Chemosphere 238 (2020) 124637

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Effects of powdered activated carbon on the coagulation-flocculation

process in humic acid and humic acid-kaolin water treatment
Xin Huang a, Yang Wan a, b, Baoyou Shi a, c, *, Jian Shi b, d
a
Key Laboratory of Drinking Water Science and Technology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, No. 18
Shuangqing Rd, Beijing, 100085, China
b
School of Textiles, Nantong University, Nantong, 226019, China
c
University of Chinese Academy of Sciences, Beijing, 100049, China
d
Analysis & Testing Center, Nantong University, Nantong, 226019, China

h i g h l i g h t s g r a p h i c a l a b s t r a c t

The effects of PAC dosing strategies

on coagulation behavior were
compared.

Floc properties were affected by the
PAC dosing method in the HA water
treatment.

Effects of PAC dosing method on floc
properties were small for the HA-
kaolin water.

Floc size of PAC-AS was larger than
that of the PAC2h-AS in the HA water
treatment.

Floc strength and compactness of
PAC2h-AS were higher in HA water
treatment.

a r t i c l e i n f o a b s t r a c t

Article history: The addition of powdered activated carbon (PAC) to remove micropollutants is a commonly used
Received 5 May 2019 technology to improve drinking water quality. However, the effects of PAC dosing strategy on the
Received in revised form coagulation-flocculation process of water treatment have not been well understood, especially for water
20 August 2019
with low amounts of inorganic particles. Therefore, the current research aimed to comprehensively study
Accepted 21 August 2019
Available online 22 August 2019
the effects of simultaneous addition of PAC and aluminum sulfate (AS) coagulants (denoted as PAC-AS) or
adding PAC 2 h before coagulation (denoted as PAC2h-AS) on the coagulation behavior in humic acid
Handling Editor: Xiangru Zhang (HA) and HA-kaolin water treatment. The results showed that the floc size, growth rate, breakage factor,
and fractal dimension were all enhanced by PAC-AS and PAC2h-AS for HA but not for HA-kaolin water
Keywords: treatment. In HA water treatment, PAC-AS reached a larger floc size and faster growth rate, while PAC2h-
Powdered activated carbon AS achieved a larger floc breakage factor and fractal dimension value. For PAC2h-AS, the pre-adsorption
Dosage strategy of HA onto PAC would lower the initial particle concentration and reduce the collision probability during
Coagulation-flocculation HA water coagulation process; thus, the DOC removal efficiency, floc size, and growth rate of PAC2h-AS
Floc properties
were relatively smaller than those of PAC-AS. For the floc strength and floc fractal dimension, the pre-
Water type
adsorption of HA onto PAC contributed to formation of stronger inter-particle bonds; thus, stronger
and more compact flocs were formed by PAC2h-AS compared with those of PAC-AS. The addition of PAC

* Corresponding author. Key Laboratory of Drinking Water Science and Tech-

nology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sci-
ences, No. 18 Shuangqing Rd, Beijing, 100085, China.
E-mail addresses: huangxin1212@163.com (X. Huang), 18252501781@163.com
(Y. Wan), byshi@rcees.ac.cn (B. Shi), shij413@126.com (J. Shi).

https://doi.org/10.1016/j.chemosphere.2019.124637
0045-6535/© 2019 Elsevier Ltd. All rights reserved.
2 X. Huang et al. / Chemosphere 238 (2020) 124637

had a smaller impact on the floc properties in HA-kaolin water treatment owing to its higher initial
particle concentration.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction The addition of PAC makes the coagulation-flocculation process

