Synthesis of potential bio-adsorbent from Indian Neem leaves (Azadirachta indica) and its optimization for malachite green dye removal from industrial wastes using response surface methodology: kinetics, isotherms and thermodynamic studies

The response surface methodology has been extended to study the adsorption of malachite green dye (MG) on Indian Neem leaf powder (NLP), Azadirachta indica. The study was experimented by varying the solution concentration from 10 to 100 mg/l. The removal percentage of MG was determined by spectrophotometer at wavelength of 618 nm. Design Expert 11.0 was used as a tool to study the optimal conditions of MG adsorption (applying 2-factorial interaction model of central composite design) and evaluation of interaction effects of different operating parameters including pH, time, solute concentration, temperature and adsorbent dosage. Numerical optimization helped to identify the optimal conditions for achieving the targeted dye removal of 95.493% when the pH, time, solute concentration, temperature and adsorbent amount were set at the range of 6.868, 36.4 min, 27.899 mg/l, 308.477 K, and 4.3475 g/l respectively. The experimental results indicate that the maximum adsorption capacity of NLP depends on various operating parameter like, pH, dye concentration, particle size of adsorbents and it’s activation. It was noted that adsorption capacity of the NLP increases from 33.33 to 66.72 mg/g when it was activated with dilute HCl. The higher correlation coefficient value (R2) of Langmuir isotherm 0.996 and lower p value (0.01203) indicate the fitness of the response surface 2FI model developed. Kinetic studies displayed the suitability of pseudo second order reaction for this adsorption process. The thermodynamic study show that the sorption process is exothermic and spontaneous in nature. It was observed that the adsorbent NLP in the form of fine powder are very effective for the removal of MG from its aqueous solution. The satisfactory values of regression coefficients at different temperature, and fittings of adsorption isotherms indicate that Indian NLP is a promising adsorbent for treatment of textile dyes.


FTIR
Fourier transform infrared spectrum ANOVA Analysis of variance

List of symbols
Initial concentration of malachite green solution (mg/l) C t Final concentration of malachite green solution (after adsorption) (mg/l) Q o Langmuir constant (mg/g) K l Langmuir constant (l/mg) X = Q e Surface load (mg/g) R l Separation factor Q Amount of adsorbent (g)

