Flocculation Optimization of Orthophosphate with ... - ACS Publications

Subscriber access provided by University of Newcastle, Australia

Article

Flocculation optimization of orthophosphate with FeCl3 and alginate using the Box-Behnken response surface methodology Henry K. Agbovi, and Lee D. Wilson Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.6b04765 • Publication Date (Web): 06 Mar 2017 Downloaded from http://pubs.acs.org on March 9, 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.

Industrial & Engineering Chemistry Research 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 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Coagulation and flocculation of orthophosphate 338x190mm (96 x 96 DPI)

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1 2 3 4 5

Flocculation optimization of orthophosphate with FeCl3 and alginate using the Box-Behnken response surface methodology Henry K. Agbovi1 and Lee D. Wilson1*

6 7 8 9

1

University of Saskatchewan, Department of Chemistry, 110 Science Place, Thorvaldson Building (Room 165), Saskatoon, SK., S7N 5C9, Saskatoon, Saskatchewan, CANADA

10 11

*

Corresponding Author: L. D. Wilson, Email: [email protected]

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 1 ACS Paragon Plus Environment

Page 2 of 36

Page 3 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


36

Abstract

37

A coagulation-flocculation process was employed to remove orthophosphate (Pi) in aqueous

38

media using a ferric chloride (FeCl3) and alginate flocculant system. Jar tests were conducted

39

and the response surface methodology (RSM) was used to optimize the Pi removal variables. The

40

Box-Behnken design was used to evaluate the effects and interactions of four independent

41

variables: pH, FeCl3 dose, alginate dose, and settling time. The RSM analysis showed that the

42

experimental data followed a quadratic polynomial model with optimum conditions at pH 4.6,

43

[FeCl3] = 12.5 mg·L-1, [alginate] = 7.0 mg·L-1 and a 37 minute settling time. Optimum conditions

44

led to a Pi removal of 99.6% according to the RSM optimization, in good agreement with

45

experimental removal (99.7±0.7%), at an initial concentration of 10.0 mg Pi/L. The isotherm

46

adsorption data at the optimized conditions were analyzed by the pseudo-first-order (PFO) and

47

pseudo-second-order (PSO) kinetic models, and several isotherms models (Langmuir, Freundlich

48

and Sips). The PFO kinetic model and Langmuir isotherm model yielded the best fit to the

49

isotherm results. The maximum adsorption capacity of the flocculant system was 83.6 mg·g-1.

50

The flocculation process followed electrostatic charge neutralization and an ion-binding

51

adsorption mechanisms.

52 53

Keywords: Coagulation; Flocculation; Orthophosphate; Alginate; Ferric chloride; Response

54

surface methodology; Box-Behnken Design

55 56

2 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 36

57 58

1. Introduction

59

Phosphorus is an essential plant nutrient for growth and development, where it exists in

60

aquatic systems as orthophosphate, polyphosphate or organophosphate. Major sources of

61

phosphate in water and wastewater originate from human, domestic and industrial waste, as well

62

as fertilizer effluent run-off from agricultural activities1. Municipal wastewater may contain ca.

63

10 mg·L–1 to 20 mg·L–1 of total phosphorus2. Although phosphate is an essential micronutrient in

64

water, elevated levels lead to eutrophication and excessive algae growth in water3. To address

65

eutrophication, phosphorus effluent is regulated to levels below 0.05 mg·L–1 4. In Ontario,

66

Canada, a total phosphate limit of 1.0 mg·L–1 is usually imposed on industry (e.g., fertilizer,

67

detergent, municipal and institutional sewage treatment processing) discharge to surface water to

68

prevent damage to aquatic ecosystems.5 Thus, water treatment technology is required to

69

drastically reduce phosphate in municipal and industrial wastewater effluent. The removal of

70

orthophosphate (Pi) in wastewater has been examined by different methods: adsorption6–9,

71

biological

72

precipitation14–16.

removal10,11,

electro-coagulation12,

membrane

processes13

and

chemical

73

Among these various methods, coagulation-flocculation has attracted attention for primary

74

and secondary water treatment due to its efficiency, simplicity and relatively low cost17.

75

Conventional coagulation-flocculation involves addition of commercially available coagulants

76

and/or flocculants such as alum, lime, ferric chloride or ferric sulphate to the wastewater to

77

precipitate the contaminants. The solid by-products from the process are then removed by

78

sedimentation and/or filtration18. Among the inorganic coagulants, ferric chloride has a

79

widespread application because of its effectiveness in contaminant removal from water. Addition 3 ACS Paragon Plus Environment

Page 5 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


80

of ferric chloride to water results in a hydrolysis reaction to form iron (III) chloride precipitates.

81

The iron (III) chloride solid particles have a high affinity towards many contaminants in

82

wastewater such as phosphate, arsenic and sulphide species. Hsu19 has shown that Fe3+ has a

83

stronger affinity for precipitating PO43- and a stronger hydrolyzing power than Al3+. In contrast

84

with aluminium-based coagulants such as alum, ferric chloride has no significant adverse health

85

effects in wastewater treatment20.

86

Inorganic salts are widely used to precipitate phosphate due to their cost effectiveness and

87

ready availability. However, mineral salts pose several disadvantages, such as large dosage

88

requirements, ineffectiveness at lower temperatures, pH dependent

89

production and human health concerns21. Even though the inorganic coagulants have several

90

disadvantages as coagulants, studies have shown that when they are used as coagulants together

91

with other polymeric flocculants or coagulant aids, the disadvantages can be offset or eliminated.

92

A smaller dosage of the inorganic coagulants is realized in the case of binary or ternary

93

coagulant-flocculant systems. In addition, the inorganic coagulants can function synergistically

94

with the biopolymer materials, which contribute to reduced sludge volume and greater

95

mechanical stability of the sludge byproducts from the flocculation process14,22–26. The use of

96

synthetic

97

poly(diallyldimethyl) ammonium chloride and poly(styrene) derivatives for the removal of

98

contaminants in wastewater are known16. Polymer flocculants have advantages over the use of

99

inorganic salts that include greater contaminant removal efficiency, ability to form large flocs

100

with high density, higher mechanical strength, and favourable settling properties14. Polymers

101

possess useful properties over a wide pH range and yield less sludge without metal (aluminium)

102

residuals whilst maintaining the alkalinity of water. By contrast, synthetic polymers are non-

polymer

flocculants

and/or

coagulant

aids,


such

performance, sludge

as

polyacrylamide,


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

103

biodegradable, higher relative cost, and pose potential biological toxicity27. The development of

104

coagulant aid and biopolymer flocculants in binary systems have several advantages: cost

105

efficiency, biodegradability, biocompatibility, reduced toxicity, with minimal production of

106

pollutants28. The application of biopolymer flocculants such as alginate, chitosan, cellulose and

107

biomass materials offer benefits to water treatment based on their renewable nature,

108

biodegradability, innocuous waste production, and tunable efficiency14,27.

109 110 111 112 113 114

Figure 1: Chemical structure of alginate, where m and n denote the degree of polymerization.

115

Alginate has significant potential as a coagulant aid for wastewater treatment. It is a linear

116

biopolymer obtained from marine brown algae (seaweed) and capsular polysaccharides typically

117

found in soil bacteria29. Alginate is an anionic block copolymer above its pKa value (3.5), that is

118

composed of β-D-mannuronate (M)and α-L-guluronate (G) units linked by β-1→4 and α-1→4

119

glycosidic bonds, as shown in Figure 130. Alginate contains carboxylic acid groups on each

120

monomer unit that possesses diverse functionality in water treatment such as chelation with

121

reactive polyvalent metal cations31. Alginate has been studied as a flocculant in wastewater

122

treatment, where Wu et al. have studied the effect of alginate as a coagulant aid in conjunction

123

with alum23. The coagulation performance of alginate with metal salts such as ferric chloride,

124

alum, titanium tetrachloride and polyaluminium chloride was reported by Zhoa et al.25,26.

