Influence of the Mixing Energy Consumption Affecting Coagulation

Subscriber access provided by UNIVERSITY OF SOUTH CAROLINA LIBRARIES

Article

Influence of the Mixing Energy Consumption Affecting Coagulation and Floc Aggregation Yamuna Srinivasan Vadasarukkai, and Graham A. Gagnon Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b06281 • Publication Date (Web): 16 Feb 2017 Downloaded from http://pubs.acs.org on February 16, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

Environmental Science & Technology

1

Influence of the Mixing Energy Consumption Affecting Coagulation and Floc Aggregation

2

Yamuna S. Vadasarukkai and Graham A. Gagnon*

3

*

Corresponding Author: Dr. G.A. Gagnon, Dalhousie University, Halifax, NS, B3J 1Z1

4

e-mail [email protected], Tel: 902.494.3268, Fax 902.494.3108

5

Abstract

6

The operational significance of energy-intensive rapid mixing processes remains unaddressed in

7

coagulation and flocculation of insoluble precipitates (flocs), which play an important role in the

8

removal of impurities from drinking water supplies. In this study, the influence of rapid mixing

9

and associated mixing energy on floc aggregation was examined for a surface water source

10

characterized by a high fraction of aquatic humic matter. Infrared spectral analyses showed that

11

the colloidal complexes resulting from ligand exchange between iron and dissolved natural

12

organic matter (DOM) were not substantially influenced by the mixing energy input. This

13

signified that DOM removal by coagulation can be achieved at lower mixing intensity, thereby

14

reducing energy consumption. In contrast, macroscopic investigations showed the coagulation

15

mixing energy affected floc size distributions during the slow mixing stage in flocculation and,

16

to some extent, their settling characteristics. The results from analysis of floc properties clearly

17

showed that more mixing energy was expended than necessary in coagulation, which is typically

18

designed at a high mixing intensity range of 600-1000 s-1 in treatment plants. The key findings

19

from this study have practical implications to water utilities to strategically meet water quality

20

goals while reducing energy demands.

21

Abstract Art

22

1 ACS Paragon Plus Environment


23

1. Introduction

24

Urban water infrastructure systems require significant amounts of energy for proper operation of

25

major processes, which may include water intake, conveyance, treatment, and distribution

26

systems, to deliver a safe and high-quality drinking water to customers. Life cycle analysis

27

(LCA) is often used to quantify energy usage at each stage1-3 as measured by energy

28

benchmarks, such as kilowatt-hour of electricity consumed per unit volume of treated water.4

29

Accordingly, the predominant use of electricity is directly associated with pumping (e.g., low lift

30

and/or high lift pumps), which is connected to the pressure demands and operational

31

characteristics of water distribution networks. Ultimately, pumping represents 60-85% of the

32

total electricity consumption for surface water treatment plants.1, 5 Rapid mixing in coagulation is

33

described as the second largest consumer of direct energy in a conventional water treatment

34

plant.2, 5 A report from the Electric Power Research Institute (EPRI)5 estimated that rapid mixing

35

processes had a total electricity consumption of approximately 308 kW-h/day in a representative

36

surface water treatment plant with a intake of 45 million liters/day (MLD). In addition, the

37

primary energy (e.g., fossil fuels, nuclear energy, renewable energy) used in the production of

38

electricity for the operation of these systems is associated with greenhouse gas (GHG)

39

emissions.6

40 41

In water treatment practice, high energy is expended to ensure instantaneous dispersal of

42

chemicals added during coagulation with the plant influent. A rapid-mixing tank with

43

mechanical mixers is often a conventional means to accomplish mixing process in water

44

treatment practice.7 Rapid-mixing tanks are typically designed for intense mixing with a root

45

mean square velocity gradient (G-value) of 600-1000 s-1.8 Previous studies have identified

46

several disadvantages in this tank design,7, 9 stating that the practice of intense mixing in

47

coagulation is largely unnecessary. A major shortcoming is that a relatively large amount of

48

energy is utilized to achieve a typical design mixing intensity since the average power supplied

49

increases with the square of the G-value. For instance, Edzwald 7 calculated that the power

50

requirement to achieve a typical G-value of 750 s-1 was 6.25 times greater when compared to 300

51

s-1. This reflects a proportional increase in the energy cost and associated carbon emissions.

52


Page 2 of 31

Page 3 of 31


53

Experimental investigations have supported the theory that the coagulation process has a great

54

potential for energy savings with regards to rapid mixing conditions, without sacrificing water

55

quality standards.10-12 For instance, Vadasarukkai, et al.11 have proposed a low-mixing intensity

56

range of 110 s-1 < G < 450 s-1 for a combined effective removal of turbidity and dissolved natural

57

organic matter (DOM). Alternatively, a more sustainable system for coagulation mixing, such as

58

pipe and open channel static mixers, are now available to water plants.7 Nevertheless, there has

59

been no study to date that examines how energy utilized in coagulation specifically influences

60

the overall aggregation of metal hydroxide precipitates. Surface hydroxyl groups on these

61

precipitates aid in effective sorption of turbidity, pathogens, inorganics, and DOM to their

62

surfaces. Such precipitates can also grow into large sized aggregates through the formation of

63

hydroxyl bridges between metal precipitates,13 collectively termed flocs. These flocs play a vital

64

role in the fate and transport of contaminants incorporated in their structural matrix and their

65

removal in subsequent solid-liquid separation processes.

66 67

In efforts to reduce targeted impurities from drinking water supplies, it is important to

68

understand the factors that may affect floc aggregation. The objective of this research was to

69

understand how and to what degree the mixing energy consumed in coagulation influences the

70

overall characteristics of flocs, including morphology, size distributions, and settling

71

characteristics. Surface analyses of flocs formed immediately after the rapid mixing process (i.e.,

72

microflocs) were characterized using Fourier transform infrared (FT-IR) spectroscopy. This was

73

performed to obtain specific information on the mechanism of interaction(s) of sorbed

74

constituents and the metal hydroxide precipitates. Additionally, morphology of microflocs

75

formed during the rapid mixing stage was investigated using transmission electron microscopy

76

(TEM) and scanning electron microscopy (SEM) combined with energy dispersive X-Ray

77

spectrometry (EDS). A digital in-line holographic microscope (DIHM) was used to measure floc

78

size distributions during slow mixing in the flocculation stage. The results from the present study

79

provide opportunities for water utilities to strategically meet both water quality goals and energy

80

reduction demands. In particular, it is estimated that an average of 10% reduction in annual

81

energy costs in utilities across the U.S. represents a saving of more than $400 million annually to

82

the water industry.14 It is anticipated that this study will contribute to the ongoing research and



83

development in implementing technologies for an effective management of energy resources,

84

costs, and carbon footprints.

85

2. Materials and methods

86

Water quality:

87

Source water from Kelly Lake, drawn from the intake at the Louisbourg drinking water treatment

88

plant (Sydney, NS, Canada), was used in this study. This raw water source is characterized by an

89

average specific ultraviolet absorbance (SUVA) of 4 to 4.2 m-1 of absorbance per mg/L,

90

indicating a high fraction of aquatic humic matter. Table 1 provides a description of the general

91

characteristics of this source water, along with the results of elemental analysis measured using

92

the inductively coupled plasma-mass spectrometry (ICP-MS) (Thermo Scientific XSeries II ICP-

93

MS, Waltham, Massachusetts).

