Oxidative dehydrogenation of ethylbenzene using nitrous oxide over vanadia–magnesia catalysts

Abstract

A series of V–Mg–O catalysts with different loadings of vanadia were prepared by the wet impregnation method and the effect of the local structure of these catalysts on the oxidative dehydrogenation of ethylbenzene with N₂O was investigated. High styrene selectivity (∼97%) was obtained at 773 K. The characterization of catalysts with methods such as XRD, FTIR, UV-visible, TPR, NMR and Laser Raman spectroscopy suggested that magnesium orthovanadate is the predominant vanadium containing phase and the size of the orthovanadate domains increases with increasing vanadia loading. The rate of ODH of ethylbenzene per V atom increases with vanadia loading and reaches a maximum at 10 wt%. The specific activity, i.e. the conversion of ethylbenzene per unit surface area of the catalysts, also exhibited a maximum at a vanadia loading of 10 wt% leading to the conclusion that activity of these catalysts is due to the presence of very small domains of Mg₃(VO₄)₂ on the surface of MgO rather than crystallites of bulk Mg₃(VO₄)₂. The higher styrene yield in the presence of N₂O can be ascribed to the ability of N₂O to keep vanadium species at a higher oxidation state.

---

1. Introduction

Oxidative dehydrogenation of ethylbenzene (EB) has been proposed recently as an alternative method for the production of styrene, an important monomer for the synthetic polymers.^1–6 The current commercial processes, based on the dehydrogenation of ethylbenzene using an iron oxide catalyst, require high reaction temperatures (873–923 K) to achieve favorable EB conversions. An excess amount of superheated steam is required to provide the energy needed by the reaction, to dilute the reactant and to reduce catalyst coking.^7,8 The process is equilibrium-limited and energy intensive, hence development of an oxidative dehydrogenation (ODH) process is necessary. Higher degree of EB conversion can be achieved at significantly lower temperatures with the use of an oxidant since the ODH process is exothermic and the conversion would not be equilibrium limited.

CO₂^9–14 and SO₂¹⁵ were suggested as mild oxidizing agents for this process instead of oxygen, which causes deep oxidation of EB to CO_x. However catalysts deactivate rapidly in a CO₂ atmosphere due to the formation of carbonaceous deposits.¹⁶ The use of SO₂ led to toxic and corrosive by-products like CS₂, COS and benzothiophene.

There are some works reported in the literature on the use of nitrous oxide as a potential alternate mild oxidizing agent for ODH.^6,17–19 Recently, N₂O was used for the ODH of the lower alkanes, ethane and propane. López Nieto et al.¹⁷ used N₂O for the oxidative dehydrogenation of n-butane and found that selectivity to olefins was higher with N₂O as the oxidant than with molecular oxygen. Kustrowski et al.¹⁸ studied the coupling of N₂O decomposition with EB dehydrogenation and obtained promising results with respect to the EB and N₂O conversions. N₂O is also utilized for propane ODH with enhanced selectivity towards propene for the same degree of propane conversion compared to O₂.¹⁹

The N₂O usage is of additional interest because the gas, which causes global warming, is effectively utilized in this chemical reaction. N₂O, which is a byproduct in industrial plants mainly from the production of adipic and nitric acids, has an average lifetime of about 150 years and a net greenhouse effect of about 300 times greater than that of CO₂.²⁰

A number of metal oxide catalysts were reported for the EB ODH in previous literature.^21–29 The catalytic activity of transition metal oxides is proposed to be dependent on the redox properties of these oxides. Among the transition metal oxides, vanadium–magnesium oxide (V–Mg–O) is reported as one of the most active and selective catalysts for ODH of EB to styrene.^26–28 Various physicochemical techniques were used to investigate the nature and structure of active species of V–Mg–O catalysts.^{3,5,6,30–43} For the ODH of propane and n-butane, isolated VO₄ tetrahedra in both amorphous and crystalline magnesium orthovanadate^{17,31,33,35,37,43} and dimeric V–O–V like species in magnesium pyrovanadate³⁴ were proposed as active phases and the nature of V–O bonds has been suggested to be the primary factor that determines the selectivity.^33,37,38 For the ODH of EB, the proposed active species also include O=V^V–O–V^IV species,²⁷ associated as opposed to isolated vanadium complexes.²⁸ Chang et al. examined the activities of single phase magnesium vanadates for ODH of EB and proposed that Mg₃(VO₄)₂ is most likely the active phase responsible for the high activity and selectivity of the V–Mg–O catalyst.³

There has been no report, to our knowledge, about the ODH of EB to styrene using N₂O as an oxidant with the V–Mg–O catalyst. In this paper, we report the utilization of N₂O for the dehydrogenation of ethylbenzene to styrene selectively over V–Mg–O catalysts. The structure and properties of the catalysts, prepared by the wet impregnation method, were investigated by XRD, UV-visible, FTIR, Laser Raman, temperature programmed reduction, ⁵¹V NMR and correlated with the activity of the catalysts.

