Decolorization of multicomponent dye-laden wastewater by modified waste fly ash: a parametric analysis for an anionic and cationic combination of dyes

In this research study, waste fly ash (WFA) underwent acid activation and subsequent amine functionalization using ammonia solution. This treatment improves the porosity, thermal tendency and crystallinity of WFA. Modified WFA was tested under different experimental conditions to treat the wastewater consisting of different concentrations of cationic (methylene blue and rhodamine 6G) and anionic (methyl orange) dyes. As an individual, methylene blue (MB) and rhodamine 6G (Rh) showed ~ 100% and ~ 82% removal efficiencies respectively in an alkaline medium while methyl orange (MO) exhibited only ~ 20% adsorption in the same medium. An antagonistic effect was observed in adsorption when wastewater contains both cationic dyes whereas the combination of cationic and anionic dyes in solution manifested a synergistic effect. For all individual and binary dye combinations, there is a close agreement in observed and calculated uptakes when the data was fitted to the fractional order kinetic rate equation. The adsorption of all dyes is spontaneous and endothermic in nature except for MB/MO combination where the process is exothermic in nature. 24.93 mg/g, 24.83 mg/g, and 14.95 mg/g monolayer uptake capacities of MB, Rh, and MO were found respectively from isothermal analysis of single dye adsorption data. Further, extended sips model gave higher correlation coefficient (R2 = 0.99) and addressed the failed assumptions of both the Langmuir and Freundlich models. Overall, in the experimental results, the modified waste fly ash could act as successful adsorbent to treat dye bearing wastewater.


Introduction
Water is a natural strategic and economic resource that plays a crucial role in the social and economic development of any community. However, swift industrial development has intensified the interference in the ecosystem because of human activities which leads to the chronic occurrence of water adulteration and the degradation of ecology. Among different types of pollution, water adulteration reportedly is the reason for 1.4 million premature deaths in 2019 (Fuller et al. 2022). About 2 million tons of sewage from industrial and agricultural wastes is being dumped every day into natural water bodies across the world. Industrial aqueous discharge is playing a vital role in water contamination whose composition, volume, strength, and flow vary depending on the source. Typical industrial sectors participating in the generation of wastewater are petrochemicals, paper and pulp, food processing, pesticide, refineries, sugar, and textile Responsible Editor: Guilherme L. Dotto (Peters and Werner 1995). According to the Pakistan Water Sector Strategy report, industries produce 1.309 × 10 9 m 3 / year of wastewater (Murtaza and Zia 2012) wherein textile industry contributed a major share by discharging colored wastewater to nearby aquifers. A typical textile factory generates about 1000-3000 m 3 /day of wastewater for 12-20 t/ day of its product. Discharged wastewater from textile factories contains multiple reactive and azo dyes, sizing agents, organic derivatives, and a large amount of salt because of which it possesses high chemical oxygen demand (COD) and biological oxygen demand (BOD). In particular, dyes are the most dangerous because of their tendency to inhibit the growth of biota by impeding light penetration through the water's surface which affects the photosynthesis activity and life of marine creatures (Ilyas et al. 2019;Li et al. 2020;Jarrah 2017;Tahira et al. 2019).
