Sage Journals: Discover world-class research

Abstract

Climate change poses a real threat to the water security by altering the precipitation patterns and other weather variables, desirability function (df) was used to determine the interactive effects and optimize with the Central Composite Design (CCD) of important variables such as initial dyes concentration (C_o : 20–100 mg L^–1), solution pH (1.5–11), temperature (20–60 °C) and adsorbent dosage (0.1–1.3 g L^–1) for the removal of toxic Yellow Tartarazine (YT) and Methylene Blue (MB) onto Untreated Walnut Shells (UWS). The analysis of variance (ANOVA) was applied for the treatment of experimental data, fitted by a second-order polynomial equation and multiple regression analysis. The ideal conditions optimized according to the response surfaces are initial concentration (C_Oyt = 60 mg L^–1, pH_YT = 1.5, D_YT = 0.7 g L^–1 and T_YT = 40 °C for YT and (C_{o MB} = 60 mL^–1, pH_MB = 8.5, D_MB= 0.7 g L^–1 and T_MB = 40 °C for MB). Experimental confirmation allowed to accurately determine the maximum eliminations (76.15) for YT and (76.81%) for MB under optimal conditions. Several error analysis equations were used to measure the goodness of fit. Kinetic studies showed that the adsorption kinetic data are well modeled by a pseudo-second-order model. Among the different models used to fit the experimental isotherms, the Langmuir model gave the best correlation with capacities q_max of 55.55 and 50.00 mg/g, for YT and MB respectively at 40 ^oC. The effect of temperature on the adsorption isotherms allowed to arrive at the thermodynamic functions, the negative values of free energy (ΔG°), enthalpy (ΔH°) and entropy (ΔS°) indicate a spontaneous and exothermic adsorption of physical nature, confirmed by the activation energy (E_a) for both dyes. The elimination of pollutants by new-generation adsorbents has proven to be very effective by application of statistical methodology, Central composite design and the desirability function.

Keywords

walnut shells central composite design sustainability bioremediation dyes desirability function reusability

Introduction

Water pollution by various species (dyes, pesticides, heavy metals, etc.) is a growing threat and a problem that needs to be addressed immediately. In fact, various pollutants are dumped into natural waters in an anarchic manner, causing significant pollution. Therefore, environmental protection and water quality is a matter of concern nowadays. Dyes represent the greatest pollution and reducing their concentrations below the set limits is a challenge to address. Several sectors contribute to this pollution, mainly the textile sector due to the volume of effluents generated by synthetic dyes (Inyinbor et al., 2016). The widespread use of synthetic dyes has led to health problems due to their carcinogenic properties and therefore the colored effluents must be treated at source (Abbas, 2020a; Abbas, 2021a, 2021b, 2021c; Abbas, 2022; Abbas et al., 2014). Indeed, discharges from different sectors (paints, plastics, paper, leather and textiles) are becoming a growing concern where the physicochemical properties of water, namely color, odor and pH, are very altered where the mesomerism of molecules makes them difficult to mineralize (Yemendzhiev et al., 2009). Accelerated industrialization and a growing population generate organic compounds with high molecular weights and stable chemical structures. Dyes (YT) and (MB) are used in the textile sector, are difficult to biodegrade and appropriate treatments are necessary for their mineralization (Abbas and Trari, 2020a, 2020b). Many physico-chemical and biological techniques were used, including reverse osmosis (Cinperi et al., 2019), precipitation (Li et al., 2022), coagulation/flocculation (Abbas, 2021d), ion exchange (Sinha et al., 2021), membrane filtration (Nozad et al., 2022), adsorption (Abbas, 2020b, 2020c; Abbas et al., 2018; Abbas et al., 2019; Abbas et al., 2020; Abbas and Trari, 2020c) and photo-Fenton oxidation (Abbas, 2021e; Abbas and Trari, 2021) are currently used in the wastewater treatment.

Walnuts are single-seeded stone fruits produced by the walnut tree belonging to the family Juglandaceae. The native plant grows in the temperate mountainous regions of the Eurasian continent and the world production of walnuts was 3.9 million tonnes in 2022 (Martínez et al., 2003). The disposal of by-products from the various uses of walnuts generates annually more than 2.6 million tonnes of organic waste that can have harmful effects on the environment, such as the release of carbon dioxide and methane, two greenhouse gases. The main constituents of walnut shell (WS) are lignin (50.3%), cellulose (23.9%) and hemicellulose, with 22.4% (Jahanban-Esfahlan et al., 2019). The physicochemical properties of WS make it attractive as a biosorbent. It is characterized by chemical stability, mechanical resistance and can be exploited in adsorption (Cao et al., 2014), because of its specific surface area and high porosity and large (Kashif-Uddin and Nasar, 2020). Both raw and modified WS have been used to remove various pollutants, mainly dyes and heavy metals (Kamar et al., 2015; Miyah et al., 2018). Response surface methodology (RSM) is a statistical tool for developing, optimizing and improving industrial processes when many factors affect production, yield or the percentage of removal of a pollutant (Myers and Montgomery, 2002). RSM is used to determine the significant factors affecting an experiment; it aims to reduce the number of experimental trials while maximizing the result through the data generated that allow developing empirical models that correlate the response to experimental factors and analyze the effects of independent variables. It studies the interactive effect of process variables to build a mathematical model that accurately describes the overall process. The most efficient design used in response surface modeling is the Box-Behnken design which has some advantages; it requires few experimental points for its application (three levels per factor) and high efficiency (Ferreira et al., 2007). Several studies used successfully this design, one can cite the MB adsorption by kapok fiber treated by NaCl (Liu et al., 2012), Cr(VI) adsorption onto activated carbons (Özdemir et al., 2011) and degradation of Acid Red 274 using H₂O₂ in subcritical water (Kayan and Gozmen, 2012). The main objective of RSM is to optimize the response surface which is influenced by various parameters and quantifies the relationship between the controllable input parameters and the obtained response surfaces (Aleboyeh et al., 2008; Zarei et al., 2010). The central composite design procedure allows (i) to perform a series of experiments for a reliable measurement of the response of interest (ii) to develop a mathematical model of the second-order response surface with the best fit (iii) to determine the optimal set of experimental parameters that produce a maximum or minimum response value and (iv) to represent the direct and interactive effects of the process parameters via 2D contour plots of binary effects. The desirability function is used to confirm the optimized parameters. In this contribution, the combined effect of adsorbent dose, pH, initial dye concentration (C_o) and temperature on the removal of YT and MB from untreated Walnut Shells was investigated using Central Composite Design. The experimental data were analyzed based on a second-order polynomial model, statistically validated by doing analysis of variance (ANOVA) and a misfit test to test the significance of the model. In addition, YT and MB adsorption were also evaluated from the perspective of kinetics and isotherms. Environmental protection and global warming reduction are closely related to pollution. Therefore, the novelty of this research lies in the recycling of food waste and its valorization in water treatment, the low economic cost of this adsorbent compared to commercial activated carbons, the application of Central Composite Design and statistical method have made it possible to achieve high performance.

Materials and methods

Adsorbent preparation

The walnut shells used in this study were obtained from the Boumerdes region (Algeria). To eliminate unwanted particles, the shells were washed with tap water and distilled water. The washed mass was dried for 6 h at 100 °C. The material obtained was crushed and sieved; the particles with a size of ∼ 0.25 μm were selected and washed with boiling distilled water and dried at 100 ^oC (24 h). The product obtained did not undergo any chemical or physical treatment (Untreated Walnut Shells (UWS)) and was used in all adsorption experiments.

Characterizations of UWS

The morphology, arrangement of cavities and their dimensions in the prepared UWS adsorbent before and after dye adsorption were visualized with a scanning electron microscope (SEM, JEOL-JSM 7600F) to confirm the occupancy of the pores by MB and YT molecules. FT-IR spectroscopy allowed to identify the characteristic features on the UWS surface before and after adsorption, 5 mg of adsorbent was mixed with dry spectroscopic grade KBr and previously dried and compacted. FT-IR spectra were plotted with a Perkin Elmer 2000 equipment between 4000 and 400 cm⁻¹ for 16 times to increase the signal-to-noise ratio. The Zero Charge Point (pH_zpc) of UWS was obtained by the method reported elsewhere (Merrad et al., 2023). Briefly, 20 mL of KNO₃ solutions (0.01 M) were placed in closed conical flasks; the pH of each solution was adjusted between 2 and 14 by HCl and NaOH. Then, 0.1 g of UWS was added and the final pH was measured after 24 h under stirring at 25 °C; pHzpc corresponds to the point of intersection of the initial pH and final pH curves.

Biosorbates (Mb and Yt)

Yelow Tartrazine, C.I.1940, acid yellow 23, is an anionic dye (IUPAC name is 4, 5-dihyro-5-oxo-1-(4-sulfophenyl)-4-[(4-sulfophenyl) azo]-1H pyrazole-3- carboxylic acid trisodium salt, Merck) with a chemical formula C₁₆H₉N₄Na₃O₉S₂, 534.3 g mol^–1). Methylene Blue (MB), used as an adsorbate, is a Basic Blue 9 dye, (C.I. 52015, IUPAC name 3,7-bis (dimethylamino) phenothiazin-5-ium chloride, C.I. 52015., C₁₆H₁₈ClN₃S, M = 319.85 g mol^–1). The chemical structures of both molecules are illustrated in Figure 1.

Figure 1.

Structural formulas of MB and YT dyes.

The toxicity risks of YT (LD50: 2 mg kg^–1) and the side effects related to its use are hyperactivity, rhinitis, vision disorders. Methylene Blue (LD50: 1.18 mg kg^–1) is widely used in dyeing cotton, wood and silk and can lead to burns with permanent eye damage. Its inhalation can cause breathing difficulties (Charbonneau, 2013; Himri et al., 2011) and their elimination is of primary interest.

