Ind. Eng. Chem. Res. 2007, 46, 5697-5706
5697
Multicomponent Adsorption Study of Metal Ions onto Bagasse Fly Ash Using Taguchi’s Design of Experimental Methodology Vimal C. Srivastava,* Indra D. Mall, and Indra M. Mishra Department of Chemical Engineering, Indian Institute of Technology-Roorkee, Roorkee 247667, India
This paper utilizes the Taguchi optimization methodology (L27 orthogonal array) to optimize various parameters for the simultaneous removal of Cd, Ni, and Zn metal ions from aqueous solutions using bagasse fly ash (BFA) as an adsorbent. The effect of such parameters as the initial metal concentrations (C0,i), temperature, initial pH, adsorbent dosage (m), and contact time on the adsorption of the aforementioned metal ions has been studied at three levels to determine their effect on the selected response characteristic (the total amount of metal adsorbed on BFA, in terms of mg/g of BFA (qtot)). The Pareto analysis of variance shows that m is the most significant parameter, with a 53.14% and 31.25% contribution to the qtot and signal-to-noise (S/N) ratio data. The contribution of interactions between the C0,i values is also significant. Confirmation experiments have been performed to prove the effectiveness of the Taguchi technique after the optimum levels of process parameters are determined. 1. Introduction Heavy metals are toxic and resistant to biodegradation. Because of their mobility in natural water ecosystems and adsorption onto soils, sediments, thick sludge, etc., they accumulate in aquatic life forms such as fish, vegetation, and horticultural products via the food chain. When consumed, these metals accumulate in tissues, fats, and other organs of humans, causing malfunctions and damage to vital functions of the body’s system. In the past four decades, the uncontrolled discharge of heavy metals from plating/rinsing industries and various other manufacturing process industries has resulted in serious contamination of numerous sites. The heavy-metal ions are often encountered at elevated levels, and their exposure is likely to persist for a prolonged time. The Central Pollution Control Board, Ministry of Environment and Forests, Government of India (CPCB), has set minimal national standards of 1.0, 3.0, and 5.0 mg/L, respectively, for Cd(II), Ni(II), and Zn(II) for the safe discharge of the industrial effluents that contain these metal ions into surface waters.1 Several methods (e.g., physicochemical, biological, and thermal processes) have been developed for the removal of heavy metals from waters and wastewaters to decrease their impact on the environment. Physicochemical processes apply to all waste types, whereas biological methods are appropriate for dilute wastewaters that contain metals and thermal techniques are applicable to organometallics.2 The physicochemical treatment processes mainly include adsorption, precipitation, and ion exchange. Adsorption technologies for metals wastes include activated carbon and ion exchange treatment. For high-strength and low-volume wastewaters, the removal of heavy metals via adsorption, using granular/powdered activated carbon, has been widely used. Most of the activated carbons available commercially are microporous and of high surface area, and exhibit high efficiency for the adsorption of gaseous molecules. However, the adsorption efficiency of bigger and high-molecular-weight molecules on microporous activated carbon is very low. Also, adsorbent-grade activated carbon is cost-prohibitive * To whom correspondence should be addressed. Tel.: +91-1332285889. Fax: +91-1332-276535, 273560. E-mail address: vimalcsr@ yahoo.co.in.
