Introduction

Water plays a crucial role as a vital natural resource globally, and the effective management of this invaluable resource is essential to ensure its sustainability for the future, benefiting both the environment and human existence Throughout time, there has been a notable increase in contamination levels, and the utilization of certain technologies for remediation purposes has shown the potential to produce secondary contaminants or byproducts, further exacerbating environmental pollution concerns1. The origin of contaminants found in wastewater can be attributed to two primary sources. Firstly, natural processes such as volcanic activities, soil erosion, and rock weathering contribute to contamination. Secondly, human activities including waste disposal, urban runoff, mining, manufacturing of printed circuit boards, agriculture practices, metal surface treatment, fuel combustion, textile dyeing, semiconductor production, among others, also significantly impact water quality2.

To maintain the diverse range of plant and animal species that rely on water ecosystems, it is imperative to establish preventive measures against contamination originating from both organic and inorganic pollutants1. Over the years, there has been a consistent observation of substantial contamination issues, with dyes frequently identified as persistent organic and mineral pollutants of concern3. Dyes are manufactured on a large scale globally in various forms and quantities4, with classifications based on the source of their materials (natural or synthetic) and the characteristics of their chromophore or autochrome groups3. This categorization significantly influences their ability to enter water bodies, where they can impede light penetration, leading to detrimental effects on ecosystems by reducing photosynthetic activity.

Moreover, dyes present severe risks due to their highly hazardous and carcinogenic nature. The accumulation of these substances in aquatic organisms poses a significant environmental hazard, coupled with potential adverse impacts on human health such as skin irritation, allergic dermatitis, cancer development, and genetic mutations3,5. Industries generate a substantial volume of vibrant wastewater containing dyes that are often resistant to biodegradation, posing challenges to environmental sustainability6. Among the numerous artificial coloring agents, examples include aniline blue, alcian blue, basic fuchsin, methylene blue (MB), crystal violet, toluidine blue, and Congo red, with MB being a particularly harmful dye utilized across various industrial sectors7.

Out of all dye, methylene blue (MB), a cationic dye, which threaten the human health and environment, presents largely in wastewater from various industries such as pharmaceutical, food processing, paint, paper8,9. Another cationic dye, which are dumped into the environment from various industries including leathering, paint, paper, and even aquaculture, causing serious health problem to human, is malachite green (MG)10,11. Another pollution contribution from the textile industry is methyl orange (MO), which is an anionic dye, can also have similar effects to human health as malachite green12,13. Similar to MO, another anionic dye, which is another contributor to water pollution from textile industry and chemical engineering industries, is Congo red (CR)14,15. Hence, removing these dyes is necessary.

Moreover, harmful microorganisms is one of the leading causes in water pollution, besides toxic chemicals, causing various diseases in human health, leading to the prediction of water crisis in 202516,17. Water can be contaminated from various sources, especially from hospitals, which commonly consist of Staphylococcus aureus (SA) – a gram-positive bacterium, Pseudomonas aeruginosa (PA) – a gram-negative bacterium, and Salmonella enterica (SE) – a gram-negative bacterium18,19,20,21. However, these bacteria can be killed easily with various types of antibiotics. Unfortunately, with the excess usage of antibiotics, antimicrobial resistance (AMR) has become another threat to human health20,22,23. Therefore, antibacterial material that does not cause AMR would be ideal.

At first, to solve the dye contamination problem, researchers have found multiple materials that can remove the dye from the sources throughout numerous ways. Recently, the dye removal properties of g-C3N4 have been conducted throughout several articles24,25. To be more specific, by creating heterojunctions or joining with other materials, graphitic carbon nitride (g-C3N4), an organic semiconductor material with a structure similar to graphene, is recognized for its high photocatalytic performance and environmental friendliness26,27. g-C3N4 functions as a photocatalyst that efficiently breaks down dyes and, when combined with metal oxides, exhibits antibacterial qualities, as well break the dye molecules28,29. Additionally, the study revealed that g-C₃N₄ exhibited a bactericidal effect on E. coli in water when exposed to visible light30,31. Integrating g-C₃N₄ with other nanomaterials such as GO, TiO₂, AgBr, and Ag enhances disinfection efficiency due to a synergistic interaction30,32,33. Furthermore, g-C₃N₄ demonstrated strong effectiveness as both a powerful antibacterial agent—attributed to its amine groups—and as an efficient photocatalyst for removing dyes34,35,36.

Despite exhibiting antibacterial properties when activated by light through photocatalytic mechanisms, it is not inherently potent as traditional antimicrobial agents (like silver or copper). Traditionally, researchers around the world found that different types of polymers can be used to inhibit the growth and even kill bacteria. Recently, SA, PA, and SE can be inhibited using polyethylene glycol (PEG), chitosan (CS), polyvinyl alcohol (PVA), and polyvinylpyrrolidone (PVP), CS/PVA, PEG, PVP, CS/PEG, CS/PVP/PEG/PVA under various physical forms (i.e., film, coatings, aqueous, solids, gels, beads,…)17,37,38,39,40,41. In general, these polymers can be divided into two groups based on their antibacterial mechanisms – passive and active42. One of the most common passive polymers can be PEG, PVP42. On the other hand, one of the most common active polymer is CS43. However, not only polymers can be used as antimicrobial agents but also metal oxides particles (i.e., selenium, silver, iron oxide, calcium oxide, zinc oxide, copper oxide,…)44,45,46,47,48,49,50,51,52,53,54,55,56. Besides, metal oxides can combine with polymers to create antibacterial composites (i.e., iron oxide/PVA, CoFe2O4/PEG/PVP, zinc oxide (ZnO), polymeric-ZnO, copper oxide, nickel oxide,…)57,58,59,60,61,62. Out of all, the most recent polymer blend that exhibit the growth inhibition of SA is M8, which is a polymer blend consisting of PVA, PVP, PEG, and CS39. Additionally, M8 was also modified with silver nanoparticles to inhibit the growth of SA, PA, and SE. Hence, for enhancing the antibacterial characteristics of g-C3N4, another approach is coating/mixing with other better-known materials such as polymers or metal oxides. Specifically, in this study, the M8F, which is the more concentrated form of M8 were coated with metal oxide particles (ZnO, CuO, NiO, and ZNC – zinc/nickel/copper oxide). One of the reasons for using a more concentrated M8 is to have a thicker coating on the particles. Then, polymers coated with metal oxide particles were deposited on the g-C3N4. The novel composites were then called ZnO/M8F/g-C3N4, NiO/M8F/g-C3N4, CuO/M8F/g-C3N4, and ZNC/M8F/g-C3N4. These composites were used to not only remove cationic dye and anionic dye, specifically, malachite green (MG), methylene blue (MB), congo red (CR), and methyl orange (MO), but also inhibit the growth of SA, PA, and SE, which is the novelty and the impact of this research. However, since this is the comparative preliminary screening of investigated dual-function composite, the in-depth dye removal and antibacterial mechanisms were not studied, leading to future research.

Experimental methods

Materials, characterizatioin method, and synthesis methods

Polyethylene glycol – 1000 (PEG), zinc (II) sulfate heptahydrate (ZnSO4.7H2O), nickel (II) sulfate hexahydrate (NiSO4.6H2O), copper (II) sulfate pentahydrate (CuSO4.5H2O), and sodium hydroxide (NaOH) were purchased from Xilong Scientific Co., Ltd. (Shantou, China). Polyvinyl Alcohol (PVA) was purchased from Wuxi Yatai United Chemical Co., Ltd., Shanghai (China). Methyl orange (MO), chitosan (CS), congo red (CR), malachite green (MG), polyvinylpyrrolidone (PVP K30), methylene blue (MB), and melamine were purchased from Shanghai Zhanyun Chemical Co., Ltd. (Shanghai, China). Glacial acetic acid (AA) was obtained from RCI Labscan (Bangkok, Thailand). All materials were used as obtained.

Morphology, functional group, elemental composition, crystallographic structure, and magnetic property of materials were analyzed using field emission scanning electron microscope (FE-SEM, Hitachi SU8000, Tokyo, Japan), Fourier transform infrared spectrometer (FTIR, Tensor 27, Bruker, Germany), and X-ray crystallography (XRD, Bruker, D8Advance ECO A25-X1-9C2Z11G7F0, Karlsruhe, Germany) were used, respectively.

The synthesis methods can be summarized as in Fig. 1 and the detailed synthesis was shown in Supplementary Information.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Summarized synthesis process.

Dyes removal

These steps were repeated 3 times. First, 0.1 g of CuO/M8F/g-C3N4, or ZnO/M8F/g-C3N4, or NiO/M8F/g-C3N4, or ZNC/M8F/g-C3N4 were added into 15-mL falcon tube containing 10mL of 10ppm of MB or MG or MO or CR at room temperature and pH of 7 in the dark. The removal process was carried out for 48 h. The aliquots concentration was calculated based on the measurement of maximum absorbance at specific wavelength (MO, MB, MG, and CR at 464, 664, 617, and 498 nm, respectively63,64,65,66 using UV-Vis spectroscopy (Hach DR6000, CO, USA) after 48 h. The dye removal percentage was calculated using Eq. (1):

