Silica gel for stabilizing humidity

[SvanGool]

A commonly heard statement here on the forum is: humidity is affecting every not hermetically closed device and has influence on the measurements. These are long term effects and are often seasonal. The effects have an influence from sub-PPM to PPM level and take days to become apparent.
The problem is that, besides of some LS8 packages, it becomes more and more difficult to obtain TO-XX or ceramic (which are not moisture sensitive) package variants, instead of epoxy or plastic (which are moisture sensitive), of accuracy determining components.
It is being advised, once in a while, to drop a sachet of silica-gel into an enclosure, but without real explanation

I stumbled on a report from the museum world, a lot of museum artifacts, shown in display cases, are also very sensitive to moisture. This 2001 report of an experiment in the museum world is a very educative writing on the usage of silica gel to stabilize humidity: https://cool.conservation-us.org/waac/wn/wn23/wn23-2/wn23-206.html
For all the "lazy" readers: if an enclosure has minimum air leakage (less than one air change per day), the RH fluctuation can be held within a few percent, by dropping app. 30 grams per dm³ of silica gel into the box.

This was tested for "big" museum cases, but I see no reason this not to be applicable for smaller electronic enclosures.
This would indeed prove the effectiveness of moisture regulation by silica gel sachets, what do you think?

[JS]

The thing with gel is it saturates, so it only can cope with so much high humidity for so long. Now, you are aiming for humidity stability, not low humidity, so the things changes. The problem I see with that approach is if the enviroment humidity changes for too long (like moving from missisipi to california?) you still can't warrantee performance and you don't even know how long you need to wait befor it's stable to be recalibrated.

Now, with this approach the short term stability could be improved by muffling to almost eliminating weakly or monthly humidity variations, but I wouldn't trust it for 15%RH to 95%RH after a few month.

The enclousure should be pretty air tight, which will help, and can make the times constants be even longer. Could be interesting to set up a datalogger to check al this, internal and external humidity sensors, 4 samples a day could be enough so it can run from a battery without opening it for very long. Then the box could travel around the world (or at least two different places) for a year as a black box and check how it went after all that time...

JS

[Conrad Hoffman]

If the seal is good and you can back fill with dry nitrogen or argon, molecular sieve may be a better choice.

[Sceadwian]

If you can't afford the instability of humidity you should probably be environmentally controlling the entire area you're working in. Dehumidifiers aren't expensive or costly to run. Silica packets sound like a nightmare to keep fresh and their effectiveness would be completly situation dependent.

[BradC]

If the seal is good and you can back fill with dry nitrogen or argon, molecular sieve may be a better choice.

That's a great way to exclude humidity. Museums do that sometimes, although the Argon fill is mostly used to preclude oxidisation. Most of the time they just want to stabilise humidity. Silica gel is great for that as the bond with the water is loose enough that it forms a nice equilibrium. It's used as a buffer against violent swings rather than an absolute remover.

[Berni]

Silica gel does start to give off moisture back into the air once its holding onto enough water and the ambient moisture drops. Its just that they are often used by first baking the silica to get it to pretty much 0% RH and then placing it inside so the silica really really wants to grab all the moisture it can.

In something that is air tight sealed i could certainly see this being helpful in stabilizing it. Something that is air tight is not necessarily moisture tight. Most plastic materials can diffuse moisture trough them, it happens very slowly but when there is very little air volume inside it takes microscopic amounts of water to make the air moist. So a silica gel could help by greatly increasing the amount of water that the interior can hold on to so a lot more water needs to diffuse in before the air humidity significantly changes.

If you ware to drop silica gel into something with ventilation holes then you probably did not do a whole lot.

[SvanGool]

Silica gel is great for that as the bond with the water is loose enough that it forms a nice equilibrium. It's used as a buffer against violent swings rather than an absolute remover.

You got it, the silica-gel is indeed proposed here as a stabilizer, not as a moisture remover. One air change per day (named ACD in the article) is quite Leaky and the leakage in a normal electronics enclosure should be much lower.

As a stabilizer, the silica-gel reaches equilibrium and the quantity should be enough that there is no saturation under "normal" environmental circumstances at the same location, which would mean that it is not necessary to "refresh" the silica-gel, ever.
The paper describes the concept of "half-time" which is the time it takes for the RH inside a case to reach the halfway point of the ambient RH difference. The figures I stated in the first message are based on a half-time of 150 days. Increasing the half-time to a longer period (which would average seasonal effects even more, if required) would need a lower leakage than 1 ACD or more silica-gel. Both are not that difficult to realize.

