Documente Academic
Documente Profesional
Documente Cultură
PII: S0927-7757(15)30131-X
DOI: http://dx.doi.org/doi:10.1016/j.colsurfa.2015.07.063
Reference: COLSUA 20089
Please cite this article as: Safoora Rahimi, Rozita M.Moattari, Laleh Rajabi,
Ali Ashraf Derakhshan, Optimization of lead removal from aqueous solution
using goethite/chitosan nanocomposite by response surface methodology,
Colloids and Surfaces A: Physicochemical and Engineering Aspects
http://dx.doi.org/10.1016/j.colsurfa.2015.07.063
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Optimization of Lead Removal from Aqueous Solution Using
Graphical abstract
1
Highlights:
The removal efficiency of nanocomposite was greater than pure chitosan film.
Abstract
This work investigates goethite/chitosan nanocomposites for their use in lead removal from
aqueous solutions. Goethite nanoparticles were synthesized and characterized by FTIR, DLS and
SEM. Goethite nanoparticles were identified as nanospheres with the average diameter of 10-60
nm. The optimum conditions were determined using response surface methodology (RSM) based
initial solution pH, adsorbent mass and initial concentration of Pb (II) ions on the removal
efficiency for Pb (II) ions were evaluated. The optimal conditions for the lead removal were
found to be 6, 0.05 g and 74.4 mg/L, for the initial solution pH, adsorbent mass and the initial
concentration of Pb (II) ions, respectively. Under these conditions, maximum lead removal
efficiency was obtained to 98.26% that was in respectable agreement with the model (97.19%).
The modified quadratic model exhibited excellent stability for Pb(II) adsorption by
nanocomposite revealed that Pb(II) uptake was enhanced by chitosan film using goethite
nanoparticles.
1. Introduction
2
Nowadays, environmental pollution which is a result of rapid technological development is a
serious apprehension for ecosystem. Some pollutants like heavy metals rarely disappear; they are
harmful to humans, animals, and other living creatures. Lead is known as one of the most toxic
heavy metals which is typically resulted from the industrial wastes of the lead mining, lead smelting,
manufacturing, rubber production, etc. . High-level lead exposure can adversely damage the brain
and kidneys and even cause death. Because of the high toxicity of lead ion, it’s imperative to remove
the lead ion from waste water before discharging it into the environment. For this purpose, A variety
of methods have been employed to remove lead ions from industrial wastewaters, such as solvent
extraction, precipitation and coagulation, biosorption, membrane filtration, chemical absorption, low
energy reverse osmosis, adsorption and so on [1-3]. Among these mentioned methods, adsorption is
widely used due to its high removal efficiency, easy handling, high selectivity and lower operating
cost even at very low concentrations of lead. To date, different types of adsorbents such as
zeolites, metallic oxides, activated carbon, ion exchange resins, polymeric adsorbents and
different biosorbents have been employed for lead removal from waste water. Low adsorption
capacity of these adsorbents usually restricts their large-scale application in water treatment.
Therefore, there is a crucial need for new adsorbents with characteristics such as high adsorption
Polysaccharide are known as functional and biocompatible materials with a great capability to be
crosslinked with nano and micro structures *Gao. Chitosan, a natural polysaccharide-based
polymer obtained from chitin, is popular in various applications due to its nontoxicity and
biodegradability [4]. It is well known as a low cost adsorbent for heavy metal removal with a
high adsorption capacity consists a large number of functional groups such as amino (NH2) and
hydroxyl (OH) groups [5]. Nowadays, preparation of chitosan nanocomposite with better
3
mechanical and chemical properties from that of the pure chitosan film have been the focus of
attention of several research groups [6, 7]. Activated clay [8], poly vinyl alcohol, poly vinyl
chloride, kaolinite [9], perlite [10], glyoxal, formaldehyde, glutaraldehyde, epichlorohydrin, oil
palm ash [11], ethylene glycon diglycidyl ether and isocyanates [12] montmorillonite [13] and
bentonite [14] have been added as cross-linking agents to improve adsorption properties of
chitosan nanocomposites. Tao and his coworkers reported the use of TiO2 in forming
chitosan/TiO2 hybrid film for the removal of pb(II) from aqueous solution. They used Box–
Behnken model and indicated that the reaction parameter optimization using response surface
method is scientific and valid [15]. RSM is essentially a particular collection of mathematical
and statistical techniques for designing experiments, building models, evaluating the effects of
variables, and optimizing process. Its greatest advantage is reducing the number of experimental
trials required to evaluate numerous parameters and their interactions. This methodology can be
conditions on the removal process. In recent years, RSM has been applied to optimize and assess
[16-18] .