2. Catalyst characterization

The vanadium content, determined by ICP analysis, BET surface areas and VO_x surface densities of the calcined V–Mg–O catalysts are presented in Table 1. The BET surface area of the MgO support was 109 m² g⁻¹ and decreased with increasing vanadia loading. The XRD patterns of the V–Mg–O catalysts calcined at 823 K (Fig. 1) show peaks centred at 43° and 62° corresponding to the (200) and (220) diffraction lines of MgO (JCPDS 4-829). No vanadium containing phases were detected for catalysts with vanadia loading up to 20 wt%, thus indicating a high dispersion of vanadium or very small crystallite sizes. When the loading was increased to 30 wt%, peaks at about 2θ = 35° characteristic of magnesium orthovanadate, Mg₃(VO₄)₂ (JCPDS 37-351), were clearly developed in the XRD pattern. The peaks attributable to V₂O₅ were not detected at any of the vanadia loadings. The volume averaged particle sizes corresponding to different catalysts were calculated using the Debye–Scherrer equation considering the most intense peak of MgO (d = 2.11). The particle sizes are around 10–12 nm and don't vary significantly with vanadia loading (Table 1).

Table 1 Physico-chemical properties of the V–Mg–O catalysts

Sample	V content^a (wt%)	Surface area/m^2 g^−1	VOx surface density/V nm^−2	Particle size/nm	Absorption edge energy/eV	H2–TPR Tmax/K
a Actual vanadium content obtained by ICP.
MgO	0	109	0	—	—	—
2 V–Mg–O	1.39	161	1.02	10.87	3.63	791
5 V–Mg–O	2.69	147	2.16	10.88	3.62	821
10 V–Mg–O	6.07	117	6.13	11.46	3.60	867
20 V–Mg–O	8.21	94	10.33	11.35	3.22	893
30 V–Mg–O	15.09	61	29.25	10.33	3.16	925

---

Fig. 1 Powder X-ray diffraction patterns of calcined V–Mg–O catalysts as a function of vanadia loading in wt%. Only the major peaks are indicated in the diagram.

FTIR spectra, recorded at ambient conditions, of the calcined V–Mg–O catalysts are shown in Fig. 2. The spectra of all catalysts showed a band around 860 cm⁻¹. This band is very prominent and broad for 30 V–Mg–O (the different samples are designated as n V–Mg–O where n represents the vanadia loading on MgO as V₂O₅). A broad band centred at 680 cm⁻¹ is also observed for 30 V–Mg–O. From the investigation of IR spectra of magnesium vanadate reference phases, the band at 860 cm⁻¹ can be assigned to the ν₃ antisymmetric stretch of (VO₄)³⁻ anions and the band at 680 cm⁻¹ to ν_as (VOV).^{27,30,34,46,47} These bands are reported as characteristic of magnesium orthovanadate, Mg₃(VO₄)₂. Mg₂V₂O₇ and MgV₂O₆ have absorption bands in the region 500–600 cm⁻¹ that are absent in the orthovanadate spectrum.²⁷ No absorption bands were observed in this region for the present catalysts, thus excluding the presence of meta and pyrovanadates. The spectra did not show any evidence for V₂O₅ formation since the band at 1020 cm⁻¹, which is usually assigned to the stretching vibration of V⁵⁺=O species, was not observed for any of the catalysts.

Fig. 2 FTIR spectra of calcined V–Mg–O catalysts, as a function of vanadia loading in wt%. The peak at 860 cm⁻¹ is marked in the diagram.

The coordination, polymerization degree and/or the oxidation state of vanadium ions influence oxygen to vanadium charge transfer bands in the diffuse reflectance UV-visible spectra of vanadium ions. In general, the lower the coordination and polymerization of vanadium, the lower is the wavelength.^47–49 The absorption band at 400–500 nm is usually considered as characteristic of V⁵⁺ in an octahedral environment, while the band at 250–350 nm is due to V⁵⁺ in a tetrahedral environment. The d–d transitions of V⁴⁺ (d¹) in a distorted octahedral symmetry give a broad band in the 550–850 nm region.

The diffuse reflectance UV-visible spectra of calcined V–Mg–O catalysts are presented in Fig. 3. The spectra show intense absorptions around 270 nm and the maximum of this absorption shifts slightly to higher wavelengths with increasing vanadia loading. The absorptions are broader in nature, especially for 20 and 30 V–Mg–O catalysts. The band around 270 nm can be assigned to the isolated VO₄ tetrahedron associated with the presence of Mg₃(VO₄)₂ since the oxygen to vanadium CT transition is usually observed in this region for Mg₃(VO₄)₂.⁴⁸ The UV-visible spectra of NH₄VO₃ and V₂O₅ are also shown in Fig. 3 for comparison. NH₄VO₃, characterized by the presence of polymeric tetrahedral VO₄, possess an absorption band centred at 291 nm while V₂O₅, having distorted tetragonal pyramidal V⁵⁺ species, show prominent bands at 336 and 472 nm. The band at 472 nm has been tentatively assigned to charge transfer to a delocalized acceptor site such as conjugated sites in octahedral VO₆ chains and the bands at lower wavelengths to charge-transfer transitions.⁴⁸

Fig. 3 Diffuse reflectance UV-visible spectra of (A) calcined V–Mg–O catalysts as a function of vanadia loading in wt% (B) pure bulk ammonium metavanadate and vanadium pentoxide.