Various techniques have been employed for the treatment of dye-laden wastewater but the adsorption is the most competent method because of flexibility in operation and its effectiveness for dilute solutions and it is easy to scale up and there is no formation of by-products (Lv et al. 2019;Bhatti et al. 2020;Patra et al. 2020). Pal et al. (2016) prepared the nanostructures containing mixed metal oxides of titanium, silicon, and aluminum through sol-gel-hydrothermal method. The synthesized nanoparticles were tested for cationic (methylene blue and rhodamine 6G) dye removal from wastewater. The nanoadsorbent showed 162.96 mg/g adsorption capacity for methylene blue while it was only 142.91 mg/g for rhodamine 6G. Zahir et al. reported the synthesis of diammonium tartrate and urea-assisted crosslinked chitosan and used it for the abatement of Congo red from aqueous solution. It was shown that diammonium tartrate-modified chitosan (1597 mg/g) proved superior to urea-modified chitosan (~ 1447 mg/g) (Zahir et al. 2017). An activated carbon functionalized with para-aminobenzoic acid was synthesized by Naushad et al. and they applied it to eradicate malachite green dye from wastewater. They found 7, 30 mg, 25 °C, and 240 min as an optimum pH, adsorbent dose, process temperature, and contact time respectively for 66.87-mg/g uptake capacity of functionalized activated carbon (Naushad et al. 2019). Kang.et al. (Kang et al. 2020) studied the adsorption of methylene blue and methyl orange by using chitosan gel as an adsorbent. This work presented that with the shielding effect of methylene blue (MB) on exfoliated montmorillonite/chitosan gel, the maximum adsorption capacity of methyl orange (MO) was promoted from 545 to 1060 mg/g. Numerous research articles have been published addressing the adsorption from a singlecomponent solution, whereas limited literature reported the competitive sorption of pollutant dyes from multicomponent solution. Real wastewater contains multiple dyes, and research investigations based on single dyes could hardly reflect the actual situation. The adsorption mechanism, complex interaction of different variables, and properties of adsorbent conclude that the simultaneous removal of dyes is a significant challenge; that is why it has piqued the attention of researchers worldwide (Yadav et al. 2022).
Diverse kinds of materials are available and reported as an adsorbent but the development of low cost, stable, and an efficient adsorbent is still a challenge for researchers Azari et al. 2020). However, high specific surface area, tunable properties, wide range of raw materials, and easy production process for activated carbon make this material a primary option for the adsorption of pollutants of any kind (Saleem et al. 2019).
Materials with high carbon content such as coal, different agriculture wastes, fly ash, and petroleum coke are the chief sources to synthesize activated carbon. Among them, activated carbon originating from waste raw material could be attractive and viable at the same time to treat polluted water. According to the survey, the worldwide generation of waste fly ash (WFA) is 750 million tons/year. The combined total production of waste fly ash (WFA) by the USA and the European Union is 115 million tons/year (Yao et al. 2014). In addition to this, the rate of waste fly ash generation in Pakistan by the power plant at Khanote is 166,550 m 3 /h (Aziz et al. 2010). Disposal of such a huge amount of fly ash as landfills in open space is vulnerable and could cause severe environmental and health issues. Although fly ash varies in composition depending on the type of fuel source in the power plant, waste fly ash from heavy oil-fired power plants contains a high fraction of carbon (~ > 70%). Therefore, this could be considered a strong candidate owing to the high fraction of unburned carbon, low bulk density, and fine particle size (Ur Rehman et al. 2020). Labaran and Vohra (2016) produced the activated carbon (63-m 2 /g specific surface area) from oil fly ash when fly ash was treated with 40% solution of phosphoric acid at 500 °C temperature. They found that the activated carbon favored the adsorption of methylene blue when the pH was above 5. However, the adsorption tests with methyl orange showed reverse behavior and gave higher adsorption in acidic medium. Primerano and Milazzo reported recovery of valuable metals from oil fly ash and its subsequent use for the removal of methylene blue dye from wastewater (Primerano and Milazzo 2020). Large dye removal efficiency (~ 99%) of fly ash-based absorbent was observed in less than 1 h and the uptake of dye follows the Langmuir isotherm model with 40-mg/g monolayer saturation capacity. Using waste fly ash to treat the polluted water would benefit in both ways of managing the fly ash and cost-effective solution for wastewater treatment.
Therefore, the objective of this research is to produce surface-functionalized activated carbon from waste oil fly ash. The emphasized novelty for current research involved a comprehensive insight into adsorption capability of prepared adsorbent when it is applied for removal of cationic (methylene blue and rhodamine 6G) and anionic methyl orange from wastewater. Further, synergistic and antagonistic behaviors of dyes are also studied in a competitive adsorption. The prepared adsorbent has been analyzed to know the modification in its characteristics. In addition, the adsorption data is obtained by varying different experimental factors and subsequently analyzed with isotherm, thermodynamic, and kinetic models.

