Polyaluminum chloride-functionalized colloidal gas aphrons for flotation separation of nanoparticles from water

Accepted Manuscript Title: Polyaluminum chloride-functionalized colloidal gas aphrons for flotation separation of nanoparticles from water Authors: Ming Zhang, Xiaoli Lu, Qi Zhou, Li Xie, Changming Shen PII: DOI: Reference:

S0304-3894(18)30817-3 https://doi.org/10.1016/j.jhazmat.2018.09.022 HAZMAT 19746

To appear in:

Journal of Hazardous Materials

Received date: Revised date: Accepted date:

24-6-2018 22-8-2018 7-9-2018

Please cite this article as: Zhang M, Lu X, Zhou Q, Xie L, Shen C, Polyaluminum chloride-functionalized colloidal gas aphrons for flotation separation of nanoparticles from water, Journal of Hazardous Materials (2018), https://doi.org/10.1016/j.jhazmat.2018.09.022 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Polyaluminum chloride-functionalized colloidal gas aphrons for flotation separation of nanoparticles from water

a

IP T

Ming Zhang a,b *, Xiaoli Lu a, Qi Zhou a, Li Xie a, Changming Shen c

State Key Laboratory of Pollution Control and Resources Reuse, Key Laboratory of

SC R

Yangtze River Water Environment, Institute of Biofilm Technology, College of

Environmental Science and Engineering, Tongji University, 1239 Siping Road, Shanghai 200092, China

College of Environment, Zhejiang University of Technology, Hangzhou 310014,

U

b

Shanghai Tongji Environmental Engineering and Technology CO., LTD, Shanghai

A

c

N

China

ED

M

200092, China

Corresponding author:

PT

Dr. Ming Zhang,

CC E

E-mail address: [email protected]

A

Graphical abstract

1

IP T SC R U N A M

Highlights

Optimum SNP removal and CCGA stability were achieved at the pH near the IEP(C15B).

ED



Additional salt content should be low for high CCGA stability and SNP removal.



The CCGA-induced flocs were of sphere-like morphology and smooth surface.



The regular chemical-dosing unit could be omitted from the CCGA-flotation.



Modifying PACl on CGAs led to higher SNP removal than adding it directly in

CC E

PT



water.

A

Abstract

The present work used the coagulative colloidal gas aphron (CCGA)-involved flotation as a robust technology to efficiently remove the typical engineered nanoparticles – silica nanoparticles (SNPs) from water. The inorganic polymer coagulant – polyaluminum chloride (PACl) was used to surface-functionalize the 2

zwitterionic surfactant (C15B)-based CGAs. Results denote that the physicochemical conditions of PACl/C15B mixed solution markedly influenced the flotation behaviors by changing the properties of CCGAs. The C15B molecules showed different dissociated states and interaction behaviors with Al species with the variation of pH. The addition of salt into the PACl/C15B mixed solution decreased the foamability of

IP T

solution, and the bubbles collapsed before they could efficiently capture SNPs in their rising trajectory. The optimum SNP removal (87.2%) was obtained when the pH and

SC R

the additional ionic strength of PACl/C15B mixed solution were ~4.7 and ≤ 1.0 g

NaCl/L, individually, and the pH of SNP suspension was ~9.4. Importantly, modifying

U

PACl on microbubbles took greater advantages than directly using it as coagulant in

N

terms of SNP removal and PACl utlization. The CCGAs were robust since their

A

colloidal attraction and collision efficiency with SNPs were simultaneously enhanced.

M

The PACl was more efficiently utilized during flotation whilst the regular

ED

chemical-dosing unit was omitted.

Keywords: Coagulative colloidal gas aphrons; Microbubble surface-functionalization;

CC E

PT

Flotation; Nanoparticle removal

1. Introduction

With the rapid development of nanotechnology, the engineered nanoparticles (ENPs)

A

exponentially appear in both of the consumer products and the industrial applications [1-3]. When being directly discharged, the ENPs in waste streams can result in high levels of NP concentration in the environment. The aquatic system is among the most significant acceptors of those released ENPs. Once entering the aquatic environment and being intaken by organisms and even human beings, the ENPs may lead to not 3

only the important eco-toxicological risks but also the combined hazardous effects with other contaminants [4-6]. Hence, it is of high importance to remove ENPs through the water and wastewater treatment processes. Traditional technologies were constrained by some limitations though they may be used to treat the ENP-polluted water and wastewater. The metal ENPs cannot be completely removed by activated

IP T

sludge or biofilm and probably inhibit their capacities of nutrient removal [7].

Nanofiltration membranes and reverse osmosis suffer from severe fouling and

SC R

concentration polarization, which leads to high cost and energy usage [8,9]. The

traditional coagulation-sedimentation process is effective but restrained by the high

N

U

coagulant consumption and sludge production [10].

A

The flotation with surface-functionalized microbubbles might be a promising

M

approach of improving the removal efficiency of ENPs [11,12]. It has been established by Malley that coating bubbles with cationic or nonionic polymers could

ED

change both the mobility and charges of bubbles and thus improve their flotation performance in particle separation especially for the low turbidity water [13]. Metal

PT

coagulants, polymers and surfactants can be the options of surface-modifiers which are beneficial for increasing the efficacy of capturing and separating pollutants by

CC E

bubbles [14]. When the surfactants were dosed into the saturator for bubble production, the removal of algae cells or particles with the size of 25 μm - 60 μm kept

A

stable above 95%; the upstream coagulation was no more needed and the sludge volume got reduced [15]. In the study of Shi et al. [16,17], the flotation with chitosan-modified bubbles achieved higher removal of natural organic matter than that using chitosan alone as coagulant aid. Further, compared with employing either polymers or surfactants as modifiers, the combination of surfactants and polymers 4

was found to be more effective in enhancing the collision and attachment efficiency between microbubbles and pollutants (such as algae) [18,19]. Nevertheless, it has to be pointed out that all the abovementioned research was based on dissolved air flotation, in which the air hold-up of the bubble suspension is low – 1%-2% [20,21] and the size of target particulate pollutants was in micron scale. For the ENP

IP T

pollutants, a special flotation process with higher air content is demanded.

SC R

The colloidal gas aphrons (CGAs), generated from the solution of surface active agents [22], possess the potential of robust bubbles for surface-modification and

U

flotation separation of ENPs. On the one hand, the surface-modified CGAs may be

N

obtained via either functionalizing surfactants with different coagulants/ flocculants or

A

modifying the latter with hydrophobic groups forming surface active polymers. Thus,

M

the colloidal attraction and attachment efficiency between CGAs and target particles can then be reinforced. For instance, the cationic polymer of poly(dimethylaminoethyl

ED

methacrylate) was functionalized with hydrophobic pendant groups. Positively surface-charge bubbles were then produced from the solution of the synthesized

PT

polymers, which presented strong attachment with the cyanobacteria cells during flotation [18]. On the other hand, it is known that, in addition to the colloidal forces,

CC E

the Brownian diffusion is a predominant mechanism determining the flotation rate of particles smaller than 100 nm [24,25]. The collection of ENPs could be enhanced by

A

adopting small bubbles, such as the typical microbubbles of CGAs. Besides, the tiny size (10-100 μm in diameter) and high air hold-up (>50%) of CGAs can result in a great specific area and large bubble density for the capture of ENPs.

