Bentonite blended with bagasse ash as an adsorbent for reactive red 198 dyes

Adsorption offers an efficient, cost-effective, and eco-friendly method for the treatment of dye-laden wastewater. This work presents reactive red 198 (RR198) removal by adsorption using bentonite clay (BC) blended with sugar cane bagasse ash (SCBA). The adsorbent’s surface morphologies, crystalline phase structures, functional groups, and specific surface before and after adsorption were examined using SEM, XRD, FTIR, and BET respectively. Central composite design (CCD) under response surface methodology (RSM) was applied to optimize independent and dependent variable values. The optimal parameters for RR198 removal using the blended adsorbent were 107 minutes contact time, 0.934 g/L adsorbent dose, and 15 mg/L initial dye concentration, and 85.2% RR198 removal efficiency was achieved. The sorption isotherms and kinetics were evaluated using various existing models. The Freundlich isotherm model (R1⁄4 0.95) and the pseudo-second-order equation best described the adsorption parameters and the RR198 adsorption kinetic mechanism, respectively. Desorption and reusability experiments in batch study confirmed that BC blended with SCBA can be used multiple times for dye removal from wastewater.


INTRODUCTION
Dye-laden industrial effluents bring several environmental challenges with them. Both reactive and basic dyes are consumed in the textile, paper, plastic, and other industries as coloring agents. The textile industry, uses about 56% of total annual synthetic dye production worldwide ( Jain & Gogate 2018). Because of incomplete binding onto materials and inadequate control, huge volumes of dyes are discharged into the environment in untreated wastewater, leading to serious health and environment issues.
Dyes are harmful to living organisms due to their recalcitrant and non-biodegradable properties (Moussavi & Khosravi 2011). Aquatic organisms and humans can be affected by carcinogenic and teratogenic effects of dye (Song et al. 2017). Dyes can also affect various human organs including the kidney, liver, and nervous system (Zhou et al. 2015). Discharge of wastewater containing dyes into water bodies reduces the photosynthetic activity of aquatic plants and, in turn, the natural equilibrium.
The binding rates of reactive dyes onto material fibers are estimated to be between 60 and 70%, so that about 40% of that used is released into the environment (Alimohammadi et al. 2016). Dyes are water soluble and inhibit light transmission through water, and cannot be removed easily using conventional treatment techniques. Reactive dye-laden effluents are toxic and mutagenic, and enhance water body hazards as well as being difficult to remove by biological treatment. Thus, it is necessary to design efficient wastewater treatment technologies for dye removal or remediation before effluents are released into the environment. Known technologies including the advanced oxidation process, ion exchange, microbial fuel cells, ozonation, coagulation, flocculation, and adsorption have been applied to remediate or remove different dyes from industrial effluents (Adebayo et al. 2020). The limitations of these technologies include cost, fouling issues, regular maintenance, and complicated operational procedures (Garg et al. 2003;Dizge et al. 2008).
Adsorption is still one of the most widely used dye treatment methods, taking the lead in high removal efficiency, operating simplicity, low operating cost, eco-friendliness, and the availability of a large variety of adsorbents (Gupta et al. 2004;Khalilzadeh Shirazi et al. 2020). Commercial activated carbon is commonly used as an adsorbent for dye removal from wastewater due to its adsorptive efficiency and capacity. Its high associated cost and difficulty in regeneration offer disadvantages (Srivastava et al. 2007). Various adsorbents have been investigated in their natural form for dye removal from wastewater including bagasse fly ash (Mall et al. 2005), bentonite/charred dolomite (Khalilzadeh Shirazi et al. 2020), natural bentonite (Carvalho et al. 2019), fly ash (Dizge et al. 2008), and pine cone (Dawood & Sen 2012). Many of these suffer from long synthesis procedures, however, or short lifespans making their use impractical in many parts of the world. Consequently, efforts have been made to obtain easily accessible and efficient adsorbents that can be used to treat dye-laden wastewaters in low-income countries such as Ethiopia.
Adsorption can be enhanced by blending with bio-waste materials, which enhance or increase an adsorbent mineral's capacity. Sugar cane bagasse ash (SCBA) is used in landfill and has negligible value (Huabcharoen et al. 2017) but creates environmental and human health problems when handled improperly and disposed into the soil, air or water (Farirai et al. 2020). Thus, agricultural wastes used as adsorbent reduce solid waste volumes and follow the principle of generating money from waste.
Little experimental data or literature are available on reactive dye removale.g. reactive red 198 (RR198)using bentonite blended with SCBA. Bentonite is widely available and with its dye adsorption capacity (Pandey & Ramontja 2016) can be used as a low-cost dye removal adsorbent (Toor et al. 2015;Carvalho et al. 2019).
In this study, adsorption experiments were carried out to investigate the application of a novel blended adsorbent systembentonite and SCBAfor reactive dye sorption.

