DID WE JUST MAKE THE WORLD'S BEST THERMAL PASTE!?

hi there interface materials what are they and why are they important [Music] [Music] if you're familiar with this type of substance it's probably because you either upgraded or repaired the cpu in a computer and needed something to facilitate the heat removal into the heat sink but the global market for this type of material is much broader than just computers the automotive marine aerospace and industrial markets all utilize high-powered electronics just like the cpu that are too small to effectively radiate away or convectively lose heat into the environment and need to be able to transfer that heat into a heat sink or some sort of a heat receiver and because of the broad market of these materials certain properties often are emphasized in different types of materials last year when we did our video on the thermal epoxy we manufactured we formulated a thermal adhesive that was thermally conductive electrically insulating but acted also as an adhesive that holds the electronic component to the heatsink sort of two birds with one stone in the automotive industry you might give up a little bit of performance in order to be able to have something that can repeatedly cycle up to an above 150 degrees celsius in an engine compartment in the marine industry you got to think about salt water in the aerospace and industrial markets you have to think about both high and low temperatures as well as low pressures and even vacuum operation but in all cases the fundamental principle is the same when you take a highly conductive enclosure for an electronic and a heat sink and you machine the surfaces and you put them together they actually are only contacting at a few very very small points or ridges the major volume of the material between those two pieces of metal is air and air has a horrible thermal conductivity it's about 110 000 of aluminum so what you want to do is you want to get rid of that air now if you had access to a silicon foundry or a high quality optical shop and could manufacture surfaces that were flat to better than a wavelength of light and smooth down in the single digit nanometers in terms of average roughness you actually will get away without using any kind of a thermal interface material at all because the surfaces can get so close that the few remaining air molecules will not inhibit heat transfer very much you may actually lose ground by placing a thermal interface material between them because the increased separation more than compensates for the higher thermal conductivity of the material but if you don't have access to that kind of equipment simple machine surfaces are not nearly flat or smooth enough and you're going to need to use a thermal interface material now if you can apply a sufficiently high pressure above 100 psi or about 1 megapascal a well proven and excellent material for a thermal interface is indium film this is an elemental low melting point metal that is very very soft it is so soft that you can dent it with your fingernail and if you can apply those kinds of pressures this material will literally begin to plastically flow between the irregularities on a machine surface filling the pits and the scratches and the divots and provide you a metal to metal contact that can reach 10 times the thermal conductivity of any of the common thermal pastes it's relatively inexpensive it's non-toxic it will never evaporate and it will operate all the way down to cryogenic temperatures and in a high vacuum it's excellent stuff the problem is the pressure though because the pressure limits that i gave you with the bare minimums and to see those kinds of performance numbers you'd have to apply to a cpu sized area pressures above a quarter of a ton it's not practical you need a material that will flow under reasonable pressures and what that means you're going to do is you're going to use an organic liquid like a mineral oil or a silicon oil or you're going to use a polyalcohol like polyethylene or polypropylene glycol or even glycerin and you will add to that a high conductivity powder that forms a paste that has a higher thermal conductivity than the liquid alone but still remains flowable at reasonable pressures now i got into this a couple of years ago when i was building a solar concentrator array and i was also working on some led lighting systems for our fish tanks and so i started researching all these different types of materials and i found two interesting facts one the manufacturers don't tell you what's in their materials duh even if they hint at it like an arctic silver they don't tell you the exact formula and the other thing is that the specifications for thermal conductivity the numbers associated with these products are kind of squirrely you'll find one product that shows a thermal conductivity that's twice as high as a competitor but it doesn't work as well now that doesn't necessarily mean that they're lying they might be but not not necessarily it could be due to the different types of testing methodologies and the manufacturers don't always tell you how they tested them and there's really four different ways that you can test these one is a field test where you actually take the compound you want to use between the surfaces you're actually trying to control the temperature of and you'll test it just like you do when you put it into a cpu another method is called the laser flash method where you take a laser and you produce a controlled pulse of light which turns out to be a pulse of heat hit a surface that sandwiches the thermal interface material and have a heat or a thermal mass on the other side and you look at how effectively and quickly the heat is transferred through the material if you have access to that kind of equipment the big advantage is it's very fast you can do a test in under a minute and you can cycle through a lot of tests very conveniently another method is called the parallel wire technique and it's often used with larger volumes like liquids and slurries effectively what you have is two very thin parallel wires running through the material you're going to test one acts as a resistor and you put a current