Journal of Hazardous Materials 227–228 (2012) 436–444

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials

journal homepage: www.elsevier.com/locate/jhazmat

Application of Central Composite Design approach for removal of chromium (VI)

from aqueous solution using weakly anionic resin: Modeling, optimization, and
study of interactive variables
Shailendra Bajpai a,∗, S.K. Gupta b, Apurba Dey c, M.K. Jha a, Vidushi Bajpai a, Saurabh Joshi a, Arvind Gupta a

a
Department of Chemical Engineering, National Institute of Technology, Jalandhar, India
b
Department of Chemical Engineering, Indian Institute of Technology, Delhi, India
c
Department of Bio Technology, National Institute of Technology, Durgapur, India

h i g h l i g h t s

 Use of weakly anionic divinylbenzene copolymer resin for Cr (VI) removal.

 Response surface methodology approach was used to optimize the process variables.
 A quadratic model was developed and validated experimentally.
 Resin dose was found to be most significant individual variable.
 Functional groups involved during reaction were studied using FTIR/EDX.

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, response surface methodology (RSM) approach using Central Composite Design (CCD)
Received 30 November 2011 is applied to develop mathematical model and optimize process parameters for Cr (VI) removal from
Received in revised form 12 April 2012 aqueous streams using weakly anionic resin Amberlite IRA 96. The individual and combined effect of
Accepted 3 May 2012
four process parameters, i.e. contact time, initial solution pH, initial Cr (VI) concentration and resin dose
Available online 17 May 2012
on Cr adsorption were studied. Analysis of variance (ANOVA) showed the relative significance of process
parameters in removal process. Initial solution pH and resin dose were found to be more significant
Keywords:
than contact time and initial Cr (VI) concentration. The second-order regression model was developed
Response surface methodology
Central Composite Design
to predict the removal efficiency using Design Expert software. The optimal conditions to remove Cr
Synthetic resins from aqueous solution at constant temperature of 30 ◦ C and stirring speed of 250 rpm were found to be
Ion exchange contact time 62.5 min, pH 1.96, initial Cr (VI) concentration 145.4 mg/L, and resin dose 8.51 g/L. At these
conditions, high removal efficiency (93.26%) was achieved. FTIR and EDX analysis were conducted to
interpret the functional groups involved during the Cr–resin interaction.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction
The increased discharge of metallic substances into environ-
ment by process industries is a cause of serious concern worldwide.
Chromium (Cr) is one of the most toxic metal that finds its wide
spread use in various industries including electroplating, tanning,
mining, metal and paint industries, etc. It is generally found in
two stable oxidation states, i.e. trivalent and hexavalent. Cr (III)
∗ Corresponding author.
E-mail address: bajpais@nitj.ac.in (S. Bajpai).
is sparingly soluble and less toxic in nature. However, Cr (VI) is 300
times more toxic compared to its trivalent form [1]. Hexavalent
chromium is a potential carcinogen as it breaks the DNA strand,
DNA–protein crosslink, and modifies DNA transcription process
leading to chromosomal damage. Long term exposure of this metal
to humans has proven to be harmful and may lead to diseases
such as cancer of digestive tract and lungs, stomach ulcers, nausea,
diarrhea, liver and kidney damage, and dermatitis [1–3].
The primary routes of chromium exposure to living organ-
isms include inhalation, oral ingestion, or through dermal contact.
Various regulatory bodies worldwide such as World Health orga-
nization (WHO), United States Environment Protection Agency

