- Email: [email protected]
chitosan nanocomposite by response surface methodology
Accepted Manuscript Title: Optimization of lead removal from aqueous solution using goethite/chitosan nanocomposite by response surface methodology Author: Safoora Rahimi Rozita M. Moattari Laleh Rajabi Ali Ashraf Derakhshan PII: DOI: Reference:
S0927-7757(15)30131-X http://dx.doi.org/doi:10.1016/j.colsurfa.2015.07.063 COLSUA 20089
To appear in:
Colloids and Surfaces A: Physicochem. Eng. Aspects
Received date: Revised date: Accepted date:
27-5-2015 23-7-2015 29-7-2015
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 Goethite/Chitosan Nanocomposite by Response Surface Methodology Safoora Rahimi1, Rozita. M. Moattari1, Laleh Rajabi*1, 2, Ali Ashraf Derakhshan3 1. Polymer Research Lab., Department of Chemical Engineering, College of Engineering, Razi University, Kermanshah, Iran. 2. Department of Chemistry, University of Victoria (UVic), Victoria, British Columbia, Canada. 3. Young Researchers and Elite Club, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran.
*Corresponding author: Laleh Rajabi Tel: +1 778 440 3144, +1 250 589 2031 Fax: +98 831 4274542 Email: [email protected] [email protected] Graphical abstract
1
Highlights:
Goethite nanoparticles were used as reinforcement agent in nanocomposite structure.
Goethite/chitosan nanocomposites were used as sorbents for lead removal.
Box-bhenken experimental design was used to provide adsorption model.
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 on three-variable-three-level Box–Behnken design (BBD). The effects of three variables, i.e. 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 goethite/chitosan nanocomposite. The results of adsorption study by goethite/chitosan nanocomposite revealed that Pb(II) uptake was enhanced by chitosan film using goethite nanoparticles. Keywords: Adsorption, Goethite, Chitosan, Nanocomposite, Nanoparticle, Box–Behnken design 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, battery manufacturing, printing, pigments, fossil fuels, photographic materials, explosive 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 capacity, easy fixation, and separation from water [1, 3].
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 used in developing suitable treatment technology considering the effects of operational conditions on the removal process. In recent years, RSM has been applied to optimize and assess interactive effects of independent parameters in various chemical and biochemical procedures [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 and optimize adsorption process, using Design Expert 7 software.
4
2. Experimental 2.1. Reagents and materials 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. 2.2. Preparation of goethite/chitosan nanocomposites 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 minutes at 100°C and centrifuged. Later, the modified goethite/chitosan adsorptive nanocomposites were prepared by solution-casting method. Three goethite/chitosan solutions, 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 guarantee complete dissolution of the solutes.
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
Goethite/chitosan nanocomposites were characterized using FTIR and SEM. Fourier-transform 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 DLS analysis were prepared as 5% W/V (nanoparticle/water).
2.4. Adsorption experiments 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 added to 25 ml of the required concentration of Pb (NO3)2 solutions and transferred to a 100 ml 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 solutions. Supernatants were analyzed using an atomic absorption spectroscopy (SHIMADZO 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. 2.5. Box–Behnken experimental design
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, instead of complete factorial design [21].
Pb(II) uptake by goethite/chitosan-nanocomposite was investigated using response surface methodology (RSM) based on three-variable-three-level Box–Behnken design (BBD). The 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 efficiency for Pb2+ were evaluated.
3. Results and discussion 3.1. Characterization of adsorbents 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 highlited in the Fig. 1, clearly indicating the involvement of goethite nanoparticles in 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 spherical with diameter of 10-70 nm. 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 presents the cross sectional SEM micrograph of goethite/chitosan nanocomposite. Thickness of chitosan film was found to be 450-470 µm, approximately. Fig. 4 Fig. 5 shows the the proposed schematic illustration of goethite/chitosan nanocomposite 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 for the chitosan polymer matrix. Fig. 5
3.2. RSM approach and statistical analysis Response surface methodology (RSM) is a combination of mathematical and statistical 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 replicate at the center point is used to determine the experimental error.
Table 2
3.3. Determination of the regression model and statistical evaluation
In order to fit an empirical second-order polynomial, model response function (Y) for predicting the optimal point was constructed in the form Eq. (2):
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 describe lead removal by goethite/chitosan nanocomposite, three different tests (sequential model sum of squares, lack of fit tests, and model summary statistics) were carried out in the present study and the results are presented in Table 3.
