Adsorption of methylene blue onto betel nut husk-based activated carbon prepared by sodium hydroxide activation process

In this study, activated carbon (AC) was prepared from agro-waste betel nut husks (BNH) through the chemical activation method. Different characterization techniques described the physicochemical nature of betel nut husks activated carbon (BNH-AC) through Fourier transform infrared spectroscopy (FTIR), Brunauer – Emmett – Teller (BET), scanning electron microscopy (SEM), and pH point of zero charge. Later, the produced AC was used for methylene blue (MB) adsorption via numerous batch experimental parameters: initial concentrations of MB dye (25 – 250 mg/L), contact time (0.5 – 24 hours) and initial pH (2 – 12). Dye adsorption isotherms were also assessed at three temperatures where the maximum adsorption capacity (381.6 mg/g) was found at 30 C. The adsorption equilibrium data were best suited to the non-linear form of the Freundlich isotherm model. Additionally, non-linear pseudo-second-order kinetic model was better ﬁ tted with the experimental value as well. Steady motion of solute particles from the boundary layer to the BNH-AC ’ s surface was the possible reaction dynamics concerning MB adsorption. Thermodynamic study revealed that the adsorption process was spontaneous and exothermic in nature. Saline water emerged as an ef ﬁ cient eluent for the desorption of adsorbed dye on AC. Therefore, the BNH-AC is a very promising and cost-effective adsorbent for MB dye treatment and has high adsorption capacity.

may involve eutrophication, acceleration of genotoxicity and microtoxicity (Khan & Malik ). Given the urgency to save the environment, certain methods such as precipitation (Cao et al. ), coagulation-flocculation (Kuppusamy et al. ), membrane filtration (Pereira et al. ) and adsorption (Ahmad et al. ; Kafshgari et al. ) have been implemented to remove dye. The adsorption method is considerably effective due to the feasibility and applicability of low-priced sorbents (Oyekanmi et al. ). It has great potential in significantly reducing such environmental problems, increasing productivity and helping to remediate the environment (Hoque & Clarke ). Several low-cost adsorbents, such as agricultural wastes, industrial solid wastes, biomass, clays minerals and zeolites, are usually utilized for MB removal (Rafatullah et al. ).
Activated carbon (AC) acts as an environmentally favorable and traditional adsorbent with a recognizable specific area, huge sorption capacity and various surface functional groups (Islam et al. a; Kaykhaii et al. ). Nonetheless, the high manufacturing cost of AC impedes its widespread use (Rafatullah et al. ). In this context, native, inexpensive and available waste materials may be used to produce AC, which could help to advance the technology. Numerous studies have already been conducted to produce AC from waste fibrous biomass (flux) (Illingworth et al. ), palm bark (Melliti et al. ), pine cone, walnut shell, and hazelnut shell (Kaya & Uzun ), sewage sludge (Chen et al. ), river sediment (Sang et al. ), Butea monosperma leaf (Das et al. ), palm kernel shell (Liew et al. ), rice straw (Sangon et al. ), Artocarpus integer peels (Selvaraju & Bakar ), waste biomass (Zubrik et al. ), coffee grounds (Rattanapan et al. ), oil palm waste (Rashidi & Yusup ), banana peels (Ingole et al. ), and many more. However, some biomasses are yet to be explored in terms of identifying their adsorption capacity when converted to AC material.
Betel nut (Areca catechu) is an evergreen, salinity-resistant tree and widely found in coastal areas of Bangladesh. Betel nut husks (BNH) comprise the unusable extraction and biodegradable fiber found in the betel nut fruits a chewing fruit of the Arecaceae family (Sultana et al. ). It is also grown in many tropical and sub-tropical areas like India, China, Indonesia, Malaysia, Myanmar and Nepal (Yusriah & Sapuan ). BNH comprise betel nut fruits, which contain around 60-80% of total volume and weight, and a single fruit generates approximately 2.50-2.75 g BNH (Hassan et al. ). It consists of α-cellulose, hemicellulose, lignin, pectin and proto-pectin, ash, and other materials of which the proportions of hemicellulose and lignin vary with its maturity (Hassan et al. ; Yusriah & Sapuan ). At the immature stage, both the husk and nut part are soft, where the outer layer of the fruit looks green (Yusriah et al. ). The juicier liquid husk becomes spongier when it ripens and the green color changes to brown and the fibers become coarser with maturation (Yusriah et al. ). With these physical and chemical properties, BNH are recognized as value-added products like composite materials, cushion, hardboards, and non-woven fabrics and housing insulation materials, although there are many unmanaged BNH which generate bad odor and landfill complications after harvesting (Borah & Dutta ). As a consequence, agro-waste like BNH requires better management with an emphasis on the waste-toresource concept. Therefore, if pyrolysis of the BNH increases its adsorption capacity of MB, it will efficiently perform as a recycled product.
The adsorption capacities and textural properties along with its functional group of AC strongly depend on the mode of preparation, i.e. starting material, activation method, type of activator, and preparation conditions ( Jawad et al. ; Rashid et al. ). AC can be generated by two activation methods: physical and chemical. Chemical activators can abate the activation temperature and time required compared to the physical activation method ( Jawad et al. ). Both carbonization and activation are performed synchronously in this way in order to save cost and energy because low activation temperature is required to produce upgraded porous materials as well as high yield (Islam et al. a; Jawad et al. ). NaOH is widely used as an effective chemical activator in preference to all other alkaline activating agents due to its availability and capacity to produce numerous pores on the AC's surface area (Heidarinejad et al. ).
There are many common natural feedstocks (especially, lignocellulosic materials) utilized as AC for removal of the cationic dye (MB). In the present study, we focused on inexpensive agro-waste: BNH. Dey et al. () figured out the kinetic and isotherm characteristics of BNH over MB adsorption. However, for the first time, BNH, a native and very available agro-waste, was used to prepare BNH AC by using widely used effective chemical (NaOH) activation. Most importantly, the AC derived from BNH might attenuate the high production cost of commercial AC. The efficiency of obtained materials was assessed by how much MB was removed. Consequently, the main objective of this analysis was to produce a feasible and low-cost BNH-AC by NaOH activation that is also an environmentally friendly as well as economically viable adsorbent for the dispelling of model cationic MB dye. To compute the adsorption capacity, removal percentage, adsorption kinetics, isotherms, thermodynamics over MB solution as adsorbate were the major issues investigated here. To fit adsorption kinetic data, pseudo-first-order, pseudo-second-order kinetic models, intraparticle diffusion and film diffusion were determined. For analyzing the equilibrium data, the Langmuir and Freundlich isotherm models have been selected.

