One of the ten most important challenges facing the semiconductor industry is to obtain an accurate quantitative two-dimensional dopant profiling technique with high spatial resolution. Secondary electron (SE) imaging in the scanning electron microscope (SEM) has been shown to be a very promising technique for dopant profiling in the past. However, the limited accuracy obtained with SE imaging (> ±20% dopant sensitivity) has hindered this technique from being used for quantitative studies.

In this dissertation, we develop the novel technique of SE energy-filtering in a field-emission gun SEM (FEG-SEM). We demonstrate that energy-filtered imaging can be used to increase the prediction accuracy of dopant concentrations (~8.5 % dopant sensitivity) because it allows one to measure the shift in SE energy distributions that are caused by the potential difference across the pn-junction. The surface potential difference across a pn-junction is a function of the dopant concentration change across the device, hence enabling us to quantify the dopant concentration. This method provides much more accuracy and reproducibility compared to conventional SE imaging where often all SEs are allowed to be detected. In addition, we explore the spatial resolution limits of dopant profiling with the FEG-SEM.

The effect of some of the imaging conditions on the dopant contrast (that is SE yield differences) observed from SE images has been examined, and recommendations for optimum conditions for the quantification of dopant profiling are given. Also, the potential of site-specific sample preparation for dopant profiling in the SEM using a focused ion beam (FIB) workstation has been explored. Contrast was observed in SE images in the SEM after FIB milling. These experimental results have provided insight into some of the SE emission mechanisms responsible for the observed dopant contrast.