preferred thermal paste

You are completely full of yourself, aren't you? As much as you would love to believe that you are super smart and knowledgeable about this, you are incorrect. A cursory investigation into this says that you are making this all up as you go along.

This is what Coollaboratory Liquid Ultra is made from:

Four of those eight metals have any possibility to oxidize with any compounds in the atmosphere when in pure form. Those are gallium, zinc, bismuth, and copper. Their oxidation is only on the exposed surface and forms a hard layer protecting the underlying metal.

I'm no chemist, but I can read up on those elements and many, even the ones that can have slight surface oxidation as a pure element, are used in alloys to make them highly resistant to oxidation.

On top of that, even if this alloy was to be at any risk of oxidation from contact with the air, how on earth is the area between the heatsink and die supposed to oxidize if it is not exposed to the air?

 

I wrote a long reply that was too dry and technical, so I'm going to make it optional reading for people who don't want to read a dry wall of text.

Let me just cut to the chase and ask you this: Liquid Ultra does eventually degrade no? So could you explain to me how that would be possible if what you said is all true?

Spoiler

Silver slowly converts to silver sulfide in air and stannous is tin, and actually the proper noun form should be stannum. Of the metals listed, only rhodium could be considered inert under the conditions likely to be encountered in Liquid Ultra's lifetime, and I guarantee it is a very minor component considering it is THE most expensive metal in the entire periodic table, its value exceeding that of gold and platinum. All other metals will either oxidize, or form halides and sulfides from trace elements in air. While some metals could be considered unreactive at ambient temperature, when heated to 70 or 80C their propensity to react becomes much greater.

As to your comment about a protective oxide layer, yes that is certainly truly for the element in its pure, bulk state. But first of all we're not dealing with a sold mass here, we're dealing with what is effectively a liquid suspension, and the effective surface area is many times larger, not to mention a liquid suspension is mobile instead of static. (on a microscopic scale anyway due to Brownian motion) And then there's the fact that we're not dealing with individual elements in their isolated forms, but one single amalgamated alloy, which is likely to have entirely different properties compared to its individual constituents. To give a specific example, look up how sodium amalgam (alloy of sodium and mercury) behaves compared to the individual elements.

The fact that the heatsink-die interface is not directly exposed to air is precisely why it takes so long for degradation to happen -- over the course of a year or two instead of weeks. Laptops aren't air tight, and the heatsink-die interface isn't a hermetic seal. In fact, the very purpose of a TIM is to fill in the voids between the heatsink and the processor die, so of course it's going to be exposed to air. You don't need bulk exposure for chemistry to happen, trace amounts + high temperature is more than enough to start the reaction on its own. Slow rate =/= doesn't happen.

Also, if you bothered reading through the entire datasheet that you quoted yourself, you'll see in section 13 it says:

CooLaboratory themselves recommend cleaning up the bulk paste with HCl in the event that it needs to be disposed of.

 

Lol, whenever people bring up Coollaboratory Liquid Pro, I can't help but giggle at the fact people are basically putting a eutectic alloy akin to Dental Amalgam on their heatsinks.

One of these days, I'll repaste a machine with some of my Dental Amalgam and report the results for you guys. It's actually a pretty good TIM come to think of it since it's designed to be highly wetting to create a seal with the tooth, plus additionally, it's highly packable so you can apply on just about any surface, and finally, it expands slightly on setting to complete the interface. Not so sure about stability though

 

So, take your laptop to your next dental cleaning?

 

You know it, nothing like blingy grilles to keep your laptop healthy and happy. Like this --->

 

Ooops, wrong thread.

 

The current spot price of rhodium is below that of gold and platinum, both of which are other metals widely used in industry.

I can't find any shred of evidence that Liquid Pro or Liquid Ultra have any noticeable oxidation in air. None whatsoever. There is evidence that Liquid Pro very readily reacts with pure aluminum, and that Liquid Ultra may in some long-term cases react with aluminum and copper, but I have been completely unable to find anything about it being anything other than completely stable in air. In addition, 70, 80, or 90 Celsius is far too low to illicit any reaction with air. 200, 300, 400, 500 Celsius would have to be reached for those other metals that I stated are stable in elemental form in air. If have any pictures or anything valuable to show where Liquid Ultra had a reaction with the atmosphere, please show it.

Intel has used an alloy with low-melting point metals to seal bridge the gad between the IHS and the die on desktop and server processors. It is closer to the metal pads that Coollaboratory sells than the Liquid Ultra, but the composition is not too dissimilar, being made with metals that you say oxidize in the air. There has been no issue there, even though it is exposed to air (the IHS is glued down in the corners, but not sealed from outside air). The IHS that Intel uses is nickel-plated copper, however, and not copper or aluminum coming into contact with the TIM.

