Microwave-assisted preparation of almond shell-based activated carbon for methylene blue adsorption

2.1 Raw material

The AS was purchased from the local market, Shihezi, China, washed with distilled water, and dried in an oven at 100°C for 24 h to remove the moisture content. The dried AS were crushed and sieved to a uniform particle size of 100 meshes. The proximate analyses of AS were ash 1.67%, fixed carbon 29.40%, moisture 5.31%, and volatiles 63.62%. All analyses were performed in three replicates.

2.2 Preparation of ASAC

The microwave-assisted process was performed using a modified microwave oven (MM823LA6-NS, Midea, China) at a frequency of 2.45 GHz. Mixtures (4.0 g) with different mass ratios (X₁) of ZnCl₂ (Zhiyuan, Tianjin, China) and AS and particle size of 100 mesh were put in a quartz tube reactor placed in a microwave oven. The activation processes were then subjected to various microwave heating times (X₂, min) and microwave radiation powers (X₃, W) under 100 cm³/min of nitrogen flow. The resulting material was washed repeatedly with hot distilled water until solution pH reached 7.0. The obtained ASAC were dried at 80°C for 24 h and then stored for further analyses.

The percentage of ASAC yield was calculated using the following equation:

where W is the weight of the prepared ASAC after activation and W₀ is the weight of the raw materials before activation.

2.3 Experimental design and response surface methodology

RSM was applied to activation process optimization as it is a useful statistical tool for modeling and analysis of multivariables. This method was chosen for fitting a quadratic surface with the fewest number of experiments. In this step, BBD method was applied in three levels in order to design the activation experiments. Table 1 shows the ranges and levels of the experimental parameters: mass ratio (X₁), activation time (X₂), and microwave power (X₃). MB adsorption (Y₁) was taken as the process responses.

In order to study the relationship between the response variable (MB number and yield) and independent variables, the experimental data were fitted to a second-order polynomial regression model, expressed by Eq. (2)

where Y is the response variable, χ_i and χ_j are the independent variables, β_0,β_i,β_ii, and β_ij are the regression coefficients (β₀ is constant term, β_i is the linear effect term, β_ii is quadratic effect term, and β_ij is the interaction effect term).

Design Expert software (version 8.0.6, Stat-Ease Inc., Minneapolis, MN, USA) was used to elaborate the experimental design, evaluate the model, and estimate the subsequent regression, variance, and response surface analysis.

2.4 Characterization of ASAC

The physical properties, textural morphology, and chemical properties of ASAC were measured using renowned analytical techniques. The surface physical properties of the ASAC were characterized with Quanta Chrome Instruments Quadrasorb SI (USA) and N₂ adsorption measurements at 77 K. The S_BET was calculated with the Brunauer-Emmett-Teller (BET) equation. The nitrogen volume held at the highest relative pressure (P/P₀=0.95) was used to calculate the total ASAC pore volume. The micropores surface area, volume, and external surface area were obtained using the t-plot method.

SEM was also employed to study the textural structure of the raw material and adsorbent after the activation process. The morphology of AS and ASAC were characterized by SEM (JEOL, JSM-6490LV, Japan) with a secondary electron beam and an acceleration voltage of 3 kV. The samples were coated with gold to ensure that the particles have suitable conductivity after having been dried overnight at 80°C under vacuum.

Chemical characterization of surface functional groups was investigated using FTIR (Magna-IR 750, Nicolet, USA) in the scanning range of 4000–400 cm^-1; 10 scans were taken at a resolution of 1 cm^-1. The samples were mixed with KBr powder, and the mixtures were pressed into a pellet.

Meanwhile, an analysis of the chemical element content at the surface of AS and ASAC was performed using XPS. The XPS analysis was conducted using AMICUS/ESCA 3400 spectrometer (Japan), Mg radiation at 240 W and 12 kV under ultrahigh vacuum. Pass energy was set at 276 and 300 eV for survey and high-resolution spectra, respectively. The XPS spectra were calibrated by taking the graphitic peak as 284.5 eV. The surface atomic concentrations were calculated from the survey spectra after correcting the relative peak areas by sensitivity factors.

