Adaptive neuro-fuzzy interference system modelling for chlorpyrifos removal with walnut shell biochar

2.1 Preparation of walnut shell biochar

Walnut shell (WS) was obtained directly from the trees grown in Central Anatolia. At the first stage, the shells were washed three times with distilled water (Millipore, Q-8, Germany), then dried in an oven (Memmert, UNB 400, Germany) at 105 °C for 24 h. The dry WS was grinded, sieved through a 0.5 mm sieve, and stored in an airtight plastic box until reused. Biochar (BC) samples were derived in a rotary oven (Protherm, RTR 11/100/500, Germany) at the various temperatures of 450, 600, 750, and 900 °C at N₂ atmosphere (1 L/min) for 1 h by pyrolysis. The heating program was 100 °C (10 min), 200 °C (10 min), 300 °C (10 min), 450 °C (60 min), 600 °C (60 min), 750 °C (60 min), and 900 °C (60 min) with 10 °C/min heating rate. The obtained samples were sieved again from 0.25 mm sieve and nomenclature as WSBC@450, WSBC@600, WSBC@750, and WSBC@900.

2.2 Characterizations of biochar samples

The characterization of the biochar samples was performed by various analyses. The point of zero charges (pzc) described as the pH at which the net charge of the total particle surface (i.e. absorbent's surface) is equal to zero, which is typically obtained by acid-base titrations of colloidal dispersions while monitoring the pH (Hach, HQ411D, Germany) of the suspension. The particle sizes of the samples were measured with Cilas-1190 instrument according to Mie Theory. The surface functionality of the biochars was examined by Fourier Transform Infrared Spectrophotometer (FTIR, Thermo Scientific Nicolet iS20, USA) at the 4000–400 1/cm region with a resolution of 1/cm after the averages of 3 scans. Elemental compositions were determined using Scanning Electron Microscopy (SEM) instrumentation (Hitachi, SU 1510, Japan) equipped with the energy-dispersive X-ray (EDX) detector. The BET is an important analysis technique for the measurement of the specific surface area of materials. Surface areas were established by the Brunauer Emmett Teller (BET, Quantachrome, Quadrasorb Evo 4, USA) multilayer adsorption method using micromeritics. The textural characterization of WSBC was carried out by N₂ adsorption–desorption isotherms at 77 K using a surface area analyzer Quantachrome. The surface area, SBET, was determined by N₂ isotherms using the Brunauer–Emmett–Teller equation (BET). Before adsorption measurement, the samples were outgassed at 150 °C for 3 h.

2.3 Adsorption experiment and chlorpyrifos analysis

Adsorption tests were conducted for each biochar sample at room temperature. 250 mL of CP (C₉H₁₁Cl₃NO₃PS, Sigma-Aldrich) solution with an initial concentration of 100–1500 µg/L was added to 500 mL Erlenmeyer flasks and water completed to 500 mL with the adjustments of 10–500 mg/L of each biochar materials. The prepared samples were shaken on a rotating shaker (Zhicheng, ZHWY-221D, Germany) at a constant contact time of 3 h. Control samples containing µg CP/L without biochar were also included. All the tests were triplicated. After equilibration, samples were centrifuged (Beckman Coulter, Allegra X-12, USA) at 3200 rpm for 10 min. and then filtered by using a 0.45 µm syringe filter. The supernatant from each sample was collected and analyzed for quantification of the concentration of CP using a UV–visible spectrophotometer (Shimadzu 1280 UV–VIS, Japan) according to the maximum absorbance at 290.5 nm of wavelength (y = 0.0188x + 0.0033, R² = 0.9992) (Aswathi et al., 2019; Lockridge et al., 2019). A calibration curve was drawn by diluting a stock solution of CP (5000 µg/L) in isopropanol: DI water (30:70).

