Introduction
The growing demand for effective wastewater management strategies is underscored by the increasing volume of saline wastewater generated from various industrial and clinical operations, including those in dental practices [1,2]. Dental practices generate wastewater laden with sodium chloride, fluoride, and other contaminants, creating a pressing need for effective treatment solutions [3]. Traditional wastewater treatment methods may prove to be inefficient when it comes to handling saline effluents [4–6]. One promising technology is membrane distillation (MD), which employs hydrophobic membranes to enable the separation of water vapor from saline feeds without the direct passage of liquid water. The process typically occurs at relatively low temperatures, making it energy-efficient compared to other evaporation-based methods [7–9].
The efficacy of MD relies heavily on the properties of the membranes employed [10]. Superhydrophobic membranes, characterized by their remarkable water-repelling capabilities, provide significant advantages for this application [11]. Such membranes exhibit water contact angles exceeding 150 degrees, which effectively inhibits water intrusion into the membrane structure, enhancing the distillation process’s efficiency [12, 13]. The preparation of superhydrophobic membranes typically involves a combination of material selection, surface modification techniques, and the incorporation of hierarchical structures that promote non-wetting characteristics [14].
Various materials are suitable for the fabrication of super-hydrophobic membranes. Polymeric materials such as polytetrafluoroethylene (PTFE), polyvinylidene fluoride (PVDF), and polypropylene (PP) are commonly utilized due to their inherent hydrophobicity and chemical stability [15,16]. For instance, the surface of PP membranes can be treated with fluorinated compounds or silanes, creating a layer that imparts additional hydrophobicity [17]. Such modifications enable the surfaces to assemble unique molecular structures that repel water while maintaining desired permeability for water vapor. The incorporation of nanostructured materials, such as silica nanoparticles or carbon nanotubes, can further enhance the surface roughness, contributing to the desired superhydrophobic behavior by promoting the Cassie-Baxter state [18, 19]. Surface modification techniques play a crucial role in achieving superhydrophobicity. Chemical treatments involving fluorination or salinization can improve the hydrophobic nature of membrane surfaces [20, 21]. For instance, applying fluorinated silanes can significantly enhance surface energy characteristics and promote the essential roughness required for superhydrophobic behavior [22].
On the other hand, inorganic scaling adds resistance to mass and heat transfer by forming a scaling layer on the membrane surface and clogging pores [23–26]. Once scaling occurs, MD performance quickly declines due to pore blockage, surface obstruction, and potential membrane damage [27–30]. Superhydrophobic membranes can enhance fouling resistance in water treatment. Their non-wetting properties significantly reduce foulant adhesion, enabling longer operation without frequent cleaning or maintenance [31]. This durability increases the economic viability of membrane distillation systems, making them an attractive choice for scaling up saline wastewater treatment applications.
This study aimed to develop superhydrophobic membranes for membrane distillation to treat saline dental wastewater, enabling clean water recovery and promoting sustainable practices in dentistry while minimizing environmental waste. We first prepared four superhydrophobic membranes and analyzed their characteristics such as hydrophobicity, surface morphology, roughness, pore size distribution, and functional groups. Then, we examined how saline concentrations and calcium ions affect normalized flux and salt rejection during the MD process, clarifying the anti-fouling mechanism.
Materials and Methods
2.1. Materials
The commercial pristine membranes used in this study were PTFE membranes with a pore size of 0.2 μm and a thickness of 50 μm, purchased from the Shengju company in China. The analytical reagents sodium chloride, calcium chloride, hydrochloric acid, glutaraldehyde (GA), ethyl silicate, hexamethyl triethoxysilane, ethoxyacetic acid, perfluorodecyltriethoxysilane, absolute ethanol, and polyvinyl alcohol (PVA) were purchased from Damao Reagent Co., Ltd., China. Silica nanoparticles were purchased from XF-nano Reagent Co., Ltd., China, with the diameter of 10 nm ~20 nm. High-salinity wastewater containing sodium chloride and calcium chloride were used as feed in this study.
