Liquid metal stains on cpu and gpu die

So, the other day my notebook started freezing and buzzing so I had to take it apart only to discover that it was the hdd. Before taking apart my temps were just great 73c max on cpu and 66c max for gpu with lm, cleaning up the old lm with 90% alcohol and applying new lm raised the temps by 7c! Worse than thermal paste, so I tried to add more but still no luck and at the end I removed the lm again and applied thermal paste. The thing that I noticed that the surface of the die of the gpu and cpu is scratched up and has stains on it which doesn't come off with the alcohol, any ideas if it can be removed? Also is this much staining normal in 2 weeks?

 

Stains, normal
Scratches, not so much

Galinstan is a mix of 3 metals, some of the galium is lost from the mix as it seeps into and alloys with the copper heatsink leaving behind a semi solid (Al/Sn) residue. Initial liquid metal applications really should be checked within 3 months or, if you're dangerous like I am, use more of it than the thin painted layers you see on all the howto videos, bead it up a bit (with your barriers in place of course, foam, nail polish, tape, etc)

The stain is an extremely thin layer that you can sand off if you want, but it's just cosmetic so there's really no point, especially if you are reapplying liquid metal

Was it really dry looking when you opened it up? The solid flecks of residue might have been dragged across the gpu die as the heatsink was removed and may also be messing up the fit afterwards leading to the bad temps

 

I applied it twice and bad results each time, 80c cpu and 72c gpu, previously gpu only hit 66c and cpu 73. I'm aware of the copper staining but not the cpu and gpu die, it just doesnt come off.

 

This mean bad fit between die and heatsink. How to remove Liquid metal... http://forum.notebookreview.com/thr...and-how-to-use-it.809332/page-8#post-10737180

 

But bad fit how, it worked the 5 times I opened it before this, and now bad fit 3 times in a row?

 

The staining from LM is only cosmetic so this isn't your problem (The staining from LM is all too thin to make an difference on temp). If you go for Liquid metal once more you can polish the heatsink with 3000 grit sandpaper or use the the mentioned ways in my link. Only to get an flat and even finish (not sand everything off). The stain on the copper can should be there. Next round of Liquid metal will suck less into the copper. Aka it will last longer even with a not perfect heatsink.

Do the same if you go with conwentional thermal paste. You can also sand very careful with 3000 grit sandpaper on die. But not much.

Edit...
You should measure the fit between heatsink and die (pressure testing)... You can use thin paste or as I in this link... http://forum.notebookreview.com/threads/clevo-overclockers-lounge.788975/page-1618#post-10767615

The thermal pads can also be a culprit. Check the imprint on pads from the components. Too much and the heatsink will be lifted off the die. Same as above... Pressure testing. Post pict if you check it.

 

Go to walmart and buy a bottle of Brasso and rub off the old liquid metal that turned into the residue on your components and then use rubbing alcohol to wipe the surfaces down. Re-apply liquid metal and try again.

I went through this process with my 8700K and its IHS a couple weeks ago where a re-mount with new liquid metal raised the temps because it wasn't clean. Works great now.

 

Oh cool, thanks! Can I use it on the die too?

 

I used a very tiny amount and rubbed it on the die itself also since there was a little bit stuck on it. You should be okay as long as you clean up all the Brasso / metal polish with rubbing alcohol afterwards.

Also note it did not remove 100% of the "staining" but it did take out all the residue that made the surface rough/uneven causing the higher temps. You should have a smooth finish afterwards.