Patent Application
20190014688
The present invention generally relates to heat dissipation for electronics applications. The invention particularly relates to vapor chamber heat spreaders having dual layer evaporators for improved heat dissipation over relatively large heat input areas.
The intensifying electrification of transportation systems and clean energy production technologies has dramatically increased the waste heat load that needs to be dissipated from high-density power electronic devices. This trend has pushed conventional air-cooling thermal management architectures to the limit; a reliance on conduction heat spreading from devices to the heat rejection surfaces incurs an overly large thermal resistance at power levels well below the inherent electrical power density limits of current and next generation devices. The integration of active liquid cooling architectures, such as high-performance single- and two-phase jet impingement or microchannel heat sinks, can alleviate these thermal limitations by eliminating the need for a direct low thermal resistance conduction pathway to the ultimate heat rejection surfaces. However, these systems require dedicated auxiliary components (for example, compressors/pumps, fluid connects, filters, etc.) which may be weak points for overall system reliability. In comparison, air cooling remains a proven and reliable approach for which heat transfer surfaces can be readily and accurately implemented.
Development of heat spreading technologies in next-generation systems are desired to enable air cooling of hotspots, that is, heat input areas with a relatively high temperature in comparison to its surroundings, for example, up to one kW/cm2 and possibly more. Although solid heat-conduction spreaders are fundamentally limited to a linearly decreasing performance (increasing thermal resistance) with effective heat transfer distance, vapor chamber heat spreaders, also referred to herein as vapor chambers, implemented as a base of a solid heat-conduction spreader, such as a heat sink, offers promising solutions.
While vapor chamber heat spreaders have been used extensively in electronics cooling applications, they have typically been utilized to transport heat over relatively long distances from low-heat-flux sources, where evaporation from a wick structure occurs in a predictable manner, that is, the total heat input is limited by capillary resupply to the evaporator at a relatively low surface superheat. The investigation of vapor chambers for high-heat-flux spreading applications has resulted in a paradigm shift where capillary limits are not encountered until a large surface superheat has been reached, and boiling occurs in the wick structure. The stochastic network of interconnected pores in sintered wick structures can continue to feed liquid to the heat source during boiling, and sustain a so-called “capillary-fed boiling” regime.
Capillary-fed boiling was first anecdotally described during measurements of thermal resistance across evaporator wick structures. Boiling was either directly observed or inferred from a sharp reduction in the thermal resistance upon incipience. While the single-phase capillary limit and boiling limit were treated as distinct phenomena before these studies, it was eventually determined that these limits are fundamentally interrelated and that boiling in the sintered wick structures incurred a large pressure drop that induced a premature capillary limit. Subsequent investigations specifically investigated capillary-fed boiling from high-heat-flux hotspots in simulated vapor chamber environments using sintered screen meshes, sintered powders, biporous sintered powders, and microposts. These investigations demonstrated that sintered wick structures could dissipate unprecedented heat fluxes of over 0.5 kW/cm2 from small hotspots (0.04-0.25 cm2) if operated in a capillary-fed boiling regime. The inherent similarities between the vapor-formation regimes allowed generalized empirical correlation of the heat transfer coefficient by assuming that pseudo-vapor columns exist within the pore spaces during capillary-fed boiling. Parametric investigation of the wick thickness, porosity, and pore sizes revealed that improved thermal performance was obtained for morphologies that eased vapor removal from the wick structure and increased the area for interstitial phase change. Homogeneous wick structures are inherently restricted by a tradeoffbetween these characteristics. Therefore, multi-scale heterogeneous wicks have since been designed to either separate the liquid feeding and vapor extraction flow paths or distribute liquid to a thin, low-permeability layer. By providing vapor a high-permeability path to preferentially exit the wick structure, a reduced vapor-phase pressure drop lowers the surface superheat. A thin wick layer imposes an extremely low thermal resistance but has not been capable of sustaining liquid resupply. Thick sintered-particle arteries have demonstrated a four-fold increase in the maximum supported heat flux.
