Water Research 36 (2002) 4064–4078

Laboratory study of electro-coagulation–flotation

for water treatment
Jia-Qian Jianga,b,*, Nigel Grahama, Cecile Andre! a, Geoff H. Kelsallc,
Nigel Brandonc
a
Department of Civil and Environmental Engineering, Imperial College, London SW7 2BU, UK
b
Department of Civil Engineering, Centre for Environmental Health Engineering, University of Surrey, Guildford,
Surrey GU2 7XH, UK
c
Department of Chemical Engineering and Chemical Technology, Imperial College, London SW7 2BY, UK
Received 5 December 2000; received in revised form 8 January 2002; accepted 11 March 2002

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 m
2. 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

1. Introduction pended particles or precipitates and adsorb dissolved

In the electro-coagulation–flotation process for water
treatment, Al or Fe electrodes are dissolved by electro-
lysis, forming a range of coagulant species and metal
hydroxides, which destabilise and aggregate the sus-
*Corresponding author. Department of Civil Engineering,
Centre for Environmental Health Engineering, University of
Surrey, Guildford, Surrey GU2 7XH, UK.
E-mail address: j.jiang@surrey.ac.uk (J.-Q. Jiang).
contaminants (e.g. dissolved organic matter). As de-
scribed below, metal anode dissolution is accompanied
by hydrogen gas evolution at cathodes, the bubbles
capturing and floating the suspended solids formed and
thus removing contaminants.
Few reports have been published in the scientific
rather than technical literature on electrocoagulation
combined with electroflotation for treating potable
water, industrial effluents and wastewaters; this con-
trasts with an extensive literature on the direct and

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
Nomenclature
A electrode area (m2)
E electrode potential vs. reference electrode (V)
F Faraday constant (charge on 1 mol of elec-
trons) (=96485) (C mol
1)
j current density (A m
2)
I; I t current, total current (A)
km mass transport rate constant (m s
1)
M molar mass (g mol
1)
n charge number of electrode reaction (number
of moles of electrons per mole of electroactive
species) (mol e
(mol of species)
1)
n0 cell numbers (dimensionless)
indirect electrochemical oxidation of pollutants. Electro-
coagulation–flotation has been reported (e.g., [1–3]) to
result in the greater removal of chemical oxygen demand
(COD) and suspended solids from effluents in compar-
ison to their treatment by conventional coagulation.
Treating simulated wastewater containing three types of
surfactants (ammonium laurylether sulphate, trimethy-
lalkyl ammonium chloride and fatty ethoxy amine) by
electrocoagulation for 10 min at a current density of
10 A m
2 was reported to result in 40–60% COD
removal [4]. The electrocoagulation kinetics were
examined using a stirred batch electrocoagulation
system with continuously flowing kaolinite suspensions,
and the flow rates were chosen such that residence
times in each of the cells could be set between 20 and
160 min [5]. The results showed that the electrocoagula-
tion kinetics were of second order with respect to
the electrophoretic movement of the particles towards
the anode. In their evaluation of electrocoagulation
for potable water treatment, Vik et al. [6] demonstrated
that, for an equivalent treated water quality, the
amount of coagulant ion required and the sludge
formed with electrocoagulation was much less than
with conventional coagulation. Electrocoagulation
was also shown to be capable of depleting E. coli
bacteria to concentrations below their detection
limit [7].
The application of low frequency (25–75 Hz) sonic
fields at current densities of 0.1 or 0.4 A m
2 has been
reported to increase the quantity of iron hydroxide, and
hence, the quantity of contaminant removed from
wastewater, in model electro-coagulation–flotation
systems using carbon steel electrodes [8]. Presumably,
the sonic field increases mass transport rates, destroys
the compact layer formed at the electrode surfaces,
activates the electrode surfaces by means of generation
of defects in the crystal lattices of the electrodes,
and hence, increases the quantity of iron hydroxide
produced.
4065
P partial pressure of gas (atm.)
Q charge (C)
R gas constant (=8.31441) (J mol
1 K
1)
seec specific electrical energy consumption (kWh
(kg Al)
1)
t time (S)
T absolute temperature (K)
u aqueous volumetric flow rate (m3 s
1)
U electrochemical cell terminal voltage (V)
v electrolyte velocity (m s
1)
V volume (m3)
F (fractional) current efficiency (1)
Z overpotential ðE
Er Þ (V)
A combined electrocoagulation and electroflotation
process was applied to restaurant wastewater treat-
ment [9]. The effects of operational variables on
performance were examined. The electrical charge
loadings of the electrocoagulation and electroflota-
tion are the most important variables. Initial pH
and conductivity had no significant effects on pollu-
tant removal. However, at pH o4, suspended solids
removal decreased sharply. The proposed electrocoagu-
lation and electroflotation process was shown to
have two significant advantages, namely a shorter
retention time and lower water content in the sludge
produced. The power requirement was only
0.5 kWh m
3 under typical operating conditions. The
amount of dried sludge was 0.20–0.37 kg kg
1 COD
removed.
Whilst a number of studies have focused on electro-
coagulation performance for water and wastewater
treatment, Jiang [10] studied the mechanism of anodic
passivation in electrocoagulation, and suggested that
anodic passivation could be prevented using electrodes
with reversible polarity, and by optimising the design of
the electrolytic reactors, as well as by adding chloride
ions to the influent.
In the previous studies, little attention was paid
in determining the current efficiency for aluminium
dissolution and evaluating the treatment performance
of the reactors with different hydraulic designs. There-
fore, the objectives of the work reported here were as
follows:
(a) to quantify hitherto neglected figures of merit for
electrocoagulation: the current efficiency for Al
dissolution and specific electrical energy consump-
tion of the electrocoagulator with Al electrodes;
(b) to evaluate the water treatment performance for
different electrocoagulation reactor configurations
and to establish the best design to maximise
treatment efficiency; and
4066 J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078

