Insights into the catalytic valorization of industrial high-concentration nitrous oxide for propylene synthesis

Abstract

Industrial nitrous oxide (N₂O) emissions at high concentrations pose a significant challenge to climate change, while the oxidative dehydrogenation of propane (ODHP) with N₂O presents a promising strategy for N₂O valorization. However, the activity-selectivity trade-off and rapid deactivation by coke deposition present challenges for future applications. Herein, we select Pd/TiO₂ as a model catalyst and design spatially separated redox centers on the rutile phase catalyst (Pd/R-TiO₂), which enable parallel adsorption and independent activation of C₃H₈ on metallic Pd sites and N₂O on oxygen vacancies. Furthermore, the catalyst suppresses product overoxidation by reducing reactive oxygen species concentration near C₃ intermediates, enhancing C₃H₆ selectivity while inhibiting enolic intermediate and coke formation to ensure catalyst stability. Hydrogen spillover from Pd sites to the TiO₂ surface connects the separated redox reactions, completing the catalytic cycle. This work presents a rational strategy for catalyst design in N₂O-ODHP, contributing to the sustainable mitigation of non-CO₂ greenhouse gases.

---

1. Introduction

For the first time, global warming exceeded 1.5 °C across an entire year above pre-industrial levels in 2024, emphasizing the urgent necessity for immediate and sustained efforts to combat climate change.^1,2 Beyond CO₂, attention should be given to non-CO₂ greenhouse gases, including CH₄, N₂O, and fluorinated gases (F-gases). Among these, N₂O is notably the third largest greenhouse gas for climate change, with a global warming potential 273 times higher than that of CO₂.³ Human activities have accelerated N₂O emissions, with an estimated increase rate of 2% per decade and a lifetime exceeding 110 years.^4,5 Worse still, it is the single greatest stratospheric ozone-depleting substance of the 21st century, posing a severe environmental threat.⁶ Nevertheless, N₂O has received less attention compared to CO₂ and CH₄. To achieve the goal of carbon neutrality, efficiently mitigating N₂O emissions from both low- and high-concentration sources is crucial but challenging. N₂O direct decomposition, selective catalytic reduction (SCR), and N₂O revalorization as an oxidant are three major technologies for mitigating N₂O emissions.^7,8 For low-concentration scenarios, the first two technologies can achieve efficient abatement of N₂O, but for high-concentration scenarios, where N₂O can be considered for waste recovery, direct decomposition or SCR would undoubtedly result in resource waste. Specifically, industrial N₂O emissions are expected to reach 1.4 Mt N₂O–N per year by 2050,⁵ a challenge that will escalate in the coming decades. For example, N₂O generated at high concentrations and in large quantities as a by-product in the production of adipic acid (30–40 vol%) or caprolactam (10–20 vol%) can serve as a low-cost and ideal N₂O source.^9,10 Valorizing N₂O offers significant advantages for mitigating the environmental impact of its industrial emissions.¹¹

On the other hand, the global demand for propylene is experiencing rapid growth, leading to a “propylene gap”,¹² which traditional methods, such as steam cracking and fluid catalytic cracking, are struggling to fill.^13–16 Direct propane dehydrogenation (PDH) has been studied and industrialized to address this issue,^17,18 driven by a sudden increase in propane supply from the shale gas revolution.¹⁹ The harsh conditions required (typically ≥650 °C) due to its endothermic nature tend to cause excessive energy consumption. Specifically, it is estimated that the potential energy savings for moving to oxidative dehydrogenation of propane (ODHP) would be ∼45% of the energy consumption.¹³ Compared to the common oxidant O₂, which tends to cause overoxidation, harnessing N₂O as a milder oxidant can improve the selectivity of the reaction. Therefore, ODHP using N₂O as a mild oxidant (N₂O-ODHP: C₃H₈ + N₂O → C₃H₆ + N₂ + H₂O) represents an emerging candidate technology for propylene production and N₂O valorization.^20–22 However, inadequate atomic-level understanding of the underlying reaction mechanisms has impeded the development of highly efficient catalysts in N₂O-ODHP. Significant challenges remain in breaking the trade-off between activity and selectivity while addressing the tendency for rapid deactivation caused by coke deposition, with current strategies primarily relying on trial-and-error methods, resulting in high costs and limited success rates.²³

Herein, a series of Pd/TiO₂ catalysts were experimentally and theoretically studied as model catalysts in N₂O-ODHP. Anatase and rutile titanium oxide supports are often chosen for mechanistic studies due to their well-understood properties and versatility.^24–26 Palladium (Pd) was selected as the primary active metal because of its well-documented C–H cleavage ability and use in propane dehydrogenation.^27–29 The results demonstrated that the Pd/R-TiO₂ catalyst significantly boosted propylene production, although it had the same chemical composition as others. This begs critical questions: What is the structural feature of Pd/R-TiO₂? Does the performance stem from the presence of redox center separation? Our findings highlight that redox center separation can break the activity-selectivity trade-off while ensuring high catalyst stability for N₂O-ODHP. This study paves the way for the rational design of efficient catalysts, advancing future application of N₂O-ODHP for sustainable N₂O valorization and propylene production.

2. Methods

2.1. Chemicals and materials

All chemicals were purchased from Aladdin Chemical Reagent Company, while all gases utilized in this study were sourced from Hangzhou Jingong Special Gases Co., Ltd. All chemicals used in the experiments were reagent grade or higher and used as received without further purification. Experimental solutions were prepared using deionized (DI) water.

2.2. Synthesis of catalysts

Pd/A-TiO₂ and Pd/R-TiO₂ catalysts were prepared via sol immobilization.³⁰ Only the TiO₂ supports were varied, including anatase and rutile TiO₂. Aqueous solutions of the metal precursors, Pd(NO₃)₂, were added to deionized water (140 mL) and stirred vigorously. PVA was then added to the solution, with a mass ratio of PVA to Pd of 1 : 1. Subsequently, a freshly prepared NaBH₄ solution (0.15 M, with a NaBH₄ to total metal ratio of 4 : 1, mole per mole) was added to the solution immediately, resulting in the formation of a sol. Following this, the support material (0.5 g) was added to the colloidal solution while stirring, facilitating the immobilization of the metal nanoparticles. After 30 min, the solid catalyst was recovered by filtration and washed with 500 mL of distilled water to remove Na⁺, NO₃⁻, BH₄⁻, and BO₂⁻ ions. Subsequently, the catalysts were dried overnight under vacuum at 80 °C. Finally, the as-obtained powders were calcined at 550 °C for 3 h in flowing Ar to prepare Pd/A-TiO₂ and Pd/R-TiO₂.

PdO/A-TiO₂ and PdO/R-TiO₂ catalysts were prepared using the impregnation method, with only the TiO₂ supports being varied. 1.0 g of TiO₂ support and a defined amount of Pd(NO₃)₂ were mixed in 140 mL of H₂O under sonication for 30 min and stirring for 4 h. The catalysts were then dried overnight. The obtained powders were calcined at 550 °C for 3 h in static air to prepare PdO/R-TiO₂ and PdO/A-TiO₂ catalysts.

