent germanium, which is subsequently
stabilized by the addition of a protective colloid, such as gelatin.
This method also was followed by the authors, for the quantitative
analysis of germanium in dilute aqueous solutions, presenting high
sensitivity and precision (42).
Special mention also has to be given for the proper selection of
conditions for careful ashing of the coal samples (gradual increase
of temperature), in order to obtain reliable results for germanium
determination (43), as well as for the careful preparation of
standard germanium solutions from GeO2 (44).
The required preliminary separation of elements, which otherwise
could inhibit the previously mentioned analytical reaction, may be
carried out by
Page 522
using a preliminary distillation step of germanium tetrachloride
from 6 N HCl solution, or by extraction of this in carbon
tetrachloride from 89 N HCl. The latter method presents some
advantages over the first one (45). Fusion of the sample solid
material, using mainly sodium carbonate, was usually the
preceding stage of the chemical analysis, in order to dissolve the
contained germanium from the solid phase. The disintegration of
mineral samples by hydrofluoric or phosphoric acids has also been
proposed. For samples containing carbon, a different treatment was
proposed, according to the type (e.g., anthracite and lignite) and the
origin of coal sample to be analyzed.
A wet ashing analytical method reported earlier (46) was finally
applied for the determination of germanium in lignite fly ashes.
Lignite is the main type of coal existing (hence, used mainly for
electricity production) in Greece. This method involves flameless
A.A.S. using a graphite furnace, after appropriate preconcentration
by applying solvent extraction into dicyclo-hexyl-18-crown-6
(dissolved in a methyl-isobutyl ketone solution). Several other
solvents also were tested for the extraction of germanium. The
possible interferences of various ions during the determination
were examined by introducing them into the solution. The
proposed technique was further verified by standard additions of
the examined metal to the fly ash samples, before ashing and the
determination took place afterward. It was realized that there were
certain difficulties in the chemical analysis of germanium, mainly
produced by the presence of silicon and/or carbon constituents
(47).
Aqueous solutions of germanium(IV) at low concentrations (below
10-2 M) and at pH values from 2 up to 7 mainly contain molecules
of the weakly ionized metagermanic (H2GeO3) and pentagermanic
acids (H2Ge5O11) (48,49). Ionic forms of germanium (cationically
or anionically charged) are known to exist only under specific
conditions, e.g., at pH lower than 2 univalent Ge(OH)3+ cations
appear. The equilibrium between the aforementioned forms
depends on the pH value and on the total germanium concentration.
At pH values lower than 7, cations of germanium, like Ge4+,
GeO2+, or Ge(OH)22+, also may appear under specific conditions.
As an example, the equilibrium diagram of the several germanium
species, constructed by means of a specific computer program, is
presented in Figure 1 (50,51). The dissociation constants of
metagermanic acid, which is of specific interest for the present
chapter, are the following (49):
In strong hydrochloric solutions of germanium dioxide, there is an

equilibrium established between germanic acid (or germanium
dioxide) and the
Page 523
Figure 1
Germanium ionic equilibria fraction diagram against pH for initial concentration 1 × 10-5 M.
anions GeCl5- or GeCl62-. It is also possible for some other intermediate

complexes to be precipitated in this equilibrium, such as the following:
The solubility of germanium dioxide in water is reported to be on the order of 25 g

L-1 at a pH range of 2 to 6 and temperature of 293298 K. The complexes of
germanium with oxalate (52), polyhydric alcohols (53), quinalizarin (54), dibasic
carboxylic acids (55), and hydroxy-acids (56) also have attracted attention.
Noteworthy complexes also are formed between germanium and humic acids, and
in the geochemistry of germanium the accumulation of this element in coals was
attributed to humic acids (57).
III
Extraction of GermaniumLeaching or Other Recovery Methods
In pyrometallurgical tests made for the extraction of germanium from fly ashes,
the iron content was shown to have deleterious effects, due to its
Page 524
oxidative power. Various ways of coal gasification were reported,
based mainly on laboratory work, for the recovery both of
germanium and gallium (58). It was suggested that certain fly ashes
were especially suitable for processing, in order to recover the rare
metal values, which they are containing. Another promising
method from the economic point of view was found to be the
preliminary sample agglomeration with freshly prepared CaSx (8%
sulfur) and the subsequent heating of the product at a temperature
of 1300 K, in order to gasify the contained germanium as the
volatile germanium sulfide. The sintering of the remaining material
(ash having a diameter lower than 40 µm), previously mixed with
alkali chlorides, was followed for gallium removal and recovery.
The described sulfurization process can be presented with the
following chemical reactions:
It was proved from thermodynamic data that the first chemical

reaction is the most probable.
Separation processes, most commonly applied for the recovery of
valuable chemicals from aqueous effluents, include several
methods, like solvent extraction, adsorption, and membrane
processes (59). For many years the application of these methods to
the recovery of germanium from fly ash or other germanium-
containing minerals has been researched intensively (6065).
Special mention is given to certain patents (6670). The last patent
(70) proposed the sintering of ash with alkali chloride and
afterward the processing/leaching with sulfuric acid. From the
resulting solution, germanium was precipitated out as a complex
salt following the addition of tannin. A 0.1% germanium content of
the ash was foreseen as economically satisfactory by that time.
Another interesting suggestion was to use hydrofluoric acid in the
preliminary stage and by heating to get rid of silicon dioxide (71).
The coprecipitation of germanium with copper hydroxide and the
subsequent removal of copper by extraction with carboxylic acids
has also been proposed for the germanium concentration (72,73).
Apart from precipitation, other separation methods of germanium
from aqueous solutions are available, as for example ion exchange
(74,75), solvent extraction (7681), sorption (82,83), adsorption on
activated carbon (8486), or other modified adsorbents (8791) and
flotation, which will be analyzed later. The development of
hydroxy-oximes (LIX reagents) has already found application in
solvent extraction of germanium (92). It is true
Page 525
that in the 1970s a real evolution in hydrometallurgy was realized
(2)this also gave the idea for the present work.
A rather complete processing for germanium (and gallium)
recovery from fly ash was based on a pyrometallurgical process
(93). The initial sintering of English coal ash with sodium
carbonate and calcium oxide resulted in the removal of over 90%
of the initial contained germanium. The obtained melt was received
in a dilute ferric chloride solution and further processed by a
stream of chlorine (exothermic reaction). In the products,
concentrated sulfuric acid (about 7 N) was added, and the solution
was then distilled. The distillate due to its acidity was separated
into two layers, the lower containing germanium tetrachloride
(having a boiling point of 357 K) and about 20% arsenic trichloride
(b.p. 403 K) (94). The subsequent separation of germanium from
arsenic was done by adiabatic fractional distillation. The obtained
germanium tetrachloride by this method is then hydrolyzed toward
the respective dioxide. This chemical reaction occurs by stirring
and after the addition of water to give a final solution of around 5
N HCl. Precipitated in this way, germanium dioxide is filtered,
washed, and dried. By heating it in a furnace and under the proper
conditions, germanium is produced in elemental form from its
oxide. The final purification, if needed, for the application of
germanium in the electronics industry usually would be followed
by applying zone refining.
It is a fact that in many countries (like Czechoslovakia or Poland)
the industry's requirements for germanium have been organized by
utilizing as resource material the fly ash produced from coal-fired
power plants. In the former USSR much work was conducted on
the leaching of germanium from fly ash. The fly ash leachate by
both acidic (95) or basic solutions (96) was attempted and gave a
92% maximum germanium recovery in the solution from alkaline
ashes.
Table 3 presents some selected preliminary laboratory results on
germanium leaching from two different fly ash samples (originated
from two Greek power plants, burning lignite of different quality)
by using comparatively dilute solutions of sulfuric acid (0.5% w/v)
or sodium hydroxide (0.5% w/v). The idea was to co-extract and
hence to obtain in the leachate solution as little of the other
constituents as possible, so as to avoid during the next processing
stage (being flotation) possible depression effects. Three
temperatures (343, 373, and 473 K) were tried at various leaching
conditions shown in Table 3, with or without previous roasting of
the samples (873 K for 2 days). More details about the leaching
procedure are presented elsewhere (47). It was observed that the
finally obtained aqueous solutions had germanium concentrations
generally on the order of 10 g m-3. Therefore,
Page 526
Table 3 Selected Results of Leaching Experiments
(Laboratory Tests).
Ge content
Leaching (g m-3 fly
OriginaMedium-% temperature (K) ash)
K1 H2SO4-0.5 473 112.0
M H2SO4-0.5 473 166.0
K1 NaOH-0.5 473 248.5
M NaOH-0.5 473 128.0
K1r NaOH-0.5 373 140.0
Mr NaOH-0.5 373 243.5
K2r H2SO4-0.5 343 650.5
Mr H2SO4-0.5 343 486.0
K2r NaOH-0.5 343 199.0
Mr NaOH-0.5 343 134.0
aThe subscript r denotes sample preparation before
leaching. These samples were previously roasted
at 873 K for more than 2 days in order to get rid of
the contained unburnt carbon, which was thought
to be causing problems, because of its adsorptive
properties.
Reprinted with permission from Ref. 47.
leaching as the first stage for germanium recovery from fly ashes
seems to be a possible hydrometallurgical route.
IV
Combination of Hydrometallurgy and Flotation
During the last decade, hydrometallurgy has continued to present
important advances, even replacing some more traditional (i.e.,
pyrometallurgical) processes. For instance, it is the only method for
extracting metals that in some cases by-passes the usual mineral
dressing operations, applied for ore enrichment. Hydrometallurgy
successfully competes in certain cases with pyrometallurgy (on
economic terms), as in the treatment of low-grade ores (such as
uranium) or in the processing of complex sulfide ores.
Hydrometallurgy has also been suggested as an alternative to other
older established processes, in order to reduce some serious
pollution problems that are caused when these methods are applied
(especially atmospheric pollution caused by smelter emissions).
However, significant amounts of liquid wastes are often produced
that may pose serious treatment and disposal problems (1). Other
problems involved with hydrometallurgical processing follow:
Page 527
1. It may be energy intensive, particularly if electrowinning is a
part of the overall process.
2. It requires the recovery of usually expensive chemical reagents
(solvents).
Various processes different from conventional smelting have been
applied, aiming at producing metals from complex or impure
concentrates (97,98). It was pointed out upon examining the
economics of smelting sulfide concentrates that it was doubtful if
large-scale smelters using the best technology could be justified
even for clean concentrates. It is true that the base metal prices are
extremely depressed.
An intermediate approach, similar in its objectives with some
suggested smelter modifications, would consist in preprocessing
complex ores or impure concentrates in order to bring them to a
situation that is amenable for subsequent smelting. This proposal
still could be carried out in conventional existing units. Such an
approach is not new, and various examples can be found, as in the
preliminary preparation of impure lead concentrates at the Kosaka
smelter of the Dowa Company (99).
The process developed by Cominco is also interesting, since it
attempts to combine hydrometallurgical with physical separation
processes for a more efficient overall separation scheme and an
optimum utilization of installations (100). For instance, two
approaches were applied in order to leach the ore directly at larger
than usual particle sizes, either with ferric chloride for galena (lead
sulfide) dissolution, leaving chalcopyrite, sphalerite, and pyrite to
be separated subsequently by flotation, or with a ferric sulfate
leaching step for sphalerite (zinc sulfide) dissolution, leaving lead
as sulfate and iron as jarosite. The latter process is selective with
respect to chalcopyrite (98).
Such types of processes were already used at full industrial scale
for the recovery of lead and silver from the residues of zinc
concentrate processing, when the RLE process was applied.
Various installations incorporated such a facility, such as
Outokumpu (101) and Hoboken (102). The process used in these
installations was the flotation of lead and silver minerals, either
directly or after sulfidization. Generally, the decisions, which have
to be taken (on any case), depend finally on economics and result
mainly from the mineralogical properties of the ore to be treated.
Therefore, the trend during the last decade has been the
development of combined processes, including physical separation
methods (such as froth flotation) or hydrometallurgy in order to
achieve maximum efficiency in recovering mineral values into
concentrates, which are subsequently treated by conventional
existing smelters (103).
Page 528
The main scope of the present work was to examine systematically
an alternative flotation technique, termed (in the broad sense) ion
flotation, applied on a dilute solution, such as that of germanium,
and to separate this element selectively from other elements that
coexist in the solution as metal ions (arsenic(III), zinc, and lead),
although information about the flotation of germanium alone has
previously appeared in the literature. More information on
terminology and available methods, as well as techniques and
process fundamentals can be found in Chapter 1 of this book.
Perhaps it is worth noting here that other precious metals, like
gallium, gold, and indium, were also recovered selectively by their
dilute aqueous solutions by applying flotation (104106).
V
Ion Flotation
A
General
This flotation method was first introduced in 1959 (107). A number
of informative reviews have appeared recently (108,109). In
general, it involves the removal of surface-inactive ions (termed
colligends) from dilute aqueous solutions by adding surfactants (or
collectors) that have the opposite charge with the ion to be removed
and the subsequent passage of air bubbles through the solution. As
a result of this procedure, a solid particle containing the surfactant
as a chemical constituent, termed sometimes sublate, appears on
the surface of the solution. This permits the separation and
concentration of the ionic species from the initial solution in the
smaller volume of collapsed foam, being around 1% or less of the
initial solution. If necessary, the foam layer can be broken down,
by using various chemical, thermal, or mechanical methods.
As opposed to froth flotation, this method requires aeration at
lower gas flow rates (usually less than 16 cm3 s-1). Different
bubble generation techniques may be followed (109). The
surfactants are generally recoverable from the ion-surfactant
product, thus enhancing the process economics (110).
As the initial concentration of the ion to be removed is usually low,
the sublate may comprise several groups of ions, held to the surface
of the bubbles due to the surface activity of the surfactant. Hence,
flotation occurs from a true solution. Raising the initial ion
concentrations, however, a precipitation of these may appear at
certain pH ranges. In this case the process is termed precipitate
flotation (111). Adsorbing colloid flotation can also be classified in
this group (112). This process generally involves the removal of an
ion by means of a colloid (such as ferric or aluminum hydroxide)
after adsorption, coprecipitation, occlusion, or other related
processes and the subsequent flotation of the carrier colloids.
Page 529
The removal or recovery of a constituent from its solution by
means of flotation (Re%) is usually expressed with the following
equation:
where C0 denotes the initial and Ct the concentration at a