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 check through analysis of variance (ANOVA).
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 modeling, by removing a repressor variable, the coefficient of determination decreases and a 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).
Y = 80.80+ 10.32X1 +
2.67X2 - 6.59X3 - 2.26X1X2 + 3.29X1X3 – 7.89 X22 - 3.23X32
(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 agreement with contribution percentages presented in Fig. 6.
Table 4 11
Table 5
Fig. 6
3.4. Diagnostic analysis of fitted model adequacy
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 data can be expressed as follow:
Observed data value = true model + random
(4)
Observed data value = fitted model + residual
(5)
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 density is centralized at zero. 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 functions of two variables at the center level of other variables, respectively. 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/chitosannanocomposite. 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 pH at low Cn values is more considerable in comparison to high Cn values.
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]. Figures 10b and 11b well-describe this fact. 3.8. Mechanism of adsorption by PABF/chitosan nanocomposite 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 3.9. Optimization of variables for removal of Pb(II) 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 between the predicted and the experimental values.
3.10. Adsorption by pure chitosan film
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: Physicochemical and Engineering Aspects, 419 (2013) 69-79. [2] L. Jiang, P. Liu, S. Zhao, Magnetic ATP/FA/Poly (AA-co-AM) ternary nanocomposite microgel as selective adsorbent for removal of heavy metals from wastewater, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 470 (2015) 31-38. [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 Surfaces A: Physicochemical and Engineering Aspects, (2015). [4] Y. Yang, Q. He, L. Duan, Y. Cui, J. Li, Assembled alginate/chitosan nanotubes for biological application, Biomaterials, 28 (2007) 3083-3090. [5] Q. Zhou, Q. Gao, W. Luo, C. Yan, Z. Ji, P. Duan, One-step synthesis of amino-functionalized attapulgite clay nanoparticles adsorbent by hydrothermal carbonization of chitosan for removal of methylene blue from wastewater, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 470 (2015) 248-257. [6] M.-S. Chiou, P.-Y. Ho, H.-Y. Li, Adsorption of anionic dyes in acid solutions using chemically cross-linked chitosan beads, Dyes and Pigments, 60 (2004) 69-84. [7] Z. Zhu, J. Wang, A. Munir, H.S. Zhou, Direct electrochemistry and electrocatalysis of horseradish peroxidase immobilized on bamboo shaped carbon nanotubes/chitosan matrix, 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 composites, Applied Clay Science, 48 (2010) 522-526. [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 literature, Progress in polymer science, 33 (2008) 399-447. [13] L. Wang, J. Zhang, A. Wang, Removal of methylene blue from aqueous solution using chitosan-g-poly (acrylic acid)/montmorillonite superadsorbent nanocomposite, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 322 (2008) 47-53. [14] W.S.W. Ngah, N.F.M. Ariff, M.A.K.M. Hanafiah, Preparation, characterization, and environmental application of crosslinked chitosan-coated bentonite for tartrazine adsorption from 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 chitosan/TiO2 hybrid film, Journal of hazardous materials, 161 (2009) 718-722. [16] B.K. Körbahti, M. Rauf, Response surface methodology (RSM) analysis of photoinduced decoloration of toludine blue, Chemical Engineering Journal, 136 (2008) 25-30.
18
[17] S. Sharma, A. Malik, S. Satya, Application of response surface methodology (RSM) for optimization of nutrient supplementation for Cr (VI) removal by Aspergillus lentulus AML05, Journal of hazardous materials, 164 (2009) 1198-1204. [18] M. Amini, H. Younesi, N. Bahramifar, A.A.Z. Lorestani, F. Ghorbani, A. Daneshi, M. Sharifzadeh, Application of response surface methodology for optimization of lead biosorption 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 media, Journal of Industrial and Engineering Chemistry, (2014). [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 Surfaces A: Physicochemical and Engineering Aspects, 453 (2014) 68-77. [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 hazardous materials, 171 (2009) 551-562. [22] S. Liao, J. Wang, D. Zhu, L. Ren, J. Lu, M. Geng, A. Langdon, Structure and Mn2+ adsorption properties of boron-doped goethite, Applied Clay Science, 38 (2007) 43-50. [23] H. Ruan, R.L. Frost, J.T. Kloprogge, L. Duong, Infrared spectroscopy of goethite dehydroxylation. II. Effect of aluminium substitution on the behaviour of hydroxyl units, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 58 (2002) 479-491. [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 ultrafiltration membranes, Journal of membrane science, 227 (2003) 207-217.
