Comparison of Heavy Metals Removal from Aqueous Solution by Moringa oleifera Leaves and Seeds

: In this work, biomass obtained from seeds (S-MO) and leaves (L-MO) of the Moringa oleifera plant were used as low-cost biosorbents to remove the Pb(II), Cd(II), Co(II), and Ni(II) from aqueous solutions. The biosorption of the heavy metal ions was done using the batch technique. The effects of contact time (30–1440 min), biosorbent dosage (10–50 g/L) (0.1–0.5 g), and initial concentration of metals (10–500 mg/L) on the sorption capacity of metal ions were investigated. The S-MO and L-MO samples’ characterization was performed using pH pzc , X-ray Diffraction (XRD), and Fourier Transform Infrared Spectroscopy (FTIR). It was found that the pH pzc was notably different between the seeds and leave-derived biosorbents. The removal process’s experimental kinetic data for both S-MO and L-MO were best described by the pseudo-second-order model for all metal ions, with R 2 above 0.997 in all cases. Langmuir and Freundlich’s models were also used to analyze the isotherms parameters. Based on the Langmuir model, the maximum sorption capacities (Qm) for L-MO were found as follows: L-MO-Pb > L-MO-Cd > L-MO-Co ≥ L-MO-Ni, and for S-MO, the values of Qm values presented the following order: S-MO-Pb > S-MO-Co > S-MO-Cd > S-MO-Ni.


Introduction
Heavy metals, like Pb, Cd, Co, and Ni, are considered a primary source of contamination in aquatic ecosystems and a major environmental problem due to their elevated toxicity and their persistence in the natural waters. Beyond even a low concentration, their presence in aquatic and soil matrixes can cause undesirable effects on human and animal health. Therefore, the United States Environmental Protection Agency (US EPA) established that the maximum concentration of the aforementioned heavy metal ions in wastewater must be 1.7 µg/L for Co(II), 5 µg/L for Cd(II), 15 µg/L for Pb(II), and 20 µg/L for Ni(II) [1,2]. However, concentration values up to 200 µg/L, 20 µg/L, 900 µg/L, and 670 µg/L of Pb(II), Cd(II), Co(II), and Ni(II), respectively, were reported in different effluents due to the uncontrolled discharge of these pollutants in the wastewater [1,2].

Characterization Techniques
The as-received and dried samples of seeds and leaves of Moringa oilefera were characterized by Fourier transform infrared spectroscopy (Nicolet Nexus 670 FTIR, Waltham, MA, USA) within the range of 4000-400 cm -1 using the KBr standard method. Spectra were collected for the pristine materials and the exposed ones. The L-MO and S-MO samples were also characterized by employing an X-ray diffractometer. The samples' XRD patterns were recorded at 30 kV and 25 mA in the range of 2θ from 5 • to 70 • at a 2 • /min scan rate.

Point of Zero Charge (pH PZC )
The following experiment process determined the pH pzc values of L-MO and S-MO biomass: an amount of 0.1 g of each biomass was added to a 50 mL of 0.01M NaCl solution at various initial pH values from 2 to 12. The mixtures were agitated for 72 h at 25 • C, followed, the phases were separated by decantation, and each solution's pH was measured using a pH-meter (Thermo Scientific ORION 3 star pH Benchtop, Waltham, MA, USA). The pH pzc value of the biosorbent is obtained by the intersection of the initial experimental pH vs. the final pH graph with the theoretical curve (pH initial = pH initial ) [32].

