Aluminum-Modified Graphene Oxide Composite Adsorbent for Humic Acid Removal from Water

Abstract

Among the pollutants that affect water quality, being also a problem in water treatment facilities, is natural organic matter (NOM), the largest percentage of which is humic acid (HA). In the present work, a new aluminum-modified graphene oxide adsorbent (henceforth abbreviated GO-Al) was produced for the elimination of HA. The factors affecting the adsorption process, such as pH, adsorbent dosage, initial HA concentration and contact time were examined. It was revealed that at pH 2.0 ± 0.1, by applying 1.0 g L⁻¹ GO-Al to 5 mg L⁻¹ HA, 91% was removed after 24 h, but equilibrium was almost reached after 30 min (82% removal). Comparable results with GO exhibited that the modification with AlCl₃⋅6H₂O enhanced the removal. The relative results associated slightly more with the pseudo-second-order kinetic model (PSO), and the Langmuir isotherm model, indicating that the process was closer to chemisorption. The maximum adsorption capacity (Q_m) conferring to the Langmuir model was considered to be 5.91 mg g⁻¹. Thermodynamics revealed that the process occurred spontaneously, while a adsorption–regeneration study up to 10 cycles confirmed the effectiveness of GO-Al material.

1. Introduction

Humic acid (HA) is the main constituent of high-molecular-weight natural organic matter (NOM) [1], which is one of the most significant contaminants affecting water quality [2]. HA is a complex compound, without a specific shape or type, formed by the decomposition of organic matter, such as microorganisms, plants, or animals, through biological and chemical processes and is usually very stable, persistent, and non-biodegradable [3]. The concentration of HA in natural surface waters (lakes or rivers) ranges from 0.1 to 20 mg L⁻¹ and is higher than that in seawater, as freshwater has a higher part of organic matter. Additionally, NOM concentration in urban wastewater ranges between 118 and 228 mg L⁻¹ and around 42% of this represents HA [4].

There are a plethora of traditionally used approaches for the removal of HA from wastewater and among these are both chemical processes such as flocculation, filtration, or oxidation, as well as biological processes [5,6,7,8]. However, compared to other techniques, adsorption is generally used to remove a plethora of chemical pollutants from water, as it has been found to be greater in terms of both initial cost as well as flexibility and simplicity in design. In addition, ease of operation and high efficiency are further advantages [9].

Several advanced adsorbents are used for the removal of HA from aquatic environments [3,10,11]. Among them, carbon-based materials, such as activated carbon [12,13] and graphene-based materials [14,15] are highly effective. Furthermore, graphene oxide (GO) is of interest as an adsorbent for water treatment due to its remarkable properties and flexibility [16]. Its large surface area and the presence of various oxygen-containing functional groups provide abundant active sites for chemical reactions [17] so that it has the potential for the effective removal of a wide range of contaminants, including organic compounds [18]. On the other hand, the possible channeling of graphene oxide into water, as well as the lack of large-scale fabrication of GO nanomaterials, are some of the disadvantages of its application [19]. Therefore, the improvement of the capacity is a target and thus modification of GO with various selective metals [20] can further improve the adsorption properties of GO by providing additional active sites [21].

In this study, the modification of GO with AlCl₃ is examined and the resulting GO-Al composite is examined for the removal of HA from aqueous solutions. In previous studies, aluminum chloride (AlCl₃⋅6H₂O) was used as a coagulant [22,23] for humic acids removal, but in this case, the risk of residual aluminum concentration in the water, which is harmful to human health, is increased [24]. For this reason, the ability of aluminum to remove HAs is tested, but incorporated into the structure of the GO, which is a combination that has not been previously tested in the existing and recent literature for the removal of NOMs. The main oxygen-containing functional groups of HA, i.e., hydroxyl, carboxyl, and phenol, were found in the literature to bind with trivalent aluminum (Al³⁺) during adsorption [1]. pH, adsorbent dosage, initial HA concentration, and contact time were investigated as factors affecting adsorption. Furthermore, the data were assessed and modeled by means of adsorption kinetics and adsorption isotherm equations.