The structural information can be deduced in a more quantitative manner, using absorption band edge energies, obtained using Tauc's law.^50,51 The edge energies for the allowed transitions were determined by finding the intercept of the straight line in the low-energy rise of a plot of [F(R_α)hν]² against hν, where F(R_α) is the Kubelka–Munk function and hν is the incident photon energy.^51–53 The edge energy values reported in the literature for the reference compounds Mg₃(VO₄)₂, Mg₂V₂O₇, and MgV₂O₆, with isolated, dimerised and two-dimensionally polymerized vanadium(V) species, are around 3.4, 3.2 and 2.4 eV respectively.^52,53 The edge energy of bulk V₂O₅ having distorted tetragonal pyramidal coordination of vanadium sharing edges and corners, which can be regarded as polymerized VO₅ units, is reported as 2.2–2.4 eV.^52,53 The absorption band edge energies determined for the V–Mg–O catalysts in the present study are given in Table 1. The edge energies fall in the narrow range of 3.2–3.6, suggesting that the molecular structure of the surface vanadium oxide is similar, though the vanadium loading varies. The closeness of edge energy values to that of magnesium orthovanadate suggests that the structure of surface vanadium species in the present series of catalysts is essentially similar to that of orthovanadate, composed of isolated, tetrahedrally coordinated vanadium(V) species.^43,52 The decrease in edge energy with increasing vanadia loading may be indicative of the increase in size of orthovanadate domains.⁵¹ The edge energy of magnesium pyrovanadate is also in the similar range, however, other characterization techniques exclude its presence.

Fig. 4 shows the TPR patterns of calcined V–Mg–O catalysts. The reduction peak is observed at 791 K for 2 V–Mg–O and the peak maximum shifts to higher temperatures with increasing vanadia loading (Table 1). This peak can be attributed to the reduction of V⁵⁺ in Mg₃(VO₄)₂.^43,54,55 The shift in T_max to higher temperatures with increasing vanadia loading may be due to the increased difficulty in reduction as the size of the Mg₃(VO₄)₂ domains increases. Similar shift in T_max has been reported previously and has been ascribed to the increase in the bulk-like character of Mg₃(VO₄)₂ domains with the increasing vanadia loading, since the bulk orthovanadate is more difficult to reduce.⁴³

Fig. 4 Temperature programmed reduction profiles of calcined V–Mg–O catalysts as a function of vanadia loading. 10% H₂–Ar; heating rate, 5 K min⁻¹; flow rate, 40 ml min⁻¹.

Fig. 5 shows the laser Raman spectra of V–Mg–O catalysts with different vanadia loadings. No characteristic peaks were observed for 2 V–Mg–O while the catalysts with 5–30 wt% vanadia loading showed well-defined bands. For the 5 V–Mg–O catalyst, the band was observed at around 834 cm⁻¹ which can be assigned to V–O–V or V–O–Mg vibrations of tetrahedrally coordinated VO₄ species.^43,56 The maximum of the band shifted to around 856 cm⁻¹ for higher vanadia loadings. The intensity of this band also increased with increasing vanadia loading. The position of the band agrees well with that reported for magnesium orthovanadate, Mg₃(VO₄).^{39,43,47,57–63} The band is relatively narrow for 30 V–Mg–O compared to catalysts with lower vanadia loadings. The band broadening at lower loadings could probably be due to the distortion of the crystalline network or due to the high dispersion of the vanadate phase.^63,64 The peaks characteristic of isolated vanadyl species (∼1000 cm⁻¹), Mg₂V₂O₇ (902 cm⁻¹) or V₂O₅ were not observed at any vanadia loading.

Fig. 5 Laser Raman spectra of calcined V–Mg–O catalysts as a function of vanadia loading.

The ⁵¹V NMR spectra of the magnesium vanadates were reported to be of different nature depending on the local environment of vanadium atoms in individual vanadates.^64,65Magnesium orthovanadate, which is composed of isolated nearly tetrahedral VO₄ species, gives an isotropic NMR spectrum. In pyrovanadates, the nonequivalent vanadium atoms in V₂O₇²⁻ fragments lead to spectra consisting of two different lines. In the case of metavanadate, the vanadium atom is in a strongly distorted octahedral environment with respect to oxygen and the spectrum of this compound has three-dimensional anisotropy of the chemical shift tensor. Fig. 6 shows the ⁵¹V MAS NMR spectra of the present series of catalysts. The spectra of all the samples show a single narrow intense peak; the isotropic chemical shift values, determined by spinning at two different rates, are around −556 ppm. This peak can be assigned to the presence of the Mg₃(VO₄)₂ phase since the isotropic nature and chemical shift of the peak resemble that of the bulk Mg₃(VO₄)₂ reference compound.^34,43,64,65

Fig. 6 ⁵¹V NMR spectra of calcined V–Mg–O catalysts as a function of vanadia loading.

