ACCEPTED MANUSCRIPT This is an early electronic version of an as-received manuscript that hasbeen accepted for publication in the Journal of the Serbian Chemical Society but has not yet been subjected to the editing process and publishing procedure applied by the JSCS Editorial Office. Please cite this article as M. Marković, M. Gorgievski, N. Štrbac, K. Božinović, V. Grekulović, A. Mitovski, M. Zdravković, J. Serb. Chem. Soc. (2023) https://doi.org/10.2298/JSC221018014M This “raw” version of the manuscript is being provided to the authors and readers for their technical service. It must be stressed that the manuscript still has to be subjected to copyediting, typesetting, English grammar and syntax correc- tions, professional editing and authors’ review of the galley proof before it is published in its final form. Please note that during these publishing processes, many errors may emerge which could affect the final content of the manuscript and all legal disclaimers applied according to the policies of the Journal. https://doi.org/10.2298/JSC221018014M J. Serb. Chem. Soc.00(0)1-19 (2023) Original scientific paper JSCS–12110 Published DD MM, 2023 1 Copper ions biosorption onto bean shells: kinetics, equilibrium, and process optimization studies MILJAN MARKOVIĆ*, MILAN GORGIEVSKI, NADA ŠTRBAC, KRISTINA BOŽINOVIĆ, VESNA GREKULOVIĆ, ALEKSANDRA MITOVSKI, MILICA ZDRAVKOVIĆ University of Belgrade Technical Faculty in Bor, Vojske Jugoslavije 12, Bor, Serbia (Received 18 October 2022; Revised 21 February 2023; Accepted 21 March 2023) Abstract: The removal of copper ions from aqueous solutions using bean shells as an adsorbent is presented in this paper. The influence of the solution solution pH on the biosorption capacity was investigated. The biosorption capacity increased with the increase in the solution pH. The pseudo-second order kinetic model showed the best agreement with the analyzed experimental data, indicating that chemisorption could be a possible way of binding the copper ions to the surface of the bean shells. The Langmuir isotherm model best fitted the analyzed isotherm data. The SEM-EDS analysis was performed before and after the biosorption process. The change in the morphology of the sample after the biosorption process was evident, whereby K, Mg, Si, and Ca were possibly exchanged with copper ions. Response Surface Methodology (RSM) based on the Box-Behnken design (BBD) was used to optimize the biosorption process, with the selected factors: the solution pH, initial copper ions concentration, and contact time. The optimum biosorption conditions were determined to be: pH = 3-4, initial copper ions concentration 100 mg dm-3, and contact time 10-30 minutes. Keywords: biosorption, copper ions, bean shells, kinetics, Box-Behnken design INTRODUCTION Concentrations of heavy metal ions in wastewaters originating from various industrial systems are often significant. These wastewaters can pose a serious threat to the surrounding ecosystems when discharged without previous treatment.1 Wastewaters polluted with heavy metals are treated by well-known conventional technologies, such as: adsorption, coagulation and flocculation, ion exchange, membrane filtration, precipitation, and others.2,3,4 *Corresponding author E-mail: mmarkovic@tfbor.bg.ac Phone: +381628264336 https://doi.org/10.2298/JSC221018014M A cc ep te d m an us cr ip t mailto:mmarkovic@tfbor.bg.ac https://doi.org/10.2298/JSC221018014M 2 MARKOVIĆ et al. Biosorption is a possible alternative method for heavy metal ions removal. This process is considered “user-friendly” with various advantages, including specific affinity, low cost, and simple design. It is based on adsorption with agricultural or industrial by-products being used as adsorbents. These by-products are convenient for this use due to their abundant availability, favorable physical, chemical, and surface characteristics, and their low cost.4 Copper is a widely used metal due to its excellent electrical and thermal conductivity, excellent corrosion resistance, and good resistance to strength and fatigue. Pure copper is widely used in the production of cables and wires and many other parts in the electrical industry. Due to its excellent anti-corrosion properties, copper is used for pipes, valves, and fittings in systems that carry drinking water, process water, or some other type of water. Copper extraction, as well as its production, is the main source of pollution with this heavy metal. Excessive copper concentrations in the environment are highly toxic for living organisms. It inhibits