Residual concentrations were analyzed by linear interpolation using a calibration curve (0–10 mg L^–1) prepared from the stock solution obtained by dissolving the required amounts in distilled water. The wavelengths of YT (= 428 nm) and MB (= 675 nm) were determined by spectral scanning using a UV/Visible spectrophotometer (Jasco-V-530).

Adsorption experiments

Batch equilibration for optimization was followed according to CCD matrix. An exact amount of adsorbent was contacted to 100 mL of YT and MB solution under stirring (150 rpm) for 90 min obtained from kinetic study. The separation was performed by centrifugation (6000 rpm, 10 min) and the concentrations retained in the biosorbent phase were titrated by spectrophotometry. The experiments were realized by changing pH, dye concentration (C_o), adsorbent dose and temperature; HCl and NaOH (1.0 mol L^–1) were used to adjust the pH. The isotherms were performed under optimal conditions by changing the temperature (20, 30 and 40°C), the initial concentration (C_o: 20, 30, 40, 50 and 60 mg L^–1 for binary solutions; 40, 60, 80, 100 and 120 mg L^–1 for simple solutions). The amount of dye to be removed was calculated as follows:

q_{t} = \frac{(C_{0} - C_{t}) . V}{m}

(1)

q_{e} = \frac{(C_{0} - C_{e}) . V}{m}

(2)where q_t and q_e are the amount of dye (mg g^–1) at time t, and t_eq respectively; V the volume of the solution (0.1 L) and m is the adsorbent weight (g).

Adsorption process optimization by response surface method

Two types of variables exist in multivariate optimization: responses (dependent variables) and factors (independent variables, qualitative or quantitative). Two responses are measured: the adsorbed quantity (mg g^–1) and the removal efficiency (%). After choosing the design matrix and performing the experiments, the data are processed by a specific software that develops a model and a construction of the response surfaces. Consequently, multivariate optimization by the response surface method includes six essential steps:

Selection of independent factors and selection of desired responses.

Choice of matrix for experimental design.

Execution of adsorption experiments.

Adjustment of mathematical model.

Confirmation and validation of model.

Process optimization.

Having an idea of the measurement response is an essential step for the application of the RSM method, which explains performance and the influence of operational parameters. Two responses are studied in adsorption: removal and/or adsorbed quantity. The next step specifies the number of tests and the quality of the model developed. RSM optimization involves estimation coefficients and predicts the response in a mathematical model. In our case, a central composite design (24) allowed to evaluate the influence on adsorption of YT and MB mixed at equivalent concentrations. RSM is basically a specific mathematical tool, which combines the experiment, models the construction process and controls the impacts of operating variables (Garg et al., 2009; Zabeti et al., 2009). In this study, the significant variables in the adsorption of the two dyes (YT and MB) were considered, pH (X₁), temperature (X₂), initial concentration (X₃) and adsorbent dose (X₄), to evaluate their impacts on dye decolorization. The low, central and high levels of each variable are denoted by −1, 0 and 1, respectively (Table 1). A total of 30 tests were designed by RSM and optimized by CCD. Each test series was doubly analyzed to minimize the effects of external factors, and the average values were considered.

Table 1.

Independent variables and their levels used for central composite design.

Variable, unit		Levels
	Factors	-α	−1	0	1	₊α
pH solution	X₁	1	3.5	6	8.5	11
Temperature T (°C)	X₂	20	30	40	50	60
Initial concentration of each dye in mixture C₀ (mg L⁻¹)	X₃	20	40	60	80	100
Adsorbent dose D (g L⁻¹)	X₄	0.1	0.4	0.7	1	1.3

The percentage of dye removal was given with an accuracy of ±0.5%. The second-order polynomial response was derived to explain the system behavior:

Y = a_{o} + \sum_{i = 1}^{K} a_{i} x_{i} + \sum_{i = 1}^{k} \sum_{j = 1}^{k} a_{i j} x_{i} x_{j} + \sum_{i = 1}^{k} a_{i i} x_{i}^{2}

(3)where Y is the predicted response that actually represents the removal efficiency R (%) of dye biosorption in binary solution, a₀ is coefficient, x _i and x _j variables a _i , a _ii and a _ij are interaction coefficients for linear, quadratic and second order terms, respectively; the color removal efficiency by UWS was calculated as follows:

R (%) = \frac{(C_{0} - C_{e}) . 100}{C_{0}}

(4)

The CCD experimental matrix was generated by Minitab 16 software, also used for regression analysis of the designed model and graphical analysis of each response. The CCD matrix is established from 16 factorial points, 6 repetitions at the center and 8 axial points. The total number of experiments (= 30) is calculated from the following equation: N = 2 ^k +2k +n_c, N is the total number of runs, k is the number of independent variables, n_c is the number of repeated experiments at the center points and α the distance from the axial point to the center. ANOVA is used to analyze the statistical significance of each variable as well as the application of the regression equation obtained. Table 2 gives the experimental design matrix found by CCD as well as the corresponding response.

Table 2.

Central composite matrix, experimental response R_exp (%) and predicted response R_pred (%) for elimination of YT and MB onto UWS.

Parameters					R_exp (%)		R_pred (%)
Experimental, Run	X₁	X₂	X₃	X₄	YT	MB	YT	MB
1	3.5	30	40	0.4	70.435	49.342	70.195	50,064
2	8.5	30	40	1.0	59.190	66.114	57.923	66,347
3	8.5	50	80	1.0	58.310	62.596	58.709	62,839
4	3.5	50	40	1.0	73.566	54.952	72.431	54,991
5	3.5	50	80	1.0	74.197	52.952	73.718	53,694
6	3.5	50	80	0.4	71.821	51.128	73.203	51,034
7	8.5	30	80	1.0	57.041	61.890	56.278	62,051
8	6.0	40	60	0.7	70.443	71.129	69.536	72,040
9	8.5	50	80	0.4	59.845	66.103	58.643	66,254
10	8.5	50	40	0.4	58.897	68.036	58.382	68,871
11	8.5	50	40	1.0	58.043	66.553	58.610	66,436
12	6.0	40	60	0.7	69.004	72.781	69.536	72,040
13	8.5	30	80	0.4	55.010	63.138	56.259	63,238
14	8.5	30	40	0.4	56.883	66.831	57.722	66,554
15	8.5	40	60	0.7	60.881	76.813	60.572	74,907
16	3.5	30	80	1.0	70.053	54.771	70.406	53,935
17	3.5	30	40	1.0	69.558	55.756	70.844	55,930
18	3.5	50	40	0.4	70.611	51.048	71.734	51,352
19	3.5	30	80	0.4	70.547	48.928	69.938	49,048
20	6.0	40	60	0.7	69.864	72.813	69.536	72,040
21	6.0	60	60	0.7	69.864	57.396	69.676	56,470
22	6.0	40	100	0.7	66.283	61.668	66.244	61,686
23	6.0	40	60	0.7	69.594	74.909	70.113	75,809
24	6.0	40	60	1.3	65.053	70.985	65.661	70,896
25	6.0	20	60	0.7	65.854	54.004	65.704	54,394
26	11	50	60	0.7	47.390	70.117	47.941	70,398
27	6.0	40	60	0.7	69.721	75.094	70.113	75,809
28	6.0	40	60	0.1	65.754	68.822	64.946	68,444
29	1.0	40	60	0.7	76.149	49.094	75.475	48,823
30	3.5	30	20	0.4	68.278	46.123	68.064	45,478

R_exp (%) is determined experimentally; R_pred (%) is predicted by the proposed model.

Modeling and statistical analysis

To verify the quality of the mathematical model obtained from the experiments, we used the Student Test (T-value), Fisher Test (F-value) and the coefficient of determination (R²) using Minitab 16 software. The Student T-test is used to statistically verify the hypothesis of equality of the expectation of two random variables according to a normal distribution and an unknown variance as well as the nullity of a coefficient in the context of a linear regression. The statistical hypothesis (F-test) enables to verify the equality of two variances thanks to the ratio of the two variances which must not exceed a certain theoretical value, as given in the Fisher table. In this case, the software gives the ratio between the adjustment variance and the experimental error variance. The determination coefficient (R²) is an indicator for judging the quality of a linear regression, simple or multiple. For a value between 0 and 1, it measures the adequacy between the model and the observed data. Of course, the R² has its imperfections, but remains relatively interesting because of its simplicity.

Variance analysis

ANOVA tests the statistical significance of the experimental design by establishing the interactions between the independent variables and the response effects (Gadekar and Ahammed, 2019; Karimifard and Moghaddam, 2019). It subdivides the total variation into two components, the variation associated with the model or the experimental error (Danmaliki et al., 2017; Zhao et al., 2020). The F and p values (prob>F, confidence interval = 95%) for the model were obtained. ANOVA is a mathematical and statistical approach that subdivides the total alteration of the data set into parts of ingredients concomitant with a defined source. The latter tests the target hypotheses on the model parameters (Hu et al., 2007) and checks the statistical significance. Fisher's F test, i.e., the ratio between the mean square of the model and the residual error, allows for a comparison (Kasiri et al., 2013; Khataee et al., 2010).

Adsorption kinetic study

Many models were proposed to study the mechanisms controlling the adsorption. In this study, the experimental data of dyes adsorption are examined using a pseudo-first kinetic (Ho and McKay, 1998), the pseudo-second-order kinetic (Juang and Chen, 1997), the Elovich kinetic equation (Woodbury, 1997) and intra-particle diffusion (Weber and Morris, 1963) given respectively by:

\log (q_{e} - q_{t}) = {logq}_{e} - \frac{k_{1}}{2.303} \cdot t

(5)

\frac{t}{q_{t}} = \frac{1}{K_{2} \cdot q_{e}^{2}} + \frac{1}{q_{e}} \cdot t

(6)

q_{t} = (\frac{1}{β}) L n α . β + (\frac{1}{β}) L n t

(7)

q_{t} = K_{i n} \sqrt{t} + C

(8)where q_t (mg g^–1) is the mass of adsorbed dye onto UWS at time t (min), K₁ (min⁻¹) and K₂ (g mg^–1min^–1) the pseudo-first-order and pseudo-second-order kinetics constants respectively, α (mg g^–1 min^–1) the initial adsorption rate, β (mg g^–1) the relationship between the degree of surface coverage and the activation energy involved in the chemisorption, K_id the intra-particle diffusion rate constant (mg/g⁻¹ min^−1/2) and C (mg g^–1) the intercept.

