Overclock.net banner

Is it safe to use Coollaboratory Liquid Pro between the CPU and copper plate?

1870 19
I have a Coolermaster water cooling and the Intel 11th Gen.
Is it safe to use the liquid pro between the CPU and the copper plate of the water cooling?

My other option is Gelid GC-Extreme thermal paste.

Thanks.
1 - 20 of 20 Posts

· Newb to Overclock.net
Joined
·
5,172 Posts
After a few weeks, you may notice alloying to the copper plate. Just sand it smooth, but leave a layer of the alloy and it should be good forever.
 

· Registered
Joined
·
891 Posts
Discussion Starter · #5 ·
After a few weeks, you may notice alloying to the copper plate. Just sand it smooth, but leave a layer of the alloy and it should be good forever.
Does the liquidPro thermal paste remove the fine etching on the CPU plate? i.e., the model, and other details?
 

· Newb to Overclock.net
Joined
·
5,172 Posts
Does the liquidPro thermal paste remove the fine etching on the CPU plate? i.e., the model, and other details?
CLU and conductonaut has done that on my CPUs. I'm sure CLP will too.
 

· Iconoclast
Joined
·
34,107 Posts
Bare copper will alloy with gallium and can cause liquid metal TIMs to 'dry up' over time.

If I want to use these TIMs on bare copper, I try to make sure the alloying process is already well underway before final mounting, by applying a layer of the liquid metal to the clean copper surface, heating it, letting it sit a while, wiping away excess, then applying another layer. Depending on how rough the surface was I might do this four of five times over the course of a week, with a light burnishing before applying the the final coat and mounting it to the part.

Does the liquidPro thermal paste remove the fine etching on the CPU plate? i.e., the model, and other details?
Nickel plated copper doesn't have the alloying issue, but most liquid metal TIMs will still stain a nickel plated heatspreader. You should still be able to read the etching on the part after wiping it off, but you'll probably be able to tell you used liquid metal on it.
 

· Registered
Joined
·
891 Posts
Discussion Starter · #8 ·
Bare copper will alloy with gallium and can cause liquid metal TIMs to 'dry up' over time.

If I want to use these TIMs on bare copper, I try to make sure the alloying process is already well underway before final mounting, by applying a layer of the liquid metal to the clean copper surface, heating it, letting it sit a while, wiping away excess, then applying another layer. Depending on how rough the surface was I might do this four of five times over the course of a week, with a light burnishing before apply thing the final coat and mounting it to the part.



Nickel plated copper doesn't have the alloying issue, but most liquid metal TIMs will still stain a nickel plated heatspreader. You should still be able to read the etching on the part after wiping it off, but you'll probably be able to tell you used liquid metal on it.
On my old custom water cooling, I used CLP and it's there for 6 years now...

This week, I won a gaming pc with intel 11th Gen Core i9 and CoolerMaster water cooling AIO. The plate has nano channels. I was surprised. I'm used to mirror finish.

I think I'll just use the Gelid thermal paste and avoid the risk.
 

· Iconoclast
Joined
·
34,107 Posts
On my old custom water cooling, I used CLP and it's there for 6 years now...

This week, I won a gaming pc with intel 11th Gen Core i9 and CoolerMaster water cooling AIO. The plate has nano channels. I was surprised. I'm used to mirror finish.
A mirror finish is much better for liquid metal on bare copper as there is very little gallium that can be absorbed. Currently the only bare copper part I'm running with liquid metal on it is my PowerColor 6800 XT and while flat, the base is not very smooth. The relatively high porosity is why I 'treated' it with liquid metal repeatedly before mounting it. I don't want to have to take the card apart in a year...which is usually what's happened in the past when I used liquid metal on less than pristine copper surfaces.

I think I'll just use the Gelid thermal paste and avoid the risk.
Likely for the best, since if anything goes wrong with the AIO, support may take issue with liquid metal being used on it.

I'm planning do to that myself, currently testing long term behavior on some old HS to see how bad staining will be. Is two weeks on air at 30C enough to give the impression?
Probably won't reach full thickness, but you should see it stain the part and start to crystalize or pit in some areas.
 
