Engineering a Coordinatively Unsaturated Au–O–Ti3+ Structure Toward Unprecedented H2 Efficiency for Low-Temperature Propene Epoxidation with H2 and O2

Abstract

Since 1998, the Au–O–Ti⁴⁺ sites of Au/Ti-based catalysts have been widely accepted as the active sites for propene epoxidation with H₂ and O₂ at a relatively high temperature, although they are limited by poor H₂ efficiency. Herein, we demonstrate a novel Au–O–Ti³⁺ active site aiming at low-temperature propene epoxidation. Notably, this active site results in a sharp shift in the optimum temperature, from 200 to 138 °C, and allows the catalyst to maintain an unprecedented H₂ efficiency of 43.6%, a high propylene oxide (PO) selectivity of 90.7%, and a stability of over 100 h. The Au–O–coordinatively unsaturated Ti³⁺ active site is quantitively constructed by tuning the amount of Si–OH and Bu₃NH⁺ in post-treated silicalite-1 seeds. Through operando ultraviolet–visible (UV–vis) spectroscopy, the dynamic evolution of the Ti–OOH intermediate was investigated. It was found that the Ti–OOH generation rate is higher on Au–O–Ti³⁺ than on conventional Au–O–Ti⁴⁺ sites. Moreover, ammonia temperature-programmed desorption (NH₃-TPD) and X-ray photoelectron spectroscopy (XPS) characterizations, together with density-functional theory (DFT) calculations, demonstrated that the coordinatively unsaturated Ti³⁺ sites promote electron transfer between Au and Ti³⁺, thereby enhancing the O₂ adsorption ability of the catalyst and promoting the in situ formation of H₂O₂ and the Ti–OOH intermediate, even at a low temperature. The insights and methodology reported here not only shed new light on maximizing H₂ efficiency over a coordinatively unsaturated Ti³⁺ structure of titanium silicate-1 but also open up new opportunities for industrial direct gas-phase propene epoxidation in a low temperature range.

Content

《1. Introduction》

1. Introduction

As an essential organic raw material, there is a tremendous market for propylene oxide (PO) in the chemical industry, where it is used for the manufacturing of polyether polyol, propylene glycol, and dimethyl carbonate. In this scenario, the utilization of propene, hydrogen (H₂), and oxygen (O₂) to manufacture PO provides a greener, simpler, and more sustainable route than traditional PO synthesis methods. In particular, this green process could liberate the industry from its reliance on traditional chlorohydrin and several organic hydroperoxide processes [1,2].

Since it was first reported in 1998 that gold (Au)/titanium dioxide (TiO₂) catalysts are efficient in direct propene epoxidation with H₂ and O₂ and possess the advantage of superior PO selectivity [3], many studies have focused on the development of Au/titanium (Ti)-containing support catalysts [4–11] to enhance their catalytic performance in industrial applications [12,13]. Au/titanium silicate-1 (TS-1), which possesses Au–O–Ti⁴⁺ active sites, has attracted a great deal of attention due to its remarkable high-temperature propene epoxidation performance with good C₃H₆ conversion (> 5.0%) and PO selectivity (> 90%). Unfortunately, Au/TS-1 catalysts still have the severe limitation of a low H₂ efficiency (only ~22.7%) [6,14–16]. Therefore, the precise construction of Au/Ti bifunctional catalysts and the identification of Au–Ti synergistic effects are highly desirable to overcome the above limitations.

It is notable that the in situ generation and decomposition of hydrogen peroxide (H₂O₂) can directly affect the H₂ efficiency of direct propene epoxidation with H₂ and O₂. A widely accepted reaction pathway has been provided, as follows: First, H₂O₂ is generated from H₂ and O₂ on Au nanoparticles [17,18]; subsequently, the H₂O₂ diffuses to nearby Ti active sites to generate Ti–OOH intermediates that epoxidize the propene to PO [19–21]. Thus, H₂ efficiency can be enhanced by promoting the synergy between Au nanoparticles and Ti sites. In recent years, in addition to the isolated tetrahedral Ti⁴⁺ species, other Ti species have been observed [22–27] and are assumed to be the active sites for activating H₂O₂. However, there are few studies on the synergistic effects of Au and coordinatively unsaturated Ti³⁺ sites on the in situ formation of H₂O₂. Along this route, there is an urgent need to precisely engineer novel Au/Ti bifunctional sites toward the development of industrially attractive Au/Ti bifunctional catalysts for the direct propene epoxidation reaction.

Herein, we devise a strategy to quantitively construct coordinatively unsaturated Ti³⁺ sites on a novel TS-1 (TS-1-cS) support by tuning the amount of Si–OH and Bu₃NH⁺ in post-treated silicalite-1 (S-1) seed. Based on operando ultraviolet–visible (UV– vis) spectroscopy, in situ Fourier-transform infrared (FT-IR) spectroscopy, X-ray photoelectron spectroscopy (XPS), and densityfunctional theory (DFT) calculation, it is confirmed that the presence of coordinatively unsaturated Ti³⁺ sites promotes electron transfer between Au and Ti, thereby facilitating the formation of Au–O–Ti³⁺ active sites. Compared with traditional Au–O–Ti⁴⁺ sites, these Au–O–Ti³⁺ active sites are more conducive to O₂ adsorption and thus effectively promote the in situ generation of H₂O₂ and the formation of the active intermediate Ti–OOH. As a result, the Au/TS-1-cS catalyst has an excellent H₂ efficiency (43.6%) at a relatively low reaction temperature. We further provide a plausible structure–performance relationship between coordinatively unsaturated Ti³⁺ sites and H₂ efficiency. The insights and methodology reported here open up new opportunities for optimizing H₂ efficiency over the coordinatively unsaturated Ti³⁺ structure of TS-1 for novel low-temperature propene epoxidation.