Adsorption equilibrium isotherms

Adsorption isotherms are very important for the optimization of the adsorption system. The equilibrium relationship between adsorbent and adsorbate is best explained, and the maximum uptake capacity can be determined by the adsorption isotherm. To assess the performance of adsorbent, different equations and isotherms exist, out of which the Langmuir (Langmuir, 1918), Freundlich (Freundlich, 1906), Temkin (Temkin and Pyzhev, 1940) and Elovich (Ghaedj et al., 2013) isotherms were used that have been presented in the following equations respectively. Besides, the isotherm models were applied at optimal conditions of the parameters. The Langmuir model is the best known and most widely applied, it is represented by the nonlinear form:

q_{e} = \frac{q_{m a x} . K_{L} . C e}{1 + K_{L} . C e}

(9a)

R_{L} = \frac{1}{1 + K_{L} \cdot C_{0}}

(9b)

q_{e} = B_{T} . L n (A_{T} C_{e})

(10a)

B_{T} = R . T . q_{m a x} / E_{T}

(10b)

q_{e} = K_{F} . C_{e}^{1 / n}

(11)

\frac{q_{e}}{q_{m a x}} = K_{E} . C_{e} . e x p (- \frac{q_{e}}{q_{m a x}})

(12)

K_L (L mg^–1), A_T, B_T, ΔE_T (kJ mol^–1), K_F, n and K_E (L mg^–1) are the Langmuir, Temkin, Freundlich and Elovich constants.

Regression analysis, goodness of fit measure and model comparison

All the model parameters were evaluated by nonlinear regression using Mattlab Software. A part from the determination coefficient (R²), the chi-square statistic (χ²) was also used to measure the goodness of fit. The error functions employed are as follows:

The sum of the square of the errors

SSE = \sum_{i = 1}^{n} (q_{i, \exp} – q_{i, cal})^{2}

(13)

The determination coefficient (R²)

R^{2} = 1 - \frac{S S E}{S S T}

(14)

SST is the total of sum square

The chi-square test (χ²)

^{2} = \sum_{n = 1}^{n} [\frac{(q_{e, \exp} - q_{e, cal})^{2}}{q_{e, cal}}]

(15)

where n is the number of experimental samples, q_e,_exp and q_e,_cal (mg g^–1) are the experimental and model calculated values of adsorption capacity respectively.

Thermodynamic characterization of adsorption

The thermal effect on the adsorption capacity of UWS was investigated in the range temperature (293–313 K). The adsorption of dyes onto UWS surface implies a variation of the free energy (ΔG°) between the initial and final states. The feasibility of the reaction is defined by the variation of ΔG° (kJ/mol) at constant pressure:

Δ G \circ = Δ H \circ - T Δ S \circ

(16)

The free energy is composed of the enthalpy ΔH° which gives the energies of interactions molecules/adsorbent surface, and entropy ΔS° which expresses the modification and arrangement of the molecules in the liquid phase and on the surface. The thermodynamic behavior for the MB and YT adsorption on UWS was further investigated and the parameters ΔG°, ΔH°, and ΔS° are calculated from the relation:

Δ G \circ = - R T \ln K_{o}

(17)where K_o is the apparent equilibrium constant, the thermodynamic functions ΔH° and ΔS° have been estimated by the intercept and slope of the relation between lnK₀ and 1/T.

\ln K_{o} = - \frac{Δ H \circ}{R T} + \frac{Δ S \circ}{R}

(18)

The values of lnK₀ are deduced from equilibrium constant (K_s):

K_{o} = \frac{q_{e}}{C_{e}} \times \frac{γ_{1}}{γ_{2}}

(19) γ ₁ is the activity coefficient of the adsorbed solute, and γ ₂ the solute in the equilibrium suspension. The ratio of activity coefficients is uniform for dilute solutions. The activity coefficient → 1 for infinitely diluted solution:

{lim}_{C_{e \to 0}} \frac{q_{e}}{C_{e}} = K_{0}

(20)

The constant K_o determined from intercept, by plotting ln(q_e/C_e) versus C_e and extrapolating to C_e = 0 (Li et al., 2010). The rate constant obeys the known relation:

l n K_{2} = l n A - \frac{E a}{R T}

(21)

A is a pre-exponential factor and K₂ (g mg^–1.min^–1) is the pseudo-second-order kinetics constant at different temperatures. The activation energy (E_a) easily deduced by plots lnK₂vs T⁻¹.

Results and discussion

Analyses of the UWS surface morphology

The Scanning Electron Microscope (SEM) micrographs of UWS before and after adsorption are illustrated in Figure 2. UWS exhibits a microporous structure with different pore diameters and a rough surface with protrusions. After adsorption, the surface became smoother with significantly reduced roughness and less visible pores, thus confirming the adsorption of YT and BM both on the surface and inside the pores. Indeed, the images show an outer surface of UWS formed by more or less homogeneous cavities with different sizes and shapes. These cavities depend on the nature of the carbon and the activating agent. These cavities represent the external pores of the access channels to the inner surface of the activated carbon. A higher magnification shows that the adsorbent surface contains pores with a high probability of absorption.

Figure 2.

Microscopic images of the adsorbent (UWS) (a, b) before adsorption and (c, d) after adsorption of pollutants MB and YT.

Analyses of the UWS surface chemical

FT-IR spectra (Figure 3) elucidate the chemical structure of the film where the bands of UWS are assigned to the vibrations of C-H, C=O and O-H groups, indicating that UWS is formed by hydrocarbons, alcohol, ketone and other functional chains. The presence of bands at 3400 cm⁻¹ assigned to NH₂ and –OH stretching functional groups due to hydrogen bonds to UWS (Wang 2013), 2929 cm⁻¹ correspond to –CH and –CH₂ stretching (Zaku et al., 2011). The intramolecular bond (OH) remains unchanged before and after adsorption with high intensity. The new peak at 1654 cm⁻¹ after MB adsorption is characteristic of amide (Wang et al., 2010), while the peaks at 685, 1385 cm⁻¹ after YT adsorption are due to the C=O and C-H deformation in the aldehyde. The peaks 685 and 795 cm⁻¹ attributed to the -CH deformation bands and -OH groups, confirm the adsorption of MB. The peaks at 1068 and 1029 cm⁻¹ correspond to the C-O-C stretching vibration (Zaku et al., 2011) while 1699 cm⁻¹ before and after MB and YT adsorption is assigned to the C=O group. The bands centered at 1220 are due to the C-O stretching vibrations in alcohols, phenols or ether groups after YT adsorption.

Figure 3.

FTIR spectrum of the UWS before and after adsorption of the MB and YT.

Experimental design analysis by central composite design

The complete experimental design and results are illustrated in Table 2. Based on our results, the relation between the response and independent process variables is given by a second-order polynomial:

\begin{aligned} R_{YT} (%) & = 46.71266 + 1.892 X_{1} \\ + 0.452 X_{2} + 0.257 X_{3} \\ + 19.366 X_{4} 0.338 X_{1}^{2} 0.006 X_{2}^{2} \\ \times 0.002 X_{3}^{2} 13.326 X_{4}^{2} 0.004 X_{1} \\ \times X_{2} 0.006 X_{1} X_{3} 0.158 X_{1} X_{4} \\ + 0.002 X_{2} X_{3} + 0.024 X_{2} \\ \times X_{4} 0.004 X_{3} X_{4} \end{aligned}

\begin{aligned} R_{MB} (%) & = 97.11873 + 11.059 X_{1} \\ + 4.143 X_{2} + 0.891 X_{3} \\ + 47.937 X_{4} 0.567 X_{1}^{2} 0.051 \\ \times X_{2}^{2} 0.007 X_{3}^{2} 17.053 X_{4}^{2} \\ + 0.010 X_{1} X_{2} 0.011 X_{1} \\ \times X_{3} 2.024 X_{1} X_{4} + 0.001 X_{2} \\ \times X_{3} 0.186 X_{2} X_{4} 0.041 X_{3} X_{4} \end{aligned}

where R is the removal percent of dyes; X₁, X₂, X₃ and X₄ are the coded values of pH, temperature T (^oC), initial dye concentration (mg L^–1) and adsorbent dose (g L^–1) respectively. The Box–Behnken responses were analyzed and the results of ANOVA, and RSM were analyzed (See experimental section) for adsorption study of YT and MB (Table 3). The p-values were used to estimate the significance of coefficients. A coefficient is significant, if it is, for a given risk, very different from zero and both T and p-value were used to evaluate the adequacy of the model. The large T-value and small p-value indicate a high significance of regression models. Except coefficients of interactions for YT dye and some interactions for MB dye (X₁ X₂, X₂ X₃ and X₃ X₄), the p-value of less than 0.05 indicates that the experimental factors are significant at 95% confidence level. The resulting modified RMS models equations are the following:

\begin{aligned} R_{YT} (%) & = 46.7126 + 1.892 X_{1} + 0.452 X_{2} \\ + 0.257 X_{3} + 19.366 X_{4} 0.338 \\ \times X_{1}^{2} 0.006 X_{2}^{2} 0.002 X_{3}^{2} 13.326 X_{4}^{2} \end{aligned}

\begin{aligned} R_{MB} (%) & = - 97.11873 + 11.059 X_{1} \\ + 4.143 X_{2} + 0.891 X_{3} \\ + 47.937 X_{4} 0.567 X_{1}^{2} 0.051 \\ \times X_{2}^{2} 0.007 X_{3}^{2} 17.053 X_{4}^{2} 0.011 \\ \times X_{1} X_{3} 2.024 X_{1} X_{4} 0.186 X_{2} X_{4} \end{aligned}

Table 3.

