Cryogenic Experiments on Passive and Active Electronic Components

In this episode, Shahriar investigates the theory and experimental results of the impact of extreme low temperatures on passive and active components. Liquid Nitrogen in used in a transparent glass Dewar where different components can be fully submerged in the liquid.  Various types of resistors are compared for their temperature stability. An electromagnet which uses Copper coils is used to generate a magnetic field at a constant power consumption at both extreme temperatures. The impact of liquid nitrogen on the junction voltage of an NPN device is measured as well as the frequency shift of a CMOS ring oscillator. Finally, the wavelength shift of an LED submerged in liquid nitrogen is studied. There is a puzzle at the end of this video, please share your thoughts in the comments section. All documents can be downloaded from here.



  1. Barney says:

    Hi, its nice post about media print, we all know media is a impressive source of information.

    Look at my blog: click to investigate

  2. Rami Jabakhanji says:

    Hello Shahriar,

    I want to start by saying thank you very much for your blog and videos. I discovered your site a few days ago and cannot stop watching.

    With regards to the resistors experiment, the relationship suggests that a decrease in temperature leads to a decrease in resistance, which I thought was the case. However, the results of the experiment displayed the opposite behavior. Is there anything I am missing?



  3. john senchak says:

    When the LED’s where placed in the liquid nitrogen, the current flow should have been measured to see if it increased or decreased . Also I would have liked to see a capacitor (change in capacitance ) or even a power transistor (to-220 unit ) used to see what the results would be.

  4. john senchak says:

    Was the .6 volt drop between the base to emitter PN junction of the diode (transistor) the same voltage needed in the class A/B amplifier video that keeps it right at the state of being turned on (biasing), so that it becomes linear?

  5. psywiped says:

    Don’t materials get denser as they get colder? Wouldn’t this cause the gap to shrink causing the color change?

  6. Rusala says:

    Hello. Very interesting video. As for the strange LED behavior we’ve discussed it shortly with my colleagues (we work in mobile telecommunication but some people have either physics or semiconductors background). The suggestion is that LED manufacturer probably doesn’t bother to keep the material very pure when it comes to such simple LEDs so that effect could be in fact caused by indirect band gap occurring in low temperature when some other bands come into account.

  7. […] Shahriar from The Signal Path super-cools an LED…and it changes color! Awesome! […]

  8. Ryan says:

    Hey Shahriar, another awesome video!

    I’m very interested in the band gap problem with that green LED, that’s a very curious result.

    I’m currently working on a set of undergraduate labs for a new nanoelectronics course at my university, and I’d be delighted to incorporate these experiments into the material. In addition, my professor is a semiconductor specialist so he may have some unique insight into the bizarre behavior of that green LED (and I have a few additional experiments in mind that may shed some light). If you haven’t sent those electronic popsicles to anyone yet, I would be honored to put them to good use. I’m pretty sure we have a glass dewar somewhere in the lab, too.

    Thanks for all the great videos.

    • Shahriar says:

      Hey Ryan,

      Thanks for looking into this interesting LED effect. I can definitely send you the parts. Send me an email (found at the bottom of this page) with your shipping address.

  9. Ning says:

    Hi Professor, this video is perfect. I’m a student of Columbia University who took your EE4312 course last year. I have some questions. May I have your email address? I send email to your Columbia mailbox and look forward to your reply.

  10. Alvaro says:

    Did the LED go back to green when it was back at room temperature? Maybe it was a permanent change in the crystal structure which caused the bandgap to drop. Direct bandgap means that the lower eedge of conduction band lines up with the upper edge of valence band. I would imagine the position of these valleys and peaks shift with temperature, perhaps making it an indirect band gap and thereby changing the energy of emitted photons.
    This could really be a shift in either direction though, and doesn’t really explain the shift towards longer wavelengths.

  11. Gerd says:

    Fun stuff this liquid nitrogen. If you play with it again – could you try a crystal oscillator, e.g. a standard hc49 like used in most microcontroller circuits. I’m interested if it still works or if the freezing will prevent it from swinging.