predefined time.
A typical example of ion flotation is that of chromium(VI) by
dodecylamine as cationic surfactant (113). Examining the aquatic
chemistry and speciation of chromium, it was realized that in acidic
solutions Cr2O72- is the dominant form at the used concentration
range. The amount of surfactant needed in this method is
theoretically at least the stoichiometric one, as it is related directly
to the amount of the metal to be removed (surface-inactive ion) and
to its state. This method is generally not influenced by the amount
of solution to be processed (114).
Nevertheless, it is possible that the metal ion to be removed or
recovered (for instance, in the form of a slightly dissociated acid)
may require as a first step the reaction with a proper activator (like
a ligand) in order to form a stronger ionized complex compound.
Then, this complex could be floated more easily, by adding a
surfactant of opposite charge. However, it can be argued that this is
another form of ion flotation, although not usually applied, and the
recovery of germanium is described as a rather typical example
(108,109).
The concentration and the subsequent selective separation of
germanium from its dilute aqueous solutions where it is known that
it coexists with other metals, like zinc, arsenic, and lead, is of great
technological importance, but the commonly available methods
suffer from a number of disadvantages. For instance, although
precipitation is considered to be a key step of hydrometallurgy, the
precipitation of germanium using tannin substances, a method that
was proposed earlier, presents the unfavorable effect of tannin on
zinc electrolysis (115).
B
Flotation of Germanium Ion
Germanium is known to form complexes with gallic and tannic
acids, which were shown to float by rosin amine acetate (116). A
more stable complex is formed with tannic acid, so that it gives
greater recoveries. It was also found that the flotation of these
complexes offered certain advantages over precipitation, such as
the simplicity of the proposed method and, particularly, higher
recovery rates. It is noted that a usual technique for germanium
separation from aqueous solutions constitutes its precipitation by
tannins (69). Other complexation agents of germanium like citric
and tartaric acids also were applied, followed by flotation using
rosin amine acetate (117). In both previous cases high germanium
recoveries were noticed.
Page 530
More specific information on several flotation reagents and their
applications on different flotation systems can be found in the
literature (118).
Aliphatic amines were also studied as surfactants for flotation of
germanium complexes with tartaric acid (119). Primary, secondary,
and tertiary amines were tested. The number of carbons in the
aliphatic chain of the examined amines, in order to obtain
successful flotation recoveries, was varied each time, depending
mainly on the nature of amine, from 10 carbon atoms in the
aliphatic chain, when primary amines were used, up to 15 carbon
atoms when quaternary ammonium salts were applied.
Trihydroxy-fluorones carrying different substitutes were reported
as suitable chemical agents for separation of germanium, forming
selectively insoluble complexes with this and enhancing its
flotation properties (120). In this way, these compounds were
considered as suitable ion exchangers for germanium
concentration. The surfactant used for the flotation of these
complexes was rosin amine acetate.
The complexes of germanium with catechol and pyrogallol have
also been examined, with their subsequent effective flotation by
using dodecylamine or cetyl trimethyl-ammonium bromide (121).
The examined parameters in these systems were the pH of solution
and the initial concentrations of germanium, surfactant, and
complexing ligand (activator). The aforementioned paper actually
may be considered as the background for the present work to be
further examined. The selected reagents (pyrogallol and
dodecylamine) are common, cheap, and can be found easily.
Another reason was that the studied system gave very promising
recoveries during preliminary experiments (122). The selective ion
separation of germanium from other coexisting metals was
subsequently investigated.
The preconcentration of several elements, including germanium,
from natural or sea waters by adsorbing colloid flotation was also
studied (123). Ferric hydroxide was the ''adsorbent" phase and
sodium dodecylsulfate used as surfactant. The aim of the last paper
was mainly the analytical determination of these elements. Due to
its advantages, ion flotation often has been applied for analytical
purposes as a selective separation step. Flotation has been applied
also as a preconcentration technique for the analytical
determination of germanium in industrial materials (124) or in coal
fly ash (125).
Completing this short literature review, special mention has to be
given to Sebba (107), who proposed the theoretical possibility of
germanium removal by ion flotation as polymerized ions or
GeCl62-, using cationic surfactants. It was noted also that the
separation of germanium from arsenic, where these elements are
coexisting in aqueous solutions, is difficult. Therefore, their
transformation was suggested to the respective chloride
Page 531
complexes, by applying concentrated hydrochloric acid, hoping for
a selective flotation due to their different carrying ionic charge.
On the other hand, in order to examine in more detail the aqueous
chemistry of the system germanium/pyrogallol/dodecylamine, it is
noted that pyrogallol (a poly-hydroxy-compound) forms a strong
complex acid with germanium in dilute solutions, having
composition H2GeL3, where L is the pyrogallol ligand. It was
assumed that this complex acid, which dissociates forming GeL32-
anions, may be floated with amine-type surfactants (dodecylamine,
in this case). So if S+ denotes the surfactant cation, ion flotation of
germanium was shown to comply with the following reaction
mechanism:
which was certified also by elemental analysis (of N, C, and H) of

the product, giving 3.28% N, 61.34% C, and 8.68% H. This means
that the stoichiometric ratio in this case is the following:
resulting in a product having molecular type C42H68N2O9Ge.

