Depth-resolved cathodoluminescence in undamaged and ion-implanted GaAs, ZnS, and CdS

Publisher Website
Google Scholar
Abstract
Here we report a variety of results obtained by using sequences of luminescence spectra excited by 1–20‐keV electron irradiation to carry out in situ studies of depth‐dependent optical activity in luminescent crystals. Data are shown for various samples subjected to localized damage from ion implantation: GaAs implanted with Cu+, ZnS implanted with Ar+ and Cu+, and CdS implanted with Ar+ and H+. Semiquantitative interpretation of the results shows that the depth‐resolved cathodoluminescence measurements can have unique value in characterizing the effects of ion‐implantation lattice damage. In this case cathodoluminescence can be excited from depths ranging from much shallower to much deeper than typical implant depths. In addition, the use of depth‐resolved measurements on nominally undamaged ZnS crystals reveals the presence of weak near‐surface luminescence bands despite careful surface preparation. This result makes it clear that luminescent center profiling by layer removal methods can lead to erroneous results when recurring spectral features result from inherent near‐surface conditions. More importantly, in several ZnS and CdS samples we find that unexpectedly sharp near‐surface depth resolution (of the order of several hundred Å) can be obtained even though bulk carrier diffusion lengths are of the order of microns. While this effect is presently not understood in detail, the result should permit depth‐resolved measurements on sputtering damage and very shallow implants. The various samples we have studied show that modifications in spectrum from injection‐level effects (particularly GaAs) or from exposure to electron irradiation (particularly CdS) may appear in depth‐resolved or other luminescence measurements. Finally, the depth‐resolved cathodoluminescence measurements are compared with single‐wavelength photoluminescence spectra on the same samples for cases where the photoexcitation light is or is not strongly absorbed. It is pointed out that the cathodoluminescence measurements are more adaptable to in situ depth profiling because the irradiation energy (excitation depth) and current (injection level) can be easily varied over wide ranges, and because the excitation depth is not affected by the optical absorption coefficient (which is often uncertain, particularly in implanted layers). In addition, we find that cathodoluminescence can be more sensitive than photoluminescence for studying optical centers introduced by shallow implants, even when the sample is strongly absorbing to the photoexcitation light.

This publication has 14 references indexed in Scilit: