**Liquid Metal: A Game-Changer in Heat Transfer?**
The concept of liquid metal has been gaining attention in the tech community, particularly among enthusiasts who are passionate about computer hardware and power electronics. In this article, we'll delve into the world of liquid metal, exploring its properties, advantages, and potential applications.
**What is Liquid Metal?**
Liquid metal is a type of thermal interface material (TIM) that uses liquid gallium or other metals to transfer heat between components. Unlike traditional TIMs like thermal paste or pads, which rely on solid-state conduction, liquid metal takes advantage of its high thermal conductivity and fluidity to efficiently dissipate heat.
**Benefits of Liquid Metal**
The use of liquid metal in heat transfer has several benefits. For instance, it can reduce the temperature difference between a component and its heat sink, leading to improved system reliability and lifespan. Additionally, liquid metal can be applied with precision, making it ideal for small-scale electronics or high-power applications.
**Real-World Applications**
While liquid metal shows promise in various fields, its effectiveness depends on the specific use case. For example:
* In computer hardware, liquid metal can be used to improve heat transfer between a processor and its heat sink, potentially leading to better performance and reduced power consumption.
* Power electronics, on the other hand, may not benefit as much from liquid metal due to lower heat flux densities.
**Challenges and Limitations**
Despite its potential benefits, liquid metal also comes with some challenges. For instance:
* Liquid metal is electrically conductive, which means it can short-circuit components if applied incorrectly.
* The high cost of liquid metal may make it less accessible for widespread adoption.
* Some materials used in liquid metal have specific requirements for handling and storage.
**Conclusion**
Liquid metal offers a promising solution for heat transfer applications, but its effectiveness depends on the specific use case. While it shows promise in computer hardware and power electronics, further research is needed to fully understand its benefits and limitations. By exploring the properties and potential applications of liquid metal, we can better determine when to use it and how to overcome any challenges that may arise.
**References**
* The content of this article was taken from a video presentation on YouTube.
* Additional information on liquid metal can be found in various online sources, including academic papers and technical articles.
WEBVTTKind: captionsLanguage: enThis is liquid metal and I first heard aboutit while watching teardowns of the new PlayStation5.In a nutshell it sits between the CPU or whateveris producing the most heat in your systemand the heat-sink.And its job is to basically conduct the heatfrom one part to the other as fast as possible.Now you might be thinking: “Wait, don'twe already have thermal paste for that whichdoes the same job and worked just fine fordecades?”And the answer would be Yes!Thermal paste does the same thing.But when it comes to how fast such a substancecan transfer heat then my used thermal pasteonly comes with a thermal conductivity of8.5W per mK while the liquid metal featuresa conductivity of 73W per mK which is quitea difference.So the logical conclusion would be that liquidmetal is superior and we should simply abandonthermal paste.But that conclusion is most of the time nottrue and in this video I will show you exactlywhy by more or less playing around with liquidmetal and doing a couple of tests.Let's get started!This video is sponsored by Keysight.Knowing how to use an oscilloscope is an essentialskill, and Keysight University’s Oscilloscopes101 course teaches you everything you needto become an oscilloscope-pro.Learn the basics, probing, and how to useadvanced capabilities that will speed up yourdesign and testing.Use the link in the description to sign upfor Oscilloscopes 101 or one of the other100 engineering courses available on KeysightUniversity.First off let me tell you that getting liquidmetal is not really hard.I got mine from Amazon for around 12€ whichis more expensive than normal thermal pastebut I think still affordable.Now when opening up such a pack of thermalliquid then you are greeted not only withthe syringe filled with metal, but also withtwo tips for placing the metal and removingit, a manual, two q-tips for distributingthe metal onto the surface and finally alcoholpads for cleaning your CPU surface or similarwhich I will not be needing.You also get a big fat warning letting youknow that you can not use liquid metal withaluminium heat-sinks which is a shame becauseas a power electronics engineer those arethe kinds of heat-sink I have lying aroundthe most.But just for fun let's try it out anyway andas you can see after a short amount of timethe liquid metal reacted with the aluminiumand turned into a crumby alloy that will certainlynot conduct heat well any more.The good news for all power engineers thoughis that the liquid metal does not react withthe exposed metal part of MOSFETs or otherpower switches.That means the usage in power supplies ispossible if we would replace the aluminiumheat-sink in them.And in case your are wondering even thoughliquid metal is obviously liquid, it stickswell to metal and thus should make a verticalmounting position possible.But enough with power electronics for nowand instead let's focus on Computer applicationsof liquid metal for your CPU and GPU.Now I hate to disappoint you but I am notdoing any tests with my work horse computerbecause I need it on a daily basis and thereare already great benchmarks out there ongamernexus with CPU and GPU temperatures usingliquid metal instead of thermal paste.Instead why not do a comparison with a RaspberryPi of which I have a few with the newest onebeing the Raspberry Pi 4 whose main chip evenlooks similar to