1. Introduction
The industrial sector produces approximately 1.65 million tons of dyes annually, of which 10–15% is released to the environment as wastewater [1–3]. Methylene blue (MB) and methyl green (MG) are two cationic dyes widely used in the textile and printing industries, and they play a crucial role in water pollution [4–6]. These dyes are toxic and persistent in the environment, where exposure may lead to skin diseases, allergic reactions, and even cancer [7]. Various technologies are available for removing dye pollutants in wastewater [8, 9], including coagulation/flocculation [10], ozonation [11], membrane filtration [12], chemical oxidation [13], and adsorption [14]. Each method varies in efficiency depending on the nature of the wastewater [15]. For this reason, adsorption has been the method of choice due to its effectiveness, simplicity, and low cost [16–18]. The economic viability of material production is critical for developing specialized adsorbents. Various adsorbent materials have been investigated for removing dyes from wastewater, including biochar [17], metal-organic frameworks [18], carboxymethylcellulose [19], activated carbon [20], and graphene [21].
Graphene oxide (GO) exhibits excellent potential for the adsorption of dyes. Among a variety of available techniques for dye removal, adsorption is known to be one of the most effective processes, which led to the emergence of GO as one of the most promising candidates due to its tunable surface chemistry [22]. The use of GO for environmental remediation has gained considerable attention due to its outstanding properties, such as a large surface area, high chemical stability, and the ability to adsorb organic pollutants from wastewater [23–26]. Oxygen-containing functional groups on the GO include carboxylic acid, carbonyl, epoxy, and hydroxyl [27].
Researchers have investigated modifying the classical GO manufacturing processes to enhance dye adsorption efficiency. The Hummers method, one of the most common GO production methods, requires several oxidation and reduction stages that prolong the process [28, 29]. Although this process produces high-quality GO with a large specific adsorption capacity for pollutants (e.g., cationic dyes), it is time-intensive and leads to structural defect deterioration if handled improperly [30, 31]. Tuning the synthesis protocol to achieve a compromise between the rapid production speed and the GO’s adsorption performance is recommended. A viable approach is to control both the drying temperature and the drying time during the synthesis process. Using lower temperatures for drying GO helps preserve its structure while speeding up its fabrication. Tuning the synthesis parameters may improve the surface property of GO, further enabling its ability to adsorb organic dyes [32].
The oxidation and exfoliation of graphite were improved by synthesizing GO at room temperature for 24 hours, as supported by Field Emission Scanning Electron Microscopy (FE-SEM) analysis. A maximum MB adsorption efficiency of 97.4% was achieved after 120 minutes of adsorption when MB adsorption onto GO was combined with Cu-CP (Cu-CP@GO) [18]. GO was dried at 60°C and yielded crumpled sheets. SEM analysis revealed that the adsorbent was selective for MB removal, achieving an efficiency of 88% using Mg(OH)2 [33]. Additionally, a Brunauer-Emmett-Teller (BET) surface area of 112.9 m2/g and an MB dye adsorption capacity of 9.46 mg/g were obtained for Cy/GO/Fe2O3 synthesized at a drying temperature of 100°C [34].
Several methods applied to the dye removal process generally involve a physical adsorption process, which commonly has drawbacks such as low adsorption capacity, tedious synthesis of catalysts, and labor-intensive quality. Thus, our research aims to develop a new, more environmentally friendly methodology using carbon-based compounds as adsorbents and to enhance the properties of GO through a novel synthesis approach, which involves shortening the synthesis time, eliminating the use of hazardous chemicals, and utilizing hydrothermal methods. Evaluate the efficacy of GO as an efficient adsorbent for cationic dyes as part of a dye removal methodology. Additionally, this study presents experimental findings and theoretical correlations.
2. Materials and Methods
2.1. Materials
The synthesis of GO was carried out using Hummers’ method, and all the chemicals were obtained from Sigma-Aldrich: graphite powder, phosphoric acid (H3PO4 98%), sulfuric acid (H2SO4 98%), potassium permanganate (KMnO4 98%), and hydrogen peroxide (H2O2 30%). Reagents of the highest purity grade were used without further purification. MB and MG were obtained from Bio Super Star Co., Ltd., Taiwan. All pure chemicals/reagents were used as received, and deionized (DI) water was used in all experiments.
2.2. Preparation of GO
A modified Hummers method was used to synthesize GO from natural graphite powder. Unlike most studies that synthesize GO traditionally (Hummer method) using many chemicals and a lengthy synthesis process, this study takes a different approach. It modifies the traditional GO synthesis to be more eco-friendly by minimizing the use of chemicals and shortening the synthesis process while still producing high-quality GO [35, 36]. The GO synthesis begins with the combination of graphite powder (5 g) and H3PO4 (2 g) in 98% concentrated H2SO4 (90 mL) in a beaker. The mixture was stirred in an ice bath for 15 minutes at below 5°C. KMnO4 (10 g) was added slowly while stirring for 45 minutes. The sample was stirred at 40 °C in an ice bath for 40 minutes. Gradually, 500 mL of DI water is added over 20 minutes at approximately 80°C, causing the sample to turn dark brown. After that, 15 ml of 30% H2O2 is added while stirring for 30 minutes to turn the color light brown. After that, the sample was allowed to cool before performing the hydrothermal method. The resulting graphite oxide paste was then subjected to hydrothermal treatment at 190 °C for 2 hours. Next, the sample was washed repeatedly using a centrifugation technique with DI water at 6000 rpm for 15 minutes until it reached a neutral pH.
Three independent batches were prepared under identical hydrothermal conditions to assess the reproducibility of the GO synthesis process. Characterization results, including X-ray photoelectron spectroscopy (XPS), combined with SEM and Energy-dispersive X-ray Spectroscopy (EDS), Fourier-transform infrared (FTIR) spectroscopy, zeta potential analysis, and BET surface area measurements, confirmed consistent material properties across batches, with minimal variation observed. These results demonstrate that the developed synthesis method is highly reproducible and reliable for consistent GO production.
