Professional Documents
Culture Documents
Laboratory Study of Electro-Coagulation - Otation For Water Treatment
Laboratory Study of Electro-Coagulation - Otation For Water Treatment
Laboratory Study of Electro-Coagulation - Otation For Water Treatment
Abstract
An electro-coagulation–flotation process has been developed for water treatment. This involved an electrolytic
reactor with aluminium electrodes and a separation/flotation tank. The water to be treated passed through the reactor
and was subjected to coagulation/flotation, by Al(III) ions dissolved from the electrodes, the resulting flocs floating
after being captured by hydrogen gas bubbles generated at cathode surfaces. Apparent current efficiencies for Al
dissolution as aqueous Al(III) species at pH 6.5 and 7.8 were greater than unity. This was due to additional reactions
occurring in parallel with Al dissolution: oxygen reduction at anodes and cathodes, and hydrogen evolution at
cathodes, resulting in net (i.e. oxidation+reduction) currents at both anodes and cathodes. The specific electrical energy
consumption of the reactor for drinking water treatment was as low as 20 kWh (kg Al)1 for current densities of 10–
20 A m2. The water treatment performance of the electrocoagulation process was found to be superior to that of
conventional coagulation with aluminium sulphate for treating a model-coloured water, with 20% more dissolved
organic carbon (DOC) being removed for the same Al(III) dose. However, for a lowland surface water sample, the two
processes achieved a similar performance for DOC and UV-absorbance removal. In addition, an up-flow
electrocoagulator configuration performed better than a horizontal flow configuration, with both bipolar and
monopolar electrodes. r 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Electrocoagulation; Current efficiency; Specific electrical energy consumption; Coagulation; Aluminium sulphate; Water
treatment
0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 1 1 8 - 5
J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078 4065
(c) to define the reactor’s optimum operating condi- schematically in Fig. 1. The electrocoagulator reactors
tions, to maximise the amount of impurities were designed to provide various hydraulic and electro-
removed with minimum specific energy consump- lytic configurations. In principle, the reactor consisted of
tion. aluminium (minimum Al 95.45%, maximum Al 97.35%,
J. Smith & Sons Ltd., Delta) plate anodes and cathodes
(210 mm 46 mm) in a PMMA electrolysis chamber
connected to a floc separation unit. The reactor could be
2. Experimental procedures
configured with water up-flow or cross-flow, and with
bipolar electrodes connected in series via the water, and
2.1. Electrochemical studies
monopolar electrodes connected in parallel via copper
connectors, as shown schematically in Fig. 2, for the
The laboratory test system consisted of an electro-
case of multiple pairs of electrodes.
coagulator unit with a 500 W (70 V/30 A) DC power
In the case of monopolar electrodes, a voltage U is
supply (Wayne Kerr Electronics Ltd., Model AP7030), a
connected between n0 pairs of anodes and cathodes,
peristaltic pump (Watson-Marlow 505-S, maximum
connected in parallel via copper bus bars, causing a
capacity 14 dm3 h1), a flow meter (Platon, 1–
current n0 I to pass across electrode/solution interfaces
16 dm3 h1), a glass raw water reservoir (10 dm3) and
via the bulk solution, I being the current that passes
an outlet reservoir (polyethylene, 10 dm3), as shown
between an individual anode/cathode pair. By contrast,
if a voltage n0 U is applied between two feeder electrodes,
between which ðn0 1Þ bipolar electrodes are interposed,
Power supply current would flow across one feeder/solution interface,
Al + - Al through the solution, sequentially into and out of each
_ _
2 Al → 2Al 3 + + 6 e 6 H + + 6e → 3H2 of the bipoles, and finally to the second feeder electrode.
As the same current passes through n0 electrode pairs,
the maximum theoretical current efficiency is also n0 ;
rather than unity. Bipolar reactors have advantages of
Hydrogen bubbles being easier to engineer, as electronic connections need
be made only to the feeder electrodes, so inter-electrode
gaps can be smaller, and costs of higher voltage/lower
current transformer rectifiers are lower for the same
total power output than for the lower voltage/higher
current required for monopolar reactors. However, the
large terminal voltage between a pair of feeder electrodes
Inlet Outlet also causes a proportion of the current to flow in the
Reservoir Rservoire
solution, bypassing the bipolar electrodes, as shown in
Fig. 2.