a ballasted-aggregation process. However, in previous studies, the
The increasing occurrence of odor compounds, microcystins, relevant parameters, such as the concentration and size distribu-
pharmaceuticals, personal care products, and other micro- tion of the ballasting agent and the velocity gradient, have been
pollutants threatens source water quality (Yang et al., 2017; Chen optimized according to the type of water to be treated (Levecq et al.,
et al., 2019; Jian et al., 2019). Thus, to improve the removal of 2007; Sieliechi et al., 2016a,b). The initial particle concentration and
micropollutants, the addition of powdered activated carbon (PAC) type will also significantly affect the aggregation process, and
in advanced conventional water treatment processes is necessary finally influence the solid/liquid separation process (Sun et al.,
and commonly used in full-scale water treatment plants (Altmann 2019). However, there is very little scientific literature on the
et al., 2015; Bonvin et al., 2016; Lompe et al., 2018; Jiang and Zhang, detailed mechanism of ballasted floc formation in different waters
2018; Jiang et al., 2018). It has been proven that the combination of containing different particles. Micro-polluted water in China usu-
adsorption and flocculation processes can effectively improve the ally contains very small amounts of inorganic particles and organic
removal of low molecular weight dissolved organic matter (DOM) carbon (Z. Liu et al., 2019). In addition, there are two general
and micropollutants that cannot be effectively removed by methods for applying PAC, namely simultaneously adding PAC with
coagulation-flocculation alone (Kristiana et al., 2011; Seckler et al., coagulants/flocculants and adding PAC several hours before
2013; Tomaszewska et al., 2004; J. Liu et al., 2019). Kristiana et al. coagulation-flocculation. Therefore, the effects of these two dosage
found that the addition of PAC improved the NOM removal and led strategies on the treatment process of water containing different
to a significant reduction (80e95%) in the formation of disinfection particles require further study.
by-products (Kristiana et al., 2011). Ho and Newcombe (2005) and To better understand the PAC-flocculation mechanism, this
Seckler et al. (2013) studied the effect of adding PAC in the removal study considered two synthetic waters (humic acid (HA) water and
of 2-methylisoborneol through the coagulation of aluminum and HA-kaolin water) with different particle concentrations and
iron salts. comprehensively investigated the effects of two PAC dosing stra-
Although the addition of PAC may contribute to an increase in tegies on the adsorption-coagulation performance and floc prop-
the removal efficacy of dissolved organic contaminants, it may erties. The adsorption-coagulation performance was evaluated by a
affect the effluent turbidity and floc formation process in the jar test, and the floc size, kinetics of floc growth, floc strength, re-
coagulation-flocculation process. Moreover, to avoid the release of covery ability, and fractal structural characteristics were monitored
adsorbed micropollutants back into treated water, the PAC should by a laser sizer. The nature of the aggregates was also characterized
be effectively separated by subsequent flocculation, sedimentation, by microscopy.
and filtration processes. The physical properties of the flocs formed
during coagulation and flocculation, including the floc size,
strength, density, and compactness, are important to the solid/
2. Materials and methods
liquid separation techniques for the efficient removal of turbidity,
organics, and other contaminants (Jarvis et al., 2004; Boller and
2.1. Test water and reagents
Blaser, 1998; Cao et al., 2011; Chu et al., 2018). Therefore, obtain-
ing flocs with a better separation ability, which usually means that
The PAC used in this study was derived from coconut shells and
they have a high density, large size, and are strong and compact, is
had an iodine value of 958 mg/g. It was purchased from Ningxia
of great importance for the subsequent separation processes.
Guanghua Co., Ltd. (China). The specific properties of the PAC were
Generally, flocs are the basic aggregates of particles, colloids, or-
as follows: particle size of 200 mesh, specific area of 944 m2/g,
ganics, and other components in water during the coagulation and
average pore size of 3.07 nm, and total pore volume of 0.57 mL/g.
flocculation process. They are usually highly porous, irregularly
The specific surface area was determined according to the
shaped, and loosely connected (Jarvis et al., 2006; Yu et al., 2011).
Brunauer-Emmett-Teller model. The density functional theory was
The addition of PAC before or during coagulation and flocculation
used to calculate the pore volume and pore size distribution from
will affect the floc characteristics, which will also affect the sepa-
the nitrogen desorption isotherms with software supplied by
ration of the PAC. However, most previous studies have focused on
Quantachrome, USA. The PAC was dispersed in deionized water at a
the adsorption efficiency of PAC, and very few studies have inves-
concentration of 1 g/L for use. Aluminum sulfate (AS) was applied
tigated its flocculation behavior, especially the floc properties.
as the coagulant in this study.
Younker and Walsh (2016) studied the impacts of adding PAC with
HA stock solution was prepared by dissolving 1 g of HA (Sigma,
ferric chloride (FeCl3) in drinking water and industrial wastewater.
USA) and 0.4 g of NaOH in 1 L of deionized water. Approximately 5 g
They found that the addition of PAC could significantly reduce the
of kaolin (<11 mm; Aladdin, China) was dispersed in 800 mL of
floc size, although it did not affect the floc strength or re-growth
deionized water under rapid mixing for 0.5 h. The solution was then
potential after exposure to a high shear rate. Krahnsto €ver and
transferred into a 1 L measuring cylinder and settled for 0.5 h. The
Wintgens optimized the flocculation parameter of PAC in waste-
upper 500 mL was extracted as the kaolin stock solution. The HA
water treatment by dosing iron salt to obtain the highest separation
synthetic water contained 10 mg/L of HA in tap water, and the HA-
of PAC from water (Krahnsto €ver and Wintgens, 2017). The effects of
kaolin synthetic water contained 10 mg/L of HA and had an initial
other adsorbents, including clay and pozzolana particles on the floc
turbidity of 15 ± 0.5 NTU, which was adjusted by the addition of
properties were also studied by Demirci et al. (1998) and Sieliechi
kaolin. The HA and HA-kaolin test waters were all prepared in tap
et al. (2016a,b), respectively.
water.
 X. Huang et al. / Chemosphere 238 (2020) 124637 3