While heat fluxes of up to nearly one kW/cm2 have been noted in the art under capillary-fed boiling conditions, there is a strong dependency of the maximum supported heat flux on the size of the heat input area. For example,
Thus, there is an ongoing desire for vapor chamber heat spreaders for applications with power densities of one kW/cm2 over relatively large heat input areas, especially in circumstances where there may be a relatively low temperature differential (gradient) across the vapor chamber heat spreader, for example, temperature differentials of about 40 K and less.
The present invention provides vapor chamber heat spreaders suitable for applications with power densities of one kW/cm2 or greater over heat input areas of at least 0.25 cm2 or more, and preferably one cm2 or greater.
According to one aspect of the invention, a vapor chamber heat spreader is provided that includes a housing having a thermally conductive substrate, a working fluid contained within the housing, a base layer formed of a porous thermally conductive material and located on and in thermal contact with the substrate, a cap layer formed of a porous thermally conductive material having through-holes formed therein defining vapor vents, and a plurality of conduits connecting the cap layer and the base layer with interstitial gaps between the conduits. The conduits are functionally operable to convey the working fluid from the cap layer to the base layer. The working fluid is capable of being evaporated by heat entering the base layer, condensing, and then flowing through the conduits to replenish the base layer.
According to another aspect of the invention, a method of making a vapor chamber heat spreader is provided that includes providing a thermally conductive substrate, depositing a thermally conductive powder onto the substrate, sintering the powder to form a porous thermally conductive material on the substrate, processing the material to form a base layer on the substrate and a plurality of conduits extending from the base layer with interstitial gaps therebetween, attaching a cap layer formed of a porous thermally conductive material to ends of the conduits oppositely disposed the base layer wherein the base layer, the conduits, and the cap layer define an evaporator, and sealing the evaporator and a working fluid in a housing, and the n sealing the evaporator and a working fluid in a housing. The cap layer has through-holes formed therein defining vapor vents that expose the interstitial gaps to a cavity on a side of the cap layer oppositely disposed from the conduits. The conduits are functionally operable to convey the working fluid from the cap layer to the base layer via capillary action. The working fluid is capable of being evaporated by heat entering the base layer, condensing, and then flowing through the conduits to replenish the base layer.
Technical aspects of the vapor chamber heat spreader and method described above preferably include the ability to provide reliable capillary-fed boiling at the highest temperature hotspot within the vapor chamber heat spreader for applications with power densities of one kW/cm2 or greater over a heat input area of at least 0.25 cm2, and preferably one cm2 or greater.
Other aspects and advantages of this invention will be further appreciated from the following detailed description.
The vapor chamber 30 will be described herein as utilizing water as a working fluid since the thermophysical properties of water are believed to be suitable for operating temperatures ranging from room temperature (about 20° C.) to about 200° C., which is an expected operating temperature range of current electronics devices. However, it is within the scope of the invention that the working fluid may a fluid other than water.
The evaporator 36 includes a base layer 42, an array of posts 44, and a cap layer 46. Interstitial gaps 45 are located around the posts 44 and between the base layer 42 and the cap layer 46. During use, heat from the heat input device 31 is conducted through the evaporator-side wall 32 to yield a heat input area (hotspot) in the base layer 42, generally corresponding to the surface area of the evaporator-side wall 32 contacted by the device 31. As the temperature increases, working fluid in the form of a liquid 24 in the base layer 42 may boil and evaporate as vapor 20. The vapor 20 rises and enters the vapor core 40 through vapor vents 48 present in the cap layer 46. As the vapor 20 flows to cooler areas within the vapor chamber 30, it condenses to reform the liquid 24 and enters the porous cap layer 46. Condensation and transport of the liquid 24 to the cap layer 46 may be promoted by the condenser 38. Due to capillary action, the liquid 24 is then transported through the cap layer 46, to and through the posts 44, and back to the base layer 42, preferably at a rate sufficient to maintain capillary-fed boiling at the hotspot in the base layer 42. The posts 44 may be referred to as “conduits” in light of their function of conducting or transporting the working fluid from the cap layer 46 to the base layer 42. The direction of flow of the working fluid, in both the vapor state 20 and liquid states (20 and 22), is represented in
The base layer 42 is preferably an ultra-thin, powder wick structure having a thickness on the order of about 200 micrometers, for example, about 25 to about 500 micrometers and more preferably about 50 to about 250 micrometers. The base layer 42 preferably provides a relatively low thermal resistance (about 0.03 Kcm2/W or below) under capillary-fed boiling (relatively high heat transfer coefficient). Preferably, the base layer 42 is sufficiently thin to promote reduction of the surface superheat temperature, for example, about 40 K or less.