  • Rep+
Reactions: wuttman

· Iconoclast
Joined
·
34,107 Posts
How much did you drop?
Only a few C (2-3C off hotspot temps, 1-2C off edge), but since I'm running the card at 2.55GHz and only cooling it with ducted case fans (I removed the stocks fans and fan shroud), every degree helps.

Btw I'm also testing how silver coating would resist to LM. Found some source stating zero reaction between silver and gallium.
Interesting. I've read of gallium-silver alloys, but I'm not a chemist or metallurgist and don't know the details or what to expect with lower temperature reactivity between the two. Let me know how the experiment turns out.
 

· Facepalm
Joined
·
11,650 Posts
A mirror finish is much better for liquid metal on bare copper as there is very little gallium that can be absorbed. Currently the only bare copper part I'm running with liquid metal on it is my PowerColor 6800 XT and while flat, the base is not very smooth. The relatively high porosity is why I 'treated' it with liquid metal repeatedly before mounting it. I don't want to have to take the card apart in a year...which is usually what's happened in the past when I used liquid metal on less than pristine copper surfaces.



Likely for the best, since if anything goes wrong with the AIO, support may take issue with liquid metal being used on it.



Probably won't reach full thickness, but you should see it stain the part and start to crystalize or pit in some areas.
Your advice about repeated heating/removing is really good here.
But your advice about mirror finish, that is not my findings.
The entire reason I bought sanding blocks awhile back was for sanding my laptop heatsinks. And I experimented with this many times.
I found that mirror finish actually causes some serious problems with LM. It's great (perfect actually) for normal paste.

The first problem is, a mirror finish (ex: 7000 grit polished) surface literally -prevents- LM from wetting. I tested this on several surfaces--IHS (delid 9900k), laptop heatsink (BGA throttlebook), etc.
It was almost impossible to get the LM to even stop sticking to itself and remaining in a ball.

Eventually I did get it able to separate and stick to the surface. But then I noticed other problems.
1) With the laptop heatsink, i saw very fast absorption of the gallium and after a few weeks, there was nothing but a hardened layer. I also noticed core temps on the 7820HK get worse in a matter of days with the fully polished heatsink.

With the 9900k, I noticed, in a matter of a week, that core temp deltas were getting worse. It had gone from 7C from hottest to coldest cores, to 11C. In just 1 week, after a lot of prime95 and other thermal testing.
I then delidded the IHS and noticed that there were spots on the die and IHS with no LM on it. Even though there was plenty (with both sides wetted) on application.

so then I did some more testing.
I took a second laptop heatsink and polished it to a mirror finish, then I applied LM. Then I kept wiping it to thin the layer. What I noticed was, with time, the LM seemed to "stop" sticking to parts of the heatsink and started making 'gaps' where as it thinned out, it started migrating to itself. I then remembered what some people on notebookreview forums said some years ago, that "LM hates polished surfaces. Roughen up the heatsink a bit first".

So then I did another test. I wiped the heatsink with 1500 grit and then reapplied the LM. (after making sure it was fully clean).
The first thing I noticed was that the LM wetted instantly. Literally instantly. It did not stick to itself at all. It went right to the heatsink.
The second thing I noticed was, as I kept wiping and thinning it out, it did NOT migrate to itself. It seemed to eventually get more vicious, almost like not as watery, but it remained on the surface perfectly.

So then I did this same thing to both the 9900k die and the underside of the IHS, as well as the top of the IHS (because rip warranty anyway right).
Applied on both, spent about 15 minutes wiping the first layer over and over and over around the underside of the IHS and the CPU, then applied a second layer after the first layer felt thicker and would no longer 'run' anywhere even if i shook it around, and put on my NH-D15.

Result: after 2 weeks, temp deltas did not change at all. They remained 6-7C.
I kept testing and checking and a month later after multiple prime95 small FFT FMA3 tests: 6-7C.

My last test before I retired the 9900k, when I upgraded to 10900k: 6-7C.

When I did this on my laptop 7820HK, my cpu temps have remained stable for a year, unlike when I was doing polishing. 2C temp delta on prime95 small FFT no AVX, 1 year later. I'm happy with that result.