2.2. Jar test

d3  d2
Recovery factor ð%Þ ¼  100 (2)
A series of jar tests were conducted on a program-controlled jar d1  d2
test apparatus (TS6-1, Wuhan Hengling Technology Ltd.) in six 1.5 L
where d1 is the median floc size of the steady phase before
square beakers with flat paddle impellers using the HA and HA-
breakage, d2 is the median floc size after the floc breakage period,
kaolin synthetic waters. The dosage of Al2(SO4)3 was 8 mg/L as
and d3 is the median floc size after re-growth to the new steady
aluminum, which was optimized by a series of pre-experiments, as
phase (Jarvis et al., 2005a,b).
shown in Fig. S1. The PAC was dosed according to two strategies,
The floc fractal dimension (Df) can be measured by laser light
namely 1) adding PAC 2 h before coagulation (denoted as PAC2h-
scattering. The light scattering technique involves measurement of
AS) and 2) adding PAC and the coagulants simultaneously (deno-
the light intensity (I) as a function of the scatter vector (Q). Q is
ted as PAC-AS). Moreover, coagulation by aluminum salts alone
defined as the difference between the incident and scattered wave
without PAC addition was also performed for comparison and was
vectors of the radiation beam in the medium, which is given by Eq.
denoted as AS. For the PAC2h-AS strategy, a certain amount of PAC
(3) as follows:
was added to a 1 L water sample under rapid mixing at 200 rpm for
2 h, and the solution was then introduced for the coagulation test. 4pnsin ðq=2Þ
For the PAC-AS strategy, the water sample (1 L) was first mixed at Q¼ (3)
l
200 rpm for 30 s, and then a certain amount of PAC and Al2(SO4)3
were added and rapidly mixed for 1.5 min. The water samples were where n is the refractive index of the suspending medium, q is the
then slowly stirred at 40 rpm for 15 min and settled for 20 min. scattering angle ( ), and l is the wavelength of the radiation in a
After settling, water samples of about 200 mL were collected from vacuum (nm).
2 cm below the solution surface for subsequent measurements. Flocs are mass fractal objects that can be described by the
Each experiment was repeated three times to ensure accuracy. An relationship between their mass (M), a characteristic measure of
unfiltered water sample was used for turbidity and zeta potential size (L), and the mass Df. For independently scattering aggregates, I
measurements with a Hach turbidimeter (Hach Company, Love- is related to Q and Df, as expressed by Eq. (4):
land, USA) and a Zetasizer (Nano-ZS90, Malvern), respectively. A
water sample filtered through a 0.45 mm glass filter membrane was IfQ Df (4)
used for absorbance analysis at 254 nm (UV254) and dissolved
organic carbon (DOC) analysis using a UV-6100 double beam where Df is the mass fractal dimension, which can be determined
spectrophotometer and a TOC analyzer (TOC-VCPH, Shimadzu, by the slope of a plot of I as a function of Q on a log-log scale. A
Japan), respectively. larger Df value usually indicates flocs with a more compact struc-
ture compared with those with small Df values (Jarvis et al.,
2.3. Imaging of flocs 2005a,b).

Microscope observations of flocs formed by AS, PAC-AS, and 3. Results and discussion
PAC2h-AS were performed with a Nikon confocal microscope. The
flocs were carefully taken out by a micropipette with a large 3.1. Effects of powdered activated carbon on turbidity and dissolved
opening to limit floc aggregation immediately after 15 min of slow organic matter removal
stirring. No floc coalescence was observed during the microscope
observations (Lapointe and Barbeau, 2016). The effects of adding PAC simultaneously and 2 h before coag-
ulation on the effluent residual turbidities, UV254, and DOC removal
efficiencies for these two types of water treatments are shown in
2.4. Online monitoring of floc dynamic variation during floc Fig. 1. The inconformity of the PAC dosage of 0 mg/L was owing to
formation, breakage, and re-growth the two sets of experiments that were conducted. The addition of
PAC significantly increased the residual turbidity in the treatment
Dynamic floc sizes during the floc formation, breakage, and re- of both the HA and HA-kaolin synthetic water. Moreover, the re-
growth stages were measured using a Mastersizer 2000 (Malvern, sidual turbidities increased with the PAC dosage. The residual tur-
UK). The size range of the Mastersizer 2000 was 0.02e2000 mm. bidities for the HA and HA-kaolin water treatments increased from
The test method was similar to that of the jar test mentioned above. 2.5 NTU and 3.0 NTU to approximately 6.0 NTU and 8.0 NTU,
After 15 min of slow stirring, a high shear rate (200 rpm) was respectively, as the PAC dosage increased from 0 mg/L to 50 mg/L.
introduced for 5 min to break the floc, and then another 15 min of Compared with these two PAC dosage strategies, the residual tur-
slow stirring at 40 rpm was applied for floc re-growth. Size mea- bidities for the HA water treatment were almost the same. For the
surements were taken every 0.5 min and logged on a computer. HA-kaolin water treatment, the residual turbidity of PAC2h-AS was
Floc size was expressed as an equivalent volumetric diameter and slightly higher than that of PAC-AS. For the removal of UV254 and
as D50, which referred to the median floc size selected as the DOM, the UV254 removal efficiencies of the two PAC dosage stra-
representative floc size. tegies were almost the same for the HA and HA-kaolin water
Floc breakage and recovery factors indicate the ability of flocs to treatments. In addition, the UV254 removal efficiencies were all
resist rupture by a certain velocity gradient and to recover after above 95% in all conditions. Unlike the UV254 removal efficiency, the
exposure to a high shear rate, respectively. Generally, larger values DOC removal efficiencies first increased and then decreased with
of breakage and recovery factors indicate that flocs are stronger and increasing PAC dosage, and the maximum value was attained for
easier to recover than those with lower values. The breakage and the PAC dosage of 20e40 mg/L. The DOC removal efficiencies were
recovery factors were calculated by the following equations: similar for the HA and HA-kaolin water treatments, while there
were slight differences between the two PAC dosage strategies.
d2
Breakage factor ð%Þ ¼  100 (1) Moreover, the DOC removal efficiencies for PAC-AS were slightly
d1 higher than those of PAC2h-AS in both the HA and HA-kaolin water
treatments. These results differed from the general understanding
4 X. Huang et al. / Chemosphere 238 (2020) 124637