The point of zero charge (pH_pzc) was determined by adjusting the pH of a 50 ml 0.1 mol/l NaCl (Hengxing, Tian Jin, China) solution (prepared in boiled water to eliminate CO₂) to a value between 2 and 12. Approximately 0.15 g of ASAC was added, and the final pH was measured after 48 h under agitation. The pH_pzc is the point where pH_initial-pH_final=0.

2.5 Adsorption capacity

About 0.1 g of ASAC was put into a 150 ml Erlenmeyer flask with stopper containing 50 ml of a 1.5 g/l MB solution at pH=6.16 in order to determine the adsorption capacity of ASAC. The flask was vibrated at 170 rpm in a thermostatic water bath shaker at around 20°C. After adsorption equilibrium time of 30 min, which was obtained by a series of preliminary experiments, samples were separated by filtration; the MB solution concentration was analyzed by absorbance measurements using a double-beam UV-75N (Yuanxi, Shang Hai, China) at a wavelength of 665 nm. The MB uptake of ASAC, q (mg/g), was calculated by Eq. (3):

where C₀ (mg/l) and C_e (mg/l) are the initial and equilibrium concentrations of MB, respectively, V (l) is the volume of solution and M (g) is the dry mass of ASAC.

2.6 Adsorption kinetics

The adsorption kinetics procedures were carried out at initial MB concentrations of 300 and 500 mg/l at 20°C. About 0.1 g ASAC was put into 150 ml conical flasks with 50 ml MB solution; the samples were separated by filtering at different contacting time intervals (2, 5, 10, 15, 30, 45, 60, and 120 min). The residual concentrations of MB were measured, and adsorption amounts of ASAC were calculated with the above method.

To understand the adsorption dynamics of MB on ASAC in relation to time, the experimental data were analyzed using pseudo-first-order [14] and pseudo-second-order [15] kinetic models. The expression of the pseudo-first-order kinetic model is given by Eq. (4):

where q_e (mg/g) is the amount of MB adsorbed at equilibrium, k₁ (min^-1) is the pseudo-first-order rate constant, and t (min) is the contact time. The form of the pseudo-second-order kinetic model is given by Eq. (5):

where k₂ is the pseudo-second-order rate constant (g/mg min).

2.7 Adsorption isotherms

The Langmuir and Freundlich adsorption models were run as discussed below to analyze the validity of the adsorption data. The adsorption isotherms of MB on AC were investigated by determining the adsorption amounts of MB in different concentrations (50, 200, 300, 550, 850, 910, and 1000 mg/l) at the equilibrium adsorption time of 30 min.

The Langmuir isotherm equation [16], which is the most commonly used for monolayer adsorption on to a surface with a finite number of identical sites, was represented by the following Eq. (6):

where C_e (mg/l) is the equilibrium concentration of the MB dye, q_m (mg/g) is the maximum adsorption capacity, and K_L (l/mg) is the Langmuir adsorption equilibrium constant, which is related to the affinity of the binding sites.

The Freundlich isotherm equation [17], the most important multilayer adsorption isotherm for heterogeneous surfaces, is described by the following Eq. (7):

where K_F [(mg/g)·(l/mg)^-1/n] and n are Freundlich parameters related to the adsorption capacity and adsorption intensity, respectively.

3.1 RSM and model fit

BBD was used to elaborate the correlation between the AC preparation variables to the MB adsorption and yield. Seventeen experiments were performed for the RSM optimization of three parameters at three levels. The results of these experiments are presented in Table 1. Runs 13–17 at the center point were used to check the experimental error. The q_m values for ASACs, which were prepared according to the experimental conditions mentioned above, ranged from 164.82 to 314.20 mg/g, while the carbon yield ranged from 31.0% to 49.7%. The quadratic model was selected for the response of MB adsorption value, as suggested by the software. The final empirical model in terms of coded factors for MB adsorption value (Y₁) is shown in Eq. (8):