Sorption Experiments
In this work, the sorption experiments, such as the effects of contact time, biosorbent dosages, and the initial concentration of metallic ions, were performed by the batch system. For the kinetic study, a level of 0.10 g of each biomass (L-MO and S-MO) was added in conic tubes to 10 mL of each metallic ion with an initial concentration of 100 mg/L and at pH 6 in order to avoid metal precipitation. The mixtures were shaken at 25 • C in a rotary shaker at 150 rpm. The samples were centrifuged for 5 min at 3500 rpm to separate the solid and liquid phases after finalizing each adsorption experiment. The concentrations of Pb, Co, Cd, and Ni ions C f (mg/L) were determined by using the FAAS technique (Thermo Scientific). Equation (1) gives the amount of metallic ions sorbed per gram of adsorbent at each contact time t, q t (mg/g) by L-MO and S-MO: where C f (mg/L) is the concentration of adsorbate at time t (min), C i (mg/L) is the initial adsorbate concentration, V(L) is the volume of the adsorbate solution, and m(g) is the adsorbent's mass. In this case, C i = 100 mg/L, V = 0.01 L, and m = 0.1 g. The effect of biosorbent dosage in removing metallic ions was done by the variation of biosorbent mass from 0.1 g to 0.5 g/10 mL, keeping the other parameters constant (contact time at 300 min, solution pH at 6, and initial concentration of metal ions at 100 mg/L). The effect of the initial concentration of metal ions on the sorption capacity of the biosorbents was carried out by varying C i from 10 to 500 mg/L. Throughout these experiments, the contact time, solution pH, and mass biosorbents were fixed at 300 min, 5.5, and 0.1 g, respectively. All experiments were conducted in duplicate to assure the repeatability of the results.
The description sorption process was investigated by fitting the experimental data to the pseudo-first-order and pseudo-second-order models. These models are described by the following equations (Equations (2) and (3)), respectively: where q t (mg·g −1 ) and q e (mg·g −1 ) are the sorption capacities at time t and equilibrium, respectively; k 1 (min −1 ) is the pseudo-first-order, and k 2 (g mg −1 ·min −1 ) is the pseudosecond-order adsorption rate constant. Equations (4) and (5) describe the Langmuir and Freundlich models, respectively: where Q m is the maximum adsorption capacity (mg/g) and K L is the Langmuir constant (L/mg), q e is the amount of each heavy metal sorbed per unit mass of biosorbents, and C e is the amount of the heavy metals in solution at the equilibrium. K F (mg·g −1 ·(mg·L −1 ) −1/n ) and n are Freundich constants.

Point of Zero Charge (pH PZC )
The plot of pH pzc of L-MO and S-MO is depicted in Figure 1. The pH PZC of leaves and seeds from Moringa-derived biosorbent was 6.4 and 4.8, respectively. Thus, at pH > pH PZC , the heavy metal ions are attracted by the negative surface charge of biosorbents, whereas at pH < pH PZC , the adsorption of metal ions is not favored since the charge in the biosorbent surface is positive, which caused an electrostatic repulsion with metal ions [33]. This section may be divided by subheadings. It should provide a concise and precise description of the experimental results, their interpretation, as well as the experimental conclusions that can be drawn.  Figure 2a,b shows the X-ray diffractograms of L-MO and S-MO powders before and after the sorption of heavy metals. The XRD pattern of L-MO shows weak and poorly resolved peaks indicating the predominance of the amorphous nature of the biosorbent. This pattern has been observed for other biosorbents where the heavy metals could be more easily sorbed in the surface [34,35]. After L-MO interactions with Co(II) and Cd(II), the XRD patterns showed an increase in the peaks' intensity, suggesting the deposition of heavy metals and/or that those specific interactions took place. In Figure 2b, the XRD patterns of Pb(II), Cd(II), Co(II), and Ni(II) loaded S-MO exhibited a similar broad peak at around 19.6° and 22.4°. The presence of this peak in Moringa seeds was also reported in previous studies, which was associated with the presence of cellulose, proteins, or lipids [36].