Laboratory bench scale tests (100 cm3 solutions) were performed
using dispersed-gas (nitrogen) flotation for bubble formation,
instead of air because pyrogallol is known to be sensitive to
oxidation by atmospheric oxygen. A 10% excess of reagents over
the theoretical ratio was used (126). The initial germanium
concentration was in most cases 1 × 10-4 M (around 7.3 mg L-1),
unless otherwise stated. Other reagents (complexing agents,
collectors, etc.) were also applied, but unsuccessfully (42). A
typical scheme of the apparatus constructed has been presented
already (127).
C
Parameters Affecting the Process
The main parameters influencing the removal of metal ions by
flotation, being extensively reviewed recently (108,109,114),
follow:
1. solution pH value
2. bubble size, flotation time
3. reagent concentrations (as described also above)
4. ionic strength
The solution pH is perhaps the most significant parameter in ion
flotation (Fig. 2). It was shown that the best flotation recoveries for
germanium (over
Page 532
Figure 2
Influence of solution pH on germanium recovery with flotation. Initial concentrations: Ge 1 × 10-4 M,
pyrogallol 3.3 × 10-4 M, dodecylamine 2.2 × 10-4 M, flotation time 300 s, gas (nitrogen) flow rate
200 cm3 min-1, ethanol 0.5% v/v.
90%) occur at a pH value of approximately 7. A decrease in germanium recovery was
noticed when the flotation was taking place in higher temperatures, possibly due to the fact
that the foam stability is inversely influenced (127). The investigated temperature range
was from ambient (around 293 K) up to 333 K. The gas flow rate has some effect on the
process, and, as such, 3.33 cm3/s (at a pressure difference of 26.6 kN/m3) was selected.
The retention time in the flotation microcell was 300 s (Fig. 3). Several diffusers, being
Schott glass frits of varying porosities, have been tried before selecting D4 (i.e., a porosity
of 1016 µm). With the latter, it was calculated theoretically (without taking into
consideration the presence of surfactant) that the produced bubbles would be around 800
µm in diameter, being well over the size encountered in dissolved-air flotation (42). It was
observed that with increasing diffuser porosity higher recoveries were obtained in shorter
times, i.e., the different frit porosity greatly influenced the recovery kinetics. The flotation
kinetics study showed that the first-order equation was followed (127). The surface tension
of the solution was measured to be around 60 dyn cm-1.
Page 533
Generally, for treatment by flotation of effluents containing fine particulate matter, fine
bubbles are required for a satisfactory separation. This is also one of the reasons why
frothers are applied in flotation, having as a main result the lowering of surface tension,
as well as the decreasing of bubble size (128). The surfactant used in the present
(dodecylamine) was conveniently fed as an ethanolic solution (0.5% v/v). Aqueous
solutions were also examined, resulting in lower germanium recoveries (Fig. 4).
Ionic strength is certainly the next important parameter (Fig. 5). Experiments were
carried out by the addition of several ammonium salts, having three different, but
common, anions, because the presence of foreign salts is sometimes advocated as the
main disadvantage of the ion flotation method. The interfering action of the examined
anions was found to follow the series PO43- > SO42- > Cl-. The pH value, which was
corrected properly (Table 4), as well as the use of surfactant in excess (Table 5) were
observed in both cases to improve substantially the respective flotation recoveries,
which were previously reduced.
Various primary aliphatic amines were attempted, in order to further optimize the
germanium flotation recovery, with varying the number of carbon atoms in the aliphatic
chain from 8 to 16. Dodecylamine was proved
Figure 3
Effect of flotation time on germanium recovery, using various diffuser porosities at pH 7. The other
parameters were kept the same as in Figure 2.
Page 534
Figure 4
Effect of surfactant (dodecylamine) concentration on germanium recovery from solutions having constant
ionic strength equal to 0.3 M (based on NH4Cl). The other parameters were kept the same as in Figure 2.
Figure 5
Ionic strength influence on germanium recovery. When NH4Cl was used, the pH value was around
5.76.3; with (NH4)2SO4, pH was 5.96.1; and with (NH4)3PO4, pH was 9.39.7. The other parameters
were held the same as in Figure 2.
Page 535
Table 4 Effect of pH Value Correction on Germanium
Recovery at Constant High Ionic Strength (0.3 M).a
pH Re% (comparison)
variation
Ionic strength 0.3 M Before pH After pH
Addition of From To correction correction
NH4Cl 5.8 7.1 68 87
(NH4)2SO4 6.0 7.0 47 54.3
(NH4)2PO4 9.5 7.2 18 37.8
aThe other parameters were held constant as in Figure 2.
Reprinted with permission from Ref. 126.
to be the best, under the applied conditions. The aging of solutions

for 24 h gave no significant different results, neither did the
stagewise addition of surfactant upon the germanium recovery.
However, the surfactant excess showed an appreciable negative
influence. This was counteracted by increasing the flotation time.
Bearing in mind the extra operating costs, the added amount of
surfactant must be carefully controlled (126).
To confirm the flotation mechanism, the necessary amount of the
activator (being pyrogallol) was subsequently examined (127).
Holding constant the used concentration of dodecylamine (twice
that of germanium present
Table 5 Effect of Surfactant
(Dodecylamine) Concentration on
Germanium Recovery from Solutions
Having Constant Ionic Strength 0.3 M,
Made with the Appropriate Salt
Additions.a
Salt Dodecylamine (M × Re%
addition 104) Ge
NH4Cl 2.2 70
NH4Cl 4.4 82.1
NH4Cl 22 94.5
(NH4)2SO4 2.2 47
(NH4)2SO4 4.4 70.7
(NH4)2SO4 22 74.9
(NH4)3SO4 2.2 20
(NH4)3PO4 4.4 53.5
(NH4)3PO4 22 57
aThe other parameters were held
constant as in Figure 2.
Reprinted with permission from Ref.
126.
Page 536
in solution), the effect of increasing the concentration of pyrogallol
was studied. High recoveries were obtained when this
concentration was three times greater than that of germanium, a
fact that was expected from the stoichiometry of the chemical
reaction. A further increase in pyrogallol concentration did not
have any appreciable influence.
Another parameter investigated was the germanium feed
concentration (126). Higher recoveries were obtained at initial
germanium concentrations of 1 × 10-4 to 2 × 10-4 M in deionized
water; hence, dilute solutions are best applied.
D
Selective Separation
Selective separation of metal ions from their aqueous solutions
presents a valid reason for the future application of ion flotation to
hydrometallurgy. Special mention is given to the selective
continuous precipitate flotation of copper and zinc ions as sulfides,
using dodecylamine as surfactant (129), and of copper, zinc, and
arsenic(V) ions (130). In this mixture, copper ion was first
floated/recovered by potassium ethyl xanthate followed by zinc ion
by dialkyl-dithiocarbamate, leaving finally only arsenic(V) in the
solution. As bubble generation techniques, dispersed-air flotation
was applied in the former, and dissolved-air flotation was applied
in the latter case.
Perhaps, one of the few reports dealing with the full industrial scale
applications of ion flotation, according to our knowledge, is due to
Gol'-man, who presented the recovery of two metals (molybdenum
and tungsten) as examples (131). The problems occurring with the
transfer of laboratory or pilot scale knowledge to full industrial
scale (e.g., the handling of the foam product or the fact that there
are only limited surfactants manufactured on commercial scale)
were solved successfully. As a rule, these surfactants proved far
from optimal for the purpose of effective ion flotation. The kinetic
potentials were also discussed and a model was suggested, in order
to specify the necessary number of cells in the flotation machine.
The selective separation of germanium from arsenic(III) in dilute
aqueous solutions, where these two elements coexisted, was the
first examined system after the study of germanium flotation. For
many years this separation has been attempted applying solvent
extraction (132), ion exchange resins (133), or selective adsorption
on carbons impregnated with ferric hydroxide and tartaric acid
(134). In the following, selected results from the application of two
flotation methods (ion and adsorbing colloid flotation) are
presented.
Tables 6 and 7 give some experimental results concerning the
successful separation of germanium from arsenic(III) by ion
flotation, using the aforementioned reagent ratio (135). A selective
recovery of germanium from
Page 537
Table 6 Influence of Different Initial
As(III) Concentrations (M × 104) on
Germanium Recovery from Dilute
Aqueous Mixtures GeAs(III).a
[As] : [Ge] Re% Ge
0.61 : 1 100.0
1.22 : 1 99.0
2.43 : 1 100.0
0.87 : 0 0.0
0.00 : 1 98.5
aReagent addition (pyrogallol and
dodecylamine) in stoichiometric
concentrations for the summary of
Ge and As(III) concentrations, while
the other conditions being constant
as in Figure 2.
135.
dilute solutions in the foam layer was noticed. Experiments dealing

