Speaker
Description
\documentclass[11pt,a4paper]{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{amsmath}
\usepackage[margin=2.5cm]{geometry}
\usepackage{hyperref}
\pagestyle{empty}
\setlength{\parindent}{1.5em}
\setlength{\parskip}{0.3em}
\setlength{\emergencystretch}{2em}
\begin{document}
\begin{center}
{\large\bfseries Silicon incorporation and dislocation-related traps in GeSn and GeSiSn layers on Si: an impedance and photocurrent study\par}
\vspace{1em}
A.A. Pocherpailo$^{1}$, S.V. Kondratenko$^{1}$, O.I. Datsenko$^{1}$, S.O. Krakovskyi$^{1}$, P.O. Kovanzhi$^{1}$, P.M. Lytvyn$^{2}$, Yu.I. Mazur$^{3}$, G.J. Salamo$^{3}$
\vspace{0.7em}
{\small
$^{1}$Physics Faculty, Taras Shevchenko National University of Kyiv, Kyiv, Ukraine\
$^{2}$V. E. Lashkaryov Institute of Semiconductor Physics, NAS of Ukraine, Kyiv, Ukraine\
$^{3}$Institute of Nano Science and Engineering, University of Arkansas, Fayetteville, AR, USA\[0.3em]
Corresponding author: \href{mailto:keepcalmandenjoylondon@gmail.com}{keepcalmandenjoylondon@gmail.com}\[0.3em]
\textbf{Contribution type:} Poster
}
\end{center}
\vspace{0.5em}
GeSn and GeSiSn are promising group-IV materials for Si-integrated infrared optoelectronics. Incorporating Sn shrinks the bandgap and, at sufficient content, makes it direct [1]. Because Sn is poorly soluble in Ge, these alloys must be grown far from equilibrium. This generates misfit and threading dislocations whose cores create band-like gap states that control carrier recombination [2]. Si is added to compensate the Sn-induced lattice expansion, but its effect on electrically active defects cannot be deduced from structural data alone.
We combine impedance spectroscopy, lateral photocurrent (PC) spectroscopy and surface imaging to compare two layers with identical Sn content grown on Si(001) via a Ge buffer: a 44~nm binary Ge$_{0.934}$Sn$_{0.066}$ layer and a 65~nm ternary Ge$_{0.817}$Si$_{0.117}$Sn$_{0.066}$ layer.
SEM and AFM show that both films are continuous and grew two-dimensionally. Their RMS roughness over $2\times2~\mu\mathrm{m}^2$ is 0.40~nm for GeSn and 0.31~nm for GeSiSn. Neither film shows Sn segregation, so metallic shunting paths can be ruled out. Both surfaces share a buffer-inherited undulation (period 2--3~$\mu$m, amplitude $\sim$10~nm) and 10--15~nm pits at densities of $\sim 10^{10}~\mathrm{cm}^{-2}$. GeSiSn additionally shows isolated 150--200~nm islands at $(7\pm2)\times10^{7}~\mathrm{cm}^{-2}$ ($\sim$2\% coverage), which are too sparse to affect lateral transport. The electrical responses are therefore directly comparable.
In the lateral geometry, the conductivity relaxation time $\tau_{max}$ and the dc conductance $G_{DC}$ are thermally activated at high temperatures. They follow $\exp[\pm E/(kT)]$ with $E = 90$--$120$~meV for GeSn and $\sim$60~meV for GeSiSn. Weaker low-temperature dependences are tentatively attributed to mobility variation. The binary-layer energies agree with dislocation-related levels near $E_V + (85$--$90)$~meV, the dominant Shockley--Read--Hall centres in Ge-rich GeSn [3].
The PC spectra show an onset above 0.7~eV from the Ge(Si)Sn absorption edge, a Ge-buffer contribution above 0.8~eV, and a strong Si-substrate band above 1.1~eV. The PC at 0.95~eV drops sharply above $\sim$220~K. The quenching energies, 280~meV for GeSiSn and $\sim$140~meV for GeSn, exceed the impedance energies, as expected, since PC quenching reflects competition between collection and trap-assisted recombination, whereas impedance probes thermal emission from the same states.
The lower activation energy of the ternary alloy indicates that Si does not eliminate dislocation-related traps but broadens their energy distribution.
This work was supported by the National Research Foundation of Ukraine (project 2023.03/0060).
\vspace{0.5em}
\noindent{\small
{[1]} D. Gr\"{u}tzmacher \textit{et al.}, Appl. Phys. A \textbf{129}, 1 (2023).\
{[2]} A. Giunto, A. Fontcuberta i Morral, Appl. Phys. Rev. \textbf{11}, 041333 (2024).\
{[3]} S. Gupta \textit{et al.}, Appl. Phys. Lett. \textbf{113}, 022102 (2018).
}
\end{document}
| Type of presence | Presence online |
|---|---|
| Figure caption | Temperature-dependent conductivity relaxation time (a) and dc conductance (b) of the studied structures. The lines are functions ∝ exp[±E/(kT)] |