Removal of Ni(II) and Cu(II) in Aqueous Solutions Using Treated Water Hyacinth (Eichhornia crassipes) as Bioadsorbent

Abstract

Phytoremediation consists of taking advantage of the capacity of certain plants to absorb, accumulate, or metabolize contaminants. In this study, Eichornia crassipes (water lily) treated with water (WLW) and NaOH (WLN) was investigated as an adsorbent for removal of Ni(II) and Cu(II) present in aqueous solution, focusing on determining the most efficient conditions (adsorbent concentration, contact time, pretreatment, temperature). The results showed that equilibrium adsorption was favorable and carried out by a multilayer physical process with both bioadsorbents. The maximum adsorption at 30 °C in WLW and WLN was 349 and 293.8 mg/g of Ni(II), respectively, and 294.1 and 276.3 mg/g of Cu(II), respectively. The thermodynamic analysis indicated that the removal in both metals was spontaneous and exothermic. The Avrami model was the most adequate in the kinetic study of Ni(II) and Cu(II) removal in both treatments, which revealed that the adsorption process was carried out by several mechanisms. In the characterization of the adsorbents, it was determined that the functional groups of WL as well as the attractive forces on the surface of the materials participated in the metal removal process.

1. Introduction

At present, due to great growth in the use of various chemical products in industry, their waste has increased considerably, and consequently, the concentrations of contaminants in the water, such as dyes, organic waste, heavy metals, and others, have increased [1,2]. Heavy metals are a major source of concern among the various contaminants found in industrial effluents. In underdeveloped countries, the treatment of effluents containing heavy metals before discharge into the sewer is often inadequate [3,4,5]. Heavy metals are non-biodegradable, exhibit high mobility in aquatic systems, and are soluble in water, which allows them to be adsorbed in cells and accumulate in the tissues of organisms, causing serious damage to the composition of blood and the respiratory and nervous systems [2,4,5,6,7]. Among the heavy metals that cause the most concern are Ni, Cu, and Co, which are used in electroplating, paint, fertilizers, tanning, metallurgy, etc. [1,2,4]. In addition, Ni, Cu, and Co are essential metals for animals, plants, and humans since they stimulate the production of red blood cells [1,2,8]. At concentrations higher than 15 ppm, they are usually toxic and cause diseases such as kidney damage, liver, cyanosis, and lung cancer, among others [2,4,7].

Due to this hypertoxicity of heavy metals, removal methods have been developed such as those that use physical barriers (reverse osmosis membranes), electrochemical processes (electrolytic extraction and electrodialysis), chemical precipitation, ion exchange, and other techniques [2,3,7,9], which are effective but turn out to have high costs for infrastructure and control and require too much energy [1,2,3,5,6,9]. Adsorption is a cost-effective and efficient method used to remove pollutants such as dyes (methylene blue, methyl orange, phenol red) and heavy metals such as Cr, Ni, Pb, Cd, Cu, Zn, and Fe, among others [2,3,4,5,10,11,12]. Different types of adsorbents, including clays, mesoporous materials, zeolites, and reducible oxides, can be used with this method [10,11,12]. In recent years, there has been growing interest in finding low-cost and abundant adsorbents with comparable adsorption capacities to conventional options. Agroindustrial residues, including orange peel, grapefruit, rice, wheat, egg, and walnut, among others, have emerged as promising alternatives [13,14,15,16,17]. Among the alternatives, phytoremediation employs plants that are able to adsorb and accumulate compounds present in the aquifers where these plant bodies are found [18,19,20,21]. Plants such as Nymphaea alba (white nymph), Nymphaea nucifera (sacred lotus), Pistia stratrioes (water lettuce), and Oesdogonium [18,19,20,21,22,23,24], among others, have presented very good results for the adsorption of contaminants from water. Among these plants, the water lily (Eichhornia crassipes) is considered a critical problem as it represents one of the main aquatic weeds. The water lily causes great economic losses due to its accelerated and uncontrollable growth, causing stagnation of water, and its long roots suspended in the water manifest a comfortable environment for aerobic microorganisms, which encourages a decrease in oxygen levels present in the water, affecting all aquatic species and causing various health problems, just to mention a few problems [21,22,23,24,25]. However, E. crassipes has shown high adsorption efficiency towards heavy metals such as Cr, Pb, Cu, etc., removing 90% of the metals present in water and it can be reused several times in the adsorption process [21,23,24]. In the literature, there are several reports of pretreating water lily before its use in the adsorption of contaminants, which results in better outcomes compared to commercial adsorbents [26,27,28]. There are studies on the parameters (contact time, pH, initial concentration, among others) that reduce the adsorption of heavy metals with water lily. In addition, the removal process is carried out by one or two active sites on the surface of the absorbent [18,19,20,21,22,23,24,25,26,27,28].

Based on this premise, the objective of this work was to investigate the potential of water lily (WL) extracted from the Yuriria lagoon, Guanajuato, Mexico, as a bioadsorbent for removing Ni(II) and Cu(II) from aqueous solutions. The study considered the impact of surface modification by treating water lily with water and NaOH and explored the effects of some parameters on heavy metal adsorption, including temperature, contact time, initial metal concentration, and adsorbent concentration. Through these experiments, the adsorption mechanisms of both metals were determined, and the biomaterial was characterized during the metal adsorption process.

2. Materials and Methods

2.1. Reagents

All reagents used were analytical grade. The water used for all solutions in the experiments was deionized. Hydrochloric acid (HCl) from J. T. Baker (CAS: 7647-01-0, 36.5–38%), sodium hydroxide (NaOH) from J. T. Baker (CAS: 1310-73-2, 98%), cupric sulfate pentahydrate (CuSO₄ 5H₂O) from Merck (CAS: 7758-99-8, 98%), and nickel nitrate hexahydrate (Ni(NO₃)₂ 6H₂O) from Sigma-Aldrich (CAS:13478-00-7, 99.99%) were the reagents used for the different experiments.

2.2. Pretreatment of Water Lily (E. crassipes)

The water lily (WL) was washed with running water at room temperature, dried in a forced convection oven (Shel Lab CE5F) at 80 °C for 24 h, and crushed using an industrial blender (Tapisa T12L) to obtain a fine powder (100 mesh) for subsequent pretreatment. The water lily was subjected to treatment with deionized water (WLW) a a ratio of 30 g/L (P/V) at 75 °C and constant stirring for 1 h, followed by subsequent filtration with a vacuum pump (Thomas 1CZC8). This process was repeated until a crystalline filtrate was obtained. WL was treated with 0.5 M sodium hydroxide (NaOH) (WLN) at a ratio of 30 g/L (P/V), with an initial pH of 13.65, and it was stirred for 2 h at 60 °C. Subsequently, it was filtered, and the solid was mixed with 1 M HCl solution (pH ~1) while stirring for 1 h at room temperature. Subsequently, the solid was filtered and washed 10 times with the amount of water used for the HCl solution. Finally, it was dried at 85 °C overnight in a forced convection oven [29].

2.3. Adsorption Isotherms

For the equilibrium study of the adsorption of the metals, an adsorbent/volume mass ratio of 0.5 g/L solution was used and the concentration was varied from 0 to 5000 ppm of Ni(II) and Cu(II). Next, the mixtures were shaken in a shaker (ZHWY-200D) at 200 rpm and at 30, 45, and 60 °C until equilibrium was reached (~12 h of contact time). Then, the samples were centrifuged (Generic 6-TRPR) at 6000 rpm for 10 min. The heavy metal concentration was determined using an atomic absorption spectrometer (Analyst-100, PerkinElmer). The amount of dye removed by adsorbent, q_e, was obtained with the following expression [30,31]:

$${\mathbf{q}}_{\mathbf{e}}\mathbf{=}\frac{\left({\mathbf{C}}_{\mathbf{0}}-{\mathbf{C}}_{\mathbf{e}}\right)\mathbf{V}}{\mathbf{m}}$$

(1)

where C₀ and C_e represent the initial concentration in equilibrium (mg/L), V is the volume of solution (L), and m is the mass of WL (g). The equilibrium models proposed to fit the experimental data for metals are in the following table.

The removal percentage, %R, was calculated as follows [30,31]:

$$\mathbf{\%}\mathbf{R}\mathbf{=}\frac{\left({\mathbf{C}}_{\mathbf{0}}-{\mathbf{C}}_{\mathbf{e}}\right)}{{\mathbf{C}}_{\mathbf{0}}}\ast \mathbf{100}$$

(2)

The different models of the isotherms are shown in

Table 1. Non-linear equilibrium (isotherm) adsorption models [30,31].

