The manufacturing of micro-electronic devices depends on chemical vapor deposition (CVD), a chemical engineering process. In particular, designing CVD reactors requires knowledge of chemical kinetics, chemical reactor engineering, and transport phenomena. In this computational experiment, our focus is on the so-called boat reactor.
You can download Dr. Yang's COMSOL Introduction presentation here.
You can download the Pre-Lab questions for this experiment here.
Our intent is to design, via simulations, CVD reactors with increasing sophistication (i.e., more realistic features). The COMSOL model is a simple isothermal reactor but it's nevertheless a good starting point. Specifically, the goals of this experiment are
- to extract from published literature the chemical kinetics of silicon deposition using silane,
- to establish a CVD reactor model with the proper assumptions, chemical kinetics, and conservation equations,
- to build and understand the computational models used by COMSOL,
- to apply the results to design a low pressure CVD reactor (LPCVD), and
- to study the effects of multiple variables using statistical Design of Experiment (DoE) principles.
That said, most of the files you need for the actual CVD simulation have already been created. The files you'll create are intended to introduce you to COMSOL; the files you're provided for the CVD simulation are intended to introduce you to the multi-physics capabilities of COMSOL so that you can rapidly design the final LPCVD reactor.
When you perform a literature search for LPCVD or CVD reactor design, stay with the older articles! Newer articles--those published in about 2002 or later--tend to be more advanced than is useful for this experiment. There are two areas that you should focus on to demonstrate your understanding in the Background and Theory sections of your written report: CVD reactors, and the kinetics of silane decomposition and deposition.
The following articles and references may be helpful towards these ends:
- Your reaction engineering texts, particularly Fogler. A good introduction to CVD in micro-electronics is provided in Section 10.5.
- Micro-reaction engineering: Applications of reaction engineering to processing of electronic and photonic materials. This is a reference used by Fogler with good background information.
- Transport properties for the silane deposition system are available here.
- A great extension of the 1-D analysis presented by Fogler is available here.
- Insight into the boat reactor design is available here, here, and here.
- The reference cited by the COMSOL tutorial is this one.
- Details of a production-scale, three-zone CVD reactor are available here.
Standard Operating Procedure
There are no chemical used in this computational experiment. You're using a computer so you should observe all the usual precautions regarding ergonomics: sit up straight, keep your feet flat, and keep the monitor at eye level (a good reason to keep your old textbooks around). Emptying the cache and occasionally talking to the computer, perhaps about its weekend plans, are good ways to keep it running at optimal speed.
There are four things you should do in preparation for this lab (i.e., before you get to lab):
- Review your reaction engineering text as noted above.
- Complete the Pre-Lab questions so that you have an idea of how numerical approximation works since that's the entire purpose of COMSOL.
- Read the COMSOL LPCVD boat reactor tutorial, which you can download here. Focus on understanding the physics that are used and the meaning of the plotted data; the actual tutorial aspect will be more clear once you're in the lab.
- Collect from literature ranges of realistic values for the following parameters, which we expect to read about somewhere in your report (logically in the Background or Theory section):
- CVD reactor geometry,
- transport properties (diffusion coefficients, viscosity, thermal conductivity, etc),
- operating conditions (pressure, temperature, flow rates, etc).
Day 1 is all about learning COMSOL and getting up to speed with the basics of a boat reactor. Consequently, the steps to be accomplished today look about the same for everyone:
- Check out the tutorial video below to learn more about COMSOL and the boat reactor model.
- Load COMSOL's boat reactor model (the file location is provided in the boat reactor tutorial above, or you can download it here.) and reproduce some of the figures to verify its functionality.
- Produce at least one plot of the effectiveness factor (which you can read about here). You can find the MATLAB code to do so here and a sample data set for use with this MATLAB code here (Note: MATLAB code updated on 3 Jan 2018).
Load the 3-zone model with Langmuir-Hinschelwood kinetics (3LH, available here) and explore the effects of various operating parameters (available here, but they should already be part of the 3LH model file) on the effectiveness factor and silicon deposition rate. The parameters you choose to vary are up to you but since there are four of you (in most cases) you should be able to explore quite a range of operating conditions. Remember that part of your task is to figure out how this whole setup works, and to explain it in your report.
Also remember that we're more interested in your interpretation of the results rather than whether or not you can find the "best" operating conditions. Different manufacturing requirements might stipulate different definitions of "best" anyway, so you should be able to explain why different parameters have different effects.
If you've got three days in the lab then you should use the third day to perform a Design of Experiment (DoE) analysis on the 3LH model. DoE is a very powerful statistical tool that lets you explore not just direct effects of varying a single parameter but interactive effects of many parameters. You can find a decent introduction to DoE procedures here.
Again, it's a computer. You should know this part by now.
Common Report Mistakes and Suggestions
The evaluator(s) for this experiment was asked to list three or more common mistakes often made on the reports for this experiment, or to provide three or more suggestions that could improve the quality of reports for this experiment. The responses were as follows:
- Use abbreviations like LPCVD (CVD) consistently. Once you defined it, then use it and do not define it again.
- The Appendix should not be used for figures which display primary or important results, or to circumvent the 10 page limit.
- Give reference for the COMSOL Multiphysics. https://www.comsol.com/
- Fogler, H.S., Essentials of Chemical Reaction Engineering; Pearson Education: Boston, 2011.
- Jensen, K.F. Micro-reaction engineering applications of reaction-engineering to processing of electron and photonic materials. Chem. Eng. Sci. 1987, 42, 923-958.
- Kleijn, C.R.; van der Meer, T.H.; Hoogendoorn, C.J., A mathematical model for LPCVD in a single wafer reactor. J. Electrochem. Soc. 1989, 136, 3423-3433.
- Komiyama, H.; Shimogaki, Y,; Egashira, Y., Chemical reaction engineering in the design of CVD reactors. Chem. Eng. Sci. 1999, 54, 1941-1957.
- Roenigk, K.F.; Jensen, K.F., Analysis of multicomponent LPCVD processes: Deposition of pure and in situ doped poly-Si. Electrochem. Soc.: Solid State Sci. Tech. 1985, 132, 448-454.
- Zambov, L.M., Optimum design of LPCVD reactors. J. Phys. IV 1995, 5, 269-276.
- Sachs, E.; Prueger, G.H.; Guerrieri, R., An equipment model for polysilicon LPCVD. IEEE Trans. Semiconduct. Manufactur. 1992, 5, 3-13.
- Voutsas, A.T.; Hatalis, M.K., Structure of as-deposited LPCVD silicon films at low deposition temperatures and pressures. J. Electrochem. Soc. 1992, 139, 2659-2665.
- Setalvad, T.; Tachtenberg, I.; Bequette, B.W.; Edgar, T.F., Optimization of a low-pressure chemical vapor deposition reactor for the deposition of thin films. Ind. Eng. Chem. Res. 1989, 28, 1162-1170.