Hi Boris,

please have a look at the Hamamatsu-link. There you will find 10x10mm ones, having a ceramic housing and owing different photo-sensitivites up to NIR, but IR suppressed. Please have a look.

https://www.hamamatsu.com/jp/en/product/alpha/P/4103/index.html#sipd_pkg1=1

Ah yeah, I'm on Hamamatsu already.. But unfortunately their standard packaging seems to include the troublesome epoxy encapsulation of the wire bonds. But I'm waiting to hear back from their technical support.

Thank you for your suggestion!

Hi Boris,

please have a look at the S1227-1010BQ. There is no resin potting, but quartz window. But you are right; there is no hint on how the leads are fixed.

https://www.hamamatsu.com/resources/pdf/ssd/s1227_series_kspd1036e.pdf

Dear Boris,

May be one of the solutions is to use Si photo diode chips and mount them your self such that the mount can withstand the required low temperature 80-100 K.

In an experiment to measure the transistor chip  we mounted the chip on a PCB cupper clad board with silver epoxy and bonded gold wires between the pcb bond pads and the bond pads on the chip to access the diode electrically. Then wires are soldered to the terminals of the diode at the PCB boards. 

For more information about the mount please refer to the link:https://www.researchgate.net/publication/226388117_The_dependence_of_diffusion_length_lifetime_and_emitter_Gummel-number_on_temperature_and_doping

You can get the photo diode chips following the link:http://www.vishay.com/photo-detectors/pin-photo

or any other vendor

Best wishes

Article The dependence of diffusion length, lifetime and emitter Gum...

Abdelhalim - thank you for the suggestion. That is an option we are looking into (although using OSI chips instead - they come in large formats). Even though we have the capability to build our own packages, I was hoping to avoid the hassle and just buy a ready product.

Dear Boris ,

There is no hassle at all. Even you can cut a very small solar cell from a commercial solar cell and use it as a photodiode. It will work satisfactory. Silver epoxy is available and also ultrasonic wire bonders ar to find in many semiconductor laboratories. I think the commertial products are normally designed to work from -55 degree centigrade upwards.

wish you success

As Abdelhalim mentioned, the commercial products are not designed to working at  cryogenic temperatures (80 K - 100 K), even you can find a commercial device with structure matching your requirement, they will be very easy to failure when operated at cryogenic temperatures.

The device must be special designed for cryogenic temperatures.

For the benefit of future searchers - OSI Optoelectronics can supply such photodiodes (no window cover, no wirebond encapsulation) as a special order.

Boris,

many thanks for the feedback.

Dear all,

I guess that the AXUV-series from optodiodes may work as well. They can be ordered window-free, and do not have epoxy-bonding as far as I remember ... Check e.g.

http://optodiode.com/pdf/AXUV100Ti-C2.pdf

or http://optodiode.com/products.html

Good luck, Dirk

I tried to reach out to Optodiode to confirm, and never got a good answer. Also, I think junction capacitance, noise, and NEP of their sensors aren't as good as OSI and Hamamatsu.

We have some windowless AXUV diodes (with 10x10 mm^2 and larger areas) and they behave very well. They were specified to work down to -20°C, and I should mention that we have used them for X-ray fluorescence measurement for samples kept at liquid nitrogen temperature. So they came quite close to the low temperature reservoir, and their temperature was certainly below -30°C during our measurements. So it may be worth to give them a try. We did bought them from Optodiodes, but from IRD radiation detectors:

International Radiation Detectors, Inc.

2527 West 237th Street Unit A

Torrance, CA 90505-5243, USA

However I feel that their website (www.ird-inc.com) is dead. maybe Optodiode took them over? bloomberg als gives an alternative address

750 Mitchel Road, Suite I, Newbury Park, CA 91320, USA

Checking the internet I found another resource for the AXUV series (http://www.rad-dvc.co.jp/images/AXUV100GX_PR.pdf), there they claim that expoxy on wirebonds is used. So maybe not what you want ... Good luck in any case, Dirk

Hey, thanks for doing the background research, Dirk! I've been in touch support at Optodiode itself, so I think you can buy the AXUV photodiodes from them directly (in case you need more).

We don't really understand the epoxy's mode of failure very well. It would take a concerted effort to figure out exactly when and how it fails. While we do cool it down to about 90 K, we've started controlling the cooling descent so it's not as abrupt as simply dumping LN2 into the cryostat. The diode I'm using now has survived about a dozen thermal cycles.

Then again, with the previous failure, there didn't seem to be any signs of fatigue degradation. One day it just popped and was dead.

Oh, there was another reason. AXUV diodes are optimized for X-Ray and UV detection, which means the silicon layer is relatively thin. We are working in NIR, and the reduced sensor thickness significantly and negatively impacts detectivity.