Model	Equation
Langmuir	${\mathrm{q}}_{\mathrm{e}}=\frac{{\mathrm{q}}_{\mathrm{m}}{\mathrm{K}}_{\mathrm{L}}{\mathrm{C}}_{\mathrm{e}}}{1+{\mathrm{K}}_{\mathrm{L}}{\mathrm{C}}_{\mathrm{e}}}$	qe is equilibrium adsorption capacity (mg/g). Ce is the equilibrium concentration of the metal in the liquid (mg/L). V (L) is the volume of the dye solution. m (g), is the mass of the adsorbent. qm, is the maximum adsorbed capacity (mg/g). KL is the Langmuir constant related to adsorption energy (L/mg). KF is Freundlich constant related to binding energy (mg/g)(L/mg)1/n. n is the constant that is related to the linearity of the adsorption (dimensionless). B is the Temkin constant related to heat of adsorption (kJ/mol). A is a constant in the equilibrium bond. KR and aR are RP constants (L/g) and (L/mg)β, respectively. β is the RP exponent (dimensionless). KS is the Sips constant (L/mg). nS is the Sips exponent (dimensionless). kDR is the activity coefficient related to the adsorption energy (mol/J)2. ε is the Polanyi potential. E is the energy of adsorption $\mathsf{\epsilon }=\mathrm{R}\mathrm{T}\ln \left(1+\frac{1}{{\mathrm{C}}_{\mathrm{e}}}\right)$ $\mathrm{E}=\frac{1}{\sqrt{2{\mathrm{k}}_{\mathrm{D}\mathrm{R}}}}$
Freundlich	${\mathrm{q}}_{\mathrm{e}}={\mathrm{K}}_{\mathrm{F}}{\mathrm{C}}_{\mathrm{e}}^{\frac{1}{\mathrm{n}}}$
Temkin	${\mathrm{q}}_{\mathrm{e}}=\mathrm{B}\ln \left(\mathrm{A}{\mathrm{C}}_{\mathrm{e}}\right)$
Redlich–Peterson (RP)	${\mathrm{q}}_{\mathrm{e}}=\frac{{\mathrm{K}}_{\mathrm{R}}{\mathrm{C}}_{\mathrm{e}}}{1+{\mathrm{a}}_{\mathrm{R}}{\mathrm{C}}_{\mathrm{e}}^{\mathsf{\beta }}}$
Dubinin–Radushkevich (DR)	${\mathrm{q}}_{\mathrm{e}}={\mathrm{q}}_{\mathrm{m}}\mathrm{e}\mathrm{x}\mathrm{p}\left(-{\mathrm{k}}_{\mathrm{D}\mathrm{R}}{\mathsf{\epsilon }}^2\right)$
Sips	${\mathrm{q}}_{\mathrm{e}}=\frac{{\mathrm{q}}_{\mathrm{m}}{\left({\mathrm{K}}_{\mathrm{s}}{\mathrm{C}}_{\mathrm{e}}\right)}^{{\mathrm{n}}_{\mathrm{S}}}}{1+{\left({\mathrm{K}}_{\mathrm{s}}{\mathrm{C}}_{\mathrm{e}}\right)}^{{\mathrm{n}}_{\mathrm{S}}}}$

. The regression coefficient was calculated to evaluate the fit of each nonlinear model and the separation factor, R_L, which allowed prediction of the affinity between the adsorbent and adsorbate using the following equation [30,31]:

$${\mathbf{R}}_{\mathbf{L}}\mathbf{=}\frac{\mathbf{1}}{\mathbf{1}\mathbf{+}{{\mathbf{C}}_{\mathbf{0}}\mathbf{K}}_{\mathbf{L}}}$$

(3)

where K_L is the constant of the Langmuir model and C₀ is the initial concentration of Ni(II) and Cu(II). To understand the thermodynamics of the adsorption process, thermodynamic parameters such as the apparent Gibbs free energy were determined.

$$\mathbf{∆}\mathbf{G}\mathbf{=}\mathbf{-}\mathbf{R}\mathbf{T}\mathbf{l}\mathbf{n}\left(\mathbf{55.5}{\mathbf{K}}_{\mathbf{L}}\right)$$

(4)

where K_L is the constant of the Langmuir model) (L/mol), R the ideal gas constant, and T is the absolute temperature (K).

$$\mathbf{∆}\mathbf{G}\mathbf{=}\mathbf{-}\mathbf{R}\mathbf{T}\mathbf{l}\mathbf{n}\left(\mathbf{55.5}{\mathbf{K}}_{\mathbf{L}}\right)$$

(5)

The values of ΔH and ΔS were determined by the slope and sorted to the origin of the ΔG chart as a function of 1/T.

2.4. Batch Dye Removal (Adsorption Kinetics)

The adsorption kinetics were performed using 2500 ppm solutions for both metals and the adsorbent concentration varied from 0 to 5 g/L. The mixtures were shaken in a shaker at 200 rpm and at 30, 45, and 60 °C, with a contact time of 9 h. An aliquot was taken every 1.5 h, which was centrifuged at 6000 rpm for 10 min. The heavy metal concentration was determined using an atomic absorption spectrometer (Analyst-100, PerkinElmer). Experimental data were adjusted with the adsorption kinetics models found in

Table 2. Adsorption kinetic models used in the analysis of experimental data [31].

Model	Equation
Pseudo first order (PPO)	$\mathrm{q}={\mathrm{q}}_{\mathrm{m}\mathrm{a}\mathrm{x}}\left[1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{\mathrm{k}}_1\mathrm{t}\right)\right]$	qt is the adsorption capacity (mg/g). C0 is the initial concentration of the dye in the liquid (mg/L). V (L) is the volume of the dye solution. m (g) is the mass of the adsorbent. qmax is the maximum adsorbed capacity (mg/g). k1 (1/h) is the speed constant of the PPO model. k2 (g s/mg) is the speed constant of the PSP model. kext is the Avrami constant (h−1). nA reflects the changes of the mechanism during the adsorption process. kInt (mg/g h) is the speed constant of the ID model. kExt (1/h) is the speed constant of the ED model.
Pseudosecond order (PSO)	$\mathrm{q}=\frac{\mathrm{t}}{\frac{1}{{\mathrm{k}}_2\ast {\mathrm{q}}_{\mathrm{m}\mathrm{a}\mathrm{x}}^2}+\frac{\mathrm{t}}{{\mathrm{q}}_{\mathrm{m}\mathrm{a}\mathrm{x}}}}$
Avrami	$\mathrm{q}={\mathrm{q}}_{\mathrm{m}\mathrm{a}\mathrm{x}}\left[1-\mathrm{e}\mathrm{x}\mathrm{p}{\left(-{\mathrm{k}}_{\mathrm{A}}\mathrm{t}\right)}^{{\mathrm{n}}_{\mathrm{A}}}\right]$
Intraparticle diffusion (ID)	$\mathrm{q}={\mathrm{k}}_{\mathrm{I}\mathrm{D}}{\mathrm{t}}^{0.5}$
External diffusion (ED)	$\mathrm{q}=\frac{{\mathrm{C}}_{0\mathrm{*}}\mathrm{V}}{\mathrm{m}}\left[1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{\mathrm{k}}_{\mathrm{e}\mathrm{x}\mathrm{t}}\ast \mathrm{t}\right)\right]$

. In addition to using the coefficient of determination to compare the efficiency of the different kinetic and equilibrium models, the standard deviation, Δq%, was calculated [30,31]:

$$∆\mathrm{q}\mathrm{\%}=\sqrt{\frac{{\left(\frac{{\mathrm{q}}_{\mathrm{e}\mathrm{x}\mathrm{p}}-{\mathrm{q}}_{\mathrm{c}\mathrm{a}\mathrm{l}}}{{\mathrm{q}}_{\mathrm{e}\mathrm{x}\mathrm{p}}}\right)}^2}{\mathrm{N}-1}}\ast 100$$

(6)

where N is the number of data, and q_exp and q_cal (mg/g) are the experimental and calculated values of the removed dyes, respectively. The kinetics models proposed to fit the experimental data for the metals are in the following table.

2.5. Characterization of Bioadorbent

Attenuated total reflectance–Fourier-transform spectroscopy (ATR–FTIR) analyses were carried out over the wavenumber range of 4000–400 cm⁻¹ using a Thermo Scientific Nicolet iS10 analyzer before and after the adsorption of DNS. A total of 32 scans were obtained with a resolution of 4 cm⁻¹. X-ray diffraction patterns (XRD) were obtained using a diffractometer (Ultima IV Rigaku). To determine the isoelectric point (pH_zpc), 0.05 g of adsorbent in 50 mL of water was stirred at 200 rpm for 24 h, and the pH was determined using a potentiometer (Science Med SM-25CW). An additional 0.05 g of adsorbent was added every 24 h until the pH did not change. Scanning electron microscopy and X-ray energy dispersion spectroscopy (SEM–EDS–EDX) were performed using a JOEL spectrometer (6510 pus, Peabody, Boston, MA, USA).

