Surface-suppressed electron resonance spectroscopies

Abstract
Surface‐suppressed electron resonance spectroscopies (SSERS) refers to the phenomenon in which stable paramagnetic radicals adsorbed on clean (noble) metal surfaces have their ESR signal suppressed. This phenomenon is studied in some detail by ultra‐high‐vacuum ESR (UHV‐ESR) and cyclotron resonance from microwave‐induced secondary electron emissions (CREMSEE) in combination with conventional thermal desorption. The UHV‐ESR is performed in situ on the inner surface of the microwave cavity after leaking in stable nitroxides as previously described by Nilges and Freed. In the SSERS phenomenon, the first layers of nitroxide deposited on the metal surface do not give rise to an observable ESR signal, even though a significant decrease in the CREMSEE microwave power threshold, Pt, is observed. It is consistently found (for several nitroxides and noble metals) that an ESR signal is observed only when Pt has dropped to of the order of 20 (±10)% of its initial value. The ESR signal then increases monotonically with increased dosage. Also, a correlation between these measured reductions in Pt, and the estimated surface coverages required for initial observation of the ESR signal, is suggested by our results. It may be that SSERS is (partly) due to the first layer of nitroxide interacting strongly by exchange forces to the surface conduction band of the metal, so that the ESR signal is too broadened to be observed. Subsequent nitroxide layers may also be affected through their interaction with previous layers by weaker exchange forces. This is consistent with experiments in which a related but diamagnetic species is utilized to form the initial adsorbed layers, and it is found to act as an ‘‘insulator’’ for the subsequent nitroxide layers. (On the other hand, surfaces pretreated with O2 or O2/H2O mixtures had very little effect on SSERS observed with Ag surfaces, although it had some effect, in the sense of a weak insulator, with Cu surfaces.) The change in ESR signal upon warming was correlated with the observed pressure changes. In some cases there are unusual nonmonotonic variations of the ESR signal strength, inconsistent with observed desorption of nitroxide, that are also believed to be due to SSERS. The possible role of temperature‐dependent surface wetting effects is briefly considered.