19
[25] M. Pradhan, C. Biswas, Modeling and Analysis of process parameters on Surface Roughness in EDM of AISI D2 tool Steel by RSM Approach, International Journal of Mathematical, Physical and Engineering Sciences, 3 (2009). [26] A. Habibi, F. Vahabzadeh, Formaldehyde degradation by Ralstonia eutropha in an 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) 13881393. [28] K. Zhang, W. Cheung, M. Valix, Roles of physical and chemical properties of activated carbon in the adsorption of lead ions, Chemosphere, 60 (2005) 1129-1140. [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 low-cost material, Journal of hazardous materials, 170 (2009) 230-238. [30] G. Zolfaghari, A. Esmaili-Sari, M. Anbia, H. Younesi, S. Amirmahmoodi, A. GhafariNazari, 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 Figure 2. SEM micrographs of goethite nanoparticles Figure 3. SEM image of goethite/chitosan nanocomposite surface 20
Figure 4. Cross-sectional SEM image of goethite/chitosan nanocomposite Figure 5. The proposed structure of goethite /chitosan nanocomposite made from chitosan biopolymer and goethite nanoparticles Figure 6. Percent contribution of various parameters on lead removal efficiency by goethite/chitosan nanocomposite. Figure 7. Histogram plot of errors, Mean: mean of errors, StDeV: standard deviation, N: number of experiments Figure 8. Correlation of observed and predicted lead adsorption 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 Pb(II) ions (Cn, X3) Figure 12. Mechanism of Pb (II) adsorption onto the goethite/chitosan nanocomposite, (a) complex formation mechanism and (b) ion exchange mechanism Figure 13. Comparison between removal efficiency and capacity of pure chitosan and goethite/ chitosan
films
(pH=6,
C0=
50 21
and
contact
time=24
h)
Figure. 1
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
R%
C 79
A
B
B
C
69
59
-1.000
-0.500
0.000
0.500
1.000
Deviation from Reference Point (Coded Units)
Figure. 9
30
Figure. 10
31
Figure. 11
32
Figure. 12
33
Figure. 13
Tables Table 1. Synthesis details of Goethite/chitosan nanocomposite 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
Table 5. ANOVA for Response Surface Reduced Quadratic Model
35
Table 1 Total weight (g)
10
10
15
Nanoparticle mass (g)
0.046
0.09
0.003
Chitosan mass (g)
0.2
0.2
0.3
Acid mass (g)
9.754
9.71
14.697
36
Table 2 Run pH, x1 Coded 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
-1 1 -1 1 -1 1 -1 1 0 0 0 0 0 0 0
Actual 3 6 3 6 3 6 3 6 4 4.5 4.5 4.5 4.5 4.5 4.5
Independent variables Nanoparticle initial dose, x2 concentration, x3 Coded Actual Coded Actual (g) (mg/l) -1 0.002 0 100 -1 0.002 0 100 1 0.09 0 100 1 0.09 0 100 0 0.046 -1 50 0 0.046 -1 50 0 0.046 1 150 0 0.046 1 150 -1 0.002 -1 50 1 0.09 -1 50 -1 0.002 1 150 1 0.09 1 150 0 0.046 0 100 0 0.046 0 100 0 0.046 0 100
37
Removal efficiency (%) Observed Predicted 64.94 88.18 72.61 86.79 82.43 98.43 60.58 89.72 71.98 78.48 59.84 68.43 80.97 80.40 81.01
62.87 88.04 72.74 88.86 82.35 96.42 62.59 89.80 73.60 78.95 60.42 65.76 80.80 80.80 80.80
Table 3
Source Mean vs. Total Linear vs. Mean 2FI vs. Linear Quadratic vs. 2FI Cubic vs. Quadratic Residual Total
Sum of squares 90447.09 1256.98 64.78 390.14 25.90
Sequential model sum of squares DOF Mean square F-value P-value > Comme F nt 1 90447.09 3 418.99 9.58 0.0021 3 21.59 0.41 0.7470 3 130.05 24.89 0.0020 Suggest ed 3 8.63 74.74 0.0132 Aliased
0.23 92185.13
2 15 DO
Source Linear 2FI Quadratic
Sum of squares 480.82 416.04 25.90
Cubic Pure Error
0.000 0.23
0.12 6145.68 Lack of Fit Tests Mean square F-value
F 9 6 3 0 2
Std. Dev.
