Fullscreen HTML5Med Res (??MB)HTML4

Heat pipes, vapour chambers and thermosiphons Heat pipes, vapour chambers and thermosiphons are “two-phase” cooling (or heating) devices that transfer heat from one place to another very efficiently. They are simple and inexpensive, with no moving parts and do not require external power. Yet they conduct heat so much better than metals like aluminium or copper that they can be considered ‘heat superconductors’. T he t e r m i n o lo g y regarding these devices is somewhat confusing. According to some definitions, vapour chambers and thermosiphons are simply variations of the standard heat pipe. That is logical since they all operate on the same principle. For clarity, we will refer to the most common type of heat pipe as a constant conductance heat pipe (CCHP), to distinguish it from the other types of heat pipe such as the thermosiphon and vapour chamber. These devices all operate on similar principles, with some differences as follows: • CCHPs can operate in any orientation, transferring heat from one place to another and are generally in the form of a cylindrical pipe. • Thermosiphons are similar to CCHPs but operate with the assistance of gravity, and thus can only work correctly in a particular orientation. • Vapour chambers distribute heat evenly over an area instead of transferring heat from one location to another. In their modern form, CCHPs were initially developed for space applications but are now widely used in many areas, especially electronics. As computer chip component density and speed becomes higher and higher, the amount of heat generated becomes difficult to remove, even with huge aircooled or liquid-cooled solid copper or aluminium heat sinks. Consider that an integrated circuit like the Nvidia GA102, with over 28.3 billion transistors onboard, has an area of just 628mm2 – about the size of a postage stamp – yet dissipates up to 450W in operation! Traditional heatsinks have no hope of removing that much heat without the silicon junction temperature greatly exceeding 100°C, therefore another solution is needed. Enter heat pipes Heat pipes are used either when a traditional heatsink cannot efficiently remove the heat from a device or when weight or size targets can’t be met with conventional heatsinks. Commonly, these considerations apply to modern computers. Water cooling (via a water block, pump and radiator with fans) is another possible solution in some cases. Still, it introduces complications like pumps, pump noise, potential pump failures and the possibility of water leaks. These problems do not occur with By Dr David Maddison siliconchip.com.au Australia's electronics magazine May 2022  19 heat pipes which are now important elements of the CPU (central processing unit) and GPU (graphics processing unit) cooling assemblies in many desktop and laptop computers, plus many other electronic devices. Heat pipes and vapour chambers are even used in some smartphones. Without adequate cooling, modern CPUs and GPUs would be destroyed in seconds if they didn’t have internal overheating protection to shut them down. Two-phase cooling devices are also used for high-power IGBTs (insulated-­ gate bipolar transistors) in wind turbines, electric vehicles, data centres and solid-state lasers, among other applications. Heat pipe construction Constant conductance heat pipes, vapour chambers, thermosiphons and related two-phase devices are sealed hollow metal tubes or cavities that have been evacuated of most air (to a low pressure), into which a tiny amount of liquid has been placed. Some of this liquid evaporates to its vapour form given the low pressure inside the tube. Depending upon the application, the typical liquids used are water, ammonia, alcohols such as methanol or ethanol, R134a refrigerant or liquid alkali metals such as sodium. We will Expansion Vessel Furnace Coils heat source by capillary action, gravity or some other force, and the process is endlessly repeated, removing heat from the object to be cooled. Heat pipes such as CCHPs and thermosiphons are typically used for cooling as described above, but the process is equally applicable to supplying heat to an area. It depends on which end of the heat pipe the object to be heated or cooled is located. In the case of vapour chambers, they can be used to evenly distribute heat as well as cold. History of heat pipes Initial Venting Filling Fig.1: the Perkins System of heating from British Patent 6146, dated 30th July 1831. later discuss what liquids are used in different applications. Heat pipe operation During operation, liquid at the heat source (evaporator end) absorbs heat and evaporates. The vapour migrates to another area of the pipe (usually the other end, called the condenser). There, it condenses into a liquid and releases ‘latent heat’ (described later) into the surrounding environment. This latent heat represents a large amount of energy. The liquid then migrates back to the Angier March Perkins (son of Jacob) invented what