Methylene blue removal using prepared activated carbon from grape wood wastes: adsorption process analysis and modeling


 An adsorption study has been conducted for activated carbon obtained from grape wood wastes to assess their capability to remove methylene blue (MB) from the aqueous solutions. The properties of prepared activated carbon were characterized using FTIR, BET and SEM analyses. The effects of independent variables such as initial concentration of MB (100–500 mg L−1), initial pH of solution (3–11), adsorbent dosage (0.25–12.25 g L−1) and contact time (10–90 min) on the MB adsorption have been optimized using response surface methodology. The highest MB removal efficiency was 98% when pH, MB and adsorbent dosage were 11, 100 mg L−1 and 12.25 g L−1, respectively. The experimental data have been tested using Langmuir and Freundlich isotherm models, and the achieved data were fully fitted with the Langmuir model (R2 = 0.99), which indicates the monolayer adsorption. The adsorption kinetics well followed by the pseudo-second-order model with R2 of 0.99. This prepared activated carbon as a low-cost and eco-friendly adsorbent can be used widely for water and wastewater treatment.

GRAPHICAL ABSTRACT ABBREVIATIONS Code Descriptions RSM Response surface methodology CCD Central composite design C 0 Initial concentration of dye solution (mg L À1 ) q e Equilibrium adsorption capacity (mg g) C e Dye concentration (mg L À1 ) at equilibrium V Volume of solution (L) W Weight of adsorbent (g) q Maximum adsorption capacity reflected a complete monolayer (mg g À1 ) in the Langmuir isotherm model K 2 Rate constant of pseudo-second-order adsorption (g mg À1 min À1 ) K F Isotherm constant that indicates the capacity parameter (mg g À1 ) related to the intensity of the adsorption R 2 Correlation coefficient %R Percentage of adsorption process efficiency b Langmuir constant or adsorption equilibrium constant (Lmg À1 ) that is related to the apparent energy of sorption Y Predicted response