First, in adsorption experiments, the most suitable biochar type was determined among the biochar samples that were pyrolyzed at different temperatures, and then, biochar dosage (10, 250, 500 µg/L), CP concentration (100, 250, 1000, and 1500 µg/L) pH (3, 5, 7 and 10) and operating times (0–300 min.) for optimal biochar were examined. It was applied to other biochar samples under specified optimum conditions. In this study, three different kinetic models (pseudo-first-order, pseudo-second-order. and intra-particle diffusion) and two isotherms models (Langmuir and Freundlich) were used to find the adsorption capacity, rate, and mechanism. The amount of CP adsorbed by the biochar samples (q_e) was determined by Eq. (1):

2.4 Development of the adaptive neuro-fuzzy interference system model using experimental variables

Fuzzy systems that establish a relationship between input variables and output variables using fuzzy sets can model even variables that do not have a numerical equivalent with the help of expectations (Alves and Aguiar, 2021). In this work, the Sugeno type model with optimized parameters (ANFIS) was established. The output parameters have the membership functions of the input variables. The Sugeno type ANFIS structure is given in Fig. 1.

Close

Fig. 1

A Sugeno type adaptive neuro-fuzzy interference system model structure with 5 inputs, 1 output, and 243 rules.

Export to PPT

In Layer 1, input signals from each node are transferred to other layers. The output for each node i is defined as Eq. (2):

Layer 2 is called the blur layer. Jang's ANFIS model uses the Bell activation function, which is generalized as a membership function, in separating input values into fuzzy sets. The output of each node consists of membership degrees that depend on the input values and the membership function used. Membership degrees obtained from Layer 2 are shown as ${\mu }_{\mathit{Ai}}\left(x\right)$ . Layer 3 is the rule layer. Each node in this layer represents the number and number of rules established according to Sugeno's fuzzy logic inference system. Output µ_i of each rule node is the product of membership degrees from Layer 2. The derivation of the μ_i values is obtained as (j = 1,2) and (i = 1,.,n) as in Eq. (3).

Layer 4 is the normalization layer. Each node in this layer accepts all nodes from the rule layer as the input value and calculates the normalized firing level of each rule. The normalized firing level $\overline{\omega }$ is calculated according to Eq. (4).

Layer 5 is the clarification layer. The weighted result values of a given rule are calculated at each node in the clarification layer. The output value of node i in the Layer 5 is calculated according to Eq. (5).

Layer 6 is the total layer. This layer has only one node and is labeled with ∑. Here, by adding the output value of each node in the 4th layer, the actual output value of the ANFIS system is obtained. The calculation of y, which is the output value of the system, is according to Eq. (6). The details of the equations and other things of ANFIS structure can be found in Alver et al. (Alver et al., 2020).

The learning algorithm of ANFIS based on the Sugeno System is a hybrid learning algorithm that consists of using the least-squares method and the backpropagation learning algorithm together. This learning algorithm is based on error backpropagation. A step in the learning process has two parts; In the first part, input samples are produced and the predecessor parameters are accepted as constant and the best result parameters are determined with the least mean square method. In the second part, the input samples are regenerated and the predecessor parameters are changed by the gradient descent method, assuming the final parameters constant. Then, this process is repeated. In this study, 143 data obtained from laboratory-scale experiments were modelled using ANFIS, and CP removal efficiency has been estimated based on parameters such as adsorbent dosage, adsorbate concentration, initial pH, contact time, and the number of recycling, which affect adsorption efficiency. Training, control, and testing data sets were created by randomly selecting data at different percentages. The correlation coefficient (R²) and the root of mean square error (RMSE) value given in Eqs. (7) and (8) were used as the performance evaluation criteria of the ANFIS model.

3.1 Characterization of biochar samples

In order to understand the connection of the adsorption mechanism with pH, it was necessary to know the zero charge point of the adsorbents. The particle sizes WSBC@900 are determined by pH_PZC measurements at the pH range from 2 to 10 shown in Fig. 2. The point of zero charges of biochar was 8.35. The surface charge of the biochar depends on the functional groups it contains, such as aliphatic, aromatic, and carboxyl groups. Carboxyl groups in the structure are the main ion-exchange sites for the adsorption or desorption of pesticides. The results of FTIR analysis performed to determine the functional groups on the biochar surface and to observe their changes after adsorption are given in Fig. 3. The —OH stretching peak is seen at 3627 1/cm as broadband. The peak at 1566 1/cm is belonged to C⚌C stretching while the C—H bending peaks of the biochar are seen at the 1200–1465 1/cm region. The peaks at 1200–1075 1/cm are causing possible Si—O stretching and the C⚌C bending peak at 740 1/cm is observed. The intensities of the FTIR peaks are becoming weaker when the pyrolysis temperature increase as a result of the probable removing of volatile molecules which are rich in functional groups. The FTIR analysis was performed after the adsorption for WSBC@900 as well, which showed the best adsorption characteristics. The weak peaks have become stronger and resembling CP peaks since CP has adsorbed on the surface pH value at the point of zero charge.