Materials
Methylene blue (C 16 H 18 CIN 3 S, MW 319.85 g/mol, dye content 82%, λ max = 660 nm), methyl orange (C 14 H 14 N 3 NaO 3 S, MW 327.34 g/mol, λ max = 460 nm), and rhodamine 6G (C 28 H 30 N 2 O 3 HCl, MW 479.02 g/mol, λ max = 523 nm) were provided by LOBA Chemie and BDH Chemicals respectively and were used without prior purification. Analytical grades of nitric and phosphoric acids were obtained from Panreac Company, Spain. Ammonia solution (32% w/w) of density 0.9 g/cm 3 was delivered by Scharlau Company, Spain. The power plant waste fly ash (WFA) used in the experiments was obtained from the Rabigh Power Plant located in Jeddah, Saudi Arabia.

Modification of waste fly ash
Known weight of plant waste fly ash (WFA) was passed through 45-mesh # BSS Taylor Sieve and picked out 15 g of undersize for modification. In a typical run, WFA was blended in an acid mixture (HNO 3 to H 3 PO 4 , 30:120, v/v) and 150-mL water was further added into it. The resultant mixture was boiled for 4 h under total reflux conditions. Activated WFA was filtered and washed with deionized water until the pH of filtrate became neutral. Subsequently, the solid residue was dried in an oven. After the material gets dried, 10 g of aliquot was soaked in 75 mL of ammonia solution which was further diluted by adding 200-mL distilled water and stirred at 200 rpm for 2 h. After 2 h, the mixture was filtered and residue was washed several times and later dried in an oven. The material was so named as "modified WFA."

Characterization techniques
The zeta potential measurement of WFA and modified WFA was made using Zeta Plus Instrument (Brookhaven Instrument Corporation, USA). The instrument employed the electrophoretic light scattering (ELS) technique to determine the zeta potential of charged particles in colloidal suspensions. The crystalline structure and phases of waste fly ash in raw form and after modification were analyzed by X-ray diffraction (XRD) on X′ Pert Pro PANalytical diffractometer. It was operated at K α1 = 1.540598 Ǻ, 45 kW, and 40 mA, with a scanning range of 2θ from 5 to 120°, a scan speed of 0.05 s/step, and step size of 0.02°. The rate of change in the weight of WFA as a function of temperature was measured using thermo-gravimetric analysis (TGA). Analysis was carried out through an SDT analyzer (Q600; TA Instruments, USA). Results were obtained by heating the 5-10-mg sample from 300 to 1100 K at the rate of 10 °C/min with a nitrogen flow of 20 mL/min. The surface morphology of WFA and modified WFA was observed using electron microscope (JEOL-JXA 840A, Japan) and dried powder of each specimen was gold coated in a sputter coater (Quorum Technologies, Q300T T Plus, UK) before capturing image. Fourier transform infrared (FTIR) spectrum of the adsorbent sample was obtained using fiber probe coupler (FPC) FTIR Perkin Elmer spectrophotometer.

Binary adsorption tests and data analysis
Various concentrations of aqueous solutions comprising two different combinations (Table 1) were used in the batch experimentation. These were contacted with modified WFA under different experimental conditions in a batch adsorption run. Operational parameters pH, initial concentration of dyes, adsorbent dosage, and temperature were varied for each combination to see the effect on the removal of the pollutant dyes. The volume of the pollutant solutions was kept constant at 100 mL in all runs. A predetermined amount of modified WFA was mixed with single and binary dye solutions in an Erlenmeyer flask, and the capped flasks were stirred at 150 rpm on an orbital shaker (Thermo Fischer-Scientific, UK). After equilibrium, the solid and liquid were separated by centrifugation. The supernatant was analyzed through a UV-visible spectrophotometer (UV 1601, Shimadzu, Japan), and the concentration of each dye was calculated by the following formulas (Eqs. 1 and 2). Both equations need absorbance values corresponding to λ i and λ j and the individual and cross-calibration constants (Yazdani et al. 2012). The calibration curve of each dye was drawn by measuring the absorbance of MB, MO, and rhodamine 6G (Rh) at 660 nm, 460 nm, and 523 nm respectively.
