Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Power Upgrades Hardware

Liquid Metal CPU Heatsink Beats Water Cooling 218

unassimilatible writes "Bios Magazine is reporting that the world's first commercially available liquid-metal based CPU cooler is about to ship. Danamics, a Danish company, claims that its LM-10 outperforms standard air-cooled heatsinks and most watercooled systems with a mere 1W power draw. 'The liquid metal is a key component in Danamics cooling systems. Liquid metal has two major advantages when cooling high power density heat sources: Firstly it has superior thermo physical properties that decrease temperature — and temperature non-uniformity — on die and across chips. Secondly, the electrical properties of the liquid metal enables efficient, reliable and ultra compact electromagnetic pumping without the use of moving parts, shafts, seals, etc.' Awesome technology, if it actually works and is affordable. The submitter requests that the moderators terminate all T-1000 jokes."
This discussion has been archived. No new comments can be posted.

Liquid Metal CPU Heatsink Beats Water Cooling

Comments Filter:
  • Toxicity? (Score:5, Interesting)

    by i_ate_god ( 899684 ) on Sunday July 20, 2008 @08:21PM (#24267891)

    And just how good for the environment is this liquid metal? Or for your health?

    • Re:Toxicity? (Score:5, Informative)

      by strelitsa ( 724743 ) * on Sunday July 20, 2008 @08:24PM (#24267935) Journal
      There was a nice discussion about this in Firehose before this made it to the Big Page. A couple of the candidates for the liquid metal that might be used in this thing are environmentally neutral (bismuth, tin, etc.).
      • Re:Toxicity? (Score:5, Informative)

        by Sangui5 ( 12317 ) on Sunday July 20, 2008 @08:29PM (#24267997)

        From my Firehose post:

        It's mostly likely using Field's metal (http://en.wikipedia.org/wiki/Field%27s_metal), Rose's metal (http://en.wikipedia.org/wiki/Rose_metal), Galinstan (http://en.wikipedia.org/wiki/Galinstan), or one of the other low-melting point low toxicity alloys, NOT mercury.

        • by MdotCpDeltaT ( 744490 ) on Sunday July 20, 2008 @08:41PM (#24268111)

          I think the liquid metal is their server.

        • Re:Toxicity? (Score:5, Informative)

          by jlarocco ( 851450 ) on Sunday July 20, 2008 @08:48PM (#24268179) Homepage

          For the lazy:

          Field's metal [wikipedia.org]
          Rose metal [wikipedia.org]
          Galinstan [wikipedia.org]

          • Of the three you listed, Galinstan is the only one that's liquid at room temperature, so I'd bet on it being that.
            • Re:Toxicity? (Score:5, Interesting)

              by Xtifr ( 1323 ) on Monday July 21, 2008 @02:29AM (#24270499) Homepage

              If it's being used to cool CPUs, I don't think it needs to be liquid at room temperature, since the area near most modern CPUs is considerably hotter than that! If your CPU is running at room temperature, you probably don't need much cooling (and if you do, you're going to need a much more elaborate system than merely one based on a liquid metal, since a passive heat-sink isn't going to take you anywhere below room temperature).

              If the liquid metal you're using for cooling tends to freeze at room temperature, that's going to present a few engineering challenges, but I don't think they're insurmountable, and, given the relative rarity of metals that are liquid at room temperature (and the unpleasant nature of some of them, e.g. mercury) it may well be worth it.

              Pure gallium [wikipedia.org] melts at body temperature and is generally considered non-toxic. If it weren't for the expense, it might be a decent option itself. That said, Galinstan does sound like a promising option, though one of its ingredients (indium) is even more expensive than gallium, which could be a problem. Tin's still relatively cheap, though. :)

              • Re: (Score:3, Insightful)

                by Alastor187 ( 593341 )

                If it's being used to cool CPUs, I don't think it needs to be liquid at room temperature, since the area near most modern CPUs is considerably hotter than that! If your CPU is running at room temperature, you probably don't need much cooling (and if you do, you're going to need a much more elaborate system than merely one based on a liquid metal, since a passive heat-sink isn't going to take you anywhere below room temperature).