In the V₂O₅–MgO system, four magnesium vanadates (orthovanadate, metavanadate and pyrovanadates) are stable at room temperature. The structural characterization discussed above suggests that, among these vanadates, only orthovanadate domains were formed on the surface of present catalysts and the size of these domains increases with increasing vanadia loading, finally leading to the bulk-like Mg₃(VO₄)₂ on 30 V–Mg–O. The presence of orthovanadate domains was supported by the appearance of FTIR bands at 860 cm⁻¹, UV-visible absorption bands in the region of 270 nm, absorption edge energy values in the range of 3.2–3.6 eV, TPR peaks between 784 and 925 K, laser Raman bands in the range of 825 and 860 cm⁻¹ and ⁵¹V NMR resonance around −556 ppm. At low coverages, XRD doesn't detect any vanadium phase because of high dispersion while the catalyst 30 V–Mg–O shows peaks characteristic of Mg₃(VO₄)₂ indicating the formation of bulk-like orthovanadate at this loading. There is no evidence for V₂O₅ domains by any of the characterization techniques.

3. Catalytic activity

Table 2 shows the styrene yields observed at 773 K. High styrene selectivity (∼97%) was observed with all V–Mg–O catalysts (only the data with the 30 V–Mg–O catalyst is shown in the Table). The styrene yield in the presence of nitrous oxide is at least 5 times higher than that in N₂. The styrene yield decreased with time on stream for both N₂O and N₂ atmospheres, however it was considerably higher in N₂O atmosphere throughout the period studied. The combined selectivity for benzene and toluene was lower in a nitrous oxide atmosphere, indicating a lower rate of dealkylation of ethylbenzene.

Table 2 Dehydrogenation of ethylbenzene under nitrous oxide and nitrogen flows over a V–Mg–O catalyst^a

N2O atmosphere						N2 atmosphere
TOS/h	EB conversion (%)	Styrene selectivity (%)	Styrene yield (%)	(Benzene + toluene) selectivity (%)	Styrene oxide selectivity (%)	EB conversion (%)	Styrene selectivity (%)	Styrene yield (%)	(Benzene + toluene) selectivity (%)	Styrene oxide selectivity (%)
a Reaction conditions: catalyst, 30 V–Mg–O, temperature = 773 K and N2O/EB = 15 (molar ratio).
2	65.3	97.2	63.5	2.1	0.7	11.4	78.6	9.0	21.2	0.6
3	64.5	97.7	63.0	2.1	0.2	10.5	78.4	8.3	21.5	0.8
4	58.4	97.4	56.8	2.0	0.6	9.8	82.3	8.1	17.7	0.5

---

The effect of reaction temperature on the dehydrogenation of ethylbenzene for a 30 V–Mg–O catalyst at a reaction time of 3 h is shown in Fig. 7. The EB conversion is low at 623 K, being 9.6%, and increased to 64.5% at 773 K. The styrene yield also increased with increasing reaction temperature. The dealkylation of ethylbenzene leading to the formation of benzene and toluene increased slightly with increasing temperature.

Fig. 7 Effect of reaction temperature on conversion of ethylbenzene, styrene yield and selectivities of products. Catalyst, 30 V–Mg–O; reaction time, 3 h; N₂O/EB = 15 (molar ratio); (●) ethylbenzene conversion; (■) styrene selectivity (▲) styrene yield; (★) sum of benzene and toluene selectivities; (◆) styrene oxide selectivity.

Fig. 8 shows the effect of space velocity on the EB dehydrogenation for 10 V–Mg–O at 723 K. The feed rate of EB was changed in these experiments, keeping the amount of catalyst constant, to vary the space velocity. With increasing space velocity, EB conversion and styrene yield decreased. The selectivity of styrene oxide increased while the selectivities of benzene and toluene were decreased slightly. The total yield of by-products, benzene, toluene and styrene oxide did not exceed 4% at all the space velocities studied.

Fig. 8 Effect of variation of space velocity on the conversion of ethylbenzene, styrene yield and selectivities of products. Catalyst, 10 V–Mg–O; reaction temperature, 723 K; reaction time, 3 h; N₂O flow rate (GHSV), 1200 h⁻¹; (●) ethylbenzene conversion; (■) styrene selectivity (▲) styrene yield; (★) sum of benzene and toluene selectivities; (◆) styrene oxide selectivity.

Fig. 9 shows the initial rates of ODH of EB per vanadium atom at 623 K plotted as a function of vanadia loading. The turn-over rate increased with increasing vanadia loading up to 10 wt% (surface density from 1 to 6 V nm⁻²) and then decreased with further increase in vanadia loading. Since the catalysts contain magnesium orthovanadate as the vanadium-containing phase, the difference in activity may be attributed to the size of the orthovanadate domains, which varies as a function of vanadia loading. A similar trend of activity with surface density was observed for the ODH of propane on V–Mg–O, VO_x/ZrO₂ and VO_x/Al₂O₃ catalysts.^44,45 It is proposed for these catalysts that the decrease in specific activity after reaching the maximum is due to the multilayer formation of the active form of vanadia which prevents the accessibility of V atoms lying below the surface. This is the most probable explanation in the present case also. The specific activity at 623 K, i.e. conversion of ethylbenzene per unit surface area of the catalysts, increased rapidly with increasing vanadia loading up to 10 wt% (surface density up to about 6 V nm⁻²), then decreased and remained constant with further increase in vanadia loading (Fig. 9). This confirms that formation of larger Mg₃(VO₄)₂ crystallites occurs when the surface density exceeds 6 V nm⁻² due to which all the vanadium atoms are not accessible for the reactants. This is similar to the trend observed earlier for propane ODH using V–Mg–O catalysts.⁴³ Hence it can be concluded that activity of the catalysts studied here is due to the presence of very small domains of Mg₃(VO₄)₂ on the surface of MgO rather than crystallites of bulk Mg₃(VO₄)₂.

