Low-temperature oxidative dehydrogenation of propane over NiV mixed oxides derived from LDH precursors

Propylene is a key raw ingredient for three important synthetic materials: plastics, rubber, and chemical fibers. The market need for propylene derivatives has grown quickly.¹ Propylene is manufactured primarily nowadays by the processes of steam cracking and fluidized catalytic cracking of petroleum by-products, such as naphtha and light diesel, as well as through the dehydrogenation of propane.² As the cost of alkanes from shale gas extraction drops, oxidative dehydrogenation propane (ODHP), which has the benefits of low energy consumption, low carbon deposit, and thermodynamic permissibility, has drawn increasing interest.^3,4 However, how to suppress deep oxidation as well as improve catalyst activity and olefin selectivity are still key issues for the industrialization of ODHP.⁵ 

The transition metal oxides, which include V-based, Cr-based, Ni-based, Mo-based, rare earth-based, and so on, are the majority of ODHP catalysts. Although Mo-based and V-based catalysts have high propylene selectivity, these catalysts require rather high reaction temperatures (T > 500 °C). Supported vanadium oxide-based catalysts are the most widely studied in ODHP,^6,7 but the acid–base properties of the supports must be controlled, and the basic supports that are conducive to propylene desorption can be used to obtain the catalysts with the better catalytic performance.⁸ In recent years, nickel oxide catalysts have been favored by researchers due to their ability to activate alkanes at low temperatures.⁹ Compared with other catalyst systems, nickel-based catalysts have the characteristics of low temperature, high activity, and low cost, being considered a kind of catalyst system with industrialization potential. Nieto’s group has extensively researched the oxidative dehydrogenation of ethane (ODHE) over nickel oxide-based catalysts and has put forth numerous strategies to increase the selectivity of NiO-based catalysts to olefins, such as doping second metals and loading on various supports.^10,11,12 They compared the effect of various dopant-doped NiO catalysts on ODHE and ODHP, respectively, and discovered that the same catalyst performed catalytically for propane much less effectively than it did for ethane.¹⁰ This phenomenon can be explained thermodynamically because the C–H bond energy of propane is lower than that of ethane.¹³ Therefore, propane is more easily oxidized to CO_x, making Ni-based catalysts with high olefin selectivity a challenging topic in ODHP. Indeed, our group has been committed to this research for many years.^14,15,16,17 It has been shown that the amount of V doping can change the conductivity type of the Ni_xV_1−xO_y system and thus affect the catalyst activity, and a small amount of vanadium oxides (x = 0.95–0.85) improved propylene selectivity significantly due to the synergetic effect between NiO and Ni₃V₂O₈ phases.^15,16 However, the low dispersion of VO_x and limited number of active sites could be blamed for the unsatisfactory propane conversion and the significant drop in propylene selectivity that occurs with increasing temperature. This is directly related to the preparation methods of Ni–V–O catalysts.

Layered double hydroxides (LDH) are highly applicable as catalysts or precursors due to their unique structure and properties,¹⁸ especially the outstandingly homogeneous dispersion of metal oxygen species. The general formula of LDH is [M²⁺_1–xM³⁺_x(OH)₂]^x+(A_x/n)ⁿ⁻·mH₂O, where M²⁺ and M³⁺ represent the divalent and trivalent metal ions occupying the center of the octahedral hydroxide on the laminate, respectively, and Aⁿ⁻ is the interlayer anion.¹⁹ The thermal degradation of LDH can create highly distributed mixed metal oxides (MMO), which have great potential for application in ODHP reaction.

There have been few investigations using Ni-based LDH as precursors of MMO catalysts for ODHP, with the majority of attention going toward the catalysts employing NiAl LDH as the precursors.^14,20,21,22 Smoláková et al.²² used NiAl LDH as precursors and found that the increase of Ni content was detrimental to propylene selectivity, which was attributed to the higher proportion of Ni²⁺ leading to dealkylation rather than dehydrogenation. Recently, Gao et al.¹⁴ synthesized three NiAl MMO catalysts for ODHP by different methods, employing NiAl LDH as precursors. It was found that the widely used variable pH coprecipitation method for the preparation of the catalysts should not be the best choice in ODHP because of the aggregation of NiAl LDH lamellas during the calcination process. Valverde et al.²⁰ produced NiAlV MMO catalysts from decavanadate intercalated NiAl LDH precursors for ODHP. They found that vanadium oxides incorporated in the catalysts improved propylene selectivity under oxygen excess conditions, and the isolated V–O phases exhibited higher catalytic activity than the V–O phases from interlayer decavanadate ions. Therefore, it will be an interesting research topic that the ODHP catalytic performances of NiV MMO derived from layer-doping NiV LDH precursors, which is also conducive to the analogy study with MgV and ZnAlV MMO.^23,24