where i and j represent the dyes in a binary combination. λ i and λ j mean the absorbance at respective wavelengths of maximum absorbance of dyes i and j . k ii , k ij , k jj , and k ji are the calibration constants wherein the first letter in the subscript shows pollutant dye and the second letter stands for the wavelength of maximum absorbance. Combinations of dyes experimented in this research and respective calibration constants are summarized in Table 1. Further, the experimental data were analyzed by appropriate isothermal, kinetic, and thermodynamic models listed in Table 2. OriginPro 9.1 software was used for the non-linear regression of above said models. (1)

Characterization of modified WFA
Modifications in the crystal structure of both raw WFA and modified WFA were analyzed by X-ray diffraction and presented in Fig. 1a. The broader peak centered at 25° in raw WFA is the characteristic peak for amorphous carbon (Aslam et al. 2019b). The presence of quartz and mullite as a main crystalline phase can also be seen as minor peaks at 34° and 43° (Yaumi et al. 2013). Modified WFA yields intensified peaks which show improvement in crystallinity after chemical modification. Two additional peaks superposing at 63.9° and 77° are associated with the different facets of quartz and mullite phases supporting the fact of improving crystalline nature (Gilja and Krehula 2019). These results propose the relative increase in the content of the crystal phases obtained as a consequence of the surface modification process of raw WFA. Thermograms in Fig. 1b show the removal of free and bonded moisture from raw and modified WFA in the temperature range 350-425 K. Between 425 to 650 K, no considerable mass loss was observed in both curves because both comprise a network of high molecular weight carbon rings which is stable against this temperature range. In raw WFA, ~ 40% mass was lost beyond 650 K because of the o 1 1−n C e (mg/L) = equilibrium concentration C o (mg/L) = initial concentration n = reaction order, where n > 1 (Levenspiel 1999) For n = 1 Q e,i (mg/g) = equilibrium adsorption capacity for the component i C e,i (mg/L) = equilibrium concentration a, b (L/mg) = Langmuir constants q m,i (mg/g) = monolayer capacity N = total number of components n i = adsorption intensity for component i derived from the individual adsorption Freundlich model xi, yi, z i = constants obtained from the set of experimental values by minimizing the error in non-linear regression analysis Extended sips e,j ) m = sips model exponent derived from the individual adsorption sip model η (L/mg) = sips constants Thermodynamics Gibb's free energy ΔG = -RTlnK R = universal gas constant (8.314 Jmol −1 K −1 ) K d = distribution coefficient (Senturk et al. 2010) Enthalpy ΔG° = ΔH° − TΔS°ΔH° ((kJ.mol −1 K −1 ) = enthalpy change ΔS° (kJ.mol −1 = entropy change ΔG° (kJ.mol −1 ) = Gibbs energy change breakdown of macromolecules into smaller volatile molecules which were inherently attached to raw WFA (Tsai et al. 2001). In the modified WFA, the major mass loss is observed after 825 K which is probably because of the degradation of carboxylic and amine surface functional groups (Safwat and Matta 2018). Comparing the thermal stability of both materials, it can be observed that WFA degrades at a faster rate than the modified WFA. The total weight loss for the WFA is about ~ 45% whereas for modified WFA, it is only ~ 12% which means later is more stable. In Figure S1, SEM photograph shows raw WFA composed of spherical particles of different shapes with macropores on their surface. In comparison to WFA, lots of small-size pores are visible on the surface of modified WFA. Figure S2 presents the FTIR spectrum of both raw and modified WFA. It can be noticed that the peaks' positions at higher wavenumber are unchanged except their intensity. A new peak at 1467 cm −1 indicates the methylene functional group. A band between 1100 and 1250 cm −1 get broader in modified WFA compared to raw WFA probably because of esterification of carbonyl groups which is also evident from the appearance of new the peak at 1741 cm −1 (Ur Rehman et al. 2020).

