Optimization of Process Parameters for Methylene Blue Dye Removal Using “Eriobotrya Japonica” Grains via Box-Behnken Design Based on Response Surface Methodology

Abstract

This study intended to examine and assess the performance of raw and treated Eriobotrya Japonica seed waste for the adsorption-based removal of methylene blue dye from an aqueous solution. The effects of several factors, including pH, adsorbent dose, and initial concentration, on the elimination of this dye were examined. To optimize the removal process, a Box-Behnken design (BBD) based on response surface methodology (RSM) was employed to evaluate the influence of key variables, adsorbent dose, initial dye concentration, and pH, along with their interactions. The findings demonstrated that the statistical analysis reveals a high significance of the dye for raw and treated Eriobotrya Japonica seed waste, with extremely weak probability values (p < 0.0001). The optimal conditions achieved were the adsorbent dose = 21.21 mg, initial dye concentration = 7.54 mg/L, and pH = 10.92 for the raw waste and adsorbent dosage = 21.75 mg, initial dye concentration = 7.5 mg/L, and pH = 11.7 for the extracted waste. These conditions result in a dye removal efficiency of 99.48% and 99.88% for raw and treated Eriobotrya Japonica seed waste, respectively. The methylene blue adsorption kinetics on the adsorbent can be precisely represented by an effective pseudo-second-order equation. The Freundlich model showed a significantly better fit to the experimental results compared to the Langmuir model.

1. Introduction

In our modern world, dyes play an omnipresent role in various aspects of life, including clothing, food industries, cosmetics, and pharmaceuticals. This prevalence of dyes is largely attributed to the growing use of synthetic colorants, which offer advantages such as ease of synthesis, quick production, and a wide spectrum of dyes not easily achievable with natural dyes. Synthetic dyes have now become a flourishing industry and are integral to the advancements of modern chemistry. In the manufacturing of our clothing, textile fibers are treated with diverse coloring substances to impart them with their distinct final hues [1].

The worldwide product of synthetic dyes is estimated at 800,000 tons annually, and they today include a large variety of chemical substances. Approximately 140,000 tons of these dyes are discharged each year during textile production and dyeing operations [2]. Textile industries—particularly during the dyeing and finishing processes—use a wide range of toxic chemicals, including cancerous dyes that pollute surface and groundwater. Methylene blue (MB) is environmentally persistent, toxic, carcinogenic, and mutagenic. It is widely employed in the textile, paper, and leather sectors, resulting in a significant amount of MB-containing effluent being released into groundwater and surface water. Furthermore, MB can influence chemical oxygen demand (COD), biological oxygen demand (BOD), and oxygen requirement levels, all of which are necessary for sustaining healthy aquatic ecosystems. Figure 1 shows the structure of methylene blue, while

Table 1. Methylene blue characteristics [3].

The Main Characteristics of Methylene Blue
IUPAC name	Bis-(dimethylamino)-3,7 phenazathionium chloride-ethanol
Synonym	Basic Blue 9
Indice color	52,015
Chemical formula	C16H18N3SCl
Molecular mass	319.85 g/mol
pKa	3.8
Melting temperature	180 °C
Solubility	50 g/L (water) and 10 g/L (ethanol)

gives its characteristics.

Our research focuses on methylene blue (MB) dye, which is widely utilized in numerous sectors and is commonly detected in wastewater. Its continual discharge into ecosystems is known for causing environmental and health issues. These environmental threats range from odor and taste changes in water sources to sunlight absorption in the aquatic environment, resulting in a drop in oxygen, low biodegradability, or partial degradation, which produces metabolites that are more toxic than the parent molecules [4]. There are several techniques for treating textile effluents, such as membrane filtration [5,6], ion exchange, adsorption [7], electrocoagulation [8,9], precipitation, and oxidation techniques such as photo-catalytic oxidation [10], and ozonation [11]. In this work, we employed adsorption to remove methylene blue dye from an aqueous solution. Adsorption is a physicochemical phenomenon characterized by an interaction between adsorbates and adsorbent surfaces, thereby making it a surface phenomenon. Adsorption on a solid can be defined as the accumulation of molecules from a gaseous or liquid phase on the surface of the solid [12,13].