Negative signs in front of the terms indicate antagonistic effects, whereas the positive signs indicate synergistic effects. Variance was employed to justify the adequacy of the model. Variance for the quadratic model of MB adsorption is shown in Table 2, where the F value of 9.92 and Pr>F of 0.0032 proved that the model is significant. The value of model terms Pr>F less than 0.05 indicates that the model terms are significant. The results showed that X₁, X₃, and X₁² are all significant model terms. These significant effects signified the importance and effectiveness of the selected independent variables for the experiments. As shown in the variance analysis, the lack-of-fit test was non-significant, indicating that the model can be well fit to the studied data. The 3D response surface (Figure 1) is generally the graphical representation of the regression equation. The effect of heating time on the experimental response of q_m values can be seen in Figure 1A and C. The heating time has only a weak effect on MB adsorption capacity. The indistinctive effect of the microwave heating time on adsorption capacity values has also reported by Zhong et al. [18]. Figure 1A and B shows the effect of the mass ratio on the response. The value of q_m evidently increased with the increase in mass ratio at high microwave power levels. The use of the activating reagent in the microwave-assisted activation promoted the bond breaking and elimination of some volatile species in the precursors [19]. ASACs with increased surface area and higher adsorption capacity were obtained as a result. Our results on the effect of microwave power (Figure 1B and C) demonstrated that the highest q_m values were obtained at microwave powers over 820 W. This trend was attributed to the fact that the high microwave power increases system temperature, reducing the activation energy of the carbon-reagent reaction and promoting the development of pores on the materials [20].

Figure 1:

Three-dimensional graphic of response surface for response: mass ratio and activation time (A), mass ratio and microwave power (B), and activation time and microwave power (C).

3.2 Process optimization

Adsorption efficiency toward organic compounds, such as MB dye, is the most important aspect of ASAC economic feasibility and marketing. Design Expert software version 8.0.6 was used to calculate an optimized value by selecting the highest responses from the experimental results. The optimum calculated conditions were a mass ratio of 3:1, an activation time of 15 min, and a microwave power of 900 W. Under optimal conditions, a carbon yield of 41.19% and an MB adsorption capacity of 285.69 mg/g were predicted. Meanwhile, the practical repeat experiments were carried out to check the accuracy of the predicted results which showed the carbon yield of 39.67% and an MB adsorption capacity of 314.20 mg/g. The relative errors between the experimental and predicted values for the carbon yield and MB adsorption capacity are 3.69% and 9.07%, respectively, indicating that the process optimization is sufficiently accurate. The present study shows that ASAC’s adsorption capacity for MB (Table 3) is quite remarkable when compared to previously reported adsorbents [11, 21, 22, 23].

3.3 Characterization of ASAC

3.3.1 Specific surface area and pore structure of ASAC

The N₂ adsorption-desorption isotherm at 77 K is shown in Figure 2A. The isotherm resembles a combination of type I and type II isotherms, as defined by the International Union of Pure and Applied Chemistry; this adsorption behavior exhibits characteristics of microporous-mesoporous structures [24]. The pore size distribution of ASAC is shown in Figure 2B. The sharpest peak is observed at a pore diameter between 0.5 and 1.0 nm, with an average pore size of 0.83 nm, which indicated the essentially microporous texture of ASAC. The specific surface area of 839.60 m²/g was calculated using the BET equation. The structural characteristics of ASAC are listed in Table 4. ASAC exhibited a relatively high S_BET of 839.60 m²/g and a V_T of 0.406 cm³/g.

Figure 2:

N₂ adsorption and desorption isotherms at 77 K (A) and pore size distribution (B) of ASAC.

Table 4:

Surface physical characteristics of the ASAC and other ACs.

Precursors of ACs	SBET (m2/g)	Micropore area (m2/g)	VT (cm3/g)	Micropore volume (cm3/g)	Average pore width (nm)	References
Almond shell	839.60	383.64	0.406	0.152	0.83	This work
Pistachio shell	700.53	–	0.375	–	2.14	[25]
Rice husk	752	346	0.64	0.26	3.41	[26]
Cotton stalk	794.84	156.69	0.63	0.083	3.20	[23]

3.3.2 SEM analysis of ASAC

Figure 3 shows SEM images of AS and ASAC prepared under the optimum conditions (i.e. mass ratio of 3:1, activation time of 15 min, and microwave power of 900 W). The AS surface is planar and undulating and mainly consists of macropores. The number of pores of the ASAC significantly increased after the activation process. The ASAC surface exhibited well-organized, pronounced, and uniform porosity (Figure 3B), which significantly increased its surface area and resulted in a higher number of pores available for MB adsorption. This observation was supported by the physical parameters, as summarized in Table 3, which highlighted the pore development during the microwave irradiation period. A similar phenomenon was also observed by Njoku et al., who reported the surface morphology of rambutan peel-based ACs which have a series of uniform cavities around the surface that are caused by KOH activation [4].