  12. dude says:

    ah, but a transitor is

  13. dude says:

    a pn-junction/diode is a passive device, not active

  14. dude says:

    B = μIN/l
    just for precision sakes…, I mean for the units alone…

  15. Wartex says:

    hey I’m from NS but I’m in Toronto for a couple of days, I’m a big fan of your blog, I’m trying to get into electronics/academia again and have some questions, do you have time to go for a beer somewhere this week? I treat!

    • Shahriar says:

      Hi, Thank you for the kind comments. I actually do not live in Toronto any more. I live in the USA. Feel free to drop me an email, hopefully I can offer some help.

  16. […] Shahramian] is playing with some liquid nitrogen in order to see how various components react to extremely low temperatures. After the break you will find forty-one minutes of video in which he conducts and explains each […]

  17. enclis says:

    Nice experiments! Why don’t you test capacitors with different types of dielectric materials?
    Measurements of wavelength change vs temperature of green LED already have been done here – . Maybe you should take another type of green LED for completeness of the experiment.

    • Shahriar used a GaP:N (nitrogen-doped gallium phosphide) green diode, as can be seen from the yellow-green emission at room temperature, as opposed by the InGaN blue-green LEDs.

      In the absence of LN2 I sometimes show students the opposite effect: letting a green GaP:N diode heat up by too high current will _also_ shift the emission to longer wavelengths, i.e. the green LED turns yellow. And this can be easily explained by the narrowing of the bandgap at higher temperatures.

      Last year I supervised a student project where the students studied the emission spectrum of a white LED at different temperatures – and there as well, the shift of the underlying blue emitter followed the trend of a narrower bandgap at higher temperatures.

      However, it is very interesting to see your experiment, and yes: you should definitley try different LEDs. Did you keep the current through the LED constant? If not: does this influence the observation?

      Somewhere in the video you say that the sample will have room temperature inside the Dewar as long as it is not immersed in the LN2 – this is a very rough estimation… In my previous university we controlled the temperature of a sample in a Dewar by adjusting the height of the sample above the LN2 level, and as long as you are inside the Dewar it’s still pretty damn cold! It might be interesting to study the color shift in more detail at different temperatures – just also mount your temperature-sensor 2N2222A on the same holder as the LED and keep it some distance above the LN2. Keep the current through the LED(s) constant and watch the shift. Our eyes are very sensitive to minute wavelength changes in the green-yellow-orange part of the spectrum!

      GaP:N is an indirect-bandgap semiconductor, which is somewhat special. The green emission is not a band-to-band emission, but a trap-assisted emission through the nitrogen impurity states. This allows for a device structure where the emitted light is less likely re-absorbed in the material, if the nitrogen-doped region is kept small. See from [E.F.Schubert, Light-Emitting Diodes].

      I am not sure and only speculating here, but it appears possible that the room-temperature emission is determined by a partly empty valence band, i.e. it goes from the nitrogen level to a level somewhat lower than the top of the valence band. At lower temperatures the dopant atomes freeze out, making it more difficult for electrons to be excited from the valence band and thus the transistion will be to the top of the valence band, resulting in a smaller energy difference and hence longer wavelength.

  18. ballooney says:

    Hi Shahriar,

    I recently had to make a precision constant current reference for some sensors which were to fly on a high altitude balloon. I bought some precision 0.01% accurate resistors, which were quite expensive (thankfully only needed a couple) at about $15USD each. Their datasheet advertised them as having a temperature coefficient which was ‘not measurable’. I’d be very interested to see if you could actually measure a temperature change with this experiment. I presume you’d have to use a kelvin connection to remove the contribution from the test leads.

    A random guess about the LED colour thing, which is very ill-informed because I’m not a materials person by any stretch, but recalling my materials science course on my degree, might the crystal structure or packing structure of the materials in the LED be changing with temperature? I remember the classic example of the properties of steel changing when it transforms from a Face-centred cubic structure to a body-centered cubic structure below a certain temperature.

    Thank you very much for this video.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

  • Social Media

  • Administrator

  • Archives

This site is protected by Comment SPAM Wiper.