Engineering oxygen vacancies in TiO₂ for isotope-selective catalysis: a strategy for hydrogen isotope separation

Abstract

Deuterium and tritium are key strategic materials for the development of nuclear energy, and it is particularly important to properly treat wastewater containing deuterium (D)/tritium (T). Due to the atomic mass of deuterium and tritium being greater than that of hydrogen (H), their chemical bond vibration frequencies are lower, and their photolysis rates are usually slower. This characteristic has made the photocatalytic H₂–H₂O method a new hotspot in the field of hydrogen isotope separation. Here, we introduce a blue titanium dioxide catalyst (TiO_2−x) with an exposed highly active (001) surface. By optimizing the band structure through oxygen vacancy engineering, the separation and transfer efficiency of photogenerated carriers has been significantly improved. Theoretical calculations indicate that TiO_2−x exhibits significant selectivity towards H₂O on the (001) crystal plane. A series of experiments have verified the high efficiency of TiO_2−x in separating hydrogen isotopes under illumination conditions. When the ratio of H₂O to D₂O is 10 : 90, the yield of H₂ reaches 10.96 mmol g⁻¹ h⁻¹, and the H/D separation coefficient is 5.31. This study achieves selective catalysis and separation of hydrogen isotopes by selecting appropriate catalytic materials, providing a powerful supplement to traditional hydrogen isotope separation methods.

---

1. Introduction

The considerable rise in global energy demand and a multitude of environmental pollution concerns have led to an increased necessity for clean production and resource utilization. Deuterium (D) and tritium (T) are of great significance in the field of energy, particularly in the context of nuclear energy production.^1,2 However, deuterium or tritium contaminated water poses a threat to human health and the environment, and special safety measures are required to protect humans and the environment due to the radioactive and half-life characteristics of tritium. Furthermore, D has become an indispensable component in a multitude of industrial and scientific research fields, including neutron scattering technology, drug discovery and development, nuclear magnetic resonance spectroscopy, and isotope tracing.^3–5 Despite the abundance of hydrogen in the Earth's crust, the proportion of deuterium in natural reserves is only 0.015%. Furthermore, the similar molecular size and physicochemical properties of D and H present significant challenges to the separation of D and H.^6–9 The growing demand for D makes the effective separation of D and H isotopes a matter of great significance. It should be noted that D and T are not directly obtained as separate isotopes, so methods need to be used to separate them from more common and lighter isotopes. At present, research based on H/D separation includes water distillation,¹⁰ chemical exchange,^11,12 component adsorption separation,^13–15 water electrolysis,^13,16–22 and combined electrolysis catalytic exchange (CECE).^23–25 Electrolytic separation with a high separation coefficient is considered a promising method for treating tritium/deuterium-containing wastewater. However, at present, catalysts mainly used for electrolytic separation are precious metals such as Pt. In contrast, pursuing efficient separation requires a higher current density, which leads to a rapid decline in the service life of most catalysts. Therefore, highly selective hydrogen isotope separation under mild conditions remains a huge challenge.

In recent years, research has shown that photocatalytic water splitting can convert water into hydrogen (H₂, D₂, and T₂) through photocatalysts under sunlight, and it has been demonstrated that using semiconductor photocatalysts can effectively and sustainably produce green hydrogen.^26–28 The photochemical process is expected to achieve effective deuterium–hydrogen separation. Therefore, the photocatalytic hydrogen strategy for isotope separation deserves further investigation. Titanium dioxide is widely used as an excellent photocatalyst due to its outstanding performance in environmental and energy issues, such as the degradation of organic matter, reduction of heavy metals,²⁹ gaseous hydrogen isotope separation,³⁰ and photocatalytic hydrogen production.³¹ However, the wide bandgap of titanium dioxide (3.0–3.2 eV) limits its photocatalytic applications in the ultraviolet region. Many reports have been published on the modification of semiconductor titanium dioxide to enhance its photocatalytic activity. Surface scientists have proven that the average surface energy is 0.90 J m⁻² for (001) > 0.53 J m⁻² for (100) > 0.44 J m⁻² for (101).^32,33 Although it can be expected that high surface energy (001) surfaces have higher chemical activity. Xie et al.³⁴ reported a simple hydrothermal approach to synthesize sheet-like rutile titanium dioxide, utilizing the specific stabilizing effect of fluoride ions on the (001) surface, which has exposed highly active (001) surfaces. Sinhamahaputra et al.³⁵ reported a new controllable magnesium thermal reduction method that can synthesize reduced black titanium dioxide with excellent catalytic hydrogen production ability in a 5% H₂/Ar atmosphere. It is necessary to identify the various performance factors, including surface lattice disorder, oxygen vacancies, Ti³⁺ ions, Ti–OH and Ti–H groups, and band edge displacement. Despite the notable enhancement in the optical absorption of hydrogenated titanium dioxide materials, many of them fail to demonstrate the anticipated efficiency in visible light-assisted water splitting. This can be attributed to the presence of several control factors, which can also have a detrimental impact on photocatalytic performance. For instance, surface defects or oxygen vacancies are regarded as significant property factors, acting as electron donors, enhancing donor density, and facilitating charge transport in black titanium dioxide.³⁵

Herein, we report the thermal reduction of titanium dioxide nanoparticles in the presence of H₂/Ar to obtain blue titanium dioxide nanoparticles, which enhance the photocatalytic hydrogen production ability in methanol–water system and achieve effective hydrogen isotope separation. The prepared TiO_2−x catalyst exhibits a maximum hydrogen production rate of 10.96 mmol h⁻¹ g⁻¹ and a separation efficiency of 5.31 in the presence of 2 wt% Pt as a co-catalyst across the entire solar and visible light wavelength ranges. The excellent catalytic and separation performance indicates the achievement of a balanced combination of different factors such as Ti³⁺, surface defects, oxygen vacancies, and recombination centers, as well as optimized band gaps and positions during the preparation process. Our research provides a reference basis for the design and development of better hydrogen isotope separation photocatalysts.