2.2. Preparation of Superhydrophobic Membranes
The preparation process of silica modified superhydrophobic membranes was as follows: the ethyl silicate, hexamethyl triethoxysilane, and deionized water were diluted in an ethanol solvent and the pH was adjusted with diluted hydrochloric acid. This mixture was stirred in a water bath at 60°C for 1 hour. Following this, a solvent exchange was performed using a mixture of anhydrous ethanol and ethyl silicate for 48 hours, and then the product was dried at room temperature and atmospheric pressure to obtain hydrophobic SiO2 aerogel. The aerogel was then sonicated for 30 minutes and loaded onto a PTFE membrane using vacuum filtration. The loading amounts were 0.4 mg/cm2, 0.5 mg/cm2, and 0.6 mg/cm2, resulting in three types of superhydrophobic membranes, labeled SM1, SM2, and SM3 after drying (Fig. S1).
The preparation of fluorinated superhydrophobic membrane involves mixing perfluorodecyltriethoxysilane and ethyl silicate in a volume ratio of 1:2, stirring for 15 minutes until homogeneous. Silica nanoparticles were then slowly added to the mixed solution for in-situ fluorination. Next, 4 mL of fluorinated solution was combined with 25 mL of ethanol and stirred for 30 minutes to form a uniform SiO2 dispersion. Afterward, 2 mL of PVA solution was added, and the mixture was stirred for another 10 minutes to obtain a PVA-containing SiO2 dispersion. This dispersion was loaded onto a PTFE membrane using vacuum filtration. The membrane was thoroughly rinsed with deionized water and dried at 90°C for 10 minutes. A 100 mL mixture of 0.5% glutaraldehyde and 1% hydrochloric acid was prepared, and the dried membrane was placed in this solution at 70°C for 1 hour to crosslink. After removal, it was rinsed thoroughly with pure water and vacuum dried at 90°C for 10 minutes, resulting in the fluorinated super-hydrophobic membrane, designated as SMf.
2.3. Membrane Characterization
Water contact angles were performed with a contact angle goniometer (OCA15EC). A scanning electron microscopy (SEM, JSM-7001F) was used to observe the membrane surface morphology. Atomic force microscopy (AFM, FASTSCANBIO) was used to measure membrane surface roughness. Attenuated total reflectance Fourier transform infrared spectroscopy was performed to analyze the functional groups of membranes. The pore size distribution of the pristine and superhydrophobic membranes were tested using a 3H-2000PB pore size analyzer.
2.4. Experimental Method
The MD system features a flat membrane with an effective area of 40 cm2 on the feed side. A peristaltic pump (BT601S, Leadfluid, China) circulates the feed, while a vacuum pump (XZ-4B, Linhai, China) and a condenser create negative pressure and collect the permeate, respectively (Fig. S2). The experiment was conducted at a vacuum level of 0.095 MPa, a feed temperature of 35°C, a condensate temperature of 3°C. The feed flow rate was set at 6 cm/s (Fig. S3).
The flux and the salt rejection were calculated as:
where J represents the flux (kg/(m2·h)), Δm is the mass reduction of the feed (kg), S is the effective membrane area (m2), t is time (h), R is the rejection of salt (%), and cf and cp are the conductivities of the feed and permeate (mS/cm), respectively.
Results and Discussion
3.1. Membrane Characterization
3.1.1. Hydrophobicity of membranes
Fig. 1 showed the water contact angles of four superhydrophobic membranes. All modified membranes exhibited water contact angles greater than 140°. As the SiO2 loading increased, the water contact angle of the superhydrophobic membranes also increased, with the SM3 membrane reaching a contact angle of 161.1° (Fig. 1c), indicating its excellent hydrophobicity. The contact angle of the SMf membrane was 158.6°, also demonstrating good hydrophobicity, though it was slightly lower than that of the SM3 membrane. This may be attributed to the introduction of fluorides, which alters the arrangement of SiO2 particles on the membrane surface.
3.1.2. Surface morphology
As shown in Fig. 2, the SiO2 adhered to the PTFE surface forms a dense hydrophobic layer. The membranes’ surface was uneven and exhibits a regular stacking structure. This is due to SiO2 bonding through Si-O-Si linkages, with the unique amorphous structure of β-SiO2 and its random arrangement integrating into the fibrous structure of the PTFE membrane. The bonding between SiO2 and the PTFE membrane is relatively tight, with tiny gaps between the SiO2 particles (Fig. 2a). As the loading amount increases, channels of varying sizes appear on the membrane surface, leading to a multilayered coverage structure (Fig. 2b). Subsequently, the amount of SiO2 supported within the limited fibrous space of the PTFE membrane increases, gradually revealing a layered structure, with noticeable stacking and agglomeration of the nano-sized SiO2 particles (Fig. 2c), resulting in particle sizes that can reach the micrometer range. The Mf membrane surface forms a network structure (Fig. 2d), which not only preserves the superhydrophobic properties of the membrane surface but also facilitates enhanced membrane permeability.
