w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 9 5 9 e3 9 6 5

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Breakage and re-growth of flocs formed by charge

neutralization using alum and polyDADMAC

Wenzheng Yu 1, John Gregory*, Luiza C. Campos

Department of Civil, Environmental and Geomatic Engineering, University College London, Gower Street, London WC1E 6BT, UK

article info abstract

Article history: The formation, breakage and re-growth of flocs were investigated using alum and poly-
Received 7 February 2010 DADMAC to explore the reversibility of floc breakage. There is a significant reversibility of
Received in revised form the breakage process, i.e. the broken flocs can re-grow to the size before breakage, when
24 April 2010 charge neutralization dominates the coagulation mechanism. However, for higher alum
Accepted 26 April 2010 dosage, the break-up process displayed a distinct irreversibility. When coagulated in
Available online 20 May 2010 charge neutralization, the re-growth process of alum was nearly the same as that of pol-
yDADMAC. The average size, coagulation rate and fractal dimension of flocs before and
Keywords: after breakage were nearly the same, including alum and polyDADMAC. While at higher
Charge neutralization alum dosage, the average size, coagulation rate and fractal dimension of flocs after
Floc breakage breakage were much lower than that before breakage. Most important is that the number
Re-growth of small flocs after breakage and re-growth was much less than before breakage when
Alum charge neutralization dominated the coagulation mechanism. On the contrary, at higher
PolyDADMAC alum dosage, the small flocs, after breakage and re-growth, increased. The fractal
dimension of flocs with alum increased as coagulation time increased until a limiting floc
size was reached, while for higher alum dosage, it decreased, whether before or after
breakage. The determining parameter for floc re-growth is probably not the fractal
dimension, but rather the chemical characteristics of the flocs surface.
ª 2010 Elsevier Ltd. All rights reserved.

1. Introduction breakage. NOM flocs re-growth after exposure to high shear

Floc removal efficiency in water and wastewater treatment is
influenced by the size, fractal dimension and density of flocs.
The strength and re-growth ability of flocs can also influence
the characteristics of flocs if they are exposed to high shear. In
addition, if the size of the broken flocs can return to their size
before breakage, its sludge can be potentially reused and so
decrease the usage of coagulant.
Some works (Xu et al., 2010; Wang et al., 2009) on flocs
formed with natural organic matter (NOM) showed that it was
difficult for the broken NOM flocs return to their size before
was found limited for waters coagulated with alum, ferric
sulfate and polyDADMAC (Jarvis et al., 2005). Wang et al. (2009)
considered that there was limited re-growth of broken humic
acid flocs when the coagulant was alum or polyaluminum
chloride (PACl), although alum formed larger and stronger
flocs than those with PACl. When water contains particles
such as kaolin, whether the flocs can return to the original size
before breakage is decided by the nature of broken flocs
(Yukselen and Gregory, 2004a; Li et al., 2007).
Yukselen and Gregory (2004a,b) found that only limited re-
growth of kaolin flocs occurred indicating a significant

* Corresponding author. Tel.: þ44 20 7689 7818; fax: þ44 20 7380 0986.
E-mail addresses: yuwenzheng_163@163.com (W. Yu), j.gregory@ucl.ac.uk (J. Gregory), l.campos@ucl.ac.uk (L.C. Campos).
1
State Key Laboratory of Urban Water Resource and Environment (SKLUWRE), School of Municipal & Environmental Engineering,
Harbin Institute of Technology, No 73 Huanghe Road, Nangang District, Harbin 150090, China.
0043-1354/$ e see front matter ª 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2010.04.032
3960 w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 9 5 9 e3 9 6 5
irreversibility of the floc break-up process; and residual
turbidity increased after floc breakage and re-growth when
the coagulant was alum or PACl. While for cationic poly-
electrolytes, the re-growth of broken flocs occurred to a much
greater extent and flocs breakage was almost fully reversible.
