Single and Binary Equilibrium Studies for Ni2+ and Zn2+ Biosorption onto Lemna gibba from Aqueous Solutions

The biosorption ability of Lemna gibba for removing Ni2+ and Zn2+ ions in aqueous batch systems, both individually and simultaneously, was examined. The influences of solution pH and initial single and binary metal concentrations on equilibrium Ni2+ and Zn2+ biosorption was explored. The optimal solution pH for Ni2+ and Zn2+ biosorption was 6.0, for both the single and binary metal systems. Ni2+ and Zn2+ biosorption capacities increased with increasing initial metal concentrations. The presence of Zn2+ ions more adversely affected the biosorption of Ni2+ ions in the binary metal systems than vice versa. The single and binary biosorption isotherms of Ni2+ and Zn2+ revealed that L. gibba’s affinity for Zn2+ ions was higher than that for Ni2+ ions. The Redlich–Peterson and Freundlich isotherm models fit well to the experimental equilibrium data of Ni2+ ions, whereas Redlich–Peterson and Langmuir models better described the equilibrium data of Zn2+ ions in single metal systems. The modified Sips isotherm model best fit the competitive biosorption data of Ni2+-Zn2+ on L. gibba. FTIR analyses suggest the involvement of hemicellulose and cellulose in the biosorption of Ni2+ and Zn2+. The presence of Ni2+ and Zn2+ on the L. gibba surface was validated by SEM-EDX.


Introduction
Soil, surface water, and groundwater pollution by divalent nickel (Ni 2+ ) and zinc (Zn 2+ ) ions can occur by natural processes, but is predominantly due to anthropogenic activities [1], especially the Ni 2+ -and Zn 2+ -contaminated discharges from industrial and commercial processes, including mining; mineral processing; electroplating; production of paints, batteries, coins, and electronic commodities; and many other applications [2,3].
Ni 2+ and Zn 2+ are essential minerals for some living beings [4]. However, at high concentrations, they are of serious environmental, public health, human well-being, and economic concern due to their toxic effects on cell physiology and metabolism, which can decrease the viability of living organisms or even cause their death [5]. Human exposure to high Ni 2+ and Zn 2+ levels has been associated with serious acute and chronic health effects. Ni 2+ induces hematotoxic, inmunotoxic, hepatotoxic, pulmonary toxic, nephrotoxic, and embryotoxic effects, allergic reactions, and contact dermatitis. In addition, Ni 2+ impairs the natural balance of essential minerals and is a carcinogenic and

Ni 2+ and Zn 2+ Stock and Test Solutions
Some 40 mM Ni 2+ and Zn 2+ stock solutions were produced by accurately weighing the required amount of NiSO 4 ·6H 2 O (J.T. Baker ® , Monterrey, Mexico; purity >99.1%) or ZnSO 4 ·7H 2 O (J.T. Baker ® , Monterrey, Mexico; purity >99.9%) and dissolving it in distilled deionized water. For the single and binary metal biosorption experiments, different initial concentrations of each heavy metal ion were obtained by diluting the stock Ni 2+ and Zn 2+ solutions.

Preparation of Biosorbent
L. gibba was harvested in Xochimilco Lake, Mexico City, Mexico. The macrophyte was exhaustively washed with tap and distilled deionized water to completely remove aquatic plants, other than L. gibba, dead insects, garbage, and debris. The duckweed samples were then dried at 60°C until reaching a dry weight constant; thereafter, the macrophyte was ground using a hammer mill, and the obtained powder was sifted through U.S. ASTM sieves to collect the fraction with particle sizes ranging from 0.3 to 0.5 mm, which was employed further in biosorption experiments.

Proximate Chemical Analysis of Biosorbent
Proximate chemical analysis of L. gibba samples was performed in triplicate, following the procedures described in the AOAC handbook [22]. We applied the Soxhlet method to quantify ether extract (EE). Total protein (TP) was determined using the Kjeldahl method, with 6.25 as the conversion factor from total nitrogen to total protein. Crude fiber (CF) was analyzed in terms of the loss in ignition of dried lipid-free residues, following digestion with H 2 SO 4 and NaOH standard solutions. Ash (A) was verified by burning samples in a muffle furnace at 600°C, until the weight was constant. Nitrogen-free extract (NFE) was quantified as the difference between dry matter and the sum of ash, total protein, ether extract, and crude fiber (Equation (1)):

