Platinum Metals Rev., 1998, 42, (4), 141
Formation and Decomposition of Palladium Hydride Particles
Imaging Pictures on the Nanometre Scale
Nanoscale changes in surface structure that accompany the low temperature exposure of palladium to hydrogen are reported. Field Emission Microscopy, a method for rapid in situ imaging of surface processes, has been used to examine palladium tips of radius ∼ 200 nm, produced by a novel technique. Images are presented of the initial stages of the uptake of hydrogen. Subsurface hydrides were initially formed when palladium tips were exposed to hydrogen gas at low temperatures, starting at highly open surfaces present on the tip. Extruding PdH particles were also formed on top of the palladium tip and their growth was observed to proceed in a ‘staccato’-like manner. Palladium crystallites remained on the surface after most of the hydrogen had been removed from the palladium sample by heating in vacuum. On heating the crystallites remained quite stable up to a temperature of ∼ 700 K, but then melted back into the tip.
A detailed understanding of the process of hydrogen absorption by metallic phases (metals, alloys and intermetallic compounds) is important for the development of new materials for hydrogen storage. The interaction of hydrogen with palladium in particular has been extensively studied (1,2), since the first report of the absorption of hydrogen into palladium in 1866 and since the first measurements of the palladium-hydrogen pressure-constitution-temperature relationship in 1895 (3). Two non-stoichiometric hydride phases can co-exist below the critical temperature (∼ 300°C), and on going from the α-to the β-phase there are large increases in volume, with the lattice constant increasing by ∼ 3.3 per cent, which is a volume increase of ∼ 11 per cent. These changes occur continuously over the phase transition.
It is surprising that comparatively few studies have been focused on the structural changes occurring at the metal surface under the strain of the expanding lattice. Triangular shapes, of ∼ 10 to 20 μm in size, have been seen developing during the early stages of hydrogen loading of a Pd(111) single crystal (4). In addition, networks in parallel lines have been observed on large single crystals by Sugeno and Kawabe (5). These patterns became apparent on complete transformation to the β-phase and could sometimes be seen after desorption of hydrogen (5).
Field Emission Microscopy (FEM) (resolution ∼ 2 nm) images of surface processes were observed in situ during exposure of stable clean palladium tips to hydrogen, see Figure 1. The tips, produced by a novel technique (6), were held in hydrogen at a pressure of 2.6 x 10-4 mbar and 147 K. Figure 1(a) is characteristic of a clean palladium tip with the (110) plane in its centre. In general, a FEM pattern represents a work function map of the various crystal faces on the end of a hemispherically shaped single crystal tip. Figure 1 (b) shows the growth of high intensity patches (that is, patches of increased work function) first occurring around (551) surfaces, which are basically (110) terraces separated by (111) steps. Similar structures were found to grow on a very open (320) face, and were also observed on the (553) surfaces. In Figurea 1(c) the case at saturation is represented, where these high intensity patches stop growing.
Surfaces defects have been found to be very important for the low temperature uptake of hydrogen below 147 K (7, 8). At 147 K, no build-up of such high electron emission centres has been observed at pressures below the one primarily studied here.
We have examined in detail the ‘staccato’-like growth of these novel structures. There appears to be overall continuous slow growth of particles, separated by periods of acceleration. The shape of most of the particles was rectangular or triangular, with the latter having previously been seen on a larger scale on the Pd(111) surface (4). After evacuation of hydrogen from the gas phase, such structures formed on the surface at 147 K still remained intact.
The effects of heating the hydrided tip are shown in Figure 2. The majority of the hydrogen desorbs from the tip at-190 K, but many of the structures formed have remained on the surface (6). There are only slight changes in the palladium tip between 300 K (Figure 2(a)) and 600 K (Figure 2(b)), with Figure 2(b) showing the situation after several minutes at 600 K. For the low-temperature H2-loaded palladium wire and for a Pd(110) surface, no high temperature β-desorption peak has been observed, suggesting that there was no hydrogen penetration into the bulk (9-11). However, the β-desorption peak does appear when hydrogen is loaded at a temperature of 300 K. With the palladium wire, heated at a rate of 5 K s-1, the β-peak appears at 620 K (11).
