Thomas Research Group

Organometallic Chemical Vapour Deposition (OMCVD)

A generalized deposition process such as OMCVD presents a well controlled method for the fabrication and precise location of nanostructures. This ability will prove crucial for future device fabrication; control of both the size and location of these quantum structures is vital. To this end, we intend to accelerate the OMCVD program that we have started and investigate new OM precurosr molecules and new deposition procedures. The success of our longterm goal to understand the energetics of the thermal and electron induced decomposition of organometallics is perhaps best measured by fabricating the essential building blocks of useful quantum devices. A viable target is to fabricate 3 quantum dots, 30 nm in diameter and separated by 30 nm, formed from either two or three different materials. A sequence of three such dots, "metal -wide bandgap semiconductor - metal" would demonstrate our ability to control these decomposition reactions with sufficient accuracy to create quantum devices with which we can observe such phenomena as a Coulomb Staircase or to successfully photoinject electrons into a single barrier quantum structure.

We hope to move our work forward and several fronts. First, the fabrication of high quality compound semiconductors requires very precise control over the relative concentration of the reactants. Exact stoichiometry is required for superior material properties achievable in defect-free substrates. One way to improve this is to develop single source precursor molecules. Here, the OM molecule contains both atoms for the deposition process and it decomposes in such a manner as to directly control the stoichiometry of the final product. We have worked with Dr. Darrin Richeson of the University of Ottawa who has developed a precursor for GaN. This material is extremely exciting right now. It is a compound semiconductor with a wide band gap such that it naturally emits light in the blue region of the visible spectrum. This is the region of interest as the development of a blue solid state laser would spur a dramatic increase in storage density for magnetoptical and CD-ROM memory devices. Dr. Richeson's molecule, Ga(N(SiCH₃)₂)(OSiCH₃)₂Py, has been shown to thermally decompose to give cubic GaN. We want to investigate its behaviour when exposed to a heated Si surface in vacuum. The direct production of GaN would be a dramatic contribution in fabricating this material and the devices we hope to make from it. Another material of potential interest might be the fabrication of something such as ZrO₂ from the zirconium alkoxides. These oxides have very large dielectric constants and may prove useful for fabricating nanoscale dimension capacitors and the like.

One of the first steps in studying the behaviour of these molecules is to understand their adsorption and desorption characteristics. We can apply Thermal Desorption Spectroscopy (TDS), carried out under ultrahigh vacuum conditions, to examine the bonding and reaction properties of these molecules on substrates such as Si(111) and Si(100). Questions regarding the strength of the adsorbate-surface bond, the state of the OM upon adsorption, and the feasability of intact molecular thermal desorption can be answered.

Molecular Beam Reaction Spectroscopy (MBRS) provides additional insight into the thermal decomposition process and will help us to control that pathway should it prove deleterious. By monitoring the desorption products obtained while dosing the sample surface at elevated temperatures, we will answer questions regarding the activation energy of the thermal decomposition pathway, the nature of the decomposition fragments, probe questions regarding the time scale (greater than a few seconds) kinetics of the system. Specifically, autocatalytic properties which are sometimes observed in these decompositions, will be measured.

Electron Induced Dissociation (EID) of the OM precursors is crucial to our fabrication scheme. We are particularly interested in how low energy electrons (a few electron volts) will affect these species. Successful dissociation with low energy electron beams will permit the use of the STM as a lithogrpahic tool with its extremely high spatial resolution. A first step will be to investigate the EID properties by exposing a monolayer of the substance toa low energy electron shower from an appropriate electron gun. Dr. Paul Rowntree from the Université de Sherbrooke in Québec has a long history in EI of organic molecular reactions and has shared with us a recent design for a low energy gun. We will copy this design and start to study these reactions. TDS studies, along with Low Energy Electron Diffraction (LEED) and Auger Electron Spectroscopy (AES) will help to clarify the reaction products and elucidate the dissociation pathways as a function of beam energy.

Our approach to nanolithography is to use the probe tip of an STM as a source of low energy electrons to locally induce the decomposition of the precursor molecule. This process has been used several times for the deposition of metallic structures but its application to the fabrication of compound semiconductors or insulating oxides from single course precursors is both novel and important. With low energy electrons of a sufficiently narrow energy bandwidth, we may be able to resonantly activate the OM molecules for dissociation with the hope of achieving preferential bond scission to lead to the preferred product state. This direct writing of nanostructures is the most straightforward approach to their fabrication. Another pathway is also possible as it has been show that the STM electron beam can be used to locally desorb the hydrogen termination atoms from a Si surface. The exposed Si dangling bonds are much more reactive than those saturated by H, so that a subsequent exposure to the OM precursor while holding the substrate at an elevated temperature could induce the thermal decomposition of the OM at the localized sites where the H had been removed. In this fashion, a patterned surface exploiting only the thermal properties of the OM is still possible, should the EID approach prove difficult.

With these tools in place, we would have at our disposal all that is necessary to fabricate the target structure mentioned at the beginning of this article.