2.1 Catalyst preparation

2.2 Catalysts characterization

Powder X-ray diffraction patterns of the synthesized catalysts were recorded on a Ultima-IV X-ray diffractometer (M/s. Rigaku Corporation, Japan) using Ni filtered Cu Kα radiation (λ = 1.5406 A°) with 2θ scan speed of 2° min⁻¹ and in the 2θ range of 10–80° operated at 40 kV and 30 mA. Crystalline phases identified were compared with the reference data from ICDD files. N₂ adsorption–desorption studies were conducted at liquid N₂ temperature on a Quadrasorb-SI (M/s. Quantachrome Instruments, USA). Prior to the measurement, the samples were preheated at 373 K for 3 h in order to expel the interlayer water and physisorbed gases. The transition electron microscope images were acquired by using FEI-Talos-F200X apparatus (M/s. Thermo Fisher Scientifics, USA) operated at 200 kV. Prior to the TEM analysis, the samples were ultrasonicated in ethanol for 1 h and one drop of the solution was placed on a carbon-coated copper grid. The solvent was then evaporated in an air oven at 353 K for 6 h. TPR analysis were done on a homemade setup equipped with an SS reactor (i.d. = 6 mm, length = 300 mm) interfaced to a thermal conductivity detector (TCD) equipped gas chromatograph (M/s. Agilent 6890 series, USA). 50 mg of the catalyst sample was loaded at the isothermal zone of the reactor and heated by an electric furnace at a rate of 10 K min⁻¹ in helium gas flowing at a rate of 30 cm³ min⁻¹ for 1 h to facilitate the desorption of physically adsorbed water. After pretreatment the furnace temperature was cooled to 373 K and then helium gas was replaced with a reducing gas (5% H₂ in argon) at a flow rate of 30 cm³ min⁻¹ and the temperature was linearly increased to 1273 K at a ramping rate of 10 K min⁻¹ and maintained at the same temperature for 15 min. The H₂ consumption was monitored by standard GC software. X-ray photoelectron spectroscopy (XPS) analysis of the catalyst was carried out by a Kratos analytical axis supra spectrophotometer (M/s. Kratos Analytical Ltd, UK), with Mg Kα monochromatic excited radiation (1253.6 eV). The residual pressure in the analysis chamber was around 10⁻⁹ mbar.

2.3 Catalytic activity measurements

Catalytic activity studies were performed in a fixed bed down flow quartz reactor at atmospheric pressure. In each run, 1 g of the catalyst (crushed and sieved in to 60–80 mesh size) was loaded into the reactor with the same amount of quartz beads sandwiched between two plugs of quartz wool placed at the center of the reactor. Prior to the reaction, the catalyst was subjected to pre-treatment with N₂ flow for 0.5 h at 873 K. EB was introduced by a liquid feed pump (M/s. B. Braun, Germany) at a feed rate of 8.2 mmol h⁻¹. The dehydrogenation reaction was performed in the range of 723–923 K with 50 K interval. The reaction was tested with different gas environments of N₂, CO₂ and mixture of N₂ and CO₂ at a total gas flow of 30 cm³ min⁻¹. The overall carbon balance in the reaction is >98%, only carbon amount that was deposited on the catalyst as carbonaceous species is lacking from the carbon balance by GC. The products were collected at regular intervals in an ice cooled trap and analyzed on a flame ionization detector (FID) equipped gas chromatograph, GC-17A (M/s. Shimadzu Instruments, Japan) using a capillary column, OV1-G43 (30 m × 0.53 mm × 5.0 µm). The products were confirmed by GC–MS QP-2010 (M/s. Shimadzu Instruments, Japan) with EB-5MS capillary column (30 m × 0.25 mm × 0.25 µm). The Conversion and selectivity were calculated by the following equations.

• Conversion (%) = (Reactant_in − reactant_out)/(reactant_in) × 100

• Selectivity (%) = (Product_out)/(Reactant_in − reactant_out) × 100.

The major products observed in the analysis were styrene, toluene and benzene along with unreacted ethylbenzene.

3.1 Characterization results

In Fig. 1 wide angle XRD patterns of bulk catalyst (MN) exhibited three broad diffraction peaks at 2θ values of 37.6 (1 1 2), 43.03 (2 0 0) and 64.2 (2 0 4) which corresponds to face centered cubic structure of γ-Mo₂N phase (ICDD No.: 751150). 30MNA catalyst also showed the signals at the same 2θ values along with diffraction peaks of γ-Al₂O₃ at 2θ values of 37.5 (3 1 1), 45.6 (4 0 0), 60.5 (5 1 1) and 66.6 (4 4 0) (ICDD No.: 50-0741). These results are signifying that nitride phase in 30MNA catalyst is present in its crystalline form. But in the case of 20MNA and 10MNA catalysts, the diffraction peaks correspond to γ-Al₂O₃ the only detected and the peaks correspond to nitride were not observed. The reason expected to be due to the fact that the crystalline sizes of MN in 10MNA and 20MNA catalysts are below the detection limit of XRD or the nitride species may be highly dispersed over the surface of the support HSA (Logan et al., 1998).

Close

Fig. 1 Wide angle powder XRD patterns of (a) HSA, (b) 10MNA, (c) 20MNA, (d) 30MNA and (e) MN catalysts.

Export to PPT

Multipoint BET surface area and corresponding normalized BET surface area were presented in Table 1. The specific surface areas of the samples were calculated by using the standard BET theory. As expected the BET surface areas of the prepared catalysts relative to the support HSA are in decreasing fashion. Normalized BET surface areas gives an insight approach for the dispersion of the nitride species over the support surface and calculated using the following formula. $${\text{NS}}_{\text{BET}}={\text{S}}_{\text{BET}}\text{of}\text{catalyst}/[\left(1-\text{y}\right)\times {\text{S}}_{\text{BET}}\text{of}\text{support}]$$ where S_BET is the BET surface area of the catalyst and y is the weight fraction of the guest phases.

As per the equation, the NS_BET of the prepared catalysts are approaching to digit one, signifying that mainstream of nitride species are existing within the pore structure of the support and suggests homogeneous distribution of the catalyst species over the support (Gurram et al., 2016). No diffraction peaks corresponding to nitride phase (except in 30MNA) in wide-angle XRD patterns of the synthesized catalyst strengthens the NS_BET approach. The nitrogen sorption isotherms and corresponding pore size distribution curves of the samples measured at liquid nitrogen temperature are shown in Fig. 2. The pore size distribution was calculated by BJH method. A strong uptake of N₂ as a result of capillary condensation was observed at a relative pressure (P/P_O) of 0.5 and reaches a turning point at 0.9, suggesting that the materials belong to the mesoporous family showing type IV hysteresis loops (Zhang et al., 2006; Enumula et al., 2016). Even though no structure directing template was used in the present study, all the catalysts exhibited a remarkable surface area which shows the significance of the preparation method.

Close

Fig. 2 Pore size distribution curves and corresponding isotherms of (a) HSA, (b) 10MNA, (c) 20MNA and (d) 30MNA catalysts.

Export to PPT

TEM image, HRTEM image and corresponding histogram of 20MNA catalyst are shown in Fig. 3. These images depict that most of the population of Mo₂N particles are present in the range of 5–8 nm range with mean particle size of 6.4 nm, estimated from the histogram of the TEM image. From the results of XRD, NS_BET approach and TEM image we conclude that the nitride particles are highly dispersed over the surface of the support HSA, which are responsible for high EB dehydrogenation activity over 20MNA catalyst.

Close

Fig. 3 (a) TEM image, (b) HRTEM image and (c) histogram of 20MNA catalyst.

Export to PPT

Fig. 4 illustrates the temperature programmed reduction (H₂-TPR) profiles of the synthesized catalysts. All the catalysts shown two T_max values, one low temperature signal at ∼623 K and another high temperature signal at ∼1123 K. The low temperature reduction signal is due to the reduction of surface NH_X species whereas the high-temperature signal corresponds to nitride phase reduction (Hada et al., 2002; Madhavi et al., 2014). In the case of 20MNA and 30MNA catalysts, the low-temperature signal remains at the same temperature as in 10MNA but the high-temperature signal shifted to lower values. This shift in reduction temperature may be attributed to the formation of bigger nitride particles. Pristine MN exhibited similar kind of reduction behavior as supported catalysts. However, the high temperature reduction signal is shifted to further lower temperatures by a margin of ∼100 K compared to the supported catalysts. This shift in temperature is due to formation of bulk nitride particles in bare MN catalyst.

