Jaideep Mavoori wrote on 2 Aug 1999:
> Vijay,
> A good non-contact method to measure the temperature of a wafer is
> by a pyrometer.  They do face the limitation of not working well at
> the lower temperature range and assume that you know the emissivity
> of the substrate.
>
> In your case, wafers with polished/non-polished and with thick
> oxide, would have varying emissivities.  An equipment vendor I know
> of (I think most of this technology is patented by the vendor),
> solves the problem by using a highly reflective area on the back
> side of the wafer, in effect, creating an approx ideal black body.
> Of course, it is not perfectly ideal, and non-idealities are
> compensated for by using a emissometer.
>

Hallo,
I am sorry I missed to answer this posting sooner - I had intended to
do so.
We are using an Agema thermographic system with microscope
lens and working in the far infrared range to characterize our
devices. We have made good experiences with this system and have even
been able to measure the temperature distribution on 20 micrometer
thick membranes with an accuracy (not resolution) of 1 K. We have
found, though, that the emittance of nearly every specimen is
different since it depends on many factors. A further difficulty lies
in the fact that silicon is partially transparent in the infrared so
that radiation from background is measured, too. We therefore start
measuring the emittance first which requires exact calibration. In
the case I mentioned, this was only possible by using integrated
thermoresistors on that membrane to define the very value of the
calibration temperature.
If anyone needs a measurement on a device, we would be glad to help
out.
Yours sincerely, Kai Hiltmann
--
mailto:Kai.Hiltmann@HSG-IMIT.de    http://www.HSG-IMIT.de
K. Hiltmann; HSG-IMIT; Sensorics Section
W.-Schickard-Str. 10; D-78052 Villingen-Schwenningen
Phone ++49-7721-943-132; Fax ++49-7721-943-210