Fabrication of Catalytic Distillation Membranes with Atomic Layer Deposition

Abstract

The integration of catalysts onto the surface of membranes enables simultaneous physical separation and catalytic transformation of constituents in a feed stream, facilitating improved contaminant removal and fouling mitigation. Distillation membranes are a particularly attractive platform for catalytic membranes because they reject nonvolatile species and exhibit exceptional resistance to oxidative and radical-driven degradation. However, imparting catalytic functionality onto hydrophobic, porous distillation membranes has proven challenging since the membranes used are chemically inert and difficult to modify. Furthermore, catalysts on the membrane surface can decrease hydrophobicity and increase the membrane’s susceptibility to pore wetting and failure. In this work, we create a catalytic distillation membrane by coating a polytetrafluoroethylene membrane surface with titanium dioxide (TiO₂) via plasma-assisted atomic layer deposition (ALD). By precisely tuning the ALD parameters, we demonstrate localized growth of TiO₂ near (within approximately 1 μm) the surface of polytetrafluoroethylene membranes, forming a catalytically active interface while preserving the underlying hydrophobic pore structure. Localized growth of TiO₂ is confirmed by electron microscopy and spectroscopy techniques, and membranes coated with 500 cycles of ALD show pressure tolerance up to 12.8 bar and higher than 95% salt rejection in pressure-driven distillation. Photocatalytic activity is demonstrated via the degradation of methylene blue dye under UV irradiation, where increasing TiO₂ loading leads to an enhancement in dye degradation. These results establish a general strategy for integrating catalytic functionality into chemically inert, hydrophobic membranes without compromising distillation performance, providing a pathway toward multifunctional membranes that couple advanced oxidation with membrane separation for water treatment.

Introduction

Increasing global water scarcity is driving the treatment of more complex and contaminated water sources, such as municipal and industrial wastewaters. − In these waters, trace organic contaminants have emerged as a major challenge for advanced treatment systems due to their persistence, negative health effects, and resistance to conventional treatment processes. − At the same time, the high organic content of many impaired waters increases fouling propensity in membrane-based advanced treatment processes, leading to performance losses, shortened membrane lifetimes, and increased operational costs. − These challenges motivate the development of treatment technologies that can simultaneously achieve high selectivity, robust operation, and in situ degradation of problematic organic compounds.

Catalytic membrane processes have attracted growing interest as a strategy for coupling physical separation with chemical transformation. In these systems, membrane filtration, such as nanofiltration or reverse osmosis, is combined with catalytic reactions at the membrane surface to generate reactive species, such as hydroxyl radicals, sulfate radicals, or other oxidants, that can degrade organic contaminants. − Such hybrid approaches offer the potential to reduce chemical dosing requirements, mitigate fouling through oxidative cleaning, and enable compact treatment architectures; however, existing catalytic membrane platforms face limitations. Reactive oxygen species generated during advanced oxidation can chemically degrade polymeric membrane materials, particularly polyamide active layers used in nanofiltration and reverse osmosis. − In addition, catalytic coatings and embedded catalytic nanoparticles frequently reduce effective membrane permeability by blocking transport pathways or increasing hydraulic resistance. , Finally, oxidation of larger organic molecules often produces low-molecular-weight byproducts that readily pass through conventional membranes, compromising water quality in the membrane permeate and raising concerns about transformation-product release. ,

Distillation-based membranes provide a compelling alternative platform for catalytic membranes. Distillation membranes rely on vapor-phase transport across hydrophobic porous media, enabling inherently high rejection of nonvolatile species, including low-molecular-weight byproducts of catalytic oxidation. − Distillation membranes are also commonly fabricated from chemically robust polymers such as polytetrafluoroethylene (PTFE), which exhibit exceptional resistance to oxidative and radical-driven degradation compared to polyamide-based membranes. Because transport occurs through gas-filled pores rather than dense polymers, distillation membranes can tolerate partial surface modification without obstructing permeation pathways. ,, Together, these properties suggest that distillation membranes offer a unique opportunity to create catalytic membranes that are highly selective and robust.