0304-3894/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jhazmat.2012.05.016
 S. Bajpai et al. / Journal of Hazardous Materials 227–228 (2012) 436–444
Nomenclature
b constants: linear, interactive, quadratic coefficients
ε error
n number of variables
qt amount of Cr (VI) adsorbed on resin (mg/g)
x coded variable
X1 contact time
X2 pH of aqueous solution
X3 initial concentration of Cr (VI) in solution (mg/L)
X4 adsorbent doses (g/L)
Y response of the system
(USEPA), Central Pollution Control Board (CPCB), India have pre-
scribed maximum emission limit of hexavalent chromium in
surface water [4–6]. In India, the maximum allowable concentra-
tion of Cr (VI) prior to disposal in receiving water is below 0.1 mg/L;
whereas it is 5 mg/L for trivalent chromium [7]. In order to comply
with the strict environmental standards, it is essential for process
industries to treat effluents to desired levels before discharging into
the natural environment. Various researchers have used different
techniques for chromium extraction including electro coagulation
[8,9]; membrane processes [10]; low cost adsorbents including
agriculture waste [11–13]; solvent extraction [14]; chemical pre-
cipitation and redissolution [15]. Amongst these techniques, two
step chemical reduction and precipitation method is more popular
in India. In this process, Cr (VI) is first reduced to the trivalent form,
which is further treated with an alkali such as MgO to get insolu-
ble Cr(OH)3 . However, this process lacks the quality of recovered
chromium. Removal using low cost adsorbents is also very pop-
ular amongst researchers, but this process lacks in the yield and
disposal of toxic sludge is a serious problem. Although membrane
separation process shows higher selectivity and excellent removal
of heavy metal, yet it is more energy intensive and requires high
initial capital investment.
Ion exchange process promises an attractive alternative. It
involves reversible chemical reaction in which one ion form from
solution gets exchanged with other similar charged ion attached to
the insoluble solid matrix of the resin. These solid ion exchanger
particles are either naturally occurring inorganic zeolites or syn-
thetically produced organic resins. Resins are either cationic or
anionic in nature. Cationic exchange resin has the mobile ion in
the form of Na+ or H+ . Anionic exchange resin has Cl− or OH− as
mobile ions. A generalized ion exchange reaction at equilibrium
can be written as:
AR + B±n  BR + A±n
where ion B from solution is exchanged by the ion A attached
to the resin R. The pollutant B is captured by the resin AR
through ion exchange mechanism. Once the resin is exhausted
(at BR form), it needs to be regenerated. Synthetic resins show
good physical, chemical and thermal stability, higher selectiv-
ity, produce no toxic sludge, and are efficient at higher metal
concentration.
The published information in literature about adsorption of Cr
with synthetic resins is much detailed [16,17]. Most of such studies
discussed the sorption mechanism, kinetics, equilibrium and ther-
modynamics at length. Neagu and Mikhalovsky [16] discussed the
removal of Cr (VI) onto functionalized pyridine copolymer resins.
Batch adsorption studies were carried out to determine the effect
of concentration, adsorbent dose, pH, temperature and the pres-
ence of sulfate anions on removal. Pehlivan and Cetin [17] studied
Cr (VI) sorption with macroporous resins containing quarternary
amine groups (Lewatit MP 64 and Lewatit MP 500). Detailed batch
437
study was conducted for varying Cr (VI) concentration, adsorbent
dose, pH, contact time and temperature. The optimum pH for Cr (VI)
adsorption was found as 5.0 for Lewatit MP 64 and 6.0 for Lewatit
MP 500. The sorption capacity obtained was 0.40 and 0.41 mmol/g
resin for Lewatit MP 64 and Lewatit MP 500 anion exchangers,
respectively.
However in these studies, optimization of process parameters
was done by conventional approach, i.e. by changing one variable
at a time while keeping all others at constant level. This classi-
cal approach is time consuming, requires lots of experiments and
could not depict the interactive effects of process variable. More-
over, optimized process condition obtained using this method is
not reliable. We have used response surface methodology (RSM)
approach for optimization in this work. RSM is a collection of statis-
tical and mathematical techniques useful for modeling and analysis
of problems in which a response of interest is influenced by sev-
eral variables [18]. In literature, RSM based on Central Composite
Design (CCD) has been used in different applications of engineering
including heavy metal removal, dye removal, electrocoagulation,
ethanol production, etc.
Garg et al. [19] studied the removal of nickel from aqueous
medium using an agricultural waste biomass – sugarcane bagasse.
The influence of adsorbent dose, pH and agitation speed on the
removal was studied by using RSM approach. The central compos-
ite face-centered experimental design was used for designing the
experiments as well as for full response surface estimation. The
optimum conditions for maximum removal of nickel from an aque-
ous solution of 50 mg/L were adsorbent dose (1500 mg/L), pH (7.52)
and stirring speed (150 rpm).
Bhatti et al. [9] studied to optimize process variables, i.e. elec-
trolysis voltage and treatment time for the removal of hexavalent
chromium using electrocoagulation. RSM with CCD was used to
achieve energy efficient removal of Cr (VI). Optimal Cr (VI) removal
efficiency occurred at 11 V and 18.6 min treatment time to give 50%
Cr (VI) removal efficiency with a consumption of 15.46 kWh m−3
energy.
Sasikumar and Viruthagiri [20] used five level CCD to study
the effects of four process variables, i.e. substrate concentration,
pH, incubation temperature and fermentation time for the pro-
duction of ethanol from pre-treated sugarcane bagasse by cellulase
and yeast Kluyveromyces fragilis. Maximum ethanol concentration
(32.6 g/L) was obtained from 180 g/L pre-treated sugarcane bagasse
at the optimized process conditions (temperature 35 ◦ C, pH 5.5) in
72 h aerobic batch fermentation.
RSM approach was used for predictive model building and opti-
mization of chromium adsorption on activated carbon obtained
from tamarind wood [21]. In this work the application of RSM
was presented for optimizing the removal of Cr (VI) ions from
(1)
aqua solutions using activated carbon as adsorbent. The opti-
mum adsorbent dose, temperature, initial concentration of Cr
(VI) and initial pH of the Cr (VI) solution were found to be
4.3 g/L, 32 ◦ C, 20.15 mg/L and 5.41, respectively. Under optimal
value of process parameters, high removal (>89%) was obtained
for Cr (VI).
In literature, there is little information available on the opti-
mization of Cr adsorption on synthetic resins. Therefore, we
decided to study the individual and combined effect of four
process parameters, i.e. contact time, initial solution pH, ini-
tial Cr (VI) concentration and resin dose on Cr sorption using
weakly anionic resin. The RSM approach using CCD was used
to develop the mathematical model and study the interactive
effect of process parameters. The present study was aimed at
maximizing the removal of hexavalent chromium using param-
eter optimization. EDX and FTIR analysis were conducted to
identify the functional groups involved during the ion exchange
process.
438 S. Bajpai et al. / Journal of Hazardous Materials 227–228 (2012) 436–444

2. Materials and methods Table 1

Independent variables and levels.

2.1. Simulated Cr (VI) solution Variables Levels

−˛ −1 0 +1 +˛
A stock solution of Cr (VI) was prepared (1000 mg/L) by dis-
pH 1 2.5 4 5.5 7
solving 2.828 g of analytical grade potassium dichromate (K2 Cr2 O7 )
Adsorbent dose 0.2 2.65 5.1 7.55 10
obtained from S.D. Fine-Chem. Ltd., India in double distilled water. Cr (VI) concentration 50 162.5 275 387.5 500
The stock solution was further diluted to obtain desired concentra- Treatment time 0 30 60 90 120
tion ranging from 50 to 500 mg/L.

