A multifunctional graphene oxide–ZnO nanohybrid for rapid and highly efficient malachite green adsorption and strong broad-spectrum antimicrobial activity

Abstract

A multifunctional graphene oxide–zinc oxide (GO–ZnO) nanohybrid was developed to tackle two significant challenges in wastewater treatment: the removal of malachite green (MG) and the inactivation of pathogenic microorganisms. Under optimized conditions (25 mg L⁻¹ MG, pH = 9.0, 18 mg adsorbent, 25 °C, 12 min), this material achieved a remarkable 96.54% removal of MG. The adsorption process was well-described by the Langmuir isotherm (Q_max = 131.91 mg g⁻¹) and followed pseudo-second-order kinetics, indicating that the mechanism involved chemisorption through electrostatic attraction, π–π interactions, and hydrogen bonding. The process was spontaneous, endothermic, and driven by entropy changes. Notably, the nanohybrid maintained an adsorption efficiency of over 87% after four regeneration cycles, using only 1 mL of methanol per cycle. It also demonstrated robust performance in real water samples, including tap water, river water, and industrial wastewater, retaining an MG removal efficiency of over 88% despite the complexity of these matrices. In addition to its adsorption capabilities, the nanohybrid exhibited potent broad-spectrum antimicrobial activity, with minimum inhibitory concentrations (MICs) as low as 0.313 mg mL⁻¹ against a clinical isolate of Acinetobacter baumannii, 0.625 mg mL⁻¹ against Escherichia coli (ATCC), and 1.25 mg mL⁻¹ against Staphylococcus aureus and fluconazole-resistant Candida albicans. This dual functionality stems from synergistic mechanisms: ROS generation mediated by ZnO and physical membrane disruption induced by GO. Overall, the ultrafast kinetics, high reusability, environmental resilience, and integrated detoxification-disinfection capabilities highlight the GO–ZnO nanohybrid as a promising and sustainable solution for advanced water purification, meeting the urgent demand for multifunctional materials in hazardous contaminant management.

Introduction

The increasing release of synthetic dyes and pathogenic microorganisms into aquatic ecosystems is among the most urgent environmental and public health issues of the 21st century^1–3. Among synthetic dyes, malachite green (MG), a cationic triphenylmethane dye, is widely used in the textile, paper, and aquaculture industries because of its vibrant color and low cost^4,5. Despite its industrial usefulness, MG is classified as a persistent organic pollutant (POP) and is known for its high toxicity and potential carcinogenicity. Even at trace levels (0.1–10 mg L⁻¹), it has been shown to cause mutagenic, hepatotoxic, and endocrine-disrupting effects in both aquatic life and humans⁶. Moreover, conventional wastewater treatment technologies often struggle to effectively degrade or remove such persistent dyes, leading to their accumulation in surface and groundwater resources⁷. At the same time, the presence of pathogenic microorganisms, such as bacteria (e.g., Escherichia coli and Staphylococcus aureus) and fungi (e.g., Candida albicans), in untreated or partially treated wastewater poses a serious health risk, especially in areas lacking adequate sanitation infrastructure⁸. The rising prevalence of multidrug-resistant (MDR) pathogens, including fluconazole-resistant Candida species and carbapenem-resistant Acinetobacter baumannii, exacerbates the microbial contamination problem, highlighting the need for innovative, non-antibiotic disinfection methods for effective wastewater treatment⁹.

In response to these dual threats, there is a growing demand for multifunctional materials capable of simultaneously removing organic contaminants and inactivating pathogenic microorganisms in a single treatment step^10,11. Adsorption is widely recognized as one of the most efficient, cost-effective, and scalable methods for dye removal, provided that the adsorbent possesses a high surface area, numerous active sites, rapid adsorption kinetics, and excellent reusability^12,13. However, conventional adsorbents, such as activated carbon, clay, and biochar, typically lack intrinsic antimicrobial activity, necessitating a separate disinfection stage (e.g., chlorination or UV irradiation), which adds to process complexity and operational costs^14,15. Conversely, many antimicrobial agents (e.g., silver nanoparticles, quaternary ammonium compounds) exhibit limited efficacy against organic dyes and are often associated with high toxicity, poor chemical stability, or rapid deactivation in complex aqueous environments^14,16. This technological limitation has spurred intense research into engineered nanohybrids that integrate both adsorptive and antimicrobial functionalities within a single platform^17,18. Recent advances in metal–oxide/carbon hybrid materials have demonstrated enhanced adsorption and sensing performance, further supporting the development of multifunctional nanohybrids¹⁹.

Among emerging nanomaterials, graphene oxide (GO) has attracted significant attention as an ideal adsorbent platform due to its two-dimensional structure, high surface area, and abundant oxygen-containing functional groups^20,21. The basal planes and edges of GO are functionalized with groups such as carboxyl (–COOH), hydroxyl (–OH), and epoxide (C–O–C), which facilitate strong electrostatic interactions, hydrogen bonding, and π–π stacking with cationic dyes like MG²². However, pristine GO tends to aggregate in aqueous solutions, lacks mechanical stability, and lacks significant inherent antimicrobial properties²³. In contrast, zinc oxide (ZnO) is a well-known semiconductor material with potent antimicrobial activity²⁴. Under light exposure—or even under ambient conditions—ZnO can generate reactive oxygen species (ROS), including hydroxyl radicals (·OH), superoxide anions (·O₂⁻), and hydrogen peroxide (H₂O₂), which cause oxidative stress, lipid peroxidation, protein denaturation, and DNA damage in microbial cells. Despite its antimicrobial efficacy, ZnO nanoparticles are prone to agglomeration, exhibit low adsorption affinity for organic dyes, and may release Zn²⁺ ions under acidic conditions, raising environmental concerns^25,26.

To overcome these individual limitations, researchers have increasingly turned to graphene oxide–zinc oxide (GO–ZnO) nanohybrids, in which ZnO nanostructures are anchored onto GO sheets through chemical bonding or physical interactions²⁷. This hybrid approach produces a composite material with synergistic functionalities: GO provides a high-surface-area, well-dispersed platform that suppresses ZnO aggregation, enhances dye adsorption through its surface functional groups, and stabilizes ZnO against deactivation or leaching^27,28. Concurrently, ZnO contributes potent antimicrobial activity while structurally reinforcing the GO matrix and mitigating restacking. Importantly, the oxygen-containing groups on GO, particularly carboxyl (–COOH) and hydroxyl (–OH), serve as nucleation sites for Zn²⁺ ions during synthesis, promoting strong interfacial bonding and uniform deposition of ZnO nanostructures, often in the form of nanorods or nanoparticles, as confirmed by FTIR, XRD, and TEM analyses^28–30.

The antimicrobial mechanism of GO–ZnO nanohybrids is governed by a dual-action mode. First, ZnO-mediated ROS generation disrupts microbial redox balance, resulting in oxidative stress and irreversible cellular damage. Second, the sharp nanoscale edges of GO physically pierce and fragment microbial membranes, a phenomenon often referred to as the “nano-knife” effect. This combined mechanical and oxidative assault significantly enhances antimicrobial efficacy against both Gram-positive and Gram-negative bacteria, as well as fungal pathogens such as Candida species. Importantly, this broad-spectrum biocidal mechanism reduces the likelihood of resistance development, offering a promising alternative to conventional antibiotics and antifungals in the context of antimicrobial resistance (AMR)^31–33.

Despite these promising attributes, most existing studies on GO–ZnO nanohybrids have focused exclusively on either dye adsorption or antimicrobial activity, often relying on reference microbial strains and idealized laboratory conditions^33–35. Comprehensive assessments against clinically relevant, drug-resistant pathogens (e.g., fluconazole-resistant Candida species) remain limited, and few investigations have explored the material’s efficacy in realistic wastewater environments containing competing ions, organic matter, and suspended particulates. Furthermore, systematically optimizing operational parameters using statistical tools like response surface methodology (RSM), alongside detailed mechanistic studies such as isotherm modeling, kinetic and thermodynamic analysis, and long-term recyclability testing—an area that is still relatively underexplored—has been shown to improve the performance of similar hybrid carbon-based composites through optimized nanoscale engineering. Additionally, recent advancements in GO-based functional systems further underscore the versatility of graphene oxide as a platform for catalytic and environmental applications^36,37.

