Electronic components suitable for use in vacuum?

[Johnny85]

One for the professional electronic engineers I reckon...

I need to make a preamplifier for a scanning electron microscope (SEM).  The circuit will take very low (nanoamp or even picoamp) AC input and amplify it (using either LF357 or LT1115 op amp IC) with a closed-loop gain about 10^5.  The circuit inputs can be bypassed with a 2PDT relay and the output is selected with a 4PDT relay.  You can get the gist of what I mean with this "beautiful" drawing I did in MS Paint:



This circuit will be installed in the SEM itself, and therefore it will be in an environment of very low pressure, about 10^(-6) Torr.

So my question is, does anybody have experience of building electronic circuits in vacuum?  What sort of materials are suitable?  I am going with standard epoxy FR4 stripboard, but how about the resistors and capacitors insulating materials?

I am particularly concerned about the relays.  They should be very low noise (particularly the input 2PDT), but what about the outgassing?  I am trying without much success to get information out of various manufacturers' technical departments.

Really I just would like to hear from anybody with experience in making vacuum-compatible circuits.

PS I hope it doesn't annoy anyone that this is not a post about a pedal project - I was going to put it in the Member's Only section but I saw a post there from Mark Hammer saying "Why do we have this section, it is unneccessary" with lots of people agreeing, so I put it here instead, hope that's OK!

[Jarno]

I don't have experience in designing stuff for vacuum, but the company I work for has a relay division, and we have an engineers forum. The forum is not visited that often I guess, posted a question on there but got zero replies. But I'll try, I'll let you know if someone replies.

[Johnny85]

Thanks Jarno, that is kind of you.

[R.G.]

I always wanted to design electronics, but never got the chance, so this speculation and not informed advice.

You're correct in looking for low-outgassing materials. Carbon film, metal oxide, and wirewound resistors are generally inert conductors on ceramic substrates, so they have the potential to be vacuum-inert, at least for short periods, if suitably preconditioned by having any adsorbed water vapor/gases removed by slow heating (I speculate). As for caps, I'd say to stay away from electros entirely, and probably oil-containing types. You should be OK with film types. Likewise, precondition.

On the relays, you may run into dissimilar metal voltages and noise. Simply putting two dissimilar metals together creates both a minor thermal battery (thermocouples are just dissimilar metals welded) and a noise source. This is a big issue in equipment design, particularly in the area of automated semiconductor testers where the test points are connected/isolated by relays. As for vacuum stability, you're in luck. Reed relays, being entirely encapsulated in a gas-tight glass and metal envelope which should be heremetically sealed, so they should work fine in a vacuum.

You'd probably find all you need in the old military specification documents for electronic components. I searched that stuff out a few years ago, and it's on line, or was. I dimly remember finding it at the Rome Air Defense Center site. These docs are no longer the current requirements for "Mil-spec", but still have the basic info in them. Maybe a little googling would still turn them up.

[wavley]

I actually have experience with this because my day job is building Cryogenic Low Noise Amplifiers.  Everything I build is put in a vacuum and thing cooled to about 14 Kelvin.

Now, when you add cryogenics and frequencies in the GHz it's kind of a different animal, but what I can tell you is what we use for parts and maybe it will help. 

Big parts for us are pretty much just SMD stuff with gold terminals to enable wire bonding and are generally of a ceramic persuasion, but we do use some film type capacitors here and there.  We solder the magnet wire for the DC bias stuff and the waveguide probes but everything else is attached via H20E conductive silver epoxy cured on a hot plate. We also use capacitors and resistors with a quartz substrate, the caps are just parallel plate and the resistors are NiCr sputtered, though this is probably moot for you because it's high frequency signal path stuff.  For our bias cards in the dewars we use run of the mill Mil Spec resistors and diodes and have been running for years with no problems, and these things go through vacuum and thermal stress two and three times a day, we also use big wire wound resistors as heaters and very rarely have problems with those.  All of our amps are channelized gold plated brass bodies with tight fitting lids and MDM connectors, but there is a hole in the side for the power for the illumination LED's (unpackaged HEMT's need illumination for quicker recovery from saturation)  so our electronics are hardly hermetically sealed.  One weekend when I was doing a cryogenic test out of curiosity I cooled a 12AU7 in the dewar with my amp in a very good vacuum with no ill effects.

Long story short, we use all sorts of Mil Spec stuff in a vacuum all the time and very rarely have problems and that is mostly due to thermal stress.

Oh yeah, don't use lead free solder, it has a tendency to grow whiskers in a vacuum, we aren't allowed to use it on anything we give to NASA and we don't use it for ourselves because it's inefficient to build things that will fail. 

[~arph]

In advance, sorry for my n00b suggestion.

Would potting the entire circuit in a strong epoxy help keep the vacuum out?
I could see some trouble with the wires goiing in/out.