2.3. Characterizations of Materials
Different analytical methods were used to investigate the qualitative contents of the synthesized GO materials as adsorbents. XPS surface composition was analyzed using the Thermo Fischer Scientific ESCALAB instrument (USA) at CSIR-NEIST, Jorhat, India. The JEOL JSM-6510LV microscope was used at 20 kV accelerating voltage for SEM analysis (Japan) and extensive material characterization, paired with an Oxford Instruments INCA Energy 250 EDS Microanalysis System (United Kingdom). FTIR was performed using a Bruker ALPHA FTIR spectrometer (Germany) to characterize GO based on the oxygen-containing functional groups in the GO samples. The colloidal system’s surface charge and stability were assessed by measuring the zeta potential (ζ) using a Malvern Zetasizer ZS (United Kingdom), with values calculated based on particle electrophoretic mobility. Dynamic light scattering (DLS) was used to determine the Z-average (Z-Avg) by monitoring the Brownian motion of the particles. BET-specific surface area measurements were conducted using an Autosorb-1 Quantachrome instrument (USA). Residual dye concentrations were examined using an OPTIZEN POP UV-Vis spectrophotometer (South Korea), and the pH of the solutions was analyzed with a JENKO 6173 spectrophotometer (Taiwan).
2.4. Batch Adsorption Experiments
In this study, dye concentration intervals ranging from 10 to 1000 ppm were selected to cover both typical and highly polluted industrial wastewater scenarios. While most industrial wastewater discharges contain dyes at concentrations between 10 and 200 ppm, certain effluents, especially those from the textile and printing industries, have been reported to reach 700–800 ppm concentrations. The industry generates around 100,000 types of commercial dyes at an annual production of 700,000 tons [37, 38]. Testing up to 1000 ppm enables a thorough assessment of the adsorption performance of exfoliated GO under both normal and extreme contamination conditions. This extended range ensures a comprehensive evaluation of the material’s adsorption capacity, kinetics, and isotherm behavior, which is critical for understanding its scalability and feasibility in real-world industrial wastewater treatments [1, 6].
The target dye samples in this study were MB and MG. Adsorption tests were conducted in a static batch system at approximately 30 °C to determine the optimal concentration and to establish the range of GO’s adsorption of MB and MG. All adsorption experiments were conducted at approximately 30 °C, corresponding to standard laboratory conditions and reflecting typical ambient wastewater temperatures in tropical regions. This fixed temperature was selected to minimize external thermal influences and to isolate the effects of pH and dye concentration on adsorption performance. Dye concentrations were adjusted during three intervals of increasing scale; the first interval used 10, 50, and 100 ppm dye concentrations. The second interval used 100, 300, and 500 ppm, and the third interval used 500, 700, and 1000 ppm. All adsorption experiments were performed in 250 mL glass containers containing 50 mL of dye solution with GO powder (0.02 or 0.05 g). The optimal dose and concentration were selected based on kinetics, isotherm models, and regeneration studies. The MB and MG post-adsorption concentrations at 664 and 633 nm were measured using an Optizen Pop spectrometer. The removal amount (R, ppm) and removal efficiency (Re, %) were calculated using Eqs. (1) and (2) [39, 40].
where C0 and Ce are the initial and equilibrium concentrations (ppm) of the adsorbate in the solution, respectively.
Additionally, the effect of pH on the removal efficiency and adsorptive capacities of MB and MG was evaluated. Both dyes were prepared at pH 3, 6, and 9. Both kinds of dye adsorption kinetic experiments were performed under the same conditions, and the optimal dose of GO and a suitable concentration were selected for each time interval. Adsorption kinetics were determined over 30, 60, 90, 120, 150, 180, 210, and 240 minutes. At each time point, the amount adsorbed (qt, mg/g) was calculated using Eq. (3) [41].
where qt is the amount of dye adsorbed on GO at any time (mg/g), Ct is the remaining dye concentration in the solution at time t, W (g) is the mass of the GO powder used, and V (L) is the volume of the dye solution.
2.5. Regeneration Ability Study
Adsorption-desorption experiments for the GO powder were performed over six cycles to evaluate the potential reuse of the GO powder as an adsorbent for MB and MG. At the end of each cycle, the GO powder containing the dye was suspended in ethanol, followed by stirring for 12 hours. The dye concentration in the 90% ethanol solution was measured using a UV-Vis spectrophotometer. The GO powder was then filtered, rinsed several times with DI water, and dried at 60°C for 10 hours before dye adsorption. This also allowed us to assess the efficiency of the powder in removing dyes after multiple applications.
2.6. Statistical Analysis
This study’s MB and MG removal data, as well as desorption efficiency, are presented as mean ± standard deviation of three replicates. Data calculations and analysis were performed using Microsoft Excel 365 and OriginLab version 2024, with OriginLab specifically used for statistical analysis. ANOVA and least significant difference (LSD) tests (p <0.05) were conducted to compare the effects of GO adsorbent dosage, pH, and initial concentration on the reduction of MB and MG at each interval. The final results of these tests were used to calculate the adsorption capacity according to various models.
3. Results and Discussion
3.1. Characterization of GO
It is noted that while this study extensively characterized the synthesized GO using XPS, SEM-EDS, FTIR, zeta potential, and BET analyses, X-ray diffraction (XRD) analysis was not conducted. Although XRD could provide additional insights into the crystallinity and phase structure of the material, the characterization techniques used herein sufficiently confirmed the surface functionalization, morphological changes, and surface area enhancement critical to the adsorption applications of GO. Future studies may include XRD measurements to further complement the understanding of the material’s structural properties.
3.1.1. XPS analysis
The XPS analysis in Fig. 1 shows that the surface chemistry between graphite and GO differs. C1S spectrum of the raw graphite (Fig. 1a) showed a comparison peak at the binding energy of 284.33 eV, which was assigned to C-C bondings due to the graphitic structure with 66.77% atomic concentration. The C1S spectra of the GO sample (Fig. 1b) were more complex, with five peaks: 284.40 eV (C-C), 286.88 eV (C-O), 292.68 eV (CF2), 293.34 eV (CF3), and a π -π* satellite peak at 295.88 eV. These differences were attributed to the functionalization of graphite to GO when oxygen-containing groups were added, and fluorine was substituted. As indicated in Table 1, after oxidation, the atomic concentration of the C-C bond in GO decreased from 81.50% to 59.80%, while the oxygen-containing groups in GO increased, particularly C-O, which reached 53.81%.
