Influence of Hydrogen-Incorporation on the Bulk Electronic Structure and Chemical Bonding in Palladium

TL;DR

This study addresses the challenge of probing the bulk electronic structure of metal hydrides by employing ambient-pressure hard X-ray photoelectron spectroscopy (AP-HAXPES) on Pd thin films under 200 mbar H$_2$. By coupling AP-HAXPES with in-situ XRD and neutron reflectometry, the authors correlate hydrogen loading with lattice expansion and electronic-structure changes, then interpret these results with DFT PDOS calculations. The findings show hydrogen occupancy primarily in octahedral sites, a narrowing of the Pd $d$-band due to reduced Pd–Pd overlap and Pd–H hybridization, and the emergence of hydrogen-induced bonding states below the valence band, providing bulk evidence of Pd–H interactions in PdH$_x$. The work demonstrates AP-HAXPES as a powerful bulk-sensitive probe for metal hydrides under realistic hydrogen pressures and temperatures, with implications for understanding hydrogen storage, catalysis, and related materials properties.

Abstract

Palladium hydride is a model system for studying metal-hydrogen interactions. Yet, its bulk electronic structure has proven difficult to directly probe, with most studies to date limited to surface-sensitive photoelectron spectroscopy approaches. This work reports the first in-situ ambient-pressure hard X-ray photoelectron spectroscopy (AP-HAXPES) study of hydrogen incorporation in Pd thin films, providing direct access to bulk chemical and electronic information at elevated hydrogen pressures. Structural characterisation by in-situ X-ray diffraction and neutron reflectometry under comparable conditions establishes a direct correlation between hydrogen loading, lattice expansion, and electronic modifications. Comparison with density functional theory (DFT) reveals how hydrogen stoichiometry and site occupancy govern the density of occupied states near the Fermi level. These results resolve long-standing questions regarding PdH and establish AP-HAXPES as a powerful tool for probing the bulk electronic structure of metal hydrides under realistic conditions.

Figures (5)

• Figure 1: Depiction of the unit cell for PdH$_x$ with $x$ = 1, with (a) H occupying the octahedral sites, and (b) H occupying the tetrahedral sites, including Pd--H bond lengths in Å. Both structures are those obtained following geometry optimisations using density functional theory, where the lattice parameters were also allowed to relax. The corresponding lattice parameters are given in Table \ref{['tab:lattice']}. . For $x$$<$ 1, the interstitial sites are only partially occupied and the lattice constant is smaller.
• Figure 2: In-situ structural characterisation of the sample composed of a 4 nm Ti and a 50 nm Pd layer on a Si wafer. The measurements were performed by stepwise lowering the temperature from $T$ = 200 ° C to the temperature indicated under $P_{H2}$ = 200 mbar. (a) Diffraction patterns around the (111) reflection (full diffraction pattern available in Fig. 2 in the Supplementary Information). The continuous lines indicate fits of a pseudo-Voigt function to the data. (b) Neutron reflectivity as a function of the momentum transfer $Q$. The continuous lines represent fits of a three-layer model to the data. (c) The scattering length density (SLD) profiles corresponding to the fits to the data presented in (b). Based on the SLD profiles and Equ. 1, the hydrogen-to-metal ratio and layer expansion of the PdH$_x$ layer have been determined. (d) Out-of-plane $d_{111}$-spacing expansion of the film due to the absorption of hydrogen. At each temperature, the $d_{111}$ spacing at $P_{H2}$ = 200 mbar is normalised to the spacing as measured in vacuum, $d_{vac}$. (e) Temperature dependence of the hydrogen-to-metal ratio and (f) layer expansion at $P_{H2}$ = 200 mbar.
• Figure 3: AP-HAXPES data collected during the heating of a hydrided Pd film under 200 mbar H2. (a) Pd 3d$_{5/2}$ core state spectra. Dashed lines and the arrow illustrate shifts in the main peak position. (b) Pd 3d$_{5/2}$ AP-HAXPES core state spectrum after heating to the maximum temperature (T$_{max}$) with depiction of the defined peak widths, where FWHM is the total full width at half maximum, and HWHMlow and HWHMhigh are the half width at half maximum for the low and high binding energy (BE) side of the binding energy centre (BEcentre). (c) Valence band maximum (VBM) spectra. Dashed lines and arrows illustrate shifts in the VBM position. The position of the Fermi energy $E_F$ at 0 eV is indicated. (d) Extracted Pd 3d$_{5/2}$ peak widths. (e) Ratio of the two Pd 3d$_{5/2}$ HWHMs compared to the single point signal intensity at $E_F$, I ($E_F$).
• Figure 4: Comparison of PDOS for PdH0.75 with different photoionisation cross-section, $\sigma$, correction strategies, including (a) PDOS without correction, (b) using Scofield $\sigma$ for Pd, (c) using Scofield $\sigma$ with In correction, and (d) comparison of all three theoretical PDOS correction approaches with the experiment at highest hydrogen loading. The position of the Fermi energy $E_F$ at 0 eV is indicated in all Subfigures, and Roman numerals are used to indicate the main spectral features observed. In (d), PDOS and VB spectrum are normalised to the height of feature I.
• Figure 5: Comparison of AP-HAXPES valence band (VB) spectra with calculated PDOS after one-electron photoionisation cross-section correction, including (a) photoionisation cross-section corrected sums of theoretical PDOS for the Pd to PdH series, (b) AP-HAXPES VB spectra collected after exposure to 200 mbar H2 followed by heating to 200 $^\circ$C, and (c) comparison of the calculated PDOS for PdH0.75 using the Scofield In $\sigma$ correction and the AP-HAXPES VB spectrum with the highest hydrogen loading. The position of the Fermi energy $E_F$ at 0 eV is indicated in all Subfigures, and Roman numerals are used to indicate the main spectral features observed. In (c), PDOS and VB spectrum are normalised to the height of feature I.