a normal CPU.And best of all you can even get such a cutelittle heat-sink for it onto which it shouldbe possible to apply thermal paste and liquidmetal.So my idea is to stress test the RaspberryPi's processor without heat-sink, with heat-sinkand thermal paste and with heat-sink and liquidmetal and then compare the temperatures toone another to see if there is a difference.To do that I firstly installed the RaspberryPi OS onto a micro SD card which I then pluggedinto the Pi before powering it up.Through the help of Putty I then connectedto the Pi in order to not only update itssoftware but also to install stressberry whichis the amazing piece of software that basicallyautomates the stress testing for me and willoutput such easy to understand graphs.And as soon as that was done, all I had todo was to type in a specific command in orderto start the first stress test.By the way during all tests in this videoI always tried to keep the room temperatureat around 20 to 21 degree Celsius so thatwe can fairly compare the results.With that being said here are the resultsof the first run with a peak temperature ofalmost 80 degrees.To lower that I completed the assembly ofthe mini heat-sink which I then secured inplace on top of the processor after applyinga thin layer of thermal paste onto it.After then plugging in 5V power for the fanit was time for the same test as before whichaccording to the output graph lowered thetemperature down to around 33 degree whichis a huge difference to before.But now the problem is that with such lowtemperatures we will probably not see a temperaturedifference with liquid metal since the morelimiting factor in such a low temperaturesetup is the heat-sink design rather thanthe thermal conductivity between processorand heat-sink.So what I tried next was overclocking thePi which initially worked out just fine andbumped up the temperatures to around 36 degreeat a clock rate of 2MHz.But while trying to go faster and thus reachinghigher temperatures my Pi always crashed andfroze during testing.That is why for my last attempt I simply unpluggedthe fan and went with the original 2MHz overclockingagain which to my surprise also crashed whileit was reaching around 53 degree.So without really any hope to see a differenceI removed the heat-sink, cleaned it up andof course only noticed at that point thatit was also using aluminium.The solution for that problem was using mydremel to get the aluminum away from the copperand afterwards I not only put a bit of painterstape around the processor but also appliedthe liquid metal onto it.In case you are wondering why the painterstape protection then let me tell you thatliquid metal is actually electrically conductive.After testing this property for a bit I candefinitely confirm that which means that ifthe liquid metal gets misplaced it can shortcomponents and or traces and thus fry yourcomplete system.In my case though the liquid metal didn'tfry anything and the Pi started successfullyafter mounting the heat-sink and pluggingin the fan.The first stress test I then did was with2MHz overclocking and active fan which surprisinglythis time also crashed but delivered a constanttemperature before that of around 36 degreewhich is pretty much identical to the thermalpaste test.What was interesting though was that withoutthe fan, the liquid metal setup reached atemperature of 50 degree in around 8 minuteswhile the thermal paste before only needed5 minutes to get that hot.Seems like the better thermal conductivitytransferred the heat faster to the heat-sinkand thus gave the heat-sink more time to emitit meaning it reached the 50 degree Celsiusmark later.That doesn't sound too bad but for a properverdict and to make my point crystal clear;let's do one last experiment with this transistorand this heat-sink which also comes with afan.After drilling a hole into it and creatinga thread I firstly used liquid metal for theheat transfer between the transistor and heat-sink.According to this very simply schematic Ithen added a resistor and powered the transistorwhich in this configuration now creates aheating power of around 16W.Next I measured how hot the transistor becomesover a time period of 2 minutes.And afterwards I cleaned up the liquid metalwhich is actually easier to do than I thoughtbefore making this video and applied thermalpaste in order to repeat this exact same experiment.In this graph you can see that the transistorwith liquid metal was around 3 degrees colderat the start in comparison to thermal paste.But this difference shrank more and more overtimeand at the end both setups basically camewith the same end temperature.And believe it or not but this graph actuallysolves all liquid metal application questionsfor me.You see according to the first temperaturedifferences it is clear that liquid metalreally gets the heat faster from the componentto the heat sink but the end temperature clearlyshows that for many setups that does not matterbecause the heat sink can not get rid of theheat fast enough and thus is the limitingfactor.So only if you have something like a processorwith tons of transistors inside which getsmaller and smaller over the years in combinationwith a well designed heat-sink; then it makessense to utilize liquid metal to transferthe heat as fast as possible.For power electronics on the other hand itdoes not make that much sense since the heatflux density is not that big unless you areworking with really high power stuff thateven requires active cooling.So in the end I think liquid metal does comewith significant advantages but also disadvantagesmeaning the usage of it always depends onthe kind of system you are working on.But I think since you watched this video youshould now have a better idea of when to useit.If you liked this video and want to see morethen consider supporting me through Patreon.As always don't forget to like, share, subscribeand hit the notification bell.Stay creative and I will see you next time.