This paper reports on chitosan, containing new functional groups in order to increase the density
of adsorption sites. Goethite nanoparticles were synthesized and applied as the reinforcing agent
to improve the adsorption properties of the chitosan film. Three operating parameters including
initial solution pH (pH), nanoparticle dose (Cn) and initial concentration of Pb(II) ions (C0) were
investigated. The Box–Behnken model was used to statistically design the experiments to assess
4
2. Experimental
Chitosan powder with low molecular weight was purchased from Sigma-Aldrich (USA). NaOH,
HNO3, acetic acid, ethanol and lead stock solution [NIST Pb(NO3)2 in HNO3 1000 ppm] were
purchased from Merck, Germany. All other reagents used, were of analytical grade and
purchased from Merck, Germany. All the solutions were prepared with deionized water.
Goethite was synthesized from the reaction of Fe(NO3)3·9H2O and KOH [19]. The obtained
suspension was sonicated for 30 minutes at room temperature and then placed in the oven for 70
each containing various amounts of goethite were prepared in the following manner. As
described in detail in Table 1, A certain amount of goethite and 0.2 g of chitosan powder were
dissolved in 1% (v/v) of acetic acid. The resulting suspension was bath-sonicated for 10 min to
Table 1
The resulting solutions were incubated at room temperature under 180 rpm for 24 h and placed in
the pre-heated oven at 50°C for 48 h to help releasing the entrapped air bubbles. The obtained
solutions were casted on glass plates using a casting knife and dried at room temperature. The
films were immersed into alkaline solution (1M NaOH) at ambient temperature for 1 h in order
to minimize the solubility of the films in water and also neutralize the excess acid. The casted
films were thoroughly washed with deionized water and dried at room temperature for 24 hours.
5
Pure chitosan films were also prepared using the same procedure. The choice of goethite for
preparation of nanocomposite against lepidocrocite is discussed later in the results and discussion
section.
2.3. Characterization
infrared (Bruker alpha, German) spectra were recorded between 400 and 4000 cm-1 with KBr
pellets at room temperature. SEM images of all samples were taken, using Philips XL-30S FEG
and LEQ 1450 VP. Dynamic light scattering analysis (DLS, Malvern Instruments, UK) was
carried out to determine size distribution and average particle size of nanoparticles. Samples for
Dilution of the standard lead solution (1000 mg L-1) with deionized water provided required
initial solutions with appropriate concentration. A known amount of the adsorbent (0.02 g) was
Erlenmeyer flask with a constant agitation rate of 180 rpm at ambient temperature. The pH of the
solutions was adjusted to the required value through the addition of HNO3 (1 M) and NaOH (1
M) solutions. The samples were then filtered to separate the nanocomposite from aqueous
AA-6300). Lead removal (%) by nanocomposite was determined according to Eq. (1):
6
Where R% is removal percentage and q t is the amount of lead uptake by the adsorbent in mgg−1.
C0 and Ct are the initial and final metal ion concentrations in mgL−1 in the solution, respectively.
V and m indicate the volume of solution (L) and weight of the adsorbent (g), respectively.
As one of the RSM designs, Box–Behnken design is known as a modified central composite
experimental design [20]. It is demonstrated that Box–Behnken design is more efficient and
requires fewer experiments in comparison to other RSM designs. The benefits of Box–Behnken
designs include the fact that they are all spherical designs and require factors to be run at only
three levels. It can be noted that a number of additional experiments as well as time consuming
and laborious laboratory studies will be eliminated by selecting the Box–Behnken experimental
Design Expert 7 software was used for regression and graphical analysis of the obtained data.