94

Coagulation-flocculation experiments:

95

Coagulation and flocculation experiments were tested at two pH conditions, a slightly acidic pH

96

of 4.80 and a near neutral pH of 6.85, following a procedure similar to that of Vadasarukkai, et

97

al..11 In that study, which used the same source water from Kelly Lake, an iron dose of 0.33 mM

98

and 0.66 mM was determined as optimal coagulant concentration (OCC) of ferric sulphate at

99

pHs 4.80 and 6.85, respectively. At the OCC, the authors showed total organic carbon (TOC)

100

concentration and residual turbidity decreased following settling in the treated water. These

101

optimal chemical conditions were used for enhanced coagulation in the present study to form

102

iron-DOM floc aggregates.

103 104

All experiments were carried out at a room temperature (21 ± 1oC) in a 1-L square jar test

105

beaker, fitted with a flat paddle impeller (76 x 25 mm). A compact digital mixer (Cole-Parmer,

106

Vernon Hills, Illinois), which had a mixing capacity ranging from 50 to 2500 rpm, was used to

107

mix the water sample. The equivalent G-value for the flat paddle impeller configuration was

108

determined from the laboratory velocity gradient calibration chart (G-curve) of Cornwell and

109

Bishop.15 The solution was adjusted to the required pH (4.80 or 6.85) by adding a pre-determined

110

volume of the pH adjusting chemicals (i.e., 1 N NaOH or 0.2 N HNO3 solutions). All reagents

111

used were of analytical grade and stock solutions were prepared using deionized (DI) water.


Page 4 of 31

Page 5 of 31


112 113

Energy calculations were based on a representative surface water treatment facility, with an

114

average flow of 120 million liters/day. Equation 1 was used to calculate the daily consumption of

115

energy during the rapid mixing operation in the coagulation process, expressed in kilowatt-

116

hour/day. The average power supplied to stir the mechanical mixer was calculated from the

117

previously determined G-value, assuming a motor efficiency of 85%.7 A water temperature of

118

15°C, a retention time of 20 s in the rapid mixing tanks, and a power cost of $0.12 per kW-h was

119

used as the plant design conditions.7

120

 G 2  Qavg  t m    E  Pavg  t     t    

(1)

121

where Pavg is the average power supplied in a mixing tank (J/s), t is the operating time period of

122

the mixer (assumed as 24 hours), G is the root mean square velocity gradient (s-1), μ is the

123

dynamic viscosity of water (kg/m.s), η is the motor efficiency, Qavg is the average daily plant

124

flow rate (m3/s), and tm is the mean hydraulic retention time (s).

125 126

Coagulation experiments were conducted at rapid mixing speeds ranging from 0 to 750 rpm,

127

which gave an equivalent G-value range from 0 to 1450 s-1. An optimized dose of ferric sulphate

128

and a calculated volume of pH adjusting chemicals were added at the start of the rapid mix

129

process. Coagulation was conducted for 90 s at a specific G-value. Flocculation was achieved at

130

a fixed G-value of 46 s-1 for 18 min. The energy consumed in the slow-mixing stage was not

131

included in the total energy consumption of the plant, but accounts for 103 kW-h/day. The

132

operating conditions used in this study were within the recommended standards followed in

133

water treatment plants,8, 16 except for the G-value and 90 s of mixing time specified in the

134

coagulation experiments. The use of more than one chemical (e.g., ferric sulphate, 1 N NaOH or

135

0.2 N HNO3 solutions) in the present case required this additional retention time to simulate

136

plant conditions.8

137

Characterization of micro- and macro- structures of iron-DOM floc aggregates:

138

Rapid mixing stage: Analyses of the functional groups in the resulting microflocs was performed

139

using an attenuated total reflectance (ATR) Fourier transform infrared spectroscopy (FT-IR) in

140

the mid-infrared region of 4000-650 cm-1 in the absorption mode (Cary 630 ATR FT-IR, Agilent 5 ACS Paragon Plus Environment


141

Technologies, California). An approximate volume (~200 mL) of floc aggregates was collected

142

following coagulation. The floc suspension was left undisturbed to settle for an hour in an acid-

143

rinsed crucible while supernatant was carefully discarded using a 25 mL glass pipette.

144

Flocs contain water molecules, which can absorb a significant portion of the infrared (IR) band

145

due to the hydroxyl group (-OH). Therefore, water molecules can interfere with the spectral

146

analysis of organic compounds (e.g., carboxylic, phenolic groups of DOM), as residual water are

147

reported to absorb strong signals at 3700-3000 cm-1 and 1700-1600 cm-1.17 For this reason, many

148

studies have concentrated and dried the floc samples .17-19 Fewer studies have preferred a diffuse

149

reflectance infrared Fourier transform (DRIFT) spectroscopy for investigations of aquatic humic

150

substances .20 Other adsorption studies have applied a thin layer of metal oxides on the crystal

151

surface and then filled the ATR cell with organic acid solutions .21 In the present study, sample

152

was oven dried at 35-40 °C for at least 3 days (or until the water was evaporated from the

153

sample). The oven-dried floc aggregates were gently scraped and stored in a vial at 4 °C until

154

analysis. A more specific description of the sample preparation method is illustrated in the

155

Supporting Information (SI).

156 157

An attenuated total reflectance flow cell with a zinc selenide (Zn-Se) crystal was used for this

158

purpose. The Happ-Genzel method was used with 200 scans at a resolution of 2 cm-1.

159

Backgrounds were taken with the crystal alone. Analysis of the spectral data was performed

160

using optical spectroscopy software (Spekwin32, Germany).22

161 162

Transmission electron microscopy (TEM) and scanning electron microscopy (SEM) combined

163

with energy dispersive X-Ray spectrometry (EDS) were used to take images of flocs generated

164

from coagulation. Approximately 20 mL sample was collected from the rapid mixing stage and

165

stored at 4oC until the analysis. As both electron microscopy techniques work in vacuum, a large

166

quantity of water in floc sample must be removed as a sample preparation step. Initial

167

investigation of floc images identified that flocs dried at 30-450C were irregular flat-plates,

168

coarse/rough textured aggregates with sharp edges (data not shown), which may not effectively

169

represent an amorphous floc precipitate commonly seen in water treatments. It was therefore

170

concluded that the sample preparation at an elevated temperature was unsuitable for the SEM

171

observation of floc structures. 6 ACS Paragon Plus Environment

Page 6 of 31

Page 7 of 31


172 173

Alternatively, Cornelissen et al.23 found the resulting damage from drying of flocs at room

174

temperature was not apparent at smaller floc sizes imaged using TEM. Therefore, a droplet of the

175

TEM sample was carefully placed on a 200-mesh copper grid from Electron Microscopy

176

Sciences (Hatfield, Pennsylvania). The grid was then allowed to dry at room temperature for 24

177

hours. This allowed the water from the surface of floc structures to evaporate and prevented

178

crystallization, which occurs at high-temperature (e.g., 30-104 °C) drying.17,19 A FEI Tecnai-12

179

TEM (Hillsboro, Oregon) equipped with a Gatan 832 CCD camera and Gatan Microscopy suite

180

software was used to take floc images. Duplicate samples were prepared for each analysis.