《2. Material and methods》

2. Material and methods

《2.1. Preparation of the catalysts》

2.1. Preparation of the catalysts

The synthesis of S-1 seed was performed following the process reported by Cundy et al. [28]. Subsequently, the S-1 seeds were processed and etched using n-butylamine at different concentrations (0.6, 1.2, and 1.8 mol·L^–1 ), as follows: A certain amount of n-butylamine (99.5 wt%; Sinopharm Chemical Reagent Co., Ltd., China) and 2 g of S-1 seeds were dissolved in 30 mL of deionized water under stirring at room temperature for 0.5 h. The resulting mixture was transfused into a Teflon-lined stainless steel autoclave and hydrothermally crystallized at 443 K for 24 h. Finally, the crystalline solid was centrifuged and washed four times, and then dried at 303 K for 24 h. The as-prepared S-1 seed etched by x mol·L^–1 n-butylamine is denoted as S-1-xM; for example, S-1-0.6M represents S-1 seed etched by 0.6 mol·L^–1 n-butylamine.

The TS-1-cS support was prepared as follows: 5 g of n-butylamine (99.5 wt%) and 3 g of tetrapropylammonium hydroxide (TPAOH, 25 wt%; Sigma-Aldrich Chemical Inc., USA) were mixed in deionized water under stirring. Subsequently, 10 g of colloidal silica (Sinopharm Chemical Reagent Co., Ltd.) was added to this mixture. In the meantime, 0.4 g of titanium(IV) tetrabutoxide (TBOT, 99 wt%; Sinopharm Chemical Reagent Co., Ltd.) was dissolved in 20 mL of isopropanol (IPA, 99.5 wt%; Sinopharm Chemical Reagent Co., Ltd.). The Ti-containing mixture was added dropwise to the Si-containing mixture, while stirring. Afterward, 1 g of the aforementioned treated S-1 seed was poured into the mixture described above. The final solution was poured into a Teflon-lined stainless steel autoclave and hydrothermally crystallized at 443 K for 72 h. Finally, the crystalline solid was centrifuged and washed four times, dried at 393 K for 12 h, and then calcined at 823 K for 6 h. The as-prepared TS-1-cS support prepared with S1-xM seed is denoted as TS-1-cS-xM. For example, TS-1-cS-0.6M represents TS-1-cS support prepared using S-1-0.6M seed. A conventional TS-1 with a similar Si/Ti ratio was synthesized using the hydrothermal method developed by Khomane et al. [29].

The deposition–precipitation method was employed to prepare a series of Au/TS-1-cS catalysts with a unified Au loading of 0.10 wt% [30]. In brief, 0.2 g of hydrogen tetrachloroaurate(IV) trihydrate (HAuCl_4·4H₂O, 99.99%; Sigma-Aldrich Chemical Inc.) was dissolved in 50 mL of deionized water. Subsequently, 0.5 g of TS-1-cS powder was dissolved in the solution with stirring. Subsequently, the solution was maintained at a pH of 6.5–7.5 by adding 1.0 and 0.1 mol·L^–1 NaOH (Sinopharm Chemical Reagent Co., Ltd.) solution with stirring at 298 K for 9 h. The final seriflux was centrifuged and washed by centrifugation (5000 r·min^–1 for 5 min) with ice deionized water, and then dried at 298 K for 12 h under vacuum.

《2.2. Characterizations》

2.2. Characterizations

²⁹Si magic-angle spinning (MAS) nuclear magnetic resonance (NMR) measurements were collected on a Varian Unity INOVA (600 MHz) spectrometer (Varian, USA). UV–vis spectra were recorded on a PerkinElmer LAMBDA 35 spectrophotometer (PerkinElmer, USA) over a wavelength range of 200–800 nm with pure BaSO₄ plate as the reference. The peak area at 220 nm was obtained through fitting and integrating according to the symmetry of the peak. Ammonia temperature-programmed desorption (NH₃-TPD) spectra were obtained on a Quantachrome ChemBET TPR/TPD (Quantachrome, USA). X-ray diffraction (XRD) patterns were collected on an X’pert PRO MPD diffractometer instrument (Panalytical, the Netherlands) using Cu Kα radiation. ³¹P MAS NMR experiments were performed on a 9.4 T Bruker AVANCE III spectrometer (Bruker, Switzerland) using 4 mm rotors at a spinning frequency of 10 kHz (operating at a Larmor frequency of 161.7 MHz for ³¹P). A pulse width of 3.0 μs corresponding to a π/2 flip angle, a recycle delay of 30 s, and a total of 1024 scans were used to collect single-pulse ³¹P MAS NMR spectra. Highresolution transmission electron microscopy (HRTEM) images were taken on a JEOL JEM-2100F microscope (JEOL, Japan). Operando UV–vis spectra were collected on an AvaSpec-2048 spectrometer (Avantes, the Netherlands) equipped with a transmission dip probe. In situ FT-IR spectra were recorded on a Nicolet iS20 instrument (Thermo, USA) equipped with an HgCdTe (MCT) detector cooled by liquid nitrogen. XPS spectra were obtained using a PerkinElmer PHI 5000C ESCA system (PHI, USA), and the carbonaceous C 1s line (284.6 eV) was used as a reference to calibrate the binding energies.

