DS0710 - Sciences et technologies des composants nanoélectroniques et nanophotoniques

Advanced aNalysis of III-V/Si nucleaTIon for highly integrated PhOtonic Devices – ANTIPODE

ANTIPODE : Photonic integration on silicon, toward the control of the III-V/Si interface

ANTIPODE is a fundamental research project which aims to deeply understand the formation of III-V/Si semiconductor interfaces in order to better control the defect generation at the interface during epitaxial growth. The project proposes to investigate three different III-V semiconductor materials, not only because they are all highly relevant for photonic applications (from UV to IR), but also because they allow to span the nucleation strain from compressive to tensile and close to zero.

Monolithic integration: the interfacial issue

On December 10, 2012 IBM announced a breakthrough optical communication technology which has been verified in a manufacturing environment. This technology breakthrough allows the integration of different optical components side-by-side with electrical circuits on a single silicon chip, for the first time, in standard 90nm semiconductor fabrication. Despite the very promising results obtained with this “combined front-end” optical on-chip integration, the development of laser sources on chip is still the limiting issue in this integration scheme. To increase the level of photonics integration in a mid-term perspective, one of the most powerful and economical way is to perform the direct epitaxy of III-V semiconductors on the silicon chip. This idea has been widely studied in the early 80’s, and revisited recently by three French laboratories which demonstrated advanced III-V optical functionalities/emitters on silicon with the growth of respectively the GaSb/AlSb (at IES Montpellier), GaP/AlP (at FOTON Rennes) and GaN/AlN (at CRHEA Valbonne) materials systems on Si. From their previous experience, IES, FOTON and CRHEA acknowledge the same fact: the overall III-V materials quality is fully determined by the control and quality of the interface, i.e. by the initial Si surface morphology, and the nucleation of the first III-V monolayers. <br /> <br />The ANTIPODE project is built around the three following objectives: <br />- understanding the 3D nucleation mechanism of III-V semiconductors on silicon (including generation of defects during coalescence) and the associated strain relaxation mechanisms. <br />- understanding the nature and role of the interfacial charges on the growth and defects generation. <br />- understanding of the influence of the initial silicon surface.

This project mainly dedicated to the study of defects through state of the art characterization and modeling tools, aims to provide a helpful and deep understanding of III-V/Si semiconductor interfaces and their electronic properties and if possible to propose a unified description of III-V/Si.
Partners of the ANTIPODE project are studying the initial stages of III-V/Si nucleation by MBE. To this end, they benefit from the strong growth expertise developed in IES (III-Sb/Si), FOTON (III-P/Si) and CRHEA (III-N/Si). They also benefit from the control of the initial Si surface (UHVCVD-MBE growth cluster), state-of-the-art advanced structural/electronic characterizations in three recognized laboratories (HRTEM or STEM-HAADF available at CEMES-Toulouse and LPN-Marcoussis, synchrotron XRD or STM-BEEM available at IPR-Rennes) and theoretical support (atomic relaxation models, DFT, local charges and dipole).

The project has underlined some fundamental differences between the three considered materials, but has also allowed to evidence some obvious similarities, that are linked with the three objectives previously described:

-Demonstration of the strong impact of the silicon substrate preparation before the III-V/Si epitaxy and influence of carbon clusters.
-STM observations of the GaP/Si(001) surface morphology during the first growth steps : identification of the different summital crystal facets on the nominal and vicinal substrates.
-Atomic scale characterizations of III-P, III-Sb and III-N/Si interfaces.
-Understanding of the 2D/3D nucleation on the different materials systems, by considering surface and interface energies. Confrontation to DFT calculations, TEM analysis and STM.
-Demonstration of the coalescence influence on the crystal defects generation, and interaction between domains boundaries and markers.

The ANTIPODE project has already allowed demonstrating significant similarities between the different materials studied. These similarities have been used to infer general concepts (2D/3D nucleation, interaction between defects and markers, influence of the substrate vicinality, etc ...). But these concepts need to be unified, with a theoretical description. This frame should allow to understand the very beginning of the III-V/Si crystal growth, either by MBE or by MOCVD, and to propose improvments based on these models. It is the main objective of members of the ANTIPODE project, which can be useful as well for the whole scientific community working on the monolithic photonic III-V/Si integration.