2.2. Adsorbent: Amberlite IRA 96 resin the response and validation of the model. The relation between
desired response and independent variables can be written as:
Amberlite IRA 96, a weakly anionic resin, was obtained from Y = f (X1 , X2 , X3 , . . . . . . Xn ) + ε (3)
Rohm and Haas, USA. The resin beads have matrix containing
polystyrenedivinylbenzene copolymer with tertiary amine (>85%) where Y represents the response, X are independent variables, n is
as functional group. Higher porosity makes it suitable for adsorp- the number of factors being studied and ε is experimental error.
tion process. Resin exhibits good physical and osmotic stability. The RSM based CCD works with the coded value for process variables.
tan color spherical beads vary between 0.55 and 0.75 mm sizes, The relation between coded form and actual value can be written
specific gravity in the range 1.040–1.060. The resin shows mois- as:
ture holding capacity of 57–63%. It is resistant to decomposition by Xi − Xavg
chemical attack due to vinylbenzene cross-linkage and can be used Coded value = xi = (4)
X
for a wide range of chemicals.
where Xi is actual value of the ith factor in the actual units, Xavg is
the average of the low and high values for the ith factor, and X rep-
2.3. Batch experiment
resent the step change. In order to get true functional relationship
between independent variables and the response, a second order
Batch experiments were conducted to study the effect of various
polynomial (Eq. (5)) was used to describe the effect of variables in
operating parameters on adsorption in an incubator-shaker (Nova
terms of linear, quadratic and cross product terms.
Instruments Ltd., India) at 30 ◦ C with a constant agitation speed of
250 rpm in 250 mL Erlenmeyer flask. The adsorption of Cr (VI) on 
n

n

n 
n

resin beads was performed for different intervals (0–120 min), ini-
tial solution pH (1–7), initial Cr (VI) concentrations (50–500 mg/L)
and resin dose (0.2–10 g/L) in 100 mL aqueous solution. The pH
was adjusted by the addition of 0.1 N HCl or 0.1 N NaOH using
pH meter (ORION 2 star pH bench-top, USA). At the end of each
experiment, samples were withdrawn, filtered and suitably diluted
for further steps. The concentration of Cr (VI) in the filtrate was
measured spectrophotometerically (Lamottee, USA) as per stan-
dard APHA method [22] using 1,5-diphenylcarbazide complexing
agent at 540 nm wavelength. The solution of 1,5-diphenylcarbazide
was prepared by dissolving 1 g of reagent into 100 mL acetone with
few drops of glacial acetic acid. All experiments were conducted in
doublets and the results are presented as average. The % removal
of Cr (VI) was calculated using the following equation:
initial conc. of Cr(VI) − final conc. of Cr(VI)
% removal = × 100
initial conc. of Cr(VI)
(2)
2.4. Response surface methodology
Various statistical methodologies are available for optimizing
process variables such as Taguchi, iterative mathematical search
technique, heuristic search technique, metaheuristic search tech-
nique, genetic algorithm, tabu search, simulated annealing [23] and
response surface methodology (RSM) [24].
RSM can be defined as a statistical method that uses quantitative
data from appropriate experiments to solve multivarient equation.
Main aim of RSM is to obtain an optimal response. Researchers
have used various techniques in RSM including Central Composite
Design (CCD) [25,26], Box–Behnken statistical experiment design
(BBD) [27,28], and two-level full factorial design (FFD) [29] to pre-
dict the ultimate response.
CCD was first introduced by Box and Wilson in year 1951 and is
still the most popular class of response surface methodology (RSM).
CCD involves three steps: performing the designed experiments,
estimating the coefficients in a mathematical model and predicting
Y = b0 + bi xi + bii xi2 + bij xi xj + ε (5)
i=1 i=1 i<j j
where i and j are linear, quadratic coefficients; b0 is constant coef-
ficient, bi is linear coefficient, bii is interactive coefficient and bij
is quadratic coefficient. Each variable is investigated for individual
and interactive effect on removal process. Eq. (5) can be written
for four independent variables with Y as ultimate response in their
coded values in following equation:
Y = b0 + b1 x1 + b2 x2 + b3 x3 + b4 x4 + b12 x1 x2 + b13 x1 x3 + b14 x1 x4
+ b23 x2 x3 + b24 x2 x4 + b34 x3 x4 + b11 x12 + b22 x22 + b33 x32 + b44 x42
(6)
CCD requires minimum number of test as the standard 2n facto-
rial with its origin at the center, 2n points fixed axially at a distance
˛ and replicate tests at the center [30]. These designs consist of a
2n factorial (coded to the usual ±1 notation), 2n axial points (±˛,
0, 0. . ., 0), (0, ±˛, 0. . ., 0),. . ., (0, 0. . ., ±˛), and nc center points (0,
0, 0,. . ., 0). In the present work, four process variables, i.e. contact
time (X1 ), initial solution pH (X2 ), initial Cr (VI) concentration (X3 )
and adsorbent dose (X4 ) were studied using CCD with the help of
Design Expert software (Stat Ease, USA). The experimental parame-
ters and their levels are given in Table 1.The range of ˛ lies between
√
1 and n.
It is also important to check the adequacy of the developed
model. The analysis of variance (ANOVA) technique was applied
to identify significant variables and their individual and interactive
effect on the Cr removal process. Significance of process variables
was checked by p-value and F-value.
3. Results and discussion
3.1. Experimental design and development of regression model
Batch experiments were performed for the removal of hexava-
lent chromium from aqueous solution using weakly anionic resin
 S. Bajpai et al. / Journal of Hazardous Materials 227–228 (2012) 436–444 439

Table 2
Independent variables and response of adsorption.