《2.3. Catalytic testing》

2.3. Catalytic testing

The direct propene epoxidation reaction was reacted in a quartz tubular reactor with an inner diameter of 8 mm, containing 0.15 g of Au/TS-1-cS catalyst and utilizing a feed containing propene (C₃H₆), H₂, O₂, and nitrogen (N₂) with a flow rate of 3.5, 3.5, 3.5, and 24.5 mL·min^–1 , respectively. The space velocity was 14 000 mL·h^–1 per gram of catalyst. These conditions are consistent with the literature [1–4,14,15,19,25,29]. Subsequently, the series of Au/TS-1-cS catalysts were tested under atmospheric pressure at the relatively low temperature of 138 °C. The products were analyzed by means of an online gas chromatograph (Agilent 6890, Agilent, USA). A flame ionization detector (FID) with a Porapak T column was used to analyze the PO, ethanal, propanal, acetone, acrolein, and other oxygenates, while a thermal conductivity detector (TCD) was used to analyze the hydrocarbons, H₂, O₂, N₂, carbon dioxide (CO₂), and water (H₂O). Blank evaluations showed that no PO was produced in the catalyst-free reactor. The catalytic performances were calculated using a normalization method defined as follows:

where n_PO, and are the moles of C₃-oxygenates, ethanal, CO₂, C₃H₆ in the feed, PO, and the converted H₂.

《2.4. Density-functional theory calculations》

2.4. Density-functional theory calculations

All calculations were performed using the DMol³ module of Materials Studio 8.0. The meta generalized gradient approximation (meta-GGA) in the form of M06-L was selected to describe the correlation and exchange effects, and double numerical plus polarization (DNP) was used to consider the electronic polarization effect. Brillouin-zone integrations with a smearing of 0.005Ha were used to improve the precise electronic convergence in the calculation. All the energies were corrected by means of the zero-point vibrational energy (ZPE). The allowable deviations of the displacement, gradient, and total energy were 0.0005 nm, 4.35 × 10^–17 J·nm^–1 , and 4.35 × 10^–23 J, respectively. In the calculation process, a 46T TS-1 model (where T refers to Si or Ti) was used to represent the titanium silicon zeolite with an MFI structure. For the TS-1-Ti⁴⁺model, T12 was replaced by a Ti atom to form the perfect tetracoordinated Ti species (Ti⁴⁺) structure. For the TS-1-Ti³⁺ model, when T12 was replaced by a Ti atom, two –OH groups were added to form a type of coordinatively unsaturated hexacoordinated Ti³⁺ structure. Then, the Au₄ cluster was supported on the TS-1-Ti⁴⁺ and TS-1-Ti³⁺ to form the Au₄/TS-1-Ti⁴⁺ and Au₄/TS-1-Ti³⁺ models. The binding energy (BE) or adsorption energy was calculated using the following formula:

where E_mod, E_sub, E_sub+mod, and E_a are the energy of the bare model, the energy of the adsorbed or bound substrate, the total energy of the substrate on the calculation model, and the adsorption energy of adsorbate, respectively.

《3. Results and discussion》

3. Results and discussion

《3.1. A novel strategy to quantitively construct coordinatively unsaturated Ti3+ sites》

3.1. A novel strategy to quantitively construct coordinatively unsaturated Ti³⁺ sites

A series of TS-1 supports with controllable coordinatively unsaturated Ti³⁺ sites (TS-1-cS) were developed to anchor Au precursors as bifunctional catalysts. First, TS-1-cS supports with controllable Ti³⁺ contents were prepared by regulating the properties of S-1 seeds. The²⁹Si MAS NMR spectra of S-1 seeds treated by n-butylamine with different concentrations (i.e., 0.6, 1.2, and 1.8 mol·L^–1 ) are shown in Fig. 1(a). The Q3 peak at –102 ppm is ascribed to the (SiO)₃SiOH groups, the higher Q4 peak at –111 ppm is ascribed to the Si(OSi)₄ groups, and the peak at –114 ppm is ascribed to the monoclinic structure. The relative Q3/Q4 ratio represents the relative content of (SiO)₃SiOH in the S-1 seeds. As can be seen from Fig. 1(b), the relative Q3/Q4 ratio of the S-1 treated by 0.6 mol·L^–1 n-butylamine is 0.205. With an increase in n-butylamine concentration, the relative Q3/Q4 ratio shows an upward trend, demonstrating a regular increase in the Si–OH proportion in the S-1 seeds, and the relative Q3/Q4 ratio of the S-1 treated by 1.8 mol·L^–1 n-butylamine reaches 0.410. Thus, treating the S-1 seeds with n-butylamine with different concentrations resulted in different Si–OH proportions.

《Fig. 1》

Fig. 1. (a) ²⁹Si MAS NMR spectra of S-1 seeds treated by different concentrations of n-butylamine. (b) Relative Q3/Q4 ratio of ²⁹Si MAS NMR spectra for the S-1 seeds. (c) UV–vis spectra and (d) relative ratio of UV–vis spectra of the TS-1-cS supports. (e) ³¹P MAS NMR and (f) NH₃-TPD spectra of corresponding TS-1-cS supports. (g) Plausible schematics of the TS-1-cS synthesis mechanism.

The amount of n-butylamine interacting with the S-1 seeds was characterized by means of inorganic elements analysis to test the corresponding nitrogen (N) element contents. The results showed that the N element contents of S-1-0.6M, S-1-1.2M, and S-1-1.8M were 0.89%, 1.17%, and 1.35%, respectively, indicating that the interaction between n-butylamine and S-1 seeds was strengthened as the concentration of n-butylamine increased. The pore structures of the S-1 seeds treated by n-butylamine with different concentrations were also investigated via N₂ physisorption characterization. The results are provided in Table S1 in Appendix A, which shows that the surface area and pore volume of S-1-xM seeds increase with an increase in n-butylamine treatment concentration. This difference in seed pore structure can lead to a change in the content of Si–OH defect sites, which is consistent with the results obtained via ²⁹Si MAS NMR. Fig. S1 in Appendix A shows transmission electron microscopy (TEM) images of the three S-1 seeds. The average crystal size of the three S-1 seeds is quite similar, at about 140 nm. The as-formed S-1 seeds were used to synthesize TS-1-cS.