2. Materials and Methods

2.1. Materials

All the materials used, except one, are supplied by Sigma-Aldrich, Merck KGaA, Darmstadt, Germany and deionized water was used in any case. First, 0.01 g of humic acid was dissolved in 50 mg L⁻¹ of distilled water and 2 mL of 0.5 M NaOH were added to dissolve the humic acid (stock aqueous solution). AlCl₃·6H₂O was used to modify graphene oxide (GO). Finally, graphite flakes were used for the preparation of graphene oxide. For pH adjustment HCl (37% HCl (Panreac, AppliChem, Barcelona, Spain) and NaOH solutions (≥97.0% ACS NaOH pellets) were used.

2.2. Synthesis of GO-Al

The GO was synthesized according to the modified Hummer’s method [25], as enhanced by Debnath et al. (2014) [26]. For the synthesis of GO-Al composite, 1 g GO and 1 g AlCl₃·6H₂O (1:1) were added in 500 mL of water and stirred for 2 h at 298 K and then sonicated. Afterward, the resulting composite was filtered and water-washed. The formed adsorbent, after cooling to room temperature, was ready for use.

2.3. Analytical Determinations

The residual humic acid concentration was calculated by fitting the absorbance obtained by a UV-Vis spectrophotometer (WTW Spectroflex 6100, Weilheim, Germany) at λ_max = 254 nm, to the calculated calibration curve.

2.4. Adsorption Experiments

In total, 10 mL of HA solution were added in falcon tubes by adding the proper amount of the adsorbent and mixed in a Trayster overhead shaker and Loopster rotator at 80 rpm. Several experimental parameters were run self-sufficiently, maintaining others constant, such as pH (2.0 to 11.0), initial concentrations HA (2 to 50 mg L⁻¹), dosages (0.2, 0.5, 0.8, 1.0 and 3.0 g/L), and contact time (2 to 240 min). The removal percentage R, %), is given by the subsequent equation (Equation (1):

$$R\left(\%\right)=\left({\displaystyle \frac{C_0-C_f}{C_0}}\right)\times 100\mathrm{\%}$$

(1)

where C₀ = initial HA concentration (mg L⁻¹), C_f = residual HA concentration (mg L⁻¹).

For the adsorption capacity of GO-Al, Q_e (mg g⁻¹) the subsequent Equation (2) was used:

$$Q_e={\displaystyle \frac{(C_0-C_e)\times V}{m}}$$

(2)

where C_e = equilibrium concentration of HA (mg L⁻¹) at equilibrium, V = volume (L), and m = mass of the adsorbent (g).

2.4.1. Equilibrium Experiments

A fixed amount of GO-Al adsorbent (g) was added to 10 mL of HA solution (2.0 to 50 mg L⁻¹) in falcon tubes. For the appraisal of the results, Langmuir [27] (Equation (3)) and Freundlich [28] (Equation (4)) isotherm models were applied. Adsorption isotherms are used to find out the mechanism of adsorption and to relate the residual concentration of HA with the adsorption capacity of GO-Al adsorbent. The Langmuir isotherm assumes monolayer adsorption, representing a homogenous adsorption process. In contrast, the Freundlich isotherm describes heterogeneous surface adsorption involving multiple layers.

$$Q_e={\displaystyle \frac{Q_m{K}_L{C}_e}{1+K_L{C}_e}}$$

(3)

$$Q_e=K_F{C}_e^{1/n}$$

(4)

where Q_e = concentration of HA adsorbed in GO-Al (mg g⁻¹) at equilibrium, Q_m = maximum adsorption capacity (mg g⁻¹), and K_L = the relative energy for HA adsorption (L mg⁻¹). K_F = constant that refers to the adsorption capacity, 1/n is a constant estimating the adsorption ability or surface heterogeneity.