$$\:\text{\%}\text{R}\text{e}\text{m}\text{o}\text{v}\text{a}\text{l}=100\times\:\frac{{\text{C}\text{o}\text{n}\text{c}\text{e}\text{n}\text{t}\text{r}\text{a}\text{t}\text{i}\text{o}\text{n}}_{\text{f}\text{i}\text{n}\text{a}\text{l}}-{\text{C}\text{o}\text{n}\text{c}\text{e}\text{n}\text{t}\text{r}\text{a}\text{t}\text{i}\text{o}\text{n}}_{\text{i}\text{n}\text{i}\text{t}\text{i}\text{a}\text{l}}}{{\text{C}\text{o}\text{n}\text{c}\text{e}\text{n}\text{t}\text{r}\text{a}\text{t}\text{i}\text{o}\text{n}}_{\text{i}\text{n}\text{i}\text{t}\text{i}\text{a}\text{l}}}$$
(1)

Determining the Inhibition percentage

This experiment was repeated 3 times. In this experiment, the bacteria (Staphylococcus aureus strain ATCC 29523, Salmonella enterica ATCC 14028, Pseudomonas aeruginosa strain ATCC 9027) and the incubation method were based on previous publications without any modifications39.

To obtain the antibacterial agents, first, 0.1 g dried CuO/M8F/g-C3N4, or ZnO/M8F/g-C3N4, or NiO/M8F/g-C3N4, or ZNC/M8F/g-C3N4 was mixed with 90mL of DI. After sonicating for 2 h, the minimum inhibitory concentration (MIC) against SA, PA, and SE were performed using the same method and instrument as previous publication39. To determine the morphology, the functional groups, and the types of particles, the other 30mL of the mixture were dried overnight at 80 °C.

The percentage of inhibition was calculated using Eq. (2)39:

$$\:\text{\%}\:\text{i}\text{n}\text{h}\text{i}\text{b}\text{i}\text{t}\text{i}\text{o}\text{n}=\left(1-\left(\frac{{{\upepsilon\:}}_{\text{t}}-{{\upepsilon\:}}_{\text{t}0}}{{{\upepsilon\:}}_{(-)\text{t}}-{{\upepsilon\:}}_{(-)\text{t}0}}\right)\right)\times\:100$$
(2)

With \(\:{\epsilon\:}_{t}\) and \(\:{\epsilon\:}_{t0}\) is the optical density at 600 nm of the test tube at 24 h and 0 h post-inoculation, respectively. Additionally, \(\:{\epsilon\:}_{(-)t}\) and \(\:{\epsilon\:}_{(-)t0}\) is the optical density at 600 nm of the negative control tube at 24 h and 0 h post-inoculation, respectively. After determining the inhibition percentage, bactericidal effect was investigated by spreading the solutions from each tube on MH agar plates.

Results and discussion

Characterization of materials

FE-SEM

The morphology of ZnO and ZnO/M8F/g-C3N4 were shown Figure S1. According to Figure S1a and Figure S1b, the particles of ZnO are aggregated and it is found that the morphology is not uniform in shape. ZnO particles can be seen to have inconsistent shape, which some have kind of the round-corner shape; however, it is difficult to identify the number of edges on the particles. On the other hand, as shown in Figure S1c and Figure S1d, ZnO/M8F/g-C3N4 has the bigger sized particles compared to ZnO particles. The explanation can be concluded that the g-C3N4 has been in aggregation with ZnO.

ZnO has the average size of 165.8 ± 31.1 nm, as shown in Figure S2a. While according to Figure S2b, ZnO/M8F/g-C3N4 has a similar size compared to ZnO which has been measured to be 166.0 ± 45.7 nm and the distribution can be concluded to be uneven in size.

On the other hand, the morphology of CuO and CuO/M8F/g-C3N4 were shown in Figure S3. According to the data presented in Figure S3a and Figure S3b, it is evident that the CuO particles exhibit a tendency to aggregate, leading to a lack of uniformity in their morphology. The irregular shape of the CuO particles is notable, with some displaying a round-corner shape; however, precise quantification of the number of edges present on these particles proves to be challenging based on the visual information provided. Conversely, in the case of CuO/M8F/g-C3N4, the particles appear larger in size when compared to the standalone CuO particles based on Figure S3c and Figure S3d. This difference in size can be attributed to the presence of g-C3N4, which seems to be involved in the aggregation process with CuO, influencing the overall morphology of the composite material.

CuO exhibits an average size of 117.3 ± 36.0 nm, as illustrated in Figure S4a. It is worth noting that the size of CuO/M8F/g-C3N4, as depicted in Figure S4b, appears to be bigger to that of pure CuO, with measurements indicating a size of 265.6 ± 76.7 nm. The distribution can be seen to be varied and inconsistent in size among the particles due to uneven grinding.

