Electrolytic manganese metal production from manganese carbonate precipi-
tate

Jianming Lu, David Dreisinger, Thomas Gluck

PII: S0304-386X(16)30011-1
DOI: doi: 10.1016/j.hydromet.2016.01.010
Reference: HYDROM 4263

To appear in: Hydrometallurgy

Received date: 12 September 2015

Revised date: 17 November 2015
Accepted date: 14 January 2016

Please cite this article as: Lu, Jianming, Dreisinger, David, Gluck, Thomas, Electrolytic
manganese metal production from manganese carbonate precipitate, Hydrometallurgy
(2016), doi: 10.1016/j.hydromet.2016.01.010

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Electrolytic Manganese Metal Production from Manganese Carbonate Precipitate

Jianming Lu1  David Dreisinger1,2 Thomas Gluck2

1
Department of Materials Engineering, University of British Columbia,
309-6350 Stores Road, Vancouver, BC, Canada, V6T 1Z4
2
Baja Mining Corporation, Suite 500 200 Burrard St., Vancouver, Canada, V4R 2T9

PT
Abstract

RI
The recovery of manganese metal from manganese carbonate precipitate by leaching-

SC
purification-electrowinning was studied. The manganese carbonate precipitate from Baja
Mining Corp.’s El Boleo project was readily leached into acidic ammonium sulfate

NU
solution. The manganese extraction reached 99.7%. The manganese leachate was purified
using ammonium sulfide to remove harmful impurities (Ni, Co, Cd, Cu and etc.).
MA
Manganese electrowinning was conducted in a diaphragm cell designed to reduce edge
effects and improve manganese deposition. The addition of polyacrylamide polymer had
a significant leveling effect on manganese electrodeposition. However it increased the
D

deposit internal stress and even resulted in cracking of manganese deposits at a high
TE

dosage. With increasing polyacrylamide concentration, current density, and pH, the
manganese current efficiency first increased, reached a maximum value and finally
P

decreased. The manganese current efficiency decreased with increasing deposition time
CE

as the deposit became rougher and the real current density deviated from its ideal value.
A reasonable catholyte circulation rate is important to maintain the optimum manganese
AC

electrodeposition. The presence of chloride in solution has a little effect on manganese

deposition in its concentration range from 0 to 2 g/L. The diaphragm selection as an
important part of the cell design was analyzed.

Key words: Manganese electrowinning  Manganese Carbonate  Ammonium sulfate

leaching  Edge effect reduction

1 Introduction

The standard potential for Mn2+/Mn is –1.18 V vs. SHE. Manganese is the least noble
metal that can be electrodeposited from aqueous solutions on a commercial scale.
Manganese metal is typically produced by the electrolysis of manganese and ammonium

1
 ACCEPTED MANUSCRIPT

sulfate solution. The most commonly used technology involves reductive calcining to
convert higher valent manganese oxides such as MnO2, Mn3O4, and Mn2O3 to MnO
(Zhang and Cheng, 2007). The calcine is subsequently leached in spent electrolyte from
manganese electrowinning to produce a neutral manganese sulfate solution.

PT
MnO+H2SO4  MnSO4+ H2O (1)

RI
This neutral solution is purified with ammonium sulfide to precipitate metal
impurities contained in the solution. The purified solution is then clarified, dosed with

SC
either selenite or sulfite, and used in electrowinning. In a typical cell, catholyte flows
through the diaphragm to the anode where acid and oxygen are formed; manganese is

NU
formed at the cathode. The associated electrolytic reactions are as follows:

Cathode: MnSO4 + 2e-  Mn + SO42- (2)

MA
Anode: H2O  2H+ + 1/2O2 + 2e- (3)

Overall: MnSO4 + H2O  Mn + H2SO4 + 1/2 O2 (4)

D

Besides the above major reactions, hydogen evolution and the oxidation of Mn(II) to
TE

MnO2 also take place at the cathode and anode respectively. Hydrogen evolution reduces
the current efficiency of manganese electrodeposition while the oxidation of Mn(II) to
P

MnO2 consumes the manganese ions.

CE

Cathode: 2H2O + 2e = H2 + 2OH- (5)

AC

Anode: MnSO4 + 2H2O = MnO2 + H2SO4 + 2H+ (6)

The current efficiency of manganese electrodeposition from such purified solution is

low, and the quality is poor. Even without any impurity, it is difficult to produce high
quality electrolytic manganese. As manganese is plated, manganese deposits become
rougher, and nodules grow up and outward, forming a cauliflower-like structure or even
tree-like dendrites, resulting in trapping of electrolyte, difficulty in stripping of the
manganese deposit, and even difficulty in removal of cathodes from cells. A short plating
cycle has to be used, resulting in a low overall process efficiency. The catholyte pH is
normally maintained at 7 or higher and Mn(II) is gradually oxidized to MnOOH by air,
affecting manganese deposition and even plugging diaphragm. Some additives have to be
used to counteract the harmful effects of impurities, improve manganese deposit quality

2
 ACCEPTED MANUSCRIPT

and stabilize the catholyte. Without the use of additives, it is very difficult or even
impossible to produce electrolytic manganese commercially.

Sulfur dioxide was first used to produce electrolytic manganese commercially with a
reasonable current efficiency (60-65%). The effects of SO2 are: (1) increasing the

PT
overpotential of hydrogen evolution and hence the manganese current efficiency; (2)
promoting the formation of stable -manganese, (3) counteracting the harmful effects of

RI
impurities, and (4) suppressing the oxidation of Mn(II) to MnOOH by air. The addition of

SC
SeO32- is more effective to improve the manganese electrodeposition. The manganese
current efficiency can reach as high as 96% in the presence of SeO32- (Lu et al., 2014).