This work addresses these critical gaps by presenting a multifunctional GO–ZnO nanohybrid engineered for integrated water purification. We systematically investigate its dual capability to (1) remove MG with high efficiency under statistically optimized conditions and (2) inactivate a panel of reference and clinical bacterial and fungal strains, including multidrug-resistant (MDR) isolates, through synergistic mechanisms involving ROS generation and membrane disruption. Using RSM, we identify the optimal operational parameters for maximum dye removal and validate the adsorption mechanism through the Langmuir isotherm model, pseudo-second-order kinetics, and thermodynamic analysis. The material’s practical viability is further demonstrated through seven reuse cycles with minimal solvent use while maintaining high performance in real water samples (tap, river, and industrial wastewater). By bridging materials science, environmental engineering, and microbiology, this study offers a robust, scalable, and sustainable solution for next-generation water treatment systems that require simultaneous pollutant removal and disinfection.

Materials and methods

Materials and reagents

Graphite powder (≥ 99%), zinc acetate dihydrate (Zn(CH₃COO)₂·2H₂O), potassium permanganate (KMnO₄), sulfuric acid (H₂SO₄, 98%), hydrogen peroxide (H₂O₂, 30%), sodium nitrate (NaNO₃), hydrochloric acid (HCl, 37%), and ethanol (≥ 99.8%) were procured from Merck (Germany). Malachite Green (MG, C₂₃H₂₅ClN₂, ≥ 95%) was purchased from Sigma-Aldrich (USA). All reagents were of analytical grade and used without further purification. Deionized (DI) water with a resistivity of ≥ 18.2 MΩ·cm was used in all experiments.

Materials characterization

The structural and morphological properties of the synthesized materials were characterized using multiple analytical techniques. Fourier transform infrared spectroscopy (FTIR) was conducted on a Bruker Tensor II spectrometer (Germany) in the wavenumber range of 4000–400 cm⁻¹, using KBr pellets. X-ray diffraction (XRD) patterns were obtained using a Philips PW1730 diffractometer (Netherlands) with Cu K_α radiation (λ = 1.5406 Å) over a 2θ range of 5–80° at a scan rate of 0.02°/s. Transmission electron microscopy (TEM) images were captured using a Zeiss EM900 microscope (Germany) operating at 80 kV; samples were prepared by drop-casting diluted dispersions onto carbon-coated copper grids. Atomic force microscopy (AFM) topography and height profiles were measured in tapping mode using a NanoSurf FlexAFM system (Switzerland) with silicon cantilevers having a spring constant of ~ 40 N m⁻¹.

Preparation of GO–ZnO nanohybrid

Synthesis of GO nanosheets

GO nanosheets were synthesized using a modified Hummers method³⁸. Briefly, 3 g of graphite was mixed with 360 mL of concentrated H₂SO₄ and 40 mL of H₃PO₄ while stirring. To control the highly exothermic oxidation reaction, the flask was placed in an ice bath before the slow addition of 18 g of KMnO₄. The mixture was stirred magnetically for 24 h. Subsequently, 400 mL of DI water was added dropwise, followed by 15 mL of 30% H₂O₂, which caused a color change from dark green to golden-brown. The mixture was then centrifuged at 4500 rpm for 15 min. The precipitate was washed three times with 5% HCl and then dialyzed until a neutral pH was achieved. Finally, the GO suspension was ultrasonicated for 2 h to exfoliate the layers and obtain GO nanosheets.

Synthesis of ZnO nanoparticles

ZnO nanoparticles were synthesized using a co-precipitation method^39,40. A 0.1 mol L⁻¹ solution of zinc acetate (60 mL) was ultrasonicated and stirred magnetically. A second solution, containing 0.1 mol L⁻¹ NaOH (50 mL) in DI water (50 mL), was added dropwise to the zinc solution while maintaining continuous mixing. After 30 min at room temperature, a white precipitate formed, which was subsequently washed several times with distilled water and then freeze-dried for 48 h.

Preparation of GO–ZnO nanohybrid

To prepare the nanohybrid with a 2:1 weight ratio (wt/wt%) of ZnO to GO, 36 mL of a 0.1 mol L⁻¹ zinc acetate solution was combined with 22 mL of a GO suspension (2.556 mg mL⁻¹) and then subjected to ultrasonication. Simultaneously, 50 mL of 0.1 mol L⁻¹ NaOH (hereafter referred to as Solution 1) was added dropwise to the reaction mixture under continuous stirring. A brownish suspension, in which ZnO nanoparticles were deposited on GO nanosheets, was formed after 30 min. The suspension was washed with distilled water and then freeze-dried for 48 h.

Microbial strains and clinical isolates

The antimicrobial activity of the GO–ZnO nanohybrid was evaluated using both reference strains and clinically derived isolates. Reference microbial strains, including Escherichia coli (E. coli, ATCC 25922), Staphylococcus aureus (S. aureus, ATCC 25923), Acinetobacter baumannii (A. baumannii, ATCC 19606), and Candida albicans (C. albicans, ATCC 10231), were sourced from the Iranian National Center for Research in Medical Sciences. Clinical isolates of E. coli, S. aureus, A. baumannii, and fluconazole-resistant C. albicans were previously collected from clinical specimens (blood, urine, and vaginal swabs) in ethically approved studies and stored in sterile distilled water at 4 °C until further use^8,41.

Preparation of microbial suspensions

For antifungal assays, clinical isolates of C. albicans were subcultured onto Sabouraud Dextrose Agar (SDA) and incubated at 35 °C for 24 h to confirm colony purity and viability. A standardized inoculum was prepared by transferring a small portion of a fresh colony (less than half of a calibrated 10 µL loop) into a sterile Falcon tube containing sterile distilled water, under aseptic conditions near a Bunsen burner flame. The suspension was vortexed vigorously and adjusted spectrophotometrically to an optical density (OD) of 0.08–0.1 at 530 nm, corresponding to a yeast cell density of ~ 1–5 × 10⁶ CFU/mL for the C. albicans isolates. This primary suspension was then diluted 1:1000 in Roswell Park Memorial Institute (RPMI) 1640 medium, supplemented with L-glutamine, devoid of sodium bicarbonate, and buffered to pH 7.0 with 3-(N-morpholino)propanesulfonic acid (MOPS), to achieve the final working inoculum⁴¹.

Antimicrobial susceptibility testing

Minimum inhibitory concentration (MIC) assays were performed in accordance with the Clinical and Laboratory Standards Institute (CLSI) guidelines, specifically M27-A3 for yeasts and M07-A10 for bacteria. For bacterial strains, the broth microdilution method was employed using Mueller–Hinton broth (MHB) in sterile, 96-well flat-bottom microtiter plates. Serial two-fold dilutions of the GO–ZnO nanohybrid were prepared in MHB, and 100 µL of each dilution was dispensed into individual wells. Subsequently, 100 µL of a standardized bacterial suspension (adjusted to ~ 5 × 10⁵ CFU/mL) was added to each well. The plates were incubated at 37 °C for 18–24 h. Three wells served as negative controls (medium only, no inoculum), and three as positive growth controls (inoculum without the nanohybrid). Following incubation, microbial growth was quantified by measuring optical density (OD) at 630 nm using a microplate reader (i.e., an ELISA reader)⁴¹.

For antifungal testing, the broth microdilution protocol was conducted using RPMI 1640 medium (with L-glutamine, without sodium bicarbonate, buffered with MOPS, pH 7.0), in accordance with CLSI M27-A3 guidelines. Due to the poor aqueous solubility of the GO–ZnO nanohybrid, a stock suspension (10,000 µg mL⁻¹) was first prepared by dispersing 10 mg of the nanohybrid in 1 mL of dimethyl sulfoxide (DMSO) with vortexing and brief sonication to ensure homogeneity. Similarly, a stock solution of fluconazole (1600 µg mL⁻¹) was prepared by dissolving 1.6 mg of the drug in 1 mL of DMSO. Serial two-fold dilutions of the nanohybrid, alone or in combination with fluconazole, were prepared in RPMI medium. Subsequently, 100 µL of each drug dilution was dispensed into sterile 96-well flat-bottom microtiter plates, followed by 100 µL of the standardized Candida suspension (adjusted to ~ 0.5–2.5 × 10³ CFU mL⁻¹, as recommended by CLSI). The plates were incubated at 35 °C for 48 h. The MIC was defined as the lowest concentration of the nanohybrid that visually inhibited fungal growth, in accordance with CLSI M27-A3 interpretive criteria⁴¹. All antimicrobial assays were performed in triplicate to ensure reproducibility and minimize experimental error. Certified reference strains were included in each assay as internal quality controls. The final DMSO concentration in all test wells did not exceed 1% (v/v), a level that was verified in preliminary tests to have no inhibitory effect on fungal growth.