FTIR Spectra
As shown in Figure 3a,b, the FTIR spectra of L-MO and S-MO before and after heavy metals sorption do not change significantly. The FTIR spectrum of Moringa leaves-derived material shows a broad band centered at 3281 cm −1 , attributed to the presence of hydroxyl (-OH) [30]. This band was also observed in the IR spectrum of Moringa seedsderived material (Figure 3b), indicating the presence of OH groups of lignocellulosic structures in the leaves and seeds of Moringa [30,36]. The band at 2917 and 2848 cm −1 (Figure 3a), as well as 2923 and 2852 cm −1 (Figure 3b), are due to the asymmetrical and symmetrical stretching vibrations of the C-H group present in lignocellulosic materials and/or proteins, respectively [36]. The presence of a vibrational band around 1538 cm −1 (Figure 3b) is attributed to the amidic group characteristic of Moringa seeds' proteins [37].  Figure 2a,b shows the X-ray diffractograms of L-MO and S-MO powders before and after the sorption of heavy metals. The XRD pattern of L-MO shows weak and poorly resolved peaks indicating the predominance of the amorphous nature of the biosorbent. This pattern has been observed for other biosorbents where the heavy metals could be more easily sorbed in the surface [34,35]. After L-MO interactions with Co(II) and Cd(II), the XRD patterns showed an increase in the peaks' intensity, suggesting the deposition of heavy metals and/or that those specific interactions took place. In Figure 2b, the XRD patterns of Pb(II), Cd(II), Co(II), and Ni(II) loaded S-MO exhibited a similar broad peak at around 19.6 • and 22.4 • . The presence of this peak in Moringa seeds was also reported in previous studies, which was associated with the presence of cellulose, proteins, or lipids [36].  Figure 2a,b shows the X-ray diffractograms of L-MO and S-MO powders before and after the sorption of heavy metals. The XRD pattern of L-MO shows weak and poorly resolved peaks indicating the predominance of the amorphous nature of the biosorbent. This pattern has been observed for other biosorbents where the heavy metals could be more easily sorbed in the surface [34,35]. After L-MO interactions with Co(II) and Cd(II), the XRD patterns showed an increase in the peaks' intensity, suggesting the deposition of heavy metals and/or that those specific interactions took place. In Figure 2b, the XRD patterns of Pb(II), Cd(II), Co(II), and Ni(II) loaded S-MO exhibited a similar broad peak at around 19.6° and 22.4°. The presence of this peak in Moringa seeds was also reported in previous studies, which was associated with the presence of cellulose, proteins, or lipids [36].

FTIR Spectra
As shown in Figure 3a,b, the FTIR spectra of L-MO and S-MO before and after heavy metals sorption do not change significantly. The FTIR spectrum of Moringa leaves-derived material shows a broad band centered at 3281 cm −1 , attributed to the presence of hydroxyl (-OH) [30]. This band was also observed in the IR spectrum of Moringa seedsderived material (Figure 3b), indicating the presence of OH groups of lignocellulosic structures in the leaves and seeds of Moringa [30,36]. The band at 2917 and 2848 cm −1 (Figure 3a), as well as 2923 and 2852 cm −1 (Figure 3b), are due to the asymmetrical and symmetrical stretching vibrations of the C-H group present in lignocellulosic materials and/or proteins, respectively [36]. The presence of a vibrational band around 1538 cm −1 (Figure 3b) is attributed to the amidic group characteristic of Moringa seeds' proteins [37].

FTIR Spectra
As shown in Figure 3a,b, the FTIR spectra of L-MO and S-MO before and after heavy metals sorption do not change significantly. The FTIR spectrum of Moringa leaves-derived material shows a broad band centered at 3281 cm −1 , attributed to the presence of hydroxyl (-OH) [30]. This band was also observed in the IR spectrum of Moringa seeds-derived material (Figure 3b), indicating the presence of OH groups of lignocellulosic structures in the leaves and seeds of Moringa [30,36]. The band at 2917 and 2848 cm −1 (Figure 3a), as well as 2923 and 2852 cm −1 (Figure 3b), are due to the asymmetrical and symmetrical stretching vibrations of the C-H group present in lignocellulosic materials and/or proteins, respectively [36]. The presence of a vibrational band around 1538 cm −1 (Figure 3b) is attributed to the amidic group characteristic of Moringa seeds' proteins [37]. The bands at 1611 cm −1 (Figure 3a) and 1649 cm −1 (Figure 3b) are associated with the vibration of the C=O bond of the fatty acid and protein structures present in the leaves and seeds from Moringa. The region between 1420 (carboxylate) and 1236 cm −1 (C-N) is attributed to carboxylic acids and amides in the Moringa leaves, respectively [38], whereas the band around 1050 cm −1 is attributed to the C-O vibrations, which are characteristic of the lignin structure of the L-MO and S-MO [38]. Overall, the presence of the difference confirms that the amino acids, alkaloids, flavonoids, and phenolics are the dominant surface functional groups in the seeds and leaves of this plant [36][37][38]. As observed in Figure 3, the band-shift after the metallic ions' adsorption was not remarkable, indicating that the complex's possible formation with the metallic ions should not have occurred. Therefore, the metallic ions could have some ion exchange with the functional groups of L-MO and S-MO biosorbents. These results were also corroborated the XRD data after the adsorption, where no new crystalline phase was observed. The bands at 1611 cm −1 (Figure 3a) and 1649 cm −1 (Figure 3b) are associated with the vibration of the C=O bond of the fatty acid and protein structures present in the leaves and seeds from Moringa. The region between 1420 (carboxylate) and 1236 cm −1 (C-N) is attributed to carboxylic acids and amides in the Moringa leaves, respectively [38], whereas the band around 1050 cm −1 is attributed to the C-O vibrations, which are characteristic of the lignin structure of the L-MO and S-MO [38]. Overall, the presence of the difference confirms that the amino acids, alkaloids, flavonoids, and phenolics are the dominant surface functional groups in the seeds and leaves of this plant [36][37][38]. As observed in Figure  3, the band-shift after the metallic ions' adsorption was not remarkable, indicating that the complex's possible formation with the metallic ions should not have occurred. Therefore, the metallic ions could have some ion exchange with the functional groups of L-MO and S-MO biosorbents. These results were also corroborated the XRD data after the adsorption, where no new crystalline phase was observed. Based on XRD and FTIR results, the mechanism of adsorption following Scheme 1 is proposed. Based on XRD and FTIR results, the mechanism of adsorption following Scheme 1 is proposed.

