This is not a PCB question and asks only for a single MOSFET. I always wondered if there is a way that a single MOSFET would work at (any) very low temperature. I tried to find some source that answers this question and returned with none satisfactory.

Would junction diffusion process stop in very low temperature or something like that?

  • 4
    \$\begingroup\$ define "very low" \$\endgroup\$ – W5VO Nov 22 '15 at 13:54
  • \$\begingroup\$ way below room temperature. \$\endgroup\$ – Dagaka Mademois Nov 22 '15 at 13:58
  • 5
    \$\begingroup\$ Define "way below". \$\endgroup\$ – Nick Johnson Nov 22 '15 at 14:12
  • 1
    \$\begingroup\$ Can you give and aproximate of such temperature in Celsius degrees? \$\endgroup\$ – jotadepicas Nov 22 '15 at 14:40
  • \$\begingroup\$ "Way below" is unhelpful. Have you tried looking at a few data sheets? \$\endgroup\$ – Andy aka Nov 22 '15 at 20:21

Well if you take one example, say a IRF740 power MOSFET.

Look at the datasheet it will give you a maximum operating temperature, and a storage temperature. The low end of the storage temperature is -65deg.

Of course this would be a starting point. As soon as you start operation and switch a load through the FET, current will flow and inevitably cause the junction temperature to rise. The more current that flows the greater temperature rise. Generally electronics likes 'cold' temperatures, as in just above zero, but in some applications very cold ( -100 to -50 ) or even colder are needed to reduce thermal noise. Take a low noise RF amp as an example.

Anyway to answer your question, yes they can operate at very low temperatures. And can get better efficiency at low temperatures because your Rds-on will be lower at lower temperatures.

|improve this answer|||||

Yes, Silicon FETs can operate 'easily' down to around 77 K (-196 C), although the characteristics will (obviously) change -- some (e.g. on-resistance) for the better. Threshold voltages will be significantly (~400 mV) higher.

Diffusion processes don't stop at cold temperatures -- but the doping (which makes it n- pr p-type) freezes out which basically means it looses effectiveness. Conversely, the impediments to carrier motion which degrade conductivity decrease. This is what makes the resistance lower.

|improve this answer|||||

At 0 Kelvin: everything stops so no, a MOSFET will not work.

Above 0 Kelvin: it will probably work.

Also see the datasheet of your favourite MOSFET, the manufacturer will list the minimum (guaranteed) operating temperature. That does not mean the MOSFET will not work, it's just not guaranteed (and not tested) to work.

|improve this answer|||||
  • \$\begingroup\$ Luckily nothing can ever reach absolute zero \$\endgroup\$ – Robert Stiffler Nov 23 '15 at 1:50
  • \$\begingroup\$ @RobertStiffler Are you sure ? Can you prove it ? In a practical sense we can get close enough to 0 Kelvin. It's like this one: reddit.com/r/Jokes/comments/2ijugj/… \$\endgroup\$ – Bimpelrekkie Nov 23 '15 at 13:20
  • \$\begingroup\$ See the third law of thermodynamics. No finite number of steps can reduce the temperature of a system to absolute zero. We can get arbitrarily close, however. Even still, at absolute zero there is still quantum ground state energy, just no thermal energy. \$\endgroup\$ – Robert Stiffler Nov 23 '15 at 13:30

Not the answer you're looking for? Browse other questions tagged or ask your own question.