Materials
BNH was collected from a local vendor (Gollamari, Khulna, Bangladesh) and washed to remove all undesirable materials. Then the moist BNH was dried in an oven at 105 C until a constant weight was achieved. The dried BNH was ground in an electrical grinder to obtain the stock raw materials with a 2 mm particle size for BNH-AC preparation. MB dye, having a molecular weight 319.85 g/mol, chemical formula C 16 H 18 ClN 3 S.H 2 O, and solubility 40 g/L, was recognized as adsorbate and prepared 500 mg/L of stock solution. All chemicals used in this study were laboratory grade and commercially available.

Preparation of activated carbon
AC was obtained through pyrolysis and activation of raw BNH. The activation process was conducted by NaOH pellets (98% pure). Prepared BNH was mixed with powdered NaOH at three (1:1, 1:2 and 1:3) ratios (BNH:NaOH, w/w%) for 24 h at room temperature. The mixtures were then put into a crucible and placed into an air tight stainless-steel vertical tubular reactor. The temperature was 500 C with a heating rate of 20 C/min for 1 h. N 2 gas was simultaneously passed through the whole system with a constant flow rate. The reactor was allowed to cool at ambient temperature after ending its dwelling time. Then the activated chars were washed continuously utilizing deionized water till the pH of the products reached 6-7. The approximately neutral products were dried at 105 C. After that, a preliminary study was carried out where the dye removal percentages were 56.93% (1:1), 62.08% (1:2) and 70.04% (1:3). So the 1:3 ratio was stored as stock absorbent since it had the highest removal percentage. This finalized product is referred to as BNH-AC.

Characterization of activated carbon
The Brunauer-Emmett-Teller (BET) surface area of BNH-AC was obtained from the N 2 adsorption-desorption isotherm at 77 K with a volumetric sorption analyzer (Micromeritics, Model ASAP 2060, USA). On the other hand, the surface morphology of the samples was observed through the scanning electron microscope (SEM) (Jeol, JSM-7610F). Surface organic structural groups on raw BNH and BNH-AC were studied by Fourier transform infrared (FTIR) spectra using Perkin-Elmer spectrometer (Model 2000, USA) in the 4,000-500 cm À1 range before and after adsorption of MB dye.

Determination of lignocellulosic and physical characteristics
The lignocellulosic content (cellulose, hemicellulose, lignin) was determined from extractive-free BNH. After proceeding with solvent extraction at 60 C to determine the hemicellulose content, 1 g of extractive-free dried (at 110 C) biomass was added to 150 mL of NaOH solution (20 g/L) and boiled for 3.5 h with distilled water. The residue was filtered, rinsed to remove Na, dried and finally weighted. The difference before and after treatment was the amount of hemicellulose. Lignin was calculated by the Klason method (Theander et al. ) where 15 mL of H 2 SO 4 (72%) was added to 1 g of extractive-free dried biomass and heated for 2 h. The mixture was diluted up to 4% H 2 SO 4 concentration and boiled for 4 h. Then the residue was filtered, washed, dried and weighted. The variation between before and after experiment revealed the lignin content. The amount of cellulose content was analyzed from the weight difference of extractives, hemicellulose and lignin.
A proximate analysis was carried out in both cases of raw BNH and BNH-AC. A measuring cylinder of 10 mL size was washed, dried and weighted. After filling the cylinder with both BNH and BNH-AC, it was further weighed. Likewise, for the determination of moisture content, at 105 C, 1 g of both samples were oven dried for 24 h and weighed. The bulk density and moisture content of BNH and BNH-AC were calculated by using the following equations: where, M f demonstrates the weight of ground BNH and BNH-AC filled container and M c is the weight of the container, and V c is the volume of the container. M 1 and M 2 are the initial and oven-dried weight of the samples, respectively. Ash content (%), volatile matter (%) and fixed carbon (% yield) of BNH and BNH-AC were obtained with a muffle furnace (HYSC, Model: MF -05, South Korea) and were calculated by following the mathematical expressions below: where, W c is the crucible weight (g), W rc is the combined weight of crucible and samples (g): both BNH and BNH-AC, W ac is the total weight of crucible and ash (g) that was determined after the experiment and W s is the samples weight (g). Temperature was fixed at 550 C for 4 h in case of determining the ash content, whereas analyzing the volatile matter was conducted at 900 C for 7 min. The amount of MB dye adsorbed by BNH-AC, q e , (mg/g), kinetics q t (mg/g) and removal percentage R (%) were calculated employing the following equations: where C 0 is the initial MB concentration (mg/L), C e and C t are the concentration of MB (mg/L) at the equilibrium and time t, respectively, V is the volume of MB solution (L), and W is the mass of BNH-AC (g). By carrying out the fit of the experimental values to the non-linear equations, all the kinetics and isotherms models were analyzed using a scientific data analysis. The pH point of zero charge (pH pzc ) was carried out by pH drift method. Patawat et al. () recently determined the pH pzc of AC derived from Dipterocarpus alatus fruit by adopting this method. The pH of a series of 0.01 M NaCl was adjusted at a different initial pH value ranging from 2 to 12; 0.1 M H 2 SO 4 and 0.1 M NaOH were used for the adjustment. Then, 0.20 g of BNH-AC was added to each adjusted solution (50 mL). The dispersions were shaken for 48 h at 40 C like the above stated procedure for pH, and the final pH of the solutions (pH f ) was determined by a pH meter. The pH pzc is the point where the curve pH f vs. pH i intersects the line pH f ¼ pH i . Its main purpose is to identify the anionic or cationic surface charge of the adsorbent.