Run Time (min) pH Conc’n (mg/L) Doses (g/L) % removal

Actual Predicted

1 60 4 500 5.1 44.98 52.45

2 90 5.5 387.5 2.65 36.97 37.10
3 60 1 275 5.1 91.56 84.26
4 90 2.5 162.5 2.65 49.46 52.18
5 90 2.5 387.5 7.55 77.98 78.65
6 90 5.5 162.5 2.65 34.76 32.762
7 60 4 275 5.1 51.98 52.18
8 120 4 275 5.1 52.86 44.18
9 60 4 275 5.1 48.64 52.18
10 60 4 275 10 61.67 69.92
11 60 4 275 5.1 49.99 56.35
12 60 4 275 5.1 46.96 52.18
13 90 2.5 387.5 2.65 55.78 56.80
14 60 4 275 5.1 52.56 52.18
15 90 5.5 387.5 7.55 42.78 42.77
16 30 2.5 162.5 7.55 95.98 88.52
17 90 5.5 162.5 7.55 58.86 50.70
18 60 4 275 0.2 20.34 23.48
19 30 5.5 387.5 7.55 40.64 27.25
20 60 7 275 5.1 8.45 17.14
21 30 5.5 387.5 2.65 30.86 14.92
22 30 2.5 162.5 2.65 51.56 47.76
23 60 4 50 5.1 78.56 75.85
24 30 5.5 162.5 2.65 30.55 26.50
25 60 4 275 5.1 56.86 52.18
26 30 2.5 387.5 2.65 32.98 33.53
27 90 2.5 162.5 7.55 97.45 105.35
28 30 5.5 162.5 7.55 54.56 49.15
29 0 4 275 5.1 0 3.07
30 30 2.5 387.5 7.55 62.87 60.78

applying RSM based CCD. A 24 CCD model was generated. This
model consists of 16 factorial points, 8 axial points and 6 replicates
at center. The 30 runs (Eq. (7)) as per experimental design obtained
from Design Expert software are presented in Table 2.
N = 2n + 2n + nc = 24 + (2 × 4) + 6 = 30
The quadratic regression model was developed in terms of
process variables in coded values (Eq. (8)). Quadratic model was
selected for model development as suggested by the software.
Based on the experimental response, the minimum and maximum
Cr (VI) adsorption was found between 0% and 97.45%. The positive
sign of the coefficients in Eq. (8) indicates the synergistic effect,
whereas negative sign suggests antagonistic effect. It is clear from
Eq. (8) that individual operating variables initial solution pH and
initial Cr (VI) concentration have net negative effect on chromium
removal, whereas contact time and resin dose have a net positive
effect.
Removal% = 44.21271+(0.71240X1 )−(7.84705X2 ) − (0.21915X3 )
+ (15.88276X4 ) − (0.0130X1 X2 ) + (1.17955E − 00 X1 X3 )
− (0.022634X1 X4 ) + (5.75924E − 003 X2 X3 ) − (1.09992 X2 X4 )
− (0.011127 X3 X4 ) − (5.56975E − 003X12 ) + (0.39154 X22 )
+ (2.36541E − 004 X32 ) − (0.22808 X42 )
and the variable is said to be significant. Sum of square should also
be checked while considering the significance of a particular vari-
able [31,32]. As the value of SS increases the significance of that
variable also increases.
It is clear from Table 3 that model F value is 7.7 with corre-
(7) sponding p value of 0.0002 and high SS (13,080.82). It implies that
the model is significant and can appropriately explain the rela-
tionship between response and independent variables. The linear
coefficients were found to be more significant than quadratic and
interacting coefficients. ANOVA study suggested that initial solu-
tion pH (p = 0.0001, SS = 4715, F = 38.88) has the most significant
effect on chromium removal followed by resin dose (p = 0.0001,
SS = 3197, F = 26.36). Initial Cr (VI) concentration (p = 0.0138) and
contact time (p = 0.0176) have comparatively less significant effect
on the response.
Fig. 1 shows the relationship between actual and the predicted
values of Cr (VI) removal from simulated solutions using resin. The
actual data are the original measure of chromium concentration in
solution that was estimated experimentally using Eq. (2). On the
other hand, predicted values were generated using Eq. (8). Fairly
moderate value of correlation coefficient R2 (0.88) was obtained
between experimental and predicted response. This could be due
to coverage of wide range of process variables in limited number of
experiments and/or contribution of insignificant terms of Table 3
[33].
(8)

3.3. Combined effect of process variables

3.2. Statistical analysis
ANOVA technique was applied to the regression model (Eq. (8))
and the significance of each term is presented in Table 3. It is deter-
mined by the corresponding F value, p value and sum of square (SS).
Larger the F value, the more significant is the corresponding term.
However, small p value indicates the rejection of null hypothesis
Three dimensional (3D) response surface plots were generated
to investigate the effects of four process parameters, i.e. contact
time, initial solution pH, initial Cr (VI) concentration and resin
dose on chromium removal. The interactive effect of adsorbent
dose and contact time on percent removal of Cr (VI) using resin
is shown in Fig. 2. More than 75% removal was achieved using
440 S. Bajpai et al. / Journal of Hazardous Materials 227–228 (2012) 436–444

Table 3
ANOVA table for response surface quadratic model.