The GO sample contains fluorine atoms, which may enhance the material’s water-repellent properties and durability, as evidenced by the presence of CF2 and CF2 functional groups. These π -π* satellite peaks indicate that the sp2 hybridized carbon network was partially restored in GO, confirming the recovery of graphite-like characteristics. The appearance of fluorine peaks (CF2 and CF3) in the XPS spectra, although not expected from the listed synthesis reagents, might be attributed to trace contamination. Possible sources include environmental fluorine contamination during sample handling, or residual fluorine compounds originating from the chemical processing equipment or environment. Similar incidental fluorine detection in GO samples has been reported in previous studies [42, 43]. Although the detected fluorine content was relatively low, it did not significantly alter the adsorption behavior of GO toward cationic dyes. During the reduction of GO, the oxygen-containing functional groups significantly decrease, reducing the oxygen surface content from approximately 38% in pristine GO to about 5% in rGO at 900 °C. In comparison, the proportion of sp2-hybridized carbon atoms increases from 46% to 87% [42]. The C1S spectrum showed distinct peaks corresponding to the C-O and C=O groups, confirming graphite oxidation and aligning with previous studies by Luo et al. [44]. These functional groups enhance GO’s adsorption ability by physically interacting with the positively charged dyes (i.e., through hydrogen bonding or electrostatic forces).
3.1.2. FTIR analysis
Fig. 1c FTIR Spectroscopy Analysis reveals significant differences in functional group compositions between graphite and GO. In the case of graphite, an O-H stretching vibration appeared at 3434 cm−1, likely due to moisture absorption. The GO spectrum showed a shift of the O-H stretching peak to 3439 cm−1, indicating increased hydroxyl groups originating from oxidation. The GO exhibits a significant C=O stretching peak at 1634–1635 cm−1, likely due to the introduction of carbonyl groups during the oxidation process. In graphite and GO, C-O stretching vibrations of epoxy or alkoxy groups were observed at 1060 cm−1 and 1068 cm−1, respectively. The O-H deformation vibrations in C-OH structures were also observed, as evidenced by a broad peak around 1452 cm−1 in GO, which indicates that more hydroxyl groups might exist in the GO structure.
C-O, C=O, and O-H groups have been observed to appear, with one study reporting shifts in O-H and C=O stretching vibrations indicating oxidation of GO [45]. Broad peaks in 1084 and 1053 cm−1 are involved with C-O stretching vibrations, which specify the presence of oxygen functional groups [46]. The presence of the O-H group deformation vibrations in GO, at 1452 cm−1, demonstrates the successful oxidation and functionalization of GO [47]. GO has more hydrophilic functional groups than conventional graphite, improving dispersion in polar solvents such as water. Modifying and expanding peaks in the FTIR spectrum confirmed the successful conversion of the graphite into GO, enhancing the chemical reactivity and the ability to adsorb dyes such as MB and MG.
3.1.3. Zeta potential (ζ) analysis
Fig. 1d shows the zeta potential measurements, which indicate the variation of surface charge for graphene and GO suspensions as a function of pH from 2 to 12. Both materials exhibited negative zeta potentials, which increased collectively as the pH increased. The conductivity of graphite and GO at pH 12 was −38.1 and −41.53 mV, respectively. The phosphorescence peak at around 890 nm upon exposure to acidic groups indicates the tendency for GO to carry negative charges, as weakly ionizable acidic functionalities, such as carboxyl and hydroxyl groups on the GO surface, increase the negative charge of the GO with increasing pH. The data indicate that GO’s higher surface charge density could help explain why GO exhibits enhanced colloidal stability in alkaline media compared to graphite.
The zeta potential values of GO were negative, which aligns with previous studies reporting that GO sheets are strongly negatively charged when dispersed in water [48]. At elevated pH, more carboxyl and hydroxyl groups ionize, resulting in a higher surface charge and increased electrostatic repulsion between particles, which helps stabilize the colloidal suspension. A high GO zeta potential value indicates a stable colloidal system, specifically at pH 12, meaning the GO dispersion is stable [49]. At high pH, the enhanced ionization of surface functional groups increased inter-particle repulsion, thereby improving colloidal stability in these environments. Moreover, the drying temperature has a significant impact on the structural and chemical properties of GO, including its zeta potential and colloidal behavior across various environments.
3.1.4. BET analysis
The BET analysis in Fig. 1e presents a small surface area of 15.90 m2/g for the graphite, which aligns with its intact structure. In contrast, as shown in Fig. 1f, the extraction of GO increased the surface area to 167.36 m2/g, which is exceedingly high, mainly due to the separation of graphite into thinner GO sheets during oxidation and drying. The pore size distribution analysis shows that GO exhibits smaller average pore diameters (3.33 nm) than graphite (13.48 nm), indicating that the exfoliation process enhances surface area and modifies porosity characteristics.
The graphite exfoliation through oxidation processes results in larger surface areas due to added oxygen-containing functional groups [50]. This study’s measurement of 167.36 m2/g exceeded values reported by Soltani and Lee [50], who found surface areas of 112 m2/g and 66 m2/g for GO prepared using lower-intensity ultrasonic irradiation. These results suggest that the oxidation and drying methods applied to these GO nanofibers are more effective in inducing interlayer spacing and increasing the physical surface area. Compared with graphite, the most substantial contribution of surface area is associated with the effective exfoliation of GO, providing more surface sites accessible for adsorption or catalysis. In addition, the significant increase in surface area from graphite to GO is consistent with previous studies, where exfoliation and oxidation introduce oxygenated functional groups and expand interlayer spacing, leading to higher surface areas [50]. Therefore, the measured values are reasonable and confirm graphite’s successful oxidation and exfoliation into GO. Moreover, the eco-friendly hydrothermal method developed in this study exhibits strong potential for scalability. The relatively low operating temperature (190 °C), reduced chemical consumption, and shortened synthesis time compared to conventional methods make scale-up feasible. The process can be readily adapted to larger reactors by proportionally adjusting the reaction volumes without significantly changing synthesis parameters. This suggests that large-scale production of high-quality GO for industrial wastewater treatment applications is achievable.
