A Quantitative Comparison Between Chemical Dosing and Electrocoagulation

A quantitative comparison between chemical dosing and
electrocoagulation
Peter K. Holt
a
, Geoffrey W. Barton
a,
*, Mary Wark
a
, Cynthia A. Mitchell
b
a
Department of Chemical Engineering, University of Sydney, Sydney, NSW 2006, Australia
b
Institute for Sustainable Futures, University of Technology Sydney, PO Box 123 Broadway, Sydney NSW 2007, Australia
Received 9 November 2001; accepted 12 June 2002
Abstract
A renewed interest in electrocoagulation has been spurred by the search for reliable, cost-effective water treatment
processes. This technology delivers the coagulant in situ as the sacrifcial anode corrodes, due to an applied potential,
while the simultaneous evolution of hydrogen at the cathode allows for pollutant removal by flotation. By comparison,
conventional chemical dosing typically adds a salt of the coagulant, with settling providing the primary pollutant
removal path. This paper provides a quantitative comparison of these two approaches based on turbidity removal
associated with a clay pollutant. Chemical coagulation was evaluated via jar tests using aluminium sulphate (alum).
This proved more effective than electrocoagulation under acidic conditions (pH /4) and low coagulant levels
(4 mg-Al l
1
being the minimum able to effectively destabilise the colloidal clay particles). Highly effective coagulation
was observed at intermediate alum dosage levels (4/20 mg-Al l
1
), where the isoelectric point occurred at pH /7.8.
Three operating stages (lag, reactive and stable) were identified in a batch electrocoagulation reactor with the operating
current determining the pollutant removal rate. At the isoelectric point, which occurs during the reactive stage, the
greatest turbidity reduction occurs, indicating aggregation by a sorption mechanism (compared to the charge
neutralisation as in the case of chemical coagulation). During the stable stage, continued precipitation of aluminium
hydroxide and a decrease in turbidity indicated a sweep coagulation mechanism. The highest current (2 A) reduced the
pollutant level in the shortest time, 1% residual turbidity after 30 min, though the highest efficiency (in terms of
pollutant removed per unit of aluminium added) was achieved at the lowest current (0.25 A).
# 2002 Elsevier Science B.V. All rights reserved.
Keywords: Electrocoagulation; Alum dosing; Water treatment; Electrochemical; Aluminium electrodes
1. Introduction
1.1. Historical perspective
Electrocoagulation is currently experiencing a
renaissance. At the turn of the nineteenth century,
it was seen as a promising technology, in fact, a
* Corresponding author. Tel.: /61-29351-3780; fax: /61-
29351-2854
E-mail address: [email protected] (G.W.
Barton).
Colloids and Surfaces A: Physicochem. Eng. Aspects 211 (2002) 233/248
www.elsevier.com/locate/colsurfa
0927-7757/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 7 - 7 7 5 7 ( 0 2 ) 0 0 2 8 5 - 6
water treatment plant was successfully commis-
sioned in London at this time [1]. In the following
decades, plants were also commissioned in the
United States to treat municipal wastewater. By
the 1930s, however, all such plants had been
abandoned due to perceived higher operating costs
[2] and the ready availability of mass-produced
alternatives for chemical coagulant dosing.
In recent years, however, smaller scale electro-
coagulation processes have found a niche in the
water treatment industry, proving to be reliable
and effective technologies [1/7], though requiring
greater technical understanding for their potential
to be fully exploited. Only recently has research
aimed at a quantitative understanding of electro-
coagulations relatively complex pollutant removal
mechanisms [3].
1.1.1. Chemical coagulation
The stability of a pollutant is determined by its
physicochemical properties. Many pollutants are
composed of similarly charged particles that repel
each other, with the repulsive forces creating a
stable, colloidal system. In order to maintain
electroneutrality, oppositely charged ions, typi-
cally hydroxyl (OH
) or hydrogen ions (H
),
are attracted to the (charged) pollutant particles.
The attraction of counter ions to a negatively
charged pollutant forms an electric double layer
divided into a Stern and diffuse layer [8/11]. It is
difficult to measure the charge at the colloid
surface, due to the charge concentrations in the
Stern and diffuse layers [8]. Consequently, the zeta
potential is used as an experimental measure of the
effective charge of the particle as it moves
through the solution, thus providing a direct
indicator of solution stability [11].
Electrostatic repulsion between electric double
layers drives particles apart, whilst van der Waals
forces act to bring them together. The interaction
between these forces of attraction and repulsion
can be described by the Deryaguin and Landau,
Verwey and Overbeek (DLVO) theory [12,13]. The
energetics are such that attraction dominates at
small separations. However, to arrive at small
separations, a repulsive energy barrier must first
be overcome.
The coagulants role here is to destabilise the
colloidal suspension by reducing the attractive
forces, thereby lowering the energy barrier and
enabling particles to aggregate. A number of
coagulation mechanisms including charge neutra-
lisation, double layer compression, bridging and
sweep, have been postulated dependent on the
physical and chemical properties of the solution,
pollutant and coagulant [9,11,14].