Source Sum of squares df Mean square F value p value prob > F

Model 13,080.82 14 934.34 7.7 0.0002

X1 – time 862.69 1 862.69 7.11 0.0176
X2 – pH 4715.57 1 3197.57 38.88 0.0001
X3 – conc 942.77 1 942.77 7.77 0.0138
X4 – dose 3197.51 1 4715.51 26.36 0.0001
X1 X2 5.38 1 5.38 0.044 0.0836
X1 X3 284.21 1 284.21 2.34 0.1467
X1 X4 41.89 1 41.89 0.35 0.5655
X2 X3 18.84 1 18.84 0.16 0.6991
X2 X4 257.11 1 257.11 2.12 0.166
X3 X4 170.65 1 170.65 1.41 0.254
X12 696.844 1 696.844 5.74 0.03
X22 21.52 1 21.52 0.18 0.6796
X32 311.07 1 311.07 2.56 0.1301
X42 51.98 1 51.98 0.43 0.5226
Residual 1819.49 15 121.3
Lack of fit 1757.53 10 175.75 14.18 0.0046
Pure error 61.96 5 12.39
Cor total 14,900.32 29

rapid. Once the resin gets saturated with chromate anions, the
removal increases slowly with time. On the other hand, removal
was found to increase with increasing resin dose. Higher adsor-
bent dose increases the number of active sites and overall surface
area of the adsorbent particles that removes Cr anions from
solution.
Fig. 3 shows the 3D response surface plot to show the combined
effect of initial solution pH and adsorbent dose on the of chromium
(VI) removal at contact time of 60 min and initial concentration of
275 mg/L. Total removal can be achieved at initial solution pH 1.6
and dose of 8.8 g/L. The interaction between chromate anions and
weakly anionic resin can take place through following ion exchange
mechanism [34]:

(R3 NH+ )Cl− + HCrO− + − −

4  (R3 NH HCrO4 + Cl ) (9)
+ − + −
2(R3 NH )Cl + CrO2−
4  (R3 NH )CrO2−
4 + 2Cl (10)
Fig. 1. The actual and predicted response plot of chromium (VI) adsorption.
2(R3 NH+ )Cl− + Cr2 O2− + 2−
7  (R3 NH )2 Cr2 O7 + 2Cl
−
(11)
−
aqueous solution of initial Cr (VI) concentration 275 mg/L, initial (R3 NH +
)HCrO−
4
+
+ NaCl  (R3 NH )2 Cl + NaHCrO−
4 (12)
solution pH 4, resin dose 5.8 g/L and contact time 70 min. It was In the above reactions, R3 NH+
is the polymer phase of the
observed that adsorption increases rapidly with increasing time resin attached with tertiary amine functional groups. It is clear
initially and attains a constant value later. In the initial stage of from Fig. 3 that the Cr removal increases with decreasing pH. This
adsorption, the concentration difference of Cr (VI) between solid could be explained as the dominant form of Cr (VI) at initial pH < 2
surface and the solution is large, therefore sorption was fast and was HCrO4 − and increase in solution pH shifts concentration of

Fig. 2. The combined effect of contact time and resin dose on chromium (VI) removal Fig. 3. The combined effect of initial solution pH and resin dose on chromium (VI)
at constant initial Cr (VI) concentration 275 mg/L and initial solution pH 4. removal at constant initial Cr (VI) concentration 275 mg/L and contact time 60 min.
 S. Bajpai et al. / Journal of Hazardous Materials 227–228 (2012) 436–444 441

Fig. 4. The Intraparticle diffusion kinetic model for the uptake of chromate ions. Fig. 5. The combined effect of contact time and initial solution pH on adsorption of
chromium (VI) removal at constant initial Cr (VI) concentration 275 mg/L and resin
dose (5.1 g/L).

