The presence of a gas in the specimen chamber of low-vacuum scanning electron microscopes (SEMs) gives rise to a number of interactions that are not found in their high-vacuum counterparts. Many of these interactions are integral to the specimen charge neutralizing capabilities of these instruments. However, these interactions, and the electronic charge state of the specimen, give rise to a number of processes that influence, and sometimes even dominate, secondary electron contrast. As many of the processes are stochastic in nature, and depend on discrete interactions between electrons, gaseous ions, neutral molecules, and the specimen surface, modeling can offer important insights for contrast interpretation. This paper discusses these interactions and the mechanisms through which contrast is influenced. Copyright
MODELING OF IMAGING PROCESSES IN THE LOW-VACUUM SEMLow-vacuum scanning electron microscopy is emerging as a powerful new technique for the study of poorly conducting materials and other difficult samples.1 -3 Because of these attributes, it is also being considered as a candidate for the next generation of sub-10 nm critical dimension metrology tools for the semiconductor manufacturing industry. 4,5 Widespread acceptance of the technique is hindered because the image formation process is considerably more complicated than in high-vacuum scanning electron microscopes (SEMs). Interactions between electrons and gas molecules, as well as the behavior of positive gaseous ions, all influence the appearance and information content of secondary electron (SE) images. Furthermore, SE images obtained from dielectric materials appear to contain more information than their high-vacuum counterparts, related to the local dielectric properties of the material. 3,6 Considerable work needs to be done before it will be possible to make routine interpretations of the SE contrast. Many of the processes that influence the information content of SE images have complicated nonlinear interdependencies, which make simple analytic descriptions unlikely. The situation is exacerbated by the commercial introduction of detectors with increasingly complicated physical and electromagnetic field geometries. However, since many of the processes are stochastic in nature, computer modeling can offer insights. The processes that contribute to image formation can be divided into three categories: cascade amplification/ion production, intrinsic SE emission from dielectrics, and modulation of SE emissions by the gaseous ions. Most commercial low-vacuum SEMs use some degree of gas cascade amplification to amplify the secondary electron emission current, regardless of the form of the signal that is actually measured. Another critical function of the cascade is to produce a flux of positive ions sufficient to stabilize electronic charging of insulating samples. In its simplest form, the gas cascade is manipulated by placing a positively biased anode (typically a few hundred volts positive) in the vicinity of the specimen. Secondary electrons ...