New thermal material provides 72% better cooling than conventional paste

zohaibahd

Posts: 841   +17
Staff
Why it matters: Data centers are hot, both figuratively and literally. As we feed more and more data and processing demands into these server farms, keeping them from overheating is becoming an increasingly expensive and energy-intensive challenge. But researchers at the University of Texas may have a cool solution – a new thermal interface material that can whisk heat away from processors better than the likes of Thermalright and Thermal Grizzly.

Thanks to a mechanochemically engineered combination of the liquid metal alloy Galinstan and ceramic aluminum nitride, this thermal interface material, or TIM, outperformed the best commercial liquid metal cooling products by a staggering 56-72% in lab tests. It allowed dissipation of up to 2,760 watts of heat from just a 16 square centimeter area.

The material pulls this off by bridging the gap between the theoretical heat transfer limits of these materials and what's achieved in real products. Through mechanochemistry, the liquid metal and ceramic ingredients are mixed in an extremely controlled way, creating gradient interfaces that heat can flow across much more easily.

"Cooling accounts for about 40% of data center energy usage, or 8 terawatt-hours annually."

Beyond just being better at cooling, the researchers claim that the higher performance reduces the energy needed to run cooling pumps and fans by up to 65%. It also unlocks the ability to cram more heat-generating processors into the same space without overheating issues.

It's an important breakthrough in a world where data centers burn through around 8 terawatt-hours per year just on cooling, which is about 40% of their total energy usage. The UT Austin researchers estimate that their new TIM could reduce cooling energy needs by 13% across the industry, cutting overall data center energy usage by at least 5%. That should translate to a gigantic reduction in operating costs and carbon emissions.

As for how you can get your hands on the material: it's yet to make it out of the labs. The UT team has so far only tested it successfully at small scales but is now working on producing larger batches to put through real-world trials with data center partners.

Of course, that means even if it lives up to the hype, you'll probably see this liquid metal thermal grease keeping huge server farms frosty before it ever hits the shelves for you to slather on your home PC's Intel or AMD processor.

"The power consumption of cooling infrastructure for energy-intensive data centers and other large electronic systems is skyrocketing," said Guihua Yu, professor in the Cockrell School of Engineering's Walker Department of Mechanical Engineering and Texas Materials Institute. "That trend isn't dissipating anytime soon, so it's critical to develop new ways, like the material we've created."

More details about this new magic material can be found in the paper published in Nature Nanotechnology.

Masthead credit: Cockrell School of Engineering

Permalink to story:

 
And what's the catch? Mass production? It eats metal faster than normal liquid metal?

If I were to guess, that special mixing process will cost more, but the question is how much more at scale. It could still be worth it.
(And then the process is another thing to be improved with future research for additional savings.)
 
How does a better thermal interface reduce the need for fans and pumps? It doesn't. You still need a mechanism to remove the heat that's been transferred.
The point is that, given a configuration where at some temperature difference the heat flow through the heatsink balances the dissipated electrical power, if you decrease the thermal resistance of the layer between the chip and the heatsink then the heat flow is greater for the same temperature gap. Hence the chip cools down until the T gap is reduced to a point where the (reduced) heat flow again balances the dissipated electrical power. TLDR, the chip is cooler and/or the heatsink is smaller.
 
If I were to guess, that special mixing process will cost more, but the question is how much more at scale. It could still be worth it.
(And then the process is another thing to be improved with future research for additional savings.)
I would pay $50 for an ounces if the performance is indeed that good. It would totally worth it to reduce the noise coming from my 64 core PC.
 
That 72% sounds like a great number, but in practice, it may just translate to 1 or 2 degree Celsius improvement in temperature. Cost wise, it may double. So to be honest, I don't see the point. You want a good thermal pad/ paste, go with the Honeywell PTM7950.
 
It takes heat away from the processors better, so what? You'll still be pumping the same amount of heat into the atmosphere.

At the end of the day, you'll likely be making more heat, since this will allow you to run CPUs at higher over clocks.
 
Percentages are the ultimate trickster. 72% of zero is what? How about using real numbers, and requiring the same from those you are reporting on. We need much less 'fake news' and more 'real news'.
 
Percentages are the ultimate trickster. 72% of zero is what? How about using real numbers, and requiring the same from those you are reporting on. We need much less 'fake news' and more 'real news'.

TFA gave a figure in watts, 2760,
It allowed dissipation of up to 2,760 watts of heat from just a 16 square centimeter area.

therefore I am assuming (and hoping not to make an *** out of me and, well, me) that this means that 'typical' liquid metal can dissipate around 1600watts in 16cm^2, which x1.72 is the ~2760watts figure given.
 