NU
However, the addition of SeO32- results in the contamination of manganese metal and
contributes to environmental issues at the manganese plant and the product use site. Some
MA
organic additives such as carboxylic acids (Griffith, 1950), thiourea (Hammerquist,
1951], water-soluble polyacrylamide polymer (Jacobs, 1962; Goddard, 1979; Coleman
and Griffin, 1984) were used in combination with SO2 and SeO32- to improve the
D

manganese deposition. However there is very little detailed information available.

TE

The effects of the catholyte composition, current density, temperature and pH on the
manganese electrodeposition from sulfate media have been discussed by Jacob et al.
P

(1946), Louis and Martin (1976) and Araujo et al. (2006). The optimum catholyte
CE

manganese concentration is 12 g/L. A lower manganese concentration resulted in a lower

current efficiency. A higher manganese concentration (above 15 g/L) resulted in
AC

precipitation of manganese hydroxide and then a lower current efficiency. An optimum

ammonium sulfate concentration is around 125 g/L (Araujo et al., 2006). In the industry,
the ammonium sulfate concentration ranges from 130 to 150 g/L. Due to diffusion and
migration of ions, the ammonium sulfate concentration is 20 to 30 g/L higher in the
catholyte than that in the anolyte or in the feed. This phenomenon was also observed in
this study. The optimum temperature reported varies from 30 to 40 C for different
authors (Jacobs et al., 1946; Louis and Martin, 1976) since it is related to the current
density, catholyte cooling efficiency and energy consumption. At a temperature below 35
C, the manganese deposit is light gray, smooth and fined-grained while at a temperature
above 35C, the manganese deposit becomes more nodular and dendritic with increasing
temperature (Jacobs et al., 1946). At a temperature above 45 C, dendrites form readily

3
 ACCEPTED MANUSCRIPT

and manganese re-dissolution takes places at the base of trees. However a temperature
slightly higher than 40 C is used in some industrial plants for a higher production
efficiency and a lower energy consumption. The optimum current density ranges from
450 to 600 A/m2 (Jacobs et al., 1946; Louis and Martin, 1976), which is in turn related to

PT
the other operation parameters. A detailed review about the current state of the art in
manganese electrodeposition technology was provided by Lu et al. (2014).

RI
The El Boleo project (originally developed by Baja Mining Corp, Vancouver,

SC
Canada) has the option to recover manganese carbonate from solvent extraction raffinate
solution as a solid precipitate. This may be done by adding sodium carbonate to an

NU
impure manganese sulfate solution (Dreisinger et al., 2008) to selectively precipitate
manganese away from calcium and magnesium. It is possible to produce manganese
MA
metal from this manganese carbonate product. The chemical reaction involved in
leaching of manganese carbonate is:

MnCO3 + H2SO4  MnSO4 + H2O + CO2 (7)

D

The objective of this study was to test the feasibility of production of high purity
TE

manganese metal from chemically precipitated manganese carbonate in ammonium

P

sulfate solutions with the use of alternative electroplating additives.

CE

2 Experimental
AC

2.1 Leaching

Manganese leaching was conducted in a 2-L baffled glass reactor. The temperature was
controlled at 50  0.5 C using a heating mantle and temperature controller. The reactor
was sealed with a lid with several openings. These openings allowed the insertion of an
impeller, temperature sensor, and pH and redox potential probes. Agitation was provided
by a 45 pitched-blade titanium impeller with a 5.7 cm diameter. The impeller was driven
via an overhead motor and was suspended around 1.5 cm from the bottom of the leach
vessels. The rotational speed was controlled at 900 rpm. The feed solution with 12 g/L
Mn(II), 36 g/L H2SO4 and 130 g/L (NH4)2SO4 was prepared using reagent grade
manganese sulfate, sulfuric acid and ammonium sulfate.

4
 ACCEPTED MANUSCRIPT

2.2 Purification

The leachate purification was conducted in a 2-L water-jacked, baffled reactor using
ammonium sulfide. The temperature was controlled at 45  0.2 C using a water bath.
The reactor was sealed with a glass lid with several openings for different purposes.

PT
Nitrogen gas was used to prevent manganese (II) oxidation. Agitation was provided by a
45 pitched-blade titanium impeller with a 5.7 cm diameter. The impeller was driven via

RI
an overhead motor and was suspended around 1.5 cm from the bottom of the reactors.

SC
The rotational speed was controlled at 900 rpm. The purified leachate was pressure-
transferred to a 9-L thickener and the sulfide precipitates settled under a nitrogen

NU
atmosphere for 4 days. The mostly clear solution was filtered using a 0.2 m membrane
filter under a nitrogen atmosphere. MA
2.3 Manganese electrowinning

Manganese electrowinning was conducted in a diaphragm cell with one 316 stainless
D

steel cathode and one lead anode with 1% silver. The electrolytic cell was placed in a
TE

water bath to maintain the catholyte temperature at 42 ± 0.5 C. The catholyte was
circulated from the bottom to the top using a peristaltic pump to maintain a uniform
P

catholyte in the cell. The feed solution was added to the cell at the top using a peristaltic
CE

pump and circulated to the bottom. The cathode with a plating area of 13.2 x 5.8 cm was
polished using 1200 grit sandpaper, washed with soap, and dried. The cathode was
AC

weighed prior to and after electrowinning to determine the amount of manganese

deposited. When a cathode was pulled out of the cell, it was dipped in 2% potassium
dichromate solution to passivate the deposit surface, washed with water, dried, and
stripped. The stripped cathode was pickled in anolyte to dissolve the residue manganese.
The anode was sandblasted to achieve an average roughness of 30 m to reduce the lead
dissolution. The active area of the anode was 1/2.5 of the cathode plating area to maintain
a higher current density at the anode and suppress the formation of MnO2. The non-active
area was blanketed using CPVC plastic.
The cathode and anode were placed 5.5 cm apart. A high precision DC power supply
was used to maintain a constant current. The cell voltage and current were measured

5
 ACCEPTED MANUSCRIPT

using a data acquisition module and DC Ammeter shunt. The pH measurements were
taken during a run at varying time intervals.

