Journal of Engineering

Volume 2014 (2014), Article ID 278075, 10 pages

http://dx.doi.org/10.1155/2014/278075

## Modeling and Optimization for Production of Rice Husk Activated Carbon and Adsorption of Phenol

^{1}Department of Water Resources and Environmental Engineering, ABU, Zaria, Nigeria^{2}Samaru College of Agriculture, Division of Agricultural Colleges, ABU, Zaria, Nigeria^{3}Department of Textile Science and Technology, ABU, Zaria, Nigeria

Received 8 August 2014; Revised 6 December 2014; Accepted 8 December 2014; Published 24 December 2014

Academic Editor: Sreekanth Pannala

Copyright © 2014 Y. S. Mohammad et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

#### Abstract

Modeling of adsorption process establishes mathematical relationship between the interacting process variables and process optimization is important in determining the values of factors for which the response is at maximum. In this paper, response surface methodology was employed for the modeling and optimization of adsorption of phenol onto rice husk activated carbon. Among the action variables considered are activated carbon pretreatment temperature, adsorbent dosage, and initial concentration of phenol, while the response variables are removal efficiency and adsorption capacity. Regression analysis was used to analyze the models developed. The outcome of this research showed that 99.79% and 99.81% of the variations in removal efficiency and adsorption capacity, respectively, are attributed to the three process variables considered, that is, pretreatment temperature, adsorbent dosage, and initial phenol concentration. Therefore, the models can be used to predict the interaction of the process variables. Optimization tests showed that the optimum operating conditions for the adsorption process occurred at initial solute concentration of 40.61 mg/L, pretreatment temperature of 441.46°C, adsorbent dosage 4 g, adsorption capacity of 0.9595 mg/g, and removal efficiency of 97.16%. These optimum operating conditions were experimentally validated.

#### 1. Introduction

The increase in industrial, agricultural, and domestic activities has led to the discharge of large amounts of wastewater containing toxic pollutants. The growing awareness of the adverse effects of the presence of these water pollutants has led to increased strict regulation of water pollution, hence making the treatment of wastewater generated from industrial activities a high priority [1]. Phenols are among the most common water pollutants that can cause hazards including health hazard which may lead to death. Amongst hydrocarbons present in refinery wastewater, phenol is one of the main dissolved components and it is also one of the difficult hydrocarbons to degrade biologically [2, 3]. Consequently, pollution control and management have evolved many technologies for the treatment of wastewater [4, 5]. These technologies and methodologies which differ in their performance and effectiveness include coagulation, filtration, ion exchange, sedimentation, solvent extraction, adsorption, electrodialysis, chemical oxidation, disinfection, chemical precipitation, and membrane separation [6, 7]. Among the various available technologies for water pollution control, adsorption process is considered relatively better because of its convenience, ease in operation and simplicity of design [8–10]. The process can remove different types of pollutants due to the availability of wide range of adsorbents especially activated carbon which make it to have a wider applicability in water pollution control [7, 11–13]. Rice husk is an agricultural waste that is readily available as a by-product of rice processing but requires pretreatment to produce activated carbon for better performance as an adsorbent.

Response Surface Methodology (RSM) had been known as a collection of mathematical and statistical techniques for modeling and analyzing problems in which a response of interest is influenced by several variables. Basically, it had been used in multivariate experimental design, statistical modeling and process optimization [14, 15]. Thus process optimization was observed to be important in determining the values of factors for which the response is at maximum. The application of statistical experimental design techniques in adsorption process was found to result in reduced process variability combined with the requirement of less resources (time, reagents and experimental work) [16, 17]. It had been reported that RSM has several classes of designs, among which are; central composite design, box-behnken design and three-level factorial design being the most widely used and the experimental data required are dependent on the chosen design [18, 19]. There is little or no information on the optimization of carbonization temperature for the production of rice husk activated carbon for phenol adsorption using response surface methodology. In addition, the usual approach to modeling and optimization has been on either the process variable for the production of activated carbon or that of the batch adsorption process. This paper, however, highlights a novel approach involving the simultaneous modeling and optimization of process variables for the production of rice husk activated carbon and batch adsorption of phenol using response surface methodology.

#### 2. Methods

##### 2.1. Production of Activated Carbon

The natural precursor used in the preparation of adsorbent was rice husk which was collected from National Cereal Research Institute, Badeggi. When collected from NCRI Rice Mill, the rice husk was washed with distilled water to remove dirt and surface impurity, then oven-dried at 100°C for 24 h in accordance with the procedure of Kudaybergenov et al. [20]. In the thermal pretreatment, rice husk was placed on a ceramic flat surface, charged into a furnace and heated to a temperature of 300°C at a heating rate of 20–25°C/min and residence time of 1 h. The charred residue was collected and cooled at room temperature. The procedure above was repeated for 400°C and 500°C. The thermal pretreated rice husks were labeled RH300, RH400 and RH500. In the chemical pretreatment, each of the carbonized rice husk (charred residue) was activated with 1 M H_{3}PO_{4} for 3 h at impregnation ratio of 2 : 1 (volume mL of acid/mass g of rice husk) and later oven-dried overnight at 200°C to ensure proper drying [21, 22]. The material was then removed from the oven, cooled for 2 h and then washed with distilled water to bring the pH to 7.0 and again oven-dried overnight at 100°C [21].