Adsorbents and Adsorbate
Bentonite and SCBA were collected from Gewane (Afar region) and Wonji Shewa sugar factory (Oromia region), Ethiopia, respectively. The bentonite was crushed, washed using distilled water, dried in an oven (FAITHFUL) at 105°C for 12 hours, and ground to a powder. The latter was screened in a sieve shaker to 125 μm particle size. The SCBA was also washed using distilled water and dried in an oven (FAITHFUL) at 105°C for 12 hours before screening to 125 μm particle size. The adsorbent materials were stored in glass bottles. The RR198 dye was obtained from the Yirgalem Addis textile factory, Addis Ababa, Ethiopia. The dye's characteristics and structure are illustrated in Table 1 and Figure 1, respectively. A 1,000 mg/L stock dye solution was prepared using distilled water and diluted as necessary to obtain dye solutions required for batch experiments.

Adsorbent characterization
Scanning electron microscopy (SEM) (INSPECT, F50, USA) was used to examine the surface morphology of adsorbent materials. Fourier transform infrared spectroscopy (FTIR) (Perkin Elmer, USA) was used to analyze the adsorbents' functional groups. Infrared spectra were recorded in the 4,000-400 cm À1 region at 4 cm À1 resolution. Brannuer-Emmett-Teller (BET) (SA-9600 series, Japan) was used to determine raw adsorbent specific surfaces. The structure of the crystalline phases of the bentonite and SCBA were identified using XRD (XRD-7000, Japan). UV A spectrophotometer (JASCO V-770) was used to determine the concentration of RR198 in the solution by recording the UV absorbance at λmax 520 nm.

Batch experiments
The effect on adsorption of parameters such as solution pH (2 to 12), contact time (0 to 240 minutes), adsorbent dosage (0.5 to 3 g/L), and initial dye concentration (15 to 100 mg/L) were investigated. Batch experiments were carried at 25°C using a magnetic stirrer at 200 rpm. The desired quantity of the blended adsorbents was added in 250 mL Erlenmeyer flasks (working volume 200 mL) containing 50 mg-RR198/L. The solution pH was adjusted by adding 0.1M HCl or NaOH. After mixing for a given contact time, the mixture was filtered using Whitman filter paper (185 mm diameter Cat No 1004 185) and the filtrate analyzed. An adsorbent recycling study was also carried out. The absorbance concentration profile was obtained by plotting the dye absorbance calibration curve from the dye absorbance and concentrations graph. The absorbance for each sample was converted to the final dye concentration using the calibration factor obtained from the calibration curve. The amount of dye adsorbed on the adsorbent surface at time (t) can be calculated from the mass balance Equation (1).
where, q t is the amount of dye adsorbed per unit mass of adsorbent (mg/g), C o the initial dye concentration (mg/L), C t the equilibrium dye concentration (mg/L), V the dye solution volume (mL), and m the adsorbent mass (g). Proportional dye removal (%) was calculated using Equation (2). (2)

Kinetic study
Three known established kinetic modelsthe pseudo-first-order (PFO), pseudo-second-order (PSO) and intraparticle diffusion (ID) modelswere chosen to analyze adsorption capacity at different contact times and reaction rate constants involved during adsorption. All of the models were determined according to linear Equations (3)-(5) (Alemayehu et al. 2011;Singh et al. 2020).
where, q e is the mass of dye adsorbed at equilibrium (mg/g), q t the mass of dye adsorbed at time t (mg/g), K 1 the PFO constant (min À1 ), K 2 the PSO constant (g/mg/min), and K p the ID constant (mg/g/min 0.5 ).