2.3. Characterizations

The actual metal loadings of the catalysts were determined using inductively coupled plasma-optical emission spectroscopy (ICP-OES) with a Thermo Fisher iCAP PRO instrument. Transmission electron microscopy (TEM) was conducted using a Tecnai G2 F20 S-TWIN instrument from FEI Company, USA, operating at an accelerating voltage of 200 kV. Powder X-ray diffraction (XRD) patterns were recorded using an X'Pert3 Powder analytical diffractometer (B.V., Netherlands) at a scan rate of 5 °C min⁻¹ with monochromatized Cu Kα radiation. Ultraviolet-visible diffuse reflectance spectra (UV-vis DRS) were obtained using a UV-2600 Shimadzu spectrophotometer equipped with an integrating sphere assembly at wavelengths ranging from 200 to 800 nm, with a resolution of 0.5 nm. BaSO₄ was used as the reflectance standard under ambient conditions. X-ray photoelectron spectroscopy (XPS) experiments were carried out on an ESCALAB 250 (Thermo, USA) using standard Al Kα X-ray radiation. All spectra were calibrated using the C 1s peak at 284.8 eV. Raman spectra were recorded on a HORIBA Scientific LabRAM HR Evolution spectrometer using a He/Cd excitation laser at 532 nm. Electron paramagnetic resonance (EPR) spectra were obtained using a Bruker EMXplus spectrometer at liquid nitrogen temperature (77 K).

Operando NAP-XPS spectra were recorded using a Thermo Scientific ESCALAB 250Xi instrument. Monochromatized Al Kα irradiation (1486.6 eV) with a pass energy of 20.0 eV was employed to collect all spectra. A pretreatment chamber was utilized to expose the catalysts to specific treatments. After treatment, the chamber was evacuated, and the sample was directly transferred to the vacuum analysis chamber without exposure to air.

H₂-TPR experiments were conducted with a TP-5089 (Tianjin Xianquan Industry and Trade Development Co., China). The samples were pretreated prior to measurements at 200 °C for 1 h with pure He and then cooled to 50 °C. The baselines were allowed to sufficiently stabilize in 5% H₂/Ar flow at 50 °C before commencing temperature programming, followed by heating to 800 °C in 5% H₂/Ar flow at a heating rate of 5 °C min⁻¹. The consumption of H₂ was monitored using a thermal conductivity detector.

In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) experiments were conducted in the range of 4000–400 cm⁻¹ with a spectral resolution of 4 cm⁻¹ using a Bruker Tensor 27 spectrometer equipped with ZnSe windows. The spectrometer also featured a high-sensitivity MCT detector cooled by liquid nitrogen and a high-temperature reaction chamber under ambient pressure. All spectra were obtained by averaging 64 scans with a resolution of 4 cm⁻¹. Before each experiment, the sample was pretreated in He at 400 °C for 30 min to degas (i.e. gaseous H₂O and hydroxy species), followed by the collection of a background spectrum.

Thermogravimetric analysis (TGA) and derivative thermogravimetry (DTG) experiments were conducted with a NETZSCH STA 449F3.³¹ TGA and DTG measured the amount and rate of change in the weight of a material as a function of temperature or time in a controlled atmosphere. Here, the sample was heated from 30 to 600 °C in a platinum crucible with a heating rate of 10 °C min⁻¹ in flowing air.

2.4. Catalytic evaluations

The N₂O-ODHP reactions were conducted in a fixed-bed reactor constructed from stainless steel. In a typical experiment, 500 mg of sieved sample (40–60 mesh) was loaded between two layers of quartz wool, and the designed weight hourly space velocity (WHSV) was set at 12 000 mL g_cat⁻¹·h⁻¹. The feed gas consisted of 16.7% C₃H₈, 16.7% N₂O, balanced with He; 2% Ar in the flow was used as the inner standard. The quantitative N₂O analysis of the feed and effluent gas was measured with an FTIR spectrometer (G210 analyser, 0.1–100%), purchased from Geotechnical Instruments (UK) Ltd. The products of the reaction were analyzed with an online gas chromatograph (GC, Agilent 7890A) equipped with a thermal conductivity detector (TCD) using He as the carrier gas with Porapak Q and TDX-01 packed columns and a flame ionization detector (FID) with a HP-PLOTQ capillary column. The probable products of the N₂O-ODHP were C₃H₆, C₂H₆, C₂H₄, CH₄, CO, and CO₂. The carbon balance between the reactant (C₃H₈) and products (i.e., CO, CO₂, CH₄, C₂H₄, C₂H₆, and C₃H₆) reached ≥98%, as determined by GC.

2.5. Computational details

The geometrical optimization and transition state (TS) calculations were performed using the Vienna ab initio simulation package (VASP) code (version 5.4.4).^32,33 Other details, including catalytic evaluations and DFT calculations, are shown in the SI.

3. Results and discussion

3.1. Synthesis and characterizations

Rutile (denoted as R) and anatase (denoted as A) TiO₂ were used as the support. The home-made anatase displayed a uniform distribution of Ti and O atoms, as determined by scanning transmission electron microscopy (STEM) and elemental mapping (Fig. S1). Anatase and rutile had surface areas of 51.7 and 48.0 m² g⁻¹ (Fig. S2), respectively, providing comparable surfaces for 0.5 wt% Pd loadings. PdO/TiO₂ catalysts were prepared for comparison using wet impregnation (Fig. 1a), labeled as PdO/R-TiO₂ (Fig. 1b) and PdO/A-TiO₂ (Fig. 1c). Pd_NP/TiO₂ catalysts were synthesized by the sol immobilization method (Fig. 1d),^30,34 denoted as Pd/R-TiO₂ (Fig. 1e) and Pd/A-TiO₂ (Fig. 1f). X-ray diffraction (XRD) showed comparable diffraction patterns with rutile TiO₂ (JCPDS No. 21-1276) and anatase TiO₂ (JCPDS No. 21-1272) for R-TiO₂ and A-TiO₂, respectively (Fig. S3). No reflections corresponding to crystalline Pd or PdO were observed, likely due to the low content and high dispersion of Pd species on the support. Notably, we chose NaBH₄ instead of H₂ as the reductant³⁰ because chemical reduction at room temperature avoids high-temperature hydrogen-mediated thermal reduction that may cause a phase transformation from anatase to rutile, which would obscure the effects of the TiO₂ crystalline phase on the Pd–TiO₂ synergy. This is supported by the XRD profile of H₂-reduced PdO/A-TiO₂ at 550 °C (Fig. S4), showing a mixture of anatase and rutile crystalline regions.

Fig. 1 (a) Schematic of the impregnation method and (b and c) TEM images of PdO/R-TiO₂ (b) and PdO/A-TiO₂ (c) and (d) schematic of the sol immobilization method and (e and f) TEM images of Pd/R-TiO₂ (e) and Pd/A-TiO₂ (f); (g) XPS spectra of Pd 3d for all the Pd/TiO₂ catalysts; (h) UV-vis DRS plots of TiO₂ and Pd/TiO₂ catalysts; (i) CO-DRIFTS spectra for all the Pd/TiO₂ catalysts. Pd, mazarine; Ti, grey; O, red.

Transmission electron microscopy (TEM) images showed the exposed facets of TiO₂ were anatase (101) and rutile (110), with lattice distances of 0.35 nm and 0.33 nm, respectively. The Pd species on both PdO/R-TiO₂ and PdO/A-TiO₂ were PdO, exhibiting an exposed facet of PdO(002) with a lattice distance of 0.27 nm. For Pd/A-TiO₂ and Pd/R-TiO₂, metallic Pd NPs with exposed Pd(111) surfaces with a lattice distance of 0.23 nm were observed, all with a similar size of ∼10 nm, as confirmed by line scanning (Fig. S5). Additionally, elemental mapping on Pd/R-TiO₂ is also presented (Fig. S6), excluding the presence of TiO_x overlayers due to strong metal-support interaction.^18,35

In Pd 3d X-ray photoelectron spectroscopy (XPS, Fig. 1g), peaks at 334.4 and 339.7 eV, and at 335.2 and 340.5 eV, can be assigned to Pd⁰ and Pd²⁺ states, respectively.³⁶ The results suggest that the Pd species on Pd/A-TiO₂ and Pd/R-TiO₂ predominantly exist as metallic Pd⁰ species, while the Pd species on both PdO/R-TiO₂ and PdO/A-TiO₂ are mainly in the oxidized Pd²⁺ form. Ultraviolet-visible diffuse reflectance spectra (UV-vis DRS, Fig. 1h) showed an absorption in the range of 200–400 nm that was associated with anatase and rutile TiO₂, while a broader absorption peak at 450–550 nm was attributed to a d–d transition of PdO species for PdO/TiO₂.³⁷ Moreover, a significant absorption in the visible region above 600 nm was observed, which was attributed to a characteristic signal for Pd NPs for Pd_NP/TiO₂.³⁸ CO-adsorption diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) showed a broad band at 1900–2000 cm⁻¹, attributed to Pd NPs,³⁹ for Pd_NP/TiO₂ (Fig. 1i). A Pd²⁺–CO band at 2185 cm⁻¹ was also detected for PdO/TiO₂. This suggested that the Pd species in Pd_NP/TiO₂ were mainly Pd NPs, while those in PdO/TiO₂ were primarily PdO species.