cell growth, impacts metabolism, and other processes. For this reason, it is very important to develop a sustainable, green remediation technique for copper removal, that is also economical, efficient, and environment-friendly.5 The aim The aim of this work is to determine whether bean shells can be used as an adsorbent for copper ions biosorption from aqueous solutions. Successful usage of bean shells as an adsorbent for lead ions biosorption produced the idea of investigating the potential use of this biomass as an adsorbent for other heavy metals.6 For this purpose, kinetics, isotherm, and SEM-EDS analyses of the data were obtained and shown in this paper. The process was also modeled by response surface methodology using Box-Behnken design to analyze the influence of three variables on the biosorption process and to determine their optimal values. EXPERIMENTAL Bean shells collected on the fields in the village Rudna Glava (Eastern Serbia) were used as an adsorbent for copper ions biosorption experiments. 0.5 g of been shells were used as samples for biosorption experiments. The bean shells samples were rinsed with 200 mL of distilled water, prior to the biosorption experiments, in order to remove the physical impurities. Biosorption experiments were conducted using synthetic copper ions solutions prepared with CuSO4 ∙ 5H2O (p.a purity). The solutions were prepared by mixing different amounts of the copper-sulfate with distilled water, in order to obtain the concentrations needed for the specific experiment. The stock solution concentrations varied based on the specifics of the experiment. The solution pHs was adjusted with 0.1 M HNO3 and 0.1 M KOH solutions. All experiments were performed in batch conditions. Cu(II) content was determined on a spectrophotometer (Spectroquant Pharo 300 - Merck, Rahway, New Jersey, USA), by forming a complex with NH4OH (p.a. purity), at 610 nm wavelength. The SEM-EDS analysis was performed on a SEM scanning electron microscope (VEGA 3 LMU, Tescan, Brno, Czech A cc ep te d m an us cr ip t BIOSORPTION OF COPPER IONS BY BEAN SHELLS 3 Republic) with an integrated energy-dispersive X-ray detector (X act SDD 10 mm2, Oxford Instruments, Abingdon, UK). The full characterization and preparation of the biosorbent are reported in a previous publication.6 Process parameters, such as: process time, initial copper ions concentration, initial solution pH, and temperature, were adjusted depending on the performed experiment. The biosorption capacity and the % of removal were calculated using the following equations: 𝑞𝑡 = 𝐶𝑖−𝐶𝑡 𝑚 × 𝑉 (1) % 𝑟𝑒𝑚𝑜𝑣𝑎𝑙 = (1 − 𝑐𝑡 𝑐𝑖 ) × 100 (2) where: qt is the adsorbent capacity defined as mass of the adsorbed metal per unit mass of the adsorbent (mg g−1) at time t; ci is the initial metal ion concentration in the solution; ct is the metal ion concentration in the solution at time t; m is the adsorbent mass; V is the volume of the solution; % removal is the degree of the adsorbed copper ions. RESULTS AND DISCUSSION The influence of the solution pH on the adsorption capacity To determine the influence of the solution pH on the biosorption capacity (Figure 1), a number of experiments were performed, and the initial solution pHs was adjusted in the range from 2 to 5. 50 mL of copper ions solutions (initial concentration 200 mg dm-3) was brought into contact with 0.5 g of been shells for 60 minutes. The experiments were performed in batch conditions, at room temperature, on a magnetic stirrer (with the stirring rate set at 300 rpm). As can be seen from Figure 1, the solution pH had a significant effect on the biosorption capacity. An increase in the biosorption capacity with the increase of the solution pH could be noted. At pH = 2, the biosorption capacity was 1.739 mg g-1, while the maximum capacity of 12 mg g-1 was achieved at pH = 5. A lower biosorption capacity at lower solution pH could be a result of a higher concentration of H+ ions, which occupied the active sites in the structure of bean shells and suppressed the already adsorbed Cu2+ ions. At higher pH, the concentration of H+ ions in the solution were lower, resulting in a higher biosorption capacity.7 A cc ep te d m an us cr ip t 4 MARKOVIĆ et al. Fig. 1. The influence of solution pH on the adsorption capacity Adsorption kinetics Kinetic models are often used for analyzing the experimental data to determine the biosorption rate, the step that dictates the rate of the process and its mechanism.8 In this paper, pseudo-first order kinetic model, pseudo-second order kinetic model, intraparticle