Simultaneous adsorption of heavy metals from water by novel lemon-peel based biomaterial

Abstract Simultaneous adsorption of heavy metals in complex multi metal system is insufficiently explored. This research gives results of key process parameters optimization for simultaneous removal of Cd(II), Co(II), Cr(III), Cu(II), Mn(II), Ni(II) and Pb(II) from aqueous solution (batch system). New lemon peel-based biomaterial was prepared and characterized by infrared spectroscopy with Fourier transformation (FTIR), scanning electron microscopy (SEM), electron dispersive spectroscopy (EDS), while the quantification of metals was made by atomic absorption spectrometry (AAS). Simultaneous removal of seven metals ions was favorable at pH 5 with 300 mg/50 mL solid-liquid phase ratio, within 60 min at room temperature with total obtained adsorption capacity of 46.77 mg g−1. Kinetic modeling showed that pseudo-second order kinetic and Weber-Morris diffusion models best describe the adsorption mechanism of all seven heavy metals onto lemon peel.


INTRODUCTION
Heavy metal ions are the most common pollutants found in water/wastewaters due to their high mobility and persistence in the environment, having serious environmental impacts worldwide 1 . Traditional methods of industrial wastewater treatment, such as chemical precipitation, coagulation, the use of activated carbon or polymer resin etc. have several drawbacks, which include sludge formation and high cost 2 . Therefore, the use of non-eatable agricultural parts that have no nutrition value, considered as green, non-toxic and biodegradable adsorbents, has becoming alternative to traditional treatment methods for heavy metal removal.
One group of such low-cost agricultural waste, generated in high quantities mostly by the juice producing industry, is citrus peels 3 . There is some indication that the carboxyl groups of pectic acid (a biopolymer known to be present in citrus peels) as well as functional groups of cellulose are acting as binding sites for positively charged metal resulting in good metal uptake. Several types of citrus peels, native or chemically modifi ed, have been recently explored for the removal of various heavy metals from wastewaters 4-11 .
Wastewaters contain more than one metal ion that could mutually interfere affecting each other's removal [12][13] . Therefore, it is important to investigate and optimize insuffi ciently explored adsorption in multi-metal systems, which is the novelty of the present study. To the best of our knowledge, Citrus limon peel has never before been used as biosorbent for simultaneous removal of seven (Cd(II), Co(II), Cr(III), Cu(II), Mn(II), Ni(II) and Pb(II)) heavy metal ions from aqueous solutions.
Regarding this, used biosorbent was characterized and applied in batch experiments in order to optimize key process parameters (sample pH, biosorbent dose, sample volume, contact time, temperature and initial metal concentration).
The research has been carried out at the Faculty of Science University of Sarajevo, Bosnia and Herzegovina, in 2019. steel sieve (Ø 0.25 mm). One part was used as native lemon peel in this form and the rest was subjected to six different modifi cation processes.
Selection of modifi cation processes was focused on two criteria: (i) avoidance of using hazardous reagents and/or their high amounts as well as any complicated procedures; (ii) removal of excess cations such as calcium or magnesium that could interfere with metal sorption and to creation of a more defi ned sorbent material.
Therefore, according to procedures described by Šabanović et al. 14 native peel (A) and (B) peel treated with 0.25 mol L −1 nitric acid followed by 0.1 mol L −1 sodium hydroxide were prepared. Furthermore, (C) material was obtained following the treatment with 0.1 mol L −1 HNO 3 described by Schiewer and Patil 15 , while (D) modifi cation was same as (B) only applied to different starting size fraction (< 250 μm instead to 0.5-1.0 cm pieces). Additionally, (E), (F) and (G) materials were obtained according to Mallampati and Valiyaveettil 16 , Memon et al. 17 and Özer and Pirinççi 18 , using 2-propanol, excess of 1 mol L −1 NaOH and concentrated sulphuric acid, respectively. In the fi nal stage these materials were dried, grounded and sieved (Ø 0.25 mm).