through it and it begins to warm the other wire next to it acts as a variable resistor and as it warms its resistance increases it's acting as a thermometer and so you measure the effectiveness of the heat transfer from wire one wire to the other based on the type of material that you put the two wires into that what you immerse them into the third and probably the most common technique is called the contact method and it's basically what you're doing when you're testing a cpu and it's what does guided the design for this piece of equipment that i built here the reason i built this was because i anticipated i was going to be testing a large number of material maybe doing dozens of tests and so as a consequence i wanted something that would not require me to be manhandling and cranking down on an expensive cpu as it turned out this little guy and me we've done hundreds of tests together because after researching this type of material and realizing it's really just a few common liquids and some conductive powders that you mix together as a paste i thought maybe i could do this myself i mean how hard could it be warning to engineers and you inventors out there whenever you use the phrase how hard could it be just imagine sitting in the corner of the room is murphy rubbing his hands together looking for a victim because it turns out it's a lot harder than you might think the devil is in the details nevertheless i struggled and i was able to produce two different thermal compounds one is very efficient and very effective and extremely inexpensive and the other material competes with the very best in the world and so i'm going to show you the relative comparison of those materials along with some of the commercial materials and give you an idea of just how they how they perform now to get an idea of what this equipment does let me kind of go through the overall sort of layout of the process here what we have is a lab chiller this is a device that contains both a refrigeration and a heating unit and it contains a large reservoir of thermally controlled water and holds the temperature of that water to better than a tenth of a degree celsius it pumps about five liters per minute through these hoses and into this aluminum block which is insulated on the bottom and clamped down and maintains the temperature of this block to within a tenth of a degree celsius so it acts as sort of an infinite heat sink on top of this block there is a two and a half centimeter or one inch diameter aluminum rod which stands up from the block and between the surface of the block and the bottom of the rod is the thermal interface material that we're going to test these four springs that are mounted around the periphery provide a controlled amount of pressure on the interface material and that pressure is approximately one atmosphere which is typical of how much pressure you would apply if you well crank down on a cooler on a cpu the stainless steel is an interesting material because it has one of the lowest thermal conductivities of any metal and when fabricated into a very long wire in the form of a spring very little heat is transferred through these springs but as i'll explain even that small amount of heat is controlled within the apparatus at the far end of this rod are four resistors that are bonded to the rod and are powered by a power supply which delivers 32.2 watts to this end of the rod which is typical for the area of a cpu of delivering about 150 watts of loading so again in the same neighborhood as the heat moves down through the rod it passes about one inch or two and a half centimeters above the interface a small hole that hole that is bored in the rod which we've mounted a thermal probe in that measures the temperature of the rod at that point so the basic principle is as this end heats and the heat is transmitted down to this heat sink the more effective the thermal interface material the less the temperature rises at the point of the probe and so this is what's going to tell us what the relative performance of the material is now you can see the surface of this block looks sort of mirror smooth and that's important because in any test of any ti any application or use of any kind of thermal interface material the flatter and the smoother you get the surfaces that are being contacted the better it's going to perform and so again last year with the thermal epoxy video i showed you some very simple inexpensive and takes five or ten minutes techniques for creating a very good smooth surface which will enhance the performance of anything that you use basically we take a machine surface we grind it even flatter and then we bring it to about a four to six thousand grit polish and that is done on this surface as well as the mating surface on the bottom of the rod the other thing you'll notice here is the big radiator because of the over capacity of the chiller i divert some of the water into this radiator on which i've mounted a couple of low-powered fans that are powered from a digital power supply this provides a slow circulation of air within the insulated box so that the air temperature within the box and surrounding the apparatus stays within 0.2 degrees celsius all the time day or night summer or winter and it is one of the advantages of doing a test like this rather than say in the cpu and a computer that may be in a room that's varying a lot in terms of temperature it's more controllable and then in addition to that one of the issues with any device like this any contact device is that it is very very slow it would take us about 90 minutes for this apparatus to reach equilibrium as it's slowly creeping up to a point where the temperature doesn't change at all and that's impractical so what we do is a timed test we bring the temperature of the interface or the probe down to 21.0 degrees and then as soon as it hits that temperature i switch on the power and switch on the timer and i run the test for exactly 12 minutes 720 not 721 seconds and at that point the temperature that we reach here measures the relative performance of our material in addition i never use the first test in a series of tests so if i'm going to do eight or nine