2.4.2. Kinetics Experiments

Pseudo-first order (PFO) (Equation (5)) and pseudo-second order (PSO) [29] (Equation (6)) kinetic models were used to study the result of interaction time in the adsorption process.

$$Q_t=Q_e(1-e^{-k_{1}t})$$

(5)

$$Q_t={\displaystyle \frac{k_2{Q}_e^{2}t}{1+k_2{Q}_{e}t}}$$

(6)

where Q_t and Q_e are related to the quantity of HA adsorbed (mg g⁻¹) during t (min) and at equilibrium. The rate constants k₁ (1/min) and k₂ (g (mg min)⁻¹) denote the rate of adsorption for the PFO and PSO models.

2.5. Thermodynamics

The evaluation of changes in Gibbs free energy (ΔG°, kJ/mol), enthalpy (ΔH°, kJ/mol), and entropy (ΔS°, kJ/mol·K) constitute the thermodynamic analysis of the adsorption process at four temperatures in this study (303, 313, and 323 K) and is achieved by using Equations (7)–(10) for the corresponding calculations [30].

$$\mathsf{\Delta }G°=\mathsf{\Delta }H°-T\mathsf{\Delta }S°$$

(7)

$$\mathsf{\Delta }G°={-RT\mathrm{l}\mathrm{n}(K}_c)$$

(8)

$$K_c={\displaystyle \frac{C_s}{C_e}}$$

(9)

$${ln(K}_c)=\left(-{\displaystyle \frac{\mathsf{\Delta }H°}{R}}\right)+{\displaystyle \frac{\mathsf{\Delta }S°}{R}}$$

(10)

where K_C = thermodynamic constant, R = 8.314 J/mol·K denoted to the universal gas constant, and T = temperature (K). C_s (mg/L) = the quantity adsorbed at equilibrium, and C_e = the concentration (mg/L) at equilibrium.

ΔG° can be received from Equation (8), and ΔH° and ΔS° from the slope and intercept of ln(K_C) − 1/T plot.

2.6. Characterization Techniques

Scanning Electron Microscopy (SEM) (Jeol JSM-6390 LV, Tokyo, Japan scanning electron microscope) was used for the characterization of GO-Al surface, before and after HA adsorption.

2.7. Regeneration Study

Several adsorption–desorption experiments were conducted by adding 0.01 M NaOH as the regenerant [31,32] up to 10 cycles. After the first cycle, the saturated GO-Al with HA was shacked at 80 rpm in a NaOH solution at optimum pH value and contact time, and then, through a membrane filter, the GO-Al was separated and heated in an oven at 80 °C for 2–3 h to be reused for the following cycle. Following this method, the recycled GO-Al was reused up to 10 cycles.

3. Results and Discussion

3.1. Effect of Experimental pH Solution

One of the most significant variables that affects adsorption is the experimental pH of the solution. Therefore, this effect was studied in the pH range of 2.0–11.0 ± 0.1 by adding 1.0 g L⁻¹ GO-Al or GO (for comparison purposes) to a 5 mg L⁻¹ HA solution at room temperature for 24 h. According to the results obtained (Figure 1), the effectiveness of GO-Al over pure GO is obvious.

Furthermore, the optimum pH was found to be 2.0 ± 0.1, showing 91% HA removal with the addition of GO-Al, while the relative removal of pure GO was found to be 81%. It is detected that with the increase in the solution pH, the efficiency decreases, which can also be interpreted by determining the point of zero charge (pH_pzc) of the surface of the materials (estimated by the pH drift method [33]). Specifically, the pH_pzc of GO-Al was found to be 6.36 and that of GO was 6.87 (Figure 2). Below this value, the surface of the materials is being positively charged, resulting in the interaction with the already negatively charged molecules of HA causing the increase in the removal percentage. When the pH increases, the surface of the material converts to negative and due to the abundance of hydroxide ions at higher pH, the option of repulsive forces among the negative adsorbent surface, and the also negative charge of humic acid, is higher. This may be the reason for lower humic acid adsorption at higher pH. Particularly at pH values above pH_pzc, the percentage drops to 31% for GO-Al and 17% for GO. Previous studies [34] confirm the presented results.