Analysis of variance (ANOVA) for response surface quadratic model.

	Yellow Tartarazine						Methylene Blue
Source	df	Coef	T-val	Adj MS	F-val	P-val	Coef	T-val	Adj MS	F-val	P-val
Const		45.301	7.549				−93.349	−18.984
Regression	14			94.735	77.20	0.000			175.624	213.12	0.000
Linear	4			10.047	8.19	0.001			217.601	264.05	0.000
X₁	1	2.092	3.162	12.271	10.00	0.007	11.060	20.396	342.828	416.02	0.000
X₂	1	0.504	2.385	6.982	5.69	0.032	4.143	23.902	470.809	571.32	0.000
X3	1	0.236	3.076	11.612	9.46	0.008	0.891	14.140	164.754	199.93	0.000
X4	1	20.494	3.418	14.333	11.68	0.004	47.937	9.755	78.416	95.16	0.000
Square	4			35.988	29.33	0.000			266.533	323.43	0.000
X₁* X₁	1	−0.335	−9.286	105.823	86.24	0.000	−0.566	−19.163	302.626	367.23	0.000
X₂* X₂	1	−0.006	−2.778	9.468	7.72	0.015	−0.050	−28.509	669.765	812.75	0.000
X₃* X₃	1	−0.002	−4.056	20.186	16.45	0.001	−0.007	−15.469	197.204	239.30	0.000
X₄* X₄	1	−13.359	−5.533	37.571	30.62	0.000	−17.052	−8.619	61.216	74.28	0.000
Interaction	6			0.921	0.75	0.619*			8.868	10.76	0.000
X₁* X₂	1	−0.008	−0.838	0.861	0.70	0.416*	0.010	1.194	1.176	1.43	0.252*
X₁* X₃	1	−0.006	−1.128	1.562	1.27	0.278*	−0.011	−2.624	5.673	6.88	0.020
X₁* X₄	1	−0.149	−0.409	0.205	0.17	0.689*	−2.024	−6.748	37.527	45.54	0.000
X₂* X₃	1	0.002	1.612	3.190	2.60	0.129*	0.001	0.797	0.524	0.64	0.439*
X₂* X₄	1	0.003	0.043	0.002	0.00	0.966*	−0.185	−2.475	5.047	6.12	0.027
X₃* X₄	1	−0.007	−0.170	0.035	0.03	0.868*	−0.040	−1.117	1.028	1.25	0.283*
Resi- Error	14			1.227					0.824
Lack-of-Fit	11			1.466	4.16	0.134			0.879	1.41	0.434
Pure Error	3			0.352					0.624
Total	29
S			1.124			S			0.907
R²			98.69%			R²			99.53%
R², predicted			91.42%			R², predicted			98.28%
R², adjusted			97.29%			R², adjusted			99.04%

Significance: * Parameter is not significant.

According to the model, all of the independent variables (singular terms of the designed model X₁, X₂, X₃ and X₄) for both dyes (YT and MB) have a positive sign, showing a positive impact on the response, and this means that an increase in each of them increases the percentage of elimination of YT and MB. Analysis of variance allowed us to identify the significance of multiple regressions. ANOVA (Table 4) provides information on the fit as well as the significance of the factors based on the F and P values. For values: P < 0.05 and F > F.0.05, the factor is significant. High F-value and a low P-value indicate significant models supported by the quality of fit; the coefficient (R²) for a significant model should tend towards 1 (Segurola et al., 1999). For our models, R² values are equal to 0.9869 (YT) and 0.9953 (MB) for adsorption (Table 4). This means that more than 98.69% of the experimental data were explained by these models and only less than 1.3% were not. These results are confirmed by the high values of R² (0.9142 for YT and 0.9828 for MB) and the adjusted R²adj (0.9729, YT) and (0.9904, MB). The validity of the model can be verified by the correlation between the experimental data and those given by the model. The coefficients (R²) are 0.985 (YT) and 0.956 (MB), indicating a good fit (Figure 4).

Figure 4.

Plot of the experimental and predicted response. (R_exp (%) is determined experimentally, R_pred (%) is calculated according to the polynomial formula).

Table 4.

Different parameters for adsorption kinetic models for the removal of MB and YT onto UWS.

Model	Parameters	Yellow Tartarazine			Methylene Blue
Model	T (°C)	20	30	40	20	30	40
Pseudo-First-Order	Q_e,_exp (mg g^–1)	17.1888	33.700	51.142	17.5155	26.429	48.279
	K₁ (min^–1)	0.0660	0.0793	0.0724	0.0642	0.0598	0.0373
	Q_e (mg g^–1)	14.755	21.646	46.433	10.614	14.644	40.213
	R²	0.9926	0.9942	0.9728	0.9769	0.9247	0.9649
	χ²	0.346	1.324	1.894	0.875	4.990	6.275
Pseudo-Second- Order	K₂×10^–4 (gmg^–1 min^–1)	43.736	63.360	79.307	104.2755	123.482	144.340
	q_e (mg g^–1)	20.492	36.101	52.356	18.832	27.473	48.544
	R²	0.9968	0.9989	0.9996	0.9933	0.9995	0.9996
	χ²	0.019	0.011	0.037	0.009	0.02	0.021
Elovich	α (mg gmin^–1)	4.639	33.769	524.579	10.969	51.905	1342.423
	β (g mg^–1)	0.228	0.167	0.168	0.289	0.240	0.363
	R²	0.9916	0.9901	0.9673	0.9733	0.9272	0.9848
	χ²	0.073	0.017	0.014	1.770	1.697	0.345
Intraparticle diffusion	K_in (mg g^–1 min^1/2)	1.886	2.487	2.091	1.853	1.120	1.211
	C (mg g^–1)	3.764	21.879	38.180	5.449	18.201	38.118
	R²	0.9311	0.9243	0.9265	0.8901	0.9124	0.8977
	χ²	0.207	2.825	1.242	2.823	1.891	2.489

The adsorption of YT (98.5%) and MB (95.6%) is explained by the independent variables; the model does not explain only 1.5% (YT) and 4.4% (MB). It is important to respect the normality assumption for a good reliability of the results which is verified by evaluating the normal distribution of the residuals (Figure 5). For the dyes, the plot shows that the data follow a straight line, with a normal distribution and the absence of evidence of non-normality or undefined variables, which also confirms that the normality assumption is valid, and thus the validity of the proposed models.

Figure 5.

Normal evaluation of the normal probability.

Effect of interactive variable

For determining the comparative significance of each term of the model, Pareto analysis was applied using the following equation:

P_{i} = (a_{i}^{2} / \sum a_{i}^{2}) . 100

(22)

where a_i is the regression coefficient of individual variables according to Pareto analysis, the order of relative importance of the factor is as follows: Adsorbent dose (g L^–1) >> pH solution >> T (^oC) >> (MB, YT) Concentration C₀ (mg L^–1). A singular term X₄ (adsorbent dose), with the highest coefficient (a_{4 (YT)} = 19.366, a_4(MB) = 47.937) has the maximum influence on the response for YT and MB adsorption, while the binary terms X₁X₂, X₁X₃, X₁X₄, X₂X₃, X₃X₄ and X₁X_2, X₂X_3, X₃X₄ have no impact on the YT and MB elimination. X₁X_3, X₁X₄ and X₂X₄ have a negative sign, indicating a negative impact on the response, and indicates that an increase of each of them should decrease the MB removal. pH is among the most influencing parameters with a high impact on the dyes uptake. In addition to its impact on the solubility of dyes; it has the capability to replace some of positive ions in the active sites and affects the ionization of the adsorbate. Indeed, pH not only has an impact on the surface charge of the adsorbent UWS but also affects the functional groups of MB and YT. The influence of pH on the dyes adsorption onto UWS is interpreted by pHpzc (Deng et al., 2011).

The basic or acidic nature of the UWS surface controls the retention of the dye. However, the chemical properties of the adsorbent are intimately related to the functional groups on its surface. The surface charge of the adsorbent depends on the ionic strength and pH of the solution. Thus, an important characteristic of the surface is the pHpzc value (= 4.14) (Figure 6), the latter characterizes the acidity or alkalinity of the UWS adsorbent. The surface charge is positive below pHpzc where the oxygenated groups are in cationic form, thus providing a favorable medium for the adsorption of YT, an anionic dye. This is due to the electrostatic attractive forces catalyzing the removal of YT, which peaks at pH 2. On the other hand, above pHpzc, the surface of the adsorbent and the YT are negatively charged, with repulsive forces that reduce the removal of YT. MB is a cationic dye, above pHpzc, the surface of the adsorbent is negative, with a favorable adsorption of MB due to the attractive forces that increase the removal of MB, the maximum is obtained for a pH of 8; the opposite effect occurs below pH 8.

Figure 6.

Determination of charge predominance zones (pHpzc) in the UWS adsorbent.

Optimization by RSM and with desirability function

Minitab software provides a graphical route for the model through 2D contours, suitable for studying binary interactions of factors. The contour plots of the second-order polynomial equation within two independent variables (pH and T, pH and Co, pH and adsorbent dose for YT and MB, while all others are kept at fixed values).

Figure 7 provides the information on their connections and relationships, which includes the main factor and the interaction impacts of these two independent factors (Adinarayana and Ellaiah, 2002; Harrington, 1965). Examination of the 2D curves of the binary variables allows delineating of the optimal parameters for maximum dye removal efficiency (For YT: pH 1.5; T = 40°C; C₀ = 60 mg L^–1 and adsorbent dose = 0.7 g L^–1, and for MB: pH 8.5; T = 40 °C; C₀ = 60 mg L^–1 and adsorbent dose = 0.7 g L^–1).