The initial catholyte was prepared using purified feed, and reagent grade ammonium
sulfate and ammonium hydroxide, and sodium sulfite. The initial catholyte composition

PT
was: 12 g/L Mn(II), 150 g/L (NH4)2SO4, 0.3 g/L SO2 and pH 8.3. The typical purified
feed composition was: 32 g/L Mn(II), 130 g/L (NH4)2SO4, 0.3 g/L SO2 and pH 6.7 to

RI
7.0. The anolyte initial composition was the same as the catholyte. The catholyte pH was

SC
measured at room temperature.

NU
2.4 Chemical analysis

Manganese was determined by potentiometric titration of manganous ion with

MA
permanganate ion in neutral pyrophosphate solution. Minor and trace elements were
analyzed using inductively coupled plasma mass (ICP-MS) spectrometry.
D
TE

3 Results and discussion

3.1 Leaching of manganese carbonate in ammonium sulfate solution

P
CE

The manganese carbonate precipitate was pinkish, indicating that there was very little or
no oxidation of manganese carbonate to higher valent oxides (MnO2 and Mn2O3). The
precipitate had a particle size of P80 of 37 m (80% passing 37 m). The leaching results
AC

are summarized in Table 1. The pH and redox potential as a function of time are shown in
Fig. 1. The manganese extraction reached 99.7%. The extractions of Cd, Ca, Co, Mg, Ni,
Ag, Na, Sr and Zn were very high (above 94%). The extractions of Al and Cu were 61.2
and 68.3% respectively, which was dependent on the final pH of the leachate. Iron and
lead were practically not extracted into solution.

As shown in Fig. 1, the solution potential rapidly increased to about 0.7 V vs.
Ag/AgCl, then decreased quickly, and finally slowly. The rapid increase in the potential
resulted from the dissolution of Mn(III) from MnO2 and Mn2O3. The decrease in the
potential was due to the increase in the pH since the redox potential of the manganese
species is dependent on the pH. The pH increased rapidly in the first 30 minutes, and then

6
 ACCEPTED MANUSCRIPT

slowly and finally stabilized. The slurry became darkish brown, indicating the manganese
carbonate precipitate contained some MnO2 or Mn2O3.

The anolyte is used to leach manganese carbonate. Its high redox potential may
inhibit the leaching of manganese carbonate as Mn2O3 or MnO2 can be produced on

PT
manganese carbonate particle surfaces by oxidation, hindering the dissolution of
manganese carbonate. Potassium permanganate and manganese dioxide were added to

RI
investigate the effect of the feed solution redox potential on leaching manganese

SC
carbonate. The addition of 0.05 g/L KMnO4 to the feed solution increased the initial
redox potential to 1.08 V vs. Ag/AgCl. The manganese extraction decreased by 0.8%

NU
mainly due to the following the reaction:

2MnO4- + 3Mn2+ + 2H2O = 5MnO2 + 4H+ MA (8)

The extractions of minor metals were not practically affected by the addition of 0.05
g/L KMnO4. 10 g/L MnO2 was added to the feed solution and the slurry was stirred until
the slurry redox potential did not change with time. The manganese dioxide was
D

separated from the solution by filtration. The filtrate redox potential was 0.978 V vs.
TE

Ag/AgCl, which was slightly lower than that with the addition of KMnO4. The
extractions of manganese and minor metals were not affected by the addition of MnO2.
P
CE

Table 1 Results for the leaching of Baja Mining manganese carbonate

Feed solution Leachate Head Residue Extraction
AC

Amounts 1.5 L 1.5 L 71.1 g 1.05 g

Assay unit mg/L mg/L % % %
Mn 12.6 g/L 31.4 g/L 41.0 8.26 99.7
Cl 1.2 11.8 0.0168
Al 0.032 182 0.45 16.5 61.2
Cd 0.0206 0.0584 0.000082 0.000017 99.7
Ca 2.16 518 1.17 0.14 99.8
Cr 3.45 7.65 0.0019 0.0319 95.0
Co 0.0288 1.2 0.00069 0.00172 99.0
Cu 0.0252 0.0747 0.00026 0.00328 68.3
Fe <0.01 1.11 <0.01 0.24
Pb 0.0515 0.0365 0.00207 0.00913
Mg 2.67 117 0.24 0.05 99.7
Ni 0.0410 9.78 0.0131 0.0934 93.7
Na 1.09 353 0.68 0.10 99.8
Zn 1.24 12 0.0185 0.0131 99.2

7
 ACCEPTED MANUSCRIPT

4.5 0.7

4.0

Potential / V vs. Ag/AgCl

PT
3.5
0.6
3.0 pH

RI
pH

Potential
2.5

SC
0.5
2.0

NU
1.5

1.0 0.4
MA
0 50 100 150 200 250
Time / Minute
Fig. 1 pH and solution potential as functions of reaction time at 50 C.
D
TE

3.2 Purification of manganese leach solution

P

The results for purification of manganese leachate are summarized in Table 2. The
CE

manganese precipitation increased from 0.12 to 3.12% with increasing ammonium sulfide
dosage from 0.5 to 4% of stoichiometric amount relative to manganese (Fig. 2). The
manganese precipitation was lower than the stoichiometric amount. The manganese
AC

sulfide precipitated was probably partially oxidized during filtration and the manganese
was redissolved.