Biosorption Studies
Biosorption tests were conducted to examine the effects of metal solution pH and the initial concentrations of Ni 2+ and Zn 2+ ions on equilibrium metal biosorption by L. gibba. Previous kinetic studies showed that equilibrium times for Ni 2+ and Zn 2+ biosorption were less than 48 h (data not shown). Thus, contact time between L. gibba and Ni 2+ or Zn 2+ , or mixtures of Ni 2+ and Zn 2+ solutions, of 48 h was used in equilibrium biosorption studies.
All single and binary equilibrium biosorption tests were conducted in Erlenmeyer flasks (500 mL) containing 1 g (dry weight) L −1 of L. gibba biomass and 100 mL of metal solution, with a constant agitation (Cole Parmer ® , Vernon Hills, IL, USA) of 180 rpm and a temperature of 20°C.
In the single metal biosorption systems, the effects of the solution pH level on L. gibba Ni 2+ and Zn 2+ biosorption were explored in metal solutions at 2 mM initial Ni 2+ or Zn 2+ concentration, with solution pH values ranging from 1.0 to 6.0. Similar experiments covering the same range of solution pH values were performed to explore the influence of metal solution pH levels on the simultaneous biosorption of Ni 2+ and Zn 2+ ions onto L. gibba, using an equimolar initial concentration of 2 mM of both heavy metals. Metal solution pH values higher than 6.0 were not assayed to prevent the precipitation of insoluble Zn(OH) 2 in the single and binary metal biosorption experiments [23]; 0.1 M NaOH or HCl solutions were used to adjust the pH of each monometal and bimetal solution to the desired value.
To investigate the influences of initial Ni 2+ or Zn 2+ concentration on their removal by L. gibba in single-metal systems, the initial concentrations of Ni 2+ or Zn 2+ ions were varied over the 0.2-30 mM range (0.2, 0.4, 1, 2, 4, 6, 10, 20, and 30 mM). For the two-metal systems, a series of tests were conducted with a constant concentration of the first metal within the interval from 0.2 to 10 mM (0.2, 0.4, 1, 2, 4, 6, and 10 mM), while varying the concentration of the second metal (0.2, 0.4, 1, 2, 4, 6, and 10 mM). A total of 49 different bimetallic systems were assayed combining the two heavy metals studied. In the two-metal systems, no experiments were performed at initial concentrations of each heavy metal greater than 10 mM due to interferences in the quantification of the two metals.
The flasks were taken out from the shaker bath (Cole Parmer ® , Vernon Hills, IL, USA) at the end of the batch biosorption experiments, and the biosorbent was separated from the aqueous solutions by filtration (Whatman ® grade 42, St. Louis, MI, USA). Filtrates were further used to quantify residual Ni 2+ and/or Zn 2+ concentrations.
Control experiments without L. gibba biomass were carried out simultaneously using the same conditions as those employed for the Ni 2+ and/or Zn 2+ biosorption tests to determine the loss of heavy metal ions by precipitation and/or adsorption onto the glass wall. No statistically significant differences were encountered in the Ni 2+ and/or Zn 2+ concentrations of the controls without L. gibba biomass, thereby corroborating that the removal of Ni 2+ and/or Zn 2+ ions in the experiments using L. gibba was caused solely by the macrophyte biosorbent.
The equilibrium Ni 2+ and Zn 2+ biosorption capacities (q e , mmol g −1 ) were quantified according to the following equation (Equation (2)): where C 0 and C e are the initial and equilibrium Ni 2+ or Zn 2+ concentrations of the liquid phase (mmol L −1 ), respectively; and M is the L. gibba concentration (g L −1 ). The quantitative simultaneous biosorption effect of both the heavy metals on L. gibba was evaluated according to global index of behavior (ξ), and ratio of biosorption capacities (R q ), which were estimated as follows: where ξ Ni and ξ Zn are the global indexes of behavior for Ni 2+ and Zn 2+ biosorption at different initial concentrations of Zn 2+ and Ni 2+ , respectively; (q eNi dC 0 Ni) m and (q eNi dC 0 Ni) b are the variations of the equilibrium biosorption capacity for Ni 2+ over initial Ni 2+ concentration in single and binary metal experiments, respectively; (q eZn dC 0 Zn) m and (q eZn dC 0 Zn) b are the variations of the equilibrium Zn 2+ biosorption capacity over initial Zn 2+ concentration in single and binary metal experiments, respectively; q eNi,m and q eZn,m are the equilibrium Ni 2+ and Zn 2+ biosorption capacities in the single metal systems respectively; and q eNi,b and q eZn,b are the equilibrium Ni 2+ and Zn 2+ biosorption capacities in the binary metal systems, respectively. The experimental curves represented by (q eNi dC 0 Ni) m , (q eNi dC 0 Ni) b , (q eZn dC 0 Zn) m , and (q eZn dC 0 Zn) b were integrated using scientific and statistical software (GraphPad Prism 8.4, GraphPad Software, San Diego, CA, USA, 2020). ξ = 0 indicates that the biosorption of one heavy metal is not affected by the other heavy metal present in the binary metal solution; if ξ > 0, that signifies an improvement in the biosorption of a heavy metal in the bimetal system as compared to the monometal system (i.e., positive or synergic effect); and if ξ < 0, that denotes a diminution in the biosorption of a heavy metal in the bimetal system compared to the monometal system (i.e., negative or antagonist effect). Furthermore, R q = 1.0 indicates that the biosorption of each heavy metal is independent (i.e., non-interaction between Ni 2+ and Zn 2+ ions); if R q > 1.0, that indicates that the biosorption of one heavy metal is enhanced by the presence of the other heavy metal (i.e., synergic effect); and if R q < 1.0, it denotes that the biosorption of one heavy metal is suppressed by the other heavy metal present in the metal mixture (i.e., antagonist effect).

Modeling of Single and Binary Equilibrium Biosorption Isotherms of Ni 2+ and Zn 2+ Ions
Modeling experimental equilibrium biosorption isotherm data is an essential tool for trustworthy derivation of design parameters, the comparison of quantitative biosorption performances of distinct biosorption systems, and for diverse conditions in each biosorption system. These equilibrium biosorption isotherm parameters frequently provide some understanding of the affinity and surface characteristics of the biosorbent, in addition to the interactions between the biosorbent and adsorbate, and are consequently required to understand the biosorption mechanism [24]. Therefore, it is critical to establish the most suitable biosorption isotherm model that can be used for optimizing the operation conditions and designing of future large-scale biosorption facilities [24].

Mono-Component Isotherm Models
Equilibrium data for biosorption of single components can be commonly interpreted by the Freundlich, Langmuir, Redlich-Peterson, and Sips isotherms.