1.1.2. Coagulation by metal cations
Metal salt coagulants are commonly used in
water treatment and in this study aluminium
sulphate (Al
2
(SO
4
)
3
/ 18H
2
O) is used as the chemi-
cal coagulant. Aluminium is also frequently used
as the sacrificial anode in an electrocoagulation
reactor. In both cases, chemical dosing and
electrocoagulation, the metal cation hydrolyses
upon addition to solution. Mononuclear com-
plexes are initially formed as described by Eqs.
(1)/(4).
Al
3
H
2
O 0AlOH
2
H
(1)
AlOH
2
H
2
O 0Al(OH)
2
H
(2)
Al(OH)
2
H
2
O 0Al(OH)
0
3
H
(3)
Al(OH)
0
3
H
2
O 0Al(OH)
-
4
H
(4)
The extent of hydrolysis depends upon total
metal concentration and pH, as well as the amount
of other species present in solution. Fig. 1 is the
solubility diagram for aluminium hydroxide,
Al(OH)
3
(s), assuming only mononuclear species
[data by Spositio, 15]. The solubility boundary
denotes the thermodynamic equilibrium that exists
between the dominant aluminium species at a
given pH and solid aluminium hydroxide. The
minimum solubility, 0.03 mg-Al l
1
, occurs at pH
6.3, with solubility increasing as the solution
becomes more acidic or alkaline [11,16,17].
However, as the aluminium concentration in-
creases and/or the solution ages, polynuclear
aluminium complexes are formed and aluminium
hydroxide precipitates, as shown below.
Al
3
0Al(OH)
(3n)
n
0Al
2
(OH)
4
2
0Al
13
complex 0Al(OH)
3
(5)
The speciation of aluminium systems has been
well documented in the literature [15,18], although
P.K. Holt et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 211 (2002) 233/248 234
little information is available on the kinetics that
determines the rate at which thermodynamic
equilibrium is attained. The complexity of the
aqua-aluminium system is neither completely un-
derstood nor completely quantified [16,17]. This
paper will assume mononuclear hydrolysed species
adequately predict aluminium hydroxide precipi-
tation.
1.2. Electrocoagulation
Electrocoagulation is a complex process, with a
multitude of mechanisms operating synergistically
to remove pollutants from the water. A wide
variety of opinions exist in the literature for both
the key mechanisms and the best reactor config-
urations. Design variations include a fluidised-bed
reactor employing aluminium pellets [4], bipolar
aluminium electrodes [5], mesh electrodes [1], as
well as simple plate electrodes [2,6]. There is
certainly no dominant electrocoagulation reactor
in use. Reported operating conditions and perfor-
mance mirror the wide variation in design, with
reactors invariably being tuned to best suit a
specific application. These empirical approaches
invariably prove the viability of the technology,
but singularly fails to fully capitalise on its
potential. This is due to a lack of fundamental
understanding of the system and hence the in-
ability to accurately predict performance.
Fig. 2 begins to show the complex, interdepen-
dent nature of the electrocoagulation process [3,7].
A sacrificial metal anode (usually aluminium, but
sometimes iron) is used to dose polluted water
with a coagulating agent [1/7]. Simultaneously,
electrolytic gases (mainly hydrogen at the cathode)
are generated.
It is possible to identify three separate categories
of mechanistic processes */ electrochemistry,
coagulation, and hydrodynamics */ that form
the basis of electrocoagulation. The fact that these
processes are difficult to investigate separately in
an operational reactor goes some way towards
explaining the lack of a detailed technical literature
on electrocoagulation.
1.3. Present study
Chemical dosing delivers the coagulant as a salt
that dissociates in solution with hydrolysis of the
aluminium cation (and associated anions) deter-
mining solution speciation and pH. Alum addi-
tion, for example, acidifies the water. By contrast,
aluminium added via electrocoagulation does not
bring with it any associated salt anions, with the
result that the pH typically stabilises in the alka-
Fig. 1. Solubility diagram of aluminium hydroxide Al(OH)
3
(s) considering only mononuclear aluminium species.
line range. The fact that electrolytic gases are also
being produced has also been seen as an unneces-
sary complication [5]. Alternatively, these gases
can be seen as being beneficial to the separation
process. The coagulants delivery and its nature
influence the coagulation and separation processes
by its speciation, removal path and associated by-
products (salt anions for chemical coagulation and
electrolytic gases for electrocoagulation).
Although electrocoagulation is seen as a pro-
mising technology, there is a perception that it is
simply chemical dosing by another name, with the
only difference being the way in which the
coagulant is delivered. The objective in this paper
is to quantitatively compare chemical dosing and
electrocoagulation to determine dominant coagu-
lation mechanisms and separation processes. The
two approaches were evaluated by monitoring key
solution characteristics (such as coagulant level,
pH and zeta potential) in a series of chemical
dosing jar tests and electrocoagulation runs car-
ried out in a 7 l batch reactor.
2. Experimental method
2.1. Chemical coagulation experiments
Chemical coagulation was evaluated using an
adaptation of the standard jar testing technique
[19], with AR grade (99.8% pure) aluminium
sulphate as the chemical coagulant. Sodium hy-
droxide and hydrochloric acid solutions were
added for any subsequent pH adjustment. The
pollutant used was a potters clay, comprising
kaolinite (67%), quartz (25%), illite/mica (3%),
feldspar (3%) and other trace elements (2%), as
characterised by X-ray diffraction analysis.