The morphology of NiO and NiO/M8F/g-C3N4 were then shown in Figure S5. The morphology shown in Figure S5a and Figure S5b clearly indicates a tendency for NiO particles to aggregate, causing a lack of uniformity in their morphological features. It appears to have different shapes of particles which give an approximate round shape or even n-shaped appearance because of uneven pounding. Additionally, in accordance with Figure S5c and Figure S5d, the distribution tends to be bulkier compared to NiO themselves. However, the appearance of big particles can be understandable because the uneven size of NiO that led to the bigger particles have a higher surface area to aggregate with g-C3N4.

NiO demonstrates an average size of 97.0 ± 29.3 nm, as illustrated in Figure S6a. It is important to highlight that the dimensions of NiO/M8F/g-C3N4, as shown in Figure S6b, seem to be smaller compared to that of pure NiO, with measurements indicating a size of 73.5 ± 20.6 nm. Particles are a result of the heterogeneous grinding conditions employed during the preparation process.

On the other hand, the morphology of ZNC and ZNC/M8F/g-C3N4 were shown in Figure S7. According to Figure S7a and Figure S7b, there are some enormous particles displayed in the image compared to nearby small particles. The assumption of different size and shape in ZNC can be noted due to the uncontrollable agglomeration among ZnO, CuO, and NiO particles while conduction co-precipitation method. However, another explanation can be known to be due to the fluffy appearance of NiOPs while synthesizing separately. From this, the connection between particles seems to be denser compared to other oxides that when attaching with ZnO and CuO, it tends to form a bigger particles. On the other hand, Figure S7c and Figure S7d displayed the aggregation of g-C3N4 on the ZNC particles which can be easily seen on the additional small particles placed on the big-sized particles.

ZNC exhibits an average size of 108.3 ± 73.4 nm, as depicted in Figure S8a. The aggregation of g-C3N4 on ZNC particles have led to the bigger in size which is measured to be 123.4 ± 97.4 nm, as given in Figure S8b.

As shown in Fig. 2, the morphologies of ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4, the metal oxide particles can be seen as small dots on a surface (M8F/g-C3N4).

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

FE-SEM images of of (a) ZnO/M8F/g-C3N4, (b) CuO/M8F/g-C3N4, (c) NiO/M8F/g-C3N4,and (d) ZNC//M8F/g-C3N4.

Based on the morphology, the proposed structure of g-C3N4 was a scaffold for metal oxide particles to deposit on the surface while M8F covered the particles as well as g-C3N4. To confirm the attachments of metal oxide particles on the M8F/g-C3N4, XRD and FTIR must be used.

FTIR

Using the FTIR from 400 to 4000 cm−1, the functional groups of M8F, g-C3N4, metal oxide (MeO), MeO/M8F/g-C3N4 were determined, as illustrated in Fig. 3.

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

FTIR spectra of (a) M8F, g-C3N4; (b) ZnO, ZnO/M8F/g-C3N4; (c) NiO, NiO/M8F/g-C3N4; (d) CuO, CuO/M8F/g-C3N4 (e) ZNC, ZNC/M8F/g-C3N4.

As illustrated in Table S1 and Fig. 3, the synthesis of g-C3N4 was successful based on the existence of the tris-triazine rings in the FTIR spectra67. Additionally, ZnO, CuO, and NiO had all necessary functional groups from 400 to 600 cm−1 to prove the chemical formula of these metal oxide particles. Additionally, the ZNC particles had the FTIR functional groups which at 450–600 cm−1, which overlapped with ZnO, CuO, and NiO, indicating the presence of these metals68. Moreover, the overlapping of vibrational band from 1630 to 3200 cm−1 indicated the functional groups of M8F (C = O, O–H) and g-C3N4 (C–N) appeared in the FTIR’s result of synthesized material denoted the complete attachment between materials. To sum up, combining with the FTIR peaks of M8F and g-C3N4, the composite of ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 were proven to be successfully synthesized.

XRD analysis (Figures S9-S13)

As illustrated in Figure S9-S13, the XRD analysis of g-C3N4, ZnO, CuO, NiO, ZNC, ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 were illustrated.

The hkl indices were calculated from the peaks 2θ, as shown in Table 1.

Table 1 XRD analysis of (%) of g-C3N4, ZnO, CuO, NiO, ZNC, ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4.
Full size table

As shown in Figure S9 and Table 1, from the hkl indices and the 2θ of g-C3N4, comparing to literature, in this research, g-C3N4 was successfully synthesized. Because the outer layer of the particles was coated by M8F, that might cause the decreases in the peak intensity of metal oxide and g-C3N4. As shown in Table 1, the hkl indices of ZnO indicating that the structure of ZnO was wurtzite75. Additionally, the hkl indices were also confirmed the chemical formula of CuO, and NiO. Additionally, XRD analysis also confirmed that ZNC particles were successfully synthesized68.