diffusion (Webber-Morris) kinetic model, and Elovich kinetic model were used to analyze obtained experimental data. In order to obtain the biosorption kinetic data, 50 mL of copper ion solutions (initial concentration 200 mg dm-3) were brought into contact with 0.5 g of bean shells, for different process time (ranging from 1 to 90 minutes). The change in the biosorption capacity with time is shown in Figure 2. It can be noted that the biosorption capacity increased rapidly at the beginning of the process (first 5 minutes), as a result of a large number of available active sites in the structure of the bean shells.9 After this initial period, a slower increase in the biosorption capacity was noted (10-90 minutes), reaching a constant value after 90 minutes of the process. Pseudo-first order kinetic model This model is based on the assumption that adsorption is a reversible process.10 The pseudo-first order kinetic model is given by:6 𝑑𝑞(𝑡) 𝑑𝑡 = 𝑘1(𝑞𝑒 − 𝑞(𝑡)) (3) where: q(t) – is the adsorbent capacity defined as the mass of the adsorbed metal per unit mass of the adsorbent (mg g-1) at time t; qe – is the adsorption capacity defined as mass of the adsorbed metal per unit mass of the adsorbent (mg g-1) at A cc ep te d m an us cr ip t BIOSORPTION OF COPPER IONS BY BEAN SHELLS 5 equilibrium; k1 – is the adsorption rate constant for the pseudo-first order kinetic model (min-1). Integrating the equation (3), it follows: log(𝑞𝑒 − 𝑞(𝑡)) = log(𝑞𝑒 ) − 𝑘1 2,303 × t (4) The plot log(qe-q(t)) vs. t gives a linear dependence that serves as a base to determine the first-order kinetic model parameters. The obtained experimental data shown in Figure 2 were linearized using Eq. (4), and the obtained plot is shown in Figure 3a. The obtained kinetic parameters corresponding to this model are given in Table 1. Pseudo-second order kinetic model Pseudo-second order kinetic model is based on the assumption that adsorption and ion exchange takes place on the surface of the adsorbent, and that the adsorbate is bound to the adsorbent surface by chemisorption.11 The pseudo-second order kinetic model can be expressed as:12 𝑑𝑞(𝑡) 𝑑𝑡 = 𝑘1(𝑞𝑒 − 𝑞(𝑡)) 2 (5) where: q(t) - is the adsorbent capacity defined as the mass of the adsorbed metal per unit mass of the adsorbent (mg g-1) at time t; qe - is the adsorption capacity defined as mass of the adsorbed metal per unit mass of the adsorbent (mg g-1) at equilibrium; k2 - is the adsorption rate constant for the pseudo-second order kinetic model (g mg-1 min-1).13 The linear form of Eq. (5) is: 1 (𝑞𝑒−𝑞(𝑡)) = 1 𝑞𝑒 + 𝑘2𝑡 (6) Plot t/q(t) vs. t, shown in Figure 3b was used to determine the kinetic parameters for this model, which are given in Table 1. Intraparticle diffusion kinetic model (Webber-Morris model) This model assumes that the adsorption does not occur only on the surface of the adsorbent, but diffusion and adsorption inside the adsorbent structure are also present.13 The intraparticle diffusion kinetic model is given as:6 𝑞(𝑡) = 𝑘𝑖 𝑡 1/2 + 𝐶𝑖 (7) where: q(t) - is the adsorption capacity defined as the mass of the adsorbed metal per unit mass of the adsorbent (mg g-1) at time t; ki -is the internal particle diffusion rate constant (mg g-1 min-0.5); and Ci – is a constant that provides insight into the thickness of the boundary layer. If the Ci value is higher, the boundary layer effect is greater, so the effect of surface adsorption in controlling the process speed is A cc ep te d m an us cr ip t 6 MARKOVIĆ et al. greater (mg g-1). The plot t 1/2 vs q(t) shown in Figure 3c is used to obtain the intraparticle diffusion kinetic model parameters, shown in Table 1. Elovich kinetic model This model was primarily used to analyze gas chemisorption onto solid adsorbents but was later successfully applied on the adsorption of toxic materials from aqueous solutions.14 Elovich kinetic model is given in the following form:6 𝑑𝑞(𝑡) 𝑑𝑡 = αe−𝛽𝑞(𝑡) (8) where: α - is the starting adsorption rate (mg g-1 min-1); β - is the parameter that expresses the degree of surface coverage and activation energy for chemisorption (g mg-1); q(t) - is the adsorption capacity defined as the mass of the adsorbed metal per unit mass of the adsorbent (mg g-1) at time t. From the plot q(t) = f(lnt), shown in Figure 3d, Elovich kinetic model parameters were determined and given in Table 1. Fig. 2. Change in the adsorption capacity with time A cc ep te d m an us cr ip t BIOSORPTION OF COPPER IONS BY BEAN SHELLS 7 Fig. 3. (a) pseudo-first order kinetic model; (b) pseudo-second order kinetic model; (c) intraparticle diffusion kinetic