Methodology
Prepared lemon peel based biosorbent (50-300 mg) was added into a multi-element solution (50-100 mL) of a defi ned pH value (3)(4)(5) containing seven heavy metals (5-100 mg L −1 , each). The simultaneous removal of tested ions was carried out at 250 rpm and fi ltrates were collected at suitable time intervals (0-180 min) at four different temperatures (room, 30, 40 and 50 o C). Adsorption capacity (Eq. 1) i.e. the adsorbed amount of each metal ion (q, mg g −1 ) was calculated by subtracting initial and fi nal ion concentration at the moment of measurement according to Vanderborght and Van Griekenm 19 : (1) where C i and C f are initial and fi nal concentrations of each metal ion at the fi ltrate (mg L −1 ), V is the volume of solution (L) and m is the mass of the biosorbent (g). Heavy metal removal effi ciency (Eq. 2) was expressed as: Removal effi ciency (%) = (C i -C f ) . 100/C i (2) where Removal effi ciency (%) is the amount of analyte retained on the biosorbent.
To ascertain whether there is a loss of analytes due to their possible adsorption on the walls of used glass fl asks and/or other processes occurring within i.e. adjusting the pH value, etc., under identical conditions only without biosorbent, a control sample or a blank was subjected to all steps of the experiment. Concentration of each metal ion measured in the control sample was taken as the initial concentration (C i ).
All experiments were made in triplicates and the results were expressed as mean value ± RSD.

Kinetic modeling
Pseudo-fi rst (Eq. 3) and pseudo-second order model (Eq. 4) developed by Lagergren 20 , Ho and McKay 21 , respectively, as well as intraparticle diffusion model (Eq. 5) developed by Weber and Morris 22 were fi tted to the obtained results. Based on the coeffi cient of correlation, as a measure of experimental data matching the proposed model, the best model was chosen. The integrated forms of used models can be expressed as follows: (3) (4) (5) where q e and q t (mg g −1 ) are the amounts adsorbed at equilibrium and at time, t, respectively; k 1 (min −1 ) and k 2 (g mg −1 min −1 ) are the rate constant of the pseudo--fi rst and pseudo-second order adsorption, respectively; k in represent the diffusion rate constants within the biosorbent particle (mg g −1 min −1/2 ) and c is the constant of the diffusion model in the function of the boundary layer thickness (mg g −1 ).

Error analysis
The correlation coeffi cient (R 2 ) and statistical functions (Eqs. 6, 7 and 8): Chi square test (X 2 ), normalized deviation (ND) and normalized standard deviation (NSD) between experimental and calculated values were applied as suitable error tools to evaluate the signifi cance of the kinetic models applicability as follows: where N is the number of data points.

Lemon peel modifi cation screening
In order to enhance the biosorption characteristics of native lemon peel, six different modifi cations were applied and the results are shown at Fig. 1. Shown values are given as average of all seven tested ions for each used material. A-G marks at x axis are clarifi ed in Preparation and modifi cation of biosorbent Section.
It can be assumed that acid treatments increased the number of positive acidic oxygen functional groups 23 , Figure 1. Lemon peel modifi cation screening. Conditions: pH 5; total metal concentration 350 mg L −1 ; amount of biomass 100 mg; volume of the solution 50 mL; room temperature; equilibrium time 60 min at 250 rpm carboxyl group, this peak disappeared in case of modifi ed lemon peel as a result of protein, pectin and some organic substances removal 11, 26, 29 . This was additionally supported by disappearance of stretching vibration of C=C in aromatic rings of lignin which in native lemon peel spectrum was obtained at 1515 cm -1 30 . The peak around 1622 cm -1 was indication for stretching vibration of carboxylate ions (COO -) and strengthening of this peak, in case of modifi ed peel, was due to increased number of carboxyl group in cellulose or pectin 28 . The peaks in the range of 1300-1020 cm -1 could be attributed to the compounds containing a C-O group. Also, a new, less intense and sharp band at 894 cm -1 in modifi ed peel spectrum was characteristic α-glycoside bonds between anhydrous glucose units in celluloses structure after removal of hemi cellulose and lignin 31 . This was additionally supported by the increase of cellulose content according to the band intensity increase, due to HCH and OCH in plane bending vibration in cellulose around 1423 cm -1 32 . By peak position and peak intensity comparison within spectra before and after metal ion adsorption, FTIR technique is a very useful tool in the confi rmation of removal of metal ions. FTIR spectrum of modifi ed lemon peel (b) showed cellulose as main constituent, having functional groups such as carboxyl and hydroxyl, capable to bond metal ions in aqueous solutions 26, 27 . Observed FTIR spectrum of modifi ed lemon peel loaded with 7 metal ions (c) shows that all peaks regarding carboxyl and hydroxyl groups have shifted to higher wave numbers and their intensities were lower compared to spectrum of unloaded modifi ed lemon peel (b). From this point of view it could be concluded that metal ions were incorporated within the modifi ed lemon peel via interaction with COOand OHgroups.