Fig. 1. Effect of PAC on the residual turbidity: (a) HA water and (b) HA-kaolin water, UV254 removal efficiency: (c) HA water and (d) HA-kaolin water, and DOC removal efficiency: (e)
HA water and (f) HA-kaolin water.

that PAC addition long before coagulation improves the removal of efficiency were not due to the changes in the charge neutralization
DOM. effect.
The zeta potentials after the addition of PAC were significantly The UV254 and DOC removal efficiencies by single PAC adsorp-
reduced from approximately 15 mV to 30 mV. After the addition tion versus time are shown in Fig. S2, and the molecular weight of
of coagulants, the zeta potentials of HA were neutralized and the HA used in the study is shown in Fig. S3. After 2 h of adsorption
reached approximately 12 mV. The zeta potentials for the two by PAC, UV254 and DOC for both the HA and HA-kaolin water
dosage strategies in the HA and HA-kaolin water treatments were treatments were only reduced by less than 20%, and the UV254
almost the same; thus, it could be inferred that the addition of PAC removal efficiency was only 15%. Thus, it could be inferred that the
had little effect on the charge neutralization effect of coagulants adsorption capacity of the PAC for HA was limited, so a longer PAC
and that the differences in the residual turbidity and DOC removal adsorption time and a higher PAC dosage would not contribute to
 X. Huang et al. / Chemosphere 238 (2020) 124637 5

higher UV254 and DOC removal efficiencies. The higher DOC coagulation alone were slightly higher in the HA-kaolin water
removal by PAC-AS may have been attributed to the complex treatment process.
functioning of PAC, coagulants, and HA in water. In the case of Fig. 4 illustrates the effects of various PAC concentrations on the
PAC2h-AS, a small portion of the HA was already adsorbed onto the floc growth stage in the HA and HA-kaolin water treatments for
PAC before coagulation, and thus the probability of collision may these two PAC dosing strategies. The PAC concentration had
have been slightly lower, which would decrease the DOM removal different effects on the floc size evolution in the HA and HA-kaolin
by adsorption and sweeping. The turbidity increased with the PAC water treatments. For the HA water treatment, the floc size
dosage, which may have resulted from the relatively low density increased with the PAC concentration for the two PAC dosage
(0.33 g/mL) of the PAC compared with those of the HA and kaolin. strategies. Nevertheless, for the HA-kaolin water treatment, the floc
Thus, the addition of PAC would result in flocs with a poor sedi- growth stage was not significantly affected by the addition of PAC at
mentation ability. Moreover, with the increasing addition of PAC, different concentrations, and the floc growth curves were very
the residual turbidity increased, possibly because there was too similar for various PAC concentrations. In the HA water treatment,
much PAC to be coagulated, and the residual PAC in the water led to the PAC increased the particle concentration and formed larger
the high turbidity. Above all, the high residual turbidity after the flocs, while for the HA-kaolin water treatment, the presence of
addition of PAC may have been attributed to poor floc sedimenta- kaolin provided a sufficient particle concentration for collisions and
tion ability and residual PAC. Further floc properties are discussed the formation of flocs with similar sizes.
in Section 3.2. Floc size can reflect the balance state between floc formation
and breakage. For the HA water treatment, the addition of PAC
3.2. Effects of powdered activated carbon on floc size and growth could increase the initial particle concentration, which would result
rate in a higher probability of collision. Finally, the floc formation rate
was higher and larger flocs were formed after the addition of PAC.
The effects of adding PAC simultaneously and 2 h before coag- Meanwhile, when adding PAC 2 h before coagulation, a small
ulation on the floc growth curves for the HA and HA-kaolin water fraction of the HA had been adsorbed onto the PAC; therefore, the
treatments are shown in Fig. 3. For the HA water treatment, the initial particle concentration during the coagulation process was
addition of PAC significantly enlarged the floc size, especially when lower than that when PAC and the coagulant were added simul-
the PAC and coagulant were added simultaneously. The floc sizes at taneously. Thus, the floc size and growth rate of PAC2h-AS were
the end of the growth stage for PAC-AS and PAC2h-AS were
approximately 500 mm and 400 mm, respectively, which were
significantly higher than those of AS alone. The results did not agree
well with those of Younker and Walsh (2016), who found that the
addition of an adsorbent significantly decreased the floc size.
However, in their study, they used FeCl3 as a coagulant to treat
synthetic water, which generally forms larger flocs than aluminum
salts. Moreover, as the floc size increased proportionally with time,
the growth rate could be calculated from the slope of the linear
part. After the addition of PAC, the floc growth rate was also
enhanced from 23 mm/min to 32 mm/min and 28 mm/min by PAC-AS
and PAC2h-AS, respectively. For the HA-kaolin water treatment, the
floc sizes under these three conditions were approximately 400 mm.
Comparing the same dosing strategy in different water treatment
processes, the floc size of PAC-AS was slightly higher in the HA
water treatment process, while the floc sizes of PAC2h-AS and