For statistical calculations, the three independent variables were designed as X1, X2 and X3 with
the coded values at three levels: -1, 0 and +1. The effects of three variables on the removal
Fig. 1 shows the FTIR spectra of goethite nanoparticle, pure chitosan film and goethite/chitosan
nanocomposite. For the goethite spectrum, the two index peaks at 3400 and 3150 cm-1 are
7
assigned to the H-O-H vibration which are related to non-stoichiometric hydroxyl units (excess
water) in the goethite structure. The band at 1634 cm−1 was assigned to the water bending
vibration. Strong absorption peaks at 890 cm−1 and 790 cm−1 caused by the in-plane bending of
surface hydroxyl of Fe-OH-Fe, were similar to the one reported in literature [22, 23]. The main
index peaks of goethite nanoparticles in the spectrum of the gothite/chitpsan nanocomposite are
nanocomposite structure.
Fig. 1
The average particle size of goethite nanoparticles had been measured to be 145.6 nm through
DLS analysis, smaller than what was reported by Roze and his co-workers [24].
Fig. 2 shows SEM micrographs of goethite nanoparticles. The goethite nanoparticles are
Fig. 2
Fig. 3 shows the SEM micrographs from the surface of the goethite/chitosan nanocomposite and
also the three dimensional structure of it. Goethite nanoparticles are dispersed in chitosan
polymeric matrix and the size of particles, including the aggregated ones is given in the picture.
Fig. 3
Fig. 4
structure. Goethite nanoparticles act as cross-linking agents to link the biopolymer chitosan
chains through the formation of both strong primary covalent bonds as well as hydrogen
8
bondings. As a matter of fact, goethite nanoparticles act as both reinforcing agent and adsorbent
Fig. 5
techniques that are useful for modelling and analysis of problems in which output or response is
influenced by several input variables and the objective is to find the correlation between the
response and the variables investigated [25]. A polynomial regression equation was developed
by using Box–Behnken design to analyze the factor interactions. The complete design matrix
together with observed and predicted experimental response values are given in Table 2. Three
Table 2
In order to fit an empirical second-order polynomial, model response function (Y) for predicting
Where Y is a response variable of removal efficiency (%); b 0 is the constant coefficient, bi are
the regression coefficients for linear effects, bii and bij are the square and interaction effects,
respectively. Xi and Xj are the coded experimental levels of the variables and k is the number of
9
the independent variables . The software Design Expert 7 and Minitab 16, were used for the
experimental design, determination of the coefficients, the data analysis and the graph plotting.
By comparing the experimental and predicted values the reliability of the model and the
credibility of the statistical evaluations were determined. The effects of process variables
including initial pH, nanoparticle dose, and initial lead concentration on the lead removal
efficiency were investigated using RSM according to BBD. Different response terms such as
linear, interactive, quadratic and cubic models were used to correlate the experimental data and
to obtain the regression equation. To decide about the competence of the obtained models to
model sum of squares, lack of fit tests, and model summary statistics) were carried out in the
Table 3
From Table 3, it is evident that quadratic model is the most suitable for the removal of Pb(II) by
the nanocomposite. The competence, significance and compatibility of the model was further
The ANOVA for the quadratic model for lead inos adsorption onto nanocomposite is tabulated in
Table 4. All terms in the regression models were not equally important. The significance of each
coefficient was determined by F-value and p-values, which are listed below. In general, the
larger the magnitude of the F-value and the smaller the p-value, the more significant is the
corresponding
coefficient term. The Model F-value of 36.40 implied its significance. In this case initial pH
(X1), nanoparticle dose (X2), initial concentration of Pb 2+ ions (X3) and interactions X1X3, X12,
10
X22 and X32 were significant model terms. The p-values ≥0.050 indicated the model terms that
were not significant. So it is better model reduction was employed to improve model; for this
purpose X1X2 and X2X3 were ruled out. It was observed that, model F-value reduced, so X1X2
was reconsidered. With this improvement, model F-value was 47.15 (Table 5). The correlation
coefficients R2 and R2adj were computed to check the adequacy of the model. In statistical
large value of R2 does not necessarily imply that the regression model is a good one. Hence, R2adj
is preferred to be used to determine the fit of a regression model, as it does not always increase
when variables are added [26, 27]. In the current work, the high value of R2 (0.984)
demonstrated a high dependence and correlation between the observed and the predicted values
of response. The value of R 2adj (0.963) indicated that the total variation of about 96% for lead
removal was attributed to the independent variables and only about 4% of the total variation
cannot be explained by the model. Eq. (3) shows the response functions with the determined
coefficients for Pb(II) removal; the initial pH of solution (X1), nanoparticle dose (X2) and initial
concentration (X3) are represented in terms of coded factors (-1, 0 and +1).