181 182

The sample preparation was kept consistent for both the electron microscopy so to produce

183

compatible results at different mixing conditions. That is, using a 10 mL pipette, a droplet of floc

184

sample was carefully placed on a SEM sample mount, and the sample was dried at room

185

temperature before image analysis. The samples were sputter coated with gold-palladium to

186

prevent charge buildup on poorly conducting samples, such as flocs in the present case. A

187

Polaron SC7620 Mini Sputter Coater (Quorum Technologies, Lewes) with a plasma current of

188

18 mA for 120s was used for this purpose. The samples may ineffectually dissipate the resulting

189

electron charge without such coating procedures, which can cause imaging artifacts.24 Both the

190

elements were removed during the energy dispersive x-ray spectroscopy (EDS) analysis of

191

elemental compositions. A Hitachi S-4700 FEG SEM (Hitachi, Tokyo) equipped with an energy

192

dispersive x-ray spectroscope was used to provide analysis of elemental composition of samples.

193

Each EDS spectrum was run for at least 30s, with an accelerating voltage of 20 kV and an

194

emission current of 8.5 µA. A minimum of four EDS spectra per image was analyzed and the

195

most representative images are presented. Each reported value was calculated by averaging the

196

elemental composition from various representative SEM images (4-7 images), with at least four

197

EDS spectra per image.

198 199

Flocculation stage: A digital in-line holographic microscope (DIHM) from 4deep Inwater

200

Imaging (Halifax, Nova Scotia, Canada) was used to measure the dynamic growth of floc sizes

201

during the coagulation-flocculation experiments. The microscope was vertically mounted to one

202

of the corners of the jar beaker, minimizing any interruption to the impeller rotations. Holograms 7 ACS Paragon Plus Environment


Page 8 of 31

203

of flocs were recorded continuously at 2 frames per second (fps), and they were stored as digital

204

bitwise images. The formation of microflocs was not recorded at coagulation G-values > 450 s-1

205

due to bubbles formed near the sampling port, which can cause significant errors in size

206

measurements. Numerical reconstruction of holograms was performed to get images of floc

207

aggregates, as described in Vadasarukkai, et al.25. Equivalent circular diameter (d) of floc size

208

was calculated from the area of the identified flocs, which was expressed as a volume-based floc

209

equivalent diameter.

210

Settling properties:

211

The treated water from the jar test beaker was poured into an Imhoff cone (Wheaton, Millville,

212

New Jersey) immediately after the slow stirring operation in flocculation. It was allowed to settle

213

quiescently for 24 hours. Samples were collected periodically from 10.3 cm below the water

214

surface using a wide-mouth glass pipette. Residual turbidity measurements were recorded using

215

a HACH 2100AN laboratory turbidimeter (Hach Company, Loveland, Colorado). The volume of

216

settleable solids was recorded at the end of the experiment. This study did not account for a drop

217

in water level height due to the withdrawal of approximately 40 mL of water samples during

218

turbidity measurements, which can result in an estimated 5% discrepancy for the high settling

219

velocity and at approximately a 30% discrepancy for the low settling velocity values.26 The

220

linear range of the curve was identified, and it was used to determine the sedimentation rate27, as

221

expressed in Equation 2.

222

Slope( NTU,min) 

y 2  y1 x 2  x1

(2)

223

where y1 and y2 are the values of turbidity in NTU of the linear slope. The slope of the regression

224

line was calculated between the time period 0 (x1) and 30 (x2) minutes of the settling period.

225

3. Results and discussion

226

Floc morphology and size distributions at optimal coagulant concentrations (OCC):

227

The iron to carbon ratio (i.e. > 0.8 mg Fe/mg TOC) used as OCC in the present study resulted in

228

a rapid growth of ferric flocs in the rapid mixing stage itself kept at a G of 175 s-1, which was

229

well within the optimal operating range (i.e., 110 s-1 < G < 450 s-1) of mixing energy input

230

recommended for enhanced coagulation.11 This corresponds to an energy usage of 27 kW-h/day

231

for a 120 million liters/day representative treatment facility. 8 ACS Paragon Plus Environment

Page 9 of 31


232 233

TEM micrographs in Figure 1 showed three sample population of flocs at the nanoscale,

234

illustrating that a variety of floc morphologies were present at the optimal coagulant

235

concentration. The darker regions of the TEM images in Figure 1b and 1c showed dense,

236

spheroid-like or needle-like clusters that were closely packed to one another. The darker regions

237

represented a floc size of 60 to 180 nm, as shown in the inserts. According to previous

238

microscopic investigations,28-29 the association of coagulant species and humic networks was

239

hypothesized to form similar electron-dense spots having a variety of shapes, including needle-

240

like, globular or net-like structures. In the present case, this network of clusters was

241

interconnected by what appear to be different sizes of granular structures that were lighter in

242

appearance. The lighter regions were prominent near the edges of floc aggregate. Another type of

243

floc morphology shown in Figure 1a was often seen in TEM micrographs at different scales of

244

observation. The shape of the aggregate can be best described as an irregular, flaky or relatively

245

tenuous precipitate made up of many particles of darker and lighter shades. Few of the darker

246

particles were closely packed, while many of them were loosely joined to one another, as one can

247

visually see from a large amount of void spaces in their structure compared to Figure 1b and c.

248

Many larger floc aggregates observed under a different magnification (e.g., Figures S2 in SI)

249

showed similar irregular geometric resemblance to the aggregate shown in Figure 1a. This visual

250

observation of self-similarity in the shape of floc aggregates, regardless of the scale of

251

observation, is consistent with the fractal nature of floc precipitates.23.30

252 253

Differences in the optimal coagulant dose at pH 4.80 and 6.85 (i.e., an iron dose of 0.33 mM vs

254

0.66 mM, respectively) resulted in a major difference in the overall floc size distributions

255

between the two pHs, as illustrated in Figure 2. A significant difference in the variation of floc

256

size distributions between the two experiments was confirmed using Levene’s test (α=0.05; N

257

>30). The volume fraction of floc sizes ranged from 100 to 1000 μm at pH 4.80. The median floc

258

diameter (D50) was 424 μm. At the higher iron dose (0.66 mM, pH 6.85), floc sizes were more

259

variable and had a greater D50 value of 581 µm.

260 261

Unlike the distinct differences in floc sizes, it was difficult to interpret the inherent differences in

262

the floc morphology between the two experiments because of fractal nature of floc aggregates. 9 ACS Paragon Plus Environment


263

Details on micrographs are included in the Supporting Information (SI) (i.e. Figures S2 in SI).

264

Most of the TEM micrographs showed irregular, flake-like geometries at both pHs. At pH 6.85,

265

many sections in the TEM grid had partially transparent patches of approximately 100 to 500 nm

266

in size (inserts shown in Figure S2c and S2d in SI) that resembles a sheet of fabric attached to

267

the precipitate. This TEM observation may represent amorphous hydroxide precipitates favoured

268

at around neutral pH. 31

269

Influence of mixing energy utilized in rapid mixing on coagulation, flocculation, and

270

settling processes:

271

This section discusses the influence of variation in the mixing energy consumed in coagulation

272

on the overall characteristics of floc properties, including morphology, size distributions, and

273

their settling characteristics. Energy calculations are representative of a 120 million liters/day

274

surface water treatment facility.