Raman spectra were also obtained (Fig. S7). All characteristic Raman shifts of anatase TiO₂ were observed at 144 cm⁻¹ (E_g), 198 cm⁻¹ (E_g), 398 cm⁻¹ (B_1g), 520 cm⁻¹ (A_1g), and 638 cm⁻¹ (E_g).²⁵ The shifts to higher wavenumbers of the E_g peaks indicated the formation of the strongest Pd–TiO₂ interactions and the most abundant oxygen vacancies. By contrast, four typical vibrational modes were observed for rutile TiO₂ around 145 cm⁻¹ (B_1g), 445 cm⁻¹ (E_g), 610 cm⁻¹ (A_1g), and 240 cm⁻¹ for second-order effects.⁴⁰ The redshifts to lower wavenumbers of the A_1g and E_g vibration modes for Pd/R-TiO₂ indicated the weakened bond strength of O–Ti–O. This implied facile formation of oxygen vacancies in TiO₂, which could significantly facilitate the activation of N₂O.

3.2. Catalytic performance and the identification of key intermediates

N₂O-ODHP catalytic performance was evaluated in a fixed-bed reactor under a WHSV of 12 000 mL g_cat⁻¹ h⁻¹, at a C₃H₈/N₂O/He ratio of 1/1/4, where 16.7% N₂O concentration is in the range of the tail gas emitted from caprolactam production (10–20 vol%).¹⁰ As shown in Fig. 2a–d, as the reaction temperature increased, both C₃H₈ and N₂O conversions improved, but the selectivity for C₃H₆ declined. Furthermore, an increase in the CO ratio and the production of other C₂ and C₁ byproducts were observed. Therefore, an optimal reaction temperature was determined as 500 °C. Pd/R-TiO₂ exhibited the highest C₃H₆ selectivity (67.0%) and C₃H₆ space-time yield (STY) of 74.2 kg_C3H6 kg_Pd⁻¹ h⁻¹ with a C₃H₈ conversion of 14.2%. The catalytic performance of 9.5% yield outperforms those of previously reported benchmark catalysts (Table S1).^20,41–43 PdO/R-TiO₂ showed high C₃H₆ selectivity (33.1%) and STY (45.7 kg_C3H6 kg_Pd⁻¹ h⁻¹) at higher C₃H₈ conversion (18.9%). In contrast, PdO/A-TiO₂ demonstrated a high conversion (53.9%) but low C₃H₈ conversion (7.0%) and C₃H₆ selectivity (22.9%), resulting in a low STY (11.3 kg_C3H6 kg_Pd⁻¹ h⁻¹). Pd/A-TiO₂, while achieving the lowest N₂O conversion (17.4%), showed slightly better C₃H₆ selectivity (24.5%) and STY (19.3 kg_C3H6 kg_Pd⁻¹ h⁻¹). Overall, the C₃H₆ STY follows the order Pd/R-TiO₂ > PdO/R-TiO₂ > Pd/A-TiO₂ > PdO/A-TiO₂. Additionally, the O₂-ODHP performance of Pd/R-TiO₂ was examined (Fig. S8), showing C₃H₆ selectivities of less than 10% at different temperatures and a C₃H₈/O₂ ratio of 1/1. The results showed lower C₃H₆ selectivity and overoxidation when using O₂ as the oxidant, suggesting the significant role of N₂O in increasing the selectivity.

Fig. 2 (a–d) Catalytic performance of PdO/A-TiO₂ (a), Pd/A-TiO₂ (b), PdO/R-TiO₂ (c), and Pd/R-TiO₂ in N₂O-ODHP at 400–550 °C (d). Histogram showing the selectivity for different products; hollow circle, C₃H₈ conversion; solid circle, N₂O conversion; hollow green star, C₃H₆ STY. (e and f) Catalytic stability of PdO/R-TiO₂ (e) and Pd/R-TiO₂ (f) for N₂O-ODHP at 500 °C. Typical reaction conditions: 100 mL min⁻¹ reaction feed of 16.7% C₃H₈, 16.7% N₂O, He balance, Ar (2%) as the inner standard; 500 mg of catalyst; 12 000 mL g_cat⁻¹ h⁻¹. (g and h) In situ DRIFT spectra of adsorbed species on the PdO/R-TiO₂ (g) and Pd/R-TiO₂ (h) catalysts as a function of time at 500 °C under a flow of 16.7% C₃H₈, 16.7% N₂O, and He balance.

To assess potential structural or compositional changes during the N₂O-ODHP reaction, the catalysts were subjected to the standard reaction conditions (500 °C, 100 mL min⁻¹ feed of 16.7% C₃H₈, 16.7% N₂O, balance He) for 6 h. After cooling to room temperature, the system was restarted under the same conditions,⁸ and the catalytic performance of the used catalysts was re-evaluated (Table S2). Compared to their fresh performance, the used PdO/A-TiO₂ exhibited a slight decrease in N₂O conversion (from 53.9% to 43.7%) and C₃H₈ conversion (from 7.0% to 6.7%), along with a drop in C₃H₆ selectivity (from 22.9% to 18.6%), resulting in a reduced propylene yield from 1.6% to 1.2%. Similarly, PdO/R-TiO₂ showed a more significant decline in overall performance after the reaction, with C₃H₈ conversion decreasing from 18.9% to 16.0%, C₃H₆ selectivity from 33.1% to 21.7%, and propylene yield from 6.3% to 3.5%. In contrast, for Pd/A-TiO₂, the catalytic performance remained largely stable, with a minor decrease in C₃H₈ conversion (from 9.9% to 9.6%) and a slight improvement in selectivity (from 24.5% to 27.1%), leading to a comparable propylene yield (2.4% fresh vs. 2.6% used). Notably, Pd/R-TiO₂ maintained good catalytic stability, with nearly unchanged C₃H₈ conversion (14.2% fresh vs. 13.7% used), selectivity (67.0% vs. 67.2%), and C₃H₆ yield (9.5% vs. 9.2%), confirming its robustness and potential applicability under practical conditions.

We then recorded the time course of the STY of C₃H₈ for PdO/R-TiO₂ and Pd/R-TiO₂ to test the long-term continuous stability at 500 °C (Fig. 2e and f). Rapid deactivation occurred within 8 h, resulting in a decrease of C₃H₆ STY from 45.7 to 20.7 kg_C3H6 kg_Pd⁻¹ h⁻¹. After 30 min of regeneration in N₂O flow, the activity recovered to 40.7 kg_C3H6 kg_Pd⁻¹ h⁻¹, lower than the initial value of 45.7 kg_C3H6 kg_Pd⁻¹ h⁻¹. This suggested that the deactivation was primarily caused by coke formation. N₂O flow effectively removed these coke species, thereby restoring the catalytic activity. However, rapid deactivation was observed in the second reaction cycle, with the STY dropping to 21.1 kg_C3H6 kg_Pd⁻¹ h⁻¹ within 5 h, and an even faster deactivation occurred in the third reaction cycle within only 2 h. This indicated poor stability for PdO/R-TiO₂. Furthermore, the activity of PdO/R-TiO₂ decreased after regeneration, which might be attributed to either irreversible sintering or the accumulation of coke.¹² To verify this, we conducted TEM analysis on the PdO/R-TiO₂ sample after the reaction and regeneration (Fig. S9). The results showed the presence of apparent coke deposits surrounding the PdO species, suggesting that the decline in activity during the second cycle is mainly due to incomplete coke removal rather than sintering. Moreover, no significant increase in particle size was observed, further suggesting that sintering is not the dominant cause.