Freundlich Isotherm Model
The Freundlich model describes non-ideal, multi-layer biosorption onto heterogeneous surfaces, indicating that biosorption sites are neither independent nor equivalent [25,26]. The Freundlich isotherm model is provided below (Equation (7)): where the Freundlich parameters 1/n F and K F are relative indicators of intensity and capacity of biosorption, respectively [25].

Langmuir Isotherm Model
This isotherm model has the following assumptions: (1) each biosorption site can adsorb only one molecule of adsorbate, and consequently, only one layer of biosorbed molecules can be formed on the biosorbent surface; (2) there are no interactions between biosorbed species; (3) all the biosorption sites are identical; (4) the capability of a molecule to be biosorbed onto a given biosorption site is not dependent on the occupancy of its neighboring sites; and (5) the biosorptive forces are similar to those found in chemical interactions [26,27]. The Langmuir equation can be expressed as follows (Equation (8)): where q max is the theoretical maximum biosorption capacity of the biosorbent; q e is the biosorption capacity at the equilibrium; C e is the equilibrium adsorbate concentration in the liquid phase; and B L is the Langmuir constant that represents the reciprocal of the adsorbate concentration at which half saturation of the biosorbent is reached [25].

Sips Isotherm Model
The Sips equation is a hybrid model of the Freundlich and Langmuir isotherms, used extensively to interpret equilibrium biosorption data. It has the following form (Equation (9)) [28]: q e = q max B S C e n S 1 + (B S C e n S ) (9) where q max is the maximum biosorption capacity; B s is the Sips constant; and n s is the Sips parameter that characterizes the system's heterogeneity.

Redlich-Peterson Isotherm Model
The Redlich-Peterson isotherm equation is shown in Equation (10) [27]: where K RP , A RP , and b RP are the model parameters. This model reduces to the Langmuir isotherm when b RP = 1, to the Freundlich isotherm at high concentration of the adsorbate, and to a linear isotherm at low surface coverage.

Multi-Component Isotherm Models
In the case of binary metal or multi-metal systems, as is representative of industrial wastewaters, there may be interference and competition between different components for biosorption binding sites, resulting in a more complex mathematical description of the dynamic equilibrium state [29]. Therefore, the isotherm models proposed for single-component biosorption equilibrium have been modified and extended to represent the binary component and multi-component biosorption equilibria over a wide range of solution concentrations [30]. However, despite its great significance for the design of treatment systems, the equilibrium modeling of binary component and multi-component biosorption has largely been neglected [29].
Seven different multi-component isotherm models were applied to describe the simultaneous biosorption of Ni 2+ and Zn 2+ ions onto L. gibba, namely, non-modified competitive Langmuir, modified competitive Langmuir, modified Langmuir with an interaction factor, extended Freundlich, non-modified competitive Redlich-Peterson, modified Redlich-Peterson with interaction factor, and modified Sips isotherm models.

Non-Modified Competitive Langmuir Isotherm Model
This model has the same assumptions as the monocomponent Langmuir isotherm equation and can be expressed for components Ni 2+ and Zn 2+ in a mixture of both heavy metals as follows [31]: where q eNi and q eZn are the equilibrium biosorption capacities of Ni 2+ and Zn 2+ in the bimetal system; C eNi and C eZn are the equilibrium Ni 2+ and Zn 2+ concentrations in binary metal solution; B Ni and B Zn are the constants of the Langmuir model for Ni 2+ and Zn 2+ biosorption in monometal systems; and q maxNi and q maxZn are the theoretical maximum biosorption capacities of Ni 2+ and Zn 2+ in the single metal systems, respectively [32].

Modified Competitive Langmuir Isotherm
Assuming the active biosorption sites are distributed uniformly on the biosorbent surface, and all the adsorbates in the liquid phase compete for the same biosorption sites of the biosorbent, the Langmuir model for single-component systems was extended for multicomponent systems as follows [30]: The parameters q maxNi , q maxZn , B Ni , and B Zn can be obtained from the fitting of Equations (13) and (14) with the experimental equilibrium biosorption data of Ni 2+ and Zn 2+ in the binary metal system [30].

Modified Langmuir Isotherm with an Interaction Factor η
In multicomponent biosorption systems, the biosorbent affinity for the adsorbates changes due to biosorbent-adsorbate interactions and the competition of the heavy metal ions species for biosorption binding sites. Considering these effects, an interaction term (η) for each adsorbate has been introduced into Equations (11) and (12) to yield Equations (15) and (16) [31].
where η Ni and η Zn are characteristic of Ni 2+ and Zn 2+ ions and dependent on the concentrations of the other components in the solution [30]. These characteristic parameters can be obtained from the experimental competitive biosorption data of Ni 2+ and Zn 2+ ions.

Extended Freundlich Isotherm
The single-component Freundlich equation has been extended to represent the binary component biosorption system, and is expressed according to Equations (17) and (18) for each heavy metal of the binary mixture of adsorbates [30]: where K FNi , K FZn , n FNi , and n FZn are Freundlich parameters that are obtained from the corresponding single-component Freundlich isotherm models; whereas x Ni , y Ni , z Ni , and x Zn , y Zn , and z Zn are the competitive Freundlich biosorption parameters for Ni 2+ and Zn 2+ ions, respectively.

Modified Sips Isotherm
The single-component Sips isotherm model can be extended to describe the biosorption in binary-component systems [32]: q eZn = q maxZn B sZn C eZn n sZn 1 + B sNi C eNi n sNi + B sZn C eZn n sZn (20) where the modified Sips model parameters q maxNi , q maxZn , B sNi , B sZn , n sNi , and n sZn are obtained from binary component biosorption data [32].

Non-Modified Competitive Redlich-Peterson Isotherm
The competitive Redlich-Peterson isotherm models for binary component biosorption systems are given in Equations (21) and (22): where K RPNi , K RPZn , A RPNi , A RPZn , b RPNi , and b RPZn are the single-component Redlich-Peterson model parameters for the Ni 2+ and Zn 2+ ions components [33].

Modified Redlich-Peterson Isotherm with Interaction Factor η
The competitive Redlich-Peterson isotherm model can be extended with an interaction factor η, in a manner analogous to that of the modified Langmuir isotherm model, and can be expressed for components Ni 2+ and Zn 2+ in a bimetal system [34] as follows: where η Ni and η Zn are the correction coefficients for Ni 2+ and Zn 2+ ions estimated from binary metal biosorption data; and K RPNi , K RPZn , A RPNi , A RPZn , b RPNi , and b RPZN are derived from the corresponding single-component biosorption data [34].

Determination of Isotherm Parameters and Statistical Analysis
Batch biosorption experiments were conducted in triplicate to confirm the reproducibility of the results and for statistical analysis. The data reported herein represent the mean values ± standard deviations. Biosorption data of Ni 2+ and Zn 2+ were analyzed by variance analysis and Tukey's test (p < 0.05), using GraphPad Prism 8.4 (GraphPad Software, San Diego, CA, USA, 2020).
The isotherm parameters of all the single-metal and binary-metal models were obtained by nonlinear regression analysis using the software MATLAB ® (version 2020a; The Mathworks, Inc., Natick, MA, USA, 2020). The standard deviation of the residuals (RMSE and Sy.x), the sum of squares error (SSE), and the coefficient of determination (R 2 ) were used to determine the adequacy and accuracy of the fit of the various isotherm models assayed with the experimental data. Small RMSE, Sy.x, SSE values, and a value of R 2 near 1.0 indicate a better description of experimental biosorption data by a particular model [28].

FTIR Analysis
FTIR spectroscopy was employed to gain information about the main chemical functional groups present on the L. gibba surface, and to detect possible modifications of active biosorption sites on the L. gibba surface during the biosorption of Ni 2+ , Zn 2+ , or the binary mixture of Ni 2+ and Zn 2+ ions.
L. gibba (1 g L −1 ) samples were mixed with aqueous solutions of 10 mM Ni 2+ , 10 mM Zn 2+ , or with 10 mM solutions of equimolar mixtures of Ni 2+ and Zn 2+ ions at pH 6.0 ± 0.1 for 24 h, with constant agitation at 180 rpm and 20 ± 1°C to saturate the active biosorption sites with Ni 2+ , Zn 2+ , or with both Ni 2+ and Zn 2+ ions. The resulting suspensions were subsequently centrifuged for 5 min at 5000 rpm to pellet the L. gibba biomass. The pellets collected were washed thoroughly with distilled deionized water to remove the unbound heavy metals. After centrifuging the resulting suspensions, the obtained heavy metal-loaded L. gibba biomass was dried at 105°C to remove any retained water which could interfere with the detection of hydroxyl functional groups on the L. gibba surface [35].
Finely ground raw (i.e., not loaded with Ni 2+ and/or Zn 2+ ions) and heavy metal-loaded L. gibba samples were mixed with dried potassium bromide (Merck™, Germany; spectroscopic grade) in a ratio of 1:5 and examined by diffuse reflectance FTIR spectroscopy in a Perkin-Elmer FTIR spectrophotometer (Spectrum 2000, Perkin-Elmer, Inc., Waltham, MA, USA). FTIR spectra data were collected over the wavenumber range of 4000-400 cm −1 , at a resolution of 4 cm −1 , for a total of 16 scans.

SEM and SEM-EDX Analysis
The surface structures of L. gibba samples, prior to and following saturation with Ni 2+ and/or Zn 2+ ions, were studied using a scanning electron microscope (SEM; FEI Quanta Dual Beam Instrument, FEI Company, Hillsboro, OR, USA) with an accelerated voltage of 15 kV. The analysis of chemical elements was conducted by energy dispersive X-ray spectrometry (EDX) using an Apollo X detector (EDAX, AMETEK, Inc., Berwyn, PA, USA). For the analysis, samples remained uncoated at low vacuum conditions to improve electron conductivity and image quality.

Analytical Techniques
Ni 2+ and Zn 2+ concentrations in liquid solution were assessed colorimetrically using a UV-Vis Evolution 201 (Thermo Fisher Scientific, Whaltman, MA, USA) spectrophotometer, following the methodology outlined in the Hach Water Analysis Handbook [36]. Ni 2+ and Zn 2+ concentrations were determined from calibration curves established for each heavy metal with a minimum of seven concentration levels.

Proximate Composition of L. gibba
The analysis results of the proximate chemical composition of L. gibba are as follows (on a dry matter basis): total carbohydrates, 46.57 ± 1.5%; total proteins, 26.29 ± 0.51%; total ashes, 19.04 ± 0.41%; and total lipids, 8.10 ± 0.05%. It is apparent that L. gibba is rich in carbohydrates and proteins and has a low fat content. These findings concur with those reported for the Lemnaceae by Schmid and Landolt [37] and Landesman [38]. The polymeric compounds present in L. gibba biomass have many chemical functional groups that can be binding sites for the biosorptive removal of toxic metal ions from aqueous solutions.