3.1.3. Surface roughness
Roughness in membranes can affect both the contact angle and the effective surface area. The roughness of the SM1, SM2, and SM3 membranes was 95 nm (Fig. 3a), 105 nm (Fig. 3b), and 108 nm (Fig. 3c), respectively. This represents an increase in roughness compared to the PTFE membrane, as the attachment of SiO2 nanoparticles enhances the curved structure of the SiO2 spherical particles on the membrane’s surface. The SMf membrane exhibited a roughness of 174 nm, significantly higher than the other three membranes (Fig. 3d). A rougher membrane surface typically increases surface area due to the micro-scale features that characterize the membrane. This increased surface area can enhance the evaporation rate, as a larger area facilitates more vapor formation per unit time.
3.1.4. Pore size distribution
The pore size distribution showed that, compared to the pristine membrane, the average pore size of all hydrophobic membranes has decreased (Fig. 4a). This change was beneficial for enhancing the membrane’s antifouling performance. The smallest pore size distribution of the SM3 membrane was already below 60%, indicated that it has a moderate pore size and average pore size distribution. This is because, with the increase in SiO2 loading, the surface of the membrane has become characterized by an uneven pore size distribution. The smallest pore size of the SMf membrane has only seen a slight decrease compared to the pristine membrane. This can be attributed to the fluorinated particles forming a dual hydrophobic layer on the fibrous PTFE, without significantly entering the membrane pores.
3.1.5. F9unctional groups on the membrane surface
The infrared spectrogram showed that silica peaks were concentrated in the 1000 cm−1 to 1200 cm−1 and 800 cm−1 to 1000 cm−1 regions (Fig. 4b). The peaks in the 800 cm−1 to 1000 cm−1 range arise from the expansion and contraction vibrations of silicon- oxygen bonds, with a Si-O-Si antisymmetric stretching peak near 1090 cm−1 confirming the successful incorporation of SiO2 particles into the membrane. Additionally, peaks between 700 cm−1 and 850 cm−1 result from the planar rocking and scissor bending vibrations of C-H in -CH3, indicating the presence of hydrophobic -CH3 groups on the membrane and further evidencing the loading of hydrophobic SiO2 particles [32]. A small peak in the higher frequency range of 2640 cm−1 to 2350 cm−1 corresponds to -OH vibrations, while the characteristic absorption peak at 2400 cm−1 represents the expansion vibration of -COO- and -OH bonds in hydroxyl groups.
3.1.6. Liquid entrance pressure of the membranes
The liquid entrance pressure (LEP) of the membranes was measured using a pressure device (Fig. S4). The device connects to the steel membrane filtration assembly via a nitrogen cylinder. The membrane is positioned on the module, and deionized water is added to fully cover it. The nitrogen cylinder’s pressure divider is then gradually opened, determining the membrane’s liquid entry pressure by observing for any liquid overflow on the non-deionized water side. When there is a liquid spillage, the pressure value of the nitrogen cylinder released to the device at this time is recorded, which is recorded as the LEP of the membrane. The LEP values of the PTFE membrane, SM1, SM2, SM3, and SMf were 2.12, 2.42, 2.51, 2.74, and 2.66 bar, respectively. The results showed that the modified membrane’s LEP value has significantly improved, enhancing both the wettability of membrane distillation and the stability of MD.
3.2. Desalination Performance Using Superhydrophobic Membranes
3.2.1. Influence of saline concentrations
Fig. 5 showed the normalized flux and salt rejection of the PTFE membrane and four types of superhydrophobic membranes when treating salt containing dental wastewater. When the salt concentration was 5 g/L, the normalized flux of the PTFE membrane sharply decreased to 0.82 within 4 hours, with a continued slow decline thereafter. After 20 hours, it dropped to 0.74 (Fig. 5a), while the salt rejection reached 99.41% (Fig. 5d). This decline was attributed to the adhesion of salts from the feed solution to the surface of the PTFE membrane, increasing mass transfer resistance and reducing the area of the membrane available for steam to pass through, thereby resulting in a rapid decrease in normalized flux.