HCrO4 − to other forms, i.e. CrO4 − and Cr2 O7 −2 . The increase in Cr

sorption at acidic pH could be attributed to strong electrostatic concentration reduces the removal efficiency. It could be explained
attraction between positively charged amine groups of resin and as the resin beads have limited number of active adsorption sites to
negatively charged HCrO4 − ions (Eq. (9)). However, the decrease accommodate Cr (VI) ions. However, increasing Cr (VI) concentra-
in removal at higher pH may be attributed to decrease of electro- tion increases the adsorbate ions to be accommodated and hence
static attraction (Eqs. (10) and (11)), and competitiveness between the removal efficiency of adsorption reduces.
anionic species (CrO4 − , Cr2 O7 −2 ) and the OH− ions in the bulk for
sorption on active sites of resin. It is concluded that the active form 3.4. Optimization and model validation
of Cr (VI) that can be adsorbed by Amberlite IRA 96 at pH < 2 was
largely HCrO4 − ions [35,36]. The exhausted resin can be regener- It is important to find out the optimal conditions to maxi-
ated as per Eq. (12). mize chromium removal using developed mathematical model.
It is a known fact that various mechanism and steps in ion The model was optimized using quadratic programming with the
exchange phenomenon can control the kinetics. The following four objective to maximize the removal within defined experimental
steps are generally cited [37] in literature: (1) mass transfer of range. The optimal conditions for chromium removal using Amber-
solute from solution to boundary layer; (2) mass transfer of solute lite IRA 96 resin in batch process at temperature 30 ◦ C were contact
ion from boundary layer to the surface of resin; (3) sorption and time 62.56 min, initial solution pH 1.96, initial Cr (VI) concentra-
ion exchange of ions onto sites; and (4) internal diffusion of solute. tion 145.4 mg/L and adsorbent dose 8.51 g/L. At these conditions,
The first step is excluded as the system was stirred. The third step 94.90% removal was predicted by the model. This result was val-
is assumed to be non rate limiting as sorption is considered to be a idated experimentally (93.26% Cr removal), which is close to the
rapid phenomenon. To know if the internal film diffusion is the rate predicted value as shown in Table 4.
limiting step or not, the Cr uptake data were plotted as per Fickian
diffusion law [38]: 3.5. FTIR and EDX analysis of resin
0.5
qt = Ki t +I (13)
FTIR analysis helps in understanding the reaction mechanism
where qt is the uptake rate in (mg/g) and t is the time in min, Ki is as it identifies the functional groups involved during interaction
the intraparticle diffusion rate (mg/g min−0.5 ) and I is the inter-
cept. From Fig. 4, it is observed that two straight line portions
with different slopes and intercept values were obtained. Both lin-
ear portions do not pass through the origin. Therefore, it can be
concluded that the overall rate of sorption for resin is controlled
by the intraparticle diffusion along with the boundary layer diffu-
sion. This conclusion is in line with the information available in the
literature [39].
The combined effect of contact time and initial solution pH on
adsorption of Cr (VI) is shown in Fig. 5. More than 90% removal was
achieved at resin dose of 5.1 g/L, contact time 80 min, initial solution
pH 1.2 and initial Cr concentration of 275 mg/L. It can be seen from
Fig. 5 that % removal increases sharply by decreasing solution pH,
while increase in contact time does not affect the adsorption at
higher level.
Fig. 6 shows the combined effect of contact time and initial Cr
(VI) concentration at initial solution pH 4 and resin dose of 5.1 g/L.
About 75% Cr (VI) removal can be achieved at initial solution pH 4,
resin dose 5.1 g/L, initial Cr (VI) concentration 50 mg/L and 65 min Fig. 6. The combined effect of contact time and initial Cr (VI) concentration on
of contact time. It is clear from Fig. 6 that increases in initial metal chromium (VI) removal at constant resin dose 5.1 g/L and initial solution pH 4.
442 S. Bajpai et al. / Journal of Hazardous Materials 227–228 (2012) 436–444

Table 4
Model validation.

pH Adsorbent Time Initial Cr (VI) % removal

dose(g/L) (min) conc’n (mg/L)

Predicted Actual

1.96 8.51 62.5 145.4 94.99 93.26

Fig. 8. FTIR spectra of IRA 96 after chromium adsorption.

Fig. 7. FTIR spectra of resin IRA 96 before chromium adsorption.
between the resin and chromium. The spectra were recorded in
region of 4000–400 cm−1 with a resolution of 2 cm−1 in an IR
spectrophotometer (Perkin Elmer, RZX). The FTIR spectra of resin
before chromium adsorption (Fig. 7) shows peak at 3405 cm−1 due
to presence of O H and N H stretching vibrations, while peak
at 2923 cm−1 is due to C H group of alkyl chain. The peaks at
1609 cm−1 and 1648 cm−1 are due to presence of carbonyl groups.
The peak at 1218 cm−1 shows the presence of C N (s) vibration. The
bands at 854 cm−1 and 811 cm−1 are characteristic of C Cl (s) vibra-
tions. The peaks at 764 cm−1 and 705 cm−1 are due to presence of
N H vibrations due to secondary amines. Comparing the FTIR spec-
tra of resin before and after chromium adsorption (Figs. 7 and 8),
there was a marked shift in the band from 3405 cm−1 to 3458 cm−1 .
This confirms the participation of O H and N H groups in Cr–resin
interaction.
The energy dispersive X-ray spectroscopy (FE-SEM JEOL, JSM-
6700F) analysis of resin before and after Cr adsorption is shown in
Figs. 9 and 10. Fig. 9 reveals that resin IRA 96 has the carbon, nitro-
gen and chlorine elements on its surface initially. However, new
chromium peaks were observed on resin surface after the reac-
tion (Fig. 10), which confirmed the chromium sorption onto resin
Amberlite IRA 96.

Fig. 9. EDX analysis of IRA 96 before chromium (VI) adsorption.