(3
It can be seen from the coefficients in Eq. (3) that removal efficiency increases with the pH (X1)
and Cn (X2) and decreases with C0 (X3). Initial pH (X1) has a more profound effect on lead
adsorption as compared to nanoparticle dose (X2) and initial concentration (X3), which is in
Table 4
11
Table 5
Fig. 6
The statistical analysis above verified the appropriate fit of model to the observed lead removal
efficiency. However, these values do not imply the adequacy of the model for its intended
application. This would require a basic diagnostic check of model adequacy. Generally, observed
Comparison of Equations (4) and (5) suggest that the fitted model is close the true model when
the residuals are close to random errors. Random errors are defined as a sequence of independent
and normally distributed observations. The randomness and normality diagnosis of the residuals
from the fitted model and observed adsorption data form the basis for judging the
appropriateness of the fitted model [28]. Fig. 7 shows histogram error plot indicating error
Fig. 7
The fitted quality of Eq. (3) was also expressed by comparing lead removal efficiency between
experimental and predicted values, as shown in Fig. 8. The better the fit of the model, the smaller
the values of residuals is, more to the point, residuals should be normally distributed [29]. It is
clear from Fig. 8 that the predicted values are quite close to the actual experiment, thus
12
confirming that the regression model exhibits excellent stability for Pb(II) adsorption on
goethite/chitosan nanocomposite. Therefore, it can be concluded that the response surface model
developed in this study (Eq. (3)) was considered to be satisfactory for the prediction of Pb(II)
adsorption system.
Fig. 8
3.5. Comparative effects of media components on Pb (II) removal efficiency
The perturbation plot was used to compare the effects of all the parameters at a point in the
design
space on the response (Fig. 9). A sharp slope for solution pH shows that the response of Pb(II)
removal efficiency was very sensitive to this parameter. The nearly flat curves for nanosorbent
dose indicated that removal efficiency was insensitive to this factor as compared to the solution
pH. Furthermore, perturbation plot for initial concentration (C) shows that this factor can affect
adsorption process considerably; although this is not as influential as pH. It was clear from the
perturbation plot that the most significant factor on the response was the solution pH.
Fig. 9
3.6. Contour plots and response surface analysis
The three-dimensional response surface plots and two-dimensional contour plots are the
graphical representations of the regression equation. These types of plots demonstrate the effects
of two factors on the response at a time. Therefore, in this work 3D response surface plots for the
measured responses were formed based on the model Eq. (3). The relationship between the
dependent and independent variables was further illuminated by constructing contour plots.
Figures 10 and 11 show the 3D response surfaces and the corresponding contour plots as the
13
Fig. 10
Fig. 11
3.7. Effects of model components and their interactions on Pb(II) removal efficiency
The solution pH value plays an important role in adsorption process and specifically on the
adsorption capacity of the adsorbent. Solution pH would affect both aqueous chemistry and
surface binding sites of the adsorbents, thus, changing solution pH could modify the surface
charge of an adsorbent. Based on the electron donating nature of the amine (–NH2) and hydroxyl
(–OH) groups in chitosan and the electron accepting nature of Pb2+ ions, it seems that the ion
exchange mechanism could be preferentially considered. In the lower pH region the positively
charged sites dominate and the H+ ions compete with Pb2+ cations for the exchange sites on the
sorbent surface. While the solution pH increases, the number of negatively charged sites
increases, which results in a lower coulombic repulsion of the sorbing metal. In this part of the
the current work, the pH range of 3 - 6 was assayed for lead removal by goethite/chitosan-
nanocomposite. pH 3 was chosen for lower bound due to solubility of chitosan film in water at
pHs lower than 3. When the initial pH of the lead solution with concentration of 150 ppm was
adjusted to values higher than 6.3, lead precipitation (Pb (OH)2) occurred due to the existence of
OH− ions in the adsorption medium. Thus the pH 6.3 was selected as the upper limit bound.
Fig. 10a and b shows the combined effects of pH correspond with nanoparticle dose and initial
concentration, respectively. As seen in fig. 10b at high lead concentrations, altering the solution
pH significantly affected removal efficiency, and this phenomenon is attributed to the influence
of pH in the presence of OH3+ cations. Furthermore, interaction of nanoparticle dose and solution
14
It can be depicted from the response graphs that the metal removal efficiency is dependent on the
initial metal concentration, in a way that, as the concentration increased the removal efficiency
decreased. At high lead concentrations, metal ions occupy active sites of adsorbent quickly, thus
reducing the number of available adsorption sites. It can be said that, the adsorbent surface is
saturated by lead ions, and this fact prevents the efficient ion adsorption by adsorbent [30].
Fig. 12 shows the mechanism of Pb(II) adsorption onto the goethite/chitosan nanocomposite.
Both ion exchange and complex adsorptions may occur through the adsorption process. Fig. 12a
and b shows complex formation and ion exchange adsorption mechanisms, respectively.
Fig. 12
The optimum values of the selected test variables were obtained by solving the Eq. (3) and also
by analyzing the response surface contour plots. The optimum variables were found to be 6 (X1=
1) for initial pH of the solution, 74.4 ppm (X2 = -0.51) for initial concentration of Pb(II) ions,
and 0.05 g (X3 =0.026) for nanoparticle dose with a predicted Pb(II) removal efficiency of about
97.19%. Later, confirming experiment was carried out to assess the predicted result. The
experimental value was 98.26%, which was in well agreement with the predicted value. The
result showed a very good consistency and this indicates that there is a good concurrence
15
Pb(II) adsorption experiment by pure chitosan film at pH=6, C0= 50, contact time=24 h, with
agitation rate of 180 rpm at ambient temperature was performed. The removal efficiency and the
adsorption capacity were found to be 58.58 and 36.61, respectively; while removal efficiency
and the adsorption capacity in the same condition for goethite/chitosan-nanocomposite (Cn=
0.046) were 98.42 and 61.51, respectively (run 6 of Table 2). A comparison between these data
shows an improvement in Pb(II) uptake by chitosan film using goethite nanoparticles (Fig. 13).
Fig. 13
4. Conclusions
Goethite nanoparticles acted as both nanofiller and adsorbent for the chitosan polymer matrix.
The modified quadratic model exhibited excellent stability for Pb(II) adsorption by
goethite/chitosan nanocomposite. The response surface model developed in this study was
considered to be satisfactory for the prediction of Pb(II) adsorption system R2=0.984. Increasing
the solution pH significantly increased removal efficiency. Pb(II) uptake was enhanced by
chitosan film using goethite nanoparticles. The obtained experimental value in optimum
conditions was 98.26% that was in well agreement with the predicted value (97.19%).
16
References:
[1] M.A. Salam, Coating carbon nanotubes with crystalline manganese dioxide nanoparticles and
their application for lead ions removal from model and real water, Colloids and Surfaces A:
microgel as selective adsorbent for removal of heavy metals from wastewater, Colloids and
[3] S. Duan, R. Tang, Z. Xue, X. Zhang, Y. Zhao, W. Zhang, J. Zhang, B. Wang, S. Zeng, D.
Sun, Effective removal of Pb (II) using magnetic Co0.6 Fe2.4O 4 micro-particles as the adsorbent:
Synthesis and study on the kinetic and thermodynamic behaviors for its adsorption, Colloids and
[4] Y. Yang, Q. He, L. Duan, Y. Cui, J. Li, Assembled alginate/chitosan nanotubes for biological
[5] Q. Zhou, Q. Gao, W. Luo, C. Yan, Z. Ji, P. Duan, One-step synthesis of amino-functionalized
of methylene blue from wastewater, Colloids and Surfaces A: Physicochemical and Engineering
[6] M.-S. Chiou, P.-Y. Ho, H.-Y. Li, Adsorption of anionic dyes in acid solutions using
[7] Z. Zhu, J. Wang, A. Munir, H.S. Zhou, Direct electrochemistry and electrocatalysis of
Colloids and Surfaces A: Physicochemical and Engineering Aspects, 385 (2011) 91-94.
17
[8] M.-Y. Chang, R.-S. Juang, Adsorption of tannic acid, humic acid, and dyes from water using
the composite of chitosan and activated clay, Journal of Colloid and Interface Science, 278
(2004) 18-25.
[9] H.-Y. Zhu, R. Jiang, L. Xiao, Adsorption of an anionic azo dye by chitosan/kaolin/γ-Fe2O3
[10] S. Kalyani, J.A. Priya, P.S. Rao, A. Krishnaiah, Removal of copper and nickel from aqueous
solutions using chitosan coated on perlite as biosorbent, Separation science and technology, 40
(2005) 1483-1495.