Figure 7.

(a) The contour plots of the solution pH and temperature (time 90 min, adsorbent dosage 0.7 g/L, YT concentration 80 mg/L, agitation speed: 300 trs/min). (b) The contour plots of the solution pH and YT concentration (time 90 min, adsorbent dosage 0.7 g/L, temperature 40 ^oC, agitation speed: 300 trs/min). (c) The contour plots of the solution pH and adsorbent dosage (time 90 min, YT concentration 80 mg/L, temperature 40 ^oC, agitation speed: 300 trs/min). (d) The contour plots of the solution pH and adsorbent dosage (time 90 min, BM concentration 80 mg/L, temperature 40 ^oC, agitation speed: 300 trs/min). (e) The contour plots of the solution pH and MB concentration (time 90 min, adsorbent dosage 0.7 g/L, temperature 40 ^oC, agitation speed: 300 trs/min). (f) The contour plots of the solution pH and temperature (time 90 min, adsorbent dosage 0.7 g/L, MB concentration 80 mg/L, agitation speed: 300 trs/min).

The desirability function allows to arrive at the optimal parameters of the input variables simultaneously which determine the performance for one or more responses and takes place in two stages:

(i) Find the levels of the independent variables that simultaneously produce the desirable responses on the dependent variables. (ii) Maximize the overall desirability with respect to the controllable factors. The desirability function approach was initially introduced by Harrington (Derringer and Suich, 1980). Depending on whether one wishes to maximize, minimize or assign a target value to a particular response y_i, the different desirability functions d_i(y_i) used are those proposed by Derringer (Reddad et al., 2002). For comparative purposes, a desirability function (d = 0.8) was performed (Figure 8), the maximum dye removals proposed by the software are given at the optimal conditions obtained for both dyes: (for YT: pH = 1.808; T = 40.70 °C; C₀ = 68.48 mg L^–1 and adsorbent dose = 0.75 g L^–1, and for MB: pH = 8.474; T = 41.01 °C; C₀ = 54.745 mg L^–1 and adsorbent dose = 0.69 g L^–1).

Figure 8.

Predicted responses using desirability function for (a) YT and (b) MB removal.

Adsorption kinetics and isotherms

The kinetics of YT and MB uptake are important for conditioning them for optimized operation at large-scale batch processes. It gives the adsorbate absorption, controls the residual time of the overall process and predicts the adsorption rate and its design. A few models have been suggested to study the behavior of adsorbents and thus propose adsorption mechanisms. The slope and intercept of the ln(q_e−q_t) plots versus (t) enable us to calculate the first-order rate constants K₁ and q_ecal while the plot of t/q_e vs (t) is used for the determination of K₂ and q_ecal of the pseudo-second-order Table 4. A closer look at the kinetic modeling shows that the Elovich and pseudo-first-order models deviate from linearity, with R² coefficients far from unity. Moreover, the high error functions, between experimental and calculated values (>5%), confirm the invalidity of these models in our case. On the contrary, the values of R² (=0.999) and the lowest values (0.001<χ² <0.037) are found for the pseudo-second-order kinetic model. In addition, the q_e,exp and q_e,cal values are also close to each other. The adsorption kinetics follows a pseudo-second-order model (Denis and Tezel, 2022; Nath et al., 2021). It suggests that the removal of YT and MB is a physisorption, with electron exchange with UWS where the dye molecules are attached to the adsorbent surface by hydrogen or Van Der Waals bonds.

The transport of dyes on the particle surface was studied using the intraparticle diffusion model. The line q_t versus t^1/2 shows multi-linearity, implying a two-step process and the calculated constants K_in (mg g^–1 min^–1/2) and C (mg g^–1). The adsorption is complex but intraparticle diffusion is important in the early stages: (i) The early lines are due to intraparticle diffusion. (ii) The slopes are characteristic of the adsorption rate in the intraparticle diffusion region.

The constants of the different models, which describe the intraparticle diffusion and, based on literature reports, the C high value (>1) is synonymous with the appearance of a significant boundary layer effect within the system. Thus, the intraparticle diffusion cannot be regarded as a probable sorption mechanism in the present case.

To investigate the dyes uptake, the adsorption isotherms were used to evaluate the amount of MB and YT accumulation on the adsorbent surface and the type of interaction dye-adsorbent. The graphic representation of the linearized forms of the four isothermal models applied makes it possible to calculate the constants of the different models, grouped in. Fitness evaluation of experimental equilibrium data by different conventional models, based on the R² value as criterion, shows the applicability of the Langmuir and Temkin models for the interpretation of experimental data with a maximum adsorption capacity $q_{m a x}$ (=55.55.30 mg g^–1, R² = 0.999 for the YT) and (= 50. Mg g⁻¹, R² = 0.999 for MB); the Langmuir model was best suited to fit the experimental data. Noted that $R_{L}$ indicates the type of isotherm: Irreversible ( $R_{L}$ = 0), Favourable (0 < $R_{L}$ <1), Linear ( $R_{L}$ = 1) or Unfavourable ( $R_{L}$ > 1). In our case, the $R_{L}$ values are smaller than 1, thus confirming that the adsorption is well described by the Langmuir isotherm frequently suggests one of the following three main hypotheses:

In a first step and over a short period, it is assumed that YT and MB are transported to the external surface of the adsorbent by film diffusion at a high velocity. After saturation, YT and MB molecules penetrate into the adsorbent by intraparticle diffusion through the pores and diffusion to the internal surface until equilibrium is established, as shown in the second line. The high C-values (>1) are synonymous with the occurrence of a significant boundary layer effect within the system. Therefore, the intraparticle diffusion cannot be considered as a likely sorption mechanism in our case. Isotherms are used to evaluate the adsorbed quantities of MB and YT on the adsorbent surface and the type of “dye-adsorbent” interaction. The linearized forms of the four applied isotherm models (Figure 9) allow us to calculate the constants of the different models (Table 5). The experimental data based on the R² coefficient show the validity of the Langmuir and Temkin models with maximum adsorption q_max (= 55.55.30 mg g^–1, R² = 0.999 for YT) and (= 50 mg g^–1, R² = 0.999 for MB); the Langmuir model is appropriate to model the experimental data. Note that R_L indicates the type of isotherm: Irreversible (R_L = 0), Good (0 < R_L < 1), Linear (R_L = 1) or Bad (R_L > 1). In our case, R_L is less than 1, clearly indicating that the adsorption is suitably described by the Langmuir modem frequently, with one of the following three main hypotheses:

The adsorption sites have the adsorption energy,

The adsorbate occupies only one site on the surface, with no interaction “adsorbate-molecules”,

The adsorption is mainly confined to the adsorbent surface.

Figure 9.

Isotherms adsorption for the different models in the optimum conditions.

Table 5.

Different parameters for adsorption isotherm models for the removal of MB and YT onto UWS at temperature 40 °C.

Model	Parameters	YT	MB
Langmuir	q_max (mg g^–1)	55.55	50.00
	K_L (L mg^–1)	0.529	0.307
	R ²	0.999	0.999
	χ²	0.009	0.009
Freundlich	K_F (mg g^–1) (L^1/nmg^–1/n)	16.527	11.588
	1/n	0.398	0.457
	R ²	0.889	0.908
	χ²	6.726	6.082
Temkin	A_T	7.44	3.797
	B_T	10.06	9.71
	E (kJmo^–1l)	14.349	13.409
	R ²	0.983	0.989
	χ²	1.623	1.241
Elovich	Q _max (mg g^–1)	17.54	21.276
	K _E	2.963	1.09
	R ²	0.89	0.895
	χ²	13.522	13.378

Effect of contact time and temperature

Figure 10 shows a rapid adsorption that saturates after 90 min which characterizes the formation of a monolayer. The adsorption capacity increases with Co, due to the diffusion of the dye towards the adsorbent, and which is accelerated by the concentration gradient. It can be predicted that the adsorption of dyes on UWS occurs in three stages:

Rapid adsorption of the dye due to the free sites on the UWS surface which reflects a linear increase in the adsorption capacity (0 to 30 min).

Slowing down of the adsorption (30–90 min), with a decrease in the adsorption capacity attributed to a lower concentration of dye and to the lower number of available sites

A tendency towards saturation of the unoccupied sites (90–100 min), reflected by a flattened region. The adsorption reaches a maximum within 60 min and then tends towards a constant value, where no molecule is adsorbed with equilibrium at 90 min.

Figure 10.

Effect of contact time and temperature for MB and YT adsorption.

The variation of the adsorption of YT and MB at equilibrium q as a function of temperature is illustrated in Figure 10. Increasing the temperature from 293 to 313 K enhances the adsorption (4.29 and 5.00 mg g^–1 for YT) to (5.11 and 6.89 mg g^–1 for MB) and this shows the endothermic nature of the adsorption for YT and MB. They also show that if their solution concentrations are high, there are very few molecules that diffuse to the surface of the unoccupied sites, which leads to an increase in the retention of MB and YT.

Thermodynamic properties modeling studies

The elimination dyes on adsorbent surface implies a variation of the free energy between the initial and final states. The adsorption of UWS increases with increasing temperature in the range (293–313 K) indicating a favorable adsorption for YT and MB (Figure 10) (Guo et al., 2020). Temperature has a dual effect, known to increase the mobility and the diffusion rate of molecules across the boundary layer and in the internal pores of the adsorbent particle, due to the decreased viscosity of the solution (Onal et al., 2007). The ratio of activity coefficients was assumed to be uniform in dilute solutions. As the dyes concentration in the solution approaches zero, the activity coefficient tends to unity. The constant K_e was determined from the plot ln(q_e/C_e) versus C_e. The functions ΔH^o, ΔS^o and ΔG^o (Figure 11(a)) obtained from the slope and intercept of the plots lnK_o versus 1/T (Equations 18 and 16), are listed in Table 6. Furthermore, ΔH° (−59.59 and −107.19 kJ mol^–1) for YT and MB indicate that adsorption is controlled by physical adsorption.

