| Page 1 | Page 2 | Page 3 | Page 4 | Page 5 | Page 6 | Home Page |
of a HEMT structure. The surface map is generated from a 55 point sampling pattern. The coefficient of variation for this wafer is 1.2%. Cross-Check and Mobility Measurement The sheet resistance measurement capability of the LEI Model 1400 EMA mobility instrument enables a cross-check against the non-destructive sheet resistance mapping instrument. The mobility instrument, which is based on an invention at IBM Watson Research Center by Dr Norman Braslau, with improvements during and after the Ben Franklin Program at Wilkes University, enables non-destructive mobility measurements on production wafers. Feedback from Litton from their use of the instrument during and after their MIMIC programme has been used to enable measurement of the 2DEG mobility. Metallization Monitoring An instrument with three-range (high, low, extra-low) capability provides for measurements of films ranging from Ni thin films with high sheet resistance to interconnect metallizations (Ti, Pt, Au) with very low sheet resistance. The non-contact RF measurement eliminates four-point damage and metal contamination from the points and enables measurement through oxides and coatings. Since there are no points to lower and raise during multi-point measurements, the measurement time for a 55 point plan can be as fast as 3.5 minutes, with improvements being implemented. |
Measurements may be made on structured and patterned wafers and materials with grain boundaries without concern for where the points land. Very repeatable measurements have been made on 200 mm silicon wafers with poly, SiO2, Si T, Al, and WSi thin films. In-line measurements (load lock) may be made for immediate feedback and correction of thin film deposition before removing the wafer to room atmosphere. Financing It should be understood that the cost savings mentioned above were realized in a production environment. Research and development applications with smaller quantities have reported savings of over $30,000, based on the detection of substandard epitaxial wafers before additional (and expensive) labour and materials have been added. These potential savings, combined with our records which show that most purchases require a six-month to one-year minimum waiting period, can result in interim losses of $100,000 - $750,000 in terms of possible yield improvements. It is our hope that a "win-win" situation can be achieved, whereby the instrument manufacturer can be paid a fair percentage of the yield improvement in return for earlier instrument delivery to the process engineer, without the risk of taking a loss on the sale of the instrument. Such a financing agreement would benefit all parties involved. Conclusions Many of the applications described for GaAs are applicable to silicon and other conductive materials as long as the sheet resistance of the processed layer is less than the substrate sheet resistance by a factor of at least 10. The instrument size can be varied to accommodate the size of the material being measured. Therefore, the yield improvements described are not limited just to wafers. Acknowledgements Many thanks to the original authors of the excerpts used and the conferences and symposia where the papers were originally presented. We want to thank everyone who contributed to, typed and proofed this paper. We also want to acknowledge the hard work of the employees who have helped in the design and manufacturing of the instruments that have made the results shown possible. We appreciate the feedback from our customers of ideas for improvements. |
| Page 1 | Page 2 | Page 3 | Page 4 | Page 5 | Page 6 | Home Page |