Generally, relatively thin sintered powder base layers produce extremely low maximum supported heat flux due to the confined path of capillary-fed liquid supply. To address this limitation, the vertical liquid-feeding posts 44 are configured to distribute a capillary-fed supply of the liquid 24 across the hotspot from the cap layer 46 to prevent dryout within the base layer 42 at high-heat-flux operation. Although the posts 44 are represented as being vertical cylinders, it is within the scope of the invention that the posts 44 may be other shapes and have other orientations.
Alternative strategies for lateral liquid feeding within a vapor chamber, for example, with wedge-shaped or arterial wick structures, inherently reduce the available boiling surface area of the thin base layer 42 due to the need to provide an adequate cross sectional wicking area within a confined height. In the vapor chamber 30 represented in
In addition to capturing and transporting liquid condensed from vapor, the cap layer 46 may provide a mechanism for proximate capture of droplet spray during intense capillary-fed boiling. This is beneficial as it has been observed that the maximum heat flux may be reduced due to liquid loss by this mechanism. It should be noted that while the proposed configuration may initially appear similar to other hybrid wicking structures that have a wick cap layer over the top of microposts, these designs have significantly different compositions and constructions, with the layers of the wick having significantly different forms and functions, wherein the interstitial gaps between the microposts are to remain filled with liquid and provide the primary high-permeability capillary pathways to a top evaporating cap layer. Such a cap layer does not provide any liquid feeding function, but is a very thin woven mesh attached to the top of the microposts to enhance the evaporation rate, and such microposts are solid and do not assist in capillary liquid supply. Consequently, the form, function, and intended operation of prior hybrid wicking structures drastically differ from the two-layer evaporator 36 proposed herein. An example of a prior hybrid wicking structure is reported in Oshman et al., “The Development of Polymer-Based Flat Heat Pipes,” Journal of Microelectromechanical Systems, Vol. 20, No. 2, pp. 410-417 (April 2011).
Despite the low thermal resistance of the base layer 42, any gradient in the saturated vapor pressure exiting the evaporator 36 can lead to a significant surface temperature rise. Therefore, the vapor vents 48 are incorporated into the cap layer 46 to provide a high-permeability pathway for vapor extraction from the interstitial gaps 45 between the posts 44. These gaps 45 will initially be flooded with liquid, but the menisci (that is, the upper surface of the liquid) will immediately recede upon marginal power input due to the negligible capillary pressure provided by the large-scale pores, leading to a regime of operation where liquid feeding and vapor venting pathways are completely separated, and pressure drop in the vapor phase in minimized.
The liquid-feeding posts 44 and the vapor vents 48 may have various sizes, shapes, and dimensions.
The vapor chamber 30 and its components may be formed of various materials. The housing of the vapor chamber 30 is preferably a highly thermally conductive metal or alloy, such as but not limited to copper, which is suitable for conducting heat from the heat input device 31. The evaporator 32 and the condenser 38 are porous wick structures formed of hydrophilic (if water is the working fluid) and preferably conductive materials, a nonlimiting example being a sintered copper powder, that preferably comprising superhydrophilic nanostructures. As used herein, the wetting of a liquid with a surface of a solid material will be described in relation to a contact angle at which the liquid-vapor interface meets the solid-liquid interface. A wettable surface (for example, hydrophilic if water is the working fluid) is any surface with a contact angle of less than 90° (low contact angle) which indicates that wetting of the surface is very favorable, and a liquid will likely spread over the surface and in the case of a porous material, may spread into the material. A nonwettable surface (for example, hydrophobic if water is the working fluid) is any surface with a contact angle of greater than 90° (high contact angle) which indicates that wetting of the surface is unfavorable, so a liquid will likely minimize contact with the surface and form a compact liquid droplet on the surface. A superphilic surface (for example, superhydrophilic if water is the working fluid) refers to a surface on which a liquid will uniformly spread such that it forms a thin conformal liquid layer rather than a droplet with a measurable contact angle. Therefore, the above-mentioned superhydrophilic nanostructures are structures that either have superhydrophilic surfaces, or in combination form a superhydrophilic surface.
