Atomic Engineering Corp.
P.O. Box 3342
Gaithersburg, MD 20885-3342
Tel: (301) 330-2902
aec at AtomicEngineeringCorp.com
About AEC - Advanced Software Products - Monitoring and Characterization Systems - Laser Remote Sensing - Track Record - Contact - Order Form - Download and Demonstrations - Feedback- Links

Real-Time Characterization of Semiconductor Etching Process in a Reactor Measured by Optical Emission Spectrometer

In this example, ACAS is used to demonstrate quantitative determination of spectra drifts in the semiconductor etching process in a reactor as measured by optical emission spectrometer. First, quantitative wavelength and intensity calibration will be carried out.

A. Calibration of Raw Data

When the user has a file containing raw data, he or she may need to calibrate it against standard reference data in order to find the intensity. If the user knows which element is represented in the raw data, he or she should graph the raw data and the standard reference data of the same element. By setting the scales of the two graphs the same, the user can see where the peaks line up. In this example, a raw data file containing mercury is graphed along with the library reference data file of mercury, so that the peaks may easily be compared. The user can now clearly see which peaks are mercury, and which of these peaks are the most intense. The extra peaks in the raw data that do not line up with mercury are other elements, which we are not concerned with. A search should now be conducted, and the wavelengths can be calibrated. The results show that the average difference after calibration is .3933 nm. The user can now determine the intensity of the wavelengths.

B. Drifting Optical Emission Spectra

It can be difficult to measure the drift in the etching process because of spectral shifts and degradation caused by environmental matrix effects. The drifting of optical emission spectra can be seen by examining the graph of measurements taken from three different positions in the same reactor. (After graphing the raw data, annotations can easily be added after searching via the Graphing Window.) Minor variations are visible when the graphs are viewed here on a normal scale. However, the differences become much more apparent when a small area is enlarged. The user can also compare the Peak Data Tables.

C. Quantitative Identification

The concentration of the sample can be determined in terms of the intensity calibration using normalized reference data. The concentration of the sample is the sum of the sub-concentration of all of the "identified elements" within the sample. The sub-concentration is determined by taking the ratio of the sum of the input spectra lines with the sum of the corresponding intensities of the reference data for the "identified elements." To determine the concentration, use the Concentration Analysis Report.

Copyright 2005