Influences of Metal Solution pH on the Individual and Simultaneous Biosorption of Ni 2+ and Zn 2+ onto L. gibba
The pH of a metal ion solution is among the crucial environmental factors that affect heavy metal biosorption from liquid solutions due to its ability to influence the chemical speciation, degree of ionization, and aqueous solubility of heavy metals. Metal ion solution pH also influences competition with other heavy metal ions present in the liquid phase and the dissociation degree of chemical functional groups present on the surface of biosorbents [11,12].
To find the optimum value of solution pH for Ni 2+ and Zn 2+ biosorption onto L. gibba for the single metal and binary metal systems, biosorption tests with the solution pH varying from 1.0 to 6.0 were performed. The influences of the pH of the aqueous metal solution on the equilibrium Ni 2+ and Zn 2+ biosorption capacities for the single and binary heavy metal systems are illustrated in Figure 1. It is apparent that the equilibrium biosorption capacities of Ni 2+ and Zn 2+ ions increased as the solution pH increased and reached its highest value at pH 6.0, both for the mono-metal and bi-metal systems. This biosorption behavior may be due to the fact that at low solution pH levels, the active binding sites available on the L. gibba surface are mostly protonated, resulting in electrostatic repulsion between the Ni 2+ and Zn 2+ cationic species and the positive electrical surface charges of L. gibba, and the competition between H + ions and Ni 2+ and Zn 2+ ions for occupancy of the binding sites. Contrastingly, the extent of Ni 2+ and Zn 2+ biosorption increases with increasing solution pH presumably due to a decrease in competition between H + ions and Ni 2+ and Zn 2+ ions for the biosorption binding sites, and also due to an increase in the number of negative electrical charges on the L. gibba surface, which results in a greater electrostatic attraction between the L. gibba surface and Ni 2+ and Zn 2+ ions [11].
Previous studies have reported an optimal solution pH of 6.0 for Ni 2+ and Zn 2+ biosorption in single metal systems [39][40][41]. However, to the best knowledge of the authors, no works have investigated the influence of pH of the metal solution on the simultaneous biosorption of Ni 2+ and Zn 2+ ions from an aqueous solution.
The capacities of Zn 2+ biosorption were higher than those of Ni 2+ at pH values greater than 3.0 and 4.0, for the single metal and binary metal systems, respectively. At solution pH values of 5.0 and 6.0, no statistically significant difference was observed in Zn 2+ biosorption capacities between single and binary metal systems, which indicates that Zn 2+ biosorption onto L. gibba was not significantly affected by the presence of Ni 2+ ions at the studied conditions. Contrastingly, a noticeable decrease in biosorption capacity of Ni 2+ was observed when Zn 2+ ions were present in the metal solution with pH values between 3.0 and 6.0.
The results show that the solution pH of 6.0 is most suitable for the single and binary Ni 2+ and Zn 2+ biosorption onto L. gibba. Consequently, all subsequent experiments were conducted with a It is apparent that the equilibrium biosorption capacities of Ni 2+ and Zn 2+ ions increased as the solution pH increased and reached its highest value at pH 6.0, both for the mono-metal and bi-metal systems. This biosorption behavior may be due to the fact that at low solution pH levels, the active binding sites available on the L. gibba surface are mostly protonated, resulting in electrostatic repulsion between the Ni 2+ and Zn 2+ cationic species and the positive electrical surface charges of L. gibba, and the competition between H + ions and Ni 2+ and Zn 2+ ions for occupancy of the binding sites. Contrastingly, the extent of Ni 2+ and Zn 2+ biosorption increases with increasing solution pH presumably due to a decrease in competition between H + ions and Ni 2+ and Zn 2+ ions for the biosorption binding sites, and also due to an increase in the number of negative electrical charges on the L. gibba surface, which results in a greater electrostatic attraction between the L. gibba surface and Ni 2+ and Zn 2+ ions [11].
Previous studies have reported an optimal solution pH of 6.0 for Ni 2+ and Zn 2+ biosorption in single metal systems [39][40][41]. However, to the best knowledge of the authors, no works have investigated the influence of pH of the metal solution on the simultaneous biosorption of Ni 2+ and Zn 2+ ions from an aqueous solution.
The capacities of Zn 2+ biosorption were higher than those of Ni 2+ at pH values greater than 3.0 and 4.0, for the single metal and binary metal systems, respectively. At solution pH values of 5.0 and 6.0, no statistically significant difference was observed in Zn 2+ biosorption capacities between single and binary metal systems, which indicates that Zn 2+ biosorption onto L. gibba was not significantly affected by the presence of Ni 2+ ions at the studied conditions. Contrastingly, a noticeable decrease in biosorption capacity of Ni 2+ was observed when Zn 2+ ions were present in the metal solution with pH values between 3.0 and 6.0.
The results show that the solution pH of 6.0 is most suitable for the single and binary Ni 2+ and Zn 2+ biosorption onto L. gibba. Consequently, all subsequent experiments were conducted with a Zn 2+ ) metal systems.
It is apparent that the equilibrium biosorption capacities of Ni 2+ and Zn 2+ ions increased as the solution pH increased and reached its highest value at pH 6.0, both for the mono-metal and bi-metal systems. This biosorption behavior may be due to the fact that at low solution pH levels, the active binding sites available on the L. gibba surface are mostly protonated, resulting in electrostatic repulsion between the Ni 2+ and Zn 2+ cationic species and the positive electrical surface charges of L. gibba, and the competition between H + ions and Ni 2+ and Zn 2+ ions for occupancy of the binding sites. Contrastingly, the extent of Ni 2+ and Zn 2+ biosorption increases with increasing solution pH presumably due to a decrease in competition between H + ions and Ni 2+ and Zn 2+ ions for the biosorption binding sites, and also due to an increase in the number of negative electrical charges on the L. gibba surface, which results in a greater electrostatic attraction between the L. gibba surface and Ni 2+ and Zn 2+ ions [11].
Previous studies have reported an optimal solution pH of 6.0 for Ni 2+ and Zn 2+ biosorption in single metal systems [39][40][41]. However, to the best knowledge of the authors, no works have investigated the influence of pH of the metal solution on the simultaneous biosorption of Ni 2+ and Zn 2+ ions from an aqueous solution.
The capacities of Zn 2+ biosorption were higher than those of Ni 2+ at pH values greater than 3.0 and 4.0, for the single metal and binary metal systems, respectively. At solution pH values of 5.0 and 6.0, no statistically significant difference was observed in Zn 2+ biosorption capacities between single and binary metal systems, which indicates that Zn 2+ biosorption onto L. gibba was not significantly affected by the presence of Ni 2+ ions at the studied conditions. Contrastingly, a noticeable decrease in biosorption capacity of Ni 2+ was observed when Zn 2+ ions were present in the metal solution with pH values between 3.0 and 6.0.
The results show that the solution pH of 6.0 is most suitable for the single and binary Ni 2+ and Zn 2+ biosorption onto L. gibba. Consequently, all subsequent experiments were conducted with a solution pH of 6.0.