The specific environment of the Ti species and the acidity of the TS-1-cS samples were studied in detail using inductively coupled plasma (ICP), UV–vis, ³¹P MAS NMR, and NH₃-TPD characterizations. The Ti contents of TS-1-cS-0.6M, TS-1-cS-1.2M, and TS-1-cS-1.8M were found to be 2.07%, 2.10%, and 2.08%, respectively, indicating that the three supports had almost the same Ti content. Fig. 1(c) shows the detailed composition of the Ti species in the TS-1-cS samples. The peak at 220 nm can be attributed to the isolated tetracoordinated Ti species (Ti⁴⁺), whereas the peak at 258 nm can be attributed to the pentacoordinated or hexacoordinated Ti species (Ti³⁺) [22,25]. Therefore, the normalized area ratio (A_258nm/A_220nm+258nm, equal to ) can be used as a descriptor to reflect the Ti³⁺ species proportion. As can be seen from Fig. 1(d), with an increase in Si–OH content in the S-1 seeds, the ratio in the corresponding TS-1-cS samples gradually increases from 0.133 to 0.228, indicating that the relative content of the Ti³⁺ species regularly increases from TS-1-cS-0.6M to TS-1-cS-1.8M. This result is consistent with the result from XPS in Section 3.3; that is, the Ti³⁺ species content shows a regular increasing trend from TS-1-cS-0.6M to TS-1-cS-1.8M.

In addition, various titanium species can be distinguished by the different phosphorus signals of trimethyl-phosphine (TMP) on the zeolite. Fig. 1(e) shows the ³¹P MAS NMR spectra of the TMP adsorbed on the series of TS-1-cS samples. The resonance peak at –4.5 ppm is assigned to the TMP adsorbed onto the sites (e.g. Si–OH–Ti, Ti–OH, or Si–OH) due to the Brønsted acidic properties of the TS-1-cS samples [31–33]. The resonance peak at –33.3 ppm is assigned to the TMP bound to the coordinatively unsaturated Ti sites [34,35], while the resonance peak at –62.5 ppm is ascribed to the physisorbed TMP [36]. It was found that the relative amount of coordinatively unsaturated Ti sites (A_–33.3ppm/A_{(–33.3ppm)+(–4.5ppm)}) shows a regular upward trend from TS-1-cS-0.6M to TS-1-cS-1.8M. Combined with the UV–vis characterization results described above, it can be found that the change trend in the Ti³⁺ species matches that of the coordinatively unsaturated Ti sites, which means that the coordinatively unsaturated sites are mainly composed of Ti³⁺ species.

NH₃-TPD was adopted to determine the total acidity of the TS1-cS samples, as shown in Fig. 1(f). All the TS-1-cS samples exhibited two NH₃ desorption peaks at 139 and 334 °C, which are ascribed to weak and medium acid centers [37,38]. Although the desorption temperatures of the three samples are the same, the peaks of the TS-1-cS samples differ greatly in intensity. In addition, the amount of weak and medium acid sites gradually increases from TS-1-cS-0.6M to TS-1-cS-1.8M.

The synthesis mechanism of TS-1-cS was further analyzed, as shown in Fig. 1(g). It has been reported that, during the hydrothermal synthesis process, isomorphous substitution occurs between Si atoms in the seed framework and Ti atoms in the precursor solution [39–41]. Therefore, it can be assumed that the higher the content of Si–OH vacancies in the seed framework, the higher the content of coordinatively unsaturated Ti vacancies in the TS-1-cS that can be obtained via isomorphous substitution. Moreover, the surfaces of the S-1 seeds are grafted with Bu₃NH⁺ after treatment with n-butylamine [42] in a strong alkali environment. Therefore, due to the insufficient amount of TPAOH, the Bu₃NH⁺ in the seeds can play an auxiliary guiding role of a template agent in the synthesis process of TS-1-cS [43]. Furthermore, the Ti-containing tetrahedron structure is unstable because of the large ionic radius of Ti [44]. Therefore, coordinatively unsaturated Ti³⁺ species are likely to form with a weak template effect of Bu₃NH⁺ [45]. In summary, the presence of Si–OH and Bu₃NH⁺ in the S-1 seeds can induce the formation of coordinatively unsaturated Ti³⁺ species.