In this study, NiV LDH precursors were synthesized by a constant pH coprecipitation method, and two variables, the Ni/V ratios and the calcination temperature, were chosen for ODHP tests. Through various types of characterization, an attempt was made to explain the catalytic results of NiV MMO catalysts in ODHP.

NiV LDH precursors (Ni/V molar ratios = 2:1, 3:1, 4:1, 6:1, and 8:1, denoted as xNiV LDH) were prepared by a constant pH coprecipitation method.¹⁴ Taking 4NiV LDH as an example, an aqueous solution (50 ml) of 16 mmol of NiCl₂·6H₂O (98.0%, Aladdin) and 4 mmol of VCl₃ (97% URChem) was slowly added dropwise to 1 mmol of Na₂CO₃ solution in N₂ atmosphere, while the dropping speed of the base solution (50 mmol NaOH and 12.5 mmol Na₂CO₃) was controlled to keep the pH around 10. The coprecipitation process was carried out in a water bath heated to 60 °C with vigorous stirring. Then the resulting suspension was transferred to a Teflon-lined stainless steel autoclave and subjected to hydrothermal treatment at 120 °C for 12 h. After the autoclave was cooled down to room temperature, the precipitate was collected and washed three times with deionized water and then dried at 80 °C in a vacuum drying oven to obtain 4NiV LDH. The NiV MMO catalysts were obtained after calcination of xNiV LDH in the air at 400, 500, 600, or 800 °C for 3 h, which are represented by 4NiV-T, x, and T being the Ni/V molar ratios and calcination temperature, respectively. For comparison, pure NiO was also prepared by the same procedure without the addition of VCl₃, denoted as NiO-500.

The crystallographic analysis of the samples was performed on a Rigaku Ultima IV multifunctional x-ray diffractometer using Cu Kα1 radiation (λ = 0.154 06 nm, 40 kV, 40 mA). The data were gathered in step scan mode with a scan speed of 10° min⁻¹ throughout the range of 5°–80°. The Scherrer equation was used to determine the average grain sizes of the samples [Eq. (S1)]. Calculations of the specific surface area and pore structure of the catalysts were carried out using N₂ adsorption–desorption measurements taken at 77 K in a Micromeritics ASAP 2020 device. Before the adsorption–desorption tests, ∼100 mg of the sample was degassed in a vacuum at 100 °C for 5 h. The specific surface area and pore volume were calculated, respectively, using the Brunauer-Emmett-Teller (BET) equation and the Barrett-Joiner-Halenda (BJH) model. The Ni/V ratios in the samples were measured by an inductively coupled plasma mass spectrometer (ICP-MS), Agilent Technologies 7900, and the results were in agreement with the theoretical values (Table I). Transmission electron microscopy (TEM) and high-resolution TEM (HR-TEM) images of the samples were obtained on an FEI Talos 200A instrument. H₂-temperature-programmed reduction (H₂-TPR) was tested on a multifunctional adsorption apparatus (Quantachrome Chem BET TPR/TPD Chemisorption Analyzer) equipped with a thermal conductivity detector (TCD). The catalyst (∼50 mg) was put into a U-shaped quartz reactor and pretreated by Ar at 300 °C for 1 h in order to remove the physisorbed impurity gas and water. After changing the gas to 10 vol. % H₂/Ar and bringing the temperature down to 80 °C, the catalyst was heated at a ramp rate of 10 °C min⁻¹ to 800 °C. The H₂ consumption was determined from the peak area using a calibration curve of standard H₂/Ar gas at 10 vol. % concentration. On a Thermo ESCALAB 250Xi multipurpose imaging electron spectrometer, x-ray photoelectron spectroscopy (XPS) analysis was performed. The binding energy was determined to be calibrated to the C1s peak at 284.8 eV. Measurements of UV–Vis diffuse reflectance spectroscopy (UV–Vis DRS) in the wavelength range of 220–800 nm were carried out using a Shimadzu UV-3600plus spectrophotometer. The diffuse reflectance electron spectra are expressed in coordinates of the Kubelka–Munk function [F(R)]-wavelength.