Sorption isotherms and kinetics
The relationship between the amount of MB adsorbed over per unit mass of the sorbent and the different concentrations of dye stuff can be resolved by isotherm analysis. Along with this, it is easy to understand the adsorption performance of BNH-AC and adsorption dynamics of MB with time. Likewise, the kinetic studies can work out the kinetic mechanisms of the adsorption process. Both studies were carried out to assess the equilibrium experimental parameter states in the context of the maximum removal of dye. Batch adsorption experiments were done in 250 mL Erlenmeyer flasks holding 200 mL of MB solution (concentrations: 25 to 250 mg/L). The pH of the solutions was 6.4, close to the neutral value. Then, 0.2 g of adsorbent material (BNH-AC) was added to the flasks, which were placed in a temperature controlled horizontal shaker with agitating speed of 150 rpm. The temperature was 40 C and after reaching the equilibrium time, the samples were extracted from the flasks and supernatant solutions were measured by using double beam UV-vis spectrophotometer at a wavelength of 668 nm. The amount of the uptake of MB by BNH-AC at optimum time was found by mathematical Equation (7). From the experimental data, the adsorption mechanism was recognized by two diffusion models: intraparticle and film. Moreover, a thermodynamic study was conducted using the same procedure beyond the temperature ranges of 30 C to 50 C. It is, basically, used to determine the effect of temperature on the nature and suitability of MB adsorption onto BNH-AC.

Physical and chemical characteristics of adsorbent
The ratio of lignin, cellulose and hemicellulose of raw BNH and physical constituents of both BNH and BNH-AC are computed in Table 1. Bulk density of raw material was 0.245 g/mL. Table 1 reveals the compositional properties of BNH, including the maximum ratio of cellulose (46%), which is virtually identical to a previous study of areca nut husk (Chandra et al. ). AC's adsorptive nature may be concerned with ash content, because it exposes not only an inhibiting but also a stimulating characteristic in the case of adsorption of different organic chemicals (Wang et al. ). Moreover, it is directly related with AC pore structure (Anisuzzaman et al. ). In the current study, volatile matter of BNH-AC has a lower value compared to the raw BNH. Conversely, the ash content shows a higher ratio. Differently, the carbon content discloses the higher value in the case of BNH-AC (Table 1). Thus, this material is a good precursor.

Characterization of activated carbon
The textural properties of BNH-AC are determined by the BET surface area analysis. The surface area and total pore volume were found to be 81.73 m 2 /g and 0.722 cm 3 /g, respectively, and furthermore are larger than pyrolyzed areca husk (Subramani et al. ). The mean pore diameter was 3.533 nm. According to International Union of Pure and Applied Chemistry (IUPAC) classification, this pore diameter indicates the mesoporous properties (2.0 nm < d < 50 nm) (IUPAC ; Sing et al. ) and according to the new updated isotherm classification by IUPAC (Thommes et al. ), the isotherm is deemed to be a type IVa isotherm, one that indicates capillary condensation accompanied by hysteresis. Furthermore, type H 4 hysteresis loop was observed ( Figure 1) which is very common in mesoporous carbon materials. Figure 2 reveals the SEM micrographs of BNH, BNH-AC and the state of BNH-AC after adsorption at 3,000 × magnification. Due to NaOH activator the pore areas are developed in Figure 2 shows a plane surface area. The activators' dehydration effect is responsible for the improved porosity. NaOH splits the C-O-C and C-O bonds of materials and is converted to Na (metallic), H 2 (gas) and sodium carbonate (Na 2 CO 3 ). The equation is expressed as: Due to high temperature, in the existing N 2 gas, Na 2 CO 3 is degraded into Na, CO and CO 2 . Consequently, an ample number of voids has been constituted after trapping Na molecules in them (Hassan & Youssef ).

