Public repository of databases for modeling the Chemical Vapor Deposition of some ceramics with CFD-Ace
General disclaimer : these databases have been developped in the SIMAP laboratory (France), between 2006 and 2023 for internal use with custom CVD reactors. The related papers in which these models have been used are given in description. Use at your own risk. These models have been used essentially for simulating epitaxial growth (microelectronics industry) and hard coatings for harsh Environments (Nuclear and Solar Energy).
All codes are to be used with CFD-Ace from Applied Materials, a general CFD modeling software suite. To import the databases for transport, properties and chemical reactions, just open the files and copy data from the local .DTF database to your local database. Codes converge with the 2023.0 version of the software. All models include transport properties, thermodynamical properties, gas-phase and surface kinetics. They can be run as it. GEOM files are not provided as they depict our custom CVD reactors and are not of prime interest.
The theories used to estimate kinetics and transport properties as well as the formalism for gas-phase and surface kinetics are detailed in my lectures given at Phelma School of engineer (France). Source material for teaching can be obtained on request.
Supporting data for paper https://doi.org/10.1016/j.surfcoat.2020.126102
This model is the oldest but it has been extensively used and modified until 2019. It embeds chemistry of the Si-C-H-Cl system in gas phase and surface to perform simulations of SiC growth from a mixture of silane, propane and chlorine. Chlorine is used here only to reduce the elementary silicon vapor pressure and avoid condensation of pure silicon during growth. Last use of this code was for simulation SiC growth on refractory metal tiles for Concentrated Solar Panel (CSP) applications. It is not the most complicated set of chemical reactions that can be written in this system but it allows a robust modeling of the silane-propane route for SiC growth. We are now developping this chemical dataset with ANSYS-Fluent.
Supporting data for paper https://doi.org/10.1007/s00214-013-1419-8
This model was initially developped between 2007 and 2010 for epitaxial growth on sapphire, then adapted for growth on refractory metals around 2015. It uses AlC3 and NH3 as main precursors and is inspired by growth mechanism of BN with BCl3 and NH3 which is quite similar. The gas phase reactions involving the Al-Cl-N-H system are numerous but the relative chemical stability of AlC3 and NH3 in gas phase allows to simplify the mechanism to a very reduced set of gas-phase and surface reactions. This model is ultra-robust and has been tested and fitted with hundreds of experimental data points in various conditions. It was used daily between 2007 and 2019. The fashion for thick AlN by CVD is a bit passed now. In particular the deceptive protective properties of AlN as thermal barrier in air at high temperature was a huge disappointement on our side (the alumina layer formed during oxydation is porous due to trapped N2 bubbles). Design of experiment optimization of the AlN growth based on the exact same process (as well as more technical details) can be found here.
This model was developped around 2009 to perform the CVD of BN from BCl3 and NH3 on various substrates (graphite, sapphire, SiC). The gas phase reactions involves the B-Cl-N-H system and are inspired from a study from M.D. Allendorf. NH3 pyrolysis reactions were skipped because too slow at this pressure/temperature. Surface reaction mechanism was inspired from similar mechanism published by A. Dollet on AlN and tuned with experimental data. This model was used in the thesis of Nicolas Coudurier to assess the effect of local gas saturation on BN texture/epitaxy on various substrates. The model per se was unpublished, so use at your own risks.
Supporting data for paper https://doi.org/10.1039/C9CE00488B
This model was intended to understand the relationship between growth conditions and color of the TiN. Indeed, TiN can be grown as a gold colored refractory and ultra-hard layer in certain conditions. This model allowed us to conclude that the only parameter playing a role on the aspect of TiN grown by CVD was the growth rate/supersaturation. Slowly grown TiN at low supersaturation is always compact and shiny, despite the variation of other growth parameters. Crystal preferential orientation or Ti/N ratio in gas phase plays no role. This model is particularly simple as there in no need for gas phase reactions to explain the reaction kinetic: only surface reactions are involved. Complete modeling of the Ti-Cl-N-H system in gas phase offers indeed no gain in terms of comprehension and modeling accuracy. The TiN obtained by chlorine based CVD, despite its shiny aspect and huge hardness, offers poor protection to corrosion in humid air due to the less than everage barrier properties of TiO2.
Supporting data for paper https://doi.org/10.3390/coatings8060220
This model was developped to simulate the growth of CrC (in real an amorphous CrxCy compound) into long narrow tubes for energy applications. It simulates the MOCVD of bis(benzene)chromium dissolved in toluene. The reaction mechanism is quite simple as bis(benzene)chromium undergoes an easy thermal degradation in mono(benzene)chromium which is a quite sticky molecule. The challenge here was to compute the thermodynamical and transport properties of mono and bis(benzene)chromium as it was never done before. The model was fitted on dozens of experimental data points in various geometrical reactor configurations. The glassy layers obtained have very low oxydation kinetic in water vapor at high temperature and pressure, which was the expected property. Chemistry of toluene is not involved in this database but it undergoes pyrolysis from about 600°C (as an observed parasitic black carbon deposit on solid surfaces).
This database is intended to be updated with time