《3.2. Low-temperature propene epoxidation performance at Au–O–Ti3+ sites》

3.2. Low-temperature propene epoxidation performance at Au–O–Ti³⁺ sites

The series of Au/TS-1-cS catalysts were evaluated in lowtemperature direct propene epoxidation with H₂ and O₂ (Fig. 2) at only 138 °C, which is much lower than the traditional temperature of 200 °C. As the Au/TS-1-cS catalyst with the least amount of coordinatively unsaturated Ti³⁺ species, Au/TS-1-cS-0.6M exhibited a stable PO formation rate of only 75.4 g·h^–1 per kilogram of catalyst (kg_cat) and a H₂ efficiency of 29.8%. Obviously, with an increase in the content of coordinatively unsaturated Ti³⁺ species, the PO formation rate and H₂ efficiency of these Au/TS-1-cS catalysts gradually increased. As the content of the coordinatively unsaturated Ti³⁺ species in the TS-1-cS reached 22.8%, the PO formation rate and H₂ efficiency of Au/TS-1-cS-1.8M significantly rose to 135.7 g·h^–1 · and 43.6%, respectively. It is worth mentioning that the stable and high PO formation rate of Au/TS-1-cS-1.8M at the relatively low reaction temperature of 138 °C is comparable with that of a conventional Au/TS-1 catalyst with isolated Ti⁴⁺ species at the higher reaction temperature of 200 °C (155.1 g·h^–1 per kilogram of TS-1; Fig. S2 in Appendix A). Moreover, a H₂ efficiency of 43.6% is unprecedented in low-temperature direct propene epoxidation with H₂ and O₂. However, with a further increase in the content of the coordinatively unsaturated Ti³⁺ species, the PO formation rate and H₂ efficiency of Au/TS-1-cS-2.4M showed a volcano trend, probably because an excessive number of Ti³⁺ sites can aggravate side reactions—that is, the ring-opening of PO. In addition, according to the thermogravimetric analysis (TGA) and derivative thermogravimetric (DTG) curves of the Au/TS-1-cS catalyst after the reaction at 138 °C for 40 h (Fig. S3 in Appendix A), the excellent stability of the Au/TS-1-cS catalysts is likely to be a result of the inhibition of coke formation due to the unique structure of TS-1-cS, which enhances the mass transfer capacity, as reported in our previous work [15,46,47].

《Fig. 2》

Fig. 2. (a) Reaction process diagram. (b) PO formation rate, (c) H₂ efficiency, and (d) PO selectivity of the Au/TS-1-cS-0.6M, Au/TS-1-cS-1.2M, Au/TS-1-cS-1.8M, and Au/TS-1- cS-2.4M catalysts at the low reaction temperature of 138 °C. kg_cat: kilogram of catalyst.

Moreover, Fig. 2(d) shows that all the catalysts have a comparable PO selectivity of about 91%, indicating that the coordinatively unsaturated Ti³⁺ sites are not very sensitive to the occurrence of side reactions in low-temperature propene epoxidation. As shown in Fig. S4 in Appendix A, as the reaction temperature rises from 138 to 168 °C, the PO selectivity and H₂ efficiency rapidly plummet, while the PO formation rate rises constantly, indicating that the in situ generation rate of H₂O₂ continues to increase. However, a high proportion of H₂O₂ decomposes into H₂O, resulting in the plummeting in H₂ efficiency. The resulting large amounts of H₂O can cause a ring-opening side reaction of PO. As the reaction temperature continues to rise to 200 °C, the PO formation rate, PO selectivity, and H₂ efficiency all rapidly plummet further; the reaction quickly deactivates, and the side reaction is dominant.

At different reaction temperatures, the differences in the PO decomposition ability of the Au/TS-1-cS catalysts with different contents of coordinatively unsaturated Ti³⁺ were analyzed via in situ FT-IR characterization, with the results shown in Fig. 3. The bands at 2976.4, 2938.6, and 2886.7 cm^–1 are assigned to the C–H stretching vibrations of the bidentate propoxy species, which can result from PO decomposition at acidic Ti sites [17,48]. The bands at 1720 cm^–1 are assigned to the C=O stretching vibrations of oxides containing a C=O group (i.e., acetone, aldehyde, or propanal), which can also result from PO decomposition [48]. However, no significant differences are observed between the three catalysts at 1720 cm^–1 (Fig. S5 in Appendix A), because the main byproducts produced from PO decomposition at the acidic Ti sites are bidentate propoxy species, rather than oxides.

As the reaction temperature increased from 140 to 200 °C, the content of the bidentate propoxy species in all samples increased, indicating that PO decomposition intensified with an increase in the reaction temperature. In addition, it is worth noting that the content of coordinatively unsaturated Ti³⁺ sites in the catalysts had little effect on PO decomposition at a lower reaction temperature (e.g., 140 °C). This finding suggests that a low temperature can significantly inhibit the decomposition of H₂O₂ and dramatically suppress the cleavage of C–O and C–C bonds at the Ti³⁺ acidic sites [25,38,49,50].

It is notable that H₂ efficiency is an important factor affecting catalytic performance in this reaction [25,51,52]. Due to the boosted H₂ efficiency, the Au/TS-1-cS catalyst is more effective at forming Ti–OOH and subsequent PO. The higher H₂ efficiency is likely to be affected by both the Au sites and the coordinatively unsaturated Ti³⁺ species [53–55]. Therefore, it was preliminarily speculated that the coordinatively unsaturated Ti³⁺ species may promote metal– support interaction, thereby promoting the formation of H₂O₂. This speculation will be discussed in detail later in Section 3.3.

《Fig. 3》

Fig. 3. In situ FT-IR spectroscopy results of the (a) Au/TS-1-cS-0.6M, (b) Au/TS-1-cS-1.2M, and (c) Au/TS-1-cS-1.8M catalysts under propene epoxidation conditions (the relative flow rate of C₃H₆, H₂, O₂, and N₂ is 1:1:1:7). (d) In situ FT-IR peak intensity of the bidentate propoxy species of the three Au/TS-1-cS catalysts at different reaction temperatures.