Despite the promise of integrating catalytic functionality with distillation membranes, this technology has remained largely unexplored. A key barrier is the intrinsic chemical inertness of hydrophobic polymers used in distillation, which complicates catalyst attachment. − Moreover, many catalytic materials, such as metal oxides, are hydrophilic and can compromise membrane hydrophobicity, leading to pore wetting and membrane failure. − Existing coating approaches often coat the entire membrane, rendering the pores hydrophilic and accelerating membrane failure via pore wetting. Thus, strategies that limit the infiltration of chemical precursors into the membrane pores are needed to selectively modify only the membrane surface while preserving the hydrophobic porous matrix. − Overcoming these fabrication challenges is essential for realizing the catalytic distillation membranes.

In this work, we demonstrate catalytically active distillation membranes based on plasma-assisted atomic layer deposition (ALD) of TiO₂ onto commercial PTFE membranes. Using short precursor exposure times and surface-activated growth, we achieve TiO₂ coatings with controlled penetration depth that preserve membrane porosity and vapor transport pathways. With a combination of electron microscopy, spectroscopy, and wettability characterization, we quantify coating thickness, chemical composition, and interfacial properties. We further demonstrate that TiO₂-coated membranes maintain wetting resistance and desalination performance in pressure-driven distillation (PD) and exhibit catalytic activity, establishing a new method for integrating advanced oxidation with membrane separation. This approach provides a generalizable route toward chemically robust, multifunctional membranes capable of addressing emerging challenges in advanced water treatment.

Materials and Methods

Membranes and Chemicals

Commercial polytetrafluoroethylene (PTFE) membranes (model PTF002LH0P, nominal pore size 20 nm) were obtained from Cytiva. The membranes were hydrophobic and supported on a nonwoven backing layer, as provided by the manufacturer. For ALD reactions, trimethylaluminum (Sigma-Aldrich, USA), titanium tetrachloride (Sigma-Aldrich, USA), and deionized (DI) water were used as reaction precursors. Sodium chloride (Fisher Chemical, USA) dissolved in DI water was used as the feed solution in liquid entry pressure tests. Ultrapure water (resistivity ≥ 18.2 MΩ·cm) produced using a Milli-Q purification system was used for all experimental preparations and cleaning steps. Methylene blue (Sigma-Aldrich, USA) was dissolved in DI water and diluted to 5 mg/L for catalytic degradation tests.

Atomic Layer Deposition

Plasma-assisted ALD was performed at 165 °C in a Picosun Sunale R150 ALD. Two reactions were programmed for deposition onto the PTFE surface. The first was a remote oxygen plasma step followed by the introduction of trimethylaluminum (TMA). For the plasma-TMA cycle, five cycles of remote oxygen plasma with a power of 2000 W, a pulse time of 30 s, and a purge time of 10 s were performed, and then five cycles of TMA with a pulse time of 0.1 s and a purge time of 6 s were performed. The second reaction was the TiO₂ formation with titanium tetrachloride and water. TiCl₄ was dosed into the ALD chamber with a pulse time of 0.1 s and a purge time of 6.0 s. Then, water was dosed with a pulse of 0.2 s and a purge of 8.0 s.

Membrane Characterization

The membrane morphology and surface topography were viewed with scanning electron microscopy (SEM) on a Hitachi SU8010 microscope. Prior to SEM imaging, platinum was sputtered onto the PTFE membranes until a thickness of 3 nm was achieved.

Cross-sectioning and imaging of the PTFE membranes was performed on a Thermo Fisher Scientific Scios 2 focused ion beam/scanning electron microscope (FIB/SEM) with a Leica liquid nitrogen-cooled SEM stage. A 7 nm gold thin film was sputtered over the entire sample on an SEC MCM-100P ion sputter coater under vacuum to reduce surface charging under the electron and Ga-ion beam. The samples were cooled on the cryo-stage under the high-vacuum environment of the SEM to −150 °C for FIB cross-sectioning. Organometallic Pt precursor was deposited in the FIB and cured using the Ga-ion beam (30 kV, 1.0 nA) to yield a 1.5 μm thick protective Pt layer on the surface of the PTFE membranes. Ga-ion beam voltage was kept at 30 kV, but currents were kept below 5 nA for cuts and were polished with a current of 1.0 nA to prevent damage to the polymeric matrix. All energy-dispersive X-ray spectroscopy (XEDS) spectra were drift corrected and taken with an E-beam voltage and current of 10 kV and 0.2 nA, respectively, with a pixel dwell time of 1 ms. XEDS maps were taken using an EDAX Octane Elite detector.