An ODHP catalytic performance test was carried out by adding 100 mg of catalyst and an appropriate amount of quartz sand to a fixed-bed quartz tube reactor, maintaining a filling height of 1.5 cm, and then connected to a temperature-controlled oven. The catalyst was activated for 1 h at 350 °C using a mixture of N₂ and O₂ at a flow rate of 36 ml min⁻¹. The reactor was then cooled to 250 °C, and the feed gas (C₃H₈/O₂/N₂ volume ratio = 1:1:8, GHSV = 24 L g⁻¹ h⁻¹) was introduced. The reaction temperature was varied in the range of 250–500 °C at 25 °C intervals. The relationship between propane conversion and propylene selectivity at 325 °C was investigated by varying the catalyst dosage (50–150 mg) or the total flow rate (20–80 ml min⁻¹). The product analysis was performed by a Shimadzu GC-2010Plus gas chromatography equipped with a barrier discharge ionization detector (BID) using a ShinCarbon ST 80/100 mechanical microfilled column.

The XRD patterns of NiV LDH with different molar ratios are shown in Fig. S1. It shows that the xNiV LDH precursors (x = 2, 3, and 4) possess the LDH structure, and the intensity of diffraction peaks increases with the increase of the Ni/V ratio. However, no LDH diffraction peaks are seen when the molar ratio of Ni/V increases from 6 to 8; on contrary, the diffraction peaks of nickel hydroxides are present. Therefore, the cation ratio (M²⁺/M³⁺) in LDH precursors should be between 2.0 and 4.0; otherwise, the LDH crystalline phase cannot be formed. After pretreatment at 500 °C, the XRD patterns of NiV MMO catalysts are shown in Fig. 1(a). The only crystalline phase detected for all catalysts belongs to NiO, with diffraction peaks at 2θ = 37.2°, 43.2°, 62.8°, 75.4°, and 79.4°, corresponding to the (1 1 1), (2 0 0), (2 2 0), (3 1 1), and (2 2 2) crystalline planes of NiO (JCPDF No. 47-1049). No crystalline phases associated with V species are observed, even for larger V contents. It indicates that the crystalline phase formed at a given calcination temperature seems to be independent of the Ni/V ratio in the layered precursors. The grain size of (1 1 1) crystallographic planes in NiO was estimated using the Scherrer equation [Table I, Eq. (S1)]. Table I presents that the grain sizes of the NiO in the xNiV-500 catalysts are significantly smaller than that of pure NiO. This suggests that the presence of V can hinder the formation of large grains of NiO.²⁵  Figure 1(b) shows the XRD patterns of the 4NiV MMO catalysts after calcination at different temperatures. 4NiV-400 and 4NiV-500 show only the diffraction peaks of the NiO phase. 4NiV-600 and 4NiV-800 show the phase of Ni₃V₂O₈ (JCPDS No. 74-1485) in addition to the NiO phase. The Ni₃V₂O₈ diffraction peaks are gradually sharpened with the increase of calcination temperature, especially for the samples calcined at 800 °C. Moreover, the NiO grain size increased significantly in 4NiV-800, which reflects the importance of calcination temperature on crystal phase and grain size. The elevated temperatures seem to favor the interaction between the amorphous vanadium-containing phase and NiO.^26,27

