Laboratoire de Physique des Interfaces et des Couches Minces

CNRS - École polytechnique - Institut Polytechnique de Paris

Plasma Processing for III-V Materials

Written by : Karim Ouaras

Pere Roca i Cabarrocas,

François Silva,

In last twenty years, researchers at LPICM have been deeply invested in the field of low temperature plasma epitaxy, addressing both plasma / material fundamental aspects, including ab initio molecular dynamics modelling, and applications of silicon for PV and microelectronics. These studies have been limited so far to Silicon and Germanium epitaxial growth. From the gained experience, efforts have recently been made to develop plasma processes for other semiconductors of superior / complementary optoelectronic properties such as III-V materials to combine them with silicon.

Taking benefit from the intrinsic features of plasmas, i.e. low temperature and cost effective process (less gas / equipment consuming), in comparison with conventional epitaxial techniques (MOCVD, MBE), we aim at achieving low cost III-V (GaAs and GaN) epitaxy. This marks the start of a new field of fundamental and applied research at LPICM. It is structured into two axes, each one being complementary:  Epitaxial growth of GaAs using Plasma Enhanced Chemical Vapor Deposition (PECVD) and Epitaxial growth of GaAs and GaN using Physical Vapor Deposition (PVD) .

Epitaxial growth of GaAs using Plasma Enhanced Chemical Vapor Deposition

III-V materials, such as GaAs or GaN, are conventionally deposited by MOCVD. This process operates at relatively high temperatures (between 800 and 1000 ° C) and at pressures close to the atmosphere. At such high pressures, it is necessary to use high gas flow rates and achieve a uniform distribution of precursor on the surface of the substrate. This induces a large consumption of precursor gases but also makes it necessary to treat a large quantity of effluents at the pump outlet. This is one of the reasons why the current cost of III-V materials is so high. In addition, the high growth temperature considerably limits the possibility of achieving heteroepitaxy with materials whose thermal expansion coefficient is very different (appearance of thermal stresses during cooling which can lead to the destruction of the substrate).

A new approach aimed at overcoming these limitations has been undertaken at LPICM. This new epitaxial growth process operating at low pressure and low temperature is based on the use of plasma assistance by an ICP source (see Figure 1). The latter makes it possible to obtain high density plasmas while maintaining a low plasma potential (low energy of the ions), unlike most PECVD deposition reactors where the substrate is subjected to ion bombardment inducing the formation of point defects at the surface that can lead to the loss of epitaxy (epitaxy breakdown). In our case, the plasma is used to provide radicals to assist the process. Numerous chemical reactions in the gas phase (decomposition of precursors) and on the surface (creation of radical sites, desorption of reaction products) can be assisted by atomic hydrogen generated by the ICP source. This free radical assistance of the process makes it possible to significantly reduce the growth temperature (by several hundred degrees) while maintaining good quality of the material deposited. In addition, the low working pressure (of the order of 1 mbar) considerably reduces the gas flow required to transport the precursor (the diffusion being much more efficient at low pressure).

This new process should therefore make it possible to drastically reduce the costs of manufacturing the material (due to the reduction in the consumption of precursor) but also to consider widening the choice of substrates for heteroepitaxy (in particular epitaxial growth of the GaAs on silicon).

Figure 1. Schematic of the GaAs PECVD reactor set-up.

Epitaxial growth of GaAs and GaN using Physical Vapor Deposition

The use of plasma process under physical vapor deposition mode offers a way to overcome the thermal and / or cost issues experienced by other methods (MOCVD or MBE). Indeed, the application of an electric field brings the energy to the electrons to enhance the reactivity of the gas to create a highly reactive media made of ions, excited species, radicals and eventually nanoparticles, while keeping a global (gas) temperature ranging from RT to 200 °C. Apart from this feature that makes unique plasma processes, in its PVD version, the use of a target made of the material of interest in front of a reactive plasma enables the synthesis of thin films of same nature as the target without resorting to toxic gases. The reactor will work under usual plasma conditions, i.e. Radiofrequency excitation (13.56 MHz) and low pressure (tens to hundreds of Pa range). Basically, in the case of GaAs for example, a solid target of GaAs will face an inert Argon plasma to produce GaAs thin film without resorting to trimethylgallium (TMGa) and Arsine(AsH3) as MOCVD does.

Figure 2 shows a scheme of the first pilot reactor dedicated to GaN epitaxy that has been designed at LPICM. For carrying out the fundamental part of this study that targets the understanding of epitaxial growth mechanisms, several optical diagnostics (OES, THz spectroscopy, TALIF …) will be implemented.


Figure 2. GaN PVD plasma reactor 3D architecture.