Jury

Isabelle Berbezier: Research Director, CNRS-IM2NP, rapporteur

Matthieu Jamet: Director of Research, CEA-SPINTEC, rapporteur

Adrien Michon: Research Fellow, CNRS-CRHEA, examiner

Fabrice Oehler: Research Fellow, CNRS-C2N, examiner

Pascal Roussel: Research Director, CNRS-UCCS, examiner

Xavier Wallart: Research Director, CNRS-IEMN, directeur de thèse

Summary:

Two-dimensional materials (2MDs) offer unique properties mainly due to their structure
crystallographic, composed of loosely coupled layers made up of one or more atomic planes where the
atoms are covalently bonded. This specific structure means that there are no dangling bonds on the surface,
allowing the formation of heterostructures without the need to adapt the lattice parameter. Among the 2MDs, the
Transition metal dichalcogenides (DMTs), of formula MX2 (M=metal, X=Se, Te or S), are particularly well suited to the production of dichloromethane.
interesting because some of them have band gaps ranging from 0.5 to 2 eV and offer a wide variety of properties.
strip alignments. These characteristics make them promising candidates for the development of various
components, particularly those based on inter-band tunneling. To date, most devices have
have been manufactured using exfoliated or transferred layers or flakes, introducing significant problems
concerning the quality of the interfaces and the reliability of the process.

Molecular beam epitaxial growth (MBE) has attracted a great deal of interest over the last decade.
for the development of heterostructures based on DMTs, particularly on 2D substrates such as graphene,
h-BN or mica due to the natural van der Waals interface between the epitaxial layer and the substrate. In parallel,
a new generation of heterojunctions has emerged, using 3D substrates such as Al2O3(0001) or AlN(0001),
underlining the crucial importance of surface preparation prior to growth. The growth of a
monolayer of WSe2 or that of a MoSe2/WSe2 heterostructure were produced on a GaAs(111)B
previously passivated under selenium, sparking new interest in the use of III-V semiconductors
as substrates for the development of devices based on DMTs. However, Monte Carlo simulations
have highlighted the need for high growth temperatures in order to improve the grain size of DMTs, making
makes GaP a suitable alternative to GaAs, thanks to its much greater thermal stability.

In this thesis work, we explore the growth by EJM of a few layers of WSe2 and HfSe2, as well as the growth by EJM of a few layers of WSe2 and HfSe2.
that of the HfSe2/WSe2 heterostructure on a GaP(111)B substrate. Each stage of the process was studied, with a
particular attention to the preparation of the substrate, the crystalline quality of the layers obtained and their preservation
the integrity of the heterojunction interface. We show that the use of a combination of hydrogen
and cracked phosphine in a three-step sequence to produce a smooth GaP(111)B surface
and completely deoxidised. To encourage the formation of a van der Waals quasi-gap between the substrate and the
DMT layer, the GaP(111)B surface is then selenium terminated. Using a combination of
experimental techniques, we demonstrate the epitaxy of WSe2 and HfSe2 on GaP(111)B-Se, highlighting
the crucial role of growth temperature and annealing under Se on the formation of polytypes, the properties
structural and morphological properties. The resulting layers show clear epitaxial relationships with the substrate.
GaP(111)B and a well-defined van der Waals interface. WSe2 has p-type doping and an alignment of
type II bands with the GaP substrate, while HfSe2 reveals a type I band alignment and a
heavily doped n. Finally, we discuss the challenges of developing a GaP/HfSe2/WSe2 heterostructure in
due to the formation of a (W/Hf)Sex compound, while the GaP/WSe2/HfSe2 stack showed results of
promising with a type II alignment and a clear p-n junction.

Abstract:

Two-dimensional materials (2MDs) offer unique properties mainly due to their
crystallographic structure, consisting of loosely coupled layers of one to a few atomic planes where the atoms are
atoms are covalently bonded. This specific structure means that there are no dangling bonds at the surface,
allowing the formation of heterostructures without the need to adapt the lattice parameter. Among the 2MDs, the
transition metal dichalcogenides (DMTs), of formula MX2 (M=metal, X=Se, Te or S), are particularly interesting because some of them
interesting because some of them have band gaps ranging from 0.5 to 2 eV and offer a wide variety of band alignments.
of band alignments. These characteristics make them promising candidates for the development of various
components, particularly those based on inter-band tunnelling. Until now, most devices have been fabricated using
been fabricated using exfoliated or transferred layers or flakes, introducing significant problems
interface quality and process reliability.
Over the last decade, molecular beam epitaxial growth (MBE) has attracted considerable interest for the development of heterostructures.
in the development of DMT-based heterostructures, particularly on 2D substrates such as graphene,
h-BN or mica because of the natural van der Waals interface between the epitaxial layer and the substrate. In parallel,
a new generation of heterojunctions has emerged, using 3D substrates such as Al2O3(0001) or AlN(0001),
underlining the crucial importance of surface preparation prior to growth. For example, the growth of a
WSe2 monolayer or a MoSe2/WSe2 heterostructure were grown on a GaAs(111)B
surface previously passivated under selenium, generating new interest in the use of III-V
semiconductors as substrates for the development of DMT-based devices. However, Monte Carlo simulations
simulations have highlighted the need for high growth temperatures in order to improve the grain size of DMTs, making GaP a suitable alternative to GaC.
GaP a suitable alternative to GaAs, due to its significantly higher thermal stability.
In this thesis, we explore the EJM growth of a few layers of WSe2 and HfSe2, as well as that of the HfSe2 heterostructure.
HfSe2/WSe2 heterostructure on a GaP(111)B substrate. Each stage of the process was studied, with particular
preparation of the substrate, the crystalline quality of the layers obtained and the preservation of the
integrity of the heterojunction interface. We show that the use of a combination of atomic hydrogen
and cracked phosphine in a three-step sequence results in a smooth and fully deoxidised GaP(111)B
surface. To encourage the formation of a van der Waals quasi-gap between the substrate and the
DMT layer, the GaP(111)B surface is then selenium terminated. Using a combination of
of experimental techniques, we demonstrate the epitaxy of WSe2 and HfSe2 on GaP(111)B-Se, highlighting the crucial role of the
growth temperature and annealing under Se on polytype formation, structural and morphological properties.
morphological properties. The resulting layers show clear epitaxial relationships with the
GaP(111)B substrate and a well-defined van der Waals interface. WSe2 exhibits p-type doping and a type II band alignment with the GaP(111)B substrate.
band alignment with the GaP substrate, while HfSe2 reveals a type I band alignment and a highly n-doped
highly n-doped character. Finally, we discuss the challenges of developing a GaP/HfSe2/WSe2 heterostructure due to the formation of a GaP/HfSe2/WSe2 compound.
formation of a (W/Hf)Sex compound, while the GaP/WSe2/HfSe2 stack shows promising results with a type II
promising results with a type II alignment and a clear p-n junction.