The surface-functionalization of CGAs is feasible on the premise that the complex of 5

surfactant and coagulant/flocculant is successfully formed. The attractive interaction between surfactant molecules and polymers (or polyelectrolytes) can be achieved by electrostatic attraction, excluded volume, van der Waals and other contributions [25]. Particularly, the zwitterionic surfactants are of great capacity in forming complex with polymers or other surfactant molecules [27,28]. This study selected the zwitterionic

IP T

surfactant – betaine to complex with the inorganic polymer coagulant – polyaluminum chloride (PACl). The former shows the characteristics of cationic or anionic surfactant

SC R

under different conditions due to its quaternary-N functional group of and carboxyl functional group. The latter contains the well-known polymeric Al species of Al13

U

polycation which has been widely used for the destabilization of particulate matter in

N

water treatment plants [28]. The unique and preferable properties endow the two types

A

of compounds with high potential of bubble surface-functionalization. The so-created

M

bubbles are named as coagulative CGAs (CCGAs). However, the characteristics of CCGAs, the performance of CCGAs-involved flotation (briefly named as

ED

CCGA-flotation) as well as the related mechanisms have not been sufficiently

PT

explored to the authors’ knowledge.

The flotation efficacy of PACl-modified CGAs was evaluated in removing ENPs from

CC E

aquatic suspension in the current investigation. The silica nanoparticle (SNP) was used as target ENP pollutant, which has been reported to be among the most

A

manufactured ENPs [6] and largely appears in the chemical mechanical polishing (CMP) wastewater [2,30]. The CCGA characteristics and their influence on the flotation performance were investigated and elaborated under different physiochemical conditions. Further, the SNP removal was compared between two ways of using coagulant: 1) being surface-modified on CGAs and 2) being directly 6

dosed as traditionally. The mechanisms involved in the CCGA-flotation were analyzed and discussed according to the specific flotation behaviors and floc properties.

2. Materials and methods

IP T

2.1. Chemicals

The SNP suspension was provided by Beijing DK nanotechnology Co., LTD. All the

SC R

other reagents were purchased from Sinopharm Chemical Reagent Co., Ltd, including pentadecyl dimethyl betaine (C15B, C15H31N+(CH3)2CH2COO-), sodium bicarbonate

U

(NaHCO3), Aluminum chloride hexahydrate (AlCl3·6H2O), sodium chloride (NaCl),

N

sodium hydroxide (NaOH), hydrochloric acid (HCl) and hydrofluoric acid (HF). The reagents mentioned above were of analytical grade, excluding those for instrumental

M

A

analyses. Milli-Q water (18.2 MΩ cm) was used to prepare PACl, and all the other solutions and suspensions were prepared with deionized water. All the experiments

ED

were performed in triplicate at around 25 °C.

PT

2.2. Production and characterization of CCGAs The inorganic polymer coagulant – PACl was prepared by slow base titration [30] and

CC E

used as CGA modifier: the NaHCO3 solution (1 mmol/L) was titrated into the AlCl3 solution (0.4 mol/L as Al) at the speed of 0.4 mL/min by peristaltic pump (BT300-2J,

A

Longe CO., China), during which the latter solution was rapidly stirred to prevent forming aggregates. The Al concentration and the molar OH/Al ratio were controlled to be 0.2 mmol/L and 2.0, respectively, in the final PACl solution. The PACl/C15B mixed solution (the CCGA generation solution) was then prepared by carefully adding PACl into the C15B solution, mildly shaking for 60 s and aging for about 30 min. In 7

this study, the concentrations of PACl and C15B in the CCGA generation solution were fixed to be 1.17 mmol /L as Al and 0.34 mmol/L, respectively.

The CCGA suspension was created via the classical CGA production method – high speed agitation. The CGA generator was fabricated in laboratory scale according to

IP T

the classical design [31]. The inner wall of the perspex beaker was equidistantly

mounted with three baffles. One wide side of each baffle was fixed to the bottom of

SC R

the beaker. A high-speed agitator (SF1100, Wanjiangweihong Plant, China) was used, whose jagged-edged disk could fiercely shear the surfactant solution and produce

U

dense bubbles as CGAs. The CCGAs were created by agitating the mixed solution of

N

surfactant and coagulant at the speed of 6000 rpm for 2 min. The characteristics of

A

CCGAs were characterized at varied pH and additional ionic strength, individually, so

M

as to interpret the flotation behaviors of CCGAs. The analytical procedures of half-life time, air hold-up, zeta potential, size and morphological structure have been presented

CC E

PT

ED

previously [12,32].

2.3. CCGA-flotation experiments for SNP removal 2.3.1. Characterization of SNP aqueous suspension (simulated SNP-polluted

A

wastewater)

The SNP suspensions for the CCGA-flotation trials were obtained by diluting the commercial SNP suspension until the concentration of silica reached the predetermined value. Prior to use, all the suspensions in this study were ultrasonically vibrated for 20 min in order to well disperse SNPs. The target SNPs were 8

characterized by dynamic light scattering (Nano ZS90, Malvern Instrument, Ltd., U.K.) and transmission electron microscope (TEOL2010, JEM, Japan), respectively (Fig. S1). The silica concentration was measured by means of the HF transformation combined with molybdenum blue method [33]. The zeta potential of SNPs was investigated via Zetasizer (Nano ZS 90, Malvern Instrument, Ltd., U.K.). The pH and

IP T

conductivity of SNP suspension were measured by pH meter (PHC101, Hach, U.S.A.) and conductivity meter (HQ30d, Hach, U.S.A.), individually. The main features of the

SC R

SNP suspension are summarized in Table 1.

SNP concentration

Zeta potential

of SNPs (nm)

(mg/L)

(mV)

21.0±1.3

2898±84

-29.4±4.2

A

8.9±0.7

Conductivity (μS/cm) 92.9±4.6

M

2.3.2. CCGA-flotation trials

pH

N

Average diameter

U

Table 1. Physicochemical properties of simulated SNP-polluted wastewater

ED

The CCGA-flotation was adopted to separate SNPs from water at batch mode. Three 1-L flotation-test beakers ran simultaneously. The SNP suspension was first added

PT

into each flotation beaker before flotation, and the CCGA suspension was then

CC E

injected with peristaltic pump during flotation. The volumes of SNP suspension and CCGA suspension were fixed to be 500 mL and 300 mL, individually. The retention time of the flotation separation process was controlled to be 3 min according to the

A

results of the preliminary tests (Fig. S2).

Particularly, the ways of using PACl were compared between surface-modifying it on CGAs and adopting it as coagulant. For the latter, the experiments were carried out using a programmable jar-test apparatus (ZR4-6, China) with the following procedure: 9

1 min rapid mix (150 rpm), 15 min slow mix (30 rpm) and 20 min settling period. PACl was added at the beginning of rapid mix. The comparative experiments were carried out at equivalent PACl dosages.

2.3.3. Evaluation of CCGA-flotation performance

IP T

The samples of clarified water were taken from the bottom of the flotation beakers.