At a dosage of 0.5 to 4% of stoichiometric sulfide relative to manganese, copper,

cadmium and lead was practically completely removed and their concentrations were
below their detection limits. The cobalt precipitation increased from 75.5 to 93.1% with
increasing ammonium sulfide dosage from 0.5 to 1% of stoichiometric amount relative to
manganese, and the cobalt concentration decreased to around 0.10 mg/L. The cobalt
precipitation did not increase with further increasing ammonium sulfide dosage. The
nickel precipitation increased progressively from 55.4 to 99.9% with increasing
ammonium sulfide dosage from 0.5 to 4% of stoichiometric amount relative to
manganese. The zinc precipitation was not sensitive to the ammonium sulfide dosage.

8
 ACCEPTED MANUSCRIPT

Around 93% of zinc was precipitated and its concentration decreased to 0.93 mg/L. The
iron concentration decreased from 1.1 to 0.81 mg/L at a sulfide dosage of 0.5% of
stoichiometric amount relative to manganese and practically did not further decreased
with increasing dosage. The removal efficiencies of Cu, Pb, Cd, Ni, Co and Zn were

PT
related to their sulfide solubilities, their own precipitation kinetics and the precipitation
conditions. Harmful impurities (Cu, Cd, Co, Ni, and Fe) can be reduced to a very low

RI
level by sulfide precipitation.

SC
The amount of ammonium sulfide added at dosages of 0.5, 1.0, 2.0 and 4.0% of
stoichiometric sulfide relative to manganese was 7, 14, 28 and 56 times that required to

NU
precipitate metal impurities respectively. Some impurities were not efficiently
precipitated by the formation of their sulfides. When ammonium sulfide was added to
MA
manganese sulfate solution, pink manganese sulfide was first precipitated. The solution
was saturated with manganese sulfide. As sulfide ions reacted with impurities,
manganese sulfide was redissolved. The color of manganese sulfide gradually became
D

dark. Some impurities were probably precipitated on the surface of manganese sulfide
TE

particles, suppressing the further dissolution of manganese sulfide and the further
precipitation of impurities. When ammonium sulfide was slowly added, the impurities
P

were removed more efficiently because the solution was always saturated with
CE

manganese sulfide.

As expected, arsenic, calcium, chromium, magnesium and sodium did not precipitate
AC

and their concentrations were not given in Table 2. Aluminum concentration decreased
from 180 to about 15 mg/L due to co-precipitation as its hydroxide with manganese
sulfide, and it practically did not further decrease at a dosage above 0.5%. The pH of the
purified solution was 6.7 to 7.0.

9
 ACCEPTED MANUSCRIPT

Table 2 Concentrations of impurities (mg/L) in manganese leachate after purification at

different dosages of ammonium sulfide relative to the stoichiometry for Mn(II)
Feed 0.5% 1.0% 2.0% 4.0%
Al 182 14.5 14.9 15.8 16.1
Cd 0.0584 <0.00005 <0.00005 <0.00005 <0.00005
Co 1.2 0.272 0.091 0.121 0.118

PT
Cu 0.0747 <0.0001 <0.0001 <0.0001 <0.0001
Fe 1.11 0.812 0.832 0.805 0.803
Pb 0.0365 <0.0001 <0.0001 <0.0001 <0.0001

RI
Ni 9.78 4.12 0.477 0.0211 0.0086
Zn 12.0 0.934 0.931 0.943 0.910

SC
NU
100
MA
80
Precipitation / %

60
D

Mn Cd
TE

Co Cu
40 Ni Zn
P
CE

20
AC

0
0 1 2 3 4
(NH4)2S stoichiometric dosage / %

Fig. 2 Precipitation of impurities as functions of ammonium sulfide stoichiometric

dosage relative to Mn(II).

3.3 Reduction of edge effect for manganese electrowinning

The morphology of a metal deposit varies as the local current density changes from point
to point on an electrode surface due to the geometry of the system, the conductivity of the
solution and electrodes, the activation overpotential, the diffusion overpotential and the
hydrodynamics of the system. In electrowinning and electrorefining of metals, a cell with

10
 ACCEPTED MANUSCRIPT

two equal parallel-plane electrodes represents the elementary cell of electrode

arrangement. If the electrode edges do not touch the side walls of the cell, the current
density is higher at the edges than that at the center of the electrode since the current
passes partially in the region outside the gap between the two electrodes. In some cases

PT
the deposit in the central part of the cathode is compact and flat while dendrites are
produced at the edges. The formation of dendrites at the edges of the cathodes can be a

RI
serious problem because the growing dendrites could cause short circuits followed by a
decrease in the current efficiency. In manganese electrowinning, the edge effect is so

SC
significant due to a low solution electrical conductivity and the kinetics of manganese
electrodeposition that it can cause many problems such as bad deposits and a low current

NU
efficiency.
MA
In the industrial electrowinning, the edge effect can be reduced using a smaller anode
and a larger cathode. However it is not effective in a lab cell due to its limited size. If the
edge effect is not reduced or eliminated in the lab cell, the results obtained from the lab
D

cell have limited value in relation to industrial applications. In this study, the edges of the
TE

cathode plating area touched the surfaces at the sides, top and bottom of the cell. There
was only a small opening area in the top of the cell for various purposes. The four edges
P

of the cathode plating area were equal and parallel to the corresponding edges of the
CE

diaphragm. The anode was positioned in such a way to make the cathode current
distribution as uniform as possible.
AC

The manganese deposits with and without the reduction of edge effect are shown in
Figs. 3 and 4 respectively. Without the reduction of edge effect, the nodules at the edges
of the deposit were so significant that they were very close to the diaphragm. With further
electrodeposition, the nodules touched and then damaged the diaphragm. The current
efficiency was 57.5%. With the reduction of edge effect, the edge effect was practically
eliminated, there was no nodule at the edge of the deposit, and the current efficiency
increased to 63-65%.