275

(a) Aggregation behaviour of primary particles (DOM) during coagulation

276

Initial population has a strong influence on floc size formation.32 In the present study, a major

277

constituent of primary particles was comprised of dissolved organic matter (DOM), which

278

played an important role in the characteristics of flocs resulting from coagulation. Results from

279

the FT-IR spectral analyses (Figure S3 in SI) showed similar surface complexation-ligand

280

exchange reactions between DOM and iron hydroxides at both pHs, despite the differences seen

281

in the floc size distributions. This phenomenon of adsorption of DOM on to metal oxide surfaces

282

has been demonstrated previously in natural soil and aqueous systems.33-35 It resulted in similar

283

IR peaks at 1559 and 1375 cm-1 at both the pHs in the present study (spectral results are

284

presented in Figure S3 in SI). These absorption bands represented the stretching vibrations of the

285

carboxylate groups (i.e., COO-) adsorbed onto the surface active hydroxyl functional groups of

286

iron hydroxides (i.e., Fe-OH2 and Fe-OH stretches), which has been previously noted for the

287

adsorption process on iron hydroxides.36

288 289

The adsorption of DOM on iron hydroxides were influenced by the pKa values of functional

290

groups in the raw water, which had approximate values of 2.85 and 7.03 from acid-base titration

291

curves (data not shown). At pH 4.80, disassociation of protons was expected, favouring the

292

COO- functional groups to replace the dominant OH2 groups from surfaces of iron.37 Therefore,

293

an addition of a 0.33 mM of iron at pH 4.80 resulted in a partial neutralization of the net positive 10 ACS Paragon Plus Environment

Page 10 of 31

Page 11 of 31


294

charge on iron hydroxide surfaces and formed insoluble iron-DOM complexes. Alternatively,

295

when the pH was near the pKa value of 7.03, more protons in the humic networks were

296

disassociated, leading to an increase in the iron dose from 0.33 to 0.66 mM at pH 6.85.

297 298

A major difference between the two pHs occurred in the fingerprinting region below 1100 cm-1,

299

which can be explained by the adsorption behaviour of sulphate. Peak, et al.38 had shown the

300

spectrum of sulfate adsorbed on goethite had a sharply defined peak at 976 cm-1 for pH less than

301

6, forming an inner-sphere sulfate complex with two distinctive splits formed at 1055 and 1133

302

cm-1. These peaks were similar to the IR peaks formed at pH 4.80 in the current study at 975,

303

1040, and 1111 cm-1. At pH 6.85, the sharp peak at 975 cm-1 was broadened and shifted to 969

304

cm-1. Aromatic ethers, carbohydrate, and polysaccharide groups also exhibit a broad band near

305

1120-980 cm-1 with a characteristics peak near 1034 cm-1.39 It is possible that sulphate may have

306

interacted with the co-adsorbed functional groups of DOM macromolecules and formed

307

complexes, instead of being directly incorporated on to the surface of iron hydroxides. This

308

observation was previously described for enhanced adsorption of trace metals onto iron oxide

309

surfaces through the formation of ternary complexes in the presence of high sulphate

310

concentrations.40

311 312

The relative peak intensity ratios in Table 2 provided a quantitative comparison of IR peaks

313

formed at different amounts of energy used in coagulation. Details on the spectral comparison at

314

different mixing energy input are included in Figure S4 in SI. The ratios for the three prominent

315

peaks of adsorbed constituents (i.e., 1559, 1375 and 1039 cm-1) and the structural O-H vibration

316

(i.e., 3177 cm-1) in the lattice structure of flocs were calculated as I1559/I3177, I1375/I3177, and

317

I1039/I3177. It showed a fairly constant ratio for the adsorbed carboxylate groups (i.e., COO-) peaks

318

1559 and 1375 cm-1 at different mixing energy inputs. This suggests iron-DOM complexes

319

formed in the hypothesized size range of 60 to 180 nm were less vulnerable to the coagulation G-

320

values. In other words, it implied DOM removal by coagulation can be achieved at reduced

321

energy consumption. This disproves the theory that intense mixing is necessary to bring about

322

particle-particle collision for effective interactions. However, a relatively large standard

323

deviation for the adsorbed sulphate peak near 1039 cm-1 was evident at both pHs. This implies



324

that the coagulation G-values caused some changes to the shape of absorption spectra of sulphate

325

complexes, indicating changes in their composition.

326

(b) Microstructure and their interactions

327

A scanning electron microscope with an energy dispersive X-ray spectroscopy (SEM-EDS)

328

analysis was used to compare the elemental composition of flocs formed at different mixing

329

energy inputs during coagulation. Although a consistent sample preparation technique was used

330

for both electron microscopy methods, floc images taken using SEM looked significantly

331

different from those observed by TEM, which is likely an artefact of drying process. This

332

suggests that drying at room temperature resulted in some perturbation of hydrated floc samples

333

for SEM observation. Cornelissen et al.23 found cryogenic freezing of the samples can improve

334

imaging of larger (1-100 µm) floc particles using SEM, without the loss of structural information

335

resulting from drying.

336 337

TEM images were more representative of flocs because of insignificant damage caused by the

338

drying process, which is likely due to the small size range of flocs analyzed using TEM.23 Floc

339

morphology can be evaluated using SEM despite the artefacts introduced by drying. SEM images

340

revealed qualitative differences in floc morphology as a function of mixing condition. For

341

instance, the SEM analysis showed the flocs resulting from a low mixing energy had flat plate-

342

like structures that were assembled in an elongated pattern. At higher mixing energies, smaller

343

flocs composed of globular aggregates with a compact structure were observed (Figure S5 in SI).

344

This phenomenon can be related to flocculation mixing studies.32,41 In those studies, a consistent

345

trend of reduction in floc sizes with increasing the mixing intensity was shown to form small and

346

compact flocs. The floc samples were investigated further using EDS analysis to understand the

347

elemental compositions that may have resulted in varied floc structures.

348 349

The EDS analysis are semi-quantitative as the elemental compositions are normalized to 100%

350

for the selected elements. A highly irregular surface of flocs resulting from the drying method

351

may cause increased variations at specific spots and between samples. A low concentration of

352

iron or DOM could be due to differences in surface area (e.g., small/globular, elongated/plate-

353

like structures), such uneven surfaces are suggested to increase variations in electron scattering

354

and absorption due to matrix effects.42 Therefore, the EDS results were not directly compared, 12 ACS Paragon Plus Environment

Page 12 of 31

Page 13 of 31


355

but it was used to evaluate the percent weight trend (e.g., increasing/decreasing) of major

356

elements at different mixing conditions.

357 358

The results from the EDS analysis for coagulation experiments carried out at the mixing energy

359

input of 2 and 183 kW-h/day are illustrated in Figure 3. The elemental composition of flocs was

360

comparable between the two mixing conditions, with oxygen detected as a major element along

361

with carbon, iron, sulphur, and sodium co-existing in their structural matrix. Even though EDS

362

spectra detected appreciable percent weight of aluminium (Al) in the floc samples, it was not

363

included in the analysis because aluminium was also present in the SEM sample mount. Other

364

elements, such as calcium and magnesium, were detected at a low concentration. The relatively

365

higher percent weight of oxygen may have been a result of iron hydrolysis (e.g., Fe-OH and Fe-

366

OH2) and precipitation reactions (e.g., iron hydroxides), in addition to the oxygen present in

367

organic matter.