Li et al. (2007) also showed that broken kaolin flocs were
difficult to re-grow when alum was used. All of the above
studies were concerned with sweep coagulation. By contrast,
Yu et al. (2009) found that flocs formed from kaolin by charge
neutralization could fully re-grow and the residual turbidity
was lower than that before breakage when both alum and
PACl were used as coagulant. These different results may be
due to different coagulation mechanisms.
Normally, sweep coagulation is the dominant mechanism
at optimal dosage of alum or PACl, and it is difficult for broken
flocs to fully re-grow to their initial size before breakage.
Therefore, this work aimed to evaluate the re-growth ability of
flocs when coagulation is carried out by alum and poly-
DADMAC in the charge neutralization zone. Floc characteris-
tics, including size and fractal dimension before and after
breakage were investigated.
The results are of some practical interest, since floc
breakage is inevitable in most water treatment processes.
Improved knowledge of the factors affecting the re-growth
ability of broken flocs and the effect of floc breakage and re-
growth on particle separation processes would be of consid-
erable technical value.
2. Methodology
2.1. Suspension
Kaolin clay (Imerys, St Austell, Cornwall, UK) was used as the
model suspension. 200 g of kaolin was dispersed in 500 ml of
deionized water (Purite) in a high-speed blender. To obtain full
dispersion it was necessary to raise the pH of the suspension
to about 7.5, which was achieved by adding 5 ml of 0.1 M
NaOH. After blending at 4000 rpm for 10 min the clay
suspension was diluted to 1 L with deionized (DI) water and
allowed to stand overnight in a measuring cylinder. The top
800 ml was decanted and its solids content was determined
gravimetrically and found to be 133 g/L. For the flocculation
tests, the stock suspension was diluted to give a final clay
concentration of 50 mg/L. The water was DI water with 5 mM
NaHCO3 and adjusted to pH 5.0.
2.2. Coagulants
Aluminum sulfate hydrate (Al2(SO4)3 16H2O; Fisons, >96%)
‘alum’ was used. Stock alum solutions were prepared at
a concentration of 0.1 M. One commercial cationic
polyelectrolyte from Ciba Specialty Chemicals was used:
Magnafloc 1697 (poly(diallyldimethylammonium) chloride,
polyDADMAC), with an intrinsic viscosity (IV) of about 0.2,
corresponding to a molecular weight of around 50,000, and
a high charge density (about 6 meq/g). This was supplied as
a 40% aqueous solution and was diluted to give a stock solu-
tion of 0.1% or 1 g/L.
2.3. Jar test
The test solutions were prepared using DI water with addition
of 5 mM NaHCO3. All reagents used were of analytical grade,
and the solution pH was kept at 5. For the flocculation tests,
the stock suspension was diluted in the water to give a clay
concentration of 50 mg/L. The pH of the solution was adjusted
to 5 by addition of 0.1 M NaOH or HCl. The experiments were
conducted at a temperature of 25  1 .
A known amount of alum was added into the test
suspension (800 ml) and at the same time the stirring of
flocculator (Flocculator 2000) started. The stirring speed was
set as 200 rpm (184 s1) for 1 min and 50 rpm (23 s1) or 10 min,
and increased to 200 rpm for 1 min to break the flocs and then
back to 50 rpm for 10 min for flocs to re-grow.
Experiments on the kinetics of formation, breakage and
subsequent re-growth of flocs were performed using the
“turbidity fluctuation” technique, as used in the Photometric
Dispersion Analyzer (PDA-2000, Rank Brothers, UK). The
experimental procedure was similar to that of Yukselen and
Gregory (2004a). In this method, the average transmitted
light intensity (dc value) and the root mean square (rms) value
of the fluctuating component are measured. The ratio (rms/dc)
is often termed the Flocculation Index (FI) and it provides
a sensitive measure of particle aggregation (Gregory and
Nelson, 1986). It significantly increases as aggregation
occurs, and decreases when aggregates are broken. In this
work, after the FI value reached an initial steady value, coag-
ulant was added into the suspension and the FI value was
recorded by a PC data acquisition system (Pico ADC-11, Pico
Technology, UK) at 1 s intervals.