INTRODUCTION
Nowadays, activated carbon (AC) due to its well-developed porous structure, massive surface area, fast removal speed and excellent adsorption ability is one of the most well-known and widely used adsorbents in water and wastewater treatment (Kosheleva et al. 2019;Li et al. 2020). AC has several advantages over other adsorbents, such as high efficiency to remove odor and taste, simple process design, high selectivity at the molecular level, low energy consumption, reusability, high adsorption potential, and resistance in corrosive and toxic environments (Seidmohammadi et al. 2015;Ao et al. 2018;). One of the major disadvantages of AC is its high cost of production, which limits its application extensively (Selvanathan et al. 2015). In general, any abundantly available, cheap and safe organic matter such as agricultural wastes can be considered as raw materials (precursor) for AC production (Hameed et al. 2017;Kosheleva et al. 2019). Estimates show that significant amounts of agricultural wastes are generated annually at the time of harvesting and processing of various agricultural products around the world. Recent studies have shown that many of these by-products, due to their high purity and very good properties, can be used as precursors to produce high-quality AC to adsorb gases and solutes from aqueous solutions (Suzuki et al. 2007). The main components of woody plants and agricultural residues consist of three fractions: cellulose (40-50%), hemicellulose (20-30%) and lignin (10-25%), which is a suitable structure, for the production of AC (Pérez et al. 2002). In recent years, researchers have used various agricultural wastes such as peach stones (Torrellas et al. 2015), date stones (Danish et al. 2014), plum kernels (Tseng 2007), peanut shells (Wu et al. 2013), wasted orange peels (hydrochars) , teak sawdust-hydrochars (Duy Nguyen et al. 2019), rice husks (Lin et al. 2013) and corncob (Wu et al. 2001) as precursors of AC. In this study, the grapevine wastes were used to produce AC. Grapes are one of the most commonly consumed fruits in the world. Approximately 7.5 ha millions of lands around the world in 2017 are under grapevine cultivation (Sun et al. 2020). On the other hand, grape wood is mainly composed of cellulose, tannin, hemicelluloses and lignin, which has a good potential for AC production (Prozil et al. 2012). A significant amount of agricultural wastes can be produced during pruning the branches of vines to produce higher quality grapes and using grape wood to make adsorbent is an efficient management strategy to control agricultural wastes. The carbon structure of AC is composed of different main functional groups such as phenol, carboxyl, lactone, carbonyl and quinone. In general, the main functional groups in the carbon structure of AC are responsible for the adsorption of pollutants (Bhatnagar et al. 2013). Also, available oxygen, hydrogen, sulfur and nitrogen can participate in the structure of AC in the form of functional groups or chemical atoms (Heidarinejad et al. 2020). The existence of functional groups as chemical properties of adsorbents depends on the activation process using different types of activating agents (Oginni et al. 2019). The activation process can be performed by either dry oxidation or wet oxidation. The dry oxidation involves the reaction of carbon with hot oxidizing gases such as steam and CO 2 or a mixture of gaseous and steam at temperatures of above 700°C (Yahya et al. 2015). But in the wet oxidation, the process is carried out through two steps: chemical activation using chemicals such as phosphoric acid (H 3 PO 4 ), zinc chloride (ZnCl 2 ), potassium hydroxide (KOH), hydrogen peroxide (H 2 O 2 ), ammonium persulphate ((NH 4 ) 2 S 2 O 8 ), nitric acid (HNO 3 ) and potassium permanganate (KMnO 4 ), and then thermal activation with increasing temperature in the range of 400-900°C (Al-Qodah & Shawabkah 2009). Among the activating agents, sulfuric acid (H 2 SO 4 ) is a highly reactive chemical activator, which is able to dissolve many organic compounds (such as carbohydrates and other organic materials), minerals and impurities from the AC precursors (Al-Qodah & Shawabkah 2009). The use of sulfuric acid to activate carbon, in addition to low cost, causes medium and large porosity in the surface of AC (Olivares-Marín et al. 2012). Therefore, the activation with sulphuric acid was the main aim of the current study to generate high-quality AC from grape wastes.
Methylene blue (MB) is a cationic dye with many applications in the chemical, biology, medical and dyeing industries (Pathania et al. 2017). The cationic dyes are more toxic compared with anionic dyes (Etim et al. 2016). The presence of MB in the water sources causes to damage the eyes of humans and animals, burning nausea, vomiting, excessive sweating, mental disorder and methemoglobinemia (Sartape et al. 2015). Therefore, the treatment of effluent containing MB is necessary due to its harmful effects on human health and environment.
This study reports the synthesis of AC from grape wood and its application in the removal of dye from aqueous solutions. The central composite design (CCD) based on response surface methodology was used to evaluate and optimize the effects of independent variables namely initial pH, MB concentration, AC dosage and contact time on the adsorption process. The experimental data have been used to develop a model using the mathematical-statistical technique. Furthermore, the mechanism and thermodynamics of MB adsorption will be also investigated using isotherm, kinetic models and parameters. The physical and chemical properties of the prepared adsorbent will be characterized using scanning electron microscope (SEM), Brunauer, Emmett and Teller (BET) and Fourier transform infrared (FTIR) spectroscopy.

Feedstock and chemical
The cationic dye, MB (Figure 1, chemical formula ¼ C 16 H 18 ClN 3 S, molecular weight ¼ 319.85 g mol À1 , max ¼ 665 nm), was supplied by Merck Company, Germany (Kuang et al. 2020). The MB, sodium hydroxide, hydrochloric acid and sulfuric acid were all analytical grades and were used without further purification. A stock solution of MB was prepared in deionized water, and the required working dye solutions were obtained from the stock solution. The chemical activation of adsorbent was carried out by using 98% sulfuric acid. The pH of solution was adjusted using NaOH (0.1 M) and HCl (0.1 M), which were purchased from Merck company.

Synthesis of AC
Grape wood wastes were collected from the Kermanshah rural area in the western region of Iran and transferred to the laboratory for further steps. The raw materials were mechanically cut into pieces with an average size of 2-3 cm. Grape wood wastes were immersed in the distilled water for about 2 h and then washed to remove any impurities such as toxins and dust. Then all washed raw materials were dried at 100°C for 4 h. The sulfuric acid (with 98% purity) was used for the chemical activation during 8 h with the weight ratio of 1:10. The chemical activated wood was dried at oven model Memmert 854-Germany for about 4 h at 100°C. The thermal activation of chemical activated wood was carried out using oven at 750°C for 1 h. After cooling, the AC was washed with deionized water to adjust the pH value at the level of ∼7. Finally, the washed AC was dried in a vacuum oven at 105°C overnight.