Zeta potential and pH value were measured before and after flotation with Zetasizer

SC R

(Nano ZS 90, Malvern Instruments Ltd, U.K.) and pH meter (PHC101, Hach, USA), respectively. The contents of Si and Al in the flotation effluent (spare liquid) were

U

analyzed by molybdenum blue method and inductively coupled plasma atomic



(SNPs, %), was then calculated and the dilution effect caused

A

efficiency of SNPs,

N

emission spectroscopy (ICP-AES, Agilent720ES, U.S.A.), respectively. The removal

  SN Ps, %





M

by the CCGA suspension was considered:

C 0 ,SN Ps  V SN Ps  C t ,SN Ps  V SN Ps  V C C G As



C 0 ,SN Ps  V SN Ps

   100%  

ED

(1)

 C t ,SN P s C t ,SN P s  V C C G A s  100%  1    C 0 ,SN Ps C 0 ,SN Ps  V SN Ps 

PT

where, C 0 , S N P s (mg/L) and C t , S N P s (mg/L) refer to the initial and final concentrations of SNPs in the flotation beaker, respectively; V S N P s (mL) and V C C G A s (mL) represent

CC E

the volumes of SNP suspension and CCGA suspension before flotation tests, which have been predetermined to be 500 mL and 300 mL, individually. Thereby, Eq. (1)

A

could be expressed as follows:   SN Ps, %



 C t ,SN Ps 3C t ,SN Ps  1    C 0 ,SN Ps 5 C 0 ,SN Ps 

   100%  

(2)

Similarly, the utilization efficiency of PACl,

10



(PACl, %), can be assessed as Eq.(3):

  PAC l, %





C 0 ,PAC l  V C C G As  C t ,PAC l  V SN Ps  V C C G As C 0,PACl  V CCGAs



 C t , P A C ls 5 C t ,PAC l  100%  1    C 0 , P A C ls 3C 0,PAC l 

  100%  

(3)

where, C 0 , P A C l (mg/L) and C t , P A C l (mg/L) refer to the initial and final concentrations of PACl in the CCGA suspension and the flotation beaker, respectively.

IP T

The SNP aggregates in the surface of the flotation suspension were carefully collected by plastic Pasteur pipette immediately after flotation. The SNP-flocs from the

SC R

coagulation process were obtained as well for comparison. The freeze-dried SNP aggregates were observed via scanning electron microscope (SEM, XL30FEG*,

U

Philips, Holland) for morphological characteristics. Besides, the wet aggregates were

N

analyzed by a laser light scattering instrument (Mastersizer 3000, Malvern

A

Instruments Ltd, U.K.). The collected flocs were first dispersed in the deionized water,

M

and the resulted suspension was then drawn through the optical unit of the Mastersizer by a peristaltic pump with plastic tube of 5 mm inner diameter. The flow rate was

ED

controlled to be 25 mL/min due to the delicate nature of flocs. Particularly for the size determination, each measurement took 30 s. The mean diameter, d50, was selected as

PT

the representative aggregate size. The principles for the measurement of mean

CC E

diameter, size distribution and fractal dimension can be found in the former study [34].

A

3. Results and discussion 3.1. Optimization of CCGA flow rate in CCGAs-flotation process For the separation of SNPs, the CCGAs were immediately transferred into the flotation beaker by peristaltic pump after they were generated. The flow rate of CCGA suspension might influence the interaction between bubbles and particles, and change 11

the capture and removal efficiency of SNPs. Thus, the effect of CCGA flow rate on the flotation performance was first examined.

As shown in Fig. 1(a), the SNP removal efficiency rose from ~69.5% to ~76.4% when the flow rate of CCGA suspension increased from 115 mL/min to 230 mL/min, kept

IP T

stable at around 76.6% at the flow rates between 230 mL/min and 689 mL/min, and

got reduced at higher flow rates. The variation of PACl consumption efficiency with

SC R

flow rates was consistent with that of the SNP removal. Meanwhile, over the

investigated flow rate range, the average size of flotation flocs was in the range of 9.1

U

μm – 15.7 μm. The flocs presented two populations of size at low flow rates but were

N

apt to be monodisperse at higher flow rates (Fig. 1(b). During the CCGA-flotation

A

process, when the injecting speed (or, flow rate) of CCGAs was low, such as 115

M

mL/min or 230 mL/min, the bubbles might collapse before sufficiently interacting with SNPs, and hence could not capture SNPs effectively. In this scenario, the SNPs

ED

might be collected by both of the CCGAs and the free molecules of PACl and C15B, which gave rise to the polydisperse flocs. In addition, the tube used for transferring

PT

CCGAs was 0.79 cm in inner diameter (17# Tube, Langer Peristaltic Pump, China) and 100 cm in length. Hence, the time for transferring the CCGA suspension was 25.6

CC E

s - 2.6 s as the CCGA flow rates increased from 115 mL/min to 1377 mL/min. By considering both of the SNP separation efficiency and the stability of CCGAs, the

A

optimum flow rate of CGA suspension was thus determined to be 689 mL/min for all the following experiments.

12

(b) 8

(a) 100

SNP removal

7

PACl consumption

90 80

Volume (%)

Efficiency (%)

6

70

5

QCCGA suspension=115 mL/min QCCGA suspension=230 mL/min QCCGA suspension=459 mL/min QCCGA suspension=689 mL/min QCCGA suspension=918 mL/min QCCGA suspension=1377 mL/min

4 3

60

2 50

1

200

600

400

800

1000

1200

1400

0.1

1

QCCGA suspension (mL/min)

10 Size (m)

100

1000

IP T

0 0.01

40

SC R

Fig. 1. (a) SNP removal, PACl consumption and (b) floc size distribution varying with flow rates of CGA suspension (experimental conditions: initial pH of PACl/C15B solution: ~4.7, additional NaCl concentration in PACl/C15B solution: 0 g/L, initial pH of SNP suspension: ~8.9, and concentrations of PACl and C15B in PACl/C15B solution: 1.17 mmol /L as Al and 0.34 mmol/L, respectively).

U

3.2.Flotation performance influenced by CCGA characteristics

N

The properties of PACl/C15B mixed solution can greatly impact the flotation removal

A

of SNPs by changing the characteristics of CCGAs. Therefore, the flotation

ED

generation solution was studied.

M

performance at different pH values and additional salt concentrations of CCGA

PT

3.2.1. Effect of pH of PACl/C15B mixed solution The CCGAs in this study exhibited the typical morphology of CGAs (Fig. 2(a)) as

CC E

reported elsewhere [12,35,36]. But they showed distinct flotation behaviors with different pH values of the bubble generation solution. When the pH of PACl/C15B

A

solution was approximately 4.7, the optimum SNP removal (~87.2%) and PACl consumption efficiency (~86.1%) were obtained and meanwhile, the CCGAs were of the highest stability (half-life time = ~98.0 s). In contrast, the CCGAs performed worst at the pH of 6.7 – the SNP removal and PACl utilization were approximately 32.8% and 39.0%, respectively, and the half-life time was only 26.4 s. It was found in Fig. 2(a) that the SNP separation efficiency was in accordance with the bubble 13

stability when the PACl/C15B solution was acidic and neutral; but for the basic bubble generation solution, the SNP removal was poor even though the half-life time of CCGAs was 75.2 s (at pH 8.9) and 53.1 s (at pH 10.7).

The zeta potential values of CCGAs and the corresponding flotation flocs were

IP T

measured at different pH values of PACl/C15B mixed solution and given in Fig. 2(b)

except for that of CCGAs generated at pH 6.7 whose half-life time was too short to be

SC R

measured. The zeta potential of CCGAs was lower than +5.0 mV when the

PACl/C15B mixed solution was highly acidic (such as pH 2.9) or highly basic (such

U

as pH 10.9); in the pH range of 4.5-9.0, it turned to be > +17.0 mV. This could be

N

attributed to the surface modification of the Al polycations in PACl. The positively

A

surface-charge bubbles should have been of great potential in capturing the negatively

M

surface-charged SNPs (see in Table 1) through electrostatic attraction. However, in fact, the zeta potential of flotation flocs was negative when the pH of PACl/C15B

ED

mixed solution was higher than 2.9. Thus, it could be inferred that the SNP removal efficacy was predominately determined by the coagulation mechanisms of bridging,

PT

sweeping and adsorption.