Close

Fig. 4 TPR patterns of (a) 10MNA, (b) 20MNA, (c) 30MNA and (d) MN catalysts.

Export to PPT

The information about the binding energies of elements present in the 20MNA catalyst obtained from XPS analysis are shown in Fig. 5. The spectrum (Fig. 5(a)) shows electron binding energies corresponding to photoelectron peak representing Mo3d of 20MNA catalyst. From the analysis two peaks observed at 234.3 eV and 237.1 eV represents the typical doublet corresponds to Mo3d_5/2 and Mo3d_3/2 spin-orbit components of Mo⁺⁴ and Mo⁶⁺ ions respectively (Ashok et al., 2013). Mo3d of molybdenum nitride shows different binding energy regions in reports, (Chen et al., 2004; Shi et al., 2004; Shi et al., 2005) which depends upon the molybdenum environment, presence in its monometallic form or bimetallic form.

Close

Fig. 5 XPS spectra for (a) Mo 3d and (b) Survey spectrum for 20MNA catalyst.

Export to PPT

3.2 Activity studies

Fig. 6 shows the EB dehydrogenation activity studies, carried out in the range of 723–923 K in a fixed bed reactor system over MN, 10MNA, 20MNA and 30MNA catalysts. In the presence of CO₂ gas feed, MN exhibits very low EB conversion at 873 K, while in the case of 10MNA catalyst, a threefold increase in the conversion of EB was observed at 873 K. 20MNA catalyst showed a maximum EB conversion of 79% with 95.7% selectivity to ST. The reason behind the improved conversion in 20MNA is due to the presence of more number of highly dispersive and active nitride species over the surface of high surface area γ-Al₂O₃ in 20MNA catalyst. In the case of 30MNA catalyst, a decrease in the EB conversion was observed which is due to the formation of Mo₂N crystalline phase, observed from the wide-angle XRD patterns. The conversion of EB at 873 K follows the order 20MNA > 10MNA = 30MNA > MN. Based on these results 20MNA catalyst was chosen to be optimum catalyst and 873 K was optimum reaction temperature for further reactions.

Close

Fig. 6 Ethylbenzene conversion and styrene selectivity over MN, 10MNA, 20MNA and 30MNA catalysts (Closed symbols represent EB conversions and open symbols represents ST selectivities). Reaction conditions: Catalyst-1.0 g; GHSV: 154.5 h⁻¹, atmospheric pressure.

Export to PPT

To study the effect of feed gas we had performed EB dehydrogenation reaction with CO₂, and N₂ as feed gases over 20MNA catalyst and the results were depicted in Fig. 7. In the presence of N₂ flow 60% EB conversion was observed. It is already proved that the nitride catalysts are well-known for ammonia synthesis, and the nitrogen species in the nitride catalysts are exchangeable with N₂ (Aika and Ozaki, 1969; Kojima and Aika, 2001). In the case of CO₂ feed, the EB conversion was slightly enhanced than with N₂ gas due to RWGSR. The improvement in the EB conversion in the presence of CO₂ flow than N₂ is due to governing role played by RWGSR in the dehydrogenation process.

Close

Fig. 7 Effect of feed gases (CO₂, and N₂) on the ethylbenzene dehydrogenation reaction over 20MNA catalyst. Reaction conditions: Catalyst-1.0 g; GHSV: 154.5 h⁻¹, atmospheric pressure.