                This may be true when the machine is running, but what about when it is initially started? As the processor begins to warm up the cooling medium will still be solid. Since solids are hard to pump the heat generated by the processor must be transported by conduction only. If the designer goes through the trouble of using a liquid cooler, it is probably safe to say that conduction only heat removal is not sufficient. So this presents some serious transient start-up issues.

        • Galistan (http://en.wikipedia.org/wiki/Galinstan [wikipedia.org]) seems the only one usable, the others have a melting point of around 70 degrees Celcius. Which won't work if they cool better than air or water (which usually keeps the temps below 50 degrees celcius).
        • Re:Toxicity? (Score:5, Informative)

          by dhovis ( 303725 ) * on Sunday July 20, 2008 @08:54PM (#24268243)
          I doubt they're using any of those. Rose's metal uses lead, and the other two contain large percentages of indium and gallium, both of which are getting pretty expensive. Much of the world's gallium goes into GaAs and GaN, whereas the indium goes into indium tin oxide (or ITO), which is a transparent conductor that goes into all LCD screens.
          • Do you know of any other suitable liquid metal they could be using, then, as I don't - mercury is just too toxic (compared to Galinstan, anyway) and NaK is a fire and explosion hazard, so even if gallium alloys are getting more expensive, they're still the less toxic, cheapest option.
        • Needing to be 62 C just to stay liquid, Field's metal probably wouldn't work. Maybe Galinstan would work, but the article doesn't discuss cost, maybe it's not so bad.

        • Re:Toxicity? (Score:5, Interesting)

          by mrmeval ( 662166 ) <.jcmeval. .at. .yahoo.com.> on Sunday July 20, 2008 @09:01PM (#24268301) Journal

          http://en.wikipedia.org/wiki/Liquid_metal_embrittlement [wikipedia.org]

          Hope they figured this out or mitigate it somehow.

          • Re:Toxicity? (Score:5, Informative)

            by Kazymyr ( 190114 ) on Sunday July 20, 2008 @09:51PM (#24268691) Journal

            Mod parent up - I was about to post on the same lines. Looks like all the likely candidates contain gallium, and gallium is not known to play nice to other metals - corrodes them etc.

            • Re: (Score:2, Informative)

              by rachit ( 163465 )

              ... Looks like all the likely candidates contain gallium, and gallium is not known to play nice to other metals - corrodes them etc.

              I was actually wondering as well, so I actually read the linked wiki article on Liquid metal embrittlement: From the article:

              Exceptions to this rule have been observed, as in the case of aluminium in the presence of liquid gallium.[1]

              • Re: (Score:3, Interesting)

                I work with gallium-indium-tin eutectic metals routinely. They definitely do not play well with some metals, specifically those commonly used in heat sinks. It will readily react with both aluminum and copper resulting in a nasty grayish black compound...not so good for moving heat. However, I do use it against nickel plated copper surfaces with no problems provided that the nickel strike is well done; i.e. it's not porous. The metals it does play well with are typically stainless steel, but stainless i
        • Re:Toxicity? (Score:5, Interesting)

          by WolfWithoutAClause ( 162946 ) on Sunday July 20, 2008 @09:43PM (#24268633) Homepage

          NaK melts at -13 C. It could be that in principle. It would burn if released, but in the intended use it should be safe enough.

    • Depends... (Score:5, Interesting)

      by ArchieBunker ( 132337 ) on Sunday July 20, 2008 @08:26PM (#24267959)

      Too much of anything can kill you. Water, salt, mercury, oxygen etc etc. They probably use some alloy with an extremely low melting temperature. The article is a bit short on the specifics.

      • Re: (Score:3, Interesting)

        by 4D6963 ( 933028 )

        Too much of anything can kill you.

        Semi-obligatory [youtube.com] :

        -Yes but too much [tobacco] is bad for you.
        -Well of course too much is bad for you, that's what "too much" means you blithering twat. If you had too much water it would be bad for you, wouldn't it? "Too much" precisely means that quantity which is excessive, that's what it means. Could you ever say "too much water is good for you"? I mean if it's too much it's too much. Too much of anything is too much. Obviously. Jesus.