Batch adsorption experiments

The adsorption performance of the GO–ZnO nanohybrid as an adsorbent was investigated using a batch technique for the removal of MG. A total of 32 experiments were conducted via RSM using a CCD. As shown in Table 1, the effects of initial pH (3.0–11), initial MG concentration (6–30 mg L⁻¹), adsorbent dosage (6–30 mg), temperature (283.15–323.15 K), and contact time (3–15 min) on the adsorption of MG by the GO–ZnO nanohybrid were systematically studied through batch adsorption experiments. A precisely weighed amount of GO–ZnO nanohybrid was added to 50 mL of MG solution (25 mg L⁻¹) in a 100 mL conical flask. The pH of the MG solution was adjusted using 0.1 mol L⁻¹ HCl or NaOH. The flask was then placed in an ultrasonic bath shaker at the preset temperature for the designated reaction time. Upon completion of adsorption, the adsorbent was separated from the MG solution by centrifugation. The residual concentration of MG in the supernatant after the adsorption process was quantified using a UV–Vis spectrophotometer (DR6000, HACH, USA) at a λ_max of 617 nm, and the equilibrium concentration of MG was calculated using the equation from our previous study⁴². The performance of the GO–ZnO nanohybrid was expressed in terms of MG removal percentage (% removal) and equilibrium adsorption capacity (q_e, mg g⁻¹)⁴³. All experiments were carried out in triplicate to ensure reproducibility.

Table 1.

Independent variables and their corresponding levels.

Independent variables	Unit	Levels (α = 2)
		− α	Low (− 1)	Center (0)	High (+ 1)	+α
(X1) MG concentration	mg L−1	6.0	12	18	24	30
(X2) pH	–	3.0	5.0	7.0	9.0	11
(X3) Nanohybrid adsorbent	Mg	6.0	12	18	24	30
(X4) Temperature	°C	10	20	30	40	50
(X5) Contact time	min	3.0	6.0	9.0	12	15

Experimental design using RSM

The removal efficiency of MG by the GO–ZnO nanohybrid was found to be governed by five critical process variables: X₁ (initial MG concentration, mg L⁻¹), X₂ (solution pH), X₃ (adsorbent dosage, mg), X₄ (temperature, °C), and X⁵ (contact time, min). Evaluating the interactive effects among these parameters via full factorial experimentation would require 5⁵ = 3125 individual runs, an impractical and resource-intensive approach that generates excessive chemical waste and dramatically increases experimental cost. To overcome this limitation while preserving statistical rigor, RSM based on a CCD was employed to construct an optimized experimental matrix requiring only 32 strategically selected runs (Table 2).

Table 2.

Experimental design matrix showing the actual and predicted MG removal (%) by the GO–ZnO nanohybrid.

Std	Run	Block	Independent variables					R% MG
			X1	X2	X3	X4	X5	Actual value	Predicted value
1	13	Block 1	12	5.0	12	20	12	56.45	55.75
2	7	Block 1	24	5.0	12	20	6	6.29	4.645
3	12	Block 1	12	9.0	12	20	6	63.45	63.16
4	8	Block 1	24	9.0	12	20	12	85.23	85.21
5	4	Block 1	12	5.0	24	20	6	47.45	46.99
6	5	Block 1	24	5.0	24	20	12	67.56	67.37
7	2	Block 1	12	9.0	24	20	12	95.45	96.61
8	16	Block 1	24	9.0	24	20	6	87.56	87.77
9	10	Block 1	12	5.0	12	40	6	67.56	66.95
10	18	Block 1	24	5.0	12	40	12	42.23	41.89
11	20	Block 1	12	9.0	12	40	12	92.57	93.58
12	6	Block 1	24	9.0	12	40	6	65.34	65.41
13	25	Block 1	12	5.0	24	40	12	58.34	59.18
14	3	Block 1	24	5.0	24	40	6	73.23	73.13
15	30	Block 1	12	9.0	24	40	6	89.36	90.61
16	23	Block 1	24	9.0	24	40	12	79.02	80.54
17	22	Block 1	6.0	7.0	18	30	9	96.54	95.65
18	15	Block 1	30	7.0	18	30	9	78.48	78.94
19	32	Block 1	18	3.0	18	30	9	29.00	30.82
20	26	Block 1	18	11	18	30	9	94.81	92.57
21	31	Block 1	18	7.0	6.0	30	9	43.56	45.04
22	29	Block 1	18	7.0	30	30	9	78.34	76.44
23	19	Block 1	18	7.0	18	10	9	64.50	65.69
24	11	Block 1	18	7.0	18	50	9	83.26	81.64
25	1	Block 1	18	7.0	18	30	3	58.91	59.91
26	27	Block 1	18	7.0	18	30	15	81.70	80.27
27	17	Block 1	18	7.0	18	30	9	82.23	83.78
28	9	Block 1	18	7.0	18	30	9	82.98	83.78
29	24	Block 1	18	7.0	18	30	9	85.23	83.78
30	28	Block 1	18	7.0	18	30	9	83.45	83.78
31	21	Block 1	18	7.0	18	30	9	84.71	83.78
32	14	Block 1	18	7.0	18	30	9	83.68	83.78

All experiments were designed and analyzed using the trial version of Design-Expert^® 10.0.1 software (Stat-Ease, Inc.). The collected data were subjected to sequential sum of squares analysis and lack-of-fit testing to select the most statistically adequate model⁴⁴. Analysis of variance (ANOVA) tables were generated for candidate models, and non-significant terms (p > 0.05) were systematically eliminated based on their F-values. Model significance was evaluated at the 95% confidence level (α = 0.05). The model’s goodness-of-fit was assessed using the coefficient of determination (R²) and adjusted R² values, thereby ensuring robustness and predictive capability⁴⁵.

Results and discussion

Structural and morphological characterization

Figure 1 presents the FTIR spectra of (a) GO nanosheets, (b) ZnO nanoparticles, and (c) the GO–ZnO nanohybrid, recorded over the wavenumber range of 4000–400 cm⁻¹. These spectra provide critical insights into the functional groups present on the surface of each material and confirm successful hybridization through characteristic peak shifts, intensity changes, and peak attenuation^46,47.

Fig. 1.

FTIR spectra of a GO, b ZnO, and c the GO–ZnO nanohybrid.

In the spectrum of pristine GO (Fig. 1a), several characteristic absorption bands are observed. The band at 1732 cm⁻¹ corresponds to the stretching vibration of carbonyl groups (C=O), while the 1627 cm⁻¹ band is attributed to aromatic C=C skeletal vibrations. A weaker band at 1400 cm⁻¹ results from C–OH bending, and the 1090 cm⁻¹ band arises from epoxy C–O–C stretching vibrations⁴⁶. The broad and intense absorption near 3400 cm⁻¹ stems from O–H stretching vibrations of hydroxyl groups, carboxyl groups, and adsorbed water molecules, indicating the oxygen-rich nature of GO^46,47. The high presence of oxygenated functional groups not only enhances hydrophilicity but also provides nucleation sites for Zn²⁺ anchoring and active sites for dye adsorption and antimicrobial activity. This aligns with theoretical studies demonstrating that pollutant adsorption on 2D surfaces is strongly affected by physisorption and surface chemistry⁴⁸.

For pristine ZnO nanoparticles (Fig. 1b), a sharp peak at 530 cm⁻¹ is assigned to the Zn–O stretching mode of the ZnO lattice. Additional bands include one around 1567 cm⁻¹, attributed to adsorbed water or surface hydroxyl bending, and a broad feature near 3400 cm⁻¹, also due to surface hydroxyl groups. A weak shoulder close to 1400 cm⁻¹ can be assigned to carbonate species (CO₃²⁻), resulting from atmospheric CO₂ adsorption^46,47. The presence of hydroxyl and carbonate groups not only confirms surface reactivity but also provides anchoring sites for dye molecules. These ZnO sites enhance MG adsorption by providing additional surface interaction points and contribute to antibacterial and antifungal activity via ROS generation.

