Accepted Manuscript Hydrophobically-associating cationic polymers as micro-bubble surface modifiers in dissolved air flotation for cyanobacteria cell separation R.K. Yap , M. Whittaker , M. Diao , R.M. Stuetz , B. Jefferson , V. Bulmuş , W.L. Peirson , A. Nguyen , R.K. Henderson PII:

S0043-1354(14)00385-6

DOI:

10.1016/j.watres.2014.05.032

Reference:

WR 10686

To appear in:

Water Research

Received Date: 17 February 2014 Revised Date:

15 May 2014

Accepted Date: 18 May 2014

Please cite this article as: Yap, R.K., Whittaker, M., Diao, M., Stuetz, R.M., Jefferson, B., Bulmuş, V., Peirson, W.L., Nguyen, A., Henderson, R.K., Hydrophobically-associating cationic polymers as microbubble surface modifiers in dissolved air flotation for cyanobacteria cell separation, Water Research (2014), doi: 10.1016/j.watres.2014.05.032. 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.

Bubble

Anionic charge

AC C

Cyanobacterial cell

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Cationic charge

Polymer

Hydrophobic group

ACCEPTED MANUSCRIPT

Hydrophobically-associating cationic polymers as micro-bubble surface

2

modifiers in dissolved air flotation for cyanobacteria cell separation

3

Yap, R.K.,1,2 Whittaker, M.,3 Diao, M.,4 Stuetz, R.M.,1 Jefferson, B.,5 Bulmuş, V.,6 Peirson,

4

W.L.,7 Nguyen, A.,4 and Henderson, R.K.8,*

5

1

6

University of New South Wales, Sydney NSW 2052, Australia.

7

2

8

University of New South Wales, Sydney NSW 2052, Australia.

9

3

UNSW Water Research Centre, School of Civil and Environmental Engineering, The

SC

Centre for Advanced Macromolecular Design, School of Chemical Engineering, The

M AN U

Monash Institute of Pharmaceutical Sciences (MIPS), Monash University, IVC 3052,

10

Australia.

11

4

12

Australia

13

5

14

Bedfordshire, MK43 0AL, UK.

15

6

16

Turkey.

17

7

18

of New South Wales, Manly Vale NSW 2093, Australia.

19

8

20

Australia

21

*Corresponding author: [email protected]

School of Chemical Engineering, The University of Queensland, Brisbane, Queensland 4072,

TE D

Cranfield Water Science Institute, School of Applied Sciences, Cranfield University,

EP

Department of Chemical Engineering, Izmir Institute of Technology, Urla 35430 Izmir,

AC C

Water Research Laboratory, School of Civil and Environmental Engineering, The University

School of Chemical Engineering, The University of New South Wales, Sydney, NSW 2052,

22 23

RI PT

1

ABSTRACT

ACCEPTED MANUSCRIPT Dissolved

air

flotation

(DAF),

an

effective

treatment

method

for

clarifying

25

algae/cyanobacteria laden water, is highly dependent on coagulation-flocculation. Treatment

26

of algae can be problematic due to unpredictable coagulant demand during algae blooms. To

27

eliminate the need for coagulation-flocculation, the use of commercial polymers or

28

surfactants to alter bubble charge in DAF has shown potential, termed the PosiDAF process.

29

When using surfactants, poor removal was obtained but good bubble adherence was

30

observed. Conversely, when using polymers, effective cell removal was obtained, attributed

31

to polymer bridging, but polymers did not adhere well to the bubble surface, resulting in a

32

cationic clarified effluent that was indicative of high polymer concentrations. In order to

33

combine the attributes of both polymers (bridging ability) and surfactants (hydrophobicity),

34

in this study, a commercially-available cationic polymer, poly(dimethylaminoethyl

35

methacrylate) (polyDMAEMA), was functionalised with hydrophobic pendant groups of

36

various carbon chain lengths to improve adherence of polymer to a bubble surface. Its

37

performance in PosiDAF was contrasted against commercially-available poly(diallyl

38

dimethyl ammonium chloride) (polyDADMAC). All synthesised polymers used for bubble

39

surface modification were found to produce positively charged bubbles. When applying

40

these cationic micro-bubbles in PosiDAF, in the absence of coagulation-flocculation, cell

41

removals in excess of 90% were obtained, reaching a maximum of 99% cell removal, thus

42

demonstrating process viability. Of the synthesised polymers, the polymer containing the

43

largest hydrophobic functionality resulted in highly anionic treated effluent, suggesting

44

stronger adherence of polymers to bubble surfaces and reduced residual polymer

45

concentrations.

AC C

EP

TE D

M AN U

SC

RI PT

24

46 47

Keywords: Algae separation; cationic bubbles; cyanobacteria; flotation; PosiDAF;

48

water soluble polymers; water treatment.

ACCEPTED MANUSCRIPT 49

1. INTRODUCTION

51

Dissolved air flotation (DAF) is a solid-liquid separation process in which nucleated

52

microbubbles are introduced to a suspension comprising flocculated particles. Collision and

53

attachment of bubbles and particles create low density bubble-particle agglomerates which

54

rise to the surface to form a float layer and can then be removed mechanically or

55

hydraulically. In water and wastewater treatment plants (WTPs/WWTPs), DAF is used for

56

the removal of low density contaminants such as algae and natural organic matter (NOM)

57

from reservoir water or waste stabilisation ponds (WSPs).

58

conventionally applied to reduce particle and colloid charge, increase particle sizes, complex

59

with NOM and ensure bubble-particle interactions and subsequent removal efficiencies are

60

optimal (Edzwald 2010).

M AN U

SC

RI PT

50

TE D

61

Coagulation-flocculation is

The presence of algae and cyanobacteria in raw water can present a significant challenge for

63

WTP/WWTP operators and DAF is becoming a popular process option to improve

64

treatability (Edzwald 2010, Teixeira and Rosa 2006). However, coagulation-flocculation

65

remains difficult to optimise due to highly variable population densities, morphologies and

66

cell motility as well as interferences by algogenic organic matter (AOM) and, consequently,

67

flotation can be rendered ineffective (Henderson et al. 2010a, Pieterse and Cloot 1997). This

68

can lead to a number of downstream problems such as turbidity breakthrough, filter clogging

69

(Buisine and Oemcke 2003), the presence of toxins (Al-Tebrineh et al. 2010) and the

70

formation of harmful disinfection by-products on disinfection (Chen et al. 2008). Hence, to

71

avoid treatment problems during algal and cyanobacterial blooms, further optimisation of the

72

DAF process is required.

AC C

EP

62

ACCEPTED MANUSCRIPT 73 Similar to influent particles and colloids, DAF microbubbles are negatively charged, likely

75

due to asymmetric dipoles of water molecules at bubble gas liquid interfaces (Oliveira and

76

Rubio 2011). The manipulation of the bubble surface charge, as opposed to that of particles,

77

has received attention as an alternative to coagulation-flocculation (Han et al. 2006,

78

Henderson et al. 2009). Specifically, controlling the bubble surface charge in DAF has been

79

investigated via two methods: 1) Altering the ion content or pH of water in which bubble are

80

introduced (Han et al. 2006), or 2) by using a chemical additive dosed into the air saturated

81

water stream (Henderson et al. 2008c, 2009, Karhu et al. 2014, Malley 1995, Oliveira and

82

Rubio 2012). For example, Karhu et al. (2014) recently demonstrated the use of modified-

83

bubbles for the treatment of oil-in-water emulsions using cetyl trimethylammonium bromide

84

(CTAB), poly(diallyldimethyl ammonium chloride) (polyDADMAC) and epichlorohydrin–

85

dimethylamine copolymer (Epi-DMA). In this case, CTAB was found to perform poorly in

86

water treatment; however, using polyDADMAC and Epi-DMA as bubble modifiers resulted

87

in >99% removal of hydrophobic particles. In the treatment of algae- and cyanobacteria-

88

laden water, Henderson et al. (2008c, 2010b) used a range of surfactants and water treatment

89

polymers to modify bubble surfaces and subsequently float unflocculated cells by dosing

90

chemicals into the recycle stream. When using surfactants (Henderson et al. 2008c), it was

91

found that cell removal for a range of species matched modelled data, reaching a maximum

92

of 64% removal of Microcystis aeruginosa.