Fig. 1. Schematic of laboratory-scale electrolytic reactor sys- Current efficiencies for Al dissolution were deter-
tem. mined using a glass electrochemical cell containing a
Monopolar Bipolar
bipolar
Cathodes - + electrodes -
Anode -+ -+ -+ -+ Cathode
feeder feeder
electrode electrode
Anodes +
Bypass current
Cell voltage, U Cell voltage, n’U
Cell current, n’I Cell current, I
Parallel connections to Series connections to bipoles
each electrode made via bus bar made via electrolyte
1 cm2 Al plate, a platinum plate and a saturated calomel Electrocoagulation was operated under the following
electrode, as working, counter, and reference electrodes, conditions:
respectively. The anolyte and catholyte were separated
by a porous sintered glass disc. The working electrode * water flow rate of 102 m3 h1;
potential was controlled by a potentiostat (EG&G * two electrode arrangements: bipolar electrodes con-
Instruments, 273A); its facility to compensate the nected in series via the water and monopolar
applied potential for any IR potential drop between electrodes connected in parallel;
the working electrode and the tip of the glass Luggin * operating electrical current densities of 3–25 A m2,
probe containing the reference electrode was used, as the which produced Al(III) concentrations of 2.4–
solutions had low ionic conductivities. 14 g m3, determined as the sum of the concentra-
In subsequent experiments at constant current density tions in the treated effluents and floc sludge; and
(20, 30 and 50 A m2 for the anode and 20 A m2 for the * pH values of 6.5 and 7.870.2.
cathode), the cell voltage was recorded over electrolysis
For the purpose of comparison, the coagulation
times between 3 and 16 h. Samples of catholyte and
performance of dosed aluminium sulphate (AS) with
anolyte were taken and the dissolved aluminium
the two raw waters were also evaluated, using the same
concentrations determined by atomic absorption spec-
pH values of 6.5 and 7.870.2 as for the electrocoagula-
troscopy (Perkin-Elmer, Zeeman 5100PC). The current
tion tests, and Al(III) concentrations of 2–14 g m3.
efficiency ðFÞ for the production of dissolved Al(III) by
The experimental procedures for the coagulation jar
the passage of electrical charge Q (C) in London tap
tests, the analysis of water samples, the preparation of
water of volume V (m3) was calculated using Faraday’s
model-coloured water, and the sources of coagulants,
law:
chemical reagents and humic substances were the same
3FV ½AlðIIIÞ as that described previously [11].
F¼ : ð1Þ
Q
The electrical conductivity and pH of the water were
measured as a function of electrolysis time from initial 3. Results and discussion
values of 7 102 S m1 and pH 7.8. The specific
electrical energy consumption was calculated as a 3.1. Current efficiencies for Al dissolution
function of applied cell voltage ðUÞ:
As shown in Table 2, apparent current efficiencies,
1 nFU defined by Eq. (1), for dissolution of the Al anodes were
seec ðkWh ðkg AlÞ Þ ¼ ; ð2Þ
3:6 103 MAl F greater than unity. It was assumed that Al(III) species
were formed directly by a single reaction such as (3) and
where n is the charge number of the reaction, i.e. moles
is discussed below.
of electrons per mole of aluminium dissolved, and was
Table 3 lists equivalent data for Al dissolution from
taken to be 3.
the cathode at both constant current density and
Table 1
Properties of the two raw waters
Table 3
Charge for Al dissolution at the cathode
1.4
1.2 O2
1.0
0.8
H2O
0.6
Electrode potential (SHE) / V
0.4 jA = 20A/m 2
0.2
0.0 H+ 1
-0.2
-0.4 H2
-0.6 Al(OH)3
-0.8
-1.0 (Amorphous) 2
-1.2 Al 3+ jC = -20A/m2
-1.4
-1.6 Al(OH)4-
-1.8
-2.0
-2.2
-2.4 Al
-2.6
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
pH
Fig. 3. Variations in solution pH and potentials of Al anode and cathode at a current density of 20 A m2 for 3 h (potential–pH
diagram calculated from the data of Bard et al. [18] for a dissolved aluminium activity of 104 M and 298 K).