Export to PPT

Generally a faster catalytic deactivation often encountered in most of the catalytic systems in ODH reactions due to coke formation. Thus, time-on-stream studies are very useful to understand the behavior of the catalysts. The major product observed in the reaction was ST along with some minor products such as toluene, benzene with some gaseous products CO, NH₃ (in the case of N₂ feed gas) and traces of CH₄, H₂, C₂H₅ were detected. The major reason behind the study of the effect of the composition of CO₂ and N₂ is to know the mutual effect of N₂ and CO₂ on the EB dehydrogenation activity. The pooled effect of feed gasses CO₂ and N₂ with different compositions (CO₂, 15%N₂-85%CO₂, 50%N₂-50%CO₂, 85%N₂-15%CO₂ and N₂) were studied over 20MNA catalyst. Fig. 8 reveals the outcome of the above gas compositions. In the case of pure CO₂, conversion initially increased from the first hour to the second hour and then gradually decreased whereas selectivity increased linearly and gets stabilized. With the addition of slight amount (15%) of N₂ as co-feed gas, the conversion elevated to a maximum value of 90% and drops down to 68% within a period of 24 h and selectivities are linearly increased and stabilized at 97%. These results suggest that the addition of N₂ to CO₂ enhances the dehydrogenation activity to maximum, because of unified effects of RWGSR and ammonia synthesis. The acidic sites present over Al₂O₃ are prone for coking when contact with organic moiety at high temperature. The role of NH₃ produced in this reaction is to neutralize the acidic sites present over the catalyst surface associated with Al₂O₃ support. Further additions of N₂ to CO₂ showed a decline in the EB conversion but the deviation in the conversion was lower compared to earlier cases suggesting that N₂ is stabilizing the EB conversion by neutralizing the acidic sites present on the surface of the catalyst and preventing additional coke formation. The same criterion was observed with N₂ and with higher N₂ compositions.

Close

Fig. 8 Time on stream effect of CO₂ & N₂ feed gas compositions over 20MNA catalyst. Reaction conditions: Catalyst-1.0 g; EB- 1.0 ml h⁻¹; GHSV: 154.5 h⁻¹, atmospheric pressure.

Export to PPT

The amount of ammonia formation was estimated by separate experiment at the reaction temperature with N₂ and H₂ in 1:3 mol ratio without EB feed. The amount of ammonia formed on nitride catalysts was shown in Fig. 9(a). N₂ and H₂ in 1:3 mol ratio were impelled onto the catalyst and at the outlet of the reactor a liquid trap containing 5 cm³ of pre-prepared 0.0005 N H₂SO₄ was taken to neutralize ammonia. Methyl orange was taken as a color indicator to know the end point. The end point time was noticed by a stopwatch by taking time interval between experiment starting and end point times. The endpoint times were noticed at regular intervals of 0.5 h by replacing the trap with a fresh one. The amount of ammonia formation increased with time and then stabilized after 1 h of experiment in all the loadings. The formed NH₃ (Fig. 9(a)) for the 10MNA, 20MNA and 30MNA at the end of the experiment were 0.248, 0.441 and 0.622 mmol h⁻¹ g⁻¹ nitride respectively. With the increase in the nitride loading, the amount of ammonia formation increased, which is due to the increase in the nitrogen percentage in the catalyst.

Close

Fig. 9 (a) NH₃-formation over (a) 10MNA, (b) 20MNA and (c) 30MNA catalysts. Reaction conditions: Catalyst-1.0 g; N₂ to H₂ mole ratio: 1:3, atmospheric pressure. (b) CO₂ conversions over 20MNA catalyst. (c) XRD spectrum of spent 20MNA catalyst in 85%CO₂ + 15% N₂ flow. (d) TGA analysis of 20MNA fresh and 20MNA spent (20MNAU) catalysts.

Export to PPT

The CO₂ conversions are mainly based on the CO₂ adsorptions on the catalyst surface (Reddy et al., 2009; Burri et al., 2015). The CO₂ conversions during the reaction process of 20MNA catalyst with 15%N₂-85%CO₂ gas feed were depicted in Fig. 9(b). With time there is a slight decrease in the CO₂ conversion was observed, which is due to the formation of coke species during time on stream process. The decrease in EB conversions can be ascribed to the coke formation over the catalyst surface. The deposited coke is the major cause for the deactivation of the catalyst. The coke obstructs the adsorption of reactants on the catalytic active sites and further improves the cracking process by adsorption of the reactant on the coke, (Burri et al., 2015) which results in loss of reactant conversions when the process runs for long time. But by using CO₂ as one of the feed gases the formation of coke over the catalyst surface can also be minimized by Boudouard reaction (CO₂ + C ↔ 2CO).