[11] M. Hasan, A. Ahmad, B. Hameed, Adsorption of reactive dye onto cross-linked chitosan/oil
palm ash composite beads, Chemical Engineering Journal, 136 (2008) 164-172.
[12] G. Crini, P.-M. Badot, Application of chitosan, a natural aminopolysaccharide, for dye
removal from aqueous solutions by adsorption processes using batch studies: A review of recent
[13] L. Wang, J. Zhang, A. Wang, Removal of methylene blue from aqueous solution using
[14] W.S.W. Ngah, N.F.M. Ariff, M.A.K.M. Hanafiah, Preparation, characterization, and
aqueous solutions, Water, Air, and Soil Pollution, 206 (2010) 225-236.
[15] Y. Tao, L. Ye, J. Pan, Y. Wang, B. Tang, Removal of Pb (II) from aqueous solution on
[16] B.K. Körbahti, M. Rauf, Response surface methodology (RSM) analysis of photoinduced
18
[17] S. Sharma, A. Malik, S. Satya, Application of response surface methodology (RSM) for
in an aqueous solution by Aspergillus niger, Journal of hazardous materials, 154 (2008) 694-702.
[19] S. Rahimi, R.M. Moattari, L. Rajabi, A.A. Derakhshan, M. Keyhani, Iron oxide/hydroxide
(α, γ-FeOOH) nanoparticles as high potential adsorbents for lead removal from polluted aquatic
[20] J. Ahmad, S.R. Mir, K. Kohli, S. Amin, Effect of oil and co-surfactant on the formation of
Solutol HS 15 based colloidal drug carrier by Box–Behnken statistical design, Colloids and
[21] K. Yetilmezsoy, S. Demirel, R.J. Vanderbei, Response surface modeling of Pb (II) removal
from aqueous solution by Pistacia vera L.: Box–Behnken experimental design, Journal of
[22] S. Liao, J. Wang, D. Zhu, L. Ren, J. Lu, M. Geng, A. Langdon, Structure and Mn2+
[23] H. Ruan, R.L. Frost, J.T. Kloprogge, L. Duong, Infrared spectroscopy of goethite
[24] M.a.M. Cortalezzi, J. Rose, G.F. Wells, J.-Y. Bottero, A.R. Barron, M.R. Wiesner, Ceramic
membranes derived from ferroxane nanoparticles: a new route for the fabrication of iron oxide
19
[25] M. Pradhan, C. Biswas, Modeling and Analysis of process parameters on Surface
immobilized cell bioreactor, Journal of Environmental Science and Health, Part A, 48 (2013)
1557-1572.
[27] Z. Zhang, H. Zheng, Optimization for decolorization of azo dye acid green 20 by ultrasound
and H2O2 using response surface methodology, Journal of hazardous materials, 172 (2009) 1388-
1393.
[28] K. Zhang, W. Cheung, M. Valix, Roles of physical and chemical properties of activated
[29] M.A. Islam, V. Sakkas, T.A. Albanis, Application of statistical design of experiment with
desirability function for the removal of organophosphorus pesticide from aqueous solution by
Nazari, Taguchi optimization approach for Pb (II) and Hg (II) removal from aqueous solutions
using modified mesoporous carbon, Journal of hazardous materials, 192 (2011) 1046-1055.
Figures
Figure 1. FTIR spectra of goethite nanoparticle, pure chitosan film and goethite/chitosan
nanocomposite
20
Figure 4. Cross-sectional SEM image of goethite/chitosan nanocomposite
Figure 5. The proposed structure of goethite /chitosan nanocomposite made from chitosan
goethite/chitosan nanocomposite.