model; (d) Elovich kinetic model TABLE I Kinetic model parameters for copper ions biosorption onto bean shells Model Parameters Values Pseudo-first order kinetic model k1 / min -1 0.023 qe,exp / mg g -1 19.07 qe,cal / mg g -1 4.97 R2 0.596 Pseudo-second order kinetic model k2 / g mg -1 min-1 0.077 qe,exp / mg g -1 19.07 qe,cal / mg g -1 19.16 R2 0.999 Intraparticle diffusion kinetic ki1 / g mg -1 min-0,5 11.722 A cc ep te d m an us cr ip t 8 MARKOVIĆ et al. model Ci1 / mg g -1 0.285 R1 2 0.989 ki2 / g mg -1 min-0,5 0.360 Ci2 / mg g -1 15.995 R2 2 0.929 Elovich kinetic model α / mg g-1 min-1 14.652 β / g mg-1 1.086 R2 0.837 Based on the obtained correlation coefficients, it can be concluded that the adsorption kinetics could be fairly modeled with the pseudo-second order kinetic model, which led to the conclusion that chemisorption was a possible way of binding copper ions onto active sites in the adsorbent structure. This statement was also supported by the negligible difference in the values of calculated and experimentally obtained adsorption capacity (qe,cal and qe,exp). Adsorption isotherms Adsorption isotherms are used to gain insight into the mechanism of the adsorption process, as well as to determine the maximum adsorption capacity.15 In this paper, the linear Langmuir, Freundlich, and Temkin isotherm models were used to describe copper ions biosorption onto bean shells. Biosorption isotherm data was obtained by performing the following experiment: 0.5 g of bean shells samples were brought into contact with 50 mL of copper ions solutions, of different initial Cu2+ concentrations (in the range from 50 to 500 mg dm−3). The suspension was stirred on a magnetic stirrer, at room temperature, for 90 minutes. Obtained experimental adsorption isotherm data for copper ions adsorption onto bean shells is shown in Figure 4a. Langmuir isotherm model This model is based on the assumption that the adsorption process occurs on specific homogenous sites inside the adsorbent structure.16 The Langmuir model can be expressed as:6 𝑞𝑒 = 𝑞𝑚𝐾𝐿 𝐶𝑒 1+𝐾𝑙𝐶𝑒 (9) where Ce - is the equilibrium concentration of metal ions (mg dm-3); qe - is the equilibrium adsorption capacity (mg g-1); qm - is the maximum adsorption capacity (mg g-1); and KL - is the Langmuir equilibrium constant (dm3 g-1). Linearizing the equation (9) the following is obtained: 𝐶𝑒 𝑞𝑒 ⁄ = 1 𝐾𝐿 𝑞𝑚 + 1 𝑞𝑚 𝐶𝑒 (10) The Langmuir isotherm data were calculated from the plot Ce vs. Ce/qe shown in Figure 4b, and given in Table 2. A cc ep te d m an us cr ip t BIOSORPTION OF COPPER IONS BY BEAN SHELLS 9 Freundlich isotherm model Freundlich model represents the earliest known relationship that describes the non-ideal and reversible adsorption. This model can also be used to study multilayer adsorption [18]. This model can be expressed as:6 𝑞𝑒 = 𝐾𝑓 𝐶𝑒 1/𝑛 (11) where Ce - is the equilibrium concentration of copper ions in the solution (mg dm-3); qe - is the adsorbent capacity defined as mass of the adsorbed metal per unit mass of the adsorbent (mg g-1) at equilibrium; KF - is the Freundlich equilibrium constant ((mg g-1) (dm3 mg-1)1/n); and 1/n -is the coefficient of heterogeneity in the Freundlich adsorption isotherm equation. Linear form of equation (11) is: 𝑙𝑜𝑔𝑞𝑤 = 𝑙𝑜𝑔𝐾𝑓 + 1 𝑛 𝑙𝑜𝑔𝐶𝑒 (12) The plot log qe vs. log Ce (Figure 4c) provided the required data for the Freundlich model isotherm data calculation, which is shown in Table 2. Temkin model This model assumes that the heat of sorption of all molecules linearly increases with the coverage of the adsorbent surface and that there is a uniform distribution of binding energies up to maximum binding energy.18 The Temkin isotherm model is given as: 𝑞𝑒 = 𝐵𝑙𝑛(𝐾𝑇 𝐶𝑒 ) (13) where: B = RT/b - is the Temkin constant, which refers to the adsorption heat (J mol-1); b - is the variation of adsorption energy (J mol-1); R - is the universal gas constant (J mol-1 K-1); T- is the temperature (K); KT - is the Temkin equilibrium constant (dm3 g-1); qe - is the adsorption capacity (mg g-1) at equilibrium; and Ce - is the equilibrium concentration of metal ions in the solution (mg dm-3).6 Linear form of equation (13) is: 𝑞𝑒 = 𝐵𝑙𝑛𝐾𝑇 + 𝐵𝑙𝑛𝐶𝑒 (14) Temkin constants B and KT were determined from the plot ln Ce vs. qe (Figure 4d), and given in Table 2. Based on the analyzed data and the obtained results (Table 2), it can be concluded that the Langmuir isotherm model was the best fit for the experimental data (R2 = 0.986), which indicated that the surface of the adsorbent was homogenous, and the biosorption of copper ions onto bean shells occurred in a