AC characterization
Microscopic and chemical properties of prepared AC were studied by SEM, BET and FTIR methods. Microscopic images of adsorbent were obtained by SEM (Jeol JSM 840A, Japan). The BET surface area and pore structure characteristics of the carbons were determined by N 2 adsorption/desorption at 77 K using a surface area analyzer (Quantachrome Corporation, USA). FTIR spectroscopy has been prepared using Shimadzu IRPrestige, Japan model to determine the vibrational frequency changes in the functional groups of the AC within the range of wave number of 400-4,000 cm À1 .

Experimental design and empirical modeling
In this study, Design-Expert software (version 11.00) was used to design experiments and model experimental data through RSM using the CCD technique, and the results were completely analyzed using the analysis of variance (ANOVA) (Mousavi et al. 2011). Response surface methodology is a mathematical and statistical method that uses the quantitative data from experiments to analyze the effects of independent variables and to investigate the interaction of several parameters affecting the process by varying them simultaneously (Mourabet et al. 2017). The mechanism of this statistical method is based on the statistical design of the experiments, evaluation of the coefficients of mathematical models, prediction of results and investigation of the accuracy of the model (Hermawan et al. 2015;Mousavi & Nazari 2017). The effects of four independent variables namely contact time (X 1 ), initial pH (X 2 ), adsorbent dosage (X 3 ) and initial dye concentration (X 4 ) were investigated by means of CCD. The effect of input factors on the responses was investigated by ANOVA through a statistical evaluation of P-value and F-value of regression coefficients (P , 0.05). In addition, the validity of the model was reported in terms of the coefficient of determination (R 2 ), adjusted coefficient of determination (Adj. R 2 ) and sufficient accuracy (AP). Finally, the three-dimensional response level diagrams have been developed to show the interrelationship between independent factors and their related effects on the response. All the experiments were carried out in three duplicate (78 runs), and the means values were used for calculation. The average of each run, except for the six central runs, is presented in Table 1. Analytical errors and standard deviation were calculated at all stages, and the standard deviation from the mean values in all runs was less than +5%. The quadratic equation model (Equation (1)), which includes all interaction terms, was used to calculate the predicted response: Uncorrected Proof The quadratic model based on Equation (1) was applied to evaluate the coefficients of the statistical model, where Y is the response, β 0 is the constant coefficient, β i is the linear coefficients, β ij is the interaction coefficients, β ii is the quadratic coefficients, x i and x j are the coded values of the investigated variables and e is the statistical error term.

Batch equilibrium studies
Batch experiments were performed to determine the AC capabilities to adsorb MB from aqueous solutions. In this study, different values of independent variables, including pH (3-11), adsorption dosage (0.25-12.25 mg L À1 ), contact time (10-90 min) and initial dye concentrations (100-500 mg L À1 ) have been applied during the adsorption process based on designed conditions using CCD. MB with 98% purity was used for preparing the stock solution. 100 mL of MB aqueous solution with desire amount of adsorbent was added to 250 mL Erlenmeyer flasks, which it was placed on the stirrer (model Shimi fan, Iran) that adjusted at 200 rpm when temperature was 25 + 2°C. The pH of the solution was adjusted at the desire value by adding HCl (0.1 M) and NaOH (0.1 M). The samples were withdrawn at regular intervals. The samples were centrifuged at 4,000 rpm for 10 min. The MB concentrations of the centrifuged solutions were analyzed using a UV-Vis Uncorrected Proof spectrophotometer (Germany, Jenway 6305) at a wavelength of maximum absorbance at 668 nm to determine removal efficiency. The amount of adsorbed dye and the MB removal efficiency (%) were calculated using the following equations (Equations (2) and (3)) (Foo & Hameed 2012): where C 0 and C e are the initial and final dye concentrations (mg L À1 ), respectively; q e is the amount of adsorbate per mass of the adsorbent (mg g À1 ), V is the volume of solution (L) and W is the mass of adsorbent (g).