2. Experimental section

2.1. Chemicals and reagents

All chemicals and reagents are described in Text S1.

2.2. Synthesis of TiO_2−x

Tetrabutyl titanate (5 mL, Ti(OBu)₄) was mixed with hydrofluoric acid (HF, 0.8 mL) and stirred at high speed for 3 hours to ensure sufficient contact. The mixture was transferred to a high-pressure reaction vessel lined with polytetrafluoroethylene and maintained at 200 °C for 24 hours for hydrothermal reaction. After the reaction was completed and cooled to room temperature, the white precipitate was obtained by washing with ethanol and distilled water several times and separating it by high-speed centrifugation. The powder obtained by vacuum drying the precipitate at 60 °C for 12 h is TiO₂ nanoparticles.

An appropriate amount of TiO₂ powder was treated in a tube furnace containing H₂/Ar (5% H₂) at 450, 500, 550, and 600 °C for 3 h (heating time of 100 minutes). Then a series of oxygen vacancies in TiO₂ were obtained, denoted as TiO_2−x-t (t = 450, 500, 550, 600 °C, representing processing temperatures).

2.3. Characterization

All the instruments required for characterization testing, electrochemical testing details, zeta potential, DFT calculation details, and in situ infrared testing details are provided in Text S2–S6 of the SI.

2.4. Photocatalytic activity test

The light source required for photocatalytic performance testing was provided by a 300 W Xenon lamp. Using a water-cooled photocatalytic hydrogen production testing system (as shown in Fig. S1), the composition and content of the gas produced were analyzed by gas chromatography at the same time interval. During the photocatalytic reaction, the ambient temperature was maintained at 4 °C. 8 mg of TiO₂ catalyst sample was added to a quartz photocatalytic reactor, and 30 mL of aqueous solution containing 10% methanol (sacrificial agent) was added to evaluate the photocatalytic hydrogen production activity of each sample. First, the photocatalytic hydrogen production rate was evaluated under different amounts of Pt (1, 2, 4, and 6 wt%) co-catalysts. Finally, photocatalytic hydrogen isotope separation experiments were conducted in a series of H₂O : D₂O (100 : 0, 90 : 10, 75 : 25, 50 : 50, 25 : 75, 10 : 90 and 0 : 100) ratio solutions to evaluate the separation efficiency of the catalyst.

3. Results and discussion

3.1. Preparation and characterization

Surface scientists have demonstrated that fluoride ions may play a crucial role in the formation of exposed (001) surfaces. In a recent study, Han et al.³⁴ described a straightforward hydrothermal method for synthesizing sheet-like anatase TiO₂ with the highly reactive (001) facets exposed. This approach leverages the unique stabilizing effect of fluorine ions on the (001) facets. As shown in Fig. 1a, sheet-like anatase TiO₂ was synthesized at 200 °C for 24 h via a straightforward hydrothermal route utilizing tert-butyl titanate Ti(OBu)₄ as the Ti source and 47% hydrofluoric acid solution as the solvent. The high-resolution transmission electron microscopy (HRTEM) images of TiO₂ reveal that the crystal assumes the shape of a nanosheet, exhibiting prominent (001) planes of considerable dimensions, measuring approximately 100 nm in width and 8 nm in thickness (Fig. 1b–d). Transmission electron microscope analysis showed that the prepared catalytic material presented a typical sheet structure, and its more refined morphological characteristics are shown in Fig. S4 of the SI. The lattice spacing of the larger surface is 0.352 nm, and the lattice spacing of the relatively smaller surface on the side is 0.237 nm. This conclusion confirms that the synthesized TiO₂ has the (101) and (001) planes, and the titanium dioxide nanosheets surrounded by the (001) planes are expected to have good photocatalytic hydrogen isotope separation efficiency. Then, the synthesized TiO₂ was subjected to continuous hydrogenation for 3 hours under H₂/Ar (H₂ 5%) flow at 450–600 °C for oxygen vacancy engineering, resulting in hydrogenated titanium dioxide nanocrystals (TiO_2−x) of different colors (Fig. S2). A series of V_O-rich TiO_2−x was synthesized following a V_O engineering strategy by heating at 450 °C, 500 °C, 550 °C and 600 °C.³⁶ After O vacancy engineering, we note that the thin, planar, square shape of TiO₂ has changed into one with enhanced thickness; the titanium dioxide nanosheets became larger (about 110 nm) with a thickness of about 10 nm (Fig. 1f–h and S3).^34,37 In addition, TiO_2−x exhibits a rougher surface morphology than TiO₂. Energy dispersive X-ray spectroscopy (EDS) elemental mappings display a uniform distribution of Ti and O elements throughout TiO_2−x (Fig. 1e and i). The O : Ti ratio in TiO_2−x catalyst is close to 0.7 : 1, while TiO₂ is about 2 : 1 (Table S1).

Fig. 1 (a) Schematic diagram of the synthesis pathway of TiO₂ and TiO_2−x. (b) HRTEM image of TiO₂, with (c) representing the stripe spacing of the (101) crystal plane, (d) representing the HRTEM image of the (100) crystal plane (corresponding to the lattice stripe spacing), and (e) representing the EDS image of TiO₂. (f) HRTEM image of TiO_2−x, with (g) representing the stripe spacing of the (101) crystal plane, (h) representing the HRTEM image of the (100) crystal plane (corresponding to the lattice stripe spacing), and (i) representing the EDS image of TiO_2−x.