A concentrated aluminium sulphate solution
(1 mg-Al ml
1
) was added to a clay mixture
Fig. 2. Interactions occurring within an electrocoagulation reactor.
(at 1 g l
1
), the mixture rapidly mixed (at more
than 140 rpm) for a period of 30 s, with the
solution then left to settle [20]. No further mixing
occurred. In the runs carried out, the initial clay
concentration was always at 1 g l
1
, while
aluminium sulphate and pH were varied. Once
rapid mixing had ceased (this being defined as time
zero), samples were taken at 0, 1, 3, 5, 10, 15, 20,
25, 30, 45 and 60 min. Turbidity, zeta potential
and pH were then measured off-line using a Merck
Turbiquant 1500T (tungsten lamp), Malvern Ze-
tasizer and a calibrated pH meter, respectively. All
experiments were conducted at ambient tempera-
ture (nominally 20 8C).
2.2. Electrocoagulation experiments
The batch electrocoagulation reactor used (see
Fig. 3) is constructed of perspex and has a
maximum capacity of 7.1 l. The pollutant used
here was the same batch of potters clay described
above. Fig. 4 is a schematic diagram of the
electrode arrangement; five stainless steel cathodes
are interspersed with four aluminium anodes, with
brass rods used to connect the parallel plate
electrodes.
Electrical conductivity, total dissolved solids
(TDS), salinity and temperature were all measured
using a WTW LF340 probe, while pH was
measured in a flow-through cell drawing solution
from the reactor at a rate of 0.30 l min
1
using a
peristaltic pump. The current flowing through the
cell and the voltage across the electrodes were both
recorded using a data logger. The current was held
constant for each run. Current was investigated
over the range 0.25/2.0 A. Aluminium concentra-
tion was measured using a Varian Atomic Absorp-
tion Spectrophotometer. Again, turbidity and zeta
potential were measured off-line using a Merck
Turbiquant 1500T (tungsten lamp) and Malvern
Zetasizer.
Fig. 3. Experimental electrocoagulation reactor.
Fig. 4. Schematic diagram of electrode arrangement.
All electrocoagulation experiments were con-
ducted under standard conditions, that is, initial
clay pollutant loading was always 1.0 g l
1
, in 6 l
of de-ionised water, with 200 mg l
1
sodium
chloride to enhance conductivity.
3. Results and discussion
Three broad aluminium concentration ranges
*/ low (less than 4 mg-Al l
1
), medium
(4/40 mg-Al l
1
) and high (greater than 40 mg-
Al l
1
) */ were used as a basis for comparing the
effectiveness of chemical dosing and electrocoagu-
lation at equivalent aluminium concentrations.
3.1. Clay removal
3.1.1. During chemical dosing
Normalised turbidity values as a function of
settling time, that is, since the cessation of rapid
mixing, for the dosage range 0/200 mg-Al l
1
studied are presented in Fig. 5. Intermediate
aluminium concentrations (4/40 mg-Al l
1
) pro-
duced effective and very similar turbidity reduc-
tion profiles. Overdosing (i.e. 200 mg-Al l
1
) or
underdosing (B/4 mg-Al l
1
) produced less effec-
tive turbidity removal. Poor coagulation perfor-
mance at low alum concentrations (0 and 2 mg-Al
l
1
) suggests insufficient coagulant for pollutant
destabilisastion. At the highest alum dosage tested
(200 mg-Al l
1
), turbidity removal was less
effective than at lower dosages, most likely in-
dicating some measure of colloidal restabilisation
was occurring [17].
The pH for these runs was monitored but not
controlled. In these jar tests; the initial pH was 5.5
with the pH decreasing rapidly upon alum addi-
tion. At an alum dosage level of 4 mg-Al l
1
, the
final pH was 4.4, while at 200 mg-Al l
1
the pH
was 3.6. Alum dosing acidifies the solution, with
the pH dropping as the dosage level increases.
3.1.2. During electrocoagulation experiments
Fig. 6 is a semi-log plot showing the normalised
turbidity and pH profiles for a typical electro-
coagulation run, where the initial clay concentra-
tion was 1.0 g l
1
, and the operating current was
kept constant at 0.5 A. Three stages can clearly be
identified-namely, a lag, a reactive and a stabilis-
ing stage. Little or no turbidity change is observed
in the lag, with the majority (/95%) of turbidity
removal occurring during the reactive stage. As
time progresses, the rate of turbidity reduction
decreases, with the turbidity eventually levelling
out above zero. The pH stabilised to 8.5 after 10
Fig. 5. Normalised turbidity for 1 g l
1
clay solutions as a function of time and chemical dosing level.
min of applied current. Note that while the system
pH was monitored, it was not controlled, thus, the
stabilisation of the pH is a characteristic of the
electrocoagulation reactor.
For this clay pollutant, batch electrocoagulation
approaches a stable alkaline pH, which might be
taken as implying stable speciation within the
solution. Yet the process is inherently dynamic,
with aluminium entering the system, being hydro-
lysed in a complex manner and being removed
from the system. Simultaneously water is being
reduced adding hydroxyl ions into solution. Such
dynamic changes in concentration affect both the
solution speciation and the equilibrium state
towards which the system is moving at any one
time. Although jar tests (using alum) may also be
seen as a time-varying process, the dynamics (to a
first approximation) might be regarded as less
severe, with coagulant addition occurring just
once, and thus the system moving towards just
one equilibrium state.
