Mechanism and isotherm modeling of effective adsorption of malachite green as endocrine disruptive dye using Acid Functionalized Maize Cob (AFMC)

Cationic Malachite green has been identified as a candidate for the endocrine disruptive compound found in the environment. In this study, the mechanism and isotherm modeling of effective adsorption of cationic malachite green dye onto acid-functionalized maize cob (AFMC) was investigated by batch technique. The operational parameters such as initial concentration (100–600 mg/L); contact time (10–120 min) and pH (3–10) influenced the removal efficiency and quantity adsorbed. A maximum of 99.3% removal efficiency was obtained at optimum conditions. AFMC physicochemical properties (surface area 1329 m2/g and particle size 300 μm < Ф < 250 μm) enhanced its efficiency. Based on R2 > 0.97 and consistently low values of adsorption statistical error functions (ASEF), equilibrium data were best fitted to Freundlich isotherm. Kinetic data were best described by a pseudo-second-order model with consistent R2 > 0.98 and validated by ASEF. The mechanism of the process was better described by intraparticle diffusion. Evidence of the adsorption process was confirmed by the change in morphology via Scanning Electron Microscopy (SEM) and surface chemistry by Fourier Transform infrared (FTIR). The performance of AFMC enlisted it as a sustainable and promising low-cost adsorbent from agro-residue for treatment of endocrine disruptive dye polluted water.


Acid Functionalized Maize Cob (AFMC) as low-cost adsorbent. Purposive and simple random sam-
pling technique which is the best time saving technique was used for the collection of the maize cob agro-waste from the dumpsite of the University being an Agricultural-based University. Maize cobs obtained from Landmark University (Agro-based University) were screened and cleaned thereafter dried at 105 °C for 5 h in Gen lab oven, crushed, grounded, and screened to 106 µm. Acid activation was carried out following the procedure in our previous study 30 and elsewhere in other literature 31  Physicochemical and spectroscopic characterization of AFMC. Determination of pH of AFMC. pH determination of AFMC was done by boiling 1 g AFMC in 100 mL distilled water for a period of 5 min. This was allowed to cool and its pH value was measured using an ATP-6 pH meter.
Determination of AFMC bulk density. Weight difference divided by the volume as depicted in Archimedes' principle was used for bulk density determination as depicted in Eq. (1) 32 W 1 = Weight of empty measuring cylinder, W 2 = combination of AFMC mass and the crucible, V = volume.
Determination of AFMC moisture content. Moisture content was determined typically by introducing 5 g AFMC into the initially weighed crucible and heated for 1 h at 105 °C. Evaluation of the moisture content was done using Eq. (2) 42 W 1 = Weight of crucible, W 2 = Initial weight of crucible with sample, W 3 = Final weight of crucible with sample.
Determination of AFMC surface area by Saer's method. The AFMC surface area was determined using Sear's method. This involves acidifying 0.5 g of each adsorbent with 0.1 M HCl to a pH of 3-3.5. The volume was made up to 50 mL with distilled water after the addition of 1 g of NaCl. The titration was carried out with standard 0.1 M NaOH at 298 K to pH 4, and then to pH 9.0 following the procedure reported in the literature 33,34 . The volume needed to raise the pH from 4 to 9 was noted and surface area evaluated using Eq. (3):  1) solution was prepared by dissolving 1 g MG salt in 1000 mL distilled water. A lower working concentration was prepared (100-600 mg/L) by serial dilution.
Biosorption operational parameters. Various Operational parameters relevant to this study were carried out following reported method 3,4,35 . The effect of pH was determined by varying the pH values between 3 and 10 via dropwise addition of 1 M HCl or NaOH where applicable. The effect of initial MG concentration was investigated by the introduction of 1 g AFMC into different concentrations of MG dye (100-600 mg/L). Variation of time as done to investigate the effect of contact time from 10 to 120 min. All through the study, the adsorbateadsorbent system was agitated on the Orbital shaker to increase effective collision in the system. Measurement of residual concentration at maximum wavelength of 617 nm was done using double beam Libra Biochrom 5505 v1.0.4 PCB 1500 coupled with water Peltier system UV-Vis spectrophotometer.
Theory. Biosorption isotherm and kinetic modeling and statistical error validity. Equilibrium biosorption data obtained from the study of were analyzed using six of two-parameter models (Freundlich 36 , Langmuir 16 , Temkin 37 , Dubinni-Raduskevich 38 , Halsey 39 and Jovanovic 40 ). Similarly, both kinetics and mechanism models were fitted to Pseudo first-order 41 , Pseudo-second-order 42 , Elovich 43 , Fractional power 44 , Intraparticle 45 and liquid film 46 diffusion models. Estimation of the quantity adsorbed and percentage removal efficiency was done using Eqs. (4) and (5) [47][48][49] Presented in Tables 1 and 2 are the descriptions of both isotherm, kinetics, and mechanism models used in this study.