CC E

As for the properties of flotation flocs (Fig 2(c) and (d)), the average size of flocs obtained at pH 2.9 and 4.7 was around 16.8 μm and 15.2 μm, individually, falling into

A

the size range of effective microbubble capture [37]. The flocs resulting from the flotation with the highest SNP removal were more dense and compact compared with those obtained at pH 2.9 of the PACl/C15B mixed solution. For the neutral and basic PACl/C15B mixed solution, the flocs were too small to be analyzed by small angle laser light scattering and therefore their fractal dimension information is not available. 14

The isoelectric point (IEP) of C15B is at pH 4.9-5.1 [38,39]. The positively charged C15B is a cationic surfactant at pH ≤ IEP; and the electrically neutral C15B molecules are ionized to form zwitterions or internal salt at pH ≥ IEP [39]. When the pH of the PACl/C15B mixed solution was equal or quite close to the IEP, such as 4.7, the C15B

IP T

molecules would be weakly positively charged or in the form of internal salt. Thus, it can be inferred that the Al species, especially the amorphous species, of PACl

SC R

interacted with C15B molecules through bridge, sweep, adsorption, Van der Waals

force and/or hydrogen bond without destroying the foamability of the surfactant. The

U

CCGAs created from such solutions could be well modified with PACl and keep a

N

comparatively high stability. When they were adopted in flotation, the efficient

A

removal of SNPs and the high utilization of PACl could be achieved simultaneously in

M

the short retention time of 3 min. At the pH value of around 7.0, the Al species and the ionized C15B molecules tended to firmly attach with one another due to the

ED

electrostatic attraction. The amphiphilicity and foamability of C15B and the coagulation ability of PACl might be negatively affected, resulting in the unstable

PT

CCGA system and the low SNP capture efficiency. When the PACl/C15B mixed solution was basic, the repulsion between the hydroxylated Al species and the COO-

CC E

of C15B molecules occurred. Thereby, it would be hard to modify the CGA surface with the dominant Al species – colloidal Al species [40]. The resulting C15B-CGAs

A

might re-gain the stability to some extent compared with the scenario at pH 6.7. Nevertheless, the coagulation of SNPs induced by the free Al species of PACl could not be sufficiently conducted within the short retention time of CCGA-flotation, and thus the SNPs were hard to be captured and separated by bubbles.

15

(a) 100 SNP removal

PACl consumption

100 Half-life time of CCGAs 100 80

80 60

80 60

60 40

60 40

40 20

40 20

20 0

2

3

4

2

3

4

0

5

6

7

8

9

10

11

pH of PACl/C15B solution 5 6 7 8 9

10

11

(b) 20 10 Half-life time (s)time (s) Half-life

(a) 100 80

0 Flocs CCGAs

-10 -20 -30

20 0

-40

0

-50 2

3

4

8

9

10

11

15 10

SC R

2.0

U

20

2.5

N

Volume (%)

25

(d) 3.0

Fractal dimension

Floc size at pH 2.9 Floc size at pH 4.7 Floc size at pH 6.7 Floc size at pH 8.9 Floc size at pH 10.7

30

7

6

pH of PACl/C15B solution

pH of PACl/C15B solution

(c)

5

IP T

Half-life time of CCGAs

Zeta potential (mV)

PACl consumption

Efficiency Efficiency (%) (%)

SNP removal

1.5

0

A

5

1.0

10

1000 Size (nm)

10000

100000

M

1

2

3

4

5

6

pH of PACl/C15B solution

CC E

PT

ED

Fig. 2. Performance of CCGAs-flotation at different pH values of PACl/C15B solution: (a) flotation efficiency versus CCGA stability, (b) zeta potential of flotation flocs and CCGAs, (c) floc size distribution, and (d) fractal dimension (experimental conditions: flow rate of CCGA suspension: 689 mL/min, additional NaCl concentration in PACl/C15B solution: 0 g/L, initial pH of SNP suspension: ~9.4, and concentrations of PACl and C15B in PACl/C15B solution: 1.17 mmol /L as Al and 0.34 mmol/L, respectively).

3.2.2. Effect of additional salt concentration in PACl/C15B mixed solution

A

The characteristics and the flotation performance of CCGAs were explored after salt (NaCl, herein) was added into the PACl/C15B solution. The half-life time of CCGAs was approximately 98.0 s without any additional salt, and the stability got decreased with the increase of salt concentration. When the additional NaCl concentrations were 0.1 g/L and 1 g/L, the half-life time was 81.9 s and 74.4 s, individually (Fig. 3(a)). The decreased life time of CCGAs was still sufficient to support the SNP separation 16

Fractal dimensio

within the retention time (≥ 80.0%). Within this NaCl concentration range, the PACl utilization rate was 65.5-69.6%, being lower than that obtained from the scenario without any additional salt in the PACl/C15B solution (~86.1%). Meanwhile, the zeta potential values of CCGAs were around +5.5 mV (Fig. 3(b)), being much lower than that at 0 g/L additional NaCl (~+19.6 mV). This could be attributed to (i) the

IP T

adsorption of Cl- ions on the bubble-liquid interface due to diffusion and electrostatic

interaction, and (ii) the thickness compression of the CCGA electric double layer [41].

SC R

The flotation flocs were negatively surface-charged, revealing that the leading SNP

removal mechanism was not electrostatic attraction of CCGAs and particles. Besides, as demonstrated in Fig. 3(c) and (d)), the size and the fractal dimension were

U

0.49-15.7 μm and 2.20, respectively, whilst the flocs without any NaCl addition were

A

N

11.6 μm in average size and 2.47 in fractal dimension, individually. The addition of

M

NaCl into the PACl/C15B solution resulted in loose and open flocs.

ED

At the NaCl concentration ≥ 10 g/L, the CCGAs became so unstable that their life time was less than 20 s which was too short for the zeta potential measurement. The

PT

addition of salt prevented the formation of micelles, which increased the rate of attaining the equilibrium surface tension and consequently decreased foaming [42].