In this research, we have also focused on evaluating a lignocellulosic adsorbent known as Eriobotrya Japonica, or loquat. Native to China and Japan, Eriobotrya Japonica is a subtropical fruit tree belonging to the Rosaceae family. Various parts of the loquat tree, including its leaves, bark, and fruit, have been shown to offer numerous health benefits. In traditional Chinese medicine, loquat is widely used to treat a range of ailments such as chronic bronchitis, inflammatory lung diseases, diabetes mellitus, and gastrointestinal disorders.

The plant contains a wealth of bioactive compounds, including carotenoids, vitamins, and polyphenolic compounds, which contribute to its diverse biological properties. Our work, however, is rooted in the context of sustainable development, where we aim to valorize loquat seeds by studying their adsorption potential under varying conditions [14].

Response surface methodology (RSM) has become a valuable asset for streamlining intricate processes and systems across various fields such as engineering, chemistry, and agriculture. It enables researchers to analyze the combined effects of multiple independent variables on one or more dependent variables concurrently, facilitating the identification of optimal operating conditions for the studied system [15,16,17]. Commonly employed RSM techniques for modeling physicochemical processes include Central Composite Design (CCD) [18], Box-Behnken Design (BBD), User-Defined Design, and the Customized Historical Data Model [19]. BBD falls under the category of rotatable or almost-rotatable second-order designs, utilizing three-level inadequate factorial designs [20].

The objective of this study is to optimize the removal of methylene blue dye by adsorption on raw and treated Eriobotrya Japonica seeds, using response surface methodology (RSM) based on the Box-Behnken design (BBD). This approach was chosen because it allows for better prediction of experimental responses. RSM was also used to determine the optimal conditions for process variables, such as adsorbent dose, initial concentration, and pH, and to generate a predictive model equation for the adsorption process. Additionally, the isotherm and kinetics of the adsorption process were studied.

2. Materials and Methods

2.1. Adsorbent: Eriobotrya Japonica Seeds

To prepare the adsorbent, the raw waste (RW) of Eriobotrya Japonica seeds was crushed and sieved to obtain specific particle size fractions (50–125–250–300 µm). The grains were then cold-macerated for three days with 99% ethanol as the extraction solvent. After filtration, the extract was concentrated by removing the solvent with a rotary evaporator at 60 °C. The extraction efficiency was estimated using Formula (1) [21,22], and the yield obtained was 53%. The extracted waste of Eriobotrya Japonica seeds (EW) was crushed and sieved into the following aggregate sizes: 50–125–250–300 µm.

$$R_d={\displaystyle \frac{m_{ech}}{m_{exp}}}\times 100$$

(1)

where R_d denotes the yield in percentage; m_ech is the dry mass of the sample before extraction; and m_exp is the mass of the extract after solvent evaporation.

2.2. Batch Adsorption Experiments

Batch adsorption experiments were conducted using various quantities of adsorbent and a 20 mL dye solution with a specified concentration. The pH of the solution was changed by adding 0.1 M HCl or 0.1 M NaOH. The dye solution and Eriobotrya Japonica seeds were shaken for an hour at 300 revolutions per minute. Afterward, the mixture was centrifuged at 3500 rpm for 6 min to separate the solid adsorbent from the solution. A UV–visible spectrophotometer at 664 nm was used to evaluate the supernatant. The remaining equations were used to compute methylene blue’s removal effectiveness (R %) (2) and adsorption capacity (Q_e) (3).

$$R={\displaystyle \frac{Ci-Cf}{Ci}}\times 100$$

(2)

$$Qe={\displaystyle \frac{Ci-Cf}{m}}\times V$$

(3)

where Ci is the initial dye concentration in mg/L; Cf represents the final dye concentration in mg/L; m represents the adsorbent mass in mg; and V is the solution volume in L.

