appeared in UVSOR Activity Report, UVSOR Facility, Institute for Molecular Science, UVSOR-23 (1995) 164.
The process of the Desorption Induced by Electronic Transitions (DIET) in rare gas solids (RGS) has been extensively studied in these 10 years [1]. RGS is a model system to investigate the dynamic processes of DIET because of its simplicity and similarity of the electronic structure to the isolated atom.
The processes of neutral atom desorption by the excitonic excitation from pure RGS can be explained using two mechanisms, excimer dissociation (ED) and cavity ejection (CE) [2] . The desorption via ED process is due to a dissociation of a molecular-type self-trapped exciton (m-STE) similar to the dissociation of an excited dimer (excimer) in the gas phase. As to the CE mechanism, negative electron affinity of the matrix is known to be essential to have a repulsive interaction between the excited atom and the surrounding ground state atoms. This corresponds well to the experimental facts, i.e., the desorption species via CE process can be observed for Ne and Ar solids, whose electron affinities are negative, but not for Kr and Xe solids because of their positive electron affinities in the bulk due to the large polarizabilities. Further evidence has been obtained by using adsorbed and mixed rare-gas systems. Weibel et al. [3] measured Ne metastable desorption from thin layer of Ne on bulk Ar, Kr and Xe. They concluded that the CE mechanisms still worked in such adsorbed systems. Runne et al. [4] revealed the systematic correlation between the electron affinity and the existence of the desorption atoms via CE process for pure and doped rare-gas solids.
The experiments have been carried out at the beam line BL-5B at UVSOR in Institute for Molecular Science, Okazaki. Details of the setup and the experimental procedure have been already published elsewhere [5, 6] . Briefly, a liquid He cryostat is installed in an UHV chamber (base pressure < 10-8Pa). A Pt(111) crystal is attached to the cryostat, and the temperature is kept at about 6K or less. The thickness of the rare gas solid is about 800 layers or more. Monochromatized synchrotron radiation is pulsed using a mechanical chopper, whose width and the interval of the light are 15µsec and 2.5msec, respectively. The photon beam is incident at 30 deg from the normal direction of the sample surface.
The desorbed metastable atoms (Ne*, 2p53s 3P0,2) are detected by a MCP with 40mm diameter (Galileo Electro-Optics Co.). Charged particles are rejected by applying suitable potentials on the input surface of the MCP and the retarding grids in front of the MCP. The detector is fixed at the distance of 358mm from the sample in the normal direction of the sample. The distance is about 5 times longer than our previous experimental setup, which results in the improvement of the resolution in the time-of-flight (TOF) spectra.
The dependence of Ne* (CE) yield from pure Ne solid on the excitation photon energy is shown in fig.1. The position of each peak corresponds to the excitation energy of the surface (S1, S') and the bulk (B1, B2, B3) excitons. The time-of-flight and kinetic energy spectra measured at the incident photon energies of 17.1eV (S1), 17.6eV (B1), 19.0eV (S') and 20.3eV (B2) are shown in fig.2 and 3, respectively. Detailed analysis of these spectra have already been given [6] . Present TOF spectrometer with the flight length longer than the previous one, which results in better time resolution, enabled us to obtain information on the desorption spectra in more detail. The strongest peak at time zero in fig.2 is due to scattered and emitted light. The higher kinetic energy peak (tf Å 100 µsec, Ek = 1.4 ± 0.1 eV) is due to the excimer dissociation (ED) process (not shown in fig.3). The present kinetic energy is in good agreement with our previous ESD results [7] using a channel electron multiplier with small acceptance angle, while it is significantly larger than our previous PSD results (Ek=1.15eV) using a MCP with large acceptance angle (61 deg.) and short flight length [6]. This discrepancy can be attributed to large uncertainty of flight time in the previous PSD experiments for ED peak because of its very broad angular distribution.
The kinetic energies of CE peaks (tf Å 250 µsec) for S1, B1, and B2 excitation are found to be the same within the experimental uncertainty (Ek = 0.18±0.02 eV), while that for S' excitation, Ek = 0.20±0.02 eV. This fact may suggest that the desorbed Ne*(CE) at S1, B1, and B2 excitation are, at least when leaving from the surface, all in the same electronic states, probably in 2p53s (3P0,2). The initial electronic state of B2 exciton (2p54s) is known to relax very rapidly (Å 10-13 sec) to the 1st order excitonic state [8] , which is consistent with the discussion above. Additional shoulder appeared in the higher energy side in B2 spectrum (tf Å 180 µsec), whose kinetic energy is 0.36±0.04eV, can be the contribution of Ne* in 2p54s state at the desorption. The Ne*(CE) at S' excitation is in 2p53p state at the desorption, and decay to 2p53s state in vacuum [9] .