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
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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,
4 ACS Paragon Plus Environment
such
performance, sludge
as
polyacrylamide,
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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
8 ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
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
9 ACS Paragon Plus Environment
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
Industrial & Engineering Chemistry Research
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
10 ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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
12 ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
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
13 ACS Paragon Plus Environment
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
Industrial & Engineering Chemistry Research
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
14 ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
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.
15 ACS Paragon Plus Environment
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
Industrial & Engineering Chemistry Research
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
16 ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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.
18 ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
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
19 ACS Paragon Plus Environment
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
Industrial & Engineering Chemistry Research
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
20 ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
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
21 ACS Paragon Plus Environment
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
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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).
23 ACS Paragon Plus Environment
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
Industrial & Engineering Chemistry Research
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.
24 ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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:
26 ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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
28 ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
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
29 ACS Paragon Plus Environment
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
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
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
Industrial & Engineering Chemistry Research
(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
Industrial & Engineering Chemistry Research
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
35 ACS Paragon Plus Environment