monolayer.19 A cc ep te d m an us cr ip t 10 MARKOVIĆ et al. Fig. 4. (a) Experimental adsorption isotherm data, (b) Langmuir adsorption isotherm model, (c) Freundlich adsorption isotherm model, (d) Temkin adsorption isotherm model TABLE II Adsorpton isotherm model parameters for copper ions biosorption onto bean shells Langmuir Freundlich Temkin KL / dm3 mg-1 qexp / mg g-1 qm / mg g-1 R2 KF 1/n R 2 B / J mol-1 KT / dm3 g-1 R2 24.941 17.2 18.31 0.986 1.423 0.363 0.919 0.363 3.276 0.919 A cc ep te d m an us cr ip t BIOSORPTION OF COPPER IONS BY BEAN SHELLS 11 The performance of the adsorbent is usually defined by the maximum biosorption capacity. Based on the results in copper removal with various biosorbents reported in other works (shown in Table III), it can be concluded that bean shells could play an important role as a cost-effective biosorbent for copper ions removal. TABLE III Cu2+ ions biosorption on bean shells in comparison with other adsorbents Biosorbent Maximum biosorption capacity (qm, mg g -1) Work Bean shells 18.31 This work Wheat straw 4.3 12 Sawdust of deciduous trees 9.9 20 Myrica esculenta 39.37 21 Activated sawdust powder 10.35 22 Rosa damascena leaves 25.13 23 Carbonized sunflower stem 20 24 Combretum indicum 12.1 25 SEM-EDS analysis The SEM-EDS analysis was performed on samples before and after the biosorption of copper ions. The obtained results are shown in Figure 5. Before the biosorption of copper ions (Figure 5a) a porous structure was noticed, with visible cavities and macro-pores. The presence of these pores and cavities facilitated the penetration of the aqueous phase into the adsorbent structure.26 The EDS spectrum of the sample before biosorption of copper ions (Figure 5b) suggested the presence of O, Mg, Si, K, and Ca. After the biosorption process, the SEM analysis (Figure 5c) showed a more compact structure, with the absence of macro-pores and cavities, as a result of the incorporation of copper ions into the molecular structure of the bean shells. The obtained EDS spectrum after the biosorption of copper ions (Figure 5d) demonstrated the presence of O, Ca, and Cu. The absence of Mg, Si, K, and detected lower levels of Ca, indicated that any of these ions could be exchanged with copper ions during the biosorption process. A cc ep te d m an us cr ip t 12 MARKOVIĆ et al. Fig. 5. SEM-EDS analysis before and after the biosorption of copper ions Optimization of the biosorption conditions – Box-Behnken experimental design Copper ions biosorption onto bean shells was optimized using an experimental design, in order to determine the effects of three selected independent variables on the percentage of Cu2+ ions removal (dependent variable).27 The optimum biosorption conditions were determined by the means of Box-Behnken Design (BBD) and Response Surface Methodology (RSM). The RSM is a set of techniques useful for evaluating the relationships between a number of experimental factors and measured responses.28 The BBD was applied, comparing three factors: solution pH (X1), initial copper ions concentration (X2), and contact time (X3). The chosen experimental ranges and levels in the design are given in Table 4. The experimental design matrix, as well as the response Y (adsorption degree), are given in Table 5. All the experiments are performed in batch conditions, at room temperature, on a magnetic stirrer (with the constant stirring rate, set at 300 rpm). TABLE IV Experimental ranges and levels in the experimental design Factors Range level -1 0 1 X1 – solution pH 2 3 4 X2 – Initial metal ion concentration, mg/L 100 500 1000 A cc ep te d m an us cr ip t BIOSORPTION OF COPPER IONS BY BEAN SHELLS 13 X3 – Contact time, min 10 30 60 TABLE V Box-Behken Design matrix for three factors along with observed response for Cu2+ biosorption onto bean shells Run X1: solution pH X2: Initial Cu 2+ ions concentration X3: Contact time Y: % removal 1 2 100 30 32.950 2 4 100 30 86.515 3 2 1000 30 22.576 4 4 1000 30 11.890 5 2 500 10 30.550 6 4 500 10 23.836 7 2 500 60 20.860 8 4 500 60 39.748 9 3 100 10 87.790 10 3 1000 10 6.922 11 3 100 60 69.700 12 3 1000 60 7.102 13 3 500 30 25.072 14 3 500 30 31.360 15 3 500 30 38.860 The correlation between the following independent variables: linear (β1, β2, β3), quadratic (β11, β22, β33), interaction terms (β12, β13, β23), and the response (Y), was described by fitting the following polynomial equation:27 𝑌 = 𝛽0 + 𝛽1𝑋1 + 𝛽2𝑋2 + 𝛽3𝑋3 + 𝛽11𝑋1𝑋1 + 𝛽22𝑋2𝑋2 + 𝛽33𝑋3𝑋3 + 𝛽12𝑋1𝑋2 + 𝛽13𝑋1𝑋3 + 𝛽23𝑋2𝑋3 (15) The obtained results are displayed in Table 6. The biosorption of