In the FTIR spectrum of the GO–ZnO nanohybrid (Fig. 1c), several changes confirm successful chemical interaction between GO and ZnO. Most notably, the carbonyl (C=O) peak at 1732 cm⁻¹, prominent in pristine GO, is significantly reduced or absent, indicating that carboxylic acid groups of GO participated in coordination/complexation with Zn²⁺ ions during nanohybrid formation. Simultaneously, the Zn–O stretching band appears around 470–530 cm⁻¹, confirming the integration of ZnO crystallites with the GO framework. Slight shifts and broadening of Zn–O vibrations further suggest strong interfacial coupling between ZnO and oxygenated groups on GO.

Overall, FTIR analysis demonstrates that the GO–ZnO nanohybrid consists of an interconnected network of oxygenated GO nanosheets and ZnO nanocrystals, in which oxygen functionalities act as anchoring sites for ZnO. This hybrid architecture provides abundant active sites for dye adsorption via hydrogen bonding, electrostatic interactions, and π–π stacking, while also imparting enhanced antibacterial and antifungal activities through ROS generation and synergistic interactions between GO and ZnO.

Figure 2 displays the XRD patterns of (a) GO and (b) the GO–ZnO nanohybrid, recorded over a 2θ range of 5°-80°. These profiles provide information regarding crystallinity, phase purity, and structural changes upon hybridization.

Fig. 2.

XRD patterns of a GO and b the GO–ZnO nanohybrid.

In the XRD pattern of pristine GO (Fig. 2a), a broad diffraction peak centered at ~ 10.5° (assigned to the (001) plane) is observed, indicative of an expanded interlayer spacing of GO sheets due to oxygen-containing functional groups (epoxides, hydroxyls, carboxyls). The broadness and low intensity of this peak indicate the largely amorphous or poorly crystalline nature of GO, consistent with extensive oxidation and structural disorder introduced during the exfoliation process^46,49.

The XRD pattern of the GO–ZnO nanohybrid (Fig. 2b) exhibits multiple intense peaks indexed to the hexagonal wurtzite phase of ZnO (JCPDS No. 36-1451) at 2θ ≈ 31.7°, 34.4°, 36.2°, 47.5°, 56.5°, 62.8°, 67.9°, and 69.0°, corresponding to the (100), (002), (101), (102), (110), (103), (200), and (201) planes, respectively. The high intensity and narrow full width at half maximum (FWHM) values indicate highly crystalline ZnO nanoparticles within the hybrid structure^46,50,51. The well-crystallized ZnO provides additional adsorption sites for MG molecules and contributes to antibacterial and antifungal activity through ROS generation and enhanced interfacial reactivity.

Notably, the characteristic (001) peak of GO at ~ 10.5° is absent, indicating either exfoliation of the GO layers or disruption of their stacked arrangement due to strong interactions with ZnO nanoparticles, possibly via coordination bonding between GO oxygen functionalities and Zn²⁺ ions. No additional phases were detected, confirming high phase purity⁵⁰. The intimate contact between ZnO and GO enhances MG adsorption and creates a heterogeneous surface capable of interacting with bacterial and fungal cells, improving antimicrobial efficacy.

The coexistence of well-defined ZnO peaks with the disappearance of the GO basal reflection confirms the successful integration of ZnO nanocrystals onto the GO matrix, forming a structurally distinct nanohybrid. By combining high surface area, abundant oxygen-containing functional groups, and crystalline ZnO domains, the GO–ZnO nanohybrid exhibits dual functionality: efficient MG adsorption and potent antibacterial and antifungal activity.

Figure 3 presents representative TEM images of pristine GO and the GO–ZnO nanohybrid, highlighting their morphological characteristics and interfacial structure.

Fig. 3.

TEM images of a, b GO and c, d the GO–ZnO nanohybrid.

In Fig. 3a and b, the thin, sheet-like, and flexible morphology of pristine GO nanosheets is clearly evident. The sheets exhibit wrinkled and folded regions, reflecting both the compliance of the carbon lattice and structural distortions introduced by oxygen-containing functional groups. The lighter (bright) regions correspond to thin, single- or few-layer domains, whereas the darker areas indicate thicker, stacked, or overlapping regions. These observations are consistent with the exfoliated structure of GO derived from chemically oxidized graphite⁵².

In contrast, Fig. 3c and d display the morphology of the GO–ZnO nanohybrid. Rod-like ZnO nanocrystals are clearly visible as dark, elongated structures uniformly anchored onto the lighter GO sheets. Their uniform dispersion across the GO surface indicates successful in situ nucleation and growth of ZnO within the GO matrix. The intimate attachment of ZnO nanorods to the GO surface suggests strong interfacial interactions, likely mediated through the coordination of Zn²⁺ ions with oxygenated functional groups (carboxyl, hydroxyl, and epoxy) on GO.

Quantitative analysis of multiple TEM images shows that the ZnO nanorods have an average diameter of 200–250 nm and a length of 500–600 nm. Their dimensions can be adjusted by modulating synthesis parameters, such as the GO-to-precursor ratio, reaction time, and solution pH. This structural tunability enables optimization of the nanohybrid’s surface area and active site density, directly enhancing its adsorption capacity and antimicrobial performance.

The homogeneous distribution and strong interfacial anchoring of ZnO nanorods on GO nanosheets prevent nanoparticle agglomeration, thereby generating abundant and stable active sites. This unique morphology is critical for enhancing dye adsorption through multiple interactions (electrostatic attraction, π–π stacking, and hydrogen bonding) and for improving antimicrobial performance via reactive oxygen species (ROS)-mediated and membrane disruption mechanisms.

Figure 4 shows AFM topographic images and corresponding height profiles of pristine GO and the GO–ZnO nanohybrid, offering quantitative information into surface roughness, sheet thickness, and nanoparticle incorporation at the nanoscale.

Fig. 4.

AFM images and corresponding height profiles of a, b GO and c, d the GO–ZnO nanohybrid.

In Fig. 4a and b, the AFM image of pristine GO reveals ultrathin, flexible sheets with micrometer-scale lateral dimensions and a relatively smooth yet wrinkled surface morphology. The corresponding height profile (Fig. 4b), measured along the white line in Fig. 4a, indicates an average thickness of ~ 1.8–2 nm, characteristic of few-layer GO (typically 2–5 layers). This value is consistent with literature reports for chemically exfoliated GO and confirms successful oxidation and delamination of graphite with minimal restacking.

In contrast, Fig. 4c and d illustrate the surface morphology of the GO–ZnO nanohybrid. Bright protrusions are observed across the GO substrate, corresponding to ZnO nanostructures anchored onto the surface. The height profile (Fig. 4d) across one such protrusion shows a sharp elevation of ~ 45–50 nm, far exceeding the underlying GO sheet thickness (~ 2 nm). This observation confirms that the protrusions are vertically oriented ZnO nanostructures rather than imaging artifacts.

The step-like height transition between GO and ZnO domains indicates strong interfacial adhesion and suggests preferential nucleation of ZnO at defect sites or oxygen-functionalized regions on GO. The root-mean-square (RMS) roughness of the hybrid surface is markedly higher than that of pristine GO, reflecting the increased topographical complexity imparted by ZnO deposition.

Taken together, these AFM results, in agreement with TEM and XRD analyses, confirm that ZnO forms discrete nanostructures rather than as a continuous conformal layer. The vertical orientation of ZnO features increases accessible surface area and may facilitate mass transfer during dye adsorption and microbial contact, thereby contributing to the enhanced multifunctional performance of the nanohybrid.

The nitrogen adsorption–desorption isotherm of the GO–ZnO nanohybrid (Fig. 5) exhibits a Type IV isotherm with an H3 hysteresis loop, characteristic of mesoporous materials with slit-shaped pores formed by the aggregation of plate-like particles. The steep uptake at low relative pressures (P/P₀ < 0.1) indicates strong micropore filling, while the gradual increase at higher pressures (P/P₀ > 0.5) reflects multilayer adsorption on mesoporous surfaces. The calculated Brunauer–Emmett–Teller (BET) specific surface area is 52.39 m² g⁻¹, which is significantly higher than that of pure ZnO nanoparticles^53,54 and comparable to many reported GO-based composites, confirming that ZnO incorporation does not compromise the high surface area of the GO scaffold. Instead, the anchoring of ZnO nanorods onto GO sheets generates additional interfacial voids and further enhances overall porosity^55,56.