3. Results and Discussions

3.1. Effect of the Initial Concentration of Metals on Adsorption with WL

The analysis of the parameters that influenced the adsorption of metal ions, such as the contact time, the concentration of the bioadsorbent and the initial concentration of the metals, allowed us to find the optimal parameters to carry out the adsorption process with WL [29]. In terms of the effect of the initial concentration of Ni(II) and Cu(II), the results allows us to identify when the mass transfer resistance between the liquid and solid phases had been overcome due to the modification of the driving force [32]. In Figure 1, it can be seen that regardless of the treatment carried out on the surface of WL, there was a rapid increase in the adsorption capacities of Ni(II) and Cu(II) until reaching equilibrium at an initial concentration of 4000 ppm. This indicated that increasing the concentration of metal ions led to an increase in bioadsorption, which was directly related to the increased difference in concentrations in the solid and liquid phases, which allowed us to avoid limitations due to mass transfer of the process [33,34,35,36,37,38].

In terms of the percentage of removal of metal ions (Figure S1), it decreased with the increase in concentration, with removal decreasing from 53% to 41% and from 39% to 24% for Ni and Cu, respectively, with WLW. In the case of WLN, the percentage of removal decreased from 58% to 37% for Ni and from 37% to 30% for Cu. This was due to the effect that occurred when the concentration of sites for the adsorption of metal ions remained constant, resulting in greater ease of disposition at these sites at low metal concentrations but preventing further metals from being removed from solution when the available sites became saturated [39,40,41,42,43,44,45].

3.2. Effect of the Amount of Adsorbent on the Adsorption of Ni(II) and Cu(II)

The quantity of the adsorbent is a relevant parameter in the biosorption process since it allows determining the viability of the adsorbent to renew the active sites for the capture of metal ions [34,44,46]. The amount of bioadsorbent is directly related to the concentration of available active sites. In addition, an increase in the amount of adsorbent causes an increase in the area of the bioadsorbent [37,46]. Figure 2 shows the effect of the bioadsorbent concentration on the adsorption process of Ni(II) and Cu(II). It was noted that the adsorption capacity of metal ions in WLW decreased significantly, with maximum losses in the adsorption capacity of 64.7% and 61.9% for Ni and Cu, respectively. In the case of WLN, there were maximum losses in the adsorption capacity of 47.23% and 67.14% for Ni and Cu, respectively. This was due to the fact that the adsorption capacity was affected by the amount of bioadsorbent present in the adsorption process, which agreed with the results reported for different adsorbents such as biocarbon from wheat and rice [47], activated carbon [36], almond shell [39], water lily treated with citric acid [48], modified sediments [46], and residual sludge [49].

In terms of the percentage of removal of metal ions, it increased with the increase in the concentration of the bioadsorbent, as shown in Figure S2, regardless of the treatment performed on the surface of WL. It was observed that the percentage of adsorption increased rapidly due to the increase in the surface area of the bioadsorbent and the concentration of available sites. This may also have been due to the electrostatic forces present on the surface of WL [34,36,38,41,44,46,47,48,49,50]. Taking the parameter of bioadsorbent quantity into account is important because an insufficient amount of bioadsorbent may not effectively remove metals from the solution. However, an excess amount of bioadsorbent would cause unnecessary waste of resources and increase the costs associated with the adsorption process [39].

3.3. Ni(II) and Cu(II) Adsorption Isotherms Using WLW and WLN

The equilibrium adsorption data provided information on the relationship between the metal concentration in solution and the surfaces of WLW and WLN. This, coupled with the adjustment made to the experimental data using the different isotherm models (Table 1), helped us to understand the mechanism of the metal removal process [29]. Figure 3 shows the adjustments of the adsorption isotherms at different temperatures for Ni (Figure 3a) and Cu (Figure 3b) using WLW as a bioadsorbent. It was observed that the adsorption capacities of both metals decreased with the temperature increase, which indicated that the removal of metals was favored at low temperatures. The parameters of the different adsorption models were obtained using the Sigmaplot^® v12 program, as shown in

Table 3. Equilibrium parameters of Ni and Cu adsorption data using WLW.

Models	Ni			Cu
T, °C	30	45	60	30	45	60
Langmuir
qmax, mg/g	567.4	476.8	386.1	363.64	361.12	716.78
KL, L/mg	0.0001	3.5 × 10−5	1 × 10−6	0.0012	0.0005	1.5 × 10−4
RL	0.67–0.91	0.88–0.97	0.99–1.00	0.14–0.45	0.29–0.67	0.57–0.89
R2	0.9302	0.9229	0.9036	0.9776	0.9437	0.8671
∆q, %	34.090	40.672	43.484	8.590	18.789	34.015
Freundlich
KF,	0.3836	0.1155	0.0773	6.6101	11.105	3.7164
n	1.1550	1.0042	0.9783	2.0940	2.7391	2.1260
R2	0.9229	0.9210	0.9037	0.9710	0.9749	0.9726
∆q, %	19.214	19.443	18.259	10.687	7.0881	7.0716
Temkin
A	3.9 × 10−16	3.5 × 10−14	1.0 × 10−14	5.1 × 10−14	2.6 × 10−14	3.3 × 10−15
B	32.1691	27.7897	24.2122	30.9096	23.2901	17.9915
R2	0.5789	0.5289	0.5033	0.7953	0.8777	0.8157
∆q, %	12.243	14.807	15.1834	5.1981	2.5145	5.4477
Sips
qmax, mg/g	430.79	355.22	301.65	344.39	247.47	206.61
Ks	0.0007	0.0008	0.0006	0.0012	0.0013	0.0009
ns	2.4118	3.5521	4.2608	1.4119	1.1630	1.3604
R2	0.9567	0.9730	0.9785	0.9946	0.9916	0.9905
∆q, %	8.5114	5.0642	3.7657	6.5693	6.8936	7.5288
RP
aR	0.0001	0.0002	0.0003	0.0008	0.0012	0.0006
KR	0.1743	0.1226	0.0983	0.3288	0.3101	0.1548
β	1.0000	0.8806	0.7436	1.0000	1.0000	1.0000
R2	0.9302	0.9222	0.9035	0.9902	0.9910	0.9876
∆q, %	17.866	13.434	15.341	4.7529	2.4373	4.9916
DR
qmax, mg/g	428.98	400.55	354.91	279.25	200.87	166.39
kDR	0.3343	0.2918	0.2155	0.0566	0.0350	0.0286
R2	0.8788	0.9288	0.9423	0.8886	0.8709	0.8845
E, kJ/mol	1.2230	1.3091	1.5232	2.9722	3.7529	4.1812
∆q, %	8.3567	9.5593	9.9136	2.3351	1.9125	1.4784

, taking as criteria the deterministic coefficient (R²) and the normalized standard deviation (∆q%) to establish the best adjustment to the experimental data.

Based on these established criteria, the Sips model presented the best fit for both metals, which indicated that the equilibrium adsorption process of the heavy metals was carried out on a surface abundant with different adsorption energies. This allowed having more than one layer to capture the Ni and Cu ions. In addition, the process was physical (n_S > 1), which indicated that there was heterogeneity in the surface of WLW [5,29,51]. The Freundlich (n > 1) and DR (E < 8 kJ/mol) models confirmed that a physical adsorption process of metals took place on the surface, as indicated by the parameters related to the type of adsorption process (physical or chemical) [51,52,53]. In addition, the energy necessary to remove the metal ions from solution (E) increased with the increase in temperature due to the nature of the removal process of both metals, thus favoring adsorption at low temperatures, as can be seen in Table 3. On the other hand, both metals having R_L values between 0 and 1 revealed that equilibrium adsorption was favorable at all temperatures [51,52,53,54,55]. The adsorption capacities of Ni(II) in WLN were 348.92, 315.01, and 276.25 mg/g at 30, 45, and 60 °C, respectively. In comparison, the adsorption capacities of water lily treated with water were between 27.88 and 44.3 mg/g. Notably, these values were lower and the models that described the adsorption process were Freundlich [41] and Langmuir [54], respectively.

In other studies that have used different adsorbents, the adsorption capacity has been found to be between 0.256 and 190.38 mg/g using CoFeO₂/SO₂ nanoparticles [54], magnetized graphene [56], magnetite–bentonite nanocomposites [57], sapropel humic acids [58], calcium carbonate in bacterial magnetosomes [36], hydroxyapatite [59], lemon peel [60], bamboo modified with mercaptoacetic acid and carbon disulfide [61], wheat and rice biochar [46], charred bones [62], and charcoal [63]. ©. These results indicated that WLW had greater efficiency for the adsorption of this metallic ion in solution, and there was a maximum removal percentage of 51.31%.