3.1.5. SEM-EDS analysis
SEM-EDS characterization of graphite (Fig. 2a) and GO (Fig. 2b) revealed interesting structural differences. Graphite exhibits a highly laminated columnar texture with crystal aggregates resembling platelets, characteristic of its graphitic nature. The morphology of GO, as shown in SEM images, shows fractured and delaminated shapes, indicating the rupture of van der Waals interaction, followed by the dispersion of single or few-layer GO sheets.
An EDS analysis confirmed the compositional changes. Graphite is composed primarily of carbon (C, 93.2%), with small amounts of oxygen (O2, 4.1%), silicon (Si, 1.5%), and sulfur (S, 1.2%). In comparison, the elemental composition of GO is more complex, with high concentrations of C (28.7%) and O2 (66.5%), followed by Si (2.3%), aluminum (Al, 1.0%), and zirconium (Zr, 1.5%). Such compositional differences indicate the incorporation of additional functional groups synthesized during GO production, primarily oxygen-containing entities critical for adsorption. The unexpected detection of Zr in the EDS analysis could be due to contamination during the synthesis or sample preparation processes. Zr-based residues can originate from container surfaces or mechanical handling instruments, especially when using laboratory equipment made of zirconium alloys. Such minor contamination has been documented in similar GO synthesis studies [51]. The Zr concentration detected was very low and is unlikely to significantly affect the prepared GO material’s adsorption properties.
The transformation of GO in the SEM images was consistent with previous literature that demonstrated that the exfoliation of oxidized graphite caused a disturbed layered structure [50]. The GO sample is essential for increasing the surface area and enhancing the availability of functional groups for adsorption applications through the exfoliation of graphite layers. Ultrasonication in the presence of chemical oxidation resulted in the conversion of graphite particles into two-dimensional GO sheets, as indicated by a transition of the three-dimensional structure to a two-dimensional material structure, similar to the KHGO morphology found in this study [52]. Zr and Al are identified in GO samples in studies of metal doping or contamination in the synthesis of GO, impacting the material’s electronic and adsorption properties [51].
3.2. Removal of Dyes Using Adsorbent GO
Synthetic dyes such as MB and MG are used as targets for removal in this study, as shown in Fig. 3. GO is used as an adsorbent to remove both types of dyes. Both types of dyes use the same initial concentration, which is divided into three intervals. The initial concentrations for interval 1 are 10, 50, and 100 ppm; for interval 2, are 100, 300, and 500 ppm; and for interval 3, are 500, 700, and 1000 ppm. The dosage of GO used as an adsorbent for MB and MG is 0.02 and 0.05 g, respectively, with pH values of 3, 6, and 9. The division into three intervals aims to identify the highest clearance rate for each concentration level. The initial concentration with the highest removal ratio was then used as the basis for various models, including kinetic and isothermal models, as well as studies on the reuse of materials for the adsorption of MB and MG dyes.
3.2.1. MB removal
The removal of MB dye from the aqueous solution was studied at each dose of GO adsorbent within a preset pH range, as shown in Figs. 3a–f. Figs. 3a–c presents the removal and efficiency of MB at a GO dose of 0.02 g for the three intervals. Meanwhile, Figs. 3d–f represents the removal and efficiency of MB at a GO dose of 0.05 g with similar intervals. Overall, pH 6 provides better MB removal than other pHs and continues to increase as the initial dye concentration increases. The maximum MB removal for both GO doses occurred at intervals 1 and 2, with an initial concentration of 100 ppm, and at interval 3, with a concentration of 500 ppm. At intervals 1 and 2, using a GO dose of 0.02 g (Figs. 3a–b), it successfully reduced the residual MB concentration to 5.49 ± 0.43 ppm (corresponding to 94.52% removal efficiency). Meanwhile, in the third interval with the highest removal, which occurred at a concentration of 500 ppm, MB was reduced to 90.81 ± 18.29 ppm (81.84% efficiency) (Fig. 3c). For the GO 0.05 g dose, the highest efficiency in intervals 1 and 2 was observed at a concentration of 100 ppm (Figs. 3d–e), yielding a residual concentration of 3.40 ± 0.30 ppm or 96.60% removal efficiency. At interval 3 (Fig. 3f), the highest removal achieved was 70.90 ± 14.10 ppm (85.82% efficiency). Based on these results, initial concentrations of 100 and 500 ppm and a GO dose of 0.05 g were selected to calculate kinetic and isothermal models. Regarding the observation of material reuse, a concentration of 100 ppm was employed.
Increased concentrations may lead to the competition of ions and saturation of the adsorbent sorption sites, preventing further MB sorption [53]. The combination of GO and hollow ferrite nanospheres within a corn stalk (CS)-ordered porous structure enables proper immobilization and dispersion. Analysis of the GO/CS/CuFe2O4 composite revealed a complex pore structure with a specific surface area of 289.85 m2/g, which can enhance its MB adsorption performance [6]. The one-step synthesis of graphene nanosheet (GNS) adsorbents by chemical vapor deposition with a wastewater-derived catalyst can remove 99% of MB [54]. GNS adsorbents achieved the best removal efficiencies of 97.3%, 96.2%, and 95.5% within 5 min at the initial MB concentrations of 500, 600, and 700 mg/L, respectively. A 99% removal efficiency is achieved at all concentrations within 20 minutes, indicating that the adsorption equilibrium of all concentrations is reached [54, 55].