Additionially, the crystall lattice size as well as microstrain can be calculated using Williams-Hall (W-H) and Debye-Scherrer (D-S) methods, as shown in Eqs. (3) and (4), respectively80,81.

$$\:\beta\:cos\theta\:=\left(\frac{k\lambda\:}{D}\right)+\eta\:sin\theta\:$$
(3)
$$\:D=k\times\:\frac{\lambda\:}{\beta\:cos\theta\:}$$
(4)

In Eqs. (3) and (4), D, \(\:\theta\:\), \(\:\beta\:\), k, \(\:\eta\:\) is the average crystalline size of the composites, Bragg angle (in radians), full width at half maximum (FWHM) in radians, shape factor, and microstrain in the crystalline phase, respectively. To determine the average crystalline size accurately using the W-H and D-S methods, the shape factor must be determined, and the shape of the composites must be consistent and well-structured. Hence, in this study, the crystalline size of the composites was not calculated using W-H and D-S methods.

Removal of dye

As shown in Fig. 4, ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4, had the ability to remove 5.22 ± 0.87%, 6.17 ± 1.14%, 9.91 ± 2.72%, and 11.29 ± 0.71% of MB, respectively.

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Using ZnO/M8F/g-C3N4 (orange), CuO/M8F/g-C3N4 (green), NiO/M8F/g-C3N4 (blue), and ZNC/M8F/g-C3N4 (purple), the removal percentage of (a) MB, (b) MG, (c) MO, (d) CR.

Hence, as shown in Fig. 4, out of all MB removal agents, ZNC/M8F/g-C3N4 is the most effective to remove MB since the composite as the highest removal dye percentage. Additionally, out of all MG removal agents, ZNC/M8F/g-C3N4 is most effective to remove MG since the composite as the highest removal dye percentage, which can remove 37.28 ± 4.15% of the dye. However, ZnO/M8F/g-C3N4 can remove 35.57 ± 2.67% of MG, which is similar to ZNC/M8F/g-C3N4, statiscally. Additionally, out of all MO removal agents, ZnO/M8F/g-C3N4 is most effective to remove MO, which can remove 29.77 ± 0.31% of the dye. Additionally, out of all CR removal agents, ZNC/M8F/g-C3N4 is the most effective to remove CR since the composite as the highest removal dye percentage, which can remove 80.51 ± 4.36% of the dye. However, ZnO/M8F/g-C3N4 has similar removal percentage (70.6 ± 13.2%) to ZNC/M8F/g-C3N4, statistically. As shown in Fig. 4, all materials can be used as dye removal agent, in which they can remove CR the most. However, in the case of cationic dye, ZNC/M8F/g-C3N4 has the highest removal ability. In the case of anionic dye, ZnO/M8F/g-C3N4 has the highest removal ability.

In this study, g-C3N4 was modified with metal oxides and a polymer blend. The dye removal ability may arise from g-C3N4 itself, the metal oxides, or their synergistic interaction. For g-C3N4, dye removal under UV or visible light is generally attributed to photocatalytic degradation, since g-C3N4 is a well-established photocatalyst82. Another possible mechanism is adsorption, which can occur under both light and dark conditions. In this work, dye removal was conducted in the dark; therefore, adsorption is considered the most plausible mechanism. Adsorption onto solid surfaces may proceed via either physisorption or chemisorption83,84,85,86,87. However, determining the precise adsorption mechanism requires isotherm and kinetics modeling84,85,87,88,89.

Because most studies on g-C3N4 have emphasized its photocatalytic degradation pathway, adsorption mechanisms are often overlooked. A common approach is to first allow complete adsorption of the dye in the dark, followed by light or UV irradiation to initiate degradation. Consequently, adsorption processes have not been systematically investigated. Recently, several studies on the dye adsorption using g-C3N4 and g-C3N4 composites were reported, as shown in Table 2.

Table 2 Adsorption of various dye using g-C3N4 and g-C3N4 composites.
Full size table

As shown in Table 2, no universal trend in isotherm or kinetic models can be established for the adsorption of cationic and anionic dyes using g-C3N4 and its composites. In general, for cationic dyes such as methylene blue (MB) and malachite green (MG), adsorption is often described by the Langmuir or Freundlich isotherm (with Langmuir being more common) and typically follows pseudo-second-order kinetics. Specifically, for MB, pseudo-second-order kinetics is consistently reported, while both Langmuir and Freundlich models have been applied to describe equilibrium. For MG, pseudo-second-order kinetics is again the most common, with the Langmuir isotherm providing the best fit.