tests in an evening i will not include the first one because test two three four five six etcetera always follow another test so i'm not including the cold test or the first test in order to try to again produce a very reliable result now i can't attest to the exact accuracy of this piece of equipment or how reliable it is but i've got reasons to believe it's pretty good first of all every time that i test the same material no matter when i do it i get the exact same result and secondly if you look at third-party sources like tom's hardware and look at the performance of commercial materials when they find something is better i find something is better so that doesn't prove anything but it means that probably if i have systematic errors here they're relatively small so what we're going to do is i'm going to go ahead and i'm going to fire up this unit and get it warmed up and then we'll get to the point of loading some of the material and show you how an actual test is done all right we've gone ahead warmed everything up and effectively i've done test number one the one that we're not going to include and so as this device is beginning to cool down i'm going to go ahead and we're going to test the first material so let me show you how we load this in here i take apart the insulating box so that i can access the equipment easily and then what i'm going to do is disconnect the springs so that i can get to the interface material this pliers is very useful because the two of these springs are pretty high tension and so it grabs the loop and allows me to pull the spring off more easily than i could with my fingertips and i'll get the other heavy duty one off then the light duty ones off and then finally separate the two components now the first thing to do is to get rid of the bulk of the material just take a piece of clean paper towel and wipe 95 percent of the mass of the material off the surfaces to begin the cleaning operation then i take some isopropyl alcohol this is not walmart 91 alcohol i like this squirt bottle and the reason i put the little tie wrap around here is actually i have some reagent grade 99.98 percent pure isopropyl in here but i like the squirt bottle and isopropyl is really good it will remove almost all of the thermal interface materials out there and it works very quickly so i dampen a piece of paper towel and then i wipe an initial wipe to get rid of most of the material [Music] and i'll do that a couple of times just to make sure and when you've got the bulk of the material off or almost all of it off if you look at the cloth it's completely white you should see no staining if you do keep going and that should be true of both surfaces then because the alcohol does still takes a little while to evaporate i then take a dry piece of paper towel and i dry off the mating surfaces and i always check to make sure i don't get any residual and then i take an organic solvent now i've tried dichloromethane hexane toluene they all work well and remove any kind of residual organic material that might be left behind and i like the toluene the best because the hexane evaporates too fast and it smells strong enough that if i forget to take the put the knot the cap back on i know i've left it open so i like using the toluene so what i'll do is i'll take another piece of paper towel and the gloves and we'll start with gloves from this point on and for you sticklers out there when i take the toluene out of here this is not how you should handle a reagent but this bottle is always used for cleaning so i'm never going to use this for some sort of chemical reaction and i'm going to wipe down one more time the surfaces still clean and one more thorough wipe around here then i'm going to wait about 30 seconds just to make sure that all the toluene has evaporated from the surfaces and then i'm going to place the compound on here now there i know there are some skeptics out there that say this whole obsession with thermal interface materials and pace it doesn't matter you could use anything you want it anything will work even toothpaste well will it so let's go ahead and start the first test by having a little bit of fun let's try to test some toothpaste so you can apply this in a number of different ways the way that i prefer is the dot method where you basically put down a little blob of the material and then you depend on the pressure between the two surfaces to squeeze the material out you could for example put a dot on there and a dot on there smear it around with your finger take a stick produce a grid pattern i found any way that i do this i get exactly the same result as long as i put enough there that i get a complete squeeze out of the material around the periphery of the rod so that i know that i've got contacted all of the surface area with the thermal interface material now i'll place this over the material and apply the first spring as i hold this in position and then the second spring these are easier to put on these springs back there have a little more tension so again it's easier if i use the clip to be able to pull these things up and over the bolts and then while i'm holding this thing down i can manually add this spring over here okay you see the squeeze out we got good squeeze out of the toothpaste and please in the comment section don't put down that i should have used colgate or one of the special whitening brands you can believe me that crest is one of the best thermal interface toothpastes out there put on the box put on the cover and as i said before this device is cooling off from the previous test so as the thermal as the temperature goes down and the thermal probe goes down toward 21.0 we'll get prepared to turn the power on and the timer so now we just have to wait now as you can imagine there's a lot of waiting in this process because it's 12 minutes of test then it's often maybe 12 or 10 12 15 minutes in the cooling operation between tests so this is a great time for you to listen to audio books or some music or read papers or think about a new project you might want to do you can do things as long as you keep your clock with you you can do a lot of stuff around this but it does involve a