Desorption procedure
To reconstitute, the first 100 mL of MB aqueous solution with a concentration of 300 mg L À1 along with 1 g of adsorbent was added to 250 mL Erlenmeyer flasks. The Erlenmeyer was then placed on a shaker at a speed of 200 rpm for 180 min. In the next step, the MB concentration of the centrifuged solution is measured by a UV-Vis spectrophotometer. The spent AC was separated from the solution and washed with distilled water and then dried immediately at 110°C for 3 h. After drying, it transfers to an Erlenmeyer flask containing 100 mL of 95% ethanol. It was then placed on a shaker under the same conditions as the previous for 4 h. To investigate the efficacy of ethanol regeneration, the desorbed concentration was measured with a spectrophotometer and the desorbed percentage was calculated according to Equation (4). The regenerated AC was again used for MB adsorption.
where C ad is the concentration of solution that is adsorbed on the adsorbent, and C de is the concentration desorption.

RESULTS AND DISCUSSION
AC characterization Figure 2 shows the SEM images of the prepared AC from grape wood wastes. The large number of honeycomb pores is clearly visible on the AC surface. The SEM image represents a porous structure for the prepared AC, which causes to increase in MB adsorption on the surface of AC resulting in high MB removal (Gao et al. 2013;Inyinbor et al. 2016). The FTIR method can identify the adsorbent functional groups responsible for the adsorption of the adsorbate on the adsorbent (Chatterjee et al. 2012). Different functional groups have been determined on the surface of prepared adsorbent such as hydroxyl, carboxyl, lactonic and carbonyl groups. The peak at about 960 cm À1 may be due to the vibration of the C-C or C-H Uncorrected Proof bands . The area between 1,000 and 1,200 cm À1 is related to the C-O functional group. The peaks at 1,700-1,800 cm À1 are associated with the vibrational band C ¼ O, and the band at 3,200-3,600 cm À1 corresponds to the O-H tensile vibration of the hydroxyl functional groups (Pirsaheb et al. 2016;Nayeri et al. 2019). The peak at 2,852 cm À1 may be due to corresponds to the presence of -CH 2 stretching of aliphatic groups. The peaks at 1,637 and 1,321 cm À1 indicate the presence of C ¼ C stretching of the phenol functional group and C-N stretching of the amine functional group, respectively (Banerjee & Chattopadhyaya 2017).
Based on BET analysis, the surface area of AC is determined as 119.084 m 2 g À1 (Figure 3).

Adsorption process analysis and modeling
The results of ANOVA for studding response (MB removal efficiency) were represented in Table 2. The F-value of model (55.23) indicates that the model has a significant level. Only 0.01% of the 'F-value of model' is likely to be due to noise. The P-value is used to determine the significance of each parameter. Pred. R 2 and Adj. R 2 for dye removal are 0.8797 and 0.9079, respectively, which confirms that there is a good match between predicted data and experimental data. When the Uncorrected Proof P-value is less than 0.05, the parameters or their interactions are statistically significant (Mousavi & Ibrahim 2016). As shown in Table 2, the P-values of X 1 , X 2 , X 3 , X 4 , X 1 X 3 , X 1 X 4 , X 2 X 4 and X 3 X 4 are less than 0.05, indicating that the effects of these variables are significant in the MB removal efficiency. The results of data modeling (Equation (5) The same experiments in optimal conditions were repeated and its results indicated the ability of model to predict the MB removal efficiency with a relative standard deviation of less than 3% (Coruh & Elevli 2014;. The lack of fit value of 2.69 is not significant and confirms that the model is adequate. Adequate precision measured the signal to noise ratio that a value of this parameter greater than 4 is generally essential. In this work, the obtained adequate precision was 27.96 for the degradation of MB that confirmed an adequate signal; thus, the obtained model in this work could be used to navigate the design space (Equation (6)).