(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 h
1), a flow meter (Platon, 1–
current n0 I to pass across electrode/solution interfaces
16 dm3 h
1), 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

Fig. 2. Monopolar and bipolar electrode connections in the electrocoagulation reactor.

 J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078 4067

1 cm2 Al plate, a platinum plate and a saturated calomel
electrode, as working, counter, and reference electrodes,
respectively. The anolyte and catholyte were separated
by a porous sintered glass disc. The working electrode
potential was controlled by a potentiostat (EG&G
Instruments, 273A); its facility to compensate the
applied potential for any IR potential drop between
the working electrode and the tip of the glass Luggin
probe containing the reference electrode was used, as the
solutions had low ionic conductivities.
In subsequent experiments at constant current density
(20, 30 and 50 A m
2 for the anode and 20 A m
2 for the
cathode), the cell voltage was recorded over electrolysis
times between 3 and 16 h. Samples of catholyte and
anolyte were taken and the dissolved aluminium
concentrations determined by atomic absorption spec-
troscopy (Perkin-Elmer, Zeeman 5100PC). The current
efficiency ðFÞ for the production of dissolved Al(III) by
the passage of electrical charge Q (C) in London tap
water of volume V (m3) was calculated using Faraday’s
law:
3FV ½AlðIIIÞ
F¼ : ð1Þ
Q
The electrical conductivity and pH of the water were
measured as a function of electrolysis time from initial
values of 7  10
2 S m
1 and pH 7.8. The specific
electrical energy consumption was calculated as a
function of applied cell voltage ðUÞ:

1 nFU
seec ðkWh ðkg AlÞ Þ ¼ ; ð2Þ
3:6  103 MAl F
where n is the charge number of the reaction, i.e. moles
of electrons per mole of aluminium dissolved, and was
taken to be 3.
Electrocoagulation was operated under the following
conditions:
* water flow rate of 10
2 m3 h
1;
* two electrode arrangements: bipolar electrodes con-
nected in series via the water and monopolar
electrodes connected in parallel;
* operating electrical current densities of 3–25 A m
2,
which produced Al(III) concentrations of 2.4–
14 g m
3, determined as the sum of the concentra-
tions in the treated effluents and floc sludge; and
* pH values of 6.5 and 7.870.2.
For the purpose of comparison, the coagulation
performance of dosed aluminium sulphate (AS) with
the two raw waters were also evaluated, using the same
pH values of 6.5 and 7.870.2 as for the electrocoagula-
tion tests, and Al(III) concentrations of 2–14 g m
3.
The experimental procedures for the coagulation jar
tests, the analysis of water samples, the preparation of
model-coloured water, and the sources of coagulants,
chemical reagents and humic substances were the same
as that described previously [11].
3. Results and discussion
3.1. Current efficiencies for Al dissolution
As shown in Table 2, apparent current efficiencies,
defined by Eq. (1), for dissolution of the Al anodes were
greater than unity. It was assumed that Al(III) species
were formed directly by a single reaction such as (3) and
is discussed below.
Table 3 lists equivalent data for Al dissolution from
the cathode at both constant current density and

2.2. Water treatment performance

The treatment performance of the electrocoagulator Table 2

was evaluated using a model upland-coloured water Anode current efficiency for the Al dissolution
and a surface water taken from the River Thames (at Anode current Al(III) Apparent anode
Putney Bridge); the principal water quality character- density, j concentration current efficiency,
istics are listed in Table 1. The quality of the treated (A m
2) (g m
3) Fa
waters was measured in terms of dissolved organic
carbon (DOC), ultra-violet absorbance at 254 nm 20 11.7 1.376
30 15.2 1.382
(UV254-abs), colour (represented as visible absor-
50 24.8 1.148
bance at 420 nm) (Vis420-abs), pH and conductivity.

Table 1
Properties of the two raw waters

DOC (g C m
3) UV254-abs (m
1) Colour (Vis420-abs) (m
1) Conductivity (S m
1) pH

Model-coloured water 7.4 34 4.5 150 7.8

Thames river water 5.9 11 0.9 800 7.6
4068 J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078

Table 3
Charge for Al dissolution at the cathode

Cathode Total charge, Q Charge for Al Fractional charge used for Al

(C m
2) dissolution as Al(III) dissolution as Al(III) species
species (C m
2)

Constant current density, j ¼ 20 A m
2
1.152  106
6.41  105 0.56

Constant cathode potential =
1.7 V (SCE)
3.15  106
8.85  105 0.28

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 m
2 for 3 h (potential–pH
diagram calculated from the data of Bard et al. [18] for a dissolved aluminium activity of 10
4 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 m
2 over 3 h, the arrows in the Faraday constant (96 485 C mol
1), MAl the molar
Fig. 3 correspond to the potential–pH trajectories for mass of aluminium (26.982 g mol
1), and u the aqueous
anodes (region 1) and cathodes (region 2), respectively, volumetric flow rate (m3 s
1). 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 m
3 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- > EAlðOHÞ
4 =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 s
1) 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 m
3 [14], Eq. (15) predicts a current density
A
jAl Al-Al3þ þ 3e
; ð3Þ for reaction (5) of order 1 A m
2. Hence, at an applied
current density of 20 A m
2 at an Al anode, 21 A m
2
A
jH 2
2H2 O þ 2e
-H2 þ2OH
; ð4Þ could be used for anodic dissolution and 1 A m
2 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 m
3 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

Electrode potential (SHE)/V

0.8
0.6
0.4
0.2
H+ Al(OH) 3 (am)
0.0
-0.2 H2 Al(OH) 4 -
-0.4 Al3+
-0.6
-0.8
-1.0
AlH 2+
-1.2
-1.4
-1.6
-1.8 AlH 3 (c)
-2.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
pH
Fig. 5. Metastable potential–pH diagram considering amorphous Al(OH)3 as the solid Al(III) phase and hydride species and phases
for a dissolved aluminium activity of 10
3 M and 298 K (data from Bard et al. [18] and Perrault [15]).