Close

Fig. 2

pH at the point of zero charge value of biochar pyrolyzed at 900 °C.

Export to PPT

Close

Fig. 3

Fourier transform infrared spectrophotometer analysis of biochars and chlorpyrifos.

Export to PPT

Another important parameter in biochar characterization and adsorption is the volume of macro-, meso- and micro-sized pores of the biochar to be used as a sorbent. The BET analysis results are summarized in Table 1. The pristine WS has a surface area of 2.77 m²/g. The surface area of WSBC@450 is the smallest as 12.9 m²/g due to the possible insufficient removal of volatiles from the surface. The surface areas increase with an increase in pyrolysis temperature to 353.3 m²/g. Corresponding pore volumes are increasing from 0.0137 cc/g to 0.2015 cc/g with pyrolysis temperature which could be the result of enlargement or coalescence of internal cavities by the further release of volatile compounds (Sizmur et al., 2017).

SEM-EDX analysis is an effective way to describe and examine the changes in the surfaces and structures of biochar and its properties. Fig. 4 shows the SEM and EDX analysis of the raw WSBC@900 sample and the CP sorbed on the WSBC@900. The rate of adsorption of Cl, which was not included in the elemental analysis of the biochar, but was found in CP due to the percentage increase on the WSBC@900 surface, was proven by EDX analysis.

Close

Fig. 4

Scanning Electron Microscopy and energy-dispersive X-ray analysis of the biochar pyrolyzed at 900 °C and the chlorpyrifos sorbet on the biochar pyrolyzed at 900 °C.

Export to PPT

As seen in Fig. 4, as a result of the heat process applied to the walnut shell, it was observed that the porosity and surface roughness increased, resulting in macro-sized pores on the surface as a result of the carbonization process. The WSBC@900 surface, which appears to be quite rough and porous, appears to be suitable for adsorbing macro and micropollutants. Also, the macropores formed on the surface of the WSBC have partial meso and micropores that open towards the interior, but it is also seen that they are not sufficient. Chemical and physical activation methods can be applied to increase the volume of meso and micropores. It is seen that the pores on the surface of WSBC@900 are filled with CP after adsorption. Besides, 13.12% Cl^- increase on WSBC@900 surface, which is one of the best evidence of CP adsorption, is shown in Fig. 4 with green color. As can be seen, observed in Fig. 4, before the CP removal from the structure, the percentages of C, O, Ca, Na, K, and Mg were 86.62%, 11.50%, 1.46%, 0.006%, 0.24%, and 0.10%, respectively. However, after the removal, the elemental composition of CP changed to: 79.39% for C, 5.61% for O, 0.54% for Ca, 0.74% Na, 0.50 % K, and 13.12% for Cl (see Fig. 4).

3.2 The effect of pyrolysis temperature

The CP removal efficiencies are shown in Fig. 5 at a constant adsorbent concentration of 250 µg/L for all the WSBCs. The removal ratio is increasing with pyrolysis temperature from 450 to 900 °C and according to the surface area. The efficiency is becoming more visible at longer retention times that sorbate concentration could decrease from 1000 till 170 µg/L or q_e values could increase from 0.2 to nearly 3.5 µg/mg. The surface area/porosity could be a leading factor affecting the CP adsorption than surface functionality since qe values are very close at the beginning of the experiments.

Close

Fig. 5

The pyrolysis temperature effect on adsorption (biochar pyrolyzed at 900 °C = 250 µg/L, chlorpyrifos = 1000 µg/L, pH = 7.07, time = 0–300 min.)

Export to PPT

3.3 The effect of biochar dosage

Fig. 6 shows the effect of adsorbent concentration on the adsorption evaluated for WSBC@900 as the best adsorbent among BCs having the highest qe values. The adsorption capacities of WSBC@900 biochar at 10, 250, 500 µg/L at a dose of 1000 µg/L CP were determined at pH 7.07 for 300 min. The q_e values were quantified as 72.87, 3.42, and 2.00 µg/mg, respectively. The CP removal reaches equilibrium after 150 min. for all three adsorbent dosage eventual filling available pores gradually. In the isotherm studies of this study although the q_e value was found to be 4.368 µg/mg in the pseudo-second-order model adsorption reaction (Table 2). It was observed that the max q_e capacity increased to 72.87 µg/mg. Although it was determined that the most efficient sorbent was 10 mg/L WSBC@900 (271.27 µg/L) at the end of operation periods. It was determined that it did not meet the receiving environment limits. Therefore 250 µg/L WSBC@900 sorbent dose was chosen as the optimum dose.