11
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
Fig. 3 Manganese deposit without the reduction of edge effect at 550 A/m2, 42 C and 24
MA
hours.
D
P TE
CE
AC

Fig. 4 Manganese deposit with the reduction of edge effect at 550 A/m2, 42 C and 24
hours.

3.4 Effect of electrodeposition time

As shown in Fig. 5, with increasing time, the current efficiency decreased with increasing
time while the pH increased. As more manganese was deposited, the surface of the
deposit became rougher and harmful effect of impurities became more significant,
resulting in a lower current efficiency. This phenomenon was similar to those reported by

12
 ACCEPTED MANUSCRIPT

Jacobs et al. (1946) and Araujo et al. (2006). The manganese concentration was
maintained at 12 g/L by decreasing the feed flow rate, resulted in an increased retention
time of the catholyte in the cathode compartment. The catholyte pH was determined by
the difference between the amount of hydroxide generated in the cathode compartment

PT
and the amount of acid transferred through the diaphragm by diffusion and migration.
The hydroxide generated in the cathode compartment was dependent on the manganese

RI
current efficiency and the retention time. The amount of acid transferred to the cathode
compartment was dependent on the diaphragm properties such as water permeability. As

SC
the manganese current efficiency decreased, more water was discharged, generating more
hydroxide per unit volume of catholye, the pH of the catholyte increased if the diaphragm

NU
properties did not change. The surface of the manganese deposit produced in 24 hours
(Fig. 4) was much rougher than that produced in 6 hours (Fig. 6). The real current density
MA
decreased with increasing time. Therefore, the manganese current efficiency effectively
decreased with decreasing current density at a current density below a certain value as
D

discussed in the later section.

P TE
CE
AC

Fig. 5 Current efficiency and catholyte pH as functions of time at 550 A/m2 and 42 C.

13
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
Fig. 6 Manganese deposit with the reduction of edge effect at 42 C, 550 A/m2 and 6
MA
hours.

3.5 Effect of additives on manganese electrowinning

D

Many additives have been tested in this study. The effect of water-soluble
TE

polyacrylamide polymer is discussed here. Polyacrylamide polymer was used with SO2
since polyacrylamide has no reducing power to stabilize the catholyte. As shown in Fig.
P

7, the manganese current efficiency for six hours of electrodeposition increased from 68
CE

to 70.6 % with the addition of 1 mg/L polyacrylamide polymer, then increased very
slightly with increasing dosage from 1 to 3 mg/L, and finally decreased with further
AC

increasing dosage. With increasing addition of polyacrylamide polymer, the manganese

deposit became smoother and denser due to the inhibition effect of polyacrylamide on
manganese deposition. The stronger adsorption of polyacrylamide polymer on the deposit
projections suppressed the nodule growth. Compared with the deposits with no additive
(Fig. 4), the deposit nodules were significantly suppressed in the presence of 2 mg/L
polyacrylamide (Figs. 8 and 9). Without additives, the deposit was not compact,
trapping the electrolyte while with addition of 2 mg/L polyacrymide polymer, the deposit
became nearly fully dense. However the internal stress also increased, resulting in
cracking of deposits. With the addition of 3 mg/L of polyacrylamide polymer, the deposit
cracked and had many pinholes and pits (Fig. 10). The pinholes and pits were
concentrated along these cracking lines probably because these areas had a higher energy

14
 ACCEPTED MANUSCRIPT

which facilitated pitting. The internal stress was tensile since the deposit was bent toward
to the anode. The formation of pinholes and pits is mainly caused by the attachment of
hydrogen bubbles on a cathode. A hydrogen bubble has a screening effect on the deposit.
The screening effect increases with increasing bubble size. When a bubble is sufficiently

PT
large, the cathode is not in electrical contact with the electrolyte and the cathodic
protection is lost. The screened manganese under hydrogen bubbles starts to re-dissolve

RI
significantly. The deposition of impurities (Ni, Co and etc.) is less affected, resulting in
an impurity-rich area, which accelerates the re-dissolution of manganese because of

SC
lower hydrogen overpotential on these impurities. The detachment of a gas bubble is
determined by the surface and interfacial tensions as discussed by Bozhkov et al. (1990).

NU
The addition of polyacryamide polymer caused the catholyte viscosity to increase and the
surface and interfacial tension to change to such a degree that pinholes and pits were
MA
eventually formed.
D

72
TE

70
Current efficiency / %

P
CE

68
AC

66

64

62
0 1 2 3 4
Dosage / mg/L

Fig. 7 Current efficiency as a function of polyacrylamide polymer dosage at 550 A/m2,

42 C and 6 hours.

15
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
Fig. 8 Manganese deposit with the addition of 2 mg/L polyacrylamide polymer at 550
A/m2, 42 C, and 24 hours. MA
D

(a)
TE

(b)
Fig. 9 Manganese deposits with no additive (a) and with 2 mg/L polyacrylamide polymer
P

(b) at 550 A/m2, 42 C, and 24 hours.

CE
AC

Fig. 10 Manganese deposits with the addition of 3 mg/L polyacrylamide polymer at 550
A/m2, 42 C, and 24 hours.