In general, for anionic dyes such as methyl orange (MO) and Congo red (CR), adsorption behavior is more system-dependent, with both Langmuir and Freundlich isotherms reported, and no single model being dominant. Kinetic analyses most often follow pseudo-second-order models, although pseudo-first-order behavior has also been observed in some composites. This suggests that anionic dye adsorption on g-C3N4-based materials is governed by a combination of electrostatic attractions, multilayer adsorption on heterogeneous surfaces, and interactions introduced by surface modifications. For methyl orange (MO), adsorption behavior varies depending on the type of g-C₃N₄ composite used. Both Langmuir and Freundlich models have been reported, with Freundlich suggesting multilayer adsorption on heterogeneous surfaces, and Langmuir implying monolayer coverage on homogeneous active sites. Kinetic analyses for MO most often align with pseudo-second-order models, although pseudo-first-order behavior has also been observed in some cases, particularly for polyaniline/g-C3N4 composites100. These differences suggest that the adsorption mechanism of MO is strongly dependent on the surface chemistry of the composite, with electrostatic attraction playing a major role in some systems, and multilayer adsorption or heterogeneous binding sites dominating in others. For Congo red (CR), fewer reports are available, but the existing studies reveal similarly diverse adsorption characteristics. Both Freundlich and pseudo-second-order models have been used to describe CR adsorption onto La-NiO/g-C3N4102, indicating a multilayer adsorption process on a heterogeneous surface with multiple anchoring interactions. In contrast, polyaniline/g-C3N4 composites follow pseudo-first-order kinetics100, dominated mainly by electrostatic attractions. Taken together, these findings show that CR adsorption on g-C₃N₄ composites is system-specific, with adsorption controlled by both surface modification and the interplay of electrostatic and multilayer interactions.

Additionally, adsorbing the dye at different pH is also important since the solution pH directly affects both the ionization state of the dye molecules and the surface charge of the adsorbent102. This relationship is governed by the point of zero charge (pHpzc) of the material102. For instance, the pHpzc of g-C3N4 adsorbing MB, which is an cationic dye, is 591. However, when adsorbing methyl oragne, which is an anionic dye, the adsorption amount changed negligibly90. When the pH of the solution is below the pHpzc, the surface acquires a positive charge, thereby enhancing the adsorption of anionic dyes but reducing the affinity for cationic dyes102. In contrast, when the pH is above the pHpzc, the surface becomes negatively charged, favoring the adsorption of cationic dyes while repelling anionic species102. Furthermore, changes in pH can influence the degree of protonation or deprotonation of functional groups on both the adsorbent and the dye, thereby altering the strength of electrostatic interactions, hydrogen bonding, and even π–π stacking102. Thus, pH optimization is a critical factor in determining adsorption efficiency and selectivity for different classes of dyes102.

These findings indicate that adsorption mechanisms must ultimately be confirmed through detailed isotherm and kinetic studies. While the mechanisms summarized in Table 2 (e.g., electrostatic interactions, π–π stacking, hydrogen bonding, chemisorption, and multilayer adsorption) provide useful insight from the literature, they should at this stage be regarded as hypotheses rather than definitive conclusions. The actual dominant mechanism in a given system depends strongly on surface chemistry, morphology, and the physicochemical properties of the dye molecules, which can vary across composites. In the present work, the primary objective was to identify the most suitable metal oxide component for g-C3N4-based composites in future applications. Therefore, adsorption isotherm and kinetic experiments were not performed, and the mechanistic assignments should be interpreted as tentative and subject to experimental validation in subsequent studies.

Antimicrobial activities

As shown in Fig. 5, the bacterial inhibition percentage of g-C3N4, ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 against SA, PA, and SE were illustrated.

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

Inhibition percentage of g-C3N4 (black), ZnO/M8F/g-C3N4 (orange), CuO/M8F/g-C3N4 (blue), NiO/M8F/g-C3N4 (green), and ZNC/M8F/g-C3N4 (purple) against (a) SA, (b) PA, and (c) SE.

As shown in Fig. 5, for SA, the concentration that inhibits the most is at 50% dilution for g-C3N4, ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4. Specifically, at 50% dilution, the growth inhibition of g-C3N4, ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 is 29.1 ± 5.7, 53.6 ± 9.3, 35.2 ± 5.3, 35.2 ± 6.2, and 24.0 ± 14.0%, respectively. However, at 25% dilution, the growth inhibition of NiO/M8F/g-C3N4 is 33.5 ± 7.5%. Hence, with only ~ 2% increase in growth inhibition, in the case of SA, if NiO/M8F/g-C3N4 is in used, the concentration at 25% dilution should be more economical.