[wavley]

[aziltz]

I work in vacuum a bit an I know that solder is something that can pollute the vacuum. That's all the advice I have unfortunately.

[MikeH]

I was hoping this thread would be about playing rock and roll in outer space.

[R.G.]

I was hoping this thread would be about playing rock and roll in outer space.

I'm guessing that tickets to that concert would be a tad expensive. 

[MikeH]

I was hoping this thread would be about playing rock and roll in outer space.

I'm guessing that tickets to that concert would be a tad expensive. 

And overpriced since, well you know- "In space no one can hear you scream"

[wavley]

I was hoping this thread would be about playing rock and roll in outer space.

I'm guessing that tickets to that concert would be a tad expensive. 

And overpriced since, well you know- "In space no one can hear you scream"

This is starting to ring of the "Stars are giant transistors" thread

[Johnny85]

I was hoping this thread would be about playing rock and roll in outer space.

I'm guessing that tickets to that concert would be a tad expensive. 

And overpriced since, well you know- "In space no one can hear you scream"

Keep it down you guys, what if Bono hears you? Judging by the last U2 tour he'll be on the phone to NASA in no time...  

Seriously though, thanks a lot to RG and Wavley.  Since 10^(-6) Torr is not such a high vacuum, pre-baking is probably not needed in this application, but RG's speculation is definitely right for UHV applications.  

Thanks for the detailed info Wavley, v. useful stuff, particularly the info on solder, you have saved me a potential headache with that one.  What sort of relays do you use, if any?  And does all that internal metallisation (I hope you understand what I'm trying to express here) in your pictures prevent whatever is inside them being exposed to vacuum?

[wavley]

I was hoping this thread would be about playing rock and roll in outer space.

I'm guessing that tickets to that concert would be a tad expensive. 

And overpriced since, well you know- "In space no one can hear you scream"

Keep it down you guys, what if Bono hears you? Judging by the last U2 tour he'll be on the phone to NASA in no time... 

Seriously though, thanks a lot to RG and Wavley.  Since 10^(-6) Torr is not such a high vacuum, pre-baking is probably not needed in this application, but RG's speculation is definitely right for UHV applications. 

Thanks for the detailed info Wavley, v. useful stuff, particularly the info on solder, you have saved me a potential headache with that one.  What sort of relays do you use, if any?  And does all that internal metallisation (I hope you understand what I'm trying to express here) in your pictures prevent whatever is inside them being exposed to vacuum?

Are you speaking of the rectangular tubing?  That is actually waveguide, for the frequencies we work it has way lower loss than coax cable http://en.wikipedia.org/wiki/Waveguide This particular amp is an EVLA K-Band amp 18 to 26.5 GHz

The flat gold plates on the bottom are part of the refrigeration system, there is a cryo pump and cold finger that compress scientific grade helium against that plate taking it down to near absolute zero, 14 Kelvin is -434.2 F

In the first picture on the right and side there are rectangular bare pc boards, what you're seeing is the solder side but it is stuffed with diodes and resistors that we never have a problem with, of course they aren't directly coupled to the cold plate, but it is still pretty cold, if we lost vacuum during refrigeration the air would instantly freeze into a solid block of ice that takes about two weeks to melt.

[wavley]

Oh yeah, the thing with the light and the waveguides going to it is the amp.  30 Db of gain from 18 to 26.5 GHz with a noise temp of less than 8 kelvin across the band, currently the lowest noise amplifiers in that frequency range in the world, and they have been for the last 10 or so years.

I don't know much about vacuums to compare to yours, but we have a Hastings Vacuum Gauge and it read 3 microns of mercury.

[aziltz]

The flat gold plates on the bottom are part of the refrigeration system, there is a cryo pump and cold finger that compress scientific grade helium against that plate taking it down to near absolute zero, 14 Kelvin is -434.2 F

oh man that is cold!  don't touch that!

while we're on the subject of sharing our vacuum apparatuses (apparatii?)  here's mine.  it goes to nK using magnetic and laser trapping and cooling.  The water lines are there to keep the coils from over heating.

[G. Hoffman]

Just curious; wouldn't it be easier to put the electronics outside the vacuum, and route the wires to whatever sensors need to be inside?  It might take a couple of expensive connectors, as they would have to be air tight, but then you can build with components rated for normal conditions.


Gabriel

[PRR]

In VACUUM TUBE fabrication, absorbed gas is a major problem. It is trapped inside forever, and bends and binds the electrons.

Copper, Nickel, Tungsten are pre-cleaned in acid then bottled and held at high vacuum and inductively heated for hours on the pump. Carbon has thermal advantages and does not cling gas, but is typically thick so it still takes hours to get most of the gas out. When pressure finally falls as low as it is going to, the stem is sealed, then a capsule of Barium and other sucky metal is burst onto the glass as a long-term gas absorber.