Also, I have been looking at alloys of the various metals, and while I really can't speak to a holistic view of Liquid Ultra, each metal that could oxidize in air also has at least one other metal in the Liquid Ultra mix that when alloyed makes it highly resistant to oxidation.

What does cleaning with HCl have to do with anything? You aren't going to clean up Liquid Ultra with soap and water. An acid is recommended for cleaning up most of those metals in their pure form.

 

You know if you wanted to have a discussion I'd appreciate you didn't take such an aggressive tone and accuse me of talking out of my butt and making up BS. And after all that you have still yet to answer my question as to how and why Liquid Ultra eventually degrades, given the Liquid Ultra is supposed to be highly resistant to degradation as you claim.

Spoiler

Just because you can't find anything doesn't mean it doesn't happen. It could be as simple as the average Joe not caring what actually happens to it other than trying to clean off as much of it as possible when they want to repaste. Hardware review sites sure as hell aren't going to be doing a degradation test and then analyzing the composition. The only place where you might get that data is from the manufacturer, but generally that's not released to public. That said, since you're hellbent on seeing pictures, knock yourself out with these photos I posted earlier in this thread.

And again slow reaction =/= no reaction. Once the activation barrier is overcome a reaction happens. Of course you're not likely to see Liquid Ultra oxidize instantly or even over a short timespan (otherwise it'd be quite worthless...), but over a period of 1-2 years significant amounts of oxidation can happen, especially when you factor in heat. That was my point. I mean, if you stared at your fingernails for 2 hours and didn't see them grow any longer, does that mean they're not growing? Not the best analogy but it gets the point across.

I don't know the exact composition of Intel's fluxless solder so I can't comment on it. Also, I hope you realize that fluxless solder between the IHS and die becomes solid once it's been cast, and requires baking in an oven to melt again (that's how some extreme overclockers delid a soldered IHS). So not a proper comparison at all to Liquid Ultra that remains liquid indefinitely as long as temp is >8 °C.

I'm no expert when it comes to solder, but I know they do and can oxidize, and the rate and extent to which that happens depends on the solder composition. Additionally, one of the primary purposes of solder flux is to prevent oxidation. So fluxless solder is... interesting to say the least. However in this case because the fluxless solder is a solid, it may actually become protected by the passivation effect that you describe. Once all exposed surface is oxidized, no further oxidation can take place because as a solid, the fluxless solder is immobile. Unless you were to cut the fully passivated solder in half, then you'd create a fresh surface where oxidation could occur again.

However, Liquid Ultra is obviously a liquid, and is in constant motion with its surface is being constantly molded. Sure you may not observe anything macroscopically, but on a microscopic scale there is constant Brownian motion as the particles rotate, move around and collide with each other.

And for all we know the fluxless solder could very well be oxidizing over the years, but because of the passivation effect mentioned above, the amount of oxidation simply is not enough to cause an appreciable change in termperature, and thus goes unnoticed. But this doesn't change the fact that it still happened, just that it was inconsequential to the end user so they simply don't pick up on it.

Then maybe you shouldn't? While it's true alloys can have some variation in their composition, mixing metals together in a random fashion without regard to proportion does not an alloy make. Liquid Ultra contains 8 metals, and short of pairing and encapsulating the appropriate metals into discrete particles, you're going to be looking at an "alloy" of 8 metals, which is going to behave very differently both physically and chemically than simple bi- and tri-metallic alloys. On top of that, 3 of the metals in that composition -- tin, indium, and zinc, are metals with fairly high reduction potentials. Were they to "alloy" any of the other 4 metals, they'd essentially behave as a sacrificial anode, and their oxidation would be accelerated. So yes the other 4 metals would be protected while those 3 are hung out to dry. And once those 3 metals are gone, the other metals will oxidize on their own, as well as start cannibalizing whichever metal in the remaining 4 has the highest reduction potential (most likely gallium).

It's to show that cleaning up Liquid Ultra with HCl is not some BS I made up, and is in fact recommended by the manufacturer themselves.

 

But you are talking about things as if you know the answers, when really you don't. I'm not pretending to know the answers, but just want to cut through the BS. You are claiming that the performance of Liquid Ultra degrades over time because of oxidation between it and compounds in the atmosphere, but this is just you hypothesizing without any real evidence in support of this.

Based on the scant evidence I can find, I believe that Liquid Ultra does in fact react when in contact with certain metals for a long time.

One person delidded his desktop processor and applied Liquid Ultra between the CPU and heatspreader and between the heatspreader and copper heatsink. Several months later, he noticed that his temperatures were higher than they had been before. He removed the heatsink and found the liquid ultra between the heatsink and heatspreader to be a white crust that was stuck to the copper heatsink but not the nickel plated heatspreader. Between the heatspreader and the exposed die, however, the Liquid Ultra looked brand new.