368 369

The EDS data, along with the FT-IR analyses, reinforces the interpretation that hydrolyzed iron

370

species interacted with DOM macromolecules, resulting in flocs with DOM-iron compositions

371

within their structure. However, the elemental analysis cannot confirm the nature of the

372

interactions, which may require investigations using X-ray photoelectron spectroscopy (XPS)

373

and/or solid-state nuclear magnetic resonance (NMR) results.43 There was a marginal difference

374

in the percentage composition of major elements present between the two mixing conditions at

375

pH 4.80. That is, carbon and oxygen were detected at slightly higher amounts in the floc samples

376

taken at 183 kW-h/day, while the percent weight of iron and sulphur contents was lesser when

377

compared to floc samples taken at 2 kW-h/day. Likewise, floc aggregates formed at pH 6.85

378

showed similarities between the elemental compositions at 2 and 183 kW-h/day (Figure 3

379

(bottom panel)). In this case, an increased fraction of iron (e.g., 8.3-18.2 %) was observed in

380

those flocs, which may have experienced more rapid precipitation reaction under higher pH

381

conditions.

382 383

The EDS analysis of elemental composition of flocs supported the hypothesis that the hydrolysis

384

of iron and their subsequent aggregation with DOM particles occurred irrespective of the mixing

385

condition. Previous microscopic investigations have found that iron and DOM interactions 13 ACS Paragon Plus Environment


386

resulted in aggregates of 60- 300 nm in size 28-29,44 Perikinetic or Brownian diffusion mechanism

387

initiated by thermal energy was likely to control particles-coagulant collisions during coagulation

388

that may result in successful attachment of particles ,45 while the importance of coagulation

389

mixing has been related to settling characteristics of flocs.29,46 A model proposed by Siéliéchi et

390

al.29 has indicated that low-intensity mixing favours small sediment volume due to intra-humic

391

colloid shrinkage and high-intense mixing is associated with a high collision rate and large

392

sediment volume.

393

(c) Floc size distribution

394

Mixing energy input in coagulation governed the floc growth during the slow mixing stage in

395

flocculation, as illustrated in Figure 4. Unless stated, the data represents the volume fractions of

396

floc size distributions measured using the in-line holographic microscopy in flocculation at 46 s-

397

1

398

floc size distribution at pH 4.80. On the contrary, a wide range of floc sizes were formed at pH

399

6.85 even without having a proper mixing condition in coagulation. The slow mixing energy

400

provided at a 103 kW-h/day in flocculation compensated for the no mixing operation in

401

coagulation and achieved necessary bridging between precipitates in that case. On a similar note,

402

a recent study had shown the overall water treatment performance of a full-scale drinking water

403

treatment plant remained unaffected during the accidental failures of mechanical mixer in the

404

rapid mixing tank.10

. It was apparent that no mixing energy during coagulation resulted in an obvious decline in the

405 406

For an average flow of 120 million liters/day, the energy consumed to operate at a specified

407

designed G-values of 600-1000 s-1 is in the range of 342 to 913 kW-h/day. As seen from Figure

408

4, operating at a mixing energy > 200 kW-h/day did not show any significant improvement in the

409

evolution of floc sizes in flocculation, in terms of median (D50) and 90th percentiles (D90).

410

Conversely, there was a slight drop in the median floc diameter at higher mixing energy. That is,

411

the decrease in the median floc size was from 375 to 275 μm at pH 4.8 and it dropped from 550

412

to 325 μm at pH 6.85. Likewise, Pivokonsky et al.41 found for a pilot-scale set up that a shallow

413

reduction in the average floc diameter was observed at G ≥ 200 s-1, with the value of the slope

414

being 0.34. This type of fragmentation of iron hydroxide flocs has been previously indicated for

415

tannery waste samples at > 170 s-1.47

416 14 ACS Paragon Plus Environment

Page 14 of 31

Page 15 of 31


417

A general trend of decrease in floc sizes occurred in the case of higher energy usage,

418

representing aggregate breakage. This phenomenon has been demonstrated previously, with floc

419

aggregates becoming more compact and homogeneous in size with increasing G-values.32,41 Floc

420

breakage could be a result of variations in local G-value,48-49 which was not accounted for in the

421

present study. High mixing energy may contain maximum local G-values several orders of

422

magnitude higher than the average G-value. Bridgeman et al.48 found that a 500 μm particle was

423

subjected to 9.4 times the average G-value of 165 s-1 at every 5 seconds of rapid mixing. This

424

suggests that, in the present case, the frequency at which a floc was exposed to a maximum

425

(peak) local velocity gradient is expected to increase with increasing mixing energy. Further

426

analysis, involving floc size measurements are required to investigate whether shortening the

427

mixing time is necessary to promote floc aggregation and subsequently, improve particle

428

removal50 at different mixing conditions.

429 430

Floc breakage can pose an important practical problem by causing resuspension of impurities

431

and increased iron concentrations in settled water. An overall shift in the floc size distributions to

432

the left support this theory of an increase in the volume fractions of smaller flocs less than 100

433

μm in diameter. While an increase in small flocs population with increasing mixing regimes was

434

indicated, the size detection limit of DIHM was limited to provide a quantitative estimation of

435

small flocs less than 17 µm. The size range of the DIHM was suitable to measure flocs larger

436

than 50 μm, which are the major source contributing towards total surface area, volume, and

437

mass of flocs.51 Further analysis, using a Malvern Mastersizer 3000 52 and/or Zetasizer nano 44 ,

438

is recommended to examine the small floc size distribution that may have resulted at high energy

439

usage, as carryover of these flocs can pose significant challenges on downstream separation

440

processes (e.g., turbidity breakthrough in filters, reduced filter runtimes, increased organic

441

loading etc.). Recent work by Yu et al.13 suggests that broken flocs may not grow into large

442

aggregates due to a net reduction in the number of active surface hydroxyl groups. This may

443

greatly reduce the effectiveness of separation processes in the present case.

444

(d) Settling kinetics of flocs:

445

Comparison of slope values in Table 3 indicated the sedimentation rates were higher for the

446

optimal iron dose of 0.66 mM at pH 6.85 than for 0.33 mM at pH 4.8. This implies a faster rate



447

of floc formation and subsequent settling rates were achieved at the higher particle concentration

448

27

in the case of the near neutral pH.

449 450

The settled water turbidity curve in Figure 5 provides a comparative evaluation of the effect of

451

coagulation mixing energy on the settling characteristics of flocs. An overall decreasing trend of

452

residual turbidity with time indicated the settling process had a dampening effect on the

453

differences in floc characteristics due to coagulation mixing energy. In other words, there was no

454

significant difference in the settling kinetics after 60 min, particularly in the recommended

455

mixing energy range of 11 to 181 kW-h.11 The compact flocs (essentially smaller size particles)

456

resulting from a higher coagulation mixing energy were able to settle given enough time, except

457

for the data shown at 0 and 1879 kW-h/day. As expected, when the coagulation mixing energy

458

was at 0 kW-h/day (i.e., G = 0 s-1) and 1879 kW-h/day (i.e., G = 1450 s-1), the residual turbidity

459

was fairly high (> 2 NTU) even after a settling time of 90 min (at an overflow rate of 0.1

460

cm/min). This demonstrates that a minimum amount of energy is necessary to form settleable

461

flocs and for effective separation of those aggregates from the settled water in the subsequent

462

separation processes.