The X-ray diffraction (XRD) spectra of a series of samples are presented in Fig. 2a. The XRD peaks of all samples are consistent with those of titanium dioxide in the anatase phase (JCPDS no. 21-1272), which indicates the successful synthesis of TiO₂.^38,39 Further details on the volume phase changes examined using XRD are provided below. The anatase peak of TiO₂ was observed to persist following hydrogenation at temperatures between 450 and 550 °C. In comparison to TiO₂, no discernible alterations were discerned in the XRD peak position of the anatase peak (101) of TiO_2−x, thereby indicating that there was no substantial hydrogen-induced distortion in the anatase bulk phase under the specified hydrogenation conditions. Nevertheless, the application of hydrogenation at 600 °C has been observed to induce the emergence of additional XRD peaks, which have been assigned to the rutile phase in addition to the previously observed anatase phase (JCPDS no. 21-1276). It has been previously reported that anatase TiO₂ undergoes a phase transition from anatase to rutile upon annealing at temperatures exceeding 550 °C. In the present case, a phase transition occurred at a higher temperature (600 °C), resulting in the appearance of a signal belonging to the rutile phase in the XRD pattern, which was not observed at 550 °C. Unless otherwise stated in the subsequent description, TiO_2−x-550 will be referred to as TiO_2−x. The presence of Ti and O elements in the catalyst was confirmed by X-ray photoelectron spectroscopy (XPS) (Fig. 2b).^35,37 Concomitantly, the data demonstrate that as the temperature rises, the signal attributed to F 1s in the full spectrum gradually diminishes. As shown in Fig. 2c, the O 1s XPS spectra of TiO₂ and TiO_2−x can be attributed to Ti–O–Ti (529.7 eV). The shoulder peak at 531.1 eV is attributed to the O atom in proximity to the O vacancy. It is postulated that OVs may induce the dissociation of H₂O and the chemical adsorption of OH at the OV sites. In comparison to the original TiO₂, TiO_2−x exhibited a greater peak area at 531.1 eV, suggesting the presence of a higher density of OVs on its surface relative to the other samples. Fig. 2d illustrates that the two peaks at binding energies of 458.5 and 464.3 eV are attributed to Ti 2p_3/2 and Ti 2p_1/2 of Ti⁴⁺, respectively. These signals were scarcely observed on the surface of TiO₂. In contrast, OVs are present on the surface to maintain charge equilibrium. Fig. 2e illustrates the changes from TiO₂ to TiO_2−x, as evidenced by the EPR spectrum. The test results of oxygen vacancy quantitative data are shown in Table S2, and the comparison of the integrated area of the EPR signal, as shown in Fig. S5, is used as auxiliary and semi-quantitative evidence for the change in oxygen vacancy concentration. From the relative strength trend (SI, Fig. S5), it is highly consistent with the change trend of oxygen vacancies obtained by XPS O 1s analysis, which mutually confirms the change in oxygen vacancy concentration at different treatment temperatures. In addition to the contribution of oxygen vacancies typically located at g = 2.002, another broad absorption peak was observed near g = 1.96. The study reported that hydrogen treatment of rutile TiO₂ and fluoride-doped TiO₂ can induce various shapes of Ti³⁺-related EPR signals at g = 1.96, with peaks near g = 1.96 belonging to the paramagnetic Ti³⁺ center. The signal at g = 2.002 is typically attributed to O-species. Surface O-species may be generated by the interaction between O₂ and surface Ti³⁺ species, which is related to surface oxygen vacancies. Surface O atoms are eliminated through the formation of surface OH species. When the temperature exceeds 500 K, surface OH species in the form of H₂O desorb from the surface, leaving Ti³⁺ species and Ti³⁺ vacancies on the surface. Subsequently, interaction with O₂ in the air ensues, initially resulting in the production of O²⁻, and subsequently, the production of O⁻ species. The excess surface charge associated with Ti³⁺ species contributes to the formation of surface O⁻ species from O²⁻, which are stabilized by the electrophilic characteristics of oxygen and Coulomb interactions with Ti⁴⁺ sites. At lower temperatures (approximately 500 °C), the formation of oxygen vacancies on the surface is favored. Once surface oxygen vacancies with adjacent Ti³⁺ sites are exposed to air, they interact with molecular O₂ to produce O²⁻, which can further react and trap O⁻ species in the surface oxygen vacancies. Furthermore, the hydrogenation of TiO₂ resulted in the removal of residual surface F atoms that had remained on the surface of the TiO₂ nanocrystals throughout the fabrication process. As the temperature of the hydrogenation treatment increased, the signals at g = 1.96 and g = 2.002 exhibited enhanced overall intensity.⁴⁰ The specific surface areas were investigated by N₂ adsorption–desorption measurements. As shown in Fig. 2f, the curve is a typical type IV curve. And the BET surface area of the TiO_2−x catalyst gradually increased from 10 m² g⁻¹ to 38 m² g⁻¹. A larger specific surface area helps to expose reactive sites and increase the probability of contact, which is very beneficial for photocatalytic reactions.

Fig. 2 (a) XRD patterns of TiO₂ and TiO_2−x. (b) XPS spectra of TiO₂, TiO_2−x-450, TiO_2−x-500, TiO_2−x-550 and TiO_2−x-600. (c) O 1s spectra. (d) Ti 2p spectra. (e) EPR spectra recorded in the dark for Ti³⁺ and oxygen vacancies. (f) N₂ adsorption–desorption isotherms of TiO₂, TiO_2−x-450, TiO_2−x-500, TiO_2−x-550 and TiO_2−x-600.