Effects of initial pH and contact time
The initial pH of the solution can affect the adsorbent surface charge and either promote or prevent the adsorption of dye on the adsorbent surface (Da Silva Lacerda et al. 2015). Figure 5 shows the interaction effects of initial pH and contact time as a function of the response in the MB adsorption process. The elliptical plotter counter shows the effect of a meaning interaction between two variables. In order to study the effect of pH on the MB dye removal efficiency, different initial values of pH (3, 5, 7, 9 and 11) were examined. As observed, the value of MB removal is low at a pH of 3-7, while its value increases up to 98% at a pH of 7-11. The pH ZPC plays an important role in the performance of any adsorbent by determining the pH at which the surface is net electrical neutrality. The results of various studies have shown that the presence of negatively charged functional groups on the adsorbent surface is necessary for the adsorption of the basic dye (Parthasarathy & Arivoli 2018). According to the results of studies, at low pH when pH , pH ZPC (pH , 3.8), the surface of AC is surrounded by H 3 O þ ions (Geçgel et al. 2013;Boumediene et al. 2018). Therefore, due to the competition between H 3 O þ ions and also the electrostatic repulsion between the dye molecules and the positively charged active adsorption sites on the surface of AC, the adsorption of dye molecules is reduced (Zhang et al. 2015a). In contrast, at higher pH values when pH . pH ZPC (pH . 5), the amount of hydroxyl ions (OH À ) in the solution increases, which cause to increase the number of negatively charged sites and increase interaction between positive dye molecules and negative surface of the prepared AC (Table 3) (Hassan et al. 2017;Mousavi et al. 2020). Zhang et al. (2015) reported that the MB adsorption rate over natural palygorskite adsorbent improves with increasing the pH of solution (Zhang et al. 2015b).  was also observed that the maximum MB adsorption over AC made from Nasturtium microphyllum was observed at alkaline pH .
To recognize the kinetics of the adsorption process and to determine the equilibrium time for maximum adsorption, the contact time between dye and AC in the range of 10-90 min has been investigated. In Figure 5, the effect of contact time on the % removal of MB dye is shown at 6.25 g L À1 adsorbent dosage and initial concentration of 100 mg L À1 . Based on the results, the adsorption rate of MB increased with increasing contact time, and the maximum value of MB adsorption of about 98% was achieved at 90 min. Because with increasing contact time, the adsorbent binding sites have enough time for the MB adsorption . Dahri et al. (2015) applied Casuarina equisetifolia as an adsorbent for the removal of malachite green and MB from aqueous solutions. They found that the adsorption efficiency for both of the  Uncorrected Proof dyes was improved with the enhancement of contact time (Dahri et al. 2015). Kaykioglu and Guneşreported the same observations for MB removal using rice husk ash (Kaykioglu & Güneş2016).

Effect of adsorbent dosage and dye concentration
Adsorption experiments were carried out to investigate the combined effect of dye concentration and adsorbent dosage on the MB dye removal (Figure 6). The influence of the initial MB dye concentration in the range of 100-500 mg L À1 on the MB adsorption capacity of AC was investigated. The results showed that with increasing MB concentration, the adsorbent efficiency in the removal of MB decreased. However, when the initial concentration of the MB was at the lowest amount (100 mg L À1 ), the value of removal efficiency increased to 98%. But, with increasing the initial concentration of MB, the required active sites to adsorb MB molecules are not available, and adsorption sites on the surface of adsorbent are saturated using dye molecules; therefore, the efficiency of system decreased (Etim et al. 2016). Rahman et al. (2012) and Geçgel et al. (2013) reported the same trends for MB removal using AC produced from rice husk and peas' skin, respectively (Rahman et al. 2012;Geçgel et al. 2013). In Figure 6, the effect of adsorbent dosage on the % removal of MB dye is shown at a pH ¼ 11 and contact time ¼ 90 min. The results according to Figure 6 showed that with increasing adsorbent dosage, the MB removal increased. The maximum removal rate of 98% was observed when the dosage of adsorbent was 12.5 g L À1 . An increase in adsorbent dose to 12.5 g L À1 would cause a corresponding increase in pores and active sites on the surface of adsorbent (Abechi et al. 2011). This trend continues until all the dye molecules are adsorbed on the adsorbent active sites (Sartape et al. 2017). Mulugeta & Lelisa (2014) also reported an increase in the removal efficiency of MB onto untreated Parthenium hysterophorus weed when the adsorbent dosage was increased from 0.1 to 1 g L À1 (Mulugeta & Lelisa 2014).

