Find Cryogenic Components: 1 uF Capacitors & More

[davidrit]

I'm looking for components specified for cryogenic temperatures. Capacitors on the order of 1 uF to start with but other components too. Ideally they will be rated to less than 80 K operation with low outgassing.

Anyone suggestion on any companies that make or sell such low temperature parts?

 

[f95toli]

I don't know of any company that sells cryogenic components like that. As long as you use some common sense you can use most room temperature components even at very low temperatures (by "common sense" I mean that you should obviously stay away from wet dielectrics etc, and you can't use Si semiconductors below the freeze-out temperature of the carriers etc).

The properties of resistors,capacitors etc will obviously change when you cool them so you will have to characterize them at whatever temperature you want to use them at.

A good way to get started is to look in journals such as Review of Scientific Instruments where you can find quite a few papers describing circuits designed to operate at low temperatures, in many cases the circuit diagrams will include the exact type (including manufacturer) of component used.

 

[Mike_In_Plano]

You might try Presidio components components for ceramic caps. I've also had good luck with AVX's glass capacitors (rated to -75C) and these parts are very clean.

You need to either spot weld the leads or use special solder for low temps. Kester solder can set you up with the correct alloys / process.

As for outgassing, Polaris Electronics can set you up with packaging / welding to keep your vacuum side away from the electronics.

As for app notes and possible contacts, Interfet supported a good many scientific endeavors and can likely lead you to some good information on cryo electronics.

Best Luck,

Mike

 

[f95toli]

There is no need to weld or use special solder at low temperatures, ordinary solder works well (at least down to 15mK, I never work below that temperatures). You only have to pay attention to the fact that ordinary 60-40 becomes superconducting at just over 4K, which isn't always desirable (and the thermal conductivity drops).

The only exception I can think of would be if you need a very good vacuum (<1e-7 mBar) at the same time and are working at relatively high temperatures (above 77K) where cryopumping isn't very efficient. Or, if you are doing something where contamination is an issue (such as low temperature STM).

 

[mgb_phys]

We always avoided a silkscreen on the boards to reduce outgassing.

 

[davidrit]

Thanks! Definitely some good tips and sources I didn't come across in web searching.

I don't have access to the Review of Scientific Instruments on-line right now but I'll look into subscription. It looks like a really good resource, I can't believe I haven't looked at it before, it probably would have saved a lot of learning the hard way.

I was thinking ceramic NPO type capacitors. The NPO should keep the capacitance close enough for my needs and I would expect ceramic materials to have low outgassing. From what I've read though most people use X7R for this type of application for some reason, even though it has worse change in C over temperature.

AVX and Polaris have some capacitors with gold terminations and I have an IC die that will be gold wire bonded so I thought wire bonding might work for the capacitors as well. I would have to use some sort of die attach material but that isn't really a penalty because the IC will need it too. An advantage is that in case of a capacitor failure it would be easy to break the wire bond to remove it from the circuit. It would be much better for my circuit to fail open then to short as a short could possibly damage other expensive and hard to rework things.

One problem I have is I don't really know what failure mechanism I should primarily worry about - the capacitors themselves or the connection to the substrate and if I should be more concerned about the operating temperature, the temperature cycling, or thermal shock. Any thoughts on failure modes to worry about and if ceramic chip capacitors would fail open or closed?

Any comments on wire bonding vs soldering?

 

[0xDEADBEEF]

Wirebonding is not that great. The wire is very thin and breaks easily, so you need to start gluing, with solder you can just use the normal electronics design process. The thermal cycling is probably the most harmful of the damaging effects, but if your component values are good, normally the circuit will work for a long time.

 

[f95toli]

May I ask what it is you are doing? And at what temperature? I am guessing you are working below 4K since you need a good isolation vacuum (at higher temperatures you usually WANT some gas in the cryostat to improve cooling, e.g. a few mBar of He).

As I said above, unless you REALLY need a high vacuum there is no need to worry about outgasing. I use all sorts of strange components (microwave componets etc, filters etc) in my cryostats, the only time outgasing is a problem is when preparing the system since the pump down time goes up somewhat, although something like a capacitor would have negligable effect due to its small size. That said, I work at 4K or below and use sorbtion pumps etc so I don't have to worry about the vacuum going soft unless there is an actual leak into the cryostat(I use He as exchange gas to cool to 4K, the sorb then kicks in and creates a good vacuum and will take care of any outgasing etc).

Electrical components tend to break due to the thermal stress (not due to the cold as such), it is usally a good idea to cycle all components a few times (i.e. dunk them in liquid nitrogen) before actually installing them. If it doesn't break then it will usually be fine.

One thing you need to watch out for are crimped connectors, a connector that works well at RT can stop working at low temperatures because some parts have shrunk more than others; it can be a pain to troubleshoot (this is another reason why the LN2 test is useful).

Also, I personally always use SMD components whenever I can, they seem to be somewhat more mechanically stable; presumable because of their smaller size and the fact that there are no leads.

 

[davidrit]

Ok, I guess I was being paranoid about outgassing. I had worked on a project at 80 K a while back where outgassing was a problem but now that I think about it the problem was the device was sealed and the outgassing had to be kept down so it could hold below a certain level (I don't remember) of vacuum for years.

I'm just doing lab testing now so the pump can stay on or the device could be pumped down again if needed. So lower outgassing is better to lower the pump down time but not a big issue.

The circuit won't be near 4 K. The silicon will probably have problems below 50 K so somewhere in that neighborhood. The silicon is cooled to lower noise and will be mounted on a small substrate with the capacitors and possibly some other components.

Liquid nitrogen dunk sounds like a good idea, thanks.

Regarding the wirebonding, I thought that the loop height would be able to tolerate the thermal expansion/contraction by changing height and that would minimize stress on the bonds. Kind of like placing extra length in a power line at average temperature so it can contract in the cold without stressing the connections to the poles.

 

[f95toli]

So how are you cooling the sample? A cryocooler? 50K is a bit of a "strange" temperature since you can't reach it using a cryogenic liquid (pumping on LN2 won't get anywhere near 50K)

Bonding works well at low temperatures (I bond all my samples), but you will obviously need to glue whatever you want to bond first; bond wires are very thin and although they are surprisingly strong you can't use them for mechanical strength.

That said, if you can just solder the components. As long as you take the usual precautions while soldering (same as for RT) it will be fine. The only time there are any serious "mechanical" problems at low temperatures is if you need to be able to physically move or rotate something, and then mainly because normal lubricants etc won't work.

I realize that 50-80K might seem a bit "exotic" if you haven't worked much with cryogenics before, but there really isn't much to worry about when it comes to things breaking as long as you are careful (a bad joint will of course eventually break if you cycle it, but that joint would probably eventually break even at RT).

Also, I guess you know this but be careful when pumping on cold cryostats. A cryostat at 50K will cryopump air etc quite efficiently meaning if you were to use e.g. a small rotary vane pump (which might only reach say 10^-3 mBar) you might actually pump "backwards", i.e. actually sucking air FROM the pump. You should definitely use a turbopump for something like this.
Personally I wouldn't bother unless I had reason to believe that there was something wrong.

 

[davidrit]

I don't know anything about the cooler yet. I'm jut going to do the electronics portion, 99% of which will be outside the cold environment. Someone else will handle the cooling/vacuum/mechanical design and the project won't really start for a month or so.

I'm trying to get a head start on the cold/vacuum portion of the electronics since that's where my ignorance is. Hopefully I'll learn enough to handle the pump and cooler selection and some mechanical design on future projects.

Thanks for the help.