The original material of vanadium oxide was obtained from Alpha chemicals at a purity of 99.99%. In order to prepare the vanadium metal carbide crystallites, the reactor used were made of a quartz tube with a cooling water jacket.

We have synthesized the eight different vanadium carbide crystallites based on the general statistical method as shown in Table 1. These vanadium carbide materials were prepared to scrutinize the influence of the molar hourly space velocity, the heating rate, and the reaction gas composition on the synthesis of vanadium carbide crystallites.

In this case, MHSV is the ratio of the methane molar flow rate to the moles of corresponding transition metal oxides (here, vanadium oxides in this study).

The experimental procedures are as follows. First of all, the furnace was heated to 1,073 K in 90 minutes. And then, it was raised to 1,313 K and was held for 1 hour. After this heating, the reactor was then cooled rapidly by removing it from the furnace. When it reached room temperature, the gas was switched to He for 30 minutes, then to 1% O₂/He for a 2 hour passivation period. The vanadium carbide crystallite was then collected and characterized. The other transition metal nitrides (VN) were also prepared using the similar methods.

A Quantasorb^Ⓡ machine has been used to measure of BET surface areas. And then, the attenuation setting of the Quantasorb^Ⓡ machine was used to give a peak height of approximately 60-90% of full scale. A peak height and count were determined by the Quantasorb^Ⓡ machine. Using these procedures, the peak height and the count were determined. Finally, the catalyst sample was removed from the machine and weighed.

We have also performed the oxygen uptake experiment by using the same Quantasorb^Ⓡ machine. These O₂ uptake experiments were then carried out at 351 K. The Quantasorb^Ⓡ machine attenuation was set at 2 and 0.5 cc injections of 9.98%, then O₂/He gas were made. The injections were repeated until the machine exhibited a constant count for about 5 straight injections. And then, we removed the sample from the machine and weighed.

Usually, we have loaded 0.2 g of vanadium carbide crystallites in the reactor to measure the reactivities of ammonia decomposition reaction. The reduction of the given sample was performed using H₂ throughout the course of temperature rise, starting from room temperature to 673 K at a rate of 0.033 K/s. After this, it was held at 673 K for 14 hours before being cooled to the reaction temperature before measuring reactivity. In most cases, the reactivities were obtained at temperatures between 623 and 823 K. We have analyzed the reactor effluent using an on-line gas chromatograph (SD6200) equipped with both flame ionization and thermal conductivity detectors. The products were separated with Porapak Q packed columns (80/100,8'x1/8", CRS) which was connected to a gas chromatography detector.

In the present work, the eight vanadium carbides, vanadium nitride, and platinum supported on carbon have been synthesized for comparison. It is generally known that in the synthesis of transition metal carbide crystallites, the most important factor is the gas dispersion through the plug of solid material and heating rate. Therefore, we can prepare the materials with small crystallite size using large heating rates and low MHSVs. These might be related to the fast temperature rise. If we want to prepare carbides which have larger crystallite sizes, the opposite trends must be considered. As shown in Table 1, our goal is to react the material before the gas penetration and saturation of solid reactant occurs at the same time.

Here, in this study we have synthesized, the vanadium carbide crystallites using vanadium oxide and CH₄ gas or methane-hydrogen mixture and their BET surface areas ranged from 5.2 m²/g to 29.9 m²/g. For the case of the vanadium carbide crystallites synthesized in pure CH₄, we have obtained the highest specific surface areas (29.9 m²/g) with the low heating rate of 30 K/h and high space velocity of 80 hr^-1.

As we can see the sorption properties in Table 2, the highest surface areas were observed for the vanadium carbides synthesized in an CH₄ gas only. The influence of greater heating rate on surface area is that the reaction rate increased, as the conversion of vanadium oxide to vanadium carbide crystallites is increased. There are several similarities in the preparation of transition metal carbide crystallites including titanium and niobium carbide crystallites and previously published by the other research group^3,5,6).

We can also see the sorption properties over vanadium carbide crystallites in Table 2. It is generally accepted that the oxygen uptake was taken as a measure of the number of potential active sites on the vanadium carbide crystallites. We can also apply these results for other transition metal carbides (WC and MoC). As the different vanadium carbide crystallites vary, the different values of oxygen uptake were obtained in this case.

These results implied that the vanadium carbide has a properties that the reactivities is totally dependent on the synthesis conditions. By the same token, in the case of vanadium carbides the sorption properties of BET surface areas and oxygen uptakes are strongly relying on preparation conditions. It is general that the oxygen site densities were determined from the O₂ uptakes and the BET surface areas. For the case of vanadium carbides, the O₂ site density averages and 18.3×10⁷ O₂/m². Assuming a 1:1 V:O stoichiometry, surface coverages averaged ～16%. These values of site density and surface coverage for vanadium carbides are greater than those for tantalum and titanium carbides.