125

Devrimic et al.31 have used calcium alginate to reduce turbidity in water, while alginate was

126

studied for the color removal in textile effluent as a coagulant for dye species22. These reports 5 ACS Paragon Plus Environment

Page 6 of 36

Page 7 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


127

indicate that combined use of alginate and metal ions have synergetic effects on contaminant

128

removal efficacy.

129

The algal form of alginate is a coagulant aid for wastewater treatment in the removal of

130

phosphate has not been reported in the open literature. This study aims to optimize the

131

coagulation-flocculation potential of alginate in conjunction with ferric chloride for the removal

132

of phosphate along with the Box-Behnken Design (BBD) and the response surface methodology

133

(RSM). The BBD method is an efficient and widely used approach in experimental design32.

134

RSM is aimed at the factorial design for determining the optimal operational conditions for a

135

given system to meet regulatory requirements33. In addition, binary-coagulant systems such as

136

alginate with ferric chloride have potential cost reduction advantages due to the reduced dosage,

137

enhancement of the flocculation activity, and potential risk reduction to human health.

138 139

2. Materials and methods

140

2.1 Materials

141

All chemicals were of analytical reagent (AR) grade. Hydrated ferric chloride (FeCl3·6H2O),

142

anhydrous monobasic potassium phosphate (KH2PO4), NaOH, HCl, vanadate molybdate reagent

143

and medium viscosity sodium alginate were purchased from Sigma-Aldrich (Oakville, ON.

144

Canada). The coagulation-flocculation experiments were carried out using a six-plate

145

conventional Jar-test apparatus, originally developed by Phipps & Bird with six parallel stirrers.

146 147 148 149 6 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 36

150 151 152 153 154

Table 1: Levels of each variable for the Box-Behnken experimental design. Independent variable pH Alginate dose Settling time FeCl3 dose

Unit

Symbol Low (-1) –1

mg·L minutes mg·L–1

x z m n

2 0.5 10 10

Coded level Middle (0) High (+1) 6.1 7.8 35 12

10 15 60 15

155 156 157 158

2.2 Coagulation-flocculation experiment

159

Standard orthophosphate (Pi), ferric chloride and alginate stock solutions were prepared by

160

dissolving in deionised and distilled water. A calibration curve of Pi was obtained using a

161

vanadate-molybdate colorimetric method at a wavelength of 420 nm at pH 6.534. The

162

coagulation-flocculation was performed, based on the experimental design matrix obtained from

163

the BBD and the RSM, using a program-controlled conventional jar test apparatus with six-2 L

164

jars and stirrers. Approximately, 1 L of simulated Pi containing sample was added to the jar

165

tester and 5.0 ml of the Pi solution was sampled and used for UV-Visible analysis by adding 1.0

166

ml of vanadate-molybdate reagent. After addition of the reagent to the Pi sample, a yellow

167

coloured complex developed after 20 minutes before recording UV-vis absorbance. The pH of

168

the Pi solution was adjusted using 0.1 M NaOH or 0.1 M HCl to the desired value according to

169

the experimental design. The coagulation-flocculation experiment was carried out by adopting 7 ACS Paragon Plus Environment

Page 9 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


170

the procedure from a previous study35,14. A pre-determined amount of the coagulant (ferric

171

chloride) was added to the solution, followed by rapid stirring for 3 minutes at 295 rpm.

172

Thereafter, the stirring rate was reduced to 25 rpm for 20 minutes. During this period, the

173

alginate (flocculant) was added within the first 5 minutes. The stirring was then stopped where

174

different settling times (10 – 60 minutes) were tested. Following the settling time, a 5.0 ml

175

aliquot of the solution was sampled from the top and prepared for UV-vis analysis for Pi. The

176

colorimetric determination of Pi using the vanadate-molybdate method involved absorbance

177

using a double beam UV-vis spectrophotometer at λmax = 420 nm. The pH of all the solutions

178

was measured after the appropriate settling period. The aqueous concentration of Pi before (co)

179

and after (ce) the coagulation-flocculation experiment was determined, as described above. The

180

experiments were performed in triplicate, where the average value and the standard deviation are

181

reported. The orthophosphate removal (Pi

182

calculated by equations (1) and (2), respectively34.

183

% Pi r emoval =

184

qe =

removal,

%) and the removal capacity (Pi; mg·g-1) was

co − ce x 100% co

(1)

(c0 - ce ) xV m

(2)

185

Here, c0 and ce are the initial and equilibrium concentrations of Pi in solution (mg·L−1), V is the

186

volume (L) and m is the mass of the adsorbent (g).

187 188

2.3 Box-Behnken experimental design

189

The traditional method of coagulation-flocculation involves changing one factor at a time

190

and requires many experiments which may be time-consuming, often leading to low optimization

191

efficiency. To address this problem, the design of experiment (DOE) was used to study the effect



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 36

192

of variables and their responses using a minimum number of experiments. RSM is a collection of

193

statistical and mathematical methods which are useful for developing, improving, and optimizing

194

processes36–38. The selection of an adequate experimental design is a key consideration in

195

experimental optimization. The Box-Behnken Design (BBD) method was employed to obtain the

196

optimum Pi,removal. The BBD is an independent, rotatable quadratic design with no embedded

197

factorial or fractional factorial points where the variable combinations are at the mid-points of

198

the edges of the variable space and at the center39. The BBD requires fewer treatment

199

combinations than the central composite design, especially in cases with three or four factors,

200

and is less computationally demanding to perform. BBD allows efficient estimation of the first-

201

and second-order coefficients and does not have axial points. Thus, it is certain that all design

202

points fall within a safe operating zone. The BBD approach ensures that not all factors are set at

203

their high levels at the same time. This affords identification of significant effects of interaction

204

for batch studies39–42.

205

Preliminary experiments indicate that important variables affecting Pi,removal include ferric

206

chloride dosage, alginate dosage, pH and settling time. These variables were evaluated to

207

optimize Pi,removal. FeCl3 dosage varied between 10 mg·L-1 and 15 mg·L-1, alginate dosage (0.5

208

mg·L-1 to 15 mg·L-1), pH (2 to 10), and settling time (10 to 60 minutes). In Table 1, the

209

experimental design had four variables (x, z, m, and n), each at three levels, coded as -1, 0 and

210

+1, for low, middle and high values, respectively. Twenty-seven experiments were carried out

211

according to the statistical matrices developed by the RSM, in order to account for variability of

212

the independent variables on Pi,removal. The experimental data was fit using a non-linear

213

regression method using a second order polynomial to identify significant coefficient terms. The

214

application of RSM provides an empirical relationship between the response function and the


Page 11 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


215

independent variables. The quadratic response model is based on all linear terms, square terms,

216

and linear interaction terms, according to equation (3)43. Y = b° + ∑ bi xi + ∑ bii xii 2 + ∑ bij xij

217

(3)

218

Here, Y is the predicted response (Pi,removal efficiency) bo is the model constant, bi is the linear

219

coefficient, bii is the quadratic effect of the input factor xii, bij is the linear interaction effect

220

between the input factors xi and xj.

221

The response function coefficients were determined by regression using the experimental data

222

and the Statgraphics XVII-X64 regression program. The response functions for phosphate

223

removal (%) was approximated by the standard quadratic polynomial equation in equation (4).

224

Equation (4) describes the regression model of the system, including the interaction terms.43

225

Y = b0 + b1 x + b2 z + b3m + b4 n + b11 x 2 +b12 xz + b13 xm + b14 xn + b22 z 2 + b23 zm + b23 zn + b33m 2 + b34 mn + b33n 2

(4)

226

Y is the predicted response, phosphate removal (%); x, z, m and n are the coded levels of the

227

independent variables: pH, alginate dosage, settling time and FeCl3 dosage, respectively. The

228

regression coefficient, b0 denotes the intercept term; b1, b2, b3 and b4 represent the linear

229

coefficients; b12, b13, b14, b23, b24, b34 represent the interaction coefficients and b11, b22, b33 and

230

b44 denote the quadratic coefficients. Analysis of variance (ANOVA) was employed to perform

231

diagnostic tests on the adequacy of the proposed model. The ANOVA test estimates the

232

suitability of the response functions and the significance of the effects of the independent

233

variables.