93

(Henderson et al. 2010b), cell removals for the same M. aeruginosa strain reached 98 %,

94

indicative of process enhancement via polymer bridging. However, this was not achieved

95

with other species, attributed to competing AOM-polymer and polymer-bubble interactions.

96

Moreover, the positive zeta potential in the treated water was suggestive of high polymer

97

concentrations which are also undesirable.

AC C

EP

TE D

M AN U

SC

RI PT

74

Interestingly, when using polyDADMAC

ACCEPTED MANUSCRIPT 98 99

It has been suggested that more robust flotation of cells may be possible by combining the

100

attributes of both the surfactants and polymers to facilitate greater adherence to bubbles,

101

achieved by incorporating hydrophobic components in a cationic polymer (Henderson et al.

102

2010b).

103

molecular weight, branched water soluble polymers with no groups that would be

104

conventionally identified as hydrophobically functional (Bolto and Gregory 2007). Hence, as

105

far as the authors are aware, there have not been any polymers designed to adhere to bubble

106

surfaces in DAF.

107

specifically designed hydrophobically-associating cationic polymers, in comparison with

108

commercially available polyDADMAC, for the alteration of bubble surface properties in

109

DAF – a process termed ‘PosiDAF’.

SC

RI PT

The development of water treatment polymers has generally targeted higher

110

M AN U

This research therefore investigates the application of a range of

The aim of this paper was to investigate the impact of increasing polymer hydrophobic

112

functionality on the efficacy of bubble coating and link results with the presence of polymer

113

residuals in PosiDAF treated effluent.

114

associated algogenic organic matter (AOM) were used as model contaminants. The cell

115

separation obtained using conventional coagulation-flocculation and DAF was also assessed

116

for comparison.

117

surfaces was assessed and the mechanisms of interaction between the bubbles, functionalised

118

polymers, cells and AOM discussed.

TE D

111

EP

AC C

The optimal polymer functionalisation for the modification of bubble

119

120 121

To achieve this aim, M. aeruginosa cells and

2. MATERIALS AND METHODS

ACCEPTED MANUSCRIPT 2.1.

Hydrophobically functionalised polymers

123

Homopolymers of 2-(N,N-dimethylamino)ethyl methacrylate (DMAEMA) (Aldrich,

124

Australia) were first synthesised as a cationic backbone, controlling the polymer molecular

125

weight by varying the concentration of free radical initiator, azobisisobutyronitrile (AIBN), in

126

a classical free radical polymerisation. DMAEMA was selected as it can be polymerised

127

under mild conditions, produces linear polymers and can be easily functionalised. Three base

128

polymers of high, medium and low molecular weights were synthesised and further

129

functionalised by quaternising the tertiary amines with iodomethane, 1-bromopentane, 1-

130

bromodecane or 1-bromopentadecane at a range of concentrations. This resulted in increased

131

cationic charge and associated hydrophobic pendant groups. The resulting 36 synthesised

132

polymers were named according to their molecular weight (L, M and H for low medium and

133

high molecular weight, respectively), the hydrocarbon chain length of quaternising alkyl

134

halide (C1, C5, C10 and C15, indicating number of carbons) and the concentration of

135

alkylhalide used in the quaternisation reaction (l, m and h for low (10%), medium (50%) and

136

high (75%) conversions, respectively). For example, a low molecular weight polymer with a

137

high concentration of 1-bromopentane was designated LC5-h.

138

Analysis of the polymers included charge density using a PCD-04 Travel Charge Demand

139

Analyser (BTG, Switzerland) and surface tension using a NIMA Surface Tensiometer

140

equipped with a du Noüy ring (Biolin Scientific, Sweden). In this study, nine of the 36

141

functionalised polymers were selected for investigation to include low, median and high

142

surface tensions in each molecular weight range (Table 1). Pictorial representations can be

143

found in the Table S1.

AC C

EP

TE D

M AN U

SC

RI PT

122

144 145

2.2.

Commercially available chemicals

ACCEPTED MANUSCRIPT 146

Cetyl trimethylammonium bromide (CTAB) (Sigma Aldrich, Australia) and low molecular

147

weight (MWw 100-200 kDa) poly(diallyldimethylammonium chloride) (polyDADMAC)

148

(Sigma Aldrich, Australia) were used as standard commercially available chemicals to

149

compare the performance of the synthesised polymers.

150

flocculation-DAF experiments, aluminium sulphate (Sigma Aldrich, Australia) was used as a

151

coagulant.

RI PT

In conventional coagulation-

152 2.3.

Cyanobacteria

154

M. aeruginosa (CS-564/01) was obtained from the Commonwealth Scientific and Industrial

155

Research Organisation (CSIRO) Australian National Algae Culture Collection, Hobart,

156

Australia, and recultured in MLA media (Bolch and Blackburn 1996).

157

subjected to a 16/8 hour light/dark cycle at 21°C, in a 500 L, PG50 incubator with a

158

photosynthetic photon flux output of 600 ± 60 µmol m-2s-1 (Labec, Australia). Cultures were

159

grown in 100 mL batches in 250 mL conical flasks and agitated frequently to ensure

160

homogeneity of the cultures. Cells were harvested at the end of the exponential growth

161

phase, as determined by cell counting via a Leica DM500 light microscope (Leica

162

Microsystems Ltd, Switzerland) and a haemocytometer. An example of a growth curve can

163

be found in the Supplementary Information (Figure S1). Cell size was measured using a

164

Mastersizer 2000 (Malvern, UK), charge demand with a Mütek PCD-04 particle charge

165

detector (BTG, Switzerland), zeta potential using a Zetasizer Nano ZS (Malvern, UK;

166

specified zeta potential measurement range of 3.8nm – 100 µm) and AOM concentration

167

using a TOC–VCSH Analyser (Shimadzu, Australia).

EP

AC C

168 169

Cultures were

TE D

M AN U

SC

153

2.4.

Conventional flotation jar testing

ACCEPTED MANUSCRIPT A DAF Batch Tester, Model DBT6 (EC Engineering, Canada), was used for conventional

171

flotation experiments incorporating the coagulation-flocculation process. Cultured cells were

172

diluted to 7.5 × 105 cells mL-1 with Milli-Q water buffered with 0.5 mM NaHCO3 and

173

brought to an ionic strength of 1.8 mM using NaCl to facilitate comparability with previous

174

studies (Henderson et al. 2008c, 2009, 2010b). The saturated water consisted of Milli-Q

175

water also containing 0.5 mM NaHCO3 and made up to an ionic strength of 1.8 mM with

176

NaCl adjusted to pH 7. Industrial grade air was used to pressurise the saturator to 450 kPa.