Fig. 1 also shows that there is a linear relationship between the oxygen uptake and BET surface area of vanadium carbides.

Fig. 1. 
Surface area and oxygen uptake for vanadium carbides

This result suggests that oxygen was a nonselective adsorbate to the vanadium carbide crystallites. The linear relationship between oxygen uptake and BET surface area in vanadium carbide crystallites is similar to that in other transition metal crystallites¹⁰⁾. We have considered that these kinds of findings are very important since the reaction properties of the transition metal carbide and nitride are related to their preparation conditions with the different sorption properties.

It has been shown that all of the vanadium carbide crystallites prepared in the present work were active for ammonia decomposition reaction. For the most of cases, the freshly prepared vanadium carbides exhibited the highest initial conversion. However, their activities were gradually decreased as times went by.

Here, as shown in Fig. 2, the NH₃ decomposition reaction rates for vanadium carbides were rapidly decreased to the steady-state activities and remained constant for several hours. Fig. 3 shows the influence of reaction temperature (at 600, 650, 750, 800, and 850 K) on the conversion. These results show clearly that the conversion increased with the increase of temperature. The maximum conversion was obtained at 700 K. However, no serious temperature effects were found after 750 K for vanadium carbide crystallites.

Fig. 2. 
NH₃ decomposition conversion vs. reaction temperature of VC

Fig. 3. 
NH₃ decomposition conversion vs. reaction temperature of WC

In Table 2, the steady state activities for all vanadium carbide crystallites were described. We can see that all of the vanadium carbide crystallites exhibited the linear relationship of surface area with reactivity of ammonia decomposition. Based on these results, it is believed that the surface area is linearly related to the reactive site of these vanadium carbide crystallites for ammonia decomposition.

These kinds of linear relationship between the surface area and the reactivity suggested the existence of structure sensitivity for the ammonia decomposition reaction in vanadium carbide crystallites.

In transition metal carbide crystallites, it was known that the concept of structure sensitivity frequently appeared in several catalytic reactions. Previous experimental results suggested that ammonia decomposition reaction was structure-sensitive by molybdenum and titanium carbide crystallites^11,12).

In the present study, amongst all the prepared vanadium carbide crystallites VC-7 had the highest steady state reactivity for NH₃ decomposition. From all of the synthesized vanadium carbides, the VC-7 crystallites was ~2.58 times more reactive than VC-6 which had the lowest activity. These results indicated that the reactive species in the VC-6 crystallites were different from that in the VC-7 crystallites. The similar results were observed that the molybdenum nitride crystallites which has high surface area exhibited the highest activity over pyridine hydrodenitrogenation¹³⁾.

Fig. 4 shows the XPS pictures of carbon species in the vanadium carbides. The C1 peak of carbidic carbon in Fig. 4 indicates that vanadium carbides prepared from vanadium oxides. The other two carbon peaks of C2 and C3 shows the free carbon and adsorbed carbon in the vanadium carbides surface, respectively. Therefore, depending upon the kinds of carbidic carbons, the final ammonia decomposition reactivity would be different.

Fig. 4. 
XPS pictures of carbon species in the vanadium carbides

Fig. 5 shows the XRD results of vanadium oxide (Fig. 4[a]) and carbides (Fig. 4[b]), respectively. Similarly, it is clear that vanadium oxides and carbides exhibit the different pictures of XRD, indicating the differences of bulk structure in these materials.

Fig. 5. 
XRD results for (a) vanadium oxides and (b) carbides

In summary, it is noted that for vanadium oxides and carbides, these kinds of physico-chemical properties should be related to the final reactivities for NH₃ decomposition.

The reactivities of all of the vanadium carbides are more active than that of vanadium nitrides as shown in Table 2. The reason for these differences might be ascribed to the electronic and structural differences. Because of the differences of electronegativities, it was assumed that the vanadium carbides contained a different electron character at the surface from the vanadium nitrides. Since the nitrogen is more electronnegative than carbon, vanadium carbides should be less acidic than vanadium nitrides. For ammonia decomposition reaction, the ammonia as a base adsorb more strongly on nitrides than on carbides. Based on this scenario, this difference might be allowing vanadium carbides to contain the higher activities. And also, as shown in Fig. 5, the crystallographic structure of vanadium carbides is different from that of vanadium oxides and nitrides. This structural difference also appeared to be involved in reactivity difference.

By the way, we also compared the reactivity of Pt/C with those of vanadium carbides. In the current study, the steady state reaction activities for vanadium carbides were comparable to or even higher than that of the Pt/C sample. According to these results, it was assumed that the characteristics of reactive sites in the vanadium carbide crystallites were similar to those in the group 8-10 metal based materials¹⁴⁾. Moreover, it is suggested based on these results that the vanadium carbide crystallites^15-17) can be utilized for various areas including many petroleum industries as the substitutes of the precious metal crystallites including Pt and Pd.