Effect of the Dosage
The mass effect of the leaves and seeds of Moringa oilefera-derived biosorbent on the biosorption capacity for Pb(II), Cd(II), Ni(II), and Co(II) is shown in Figure 4a-d. The removal percentage of all metallic ions increases with the augmentation of biosorbent mass from 0.1 to 0.5 g. This sentence should be changed by the following: This effect was more notable for S-MO than L-MO: for instance, Cd(II), Ni(II) and Co(II) were removed at 40%, 30% and 30%, respectively, whereas this was not significant for Pb(II) (1.59 %). The increase in the percentage removal of Cd(II) and Pb(II) with increasing L-MO mass was insignificant (1.94% and 6.59%, respectively), whereas a slight increase presented in the case of Co(II) (14.37%) and a significant increase presented for Ni(II) (31.83%). In various investigations, similar behavior was observed and attributed to the great accessibility of active adsorption sites due to the biosorbent mass increase [39,40]. Furthermore, the differences observed also depend on the chemical nature of the heavy metals involved and their affinity to the biosorbents.

Effect of the Dosage
The mass effect of the leaves and seeds of Moringa oilefera-derived biosorbent on the biosorption capacity for Pb(II), Cd(II), Ni(II), and Co(II) is shown in Figure 4a-d. The removal percentage of all metallic ions increases with the augmentation of biosorbent mass from 0.1 to 0.5 g. This sentence should be changed by the following: This effect was more notable for S-MO than L-MO: for instance, Cd(II), Ni(II) and Co(II) were removed at 40%, 30% and 30%, respectively, whereas this was not significant for Pb(II) (1.59 %). The increase in the percentage removal of Cd(II) and Pb(II) with increasing L-MO mass was insignificant (1.94% and 6.59%, respectively), whereas a slight increase presented in the case of Co(II) (14.37%) and a significant increase presented for Ni(II) (31.83%). In various investigations, similar behavior was observed and attributed to the great accessibility of active adsorption sites due to the biosorbent mass increase [39,40]. Furthermore, the differences observed also depend on the chemical nature of the heavy metals involved and their affinity to the biosorbents.

Effect of the Dosage
The mass effect of the leaves and seeds of Moringa oilefera-derived biosorbent on the biosorption capacity for Pb(II), Cd(II), Ni(II), and Co(II) is shown in Figure 4a-d. The removal percentage of all metallic ions increases with the augmentation of biosorbent mass from 0.1 to 0.5 g. This sentence should be changed by the following: This effect was more notable for S-MO than L-MO: for instance, Cd(II), Ni(II) and Co(II) were removed at 40%, 30% and 30%, respectively, whereas this was not significant for Pb(II) (1.59 %). The increase in the percentage removal of Cd(II) and Pb(II) with increasing L-MO mass was insignificant (1.94% and 6.59%, respectively), whereas a slight increase presented in the case of Co(II) (14.37%) and a significant increase presented for Ni(II) (31.83%). In various investigations, similar behavior was observed and attributed to the great accessibility of active adsorption sites due to the biosorbent mass increase [39,40]. Furthermore, the differences observed also depend on the chemical nature of the heavy metals involved and their affinity to the biosorbents.