Influence of Initial Metal Concentration in the Single and Binary Biosorption Systems of Ni 2+ and Zn 2+ onto L. gibba
The influences of initial concentrations of Ni 2+ and Zn 2+ on the L. gibba equilibrium biosorption capacities of the heavy metals in monometal and bimetal systems are displayed in Figures 2 and 3, respectively. The capacity of metal biosorption of L. gibba increased as the initial concentration of Ni 2+ or Zn 2+ in single ( Figure 2) and binary metal systems increased (Figure 3). These results can be attributed to the fact that the greater the initial concentration of Ni 2+ or Zn 2+ ions, the greater the number of Ni 2+ or Zn 2+ ions in the aqueous metal solution, which results in a greater number of interactions between the metal ions and the L. gibba biomass. Additionally, a higher initial Ni 2+ or Zn 2+ concentration also causes metal concentration gradient increases, resulting in a faster transfer of Ni 2+ or Zn 2+ ions from the bulk of the liquid phase to the L. gibba surface due to an enhanced mass transfer coefficient, along with providing the driving force to overwhelm the mass transfer resistances of Ni 2+ or Zn 2+ ions between the liquid phase and the L. gibba biomass [11,12]. Figure 2 shows that L. gibba's Zn 2+ biosorption capacity was higher than Ni 2+ biosorption capacity up to a 15 mM initial metal concentration in the single metal systems. At higher initial metal concentrations, the Ni 2+ and Zn 2+ biosorption capacities were similar. However, the initial slope of the curve that relates Zn 2+ biosorption capacity as a function of initial Zn 2+ concentration was higher than the initial slope of the curve relating biosorption capacity of Ni 2+ ions to initial Ni 2+ concentration (Figure 2), which indicates that L. gibba biomass has a greater affinity for Zn 2+ ions than for Ni 2+ ions. Similarly, the results of simultaneous biosorption of Ni 2+ and Zn 2+ ions also show that the biosorption of a heavy metal was affected to a differing extent by the other heavy metal present in the metal solution, and that L. gibba biosorption capacity for Zn 2+ ions was higher than that for Ni 2+ ions ( Figure 3).
Regardless of the initial concentration of Ni 2+ assayed, the L. gibba Ni 2+ biosorption capacity at equilibrium decreased with the increasing of initial Zn 2+ concentration (Figure 3a). Similarly, the global behavior index for the biosorption of Ni 2+ ions decreased from −2.1% to −60.3% as the initial concentration of Zn 2+ ions increased from 0.2 to 10 mM (Table 1). This indicates that the greater the initial Zn 2+ concentration, the higher the adverse effect on Ni 2+ biosorption. Furthermore, Table 2 displays that R qNi values are far less than 1.0 at almost all initial Zn(II) concentrations, indicating that Zn 2+ interfered with Ni 2+ biosorption onto L. gibba. These results show that Zn 2+ inhibited biosorption of Ni 2+ ions. Similarly, Figure 3b and global behavior index data (Table 3) of Zn 2+ biosorption in the presence of Ni 2+ ions revealed that the initial concentrations of Ni 2+ in the range of 0.2 to 2.0 mM have a far lower adverse effect on Zn 2+ biosorption onto L. gibba than initial Ni 2+ concentrations higher than 2.0 mM. Furthermore, the inhibitory influence of Ni 2+ on Zn 2+ biosorption was only evident when the initial concentration of Ni 2+ was greater than the equimolar ratio (Table 4). Hence, these results confirm that L. gibba biomass has a higher preference for Zn 2+ ions than for Ni 2+ ions, and that Zn 2+ biosorption was almost free from interferences from Ni 2+ ions at concentration ratios lower than or equal to the equimolar ratio, at the reported working conditions.
The higher biosorption preference of L. gibba for Zn 2+ ions over Ni 2+ ions could be attributed to the higher ionic radius (Zn, 74 pm; Ni, 69 pm) and atomic weight (Zn, 65.4; Ni, 58.7) of Zn 2+ than of Ni 2+ . Generally, the higher the ionic size and atomic weight of a heavy metal, the higher the affinity of a biosorbent for the heavy metal [42,43]. Consequently, the preference of a biosorbent for biosorbing the heavy metal will be maximum [44]. This may be due to the fact that the motion of a heavy metal has a greater atomic mass generating a greater energy of momentum, which can facilitate the binding of the heavy metal to the biosorbent by increasing the probability of interaction between the binding sites of the biosorbent and the heavy metal, leading to a higher biosorption preference by a biosorbent [44].
A higher affinity for the biosorption of Zn 2+ ions than for Ni 2+ ions has been previously reported [45].

