Courses:

The purpose of the term project in 6.780 is for students to work with actual data from some aspect of semiconductor fabrication and apply the tools and methods (or extensions of them) that the class has discussed this semester. Students are responsible for identifying the topic of the term project, understanding the data associated with the topic, finding and understanding methods appropriate to the problem or analysis they wish to perform, completing the analysis, and presenting their results. The term project report will have two components: an oral presentation and a written component.

Two student team projects are encouraged. A single grade for the team will be assigned, based on both the written report and oral presentation.

The written report should be submitted electronically. The paper should be approximately 6-8 pages in length, formatted in two columns per page format consistent with journal publications. The IEEE has a Word template that may be helpful.

The presentation should be 10 minutes in length, and may be presented either electronically (using the PC projector) or on overhead transparencies. It will be very important to practice the presentation and not run over the alotted time. At the conclusion of the talk, there will be five minutes for questions.

Important dates:

• Monday, April 28 - submit one page proposal identifying the topic, data set, and approach
• Monday May 12 and Wednesday May 14 - oral presentation (10 minutes plus 2 minutes for questions)
• Wednesday May 14 - written component of term project due

Term Project - Potential Data Sets

• Students may be able to identify a good potential dataset based on other work within their research group, or within their laboratory or previous experience. This will be an excellent way to have relevant and "fresh" data for use in the project.
  
• Another option is datasets from the book Statistical Case Studies for Industrial Process Improvement by V. Czitrom and P. Spagon. In the past, students have found the data content in the book to be somewhat limited.
  
• The instructor also gathers a number of data sets from his group's research related to semiconductor manufacturing control and modeling that may form the basis for a good term project.
  • Ring oscillator frequencies in a variation test chip
    A rich set of data (about 2000 ring oscillators per chip) relating frequency to different layout designs. Potential for spatial modeling, anova modeling, nested variance analysis.
  • Spatial epitaxial thickness and resistivity DOE
    A fairly large fractional factorial experiment on epi growth was performed, with spatial measurements of thickness and resistivity. Interesting spatial modeling and uniformity might be possible.
  • Real time signals from plasms etch spectral OES system
    About 1000 channels of time-series data, for a collection of plasma etches. Time-series analysis, combined with multivariate models, might be of interest.
  • CMP experimental design
    A design of experiments in a number of process parameters, with wafer film thickness measurements gathered for approximately nine sites. Comparison of site vs. overall modeling approaches might be possible.

Term Project Reports

The following are term project reports and presentations for the course 6.780, Semiconductor Manufacturing, presented in the Spring Term 2003.