P-Doped V–Cr composite oxide catalysts for the ammoxidation of 2,6-dichlorotoluene

Abstract

This study presents a novel P-doped V–Cr composite oxide catalyst for the ammoxidation of 2,6-dichlorotoluene. P-doped V–Cr composite oxides, which crystallized in the hexagonal CrVO₃ structure, were synthesized by the solvothermal method with H₃PO₄ addition. The V–Cr composite oxide particles were composed of numerous nanorods (∼10 nm wide, 100–200 nm long), which assembled into compact microspheres (0.5–1 µm in diameter). The addition of H₃PO₄ led to the formation of loosely packed irregular and porous nanoparticles with increased specific surface area and pore sizes. Elemental mapping revealed a homogeneous surface distribution of V, Cr, and P. As the molar ratio of V, Cr, and P was set as 1 : 1 : 0.1 in the solvothermal reaction system, the resulting VCrP_0.10-N600 product exhibited a maximum specific surface area of 38.7 m² g⁻¹ and an average pore size of 7.8 nm. In contrast to the P-free VCr-N600, medium-strong acid sites predominated on the VCrP_0.10-N600 product, with their concentration exceeding those of weak and strong sites by over an order of magnitude. The catalytic performance of P-doped V–Cr composite oxides was evaluated for the ammoxidation of 2,6-dichlorotoluene to 2,6-dichlorobenzonitrile. The VCrP_0.10-N600 catalyst, featuring the optimal P content, achieved a conversion of 92.7%, a molar yield of 78.7% and a selectivity of 82.7%. These values represented improvements of 5.5%, 7.5% and 4.8%, respectively, compared with the P-free VCr-N600 catalyst. First principles methods calculations confirmed that P doping broadened the density of states (DOS) near the Fermi level, thereby increasing the yield, while the introduction of highly localized empty states enhanced selectivity. These theoretical findings were consistent with our experimental results.

Incorporating phosphorus into the V–Cr composite oxide modulates the electronic structure and increases both surface area and medium-strength Lewis acidity, leading to superior catalytic performance in the ammoxidation of 2,6-dichlorotoluene.

1. Introduction

Aromatic nitriles are recognized as crucial intermediates in organic synthesis and are widely employed in the production of pharmaceuticals, agrochemicals, and fine chemicals.¹ Conventional synthetic routes have limited their wide application due to the drawbacks of stringent reaction conditions, the use of hazardous cyanide reagents, and the generation of stoichiometric byproducts.² Ammoxidation is a catalytic one-step process wherein a methyl (or alkyl) group undergoes oxidative conversion with ammonia and oxygen to directly form a cyano group, thereby producing nitrile compounds.³ Currently, ammoxidation represents the most economical, environmentally benign, and efficient industrial-scale method for the synthesis of nitrile compounds. The key challenge is developing ammoxidation catalysts with both high activity and selectivity – a critical technological breakthrough required for this process. Extensive research and industrial practice have demonstrated that vanadium-based catalysts exhibit exceptional efficiency for the ammoxidation of methyl aromatics, with vanadium species serving as an indispensable component.⁴ These vanadium catalysts are particularly effective for the gas-phase ammoxidation of methyl-substituted aromatics (e.g., toluene, xylene, and halogenated toluenes)⁵ and heteroaromatics (e.g., methylpyridine, methylpyrazine)⁶ to their corresponding nitrile compounds. In this process, vanadium in the vanadia species undergoes reduction from the V⁵⁺ to V⁴⁺ oxidation state, with lattice oxygen being extracted via the Mars–van Krevelen mechanism.⁷ However, single-component V₂O₅ catalysts often show limited catalytic performance due to the deep oxidation behavior.⁸ The introduction of secondary elements enables the formation of binary composite-oxide systems with markedly improved catalytic properties. Thus, various vanadium-based binary composite-oxide catalysts, including V–Cr,⁹ V–P,¹⁰ V–Ti,¹¹ V–W¹² and V–Mo¹³ have been extensively studied. The catalytic behavior of vanadium-based composite oxides is typically governed by a bifunctional mechanism:¹⁴ VO_x acts as the redox center, while the secondary component provides suitable acidic sites for NH₃ adsorption. Introducing phosphorus as a third component induces profound changes in the redox behavior and acid site distribution of the V–Cr system, which collectively contribute to the improved catalytic performance.

In the case of the binary system, the V–Cr composite oxide is currently considered the most promising candidate. The incorporation of Cr promotes the dispersion of active vanadium sites and supplies appropriate acidic adsorption sites for NH₃,¹⁴ thereby significantly improving the ammoxidation activity toward chlorotoluenes. However, the activity and selectivity of V–Cr binary composite oxides remain unsatisfactory, particularly for ortho-substituted substrates like 2-chlorotoluene and 2,6-dichlorotoluene. The chlorine atoms next to methyl group in these substrates sterically hinder the adsorption and reaction, thereby lowering the ammoxidation activity.¹⁵ The introduction of a third component as a promoter may induce marked changes in both the structural characteristics and catalytic properties of the V–Cr system.

The solvothermal method followed by thermal treatment is a facile route to prepare various vanadium-based composite oxides such as CrVO₄,¹⁶ Cr₂V₄O₁₃,¹⁷ VOPO₄ (ref. 18) and MoV₂O₈.¹⁹ The crystal structure, morphology, and particle size can be tuned by adjusting the reaction conditions, such as the solvent, solvothermal duration, heating temperature, and types and amount of additives. Phosphate ions (PO₄³⁻) exhibit strong coordination capabilities, enabling selective adsorption onto specific crystal facets.²⁰ Furthermore, the introduction of phosphoric acid can create acid sites on the surface of composite oxides,²¹ influencing surface charge and reactant adsorption behavior.

In this study, CrVO₃ nanomaterials were prepared through a solvothermal method using inexpensive starting materials, including V₂O₅, CrO₃, and methanol. Phosphorus acid was incorporated as a third component in the solvothermal reaction system. The influence of phosphorus incorporation on the microstructure of CrVO₃ was systematically investigated. The potential of phosphorus to enhance the catalytic performance of the system was evaluated in the ammoxidation of 2,6-dichlorotoluene. This work developed a P-doped V–Cr composite oxide catalyst with high activity, offering valuable insights for the design of industrial catalysts.

2. Experimental section

2.1. Catalyst preparation

The V–Cr composite oxide was synthesized via a methanol-based solvothermal method. Specifically, V₂O₅ (Aladdin, 99%) and CrO₃ (Aladdin, 99%) were used at a 1 : 2 molar ratio (1 : 1 atom ratio of V and Cr). V₂O₅ was first dissolved in methanol (Aladdin, 99%) under stirring, followed by slow, portion-wise addition of CrO₃ due to its strong oxidative reactivity. After 30 min homogenization, the solution was transferred to an autoclave for solvothermal treatment at 180 °C for 24 h. The resulting precursor was collected by filtration, washed sequentially with deionized water and ethanol (Aladdin, 99%) three times, and dried at 60 °C for 12 h. The precursor was calcined in a tube furnace under N₂ atmosphere, employing a heating rate of 5 °C min⁻¹ to final temperatures of 500, 600, and 700 °C, yielding catalysts designated as VCr-N500, VCr-N600, and VCr-N700, respectively.

P-Doped V–Cr composite oxides were synthesized with a similar procedure, with H₃PO₄ added at varying V : P ratios (1 : 0.01, 1 : 0.05, 1 : 0.1, 1 : 0.15, 1 : 0.2). After the homogenization of the V/Cr precursor solution, specified amounts of phosphoric acid were added and stirred for 30 min. The mixture was then subjected to solvothermal treatment (180 °C, 24 h) in a Teflon-lined autoclave. The product was filtered, washed, and dried at 60 °C for 12 h to obtain the precursor. The precursors were calcined in a tube furnace under N₂ atmosphere at a heating rate of 5 °C min⁻¹ and held for 2 h at final temperatures of 500, 600, or 700 °C, yielding catalyst series designated as VCrP_x-N500/600/700, where x represents the initial V : P molar ratio (0.01, 0.05, 0.1, 0.15, 0.2).

2.2. Catalyst characterization

The morphology of the samples was characterized using field-emission scanning electron microscopy (FE-SEM) (SU801 microscope, acceleration voltage: 5–15 kV). High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and TEM analyses, along with energy dispersive spectrometer (EDS) elemental mapping, were performed on a Thermo Fisher Talos F200X STEM microscope operating at 200 kV.

The structural properties of the P-doped V–Cr composite oxide samples, including the specific surface area, average pore size, and pore volume, were analyzed using a Quantachrome Autosorb-iQ-MP analyzer via N₂ adsorption–desorption at −196 °C.

X-ray diffraction (XRD) patterns of the samples were recorded by a Bruker D8 Advance X-ray diffractometer equipped with a Cu Kα radiation source (λ = 1.5406 Å). The scan range was from 10 to 80° with a scanning step size of 0.0167°.

Fourier Transform Infrared spectra (FT-IR) were acquired using a Nicolet Nexus 470 spectrometer. Samples were prepared by mixing catalyst and KBr in a ratio of 1 : 100, followed by grinding, drying, and forming pellets for testing.