TFA gave a figure in watts, 2760,


therefore I am assuming (and hoping not to make an *** out of me and, well, me) that this means that 'typical' liquid metal can dissipate around 1600watts in 16cm^2, which x1.72 is the ~2760watts figure given.
Listen, I think that the term "dissipate" is deceptive. See, because if if could "dissipate" that amount of heat, you wouldn't need a cooler on top of it, now would you? I would choose, "transmit",. While that may be an "overthink" on my part, does any thermal paste really dissipate heat?

So, throw away your coolers boyz, girlz, and all genders in between, the cure is here, this cooler paste has you covered. pun intended (See what I mean)?
 
Bit of a strange one when they say "conventional paste" - The liquid metal paste is already at 50% above conventional thermal paste, so this might not be that revolutionary
 
And what's the catch? Mass production? It eats metal faster than normal liquid metal?

that catch is this is a scam.

Energy cannot be created or destroyed .. so it doesn't matter if the goo ''dissipates'' better.. that just means the processor can now run FASTER and consume MORE energy and produce MORE heat ... it is just better at putting the heat into the surrounding environment. not saying it is bad product .. but claiming to reduce thermal output is a bold azz lie.
 
that catch is this is a scam.

Energy cannot be created or destroyed .. so it doesn't matter if the goo ''dissipates'' better.. that just means the processor can now run FASTER and consume MORE energy and produce MORE heat ... it is just better at putting the heat into the surrounding environment. not saying it is bad product .. but claiming to reduce thermal output is a bold azz lie.
It depends. For servers that run with passive heatsinks and large coolers that blow through the entire server blade, this could help reduce temps a lot. The difficulty of cooling something increases exponentially because of the physical limits.

The problem is that most servers don't use conductive TIM like liquid cooling because of how risky it is.

Realistically, they did say that it could reduce power usage of data centers by 5%.
 
Better cooling may not translate to power saving because modern processors will throttle performance when it runs too hot. The general behaviour is that the CPU will either maintain a more consistent clock speed or it may see it as an opportunity to boost clock speed.
 
Better cooling may not translate to power saving because modern processors will throttle performance when it runs too hot. The general behaviour is that the CPU will either maintain a more consistent clock speed or it may see it as an opportunity to boost clock speed.
On servers you have much tighter standards. The CPU won't boost past its specs unless it is set to do so (but I doubt anybody does this since stability is king).
 
Kidz, Kidz, Kidz, this whole thing is a scam, from day one.

Thermal paste doesn't "DISSIPATE" any heat whatsoever. How about if we put down our game controllers and put on our thinking caps?

Any thermal paste can only >> remove << heat from a CPU, and >> transfer << it to a cooler_full stop. The only variable involved in doing that is the speed at which it can occur.

So, if a CPU is generating a 100 watts of heat, all the thermal paste can accomplish is presenting it to the cooler sooner. The cooler then "dissipates", (how about if we call it "redistributes"), the exact same 100 watts of heat into the atmosphere. And again_full stop.

So then how do I explain air conditioners, they cool things, right? NO, they don't. They "redistribute" heat from indoors to outdoors. No heat is lost, just moved.

And as a matter of fact, air conditioners create more heat than they redistribute Why, because the electric motors used to drive the compressors are not 100% efficient, So, however many BTU are created by the compressor drive motor are "redistributed" to the atmosphere.

How can we fix this? How about if we smear this snake oil on the motor;s drive shaft. That'll help right? .WRONG.

And for any of you old timers present, "the duck didn't come down" on this nonsense.
 
Kidz, Kidz, Kidz, this whole thing is a scam, from day one.

Thermal paste doesn't "DISSIPATE" any heat whatsoever. How about if we put down our game controllers and put on our thinking caps?

Any thermal paste can only >> remove << heat from a CPU, and >> transfer << it to a cooler_full stop. The only variable involved in doing that is the speed at which it can occur.

So, if a CPU is generating a 100 watts of heat, all the thermal paste can accomplish is presenting it to the cooler sooner. The cooler then "dissipates", (how about if we call it "redistributes"), the exact same 100 watts of heat into the atmosphere. And again_full stop.

So then how do I explain air conditioners, they cool things, right? NO, they don't. They "redistribute" heat from indoors to outdoors. No heat is lost, just moved.

And as a matter of fact, air conditioners create more heat than they redistribute Why, because the electric motors used to drive the compressors are not 100% efficient, So, however many BTU are created by the compressor drive motor are "redistributed" to the atmosphere.

How can we fix this? How about if we smear this snake oil on the motor;s drive shaft. That'll help right? .WRONG.

And for any of you old timers present, "the duck didn't come down" on this nonsense.
Rhetoric.
 
Rhetoric.
Did you know you could get a college degree in "rhetoric"?

From Google search|: "Rhetoric degree programs are offered at multiple degree levels and programs often combine the study of rhetoric with writing or communications"

Beyond that, "so what", and can you prove me wrong?.

Maybe you could try this,."wul, if they said it, it must be true" Nice try.
 
Back