Methodology for improved semiconductor process monitoring using optical

Abstract

In a semiconductor process which utilizes a plasma within a process tool chamber, a method of using optical emission spectroscopy (OES) to monitor a particular parameter of the process is disclosed. A first wavelength present in the plasma is determined which varies highly in intensity depending on the particular parameter by observing a statistically significant sample representing variations of the particular parameter. A second wavelength of chemical significance to the process is also determined which is relatively stable in intensity over time irrespective of variations of the particular parameter, also by observing a statistically significant sample representing variations of the particular parameter. These two wavelengths may be determined from test wafers and off-line physical measurements. Then, the intensity of the first and second wavelengths present in the plasma is measured on-line during normal processing within the process tool chamber, and the ratio between the first and second wavelength's respective intensities generates a numeric value which is correlated to the particular parameter. As an example, such a method may be used to generate a reliable alarm signal indicating the presence of etch stop conditions within a plasma oxide etcher, as well as to indicate the oxide etch rate.