(143a) Coverage-Dependent Adsorption of Hydrogen on Fe(100): Using Lattice Gas Parameters As Design Variables to Enhance Fe’s Resistance to Oxidative Deactivation

Authors: 
Hensley, A., Washington State University
McEwen, J. S., Washington State University
Collinge, G., Washington State University
Bray, J., Washington State University

            Hydrogen
adatoms are known to significantly affect the adsorption configurations of
co-adsorbates at high coverages,[1, 2] and yet many atomistic models of these
reactions tend to take the presence (or absence) of adsorbed hydrogen for
granted. Since reaction pathways can be significantly altered due to the coverage
distribution of hydrogen,[2] it is important to quantitatively capture such
distributions under realistic reaction conditions. Here, we determine how hydrogen-hydrogen
interactions affect the ordered structures that are formed on Fe(100), a system
of particular relevance to major industrial reactions such as ammonia
synthesis, Fischer-Tropsch, and hydrotreating. We do this by parameterizing the
surface energy―calculated with density functional theory (DFT)―for
each H/Fe(100) configuration at a given coverage using a lattice gas cluster
expansion (LG CE). This H/Fe(100) LG CE was fit to 950 unique H/Fe(100)
configurations and has a leave-multiple-out cross validation score of 3.8
meV/site,[3] clearly demonstrating a high level of predictability in our final model.
Furthermore, the thorough scan of the H/Fe(100) configuration space allows for
the identification of the electronic ground state systems, i.e. the lowest
energy configurations as a function of coverage. Nine ground states were found
for H/Fe(100) over the entire coverage range. The ground states for coverages
up to 1.000 monolayer (ML)―saturation coverage at ultra-high vacuum―reproduce
H/Fe(100) temperature programmed desorption and electron energy loss spectroscopy
experimental results.[4] We have subsequently determined the catalytic relevance
of each ground state by calculating the Gibbs free energy of each system as a
function of temperature at 1 bar H2 pressure (Figure 1A). The
resulting phase diagram shows that the saturation coverage is 1.000 ML and the dominant
ground states are the 1.000, 0.875, and 0.400 ML systems (Figure 1A). Our LG CE
reveals that this result is largely due to a major lack of interactions
between hydrogen up to 1.000 ML, with major repulsions beginning to take effect
at coverages greater than 1.000 ML. Given this insight, we can conclude that
researchers would be remiss to exclude high coverages (up to 1.000 ML) of
hydrogen in their models of catalytic reactions without justification.

            We
have further compared the ground states of H/Fe(100) with those of O/Fe(100),
previously studied by Bray et al.[5] Our analysis shows that both the
adsorption energy and the energetic coverage dependence for oxygen are ten
times larger than for hydrogen. Further, it is worth noting that the most
attractive interactions in the O/Fe(100) system are longer-range in nature and
thus much more likely to be labile to disruption by dopants as posited in
Figure 1B. Electronically, the adsorption of both hydrogen and oxygen on
Fe(100) withdraws electrons from the surface (Bader charges of -0.4 and -1.2
electrons/adspecies, respectively), shifting the d-band center away from the
Fermi level. As the d-band dependence on coverage for oxygen is twice as
large as that for hydrogen, this suggests that surface modification leading to
shifts in the Fe d-band will have an even greater destabilization effect
for oxygen as compared to hydrogen.

 

[1]
Ferguson, G.A.; Vorotnikov, V.; Wunder, N.; Clark, J.; Gruchalla, K.;
Bartholomew, T.; Robichaud, D.J.; Beckham, G.T. J. Phys. Chem. C, 2016,
120, 26249-26258.

[2]
Wang, S.; Vorotnikov, V.; Vlachos, D.G. ACS Catal., 2015, 5, 104–112.

[3]
Baumann, K. Trends Anal. Chem., 2003, 22, 395-406.

[4]
Merrill, P.B.; Madix, R.J., Surf. Sci., 1996, 347, 249-264.

[5]
Bray, J.; Collinge, G.; Stampfl, C.; Wang, Y.; McEwen, J.-S. Top. Catal.,
2018, 61, 763-775.

Figure 1. (A) Ab initio phase diagram
for the H/Fe(100) electronic ground state systems as a function of temperature
at 1 bar H2 pressure, showing that saturation coverage is 1.000 ML
(insert). The dominant ground state changes at 810 K to the 0.875 ML system, 1000
K to the 0.200 ML system, and 1050 K to clean Fe(100). (B) Schematic showing the
potential modification of oxygen and hydrogen distributions on Fe(100) via
alterations in surface composition and d-band, which disrupts the most attractive
interactions as determined from a cluster expansion (blue lines). The gold,
silver, red, white, and pink spheres represent Fe, modified surface species
(dopant metal or electronically altered Fe), oxygen, hydrogen in 4-fold hollow,
and hydrogen in 3-fold hollow, respectively.