Improved Defect and Fourier Transform Infrared Spectroscopy Analysis for Prediction of Yield for HgCdTe Multilayer Heterostructures

Abstract

The ability to accurately predict HgCdTe focal plane array (FPA) performance using nondestructive, postgrowth wafer analysis is of great importance. These predictions, if accurate, reduce costs by screening the wafers prior to processing, and selecting only those wafers that are most likely to yield FPAs that meet program specifications. In this paper, we examine the use of a macrodefect inspection tool, the NSX 1255, from August Technology. This inspection tool has the ability to measure defects 0.5 μm and larger and store the location and size data to a file. We have then, through the use of custom written software, been able to analyze these data on a wafer by wafer basis. We have also incorporated the use of a thin film transmission matrix model to analyze room-temperature Fourier transform infrared spectroscopy (FTIR) transmission spectra. This technique, which is applied to the entire wafer surface, can be used to determine the individual layer thicknesses as well as their compositions. Then, using analytical expressions for bandgap, absorption, and index of refraction, we can predict responsivity and quantum efficiency. Through the use of these two inspection tools and our analysis software, we are able to overlay FPA die information and perform statistics on a die-per-die basis. This allows us to effectively “pass” or “fail” each FPA based on the program specifications. We are then able to set a minimum criterion for the number of FPAs that pass on any given wafer. That wafer is then sent off to processing if it meets this criterion. Furthermore, knowing why a wafer fails before it reaches processing allows for real time feedback to the epilayer growth process. This allows for run-to-run adjustments in order to keep as many wafers within specifications as possible and increases yield overall.