Close

Fig. 6

Adsorbent dosage effect on adsorption (Biochar pyrolysis temperature = 900 °C, chlorpyrifos = 1000 µg/L, pH = 7.07, time = 0–300 min.)

Export to PPT

3.4 The effect of chlorpyrifos concentration

The removal efficiencies were examined at 100, 250, 1000, and 1500 µg/L CP initial concentrations by 300 min of retention time and 250 µg/mg WSBC@900. As shown in Fig. 7, when the initial concentration of CP increased from 100 µg/L to 1000 µg/L, the amount of pesticide adsorbed on WSBC@900 increased from 0.4 µg/mg to 3.42 µg/mg after 300 min. Increasing the pesticide concentration further to 1500 µg/L causes a partial decrease in the adsorbed amount of CP but then reaches 3.06 µg/mg after 150 min. These would be the result of the adsorption limit of the adsorbent as 250 µg/mg which had a restrictive effect on sorption for CP concentrations above 1000 µg/L coherent with literature (Rodríguez-Romero et al., 2020). Therefore, the initial CP dose was decided as 1000 µg/L for further kinetic studies for this work.

Close

Fig. 7

Initial sorbate concentration effect on adsorption (biochar pyrolyzed at 900 °C = 250 µg/L, pH = 7.07, time = 0–300 min.)

Export to PPT

3.5 The effect of pH

The pH is one of the most important parameters affecting adsorption (Wu et al., 2021). The pH of each solution was adjusted to the required value by adding dilute acid and base solutions. The optimum pH on the adsorption efficiency was determined at 250 µg/L WSBC@900 and 1000 µg/L CP, 300 min. residence time, and at various pH values of 3, 5, 7, and 10 (Fig. 8). With the change of solution pH, H⁺ or OH^– ions in the solution are bound to the biochar surface by electrostatic attractions. While the solution pH decreases that the H⁺ concentration increases, the electrostatic interactions become favourable on the negatively charged biochar surface (Fig. 4). Furthermore, hydraulic stability of CP decreases in pH decreases (John and Shaike, 2015) which causes the pushing force between the pesticide and biochar. Thus, the adsorption efficiency decreased at very low pHs. Nonetheless, the adsorption amount increased significantly by changing the pH from 3 to 5, then it varied a little with the continuous increase to pH 7.0. The decrease in the concentration of hydronium ions in the environment instead of the increase in the concentration of negatively charged hydroxyl ions brought about an increase in the adsorption capacity for positively charged CP. The electrostatic attraction between the negatively charged adsorbent surface and the positively charged CP in the basic environment also increased. The q_e values were determined as 1.2, 3.1, 3.4, and 3.3 for pHs of 3, 5, 7, and 10, respectively. pH 7 was examined as the optimum value with the adsorption efficiency as 86%.

Close

Fig. 8

Initial pH effect on adsorption (biochar pyrolyzed at 900 °C = 250 µg/L, chlorpyrifos = 1000 µg/L, time = 0–300 min.)

Export to PPT

3.6 Contact time

Adsorption is a process that continues until a dynamic balance is established between the concentration of solute remaining in solution and the concentration of solute held on the solid surface (Balci and Elçin, 2017). To find the equilibrium time, the concentration of CP remaining in the solution was determined in the samples taken at a certain time interval and the optimum contact time was determined. The data obtained after 300 min of operation under optimum conditions (adsorbent = 250 µg/L, Sorbate = 1000 µg/L, and pH = 7.07) were given in Fig. 6 and Fig. 7. The adsorption capacity values increase as the operating time increases. Experimental results show that the adsorption rate increases more rapidly, especially after an average of 15 min after the start of the process. Initially, the excess active empty areas required for adsorption on the adsorbent surface cause the adsorption to occur faster. However, as the time approaches equilibrium, the interaction between the sorbate and the adsorbent decreases (Changmai et al., 2018). Equilibrium mainly depends on the saturation of active areas on the adsorbent surface and slow pore diffusion (Jacob et al., 2020). It can be concluded that the optimum operating time would be 150 min at which approximately 73% removal efficiency was achieved.

