It doesn't look like anyone here tried this so far. How about coating the contact surface of your heatsink with pure silver? After all, we're all using silver paste for contact. I know it was done in a commercial heatsink (the Silver Mountain, a couple of years back), which was a decent heatsink at the time, but the idea seems to have tanked comercially. Too expensive I guess.
I decided to try. And it doesn't cost a dime!The method that I'll describe works for copper only, not for aluminum (aluminum chemistry is a whole different beast). You can either prepare the silver solution yourself if you've got the chemicals, or you can get it ready to use at any photo processing lab. The stuff I'm talking about is used photographic fixing solution. That contains a lot of silver as silver thiosulfate, which is ideal for deposition on copper. So just find a friend who works at a processing lab and get a pint of the stuff - and I repeat, *used* fixer, not new one (the fresh liquid doesn't have any silver). Or if you want to make it, just combine some 10% silver nitrate solution with an equal volume of 20% sodium thiosulfate (or fresh photo fixing solution) and a few drops of ammonium hydroxide. Wear rubber gloves, and be aware that it'll stain your clother irreversibly if it gets on them.
Now onto the silver deposition. I used an older Thermaltake Volcano 6 Cu, which is aluminum with a copper insert at the base. Not a bad HSF for its time, but way too weak these days. Anyway, I thought it'd make a good test case since the CPU will be running on the hot side, and the difference will be easier to notice with/without silver.
I first lapped the copper insert with 600g paper, then polished it a bit more with a wet rag. Not to a shiny mirror (I know all about the mirror/no mirror dispute). Cleaned the surface first with methylene chloride to remove all trace of grease, then with pure isopropyl alcohol. Cleanliness is very important, or else the silver layer will have streaks - not good! I applied slowly 4 ml of the silvering solution onto the copper plate (I forgot to say I first diluted the solution 3 times with pure water) and watched the silver layer appear. You want the solution diluted enough so that the silver layer grows rather slowly, over 3-4 minutes, not at once. I removed the spent solution, wiped the disk with a clean cloth, and applied diluted solution once more for another 5 minutes. At the end I washed everything thoroughly with distiled water, and gave the silvering a slight polish by rubbing several minutes against a sheet of ordinary printer paper laid flat on glass. Again, I didn't give it a mirror shine.
Here's an "after" image:
And a closeup of the silvered copper disc:
Now why would copper+silver layer+silver paste work better than copper+silver paste alone? Well the theory is that the silver paste penetrates the small crevices in the copper and maximizes thermal transfer surface. The silver layer instead grows directly on the copper base, and contacts it everywhere, even in the smallest cracks where silver paste doesn't penetrate. And the pure silver deposited is much softer than even copper, so when it's pressed against the CPU core it will easily mold onto it. And of course a thin layer of silver paste will cover every imperfection at the surface of the silver coat. It may work, or not. I have it ready to go, and took a set of readings last night, with the heatsink lapped but not silvered. Tonight when I get home I'll try the heatsink with the silver layer, then post the numbers. I expect to see maybe a 1-2 degrees improvement. But hey, every degree counts, right?
I decided to try. And it doesn't cost a dime!The method that I'll describe works for copper only, not for aluminum (aluminum chemistry is a whole different beast). You can either prepare the silver solution yourself if you've got the chemicals, or you can get it ready to use at any photo processing lab. The stuff I'm talking about is used photographic fixing solution. That contains a lot of silver as silver thiosulfate, which is ideal for deposition on copper. So just find a friend who works at a processing lab and get a pint of the stuff - and I repeat, *used* fixer, not new one (the fresh liquid doesn't have any silver). Or if you want to make it, just combine some 10% silver nitrate solution with an equal volume of 20% sodium thiosulfate (or fresh photo fixing solution) and a few drops of ammonium hydroxide. Wear rubber gloves, and be aware that it'll stain your clother irreversibly if it gets on them.
Now onto the silver deposition. I used an older Thermaltake Volcano 6 Cu, which is aluminum with a copper insert at the base. Not a bad HSF for its time, but way too weak these days. Anyway, I thought it'd make a good test case since the CPU will be running on the hot side, and the difference will be easier to notice with/without silver.
I first lapped the copper insert with 600g paper, then polished it a bit more with a wet rag. Not to a shiny mirror (I know all about the mirror/no mirror dispute). Cleaned the surface first with methylene chloride to remove all trace of grease, then with pure isopropyl alcohol. Cleanliness is very important, or else the silver layer will have streaks - not good! I applied slowly 4 ml of the silvering solution onto the copper plate (I forgot to say I first diluted the solution 3 times with pure water) and watched the silver layer appear. You want the solution diluted enough so that the silver layer grows rather slowly, over 3-4 minutes, not at once. I removed the spent solution, wiped the disk with a clean cloth, and applied diluted solution once more for another 5 minutes. At the end I washed everything thoroughly with distiled water, and gave the silvering a slight polish by rubbing several minutes against a sheet of ordinary printer paper laid flat on glass. Again, I didn't give it a mirror shine.
Here's an "after" image:
And a closeup of the silvered copper disc:
Now why would copper+silver layer+silver paste work better than copper+silver paste alone? Well the theory is that the silver paste penetrates the small crevices in the copper and maximizes thermal transfer surface. The silver layer instead grows directly on the copper base, and contacts it everywhere, even in the smallest cracks where silver paste doesn't penetrate. And the pure silver deposited is much softer than even copper, so when it's pressed against the CPU core it will easily mold onto it. And of course a thin layer of silver paste will cover every imperfection at the surface of the silver coat. It may work, or not. I have it ready to go, and took a set of readings last night, with the heatsink lapped but not silvered. Tonight when I get home I'll try the heatsink with the silver layer, then post the numbers. I expect to see maybe a 1-2 degrees improvement. But hey, every degree counts, right?