Fig. 9 Effect of vanadia loading on the initial rate of ethylbenzene conversion normalized per V atom (▲), per unit surface area of the catalyst (●) and the selectivity of styrene at a 10% conversion level (■). Reaction temperature, 623 K.

The previous reports mentioned that the Mg₃(VO₄)₂ phase is mainly responsible for the activity and selectivity for EB dehydrogenation, in the presence of oxygen³ and CO₂.¹² Moreover, the most active species for propane ODH is also suggested as VO²⁻₄ species, existing in a surface structure that is essentially identical to that present in Mg₃(VO₄)₂, for V–Mg–O catalysts prepared by conventional and thermolytic methods.^{17,31,33,35,37,43}

Fig. 9 also shows the styrene selectivity at an EB conversion level of 10% as a function of the vanadia loading. The selectivity to styrene increased with vanadia loading up to 10 wt% (6 V nm⁻²), after which it decreased. The observed trend in selectivity with the apparent surface density of VO_x is very similar to that reported previously for vanadia dispersed on MgO, ZrO₂ and Al₂O₃.^44,45 Reaction data for prolonged runs show that the catalyst does not undergo deactivation rapidly and exhibits a stabilized performance (Fig. 10).

Fig. 10 Effect of time on stream on the EB conversion to study the catalyst stability. Catalyst, 30 V–Mg–O, temperature = 773 K and N₂O/EB = 15 (molar ratio).

4. Characterization of spent catalysts

The XRD patterns of the catalyst 30 V–Mg–O after dehydrogenation reaction at 773 K for 4 h in N₂O and N₂ atmospheres are shown in Fig. 11. The pattern of the sample after dehydrogenation reaction in a N₂O atmosphere shows peaks characteristic of magnesium orthovanadate, though the intensity of these peaks reduced considerably compared to that of fresh catalyst. This might be due to some sort of restructuring on the catalyst surface, whereby the large particles of magnesium orthovanadate are decomposed into small particles that cannot be detected by XRD. However, magnesium orthovanadate peaks completely disappeared after the dehydrogenation reaction in a N₂ atmosphere.

Fig. 11 Powder X-ray diffraction patterns of 30 V–Mg–O after (a) dehydrogenation reaction under N₂O flow at 773 K for 4 h, (b) dehydrogenation reaction under N₂ flow at 773 K for 4 h.

The diffuse reflectance UV-visible spectra of 30 V–Mg–O catalysts after various treatments are shown in Fig. 12. After EB dehydrogenation in N₂, the catalyst showed a new broad and weak band with absorption maximum between 500 and 700 nm, which is characteristic of the V⁴⁺ d–d transitions (Fig. 12b). This band can be observed prominently after the reduction of the fresh 30 V–Mg–O catalyst with H₂ at 773 K (Fig. 12c). Hence a higher extent of reduction of vanadium species occurs during the reaction in a N₂ atmosphere. The reduction during the reaction is reversible since the band at 270 nm reappeared and the band due to V⁴⁺ disappeared when the reduced catalyst is treated in a N₂O atmosphere at 773 K for 4 h.

Fig. 12 Diffuse reflectance UV-vis spectra of 30 V–Mg–O after (a) dehydrogenation reaction under N₂O flow at 773 K for 4 h, (b) dehydrogenation reaction under N₂ flow at 773 K for 4 h, (c) after reduction with H₂ at 773 K for 4 h.

Fig. 13 shows the FTIR spectra of the 30 V–Mg–O catalyst after different treatments. The intensity of the band at 860 cm⁻¹, characteristic of Mg₃(VO₄)₂, was decreased after dehydrogenation reaction in a N₂O atmosphere (Fig. 13a) while the band disappeared after reaction in a N₂ atmosphere (Fig. 13b). This again shows a greater extent of reduction for the catalyst when the reaction is carried out in a N₂ atmosphere. The band disappeared after treating 30 V–Mg–O in H₂ at 773 K and reappeared when the reduced catalyst was treated in a N₂O atmosphere at 773 K for 4 h.

Fig. 13 FTIR spectra of 30 V–Mg–O after (a) dehydrogenation reaction under N₂O flow at 773 K for 4 h, (b) dehydrogenation reaction under N₂ flow at 773 K for 4 h, (c) after reduction with H₂ at 773 K for 4 h, (d) after treating the reduced catalyst under N₂O flow at 773 K for 4 h.