463

4. Practical implications for water treatment:

464

The key findings from the present study, which was done at bench-scale, are also critical to pilot-

465

scale and full-scale implementations. In most treatment plants, physicochemical conditions

466

related to the coagulant type and dose, temperature, and pH are routinely optimized at

467

laboratory-scale using jar tests and at pilot-scale. Mixing energy input associated with the tank

468

geometry, impeller types, and speed are not often considered for floc growth in existing testing

469

protocols. This shortcoming in the design of rapid mixing systems is too big to ignore in the

470

current scenario of rising demands for energy and water and consequent carbon emissions. As

471

shown in this study, more mixing energy is expended in current treatment plants than is

472

necessary if they operate at recommended design G-values (i.e., 600-1000 s-1). In the case of a

473

120 MLD water treatment system, reducing energy consumption from a 502 kW-h/day,

474

representing a G-value of 750 s-1, to less than 200 kW-h/day, representing a G-value of 450 s-1,

475

would result in an annual energy savings of nearly 170 MW-h/year without compromising

476

effluent water quality. That is, energy savings of over 90% can be achieved depending on plant 16 ACS Paragon Plus Environment

Page 16 of 31

Page 17 of 31


477

flow rate. It is pertinent that stakeholders and decision-makers consider the key results from this

478

study to capitalize on energy efficiency opportunities.

479 480

Flocs play a significant role in the transportation of dissolved natural organic matter (DOM) and

481

other impurities within engineered treatment systems and their removal from drinking water

482

supplies. In this study, floc characteristics for a surface water source were investigated at two

483

pHs from within a range (pH 4 to 6) associated with coagulation of high colour and TOC

484

waters.53 An important inference drawn was that the mixing energy did not have an influence on

485

the iron-DOM colloidal complexes formed in coagulation. While this phenomenon of

486

complexation of DOM on to the surface of metal oxide is well documented in early mechanistic

487

studies,33,34 it is expected the impact of the reported finding is applicable to other water sources

488

and coagulant types. Further investigations on the effect of different particle concentrations are

489

recommended.

490 491

The major difference in the overall floc size distributions resulting at the optimal coagulant

492

concentrations for the two pHs studied would have potential implications for a number of

493

treatment plants. Prior investigation of the spatial distribution of floc sizes at a direct filtration

494

plant suggests large, irregular shaped floc aggregates can reduce the effective filter run times.25

495

The data acquired in the present study on floc sizes can be calibrated to a treatment plant’s

496

specific case by operating at a definite pH, coagulant dose, and mixing energy. For plants

497

requiring dense and pin-point flocs, such as in a direct filtration facility, a small standard

498

deviation in floc sizes would be beneficial; this can be achieved at a slightly acidic pH condition.

499

While ferric salts based coagulation have been practiced widely in the pH 5s, plants that

500

coagulate at this pH and less may require pH adjustments of treated water prior to filtration.

501

Alternatively, in treatment plants with clarification processes (e.g. sedimentation) it may be more

502

desirable to operate at a higher pH since this resulted in a relatively high standard deviation in

503

floc size distributions which would aid in differential settling processes.

504



505 506

5. Associated content Author Information

507

Corresponding Author

508

*Phone: 902.494.3268; fax: 902.494.3108; e-mail: [email protected]

509

ORCID

510

Yamuna S. Vadasarukkai: 0000-0003-4689-6888

511

Notes

512

The authors declare no competing financial interest

513

6. Acknowledgements

514

The authors acknowledge the financial support of both NSERC and Halifax Water through the

515

NSERC/Halifax Water Industrial Research Chair program. The authors acknowledge the

516

Institute for Research in Materials (IRM) at Dalhousie University. The authors thank the

517

technical and in-kind supports of Patricia Scallion (IRM, Dalhousie University) during SEM

518

analysis and Dr. Ping Li (Biology Dept., Dalhousie University) in executing TEM analysis. We

519

thank Jon-Paul Sun (Ph.D. Student) from the department of Physics at Dalhousie University for

520

allowing us to use the FTIR instrument. The authors also acknowledge the technical support of

521

Heather Daurie at Dalhousie University’s Clean Water lab. We would like to thank the

522

Louisbourg water treatment plant authority, Nova Scotia, Canada for providing access to the

523

water treatment for sampling.

524

7. Supporting Information Available

525

Five figures (FT-IR spectral data, Figure S1; TEM micrographs, Figure S2; FT-IR spectral data,

526

Figures S3 and S4; SEM-EDS analysis, Figure S5) addressing additional experimental data are

527

included.

528

8.

529 530 531 532 533 534

1. Santana, M. V. E.; Zhang, Q.; Mihelcic, J. R. Influence of water quality on the embodied energy of drinking water treatment. Environ. Sci. Technol. 2014, 48 (5), 3084-3091.

References

2. Plappally, A. K.; Lienhard, J. H. Energy requirements for water production, treatment, end use, reclamation, and disposal. Renewable & Sustainable Energy Reviews 2012, 16 (7), 48184848. 18 ACS Paragon Plus Environment

Page 18 of 31

Page 19 of 31

535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580


3. Stokes, J.; Horvath, A. Life cycle energy assessment of alternative water supply systems. International Journal of Life Cycle Assessment 2006, 11 (5), 335-343. 4. New York State Energy Research & Development Authority, Water and Wastewater Energy Management Best Practices Handbook; New York State Energy Research & Development Authority: Albany, New York, 2010. 5. Electric Power Research Institute, Water & Sustainability (Volume 4): U.S. Electricity Consumption for Water Supply & Treatment- The Next Half Century; Electric Power Research Institute Inc.: Palo Alto, CA, 2002. 6. Mo, W.; Wang, R.; Zimmerman, J. B. Energy-water nexus analysis of enhanced water supply scenarios: A regional comparison of Tampa Bay, Florida, and San Diego, California. Environ. Sci. Technol. 2014, 48 (10), 5883-5891. 7. Edzwald, J. K. Coagulant mixing revisited: theory and practice. Journal of Water Supply Research and Technology-Aqua 2013, 62 (2), 67-77. 8. American Water Works Association; American Society of Civil Engineering. Mixing, Coagulation, and Flocculation. In: Water Treatment Design; Fifth edition; Randtke, S.J., Horsley, M.B., Eds.; McGraw Hill, AWWA and ASCE: New York 2012. 9. Kawamura, S. Integrated Design and Operation of Water Treatment Facilities; Second edition; John Wiley & Sons: New York 2000. 10. Allerdings, D.; Foerster, G.; Vasyukova, E.; Uhl, W. The practical influence of rapid mixing on coagulation in a full-scale water treatment plant. Water Science and Technology 2015, 71 (4), 566-571. 11. Vadasarukkai, Y. S.; Gagnon, G. A. Application of low-mixing energy input for the coagulation process. Water Res. 2015, 84, 333-341. 12. Amirtharajah, A.; Mills, K.M. Rapid-mix design for mechanisms of alum coagulation. Journal American Water Works Association 1982, 74 (4), 210-216. 13. Yu, W.; Gregory, J.; Graham, N. Regrowth of broken hydroxide flocs: Effect of Added Fluoride. Environ. Sci. Technol. 2016, 50 (4), 1828-1833. 14. U.S Environmental Protection Agency, Ensuring a Sustainable Future: An Energy Management Guidebook for Wastewater and Water Utilities; Environmental Protection Agency: Washington, D.C., 2008. 15. Cornwell, D. A.; Bishop, M. M. Determining velocity-gradients in laboratory and full-scale systems. Journal American Water Works Association 1983, 75 (9), 470-475.