2.3. RSM Statistical Analysis Method

RSM, the statistical method employed in this study, aims to delineate the cause-and-effect association across real average ratings and input control parameters. The underlying principle of RSM entails executing a sequence of tests with the objective of attaining an ideal response [23]. In this investigation, the RSM design, specifically the Box-Behnken design (BBD), was employed. The Design Expert software (version 13) generated 17 experiments to assess the relationship between the responses of the RW removal rate of MB (Y1), the EW removal rate of MB (Y2), and the independent variables (adsorbent dose (A), initial concentration (IC) (B), and pH (C)).

Table 2. Independent variables, their levels, and symbols for Box-Behnken experimental design.

Factor	Name	Units	Min	Max
A	Adsorbent dose	Mg	5	50
B	IC	mg/L	5	50
C	pH	-	2	12

displays the variables of the process employed in this experimental design, with three factors and the response each having two selected levels. The design matrix resulting from the application of BBD is presented in

Table 3. Variables of BBD matrix and experimental data for MB removal rate using RW and EW.

Run	Factor 1 A: Adsorbent Dose (mg)	Factor 2 B: IC (ppm)	Factor 3 C: pH	Eriobotrya Japonica Seed Waste
				Response 1: RW (Y1) TR (%)	Response 1: EW (Y2) TR (%)
1	50	27.5	2	26.67	26.67
2	50	27.5	12	92.98	94.23
3	5	27.5	2	12.12	12.12
4	27.5	5	12	97.31	98.34
5	5	27.5	12	81.52	84.13
6	27.5	27.5	7	87.02	83.45
7	27.5	27.5	7	87.02	83.45
8	27.5	50	2	51.67	51.67
9	27.5	27.5	7	87.02	83.45
10	50	50	7	90.51	87.99
11	27.5	5	2	18.75	19.45
12	5	5	7	73.21	71.4
13	5	50	7	83.67	79.67
14	27.5	27.5	7	87.02	83.45
15	27.5	27.5	7	87.02	83.45
16	27.5	50	12	87.59	89.7
17	50	5	7	98.98	98.71

. In this research, Equation (4) introduces the utilization of a quadratic polynomial model.

$$y=a_0+{\displaystyle \sum _{i=1}^n}{a}_i{x}_i+{\displaystyle \sum _{i=1}^n}{a}_{ii}{\displaystyle \underset{i}{\overset{2}{x}}}+{\displaystyle \sum _{i=1}^n}{\displaystyle \sum _{j=i+1}^n}{a}_{ij}{x}_i{x}_j$$

(4)

The response variable in this study is Y (%), which stands for the percentage of methylene blue dye clearance. The model constant is pictured by a₀, while Xi and Xj indicate the selected independent variables. The linear, quadratic, and interaction terms within the model are represented by the regression coefficients ai, aii, and aij, in that order. An analysis of variance (ANOVA) was executed on the acquired outcomes to evaluate the model’s efficacy. Model significance and insignificance are common approaches for evaluating the alignment between experimental data and the expected response prediction model, as current research suggests [24,25].

2.4. Model of Experimental Data

Various models—such as pseudo-first-order and pseudo-second-order kinetics as well as Freundlich and Langmuir isotherms—were utilized to analyze the adsorption kinetics and isotherms (

Table 4. Model analysis for adsorption phenomena.

NonLinear Models		Expression	Characteristic Parameters
Kinetic models	Pseudo-first-order model	$q_t=q_e\left(1-e^{-k_{1}t}\right)$	qe (mg/g) qt (mg/g) k1 (min−1)
	Pseudo-second-order model	$q_t={\displaystyle \frac{k_2{q}_e^{2}t}{1+k_2{q}_{e}t}}$	qe (mg/g) qt (mg/g) k2 (g/mg·min)
Isotherm models	Freundlich model	$Q_e=K_f{C}_e^{{\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$n$}\right.}}$	Qe (mg/g) Kf (min−1)
	Langmuir model	$Q_e={\displaystyle \frac{Q_m{K}_L{C}_e}{1+K_L{C}_e}}$	Qm (mg/g) Qe (mg/g) KL

) [22,23,24,25]. The experimental data from the adsorption experiments were fitted to the nonlinear forms of each model. The aim of this analysis was to identify the most suitable model capable of accurately predicting the adsorption behavior of methylene blue on Eriobotrya Japonica seed waste.