In terms of Cu(II) adsorption in WLW, adsorption capacities of 293.82, 209.46, and 171.88 mg/g were obtained at 30, 45, and 60 °C, respectively, with a maximum removal percentage of 51.8%. The adsorption capacities of various adsorbents, including ZnCl₂-activated carbon [64], almond shell [39], CoFeO₂/SO₂ nanoparticles [55], bamboo modified with mercaptoacetic acid and carbon disulfide [61], magnetite-bentonite nanocomposites [15], lime shell [65], peanut shell [40], charcoal [60,61], hydroxyapatite [59], charred bones [62], sapropel humic acids [58], calcium alginate beads immobilized on rice bran [66], residual sludge [46], bentonite [43], and bentonite pretreated with CaCl₂ [52], ranged from 2.44 to 128.21 mg/g. These adsorbents used the Langmuir model to describe the adsorption of Cu(II) on their surfaces. Regarding the use of other isotherm models to analyze the adsorption of Cu(II) on the surfaces of adsorbents, the Freundlich model for the adsorption of Cu on orange peel and banana peel [33] showed an adsorption capacity of 309 mg/g. The Sips model was used to analyze the adsorption of Cu on DCPD [35], resulting in an adsorption capacity of 309 mg/g, while the Temkin model was used to analyze the adsorption of Cu on chitosan [34], resulting in an adsorption capacity of 166.6 mg/g.

The adsorption capacity of water lily treated with water as an adsorbent has been reported as 0.65 mg/g [29,31], 2.04 mg/g [43], and 22.7 mg/g [16]. Carbon obtained from water lily has shown an adsorption capacity of 19.62 mg/g [15], 32 mg/g [66], and 48.2 mg/g [50] for Cu(II). The Freundlich and Langmuir models are reported to provide the best fit for the experimental data of Cu(II) adsorption. Generally, the Langmuir and Freundlich isotherms are considered to be the best models for describing the adsorption process of Ni(II) and Cu(II). The Sips model, which is the combination of these two isotherms, also allows the adjustment of the experimental data of the adsorption of these two metals in equilibrium, as obtained in this work.

In terms of the removal of Ni and Cu using WLN as a bioadsorbent, Figure 4 shows the experimental data at different temperatures with the adjustments made with the isotherm models. It was possible to observe that as the temperature increased, the adsorption capacities of both metals decreased, which revealed that the elimination of both metals was favored in the same way as using WLW; that is, there was better removal of these metals at low temperatures [5,29,51].

Table 4. Equilibrium parameters of Ni and Cu adsorption data using WLN.

Models	Ni			Cu
T, °C	30	45	60	30	45	60
Langmuir
qmax, mg/g	312.61	403.46	225.13	325.79	300.19	287.93
KL, L/mg	0.0009	0.00035	0.0001	0.0009	0.0006	0.0004
RL	0.18–0.52	0.37–0.74	0.67–0.91	0.19–0.53	0.26–0.64	0.33–0.72
R2	0.7828	0.8535	0.7347	0.8128	0.9550	0.9621
∆q, %	5.2122	17.753	1.5812	8.5156	11.657	15.309
Freundlich
KF,	0.1420	0.0467	0.0615	0.1314	0.1419	0.0824
n	1.0581	0.9439	0.9940	1.0604	1.1016	1.0554
R2	0.9393	0.9371	0.9429	0.9689	0.9493	0.9584
∆q, %	16.943	16.639	14.711	15.567	14.023	12.577
Temkin
A	5.2 × 10−15	1.2 × 10−14	1.5 × 10−13	9.3 × 10−16	1.7 × 10−14	6.5 × 10−14
B	24.574	20.809	18.843	23.137	19.918	16.582
R2	0.5512	0.5101	0.5341	0.5831	0.5735	0.5627
∆q, %	14.314	16.673	15.783	15.136	13.809	15.255
Sips
qmax, mg/g	309.87	273.03	253.64	386.17	265.19	222.21
Ks	0.0007	0.0006	0.0005	0.0005	0.0006	0.0005
ns	3.0935	3.4525	3.0194	1.8674	2.5436	2.6258
R2	0.9924	0.9912	0.9840	0.9830	0.9871	0.9951
∆q, %	4.8578	4.8737	6.4329	14.182	7.2945	6.6089
RP
aR	0.0002	0.0001	0.0001	0.0002	0.0015	0.0001
KR	0.1089	0.0844	0.0754	0.0958	0.0840	0.0630
β	0.8636	0.9054	0.9309	0.9171	0.9208	0.9216
R2	0.9439	0.9358	0.9428	0.9712	0.9545	0.9617
∆q, %	7.9651	52.825	12.612	7.9567	19.085	10.931
DR
qmax, mg/g	328.94	304.68	271.51	314.37	261.34	221.19
kDR	0.3515	0.3019	0.2659	0.3150	0.3060	0.2846
R2	0.9682	0.9709	0.9496	0.9089	0.9453	0.9677
E, kJ/mol	1.1926	1.2869	1.3713	1.2599	1.2783	1.3255
∆q, %	7.1690	9.0071	8.9545	7.2009	6.7331	6.2455

presents the isotherm model parameters for both metals. Based on the criteria of R² and ∆q%, the Sips model showed the best fit. This suggested that the equilibrium adsorption of heavy metals occurred on a heterogeneous surface with varying adsorption energies, which allowed for the formation of more than one layer for the capture of Ni and Cu ions. Furthermore, the process was physical (n_S > 1), indicating the presence of different active sites on the surface of WLN [51,53]. This was confirmed by the Freundlich (n > 1) and DR (E < 8 kJ/mol) models, in which the parameters related to the type of adsorption process (physical or chemical) taking place on the surface indicated that there was a physisorption of the metals. This behavior was similar to the adsorption with WLN with respect to the energy needed to remove metals from the aqueous solution towards the surface of WL [51,52,53,67]. On the other hand, the fact that both metals had R_L values between 0 and 1 revealed that equilibrium adsorption was favorable at all temperatures [5,51,53].

The adsorption capacities of Ni(II) in WLN at 30, 45, and 60 °C are shown in Table 4. The values obtained were 276.25, 243.32, and 217.16 mg/g, respectively. The maximum removal percentage achieved was 39.4%. These results indicated that treating WL with NaOH caused a loss of removal efficiency of around 21.67% for Ni(II) adsorption. Studies conducted on various adsorbents treated with NaOH have indicated that the Langmuir, Freundlich, and Sips models offer better descriptions of the Ni(II) equilibrium adsorption process. In these studies, the adsorption capacities for Ni(II) were found to be 0.2, 0.123, and 41.75 mg/g using pineapple shell [64], peanut shell [44], and palm biochar [40], respectively. It was determined that the Langmuir model best fits the experimental data for the adsorption capacity of WL treated with HCl [32], methanol and acetonitrile [54], and citric acid [48]. The adsorption capacities of Ni(II) for each treatment were 0.29, 26.5, 53, and 77.98 mg/g, respectively. This indicated that WLN had greater efficiency in the adsorption of Ni(II) in solution than materials reported in the literature using NaOH in the pretreatment of different agro-industrial residues or water lily treated with organic compounds and acids.

On the other hand, the adsorption capacities of Cu(II) in WLN were 201.95, 158.52, and 107.57 mg/g at 30, 45, and 60 °C, respectively, which represented efficiency losses of 31.27%, 24.32%, and 37.41%, respectively. The maximum removal percentage obtained was 49.68%, which meant that despite losing efficiency in Cu(II) adsorption capacity, the removal percentage was not significantly affected by the treatment in WL. Palm biochar [40] and rice husk [68] exhibited adsorption capacities of 0.2 and 12.66 mg/g, respectively, when treated with NaOH. Furthermore, modifying the surface of orange peel with mercaptoacetic acid [69] resulted in an adsorption capacity of 70.67 mg/g. For water lily treated with acetic acid [48] and nitric-perchloric acid [42], the models that described the equilibrium adsorption process were Langmuir and Freundlich, respectively, and the adsorption capacities were 59.64 and 96.67 mg/g, respectively. These results indicated that despite the loss of efficiency in the adsorption of Cu(II), there was still very good removal taking into account that there was a high concentration (5000 ppm) of the heavy metal.

3.4. Study of the Thermodynamics of Ni and Cu Removal Process in Pretreated WL

The nature of the equilibrium removal of Ni and Cu was studied by determining the thermodynamic parameters, as shown in

Table 5. Thermodynamic parameters of Ni and Cu removal using WLW and WLN.

Ni
WLW				WLN
T, °C	−∆G, kJ/mol	−∆H, kJ/mol	−∆S, kJ/mol K	−∆G, kJ/mol	−∆H, kJ/mol	−∆S, kJ/mol K
30	26.195	193.13	0.551	8.212	61.323	0.158
45	18.617			3.352
60	9.659			1.003
Cu
30	26.854	58.021	0.102	13.575	22.744	0.031
45	25.866			12.993
60	23.751			12.673

. It was observed that independent of the pretreatment carried out with WL, the adsorption process of Ni and Cu was spontaneous (ΔG < 0). Regarding the Gibbs free energy, the energy increased with increasing temperature, and this behavior was directly related to the energy needed to remove metal ions from the solution [31,34,48,50,56,67,70]. In addition, the process was exothermic (ΔH < 0), which confirmed what was observed in the adsorption isotherms, that is, the adsorption capacities decreased with increasing temperature in both WLW and WLN. This same behavior was reported in various papers [31,34,39,48,50,70]; however, some reports determined that the process was exothermic [56,67]. A random change in the fluid–solid interface (ΔS < 0) was also observed in the metal removal process, causing a delay in the interaction between Cu and Ni ions with the surfaces of WLW and WLN. This caused a decrease in the adsorption of the metals [39,56,65], although when using magnetized graphene [34], CaMg phosphate [70], and water lily carbon [15], the adsorption process increased due to the increase in randomness in the solid–fluid interface.