Difference IR spectra were collected as follows. The sample was heated in air to 350 °C and this temperature was maintained for 20 min. Then, the catalyst was cooled to room temperature and background IR spectra of the clean surface was collected. The quantitative analysis was carried out after 1 h pyridine adsorption and followed by evacuation at 250 °C and for 10 min in vacuum. To determine changes in acid sites due to desorption of pyridine, the temperature was raised to 350 °C and then cooled to room temperature before collecting another IR spectra. The same procedure followed for VCr-N600 and VCrP_0.10-N600 catalysts.

The reducibility of the samples was evaluated by temperature-programmed reduction with hydrogen (H₂-TPR) using a Tianjin Xianquan TP5080 system. Each sample was pretreated in a Ar flow at 300 °C for 60 min, cooled to 50 °C, and then analyzed under 5% H₂/Ar with 10 °C min⁻¹ ramp to 800 °C while monitoring thermal conductivity detector (TCD) response. NH₃ temperature-programmed desorption (NH₃-TPD) analysis was performed using a Micromeritics AutoChem II 2920 system. Samples were pretreated in He atmosphere 300 °C for 60 min, cooled to 50 °C for NH₃ adsorption, purged, then heated to 700 °C at 10 °C min⁻¹ under He flow with TCD detection.

2.3. Catalytic experiment

Catalyst evaluation was performed using a self-built fixed-bed microreactor system (9 mm inner diameter quartz tube), which was described in detail elsewhere.¹⁷ 1.0 g of catalyst was loaded into the fixed-bed reactor with a thermocouple inserted into the catalyst bed. The reaction temperature was maintained between 390 and 440 °C. Air and ammonia, regulated via mass flow controllers, were fed into the reactor. The liquid reactants were fed at controlled rates via a micro-syringe pump and injected directly into the reaction tube. The effluent products were dissolved in ethanol and analyzed using a Shimadzu GC-2014Plus gas chromatograph equipped with FID detector and nitrogen carrier gas. Conversion of 2,6-dichlorotoluene (2,6-DCT) and yield and selectivity of 2,6-dichlorobenzonitrile (2,6-DCBN) were calculated by using the following formulae:Conversion (%) = (n_2,6-DCT,in − n_2,6-DCT,out)/n_2,6-DCT,in × 100Yield (%) = n_2,6-DCBN,out/n_2,6-DCT,in × 100Selectivity (%) = n_2,6-DCBN,out/(n_2,6-DCT,in − n_2,6-DCT,out) × 100

2.4. First principle calculation

The electronic structure of CrVO₃ and P doped CrVO₃ were calculated through first principle methods. The kinetic energy cutoffs for the wavefunctions and the charge density were set to 50 Ry and 500 Ry, respectively. Integration over the Brillouin zone was performed using a 5 × 5 × 2 k-point mesh for SCF calculations. For density of states (DOS) calculations, a denser grid of 12 × 12 × 6 was employed in the NSCF step. The SCF calculation was performed using the lattice constants obtain from the refine of XRD pattern using the GASA code. The DOS patterns of CrVO₃ and 10% P doped CrVO₃ were studied using the virtual crystal approximation(VCA) via the quantum espresso code.

3. Results and discussion

3.1. Catalyst characterization

The crystallinity and phase composition of the samples were characterized. Fig. 1 displays the XRD patterns of VCrP_x-N500, VCrP_x-N600, and VCrP_x-N700 samples prepared with different phosphorus concentrations. All patterns exhibit distinct diffraction peaks. The characteristic reflections, positioned at 2θ values of 24.315° (012), 33.269° (104), 35.995° (110), 41.177° (113), 49.612° (024), 54.303° (116), 62.984° (214), and 64.710° (300), correspond to the hexagonal CrVO₃ pattern (PDF#71-0866). This pattern is consistently observed across all calcination temperatures. The intensity ratio of the (1 1 0) peak to the (1 0 4) peak is clearly larger than the standard value of that (PDF#71-0866), indicating that the crystal obtained under the present synthetic conditions has special anisotropic growth along the (1 0 4) plane.²² No significant shift in the peak position was observed for the VCrP_x samples with increasing P content. This indicates the existence of PO₄³⁻ in the grain boundaries of CrVO₃ crystallites rather than going into the lattice.²³ As illustrated in Fig. 1a, the diffraction peak intensities of P-doped VCrP_x samples calcined at 500 °C are markedly reduced relative to those of the P-free VCr-N500 sample. This reduction is accompanied by a slight peak broadening, indicating that the P species markedly slows down the crystal-growth rate of the CrVO₃ phase. In addition, several weak diffraction peaks at approximately 27.7 and 36.9° can be discerned, corresponding to the CrVO₄ phase (PDF#15-0296). Additionally, at elevated P contents (x > 0.05), the CrVO₄ phase is barely observed in the XRD patterns, implying that phosphorus addition effectively restrains the formation of CrVO₄. As shown in Fig. 1b, elevating the calcination temperature to 600 °C enhances the diffraction peak intensities while preserving their positions, which is attributed to the enhanced crystallization of CrVO₃. For the VCrP_0.10-N600, VCrP_0.15-N600 and VCrP_0.20-N600 samples, diffraction peaks associated with the CrVO₄ phase are negligible. This suggests that moderate phosphorus doping effectively inhibits the formation of the CrVO₄ phase. The P species tend to segregate to the surface of the CrVO₃ primary particles, forming a core–shell structure.²⁴ Charge compensation for P doping requires oxygen incorporation near the dopant sites, resulting in a surface-enriched oxygen profile due to the heterogeneous distribution of P species. Because synthesis under a N₂ atmosphere fails to provide sufficient oxygen to the surface of the P-doped Cr–V composite oxides, oxygen vacancies are formed. Consequently, the incorporation of an amount of phosphorus effectively suppresses the formation of the CrVO₄ phase by stabilizing the lattice. Fig. 1c reveals that at elevated temperatures (700 °C), a fraction of the CrVO₃ particles decompose into Cr₂O₃ and the CrVO₄ phase.

