Introduction

Nanomaterials have emerged as promising candidates to play important roles in wastewater treatment, contributing to the development of new membrane materials, adsorbents, and catalysts1,2,3,4. Nanomaterials, including carbon allotropes (such as graphene, CNTs, and MWCNTs)5, graphitic carbon nitride6, and metal oxides7, are increasingly vital in wastewater treatment, offering favorable properties such as small size, large surface area, high porosity, high catalytic activity, and tailorable physical and chemical properties, making them attractive for wastewater treatment8,9,10. Their easily-modifiable nature enhances nanoscale remediation applications11, effectively treating pollutants and leading to prevention or minimization12,13,14,15. Engineered nanoparticles and nanocomposites can effectively remove a range of contaminants, including dyes, organic molecules, and pathogens in industrial wastewater16,17,18. Additionally, nanocomposite hydrogels are emerging as efficient adsorbents for hazardous substances in wastewater19. More recently, a family of 2D nanomaterials named transitional metal dichalcogenides (TMD) has gained significant attention20. Among them, molybdenum disulfide (MoS2) has been identified as a potential material for water treatment applications. Studies have demonstrated the ability of MoS2 to act as a photocatalyst in the degradation of organic pollutants21, a process enhanced by its high surface area and active edge sites22. MoS2 has also been utilized as a sorbent for heavy metals, owing to its strong affinity for various ionic species23. Moreover, the versatility of MoS2 allows for its integration into composite materials that synergistically improve the efficiency of contaminant removal from wastewater24.

The advancement in nanotechnology has led to the precise control of the size, structure, properties, and morphology of nanomaterials25,26. Various synthesis methods, such as chemical reduction, template-based approaches, and hydrothermal reactions, have been developed to produce nanomaterials with controlled morphology. These methods demonstrate the capability to finely tune their dimensions and enable the formation of nanoparticles with different shapes, including spheres, rods, cubes, and plates27,28. Material functionality has also been improved as a result of employing nanotechnology. For example, the superior catalytic performance of platinum nanoparticles is attributed to their unique size-dependent properties29.

The incorporation of carefully designed nanomaterials into membranes brings opportunities to overcome the classic trade-offs between the separation of contaminants and water permeate and improve the membrane’s overall performance by introducing exceptional characteristics such as high surface area, tunable porosity, and selective permeability30. The resultant nanomaterials incorporated membrane led to optimizing their performance for enhanced filtration and purification processes31,32. MoS2 nanoparticles are a prime example, offering improved molecular separation due to their unique layered structure and chemical properties, thereby improving the efficiency of the processes even further. Adsorptive polymeric membranes have demonstrated significant promise in the removal of heavy metals from wastewater. For instance, a novel complexing membrane containing poly(ethyleneimine) in a poly(vinyl alcohol) matrix exhibited remarkable removal efficiencies for heavy metals such as Cd²⁺ (96-99.5%), Cu²⁺ (94-99%), and Pb²⁺ (80-96%)33. However, these membranes reach saturation quickly, limiting their long-term effectiveness. Chitosan, a polymer known for its high adsorption capacity, can similarly reach equilibrium fast, necessitating frequent regeneration34,35. On the other hand, nanomaterial-enhanced adsorptive membranes have shown even greater potential. Studies have shown that nanomaterial-enhanced membranes, such as those incorporating GO, exhibit superior adsorption capabilities due to their high surface areas and functionalized surfaces. For instance, GO-based membranes have achieved Cd²⁺ removal efficiencies of up to 95% from aqueous solutions36. The mechanism of adsorption in these membranes involves several processes, including electrostatic attraction, complexation, and ion exchange. Nanomaterials provide a high density of functional groups, such as hydroxyl, carboxyl, and amine groups, which can bind heavy metal ions through these mechanisms37. The high surface area of these nanomaterials allows for a greater number of binding sites, which increases the adsorption capacity and efficiency. However, adsorption equilibrium is a limiting factor of continuous and progressive removal of heavy metals by many adsorptive-based membranes.

In the synthesis of MoS2 nanomaterials, precise control over reaction conditions, including temperature, reactant concentration, pH, reaction time, and sulfur source, is essential for tailoring its morphology and characteristics. Investigating the impact of growth temperature variations on MoS2 morphology during synthesis has yielded valuable insights. Lower temperatures tended to favor the creation of MoS2 quantum dots, while temperature increments promoted the growth of nanosheets38. In contrast, higher temperatures facilitated the formation of larger continuous sheets, accompanied by distinct morphological transitions39, and elevated temperatures further induced the growth of nanopetals and nanorods exerting a profound influence on the chemical and optical properties of the synthesized MoS2 materials38,40. On the other hand, increasing precursor concentration leads to a transformation from irregular, aggregated nanoflower-like morphologies to nearly spherical shapes associated with enhanced crystallinity and enlarged nanoflower sizes41. The influence of pH on MoS2 morphology resulted in the formation of nanosheets, nanorods, and even hollow nanorod structures as the acidic conditions increased42.