In contrast, Pd/R-TiO₂ retained a high C₃H₆ STY (∼70 kg_C3H6 kg_Pd⁻¹ h⁻¹) during the stability test, showing a reliable durability compared to that of PdO/R-TiO₂. Thermogravimetric analysis (TGA) and derivative thermogravimetry (DTG) experiments showed PdO/R-TiO₂ exhibited approximately 1.3 wt% weight loss (Fig. S10), which could be attributed to both water desorption (<100 °C) and coke formation (200–350 °C).⁴⁴ In contrast, Pd/R-TiO₂ showed almost no weight loss without coke formation under working conditions.

To evaluate the long-term durability, we extended the Pd/R-TiO₂ catalytic test to 24 h (Fig. S11). A decrease in C₃H₆ STY was observed after 14 h, likely due to coke deposition. After a 1 hour regeneration in N₂O flow, the catalyst's performance was fully restored and maintained for another 10 h. These findings confirm the catalyst's regenerability and demonstrate potential for industrial application. Nevertheless, the observed deactivation highlights the demand to further optimize catalyst design based on the redox center separation strategy in future catalyst development.

In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) was conducted under C₃H₈/N₂O co-feed conditions with a ratio of 1/1 at 500 °C on PdO/R-TiO₂ and Pd/R-TiO₂ in Fig. 2g and h. Bands attributed to C₃H₈ (2980, 2967, 2959, 2901, 2887, 2796 cm⁻¹),⁴⁵ and N₂O (2235, 2213, 1299, 1271 cm⁻¹),⁴⁶ were all observed for both samples. Additionally, peaks attributed to both C₃H₆ (1516, 1439 cm⁻¹) and CO₂ (2359 cm⁻¹) were observed, consistent with the products detected in the catalytic evaluation. However, the peaks of acrylate species (1644, 1301 cm⁻¹), acetates (1565, 1460 cm⁻¹), and formate (1592, 1376 cm⁻¹),⁴⁷ were not observed, likely ruling out the formation of these potential intermediates. By contrast, a notable peak at 1620 cm⁻¹ appeared over PdO/R-TiO₂ (Fig. 2g), which was assigned to the characteristic peak of enolic species.^48,49 This indicated C₃H₈ was partially oxidized to enolic species, leading to low C₃H₆ selectivity. In contrast, no enolic species signal was observed on Pd/R-TiO₂ (Fig. 2h), indicating that Pd NPs inhibited N₂O-ODHP from proceeding through the enolic path. Moreover, the oxygen-containing enolic intermediate was likely to serve as a precursor for coke formation by polymerization,⁵⁰ causing poor stability.

3.3. Identifying the geometric and electronic structures of Pd/TiO₂

Pd/TiO₂ was selected as a model catalyst for density functional theory (DFT) calculations for N₂O-ODHP (Fig. 3a). The charge density of PdO on PdO/TiO₂ showed positive values near TiO₂ and negative values near PdO, indicating electron transfer from PdO to TiO₂ (Fig. S12). For Pd NPs on Pd_NP/TiO₂, the charge transfer direction was reversed, with electron transfer occurring from TiO₂ to Pd NPs (Fig. 3b); that is, a reversed charge transfer.⁵¹ This might be one of the reasons for the improved performance on Pd/R-TiO₂.

Fig. 3 (a) Optimized atomic structures of different Pd/TiO₂ models for PdO/A-TiO₂, Pd/A-TiO₂, PdO/R-TiO₂, and Pd/R-TiO₂. (b) Calculated charge density difference and planar-averaged charge density difference Δρ(z) of Pd/R-TiO₂. Cyan indicates holes, and yellow indicates electrons. (c) Energy diagrams of O_v formation on different sites on Pd/TiO₂. (d) Possible configurations for N₂O adsorption on the Pd/R-TiO₂ model. Pd, mazarine; Ti, grey; O, red; N, blue.

N₂O reduction could readily occur at O_v sites. Hence, O_v formation energy was calculated (Fig. 3c). Notably, the O_v formation on PdO on PdO/A-TiO₂ was even exothermic (−1.61 eV), indicating that PdO/A-TiO₂ would exhibit highly reactive O_v sites and, consequently, the highest N₂O conversion. O_v formation on PdO on PdO/R-TiO₂ was only 0.34 eV, smaller than those on the TiO₂ surface, suggesting that the step O_v + N₂O → * + N₂ on PdO occurred more readily than on the TiO₂ surface, whether anatase or rutile. As evident from the DFT results, the most efficient O_v formation on the TiO₂ surface was observed with Pd/R-TiO₂, with the lowest energy of 0.83 eV. The order of O_v formation on the TiO₂ surface was Pd/R-TiO₂ > Pd/A-TiO₂ (1.11 eV) > PdO/A-TiO₂ (2.66 eV) > PdO/R-TiO₂ (3.27 eV). However, given the facile O_v formation on PdO, the overall order of O_v formation tendency was PdO/A-TiO₂ > PdO/R-TiO₂ > Pd/R-TiO₂ > Pd/A-TiO₂, in good agreement with XPS O 1s results.

To assess the ability of Pd NPs for N₂O activation, various initial configurations of N₂O on Pd NPs were examined,⁵² including side-on and end-on modes (Fig. 3d). The N₂O molecule energetically preferred a side-on coordination with an adsorption energy of −0.81 eV, which was relatively weak for facilitating further N–O bond cleavage. As a result, for PdO/TiO₂, both C₃H₈ and N₂O activation occurred on the PdO species. In contrast, for Pd_NP/TiO₂, C₃H₈ was dissociated on the Pd NPs and N₂O reduction occurred on the O_v sites on TiO₂ with the recovery of O_v, achieving the separation of redox centers.

To further support the DFT results regarding reactive oxygen species, EPR spectra were recorded for fresh and used Pd/R-TiO₂ samples (Fig. 4a). Evidently, a greater number of O_v sites were formed during the N₂O-ODHP with more Ti³⁺ species.⁵³ In the deconvoluted O 1s spectra, lattice oxygen (O_lat) appears at ∼530.0 eV and surface-adsorbed oxygen species (O_ads) at ∼531.7 eV.⁵⁴ For fresh samples (Fig. 4b), the O_ads ratios followed the order PdO/A-TiO₂ (17.1%) > PdO/R-TiO₂ (16.2%) > Pd/R-TiO₂ (15.8%) > Pd/A-TiO₂ (11.3%). The O_ads species are considered reactive intermediates that can readily participate in surface reactions, leading to the formation of O_v sites.⁵⁵ Thus, the O_ads ratio serves as an indicator of the oxygen vacancy formation tendency on the TiO₂ surface or PdO species. To further verify the dynamics of oxygen vacancy formation, operando near-ambient-pressure X-ray photoelectron spectroscopy (NAP-XPS) measurements were conducted at 500 °C under a C₃H₈/N₂O (1 : 1) flow. When measured under N₂O-ODHP conditions, the O_ads ratios changed to 9.0%, 14.4%, 10.6%, and 12.9% for PdO/A-TiO₂, Pd/A-TiO₂, PdO/R-TiO₂, and Pd/R-TiO₂, respectively (Fig. 4c). The decrease in O_ads content reflects the formation of oxygen vacancies via H* species generated during C₃H₈ dehydrogenation. The observed O_ads reduction followed the order PdO/A-TiO₂ (−8.1%), PdO/R-TiO₂ (−5.6%), Pd/R-TiO₂ (−2.9%), Pd/A-TiO₂ (+3.1%), consistent with the DFT-predicted trend in vacancy formation energies. Notably, the slight increase in O_ads observed for Pd/A-TiO₂ might result from the generation and accumulation of O* species derived from activated N₂O, which remain strongly adsorbed on the surface.