Adsorption isotherms
The Langmuir and Freundlich adsorption isotherm linear models were used to analyze adsorbate/adsorbent interaction at equilibrium. The Langmuir isotherm model is used to establish the relationship in monolayer coverage of adsorbate molecules and homogeneous active sites on the adsorbent surface, and that there are no interactions between adsorbed species (Dehvari et al. 2017). The assumption underlying the Freundlich model is that adsorption is onto a heterogeneous adsorbent. The Langmuir linear equation model (Carvalho et al. 2019) was evaluated using Equation (6) and the Freundlich linear equation model (Das et al. 2020a) using Equation (7).
where, q e is the specific amount of dye adsorbed (mg/g) and C e the dye concentration in the liquid phase at equilibrium (mg/L). The Langmuir parameters are q max (mg/kg), which is related to adsorption density, and b (L/mg), which indicates the adsorption energy. Both are calculated from the plot 1=q e vs 1=C e .
The parameters K (L/kg) and n (dimensionless) are the Freundlich constants, and are related to the total adsorption capacity and intensity of adsorption, respectively. The Freundlich constant K f is used to evaluate the relative capacity or power of adsorption of an adsorbent.

Data optimization model
Central composite design (CCD) is commonly applied under response surface methodology (RSM) to analyze experimental data. CCD was used to analyze the data obtained. Stat-Ease version 11 (Design Expert, USA) was used for the data processing. Table 2 shows three independent parameters: contact time, adsorbent dose, and initial dye concentration. The 3D plots and predicted vs. actual values were evaluated to interpret the interaction between the independent parameters, and their impact on the adsorption process. The general formula of the second-order mathematical model with interactions according to the dependent variable (y i) in the response surface analysis (Asaithambi et al. 2017 where, y i is the dye removal value, b 0 the constant coefficient, b i (i ¼ 1,2 and 3) the linear coefficients, b ii (i ¼ 1,2 and 3) the quadratic coefficients, b ij (i ¼ 1,2, and 3; j ¼ 1,2 and 3) the interaction coefficients and x i , x j the coded factor values.

Desorption experiments
Desorption was done to enable adsorbent recycling and reuse multiple times, which will reduce costs (Muralikrishnan & Jodhi 2020). Before the desorption experiments, 3 g of bentonite blended with SCBA was added to a 250 mL flask containing 200 mL of 30 mg-RR198/L solution at a pH 2 and 25°C. The flask was shaken at 200 rpm for 120 minutes until adsorption reached equilibrium and the residual dye concentration determined. The separated residue was added to a 250 mL flask containing 100 ml of 1M NaOH, which was shaken at 100 rpm for 7 hours. The desorption capacity (q e ) and efficiency (Q) were calculated using Equations (9) and (10). where, q e is the amount desorbed (mg/g), V the solvent solution volume (L), M the weight of spent adsorbent (g) and C f the dye concentration in the desorbing solution (mg/L). Q(%) ¼ q e,desorption q e,sorption Ã100 (10)
The FTIR results for both adsorbents are shown in Figure 3(a) and 3(b). The raw bentonite absorption band at 3,630 and 1,638 cm À1 indicates O-H group bending and stretching in the hydration water on the bentonite surface. Al-Essa (2018) observed peaks at 3,450 and 1,650 cm À1 indicating the presence of OH groups on the bentonite surface. Özcan & Özcan (2004) observed peaks at 3,446 and 3,629 cm À1 , because of OH group stretching bands. While the band at 1,640 cm À1 corresponds to OH group deformation on natural bentonite.
The very strong absorption band at 1,009 cm À1 is recognized as representing Si-O band stretching vibration, providing strong evidence for the presence of a silicate structure. The peaks at 779 and 507 cm À1 show the presence of quartz in the bentonite. Zhirong et al. (2011) report that the band at 796 cm À1 confirms the presence of quartz.
The FTIR analysis of SCBA shows a peak at 3,315 cm À1 representing the presence of OH on the ash surface.   Figure 4(a) and 4(b) show that both adsorbents are rough and dense before adsorption, with pores and active sites available; the internal surface can also be seen. This confirms  successful adsorbent formation, with homogeneous component dispersion and porous structures on the surfaces, which increase their dye sorption capacity. However, Figure 4(c) shows a rather smaller proportion of irregular structures, including pores, which can be attributed to dye adsorption on the adsorbent surfaces with active sites occupied. The surface texture of SCBA is similar to that observed by others (Alves et al. 2017;Farirai et al. 2020). The bentonite surface morphology is also similar to that reported by others (Javed et al. 2018;Carvalho et al. 2019).