Coagulation rate is here defined as the rate of change of the
FI value since this is directly related to the rate of change of
floc size (Xiao et al., 2008).
2.4. Electrophoretic mobility
Electrophoretic mobility (EM) was measured on the kaolin
suspension (before coagulation) and coagulated sample after
1 min of rapid mixing by a Particle Electrophoresis Apparatus
Mk 2 (Rank Brothers Ltd, Cambridge, UK). In addition, the EM
of flocs after 1 min of breakage by high stirring speed was
measured. The delay time of the measurement was about 40 s.
The average EM value for a sample was determined from 20
measurements.
2.5. Image analysis
During the slow mixing period, samples of flocs were taken
from below the surface of the suspension with a hollow glass
tube with an inner diameter of 5 mm. One end of tube was
inserted 3.0 cm below the surface and the other end was
covered by a finger, and then the samples were withdrawn
carefully. After transferring the sample onto a flat microscope
slide, the image of flocs in the sample was captured by an
optical microscope with a CCD camera (GE-5, Aigo, China). The
camera has a sensor matrix consisting of 1280 (horizon-
tal)  1024 (vertical) pixels. Each pixel is recorded using bmp
files. To interpret the image sizes correctly, a graduated
microscale was photographed to determine the number of
 w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 9 5 9 e3 9 6 5
pixels corresponding to a given standard length. Images were
obtained from an interrogation window of about
2230 mm  1784 mm with a resolution of 574 pixels/mm were
achieved. Thus 1 pixel corresponds to about 1.7 mm. All data
were recorded on the hard drive of a PC and a public domain
software package, Scion-image (Scion Corporation, Frederick,
MD), was used to analyze the captured images.
The two-dimensional fractal dimension is defined by
a power law relation between projected area (As) and the
characteristic length of the aggregates, l. According to Logan
and Kilps (1995), Grijspeerdt and Verstraete, 1997 and
Chakraborti et al. (2000) it can be derived from:
As w1D2
where D2 is the two-dimensional fractal dimension.
The two-dimensional fractal dimension can only take
values between 1 and 2. The value of D2 essentially reflects the
shape of the projected floc image. If the image is circular, then
D2 ¼ 2, indicating a spherical floc. With more elongated flocs,
D2 takes lower values and D2 ¼ 1 would indicate a linear
structure.
In this work the real l and area were substituted by the long
dimension (in pixels) and the number of pixels in the floc
image respectively. The fractal dimension was determined for
flocs larger than 30 pixels (52 mm), and the average size for
flocs larger than 3 pixels (5 mm).
3. Results and discussion
3.1. EM of flocs before and after breakage
The effect of alum and polyDADMAC concentration on the EM
of flocs before and after breakage is shown in Fig. 1. Prior to
dosage, the EM of kaolin particles was negative. The addition
of low dosage of alum caused a rapid increase of EM and the
charge became positive. At dosages higher than 0.1 mM of
alum, the EM was relatively constant with alum dosage. For
Fig. 1 e Electrophoretic mobility of flocs before and after
breakage.
3961
PolyDADMAC, the EM also increased with increasing dosage
and the flocs became positive beyond 0.2 mg/L.
The EM obtained 1 min after alum addition was found to be
nearly the same as that obtained 1 min after breakage of the
formed flocs, as shown in Fig. 1. This indicated that most
particles had been neutralized by hydrolyzed Al species
during the initial 1 min after alum dosage. The average EM of
particles (flocs) did not change when the flocs were broken.
Although the average EM did not change during the break-up
process, the newly exposed surface of aggregates may have
a negative, positive or neutral charge (McCurdy et al., 2004). As
a result, charge re-distribution may cause different results for
floc re-growth by charge neutralization and sweep
(1) coagulation.