Furthermore, the duration of the synthesis process, or reaction time, plays a fundamental role—specifically, longer reaction times allowed for the formation of a nanoflower-like MoS2 morphology with identifiable petal-like features. Conversely, a shorter reaction time yielded material exhibiting a mixed-phase structure43. Additionally, the choice of the sulfur source significantly affects MoS2 morphology and characteristics. MoS2 samples synthesized from l-cysteine led to the formation of nanosheets and intermediary nanorods characterized by higher crystallinity. In contrast, the utilization of thiourea as the sulfur source resulted in MoS2 with a distinct morphology with petal-like features43. These multifaceted factors collectively contribute to the versatile design and tailoring of MoS2 nanomaterials, enabling their applicability across various domains.

The careful trials and evaluation of the effects of different morphologies of MoS2 additives in the membrane, such as nanosheets, nanotubes, or nanoparticles, are critical for optimizing the formation of nanocomposite membranes. These morphologies, individually or in combination, can lead to membranes with improved removal rates, anti-fouling properties, higher flux, and greater stability44,45,46. Different from the adsorption process where the membrane or adsorbent is submerged constantly in a solution, this experiment was conducted by passing the heavy metal solution through the adsorptive-catalytic membrane. Adsorption occurred as the heavy metal ions in the solution contacted the membrane, allowing for continuous wastewater treatment and easier operation. The catalytic properties of the membrane also generated reactive oxygen species (ROS), which enhanced the breakdown and removal of heavy metals, thereby assisting the adsorption process and improving overall efficiency. Despite numerous studies demonstrating the potential of MoS2 in enhancing membrane performance, current literature lacks comparison and understanding of how different morphologies of MoS2 nanoparticles influence the formation and efficacy of polymer-based membranes. Hence, this study aimed to address this by synthesizing three different morphological forms of MoS2 nanoparticles, including nanospheres, nanoplatelets, and nanosheets, under various reaction conditions and integrating them into a chitosan matrix. The effects of each morphology on the membrane characteristics and membrane performance in removing heavy metal contaminants from wastewater were systematically investigated. Chitosan is a polysaccharide biopolymer and is known as an environmentally friendly, biodegradable, renewable, and non-toxic green membrane material. Chitosan was chosen for this study because it can be well dissolved in an aqueous solution with acetic acid and is highly compatible with MoS2 water-based dispersions, whereas more common polymers such as Polyvinylidene fluoride (PVDF), cellulose acetate (CA) only can be dissolved in organic solvent, not in water.

Result and discussion

Synthesis of nanomaterials and membrane fabrication

MoS2 nanoparticles were synthesized under precisely controlled conditions; each synthesis was conducted with different reactant quantities and conditions, resulting in nanoparticles with distinctively different morphologies. Reaction conditions, particularly pH value and precursor concentration, played a crucial role in the resultant morphology. Lower pH (<1) favored the formation of spherical nanoparticles, while higher pH (=2) favored platelet structure.

The nanomaterial aqueous dispersions used in membrane fabrication are shown in Fig. 1a–c. The corresponding fabricated membranes are depicted in Fig. 1d–f. All the fabricated membranes had a uniform smooth appearance with a slight color variation. The S-CM nanosphere membrane exhibited a dark brown color, and the P-CM nanoplatelet membrane, characterized by its blackish color, were both fabricated using a higher amount of MoS2, specifically 190.4 mg. In contrast, the T-CM Nanosheet membrane, with its light gray appearance, was produced using a lesser amount of MoS2, only 32.1 mg. The amount of MoS2 in each case corresponded to the amount of MoS2 synthesized per batch of hydrothermal reaction.

Fig. 1: Nanomaterial solutions and resultant membranes.
figure 1

a MoS2 nanosphere solution, (b) nanoplatelet solution, and (c) nanosheet solution. Corresponding fabricated membranes are illustrated in (d) as the S-CM Nanosphere Chitosan Membrane, (e) the P-CM Nanoplatelet chitosan membrane, and (f) the T-CM nanosheet chitosan membrane, each exhibiting unique characteristics derived from their respective nanomaterial formulations.

Full size image

Nanomaterials and membranes characterization

The distinct morphologies of nanospheres, nanoplatelets, and nanosheets are shown in Fig. 2a, d, g. Low magnification images initially displayed the aggregated nature of the nanospheres, clearly showcasing their spherical morphology with an average size ranging from approximately 300 to 500 nm (Fig. 2a). However, the SAED pattern depicted in Fig. 2b and the FFT performed on a high-resolution TEM image demonstrated in the inset of Fig. 2c suggested a very weak crystalline structure in these nanospheres. The TEM image depicted in Fig. 2d exhibited a plate-like structure for the nanoplatelets, which were observed to be loaded with some nanospheres. These predominantly rectangular nanoplatelets were larger in size but maintained a flat and thin structure, resembling opened and expanded versions of the nanospheres. It was noted that these nanoplatelets revealed a tendency to curl up and fold, presenting a morphology distinct from the more structurally stable nanosphere-like morphologies. Whereas the nanosheets appeared ultra-thin, nearly transparent and presented markedly highest crystallinity among the studied morphologies (Fig. 2g). This superior crystallinity of the nanosheets was demonstrated by the FFT patterns, which elucidated the crystal structure of the particles as observed in the inset depicted in Fig. 2i. The thickness of these nanostructures exhibited a progressive decrease from spheres to platelets and subsequently to sheets, as illustrated by their corresponding SAED patterns shown in Fig. 2b, e, h, respectively. Additionally, an enhancement in the degree of crystallinity from the lowest of spheres to the highest of sheets was evident, as depicted by the FFT performed on high-resolution TEM images in the inset of Fig. 2c, f, i.

