Correlation of Electronic Structure with Catalytic Activity: H2-D2 Exchange across CuxPd1-x Composition Space
G. Gumuslu, P. Kondratyuk, J. R. Boes, B. Morreale, J. B. Miller, J. R. Kitchin and A. J. Gellman, “Correlation of Electronic Structure with Catalytic Activity: H2-D2 Exchange across CuxPd1-x Composition Space”, ACS Catalysis 5(5), 3137-3147 (2015). https://doi.org/10.1021/cs501586t
This summary was written by a language model from the paper and has not been checked line by line. Rely on the published paper for details.
Question
Computational studies routinely correlate the d-band center of a metal or alloy with barriers for elementary surface steps, but testing such relationships experimentally requires many catalysts of finely spaced composition, which is slow to prepare one sample at a time. This paper asks whether a continuous library of Cu-Pd alloy compositions can be used to measure both a valence band energy and the barriers for H2 adsorption and desorption at the same compositions, and so obtain an experimental barrier versus electronic structure correlation across composition space (Gumuslu et al. 2015). The authors present it primarily as a demonstration of a general methodology.
Approach
CuxPd1-x composition spread alloy films (CSAFs), about 100 nm thick, were deposited on Mo substrates with an offset-filament evaporation tool and annealed at 800 K. Spatially resolved XPS mapped the near-surface composition, and the region probed by a 10 × 10 array of microreactors spanned x = 0.30 to 0.97. Spatially resolved He I UPS gave the density of filled valence states at each point, and the mean energy of the filled valence band, εv, was taken as an experimental proxy for the DFT d-band center. Bulk alloy DFT calculations (VASP, PBE) at nine compositions from pure Pd to pure Cu provided the filled d-band and total valence band centers for comparison. H2-D2 exchange was run in a 100-channel microreactor array at atmospheric pressure from 333 to 593 K under three H2/D2/Ar feed conditions, with HD production measured by mass spectrometry. A two-step microkinetic model (dissociative adsorption and recombinative desorption on two sites, no isotope effects, coverage-independent parameters) taken from earlier foil-catalyst work was fit to the conversion data at each composition to obtain pre-exponents and barriers for adsorption and desorption (Gumuslu et al. 2015).
Key findings
The UPS valence band center moved monotonically from about -3.9 eV at x = 0.97 to -3.3 eV at x = 0.30, shifting toward the Fermi level as Cu content fell, and extrapolation suggested about -3.0 eV for pure Pd. DFT reproduced the slope of this shift with an offset of roughly 0.8 eV, which the authors attribute partly to incomplete secondary electron background subtraction, and showed that the total filled-state center tracks the d-band center, supporting εv as a proxy. H2-D2 exchange activity decreased monotonically with increasing Cu content, reaching a conversion near 0.4 at x ≈ 0.3 and 593 K.
The fitted pre-exponents were roughly composition independent (log νads about -4.0 and log νdes about 6.0 in the units given) and consistent with transition state theory, so they were fixed while the barriers were refit. The adsorption barrier was about 0.15 eV for x < 0.6 and rose to about 0.4 eV for x > 0.85; the desorption barrier rose from about 0.55 to 0.65 eV over x = 0.3 to 0.8. For x < 0.7 desorption is rate limiting, and for x > 0.8 the desorption barrier became indeterminate within its uncertainty, indicating a switch to adsorption-limited kinetics near x ≈ 0.8. Barriers near x = 0.53 matched the B2 phase measured earlier on foils, suggesting the film retains its stable bulk structure under reaction. Plotting barriers against εv, the adsorption barrier fell at about 1.5 eV per eV from Cu toward roughly Cu0.5Pd0.5 and then stayed constant; the desorption barrier was much less sensitive (about 0.2 eV per eV) and correlated more poorly. The authors propose that hydrogen-induced Pd segregation saturates the top layer of Pd-rich alloys, so the adsorption barrier stops tracking a near-surface εv that still follows bulk composition (Gumuslu et al. 2015).
Significance
The work is presented as the most comprehensive experimental measurement to date of the kind of electronic structure versus barrier relationship predicted by DFT, obtained across 100 compositions in a single library. It shows that the expected linear trend holds only over part of composition space and points to adsorbate-induced segregation as a factor that standard DFT correlations omit. The authors note the method is general but limited to reactions with turnover above about 1e-5 per second, temperatures below 600 K, pressures below 1 atm, mass spectrometric detection, and mechanisms simple enough for microkinetic fitting (Gumuslu et al. 2015).