177

Coagulation was performed by adding aluminium sulphate to the jar and rapidly mixing for

178

180 seconds at 200 rpm. Immediately following the addition of coagulant, the pH was

179

adjusted to pH 7 using 1 M HCl and 1 M NaOH solutions. A pH210 Microprocessor pH

180

Meter (Hanna Instruments, USA) was used to monitor the pH during rapid mixing. The

181

samples were then flocculated for 10 minutes at 30 rpm followed by flotation for 10 minutes

182

with an equivalent recycle ratio of 10%, as per typical DAF operation (Edzwald 2010).

183

Treated water analysis included cyanobacteria cell concentration achieved by cell counting

184

and zeta potential analysis (as described in Section 2.3). Each analysis was conducted in

185

triplicate.

EP

186

TE D

M AN U

SC

RI PT

170

2.5.

Bubble charge measurements

188

Bubble surface charge was measured to determine whether the polymers altered the surface

189

properties of the bubbles.

190

Engineering at the University of Queensland. A Microelectrophoresis Apparatus Mk II

191

(Rank Brothers Ltd., UK), consisting of a rectangular cell (10 mm × 1 mm) and platinum

192

electrodes was used in the measurement of microbubbles. The generation of microbubbles

193

and measurement was adapted from that described by Qu et al. (2009). Specifically, nitrogen

194

was dissolved into a Milli-Q solution comprising 0.5 mM NaHCO3 and made up to an ionic

AC C

187

Measurements were undertaken by the School of Chemical

ACCEPTED MANUSCRIPT strength of 1.8 mM with NaCl, corrected to pH 7 with 1.6 mg L-1 of polymer, at 450 kPa by

196

leaving overnight. The polymer concentration was based on saturator concentrations applied

197

in PosiDAF jar testing. About 100 mL of the oversaturated solution was introduced to the

198

glass cell of the microelectrophoretic unit. At room pressure, micro-bubbles that formed

199

were subject to an electrical field of 40 V/m. The motion of the bubbles were then recorded

200

with a CCD camera and their electrophoretic mobilities and zeta potentials were calculated

201

using the von Smoluchowski equation (Hunter 1981), for which 60 to 100 bubble

202

measurements were obtained per polymer tested.

SC

RI PT

195

204

2.6.

205

The same DAF Batch Tester, Model DBT6 (EC Engineering, Canada), and cell suspensions

206

were used for PosiDAF jar testing as in Section 2.4. The recycle water composition was also

207

the same as that described in Section 2.4; however, the various polymers at a range of

208

concentrations up to 3 mg L-1 were now added to the buffered solution. Industrial grade air

209

was again used to pressurise the saturator to 450 kPa but, in these experiments, an equivalent

210

recycle ratio of 20% was applied to ensure a high bubble to particle ratio was maintained

211

given that coagulation-flocculation would not lower particle numbers.

212

conducted for 10 minutes prior to residual sampling without any coagulation-flocculation.

213

Treated water analysis was undertaken as previously described for conventional flotation

214

experiments. Zeta potential measurements were anticipated to give an indication of the

215

presence of polymer in the treated water, as any residual cationic polymer will either complex

216

with or adsorb onto oppositely charged colloids/particles or remain free in solution. Note that

217

the hydrodynamic diameters of the polymers in solution were found to be 12.2 to 1130 nm

218

(Table S2), greater than the 3.8 nm zeta potential measurement limit of the instrument.

Flotation was

AC C

EP

TE D

PosiDAF Jar testing

M AN U

203

ACCEPTED MANUSCRIPT 219

Overall, the presence of cations is expected to reduce the magnitude of the measured negative

220

charge in the treated water.

222

3. RESULTS

223 224

3.1.

225

On microscopic evaluation of M. aeruginosa cultures harvested at the end of the exponential

226

growth phase, the cells were found to be spherical and unicellular with an average diameter

227

of 3.0 ± 0.7 µm (Table 2). The charge density and zeta potential of the cell culture were

228

determined to be -1.51 × 10-9 ± 7 × 10-11 meq cell-1 and -31.6 ± 1.6 mV, respectively.

229

Average cell concentration and concentration of AOM at this phase of growth was 2.1 × 107

230

± 2 × 106 cells mL-1 and 8.04 × 10-10 ± 4.4 × 10-11 mg C cell-1, respectively (Table 2).

231

Notably, Henderson et al. (2008b) found that cell sizes obtained for the UK strain of M.

232

aeruginosa (CCAP 1450/3) used in PosiDAF research were larger (5.4 µm) and had a much

233

lower charge density per cell of 2.0 × 10-12 meq cell-1. If it is assumed that all AOM was

234

associated at the cell surfaces, this indicates that the Australian strain (CS-564/01) had a

235

much more negative cell surface charge density at -57 meq m-2 compared to -0.04 meq m-2

236

for CCAP 1450/3 (Henderson et al. 2008b). Similar to the surface charge density, the zeta

237

potential of the Australian strain was also more negative.

AC C

EP

TE D

M AN U

SC

Microcystis aeruginosa Characterisation

RI PT

221

238 239

3.2.

Cell Removal with Conventional DAF

240

Conventional DAF with coagulation-flocculation pre-treatment upstream of flotation resulted

241

in high cell removal efficiencies that were dependent on effective coagulation (Figure 1). For

ACCEPTED MANUSCRIPT example, it was observed that a dose of 1 mg L-1 as Al (or 11.1 mg L-1 Al2(SO4)3•14H2O) was

243

required to achieve cell removals greater than 95%, coinciding with a lowering of the

244

magnitude of the zeta potential. At a dose of 5 mg L-1, a maximum cell removal of 99% was

245

obtained with large flocs observed (Figure 2).

RI PT

242

246 3.3.

248

Bubbles coated with hydrophobically functionalised polyDMAEMA were confirmed to be

249

cationic at pH 7, with zeta potentials ranging from between +38.6 mV to +63.8 mV. These

250

values were comparable with the charge of bubbles modified with polyDMAEMA

251

homopolymer, polyDADMAC and CTAB of +39 ± 10 mV, +44 ± 9 mV and +44 ± 7 mV,

252

respectively (Figure 3). It was found that zeta potentials observed in the current study were

253

consistently more positive than those observed in prior work. For example, Cho et al. (2005)

254

used a range of cationic surfactants to modify nanobubbles, resulting in bubbles with a

255

maximum zeta potential of +30 mV at pH 7. Similarly, Han et al. (2006) generated bubbles

256

with a zeta potential of +30 mV using aluminium hydroxide, although high standard

257

deviations were observed.

M AN U

TE D

EP AC C

258

Modified Bubble Properties

SC

247

259

Overall, bubble zeta potential did not vary to the same extent as polymer charge density. It

260

was observed that the modification of bubbles with polymers quaternised with a high

261

concentration of C10 groups resulted in bubbles with less positive zeta potentials for each of

262

the molecular weight ranges (specifically, polymers LC10-h, MC10-h and HC10-h). It was

263

found that LC10-h had a statistically different zeta potential compared to both LC5-l and

264

LC5-m (P-value 10%) formed gels and thus could not be analysed in

271

solution.

RI PT

Bubbles modified with LC10-h, MC10-h and HC10-h had more negative

SC

272

On examining the average zeta potential for each of the molecular weight groups, it was

274

revealed that lower molecular weight polymers resulted in more positive bubble zeta

275

potentials compared to the medium and high molecular weight polymers (P-values 0.016 and

276

0.025, respectively). For example, the average bubble zeta potentials for each of the polymer

277

molecular weight groups (low, medium and high) were +60 ± 13 mV, +47 ± 10 mV and +42

278

± 12 mV, respectively.