and both regeneration and disposal of the used carbon is often very difficult. Therefore, the search for unconventional and lessexpensive adsorbents such as bagasse fly ash (BFA), rice husk ash, silica, peat, lignite, bagasse pith, wood, saw dust, etc. for the removal of various pollutants from industrial effluents has attracted the attention of several investigators. BFA, which is a waste that is collected from the particulate collection equipment attached upstream to the stacks of bagassefired boilers, causes disposal problems. It is mainly used for land filling, and it is used in part as a filler in building materials and paper and wood boards. BFA has good adsorptive properties and has been used for the removal of chemical oxygen demand (COD) and color from paper mill effluents.3 Various researchers have utilized it for the adsorptive removal of phenolic compounds,4 pyridine,5 dyes,6-8 and metals.9-12 This waste material is a potent low-cost adsorbent for the removal of heavy-metal ions from industrial aqueous effluents. Much of the work on the adsorption of heavy-metal ions by various types of adsorbents has focused on the uptake of single metals. Because of the fact that industrial effluents generally contain several metals, it is necessary to study the simultaneous sorption of two or more metals and also to quantify the interference of one metal with the sorption of the other. No information is available in the literature for the simultaneous removal of Cd(II), Ni(II), and Zn(II) ions by BFA. However, it has also been reported that the adsorption of metal ions from aqueous solution by any adsorbent is drastically affected by several factors, such as the initial concentration of metal ion (C0), temperature (T), initial pH (pH0), adsorbent dosage (m), and contact time (t). Generally, “one-factor-at-a-time” experiments have been conducted in most of the previous studies to determine the operating conditions of optimum metal removal. One-factor-at-a-time designs often overlook interactive effects of the factors on the sorption process. Fractional factorial design based on Taguchi’s orthogonal array (OA) can be a very effective methodology to investigate the effects of multiple factors, as well as potential interactions between these factors, in a time- and cost-effective manner.13 To date, no work is available in the literature on the optimization of process parameters based on Taguchi’s OA experimental design for the simultaneous removal of metal ions. The objective of this paper is to apply Taguchi’s fractional factorial experimental design
10.1021/ie0609822 CCC: $37.00 © 2007 American Chemical Society Published on Web 07/14/2007
5698
Ind. Eng. Chem. Res., Vol. 46, No. 17, 2007
to screen significant factors, which would have a great impact on the multicomponent adsorption efficiency of metal ions from aqueous solution using BFA as an adsorbent. The objective of the present study is to maximize the selected response characteristic (the total amount of metal ions adsorbed onto BFA in terms of milligrams per gram of BFA (qtot)) by optimizing the various parameters that affect the simultaneous removal of Cd(II), Ni(II), and Zn(II) metal ions from aqueous solutions by BFA. The effect of individual process parameters and their interactions on qtot will be examined using the standard procedure suggested by Taguchi. The mean or the average values and S/N ratio of the quality/response characteristics for each parameter at different levels have been calculated from experimental data. For the graphical representation of the change in value of quality characteristic, and that of S/N ratio with the variation in process parameters, the response curves will be plotted. These response curves are used to examine the parametric effects on the response characteristics. The analysis of variance (ANOVA) will be performed for the raw and S/N data to identify the significant parameters and to quantify their effect on the response characteristics. The mostfavorable conditions (optimal settings) of process parameters, in terms of the mean response of characteristics, would be established by analyzing the response curves and the ANOVA tables. 2. Materials and Methods 2.1. Adsorbent and Adsorbates. BFA was obtained from a nearby sugar mill (Deoband Sugar Mill, U.P., India) and used as an adsorbent without any pretreatment (except sieving). Detailed physicochemical characteristics of the BFA have already been presented elsewhere.4,9 All the chemicals used in the study were analytical reagent grade. Nickel chloride hexahydrate (NiCl2·6H2O) was procured from Qualigens Fine Chemicals, Mumbai, India. Cadmium sulfate octahydrate (3CdSO4·8H2O), zinc sulfate heptahydrate (ZnSO4·7H2O), NaOH, HCl, HNO3, H2SO4, and CH3COOH were obtained from s.d. fine chemicals, Mumbai. Stock solutions of Cd(II), Ni(II), and Zn(II) metal ions (1 g/L strength) were prepared by dissolving exact amounts of 3CdSO4·8H2O, NiCl2· 6H2O, and ZnSO4·7H2O separately in double-distilled water. The stock solution for each metal salt was diluted to give metal-ion concentrations in the range of 0-100 mg/L for use in the experiments. 