3.5. Effect of Contact Time on Metal Adsorption

For the metal ion adsorption process to be considered profitable, a very important parameter must be analyzed, which is the contact time between the solution and the bioadsorbent’s surface [63]. In Figure 5 and Figure 6, it was possible to observe the adsorption process for Ni(II) and Cu(II), respectively, in which the adsorption was carried out in three stages: the first being a rapid increase of the adsorption capacity, followed by a slow stage to reach equilibrium, and stage 3 in which the adsorption of metals was almost constant [34,35,36,44,67,70,71]. This behavior was due to the great availability of sites at the beginning of the process, but since the adsorption was gradual over time, the sites became depleted, which caused a slowdown in the adsorption of the heavy metals [32,36,37,41,46,48,63,65]. The analysis of this parameter in the adsorption of Ni and Cu had the same behavior reported with different adsorbents, such as chitosan [34], peanut shell [44], DCPD [35], rice hull [68], CaMg phosphates [70], charcoal [63], lime peel [65], peanut peel [37], banana peel [71], and water lily treated with water [32,41,43], HCl [32], citric acid [48], and charred [36,67]. In addition, the adsorption capacity decreased as the amount of bioadsorbent increased, as can be seen in Figure 2. Figures S3 and S4 show the removal percentages of Ni(II) and Cu(II) at different temperatures, where it was possible to appreciate that the removal of metal ions increased by increasing the contact time between the solution and the bioadsorbent’s surface until reaching equilibrium, which occurred around 5 h after the process started. This behavior agreed with that reported in the literature [37,38,39,40,41,43,46,49,54].

3.6. Kinetic Mechanism of Ni(II) and Cu(II) Adsorption

The kinetic study of Ni(II) and Cu(II) adsorption at different temperatures using WLW and WLN enabled us to identify the process that controlled the removal of these metals from the solution. To achieve this, we analyzed the experimental data presented in Figure 5 and Figure 6 by applying the kinetic models outlined in Table 2. To better understand the adsorption of Ni(II) and Cu(II) in WLW,

Table 6. Kinetic parameters obtained from the fit of the experimental data of Ni adsorption in WLW.

Cads, g/L	1			2			3			4			5
Model	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C
PFO
qm, mg/g	262.41	408.96	482.09	209.30	202.78	200.49	206.25	171.19	260.55	156.28	166.43	120.25	118.88	90.000	70.753
ki, L/mg	0.7104	0.2620	0.1441	0.7095	0.3419	0.2945	0.3710	0.2380	0.1055	0.4287	0.1658	0.1909	0.4017	0.2270	0.2514
R2	0.9966	0.9879	0.9880	0.9906	0.9951	0.9939	0.9941	0.9935	0.9857	0.9978	0.9893	0.9888	0.9921	0.9139	0.8608
∆q%	6.1249	22.268	35.269	1.0645	7.6147	10.3706	5.9299	15.103	10.3706	3.6861	27.644	21.665	4.0899	8.1801	2.1210
PSO
qm, mg/g	323.87	636.45	829.46	259.68	295.17	302.13	292.02	269.41	464.83	213.42	279.91	79.753	165.26	100.00	69.753
ki, L/mg	0.0023	0.0003	0.0001	0.0029	0.0009	0.0007	0.0010	0.0006	0.0001	0.0017	0.0004	0.0089	0.0020	0.0027	0.0056
R2	0.9918	0.9849	0.9872	0.9806	0.9920	0.9913	0.9926	0.9917	0.9854	0.9971	0.9884	0.8538	0.9917	0.8527	0.7491
∆q%	2.0023	57.363	68.888	18.251	29.685	25.868	25.462	47.191	36.326	21.613	74.330	0.6198	21.613	13.625	1.5140
Elovich
α	0.0138	0.0072	0.0080	0.0168	0.0143	0.0146	0.0145	0.0182	0.0183	0.0197	0.0217	0.0286	0.0252	0.0000	0.0285
β	508.16	237.93	177.63	380.83	150.19	130.11	169.49	94.372	74.861	152.39	69.022	56.461	105.32	3470.4	34.475
R2	0.9859	0.9914	0.9831	0.9719	0.9953	0.9957	0.9939	0.9924	0.9719	0.9965	0.9832	0.9837	0.9939	1.0000	0.9268
∆q%	40.824	40.824	40.623	40.823	40.822	39.756	40.820	40.818	37.967	40.613	40.817	39.075	40.821	40.814	40.734
ID
kint, h−1	119.64	128.68	104.03	95.291	72.888	67.105	76.885	51.098	44.515	61.721	39.592	31.418	45.788	26.665	22.598
R2	0.9360	0.9508	0.9252	0.9191	0.9736	0.9664	0.9814	0.9577	0.9120	0.9891	0.9349	0.9466	0.9822	0.8343	0.7909
∆q%	2.0719	7.8031	7.1285	5.8919	1.8243	1.932	1.8676	0.0645	1.1473	1.5511	0.9273	0.8316	1.5511	5.1438	7.2283
ED
kext, L/mg	0.0300	0.0334	0.0268	0.0505	0.0381	0.0349	0.0641	0.0406	0.0352	0.0693	0.0423	0.0327	0.0634	0.0357	0.0301
R2	0.5887	0.9238	0.9707	0.6095	0.8871	0.9130	0.8903	0.9458	0.9796	0.8611	0.9704	0.9581	0.8718	0.9753	0.9617
∆q%	2.7627	15.236	1.5218	2.0945	8.0160	7.8564	7.4510	6.2885	7.4443	7.0181	1.0095	5.4875	7.0180	1.1888	1.1359
Avrami
qm, mg/g	320.56	274.20	257.85	205.34	175.78	162.04	185.13	124.97	118.54	156.28	98.531	77.699	108.75	73.995	65.657
kA, L/mg	0.6903	0.3065	0.2091	0.6665	0.3727	0.3356	0.4071	0.2828	0.1710	0.4451	0.2235	0.2594	0.4308	0.2222	0.1916
nA	1.0291	0.8549	0.6894	1.0646	0.9173	0.8776	0.9113	0.8415	0.6168	0.9631	0.7417	0.7359	0.9324	0.9937	0.9243
R2	0.9966	0.9879	0.9880	0.9906	0.9951	0.9939	0.9941	0.9935	0.9857	0.9978	0.9893	0.9888	0.9921	0.9164	0.9024
∆q%	1.6968	1.4774	6.7289	0.2720	1.1650	1.0479	1.1422	0.0028	0.5523	0.8607	0.2896	0.4476	0.2627	0.5345	0.9721

and

Table 7. Kinetic parameters obtained from the fit of the experimental data of Cu adsorption in WLW.