Single-Metal Biosorption Systems
The equilibrium biosorption isotherm provides important information on the interactions between the biosorbent and the adsorbate (e.g., heavy metal) and is consequently needed to understand the mechanism of biosorption [25]. Figure 4 presents the experimental biosorption isotherm of Ni 2+ (Figure 4a) and Zn 2+ (Figure 4b) ions at pH 6.0 and at 20°C, for the single metal systems.
It is evident that the biosorption capacities at the equilibrium of Ni 2+ and Zn 2+ increased as the equilibrium concentrations of Ni 2+ and Zn 2+ ions in the solution increased. The initial slopes of the Ni 2+ and Zn 2+ biosorption isotherms steadily decrease with increasing equilibrium metal concentrations because metal ions have more difficulty finding vacant active biosorption sites with the progressive covering of the L. gibba surface. This dynamic indicates a progressive occupation of L. gibba's active sites by Ni 2+ or Zn 2+ ions [11]. In addition, the initial slope of the Zn 2+ isotherm was steeper than that of the Ni 2+ isotherm, which is indicative of a higher affinity of L. gibba biomass for Zn 2+ ions than for Ni 2+ ions.
Furthermore, the analysis of biosorption equilibrium data is essential to establish the most suitable isotherm equation that characterizes precisely the process of biosorption, which can also be employed to optimize the design and operating conditions of a biosorption system [11,25].
Isotherm models with two (Freundlich and Langmuir) and three (Redlich-Peterson and Sips) parameters were assessed for their ability to fit the experimental data of Ni 2+ and Zn 2+ equilibrium biosorption onto L. gibba. The biosorption parameters for each metal ion obtained from the non-linear regression analysis of the isotherm models employed in this work are shown in Table 5, together with the values of R 2 , Sy.x, RMSE, and SSE. Likewise, Figure 4 displays the isotherm curves predicted by the four isotherm models. The Freundlich, Redlich-Peterson, and Sips isotherm equations best described Ni 2+ biosorption data, as evidenced by the lowest Sy.x, RMSE, and SSE and the highest R 2 values of these models. Nonetheless, the Sips model was not capable of predicting the experimental maximum Ni 2+ biosorption capacity (q max = 0.91 mmol g −1 ). Therefore, the Sips model is unsuitable for interpreting the experimental equilibrium isotherm for the biosorption of Ni 2+ ions onto L. gibba. The Freundlich model parameter n F , which is related to the biosorption intensity, was found to be 2.59. This value is within the range of 1 to 10, which suggests that Ni 2+ biosorption by L. gibba is favorable at the conditions being studied [11]. Furthermore, according to the error function (R 2 , RMSE, Sy.x, and SSE) values, the Langmuir, Redlich-Peterson, and Sips isotherm equations better represented the experimental equilibrium data of Zn 2+ biosorption onto L. gibba than the Freundlich model. However, the Sips model was unable to adequately predict the maximum equilibrium Zn 2+ biosorption capacity obtained experimentally (0.88 mmol g −1 ). Thus, the Sips model is not appropriate for describing the biosorption isotherm of Zn 2+ ions onto L. gibba. Contrastingly, the Langmuir model predicted a maximum biosorption capacity of Zn 2+ ions of 0.852 mmol g −1 , which closely matched the experimental value. Similarly, the exponent value of the Redlich-Peterson (b RP = 0.885) model was close to 1.0. In this case, the Redlich-Peterson model is transformed to the Langmuir equation [28].
Considering these results, it is concluded that the Freundlich and Redlich-Peterson models provided the most suitable representations of the biosorption isotherm of Ni 2+ ions by L. gibba, whereas the Langmuir and Redlich-Peterson models best described the equilibrium biosorption pattern for Zn 2+ biosorption.
Furthermore, the values of the constants B L of Langmuir equation and n F of Freundlich equation for Zn 2+ biosorption are higher than those for Ni 2+ biosorption. These results confirm that L. gibba has a greater affinity for Zn 2+ ions than for Ni 2+ ions, as the higher the values of B L and n F , the higher the affinity between the adsorbate and biosorbent [30,40].

Binary-Metal Biosorption Systems
For the purpose of representing the equilibrium data of binary biosorption of Ni 2+ and Zn 2+ on L. gibba in a suitable graphical method, three-dimensional biosorption surfaces were built, plotting the capacity of Ni 2+ or Zn 2+ biosorption at equilibrium (z-axis) versus the Ni 2+ and Zn 2+ concentrations at equilibrium (x-and y-axis). Figure 5 displays the three-dimensional biosorption isotherm surfaces of the Ni 2+ -Zn 2+ bimetal system. 3D plots confirmed that biosorption of Ni 2+ ions was negatively influenced to a great extent by Zn 2+ ions present in aqueous solution (Figure 5a), whereas the presence of Ni 2+ in the liquid phase had a less pronounced influence on Zn 2+ biosorption (Figure 5b). This difference can be attributed to the competition between Ni 2+ and Zn 2+ ions for the same active biosorption sites on the L. gibba biomass surface and to the higher affinity of L. gibba for Zn 2+ than for Ni 2+ . Moreover, these results clearly show that an antagonistic competitive effect occurs in the biosorption of each single heavy metal ion (Figure 5a,b), particularly at the high concentration range. Table 6 reports the calculated parameter values of the different multicomponent isotherm models assayed in this work for the binary biosorption of Ni 2+ and Zn 2+ ions onto L. gibba, together with the corresponding error functions (i.e., R 2 , Sy.x, RMSE, and SSE values). The comparative analysis of the error function values indicates that several of the applied models can be used for the estimation of the biosorption behavior of Ni 2+ -Zn 2+ binary mixtures on L. gibba. However, the modified Sips model gives the best fitting for this binary metal system. The agreement between experimental results and those calculated by the modified Sips model is shown in Figure 5 (points and mesh surfaces, respectively). The modified Sips model predicted maximum capacities for Ni 2+ and Zn 2+ biosorption of 0.75 and 0.764 mmol g −1 , respectively, for the binary metal system.