Fig. 2: TEM characterization of MoS2 nanostructures showing the morphological and crystalline distinctions among MoS2 nanospheres, nanoplatelets, and nanosheets.
figure 2

Image (a) shows the low-magnification (LM) image of MoS nanospheres, (b) the selected area electron diffraction (SAED) pattern of nanospheres, and (c) the high-resolution TEM (HRTEM) image along with the fast Fourier transform (FFT) pattern (inset) of nanospheres indicating relatively low crystallinity, (d) presenting the LM image, of the nanoplatelets, (e) the SAED pattern, and (f) the HRTEM image of nanoplatelets, including the FFT pattern (inset), revealing notably better crystallinity compared to the nanospheres, (g) is the LM image, (h) the SAED pattern, and (i) the HRTEM image with its FFT pattern (inset) of nanosheets, displaying the highest crystallinity level among the three nanostructures, evidencing their distinct structural qualities.

Full size image

Three membrane samples were analyzed by SEM to compare their surface and internal structures (Fig. 3). The top surfaces of all three membranes were prominently displayed with MoS2 nanomaterials (Fig. 3a, d, g), although the nanosheets were less visible than nanospheres and nanoplatelets. The cross-sectional images (Fig. 3b, e, h) highlighted the co-existence of MoS2 nanoparticles within the chitosan polymeric membrane matrix. Higher magnification revealed that the incorporated MoS2 were present as individual particles without large agglomeration; this confirmed not only the high quality of synthesized nanomaterials but also the successful fabrication of uniform composite membrane. EDX spectra analysis confirmed the presence of Mo and S in the membranes (Fig. 3i).

Fig. 3: SEM analysis of chitosan membranes embedded with MoS2 nanostructures providing a comprehensive look at the surface and cross-sectional SEM views of the S-CM, P-CM, and T-CM membranes.
figure 3

For the S-CM membrane, (a) depicts the surface morphology at a 10 µm scale, while (b) and (c) show the cross-sectional views at 20 µm, highlighting embedded nanospheres. In the P-CM membrane series, (d) illustrates the surface at a 10 µm scale, and (e) and (f) provide cross-sectional views at 20 µm, revealing the integration of nanoplatelets. The T-CM membrane is examined in (g) with a surface view at 5 µm, (h) showing a cross-section at 10 µm scale. i features the EDX plot for the S-CM membrane, affirming the presence of molybdenum (Mo) and sulfur (S), key elements in MoS2 nanomaterials, evidencing their successful incorporation into the membrane structure.

Full size image

Both MoS2 nanomaterial morphology and membrane topography were studied by Atomic Force Microscopy (AFM) to provide insights into nanoparticle height profiles and membrane surface roughness (Fig. 4). Profile analysis along the nanomaterials in AFM confirmed their dimensional parameters (Fig. 4a–c), where the height profiles of the nanosphere and nanoplatelets were 250 nm and 350 nm, respectively, while the nanosheet had the lowest height profile of about 145 nm. This information was consistent with the observation by TEM. Figure 4d–f depicts membrane topography, with Fig. 4g–i showing 3D representations. The S-CM membrane exhibited the highest roughness (RMS: 203.06 ± 0.1 nm), followed by P-CM (RMS: 157.22 ± 0.2 nm), and T-SM (RMS: 96.41 ± 0.1 nm).

Fig. 4: AFM imaging of MoS2 nanoparticles and chitosan membranes presents a series of 2D and 3D AFM images, showcasing the distinct height profiles and surface morphologies of MoS2 nanoparticles and the corresponding membranes.
figure 4

(a) The height profiles of MoS2 nanosphere, (b) nanoplatelets, and (c) nanosheets. 2D images highlight the surface morphology of (d) S-CM, (e) P-CM, and (f) T-CM membranes, offering a visual understanding of the textures influenced by embedded MoS2 nanostructures. 3D images reveal the surface roughness of the (g) S-CM, (h) P-CM, and (i) T-CM, further illustrating the impact of the nanostructures on the membranes’ topographical features.

Full size image

The streaming potential of membranes indicates their surface charge properties, which are crucial for interactions with solutes and filtration performance. The fabricated nanocomposite membranes showed varying isoelectric points (IEP), reflecting the influence of different MoS2 structures. S-CM had an IEP of 6.4, P-CM a lower IEP at 5.4, and T-CM a higher IEP of 7.5 (Fig. 5a–c). The lower IEP of nanoplatelet membrane P-CM suggested a more negatively charged surface at a solution with a pH of 7, possibly due to nanoplatelets’ larger surface area and specific edge properties. On the other hand, the charge of three MoS2 incorporated in the membrane was also measured. It was found that the platelet nanomaterial solution had the highest negative charge (−41.7 mV), which supported the membrane’s IEP results. Conversely, nanosheet membrane T-CM’s higher IEP indicated a less negatively charged surface; this was because that nanosheet solution had the lowest negative charge (−11.9 mV) due to nanosheets’ stacking nature limiting surface charge exposure; in addition, the higher crystallinity of the nanosheets possessed less charge inducing defects on the surface and edges. The nanosphere membrane S-CM exhibited a balanced surface charge with its IEP in the middle range, likely due to the spherical morphology of MoS2 nanospheres providing moderate surface charge distribution, consistent with the nanosphere solution’s charge (−20.5 mV).