XPS survey spectra (0–1200 eV) were acquired on a Kratos Supra spectrometer using an Al Kα source at 1486.7 eV and 15 mA emission current; charge neutralization parameters were determined via continuous C 1s scanning. High-resolution elemental core scans were collected on a Kratos Axis Ultra DLD spectrometer with a monochromatic Al Kα source at 120 W (1486.6 eV), a base pressure of 3 × 10^–9 Torr, and a pass energy of 20 eV; charge compensation was achieved using low-energy electrons, and all spectra were referenced by setting the C 1s peak to 284.6 eV. Polymer samples were isolated on the sample plate to ensure consistent charging, and data were processed using CasaXPS. Survey spectra were used to determine elemental atomic percentages. Measurements were taken at four locations per sample across three replicates, and error bars reflect the standard deviation of these measurements.

Contact angles were measured with a Biolin Scientific tensiometer using the sessile drop technique. For all measurements, a droplet of DI water (less than 10 μL) was deposited onto the active side of the membrane surface using a pipette.

Grazing-incidence wide-angle X-ray diffraction profiles were acquired using Cu Kα radiation (Xenocs Xeuss 3.0) on the 1000TiO₂ sample. Samples were mounted onto a silica substrate. The X-ray angle of incidence was 0.2° to maximize sensitivity to the membrane surface structure.

The presence and morphology of TiO₂ on the membrane surface were investigated using X-ray computed tomography (XCT) imaging in the absorption contrast imaging modality of a Zeiss Xradia Ultra 810 lab instrument with large field of view (LFOV) optics. The PTFE membrane was prepared with a cold fracture method. The membrane was submerged into liquid nitrogen and subsequently sliced with a razor blade into a wedge with a wide base and very thin point (<100 μm). 2-part epoxy was used to affix this wedge to a sample pin for imaging. The thin point extended slightly above the pin to isolate it for imaging while maintaining stability during rotation.

In pure absorption mode, the sample was rotated through 1301 projections spanning 180° with 15 s of X-ray exposure at each projection. Detector pixels in LFOV are 64 nm, and camera binning of 2 was used for all measurements to yield reconstructed voxel sizes of 128 × 128 × 128 nm and a subsequent resolution of ∼500 nm. Imaging was performed with monochromatic X-rays at 5.4 keV.

Pressure-Driven Distillation Testing

Distillation experiments were conducted in a bench-scale dead-end filtration setup with a constant applied hydraulic pressure of 3.4 bar (50 psi), which is below the reported liquid entry pressure of 13.8 bar for the 20 nm PTFE membranes used in this study. , All PD experiments were conducted at 22 ± 1 °C. The feed solution was prepared by dissolving NaCl at a concentration of 10 mM in ultrapure water, and the permeate side of the membrane was initially in contact with ultrapure water. Four membrane samples were tested: pristine PTFE, and PTFE coated with 250, 500, and 1000 cycles of TiO₂. Circular membrane coupons with an active diameter of 1 cm, corresponding to an active area of 0.79 cm², were used for all tests. Feed pressure was supplied using compressed nitrogen gas. The filtration setup consisted of a custom-fabricated polyether ether ketone (PEEK) pressure cell fitted with a membrane support and an O-ring seal to prevent bypass. Permeance was monitored by measuring the change in water volume on the permeate side after 20 h. Membrane salt rejection was assessed by measuring the conductivity in the feed and permeate to verify membrane integrity and identify potential wetting.

Model Dye Degradation to Probe Catalytic Activity

The photocatalytic performance of the TiO₂–PTFE membranes with increasing cycles of ALD and a pristine PTFE as a control was evaluated by measuring the decrease in the concentration of methylene blue (MB) using a UV spectrophotometer (Cary UV-4000) following exposure to UV light irradiation (365 nm, 8.0 W). Membranes were cut into 16 mm diameter circles and mounted in Petri dishes with Kapton tape to ensure a flat surface. The membranes were immersed in a solution containing 5 mg/L MB. The lamp was positioned 13 cm above the membrane. Prior to starting the UV lamp, the membranes sat in the dye solution in the dark for 1 h to come to light absorption equilibrium. Then, for an additional hour, the membranes remained submerged in the MB solution under dark conditions to establish their behavior. Absorbance measurements were taken at each of these points with a UV–vis spectrometer. Once UV light exposure commenced, absorbance measurements were taken every 30 min for 2 h. After the initial UV exposures, the samples were left in the dark for 16 h, and their absorbance measurements were taken. Similarly, UV exposure was commenced after the 16 h dark period, and absorbance measurements were taken every 45 min for an additional hour and a half.

