Browse Prior Art Database

Thick Film Optical Measurement

IP.com Disclosure Number: IPCOM000086248D
Original Publication Date: 1976-Aug-01
Included in the Prior Art Database: 2005-Mar-03
Document File: 2 page(s) / 40K

Publishing Venue

IBM

Related People

Ewalt, AB: AUTHOR [+3]

Abstract

The thickness of a composite transparent film layer on a silicon wafer is determined by Fourier transforming the reflectivity vs. wavelength spectrum of the sample wafer, and finding the peak in the power spectrum. This technique has advantages over similar optical measurements in that it is more immune to noise, less sensitive to wavelength calibration errors, and extends the maximum total thickness which can be measured from ~ 55,000 Angstroms to 80,000 Angstroms without requiring any hardware adjustments.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 57% of the total text.

Page 1 of 2

Thick Film Optical Measurement

The thickness of a composite transparent film layer on a silicon wafer is determined by Fourier transforming the reflectivity vs. wavelength spectrum of the sample wafer, and finding the peak in the power spectrum. This technique has advantages over similar optical measurements in that it is more immune to noise, less sensitive to wavelength calibration errors, and extends the maximum total thickness which can be measured from ~ 55,000 Angstroms to 80,000 Angstroms without requiring any hardware adjustments.

The thickness range measurable with this technique is 15,000 Angstroms to 80,000 Angstroms with a 30 precision of better than 3%. The analysis procedure is flow charted in the accompanying figure. Once the sample reflectivity spectrum R(lambda) vs. lambda is measured, an approximate thickness TAPRX is computed: TAPRX = - 6.9 + 3.1x - 0.09x/2/ + (2.26 - 10/-3/ - (9.4.10/-6/)x/4/ where x = number of extrema in the reflectivity spectrum from 400 nm to 680 nm.

The thickness approximation is used to define the domain in frequency space over which the power spectrum is to be computed. The frequency domain consists of 45 equally spaced W-points in the interval;, W Sigma [.8 x TAPRX,
1.2 x TAPRX]

In order to obtain a final result in thickness units, a change of variables is made before finding the Power Spectrum: x absolutely = 4 pi n(lambda) over lambda. R = R(x)

One more preprocessing step is performed before computing the power spect...