*P O L L* Liquid Metal damage ---> Integrated Heat Spreader, Copper & Nickel

[Nicholas Peyton]

wow that looks worse than mine; like a lot of "crud" or "bird shit" as you say lol.

on mine, you can still see spots of copper through the "crud" and its still smooth.

what kind of sand paper are we talking? I've got some old emery cloth laying around with 180R/R printed on the back.. it's smoothest out of everything I've got at home already .

I only want to do it enough to take off the bird shit, so minimal possible lapping

Hopefully also keep warranty intact on block.

 

[Mr.Scott]

Hopefully also keep warranty intact on block.

Nope. That warranty will be history. It probably already is just from the TIM.

 

[Vario]

Probably could try some copper polish first such as Brasso on a rag. If that doesn't work try this method:
http://www.overclockersclub.com/guides/lapping/

edit:

or this one
https://www.techpowerup.com/articles/cooling/39

Its not like sanding a piece of wood.

I'm not a lapping expert but I have done it a few times over the years
some tips:
1) use a flat surface like a sheet of glass or a counter that you check with a straight edge for flatness
2) tape the sandpaper down all sides
3) put a drop of soap or water on the paper
4) go one direction only and do 10 passes, rotate 90 degrees, 10 passes, and so on.
5) I don't sand back and forth, instead go one direction, like away from you, pick it up, return to center, push away, pick it up, return to center etc. Might have to clean the paper too so the abrasive surface doesn't clog with small metal particles.
6) go up a grit after a set number of passes, might be 10 passes each 90 degree direction, up to you I guess.
7) try to slide it across, don't want to press it hard into the paper or you won't get a flat result.

 

[Nicholas Peyton]

thanks, good article, but for just removing some old paste would be better to just go straight to the grit 2000? avoiding the harsher lower grades?

 

[Vario]

thanks, good article, but for just removing some old paste would be better to just go straight to the grit 2000? avoiding the harsher lower grades?

You could try it first yeah, might take the alloy off. Sorry for ninja edits, I added some tips above.

 

[Nicholas Peyton]

You could try it first yeah, might take the alloy off. Sorry for ninja edits, I added some tips above.

okay excellent thank you


Edit:

Interesting results to the poll by the way:
For copper:
27.3% said it either corroded or discoloured.

But 13.6% said LM caused no permanent damage at all to their copper.

The highest % was those who only used for delid or never used at all.
The 2nd highest % people actually said it caused no damage to their IHS at 27.3% (which is good to hear) although 22.7% said it did indeed either corrode or discolour their IHS

The amount of time it took me to scrub it off my IHS (1hr+ with non-abrasive cloth) would indicate that it's quite possible that for the 13% who said it discoloured their IHS it may be possible to correct that with a little more patience. I could be wrong though as 9% did say it actually corroded their IHS.

Nickel seemed to be the least affected; although that could be in part more due to participation levels in the poll and fewer people actually owning nickel blocks...

Out of those who did use it on Nickel. 3/4 said it caused some corrosion or discolouration only 1/4 said it caused no damage at all.

 

[jaggerwild]

Have you tired liquid nail polish remover to remove it?

 

[Nicholas Peyton]

Have you tired liquid nail polish remover to remove it?

No. But thanks for suggestion; might give that a try

 

[EntropyZ]

Have you tired liquid nail polish remover to remove it?

That has acetone in it doesn't it? Would liquid metal residue be affected by it? I know acetone is good for removing glue residue, but it does damage some plastics.

He should give it a try though. Just might be enough to get rid of the crap stuck to the copper.

Nail polish shouldn't be pure acetone though, which means it will be a mild cleaning solvent at best.

Liquid Metal paste is definitely not worth the gamble. Between IHS and heatsink it doesn't make much of of a difference between a high-end normal thermal paste, it's just not worth it.

If nothing else works, you'll just have to sand the surface off until it's clean. The copper might have corroded (very very slim chance, they couldn't really sell the thermal paste if it was harmful) or it's just that the non-smooth surface was completely filled up by the now dried up liquid metal.

Still this doesn't look as bad as a battery leakage.