While copper is intrinsically hydrophilic, careful control of the environmental conditions (viz., a vacuum or reducing atmosphere) is preferred in vapor chamber fabrication and processing steps in order to maintain repeatable wetting characteristics. To alleviate these process control restrictions, the wick structures can be functionalized to maintain or enhance their intrinsic wettability. While a variety of approaches have been proposed that impart or promote wettability by producing or processing the wick structure to have textured or nanostructured surfaces (i.e., comprising surfaces with nanostructure features), these techniques often require line-of-sight processes for catalyst deposition or nanowire growth. Alternatively, thermal or solution-immersion chemical processes can be used to grow conformal layers of stable copper oxide. Thermally grown Cu2O can provide controlled hydrophilicity, but the relatively smooth layer does not provide geometric enhancement of wettability. CuO nanostructures grown chemically in alkali solutions can yield excellent superhydrophilic wetting behavior. Superhydrophilic copper oxide nanostructures grown chemically on the surface renders the sintered copper surface superhydrophilic and the thin oxide layer has a negligible thermal resistance.
A feasibility analysis was performed to demonstrate that the vapor chamber 30 is capable of yielding at least about one kW/cm2 heat dissipation over at least about one cm2 area for which the two-layer evaporator 36 can sustain capillary liquid supply to a hotspot, and the temperature of the evaporator 36 remains below a target value, for example, a temperature rise of 40 K or less. The schematic diagram shown in
In order to study the performance of the proposed two-layer evaporator 36, a reduced-order model was developed. The model included a two-layer evaporator containing a square array of the liquid feeding posts 44 over a heat input area (hotspot). The model assumed that the liquid-phase pressure drop returning from peripheral regions was negligible due to the bulk wick structure thickness in this region and the presence of liquid return posts (not shown) throughout the vapor core 40 (which are typically included in vapor chambers for mechanical support). The pressure drop in the cap layer (ΔP2) was calculated assuming axisymmetric inward flow in a representative disk of porous media having the same footprint area as the evaporator 36, with uniform mass loss to the liquid-feeding posts 44. The effective thickness of this disk was reduced based on the percentage area of vapor vents 48. Pressure drop across the liquid-feeding posts (ΔP3) was calculated based on uniform one-dimensional flow through an equivalent cross-sectional area. There are no available methods for mechanistic prediction of the two-phase pressure drop in the base layer 42 (ΔP4), therefore this pressure drop was estimated by extrapolating previous empirical data for which the pressure gradient was correlated to the superficial vapor velocity exiting the wick structure at a given thickness of 200 micrometers.
Thermal resistances for the model were calculated assuming uniform heat rejection through the surface of the condenser 38. Conduction resistances (Rcond,1+2) included the vapor chamber walls and saturated condenser 38. While semi-empirical models can be used to correlate capillary-fed boiling thermal resistances (Rboii), the model instead assumed a fixed value of 0.03 Kcm2/W based on experimental observations for a 200 micrometer-thick wick structure. The pressure drop (ΔPvap,1) across the vapor vents 48 was calculated assuming Hagen-Poiseuille flow (neglecting minor losses). The pressure drop in the vapor core 40 (ΔPvap,2) was calculated assuming axisymmetric viscous outward flow in an equivalent disc-shaped region. The associated vapor saturation temperature gradients (Rvap,1+2) were calculated using the Clausius-Clapeyron relation.