The important point here is that in a jar test the
coagulant addition is a discrete event, with the
system behaving as a (shot-fed) batch reactor
moving towards a defined final equilibrium state.
The (non-continuous flow) electrocoagulation re-
actor behaves more like a fed-batch system with
coagulant being added continuously, at a rate
dependent on electrode/solution properties and
the applied current. As a result, the equilibrium
towards which such a reactor is moving, is itself
constantly shifting. These differences in dynamic
behaviour contribute significantly to the distinc-
tion between chemical dosing and electrocoagula-
tion.
3.2. Zeta Potential measurements
As previously discussed, colloids are maintained
in suspension by electrostatic repulsion between
particles. The zeta potential provides an effective
measurement of the charge on a particle. Addition
of aluminium coagulant can suppress the electric
double layer around colloidal particles, thus en-
couraging aggregation of the pollutant [11].
3.2.1. Zeta potential for chemical dosing
Fig. 7 shows the zeta potential of clay at low
alum concentrations (0, 2 and 4 mg-Al l
1
) for pH
adjusted from 3/11. The colloidal clay system
(with no alum addition) is stable, with the particles
being negatively charged. Alum addition (to 2 and
4 mg-Al l
1
) increased the zeta potential, although
it remained below zero. At these concentrations,
there is insufficient coagulant to effectively reduce
the pollutants electric double layer to enable
Fig. 6. Normalised turbidity and pH for 1 g l
1
clay solution subjected to electrocoagulation at a constant current of 0.5 A.
destabilisation. The ineffective destabilisation is
reflected by the sub-optimal turbidity removal at 0
and 2 mg-Al l
1
alum (Fig. 5). At an alum
concentration of 4 mg-Al l
1
, particles ap-
proached the isoelectric point and coagulation
performance improved markedly (also see Fig.
7). Generally, the zeta potential decreased with
increasing pH. This is probably due to an increase
in negatively charged ions, specifically the hydrox-
yl ion added to increase the pH, and the aluminate
ion, Al(OH)
4
(aq), which is dominant above pH
/9 [11,17].
At intermediate alum concentrations (10 and 20
mg-Al l
1
), the zeta potential increases from
/3.5 mV at low pH values, to a maximum of /
14.5 mV at pH /5.1, before decreasing as the
solution becomes more alkaline. Fig. 8 shows that
the suspension passes through two isoelectric
points at approximate pH values of 4.0 and 7.8.
Effective turbidity removal is observed at pH 4.1/
4.2 which corresponds to the isoelectric point (see
Fig. 5). At this pH, the highly charged aluminiu-
m(III) cation, Al
3
(aq), is the dominant species
[11,15] that binds to the negatively charged clay
pollutant. Thus the solution is destabilised by a
charge neutralisation mechanism at alum concen-
trations of 10 and 20 mg-Al l
1
.
3.2.2. Zeta potential for electrocoagulation
Fig. 9 presents the zeta potential measurements
for the electrocoagulation reactor as a function of
time for an applied current of 0.5 A, along with
normalised turbidity measurements relative to the
initial value. Initially, there is no change observed
in the zeta potential */ a response that mirrors the
lag stage in the reactor. As the amount of
coagulant dissolved at the anode increases, a point
is reached where there is a sudden decrease in
turbidity and the zeta potential increases rapidly,
passing through the isoelectric point. The pH
stabilises to 8.5 before the reactive stage and its
profile is presented in Fig. 6. As the run pro-
gresses, turbidity, zeta potential and pH stabilise.
The beginning of the reactive stage may be
defined as the point where a rapid decrease in
turbidity commences. Fig. 9 shows that the iso-
electric point occurs during this reactive stage.
This isoelectric point occurs at pH 8.5 which is
comparable to results obtained from chemical
coagulation with the occurrence of an isoelectric
point at pH /8 (see Fig. 8). Assuming mono-
nuclear aluminium speciation, the aluminate ion,
Al(OH)
4
(aq), is the dominant specie at pH 8.5.
The negatively charged clay could not be destabi-
lised by a negative aluminium ion hence coagula-
tion by charge neutralisation appears implausible.
Nevertheless, the aluminium solution chemistry
model has assumed the presence of mononuclear
species only. Extension of this model (using Out-
okumpu HSC Chemistrys thermodynamic pack-
age) to include polynuclear species, specifically the
aluminium tri-meric specie, Al
3
(OH)
5
4
(aq), shifts
the pH from 6.6 to 10.7. This pH range spans the
region of interest thus suggesting that this highly
charged positive ion might be responsible for
destabilisation and coagulation. Regardless, the
overlap, that is the correspondence of the iso-
Fig. 7. Zeta potential for 1 g l
1
clay solutions subjected to
chemical dosing at 0, 2 and 4 mg-Al l
1
of alum.
1
clay solution subjected to
chemical dosing at 10 and 20 mg-Al l
1
of alum.
electric point and the rapid turbidity reduction,
together with the nature of the aluminium coagu-
lant and clay pollutant, strongly suggests that
some type of sorption mechanism is the primary
coagulation mechanism here [11,17].