CC E

Under such conditions, the liquid drainage of CCGAs was so fast that the bubbles collapse before they could efficiently capture SNPs in their rising trajectory. The free

A

Al species and SNPs formed flocs by coagulation in the flotation beaker; however, without stable bubbles, part of those flocs settled down to the bottom of the flotation beaker whereas others floated up on the surface of suspension as shown in the photo of Fig. 3(a). It was then difficult to get clarified effluent and accurately evaluate the efficiency of SNP removal and PACl utilization at such high additional salt 17

100

80

80

60

60

Half-life time (s)

Efficiency (%)

(a) 100

concentrations. Hence, the CCGA-flotation was feasible for SNP removal only when 40 40 the 20salt concentration in the PACl/C15B 20 solution was controlled to be low. 0

0 0.1 1 10 100 Additional salt concentration in PACl/C15B solution (g/L) SNP removal

PACl consumption

Half-life time of CCGAs

(a) 100

(b)

100

20

40

40

20

20

0 0.1

1

-10

0

PACl consumption

(c)

Additional salt concentration in PACl/C15B solution (g/L)

Half-life time of CCGAs

(d) 3.0

Additional salt concentration=0.1 g/L Additional salt concentration=1 g/L

8

A

Fractal dimension

4

M

Volume (%)

Fractal dimension

N

2.5

6

0.1

1

ED

2

0 0.01

100

U

SNP removal

10

Flocs CCGAs

-20 -30

0 0.1 1 10 100 Additional salt concentration in PACl/C15B solution (g/L)

IP T

60

SC R

60

Zeta potential (mV)

80

Half-life time (s)

Efficiency (%)

10 80

10

100

2.0

1.5

1.0

1000

0.1

Size (m)

1

Additional salt concentration in PACl/C15B solution (g/L)

A

CC E

PT

Fig. 3. Performance of CCGAs-flotation at different additional salt concentration in PACl/C15B solution: (a) flotation efficiency versus CCGA stability, (b) zeta potential of flotation flocs and CCGAs, (c) floc size distribution, and (d) fractal dimension (experimental conditions: flow rate of CCGA suspension: 689 mL/min, initial pH of PACl/C15B solution and SNP suspension: ~4.7 and ~9.4, respectively, and concentrations of PACl and C15B in PACl/C15B solution: 1.17 mmol /L as Al and 0.34 mmol/L, respectively).

3.3.Flotation performance influenced by pH of SNP suspension It is necessary to study the influence of initial pH of SNP suspension on the CCGA-flotation performance since the pH condition is closely related to the characteristics and interfacial behaviors of NPs [43]. As displayed in Fig. 4, for the 18

acidic and neutral SNP suspension, the SNP removal efficiency was between 27.8% and 42.6%, respectively. The PACl was almost not consumed at pH 2.8 and 5.2 while ~13.0% of PACl was utilized at pH 7.1. Correspondingly, the zeta potential of flocs grew from +5.3 mV to +17.1 mV with the pH increasing from 2.8 to 7.1. The highest SNP removal and PACl utilization (87.2% and 86.1%, individually) were achieved at

IP T

pH 9.4 of SNP suspension where the zeta potential of flocs dropped to -26.8 mV. This pH condition was very close to the original pH value of SNP suspension, which was

SC R

obviously suitable for the flotation treatment in the present work. The CCGA-flotation was useless for the highly basic SNP suspension (i.e. at pH 10.9) whose pH condition

U

was not suitable for the PACl-involved coagulation [40].

A

N

The average diameters of flotation flocs were approximately 15.7 μm and 37.5 μm at

M

pH 5.2 and 7.1, respectively, with dense and close structure (Fig. 4(c) and (d)). Those SNP flocs in micron scale should have facilitated the effective collision and

ED

attachment between SNPs and bubbles [37]. However, it should also be emphasized that the total capture efficiency of particles in flotation was determined by the stability

PT

efficiency of the bubble-particle aggregate in addition to the collision and attachment efficiency [44]. In the acidic SNP suspensions (pH ≤ 5.2), the particles were neutral or

CC E

weakly negatively charged. After adding the positively surface-charged CCGAs, the aggregates of “CCGAs-SNPs” were not stable since the electrostatic attraction

A

between SNPs and bubbles were not strong enough. Thus, the high flotation efficiency could not be achieved. The Al in the caducous “Al species-SNPs” aggregates was then detected in the flotation effluent so the PACl utilization was found to be extremely low.

19

(a)

100

(b)

SNP removal PACl consumption

20 10

Zeta potential (mV)

Efficiency (%)

80

60

40

0 -10 -20 -30

20

Flocs SNPs

-40

0 4

6

8

10

2

12

4

Initial pH=5.2 Initial pH=7.1 Initial pH=9.5

3.0

2.5

6 4

2.0

1.5

100

1000

Size (m)

1.0

A

10

N

2

1

12

U

Fractal dimension

Volume (%)

(d)

0.1

10

SC R

(c) 10

0 0.01

8

Initial pH of SNP suspension

Initial pH of SNP suspension

8

6

IP T

2

5

6

7

8

9

10

Initial pH of SNP suspension

PT

ED

M

Fig. 4. Performance of CCGA-flotation influencing by initial pH of SNP suspension: (a) SNP removal and PACl consumption, (b) zeta potential, (c) Flc size distribution, and (d) fractal dimension (experimental conditions: flow rate of CCGA suspension: 689 mL/min, initial pH of PACl/C15B solution: ~4.7, additional NaCl concentration in PACl/C15B solution: 0 g/L, and concentrations of PACl and C15B in PACl/C15B solution: 1.17 mmol /L as Al and 0.34 mmol/L, respectively).

CC E

3.4 Comparison SNP removal efficiency between PACl-modified CGAs and PACl alone

To verify the advantage of PACl-modified CGAs in flotation separation of SNPs from

A

water, the ways of using coagulant were explored and compared between modifying coagulant on the CGA surface (Mode I) and directly adding coagulant into water (Mode II).

For both Mode I and II, the highest SNP removal and the PACl consumption were 20

Fractal dimension

obtained at the coagulant dosage of 0.70 mmol/L as Al in the SNP suspension (Fig. 5(a)). As much as 87.2% of SNPs and 86.1% of PACl could be separated by PACl-functionalized CGAs while 31.7% of SNPs and 35.4% of PACl were simultaneously removed during the PACl coagulation. In Fig. 5(b), the zeta potential of flocs induced by Mode I was greater than that of flocs resulting from Mode II. It

IP T

reveals that the charge neutralization capacity of PACl-modified CGAs was stronger than that of PACl alone. Moreover, in terms of zeta potential, it is obvious that the

SC R

optimum removal efficiency was obtained when the suspension was stable. Thus, the predominant mechanisms of SNP removal should include bridging, sweeping or

N

U

adsorption between CCGAs and SNPs as well.

A

The characteristics of flocs induced by different ways of PACl adoption were also investigated in terms of size and fractal dimension (see in Fig. 5(c), (d) and (e)). For

M

Mode I and II, the average diameters of flocs were 0.39 μm and 0.22 μm at the lowest

ED

PACl concentration of 0.55 mmol/L as Al, respectively. The CCGA-induced flocs were no larger than 15 μm for the investigated PACl concentrations; as comparison,

PT

the size of flocs resulting from Mode II kept increasing with the increase of PACl dosage, reaching 81 μm at the highest PACl concentration of 1.50 mmol/L as Al. At

CC E

the PACl dosage of 0.70 mmol/L as Al, the fractal dimension of flocs obtained from Mode I and II was 2.47 and 2.25, individually. The former process created more

A

compact flocs compared with the latter. The floc morphology was further observed by SEM (see the photo in Fig. 5(e)). The surface of flocs induced by PACl-modified CGAs was uniform, smooth and dense whereas that resulting from the direct PACl adoption was coarse and loose. Given to the results of both fractal dimension and SEM, for the optimum efficacy of SNP removal, the flocs retained from Mode I were 21

sphere-like whilst those attained from Mode II were stretched and open. It has been reported that the flocs with close conformation rather than the open structure are more beneficial to the flotation separation of NPs [34].

As illustrated in Fig. 6, in Mode I, most of the Al species were first modified on the

IP T

CGA surface and then used in flotation, by which the coagulant could sufficiently interact with and finally capture the target SNPs. In contrast, when being dosed

SC R

directly into the SNP-polluted wastewater during coagulation, it was probable that the

Al species were not well dispersed in water and thus insufficiently interact with SNPs.