R2
Source Linear 2FI Quadratic
6.61 7.21 2.29
0.7232 0.7605 0.9850
Cubic
0.34
0.9999
53.42 69.34 8.63
462.53 600.33 74.74
0.12 Model Summary Statistics R2adj Predicted R2 0.6477 0.4243 0.5809 -0.2037 0.9579 0.7613 0.9991
38
-
-
-
P-value > F 0.0022 0.0017 0.0132
Remark
-
Suggest ed Aliased -
PRESS
Remark
455.24 485.90 269.04
Suggest ed Aliased
-
Table 4 Source Model X1-pH X2-C n X3-C0 X1X2 X1X3 X2 X3 X12 X22 X32 Cor Total
Sum of squares 1711.91 852.15 57.15 347.68 20.48 43.21 1.09 100.59 229.67 38.42 1738.04
DOF
Mean square 190.21 852.15 57.15 347.68 20.48 43.21 1.09 100.59 229.67 38.42 -
9 1 1 1 1 1 1 1 1 1 14
R2 = 0.9850, R2adj =0.9579, R2pre =0.7613
39
F-value
p-value > F
36.40 163.07 10.94 66.54 3.92 8.27 0.21 19.25 43.95 7.35 -
0.0005 < 0.0001 0.0213 0.0004 0.1046 0.0348 0.6673 0.0071 0.0012 0.0422 -
Table 5
Source Model X1-pH X2-Cn X3-C0 X1X2 X1X3 X21 X22 X23 Residual Lack of Fit Pure Error Cor Total
Sum of squares 1710.82 852.15 57.15 347.68 20.48 43.21 100.59 229.67 38.42 27.22 26.98 0.23 1738.04
DOF
Mean square 213.85 852.15 57.15 347.68 20.48 43.21 100.59 229.67 38.42 4.54 6.75 0.12 -
8 1 1 1 1 1 1 1 1 6 4 2 14
R2 = 0.9843, R2adj = 0.9635, R2pre = 0.8225.
40
F-value
p-value > F
47.15 187.87 12.60 76.65 4.52 9.53 22.18 50.63 8.47 58.41 -
< 0.0001 < 0.0001 0.0121 0.0001 0.0777 0.0215 0.0033 0.0004 0.0270 0.0169 -
S0927-7757(15)30131-X http://dx.doi.org/doi:10.1016/j.colsurfa.2015.07.063 COLSUA 20089
To appear in:
Colloids and Surfaces A: Physicochem. Eng. Aspects
Received date: Revised date: Accepted date:
27-5-2015 23-7-2015 29-7-2015
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 Goethite/Chitosan Nanocomposite by Response Surface Methodology Safoora Rahimi1, Rozita. M. Moattari1, Laleh Rajabi*1, 2, Ali Ashraf Derakhshan3 1. Polymer Research Lab., Department of Chemical Engineering, College of Engineering, Razi University, Kermanshah, Iran. 2. Department of Chemistry, University of Victoria (UVic), Victoria, British Columbia, Canada. 3. Young Researchers and Elite Club, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran.
*Corresponding author: Laleh Rajabi Tel: +1 778 440 3144, +1 250 589 2031 Fax: +98 831 4274542 Email: [email protected] [email protected] Graphical abstract
1
Highlights:
Goethite nanoparticles were used as reinforcement agent in nanocomposite structure.
Goethite/chitosan nanocomposites were used as sorbents for lead removal.
Box-bhenken experimental design was used to provide adsorption model.
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 on three-variable-three-level Box–Behnken design (BBD). The effects of three variables, i.e. 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 goethite/chitosan nanocomposite. The results of adsorption study by goethite/chitosan nanocomposite revealed that Pb(II) uptake was enhanced by chitosan film using goethite nanoparticles. Keywords: Adsorption, Goethite, Chitosan, Nanocomposite, Nanoparticle, Box–Behnken design 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, battery manufacturing, printing, pigments, fossil fuels, photographic materials, explosive 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 capacity, easy fixation, and separation from water [1, 3].
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 used in developing suitable treatment technology considering the effects of operational conditions on the removal process. In recent years, RSM has been applied to optimize and assess interactive effects of independent parameters in various chemical and biochemical procedures [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 and optimize adsorption process, using Design Expert 7 software.