16
 ACCEPTED MANUSCRIPT

3.6 Effect of pH

Fig. 11 shows the manganese current efficiency as a function of pH at a plating time of 6

hours. The catholyte pH was controlled by addition of sulfuric acid or ammonia
hydroxide. With increasing pH, the manganese current efficiency first increased quickly,

PT
then slowly, reached a maximum value at pH 9.2, and finally decreased. The increase in
the pH initially suppressed the hydrogen evolution, resulting in a higher manganese

RI
efficiency. After the pH exceeded a value (9.2), the increase in the pH also inhibited the

SC
manganese deposition probably due to the formation of manganese oxide on the surface
according to Jacob et al. (1946), giving a lower current efficiency. This relationship

NU
between the current efficiency and pH was similar to that reported by Louis and Martin
(1976), but they did not report the current efficiency at a pH above 9.2. As shown in Fig.
MA
12, the morphology of the deposit changed very slightly at pH 8.2 to 9.2. It was quite
different at pH 9.5 and became significantly different at pH 9.8. At pH 9.5, there were
some very shallow tailed pits (or depressions) due to the adhesion of hydrogen bubbles to
D

the deposit. At pH 9.8, there were numerous tiny pits (or depression). Although the
TE

highest current efficiency is reached at pH 9.2, the loss of ammonia (mainly due to
evaporation) is significant and Mn(II) is readily oxidized by air. Therefore the operation
P

pH should range from 8.5 to 9.0 for industrial application.

CE

74
AC

72

70
Mn CE / %

68

66

64

62
8 8.5 9 9.5 10
pH

Fig. 11 Current efficiency as a function of catholyte pH at 2 mg/L polyacrylamide

polymer at 550 A/m2, 42 C and 6 hours.

17
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
(a) (b)
MA
D
P TE
CE
AC

(c) (d)
Fig. 12 Manganese deposits prepared at pH 8.2 (a), pH 9.2 (b), pH 9.5 (c) and pH 9.8 (d)
with the addition of 2 mg/L polyacrylamide polymer at 550 A/m2, 42 C and 6 hours.

3.7 Effect of current density

The current efficiency first increased with increasing current density, reached a maximum
value at 550 Am-2, and decreased with further increasing current density (Fig. 13). This
result is in good agreement with that reported by Louis and Martin (1976). The hydrogen
evolution can be divided into two parts. The first part is due to the discharge of water
while the second part results from the re-dissolution of manganese (Gamali and
Vorozhko, 1980). The manganese current efficiency is determined by the rate ratio of

18
 ACCEPTED MANUSCRIPT

manganese deposition to hydrogen evolution. At a current density below 550 A/m2, the
rate of manganese deposition increased with increasing current density faster than that of
hydrogen evolution, resulting in an increase in the manganese current efficiency. At a
current density above 550 A/m2, the rate of manganese deposition increased slower than

PT
that of hydrogen evolution, resulting in a decrease in the manganese current efficiency.
The deposit was less dendritic at a lower current density than that at a higher current

RI
density. As shown in Fig. 14, the deposit produced at 450 Am-2 was compact.

SC
72

NU
71

70
MA
Mn CE / %

69
D

68
TE

67
P

66
CE

400 450 500 550 600 650

-2
Current density / Am
AC

Fig. 13 Current efficiency as a function of current density with the addition of 2 mg/L
polyacrylamide polymer at 550 A/m2, 42 C and 6 hours.

Fig. 14 Manganese deposit produced with the addition of 2 mg/L polyacrylamide

polymer at 450 Am-2, 42 C, and 24 hours.

19
 ACCEPTED MANUSCRIPT

3.8 Effect of chloride

Baja Mining manganese carbonate contains some chloride as the solution used in
leaching and hence manganese precipitation is seawater based. The chloride will

PT
accumulate in the leach solution. Therefore manganese electrowinning was conducted at
different chloride concentrations to investigate the effect of chloride on manganese

RI
electrodeposition. The manganese current efficiencies at different chloride concentrations

SC
are summarized in Table 3. The manganese current efficiency increased slightly with
increasing chloride concentration from 50 to 100 mg/L and then was insensitive to the

NU
chloride concentration. The morphology of the deposit practically did not change with
increasing chloride concentration. However, the anode corrosion increased with
MA
increasing chloride concentration.

Table 3 Manganese current efficiency for 24-hour deposition at different chloride

D

concentrations (2 mg/L polyacrylamide polymer).

Cl / mg L 50 100 200 500 1000 2000
TE

Current efficiency / % 67.1 68.3 68.5 68.9 68.9 68.3

P

3.9 Effect of catholyte circulation rate

CE

The catholyte should be circulated at an appropriate flow rate to maintain the uniform
catholyte composition and control the catholyte temperature in the cathode compartment.
AC

Manganese electrowinning was carried out at different circulation rates. The manganese
current efficiency increased by 2.4% with increasing catholyte rate from 0.01 to 0.02
L/min. and it was practically insensitive to the circulation rate at a higher circulation rate
(Fig. 15). Further increasing catholyte circulation rate did not improve the manganese
deposition. The hydrogen evolution was so significant that increasing circulation rate to
0.20 L/min. did not contribute to the mass transfer very much. In another set of tests
(without the addition of polyacrylamide polymer), with increasing circulating rate from
0.1 to 0.2 L/min., the current efficiency practically did not change. However with
increasing circulation rate from 0.2 to 4.0 L/min., the current efficiency and the catholyte
pH decreased from 66 to 62% and 8.5 to 8.3 respectively. The decrease in the pH
indicates that the transfer of protons to the catholyte was enhanced at a higher circulation

20
 ACCEPTED MANUSCRIPT

rate in the way that will be discussed in Section 3.11. As the catholyte pH was below 9.2,
the current efficiency decreased with decreasing pH.

PT
RI
SC
NU
MA
D
TE

Fig. 15 Current efficiency as a function of catholyte circulation rate with the addition of 2
mg/L of polyacrylamide polymer at 550 A/m2, 42 C and 6 hours.
P
CE

3.10 Chemical composition of manganese deposit

The analysis of manganese deposit was conducted at Manganese Metal Company, South
AC

Africa. The typical results are summarized in Table 4. The manganese deposit was very
pure. A low oxygen content mainly resulted from dense deposition of manganese because
of inclusion of less electrolyte and hydroxide.

