The present invention generally relates to thermal management processes and equipment. The invention particularly relates to vapor chamber devices capable of spreading heat, including heat from sources that generate large total heat loads and/or high-power-density hotspots.
Vapor chamber devices have long been used in thermal management applications. Such devices passively spread heat from a localized input area to a relatively larger output area, where the heat can then be rejected, as nonlimiting examples, to a heat sink or cold plate. As a nonlimiting example, when applied to thermal management of electronic devices, vapor chamber devices can be subjected to non-uniform power maps that include hotspots (discrete areas with high heat fluxes). Vapor chamber devices typically comprise a vacuum-sealed envelope or housing having an internal cavity therein that stores a working fluid (e.g., water). The cavity is lined on its interior by porous wick structures that define what is referred to as a vapor core that generates a capillary pressure to recirculate the working fluid between cooler and warmer areas of the housing (e.g., an evaporator side and a condenser side).
As a nonlimiting example, use of such vapor chamber devices may include positioning the device such that a bottom surface thereof thermally contacts a heat source and a top surface thereof thermally contacts a heat sink. As the heat is conducted from the heat source to the vapor chamber device, some of the working fluid vaporizes and travels to cooler areas such as internal surfaces of the top surface. The heat sink absorbs heat at the top surface causing the vaporized working fluid to condense and return to liquid form. This liquid working fluid is then reabsorbed by the wick structures and distributed through capillary action to warmer areas such as internal surfaces of the bottom surface where the cycle repeats. As used herein, a portion of the porous wick structure adjacent the heat input area where the working fluid vaporizes due to heat absorption is referred to as an evaporator wick, and a portion of the porous wick structure adjacent the heat output area where the working fluid condenses due to heat rejection is referred to as a condenser wick.
In a conventional vapor chamber device having a single vapor core, there is a design tradeoff between increased power handling (via increased liquid feeding) and reduced conduction thermal resistance. Conventional vapor chamber devices having a single vapor core require a relatively thick evaporator wick to avoid a capillary limit at high total heat loads. Even though high heat flux hotspots may contribute only a small portion of the total power, as the thickness of the evaporator wick is increased to handle higher total heat loads, a dramatic increase in the temperature of the hotspots may occur due to an increased conduction resistance across the evaporator wick.
Recent investigations reported in the literature have focused on the structure of the evaporator wick within a vapor chamber device to address the removal of high heat fluxes over differing heat input areas. Several of these studies focus on the development of hybrid or micro-patterned wick structures that aim to maintain low surface temperatures and adequately feed liquid over relatively large areas of the evaporator wick having uniform heat fluxes. However, previous evaporator wicks and vapor chamber designs have not been developed or demonstrated with the goal of simultaneously managing both high total powers (over large areas) and randomly located hotspots, despite the importance and prevalence of this power map.
The present invention provides vapor chamber devices suitable for use in thermal management applications and methods for their use.
According to one aspect of the invention, a vapor chamber device having a heat input side and a heat output side includes a first vapor core configured to passively spread heat from a localized first input area to a relatively larger first output area adjacent to and in thermal contact with the heat output side of the vapor chamber device, and a second vapor core configured to passively spread heat from a localized second input area adjacent to and in thermal contact with the heat input side of the vapor chamber device to a relatively larger second output area in thermal contact with the first input area of the first vapor core. The first vapor core and the second vapor core are sealed from each other and hydraulically independent. The second vapor core is configured to attenuate high heat flux hotspots on the first input area before the heat fluxes thereof pass through the second output area of the second vapor core to the first input area of the first vapor core.