4
2. Experimental 2.1. Reagents and materials 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. 2.2. Preparation of goethite/chitosan nanocomposites 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 minutes at 100°C and centrifuged. Later, the modified goethite/chitosan adsorptive nanocomposites were prepared by solution-casting method. Three goethite/chitosan solutions, 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 guarantee complete dissolution of the solutes.
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
Goethite/chitosan nanocomposites were characterized using FTIR and SEM. Fourier-transform 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 DLS analysis were prepared as 5% W/V (nanoparticle/water).
2.4. Adsorption experiments 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 added to 25 ml of the required concentration of Pb (NO3)2 solutions and transferred to a 100 ml 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 solutions. Supernatants were analyzed using an atomic absorption spectroscopy (SHIMADZO 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. 2.5. Box–Behnken experimental design
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, instead of complete factorial design [21].
Pb(II) uptake by goethite/chitosan-nanocomposite was investigated using response surface methodology (RSM) based on three-variable-three-level Box–Behnken design (BBD). The 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 efficiency for Pb2+ were evaluated.
3. Results and discussion 3.1. Characterization of adsorbents 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 highlited in the Fig. 1, clearly indicating the involvement of goethite nanoparticles in 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 spherical with diameter of 10-70 nm. 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 presents the cross sectional SEM micrograph of goethite/chitosan nanocomposite. Thickness of chitosan film was found to be 450-470 µm, approximately. Fig. 4 Fig. 5 shows the the proposed schematic illustration of goethite/chitosan nanocomposite 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 for the chitosan polymer matrix. Fig. 5
3.2. RSM approach and statistical analysis Response surface methodology (RSM) is a combination of mathematical and statistical 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 replicate at the center point is used to determine the experimental error.
Table 2
3.3. Determination of the regression model and statistical evaluation
In order to fit an empirical second-order polynomial, model response function (Y) for predicting the optimal point was constructed in the form Eq. (2):
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 describe lead removal by goethite/chitosan nanocomposite, three different tests (sequential model sum of squares, lack of fit tests, and model summary statistics) were carried out in the present study and the results are presented in Table 3.
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 check through analysis of variance (ANOVA).
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 modeling, by removing a repressor variable, the coefficient of determination decreases and a 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).
Y = 80.80+ 10.32X1 +
2.67X2 - 6.59X3 - 2.26X1X2 + 3.29X1X3 – 7.89 X22 - 3.23X32
(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 agreement with contribution percentages presented in Fig. 6.
Table 4 11
Table 5
Fig. 6
3.4. Diagnostic analysis of fitted model adequacy
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 data can be expressed as follow:
Observed data value = true model + random
(4)
Observed data value = fitted model + residual
(5)
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 density is centralized at zero. 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 functions of two variables at the center level of other variables, respectively. 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/chitosannanocomposite. 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 pH at low Cn values is more considerable in comparison to high Cn values.
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]. Figures 10b and 11b well-describe this fact. 3.8. Mechanism of adsorption by PABF/chitosan nanocomposite 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 3.9. Optimization of variables for removal of Pb(II) 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 between the predicted and the experimental values.
3.10. Adsorption by pure chitosan film
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: Physicochemical and Engineering Aspects, 419 (2013) 69-79. [2] L. Jiang, P. Liu, S. Zhao, Magnetic ATP/FA/Poly (AA-co-AM) ternary nanocomposite microgel as selective adsorbent for removal of heavy metals from wastewater, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 470 (2015) 31-38. [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 Surfaces A: Physicochemical and Engineering Aspects, (2015). [4] Y. Yang, Q. He, L. Duan, Y. Cui, J. Li, Assembled alginate/chitosan nanotubes for biological application, Biomaterials, 28 (2007) 3083-3090. [5] Q. Zhou, Q. Gao, W. Luo, C. Yan, Z. Ji, P. Duan, One-step synthesis of amino-functionalized attapulgite clay nanoparticles adsorbent by hydrothermal carbonization of chitosan for removal of methylene blue from wastewater, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 470 (2015) 248-257. [6] M.-S. Chiou, P.-Y. Ho, H.-Y. Li, Adsorption of anionic dyes in acid solutions using chemically cross-linked chitosan beads, Dyes and Pigments, 60 (2004) 69-84. [7] Z. Zhu, J. Wang, A. Munir, H.S. Zhou, Direct electrochemistry and electrocatalysis of horseradish peroxidase immobilized on bamboo shaped carbon nanotubes/chitosan matrix, 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 composites, Applied Clay Science, 48 (2010) 522-526. [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 literature, Progress in polymer science, 33 (2008) 399-447. [13] L. Wang, J. Zhang, A. Wang, Removal of methylene blue from aqueous solution using chitosan-g-poly (acrylic acid)/montmorillonite superadsorbent nanocomposite, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 322 (2008) 47-53. [14] W.S.W. Ngah, N.F.M. Ariff, M.A.K.M. Hanafiah, Preparation, characterization, and environmental application of crosslinked chitosan-coated bentonite for tartrazine adsorption from 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 chitosan/TiO2 hybrid film, Journal of hazardous materials, 161 (2009) 718-722. [16] B.K. Körbahti, M. Rauf, Response surface methodology (RSM) analysis of photoinduced decoloration of toludine blue, Chemical Engineering Journal, 136 (2008) 25-30.