Adsorption statistical error function (ASRF) models. In most cases, determination of best fitting relationship and finalizing the best isotherm and kinetics model have always been through the use of linear correlation coefficient (R 2 ) values. However. Owing to inherent bias from this transformation, the following four rigorous statistical error function models were used: Sum of square error (SSE) 50 ; Hybrid fractional error function (HYBRID) 39 ; Nonlinear chi-square test (χ 2 ) 51 ; Marquardt's Percent Standard Deviation (MPSD) 52 , Presented in Table 2 are the equation of the Adsorption Statistical Error Function (ASRF) Models from Eqs. (24)- (27).
K L is the Langmuir isotherm constant (L/mg) related to the binding energy of adsorption.Q max is the maximum monolayer coverage capacity (mg/g), R L dimensionless separation factor indicating the nature and favourability of adsorption process. From slope and intercept of linear plot of Ce/Qe vs 1/Ce, K L and Q max were determined Freundlich Q e = K F C e logQ e = logK F + 1 nF logC e (8) C e equilibrium concentration of the MG dye adsorbate (mgL -1 ); Q e amount of MG dye adsorbed at equilibrium per unit weight of AFMC (mg g -1 ); K F Freundlich indicator of adsorption capacity,1/n F Intensity of the adsorption indicating the surface heterogeneity and favourability of the adsorption process 1/n F and K F were determined from slope and intercept of linear plot of log Qe vs log Ce Temkin Q e = RT bT ln(A T C e ) Q e = RT bT lnA T + RT bT lnC e (9) b T is the Temkin isotherm constant related to the heat of adsorption and A T is the Temkin isotherm equilibrium binding constant (L/g) R = universal gas constant (8.314 J/mol/K) T = absolute Temperature in Kelvin B = RT/b T = Constant related to heat of sorption (J/mol) obtained either from intercept or slope Q DKR is the theoretical adsorption isotherm saturation capacity (mg/g) obtained from intercept. A DkR is the D-R isotherm constant (mol 2 /kJ 2 ) related to free sorption energy obtained from the slope. Ɛ is Polanyi potential determined by the expression = RT ln(1 + 1/C e ). E is the mean adsorption free energy helpful in determining the adsorption nature (physisorption or chemisorption of the adsorption process). Q D-R and A D-R were determined from intercept and slope of linear plot of ln q e vs Ɛ 2 Jovanovic K J is Jovanovic isotherm constant (L/g) determined from the slope of plot of ln q e against C e Table 2. Kinetics and mechanism modeling of adsorption [40][41][42][43][44][45][46][47][48][49][50][51] .

Kinetic and mechanism models Linear expression Parameters nomenclature and description
Pseudo first order (PFO) log q e − q t = logq e − K1t 2.303 (15) h 1 = k 1 q e (16) q e is the quantity of adsorbate at equilibrium per unit weight of the adsorbent (mg/g), q t is the amount of adsorbed at any time (mg/g) and k 1 is the pseudo first-order rate constant (min −1 ) and h 1 initial pseudo first-order rate constant (mg/g/min). q e and k 1 were determined respectively from intercept and slope of the linear plot of loq e -q t vs t Pseudo second-order (PSO) k 2 is the pseudo second-order rate constant (min −1 ) h 2 is initial pseudo second-order adsorption rate constant (mg/g/min). q e and k 2 were determined respectively from slope and intercept of the linear plot of t/q t vs t q t is the amount of adsorbate per unit mass of adsorbent at time (t), and α and β are the constants slope and intercept of the determined from the linear plot of q t versus ln(t). ∝ is the initial adsorption rate (mg/g-min); β is the desorption constant (g/mg) during any one experiment. The slope is 1/β while the intercept is 1/β ln(αβ) Fractional power (power function) log (q t ) = log(k) + vlog(t) (20) q t is the amount of adsorbate per unit mass of adsorbent, k is a constant, t is time, and v is a positive constant (< 1). The parameters v and k are obtainable from slope and intercept of a linear plot of log (q t ) versus log (t) Intraparticle diffusion (IPD) q t = k id t 0.5 + C (21) k id is the intraparticle diffusion rate constant (mg.g −1 min 0.5 ) and C is the thickness of the adsorbent determined from slope and intercept of linear plot of q t vs t 0.5 F is fractional attainment to equilibrium and K LFD is the rate coefficient for particle-diffusion controlled process corresponding to the particle size of the adsorbent.