with the adsorbing colloid flotation of arsenic(III)-containing
solutions had been previously presented (122), resulting in high
arsenic(III) removals.
Adsorbing colloid flotation of the solutions containing germanium
and arsenic(III) was also attempted (Fig. 6). This method was
found to be able to concentrate both elements at the surface of the
solution, but without any selectivity obtained. A ratio of [Fe(III)]:
[sodium oleate] equal to 60:15 mg L-1 was applied. Increasing the
reagents ratio, some preferential removal of arsenic was noticed at
low Fe(III) addition (Fig. 7).
Table 7 Influence of Different Initial
As(III) Concentrations (M × 104) on
Germanium Recovery from Dilute
Aqueous Mixtures Ge/As(III).a
[As] : [Ge] Re% Ge
1.74 : 0 0.0
0.63 : 1 96.4
1.26 : 1 95.8
1.89 : 1 95.8
2.51 : 1 96.2
3.14 : 1 96.6
6.29 : 1 96.4
9.43 : 1 97.4
aFlotation reagent addition (pyrogallol
and dodecylamine) sufficient for the
germanium concentration only.
135.
Page 538
Figure 6
Adsorbing colloid flotation of the mixture germanium/arsenic(III). Effect of initial germanium
concentration on its separation from arsenic(III). The following concentrations were used:
As(III) 10 mg L-1, Fe(III) 60 mg L-1, surfactant (sodium oleate) 15 mg L-1.
Figure 7
Adsorbing colloid flotation of the mixture germanium/arsenic(III). Effect of different initial reagent
concentration, but at a constant ratio [Fe] : [Sodium Oleate] = 4 : 1, on recoveries of germanium and
arsenic(III). Initial concentrations of the metal ions were 10 mg L-1 each; the pH value in all the
experiments was around 7.
Page 539
Another coexisting metal with germanium in certain industrial
liquors is certainly zinc. It is known that zinc ions in aqueous
solutions of concentration 1500 mg L-1 can be separated easily by
precipitate flotation as the respective hydroxide Zn(OH)2, when the
solution pH is 99.5. The results presented as recovery curves
against pH of solution were compared with theoretical data of
solubility of zinc, showing a good correlation (136). A scale-up of
this technique was also tried (Fig. 8). In the larger flotation cell,
having five times greater volume than in previous experiments,
comparatively lower surfactant (dodecylamine) was required.
Scale-up experiments on germanium ion flotation were also
presented elsewhere (127).
When the ion flotation technique (with pyrogallol/dodecylamine)
was tried on zinc/germanium and later on lead/germanium
mixtures, it was shown to float the coexisting metals together,
without achieving any selectivity. Because of this reason, the
precipitate flotation was then applied (in bench scale), resulting in
the selective separation of zinc from germanium (Fig. 9) (42).
The separation at pH 10 of the system lead/germanium was already
published (136). Figures 10 and 11 present some of these results,
where the effect of the initial lead content was examined, using
dodecylamine as surfactant. These laboratory experiments were
conducted with a liquid volume of about 2 L, hence a scale-up, in
comparison with the previous experiments. It is noted that the ratio
ø, often used by flotation chemists, represents the ratio of
surfactant to colligend to be floated (here, lead).
Finally, an overall separation flowsheet was presented between the
aforementioned binary separations (i.e., germanium, arsenic(III),
lead, and zinc) giving in two stages the selective recovery of
germanium and summarizing all the foregoing results (Fig. 12)
(136).
VI
Conclusions
As the reserves of metal ores are being exhausted, recovery of
metals from ores of progressively lower grade is becoming
necessary, and the application of flotation for further concentration
and separation is considered to be promising, particularly in
recovering relatively expensive materials from spent or leach dilute
solutions. Ion flotation was found satisfactory for germanium
recovery from aqueous solutions. The aspect of selective separation
by ion flotation constitutes perhaps the most important potential of
this process, contributing in this way in the recovery, hence the
recycling, of metal ions from their dilute aqueous solutions. This
may open new horizons in hydrometallurgy, following the
combination of it with flotation.
Page 540
Figure 8
Precipitate flotation removal of zinc. Effect of solution pH using the following concentrations: zinc 1 × 10-4
M and dodecylamine 2 × 10-4 M. The dotted line is a theoretical curve, showing the precipitation of zinc as
Zn(OH)2, i.e., the free ion concentration remaining in the solution after the precipitation.
Figure 9
Effect of initial germanium concentration on zinc/germanium separation. Concentrations used: zinc
1.53 × 10-4 M, dodecylamine 1 × 10-4 M, pH value 9.5, flotation time 300 s, ethanol 0.5% v/v.
Page 541
Figure 10
Effect of increasing initial lead concentration on lead/germanium separation with flotation. Germanium
concentration 1 × 10-4 M, pH value 10.010.3, concentration ratio of surfactant (dodecylamine) to
germanium ø equal to 2, gas flow rate 800 cm3 min-1, flotation time 1.8 ks, ethanol 0.2%.
Figure 11
Separation of lead from germanium with flotation. The same conditions were applied as in Figure 10,
except that the ratio ø was in this case equal to 0.1.
Page 542
Figure 12
Suggested flowsheet for the separation of germanium from a mixture
with coexisted arsenic(III), lead, and zinc ions.
References
1. F. Habashi, Hydrometallurgy, Chem. Eng. News, 60: 46 (Feb. 8,
1982).
2. R. G. Bautista, Ed., Hydrometallurgical Process Fundamentals,
Plenum Press, London (1984).
3. R. C. Corey, J. W. Myers, C. H. Schwartz, F. H. Gibson, and P. J.
Colbasam, Occurrence and determination of germanium in coal ash
from power plants, Report Invest. No. 575, U.S. Dept. Interior,
Bureau of Mines (1958).
4. D. F. C. Morris, A. Gharib, and H. Saber, Distribution of
lanthanium, europium, germanium, tin, phosphorus and sulfur in
nickeliferous lateritic profiles, Miner. Mag., 43: 816 (1980).
5. V. A. Glembotskii, V. I. Klassen, and I. N. Plaskin, Flotation,
Primary Sources, New York (1972), p. 521. (Engl. transl.)
6. P. B. Queneau, P. V. Avotins, and L. F. Farris, Germanium
recovery from Fargo Oil's Lang Bay property, CIM Bull., 79: 92
(1986).
7. M. T. Kazakova and V. N. Lenskaya, Determination of
germanium in Kashpir shale ash and slag, J. Appl. Chem. USSR
(transl. of Zh. Prikl. Khim.), 39: 2043 (1966).
8. K. A. Matis and A. I. Zouboulis, Germanium existence and
recovery from
Page 543
fly ash, Tech. Chron. (Chem. Eng.), 4: 31 (1984). (In Greek with
extended English summary.)
9. D. M. Jacobson, Special features of germanium and gallium
supply and demand, Trans IMM, 97: C45 (1988).
10. J. H. Adams, Germanium, Eng. Min. J., 181: 155 (1980).
11. W. Holland, K. A. Wilde, J. L. Parr, P. S. Lowell, and R. F.
Pohler, Environmental Effects of Trace Elements from Ponded Ash,
and Scrubber Sludge, Report NTIS PB-252090, Radian Corp.
(1975).
12. S. Torrey, Ed., Coal Ash Utilization: Fly Ash, Bottom Ash, and
Slag, Noyes Data Corp., Park Ridge, New Jersey (1978).
13. K. A. Matis and E. Deliyianni-Yannakoudaki, Portland cement
with lignites fly ash, Chim. Chron. (Gen. Ed.), 48: 39 (1983). (In
Greek with English abstract.)
14. Oak Ridge National Laboratory, Evaluation of potential
processes for recovery of metals from coal ash, Report CS-1992,
Proj. 1404-2, Elect. Power Res. Inst., Tennessee (1981).
15. R. M. Cannon, T. M. Gilliam, and J. S. Watson, Evaluation of
Potential Processes for Recovery of Metals from Coal Ash, Report
EPRI CS-1992, USA (1981).
16. Y. T. Hung, Utilization of Fly Ash in Polishing Oxidation Pond
Effluents, Report NTIS, EPA PB-267307, Virginia (1976).
17. A. I. Zouboulis and R. Tzimou-Tsitouridou, Fly ash utilization
in environmental engineering: the case of Greece, Reclamation,
Treatment and Utilization of Coal Mining Wastes (A. K. M.
Rainbow, Ed.), Balkema, Rotterdam (1990), p. 493.
18. A. I. Zouboulis and P. Mavros, Use of fly ash for the removal
of nickel ions from waste waters, Fresenius Environ. Bull., 1: 485
(1992).
19. K. A. Matis, Possibilities for metals recovery from fly ash,
Chim. Chron. (Gen. Ed.), 52: 107 (1987). (In Greek.)
20. K. Steward, Germanium in coal ash, Gas World, 137: 794
(1953).
21. H. R. Harner, Germanium, KirkOthmer Encylcopedia of
Chemical Technology, 2nd ed., Wiley New York (1972), Vol. 10,
pp. 519527.
22. M. K. Chaudhuri, H. S. Das Gupta, and D. T. Khathing, North-
east Indian coal: An assessment of its potential as a germanium
source, Chem. Ind., 475 (1979).
23. R. D. Smith, The trace element chemistry of coal during
combustion, and the emissions from coal-fired plants, Prog. Eng.
Combustion Ser., 6: 53 (1980).
24. V. Valkovic, Ed., Trace Elements in Coal, CRC Press, Boca
Raton (1983).
25. C. L. Waring and W. P. Tucker, Effect of ashing temperature on
the volatility of germanium in low-rank coal samples, Anal. Chem.,
26: 1198 (1954).
26. J. Jastzebski and S. Rosinki, Coal as a raw material for the
production of germanium, Koks Smola Gaz., 13: 185 (1968).
(English translation.)
27. A. Lange and H. F. Hadamovsky, Beitrag zur Kenntnis der
Bindungsformen des Germaniums in metallurgischen Flugstauben,
Z. Bergakad., 13: 47 (1961).
28. M. Ya. Shpirt, A. Z. Yurovskii, and T. I. Sendul'skaya,
Germanium com-
Page 544
pounds which form in the process of coal combustion, Sov. J.
Nonferrous Metals, 7: 67 (1966). (English translation.)
29. C. Lambrou, ''Fly ash from lignite combustion: Thermal
behavior study using TG-DTA and X-ray diffraction," Ph.D.
Thesis, Dept. of Chem. Eng., Aristotelian University, Thessaloniki,
Greece, 1983 (in Greek with German abstract).
30. D. C. Manning, The determination of boron, beryllium,
germanium and niobium using the nitrous oxide-acetylene flame,
At. Abstr. Newslett., 6: 35 (1967).
31. R. E. Popham and W. G. Schrenk, Atomic absorption
characteristics of germanium, Spectrochim. Acta, 23B: 543 (1968).