-1 Al 3+ Al(OH)3 (amorphous)
-2
-3 AlOH 2+
log (Al(III) activity)
-4 Al(OH) 4-
-5
AlO +
-6
-7
-8 Al(OH) 3 (aq)
-9
-10
0 2 4 6 8 10 12 14
pH
Fig. 4. Activity–pH diagram for Al(III) species in equilibrium with Al(OH)3 (amorphous) (data from Bard et al. [18]).
J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078 4069
constant cathode potential. For electrolysis at constant i.e. the measured current I ¼ AA jnet pjAl
A
AA ; where F is
current densities of 720 A m2 over 3 h, the arrows in the Faraday constant (96 485 C mol1), MAl the molar
Fig. 3 correspond to the potential–pH trajectories for mass of aluminium (26.982 g mol1), and u the aqueous
anodes (region 1) and cathodes (region 2), respectively, volumetric flow rate (m3 s1). According to Fig. 3,
superimposed on a (meta-stable) potential–pH diagram EHþ =H2 oEA oEO2 =OH ; so hydrogen evolution by reac-
for the Al–H2O system, for which amorphous Al(OH)3 tion (4) did not occur on the Al anodes at the pHs used,
was the Al(III) phase considered. For kinetic reasons, whereas reaction (5) accompanied (3), resulting in
the most stable Al(III) oxide phase, a-Al2O3, is not A
jnet A
¼ jAl þ jOA2 : ð9Þ
formed by precipitation from aqueous solutions. The
ions Al13O4(OH)7+ 24 (e.g. [12]) were not included in the Al cathode reactions:
calculations, as their Gibbs energy of formation appears C
Al þ 4OH -AlðOHÞ
jAl 4 þ3e ; ð10Þ
somewhat uncertain. Fig. 4 shows the equivalent dia-
gram predicting the solubility of amorphous Al(OH)3, C
jH 2H2 O þ 2e -H2 þ 2OH ; ð11Þ
2
most likely to be the initial product of hydrolytic
precipitation of Al(III) species from aqueous solutions. jOC2 O2 þ2H2 O þ 4e -4OH ; ð12Þ
Even so, the predicted Al(III) solubilities at the bulk
C C C
solution pH of 6.5 and 7.8 are predicted to be so small jnet ¼ jAl þ jH 2
þ jOC2 ; ð13Þ
that strong passivation of anodic dissolution would be
C
expected, due to adsorption/precipitation at the anode. jAl AC MAl jnet AC MAl
½AlIII C ðg m3 Þ ¼ o : ð14Þ
Several possibilities could explain the apparent super- 3Fu 3Fu
Faradaic yield of dissolved aluminium species: Hence, some of the dissolved Al(III) is predicted to be
* dissolution of air-formed oxide films, at short times, produced by hydrogen-evolving cathode surfaces, if
but this extra source of dissolved Al(III) species their operating potential at a given current density is
would become less significant with increasing electro- > EAlðOHÞ4 =Al :
lysis time; The oxygen overpotential at the cathode
* the applied current densities were due to the sum of (ZC
O2 ¼ EC EO2 =H2 O ) is so large at the respective
anodic oxidation and cathodic reduction processes; operating potentials (Fig. 5) that the oxygen reduction
and current densities would probably have been controlled
* the aluminium oxidation state of the dissolving by mass transport:
species was less than that implied by reaction (3). jLC ðO2 Þ ¼ 4Fkm
C
½O2 ; ð15Þ
C
For aluminium anodes and cathodes of area AA and where km is the (mean) mass transport rate coefficient
AC ; respectively, the principal electrolysis reactions and (ca. 10 m s1) at the cathode, which is determined
5
their respective partial current densities ðji Þ are believed largely by the rate of gas evolution (e.g. [13]) rather than
to be as follows: imposed convection flow. As the aqueous solubility of
oxygen at atmospheric oxygen pressures is about
Al anode reactions:
0.25 mol m3 [14], Eq. (15) predicts a current density
A
jAl Al-Al3þ þ 3e ; ð3Þ for reaction (5) of order 1 A m2. Hence, at an applied
current density of 20 A m2 at an Al anode, 21 A m2
A
jH 2
2H2 O þ 2e -H2 þ2OH ; ð4Þ could be used for anodic dissolution and 1 A m2 for
cathodic reduction of oxygen, neglecting any hydrogen