Cads, g/L	1			2			3			4			5
Model	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C
PFO
qm, mg/g	354.37	211.12	175.87	185.59	186.15	163.97	153.25	122.11	108.36	105.53	98.802	96.753	88.839	121.43	116.44
ki, L/mg	0.2933	0.5997	0.6142	0.5478	0.4647	0.5060	0.4651	0.6270	0.6297	0.4645	0.4344	0.3219	0.3229	0.1885	0.1417
R2	0.9854	0.9984	0.9995	0.9978	0.9917	0.9978	0.9976	0.9958	0.9950	0.9943	0.9973	0.9945	0.9816	0.9902	0.9861
∆q%	9.0739	1.2717	1.4930	2.4629	2.6962	2.3663	3.7680	1.2944	1.9897	3.9201	3.9823	7.7887	8.1726	22.707	34.633
PSO
qm, mg/g	291.77	201.77	223.06	242.07	246.77	216.25	207.47	154.88	137.48	142.61	136.44	139.68	127.91	200.66	65.765
ki, L/mg	0.0029	0.0071	0.0027	0.0021	0.0017	0.0021	0.0019	0.0040	0.0045	0.0028	0.0027	0.0018	0.0019	0.0006	0.0093
R2	0.8966	0.9456	0.9960	0.9928	0.9952	0.9970	0.9934	0.9905	0.9865	0.9901	0.9948	0.9950	0.9816	0.9888	0.8393
∆q%	0.2593	0.5926	12.848	11.202	15.636	16.137	19.644	12.598	13.496	19.642	21.052	29.357	29.172	64.159	1.7939
Elovich
α	0.0093	0.0162	0.0192	0.0171	0.0175	0.0193	0.0197	0.0275	0.0307	0.0289	0.0305	0.0322	0.0360	0.0276	0.0333
β	258.42	323.30	267.11	231.91	210.73	191.49	155.24	185.11	162.66	108.46	96.601	72.981	69.109	55.148	42.151
R2	0.9795	0.9967	0.9924	0.9900	0.9952	0.9964	0.9918	0.9860	0.9803	0.9889	0.9943	0.9920	0.9745	0.9896	0.9799
∆q%	40.814	39.648	38.756	40.065	39.648	38,764	39.965	38.957	37.520	39.056	40.005	37.956	38.856	38.856	37.895
ID
kint, h−1	118.74	92.634	77.573	79.358	75.749	68.529	62.239	54.116	48.034	42.843	39.587	33.885	31.168	31.433	24.805
R2	0.9849	0.9713	0.9613	0.9668	0.9925	0.9824	0.9763	0.9516	0.9423	0.9742	0.9829	0.9875	0.9770	0.9393	0.9203
∆q%	0.1297	4.4202	4.9002	4.5155	2.4536	3.3909	3.6077	4.8989	5.6642	3.6667	9.1786	0.8785	1.2845	0.5440	1.4496
ED
kext, L/mg	0.0305	0.0229	0.0190	0.0413	0.0394	0.0352	0.0499	0.0422	0.0369	0.0452	0.0415	0.0351	0.0410	0.0419	0.0324
R2	0.9049	0.6743	0.6514	0.7352	0.8045	0.7691	0.8118	0.6742	0.6534	0.8027	0.8262	0.8991	0.8884	0.9636	0.9711
∆q%	6.2943	10.396	11.053	10.393	10.396	9.3493	9.4748	10.638	11.516	9.5593	3.1461	6.9079	7.3073	5.7186	4.9920
Avrami
qm, mg/g	290.78	205.21	167.99	175.59	173.87	154.32	175.87	120.99	108.36	95.583	92.867	82.632	75.061	77.985	63.632
kA, L/mg	0.3374	0.6252	0.6168	0.5475	0.4995	0.5198	0.4766	0.6115	0.6130	0.4720	0.4560	0.3640	0.3301	0.2448	0.2028
nA	0.8692	0.9592	0.9957	1.0007	0.9302	0.9735	0.9760	1.0255	1.0273	0.9842	0.9528	0.8842	0.9784	0.7701	0.6984
R2	0.9854	0.9984	0.9995	0.9978	0.9917	0.9978	0.9980	0.9958	0.9950	0.9943	0.9973	0.9945	0.9816	0.9902	0.9861
∆q%	0.1198	0.0933	0.4030	0.1305	0.2707	0.1756	1.2246	0.9081	0.2229	0.2974	1.2908	0.6936	0.5736	0.0234	0.4116

provide kinetic parameters that reveal multiple models that show a good fit. By using the criteria R² and ∆q% to select the best fit for the experimental data, the Avrami model was determined to best describe the adsorption of Ni(II) and Cu(II) regardless of temperature. This suggested that the adsorption process underwent changes in the adsorption mechanism (n_A < 1) [71,72,73,74]. This allowed us to infer that in addition to the physical adsorption of metals on the WLW surface, other mechanisms may have intervened (ion exchange, intervention of functional groups, etc.), in addition to not having external and internal mass transfer limitations [59]. The maximum capacities for Ni(II) obtained at 30, 45, and 60 °C were 308.73, 264.62, and 259.11 mg/g, respectively, and for Cu(II), they were 289.93, 204.74, and 169.67 mg/g at 30, 45, and 60 °C, respectively. These results confirmed the nature of the equilibrium adsorption process for Ni(II) and Cu(II) as well as the thermodynamic analysis of the process. The fact that the maximum adsorption capacity was observed at 30 °C suggested the elimination of an endothermic process.

For the adsorption of Ni, there are reports in the literature about the use of different adsorbents for the removal of Ni(II), such as chitosan [34], magnetized graphene [54], sapropel humic acids [58], modified sediment [46], charred bones [62], and palm activated carbon with magnetite particles [38], with adsorption capacities ranging from 3.66 to 50.68 mg/g. In the reports where they used WLW, adsorption capacities of 0.2899 [40] and 39.6 mg/g [71] were obtained. In terms of what has been reported in the literature for Cu(II) adsorption using adsorbents such as Zn–Cl2-activated carbon [64], chitosan [34], DCDP [34], sapropel humic acids [58], almond [39], peanut shell [37], charred bones, and bentonite [52], the reported adsorption capacities were between 6.07 and 195.88 mg/g. When comparing the adsorption capacities of WLW and carbonized water lily as adsorbents, the latter had a slightly higher adsorption capacity of 24.3 mg/g [73], compared to WLW with an adsorption capacity of 23 mg/g [67]. However, the results obtained in this study showed a significant increase in the removal efficiency of Ni(II) and Cu(II) compared to other adsorbents and WLW. Specifically, there were increases in Ni(II) removal efficiency of 83.6% and 87.2% compared to other adsorbents and WLW, respectively. Similarly, there were increases in Cu(II) removal efficiency of 32.43% and 91.44% compared to other adsorbents and WLW, respectively. These results demonstrated that the adsorption capacity of water lily as an adsorbent could be significantly improved with appropriate modification and activation.

The kinetic study of Ni(II) and Cu(II) adsorption in WLN at different temperatures yielded the kinetic parameters presented in

Table 8. Kinetic parameters obtained from the adjustment of the experimental data of Ni adsorption in WLN.

Cads, g/L	1			2			3			4			5
Model	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C
PFO
qm, mg/g	313.51	291.36	247.76	309.69	274.19	224.05	294.92	240.83	296.76	265.11	177.94	316.49	269.34	226.20	555.78
ki, L/mg	0.3249	0.2882	0.3032	0.2299	0.2360	0.2495	0.1946	0.1856	0.1102	0.1767	0.2125	0.0754	0.1306	0.1046	0.0285
R2	0.9954	0.9930	0.9864	0.9923	0.9917	0.9899	0.9919	0.9900	0.9884	0.9859	0.9882	0.9839	0.9778	0.9816	0.9670
∆q%	8.5709	10.546	73.499	15.802	15.576	37.298	20.985	22.118	48.265	25.165	19.070	43.182	39.599	52.238	84.841
PSO
qm, mg/g	461.68	440.39	374.52	483.88	430.19	349.02	476.85	389.21	529.05	442.82	288.31	110.76	477.05	100.76	113.89
ki, L/mg	0.0005	0.0005	0.0006	0.0003	0.0004	0.0005	0.0003	0.0003	0.0001	0.0002	0.0005	0.0051	0.0002	0.0060	0.0001
R2	0.9926	0.9908	0.9824	0.9915	0.9902	0.9878	0.9912	0.9898	0.9880	0.9846	0.9862	0.8092	0.9768	0.8092	0.9671
∆q%	31.916	36.822	17.398	47.651	47.665	13.484	59.114	60.899	118.01	69.401	56.220	137.99	101.62	8.8557	149.13
Elovich
α	0.0093	0.0101	0.0114	0.0106	0.0116	0.0138	0.0118	0.0149	0.0155	0.0132	0.0178	0.0001	0.0001	0.0199	0.0002
β	221.86	185.61	160.04	172.75	152.76	128.98	142.74	113.15	88.383	115.52	88.454	303.40	25301	62.406	392.65
R2	0.9959	0.9964	0.9919	0.9832	0.9861	0.9864	0.9811	0.9791	0.9742	0.9792	0.9891	1.0000	1.0000	0.9855	1.0000
∆q%	40.824	40.822	41.956	40.532	40.005	40.005	40.321	39.756	38.945	39.065	39.745	38.756	37.856	40.003	37.856
ID
kint, h−1	109.93	96.459	84.059	90.737	81.485	68.691	78.039	61.845	52.449	65.822	49.642	41.136	54.060	38.296	30.133
R2	0.9726	0.9654	0.9530	0.9646	0.9594	0.9573	0.9535	0.9528	0.9151	0.9337	0.9423	0.8906	0.9052	0.8953	0.8398
∆q%	1.6002	4.4667	1.5287	0.3486	0.2328	0.7084	0.7614	1.2314	2.2568	0.6924	6.909	4.7692	1.2847	2.2314	3.3882
ED
kext, L/mg	0.0280	0.0244	0.0211	0.0490	0.0434	0.0359	0.0659	0.0504	0.0422	0.0760	0.0545	0.0445	0.0788	0.0527	0.0407
R2	0.8908	0.9094	0.8901	0.9526	0.9462	0.9335	0.9720	0.9680	0.9827	0.9737	0.9576	0.9828	0.9745	0.9783	0.9670
∆q%	7.9313	7.2242	8.121	0.6741	15.921	7.0326	4.6553	27.413	3.9667	11.246	6.141	3.1265	4.1976	2.9851	3.1548
Avrami
qm, mg/g	264.87	234.87	201.45	224.85	198.54	167.05	197.92	154.62	137.32	164.99	122.99	106.89	138.65	101.03	81.642
kA, L/mg	0.3602	0.3360	0.3400	0.2730	0.2779	0.2941	0.2461	0.2505	0.1748	0.2299	0.2645	0.1434	0.1914	0.1732	0.0851
nA	0.8994	0.8576	0.8919	0.8423	0.8490	0.8485	0.7908	0.7408	0.6304	0.7688	0.8034	0.5259	0.6824	0.6038	0.3347
R2	0.9954	0.9930	0.9864	0.9923	0.9917	0.9899	0.9919	0.9900	0.9884	0.9859	0.9882	0.9839	0.9778	0.9816	0.9672
∆q%	0.9074	0.5861	0.6134	0.2878	0.0148	0.4098	0.6387	0.4133	0.4008	0.2045	0.5736	0.7468	0.5756	0.7389	0.5844

and

Table 9. Kinetic parameters obtained from the fit of the experimental data of Cu adsorption in WLN.