581 582 583

16. Ten State Standards of the Great Lakes-Upper Mississippi River Board of State and Provincial Public Health and Environmental Managers, Recommended Standards for Water Works; Health Research Inc., Health Education Services Division: Albany, New York, 2012.

584 585

17. Sari, M.A.; Chellam, S. Mechanisms of boron removal from hydraulic fracturing wastewater by aluminum electrocoagulation. J. Colloid Interface Sci. 2015, 458, 103-111.

586 587 588

18. El Samrani, A.G.; Lartiges, B.S.; Montarges-Pelletier, E.; Kazpard, V.; Barres, O.; Ghanbaja, J. Clarification of municipal sewage with ferric chloride: the nature of coagulant species. Water Res. 2004, 38 (3), 756-768.

589 590

19. Goodwill, J.E.; Jiang, Y.; Reckhow, D.A.; Gikonyo, J.; Tobiason, J.E. Characterization of Particles from Ferrate Preoxidation. Environ. Sci. Technol. 2015, 49 (8), 4955-4962.

591 592 593

20. Niemeyer, J.; Chen, Y.; Bollag, J.M. Characterization of humic acids, composts, and peat by diffuse reflectance fourier-transform infrared-spectroscopy. Soil Science Society of America Journal 1992, 56 (1), 135-140.

594 595 596

21. Guan, X.; Shang, C.; Chen, G. ATR-FTIR investigation of the role of phenolic groups in the interaction of some NOM model compounds with aluminum hydroxide. Chemosphere 2006, 65 (11), 2074-2081.

597 598

22. Menges, F. Spekwin32 - Optical Spectroscopy Software, Version v1.72.0; 2016; http://www.effemm2.de/spekwin/.

599 600 601

23. Cornelissen, A.; Burnett, M.G.; McCall, R.D.; Goddard, D.T. The structure of hydrous flocs prepared by batch and continuous flow water treatment systems and obtained by optical, electron and atomic force microscopy. Water Sci.Technol. 1997, 36 (4), 41-48.

602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618

24. Fischer, E.R.; Hansen, B.T.; Nair, V.; Hoyt, F.H.; Dorward, D.W. Scanning Electron Microscopy. In: Current Protocols in Microbiology; John Wiley & Sons, Inc., 2012. doi: 10.1002/9780471729259.mc02b02s25 25. Vadasarukkai, Y. S.; Gagnon, G. A. Characterization, fate and transport of floc aggregates in full-scale flocculation tanks. Environmental Science-Water Research & Technology 2016, 2 (1), 223-232. 26. Wong, K. B.; Piedrahita, R. H. Settling velocity characterization of aquacultural solids. Aquacult. Eng. 2000, 21 (4), 233-246. 27. Tassinari, B.; Conaghan, S.; Freeland, B.; Marison, I. W. Application of turbidity meters for the quantitative analysis of flocculation in a jar test apparatus. J. Environ. Eng. 2015, 141 (9), 04015015-1-04015015-8. 28. Myneni, S. C. B.; Brown, J. T.; Martinez, G. A.; Meyer-Ilse, W. Imaging of humic substance macromolecular structures in water and soils. Science 1999, 286 (5443), 1335-1337. 20 ACS Paragon Plus Environment

Page 20 of 31

Page 21 of 31

619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662


29. Sieliechi, J. M.; Lartiges, B. S.; Kayem, G. J.; Hupont, S.; Frochot, C.; Thieme, J.; Ghanbaja, J.; de la Caillerie, J. B. D.; Barres, O.; Kamga, R.; Levitz, P.; Michot, L. J. Changes in humic acid conformation during coagulation with ferric chloride: Implications for drinking water treatment. Water Res. 2008, 42 (8-9), 2111-2123. 30. Gorczyca, B.; Ganczarczyk, J. Structure and porosity of alum coagulation flocs. Water Qual. Res. J. Can. 1999, 34 (4), 653-666. 31. Duan, J.M.; Gregory, J. Coagulation by hydrolysing metal salts. Advances in Colloid and Interface Science 2003, 100, 475-502. 32. Coufort, C.; Bouyer, D.; Line, A. Flocculation related to local hydrodynamics in a TaylorCouette reactor and in a jar. Chemical Engineering Science 2005, 60 (8-9), 2179-2192. 33. Gu, B. H.; Schmitt, J.; Chen, Z.; Liang, L. Y.; Mccarthy, J. F. Adsorption and desorption of different organic-matter fractions on iron-oxide. Geochim. Cosmochim. Acta 1995, 59 (2), 219229. 34. Guan, X.; Shang, C.; Chen, G. ATR-FTIR investigation of the role of phenolic groups in the interaction of some NOM model compounds with aluminum hydroxide. Chemosphere 2006, 65 (11), 2074-2081. 35. Vinkler, P.; Lakatos, B.; Meisel, J. Infrared spectroscopic investigations of humic Substances and their metal-complexes. Geoderma 1976, 15 (3), 231-242. 36. Rout, K.; Mohapatra, M.; Anand, S. 2-Line ferrihydrite: synthesis, characterization and its adsorption behaviour for removal of Pb(II), Cd(II), Cu(II) and Zn(II) from aqueous solutions. Dalton Transactions 2012, 41 (11), 3302-3312. 37. Liu, C., Huang, P. M. The influence of catechol humification on surface properties of metal oxides. In Humic substances structures, models and functions; Ghabbour, E. A.; Davies, G., Eds.; The Royal Society of Chemistry: Cambridge, UK, 2001; pp 253-270. 38. Peak, D.; Ford, R. G.; Sparks, D. L. An in situ ATR-FTIR investigation of sulfate bonding mechanisms on goethite. J. Colloid Interface Sci. 1999, 218 (1), 289-299. 39. Amir, S.; Jouraiphy, A.; Meddich, A.; El Gharous, M.; Winterton, P.; Hafidi, M. Structural study of humic acids during composting of activated sludge-green waste: Elemental analysis, FTIR and C-13 NMR. J. Hazard. Mater. 2010, 177 (1-3), 524-529. 40. Webster, J. G.; Swedlund, P. J.; Webster, K. S. Trace metal adsorption onto an acid mine drainage iron(III) oxy hydroxy sulfate. Environ. Sci. Technol. 1998, 32 (10), 1361-1368.



663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697

41. Pivokonsky, M.; Bubakova, P.; Pivokonska, L.; Hnatukova, P. The effect of global velocity gradient on the character and filterability of aggregates formed during the coagulation/flocculation process. Environ. Technol. 2011, 32 (12), 1355-1366.