Effect of pH
The solution pH is an essential factor in sorption because it can alter the adsorbate-adsorbent interaction surface charge, enhancing attraction or repulsion (Akar et al. 2018). In this study, the batch runs were carried out at various dye solution pH values (2, 4, 6, 8, 10, and 12) at 25°C and stirred at 200 rpm. Figure 5 shows the impact of pH on RR198 sorption onto the blended adsorbents dosed at 2 g/L, initial dye (30 mg/L), and contact time 60 minutes. As can be seen, removal efficiency fell from 86 to 24% with increasing pH. The maximum dye removal efficiency and sorption ability were achieved at pH 2. Acidic conditions (low pH) enhance H þ ion concentrations, which develop a positive surface charge by binding the H þ ion onto the adsorbent surface. Negatively charged groups on the dye surface were attracted by the positive groups on the adsorbent surface (Alimohammadi et al. 2016). The effect of pH on the adsorption of reactive dyes such as reactive red 2, reactive orange and reactive blue 4, is to yield maximum removal efficiency at low values. For instance, when using coir pith as the adsorbent Çolak et al. (2009) andSalari et al. (2021) also reported the highest RR198 removal efficiency at pH 2.

Effect of adsorbent dosage
A series of batch runs was carried out using different adsorbent doses; that is, 0.5, 1, 1.5, 2.5, and 3 g/L, other conditions remaining the same (stirred at 200 rpm, temperature 25°C, contact time 60 minutes, initial dye concentration 70 mg/L, and solution pH 2). Figure 6 shows that as the adsorbent dose was increased from 0.5 to 3 g/ L, the proportional amount of dye adsorbed (q e ) fell from 8.24 to 3.4 mg/g, while dye removal efficiency rose from 27.4 to 73%. Other studies have shown similar phenomena; for example, Carvalho et al. (2019) used activated carbon and bentonite to adsorb direct orange 39 dye, and Das et al. (2020b) removed malachite green dye using bio-degradable sodium alginate/NaOH-treated activated sugar cane bagasse charcoal. Increasing the adsorbent dose increases the number of active sites available (Dehvari et al. 2017).

Effect of initial dye concentration
A similar series of runs to those reported in sub-section 3.3 involved different initial dye concentrations; that is, 15, 30, 50, 70, and 100 mg/L, other conditions remaining the same, with contact time 150 minutes, adsorbent dose 3 g, and pH 2. Figure 7 shows the results. As can be seen, a higher initial dye concentration leads to a decrease in proportional dye removal. The amount of dye adsorbed increased from 0.89 mg/g at 15 mg/L initial dye concentration to 3.8 mg/g at 100 mg/L initial dye concentration, while proportional dye removal fell from 89% at 15 mg/L to 57% at 100 mg/L initial concentration. If active sites are available on the adsorbent surface, proportional dye removal increases with increasing initial concentration, while, if active sites are saturated, proportional dye removal decreases (Zhou et al. 2019). Thus, the higher proportional dye removal at lower initial concentration can be attributed to increased probability of binding all dye molecules to vacant adsorbent sites. Mahanna & Samy (2020) reported similar results in reactive red 195 dye removal using acid-modified soybean leaves.