Fig. 5.

N₂ adsorption–desorption isotherm (BET) and BJH pore size distribution (inset) of the GO–ZnO nanohybrid.

The Barrett–Joyner–Halenda (BJH) pore size distribution, derived from the desorption branch (inset in Fig. 5), reveals a broad mesopore range centered around ~ 10–50 nm, with a peak near ~ 20 nm. This hierarchical pore structure, comprising micropores (from GO wrinkles) and mesopores (from ZnO-GO interfacial spaces), facilitates rapid diffusion of dye molecules to active sites, contributing to the observed high adsorption kinetics. The presence of a distinct hysteresis loop further indicates the existence of interconnected pore networks, which enhance accessibility and reduce mass transfer resistance during MG adsorption.

Optimization of MG adsorption using CCD-RSM

Statistical analysis and model fitting based on the RSM

A second-order polynomial regression was identified as the most appropriate empirical model to describe the relationship between the independent variables and the response (% MG removal), as validated by sequential model sum-of-squares analysis and the absence of aliasing. The final reduced quadratic equation, expressed in actual factor units, is presented below:

1

Where, X₁, X₂, X₃, X₄, and X₅ denote the initial MG concentration (mg L⁻¹), solution pH, adsorbent dosage (mg), temperature (°C), and contact time (min), respectively. The predicted removal efficiency of MG dye was calculated using the above equation, and the results are summarized in Table 2.

The adequacy and statistical significance of the quadratic regression model were rigorously evaluated through analysis of variance (ANOVA), as presented in Table 3. The model demonstrated a very high level of significance, with an F-value of 184.7 and a corresponding p-value < 0.0001, confirming its reliability in describing the relationship between the independent variables and MG removal efficiency⁵⁷. The coefficient of determination (R² = 0.997) and adjusted R² (0.992) indicate that the model explains nearly 99.7% of the response variability, reflecting excellent fit and predictive capability across the design space. Furthermore, the predicted R² (0.932) demonstrates strong model robustness and generalizability to new experimental conditions⁴².

Table 3.

ANOVA results for MG adsorption onto the GO–ZnO nanohybrid.

Source of variation	SS	DF	MS	F value	P value
Model	13,230	20	661.4	184.7	< 0.0001
X1	419.1	1	419.1	117.1	< 0.0001
X2	5719	1	5719	1598	< 0.0001
X3	1479	1	1479	413.2	< 0.0001
X4	381.9	1	381.9	106.7	< 0.0001
X5	622.1	1	622.1	173.8	< 0.0001
X1 × 2	17.66	1	17.66	4.933	0.04828
X1 × 3	596.7	1	596.7	166.7	< 0.0001
X1 × 4	63.4	1	63.4	17.71	0.0015
X1 × 5	2.747	1	2.747	0.767	0.3998
X2 × 3	53.47	1	53.47	14.94	0.0026
X2 × 4	297.6	1	297.6	83.14	< 0.0001
X2 × 5	17.04	1	17.04	4.759	0.0517
X3 × 4	184.6	1	184.6	51.57	< 0.0001
X3 × 5	315.7	1	315.7	88.18	< 0.0001
X4 × 5	949.7	1	949.7	265.3	< 0.0001
X12	22.61	1	22.61	6.315	0.02883
X22	894.8	1	894.8	249.9	< 0.0001
X32	973.8	1	973.8	272	< 0.0001
X42	187.7	1	187.7	52.43	< 0.0001
X52	343.7	1	343.7	96.01	< 0.0001
Residual	39.38	11	3.58
Lack of Fit	33.28	6	5.546	4.545	0.0590
Pure Error	6.102	5	1.22
Corr. Total	13,270	31
Summary statistics of the quadratic model
Standard deviation (SD%)	1.892	R2	0.997
Mean	71.51	Adjusted-R2	0.992
Coefficient of variation (C.V. %)	2.646	Predicted-R2	0.932
PRESS	898.6	Adequate precision	60.00

All five main factors, initial MG concentration (X₁), pH (X₂), adsorbent dosage (X₃), temperature (X₄), and contact time (X₅), were highly significant (p < 0.0001), with pH (X₂) contributing most strongly (SS = 5719, F = 1598). This dominance of pH reflects its dual role in controlling both the surface charge of GO and the electrostatic interactions with cationic MG molecules. Among the interaction terms, X₁ × ₃ (MG concentration × adsorbent dosage), X₂X₄ (pH × temperature), and X₄ × ₅ (temperature × contact time) exerted particularly strong influences (all p < 0.0001), indicating that synergistic and/or antagonistic effects among these variables significantly affect adsorption efficiency. Quadratic effects (X₁² to X₅²) were also statistically significant (p < 0.05 for all, except X₁² at p = 0.0288), demonstrating curvature in the response surface and justifying the selection of a second-order polynomial model to capture nonlinear behavior⁵⁸.

The lack-of-fit test yielded a p-value of 0.0590 (> 0.05), indicating no significant deviation between the model and the experimental data. This confirms that the model sufficiently accounts for experimental variability without systematic deviations. Additional validation metrics, including the low coefficient of variation (C.V. = 2.646%) and the high adequate precision value (60.00), further confirm an excellent signal-to-noise ratio and robust model reliability across the design space^59,60. Collectively, these results validate the predictive power of the RSM-CCD model for optimizing MG adsorption by the GO–ZnO nanohybrid.

The Pareto chart of standardized effects (Fig. 6) provides a visual ranking of variable importance, with bars extending beyond the 95% confidence threshold (p = 0.05) indicating statistically significant effects. As shown, pH (X₂, linear term) exerts the strongest positive effect on dye removal, followed by adsorbent dosage (X₃, linear term) and its quadratic effect (X₃²), emphasizing the combined importance of solution chemistry and adsorbent availability. This behavior reflects enhanced electrostatic interactions between deprotonated GO functional groups and cationic MG molecules at higher pH, as well as the greater availability of active binding sites at higher adsorbent dosages.

Fig. 6.

Pareto chart of standardized effects for the response variable R% MG, showing the relative magnitude and statistical significance (p = 0.05) of main effects, quadratic terms, and two-factor interactions. Bars extending beyond the red line indicate significant contributions to dye removal efficiency.

Other important contributors include contact time (X₅, linear) and pH (X₂², quadratic), indicating that removal efficiency increases with contact duration but eventually approaches equilibrium, while nonlinear pH effects reflect complex protonation–deprotonation dynamics. MG concentration (X₁, linear) was moderately significant, as higher initial dye loading decreases the percentage removal due to the fixed adsorbent capacity, yet remains a critical operational parameter.

Notably, several interaction terms, such as X₄ × ₅ (temperature × contact time) and X₁ × ₃ (concentration × dosage), exhibited substantial effects, underscoring the multidimensional nature of the adsorption process and the necessity of considering these variable interactions when optimizing conditions. In contrast, some higher-order interactions and quadratic terms for MG concentration and temperature were not statistically significant, suggesting limited curvature or interaction effects in those regions of the design space.

Overall, this analysis validates the robustness of the quadratic regression model and highlights pH and adsorbent dosage as the most influential parameters. These findings provide actionable guidance for process optimization and scale-up, demonstrating that careful adjustment of pH and adsorbent dosage is essential for maximizing MG removal efficiency in practical applications using the GO–ZnO nanohybrid.

Interaction of variables and model validation

The three-dimensional response surface plots in Fig. 7 illustrate the interactive effects of key operational parameters on MG removal efficiency by the GO–ZnO nanohybrid. As shown in Fig. 7a, the combined influence of pH and adsorbent dosage reveals a distinct maximum in removal efficiency at alkaline pH values (~ 9.0–11) and moderate-to-high adsorbent dosages (4–8 mg). This trend aligns with the cationic character of MG, whose protonated tertiary amine groups carry strong positive charges, and the negatively charged oxygen-containing functionalities on GO (e.g., –COO⁻, –O⁻), which become increasingly deprotonated at elevated pH, thereby enhancing electrostatic attraction⁶¹. Under acidic conditions (pH < 5), protonation of these groups reduces the negative charge density of GO, weakening electrostatic interactions and decreasing adsorption efficiency. The quadratic curvature observed further indicates that, beyond the optimal pH (~ 10), excessive alkalinity may alter surface speciation or introduce repulsive interactions, slightly diminishing dye uptake^62,63.