U

The strong floatability of CCGA-induced flocs and the weak settleability of

N

coagulation flocs might also be the possible reasons for the markedly different

A

performance between the two ways of using PACl. In sum, the comparison between

M

Mode I and II confirmed a more robust treatment process of using PACl-modified

SNP removal PACl consumption

(b) 30

PACl PACl

20

PT

60

CC E

Efficiency (%)

80

PACl-modified CGAs PACl-modified CGAs

Zeta potential (mV)

(a) 100

ED

CGAs for SNP separation.

40

20

A

0 0.4

0.6

0.8

PACl-modified CGAs PACl

10 0 -10 -20 -30

1.0

1.2

1.4

1.6

0.4

PACl dosage (mmol/L as Al)

0.6

0.8

1.0

1.2

PACl dosage (mmol/L as Al)

22

1.4

1.6

(d) 12 C(PACl)=0.70 mmol/L as Al C(PACl)=1.50 mmol/L as Al

10

6 4 2 0 0.01

Flocs induced by PACl C(PACl)=0.55 mmol/L as Al C(PACl)=1.20 mmol/L as Al

8 6 4 2

0.1

1

10

100

0 0.01

1000

0.1

10

2.8

2.4 2.2

1000

U

2.0

Flocs induced by PACl-modified CGAs Flocs induced by PACl

0.8

1.0

1.2

1.4

PACl dosage (mmol/L as Al)

1.6

A

0.6

N

1.8 1.6 0.4

100

SC R

2.6 Fractal dimension Df

1

Size (m)

Size (m)

(c) (e)

C(PACl)=0.70 mmol/L as Al C(PACl)=1.50 mmol/L as Al

IP T

Volume (%)

8

Flocs induced by PACl-modified CGAs C(PACl)=0.55 mmol/L as Al C(PACl)=1.20 mmol/L as Al

Volume (%)

(c) 10

PT

ED

M

Fig. 5. Comparison of SNP separation efficacy between PACl-modified CGAs and PACl alone: (a) SNP removal and PACl consumption, (b) zeta potential, (c) and (d) floc size distribution, and (e) floc morphology (other experimental conditions for CCGA-flotation: flow rate of CCGA suspension: 689 mL/min, initial pH of PACl/C15B solution: ~4.7, additional NaCl concentration in PACl/C15B solution: 0 g/L and and initial SNP suspension pH of ~9.4).

Furthermore, the CCGA-flotation was compared with other flotation and/or

CC E

coagulation technologies in SNP removal (Table 2). It is worth noting that the target SNPs in this study were apparently smaller than those investigated in others’

A

work [12,45-47]. For the flotation separation of particles in such a small size scale, colloidal forces and Brownian diffusion predominate the collection efficiency by bubbles [24,48]. The regular coagulation worked less efficiently on the removal of NPs than it did on the separation of larger particles (i.e. 3 μm in diameter) [49]. Moreover, the SNP concentration herein was as high as ~2898 mg/L and the regular CGAs without coagulants might perform poorly even thought the surfactant addition 23

increased. Given to the abovementioned concerns, the surface-modification of

CGAs by the highly efficient coagulant was developed and adopted in the flotation separation of SNPs. The CCGAs were of strong ability in capturing SNPs by colloidal attraction, and meanwhile, the CCGAs in micron scale ensured their high collision and separation efficiency. The CCGA-flotation achieved the high

IP T

removal of ~87.2% though the tiny size of NPs added the difficulties of flotation separation. Such coagulant-functionalized CGAs took great advantages in

SC R

removing SNPs with respect to sound efficacy, less coagulant consumption

[12,45,46] and much shorter retention time [47] compared with the regular way of

A

CC E

PT

ED

M

A

N

U

coagulant adoption.

24

I N U SC R A M

CCGAs

SNPs

CCGAs

CC E

PT

ED

Al species of PACl

C15B

Monomeric Al species

CCGA (PACl-modified CGA) Al13

Amorphous Al species

C15B molecule

CCGA capturing SNPs SNP

A

Fig. 6. Schematic of CCGA formation and CCGA-SNP interaction.

25

I N U SC R

Table 2. Comparison of SNP removal by different coagulation-flotation processes.

55-220

8571

Regular coagulation

CPACl=1.2 mmol/L as Al =0.01 mg Al/mg SNPs; Initial pH=3.8-6.2 CBS12=30 mg/L =0.023 mg/mg SNPs; Pore diameter of gas distributor=100 μm; Qair=800 mL/min

Flotation

A

1307

Optimum conditions

M

106

Technology

ED

42-35000

Average initial CSNPs (mg/L) 3386

21

90.8±4.5

Description of treatment

Reference

Confirming the Al-coagulant being very effective in removing SNPs

[45]

- Using the environmentally friendly BS12; - Directly adding BS12 into the SNP wastewater; - Adopting the regular-decagonal hollow frustum to enhance the foam drainage; - Mixing microbubbles (15.0±0.8 um) with macrobubbles to improve the flotation efficiency.

[46]

CPACl=1.9-2.2 mmol/L as Al =0.006-0.007 mg Al/mg SNPs; CNaOl=5-10 mg/L =0.0006-0.001 mg NaOl/mg SNPs; HRTa=1 h

96.4±3.4 - Conducting coagulation and flotation, successively; - Adding flotation activator (coagulants) and collector (surfactants) together into wastewater for coagulation; - Representing the flotation efficiency in terms of turbidity removal.

[47]

353

Continuous CGA-flotation

CPACl=0 mg Al/mg SNPs; CCTAB=0.32 mmol/L=0.33 mg CTAB/mg SNPs

96.0±2.8

[12]

2898

Batch CCGAflotation

CPACl=0.7 mmol /L as Al =0.006 mg Al/mg SNPs; CC15B=63 mg/L=0.022 mg/mg SNPs; Initial pH=~4.7; Qair=0 mL/min; HRT=3 min.

87.2±1.0

CC E

A

30

SNP removal (%) > 95

Nanobubble flotation

PT

Size of SNPs (nm)

Note: a. HRT – hydraulic retention time. 26

- Removing small SNPs; - Continuous flotation treatment was performed; - High removal efficiency was achieved upon the SNP polluted water at a low concentration. - Removing smaller SNPs than other study did; - No upstream coagulation unit; - Less coagulant and surfactant per SNP weight; - Using the environmentally friendly C15B; - Shorter HRT.

This study

4. Conclusions The robust flotation with the inorganic polymer coagulant (PACl)-functionalized CGAs was adopted to efficiently remove SNPs from water. The optimum SNP removal (~87.2%) and PACl consumption efficiency (~86.1%) were obtained when the PACl/C15B mixed solution and the SNP suspension were used as original in terms

IP T

of pH and ionic strength. The flotation flocs resulting from the PACl-functionalized

CGAs were ~15 μm in diameter with sphere-like morphology and smooth surface. It

SC R

was found that the variation of SNP separation efficiency was consistent with that of bubble stability when the PACl/C15B solution was acidic and neutral. That was

U

caused by the dissociated states of C15B molecules and their interaction with Al

N

species under different pH conditions. The addition of salt into the PACl/C15B mixed

A

solution decreased the foamability of the CCGA generation solution and the bubbles

M

collapsed before they could efficiently capture SNPs in their rising trajectory. In addition, the pH of the SNP suspension affected the performance of CCGA-flotation.