Fig. 3. Effect of PAC dosing strategy on the floc size and growth for (a) HA water
Fig. 2. Effects of PAC dose strategy on the floc zeta potential during coagulation for HA treatment and (b) HA-kaolin water treatment (PAC concentration:10 mg/L; AS dosage:
and HA-kaolin water treatments. 8 mg/L).
6 X. Huang et al. / Chemosphere 238 (2020) 124637

Fig. 4. Effect of PAC concentration on floc growth in HA water ((a) PAC-AS and (b) PAC2h-AS) and HA-kaolin water ((c) PAC-AS and (d) PAC2h-AS) treatments.

and Barbeau, 2016). They found that a larger floc size usually means
a lower residual turbidity, which is not consistent with the results
in this study. This difference may be due to the different ballast
materials applied in this research. Although the addition of PAC
could increase the probability of collision during coagulation and
contribute to the formation of larger flocs, PAC with a relatively low
density was adverse to floc sedimentation compared with micro-
sand with a high density. Thus, the floc size increased but the re-
sidual turbidity was high after the addition of PAC.

3.3. Effects of powdered activated carbon on floc strength and re-

growth properties

The floc breakage and recovery factors were calculated accord-

ing to Eq. (1) and Eq. (2); these factors are indicators of floc strength
and recovery ability, as displayed in Fig. 5. Fig. S4 shows the median
floc size for coagulation alone and for the PAC-AS and PAC2h-AS
Fig. 5. Effect of PAC dosing strategy on the floc breakage factor and recovery factor for
strategies during growth, breakage, and re-growth in the HA and
(a) HA water treatment and (b) HA-kaolin water treatment (PAC concentration:10 mg/
L; AS dosage: 8 mg/L). HA-kaolin water treatments. Similar to the floc size and growth
rate, the floc breakage and recovery factors of the different dosing
strategies for the HA and HA-kaolin water treatments showed
slightly smaller than those of PAC-AS. For the HA-kaolin water different trends. For the HA water treatment, the floc breakage and
treatment, the water with a high turbidity already contained recovery factors were enhanced by the addition of PAC, especially
enough particles, so the floc size after the addition of PAC was for PAC2h-AS, while for the HA-kaolin water treatment, the floc
similar to that when the coagulant was used alone. Moreover, a breakage and recovery factor were not affected by the addition of
previous study by Lapointe and Barbeau showed that for the bal- PAC. After the addition of PAC in the HA water treatment, the
lasted flocculation process combined with micro-sand, floc size is breakage factor was increased from 20% to 30% and 40% by PAC-AS
the most important factor affecting the residual turbidity (Lapointe and PAC2h-AS, respectively. Meanwhile, the improvement of the
recovery factor was relatively small for both PAC-AS and PAC2h-AS.
 X. Huang et al. / Chemosphere 238 (2020) 124637 7