32. D. J. Johnson, T. S. West, and R. M. Dagnall, The
determination of germanium by atomic absorption spectrometry
with a graphite tube atomizer, Anal. Chim. Acta, 67: 79 (1973).
33. K. C. Thompson and D. R. Thomerson, Atomic-absorption
studies on the determination of antimony, arsenic, bismuth,
germanium, lead, selenium, tellurium and tin by utilising the
generation of covalent hydrides, Analyst, 99: 595 (1974).
34. S. Shimomura, H. Sakurai, H. Morita, and Y. Mino,
Determination of germanium by atomic absorption spectrometry
after solvent extractionEnhancement of sensitivity by a nebulizer
effect, Anal. Chim. Acta, 96: 69 (1978).
35. M. O. Andreae and P. N. Froelich, Determination of germanium
in natural waters by graphite furnace atomic adsorption
spectrometry with hydride generation, Anal. Chem., 53: 287
(1981).
36. J. R. Castillo, J. Lanaja, M. A. Belarra, and J. Aznarez, Flame
atomic adsorption determination of germanium in lignite ash with
extraction of GeCl4 into n-hexane, At. Spectrosc., 2: 159 (1981).
37. J. R. Castillo, J. Lanaja, and J. Aznarez. Determination of
germanium in coal ashes by hydride generation and flame atomic
adsorption spectrophotometry, Analyst, 107: 89 (1989).
38. H. J. Cluley, The determination of germanium, Analyst, 76: 517
(1951).
39. C. L. Luke and M. E. Campbell, Photometric determination of
germanium with phenylfluorone, Anal. Chem., 28: 1273 (1956).
40. E. B. Sandell, Colorimetric Determination of Traces of Metals,
Interscience, New York (1965), pp. 482493.
41. S. K. Tobia, M. F. El-Shahat, and E. A. Saad,
Spectrophotometric determination of germanium by
phenylfluorone, Microchem. J., 23: 525 (1978).
42. A. I. Zouboulis, "Germanium recovery from dilute aqueous
solutions and its separation from arsenic, lead, and zinc by flotation
techniques," Ph.D. Thesis, Sci. Annals No. 4, Aristotle University,
Thessaloniki, Greece (1986) (in Greek with English abstract).
43. M. A. Menkovskii and A. N. Aleksandrova, The selection of
conditions for ashing coal for germanium determination, Zavod.
Lab., 29: 853 (1963). (English translation).
Page 545
44. S. Shimomura, H. Sakurai, H. Morita, and Y. Mino, Preparaton
of standard germanium solution, Anal. Chim. Acta, 91: 421 (1977).
45. V. A. Nazarenko, N. V. Lebedova, and R. V. Ravitskaya, A
method for determination of germanium in ores, coal and industrial
wastes, Zavod. Lab., 24: 6 (1958). (English translation.)
46. I. N. Papadoyiannis, K. A. Matis, and A. I. Zouboulis,
Extraction and flameless A. A. S. determination of germanium in
lignite fly ash, Anal. Lett., 18: 2467 (1985).
47. A. I. Zouboulis, I. N. Papadoyiannis, and K. A. Matis,
Possibilities of germanium recovery from fly ash, Chim. Chron.
(New Ser.), 18: 85 (1989).
48. V. A. Nazarenko and A. M. Andrianov, Complex compounds of
germanium and its state in solution, Russian Chem. Rev. (Engl.
transl. of Uspek. Khim. USSR), 34: 547 (1965).
49. C. F. Baes and R. E. Mesmer, The Hydrolysis of Cations, Wiley,
New York (1976), p. 343.
50. I. Puigdomenech, INPUT, SED and PREDOM: Computer
Programs Drawing Equilibrium Diagrams, Interim Report TRITA-
OOK-3010, Royal Inst. of Tech., Sweden (1983).
51. M. Pourbaix, Ed., Atlas of Electrochemical Equilibria in
Aqueous Solutions, Pergamon, Oxford (1966), p. 464.
52. D. A. Everest, Studies in the chemistry of Ge(IV). Part III. Ion-
exchange studies of solutions containing germanium and oxalate, J.
Chem. Soc., 2: 4415 (1955).
53. D. A. Everest and J. C. Harrison, Studies in the chemistry of
Ge(IV). Part VII. Further studies of the complexes formed between
germanium and polyhydric alcohols, J. Chem. Soc., 7: 1745 (1960).
54. G. V. Flyatikova, V. A. Nararenko, and I. G. Kostenko,
Reaction of germanium with quinalizarin in aqueous solution, J.
Anal. Chem. USSR (transl. of Zh, Anal. Kh.), 30: 684 (1974).
55. G. Mattock, The complex-forming behaviour of tin, germanium
and titanium with some dibasic carboxylic acids, J. Chem. Soc., 1:
989 (1954).
56. A. Pflugmacher and I. Rohrmann, Ueber
Komplexverbindungen des Germaniuns mit organischer
Hydroxysauren, Angew. Chem., 69: 778 (1957).
57. F. Glocking, The Chemistry of Germanium, Academic, London
(1969).
58. R. F. Waters and H. Kenworthy, Extraction of germanium and
gallium from coal fly ash and phosphorus furnace flue dust, Report
of Invest. No. 6940, Bureau of Mines, U.S. Dept. of Interior (1967).
59. S. J. Parekular, M. M. Sharma, J. B. Joski, and Y. T. Shah,
Separation processes for recovery of valuable chemicals from
aqueous solution, Separ. Process. Tech., 3: 2 (1982).
60. M. Reynolds, Gallium and germanium extraction from flue
dust, Chem. Products, 13: 152 (1950).
61. A. P. Thomson and J. R. Musgrave, Germanium, produced as a
byproduct, has become of primary importance, J. Metals, 4: 1132
(1952).
62. W. Schreiter, Die Gewinnung von Germanium, Chem. Technol.,
6: 141 (1954).
Page 546
63. C. K. N. Nair and K. S. Rajan, Extraction of germanium from
Indian minerals, J. Sci. Ind. Res., 13B: 37 (1954).
64. L. Zahradnik, Z. Formanck, M. Stovik, J. Tyroler, and Z.
Vondrakova, Properties of light ashes from boiler houses and their
application in germanium production, Chem. prumyls, 11: 337
(1961) (in Czech.), CA 55: 23947.
65. I. Kh. Volodarskii, I. T. Grekhov, V. M. Ratynskii, M. Ya.
Shpirt, and A. Z. Yurovskii, Preparation of concentrates of rare and
trace elements from fuel coals, Tr. Inst. Goryuch. Iskop. (Moscow),
23: 64 (1966). (In Russian.)
66. Zaidan Hojin, Sekitan Sogo, and Kenkyujo, Improvements in,
or relating to the recovery of germanium from coal, Brit. Patent
733, 241, July 6, 1955.
67. H. Nakaya, Metallic germanium from germaniumcontaining
substances, Japan Patent 4211, June 5, 1956.
68. J. Memel, Improvements in or relating to the extraction of
germanium from flue dust, Brit. Pat. 774,706, May 15, 1957.
69. C. F. Reed, Germanium compound recovery process, U.S.
Patent 2,898,188, Aug. 4, 1959.
70. E. S. Pilkington, Improvements in and relating to the extraction
of germanium, Australian Patent 235,803 (1962).
71. N. S. Nikolaev, I. I. Tychinskaya, and A. A. Opalovskii,
Separation of germanium and silicon in liquid hydrogen fluoride, J.
Appl. Chem. USSR, 40: 681 (1967).
72. A. M. Andrianov and V. E. Poladyan, Reaction of germanium
containing hydroxide precipitates with carboxylic acid solutions, J.
Anal. Chem. USSR, 30: 1984 (1974).
73. A. M. Andrianov and V. E. Poladyan, Concentration of
germanium by coprecipitation with copper hydroxide and removal
of copper by extraction with carboxylic acids, J. Appl. Chem. of
USSR, 31: 1364 (1975).
74. S. Ziegenbalg and E. Scheffler, Die technische Anwendung von
Ionenaustauschern zur selektiven Gewinnung von Germanium, Z.
Bergakad., 13: 8 (1961).
75. V. P. Meleshko and G. E. Dubrovskaya, Some characteristics of
sorption of germanium by anion exchangers, J. Appl. Chem. USSR,
42: 2680 (1967).
76. G. H. Morrison, E. G. Dorfman, and J. F. Cosprove,
Purification of germanium tetrachloride by solvent extraction, J.
Am. Chem. Soc., 76: 4236 (1954).
77. N. N. Murach, V. V. Krapukhin, F. S. Kulikov, V. N. Chernyaev,
and L. G. Nekhamkin, Certain relationships in the extraction of
germanium chloride, J. Appl. Chem. USSR, 34: 2080 (1961).
78. A. M. Andrianov and L. M. Aviasovich, Extraction of
germanium by tri-n-octylamine, J. Appl. Chem. USSR, 41: 2191
(1968).
79. B. I. Petrov, V. P. Zhiropitsev, Yu. M. Burmantov, E. T.
Bobovskaya, and T. B. Cherepanova, Extraction of
trioxalatogermanic acid by the diantipyrylmethanes, J. Appl. Chem.
USSR, 30: 1462 (1975).
80. S. Kalyanaraman and S. M. Khopkar, Solvent extraction of
germanium using tributyl-phosphate, Indian J. Chem., 15: 1031
(1977).
81. Y. Sohrin, Liquidliquid extraction of organometallic and
inorganic germanium as the chloride complex, Appl. Chem., 63:
811 (1991).
Page 547
82. T. I. Denisova and I. A. Sheka, Sorption of germanium from
sodium chloride solutions, J. Appl. Chem. USSR, 52: 1004 (1979).
83. V. S. Pakholkov, G. V. Rashchupkin, and V. F. Markov,
Sorption of germanium ions from aqueous solutions by granulated
hydroxides of iron, titanium and zirconium, J. Appl. Chem. USSR,
55: 58 (1982).
84. D. Ya. Evdokimov and A. P. Kostynk, Adsorption of
germanium compounds by activated charcoal from solutions, J.
Appl. Chem. USSR, 38: 757 (1965).
85. D. Ya. Evdokimov and A. P. Kostynk, Flow method for
adsorption of germanium from solutions by a fluidized bed of
activated carbon, J. Appl. Chem. USSR, 39: 2083 (1966).
86. D. Ya. Evdokimov and A. P. Kostynk, Desorption of
germanium from activated wood charcoal under static conditions
and by centrifuging, J. Appl. Chem. USSR, 40: 681 (1967).
87. A. M. Egorov, Z. K. Odinets, and G. E. Evseeva, Some cases of
adsorption of germanium by modified adsorbents, J. Appl. Chem.
USSR, 40: 360 (1967).
88. D. Ya. Evdokimov and E. A. Kogan, Sorption of germanium on
ferric hydroxide and on wood charcoal treated with Fe(OH)3, J.
Appl. Chem. USSR, 41: 2510 (1968).
89. D. Ya. Evdokimov and E. A. Kogan, Sorption of germanium
from solutions by activated carbon impregnated with citric acid, J.
Appl. Chem. USSR, 43: 2032 (1970).
90. D. Ya. Evdokimov, E. A. Kogan, and Z. P. Sheikina, Dynamic
sorption of germanium in fluidized beds of activated carbons
modified with citric acid and 8-hydroxyquinoline, J. Appl. Chem.
USSR, 46: 1409 (1973).
91. P. Lievens and J. Hoste, Selective removal of germanium by
retention on silica gel, Appl. Chim. Acta, 70: 462 (1974).
92. A. De Schepper, and A. Van Peteghem, Liquidliquid extraction
of germanium from aqueous solution using hydroxy-oximes, U.S.
Patent 3,883,664, May 13, 1975.
93. A. R. Powell, F. M. Lever, and R. E. Walpole, The extraction
and refining of germanium and gallium, J. Appl. Chem., 1: 41
(1951).
94. E. R. Allison and J. H. Muller, The behavior of germanium
tetrachloride and related chlorides, especially arsenious chloride,