Fig. 1. XRD patterns of VCrP_x catalysts calcined at 500 °C (a), 600 °C (b), and 700 °C (c) under N₂ atmosphere.

As shown in Fig. 2, FT-IR spectra of VCrP_x-N600 catalysts (x = 0–0.20) exhibit identical characteristic bands: 3470 cm⁻¹ (ν_O–H), 1710 cm⁻¹ (δ_H–O–H), 1012 cm⁻¹ (ν_{V O}), 573 cm⁻¹ (δ_V–O–V), and 750 cm⁻¹ (ν_Cr–O). The consistent patterns suggest the incorporation of phosphorus does not significantly alter the V–Cr–O framework, though P-related bands (expected at 1050–1150 cm⁻¹ for ν_{P O}) were too weak to be detected.

Fig. 2. FTIR spectra of VCrP_x-N600 catalysts (x = 0–0.20).

The difference IR spectra of pyridine adsorbed onto VCr-N600 and VCrP_0.10-N600 catalysts, followed by evacuation at 250 °C and 350 °C, are shown in Fig. 3. Pyridine can react with a Brønsted acid to form pyridinium ion with IR features at 1539 cm⁻¹ and 1637 cm⁻¹, and react with Lewis acid to form a pyridine-Lewis coordination complex with IR features at 1446 cm⁻¹ and 1622 cm⁻¹.²⁵ The concentration of Brønsted and Lewis-acid sites referenced to unit weight of catalyst has been obtained and shown in Table 1.

Fig. 3. FT-IR difference spectra recorded after pyridine adsorption on VCr-N600 and VCrP_0.10-N600, followed by evacuation at 250 °C and 350 °C.

Table 1. Results of acid sites quantification.

Catalyst	T (°C)	Brønsted-acid sites (µmol g−1)	Lewis-acid sites (µmol g−1)	Total acid sites (µmol g−1)	B/L
VCr-N600	250	26.97	7.59	34.56	3.55
	350	17.16	4.50	21.66	3.81
VCrP0.10-N600	250	54.37	43.89	98.26	1.24
	350	42.10	23.22	65.32	1.81

SEM analysis of the VCr-N600 and VCrP_x-N600 (x = 0.05, 0.10 and 0.15) catalysts reveal distinct phosphorus-dependent morphological evolution. Fig. 4 demonstrates that increasing P content leads to particle size reduction and increased surface roughness. The VCr-N600 catalyst exhibits well-defined compact microspheres (0.5–1 m), composed of tightly stacked nanorods (Fig. 4a–c). Following P doping, the VCrP_0.05-N600 catalyst (Fig. 4d–f) maintains a microsphere structure, but with irregular morphology composed of quasi-spherical nanoparticle clusters (50–100 nm). The VCrP_0.10-N600 catalyst (Fig. 4g–i) consists of densely assembled nanoparticles (20–50 nm), exhibiting obviously higher surface area compared to VCr-N600. Fig. 4j–l presents the SEM images of VCrP_0.15-N600 samples, revealing dense spherical aggregates with reduced porosity, indicative of phosphorus-mediated particle fusion at a high level of loading.

Fig. 4. SEM images of VCr-N600 (a)–(c), VCrP0.05-N600 (d)–(f), VCrP0.10-N600 (g)–(i), VCrP0.15-N600 (j)–(l).