large footprint in terms of time so once we get to 21.0 we'll begin the test all right you can see we're getting pretty close we're at 21.1 and as soon as we hit a stable 21.0 i'm going to hit the power and i'm going to hit the timer all right that's it okay so power on time on and we begin 32.2 watts and 12 minutes and the final temperature we get here will be the temperature for toothpaste now an interesting thing about toothpaste is that it is surprisingly good as you'll see it's not bad but what's interesting about it is that the toothpaste has a property of polishing metal it can actually be used to as a low-grade polish to polish aluminum and if you rub it on you use the smear technique when you're putting it on you'll actually get some black material on the on your finger or on the stick and so you're really doing a bit of a surface prep when you're doing the toothpaste and that might be part of the reason it has a reputation for being not a bad thermal compound right so we're getting close we're at 11 minutes and 35 seconds and the temperature is at 30.9 degrees centigrade and as soon as we hit 12 minutes i'll turn the power off but the important thing is what does this number reach three two one that's it 31.0 degrees that represents toothpaste now that number really doesn't have any meaning outside of this machine it's idiosyncratic to the type of setup we have but it represents the performance of toothpaste in this machine so as we go through progressive tests you'll see what the temperatures are and the lower the temperature the better the material and that's how we're going to compare the performance of these different substances so now during the cooling operation is when i will replace the material so this is gonna be test three for me test two for you so i take off the box and i'm gonna go ahead and disconnect this clean this up and keeping with the whole sort of healthcare idea i'm gonna have fun with one more substance i'm going to go ahead and test honey and see how it compares to the toothpaste smells nice and minty nice edible thermal compounds so go ahead and squeeze these together and just like the temperature going up and going down it takes longer for the material to cool if it isn't as good or for the apparatus to cool if the material isn't as good so we have longer intervals for the cooling process because of the fact that the thermal conductivity of these first materials is not that good so we're going to wait for this to again drop down to 21.0 we're going to start the timer once again at zero seconds and give 12 minutes to testing the honey okay here we go power on timer on so 12 minutes for honey now an interesting thing about the honey and the toothpaste is that as you will see they actually perform surprisingly well and one of the reasons for that is the fact that they contain water and water is a very unusual liquid except for the very strong mineral acids like hydrofluoric and nitric acid and obviously mercury and gallium water conducts heat about three times better than any of the other compounds that you might use as the liquid in the thermal interface material the alcohols the oils three times better and so that's largely what's carrying these products into a pretty competitive range the problem with water though obviously is evaporation so that after a period of time as the water evaporates and leaves a hard crust behind it what fills the voids is air so you might get some decent testing on the wall on the water content in the honey and the the toothpaste initially but it's not going to last so it's not a practical long-term test but it's an interesting fact so here we go we're just a little over 12.12 8 and we get 30.6 degrees 30.6 at 12 minutes so that's honey now the reason the honey probably performs a little bit better than the toothpaste is simply because it's very low viscosity it has very high lubricity it allows flow very effectively and so the two surfaces likely approximate a little bit better than they do with the toothpaste so that's probably one of the reasons why it performs as well as it does in any case now that i'm beginning the cooling process what we're going to do is i'm going to get more serious and the next thing that we're going to test is arctic silver's mx4 this is their sort of low-cost version of a thermal interface paste this is a commercial material i'm going to go ahead and load this in here and we're going to begin a 12-minute test on the mx-4 all right completely pumped out good so now we're going to put this on here and wait for this to cool down to 21.0 to do the mx4 test i've done this test so many times looks like we're there nope not quite go time all right so let's see how this performs and it looks like we're going to get to 30.5 degrees so the mx4 actually performs just about as well as the honey kind of interesting now what we're going to do is i'm going to use some of the arctic silver this has sort of been the gold or silver standard for decades and we're going to replace the mx4 with the arctic silver and see how well it performs on there nice squeeze out and now we got to wait for this to cool one nice thing about the higher performance materials is they cool faster so it saves a little bit of time in this part of the cycle but we got to wait to 21.0 and then we're going to begin the test for arctic silver 5. so here we are we're at 21.1 getting ready to stabilize at 21.0 good okay so this is arctic silver 5 the classic arctic silver and we're going to do a 12-minute test and see what kind of temperature we got so basically 30.1 degrees at 12 minutes for arctic silver 5 30.1 degrees now the next material that i'm going to test is our very low cost thermal interface material the one that we manufactured and we'll compare that so i'm going to go ahead and load it into the test rig set the timer and just wait for it to cool and wait all right so we're just waiting for 21.0 and then we'll start the test on our material all right got it okay so here we go 12 minutes on our low cost material 11 56 57 58 59 so basically 30.1 30.1 and a half just a little over 30.1 so it's basically about the same as the arctic silver the difference is that