Wettability Tests with Liquid Entry Pressure

Liquid entry pressure measurements were taken to establish the wetting resistance of the membranes following ALD fabrication. The membranes were cut into 13 mm diameter circles and loaded into a stainless steel dead-end membrane cell. The feed side was loaded with a 5 mM NaCl solution and was gently tapped to ensure no bubbles. The cell was attached to a nitrogen gas cylinder, and the pressure was slowly increased to 3.5 bar and maintained there for 30 min. Following this initial pressurization, the pressure was increased every 10–15 min by 0.7 bar. Once water was observed on the permeate end of the cell, the pressure was recorded, and a needle valve was used to release gas from the system.

Results and Discussion

Atomic Layer Deposition of TiO₂ on PTFE Membranes

TiO₂ was grown on PTFE membranes by using a remote plasma ALD process (Figure ). Because PTFE consists of a fully fluorinated carbon backbone and is therefore highly chemically inert, remote plasma was employed to introduce reactive oxygen-containing functional groups, such as hydroxyls or carbonyls, onto the surface prior to catalyst deposition. − , Following a brief remote plasma activation step, the reaction chamber was sequentially dosed with titanium tetrachloride (Figure , step 1) and water (Figure , step 2), with a purge step in between the introduction of each chemical. One complete sequence of precursor dosing and purging constituted a single ALD cycle, and this study performed 250, 500, and 1000 cycles of TiO₂ ALD on the PTFE membranes.

1.

Atomic layer deposition of TiO₂ on a porous surface: (1) Exposure to titanium tetrachloride, (2) exposure to water vapor, and (3) formation of titanium dioxide on the surface. Repeated cycles are used to deposit layers of a desired thickness.

Following ALD deposition, the top surfaces of the membranes were viewed under SEM to evaluate the morphological changes (Figure ). Unmodified PTFE membranes exhibit a characteristic fibrous and nodular structure. ,, From 250, 500, and 1000 ALD cycles, the apparent PTFE fiber diameters increased by 60, 172, and 189%, respectively, relative to the unmodified fibers (Table S1). The membranes exposed to 1000 ALD cycles also exhibited regions where large clusters of TiO₂ blocked voids between the PTFE fibers (Figure S1). At higher magnifications, the unmodified membranes displayed fibers with relatively smooth surfaces. In contrast, ALD treatment led to observable particulate growth on PTFE fibers and nodules. Particulates grown at 250 cycles were relatively small and discretely distributed; at 500 cycles, the particulates began to coalesce; and after 1000 cycles, a smooth fiber morphology was observed. Xu et al. observed similar particulate growth on PTFE fibrils following ALD growth of aluminum oxide on PTFE and attributed the observed surface smoothing of the fibers to filling of the spaces between adjacent particles at higher ALD cycles, consistent with our observations.

2.

Scanning electron micrographs of pristine PTFE membrane (A) and samples coated with TiO₂ via ALD with 250 cycles (B), 500 cycles (C), and 1000 cycles (D).

Analysis of the top surface of the membrane via XPS confirmed the chemical composition of the ALD coating. The survey scan (Figure A) shows that with increasing ALD cycles, the Ti 2p peak at 458.5 eV systematically increases, while the C 1s peak at 291 eV and the F 1s peak at 688 eV both decrease in intensity. In the carbon core scans (Figure B), we see the C–F bond peaks at 290–292 eV disappear as we increase from 250 to 1000 cycles but begin to observe more pronounced peaks for C = O functional groups at 288–290 eV, C–O at 286 eV, and C–C at 284 eV. Similarly to the survey scans, we see the Ti 2p_1/2 and Ti 2p_3/2 peaks become more pronounced with increasing cycles (Figure C). Since increasing ALD cycles introduce a growing amount of TiO₂ on the membrane surface, the eventual disappearance of the C–F bond on the surface is consistent with the depth limit of XPS measurements, where only the top 10 nm of the surface can be analyzed. We assume that the increasing signals for C = O, C–O, and C–C in the core scans are likely from adventitious carbon due to the surface’s exposure to air. The oxygen core scans (Figure D) show two distinct peaks, the first at 529 eV corresponds to lattice oxygen present in TiO₂, and the second at 531 eV, which is attributed to surface hydroxyl groups or oxygen bound to adventitious carbon. Table S2 summarizes survey scan atomic percentages and peak-fitting results from the core scans. Interestingly, even after 1000 cycles, we observed a small fluorine signal in some areas of the analyzed membrane in the XPS survey scan. This may be due to potentially untreated PTFE fibers that lie beneath the coated surface being analyzed during XPS or potential inhomogeneous coverage of the membrane during ALD, which is observed in SEM (Figure S1).