3.2. Formation, breakage and re-formation of
aggregates
The formation, breakage and re-formation of aggregates by
different dosage of alum and polyDADMAC at pH 5 are shown
in Fig. 2. The FI value is higher in charge neutralization than at
higher alum dosages (higher than 0.1 mM) at pH 5, as shown in
Fig. 2a. When the dosage of alum was higher than 0.1 mM, FI
value of alum was nearly the same. The most important
feature is that the FI value of broken flocs can return to that of
flocs before breakage in charge neutralization, while at higher
Al dosages, there is a distinct irreversibility of the break-up
process. However for polyDADMAC, the FI value can return to
Fig. 2 e Formation, breakage and re-formation of
aggregates for different dosage of (a) alum (mmol) and (b)
polyDADMAC (mg/L) at pH 5.
3962 w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 9 5 9 e3 9 6 5
the same value as before breakage, and this was not influ-
enced by the dosage of polyDADMAC.
The FI value of broken flocs at breakage time, besides the FI
value before breakage, increased initially and then decreased
with increasing dosage of both alum and polyDADMAC, indi-
cating that the strength of the flocs (between clusters) was
different. As shown in Fig. 2a, the FI value of broken flocs by
charge neutralization was higher than that at higher Al
dosages. Kobayashi et al. (1999) considered that, the discrete
number of clusterecluster bond(s) and the magnitude of
cohesive force dominate the breakage of a floc. The large flocs
will be then be broken points corresponding to clusterecluster
bonds, depending on the relative magnitude of the cohesive
force and the applied shear force. The large reduction of FI
value on increasing the stirring rate strongly suggests frag-
mentation of flocs, rather than surface erosion as the domi-
nant breakage mechanism.
As suggested by Xiao et al. (2008), the coagulation rate, as
well as the highest FI value, was used to describe coagulation
before and after breakage. As shown in Fig. 3, the changes of
coagulation rate were distinctly different between alum and
polyDADMAC. When alum was used, there was apparently an
optimal dosage (0.02 mM) for which coagulation rate was
maximum (4.25  105 s1) and then started to decrease until
a constant value was reached over the dosage range of
0.1e0.4 mM. For alum dosages lesser than 0.03 mM, the
Fig. 3 e Coagulation rate of flocs before and after breakage
with dosage of (a) alum (mmol/L) and (b) polyDADMAC
(mg/L) at pH 5.
coagulation rate was nearly the same before and after
breakage. However, for alum dosages higher than 0.04 mM,
the coagulation rate after breakage decreased more sharply
than that before breakage, and the FI value (Fig. 2a) and
coagulation rate (Fig. 3a) after breakage were lesser than
before breakage. According to Fig. 1, EM values of flocs do not
change as a result of breakage. When alum was used as
coagulant, there was significant reversibility of break-up
process in charge neutralization, while at higher dosages, only
limited re-growth occurred. Irreversibility of the break-up
process is probably caused by the different surface charac-
teristics of broken flocs.
When polyDADMAC (Fig. 3b) was used, the highest coag-
ulation rate occurred when EM was near zero. The coagulation
rate of the flocs before and after breakage was nearly the same
for all dosages (Fig. 3b). Comparing Fig. 3a with b, for alum
dosages lower than 0.03 mM, the coagulation rates of flocs
before and after breakage were almost the same, and the
result was similar with those of polyDADMAC flocs before and
after breakage with dosage from 25 to 400 mg/L. These results
suggest that if charge neutralization dominates the coagula-
tion mechanism, there is full re-growth of broken flocs.