Fig. 4 (a) EPR spectra of fresh and used Pd/R-TiO₂. (b) XPS spectra of O 1s. NAP-XPS spectra of O 1s (c), Ti 2p (d), and Pd 3d (e) recorded after catalysts were exposed to C₃H₈/N₂O = 1/1 at 500 °C. (f) H₂-TPR plots for PdO/A-TiO₂, Pd/A-TiO₂, PdO/R-TiO₂, Pd/R-TiO₂.

We also recorded the Ti 2p XPS spectra under reaction conditions. Prior to the reaction, all catalysts exhibited a Ti³⁺ component of less than 10% (Fig. S13), indicating a well-preserved crystalline structure of TiO₂, consistent with the XRD results (Fig. S3). After the reaction, the Ti³⁺ proportion significantly increased, following the order Pd/R-TiO₂ (40.5%) > Pd/A-TiO₂ (38.3%) > PdO/A-TiO₂ (35.1%) > PdO/R-TiO₂ (24.2%) (Fig. 4d). This trend aligns well with the DFT-predicted tendency for O_v formation on TiO₂ surfaces.

However, since the formation of O_v requires H* species generated from propane dehydrogenation, it suggests that PdO species may directly trap these H* species via their surface oxygen, thereby suppressing H* migration to the TiO₂ support and inhibiting O_v formation. In contrast, metallic Pd facilitates hydrogen spillover, promoting vacancy generation on the TiO₂ surface. To further validate this hypothesis, we collected NAP-XPS spectra of Pd 3d under reaction conditions (Fig. 4e). For both PdO/TiO₂ catalysts, a portion of Pd⁰ species was observed, indicating partial reduction of PdO during the reaction, consistent with the loss of surface oxygen. In contrast, the Pd/A-TiO₂ and Pd/R-TiO₂ samples maintained a fully metallic Pd⁰ state throughout. Taken together, these findings support the DFT predictions regarding the oxygen vacancy formation tendencies and demonstrate that hydrogen spillover from Pd to the TiO₂ surface plays a crucial role in O_v dynamics during the reaction.

Further, we conducted H₂-temperature programmed reduction (H₂-TPR) to investigate the dynamic behavior of oxygen species and the ability of the catalysts to dissociate H₂ (Fig. 4f). For all samples, a slight but noticeable negative peak was observed below 100 °C, which can be attributed to H₂ adsorption. Among them, Pd/R-TiO₂ exhibited the strongest H₂ adsorption capacity. For both PdO/TiO₂ samples, a broad H₂ consumption peak appeared below 400 °C, corresponding to the reduction of PdO species. Upon further increasing the temperature above 700 °C, an additional peak was observed, which can be attributed to the removal of lattice oxygen from the TiO₂ support. Notably, PdO/A-TiO₂ showed a surprisingly high H₂ consumption above 500 °C, which is likely associated with the anatase-to-rutile phase transformation; a process known to involve substantial hydrogen participation. This observation is consistent with the XRD results (Fig. S4). In contrast, although Pd/A-TiO₂ also exhibited a high-temperature signal above 700 °C, it was much less pronounced than that of PdO/A-TiO₂. This suggests that the interaction between Pd nanoparticles and anatase TiO₂ may help stabilize the crystal phase against transformation. Moreover, for both Pd/A-TiO₂ and Pd/R-TiO₂, a reduction peak appeared around 450 °C, which can be assigned to H₂ dissociation on metallic Pd NPs. This indicates that at our reaction temperature of 500 °C, the catalysts are fully capable of facilitating H*-involved processes, which are essential for bridging propane dehydrogenation and N₂O activation.⁵⁶

3.4. Theoretical insights into the reaction pathway

We then calculated the reaction pathways (Fig. 5). The initial C–H bond activation in C₃H₈ is a crucial step in the ODHP for all models, where a hydrogen atom from methyl was eliminated and adsorbed onto the catalyst surface. The activation of C₃H₈ showed high energy barriers of 2.65 eV and 2.22 eV for PdO/A-TiO₂ and Pd/A-TiO₂, respectively, which strictly impeded subsequent dehydrogenation steps. In contrast, C₃H₈ activation was facile for both PdO/R-TiO₂ and Pd/R-TiO₂ with low barrier energies of 1.40 eV and 1.21 eV, respectively.

Fig. 5 (a and b) Energy diagrams of PdO/R-TiO₂ (a) and Pd/R-TiO₂ (b) for C₃H₈ dehydrogenation through the propylene and enolic pathways. (c) Summary of the energy barriers for H* spillover on the PdO/R-TiO₂ and Pd/R-TiO₂ models. (d and e) Energy diagrams of PdO/A-TiO₂ (c), Pd/A-TiO₂ (d) for C₃H₈ dehydrogenation through the propylene and enolic pathways. TS, transition states. Pd, mazarine; Ti, grey; O, red; C, dark grey; H, white; H*, green. (f) Plots of the activation energy for the first dehydrogenation step of propane against e^−W.

For PdO/R-TiO₂, as shown in Fig. 5a, the energy barriers for C–C scission and C–H cleavage of C₃H₇* species were determined to be 1.50 eV and 2.15 eV, respectively, facilitating the formation of enolic species via the enolic pathway. Subsequently, the C₂H₅* species underwent dehydrogenation to C₂H₄*, followed by oxidation by oxygen atoms on PdO to form enolic species, as observed by in situ DRIFTS. The formation of enolic species required overcoming an energy barrier of 1.73 eV, suggesting that it was a key step for overoxidation. In contrast, for Pd/R-TiO₂ in Fig. 5b, the energy barriers for C–C scission and C–H cleavage of C₃H₇* species were 2.64 eV and 1.86 eV, respectively. These values indicated a stronger propensity to generate C₃H₆ rather than enolic species, leading to higher selectivity. Besides, the dissociated H* species could directly react with the PdO or TiO₂ surface after migrating from the Pd species. However, the oxygen atom within PdO on PdO/R-TiO₂ strictly hindered H* migration, with a higher barrier of 0.83 eV compared to that of 0.21 eV on Pd/R-TiO₂ (Fig. 5c).

To shed light on the atomic-level origin of the higher propylene selectivity of Pd/R-TiO₂ compared to Pd/A-TiO₂, we performed comparative DFT calculations of the ODHP reaction pathway on PdO/A-TiO₂ and Pd/A-TiO₂, as shown in Fig. 5d and e, respectively. In the oxidative dehydrogenation reactions, product selectivity is generally governed by two key factors: (i) suppression of undesired C–C bond cleavage, and (ii) facile desorption of the target product, propylene.²¹ Our results reveal that anatase TiO₂ does not significantly promote C–H bond activation, as reflected by the high C₃H₈ activation barriers (2.55 eV and 2.22 eV). Moreover, C₃H₆ binds weakly on anatase-supported Pd surfaces (1.26 eV and 0.96 eV), facilitating its desorption. However, despite the weak binding, anatase-based catalysts still exhibit low C₃H₆ selectivity, suggesting that propylene adsorption is not the determining factor for high yield. Instead, the competition between C–H bond scission and C–C bond cleavage of the surface intermediate C₃H₇* is more critical. For both PdO/A-TiO₂ and Pd/A-TiO₂, the C–C bond cleavage pathway shows a lower barrier (3.55 eV vs. 4.05 eV, and 2.74 eV vs. 3.12 eV, respectively) than the pathway toward C₃H₆ formation. Therefore, it is reasonable to propose that suppressing C–C bond cleavage is key to improving propylene selectivity.