Figure 11.

Estimation of thermodynamic parameters.

Table 6.

Estimation of thermodynamic parameters.

Dyes	T (K)	LnKe	LnK₂ (g mg^–1min^–1)	ΔG^o (kJ mol^–1)	ΔS^o (kJ K^–1mol^–1)	ΔH^o (kJ mol^–1)	R ²	Ea (kJ mol^–1)	R ²
YT	293	11.96	−5.4323	−148.036	−0.302	−59.55	0.998	22.75	0.970
	313	12.81	−5.0615	−154.076
	333	13.52	−4.8371	−160.123
MB	293	10.32	−4.5634	−239.616	−0.452	−107.18	0.999	12.40	0.999
	313	11.75	−4.3943	−248.656
	333	13.13	−4.2382	−257.696

The standard entropy (ΔS°) is used to describe the randomness of the adsorption; the negatives values (−0.302 and −0.452 kJ mol^–1K^–1) for YT and MB imply an ordered nature of the dyes adsorption. As the adsorption time increases, the dyes move from an ordered state within the solution, resulting in an increased rate of diffusion and removal. The negative free energy (ΔG°) at the investigated temperatures indicates a spontaneous and favorable YT and MB elimination (Joseph et al., 2018) for the dyes.

The linearized form (Equation 21), LnK₂ = f (1/T) (Figure 11(b)) enabled us to calculate the activation energies (E_a) (Table 6). The energies E_a (12.40 kJ mol^–1 for MB, and 22.75 kJ mol^–1 for YT), are characteristic of a physisorption since the E_a value for the chemical reaction is in the range (65–250 kJ mol^–1) (Al-Rashed and Al-Gaid, 2012), this result is in good agreement with the obtained results (ΔH° = −59.55 kJ mol^–1 for MB and −107.18 kJ mol^–1 for YT), these values are less than 20 kcal mol^–1, thus confirming a physisorption process.

Performance and regeneration of the UWS

In order to reveal the efficiency of our adsorbent for MB and YT removal, we undertook a comparison with other adsorbents (Table 7), where the maximum Langmuir adsorption capacities (q_max) are reported. We can see that the adsorption of MB (55.55 mg g^–1) and YT (50.00 mg g^–1) is well positioned compared to other works. Due to pHpzc, UWS is an attractive adsorbent for basic and acidic dyes. Other factors were optimized such as pH, temperature, stirring speed, adsorbent dose, dye concentration and contact time.

Table 7.

The maximum adsorption capacity for Methylene Blue and Yellow Tartarazine of various adsorbent materials.

Adsorption of Methylene Blue onto adsorbent	q_max (mg g^–1)	Ref
Activated Carbon from Rise Husk	09.83	(Sharma and Uma, 2010)
Natural Core-Shell structure AC derived from LitseaGlutinosa Seed	24.68	(Sharma and Uma, 2010)
PenicilineGlabrum	16.67	(Bouras et al., 2021)
Pristine Seeds Nigella Saltiva L.	18.79	(Thabed et al., 2021)
AsperillusCarbonarius	21.88	(Dao et al., 2021)
Sawdust Bbiohar	27.47	(Nath et al., 2021)
Carica Papaya Wood	32.25	(Rangabhashiyam et al., 2018)
Passifloraedulis. f. FlaviscarpaDegener	44.70	(Pavan et al., 2008)
PaspalumMaritimum	56.18	(Siva et al., 2019)
Papaya Bark Fiber	66.67	(Nipa et al., 2023)
Palm Spath	74.29	(Djelloul et al., 2021)
Sawdust derived Biohar Modified using Okra Nucilage	78.13	(Nath et al., 2021)
TerminaliaCatappa Shell	86.22	(Hivira et al., 2021)
Untreated Walnut Shell (UWS	50.50	This study

Adsorption of Yellow Tartarazine onto adsorbent	q_max (mg g^–1)	Ref
Botton Ash	10.10	(Mittal et al., 2006)
De-Oiled Soya	12.90	(Mittal et al., 2006)
Hen Feather	74.81	(Mittal et al., 2007)
Soil (Romanium)	01.22	(Caliman et al., 2009)
Sawdust	04.71	(Banerjee and Chattopadhyaya,2017)
Chitin	30.00	(Dotto et al., 2012)
Soil (Negeran)	83.33	(Dawodu and Akpomie, 2016)
Plant derived Activated Carbon (AC)	46.37	(Torres-Perez et al., 2018)
Cassava Sievate Biomass	20.83	(Chukwuemeka-Okorie et al., 2021)
Untreated Walnut Shell(UWS)	55.55	This study

The regeneration of adsorbents is of utmost importance for the removal of dyes, not only for its reusability but also for its viability. Desorption is imperative and constitutes an intermediate step for new cycles. For the regeneration of adsorbents, some techniques are currently used such as electrochemistry, thermal and chemical methods. We chose the chemical method for economic reasons (Momina and Kafeel, 2023, 2024). For this, three different solvents: H₂O, NaOH and HCl were used (Figure 12). The best desorption was obtained in HCl solution with a rate of 70.56%, followed by NaOH (58.24%) and H₂O (35.23%. 2 cycles (R₁ = 70.58%, R₂ = 45.26%) were tested because of physisorption.

Figure 12.

Regeneration of the adsorbent.

Adsorption mechanism probable

The possible mechanism followed in the MB and YT removal by using UWS according to the kinetic results is shown in (Figure 13). The adsorption mechanism includes π-π stacking, hydrogen bonding and electrostatic interaction. UWS has a positive charge at below pHpzc (=4.14), on its surface and YT is an anionic dye, and for pH>pHpzc while it acquires a negative charge on its surface and MB is a cationic dye leads to interaction between UWS and the molecules. The OH⁻ions present on the UWS surface are involved in the hydrogen bonding of N, S, O atoms of MB and YT molecules as evidenced from the FT-IR analysis, the broad hydroxyl peak in UWS before adsorption is decreased due to the involvement of -OH groups in the adsorption mechanism. Also, the adsorption is catalyzed by π-π interactions among the structure of UWS and the aromatic rings of MB and YT molecules and these possible mechanisms facilitate the MB and YT adsorption.

Figure 13.

Probable mechanism for adsorption of MB and YT dyes by UWS.

Conclusions and future work

Globally, there is a need for a sustainable approach and the development of efficient processes for water treatment. Dye removal by new-generation adsorbents has been shown to be highly effective. In this study, the statistical methodology, Response Surface Design, was found to be effective and reliable in optimizing the adsorption conditions, which showed significant effects on the removal of YT and MB. Response surface plots were used to estimate the interactive effects of four independent variables (initial dye concentration, pH, temperature, and adsorbent dose) on the adsorption percentage. The second-order mathematical model was developed by regression analysis of experimental data obtained from 30 batch tests. The optimum conditions according to the response surfaces were evaluated using the desirability function (C_YT = 60 mg L^–1, pH_YT = 1.5, D_YT = 0.7 g L^–1) and (T_YT = 40 °C and C_MB = 60 mg L^–1, pH_MB = 8.5, D_MB = 0.7 g L^–1 and T_MB = 40 °C). A confirmation test was undertaken to determine the accuracy and reliability of the optimization with maximum removals of YT (76.149%) and MB (76.81%). Error analyses were used to measure the goodness of fit. The adsorption kinetic data obey a pseudo-second-order model. The Langmuir, Freundlich, Temkin and Elovich equations were used to fit the experimental isotherms. The Langmuir model gave a better correlation with q_max capacities (55.55 mg g^–1, YT) and (50.00 mg g^–1, MB). The free energy (ΔG°); enthalpy (ΔH°) and entropy (ΔS°) predict the character of dye adsorption. Their negative values predict spontaneous and exothermic adsorption with physisorption for both dyes. The present study showed that UWS requires chemical and physical activations with ZnCl₂ as activating agent for high affinity towards cationic and anionic pollutants through increasing porosity.

Footnotes

Acknowledgements

The authors gratefully acknowledge support from the Laboratory of Applied Chemistry and Materials (LabCAM), University of M’hamed Bougara of Boumerdes, Avenue de l'Indépendance Boumerdes, 35000. Algeria and the Laboratory of Storage and Valorization of Renewable Energies, Faculty of Chemistry (USTHB) Algeria.

Author contributions

M. Trachi and M. Abbas: preparation, conceptualization, formal analysis investigation writing-original draft and methodology. M. Trari and M. Abbas: validation data curation, supervision and writing-review and editing. All authors have read and agreed to the published version of the manuscript.

Data availability

All data generated or analyzed during this study are included in this published article. No AIF related data was generated in this study.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Moussa Abbas

Mohamed Trari

References

Abbas

(2020a) Performance of apricot stone to remove dyes from aqueous solutions – equilibrium, kinetics, isotherms modeling and thermodynamic studies. Materials Today: Proceedings 31: 437–443.

Abbas

(2020b) Experimental investigation of titanium dioxide as an adsorbent for removal of Congo red from aqueous solution, equilibrium and kinetics modeling. Journal of Water Reuse and Desalination 10(3): 251–266.

Abbas

(2020c) Removal of brilliant green (BG) by activated carbon derived from medlar nucleus (ACMN) – kinetic, isotherms and thermodynamic aspects of adsorption. Adsorption Science and Technology 38(9–10): 464–482.

Abbas

(2021a) Mass transfer processes in the adsorption of lead (Pb²⁺) by apricot stone activated carbon (ASAC): Isotherms modeling and thermodynamic study. Protection of Metals and Physical Chemistry of Surfaces 57: 687–698.

