Metal Phosphides for Selective Alkane Dehydrogenation

Increased shale gas production in the United States has resulted in an abundance of light alkanes, which can be converted to light alkenes via catalytic dehydrogenation and used as building blocks for transportation fuels and chemicals. The challenges of alkane dehydrogenation are (1) to suppress coke formation and (2) to replace current expensive Pt-based catalysts. One particular class of understudied, yet highly tunable catalysts are transition metal phosphides due to their inherent thermal stability and the ability to structurally and electronically tune their catalytic performance for hydrogenation and dehydrogenation reactions.^1,2

In this contribution, we will discuss recent experimental and computational methods to evaluate a range of metal phosphides for ethane and propane dehydrogenation. To date, we have evaluated a number of metal phosphides supported on SBA-15 for both propane and ethane dehydrogenation using time on stream experiments in a flow reactor. Catalysts have been characterized using XANES, EXAFS, XPS, TEM, N₂ physisorption, as well as other techniques in order to correlate structure with catalytic performance. We have determined that P addition vastly improved the alkene selectivity from Ni to Ni-P, going from 0% to > 99% selectivity for propane dehydrogenation. The P is most likely acting in a manner similar to Sn in current industrial Pt-Sn dehydrogenation catalysts, helping to prevent the formation of large Ni ensembles which favor side reactions such as hydrogenolysis. In order to better understand the effects of P addition we employed density functional theory (DFT) calculations to compare the reactivity for dehydrogenation of Ni₂P with that of Ni. Firstly, we compared adsorption energies and site preferences of C1 and C2 hydrocarbon species on Ni(111) and Ni₂P(001). They generally exhibit weaker adsorption strength with Ni₂P compared to Ni. Certain species, however, tend to favor binding at P sites over Ni sites, providing further stabilization. We then predicted ethylene selectivity by comparison of the activation barrier for ethylene dehydrogenation and ethylene desorption energy on both surfaces. While both surfaces bind ethylene nearly isoenergetically, ethylene dehydrogenation to vinyl is considerably more activated on Ni than on Ni₂P. These results imply a higher selectivity towards ethylene on Ni₂P over Ni.

As a means to enhance the stability of Ni-P catalysts, we evaluated the effects of Cs doping on the catalyst performance. For ethane dehydrogenation, Cs-Ni-P improved the ethylene selectivity over Ni-P, going from 67% for Ni-P to >99% for Cs-Ni-P due to the reduced acidity of the catalyst which also reduces side reactions.³ We have also evaluated the performance of other mono- and bimetallic transition metal phosphides for dehydrogenation reactions. When comparing Mo-P and Ni-Mo-P for propane dehydrogenation, we found that Ni-Mo-P displayed a higher propylene selectivity than Mo-P, >99% and 97%, respectively. Both Fe-P and Ni-Fe-P showed selectivity to propylene of >99% as well. Using a first order deactivation model, the stability of the bimetallic materials was significantly higher than the monometallic catalysts. The observed enhancement is most likely due to both structural and electronic differences that result from the incorporation of Ni.

Overall, our results provide evidence that a variety of metal phosphides are selective dehydrogenation catalysts, and that bimetallic compositions can further enhance the alkene selectivity and stability of metal phosphides. This presentation will highlight our current efforts in identifying new catalyst compositions for light alkane dehydrogenation reactions.

References

(1) S. T. Oyama and T. Gott et al., Catal. Today. 2009, 143, 94–107. (2) Y. Xu and H. Sang et al., Applied Surface Science. 2014, 316, 163-170. (3) Sattler, J.J.H.B et al., Chem. Rev. 2014, 114, 10613-10653.