##### 2.2. Preliminary Batch Experiment

In the preliminary Batch Adsorption experiment, 2 g of the pretreated adsorbent was added to 100 mL of standard solution of phenol with initial concentration of 10, 30, and 50 mg/L in 250 mL conical flask. The mixture in the flask was placed on magnetic stirrer at 150 rpm [23, 24] for 90 min at ambient temperature. Using a 5 mL syringe, samples were withdrawn at predetermined time interval (90 min) and centrifuged at 3000 rpm for 20 min. The supernatant solution was collected from the centrifuge by decantation and filtered using a micro filter attached to a 5 mL syringe. The procedure above was repeated for 3 and 4 g of pretreated adsorbents. The analysis of phenol in each sample filtrate was carried out using UV spectrophotometer set at wave length of 270 nm.

##### 2.3. Determination of Adsorption Capacities and Removal Efficiency

Adsorption capacity at equilibrium was determined using the equation [25, 26]: Adsorption removal efficiency was determined using the equation [27–29]: where = adsorption capacity at equilibrium (mg/g), = initial concentration of solute (mg/L), = equilibrium concentration of solute (mg/L), = volume of solution (L), = mass of activated carbon used (g), and = removal efficiency (%).

##### 2.4. Modeling and Optimization

###### 2.4.1. Experimental Design

The Central Composite Design (CCD) was applied in this work to study the interaction of variables involved in the preparation of rice husk activated carbon as well as batch adsorption process of phenol using the activated carbon prepared. The CCD is widely used for modeling and optimization and it requires only a minimum number of experiments. Generally, the CCD consists of three kinds of runs; they are factorial runs (), axial runs () and center runs () [30, 31]. This design consist of a factorial (coded to the usual ±1 notation) augmented by axial points , , , and center points [1, 32], where is the distance of the axial point from the center [33]. The center points are used to determine the experimental error and reproducibility of the data [34, 35] and the axial points are chosen such that they allow rotatability which ensures that the variance of the model prediction is constant at all points equidistant from the design center [36]. Therefore, according to Abbas [37] and Arulkumar et al., [38], the number of experimental runs required is given by the equation: where = total number of experimental runs, = number of independent variables (factors), and = number of center points.

Three variables were considered in this study, they are (i) pretreatment temperature “”, a rice husk activated carbon production variable, (ii) initial phenol concentration “”, a batch adsorption process variable, and (iii) adsorbent dosage “”, a batch adsorption process variable.

For three variables, the number of center run was six, therefore number of experimental run required was computed as: This implied that 20 experimental runs consisting of 8 factorial runs, 6 axial runs and 6 center runs were required. The two response variables considered in the study were: (i) removal efficiency () and (ii) adsorption capacity ().

###### 2.4.2. Regression Model

Each response was used to develop an empirical model that correlate the response to the three factors that is, rice husk activated carbon preparation variable () and batch adsorption process variables (, ), using second-order polynomial equation [39–41]: where is the predicted response, is the constant coefficient, is the linear coefficient, is the interaction coefficient, is the quadratic coefficient, and , are the coded values for the factors.

###### 2.4.3. Statistical and Graphical Analysis

Significance of the model equations and their terms were evaluated using statistical tools such as coefficient of determination (-squared), Fisher value (-value), probability ( value), and residual [42–44]. Graphs were employed to analyze the combined effect of factors on responses using 3D plots and to also analyze the predicted versus actual value plots of the response variables.

###### 2.4.4. Optimization

Optimization technique was employed to determine the optimum operating conditions for the process variables under consideration. To achieve this, goals were set with constraints. For each of the factors, goal was set “in range” with constraints 300–500, 10–50, and 2–4 of lower-upper level for factors , , and , respectively. For the response surface, the goal for was also set “in range” with constraint 0.18–2.393 as lower-upper levels, while the goal for was set “maximize” at 100. Therefore removal efficiency becomes the objective function or performance index.

###### 2.4.5. Model Validation

Model validation was carried out by conducting batch experiment under optimum operating conditions. In order to evaluate the validity of the model, experimental values obtained were compared with the model predicted values.

###### 2.4.6. Software Application

Design Expert Software (version 8) was used for the design of the experiment, regression, statistical analysis, optimization, and graphical analysis.

#### 3. Results and Discussion

##### 3.1. Responses Obtained from the Experiment

Table 1 shows the design matrix consisting of types of run, coded and actual factors as randomized by the software, and respective response obtained from the experiment. From the results, it could be observed that the highest removal efficiency () of 95.72% was obtained and this was followed by the removal efficiency of 94.17%. In terms of adsorption capacity (), the highest value of 2.39 mg/g was obtained and this was followed by adsorption capacity of 2.21 mg/g. In comparison, recent studies by Kalderis et al., [13], Kermani et al., [45], Mahvi et al., [46], and Daffalla et al., [47] on adsorption of phenol onto rice husk activated carbon shows that adsorption capacities of 27.58 mg/g, 0.886 mg/g, 0.95–1 mg/g, and 0.98–46.19 mg/g were obtained, respectively. The differences in adsorption capacities could be as a result of the influence of the processes employed in the production of the activated carbon which was reported to have significant influence on its performance in adsorption process [48].