Abbas

(2021b) Modeling of adsorption isotherms of heavy metals onto apricot stone activated carbon: Two-parameter models and equations allowing determination of thermodynamic parameters. Materials Today: Proceedings 43: 3359–3364.

Abbas

(2021c) Factors influencing the adsorption and photocatalysis of direct red 80 in the presence of a TiO₂: Equilibrium and kinetics modeling. Journal of Chemical Research 45(7-8): 694–701.

Abbas

(2021d) Contribution of the electrocoagulation method in the removal of fluorine ions from aqueous solutions. Desalination and Water Treatment 226: 177–183.

Abbas

(2021e) Photolysis of bromophenol blue in aqueous solution under UV-irradiation: Optimization of the parameters influencing the kinetics of degradation. Desalination and Water Treatment 215: 167–173.

Abbas

(2022) Adsorption of methyl green (MG) in aqueous solution by titanium dioxide (TiO₂): Kinetics and thermodynamic study. Nanotechnology for Environmental Engineering 7(1): 713–724.

10.

Abbas

Trari

(2020a) Photocatalytic degradation of asucryl red (GRL) in aqueous medium on heat-treated TiO₂ powder - effect of analytical parameters and kinetic modeling. Desalination and Water Treatment 180: 398–404.

11.

Abbas

Trari

(2020b) Contribution of adsorption and photo catalysis for the elimination of black eriochrome (NET) in an aqueous medium-optimization of the parameters and kinetics modeling. Scientific African 8: e00387.

12.

Abbas

Trari

(2020c) Photocatalytic degradation of asucryl red (GRL) in aqueous medium on heat-treated TiO2 powder – effect of analytical parameters and kinetic modeling. Desalination and Water Treatment 180: 398–404.

13.

Abbas

Trari

(2021) Contribution of photocatalysis for the elimination of methyl orange (MO) in aqueous medium using TiO2 catalyst, optimization of the parameters and kinetics modeling. Desalination and Water Treatment 214: 413–419.

14.

Abbas

Kaddour

Trari

(2014) Kinetic and equilibrium studies of cobalt adsorption on apricot stone activated carbon. Journal of Industrial and Engineering Chemistry 20(3): 745–751.

15.

Abbas

Aksil

Trari

(2018) Removal of toxic methyl green (MG) in aqueous solutions by apricot stone activated carbon- equilibrium and isotherms modeling. Desalination and Water Treatment 125: 93–101.

16.

Abbas

Harrache

Trari

(2019) Removal of gentian violet in aqueous solution by activated carbon equilibrium, kinetics, and thermodynamic study. Adsorption Science and Technology 37(7–8): 566–589.

17.

Abbas

Harrache

Trari

(2020) Mass-transfer processes in the adsorption of crystal violet by activated carbon derived from pomegranate peels: Kinetics and thermodynamic studies. Journal of Engineered Fibers and Fabrics 15: 1–11. DOI: https://doi.org/10.1177/1558925020919847.

18.

Adinarayana

Ellaiah

(2002) Response surface optimization of the critical medium components for the production of alkaline protease by a new isolated Bacillus sp. Journal of Pharmacy and Pharmaceutical Sciences 5(3): 272–278.

19.

Aleboyeh

Daneshvar

Kasiri

(2008) Optimization of C.I. acid red 14 azo dye removal by electrocoagulation batch process with response surface methodology. Chemical Engineering and Processing: Process Intensification 47(5): 827–832.

20.

Al-Rashed

Al-Gaid

(2012) Kinetic and thermodynamic studies on the adsorption behavior of rhodamine B dye on duolite C-20 resin. Journal of Saudi Chemical Society 16(2): 209–215.

21.

Banerjee

Chattopadhyaya

(2017) Adsorption characteristics for the removal of a toxic dye, tartrazine from aqueous solution by a low-cost agricultural by–product. Arabian Journal of Chemistry 10(2): S1629–S1638.

22.

Bouras

Isik

Arikan

, et al. (2021) Biosorption characteristics of methylene blue dye by two fungal biomasses. International Journal of Environmental Studies 78(3): 365–381.

23.

Caliman

Apostol

Bulgariu

, et al. (2009) Sorption of acid yellow 23 from aqueous solution onto soil. AfinidAd 544: 465–473.

24.

Cao

Lin

Fang

, et al. (2014) A new absorbent by modifying walnut shell for the removal of anionic dye: Kinetic and thermodynamic studies. Bioresource Technology 163: 199–205.

25.

Charbonneau

(2013) Toxicité sérotoninergique résultant d’une interaction médicamenteuse entre le bleu de méthylène et les inhibiteurs de la recapture de la sérotonine. Canadian Journal of Hospital Pharmacy 66(4): 241–252.

26.

Chukwuemeka-Okorie

Ekuma

Akpomie

, et al. (2021) Adsorption of tartrazine and sunset yellow anionic dyes onto activated carbon derived from cassava sievate biomass. Applied Water Science 11(27): 1–8..

27.

Cinperi

Ozturk

Yigit

, et al. (2019) Treatment of woolen textile wastewater using membrane bioreactor, nanofiltration and reverse osmosis for reuse in production processes. Journal of Cleaner Production 223: 837–848.

28.

Danmaliki

Saleh

Shamsuddeen

(2017) Response surface methodology optimization of adsorptive desulfurization on nickel/activated carbon. Chemical Engineering Journal 313: 993–1003.

29.

Dao

Hoang

, et al. (2021) Natural core-shell structure activated carbon beads derived from litseaglutinosa seed for removal of methylene blue: Facile preparation, characterization, and adsorption properties. Environmental Research 198: 110481.

30.

Dawodu

Akpomie

(2016) Evaluating the potential of Negerian soil as an adsorbent for tartrazine dye: Isotherm, kinetic and thermodynamic studies. Alexandria Engineering Journal 55(4): 3211–3218.

31.

Deng

, et al. (2011) Adsorption of methylene blue on adsorbent materials produced from cotton stalk. Chemical Engineering Journal 172(1): 326–334.

32.

Denis

Tezel

(2022) A novel biowast based biosorbent material for effective purification of methylene blue from water environment. International Journal of Phytoremediation 24(12): 1243–1250.

33.

Derringer

Suich

(1980) Simultaneous optimization of several response variables. Journal of Quality Technology 12(4): 214–219.

34.

Djelloul

Hamdaoui

Alghyamah

(2021) Batch biosorption of the methylene blue its aqueous solution by palm spathe: Kinetic, isotherm, and thermodynamic studies. Desalination and Water Treatment 231: 389–397.

35.

Dotto

Vieira

MLG

Pinto

LAA

(2012) Kinetics and mechanism of tartrazine adsorption on chitin and chitosan. Industrial and Engineering Chemistry Research 51: 6599–6608.

36.

Ferreira

Bruns

Ferreira

, et al. (2007) Box– behnken design: An alternative for the optimization of analytical methods. Analytica Chimica Acta 597(2): 179–186.

37.

Freundlich

HMF

(1906) Uber die adsorption in losungen. Zeitschriftfür Physikalische Chemie 57: 385–470.

38.

Gadekar

Ahammed

(2019) Modeling dye removal by adsorption onto water treatment residuals using combined response surface methodology-artificial neural network approach. Journal of Environmental Management 231: 241–248.

39.

Garg

Kaur

Sud

, et al. (2009) Removal of hexavalent chromium from aqueous solution by adsorption on treated sugarcane bagasse using response surface methodological approach. Desalination 249(2): 475–479.

40.

Ghaedj

Karimi

Barrazzch

, et al. (2013) Removal of reactive orange 12 from aqueous solutions by adsorption on tin sulfide nanoparticle loaded on activated carbon. Journal of Industrial and Engineering Chemistry 19(3): 756–763.

41.

Guo

Jiang

, et al. (2020) Synthesis of MgO/Fe₃O₄ nanoparticles embedded activated carbon from biomass for high-efficient adsorption of malachite green. Materials Chemistry and Physics 240: 122240.

42.

Harrington

(1965) The desirability function. Industrial Quality Control 21: 494–498.

43.

Himri

Bellahcen

Souna

, et al. (2011) A 90 day oral toxicity study of tartrazine, a synthetic food dye, in Wistar rats. International Journal of Pharmacy and Pharmaceutical Sciences 3(3): 159–169.

44.

Hivira

Ighalo

Aziz

, et al. (2021) Terminaliacatappa shell as low cost biosorbent for the removal of methylene blue from aqueous solutions. Journal of Industrial and Engineering Chemistry 97(3): 188–199.

45.

McKay

(1998) Kinetic models for the sorption of dye from aqueous solution by wood. Journal of Environmental Science and Health, B 76: 183–191.

46.

IMC

Chem

(2007) Comparative study of various magnetic nanoparticles for Cr(VI) removal. Separation and Purification Technology 56(3): 249–256.

47.

Inyinbor

Adekola

Olatunji

(2016) Kinetics, isotherms and thermodynamic modeling of liquid phase adsorption of rhodamine B dye onto raphiahookerie fruit epicarp. Water Resources and Industry 15: 14–27.

48.

Jahanban-Esfahlan

Ostadrahimi

Tabibiazar

, et al. (2019) A comprehensive review on the chemical constituents and functional uses of walnut (Juglans spp.) husk. International Journal of Molecular Sciences 20(16): 1–37.

49.

Joseph

Chinonye

Oluchukwu

, et al. (2018) Sorption studies of dyestuffs on low-cost adsorbent. Asian Journal of Physical and Chemical Sciences 5(3): 1–19.

50.

Juang

Chen

(1997) Application of the Elovich equation to the kinetics of metal sorption with solvent-impregnated resins. Industrial and Engineering Chemistry Research 36(3): 813–820.

51.

Kamar

Nechifor

Mohammed

, et al. (2015) Removal of lead and cadmium ions from aqueous solution using walnut shells as low-cost adsorbent materials. Revista de Chimie 66(5): 615–620.

52.

