Direct Bandgap Group IV Epitaxy on Si for Laser Applications

TLDR: In this article, chemical vapor deposition of direct bandgap GeSn alloys with a high Γ- to L-valley energy separation and large thicknesses for efficient optical mode confinement is presented and discussed.

Abstract: The recent observation of a fundamental direct bandgap for GeSn group IV alloys and the demonstration of low temperature lasing provide new perspectives on the fabrication of Si photonic circuits. This work addresses the progress in GeSn alloy epitaxy aiming at room temperature GeSn lasing. Chemical vapor deposition of direct bandgap GeSn alloys with a high Γ- to L-valley energy separation and large thicknesses for efficient optical mode confinement is presented and discussed. Up to 1 μm thick GeSn layers with Sn contents up to 14 at. % were grown on thick relaxed Ge buffers, using Ge2H6 and SnCl4 precursors. Strong strain relaxation (up to 81%) at 12.5 at. % Sn concentration, translating into an increased separation between Γ- and L-valleys of about 60 meV, have been obtained without crystalline structure degradation, as revealed by Rutherford backscattering spectroscopy/ion channeling and transmission electron microscopy. Room temperature reflectance and photoluminescence measurements were performed to ...

Figures

Figure 3. (a)-(f) AFM images of the surface of differently thick Ge0.875Sn0.125 layers. Surface rms roughness dependence on the (g) layer thickness and (h) growth temperature for thick layers (800 nm - 1 µm).

Figure 5. XTEM micrographs of a GeSn layer: (a) overview, (b) surface, (c) and (d) crystallinity and diffraction pattern from the layer volume, (e) interface between GeSn and the Ge-VS, (f) defects near the interface, (g) the Burgers vector associated with a Lomer dislocation.

Figure 1. (a) RBS random and channeling spectra for thick GeSn layers with Sn concentrations of 5%, 10% and 14%. Inset: Detail of random and channeling Sn signal for a 1 µm thick Ge0.86Sn0.14 sample. (b) Min. channeling yield ぬmin of thick GeSn alloys with different Sn contents. (c) Incorporated Sn content in GeSn alloys as a function of growth temperature.

Figure 7. Electronic band structure calculation and directness, 〉E = E-EL for Ge0.875Sn0.125 layers as a function of compressive strain (lower scale) and degree of strain relaxation (upper scale).

Figure 4. X-Ray Diffraction Reciprocal Space Maps associated with (a) a 970 nm Ge0.875Sn0.125 layer and (b) various thickness Ge0.875Sn0.125 layers. (c) Strain relaxation (top) and residual compressive strain (bottom) versus GeSn layer thickness. (d) RBS channeling measurements along (011) for 46 nm thick, fully pseudomorphic and 705 nm thick, partially relaxed Ge0.875Sn0.125 layers.

Figure 2. (a): The relation between the Sn content in GeSn binary alloys as a function of mean growth rate and temperature.(inset): Arrhenius plot of the GeSn growth rate as a function of the reverse absolute temperature and the extracted activation energy. (b) GeSn mean growth rate a 350°C as a function of layer thickness (and thus growth duration).

Full text

This is a repository copy of Direct Bandgap Group IV Epitaxy on Si for Laser Applications.
White Rose Research Online URL for this paper:
http://eprints.whiterose.ac.uk/89359/
Version: Accepted Version
Article:
Von Den Driesch, N, Stange, D, Wirths, S et al. (8 more authors) (2015) Direct Bandgap
Group IV Epitaxy on Si for Laser Applications. Chemistry of Materials, 27 (13). 4693 -
4702. ISSN 0897-4756
https://doi.org/10.1021/acs.chemmater.5b01327
eprints@whiterose.ac.uk
https://eprints.whiterose.ac.uk/
Reuse
Unless indicated otherwise, fulltext items are protected by copyright with all rights reserved. The copyright
exception in section 29 of the Copyright, Designs and Patents Act 1988 allows the making of a single copy
solely for the purpose of non-commercial research or private study within the limits of fair dealing. The
publisher or other rights-holder may allow further reproduction and re-use of this version - refer to the White
Rose Research Online record for this item. Where records identify the publisher as the copyright holder,
users can verify any specific terms of use on the publisher’s website.
Takedown
If you consider content in White Rose Research Online to be in breach of UK law, please notify us by
emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request.

