Ask the Staff

james.a.cooper.1 2012-02-15 02:47 PM
Quick non-destructive measurement of oxide thickness
We'd like to use ellipsometry or similar technique to measure oxide thickness on SiC. I think in the past we've had poor results using elliposometry, due to differences in SiC optical parameters relative to silicon, which usually works well. The technique needs to be quick and non-destructive, to allow us to continue processing after the measurement. Would the Filmetrics be appropriate? Would the wide band gap and short absorption wavelength of SiC be a problem?

Any advice or suggestions would be appreciated.

geoffrey.c.gardner.1 2012-02-15 03:35 PM
REPLY: Quick non-destructive measurement of oxide thickness
The Filmetrics may be a great candidate; it is indeed quick and non-destructive.

For this tool to work white light, (~400-1000nm), needs to pass through the oxide, reflect off the SiC, pass back through the oxide a second time and be received by the detector. So depending on the oxide type and thickness (transparency), and the SiC refractive index (reflectivity), the described process may or may not be able to take place.

Please be aware the measurement area is ~3mm, so it cannot interrogate a single feature. The useful thickness range, which is dependent on the process described previously, is ~15nanometers -100 micrometers.

Can I assume you have SiO2 on single crystal SiC? How thick is the oxide?

My advice is let’s try it!

james.a.cooper.1 2012-02-16 10:24 AM
Filmetrics for SiC oxide thickness measurements
Geoff,

Thanks for the quick response!

Dallas Morisette tells me the absorption edge is probably not the determining factor, since reflections will occur at the SiO2/SiC interface due to the difference in refractive index, which is similar to that of silicon. We have a SiC sample with known oxide thickness that we can use to verify and calibrate the Filmetrics, so we'll try that. Steven Swandono has the sample, and will probably contact you to get checked out on the Filmetrics.

-- Jim