Karimifard

Moghaddam

MRA

(2019) Corrigendum to “application of response surfacemethodology in physicochemical removal of dyes from wastewater: A critical review” (Sci. Total Environ. 2018, 640, 772). Science of The Total Environment 650(prt1): 696.

53.

Kashif-Uddin

Nasar

(2020) Walnut shell powder as a low-cost adsorbent for methylene blue dye: Isotherm, kinetics, thermodynamic, desorption and response surface methodology examinations. Scientific Reports 10(16): 1–13.

54.

Kasiri

Modirshahla

Mansouri

(2013) Decolorization of organic dye solution by ozonation; optimization with response surface methodology. International Journal of Industrial Chemistry 4(1): 3.

55.

Kayan

Gozmen

(2012) Degradation of acid red 274 using H₂O₂in subcritical water: Application of response surface methodology. Journal of Hazardous Materials 201–202: 100–106.

56.

Khataee

Fathinia

Aber

, et al. (2010) Optimization of photocatalytic treatment of dye solution onsupported TiO₂ nanoparticles by central composite design: Intermediates identification. Journal of Hazardous Materials 181(1-3): 886–897.

57.

Langmuir

(1918) The adsorption of gases on plane surfaces of glass, mica and platinum. Journal of the American Chemical Society 40(9): 1361–1403.

58.

Yue

, et al. (2010) Equilibrium, thermodynamics and process design to minimize adsorbent amount for the adsorption of acid dyes onto catoinic polymer- loaded bentonite. Chemical Engineering Journal 158(3): 489–497.

59.

Rui

, et al. (2022) Synergetic removal of oppositely charged dyes by co-precipitation and amphoteric self-floating capturer: Mechanism investigation by molecular simulation. Chemosphere 296: 134033.

60.

Liu

Wang

Zheng

, et al. (2012) Adsorption of methylene blue by kapok fiber treated by sodium chlorite optimized with response surface methodology. Chemical Engineering Journal 184: 248–255.

61.

Martínez

Moiraghi

Agnese

, et al. (2003) Making and some propertiesof activated carbon produced from agricultural industrial residues fromAargentina. Anales des la Asociacion Quimica Argentina 91(4): 103–108.

62.

Merrad

Abbas

Trari

(2023) Adsorption of malachite green onto walnut shells: Kinetics, thermodynamic, and regeneration of the adsorbent by chemical process. Fibers and Polymers 24(3): 1067–1081.

63.

Mittal

Kurup

(2006) Adsorption isotherm kinetics and column operations for the removal of hazardous dye, tatrazine from aqueous solution using waste materials- bottom ash and oiled soya, as adsorbents. Journal of Hazardous Materials 136(3): 567–578.

64.

Mittal

Kurup

Mittal

(2007) Freundlich and Langmuir adsorption isotherms and kinetics for the removal of tartrazine from aqueous solution using hen feathers. Journal of Hazardous Materials 146(1–2): 243–248.

65.

Miyah

Lahrich

Idrissi

, et al. (2018) Adsorption of methylene blue dye from aqueous solutions onto walnut shells powder: Equilibrium and kinetic studies. Surfaces and Interfaces 11: 74–81.

66.

Momina

Kafeel

(2023) Feasibility of the adsorption as a process for its large scale adoption across industries for the treatment of wastewater: Research gaps and economic assessment. Journal of Cleaner Production 388: 136014.

67.

Momina

Kafeel

(2024) Synthesis of biodegradable sodium alginate-based carbon dot-nanomagnetic composite (SA-FOCD) for enhanced water remediation using ANN modelling, RSM optimization, and economic analysis. International Journal of Biological Macromolecules 263(1): 130253.

68.

Myers

Montgomery

(2002) Response Surface Methodology – Process and product optimization using designed experiments (2nd ed.). New York: John Wiley and Sons.

69.

Nath

Saikia

Goutam

, et al. (2021) Removal of methylene blue from water using okra (Abelmoschusesculentus L.) mucilage modified biohar. Bioresource Technology Reports 14: 100689.

70.

Nipa

Shefa

Parvin

, et al. (2023) Adsorption of methylene blue on papaya bark fiber: Equilibrium, isotherm and kinetic perspective. Results in Engineering 17: 100857.

71.

Nozad

Marjani

Mahmoudian

(2022) A novel and facile semi-IPN system in fabrication of solvent resistant nano-filtration membranes for effective separation of dye contamination in water and organic solvents. Separation and Purification Technology 282: 120121–120135..

72.

Onal

Akmil-Basar

Sarıcı-Ozdemir

(2007) Investigation kinetics mechanisms of adsorption malachite green onto activated carbon. Journal of Hazardous Materials 146(1–2): 194–203.

73.

Özdemir

Duranoğlu

Beker

, et al. (2011) Process optimization for Cr(VI) adsorption onto activated carbons by experimental design. Chemical Engineering Journal 172(1): 207–218.

74.

Pavan

Lima

Dias

, et al. (2008) Methylene blue biosorption from aqueous solution by yellow passion fruit wast. Journal of Hazardous Materials 150(3): 703–712.

75.

Rangabhashiyam

Lata

Balasubramanian

(2018) Biosorption characteristic of methylene blue and malachite green simulated wastewater onto Carica papaya wood biosorbent. Surfaces and Interfaces 10: 197–215.

76.

Reddad

Gérente

Andres

, et al. (2002) Adsorption of several metal ions onto a low-cost biosorbent: Kinetic and equilibrium studies. Environmental Science and Technology 36(9): 2067–2067.

77.

Segurola

Allen

Edge

, et al. (1999) Design of eutectic photoinitiator blends for UV/visible curable acrylated printing inks and coatings. Progress in Organic Coatings 37(1–2): 23–37.

78.

Sharma

Uma

(2010) Optimization of parameters for adsorption of methylene blue on a low cost activated carbon. Journal of Chemical and Engineering Data 55(1): 435–439.

79.

Sinha

Sasmal

Pal

, et al. (2021) Ammonium phosphomolybdate [(NH₄)₃PMo₁₂O₄₀] an inorganic ion exchanger for environmental application for purification of dye contaminant waste water. Journal of Photochemistry and Photobiology A: Chemistry 418: 113427.

80.

Siva

Nascimento

Brito

, et al. (2019) Biosorption of methylene blue dye using natural biosorbents made from weeds. Materials 12(15): 2486.

81.

Temkin

Pyzhev

(1940) Kinetics of ammonia synthesis on promoted iron catalysts. Acta Physicochimica, URSS 12: 327–356.

82.

Thabed

Shooto

Naido

(2021) Sorption of chromium (VI), cadmium (II) ions and methylene blue dye by pristine, defatted and carbonized Nigella Sativa L. seeds from aqueous solution. Asian Journal of Chemistry 32(2): 471–483.

83.

Torres-Perez

Huang

Hadi

, et al. (2018) Equilibrium, kinetic and optimization studies for the adsorption of tartrazine in water onto activated carbon from pecan nut shells. Water, Air, & Soil Pollution 229(73): 1–16.

84.

Wang

Duncan

Shi

, et al. (2013) WEB-based GEne SeT AnaLysis Toolkit (WebGestalt): Update 2013. Nucleic Acids Research 41(Web Server issue): 77–83.

85.

Wang

, et al. (2010) Cross- linking enzyme aggregates in the macroporos of silica gel: A practical and efficient method for enzyme stabilization. Biochemical Engineering Journal 52(2–3): 168–174.

86.

Weber

Morris

(1963) Kinetic of adsorption on carbon from solution. Journal of the Sanitary Engineering Division, American Society of Civil Engineers 89(2): 31–60.

87.

Woodbury

(1997) Physical Chemistry. New York: Brooks/Cole Publishing Company.

88.

Yemendzhiev

Alexieva

Krastanov

(2009) Decolorization of synthetic dye reactive blue 4by mycelial culture of white-rot fungi trametesversicolor 1. Biotechnology and Biotechnological Equipment 23(3): 1337–1339.

89.

Zabeti

Daud

WMAW

Aroua

(2009) Optimization of the activity of CaO/Al₂O₃ catalyst for biodiesel production using response surface methodology. AppliedCatalysis A: General 366(1): 154–159.

90.

Zaku

Emmanuel

Aguzue

, et al. (2011) Extraction and characterization of chitin, a functional biopolymer obtained from scales of common carp fish (Cyprinuscarpio L): A lesser known source. African Journal of Food Science 5(8): 478–483.

91.

Zarei

Niaei

Salari

, et al. (2010) Application of response surface methodology for optimization of peroxi-coagulation of textile dye solution using carbon nanotube-PTFE cathode. Journal of Hazardous Materials 173(1–3): 544–551.

92.

Zhao

, et al. (2020) Multiobjective optimization of a double-side linear vernier PM motor using response surface method and differential evolution. IEEE Transactions on Industrial Electronics 67(1): 80–90.

Application of central composite design to optimization of Yellow Tartarazine and Methylene Blue environmentally clean adsorption by untreated walnut shells

Abstract

Keywords

Introduction

Materials and methods

Adsorbent preparation

Characterizations of UWS

Biosorbates (Mb and Yt)

Adsorption experiments

Adsorption process optimization by response surface method

Modeling and statistical analysis

Variance analysis

Adsorption kinetic study

Adsorption equilibrium isotherms

Regression analysis, goodness of fit measure and model comparison

Thermodynamic characterization of adsorption

Results and discussion

Analyses of the UWS surface morphology

Analyses of the UWS surface chemical

Experimental design analysis by central composite design

Effect of interactive variable

Optimization by RSM and with desirability function

Adsorption kinetics and isotherms

Effect of contact time and temperature

Thermodynamic properties modeling studies

Performance and regeneration of the UWS

Adsorption mechanism probable

Conclusions and future work

Footnotes

Acknowledgements

Author contributions

Data availability

Declaration of conflicting interests

Funding

ORCID iDs

References