Band bending analysis of charge characteristics at GeO2/Ge interface by x-ray photoemission spectroscopy

Wenfeng Zhang,Xun Lou,Zijian Xie,Haixin Chang
DOI: https://doi.org/10.1088/1361-6463/aaed00
2019-01-01
Abstract:Ge complementary metal oxide semiconductor (CMOS) is promising for scaling beyond the Si CMOS due to its higher carrier mobility than Si. Analogue to classical SiO2/Si system in the Si CMOS technology, various interface/bulk charges including interface traps (Q(it)), fixed surface state charges (Q(f)), trapped positive charges (Q(pt)) and negative charges ((Q(nt)) in GeO2/Ge system are also crucial both for the device performance and reliability. Because small amounts of charges would cause photoemission peak shift characterized by x-ray photoemission spectroscopy (XPS), it offers a feasible way to evaluate various charge densities by measuring the band bending in Ge substrate from Ge 3d core-level energy shift at GeO2/Ge interface. Moreover, photoemission peak shifts as a function of x-ray irradiation time have been widely accepted for characterization of charge trapping phenomena. Here, we report a band bending analysis at GeO2/Ge interface of featuring vital charge characteristics for diverse device applications by XPS. HF-last cleaned Ge surface was verified to tend to be p-type, irrespective of the bulk conductivity. The n-Ge/GeO2 interfaces exhibit a reduction of upward band bending evolution of Ge substrate, while p-type-Ge/GeO2 interfaces indicate a reduction of downward band bending evolution when comparing the different quality GeO2/Ge interfaces. Based on the requirement of surface charge neutrality, such observation has been attributed to a dominated passivation effect to negatively charged interface traps and the positive fixed surface state charges, respectively. Moreover, a time evolution of Ge 3d and O 1s signals reveals a progressive band bending modification at GeO2/Ge interface, clarifying the thermally-grown GeO2 also contains electron traps (Q(nt)). Ultimately, the four types of charges relying on the GeO2/Ge quality were modeled to correlate with the observed Ge band bending evolution, which would impact both the device operation and reliability.
What problem does this paper attempt to address?