Of particular technological importance are the chemical reactions that occur at solid surfaces during the processing of semiconductor wafers. For example, dry processes such as reactive ion etching (RIE) and chemical vapor deposition (CVD) are routinely employed in the manufacture of microelectronic devices. Although much is known about the conditions necessary for making device structures, the mechanistic details of the surface reactions that occur during these procedures are largely unknown. By determining the underlying physics and chemistry behind these reactions, device manufacturers will be better able to design new processes.
Our research program is unique in that we employ surface science techniques to look at fundamental aspects of RIE and CVD. Semiconductor surfaces are studied after reaction with model gas-phase precursor species in UHV. A variety of surface analysis techniques are employed to probe the surfaces, both in our laboratory and at synchrotron radiation (SR) facilities. Systems are studied in the submonolayer regime, where the interaction between the bare substrate atoms and the reactant is quantified, and in the high exposure regime typical of a 'real' process. Both the geometric and electronic structures of a sample are probed, and the results are correlated into a detailed understanding of the properties of a system that is not available from the use of a single technique.
An example
of this type of research is our work on the XeF2/silicon etching
reaction. XeF2 is a convenient source of atomic fluorine, which
allows one of the most important interactions in semiconductor processing
to be probed on an atomic scale. Soft x-ray photoemission (SXPS) and photon
stimulated desorption (PSD), which are both SR-based techniques, were employed
to study Si surfaces after etching with XeF2. A reaction layer,
composed of intermediate fluorosilyl species, forms on the surface as a
result of the etching process as indicated in the figure. We showed (1)
that the fluorosilyl species have a particular arrangement on the surface
which is critical to the etching process [1], (2) that the evolution of
the reaction layer geometry is a direct result of the formation of trivalent
Si defects in the lattice [2], and (3) how the electronic structure of
the surface region affects the reaction dynamics [3].
The etching of III-V semiconductor surfaces by halogens is important in the manufacture of optoelectronic devices [4]. The surface reactions of these materials present a more complex problem than for elemental semiconductors, however, as there are competing reaction pathways. For example, it has been known that ordered structures form when halogens adsorb onto certain III-V semiconductor surfaces, while other surfaces become disordered and etch, but a physical picture that explains these observations was lacking. We have found that the initial surface structure determines the pathways that are followed by any given reaction, as some of these systems will spontaneously etch at low temperature, while others will form stable ordered overlayers [5-11]. From our most recent work on chlorine and iodine adsorption on InAs(001) and InSb(001), we have developed a mechanistic understanding of these reactions [12-14]. This mechanism is based on the ideas that (1) surface group III elements are initially more reactive to incoming halogen molecules due to their empty surface orbitals, and (2) steric limitations imposed by halogen-halogen interactions on a surface can inhibit the formation of dihalide species. This understanding of the fundamental surface interactions is critical in designing processes to be used for binary materials.
An area that is analogous to etching is deposition. Atomic layer epitaxy (ALE) is a chemical deposition technique that can produce layer-by-layer epitaxial growth of Si with monolayer control. We studied ALE processes on Si and Ge using alternating cycles of SiH2Cl2 and H2, and were the first to identify the surface intermediate reaction products for this system [15,16]. We have also looked at the growth of insulating films of GaF3 on GaAs surfaces via reaction with XeF2 [17-19]. These 10 eV bandgap films have potential for use in III-V microelectronic devices, as this method of growth involves a simple chemical reaction that makes it easily amenable to current manufacturing techniques.
Our future plans for semiconductor processing studies involve the use
of energetic, rather than thermal, reactants. The motivation here
is to better understand the mechanisms involved in plasma processing in
order to provide better control of the structures that are produced, with
the ultimate goal being the reproducible fabrication of uniform nanoscale
features. We will use controlled molecular beams of fast halogen
atoms and low energy beams of halogen ions to etch semiconductor surfaces.
We will then investigate the resulting surfaces with our array of surface
chemical and structural probes.