3.2.2. MG removal
The removal of MG dye using a GO adsorbent, which has been modified to the same concentration and pH, is also studied and illustrated in Figs. 3g–l. The initial concentration of MG is divided into three intervals to determine the highest concentration and pH of the removal value. Based on Figs. 3g–l, pH 9 proved more effective for the MG case than pH 3 and 6. The pH value affects the surface charge of GO and the ionization of dye molecules. At higher pH, the deprotonation of GO functional groups enhances negative surface charges, promoting electrostatic attraction with cationic dyes. This explains the improved removal at alkaline conditions, consistent with previous findings [56]. At intervals 1 and 2 (Figs. 3g–h), with a GO dose of 0.02 g, the highest removal occurred at 100 ppm, yielding a residual dye concentration of 1.08 ± 0.92 ppm, corresponding to a removal efficiency of 98.92%. This value reflects the remaining dye in solution post-adsorption, not the amount removed. As for interval 3 (Fig. 3i), the highest removal occurs at a concentration of 500 ppm with a residual concentration of 27.05 ± 22.06 ppm, corresponding to a removal efficiency of 94.60%. Meanwhile, for a GO dose of 0.05 g, the highest removal occurred at a concentration of 100 ppm for intervals 1 and 2 and 500 ppm for interval 3 (Figs. 3j–l). The highest efficiency of MG dye removal at intervals 1 and 2 is 0.96 ± 0.77 ppm, corresponding to a removal efficiency of 99.05% and 97.52% (12.42 ± 16.08 ppm) at interval 3.
The standard deviation value for residual dye concentration, particularly at low Ce levels, is relatively higher as a percentage of the mean due to limitations in spectrophotometric sensitivity at near-complete dye removal. However, all values fall within expected variability ranges and are statistically consistent across replicates. Increasing the GO dosage from 0.02 to 0.05 g enhances MG removal, with this trend being more pronounced at higher MG concentrations. The increase in removal is associated with an increase in pH and the presence of GO. Adsorbents made from reduced graphene oxide (rGO) have been found to exhibit the adsorption capacity of MG on rGO nanosheets, which increases with a higher pH of the medium, thereby altering the surface properties. Another study revealed that MG adsorption onto graphene sheets decorated with CoFe2O4 involves intraparticle diffusion, as demonstrated by intraparticle diffusion regression analysis [57].
3.3. Adsorption Experiment
3.3.1. Adsorption capacity
The enhancement in adsorption capacities of MB and MG, in terms of solution concentration, operation time, pore structure, and pH, is among the important factors in this phenomenon. Figs. 3m–n illustrate the adsorption capacity of both types of dye over 240 minutes. Selected concentrations of dyes, based on the greatest efficiency, are 100 and 500 ppm. The pH levels used were determined based on optimal efficiencies previously obtained: pH 6 for the MB dye and pH 9 for the MG dye. The adsorption experiment was conducted at 30 °C using 0.05 g of GO.
Both types of dye adsorption analysis showed improved adsorption capacities. As shown in Fig. 3m, MB exhibits the highest adsorption capacity, 99.99 mg/g at 100 ppm and 469.70 mg/g at 500 ppm. As shown in Fig. 3n, MG adsorption capacities were 97.80 mg/g and 435.05 mg/g at 100 ppm and 500 ppm, respectively. While the maximum adsorption was recorded at 240 minutes, this duration represents the upper limit of the experiment, and complete equilibrium may not have been fully achieved, particularly at higher concentrations. However, the adsorption curve visibly plateaus as they approach 240 minutes, indicating that the system was nearing equilibrium. The results confirm GO’s adsorbent capability for removing MB and MG at lower initial dye concentrations. Lower initial pollutant concentrations facilitate higher removal because, at higher concentrations, the GO adsorption sites become saturated. The decrease in removal efficiency at higher initial concentrations results from the saturation of GO active sites. With increasing dye molecules, competition intensifies, reducing adsorption performance. Similar behavior was observed in other adsorbent studies, including Bai et al. [56].
Previous studies have reported that the high MB adsorption capacity of 714 mg/g highlights strong electrostatic interactions between the dye and the abundant oxygen-containing functional groups on the surface of GO [32]. Higher initial concentrations would raise the adsorption driving forces and lead to the swift saturation of adsorption sites, contributing to declining efficiencies at 500 ppm. Another study showed that the adsorption of MG was endothermic by the CoFe2O4-decorated GO sheets at higher temperatures (298, 313, and 323 K), which can increase the mobility of MB molecules and the interaction between MB molecules and active sites [57]. Another study reported that MB dye adsorption capacity could reach 1111 mg/g using a TNT/LDH/OS composite adsorbent [14].
3.3.2. Removal adsorption of dye by GO
To complement the adsorption capacity data shown in Figs. 3m–n, this experiment also analyzed the dye removal from the solution phase by a kinetics model as illustrated in Figs. 3o–r. The study monitored the actual dye concentration and removal efficiency over time during a 240-minute adsorption process using 0.05 g of GO. Samples were taken every 30 minutes at 30°C for MB (pH 6) and MG (pH 9).
Longer processing durations resulted in enhanced adsorption of MB and MG dyes. At 100 ppm of MB, the concentration decreased from 88.02 ppm at 30 min to 4.22 ppm at 240 min. Meanwhile, MB concentration decreased from 423.41 to 83.29 ppm at 500 ppm (Fig. 3o). The highest removal efficiencies were observed at 240 min (Fig. 3p), where removal was 95.77% for 100 ppm and 83.34% for 500 ppm. The MG dye concentration at 100 ppm declined from 90.63 ppm to 9.20 ppm during the brownout period (240 minutes). At 500 ppm, the elimination of MG decreased from 463.12 ppm at 30 minutes to 94.94 ppm at 240 minutes (Fig. 3q). The MG removal efficiencies peaked during the second sampling, with 90.79% for 100 ppm and 81.01% for 500 ppm (Fig. 3r). The increased adsorption of MB and MG dyes using the GO adsorbent is attributed to the negatively charged surface of GO, which attracts positively charged dye molecules and enhances adsorption through electrostatic attraction. On the other hand, the removal efficiency appeared to decline at higher initial dye concentrations, indicating that high concentrations could hinder overall removal performance. The adsorption process is thermodynamically spontaneous. This process obtained a 95% MB elimination rate and a final concentration of 0.25 ppm [22]. A 90% removal of MB and MG was reported in another study after a few hours of contact time [58, 59].