Table 4 Chemical composition of manganese deposit

Mn * Mn ** S C O N H Al B Ba
/% (Met)/% /% /% /% /% /% /% /% /%
>99.8 >99.99 0.020 0.001 0.10 0.001 0.015 0.0002 0.0007 <0.0005
Ca Co Cr Cu Fe K Mg Na Ni Zn
/% /% /% /% /% /% /% /% /% /%
0.0002 <0.0001 0.0002 <0.0001 0.0005 0.0001 0.0001 0.0020 <0.0001 <0.0001
*Total manganese on a total weigh basis
**Metallic metal calculated excluding non-metallic elements

21
 ACCEPTED MANUSCRIPT

3.11 Selection of diaphragm

The use of an appropriate diaphragm is very important for successful manganese

electrowinning. An ideal diaphragm should have a proper porosity (or water

PT
permeability) to maintain the catholyte pH at a target value. If a diaphragm is too tight
(less porous), the catholyte pH will exceed the optimum value while if it is too loose

RI
(porous), the catholyte pH will be below the optimum value. The hydrodynamic factor
influences the performance of a diaphragm. The hydrodynamic factor in an industrial cell

SC
is more significant than that in a small lab cell. Therefore a tall cell with a plating area of
1m (height) x 0.086 m (width) was built to test diaphragms. In a small lab cell, the

NU
catholyte pH gradually increased and then stabilized around 8.9 while in a tall cell , the
pH decreased slowly, and then quickly to 2.2 ( with the use of Diaphragm 1 in Fig. 16)
MA
although the catholyte flow velocity in the tall cell was the same as that in the lab cell.
This was mainly related to the level difference between the catholyte and anolyte and the
D

different strengths of agitation due to gas evolution in the two cells. The level difference
is dependent on the following parameters: (1) Flow rate (feed addition rate) from the
TE

catholyte to the anolyte (a higher flow rate results in a higher catholyte level); (2)
P

Diaphragm water permeability(a lower diaphragm water permeability results in a higher

CE

catholyte level); (3) Solution viscosity (a higher solution viscosity results in a higher
catholyte level); (4) Bulk densities of anolyte and catholyte (a lower catholyte bulk
AC

density and higher anolyte bulk density result in a higher catholyte level); (5) Catholyte
circulation rate (catholyte flow velocity). A higher level difference obviously suppresses
the transfer of acid to the catholyte.

The bulk density should be: the real density × (1 - void fraction). The void fraction
depends on the volume of gas bubbles in the catholyte or anolyte. In an electrowinning
cell, the volume of gas bubbles in the catholyte or anolyte depends on the current density,
gas current efficiency, the solution viscosity, bubble size and the cell height. A higher
current density and gas current efficiency obviously results in more gas and a lower bulk
density. A higher viscosity results in a slower movement of a bubble in solution.
Therefore the retention time of a bubble in the catholyte or anolyte is longer, and the bulk
density is smaller. A smaller bubble rises more slowly than a larger one, resulting in a
lower bulk density. The bubble size depends on the gas and the solution properties. Some

22
 ACCEPTED MANUSCRIPT

additives can increase the solution viscosity and change bubble sizes, affecting the level
difference between the anolyte and catholyte.

The bulk density decreases with increasing height because more bubbles accumulate.
The effect of the cell height on the bulk density is schematically shown in Fig. 17. If the

PT
bulk density of anolyte in a 20 cm long compartment is 0.99, the bulk density of anolyte
in a 100 cm long compartment will about be 0.97. Therefore the level difference between

RI
catholyte and anolyte in a lab can be quite different than that in an industrial (or tall) cell.

SC
As the void fraction increases with increasing height, the solution effective conductivity
deceased, resulting in a lower current density at the upper part of the cathode or anode

NU
and then a thinner and smoother manganese deposit. This phenomenon was observed in
this study. MA
If the catholyte is circulated in a cell from the bottom to the top, there is a pressure
drop or solution head change. This changes the pressure difference across the diaphragm
as shown in Fig. 18. More catholyte flows through the diaphragm near the bottom. The
D

amount of the catholyte flowing through the diaphragm near the bottom can be higher
TE

than the feed rate. The anolyte can flow back to the cathode compartment near the top.
The effect of catholyte flow velocity depends on the diaphragm water permeability and
P

the height of a cell. For a lab cell, this effect is negligible while for an industrial size tall
CE

cell, it can be significant. A tighter (less porous) diaphragm lessens the effect of catholyte
circulation in flow through the diaphragm with height. As shown in Fig. 16, with the use
AC

of a tighter diaphragm (Diaphragm 2), the catholyte pH was stabilized. The catholyte
(circulation) flow should be uniformly distributed to avoid local higher pressures.

The hydrogen generated at cathodes and oxygen generated at anodes move up and act
as gas-lifts to push the catholyte and anolyte upwards respectively, affecting the solution
pressure distribution in the two compartments. The bubble sizes and current efficiencies
of hydrogen and oxygen are different and hence their effects on the pressure distribution
in the two compartments are not the same. Therefore the local pressure difference across
a diaphragm varies, affecting the local flow through the diaphragm. In a small lab cell,
this gas-lift effect can be neglected while in a tall industry cell, it can influence the local
transfer of acid through the diaphragm. The gas-lift effect is applied in some industry
manganese cells to draw the common catholyte into each cathode compartment through

23
 ACCEPTED MANUSCRIPT

the holes in the lower part of the cathode frame sides and eject the catholyte into the
common catholyte chamber through the holes in the upper part of the cathode frame
sides.

10

PT
RI
8

SC
Small cell Diaphragm 1
pH

6 Tall cell Diaphragm 1

NU
Tall cell Diaphragm 2

MA
4
D

2
0 5 10 15 20 25
TE

Time / hour
P

Fig. 16 Catholyte pH as a function of time

CE

1 1
AC
Void fraction

Bulk density

Void Fraction
Bulk density

0 0
0.0 0.2 0.4 0.6 0.8 1.0
Height from the bottom of diaphragm / m
Fig. 17 Schematic diagram of void fraction and bulk density as functions of the distance
from electrode bottom.