TE D

279

M AN U

273

3.4.

Cell Removal Using PosiDAF

281

Results from jar tests that were conducted using the nine synthesised polymers to modify

282

bubble surfaces demonstrated that cell removals in excess of 93% were achievable for each

283

polymer tested when applying doses of 0.3 mg L-1, without coagulation (Figure 4A). With a

284

maximum cell removal of 99%, using 0.3 mg L-1 of polyDADMAC, PosiDAF has the same

285

cell removal efficiency as conventional DAF (Figure 1). The resultant dose response curves

286

for all polymers are presented in the Supplementary Information (Table S3). Overall, the

287

dose response curves increased to greater than 90% cell removal, with no decrease of cell

288

removal observed even at the highest doses of polymer used in this test. All dose response

289

curves were similar despite variations of polymer charge densities and associated

AC C

EP

280

ACCEPTED MANUSCRIPT hydrophobic groups as indicated in Table 1. Comparing the polymers in terms of their charge

291

dose demonstrated similarity in the removal efficiencies for polyDADMAC and synthesised

292

polyDMAEMA samples with medium and high concentrations of quaternised residuals,

293

specifically LC5-m, LC10-h, MC1-m, MC10-h, HC1-m and HC10-h (select examples in

294

Figure 4B and full dataset in Table S3). However, polymers with low concentrations of

295

quaternised residuals (LC5-l, MC5-l and HC15-l) resulted in greater removal efficiencies at

296

low charge concentrations. As an example, the charge dose response curve for LC5-l is

297

compared to the curves for LC5-m, HC10-h, polyDADMAC and CTAB in Figure 4B. It can

298

be seen that for a polymer dose of approximately 0.3 × 10-3 meq L-1, cell removal achieved

299

by PosiDAF with LC5-m and HC10-h was 66 ± 6%, whereas with LC5-l, 97 ± 4% was

300

achieved. This indicates that polymer bridging is a dominant mechanism and elevated charge

301

may not be necessary in establishing polymer-particle attachments. However, it is known

302

that low charge density polymers have different neutralisation effects in water: Kam and

303

Gregory (2001) showed that polymers with a charge density of greater than 3 meq g-1

304

exhibited a stoichiometric neutralisation of anionic humic substances, where 1 meq of

305

polymer could neutralise 1 meq of humic substances. This was not true for polymers with

306

charge densities less than 3 meq g-1, where less than 1 meq of polymer could neutralise 1 meq

307

of humic substances.

SC

M AN U

TE D

EP

AC C

308

RI PT

290

309

Similar to the PosiDAF jar tests undertaken with the hydrophobically functionalised polymer,

310

jar tests conducted with polyDADMAC revealed that cell removal was again high, with up to

311

99% removal achieved at doses above 1.0 × 10-3 meq L-1.

312

aeruginosa (CCAP 1450/3), Henderson et al. (2009) found that 95% removal could be

313

achieved at a dose of 2.4 × 10

314

dependent. To demonstrate this, the dose was normalised to charge dose per cell charge. In

-3

Using a UK strain of M.

meq L-1, indicating the required polymer dose was strain

ACCEPTED MANUSCRIPT this work, the optimal dose was found to be 0.9 meq polyDADMAC per meq M. aeruginosa

316

whereas Henderson et al. (2009) reported the optimal dose to be 1.7 meq polyDADMAC per

317

meq M. aeruginosa for CCAP 1450/3. A major difference between these strains was the cell

318

size, where CS-564/01 cells were nearly half the diameter of CCAP 1450/3 cells (3.0 µm

319

versus 5.4 µm, respectively). Considering this, the relative dose per cell surface area is 67%

320

greater for CS-564/01 than for CCAP 1450/3 at the same cell concentration. In addition, for

321

the same comparison, the charge density was also found to be significantly greater for CS-

322

564/01 (Henderson et al. 2010a). Differences of this scale between species have previously

323

been reported (Henderson et al. 2008a, Henderson et al. 2010a); the observation that such

324

differences in charge density can also occur between different strains of the same species is

325

an important consideration.

326

M AN U

SC

RI PT

315

The highest cell removal obtained using CTAB as the bubble modifier (33 ± 7% with a dose

328

of 1.18 × 10-3 meq L-1, Figure 4) was found to be much lower than that obtained for polymers

329

and obtained by Henderson et al. (2008c) (64% with a dose of 2.2 × 10-3 meq L-1). However,

330

both the results obtained in this study and by Henderson et al. (2008c) are comparable with

331

modeled results obtained using the white water model performance equation (Haarhoff and

332

Edzwald 2004) (Equation S1). For example, assuming 100% attachment efficiency, it was

333

determined that the modeled cell removal was 30% for a particle size of 3.0 µm, and 64% for

334

a particle size of 5.4 µm.

EP

AC C

335

TE D

327

336

Interestingly, unlike observations made by Henderson et al. (2009), large bubble-cell

337

networks were observed to develop during the 10 minute flotation period of CS-564 (Figure

338

5A), creating a stable, cell-rich float (Figure 5B). These networks formed rapidly after the

ACCEPTED MANUSCRIPT introduction of the recycle stream to become clearly visible to the naked eye and may be the

340

result of extended bridging which has in turn aided cell removal. The formation of such

341

structures may be attributed to large biopolymers in AOM present in algae and cyanobacteria

342

systems (Henderson et al. 2010a) which can influence the action of the polymers PosiDAF.

343

For example, it has been demonstrated that polymers preferentially interact with AOM over

344

cells (Haarhoff and Cleasby 1989) and dissolved matter over other particles (Lurie and

345

Rebhun 1997). Furthermore, in the case of CCAP1450/3 (Henderson et al. 2010b), polymer-

346

AOM interactions were suggested to create favourable flotation conditions for cells.

SC

RI PT

339

M AN U

347 3.5.

Charge Measured in the PosiDAF Treated water

349

The charge of particles or dispersed/dissolved macromolecules measured in the treated water

350

after jar tests using the nine synthesised polymers was highly variable, ranging from -44.6

351

mV to +12.7 mV. A full set of these results can be found in the Supplementary Information

352

(Table S3). Zeta potential dose response curves for LC5-l, LC5-m and HC10-h are displayed

353

as an example of this and compared to CTAB and polyDADMAC in Figure 6. For all

354

polymers, increasing the polymer dose resulted in a decrease in the magnitude of the negative

355

zeta potentials in the treated water to some degree (Table S3). Of the polymers tested, those

356

modified with high concentrations of highly hydrophobic groups, specifically, LC10-h,

357

MC10-h and HC10-h, resulted in the most negative resultant zeta potentials upon jar testing

358

and in fact retained negative zeta potentials over the range of doses applied. Conversely,

359

polymers with the fewest quaternised groups resulted in positive resultant zeta potentials at

360

lower charge doses than the other polymers (e.g. LC5-l in Figure 6). With highly cationic

361

polymers, such as polyDADMAC and LC5-m, zeta potentials became positive with

362

increasing polymer dose. After a jar test, the remaining cells, AOM and polymer in the

363

treated water contribute to the zeta potential. As charge measurements can be used to

AC C

EP

TE D

348

ACCEPTED MANUSCRIPT determine polymer or colloid concentrations (Kam and Gregory 1999), a positive zeta

365

potential is indicative of relative residual polymer concentration when contrasted with results

366

of other polymers with similar cell removals. It was found that higher molecular weight

367

polymers resulted in negative zeta potentials at much greater doses than that needed for

368

optimal cell removal (Table S3). Similar observations have been made before when using

369

PosiDAF (Henderson et al. 2010b) and for conventional DAF (Gehr and Henry 1982).