234 235

3. Results and Discussion

236

3.1 Box-Behnken Analysis



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 36

237

The Box–Behnken Design (BBD) for statistical analysis and the response surface

238

methodology (RSM) was employed to investigate the effects of the four independent variables

239

on the response function, and determination of the optimal conditions to maximize phosphate

240

removal (%). The independent variables were pH (x), alginate dosage (z), settling time (m) and

241

FeCl3 dosage (n). The experimental design involved four variables (x, z, m, and n), each at three

242

levels, coded as -1, 0 and +1, for low, middle and high values, as shown in Table 1. The

243

optimization procedure comprises study of the response of the statistically designed

244

combinations, estimating the coefficients by fitting the experimental data to the response

245

functions, predicting the response of the fitted model and goodness-of-fit of the model39,40. The

246

results for the Pi removal efficiency are listed in Table S1. The coefficients of the response

247

function for dependent variables were determined by correlating the experimental data with the

248

response functions using Statgraphics XVII-X64 regression program. Different response

249

functions and the respective coefficients are described by equation (5).

250

Y = -95.4 + 17.2x + 0.260z + 0.611m - 1.53- 1.53x 2 + 0.172xz - 0.0022xm - 0.329xn - 0.0982z 2 - 0.00317zm + 0.0361zn -0.0117m 2 + 0.0234mn -0.8746n 2

(5)

251

Statistical testing of the model was performed with the Fisher’s statistical test for analysis of

252

variance (ANOVA). The ANOVA test for the Pi removal is presented in Table 2. The ANOVA

253

test estimates the suitability of the response functions and the significance of the effects of the

254

independent variables. Student’s t-test and p-values were checked to determine the statistical

255

significance of the BBD-RSM variables and their interactions (combination of two codes or

256

variables) at variable probability values44. A large value of F, according to ANOVA test, indicate

257

that most of the variables in the response can be explained by the regression equation, and

258

probability value (p values) less than 0.05 are considered to be statistically significant. The

259

ANOVA results indicate that the second order polynomial equation adequately represents the 11 ACS Paragon Plus Environment

Page 13 of 36

260

actual relationship between the response function (Pi removal efficiency) and the independent

261

variables (FeCl3 dose, alginate dose, pH and setting time). The F-value model was 250.82,

262

implied the model was significant. Values of p less than 0.05 indicated that the model terms were

263

significant, and values greater than 0.05 were considered statistically insignificant.

264 265 266 267 268 269 270 271 272 273

Negative

Standardized effect estimate

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Positive

p = 0.05

274 275 276

Figure 2: Pareto chart showing the standardized effects of variables on phosphate removal efficiency.

277 278

Figure 2 is a Pareto plot that depicts the standardized effect variables based on the t values of

279

the independent variables, along with quadratic and interaction effects. The length of each bar in

280

Figure 2 represents the standardized effect of that variable on the response function (Pi,removal),

281

and the alpha level of 0.05 was employed to evaluate the statistical significance. Values of p less



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 36

282

than 0.05 indicate a statistically significant variable effect. If the p-value is greater than 0.05, the

283

variable effect was considered to be statistically insignificant.

284

The selected 1st-order and 2nd-order variables were shown to be statistically insignificant

285

except pH (x), x2, according to the Pareto plot in Figure 2 and Table 2. Also, for the interactive

286

variables, xz, xn, and mn were found to be insignificant, whereas; xm, zn and zm are statistically

287

significant. The insignificant variables were identified because the Pareto bars for these variables

288

are below the reference line, as shown in Figure 2, and their p values are relatively high (p >

289

0.05), as listed in Table 2. The quadratic term for pH (x2) is the most significant component of

290

the regression model for the present application. The pH was also identified by the linear terms

291

among the independent variables that had a significant effect on the Pi,removal efficiency.

292

Moreover, each bar in the plot represents an absolute standardized effect. Values in grey display

293

positive effects while those in black have negative effects on the response variable.

294 295 296

Table 2: Analysis of variance (ANOVA) results of phosphate removal efficiency as a function of pH, FeCl3 dosage, alginate dosage and settling time. Source x z m n xx xz xm xn zz zm zn mm mn nn Total error Total (corr.)

Sum of squares 3110 8.234 18.78 1.248 3212 100.2 0.198 43.23 142.1 1.323 1.716 283.1 8.585 159.4 24.23 6609

Degree of freedom 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 12 26

Mean square 3110 8.234 18.78 1.248 3212 100.2 0.198 43.23 142.1 1.323 1.716 283.1 8.585 159.4 2.019

F-Ratio

p-value

1540 4.080 9.300 0.620 1591 49.63 0.100 21.41 70.40 0.660 0.850 140.24 4.250 78.93

0.000 0.066 0.010 0.447 0.000 0.000 0.760 0.001 0.000 0.434 0.375 0.000 0.062 0.000


Remarks Significant Insignificant Significant Insignificant Significant Significant Insignificant Significant Significant Insignificant Insignificant Significant Insignificant Significant

Page 15 of 36

297 298

The goodness-of-fit by the regression model was evaluated by coefficients of R2, adjusted

299

R2, standard error of estimate and mean absolute error. The R2 statistic indicates that the

300

model as fitted explains 99.6% of the variable Pi,removal efficiency. The adjusted R2

301

statistic (99.2%) is more suitable for comparing models with variable independent

302

variables.

303

residuals to be 1.42. The mean absolute error (MAE) was 0.77 and is the average value

304

of the residuals. The accuracy of the regression model was checked by the parity and

305

residual plots in Figure 3. Figure 3(a) illustrates the plots of the predicted Pi removal (%)

306

versus measured values. Most of the data are distributed near the straight line where the

307

measured and predicted Pi,removal are the same, and indicates that the regression model is

308

able to predict these values. These plots provide information on the fit criteria contained

309

in the residuals. In Figure 3(b), the maximum deviation between the predicted and the

310

measured Pi,removal is below 3.0% and reveals a good correlation between the predicted

311

and the measured values.

The standard error of the estimate shows the standard deviation of the

100

3

R² = 0.9963 90

Residual error (%)

Predicted Pi (%) removal

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


80 70

1 0

-1

60

-2

50

-3

40 40

312

2

50

60

70

80

90

40

100

50

60

70

80

90

100

Predicted Pi (%) removal

Oberved Pi (%) removal



313 314 315 316 317

Figure 3: (a) Parity plot showing the correlation between the predicted and experimental values of phosphate removal efficiency (b) Percent error as a function of predicted phosphate removal.

318

Figure 4 displays a plot of the main effects and the independent variables for the Pi,removal

319

efficiency. The main effect plots are used to depict a preliminary conclusion about effects of the

320

flocculation variables. The role of independent variables; pH, FeCl3 dose, alginate dose, and

321

settling time on the response variable are shown in Figure 4. The effect of pH in Figure 4(a) is

322

significant on the Pi,removal, where optimum removal was obtained at 4.7, with a sharp decrease in

101

99

91

81

71

61

(b) Pi removal (%)

(a)

Pi removal (%)

51

97

95

93

91

89 2.0

10.0

10.0

pH 99

60.0

Settling time (min) 99

(c)

97

(d) Pi removal (%)

323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350

3.2 Main effect plots

Pi removal (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 36

95

93

91

97

95

93

91 0.5

15.0

10.0

Alginate dose (mg·L–1)

15.0

–1

FeCl3 dose (mg·L )

Figure 4: Main effects plot of (a) initial pH, (b) settling time (c) alginate dose and (d) FeCl3 dose on phosphate removal efficiency.


Page 17 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


351

the slope as the pH increases. As the FeCl3 and alginate dosage increases (Figure 4 c-d), the

352

removal efficiency increases until an optimum dose is obtained at 12.5 mg·L–1 and 7.5 mg·L–1,

353

respectively. An increase in their respective dosages beyond the optimum level leads to a

354

decrease in the response variable. The settling does not appear to have any considerable

355

influence on the response function, since there is no significant variation in the Pi,removal

356

efficiency with time, as depicted in Figure 4(b). Employing the main effect to evaluate the

357

response function is problematic because the interaction between different independent factors

358

may affect the response variable. Hence, the effect of the interaction variables are discussed in

359

the next section.