A baseline case was chosen to represent the performance targets (Q=1 kW, ttotal=3 mm, A1=1 cm2, A2=100 cm2). All thermophysical fluid properties were evaluated at a saturation temperature of 100° C. All wick regions were assumed to be composed of 45 to 75 micrometer-diameter sintered copper powder due to the availability of property data (keff=56 W/mK, K=2.5×10−11 m2, Pcap=4σ/Dp where Dp=74 micrometers) obtained via microtomography based simulations of actual structures. Several geometric parameters were chosen based on convention (twick,condenser=twall=200 micrometers) and the availability of empirical data (twick,base=200 micrometers).
A thickness of 1 mm was used for the vapor core 40, a bulk thickness of 1.4 mm was used for the evaporator 36, and a thickness of 1 mm was used for the cap layer 46 (i.e., 400 micrometer-tall liquid-feed posts 44). These dimensions were chosen based on an evaluation of the model for various possible geometric configurations and estimated manufacturing constraints. A parametric evaluation of the liquid-feeding post array was performed assuming that the post spacing (S) and vapor vent diameter (Dvent) were equal. The results of this evaluation are represented for various post diameters (Dpost) at a constant spacing of S=1 mm in
To attach the cap layer 46, the exposed recesses in the wick layer 110 can be backfilled with a temporary filler material 114 (image d), such as but not limited to carbonate particles, before applying a second layer of loose copper powder thereover 116 (image e). During a lost-carbonate sintering (LCS) process, the carbonate particles decompose at the elevated sintering temperatures, leaving behind the features formed by the sintered copper particles to form the cap layer 46 (image f). Vapor vents 48 can then be formed in the cap layer 46 (image g).
Optionally, surface modifiers, such as copper oxide nanostructures, can be coated onto the as-fabricated evaporator 36 (image g) to impart wetting hydrophilic, and preferably superhydrophilic, behavior. This can be achieved with a chemical solution-immersion process that subjects the samples to an alkali solution.
For the material removal steps of images (c) and (g) of
An alternative ablation method is electrical discharge machining (EDM), a common industrial process where the electrode tool is held at a voltage differential from the substrate to be machined and dielectric breakdown causes material to be ablated from the substrate and electrode. Other conventional machining processes can also be explored. However, it is preferred that the properties of the sintered coating are at least partially if not fully maintained after material removal is complete.
The feasibility calculations suggested that millimeter-scale features can achieve desirable performance targets of 1 kW/cm2 heat flux over a 1 cm2 surface area with a temperature rise of 40 K. For images (d) through (f) of
Conventional fabrication methods (e.g., using a mold to sinter particles onto solid posts) impose fabrication constraints that limit feature size and reduces the percentage of cross-sectional area available for liquid feeding, ultimately negating any performance improvement over lateral feeding approaches. The fabrication process described herein allows for improved flexibility of the liquid-feeding post height and diameter to yield improved performance.
In view of the above, the evaporator 36 is a scalable, two-layer structure that provides low thermal resistance operation while maintaining adequate passive liquid feeding by capillary action over relatively large heat input areas and relatively high power densities, as nonlimiting examples, heat input areas of 1 cm2 or greater and power densities of 1 kW/cm2 or greater.
While the invention has been described in terms of specific or particular embodiments, it should be apparent that alternatives could be adopted by one skilled in the art. For example, the vapor chamber 30 and its components could differ in appearance and construction from the embodiments described herein and shown in the drawings, functions of certain components of the vapor chamber 30 could be performed by components of different construction but capable of a similar (though not necessarily equivalent) function, and appropriate materials could be substituted for those noted. In addition, the invention encompasses additional or alternative embodiments in which one or more features or aspects of different disclosed embodiments may be combined. Accordingly, it should be understood that the invention is not necessarily limited to any embodiment described herein or illustrated in the drawings. It should also be understood that the phraseology and terminology employed above are for the purpose of describing the disclosed embodiments and investigations, and do not necessarily serve as limitations to the scope of the invention. Therefore, the scope of the invention is to be limited only by the following claims.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/US2016/065824 | 12/9/2016 | WO | 00 |
| Number | Date | Country | |
|---|---|---|---|
| 62266330 | Dec 2015 | US |