The FI value and coagulation rate with polyDADMAC were
significantly higher than for coagulation with alum in charge
neutralization, indicating stronger flocs. It is very likely that
the increased floc strength with polyDADMAC is a result of
additional attractive forces between particles caused by
“electrostatic patch” effects (Gregory, 1973). The polyDADMAC
used was of quite low molecular weight (about 50,000) and so
polymer bridging would not be a significant factor, especially
since cationic polyelectrolytes tend to adopt a rather flat
adsorbed configuration on negatively charged surfaces
Comparing Fig. 2a with b, the time at which the FI value began
to increase (effectively, the onset of coagulation) was signifi-
cantly shorter for alum than for polyDADMAC at optimum
dosage. This is very likely due to the relatively slow adsorption
of the cationic polymer (Gregory, 1988). The electric patch
interaction is a physical bond, and there is no obvious reason
why flocs should not re-grow fully after breakage.
3.3. The average size and fractal dimension of flocs
The size and fractal dimension of flocs influences solid/liquid
separation. The Flocculation Index is strongly influenced by
large flocs (Gregory and Chung, 1995) and smaller flocs do not
greatly affect the FI value, although they do contribute to
residual turbidity. Therefore, in addition to FI values, the
number average size of flocs from image analysis was deter-
mined. A typical microscope image of flocs is shown in Fig. 4,
for alum at a dosage of 0.02 mM Al. Derived floc sizes for
different coagulant dosages are shown in Fig. 5.
The highest average size of flocs seems to be decided by the
EM of particles surface when coagulation mechanism was
charge neutralization. Comparing Fig. 5a with Fig. 3a, the
coagulation rate changed in the same way as that of the
average size. The average size of flocs after breakage was
slightly higher than that before breakage, suggesting that
there is a significant reversibility under conditions of charge
neutralization. At higher alum dosage, the average size of
flocs before breakage was smaller than of flocs formed at
 w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 9 5 9 e3 9 6 5
Fig. 4 e Microscope image of flocs produced by 0.02 mM Al
at pH 5.
around zero EM; and the average size of flocs after breakage
was smaller than that before breakage. The change of average
floc size before and after breakage with coagulation time will
be discussed later.
With polyDADMAC as coagulant (Fig. 5b), the highest
average size of flocs was found to be at a dose of 200 mg/L,
when the EM of suspension was near zero (Fig. 1), and this
coincides with the maximum coagulation rate as shown in
Fig. 5 e Variation of average floc size with dosage of (a)
alum and (b) polyDADMAC at pH 5.
3963
Fig. 3b. There was no significant difference in this case
between average floc sizes before and after breakage.
Fig. 6a shows that fractal dimension (D2) of the flocs formed
by charge neutralization with alum was different with that at
higher alum dosage. In general, D2 value was higher in charge
neutralization, in contrast to the results of Lin et al. (2008) who
found that the fractal dimension of kaolin flocs formed by
sweep coagulation with polyaluminum chloride (PACl) was
higher than that by charge neutralization. However, the
fractal dimensions reported by Lin et al. were derived from
small angle light scattering and so are more closely related to
the mass fractal dimension D3, which has no direct link to D2.
So, the apparent discrepancy is not of great significance. It can
be seen in Fig. 6b that the D2 value did not change or only
decreased slightly after breakage when charge neutralization
dominated the coagulation mechanism. However, for high
dosage of alum, D2 decreased after flocs were broken and re-
grown, especially when the dosage of alum was higher than
0.1 mM.
With polyDADMAC the D2 values before and after breakage
were very similar to those with alum at charge neutralization.
D2 values before and after floc breakage were almost the same
(1.80e1.85) and they did not change with increasing dosage of
polyDADMAC (Fig. 6b). These values were similar to those
found for alum flocs when coagulated in charge neutralization
(0.02 mM).
Changes in the average size and fractal dimension of the
flocs mean that the flocs do not retain their original
Fig. 6 e Variation of fractal dimension of flocs with dosage
of (a) alum and (b) polyDADMAC at pH 5.
3964 w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 9 5 9 e3 9 6 5
characteristics after breakage and re-growth. When formed at
higher alum dosages, the broken flocs could not fully aggre-
gate with other flocs at the breakage point and limited
reconnection is possible, as chemical bonds were broken
(Jarvis et al., 2005). It seems that broken flocs have fewer
potential connection points for re-growth than during the
initial floc growth, and the structure of flocs became relatively
more open. That may be the reason why the fractal dimen-
sions of flocs re-grown at higher alum dosages were lower
than before breakage (Fig. 6a).
