Abstract. Spin labelling techniques, specifically the use of electron-spin-polarized He(2 3S) metastable atoms coupled with energy-resolved spin analysis of the ejected electrons, are used to investigate the dynamics of He(2 3S) deexcitation at solid surfaces. Data for a clean Au(100) surface are presented that show that deexcitation occurs exclusively through resonance ionization followed by Auger neutralization. The electrons involved in Auger neutralization are observed to be correlated in spin and possible reasons for this are discussed. Results obtained at Xe and NO films adsorbed on cooled Au(100) and Cu(100) substrates, respectively, show that He(2 3S) metastable atom deexcitation is analogous to gas-phase Penning ionization. Detailed differences are apparent that can be attributed to effects associated with the underlying substrate and interactions involving neighboring atoms in the film. 79.20.Nc; 79.20.Rf; 79.80.tw Recent work in this laboratory has shown that spin-labelling techniques provide a powerful means to investigate the dynamics of reactions involving He(2 3S) atoms. This approach complements earlier studies of reaction dynamics that were based solely on analysis of the energy distributions of product electrons. Here we focus on the use of spin-labelling techniques to probe the dynamics of He(23S) atom deexcitation at clean and adsorbatecovered surfaces [1,2]. Conventional models [3] suggest that He(2 3S) atoms will be deexcited at clean high-workfunction metal surfaces exclusively by the two-step process diagrammed in Fig. la. An incident 2 3S atom first undergoes resonance ionization (RI) in which the excited 2s electron tunnels into a vacant level above the Fermi surface. The resulting He + ion continues toward the surface where it is neutralized by a conduction electron from the metal. The energy released is communicated to a second electron which, if the energy transfer is sufficient, may be ejected from the surface. This Auger neutralization (AN) process results in a relatively structureless ejected electron energy distribution that reflects, approximately, a self convolution of the local density of electronic states at the surface. At low work function surfaces RI cannot occur because, as illustrated in Fig. lb, there are no unfilled levels of appropriate energy available. The 2 3S atoms are then deexcited by an Auger deexcitation (AD) process in which an electron from the metal tunnels into the 2 3S 1 s core hole. The energy released is communicated to the excited 2s electron which may be ejected. Since AD is a quasi-one-electron process, the ejected electron energy distribution reflects directly the local density of electronic states and can contain relatively sharp features. The presence of adsorbed layers on a surface can also impede RI by preventing good overlap between the 2s electron wavefunction and vacant levels in the metal. Deexcitation may again occur by AD and, because the electron that tunnels into the 2 3S 1 s core hole originates in the adsorbate layer, the process is ...