Figure 7. Histogram plot of errors, Mean: mean of errors, StDeV: standard deviation, N: number
of experiments
Figure 9. Perturbation plots; (A) initial solution pH, (B) nanoparticle dose and (C) initial lead
concentration
Figure 10. 3-D surface plots for interactive effect of (a) pH and adsorbent dose while initial
concentration was adjusted in 100 mg/L (b) pH and initial concentration while Cn was adjusted
in 0.046 g
Figure 11. Contour plots exhibiting the interactive effects between two independent variables
(other variables were held at their respective center levels); (a) initial pH of solution (pH, X1)
and nanosorbent dose (Cn, X2), (b) initial pH of solution (pH, X1) and initial concentration of
Figure 12. Mechanism of Pb (II) adsorption onto the goethite/chitosan nanocomposite, (a)
Figure 13. Comparison between removal efficiency and capacity of pure chitosan and goethite/
22
Figure. 2
23
Figure. 3
24
Figure. 4
25
Figure. 5
26
Figure. 6
27
Figure. 7
28
Figure. 8
29
Perturbation
99
A
89
C
R%
79
A B
B C
69
59
Figure. 9
30
Figure. 10
31
Figure. 11
32
Figure. 12
33
Figure. 13
Tables
Table 2. Complete design matrix with observed and predicted experimental response values
Table 3. Sequential model fitting for the lead adsorption on goethite/chitosan nanocomposite
34
Table 4. ANOVA for Response Surface Full Quadratic Model
35
Table 1
36
Table 2
Run Independent variables Removal
pH, x1 Nanoparticle initial efficiency (%)
dose, x2 concentration, x3
Coded Actual Coded Actual Coded Actual Observed Predicted
(g) (mg/l)
1 -1 3 -1 0.002 0 100 64.94 62.87
2 1 6 -1 0.002 0 100 88.18 88.04
3 -1 3 1 0.09 0 100 72.61 72.74
4 1 6 1 0.09 0 100 86.79 88.86
5 -1 3 0 0.046 -1 50 82.43 82.35
6 1 6 0 0.046 -1 50 98.43 96.42
7 -1 3 0 0.046 1 150 60.58 62.59
8 1 6 0 0.046 1 150 89.72 89.80
9 0 4 -1 0.002 -1 50 71.98 73.60
10 0 4.5 1 0.09 -1 50 78.48 78.95
11 0 4.5 -1 0.002 1 150 59.84 60.42
12 0 4.5 1 0.09 1 150 68.43 65.76
13 0 4.5 0 0.046 0 100 80.97 80.80
14 0 4.5 0 0.046 0 100 80.40 80.80
15 0 4.5 0 0.046 0 100 81.01 80.80
37
Table 3
Sequential model sum of squares
Sum of DOF Mean square F-value P-value > Comme
Source squares F nt
Mean vs. Total 90447.09 1 90447.09 - - -
Linear vs. Mean 1256.98 3 418.99 9.58 0.0021 -
2FI vs. Linear 64.78 3 21.59 0.41 0.7470 -
Quadratic vs. 390.14 3 130.05 24.89 0.0020 Suggest
2FI ed
Cubic vs. 25.90 3 8.63 74.74 0.0132 Aliased
Quadratic
Residual 0.23 2 0.12 - - -
Total 92185.13 15 6145.68 - - -
Lack of Fit Tests
Sum of DO Mean square F-value P-value > Remark
Source squares F F
Linear 480.82 9 53.42 462.53 0.0022 -
2FI 416.04 6 69.34 600.33 0.0017 -
Quadratic 25.90 3 8.63 74.74 0.0132 Suggest
ed
Cubic 0.000 0 - - - Aliased
Pure Error 0.23 2 0.12 - - -
Model Summary Statistics
Std. Dev. R2 R2adj Predicted PRESS Remark
Source R2
Linear 6.61 0.7232 0.6477 0.4243 455.24 -
2FI 7.21 0.7605 0.5809 -0.2037 485.90 -
Quadratic 2.29 0.9850 0.9579 0.7613 269.04 Suggest
ed
Cubic 0.34 0.9999 0.9991 - - Aliased
38
Table 4
Source Sum of DOF Mean F-value p-value > F
squares square
Model 1711.91 9 190.21 36.40 0.0005
X1-pH 852.15 1 852.15 163.07 < 0.0001
X2-C n 57.15 1 57.15 10.94 0.0213
X3-C0 347.68 1 347.68 66.54 0.0004
X1X2 20.48 1 20.48 3.92 0.1046
X1X3 43.21 1 43.21 8.27 0.0348
X2 X3 1.09 1 1.09 0.21 0.6673
X12 100.59 1 100.59 19.25 0.0071
X22 229.67 1 229.67 43.95 0.0012
X32 38.42 1 38.42 7.35 0.0422
Cor Total 1738.04 14 - - -
39
Table 5
40