with concentrated hydrochloric acid, J. Am. Chem. Soc., 54: 2823
(1932).
95. M. Ya. Shpirt and V. M. Blavatnik, Mechanism of the sulfuric
acid leaching of germanium from cinders from the combustion of
powdered coals, Khim. Tverd. Topl., 2: 107 (1968). (In Russian.)
96. L. N. Aleksadrova, L. P. Ryzunof, and K. M. Stapostuna,
Optimization of the leaching of germanium from ashes based on
the layer combustion of coal by sodium hydroxide solutions, Izv.
Vyssh. Ucheb. Zaved. Tsvet. Met., 13: 146 (1970). (In Russian.)
97. G. Barbery, Recent progress in the flotation and
hydrometallurgy of sulfide ores, Flotation of Sulfide Minerals (K.
S. E. Forssberg, Ed.), Elsevier, Amsterdam (1985), p. 1.
Page 548
98. G. Barbery, Complex sulfide ores: Processing options, Mineral
Processing at a Crossroads (B. A. Wills and R. W. Barley, Eds.),
Nijhoff, Dordrecht (1986), p. 157.
99. T. Nishida, T. Konada, and S. Hashimoto, Treatment of
complex black ore, Complex Sulfides: Processing of Ores,
Concentrates and By-Products (A. D. Zunkel, R. S. Boorman, A.
E. Morris and R. J. Wesley, Eds.), TMS/AIME, New York (1985),
p. 239.
100. D. R. McKay and E. G. Parker, Hydrometallurgical upgrading
of copper concentrate, Complex Sulfides: Processing of Ores,
Concentrates and By-products (A. D. Zunkel, R. S. Boorman, A. E.
Morris, and R. J. Wesley, Eds.), TMS/AIME, New York (1985), p.
377.
101. J. A. Asteljoki, S. P. Fugleberg, and T. K. Tuominen, The
Outokumpu flash smelting process for complex sulfide ores,
Complex Sulfides: Processing of Ores, Concentrates and By-
products (A. D. Zunkel, R. S. Boorman, A. E. Morris, and R. J.
Wesley, Eds.), TMS/AIME, New York (1985), p. 595.
102. R. Maes, F. Lawers, and L. Groothaert, Processing of complex
concentrates and by-products at Metallurgie Hoboken-Overpelt,
Complex Sulfides: Processing of Ores, Concentrates and By-
Wesley, EDs.), TMS/AIME, New York (1985), p. 255.
103. M. E. Wadsworth, Recovery of metal values from low grade
copper sulfide ores, Sep. Sci. Technol., 16: 1081 (1981).
104. Y. Koide, H. Hokonohara, K. Jinnal, and K. Yamada, Studies
of collectors. VIII. The flotation of the gallium ion with oxine-type
surfactants, Bull. Chem. Soc. Jpn., 60: 2327 (1987).
105. H. Yamasita and K. Kato, Recovery of individual metals from
galliumarsenicindium semiconductor wastes by flotation, Appl.
Sci., 6: 783 (1990).
106. X. Wu and S. C. Liang, Enrichment of trace amounts of gold,
silver, palladium and platinum by precipitate flotation, Fresenius J.
Anal. Chem., 336: 120 (1990).
107. F. Sebba, Ion Flotation, Elsevier, New York (1962).
108. A. I. Zouboulis, K. A. Matis, and G. A. Stalidis, Flotation
methods and techniques in wastewater treatment, Innovations in
Advanced Flotation Technology (P. Mavros, and K. A. Matis, Eds.),
Nijhoff, Dordrecht (1992), p. 475.
109. K. A. Matis and P. Mavros, Recovery of metals by ion
flotation from dilute aqueous solutions, Sep. Purif. Meth., 20: 1
(1991).
110. B. L. Currin, F. J. Porter, D. J. Wilson, and R. H. French,
Surfactant recovery in adsorbing colloid flotation, Sep. Sci.
Technol., 13(4): 285 (1978).
111. T. A. Pinfold, Ion & precipitate flotation, Adsorptive Bubble
Separation Techniques (R. Lemlich, Ed.), Academic Press, New
York (1972), p. 75.
112. A. N. Clark and D. J. Wilson, Foam FlotationTheory and
Applications, Marcel Dekker, New York (1983).
113. A. I. Zouboulis and K. A. Matis, Electrolytic flotation of
chromium from dilute solutions, Environ. Technol. Lett., 10: 601
(1989).
Page 549
114. A. I. Zouboulis and K. A. Matis, Ion flotation in
environmental technology, Chemosphere, 16: 623 (1987).
115. T. Run-Chang, Z. Jia-Yan, and Z. Ling-Zhi, "New Technology
for Indium, Germanium and Gallium Recovery in an Electrolytic
Zinc Plant," Proceedings Mineral Processing and Extractive
Metallurgy, IMM, London (1984), pp. 615624.
116. I. I. Seiffulna, A. F. Pozharitskii, L. D. Skrylev, E. M.
Belousova, and A. S. Chistov, Separation of germanium tannate
and gallate complexes by the flotation method, J. Appl. Chem.
USSR, 46: 2077 (1973).
117. I. I. Seiffulina, E. M. Belousova, L. D. Skrylev, and A. F.
Pozharitskii, Ion flotation of citrato-, and tartrato-germanic acids
with rosin amine acetate, J. Appl. Chem. USSR, 48: 1363 (1975).
118. J. Leja, Surface Chemistry of Flotation, Plenum Press, New
York (1982).
119. V. V. Egorov and L. G. Starobinevits, Effect of the nature of
aliphatic amines on the flotation of anionic complexes of
germanium with tartaric acid, J. Appl. Chem. USSR, 53: 312
(1980).
120. I. I. Seiffulina, A. F. Pozharitskii, L. D. Skrylev, I. N.
Nazarova, and G. A. Tarasishina, Some relationships in isolation of
germainum(IV) by flotation in the form of trihydroxy-fluoronate
complexes, J. Appl. Chem. USSR, 47: 2741 (1974).
121. A. M. Gol'man, V. K. Chernov, and E. A. Shrader, Extraction
of tris(arenediolato)germanate anions from dilute solutions by the
flotation method, J. Appl. Chem. USSR, 44: 563 (1971).
122. K. A. Matis, G. A. Stalidis, and A. I. Zouboulis, Flotation
techniques for the separation of trace pollutants, Chim. Chron.
(New Ser.), 15: 136 (1986).
123. E. E. De Carlo, H. Zeitlin, and Q. Fernando, Simultaneous
separation of trace levels of germanium, antimony, arsenic and
selenium from an acid matrix by adsorbing colloid flotation, Anal.
Chem., 53: 1104 (1981).
124. G. V. Flyatikova, L. I. Korolenko, L. I. Vinarova, and T. N
Chekirda, FlotationPhotometric determination of germanium in
industrial materials, J. Appl. Chem. USSR, 32: 806 (1977).
125. J. Shida, Y. Takahashi, K. Ojima, and T. Matsuo,
Flotationspectrophotometric determination of germanium with
phenylfluorone in coal fly ash, Bunseki Kagaku, 33: E143 (1984).
126. A. I. Zouboulis, K. A. Matis, and G. A. Stalidis, Parameters
influencing flotation in the removal of metal ions, Int. J. Environ.
Stud. (Sec. B: Environ. Sci. Technol.), 35: 183 (1990).
127. K. A. Matis, G. A. Stalidis, and A. I. Zouboulis, Flotation of
germanium from dilute solutions, Sep. Sci. Technol., 23: 347
(1988).
128. A. I. Zouboulis, D. Zamboulis, and K. A. Matis, Foam
flotation of zeolitesApplication on zinc ion removal, Sep. Sci.
Technol., 26: 355 (1991).
129. G. A. Stalidis, N. K. Lazaridis, K. A. Matis, and I. N.
Papadoyannis, Continuous precipitate flotation of CuS/ZnS, Sep.
Sci. Technol., 24: 1033 (1989).
130. G. A. Stalidis, K. A. Matis, and N. K. Lazaridis, Selective
separation of copper, zinc, and arsenic(V) from solution by
flotation techniques, Sep. Sci. Technol., 24: 97 (1989).
Page 550
131. A. M. Gol'man, "Principles of Theory and Industrial
Application of Ion Flotation," Proceedings XVI Inter. Min. Proces.
Congr. (E. Forssberg, Ed.), Elsevier, Amsterdam (1988), pp.
17011713.
132. G. O. Brink, P. Kafalas, R. A. Sharp, E. L. Weiss, and J. W.
Irvine, Solvent extraction systems. I. As(III) and Ge(IV) in
hydrochloric acid and As(III) in hydrochloric acid with several
solvents, J. Am. Chem. Soc., 77: 4508 (1955).
133. F. Nelson and K. A. Kraus, Anion-exchange studies, XVIII.
Germanium and arsenic in HCl solutions. J. Am. Chem. Soc., 77:
4508 (1955).
134. D. Ya. Evdokimov and E. A. Kogan, Kinetics of separate and
mixed sorption of germanium(IV) and arsenic(III) on carbons
impregnated with ferric hydroxide and tartaric acid, J. Appl. Chem.
USSR, 42: 2601 (1969).
135. K. A. Matis, I. N. Papadoyiannis, and A. I. Zouboulis,
Separation of germanium and arsenic ions from effluents by
flotation techniques, Int. J. Miner. Process., 21: 83 (1987).
136. A. I. Zouboulis, K. A. Matis, and P. K. Spathis, Removal of
zinc from solutions by precipitate flotation, Tech. Chron. (Chem.
Eng.), 7: 5 (1987). (In Greek with extended English summary.)
137. A. I. Zouboulis, K. A. Matis, and N. K. Lazaridis, Separation
of germanium from lead in solutions by flotation, Sep. Sci.
Technol., 25: 463 (1990).
138. A. I. Zouboulis and K. A. Matis, Separation of heavy metals
from effluents by flotation, Pretreatment in Chemical Water and
Wastewater Treatment (H. H. Hahn and R. Klute, Eds.), Springer-
Verlag, Berlin (1988), p. 159.
Page 551
Index
A
Abstraction density, 223
Acid mine waste, 388
Acintol, 500
Activated sludge, 401
Adhesion, 73
Admixtures, 482
Adsorbing colloid flotation, 528, 537
Adsorption density, 96, 418
energy, 232
isochrone, 235
isotherms, 217
mechanism, 9, 25
parameter, 432
physical, 102
Adsorptive bubble separation, 32
Aerators, 310
Agitation intensity, 428
Air dissolution, 20, 494
Air feed ratio, 348, 536
Air flow rate, 372
Air slugs, 340
Air volume measurement, 495
Aluminum oxide, 56, 96
Amenability testing, 360
American Petroleum Institute separator, 395
Amines, 530 see also Collectors
Ammonium salts, influence on pyrite, 147
Analysis of traces by flotation, 32
Anemometry, 285
Anionic flotation, 479
Apatite, 210, 426
Aqueous solutions/effluents, 528
Arithmetic probability of bubbles, 403
Arsenic ion removal, 30
Arsenopyrite, 127, 484
Attachment, 46
Augite, 115
Autogenous mill, 359
Axial dispersion model, 278
Axial profile, 285
B
Backflow parameter, 282
Backmixing, 274
Page 552
Bacteria, 465
Bacterial conditioning, 458
Baffling, 306
Banks of cells, 274
Barite, 238, 461, 479
Basicity effect, 105
Bauxite, 427, 518
Beach sand minerals, 118
Bias, 291, 351
Bilayer formation, 9, 239
Bioadsorption, 464
Biochemically enhanced flotation, 32
Bioleaching, 455
Biomodification, 455
Biopretreatment, 467
Biosorption of metals, 33
Bridging, 419
Bubble column reactor, 297, 365
fractionation, 366
growth, 64
particle aggregate, 46
particle collision, 45, 70, 273
pick-up, 49
radius, 371
size, 66, 332, 347, 403, 497
surface area, 349
swarms, 311
velocity, 309
Bulk density, 317
Buoyancy velocity, 312
By-product recovery, 31, 409
C
Calcite, 221, 387, 461, 478
Carbonate fines, 389
Carrying capacity, 316, 346, 354
Cascade machines, 332
Cathodic reduction, 166
Celestine, 462
Cell, 261, 273, 333, 499
Cellulose, carboxymethyl, 11, 500
Centrifugal forces, 20