3. Data and Interpretations

3.1. Sample Characterization

3.1.1. Fourier-Transform Infrared Spectroscopy (FTIR)

Figure 2 depicts the infrared spectra of RW and EW before and after adsorption. A peak at 575 cm⁻¹ suggests the existence of bromo-alkanes. An absorption band at 850 cm⁻¹ indicates the valence vibration of para-disubstituted aromatic benzene. The adsorption peak at 1035 cm⁻¹ indicates the existence of fluoroalkanes with a typical bond. The peaks at 1450 cm⁻¹ and 1510 cm⁻¹ indicate the C=C stretching vibration, whereas the peak at 1630 cm⁻¹ is linked to the C≡N stretching vibration. The methylene group has a peak at 2990 cm⁻¹, whereas secondary and primary amines show peaks at 3100 cm⁻¹ and 3400 cm⁻¹, respectively.

3.1.2. Scanning Electron Microscopy (SEM)

Microscopic examination enables us to observe the physical structure of the raw material and the extract both before and after adsorption. Figure 3 and Figure 4 illustrate the heightened porosity following the adsorption process, indicating that both materials exhibit increased porosity after participating in the adsorption phenomenon.

3.1.3. Energy Dispersive X-Ray Analysis (EDX)

X-ray analysis (EDX) (Figure 5 and Figure 6) patterns reveal the presence of major peaks corresponding to carbon (C) and oxygen (O) in both the raw Eriobotrya Japonica seeds and the extract, along with a significant peak of silica (Si) in the raw waste that disappears after adsorption. Other elements detected include aluminum (Al), sulfur (S), potassium (K), calcium (Ca), and magnesium (Mg). Silicon appears as the dominant element in the sample, with oxygen and carbon in lower proportions, potentially originating from impurities, organic compounds, or oxides. Potassium is observed in trace amounts, possibly as a mineral inclusion or residual salt. The X-axis likely represents wavelength, energy (keV), or angle (2θ), depending on the technique used, while the Y-axis indicates signal intensity, reflecting the relative proportions of the detected elements.

3.2. RSM Modeling

The Box-Behnken design (BBD) model was used to evaluate the effects of adsorbent mass, initial concentration (IC), and pH—both separately and in combination—on the removal of methylene blue dye. As indicated in Table 4 and

Table 5. ANOVA and model coefficients for the adsorption process for the RW.

Source	Sum of Squares	Df	Mean Square	F-Value	p-Value
Model	12,481.9	9	1386.88	82.26	<0.0001	Significant
A—adsorbent dose	429.54	1	429.54	25.48	0.0015
B—IC	79.32	1	79.32	4.7	0.0667
C—Ph	7824.38	1	7824.38	464.09	<0.0001
AB	89.59	1	89.59	5.31	0.0546
AC	2.39	1	2.39	0.1416	0.7179
BC	454.54	1	454.54	26.96	0.0013
A2	125.87	1	125.87	7.47	0.0292
B2	106.95	1	106.95	6.34	0.0399
C2	3355.51	1	3355.51	199.03	<0.0001
Residual	118.02	7	16.86
Lack of fit	118.02	3	39.34
Pure error	0	4	0
Cor total	12,599.92	16

, an ANOVA (analysis of variance) was used to validate the statistical analysis of the experimental data for the removal of methylene blue dye. The two models (RW (Y1), EW (Y2)) have estimated F-values of 56.85 and 82.26, respectively. The significant importance of both models was indicated by the obtained p-values, which were both less than 0.0001 [26]. Additionally, a coefficient of determination (R²) of 0.97 showed that there was a strong correlation between the actual and expected values of methylene blue dye clearance. Given the selected conditions, the p-values of terms A, C, BC, A², B², and C² in the BBD model were found to be less than 0.05 (Prob > F < 0.0500), indicating that they were statistically significant for the removal of methylene blue dye. Thus, it was believed that these words had a big impact. On the other hand, terms in the BBD model with p-values greater than 0.05 were taken out of the quadratic polynomial equation to achieve the best fit. The relationship between the response (RW (Y1), EW (Y2)) and the parameters under study is expressed by Equations (5) and (6).