360 361 362 363

3.3 The interactive effects of variables on phosphate flocculation

364

The three-dimensional (3D) response surface plots and 2D contour plots in Figure S1 were

365

obtained from the model-predicted response by keeping two independent variables constant at

366

the optimal level while varying the other two terms within the experimental conditions. The

367

response surface is a graphical representation of the regression equation for visualization of the

368

relationship between the response and experimental levels of each variable45. These plots are

369

useful to assess the interactive relationship between the independent variables and the response

370

variable. The obvious peak in the response surfaces indicate that the optimal conditions were

371

exactly located inside the design boundary. The elliptical contour plots show significant

372

interactive effects on Pi,removal efficiency between any two variables. In other words, there are

373

significant interactive effects on the response variable between FeCl3 and alginate dosage, FeCl3

374

dosage and pH, FeCl3 dosage and settling time, alginate dosage and settling time, pH and settling



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

375

time, as well as alginate dosage and pH. The interactive effects between different variables are

376

presented and discussed below.

377

The response surface and contour plots for the interaction between pH and FeCl3 dose at the

378

optimal levels of alginate dose and settling time are presented in Figure S1(i). Between pH 1.0 to

379

4.7, the Pi,removal increased from 62.5 to 99%. However, as the pH increases further, the removal

380

efficiency decreases significantly. At pH of 10 and FeCl3 dose of 15.0 mg·L–1, the Pi,removal is

381

55%. At pH 2 and FeCl3 dose of 12 mg·L–1, the Pi,removal was 77.5%. This is expected because pH

382

plays an important role in coagulation-flocculation process. Charge neutralization on hydrolysis

383

products and the precipitation of metal hydroxides can be controlled by pH variations. Phosphate

384

has anionic surface charge, where hydrolysis products of the ferric chloride can neutralize Pi

385

charge. Therefore, charge neutralization is a likely mechanism for Pi,removal, where adsorption and

386

sweeping may be considered as possible removal mechanisms due to the presence of alginate. In

387

acidic media at low pH, protonation may occur, resulting in reduced charge density, which leads

388

to self-aggregation of Pi where less coagulant is required. For instance, at fixed pH, an increase

389

of the FeCl3 dose leads to greater Pi,removal until an optimum dose of 12.5 mg·L–1 which decreases

390

gradually as the FeCl3 dose increases further.

391

Pi,removal was studied as a function of pH and alginate dose is presented in Figure S1(iia) and

392

S1(iib) as 3D and 2D plots, respectively. The plots indicate that the optimal regions for the two

393

interacting variables are located within the design boundary. At constant alginate dosage such as

394

6.0 mg·L–1, the Pi,removal increases until an optimum pH condition is obtained. As the pH

395

increases beyond the optimum level, the Pi,removal decreases significantly. By comparison, the

396

alginate dose increased at constant pH where the Pi,removal increases steadily with an optimum

397

dose of 7.75 mg·L-1 and decreases as the alginate dose increases beyond the optimal level. As pH 17 ACS Paragon Plus Environment

Page 18 of 36

Page 19 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


398

and alginate dose were increased, the Pi,removal increased, while a decrease in Pi,removal occurs as

399

the pH and alginate dose increase beyond their optimal values.

400

Figure S1(iii) shows the effect of variable pH and settling time at the optimal levels of FeCl3

401

and alginate dosage on the Pi,removal. At constant pH, greater Pi,removal occurred with longer

402

settling time until it reached an optimal value at 37 minutes, and a sharp decrease with a further

403

increase in settling time. During the long contact time, the Pi species were destabilized which led

404

to the formation of large flocs with excellent settling ability. However, beyond the optimal

405

settling time, re-stabilization of the flocs occurred that resulted in lower Pi,removal. On the other

406

hand, an increase in pH at constant settling time yielded greater Pi,removal, where the response

407

variable decreased as the pH increased beyond the optimal level. As the pH and settling time

408

decreased or increased beyond their optimal level, the Pi,removal was lower relative to the optimal

409

conditions.

410

Figure S1(iv) represents the response surface and contour plots for Pi,removal at optimal levels

411

of pH and settling time, and variable FeCl3 and alginate dosage within the experimental design.

412

The interactive effect of the alginate and FeCl3 dosage does not appear to strongly affect Pi,removal.

413

Greater removal efficiency (> 92%) was obtained at all conditions. The response surface plot

414

shows a clear maximum, indicating that the optimum condition for Pi,removal is well within the

415

experimental design boundary conditions. At the optimum pH, metal hydroxide of the iron

416

neutralizes the phosphate through charge neutralization. The FePO4 precipitate is swept and then

417

adsorbs the alginate polyelectrolyte species through sweeping and adsorption mechanisms46.

418

Figure S1(v) depicts the response surface and contour plots for Pi,removal at optimal levels of

419

pH and FeCl3 dose, with variable alginate dose and settling time for the experimental design.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 36

420

From Figure SI1(v), at constant FeCl3 dose, the Pi,removal increases from 10 to 37 minutes.

421

However, beyond the optimal settling time, the Pi,removal remains virtually constant. Greater

422

removal (85-92%) was obtained at all working conditions. This implies that the interaction

423

between settling and FeCl3 dosage within the specified range does not have a significant impact

424

on Pi,removal. Similar results were observed for the variation between settling time and FeCl3 dose,

425

as shown in Figure S1(vi).

426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441

Table 3: Results of validation experiments under optimised conditions. Independent variable Test 1 2 3 4 5 6

442 443 444 445 446

pH 4.6 3.0 4.0 5.0 6.0 9.0

FeCl3 dose (mg·L-1) 13 10 11 12 13 14

Pi Removal (%) Settling time (minutes) 37 15 25 35 45 55

Alginate dose (mg·L-1) 7.1 3.0 5.0 7.0 9.0 12

3.4 Confirmation and validation


Experimental value

Estimated value

99.7 ± 0.7 85.4 ± 0.9 90.1 ± 1.8 94.7 ± 0. 5 94.4 ± 0.9 74.6 ± 0.9

99.6 85.7 90.5 95.1 94.3 74.8

Page 21 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


447

The validity of the statistical methodology was confirmed by carrying out additional

448

experiments in triplicate at the predicted optimized levels obtained from the solution of equation

449

(5), and five other conditions within the range of the experimental design. The selected

450

conditions for the pH, FeCl3 dosage, alginate dosage and settling time are listed in Table 3, along

451

with the predicted and the measured results. The Pi,removal determined experimentally was 99.7 ±

452

0.7%, which is significantly close to the predicted optimal Pi,removal (99.6%). As shown in Table

453

3, the measured and the estimated values of Pi,removal obtained from RSM are in good agreement.

454

The result verifies that the RSM approach is useful for optimising operational conditions of the

455

coagulation-flocculation process. As well, RSM enables prediction of an empirical relationship

456

between the response factor and the independent variables. Table S2 presents a comparison

457

between the removal of Pi using different coagulant and/or flocculant systems. The results show

458

that the Pi removal at the optimized conditions herein for the binary Fe(III)-alginate coagulant-

459

flocculant system is significantly greater when compared with other coagulant-flocculant

460

systems.

461 462

3.5 Flocculation kinetics

463

The contact time between the orthophosphate (Pi) and the flocculant is a significant factor in

464

understanding the flocculation process and the flocculation kinetics of Pi at fixed initial

465

flocculant dosage 47,48. Figure 5 illustrates the effect of contact time on the flocculation ability of

466

Fe(III)-alginate flocculant with Pi at the RSM optimized conditions. The flocculation process is

467

divided into three regions. In the first region, there is a rapid increase in the flocculation up to 15

468

minutes, it then increases steadily until a dynamic equilibrium is reached at 35 minutes, and



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 36

469

remains nearly constant thereafter. The behaviour in the first region is due to electrostatic

470

interaction between the phosphate and the ferric cations, leading to charge neutralization. Pi is

471

negatively charged and favours interaction with metal cation species to neutralize its surface

472

charge. The trivalent iron metal cation may act as a cross-linker along the alginate chain to form

473

a stable gel which leads to adsorption of Pi49,50. In the second region, there is a gradual increase

474

in the flocculation efficacy as a result of a reduced effective cationic charge on the floc surface.