    • Re:Toxicity? (Score:5, Informative)

      by Amiga Lover ( 708890 ) on Sunday July 20, 2008 @08:33PM (#24268045)
      I expect it's a metal related to these, http://www.indium.com/TIM/solutions/liquidmetal.php [indium.com] which are used as thermal interface materials in machines like Apple's 8 core Mac Pros. The heatsinks on those are wetted with a little of the liquid metal in place of stuff like arctic silver. While working on Mac Pros I found it's like mercury, but sticks to the processor heatspreader and heatsink base. It's liquid even in a cold room. There's toxicity info on that site somewhere, but I'm in a rush at the moment. No doubt someone else will find it and post.
  • by realmolo ( 574068 ) on Sunday July 20, 2008 @08:24PM (#24267947)
    Slashdot: Remember, Subby, when I promised to kill you last? Subby: That's right, Slashdot. You did. Slashdot: I lied.
  • by kaos07 ( 1113443 ) on Sunday July 20, 2008 @08:29PM (#24268001)
    I can see the Governator imposing a ban on the import of this product...
    • by Anonymous Coward on Sunday July 20, 2008 @08:43PM (#24268139)

      I can see the Governator imposing a ban on the import of this product...

      Warning:
      This product contains chemicals known by the State of California to cause cancer and/or the rise of cyborg assassins bent on enslaving mankind.

  • by actionbastard ( 1206160 ) on Sunday July 20, 2008 @08:37PM (#24268067)
    Chill out. Dickwad.
  • is what ends being your CPU if your cooler fails.

    Assuming that we are speaking here of room temperature liquid metal, which one? Some are rare, other are not liquid at normal room temperature, but could be when the cpu starts to get hot, or could be an alloy, but aren't so much choices afaik (or they don't specify as any will work).
  • What metal? (Score:4, Interesting)

    by dpbsmith ( 263124 ) on Sunday July 20, 2008 @08:39PM (#24268081) Homepage

    How odd that the article doesn't even hint at what the metal is. I wonder why not?

    • Re: (Score:3, Informative)

      by frostband ( 970712 )

      I'm not sure what the metal is either. At first I though they were talking about Liquidmetal(TM) [wikipedia.org]

      But they don't mention that it is those trademarked alloys. I still suspect it's another amorphous alloy (bulk metallic glass) even though that's not mentioned in the article and I haven't seen any other posts that mention that might be what it is. In any case, the Zirconium based Liquidmetal(TM) uses Zr, Be, Ti, Cu, and Ni. Here's the wikipedia article on Liquidmetal [wikipedia.org]

      I've done a bit of research on AA/BMG'

    • Re: (Score:3, Interesting)

      by rubycodez ( 864176 )

      a mimetic poly-alloy

  • Sodium cooling (Score:5, Informative)

    by bmo ( 77928 ) on Sunday July 20, 2008 @08:41PM (#24268105)

    " Awesome technology, if it actually works and is affordable."

    It works. It's worked for decades under the sea cooling nuclear reactors and in piston aircraft engines.

    http://www.enginehistory.org/air-cooled_cylinders_3.htm [enginehistory.org]

    " During his early years at McCook Field the ever-ingenious Sam Heron had observed the characteristics of various sodium compounds which are normally used in heat-treating operations. These materials are solid at room temperature and become liquid at engine operating temperatures. He observed that since these compounds wet the surface of steel alloys readily and transfer heat very well, their use should be effective in extending the life of exhaust valves. The ancestor of our present-day sodium-cooled valves had arrived, thanks to Mr. Heron, and almost ninety years later we are still enjoying the benefits of his ingenuity though even today such valves are not completely fault free."

    Also, it's not terribly expensive. Just don't go hacking into the reservoir or any of the tubes with a saw, mmmkay?

    --
    BMO

    • Re:Sodium cooling (Score:4, Informative)

      by Cal Paterson ( 881180 ) * on Sunday July 20, 2008 @08:53PM (#24268225)
      Some supercomputers already use liquid metal to supercool them (and, AFAIK, have done for a while). I believe it's normally some kind of gallium alloy, but I'm not chemist enough to say more than that.
      • by joib ( 70841 )

        Which supercomputers might that be? This is the first time I've heard of it.