Adsorption capability
Initial solution pH is an essential aspect in the adsorption as it governs the degree of ionization of solution entities. The removal efficiency for each dye as an individual and in combination with another as a function of pH is presented in Fig. 2a. The removal efficiency of MB(s) and Rh(s) has a positive trend with solution pH as compared to MO(s) where adsorption shows an inverse trend as depicted in Fig. 2a. The negative trend of MO(s) is probably due to the competition of hydroxyl ions with anionic MO resulting in decrease in removal at higher initial solution pH. MB removal by modified WFA is better than Rh and MO. It may be associated to small molecular size (i.e., 13.82 Å) of MB compared to other dyes and its resultant higher mobility in the solution and π-π interaction between adsorbate molecules (Gong et al. 2013). In an acidic medium, electrostatic repulsion between basic dyes like MB and Rh and positively charged adsorbent surfaces causes lower removal efficiency of both basic dyes. However, in an alkaline medium, the interaction between adsorbent and adsorbate increases due to de-protonation which helps to achieve greater removal efficiency for basic dyes (Karaca et al. 2008). Since the adsorbent surface gets negative charge as pH increases, therefore, removal efficiency of anionic dye MO declines towards an alkaline medium. The removal efficiency for each dye from a binary solution can be seen in Fig. 2b. The combination of both cationic dyes establishes the repulsion phenomena in binary adsorption. This antagonistic effect decreases the removal efficiency for binary dyes in comparison to the individual dye adsorption. However, MB adsorption efficiency towards adsorbent was still high in binary solution with Rh because of self-association of Rh forms dimers, trimmers, and higher aggregates which promotes a push-pull mechanism for MB removal. However, for the combination of cationic (MB) and anionic (MO) dye, the individual uptake of both dyes rises markedly as compared to a single dye solution. This means that the presence of another dye imparts a synergistic effect for competing dye in a binary solution. Hence, the results of this set of multicomponent mixture depict increase in efficiency due to electrostatic interactions of both anionic (MO) and cationic (MB) dye apart from adsorption interactions of individual dyes with adsorbent. Optimized removal efficiency for this binary dye solution was also achieved against the initial solution pH 6.4. The data of removal efficiency may also be supported by the zeta potential graphs (Fig. 2c). Generally, values between + 10 and -10 mV are counted as undesirable because of instability of the particles in this range. However, the value beyond this limit (either positive or negative) represents the stability of the adsorbent particles in colloidal suspensions. The stability of the adsorbent particles is directly related to the removal efficiency of the adsorbent. The measured values of zeta potential for both WFA and modified WFA are presented in Fig. 2c. The plot shows that both materials give the zeta potential values between − 5 and -15 mV for a pH range of 1.8 to 3.2. This indicates the unstable colloidal suspension which is not recommended for adsorption. However, there is a huge drop in zeta potential values at higher pH for modified WFA. This change in zeta potential indicates the efficacy of pollutant adsorption of modified WFA over raw WFA probably because of positive charges of quaternary amine functional groups on its surface. The graph depicts that zeta potential becomes almost constant for the 6.42-10.5 pH range which shows constancy in charge density over modified WFA's surface. The highest zeta potential value for the highest removal efficiency is -40 mV against the pH value of 6.41.
The effect of dosage of adsorbent on the removal percentage was also investigated and results are plotted in Fig. 3. According to the data, removal efficiency increases by an adsorbent amount for all three dyes. For any particular adsorbent dosage, the removal efficiency follows the order MB > Rh > MO. It can be seen that removal percentage is insignificant at higher dosage. For further experiments, 0.4 g was chosen as representative dosage.

Kinetic analysis
Kinetic data describes the rate of adsorption and the analysis of kinetic data produces the basis for the adsorber design. Figures 4 and 5 indicate that there is a sharp decrease in the concentration ratio of dyes in the initial few minutes of their exposure with modified WFA for all single and binary combinations. After fast initial uptake, the rate of adsorption became slower until an adsorption plateau was reached. Approximately, 20 min is taken by dye when it is being adsorbed from a single-component solution and after this time, there is insignificant change in concentration ratio. However, adsorption from binary dye solution takes a little longer to reach a plateau probably because of competition of dye moieties for active sites on the adsorbent surface. The fast adsorption of pollutant dyes can be attributed to the excellent texture of modified WFA which provides favorable reaction affinity sites and improved diffusion for the dyes. The slower adsorption rate at later stages might be due to less number of unoccupied sites and the resultant increase in diffusion resistance.