698 699 700 701 702 703 704

52. Lee, S.Y.; Gagnon, G.A. Growth and structure of flocs following electrocoagulation. Separation and Purification Technology 2016, 163, 162-168.

42. McNaught, H.; Trenholm, R.; Claassen, D.; Scott, C.; Kodweis, G.; Vanderford, B.J.; Rexing, D.; Wert, E.C. No-lead brass service valve performance in Colorado River water. Journal American Water Works Association. 2015, 107 (8), E416-E424. 43. Lin, J.; Huang, C.; Dempsey, B. A.; Hu, J. Fate of hydrolyzed Al species in humic acid coagulation. Water Res. 2014, 56, 314-324. 44. Rahman, M. S.; Gagnon, G. A. Bench-scale evaluation of drinking water treatment parameters on iron particles and water quality. Water Res. 2014, 48, 137-147. 45. Benjamin, M.M.; Lawler, D.F. Water quality engineering: physical/chemical treatment processes; John Wiley & Sons: New Jersey. 2013. 46. Kan, C.C.; Huang, C.P.; Pan, J.R.S. Time requirement for rapid-mixing in coagulation. Colloids and Surfaces a-Physicochemical and Engineering Aspects 2002, 203(1-3), 1-9. 47. Rossini, M.; Garrido, J. G.; Galluzzo, M. Optimization of the coagulation-flocculation treatment: Influence of rapid mix parameters. Water Res. 1999, 33 (8), 1817-1826. 48. Bridgeman, J.; Jefferson, B.; Parsons, S. Assessing floc strength using CFD to improve organics removal. Chemical Engineering Research & Design 2008, 86 (8A), 941-950. 49. Hopkins, D.C.; Ducoste, J.J. Characterizing flocculation under heterogeneous turbulence. Journal of Colloid and Interface Science 2003, 264 (1), 184-194. 50. Garland C.; Weber-Shirk M.; Lion L.W. Revisiting Hydraulic Flocculator Design for Use in Water Treatment Systems with Fluidized Floc Beds. Environmental Engineering Science 2016, ahead of print. doi:10.1089/ees.2016.0174. 51. Li, D.H.; Ganczarczyk, J. Size distribution of activated-sludge flocs. Research Journal of the Water Pollution Control Federation 1991, 63 (5), 806-814.

53. Amirtharajah, A.; Dennett, K. E.; Studstill, A. Ferric chloride coagulation for removal of dissolved organic matter and trihalomethane precursors. Water Sci. Technol. 1993, 27 (11), 113121.

705


Page 22 of 31

Page 23 of 31

706 707


For Table of Contents Only

708



Page 24 of 31

Table 1. Raw water quality for Kelly Lake

Water Quality Parameters pH Turbidity Total Organic Carbon Dissolved Organic Carbon UV absorbance at 254 nm SUVA Zeta Potential Elemental Analysis Magnesium (Mg) Iron (Fe) Sodium (Na) Aluminum (Al) Calcium (Ca) Potassium (K)

Values 6.0± 0.3 0.4± 0.1 NTU 3.8± 0.3 mg/L 3.6± 0.3 mg/L 0.15±0.02 cm-1 4.0± 0.2 L/mg.m -24± 0.4 mV Concentration (µg/L) 556 ± 2.7 330 ± 2.4 3900 ± 2.8 111 ± 1.0 1470 ± 11.3 196 ± 5.4

a

b

c

Figure 1. TEM images of microfloc formed at a G-value of 175 s-1 for the optimal coagulant concentration of 0.33 mM of Fe and pH = 4.80. Floc sizes shown in the insert were measured using ImageJ software v1.49. ACS Paragon Plus Environment

Page 25 of 31


Figure 2. Illustration of floc size distributions at a flocculation mixing intensity (G-value) of 46 s-1 for the optimal coagulant concentrations. The displayed histogram was estimated from three replicate experiments at pH 4.80 and 6.85, respectively.

ACS Paragon Plus Environment


Page 26 of 31

Table 2. Peak intensity ratios calculated from the infrared spectra at various coagulation mixing intensities (G-value) and the corresponding energy consumed in the coagulation experiments at (a) 0.33 mM of Fe and pH = 4.80 (b) 0.66 mM of Fe and pH= 6.85.

G-values (s-1) Energy consumed at 120 MLD (kW-hr/day) Assignment of peak intensity ratios:

46

175

452

1450

2

27

183

1879

(a) 0.33 mM of Fe and pH = 4.80

Adsorbed caroboxylate(as ) Structural O  H

I1559 I 3177

1.77

1.50

1.57

1.51

Adsorbed caroboxylate(s ) Structural O  H

I1375 I 3177

1.61

1.32

1.46

1.33

Adsorbed SO4 Structural O  H

I1039 I 3177

2.46

1.72

1.99

1.71

Assignment of peak intensity ratios:

(b) 0.66 mM of Fe and pH = 6.85

Adsorbed caroboxylate(as ) Structural O  H

I1558 I 3177

1.29

1.44

1.34

1.32

Adsorbed caroboxylate(s ) Structural O  H

I1374 I 3177

1.11

1.23

1.13

1.11

Adsorbed SO4 Structural O  H

I1044 I 3177

1.41

1.37

1.38

1.07



Normalized weight (%)

Page 27 of 31

Elements Figure 3. Illustration of elemental composition of floc aggregates using SEM-EDS analysis for pH 4.80 (top panel) and 6.85 (bottom panel).


Equivalent volumetric diameter (μm)

Energy consumed (E), kW-h/day


Coagulation- mixing intensity (G-value) (s-1) Figure 4. Variation in floc size distributions with the change in the coagulation mixing intensities (G-value) and the corresponding energy consumed for the coagulation experiments conducted at 0.33 mM of Fe and pH = 4.80 (top panel) and 0.66 mM of Fe and pH= 6.85 (bottom panel). ACS Paragon Plus Environment

Page 28 of 31

Page 29 of 31


Table 3. Influence of the coagulation mixing intensities (G-value) and the corresponding energy consumed on the sedimentation rate as calculated from the slope of turbidity-time regression line.

G-values (s-1) Energy consumed at 120 MLD (kW-hr/day)

0

24

112

452

1450

0

0.5

11

182

1879

Sedimentation rate (NTU/min):*

(a) 0.33 mM of Fe and pH = 4.80 -0.32

Sedimentation rate (NTU/min):*

-0.17

-0.10

-0.18

(b) 0.66 mM of Fe and pH = 6.85 -0.52

*

-0.17

-0.33

-0.32

-0.25

-0.61

The slope of the regression line was calculated between the time period 0 (x1) and 30 (x2) minutes of the settling period.


Residual Turbidity (NTU)


Coagulation- mixing intensity (G-value) (s-1) Figure 5. Influence of the coagulation mixing intensities (G-value) on settling kinetics for the coagulation experiments conducted at 0.33 mM of Fe and pH = 4.8 (top panel) and 0.66 mM of Fe and pH= 6.85 (bottom panel). ACS Paragon Plus Environment

Page 30 of 31

Page 31 of 31


For Table of Contents Only


Influence of the Mixing Energy Consumption Affecting Coagulation

Recommend Documents