3.

(A) XPS survey spectra of TiO₂-coated membranes. Core scans and peak fittings of (B) Carbon 1s, (C) Titanium 2p, and (D) Oxygen 1s.

Penetration Depth of TiO₂ Coating into PTFE Membranes

Although the successful deposition of TiO₂ onto the PTFE membrane was confirmed by imaging and spectroscopy, it is important to determine the penetration depth of the precursors into the porous membrane structure. ALD is conventionally thought of as a conformal surface modification technique and has been shown to uniformly coat the entirety of the pores in membrane materials. However, the selectivity and overall performance of distillation membranes are reliant upon their hydrophobic nature, which allows air to remain trapped in the pore even when the membrane is immersed in liquid water. , If the pore is entirely hydrophilic, water will infiltrate the membrane pores, causing nonselective liquid transport. Thus, effective catalytic distillation membranes must limit the penetration of the hydrophilic precursors into the hydrophobic porous membranes. Pathak et al. showed that they were able to control the penetration depth of ALD precursors into porous substrates by limiting dosing times of the precursors. Following this procedure, precursor dosing times were limited to 0.1 s to control penetration into the PTFE membrane.

To determine the penetration depth of the ALD precursors, cross sections of membranes coated with 1000 cycles of ALD were analyzed with SEM and X-ray energy-dispersive spectroscopy (XEDS) to evaluate the penetration of TiO₂ coatings. The cross section of the membrane was prepared with a focused ion beam (Figure A), and elemental analysis by XEDS showed clear bands of oxygen, titanium, and fluorine (Figure B–E). The XEDS mapping indicates that TiO₂ remained localized near the upper surface of the pores and did not penetrate further into the membrane. To further confirm the localization of the TiO₂, membranes with 1000 cycles of ALD were imaged with X-ray computed tomography (XCT), which provides submicrometer resolution three-dimensional images of materials’ internal pore network and morphology. , Figure F shows the reconstructed volume of a sample after 1000 cycles, where the grayscale value is directly proportional to the X-ray absorption. A thin layer of highly absorbing material, likely corresponding to TiO₂, is observed on the surface of the membrane. Throughout the rest of the PTFE, the average grayscale remains nearly constant, indicating the lack of TiO₂ penetration. Although deformation during sample preparation prevents precise thickness determination, the surface layer thickness is estimated to be approximately 1 μm (Figure S2). Together, the XCT and XEDS results demonstrate that the TiO₂ forms a thin surface layer and does not penetrate deeply into the PTFE porous structure.

4.

(A) Scanning electron micrograph of PTFE membrane cross sections after 1000 ALD cycles (1000TiO₂). (B) Corresponding energy-dispersive X-ray spectroscopy elemental map and individual maps for oxygen (C), titanium (D), and fluorine (E). (F) X-ray computed tomography of 1000TiO₂ membrane, where grayscale intensity is proportional to X-ray attenuation. The bright top surface is associated with a high X-ray absorbing material (TiO₂), while the black volume is associated with PTFE.

Water contact angle and liquid entry pressure measurements were used to assess changes in membrane wettability following TiO₂ deposition (Figure A). The water contact angle measured at the top surface of the membrane decreased from 147° for unmodified PTFE to 110, 98, and 80° with 250, 500, and 1000 cycles, respectively. Despite this reduction in surface hydrophobicity, the membranes did not readily wick water following modification, suggesting that the underlying air layer within the porous structure remained intact.

5.