According to another aspect of the invention, a vapor chamber device includes a cascaded multi-core unit having a top-tier subunit comprising a first sealed cavity lined on an interior thereof by a first porous wick structure that defines a single first vapor core configured to generate a capillary pressure to recirculate a first working fluid therein and thereby passively spread heat from a localized first input area to a relatively larger first output area of the top-tier subunit, and a bottom-tier subunit comprising a second sealed cavity lined on an interior thereof by a second porous wick structure that defines at least a second vapor core configured to generate a capillary pressure to recirculate a second working fluid therein and thereby passively spread heat from a localized second input area to a relatively larger second output area of the bottom-tier subunit. The second output area of the bottom-tier subunit is thermally coupled to the first input area of the top-tier subunit. The second vapor core is configured to attenuate high heat flux hotspots before the heat fluxes thereof pass through the second output area of the bottom-tier subunit to the first input area of the top-tier subunit.
According to yet another aspect of the invention, a method is provided for dissipating heat from a surface with a vapor chamber device having a cascaded multi-core unit comprising a top-tier subunit comprising a first sealed cavity lined on an interior thereof by a first porous wick structure that defines a single first vapor core containing a first working fluid therein, and a bottom-tier subunit comprising a second sealed cavity lined on an interior thereof by a second porous wick structure that defines at least a second vapor core containing a second working fluid therein. The method includes locating a heat input side of the vapor chamber device onto the surface, conducting heat to a second input area of the bottom-tier subunit and thereby vaporizing the second working fluid within the second vapor core and rejecting heat from a second output area of the bottom-tier subunit that is relatively larger than the second input area and thereby condensing the second working fluid within the second vapor core, generating a capillary pressure within the second vapor core to recirculate the second working fluid therein and thereby passively spread heat from the second input area to the second output area of the bottom-tier subunit, conducting heat from the second output area of the bottom-tier subunit to a first input area of the top-tier subunit and thereby vaporizing the first working fluid within the first vapor core and rejecting heat from a first output area of the top-tier subunit that is relatively larger than the first input area and thereby condensing the first working fluid within the first vapor core, generating a capillary pressure within the first vapor core to recirculate the first working fluid therein and thereby passively spread heat from the first input area to the first output area of the top-tier subunit, and attenuating high heat flux hotspots on the surface with the second vapor core before the heat fluxes thereof pass through the second output area of the bottom-tier subunit to the first input area of the top-tier subunit.
Technical aspects of the vapor chamber devices and the method described above preferably include the capability of decoupling the spreading of total background power from that of individual hotspots using a single top-tier subunit for bulk heat spreading and a bottom-tier subunit for damping of high heat fluxes. The vapor chamber devices offer significant potential enhancement in performance, possibly on the order of a magnitude lower in thermal resistance, compared to conventional solid copper heat spreaders and conventional single-core vapor chamber devices owing to a reduction in the conduction resistance across the internal wick structure.
Other aspects and advantages of this invention will be appreciated from the following detailed description.
Existing vapor chamber device designs generally lack the capability of efficiently managing power maps in which both high total powers (over large areas) and small hotpots appear simultaneously. To address, disclosed herein are vapor chamber devices capable of spreading heat from sources that may generate large total heat loads and/or high-power-density hotspots. According to certain nonlimiting aspects of the invention, such vapor chamber devices include a cascaded multi-core unit that includes at least two tiers of cascaded vapor cores, with a top-tier subunit having a vapor core for bulk heat spreading and a bottom-tier subunit with multiple vapor cores for damping of local hotspots that may be generated anywhere over a footprint area thereof.
Coverage of the heat source with the secondary vapor cores 22 promotes the likelihood that a hotspot formed in any area contacting the second input area 30 will be spread out by the corresponding secondary vapor core 22 located above the hotspot. The small size of each of the secondary vapor cores 22 appreciably reduces a pressure drop of the recirculating second working liquid by minimizing a flow length from condenser wicks to evaporator wicks in each of the secondary vapor cores 22. Consequently, each of the secondary vapor cores 22 in the bottom-tier subunit 14 can sustain operation at the same capillary-limited heat load as the primary vapor core 20 of the top-tier subunit 12, but with significantly thinner evaporator wicks. The bottom-tier subunit 14 thus attenuates the high hotspot fluxes while imposing a small conduction resistance across the thin evaporator wicks before heat is transferred into the top-tier subunit 12, which requires thicker evaporator wicks to manage the total heat load within the capillary limit.