370

CTAB demonstrated highly anionic zeta potentials at all doses in this study, which can be

371

attributed to its ability to congregate at air-water interfaces. Synthesised polymers with the

372

highest degree of functionalization, LC10-h and MC10-h and HC10-h, displayed the most

373

negative zeta potentials in PosiDAF treated effluent (Table S3). In particular, the zeta

374

potentials obtained from tests using HC10-h were most negative and very similar to that of

375

CTAB, suggesting the strongest adhesion to the bubble surface was achieved when using this

376

polymer.

377

4. DISCUSSION

379

EP

378

TE D

M AN U

SC

RI PT

364

4.1.

Bubble Coating with Hydrophobically Functionalised Polymers

381

The surfaces of microbubbles in water have been found to be negatively charged under a

382

range of pH conditions (Elmallidy et al. 2008, Han et al. 2006). At pH 7, the zeta potential of

383

a micro-bubble in water is approximately -25 to -60 mV (Oliveira and Rubio 2011, Yang et

384

al. 2001), though the value can be altered depending on background ionic conditions

385

(Oliveira and Rubio 2011, Yang et al. 2001).

386

polyDADMAC resulted in positive bubbles with little variation in the magnitude of their zeta

387

potentials at pH 7. This is in agreement with literature for CTAB (Yoon and Yordan 1986);

AC C

380

Bubbles modified with CTAB and

ACCEPTED MANUSCRIPT however, the study of the effect of various cationic polymers is limited to commercially

389

available polymers (Oliveira and Rubio 2011). With all synthesised polymers used in this

390

work, positively charged bubbles were generated; however, the bubble zeta potential was

391

observed to shift to less positive values with increases in a) polymer hydrophobic

392

functionality and b) polymer molecular weight (Figure 3).

393

With respect to hydrophobic functionality in synthesised polymers, the association of

394

hydrophobic pendant groups to cationic polyDMAEMA was expected to enhance

395

electrostatic polymer interaction at the bubble surfaces by hydrophobic association (Bütün et

396

al. 2001). Counterintuitively, the polymers with high concentrations of large hydrophobic

397

groups (LC10-h, MC10-h and HC10-h) were observed to have a less positive bubble zeta

398

potential in comparison to polymers of similar molecular weight, despite these polymers

399

having greater charge densities than the other DMAEMA-based polymers (Table 1).

400

Assuming that this was the result of lower charge concentrations at the bubble surface, it is

401

suggested that the polymer attached to the bubble surface in a flatter conformation than other

402

polymers due to increased hydrophobicity, similar to the adsorption of hydrophobic polymers

403

onto hydrophobic surfaces (Jamadagni et al. 2009). This would result in any given unimer

404

occupying a larger surface area therefore limiting the attachment of other unimers via steric

405

interactions.

406

With respect to polymer molecular weight, the shift in bubble zeta potential to less positive

407

values was surprising as Aoki and Adachi (2006) had observed that the electrophoretic

408

mobility of polystyrene latex particles with adsorbed fully quaternised polyDMAEMA was

409

constant, regardless of the polymer molecular weight. However, the adsorption of polymers

410

onto bubbles may be inhibited not only by steric interferences in the case of high molecular

411

weight polymers, but also by electrostatic repulsion in areas where cationic unimers have

412

adsorbed to form areas of high charge concentration. It is anticipated that the extension of a

AC C

EP

TE D

M AN U

SC

RI PT

388

ACCEPTED MANUSCRIPT 413

low molecular weight polymer from the surface of a bubble will be less than that of high

414

molecular weight polymers (Henderson et al. 2010b, Napper 1983), therefore it may be that

415

less steric interaction would be encountered and more polymers (and charge) could occupy

416

the same area on a microbubble surface.

417

complications can arise in modified-bubble measurements due to differing polymer

418

concentrations on bubble surfaces in a single system; for example, Oliveira and Rubio

419

(2011) observed that randomly measured bubbles do not carry the same charge in a given

420

system.

421

With respect to the toxicity of polymers in water treatment, it is generally accepted that the

422

polymers with greater cationicity are more toxic (Bolto and Gregory 2007). The inclusion of

423

hydrophobic groups to a cationic polymer can increase antibacterial activity at high

424

concentrations (van de Wetering et al. 2000); however, it has been long recognised that

425

polymers with surfactant like residuals are much less toxic than surfactants (Schmolka 1977).

426

For surfactants, surface activity measurements can be related to packing density of the

427

molecule at the air-water interface (Henderson et al. 2008c, Rosen and Milton 1978). The

428

positive bubble zeta potential measurements observed in this study (Figure 3) and in other

429

studies (Malley 1995, Oliveira and Rubio 2011) indicate polymer adsorption at the bubble

430

surface. However, highly cationic polymers (specifically LC10-h and HC10-h) had surface

431

tensions similar to water at 72 mN m-1 (Table 1) despite the hydrophobic functionality.

432

Similarly, in an investigation of polymers with low polydispersity at an air-water interface,

433

Matsuoka et al. (2004) also found that charged polymers exhibited “non-surface activity”

434

despite the presence of hydrophobic groups. Polymer accumulation at the air-liquid interface

435

is theorised to occur as hydrogen bonding with OH- or H+ groups (Yang et al. 2001), which

436

translates to air-liquid surface adsorption without surface penetration.

437

confirmed with X-ray reflectance, demonstrating that both hydrophilic-hydrophobic random

AC C

EP

TE D

M AN U

SC

RI PT

It is acknowledged, however, that further

This has been

ACCEPTED MANUSCRIPT copolymers and hydrophilic homopolymers did adsorb at the air-liquid interface in solution,

439

despite demonstrating no change in surface tension over a large concentration range

440

(Matsuoka et al. 2012). This means that packing density cannot be estimated using polymer

441

surface activity and therefore further research is required to gain a fuller understanding of the

442

mechanism for polymer adsorption and packing at a bubble surface.

443 444

4.2.

445

When chemicals are applied to bubble surfaces as opposed to particles and colloids, the

446

flotation of particles from water is dependent on their effective interaction with bubbles. In

447

this work and previous research (Henderson et al. 2008c), the use of CTAB resulted in cell

448

removals comparable to that predicted by theory and were thus dependent on cell size

449

(Haarhoff and Edzwald 2004). The use of polymers for this cyanobacteria system was able to

450

exceed the theoretical particle removal efficiency, demonstrated not only in this research, but

451

on other algae and cyanobacteria (Henderson et al. 2010b) and other synthetic raw water

452

(Malley 1995).

453

Henderson et al. (2010b), is the increase in “swept volume”, whereby the effective surface

454

area of a polymer modified bubble is greater than an unmodified bubble, facilitating further

455

cell attachments. However, as the bridging distances of the polymers used are insignificant

456

compared to cell sizes (200 nm (Henderson et al. 2010b) versus 3.0 µm, respectively),

457

additional interactions must also occur. During PosiDAF jar tests, the large bubble-cell

458

networks that were observed may be the result of AOM and polymer complexation, resulting

459

in extended bridging lengths that enhanced the capture of cells. This is plausible as it has

460

been shown that polymer preferentially interacts with AOM over cells (Haarhoff and Cleasby

461

1989).

462

presence of additional DOC from sources other than algae/cyanobacteria could alter bubble-

TE D

M AN U

SC

Mechanisms of Cell Removal in PosiDAF

RI PT

438

AC C

EP

A potential mechanism for this enhanced separation, as discussed by

When considering applications of the process in a water treatment context, the

ACCEPTED MANUSCRIPT 463

polymer-cell interactions as observed in this work and this would require investigation in the

464

appropriate context.