Results and discussion
Physicochemical characterization. Figure 1 shows the structure of Malachite green as a cationic dye.
Presented in Table S1 of the supplementary document is the physicochemical characteristics of Malachite Green (MG) indicating that it is a cationic dye having vast application. The unique physicochemical properties of AFMC were determined and summarized in Table S2. The pH determined was 6.75, surface area (1329 m 2 /g), 12% moisture content, 0.386 g/cm 3 bulk density, and approximated particle size 300 μm < Ф < 250 μm. It has been reported that for applicability, activated carbon in the range of pH 6 to 8 is acceptable. The pH of AFMC determined as 6.75 is suitable for activated carbon (AC). The amount of water bound to activated carbon is determined via the moisture content. Lower moisture content is desirable for active activated carbon because of the competition of the water vapor with the pores of AC. The moisture content of AFMC lower than commercial activated carbon (CAC) 53 is suitable. The filterability of activated carbon is determined from the bulk density. Fig. S1 is the effect of pH on biosorption of MG cationic dye onto AFMC. Ionic mobility and degree of ionization as well as the surface chemistry was influence by this operational parameter. Protonation, as well as ionic competition between H + and MG + zwitterion in aqueous solution for available sites, was observed between pH 2-5 at the acidic region. Higher quantity adsorbed and removal efficiency observed between pH 6 and 8 was due to deprotonation, low competition, and a higher aggregate of MG + . 7.425 mg/g quantity of MG was adsorbed at 100 ppm as observed at pH 6. Beyond pH 6, no further increase was observed, therefore pH 6 was chosen as an optimum pH similar to the finding of Alqadami et al. where MOF was used for adsorption of both Malachite green and methylene blue and optimum pH for highest adsorption capacity of MG was at 6.8 28 .

Effect of pH. Shown in
Effect of initial concentration. Figure S2 of the supplementary document shows the result of the effect on initial concentration on effective removal of EDC cationic MG dye using AFMC. Lower transport of the MG dye at lower concentrations led to lower adsorption due to low driving force. However, the percentage of the percentage removal efficiency increased with increase in concentration. The concentration gradient developed was due to bombardment of the MG + surrounding the active sites. It is obvious from Fig. S2 that rapid adsorption was observed at low concentration as a result of an increase in the active sites as compared to MG molecules in the bulk. Thereafter, diffusion, convection, and migration of MG molecules as a result of mass transport from the bulk lead to an increase in removal efficiency until a saturated point was reached. All the active sites were filled up at equilibrium and thereafter, no significant percentage removal efficiency observed.  Table 3a for Freundlich, Temkin, Dubinin-Raduskevich (D-R), Halsey. Equilibrium data did not fit well to Langmuir and Jovanovic considering their R 2 value less than 0.92 (Table 3). Both Freundlich and Halsey isotherm models describe the adsorption characteristic for the heterogeneous surface. The characteristics parameters of Freundlich isotherm models are K F (adsorption capacity) and 1/n F and n F (adsorption intensity) obtained from the linear plot of log Qe against log Ce. The function of the strength of adsorption of MG onto AFMC is determined from the parameter 1/n F. The value of 1/n F (2.1372) being above unity is an indication of a cooperative adsorption 54,55 . The favourability of the adsorption process of MG onto AFMC could be affirmed from the Langmuir dimensionless and separation factor (R L ). The R L value indicates the adsorption nature to either unfavorable or unfavorable. It is unfavourable if R L > 1, linear if R L = 1, favourable if 0 < R L < 1 and irreversible if R L = 0. The value of R L ranges between 0.00377 and 0.0744 and being less than one indicated favorable adsorption. There are many studies carried out on the adsorption of MG onto different adsorbents. Comparison of the Qmax monolayer capacities of adsorption of MG onto various adsorbents was presented in Table 4. Qmax of AFMC surpassed all those adsorbents compared indicating that AFMC is a better adsorbent for MG adsorption. The Dubinin-Kaganer-Raduskevich is generally applied to determine the mechanism of the MG-dye and AFMC system with a Gaussian energy distribution onto a heterogeneous surface. The R 2 > 0.98 is an indication of a better description of equilibrium data by The DKR mean energy (E) value being less than 8 kJ indicated that the mechanism is physisorption. Studies from Bello et al. on scavenging of MG onto Citrus grandis peels further supported these findings 38,56 . Statistical error validity on isotherm model. Studies have shown that the determination of the best isotherm model does not only depend on the R 2 value. Statistical validity model has been introduced to further justify the suitability of the best isotherm model to describe the adsorption process 49,59 . Table 5 has shown the values of the experimental and calculated quantity adsorbed, Qe, exp and Qe, cal, respectively. The four most used statistical validity models in adsorption studies explored are SSE, HYBRID, X 2 , and MPSD. Adsorption Statistical Error Function (ASRF) has always been the most reliable validity parameter in justifying the best isotherm model for the equilibrium studies. To determine the isotherm best model, coupled with the higher R 2 value, there must be closeness between the data of the Qe, cal, and Qe, exp alongside a low value of the ASRF 60,61 . Considering Table 5, Freundlich, Temkin, and Halsey isotherm models fit well into these conditions for fitness. From Table 4