A
jO2
O2 þ2H2 O þ 4e -4OH : ð5Þ evolution for anode potentials EHþ =H2 oEA oEO2 =OH :
This would produce an apparent current efficiency of
Bulk (neutral) solution:
1.05, which is lower than the experimental data in Table
Al3þ þ3H2 O-AlðOHÞ3 þ3Hþ : ð6Þ 2, and could be attributed to uncertainty in the mass
Hence, the net current supplied to the reactor would transport rate coefficient and the difficulty of measuring
be the algebraic sum of the partial currents due to Al(III) concentration in water sludge samples. The much
anodic oxidation by reaction (3) and cathodic reduction lower oxygen overpotential (ZA O2 ¼ EA EO2 =H2 O ) at the
by reaction (5), and at potentials EA oEHþ =H2 ; reduction anode would probably result in reaction (5) not being
by reaction (4) transport controlled, but in any case, the absence of
hydrogen evolution would result in much lower mass
A A A A
jnet ¼ jAl þ jH 2
þ jO2
: ð7Þ transport rates at the anode:
Hence, j A ðO2 Þp4Fkm
A
½O2 : ð16Þ
A
jAl AA MAl jnet AA MAl Experimental data [15] for the behaviour of alumi-
½AlIII A ðg m3 Þ ¼ X ; ð8Þ
3Fu 3Fu nium in aqueous solutions suggested the presence of
4070 J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078
1.4
1.2
O2
1.0 H 2O
both aluminium hydride phases (e.g. AlH3) and solution Regardless of whether Al(III) species are formed
species (e.g. AlH+ 2 ). Perrault’s thermodynamic data directly by reaction (3) or by reaction (17), the
were used to compute the potential–pH diagram shown concurrent reduction of oxygen and, at the cathode,
in Fig. 5, which considers the metastable amorphous hydrogen evolution, have the net effect of increasing
Al(OH)3 phase rather than the more stable corundum the local pH, thus enhancing the solubility of Al(OH)
4
(a-Al2O3) as the former phase was formed hydrolytically ions:
from aqueous solutions. In common with other very log½AlðOHÞ
4 DpH 13:49 ð22Þ
reactive elements, such as Ti, the elemental metal has no
area of stability. However, using the same data set, the while the bubbles from hydrogen evolution enhance the
Nernst equation for the reaction dispersion of Al(III) species from electrode surfaces.
This explains why anode passivation does not restrict Al
Al þ 3Hþ þe -AlH2þ þH2 ð17Þ dissolution rates more severely at neutral bulk solution
may be calculated for 298 K as pHs.
Anomalous pitting corrosion behaviour of aluminium
EAl=AlHþ2 ðSHEÞ ðVÞ ¼ 0:003 0:177pH
has been reported [16], and the rate of hydrogen
0:059 logðAlH2þ Þ evolution increases with the increase in electrode
0:059 log PH2 : ð18Þ potentials over a wide pH range. However, those effects,
again leading to anomalous Faradaic yields, appear in
Hence, at pH 7, the reversible potential for reaction (17) the pitting potential range of ca. 0.5 V (SHE), which
is ca. 1.24 V vs. SHE, which is less negative than the was not adopted by either cathode or anode in the
observed cathode potentials (Fig. 3) at 20 A m2; hence, driven electrochemical reactor used in the work reported
reaction (17) is a possible source of dissolved aluminium, here.
though, depending on the concentration achieved and
the pH of an oxygenated solution, it could be followed 3.2. Specific electrical energy consumption (seec) for
by the homogeneous reaction monopolar and bipolar connections—different electrode
2AlH2þ þO2 þ4H2 O-2AlðOHÞ3 þ4Hþ : ð19Þ arrangements
At sufficiently low potentials (Fig. 5), dissolution would Fig. 6 shows the specific electrical energy consumption
be suppressed due to the formation of the AlH3 phase (seec) vs. current density for the two electrode connec-
AlH2þ þ2Hþ þ4e -AlH3 ðcÞ; ð20Þ tion configurations: bipolar electrodes with cells in
series, and monopolar electrodes with cells in parallel.