 

[Nicholas Peyton]

I'll certainly be doing a good scrub + rinse with hot water afterwards (when I do try it).. overly harsh chemicals can damage copper. (although it is resistant to alcohol and naturally acidic chemicals found in food like ketchup, citrus (lemon juice) or vinegar).

It was awfully late and I never tried anything apart from plain hot water and hard scrubbing. So there's definitely still scope to try all these suggestions. Thanks again to everyone who suggested something. I'll let you know how I get on; judging from the results on this POLL I know the info could be valuable to many

 

[OneMoar]

lquidpro not a fault here
really shotty copper plating on the H100 is to blame too high of a aluminium content I would guess

 

[Arrakis9]

lquidpro not a fault here
really shotty copper plating on the H100 is to blame too high of a aluminium content I would guess

its just how copper reacts to it, all copper not just low grade stuff. i have a danger den block that i used with LM that did the exact same thing.

 

[OneMoar]

gallium and copper do not react gallium and aluminium do

 

[Random Murderer]

I tried the pads a while back. While it was a bit difficult to remove after it had set, it left no damage or discolouration on either of the IHSs tested and didn't damage or discolour the nickel-plated cold plate, however it did discolour the exposed copper heatpipes on one of the heatsinks tested. Surprisingly, didn't discolour the plate of my H100i which is also bare copper. Different alloys, I suppose? Still, only saw a couple of degrees (<4C) difference over decent pastes and ended up just sticking with pastes out of convenience.
What did discolour the copper plate of my H100i and the IHS it was mounted to was ICDiamond, plus it was a bitch to get an even spread and smelled like rotting fish when you removed it. Never again.

 

[jsfitz54]

When trying to clean the metal, try: Jewlers Rouge or Cymbal Polish.

http://www.bing.com/search?q=jewelers+rouge+polish&FORM=HDRSC1

http://www.bing.com/search?q=cymbal+polish&FORM=HDRSC1

 

[revin]

too high of a aluminium content

The copper plating itself should not have aluminum in it, maybe a zinc contamination or really really bad nickel substrate [if any used.]
I have not seen those base's but Gallium would tarnish/corrode a crappy contaminated copper finish faster than a "good" copper finish, or if they skimped, tried to plate directly to the aluminum [worse] a reaction would start there depending on the purity of the copper and the stability of the Gallium [ie. which way on the ph scale it is at a given state]
Temp's from using it for TIM can change the properties of it to acidic or alkaline if it a crappy batch.
Back many years ago we had to remove thermometers from key part's[pretty much all] of the production area's and zero was allowed in the Q.C area due to government contracts for satellites.
Our QC manager said Gallium in certain states can degrade most all metals but said tungsten was pretty much not effected.
I'd kinda like to break open a thermometer and stick a TIG tip in it to see what it would due.
This was our upgraded plating line from back in the 80's. nickel, yellow and clear zinc, phosphate, passivation, electro polish ect. to have certification for those contracts and all the ISO classification's

 

[phanbuey]

lap the block.

1000-1500 grit first then 2000.

 

[Nicholas Peyton]

hmm interesting.. because I'm thinking of using it on my GPU and nickel-plated GPU block.

I only get 6-8c difference between water temp and load temp on GPU (or 9-10c with Furmark) between water & load.

So I am desperate to try it on my GPU.

I'm getting 11c load GPU temps with my chiller.... trying to figure out a way to get that down to 5c load GPU temp. LM is what I'm *considering*.

Hoping it will give me better results on GPU than it did on CPU.

GPU's are all bare die these days aren't they? (it certainly doesn't have a big IHS if it is an IHS, it's small and perfectly shiny, looks a bit like a metallic glass).

 

[Arrakis9]

hmm interesting.. because I'm thinking of using it on my GPU and nickel-plated GPU block.

I only get 6-8c difference between water temp and load temp on GPU (or 9-10c with Furmark) between water & load.

So I am desperate to try it on my GPU.

I'm getting 11c load GPU temps with my chiller.... trying to figure out a way to get that down to 5c load GPU temp. LM is what I'm *considering*.

Hoping it will give me better results on GPU than it did on CPU.