475

The third region illustrates that the dynamic equilibrium for the floc formation is ca. 35 minutes,

476

in agreement with the RSM model. Similar results for the removal of dyes using such

477

polysaccharide-based flocculants were reported by Kono et al.51.

478

The kinetics of the flocculation process was examined at optimized conditions. The

479

adsorption kinetics of Pi with the Fe(III)-alginate flocculant system was investigated using

480

pseudo-first-order (PFO) and pseudo-first-order (PSO) kinetic models. The PFO kinetic model

481

assumes that the sorption rate decreases linearly as the adsorption rate increases52,53. By

482

comparison, the PSO kinetic model assumes that the rate-limiting step involves interaction

483

between the adsorbent and adsorbate is often applied to describe an adsorption process53,54. The

484

PFO and PSO kinetic models are shown in equation (6) and (7), respectively.

485

qt = qe (1 - e-

486

k2 qe 2t qt = 1+ k2 qet

k1t

(6)

)

(7)

487

Here, qe and qt are the amount of phosphate adsorbed (mg/g) at dynamic equilibrium and at time

488

t. k1 and k2 are the rate constants for the PFO and PSO kinetic models, respectively. k1, k2 and qe

489

were estimated using non-linear regression fitting according to equations (6) and (7). The “best


Page 23 of 36

490

fit” criterion between the experimental data and the kinetic models was determined by

491

minimizing the absolute sum of squares of errors (SSE) by equation (8).

SSE =

492

(qe,i - qc,i )2

å

(8)

N

493

Here, qe,i is the experimental value and qc,i is the calculated or predicted value, according to the

494

kinetic model and N is the number of experimental data points.

50 40

qt (mg/g)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


30 Oberved data

20

PFO prediction

10

PSO prediction

0 0

495 496 497

10

20

30

40

50

60

70

80

Time (min) Figure 5: Flocculation kinetics of Pi removal in aqueous solution at variable time intervals, where co = 10.0 mg·L-1.

498 499

Figure 5 illustrates the kinetic results of the flocculation of Pi with the Fe(III)-alginate

500

flocculant system. The blue and red curves in Figure 5 represent the theoretical PFO and PSO

501

kinetic models, respectively. In Figure 5, the Pi,removal is well described by both PSO and PFO

502

kinetic models. Values of k1, k2, qe and R2 obtained from the regression analysis are listed in

503

Table 4. The correlation coefficient (R2) for the PFO model (R2 = 0.998) exceeds that of the PSO

504

model (R2 = 0.988). Also, it is evident that the predicted values of qe determined by the PFO

505

model agrees well with the experimental values relative to the PSO kinetic model. In addition, 22 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 36

506

the rate constant of the PFO (k1) model is significantly greater than that of the PSO (k2) model.

507

This PFO adsorption mechanism describes the flocculation of phosphate by Fe(III)-alginate,

508

according to the PFO model. The adsorption of Pi by the flocculant system indicates that the

509

sorption rate decreases linearly as the adsorption rate increases, Similar results were reported

510

for the flocculation of anionic dyes and arsenates by polysaccharide-based flocculants

511

The rapid equilibrium process at 35 minutes is shown in Figure 5, indicating that Pi species

512

interact with the outermost surface of the flocculant system.

513 514 515

Table 4: Pseudo-First-Order (PFO) and Pseudo-Second-Order (PSO) kinetic variables for Pi removal in aqueous solution at optimized flocculation condition. co (Pi) = 10.0 mg·L-1.

Model

Qt (mg·g-1)

k · 103

R2

PFO

49.3 ± 0.3

136.7 ± 3.1 min−1

0.998

PSO

556.7 ± 1.1

3.1 ± 0.3 g·mg−1·min−1

0.988

47,48,55,56

.

516 517

3.6 Adsorption isotherms

518

Flocculation-adsorption processes can be described through graphical isotherms. Adsorption

519

isotherms provide insight concerning the interaction between adsorbates and adsorbents and the

520

optimization conditions51. The Langmuir57, Freundlich58 and Sips59 isotherm models are the most

521

frequently used to describe adsorption during a flocculation processes of colloidal particles. In

522

this study, the adsorption isotherm models were used to understand the adsorption processes to

523

elucidate the flocculation mechanism.

524 525

The Langmuir adsorption isotherm model is based on the assumption that adsorption occurs at specific homogeneous sites with the adsorbent, and is described by equation (9).


Page 25 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

526


qe =

K l qm ce 1 + K l ce

(9)

527

qm (mg·g-1) is the monolayer adsorption of Pi per unit mass of Fe (III)-alginate adsorbent and qe

528

is the amount of adsorbed Pi (mg·g-1). Kl (Lg-1) is the equilibrium adsorption constant, which is

529

related to the affinity of the binding sites and is related to a dimensionless constant called the

530

separation factor or equilibrium variable, Rl, according to equation (10). The adsorption process

531

can be categorized into four categories based on the value of Rl: irreversible (Rl = 0), favourable

532

(0 < Rl < 1), linear (Rl = 1) and unfavourable (Rl > 1).

533

Rl =

1 1 + K l c0

(10)

534

The Freundlich isotherm model accounts for multilayer sorption by assuming that the

535

adsorbent has a heterogeneous surface with non-uniform distribution of sorption sites. The

536

Freundlich adsorption isotherm is described by equation (11).

537

qe = K f ce

1

(11)

n

538

Kf is the Freundlich isotherm constant which relates to the adsorption capacity and n is a

539

dimensionless constant. The empirical exponent variable gives valuable information on the shape

540

of the isotherm, where the adsorption process may be classified as unfavourable (1/n > 1),

541

favourable (1/n < 1), and irreversible (1/n = 0).

542

The Sips isotherm model (equation (12)) shares features of the Langmuir and Freundlich

543

isotherms. At low adsorbate concentrations, it converges to a Freundlich isotherm. At high

544

adsorbate concentration (or n = 1), the model reduces to a Langmuir isotherm60.



K s qm cens qe = 1 + K s cens

545

(12)

90 75

Qe (mg·g-1)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 36

60

Experimental data

45

Langmuir

30

Freundlich

15

Sipps 0 0

1

2

3

ce

4

5

6

7

(m·gL-1)

546 547 548 549 550

Figure 6: The adsorption isotherm of phosphate by combined FeCl3 and alginate at optimized flocculation condition. Initial phosphate concentration (1.0 – 20.0) mg·L–1.

551

Figure 6 is the adsorption isotherm for the removal of Pi by Fe(III)-alginate. The experimental

552

data were fit by the Langmuir, Freundlich and Sips isotherm models. The fitted lines through the

553

data represent the goodness-of-fit of the isotherm models employed herein, where the variables

554

are listed in Table S3. According to the results in Figure 6 and Table S3, the removal of Pi by the

555

Fe(III)-alginate system is well-described by the Langmuir and Sips models. The dimensionless

556

constant, Rl obtained from the Langmuir (0.054) and Sips (0.049) models indicate favorable

557

adsorption of phosphate. Also, the monolayer adsorption capacity (qm; mg·g-1) for Pi are given in

558

parentheses at the optimized conditions: Langmuir (83.6±1.02) and Sips (80.3±1.88) isotherm

559

models. The values obtained from the Langmuir and Sips models are in good agreement, where

560

the Langmuir model is considered more reliable due to its relative simplicity and favourable

561

goodness-of-fit for the Fe(III)-alginate isotherm system. As well, the exponent term (ns) is near 25 ACS Paragon Plus Environment

Page 27 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


562

unity which implies that the Sips isotherm converges with the Langmuir model when ns=1. Thus,

563

it can be inferred that the flocculation of orthophosphate by Fe(III)-alginate follows a

564

homogeneous mechanism based on the charge neutralization reaction between the anionic Pi and

565

the ferric cation.

