大部分的學術論文都是透過一個已經知道數據的測試結構來驗證某一款寄生參數抽取軟體的準確性，又或者是研究出來一些新的經驗公式或是模擬軟體，藉著時間複雜度來說明本身的軟體是優於市面上的軟體。

在這一篇論文中，我提供一些公正的測試結構集合，而且我並沒這些測試結構的數據，但是藉由這兩款軟體，CLEVER和Raphael，來計算公正測試結構集合的寄生電容數值。這些公正的測試結構集合分成五種測試結構的型態。藉由CLEVER和Raphael所計算出來這五個測試結構型態的數值來證明這兩款軟體在某一些應用上是值得信賴的。

我在Linux的作業系統下，使用PERL寫了一個產生CLEVER跟Raphael專用的測試結構產生器。這個測試結構產生器可以產生五種測試結構的型態，而且還可以計算出來這些測試結構集合的寄生電容的數值，透過這些數值產生這兩款軟體的相對誤差清單與XGRAPH的統計圖表。透過這些資訊來評論這兩款軟體的在準確性、時效性、使用者介面的便利性以及所需的計算資源對台灣IC公司有何效益。

A lot of theses prove the accuracy of field solver through a test structure set with known data. Some theses would provide some empirical formulas or a simulator, and these theses prove their empirical formula or simulator is better than field solver through time complexity.

In my thesis, I provide a rich set of impartial test structure sets without the data of parasitic capacitance. And, my thesis uses two field solvers, CLEVER and Raphael, to measure the values of parasitic capacitance from these impartial test structure sets. The impartial test structure sets are divided into five types of test structure. According to the five types of test structure, I prove CLEVER and Raphael which are correct through the values of parasitic capacitance.

I write an automatic program using PERL in Linux operating system. This automatic program can generate CLEVER and Raphael test structures. It can also gather statistics such as the value of parasitic capacitance, errors of two pieces of software, and the corresponding XGRAPH diagrams. By these information, I comment on the accuracy, speed, user-friendliness and computing resource between CLEVER and Raphael for the benefits of the IC industry in Taiwan.

[1] Froment, B., Guichard, E., Borot, B., Hanriat, S., Cluzel, J., Schoellkopf, J. P., & Jaouen, H. (1999). “New interconnect capacitance characterization method for multilevel metal CMOS processes,” IEEE Proceedings of the International Interconnect Technology Conference. (pp. 224-226)

[2] Chang, K. J., Oh, S. Y., & Lee, K. (1991). “HIVE: An efficient interconnect capacitance extractor to support submicron multilevel interconnect designs,” IEEE Proceedings of the International Conference of Computer-Aided Design.(pp.294-297)

[3] Goel, A. K. (2004). “Capacitance extraction for the nano-scale on-chip interconnects,” IEEE Proceedings of the International Conference on Software Engineering. (pp. 112-116)

[4] Chern, J. H., Huang, J., Arledge, L., Li, P. C., & Yang, P. (1992). “Multilevel metal capacitance models for CAD design synthesis systems,” IEEE Electron Device Letters. (vol. 13, no. 1, pp. 32-34)

[5] Lin, M. S. (1990). “Measure capacitance coefficients of multiconductor microstrip lines with small dimensions,” IEEE Transactions on Componets, Hybrids, and Manufacturing Technology. (vol. 13, no. 4, pp. 1050-1054)

[6] Silvaco. (2004). CLEVER user’s manual.

[7] Synopsys. (2003). Raphael manual.

[8] Sakurai, T. & Tamaru, K. (1983). “Formulas for two- and three- dimensional capacitances,” IEEE Transactions on Electron Devices. (pp. 183-185)

[9] Yang, Z., Wang, Z. & Fang, S. (2001). “A virtual 3-D multipole accelerated extractor for VLSI parasitic interconnect capacitance,” Proceedings of the ASP-Design Automation Conference. (pp. 214-217)