3.2. Effect of Dosage

The effect of adsorbent mass was studied in batch experiments by varying the dosage of GO-Al and GO from 0.2 to 3.0 g/L at the optimum pH of 2.0 ± 0.1 (as it was found previously) in order to cover a wide range of values. The removal rate of HA is enhanced as the adsorbent dosage increases, showing a removal rate from 71% to 98% when applying 0.2 and 3.0 g L⁻¹, respectively, of the optimum GO-Al (Figure 3). The relative removal percentages for unmodified GO were 63 and 87%.

Figure 3 also shows the change in adsorption capacity (Q_e) by increasing the adsorbent’s mass. As shown, by increasing the adsorbent’s dosage, the adsorption capacity decreases due to the reduced number of active adsorptive sites on the adsorbent surface [34]. Specifically, from 17.83 mg g⁻¹ with the addition of 0.2 g L⁻¹ GO-Al, Q_e decreases to 1.64 mg g⁻¹ when 3.0 g L⁻¹ are added. Regarding the adsorption capacity, similar values were obtained using GO, i.e., 15.71 mg g⁻¹ to 1.42 mg g⁻¹. Eventually, 1.0 g L⁻¹, which shows 91% removal, is selected for the following evaluation experiments.

3.3. Effect of Contact Time—Kinetics

Figure 4 shows the effect of contact time, in the range of 10–1440 min under optimal conditions, on the adsorption of HA onto GO-Al. It is observed that 65% of HA was removed in the first 10 min, while after 30 min the percentage reached 82% and then the removal rate started to slow down, as the existing adsorptive sites, which were “uncovered/empty” at the beginning of the process, started then to become saturated [35]. Finally, after 24 h (1440 min), 91% of HA was removed.

Furthermore, the most widely used kinetic models, such as pseudo-first order (PFO) [36] and pseudo-second order (PSO) [37], were applied in this study. The comparative fits and the obtained parameters according to the relative models, are presented in Figure 5 and

Table 1. Parameters of PFO and PSO models.

Pseudo-First Order Model (PFO).
Qe.exp (mg g−1)	k1 (min−1)	Qe.cal (mg g−1)	R2
4.55	0.1175	4.27	0.9906
Pseudo-Second Order model (PSO)
Qe.exp (mg g−1)	k2 (L (mg∙min)−1)	Qe.cal (mg g−1)	R2
4.55	0.0408	4.56	0.9952

, respectively. As can be seen, the PSO kinetic model fitted to the adsorption of HA on GO-Al better, indicating a higher coefficient parameter (R² = 0.9952). However, what is observed is that the correlation coefficient for PFO is also high (R² = 0.9906), indicating that different mechanisms may occur at different stages of the adsorption process [38]. Furthermore, the experimental adsorption capacity (Q_e.exp) was 4.55 mg g⁻¹, while the calculated adsorption capacity (Q_e.cal) for the PSO model was equal (4.56 mg g⁻¹). For the PFO model, the relative value was lower (4.27 mg g⁻¹). Therefore, the fitting of the adsorption to the PSO model suggests that chemisorption might be the rate-controlling step [34], but the behavior of PFO is often observed at shorter contact times and the kinetics of PSO becomes more evident at longer contact times, as also explained in the literature [38,39].

3.4. Adsorption Isotherms

Since the concentration of HA in natural surface waters range from 0.1 to 20 mg L⁻¹ [4], the examination of the effect of initial concentration of HA, by means of isotherm study, is important. In this study, different initial concentrations of HA (2 and 50 mg L⁻¹) were prepared, and the HA removal efficiency was studied at pH 2.0 ± 0.1 after 30 min. As is shown in Figure 6, the initial concentrations of 2 and 5 mg L⁻¹ were almost completely removed (99 and 82%, respectively) within 30 min, and after 20 and 40 mg L⁻¹, the adsorption efficiency gradually decreased (from 67 to 39%).

Two isotherm models, Langmuir [27] and Freundlich [28], which are most frequently used in the literature, were applied for the adsorption of HA on GO-Al. According to the correlation coefficient (R²) (

Table 2. Langmuir and Freundlich isotherm model parameters.