 S. Bajpai et al. / Journal of Hazardous Materials 227–228 (2012) 436–444
Fig. 10. EDX analysis of IRA 96 after chromium (VI) adsorption.
4. Conclusions
In this work, response surface methodology approach using
Central Composite Design was applied to develop mathemati-
cal model and optimize process parameters for the hexavalent
chromium removal using weakly anionic resin Amberlite IRA 96.
A 24 CCD model was generated. The experimental design (30 runs)
was obtained using design experiment software. The individual and
combined effect of four process parameters, i.e. contact time, solu-
tion pH, initial Cr (VI) concentration and resin dose were studied
on Cr sorption. A quadratic model was developed to predict the Cr
removal efficiency in terms of individual and interactive operating
parameters.
The quadratic model was analyzed using ANOVA technique. The
model F value was found to be 7.7 with corresponding P value
of 0.0002, and high SS (13,080.82). It implies that the model is
significant and can appropriately explain the relation between
response and independent variables. Individual process parame-
ters were found more significant than the combined one. Initial
solution pH (p = 0.0001, F = 38.88) was found to be the most sig-
nificant parameter for chromium removal followed by the resin
dose (p = 0.0001, F = 26.36). Different response surface plots were
generated to show the combined effect of process variables on
% removal. It was found that removal increased with decreasing
pH. The increase in Cr sorption at acidic pH could be attributed to
strong electrostatic attraction between positively charged amine
groups of resin and negatively charged HCrO4 − ions. However, the
decrease in the removal at higher pH may be attributed to the net
decrease of electrostatic attraction; and competitiveness between
anionic species (CrO4 − , Cr2 O7 −2 ) and the OH− ions in the bulk for
sorption on active sites of resin. The removal was also found to
443
increase with resin dose. It could be explained as increasing resin
dose increases the total number of active sites and overall surface
area of the sorbent particles that removes chromate anions from
the solution. It was found that the overall rate of sorption was con-
trolled by the intraparticle diffusion along with the boundary layer
diffusion.
The optimum process conditions for chromium removal at tem-
perature 30 ◦ C using resin IRA 96 were contact time 62.56 min, pH
1.96, initial Cr (VI) concentration 145.4 mg/L and resin dose 8.51 g/L.
At these conditions high removal (93.26%) was obtained, which
was quite close to the predicted value (94.99%) obtained from RSM
model. The FTIR study confirms the involvement of O H and N H
groups in Cr–resin interaction. Distinct peaks of Cr (VI) in EDX after
Cr sorption confirmed the attachment of chromium onto resin sur-
face.
One of the main limitations of ion exchange process is the high
cost of resin. It is therefore important to optimize the process condi-
tions with a clear objective to achieve maximum removal efficiency
using minimum resin dose and maximum initial Cr concentration.
This study has demonstrated the use of RSM/CCD approach for pre-
dicting the optimal process conditions for the removal of Cr using
synthetic resin in batch process. Based on our results at laboratory
scale, it is possible to design a batch contactor on pilot plant scale
that can work over the broad range of pH 1–7. This batch contactor
will require less than 2 h contact time to achieve 80–95% Cr removal
for a specified range of process conditions.
References
[1] N.M. Gatto, M.A. Kelsh, D. Mai, M. Suh, D.M. Proctor, Occupational exposure
to hexavalent chromium and cancers of the gastrointestinal tract: a meta-
analysis, Cancer Epidemiol. 34 (2010) 388–399.
444 S. Bajpai et al. / Journal of Hazardous Materials 227–228 (2012) 436–444
[2] K.P. Nickens, S.R. Patierno, S. Ceryak, Chromium genotoxicity: a double-edged
sword, Chem. Biol. Interact. 188 (2010) 276–288.
[3] World Health Organization (WHO), Chromium, Environmental Health Criteria
61, Geneva, Switzerland, 1988.
[4] Agency for Toxic Substances and Disease Registry (ATSDR), Toxicological Profile
for Chromium, Atlanta, USA, 1998.
[5] U.S. Environmental Protection Agency (USEPA), Integrated Risk Information
System (IRIS) on Chromium VI, National Center for Environmental Assessment,
Office of Research and Development, Washington, DC, 1999.
[6] U.S. Environmental Protection Agency (USEPA), Toxicological Review of Hex-
avalent Chromium, National Center for Environmental Assessment, Office of
Research and Development, Washington, DC, 1998.
[7] S. Bajpai, G. Jung, A. Dey, Batch studies on chromium removal from tannery
effluent using cationic exchange resin, J. Environ. Res. Dev. 4 (2009) 77–82.
[8] I. Heidmann, W. Calmano, Removal of Cr (VI) from model wastewaters by
electrocoagulation with Fe electrodes, Sep. Purif. Technol. 61 (2008) 15–21.
[9] M.S. Bhatti, A.S. Reddy, R.K. Kalia, A.K. Thukral, Modeling and optimization
of voltage and treatment time for electrocoagulation removal of hexavalent
chromium, Desalination 269 (2011) 157–162.
[10] A. Cassano, R. Molinari, M. Romano, E. Drioli, Treatment of aqueous effluents
of the leather industry by membrane processes: a review, J. Membr. Sci. 181
(2001) 111–126.
[11] M. Bansal, D. Singh, V.K. Garg, A comparative study for the removal of hexava-