For the HA-kaolin water treatment, the breakage and recovery

factors were nearly the same for the three strategies at approxi-
mately 35% and 45%, respectively. The results of the HA-kaolin
water treatment were similar to those of a previous study con-
ducted by Younker and Walsh who also found that the addition of
adsorbents during FeCl3 treatment did not increase nor decrease
the floc strength and recovery ability, thereby suggesting that
adsorbent addition does not significantly change the breakage
response of flocs to the operational shear rate in high turbidity
waters (Younker and Walsh, 2016).
The floc strength and recovery ability rely on the inter-particle
bonds between the components of the aggregates, and increased
floc compaction can also improve floc strength (floc compaction is
discussed in Section 3.4) (Bache et al., 1997; Yu et al., 2009). Pre-
vious studies have suggested that the ease with which flocs are
broken with aggregate breakage is directly related to the strength
and number of bonds holding the floc together (Jarvis et al.,
2005a,b; François, 1987). The addition of PAC in the HA water
treatment increased the initial particle concentration in suspen-
sion, and thus increased the collision probability. Therefore, it could
be inferred that the number of inter-particle bonds in the flocs
formed by the addition of PAC was greater than that in the flocs
formed by coagulation alone, which may have resulted in stronger
flocs with a higher floc breakage factor. In addition, unlike the floc
size, the floc breakage factor of PAC2h-AS was larger than that of
PAC-AS, which may have been due to the long contact time be-
tween the PAC and the HA leading to stronger connections. For the
HA-kaolin water treatment, the number of inter-particle bonds
with and without PAC addition may have been similar, so the floc
strengths were similar.
The floc recovery ability is usually dependent upon formation
mechanisms (Yukselen and Gregory, 2004). Previous studies have
found that compared with flocs formed by the charge neutraliza-
tion effect, sweep flocs display a limited re-growth potential (Jarvis Fig. 6. Effect of PAC dosing strategy on the floc fractal dimension for (a) HA water
et al., 2005a,b; Yukselen and Gregory, 2004). In the current study, treatment and (b) HA-kaolin water treatment (PAC concentration:10 mg/L; AS dosage:
the zeta potentials of the flocs in both the HA and HA-kaolin water 8 mg/L).
treatments by these three dosing strategies were similar, as shown
in Fig. 2. It could be deduced that the addition of PAC had no in-
fluence on the charge neutralization effect of the coagulation pro-
cess. Moreover, the limited recovery ability of the flocs formed with presence of PAC. This result is in agreement with that of previous
and without PAC addition indicated that enmeshment in the studies by Sieliechi et al., who suggested that the flocs formed by
aluminum hydroxide precipitate was the aggregation mechanism ballasted flocculation with pozzolana are compact (Sieliechi et al.,
followed for PAC in the floc structure. 2016a,b). Moreover, the differences among the floc Df values of
AS, PAC-AS, and PAC2h-AS during the floc growth stage were sig-
3.4. Effect of powdered activated carbon on floc structure nificant, while the floc Df values of AS, PAC-AS, and PAC2h-AS
during the breakage and re-growth stages were similar. This
The floc Df was calculated by Eq. (4) during floc growth, result may have been because the formed flocs could re-adjust
breakage, and re-growth, and the results are shown in Fig. 6. An themselves and form a more compact structure after exposure to
increase in the floc Df value during the floc growth stage was a high shear rate. However, for the HA-kaolin water treatment, the
observed when PAC was added. The initial high Df value was due to differences among the floc Df values of these three dosing strategies
the presence of PAC, which was more compact than the formed did not change during the floc growth, breakage, and re-growth
flocs. Afterwards, there was an increase in the floc Df value during processes. In addition, for the HA water treatment, after the addi-
floc breakage and re-growth, which agreed with previous studies tion of PAC, more compact flocs with larger Df values were formed,
by Jarvis et al. (2006) and Zhao et al. (2013). The loose and open while for the HA-kaolin water treatment, slightly looser flocs with
parts of the floc were damaged and cut by the high shear rate, and smaller Df values were formed with PAC addition. The microscope
the compact and strong parts remained. Therefore, the floc Df value graphs of the flocs formed by AS, PAC-AS, and PAC2h-AS in the HA
during the breakage stage increased. When the slow shear rate was and HA-kaolin water treatments are illustrated in Fig. 7, in which
re-applied, the remaining part aggregated and formed more the floc structure could be more directly observed. In addition, the
compact flocs. Compared with these two treated waters (HA and flocs formed in the HA-kaolin water were more compact. For the
HA-kaolin water treatments), the flocs that formed in the HA- flocs formed in the HA water, the PAC-AS flocs obtained the most
kaolin water generally had larger Df values compared with those compact structure, which was consistent with the results obtained
that formed in the HA water, which meant that the flocs formed in with the laser scattering method.
the HA-kaolin water were more compact. For the HA water treat- Above all, the flocs formed in the HA water with the addition of
ment process, the higher Df values found in the case of PAC-AS and PAC were more compact, stronger, and larger. However, the sedi-
PAC2h-AS suggested that floc compaction was induced by the mentation capacity of these flocs seemed to be inferior to that of
8 X. Huang et al. / Chemosphere 238 (2020) 124637

Fig. 7. Microscopy images of flocs formed by AS, PAC-AS and PAC2h-AS for HA (a), (c), (e) and HA-kaolin (b), (d), (f) water treatment (PAC concentration:10 mg/L; AS dosage: 8 mg/L).

the flocs formed by coagulation alone (higher residual turbidity flocs composed of ballasts. Although larger, stronger, and more
after the addition of PAC). These results may have been related to compact flocs were formed with the addition of PAC, the density of
the density of the components participating in floc formation. The the flocs was significantly lower compared with that of the flocs
order of the density among HA, kaolin, and PAC was as follows: formed only by coagulation. Therefore, for the addition of PAC, the
kaolin > HA > PAC. Fig. 7 shows that the PAC was a vital part of the effluent residual turbidites were high. The use of PAC in low
flocs formed by the PAC-AS and PAC2h-AS strategies. Although the turbidity waters should receive more attention to ensure the
connections between the PAC-HA aggregations were strong, the effectiveness of the separation processes designed to prevent the
low PAC density reduced the overall density of the flocs. The den- release of adsorbed pollutants from the residual PAC in water
sity of the ballasted flocs was calculated from the proportion of the treatment processes.
 X. Huang et al. / Chemosphere 238 (2020) 124637 9