Introduction
The Neem tree (Azadirachta indica) of family Meliacae is native to the Indian sub-continent and it has been in use (especially its leaves and barks) since ancient periods to treat various human disorders and other hygienic activities (Bhattacharyya and Sharma 2003). From the ancient time, the medicinal and germicidal activities of this tree (A. indica) have been well described. A huge range of useful products like anti-malarial, spermicidal, anti-tuberculosis agents, antipyrehetics, anti-viral drugs anti-fungal agents, anti-allergic medicines, etc. has been extracted from Neem (Chopra et al. 1956;Sharma and Bhattacharyya 2005). It has been reported that fresh Neem leaves contain maximum 59.4% of moisture, 22.9% of carbohydrates, 7.1% of proteins, 6.2% of fiber, 3.4% of minerals and 1% of fats with other chemicals (Neem Foundation 1997). Various chemical process industries like textile, pulp and paper, paints and pigments, printing, food, cosmetics, petroleum, rubber, etc. use synthetic dyes extensively (Ergene et al. 2009;Kiran et al. 2009). Annually, over million tonnes of production of commercial dyes and pigments are taken place among the different existing dyes and pigments (Gupta 2009). Textile industries consume the dyes extensively higher compared to other industries, almost 50% of the dyes are lost during textile processing among which 10-15% are discharged to effluents (Yang et al. 2009). The effluents of dyeing industries use highly colored compounds having high chemical and biological oxygen demand (Garg et al. 2004). These dye containing effluents can harm the living systems of the natural water bodies as well as the human beings due to their carcinogenic, allergenic and toxic effect (Akar et al. 2009;Chakraborty et al. 2020). Since water is an indispensable part of our life (Halder et al. 2016;Roy et al. 2020), several physicochemical processes have been developed to reduce the impact of carcinogenic dyes on the environment. It was reported that among various selective and expensive treatment processes, adsorption is highly cost effective. Though commercial activated carbon is widely acceptable adsorbent due to its high porosity and large surface area, but it is expensive in comparison to naturally obtained bioadsorbent like NLP. Thus, effective utilization of NLP as an alternative adsorbent for the treatment of textile dying wastes and the investigation of its adsorption characteristics is a big challenge to the researchers till now.
Malachite green (MG), an N-methyl-di-amino-tri-phenylmethane dye, is water-soluble, organic cationic dye ( Fig. 1) is widely used for dyeing purposes (Chowdhury et al. 2011;Ahmad et al. 2017). This cationic dye is generally used for the coloring of cotton, wool, silk, paper, and leather and also in distilleries. It is also used as a therapeutic agent (fungicide, ectoparasiticide) and as antiseptic, but only for external applications on the wounds and ulcers (Crini et al. 2007). Some noticeable detrimental nature of MG dye are to effect on nervous system, brain and liver while consumed, to reduce the efficiency of food intake, growth and fertility rates (Raval et al. 2017). It is environmentally persistent and responsible to damage liver, spleen, kidney and heart etc. (Hameed and El-Khaiary 2008;Bulut et al. 2008). Hence, treatment of the effluent containing carcinogenic dye has become an area of interest to the researchers.
In classical methods of optimization, it is difficult to solve the complex interactions between the variables and responses since it can handle the variation of only one parameter at a time, keeping other factors constant Dey et al. 2019). Response surface methodology is one of the most user friendly, potable method to design experiments statistically. Development of mathematical model and its optimization to reduce the complicacy of the process is another importance of RSM (Khajeh et al. 2013(Khajeh et al. , 2016Khajeh and Gharan 2014;Khuri and Cornell 1996). It can provide the effects of several factors at various level with their influence on each other with the aim of optimizing the response based on the investigating factors (Mothgomery 1997;Myers and Mothgomery 2002;Jana et al. 2018). Major benefits of this statistical analytical method are reported by Jana et al. (2018) and Cochran and Cox (1992) as: 1. Minimum number of experimental runs are required instead of full experimental design at the same level. 2. It allows to visualize the interaction effects of a factor at several levels with other factors and their estimation. 3. It helps to obtain the surface variations of response that provides a good way of envisioning the parameter interaction.
Though dye removal from textile effluent have already been investigated using hard wood, saw dust and charcoal by Asfour et al. (1985), Hameed and El-Khaiary (2008) and Ferrero (2007). Moreover, performance efficiency of saw dust as low cost adsorbent has also been studied for the removal of toxic heavy metal, hexavalent chromium (Cr) and dyes by Siboni et al. (2011), Das et al. (2015 and Karunanithi et al. (2000). Similarly low-cost rice-husk has been used to study adsorption efficiency of hexavalent chromium by Mullick et al. (2017). Garg et al. (2004) and Ho and McKay (1998) explained the overview performance characteristics of timber wastes for dyes removal from textile wastes. In spite of the above-mentioned development, we have considered the following aspects as objectives of the present work: 1. To design the experimental and statistical analysis for the batch adsorption of malachite green dye using Neem leaf powder as low-cost adsorbent and observe the effects of five primary operational parameters (i.e. pH, time, solute concentration, temperature and adsorbent amount) on one output parameter i.e. percentage dye removal during the batch adsorption of synthetic wastewater solution. 2. To investigate the suitability of proposed design and its statistical analysis (ANOVA) in adsorption phenomena for its optimization. 3. To enlighten the adsorption characteristics of NLP for the treatment of MG dye based on the kinetic and thermodynamic study. 4. To investigate the optimal conditions of five operational parameters which maximize the dye removal quantity using RSM.

Adsorbate collection
Malachite green the chemical formula of MG is C 6 H 5 C C 6 H 4 N CH 3 2 2 Cl and mol. wt. of 364.11 (g/mol) used in this experiment was of analytical reagent grade and purchased from Merck Specialities Pvt. Ltd (Mumbai,40018). Different concentrations of MG solutions were prepared by dissolving the required amounts of MG dye in distilled water. Laboratory grade HCl and NaOH (Fisher Scientific, Kolkata) was used in this experiment.