20MNA catalyst after 24 h of TOS study in the presence of 85% CO₂ and 15% N₂ flow was characterized by XRD analysis and shown in Fig. 9(c). XRD results imply that the presence of nitride phase in spent catalyst, while the fresh catalyst did not showed any nitride peaks suggesting that crystalline phase formation in the spent catalyst. Fig. 9(d) divulges the TGA analysis of fresh and spent 20MNA (20MNAU) catalyst in the presence of air flow (10 ml/min) after TOS study in 85% CO₂ and 15% N₂ flow. The fresh catalyst has shown almost no weight reduction where as spent catalyst has shown a weight loss of 22%. The reduction in catalyst weight in 20MNAU is started from 660 K and completed by 900 K and it can be attributed to the formation of carbonaceous species over the catalyst surface. 20MNA catalyst has shown a coke formation of 0.763 mmol-carbon/h/g catalyst, which is low compared to our previous reports (Gurram et al., 2016; Itika et al., 2017). The decrease in coke formation is mainly composed of two reasons, one is due to the novel molybdenum nitride catalyst, which is able to produce ammonia from N₂ and H₂ and the second is due to neutralization of surface acidic sites by in-situ generating NH₃. The 20MNAU catalyst after 24 h TOS study was evaluated by TEM analysis and the TEM image, HRTEM image and corresponding histogram for 20MNAU catalyst are shown in Fig. 10. These images depict that most of the population of Mo₂N particles are present in the range of 9–12 nm range with mean particle size of 10.4 nm, estimated from the histogram of the TEM image, which is 4 nm higher than the fresh 20MNA catalyst. XRD and TEM characterization results of 20MNAU catalyst suggests that the nitride species are growing its crystallite size during the reaction progress.

Close

Fig. 10 (a) TEM image, (b) HRTEM image and (c) histogram of spent 20MNA catalyst.

Export to PPT

Based on the reaction results a plausible reaction mechanism was proposed and presented in Scheme 2. The rate-determining step could be the chemisorption of nitrogen retarded by adsorbed nitrogen atoms (Aika and Ozaki, 1969; Kojima and Aika, 2001). This step suggests that the nitrogen in the nitride catalyst is the active species during the ammonia synthesis process. The present catalytic reaction follows Mars-Van Krevelen mechanism, in which catalyst surface acts as active species and forms a chemical bond with the reactant (Doornkamp and Ponec, 2000). In the present study Mo⁶⁺ is the active center surrounded by nitrogen atoms. In the first step, EB chemically attacks Mo⁶⁺ ions of the catalyst surface forming a thin surface layer of Metal-Reactant. Reduction of metal ions by reactant molecule follows the second step resulting in products (styrene and ammonia). Due to loss of nitrogen in the form of ammonia Mo⁺⁵ acts as a nitrogen acceptor and regains its original structure by reacting with the molecular nitrogen. Kojima and Aika experimentally proved this step by N¹⁵ tracer and deuterium isotope test (Kojima and Aika, 2001).

Close

Scheme 2 Plausible reaction pathway for the ethylbenzene dehydrogenation reaction with (a). N₂ feed gas and (b) CO₂ feed gas over nitride catalyst.

Export to PPT

ODH process precedes via electron donation route to the acidic site of the catalytic surface and basic sites abstracts proton from reactant molecule (Mukherjee et al., 2016). Thus both acidic sites and basic sites are necessary for ODH process. Majority of the reports on EB dehydrogenation and Ethylene dehydrogenation reactions using CO₂ as soft oxidant proceeds via this redox or acid-base mechanism. These acid-base bifunctional catalysts drive the ODH process to good results. In the present study Mo⁶⁺ acts as electron acceptor from benzene ring and nitrogen acts as proton acceptor.