The characterization of catalysts after reaction reveals that the vanadium species is continuously reduced when the dehydrogenation is carried out in a N₂ atmosphere. Though reduction of vanadium species occurs during the reaction in a N₂O atmosphere also, the vanadium species is at a higher oxidation state in this case owing to the mild oxidizing property of N₂O. The higher styrene yield in N₂O than that in a N₂ atmosphere can be ascribed to the ability of N₂O to reoxidise vanadium that got reduced during the reaction, though the complete reoxidation was not observed. This is similar to the effect reported by Sakurai et al. for the dehydrogenation of EB with CO₂ in which the vanadium species was kept at a higher oxidation state during the reaction due to the oxidation capability of CO₂.¹²

5. Conclusions

The catalytic properties of V–Mg–O catalysts with different vanadia loadings were evaluated for the oxidative dehydrogenation of EB in the presence of N₂O. Structural characterization of these catalysts shows that magnesium orthovanadate domains are present at all vanadia loadings. With increasing vanadia loading, the size of orthovanadate domains increases, finally leading to the appearance of bulk-like Mg₃(VO₄)₂ on 30 V–Mg–O. The rate of EB conversion per V atom increases with increasing vanadia loading, reaches a maximum at 10 wt% and decreases when the loading is increased further. Consideration of the structural characterization data together with catalytic activity indicates that the surfaces of most active catalysts possess small domains of Mg₃(VO₄)₂. The formation of bulk-like orthovanadate on the surface at high loadings leads to the decrease in activity since the fraction of V atoms that lie below the catalyst surface is not accessible for the reactants. The styrene yield was much higher in N₂O than that in a N₂ atmosphere, since the surface vanadium species were kept in a high oxidation state when N₂O was used for the reaction, as revealed by the characterization of used catalysts. Prolonged runs show that the EB conversion remains stable. If a process can be developed using N₂O as an oxidant, it is an effective way to avoid some of the green-house gases, especially since the by-product of decomposition is N₂. Solutia has developed a process for the direct oxidation of benzene to phenol using N₂O, coming out of adipic acid plants as a byproduct.⁶⁶

6. Experimental section

V–Mg–O catalysts were prepared by impregnating MgO with NH₄VO₃ dissolved in an aqueous solution of oxalic acid with a pH of ∼1.9. After impregnation, the samples were dried in air at 383 K for 12 h and calcined in air at 823 K for 4 h.

The vanadium content was determined by ICP analysis. The different samples are designated in the text as n V–Mg–O where n is an integer corresponding to the loading of vanadia on MgO as V₂O₅.

The BET surface areas of the samples, S_BET, were measured by nitrogen physisorption at 77 K with a NOVA 1200 surface area analyzer (Quantachrome). The catalyst samples were evacuated at 573 K for 3 h prior to N₂ physisorption measurements.

The temperature programmed reduction experiments were carried out with a Micromeritics Autochem 2910 catalyst characterization system, equipped with a TCD detector. Fresh, dried samples were pretreated by passing high purity (99.9%) argon (20 ml min⁻¹) at 773 K for 3 h. After cooling to ambient temperature, argon flow was replaced by a 10% H₂/Ar mixture. The catalyst samples were heated in this atmosphere to 1273 K at a heating rate of 5 K min⁻¹. The flow rate of the H₂/Ar mixture was kept at 40 ml min⁻¹ throughout the experiment. The water produced during the reduction was condensed in a cold trap immersed in a slurry of isopropanol and liquid nitrogen. The amount of catalyst in these experiments was adjusted so that samples contain equivalent amounts of V₂O₅ in all the experiments.

Powder X-ray diffraction (XRD) patterns were obtained on a Rigaku Miniflex diffractometer with a monochromated CuKα radiation (λ = 0.15406 nm, 30 kV, 15 mA). Samples were scanned within the 2θ range of 10°–85° with a step size of 0.02.

Fourier transform-infrared spectra of the samples were recorded on a Shimadzu 8300 FT-IR spectrometer under ambient conditions. The spectra were recorded using thin self-supporting discs made by pressing the mixture of catalyst samples and KBr.

Diffuse reflectance UV-visible spectra were recorded using a Shimadzu UV-2550 spectrophotometer. The reflectance spectra were converted into the Kubelka–Munk function, F(R), which is proportional to the absorption coefficient for low values of F(R). The spectra were measured in the range of 200–800 nm in air at room temperature.

Laser Raman spectra were obtained using a Spex 1403 spectrometer using an Ar ion laser with a 514.5 nm line (Spectra Physics 165) under ambient conditions. The laser power used was approximately 40 mW. The scattered light was collected in the backscattering geometry and detected with a thermoelectrically cooled photomultiplier (RCA, type C31034). Samples were pressed into wafers about 0.1 cm thick and 0.9 cm in diameter. Before recording the spectra, samples were heated in air at 473 K for 2 h and cooled to room temperature.

⁵¹V MAS NMR experiments were performed at 131.54 MHz on a Bruker DRX-500 NMR spectrometer with a Bruker broad-band CP-MAS probe. The samples were spun at 8, 10, or 12 kHz with 4 mm-diameter zirconia rotors. A pulse length of 3 μs and a recycle delay of 500 ms were used for acquisition of the data.

Catalytic reactions were carried out in a fixed bed down-flow glass reactor operated at atmospheric pressure. 1 g of catalyst was placed at the centre of the reactor using quartz wool plugs. EB was fed into the reactor using a high pressure syringe pump (Isco 500 D). N₂O was introduced into the reactor at controlled flow rates using a mass flow controller (Bronkhorst, type E-7600-AAA).