both aluminium hydride phases (e.g. AlH3) and solution
species (e.g. AlH+ 2 ). Perrault’s thermodynamic data
were used to compute the potential–pH diagram shown
in Fig. 5, which considers the metastable amorphous
Al(OH)3 phase rather than the more stable corundum
(a-Al2O3) as the former phase was formed hydrolytically
from aqueous solutions. In common with other very
reactive elements, such as Ti, the elemental metal has no
area of stability. However, using the same data set, the
Nernst equation for the reaction
Al þ 3Hþ þe
-AlH2þ þH2
may be calculated for 298 K as
EAl=AlHþ2 ðSHEÞ ðVÞ ¼ 0:003
0:177pH

0:059 logðAlH2þ Þ

0:059 log PH2 :
Hence, at pH 7, the reversible potential for reaction (17)
is ca.
1.24 V vs. SHE, which is less negative than the
observed cathode potentials (Fig. 3) at 20 A m
2; hence,
reaction (17) is a possible source of dissolved aluminium,
though, depending on the concentration achieved and
the pH of an oxygenated solution, it could be followed
by the homogeneous reaction
2AlH2þ þO2 þ4H2 O-2AlðOHÞ3 þ4Hþ :
Regardless of whether Al(III) species are formed
directly by reaction (3) or by reaction (17), the
concurrent reduction of oxygen and, at the cathode,
hydrogen evolution, have the net effect of increasing
the local pH, thus enhancing the solubility of Al(OH)

4
ions:
log½AlðOHÞ

4 DpH
13:49 ð22Þ
while the bubbles from hydrogen evolution enhance the
dispersion of Al(III) species from electrode surfaces.
This explains why anode passivation does not restrict Al
ð17Þ dissolution rates more severely at neutral bulk solution
pHs.
Anomalous pitting corrosion behaviour of aluminium
has been reported [16], and the rate of hydrogen
evolution increases with the increase in electrode
ð18Þ potentials over a wide pH range. However, those effects,
again leading to anomalous Faradaic yields, appear in
the pitting potential range of ca.
0.5 V (SHE), which
was not adopted by either cathode or anode in the
driven electrochemical reactor used in the work reported
here.
3.2. Specific electrical energy consumption (seec) for
monopolar and bipolar connections—different electrode
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 m
2, it was only 20 kWh (kg Al)
1
 J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078 4071

Specific Electrical Energy Consumption / kWh (kg Al)-1

120

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 m
2, 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

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
[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

UV-abs and colour (m-1) and DOC (g C m-3)

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. 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 m
3) Colour (Vis420-abs) (m
1) UV254-abs (m
1)

Conc. R%a m
1 R% m
1 R%

3
(a) Dose of 3.6 g m as Al(III)
AS 7.8 5.2 29.7 1.8 60 22 36
Monopolar 3 59.4 0.5 89 12 64.7
Bipolar 2.8 62.1 0.5 89 11 67.6

AS 6.5 3.9 47.3 1 78 13 62

Monopolar 2.4 67.5 0.5 89 10 71
Bipolar 2.4 67.5 0.5 89 9 74

Model-coloured waterb 7.8 7.4 — 4.5 34 —

(b) Dose of 4.3 g m
3 as Al(III)

AS 7.8 5 32.4 1.5 67 20 41
Monopolar 2.5 66.2 0.5 89 10 71
Bipolar 2.5 66.2 0.5 89 10 71

AS 6.5 3.9 47.3 1 78 10 71

Monopolar 2.4 67.5 0.5 89 9 74
Bipolar in series 2.4 67.5 0.5 89 9 74

Model-coloured waterb 7.8 7.4 — 4.5 34 —

a
R%=removal percentage.
b
Conductivity=1.5  10
2 S m
1.