ED

Particularly, modifying PACl on CGAs took greater advantages in effective removing SNPs compared with directly using it as coagulant. The colloidal attraction and

PT

collision efficiency between SNPs and bubbles could be enhanced simultaneously in CCGA-flotation. The PACl was efficiently utilized by sufficiently interacting with the

CC E

target SNPs, and the regular chemical-dosing unit could be omitted from the flotation

A

system.

Acknowledgement This research was supported by the National Science Foundation of China (No. 51608373 and No. 51378373), the Foundation of State Key Laboratory of Pollution Control and Resource Reuse (Tongji University), China PCRRE16015, and the 27

Scientific Starting Foundation of Zhejiang University of Technology (2017129008229).

Appendix

A

CC E

PT

ED

M

A

N

U

SC R

IP T

Supplementary data associated with this article can be found, in the online version, at

28

References [1] S. Choi, M. Johnston, G.S. Wang, C.P. Huang, A seasonal observation on the distribution of engineered nanoparticles in municipal wastewater treatment systems exemplified by TiO2 and ZnO, Sci. Total Environ. 625 (2018). http://adsabs.harvard.edu/abs/2018ScTEn.625.1321C (accessed May 28, 2018).

IP T

[2] Y. Liu, M. Tourbin, S. Lachaize, P. Guiraud, Nanoparticles in wastewaters: Hazards, fate and remediation, Powder Technol. 255 (2014) 149–156.

SC R

doi:10.1016/j.powtec.2013.08.025.

[3] K. Savolainen, L. Pylkkänen, H. Norppa, G. Falck, H. Lindberg, T. Tuomi, M. Vippola, H. Alenius, K. Hämeri, J. Koivisto, Nanotechnologies, engineered

N

U

nanomaterials and occupational health and safety – A review, Saf. Sci. 48 (2010)

A

957–963.

M

[4] W. Fan, R. Peng, X. Li, J. Ren, T. Liu, X. Wang, Effect of titanium dioxide nanoparticles on copper toxicity to Daphnia magna in water: Role of organic

ED

matter, Water Res. 105 (2016) 129–137. doi:10.1016/j.watres.2016.08.060. [5] S.M. Kteeba, H.I. El-Adawi, O.A. El-Rayis, A.E. El-Ghobashy, J.L. Schuld, K.R.

PT

Svoboda, L. Guo, Zinc oxide nanoparticle toxicity in embryonic zebrafish: Mitigation with different natural organic matter, Environ. Pollut. 230 (2017)

CC E

1125–1140. doi:10.1016/j.envpol.2017.07.042.

[6] L.M. Skjolding, S.N. Sørensen, N.B. Hartmann, R. Hjorth, S.F. Hansen, A. Baun,

A

Aquatic ecotoxicity testing of nanoparticles—The quest to disclose nanoparticle effects, Angew. Chem. 55 (2016) 15224–15239.

[7] C. Walden, W. Zhang, Biofilms versus activated sludge: Considerations in metal and metal oxide nanoparticle removal from wastewater, Environ. Sci. Technol. 50 (2016) 8417-8431. 29

[8] R. Guha, B. Xiong, M. Geitner, T. Moore, T.K. Wood, D. Velegol, M. Kumar, Reactive micromixing eliminates fouling and concentration polarization in reverse osmosis membranes, J. Membr. Sci. 542 (2017) 8–17. doi:10.1016/j.memsci.2017.07.044. [9] K.W. Trzaskus, A. Zdeb, W.M. de Vos, A. Kemperman, K. Nijmeijer, Fouling

IP T

behavior during microfiltration of silica nanoparticles and polymeric stabilizers, J. Membr. Sci. 505 (2016) 205–215. doi:10.1016/j.memsci.2016.01.032.

SC R

[10] H. Xu, F. Xiao, D. Wang, Effects of Al2O3 and TiO2 on the coagulation process by Al2(SO4)3 (AS) and poly-aluminum chloride (PACl) in kaolin suspension,

U

Sep. Purif. Technol. 124 (2014) 9–17. doi:10.1016/j.seppur.2014.01.010.

N

[11] M. Zhang, Elimination de nanoparticules par des procédés de flottation, 2015.

A

INSA de Toulouse, France. http://www.theses.fr/2015ISAT0012 (accessed

M

December 26, 2017).

[12] M. Zhang, P. Guiraud, Surface-modified microbubbles (colloidal gas aphrons)

ED

for nanoparticle removal in a continuous bubble generation-flotation separation system, Water Res. 126 (2017) 399–410. doi:10.1016/j.watres.2017.09.051.

PT

[13] J.P.M. Jr, The use of selective and direct DAF for removal of particulate contaminants in drinking water treatment, Water Sci. Technol. 31 (1995) 49–57.

CC E

[14] R.K. Henderson, S.A. Parsons, B. Jefferson, The potential for using bubble modification chemicals in dissolved air flotation for algae removal, Sep. Sci.

A

Technol. 44 (2009) 1923–1940.

[15] R.K. Henderson, S.A. Parsons, B. Jefferson, Surfactants as bubble surface modifiers in the flotation of algae: Dissolved air flotation that utilizes a chemically modified bubble surface, Environ. Sci. Technol. 42 (2008) 4883–4888. doi:10.1021/es702649h. 30

[16] Y. Shi, J. Yang, J. Ma, C. Luo, Feasibility of bubble surface modification for natural organic matter removal from river water using dissolved air flotation, Front. Environ. Sci. Eng. 11 (2017) 10. doi:10.1007/s11783-017-0954-2. [17] Y. Shi, J. Ma, J. Yang, Improved dissolved air flotation performances using chitosan under different dosing schemes, Pol. J. Environ. Stud. 26 (2017)

IP T

2731–2737. doi:10.15244/pjoes/73806.

[18] R.K.L. Yap, M. Whittaker, M. Diao, R.M. Stuetz, B. Jefferson, V. Bulmus, W.L.

SC R

Peirson, A.V. Nguyen, R.K. Henderson, Hydrophobically-associating cationic polymers as micro-bubble surface modifiers in dissolved air flotation for

U

cyanobacteria cell separation, Water Res. 61 (2014) 253–262.

N

doi:10.1016/j.watres.2014.05.032.

A

[19] R.K. Henderson, S.A. Parsons, B. Jefferson, Polymers as bubble surface

M

modifiers in the flotation of algae, Environ. Technol. 31 (2010) 781–790. doi:10.1080/09593331003663302.

ED

[20] A. Azevedo, R. Etchepare, J. Rubio, Raw water clarification by flotation with microbubbles and nanobubbles generated with a multiphase pump, Water Sci.

PT

Technol. 75 (2017) 2342–2349. doi:10.2166/wst.2017.113. [21] R. Perez-Garibay, E. Martinez-Ramos, J. Rubio, Gas dispersion measurements in

CC E

microbubble flotation systems, Miner. Eng. 26 (2012) 34–40. doi:10.1016/j.mineng.2011.10.006.

A

[22] F. Sebba, Microfoams—an unexploited colloid system, J. Colloid Interface Sci. 35 (1971) 643–646. doi:10.1016/0021-9797(71)90223-2.