To better understand the electronic factors governing catalytic activity, we investigated the relationship between the work function (W) and the activation barrier (E_a) for C₃H₈ activation across four Pd/TiO₂ catalysts. The calculated W values were 4.43, 4.60, 5.26, and 5.41 eV for PdO/A-TiO₂, Pd/A-TiO₂, PdO/R-TiO₂, and Pd/R-TiO₂, respectively, consistent with previous reports.⁵⁷ Catalysts with larger W values are electron-deficient, which enables them to readily accept electrons from the bonding orbital of the C–H bond in C₃H₈, suggesting that the Pd/R-TiO₂ catalyst is particularly active for C₃H₈ activation.^58–60 This trend has been recently reported, showing that e^−W can effectively quantify the ability to activate the C–H bond, surpassing the predictive capability of the d-band center.⁶¹ We also plotted E_a against e^−W, which exhibited a linear relationship (Fig. 5f). This not only supplements existing findings but, together with our data, further confirms that the work function can serve as a reliable descriptor for the development of future dehydrogenation catalysts. The plot shows that as W increases, the energy barrier decreases, indicating that a higher work function facilitates C–H bond dissociation during C₃H₈ activation.

The combined reaction mechanisms for N₂O-ODHP are summarized in Fig. 6. The designed Pd/R-TiO₂, with separated redox centers, enables efficient activation of both C₃H₈ and N₂O while minimizing direct interaction between C₃ intermediates and reactive oxygen species. Hydrogen spillover bridges the oxidation and reduction reactions, thereby completing the N₂O-ODHP redox cycle and effectively disentangling the trade-off between activity and selectivity, leading to boosted propylene production. In contrast, PdO/R-TiO₂ fails to achieve high propylene yield as the redox centers are not spatially separated, causing both C₃H₈ and N₂O to be activated on PdO sites. In contrast, Pd/A-TiO₂ and PdO/A-TiO₂ exhibit limited C₃H₈ conversion due to their high energy barriers resulting from their low work functions.

Fig. 6 Schematic of the reaction mechanisms of Pd/TiO₂ catalysts in N₂O-ODHP.

4. Conclusion

Pd/TiO₂ model catalysts were employed to elucidate the reaction mechanisms for N₂O-ODHP, which is a promising technique for propylene production and N₂O valorization. Comprehensive DFT calculations combined with experimental efforts revealed that the Pd/TiO₂ catalyst with redox center separation can effectively avoid competition and direct contact between N₂O and C₃H₈, where O_v sites on TiO₂ dissociate N₂O and the metallic Pd sites activate C₃H₈, thereby disentangling the trade-off between activity (C₃H₈ conversion) and C₃H₆ selectivity. The separated redox reactions are linked by the hydrogen spillover from the Pd sites to the TiO₂ surface. This work clarifies the atomic-level mechanisms and catalyst design behind an innovative approach to valorizing high-concentration N₂O emissions from industrial sources. Building on our current work, future efforts should prioritize the development of advanced catalysts for N₂O valorization and the broader implementation of the redox center separation strategy.

Author contributions

Yunshuo Wu: conceptualization, methodology, formal analysis, writing – original draft. Yuxin Sun: methodology. Xuanhao Wu: formal analysis. Haiqiang Wang: conceptualization, writing – review & editing, project administration. Lianzhou Wang: project administration, writing – review & editing. Zhongbiao Wu: project administration, writing – review & editing.

Conflicts of interest

The authors declare no competing financial interest.

Data availability

The data supporting this article have been included as part of the SI.

The SI file contains detailed methods descriptions of catalysts preparation and theoretical calculations. It also includes TEM, STEM and EDX mapping images, N₂ adsorption-desorption isotherms, XRD and Raman spectra, TGA and DTG curves, XPS data, DFT results, ODHP performance data, and a comparative table of N₂O-ODHP performance against previously reported catalysts. See DOI: https://doi.org/10.1039/d5ta03582a.

Acknowledgements

The authors gratefully acknowledge the financial support of this work from the National Natural Science Foundation of China (NSFC-524B2140, NSFC-52370120), the Zhejiang Provincial “Lead Wild goose” Research and Development Project (No. 2024C03114), and the Program for Zhejiang Leading Team of S&T Innovation (Grant No. 2013TD07).