Fig. 5: Streaming potential and water contact angle measurements of MoS2-based chitosan membranes.
figure 5

a the streaming potential across pH 4-14 for S-CM (IEP 6.4), (b) P-CM (IEP 5.4), and (c) T-CM (IEP 7.5), illustrating each membrane’s surface charge characteristics essential for solute interactions and filtration performance. Water contact angle measurements display varying degrees of hydrophilicity: (d) S-CM at 72 degrees indicates moderate hydrophilicity, (e) P-CM at 65 degrees has the highest wettability, and (f) T-CM at 87 degrees shows more hydrophobic properties.

Full size image

The water contact angle measurements for the fabricated nanocomposite membranes revealed different hydrophilicity levels (Fig. 5d–f). The nanosphere membrane S-CM exhibited moderate hydrophilicity with a contact angle of 72 degrees, while the P-CM membrane displayed the highest wettability with the lowest water contact angle of 65 degrees. Conversely, the nanosheet membrane T-CM demonstrated a less hydrophilic and more hydrophobic property with an 87-degree water contact angle. The P-CM membrane’s increased hydrophilicity is linked to the larger surface area and more defects on the plane and edges of MoS2 nanoplatelets, which facilitate enhanced interactions with water molecules. On the other hand, the lowered hydrophilicity of the T-CM membrane was likely a consequence of its nanosheet stacking arrangement, potentially restricting hydrophilic site exposure. The S-CM membrane’s spherical MoS2 nanospheres contributed to a balanced, intermediate hydrophilicity.

Additionally, the incorporation of chitosan and polyethylene glycol (PEG) into the membrane matrix played a pivotal role in enhancing hydrophilicity. Chitosan, with its inherent hydrophilic properties and ability to form hydrogen bonds with water molecules, significantly increased the water-attracting nature of the membranes47. The added hydrophilic polymer PEG not only contributed to the formation of a porous structure but also enhanced the overall membrane’s hydrophilicity. This combined hydrophilic effect of chitosan and PEG was essential in overcoming the slightly hydrophobic nature of the MoS2 nanoparticles, thus elevating the overall wettability of the composite membranes.

The membrane’s porosity measurements further elucidated the structural differences among the membranes. The P-CM membrane registered the highest porosity with a value of 3.15%, followed by the S-CM membrane at 2.87%. The T-CM membrane exhibited the least porosity, recorded at 2.03%, underscoring the correlation between membrane structure and its functional properties. It was clear that when MoS2 nanomaterials were incorporated in the chitosan polymer matrix, it affected the intrinsic membrane property which can be described by membrane structure parameter S48, which was defined by support layer thickness, tortuosity (Ʈ), and porosity (Ɛ) of the membrane. For polymer-only membranes, the porosity was decided completely by the pore-forming agent and phase inversion conditions during the casting process, however, the nano-additives such as MoS2 nanoparticles in this study introduced another influencing factor to dictate the pore formation as well as porosity. Based on the above results, the nanoplatelets with bigger and thicker dimensions could result in more pores in the membrane (porosity of 3.15%), whereas the more flexible and much thinner nanosheets could lead to fewer pores formed in the membrane (porosity of 2.03%). It was interesting to know that incorporating different nanomaterials could further manipulate and control the intrinsic structure property of the membrane.

Membrane performance evaluation

The flux performance evaluation of the three composite membranes was conducted under constant transmembrane pressure of 10 bar, and notable differences were observed: S-CM exhibited a flux of 60 ± 2 L/m² h at 10 bar, the flux of P-CM significantly was higher at 90 ± 3 L/m² h at 10 bar, and T-CM’s flux was considerably lower at 12 ± 1 L/m² h at 10 bar only (Fig. 8a). This variation is primarily due to the structural differences among the membranes. Despite the same amount of pore-forming agents being used across all three membranes, the P-CM membrane’s notably higher flux can be linked to its more porous morphology that provided more water channels allowing water to pass. The larger and thicker, as well as the more irregular morphology of MoS2 nanoplatelets in the membrane, contributed to this enhanced porosity.

In contrast, the T-CM membrane exhibited a significantly lower flux due to the tendency of nanosheets to stack atop one another; in addition, their thinner thickness formed the less pronounced narrow pores, thereby impeding the transport of water molecules through the membrane matrix. On the other hand, the individually presented nanospheres created less interconnected water passages compared to a more connected nanoplatelet membrane matrix.

Another factor contributing to the differing flux values could be the increased hydrophilicity of the P-CM membrane, as evidenced by decreased contact angles in measurements (Fig. 5d–f), facilitating a more rapid transport of water, thus improving membrane flux. Additionally, the physical arrangement of nanoplatelets might have created more effective flow channels, further augmenting the permeation rate. These findings suggested that the morphology and arrangement of nanomaterials within the membrane matrix played crucial roles in determining their water filtration performance. The flux trend observed among the three membranes reflected the porosity trend measured in the earlier section.