465 With high cell removals in all polymer samples tested, the PosiDAF polymer performance

467

was indicated by the zeta potential in the treated water.

468

containing a large number of hydrophobic pendant groups, for example polymer HC10-h,

469

may further strengthen polymer adhesion to bubbles. The grouping of the hydrophobic

470

regions can also facilitate bubble nucleation via catalytic effects offered by the hydrophobic

471

zones (Lubetkin 2003), leading to bubbles formed with polymers in situ. When bubbles are

472

formed in the presence of the polymer, little to no diffusion is required to associate a polymer

473

to its surface. Though the hydrophobically functionalised polymer may not project into

474

solution as readily as a cationic polymer without hydrophobic pendant groups, it is evident

475

that bubble-polymer-AOM-cell suprastructures still form. It was therefore found that the

476

ideal polymer for PosiDAF were those comprising increased proportions of hydrophobic

477

functionalisation of long hydrophobic pendant groups, specifically a carbon chain length of

478

approximately C10.

481 482

EP

TE D

M AN U

SC

The association of polymers

AC C

479

480

RI PT

466

5. CONCLUSIONS

The following conclusions have been drawn from this study:

483

• Use of CTAB, polyDADMAC and all synthesised polymers as bubble modifiers

484

resulted in positively charged bubbles with statistically different zeta potentials and charge

485

characteristics

ACCEPTED MANUSCRIPT • The use of all synthesised polymers resulted in cell removals in excess of 90% with a

486 487

maximum of 99%.

These removals were comparable to that obtained when using

488

commercial polyDADMAC as a bubble modifier and conventional coagulation-DAF • The most negative zeta potentials in PosiDAF treated effluent were achieved using

490

synthesised polymer HC10-h, indicating a relatively low polymer concentration in the treated

491

water and therefore stronger bubble attachment. This suggests that higher polyDMAEMA

492

quaternisation/hydrophobic functionality leads to enhanced bubble attachment

SC

RI PT

489

• Large superstructures were observed to form suggesting that polymers and AOM

494

complex, leading to large networks that link bubbles, polymer, AOM and cells, enhancing

495

overall cell capture and thus exceeding removal efficiencies that are predicted by a flotation

496

model.

497

6.

498

This research was supported under Australian Research Council’s Linkage Projects funding

499

scheme (project number LP0990189) which included support from SA Water, Veolia Water,

500

United Water, Melbourne Water and Seqwater. In addition, the authors would like to thank

501

Water Research Australia for the PhD top up scholarship (project number 4025-10) provided

502

over the duration of this work.

503

7. REFERENCES

504

Al-Tebrineh, J., Mihali, T.K., Pomati, F. and Neilan, B.A. (2010) Detection of saxitoxin-

505

producing cyanobacteria and anabaena circinalis in environmental water blooms by

506

quantitative PCR. Applied and Environmental Microbiology 76 (23), 7836-7842.

AC C

EP

TE D

ACKNOWLEDGEMENTS

M AN U

493

ACCEPTED MANUSCRIPT 507

Aoki, K. and Adachi, Y. (2006) Kinetics of polyelectrolyte adsorption onto polystyrene latex

508

particle studied using electrophoresis: Effects of molecular weight and ionic strength.

509

Journal of Colloid and Interface Science 300 (1), 69-77. Bolch, C.J.S. and Blackburn, S.I. (1996) Isolation and purification of Australian isolates of

511

the toxic cyanobacterium Microcystis aeruginosa Kütz. Journal of Applied Phycology

512

8 (1), 5-13.

514

Bolto, B. and Gregory, J. (2007) Organic polyelectrolytes in water treatment. Water Research 41 (11), 2301-2324.

SC

513

RI PT

510

Buisine, F. and Oemcke, D. (2003) Seasonal influence of waste stabilisation pond effluent on

516

DAF/F (dissolved air flotation/filtration) process operation. Water Science and

517

Technology 48 (2), 357-364.

M AN U

515

Bütün, V., Armes, S.P. and Billingham, N.C. (2001) Selective quaternization of 2-

519

(dimethylamino)ethyl methacrylate residues in tertiary amine methacrylate diblock

520

copolymers. Macromolecules 34 (5), 1148-1159.

TE D

518

Chen, C., Zhang, X.j., Zhu, L.x., Liu, J., He, W.j. and Han, H.d. (2008) Disinfection by-

522

products and their precursors in a water treatment plant in North China: Seasonal

523

changes and fraction analysis. Science of the Total Environment 397 (1-3), 140-147.

525 526

Cho, S.-H., Kim, J.-Y., Chun, J.-H. and Kim, J.-D. (2005) Ultrasonic formation of

AC C

524

EP

521

nanobubbles and their zeta-potentials in aqueous electrolyte and surfactant solutions. Colloids and Surfaces A: Physicochemical and Engineering Aspects 269 (1–3), 28-34.

527

Edzwald, J.K. (2010) Dissolved air flotation and me. Water Research 44 (7), 2077-2106.

528

Elmallidy, A.M., Mirnezami, M. and Finch, J.A. (2008) Zeta potential of air bubbles in

529 530 531

presence of frothers. International Journal of Mineral Processing 89 (1-4), 40-43. Gehr, R. and Henry, J.G. (1982) The adsorption behaviour of cationic polyelectrolytes in dissolved air flotation. Water Science and Technology 14 (6-7 /7), 689-704.

ACCEPTED MANUSCRIPT 532 533

Haarhoff, J. and Cleasby, J.L. (1989) Direct filtration of Chlorella with cationic polymer. Journal of Environmental Engineering 115 (2), 348-366. Haarhoff, J. and Edzwald, J.K. (2004) Dissolved air flotation modelling: Insights and

535

shortcomings. Journal of Water Supply Research and Technology-Aqua 53 (3), 127-

536

150.

RI PT

534

Han, M.Y., Kim, M.K. and Shin, M.S. (2006) Generation of a positively charged bubble and

538

its possible mechanism of formation. Journal of Water Supply Research and

539

Technology-Aqua 55 (7-8), 471-478.

SC

537

Henderson, R.K., Baker, A., Parsons, S.A. and Jefferson, B. (2008a) Characterisation of

541

algogenic organic matter extracted from cyanobacteria, green algae and diatoms.

542

Water Research 42 (13), 3435-3445.

M AN U

540

Henderson, R.K., Parsons, S.A. and Jefferson, B. (2008b) Successful removal of algae

544

through the control of zeta potential. Separation Science and Technology 43 (7),

545

1653-1666.

TE D

543

Henderson, R.K., Parsons, S.A. and Jefferson, B. (2008c) Surfactants as bubble surface

547

modifiers in the flotation of algae: dissolved air flotation that utilizes a chemically

548

modified bubble surface. Environmental Science & Technology 42 (13), 4883-4888.

550 551

Henderson, R.K., Parsons, S.A. and Jefferson, B. (2009) The potential for using bubble

AC C

549

EP

546

modification chemicals in dissolved air flotation for algae removal. Separation

Science and Technology 44 (9), 1923-1940.

552

Henderson, R.K., Parsons, S.A. and Jefferson, B. (2010a) The impact of differing cell and

553

algogenic organic matter (AOM) characteristics on the coagulation and flotation of

554

algae. Water Research 44 (12), 3617-3624.