3.3.3. Adsorption kinetic study
This section investigated the adsorption kinetics of MB and MG dyes onto GO powder (0.05 g) in the aqueous phase. Figs. 3s–v shows that both dyes were subjected to experiments at initial concentrations of 100 and 500 ppm, with pH levels of 6 for MB and 9 for MG, respectively, at 30°C. The experimental data were interpreted using two kinetic models: pseudo-first-order (PFO) and pseudo-second-order (PSO). Table 2 provides a detailed overview of the model results (kinetic and isotherm), accompanied by a graphical representation in Figs. 3s–v and Figs. 4a–f. The expression of the PFO rate is given by Eq. (4) [60].
where qe and q are the amounts of MB and MG adsorbed on GO at a given time t (mg/g), k1 is the rate constant of the PFO model for adsorption (min−1), and k2 is the rate constant of the PSO model for adsorption (g/mg min).
The PFO and PSO models best fit the data at 100 and 500 ppm concentrations, respectively. Each model describes the adsorption kinetics with different regression coefficients (R2) values. Figs. 3s–t illustrate how PFO consistently decreases with increasing contact time, whereas Figs. 3u–v indicates that PSO increases with extended durations. At concentrations of 100 and 500 ppm MB dye, with doses of 0.02 g and 0.05 g of GO (Figs. 3s–t and Table 2), the PFO model yielded R2 values of 0.9757 and 0.9990. Furthermore, MG has given R2 values of 0.9808 and 0.9816, respectively. The PSO model (Figs. 3u–v and Table 2) yielded higher R2 values: 0.9994 and 0.9991 for MB dye and 0.9995 and 0.9990 for MG dye. These observations underscore the chemical interactions at 100 and 500 ppm. The PSO model is more effective, exhibiting a higher R2. Its high R2 values confirm its suitability for systems with more complex kinetics, providing a deeper insight into the adsorption mechanism [62]. Both the PFO and PSO models were fitted to the data to validate this conclusion, and model fit was evaluated using R2 and residual sum of squares (RSS).
The PSO model consistently showed higher R2 values (>0.999) and lower RSS compared to PFO (Table 2), confirming that PSO provides the most accurate description of adsorption kinetics. The PSO model was well-suited for the covalent attachment processes of positively charged dyes, such as MB and MG, to the surfaces of GO. The lower difference between the values of both, compared to PFO, suggests that chemisorption is a suitable fit mechanism for the adsorption process in the PSO model. Chemical bonds form during these processes, and electron exchange occurs between the dye molecules and the GO surface. The initial rapid uptake was followed by a diffusion-dominated slow phase of adsorption for both MB and MG, as observed. In kinetic studies of MB dye removal using teak wood waste as the adsorbent, the PSO model is the most effective adsorption system, with R2 values exceeding 0.9927 [63].
Composite materials show similar mass transport in the dye adsorption kinetics [64, 65]. In addition, the uptake of cationic dyes was facilitated by strong surface complexation with O2 and N-positioning functional groups. The PFO and PSO model analysis fit was R2 = 0.92 and 0.99 when eliminating MB using GO and SiO2 [66]. Another study showed the R2 values for the PSO model in an MG sample to be between 0.93 and 0.96 [57]. However, this study used only one kind of GO adsorbent and did not combine it with other adsorbents. However, the R2 for the PSO model is higher (0.99) than in previous studies.
3.3.4. Adsorption isotherm experiment
The design of a commercially usable adsorbent for a specific solute illustrates its effectiveness and feasibility, as shown on the adsorption isotherm. Adsorption equilibrium was studied using a 0.05 g GO adsorbent at an initial concentration of 100 ppm for MB and 9 ppm for MG, at pH 6 and 9, respectively, both at 30 °C. The dyes were agitated at 150 rpm in a water bath. A UV-visible spectrophotometer was used to determine the equilibrium concentration and absorbance of the dye. Adsorption isotherms indicate the distribution of adsorption molecules between the liquid and solid phases at equilibrium. Fig. 4 illustrates the linear and nonlinear Langmuir, Freundlich, and Temkin isotherm models employed to investigate the adsorption capacity and mechanisms. The model results are summarized in Table 2. The R2 values of the three isotherm models for both dyes were obtained by fitting the experimental data. To rigorously validate this conclusion, the equilibrium data were also fitted to the Freundlich and Temkin models (Table 2). The Langmuir model yielded the highest R2 values (0.9995 for MB and 0.9980 for MG) and the lowest RSS compared with the alternative models, confirming its superior applicability. Although Freundlich and Temkin provided acceptable fits, their lower correlation coefficients and higher RSS values suggest that multilayer or heterogeneous adsorption is less dominant in this system.
The Langmuir model describes physical adsorption processes involving single-layer adsorption with uniform surface coverage. According to this model, for single-layer adsorption, adsorbate molecules are attracted to the adsorbent surface in a confined area, with no interaction between adsorbed molecules. Eq. (6) represents the linear regression equation applicable to the Langmuir isotherm model [67]:
where Ce (mg/L) is the concentration of the dye solutions at equilibrium, qe (mg/g) is the number of dyes adsorbed per unit mass of the adsorbent at equilibrium, qmax is the maximum adsorption capacity of the MB and MG dyes, and KL (L/mg) is the Langmuir equilibrium constant associated with the affinity of the binding sites and the energy of adsorption. Plotting Ce/qe against Ce yields qmax and the isotherm constant (KL).
A dimensionless constant separation factor (RL), defined by Eq. (8), was used to classify the isotherms and determine the favorability of Langmuir adsorption.
where Co denotes the upper limit of the initial dye concentration, and the isotherm type can be effectively described by the parameter RL, where RL = 0 indicates irreversibility, 0 < RL < 1 indicates favorability, RL = 1 indicates linearity, and RL > 1 indicates unfavorability [69].
The Freundlich model assumes a heterogeneous surface for adsorption and utilizes an exponential distribution of adsorption sites and energies [70]. This model is significantly influenced by the constants n and KF, which represent the mechanism and rate of adsorption [69]. Eqs. (9) and (10) represent the linear and nonlinear regressions of the Freundlich model, which accounts for a range of adsorption processes. Here, 1/n is associated with the solid surface heterogeneity, and n represents the nonlinearity of the adsorption-solution concentration relation [71]. Eq. (9) expresses the linear regression form of the model.