The nanomaterial embedded in the polymeric framework has introduced new functions such as catalytic decomposition and affected the membrane structure, thus making the resultant membrane not fit in the MF, UF, and NF classifications. Nevertheless, the permeability of the three membranes reported in this work was 6, 9, and 1.2 L/m² h /bar, which were comparable to or higher than some of the membranes reported for heavy metal removal.

The filtration operation in this study was different from the continuous flow in the practical water treatment process that feeds wastewater only passed the membrane once followed by discharge. Instead, the MoS2 nanocomposite membranes’ performance was tested by passing the feed solution multiple times in a cyclic mode at 10 bar applied pressure. pH value was measured; it was found that the pH was in the range of 6-7 throughout the experiment. The continued increase of removal efficiencies observed in the multiple cyclic process reflected the possible performance in a practical water treatment process. It was found that the removal of the first pass was in the range of 8.5 ± 0.5% - 20.0 ± 0.5% for Mn2+ and 5.0 ± 0.5% - 19.0 ± 0.5% for Zn2+ by MoS2 nanocomposite membranes. To achieve the higher removal efficiency, the feed was passed through the membrane the second time and third time until a total of seven passes, the accumulated total removal of 57.1 ± 0.5%–93.0 ± 0.5% and 54.2 ± 0.5%–90.4 ± 1.5% were achieved for Mn2+ and Zn2+ respectively. The P-CM membranes exhibited superior efficacy in removing Mn2+ and Zn2+ from aqueous solutions, with removal rates of 93.0 ± 0.5% and 90.4 ± 1.5%, respectively, significantly outperforming the S-CM (88.2 ± 1.0% for Mn2+, 83.4 ± 1.0% for Zn2+) and T-CM membranes (57.1 ± 0.5% for Mn2+, 54.2 ± 0.5% for Zn2+) (Fig. 6). This enhanced performance was attributed to the MoS2 platelets in the P-CM membranes, which provided a more negatively charged surface for attracting positively charged contaminants, and their unique structural features, like crumpled and fishnet structures (Fig. 1), that created nanovoids and nanopores. These structures facilitated water molecule passage and enhanced interactions with ROS, boosting the catalytic reaction. A neat membrane without any nanomaterials was fabricated and tested for Zn2+ and Mn2+ removal to represent the adsorption-only scenario for comparison. It can be seen (Fig. 6) that only 4.2 ± 0.1%- 6.3 ± 0.1% removal was achieved for both metal ions by the neat membrane. This removal was attributed to the adsorption properties of the polymeric chitosan in the membrane. It was noted that there was no increase in the removal percentage as the number of passes increased, confirming that the adsorption limit of the membrane had been reached.

Fig. 6: Removal efficiencies of heavy metals by MoS2 chitosan membranes.
figure 6

a Mn2+ removal: S-CM (88.2 ± 1.0%), P-CM (93.0 ± 0.5%), T-CM (57.1 ± 0.5%), and neat membrane (6.3 ± 0.1%). b Zn2+ removal: S-CM (83.4 ± 1.0%), P-CM (90.4 ± 1.5%), T-CM (54.2 ± 0.5%, and neat membrane (4.2 ± 0.1%). c Comparative bar chart of removal efficiencies for Mn2+ and Zn2+ across the membranes. P-CM shows notably higher removal rates, attributable to its MoS2 platelets’ structural features. Error bars represent percentage errors.

Full size image

A commercial NP010 membrane was tested, and it showed no removal in the first pass, and the highest removal obtained was 4% at the end of 7 passes. Results are given in the Supplementary file (Supplementary Fig. 2). When compared to other nanocomposite membranes reported in the literature, the P-CM membranes exceeded those reported membranes incorporating materials such as graphene oxide (GO)49, boron nitride (BN)50, and graphitic carbon nitride (g-C3N4)51, and showed exceptional performance in both pure water flux (1.3–2.6 times higher)52,53 and heavy metals rejection (1.2 times higher) than that reported previously54. These results highlighted the P-CM membrane’s potential as an efficient water purification solution, particularly for heavy metal contaminant removal. In addition to these advantages, the P-CM membrane in this study uniquely combines the adsorptive properties of the membrane with the catalytic decomposition of heavy metals. This combination offers several benefits compared to traditional adsorbents, which consist of loose particles that can be challenging to handle and regenerate. Furthermore, traditional adsorptive membranes often reach equilibrium or their maximum adsorptive capacity relatively quickly, limiting their effectiveness over time. In contrast, the catalytic properties of the P-CM membrane ensure sustained performance by continuously breaking down contaminants and preventing the membrane from becoming saturated. This dual functionality not only enhances the overall removal efficiency of heavy metals but also extends the membrane’s operational lifespan, making it a more reliable and durable solution for water purification applications.