There was a German company built a 5-tube radio in one bottle. (There was a per-tube radio tax.) To minimize the pin-out, it was expedient to put several resistors and caps in the same bottle. They found that they had to seal most resistors in glass because otherwise they contaminated the vacuum.

Standard epoxy sounds like a high-vacuum disaster. Proprietary epoxies are full of additives and extenders. Epoxy sets-up fast but full cure may take forever. Cured epoxy is not real permeable, but not utterly impermeable either.

NASA has special problems. High-orbit vacuum is even better than a vacuum tube, and it is infinite. ALL your absorbed gasses eventually go away. On Earth we think steel on brass makes a good bearing. In space, it seizes. Turns out that what we observe on earth is not naked steel on brass, there is a monatomic oxide film. Deep space sucks that off. The naked metals weld to each other.

There's whole books about vacuum technology and practice. Mostly older, pre-IC.

So.... do you mind a little gas? If the SEM can stand a few stray ions, maybe. If the pump runs all the time, they wander out.

Also talk to your vacuum-pump guy. Whatever drek comes out of your stuff will gunk-up his pump. Especially the diffusion pump, which is really a small "still", loaded with either mercury or a special oil. Either way, the heat and recirculation may cook-up contaminants in strange new ways.

I agree with G. If you can find pass-throughs, and can stand the lead-length, put the works outside.

[amptramp]

Many years ago when I was working on spacecraft, we had a prohibition on zinc, pure tin and cadmium as plating materials because they have a high vapour pressure and the vapour tended to redeposit on optical sensors like earth and star sensors.  I would suggest you either get active devices in ceramic or metal packages if they are available or take the epoxy devices and put them into hybrid packages.  It would be a good idea to pull a vacuum on these hybrid packages as they were being sealed.  There are established reliability glass-body metal film resistors but they were $5 apiece in the 70's. I think ceramic or teflon capacitors would be OK but I would test film caps for outgassing before using them - the basic materials may have a decent vapour pressure but the additives like plasticizers may outgas severely.  Reed relays as suggested by R. G. sound good but there were also some all-metal hermetically sealed relays by Teledyne that were available in miniature packages similar to the contemporary logic devices (which also had metal cases and leads like the original silicon op amps) and were about the same size.

You may be shopping the surplus stores for some of this stuff as a lot of it may no longer be available.

It may pay to look at the NASA Preferred Parts List.  Everything there should be usable.  This is one guide:

http://nepp.nasa.gov/npsl/index.htm

which is contained in here:

http://library.gsfc.nasa.gov/Engineering/Eng.htm

but you get the idea - the information is available.  Also, since maintaining low temperatures is an essential part of your design, be sure devices are rated for more than the usual power.  A 1 watt resistor will not heat up as much as a 1/2 watt resistor even though the power may be much less, just because of the extra heat conductivity of the larger package.  Keep in mind, you have no convection and the only conduction is through the leads.

[Johnny85]

I don't know much about vacuums to compare to yours, but we have a Hastings Vacuum Gauge and it read 3 microns of mercury.

Thanks for all the info again, Wavley.  Two weeks to melt, that's insane, has it ever happened?  Well I don't think I can build a better amp than you but we have a much better vacuum, 3 micron Hg is only about 3x10^(-3) Torr, we go 3 orders of magnitude lower - our ion gauges start working at 10^(-3) Torr!   But even that is not ultra high vacuum, it's kind of "mid-high" vacuum, hence my earlier comment to RG re: pre-baking.

Just curious; wouldn't it be easier to put the electronics outside the vacuum, and route the wires to whatever sensors need to be inside?  It might take a couple of expensive connectors, as they would have to be air tight, but then you can build with components rated for normal conditions.

Thanks Gabriel, but the reason we are putting it inside the SEM chamber is that this chamber is made of several inch thick stainless steel and is therefore a near-perfect Faraday cage - we need excellent EM shielding and minimal cable length since the input signals are so small.

Cheers for lots of interesting points there, PRR.   I'm going with epoxy board because a group developing a particle detector for use in UHV (the CHICSi detector) tested several board materials and found epoxy FR4 to be the lowest outgassing material available, and they were going to a vacuum several orders of magnitude better than we will.  Our system can stand stray ions, as it is only 10^(-6) Torr as displayed on an ion gauge.

High-orbit vacuum is even better than a vacuum tube, and it is infinite.

According to wikipedia (which may be dodgy), the pressure at 100km above sea level is still 10^(-3) Torr, which is a much lower quality vacuum than ours.  The interstellar vacuum, even, is not infinite, really - there are always some bits of dust or molecules about, and even in their absence (i.e. in the spaces between them), quantum mechanics predicts a finite minimum energy density (the "vacuum energy").

It may pay to look at the NASA Preferred Parts List.  Everything there should be usable. 

Thanks for that link, amptramp, I will be checking that one out, could be v. useful, cheers.

Thanks to everyone who has responded so far! Really, thanks so much, it's great that so many people offer their help