HBT Epitaxial Material Matching and Qualification for High Volume Production

Mike Sun, Peter Zampardi, Cristian Cismaru, and Lance Rushing

Skyworks Solutions, Inc., 2427 W. Hillcrest Drive, Newbury Park, CA 91320 mike.sun@skyworksinc.com (805)480-4456

Keywords: HBT, Epitaxy Materials, SIMS, QL

ABSTRACT

Reliable and reproducible epitaxy (epi) materials are critical to the success of a high volume compound semiconductor fabrication line. An efficient epi matching and qualification procedure among different epi reactors and suppliers is paramount to supporting rapid fab capacity ramp to keep pace with market demands. Traditional methods of individual layer thickness and doping matching, such as Secondary Ion Mass Spectroscopy (SIMS), have limitations when applied to heterojunctions and subtle interface matching. Sometimes a subtle epi interface mismatch can create significant discrepancies in transistor characteristics and RF module performance. Multiple iterations of lengthy epi matching as well as RF module level characterization may be required to resolve the differences. To overcome these limitations, we developed an epi matching procedure which greatly shortened the cycle time for epi qualification. After an initial “quick-lot” matching is performed by our suppliers, extensive transistor level electrical data is collected and compared to the historical production distribution. The measured data of these qualification samples must be within 2.5 sigma of the historical distribution with the same mean. The in-line DC PCM parameters, for the technology generation, are a subset collected data. In addition, an extended set of DC and RF parameters is also matched within the same stringent criteria. After the transistor level electrical data fully matches the historical distribution of our database, the qualification samples then go through transistor level reliability and module level evaluations. Adequate transistor level reliability and comparable module level performance are required for a full production release of a new epi reactor.

11.2.PDF                      Return to TOC