Catalytic generation of reactive oxygen species

In the process of heavy metal removal by MoS2 nanoparticle-embedded membranes, a critical mechanism involves the catalytic decomposition of H2O2 by MoS2, leading to the formation of ROS, such as hydroxyl radicals (•OH). The generation of these ROS plays a pivotal role in the reduction of heavy metal ions, effectively transforming them into their elemental form, which can be captured by the membrane. This process is fundamentally characterized by electron transfer between the ROS and the heavy metal ions. The following simplified chemical equation can represent the interaction with hydroxyl radicals:

$${{\rm{H}}}_{2}{{\rm{O}}}_{2}\begin{array}{c}{\bf{Mo}}{{\bf{S}}}_{{\boldsymbol{2}}}\\ \longrightarrow \\ {\bf{Catalytic}}\end{array}2{{\rm{e}}}^{-}+2\cdot {{\rm{OH}}}^{-}+\frac{1}{2}{\rm{O}}$$
(1)
$${{\rm{Mn}}}^{2+}+2{{\rm{e}}}^{-}\to {\rm{Mn}}$$
(2)
$${{\rm{Zn}}}^{2+}+2{{\rm{e}}}^{-}\to {\rm{Zn}}$$
(3)

The unique tri-layered sulfur-molybdenum-sulfur structure55 with defects in the nanocrystalline of MoS2 significantly contributed to the enhanced generation of ROS. The substantial surface area of nanomaterials offered a multitude of interaction sites with H2O2, thereby amplifying the production of ROS. These ROS are instrumental in the reduction of heavy metal ions, facilitating their conversion into less toxic forms or species that are more amenable to separation from the aqueous environment. This interplay between MoS2’s properties, ROS generation, and heavy metal ion reduction underscores the synergistic functions of this composite membrane in advanced water purification applications.

To confirm the generation of ROS during the interaction between MoS2 and H2O2, both types and quantities of MoS2 were varied in reaction with H2O2, and coumarin. Fluorescence emission spectrometry was utilized to detect the emission intensity, and the resulting curves are presented in Fig. 7 below. An increase in MoS2 volume from 2.5 µl to 10 µl in 10 ml solution led to a corresponding increase in peak emission intensity, indicating a higher generation of ROS with increased MoS2 quantity (Fig. 7a). This trend was consistent across MoS2 nanoplatelets (Fig. 7b) and MoS2 nanosheets (Fig. 7c), however, it is observed that three nanomaterials displayed different fluorescence peak intensities with decreasing in the order of nanoplatelets > nanospheres > nanosheets. This trend aligned with the heavy metal removal efficacy observed in Fig. 6c.

Fig. 7: Fluorescence intensity confirming ROS generation in MoS2-H2O2 reaction.
figure 7

The bar chart displays fluorescence peak intensity at 666 nm for varying concentrations of MoS2. a MoS2 nanospheres: Increase in intensity with MoS2 volume from 2.5 µl to 10 µl. b MoS2 nanoplatelets: Similar trend, showing higher peak intensities compared to nanospheres. c MoS2 nanosheets: Increment in fluorescence intensity, but lower than other forms. Insets illustrate intensity vs wavelength peak variation. The fluorescence intensities correlate with the observed heavy metal removal efficiencies, with nanoplatelets exhibiting the highest efficiency.

Full size image

As known, a membrane is generally defined as a selective barrier between two phases and offers a different way of operation from the adsorption column; for example, the membrane has a more intact polymeric structure that can take higher applied pressure to achieve better separation of contaminant from water56. The membrane’s morphology or structure can determine which separation mechanisms will dominate, such as sieving, adsorption, and Donnan exclusion. Incorporating nanomaterials such as MoS2 enhances the adsorption function while maintaining the membrane polymer framework; both sieving and adsorption mechanisms work together. All adsorption columns use a lot more adsorbent particles to fill the 3D column. In contrast, one membrane sheet (100 µm thickness) uses only a fraction or less material mass. It can achieve the equivalent results of a column, particularly when the membrane offers combined adsorption and catalytic decomposition functions. Membrane technology represents a potentially less material and more sustainable water treatment technology. An adsorptive-catalytic membrane also has limitation compared to an adsorption column, i.e. the treatment capacity is much smaller, we aim to resolve this by a few strategies in the future research such as incorporating more nano-additives in the membrane or fabricate thicker membrane.

Membrane regeneration and re-use

The spent membrane after the multiple cycle experiments was subject to chemical cleaning (no cleaning was performed between each pass during the experiment). The cleaned membranes demonstrated impressive flux recovery and durability post-testing and chemical cleaning by H2O2: S-CM achieved 93.0 ± 1.0%, P-CM a slightly higher 96.0 ± 0.5%, and T-CM the lowest at 87.5 ± 2.0%. The DI water-cleaned membranes also exhibited comparable flux recoveries that 92.0 ± 2.0% for S-CM, 96.0 ± 1.0% for P-CM, and 87.5 ± 1.0% for T-CM were achieved (Fig. 8b). Despite variations in heavy metal removal rates and water flux, all membranes showed remarkable mechanical and chemical stability, maintaining structural integrity even under high pressure and extended periods of use, with no swelling observed. They also exhibited resilience to chemical cleaning, indicating their suitability for repeated use without degradation. The ease of regeneration, achieved by simple filtering with H2O2 in water, further emphasized their practicality in operation and maintenance. This combination of stability, durability, and ease of regeneration made these MoS2-incorporated membranes promising options for long-term water and wastewater treatment applications.