As shown in Fig. 5, for PA, the concentration that inhibits the most is at 50% dilution for g-C3N4, ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4. Specifically, at 50% dilution, the growth inhibition of g-C3N4, ZnO/M8F/g-C3N4, NiO/M8F/g-C3N4, CuO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 is 44.8 ± 5.9, 30.8 ± 1.2, 53.4 ± 3.4, 73.1 ± 16.1, and 65.6 ± 15.5%, respectively. Based on the standard deviation, CuO/M8F/g-C3N4 and ZNC/M8F/g-C3N4 can be considered to be unreliable antibacterial against PA at 50% dilution. Hence, the next highest growth inhibition antibacterial agent is NiO/M8F/g-C3N4 at 50% dilution. Additionally, when NiO/M8F/g-C3N4 at 25% dilution, it exhibited similar growth inhibition percentage against SA at 33.2 ± 7.01%.

As shown in Fig. 5, for SE, the concentration that inhibits the most is at 50% dilution for g-C3N4, ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4. Specifically, at 50% dilution, the growth inhibition of g-C3N4, ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 is 30.0 ± 3.3, 21.4 ± 1.8, 26.4 ± 1.5, 9.7 ± 2.2, and 17.5 ± 3.8%, respectively. Despite g-C3N4 at 50% dilution has higher inhibition percentage than NiO/M8F/g-C3N4, but, when taking account of the standard deviation, both of these materials have similar growth inhibition percentage.

Comparing between 3 bacteria – SA, PA, and SE, ZnO/M8F/g-C3N4C3N4 showed stronger growth inhibition against gram-positive bacteria. This results is similar to other publications103,104. In contrast, g-C3N4 seems like to be more effective against gram-negative bacteria. On the other hand, considering the order of inhibition from highest to lowest using NiO/M8F/g-C3N4, CuO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 as antibacterial agents, the order is PA > SA > SE. This indicates that there are no clear trends in favoring neither gram-positive nor gram-negative bacteria.

Based on Fig. 5, the MIC values were determined and shown in Table 3.

Table 3 MIC of ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 against SA, PA, and SE.
Full size table

As shown in Table 3, for SA, only ZnO/M8F/g-C3N4 exhibited MIC50. For PA, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4C3N4, and ZNC/M8F/g-C3N4 exhibited MIC50. For SE, none of the novel bacterial agents exhibited MIC50. Additionally, as shown in Table 3, for SA, SE, and PA, none of the novel bacterial agents exhibited MIC90. Additionally, as illustrated in Table 4, none of the materials that can kill SA, SE, and PA.

Table 4 Bactericidal results of ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 against SA, PA, and SE.
Full size table

Before delving into the mechanisms, the differences between gram-negative and gram-positive bacteria must be understood. The gram-negative bacteria has a bilayer negatively-charged cell wall105. This cell wall consists of inner layer and outer layer105. The inner layer consists of peptidoglycan and the outer layer, which is negatively-charged, consists of lipopolysaccharides105. On the other hand, gram-positive bacteria has much thicker cell wall and less negatively-charged cell wall, comparing to gram-negative bacteria105. In this study, the polymers were modified with metal oxide particles. Hence, the mechanisms can be contributed by polymers, metal oxide particles, or the synergistic effects between these two components.

One of the elements influencing the antibacterial activity may be polymers. Antibacterial properties are exhibited by polymeric materials, which include copolymers, functionalized polymers, and other macromolecular systems. These antibacterial polymers may be divided into two groups: passive and active, depending on the antibacterial processes. In essence, the passive polymers (PEG and PVP) reduce the bacterial adhesion to the surface42. Furthermore, due to the existence of amino groups in CS, the interaction between bacteria’s negatively-charged cell wall is interacted with CS, leading to the RNA production, and resulting in the demise of the bacterium105.

In the case of metal oxides, CuO, NiO, and ZnO have some main antibacterial mechanisms. Due to the negatively-charged cell wall of bacteria, CuO, NiO, and ZnO can release ions – Cu2+, Ni2+, and Zn2+. These ions can attract to the cell wall due to electrostatic forces. Due to the higher difference between the charges, these ions can attract to the gram-negative bacteria cell wall easier, comparing to gram-positive bacteria105. The interaction between the particles and the cell wall cause cell wall disruption, leading to the death of bacteria106,107,108. However, in gram-positive bacteria, the cell wall is multilayered and much more permeable than gram-negative bacteria, which can be used as a shield to protect the bacteria against metal oxide particles105,109. Another explanation for the antibacterial mechanisms for these metal oxide particles is the increase production of reactive oxygen species (ROS) in bacteria and cause the dysfunction of bacteria’s macromolecules105,110. To be specific, ZnO can increase the ROS, which attaching on the surface or accumulating in the SA cells cytoplasm111. However, these mechanisms are not confirmed, but just proposed by many researchers105,109,112. Based on these proposed antibacterial mechanisms of ZnO, NiO, and CuO particles, the proposed mechanism for ZNC particles is (1), the electrostatic forces between positively charged particles and negatively-charged bacteria cell wall, (2) the increase in ROS in bacteria, (3) the synergistic effects between metal ions and polymers. However, further studies should be conducted to confirm these mechanisms.