Kinetic
Figure 5a-d show the amounts of Pb(II), Cd(II), Ni(II), and Co(II) sorbed from aqueous solutions onto L-MO and S-MO at different contact times. It is observed that, for both Moringa components (seeds and leaves), the biosorption process is mainly carried out in two stages. In the first step, approximately 90% of all metal ions' maximum quantity removal occurred within the first 60 min. This fast removal can be related to the various surface chemical functionalities acting as available active sites on the L-MO and S-MO surfaces. After 60 min, the biosorption process becomes almost invariable due to the few available sites on the biosorbent surfaces [20]. The time to attain the equilibrium by using Moringa leaves and seeds is between 120 and 240 min for all metal ions.
As seen in Figure 5a-d, the Moringa leaves-derived biosorbent showed a better removal performance than the one obtained from the seeds. The amounts of Pb(II), Cd(II), Ni(II), and Co(II) sorbed on L-MO at equilibrium reached 9.46, 9.17, 7.74, and 5.07 mg/g, respectively. In S-MO's case, the metals' sorption revealed the same order, with the capacities to be 9.34, 2.76, 1.48, and 1.34 mg/g, respectively. These results can be assigned to the interactions of these metallic ions with the functional groups from L-MO and S-MO, as shown in the FTIR spectra.
The differences in the adsorption capacities between leaves and seeds could be due to their chemical composition. It has been reported that M. oleifera seeds contain a higher amount of 4-(a-L-rhamnopyranosyloxy)-benzylglucosinolate than the leaves, and the leaves also contain three monoacetyl isomers of that glucosinolate [41]. Therefore, the functional groups of those compounds represent adsorption sites available for interaction with the metallic species. The preference in the removal efficiency exhibited by the biosorbent for Pb 2+ over Cd 2+ , Co 2+ and Ni 2+ may be attributed to Pb 2+ smaller hydrated radius (Pb 2+ = 0.401 nm; Cd 2+ = 0.426 nm; Co 2+ = 0.423 nm; Ni 2+ = 0.404 nm) as well as hydration energy (Pb 2+ = −1481 kJ/mol; Cd 2+ = −1807 kJ/mol; Co 2+ = −1996 kJ/mol; Ni 2+ = −2106 kJ/mol) [42]. Cd 2+ showed higher retention capacity in comparison with Co 2+ and Ni 2+ , which could be more associated with hydration energy and electrostatic factors than with hydrated radius and steric hindrance.
The pHpzc of the surface of the biomasses from leaves and seeds (6.4 and 4.8, respectively) can also exert a vital role in the adsorption process as well as each metal's chemical species in the solution. The predominant species that could be found in aqueous solution considering 100 mg/L of each metal at pH 6 are Pb 2+ (50%) and Pb(OH)2 (50%); Cd 2+ ; Ni 2+ ; and Co 2+ [43]. Based on the results presented in Figure 5, the metallic cationic species were efficiently adsorbed by the leaves biomass of M. oilefera, which has a pHpzc near to 7 compared to the acid surface of seeds biomass. It is important to mention that in the case of Pb, the cationic exchange and the precipitation of the lead species on the surface of the biomasses could be the mechanisms to remove this heavy metal from aqueous solutions.  It is observed that, for both Moringa components (seeds and leaves), the biosorption process is mainly carried out in two stages. In the first step, approximately 90% of all metal ions' maximum quantity removal occurred within the first 60 min. This fast removal can be related to the various surface chemical functionalities acting as available active sites on the L-MO and S-MO surfaces. After 60 min, the biosorption process becomes almost invariable due to the few available sites on the biosorbent surfaces [20]. The time to attain the equilibrium by using Moringa leaves and seeds is between 120 and 240 min for all metal ions.