3.2. Energy band structure and charge-transfer properties

To gain a deeper understanding of the mechanism of O-vacancy-induced catalyst optimization, a detailed study was conducted on the relationship between the band structure and photophysical properties of a series of TiO_2−x catalysts, including bandgap energy, transient photocurrent intensity, and impedance. As shown in Fig. 3a, compared with the initial TiO₂ with a bandgap value of 3.03 eV, all TiO_2−x catalysts after vacancy engineering exhibit a narrower bandgap. Among them, the bandgap energy of TiO_2−x-550 is the smallest at 2.79 eV. Furthermore, valence band X-ray photoelectron spectroscopy (VB-XPS) was used to probe the work function and valence band position near the sample surface. From the results presented in Fig. 3b, it can be observed that the valence band (E_VB) values of TiO₂, TiO_2−x-450, TiO_2−x-500, TiO_2−x-550 and TiO_2−x-600 were 2.99, 2.55, 2.25, 2.12 and 2.65 eV, respectively. The schematic diagram of the band structure of a series of TiO_2−x catalysts is shown in Fig. 3c. TiO_2−x-550 has the most advantageous band structure for photocatalytic hydrogen production (still marked as TiO_2−x). As shown in Fig. 3d, the photoluminescence (PL) spectrum of TiO₂ and TiO_2−x shows a trend of decreasing and then increasing spectral intensity with increasing hydrogenation treatment temperature. Among them, the PL intensity of TiO_2−x-550 is the weakest, indicating that TiO_2−x-550 can effectively suppress the recombination of photo generated e⁻–h⁺ pairs. Inhibiting the recombination of e⁻–h⁺ pairs indicates that the separation and transfer of photogenerated charge carriers are more effective in the catalytic reaction process. Subsequently, further verification of its charge transfer resistance through electrochemical impedance spectroscopy showed that TiO_2−x-550 has the lowest impedance among a series of catalytic materials (Fig. 3e), which is conducive to the transfer of photogenerated carriers. Furthermore, the photoelectric conversion ability of a series of TiO₂ catalysts was evaluated by comparing the relative intensity based on the transient photocurrent response. The results indicate that TiO_2−x-550 has the strongest photoelectric conversion ability (Fig. 3f). The above results indicate that the TiO_2−x-550 catalyst exhibits the best photo-responsiveness and carrier separation and transfer ability, and will demonstrate excellent catalytic ability in the photocatalytic process.

Fig. 3 (a) Tauc plots, (b) VB-XPS spectra and (c) schematic diagram of the band structure of a series of TiO₂ catalysts obtained from Tauc plots and VB-XPS. (d) PL spectra, (e) EIS spectra, and (f) transient photocurrent responses of TiO₂, TiO_2−x-450, TiO_2−x-500, TiO_2−x-550, and TiO_2−x-600.

3.3. Photocatalytic hydrogen isotope separation performance

The study on photocatalytic hydrogen isotope separation performance of the TiO_2−x catalyst was analyzed in detail from the following aspects. First, in order to experimentally verify the structural design, the photocatalytic hydrogen production performance (Fig. S6) and hydrogen isotope separation efficiency (Fig. S7) of TiO₂, TiO_2−x-450, TiO_2−x-500, TiO_2−x-550, and TiO_2−x-600 were preliminarily evaluated under the conditions of a 6% addition of co-catalyst Pt and a 10% addition of sacrificial agent methanol. The separation factor of the series of TiO₂ catalysts showed a trend that first increased and then decreased in the range of 4.04 to 4.23 in a H₂O : D₂O ratio of 1 : 1 (Fig. S7). TiO_2−x-550 exhibited the best performance in both hydrogen production and separation. Subsequently, the optimal amount of Pt co-catalyst was studied, and with the increase in co-catalyst (1%, 2%, 4%, and 6%), the H₂ production of TiO_2−x-550 also showed a trend that first increased and then decreased (Fig. S8). As shown in Fig. 4a, the optimal catalytic hydrogen production performance was 10.96 mmol g⁻¹ h⁻¹ when the Pt addition amount was 2 wt%. As shown in Fig. 4b, under the same conditions using different catalysts (TiO₂, TiO_2−x-450, TiO_2−x-500, TiO_2−x-550, and TiO_2−x-600), the H₂ generation rate of the original TiO₂ is lower, at 8.96 mmol g⁻¹ h⁻¹. After the implementation of oxygen vacancy engineering, the evolution rate of H₂ experienced a transition from an increasing period to a decreasing period. TiO_2−x-550 showed the optimal hydrogen production rate of 10.96 mmol g⁻¹ h⁻¹. Based on the above results, we chose TiO_2−x-550 as the catalyst for the subsequent experimental group, still marked TiO_2−x. In addition, the stability of the catalyst was evaluated after six consecutive cycles. As shown in Fig. S9, after six cycles, the hydrogen production rate of the catalyst still remained above 92% of the initial efficiency, with only a slight decrease. The XRD characterization of the catalyst after the reaction (SI, Fig. S9b) showed that the crystal structure and chemical state of the catalyst did not change significantly. These results fully prove that the TiO_2−x catalyst has good stability and reusability.

Fig. 4 (a) H₂ evolution rates of TiO_2−x under different Pt addition conditions. (b) H₂ evolution rates of TiO₂, TiO_2−x-450, TiO_2−x-500, TiO_2−x-550, and TiO_2−x-600 in pure H₂O. (c) Changes in the generation of total gas, H₂, HD, and D₂ over time and kinetic fitting (TiO_2−x catalyst: 8 mg, H₂O : D₂O = 50 : 50). (d) Under different H₂O input conditions, the percentage of total gas, H₂, HD, and D₂ production under illumination. (e) Total gas, H₂, HD and D₂ production for different H inputs on TiO_2−x. (f) Seven different input hydrogen isotope separation factors and total gas production rate for H₂O : D₂O. (Unless otherwise marked, the catalyst is 8 mg TiO_2−x, Pt 2 wt%, methanol 10%, reaction time 4 h.)