Figure 3:

SEM images of AS (A) and ASAC (B).

3.3.3 Functional groups analysis of ASAC

The surface chemistry of AS and ASAC was studied by FTIR spectroscopy in the infrared spectral region of 4000–400 cm^-1 (Figure 4). AS contained much more bands than ASAC as well as some major differences. The broad band at approximately 3450 cm^-1 was observed for all samples and was attributed to the O–H stretching vibration of the hydroxyl functional groups [13, 27]. The intense band at approximately 2950 cm^-1 in AS was assigned to the C–H stretching vibration [28]. This peak disappeared in ASAC, indicating the removal of hydrogen by activation. A band at approximately 2359 cm^-1 for AS and ASAC was attributed to C≡C (alkynes), which was also reported by Foo et al. [29]. A series of complex bands, which was significantly reduced after activation, was detected in the range of 800–1630 cm^-1 in AS. This phenomenon is possibly due to the breakdown of the chemical bonds in the raw material [30]. The bands located at 1630, 1390, 1255, 1085, and 800 cm^-1 in the AS spectra were attributed to O-H bending vibrations, C-O-C (esters or phenol), C-O-C (ether) group, C-O stretching vibration, and C-H blending or benzene. These bands in ASAC were weakened in ASAC [5, 31]. The signal at 550 cm^-1 was assigned to the C-O-H bending vibration [32].

Figure 4:

FTIR spectra of AS and ASAC.

The compositional chemistry of the AS and ASAC surfaces was analyzed by XPS, as shown in Figure 5A. The spectra contained distinct peaks for C and O. The relative content of these elements was calculated and listed in Table 5. The percentage amounts of C, O, and S were 72.07%, 26.88%, and 1.05% in AS and 86.57%, 13.43%, and 0.00% in ASAC, respectively, thereby indicating a small increase in the C amount and a decrease in the O amount after activation procedure.

Figure 5:

XPS survey spectra and high-resolution spectra of AS and ASAC: (A) survey scans; (B) AS: C 1s; (C) ASAC: C 1s; (D) AS: O 1s; (E) ASAC: O 1s.

Table 5:

Relative contents of various elements and peak parameters of different C 1s and O 1s components of AS and ASAC samples based on the XPS spectra.

	Relative content (%)			Relative content (%) of C 1s				Relative content (%) of O 1s
	C	O	S	C-C	C-O-R	C=O	C-OOR	C=O	C-O	C-OOR	C-OOH
AS	72.07	26.88	1.0	13.15	41.97	25.59	19.29	22.55	32.73	24.05	20.67
ASAC	86.57	13.43	0.0	59.18	22.18	11.47	7.17	39.17	26.21	17.18	17.44

The high-resolution spectra of C 1s and O 1s were further analyzed using a curve-fitting procedure based on the Gaussian-Lorentzian function after baseline subtraction with Shirley’s method. Four components were considered to comprise the C 1s spectra [33] with chemical shifts: C-C (284.5 eV), C-O-R (285.8 eV), C=O (287.0 eV), and COOR (289.2 eV). The fitting curves of these groups are illustrated in Figure 5B and C; Figure 5D and E shows that the deconvolution of the O 1s spectra for the AS and ASAC produced four peaks at 533.3, 534.2, 535.1, and 535.7 eV; these peaks represented O=C, O-C, R-O-C=O, and C-OOH, respectively. The relative content of each group is listed in Table 5. These results suggest that the C–C group was predominant in ASAC as it made up close to 60% of the carbon species. Moreover, relative oxygen contents showed a significant decrease. The analysis for oxygen species displayed an increase in the C=O group and a decrease in other oxygen groups in ASAC (Figure 5D and E). Liu et al. have reported on a similar reduction of oxygen-containing groups in AC, which is consistent with the FTIR and XPS spectra [34].

