New Video. How low can you go?

[planters]

I bet you didn't take the bias-component into account.

Looks ok to me...

You are correct sirs. To put it better (and to avoid besmirching Dr. Lava's product), my CALCULATED input currents are far from accurate. Also, to avoid being accused of too much enthusiasm for this process, the acceleration of output with current is subtle when you take into account the threshold current. But, there is no rolling off and so I suspect the safe power limit even at this "conservative" -80C is still higher. What needs to be tabulated (and as you noticed Dan, I hate making tables/graphs within a post) is threshold current, V. drop, input current, power out and wavelength. Also, the ability of the cold diode to modulate at a useful frequency needs to be established.

This information could be obtained with a computerized data logging program as the diode is cooled. With a cryogen and without temperature control through heating of the cold finger the process of cooling is rather quick. Nevertheless, 10-12 data points down to -80C should be plenty and maybe 10 more down to -196C should do.

Research @ this level is no longer amateur. I have no doubt that a paper could be written based on these results, but then an explanation of the physical process as well as a better description of the diode design/ condition would be required. Maybe hit Mitsu up for a few bucks? Seriously though, I am surprised @ the lack of published information about cooling of non-IR diodes, but I wouldn't be surprised to learn that the major manufactures of these diodes have tested this.

[-bart-]

I wouldn't be surprised to learn that the major manufactures of these diodes have tested this.

I bet laser diode manufacturers have diode lasers running cryogenic at the most exotics wavelengths, (even between 520 and 620 ?)
I would not be surprised if the Real Challenge is obtaining reasonable lasing threshold at room-temperature.

[danielbriggs]

I would not be surprised if the Real Challenge is obtaining reasonable lasing threshold at room-temperature.

It is! Nakamura's book describes the GaN developments well.
http://www.springer.com/physics/opti...-3-540-66505-2

My own aim for LN2 is to see if it is possible to hit 600nm from a "red" diode.

@Eric, Re: datalogging:
I began to write a LabView VI for control of my Melles diode + TEC controller, with the long term aim to hook into the spectrometer output + Coherent meter power output. This should give a complete "one screen" datalogging solution, showing centre wavelength, power, diode v_f, temperature, current etc. etc. Note to self: Finish this for the LN2 setup.

I haven't ordered the thermocouple reader yet; I might try and find one with USB or RS-232 connectivity.

[planters]

Note to self: Finish this for the LN2 setup.

Well, this is promising. At this point your test set up will be the better, more complete route and so I'll hand the torch to you on the red wavelength work. I'll still peruse the MO power aspect and I'll also look at the 445's. These will shift very little, but again the power output will be interesting as will be the opportunity to eat.

[krazer]

Per the comments on temperature drop across the junction causing a wavelength shift, yes this does occur however the increasing current itself does also play a role. It is hard to tell the difference between the two (especially because the relationship between current vs wavelength and temperature vs wavelength are not very linear), however you can do pulsed measurements to sort out the two (the time constant of the current based shift will occur on the order of ps or ns, the temperature shift will be much slower). I decent paper is up at http://www.sciencedirect.com/science...21452611008362 (probably behind a paywall), which measured a 3nm shift going from 10ma to 80ma for a 450nm diode (red diodes should have an even larger shift). I am not sure what diode they used, I suspect it is something similar to the osram pl450b. Note - they were not looking at the lasing output spectrum, but rather the peak of the non-lasing florescence spectrum (ie, gain spectrum), however they cite other works which show that this is a valid method for measuring the shift in peak wavelength while lasing.

Also, I am curious to see how the diode physically handles being run at LN2 temps, from the ideal gas law we get that the pressure inside the can will be 1/3 of atmosphere, so for a 1.6mm window you are looking at about 50g of force. Then if you actually get it in equilibrium with the LN2 presumably the nitrogen in the can will start condensing on the inside of the diode...