As the system moves into the stable stage, the
rate of turbidity reduction continues to decrease,
while the zeta potential increases above the iso-
electric point before stabilising (see Fig. 9). In-
crease in zeta potential indicates re-stabilisation of
the solution, thus a sorption coagulation mechan-
ism is not possible during the stable stage. At pH
range of interest, 8.3/8.8, thermodynamics
predicts precipitation of aluminium hydroxide
when total aluminium concentration exceeds
2.1/6.7 mg-Al l
1
, respectively (Fig. 1). This
data, together with the fact that aluminium con-
tinues to be added to solution and hydrolysed with
aluminium hydroxide precipitating, would seem to
indicate a gradual shift to sweep flocculation (or
enmeshment) as the dominant pollutant removal
mechanism in the stable stage. The residual
turbidity levels were explored by operating the
electrocoagulation reactor with and without clay
present.
Fig. 10 presents the normalised turbidity within
the electrocoagulation reactor, both with pollutant
initial clay loading of 1 g l
1
and without clay.
The interesting point to note is the increase during
the reactive stage in turbidity level over time when
no clay is present. In this scenario the only
solution changes are due to the dissolution of the
aluminium anode and the cathodic reduction of
water so this turbidity increase is best explained by
the production of a precipitate. During the stabi-
lising stage, turbidity levels converged. This would
seem to suggest that the residual turbidity ob-
served in electrocoagulation runs with clay is due
to the electrocoagulation process and not the clay.
That is, the turbidity rise in the no pollutant case
representing the establishment of a steady-state
between the production and flotation of hydro-
lysed aluminium hydroxide flocs. This hypothesis
is further explored by measurements of total
aluminium both in solution and as a precipitate,
with and without clay. These results are shown in
Fig. 11. Faradays law is used to relate the current
flow (I for time t) to the amounts (m) of
aluminium and hydroxide ions generated within
the reactor, and hence provides a theoretical
amount for total aluminium.
m
I t M
ZF
(6)
Z is the number of electrons transferred in the
reaction at the electrode, M is the molecular
1
weight (g mol
1
) and F is Faradays constant
(96 486 C mol
1
).
Measured total aluminium concentration for
runs with and without clay remained below that
predicted by Faradays law (as shown in Fig. 11).
When clay is present, the total aluminium con-
centration was consistently below 5 mg l
1
. The
bulk of the coagulant must be bound to the clay
pollutant and hence removed with the pollutant.
This observation proves a binding interaction
exists between the coagulant and the pollutant
supporting a sorption mechanism.
Without clay, the total aluminium concentration
increases to 30 mg l
1
before levelling out when
the rates of formation and removal of the hydro-
lysed aluminium precipitate converge. The clay
provides binding sites and is removed with the
coagulant, yet when no clay is present, and hence
Fig. 10. Normalised turbidity for 1 and 0 g l
1
Fig. 11. Total Aluminium concentration for 1 and 0 g l
1
the absence of binding sites, the added aluminium
forms a colloidal suspension of microcrystalline
aluminium hydroxide flocs [11]. This contributes
to a residual turbidity level, observed during the
stable phase and the difference in measured
aluminium concentration. Absence of binding
sites, that is, the clay pollutant, results in the
different behaviour of the electrocoagulation sys-
tem. Thus the convergence of turbidity levels for
the two scenarios considered (with and without
clay) is coincidental with the aluminium and clay
properties dictating the residual turbidity levels.
3.3. Electrocoagulation */ change in current
After the colloidal suspension has been destabi-
lised, effective aggregation requires adequate con-
tact between the coagulant and pollutant particles.
Consequently, the transport mechanism is impor-
tant as this determines the transportation of,
including collisions between, particles coagulant
and bubbles. Thus the fluid regime clearly has a
significant impact on both the flocculation me-
chanism and the subsequent means of pollutant
removal. Hence Table 1 describes flocculation
mechanisms and associated transport properties
[11].
The current not only determines the coagulant
dosage rate but also the bubble production rate
and fluid regime (mixing) within the reactor.
Hence the collision between particles, floc growth
and the potential for material removal, both
pollutant and coagulant, by flotation is deter-
mined by the current.
Fig. 12 shows pollutant removal behaviour
when current was varied form 0.25/2 A. For all
cases, however, regardless of initial value, the pH
stabilises within 15 min to somewhere in the range
8.3/8.8. The highest current (2 A) produced the
shortest lag (10 min) and the quickest response,
with a 90% turbidity reduction occurring between
10 and 15 min. As the current decreased, the length
of each electrocoagulation stage increased.
The results of mass balances performed on the
electrocoagulation reactor are presented in Table
2. Pollutant loading was constant for all experi-
ments with current being altered. At 0.25 A, 26%
of the total pollutant was recovered at the surface
after 4 h of operation with the remaining mass
recovered as sludge at the reactor base. Conver-
sely, the highest current, 2 A, produced the largest
pollutant recovery at the surface (63% of the
total). Low current produces low (hydrogen)
bubble density, leading to a low upward momen-
tum flux, and thus poor mixing within the reactor-
conditions that encourage sedimentation over
flotation. As the current is increased, the bubble
density and the amount of mixing increase */
favouring flotation over sedimentation [7]. Hence,
the operational current has a strong influence on
the dominant pollutant removal path, that is,
flotation (surface) or settling (base) and conse-
quently sludge production.
As noted, the amount of aluminium entering
solution can be calculated as a function of time
using Faradays law (Eq. (6)). For example,
operation at 1 A for an hour corresponds to a
total aluminium dosage of 40 mg-Al l
1
for a
reactor volume of 6 l. Fig. 13 presents the same
electrocoagulation performance data as in Fig. 12,
but presented on an equivalent aluminium basis,
that is, on the basis of the total amount of
aluminium electrochemically added to solution
up to a given point in time.
Reactor behaviour is shown to be a strong
function of aluminium concentration. However,
the order of turbidity reduction has reversed
relative to the time-based data. The lower the
current, the less aluminium is used to achieve an
equivalent turbidity reduction. In addition, the
characteristic stages, that is, lag, reactive and
stabilising for each current occurs at different
Table 1
Flocculation mechanisms and associated uid regime
Flocculation Me-
chanism
Transport Mechanism Forces on par-
ticles
Perikinetic floccula-
tion
Brownian fluid motion Random mo-
tion
Orthokinetic floccu-
lation
Laminar flow Low shear
Turbulent condi-
tions
Fluctuating flows and
eddies
High shear
Differential settling Unequal settling velo-
cities
Gravity
aluminium levels. Efficiency, defined as aluminium
delivered per unit of pollutant removed, appears
highest for a current of 0.25 A. As the current
increases, the efficiency, on an aluminium basis,
decreases.
At high current, the coagulant dosage rate
increases resulting in a greater amount of pre-
cipitate in the stable stage. Likewise the bubble
1
clay solution subjected to electrocoagulation at a current of 0.25, 0.5, 1 and 2 A as a
function of time.
Table 2
Mass Balance of electrocoagulation reactor. I/0.25, 1.0 and
2.0 A; 1 g l
1
clay solution
Current 0.25 A 1.0 A 2.0 A
Clay at surface of reactor 26% 54% 63%
Clay at base of reactor 73% 42% 39%
Error in mass balance 1% 4% /2%
1
clay solution subjected to electrocoagulation at a current of 0.25, 0.5, 1 and 2 A as a
function of equivalent aluminium.
density increases resulting in a greater upwards
momentum flux and thus faster removal of pollu-
tant and coagulant by flotation [7]. At higher
currents, more coagulant (aluminium) is available
per unit time but there is a decrease in its residence
time in the reactor. Hence there is a reduction in
the probability of collision and adhesion between
the pollutant and coagulant. This results in the
observed decrease in efficiency on an aluminium
basis. Consequently the electrocoagulation reactor
operational current must be determined by a
trade-off between reaction time, aluminium effi-
ciency and desired pollutant path (flotation or
settling), in other words, the sludge production.
3.4. Comparison of electro- and chemical
coagulation
A direct comparison between electro- and che-
mical coagulation is not practicable. The systems
do not operate on an equivalent basis. In chemical
coagulation, coagulant addition is a discrete (shot-
fed) event with equilibrium determining alumi-
nium speciation and pH. In electrocoagulation the
coagulant addition is a function of current and
time. Concentration gradients continually shift as
the aluminium electrode corrodes yet the pH
stabilised at 8.3/8.8, depending on current (see
Fig. 12). Pollutant separation processes are also
different with settling the only removal path for
chemical coagulant while a combination flotation/
settling is employed by electrocoagulation. Thus, it
is important to compare turbidity reduction on a
more consistent basis.
A new set of chemical coagulation runs were
carried out in which the pH was varied by addition
of hydrochloric acid or sodium hydroxide solu-
tion. The alum dosages chosen for these runs (4
and 20 mg-Al l
1
) are significant for chemical
dosing because they span the range of optimal
turbidity reduction without pH adjustment (see
Fig. 5), and for electrocoagulation because they
span the stage during which most turbidity reduc-
tion occurs. Thereby enabling comparison at
comparable pH values and aluminium concentra-
tions. Normalised turbidity response curves for the
chemical dosing jar tests at a range of pHs are
presented in Fig. 14 and Fig. 15 for 4 and 20 mg-Al
l
1
alum dosage levels, respectively.
Alum dosing at 4 mg-Al l
1
produced average
performance and is presented in Fig. 14. A high
residual turbidity was observed with the turbidity
dropping to 10% of its initial value after settling
for 60 min. The best turbidity reduction achieved
was at an unaltered (i.e. no acid or base addition)
pH of 4.1 where the zeta potential is closest to zero
(Fig. 7). As the pH increases, zeta potential moves
away from zero indicating an increase in stability
of the solution and turbidity removal performance
declines. Indeed, under strongly basic conditions
(pH /11.4), settling performance was so poor as
to indicate substantial restabilisation of the colloi-
dal clay [17]. This sub-optimal turbidity removal
performance is more noticeable in the electrocoa-
gulation results.