EAlH2þ =AlH3 ðSHEÞ ðVÞ ¼ 1:279 0:0296pH The seec of electrocoagulation for the application of
drinking water treatment was very low; at current
þ 0:0148 logðAlH2þ Þ: ð21Þ densities of 10–20 A m2, it was only 20 kWh (kg Al)1
J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078 4071
100 bipolar
80
monopolar
60
40
20
0
0 20 40 60 80 100
Current density, j / A m-2
Fig. 6. Variation of specific electrical energy consumption with current density and electrode connection mode.
UV-abs and colour (m-1) and DOC (g C m-3)
40
35
30 Absorption at 254 nm
25
20
15
10
DOC
5
Colour
0
0 2 4 6 8 10
[Al(III)] / g m-3
Fig. 7. Performance of electrocoagulation reactor at pH 7.8 with monopolar electrodes (2 anodes and 2 cathodes).
for the monopolar electrode arrangement and gaps required to minimise ohmic potential losses in low
20–30 kWh (kg Al)1 for the bipolar electrode arrange- conductivity waters.
ment. Over the wider range of 10–60 A m2, seec values
were about 20–80 kWh (kg Al)1, essentially indepen- 3.3. Treatment performance for model-coloured water
dent of whether monopolar or bipolar electrodes were with an up-flow electrolytic reactor
used. Hence, other factors rather than seec alone, would
influence the selection of the optimal electrode config- Uncertainties in the effective charge number ðnÞ of
uration in practice, notably the more mechanically facile the Al dissolution reaction and current efficiencies greater
connection of multiple electrodes using the bipolar than unity, discussed above, preclude the use of the applied
configuration, especially with the narrow inter-electrode current as a quantitative measure of the rate of aluminium
4072 J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078
35
Absorption at 254 nm
30
25
20
15
10
DOC
5
Colour
0
0 2 4 6 8 10
[Al(III)] / g m-3
Fig. 8. Performance of electrocoagulation reactor at pH 6.5 with monopolar electrodes (2 anodes and 2 cathodes).
UV-abs and colour (m-1) and DOC (g C m-3)
40
35 Absorption at 254 nm
30
25
20
15
10 DOC
5
Colour
0
0 2 4 6 8 10 12
[Al(III)] / g m-3
Fig. 9. Coagulation performance of aluminium sulphate at pH 7.8.
dissolution according to Faraday’s Law. Hence, in performance significantly when using electrocoagula-
evaluating the water-treatment performance of the electro- tion, but a lower pH (6.5) improved the coagulation
coagulation reactor, Al(III) doses were calculated from the performance with aluminium sulphate greatly, especially
combined measurements of aluminum concentrations in in removing UV-absorbing substances; this is a well-
the treated effluent and in the floated sludge. established effect due to two factors. Firstly, a lower
Figs. 7–10 show the performance of the up-flow pH favours the existence of polymeric coagulating
electrocoagulation reactor, with either monopolar or species with a higher positive charge and offsets
bipolar electrodes, and coagulation by aluminium the acceleration of hydroxide precipitation caused
sulphate, for treatment of a model-coloured water for by a higher coagulant dose. Secondly, decreasing
pH 7.8 and 6.5. pH did not affect the treatment pH of natural coloured waters leads to a reduction of
J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078 4073
40
30
25
20
15
10
DOC
5
Colour
0
0 2 4 6 8 10 12
[Al(III)] / g m-3
Fig. 10. Coagulation performance of aluminium sulphate at pH 6.5.
Table 4
Comparative performance of aluminium sulphate (AS) and electrocoagulation
Coagulant and reactor electrical connection pH DOC (g C m3) Colour (Vis420-abs) (m1) UV254-abs (m1)
the net molecular charge of the humic and fulvic humic and fulvic acids) by Al(III)/Fe(III) coagulants.
acids, which represent well the major UV-absorbing However, of the two reasons mentioned above, the
substances. These two factors subsequently lead to a first (coagulant species) appears to be the more
greater removal of UV-absorbing substances (mainly important.
4074 J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078
6 DOC
2
Colour
0
0 2 4 6 8 10 12 14 16
[Al(III)] / g m-3
Fig. 11. Effect of Al(III) concentration from monopolar electrodes on lowland surface water properties at pH 7.6.