For the superhydrophobic membranes, the SM1 membrane showed a gradual decline in normalized flux due to its lower SiO2 load, decreasing to 0.85 after 20 hours. Both SM2 and SM3 membranes maintained a normalized flux around 0.95 after 20 hours (Fig. 5a). This suggests that the SiO2 particles loaded onto the surface of the superhydrophobic membranes can prevent salt crystallization. As the SiO2 loading increased, the channels formed by the membrane surface became smaller, leading to improved salt rejection compared to the PTFE membrane. The salt rejection of the superhydrophobic membranes increased significantly, all exceeding 99.92%. The normalized flux of the SMf membrane remained at 0.99 after 20 hours with no declining trend, while its salt rejection peaked at 99.99% (Fig. 5d).
When the salt concentration increased to 10 g/L, the normalized flux of the PTFE membrane fell to 0.68 after 20 hours, with a noticeable decrease in salt rejection rate to 99.04%. The trend in normalized flux for the four superhydrophobic membranes was similar to that observed at 5 g/L (Fig. 5b), with the SMf membrane still achieving a salt rejection of 99.97% (Fig. 5e). As the NaCl concentration rose to 35 g/L, the PTFE membrane exhibited a rapid decline in normalized flux after just 2 hours (Fig. 5c). This can be explained by the higher concentrations of NaCl more readily contaminating the membrane surface, even allowing salts to penetrate to the permeate side, reducing the rejection rate to 98.15% (Fig. 5f), indicating the PTFE membrane’s limited ability to suffer high salt concentrations. The normalized flux of the SMf membrane remained at 0.98 after 20 hours (Fig. 5c), with a salt rejection rate reaching 99.98% (Fig. 5f). Thus, it can be inferred that this membrane demonstrates good anti-fouling properties when treating high-salinity water, enabling the production of high-purity water [33].
3.2.2. Influence of saline containing calcium
With a feed containing 5 g/L of NaCl and 1 g/L of CaCl2, the normalized flux of the PTFE membrane decreased rapidly, falling to 0.77 after 5 hours. In contrast, the superhydrophobic membrane maintained the highest normalized flux at 0.97 after 20 hours, while the lowest flux of the SM1 membrane remained stable at 0.85 (Fig. 6a). The PTFE membrane achieved a salt rejection of 99.27%, whereas all superhydrophobic membranes exceeded 99.9%, with the SMf membrane reaching 99.99% (Fig. 6b). This superior performance is due to the excellent antifouling properties of the superhydrophobic membrane.
Fig. 7a illustrated that salt crystallization on the PTFE membrane surface and within fiber pores leads to a layered buildup, narrowing the vapor transport channel and increasing mass transfer resistance, which reduces membrane flux. In contrast, SiO2 particles on the superhydrophobic membrane surface remain visible, with no significant salt crystallization (Fig. 7b, Fig. 7c, Fig. 7d), highlighting the membrane’s effective anti-fouling properties. The sparse SiO2 particles on the SMf membrane were uniformly arranged in strips, displaying staggered dendritic shapes (Fig. 7e). This structure promotes a Cassie-Baxter condition, limiting points for salt accumulation. Thus, salt particles that did not fully adhere to the membrane were carried away by the feed solution, helping to maintain stable hydrophobic performance.
The XDLVO theory elucidates the superior membrane flux and salt rejection of the prepared superhydrophobic membranes. The XDLVO model also incorporates Lewis acid-base interactions and hydrophobic forces between the membrane and fouling materials [34]. Membrane fouling interactions are primarily influenced by electrochemical bilayer and van der Waals interactions. The water contact angle of the superhydrophobic membrane is significantly higher than that of the PTFE membrane, resulting in a lower polar surface tension component and indicating weaker polar forces compared to PTFE. Additionally, PTFE membranes contain electronegative fluorine atoms and exhibit a negative zeta potential. Following the modification with fluorinated SiO2 nanoparticles, the fluorine content on the SMf membrane’s surface increased, leading to a decreased zeta potential and enhanced negative charge. This contributes to the excellent fouling resistance of SMf membranes when processing saline.