FTIR spectral analysis
FTIR spectra of raw BNH, BNH-AC and BNH-AC after adsorption are shown in Figure 3(a)-3(c), respectively. FTIR analysis determines the existing important functional groups like alkanes, esters, aromatics, ketones, alcohols and other oxygen-containing functional groups because of their concentrated lignocellulosic properties (Table 1). Adding or removing any functional groups reveals the changes occurring in the adsorption process after modifying the raw BNH to BNH-AC, which is visualized by band points in spectra. Two main peak regions are focused on the spectra of raw BNH: at low wavelength (600 À 1,800 cm À1 ) and high wavelength (2,300 À 3,100 cm À1 ). The peak points of 3,681.98 cm À1 , 3,750.11 cm À1 and 3,756.08 cm À1 in the spectra indicate the O À H stretching of alcohols and phenols, which are not clearly visualized in BNH-AC and after adsorption spectra. There are two main peaks observed on the surface of BNH and these present the alkanes (C-H stretching) and nitriles  (C≡N stretching) at 2,866.06 cm À1 and 2,275.03 cm À1 , respectively. After activation, these two peaks are shown to totally disappear from the spectrum after MB adsorption. However, a new peak (2,500 cm À1 ) is observed in the BNH-AC after adsorption band corresponding to the O-H stretching of carboxylic acid (Islam et al. ).
A large number of peak points at 1,756.9 cm À1 and 1,720.76 cm À1 , 1,550.56 cm À1 , 1,335.37 cm À1 , 1,227.67 cm À1 , 1,029.99 cm À1 correspond to the C¼O stretch (carboxylic acids, aldehydes), C-C stretch (aromatic), C-N stretch (aromatic amines), C-H (alkanes) and C-O stretch (alcohols, carboxylic acids, esters, ethers), respectively (Figure 3(a)).  Several functional groups remain constant after preparing AC, while 1,651.07 cm À1 and 856.98 cm À1 are noted as new peaks that indicate -C¼Cstretch (alkanes) and ¼C-H bend (alkanes), respectively (Figure 3(b)). The dissipating of a large peak 1,029.99 cm À1 ascribed to the C-O stretch of alcohols, carboxylic acids, esters, ethers is noteworthy. In Figure 3(c), the immense band (600 À 1,800 cm À1 ) mostly vanishes after MB adsorption except for some similar smaller peaks at 1,315.45 cm À1 , 1,029.99 cm À1 and 875.68 cm À1 that represent the C-N stretch (aromatic amines), C-O stretch (alcohols, carboxylic acids, esters, ethers) and ¼C-H bend (alkanes), respectively. The departure authenticates that the C¼O and C-H are the active functional groups; perhaps these groups are engaged in the cationic MB dye and BNH-AC reaction. This type of phenomenon is confirmed by Islam et al. (b).

Adsorption studies
The MB dye's adsorption performance onto BNH-AC was conducted to determine the adsorption capacity at three different temperatures (30, 40 and 50 C) and initial concentrations (25,50,100,150,200 and 250 mg/L). In Figure 4, adsorption capacity (q t ) of BNH-AC rises rapidly to 5 h in terms of higher concentrations occupying the void areas, whereas in low dye concentrations (25 and 50 mg/L), it rises slowly. The value of q t increases when the dye concentration also increases. Overall, from all temperature graphs, we can conclude that the equilibrium capacity was obtained at 6 h, which supports other studies. According to Islam et al. (b), the MB adsorption capacity was attained at 8 h for rattan AC by NaOH activation. Available surface sites, is the main driving force for the dye removal, which remains in a large number at initial stage and is occupied with time. After arriving at equilibrium, this ratio declines, but due to the repulsion force between MB and the BNH-AC surface, some pore sites remain unfilled. Hence, the study lasted 24 h; it was valid and concurred with the findings of Ingole et al.