《3.3. Intrinsic reasons for the unique H2 efficiency promoted by the coordinatively unsaturated Ti3+ sites》

3.3. Intrinsic reasons for the unique H₂ efficiency promoted by the coordinatively unsaturated Ti³⁺ sites

H₂ efficiency is a key aspect in the direct propene epoxidation reaction system with H₂ and O₂. According to recent studies [25,56–58], the H₂ efficiency for this reaction system is defined as the ratio of moles of PO generated to those of H₂ converted. For the Au–O–Ti bifunctional active sites, both the Au and Ti species can influence the catalytic performance. The in situ generation of H₂O₂ from H₂ and O₂ on Au nanoparticles is the key step in this reaction. Therefore, the physicochemical properties of the Au in the catalysts were studied first, in order to elucidate the intrinsic reason for the significant changes we observed in catalyst activity and H₂ efficiency (Fig. 4). The ICP results showed that the Au loading of the Au/TS-1-cS-0.6M, Au/TS-1-cS-1.2M, and Au/TS-1-cS-1.8M catalysts was about 0.1 wt%, indicating that the three catalysts have similar Au loadings. Considering that the size effect of Au nanoparticles is a key element in the reaction [4,7,16,59,60], HRTEM was used to study the size distribution and lattice spacing of the Au nanoparticles in the Au/TS-1-cS catalysts. Figs. 4(a)–(c) show that all catalysts exhibit similar average sizes (3.0 nm) of Au nanoparticles, which can eliminate the influence of the Au size effect on catalytic performance [2,61–65]. In addition, Figs. 4(d)–(f) show that the lattice spacing of the Au nanoparticles in the three catalysts is 0.23 nm, corresponding to Au(111), which is a stable surface for Au nanoparticles.

The other key factors affecting the reaction are the structure and crystal morphology of the support. Figs. 4(g)–(i) show that all samples are in the shape of a cuboid with similar average crystal sizes of about 150 nm × 120 nm × 105 nm, which is due to the similar synthesis conditions. These results can eliminate the influences of zeolite crystal morphology and size on the catalytic performance. Figs. 4(j) and (k) display the XRD spectra of the three TS-1-cS supports. It is clear that all the TS-1-cS supports exhibit characteristic peaks at 7.8°, 8.8°, 23.1°, 23.9°, and 24.3°, which agree well with the peaks of a typical MFI topological structure. In addition, the single diffraction peak at 2θ = 24.3° for all three TS1-cS supports indicates a structure with orthorhombic symmetry and the incorporation of Ti in the framework [66].

The N₂ physisorption results for Au/TS-1-cS-0.6M, Au/TS-1-cS1.2M, and Au/TS-1-cS-1.8M are shown in Fig. 4(l) and Table S2 in Appendix A. It can be seen that the N₂ adsorption–desorption isotherms of the three samples are all type-IV isotherms, according to the International Union of Pure and Applied Chemistry (IUPAC)’s classification, which is typical of mesostructured materials. Moreover, with a change in the S-1-xM seed, the surface area and total pore volume of the three samples do not change significantly, which can exclude the influence of pore structure on the catalytic performance. In conclusion, after excluding the size effect of Au nanoparticles and of the structure and crystal morphology of the supports, it is speculated that the key factors affecting the catalytic performance involve the titanium species and the metal–support interaction. Therefore, the relationship between the content of the coordinatively unsaturated Ti³⁺ sites and H₂ efficiency was investigated; Fig. 4(n) shows the results, which exhibit an obvious linear relationship.

《Fig. 4》

Fig. 4. HRTEM images and Au size distributions for the (a, d) Au/TS-1-cS-0.6M, (b, e) Au/TS-1-cS-1.2M, and (c, f) Au/TS-1-cS-1.8M catalysts. (g–i) TEM images of the (g) Au/TS1-cS-0.6M, (h) Au/TS-1-cS-1.2M, and (i) Au/TS-1-cS-1.8M catalysts. (j, k) XRD patterns of the TS-1-cS supports. (l) N₂ adsorption–desorption isotherms of the TS-1-cS supports. (m) Crystal structure diagram of the TS-1-cS supports. (n) Relationship between the content of coordinatively unsaturated Ti³⁺sites and H₂ efficiency.

It is supposed that there are two kinds of Ti species around Au sites in Au/Ti-based catalysts: One kind is adjacent to isolated Ti⁴⁺species [6], and the other is adjacent to coordinatively unsaturated Ti³⁺ species. It is widely recognized that H₂O₂ formed around traditional Au–O–Ti⁴⁺ sites can migrate to the surrounding Ti⁴⁺ sites to form the Ti–OOH intermediate. However, based on the catalytic performance and structural parameters of the series of Au/TS-1- cS catalysts, we consider that the formation rate of the Ti–OOH intermediate is higher on coordinatively unsaturated Au–O–Ti³⁺ sites than on traditional Au–O–Ti⁴⁺ sites. We resorted to operando UV–vis, XPS, and DFT to explore the structure–performance relationship between different Au–O–Ti sites and H₂ efficiency. Operando UV–vis spectroscopy was employed to analyze the change in the Ti–OOH intermediates of the Au/TS-1-cS-0.6M, Au/TS-1-cS-1.2M, and Au/TS-1-cS-1.8M catalysts with different contents of coordinatively unsaturated Ti³⁺ under real reaction conditions. Fig. 5 displays the different peak intensities of the Ti–OOH intermediate, which shows up at approximately 360 nm [67] at various reaction temperatures. In the temperature range of 60–140 °C, the Ti–OOH formation rates of all the Au/TS-1-cS catalysts increase with a rise in reaction temperature. The Au/TS-1-cS1.8M catalyst with the highest content of coordinatively unsaturated Ti³⁺ sites exhibits the highest Ti–OOH intermediate formation rate in the temperature range of 60–140 °C, which is necessarily related to the highest H₂ efficiency and PO formation rate of the Au/TS-1-cS-1.8M catalyst at 138 °C. It is speculated that the outstanding Ti–OOH formation rate derives from the Au–O–Ti³⁺ bifunctional active sites, which can effectively improve the generation of H₂O₂ and the further formation of the active Ti–OOH intermediate. The form of the Au–O–Ti³⁺bifunctional active sites is provided in Fig. 5(a), where R is mainly used to show all the possible types of Ti³⁺ species. Based on the catalytic performance of the different Au/TS-1-cS catalysts, the use of different S-1 seeds may only change the number of a dominant R group and not the type of R group.