555

Henderson, R.K., Parsons, S.A. and Jefferson, B. (2010b) Polymers as bubble surface

556

modifiers in the flotation of algae. Environmental Technology 31 (7), 781 - 790.

ACCEPTED MANUSCRIPT 557 558

Hunter, R.J. (1981) Zeta potential in colloid science : principles and applications, London : Academic Press, London. Jamadagni, S.N., Godawat, R. and Garde, S. (2009) How Surface Wettability Affects the

560

Binding, Folding, and Dynamics of Hydrophobic Polymers at Interfaces. Langmuir 25

561

(22), 13092-13099.

RI PT

559

Kam, S.K. and Gregory, J. (1999) Charge determination of synthetic cationic polyelectrolytes

563

by colloid titration. Colloids and Surfaces A: Physicochemical and Engineering

564

Aspects 159 (1), 165-179.

567 568 569 570

polyelectrolytes. Water Research 35 (15), 3557-3566.

M AN U

566

Kam, S.K. and Gregory, J. (2001) The interaction of humic substances with cationic

Karhu, M., Leiviskä, T. and Tanskanen, J. (2014) Enhanced DAF in breaking up oil-in-water emulsions. Separation and Purification Technology 122 (0), 231-241. Lubetkin, S.D. (2003) Why is it much easier to nucleate gas bubbles than theory predicts? Langmuir 19 (7), 2575-2587.

TE D

565

SC

562

Lurie, M. and Rebhun, M. (1997) Effect of properties of polyelectrolytes on their interaction

572

with particulates and soluble organics. Ives, K.J. and Adin, A. (eds), pp. 93-101,

573

Jerusalem, Isr.

575 576

Malley, J.P. (1995) The use of selective and direct DAF for removal of particulate

AC C

574

EP

571

contaminants in drinking water treatment. Water Science and Technology 31 (3-4),

49-57.

577

Matsuoka, H., Hachisuka, M., Uda, K., Onishi, T. and Ozoe, S. (2012) Why ionic

578

amphiphilic block copolymer can be non-surface active comparison of homopolymer,

579

block and random copolymers of poly(styrenesulfonate). Chemistry Letters 41 (10),

580

1063-1065.

ACCEPTED MANUSCRIPT 581

Matsuoka, H., Maeda, S., Kaewsaiha, P. and Matsumoto, K. (2004) Micellization of non-

582

surface-active diblock copolymers in water. Special characteristics of poly(styrene)-

583

block-poly(styrenesulfonate). Langmuir 20 (18), 7412-7421. Napper, D.H. (1983) Polymeric Stabilization of Colloidal Dispersions.

585

Oliveira, C. and Rubio, J. (2011) Zeta potential of single and polymer-coated microbubbles

586

using an adapted microelectrophoresis technique. International Journal of Mineral

587

Processing 98 (1–2), 118-123.

RI PT

584

Oliveira, C. and Rubio, J. (2012) Kaolin aerated flocs formation assisted by polymer-coated

589

microbubbles. International Journal of Mineral Processing 106–109 (0), 31-36.

590

Pieterse, A.J.H. and Cloot, A. (1997) Algal cells and coagulation, flocculation and

M AN U

591

SC

588

sedimentation processes. Water Science and Technology 36 (4), 111-118. Qu, X., Wang, L., Karakashev, S.I. and Nguyen, A.V. (2009) Anomalous thickness variation

593

of the foam films stabilized by weak non-ionic surfactants. Journal of Colloid and

594

Interface Science 337 (2), 538-547.

597 598

Wiley, New York.

Schmolka, I. (1977) A review of block polymer surfactants. Journal of the American Oil

EP

596

Rosen, M.J. and Milton, J.R. (1978) Surfactants and interfacial phenomena, New York :

Chemists' Society 54 (3), 110-116.

AC C

595

TE D

592

599

Teixeira, M.R. and Rosa, M.J. (2006) Comparing dissolved air flotation and conventional

600

sedimentation to remove cyanobacterial cells of Microcystis aeruginosa: Part I: The

601

key operating conditions. Separation and Purification Technology 52 (1), 84-94.

602

van de Wetering, P., Schuurmans-Nieuwenbroek, N.M.E., van Steenbergen, M.J.,

603

Crommelin, D.J.A. and Hennink, W.E. (2000) Copolymers of 2-(dimethylamino)ethyl

604

methacrylate with ethoxytriethylene glycol methacrylate or N-vinyl-pyrrolidone as

605

gene transfer agents. Journal of Controlled Release 64 (1–3), 193-203.

ACCEPTED MANUSCRIPT 606

Yang, C., Dabros, T., Li, D., Czarnecki, J. and Masliyah, J.H. (2001) Measurement of the

607

zeta potential of gas bubbles in aqueous solutions by microelectrophoresis method.

608

Journal of Colloid and Interface Science 243 (1), 128-135. Yoon, R.-H. and Yordan, J.L. (1986) Zeta-potential measurements on microbubbles

610

generated using various surfactants. Journal of Colloid and Interface Science 113 (2),

611

430-438.

RI PT

609

AC C

EP

TE D

M AN U

SC

612

ACCEPTED MANUSCRIPT 1

TABLE 1. Synthesised polymers identified for use in PosiDAF; accompanying data

2

includes polymer stoichiometric quaternisation percentages (as determined by NMR),

3

charge density and surface tension Stoichiometric Quaternisation

Charge Density (meq g-1)

Surface Tension (mN m-1 at 1 mg L-1)

LC5-l LC5-m LC10-h MC1-m MC5-l MC10-h HC1-m HC10-h HC15-l polyDMAEMA polyDADMAC

5% 21% 40% 42% 2% 34% 35% 49% 8% 0% 100%

1.20 2.41 2.88 3.44 1.91 2.45 3.01 2.76 1.38 1.21 6.64

45.6 54.3 69.5 66.2 44.1 56.8 56.8 69.0 41.1 44.0 71.9

AC C

EP

SC

M AN U

TE D

4

RI PT

Sample Name

ACCEPTED MANUSCRIPT TABLE 2.

Microcystis aeruginosa cell properties

Attribute Morphology Diameter (µm) Cell Concentration (cells mL-1) AOM (mg C cell-1) Zeta Potential (mV) Charge Density (meq cell-1) ‡

CCAP 1450/3 Spherical 5.4 † 10 × 10-10 † -19.8 ‡ -2.0 × 10-12 †

TE D

M AN U

SC

Data obtained from Henderson et al. (2010a); Data obtained from Henderson et al. (2010b)

EP

†

AC C

2

CS-564/01 Spherical 3.0 ±0.7 2.1 × 107 ± 2 × 106 8.04 × 10-10 ± 4.4 × 10-11 -31.6 ± 1.6 -1.51 × 10-9 ± 7 × 10-11

RI PT

1

ACCEPTED MANUSCRIPT Cell Removal

Zeta Potential

100

30

Cell Removal (%)

10

60

0 40

-10

20

Zeta Potential (mV)

20

80

0

RI PT

-20 -30

0

1

2

3

4

Dose of Aluminium (mg L-1)

1

3

reported as concentration as Al

6

Conventional DAF with coagulation-flocculation pretreatment; dose is

SC

FIGURE 1.

AC C

EP

TE D

M AN U

2

5

RI PT

ACCEPTED MANUSCRIPT

1

Observation of flocs formed after coagulation of M. aeruginosa with alum

3

at a dose of 1 mg L-1 as Al

SC

FIGURE 2.