566 567

The dimensionless constant, n-1 in the Freundlich model exceeds unity (2.36±0.12) and

568

indicates that the adsorption process is not favourable. The Langmuir adsorption constant, Kl is

569

0.881±0.003 mg·L–1, where this describes the adsorption process of Pi adsorbed on Fe(III)-

570

alginate. This shows that the flocs formed from the interaction between Pi with Fe(III)-alginate

571

have stronger stability in solution. The flocculation of Pi species by Fe(III)-alginate system

572

follows a homogeneous mechanism based on the charge neutralization reaction between the

573

anionic Pi and the ferric cation species. The flocculation isotherm would be described by the

574

Freundlich model if other interactions occur besides electrostatic processes during flocculation.

575

The Freundlich isotherm model is often used to describe heterogeneous adsorption systems61.

576 577 578 579

4.7 Mechanism

580

The use of chemical precipitation for the removal of soluble impurities or nutrients in water

581

depends on the nature of the complexes formed. The mechanism of chemical Pi removal involves

582

a combined process of coagulation and precipitation that is influenced by solution conditions

583

such as pH initial concentration. The mechanism of Pi removal by chemical precipitation,

584

generally involves two steps62:



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

585

i.

Formation

of

metal-hydroxide-phosphate

complex,

Page 28 of 36

M(OH)3-x(PO4)x.

These

586

complexes either act as centers of precipitation of phosphate ions on the M(III)

587

hydrolysis species, or absorbed onto positively charged M(III) hydrolysis species.

588 589

ii.

Adsorption of phosphate ions onto the M(III) hydrolysis species followed by separation via sedimentation and/or filtration.

590

The mechanism involving the reaction when Fe(III) is added to wastewater to precipitate Pi

591

species is complex and does not follow any simple single steps63. The efficiency of the

592

precipitation of the oxidation product for Pi removal depends on pH, initial Pi concentration in

593

the wastewater, iron (III) dose and temperature1. Precipitation with ferric salt is most effective

594

within certain pH range. The optimum pH for ferric ion is 4.0 to 5.064. A simplified reaction that

595

occurs between Fe (III) and phosphate ions in wastewater is described by equation (13).

596

FeCl3 + PO 43− → FePO 4 ↓ + 3Cl −

(13)

597

As ferric chloride is added to wastewater containing Pi, the contaminants are predominantly

598

coagulated by charge neutralization. In this case, the positive charge on the metal ion neutralises

599

the negatively charged sites of the Pi particles. However, the use of inorganic salts requires

600

higher salt dosage with the subsequently production of large sludge volume. In addition, the pH

601

of the solution is altered when inorganic salts are used in the coagulation process. Biopolymers

602

such as alginate may be used as coagulant aids to yield enhanced precipitation of the

603

contaminants compared with the sole use of ferric chloride. Alginate acts as a destabilising agent

604

through charge neutralization and precipitation mechanism due to the electrostatic attraction

605

between the charge on the polyelectrolyte and the negative charge on Pi. Hence, the growth of

606

micro- to macro-flocs occurs via flocculation due to the reduced surface charge of the particles,

607

leading to decreased electrostatic repulsion. Effective removal of Pi is favoured when the zeta27 ACS Paragon Plus Environment

Page 29 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


608

potential of particles is close to zero65. The high charge density of alginate imparts greater charge

609

to the surface of the contaminant species leading to greater removal efficiency66. High charge

610

density polymers adsorb as a well-defined monolayer orientation with reduced bridging

611

interactions67. The charge neutralisation mechanism enables flocculation among particles with a

612

reduced surface charge (reduction of zeta potential). Hence, a decreased electrostatic repulsion

613

between colloidal particles favour van der Waals interactions to allow for initial aggregation of

614

colloidal and fine suspended materials as a microfloc66.

615

At higher dosage of alginate, there is greater Pi removal efficiency for the same FeCl3 dose

616

and pH, as reported in Table S1. This effect occurs due to adsorption of Pi species onto the

617

Fe(III)-alginate surface. The adsorption affinity between the polymer and Pi is likely appreciable

618

and outweighs the loss of entropy associated with polymer because an adsorbed chain will have a

619

more restricted configurational states than a random coil in free solution67. The mechanism of

620

this flocculation process is an ion-binding adsorption process that occurs at hydroxyl polar

621

functional groups of alginate or charged sites on the polymer backbone. Adsorption of Pi onto

622

alginate occurs when there is sufficient ferric ions present since Fe (III) can act as a bridge

623

between alginate and Pi. In the absence of a trivalent cation, limited flocculation occurs even at

624

high ionic strength, as observed in a previous report14.

625 626

4. Conclusion

627

Optimization of a coagulation-flocculation process for orthophosphate (Pi) removal in

628

aqueous media was investigated using a jar test system. The RSM along with the Box-Behnken

629

design method was used to optimize the conditions for maximum Pi removal. The FeCl3 and

630

alginate dosage, along with pH and settling time were determined as significant factors to yield



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

631

optimal Pi,removal. The predicted quadratic polynomial model by the RSM method was used to

632

estimate the Pi,removal over the range of experimental variables. Optimal conditions for the

633

coagulation-flocculation process occurred at pH 4.6, FeCl3 (12.5 mg·L-1), alginate dosage (7.3

634

mg·L-1) and settling time (37 minutes), where the Pi,removal was 99.7±0.7%. This study revealed

635

good agreement between experimental results and the RSM predictions, further illustrating that

636

RSM can be used to model and optimize coagulation-flocculation processes for phosphate

637

removal.

638

The results for optimized conditions indicate that Pi,removal occurs via adsorption of Pi by the

639

Fe(III)-alginate flocculant system. Greater Pi,removal occurred at acidic versus alkaline conditions.

640

The variable Pi,removal relates to the electrostatic interaction between orthophosphate anion species

641

and the flocculant system. Kinetic adsorption profiles reveal uptake within 37 minutes, where the

642

experimental results are well-described by the pseudo-first-order kinetic model and Langmuir

643

isotherm model at equilibrium. The maximum adsorption capacity of Fe(III)-alginate for Pi

644

obtained was 83.6 ± 1.02 mg·g-1. The results of this study will contribute to the further

645

development of advanced water treatment processes for controlled removal of other types of

646

waterborne oxoanion species. Electrostatic charge neutralization and ion-binding adsorption

647

mechanisms were found to govern the flocculation process.

648 649

Acknowledgements

650

The authors are grateful to the University of Saskatchewan (U of S), the Government of

651

Saskatchewan (Agriculture Development Fund; Project #20140260), and the Natural Sciences

652

and Engineering Research Council of Canada (NSERC) for supporting this research.

653


Page 30 of 36

Page 31 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


654

References

655

(1)

656 657

Wastewater Treatment: A Review. Environ. Pollut. 1988, 49, 183-233. (2)

658 659

Yeoman, S.; Stephenson, T.; Lester, J. N.; Perry, R. The Removal of Phosphorus during

Rybicki, S. Phosphorus Removal From Wastewater - A Literature Review. Division of Water Resources Engineering, 1997, 1-106.

(3)

Painting, S. J.; Devlin, M. J.; Malcolm, S. J.; Parker, E. R.; Mills, D. K.; Mills, C.; Tett,

660

P.; Wither, a.; Burt, J.; Jones, R.; et al. Assessing the Impact of Nutrient Enrichment in

661

Estuaries: Susceptibility to Eutrophication. Mar. Pollut. Bull. 2007, 55, 74-90.

662

(4)

Adsorbent. Desalination 2011, 275 (1–3), 231-236.

663 664

(5)

(6)

667 668

MOEE. Provision and Operation of Phosphorus Removal Facilities at Municipal, Institutional and Private Sewage Treatment Works. 1994, 1-89.

665 666

Benyoucef, S.; Amrani, M. Adsorption of Phosphate Ions onto Low Cost Aleppo Pine

Mahaninia, M. H.; Wilson, L. D. Cross-Linked Chitosan Beads for Phosphate Removal from Aqueous Solution. J. Appl. Polym. Sci. 2016, 133, 1-10.