Dual-function composite and recommendation

As discussed, the composite that can inhibit the growth of PA the most (in order from highest to lowest) is CuO/M8F/g-C3N4 and ZNC/M8F/g-C3N4. Despite of having higher growth inhibition percentage, similar MB and CR removel percentage, CuO/M8F/g-C3N4 has much lower MG removal percentage. Hence, ZNC/M8F/g-C3N4 is a more versatile dual-function composite to remove MB, MG, CR dye as well as inhibit the growth of PA. On the other hand, since only ZnO/M8F/g-C3N4 has MIC50 against SA, and ZnO/M8F/g-C3N4 can remove MO much more than other composites, ZnO/M8F/g-C3N4 should be considered to be a more versatile dual-function composite to remove MO and inhibit the growth of SA. Taking account of none of these composite have MIC50 against SE, ZNC/M8F/g-C3N4 and ZnO/M8F/g-C3N4 must be further modified to remove all of the metioned dye and inhibit the growth of mentioned bacteria better.

Additionally, as this research’s main focus was to perform comparative preliminary screening of ZnO/M8F/g-C3N4, NiO/M8F/g-C3N4, CuO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 as dual-function composite, the dye removal mechanisms and others related dye removal experiments were not performed. Specifically, to fully understand the dye removal mechanisms, further studies should be conducted. One of the possible studies can be measuring the removed dye at different initial dye concentrations at different temperature. From the proposed studies, the kinetics, isotherms, and thermodynamics models can be evaluated and determined whether the dye removal mechanism is adsorption (physically or chemically), absorption, diffusion (i.e. intraparticle diffusion), or the combination113,114. Additionally, future study can include the dye removal process, in which can be done under UV, visible, or IR light, to investigate the degradation capability of the material since g-C3N4 is a photocatalysts. Additionally, since these dyes are anionic and cationic, which is affected by the pH of the solution, future study can investigate the removal ability of these materials at different pH since different functional groups in each dye can be activated at different pH. Additionally, the pH zero-point charge of the composite must also be determined.

Conclusion

A graphitic carbon nitride (g-C3N4) and polymer blend (M8F) consists of chitosan, polyethylene glycol, polyvinyl alcohol, and polyvinylpyrrolidone were used to modified the surface of zinc oxide particles (ZnOP), copper oxide particles (CuOP), nickel oxide particles (NiOP), zinc/nickel/copper oxide particles (ZNC) which has the average size of 165.8 ± 31.1, 117.3 ± 36.0, 97.0 ± 29.3, and 108.3 ± 73.4 nm, respectively. The combination of metal oxides with g-C3N4 with M8F that had been synthesized were ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4 and their average sizes were calculated to be 166.0 ± 45.7, 265.6 ± 76.7, 73.5 ± 20.6, and 123.4 ± 97.4 nm, respectively.

In the case of cationic dye (MB and MG), ZNC/M8F/g-C3N4 has the highest removal ability. In the case of anionic dye as MO, ZnO/M8F/g-C3N4 has the highest removal ability. Out of these dyes, CR can be removed the most by these novel materials.

On the other hand, against Staphylococcus aureus (SA) and Pseudomonas aeruginosa (PA) the materials that should be used to inhibit the growth are ZnO/M8F/g-C3N4 at 50% dilution and CuO/M8F/g-C3N4 at 50% dilution, respectively. In the case of Salmonella enterica (SE), none of the materials can inhibit at least 50% of the bacteria. However, at these dilution concentration, the materials can only inhibit 50% of mentioned bacteria. Additionally, none of the materials can kill SA, PA, and SE. Hence, in terms of antibacterial agents, these materials can be considered to not have these properties.

Hence, as a preliminary study of the novel material of ZnO/M8F/g-C3N4, CuO/M8F/g-C3N4, NiO/M8F/g-C3N4, and ZNC/M8F/g-C3N4, these materials can not only remove anionic dye and cationic dye but also have the capability to inhibit the growth of SA, PA, and SE. Since this study is a preliminary study, these novel materials should be studied even further on the removal mechanisms of cationic dye, anionic dye, and other possibility to remove other dyes and also inhibit, and even kill other harmful bacteria. Additionally, further study should be conducted to enhance the antibacterial ability, as well as increasing the dye removal ability.