was to become the antecedent of the heat pipe and obtained US Letters Patent No. 888 in 1838 and UK Patent No 6146 for his invention (see siliconchip.com. au/link/abd5). Later, he and his son, Loftus Perkins, invented a hermetically-sealed boiler tube with water or another liquid as the working fluid. It was a heat transfer device; however, it was single-phase (liquid-only) and operated at high pressure (about 20 atmospheres) and high temperature (150°C or more). By comparison, a modern heat pipe uses two phases, eg, water and steam. It was highly successful for about 100 years and was known as the “Perkins System of Heating”. Many of these systems are still in use today in Fig.2: a steam locomotive built by Jacob Perkins in 1836 using his sealed steam tube patent of that same year. The device became known as the Perkins Tube. Fig.3: a Perkins steam oven displayed at the Paris Exhibition of 1867 that used a Perkins Tube. 20 Silicon Chip Fig.4: an advertisement for a Perkins steam baking oven, probably from the 1890s. Australia's electronics magazine siliconchip.com.au southern England and Wales; some are 160 years old. The chronology of heat pipe development is confusing because an important patent of Jacob Perkins from 1836 is widely misquoted as having been awarded in 1936. This is British Patent No 7059, 12th April 1836, “Steam engines; generating steam; evaporating and boiling fluids for certain purposes”. This device was a sealed vertical tube filled with water that passed over an evaporator and then a condenser. It relied on gravity for the cooled condensate to return to the heat source (see Figs.1-4) and became known as the Perkins Tube. As this device contained both water and steam, it was a twophase device, like a modern heat pipe. Perkins Tubes were first used in locomotive firebox superheaters. Another important use was “stoppedend steam tubes” in bread-making ovens, patented by Loftus Perkins in 1865. These were adopted by the British Army some years after difficulties encountered feeding troops in the Crimean War (which ended in 1856). The ovens contained a multiplicity of slightly sloping tubes above and below where the bread was baked, each hermetically sealed and filled with distilled water. The lower end of each tube was immersed in the furnace. The ovens were widely acclaimed because of the even, continuous heat they supplied, plus their economical operation. The Perkins Tube relied on gravity to return the condensed liquid; today, they would be known as a two-phase Fig.5: F.W. Gay’s thermosiphon heat pipe invention, as disclosed in US Patent 1,725,906. siliconchip.com.au thermosiphon. Later, we will discuss the various types of heat pipes in greater detail. For more on the engineering genius of the Perkins family, see siliconchip. com.au/link/abd6 Later developments In 1942, F. W. Gay developed a finned heat pipe gas-to-gas heat exchanger in the form of a thermosiphon to exchange heat between a flow of hot air and cold air (see Fig.5). The main problem with thermosiphons is that they rely on gravity, so they only operate in a particular orientation. But this problem can be solved by using very small diameter pipes called capillaries. The flow is then dominated by capillary action, which can act in opposition to the force of gravity. Simple examples of capillary action are the way paint is drawn into the bristles of a paintbrush, or how water soaks upwards in tissue paper. This action occurs because intermolecular forces dominate the liquid’s smallscale behaviour, rather than gravity. A capillary-based heat transfer device was the subject of the 1942 patent application of Richard S. Gaugler of General Motors (awarded in 1944) for a “Heat Transfer Device” – see siliconchip.com.au/link/abd7 However, nothing seems to have come of it at the time. The idea of the patent was that, unlike a thermosiphon, his capillary-based heat transfer device (which today would be called a heat pipe) could function in any orientation. Independently of Gaugler’s work, and seemingly without prior knowledge of it, in 1963 George M. Grover of the US Los Alamos National Laboratory independently discovered the heat pipe and filed a patent which was awarded in 1966 for an “Evaporation-­ Condensation Heat Transfer Device” – see siliconchip.com.au/link/abd8 He coined the term “heat pipe”, mentioned in the patent application. Apparently, the patent examiner was aware of Gaugler’s work (citing it) but awarded the patent anyway. Both inventions are almost identical, using materials such as metal powders attached to the inside of capillary tubes to enhance the capillary action by the wicking effect. But while Gaugler’s was not widely known or put to use, Grover’s was, and he became known as the “father of the heat pipe”. Grover’s work saw the heat pipe put to use in space applications by NASA. Latent heat To further understand the operation of heat pipes and related devices, we must first discuss latent heat. Latent heat is the release (or absorption) of