Pseudo-first-and second-order kinetic model, Elovich model, and intra-particle diffusion model were applied  Figure S3) to data obtained for adsorption of dye when it is being adsorbed from monocomponent solution (kaur, S., Rani, S., and Mahajan 2012; Ojedokun and Bello 2017). The regression results of the Elovich and intra-particle diffusion model are presented as Table S1. None of the models follows exactly the whole experimental range which can also be witnessed from their poor correlation coefficients. The deficient R 2 in the regression of single dye adsorption data motivated the authors to apply a fractional-order kinetic model. In this regard, the kinetic data for the adsorption of dyes is also analyzed by implementing the nth-order kinetic models. Table 3 compares the regression results of tested kinetic models for both sets of binary solution, i.e., MB/Rh and MB/MO. Results indicate significantly lower (R 2 ) values for kinetic models of order n = 1 and 2. This suggests that the experimental uptake for individual and binary dyes does not obey these two kinetic models strictly. Both integer-order kinetic models over-estimate the dye adsorption corresponding to the time which can be described by their high error values. A fractional-order kinetic equation is also tested for which close agreement between the experimental and calculated values for the entire adsorption can be confirmed by their lower error values and higher squared correlation coefficients (R 2 ). Further, the effect of the driving force is indicated by the "n" and the parameter "k n " could be an overall parameter coupling various adsorption-related factors (Wei et al. 2017). For the aqueous solution containing cations dye mixture (i.e., MB and Rh), the parameter "n" is increasing by increasing the temperature. This shows the direct relationship between the driving force for adsorption and temperature for both single and binary dyes.
Comparing the results of the cation and anions binary dye mixtures, Figs. 4 and 5 indicate that the higher equilibrium adsorption capacity for MB/MO than the single dye adsorption reflects a synergistic effect which promotes adsorption of each other. The driving force "n" is decreasing with temperature which manifests decrease in the adsorption of the MB/MO binary mixture by rising temperature. The antagonistic effect is revealed for the MB/Rh combination, which is also supported from the "n" values in the Table 3 for MB and Rh.

Thermodynamic analysis
Thermodynamic analysis is another important study for multicomponent adsorption because it helps to get an insight into the adsorption mechanism, surface properties and degree of affinity of the adsorbent. The enthalpy change (ΔH) gives an idea of the type of adsorption by evaluating its value on a scale of 1 to 60 kJ/mol. Enthalpy change below 50 kJ/mol shows physisorption process whereas a value > 60 kJ/mol represents chemisorption of the pollutant (Rida et al. 2013). Calculated results of the first set of a binary mixture of MB and Rh in Table 4 show physisorption for both single and binary dye solutions, which means van der Waals and electrostatic attraction is involved between the adsorbent and the pollutant dyes. The positive values of the (ΔH) and (ΔS) show that heat drives the process forward hence endothermic in nature. As MB is hydrophobic and Rh is a highly fluorescent hydrophilic dye, they attract each other through weak van der Waals' forces in addition to their interaction with modified WFA surface. An increase in temperature causes an increase in the motion of the particles resulting in an increase in the attraction of dyes towards the adsorbent (Fig. 6).
Negative ΔG indicates that the adsorption is spontaneous. An increase in the negativity of ΔG with rise in temperature Table 3 Order-based kinetic model regression results for adsorption from a single as well as a binary system. The initial concentration of single dye solution = 70 mg L −1 ; initial concentration of binary dye mixture = 140 mg L −1 . In the binary dye solution, dyes are in equal fractions (time = 1 h, adsorbent dose = 0.4 g, volume = 0.1 L, pH = 6.4 ± 0.1) **Where MB Rh shows MB properties when it is in combination with Rh. The same applies to other symbols Component Temperature (K) For n = 1 For n = 2 nth order k 1 (min -1 ) R 2 Error k 2 (L mg -1 min -1 ) R 2 Error k n (L mg 1−n min -1 ) n   Binary dye solution results (single dye solution is represented by the symbol "s" and the "b" symbol stands for binary dye solution, initial concentration of single dye solution = 70 mg L −1 , initial concentration of binary dye solution = 140 mg L. −1 ).
