Abstract
Nanotechnology allows for the manipulation and observation of matter on the atomic scale through various top-down and bottom-up nanofabrication techniques. Among these nanofabrication techniques, the process of atomic layer deposition shows promise due to its ability to fabricate ultra-thin films on the Nano scale (below 100 nm). This technique uses the sequential pulsing of precursor, reactant, and inert gases to deposit single atomic layers of a desired thin film species. Thin films prepared using the technique of atomic layer depositions, show superior physical, optical, thermal, and chemical properties, making it an attractive option for many applications such those found in the semiconductor, gas sensing, biomedical, energy storage and water purification industries. Thus, academic, and industrial interests are concerned with improving the atomic layer deposition process by considering various process factors. The reactor geometry plays a vital role in achieving optimal gas flow, and precursor distribution. In plug flow reactors, such as the Picosun R-200 ALD reactor, reactor designers have incorporated a perforated plate to induce more uniform flow fields.
This study uses computational fluid dynamics to investigate the incorporation of perforated plates in plug flow ALD reactors under the assumption of the continuum domain. Herewithin, the deposition of alumina is investigated using trimethyl aluminium (TMA) and ozone (O3) as precursors. The perforated plate is modelled as a porous medium to reduce the number of elements required for modelling the complex perforated geometry while still capturing the effects of the perforated plate in the atomic layer deposition process. The governing equations of mass, momentum, energy, and species' transport were subjected to a chemical mechanism supplied by CHEMKIN-PRO using the computational fluid dynamics software ANSYS Fluent. The superficial porosity formulation was used to model the porous medium and investigate the effects of varying free area ratios and the effects of changing the distance between the porous plate and the substrate surface.
vi
The results of this study were validated by comparing the obtained growth rate from previous studies. and then comparing them to the growth found in this study. Previous research predicts a growth per cycle ranging between 0.6 Å and 1.1 Å. In comparison, the growth per cycle found in Case Study 1 (investigation into varying free area ratios) ranges between 0.714 Å and 0.775 Å. The growth per cycle found in Case Study 2 (investigation into varying porous plate heights), ranges between 0.730 and 0.790. These results fall within expected values and are thus considered valid. The results show that incorporating a perforated plate removes recirculating zones caused by the sudden expansion of gas entering through the inlet manifolds (Type A recirculation) and inlet configuration (Type B recirculation). These recirculating zones are shown to impinge on the mass fraction distribution of TMA and O3 within the reactor, above the substrate surface. Four different free area ratios were investigated (0.05, 0.1, 0.2, and 0.3) and it was found that varying the free area ratio between 0.1 and 0.3, had little impact on the growth. Ultimately, a free area ratio of 0.2 was selected as the best case because it provided the best uniformity.
Three porous plate heights were investigated (12 mm, 18 mm, and 24 mm). The results show that increasing the substrate height (decreasing the distance between the substrate surface and the porous plate), decreases the volume above the substrate surface. This subsequently leads to faster purge times and greater mass fractions of precursor and reactant gases. However, the porous plate height of 12 mm showed that this decreased volume may lead to insufficient mixing of precursor and reactant gases with carrier gas and thus, may cause irregular growth. These predictive models contributed towards the understanding and identification of fundamental behaviors within atomic layer deposition. The incorporated porous medium model simplifies the study of flow-controlling components such perforated plates and showerhead components in ALD.