3.4 Adsorption studies of MB

3.4.1 Effect of pH on the adsorption of MB onto ASAC

The adsorption behavior of MB onto ASAC at different pH values is shown in Figure 6. The effect of pH was conducted by varying the pH of dye solutions of an initial concentration of 500 mg/l, from 2 to 12. Adsorption capacity increased with increasing pH. Adsorption capacity displayed a particularly significant enhancement at pH of 6–8. This may be related to the protonation of MB in acidic medium, which triggers a competition effect between excessive amounts of H⁺ and cationic dye on adsorption sites. ASAC may become more negatively charged as the solution pH increases, causing the formation of electric double layers which facilitates a change in the polarity, leading into the enhancement of adsorption capacity. The pH effect can be described in terms of pH_pzc. The pH_pzc of ASAC was found to be 6.16. When the solution pH<pH_pzc, the AC adsorbent will react as a positive surface, but when the solution pH>pH_pzc, it will act as a negative surface. At pH values above 6.16, the electric double layer changes its polarity, which causes an increase in the surface charge density and electrostatic force of attraction. This led us to the dual adsorption of MB onto ASAC: physical adsorption and chemical adsorption. The chemical adsorption mechanism includes the reaction between hydroxyl and carbonyl functional groups on the surface of the ASAC and MB. The positive charge on MB and OH^- on ASAC react as follows:

Figure 6:

Effect of pH on the adsorption of MB onto ASAC at 20°C (C₀=500 mg/l, contact time=30 min, AC mass=0.1 g).

      (9)

3.4.2 Adsorption kinetics

The experimental data obtained for the AC from the initial MB concentrations (300 and 500 mg/l) were fitted to the pseudo-first-order and pseudo-second-order models (Figure 7). Rapid adsorption occurred in the first 15 min, and equilibrium was reached after 30 min, indicating a favorable interaction between the adsorbent and adsorbate (Table 6). The data show that the adsorption of MB onto ASAC was best described by the pseudo-second-order model, R²=0.999; similar q_e values were calculated and experimentally determined. These findings suggest that the MB-ASAC system followed the pseudo-second-order model and that the overall adsorption process rate was controlled by physisorption at the beginning of adsorption and chemisorption thereinafter, which has already been explained by the pH effect on MB adsorption onto ASAC. The same behavior was also observed by Foo et al. [29].

Figure 7:

Effect of time on the adsorption of MB onto ASAC at 20°C (pH=6.16, 0.1 g of ASAC).

Table 6:

Pseudo-first-order and pseudo-second-order kinetic rate constants for the adsorption of MB onto ASAC at 20°C.

C0 (mg/l)	qe,exp (mg/g)	Pseudo-first-order			Pseudo-second-order
		K1 (min-1)	qe (mg/g)	R2	K2 (g/mg min)	qe (mg/g)	R2
300	148.59	0.770	148.45	0.785	0.050	149.25	0.999
500	249.56	0.508	248.51	0.842	0.160	250.00	0.999

3.4.3 Adsorption isotherms

The plots of the nonlinear Langmuir and Freundlich isotherm models for the adsorption of MB onto ASAC at 20°C according to Eqs. (6) and (7) are presented in Figure 8. The isotherm parameters determined from the slopes and the correlation coefficients of the isotherm models are summarized in Table 7. The equilibrium data were valid for the Langmuir isotherm model (R_adj²=0.989), demonstrating that the adsorption of MB onto ASAC from aqueous solutions proceeds via monolayer formation. This result was in accordance with the previous nitrogen adsorption-desorption result which authenticates the monolayer adsorption phenomenon. The present results are relatively consistent with previous research into the adsorption of MB onto ACs that had been prepared from pomelo skin and oil palm (Elaeis) empty fruit bunches by microwave-assisted chemical activation [32, 35].

Figure 8:

Plots of Langmuir and Freundlich models for adsorption of MB onto ASAC (pH=6.16, contact time=30 min, 0.1 g of ASAC).

Table 7:

Isotherm model parameters for the adsorption of MB onto ASAC at 20°C.

Isotherms	Langmuir			Freundlich
Constants	qm (mg/g)	KL (l/mg)	R2	n	KF (mg/g)(l/mg)1/n	R2
	500.00	0.307	0.989	2.22	112.449	0.717