4. Conclusion doi.org/10.1016/j.watres.2005.06.028.
Jarvis, P., Jefferson, B., Gregory, J., Parsons, S.A., 2005a. A review of floc strength and
breakage. Water Res. 39, 3121e3137. https://doi.org/10.1016/
The effects of PAC dosage strategies on the coagulation perfor- j.watres.2005.05.022.
mance and floc properties in the HA and HA-kaolin water treat- Jarvis, P., Jefferson, B., Parsons, S., 2004. The duplicity of floc strength. Water Sci.
ments were comprehensively investigated in this study. The results Technol. 50, 63e70.
Jarvis, P., Jefferson, B., Parsons, S.A., 2006. Floc structural characteristics using
indicated that the residual turbidities of both the PAC2h-AS and conventional coagulation for a high doc, low alkalinity surface water source.
PAC-AS strategies were higher than those of coagulation alone, and Water Res. 40, 2727e2737. https://doi.org/10.1016/j.watres.2006.04.024.
the turbidities increased with the PAC dosage in both the HA and Jarvis, P., Jefferson, B., Parsons, S.A., 2005b. Breakage, regrowth, and fractal nature of
natural organic matter flocs. Environ. Sci. Technol. 39, 2307e2314. https://
HA-kaolin water treatments. In addition, the DOC removal effi- doi.org/10.1021/es048854x.
ciency was slightly enhanced by the addition of PAC, especially for Jian, Z., Bai, Y., Chang, Y., Liang, J., Qu, J., 2019. Removal of micropollutants and
PAC-AS in both the HA and HA-kaolin water treatments. Moreover, cyanobacteria from drinking water using KMnO4 pre-oxidation coupled with
bioaugmentation. Chemosphere 215, 1e7. https://doi.org/10.1016/
the addition of PAC influenced the floc properties in the HA water j.chemosphere.2018.10.013.
treatment but not in the HA-kaolin water treatment. For the HA Jiang, J., Li, W., Zhang, X., Liu, J., Zhu, X., 2018. A new approach to controlling
water treatment, the floc size and floc growth rate of PAC-AS were halogenated DBPs by GAC adsorption of aromatic intermediates from chlorine
disinfection: effects of bromide and contact time. Separ. Purif. Technol. 203,
larger than those of PAC2h-AS, while the floc strength and floc 260e267. https://doi.org/10.1016/j.seppur.2018.04.050.
compactness of PAC2h-AS were higher than those of PAC-AS. The Jiang, J., Zhang, X., 2018. A smart strategy for controlling disinfection byproducts by
floc recovery ability was not affected by the addition of PAC. These reversing the sequence of activated carbon adsorption and chlorine disinfec-
tion. Sci. Bull. 63, 1167e1169. https://doi.org/10.1016/j.scib.2018.07.022.
results will contribute to further understanding the effect of the €ver, T., Wintgens, T., 2017. Optimizing the flocculation of powdered acti-
Krahnsto
PAC dosing strategy on the coagulation-flocculation process and vated carbon in wastewater treatment by dosing iron salt in single- and two-
provide guidance for the application of PAC in the treatment of stage processes. J. Water Process Eng. 20, 130e137. https://doi.org/10.1016/
water with various initial particle concentrations. j.jwpe.2017.10.010.
Kristiana, I., Joll, C., Heitz, A., 2011. Powdered activated carbon coupled with
enhanced coagulation for natural organic matter removal and disinfection by-
Acknowledgments product control: application in a Western Australian water treatment plant.
Chemosphere 83, 661e667. https://doi.org/10.1016/j.chemosphere.2011.02.017.
Lapointe, M., Barbeau, B., 2016. Characterization of ballasted flocs in water treat-
This work was supported by the National Natural Science ment using microscopy. Water Res. 90, 119e127. https://doi.org/10.1016/
Foundation of China (51808532) and the Major Science and Tech- j.watres.2015.12.018.
nology Program for Water Pollution Control and Treatment Levecq, C., Breda, C., Ursel, V., Marteil, P., Sauvignet, P., 2007. A new design of
flocculation tank: the Turbomix applied to weighted flocculation. Water Sci.
(2017ZX07502003). Technol. 56, 141e149.
Liu, J., Cheng, S., Cao, N., Geng, C., He, C., Shi, Q., Xu, C., Ni, J., DuChanois, R.M.,
Elimelech, M., Zhao, H., 2019. Actinia-like multifunctional nanocoagulant for
single-step removal of water contaminants. Nat. Nanotechnol. 14, 64e71.
Appendix A. Supplementary data https://doi.org/10.1038/s41565-018-0307-8.
Liu, Z., Wei, H., Li, A., Yang, H., 2019. Enhanced coagulation of low-turbidity micro-
Supplementary data to this article can be found online at polluted surface water: properties and optimization. J. Environ. Manag. 233,
739e747. https://doi.org/10.1016/j.jenvman.2018.08.101.
https://doi.org/10.1016/j.chemosphere.2019.124637. Lompe, K.M., Duy, S.V., Peldszus, S., Sauve , S., Barbeau, B., 2018. Removal of