[23] A.V. Nguyen, P. George, G.J. Jameson, Demonstration of a minimum in the recovery of nanoparticles by flotation: Theory and experiment, Chem. Eng. Sci. 61 (2006) 2494–2509. doi:10.1016/j.ces.2005.11.025. 31

[24] J.A. Ramirez, A. Zinchenko, M. Loewenberg, R.H. Davis, The flotation rates of fine spherical particles under Brownian and convective motion, Chem. Eng. Sci. 54 (1999) 149–157. [25] C.V. Ferber, H. Löwen, Complexes of polyelectrolytes and oppositely charged ionic surfactants, J. Chem. Phys. 118 (2003) 10774–10779.

IP T

doi:10.1063/1.1574782.

[26] T. Iwasaki, M. Ogawa, K. Esumi, K. Meguro, Interactions between betaine-type

SC R

zwitterionic and anionic surfactants in mixed micelles, Langmuir. 7 (1991) 30–35. doi:10.1021/la00049a008.

U

[27] Z. Ye, G. Guo, H. Chen, Z. Shu, Interaction between aqueous solutions of

N

hydrophobically associating polyacrylamide and dodecyl dimethyl betaine, J.

A

Chem. 2014 (2014) 1–8. doi:10.1155/2014/932082.

M

[28] J.L. Lin, C. Huang, C.J.M. Chin, J.R. Pan, The origin of Al(OH)3-rich and Al13-aggregate flocs composition in PACl coagulation, Water Res. 43 (2009)

ED

4285.

[29] W.-L. Chou, C.-T. Wang, W.-C. Chang, S.-Y. Chang, Adsorption treatment of

PT

oxide chemical mechanical polishing wastewater from a semiconductor manufacturing plant by electrocoagulation, J. Hazard. Mater. 180 (2010)

CC E

217–224. doi:10.1016/j.jhazmat.2010.04.017.

[30] H. Tang, Inorganic polymer flocculation theory and flocculants, first, China

A

Architecture and Building Press, China, 2006.

[31] P.G. Chaphalkar, K.T. Valsaraj, D. Roy, A Study of the size distribution and stability of colloidal gas aphrons using a particle size analyzer, Sep. Sci. Technol. 28 (1993) 1287–1302. doi:10.1080/01496399308018037. [32] M. Zhang, L. Xie, Z. Wang, X. Lu, Q. Zhou, Using Fe(III)-coagulant-modified 32

colloidal gas aphrons to remove bio-recalcitrant dissolved organic matter and colorants from cassava distillery wastewater, Bioresour. Technol. 268 (2018) 346–354. doi:10.1016/j.biortech.2018.08.010. [33] Y. Zeng, C. Yang, W. Pu, X. Zhang, Removal of silica from heavy oil

compounds with coagulation, Desalination. 216 (2007) 147–159. doi:10.1016/j.desal.2007.01.005.

IP T

wastewater to be reused in a boiler by combining magnesium and zinc

SC R

[34] M. Zhang, J.-L. Trompette, P. Guiraud, Role of humic acid in enhancing

dissolved air flotation for the removal of TiO2 nanoparticles, Ind. Eng. Chem.

U

Res. 56 (2017) 2212–2220. doi:10.1021/acs.iecr.6b04572.

N

[35] M. Arabloo, M.P. Shahri, M. Zamani, Characterization of colloidal gas

A

aphron-fluids produced from a new plant-based surfactant, J. Dispers. Sci.

M

Technol. 34 (2013) 669–678. doi:10.1080/01932691.2012.683989. [36] O. Zozulya, V. Pletneva, Influence of thermobaric conditions on size distribution

ED

of colloidal gas aphrons, Colloids Surf. Physicochem. Eng. Asp. 483 (2015) 232–238. doi:10.1016/j.colsurfa.2015.05.039.

PT

[37] J.K. Edzwald, Dissolved air flotation and me., Water Res. 44 (2010) 2077-2016. [38] J.Q. Guan, C.H. Tung, G.Z. Li, Modification of alkylbetaine by incorporation of

CC E

hydroxypropyl group: Preparation, surface activity, and biodegradability of N-alkoxy-hydroxypropyl-N, N-dimethyl-glycine betaine surfactants, J. Dispers.

A

Sci. Technol. 19 (1998) 63–76. doi:10.1080/01932699808913161.

[39] Z. Zhang, Z. Wu, G. Liu, Interfacial adsorption of methyl orange in liquid phase of foam fractionation using dodecyl dimethyl betaine as the collector, J. Ind. Eng. Chem. 28 (2015) 184–189. doi:10.1016/j.jiec.2015.01.027. [40] Y. Matsui, N. Shirasaki, T. Yamaguchi, K. Kondo, K. Machida, T. Fukuura, T. 33

Matsushita, Characteristics and components of poly-aluminum chloride coagulants that enhance arsenate removal by coagulation: Detailed analysis of aluminum species, Water Res. 118 (2017) 177–186. doi:10.1016/j.watres.2017.04.037. [41] C. Yang, T. Dabros, D. Li, J. Czarnecki, J.H. Masliyah, Measurement of the zeta

IP T

potential of gas bubbles in aqueous solutions by microelectrophoresis method, J. Colloid Interface Sci. 243 (2001) 128–135. doi:10.1006/jcis.2001.7842.

SC R

[42] S.V. Save, V.G. Pangarkar, Characterization of colloidal gas aphrons, Chem. Eng. Commun. 127 (1994) 35–54.

U

[43] F. Loosli, S. Stoll, Effect of surfactants, pH and water hardness on the surface

N

properties and agglomeration behavior of engineered TiO2 nanoparticles,

A

Environ. Sci.-Nano. 4 (2017) 203–211. doi:10.1039/c6en00339g.

M

[44] Z. Huang, D. Legendre, P. Guiraud, A new experimental method for determining particle capture efficiency in flotation, Chem. Eng. Sci. 66 (2011) 982–997.

ED

doi:10.1016/j.ces.2010.12.006.

[45] W.-H. Kuan, C.-Y. Hu, Chemical evidences for the optimal coagulant dosage

PT

and pH adjustment of silica removal from chemical mechanical polishing (CMP) wastewater, Colloids Surf. Physicochem. Eng. Asp. 342 (2009) 1–7.

CC E

doi:10.1016/j.colsurfa.2009.03.019.

[46] N. Hu, T. Shu, Z. Wu, G. Liu, Z. Li, Y. Zhao, H. Yin, D. Huang, BS12-assisted

A

flotation for the intensification of SNPs separation from CMP wastewater using a novel flotation column, J. Hazard. Mater. 344 (2018) 788–796. doi:10.1016/j.jhazmat.2017.11.028.

[47] J.-C. Tsai, M. Kumar, S.-Y. Chen, J.-G. Lin, Nano-bubble flotation technology with coagulation process for the cost-effective treatment of chemical mechanical 34

polishing wastewater, Sep. Purif. Technol. 58 (2007) 61–67. doi:10.1016/j.seppur.2007.07.022. [48] T. Miettinen, J. Ralston, D. Fornasiero, The limits of fine particle flotation, Miner. Eng. 23 (2010) 420–437. doi:10.1016/j.mineng.2009.12.006. [49] D. Reay, G.A. Ratcliff, Removal of fine particles from water by dispersed air

A

CC E

PT

ED

M

A

N

U

SC R

Chem. Eng. 51 (2010) 178–185. doi:10.1002/cjce.5450510207.

IP T

flotation: Effects of bubble size and particle size on collection efficiency, Can. J.

35