The N₂ adsorption–desorption isotherms of NiV MMO catalysts are shown in Fig. S2. A type II isotherm with an H3 hysteresis loop was found in all of the samples, which can be attributed to the space between the lamellae.²⁰ The corresponding BET specific surface areas and pore volumes are presented in Table I. Compared with xNiV-500, the specific surface area of pure NiO is rather low. The specific surface areas of xNiV-500 series catalysts vary regularly with Ni/V ratios, increasing with the Ni/V ratio and reaching a maximum (149.6 m² g⁻¹) when Ni/V = 4. The specific surface area then decreases after further increasing the Ni/V molar ratios. In general, the increase in specific surface area with increasing dopant content is almost linear after doping NiO with transition metals, which is attributed to the so-called “mutual protection effect” in the co-precipitated binary oxide system.²⁸ On the contrary, the specific surface areas of xNiV-500 present volcanic-type changes. This difference is primarily due to the lack of LDH structure and the crystallization of the second phase in 6NiV-500 and 8NiV-500.²⁹ This is also consistent with the XRD results. With the increase of calcination temperature, the specific surface areas of 4NiV-T decrease significantly. This is mostly attributable to the growth of NiO grains and the formation of the Ni₃V₂O₈ phase with an increase in calcination temperature.

A more detailed observation of the microstructure of 4NiV-500 was carried out using transmission electron microscopy (TEM). It shows that the 4NiV-500 catalyst consists of a large number of very thin nanosheets [Figs. 2(a) and 2(b)], demonstrating that the NiV MMO pretreated at 500 °C also has the lamellar characteristics like the LDH precursor. As can be seen in Fig. 2(c), more research using HR-TEM was conducted to explore its intricate microstructure. The HR-TEM image shows clear, uniform, and parallel lattice stripes. The distance between the lattice stripes is 0.242 nm, which is the crystallographic spacing of the (1 1 1) lattice plane belonging to cubic NiO.³⁰ It was discovered that some nanoparticles with particle size less than 5 nm were spread on NiO, which may be Ni–V–O species. These nanoparticles did not exhibit any noticeable lattice striations. Elemental mapping further determined that the Ni, V, and O elements were uniformly distributed on the surface of 4NiV-500 [Fig. 2(d)], verifying the high dispersion of LDH-derived MMO.

The calcination temperature can affect the crystalline phase of the NiV MMO catalyst.²⁶ We prepared a series of 4NiV-T catalysts at different calcination temperatures (400, 500, 600, and 800 °C) and tested their catalytic performance in ODHP, as shown in Fig. 3. We found that the calcination temperature affects the onset temperature of propane conversion on 4NiV-T, which is related to the grain size of NiO, the larger the grain size, the higher the onset temperature of propane conversion. The onset temperature of propane conversion at 4NiV-500 can be as low as 250 °C, and the highest initial selectivity of propylene (70%) is obtained. Indeed, particle size is not the only factor that affects the onset temperature of propane conversion. The completeness of the crystalline phase and the generation of other phases also may affect the onset temperature, which would be a complex process. The same complex situation also exists in the variety of propylene selectivity with reaction temperature. The propylene selectivity decreases as temperature increases over NiV-T catalysts, being similar to the Ni-based catalysts in the literature.³¹ However, the propylene selectivity over the 4NiV-800 catalyst is not significantly reduced within the test temperature range, although the activation temperature of propane is the highest on it. This can be attributed to the formation of a second crystalline phase, Ni₃V₂O₈, after calcination at 800 °C. In combination with the characterization, the variation of the catalytic activity of the 4NiV-T catalysts is mainly influenced by the specific surface area, particle size, and crystalline phase. The above results show that the satisfactory catalytic performance of the NiV MMO catalysts cannot be achieved only by adjusting the calcination temperature.

