Extracting surface recombination parameters of germanium–dielectric interfaces by corona-lifetime experiments

Abstract

The interest in germanium (Ge) is rising for use in field-effect transistors, (space) photovoltaics, and silicon photonics. Suppressing and understanding carrier recombination at the Ge surface are vital for the performance of Ge in these applications. In this work, we have investigated the surface recombination at various germanium–dielectric interfaces (Ge/Al2O3, Ge/SiNx, Ge/GeOx/Al2O3, and Ge/a-Si:H/Al2O3). For this purpose, we performed corona-lifetime experiments and extracted a set of recombination parameters by fitting the data with the theoretical Girisch model. To keep the model straightforward, the distributions of the capture cross sections and the interface defect density [Formula: see text] were parameterized. The importance of each parameter in these distributions was examined so that a minimum number of parameters was distilled: the so-called fundamental recombination velocities ([Formula: see text] and [Formula: see text]) and the magnitude of the [Formula: see text] near the valence and conduction band edge ([Formula: see text] and [Formula: see text]). These parameters form together with the fixed charge density [Formula: see text], the spatial distribution thereof [Formula: see text], and a minimum surface recombination velocity [Formula: see text], a set of parameters that can well describe our experimental data. Relevant insights were obtained from the experiments, with highlights including a Ge/GeOx/Al2O3 stack with virtually no fixed charge density, a highly passivating Ge/a-Si:H/Al2O3 stack, and a negatively charged Ge/SiNx stack. The findings in this study are valuable for applications where a more profound understanding of recombination at Ge surfaces is of concern, such as in photonics, photovoltaics, and nano-electronics.

Comparison of SiNx‐Based Surface Passivation Between Germanium and Silicon

“…This indicates the presence of a negative Q tot (in the range of 10 12 cm À2 ) within the thin film. A similar observation (i.e., a negative Q tot within SiN x film on the Ge surface) has been reported recently by Berghuis et al [10] The PDA seems to have almost no impact on the trend of SRV max of Si/SiN x samples as a function of Q c : both the polarity and magnitude of Q tot remain similar which is in agreement with previous studies. [24] In contrast, for the Ge/SiN x samples, the charge polarity remains the same but the SRV max behaves differently as a function of Q c after the PDA, indicating changes in the magnitude of Q tot .…”

“…This indicates the presence of a negative Q tot (in the range of 10 12 cm −2 ) within the thin film. A similar observation (i.e., a negative Q tot within SiN x film on the Ge surface) has been reported recently by Berghuis et al [ 10 ]…”

“…via PECVD SiN x have also been reported, but the data available is still rather limited. [10,11] It is rather common to enhance the passivation of Si surfaces by capping the PECVD films with another thin film, such as Al 2 O 3 . During a subsequent anneal, the Al 2 O 3 capping layer acts both as a hydrogen source as well as a hydrogen diffusion barrier, which enables efficient hydrogen diffusion to the Si-thin-film interface and consequently a high level of chemical passivation.…”

“…[ 9 ] Some preliminary results on the passivation of Ge via PECVD SiN x have also been reported, but the data available is still rather limited. [ 10,11 ]…”