AC C

EP

TE D

M AN U

2

ACCEPTED MANUSCRIPT 8

60

6

40

4

20

2

0

0

RI PT

Bubble Zeta Potential (mV)

Polymer Charge Density

Polymer Charge Density (meq g-1)

Bubble Zeta Potential

80

1 FIGURE 3.

Average bubble zeta potentials for the selection of polymers,

3

polyDADMAC and CTAB (×) and respective charge density of the polymer used in the

4

test (○); zeta potential measurements were conducted on solutions made up to 1.8 mM

5

of NaCl and 0.5 mM of NaHCO3, corrected to pH 7 with 1.6 mg L-1 of polymer

AC C

EP

TE D

M AN U

SC

2

ACCEPTED MANUSCRIPT LC5-l polyDADMAC

A

LC5-m CTAB

HC10-h

100

60

40

20

0 0.0

0.2

0.4

0.6

Dose of Polymer (mg L-1)

1 100

60 40 20 0 0.0

M AN U

80 Cells Removed (%)

0.8

SC

B

RI PT

Cells Removed (%)

80

0.5

1.0

1.5

2.0

Dose of Polymer (× 10-3 meq L-1)

TE D

2 FIGURE 4.

Dose response curves for three of the nine polymers and CTAB in

4

comparison to polyDADMAC and CTAB - graphs show cell removal versus dose as (A)

5

polymer mass and (B) dose as charge – the results for all polymers are presented in

6

Table S3

AC C

EP

3

ACCEPTED MANUSCRIPT

RI PT

1

Visual observations from PosiDAF using 2 × 10-3 meq L-1 of

FIGURE 5.

3

polyDADMAC – (A) a photograph of rising bubble networks beneath the float layer,

4

observed from the side of the jar after the introduction of saturated water (26mm lens)

5

and (B) A microscope image of the float after a jar test (10× magnification)

AC C

EP

TE D

M AN U

SC

2

ACCEPTED MANUSCRIPT LC5-l polyDADMAC

LC5-m CTAB

HC10-h

50 40 20 10

0 -10 -20 -30 -40 -50 0.0

1

0.5

1.0 1.5 2.0 Dose of Polymer (× 10-3 meq L-1)

2.5

RI PT

Zeta Potential (mV)

30

3.0

FIGURE 6.

Dose response of HC10-3, LC5-1, LC5-2, polyDADMAC and CTAB in

3

terms of zeta potential - the results for all polymers are presented in Table S3

AC C

EP

TE D

M AN U

SC

2

ACCEPTED MANUSCRIPT

Highlights • Bubbles modified with polymers were found to have a positive surface charge • Successful cyanobacteria removal from water was achieved using

RI PT

PosiDAF • PosiDAF was found to be comparable to conventional DAF

• Effluent zeta potential from PosiDAF could be manipulated by

AC C

EP

TE D

M AN U

SC

introducing hydrophobic functionality to polymer backbones

ACCEPTED MANUSCRIPT

Hydrophobically-associating

cationic

polymers

as

micro-bubble

2

surface modifiers in dissolved air flotation for cyanobacteria cell

3

separation

4

SUPPLIMENTARY INFORMATION

RI PT

1

5

Yap, R.K.,1,2 Whittaker, M.,3 Diao, M.,4 Stuetz, R.M.,1 Jefferson, B.,5 Bulmuş, V.,6

7

Peirson, W.L.,7 Nguyen, A.,4 and Henderson, R.K.8,*

SC

6

8

1

9

University of New South Wales, Sydney NSW 2052, Australia.

M AN U

UNSW Water Research Centre, School of Civil and Environmental Engineering, The

10

2

11

University of New South Wales, Sydney NSW 2052, Australia.

12

3

13

Australia.

14

4

15

4072, Australia

16

5

17

Bedfordshire, MK43 0AL, UK.

19 20

TE D

Monash Institute of Pharmaceutical Sciences (MIPS), Monash University, IVC 3052,

EP

School of Chemical Engineering, The University of Queensland, Brisbane, Queensland

Cranfield Water Science Institute, School of Applied Sciences, Cranfield University,

AC C

18

Centre for Advanced Macromolecular Design, School of Chemical Engineering, The

6

Department of Chemical Engineering, Izmir Institute of Technology, Urla 35430 Izmir,

Turkey. 7

Water Research Laboratory, School of Civil and Environmental Engineering, The

21

University of New South Wales, Manly Vale NSW 2093, Australia.

22

8

23

2052, Australia

24

*Corresponding author: [email protected]

School of Chemical Engineering, The University of New South Wales, Sydney, NSW

ACCEPTED MANUSCRIPT 25 26 Table S1 Pictorial representations of each of the polymers used in this study – blue lines indicate the polymer backbone and red projections indicated hydrophobic alkyl chains Sample Schematic LC5-l

RI PT

27 28 29

LC5-m

MC1-m MC5-l

TE D

MC10-h

M AN U

SC

LC10-h

HC1-m

EP

HC10-h

30 31

AC C

HC15-l

ACCEPTED MANUSCRIPT

RI PT

†

PI is calculated as a ratio between coefficients of the correlation function, calculated according to ISO13321:1996 E and Koppel, D.E. (1972) Analysis of macromolecular polydispersity in intensity correlation spectroscopy: The method of cumulants. The Journal of Chemical Physics 57 (11), 4814-4820. To summarise, larger numbers indicate more polydisperse samples. For example, near monodisperse standards can be expected for have a PI of 0.7 indicate a very broad distribution of particle sizes.

SC

34 35 36 37 38 39 40 41 42

Table S2 Zeta potential and hydrodynamic diameter (with polydispersity index (PI)) of polymers in buffer solution at pH 7 and a concentration of 1 mg mL-1 Sample Zeta Potential Hydrodynamic Diameter (mV) (Dh, nm) and PI† LC5-l +26.7 ± 1.7 18.7 PI: 0.55 LC5-m +27.6 ± 2.0 90.5 PI: 0.46 LC10-h +38.6 ± 0.9 110.9 PI: 0.53 MC1-m +27.8 ± 2.2 12.2 PI: 0.46 MC5-l +37.9 ± 1.8 17.3 PI: 0.18 MC10-h +29.1 ± 2.0 66.1 PI: 0.23 HC1-m +54.4 ± 1.4 235.9 PI: 0.44 HC10-h +17.2 ± 1.8 244.0 PI: 0.30 HC15-l +44.7 ± 2.6 1129.0 PI: 0.43

M AN U

32 33

AC C

EP

TE D

43

ACCEPTED MANUSCRIPT

Table S1 Dose response curves for all functionalised polymers (shapes); results for all functionalised polymers (shapes) are compared to results obtained for polyDADMAC and CTAB (lines) Zeta potential Cell removal

AC C

EP

TE D

M AN U

SC

RI PT

44 45 46 47

48

SC

RI PT

ACCEPTED MANUSCRIPT

49

Figure S1 An example of growth curves obtained for M. aeruginosa strains CS555/01 and CS-564/01 – the end of the exponential growth phase in this instance was found at day 10

M AN U

50 51 52 53 54

56

TE D

55

n p ,e

 3α η ϕ v t  = 1 − exp  − pb T b b cz  2d b  

EfficiencyOfRemoval = 1 −

58 59 60 61 62

Equation S1 White water model performance equation as developed by Haarhoff and Edzwald (2004); np,e and np,i = number of particles in the effluent and influent water respectively; αpb = the attachment efficiency; ηT = the dimensionless particle transport coefficient; φb = the bubble volume concentration; νb = the bubble rise velocity; tcz = the time the bubble spends in the contact zone; db = the bubble diameter

AC C

EP

57

n p ,i