1
Direct bandgap Group IV epitaxy on Si for laser applications
N. von den Driesch
1*
, D. Stange
1
, S. Wirths
1
, G. Mussler
1
, B. Holländer
1
, Z. Ikonic
2
, J.M. Hartmann
3
, T.
Stoica
1

, S. Mantl
1
, D. Grützmacher
1
and D. Buca
1*
1
Peter Grünberg Institute 9 (PGI-9) and JARA-Fundamentals of Future Information Technologies
(JARA-FIT), Forschungszentrum Juelich, 52425 Juelich, Germany.
2
Institute of Microwaves and Photonics, School of Electronic and Electrical Engineering, University of
Leeds, Leeds LS2 9JT, United Kingdom.
3
CEA, LETI, MINATEC Campus, F-38054 Grenoble, France and Univ. Grenoble Alpes, F-38000
Grenoble, France.
*Corresponding Authors: n.von.den.driesch@fz-juelich.de ; d.m.buca@fz-juelich.de

2
Abstract
The recent observation of a fundamental direct bandgap for GeSn group IV alloys and the
demonstration of low temperature lasing provide new perspectives to the fabrication of Si photonic
circuits. This work addresses the progress in GeSn alloy epitaxy aiming at room temperature GeSn
lasing. Chemical vapor deposition of direct bandgap GeSn alloys with a high - to L-valley energy
separation and large thicknesses for efficient optical mode confinement is presented and discussed. Up
to 1 µm thick GeSn layers with Sn contents up to 14 at.% were grown on thick relaxed Ge buffers, using
Ge
2
H
6
and SnCl
4
precursors. Strong strain relaxation (up to 81 %) at 12.5 at.% Sn concentration,
translating into an increased separation between - and L-valleys of about 60 meV, have been obtained
without crystalline structure degradation, as revealed by Rutherford backscattering/ion channeling
spectroscopy and Transmission Electron Microscopy. Room temperature transmission/reflection and
photoluminescence measurements were performed to probe the optical properties of these alloys. The
emission/absorption limit of GeSn alloys can be extended up to 3.5 µm (0.35 eV), making those alloys
ideal candidates for optoelectronics in the mid-infrared region. Theoretical net gain calculations indicate
that large room temperature laser gains should be reachable even without additional doping.
1. Introduction
Silicon and Germanium, the most commonly used group IV semiconductors, are indirect bandgap
semiconductors with the conduction band minimum at the X- or L-points, respectively, and not at the
center of the Brillouin zone, as is the valence band maximum. Electron-hole pair recombination hence
usually requires phonons in order to preserve momentum. As a consequence of this second order
recombination process, the radiative recombination probability is significantly lower than in direct
recombination processes, occurring in direct bandgap semiconductors like III-V alloys. This is the main
obstacle to the fabrication of efficient group IV light emitters, especially lasers. Despite this
fundamental limitation, progress has recently been made concerning the luminescence efficiency of

3
strain-engineered Ge by inducing high biaxial
1–3
or uniaxial
4,5
tensile strain. Tensile distortion of the
Ge lattice modifies the electronic band structure of Ge by reducing the energy difference between the -
and L-valleys. Therefore, a large electron population in the -valley becomes available for efficient
direct recombination. An alternative approach is the heavy n-type doping of Ge in order to fill up the
electronic states in the L-valley up to the -valley
6
.
The successful realization of a fundamental direct bandgap group IV semiconductor was achieved by
alloying Ge with Sn. The substitutional incorporation of Sn atoms into the cubic Ge lattice has a similar
effect as tensile strain in Ge: it reduces the energy separation between the - and L-valley, eventually
swapping their positions. For cubic GeSn crystals this transition to a fundamental direct semiconductor
occurs at a Sn concentration of approx. 9 at.%
7
.
The epitaxial growth of GeSn binaries on Si or Ge results naturally in a large compressive strain, which
increases the -L energy separation. It may even lead to an indirect behavior in a compressively strained
alloy, which would otherwise be direct if fully strain relaxed. In order to compensate for the effect of
compressive strain, larger Sn contents are required. Values up to 17 at.% for GeSn alloys grown
pseudomorphically on Ge substrates have previously been shown
8
. A practical approach towards
fundamental direct bandgap GeSn at experimentally achievable Sn contents is the realization of strain
relaxed alloys. However, for active laser materials both, a high degree of strain relaxation in thick layers
together with a low defect density is of paramount importance.
The synthesis of fully strain relaxed GeSn epilayers with device grade quality is a demanding task,
mainly due to the metastable nature of this alloy. The low solid solubility of Sn in Ge of approximately
1 at.%
9
and the large lattice mismatch of 15% between -Sn and the Ge substrate are major hurdles that
have to be overcome. Thermal annealing - the most common technique for plastic strain relaxation - is
not applicable for Sn-based systems, as severe Sn diffusion leads to poor layer quality
10
. The most
efficient way, in our opinion, to fabricate high quality, partially strain relaxed epilayers is to grow
several hundred nanometer thick GeSn layers on Ge virtual substrates (Ge-VS). Significant progress has