Therefore, we further regulated the Ni/V ratio in the samples. Figure 4 demonstrates the trends of propane conversion and propylene selectivity with temperature for the NiV MMO catalysts (xNiV-500) and NiO (NiO-500) after pretreatment at 500 °C. Table S1 summarizes the detailed results at 300, 325, and 350 °C for NiO-500 and xNiV-500 in ODHP. Evidently, compared with NiO, the NiV MMO can activate propane at a lower temperature, and doped V has little effect on propane conversion, but can significantly improve propylene selectivity within the temperature range of 250–375 °C. The smaller the NiO particle size, the more effective the catalyst is in improving olefin selectivity.³² This is consistent with the calculation of NiO grain size (Table I). Figure 4(a) shows that the propane conversion rises monotonically when the reaction temperature rises and has a volcano-type trend in the temperature range of 275–350 °C with the increment of Ni/V ratios. The highest propane conversion in this low-temperature range was observed when Ni/V = 4. However, when the Ni/V ratios continued to increase, the propane conversion decreased, and the order of propane conversion on the catalysts was as follows 4NiV > 3NiV > 2NiV > 6NiV > 8NiV. Figure 4(b) shows the corresponding propylene selectivity at different temperatures. The relationship between propylene selectivity and Ni/V ratios was consistent with propane conversion in the temperature range of 250–325 °C; the 4NiV-500 catalyst displayed the highest propylene selectivity. The relatively poor catalytic performances of the catalysts with the higher or lower Ni/V ratio (e.g., 8NiV-500 and 2NiV-500) can be attributed to large NiO grains and poor V species dispersion, respectively. In addition, the 2NiV-500 catalyst was found to have propane conversion of 100% but no selectivity to propylene at 400 °C, which is related to the higher relative amount of V in the NiV MMO, leading to high-temperature sintering of vanadium oxide. In terms of reaction kinetics, the apparent activation energies on the NiV MMOs have been calculated (see Fig. S3). The values (around 80 kJ mol⁻¹) are within the general range measured for ODHP on MMO-based catalysts derived from LDH and similar to that of the supported VO_x catalysts,^33,34 but obtained in a lower reaction temperature range.

In addition, we observed the influence of the propane conversion on the propylene selectivity at 325 °C by varying the contact time of feed gas and catalyst dosage (Fig. 5). A low influence is observed over NiO-500 and 4NiV-500, suggesting the consecutive over-oxidation of propylene should be fairly low, and the low propylene selectivity of pure NiO is the result of the over-oxidation of propane.¹⁰ It is worth mentioning that the selectivity to propylene of ∼60% over 4NiV-500 is nearly four times higher than that over NiO-500 and better than that over Nb–NiO¹¹ (∼50%, 350 °C) at a similar propane conversion, indicating that the moderate incorporation of V can greatly reduce the deep oxidation of propane on the NiO-based catalyst. The selectivity to propylene over other xNiV-500 catalysts is also significantly higher than that over NiO-500 but decreases with the increase of propane conversion, which suggests that the resulting propylene is oxidized to CO_x. Table II compares the ODHP catalytic performance at the low temperatures of our catalyst with the representative Ni-based catalysts reported in the literature. The results show that the addition of V is beneficial to reduce the activation temperature and obtain higher propylene selectivity, compared with other dopants.

By adjusting the Ni/V ratios, we found that the 4NiV-500 catalyst exhibited the best catalytic performance in ODHP compared with other xNiV-500 catalysts. Therefore, we further performed a stability test on the 4NiV-500 catalyst. It exhibited constant propane conversion (∼7%) and propylene selectivity (∼60%) during 20 h of operation at 300 °C (Fig. S4), stating clearly the great potential of 4NiV-500 as a low-temperature ODHP catalyst.

H₂-TPR was utilized in order to investigate the reducibility of the samples. As shown in Fig. 6, the H₂-TPR curves for NiV MMO catalysts can be fitted as three reduction peaks by Gaussian deconvolution (denoted as P1, P2, and P3, respectively) in the temperature range of 200–700 °C. However, only two maximum reduction peaks for pure NiO were detected at 384 and 427 °C, which can be attributed to the reduction of Ni (II) species in the surface and bulk phases, respectively [Fig. 6(a)].^29,38 In NiV MMO catalysts, the initial reduction peak around 370 °C is due to the reduction of non-stoichiometric Ni–O species on the surface, which is generated by weakly bound nickel oxide sites.³⁹ The second reduction peak appears in the temperature range of 430–470 °C and is associated with the reduction of the bulk phase Ni–O species,⁴⁰ while the reduction of the Ni–V–O species is associated with the maximum reduction peak P3. Compared to pure NiO, a shift in the reduction temperatures of the P1 and P2 peaks of the xNiV-500 catalysts toward the high-temperature region is observed, indicating that the doped V species hinders the reduction of NiO. It has been widely reported that the non-stoichiometric oxygen content can be reduced by doping NiO with a second metal, improving the olefin selectivity.^10,41,42 Higher valent metals have been found to be the most successful dopants because they can diminish the average oxidation state of Ni and the non-stoichiometric oxygen content.^12,43 The results indicate that V doping can reduce the non-stichiometric oxygen concentration, thus inhibiting the deep oxidation of propylene and improving the selectivity of propylene. This is consistent with the results of the catalytic activity evaluation.