The array of secondary vapor cores 22 are represented in
The evaporator wick of the second wick structure 26 of the bottom-tier subunit 14 is preferably thinner than the evaporator wick of the first wick structure 24 of the top-tier subunit 12. That is, the evaporator wick of the second wick structure 26 preferably has a dimension in a direction between the second input area 30 and the second output area 34 that is less than a dimension of the evaporator wick of the first wick structure 24 in a direction from the first input area 28 to the first output area 32. The second wick structure 26 of the bottom-tier subunit 14 preferably has a particle diameter that is lower than the particle diameter of the first wick structure 24 of the top-tier subunit 12. The second wick structure 26 of the bottom-tier subunit 14 also preferably has a path of return for the second working fluid from the second output area 24 to the second input area 30 (e.g., from the condenser wick to the evaporator wick) that is smaller than a path of return for the first working fluid of the first wick structure 24 of the top-tier subunit 12 from the first output area 32 to the first input area 28.
The vapor chamber device 10 and its components may be formed of various materials, such as but not limited to those known in the art for conventional vapor chamber devices. In addition, the materials used for the components of the top-tier subunit 12 and the bottom-tier subunit 14 may differ. A nonlimiting example includes forming the walls of the housing 16 from copper, forming the wick structures 24 and 26 from porous sintered copper, and using water as the working fluids.
The vapor chamber device 10 provides for methods of dissipating heat from a heat generating surface in various applications. The methods may include locating the cascaded multi-core unit onto the surface, specifically such that the second input surface 30 of the bottom-tier subunit 14 contacts the surface or is thermally coupled with and therefore capable of conducting heat from the surface. Once in this location, the vapor chamber device 10 may be used to conduct heat from the surface, through the second input area 30, and into the cavity of the bottom-tier subunit 14 and thereby vaporize the second working fluid within one or more of the secondary vapor cores 22. The heat of the vapor is rejected at the second output area 34 causing the second working fluid to condense into liquid form. A capillary pressure is generated within the one or more secondary vapor cores 22 to recirculate the second working fluid therein from a condenser side to an evaporator side of the secondary vapor cores 22 and thereby passively spread heat from the second input area 30 to a relatively larger second output area 34 of the bottom-tier subunit 14.
Heat may then be conducted from the second output area 34 of the bottom-tier subunit 14, through the first input area 28 of the top-tier subunit 12, and into the cavity of the top-tier subunit 12 and thereby vaporize the first working fluid within the primary vapor core 20. The heat of the vapor is rejected at the first output area 32 causing the first working fluid to condense into liquid form. A capillary pressure is generated within the primary vapor core 20 to recirculate the first working fluid therein from a condenser side to an evaporator side of the primary vapor core 20 and thereby passively spread heat from the first input area 28 to the relatively larger first output area 32. Heat may be conducted from the first output area 32 to a heat dissipating surface such as but not limited to a surface of a heat sink mounted to the first output area 32.
In this manner, the bottom-tier subunit 14 may be used to attenuate high heat flux hotspots on the heat generating surface with the secondary vapor cores 22 before the heat fluxes thereof pass through the second output area 34 of the bottom-tier subunit 14 to the first input area 28 of the top-tier subunit 12. Attenuation of the hotspot heat fluxes within the bottom-tier subunit 14, whose secondary vapor cores 22 have relatively thin evaporator wicks, thereby avoids a large thermal resistance that would be otherwise incurred by directly subjecting the much thicker evaporator wick in the top-tier subunit 12 to hotspots.
Nonlimiting embodiments of the invention will now be described in reference to experimental investigations leading up to the invention.