After the above experiments, the optimal amount of co-catalyst introduced was determined to be 2 wt%, and the optimal catalyst was TiO_2−x. Further detailed investigations were carried out on the hydrogen isotope separation performance of the catalyst. In a series of hydrogen isotope separation experiments, we repeated these measurements using the TiO_2−x catalyst in environments with different H₂O : D₂O ratios, and observed the differences in H₂, HD, and D₂ content in the output gas. Fig. 4c shows the changes and kinetic fitting curves of total gas, H₂, HD, and D₂ content in the output gas of the TiO_2−x catalyst in an H₂O : D₂O ratio of 1 : 1 environment over time. The variation of gas production over time under other component conditions is shown in Fig. S10 of the SI. We observed that the composition of gas products is strongly influenced by the relative concentration of hydrogen ions in the input water (H₂O/D₂O), where the sum of H and D is equal to 100%. By measuring gas products through chromatography, we quantified the percentages of H₂, HD and D₂ in the output gas, which can be easily converted into the percentages of H and D atoms at the output end, expressed as a function of the input end H or D. We found that the output fraction of hydrogen atoms is disproportionately high compared to the input fraction of protons (Fig. 4d). For example, when the ratio of protons to deuterium is 50 : 50 at the input, hydrogen accounts for about 84% of the atoms in the output stream, indicating that TiO_2−x effectively screens deuterium in situ. Fig. 4e shows the total gas, H₂, HD and D₂ production under different amounts of H₂O addition after 4 hours of continuous illumination. It is worth noting that under the same conditions, the rate of H₂ production is higher than that of D₂ production. Specifically, in pure D₂O, the rate of D₂ production is 4.38 mmol g⁻¹ h⁻¹, while in pure H₂O, the rate of H₂ production is 10.96 mmol g⁻¹ h⁻¹. This also indicates that TiO_2−x can effectively achieve hydrogen isotope separation. And the hydrogen isotope separation factor was calculated according to eqn (1).

when H₂O : D₂O = 10 : 90, the highest separation factor is 5.31 (Fig. 4f). In addition, under the same photocatalytic reaction conditions, H₂O and D₂O were added to two different reaction vessels, and the rate constants of the TiO_2−x catalyst were independently measured to be ν_H = 10.96 mmol g⁻¹ h⁻¹ and ν_D = 2.05 mmol g⁻¹ h⁻¹ for these two reactions. According to KIE = ν_H/ν_D, the kinetic isotope effect (KIE) of photocatalytic production of H₂ and D₂ is 5.35, significantly greater than 2, indicating that this reaction is controlled by first-order isotope effects. The performance comparison between the photocatalytic strategy used for liquid-phase hydrogen isotope separation in the manuscript and other methods used for liquid-phase hydrogen isotope separation is shown in Table S4.

3.4. Mechanism analysis

The water contact angle was measured on the surfaces of TiO₂ and TiO_2−x. A low contact angle value indicates that water molecules tend to diffuse and adhere to the surface, while a high contact angle value indicates that the surface tends to repel water.¹⁴ As shown in Fig. 5a and b, the contact angles of TiO₂ and TiO_2−x with H₂O droplets are smaller than those with D₂O, indicating that TiO₂ and TiO_2−x have hydrophilic surfaces. For H₂O droplets, the contact angle of TiO_2−x is slightly smaller, indicating that TiO_2−x has a higher probability of preferentially contacting H₂O. To elucidate the reaction mechanism of TiO_2−x for photocatalytic hydrogen isotope separation, a possible separation mechanism was validated through DFT theoretical analysis and characterization methods such as contact angle measurements. The binding abilities of the (101) and (001) crystal planes of TiO₂ and TiO_2−x catalysts with H₂O were compared theoretically, with each point labeled as P1, P2, P3, and P4. The adsorption energies of the four sites are −2.22, −1.91, −1.81, and −0.98 eV, respectively, indicating that H₂O in the reaction environment is more likely to bind to the (001) crystal plane of the TiO_2−x catalyst (Fig. S11). This is highly consistent with the idea that the originally designed (001) crystal surface catalyst has more advantages. Research has shown that after oxygen vacancy engineering, a charge enrichment region is formed in the vacancy area, which will help capture H₂O in the environment and preferentially participate in catalytic reactions. The zeta potential shown in Fig. S12 was compared with the zeta potential differences of the series of catalytic materials at pH 3–9. The isoelectric point of all materials is within the range of 6.0–6.5. As the concentration of oxygen vacancies in the material increases, its isoelectric point (IEP) shifts towards a higher pH direction. In contrast, positively charged surfaces are conducive to electron migration towards the surface.

Fig. 5 The contact angle characterization of (a) TiO₂ to H₂O and D₂O droplets. The contact angle characterization of (b) TiO_2−x to H₂O and D₂O droplets. (c) In situ FTIR spectra of catalytic processes on TiO_2−x samples. (d) and (e) Partial enlarged images (H₂O : D₂O = 50 : 50).