Adsorbent preparation
Neem leaves were collected from local area which were washed several times, dried at room temperature till it became crispy and grounded to fine powder, sieved to 125-250 µm, washed several times with distilled water. Then about 30 g NLP were boiled in a 250 ml beaker adding 5 ml, (5 N) HCl to activate the adsorbents and to remove dark green color and soluble impurities. The Neem leaf powder were filtered and boiled 2-3 times with distilled water again to remove the color fully and dried at room temperature for 48 h. It was kept in the oven at about 60-70°C for 6-8 h, to remove the moisture completely and dry activated NLP were kept in an air tight vessel.

Adsorption study in batch process
Batch adsorption study was carried out using magnetic stirrer at constant rpm of 300 in a 250 ml Erlenmeyer flasks by adding different concentrations of MG solution (10-100 mg/l) and NLP of 3-10 g/l. Kinetic studies were performed at 308 K, 318 K and 333 K. The concentrations of MG left in supernatant solutions after different time intervals (10-120 min) were determined using a UV-Vis spectrophotometer (Simazu-1800) at wave length of 618 nm. The amount of MG adsorbed per unit mass of NLP (Q) and the dye removal efficiency (R) were calculated according to the following Eq. (1): where C o and C t (mg/l) are the initial (before adsorption) and final (after adsorption) concentration of dye solution. The effect of pH was performed by adjusting the solution using 0.1 mol/l of NaOH or 0.1 mol/l of HCl solutions.

Scanning electron microscopy (SEM)
Scanning electron microscopy (SEM) is used to study the surface topography of any solid material (Chatterjee et al. 2012). It shows better results than optical microscopy because of its high and better resolution and enhanced depth of field. SEM photographs were taken with scanning electron microscope (Zeiss EVO 10, U.K.) at the required magnification at room temperature. The working distance of 7 mm was maintained and acceleration voltage used was 5-20 kV at different magnification, with the secondary electron image (SEI) as a detector. (1) Fourier transform infra-red (FTIR) study IR spectroscopy is used for obtaining information of functional groups present on solid NLP surface (Mullick and Neogi 2016). The principle of IR spectroscopy is based on molecular vibration in crystal lattice with specific frequencies associated with internal vibrations of groups of atoms. These frequencies occur in the IR region of the electromagnetic spectrum, i.e. in the range of 500-4000 cm −1 . The Fourier transform infrared spectrometer (FT1R) employs an interferometer instead of a monochromator. It was conducted with the help of Fourier Transform Infra-Red machine (ATR, Thermo Fischer Scientific, USA). The infrared spectra were recorded in the mid IR region (i.e. 2-25 µm, 500-4000 cm −1 ) using FTIR spectrometer.

Energy dispersive X-ray (EDX)
Energy (electron) dispersive X-ray analysis (EDX) technique is used for performing elemental analysis or chemical characterization of a sample in conjunction with scanning electron microscopy (SEM). For determining elemental content, the electron-beam strikes the surface of conducting sample (SEM) (Kibami et al. 2017). Here, it was used for the elemental/chemical characterization of the pure adsorbent A. indica (NLP). Its characterization capabilities were due to the principle that each element has a unique atomic structure, thus allowing a unique set of peaks on its X-ray spectra which was shown in Fig. 3.

Model selection
Since the pattern of combinations between the independent and response variables is unknown in most of the cases, suitable approximation is required to determine. Among various approximation functions, polynomials are such functions which are used in a large extent (Nair et al. 2014). Since a simple linear equation cannot describe the interaction effect between different parameters and also unable to determine the critical point, the full quadratic second-order function is has been used expressing their various interaction terms given in the Eq. (2).
where y is the predicted response in the system, x i is the independent variable, β 0 is the constant term, β i are the linear coefficients, β ii is the squared coefficients, the β ij is the interaction coefficient, and ϵ is the error.