Effect of pH
The pH pzc exhibits a specific pH point where the net surface charge of adsorbent corresponds to zero, and it shows the possible mechanism for electrostatic interaction between adsorbent and adsorbate (Islam et al. ). The pH pzc value of BNH-AC was obtained at 7.87 in this study ( Figure 5(a)). This value indicates that the adsorbent surface area remains positively charged below this point, enhancing the conducive situation for anions. However, according to Islam et al. (b), the electrostatic interaction is likely to be escalated above this point.
The effect of pH on the adsorption process was observed at 40 C. Basically, the main cause of this was due to the total adsorption process being in line with the Freundlich isotherm model, and the root mean square error (RMSE) at 40 C was low compared to the other two temperatures, as shown in Table 2. Moreover, the treatments of industrial dye effluent are proceeded with at high temperature (Zou et al. ). The pH effect is illustrated by Figure 5(b), where pH ranged from 2 to 12 and the initial concentration of MB was 100 mg/L. From the figure, we can see that at low pH (pH ¼ 2), the percentage removed remains low (77.59%) but increases simultaneously when the value of pH increases. The main logic behind this is the oxygen-containing functional groups (-COOH, -OH) of BNH-AC. In acidic pH, the oxygen-containing functional groups are protonated to the cationic form, like -COOH 2 þ , -OH 2 þ , which releases the number of negative charges. Thus, the surface sites do not favor binding the cationic dye and resulted in an electrostatic repulsion causing poor removal (%). Conversely, such functional groups are deprotonated to anionic form (-COO À , -O À ) with increasing the basic condition. Thus, it creates electrostatic force with the cationic dye and makes the surface more favorable to adsorption (Konicki et al. ).

Adsorption isotherms
The interaction of adsorbate with adsorbents and the equilibrium distributions of adsorbate molecules between the solid and liquid phases are known as adsorption isotherms (Islam et al. a). To determine the further adsorption equilibrium, two famous isotherm equations, Langmuir (Langmuir ) and Freundlich (Freundlich ), were applied to fit the experimental equilibrium isotherm data of MB adsorption on BNH-AC. The single layer adsorption process on consistent adsorption sites is narrated by the Langmuir isotherm model, while the multilayer adsorption process on different adsorption sites is described by the Freundlich isotherm model.  These two models are mathematically expressed as: Langmuir isotherm: q e ¼ q m K a C e 1 þ K a C e (11) Linear form: where, q e (mg/g) represents the adsorption capacity at equilibrium, C e (mg/L) is the equilibrium dye concentration, and q m is the maximum adsorption capacity. K a (1/mg) and K L (L/mg) stand for the Langmuir constant and the constant of free energy of adsorption process, respectively: Linear form: ln q e ¼ ln K F þ 1 n ln C e (14) where, K F (mg/g (L/mg) 1/n ) and n are Freundlich constants representing the adsorption capacity and the intensity of adsorption, respectively. By plotting q e and C e , the adsorption isotherm was attained and is shown in Figure 6(a). Furthermore, out of these two models, the best suited isotherm model was recognized by the R 2 value and also authenticated by the RMSE value. Both of these are numerically expressed as: Here, q e.meas (mg/g) and q e.cal (mg/g) are the experimental and predicted value of adsorption capacity at equilibrium. The number of observations is presented by 'n'. Table 2 exhibits the non-linear calculated value of two isotherm models. The Freundlich isotherm model shows the highest R 2 value and lower RMSE compared to the Langmuir model. That means the adsorption process can be better described by the Freundlich model where adsorption appeared on heterogeneous sites and revealed high adsorption capacities. Moreover, the non-linear calculated q m was found to be 381.6, 339.8 and 235.2 mg/g at 30, 40 and 50 C, respectively, which is higher than any other materials, like cotton stalk activated by KOH and K 2 CO 3 (q m ¼ 294.12 and 285.71 mg/g, respectively) (Deng et al. ), karanj fruit hull activated by KOH (q m ¼ 239.40 mg/g) (Islam et al. a), rattan palm tree activated by NaOH (359 mg/g) (Islam et al. b) and so on. We can deduce that BNH as a precursor can be feasibly employed for manufacturing an efficacious and productive AC.
The values of the Freundlich constant 1/n were less than 1, and this indicates a supportive adsorption process. If these values are near zero, the adsorption heterogenicity will be higher. This finding coincides with various studies where the Freundlich model fitted better than the Langmuir model. Examples are NaOH-activated palm date seed (Islam et al. ), ion beam-modified palygorskite (Zhang et al. ), amino-modified mesostructured silica nanoparticles (Karim et al. ) and many more. Therefore, some heterogenicity is placed on the surface sites of BNH-AC, which is responsible for MB adsorption.