Kinetic
As seen in Figure 5a-d, the Moringa leaves-derived biosorbent showed a better removal performance than the one obtained from the seeds. The amounts of Pb(II), Cd(II), Ni(II), and Co(II) sorbed on L-MO at equilibrium reached 9.46, 9.17, 7.74, and 5.07 mg/g, respectively. In S-MO's case, the metals' sorption revealed the same order, with the capacities to be 9.34, 2.76, 1.48, and 1.34 mg/g, respectively. These results can be assigned to the interactions of these metallic ions with the functional groups from L-MO and S-MO, as shown in the FTIR spectra.
The differences in the adsorption capacities between leaves and seeds could be due to their chemical composition. It has been reported that M. oleifera seeds contain a higher amount of 4-(a-L-rhamnopyranosyloxy)-benzylglucosinolate than the leaves, and the leaves also contain three monoacetyl isomers of that glucosinolate [41]. Therefore, the functional groups of those compounds represent adsorption sites available for interaction with the metallic species. The preference in the removal efficiency exhibited by the biosorbent for Pb 2+ over Cd 2+ , Co 2+ and Ni 2+ may be attributed to Pb 2+ smaller hydrated radius (Pb 2+ = 0.401 nm; Cd 2+ = 0.426 nm; Co 2+ = 0.423 nm; Ni 2+ = 0.404 nm) as well as hydration energy (Pb 2+ = −1481 kJ/mol; Cd 2+ = −1807 kJ/mol; Co 2+ = −1996 kJ/mol; Ni 2+ = −2106 kJ/mol) [42]. Cd 2+ showed higher retention capacity in comparison with Co 2+ and Ni 2+ , which could be more associated with hydration energy and electrostatic factors than with hydrated radius and steric hindrance.
The pH pzc of the surface of the biomasses from leaves and seeds (6.4 and 4.8, respectively) can also exert a vital role in the adsorption process as well as each metal's chemical species in the solution. The predominant species that could be found in aqueous solution considering 100 mg/L of each metal at pH 6 are Pb 2+ (50%) and Pb(OH) 2 (50%); Cd 2+ ; Ni 2+ ; and Co 2+ [43]. Based on the results presented in Figure 5, the metallic cationic species were efficiently adsorbed by the leaves biomass of M. oilefera, which has a pH pzc near to 7 compared to the acid surface of seeds biomass. It is important to mention that in the case of Pb, the cationic exchange and the precipitation of the lead species on the surface of the biomasses could be the mechanisms to remove this heavy metal from aqueous solutions. The values of the kinetic parameters were obtained from the slope and the intercept by plotting of ln( − ) . and . graphs. The results of the linear graphs are presented in Table 1. As observed, the experimental values of sorption capacities of Pb(II), Cd(II), Ni (II), and Co(II) sorbed at equilibrium (qe,exp) are very close to those calculated from the pseudo-second-order (qe,cal). Moreover, as seen in Table 1, the determination coefficient (R 2 ) values of the pseudo-second-order are between 0.999 and 1.0, whereas for the pseudo-first-order model, the R 2 values are between 0.249 to 0.973. Therefore, the experimental kinetics data of metal ions biosorption onto L-MO and S-MO are better expressed by the pseudo-second-order model than the pseudo-first-order model. The k2 constant depends on the plant's part (leaves or seeds) used as biosorbent and the heavy metal's chemical nature. The k2Pb(II) and k2Cd(II) are 17 and 127 times higher for L-MO than for S-MO, respectively. However, k2Ni(II) and k2Co(II) are slightly different between L-MO and S-MO (1.8% and 22 %, respectively).  The values of the kinetic parameters were obtained from the slope and the intercept by plotting of ln(q e − q t ) vs.t and t q t vs. t graphs. The results of the linear graphs are presented in Table 1. As observed, the experimental values of sorption capacities of Pb(II), Cd(II), Ni (II), and Co(II) sorbed at equilibrium (q e,exp ) are very close to those calculated from the pseudo-second-order (q e,cal ). Moreover, as seen in Table 1, the determination coefficient (R 2 ) values of the pseudo-second-order are between 0.999 and 1.0, whereas for the pseudofirst-order model, the R 2 values are between 0.249 to 0.973. Therefore, the experimental kinetics data of metal ions biosorption onto L-MO and S-MO are better expressed by the pseudo-second-order model than the pseudo-first-order model. Units of q e,exp and q e,cal (mg g −1 ); k 2 (g mg −1 min −1 ); and k 1 (min −1 ).
The k 2 constant depends on the plant's part (leaves or seeds) used as biosorbent and the heavy metal's chemical nature. The k 2Pb(II) and k 2Cd(II) are 17 and 127 times higher for L-MO than for S-MO, respectively. However, k 2Ni(II) and k 2Co(II) are slightly different between L-MO and S-MO (1.8% and 22 %, respectively). The experimental data were fitted to Langmuir and Freundlich isotherms to describe the biosorption behavior.