18
[17] S. Sharma, A. Malik, S. Satya, Application of response surface methodology (RSM) for optimization of nutrient supplementation for Cr (VI) removal by Aspergillus lentulus AML05, Journal of hazardous materials, 164 (2009) 1198-1204. [18] M. Amini, H. Younesi, N. Bahramifar, A.A.Z. Lorestani, F. Ghorbani, A. Daneshi, M. Sharifzadeh, Application of response surface methodology for optimization of lead biosorption 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 media, Journal of Industrial and Engineering Chemistry, (2014). [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 Surfaces A: Physicochemical and Engineering Aspects, 453 (2014) 68-77. [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 hazardous materials, 171 (2009) 551-562. [22] S. Liao, J. Wang, D. Zhu, L. Ren, J. Lu, M. Geng, A. Langdon, Structure and Mn2+ adsorption properties of boron-doped goethite, Applied Clay Science, 38 (2007) 43-50. [23] H. Ruan, R.L. Frost, J.T. Kloprogge, L. Duong, Infrared spectroscopy of goethite dehydroxylation. II. Effect of aluminium substitution on the behaviour of hydroxyl units, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 58 (2002) 479-491. [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 ultrafiltration membranes, Journal of membrane science, 227 (2003) 207-217.
19
[25] M. Pradhan, C. Biswas, Modeling and Analysis of process parameters on Surface Roughness in EDM of AISI D2 tool Steel by RSM Approach, International Journal of Mathematical, Physical and Engineering Sciences, 3 (2009). [26] A. Habibi, F. Vahabzadeh, Formaldehyde degradation by Ralstonia eutropha in an 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) 13881393. [28] K. Zhang, W. Cheung, M. Valix, Roles of physical and chemical properties of activated carbon in the adsorption of lead ions, Chemosphere, 60 (2005) 1129-1140. [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 low-cost material, Journal of hazardous materials, 170 (2009) 230-238. [30] G. Zolfaghari, A. Esmaili-Sari, M. Anbia, H. Younesi, S. Amirmahmoodi, A. GhafariNazari, 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 Figure 2. SEM micrographs of goethite nanoparticles Figure 3. SEM image of goethite/chitosan nanocomposite surface 20
Figure 4. Cross-sectional SEM image of goethite/chitosan nanocomposite Figure 5. The proposed structure of goethite /chitosan nanocomposite made from chitosan biopolymer and goethite nanoparticles Figure 6. Percent contribution of various parameters on lead removal efficiency by goethite/chitosan nanocomposite. Figure 7. Histogram plot of errors, Mean: mean of errors, StDeV: standard deviation, N: number of experiments Figure 8. Correlation of observed and predicted lead adsorption 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 Pb(II) ions (Cn, X3) Figure 12. Mechanism of Pb (II) adsorption onto the goethite/chitosan nanocomposite, (a) complex formation mechanism and (b) ion exchange mechanism Figure 13. Comparison between removal efficiency and capacity of pure chitosan and goethite/ chitosan
films
(pH=6,
C0=
50 21
and
contact
time=24
h)
Figure. 1
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
R%
C 79
A
B
B
C
69
59
-1.000
-0.500
0.000
0.500
1.000
Deviation from Reference Point (Coded Units)
Figure. 9
30
Figure. 10
31
Figure. 11
32
Figure. 12
33
Figure. 13
Tables Table 1. Synthesis details of Goethite/chitosan nanocomposite 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
Table 5. ANOVA for Response Surface Reduced Quadratic Model
35
Table 1 Total weight (g)
10
10
15
Nanoparticle mass (g)
0.046
0.09
0.003
Chitosan mass (g)
0.2
0.2
0.3
Acid mass (g)
9.754
9.71
14.697
36
Table 2 Run pH, x1 Coded 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
-1 1 -1 1 -1 1 -1 1 0 0 0 0 0 0 0
Actual 3 6 3 6 3 6 3 6 4 4.5 4.5 4.5 4.5 4.5 4.5
Independent variables Nanoparticle initial dose, x2 concentration, x3 Coded Actual Coded Actual (g) (mg/l) -1 0.002 0 100 -1 0.002 0 100 1 0.09 0 100 1 0.09 0 100 0 0.046 -1 50 0 0.046 -1 50 0 0.046 1 150 0 0.046 1 150 -1 0.002 -1 50 1 0.09 -1 50 -1 0.002 1 150 1 0.09 1 150 0 0.046 0 100 0 0.046 0 100 0 0.046 0 100
37
Removal efficiency (%) Observed Predicted 64.94 88.18 72.61 86.79 82.43 98.43 60.58 89.72 71.98 78.48 59.84 68.43 80.97 80.40 81.01
62.87 88.04 72.74 88.86 82.35 96.42 62.59 89.80 73.60 78.95 60.42 65.76 80.80 80.80 80.80
Table 3
Source Mean vs. Total Linear vs. Mean 2FI vs. Linear Quadratic vs. 2FI Cubic vs. Quadratic Residual Total
Sum of squares 90447.09 1256.98 64.78 390.14 25.90
Sequential model sum of squares DOF Mean square F-value P-value > Comme F nt 1 90447.09 3 418.99 9.58 0.0021 3 21.59 0.41 0.7470 3 130.05 24.89 0.0020 Suggest ed 3 8.63 74.74 0.0132 Aliased
0.23 92185.13
2 15 DO
Source Linear 2FI Quadratic
Sum of squares 480.82 416.04 25.90
Cubic Pure Error
0.000 0.23
0.12 6145.68 Lack of Fit Tests Mean square F-value
F 9 6 3 0 2
Std. Dev.
R2
Source Linear 2FI Quadratic
6.61 7.21 2.29
0.7232 0.7605 0.9850
Cubic
0.34
0.9999
53.42 69.34 8.63
462.53 600.33 74.74
0.12 Model Summary Statistics R2adj Predicted R2 0.6477 0.4243 0.5809 -0.2037 0.9579 0.7613 0.9991
38
-
-
-
P-value > F 0.0022 0.0017 0.0132
Remark
-
Suggest ed Aliased -
PRESS
Remark
455.24 485.90 269.04
Suggest ed Aliased
-
Table 4 Source Model X1-pH X2-C n X3-C0 X1X2 X1X3 X2 X3 X12 X22 X32 Cor Total
Sum of squares 1711.91 852.15 57.15 347.68 20.48 43.21 1.09 100.59 229.67 38.42 1738.04
DOF
Mean square 190.21 852.15 57.15 347.68 20.48 43.21 1.09 100.59 229.67 38.42 -
9 1 1 1 1 1 1 1 1 1 14
R2 = 0.9850, R2adj =0.9579, R2pre =0.7613
39
F-value
p-value > F
36.40 163.07 10.94 66.54 3.92 8.27 0.21 19.25 43.95 7.35 -
0.0005 < 0.0001 0.0213 0.0004 0.1046 0.0348 0.6673 0.0071 0.0012 0.0422 -
Table 5
Source Model X1-pH X2-Cn X3-C0 X1X2 X1X3 X21 X22 X23 Residual Lack of Fit Pure Error Cor Total
Sum of squares 1710.82 852.15 57.15 347.68 20.48 43.21 100.59 229.67 38.42 27.22 26.98 0.23 1738.04
DOF
Mean square 213.85 852.15 57.15 347.68 20.48 43.21 100.59 229.67 38.42 4.54 6.75 0.12 -
8 1 1 1 1 1 1 1 1 6 4 2 14
R2 = 0.9843, R2adj = 0.9635, R2pre = 0.8225.
40
F-value
p-value > F
47.15 187.87 12.60 76.65 4.52 9.53 22.18 50.63 8.47 58.41 -
< 0.0001 < 0.0001 0.0121 0.0001 0.0777 0.0215 0.0033 0.0004 0.0270 0.0169 -