Experimental design
This statistical method was used to design a model using the experimental data obtained from the findings and its optimization (Arivoli et al. 2010;Sahu et al. 2009). Basically, the analysis was done by following three steps viz., (1) conducting the experiments statistically, (2) evaluating the correlation coefficients from mathematical model and predicting responses and (3) analysing the adequacy of the proposed model (Myers and Montgomery 1995). RSM helps to enumerate the relationships Eq.
(3) between output variables called responses (O) and input variables called factors (I n s) (Sahu et al. 2009).
As a response surface methodology and central composite design (CCD) has been applied in the present work. This method is suitable for fitting a quadratic surface with analyzing the interaction between the parameters (Arivoli et al. 2010).
In this present study, response surface design was developed based on three levels two factorial central composite design (CCD) having five factors for developing the CCD model using Design-Expert 11.0. In this purpose, i 1 , i 2 , i 3 , i 4 , and i 5 were taken as input parameters and O 1 was taken as response in this model. In this perspective, as explained before, the applied variables and their levels were represented in Table 1, whereas the design matrix was depicted in Table 2.
For statistical analysis, the levels of five variables i n [i 1 , i 2 , i 3 , i 4 , and i 5 ] were coded according to the following relationship in Eq. (4):  where I n is coded value of factors i n , and i 0 at the center point and δ i represents the step change.

Results and discussions
With respect to every inputs and output, we have formulated a central composite design model for toxic dye removal using adsorption process. In order to maintain the percentage dye removal of the predictive model, we have defined some statistical parameters. The predicted capability of proposed CCD model was evaluated and compared the outputs with the experimental values. In addition, the statistical parameter such as the co-relation coefficient (R 2 ) was used to compare the predicted value with experimented values of flexible modulus.

SEM analysis
SEM micrographs of the Neem leaf powder (Fig. 2) showed that the powder was an assemblage of fine particles, which did not have regular, fixed shape and size. Some of the particles may be quite small while some others may be as big as 10 µm in diameter. All the particles have broken edges with steps and kinks contributing to the uneven topography.
The surface appeared as a fibrous material with irregular macro-pores and some expanded cavities which may allow for the diffusion of the molecules through the macro-pores of the adsorbent.

EDX analysis
Due to the low X-ray intensity, EDX images usually take a number of hours to be acquired. Elements of low atomic number are difficult to detect by EDX. EDX image analysis from Fig. 3 show that the Neem leaf sample highly consists of carbon, oxygen, and calcium at varied proportions. EDX analysis of the sample (i.e. weight percentage and atomic percentage) (table inside the Fig. 3) indicate the presence of phosphorus which carried rather good adsorbent properties particularly for this Neem leaf powder. This analysis also displayed the presence of Ca, Mg, Br, Fe, and Si in a trace amount compare to the activated carbon quantity.

FTIR analysis
The principal IR bands obtained for the sample were presented in Table 3. The results obtained from FTIR spectra of the Neem leaf powder were in conformity with the large number of constituents present in the Neem leaves. The presence of polar groups on the surface has been shown to be responsible for considerable cation exchange capacity of a

Effect of adsorbent dosage and contact time
The effect of adsorbent amount and contact time on removal of MG has been shown in Fig. 5. It also explains, that the adsorption increases with contact time, and it attends the equilibrium concentration at about 120 min.
Results also indicate that percentage removal of MG increases from 90 to 98% for the increase amount of adsorbent from 3 to 10 g/l NLP. The results agree with the findings of other investigator also (Bulut et al. 2008;Chen et al. 2014a), it implies that with increase of surface area of the adsorbent, percentage of solute removal also increases consequently.

Effect of solute concentrations and time of contact
The effect of concentration of MG dye on its removal has been described in Fig. 6. It was noted that removal quantity of dye decreases with increase in solute concentration. This was because at lower concentration, the ratio of available surface area to the moles of dye was higher. But with the increase in solute concentration, this ratio became small due to unavailability of vacant spaces on the adsorbent surfaces after formation of monolayer. Similar adsorption characteristics have been reported by other researchers also Chopra et al. 1956;Ferrero 2007).