As shown in Fig. 6(b), with the increase of calcination temperature, the reduction peaks of the samples treated at different temperatures shift toward higher temperatures, which is attributed to the agglomeration of catalysts at high temperatures so as to affect the catalyst specific surface area as well as the number of oxygen species. In addition, combined with the XRD analysis, the Ni₃V₂O₈ phase formed after calcination at high temperatures may have an interaction with the NiO phase, resulting in the obvious shift of the reduction peak. In addition to the reduction peaks position moving toward high temperature, the reduction peak area of the surface non-stoichiometric Ni (II) is significantly reduced, indicating that the activity and quantity of surface oxygen species are reduced after V doping. In other words, properly doped V can effectively regulate the activity and quantity of the surface oxygen species, which is directly related to propylene selectivity.¹⁶ Therefore, the propylene selectivity on NiV MMO is significantly higher than that on NiO, especially on 4NiV-500. Meanwhile, compared with NiO-500, the lower initial reduction temperature of the xNiV-500 catalyst is consistent with its low-temperature ODHP activity.

The H₂ consumption of the samples listed in Table III shows that pure NiO has the highest H₂ consumption (13.3 mmol g⁻¹), which is close to the predicted amount (13.4 mmol g⁻¹), suggesting that Ni²⁺ ions have been completely reduced to metallic Ni.⁴⁴ There is a significant difference in H₂ consumption between NiV MMO catalysts and pure NiO, and the total H₂ consumption of NiV MMO is relatively small. From the H₂ consumption, it can be obtained that the nickel element exists in the catalyst in the form of Ni²⁺, and the vanadium element is easily oxidized and exists in more complicated forms. Taking the H₂ consumption of the 4NiV-500 catalyst as an example, the valence of vanadium was calculated. If all V⁵⁺ is reduced to V³⁺ (all Ni²⁺ is reduced to Ni), the theoretical H₂ consumption should be 13.5 mmol g⁻¹, which is greater than the experimental value. In the second case, V⁵⁺ is reduced to V⁴⁺, or V⁴⁺ is reduced to V³⁺, and the theoretical H₂ consumption of 12.2 mmol g⁻¹ is less than the experimental value. This indicates that element V exists as mixed valence states in the catalyst.

The valence states of the elements Ni, V, and O in the 4NiV-T catalysts and the 4NiV LDH precursor were studied by XPS. Figure 7(a) shows the evolution of the core energy spectra Ni 2p_3/2. Due to the unique layered structure of LDH, the Ni species in 4NiV LDH exists as NiO₆ octahedra, and the corresponding spectrum is fitted to only two peaks at 855.9 and 861.7 eV, belonging to Ni (II) 2p_3/2 main peak and its satellite peak, respectively.^45,46 The Ni (II) 2p_3/2 main peak of 4NiV MMO becomes asymmetric with a shoulder peak at low binding energy, showing a typical oxide bimodal structure, which is the intrinsic peak shape of NiO (Fig. S5). However, the doping of transition metal V causes competition between the shielding electrons of the ligand around the Ni species and the electrons from the ligand around the neighboring transition metal V ion, which interferes with the surface environment of Ni (II), thus affecting the main peak intensity. As shown in Fig. 7(a) and Fig. S5, the Ni 2p_3/2 region of 4NiV MMO and NiO requires four peaks to reproduce the observed peak shape.⁴⁴ The main peak is divided into two peaks at 854.3 and 855.9 eV, attributing to the existence of Ni³⁺ species, Ni²⁺ vacancies, or Ni²⁺-OH sites, and the corresponding two satellite peaks are at 861.2 and 864.0 eV.^26,47,48