RW

$$\begin{array}{c}Y1=87.02+7.33\ast A+3.15\ast B+31.27\ast C-4.73\ast AB-0.7725\ast AC-10.66\ast BC\\ -5.47\ast A^2+5.04\ast B^2-28.23\ast C^2\end{array}$$

(5)

EW

$$\begin{array}{c}Y2=83.45+7.53\ast A+2.64\ast B+32.06\ast C-4.75\ast AB-1.11\ast AC-10.22\ast BC-4.75\\ \ast A^2+5.75\ast B^2-24.41\ast C^2\end{array}$$

(6)

The regression Equations (5) and (6) reveal that variables (A-B-C-B2-C2) have positive effects on the response (RW (Y1), EW (Y2)), as indicated by their positive regression coefficients. Conversely, variables AB, AC, BC, and A2 have negative effects on the response (RW (Y1), EW (Y2)). In this case, a positive number denotes a beneficial impact that facilitates optimization, whereas a negative value implies that the factor and response have an inverse relationship [27]. In addition, it should be noted that variables affecting methylene blue dye removal show a similar trend for both the raw and extracted Eriobotrya Japonica seed waste, indicating their proximity.

The fit and significance of the developed model were assessed using the analysis of variance (ANOVA), as presented in Table 5 and

Table 6. ANOVA and model coefficients for the adsorption process for the EW.

Source	Sum of Squares	Df	Mean Square	F-Value	p-Value
Model	11,980.38	9	1331.15	56.85	<0.0001	Significant
A—adsorbent dose	454.21	1	454.21	19.4	0.0031
B—IC	55.81	1	55.81	2.38	0.1665
C—pH	8223.39	1	8223.39	351.23	<0.0001
AB	90.16	1	90.16	3.85	0.0905
AC	4.95	1	4.95	0.2114	0.6596
BC	417.38	1	417.38	17.83	0.0039
A2	95.2	1	95.2	4.07	0.0836
B2	139.09	1	139.09	5.94	0.0449
C2	2508.32	1	2508.32	107.13	<0.0001
Residual	163.89	7	23.41
Lack of fit	163.89	3	54.63
Pure error	0	4	0
Cor total	12,144.28	16

for the response (RW(Y1), EW (Y2)).

Based on the ANOVA displayed in Table 5 and Table 6, we can arrange the most significant effects for RW and EW for the elimination of methylene blue dye as follows:

RW: C > C² > BC > A > A² > B² > AB > B > AC

EW: C > C² > A > BC > B² > A² > AB > B > AC

The validation of the BBD model can also be approached graphically by examining the distribution of residuals and the correlation between actual and predicted values for methylene blue dye removal. Figure 7 presents the normal probability plot of residuals for the BBD model. As shown, the data points align closely along a straight line, suggesting a desirable residual distribution and independence [28]. Furthermore, Figure 7 indicates a strong agreement between observed and predicted values for methylene blue removal, providing additional support for the model’s validity.

Figure 7 illustrates the comparison between the predicted and actual response values. It is evident that the predicted values closely and uniformly align with the actual responses for both responses, resulting in a satisfactory agreement with R² > 0.97. This suggests that the regression models created are successful in elucidating the connection between the variables and outcomes within the investigated range. The graph’s data points are spread along a linear trajectory in a uniform and consistent manner.