Effect of pH
Effect of pH on adsorption of MG has been shown in the Fig. 7. It show that adsorption increases from 80 to 95% with increase of pH from 2.46-9.76 indicating that adsorption of MG on NLP favors in alkaline media.

Effect of temperature
Effect of temperature on removal of MG on NLP has been studied and shown in the Fig. 8. It shows that percentage removal of MG decreases from 97 to 90% with increase of temperature from 308 to 333 K. Variation of adsorption capacity with temperature has been shown in Figs. 8 and 9. It also shows that adsorption capacity reduces from 33 to 31.0 mg/g of adsorbent (for raw NLP) indicating that this sorption process is exothermic in nature.

Studies on regeneration of adsorbent
Though neem leaves are available abundantly, yet in this work the used up adsorbent NLP was regenerated twice to investigate the efficiency of the regenerated NLP to regain its original adsorption capacity. For regeneration used up NLP were boiled with distilled water along with 5 N HCl in a 250 ml beaker 2-3 times until the color of MG dye was removed completely. Then the filtered residue was boiled with distilled water and was dried at room temperature followed by drying in the oven at 70-80 °C for 2-3 h. It was observed that initial adsorption was lower and it increases gradually and about 95% adsorption capacity may be achieved after 1 h adsorption with regenerated adsorbent. The Fig. 10 shows the adsorption property of the regenerated NLP for various solute concentration.

Langmuir adsorption isotherm
The experimental results obtained from Neem leaf powder-Malachite green adsorption process at 308 K have been correlated with the following rearranged Langmuir model of adsorption in Eq. (5): where Q o and K L are Langmuir constants related to the capacity and energy of the activated NLP. The plot of C e /Q e versus C e is shown in Fig. 11. The linear plot with satisfactory regression coefficient of 0.996 suggests the applicability of the Langmuir isotherm for the systems. It also indicates the formation of monolayer coverage at the outer surface of the adsorbent. The Langmuir constants values of Q o and K L are obtained as 66.72 mg/g and 2.77 l/mg respectively. K L is Langmuir constant related to the sorption energy, it is the affinity of adsorbent toward the adsorbate. High values of K L implies strong binding between adsorbent NLP and solute MG.

Freundlich isotherm
Applicability of the Freundlich isotherm for the present system has also been estimated by correlating the results empirically using the well-known Freundlich Eqs. (6) and (7) as: where X/M or Q e is surface load in (mg/g). The linear plot of log(X/M) versus log C e in Fig. 12   The separation factor (R l ) related to Langmuir isotherm was used to evaluate the feasibility of adsorption on adsorbent. It can be calculated by Eq. (8) (Mullick and Neogi 2016).
where C o (mg/l) was initial dye concentration and K l (l/mg) is Langmuir constant. The value of R l indicates the type of the isotherm: irreversible (R l = 0), favorable (0 < R l < 1), linear (R l = 1), unfavorable (R l > 1). The R l of MG adsorption onto NLP is 0.042 for solution having concentration of 100 mg/l. It can demonstrate the MG adsorption onto NLP is favorable.

Adsorption kinetics
Adsorption kinetics model are very important in this process of removal of complex dyes from industrial textile wastes or from aqueous solutions. Different reversible models are tested as below.

Lagergren pseudo first order model
The adsorption kinetics may be described by Pseudo First order reaction model (Siboni et al. 2011;Das et al. 2015) as expressed below in Eq. (9): where 'Q e ' is the amount of solute adsorbed at equilibrium per unit mass of adsorbent and 'Q' is the amount of solute adsorbed at any time 't', and 'k 1 ' is the rate constant. Applying boundary conditions and simplifying Eq. (9) we may obtain the following Eq. (10): The plot of log(Q e -Q) versus 't' was plotted at different temperature in Fig. 13. Result shows that correlation coefficients are in good agreement with pseudo first order reaction kinetics which are given in the Table 4.