Cads, g/L	1			2			3			4			5
Model	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C	30 °C	45 °C	60 °C
PFO
qm, mg/g	217.58	155.03	108.79	176.55	118.23	83.275	200.69	97.971	78.736	626.17	80.853	83.381	180.71	130.54	42.211
ki, L/mg	0.4025	0.5306	0.8190	0.3510	0.5018	0.6499	0.1772	0.3820	0.3288	0.0315	0.2620	0.1611	0.0064	0.0751	0.2494
R2	0.9952	0.9937	0.9992	0.9940	0.9937	0.9961	0.9940	0.9927	0.9935	0.9766	0.9914	0.9859	0.9738	0.9838	0.9927
∆q%	4.7928	2.9417	1.6670	7.2338	3.0934	17.804	22.175	5.3493	7.4558	21.154	13.103	27.347	74.449	77.716	14.896
PSO
qm, mg/g	299.45	202.87	125.20	253.99	156.61	103.66	323.58	138.11	75.767	110.76	65.761	55.765	65.765	280.82	63.262
ki, L/mg	0.0011	0.0024	0.0077	0.0011	0.0029	0.0066	0.0004	0.0022	0.0085	0.0041	0.0106	0.0094	0.0075	0.0469	0.0027
R2	0.9954	0.9891	0.9981	0.9915	0.9901	0.9987	0.9943	0.9912	0.9467	0.8181	0.8901	0.8906	0.7842	0.9628	0.9920
∆q%	21.957	16.447	20.344	28.313	17.358	36.836	60.752	24.266	5.6352	3.8144	3.0371	4.7684	1.1265	21.418	42.685
Elovich
α	0.0144	0.0206	0.0392	0.0169	0.0266	0.’0432	0.0185	0.0302	0.0399	0.0204	0.0370	0.0474	0.0001	0.0469	0.0782
β	204.35	190.15	302.87	137.67	135.25	149.23	92.015	81.791	61.293	60.798	47.561	35.793	488.78	28.082	24.636
R2	0.9936	0.9862	0.9933	0.9910	0.9881	0.9977	0.9812	0.9939	0.9901	0.9437	0.9930	0.9648	1.0000	0.9628	0.9849
∆q%	40.8341	37.659	41.657	40.068	38.750	37.439	39.867	39.645	36.758	38.756	38.654	37.989	38.756	40.862	37.932
ID
kint, h−1	83.921	65.681	49.113	64.237	49.249	37.289	50.042	36.970	27.857	37.459	25.443	19.502	23.218	16.933	12.648
R2	0.9912	0.9665	0.9188	0.9777	0.9722	0.9648	0.9587	0.9793	0.9889	0.8724	0.9575	0.9471	0.8567	0.8988	0.9700
∆q%	2.2727	4.5960	6.1627	2.0074	3.9947	15.696	2.3576	3.9947	2.2891	8.5783	0.7403	1.7682	4.5475	3.1576	2.7241
ED
kext, L/mg	0.0209	0.0160	0.0117	0.0331	0.0246	0.0182	0.0397	0.0283	0.0209	0.0402	0.0259	0.0197	0.0305	0.0217	0.0157
R2	0.8291	0.7084	0.4737	0.8749	0.7476	0.6300	0.9713	0.8521	0.8817	0.9798	0.9228	0.9609	0.9730	0.9802	0.9248
∆q%	8.6191	10.883	12.156	8.2301	10.184	10.494	3.5536	10.184	7.3878	31.192	7.1985	4.7725	1.8699	3.5441	6.6975
Avrami
qm, mg/g	201.56	142.78	103.99	146.32	111.11	82.891	88.048	134.76	68.915	62.634	98.731	50.879	64.827	45.879	29.698
kA, L/mg	0.4298	0.5399	0.8034	0.3722	0.5093	0.6799	02408	0.4155	0.3636	0.0867	0.3041	0.2182	0.0373	0.1386	0.2882
nA	0.9364	0.9828	1.0194	0.9429	0.9854	0.9556	0.7360	0.9193	0.9044	0.3633	0.8616	0.7282	0.1711	0.5422	0.8654
R2	0.9952	0.9937	0.9992	0.9940	0.9937	0.9961	0.9940	0.9927	0.9935	0.9799	0.9914	0.9859	0.9738	0.9838	0.9927
∆q%	1.4341	0.5166	0.2078	0.9951	0.4542	0.2797	1.4787	0.4542	1.4337	1.0339	0.9515	0.7737	0.5282	0.8371	1.6216

. The results showed that the Avrami model provided the best fit to the experimental data for both metals. This suggested that during the adsorption process, more than one mechanism occurred for removing metal ions from the solution (n_A < 1). Thus, in addition to the electrostatic forces that attracted Ni(II) and Cu(II) to the surface, the functional groups present in WLN or other mechanisms may have also participated in the adsorption process. It was possible to observe that the adsorption capacity increased with increasing temperature, confirming that the elimination of heavy metals had an endothermic nature. In the adsorption of Ni(II) at 30, 45, and 60 °C, maximum adsorption capacities of 259.11, 231.55, and 198.4 mg/g were obtained, respectively. Compared to the literature on adsorbents treated with NaOH, our study showed higher adsorption capacities for Ni(II) and Cu(II). For instance, pineapple peel [75], peanut [44], and palm carbon [40] had reported adsorption capacities of 9.28, 0.123, and 0.2 mg/g, respectively. Similarly, the use of mercaptoacetic acid and carbon disulfide to modify bamboo as an adsorbent resulted in an adsorption capacity of 8.64 mg/g [61]. Acetic acid-treated WL achieved an adsorption capacity of 16.43 mg/g [48]. All of these reports mentioned that the best fitting model was PSO. In the adsorption of Cu(II) using WLN, we obtained adsorption capacities of 259.11, 231.55, and 194.72 mg/g at 30, 45, and 60 °C, respectively. When comparing our results with the literature, there was a maximum adsorption capacity between 1.03 and 18.58 mg/g with different adsorbents and treatments, including orange peel modified with mercaptoacetic acid [33], rice pretreated with NaOH [68], bamboo modified with mercaptoacetic acid and carbon disulfide [61], palm biocarbon modified with NaOH [40], and water lily modified with acetic acid [48]. Therefore, there was an increase in the adsorption capacity of our adsorbent with respect to these materials, with increases in the adsorption of Ni(II) and Cu(II) using WLN of 93.66% and 90.46%, respectively.