copper ions onto bean shells was expressed using the following equation: 𝑌 = 31,76 + 6,88𝑋1 − 28,56𝑋2 − 1,46𝑋3 − 3,71𝑋1 ⋅ 𝑋1 + 10,42𝑋2 ⋅ 𝑋2 + +0,69𝑋3 ⋅ 𝑋3-16,06𝑋1 ⋅ 𝑋2 + 6,40𝑋1 ⋅ 𝑋3 + 4,57𝑋2 ⋅ 𝑋3 (16) The statistical significance of the model was evaluated by the analysis of variance (ANOVA), presented in Table 6. The significance of each coefficient was determined by the magnitude of the F-values and P-values, given in Table 6. The larger the F-value, and the smaller P-value, the corresponding coefficient was more significant. P-values less than 0.0500 indicated high significant regression at 95 % confidence level.28 A cc ep te d m an us cr ip t 14 MARKOVIĆ et al. TABLE VI ANOVA analysis for response surface model in relation to Cu2+ biosorption onto bean shells Source DF Adj SS Adj MS F-Value P-Value Model 9 8676.69 964.08 6.76 0.024 Linear 3 6920.46 2306.82 16.17 0.005 X1 1 378.85 378.85 2.66 0.164 X2 1 6524.53 6524.53 45.74 0.001 X3 1 17.08 17.08 0.12 0.743 Square 3 476.86 158.95 1.11 0.426 X1 2 1 50.70 50.70 0.36 0.577 X2 2 1 401.23 401.23 2.81 0.154 X3 2 1 1.76 1.76 0.01 0.916 2-Way Interaction 3 1279.36 426.45 2.99 0.135 X1·X2 1 1032.05 1032.05 7.23 0.043 X1·X3 1 163.87 163.87 1.15 0.333 X2·X3 1 83.45 83.45 0.58 0.479 Error 5 713.24 142.65 Lack-of-Fit 3 617.94 205.98 4.32 0.194 Pure Error 2 95.30 47.65 Total 14 9389.93 The suitability of the model was confirmed by the regression coefficients of the predicted and experimental responses (R 2 = 0.924 and adj-R 2 = 0.787). This suggested that 96 % of the responses were explained by the used model. The corresponding F-value (6.76) and P-value (0.024) indicated that the model was significant. P-values lower than 0.0500 in the cases of X2 (initial Cu2+ ions concentration) and X1·X2 (solution pH combined with the initial Cu2+ ions concentration) indicated that these were the significant model terms. The relationship between the experimental responses and the responses predicted by the model is shown in Figure 6. A cc ep te d m an us cr ip t BIOSORPTION OF COPPER IONS BY BEAN SHELLS 15 Fig. 6. Plot of experimental and predicted responses As can be seen from Figure 6, there was a good relationship between the experimental and predicted responses, based on the correlation coefficient (R 2 = 0.924). The contour plots showing the influence of the analyzed process parameters on the adsorption degree are presented in Figure 7. Figure 7a indicates that the higher solution pH and lower initial Cu2+ ions concentration in the solution led to higher metal ions removal. Further, the biosorption process was more favorable at lower initial Cu2+ ions concentration combined with shorter contact time (Figure 7b). Lastly, Figure 7c shows that a high solution pH combined with medium-to- longer contact time led to a higher % removal. A cc ep te d m an us cr ip t 16 MARKOVIĆ et al. Fig. 7. Contour plots showing the interaction and the influence on the adsorption rate (Y) of: solution pH and the initial metal ions concentration (a); initial metal ions concentration and contact time (b); and solution pH and contact time (c) Response Surface Methodology, based on Box-Behnken Design, was used to optimize the process of copper ions biosorption onto bean shells. The influence of three parameters (solution pH, initial metal ions concentration, and contact time) was investigated. The obtained data indicated that the used model was statistically significant. The data showed that initial metal ions concentration, as well as the combination of initial metal ions concentration, and solution pH, had a significant influence on the biosorption efficiency. Using this model, the optimum biosorption conditions were determined to be: pH = 3-4; initial metal ions concentration 100 mg/L; and contact time 10-30 minutes. CONCLUSIONS The study of biosorption, as a potential method for copper ions removal from aqueous solutions, is presented in this paper. Bean shells were investigated as BIOSORPTION OF COPPER IONS BY BEAN SHELLS 17 potential adsorbent. For that purpose, kinetics, equilibrium, SEM-EDS, and process optimization studies were performed. The solution pH was determined to have a significant influence on the biosorption capacity. An increase in the biosorption capacity with the increase in the solution pH from 2 (1.739 mg g-1) to 5 (12 mg g-1) was noted. Experimentally obtained kinetics data were analyzed using four adsorption kinetic models (the pseudo-first order kinetic model, pseudo-second order kinetic model, intraparticle diffusion kinetic model, and Elovich kinetic model). The obtained