        (Cray uses Flourinert for some machines, which is not a liquid metal).

    • Re: (Score:3, Interesting)

      by icegreentea ( 974342 )
      Sodium only works because nuclear reactors put out a lot of heat. You won't be able to sodium cool your computers. That's where the uncertainty comes in. What metal/metal-alloy is liquid below 100C, is relatively cheap, and safe. Mercury (the obvious answer for one and two) is almost certainly a nono for number three.
      • by viking80 ( 697716 ) on Sunday July 20, 2008 @09:06PM (#24268345) Journal

        An alloy of sodium (Na) and potassium (K) is liquid from 12.6 to 785 C

        it is cheap and wets most metals (good heat transfer)

        it is a little reactive, so recycling would need some special handling.

        • by Ellis D. Tripp ( 755736 ) on Sunday July 20, 2008 @09:13PM (#24268411) Homepage

          it is a little reactive, so recycling would need some special handling.

          A little reactive? It would burn pretty violently if simply exposed to air, and EXPLODE if it came in contact with water. And either event would produce highly corrosive byproducts.

          • Could be worse. Could be cesium [theodoregray.com].

          • Re: (Score:3, Funny)

            by maglor_83 ( 856254 )

            A little reactive? It would burn pretty violently if simply exposed to air, and EXPLODE if it came in contact with water. And either event would produce highly corrosive byproducts.

            Where do I sign up?!

          • Re: (Score:3, Interesting)

            by bmo ( 77928 )

            "A little reactive? It would burn pretty violently if simply exposed to air"

            You're exaggerating. You're thinking of cesium.

            It all depends on how far down the periodic table you go and how much.

            This is sodium and potassium

            http://www.youtube.com/watch?v=l9z5-mJ8NZk&feature=related [youtube.com]

            "EXPLODE"

            Yeah, but you're exaggerating more. If anything, there would be a gram or two of metal in the coolant tube to cool a CPU. Not a whole heck of a lot. We're not talking about throwing 20 pounds of the stuff off the bri

          • Re: (Score:3, Funny)

            by utnapistim ( 931738 )

            it is a little reactive, so recycling would need some special handling.

            A little reactive? It would burn pretty violently if simply exposed to air, and EXPLODE if it came in contact with water. And either event would produce highly corrosive byproducts.

            What he meant was ...
            Do not touch the operational end of The Device.
            Do not look directly at the operational end of The Device.
            Do not submerge The Device in liquid, even partially.
            Most importantly, under no circumstances should you-HHhhhHH!

            Doing so, will result

    • Re: (Score:3, Insightful)

      by Chryana ( 708485 )

      I can't speak for the piston aircraft engine cooling part, but the advantages of sodium weren't sufficient to keep using it in submarines [wikipedia.org]. Basically, if the reactor had to be shut down, it was impossible to restart, since the coolant would have by that time frozen solid. Not so desirable in the middle of the Pacific Ocean. So, as far as I know, cooling with liquid sodium is not used in any currently running nuclear submarine. Anyways, I am quite skeptical of the bold claims made in this press release. We se

    • Re: (Score:3, Informative)

      FWIW my dad's 1964 Ford had sodium-filled exhaust valves, and I'm told many higher-end Mercedes these days have the same.

      The problem comes when you're using the engine for drag-racing, and a connecting rod fails, slamming the piston into the red-hot exhaust valves. Things get very exciting.

  • by dacut ( 243842 ) on Sunday July 20, 2008 @08:41PM (#24268107)

    Google is being of limited help here. The main link I'm finding is to Liquidmetal Technologies [liquidmetal.com], which is producing Liquidmetal and Vitreloy -- zirconium-based alloys which are amorphous in structure (hence the "liquid" in the name) but are otherwise solid in appearance and use (and much stronger than stainless steel or titanium). This is not something one would be pumping through heat tubes to cool a CPU.