TEM analysis coupled with elemental mapping was conducted on the catalyst series (VCr-N600 to VCrP_0.15-N600, Fig. 5), demonstrating phosphorus-dependent nanostructural evolution and confirming homogeneous nanoscale distribution of V, Cr, P, and O species. TEM images reveal that VCr-N600 (Fig. 5a–c) consists of tightly packed nanorods (10 nm wide × 100–200 nm long), while VCrP_0.05-N600 (Fig. 5d–f) transitions to looser nanoparticle assemblies. VCrP_0.10-N600 (Fig. 5g–i) exhibits optimal porosity with 10–20 nm particles and visible voids, whereas VCrP_0.15-N600 (Fig. 5j–l) forms larger (20–30 nm) and densely packed particles. This structural progression correlates with SEM observations, with VCrP_0.10-N600's balanced porosity/nanoparticle architecture suggesting superior catalytic potential.

Fig. 5. TEM images of VCr-N600 (a)–(c), VCrP_0.05-N600 (d)–(f), VCrP_0.10-N600 (g)–(i), VCrP_0.15-N600 (j)–(l).

EDS elemental mapping (Fig. 6) confirms homogeneous distribution of V, Cr, O, and P across all phosphorus-modified catalysts. The increasing P signal intensity with higher phosphorus content (0.05 → 0.15) validates successful incorporation, while the absence of crystalline phosphorus phases in XRD suggests amorphous dispersion of P species within the oxide matrix.

Fig. 6. STEM images and EDS elemental mapping of VCrP_0.05-N600 (a), VCrP_0.10-N600 (b), and VCrP_0.15-N600 (c) samples.

Textural properties of VCr-N600 and phosphorus-modified VCrP_x-N600 (x = 0.05, 0.10, 0.15) catalysts were analyzed via N₂ physisorption and BET measurements (Fig. 7), revealing phosphorus-dependent pore structure evolution and surface area enhancement up to 150% compared to unmodified VCr-N600. Fig. 7 reveals all catalysts exhibit type IV isotherms with H3 hysteresis loops, indicating slit-shaped mesopores formed by irregular nanoparticle aggregation. The absence of adsorption plateaus suggests pore size distributions extending into the macropore range (>50 nm).

Fig. 7. Adsorption–desorption isotherms for synthesized VCr-N600 and VCrP_x-N600 (x = 0.05, 0.10, 0.15) samples.

The phosphorus-dependent textural evolution is further demonstrated in Fig. 8 and Table 2. The surface area, pore volume and size all increase with the addition of phosphorus. VCrP_0.10-N600 achieves optimal porosity with 38.7 m² g⁻¹ surface area, exposing maximal active sites. Furthermore, VCrP_0.10-N600 exhibits the largest average pore size of 7.8 nm, which is consistent with the observed results of SEM and TEM, and facilitates the desorption of product molecules. At higher phosphorus loading (VCrP_0.15-N600), both surface area (37.4 m² g⁻¹) and average pore size (3.4 nm) decrease, corroborating SEM/TEM-observed particle densification phenomenon.

Fig. 8. Pore size distributions for synthesized VCr-N600 and VCrP_x-N600 (x = 0.05, 0.10, 0.15) samples.

Table 2. Textural and structural features of synthesized samples.

Catalysts	Surface area (m2 g−1)	Micropore volume (m3 g−1)	Average pore size (nm)
VCr-N600	18.6	0.124	1.9
VCrP0.05-N600	30.9	0.137	3.4
VCrP0.10-N600	38.7	0.172	7.8
VCrP0.15-N600	37.4	0.105	3.4

H₂ temperature-programmed reduction (TPR) profiles of VCr-N600 and VCrP_x-N600 (x = 0.05, 0.10, 0.15) catalysts are presented in Fig. 9, with quantitative reduction characteristics summarized in Table 3, revealing phosphorus-dependent redox behavior shifts. In the P-free VCr-N600 sample, the TPR profile appears as a broad peak that can be deconvoluted into two peaks, with peak temperatures at 336 °C and 550 °C, corresponding to the hydrogen consumption reduction of the CrVO₄ phase and the CrVO₃ phase,²⁶ respectively. With the addition of phosphorus, the TPR peaks become narrower and more symmetric. Numerical fitting shows the two peak temperatures gradually converging, indicating the primary reduction peak becomes predominantly that of CrVO₃. This observation is consistent with the results of XRD. H₂-TPR results (Fig. 9 and Table 3) show reduction temperatures at 536 °C (VCr-N600), 544 °C (VCrP_0.05-N600), 559 °C (VCrP_0.10-N600), and 581 °C (VCrP_0.15-N600) – all significantly lower than pure V₂O₅'s first reduction step (632 °C).¹³ Increasing phosphorus content raises reduction temperatures and increases hydrogen consumption, with VCrP_0.10-N600 exhibiting optimal redox properties for ammoxidation.