SEM-EDS analysis
In the present study SEM analysis was used for characterizing the morphology and structure of the chosen F material before and after metal ions sorption. SEM micrograph of native lemon peel, F modifi ed peel within heavy metals as well as loaded with 7 heavy metals are shown at Fig. 4.
Comparing SEM images of native and modifi ed lemon peel, there are evident structural differences within shape, distribution and size of the particles, which is at the same time an indicator of successful modifi cation. Heterogeneous surface of native peels particles is noticeable (a1, a2), while modifi ed particles are bigger, homogeneous, while NaOH cleaved ester bonds and generated more hydroxyl groups 16 . As can be seen, F material treated by 1 mol L -1 NaOH had shown the best metal uptake ability both in Removal effi ciency (%) and adsorption capacity (mg g -1 ). Therefore, for all experiments within this study only this lemon peel-based sorbent was utilized.

Point of zero charge
According to Ghasemi et al. 24 the pH at the point zero charges (pH PZC ) indicate the net surface charge responsible for electrostatic repulsion or affi nity for cations. The determination of this parameter has been made following the procedure described by Sulejmanović et al. 25 which included suspending 0.10 g of biosorbent into 30 mL of 0.1 mol L -1 KNO 3 as electrolyte buffered to pH 2-10, stirred 24 h followed by fi ltration and pH measuring. As can be seen from Fig. 2, pH PZC of modifi ed lemon peel was 5.67, which means that at pH lower than 5.67 its surface is positively charged and at pH higher than this values it is negatively charged. The point of zero charge of native lemon peel was 3.95, meaning that the modifi cation made the material more alkaline and negatively charged, resulting in better affi nity to bind positively charged metal ions.

FTIR spectra
Since protein, pectin, cellulose, hemi cellulose, lignin and pigments are the most important components of lemon peel, FTIR spectra was used to detect reactive functional groups containing different donor atoms (i.e. oxygen, nitrogen, sulfur and phosphorus) 26 . Fig. 3 shows the FTIR spectra of native lemon peel (a), modifi ed lemon peel (b) and modifi ed lemon peel with 7 heavy metals (c). The expectation was that saponifi cation of protein and other organic substances of native lemon peel would result in higher quantity of carboxyl groups and reduced amount of phenolic and lactonic groups 27 . As can be seen, the wide and intense peaks among 3200 and 3600 cm -1 in both (a and b) spectra are due to the stretching vibrations of free or H-bonded hydroxyl groups of phenols, alcohols and carboxylic acids contained in cellulose, pectin and lignin 28 . The peaks observed around 2926 cm -1 were due to the stretching vibrations of the methylene, methoxy and methyl groups 14 . The characteristic peak at 1742 cm -1 in case of native lemon peel could be assigned to stretching vibrations of ester carbonyl (C=O) group. After hydrolysis of carbonyl to smooth and more compact (b1, b2). It can be seen that the modifi cation resulted in the removal of smaller particles of native material and it can be assumed that the treatment of lemon peel resulted in the reduction of "inactive" parts of the biomass and other impurities, thereby making the "active" surface of the biosorbent (surface rich in functional groups) more accessible for binding of ions. The presence of impurities, waxes and greases can interfere with sorption by masking or blocking active sites on the material surface. Furthermore, after adsorbing/incorporating heavy metals, signifi cant change in morphological structure of modifi ed peel has been occurred. Instead of expectably layers, particles and/or "islands" as separate phase on the biomaterial surface, a "sponge" like porous structure was obtained after adsorption (c1, c2). This phenomenon indicates that the sorption took place on the biosorbent surface. Additionally, EDS data confi rmed the successful modifi cation of native lemon peel (sodium was detected after modifi cation) as well as successful metal ions sorption (all seven metals were detected). At Fig. 5, EDS spectrum with respectful surface mapping (inset) of analyzed surface has been presented.
Furthermore, element weights (%) are given in detail at Table 1. Based on weight percentage quantifi cation of samples, it can be seen that the surface composition of the biosorbent is mostly carbon and oxygen. Also, the presence of K, Ca, Mg and Cu was observed as they are mineral constituents of lemon peel 33 .
The modifi cation completely removed magnesium, while the potassium and copper content signifi cantly decreased. Additionally, after simultaneous sorption of analytes, content of calcium ions becomes negligible, which could imply that ion exchange could be a kind of mechanism of interaction of this biomaterial with cationic species in general.