heat that occurs during a ‘phase transition’ such as between solid, liquid and gaseous states (see Fig.6). It can also be released or absorbed due to structural changes within a material, such as changing from one crystal structure to another. For example, consider that if you had ice at 0°C and you added heat to it, it would melt and become liquid water, but the water could still be at Fig.6: water’s energy content vs its temperature at atmospheric pressure. Energy added or removed can either change the temperature or change the phase. The change in phase at constant temperature is indicated by the horizontal areas of the graph and is due to latent heat. The sloping areas of the graph indicate changes in temperature (sensible heat). Original source: Wikimedia user Cawang (CC BY-SA 3.0) Australia's electronics magazine May 2022  21 0°C. Where did that heat energy go? It is the heat of fusion and is returned when the liquid water is re-frozen. Similarly, if you heat liquid water, you get steam at 100°C, with the added energy being the heat of vapourisation. That energy is returned when the steam condenses as the heat of condensation (making steam burns even worse than they already are). Another example is the process of sweating, which results in the body being cooled due to energy removed in the latent heat of vaporisation of water as the sweat evaporates (swamp coolers use the same effect). During the release or absorption of latent heat, two phases of the substance coexist, such as liquid water and ice or liquid water and water vapour. There is a lot of energy associated with these transitions, which is why ice keeps a drink much colder for longer than simply having the drink at a temperature close to freezing. Similarly, there is a lot more energy in steam than there is in water close to the boiling point, which is part of the reason why steam is effective for powering steam engines or turbines in power stations. Latent heat versus sensible heat for cooling Because of the large amount of energy associated with latent heat, it is much more efficient than traditional sensible heat cooling. Latent heat is shown as the horizontal regions in Fig.6, while sensible heat corresponds to the sloped sections. Note how the heat of vaporisation is considerably higher than the energy required to raise water temperature from 0°C to 100°C! Table 1: typical working fluids for heat pipes & their operating ranges. Working fluid Operating temperature range Silicon Chip Operating temperature range Helium -271°C to -269°C Ammonia -75°C to +125°C Hydrogen -260°C to -230°C Methanol -75°C to +120°C Neon -240°C to -230°C Acetone -48°C to +125°C Oxygen -210°C to -130°C Water +1°C to +325°C Nitrogen -200°C to -160°C Caesium +350°C to +925°C Methane -180°C to -100°C Potassium +400°C to +1025°C Ethane -150°C to +25°C Propylene -150°C to +60°C Pentane -125°C to +125°C Methylamine -90°C to +125°C To put it another way, it takes much less energy to boil a kettle full of water starting at 0°C than it does to convert all that boiling water into steam. You can easily observe this yourself if you force a kettle to stay on after the water is boiled for a period equal to the boiling time. Most of the water will still be liquid by the end. Elements of a heat pipe Heat pipes, and similar, essentially comprise a container (often a tube but not necessarily), a working fluid and possibly a wick or capillary structure – see Fig.7. The container must: • be easy to fabricate • be chemically compatible with the working fluid • be wettable by the working fluid • have sufficient strength and good thermal conductivity • in cases like spacecraft applications, be light Common materials used for heat pipes are copper, aluminium and stainless steel. More exotic materials Fig.7: the operation of a heat pipe. The working fluid evaporates at the hightemperature end and absorbs energy (1). It then migrates along the cavity to the low-temperature end (2) and condenses, releasing its latent heat (3). The liquid is absorbed by the wick structure and migrates back to the high-temperature end (4), repeating the cycle. Original source: Wikimedia users Zootalures & Offnfopt (CC BY-SA 3.0) 22 Working fluid Australia's electronics magazine NaK +425°C to +825°C Sodium +500°C to +1225°C Lithium +925°C to +1825°C Silver +1625°C to +2025°C such as tungsten, molybdenum, niobium and Inconel are used for the highest-­temperature applications. Among other characteristics, the working fluid must: • be able to wet any wicking material present • be able to wet the container walls • be chemically & thermally stable • have a high latent heat • have high thermal conductivity • be able to exist as a liquid and vapour over the desired temperature range • have a high surface tension to drive capillary action • have low vapour and liquid viscosity to aid flow The working fluid used chiefly depends on the desired temperature range of the heat pipe. Water is the most common working fluid, with an operating temperature range