Elettra | SuperESCA The basal plane of graphene grown on nickel dissociates water molecules Although graphene (Gr) could play a dominant role in hydrogen production and storage due to its high specific surface area, its use is actually limited by its chemical neutrality. The close interaction with a metal support, responsible for a strong charge transfer, could lead to the formation of a dipole layer capable of catalyzing chemical reactions, and, in this respect, the Gr/ Ni(111) interface has often been reported as one of the most reactive Gr/metal systems. In this study we combined X-ray Photoelectron Spectroscopy (XPS), Thermal Programmed Desorption (TPD), surface X-ray Diffraction (XRD) and Scanning Tunneling Microscopy (STM) to investigate the early interaction between water and Gr/Ni(111) at room temperature (RT). Then, we used Density Functional Theory (DFT) calculations to determine the energetics of the dissociation of single water molecules in different interface configurations. Figs. 1 (a, b) show the C1s and O1s spectra measured on the clean Gr/Ni(111) after the exposure to 0.02 ML (1 ML=106 Langmuir=1.33 mbar s) of H2O molecules at pressure of 5x10-4 mbar t RT. In addition to the C0 component, arising from the unperturbed Gr regions, the C1s spectrum shows components appearing on the high and low binding energy sides of the main peak. The same components are observed when Gr/Ni(111) is exposed to a low dose of H atoms, and thus, in analogy with this case, A1 and A2 are attributed to C atoms directly bonded to H atoms and B1 and B2 to C atoms first neighbors of the C-H bonds. The hydrogenation of Gr does imply the occurrence of water splitting. Interestingly, the corresponding spectrum measured in the O1s region Fig. 1 (b) is flat, which means that the OH counterparts of the H atoms do not remain on the Gr surface. Note that, on bare Ni(111), water molecules adsorb molecularly below 150-170 K, but require vibrational excitation to dissociate when impacting the Ni surface at RT. Then, the presence of Gr appears indispensable to dissociate water at RT. The occurrence of Gr hydrogenation is confirmed by the evolution of the C1s spectrum measured while heating the sample (Fig. 1 c) and by the corresponding TPD curves (Fig. 1 d), which show in order the disappearance of the C-H bonds and in parallel the release of molecular hydrogen around 600 K. Furthermore, surface XRD measurements excluded the occurrence of intercalation demonstrating that water dissociation initiates at Gr and not below Gr. The STM image acquired on the clean Gr/Ni(111) surface (Fig. 1 e) shows the hexagonal Gr lattice with some dispersed adatoms, assumed to be Ni atoms trapped at C vacancies that could behave as active sites for catalytic reactions. After a water dose of 4.5x10-3 ML, the Gr layer appears covered by a uniform distribution of H atoms, which remains homogeneous even near the step edge, strongly suggesting that water dissociates on the basal plane of Gr (Fig. 1 f). To understand the mechanics of wafer splitting we have evaluated the energetics of the reaction by DFT calculations. Starting from the plain Gr/Ni(111) interface it comes out that the adsorption energy Ea of the most stable configuration with H and OH fragments bonded to C atoms is +0.77 eV, which indicates an endothermic process incompatible with the prominent dissociation yield observed experimentally. Then we explored the possibility that the splitting reaction could be made easier by the presence of adsorbates, lattice defects or structural discontinuities (see structures at the bottom of Fig. 1. At first, we considered the presence of a pre-adsorbed H atom, which, inducing a sp3 character in the flat Gr layer, could enhance the local reactivity. As lattice defects we considered monovacancies obtained by removing a C atom from the Gr lattice. In the first case a preadsorbed H decreases Ea from 0.77 eV to -1.60 eV, rendering the water dissociation exothermic. However, due to the energy barrier Eb = 2.27 eV to reach the final state (Fig. 1 g), the reaction would need the support of external energy. When moving to the C monovacancy, the most stable configuration for dissociated water would be even more energetically convenient (Ea=-2.6 eV), but also in this case the high Eb of 1.5 eV is still too high to allow spontaneous splitting at RT. A stronger substrate reactivity is obtained when the C monovacancy is saturated by a Ni atom. Our calculations found that the most favorable configuration, that is with H and OH bonded to two C atoms neighboring the Ni dopant can be reached by overcoming a barrier of 0.8 eV (Fig. 1 g), a value compatible with RT dissociation at a reasonable rate. The final state is exothermic by -0.8 eV and the energy gained in the reaction can be used by H and OH to diffuse away from the defect site, that then remains free to catalyze new dissociation reactions. The limited surface density of Ni trapped at C vacancies may not account for the H coverage observed by XPS, which could perhaps only be explained by the occurrence of collective adsorption processes that we have not considered. Cooperative effects that can lower the dissociation energy barrier of water molecules adsorbing as dimers or larger clusters are likely to occur at the dosing pressure we used. However, the role of trapped Ni atoms to initiate the splitting reaction locally could be crucial to start autocatalytic reactions that are often triggered by adsorbed OH groups. Further experiments conducted as a function of dosing pressure and temperature could better elucidate the complex mechanism of water dissociation allowing the optimization of experimental parameters that maximize hydrogen production. Elettra Highlights 2024–2025 130
RkJQdWJsaXNoZXIy MjQ4NzI=