GPU's are all bare die these days aren't they? (it certainly doesn't have a big IHS if it is an IHS, it's small and perfectly shiny, looks a bit like a metallic glass).

yes, your gpu would probably benefit the most from it. just be very very careful applying it. if i were you i would put some clear nail polish or liquid electrical tape around the tiny resistors surrounding the gpu first so you make absolutely sure you wont kill the entire thing.

 

[Random Murderer]

yes, your gpu would probably benefit the most from it. just be very very careful applying it. if i were you i would put some clear nail polish or liquid electrical tape around the tiny resistors surrounding the gpu first so you make absolutely sure you wont kill the entire thing.

I've always used blue-tack or kneaded eraser, but it could just be habit leftover from my days of subzero bench runs.
I agree with @Arrakis+9 though, you definitely need something to protect against accidental shorts if you're going to be applying liquid metal to your GPU die.

 

[Nicholas Peyton]

yes, your gpu would probably benefit the most from it. just be very very careful applying it. if i were you i would put some clear nail polish or liquid electrical tape around the tiny resistors surrounding the gpu first so you make absolutely sure you wont kill the entire thing.

I've always used blue-tack or kneaded eraser, but it could just be habit leftover from my days of subzero bench runs.
I agree with @Arrakis+9 though, you definitely need something to protect against accidental shorts if you're going to be applying liquid metal to your GPU die.

I always hear people saying "use liquid electrical tape".

Can I just use normal electrical insulating tape? Like this: (the stuff I've got is 2cm wide & even listed as flame retardant)



Because... I know it will peel off easily (without leaving residue) in case I ever need to make a warranty claim....?

Don't have any experience with liquid electrical tape.. is it easy to get all of it off?

After all these reports of some batches of EVGA cards blowing up (3-5% I think it was) I always worry my card could be one of them.

And because I run with a water chiller and only use the card very "lightly"; it could take a lot longer for something to go wrong. I'm probably fine; but I don't want to risk not being able to make that warranty claim in case something did happen.

 

[Arrakis9]

I always hear people saying "use liquid electrical tape".

Can I just use normal electrical insulating tape? Like this: (the stuff I've got is 2cm wide & even listed as flame retardant)

View attachment 82681

Because... I know it will peel off easily (without leaving residue) in case I ever need to make a warranty claim....?

Don't have any experience with liquid electrical tape.. is it easy to get all of it off?

After all these reports of some batches of EVGA cards blowing up (3-5% I think it was) I always worry my card could be one of them.

And because I run with a water chiller and only use the card very "lightly"; it could take a lot longer for something to go wrong. I'm probably fine; but I don't want to risk not being able to make that warranty claim in case something did happen.

liquid electrical tape should come off easily with any sort of solvent such as acetone, same thing for removing clear nail polish. don't use electrical tape because it can leave gaps which completely negates what your trying to do in the first place.

clear nail polish would be the easiest to come by and would cost you around $1 at the local store.

 

[phanbuey]

the point is to be airtight... so silicone, nailpolish, non-conductive grease, something airtight that will keep condensation out.
https://rog.asus.com/articles/overclocking/sub-zero-overclocking-motherboard-preparation/

my favorite one from that list has to be crazy putty

 

[Nicholas Peyton]

Just read a very interesting article about Liquid Metal on Wikipedia.

The listed metals that it is able to dissolve is very interesting, as it mentions Nickel, but doesn't mention Copper:

Copy/Pasted quote reads:
"Liquid metal can dissolve most metals; however, at moderate temperatures, only some are slightly soluble, such as sodium, potassium, gold, magnesium, lead, nickel and interestingly mercury.[5] Gallium is corrosive to all metals except tungsten and tantalum, which have a high resistance to corrosion, more so than niobium, titanium and molybdenum.[6]"

That is very interesting; because I think there may be a common misperception that Copper is more at risk than Nickel. In a post I read today (while discussing this with someone at a different forum) a very experienced member even said that he's never had an issue with Nickel but has always had an issue with Copper) because the LM seeps into the pores of copper more easily. Where as nickel is a harder, more shiny surface so the LM is not absorbed as much.

The results of this POLL would even agree with that too...

Don't know if anyone has any further thoughts; or if the discussion has already been exhausted. But on the back of topic; thought it was still worth sharing.

 

[OneMoar]

yes but it isn't pure gallium its a alloy usually of gallium indium and tin