Effect of contact time at various initial concentrations. Importance relevant parameter that controls
the transfer and build-up of charges from the bulk to the pore active site in all transfer media is the contact time. Effect of contact time was studied from 10 to 120 min at six different initial concentrations from 100 to 600 mg/L as depicted in Fig. S3 of the supplementary document. Based on the results, rapid adsorption was observed in the first 30 min dues increase attractive forces between the active sites and MG molecules as a result of van der Waals forces and electrostatic attractions. Between 60 and 90 min not significant increase in adsorption capacity and removal, efficiency was observed due to attainment of saturation and equilibrium. A fast diffusion onto the external surface of AFMC was followed by fast pore diffusion into the intraparticle matrix as a result of the par- Table 3. Isotherm models' parameters and for adsorption of malachite green onto AFMC. www.nature.com/scientificreports/ ticipation of the functional groups until equilibrium was attained where 93.09% removal efficiency was achieved. The reaction was allowed to proceed till 90 min beyond which to increase was observed as depicted in Fig. 5. This finding is supported by the report of Hamdaoui et al. 57  Kinetics and mechanism model of MG sequestration. The rate of binding of MG onto AFMC was determined by the adsorption kinetics which also helps in gaining insight into the mechanism of the sorption process. Across various concentrations from 100 to 600 mg/L, the kinetic data were fitted to the following kinetics and mechanism models: Pseudo first-order (PFO) (Fig. 3A), Pseudo second-order (PSO) (Fig. 3B), Elovich (Fig. 3C), Fractional power (power function) (Fig. 3D); Intraparticle Diffusion ( Fig. 3E) and Liquid film diffusion (Fig. 3F). Based on the evaluated data presented in Table 6, correlation coefficient R 2 of pseudosecond-order (> 0.99) is highest among all the kinetics models explored. The R 2 value is consistently higher and increases as the concentration increases. The h 2 initial pseudo-second-order adsorption rate constant increases from 23.92 to 105.26 mg/g/min suggesting a rapid kinetic process. The error bars on the kinetic plots from Fig. 3A-D showed that the kinetic models were validated using statistical error functions. The consistency of the calculated adsorption capacity (Q e , cal) with the experimental adsorption capacity (Q e , exp) coupled with lower values of the statistical error function validity data as observed in SSE, HYBRID, X 2 , and MSPD further supported the PSO as the best kinetic model in this study. This result is supported by the investigation carried out by researchers 59,60 .
Presented in Fig. 3D is the Fractional power plot for adsorption of MG onto AFMC. Considering Table 6, the parameters v and k being positive, greater than unity, and increase with the increase in concentration suggested a rapid kinetic process. The close agreement between Q e exp and Q e , cal are indications of the best fitting of the kinetic data to the fractional power model. At low concentration, the R 2 values were far away from unity, however, better regression coefficients were obtained with higher concentration indicating the applicability of the adsorbent, AFMC, to the removal of pollutant at higher concentrations of MG dye. The choice of the best fit kinetic model was adjudged not only with correlation coefficient but also with the statistical error validity functions. It has been established that the model with a higher R 2 value, nearness/closeness between Q e ,exp and Q e , cal and lower data of statistical error function, would be chosen as the best descriptive model [61][62][63][64] . Pseudo second-order fit perfectly well into this condition and thus the best kinetic model to describe the sequestration of MG dye onto AFMC. Supporting this claim is the finding of Dehbi et al. 65 . Figure 3E and F show the linear plots of Intraparticle Diffusion (IPD) and Liquid film diffusion (LFD) models. The evaluated parameters are presented in Table 7. Both the rate-controlling step and the diffusion mechanism were explored using IPD because its R 2 values were consistently higher than that of the LFD. IPD would be the only rate-determining step if its plot begins from the origin. Contrary to this, the plot of q t against t 1/2 did not begin from the origin hence IPD is not the only rate-determining step. However, the value of the thickness C of the adsorbent calculated from the IPD model being greater than zero across all concentrations indicated that the thickness of the boundary layer participated in the adsorption process. It is suggested that since boundary layer, C > 0 from the evaluated parameters in Table 7, another diffusion model may be involved in determining the rate-controlling step 65 .