micropollutants by fresh and colonized magnetic powdered activated carbon.
References J. Hazard Mater. 360, 349e355. https://doi.org/10.1016/j.jhazmat.2018.07.072.
Seckler, F.F.S., Margarida, M., Rosemeire, A.L., 2013. Interference of iron as a coag-
ulant on MIB removal by powdered activated carbon adsorption for low
Altmann, J., Zietzschmann, F., Geiling, E.-L., Ruhl, A.S., Sperlich, A., Jekel, M., 2015.
turbidity waters. J. Environ. Sci. 25, 1575e1582. https://doi.org/10.1016/S1001-
Impacts of coagulation on the adsorption of organic micropollutants onto
0742(12)60231-9.
powdered activated carbon in treated domestic wastewater. Chemosphere 125,
Sieliechi, J., Lartiges, B., Skali-Lami, S., Kayem, J., Kamga, R., 2016a. Floc compaction
198e204. https://doi.org/10.1016/j.chemosphere.2014.12.061.
during ballasted aggregation. Water Res. 105, 361e369. https://doi.org/10.1016/
Bache, D.H., Johnson, C., McGilligan, J.F., Rasool, E., 1997. A conceptual view of floc
j.watres.2016.09.015.
structure in the sweep floc domain. Water Sci. Technol. 36, 49e56. https://
Sieliechi, J., Lartiges, B., Skali-Lami, S., Kayem, J., Kamga, R., 2016b. Floc compaction
doi.org/10.1016/S0273-1223(97)00418-6.
during ballasted aggregation. Water Res. 105, 361e369. https://doi.org/10.1016/
Boller, M., Blaser, S., 1998. Particles under stress. Water Sci. Technol. 37, 9e29.
j.watres.2016.09.015.
https://doi.org/10.1016/S0273-1223(98)00303-5.
Sun, H., Jiao, R., Xu, H., An, G., Wang, D., 2019. The influence of particle size and
Bonvin, F., Jost, L., Randin, L., Bonvin, E., Kohn, T., 2016. Super-fine powdered acti-
concentration combined with pH on coagulation mechanisms. J. Environ. Sci.
vated carbon (SPAC) for efficient removal of micropollutants from wastewater
82, 39e46. https://doi.org/10.1016/j.jes.2019.02.021.
treatment plant effluent. Water Res. 90, 90e99. https://doi.org/10.1016/
Tomaszewska, M., Mozia, S., Morawski, A.W., 2004. Removal of organic matter by
j.watres.2015.12.001.
coagulation enhanced with adsorption on PAC. Desalination 161, 79e87. https://
Cao, B., Gao, B., Liu, X., Wang, M., Yang, Z., Yue, Q., 2011. The impact of pH on floc
doi.org/10.1016/S0011-9164(04)90042-2.
structure characteristic of polyferric chloride in a low DOC and high alkalinity
Yang, Y.-Y., Toor, G.S., Wilson, P.C., Williams, C.F., 2017. Micropollutants in ground-
surface water treatment. Water Res. 45, 6181e6188. https://doi.org/10.1016/
water from septic systems: transformations, transport mechanisms, and human
j.watres.2011.09.019.
health risk assessment. Water Res. 123, 258e267. https://doi.org/10.1016/
Chen, Y., Xu, W., Zhu, H., Wei, D., He, F., Wang, D., Du, B., Wei, Q., 2019. Effect of
j.watres.2017.06.054.
turbidity on micropollutant removal and membrane fouling by MIEX/ultrafil-
Younker, J.M., Walsh, M.E., 2016. Effect of adsorbent addition on floc formation and
tration hybrid process. Chemosphere 216, 488e498. https://doi.org/10.1016/
clarification. Water Res. 98, 1e8. https://doi.org/10.1016/j.watres.2016.03.044.
j.chemosphere.2018.10.148.
Yu, W., Gregory, J., Campos, L., Li, G., 2011. The role of mixing conditions on floc
Chu, Y.-B., Li, M., Liu, J.-W., Xu, W., Cheng, S.-H., Zhao, H.-Z., 2018. Molecular insights
growth, breakage and re-growth. Chem. Eng. J. 171, 425e430. https://doi.org/
into the mechanism and the efficiency-structure relationship of phosphorus
10.1016/j.cej.2011.03.098.
removal by coagulation. Water Res. 147, 195e203. https://doi.org/10.1016/
Yu, W., Li, G., Xu, Y., Yang, X., 2009. Breakage and re-growth of flocs formed by alum
j.watres.2018.10.006.
 € and PACl. Powder Technol. 189, 439e443. https://doi.org/10.1016/
Demırcı, Ş., ErdoGan, B., Ozcımder, R., 1998. Wastewater treatment at the petroleum
j.powtec.2008.07.008.
refinery, Kirikkale, Turkey using some coagulants and Turkish clays as coagu-
Yukselen, M.A., Gregory, J., 2004. The reversibility of floc breakage. Int. J. Miner.
lant aids. Water Res. 32, 3495e3499. https://doi.org/10.1016/S0043-1354(98)
Process. 73, 251e259. https://doi.org/10.1016/s0301-7516(03)00077-2.
00111-0.
Zhao, Y.X., Gao, B.Y., Qi, Q.B., Wang, Y., Phuntsho, S., Kim, J.H., Yue, Q.Y., Li, Q.,
François, R.J., 1987. Strength of aluminium hydroxide flocs. Water Res. 21,
Shon, H.K., 2013. Cationic polyacrylamide as coagulant aid with titanium tet-
1023e1030. https://doi.org/10.1016/0043-1354(87)90023-6.
rachloride for low molecule organic matter removal. J. Hazard Mater. 258e259,
Ho, L., Newcombe, G., 2005. Effect of NOM, turbidity and floc size on the PAC
84e92. https://doi.org/10.1016/j.jhazmat.2013.04.044.
adsorption of MIB during alum coagulation. Water Res. 39, 3668e3674. https://