How can I purify mineral oil?

My early experiments with oil-cooling my PC lead to some chemical *discoveries*
The first was mineral oil will dissolve latex rubber.  I was using surgical tubing to pump the oil to the radiator and after a day the tubes had softened and swollen so much they wouldn't stay attached to the pump.  That and the oil turned an ugly yellow.
(* the second was warm oil reacts with PVC, and causes the insulation on usb cables to become brittle)

I'm working on the next gen of the project and I'd like to clean up the oil, but I can't quite google my way to a solution.
I have filtered the oil (removed all the bugs) though a coffee filter, and a couple of paint filters, I don't have any reason to believe the rubber is suspended and filterable, but I could be wrong.

Any chemists out there know what I can do?

sort by: active | newest | oldest
Malkaris (author) 10 months ago


So I found this article today... Unfortunately I lost most of my dirty oil to a leak, I may have to try this out if the problem comes up again.

rickharris1 year ago

Yes Oil + rubber is a bad mix.


Malkaris (author)  rickharris1 year ago

Hmmm, any suggestions how one might distill 5+ gallons of mineral oil?

You have two problems at hand:
1. Something is dissolved in the oil
2. Something is floating around in the oil

Since you don't need the oil to lube an expensive engine we could ignore the dissolved bits.
For the second bit a good idea is time.
Anything floating in the oil will seperate over time and you can pump off the oil that is clean.
Do a test with a long test tube.

But I do have to wonder why you used oil in the first place!?
Destilled water won't cause any shorts and has a good heat transfer rate.
Oil can do the same but you will need special equippment rated for oil use.
Anything that has a rubber seal need to be replaced with nitrile seals and some other things you already discovered.
PersonallyI would not bother with the oil, both in terms of cleaning it and using it.
Get new oil or use destilled water for the cooling ;)

Malkaris (author)  Downunder35m1 year ago

this is for submersion cooling. I've seen video tests with deionized water but the computer shorts after a min from something probably dissolving from the machine.

The oil gas sat for months and is still yellow so I'm inclined to believe it's dissolved contamination.

Malkaris (author)  rickharris1 year ago

It is rather tough to find a melting point for latex rubber, one poor reference to 212F (100C). I've got a feeling I need to "boil oil" for a long time to get this gunk out...

Maybe bone black, or activated carbon? I have read that bone black is used for decolorizing petrolatum, and also for sugar; i.e. turning brown sugar into white sugar.

I'm not sure what the actual procedure is for this is though. I am guessing, (1) mix the oil and solid absorbent together. Then, (2) wait for however long it takes for the impurities to sorb onto the sorbent. Then, (3) filter out the solid sorbent, using your coffee filter or whatever.

Increased temperature might make step (2) happen faster, but that's just a guess.

Anyway, this might be something fun to try for a small sample of your oil, for to see if it's color, and odor, can be improved just by soaking some activated carbon in it.

if it's in addition, and the viscosity is ok, I'd live with it.