Chemical Vapor Deposition (CVD)
The science and engineering literature on the topic of thermal or plasma-enhanced CVD is plentiful. Here, I state the main challenges and design consideration for the reactor.
The challenges are (1) uniformities of thickness and properties across the substrate surface, (2) keeping a clean bevel-edge, (3) avoiding particles generation, (4) adhesion of film to the substrate, and (5) cleaning the reactor (preferably by an in-situ method) and controlling the cleaning time by an end-pointing method.
The schematic of the reactor summarizes the ideas that would go into the design of such hot-wall reactor.
In this conceptualization the process engineer need to identify the main effects amongst:
 | temperature(s) of pedestal and showerhead |
| reactor pressure |
| gas mixtures |
| gases flow rates |
| clamping forces (whichever method is used to generate them) |
| gas flow rate around the beveled edge of the substrate |
| if one of the gases is supplied in its liquid phase then the evaporator variables might need to be considered |
Some precursors are in liquid form. In such cases an evaporator is needed to vaporize the liquid at some point before entry into the showerhead. The vapor and other injected gas(es) react on the substrate surface to form the film.
Chemical vapor deposition results in films that are conformal to the surface the film is deposited on; this is in contrast to PVD, or sputtering. The conformity of the deposited layer to the underlying surface can be achieved up to a limited surface topography, and beyond these limiting topographic features extensive experimentation is required, and perhaps a re-design of the reactor. For example, with memory devices featuring aspect ratios of 20:1 (that is, 20 units down for every horizontal unit), it is difficult to achieve coverage of the side walls and complete filling of, say, the vias without defects, as voids. A solution to this process problem is Atomic Layer Deposition (ALD).
I came across Daniel Dobkin's web site on CVD. He covers topics I was planning to include on this page. Danile's CVD page is Fundamentals of Chemical Vapor Deposition.
Plasma-Enhanced Chemical Vapor Deposition (PECVD)
PECVD requires either the shower-head or the pedestal be energized by a rf-source.
Helpful Books on CVD
- S. Sivaram, "Chemical Vapor Depsotion", Van Nostrand Reinhold, 1995.
- S. Wolf and R.N. Tauber, "Silicon Processing", Vols. 1 and 2, Lattice Press, 2000.
- R.C. Jaeger, "Introduction to Microelectronic Fabrication", Addison-Wesley, 1988.
Design of Experiments
Design-of-experiments (DOE) are recommended for these undertakings. The software packages I use are JMP from SAS and Design Expert from StatEase (depending on the business entity's preference). As for the fundamental knowledge for DOE, here is a list of the books I use (in alphabetical order by the author's last name):
- M.J. Anderson, et alia, RSM Simplified,
- G.E.P. Box, et alia, Statistics for Experimenters,
- F.W. Breyfogle, Implementing Six Sigma,
- W.W. Hines, et alia, Probability and Statistics in Engineering and Management Science,
- D.C. Montgomery, Design and Analysis of Experiments,
Home | About Me | Plasma Processes | Resources | Feedback
This site was last updated 07/11/10
© 2009 Stephen Hyatt
Web Site Administrator Novella Web Design