Pseudo second order reaction kinetics
Pseudo second order reaction model is also applicable for the sorption process and equation for this reaction may be written as Eq. (11): Integrating the Eq. (11) and applying boundary conditions as t = 0 to t >0 and Q = 0 or >0 the following Eqs. (12 and 13) may be achieved as: where h = (k 2 ·Q e 2 ) and is known as initial sorption rate. The kinetics plot of t/Q versus t as per Eq. (13) were made at different concentration (Fig. 14) and the values of correlation  coefficients, 'h' (mg/g min) and 'k 2 ' (g/g min) may be calculated from the graph. Results show ( Table 5) that this process follows pseudo second order reaction kinetics better than the first order kinetics which also agree with the results of other researchers.

Adsorption thermodynamics
Thermodynamic parameter, Gibb's free energy ΔG of a process determines the spontaneous and non-spontaneous nature of a process. Hence, this thermodynamic parameters study gives the idea of the feasibility of any adsorptive reaction. Also it provides the information regarding the mechanisms involved in the adsorption process. Considering the adsorption equilibrium constant K l (l/mg) the free energy change of the process is given by the following Eq. (14): where ΔG and ΔS are standard enthalpy and entropy change, respectively, during the adsorption process and T is the absolute temperature (K). Combining Eqs. (14) and (15), it can be written as follows: where Ro is the universal gas constant = 8.314 (J/mol K). From the slopes and intercepts of the linear plots of ln K L versus 1/T (Fig. 15) the values of ΔH and ΔS were calculated as 1.143 kJ/mol and 24.775 J/mol K. The values of ΔG calculated as − 8.8, − 9.03 and − 9.337 kJ/mol at 308 indicating a spontaneous and favorable adsorption process. The negative value of enthalpy change ΔH (−1.695 kJ/mol) for the adsorption process suggested that the process was exothermic in nature and the positive values of ( ΔS ) indicate the disorder created in the system due to adsorption. Meanwhile, the positive ΔS value implies some structural changes in the adsorbent and adsorbate and the increasing randomness at the solid-solution interface during fixation of MG onto the active sites NLP. All the thermodynamic parameters mentioned above (Table 6)

Diffusion model
The intra-particle diffusion model was (a kinetic model) proposed by Weber and Morris and is used to understand the controlling mechanism of the adsorption process and can be expressed in Eq. (17) (Weber and Morris 1963) where Q t is the amount of solute adsorbed at time t (mg g −1 ), K diff was the intra-particle-diffusion rate constant (mg g −1 min 0.5 ) obtained from the slopes of Q t versus t 0.5 plot, t was the time (min), and C was the intercept (mg g −1 ). K diff was calculated from the Fig. 16 and it was found to be 0.047 mg g −1 min 0.5 .

Response surface of 2-factor interaction (2FI) model
The interaction among the different independent variables and their corresponding effect on the response on the Malachite green dye removal was investigated by analyzing the response surface analysis. Five process parameters-time, pH, temperature, adsorbent amount and solute concentration were considered as the variable parameters for the experimental studies using Neem leaf powder as adsorbent based on response

Estimation of response surface for maximum percentage of textile dye removal
Here the goodness of fit of the model was verified by the correlation coefficient (R 2 ) between the experimental and model predicted values of the response variables in Fig. 17. It was observed that the model p value (0.01203) was highly significant model terms for malachite green dye removal. A high R 2 value of 0.9678 and model F value of 2.93 implied that the results were in agreement with the model. The value of coefficient of variance 1.46% indicated good precision and reliability of the experimentally obtained value. Here, our A deq precision measured the signal noise ration. This analysis was given in Table 7. The quadratic ANOVA model for percentage malachite green dye removal from its dilute solution was given by Table 8 based on 2-factorial interaction model. The response variable which was expressed as a function of independent variables was expressed in the form of different numerical factors in Eq. (18) and its corresponding coefficient table was given in Table 9. (18)