In the binary solution, dyes are in equal fraction (time = 1 h, adsorbent dosage = 0.4 g, solution volume = 0.1 L, pH = 6.4 ± 0.1) reveals that higher temperature favors the adsorption process because of increased mobility of the dye molecules (Manjunath and Kumar 2018). Compared to individual adsorption of MB and Rh, binary dye adsorption depicts a decline in ΔH as well as ΔG reflects the less affinity of binary dye solution towards modified WFA. At the same time, decrease in positive values of ΔS for the binary adsorption shows less randomness on the surface of adsorbent because of adsorbate. The reason behind the less affinity for a combination of MB and Rh shows an antagonistic effect. Thermodynamic findings are also shown in Table 4 for the second combination (MB/MO) under consideration. For this, Gibb's free energy of MB when it was adsorbed from monocomponent solution is − 19.56 to − 21.64 kJ/mol; however, ΔG for the same dye is 8.22 to 9.10 kJ/mol when it was adsorbed from binary dye solution. Gibb's free energy for MO is in the range of − 5.56 to − 6.15 kJ/mol while it is 8.95 to 9.90 kJ/mol for MO MB respectively. The negative ΔG values indicate spontaneity for both MO and MB. The negative value of standard enthalpy change (ΔH) indicated that the process is exothermic. MO is anionic dye where MB is cationic; they generally attract each other rather than adsorbent surface. Therefore, during adsorption, there is a decrease in the attractive forces present on the surface of the adsorbent, leading to a decrease in the surface energy of the adsorbent, which is evolved as heat. Thus, the enthalpy of adsorption is negative and process is exothermic. Entropy of the system also enhanced as shown in the tabulated values for binary solutions.

Isotherm results
Adsorption isotherms describe the phenomena which govern the retention or mobility of the adsorbate from the polluted solution to the sorbent surface at a constant temperature. Adsorption equilibrium explains the dynamic balance of the adsorbate with the interface concentration. Typically, mathematical correlations of the isotherm models constitute an important role in the operational design, modeling analysis, and applicable practice of the adsorption systems as depicted by the graphical expressions of the adsorbent capacity against the residual concentration of the pollutants in the solution. The analysis of the binary experimental data shows that, for a constant value of the initial concentration ratio, it is possible to implement different adsorption isotherm models. Langmuir, Freundlich, Temkin, and Dubinin-Radushkevich isotherm models were applied to MB, MO, and Rh adsorption from monocomponent solution (Mozaffari Majd et al. 2022). The corresponding graphical plots ( Figure S4) and summary of their regression results are presented as Table S2. Analysis of results suggests that the Langmuir model gives good results for both MB and Rh; however, it gives unrealistic value of uptake capacity for the case of MO as shown in the Table S2. Other models produced poor correlation coefficients for all dyes which make them unsuitable to further explore. For the binary solution comprising cationic dyes, a higher amount of MB adsorbed compared to Rh could be imputed to its stronger interactions with the substrate and higher hydrophobicity. The affinity of modified WFA for dyes can be determined by the extended Langmuir model. Large values of Langmuir constant "a" for the single dye in which the other dye concentration is zero show a stronger bonding with the adsorbent (Turabik 2008). It can be seen from the results that the "a" value for an individual component MB (a = 8.63) is higher than Rh (a = 6.55).
In the case of binary components solution, the Langmuir constant "a" shows the affinity of the adsorbent for the variable dye. Lower values of "a" with an increase in other dye concentration clearly indicate the competition between constituent dyes for the available surface area. This decline in the affinity of the modified WFA can be attributed to the antagonistic effect (Table 5).
However, results obtained for the MB and MO dye solution proved the exception. With this combination, the presence of MO imparts a synergistic effect over MB uptake in binary solution adsorption. This can be confirmed by the enhanced values of affinity constant "a" against the increased values of other dye concentrations. Cooperative adsorption can be associated with either alteration of the overall charge within the solution or the reorientation of adsorbed molecules. On the other hand, molecular wedging effects may have been liable for the creation of new adsorption sites on the adsorbent surface. Many researchers have also reported selective adsorption in multicomponent systems. Adsorbent preferences for a different type of dyes adsorption depend upon many factors like solution chemistry (e.g., pH, ionic strength), properties of the adsorbates (e.g., ionic nature or standard reduction potential, molecular structure, concentration, ionic size, ionic weight), and characteristics of the binding sites (e.g., surface properties, structure, functional groups) (Laskar and Hashisho 2020). That is why observed adsorption results are the combinations of all the factors. The average percentage errors between the experimental and predicted values (%) were calculated using the following equation, where N shows the number of measurements (Turabik 2008): Analysis of isotherm models is also made by the comparison