Centrifugation, 33, 417
Cetyl trimethyl-ammonium bromide, 394
Cetyl-pyridinium chloride, 58
Chalcopyrite, 165, 358, 481
fines, 389
Challenges of flotation, 30
Charge effects, 17, 69
Charge of bubbles, 68
Charge patch neutralization, 421
Chemical analysis by flotation (see Preconcentration)
Chemical flotation, 36
Chemisorption, 108, 237, 479
Chromite, 109
Chromium ion separation, 23, 387, 529
Chrysocolla, 115
Circuits, 274, 308, 332
Cleaning flotation, 267
Coagulation selective, 17 see also Flocculation
Coal, 331, 356, 462, 518
COD removal, 387
Collection zone, 291
Collectorless flotation, 38, 163
Collector adsorption, 51
amine, 33
anionic 94, 213, 242
cationic, 93
double layer, 57
effect on bubbles, 67
fatty acids, 8, 500
for oxides, 7, 92
for pyrite, 142, 144
for salt-type minerals, 8
Colligend, 528
Collision efficiency factor, 432
Collision of bubble particle, 13
Colloidal gas aphrons, 379
Column flotation, 19, 275, 291, 331
Combination with hydrometallurgy, 526
Cominco,
Polaris concentrator, 316
process, 527
Sullivan concentrator, 321
Complexation, 179
Page 553
Complexing agents, 32, 245
Computer programs, 192
Concentration diagram,
for Ca2+, 102
for Cu2+, 22
for Fe2+, 117
for Fe3+, 100
for Ge4+, 523
for Mg2+, 116
for Mn2+, 114
for Pb2+, 104
Constant capacitance model, 191
Contact angle, 7, 47, 54, 72, 110, 474
Continuous flow flotation, 29, 498
Control, automatic, 4, 358
Copper ion removal, 26, 28, 30, 442
retreatment, 352
sulfate, 146
sulfides, 331, 356
Corrosion, 157
Corrundum, 103
Counter ions, 91
Coverage, in flocculation, 433
Critical micelle concentration, 54, 91, 215
Critical surface tension, 45
Crypto-crystalline deposits, 499
Crystal structure, 486
Cuprous xanthate, 6
Cyanide ion, 7
D
DAF (see Dissolved-air flotation)
Deformation, 46
Denver cell, 500
Design aspects, 403
Detachment, 47, 74, 273
Diagram Eh-pH, 161 see also Eh measurements
Diffuser, 399
Dilute metal solution treatment, 528
Dirac delta equation, 279
Disengagement, 274, 341
Dispersed-air flotation, 63, 392, 532
Dispersion coefficient, 302
number 278, 304
Dissolved oxygen (DO), 483
Dissolved-air flotation (DAF), 15, 66, 400, 442, 493
Distribution diagrams, 196 see also Concentration diagrams
Dithiocarbamate, dialkyl, 536
Dixanthogen, 486
Dodecylamine, 93, 448, 530
Dodecylsulfate, 215
Dolomite, 8, 397, 493
Dosage of polymer, 422
Double layer, 97, 134
Downcomer, 295, 333, 344
Drift flux analysis, 65
Dumitrescu bubbles (see Air slugs)
Dynamic surface tension, 53
E
EDTA, presence, 7, 32, 246
Eh measurements, 35, 485
Electric field gradient, 396
Electrical properties of interfaces, 97, 133, 208
Electrochemical actions, 133, 474
behavior, 157
models, 476
potential, 160
techniques, 166
Electrochemistry, 157, 396
Electroflotation (see Electrolytic flotation)
Electrokinetic aspects/studies, 33, 127, 209, 473
Electrolytic flotation, 16, 37, 81, 385
continuous, 391
Electroseparation, 386
Enargite, 389
Enhancement factor, 371
Entrainment, 315
Equilibria, 189
Equipment, 261
Ethanol, as frother, 540
Page 554
Ethyl xanthate, 189
Exchange enthalpy, 211
Extraction of Ge, 524
F
Factor levels, 444
Factorial design, 441
Factorial experiment,
first-order, 448
second-order, 446
Factors influencing flotation, 2
Fatty acids (see Collectors)
Feed water, 316
Feldspar, 107
Fermi level, 479
Ferric hydroxides, 99
Film flotation, 49
Fine classification, 5
creation, 17, 25
flotation, 12, 275
problem, 12, 24, 493
Fisher criterion, 445
Flash flotation, 35, 263
Floatability, 5, 53
Floc characterization, 418
properties, 416
Flocculants, 416
Flocculation, 18, 23, 71, 415
Flocs in flotation, 495
Flotation column (see Column flotation)
Flotometric equation, 4
Fluoride activation, 106
Fluorite, 52, 209, 240, 461, 479
Fluoroapatite, 209
Fluosilisic acid, 500
Fly ash, 518
Foam flotation, 25
layer, 528
Free jet, 335
Froth depth, 321, 341
flotation, 398
formation, 2
stability, 3
zone, 314
FT-Raman measurements, 189
FTIR measurements, 188, 236
G
Galena, 19, 184, 460
Gallium, 524
Galvanic couples, 157
interactions (see Interactions, galvanic)
Gamma flotation, 35, 50
Gamma function, 281
Gangue, 319
Gas holdup, 294
liquid contactors, 332
rate, 321
slurry system, 294
Gases in flotation, 169, 477
Germanium ion removal, 31, 517
metallurgy, 517
Glass, 265
Goethite, 106
Gold ores, 263
Graft copolymers, 436
Grinding medium, 476
Guemassa mines, 270
H
Hallimond tube, 10, 5089
for electrolytic flotation, 389
Hamaker constant, 18
Hardness of solids, 429
Height, effect on mixing, 306
Hematite, 18, 356
Hemimicelle, 96
Hexametaphosphate, sodium, 10, 500
Higee, 66
High-grade machine, 262
Hogg's expression, 434
Page 555
Hydrazinium sulfate, 38
Hydrocarbon chain length effect, 109
Hydrocyclones, 273
Hydrodynamics, 73, 334, 406
Hydrometallurgy (and flotation), 32, 517
Hydrophobicity, 5, 47, 481
Hydroxamate, 116
Hydroxide precipitates, 29
Hydroxy complexes, 99
I
INCO's plant, 307
Induced-air flotation, 24
Induction time, 82
trumpet, 337
Industrial minerals, 331
Industrial scale columns, 283
Inoculum concentration, 466
Insertion stage, 443
Interaction collector salt-type minerals, 132
collector sulfides, 131, 179
galvanic, 132, 475
paths, 24
Interface, 179, 207
Interfacial free energy, 45
Ion exchangers, 247
Ion flotation, 27, 517
Ionic equilibria (see Concentration diagrams)
Ionic strength, 533
Iron hydroxide film, 483
sulfide minerals, 127 see also Pyrite
Isoelectric point, 98, 208
J
Jameson cell, 294, 331
applications, 356
Jet flotation, 17
K
Kinetics, 3, 404, 532
Kosaka smelter, 527
Krafft point, 215
L
LaMer's model, 432
Langmuir equation, 369
parameter, 372
Lattice ion concentration, 209
Laurate, 229
Laurylamine (see Dodecylamine)
Leaching of Ge, 523
solutions, 447, 526
Lead hydroxide precipitate, 21, 31, 539
sulfide 22, 185, 356
Lignite, 522
LMWF polymer, 431
M
Magnesite, 10, 397, 493
Magnetite, 356
Mass balance, semibatch, 370
steady-state, 375, 408
Mass transfer, 369
Mechanically agitated cells, 274
Membranes, in electrolytic flotation, 36
Metagermanic acid, 522
Metal ions, flotation, 517
sulfides, 181
Microbial cell flotation, 33
Microcalorimetry, 252
Microcline, 111
Microelectrophoresis, 127
Microorganisms, 455
Mineral types, 6
Mixed zones-in-series model, 273, 450
Mixing, 273, 450
zone, 338
Modifying agents, 487
Monolayer coverage, 9, 51
Page 556
N
Navier-Stokes equation, 74
Network model, 421
Nickel sulfides, 331
Nitrogen, 170, 473, 532
conditioning, 488
Nonfoaming flotation, 365
Nozzle, 64, 342
O
Oil emulsion, 389
Oil shale, 320
OK cell, 266
Oleate, 95, 237, 537
Oleate adsorption, 9
Optimization, 442
Organics removal, 365
Orifice, 64
Orthogonal design, 443
Orthokinetic flocculation, 430
Overgrinding, 263
Overview of flotation, 1
Oxidation of sulfides, 6, 159, 477
Oxidative reagents, 148
Oxide ores, 89
surfaces, 181
Oxygen, 478
P
Paint effluent, 390
Paper recycling, 31
Particle bubble (see Bubble-particle)
Particle interaction, 9
Particle size, 63
Partridge-Smith cell, 49
Peclet number, 278
Pentlandite, 18
pH (see solution pH),
Phenylfluorone, 521
Phosphate fines, 498
ion removal, 24
Pipe flow zone, 340
Plug flow, 286
Plunging jet, 337
Point of zero charge, 19, 91
Polar group of collector, 226
Polarograms, 35
Polyacrylamide, 418
Polyaphrons, 379
Polyelectrolyte, 405
Polygorskite, 423
Polymers, 419
Portland cement 518
Potential-determining ions, 90, 208
Potentiometric titrations, 33, 187
Pourbaix diagrams, 35
Pozzolan, 518
Preadaptation, 468
Precious metals, 32, 528
Precipitate adsorption, 98
flotation, 28, 447
Precipitation diagram, 217
reactions, 213
Preconcentration by flotation, 32, 530
Pressure profiling, 317
reducing device, 401
Pressurization, 494
Probability of collision, 73, 78
flotation, 72
Pulp density, 504
potential, 38, 487
Pulverized fuel ash (see Fly ash)
Pure samples, 11
Pyhosalmi mine, 269
Pyrite, 127, 356, 481
as adsorbent, 26
Cationic collectors for, 142
Dissolved-air flotation, 38
Flotation behavior, 136, 463, 477
Influence of copper ions, 28
Influence of Fe3+, 34
Pyritic sulfur, 462
Pyrogallol, 530
Pyrolusite, 111
Pyrophosphate, tetrasodium, 15, 500
Pyrrhotite, 486
Page 557
Q
Quartz, 55, 89, 248, 356, 479
R
Radial profile, 298
Rate of flotation, 406, 499
Recovery equation, 529
of froth zone, 320
Recycle in DAF, 494
Reducing modifiers, 149
Refinery's waste, 399, 495
Residence time, 265
distribution, 274
Rest potential, 164
Reverse osmosis, 388
Reynolds number, 80
Rhodonite, 113
RLE process, 527
Rotating apparatus, 20
Rotor stator, 274
speed, 266
Roughing-scavenging, 264, 308, 347
Roughness of particles, 84
RTD measurements, 274
S
Salt-type minerals, 207, 493
Saturation tanks, 496
Sauter mean bubble diameter, 348
Scale-up, 405, 539
Scheelite, 209, 251, 461
Sea water, in electrolytic flotation, 393
Sedimentation, comparison, 23, 390
Selective separation, of metal ions, 536
Selectivity index, 74
of foam separation, 28
SEM micrographs, 418
Separation factor, 375
Settling rate, 418
Sewage, 386, 388
Shear flocculation, 19
Silicate ores, 89
sodium, 500
Silver ores, 263
SK unit, 264
Skim-air flotation, 35, 261
Sliding time, 82
Sludge thickening, 20
Solid-liquid separation, 415
Solubility considerations, 182, 209
effects, 19, 95, 514
product, 222
Solute transport, 368
Solution pH, flotation modulation, 105, 503
influence on pyrite, 141
Solvent extraction, predispersed, 379
sublation, 366
Spargers, 298, 309
Speciation by flotation, 22
Sphalerite, 170, 195, 458
Statistical analysis, 441
Stokes equation, 79
number, 83
Student t criterion, 445
Sulfide minerals, 6, 157, 179
Sulfide biomodification, 457
flotation, 127, 158, 474, 498
Sulfonate, 117, 213
Sulfosuccinate, 255
Superficial velocity, 276, 345
Surface charge, 90
chemistry, 207
free energy, 52
phenomena, 17, 243
reactions, 180
tension, 45, 35
(see also Electrical properties)
Surfactant to colligand ratio, 539
Surfactants (see Collectors)
Suspended solid concentration, 387
T
Taconites, 18
Talc, 105
Tap water, 394
Tara mines, 269