of +1°C to 325°C. The lowest temperature heat pipe uses helium for a range of -271°C to -269°C and the highest temperature pipe uses silver for an operational Fig.8: several basic, straight, constant conductance heat pipes (CCHPs) of the type that can be bought online very... siliconchip.com.au range of +1625°C to +2025°C. See Table 1 for other working fluids. Correct selection of the working fluid is essential; if the temperature is too high for the fluid, it will be all gas, and if too low, it will freeze. The temperature range must accommodate the coexistence of both liquid and vapour of the chosen fluid. The velocity of vapour in a heat pipe is surprisingly high, approaching the speed of sound. The return liquid flow is at about walking speed. Figs.9(a) & (b): a CCHP with a metal sintered powder wick opened up. Source: Thermolab (http://thermolab.co.kr/) Wicks One of the defining features of a heat pipe compared to a vapour chamber or thermosiphon is the presence of a wick or wicks. The function of a wick is to transport working fluid from the condenser back to the evaporator by capillary action. Wicks come in various forms, such as sintered metal powder, grooves in the tube, a screen mesh or other porous or fibrous wicking structures, such as carbon fibre or ceramic fibres. For some examples of wicks, see Figs.9-11. Sintering is when small particles of metal are fused by heat and pressure, forming a porous solid structure with a very high surface area. Figs.10(a) & (b): a CCHP with a grooved metal wick opened up. Source: Thermolab (http://thermolab.co.kr/) Heat pipe types There are many variations on heat pipes, but we’ll concentrate on describing the more common types. Standard heat pipe (CCHP) Figs.11(a) & (b): CCHP with a metal mesh wick opened up. Source: Thermolab (http://thermolab.co.kr/) The constant conductance heat pipe (CCHP) is the most common type of heat pipe and is ‘simply’ a partially evacuated, sealed tube with a wicking material and a working fluid inside (see Figs.8 & 12). It transfers Fig.12: the operation of a typical constant ► conduction heat pipe. Heat applied to one end causes the working fluid to evaporate and flow along the centre of the tube to the cold end. It then condenses and flows back to the hot end, along the capillary wick, and the process repeats. ...inexpensively for experimentation. Other CCHPs may have bends and attachments to suit. siliconchip.com.au Australia's electronics magazine May 2022  23 Fig.13: a CPU cooling assembly (known as a “tower cooler”) with six heat pipes. Note how they are flattened to make good thermal contact with a CPU. Heat is removed from the ‘cold end’ of the heat pipes via fin stacks and one or more fans, blowing air between the fins. In this case, one fan is mounted in the middle of the two fin stacks. heat energy from the ‘hot end’ to the ‘cold end’. While a CCHP can work in any orientation, the maximum distance it can work against gravity is about 250mm for a copper/water heat pipe. In many cooling applications such as computer CPU coolers, fins are added to the heatsink, and possibly fans, to dissipate that heat (Fig.13). While some lower-end modern CPUs can be cooled with a standard finned heatsink and fans, that is not good enough at the high levels of heat generated by many modern CPUs, some of which can exceed 200W under full load. To allow transfer into and out of the heat pipe, sections of the tube can be flattened, as shown in Fig.13. These flattened sections can then be laid side-by-side and machined to form rectangular areas which make Fig.15: a Dynatron-brand R15 vapour chamber base with a copper stacked fin heatsink, recommended for use with certain CPUs in server applications. It is capable of dissipating 165W. Despite the relatively small source area (typically around 200mm2), the vapour chamber ensures an even distribution of heat across the heatsink. Source: Dynatron Corporation 24 Silicon Chip Fig.14: the structure of a vapour chamber. Note the support structure made from numerous solid copper pillars to resist the high clamping force. intimate contact with either the heat source (eg, the flat surface of a silicon chip) or the heat removal system (eg, a set of metal fins). As long as the sections are not flattened so much that they pinch off the inside of the pipe, this has little impact on their performance. Vapour chambers A vapour chamber can be thought of as a type of flattened and square CCHP (see Fig.14). Its purpose is to distribute heat uniformly, remove hot spots, and transfer high heat from a smaller area such as CPU or GPU to a larger heatsink such as the finned assembly. That finned assembly can then deal with the lower heat flux, as seen in Figs.15 & 16. A vapour chamber is constructed much the same as a heat pipe. But in addition to the capillary material lining the interior chamber, there may also be internal support posts to allow for the high clamping pressures involved. These are from the need to firmly attach the heatsink and