[danielbriggs]

Also, I am curious to see how the diode physically handles being run at LN2 temps, from the ideal gas law we get that the pressure inside the can will be 1/3 of atmosphere, so for a 1.6mm window you are looking at about 50g of force. Then if you actually get it in equilibrium with the LN2 presumably the nitrogen in the can will start condensing on the inside of the diode...

I was planning to run the P73, a cap-less diode. Or if I run another diode, de-can it.
I'm hoping that gets around such issues.

[planters]

Krazer,

I will look at that paper ( if it isn't too expensive). Do you remember if they provide a mechanism for current induced shift?

I do not think condensation will be a problem because no matter how hard we try, the temperature in the diode chamber and specifically at the diode will be above 73 K. However, I had been contemplating putting a vacuum above the LN2 if this was successful down to the LN2 range and so grab another 13K, but that might truly bring about the issue of condensation and never ending cryo-pumping into a chamber if it is less than vacuum tight.

I would like to leave the windows in place if they can stand the pressure and so when I test the 445 I'll leave the can on. I do not know if the 9mm can has a larger window than the M series and so each might have to be evaluated separately.

Dan,

While we wait for your set up, I did try a low power (80 mW) run to -80 and the wavelength was 12 nm short of HeNe (621nm) @ 1.82A I hit 2.00W with an 8 nm red shift to 629nm. I think @ -196C your goal of 600nm might be possible if the power is kept low.

[planters]

I can't wait till we have a spectrometer on this.

I was playing around and wanted to see what the thermal gradient is for my rough set up. So, neglecting the obvious calibration error in the thermometer, I measured the temperature at 2A and at "off" (50ma) where the diode is below threshold. The gradient is 10C and so I will be looking at this if all goes well with an aluminum clone of Dave's mount (unless I can convince some brave soul to fabricate it in copper). But this is not the interesting part by a long shot. The emission at threshold is noticeably more orange. No surprise here; the diode is 10C colder and the current is lower. But, I thought I was having trouble with the monochrometer because it would not snap from light to dark as I passed the peak emission. There was no obvious peak. The diode was emitting from the dark green smoothly through to the IR. There was a slow and gradual increase in intensity in the 600-630 region, but it was not overwhelming. This is very surprising to me. I thought LEDs were narrow band as in 10-2o nm emitters that sharpened up when a cavity forces them to lase. What I am seeing is a tremendous bandwidth on the order of 200nm. There are videos where an led submerged in LN2 emits at a higher frequency, but when constrained by a cavity the shift is less pronounced.

https://www.youtube.com/watch?v=U6dUBiEIBKQ

https://www.youtube.com/watch?v=_AbQh1g_1ok

I suspect if the cavity was fabricated to favor shorter wavelengths such as 595nm that AlGaAs could do this at a lower power output. If the diode was DESIGNED to operate @ LN2 temps then the shifting of the emision peak coupled with a cavity designed to take advantage of this shifted peak could produce yellow at HIGHER output than the red!

[planters]

Here is an interim update. I obtained some LN2 and after adjusting the driver to 3A @ 5V modulation, replacing what turned out to be a bad thermistor I ran another test. I set the drive to 600mA and using dry ice I cooled the diode down to -75C. Then I began to add LN2 to the cold reservoir until the diode temperature bottomed out at -183C. The color became quite orange and I would estimate the wavelength to be close to 608nm. Next I slowly ran the current up until I reached 2.42A and 2.642mW at which point the diode went LED. Just prior to its destruction the wavelength was close to 616 nm. The diode continued to conduct and the current and V drop were little changed after its failure. The LED emission had a noticeable asymmetric wing and so I am pretty sure the cause was a catastrophic facet failure.

I am waiting for a spectrometer and rather than pursue huge output power @ baseline wavelength ( the most you're going to get is a factor of two increase in power) I am going to look @ accurately measuring color shifting with a number of diodes @ moderate power levels.

[andyf97]

Hey Mr Planters, when you do this again, split the beam off and direct it 1 meter to a screen and do regular shots and measurements of the line.