Energy- and angle-resolved measurements of the Rh(4F9,2) and Rh(4F7,2)

The measurement of the energy and the angular distributions of excited atoms desorbed from ion-bombarded solids has been of long-standing interest since these distributions hold the key to understanding the excitation and deexcitation processes. Although many investigators have experimentally determined the energy distributions of atoms ejected in various excited states, all these measurements have been confined to a single angle of emission.’ Corresponding theories and models have thus concentrated on the velocity dependence of the excitation probability. In 1954, Hagstrum proposed that for excited atom fractions, the velocity dependence of the excitation probability should follow an exp( -A/au,) behavior, where ul is the component of escape velocity perpendicular to the surface and A/u is the deexcitation coefficient.’ This relation has been used to describe experimental distributions except for the low velocity regime whe’re surface binding energy effects are suggested to alter this dependence.3-5 In this Communication, we present energyand angleresolved neutral (EARN) distributions of excited Rh atoms (4F7,2 state, with excitation energy of 0.2 eV) ejected from Rh( 100) by bombardment with a 5 keV Ar + ion beam at normal incidence. In addition, we have measured the EARN distributions of atoms in the 4F9,2 ground state. In this way, for the first time the final excitation probability can be presented as a function of the emission velocity and the take-off angle of the particles. These results show that at high velocities, the dependence is indeed exp( --A/au,), although the value of A/a changes by over 50% depending on the polar and azimuthal angles of detection. Moreover, at lower velocities, the ratio becomes almost independent of velocity. We show that collisional excitation and the details of individual atomic motions are needed to account for the details of the velocity and angular dependences of the excitation probability. The EARN distributions have been obtained using a multiphoton resonance ionization (MPRI) scheme that has been described in detail elsewhere.6’7 A 200 ns pulse of 5 keV Ar + ions is first focused at normal incidence onto a 2 mm spot on the sample. A given time ( > 1.5 ps) after the ion impact, a ribbon-shaped laser pulse is used to resonantly ionize a small volume of the desorbed species. This time delay determines the velocity time-of-flight (TOF) of the probed species. Once the particles are ionized, they are accelerated toward a position-sensitive microchannel plate where they are detected. For a given azimuthal angle (p,), 30 images each corresponding to a different TOF are collected, and sorted into an intensity map of kinetic energies (E) and take-off angles (0). The energyand angle-resolved distributions of Rh atoms sputtered in the ground state and the next higher-lying excited state are presented in Fig. 1. The results correspond to ejection along two crystallographic directions, as defined in the inset to Fig. 1. For the ground state distribution, the most intense peak is seen along the e, = 0” azimuth ( (100) direction) at a polar angle of about 50”. The observed angular anisotropies are the same as has been reported previously for the ground state.8 The EARN distributions of excited atoms are qualitatively different from the ground state distributions. First, the most intense peak appears at normal ejection (8 = 00). Second, at low energies, only a shoulder exists along the p = 45” azimuth ( (110) direction). Third, the off-normal peak position occurs closer to the surface normal for the excited state than for the ground state. Fourth, the fall off with energy is much slower for the excited state.’ The relative populations in the 4Fg,2 and 4F7,2 states was approximately 20 to 1. Our probing of the next higherlying excited state (4F5,2 with excitation energy of -0.3 eV above the ground state) revealed its population to be at least two orders of magnitude smaller than that of the 4F7,2 state. Since the population density of a given level has been found experimentally to exponentially decrease with the excitation energy,’ we conclude that the populations in the ground and first excited state are not affected by decay in the gas phase from higher excited states. The observation that the population in the ground state is much larger than the excited states allows us to approximate the excitation probability by the ratio of the excited state to ground state distributions, (dN*/du)/ (dN/dv), rather than by the ratio of the excited state population to the total population. Since the excitation probability is generally thought to vary as exp( -A/~LJ,),*“~ we have plotted (dN*/du)/(dN/dv) vs l/v, in Fig. 2 for several representative angles of ejection. There are several features that are clear from inspection of Fig. 2. First, at high velocities the ratio exhibits the expected exp( -A/av,) dependence. However, the value of A/u varies from 0.58~ lo6 to 1.52~ lo6 cm/s. Second, at low velocities there is a sharp leveling off of the intensity ratios. The height of this plateau depends strongly on the polar and azimuthal angles of ejection. A somewhat similar behavior has been observed and attributed to the effect of surface binding energy.3-5 However, in the present investigation, the deviation at low velocities occurs much more