A reduced-order model was developed to evaluate thermal resistance of different intra-lid heat spreaders including, for example, a solid copper device 40 (
For a fair comparison, the design envelope and power map were kept fixed across all of the heat spreaders 10, 40, and 50, while the remaining free design parameters were adjusted independently for each heat spreader type. All parameters varied during parametric design are shown in
R
ext
=R
TIM
+R
cond (1)
This thermal resistance results in an effective heat transfer coefficient (h) of 2250 W/m2K at the heat spreader-TIM interface. The temperature at the heat spreader-TIM interface can be computed from the total die heat load (Qdie), as:
T
cond,t
=T
inf
+Q
die
R
ext (2)
The heat spreaders 10, 40, and 50 were modeled as cylindrical disks with effective radii that yield the same equivalent heat input and heat output areas as the rectangular geometry. The solid copper device 40 resistance due to conduction was calculated as a function of the geometry and boundary conditions. The thermal resistance of the vapor chamber devices 10 and 50 for a given uniform heat input was estimated based on one-dimensional conduction across the evaporator wicks and the temperature drop across the vapor cores 20, 22, and 60 due to the saturation pressure difference. The resistance due to phase change at the interface was neglected. For evaluation of the vapor pressure drop, the thermophysical properties were taken at the temperature corresponding to the heat spreader-TIM interface.
For the given non-uniform power map, the maximum die temperature and the corresponding thermal resistance occurred at the hotspot location. To calculate the maximum thermal resistance for the solid copper device 40 and the conventional vapor chamber device 50 (Rsp), the total heat load of the power map was decomposed into a 468.75 W heat input (Q1) at a uniform flux of 0.75 W/mm2 over the entire die area and a 7.25 W heat input (Q2) over the 1 mm2 hotspot. The total temperature difference between the hotspot and the heat spreader-TIM interface (ΔThs) was computed from the thermal resistances associated with the decomposed heat inputs, respectively R1 and R2 as estimated from the reduced-order models, using the principle of linear superposition:
ΔThs=Q1R1+Q2R2 (3)
This net hotspot temperature difference (ΔThs) and the heat load at the hotspot (Qhs) were employed to compute the maximum heat spreader resistance as:
For the bottom-tier subunit 14 that was located directly over the hotspot of the die, it was considered that the total heat input to this bottom-tier subunit 14 was spread uniformly over the condenser wick, and there was a uniform heat flux into the evaporator wick of the top-tier subunit 12 dictated by the cross-sectional area (Acond,b) of the bottom-tier subunit 14.
T
hs
−T
cond,chs
=Q
1,b
R
1,b
+Q
2,b
R
2,b (5)
This net temperature difference and the total heat input at the hotspot (Qhs) were employed to compute the resistance of the bottom-tier subunit 14 as:
This temperature difference and the heat load for the hotspot (Qhs) were employed to compute the thermal resistance of the top-tier subunit as:
To evaluate the maximum heat spreader thermal resistance (Rsp) for the vapor chamber device 10 using Eq. (4), the net temperature difference between the hotspot and the condenser wick of the vapor chamber device 10 (ΔThs) was calculated as:
ΔThs=Ths−Tcond,s=Qhs(Rsp,b+Rsp,t) (9)
The wick structures 28, 30, and 68 of the vapor chamber devices 10 and 50 were designed to have the minimum possible thickness without reaching the capillary limit at the required total heat load; this corresponds to the possible conduction thermal resistance. This minimum wick thickness for a given vapor chamber device (or individual vapor core within the bottom-tier subunit 14) was dictated by the balance between the total liquid pressure drop (ΔPl) and the available capillary pressure (ΔPcap). The liquid pressure drop (ΔP) for a given uniform heat input (Q) over the entire evaporator area was estimated by considering a one-dimensional radial flow through the evaporator and the condenser wicks according to Darcy's law for porous materials:
For the representative nonuniform power map, the liquid pressure drops (ΔPl) and (ΔP2) were respectively computed using Eq. (10) from the decomposed uniform heat inputs Q1 and Q2. The total pressure drop of the liquid, employing the principle of linear superposition, was estimated as:
ΔP1=ΔP1+ΔP2 (11)
The driving capillary pressure head was computed from an effective pore radius of the wick and assuming perfect wettability as:
Owing to their high capillary pressure and effective thermal conductivity, this investigation considered sintered copper particle wicks. The effective pore radius (Teff) and the permeability (K) of the wicks were estimated as a function of the wick porosity (e) and the particle diameter (D), as:
The capillary-limit-governed thickness of the wicks (tcap) was obtained by equating ΔP1 to ΔPcap, and depended on the ratio of the effective pore radius and the permeability of the wicks:
where, Ml=(ρlσlhl,v)/μl, was the liquid figure of merit. An additional constraint was imposed to ensure that the sintered copper wicks had a minimum thickness of at least three particle diameters. Hence, the wick thickness (twick) became set based on the maximum of either the capillary-limited thickness or this three-particle constraint:
t
wick=max(tcap,3D) (16)
For the vapor chamber device 10, this same design approach was adapted to the individual tiers 12 and 14 by calculating the total pressure drop (ΔPl,b) in the bottom-tier subunit 14 using the decomposed heat inputs Q1,b and Q2,b, and equating to the capillary pressure (ΔPcap,b) to design the wick thickness (twick,b). Separately, the balance between the capillary pressure (ΔPcap,t) and the total liquid pressure drop (ΔPl,t) in the top-tier subunit 12, computed with Q1,t and Q2,t, was used to design the wick thickness in the top-tier subunit (twick,t).
For each heat spreader type, a parametric optimization investigation was performed to minimize the thermal resistance (Rsp) for the same equivalent cylindrical design envelope dimensions (dcond,t) and power map.
The thermal resistance of the solid copper device 40 was governed only by its thickness (t) and cross-sectional area. The minimum resistance was obtained for the trivial case where copper occupies the entire structure.
For the vapor chamber devices 10 and 50, water was considered as the working fluid. Furthermore, the thickness of the copper walls of the housing was neglected, such that the vapor cores 20, 22, and 60 and wick structures 24, 26, and 64 occupy the entire design envelope when comparing their performance with the solid copper device 40. The total thermal resistance was dictated by one-dimensional heat conduction across the wick structures 24, 26, and 64 and heat spreading in the vapor cores 20, 22, and 60. Because the wick thickness was minimized, this determined the vapor core thickness and the associated thermal resistance. The thermal resistance resulting from heat conduction in the wicks (Rwick) was determined by the effective thermal conductivity of the sintered copper powder, the wick thickness, and the area corresponding to a given heat load (Ahl). Because the wick thickness was constrained per Eq. (16), these conduction thermal resistances were inherently a function of the porosity, particle diameter, and permeability, as:
For the conventional vapor chamber 50, because there was only a single core 60, the wick thickness (twick) set the vapor core thickness based on the available total design envelope thickness (t). The vapor core thermal resistance accounted for the in-plane heat spreading and the associated three-dimensional variation in temperature. For the design envelope thickness and power map considered in this investigation, it was experimentally demonstrated that the thermal resistance resulting from the difference in saturation pressure in the vapor core 60 (Rvap) was always orders lower than the conduction resistance across the wick structure 68 and was therefore neglected. For the parametric optimization investigation, the porosity was varied between 0.42 and 0.6, and the particle diameter was varied between 5 μm and 75 μm, corresponding to the approximate range of reasonable parameters for sintered copper wicks.
For the vapor chamber device 10, the parametric investigation was extended to allow the thicknesses of the individual tiers (tt, tb) to vary within the available design thickness (t). Furthermore, the number and diameter of the secondary vapor cores 22 (dcond,b) in the bottom-tier subunit 14 were free to vary and influence the heat flux levels at the evaporator wick of the top-tier subunit 12, which consequently affected the design of the wick thickness (twick,t, twick,b) and the vapor core thicknesses (tt, tb) of the individual tiers 12 and 14. The porosities (et, eb) and particle diameters (Dt, Db) of the wick structures 28 and 30 of the individual tiers 12 and 14 were varied between the same bounds as the conventional vapor chamber device 50. A custom software program executed the reduced-order model throughout the design space to identify the parameters which were believed to offer the lowest thermal resistance.