2.2. Analysis of Metal Ions. The concentration of Cd(II), Ni(II), and Zn(II) in the aqueous samples was determined using a flame atomic absorption spectrophotometer (GBC Avanta, Australia) with the detection limit of 0.009, 0.040, and 0.008 mg/L at wavelengths of 228.8, 232, and 213.9 nm, for Cd(II), Ni(II), and Zn(II), respectively, using an air-acetylene flame. Before the analysis, the sample was diluted, if necessary, with distilled water to a concentration in the range of 0.2-1.8 mg/L for Cd(II), 1.8-8 mg/L for Ni(II) and 0.4-1.5 mg/L for Zn(II). Metal-ion concentrations were determined with reference to the appropriate standard metal-ion solutions. 2.3. Taguchi’s Methodology of Experimental Design. Taguchi’s methodology has been used extensively in conducting experiments and devising a strategy for the quality control of products in the manufacturing industries. Taguchi’s method to improve the quality of the products is heavily dependent on designing and testing a system based on engineer’s judgment of selected materials and parts, and nominal product/process parameters based on current technology.14 Although system
Table 1. Process Parameters for Multicomponent Adsorption Study of Metal Ions onto BFA Using Taguchi’s Orthogonal Arrays (OAs) parameter A B C D E F G
initial concentration of cadmium, C0,Cd (mg/L) initial concentration of nickel, C0,Ni (mg/L) initial concentration of zinc, C0,Zn (mg/L) temperature, T (°C) initial pH of solution, pH0 BFA dose, m (g/L) contact time, t (min)
level 0
level 1
level 2
0
50
100
0
50
100
0
50
100
20 4 5 30
30 6 10 60
40 8 15 90
design helps to identify the working levels of the design parameters, process parameter design seeks to determine the parameter levels that produce the best performance of the product/process under study. The optimum condition is selected so that the influence of uncontrollable factors (noise factors) causes minimum variation to the system performance. The OAs, variance, and S/N analysis are the essential tools of parameter design. Taguchi’s method of experimental design provides the optimal selection of parametric values, based on their intraparametric interaction, to accomplish a process. This method minimizes the number of experiments to be conducted based on the statistical significance of parameters and the interactive influences of these parameters. Taguchi’s methodology, as adopted in this study, consists of four phases (with various steps), viz., planning, conducting, analysis, and validation. Taguchi’s method of design of experiments (DOE) involves the establishment of a large number of experimental situations, described as OAs, to reduce experimental errors and to enhance their efficiency and the reproducibility of the laboratory experiments. Each phase has separate objectives that are interconnected sequentially to achieve the overall optimization process. 2.3.1. Design of Experiment (Phase 1). The first step in Phase 1 is to identify the various factors to be optimized in batch experiments that have critical effect on the simultaneous adsorptive removal of Cd(II), Ni(II), and Zn(II) metal ions from aqueous solutions adsorption onto BFA. Factors were selected and the ranges were further assigned based on the detailed experiments for metal removal using BFA.9-12 Based on the experience, seven process parameters that exerted significant influence on the metal adsorption have been selected for the present experimental design. These process parameters, as well as their designations and levels, are given in Table 1. The initial concentration of one metal ion (C0,i) significantly affects the adsorption of other metal ions in the simultaneous adsorption of metal ions;9,10 therefore, it has been decided to study three two-parameter interactions between the initial concentrations of metal ions, i.e., C0,Cd × C0,Ni, C0,Cd × C0,Zn, and C0,Ni × C0,Zn. 2.3.1.1. Selection of Orthogonal Array (OA) and Parameter Assignment. The next step in phase 1 was to design the matrix experiment and define the data analysis procedure. The appropriate OAs for the control parameters to fit a specific study were selected. Taguchi provides many standard OAs and corresponding linear graphs for this purpose. The OA selected must satisfy the following inequality:
total DOF of the OA g total DOF required for the experiment
Ind. Eng. Chem. Res., Vol. 46, No. 17, 2007 5699
where DOF denotes the degrees of freedom. It was decided to study each selected process parameter at three levels to account for nonlinear behavior (if any) of the parameters of a process.15 With seven parameters each at three levels and three secondorder interactions, the total DOF required is 26 [ ) 7 × (3 1) + (3 × 4)], because a three-level parameter has DOF ) 2 (number of levels ) 1) and each two-parameter interaction term has DOF ) 4 (2 × 2). Hence, an L27 (313) OA (a standard threelevel OA) has been selected for this phase of work. The L27 OA, with the assignment of parameters and interactions, is shown in Table 2. The parameters and interactions have been assigned to specific columns of the OA using the triangular table.16 2.3.2. Batch Experimental Adsorption Studies (Phase 2). Batch adsorption experiments were