UV-abs and colour (m-1) and DOC (g C m-3)
12
Absorption at 254 nm
10
6 DOC
2
Colour
0
0 2 4 6 8 10 12 14 16
[Al(III)] / g m-3
Fig. 12. Effect of Al(III) concentration from monopolar electrodes on lowland surface water properties at pH 6.5.
Electrocoagulation performed better than aluminium electrode configurations gave a very similar treatment
sulphate; at a lower Al(III) dose (o5 g m3) there is performance.
only, for electrocoagulation, 1 point at o5 g m3, while
at higher doses, the treatment performance by electro- 3.4. Treatment performance of up-flow electrolytic
coagulation and coagulation with aluminium sulphate reactor for lowland surface water
was similar. Table 4 summarises their overall compara-
tive performance for two specific Al(III) doses, showing Figs. 11–14 show results of the treatment performance
clearly that the performance of electrocoagulation was of River Thames lowland surface water by electrocoa-
superior to conventional coagulation with aluminium gulation and coagulation by aluminium sulphate. As
sulphate in treating the model-coloured water, under the indicated, the pH of lowland surface water was either 7.6
conditions used. Table 4 also indicates that the two or adjusted to 6.5 before the evaluation of the
J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078 4075
12
DOC
6
2
Colour
0
0 2 4 6 8 10 12 14 16
[Al(III)] / g m-3
Fig. 13. Effect of aluminium sulphate concentration on lowland surface water properties at pH 7.6.
12
UV-abs and colour (m-1) and DOC (g C m-3)
Absorption at 254 nm
10
6 DOC
2
Colour
0
0 2 4 6 8 10 12 14 16
[Al(III)] / g m-3
Fig. 14. Effect of aluminium sulphate concentration on lowland surface water properties at pH 6.5.
performance of electrocoagulation. The laboratory 3.5. Comparative treatment performance for model-
study results demonstrated that decreasing water coloured water with two electrolytic reactors of different
pH did affect the treatment performance. For the hydraulic designs
aluminium sulphate (AS) coagulation and electrocoagu-
lation at the same dose of 5 g m3, lowering the water The schematic design of up-flow and horizontal
pH to 6.5 resulted in the removal of 20% and 14% more flow electrocoagulation reactors can be seen in Figs. 15a
colour and UV-abs254, respectively (Table 5). However, and b. The electrolytic reaction time is 57 s for the up-
on increasing Al(III) doses from 5 to 8 g m3, only the flow reactor and 27 s for the horizontal-flow reactor.
UV-absorbance decreased significantly for both the Table 6 compares the performance of the two config-
electrocoagulation and AS coagulation. In addition, urations of electrocoagulation reactors with monopolar
both processes performed similarly under the conditions and bipolar electrode arrangements. For treating
used. model water at a given Al(III) dosage (5 g m3), the
4076 J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078
Table 5
Comparative performance of AS and electrocoagulation for treating lowland surface water at a dose of 5 g m3 as Al(III)
Coagulant and reactor electrical connection pH DOC (g C m3) Colour (Vis420-abs) (m1) UV254-abs (m1)
Table 6
Comparative performance of horizontal- and up-flow electrocoagulators for treating model-coloured water at a dose of 5 g Al(III) m3
Up-flow Monopolar 51 76 64
Bipolar 51 74 60
a
R%=removal percentage.
Aluminum
plates Weirs up-flow reactor performed slightly better than the
horizontal flow reactor, regardless of which electrode
To the floc arrangement was used, and the reason for this can be
Inlet collector
suggested as being one or a combination of the
following:
[16] Drazic DM, Popic JP. Corrosion rates and negative dissolved air flotation: a pilot-scale trial. In: Proceedings of
difference effects for Al and some Al alloys. J Appl Innovations in Conventional and Advanced Water Treat-
Electrochem 1999;29:43–50. ment Processes, Amsterdam, The Netherlands, 2000. p. 54,
[17] Jiang JQ, Graham NJD, Andr!e C, Kelsall GH, Brandon 1–11.
NP, Chipps MJ. Comparative performance of an electro- [18] Bard AJ, Parsons R, Jordan J, editors. Standard potentials
coagulation/flotation system with chemical coagulation/ in aqueous solution. New York: Marcel Dekker, 1985.