《Fig. 5》

Fig. 5. (a) Diagram of the possible reaction path. (b) Ti–OOH intermediate intensity versus temperature of the Au/TS-1-cS-0.6M, Au/TS-1-cS-1.2M, and Au/TS-1-cS-1.8M catalysts via operando UV–vis. (c–e) Operando UV–vis spectra of (c) Au/TS-1-cS-0.6M, (d) Au/TS-1-cS-1.2M, and (e) Au/TS-1-cS-1.8M catalysts at different reaction temperatures.

In order to confirm that coordinatively unsaturated Ti³⁺ sites can anchor the Au precursor to form Au–O–Ti³⁺ bifunctional active sites, XPS was employed to demonstrate the interaction between Au and Ti. Fig. 6 shows the XPS spectra for the Ti 2p and Au 4f of fresh Au/TS-1-cS-0.6M, Au/TS-1-cS-1.2M, and Au/TS-1-cS-1.8M catalysts. The detailed fitting analysis in Figs. 6(a)–(c) show that, for each Ti valence state, there are two peaks separated by about 5.7 eV, corresponding to the 2p_1/2 (higher BE) and 2p_3/2 (lower BE) spin orbit states. Based on previous studies, the high-BE component of Ti 2p_3/2 at 460.03 eV and of Ti 2p_1/2 at 465.73 eV are assigned to framework tetracoordinated Ti⁴⁺ species, while the low-BE component of Ti 2p_3/2 at 458.02 eV and of Ti 2p_1/2at 463.72 eV are assigned to framework and/or extra-framework Ti³⁺ species [25,51,68]. It was found that the content of the Ti³⁺species exhibited a regular increasing trend from fresh Au/TS-1- cS-0.6M to fresh Au/TS-1-cS-1.8M, which is consistent with the UV–vis, ³¹P MAS NMR, and NH₃-TPD characterization results given in Section 3.1. Moreover, with an increase in the Ti³⁺ content in the catalysts, the BEs of the Ti 2p_3/2 and Ti 2p_1/2peaks in the catalysts slightly downshift, demonstrating that the charge of the Ti ions in the catalysts becomes more negative with an increase in Ti³⁺ content.

The specific fitting analysis indicated that, for each Au valence state, there are two peaks separated by about 3.7 eV, corresponding to 4f_5/2(higher BE) and 4f_7/2 (lower BE) spin orbit states (with an intensity ratio of 3/4). According to the literature, the low-BEcomponent peaks (for Au 4f_7/2) at 83.55, 84.72, and 86.11 eV are assigned to the Au⁰ , Au⁺ , and Au³⁺ species, respectively, while the high-BE-component peaks (for Au 4f_5/2) at 87.25, 88.42, and 89.81 eV are assigned to the Au⁰ , Au⁺ and Au³⁺ species, respectively [69–71]. As shown in Figs. 6(d)–(f), with an increase in the Ti³⁺ content in the catalysts, the binding energies of the peaks for the Au 4f in the catalysts upshift slightly. As this is combined with a downshift in the Ti 2p BE, it can be speculated that electrons have been transferred between Au and Ti, indicating that the introduction of coordinatively unsaturated Ti³⁺ sites enhances the immobilization of Au clusters on the TS-1-cS support. This enhancement could be conducive to activating O₂ and H₂ to generate more hydroperoxide active intermediate, which probably promotes the catalytic performance. Furthermore, the content of the Au⁺ and Au³⁺ species exhibits a regular increasing trend with an increase in the coordinatively unsaturated Ti³⁺ content in the catalysts. This could be because the coordinatively unsaturated Ti³⁺ sites are more likely to anchor the Au complexes ([AuCl_x(OH)_4–x] ^–) [72]. However, the initial valence state of the Au species has little effect on the catalytic performance, because the catalyst is reduced before the reaction.

《Fig. 6》

Fig. 6. (a–c) Ti 2p and (d–f) Au 4f XPS spectra, respectively, of fresh (a, d) Au/TS-1-cS-0.6M, (b, e) Au/TS-1-cS-1.2M, and (c, f) Au/TS-1-cS-1.8M catalysts.

In order to further investigate the effect of Ti species with different coordination forms on Au metal and the effect of Au/Ti bifunctional sites with different forms on the direct gas-phase propene epoxidation reaction, a DFT study was performed. According to the above characterization results, TS-1-Ti⁴⁺ and TS-1-Ti³⁺ were constructed to represent perfect tetracoordinated Ti species (Ti⁴⁺) and coordinatively unsaturated hexacoordinated Ti species (Ti³⁺) structures, respectively (Figs. 7(a) and (b)). The Au₄ clusters were further supported on TS-1-Ti⁴⁺ and TS-1-Ti³⁺ to obtain Au₄/TS-1-Ti⁴⁺ and Au₄/TS-1-Ti³⁺ models (corresponding to the Au/TS-1-cS-0.6M and Au/TS-1-cS-1.8M catalysts). Fig. 7(c) shows that the Au₄/TS-1-Ti³⁺ model (–153.1 kcal∙mol^–1 , 1 kcal = 4185.85 J) exhibits a much higher BE of the Au₄ cluster than the Au₄/TS-1- Ti⁴⁺ model (–44.0 kcal∙mol^–1 ), indicating that the introduction of coordinatively unsaturated Ti³⁺ sites enhances the immobilization of Au clusters on the TS-1 support. Moreover, Table 1 shows that the Au atoms in the Au₄/TS-1-Ti³⁺ model have stronger electropositivity than those in the Au₄/TS-1-Ti⁴⁺ model, which is consistent with the XPS result. This finding suggests that the coordinatively unsaturated Ti³⁺ sites promote electron transfer from the Au to Ti species, leading to stronger interaction between the Au clusters and the support.