Meanwhile, the nonlinear regression is expressed in Eq. (10).
The adsorption constant, represented by KF, is associated with the binding energy, reflecting its interconnected nature.
The Temkin isotherm model states that a linear reduction in the coverage of the adsorbent surface cannot explain the adsorption of each target species. This model, previously mentioned in the context of the phenomenon, utilizes a uniform zeta bond energy distribution peak for adsorption. The mathematical representation of the Temkin isotherm for a linear system, as expressed in Eq. (11) [72, 73], and a nonlinear system, as shown in Eq. (12) [74]. The linear form of the Temkin model is represented by Eq. (11):
Meanwhile, for nonlinear regression, as shown in Eq. (12).
T, R, KT, and bT refer to the absolute temperature and universal gas constant, with defined values of 0.0095 J/mol/K for MB and 0.0092 J/mol/K for MG. The equilibrium binding constant (KT, expressed in L/mol) and the Temkin constant for the adsorption enthalpy (expressed in kJ/mol).
Fig. 4 shows the linear and nonlinear regressions for the three models, which trend upward toward equilibrium for both types of dye. In the case of GO, the holes and pores tend to become increasingly saturated with the presence of MB and MG. The Langmuir model (Fig. 4a) yielded R2 values of 0.9995 and 0.9980 for MB and MG, respectively. In the Freundlich model (Fig. 4b), the R2 values for MB and MG are 0.9467 and 0.9481, respectively. Meanwhile, the Temkin model (Fig. 4c) yields an R2 of 0.9703 for MB and 0.9580 for MG. The Langmuir non-linear model (Fig. 4d) yielded R2 values of 0.9983 and 0.9935 for MB and MG, respectively. The Freundlich non-linear model (Fig. 4e) yielded R2 values of 0.9744 for MB and 0.9658 for MG. The Temkin non-linear model (Fig. 4f) yields R2 values of 0.9644 for MB and 0.9496 for MG. The higher average R2 value for MB dye compared to the other models implied better comparability of the models for MB than for MG in terms of adsorption behavior. The molecular features of different dyes and their interactive behavior with adsorbents under specific conditions contribute to this discrepancy. MB is significantly smaller in molecular size than MG, which has a profound impact on the subsequent adsorption of MB onto GO surfaces. Hence, the functional groups and positively charged distribution of both types of dye produced different interactions with the adsorbent surface.
The adsorption of MB onto GO is best described by the Langmuir isotherm model (Table 2), indicating that single-layer adsorption occurs at a homogeneously distributed surface site. The Langmuir model, with high R2 values (typically R2 >0.95), demonstrated that GO provided a suitable surface that promotes homogeneous adsorption. MG adsorption was also well-fitted by the Langmuir model, although it had slightly lower R2 values, indicating different affinities for adsorption depending on molecular size and charge distribution. Other researchers have shown that the Langmuir isotherm model Fitting yielded R2 values close to 1.0, indicating that this model best represents the dye adsorption onto the composite sponge and confirms the formation of a monolayer adsorption onto a surface with evenly distributed adsorption sites [6, 75]. Meanwhile, based on the theory of multilayer adsorption onto a surface with heterogeneous energy, the Freundlich model provided the lowest R2 values. However, this adsorption mechanism is often found in species with non-existent site heterogeneity. A Freundlich model value greater than one indicates a process of adsorption. Modest fits were given by the Temkin model, indicating adsorbate-adsorbent interactions, which also suggested the occurrence of chemisorption in the adsorption process [76].
3.3.5. Adsorption mechanism of MB and MG
The effective removal of these dyes via wastewater treatment depends on understanding the mechanism that occurs during the adsorption of modified GO on the MB and MG dyes. GO has unique structural and chemical properties that make it a potential adsorbent for various dye pollutants, including cationic dyes such as MB and MG. The exceptional adsorption capabilities are attributed to numerous surface functional groups, particularly hydroxyl and carboxyl units, as determined by FTIR and zeta potential investigations. Such groups facilitate strong interactions with positively charged dye molecules. The hydrophilicity of GO is effectively enhanced by O2-containing functional groups, which facilitate effective interactions with the dye solution. Due to the electrostatic interaction between the negatively charged surface of GO and the positively charged molecules of the cationic dyes MB and MG, electrostatic forces dominate the adsorption process. The π -π stacking interactions occur between the dye’s aromatic rings and the GO’s surface morphology, as determined by XPS and SEM-EDS analyses, suggesting the adsorption mechanism. A larger surface area allows more dye molecules to diffuse into the porous structure of the modified GO, thereby increasing the adsorption efficiency. Although the BET surface area of the synthesized GO was moderate (167.36 m2/g), the effective adsorption performance for large molecules, such as MB and MG, is primarily driven by abundant surface functional groups and accessible external surfaces, allowing strong electrostatic and π -π stacking interactions. BET analysis revealed that the high surface area is responsible for the dye adsorption kinetics, as it provides numerous active sites that facilitate the attachment of the dye.
3.4. Reusability Study of Material
The regeneration ability of the adsorbent, along with its reuse for removing MB and MG dyes in various cycles (from 1 to 6), was studied and presented in Fig. 5. For desorption, 0.05 g of GO adsorbent was added to a dye solution containing 100 ppm of MB and MG. The solution was then equilibrated by shaking in a temperature-controlled water bath for 12 hours at 30°C and 150 rpm. Then, the absorbent was dried at 60°C for 10 hours. The GO adsorbent was treated with 100 mL of 0.1 M HCl. The adsorbed concentrations of MB and MG in the solvent were calculated using Eq. (13).
where Cd is the concentration of desorbed (ppm), and Ca is the concentration of adsorbed (ppm).