The adsorption state of liquid water on the catalyst surface in the environment was analyzed by FT-IR. As shown in Fig. 5c–e, the TiO_2−x catalyst was exposed to a xenon lamp light source for 30 minutes in an environment with H₂O : D₂O = 50 : 50, and offline spectra were collected every 2 min.^41–44 In a series of infrared spectra, no additional signals were observed except for H₂O, HDO and D₂O (Fig. 5c). At the initial moment without light, an intensity ratio of approximately 1 : 2 : 1 was observed for H₂O, HDO, and D₂O at 1645, 1457, and 1210 cm⁻¹ (Fig. 5d). During the 30 min testing process, the bending vibration absorption peak of H₂O showed a red shift from 1651 cm⁻¹ to 1641 cm⁻¹, indicating a weakening of hydrogen bonds during this process, while the bending vibration peak of D₂O showed a blue shift from 1209 cm⁻¹ to 1217 cm⁻¹. A similar trend was observed in the 2200–3600 cm⁻¹ region of the symmetric stretching vibration peaks of H–O–H or D–O–D in H₂O and D₂O, with a more significant red or blue shift (Fig. 5e). It is worth noting that a clear shoulder peak appeared between 2300 and 2400 cm⁻¹, possibly due to isotope exchange reactions occurring in a mixed environment of H₂O and D₂O. Regarding this, as shown in Fig. S13, in an environment containing only D₂O, the bending vibration peak in the 1000–2000 cm⁻¹ fingerprint region gradually conforms to an intensity ratio of approximately 1 : 2 : 1 as time goes on, instead of initially not appearing as a bending vibration peak belonging to H₂O, HDO, and D₂O. This observation supports the evidence of isotope exchange reactions accompanying the catalytic process (eqn (S1)).

Based on the above characterization and experimental results, a detailed schematic diagram of the process is presented in Fig. 6 to provide a more intuitive understanding of the potential mechanism of hydrogen isotope separation of TiO_2−x catalysts. First, upon receiving of light, the TiO_2−x catalyst accumulates electrons with sufficient energy in the valence band, which subsequently transition to the conduction band as photogenerated electrons. These electrons can be captured by unsaturated covalent bonds near oxygen vacancies. However, the defect energy levels resulting from oxygen vacancies are shallow, leading to a transient capture. This helps to transfer more photogenerated electrons to the surface of the TiO_2−x catalyst, enabling it to participate in the photocatalytic water splitting process. It has been demonstrated through contact angle that H₂O will make a more active contact with the catalyst surface compared to D₂O. In a liquid environment with a 1 : 1 ratio of H₂O : D₂O, infrared results show that in a mixed liquid environment, the isotope exchange reaction, as shown in eqn (S1), will first rapidly occur to a stable state, followed by a blue shift of the δ(O–H) peak over time indicating a weakening of hydrogen bonds, while the opposite is true for δ(O–D). The difference in bond energy between H–O and D–O leads to a difference in reaction rate between the catalyst and D₂O, laying the foundation for effective hydrogen isotope separation in mixed systems of H₂O and D₂O. The final performance showed that the total gas production rate of the catalyst in pure D₂O was 4.38 mmol g⁻¹ h⁻¹, which was lower than the gas production rate of 10.96 mmol g⁻¹ h⁻¹ in pure H₂O, while the maximum separation factor was 5.31.

Fig. 6 Schematic diagram of the possible mechanism of TiO_2−x for photocatalytic hydrogen isotope separation.

4. Conclusion

In summary, the blue TiO_2−x catalyst was successfully prepared based on the exposed (001) crystal plane TiO₂ material through oxygen vacancy engineering. The results indicate that the exposed (001) crystal plane exhibits significant preferential selectivity towards H₂O. The presence of oxygen vacancies optimizes the band structure of the original TiO₂, promotes the separation and transfer of charges generated by photoexcitation, and ensures its performance in photocatalytic water splitting (the photolysis rate of H₂ in aquatic products is 10.96 mmol g⁻¹ h⁻¹). In the mixed system of H₂O and D₂O, the D–O bond is slightly stronger than the H–O bond, and the preferential selectivity of the exposed (001) crystal plane TiO_2−x catalyst for H₂O provides strong support for the separation of hydrogen isotopes in water. Under illumination conditions, when the H₂O : D₂O ratio is 10 : 90, the hydrogen isotope separation coefficient of the TiO_2−x catalyst reaches 5.31, achieving effective separation of hydrogen isotopes in the mixed system.

Authors contributions

Linzhen Wu: methodology, investigation, data curation, writing – original draft. Hongbo Li: methodology, formal analysis. Weiwei Wang: methodology. Shengtai Zhang: methodology. Sifan Zeng: funding acquisition, writing – review & editing, resources, supervision. Xiaosong Zhou: conceptualization, methodology, writing – review & editing, funding acquisition, resources, supervision.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supplementary information is available. See DOI: https://doi.org/10.1039/d5ta04254b.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (No. 22106151).