References

1. M. T. McCulloch, A. Winter, C. E. Sherman and J. A. Trotter, 300 years of sclerosponge thermometry shows global warming has exceeded 1.5 °C, Nat. Clim. Change, 2024, 14(2), 171–177.
2. D. I. Armstrong McKay, A. Staal, J. F. Abrams, R. Winkelmann, B. Sakschewski, S. Loriani, I. Fetzer, S. E. Cornell, J. Rockström and T. M. Lenton, Exceeding 1.5°C global warming could trigger multiple climate tipping points, Science, 2022, 377(6611), eabn7950.
3. H. Tian, R. Xu, J. G. Canadell, R. L. Thompson, W. Winiwarter, P. Suntharalingam, E. A. Davidson, P. Ciais, R. B. Jackson, G. Janssens-Maenhout, M. J. Prather, P. Regnier, N. Pan, S. Pan, G. P. Peters, H. Shi, F. N. Tubiello, S. Zaehle, F. Zhou, A. Arneth, G. Battaglia, S. Berthet, L. Bopp, A. F. Bouwman, E. T. Buitenhuis, J. Chang, M. P. Chipperfield, S. R. S. Dangal, E. Dlugokencky, J. W. Elkins, B. D. Eyre, B. Fu, B. Hall, A. Ito, F. Joos, P. B. Krummel, A. Landolfi, G. G. Laruelle, R. Lauerwald, W. Li, S. Lienert, T. Maavara, M. MacLeod, D. B. Millet, S. Olin, P. K. Patra, R. G. Prinn, P. A. Raymond, D. J. Ruiz, G. R. van der Werf, N. Vuichard, J. Wang, R. F. Weiss, K. C. Wells, C. Wilson, J. Yang and Y. Yao, A comprehensive quantification of global nitrous oxide sources and sinks, Nature, 2020, 586(7828), 248–256.
4. X. Wu, J. Du, Y. Gao, H. Wang, C. Zhang, R. Zhang, H. He, G. Lu and Z. Wu, Progress and challenges in nitrous oxide decomposition and valorization, Chem. Soc. Rev., 2024, 53(16), 8379–8423.
5. M. Konsolakis, Recent Advances on Nitrous Oxide (N₂O) Decomposition over Non-Noble-Metal Oxide Catalysts: Catalytic Performance, Mechanistic Considerations, and Surface Chemistry Aspects, ACS Catal., 2015, 5(11), 6397–6421.
6. A. R. Ravishankara, J. S. Daniel and R. W. Portmann, Nitrous Oxide (N₂O): The Dominant Ozone-Depleting Substance Emitted in the 21st Century, Science, 2009, 326(5949), 123–125.
7. H. Liu, S. Yang, G. Wang, H. Liu, Y. Peng, C. Sun, J. Li and J. Chen, Strong Electronic Orbit Coupling between Cobalt and Single-Atom Praseodymium for Boosted Nitrous Oxide Decomposition on Co₃O₄ Catalyst, Environ. Sci. Technol., 2022, 56(22), 16325–16335.
8. Y. Wu, X. Wu, J. Fan, H. Wang and Z. Wu, Insights into the Roles of Different Iron Species on Zeolites for N₂O Selective Catalytic Reduction by CO, Environ. Sci. Technol., 2024, 58(51), 22583–22593.
9. L. Li, J. Xu, J. Hu and J. Han, Reducing Nitrous Oxide Emissions to Mitigate Climate Change and Protect the Ozone Layer, Environ. Sci. Technol., 2014, 48(9), 5290–5297.
10. E. A. Davidson and W. Winiwarter, Urgent abatement of industrial sources of nitrous oxide, Nat. Clim. Change, 2023, 13(7), 599–601.
11. F. Le Vaillant, A. Mateos Calbet, S. González-Pelayo, E. J. Reijerse, S. Ni, J. Busch and J. Cornella, Catalytic synthesis of phenols with nitrous oxide, Nature, 2022, 604(7907), 677–683.
12. S. Chen, X. Chang, G. Sun, T. Zhang, Y. Xu, Y. Wang, C. Pei and J. Gong, Propane dehydrogenation: catalyst development, new chemistry, and emerging technologies, Chem. Soc. Rev., 2021, 50(5), 3315–3354.
13. J. T. Grant, C. A. Carrero, F. Goeltl, J. Venegas, P. Mueller, S. P. Burt, S. E. Specht, W. P. McDermott, A. Chieregato and I. Hermans, Selective oxidative dehydrogenation of propane to propene using boron nitride catalysts, Science, 2016, 354(6319), 1570–1573.
14. D. Zhao, X. Tian, D. E. Doronkin, S. Han, V. A. Kondratenko, J.-D. Grunwaldt, A. Perechodjuk, T. H. Vuong, J. Rabeah, R. Eckelt, U. Rodemerck, D. Linke, G. Jiang, H. Jiao and E. V. Kondratenko, In situ formation of ZnO_x species for efficient propane dehydrogenation, Nature, 2021, 599(7884), 234–238.
15. W. Wang, S. Chen, C. Pei, R. Luo, J. Sun, H. Song, G. Sun, X. Wang, Z.-J. Zhao and J. Gong, Tandem propane dehydrogenation and surface oxidation catalysts for selective propylene synthesis, Science, 2023, 381(6660), 886–890.
16. R. Almallahi, J. Wortman and S. Linic, Overcoming limitations in propane dehydrogenation by codesigning catalyst-membrane systems, Science, 2024, 383(6689), 1325–1331.
17. H. Zhou, H. Li, L. Wang, S. Chu, L. Liu, L. Liu, J. Qi, Z. Ren, A. Cai, Y. Hui, Y. Qin, L. Song, X. Qin, J. Shi, J. Hou, Y. Ding, J. Ma, S. Xu, X. Tao, L. Li, Q. Yang, B. Hu, X. Liu, L. Chen, J. Xiao and F.-S. Xiao, Cobaltosilicate zeolite beyond platinum catalysts for propane dehydrogenation, Nat. Catal., 2025, 8, 357–367.
18. S. Chen, Y. Xu, X. Chang, Y. Pan, G. Sun, X. Wang, D. Fu, C. Pei, Z.-J. Zhao, D. Su and J. Gong, Defective TiO_x overlayers catalyze propane dehydrogenation promoted by base metals, Science, 2024, 385(6706), 295–300.
19. Y. Wang, J. Wang, P. Zheng, C. Sun, J. Luo and X. Xie, Boosting selectivity and stability on Pt/BN catalysts for propane dehydrogenation via calcination & reduction-mediated strong metal-support interaction, J. Energy Chem., 2022, 67, 451–457.
20. G. M. Beshara, I. Surin, M. Agrachev, H. Eliasson, T. Otroshchenko, F. Krumeich, R. Erni, E. V. Kondratenko and J. Pérez-Ramírez, Mechanochemically-derived iron atoms on defective boron nitride for stable propylene production, EES Catal., 2024, 2(6), 1263–1276.
21. X. Jiang, L. Sharma, V. Fung, S. J. Park, C. W. Jones, B. G. Sumpter, J. Baltrusaitis and Z. Wu, Oxidative Dehydrogenation of Propane to Propylene with Soft Oxidants via Heterogeneous Catalysis, ACS Catal., 2021, 11(4), 2182–2234.
22. J. H. Carter, T. Bere, J. R. Pitchers, D. G. Hewes, B. D. Vandegehuchte, C. J. Kiely, S. H. Taylor and G. J. Hutchings, Direct and oxidative dehydrogenation of propane: from catalyst design to industrial application, Green Chem., 2021, 23(24), 9747–9799.
23. Y. Wu, X. Wu, H. Wang and Z. Wu, Advances and challenges in N₂O valorization for alkane oxidative dehydrogenation to olefins, Green Chem. Eng., 2024 DOI:10.1016/j.gce.2024.1011.1005.
24. J. Zhou, Z. Gao, G. Xiang, T. Zhai, Z. Liu, W. Zhao, X. Liang and L. Wang, Interfacial compatibility critically controls Ru/TiO₂ metal-support interaction modes in CO₂ hydrogenation, Nat. Commun., 2022, 13(1), 327.
25. Z. Tang, Y. Li, K. Zhang, X. Wang, S. Wang, Y. Sun, H. Zhang, S. Li, J. Wang, X. Gao, Z. Hou, L. Shi, Z. Yuan, K. Nie, J. Xie, Z. Yang and Y.-M. Yan, Interfacial Hydrogen Spillover on Pd-TiO₂ with Oxygen Vacancies Promotes Formate Electrooxidation, ACS Energy Lett., 2023, 8(9), 3945–3954.
26. H. Cheng and A. Selloni, Surface and subsurface oxygen vacancies in anatase TiO₂ and differences with rutile, Phys. Rev. B:Condens. Matter Mater. Phys., 2009, 79(9), 092101.
27. J. Wang, H. Chen, Z. Hu, M. Yao and Y. Li, A Review on the Pd-Based Three-Way Catalyst, Catal. Rev., 2015, 57(1), 79–144.
28. E. Araujo-Lopez, L. Joos, B. D. Vandegehuchte, D. I. Sharapa and F. Studt, Theoretical Investigations of (Oxidative) Dehydrogenation of Propane to Propylene over Palladium Surfaces, J. Phys. Chem. C, 2020, 124(5), 3171–3176.