Fig. 7.

Three-dimensional response surface plots showing the interactive effects of a pH and nanohybrid adsorbent dosage, b pH and contact time, c adsorbent dosage and contact time, and d adsorbent dosage and temperature on MG removal efficiency (R% MG). Color gradients represent response magnitude, and red circles indicate actual experimental data points.

Figure 7b highlights the interaction between pH and contact time. While extended contact times generally enhance adsorption by facilitating diffusion and equilibrium attainment, this effect is most pronounced under alkaline conditions⁶⁴. The pronounced curvature observed in the response surface, therefore, reflects a synergistic interplay between pH-dependent surface chemistry and time-dependent diffusion processes: while contact time governs the rate at which equilibrium is reached, pH dictates both the strength and number of effective adsorption sites²⁷. This dual role of pH underscores its central importance in adsorption process design, as it simultaneously controls equilibrium capacity and adsorption kinetics, thereby exerting a dominant influence on overall process efficiency⁶⁵.

In contrast, Fig. 7c and d demonstrate synergistic relationships between adsorbent dosage, contact time, and temperature. Increasing either dosage or contact time improves removal efficiency, but the combined effect eventually plateaus due to saturation of available binding sites (Fig. 7c). Temperature and adsorbent dosage exhibit a mild positive correlation (Fig. 7d), suggesting that elevated temperatures enhance molecular mobility and reduce solution viscosity, thereby facilitating faster diffusion of MG molecules to adsorption sites and accelerating adsorption kinetics. Despite this, the equilibrium adsorption capacity is largely independent of temperature, implying that temperature primarily influences the rate of adsorption rather than the ultimate uptake. Additionally, higher temperatures may slightly enhance the interaction energy between MG molecules and functional groups on GO, contributing to more rapid achievement of equilibrium²⁷.

Taken together, these response surface plots confirm that pH is the dominant factor influencing MG removal, as further supported by ANOVA and Pareto analysis. Optimization of pH, in combination with adsorbent dosage, is therefore critical to achieving optimal dye removal. The ability of the quadratic model to capture these nonlinear and interactive effects demonstrates the robustness of RSM-CCD approach for designing efficient and scalable adsorption processes for cationic dyes such as MG.

Process optimization

The desirability function-based optimization profile shown in Fig. 8 defines the optimal set of experimental conditions for maximizing MG removal efficiency while accounting for practical operational constraints. The model predicts a maximum removal efficiency of 98.97% under the following conditions: initial MG concentration of 25 mg L⁻¹, solution pH of 9.0, GO–ZnO nanohybrid adsorbent dosage of 18 mg, temperature of 25 °C, and contact time of 12 min. These parameter values, marked by the vertical red dashed lines in each subplot, correspond to the points where the individual desirability functions (green curves) reach their maximum (d = 1.0). At this intersection, the overall desirability index also attains 1.0, indicating complete alignment with the optimization criteria.

Fig. 8.

Profiles for predicted values and desirability showing the optimization of MG removal efficiency (R% MG). Vertical red dashed lines mark the optimal conditions: MG concentration = 25.0 mg L⁻¹, pH = 9.0, adsorbent dosage = 18.0 mg, temperature = 25.0 °C, contact time = 12.0 min, yielding a predicted removal efficiency of 98.97% and an overall desirability of 1.0.

Importantly, the optimization highlights the critical role of pH = 9.0, which enhances electrostatic attraction between cationic MG molecules and the negatively charged surface of the nanohybrid, consistent with earlier mechanistic discussions. This observation supports the hypothesis that surface charge interactions dominate the adsorption mechanism under alkaline conditions, where deprotonation of the functional groups on the adsorbent significantly improves dye binding affinity. The chosen adsorbent dosage of 18 mg achieves an optimal balance, maximizing dye removal while minimizing material consumption, as higher amounts offer little additional benefit once the available surface sites are saturated. The optimal temperature (25 °C) confirms that the adsorption process proceeds efficiently under ambient conditions, removing the need for energy-intensive thermal input, which is advantageous for large-scale wastewater treatment. Likewise, the short equilibrium time (12 min) reflects rapid adsorption kinetics, likely facilitated by the high surface area and accessible active sites of the GO–ZnO nanohybrid.

Adsorption isotherms, kinetic, and thermodynamic models

The adsorption behavior of MG onto the GO–ZnO nanohybrid was systematically examined through equilibrium isotherm, kinetic, and thermodynamic modeling to clarify the governing mechanisms and energetics. Among the tested isotherm models (Table 4), the Langmuir model provided the best fit, with an excellent coefficient of determination (R² = 0.992) and a maximum monolayer adsorption capacity (Q_max = 131.91 mg g⁻¹). This result indicates that MG molecules are adsorbed onto a homogeneous surface with energetically equivalent active sites, forming a uniform monolayer without significant adsorbate–adsorbate interactions. Comparable monolayer adsorption behavior has also been reported for newly engineered inorganic adsorbents⁶⁶. The dimensionless separation factor (R_L = 0.005–0.049, < 1) further confirms the highly favorable nature of adsorption across all tested concentrations. In contrast, the Freundlich (R² = 0.976) and Temkin (R² = 0.878) models exhibited weaker correlations, suggesting only limited surface heterogeneity or variations in the heat of adsorption, which is consistent with electrostatic attraction and π–π stacking interactions as the dominant binding mechanisms rather than multilayer or heterogeneous adsorption⁶⁷.

Table 4.

Parameters of isothermal adsorption models for MG adsorption onto the GO–ZnO nanohybrid.

Isotherm	Plot	Parameters	Values
Langmuir	Ce/qe vs. Ce	Qm (mg g−1)	131.91
		KL (L mg−1)	3.906
		R2	0.992
		RL=1/(1+(KL×C0))	0.005–0.049
Freundlich	ln qe vs. ln Ce	1/n	0.314
		KF (L mg−1)	7.055
		R2	0.976
Temkin	qe vs. ln Ce	B1	16.27
		KT (L mg−1)	1.00
		R2	0.878
Dubinin-Radushkevich	ln qe vs. ε2	Qs (mg g−1)	84.62
		β	− 8.0E−09
		E (kJ mol−1)	7.882
		R2	0.772

Kinetic analysis (Table 5) demonstrated that the pseudo-second-order model best described the adsorption process (R² = 0.997), with a calculated equilibrium capacity (q_e(calc) = 78.61 mg g⁻¹) closely matching the experimental value (q_e(exp) = 69.19 mg g⁻¹). This strong agreement implies that chemisorption involving valence forces, electron sharing, or exchange between MG and surface functional groups is the rate-limiting step. It also indicates a strong affinity between MG molecules and the GO–ZnO surface, supporting the presence of chemical bonding interactions in addition to electrostatic attraction⁶⁸. The intraparticle diffusion model (R² = 0.946) also exhibited good correlation, indicating that pore diffusion contributes to mass transfer, though the non-zero intercept (C = 28.24 mg g⁻¹) suggests that boundary layer resistance plays a significant role. These observations demonstrate that the adsorption process is governed by multiple mechanisms, with surface reaction being the dominant process, while diffusion resistance cannot be neglected in practical applications. The Elovich model (R² = 0.9730) further supports adsorption onto a heterogeneous surface, although its lower fit compared to the pseudo-second-order model reinforces that chemical interactions are the primary kinetic mechanism⁶⁹.

Table 5.

Kinetic model parameters for MG adsorption.