Adsorption kinetics
Kinetic studies are required to understand the adsorption mechanism as well as the rate of adsorption. Pseudo-firstorder and pseudo-second order kinetic models have been devised to describe the potential adsorption dynamics. Pseudo-first order kinetic model was suggested by Lagergren and Svenska (Lagergren ), who mathematically expressed the non-linear form as: q t and q e of this equation are adsorption capacity of BNH-AC at time, t (min) and at equilibrium, respectively, k 1 is rate constant of the adsorption (1/min) of the pseudo-first-order kinetic model. The non-linear pseudosecond-order equation (Ho & McKay ) is: where, k 2 (g/mg⋅min) is the rate constant of pseudo-secondorder adsorption. The best-fitted model between these two was obtained by calculating the coefficient of determination (R 2 ) and two non-linear error functions: normalized standard deviation (NSD) and average relative error (ARE). These are articulated as follows: R 2 ¼ 1 À P n n¼1 (q e:exp À q e:mod ) 2 P n n¼1 (q e:exp À q e:mod ) 2 NSD ¼ 100 × ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi 1 n À 1 X n i¼1 q t:exp À q t:mod q t:exp where, q t.exp (mg/g) is the experimental adsorption capacity value at time t and q t.mod (mg/g) represents the model-anticipated value of adsorbed amount MB dye at time t. The higher value of R 2 and lower value of NSD and ARE demonstrate the best-suited model. The non-linear plot of the pseudo-first-and second-order fitting and experimental kinetic values is presented in Figure 6(b). This figure displays the obtained equilibrium at 6 h, which means the dye adsorbing rate was equal to the dye desorbing rate after achieving its saturation level. The analyzed kinetic parameters using the non-linear method are summarized in Table 3. Referring to the different dye concentrations, the pseudo-second-order model shows high R 2 values with low NSD and ARE values compared to pseudo-first-order kinetics. Even the modeled value of q e (mg/g) is best fitted with experimental q e (mg/g) value in pseudo-second-order kinetics. All the outcomes cited above prove that the adsorption of MB dye on to BNH-AC follows the pseudo-second-order model, so chemisorption is supported and not physisorption. This study agrees with the findings of other analyses, such as adsorption on cabbage waste powder (Wekoye et al. ) and white sugar AC (Xiao et al. ). Table 4 represents a comparative study of several previous precursors with the present study in terms of MB adsorption. Table 4 disclosed that the q m of the Mangoosteen peel is 1,193 mg/g, which represents a higher value due to large surface area (1,622 m 2 /g) and mesoporous characteristics (Nasrullah et al. ). However, the q m of BNH-AC reveals higher value compared with other precursors (Table 4).

Adsorption mechanism
Kinetic studies cannot describe the mechanism of the adsorption process; they can merely interpret the solute- where, k id and C are the intraparticle diffusion rate constant (mg/g. min 1/2 ) and intercept related to the boundary layer effect (mg/g), respectively. In Figure 7( versus t 1/2 , the gradient and intercept values are the expression of k id and C. From the graphical representation, we can observe that the pore and intraparticle diffusion increased with an increase in the initial MB concentrations (25 mg/L-250 mg/L). The Boyd model (Boyd et al. ) could differentiate intraparticle and film diffusion and is linearly expressed as: where, k fd is the liquid film diffusion constant obtained from the linear plotting of ln (1 À q t /q e ) vs. t. Table 5 illustrates the value of intraparticle as well as film diffusion for adsorption of MB dye on to BNH-AC for different initial concentrations. As the value of C defines the boundary layer effect, the higher values confirm the higher effect. If the plotted lines pass through the origin, then it can be said that intraparticle diffusion controls the adsorption process. Nevertheless, in most cases the plotting lines do not pass through the origin, which corresponds to different mechanisms rather than intraparticle diffusion controlling the adsorption process (Figure 7(a)). Hence, film diffusion, the slow movement of solutes from the boundary layer to the adsorbent's surface, might be responsible for controlling MB adsorption. Even the R 2 values of the intraparticle diffusion model are lower than the Boyd model (Table 5) and this coincides with the findings. Similar studies were done by Njoku et al. () for carbofuran insecticide by coconut frond AC.