Effect of contact time
A further, similar set of batch experiments was carried out with different contact times; that is, 0, 30, 60, 90, 120, 150, 180, 210, and 240 minutes. All other factors were kept the same as noted above (sub-section 3.4). Figure 8 shows that dye removal began at some point within 50 minutes of start-up, and then proceeded at a fairly constant rate until almost reaching equilibrium after 150 minutes, after which removal efficiency decreased, perhaps because desorption occurred and active sites were occupied, leading to an increase in dye concentration in solution (Saha et al. 2012;Das et al. 2020b). Table 3 shows that the PSO model fitted the trial results better than the other models, with R 2 ¼ 0.98 indicating that chemical adsorption is more probable than physical adsorption. This is attributed to electron interaction between RR198 dye molecules and the blended bentonite and SCBA during chemical adsorption (Alimohammadi et al. 2016;Tayebi et al. 2016).

Isothermal study
The Langmuir isotherm model was evaluated using Equation (6), and constants q max (mg/g) and b (L/g) were analyzed from the slope and intercept of C e /q e vs C e respectively. The dye concentration range was from 15 to 100 mg/L, but no other factors changed. The values of q max and b were 4.24 mg/g and 0.097 L/g. The Freundlich isotherm model was analyzed using Equation (7), and the constant values K and n were evaluated from the   Langmuir model was 0.87 (Table 4). This showed the heterogeneous distribution of binding sites on the surface of the blended adsorbents.

RSM results
The dye removal results were evaluated using RSM. Figure 9 shows that the actual and predicted dye removal efficiencies were close for the RR198 removal studies.

Effect of contact time vs initial concentration
The interactive effect of contact time and initial adsorbate concentration on RR198 removal efficiency is illustrated in the 3D graph in Figure 10(b). This shows that RR198 removal efficiency increased with increasing contact time. The maximum calculated adsorption efficiency was 93.2%, at 197 minutes contact time and just under 19 mg/L initial dye concentration. The interaction model results indicate that adsorption efficiency was higher with longer contact time because enough time is available for interaction between dye molecules and the adsorbent surface (Brahmi et al. 2019).

Effect of adsorbent dose vs initial dye concentration
Figure 10(c) illustrates the interactive effect of adsorbent dose and initial dye concentration on RR198 removal. As can be seen, raising the adsorbent dose from 0.5 to 3 g/L increased dye removal efficiency, while the increase in initial dye concentration reduced efficiency. The highest proportional dye removal was achieved at low initial dye concentration and high adsorbent dose, and the optimal result was 90.7% dye removal efficiency at a bit more than 15 mg/L initial dye concentration and 2.9 g/L blended adsorbent dose. The adsorbent dose increase enhanced the availability of free active sites on the adsorbent surface (Saha et al. 2012;Davarnejad et al. 2020). This may occur because of the reduced ratio of active sites to dye molecules, so that there was insufficient vacant space on the adsorbent surface (Brahmi et al. 2019).
3.9. Desorption and reusability experiment Figure 11 shows the blended adsorbent's removal efficiency through three cycles of desorption-adsorption. As can be seen, the proportional removal performance fell from 86 to 84.2% between the first and third cycles. At the same time, the adsorption capacity also fell from 2.6 to 2.3 mg/g. These results confirm that bentonite blended with SCBA can be reused for RR198 removal from wastewaters.

CONCLUDING REMARKS
RR198 adsorption was investigated using bentonite blended with SCBA as the adsorbent. The adsorption isotherm model was fitted better by Freundlich than Langmuir, while the pseudo-second-order model fitted the adsorption kinetics best. CCD under RSM was used to investigate the interaction between contact time, adsorbent dose, and initial dye concentrationthree independent parameterson RR198 adsorption. Use of RSM enabled determination of the optimal parameters for removal of the dye, and the precise values of 107 minutes contact time, 0.934 g adsorbent dose, and 15 mg/L initial dye concentration were found to give 85.2% RR198 dye removal efficiency. Desorption and reusability studies were carried out in batch mode and confirmed that bentonite blended with SCBA can be used multiple times for dye removal from aqueous solution. In general, the results indicated that natural bentonite blended with sugar cane bagasse ash can be used to treat textile industry wastewaters to remove RR198 dyes. The blend is an efficient and environment-friendly adsorbent.