The contour plots represented in
For a given particle diameter, there existed a transitional porosity (noted by the dashed lines) above which the designed wick thickness was governed by the particle diameter constraint per Eq. (16) and below which the wick thickness was governed by the capillary limit. This was attributed to a reduction in the capillary-limit-governed wick thickness (tcap) because of the increase in the wick permeability. With an increase in the particle diameter, there was a reduction in the driving capillary pressure head (ΔPcap) (see Eq. (12)). Furthermore, the wick permeability increased (see Eq. (14)) with an increase in the particle diameter. Hence, with an increase in the particle diameter, the transition of the design from a capillary-limit-governed wick thickness (twick=tcap) to a particle diameter-governed wick thickness (twick=3D) occurred at a lower wick porosity.
The thermal resistance (Rsp) of the conventional vapor chamber 50 was dominated by the conduction resistance across the wicks (about 104 times the vapor core thermal resistance). The minimum thermal resistance (dashed line in
This parametric optimization investigation of the conventional vapor chamber device 50 yielded calculated optimal porosity (eopt) of 0.47, particle diameter (Dopt) of 66 μm, and wick thickness (twick,opt) of 199 μm having a calculated optimized thermal resistance of 1.76 K/W. At the calculated optimum, the thermal resistance was more sensitive to the porosity compared to the particle diameter (e.g., increases to 2.31 K/W versus 1.94 K/W with 10% increase in e and D, respectively). The thermal resistance isocontours shown in
The dependence of the thermal resistance (Rsp) of the vapor chamber device 10 on the core diameter (dcond,b) is depicted in
With an increase in the core diameter (dcond,b), there was an increase in the calculated optimum thermal resistance of the bottom-tier subunit 14 (Rsp,b,opt). However, there was a simultaneous reduction in the calculated optimum thermal resistance of the top-tier subunit 12 (Rsp,t,opt) and an increase in the heat input (Qchs) at the secondary vapor core of the bottom-tier subunit 14 above the hotspot. Consequently, the calculated optimal core diameter was governed by this tradeoff. Nevertheless, the results represented in
In another investigation leading to aspects of the present invention, it was demonstrated that the thermal resistance of a conventional vapor chamber device (e.g., analogous to the top-tier subunit 12) can be significantly reduced by the introduction of a buffer vapor chamber device (e.g., analogous to the bottom-tier subunit 14) placed below the conventional vapor chamber device to diffuse hotspots. This was achieved by characterizing the thermal resistance of a first vapor chamber device 110 that included a single vapor core 120, and a second vapor chamber device 210 having a cascading multi-core unit that included a top-tier subunit 212 having a single vapor core 220 that was identical to the vapor core 120 and a smaller footprint bottom-tier subunit 214 that included a single vapor core 222 interfaced to the top-tier subunit 212 to act as a buffer.
The above investigations evaluated the performance sensitivity to a range of parameters for the vapor chamber device 10 and thereby offered new insight into the flexibility of design in the context of manufacturing and subsequent intra-lid integration in an electronic package. Separately, the investigations demonstrated that the thermal resistance of a given commercial vapor chamber device can be reduced by interfacing it with another buffer vapor chamber device placed between the commercial vapor chamber device and the heat source. This demonstration confirmed a key benefit of the vapor chamber device 10, that a performance improvement can be achieved in the top-tier subunit 12 via a reduction in the conduction resistances across the internal wick structure 24.
While the invention has been described in terms of specific embodiments, it is apparent that other forms could be adopted by one skilled in the art. For example, the physical configurations of cascaded multi-core vapor chamber devices within the scope of the invention can differ from that shown, and materials and processes/methods other than those noted could be used. In addition, the invention encompasses additional embodiments in which one or more features or aspects of different disclosed embodiments may be omitted or combined. Therefore, the scope of the invention is to be limited only by the following claims.
This application claims the benefit of U.S. Provisional Application No. 63/030,064 filed May 26, 2020, the contents of which are incorporated herein by reference.
Number | Date | Country | |
---|---|---|---|
63030064 | May 2020 | US |