Solution pH
The effect of this parameter on metal Removal efficiency was investigated by varying the pH of multi metal solution from 3 to 5. Fig. 6 plots the Removal effi ciency values of all seven tested ions for each tested pH.
As can be seen, this biomaterial can be effi ciently applied in batch experiments at pH 5 because it showed the highest Removal effi ciency values for all ions in general (36.40% Cd(II), 12.24% Co(II), 68.46% Cr(III), 81.73% Cu(II), 12.40% Mn(II), 17.35% Ni(II) and 87.84% Pb(II)). At pH below 5, the main groups responsible for metal sorption (which are seen from FTIR spectrum of lemon peel given at Fig. 3) are protonated due to the high concentration of H + species. Regarding Table 1. EDS results of weight percentage quantifi cation of samples   Furthermore, multi metal solutions with pH above 5 were not tested in view of the risk of hydrolysis and/or precipitation. Therefore, pH 5 was chosen as optimal value for further experiments.

Biomaterial dose
Biomaterial dose is an important parameter used to fi nd out its adsorption capacity at particular initial concentration. The effect of this parameter was tested by dosing a different biosorbent mass (from 50 to 300 mg) in 50 mL of multi metal solution in which the initial concentration of each metal was 50 mg L −1 . The obtained results are presented in Table 2. The relative standard deviations for all obtained results were lower than 5.0%.
With the use of sorbent dose below 300 mg, a signifi cantly lower Removal effi ciency (R eff , %) values were obtained, especially for Co, Mn and Ni. Therefore, 300 mg was selected for subsequent work because it represents an adequate compromise in the meaning of R eff and q.

Sample volume
Sample volume, as well as sorbent dose, determines biosorbents capacity, so it is important to be optimized. To determine the best relation between the solid and liquid phase at which adsorption of investigated ions was most effective, sample volume varied from 50 to 100 mL (Table 3).
In general, a sample volume negatively correlated with R eff in the case of 5 of 7 tested ions. Obtained relative standard deviations for all results were lower than 7.0%. However, at various tested solid-liquid phase ratios there was no signifi cant difference in R eff for Pb and Cu ions (> 90%) resulting in higher adsorption capacities. Based on this, it could be concluded that this biomaterial has stronger affi nity toward Pb and Cu ions. Regarding simultaneous removal of heavy metal ions, for further testing 50 mL of multi metal solution was applied.