References

1. H. S. Taylor, New Prospects in Isotope Separation, Nature, 1939, 144, 8–9.
2. M. Liu, L. Zhang, M. A. Little, V. Kapil, M. Ceriotti, S. Yang, L. Ding, D. L. Holden, R. Balderas-Xicohténcatl, D. He, R. Clowes, S. Y. Chong, G. Schütz, L. Chen, M. Hirscher and A. I. Cooper, Barely porous organic cages for hydrogen isotope separation, Science, 2019, 366, 613–620.
3. J. Y. Kim, R. Balderas-Xicohtencatl, L. Zhang, S. G. Kang, M. Hirscher, H. Oh and H. R. Moon, Exploiting Diffusion Barrier and Chemical Affinity of Metal-Organic Frameworks for Efficient Hydrogen Isotope Separation, J. Am. Chem. Soc., 2017, 139, 15135–15141.
4. R. Muhammad, S. Jee, M. Jung, J. Park, S. G. Kang, K. M. Choi and H. Oh, Exploiting the Specific Isotope-Selective Adsorption of Metal–Organic Framework for Hydrogen Isotope Separation, J. Am. Chem. Soc., 2021, 143, 8232–8236.
5. L. Zhang, S. Jee, J. Park, M. Jung, D. Wallacher, A. Franz, W. Lee, M. Yoon, K. Choi, M. Hirscher and H. Oh, Exploiting Dynamic Opening of Apertures in a Partially Fluorinated MOF for Enhancing H₂ Desorption Temperature and Isotope Separation, J. Am. Chem. Soc., 2019, 141, 19850–19858.
6. D. Cao, H. Huang, Y. Lan, X. Chen, Q. Yang, D. Liu, Y. Gong, C. Xiao, C. Zhong and S. Peng, Ultrahigh effective H₂/D₂ separation in an ultramicroporous metal–organic framework material through quantum sieving, J. Mater. Chem. A, 2018, 6, 19954–19959.
7. F. Gao, X. Wang, W. Chen, W. Wang, W. Fan, Z. Kang, R. Wang, H. Guo, Q. Yue, D. Yuan and D. Sun, Metal–organic frameworks for hydrogen isotopes separation, Coord. Chem. Rev., 2024, 518, 216047.
8. M. Lozada-Hidalgo, S. Zhang, S. Hu, V. G. Kravets, F. J. Rodriguez, A. Berdyugin, A. Grigorenko and A. K. Geim, Giant photoeffect in proton transport through graphene membranes, Nat. Nanotechnol., 2018, 13, 300–303.
9. J. Y. Kim, L. Zhang, R. Balderas-Xicohténcatl, J. Park, M. Hirscher, H. R. Moon and H. Oh, Selective Hydrogen Isotope Separation via Breathing Transition in MIL-53(Al), J. Am. Chem. Soc., 2017, 139, 17743–17746.
10. Y. Iwai, T. Yamanishi, S. O'Hira, T. Suzuki, W. M. Shu and M. Nishi, H–D–T cryogenic distillation experiments at TPL/JAERI in support of ITER, Fusion Eng. Des., 2002, 61–62, 553–560.
11. P. Li, X. Hu, Z. Zhang, X. Zhang, N. Zeng, C. Hu, B. Yu, X. Hu, J. Song, Y. Shi, L. Zhou and W. Luo, NiAl-layered double hydroxides stabilized Pt clusters with enhanced metal–support interaction: boosting hydrogen isotope separation, Appl. Surf. Sci., 2023, 611, 155780.
12. H. Oh, S. B. Kalidindi, Y. Um, S. Bureekaew, R. Schmid, R. A. Fischer and M. Hirscher, A Cryogenically Flexible Covalent Organic Framework for Efficient Hydrogen Isotope Separation by Quantum Sieving, Angew. Chem., Int. Ed., 2013, 52, 13219–13222.
13. X. Xue, X. Chu, M. Zhang, F. Wei, C. Liang, J. Liang, J. Li, W. Cheng, K. Deng and W. Liu, High Hydrogen Isotope Separation Efficiency: Graphene or Catalyst?, ACS Appl. Mater. Interfaces, 2022, 14, 32360–32368.
14. J. Liang, X. Zhang, T. Q. Liu, X. D. Gao, W. B. Liang, W. Qi, L. J. Qian, Z. Li and X. M. Chen, Macroscopic Heterostructure Membrane of Graphene Oxide/Porous Graphene/Graphene Oxide for Selective Separation of Deuterium Water from Natural Water, Adv. Mater., 2022, 34, e2206524.
15. S. Krause, Active Separation of Water Isotopologues by Local Molecular Motion in Microporous Framework Materials, Angew. Chem., Int. Ed., 2023, 62, e202217680.
16. M. Lozada-Hidalgo, S. Hu, O. Marshall, A. Mishchenko, A. N. Grigorenko, R. A. W. Dryfe, B. Radha, I. V. Grigorieva and A. K. Geim, Sieving hydrogen isotopes through two-dimensional crystals, Science, 2016, 351, 68–70.
17. Z. Junbo, W. Kuisheng, S. Huitang and W. Shaobo, Dynamic equations of impurity hydrogen during heavy water electrolysis, Int. J. Hydrogen Energy, 2004, 29, 1393–1396.
18. K. Harada, R. Tanii, H. Matsushima, M. Ueda, K. Sato and T. Haneda, Effects of water transport on deuterium isotope separation during polymer electrolyte membrane water electrolysis, Int. J. Hydrogen Energy, 2020, 45, 31389–31395.
19. Q. Zhao, M. Pang, C. Tang, X. Xiang, X. Wang, J. Chen and C. Chen, Molybdenum disulfide nanosheets rich in edge sites for efficient hydrogen isotope separation by water electrolysis, Electrochim. Acta, 2023, 464, 142780.
20. C. W. Park, E. Jeong, I. Kim, H.-J. Kim, H.-M. Yang, Y.-H. Sihn and I.-H. Yoon, Combined water electrolysis and 2D hydron separator for enhanced hydrogen isotope separation, Chem. Eng. J., 2024, 498, 155328.
21. D. He, L. Zhang, T. Liu, R. Clowes, M. A. Little, M. Liu, M. Hirscher and A. I. Cooper, Hydrogen Isotope Separation Using a Metal–Organic Cage Built from Macrocycles, Angew. Chem., Int. Ed., 2022, 61, e202202450.
22. J. Xu, R. Li, X. Yan, Q. Zhao, R. Zeng, J. Ba, Q. Pan, X. Xiang and D. Meng, Platinum single atom catalysts for hydrogen isotope separation during hydrogen evolution reaction, Nano Res., 2022, 15, 3952–3958.