FTIR Analysis
FTIR spectroscopy analyses were performed on raw (i.e., not loaded with Ni 2+ or Zn 2+ ions), Ni 2+ -loaded, Zn 2+ -loaded, and (Ni 2+ + Zn 2+ )-loaded L. gibba samples to determine the chemical functional groups present on the L. gibba surface that might play a key role in the biosorptive removal of Ni 2+ and Zn 2+ ions from the liquid phases. To achieve this purpose, the FTIR spectrum of raw biosorbent was compared with that of Ni 2+ -loaded, Zn 2+ -loaded, and (Ni 2+ + Zn 2+ )-loaded L. gibba ( Figure 6). The abundant infrared absorption bands within the range of 4000-400 cm −1 in the FTIR spectrum of raw L. gibba is indicative of the complex chemical composition of its surface. The broad FTIR band at 3800-2800 cm −1 , with a minimum at 3294 cm −1 , corresponds to the stretching vibration of hydrogen-bonded -OH and -NH functional groups. This infrared absorption band is associated with the vibrations of the linked -OH groups in lignin, cellulose, and hemicellulose, which are present in Lemnaceae, and water adsorbed on the surface of lignocellulosic biomass [35,46]. Likewise, the -NH stretching vibration band in the FTIR spectrum of raw L. gibba indicated the presence of amino acids in the duckweed biomass, which are the building blocks of proteins. These polymeric compounds constitute approximately 29.26% of the dry biomass of L. gibba.
At approximately 2921 cm −1 , the absorption band is attributed to the stretching vibration of CH 2 and C-H functional groups found in lignin [35]. The absorption bands at approximately 1656 cm −1 and 1544 cm −1 are indicative of C=O stretching and N-H bending of amide I and amide II functional groups, respectively [35]. The C-O stretching of cellulose, hemicellulose, and proteins, and the -SO 3 stretching vibration of hemicelluloses, were detected at 1448 and 1240 cm −1 , respectively [35]. The absorption band of C-H vibration of cellulose and hemicellulose at approximately 1321 cm −1 [35,47] was also found in the FTIR spectrum of raw L. gibba. In addition, the FTIR band at 1092 cm −1 was assigned to the vibrations of O-H and C-O-C groups of polysaccharides [41].

SEM-EDX Analysis
SEM was employed to examine the physical morphology of the L. gibba surface before and after metal biosorption. SEM micrographs (Figure 7) displayed differences in surface morphology between the raw (Figure 7a) and the heavy metal-loaded L. gibba (Figure 7b-d).
The raw biosorbent showed a relatively regular surface morphology. Contrastingly, the morphology of the biosorbent surface changed after metal biosorption, becoming heterogeneous and smooth. These results are in agreement with other studies that have also found differences between the surface morphology of metal-unloaded and metal-loaded biosorbents [48,49].
Results of EDX analysis showed the presence of carbon, oxygen, nitrogen, sodium, magnesium, silicon, sulfur, chlorine, and potassium, and the absence of nickel and/or zinc on the L. gibba surface before interacting with heavy metals (Figure 7a) Furthermore, SEM and elemental mappings of Ni 2+ and Zn 2+ on a particular surface spot after heavy metal biosorption are provided in Figure 8.
Bright spots denote the presence of nickel (Figure 8a,c) or zinc (Figure 8b,d) on the L. gibba surface. It is apparent that the biosorbed Ni and Zn ions were not distributed uniformly on the L. gibba surface, indicating that only some chemical functional groups present on the L. gibba surface were capable of biosorbing the divalent metal ions from the aqueous solutions. Therefore, SEM-EDX analysis confirmed the biosorption of Ni 2+ and Zn 2+ ions onto L. gibba.

Conclusions
This study demonstrates that L. gibba is an effective and promising biosorbent for the simultaneous removal of Ni 2+ and Zn 2+ ions from liquid solutions. Solution pH value has a significant influence on Ni 2+ and Zn 2+ biosorption by L. gibba, both in the single and in binary metal systems. The maximum levels of divalent heavy metal biosorption occurred at pH 6.0. In both the single and binary metal biosorption systems, the affinity of L. gibba for Zn 2+ ions was greater than for Ni 2+ ions. In the binary metal biosorption systems, it was observed that Zn 2+ ions exhibited a strong antagonistic effect on Ni 2+ biosorption. Contrastingly, the interference of Ni 2+ ions on Zn 2+ biosorption was less significant. Redlich-Peterson and Freundlich models were found to better fit Ni 2+ equilibrium data, whereas Redlich-Peterson and Langmuir models provided the best fit of the Zn 2+ equilibrium data in single metal biosorption systems. The simultaneous biosorption of Ni 2+ and Zn 2+ ions onto the L. gibba was adequately and satisfactorily described by the modified competitive Sips model. FTIR results suggest that Ni 2+ and Zn 2+ ions bind to L. gibba's polysaccharides, and SEM-EDX analysis validated the presence of the heavy metals on the surface of duckweed. Both Ni 2+ and Zn 2+ ions could be desorbed easily from heavy metal-loaded L. gibba using acid solutions, and facilitate heavy metal recovery and biosorbent regeneration.