performed for simultaneous metal removal with BFA, using the selected 27 experimental trials, in combination with seven process factors at three levels (Table 1), and the result obtained from each set, in terms of the total amount of metal ion adsorbed onto BFA (in mg/g of BFA) (qtot) is shown in Table 2. The results presented in the table represent those for three individual determinations. For each experimental run, 150 mL of aqueous solution, having 50 mL of each metal ion solution (viz. Cd(II), Ni(II), and Zn(II)) of known concentration was taken in a 500-mL stoppered conical flask that contained a specific amount of BFA. These flasks were agitated at a constant shaking rate of 150 rpm in a temperature-controlled orbital shaker (Remi Instruments, Mumbai, India) maintained at 20, 30, or 40 °C. The initial pH (pH0) of the adsorbate solution was adjusted using 1 N (36.5 g/L) HCl or 1 N (40 g/L) NaOH aqueous solution without any further pH adjustment during the sorption process. The samples withdrawn after appropriate contact time were centrifuged using Research Centrifuge (Remi Instruments, Mumbai, India) at 5000 rpm for 5 min, and then the supernatant liquid was analyzed for the residual concentration of metal ions. The removal of metal ions from the solution and the equilibrium adsorption uptake in the solid phase, qtot (in units of mg/g), were calculated using the following relationship: 3
qtot )
∑ i)1
C0,i - Ce,i m
(1)
where C0,i is the initial metal ion concentration (given in units of mg/L), Ce,i the equilibrium metal ion concentration (also given in mg/L), and m the adsorbent dose (given in units of g/L). 2.3.3. Analysis of Experimental Data and Prediction of Performance (Phase 3). The obtained experimental data was processed with “higher-is-better” (HB) quality characteristics (i) to determine the optimum conditions for the adsorption, (ii) to identify the influence of individual factors on adsorption, and (iii) to estimate the performance (qtot) under the optimum conditions. Taguchi defines the loss function L(y) as a quantity proportional to the deviation from the nominal quality characteristic, and he found the following quadratic form to be a practical workable function, viz.,
L(y) ) k(y - mT)2
(2)
where k denotes the proportionality constant, mT is the target value, and y is the experimental value obtained for each trial. In the case of HB-type quality characteristics, the loss function
can be written as L(y) ) k(1/y)2 and the expected loss function can be represented by
() 1 y2
E[L(y)] ) kE
(3)
where E(1/y2) can be estimated from a sample of R as R
E(1/y2) )
[1/yi2] ∑ i)1 (4)
R
2.3.3.1. Signal-to-Noise (S/N) Ratio. Taguchi created a transform for the loss function that is called the signal-to-noise (S/N) ratio, which looks at two characteristics of a distribution and combines these characteristics into a single number or figure of merit. The S/N ratio combines the mean level of the quality characteristic and the variance around this mean into a single metric.17 A high S/N ratio implies that the signal is much higher than the random effects of noise factors. Process operation consistent with the highest S/N ratio always yields optimum quality with minimum variation. The S/N ratio consolidates several repetitions (at least two data points are required) into one value. The equations for calculating S/N ratios for HB-type characteristics are given as follows:16
(S/N)HB ) -10 log
( ) 1
R
∑ R i)1
1
yi2
(5)
where yi is the value of the characteristic in an observation i and R is the number of observation or number of repetitions in a trial. From among the methods suggested by Taguchi for analyzing the data, the following methods have been used in the present work: (i) plot of average response curves; (ii) ANOVA for raw data; and (iii) ANOVA for S/N data. The plot of the average response at each level of a parameter indicates the trend. It is a pictorial representation of the effect of a parameter on the response. The change in the response characteristic with the change in levels of a parameter can easily be visualized from these curves. Typically, ANOVA for OAs is conducted in the same manner as other structured experiments.16 The S/N ratio is treated as a response of the experiment, which is a measure of the variation within a trial when noise factors are present. A standard ANOVA can be conducted on the S/N ratio which will identify the significant parameters (mean and variation). 2.3.3.2. Prediction of the Mean. After determination of the optimum condition, the mean of the response (µ) at the optimum condition is predicted. This mean is estimated only from the significant parameters. The ANOVA identifies the significant parameters. Suppose, parameters A and B are significant and A2, B2 (second level of A ) A2, second level of B ) B2) are the optimal treatment conditions. The mean under the optimal condition (optimal value of the response characteristic) then is estimated as
h ) + (B h2 - T h) ) A h2 + B h2 - T h µ)T h + (A h2 - T
(6)
where T h is the overall mean of the response, and A h 2 and B h2 represent average values of response at the second levels of parameters A and B, respectively.