The link you posted earlier in this thread showed someone with a laptop that had Liquid Ultra on it for two years. Both the aluminum and copper portions of the heatsink were pitted in their centers, suggestion a reaction between the Liquid Ultra and both the copper and aluminum.

It really seems to me that Liquid Ultra is far more stable with aluminum that it used to be, but over time it will have a reaction with aluminum and possibly copper as well. I say possibly copper, because we don't know if it was pure copper or an alloy in the heatsinks that had issues, because there are lots of people that don't have any issues wither. It also looks like processor dies do not react at all with liquid ultra, and that nickel has very little reaction with liquid ultra. It would be interesting to have a long-term test with liquid ultra in many different applications.

 

Yes, because the die-IHS interface is much less exposed to air compared to the IHS-heatsink interface, hence the rate of degradation is much slower, consistent with what I've been constantly saying for the past couple of posts.

The GPU and CPU contact surfaces appear to be copper but perhaps with a certain admixture of aluminum, while the northbridge plate possibly does contain a significant amount aluminum. For the CPU and GPU contact surfaces, the area where the die makes contact is clearly much darker than the peripherals, and the CPU die makes an almost perfect impression in the liquid ultra residue. This is consistent with what I said about heat increasing rate of degradation, as the non-contact areas retain much of their lustre with only a few dark spots, while the area in contact with the die has degraded to all hell. Gallium eats right through aluminum, so I find it unlikely the northbridge plate is made of pure aluminum, otherwise after 2 years it would basically be in pieces.

Well I'm sorry but you're basically doing the same thing you're accusing me of doing here -- you look at the pictures, then draw your own conclusions. If I'm hypothesizing, then so are you.

The problem with what you're stating is that you haven't accounted for other alternatives before making the conclusion that what you're seeing is a reaction between Liquid Ultra and the metal, instead of the reaction with elements in the atmosphere. This video, the desktop example, and those laptop pictures are all very telling. Zinc(II), gallium(III), and tin(IV) oxides are all white, and indium(III) oxide is light yellow/off-white. These 4 metals have the highest reduction potentials, and will be the first to get oxidized. This appears to be consistent with the white crust described by that desktop owner, as well as the white/grayish crusty stuff you see in that video. In addition I can throw in my own experience and say after 3 months of usage, white crust does indeed form along the edges where the die contacts the heatsink, in addition to forming patchy spots between the die-heatsink interface in my Clevo laptop.

Both copper(II) oxide is black and silver(I) sulfide are black. So if I had to guess, the reason why we're seeing black patches on the heatsink in the Macbook that went for 2 years is because after such a period of time, even the less reactive metals in Liquid Ultra have degraded.

If it was a metal-on-metal reaction, I'd expect the contact surfaces to become corroded. My own experience, as well as that of Mr. Fox's (another Liquid Ultra user) both point to this NOT being the case. Other than Liquid Ultra leaving a silver finish on the heatsink, the surface itself feels as smooth as it was before using Liquid Ultra, with no evidence of corrosion. I don't know whether that desktop user reported anything about the heatsink so I don't want to assume, but my gut feeling is his heatsink was fine apart from the white crust and silver stain. Most users don't report heatsinks being damaged when using Liquid Ultra, and you even say so yourself, and that to me is even more evidence that this is not a metal-on-metal reaction we're seeing. (unless the heatsink is made of aluminum, then in that case gallium will most certainly eat right through it) As for that Macbook, one hypothesis is that after 2 years, the metal oxides/sulfides/halides/whatever have basically sintered themselves to the heatsink. The other is that the contact plates contain a certain amount of aluminum, so in addition to oxidation there is a fair amount of pitting corrosion going on here as well, but only because of the inherent reactivity of gallium towards aluminum.

I could point you to papers that study the oxidation kinetics of the metals that make up Liquid Ultra, but that still isn't the "real evidence" that you're looking for. Really, short of someone scraping off that white crusty stuff and sending it for elemental analysis, there is not going to be "real evidence". You're basically holding me to an impossible standard here. I mean, if I gave you a chemical explanation of why nitrites are undesirable as preservatives due to the way they react with certain compounds, are you going to demand I perform lab tests to back up my claims? Making reasonable deductions based on known chemical reactivity is hardly "making stuff up".

So working with what we do have, I favor atmospheric oxidation (or more generally, reaction with elements in the atmosphere) over metal-on-metal reaction as the primary degradation mechanism. (unless it's an aluminum heatsink) You're certainly entitled to your own conclusions, but I just want to set the record straight that I draw my conclusions based on what I know about the reactivity of those metals, and the balance of evidence available. I most certainly do not make thoughtless empty BS claims.

You know at the end of the day, I could turn the tables on you and say there's no real evidence that Liquid Ultra causes pitting corrosion either. But that's not going to get us anywhere is it?

 

I prefer Shin Etsu X-23-7783D, it works fine for me.