24
 ACCEPTED MANUSCRIPT

1.0
No circulation
Pressure drop 0.8 Circulation for a tall cell
0.6 Circulation for a lab cell

PT
0.4
0.2

RI
0.0

SC
-0.2
0.0 0.2 0.4 0.6 0.8 1.0

NU
Height from the bottom of diaphragm / m
Fig. 18 Schematic diagram of pressure difference across the diaphragm with and without
MA
catholyte circulation.
D

4 Conclusions
TE

Manganese carbonate produced from solvent extraction raffinate from the El Boleo
P

project can be readily leached into acidic ammonium sulfate solution with a high
CE

manganese extraction (99.7%). The extractions of trace and minor metals (cadmium,
chromium, cobalt, nickel, zinc, magnesium and sodium) were very high (over 94%) while
AC

the extractions of copper, aluminum, and iron were dependent on the final pH of the
leachate. Iron and lead were practically not extracted into the solution.

Copper, lead and cadmium can be removed completely by sulfide precipitation at a

dosage of 0.5 % of stoichiometric sulfide relative to manganese. Cobalt can be removed
to 0.1 mg/L at a dosage of 1.0 % of stoichiometric sulfide relative to manganese. With
further increasing sulfide dosage, the cobalt concentration was practically reduced. While
the nickel concentration decreased from 9.8 to 0.009 mg/L progressively with increasing
dosage to 4.0% of stoichiometric sulfide relative to manganese. A slow addition of
ammonium sulfide improved the efficiency of impurity precipitation.

The edge effect is significant in manganese electrodeposition. The edge effect can be
eliminated in an appropriately designed cell. Without edge effect, manganese was

25
 ACCEPTED MANUSCRIPT

uniformly deposited on the cathode and the manganese current efficiency increased by
5.5%. The manganese current efficiency decreased with increasing time since the deposit
became rougher and the harmful effects of impurities became more significant. The
addition of polyacrylamide polymer can significantly improve the manganese deposit

PT
quality. A compact manganese deposit can be produced with an improved current
efficiency at an appropriate dosage. At a dosage above a certain value, the manganese

RI
deposit cracked due to the increased internal tensile stress. Pits and pinholes were formed
due to the increased viscosity and change in the surface tension and interfacial tension.

SC
With increasing pH and current density, the current efficiency first increased, reached a
maximum value, and decreased.

NU
Because of the hydrodynamics issue, more acid is transferred through a diaphragm to
MA
the catholyte in a tall cell than that in a small lab cell. The final selection of diaphragm
should be conducted in an industrial size cell.
D

Acknowledgements
TE

The authors would like to thank Baja Mining Corp. for funding this project and allowing
for publication.
P
CE

References
AC

Araujo, J., Castro, M. and Lins, V., 2006. Reuse of furnace fines of ferro-alloys,
Hydrometallurgy, 84, 204-210.
Bozhkov, C., Ivanov, I., Rashkov, ST., 1990. The relationship between the growth rate of
hydrogen bubbles and the duration of the induction period in the electrowinning of
zinc from sulfate electrolytes, Journal of Applied Electrochemistry, 20,447-453.
Dreisinger, D. B., W. Murray, E. Norton, E., Baxter, K., Holmes, M., Molnar, R, 2008.
The Boleo copper-cobalt-zinc-manganese project, Can. Metall. Quart., 47, 357-367.
Coleman, T. W., Griffin, R. A., 1984. Method for electrodepositing metallic manganese,
US patent No. 4478697.
Griffith, F. S., 1950. Recovery of manganese by electrolysis, US Patent No. 2497725.
Goddard, J. B., 1979. Method for electrolytic deposition of manganese, US Patent No.
4149944.

26
 ACCEPTED MANUSCRIPT

Gamali. I. V., Vorozhko, A. V., 1980. Kinetic limitation in the electrodeposition of

manganese, Soviet J. Electrochem., 16, 714-717.
Hammerquist, W. L., 1951. Manganese electrowinning process, US Patent No. 2559874.
Jacobs, J. H., 1962. ‘Electrolytic manganese production’, US Patent No. 3034973.

PT
Jacobs, J. H., Hunter, J. W., Yaroll, W. H., Churchward, P. E. and R. G. Knikerbocker,
R. G., 1946, ‘Operations of electrolytic manganese pilot plant at Boulder City,

RI
Nevada’, (Technical Report 463, United States Department of Interior- Bureau of
Mines, Boulder City, USA).

SC
Lu, J., Dreisinger, D. B., Gluck, T., 2014. Manganese Electrodeposition – A literature
review, Hydrometallurgy, 141, 105-116.

NU
Louis. P., Martin, J. P., 1976. Increased current efficiency during the electrodeposition of
manganese from sulfate electrolysis’, (Report No. 1859, National Institute for
MA
Metallurgy, South Africa).
Zhang, W., Cheng, C., 2007. Manganese metallurgy review. Part I: Leaching of
D

ores/secondary materials and recovery of electrolytic/chemical manganese dioxide,

TE

Hydrometallurgy, 89, 137-159

P
CE
AC

27
 ACCEPTED MANUSCRIPT

Highlights

High purity Mn was produced from MnCO3 precipitate by leaching-purification-

electrowinning.
The design of a lab cell was analyzed to eliminate the edge effect.
Electrowinning conditions were investigated to optimize manganese electrowinning.

PT
A compact manganese deposit can be produced with the addition of polyacrylamide
polymer.
The diaphragm selection as an important part of a cell design was analyzed.

RI
SC
NU
MA
D
P TE
CE
AC

28