So it means that a rough surface is more durable and also more resistant to absorption IF you take time to wipe it around 15 minutes (or do the 100C oven bake method then remove then reapply, for objects that are safe to bake), then apply a second layer on top.
 
  • Rep+
Reactions: deepor and wuttman

· Facepalm
Joined
·
11,650 Posts
  • Rep+
Reactions: wuttman

· Iconoclast
Joined
·
34,107 Posts
The first problem is, a mirror finish (ex: 7000 grit polished) surface literally -prevents- LM from wetting. I tested this on several surfaces--IHS (delid 9900k), laptop heatsink (BGA throttlebook), etc.
It was almost impossible to get the LM to even stop sticking to itself and remaining in a ball.
While wetting is certainly better on rougher surfaces, I've usually been able to get it to wet to polished copper surfaces, at least if they were clean. I also tend to use Phobya's stuff, which has always seemed to wet well (even on polished silicon dies) for me.

With the laptop heatsink, i saw very fast absorption of the gallium and after a few weeks, there was nothing but a hardened layer.
A polished surface definitely shouldn't be absorbing faster than a rough one, though I suppose if you can't get as large a volume to adhere between the surfaces, that could make up for the difference.

Once it starts to alloy/stain a little, it sticks much better. The smoother surface should reduce the potential depth of the alloy layer, requiring less excess to be stable over time, which mostly matches my experience.

So it means that a rough surface is more durable and also more resistant to absorption IF you take time to wipe it around 15 minutes (or do the 100C oven bake method then remove then reapply, for objects that are safe to bake), then apply a second layer on top.
I'd definitely do more than 15 minutes and two layers. On my 6800 XT's cooler I did five or six layers, with an hour at 70C then slowly cooling down to ambient before wiping off/rubbing in the excess and applying another layer 12-24 hours later.

What do you wipe it with?
I use sterile medical pads to wipe away the excess between applications and foam tipped PCB swabs for application and spreading. The foam applicators absorb liquid metal pretty well and have a bit of 'stick' to even very smooth surfaces, which I find aids considerably in spreading liquid metal TIMs.

 

· Facepalm
Joined
·
11,650 Posts
While wetting is certainly better on rougher surfaces, I've usually been able to get it to wet to polished copper surfaces, at least if they were clean. I also tend to use Phobya's stuff, which has always seemed to wet well (even on polished silicon dies) for me.



A polished surface definitely shouldn't be absorbing faster than a rough one, though I suppose if you can't get as large a volume to adhere between the surfaces, that could make up for the difference.

Once it starts to alloy/stain a little, it sticks much better. The smoother surface should reduce the potential depth of the alloy layer, requiring less excess to be stable over time, which mostly matches my experience.



I'd definitely do more than 15 minutes and two layers. On my 6800 XT's cooler I did five or six layers, with an hour at 70C then slowly cooling down to ambient before wiping off/rubbing in the excess and applying another layer 12-24 hours later.



I use sterile medical pads to wipe away the excess between applications and foam tipped PCB swabs for application and spreading. The foam applicators absorb liquid metal pretty well and have a bit of 'stick' to even very smooth surfaces, which I find aids considerably in spreading liquid metal TIMs.

I think the main issue with the polished surface was the streaking that happened.
I didn't discover this until I actually started spreading the LM extremely thin manually, and then I saw it leaving bare copper in places, and I watched it and the LM was literally 'migrating' to itself.
Which explains the core to core temp delta degradation and hot spots on my 9900k.

When I roughened it up with 1500 grit and did the 10 min wipe method, that no longer happened.
I understand my word can't be taken as gospel here, but this is just my experience with what happened, and it of course cost me money for this stuff.
 
  • Rep+
Reactions: Blameless

· Iconoclast
Joined
·
34,107 Posts
The streaking is surface tension of the LM favoring itself over the unwetted copper.

If you can't get it to stick to a freshly polished surface it might be worth trying an acid to more subtly roughen it than sandpaper and then applying the LM before any oxidation can occur. Vinegar, cleaned off thoroughly with an anhydrous isopropyl rinse, should do the trick.

As for cost, that's one of the reasons I use Phobya (though price seems to have gone up on it) and buy my 99.95% isopropyl by the gallon.
 
1 - 20 of 20 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top