(7)

Filipkowska, U.; Jóźwiak, T.; Szymczyk, P. Application of Cross-Linked Chitosan for

669

Phosphate Removal From Aqueous Solutions. Prog. Chem. Appl. Chitin its Deriv. 2014,

670

19, 5-14.

671

(8)

Sø, H. U.; Postma, D.; Jakobsen, R.; Larsen, F. Sorption of Phosphate onto Calcite;

672

Results from Batch Experiments and Surface Complexation Modeling. Geochim.

673

Cosmochim. Acta 2011, 75, 2911-2923.

674

(9)

Yan, L. G.; Xu, Y. Y.; Yu, H. Q.; Xin, X. D.; Wei, Q.; Du, B. Adsorption of Phosphate

675

from Aqueous Solution by Hydroxy-Aluminum, Hydroxy-Iron and Hydroxy-Iron-

676

Aluminum Pillared Bentonites. J. Hazard. Mater. 2010, 179, 244-250.

677

(10)

678 679

Sathasivan, A. Water and Wastewater Treatment Technologies - Biological Phosphorus Removal Processes for Wastewater Treatment. Encycl. Life Support Syst. 2010, 1-16.

(11)

Gautam, R. K.; Banerjee, S.; Gautam, P. K.; Chattopadhyaya, M. C. Remediation

680

Technologies for Phosphate Removal From Wastewater: An Overview. Adv. Environ. Res.

681

2014, 36, 1-23.

682

(12)

683 684

Inan, H.; Alaydın, E. Phosphate and Nitrogen Removal by Iron Produced in Electrocoagulation Reactor. Desalin. Water Treat. 2014, 52, 1396-1403.

(13)

Peleka, E. N.; Mavros, P. P.; Zamboulis, D.; Matis, K. a. Removal of Phosphates from 30 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Water by a Hybrid Flotation-Membrane Filtration Cell. Desalination 2006, 198, 198-207.

685 686

(14)

Agbovi, H. K.; Wilson, L. D.; Tabil, L. G. Biopolymer Flocculants and Oat Hull Biomass

687

To Aid the Removal of Orthophosphate in Wastewater Treatment. Ind. Eng. Chem. Res.

688

2017, 56, 37-46.

689

(15)

690 691

Dunets, C. S.; Zheng, Y. Combined Precipitation / Flocculation Method for Nutrient Recovery from Greenhouse Wastewater. Hortscience 2015, 50, 921-926.

(16)

Mohammed, S. A. M.; Shanshool, H. A. Phosphorus Removal from Water and Waste

692

Water by Chemical Precipitation Using Alum and Calcium Chloride. Iraqi J. Chem. Pet.

693

Eng. 2009, 10, 1-6.

694

(17)

Quinlan, P. J.; Tanvir, A.; Tam, K. C. Application of the Central Composite Design to

695

Study the Flocculation of an Anionic Azo Dye Using Quaternized Cellulose Nanofibrils.

696

Carbohydr. Polym. 2015, 133, 80-89.

697

(18)

698 699

(19)

Pa Ho Hsu. Comparison of iron(III) and Aluminum in Precipitation of Phosphate from Solution. Water Res. 1976, 10, 903-907.

(20)

Yang, K.; Li, Z.; Zhang, H.; Qian, J.; Chen, G. Municipal Wastewater Phosphorus Removal by Coagulation. Environ. Technol. 2010, 31, 601-609.

702 703

Tripathy, T.; De, B. R. Flocculation: A New Way to Treat the Waste Water. J. Phys. Sci.

2006, 10, 93-127.

700 701

(21)

Tzoupanos, N. D.; Zouboulis, A. I. Coagulation-Flocculation Processes in Water /

704

Wastewater Treatment : The Application of New Generation of Chemical Reagents. 6th

705

IASME/WSEAS Int. Conf. HEAT Transf. Therm. Eng. Environ. 2008, 309-317.

706

Page 32 of 36

(22)

Vijayaraghavan, G.; Shanthakumar, S. Performance Study on Algal Alginate as Natural

707

Coagulant for the Removal of Congo Red Dye. Desalin. Water Treat. 2016, 57, 6384-

708

6392.

709

(23)

Wu, C.; Wang, Y.; Gao, B.; Zhao, Y.; Yue, Q. Coagulation Performance and Floc

710

Characteristics of Aluminum Sulfate Using Sodium Alginate as Coagulant Aid for

711

Synthetic Dying Wastewater Treatment. Sep. Purif. Technol. 2012, 95, 180-187.

712

(24)

Zhao, Y. X.; Gao, B. Y.; Shon, H. K.; Wang, Y.; Kim, J. H.; Yue, Q. Y.; Bo, X. W.

713

Anionic Polymer Compound Bioflocculant as a Coagulant Aid with Aluminum Sulfate

714

and Titanium Tetrachloride. Bioresour. Technol. 2012, 108, 45-54.

715

(25)

Zhao, Y. X.; Gao, B. Y.; Wang, Y.; Shon, H. K.; Bo, X. W.; Yue, Q. Y. Coagulation 31 ACS Paragon Plus Environment

Page 33 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


716

Performance and Floc Characteristics with Polyaluminum Chloride Using Sodium

717

Alginate as Coagulant Aid: A Preliminary Assessment. Chem. Eng. J. 2012, 183, 387-394.

718

(26)

Zhao, Y. X.; Wang, Y.; Gao, B. Y.; Shon, H. K.; Kim, J.-H.; Yue, Q. Y. Coagulation

719

Performance Evaluation of Sodium Alginate Used as Coagulant Aid with Aluminum

720

Sulfate, Iron Chloride and Titanium Tetrachloride. Desalination 2012, 299, 79-88.

721

(27)

Processes - An Eco-Friendly Approach. Eur. Polym. J. 2009, 45, 1337-1348.

722 723

(28)

724 725

Renault, F.; Sancey, B.; Badot, P. M.; Crini, G. Chitosan for Coagulation/flocculation

Sharma, B. R.; Dhuldhoya, N. C.; Merchant, U. C. Flocculants—An Ecofriendly Approach. J. Polym. Environ. 2006, 14 (2), 195-202.

(29)

Saranya, P.; Ramesh, S. T.; Gandhimathi, R. Effectiveness of Natural Coagulants from

726

Non-Plant-Based Sources for Water and Wastewater Treatment - A Review. Desalin.

727

Water Treat. 2014, 52, 6030-6039.

728

(30)

729 730

Polym. Sci. 2012, 37, 106-126.

(31)

731 732

Lee, K. Y.; Mooney, D. J. Alginate: Properties and Biomedical Applications. Prog.

Devrimci, H. A.; Yuksel, A. M.; Sanin, F. D. Algal Alginate: A Potential Coagulant for Drinking Water Treatment. Desalination 2012, 299, 16-21.

(32)

Ferreira, S. L. C.; Bruns, R. E.; Ferreira, H. S.; Matos, G. D.; David, J. M.; Brandão, G.

733

C.; da Silva, E. G. P.; Portugal, L. A.; dos Reis, P. S.; Souza, A. S.; et al. Box-Behnken

734

Design: An Alternative for the Optimization of Analytical Methods. Anal. Chim. Acta

735

2007, 597, 179-186.

736

(33)

Montgomery, D. C. Design and Analysis of Experiments, 8th Ed.; Wiley, 2013.

737

(34)

Patel, H.; Vashi, R. T. Removal of Congo Red Dye from Its Aqueous Solution Using

738 739

Natural Coagulants. J. Saudi Chem. Soc. 2012, 16, 131-136. (35)

Ebeling, J. M.; Sibrell, P. L.; Ogden, S. R.; Summerfelt, S. T. Evaluation of Chemical

740

Coagulation–flocculation Aids for the Removal of Suspended Solids and Phosphorus from

741

Intensive Recirculating Aquaculture Effluent Discharge. Aquac. Eng. 2003, 29, 23-42.