Contact time and biosorption kinetics
The contact time is one of the most important parameters for successful usage of the biosorption in practical and rapid adsorption application 34 . At the previously determined optimal conditions (pH 5, sorbent dose 300 mg and sample volume 50 mL), contact time was varied in the range of 0-180 min, and the results are plotted at Fig. 7. Fig. 7 shows a high Removal effi ciency percentage in the initial minutes followed by a gradual stabilization. In order to assure enough time for establishing the dynamic equilibrium, all further experiments lasted 60 min. Table 2. Effect of sorbent dose on the simultaneous removal of heavy metals. Conditions: pH 5; total metal concentration 350 mg L −1 ; sample volume 50 mL; room temperature; equilibrium time 60 min at 250 rpm  Pseudo-fi rst-order, pseudo-second-order and intraparticle diffusion kinetic models were used to simulate adsorption of metal ions onto modifi ed lemon peel in order to examine the controlling mechanism of adsorption process. In Table 4, maximum adsorption capacities (q max ), kinetic constants (k) and correlation coeffi cients (R 2 ) used in the three named fi ttings, are listed. The calculated q max of the pseudo-second-order model were closest to the experimental values with correlation coeffi cient (> 0.9970) higher than the pseudo-fi rst-order (< 0.5138) and intraparticle diffusion kinetic model (< 0.9675).
Additionally, error analysis (the lowest X 2 , ND and NSD values) supported the applicability of pseudo-second order model and chemisorption for each tested ion onto modifi ed lemon peel.
Diffusion of analyte toward the outer lemon peel surface as well as the ion exchange of the analyte with the weakest bonded and easily interchangeable ions on the biomaterial surface (fi rst stage of Webber-Morris model) are also involved, especially in the case of Cd and Cu adsorption processes.
Furthermore, the combination of intraparticle diffusion of the analyte ions into the internal channels and the cavities of lemon peel, the ion exchange with the more diffi cult interchangeable ions as well as the binding of an analyte to active centers of biomass (second stage of Webber-Morris model) occurred, primarily for Mn. Attainment of equilibrium (third stage of Webber-Morris model) is important, mainly for the process of adsorption of Co and Ni ions.
Consequently, it can be concluded that the mechanism of Cd, Co, Cr, Cu, Mn, Ni and Pb adsorption onto lemon peel is mixed process of surface chemisorption, which takes place through the boundary layer of the biosorbent particles and intraparticle diffusion 35, 36 .

Temperature
To establish the effect of temperature on the simultaneous heavy metal sorption onto used biomaterial, batch equilibrium studies at different temperatures (room, 30, 40 and 50 o C) were carried out and the result were shown in Fig. 8.
Simultaneous heavy metal sorption was temperature independent, with slight fl uctuations in removal effi ciency. The fact that the proposed adsorption process does not require heating makes it very economic and could have industrial application in wastewater treatment. Therefore, the room temperature was the most suitable.

Initial metal concentration
Effect of initial multi metal concentration on the R eff was investigated at the range from 5 to 100 mg L -1 (Fig. 9).
In general, two different trends were obtained among tested metals. Copper, chromium and lead ions could be effectively adsorbed at all tested concentrations by used biosorbent. Obtained R eff values were >90% for their initial concentrations above 40 mg L -1 . On the other side, R eff sharply decreased after 30 mg L -1 for cadmium, cobalt, manganese and nickel ions. This indicates that lemon peel has stronger affi nity toward Cu, Cr and Pb ions showing that their adsorption is not infl uenced by matrix effect or ionic strength.

Comparison of adsorption capacity
The biosorption capacities for heavy metal ions of used lemon peel and other adsorption materials 11, 37-41 were compared (Table 5). Used peel exhibited better adsorption capacity for Cd, Co, Cr, Cu, Mn and Ni than other listed biosorbents. Additionally, obtained adsorption capacity values were similar to Fe 2 O 3 -SiO 2 -PAN nanocomposite material, which is less economical, high cost and more diffi cult to prepare. Despite lower achieved lead adsorption by used lemon peel, total capacity for heavy metals was better compared to all listed materials. It is important to highlight that lemon peel is applicable for simultaneous removal of seven metal ions, which is more than the average number of metals for which the other materials were used.
Therefore, this biomass may be considered as an excellent adsorbent with promising potential applications for the treatment of waters polluted with listed heavy metals due to its numerous advantages i.e. it is inexpensive biocompatible and readily available, biodegradable an nontoxic material with minimum effort in its preparation and minimum chemical consumption for its surface modifi cation.

CONCLUSIONS
Biosorption using Citrus limon peel could make simultaneous metal removal economically feasible in situation were pollution control is otherwise insuffi cient due to prohibitively high costs. Kinetic modeling showed that pseudo-second order kinetic and Weber-Morris diffusion models best describe the adsorption mechanism of all seven heavy metals onto lemon peel with total obtained adsorption capacity of 46.77 mg g -1 .
This low-cost, easily available, eco-friendly and highly effi cient biomass can be exploited for simultaneous removal of Cd(II), Co(II), Cr(III), Cu(II), Mn(II), Ni(II) and Pb(II) ions from polluted water.

Funding
This study is fi nancially supported by the Federal Ministry of Education and Science, Bosnia and Herzegovina, Project No. 05-39-2625-1/18.

Disclosure statement
The authors declare no confl ict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.