If this solution still works under extreme permanent environmental humidity changes (e.g. a portable standard travelling from 95% to 15%), is dependent on how long the unit stays under certain circumstances and the half-time, it should indeed be tested in a measuring/logging project 

Addition: one thing is important in the process and that is, just before closing the box, the humidity should be near the point you want it to be on the long term (e.g. the average for your environment) and the silica-gel should be at that level of humidity by having it in that atmosphere for a couple of hours.

[Edwin G. Pettis]

There were older instruments that actually had 'pockets' in them for silica packets, not so much any more, it certainly wasn't because the silica wasn't doing what it was supposed to do, it may have been a younger generation of engineers knew nothing about the practice so they didn't incorporate the silica into their designs.  Silica is not only cost efficient but more over reusable, if it gets saturated, just pop it in a warm over for a few hours and they're good to go again.  A lot less trouble than trying to produce a sealed gas enclosure.

[Conrad Hoffman]

They put a packet of silica gel inside all those medium precision General Radio capacitors before they sealed them up with tar (or whatever the stuff is). No idea how long they were really good for, but most of the capacitors have stayed very stable over the decades. The high precision invar standards were filled with dry nitrogen. No idea if they used a desiccant pack in those.

IMO, nobody puts rechargeable desiccant in modern instruments because this is the age of the idiot customer and short product life. It would never be maintained. Or, not needed in most cases because everything is digital.

[sipo75]

I was logging temperature and humidity inside a Rigol DM3068 for a few days before replacing its fan. The chart showed that relative humidity decreases to 20-25% (ambient RH was about 20% higher) inside the case. Would it be valid to imply that any instrument with a sufficiently high internal temperature (above ambient) can be considered "dry"?

And on the point of humidity ingress in an air tight container. Second chart shows a plastic box (advertised as water proof) inside my refrigerator with a pack of silica inside it over a period of four months. RH stays at about 10%.

[SvanGool]

I was logging temperature and humidity inside a Rigol DM3068 for a few days before replacing its fan. The chart showed that relative humidity decreases to 20-25% (ambient RH was about 20% higher) inside the case.

Which you can explain by plugging your chart numbers in here: https://www.lenntech.com/calculators/humidity/relative-humidity.htm  Which results in a calculated RH of 24.8%, a close match to your measured number.

Would it be valid to imply that any instrument with a sufficiently high internal temperature (above ambient) can be considered "dry"?

No, I don't think so, if your environment RH was 95%, you would, under your circumstances, end up at 50% inside which is not "dry".

And on the point of humidity ingress in an air tight container. Second chart shows a plastic box (advertised as water proof) inside my refrigerator with a pack of silica inside it over a period of four months. RH stays at about 10%.

I think we here have a case of the silica-gel acting as a moisture remover AND stabilizer. There is some information missing:

• Longer term temperature and RH of the silica gel when it entered the box
• The temperature and RH of the open box when the silica gel was entered
• The amount of silica gel
• The volume of the box

Nice experiment ! 

[TUMEMBER]

For three basic chemical salts one can achieve repeatability and certainty of the values at an acceptable level. Three boxes are not so complicated with the simplicity of technical implementation.
I recommend "old" methods of implementing the "humidity standard".
Minimum outlays and good repeatability and reproducibility.
For "small" objects subject to measurement it is more convenient to use stable "chemical" humidity standards.

[intabits]

"Incredible solid state dehumidifier. No moving parts.":-



?!

[Henrik_V]

...
The high precision invar standards were filled with dry nitrogen. No idea if they used a desiccant pack in those.

No the 1404 are just filled with dry N  . We had a open GR104 at the Makerfair. I just opened two not GR (Quadtech?) 1404 cans (1nF). Due to wisker. Took 10min  with a torch. Pictures might follow.
The Invar plates don't look as nice as the GR ones. If I have the chance/time,  I will try to unmount the plates, polish and gold plate them. AND get as much tin out of the can as possible. The (copper?) can and the mounting stud has a tin plating ... planting ground for the tin wisker.

Tettex reference caps have a silica bag in the housing that that holds the N filled air can.