Fig. 9. H₂-TPR profiles for the prepared VCr-N600 and VCrP_x-N600 (x = 0.05, 0.10, 0.15) catalysts.

Table 3. The H₂ consumption in TPR for the prepared catalysts.

Catalysts	Peak temperature (°C)	H2 consumption (µmol g−1)
VCr-N600	536	0.0883
VCrP0.05-N600	544	0.0896
VCrP0.10-N600	559	0.0921
VCrP0.15-N600	581	0.0957

Ammonia temperature-programmed desorption (NH₃-TPD) profiles of VCr-N600 and VCrP_x-N600 (x = 0.05, 0.10, 0.15) are presented in Fig. 10, with acid site quantification data summarized in Table 4, revealing phosphorus-dependent acidity enhancement. NH₃ adsorption capacity, which correlates directly with acid site strength, is classified by desorption temperature as weak (<200 °C), medium (200–400 °C), and strong (>400 °C), where desorption temperature reflects binding energy.

Fig. 10. NH₃-TPD profiles of VCr-N600 and VCrP_x-N600 (x = 0.05, 0.10, 0.15) catalysts.

Table 4. Acid site distribution results from NH₃-TPD analysis.

Catalysts	Acid sites distribution, n(NH3) (mmol)
	Weak		Medium		Strong
	(g−1)	(m−2) × 103	(g−1)	(m−2) × 103	(g−1)	(m−2) × 103
VCr-N600	0.2099	11.28	0.0211	1.134	—	—
VCrP0.05-N600	0.1590	5.145	0.3727	12.06	—	—
VCrP0.10-N600	0.0314	0.811	0.4492	11.61	0.0234	0.6046
VCrP0.15-N600	0.1039	2.778	0.2227	5.954	0.5477	14.64

Peak deconvolution allows quantitative acid site determination. Fig. 10 and Table 4 reveal that the VCr-N600 catalyst exhibits NH₃-TPD desorption peaks at 97.3 °C (weak acid sites) and 358.2 °C (medium-strength acid sites), with weak acid sites dominating at a ratio of almost 10 : 1. The weak adsorption capacity results in insufficient NH₃ residence time on the catalyst surface, hindering complete reaction progression. With the addition of phosphorus, the NH₃-TPD profile of VCrP_0.05-N600 shows desorption peaks at 83.3 °C (weak acid sites) and 323.6 °C (medium-strength acid sites), with the latter dominating at a ratio exceeding 2 : 1. This indicates enhanced NH₃ adsorption capacity and prolonged surface residence time, significantly promoting catalytic performance.²⁷ With an increase in phosphorus content, the NH₃-TPD profile of VCrP_0.10-N600 exhibits three distinct desorption peaks at 89.6 °C (weak acid sites), 270.8 °C (medium-strength sites), and 646.5 °C (strong acid sites), indicating a hierarchical acidity distribution. The VCrP_0.10-N600 catalyst has the most amount of medium-strength acid sites among all these catalysts. Furthermore, medium-strength sites dominate, with their population exceeding that of weak and strong sites by more than a factor of ten. The high proportion and quantity of medium-strength acid sites along with balanced adsorption strength and residence time could optimally promote catalytic reactions. With a further increase in P content, the NH₃-TPD profile of the VCrP_0.15-N600 catalyst exhibits NH₃-TPD peaks at 378.4 °C (medium-strength sites) and 673.2 °C (strong acid sites), with strong acid sites being dominating. The excessive number of strong acid sites results in overly strong adsorption of NH₃, making desorption difficult and thereby unfavorable for the catalytic cycling process.

According to the bifunctional mechanism,¹⁴ in the surface of P-doped V–Cr composite oxide catalysts, VO_x acts as the redox center and supplies lattice oxides, while Cr and P generate Brønsted and Lewis acid sites, respectively, facilitating NH₃ adsorption. The proposed reaction mechanism for the ammoxidation of 2,6-DCT over P-doped V–Cr composite oxide catalysts is shown in Fig. 11.

Fig. 11. Proposed reaction mechanism for the ammoxidation of 2,6-DCT over P-doped V–Cr composite oxide catalysts.

3.2. Catalytic performance

Using 2,6-dichlorotoluene (2,6-DCT) as the substrate, single-factor screening experiments were conducted with the VCr-N600 catalysts to systematically evaluate the effects of catalyst loading, temperature, NH₃/2,6-DCT and air/2,6-DCT molar ratio on ammoxidation performance. The effect of 2,6-DCT loading over VCr-N600 catalyst was studied and results are presented in Fig. 12a. When the 2,6-DCT loading over the VCr-N600 catalyst is below 0.05 g g_Cat⁻¹, the conversion is high but molar yield and selectivity are low due to over-oxidation from prolonged contact time. At 0.05 g g_Cat⁻¹ loading, optimal performance is achieved with 69.2% yield and 77.9% selectivity for 2,6-dichlorobenzonitrile (2,6-DCBN). Higher loading (>0.05 g g_Cat⁻¹) reduces both conversion and yield as insufficient reaction time prevents complete conversion. As shown in Fig. 12b, low reaction temperatures (<420 °C) result in poor conversion and yield due to insufficient catalyst activity, while high temperatures (>440 °C) reduce yield and selectivity due to over-oxidation side reactions. Optimal performance was achieved at 420 °C, yielding a maximum 2,6-DCBN of 69.2%. Fig. 12c illustrates that low NH₃/2,6-DCT ratios limit the availability of reactive ammonia species, while excessive ratio decrease the effective contact time between the substrate and catalyst, thereby inhibiting full conversion. Fig. 12d reveals that low air/2,6-DCT ratios (<20) cause over-oxidation due to prolonged contact time, while high ratios (>20) limit contact, leading to incomplete conversion and reduced yields. Consequently, the optimal feed composition was established as a 1 : 3 : 20 molar ratio of 2,6-DCT : NH₃ : air.

Fig. 12. Effects of reaction conditions on the ammoxidation of 2,6-DCT catalyzed by VCr-N600. (a) Influence of loading (T = 420 °C, n(2,6-DCT) : n(NH₃) : n(air) = 1 : 3 : 20); (b) influence of reaction temperature (loading = 0.05 g g_Cat⁻¹ h⁻¹, n(DCT) : n(NH₃) : n(air) = 1 : 3 : 20); (c) influence of ammonia ratio (T = 420 °C, loading = 0.05 g g_Cat⁻¹ h⁻¹, n(DCT) : n(air) = 1 : 20); (d) influence of air ratio (T = 420 °C, loading = 0.05 g g_Cat⁻¹ h⁻¹, n(DCT) : n(NH₃) = 1 : 3).

Under optimized conditions (T = 420 °C, loading = 0.05 g g_Cat⁻¹ h⁻¹, n(2,6-DCT) : n(NH₃) : n(air) = 1 : 3 : 20), the catalytic performances of VCr-N600 and VCrP_x-N600 (x = 0.01–0.20) are systematically compared, and the results are shown in Fig. 13. It demonstrates that P-doped V–Cr composite oxide catalysts exhibit enhanced performance in 2,6-DCT ammoxidation, with significant improvements in conversion, yield, and selectivity. The degree of enhancement shows clear phosphorus-loading dependence. With increasing phosphorus content, the catalyst particle size decreases while specific surface area and pore size gradually increase, exposing more active sites. VCrP_0.10-N600 exhibits the highest surface area, pore volume, and pore size, coupled with significantly enhanced medium-strength acid sites. This optimized architecture resulted in peak catalytic activity of 92.7% conversion, 76.7% molar yield and 82.7% selectivity, correspond to significant increases of 5.5% in conversion, 7.5% in yield and 4.8% in selectivity compared to phosphorus-free VCr-N600. Further phosphorus addition triggers nanoparticle agglomeration, reducing surface area and pore size, and decreasing active site exposure. Furthermore, dominant strong acid sites overly stabilize NH₃ adsorption, disrupting catalytic cycling and degrading performance.

Fig. 13. Performance comparison chart of catalysts with different phosphorus contents.

Table 5 presents comparative performance data of some binary catalysts reported in literature for the ammoxidation of 2,6-DCT. VCrP_0.10-N600 catalyst demonstrates superior 2,6-DCBN yield and selectivity, outperforming conventional binary composite oxide catalysts and exhibiting promising potential for industrial-scale applications.

Table 5. Comparing the results of the ammoxidation of 2,6-DCT in the references.

Catalyst	T (°C)	Conversion (%)	Yield (%)	Selectivity (%)	Ref.
VCr-N600	420	87.2	69.2	77.9	This work
VCrP0.10-N600	420	92.7	76.7	82.7	This work
VPO_G0.5	410	97	73	76	18
Nano Cr2V4O13	430	87	71	80	17
Nano VCrO	335	78	70	90	11
Bulk VPO	440	92	55	60	28
VPO/SiO2	380	87	57	66	29
VPO/SiO2	400	90	57	63	30

A long-term catalytic test was conducted on the VCrP_0.10-N600 sample to evaluate its time-on-stream performance in the ammoxidation of 2,6-DCT to 2,6-DCBN. As shown in Fig. 14, the conversion of 2,6-DCT remained relatively stable over 24 hours on-stream. Meanwhile, the yield and selectivity to 2,6-DCBN remained essentially unchanged for at least 20 hours at 420 °C. These results suggest that the stability of the VCrP_0.10-N600 catalyst in the ammoxidation of 2,6-DCT requires further improvement under the applied reaction conditions.