3.3. Anti-fouling Mechanism of Superhydrophobic Membranes
The anti-fouling mechanism of superhydrophobic membranes prepared in this study includes:
Low nucleation propensity. Heterogeneous and homogeneous nucleation significantly influence molecular dynamics. The distinct structure of superhydrophobic surfaces diminishes the tendency for inorganic salts to nucleate in solution [35,36]. At the solid-liquid interface, heterogeneous nucleation occurs more rapidly than homogeneous nucleation, while at the gas-liquid interface, their crystallization energies are comparable and the air’s surface energy approaches zero. In this study, the nucleation tendency on the PTFE membrane surface was higher than on the SM membrane and at the gas-liquid interface (Fig. 8a, Fig. 8b), indicating that the SM membrane surface presents a greater nucleation challenge. The lower nucleation tendency of the superhydrophobic membrane can be attributed to its reduced contact area with the feed liquid and increased gas-liquid interface. Additionally, the unique spherical structure of SiO2 further minimizes the contact area between the feed liquid and the membrane surface, resulting in a significantly lower crystallization tendency compared to PTFE membranes.
Short solid-liquid contact time. To further clarify the mechanism, we experimentally tested the water slip angles of membranes. The PTFE membrane did not slip at inclination angles exceeding 20° (Fig. S5a), while the slip angles of SM1, SM2, SM3, and SMf membranes were 2.1° (Fig. S5b), 2.0° (Fig. S5c), 1.8° (Fig. S5d), and 0.5° (Fig. S5e), respectively. The superhydrophobic membrane’s minimal slip angle resulted in a very short contact time with the feed cycle. This phenomenon, known as the “interfacial slip effect,” means that when this effect is absent, the relative velocity between the liquid and solid is zero [37]. However, with this effect, this velocity becomes greater than zero. Under continuous feeding in membrane distillation, the reduced solid-liquid contact time decreases the ion residence time on the super-hydrophobic membrane’s surface, thus lowering the nucleation probability and influencing scaling control. In contrast, PTFE membrane did not exhibit the interfacial slip effect, leading to longer contact times with the feed (Fig. 8c). Additionally, fluid flow on the membrane surface is impeded by nucleation mechanisms, causing salt ions in the feed to accumulate and form crystal nuclei. These crystalline ions can adhere to the nuclei and gradually enter the membrane pores during the distillation process, leading to partial wetting failure of the hydrophobic membrane. Overall, the interfacial slip effect of the superhydrophobic membrane not only shortens the heterogeneous nucleation time at the liquid-solid interface but also reduces the contact time between crystalline nuclei in the liquid and the membrane.
Less solid-liquid contact area. The superhydrophobic membrane’s surface structure allows an air layer to form between the feed and the membrane, resulting in initial contact with the air instead of the membrane itself. This reduces the contact area between the feed and the membrane. For superhydrophobic membranes with a water contact angle of 160°, the droplet’s direct contact with the membrane is just 5.1%, compared to 57.8% for a PTFE membrane with a 128° contact angle [38]. The reduced contact fraction makes the superhydrophobic membrane less vulnerable to water molecule penetration, thereby decreasing fouling from salt accumulation in the membrane pores. Additionally, the reduced contact area minimizes attraction between the feed material and the membrane, enabling droplets to roll off more easily. This is evidenced by the superhydrophobic membranes’ slip angle of less than 2.1°, while the PTFE membranes exceed 20°. In the Wenzel state, when the membrane surface contacts droplets, the droplets become anchored in the membrane pores. When the feed containing Ca2+, this anchoring enhances the growth and crystallization of Ca2+ within the membrane, significantly reducing the boundary slip effect that obstructs feed movement. Additionally, the increased liquid-solid contact area fosters more opportunities for heterogeneous nucleation, crystal deposition, and adhesion [39, 40].
Conclusions
This study concludes that superhydrophobic membranes, created from silica-modified commercial PTFE membranes, effectively treated saline dental wastewater through membrane distillation. These superhydrophobic membranes demonstrated superior salt rejection compared to the pristine PTFE membrane and enabled prolonged operation by alleviating membrane scaling. In MD experiments with saline dental wastewater, membranes SM2, SM3, and SMf demonstrated stable flux and over 99.9% salt rejection at 5 g/L NaCl over 20 hours. At 10 g/L NaCl, SM2 and SM3 exhibited reduced performance after 10 hours due to increased brine concentration. When treating wastewater with Ca2+, the superhydrophobic membrane maintained stable flux for 18 hours with over 99.9% salt rejection, benefiting from its Cassie-Baxter structure. Anti-fouling mechanisms included reduced nucleation, shorter solid-liquid contact time, and minimized contact area.
PDF Links
PubReader
Full text via DOI
Download Citation 
Supplement
Print 