-P. Guillemé et al., “Antiphase domain tailoring for combination of modal and 4 ¯-quasi-phase matching in gallium phosphide microdisks” Optics Express, Vol. 24, no. 13, 14608 (2016).(FOTON, CEMES)
-J.B. Rodriguez et al., X-ray diffraction study of G

ANTIPODE is a fundamental research project which aims to deeply understand the formation of III-V/Si semiconductor interfaces in order to better control the defect generation at the interface during epitaxial growth. The project proposes to investigate three different III-V semiconductor materials, not only because they are all highly relevant for photonic applications (from UV to IR), but also because they allow to span the nucleation strain from compressive to tensile and close to zero. Although quite different, these three semiconductor materials exhibit a common 3D nucleation growth mode during the very-early stage of the growth on silicon substrates. This project mainly dedicated to state of the art characterization and modeling tools, aims to provide a unified and helpful understanding of III-V/Si semiconductor interfaces and their electronic properties.
On December 10, 2012 IBM announced a breakthrough optical communication technology which has been verified in a manufacturing environment. This technology breakthrough allows the integration of different optical components side-by-side with electrical circuits on a single silicon chip, for the first time, in standard 90nm semiconductor fabrication. Despite the very promising results obtained with this “combined front-end” optical on-chip integration, the development of laser sources on chip is still the limiting issue in this integration scheme. To increase the level of photonics integration in a mid-term perspective, one of the most powerful and economical way is to perform the direct epitaxy of III-V semiconductors on the silicon chip. This idea has been widely studied in the early 80’s, where III-V on silicon devices were realized at an advanced level especially by growing the well-known GaAs directly on silicon. In these works, crystalline defects were found to be generated mainly at the III-V/Si heterointerface, and propagating in the volume.
This route was revisited recently by three leading French laboratories which demonstrated advanced III-V optical functionalities/emitters on silicon with the growth of respectively the GaSb/AlSb (at IES Montpellier), GaP/AlP (at FOTON Rennes) and GaN/AlN (at CRHEA Valbonne) materials systems on Si. The heteroepitaxy of these materials on Si at the “front-end” level would be a very promising industrial solution provided that the III-V optical active area remains very close to the III-V/Si interface (typically below 300 nm) to increase the level of integration. In this context, the common strategies to manage the defects limitation engineering are not pertinent, as they require thick III-V buffer layers. The only solution is to control the defects generation at the III-V/Si interface and limit their propagation. From their previous experience, IES, FOTON and CRHEA acknowledge the same fact: the overall III-V materials quality is fully determined by the quality of the interface, i.e. by the initial Si surface, and the first III-V monolayers.
Partners of the ANTIPODE project will study the initial stages of III-V/Si nucleation by MBE. To this end, they will benefit from the control of the initial Si surface (UHVCVD-MBE growth cluster), state-of-the-art advanced structural/electronic characterizations in three internationally recognized laboratories (HRTEM or STEM-HAADF available at CEMES-Toulouse and LPN-Marcoussis, synchrotron XRD or STM-BEEM available at IPR-Rennes) and theoretical support (atomic relaxation models, DFT, local charges and dipole).
The ANTIPODE project is built around the three following objectives:
O-I - understanding the 3D nucleation mechanism of III-V semiconductors on silicon (including generation of defects during coalescence) and the strain relaxation mechanisms.
O-II - understanding the nature and role of the interfacial charges (in both III-V/Si interfaces and III-V defects interfaces) on the growth and defects generation.
O-III - understanding of the influence of the initial silicon surface.

Project coordinator

Monsieur Charles Cornet (Fonctions optiques pour les technologies de l'information)

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.

Partner

IES Institut d'Electronique du Sud
LPN (CNRS DR IDF SUD) Laboratoire de Photonique et Nanostructures
FOTON Fonctions optiques pour les technologies de l'information
CEMES Centre d'Elaboration de Matériaux et d'Etudes Structurales
LPN Laboratoire de Photonique et de Nanostructures
CNRS CRHEA Centre de Recherche sur l'Hétéro-Epitaxie et ses applications
IPR Institut de Physique de Rennes

Help of the ANR 499,928 euros
Beginning and duration of the scientific project: September 2014 - 36 Months

Useful links