Fig. 14. Time-on-stream performance of VCrP_0.10-N600 catalyst (reaction conditions: T = 420 °C, loading = 0.05 g g_Cat⁻¹ h⁻¹, n(2,6-DCT) : n(NH₃) : n(air) = 1 : 3 : 20).

3.3. First principle calculation

In the following, to reveal the effect of doped P, the electronic structures were studied using first principle methods. The DOS patterns of CrVO₃ and 10% P doped CrVO₃ were studied using the virtual crystal approximation (VCA) via the quantum espresso code. Although VCA may oversimplify certain details, it is capable of capturing the general trends in electronic structure and has been widely applied to study doped compounds.³¹ As shown in Fig. 15, we found that the density of states (DOS) around Fermi level were great adjusted by the doped P atoms. The significant broadening of the density of states around the Fermi level, along with the sharpening of peaks above it, reveals that the empty energy levels become highly localized compared to those in pure CrVO₃. The broadening of the DOS near the Fermi level generally indicates an expansion in the energy distribution range of electronic states. During the C–H bond activation process, the reactant molecule must donate electrons to the empty orbitals of the catalyst (σ donation) while simultaneously accepting back-donated electrons from the occupied orbitals of the catalyst (π back-donation) to effectively weaken the C–H bond.³² Therefore, broadening of the DOS near the Fermi level may enhance the electron-accepting or electron-donating ability of the catalyst surface toward reactant molecules, thereby accelerating electron transfer efficiency. This could significantly improve the catalytic performance. Meanwhile, the sharp, localized empty states above E_f may selectively interact with key reaction intermediates, and steer the reaction pathway towards the desired product. Thus, P doping not only significantly enhances the catalytic activity but also improves the selectivity, a dual enhancement that is clearly validated by our experimental results.

Fig. 15. The DOS pattern of CrVO₃ (left) and 10% P doped CrVO₃ (right).

4. Conclusions

Vanadium–chromium–phosphorus oxide catalysts with varying P content were synthesized via solvothermal method using V₂O₅, CrO₃ and H₃PO₄ as starting materials. Comprehensive characterization was performed including XRD (crystal structure), SEM/TEM (morphology), EDS (surface composition), and BET (textural properties). After adding an appropriate amount of phosphorus, calcination at 600 °C under N₂ atmosphere yielded V–Cr composite oxides characterized by a well-crystallized pure CrVO₃ phase. The VCr-N600 catalyst consists of tightly packed microspheres (0.5–1 µm diameter) formed by the stacking of nanorods with widths of ∼10 nm and lengths of 100–200 nm. Phosphorus incorporation fundamentally altered the catalyst morphology, transforming it from aggregates of nanorods into a loose, porous structure composed of packed irregular nanoparticles. This structural evolution resulted in an increased specific surface area and pore size. The V, Cr, and P elements were homogeneously distributed on the surface of the catalyst. The VCrP_0.10-N600 catalyst with V/Cr/P = 1 : 1 : 0.1 exhibits maximum surface area, pore volume, pore size, and most medium-strength acid sites along with balanced acid sites for optimal NH₃ adsorption strength. This structural optimization enhances catalytic cycling and significantly improves reaction activity. Under the reaction conditions of T = 420 °C, loading = 0.05 g g_Cat⁻¹ h⁻¹, n(2,6-DCT) : n(NH₃) : n(air) = 1 : 3 : 20 for the ammoxidation of 2,6-DCT, the catalyst demonstrated optimal performance, achieving 92.7% conversion of 2,6-DCT, 76.7% molar yield and 82.7% selectivity of 2,6-DCBN – representing increases of 5.5%, 7.5%, and 4.8%, respectively, over the phosphorus-free VCr-N600 catalyst. From our first-principles calculations, we reveal that the mechanism behind the improved yield and selectivity lies in the P-dopant-induced modification of the DOS distribution around the Fermi level.

Author contributions

W. T., G. X. and L. L. designed the experiments and wrote the original manuscript. D. Z. performed the experiments. L. S. analyzed the data. T. L. and Q. Y. corrected the paper. All authors discussed the results and commented on the manuscript.

Conflicts of interest

The authors declare no competing financial interest.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6ra03224a.