Condensed reaction products (styrene, benzene, toluene, styrene oxide) were analyzed using a Hewlett-Packard 6890 plus gas chromatograph equipped with a FID and a capillary column (BPX5, 50 m × 0.32 mm i.d.). Conversion and selectivity were calculated from the measured GC compositions and specific activities (in mol_EB h⁻¹m⁻²) and turnover rates (in mol_EB (mol V)⁻¹ S⁻¹) were calculated from conversion values of ethylbenzene (X_EB) using specific surface areas and vanadium content respectively.

Acknowledgements

We thank Ms Violet Samuel, Mr S. C. Jha, Mrs N. Jacob and Dr S. Umbarkar of the Catalysis Division, NCL, Pune, and Dr Rajmohanan of the NMR unit, NCL, Pune, for various characterization studies.

Notes and references

1. G. Emig and H. Hofmann, J. Catal., 1983, 84, 15.
2. G. E. Vrieland and P. G. Menon, Appl. Catal., 1991, 77, 1.
3. W. S. Chang, Y. Z. Chen and B. L. Yang, Appl. Catal., A, 1995, 124, 221.
4. F. Cavani and F. Trifiro, Appl. Catal., A, 1995, 133, 219.
5. W. Oganowski, W. Hanuza, H. Drulis, W. Mista and L. Macalik, Appl. Catal., A, 1996, 136, 143.
6. (a) N. R. Shiju, M. Anilkumar, S. P. Mirajkar, C. S. Gopinath, B. S. Rao and C. V. Satyanarayana, J. Catal., 2005, 230, 484; (b) J. Janas, J. Gurgul, R. P. Socha, J. Kowalska, K. Nowinska, T. Shishido, M. Che and S. Dzwigaj, J. Phys. Chem. C, 2009, 113, 13273; (c) P. M. Woods, B. Mirkelamoglu and U. S. Ozkan, J. Phys. Chem. C, 2009, 113, 10112; (d) J. Perez-Ramirez and A. Gallardo-Llamas, J. Phys. Chem. B, 2005, 109, 20529; (e) A. Gallardo-Llamas, C. Mirodatos and J. Perez-Ramirez, Ind. Eng. Chem. Res., 2005, 44, 455; (f) J. Kowalska-Kus, A. Held and K. Nowinska, Catal. Lett., 2010, 136, 199; (g) E. V. Kondratenko and J. Perez-Ramirez, Appl. Catal., A, 2004, 267, 181.
7. K. K. Kearby, in Catalysis, ed. P. H. Emmett, Reinhold, New York, 1955, vol. 3, ch.10.
8. E. H. Lee, Catal. Rev. Sci. Eng., 1973, 8, 285.
9. M. Sugino, H. Shimada, T. Turuda, H. Miura, N. Ikenaga and T. Suzuki, Appl. Catal., A, 1995, 121, 125.
10. T. Badstube, H. Papp, P. Kustrowski and R. Dziembaj, Catal. Lett., 1998, 55, 169.
11. Y. Sakurai, T. Suzaki, K. Nakagawa, N. Ikenaga, H. Aota and T. Suzuki, J. Catal., 2002, 209, 16.
12. Y. Sakurai, T. Suzaki, N. Ikenaga and T. Suzuki, Appl. Catal., A, 2000, 192, 281.
13. N. Ikenaga, T. Tsuruda, K. Senma, T. Yamaguchi, Y. Sakurai and T. Suzuki, Ind. Eng. Chem. Res., 2000, 39, 1228.
14. S. E. Park, J.-S. Chang, V. P. Vislovskiy, M. S. Park, K. Y. Lee and J. S. Yoo, 6th international conference on carbon dioxide utilization, Breckenridge, CO, USA, 2001, Preprinted Abstracts, p. 60.
15. C. R. Adams and T. J. Jennings, J. Catal., 1970, 17, 157.
16. R. Dziembaj, P. Kustrowski, T. Badstube and H. Papp, Top. Catal., 2000, 11/12, 317.
17. J. M. López Nieto, A. Dejoz, M. I. Vazquez, W. O'Leary and J. Cunningham, Catal. Today, 1998, 40, 215.
18. P. Kustrowski, M. Zbroja, R. Dziembaj and H. Papp, Catal. Lett., 2002, 80, 1.
19. E. V. Kondratenko and M. Baerns, Appl. Catal., A, 2001, 222, 133.
20. S. Kannan, Appl. Clay Sci., 1998, 13, 347.
21. A. Cortes and J. L. Seoane, J. Catal., 1974, 34, 7.
22. W. Kania and K. Jurczyk, Appl. Catal., 1990, 61, 35.
23. J. Hanuza, B. Jezowska-Trzebiatowska and W. Oganowski, J. Mol. Catal., 1978, 4, 271.
24. Z. Dziewiecki and A. Makowski, React. Kinet. Catal. Lett., 1980, 13, 51.
25. R. J. Rennard, R. A. Innes and H. E. Swift, J. Catal., 1973, 30, 128.
26. G. V. Shakhnovich, I. P. Belomestnykh, N. V. Nekrasov, M. M. Kostyukovsky and S. L. Kiperman, Appl. Catal., 1984, 12, 23.
