Active dopant characterization methodology for germanium

Abstract

In order to reach the ITRS goals for future complementary metal-oxide semiconductor technologies there is a growing interest in using germanium as an alternative substrate material in view of its higher mobility. Different species and thermal budgets are presently being investigated in order to determine the most likely candidates for the required junction formation. A key issue is the accurate determination of the achievable electrical activation, i.e., the reliable measurement of the sheet resistance and electrical depth profile. In order to be applicable to Ge-based junctions, standard techniques such as the spreading resistance probe and scanning spreading resistance microscopy (SSRM) need to be reevaluated in terms of their performance and operational conditions. First, the significantly different behavior of germanium calibration curves (versus silicon) will be discussed. Next, the shape and characteristics of the probe imprints (Ge is softer than Si) and the differences in raw data behavior will be covered, such as the presence of a junction cusp for n+-p, but not for p+-n diodes. A preliminary germanium contact model will be presented, supporting all of the present experimental data (including SSRM current-voltage curves) based on the impact of surface states. Finally, the accuracy of the quantified resistivity and carrier depth profiles using Laplace-based modeling will be evaluated in comparison with alternative low penetration sheet resistance measurements, such as variable probe spacing, and secondary-ion-mass spectroscopy dopant profiles in the case of 100% activation.