Model	Plot	Parameters	Values
First-order-kinetic	ln (q e − q t ) vs. t	k1 (min−1)	0.424
		qe (calc) (mg g−1)	184.0
		R2	0.869
Pseudo-second-order-kinetic	t/qt vs. t	k2 (min−1)	0.005
		qe (calc) (mg g−1)	78.61
		R2	0.997
		h (mg g−1 min−1)	31.82
Intraparticle diffusion	qt vs. t1/2	Kdiff (mg g−1 min−1/2)	10.08
		C (mg g−1)	28.24
		R2	0.946
Elovich	qt vs. ln t	β (g mg−1)	115.2
		α (mg g−1 min−1)	0.072
		R2	0.973
Experimental data		qe (exp) (mg g−1)	69.19

Thermodynamic analysis (Table 6) revealed that the adsorption of malachite green (MG) onto the GO–ZnO nanohybrid is both spontaneous and endothermic. The negative values of Gibbs free energy (ΔG° = − 4.55 to − 10.13 kJ mol⁻¹) confirm the spontaneity of the process across the investigated temperature range (283.15–323.15 K), with adsorption becoming increasingly favorable at higher temperatures (Fig. 9). The positive enthalpy change (ΔH° = +35.48 kJ mol⁻¹) indicates an endothermic nature, likely due to the energy required for desolvation of MG cations and/or dehydration of surface-active sites before adsorption—processes commonly observed in electrostatically driven systems. Notably, the large positive entropy change (ΔS° = +141.46 J mol⁻¹ K⁻¹) suggests a significant increase in system disorder during adsorption. This entropy gain can be mechanistically explained by two complementary contributions: (i) the release of structured water molecules from the hydrated MG cations and the adsorbent surface (desolvation), and (ii) the possible exchange of loosely bound counterions (e.g., Na⁺ or H⁺) on the GO–ZnO surface with MG⁺ ions, which liberates additional water molecules into the bulk solution. Both processes increase the degrees of freedom in the system, thereby driving the adsorption spontaneously despite the endothermic enthalpy. The excellent linearity of the van’t Hoff plot (R² = 0.996) further corroborates the reliability of these thermodynamic parameters^42,68.

Table 6.

Thermodynamic parameters for MG adsorption at different temperatures.

T(k)	kd	ΔG°( kJ mol−1)
283.15	6.91	− 4.55
293.15	11.24	− 5.90
303.15	20.27	− 7.58
313.15	29.96	− 8.85
323.15	43.39	− 10.13
ΔS° (j mol−1 k−1)	141.46
ΔH° (kj mol−1)	35.48
R2	0.996

Fig. 9.

Van’t Hoff plot for the adsorption of MG onto the GO–ZnO nanohybrid.

Overall, these results demonstrate that MG adsorption onto the GO–ZnO nanohybrid is a chemisorption-dominated, spontaneous, endothermic, and entropy-driven process. The combination of high adsorption capacity, rapid kinetics, and favorable thermodynamics positions the GO–ZnO nanohybrid as a promising adsorbent for practical dye removal. The findings are consistent with a mechanism involving electrostatic attraction, π–π stacking, and hydrogen bonding, as further confirmed by FTIR, TEM, and RSM analyses.

Antimicrobial activity: MIC analysis

The antimicrobial activity of the GO–ZnO nanohybrid was systematically evaluated against both standard reference strains and clinically isolated bacteria, as well as against C. albicans, using MIC assays in accordance with CLSI M27-A3 and M07-A10 guidelines (Fig. 10). The nanohybrid demonstrated significantly superior activity compared to its individual components, GO and ZnO, highlighting a clear synergistic effect.

Fig. 10.

MIC values of ZnO, GO, and the GO–ZnO nanohybrid against bacterial and fungal strains. Data are presented as box-and-whisker plots with individual data points (n = 8). Statistical significance was assessed by one-way ANOVA, followed by Tukey’s post hoc test: p < 0.05 (*), p < 0.01 (**), ns = not significant.

Against S. aureus (ATCC), the GO–ZnO nanohybrid displayed an MIC of 1.25 mg mL⁻¹, while GO alone showed no inhibition even at 10 mg mL⁻¹, and ZnO was active at 0.078 mg mL⁻¹. For the clinical isolate of S. aureus, the nanohybrid maintained an MIC of 1.25 mg mL⁻¹, outperforming GO (MIC = 5 mg mL⁻¹) and showing the activity of ZnO (MIC = 0.156 mg mL⁻¹), while offering superior dispersion stability, an important consideration for biomedical applications.

Against E. coli (ATCC), ZnO showed moderate inhibition (MIC = 0.156 mg mL⁻¹), GO remained ineffective (MIC = 10 mg mL⁻¹), and the GO–ZnO nanohybrid achieved complete inhibition at 0.625 mg mL⁻¹. For the clinical isolate of E. coli, the MIC of GO–ZnO was 1.25 mg mL⁻¹, again outperforming GO (10 mg mL⁻¹) and showing activity comparable to ZnO (0.313 mg mL⁻¹).

Notably, for the multidrug-resistant pathogen A. baumannii, both ATCC and clinical isolates required high concentrations of GO and ZnO (MIC = 5 mg mL⁻¹ each). However, the GO–ZnO nanohybrid significantly reduced the MIC to 0.625 mg mL⁻¹ (ATCC) and 0.313 mg mL⁻¹ (clinical isolate), representing a 16-fold improvement in both cases, underscoring its potential against hard-to-treat nosocomial infections.

The antifungal evaluation revealed similarly promising results. For both the ATCC and fluconazole-resistant clinical isolates of C. albicans, the GO–ZnO nanohybrid exhibited an MIC of 1.25 mg mL⁻¹, matching the activity of ZnO, but demonstrating substantially greater efficacy than GO (2.5–5 mg mL⁻¹). The retention of antifungal activity against drug-resistant clinical isolates highlights the nanohybrid’s potential as an alternative therapeutic strategy in cases where conventional antifungals are ineffective.

As shown on the right side of Fig. 10, the GO–ZnO nanohybrid exhibits significantly lower MIC values (mean = 0.98 mg mL⁻¹) compared to its individual components, ZnO (mean = 1.65 mg mL⁻¹) and GO (mean = 6.5 mg mL⁻¹), across the tested pathogens. This indicates that the nanohybrid is approximately 1.5-fold more potent than ZnO and 6.5-fold more effective than GO alone. The statistical analysis confirms highly significant differences between the GO–ZnO nanohybrid and both precursors (p < 0.01 for GO; p < 0.05 for ZnO), while no significant difference was observed between ZnO and GO. This synergistic effect likely arises from the combined antimicrobial mechanisms: ZnO-induced ROS generation and GO-mediated membrane disruption, which together enhance bacterial cell damage at lower concentrations. These findings underscore the benefits of nanoarchitectural design in developing next-generation antimicrobial materials with improved efficiency and reduced required dosages.

When compared with other nanomaterial-based antimicrobials reported in the literature, the GO–ZnO nanohybrid demonstrates competitive or superior MIC values. For instance, silver nanoparticles often require 1–2 mg mL⁻¹ for effective inhibition of S. aureus and E. coli, while CuO and TiO₂ nanoparticles typically exhibit weaker antimicrobial activity, particularly against Gram-negative strains^70–72. In contrast, the GO–ZnO nanohybrid achieves comparable or lower MICs, while offering enhanced stability, lower cytotoxicity risk compared to silver, and improved versatility against both bacterial and fungal pathogens. These advantages position GO–ZnO as a promising alternative nanomaterial for broad-spectrum antimicrobial applications, particularly in the context of the growing issue of multidrug resistance.

Mechanisms

The superior performance of the GO–ZnO nanohybrid in both dye removal and antimicrobial activity arises from its unique structural features and the synergistic interactions between its components, as schematically illustrated in Fig. 11.

Fig. 11.

Proposed mechanistic pathways of the GO–ZnO nanohybrid: a structural model showing ZnO nanoparticles anchored onto GO sheet via oxygen-containing functional groups; b multimodal MG adsorption mechanism involving electrostatic attraction, π–π stacking, and hydrogen bonding; c dual antimicrobial mechanism combining ROS-mediated oxidative stress and physical membrane disruption by the sharp edges of GO against E. coli, S. aureus, and C. albicans. (This schematic illustration was generated with the assistance of artificial intelligence to conceptually represent)

As shown in Fig. 11a, the GO–ZnO nanohybrid consists of graphene oxide nanosheets enriched with oxygenated functional groups (–OH, –O–, –COOH, and epoxy) that provide abundant active sites for adsorption and interaction. ZnO nanoparticles are uniformly anchored onto the GO surface, where they not only increase the effective surface area but also prevent the restacking of GO nanosheets. This hybrid configuration establishes a stable heterointerface that enhances electron transport, facilitates surface reactivity, and creates a porous and chemically robust network. Such synergistic integration improves dispersibility in aqueous media and ensures long-term durability during repeated adsorption–desorption cycles^73,74.

For dye removal, the adsorption of MG onto the GO–ZnO nanohybrid is governed by multiple physicochemical interactions that collectively enhance both affinity and adsorption capacity. As shown in Fig. 11b, electrostatic attraction is the predominant mechanism: the negatively charged oxygenated functional groups on GO (particularly carboxylate, –COO⁻, and hydroxyl, –OH) strongly interact with the positively charged nitrogen centers of MG under neutral to alkaline conditions. This charge-based interaction is further stabilized by π–π stacking between the aromatic rings of MG and the sp²-hybridized carbon domains of GO, providing strong non-covalent anchoring. Additional stabilization is achieved through hydrogen bonding between epoxide or hydroxyl groups on GO and heteroatoms (N, O) within MG^73,75. ZnO nanoparticles also contribute by providing additional surface-active sites, thereby enhancing chemisorption and facilitating electron transfer, which accelerates dye–adsorbent interactions. They also enhance structural integrity and chemical stability, thereby improving recyclability and long-term durability. The GO–ZnO synergy prevents nanoparticle aggregation and maintains highly accessible adsorption sites, thereby achieving superior adsorption efficiency⁶⁴.

Beyond adsorption, the GO–ZnO nanohybrid demonstrates broad-spectrum antimicrobial activity against Gram-negative (E. coli), Gram-positive (S. aureus), and fungal (C. albicans) strains. As shown in Fig. 11c, its biocidal efficacy operates via a dual mechanism. First, ZnO nanoparticles generate ROS, including superoxide radicals (·O₂⁻), hydroxyl radicals (·OH), and hydrogen peroxide (H₂O₂), under ambient or photoactivated conditions. These ROS induce oxidative stress, leading to lipid peroxidation, protein denaturation, DNA fragmentation, and ultimately cell death^75,76. Second, the sharp edges of GO nanosheets physically disrupt microbial membranes via a “nanoknife” effect. Moreover, the proximity of GO–ZnO to microbial membranes localizes ROS generation at the cell surface, intensifying oxidative damage. The nanohybrid also perturbs membrane potential and ion transport, further destabilizing cellular homeostasis. Its structural rigidity enhances resistance to microbial defense mechanisms, ensuring sustained antimicrobial performance^77,78. The combination of ROS-induced chemical damage and GO-mediated mechanical disruption produces a synergistic antimicrobial effect, independent of conventional antibiotic pathways, thus reducing the risk of resistance development⁷⁹.

Supporting evidence from structural characterization, including FTIR (disappearance of the C = O peak indicating Zn²⁺ coordination), XRD (confirmation of crystalline ZnO), TEM (ZnO nanorods anchored on GO), and AFM (vertical ZnO growth on GO), corroborates the proposed mechanisms. The intimate GO–ZnO interface promotes efficient electron transfer and ROS generation, while the high surface area and accessible oxygen-containing functional groups maximize dye adsorption. Furthermore, the morphology of ZnO can be tuned via synthesis conditions to optimize both adsorption kinetics and antimicrobial activity.

Taken together, these findings establish the GO–ZnO nanohybrid as a multifunctional material that not only effectively removes hazardous dyes but also inactivates pathogenic microorganisms. Such dual functionality represents a significant advancement for integrated water treatment systems, where simultaneous pollutant removal and disinfection are required. Future investigations should evaluate real wastewater applications, regeneration efficacy, and potential ecotoxicological influences to validate its broader environmental and biomedical applicability.

Application to real water samples and recyclability

The practical applicability of the GO–ZnO nanohybrid as a sustainable adsorbent was evaluated through seven consecutive adsorption–desorption cycles, using 1 mL of methanol for regeneration under optimized conditions (pH = 9.0, 25 °C, 12 min). As illustrated in Fig. 12, the material exhibited outstanding reusability: MG removal efficiency remained above 98% during the first three cycles and declined to ~ 87% by the fourth cycle. Retention of > 87.0% efficiency after repeated use underscores the structural integrity and chemical stability of the nanohybrid. The modest performance loss can be attributed to partial active site deactivation caused by incomplete desorption of strongly bound dye molecules and/or slight ZnO nanoparticle aggregation during successive washings. Importantly, the use of minimal solvent (1 mL methanol per cycle) highlights the nanohybrid’s economic and environmental advantages over conventional single-use adsorbents.

Fig. 12.

Recyclability of the GO–ZnO nanohybrid over seven consecutive adsorption–desorption cycles. Each cycle involved regeneration with 1 mL of methanol, adjustment of the pH to 9.0, and reapplication under optimized conditions. Error bars represent the standard deviation (n = 3).

Beyond recyclability, the adsorbent’s performance was tested in real-world water matrices, including distilled water, mineral water, tap water, river water, and industrial wastewater, under the same optimized conditions (Fig. 13). Remarkably, MG removal exceeded 95% in all matrices except industrial wastewater, where efficiency slightly decreased to ~ 88%. This reduction is likely due to competitive adsorption from coexisting organic and inorganic pollutants, dissolved ions, or suspended solids that interfere with active site availability or alter surface charge. Nevertheless, the ability to achieve near-complete removal (> 88%) in such a complex system demonstrates the material’s robustness, selectivity, and applicability under realistic conditions. The consistent performance in tap and river water, which naturally contain organic matter and common ionic species, further supports its suitability for municipal and industrial wastewater treatment.

Fig. 13.

MG adsorption efficiency (%) of the GO–ZnO nanohybrid in various real water matrices: distilled water, mineral water, tap water, river water, and industrial wastewater. All experiments were conducted under optimized conditions (MG concentration = 25.0 mg L⁻¹, pH = 9.0, adsorbent dosage = 18.0 mg, temperature = 25.0 °C, contact time = 12.0 min). Error bars represent the standard deviation (n = 3).

Collectively, these findings establish the GO–ZnO nanohybrid as a highly efficient, reusable, and environmentally sustainable adsorbent that performs reliably under both laboratory and real-world conditions, which are critical attributes for next-generation materials in integrated water purification systems.

Conclusions

In this study, a GO–ZnO nanohybrid was developed as a multifunctional material capable of simultaneous dye removal and microbial disinfection. Under statistically optimized conditions (pH = 9.0, 18 mg adsorbent, 12 min, 25 °C), the nanohybrid achieved 96.54% removal of MG with a high monolayer adsorption capacity of 131.91 mg g⁻¹, surpassing many previously reported adsorbents. Adsorption was confirmed to be chemisorption-dominated, spontaneous, and endothermic, governed by synergistic electrostatic interactions, π–π stacking, and hydrogen bonding. The nanohybrid also demonstrated excellent reusability, retaining > 87% efficiency after four adsorption–desorption cycles with minimal solvent consumption (1 mL methanol per cycle). Its applicability in real-world applications was validated by efficient MG removal (> 88%) from complex water matrices, including industrial wastewater. Beyond adsorption, the nanohybrid exhibited strong antimicrobial activity against Gram-positive (S. aureus), Gram-negative (E. coli and A. baumannii) bacteria, and drug-resistant fungal (C. albicans) strains, with MICs as low as 0.313–1.25 mg mL⁻¹, depending on the microbial species and strain. This antimicrobial efficacy arises from ZnO-mediated ROS generation in combination with the membrane-disrupting “nanoknife” effect of GO sheets. These results establish the GO–ZnO nanohybrid as a highly efficient, reusable, and multifunctional platform capable of simultaneously addressing two critical water quality challenges—organic dye pollution and microbial contamination—within a single, scalable system, positioning it as a strong candidate for advanced wastewater treatment applications.

Author contributions

Sh. E.: Methodology, Software; Data Curation, Writing - Original Draft., P. Z.: Methodology, Investigation, Software; Data Curation., S. N.: Methodology, Investigation, Data Curation; Writing - Original Draft., H. J.: Methodology; Validation, Investigation, Writing - original draft, Writing - review & editing., A. B.: Writing - original draft, Methodology, Investigation, Software; Formal analysis., R. Kh.: Writing - original draft, Methodology, Investigation., A. A.: Supervision, Conceptualization, Methodology; Writing - original draft, Writing - review & editing, Validation, Project administration; Resources; Funding acquisition.

Data availability

All data generated or analyzed during this study are included in this published article.

Declarations

Competing interests

The authors declare no competing interests.