Adsorption thermodynamics
The thermodynamic parameters for adsorption consist of Gibbs free energy change (G ), enthalpy change (H ), and entropy change (S ), which can be calculated by: where, R is the universal gas constant (8.314 J/mol K) and T is the absolute temperature (K), respectively. K d demonstrates the adsorption distribution constant (1/mg), which can be expressed as: where, C Ae denotes the amount of MB dye adsorbed at equilibrium and C e is the equilibrium concentrations of the MB dye (mg/L). Table 6 computes the thermodynamic parameters, where ΔG implies negative values. It supports the remarkable spontaneity of the adsorption process and significant amount of energy being generated (Njoku et al. ). As well, these values decrease when the temperature rises, demonstrating that the adsorption of MB on to BNH-AC is thermodynamically conducive at high temperature. The ΔH value also provides the negative sign indicating the exothermic nature of the adsorption process. The decreased attraction between MB molecules and the active surface sites on the BNH-AC with rising temperature could also explain the findings, which eventually decreased the adsorption capacity. The positive sign of ΔS indicates the randomness of the adsorption system, which provides superior attraction between adsorbate molecules and adsorbent surface sites. Islam et al. () reported an analogous study for palm date seed AC for ΔG , ΔH and ΔS . Basically, such kinds of results are a common outcome of physical adsorption, something that can take place through electrostatic interactions.

Desorption study
Desorption performance refers to clarifying not only the overall sorption mechanism but also recovery of the used adsorbent material (Yu et al. ). It is essential to know the desorption efficiency of separate media, as this can suggest the suitable eluent for regeneration studies. In the present work, three media were used for the MB desorption study: methanol (CH 3 OH), sodium chloride (0.1 mol/dm 3 ) in methanol, and sodium chloride (0.1 mol/dm 3 ) in water, as recommended by Moreira et al. (). The desorption study was carried out in a batch experiment mode at adsorption equilibrium state (initial concentration ¼ 100 mg/L; temperature ¼ 30 C; time ¼ 6 h). According to Figure 7(b), saline water (NaCl þ H 2 O) is the most efficient eluent of the three solutions, showing 86.07% desorption efficiency. On the other hand, a combination of NaCl and methanol shows a noticeable desorption potential (76.27%) compared to only methanol. Previously, NaCl solution has proven to be suitable media for desorption of cationic dye (malachite green) where a high concentration of NaCl solution enhanced dye desorption due to its ionic strength (Jiang et al. ). This high ionic strength splits the electrostatic bindings between positive MB dye and adsorbent, which stimulates the MB dye removal. This finding is due to using saline water, and BNH-AC can be handily regenerated for further dye removal application.

CONCLUSION
The potential of BNH as a new and economically efficient raw material for AC was explored in the present study. Various authenticated and well-known numerical models and settings were adopted to compare the experimental findings. The significant adsorption capacity (q m ) of BNH-AC was found as 381.6, 339.8 and 235.2 mg/g at 30, 40 and 50 C, respectively, by using the non-linear calculation of the conventional isotherm model. Also the Freundlich isotherm model proved to be the best-suited model for the isotherm study due to its high R 2 value. Meanwhile, pseudo-second-order kinetics explained the adsorption kinetic model as revealed in this study. Due to the electrostatic force increasing with a basic pH level, the removal of MB was higher at a pH value of 10 and this remained constant from pH 10 to 12. In accordance with the fitting on the Weber-Morris intraparticle diffusion and Boyd models, solutes' low motion from the boundary layer to the BNH-AC surface was proved, followed by film diffusion rather than intraparticle diffusion. Moreover, the thermodynamic studies indicated that the adsorption process was spontaneous and exothermic. C¼O and C-H were the active functional groups, engaged in the reaction of cationic MB dye and BNH-AC. This was authenticated by FTIR spectral analysis. Finally, the above mentioned findings confirmed that BNH-AC, using NaOH activation, can be an efficiently advanced adsorbent for removing environmental pollutants.