3.3. Response Surface (3D)

The three-dimensional response surfaces display the interactions between variables (Adsorbent dose*IC), (Adsorbent dose*pH), and (IC*pH) of the removal of methylene blue dye onto RW and EW, shown in Figure 8 and Figure 9, respectively.

The impact of pH on methylene blue dye removal is observed to be more significant compared to the dose and IC. These findings align with the results obtained from the ANOVA analysis, further validating their accuracy. The statistical significance of these results is noteworthy. Nonetheless, this study looked at the impact of pH, starting dye concentration, and adsorbent mass on MB elimination from an experimental perspective. As the quantity of adsorbent increases within a constant volume of known concentration, the accessibility of the adsorption sites diminishes, leading to elevated rates of MB removal. This is attributed to the aggregation of Eriobotrya Japonica waste particles or the overlapping of binding sites.

This aggregation becomes more pronounced with higher adsorbent dosages in the solution. Moreover, pH influences the distribution and speciation of cations and the surface charge of the involved molecules. Cationic dyes, when dissolved in aqueous solutions, typically carry a positive charge. At lower pH levels, H⁺ ions surround the adsorbent surface, impeding the interaction of MB ions with the adsorbent sites. Consequently, the formation of bonds between the MB and active regions is less likely as H⁺ ions and cationic dye compete for accessible adsorption sites and experience electrostatic repulsion. Conversely, at higher pH levels, an improved interaction between OH- ions and the dye enhances the binding between dye ions and surface sites [29]. Conversely, the initial dye concentration has a detrimental influence on the adsorption from Eriobotrya Japonica waste. The quantity of MB adsorbed per unit mass of Eriobotrya Japonica waste (Qe) increases quickly with rising initial MB solution concentration as the initial concentrations grow. Nevertheless, the low ratio of MB in solution to accessible active adsorption sites causes a decrease in the elimination rate [30].

3.4. RSM Optimization

As shown in Figure 10 and Figure 11, the Design Expert program was utilized to optimize the process parameters during the adsorption-based removal of methylene blue dye using RW and EW.

The goal of the optimization process was to maximize the elimination of methylene blue dye employing two classes of waste Eriobotrya Japonica grains (RW and EW) within the range of independent variables studied. The optimal solution is typically selected according to its maximal attractiveness or proximity to unity, with the red zone often serving as a guide to the region where maximum removal efficiency may be achieved [29,30]. For both waste products (RW and EW), the ideal values for the adsorption of MB was an adsorbent dose (21.21–21.75 mg), with the IC at 7.54–7.5 mg/L, and pH at 10.92–11.07. The methylene blue dye removal efficiency under these ideal circumstances of RW and EW as adsorbents was 99.48% and 99.88%, respectively.

3.5. Kinetic Study

Surface reaction kinetic models are frequently employed for predicting the adsorption kinetics of MBD over both adsorbents investigated in this study. These include the PFO and PSO models that are widely utilized for describing surface reactions. The kinetics of PFO and PSO can be quantified and expressed by the nonlinear formulas in Table 4.

Figure 12 and Figure 13 present the experimental results of MBD adsorption, along with the nonlinear kinetic models PFO and PSO, for the two types of Eriobotrya Japonica waste, RW and EW, respectively.

Table 7. Kinetic parameters of the MBD adsorption onto the two waste materials of Eriobotrya Japonica.

	PFO
Adsorbent	RW	EW
Qe (mg/g)	17.228	17.252
K1 (min−1)	0.3103	0.5706
RMSE	0.276	0.154
MSE	0.076	0.024
SSE	0.764	0.245
R2	0.9961	0.9971
	PSO
Qe (mg/g)	17.228	17.252
K2 (g/mg·min)	0.0524	0.208
RMSE	0.349	0.154
MSE	0.122	0.024
SSE	1.227	0.243
R2	0.9973	0.9991

summarizes the kinetic parameters of MBD adsorption for these two wastes.

The data given in Table 7 prove the PSO kinetic model accurately describes the adsorption procedure.