total mean
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2
0 0 0 1 1 1 2 2 2 0 0 0 1 1 1 2 2 2 0 0 0 1 1 1 2 2 2
0 0 0 1 1 1 2 2 2 1 1 1 2 2 2 0 0 0 2 2 2 0 0 0 1 1 1
0 0 0 1 1 1 2 2 2 2 2 2 0 0 0 1 1 1 1 1 1 2 2 2 0 0 0
0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2
0 1 2 0 1 2 0 1 2 1 2 0 1 2 0 1 2 0 2 0 1 2 0 1 2 0 1
0 1 2 0 1 2 0 1 2 2 0 1 2 0 1 2 0 1 1 2 0 1 2 0 1 2 0
0 1 2 1 2 0 2 0 1 0 1 2 1 2 0 2 0 1 0 1 2 1 2 0 2 0 1
0 1 2 1 2 0 2 0 1 1 2 0 2 0 1 0 1 2 2 0 1 0 1 2 1 2 0
0 1 2 1 2 0 2 0 1 2 0 1 0 1 2 1 2 0 1 2 0 2 0 1 0 1 2
0 1 2 2 0 1 1 2 0 0 1 2 2 0 1 1 2 0 0 1 2 2 0 1 1 2 0
0 1 2 2 0 1 1 2 0 1 2 0 0 1 2 2 0 1 2 0 1 1 2 0 0 1 2
0 1 2 2 0 1 1 2 0 2 0 1 1 2 0 0 1 2 1 2 0 0 1 2 2 0 1
expt 1, A expt 2, B expt 3, A × B expt 4, A × B expt 5, C expt 6, A × C expt 7, A × C expt 8, B × C expt 9, D expt 10, E expt 11, B × C expt 12, F expt 13, G
0.00 4.00 4.21 2.36 8.62 7.29 6.17 3.83 13.20 3.32 3.22 12.93 9.80 4.06 5.39 3.70 10.30 6.50 3.60 12.74 7.66 4.10 4.05 14.02 11.94 7.36 6.06
0.00 3.86 4.31 2.47 9.25 6.86 5.81 4.21 13.38 3.24 3.25 13.36 10.20 4.17 5.13 3.64 11.83 7.15 3.48 11.83 7.84 4.26 4.29 16.08 11.42 7.51 5.36 182.74 180.45 184.18 6.76
0.00 3.90 4.25 2.48 8.84 6.99 5.78 4.23 13.41 3.29 3.11 13.08 10.00 4.31 5.25 3.52 11.88 7.30 3.55 12.08 7.20 4.31 4.16 16.00 11.08 7.25 5.50
Experimental qtot Values R1 R2 R3
S/N ratio (dB) 0.00 11.86 12.58 7.73 18.98 16.95 15.44 12.20 22.50 10.33 10.09 22.36 20.00 12.41 14.41 11.17 21.03 16.85 10.98 21.72 17.56 12.51 12.39 23.68 21.19 17.35 14.99
Table 2. Column Assignment for the Various Factors and Three Interactions in the Taguchi’s L27 (313) Orthogonal Array (OA) and Experimental qtot Values for Multicomponent Metal Ions Adsorption onto BFA
5700 Ind. Eng. Chem. Res., Vol. 46, No. 17, 2007
Ind. Eng. Chem. Res., Vol. 46, No. 17, 2007 5701 Table 3. Average and Main Effects of qtot Values for BFA: Raw and S/N Data Raw Data, Average Value A B C D E F G A×B A×C B×C
Main Effects (Raw Data)
S/N Data, Average Value
Main Effects (S/N Data)
L1
L2
L3
L2 - L1
L3 - L2
L1
L2
L3
L2 - L1
L3 - L2
5.55 5.68 4.95 6.01 6.06 10.64 5.94 5.89 6.28 5.96
6.78 6.84 6.60 6.98 7.43 5.61 7.19 7.01 6.88 7.22
7.95 7.75 8.73 7.28 6.78 4.02 7.14 7.38 7.10 7.09
1.23 1.16 1.65 0.96 1.37 -5.03 1.25 1.12 0.60 1.26
1.18 0.91 2.13 0.31 -0.65 -1.59 -0.05 0.37 0.22 -0.13
13.14 13.05 12.15 13.81 14.14 19.05 13.84 13.71 13.95 13.69