3.7 Comparison studies of walnut shell biochar with other adsorbent

It is very hard to directly compare the other studies with our study due to the various experimental conditions that was used different adsorbents for removal CP removal. But, the experimental results of this study clearly showed that the adsorption capacity of the WSBC was nearly same as other adsorbents with CP adsorption. The maximum adsorption capacity of WSBC was comparised with other adsorbents at Table 2 for CP removal. Vigneshwaran et al (2019), investigated the initial concentration, pH, operating time and dose parameters in the removal of chlorpyrifos using a heterogeneous metal-free graphite carbon nitride (g-C₃N₄) incorporated into chitosan. They achieved 85% efficiency under optimum conditions. It was observed that the yield they obtained was almost the same as the yield in our study (86.64%). Gonçalves et al. (2019), stated that 4 g of activated biochars produced from tobacco biomass are sufficient to cure 1 L of water contaminated with chlorpyrifos. Zheng et al. (2019), stated that while the adsorption capacity for chlorpyrifos was 4.32 mg/g in the biochar samples they obtained at 300 °C, the adsorption capacity increased by 14.8 mg/g in the biochars at 600 °C. In our study, it was observed that the highest removal efficiency of biochar samples obtained at high temperatures was higher. Hamadden et al. (2021a), reported in their study that higher removal efficiencies were obtained in the first operating times in the chlorpyrifos adsorption mechanism using green adsorbent. In the study in which chlorpyrifos pesticide removal was carried out by using nanoparticles of water treatment residues, it was determined that 92% removal efficiency was obtained after 30 min of operation at pH 7, from Hamadden et al. (2021b), reported in another study.

3.8 Reusability of the adsorbent

Reusability of the sorbent is an important factor for an adsorbent for environmental benefit and the cost be minimized. The stability of WSBC@900 was investigated for the reusable test for 10 cycles as shown in Fig. 9. After each cycle, the biochar was washed three times with distilled water and dried at 70 °C, and used in the next cycle. After 5 cycles, there was no significant loss of activity.

Close

Fig. 9

The stability of biochar (biochar pyrolyzed at 900 °C = 250 µg/L, chlorpyrifos = 1000 µg/L, pH = 7.07, time = 300 min.)

Export to PPT

3.9 Adsorption isotherm kinetics and mechanism

Kinetic is an important step in determining the adsorption mechanism and understanding the adsorption steps that affect the adsorption rate. As a result of kinetic studies, the adsorption rate and the required contact time for filling the adsorbent surface with sorbate can be determined. Also, the adsorption rate is an important parameter in choosing the most suitable sorbent. In determining the adsorption rate and rate constant, the compatibility of experimental data and kinetic models such as pseudo-first-order. Pseudo-second-order and intra-particle diffusion were examined. Besides adsorption is generally described through isotherms by the concentration of sorbate adsorbed at a constant temperature as a function of the sorbate concentration on the adsorbent. Freundlich and Langmuir are the most frequently used isotherm models in the literature. Freundlich isotherm describes the concentration of dissolved sorbate on the sorbent surface by the concentration of dissolved sorbate in the liquid. This model assumes that adsorption occurs on a heterogeneous surface. The Langmuir isotherm states that the adsorption of sorbate molecules on the sorbent surface is related to the sorbate concentration on the sorbent surface at a constant temperature and it is based on the fact that the adsorption process takes place in a monolayer on the sorbent surface. The experimental data obtained from batch sorption experiments and the model coefficients were given in Table 3.

Mechanisms for sorption of CP in aquatic environments that have been demonstrated or postulated include hydrophobic bonding, electrostatic attraction, van der Waals affinity, hydrogen bonding, chelation/coordination, covalent bonding, and entrapment in WSBC micropores. Dissolving into the aquatic environment is commonly assumed to be the main mechanism of pesticide transport and entrapment into biochar micropores. Besides to the correlation between the CP sorption coefficient and the microporosity of biochars is very good (R² = 0.96), and normalization of k_d values with micropore content explains an important component of the variability of CP absorption between biochars.