3.7. Characterization of WLW and WLN

The following figure shows the ATR–FTIR spectra of WLW and WLN. It was possible to observe the different characteristic WL bands, which were related to the functional groups of the compounds present on the surface of WL. In the ATR–FTIR spectrum of WLW shown in Figure 7a, the band at 3368 cm⁻¹ corresponded to the OH- group and links associated with lignin, cellulose, and hemicellulose. The signal at 2927 cm⁻¹ was attributed to the symmetric and asymmetric vibrations of the C-H bond of the methyl and methylene groups [29,45,48,50,71,73]. The shoulder at 2851 cm⁻¹ and the band at 1550 cm⁻¹ were due to the stretching of the CH groups of lignin [29,45,48,50,71,73]. The band at 1649 cm⁻¹ corresponded to the stretching vibrations of the C=O carboxylic bond [29,45,48,50,71,73], while a shoulder at 1726 cm⁻¹ was attributed to the C=O stretching of the carboxylic group [73]. The peak at 1417 cm⁻¹ was assigned to the vibration of aliphatic compounds (-CH₂ and -CH₃) and methoxy group (O-CH₃) [29,45,48,50,71,73]. Additionally, the peaks at 1318, 1252, and 1087 cm⁻¹ corresponded to the symmetry vibration of COO- bond stretching, the presence of the C-H bond of the aromatic group, and the vibration stretching of the C-OH bond of the alcoholic, lignin, and carboxylic acid groups, respectively [29,45,48,50,71,73]. Moreover, the bands at 1153 and 1042 cm⁻¹ were attributed to the stretching vibration in the C-O bond of the lignin structure and the CO-R vibration of the alcohol groups, respectively [29,45,48,50,71,73]. The bands at 898 and 624 cm⁻¹ corresponded to the stretching of the C=O group and the β-glucosidic bonds of cellulose, respectively [29,45,48,50,71,73]. In Figure 7b, the ATR–FTIR spectrum of WLN exhibited changes in all bands compared to WLW. These changes indicated that there were alterations in the functional groups present on the surface of WLN, which in turn affected the adsorption capacity of Ni(II) and Cu(II) in solution. The bands related to the functional groups on WLN’s surface were located at 3373 cm⁻¹ and corresponded to the OH- groups and lignin, cellulose, and hemicellulose bonds. Meanwhile, the peak at 2922 cm⁻¹ was assigned to symmetrical and asymmetric vibrations of the C-H bond of the methylene and methyl groups [29,41,45,48,50,71,73]. The shoulder at 2855 cm⁻¹ and the band at 1544 cm⁻¹ were attributed to the stretching of the CH groups of lignin [29,41,45,48,50,71,73]. The band at 1654 cm⁻¹ corresponded to the stretching vibrations of the C=O carboxylic bond [29,45,48,50,71,73]. A shoulder was observed at 1731 cm⁻¹, corresponding to the C=O stretching of the carboxylic group [73]. The peak at 1434 cm⁻¹ was assigned to the vibration of aliphatic compounds (-CH₂ and -CH₃) and methoxy group (O-CH₃) [29,41,45,48,50,71,73]. The bands at 1323, 1257, and 1092 cm⁻¹ were assigned to the symmetry vibration of COO- bond stretching, the presence of the C-H bond of the aromatic group, and the stretching vibration of the C-OH bond of the alcoholic, lignin, and carboxylic acid groups, respectively [29,41,45,48,50,71,73]. Bands were also observed at 1147 and 1048 cm⁻¹, corresponding to the stretching vibration in the C-O bond of the lignin structure and the CO-R vibration of the alcohol groups, respectively [29,41,45,48,50,71,73]. Finally, the bands at 904 and 596 cm⁻¹ were attributed to the stretching of the C=O group and the β-glucosidic bonds of cellulose, respectively [29,41,45,48,50,71,73]. In the ATR–FTIR spectra of WLW and WLN after metal adsorption, the functional groups involved in the process were identified. The peaks observed corresponded directly to lignin, carboxylic acids, cellulose, hemicellulose, and alcohols [30,31]. The intensity of these bands was significantly reduced after Ni(II) and Cu(II) removal due to the interaction between the surface functional groups of the biomaterials and the metal ions in solution. It was observed that there was a greater interaction between Ni(II) ions and the functional groups than between Cu(II) ions and the functional groups. This was indicated by a greater decrease in the intensity of the bands in the Ni(II) spectrum compared to that in the Cu(II) spectrum [30,31].

The micrographs in Figure 8 depict the surface characteristics of WLW and WLN. In WLW, the surface appeared to be irregular with pores and holes resulting from water treatment, while in WLN, the surface was fibrous with fractures and pores due to NaOH treatment. These unique surface characteristics indicated the high adsorption capacity for both materials [29,31,48,50]. The micrographs also illustrated the adsorption of Cu(II) and Ni(II) in WLW and WLN. Although no significant changes in the surface of the adsorbents were observed, it could not be concluded that both metal ions were effectively adsorbed since clusters on the surface were visible in the micrographs [19,23,27,29,31].

The elemental analysis (

Table 10. Elemental analysis of WLW and WLN before and after adsorption of Ni(II) and Cu(II).

Bioadsorbents	wt,%
	C	O	Al	Si	Ca	Metal
WLW	59.64	30.05	0.2	0.72	2.08	-------
Ni	52.78	43.23	0.18	0.37	0.98	2.69
Cu	57.51	30.24	1.08	0.43	0.0	1.85
WLN	54.96	41.27	0.87	2.21	0.36	-------
Ni	53.40	40.69	0.24	0.63	0.24	2.54
Cu	55.27	41.78	0.76	0.74	0.19	1.41

) performed on the adsorbents showed changes in the C/O ratio (1.99 and 1.33 for WLW and WLN, respectively) and Ca/Si ratio (2.89 and 1.69 for WLW and WLN, respectively), revealing that there were changes in the functional groups of cellulose, hemicellulose, lignin, carboxylic acids, etc. present on the surfaces of WLW and WLN due to the treatment carried out in WL [29,31,50]. The decrease in the C/O ratio in WLN with respect to that in WLW was 33.17%, while the Ca/Si ratio decreased by 94.15% in WLN compared to that in WLW.

The analysis carried out on the adsorbents after the adsorption of Ni(II) and Cu(II) also showed changes in the C/O and Ca/Si ratios, indicating which functional groups had the greatest participation in the metal adsorption process. Table 8 shows that for the case of Ni(II) and Cu(II) adsorption using WLW, the C/O ratios were 1.22 and 1.68, respectively, which represented decreases of 38.7% and 15.6%, respectively. The Ca/Si ratios for Ni(II) and Cu(II) were 2.65 and 0, which implied decreases of 7.8% and 100%, respectfully. These results indicated that the functional groups containing Si and Ca had a greater role in the adsorption of Cu(II) compared to Ni(II), while the functional groups containing C and O had a greater role in the adsorption of Ni(II) compared to Cu(II) when using WLW. Therefore, the adsorption capacity of Ni(II) in WLW was greater than that of Cu(II) [29,31,50]. In the case of the adsorption of metals using WLN, the Ca/Si ratios were 0.38 and 0.26, respectively, which implied that there were decreases of 77.46% and 84.62% for Ni(II) and Cu(II), respectively. The C/O ratios were 1.31 and 1.32, which meant that there were decreases of 1.5% and 0.07% for Ni(II) and Cu(II), respectively. These results indicated that the adsorption of metal ions in solution was carried out mainly with the participation of the functional groups that contained Si and Ca, since the participation of the functional groups in which C and O were present was practically insignificant.

The solutions of Ni(II) and Cu(II) were determined to have initial pH values of 6.61 and 4.59, respectively, which indicated that the solutions contained positively charged metal species according to the diagrams obtained by the MEDUSA^® program shown in Figure S5. The isoelectric point value of WLW and WLN were 6.25 and 8.8. These results indicated that the adsorption of Ni(II) in WLW was favored by the electrostatic attraction of the species of this metal to the adsorbent surface because pH_sol > pH_zpc, revealing that the WLW surface was negatively charged [27,29]. The adsorption of Cu(II) took place on a positively charged surface (pH_sol < pH_zpc), which caused an electrostatic repulsion between the metal species and the WLW surface, causing a decrease in its ability to adsorb Cu(II) compared to Ni(II) for removal [27,29]. In the case of WLN, pH_sol < pH_zpc was true for both metals, which revealed that the surface of the adsorbent was positively charged [27,29]. This implied that there was an electrostatic repulsion of the Cu(II) and Ni(II) species in solution with the WLN surface, causing a lower adsorption capacity compared to WLW. Based on the results obtained from the characterization of the adsorbents, it could be stated that the removal process of Ni(II) and Cu(II) in solution was carried out through mechanisms involving the participation of functional groups present in WL (lignin, cellulose, etc.) and mainly physical attraction. In the particular case of Ni(II) adsorption in WLW, adsorption by electrostatic forces of attraction was also present, which is why greater removal of this metal with this material was observed compared to WLN.

4. Conclusions

In this work, WL treated with water and NaOH was studied to obtain a low-cost and efficient adsorbent to purify water by eliminating Ni(II) and Cu(II). The results of the analysis of different parameters showed the conditions that give the highest metal adsorption capacity. The adsorption capacity increased as the initial concentration and contact time increased. The optimal conditions for Ni(II) were a contact time of 4 h and an initial concentration of 3500 ppm, while for Cu(II), the optimal conditions were an initial concentration of 4000 ppm with a contact time of 5 h. Additionally, the percentage of removal increased as the adsorbent concentration increased, with an optimum dosage of 3 g/L. The adsorption efficiency decreased when WL was treated with NaOH compared to treatment with water for each metal ion, with decreases in the adsorption capacity of 20.1% for Ni and 31.3% for Cu. The ATR–FTIR analysis revealed the characteristic functional groups of WL, including lignin, cellulose, hemicellulose, and carboxylic acid, among others. The treatment caused changes in these functional groups, with the most significant changes observed in WLN. The changes were confirmed by elemental analysis, which showed that the content of C and O in WLW was decreased compared to that in WLN. This confirmed the participation of functional groups in the adsorption of Ni(II) and Cu(II). Therefore, the adsorption process was carried out by the mechanisms of physical adsorption and participation of functional groups, mainly considering that the nature of the adsorption process was endothermic. These results allow us to suggest that WLW is a promising adsorbent for the removal of these heavy metals.