Figure 7(b) shows that V 2p_3/2 can be fitted as three peaks at 516.0, 517.1, and 518.1 eV,^49,50 which correspond to V³⁺, V⁴⁺, and V⁵⁺, respectively. This suggests that V³⁺ was oxidized to form V⁴⁺ and V⁵⁺ during the synthesis process. The binding energies of the V 2p_3/2 peak and the relative areas of the fitted peaks are presented in Table IV, and the relationship between the areas of the fitted peaks is considered as the content of the different oxidation states of vanadium. The 4NiV-T had a higher content of V⁴⁺ compared with 4NiV LDH. Comparing the V 2p_3/2 spectra of the fresh and used 4NiV-500 (Fig. S6), it is found that the used samples showed an increase in V⁴⁺ content, as well as a slight increase in V³⁺ content and a relative decrease in V⁵⁺ content (Table S2). This indicates that the surface V species had little change in ODHP.^51,52

For all samples, the O 1s spectra show three peaks at 529.8, 531.4, and 533 eV, attributed to lattice oxygen (O²⁻), OH⁻ species, and electrophilic surface oxygen (O₂⁻, O⁻), respectively.⁵³ It is evident from Table IV that the electrophilic surface oxygen (O₂⁻, O⁻) content decreases with increasing calcination temperature due to the increase in grain size with decreasing specific surface area.⁵⁴ However, the lattice oxygen content increases with increasing calcination temperature, which is consistent with generating a new crystalline phase (Ni₃V₂O₈). Combined with O 1s results of the fresh and used 4NiV MMO (Fig. S7), the concentration change of oxygen species is not obvious, indicating that a catalytic oxidation and reduction cycle in ODHP can be completed excellently on 4NiV MMO. This is consistent with the catalytic life test results (Fig. S4).

UV–Vis DRS was used as an auxiliary characterization for XPS to explore the properties of surface NiO or Ni–V–O species. In Fig. S8, two absorption bands can be observed at 377 and 660 nm for pure NiO, corresponding to the octahedrally coordinated Ni²⁺.⁴⁵ All xNiV-500 catalysts have different spectral characteristics from pure NiO. We further determined the UV–Vis edge energy (E_g) values of the xNiV-500 catalysts for estimating the degree of polymerization of vanadium species (Fig. S9). The edge energy of 3.38 eV for 4NiV-500 corresponds to the isolated monomeric vanadium,⁵⁵ which explains the high propylene selectivity over this catalyst.

In this work, NiV MMO catalysts with highly dispersed metal sites were successfully prepared using NiV LDH as the precursors. On the one hand, the appropriate calcination temperature (500 °C) of the same LDH precursor (Ni/V = 4) effectively reduced the propane activation temperature to 250 °C on the NiV MMO catalyst. This is directly related to the smaller grain size and high dispersion of NiO in the 4NiV MMO catalyst, which is the merit of the LDH precursor. Furthermore, the Ni₃V₂O₈ phase formed at high calcination temperature is beneficial to maintain propylene selectivity at the high reaction temperature. On the other hand, the properly doped V effectively inhibited the deep oxidation of propylene and improved the propylene selectivity. This is thanks to the lower vanadium surface density, which effectively regulated the activity and quantity of the surface oxygen species. The 4NiV-500 catalyst activated propane at 250 °C with an initial propylene selectivity of up to 70% and exhibited excellent life at 300 °C, having a potential as a low-temperature ODHP catalyst. However, the maintenance of propylene selectivity at high propane conversion is the next challenge the NiV MMO catalyst must face.

See the supplementary material for the supplementary figures and tables.

This research was supported by the National Natural Science Foundation of China (Grant No. 22262026), the Science and Technology Program of Inner Mongolia Autonomous Region of China (Grant No. 2020PT0003), the Key Research Program of Science and Technology at Universities of Inner Mongolia Autonomous Region (Grant No. NJZZ22588), the Fundamental Research Funds for the Inner Mongolia Normal University (Grant No. 2022JBQN084), and the Collaborative Innovation Center for Water Environmental Security of Inner Mongolia Autonomous Region of China (Grant No. XTCX003).

The authors have no conflicts to disclose.

Zhenxing Liu: Data curation (lead); Investigation (equal); Writing – original draft (equal). Jiang Wang: Conceptualization (lead); Funding acquisition (equal); Writing – review & editing (lead). Kailu Wu: Validation (equal). Aiju Xu: Funding acquisition (equal); Supervision (equal); Writing – review & editing (equal). Meilin Jia: Funding acquisition (equal); Resources (equal).

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