29. J. J. H. B. Sattler, J. Ruiz-Martinez, E. Santillan-Jimenez and B. M. Weckhuysen, Catalytic Dehydrogenation of Light Alkanes on Metals and Metal Oxides, Chem. Rev., 2014, 114(20), 10613–10653.
30. X. Huang, O. Akdim, M. Douthwaite, K. Wang, L. Zhao, R. J. Lewis, S. Pattisson, I. T. Daniel, P. J. Miedziak, G. Shaw, D. J. Morgan, S. M. Althahban, T. E. Davies, Q. He, F. Wang, J. Fu, D. Bethell, S. McIntosh, C. J. Kiely and G. J. Hutchings, Au–Pd separation enhances bimetallic catalysis of alcohol oxidation, Nature, 2022, 603(7900), 271–275.
31. Y. Sun, Y. Wu, Y. Bai, X. Wu, H. Wang and Z. Wu, High performance iridium loaded on natural halloysite nanotubes for CO-SCR reaction, Fuel, 2024, 357, 129938.
32. G. Kresse and J. Furthmuller, Efficiency of ab initio total energy calculations for metals and semiconductors using a plane-wave basis set, Comput. Mater. Sci., 1996, 6(1), 15–50.
33. G. Kresse and J. Furthmüller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Phys. Rev. B:Condens. Matter Mater. Phys., 1996, 54(16), 11169–11186.
34. J. Chen, S. Xiong, H. Liu, J. Shi, J. Mi, H. Liu, Z. Gong, L. Oliviero, F. Maugé and J. Li, Reverse oxygen spillover triggered by CO adsorption on Sn-doped Pt/TiO₂ for low-temperature CO oxidation, Nat. Commun., 2023, 14(1), 3477.
35. J. Chen, Y. Su, Q. Meng, H. Qian, L. Shi, J. A. Darr, Z. Wu and X. Weng, Palladium Encapsulated by an Oxygen-Saturated TiO₂ Overlayer for Low-Temperature SO₂-Tolerant Catalysis during CO Oxidation, Angew. Chem., Int. Ed., 2023, 62(49), e202310191.
36. L. Kuai, Z. Chen, S. Liu, E. Kan, N. Yu, Y. Ren, C. Fang, X. Li, Y. Li and B. Geng, Titania supported synergistic palladium single atoms and nanoparticles for room temperature ketone and aldehydes hydrogenation, Nat. Commun., 2020, 11(1), 48.
37. Z. Sheng, Z. Wu, Y. Liu and H. Wang, Gas-phase photocatalytic oxidation of NO over palladium modified TiO₂ catalysts, Catal. Commun., 2008, 9(9), 1941–1944.
38. C. Wu, Q. Liu, Y. Zhan, W. Tan, X. Wei, Q. Tong, H. Wan and L. Dong, Insights into the surface Structure-Sensitive photocatalytic oxidation of gaseous toluene on Pd/TiO₂ catalysts, Chem. Eng. J., 2023, 475, 146294.
39. V. Muravev, G. Spezzati, Y.-Q. Su, A. Parastaev, F.-K. Chiang, A. Longo, C. Escudero, N. Kosinov and E. J. M. Hensen, Interface dynamics of Pd–CeO₂ single-atom catalysts during CO oxidation, Nat. Catal., 2021, 4(6), 469–478.
40. Y. Zhang, C. X. Harris, P. Wallenmeyer, J. Murowchick and X. Chen, Asymmetric Lattice Vibrational Characteristics of Rutile TiO₂ as Revealed by Laser Power Dependent Raman Spectroscopy, J. Phys. Chem. C, 2013, 117(45), 24015–24022.
41. E. V. Kondratenko and M. Baerns, Catalytic oxidative dehydrogenation of propane in the presence of O₂ and N₂O—the role of vanadia distribution and oxidant activation, Appl. Catal., A, 2001, 222(1), 133–143.
42. E. V. Kondratenko, M. Cherian, M. Baerns, D. Su, R. Schlögl, X. Wang and I. E. Wachs, Oxidative dehydrogenation of propane over V/MCM-41 catalysts: comparison of O₂ and N₂O as oxidants, J. Catal., 2005, 234(1), 131–142.
43. A. Ates, C. Hardacre and A. Goguet, Oxidative dehydrogenation of propane with N₂O over Fe-ZSM-5 and Fe–SiO₂: Influence of the iron species and acid sites, Appl. Catal., A, 2012, 441–442, 30–41.
44. B. Wang and G. Manos, A novel thermogravimetric method for coke precursor characterisation, J. Catal., 2007, 250(1), 121–127.
45. Z. Ren, Z. Wu, W. Song, W. Xiao, Y. Guo, J. Ding, S. L. Suib and P.-X. Gao, Low temperature propane oxidation over Co₃O₄ based nano-array catalysts: Ni dopant effect, reaction mechanism and structural stability, Appl. Catal., B, 2016, 180, 150–160.
46. M. L. Bols, B. E. R. Snyder, H. M. Rhoda, P. Cnudde, G. Fayad, R. A. Schoonheydt, V. Van Speybroeck, E. I. Solomon and B. F. Sels, Coordination and activation of nitrous oxide by iron zeolites, Nat. Catal., 2021, 4(4), 332–340.
47. Z. Liu, Y. An, G. Xu, Y. Yu and H. He, Insight into the Promotion Effect of Trace Pd Doping on the Catalytic Performance of Ag/Al₂O₃ for C₃H₆-SCR of NO_x, Environ. Sci. Technol., 2023, 57(39), 14760–14767.
48. Q. Wu, H. He and Y. Yu, In situ DRIFTS study of the selective reduction of NO_x with alcohols over Ag/Al₂O₃ catalyst: Role of surface enolic species, Appl. Catal., B, 2005, 61(1), 107–113.
49. Y. Yu, H. He and Q. Feng, Novel Enolic Surface Species Formed during Partial Oxidation of CH₃CHO, C₂H₅OH, and C₃H₆ on Ag/Al₂O₃: An in Situ DRIFTS Study, J. Phys. Chem. B, 2003, 107(47), 13090–13092.
50. E. T. C. Vogt, D. Fu and B. M. Weckhuysen, Carbon Deposit Analysis in Catalyst Deactivation, Regeneration, and Rejuvenation, Angew. Chem., Int. Ed., 2023, 62(29), e202300319.
51. Y. Wu, Y. Sun, X. Wu, H. Wang and Z. Wu, Reversed Charge Transfer Enables Dual Active Sites on Ir/hBN for Synergistic N₂O Valorization and Propane Selective Oxidation, ACS Catal., 2024, 14(17), 13520–13530.
52. D. Wu, K. Chen, P. Lv, Z. Ma, K. Chu and D. Ma, Direct Eight-Electron N₂O Electroreduction to NH₃ Enabled by an Fe Double-Atom Catalyst, Nano Lett., 2024, 24(28), 8502–8509.
53. Y. Gao, H. Xu, K. Yin, Z. Li, J. Fan, X. Wu and Z. Wu, Modulating Charge Transfer for Photo-driven N₂O Reduction via Electronegativity Differences Between Cu and Support, Appl. Catal., B, 2025, 125528.
54. Y. Wu, H. Wang and Z. Wu, Synthesis of a Cu@ZnO_1−x/Al₂O₃ catalyst with high-density ZnO_1−x–Cu interfacial sites for enhanced CO₂ hydrogenation to methanol, J. Mater. Chem. A, 2024, 12(44), 30692–30706.
55. H. Zhang, C. Han, C. Li, P. Wang, H. Huang, S. Wang and J. Li, Design of Cu/ZnO/Al₂O₃ catalysts with a rich Cu–ZnO interface for enhanced CO₂ hydrogenation to methanol using zinc-malachite as the precursor, New J. Chem., 2023, 47(12), 5885–5893.
56. Y. Wu, H. Wang, F.-S. Xiao and Z. Wu, Gadolinium-Mediated Oxygen Affinity Induced Efficient Covalorization of CH₄ and N₂O in an Ir–Gd₂O₃ Single-Atom Catalyst, J. Am. Chem. Soc., 2025, 147(29), 25787–25798.
57. H. Lee, M. Shin, M. Lee and Y. J. Hwang, Photo-oxidation activities on Pd-doped TiO₂ nanoparticles: critical PdO formation effect, Appl. Catal., B, 2015, 165, 20–26.
58. M. T. Greiner, L. Chai, M. G. Helander, W.-M. Tang and Z.-H. Lu, Transition Metal Oxide Work Functions: The Influence of Cation Oxidation State and Oxygen Vacancies, Adv. Funct. Mater., 2012, 22(21), 4557–4568.
59. A. Kahn, Fermi level, work function and vacuum level, Mater. Horiz., 2016, 3(1), 7–10.
60. Z. Wang, C. Wang, S. Mao, B. Lu, Y. Chen, X. Zhang, Z. Chen and Y. Wang, Decoupling the electronic and geometric effects of Pt catalysts in selective hydrogenation reaction, Nat. Commun., 2022, 13(1), 3561.
61. X. Chang, Z. Lu, X. Wang, Z.-J. Zhao and J. Gong, Tracking C–H bond activation for propane dehydrogenation over transition metal catalysts: work function shines, Chem. Sci., 2023, 14(23), 6414–6419.

---