Considering the electrocoagulation results in
Fig. 13, by 4 mg-Al l
1
no turbidity removal has
occurred regardless of the current used. Turbidity
removal began at 7 mg-Al l
1
for 0.25 A, with the
concentration necessary for the start of turbidity
reduction increasing as the current increases. At
4 mg-Al l
1
, the pH was far higher (8.3-10.1) than
the optimal (acidic) pH of 4.2 identified for
chemical dosing with alum where charge neutrali-
sation was identified as the dominant coagulation
mechanism [16]. As the pH increased, zeta poten-
tial decreased and performance deteriorated. The
two sets of results are, thus, consistent with the lag
stage observed in electrocoagulation being char-
acterised by a low coagulant level and a high pH
*/ a combination where both chemical and
electrocoagulation are far from optimal.
In the electrocoagulation reactor at 20 mg-Al
l
1
, pH has levelled out to 8.3/8.8, which is
towards the upper end of the range identified by
alum dosing (see Fig. 13) as leading to good
turbidity reduction. For 2 A, the system is still in
the reactive stage at 20 mg-Al l
1
with the
turbidity still dropping rapidly as the aluminium
concentration increases. On an equivalent alumi-
nium basis, efficiency has declined due to the
higher bubble production rate resulting in the
domination of flotation, as previously discussed.
At low current the coagulant release is slow,
bubble production low hence gentle agitation
favouring removal by settling conditions, which
more closely resemble those in chemical coagula-
tion jar tests. Thus a more realistic comparison
between chemical and electrocoagulation is at
low currents.
The normalised turbidity for alum dosage at
20 mg-Al l
1
is presented in Fig. 15. Here, pH
adjustment improved turbidity reduction with
good performance demonstrated over pH 5.2/
9.3, where residual turbidity fell to between 5/
7% after 60 min. This region coincides, to a first
approximation, to the pH range of 4/8 spanning
the isoelectric points identified in chemical coagu-
lation (Fig. 8). Thus implying the clay solution is
destabilised, to a certain extent, and coagulation
by a sorption mechanism. Outside this range,
Fig. 14. Normalised turbidity tests for 1 g l
1
clay solutions at 4 mg-Al l
1
subjected to chemical dosing as a function of time and pH.
Fig. 15. Normalised turbidity tests for 1 g l
1
clay solutions at 20 mg-Al l
1
subjected to chemical dosing as a function of time and
pH.
however, zeta potential decreased resulting in a
stable solution reflected by a decline in turbidity
removal performance. At a highly alkaline pH,
/10.6, little change in turbidity was observed, due
to restabilisation of the colloidal clay and the
dominance of the aluminate ion in solution [16,17].
Thus at alum concentration of 20 mg-Al l
1
turbidity removal performance is better and occurs
over a wider pH range than at 4 mg-Al l
1
.
At 20 mg-Al l
1
and at low currents (0.25 and
0.5 A) the electrocoagulation system operates in
the final (stable) stage. Aluminium hydroxide
precipitates removing the remaining clay particles
by sweep coagulation, resulting in a small change
of turbidity. In comparison to chemical coagula-
tion at 20 mg-Al l
1
, the electrocoagulation
reactor achieves better pollutant removal (than
alum dosing) with normalised turbidity being
reduced to about 3% for both 0.25 and 0.5 A.
Differences due to bubble density and removal by
flotation have been minimised by operating at low
currents. The remaining differences between the
systems are the presence of the sulphate ion from
alum (chemical coagulation) and the reaction time
for equivalent aluminium levels (electrocoagula-
tion). The presence of the sulphate anion,
SO
2
4
(aq), from alum has been suggested to
improve sweep coagulation by destabilising posi-
tively charged ions (usually hydrolysis products)
and adsorption into the Stern layer [11]. In this
case it appears electrocoagulation produces a
better result than chemical coagulation. Hence
the absence of sulphate anions and presence of
hydroxyl ions, OH
(aq), formed at the cathode,

appears to improve overall performance. How-
ever, to achieve equivalent levels of coagulant
dosage on an aluminium basis for 0.25 and 0.5
A, the reactor was operated for 85 and 43 min,
respectively. This reaction time has a direct impact
the settling time. It appears that operation at
0.25 A has longer time to allow clay pollutant to
settle but this is not the case at 0.50 A. Thus it is
questionable to the governing influence that has
resulted in the better performance of electrocoa-
gulation at 20 mg-Al l
1
. A multitude of factors
including, coagulant and pollutant type and con-
centration, dominant removal path (flotation and/
or settling), solution pH, zeta potential, aqueous
speciation to name a few, must be explored in
variety of dimensions, that is, time and space to
understand the complexities of electrocoagulation.
4. Conclusions
The dynamic nature of batch electrocoagulation
is characterised by lag, reactive and stable stages.
During the lag stage, no turbidity reduction occurs
while the pH (after an initial increase) begins to
stabilise in the range 8.3/8.8. During the reactive
stage, the isoelectric point is reached and turbidity
decreases rapidly, while in the stable stage turbid-
ity asymptotes out. In the case of clay removal, the
colloid is destabilised at a critical aluminium
concentration that is dependent on the applied
current. The results indicate that a sorption
removal mechanism is dominant during the reac-
tive stage. Further turbidity reduction in the stable
stage is attributed to pollutant removal by a sweep
coagulation mechanism.