(A) Liquid entry pressure (LEP) and contact angle of the pristine PTFE membrane and membranes coated with increasing cycles of TiO₂ ALD. (B) Pressure-driven distillation water permeance and salt rejection results of the pristine PTFE membrane and membranes coated with increasing cycles of TiO₂ ALD. (C) Summary of modulus data of membrane samples taken with nanoindentation. Error bars represent the mean ± 1 s.d. for measurements from at least four areas on the membrane samples.

Liquid entry pressure (LEP) measurements further demonstrated that the TiO₂ coating did not cause wetting or a loss of distillation functionality. The unmodified PTFE membrane exhibited an LEP of 13.8 bar, while membranes treated with 250, 500, and 1000 ALD cycles showed LEP values of 13.3, 12.8, and 9.4 bar, respectively (Table S1). These results indicate that the trapped air layer inside the membrane is preserved after coating, although increasing coating cycles decreases the LEP. This reduction in the LEP may be caused by changes in hydrophobicity throughout the membrane. Although XCT and SEM-XEDS indicate that TiO₂ is primarily confined to the membrane surface, subtle changes in wettability within the near-surface pore region may still contribute to changes in the LEP.

We verified that catalytic membranes were capable of desalination using pressure-driven distillation (PD) experiments, where an applied pressure of 3.4 bar was used to drive vapor transport through the membrane. ,, The pristine membrane and the membranes modified with 250 and 500 ALD cycles all maintained higher than 95% rejection of 10 mM NaCl, indicating that the ability to separate via distillation was maintained after coating. These membranes also demonstrated permeances around 5.5 g m^–2 h^–1 bar^–1, consistent with prior PD work using these membranes. In contrast, the membrane treated with 1000 ALD cycles showed a near-complete loss of salt rejection and a 10-fold increase in water permeance, indicating that this membrane experienced liquid breakthrough.

The apparent failure of the membranes treated with 1000 ALD cycles was consistent with SEM measurements taken following LEP measurements, which showed that these membranes exhibited surface cracking (Figure S3), indicating that higher ALD cycles may impact membrane mechanical properties. To probe this behavior further, nanoindentation measurements were performed to determine the elastic moduli of the membranes (Figure C). With increasing ALD cycles, the modulus increased markedly from 4.7 MPa for bare PTFE to 310 MPa for the 1000 cycles of TiO₂ on the PTFE. This indicates that TiO₂ created a rigid top layer on the membrane, explaining the cracked top surface following LEP measurements. Overall, LEP and distillation performance measurements suggest that the number of cycles should be optimized to enable catalytic behavior while preventing substantial changes to the membrane wetting or mechanical properties.

Photocatalytic Performance of Fabricated Membranes

TiO₂ is an effective photocatalyst due to its flat band potential, stability, cost-effectiveness, and low toxicity. , When exposed to light, TiO₂ exhibits photocatalytic activity and produces reactive oxygen species (e.g., hydroxyl or oxygen radicals), which aid in the decomposition of contaminants (Figure A). ,, The photocatalytic activity of the membranes under UV irradiation was evaluated via degradation of methylene blue (MB), a model contaminant commonly used to benchmark catalytic performance in the literature (Figure B). − Prior to UV irradiation, the membranes were held under dark conditions for 2 h, with absorbance measurements taken every 2 h. The coated membranes exhibited some slight physical adsorption of the methylene blue, but it was negligible in comparison to the amount removed during UV exposure (Table S3). Additionally, after running 2 h of UV irradiation, the membranes were held in dark conditions for 16 h, during which the absorbance remained unchanged.

6.

(A) Schematic diagram illustrating photocatalytic degradation of methylene blue in solution with TiO₂-coated distillation membranes. (B) UV Vis spectroscopy of methylene blue solutions after UV irradiation of TiO₂ membranes. The initial UV irradiation experiment was 2 h, with absorbance being taken every 30 min. Following the first 2 h, the membranes were kept in the dark for 16 h and then were exposed to UV light for an additional 1.5 h. (C) Grazing-incidence wide-angle X-ray diffraction of pristine PTFE and 1000TiO₂ PTFE.