Even if hydrogen had still been present in the palladium tip, it would have diffused out of the sample after several minutes at 600 K (7-11). It therefore seems possible to conclude that the patches left on the surface at this point are palladium microcrystals, still on top of the palladium tip. At 700 K however (Figure 2(c)) these crystallites have all but disappeared, the obvious interpretation being that the palladium atoms are now becoming mobile and that the palladium microcrystals are melting back into the tip.
One of the questions that remains to be answered is whether the structures formed on the palladium surface under the hydrogen dosing at low temperature (that is, 147 K), also contain hydrogen, or whether they are pure palladium crystals formed on top of a surface hydride. It is clear that the palladium atoms are mobile on the surface at low temperatures under the influence of an expanding PdH lattice, although possibly only in the topmost layers of the tip.
The staccato growth of the treated particles gives an indication that the structures are actually PdH crystallites at low temperature. The slow growth can be equated to PdH lattice expansions in α-to β-phase transitions and the accelerated growth periods can be equated to stress causing palladium atoms to radically change positions. Indeed, the fact that the tip could take on a new orientation after high temperature annealing of a sample exposed to H2 at low temperatures, also seems to indicate the influence of stress.
The use of FEM has provided a high resolution view of the kinetics of the initial stages of PdH formation at palladium surfaces exposed to H2 at low temperatures. At patches of increased emission, initiated on the more open surfaces of the palladium tip, growth is mainly slow – since the palladium lattice must expand in taking up the hydrogen atoms, but there are rapid increases in the growth rate as the strains in the β-phase cause atoms to be expelled or rearranged. At saturation the particles have grown not only two dimensionally, but also in height, such that when hydrogen is removed, microcrystals of palladium remain on top of the palladium tip. These palladium microcrystals are stable until temperatures are reached at which the palladium atoms become thermally activated. The fact that hydrogen has the ability to cause movement of the palladium atoms at such low temperatures is significant. Indeed, many recent studies have shown that such metal surfaces can be quite mobile and the present study has demonstrated that large reconstructions can occur when palladium interacts with hydrogen at 150 K. However, in the bulk of a metal in which a hydride is being formed, there are few places for the atoms to go when the lattice expands. As a consequence it would appear that more thought needs to be given to stabilising the surface in order to prevent material degradation, when metal-hydrogen reservoirs are being designed.
- 1F. A. Lewis, “The Palladium Hydrogen System”, Academic Press, 1967, London/New York
- 2F. A. Lewis, Int. J. Hydrogen Energy, 1981, 6, 319
- 3C. Hoitsema, Z. Phys. Chem., 1895, 17, 1
- 4T. J. Tiedema,, B. C. de Jong and W. G. Burgers, Proc. Kon. Ned. Akad. Wet., 1960, 63B, 422
- 5T. Sugeno and H. Kawabe, Mem. Inst. Scient. Ind. Res. Osaka Univ., 1957, 14, 25
- 6P. D. Cobden,, V. V. Gorodetskii and B. E. Nieuwenhuys, to be published
- 7R. Dùs,, E. Nowicka and Z. Wolfram, Surf. Sci., 1989, 216, 1
- 8H. Okuyama,, W. Siga,, N. Takagi,, M. Nishijma and T. Aruga, Surf. Sci., 1998, 401, 344
- 9R. J. Behm,, V. Penka,, M.-G. Cattania,, K. Christmann and G. Ertl, J. Chem. Phys., 1983, 78, 7486
- 10M.-G. Cattania,, V. Penka,, R. J. Behm,, K. Christmann and G. Erti, Surf. Sci., 1983, 126, 382
- 11O. M. Ilinitch,, F. P. Cuperus,, V. V. Gorodetskii,, M. Yu. Smirnov,, O. P. Burmatova and I. O. Ilinitch, Proc. 4th Workshop “Optimisation of Catalytic Membrane Reactor Systems, European Science Foundation, 1997, SINTEFF Materials Technology, Oslo, p. 89