this material can be produced for about a penny a gram as opposed to a dollar a gram for arctic silver a significant change in terms of price so pretty impressive for something that you'll see is very very easy to make sorry arctic silver so let's go ahead and we'll go ahead and begin by testing a little bit higher performance materials this is the noctua thermal paste i'm going to go ahead and replace ours with this all right let's go ahead and let this cool off good okay here we go 12 minutes all right so we're at 11 minutes and 45 seconds 30.0 degrees on the noctua 53 54 55 56 57 59 12 minutes so 30.0 degrees for the noctua so the next material we're going to test is the prolima tech pk3 okay i'll give that a few seconds to dry stuff comes in a very small little syringe it's kind of pricey all right okay so 21.0 for a prolimatech pk3 and there we go 12 minutes and we have 29.9 degrees for the pk-3 now the next material i'm going to test is dow sil 5622 it's a industrial thermal interface material good squeeze out yeah all right so we're going to let this cool i'm going to do it again all right we got to wait for 21.00 21.0 go so 12 minutes dow cell 5622 okay so 11 to 52 53 54 it looks like we're going to hit 30.0 degrees 12 minutes 30.0 degrees for the dow sill product so now the next material is going to be the shin etsu okay good pump out and when we get to 21.0 we'll go ahead and check out the japanese product good beginning 12 minutes so i would call this 29.9 to be fair 29.9 kind of popping between 20 30.0 and 29.9 but let's call it 29.9 to be generous all right so now what we're going to do is we're going to test the high performance material that i formulated now this material that i fabricated i'm still leaving this in the mortar and pestle and we'll take a small sample of that with a stick and apply that down there all right so let's let this cool and then we're going to test this so we got to get this down to 21.0 degrees and it's cooling pretty fast okay so we're at 21.0 just want to see it stabilize we're still seeing a little bit of 21.1 and this is going to be a test of our high performance material okay so 12 minutes let's see what kind of result we get okay so 11 59 12 minutes 29.8 degrees so not bad this is actually better than the shinetsu by a 10 10th of a degree celsius so as you can see the material that i formulated compares very favorably with some of the leading commercial materials out there in addition this costs less so in view of both those factors is this the best thermal compound in the world probably not but it might be and so we're going to do three things number one we're going to send a couple samples of this material over to linus tech last year they did a review of our thermal epoxy and a thermal adhesive and they did a good job but they did admit that it was a bit of an apples and oranges comparison because thermal adhesives don't perform as well as thermal pastes so we're going to send them some real thermal paste and if they're willing to do the test get their feedback in terms of its performance it's ease of use etc and we have no relationship with linus tech we do follow their channel but other than that all we can say is that they're obviously just a wonderful bunch of guys the second thing is in the next video i'm going to cover the techniques that i use to generate r2 materials i'll give you the formulas i'll show you the methods so that if you want to you can reproduce the material that i produced and if you don't want to spend the time or the effort or the money in order to make this yourself we're going to put this up for sale on our website so you can go ahead and buy a very high performance thermal paste for a lower cost than most of the commercial materials out there and that does two things one is it helps to support the channel support the cost of the materials and the equipment that we use for our videos but it also gives a much broader user base for feedback in terms of the performance of the material as a little aside i just want to mention that a lot of people or i would say a significant fraction of the audience follows these videos very very carefully and they'll say you know is that a new shirt or a nice haircut or you need a haircut but for those of you that don't follow quite as closely you may not have noticed the progression on our stairway over there you know it takes a long time to film these many tests but it is just a joke no actual alcoholic beverages were consumed during this flight in any case if you have a question you want to make a comment put it down in the comments section below i read them all i try to answer as many questions as possible and sometimes it gives us ideas for new videos in the future and i would like to appeal to you if you like the kind of content that we cover here whether it's computers or chemistry distillation rocket engines jet engines lasers radios etc you might consider subscribing because there are two really important reasons one it gives us positive feedback for the kind of effort that we put into these videos but it also helps to increase the apparent footprint of the channel the gravitas and when we get people applying to become employees or when we go out to outside groups like we did with mit for the thermonuclear fusion video or as we're going to do next week with alteros we're more likely to get positive feedback from companies and from universities if our channel has a larger footprint a larger potential viewing audience so if you would be willing to subscribe but don't want to be annoyed don't hit the notification bell and if you do want to be notified make sure you click on all so that youtube has less flexibility to sort of filter what we produce as to what they show you and then don't depend on that either because youtube has kind of a hit or miss reputation for notifying people about new videos on their subscriber list in any case i want to thank you very much for watching spending your time with me and i'm going to wish you a very good evening you take care [Music] you

Channel: Tech Ingredients

Views: 296,895

Keywords:

Id: Glfak2B9J5U

Length: 41min 52sec (2512 seconds)

Published: Fri Sep 10 2021