Interaction effect of pH, time, temperature, solute concentration, and adsorbent amount
In order to study the interaction among the different independent variables and their corresponding effect on the response surface plots were represented in the Figs. 18, 19 and 23. The combined effect of time and pH on percentage dye removal was shown in the surface plot of Fig. 18. Figures show that Percentage dye removal increased with increase in time as well as pH value, which was confirmed by the experimental results within the experimental limit. This type of behavior indicates that the amount of dye removal depends on the alkalinity of the medium highly. Figure 19 illustrated the interaction effect of time and solute concentration on the removal percentage of malachite green dye. Within the experimental limit, dye removal percentage increases with the time and decreases with solute concentration of the dye (MG). Similar nature was found in the experimental results. Therefore, it can be predicted that the treatment efficiency will be higher at longer time and lower solute concentration.  The surface plot in Fig. 20 illustrated the combined effect of temperature and pH on the dye removal of textile waste water using batch adsorption process. At high pH, low temperature the dye removal percentage was also high. Figure 21 illustrated the interaction effect of adsorbent amount and temperature on the removal of MG. Within the experimental limit, removal efficiency was higher at low temperature for increase amount of Neem leaf adsorbent, during adsorption process of malachite green from its solution.
The combined effect of solute concentration and adsorbent dose on percentage dye removal was represented in Fig. 22. In this case, the dye removal percentage was maximum at higher amount of adsorbent dose and lower value of solute concentration. The interaction effect of solute conc. and pH on dye removal was shown in Fig. 23. Here, the dye removal efficiency was higher at low pH when solute concentration was at high region.

Optimization analysis using the desirability function
In this study, the desired outcomes i.e. the response functions were selected from design. The overall desirability function was achieved from the combination of all the response functions. Experimentally, it was observed that the malachite green dye removal from wastewater in batch adsorption process became 95.493%. If the pH, time, solute concentration, temperature and adsorbent amount were set within the range of 6.868, 36.4 min, 27.899 mg/l, 308.477 K, and 4.3475 g/l respectively, then the maximum desirability i.e. optimum condition was achieved which was illustrated in the desirability ramp (Fig. 24). A desirability value of 0.869 indicated the good agreement with the experimental model and design.

Validation of the model
The numerical optimization of the present study indicate that the optimum values of the different independent variables such as pH: 6.868, time: 36.40 min, temperature: 308.477 K, solute concentration:, 27.899 mg/l, and adsorbent amount: 4.3475 g/l in order to achieve the targeted malachite green removal of 95.493% which is very close to the experimental value of dye removal: 94.3004%. Therefore, the response surface optimization model can be taken as valid.

Conclusion and future prospects
Present study confirmed the potential capability of Indian Neem leaf powder (A. indica) as natural bio-adsorbent to remove MG dye from aqueous synthetic solution or industrial wastes. Results show that the maximum adsorption of MG on activated NLP was 66.72 mg/g and maximum removal of 98% dye may be achieved from 100 mg/l using 3 g/l of activated NLP. The comparison of the maximum adsorption capacity of MG dye by different low cost bioadsorbents with the NLP used in this study has been shown in Table 10. The adsorbent was characterized by Fourier Transform Infrared Spectroscopy (FTIR), scanning electron microscope (SEM) and energy dispersive X-ray (EDX)   techniques. The kinetics of adsorption followed second-order rate expression. The linear plot and the satisfactory values of correlation coefficients (R 2 ) suggested the applicability of both Freundlich's as well as Langmuir's model for this adsorption. The negative values of ΔG and enthalpy changes ΔH at 308 K indicate that this adsorption process is spontaneous and favorable. The positive values of ΔS implies that the disorder created in the system due to adsorption. It was found that the process was exothermic in nature and favored at lower temperature. The maximum adsorption of malachite green dye was found at pH: 6.868, time: 36.40 min, temperature: 308.477 K, solute concentration: 27.899 mg/l and adsorbent dose: 4.3475 g/l, the overall desirability function was achieved by estimating the interaction effects of independent variables with response variables. Thus it can be concluded that activated NLP, a low-cost adsorbent has an excellent potential to remove various carcinogenic dyes from industrial wastes also to protect the environment.