15.40 15.45 15.34 15.44 15.56 14.58 16.22 15.72 15.67 15.82
16.93 16.97 17.99 16.22 15.78 11.84 15.41 16.04 15.85 15.97
2.27 2.39 3.19 1.63 1.42 -4.47 2.38 2.01 1.72 2.13
1.53 1.52 2.65 0.77 0.22 -2.75 -0.81 0.32 0.17 0.15
2.3.3.3. Determination of Confidence Interval. The estimation of µ is only a point estimate based on the average of results obtained from the experiment. Statistically, this provides a 50% chance of the true average being greater than µ and a 50% chance of the true average being less than µ. Therefore, it is customary to represent the values of a statistical parameter as a range within which it is likely to fall for a given level of confidence. This range is termed as the confidence interval (CI). In other words, the confidence interval is a maximum and minimum value between which the true average should fall at some stated percentage of confidence. The following two types of confidence intervals are suggested by Taguchi, in regard to the estimated mean of the optimal treatment condition:16 (i) Around the estimated average of a treatment condition predicted from the experiment. This type of confidence interval is designated as CIPOP (confidence interval for the population). (ii) Around the estimated average of a treatment condition used in a confirmation experiment to verify predictions. This type of confidence interval is designated as CICE (confidence interval for a sample group). The difference between CIPOP and CICE is that CIPOP is associated with the entire population, i.e., all parts ever made under the specified conditions, and CICE is associated with only a sample group made under the specified conditions. Because of the smaller size (in confirmation experiments), relative to the entire population, CICE must be slightly wider. The expressions for computing the confidence interval are given as follows:18
CIPOP ) CICE )
x
x
FR(1,fe)Ve neff
FR(1,fe)Ve
[
(7)
]
1 1 + neff R
(8)
where FR(1, fe) represents the F-ratio at a confidence level of (1 - R) against DOF ) 1 and a DOF error of fe, Ve is the error variance (from ANOVA), and neff is defined as
neff ) N (9) 1 + [total DOF associated in the estimate of the mean] where N is the total number of results. R represents the sample size for the confirmation experiment. Equation 8 shows that, as R approaches infinity (i.e., the entire population), the value 1/R approaches zero and CICE ) CIPOP. As R approaches 1, CICE becomes wider.