3.4. The size distribution of flocs formed with alum
Since low residual turbidity is the goal of coagulation, the size
distribution of flocs, especially the number of small flocs, is of
some importance. A higher FI value or a larger average floc
size does not necessarily mean fewer small particles in the
water. Although the FI value (Fig. 2a) and average size of flocs
(Fig. 5a) obtained with 0.4 mM alum were smaller than that
with 0.02 mM alum, the particle number frequency of small
flocs before breakage was relatively lower than that with
0.02 mM (Fig. 7). The size distribution of flocs before and after
breakage varied with the coagulation mechanism, and flocs
with small size had the highest frequency for both coagulation
mechanisms. Also, in general, the particle number frequency
decreased with increasing size of flocs for both coagulation
mechanisms. When coagulation occurred by charge neutral-
ization (Fig. 7a), the frequency of small flocs (6e20 mm) after
Fig. 7 e Floc distribution before and after breakage by alum
dosage of (a) 0.02 mM and (b) 0.4 mM.
breakage was smaller than that before breakage, and residual
turbidity was low (data were not shown). Yet for flocs of size
between 20 mm and 100 mm the frequency after breakage was
larger than before breakage. For flocs larger than 100 mm, the
frequency of particle number after breakage was only a little
lower than that before breakage. For higher dosage of alum
(0.4 mmol/L), the frequency of smaller flocs (<50 mm) after
breakage was higher than that before breakage, causing
higher residual turbidity. The flocs larger than 50 mm after
breakage were lower than that before, causing a lower FI value
(Fig. 2a), especially because of the reduced number of flocs
larger than 100 mm, as shown in Fig. 7b. The results show that
different coagulation mechanisms give different floc size
distributions before and after breakage.
3.5. The change of average size and fractal dimension of
flocs
As shown in Fig. 8a, the average size and fractal dimension of
flocs with alum by charge neutralization and higher alum
dosage changed during the processes of formation, breakage
and re-growth. The change of the average size of flocs (Fig. 8a)
shows a similar trend as the change of FI values (Fig. 2a).
Fractal dimension of flocs is an important parameter, since it
influences floc density. The lower the fractal dimension, the
more open the floc structure and the lower the effective density.
Fig. 8 e Variations of the (a) average size and (b) fractal
dimension of flocs with alum dosages during formation,
breakage and re-formation at pH 5.
 w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 9 5 9 e3 9 6 5
Floc density can have a major effect on solideliquid separation
processes (Gregory, 1998). D2 values before and after breakage
are shown in Fig. 8b. D2 values were found to increase with
coagulation time and were very little affected by floc breakage
and re-growth when coagulation occurred by charge neutrali-
zation. However, at higher alum dosages, D2 values were lower
and decreased with increasing coagulation time until breakage.
After floc breakage, D2 increased with time and then decreased
after 6 min. Interestingly, the broken flocs with low fractal
dimension were hard to re-grow, while the ones with higher
fractal dimension could fully re-grow when only low dosage of
alum was used. Therefore, the determining parameter for floc
re-growth is probably not the fractal dimension, but rather the
chemical characteristics of the floc surface.
4. Conclusions
The main conclusions of this work are:
1. Broken flocs can return to their initial size before breakage
when coagulation occurred by charge neutralization with
both alum and polyDADMAC.
2. The average size, coagulation rate and fractal dimension of
flocs before and after breakage were nearly the same for
each coagulant when coagulated by charge neutralization.
When coagulation occurred at higher alum dosages, the
average size, coagulation rate and fractal dimension of flocs
after breakage were lower than that before breakage.
3. The frequency of small flocs after breakage and re-growth
was smaller than that before breakage when coagulation
happened by charge neutralization with alum, while at
higher alum dosages, the reverse trend was observed.