    Obviously, mercury is out due to its toxicity. My initial thought was they're using metal bits in a suspension, but I have doubts as to whether this would actually do anything useful. Deeper searching yields this page [scitoys.com], which describes a gallium/indium/tin alloy which is liquid at room temperature. Wikipedia'a entry for gallium [wikipedia.org] concurs, saying, "It has been suggested that a liquid gallium-tin alloy could be used to cool computer chips in place of water."

    Any materials experts out there care to comment?

  • by martin-boundary ( 547041 ) on Sunday July 20, 2008 @08:41PM (#24268109)
    Fair enough. How about mimetic poly alloy jokes?

    "That's not liquid metal, that's a mimetic poly alloy!"

    "It's as if millions of nano cpus suddenly cried out in terror and were silenced"

    "Mime or mime not. There is no try."

    "T-1000, I am your mould."

  • Magnetic pump? (Score:4, Interesting)

    by lawpoop ( 604919 ) on Sunday July 20, 2008 @08:58PM (#24268273) Homepage Journal
    The article mentions that this device uses an electromagnetic pump to move the heat around. In my naivete, I suggested a similar thing [halfbakery.com], without a pump. I imagined a sort of metal lava-lamp, where at the base, next to the CPU, blobs of molten metal would rise up towards the top of the heat sink. As they rose, they would cool off ( with the help of the fan) , and then sink again to collect more heat. So, the outside of the heat sink would be copper or aluminum or something, and the inside would be some low-temperature metal like tin. Traditional fin architecture would assist in dispersing the heat.
    • Re: (Score:3, Insightful)

      by maglor_83 ( 856254 )

      Isn't that exactly what a heat pipe does?

    • So, the outside of the heat sink would be copper or aluminum or something, and the inside would be some low-temperature metal like tin.

      From what I can figure out of low temperature liquid metals, inside would probably be some gallium-tin alloy if it was to be non toxic. If so, the heat sink would need to be copper. Gallium eats through aluminium.

  • We, Denmark (Score:5, Funny)

    by ZarathustraDK ( 1291688 ) on Sunday July 20, 2008 @09:03PM (#24268313)
    We give you funny cartoons and heatsinks.

    You give us Bush, Microsoft and Michael Jackson.

    Bad deal, Denmark thinks.
  • by TheModelEskimo ( 968202 ) on Sunday July 20, 2008 @09:04PM (#24268325)
    Actually, back in my sophomore year in college, my roommates and I built a liquid-metal-cooled 6-node cluster into the back seat of an old Fiero.

    The cool thing was, power was so cheap (via government subsidies called "grants") that we eventually upgraded the cluster to run a realtime terrain modeling system that was supposed to identify a path through a network of roads that allowed for the highest average speed, given speed limits and road lengths. The terrain modeling part would determine a way to maximize the time spent going downhill.

    DARPA initially supported us, and were going to upgrade our status, give us clearances, etc. but they eventually killed our funding after Ford found out we were using a Fiero and complained about our physical safety while operating the system. But man, DARPA know of some *way* cool ghost towns and low-traffic road networks.



    The preceding is a work of fiction written by an easily-distracted procrastinator in a severe time crunch
  • He's made out of a "mimetic poly-alloy" you insensitive clod! Evil governator killing robots have feelings too ya know!
  • by X ( 1235 )

    Supposedly it is reasonably priced, but I couldn't find any information on price anywhere.

  • Im just asking (Score:2, Interesting)

    by meeya ( 1152133 )
    cant it be sodium? or something of the class?
    • by MrNaz ( 730548 )

      Yep. Fun and games when your sodium conduit breaks at the same time as the conduit on your water cooling system.

    • It can't be pure sodium. Pure sodium becomes liquid at like 95 degrees or so. Getting your CPU to run that fast to begin with would be quite a feat. I also imagine if you get a hole in the piping, then your screwed. Sodium reacts violently with water (including water vapour in the air). It cannot end well.
  • Not a joke. (Score:5, Funny)

    by MrNaz ( 730548 ) on Sunday July 20, 2008 @10:04PM (#24268805) Homepage

    "The submitter requests that the moderators terminate all T-1000 jokes."

    The only way to do that would be to wait until the thread becomes stale, assess people who made the jokes, and then send back a cybernetic soldier to kill their mothers.