For the UV exposure experiments, both the MB control solution and pristine PTFE exhibited negligible degradation (less than 1%), confirming the absence of significant photolysis or adsorption. Increasing the number of ALD cycles on the membranes led to a monotonic increase in the amount of MB degraded under UV irradiation, with 1.2, 5.2, and 9.6% removal of MB observed for membranes coated with 250, 500, and 1000 cycles of TiO₂, respectively. The degradation rates observed are lower than those of systems highly loaded with photocatalysts but similar to reported values for ultrathin, surface-bound catalytic coatings with limited active material. − Zhang et al. reported they saw a 9.6% decrease in methyl orange concentration for ceramic membranes coated with 400 TiO₂ ALD cycles at a higher temperature of 375 °C, which likely formed more catalytically active anatase species but would have degraded the polyester backing material on our membranes. , To facilitate comparison with other studies, the area-normalized degradation rate and TiO₂ mass loading per area are provided in Table S3.

Photocatalytic reactivity of TiO₂ is dependent on the crystalline phase, and photocatalytic activity is usually seen in anatase and rutile phases, with anatase exhibiting higher photocatalytic performance. , To ensure the formation of anatase TiO₂, we performed the ALD reactions at 165 °C, which is near the temperature where amorphous TiO₂ begins to transition to anatase. , To probe the crystallinity of the coatings, grazing-incidence X-ray diffraction (GIXRD) measurements were made on the 1000TiO₂ membrane. The strong reflection in both profiles at 1.3 Å^–1 arises from crystalline domains of the PTFE membrane (Figure C). The profile measured from the 1000TiO₂ has distinct diffraction peaks arising from the coating, and a broad amorphous peak centered at 0.95 Å^–1 from the amorphous TiO₂. The relatively weak reflection at 1.6 Å^–1 corresponds to a (002) reflection from anatase TiO₂. The combined amorphous and anatase scattering signatures indicate that the ALD process produces semicrystalline TiO₂ films at the surface of the PTFE membrane. Polarized optical microscopy images of the coated membranes further confirm the polycrystalline nature of the TiO₂ films on the membrane surface (Figure S4). This crystallinity was not observed on untreated PTFE. Overall, these results demonstrate that ALD-deposited TiO₂ imparts photocatalytic activity to otherwise inert PTFE membranes. Even at low deposition temperatures, sufficient anatase formation is achieved for contaminant degradation. However, future optimization may allow for more anatase TiO₂ formation and higher catalytic degradation rates.

Conclusions

This work demonstrates a new class of catalytically active distillation membranes enabled by controlled atomic layer deposition on PTFE substrates. By limiting precursor exposure time, TiO₂ growth was confined to an approximately 1 μm distance from the surface of the membranes, creating a catalytically active interface while preserving the hydrophobic air-trapping pores required for distillation. The fabricated membranes demonstrated pressure tolerances higher than 9 bar without wetting, achieved higher than 95% salt rejection in PD, exhibited chemical signatures consistent with TiO₂ coatings, and showed catalytic degradation of model dye compounds. These results highlight the ability to integrate catalytic functionality into chemically inert, hydrophobic membranes without compromising their separation performance.

The fabrication strategy presented here provides a platform for the continued development of catalytic distillation membranes that integrate advanced oxidation, high selectivity, and robustness. We show that optimization of catalyst coating parameters (such as the ALD cycle number and precursor pulse times) will be critical for balancing catalytic activity with membrane wettability and mechanical integrity. While this study focused on verifying photocatalytic activity using methylene blue as a standardized benchmark, future work should validate the catalytic removal of representative volatile contaminants and investigate the potential of catalytic coatings for fouling mitigation at the membrane surface. More broadly, future studies should examine performance under realistic desalination and complex wastewater conditions. Continued development of scalable and application-specific coating approaches will be essential for translating catalytic distillation membranes to practical water treatment applications.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.6c07890.

• Membrane fiber diameter, liquid entry pressure, and modulus data (Table S1); XPS peak-fitting results for carbon, titanium, and oxygen core-level spectra and survey scans (Table S2); SEM images showing heterogeneity of the TiO₂ coating on membrane surfaces (Figure S1); nanoCT cross-sectional image and histogram of oxide layer thickness distribution (Figure S2); SEM images of membranes after liquid entry pressure testing (Figure S3); polarized and optical microscopy images demonstrating TiO₂ crystallinity (Figure S4); and calculated TiO₂ mass loading per membrane area and methylene blue degradation rates under dark and UV conditions (Table S3) (PDF)

The authors declare no competing financial interest.