4. The fractal dimension of flocs with alum in charge
neutralization increased as the coagulation time, and it
deceased at higher alum dosages, whether it was before
breakage or after breakage.
5. The fractal dimension is probably not the determining
parameter for flocs re-growth, but may be the chemical
characteristics of floc surface.
Acknowledgements
Wenzheng Yu’s research at University College London was
made possible by funding from the China Scholarship Council
and a Simon Li Scholarship from UCL.
View publication stats
3965
references
Chakraborti, R.K., Atkinson, J.F., Benschoten, J.E., 2000.
Characterization of alum floc by image analysis. Environ. Sci.
Technol. 34, 3969e3976.
Gregory, J., 1973. Rates of flocculation of latex particles by cationic
polyelectrolytes. J. Colloid Interface Sci. 42, 448e456.
Gregory, J., 1988. Polymer adsorption and flocculation in sheared
suspensions. Colloids Surf. 31, 231e251.
Gregory, J., 1998. The role of floc density in solideliquid
separation. Filtrat. Separ. 35, 571e575.
Gregory, J., Chung, H.J., 1995. Continuous monitoring of floc
properties in stirred suspensions. J. Water Supply: Res.
Technol. AQUA 44, 125e131.
Gregory, J., Nelson, D.W., 1986. Monitoring of aggregates in
flowing suspensions. Colloids Surf. 18, 175e188.
Grijspeerdt, K., Verstraete, W., 1997. Image analysis to estimate
the settle ability and concentration of activated sludge. Water
Res. 31, 1126e1134.
Jarvis, P., Jefferson, B., Parson, S.A., 2005. Breakage, re-growth and
fractal nature of natural organic matter flocs. Environ. Sci.
Technol. 39, 2307e2314.
Kobayashi, M., Adachi, Y., Ooi, S., 1999. Breakup of fractal flocs in
a turbulent flow. Langmuir 15, 4351e4356.
Li, T., Zhu, Z., Wang, D., Yao, C., Tang, H., 2007. The strength and
fractal dimension characteristics of alumekaolin flocs. Int. J.
Miner. Process. 82, 23e29.
Lin, J.-L., Huang, C., Chin, C.M., Pan, J.R., 2008. Coagulation
dynamics of fractal flocs induced by enmeshment and
electrostatic patch mechanisms. Water Res. 42, 4457e4466.
Logan, B.E., Kilps, J.R., 1995. Fractal dimensions of aggregates
formed in different fluid mechanical environments. Water
Res. 29, 443e453.
McCurdy, K., Carlson, K., Gregory, D., 2004. Flocs morphology and
cyclic shearing recovery: comparison of alum and
polyaluminum chloride coagulants. Water Res. 38, 486e494.
Wang, Y., Gao, B., Xu, X., Xu, W., Xu, G., 2009. Characterization of
floc size, strength and structure in various aluminum
coagulants treatment. J. Colloid Interface Sci. 332, 354e359.
Xiao, F., Ma, J., Yi, P., Huang, Ju.H., 2008. Effects of low
temperature on coagulation of kaolinite suspensions. Water
Res. 42, 2983e2992.
Xu, W.Y., Gao, B.Y., Yue, Q.Y., Wang, Y., 2010. Effect of shear force
and solution pH on flocs breakage and re-growth formed by
nano-Al13 polymer. Water Res. 44 (6), 1893e1899.
Yu, W., Li, G., Xu, Y., Yang, X., 2009. Breakage and re-growth of
flocs formed by alum and PACl. Powder Technol. 189 (3),
439e443.
Yukselen, M.A., Gregory, J., 2004a. The effect of rapid mixing on
the break-up and re-formation of flocs. J. Chem. Technol.
Biotechnol. 79, 782e788.
Yukselen, M.A., Gregory, J., 2004b. The reversibility of flocs
breakage. Int. J. Miner. Process 73, 251e259.