Fig. 8: Initial and recovered DI water flux comparison.
figure 8

a DI water flux of the three nanocomposite membranes; b flux recoveries by H2O2 chemical cleaning and by DI water cleaning.

Full size image

Methods

Chemicals and reagents used in nanoparticle synthesis, membrane fabrication, characterization, and analytical investigations were of high scientific purity and employed without any alterations. Notably, l-cysteine (97%), sodium molybdate dehydrate (99%), hydrochloric acid (37%), chitosan (MW 30000), polyethylene glycol (PEG) (Mn 400), and coumarin were procured from Sigma-Aldrich. Additionally, heavy metal salts, including zinc (II) chloride and manganese (II) chloride, were sourced from Sigma Aldrich. The solvents employed in this study, namely ethanol (99.7%), acetic acid (99%), sodium hydroxide, and hydrogen peroxide (35%), were obtained from Honeywell.

Synthesis of MoS2 nanomaterials with different morphologies

Firstly, MoS2 nanospheres were synthesized using a modified bottom-up green hydrothermal process57 (Fig. 9). l-cysteine (0.605 g) was dissolved in 10 ml DI water and acidified by adding 5 ml 10 M HCl to pH < 1. Sodium molybdate dehydrate (0.29 g) was then added dropwise and stirred for 30 min at room temperature. The solution was under hydrothermal reaction at 200 °C for 30 h, yielding dark brown MoS2 nanospheres. After cooling, the nanospheres were washed alternately with DI water and ethanol five times to remove byproducts and then stored in DI water.

Fig. 9
figure 9

A Comprehensive schematic illustration of the hydrothermal synthesis process for producing diverse MoS2 nanoparticle morphologies and fabricating three nanocomposite membranes (S-CM, P-CM, and T-CM).

Full size image

Secondly, MoS2 nanoplatelets were synthesized using a similar approach, but the pH was adjusted to 2 using 1 M HCl, unlike nanospheres synthesized at pH < 1. The resulting MoS2 nanoplatelets slurry was washed and stored for further use.

Thirdly, a much smaller amount of l-cysteine (0.1005 g) and sodium molybdate dihydrate (0.0485 g) were employed to produce thin MoS2. After vigorous mixing at 90°C, the pH was adjusted to 2 with 1 M HCl before transferring the solution to a hydrothermal reactor at 210°C for 24 h—the resulting gray-colored solution contained uniformly suspended particles, sonicated before use. Figure 9 illustrates the preparation process.

Fabrication of nanocomposite adsorptive-catalytic membrane

Fabrication of MoS2 nanosphere-chitosan nanocomposite adsorptive-catalytic membrane (S-CM) involved in blending nanomaterials into a polymer solution with a pore-forming agent, then followed by phase inversion. For S-CM, chitosan (1.6 g) was mixed with a 10 wt% acetic acid solution (30 ml), PEG (1.6 g), and the full batch of synthesized MoS2 nanospheres. After thorough mixing and sonication to remove air bubbles, the solution was cast into membranes using a film-casting knife to achieve a 210 µm thickness. The membranes were air-dried for 30 min, oven-dried, neutralized with 2 M NaOH, rinsed with DI water, and cut to size for testing. The synthesized MoS2 nanoplatelets and MoS2 nanosheets were used to fabricate the MoS2 nanoplatelet membrane (P-CM) and MoS2 nanosheet membrane (T-CM) following the same procedure, respectively. This work aims to incorporate more quantity of MoS2 nanoparticles in the membrane to create significant effects, the reported concentration equivalent to the whole batch quantity of synthesis, each synthesis involves a hydrothermal reaction followed by extensive cleaning by water and ethanol alternately. It was found that adding more than this whole batch quantity of MoS2 will further reduce the water flux, and it is not a positive outcome. Table 1 reflects the composition information of the three membranes fabricated.

Table 1 Composition of the S-CM, P-CM, and T-CM adsorptive-catalytic membranes
Full size table

Nanomaterials characterization

To assess the structural differences among the three synthesized nanomaterials, the morphological differences and size variations, as well as their respective crystallinity in the three nanomaterials, were investigated using Titan 80-300 ST transmission electron microscope (TEM) (Thermo Fisher Scientific Inc.). The TEM analysis was conducted with an accelerating voltage of 300 kV. Bright-field (BF) TEM imaging was selected to capture both low magnification (LM) and high magnification (HM) images, allowing for the examination of the morphology and structure of the MoS2 nanoparticles, respectively. Additionally, structural analysis of the nanoparticles of three samples was conducted utilizing the selected area electron diffraction (SAED) technique. The SAED technique was constrained by the presence of numerous nanoparticles contributing to the diffraction pattern, primarily due to the relatively large size of the illuminated area. This characteristic posed challenges for the individualized examination of each nanoparticle. Therefore, the Fast Fourier Transform (FFT) was used in this study to address the crystallinity of the materials.

The nanomaterial dimensions and thickness were evaluated and compared using atomic force microscopy (AFM). AFM Concept Scientific Instrument (France) was used and functioned in non-contact resonant mode. The zeta potential (charge) of the nanomaterials was inspected using Zetasizer (Nanoseries ZS from Melvern).