Langmuir Isotherm Model
Qm (mg g−1)	KL (L mg−1)	R2
5.91	0.3655	0.9584
Freundlich isotherm model
1/n	KF (mg g−1) (L mg−1)1/n	R2
0.2867	2.09	0.9087

), the Langmuir model emerged to be rather more acceptable (R² = 0.9584) than the one calculated by Freundlich model (R² = 0.9087), indicating that the adsorption of HA onto GO-Al would take place in a monolayer adsorption [28]. The maximum adsorption capacity, Q_m, with respect to the Langmuir isotherm model, was calculated to be 5.91 mg g⁻¹. In the recent literature, only one article by Naghizadeh et al. [15] was found to remove HA with GO, which presented a higher adsorption capacity, i.e., 39.3 mg g⁻¹, as presented in a recent review by Alomar et al. [3], which indicates that despite the lower Q_m, the proposed material of this research presents innovation.

3.5. SEM Images of GO-Al

Scanning Electron Microscopy (SEM), supplied from Jeol JSM-6390 LV, Tokyo, Japan Scanning Electron Microscope, was used to analyze the surface morphology of net GO and GO-Al, before and after adsorption of HA. As depicted in Figure 7a, GO shows a surface with a wrinkled and layered morphology, but a rougher and more porous surface is observed after modification of GO with aluminum, before HA adsorption (Figure 7b), contributing to a suitable and beneficial surface for adsorption. After HA adsorption (Figure 7c) the surface becomes smoother, thus confirming the successful adsorption of HA. Therefore, the smoother the surface, the stronger the interaction between HA and the GO-Al composite.

3.6. Thermodynamics

The results obtained from the thermodynamic study are shown in

Table 3. Thermodynamic parameters for HA adsorption on GO-Al (C₀ of 5 mg L⁻¹; pH 2.0 ± 0.1, dose of 1.0 g L⁻¹; 298, 303, 313, K; 30 min).

Adsorbent	T (K)	∆G° (kJ mol−1)	∆H° (kJ mol−1)	∆S° (kJ mol−1∙K−1)	R2
GO-Al	298	−3.752	15.507	0.0646	0.9997
	303	−4.075
	313	−4.721

. Therefore, the calculated negative ΔG₀ values are indicative for the spontaneous adsorption of HA on the surface of GO-Al, while the positive ΔH₀ value (5.507 kJ mol⁻¹) and the positive ΔS₀ value (0.0646 kJ mol⁻¹∙K⁻¹) confirm the endothermic nature [40].

3.7. Regeneration Study

To assess the reuse and effectiveness of adsorbents, it is important to conduct a regeneration study through cycling experiments [41]. Thus, in this study, 1.0 g L⁻¹ GO-Al was added to 5 mg L⁻¹ HA at pH 2.0 ± 0.1 for 30 min. Consequently, the 1st cycle was started and, after 30 min, the solid was separated from the supernatant and 0.01 M NaOH solution was used to wash the used material. This was followed by washing with distilled water to eradicate any remaining base. The procedure was repeated for the next 10 cycles. The results are presented in Figure 8. As illustrated, the GO-Al material is effective and has been successfully regenerated up to the first five cycles (green area on the diagram). Particularly, there is only an overall diminution in efficiency of 14% up to five cycles. On the other hand, after the 6th cycle (red area on the diagram) the reduction rate increases reaching the 10th cycle (17.5%) with a total reduction of 72.8%, which may be due to the destabilization of the material and its saturation.

4. Conclusions

In this study, a proposed material containing graphene oxide and aluminum chloride (GO-Al) was investigated for the potential removal of HA from aqueous solutions. As can be seen from the results, the modification of GO with AlCl₃⋅6H₂O enhanced the removal of HA. At pH 2.0 ± 0.1, 91% was removed by adding 1.0 g L⁻¹ GO-Al. The kinetic study showed that after 30 min, almost equilibrium was reached as 82% removal was achieved. Furthermore, the PSO kinetic model was found to fit the adsorption process better. Regarding the evaluation of the isotherms, the relevant results were more aligned with the Langmuir isotherm model. Thus, it can be concluded that the adsorption process was closer to chemisorption. The maximum adsorption capacity (Q_m) was calculated to be 5.91 mg g⁻¹ GO-Al.