lent chromium from aqueous solution by agriculture wastes’ carbons, J. Hazard.
Mater. 171 (2009) 83–92.
[12] S.S. Baral, N. Das, G.R. Chaudhury, S.N. Das, A preliminary study on the adsorp-
tive removal of Cr (VI) using seaweed (Hydrilla verticillata), J. Hazard. Mater.
171 (2009) 358–369.
[13] J. Anandkumar, B. Manda, Removal of Cr (VI) from aqueous solution using Bael
fruit (Aegle marmelos Correa) shell as an adsorbent, J. Hazard. Mater. 168 (2009)
633–640.
[14] R. Mauri, R. Shinnar, M.D. Amore, P. Giordana, A. Volpe, Slovent extraction of
chromium and cadmium from contaminated soils, AIChE J. 47 (2001) 509–512.
[15] G. Busso, V. Martinez, G.C. Aguilera, S.G. Granado, E.T. Reyes, A.A. Ordaz, A
comparative study of global hexavalent chromium removal by chemical and
electrochemical process, Electrochem. Commun. 11 (6) (2009) 1097–1100.
[16] V. Neagu, S. Mikhalovsky, Removal of hexavalent chromium by new
quaternizedcrosslinkedpoly(4-vinylpyridines), J. Hazard. Mater. 183 (2010)
533–540.
[17] E. Pehlivan, S. Cetin, Sorption of Cr (VI) ions on two Lewatit-anion exchange
resins and their quantitative determination using UV–visible spectrophotome-
ter, J. Hazard. Mater. 163 (2009) 448–453.
[18] M. Amini, H. Younesi, N. Bahramifar, Biosorption of nickel (II) from aque-
ous solution by Aspergillus niger: response surface methodology and isotherm
study, Chemosphere 75 (2009) 1483–1491.
[19] U.K. Garg, M.P. Kaur, D. Sud, V.K. Garg, Removal of nickel (II) from aqueous
solution by adsorption on agricultural waste biomass using response surface
methodological approach, Bioresour. Technol. 99 (2008) 1325–1331.
[20] E. Sasikumar, T. Viruthagiri, Optimization of process conditions using response
surface methodology (RSM) for ethanol production from pretreated sugarcane
bagasse: kinetics and modeling, Bioenerg. Res. 1 (2008) 239–247.
[21] J.N. Sahu, J. Achary, B.C. Meikap, Response surface modeling and optimization of
chromium(VI) removal from aqueous solution using Tamarind wood activated
carbon in batch process, J. Hazard. Mater. 172 (2009) 818–825.
[22] APHA, Standard Methods for the Examination of Water and Wastewater, 18th
ed., APHA, Washington, DC, 1992.
[23] I. Mukherjee, P.K. Ray, A review of optimization techniques in metal cutting
processes, Comp. Ind. Eng. 50 (2006) 15–34.
[24] M. Giovanni, Response surface methodology and product optimization, J. Food
Technol. 37 (1983) 41–45.
[25] M.S. Bhatti, D. Kapoor, R.K. Kalia, A.S. Reddy, A.K. Thukral, RSM and ANN
modeling for electrocoagulation of copper from simulated waste water: multi
objective optimization using genetic algorithm approach, Desalination 274
(2011) 74–80.
[26] K. Anupam, S. Dutta, C. Bhattacharjee, S. Datta, Adsorptive removal of
chromium (VI) from aqueous solution over powdered activated carbon: opti-
mization through response surface methodology, Chem. Eng. J. 173 (2011)
135–143.
[27] M.B. Baskan, A. Pala, A statistical experiment design approach for arsenic
removal by coagulation process using aluminum sulfate, Desalination 254
(2010) 42–48.
[28] P. Tripathi, V.C. Srivastava, A. Kumar, Optimization of an azo dye batch
adsorption parameters using Box–Behnken design, Desalination 249 (2009)
1273–1279.
[29] R. Gottipati, S. Mishra, Process optimization of adsorption of Cr (VI) on activated
carbons prepared from plant precursors by a two-level full factorial design,
Chem. Eng. J. 160 (2010) 99–107.
[30] D.C. Montgomery, Response Surface Methodology, 2nd ed., John Wiley and
Sons, New York, USA, 2002.
[31] K. Ravikumar, S. Krishnan, S. Ramalingam, K. Balu, Optimization of process vari-
ables by the application of response surface methodology for dye removal using
a novel adsorbent, Dyes Pigments 72 (2007) 66–74.
[32] X. Jing, Y. Cao, X. Zhang, D. Wang, X. Wu, H. Xu, Biosorption of Cr (VI) from
simulated wastewater using a cationic surfactant modified spent mushroom,
Desalination 269 (2011) 120–127.
[33] K.J. Cronje, K. Chetty, M. Carsky, J.N. Sahu, B.C. Meikap, Optimization of
chromium(VI) sorption potential using developed activated carbon from sugar-
cane bagasse with chemical activation by zinc chloride, Desalination 275 (2011)
276–284.
[34] B. Mukhopadhyay, J. Sundquist, E. White, Hydro-geochemical controls on
removal of Cr (VI) from contaminated groundwater by anion exchange, Appl.
Geochem. 22 (2007) 370–387.
[35] V. Sarin, K.K. Pant, Removal of chromium from industrial waste by using euca-
lyptus bark, Bioresour. Technol. 97 (2006) 15–20.
[36] S. Chen, Q. Yue, B. Gao, X. Xu, Equilibrium, kinetic adsorption study of adsorptive
removal of Cr(VI) using modified wheat residue, J. Colloid Interface Sci. 349
(2010) 256–264.
[37] K.Z. Elwakeel, Removal of Cr (VI) from alkaline aqueous solution using
chemically modified magnetic chitosan resins, Desalination 250 (2010)
105–112.
[38] A.H. Bandegharaei, M.S. Hosseini, M.S. Ghadi, S. Zowghi, E. Hosseini, H.H.
Bandegharaei, Kinetics, equilibrium and thermodynamic study of Cr (VI) sorp-
tion into toluidine blue o-impregnated XAD-7 resin beads and its application
for treatment of wastewater containing Cr (VI), Chem. Eng. J. 160 (2010)
190–198.
[39] H. Zhang, Y. Tang, D. Cai, X. Liu, X. Wang, Q. Huang, Z. Yu, Hexavalent chromium
removal from aqueous solution by algal bloom residue derived activated car-
bon: equilibrium and kinetic studies, J. Hazard. Mater. 181 (2010) 801–808.