vapour chamber to the device to be cooled, so that it has sufficient thermal conductivity to the vapour chamber. An advantage of a vapour chamber is that the cooling assembly can be larger and therefore quieter than a traditional heatsink, the latter of which may require very powerful and noisy fans to remove a high heat load. (Have you ever heard a modern computer server working? They sound like a plane about to take off!) Note that heat pipes used in coolers have a similar role; they spread the heat out to a much larger area than the source, allowing many more fins to conduct the heat into the air, and larger (and thus slower spinning and Fig.16: an illustration of vapour chamber arrangement as used on a reference Nvidia GTX580 graphics card. The function of the vapour chamber is to spread heat evenly to the finned heatsink. The condensed liquid is returned via a wick structure. “GPU” is the graphics processing unit chip. Australia's electronics magazine siliconchip.com.au Fig.17: a video frame showing a vapour chamber from Razor Phone 2 with the chamber cut open to reveal the wicking and support structure. From the video titled “Razer Phone 2 Teardown - The Vapor Chamber is Incredibly Cool” at https://youtu.be/UGsICbmmfws quieter) fans to assist in that transfer. The quietness of these designs is a particular advantage for computer gamers who want quiet machines that must run for long periods under heavy 3D graphics computational loads. Another advantage of vapour chambers is that they can be used in height-sensitive devices like phones and laptops as they can be made as thin as one millimetre, much thinner than a heat pipe in the same application (see Fig.17). In such applications, heat can be distributed and ‘diluted’ elsewhere in the device, or removed via a flat outside surface such as the back cover. In a sense, this means that rather than your phone or tablet CPU getting hot under load and throttling back its frequency, the whole phone/tablet instead becomes somewhat warm. That’s because the same amount of energy is spread over a wider area, lowering the temperature and improving thermal transfer to the surrounding air. Thermosiphons Thermosiphons can be thought of as wickless heat pipes (see Figs.18 & 19) and were the subject of the original invention of Perkins. While they do not have a wick, sometimes they have grooves on the pipe’s interior walls to increase the surface area and facilitate the return of the working fluid to the evaporator. Unlike CCHPs, they rely on gravity, not capillary action, for the return of the working fluid. Therefore, they can only be used with the heat moving siliconchip.com.au from a lower area to a higher location, since gravity can only return the condensate to a lower area. So why use thermosiphons instead of CCHPs that can be used in any orientation? The advantage of thermosiphons is that they have about three times the heat transfer capacity for the same pipe diameter. They can also transfer heat over distances of tens of metres. Since thermosiphons will remove heat from the bottom of the pipe to the top, but won’t transfer heat from top to bottom, they can be thought of as analogous to a diode. This type of thermosiphon should not be confused with the natural convention and circulation of water without a pump that occurs in some solar hot water systems or older internal combustion engines. While those are classified as thermosiphons, they are not heat pipes. One variation is the loop thermosiphon, where the liquid return and vapour paths are separated. This has the advantage of removing any restriction caused by the liquid and vapour flowing in the same pipe in different directions. Fig.18: the operation of a thermosiphon heat pipe. This one is embedded in the ground and is designed to prevent the permafrost from melting around buildings in cold climates like Alaska or northern Canada. The thermosiphon can also be designed to support structures. Original source: www.researchgate. net/publication/266672789_Review_ of_Thermosyphon_Applications Thermosiphons in building construction While not a problem in Australia or New Zealand, there is permanently frozen ground known as permafrost in the far north of North America, Europe, and Russia. Any attempt to build on permafrost will result in heat from the building causing the permafrost to thaw, thus Australia's electronics magazine Fig.19: a heat pipe loop thermosiphon. Source: Celsia, Inc May 2022  25 Fig.20: thermosiphon support structures hold up the Trans-Alaska Pipeline System (TAPS). Without them, heat from the pipeline would cause the permafrost to melt, and the pipeline supports would sink into the ground. Note the finned condensers. These heat pipes use ammonia as the working fluid and steel for the pipes. Source: Dave Bezaire & Susi HavensBezaire (CC BY-SA 2.0) destabilising the foundations of the structure. The solution is to either drive piles deeply into the ground and build on top of those, build on a thick gravel pad, or use heat pipe