742

(36)

743 744

Aslan, N.; Cebeci, Y. Application of Box – Behnken Design and Response Surface Methodology for Modeling of Some Turkish Coals. Fuel 2007, 86, 90-97.

(37)

Aslani, H.; Nabizadeh, R.; Nasseri, S.; Mesdaghinia, A.; Alimohammadi, M.; Mahvi, A.

745

H.; Rastkari, N.; Nazmara, S. Application of Response Surface Methodology for

746

Modeling and Optimization of Trichloroacetic Acid and Turbidity Removal Using 32 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Potassium ferrate(VI). Desalin. Water Treat. 2016, 3994, 1-12.

747 748

(38)

Bezerra, M. A.; Santelli, R. E.; Oliveira, E. P.; Villar, L. S.; Escaleira, L. A. Response

749

Surface Methodology (RSM) as a Tool for Optimization in Analytical Chemistry. Talanta

750

2008, 76, 965-977.

751

(39)

Usharani, K.; Lakshmanaperumalsamy, P. Box-Behnken Experimental Design Mediated

752

Optimization of Aqueous Methylparathion Biodegradation by Pseudomonas Aeruginosa

753

MPD Strain. J. Microbiol. Biotechnol. Food Sci. 2016, 5, 534-547.

754

(40)

Usharani, K.; Muthukumar, M. Optimization of Aqueous Methylparathion Biodegradation

755

by Fusarium Sp in Batch Scale Process Using Response Surface Methodology. Int. J.

756

Environ. Sci. Technol. 2013, 10, 591-606.

757

Page 34 of 36

(41)

Rakić, T.; Kasagić-Vujanović, I.; Jovanović, M.; Jančić-Stojanović, B.; Ivanović, D.

758

Comparison of Full Factorial Design, Central Composite Design, and Box-Behnken

759

Design in Chromatographic Method Development for the Determination of Fluconazole

760

and Its Impurities. Anal. Lett. 2014, 47, 1334-1347.

761

(42)

Zolgharnein, J.; Shahmoradi, A.; Ghasemi, J. B. Comparative Study of Box-BehnKen,

762

Central Composite, and Doehlert Matrix for Multvariate Optimization of Ph (II)

763

Adsorption onto Robinian Tree Leaves. J. Chemom. 2013, 27, 12 -20.

764

(43)

Jain, M.; Garg, V. K.; Kadirvelu, K. Investigation of Cr(VI) Adsorption onto Chemically

765

Treated Helianthus Annuus: Optimization Using Response Surface Methodology.

766

Bioresour. Technol. 2011, 102, 600-605.

767

(44)

Yetilmezsoy, K.; Demirel, S.; Vanderbei, R. J. Response Surface Modeling of Pb(II)

768

Removal from Aqueous Solution by Pistacia Vera L.: Box-Behnken Experimental Design.

769

J. Hazard. Mater. 2009, 171, 551-562.

770

(45)

771 772

Design to Optimize Coagulation Tests. Environ. Eng. Res. 2010, 15, 63-70. (46)

773 774

777

Bratby, J. Coagulation and Flocculation in Water and Wastewater Treatment, Second.; IWA publishing: London, 2006.

(47)

775 776

Trinh, T. K.; Kang, L.-S. Application of Response Surface Method as an Experimental

Singh, R. P.; Pal, S.; Rana, V. K.; Ghorai, S. Amphoteric Amylopectin: A Novel Polymeric Flocculant. Carbohydr. Polym. 2013, 91, 294-299.

(48)

Lin, Q.; Qian, S.; Li, C.; Pan, H.; Wu, Z.; Liu, G. Synthesis, Flocculation and Adsorption Performance of Amphoteric Starch. Carbohydr. Polym. 2012, 90, 275-283. 33 ACS Paragon Plus Environment

Page 35 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

778


(49)

Rezende, R.; Bártolo, P.; Mendes, A.; Filho, R. Experimental Characterisation of the

779

Alginate Gelation Process for Rapid Prototyping. eight Int. Conf. Chem. Process Eng.

780

2007, 11, 509-514.

781

(50)

Simpson, N. E.; Stabler, C. L.; Simpson, C. P.; Sambanis, A.; Constantinidis, I. The Role

782

of the CaCl2-Guluronic Acid Interaction on Alginate Encapsulated βTC3 Cells.

783

Biomaterials 2004, 25, 2603-2610.

784

(51)

785 786

with Cellulose Ampholytes. J. Water Process Eng. 2015, 7, 83-93. (52)

Lagergren, S. About the Theory of so-Called Adsorption of Soluble Substances Zur Theorieder Sogenannten. K. Sven. Vetensk. sakademiens Handl. 1898, 24, 2474-2479.

787 788

Kono, H.; Kusumoto, R. Removal of Anionic Dyes in Aqueous Solution by Flocculation

(53)

Zhen, Y.; Ning, Z.; Shaopeng, Z.; Yayi, D.; Xuntong, Z.; Jiachun, S.; Weiben, Y.;

789

Yuping, W.; Jianqiang, C. A pH- and Temperature-Responsive Magnetic Composite

790

Adsorbent for Targeted Removal of Nonylphenol. ACS Appl. Mater. Interfaces 2015, 7,

791

24446-24457.

792

(54)

793 794

Biochem. 1999, 34, 451-465.

(55)

(56)

797 798

Kwok, K. C. M.; Koong, L. F.; Chen, G.; McKay, G. Mechanism of Arsenic Removal Using Chitosan and Nanochitosan. J. Colloid Interface Sci. 2014, 416, 1-10.

(57)

Langmuir, I. The Constitution and Fundamental Properties of Solids and Liquids. Part I. Solids. J. Am. Chem. Soc. 1916, 252, 2221-2295.

799 800

Guibal, E.; Roussy, J. Coagulation and Flocculation of Dye-Containing Solutions Using a Biopolymer (Chitosan). React. Funct. Polym. 2007, 67, 33-42.

795 796

Ho, Y. S.; McKay, G. Pseudo-Second Order Model for Sorption Processes. Process

(58)

801

Freundlich, H. M. F. Über Die Absorption in Lösungen. Z. Phys. Chem. 1909, 57, 385470.

802

(59)

Sips, R. On the Structure of a Catalyst Surface. II. J. Chem. Phys. 1948, 18, 490-495.

803

(60)

Ho, Y. S.; Porter, J. F.; Mckay, G. Divalent Metal Ions Onto Peat : Copper , Nickel and

804 805

Lead Single Component Systems. Water, Air, Soil Pollut. 2002, 141, 1-33. (61)

806 807 808

Crini, G. Kinetic and Equilibrium Studies on the Removal of Cationic Dyes from Aqueous Solution by Adsorption onto a Cyclodextrin Polymer. Dye. Pigment. 2008, 77, 415-426.

(62)

Jiang, J. Q.; Graham, N. J. D. Pre-Polymerised Inorganic Coagulants and Phosphorus Removal by Coagulation - A Review. Water SA 1998, 24, 237-244. 34 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

809

(63)

810 811

(64)

(65)

818

Kleimann, J.; Gehin-Delval, C.; Auweter, H.; Borkovec, M. Super-Stoichiometric Charge Neutralization in Particle-Polyelectrolyte Systems. Langmuir 2005, 21, 3688-98.

(66)

816 817

Suschka, J.; Machnicka, a.; Poplawski, S. Phosphates Recovery from Iron Phosphates Sludge. Environ. Technol. 2001, 22 (11), 1295-1301.

814 815

Jenkins, D.; Ferguson, J.; Menar, A. Chemical Processes for Phosphate Removal. Water Res. 1971, 5, 369-389.

812 813

Page 36 of 36

Lee, C. S.; Robinson, J.; Chong, M. F. A Review on Application of Flocculants in Wastewater Treatment. Process Saf. Environ. Prot. 2014, 92, 489-508.

(67)

Bolto, B.; Gregory, J. Organic Polyelectrolytes in Water Treatment. Water Res. 2007, 41, 2301-2324.

819 820 821 822 823 824 825 826 827


Flocculation Optimization of Orthophosphate with ... - ACS Publications

Recommend Documents