Optimization of experimental conditions
Process optimization was performed using a numerical optimization program using the DOE software (Almasi et al. 2017a). The optimization criteria were determined as shown in the overlay plot (Figure 7). According to Figure 7, the optimum condition achieved when contact time, pH, AC dosage and dye concentration were 84. 23 min, 10.88, 11.40 and 204.94 with 0.99 desirability. To explore an optimum region in the design space, an acceptable range of adsorption efficiency (.90%) was considered, and a region enclosed with reaction time (84.23 min) was determined.

ADSORPTION KINETICS AND ISOTHERM MODELS Adsorption isotherm
Adsorption isotherms are one of the basic requirements for optimizing the design of the adsorption system. In fact, the adsorption isotherm represents the relationship between the mass of adsorbed dye per unit mass of adsorbent and the liquid phase dye concentration at constant temperature (Almasi et al. 2017b). In this study, isotherm studies were carried out to determine the maximum adsorption capacity and select the best equilibrium model. Langmuir and Freundlich isotherm equations were used to investigate the adsorption balance of MB dye on the AC prepared from grape wood. Langmuir isotherm describes monolayer adsorption on homogeneous sites with negligible interaction. In the Langmuir isotherm, when a dye molecule occupies one site, no other adsorption actually occurs at that site (Sartape et al. 2017). The linear form of Langmuir isotherms is shown in the following equation: where C e (mg L À1 ) is the equilibrium concentration of MB dye in the solution, q e (mg g À1 ) is the adsorption capacity at equilibrium, Q m (mg g À1 ) is the maximum adsorption capacity and K (L mg À1 ) is the effective dissociation constant that relates to the affinity binding site. The values of Q m and K are obtained from the intercept and the slope of the linear plot of C e /q e against C e . The Langmuir model, RL value, demonstrates the type of adsorption. RL is the separation coefficient, and it can be defined as a dimensionless transition variable (Equation (8)). RL . 1 indicates undesirable adsorption type, RL ¼ 1 implies linear adsorption type, RL ¼ 0 shows irreversible adsorption and 0 , RL , 1 shows a favorable adsorption (Pal & Deb 2014): The Freundlich isotherm expresses the dye adsorption on the heterogeneous surface with interaction. In fact, in the Freundlich isotherm, the amount of dye adsorption changes with the exponential distribution of sites and adsorption energies (Equation (9)) (Malik 2004): where n and K F are the Freundlich adsorption isotherm constants, which indicate adsorption intensity and adsorption capacity, respectively. The Freundlich isotherm constants K F and 1/n can be reported based on the plot of lnq e versus lnC e , which has been presented in Figure 8.
In the adsorption process, the magnitude of the exponent, 1/n, indicates the favorability of adsorption (Malik 2004). If the values are 1/n , 1, it indicates that with the appearance of new adsorption sites, the type of isotherm to be required, the adsorption intensity and the adsorption capacity increase. On the other hand, if it is 1/n . 1, it indicates the adsorption bond weakens, which reduces the dye adsorption capacity (Kahrizi et al. 2018).
The Langmuir and Freundlich isotherms for MB adsorption are given in Figure 8(a) and 8(b), respectively. Table 6 represents the calculated parameters for Langmuir and Freundlich isotherm models. The high R 2 values (0.99) indicate that adsorption follows the Langmuir model, with Q m ¼ 5.88 mg g À1 and b ¼ 14.98 L mg À1 . The higher value of Q m (≫1) indicates a strong adsorbate adsorbent interaction (Table 4). Therefore, it can be argued that MB adsorption on the AC prepared from Uncorrected Proof grape wood waste is a monolayer adsorption type. According to the Langmuir isotherm, the amount of RL for MB adsorption was 0 , RL , 1, which approves favorable adsorption of MB.