technology to keep the ground frozen, as shown in Figs.20 & 21. In cases where the ground has thawed, it may be re-frozen and kept frozen using a variation of a thermosiphon called a thermoprobe, such as from Arctic Foundations of Canada (http://arcticfoundations.ca/). How good are heat pipes? Excellent passive heat conductors such as pure copper, aluminium, graphite, and diamond have a thermal conductivity between 250W/m.K and 1500W/m.K. In comparison, heat pipes have a thermal conductivity in the range of 5000W/m.K to 200,000W/m.K. So they range from being around three times better heat conductors to being 800 times better than solid metal! Variable conductance heat pipe (VCHP) Fig.21: a diagram showing how the Trans-Alaska Pipeline System thermosiphons shown in Fig.20 are made. 26 Silicon Chip Australia's electronics magazine Constant conductance heat pipes are linear devices in which the temperature at the evaporator end (the source of heat where evaporation occurs) drops proportionally to the difference in temperature between the evaporator end and the condenser end. Situations where the heat source is not generating much heat and/or the condenser ambient temperature is low can result in the device being excessively cooled. A variable conductance heat pipe can prevent that. In a variable conductance heat pipe, the device being cooled is, by design, kept at a relatively constant temperature even when heat dissipation from the device changes or the ambient temperature of the condenser end changes (see Figs.22 & 23). This is done by adding a non-condensable gas (NCG) to the heat pipe, in addition to the working fluid. A gas reservoir is also added at the condensing end of the heat pipe (the end remote from the heat source). When there is significant heat to be moved and the ambient temperature is not too low, the working fluid vapour pressure pushes the NCG back into the reservoir. The heat pipe then works in the usual manner, as shown at the top of Fig.22. siliconchip.com.au But when the dissipation from the device being cooled is low and/or the ambient temperature is low, meaning the device could be excessively cooled, the working fluid has a lower pressure and cannot push back the NCG as much. As shown at the bottom of Fig.22, less condensing area is exposed, and therefore, the device is not cooled as much and stays at an appropriate temperature. A VCHP can maintain the temperature of the evaporator end to within 1-2°C of the desired temperature. This is despite significant variations in the heat being dissipated by the device at the evaporator end and the ambient temperature at the condenser end. Loop heat pipes The loop heat pipe is based on the CCHP and is like a loop thermosiphon. But unlike a thermosiphon, it does not rely on gravity. Loop heat pipes can transfer more heat over longer distances than CCHPs can. They can be used in conjunction with CCHPs and VCHPs. Applications include spacecraft, avionics cooling in aircraft and aircraft de-icing – see Fig.24. Rotating heat pipes Fig.23: a variable conductance heat pipe from a spacecraft. The bulbous structure is the gas reservoir, and the distant end is the evaporator. The condenser portion is the long flange. The valve and pressure gauge are removed when the device is put into service. Source: Advanced Cooling Technologies, Inc (CC BY-SA 3.0) Fig.24: a commercial loop heat pipe system for NASA spacecraft designed by Advanced Cooling Technologies. The titanium/water heat pipes operate from 70°C to 250°C. Spacecraft heat pipes can have multiple evaporators and condensers. Source: Advanced Cooling Technologies ► Fig.25: rotating heat pipes work similarly to other heat pipes, but they use centripetal/centrifugal forces along with a tapered profile to return the working fluid after it has condensed. ► A rotating heat pipe (Fig.25) is designed to cool rotating machinery such as motors or RF rotary joints, as used in telecommunications. They work much like a CCHP, but they rely on centrifugal forces instead of relying on capillary action for the condensate return. They do this either via a tapered wall with a smaller diameter at the condenser end or by having spiral grooves similar to a rifle barrel to convey the condensate back to the evaporator. Heat can only flow in one direction in a rotating heat pipe, so it is again analogous to a diode. Fig.22: how a variable conductance heat pipe (VCHP) works. The top diagram shows its operation under optimal conditions, while at the bottom, it has reduced heat dissipation at the evaporator end (where the device being cooled is located) due to less heat being produced. This is because non-condensable gas migrates down the tube, blocking some of the condenser area and reducing its capacity. Oscillating and pulsating heat pipes Oscillating or pulsating heat pipes (OHP), are relatively new members of the heat pipe family, having been invented in the 1990s. They comprise a continuous loop of pipe or pipe-like shape laid out in a serpentine manner, containing