Few more results. After four heating cycles the liquid metal does not appear quite solid, in the near future I'm going to attempt to clean it off to see if any has attacked the surface. Just trying to persuade the aluminium to finish corroding in order to play with it in the absence of fumes and risk of fire.
Aluminium/gallium + water is similar cold, but much less impressive than when properly hot. Still, the only way of removing it does appear to be waiting for it to burn out, when you're left with a pile of grey powder.
Cheers for the warning Mike, though solder failing is the last of my worries were I to want to reuse this heatsink. It's an intel stock one with the plastic framework removed, extruded aluminium fins and what is either a push fit or soldered copper core. I suspect push fit. If this falls out, it'll give me a piece of copper to play with. Not the end of the world. That said, I have no idea what temperature a heatsink sat on a hob will reach. Oven might be a more sensible route.
Yeah, I'm enjoying myself. I really do think this liquid has sensible applications in computers.
Removing was interesting. I think its fair to say that the layer over the copper is insoluble in IPA. However, it does come off. I think its a case of the IPA reacting with the alloy to form something which doesn't adhere particularly well. Certainly the cotton buds are coming away black. I've not managed to remove all of it, though can see copper glinting through in a couple of places. Either unreactive or much less reactive with citrus degreaser, hard to tell if it's reacting or if I'm picking up the remains of the IPA reaction. Leaning towards reacts, but slowly. Suggests my ideas of passivating a heatsink with this are flawed. Anyone have a spare copper heatsink to ship my way? After this one I'm going to be experimenting with a lapped true I fear.
Conclusions as of a day or so with this stuff:
Surface must be clean before application.
Reacts well to heat, this can make spreading a thin layer much easier
Violent with aluminium and water, destructive with dry aluminium
The liquid pro is reactive. It is hard to tell whether the part that has alloyed with the copper is still reactive.
Reacts fast with IPA, slowly with citrus degreaser. Insoluble, but forms a material that can be removed by cotton buds.
Testing still to do.
Copper in absence of aluminium, water and IPA. I want to find a solvent that attacks thermal paste but not this stuff.
Results of prolonged heat treatment in coating a piece of copper. Intention is prolonged time in an oven, heat to encourage reaction with the copper, and reaction with the air to see if the latter occurs.
I'd like to turn a blowtorch on this, but don't think I'd learn much from the process.
Probably need to find a way of removing dust that occurs as a consequence of lapping.
I still think it has use in protecting bare copper. However, it cannot be washed with water or IPA. It probably reacts with the atmosphere, much like copper. However the surface oxidised protective layer is very likely to be soluble in fresh liquid metal, in which case the oxide will disperse throughout the join and have minimal effect on heat transfer. After long storage times it is probably worth relapping the surface with very fine paper. I should check the numbers relating to heat transfer, thickness of this layer may matter very little.
/offtopic
Current plan, as a reminder to myself but by all means criticise if you see flaws. Mill waterblock. Lap waterblock. Clean with IPA, then place in oven to drive out remaining solvent. Apply liquid metal, rub in thoroughly. Return to oven.
Lap cpu. Clean with IPA, leave in warm place. Not sure I'm brave enough to put an i7 into an oven, but at least it'll be cleaner than the block to start with. Apply liquid metal, leave in warm place. Consider under the grill.
Remove surplus from waterblock. Block should now be inarguably attached to liquid pro, cpu less so. Some surplus on cpu.
Mount processor/waterblock, light pressure. Run processor at 70 degrees for several hours to encourage the liquid metal to join the two together, and allow air to escape. Increase mounting pressure. Test as normal to see what change in temperatures has resulted.