The fundamental physical separation processes
are different for chemical dosing and electrocoa-
gulation, with settling being the key removal path
in the former case. In electrocoagulation, however,
the dominant removal path is determined by the
applied current. At low currents, settling domi-
nates. As the current increases, so the pollutant
fraction that is removed by flotation increases,
although the coagulant appears to be used less
effectively. Experimental data support the hypoth-
esis that the increased bubble densities associated
with higher currents are removing flocs of coagu-
lant from solution.
The mechanism of coagulant delivery is a key
difference between chemical dosing and electro-
coagulation performance. In the alum dosing
studied here, coagulant addition is a discrete event
with the system then moving towards an acidic
equilibrium. Modification of pH was shown to be
able to improve turbidity reduction under some
conditions. By contrast, in electrocoagulation the
continuous addition of aluminium and hydroxyl
ions (with the multi-step hydrolysis of the former)
results in a non-equilibrium state. Somewhat
surprisingly, however, the system appears stable,
at least with respect to pH, during the reactive and
stable stages. The relationship between thermo-
dynamic equilibrium and reaction kinetics, both in
the solution and in terms of pollutant removal, are
clearly significant for a deeper understanding of
electrocoagulation.
References
[1] M. Matteson, L. Regina, R. Glenn, N. Kukunoor, W.
Waits, E. Clayeld, Electrocoagulation and separation of
aqueous suspensions of ultrane particles, Colloids and
Surfaces A: Physicochemical and Engineering Aspects 104
(1) (1995) 101/109.
[2] E.A. Vik, D.A. Carlson, A.S. Eikum, E.T. Gjessing,
Electrocoagulation of potable water, Water Research 18
(11) (1984) 1355/1360.
[3] P. Holt, G. Barton, C. Mitchell, Electrocoagulation as a
wastewater treatment, The Third Annual Australian En-
vironmental Engineering Research Event, 1999, M:41/46.
[4] N.P. Barkley, C.W. Farrell, T.W. Gardner-Clayson, Alter-
nating current electrocoagulation for superfund site reme-
diation Air and Waste 43 (5) (1993) 784/789.
[5] N. Mameri, A.R. Yeddou, H. Lounici, D. Belhocine, H.
Grib, B. Bariou, Deuoridation of Septentrional Sahara
water of North Africa by electrocoagulation process using
bipolar aluminium electrodes, Water Research 32 (5)
(1998) 1604/1612.
[6] M.F. Pouet, A. Grasmick, Urban wastewater treatment by
electrocoagulation and otation, Water Science and
Technology 31 (3-4) (1995) 275/283.
[7] P.K. Holt, G.W. Barton, C.A. Mitchell, A step forward to
understanding electrocoagulation, characterisation of pol-
lutants fate, Sixth World Congress of Chemical Engineer-
ing, Melbourne, 2001.
[8] J. Lyklema, in: K.J. Ives (Ed.), The Scientic Basis of
Flocculation, Sijthoff and Noordhoff International Pub-
lishers, Apphen aan den Rijn, The Netherlands, 1978.
[9] D.N. Thomas, S.J. Judd, N. Fawcett, Flocculation Mod-
elling: A review, Water Research 33 (7) (1999) 1579/1592.
[10] R.J. Hunter, Introduction to Modern Colloid Science,
Oxford University Press, Oxford, 1993.
[11] R.D. Letterman, A. Amirtharajah, C.R. OMelia, Coagu-
lation and occulation in water quality and treatment, in:
R.D. Letterman (Ed.), A Handbook of Community Water
Supplies, Fifth, AWWA, McGraw-Hill, New York, 1999.
[12] B.V. Deryaguin, L.D. Landau, A theory of the stability of
strongly charged lyophobic sols and so the adhesion of
strongly charged particles in solutions of electrolytes, Acta
Physicochim, USSR 14 (1941) 633.
[13] E.J.W. Verwey, J.T.G. Overbeek, Theory of the Stability
of Lyphobic Colloids, Elsevier, Amsterdam, 1948.
[14] W.J. Eilbeck, G. Mattock, Chemical Processes in Waste
Water Treatment, Ellis Horwood Limited, Chichester,
1987.
[15] G. Sposito, The Environmental chemistry of aluminum,
Second, CRC Press Ltd, Florida, 1996.
[16] W. Stumm, J.J. Morgan, Chemical aspects of coagulation,
Journal American Water Works Association 54 (8) (1962)
971/992.
[17] W. Stumm, C.R. OMelia, Stoichiometry of coagulation
Journal American Water Works Association 60 (5) (1968)
514/539.
[18] A. Amirtharajah, K.M. Mills, Rapid-mix design for
mechanisms of alum coagulation, Journal American Water
Works Association 74 (4) (1982) 210/216.
[19] T.M. Riddick, Role of the zeta potential in coagulation
involving hydrous oxides, Tappi 47 (1) (1964) 171A/179.
[20] J. Bratby, Coagulation and Flocculation with an Emphasis
on Water and Wastewater Treatment, Uplands Press,
Croydon, 1980.

A Quantitative Comparison Between Chemical Dosing and Electrocoagulation

Uploaded by

Copyright:

Available Formats

A Quantitative Comparison Between Chemical Dosing and Electrocoagulation

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Quantitative Comparison Between Chemical Dosing and Electrocoagulation

Uploaded by

Copyright:

Available Formats

A quantitative comparison between chemical dosing and

(aq), formed at the cathode,

You might also like