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

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. 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 m
3) there is performance.
only, for electrocoagulation, 1 point at o5 g m
3, 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

UV-abs and colour (m-1) and DOC (g C m-3)

Absorption at 254 nm
10

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 m
3, 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 m
3, 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 m
3), 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 m
3 as Al(III)

Coagulant and reactor electrical connection pH DOC (g C m
3) Colour (Vis420-abs) (m
1) UV254-abs (m
1)

Conc. R%a m
1 R% m
1 R%

AS 7.6 5 15.3 0.7 18.6 8.8 16.2

Monopolar 7.6 4.6 22 0.67 22.1 9 14.3

AS 6.5 4.5 23.7 0.5 41.8 7.4 29.5

Monopolar 6.5 4.3 27.1 0.5 41.8 7.5 28.6

Lowland surface waterb 7.6 5.9 — 0.86 — 10.5 —

a
R%=removal percentage.
b
Conductivity=8  10
2 S m
1.

Table 6
Comparative performance of horizontal- and up-flow electrocoagulators for treating model-coloured water at a dose of 5 g Al(III) m
3

Hydraulic flow and electrode connection DOC Colour (Vis420-abs) UV254-abs

configuration in electrocoagulators (R%)a (R%) (R%)

Horizontal flow Monopolar 45 71 60

Bipolar 43 70 57

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:

Baffles (a) Aluminium ions were apparently not generated

Treated uniformly, and therefore, the different mixing
Raw water water characteristics in the two designs resulted in
different pH profiles between anodes and cathodes.
(a)
Enhanced cross-gap mixing would result in
Baffles more uniform spatial distributions of pH and
Weirs
this causes relatively low pH values at the catho-
Electrodes To the floc
collector des, which in turn would lead to less flocs being
formed.
(b) A lower flow residence time through the electrolytic
unit of horizontal flow reactor results in a higher
current density to be applied in order to achieve the
same Al dosage, compared with the up-flow
Inlet
Raw water Treated reactor. This subsequently causes relatively large
water (>100 mm) gas bubbles to be generated from the
cathodes and leads to less efficient flotation
(b) performance.
Fig. 15. (a) Up-flow electrocoagulation reactor and (b) hor- (c) The up-flow reactor arrangement allows more
izontal-flow electrocoagulation reactor. collisions to occur between the gas bubbles and
 J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078
the resulting Al-contaminant flocs (since there is a
longer distance to the surface), which enhance the
overall flotation performance.
Based on the above observations and discussion, the
up-flow electrocoagulation reactor was chosen for
further evaluation in pilot-scale experiments, whose
results can be seen elsewhere [17].
4. Conclusions
A laboratory electrocoagulation process system with a
treatment capacity of 10
2 m3 h
1 was designed, built
and its performance characterised.
1. For a current density of 20 A m
2, apparent current
efficiencies for Al anode dissolution were greater than
unity for a three-electron oxidation process produ-
cing Al(III) species. This was due to the applied
currents being the sum of partial currents due to
anodic oxidation and cathodic reduction of dissolved
oxygen.
2. Al cathodes also contributed to aluminium dissolu-
tion, again due to the applied current densities being
the sum of partial current densities due to the
reduction of water and dissolved oxygen and
oxidative dissolution of Al. Hence, these effects at
anodes and cathodes resulted in the total Al(III)
dissolution in the electrocoagulation process being
significantly greater than that predicted from Far-
aday’s Law for a single three-electron oxidation
process producing Al(III) species.
3. The specific electrical energy consumption (seec) of
the electrocoagulation system was about 20–
80 kWh (kg Al)
1 for current densities of 10–
60 A m
2, essentially independent of whether mono-
polar or bipolar electrodes were used.
4. Electrocoagulation performed better than conven-
tional coagulation with aluminium sulphate for
treating a model-coloured water, 20% more DOC
was removed for the same Al(III) dose.
5. For the lowland surface water sample, the two
processes achieved a similar performance for
DOC and UV-absorbance removal. Lowering the
water pH to 6.5 resulted in the removal of 20% and
14% more colour and UV-abs254, respectively, for
coagulation with aluminium sulphate and electro-
coagulation.
6. The up-flow reactor arrangement achieved an
average of 5–8% more DOC, colour and UV254
removal than the horizontal flow reactor, with
both bipolar and monopolar electrodes. In view
of this, an up-flow reactor was chosen for fur-
ther evaluation in the subsequent pilot-scale experi-
ments.
4077
Acknowledgements
The authors gratefully acknowledge the funding of
this study by the Engineering and Physical Sciences
Research Council. They also thank Dr. Q. Yin,
Imperial College, for his assistance with the mea-
surement of current efficiencies for Al dissolution,
and N. Paige of the Royal Holloway and Bedford
New College, University of London, for her assis-
tance in the measurement of dissolved Al(III) concen-
trations.
References
[1] Marson HW. Electrolytic sewage treatment. Engineer
1965;4:591–6.
[2] Pouet MF, Persin F, Rumean M. Intensive treatment by
electrocoagulation–flotation–tangential flow microfiltra-
tion in areas of high seasonal population. Water Sci
Technol 1992;25:247–53.
[3] Pouet MF, Grasmick A. Urban waste-water treatment by
electrocoagulation and flotation. Water Sci Technol
1995;31(3–4):275–83.
[4] Ciorba GA, Radovan C, Vlaicu I, Pitulice L. Correlation
between organic component and electrode material:
consequences on removal of surfactants from wastewater.
Electrochim Acta 2000;46(2–3):297–303.
[5] Matteson MJ, Dobson RL, Glenn RW, Kukunoor NS,
Waits WH, Clayfield EJ. Electrocoagulation and separa-
tion of aqueous suspensions of ultrafine particles. Colloids
Surf A—Physicochem Eng Aspects 1995;104:101–9.
[6] Vik EA, Carlson DA, Eikum AS, Gjessing ET. Electro-
coagulation of potable water. Water Res 1984;18:1355–60.
[7] Mills D. A new process for electrocoagulation. J Am
Water Works Assoc 2000;92:34–43.
[8] Kovatcheva VK, Parlapanski MD. Sono-electrocoagula-
tion of iron hydroxides. Colloids Surf A—Physicochem
Eng Aspects 1999;149(1–3): 603–8.
[9] Chen GH, Chen XM, Yue PL. Electrocoagulation and
electroflotation of restaurant wastewater. J Environ Eng–
ASCE 2000;126(9):858–63.
[10] Jiang JQ. Study on the anodic passivation of the
electrocoagulation in water treatment process. Water
Treatment 1986;3:344–52.
[11] Jiang JQ, Graham NJD. Enhanced coagulation using Al/
Fe(III) coagulants: effect of coagulant chemistry on the
removal of colour-causing NOM. Environ Technol
1996;17:937–50.
[12] Baes Jr CF, Mesmer RE. The hydrolysis of cations. New
York: Wiley, 1976.
[13] Gijsbers HFM, Janssen LJJ. Distribution of mass-transfer
over a 0.5-m-tall hydrogen-evolving electrode. J Appl
Electrochem 1989;19:637–48.
[14] Tromans D. Oxygen solubility modeling in inorganic
solutions: concentration, temperature and pressure effects.
Hydrometallurgy 1998;50:279–96.
[15] Perrault GC. The role of hydrides in the equilibrium of
aluminium in aqueous solutions. J Electrochem Soc
1979;126:808.
4078 J.-Q. Jiang et al. / Water Research 36 (2002) 4064–4078

[16] Drazic DM, Popic JP. Corrosion rates and negative
difference effects for Al and some Al alloys. J Appl
Electrochem 1999;29:43–50.
[17] Jiang JQ, Graham NJD, Andr!e C, Kelsall GH, Brandon
NP, Chipps MJ. Comparative performance of an electro-
coagulation/flotation system with chemical coagulation/
dissolved air flotation: a pilot-scale trial. In: Proceedings of
Innovations in Conventional and Advanced Water Treat-
ment Processes, Amsterdam, The Netherlands, 2000. p. 54,
1–11.
[18] Bard AJ, Parsons R, Jordan J, editors. Standard potentials
in aqueous solution. New York: Marcel Dekker, 1985.