Joao Das Couves Vol4

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Joao Das Couves Vol4

Uploaded by

Copyright:

Available Formats

title: Flotation Science and Engineering

Flotation Science and Engineering

several mineral flotation kinetic models, known from the literature

where a50 is the maximum diameter of the particle that can be

It is established that the interaction between sulfide minerals and

Cuprous xanthate and dixanthogen precipitate out from the solution.

(b) Presence of ferric ions.

As W(deformation) + W(detachment) ³ 0, it follows that the

This is the reverse condition for the spontaneous spreading of a

Using this equation, the necessary condition for flotation (or

Flotation using liquid/vapor surface tension control, also known as

Some authors specify the floatability of heterogeneous mineral

Kelebek et al. (28) found that the gc values of galena, precovered by

This approach provides an extended understanding of the surface

and the adhesion tension can be expressed by the critical surface

As the particle is not affected by dynamic adsorption processes

This demonstrates that bubbleparticle attachment is in principle

where C has a constant value of 1.378, g is the gravitational

where µ is the viscosity of the liquid, which gives good results at

where V is the bubble volume at the end of the detachment stage.

Bubble sizes produced by shear and turbulence in a bubble column

where the superscripts s and w refer to strongly or weakly floating

Because of its importance on a wide range of disciplines, the study

where a and n obtain the respective values shown in the table. A

Intermediate 2 0 < Re < 300

produced and the area swept by each bubble. Hence, it would be

and obviously k will increase markedly as the bubble radius

where ut is the terminal rise velocity of the gassolid agglomerate and

Sutherland found (also several years ago) a relatively simple

where ub is the bubble rise velocity (considering a single bubble)

The induction time is certainly valuable in that it gives a physical

where d is the diameter and r is the density. The tendency of a

A speciation diagram for 1 × 10-4 mol/L dodecylamine is presented

R represents the hydrocarbon chain

where f is the specific adsorption potential. In the case of

(b) hemimicelle formation at

mol/L CaCl2 in Figure 12 shows that CaOH+ becomes significant

3. Adsorption of aminium ions on the alumino-fluosilicate sites:

reasons, pyrite is one of the comparatively little studied sulfide

(b) mineral/grinding media.

Generally, the surface potential is not directly measurable. So, the

At the same time, Majima and Takeda (21) by electrochemical

These half reactions are connected by electron passage and exchange

(b) stationary levels.

to give the overall reaction

2. Anodic oxidation of collector through an adsorption phase:

to give an overall reaction

3. Sulfide mineral oxidation and formation of polysulfides:

4. Decomposition of metal/reagent compound by oxidation:

This decomposition is similar to that of collector to a nonactive form

6. Formation of depressant compounds with metallic species such as

Further oxidation may lead to the production of oxy-sulfur species:

Indeed, this generalization may also be used to describe the

It should be appreciated that these diagrams are much the same,

Adsorption of metal ions can occur due to either an ion exchange

According to the literature, reaction (7a) occurs mainly when M

Inorganic and organic ligands have also been reported to form

(b) PbS (galena) as a function of -log[H+].

where ME denotes a glass electrode, or ion selective electrodes

where E0 is an apparatus constant determined within each titration

where E·0S and KS were determined in separate experiments, in

where E0Pb and KPb were separately determined in experiments with

y in this equation represents the potential at the surface. According

where K is the specific capacitance (C V-1 m-2). The surface charge