3.10 The adaptive neuro-fuzzy interference system model

The process of creating and structuring the ANFIS model is one of the main problems to be solved. All these processes, such as selecting inputs, selecting the type and number of membership functions for inputs. Determining the number of rules by creating the rule layer, and determining the initial values of membership functions parameters, include the stage of developing the ANFIS model. For the selection of the ANFIS model structure expert knowledge based on experience is generally used. However, in many cases. The knowledge of the experts may be insufficient in determining the most appropriate criteria for the ANFIS model by making the right choices. It is the most applicable process to choose the most suitable variables by making different trials. In this study. ANFIS model was developed with MATLAB R2019b software to well estimate the adsorption efficiency in the range of factors used in the batch adsorption experiments performed to determine the CP removal efficiency of newly synthesized walnut shell-based biochar. The factors of initial sorbate concentration, sorbent dosage, initial pH, contact time, and sorbent reuse which affect the adsorption efficiency in batch adsorption experiments were defined as inputs to the ANFIS model and adsorption efficiency as output. All experimental data set containing 109 data (excluding characterization study) in order to training testing and checking the ANFIS model was randomly divided into 60–20-20%. 70–15-15% and 80–10-10% data, respectively. In addition to the data set ratios membership function variables and optimization methods were changed to develop the best ANFIS model structure and the performance of the model in training testing and control were determined by following RMSE and R² values and presented in Table 4. Each variable was calculated by running the model in 500 epochs.

The ANFIS model structure in the first row of Table 3 is the one with the best predictive ability. This model structure consists of 524 nodes. 243 linear parameters. 45 nonlinear parameters. 288 parameters in total. 65 training data pairs. 22 testing data pairs. 22 checking data pairs and 243 fuzzy rules. The sensitivity of the developed ANFIS model in training testing and control are given in Fig. 10. The ability of the developed ANFIS model to represent the data in the experimental data set is 98.56%. This developed model has similar or better success than many model studies performed in the literature to predict the adsorption efficiency depending on the adsorption variables it contains (El Hanandeh et al., 2021). For this reason, the ANFIS model was found to be quite successful in predicting the CP adsorption efficiency on the WSBC@900 at the experimental data limit depending on the initial sorbate concentration, sorbent dosage, initial pH, contact time, and sorbent reuse adsorption factors. Due to a large number of parameters in the data set used a fuzzy rule creation technique in which ANFIS and different optimization methods such as grid partitioning subtractive clustering fuzzy c-means clustering etc. are integrated can be used to create higher performance ANFIS models by reducing the number of fuzzy rules and parameters.

Close

Fig. 10

Correlation results analysis of adaptive neuro-fuzzy interference system model: training, testing, and checking.

Export to PPT

3.11 Walnut shell biochar cost analysis

In literature, the potential usage and the efficiency of adsorbents is frequently mentioned, but cost analysis was not utilized. This section provides a preliminary assessment of the cost of WSBC. However, the cost analysis may differ from region to region. The simple cost calculation for the production and disposal of WSBC is as follows:

1. Raw material cost:

Since the walnut shells used in the study are an agricultural waste, there is no cost of raw material.

1. Biochar production cost:

The first cost was the electrical energy consumed by the system to produce biochar and the second one is the nitrogen (N₂) gase cost were given to the system for the pyrolysis process. The most attractive biochar was obtained at 900 °C pyrolysis temperature and the cost calculations were carried out according to these conditions. (2) grams of biochar is produced from 5 g of walnut shells and the total time for heating was 240 min (4 h). The power of the system was 1.3 kW.

The cost of 1 g of biochar:

Electricity consumption cost ($) = 1.3 kW*4h*0.11$/kWh) = 0.572 cent/2g = 0.286 cent/g.

1L of N₂ gase per minute was given to the system for 1 h. Since the cost of 60 L nitrogen gase was $15.

The cost of nitrogen gas spent for the production of 1 g of biochar; 15$/2g = 7.5$/g

Since the biochar can be reused 5 times, the cost will be reduced by 1/5 and the actual cost will be 1.56 $/g (1560 $/kg).

1. Biochar disposal cost:

The cost for the disposal of 1 kg of Biochar in the landfill = 2.4$

The sum of production and disposal costs was = 1562.4$/ kg (1.5624$/g)

Therefore, the cost of WSBC is significantly more feasible and economical than commercially available activated carbons.