Isotherms
Both models well described the biosorption behavior of the heavy metals by the biosorbents according to the determination coefficients (R 2 ) presented in Table 2. Units of qe,exp and qe,cal (mg g −1 ); k2 (g mg −1 min −1 ); and k1 (min −1 ).

Isotherms
Figure 6a-d show the sorption isotherms for the two biomasses (L-MO and S-MO). The experimental data were fitted to Langmuir and Freundlich isotherms to describe the biosorption behavior.
Both models well described the biosorption behavior of the heavy metals by the biosorbents according to the determination coefficients (R 2 ) presented in Table 2.
The equilibrium constant of sorption values K L (L/mg)) obtained from the Langmuir isotherm for L-MO and S-MO were highest for Pb (15.89 × 10 −2 and 55.01 × 10 −2 ) and lowest for Co (1.23 × 10 −2 and 0.33 × 10 −2 ). According to the Qm value observed here, the biomasses adsorbed the heavy metal ions in the order: Pb(II) > Cd(II) > Ni(II) > Co(II). These results are under Zamzow et al., who investigated the adsorption onto zeolites [23].
The 1/n value was lowest for Pb by L-MO and S-MO (0.56 and 0.48), and it was highest for Co for both biosorbents (0.68 and 0.77, respectively). It is important to note that L-MO and S-MO presented a higher maximum adsorption capacity for Pb(II) than the Cd(II), Ni(II), and Co(II), respectively, in this order. In general, the biosorbent L-MO is more efficient in removing the heavy metals in solution than S-MO. According to the Q m value obtained based on the Langmuir model, the maximum adsorption capacities of each heavy metal follow the order: L-MO-Pb > L-MO-Cd > S-MO-Pb > L-MO-Co ≥ L-MO-Ni > S-MO-Co > S-MO-Cd > S-MO-Ni.
In Table 3, plenty of reported removal capacities for other biosorbents are collected for the sake of comparison with the herein studied materials. It can be seen that the maximum uptake capacities of L-MO and S-MO (Q m ) with other reported biosorbents under similar experimental conditions. It can be noted that the sorption capacities depend on the physicochemical characteristics of the biosorbents and the heavy metal (pK a , solubility, chemical speciation, among others). Additionally, the leaves-derived sample exhibits a higher efficiency than the seeds-derived sample for the removal of Pb(II), Cd(II), Ni(II), and Co(II) from aqueous solutions. The values showed in Table 3 indicate that the adsorption capacities of heavy metals on the leaves and seeds of Moringa are within the range of values reported for other biosorbents. The outperformance most worth highlighting was for a Moringa oilefera leaves-derived sample against Pb(II), with a value that, to the best of our knowledge, was the highest among the plant biomass-derived biosorbents.

Conclusions
In this study, the remediation efficiency of two biosorbents derived from Moringa oilefera leaves or seeds was explored against four heavy metals. It was found that the sorption capacity of Pb(II), Cd(II), Co(II), and Ni(II) from aqueous solution by Moringa leaves-derived sample (L-MO) is higher than that for the seeds-derived biosorbent (S-MO). Both L-MO and S-MO biosorbents presented the following removal efficiency order: Pb(II) >Cd(II) >Co(II) >Ni(II). In all cases, the metal ions' sorption capacity was dependent on the contact time, the dosage of biosorbent, and the initial metal concentration. The effect of contact time on the sorption capacity indicates the very fast metal-binding for the L-MO and S-MO biosorbent, since the equilibrium time was obtained within the first 60 min. Based on the values of correlation coefficients (R 2 values close or equal to 1), sorption kinetics was well described by the pseudo-second-order kinetic model. The Langmuir and Freundlich models were also fitted well on the sorption isotherms. The R 2 values for both isotherms are close to 1, which suggests that the sorption process of the metal ions on L-MO and S-MO satisfied the Langmuir and Freundlich assumptions. The maximum sorption capacities were found in the following order: L-MO-Pb > L-MO-Cd > S-MO-Pb > L-MO-Co ≥ L-MO-Ni > S-MO-Co > S-MO-Cd > S-MO-Ni. Finally, but no less importantly, it is worthy to point out that the Moringa oilefera leaves-derived sample presented the highest remediation performance against Pb(II) compared to various other plant biomass-derived biosorbents reported in the literature.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.