The desorption efficiencies for MB and MG dyes during the first observation (cycle 1) were 93.45% and 84.43%, respectively. Efficiency decreased after increased recycling cycles, with every six cycles (i.e., MB 44.91%, MG 34.54%). Both have a higher desorption efficiency of MB than MG. The adsorbent’s capacity to release both dyes decreased with each reuse. Compared to MG, MB-molecules have a larger and more planar configuration and thus exhibit a stronger π -π stacking interaction with the surface of GO, leading to better adsorption, as ascertained from the XPS study. Since both dyes are cationic, the larger number of N atoms and aromatic rings in MB contributes to a higher positive charge density on the dye molecules, resulting in a stronger electrostatic attraction between MB and the negatively charged sites on GO. The reduction in adsorption efficiency after six regeneration cycles is mainly due to mechanical damage to the GO sheets, partial loss of oxygen-containing functional groups, irreversible binding of dye molecules that block active sites, and agglomeration of GO layers, which reduces the surface area. Furthermore, compared to GO composites reinforced with other materials, pure exfoliated GO is more sensitive to mechanical and chemical stress, which explains the decline in performance. Despite this, the material demonstrates good potential for repeated dye removal applications. Due to a higher surface charge density, MB achieves a greater adsorption capacity than MG. Another study reported 98.73% MB removal, where MG’s removal efficiency was significantly lower under similar conditions [77].
Although the XPS, EDS, and FTIR analyses were comprehensively conducted for the freshly synthesized GO to confirm its chemical structure and functional groups, these characterizations were not repeated for the reused GO samples. This is because the regeneration process employed mild conditions (ethanol washing and drying at 60°C) designed to preserve the original chemical structure of GO. The gradual decrease in dye removal efficiency observed over reuse cycles is primarily attributed to physical fouling or partial pore blockage, rather than to significant chemical or structural degradation of the adsorbent. Accordingly, in future work, SEM, FTIR, and XPS analyses will be performed on reused GO to directly confirm any morphological and functional-group changes after multiple regeneration cycles. These analyses will also help validate the material’s structural stability during recycling applications.
3.5. Comparative Performance of GO and Similar Adsorbent Materials
To provide a comprehensive comparison with existing materials, the performance of the synthesized exfoliated GO was evaluated against similar adsorbents reported in the literature. As summarized in Table 3, the prepared GO exhibited outstanding adsorption capacities of 622 mg/g for MB and 584 mg/g for MG, surpassing many conventional and composite-based adsorbents, such as GO/CS/CuFe2O4 composites and clay/GO/Fe2O3 nanocomposites. Compared to traditional GO produced via the Hummers’ method, which often requires longer synthesis times and more aggressive chemicals, the eco-friendly exfoliated GO synthesized in this study demonstrates superior adsorption performance with a greener and more efficient production method. The results highlight the potential of the synthesized GO as a competitive adsorbent for practical dye removal applications, offering a balance of high capacity, process sustainability, and reusability.
4. Conclusions, Limitations, and Future Work
4.1. Conclusions
This work presents a practical and reproducible method for synthesizing GO through an eco-friendly modification of the classical Hummers’ method, enhanced by a hydrothermal post-treatment step. The synthesis substantially reduces environmental burden by minimizing chemical inputs and energy requirements. Specifically, the amounts of concentrated H2SO4 and KMnO4 were reduced by approximately 30%, the total synthesis duration was shortened from >12 hours to 3 hours, and post-treatment high-temperature annealing (>900 °C) was eliminated. Drying was achieved at 60 °C, representing a significantly lower energy demand than conventional methods. These modifications collectively decrease hazardous chemical waste generation and energy consumption, supporting our claim of a greener synthesis pathway. The synthesized GO achieved substantial adsorption capacities for MB (622 mg/g) and MG (584 mg/g), outperforming several reported GO-only adsorbents. Characterization confirmed strong electrostatic and π -π interactions between the dye molecules and the oxygen-functionalized GO surface, supported by extensive physicochemical analysis. The adsorption mechanism followed PSO kinetics and Langmuir isotherm behavior, indicating chemisorption and monolayer adsorption. Furthermore, GO demonstrated satisfactory regeneration capacity over multiple cycles, underscoring its potential for real-world implementation. While the novelty lies not in the material itself but in the accessible and greener synthesis process, this study contributes a valuable, scalable solution for wastewater treatment.
4.2. Limitations and Future Work
The present study highlights the role of pH on the adsorption of MB and MG onto GO and demonstrates the environmental advantages of the eco-friendly synthesis approach. While the present study systematically examined pH’s effect on MB and MG’s adsorption behavior onto GO, all experiments were conducted at a constant temperature (30 °C). This design ensured that the role of solution chemistry could be isolated without the influence of thermal variation. However, because arranging Arrhenius activation energy (Ea) requires adsorption kinetic data obtained at multiple temperatures, this parameter could not be determined within the current scope. Future investigations will therefore focus on temperature-dependent kinetic studies to derive Ea values. Such analysis will provide deeper mechanistic insight into whether the adsorption is predominantly controlled by physisorption or chemisorption under varying environmental conditions, thereby strengthening the understanding of the adsorption process.
On the other hand, a full life cycle assessment (LCA) should be incorporated to quantitatively benchmark the environmental footprint of this method against conventional GO synthesis routes, ensuring further validation of its eco-friendliness. In addition, although XPS, SEM-EDS, FTIR, zeta potential, and BET analyses collectively confirmed the successful oxidation, exfoliation, and surface functionalization of graphite into GO, the study did not employ XRD characterization. While this omission does not compromise the validation of adsorption-related properties, XRD could provide complementary insights into crystallinity and interlayer spacing. Future work will incorporate XRD analysis to strengthen the structural understanding of GO and further correlate its crystalline features with adsorption performance. Moreover, the reusability study revealed that the adsorption efficiency of GO decreased to 44.91% for MB and 34.54% for MG after six cycles, primarily due to partial mechanical damage, agglomeration, irreversible dye binding, and loss of oxygen-containing functional groups. While this performance remains promising for pure GO, future work will focus on strategies to enhance stability and reusability. These include embedding GO into composite materials (e.g., polymers, biochar, or magnetic nanoparticles) to improve structural robustness, employing gentler regeneration protocols to minimize functional group loss, functionalizing the GO surface with stabilizing groups, and incorporating GO into crosslinked hydrogel or membrane systems. Such approaches are expected to mitigate degradation during repeated use and improve the practical applicability of GO for industrial wastewater treatment.
PDF Links
PubReader
Full text via DOI
Download Citation 
Print 