kinetics parameters indicated that the pseudo-second order kinetic model best fitted the analyzed experimental data, which further suggested that chemisorption was a possible way of binding the copper ions to the surface of bean shells. Experimental biosorption isotherm data were fitted using the Langmuir, Freundlich, and Temkin adsorption isotherm models. Obtained results indicated that the Langmuir model served as the best fit for the analyzed data, leading to the conclusion that the surface of the adsorbent was homogenous, and the biosorption of copper ions onto bean shells occurred in a monolayer. The SEM-EDS analysis was performed on a bean shells sample before and after the biosorption of copper ions. The obtained SEM micrographs showed that the surface morphology of the sample changed after the biosorption process, from a porous to a more compact structure, possibly as a result of the incorporation of copper ions into the structure of the bean shells. The EDS spectrums of the samples before and after the biosorption process indicated that Mg, Si, K, and Ca could potentially be involved in the biosorption process and exchanged with copper ions. Process optimization studies were performed by the means of Response Surface Methodology based on the Box-Behnken Design. The influence of solution pH, initial metal ions concentration, and contact time was investigated and modeled. The used model was determined to be statistically significant. The data suggested that initial metal ions concentration, as well as the combination of initial metal ions concentration and solution pH, had a significant influence on the biosorption efficiency. Using this model, the optimum biosorption conditions were determined to be: pH 3-4; initial metal ions concentration 100 mg dm-3; and contact time 10-30 minutes. Acknowledgements: The research presented in this paper was done with the financial support of the Ministry of Education, Science and Technological Development of the Republic of Serbia, within the funding of the scientific research work at the University of Belgrade, Technical Faculty in Bor, according to the contract with registration number 451-03-68/2022- 14/200131. The authors express their appreciation to Ms. Sandra Vasković, an English Lecturer at the Technical Faculty in Bor, University of Belgrade, for her help in editing the manuscript. A cc ep te d m an us cr ip t 18 MARKOVIĆ et al. И З В О Д БИОСОРПЦИЈА ЈОНА БАКРА НА ЉУСКАМА ПАСУЉА: ИСПИТИВАЊА КИНЕТИКЕ, РАВНОТЕЖЕ, И ОПТИМИЗАЦИЈА ПРОЦЕСА МИЉАН МАРКОВИЋ, МИЛАН ГОРГИЕВСКИ, НАДА ШТРБАЦ, КРИСТИНА БОЖИНОВИЋ, ВЕСНА ГРЕКУЛОВИЋ, АЛЕКСАНДРА МИТОВСКИ, МИЛИЦА ЗДРАВКОВИЋ Универзитет у Београду, Технички Факултет у Бору, Војске Југославије 12, Бор, Србија У овом раду приказана је анализа уклањања јона бакра из водених раствора коришћењем љуски пасуља као адсорбенса. Испитан је утицај pH вредности раствора на капацитет биосорпције. Добијени резултати су показали да капацитет биосорпције расте са повећањем pH вредности раствора. Кинетичка испитивања су показала да модел псеудо-другог реда најбоље описује анализиране податке, што указује да је хемисорпција могућ начин везивања јона бакра за површину љуски пасуља. Испитивања равнотеже процеса су показала да Ленгмиров модел адсорпционе изотерме најбоље описује анализиране податке. СЕМ-ЕДС анализом су испитани узорци пре и након извођења процеса биосорпције. Ова анализа је показала евидентну промену у морфологији узорка након процеса биосорпције, при чему су ЕДС спектри указали на могућу измену K, Mg, Si, и Ca јона са јонима бакра. Модел Response Surface Methodology (RSM) базиран на Box- Behnken дизајну (BBD) је коришћен за оптимизацију процеса биосорпције, са изабраним факторима: pH вредности раствора, почетна концентрација јона бакра у раствору и време контакта. Помоћу модела су одређени оптимални услови за извођење процеса биосорпције, и то: pH вредности између 3 и 4, почетна концентрација јона бакра од 100 mg dm-3 и време контакта између фаза 10-30 минута. (Примљено 18. октобра 2022.; ревидирано 21. фебруара 2023.; прихваћено 21. марта 2023.) REFERENCES 1. F. Shafique, Q. Ali, A. Malik, Biol. Clin. Sci. Res. J. 1 (2020) e027 (https://doi.org/10.54112/bcsrj.v2020i1.27) 2. N. Agasti, CRGSC 4 (2021), 100088 (https://doi.org/10.1016/j.crgsc.2021.100088) 3. D. Lakhterwal, Adsorption of Heavy Metals: A Review, IJERD 4 (2014) 41 (https://www.ripublication.com/ijerd_spl/ijerdv4n1spl_08.pdf) 4. S. K. Gunatilake, Methods of Removing Heavy Metals from Industrial Wastewater, JMESS 1(2015), 12 (http://www.jmess.org/wp- content/uploads/2015/11/JMESSP13420004.pdf) 5. C. Tu, Y. Liu, J. Wei, L. Li, K. G. Sheckel, Y. Luo, Environ. Sci. Pollut. Res. Int. 25 (2018) 24965 (https://doi.org/10.1007/s11356-018-2563-4) 6. M. Marković, M. Gorgievski, D. Božić, V. Stanković, M. Cakić, V. Grekulović, K. Božinović, Rev. Chim. 72 (2021), 118 (https://doi.org/10.37358/RC.21.4.8462) 7. S. Schiewer, B. Volesky, Environ. Sci. Technol. 