Kinetic study
The kinetic study is important part of the adsorption process in water and wastewater treatment because it can provide valuable information about the reaction pathways and the mechanism of dye adsorption reactions (Inbaraj et al. 2006). In addition, through adsorption kinetics, the rate of solute adsorption can be predicted, which in turn depends on the retention time or adsorption reaction at the interface of the solid solution (Nekouei et al. 2015). Therefore, the adsorption rate or kinetic of adsorption is an important factor to select adsorbent materials, design process, control operation and predict adsorption amount (Rahman et al. 2012;Pal & Deb 2014). In this study, assuming that the measured concentrations are equal to the surface concentrations, pseudo-first-order (Equation (10)) and pseudo-second-order (Equation (11)) equations  Uncorrected Proof were used to describe the adsorption mechanism of kinetic models (Karthikeyan et al. 2010;Sarici-Özdemir & Önal 2014).
Log (q e -q t ) ¼ log q e -K 1 2:0303 t (10) where q e is the amount of dye adsorbed at equilibrium (mg g À1 ), q t is the amount of dye adsorbed at any time t (mg g À1 ), K 1 is the first-order rate constant (min À1 ) and K 2 is the pseudo-second-order rate constant (g mg À1 min À1 ). The rates of adsorption have been calculated based on both aforementioned models, and their results were illustrated in Figure 9(a) and 9(b). The obtained kinetic parameters for both of the models were also represented in Table 5. The results showed that the pseudo-second-order equation with regression coefficient (R 2 ) of 0.99, maximum adsorbent amount of 62.41 mg g À1 and rate constant of 0.02 min À1 is better matched with experimental data compared to the pseudo-first-order model.

Adsorption thermodynamics
The properties and mechanisms of adsorption can be evaluated through thermodynamic study. Thermodynamic parameters such as standard enthalpy change (ΔH o ), Gibbs free energy change (ΔG o ) and standard entropy change (ΔS o ) can be calculated using the following equations based on the equations, the values of lnK are plotted against 1/T, and the values of ΔH o and ΔS o can also be calculated from the slope and intercept of the plot (Equations (12)-(15)) (Purkait et al. 2007): where q e is the MB concentration at equilibrium onto AC (mg L À1 ), C e is the MB concentration at equilibrium in solution (mg L À1 ), T is the temperature in Kelvin, R is the ideal gas constant (8.314 Â 10 À3 kJ(mol K) À1 ) and K L is the distribution coefficient. Thermodynamic parameters at three different concentrations at 24 + 2°C were investigated and the results are summarized in Table 6. As can be observed, the AC at the three concentrations studied have negative ΔG°values; thus, it can be proved that the adsorption process at these concentrations is spontaneous in nature. In addition, the more negative the Gibbs standard energy value, the higher the adsorption driving force and consequently the higher the adsorption capacity of the adsorbent. ΔH°was positive at all concentrations, which implies that the adsorption process is endothermic. On the other hand, ΔH°has values below 40 kJ mol À1 , indicating that the adsorption is physical-type (Purkait et al. 2007;Moradi & Zare 2011).

CONCLUSIONS
This study shows that AC made from grape wood waste can be a promising and efficient adsorbent for the removal of MB from aqueous solution. The effects of initial pH, contact time, adsorbent dosage and initial concentration in the removal of dye were investigated using RSM. For the characterization of AC surfaces, FTIR and SEM methods were used. The Uncorrected Proof optimum conditions of independent variable achieved when initial MB concentration, contact time, adsorbent dose and pH were 100 mg L À1 , 90 min, 12.25 g L À1 and 11, respectively, to yield a maximum MB removal of 98%. The results showed that the adsorption of MB dye on AC as an adsorbent is consistent with the Langmuir model .Also, the pseudo-second-order kinetic model gave a better fitting of the kinetic data. Finally, the achieved results indicated that AC produced from grape wood waste has a relatively high adsorption capacity and can be used as a natural adsorbent.

ACKNOWLEDGEMENT
The authors gratefully acknowledge the Research Council of Kermanshah University of Medical Sciences for the financial support.

DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.