As reported by Dada et al. 4,62 and Boparai et al. 61 , three definite steps involved in adsorption are: intraparticle or pore diffusion, where adsorbate molecules percolates into the interior of adsorbent particles; Liquid film or surface diffusion where the adsorbate is transported from the bulk solution to the external surface of adsorbent, and adsorption on the interior sites of the adsorbent. From this study, Pseudo second order (PSO) model best described the kinetic data as a result of a rapid adsorption process which is being supported by best R 2 values and low statistical validity models. More so, the mechanism is pore diffusion dependent. Depicted in Fig. 3G is the scheme of mechanism of adsorption of malachite green onto AFMC. Adsorption is always a surface phenomenon. This scheme shows the summary of the adsorption of MG onto AFMC. Change in morphology was noticed after adsorption showing the evidence of the percolation of the MG dye into the pores and matrix of AFMC.
Surface morphology and surface chemistry post adsorption characterization. Evidence of the adsorption process was justified by morphological characterization of AFMC before and after adsorption onto MG using scanning electron microscopy (SEM). More so, surface chemistry was investigated by functional group determination using Fourier Transform Infrared (FTIR) spectroscopy. Before adsorption, dry and particle-like-crake nature with the presence of pores is evident all through the micrographs at different magni-    www.nature.com/scientificreports/ fications as portrayed in Fig. 4A,B. However, after adsorption as depicted in Fig. 4C,D, there was the disappearance of crakes, impregnated pores with MG dye solution, and robustness of AFMC adsorbent are morphology evidence of the adsorption process. Depicted in Fig. 5A,B are the FTIR spectra of AFMC before and after adsorption. The surface chemistry of AFMC before and after adsorption was investigated using FTIR. Broadband at 3390.70 cm −1 is attributed to O-H stretching of the hydrogen bonding which disappeared after adsorption as evidence of its participation in the adsorption process 66 . Aliphatic C-H stretching band at 2920.98 cm −1 was also found to decrease after adsorption. Carbonyl group, -C=O stretching vibration attributed to the lignin aromatic groups was assigned to 1714.28 cm −1 and 1667.29 cm −1 . Ascribed to -C=C-bending of the Aromatic ring are the signals observed between 1515.10 and 1427.71 cm −1 while -CH 3 bands as a result of deformation are observed at 1372.37 cm −1 . Several bands between 1200 and 800 cm −1 are ascribed to characteristic carbohydrate bands while at 1049.93 cm −1 , C-O vibrational band assigned to cellulose is observed at 1049.93 cm −1 . The shift in bands and disappearance of functional groups confirmed their participation in the adsorption process 67 .

Conclusion
This study has investigated the efficacy of Acid Functionalized Maize Cob (AFMC) as a sustainable, cost-effective, easy, unique, efficient adosrbent ultilizing adsorption as a low-cost technique for effective removal of malachite green (cationic dye). Unique physicochemical properties of AFMC vis-à-vis high surface area (1329 m 2 /g), moisture content (12%), bulk density (0.386) enhanced the adsorptive capacity. Adsorption of malachite green onto AFMC before and after was confirmed SEM and FTIR. Effective removal of malachite green was achieved at pH 6, 10-120 min contact time, six different initial concentrations from 100 to 600 mg/L concentrations at ambient temperature. A rapid and fast kinetics was attained at 90 min with 93%% removal efficiency. Based on higher R 2 > 0.97 and lower suitable statistical validity models (SSE, HYBRID, X 2 , and MSPD), the equilibrium data were best described by Freundlich and Halsey isotherm models. The Langmuir adsorption mnolayer capacity (Qmax) of AFMC being 66.52 mg/g surpassed several adsorbents previously used for adsorption of MG. Free energy value being less than 8 kJ from DRK supported a physisorption mechanism. Based on R 2 values and statistical error validity models, the kinetic and mechanism data were best fitted to Pseudo second order and supported by intraparticle diffusion. Subsequently, consideration could be given to AFMC as propitious material for environmental remediation. www.nature.com/scientificreports/

Data availability
All data generated or analyzed during this study are included in this manuscript (and its Supplementary Information files).