31 (1997) 2478 (https://doi.org/10.1021/es00012a024) 8. U. Farooq, J. Kozinski, M. Khan, M. Athar, Bioresour. Technol. 101 (2010) 5043 (https://doi.org/10.1016/j.biortech.2010.02.030) 9. B. Nagy, C. Manzatu, A. Maicaneanu, C. Indolean, B. T. Lucian, C. Majdik, Arab. J. Chem. 10 (2017) 3569 (https://doi.org/10.1016/j.arabjc.2014.03.004) 10. S. Lagergren, About the Theory of So-called Adsorption of Soluble Substances, Sven. Vetenskapsakad. Handingarl 241 (1898) 1 A cc ep te d m an us cr ip t https://doi.org/10.54112/bcsrj.v2020i1.27 https://doi.org/10.1016/j.crgsc.2021.100088 https://www.ripublication.com/ijerd_spl/ijerdv4n1spl_08.pdf http://www.jmess.org/wp-content/uploads/2015/11/JMESSP13420004.pdf http://www.jmess.org/wp-content/uploads/2015/11/JMESSP13420004.pdf https://doi.org/10.1007/s11356-018-2563-4 https://doi.org/10.37358/RC.21.4.8462 https://doi.org/10.1021/es00012a024 https://doi.org/10.1016/j.biortech.2010.02.030 https://doi.org/10.1016/j.arabjc.2014.03.004 BIOSORPTION OF COPPER IONS BY BEAN SHELLS 19 11. N. T. Coleman, A. C. McClung, D. P. Moore, Formation Constants for Cu(II)—peat Complexes, Science 123 (1956) 330 12. M. Gorgievski, D. Božić, V. Stanković, N. Štrbac, S. Šerbula, Ecol. Eng. 58 (2013) 113 (https://doi.org/10.1016/j.ecoleng.2013.06.025) 13. S. M. Mousa, N. S. Ammar, H. A. Ibrahim, J. Saudi Chem. Soc. 20 (2016) 357 (https://doi.org/10.1016/j.jscs.2014.12.006) 14. R. S. Juang, M. L. Chen, Ind. Eng. Chem. Res. 36 (1997) 813 (http://dx.doi.org/10.1021/ie960351f) 15. S. A. Sadeek, N. A. Negm, H. H. Hefni, M. A. Abdel Wahab, Int. J. Biol. Macromol. 81 (2005) 400 (https://doi.org/10.1016/j.ijbiomac.2015.08.031) 16. R. Han, J. Zhang, W. Zou, J. Shi, H. Liu, J. Hazard. Mater. 125 (2005) 266 (https://doi.org/10.1016/j.jhazmat.2005.05.031) 17. X. Chen, Information 6 (2015) 14 (https://doi.org/10.3390/info6010014) 18. O. Hamdaoui, E. Naffrechoux, J. Hazard. Mater. 147 (2007) 381 (https://doi.org/10.1016/j.jhazmat.2007.01.021) 19. G. Murithi, C. O. Onindo, G. K. Muthakia, Bull. Chem. Soc. Ethiop. 26 (2012) 181 (http://dx.doi.org/10.4314/bcse.v26i2.3) 20. D. Božić, V. Stanković, M. Gorgievski, G. Bogdanović, R. Kovačević, J. Hazard. Mater. 171 (2009) 684 (https://doi.org/10.1016/j.jhazmat.2009.06.055) 21. R. Kumar, H. J. Kumar, M. C. Vishwakarma, H. Sharma, K. S. Joshi, N. S. Bhandari, Environ. Nanotechnol. Monit. Manag. 19 (2023) 100775 (https://doi.org/10.1016/j.enmm.2022.100775) 22. N. Ilavarasan, Y. S. Sirinivasa Rao, R: Gokulan, A. Aravindan, Glob. Nest J. 25 (2023) 47 (https://doi.org/10.30955/gnj.004496) 23. M. A. Fawzy, H. M. Al-Yasi, T. M. Galal, R. Z. Hamza, T. G. Abdelkader, E. F. Ali, S. H. A. Hassan, Sci. Rep. 12 (2022) 8583 (https://doi.org/10.1038/s41598-022- 12233-1) 24. C. Sireesha, R. Subha, S. Sumithra, RASAYAN J. Chem. 15 (2022) 2267 (http://doi.org/10.31788/RJC.2022.1548035) 25. A. Tahir, M. Salman, Desalination Water Treat. 270 (2022) 127 (https://doi.org/10.5004/dwt.2022.28775) 26. G. F. Coelho, A. C. Goncalves, C. R. Teixeira Tarley, J. Casarin, N. Nacke, M. A. Fancziskowski, Ecol. Eng. 73 (2014) 514 (https://doi.org/10.1016/j.ecoleng.2014.09.103) 27. A. Choinska-Pulit, J. Sobolczyk-Bednarek, W. Laba, Ecotoxicol. Environ. Saf. 149 (2018) 275 (https://doi.org/10.1016/j.ecoenv.2017.12.008) 28. H. Turkyilmaz, T. Kartal, S. Yigitarslan Yildiz, J. Environ. Health Sci. Eng. 12 (2014) (https://doi.org/10.1186/2052-336X-12-5) A cc ep te d m an us cr ip t https://doi.org/10.1016/j.ecoleng.2013.06.025 https://doi.org/10.1016/j.jscs.2014.12.006 http://dx.doi.org/10.1021/ie960351f https://doi.org/10.1016/j.ijbiomac.2015.08.031 https://doi.org/10.1016/j.jhazmat.2005.05.031 https://doi.org/10.3390/info6010014 https://doi.org/10.1016/j.jhazmat.2007.01.021 http://dx.doi.org/10.4314/bcse.v26i2.3 https://doi.org/10.1016/j.jhazmat.2009.06.055 https://doi.org/10.1016/j.enmm.2022.100775 https://doi.org/10.30955/gnj.004496 https://doi.org/10.1038/s41598-022-12233-1 https://doi.org/10.1038/s41598-022-12233-1 http://doi.org/10.31788/RJC.2022.1548035 https://doi.org/10.5004/dwt.2022.28775 https://doi.org/10.1016/j.ecoleng.2014.09.103 https://doi.org/10.1016/j.ecoenv.2017.12.008 https://doi.org/10.1186/2052-336X-12-5