Subsequently, the adsorption of O₂ over the two models was investigated to reveal the O₂ adsorption ability. Fig. 7(d) clearly shows that a higher O₂ adsorption energy is obtained with the Au₄/TS-1-Ti³⁺ model (18.1 kcal∙mol^–1 ). Furthermore, the positive charge of the O atom in the Au₄/TS-1-Ti³⁺ model is significantly higher than that in the Au₄/TS-1-Ti⁴⁺ model. This result suggests that the back-donating interaction between Au and O₂ is greatly enhanced due to the presence of coordinatively unsaturated Ti³⁺sites. It has been demonstrated that enhancing the O₂ adsorption ability promotes the formation of O₂^–, which plays a critical role in boosting the catalytic performance in propene epoxidation. In essence, the formation of coordinatively unsaturated Ti³⁺ sites enhances the interaction between Au and the support and induces the formation of partial positively charged Au species. As a result, the adsorption of O₂ is significantly improved, granting the Au/TS-1-cS-1.8M catalysts an excellent low-temperature catalytic performance. To further confirm the results from DFT, O₂ TPD characterization was performed, as shown in Fig. S6 in Appendix A. The oxygen species can generally be identified as surface oxygen species (< 350 °C), active lattice oxygen species near the surface (350–550 °C), and lattice oxygen species (> 550 °C). As shown in Fig. S6, the desorption capacity of O₂ gradually increases as the number of coordinatively unsaturated Ti³⁺ sites gradually increases, indicating that the increase in Au–O–Ti³⁺ content is indeed conducive to the adsorption of O₂.

In summary, coordinatively unsaturated Ti³⁺ sites are conducive to anchoring Au ions and forming Au–O–Ti³⁺ bifunctional active sites. These new sites enhance the generation of H₂O₂ (i.e., the key step in the reaction) by boosting the O₂ adsorption ability, which is the main reason for the resulting high H₂ efficiency. With an increase in H₂O₂ concentration, the active Ti–OOH intermediate formation rate is increased, which effectively enhances the lowtemperature propene epoxidation performance. The insights revealed here shed new light on the design of high-efficiency catalysts for direct low-temperature propene epoxidation.

《Fig. 7》

Fig. 7. DFT calculation models of (a) TS-1-Ti⁴⁺ (perfect 4-coordination Ti) and (b) TS-1-Ti³⁺ (6-coordination Ti³⁺–OH). (c) Binding energy of Au clusters and (d) adsorption energy of O₂over Au₄/TS-1-Ti⁴⁺ and Au₄/TS-1-Ti³⁺ models.

《Table 1 》

Table 1 Mulliken charge (|e|) distribution before and after O₂ adsorption over Au₄/TS-1-Ti⁴⁺ and Au₄/TS-1-Ti³⁺ models.

《4. Conclusions》

4. Conclusions

In this work, we formulated a novel strategy to quantitively construct coordinatively unsaturated Ti³⁺ sites on a novel Au/Tibased catalyst for propene epoxidation with H₂ and O₂ with a Ti³⁺/(Ti³⁺ + Ti⁴⁺) proportion ranging from 13.3% to 22.8% by tuning the amount of Si–OH and Bu₃NH⁺ in post-treated S-1 seeds. Based on operando UV–vis, XPS, and DFT calculations, it was found that the presence of coordinatively unsaturated Ti³⁺ sites promoted electron transfer between Au and Ti, thereby facilitating the formation of Au–O–Ti³⁺ active sites. Compared with traditional Au–O–Ti⁴⁺ sites, these Au–O–Ti³⁺ active sites were found to be more conducive to O₂ adsorption; thus, they effectively promote the in situ generation of H₂O₂ and the formation of the active intermediate Ti–OOH. This enables the Au/TS-1-cS catalyst to exhibit excellent hydrogen efficiency (43.6%) and stability than that of traditional Au/Ti-based catalysts at a lower reaction temperature. The insights and strategies reported here open up new opportunities for optimizing H₂ efficiency via a coordinatively unsaturated Ti³⁺ structure on TS-1 and reveal the intrinsic structure–performance relationship of Au/Ti bifunctional catalysts for low-temperature propene epoxidation.

《Acknowledgments》

Acknowledgments

This work was supported by the National Natural Science Foundation of China (21978325 and 22122807), Outstanding Youth Fund of the National Natural Science Foundation of China (22122807), Outstanding Youth Fund of Shandong Provincial Natural Science Foundation (ZR2020YQ17), and Natural Science Foundation of Shandong Province (ZR2020KB006).

《Compliance with ethics guidelines》

Compliance with ethics guidelines

Zhaoning Song, Hao Yan, Juncong Yuan, Hongfei Ma, Jianlin Cao, Yongxiang Wang, Qiang Wang, Chong Peng, Feng Deng, Xiang Feng, De Chen, Chaohe Yang, and Yongkang Hu declare that they have no conflict of interest or financial conflicts to disclose.

《Appendix A. Supplementary data》

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.eng.2023.01.008.

Show More