The sensitivity of the parameters of covariance based response surfaces

Author

Waring, T.G. ; Walton, A.J. ; Ferguson, S. ; Sprevak, D.

Author_Institution

Dept. of Electr. Eng., Edinburgh Univ., UK

fYear

1998

fDate

35953

Firstpage

50

Lastpage

53

Abstract

The use of TCAD in combination with the response surface methodology (RSM) and design of experiments (DOE) is now well established in the development of IC fabrication processes. The key to the success of this approach is the accuracy with which the simulated data can be fitted with functions that map the parameter space. Computer based experiments should not suffer from random errors and so it is desirable that the model passes through all the simulated data points. This paper presents a method that ensures that this happens for the response surface model. The paper also examines how the estimated model parameters influence the accuracy of the prediction and compares this with results given by conventional polynomials

Keywords

covariance analysis; design of experiments; integrated circuit design; parameter estimation; semiconductor process modelling; surface fitting; IC fabrication process development; TCAD; computer based experiments; covariance based response surfaces; design of experiments; estimated model parameters; model prediction accuracy; parameter sensitivity; parameter space mapping; random errors; response surface methodology; response surface model; simulated data function fitting; simulated data points; Accuracy; Computational modeling; Computer errors; Computer simulation; Design methodology; Fabrication; Parameter estimation; Predictive models; Response surface methodology; US Department of Energy;

fLanguage

English

Publisher

ieee

Conference_Titel

Statistical Metrology, 1998. 3rd International Workshop on

Conference_Location

Honolulu, HI

Print_ISBN

0-7803-4338-7

Type

conf

DOI

10.1109/IWSTM.1998.729768

Filename

729768