Data reported in the table obviously demonstrate that the PSO system is the most appropriate for predicting the adsorption kinetics of MBD by Eriobotrya Japonica. This conclusion is supported by the very high correlation coefficient (greater than 99%), which indicates a high level of consistency of the experimental and approach data. In addition, the values estimated by the model closely match the actual experimental results, confirming the reliability of the PSO model [31].

3.6. Isotherm Study

As shown in Figure 14 and Figure 15, isothermal adsorption tests using Langmuir and Freundlich systems are applied to estimate the capability of adsorbents to remove MBD. The Langmuir and Freundlich models used in this study are nonlinear models, with their formulas presented in Table 4.

Table 8. Isotherm parameters of MBD adsorption onto the two waste materials of Eriobotrya Japonica.

	Langmuir
Adsorbent	RW	EW
Qm (mg/g)	365.04	300
KL	0.014	0.019
RMSE	2.046	2.196
MSE	4.187	4.824
SSE	41.87	48.24
R2	0.936	0.931
	Freundlich
KF	5.34	5.9
1/n	0.934	0.933
RMSE	2.048	2.135
MSE	4.197	4.561
SSE	41.97	45.61
R2	0.946	0.945

illustrates the values calculated for the constants and correlation coefficients (R²).

From Table 8, the results indicate a coefficient of determination (R²) above 0.93 for both raw and treated waste. These findings suggest that the Freundlich model accurately represents the equilibrium adsorption of DBM by RW and EW. This implies that the surface of the adsorbent promotes multilayer adsorption [32]. Additionally, the high K_F values reveal a strong adsorption capacity of the adsorbent [33]. The maximum adsorption capacity of DBM was compared to the values reported in the literature for various other adsorbents, as shown in

Table 9. Comparison of the adsorption capacity of DBM with various adsorbents reported in the literature.

Adsorbent	Experimental Adsorption Capacity Qmax (mg/g)	Reference
Sulfuric-acid-treated coconut shell (SATCS)	50.6	[34]
Fava bean peel waste (FBP)	140	[35]
Citrus limetta peel (CLP)	227.3	[36]
Waste seeds Aleurites moluccana (WAM)	178	[37]
Nanoparticles (SFH-SiO2)	70.16	[38]
Fly ash-based geopolymer (FAG)	37.04	[39]
Raw and treated Eriobotrya Japonica seed waste	365.04 and 300	This study

. It was found that the adsorption efficiency of DBM exceeded that of many other previously studied adsorbents.

4. Conclusions

This study introduced a novel and eco-friendly approach for removing methylene blue dye from aqueous solutions using raw and chemically treated Eriobotrya Japonica seeds as low-cost lignocellulosic adsorbents. The originality of this work lies in the valorization of vegetable waste combined with a statistical optimization approach through response surface methodology (RSM) based on the Box-Behnken Design (BBD).

The experimental results confirmed that the adsorption process was significantly influenced by pH, initial dye concentration, and adsorbent dose. The model developed through RSM showed strong predictive capabilities, with correlation coefficients (R²) exceeding 0.97 for both raw (RW) and treated (EW) seed materials. Optimal removal conditions were identified as follows: adsorbent dose—21.21–21.75 mg, initial concentration—7.50–7.54 mg/L, and pH—10.92–11.07. Under these conditions, the maximum removal efficiencies reached 99.48% for RW and 99.88% for EW.

Compared to other biosorbents such as Banana peels (Qmax ≈ 40 mg/g, ~85% removal), Orange peels (Qmax ≈ 78 mg/g, ~93.5% removal), or Rice husk (Qmax ≈ 49.5 mg/g, ~80% removal), Eriobotrya Japonica seeds demonstrated excellent performance, highlighting their potential as a high-efficiency and eco-friendly alternative for dye remediation in wastewater treatment.

Beyond its technical achievements, this research contributes to broader environmental and sustainable development goals by promoting green chemistry, waste valorization, and low-cost treatment technologies. Future work should focus on testing the adsorbent’s regeneration and reuse potential, as well as validating its effectiveness in continuous-flow systems and real textile wastewater.