Ultra low temp measuring techniques

Johnson noise limits the sensitivity you can get out of a thermistor (or a plantinum resistance sensor) to about 1e-6 Kelvin at room temperature.

I first came acreoss the relationship in Larsens's classic paper

Larsen N T 1968 Rev. Sci. Instrum. 39 1=9612

and published a comment that discussed the implications of the relationship some ten years later

Sloman A W 1978 J. Phys. E: Sci. Instrum. 11 967=968

You may need to think this one out again.

-- Bill Sloman, Nijmegen

From what you've said already I doubt I can contribute much, all I can suggest is that you eliminate the noise. The difficulty is distinguishing unwanted electrical noise from the thermal noise that you want to measure. Have you considered an RTD?

Do not forget Heisenberg.... In order to measure something, you WILL change it..

Google "Sub-microkelvin measurement"

Tom Davidson Richmond, VA

Convert 10^(-7) K to kilocal/mole, eV, cm^(-1)... Even if you could do it with temp-modulated Faraday rotation, how would you prevent the interrogating beam from swamping the signal?

If the temp gradient deforms the substrate (coefficient of thermal expansion),

"Ultrasensitive Beam Deflection Measurement via Interferometric Weak Value Amplification" Phys. Rev. Lett. 102, 173601 (2009)

om/uncleal/lajos.htm#a2

Thank for all helpful advice. Bill's reference to a 1e-6 Kelvin limitation appears to match my 1e-6 K limit. Uncle Al and Robert Baer made a valid point regarding the sensing device interfering with the measurements. The only idea that comes to mind is perhaps to measure the blackbody radiation emitting from the DUT; i.e., a highly sensitive non-contact FIR thermal chip. Personally I've never heard of such sensitivity from FIR thermal chips. Perhaps several hours worth of sampling in a shielded controlled system could achieve sub- microkelvin measurement?

Regards, Paul

The 1e-6 Kelvin limit isn't absolute; it's just the way the numbers work out in typical situations with resistance sensors. In my comment

Sloman A W 1978 J. Phys. E: Sci. Instrum. 11 967=968

I discussed how you might lower the Johnson noise limit. It involves dissipating appreciable power in the sensor, which quantifies the point made by Uncle Al and Robert Baer.

-- Bill Sloman, Nijmegen