  • by Jaqenn ( 996058 ) on Sunday July 20, 2008 @11:03PM (#24269279)
    Go To the Danamics website: http://www.danamics.com/technology/background.aspx [danamics.com]

    Click on Technology. Then click background:

    Cooling of semiconductor hot-spots has developed into a demanding task. In recent years the high power density of microchips has increased rapidly, thereby increasing temperature and temperature non -uniformity on die and across chips. The drive for extremely low thermal resistance if further accelerated by the trend towards smaller system enclosures which has made the cooling task even more difficult. This gives the cooling companies a new challenge. To create cooling devices that is reliable, efficient and compact. Danamics does all of those.

    Then click Pump:

    Electromagnetic pumps have been used for decades and have been seen in a variety of applications. From space crafts and satellites, on nuclear-driven surface warships and submarines and also in land based nuclear reactors and industrial applications. In space, at sea, at land and now also as a key element for hot spot cooling. An electromagnetic pump has several advantages over typical mechanical pump designs. It contains no moving parts, shafts, seals, etc., emits no noise or vibration, has unlimited MTBF and suffers no performance degrading over time. Typically electromagnetic pumps have had the drawback of a high current requirement. The Danamics patent pending multi-string electromagnetic pump eliminates this limitation and delivers a high flow with minimal current input. Due to the small size and power requirements, the multi-string pump makes a liquid metal cooler suitable for use in smaller applications such as home computers, workstations and even laptops.

    Then click Liquid Metal:

    The liquid metal is a key component in Danamics cooling systems. Liquid metal has two major advantages when cooling high power density heat sources: Firstly it has superior thermo physical properties that decrease temperature - and temperature nonâ"uniformity - on die and across chips. Secondly the electrical properties of the liquid metal enables efficient, reliable and ultra compact electromagnetic pumping without the use of moving parts, shafts, seals, etc.

    Then click Benefits:

    Liquid Metal has numerous key advantages compared to competing technologies, such as water-cooling or traditional air cooling. The most important ones are: Ultimate performance Danamics LM10 is the worldâ(TM)s best performing air-cooler. By utilizing the unique properties of liquid metal, the cooler has the lowest thermal resistance of any air-cooler in the market. The cooling capabilities exceeds most watercoolers in a single device. Easy mounting Compared to competing solutions, the Danamics LM10 is a true all-in-one product. There are no external housings, large reservoirs or bulky radiators. With liquid metal cooling you have true performance in a compact, self-contained unit. This makes mounting a breeze and the Danamics LM10 cooler can be mounted in minutes. No moving parts The superior liquid metal brings another advantage besides its efficiency. The properties of liquid metal make the use of an electromagnetic pump possible. Danamicsâ(TM) patent pending multi-string pump delivers high flow combined with a very low power-draw of less than 1W. No noise and no vibration Compared to watercooling pumps, the electromagnetic pump used in Danamics LM10 cooler has numerous benefits. It is hermetically sealed and has no moving parts, which means that no noise or vibration is added to the system from the electromagnetic pump. Also the pump is orientation independent, which ensures maximum flexibility for users and system integrators. Innovative technology Danamics LM10 is built from the ground up using innovative technology. The combination of liquid metal and an electromagnetic pump is all new in commercial cooling devices, bringing

  • ... when you open the case to add a stick of RAM and the heat sink asks you if you've "seen this boy."
  • Nice try (Score:2, Redundant)

    by AuMatar ( 183847 )

    You can terminate all the T-1000 jokes you want, but they'll be back.

  • Obviously no information about how the liquid metal is made since this is an advertisement, but it's probably a repackaging of ferrofluid in a standard Chinese water cooler.

  • Technically I understand how liquid metal CAN be more efficient than water. But then again the flow of water is aided by pumps which force a lot of water per second through the cooler, not to mention that a proper radiator is about twice the size of that thing.

    Looks to me like Biosmagazine just copied that part from Danamics' press release. Without any comparative benchmarks (hell, we don't even have a comparison to a standard cooler) I'll remain unconvinced.

Real programmers don't comment their code. It was hard to write, it should be hard to understand.

Working...