of calculated values with the experimental equilibrium data for the description of binary solution adsorption. Moreover, the error and correlation coefficient values (R 2 ) generated from the non-linear regression analysis of the binary adsorption solution play an important role as deciding factor. The comparison of binary mixtures of (MB, Rh) and (MB, MO) is represented by the parity graph, i.e., Fig. 7. Demonstrated points of the graph of the variable dye Rh and MO show the inadequacy of the dyes towards adsorbent by the deviation from the square's diagonal. This resulted because of the limited number of identical adsorption sites available for the Rh and MO in binary systems, to the interactive effects and to the competitive effects that exist in multicomponent system. However, most of the data points of the variable dye MB were distributed closely to 45° line which shows that it fits satisfactorily and adequately as compared to other dyes. These results can be verified by the highest correlation coefficient (R 2 > 0.9) and lower error values of the variable MB dye for both binary systems. It can be seen through results and graphs that the extended Freundlich model fits poorly because of its highest error values along with the lowest (R 2 ) value for both binary solutions of (MB-Rh) and (MB-MO). These results can be attributed to the insensitivity of the extended Freundlich model to competitive and interactive effects existing in multicomponent systems. The parity graph suggests that the extended Langmuir model fits relatively well with the binary solution for the adsorption of MB and Rh. But the basic assumptions of the extended Langmuir model, (1) equal independent competition between species, (2) constant energy of adsorption, and (3) no interaction among solute molecules, appear invalid in this application. However, the extended Langmuir model makes a good attempt for predicting higher capacity dye in each solution but fails to predict the correct monolayer for the lower partner in each combination. The extended sips model deals with the failed assumptions of both Langmuir and Freundlich models. It is the modified form of both models and is also known as the extended Langmuir-Freundlich model. This can be confirmed by the relatively higher R 2 values of the extended sips model and better-fitted graph to the experimental data of the binary system even for the lower capacity partner in each combination. By comparing two combinations, results suggest that the MB and MO binary system gives better regression results probably because of synergism between them. The adsorbent surface-adsorbed methylene blue has a positive charge and would act as a site for the adsorption of methyl orange by electrostatic interactions. This conclusion can be verified through smaller error between experimental and calculated uptake capacities, higher correlation coefficient, and increased affinity of MB and MO binary solution towards adsorbent. Being smaller in size, MB approaches the adsorbent sites quickly; that is why its adsorption efficiency is greater compared to MO and Rh even at lower concentration. The performance of modified WFA is also gauged by comparing its uptake with the literature reported values, and data is presented in Table 6. It can be seen that the uptake of prepared absorbent is comparable with reported adsorbents.

Conclusion
The properties of WFA have been studied along with its modification with acid and ammonia solution. Only 12% weight loss for modified WFA shows its higher thermal stability compared to raw WFA. No change has been observed in zeta potential values for a pH range of 6.42-10.5. Adsorbent showed the highest removal efficiency at − 40 mV and 6.41 pH. The kinetic analysis suggests that nth-order model parameters justify the complex nature of the adsorption mechanism and shows R 2 > 0.91 and smaller error values for both individual and binary solutions. To analyze the best adsorption results between the two binary solutions (MB/Rh) and (MB/MO) as single and binary dye, isothermal models were applied. Single dye adsorption results are good as compared to binary adsorption. Single solution systems of MB, Rh, and MO give experimental adsorption capacities of 24.93 mg/g, 24.83 mg/g, and 14.95 mg/g respectively.
For the binary dye solution, both dyes affect each other and the attraction among them decreases the adsorption ability of the adsorbent for the case of cationic dyes (MB/ Rh). However, for the cation and anion (MB/MO) solution, adsorption ability got raised because of the synergistic effect, where it can also be observed that MB is more attractive towards the adsorbent and gives more adsorption regardless of the type of solution, i.e., single or binary solution. Parity graphs give the idea that though the extended Langmuir model fitted relatively well to the binary solution for adsorption, the extended sips model deals with the failed assumptions of both Langmuir and Freundlich models with the combination of the highest correlation coefficient (R 2 = 0.99) and high affinity towards modified WFA. The thermodynamic model generated results for positive enthalpy values of range 1-50 kJ/mol describing that involved single and binary physisorption process is spontaneous and endothermic for MB/ Rh solution while it is exothermic for MB/MO solution. Overall, the modification of WFA enhances its adsorption capability and it could be treated as an economical source for wastewater treatment at larger scale.