alternating pockets of liquid ‘slugs’ and vapour bubbles which move back and forth in relation to the condenser area as they are alternatively heated or cooled – see Fig.26. siliconchip.com.au Australia's electronics magazine May 2022  27 They are often machined into a bottom plate, and a smooth top plate is bonded to that, with the item to be cooled attached to the top plate. In Fig.26, the OHP is said to be bonded to a battery pack but it could be just about anything that generates heat. A video of how an oscillating heat pipe works can be seen, titled “Pulsating Heat Pipe (PHP)/Oscillating heat pipe (OHP) -CFD analysis | Animation” at https://youtu.be/ glYguHLKRL0 Direct liquid cooling of ICs Fig.26: an oscillating heat pipe for cooling an electric vehicle battery. Original source: www.mdpi.com/1996-1073/11/3/655 Fig.27: a silicon chip with an onboard microfluidic cooling system, developed by the Swiss Federal Institute of Technology in Lausanne. The fluid inlet and outlet can be seen at the top of the device. 28 Silicon Chip Australia's electronics magazine All the above-mentioned types of heat pipes can be used to cool electronics or other devices. But a heat pipe can only ever contact the exterior of a chip or electronic device package and often requires a thermal interface material to achieve sufficient thermal conductivity between the two. That material always has some sort of thermal resistance, though. Another way to cool silicon chips that does not involve heat pipes, currently under development, is to build liquid cooling channels into the chip itself (see Fig.27). This technology is under development at the Swiss Federal Institute of Technology in Lausanne under the leadership of Professor Elison Matioli. In this case, liquid-carrying microchannels are fabricated in the silicon substrate. The size of the channels vary according to the cooling required in a particular area of the device. The channel size varies because if they were all of the same small size, a large amount of energy would be required to pump the fluid. So, like a human circulatory system to which the cooling channels have been likened, the channels are only narrow in the areas where the cooling is needed most. Cooling channels of the small size involved come under the general area of microfluidics, which we covered in the Silicon Chip article on Fluidics in the August 2019 issue (siliconchip. com.au/Article/11762). This type of system has been shown to be capable of removing 1700W/cm2 with the chip temperature limited to 60°C. That’s about ten times more effective than external liquid cooling or cooling using heat pipes. The work is significant because, until now, semiconductor device fabrication and cooling have been siliconchip.com.au considered two separate areas of design. This approach integrates the two areas. Ice cream scoops One application you might not have considered for heat pipes is ice cream scoops! Heat is transmitted from the hand via a heat pipe in the handle to the scoop, where it melts the ice cream, making it easier to scoop out (Fig.28). You can view the US Patent for this vital technology at siliconchip.com. au/link/abda Related videos ● “What’s Inside the Worlds’ Fastest Heat Conductor?” – https://youtu.be/ OR8u_ _Hcb3k ● “Liquid Crystals Painted on Heat Pipes” – https://youtu.be/Y6K7h9tbD_s ● “Heat Pipe Basics and Demonstration Video” – https://youtu. be/2vk5B6Gga10 ● “How Copper Heatpipes Are Made | China Factory Tour (Cooler Master)” – https://youtu.be/AD-4WKwCAfE Fig.29: heat pipes (labelled) as used on a NASA Kilopower experimental reactor proposed, for use in space, on the Moon and on Mars. Source: NASA Heat pipe limits Limitations are imposed on the operation of heat pipes by several factors. These include: 1) the capillary limit, where capillary action in the returning liquid is not fast enough to support the evaporation rate in the opposite direction 2) the entrainment limit whereby the velocity of the vapour near the wick is enough to restrict the return flow of the liquid 3) the sonic limit, where the vapour cannot exceed the speed of sound at the pressure inside the heat pipe whereby a shockwave may be created 4) excessive heat, causing the liquid in the wick to evaporate Conclusion Heat pipes are a vital technology for today’s high-density semiconductors. They allow waste heat to be removed to a sufficiently large fin stack for the semiconductor device to remain at a reasonable operating temperature, without the additional complexity, cost or size of a standard liquid-­cooling system. With the density of digital semiconductors continuing to increase, and greater demand for high-efficiency power semiconductors in renewable energy systems and electric vehicles, they have become an essential part of SC modern technology. Fig.28: a Thermoworks ice cream scoop that uses a heat pipe to assist in scooping the ice cream. It appears to be no longer manufactured. siliconchip.com.au Australia's electronics magazine May 2022  29