27. J. Hanuza, B. Jezowska-Trzebiatowska and W. Oganowski, J. Mol. Catal., 1985, 29, 109.
28. A. V. Simakov, N. N. Sazonova, S. A. Ven'yaminov, I. P. Belomestnykh, N. N. Rozhdestvenskaya and G. V. Isagulyants, Kinet. Catal., 1989, 30, 598.
29. J. J. Kim and S. W. Weller, Appl. Catal., 1987, 33, 15.
30. V. Sharma, A. Wokaun and A. Baiker, J. Phys. Chem., 1986, 90, 2715.
31. M. A. Chaar, D. Patel, M. C. Kung and H. H. Kung, J. Catal., 1987, 105, 483.
32. M. A. Chaar, D. Patel and H. H. Kung, J. Catal., 1988, 109, 463.
33. D. Patel, M. C. Kung and H. H. Kung, in Proc. 9th Int. Congr. Catal., Calgary, ed. M. J. Phillips and M. Ternan, The Chemical Institute of Canada, Ottawa, 2nd edn, 1988, vol. 4, p. 1554.
34. D. S. H. Sam, V. Soenen and J. C. Volta, J. Catal., 1990, 123, 417.
35. A. Corma, J. M. López Nieto, N. Paredes, A. Dejoz and I. Vasquez, in New Developments in Selective Oxidation II, ed. C. C. Corberán and S. V. Bellón, Studies in Surface Science and Catalysis, Elsevier, Amsterdam, 1994, vol. 82, p. 113.
36. D. Patel, P. J. Andersen and H. H. Kung, J. Catal., 1990, 125, 132.
37. M. C. Kung and H. H. Kung, J. Catal., 1992, 134, 668.
38. P. M. Michalakos, M. C. Kung, I. Jahan and H. H. Kung, J. Catal., 1993, 140, 226.
39. A. Corma, J. M. López Nieto and N. Paredes, J. Catal., 1993, 144, 425.
40. V. Soenen, J. M. Herrman and J. C. Volta, J. Catal., 1996, 159, 410.
41. H. H. Kung and M. C. Kung, Appl. Catal., A, 1997, 157, 105.
42. A. Burrows, C. J. Kiely, J. Perregaard, P. E. Hojlund-Nielson, G. Vorbeck, J. J. Calvino and C. Lopez-Cartes, Catal. Lett., 1999, 57, 121.
43. C. Pak, A. T. Bell and T. D. Tilley, J. Catal., 2002, 206, 49.
44. A. Khodakov, J. Yang, S. Su, E. Iglesia and A. T. Bell, J. Catal., 1998, 177, 343.
45. A. Khodakov, B. Olthof, A. T. Bell and E. Iglesia, J. Catal., 1999, 181, 205.
46. A. A. Rar, A. V. Simakov and S. A. Veniaminovv, React. Kinet. Catal. Lett., 1989, 39, 299.
47. A. Corma, J. M. López Nieto and N. Paredes, Appl. Catal., A, 1993, 104, 161.
48. G. Centi, S. Perathoner, F. Trifiro, A. Aboukais, C. F. Aissi and M. Guelton, J. Phys. Chem., 1992, 96, 2617.
49. G. Lischke, W. Hanke, H.-G. Jerchkewitz and G. Olmann, J. Catal., 1985, 91, 54.
50. J. Tauc, in Amorphous and Liquid semiconductors, ed. J. Tauc, Plenum, London, 1974.
51. R. S. Weber, J. Catal., 1995, 151, 470.
52. X. Gao, S. R. Bare, B. M. Weckhuysen and I. E. Wachs, J. Phys. Chem. B, 1998, 102, 10842.
53. A. M. A. Cruz and J. G. Eon, Appl. Catal., A, 1998, 167, 203.
54. F. Arena, F. Frusteri and A. Parmaliana, Catal. Lett., 1999, 60, 59.
55. F. Arena, F. Frusteri and A. Parmaliana, Appl. Catal., A, 1999, 176, 189.
56. G. Deo and I. E. Wachs, J. Phys. Chem., 1991, 95, 5889.
57. A. Corma, J. M. López Nieto, N. Paredes and M. Perez, Appl. Catal., A, 1993, 97, 159.
58. M. L. Occelli and J. M. Stencel, ACS Symp. Ser., 1988, 375, 195.
59. P. Concepción, J. M. López Nieto and J. Perez-Pariente, J. Mol. Catal. A: Chem., 1995, 97, 173.
60. G. Busca, G. Ricchiardi, D. S. H. Sam and S. C. Volta, J. Chem. Soc., Faraday Trans., 1994, 90, 1161.
61. X. Gao, P. Ruiz, Q. Xin, X. Guo and B. Delmon, J. Catal., 1994, 148, 56.
62. X. Gao, P. Ruiz, Q. Xin, X. Guo and B. Delmon, Catal. Lett., 1994, 23, 321.
63. R. X. Valenzuela and V. C. Corberán, Top. Catal., 2000, 11/12, 153.
64. M. L. Occelli, R. S. Maxwell and H. Eckert, J. Catal., 1992, 137, 36.
65. O. B. Lapina, A. V. Simakov, V. M. Mastikhin, S. A. Veniaminov and A. A. Shubin, J. Mol. Catal., 1989, 50, 55.
66. G. Bellussi and C. Perego, CATTECH, 2000, 4, 4.

---