4
been made in recent years in epitaxy of these alloys by Chemical Vapor Deposition (CVD)
11–13
and
Molecular Beam Epitaxy
14,15
. However, the low growth temperatures required for Sn incorporation still
lead to an epitaxial breakdown in thick layers due to a strong increase of surface roughness for Sn
concentrations > 10 at.%
14
. Very recently, the ability to grow ~500 nm thick GeSn layers with Sn
concentrations of approx. 12.5 at.% Sn has enabled the demonstration of fundamental direct bandgap
of these alloys and even lasing
7
. This breakthrough has brought about a new class of materials in
group IV, which can revolutionize the Si photonics
16
.
In this work we present details on the epitaxial CVD growth and in-depth structural and optical
characterization of direct bandgap GeSn alloys aiming for laser fabrication. Most emphasis is placed on
binaries with thicknesses up to 1 µm for a complete optical mode confinement and Sn concentrations
between 8 and 14 at.% which offer sufficient -L energy difference for efficient laser emission. Atomic
Force Microscopy (AFM) for surface morphology, Rutherford backscattering spectroscopy/ion
channeling (RBS/C) and cross-sectional Transmission Electron Microscopy (TEM) for crystallinity
together with X-Ray Diffraction (XRD) for plastic strain relaxation were used in order to gain insights
into the growth mechanism and the relaxation process of those thick GeSn layers. The optical properties
were assessed using room-temperature transmission/reflection and photoluminescence measurements.
The increase of the Sn content to 14 at.% leads to a reduction of the fundamentally direct bandgap
below 0.4 eV extending the window for optical applications above 3 µm. Furthermore, net gain
calculations show that the GeSn layers are suitable for room-temperature laser emission in
heterostructures, without the necessity of additional doping, as suggested for the tensily strained Ge
approach
6
.
2. Experimental Part
The growth of the GeSn layers was performed on cyclically annealed, 2.5 – 2.7 µm thick Ge
virtual substrates (Ge-VS) on Si(100)
17
. A 200 mm AIXTRON TRICENT
®
reduced-pressure CVD

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "Direct bandgap group iv epitaxy on si for laser applications" ?

The copyright exception in section 29 of the Copyright, Designs and Patents Act 1988 allows the making of a single copy solely for the purpose of non-commercial research or private study within the limits of fair dealing. The publisher or other rights-holder may allow further reproduction and re-use of this version refer to the White Rose Research Online record for this item. 

Q2. How did the spectroscopy reveal the crystalline structure of GeSn?

Sn concentration, translating into an increased separation between d- and L-valleys of about 60 meV, have been obtained without crystalline structure degradation, as revealed by Rutherford backscattering/ion channeling spectroscopy and Transmission Electron Microscopy. 

Q3. How much relaxation can be observed in the second GeSn layer?

Plastic relaxation of the GeSn alloy can be observed in the second, 172 nm thick GeSn layer, with adegree of relaxation of about 39%. 

Q4. What was used to determine the properties of GeSn alloys from those of elemental Ge?

A Vegard’s Law (with bowing if available) was used in order to determine the properties of GeSn alloys from those of elemental Ge or Sn. 

Q5. What is the effect of a direct bandgap on the light emission efficiency?

If a direct bandgap is a necessary requirement for lasing, a larger directness, i.e. through increase of a Sn content or strain relaxation, is expected to increase the light emission efficiency. 

Q6. What are the major hurdles that have to be overcome?

The low solid solubility of Sn in Ge of approximately 1 at.% 9 and the large lattice mismatch of 15% between g-Sn and the Ge substrate are major hurdles that have to be overcome. 

Q7. How can the emission limit of GeSn be extended?

The emission/absorption limit of GeSn alloys can be extended up to 3.5 µm (0.35 eV), making those alloys ideal candidates for optoelectronics in the mid-infrared region. 

Q8. What is the reason for the higher gain in the 14 at.% sample?

The reason for the slightly highergain in the 14 at.% sample is mainly the higher residual compressive strain, which makes the HH-LH splitting in it twice as large as in the 12.5 at.% sample, amplifying the fraction of ‘useful’ holes significantly. 

Q9. What is the energy of the peaks in the reflection spectra?

PL peaks are at the same energy as the onset of strong absorption regions in the reflection spectra, evidencing band-to-band recombination in direct bandgap alloys. 

Q10. What is the effect of the sn layer on the PL?

Tuning the directness of their layers (by increasing the Sn content or relaxing the residual compressive strain thanks to higher thicknesses), led to an even stronger PL response, demonstrating the potential of their GeSn layers as light-emitting devices. 

Q11. What is the effect of the recombination rate of direct energy transitions?

The higher spontaneous radiative recombination rates of direct energy transitions lead to a remarkable PL emission increase in the thick 12.5 at.% 

Q12. What is the effect of the increased rms roughness on the surface of the alloy?

In addition,the decreased Sn incorporation in the alloy for elevated growth temperatures (see Fig. 1(b)) reduces the strain field at the interface and subsequently the surface undulation. 

Q13. What is the angular shift of the GeSn layer?

The measured angular shift of 〉し[011] = -0.82° corresponds here to a tetragonal strain of -2.9 % in the 46 nm GeSn layer as theoretically expected for full pseudomorphic growth on a Ge-VS. 

Q14. What is the efficient strain releasing type of dislocations?

These so-called Lomer dislocations have a Burgers vector parallel to the interface and are, therefore, the most efficient strain releasing type of dislocations.