2.3.4. Confirmation Experiment (Phase 4). The confirmation experiment is the final step in verifying the conclusions drawn from the previous round of experimentation. The optimum conditions are set for the significant parameters (the insignificant parameters are set at economic levels), and a selected number of tests are conducted under constant specified conditions. The average of the results of the confirmation experiments is compared with the anticipated average, based on the parameters and levels tested. The confirmation experiment is a crucial step and is highly recommended to verify the experimental conclusions. 3. Results and Discussion Experiments were conducted for Cd(II), Ni(II), and Zn(II) ion adsorption onto BFA, according to the test conditions specified by L27 OA (see Table 2). Each experiment was repeated three times for each trial condition. The average or mean values of qtot and S/N ratio for each parameter at levels 1, 2, and 3 are calculated from Table 2. It is observed that metal adsorption is strongly dependent on the parametric conditions. 3.1. Effect of Process Parameters. The raw data for the average value of qtot and S/N ratio for each parameter at levels 1, 2, and 3, along with interactions at the assigned levels, are given in Table 3 for metal adsorption onto BFA. Various metal removal parameters (C0,i, T, pH0, m, and t) significantly affect the qtot values. The interaction effect of concentration of one metal ion with respect to (wrt) another metal ion also has significant influence on the qtot values. Individually, relative to the level stage, with qtot as the desired response characteristic, m (parameter F) has the highest influence at level 1, pH0 (parameter E) has the highest influence at level 2, and C0,Zn (parameter C) has the highest influence at level 3. The difference between level 2 and level 1 (L2 - L1) of each factor indicates the relative influence of the effect. The larger the difference, the stronger the influence. Table 3 shows that no single parameter has an overriding or predominant influence over others for the removal of Cd(II), Ni(II), and Zn(II) from aqueous solution by BFA. C0,i shows a stronger influence on qtot than that of other parameters. The qtot value increased as C0,i increased, because the resistance to the metal uptake decreased as the mass-transfer driving force increased. The response curves for the individual effects of metal adsorption parameters on the average value of qtot and respective S/N ratio for metal adsorption onto BFA are given in Figure 1. An increase in the levels of factors such as C0,i, and T, from 1 to 2 and from 2 to 3, has resulted in an increase in the qtot values. Because sorption is an exothermic process, it would be expected that an increase in T would result in a decrease in the qtot value. However, if the diffusion process controls the adsorption process, the qtot value will increase as T increases,
5702
Ind. Eng. Chem. Res., Vol. 46, No. 17, 2007
Figure 1. Effect of process parameters on qtot and S/N ratio for multicomponent adsorption of metal ions onto BFA.
because of the endothermicity of the diffusion process. An increase in T results in increased mobility of the metal ions and a decrease in the retarding forces acting on the diffusing ions. These result in the enhancement in the sorptive capacity of the adsorbents. However, the diffusion of the metal ions into the pores of the adsorbent is not the only rate-controlling step,12 and the diffusion resistance can be ignored with adequate contact time. Therefore, the increase in sorptive uptake of the metal ions with an increase in temperature may be attributed to chemisorption. An increase in the level of contact time t, from 1 to 2, results in an increase in the qtot value. However, the qtot value remains constant with a subsequent increase from level 2 to level 3. Also, the adsorption of metal ions increases as t increases until equilibrium is attained between the solute molecules in the liquid and the solid phases. During the initial stage of sorption, a large number of vacant surface sites are available for adsorption. As the sorption process progresses, its intention is to occupy the vacant sites, because of the repulsive forces between the solute molecules onto the solid surface and the bulk liquid phase. Besides, the metal ions are adsorbed into the mesopores that get almost saturated with metal ions during the initial stage of adsorption. Thereafter, the metal ions must traverse farther and deeper into the pores, encountering much-greater resistance. This results in the slowing of the adsorption during the later period of adsorption.19 An increase in pH0 has resulted in higher adsorption up to level 2 and subsequent increase resulted in the decrease in the desired characteristic (qtot). The removal of metal ions is determined to increase as pH0 increases from 4 to 6. The maximum uptake of metal ions was obtained at pH0 ∼6, and the qtot value decreased at pH0 >6. The oxides of aluminum, calcium and silicon present in the BFA develop charges when in contact with water. Except silica, all other oxides possess a positive charge for the pH range of interest, because the zeropoint charge (pHZPC) of SiO2, Fe2O3, Al2O3, and CaO are 2.2, 6.7, 8.5, and 11.0, respectively.20 A positive charge develops on the surface of the oxides of BFA in an acidic medium, because of the aqua complex formation of the oxides present, as follows: H+
-MO + H-OH 98 M-OH2+ + OH-
(10)
Metal-ion adsorption at low pH0 ( pH0 e6) is less than that at higher pH0 (∼6). This is due to the fact that the surface charge that is developed at low pH0 is not suitable for the adsorption of these metal ions. For pH0