23. A. M. Bornea, M. Zamfirache, G. Ana, L. Stefan, O. Balteanu and C. Bucur, The Study of CECE Process for Low-Tritiated Liquid Waste Prior to Experimental Phase, Fusion Sci. Technol., 2020, 76, 384–391.
24. T. Sugiyama, Y. Asakura, T. Uda, T. Shiozaki, Y. Enokida and I. Yamamoto, Present status of hydrogen isotope separation by CECE process at the NIFS, Fusion Eng. Des., 2006, 81, 833–838.
25. I. A. Alekseev, S. D. Bondarenko, O. A. Fedorchenko, T. V. Vasyanina, K. A. Konoplev, E. A. Arkhipov, T. V. Voronina, A. I. Grushko, A. S. Tchijov and V. V. Uborsky, Heavy water detritiation by combined electrolysis catalytic exchange at the experimental industrial plant, Fusion Eng. Des., 2003, 69, 33–37.
26. L. Wu, S. Zeng, W. Wang, S. Zhang, H. Li and X. Zhou, Efficient hydrogen isotope separation utilizing photocatalytic capability, J. Mater. Chem. A, 2024, 12, 33133–33141.
27. X. Feng, Q. Zang, X. Feng, B. Lv, H. Yu, T. Sun, Z. Yuan, J. Liu, Y. Yang and F. Zhang, Enhanced Solar-to-Hydrogen Conversion and Hydrogen Isotope Separation through Interfacial Hydrogen-Bond Engineering and Homolytic O–H Cleavage on Multianionic Sulfides in Large-Scale Floating Nanocomposites, ACS Catal., 2024, 14, 9077–9092.
28. S. Zeng, L. Wu, W. Bing, W. Wang, H. Li, X. Zhou and S. Peng, Constructing O doped g-C₃N₄ nanosheets as photocatalysts for photocatalytic water splitting derived hydrogen isotopic water separation, Sep. Purif. Technol., 2025, 359, 130880.
29. P. He, L. Zhang, L. Wu, S. Xiao, X. Ren, R. He, X. Yang, R. Liu and T. Duan, Synergy of oxygen vacancies and thermoelectric effect enhances uranium(VI) photoreduction, Appl. Catal., B, 2023, 322, 122087.
30. Q. Yan, J. Wang, L. Zhang, J. Liu, M. Wahiduzzaman, N. Yan, L. Yu, R. Dupuis, H. Wang, G. Maurin, M. Hirscher, P. Guo, S. Wang and J. Du, A squarate-pillared titanium oxide quantum sieve towards practical hydrogen isotope separation, Nat. Commun., 2023, 14, 4189.
31. S. El Hakim, T. Chave and S. I. Nikitenko, Deciphering the reaction mechanisms of photothermal hydrogen production using H/D kinetic isotope effect, Catal. Sci. Technol., 2022, 12, 5252–5256.
32. U. Diebold, The surface science of titanium dioxide, Surf. Sci. Rep., 2003, 48, 53–229.
33. M. Lazzeri, A. Vittadini and A. Selloni, Structure and energetics of stoichiometric TiO₂ anatase surfaces, Phys. Rev. B, 2001, 63, 155409.
34. X. Han, Q. Kuang, M. Jin, Z. Xie and L. Zheng, Synthesis of Titania Nanosheets with a High Percentage of Exposed (001) Facets and Related Photocatalytic Properties, J. Am. Chem. Soc., 2009, 131, 3152–3153.
35. A. Sinhamahapatra, J.-P. Jeon and J.-S. Yu, A new approach to prepare highly active and stable black titania for visible light-assisted hydrogen production, Energy Environ. Sci., 2015, 8, 3539–3544.
36. T. Zhao, K. Wang, S. Zhang, R. Wang, Y. Chen and S.-H. Ho, Tuned targeted catalytic engineering enables high-selective electrochemical low-concentration nitrate-to-ammonia, Appl. Catal. B Environ. Energy, 2025, 361, 124693.
37. Z. Wang, C. Yang, T. Lin, H. Yin, P. Chen, D. Wan, F. Xu, F. Huang, J. Lin, X. Xie and M. Jiang, Visible-light Photocatalytic, Solar Thermal and Photoelectrochemical Properties of Aluminium-reduced Black Titania, Energy Environ. Sci., 2013, 6, 3007–3014.
38. J. Qu, Y. Wang, X. Mu, J. Hu, B. Zeng, Y. Lu, M. Sui, R. Li and C. Li, Determination of Crystallographic Orientation and Exposed Facets of Titanium Oxide Nanocrystals, Adv. Mater., 2022, 34, e2203320.
39. H. Li, Y. Zeng, T. Huang, L. Piao, Z. Yan and M. Liu, Hierarchical TiO₂ nanospheres with dominant 001 facets: facile synthesis, growth mechanism, and photocatalytic activity, Chemistry, 2012, 18, 7525–7532.
40. X. Yu, B. Kim and Y. K. Kim, Highly Enhanced Photoactivity of Anatase TiO₂ Nanocrystals by Controlled Hydrogenation-Induced Surface Defects, ACS Catal., 2013, 3, 2479–2486.
41. M. Unger, Y. Ozaki, F. Pfeifer and H. W. Siesler, 2DCOS and PCMW2D analyses of FT-IR/ATR and FT-NIR spectra monitoring the deuterium/hydrogen exchange in liquid D₂O, J. Mol. Struct., 2014, 1069, 258–263.
42. X. Yan, Y. Song, D. Wang, T. Xia, X. Tan, J. Ba, T. Tang, W. Luo, G. Sang and R. Xiong, Direct observation of highly effective hydrogen isotope separation at active metal sites by in situ DRIFT spectroscopy, Chem. Commun., 2023, 59, 3922–3925.
43. M. Ceriotti, W. Fang, P. G. Kusalik, R. H. McKenzie, A. Michaelides, M. A. Morales and T. E. Markland, Nuclear Quantum Effects in Water and Aqueous Systems: Experiment, Theory, and Current Challenges, Chem. Rev., 2016, 116, 7529–7550.
44. M. Pastorczak, K. Duk, S. Shahab and A. A. Kananenka, Combinational Vibration Modes in H₂O/HDO/D₂O Mixtures Detected Thanks to the Superior Sensitivity of Femtosecond Stimulated Raman Scattering, J. Phys. Chem. B, 2023, 127, 4843–4857.

---