Membrane characterization

The three fabricated membranes, S-CM, P-CM, and T-CM were examined by SEM to study the surface and cross-sectional morphologies. The membrane surface roughness and height profile were also studied using AFM. The S-CM, P-CM, and T-CM membranes’ surface charge was analyzed using a SurPASS Electrokinetic Analyzer. Measurements were conducted with 0.001 M KCl electrolyte, adjusting the initial pH to around 4 using 0.2 M HCl. Titration using 0.1 M NaOH was performed to adjust pH to 14, with four readings taken and averaged per pH point. The hydrophilicity of the membranes was analyzed by measuring the water contact angles using a goniometer (Krüss Easy Drop Analyzer, Germany) equipped with Drop Analyzer software. Membrane porosity (Ɛ) (Eq. 4) is determined by soaking them in water for an hour and measuring the volume of water absorbed by the membrane.

$${\rm{\varepsilon }}=\frac{{\boldsymbol{Volume\; of\; pores}}}{{\boldsymbol{Total\; volume\; of\; membrane}}}\times 100 \%$$
(4)

Where the volume of pores is the volume of water absorbed. The total volume of the membrane was calculated after measuring the diameter and height of the membrane.

Membrane performance evaluation

The clean water flux of the membrane was tested by using a stirred cell (HP-4750 Supplementary Fig. 1), and the membranes were subjected to a constant transmembrane pressure of 10 bar. Flux was determined only using clean DI water as feed, the flux was measured based on the volume of permeate produced per unit membrane area for 10 min using the same mathematical approach as ref. 46.

The same pressure, feed volume, and membrane conditions were maintained in each multiple passes experiment, so all the experiments were conducted under the initial flux similar to DI water flux. It was observed that the flux gradually declined with the progression of the experiments. Permeability was measured by dividing the obtained flux by pressure46. This measurement method provides a comparable and consistent means of assessing membrane performance. The treatment flux was not measured during the heavy metal removal experiments due to multiple filtration cycles being conducted.

The removal of heavy metal ions of Zn2+ and Mn2+ by the three membranes was evaluated, and the effects of MoS2 nano-additives with different morphologies were compared. A neat membrane without any nanomaterials, as well as a commercial polymeric membrane (NP010), were also tested for heavy metal ions removal to compare the results with the MoS2 nanocomposite adsorptive-catalytic membranes. Each membrane was placed in a stirred dead-end cell and the test was carried out with a 50 ml feed solution consisting of 5 ppm of Zn2+ or Mn2+ ions. In addition, 30 mmol H2O2 was introduced to the feed to chemically react with the MoS2 and generate ROS. Permeate samples (approximately 4 ml) were collected at fixed time intervals by maintaining a constant transmembrane pressure of 10 bar. Once 50 ml of permeate had passed though the membrane, about 4 ml sample was collected from this for analysis and the remaining was poured back to the cell to pass through the membrane again. This process was done until the feed passed through the membrane 7 times. Permeate samples were collected at each pass. Concentrations of metal ions in both feed and permeate were quantitatively determined by using Inductively Coupled Plasma- Optical Emission Spectroscopy -(ICP-OES). The removal efficiencies of heavy metal ions were then calculated by the method as in ref. 46. After the filtration experiment of the first heavy metal Zn2+ ions removal, the membrane was thoroughly cleaned and re-used for the experiments of second heavy metal Mn2+ ions removal. The membrane was only chemically clean after the 7th pass at the end of this experiment. These test procedures were repeatedly followed for all three membranes.

Confirmation of the generation of reactive oxygen species

To confirm the ROS generation from the catalytic interaction between MoS2 nanoparticles in the membranes and H2O2; coumarin was used as a probe molecule for trap** (•OH) radicals followed by photoluminescence (PL) emission spectra analysis58,59. Initially, MoS2 nanospheres, nanoplatelets, or nanosheets were mixed with coumarin, followed by the addition of H2O2. Analysis was performed using a Perkin Elmer LS-55 Fluorescence Spectrometer with an excitation wavelength of 330 nm and a 16.2 W LED lamp emitting radiation in the 350 to 650 nm range. The impact of the three nanomaterials on ROS generation and the effects of varying MoS2 concentrations were investigated. Coumarin (11.5 ppm, 10 ml) was mixed with MoS2 solution, followed by adding H2O2 (10 µl) and stirring. MoS2 amounts varied from 2.5 µl to 10 µl. The emission intensity peaks from the fluorescence spectrometer were compared for different types and amounts of MoS2 nanomaterials.

Membrane cleaning and re-use

The used membrane was only chemically cleaned at the end of each experiment (after the 7th pass). The three fabricated membranes of S-CM, P-CM, and T-CM underwent sequential multiple experiments including water flux tests, metal ion removal studies, chemical cleaning, subsequent water flux tests, and repeated metal ion removal studies. Chemical cleaning involved passing 250 ml of 30 mmol H2O2 through the membrane, followed by rinsing with DI water. The effect of DI water-only cleaning, without chemical cleaning, was also studied. The membranes were subjected to thorough DI water cleaning followed by re-testing. The difference between the initial and final water flux values indicated the flux recovery rate of the membranes and was calculated as in ref. 46.