Interposer for system-on-chip cooling device

Abstract

A method of assembling a cooling device includes forming a centerpiece by the following steps: plating a metallic material onto a carrier to form a first metallic trace; forming a first layer of non-metallic material around the first metallic trace to form a first centerpiece layer; and removing the first metallic trace to form a first fluid channel. The method further includes connecting a first side of the centerpiece to a first manifold including a second fluid channel to fluidly connect the first fluid channel to the second fluid channel.

Description

Background Technology

[0001] Electronic printed circuit boards (PCBs), integrated circuits, and systems-on-a-chip (SoCs), such as those used in computers or other electronic systems, typically have boards on which electronic components are mounted. Electronic components can generate considerable heat due to electrical power consumption. This heat is dissipated from the components and the board to ensure proper functioning of the components and to prevent damage to any part of the SoC and the entire electronic system. Attached Figure Description

[0002] Figure 1 This is a side view of an example of a system-on-a-chip component including a cooling device; Figure 2 This is a side view of an example of a system-on-a-chip component including a cooling device; Figure 3 This is a side view of an example of a system-on-a-chip component including a cooling device; Figure 4 This is an example of a method for manufacturing a centrally located component; Figure 5 This is a side sectional view of an example of a central component connected to a microchannel manifold; Figure 6 yes Figure 5 Top view of the central component; Figure 7 yes Figure 5 Bottom view of the central component; Figure 8 This is a side sectional view of an example of a central component connected to a microchannel manifold; Figure 9 yes Figure 5 Top view of the central component; Figure 10 yes Figure 5 Bottom view of the central component; Figure 11 This is a side sectional view of an example of a central component connected to a microchannel manifold; Figure 12 This is a side view of an example of a system-on-a-chip component including a cooling device; Figure 13 This is a top view of an example of a center-mounted component; Figure 14 This is a side view of an example of a system-on-a-chip component including a cooling device; Figure 15 This is an example of a method for manufacturing a cooling device.

[0003] It should be understood that the accompanying drawings are schematic diagrams for illustrative purposes. These drawings are provided to illustrate embodiments and are expressly intended not to limit the scope of the claims. Detailed Implementation

[0004] Various techniques can be used to remove heat from a System-on-a-Chip (SOC). For example, heat can be simply transferred to ambient air via convection. However, the amount of heat transferred using this method may be relatively low. To enhance the dissipative effect of convection, a fan can be added to force airflow over the component. A heat sink or fin can be attached to the component to further enhance heat dissipation by moving heat away from the heat source and onto a large surface area where it can be dissipated into the air. The heat sink can be a solid material component with high thermal conductivity (such as metal) with fins extending from its base attached to the SOC. The fins increase the surface area of ​​the heat sink, thereby increasing the amount of heat that can be carried away by convection as air or another fluid flows over the fins. For even greater heat dissipation, a liquid-cooled heat dissipation system can be used.

[0005] A liquid-cooled system may include a pump or compressor, a cooling manifold including fluid passages, and a radiator. The cooling manifold may be connected to the state-of-the-art (SOC) to allow heat to be transferred from the cooling plates to the cooling fluid. The compressor pumps the cooling fluid through the manifold passages and through fluid-carrying pipes or conduits to the radiator located away from the SOC. In some cases, the radiator may include a compressor for actively cooling the cooling fluid.

[0006] Thermal interface materials (TIMs) can be positioned between a System-on-a-Chip (SOC) and a heatsink or cooling manifold to improve thermal conductivity and heat transfer between them. A variety of materials can be used as TIMs, including thermal paste, polymer-based TIMs, carbon-based TIMs, and metal-based TIMs. Metal-based TIMs can have the highest thermal conductivity and low thermal contact resistance, thus providing optimal heat transfer from the SOC to the heatsink or cooling manifold. These metal-based TIMs can include silver sintered paste, gallium-based liquid metals, solder materials, or TIMs comprising gold, silver, or copper bonding interfaces.

[0007] See now Figure 1The diagram illustrates a typical arrangement of a SOC assembly 100, which includes an SOC 102 bonded to a liquid-cooled cooling device 202. The SOC 102 may include processing elements 108 mounted to a substrate 110 (e.g., an organic or inorganic substrate). The cooling device 202 is bonded to the SOC 102 by a TIM 106. The cooling device 202 includes a silicon microchannel manifold (MCM) 204 bonded to a coolant distribution manifold (CDM) 208 by an adhesive layer 210. Cooling fluid (e.g., water, oil, etc.) may be supplied to the CDM 208 via a fluid inlet 211. The CDM 208 includes fluid channels that distribute the cooling fluid to smaller channels within the silicon MCM 204. The silicon MCM 204 may include multiple layers 205 of silicon (e.g., three layers as shown), each layer may be bonded to adjacent layers 205 by an adhesive. As shown in the figure, the lower layer 205a of the silicon MCM 204 includes microchannels extending through it. As shown, the ends of these microchannels are connected to channels in the intermediate layer 205b, which extend to the top layer 205c. At the top layer 205c, these channels can extend to the upper surface of the silicon MCM 204 and are fluidly connected to channels in the CDM 208. Therefore, the CDM 208 distributes fluid to the channels in the top layer 205c of the silicon MCM 204, and the liquid flows through the channels to the microchannels in the lower layer of the silicon MCM 204. After passing through the microchannels, the liquid continues to flow back to the intermediate and upper layers 205 of the silicon MCM 204, enters the CDM 208, and exits through the fluid outlet 213. The liquid can then be delivered to the heat sink of the cooling unit, where it is cooled and circulated back to the cooling device 202.

[0008] The microchannels in the lower layer 205a of the silicon MCM 204 can be the smallest channels in the silicon MCM 204, can be tightly packed, and can extend substantially parallel to and close to the lower surface of the lower layer 205a, which is bonded to the SOC 102 by TIM 106. This arrangement allows the fluid in the microchannels to absorb as much heat as possible from the SOC 102 by maximizing the surface area through which the fluid passes in close proximity to the SOC 102. Due to manufacturing constraints, it may be difficult to fabricate the channels of the CDM 208 (which can be made of molded thermoplastic material) to accurately and precisely supply fluid directly to the microchannels of the lower layer 205a of the silicon MCM 204. Instead, the channels in the silicon MCM 204 can increase in size as each layer of the silicon MCM 204 approaches the CDM 208 to interface with the relatively large channels of the CDM 208. For example, the microchannels in the lower layer 205a can each be approximately 60 micrometers wide and 600 micrometers deep. The fluid connection between the CDM 208 and the lower channel of the silicon MCM 204 can be approximately 1.4 mm wide and 6.25 mm deep. The cross-sectional area of ​​the channel in the additional layer 205 of the silicon MCM 204 increases until the channel in the upper layer 205 abuts (e.g., fluidly connected) with the channel of the CDM 208. This gradual increase in the cross-sectional area of ​​the channel allows the cooling device 202 to maintain a high coolant flow rate through the channel, uniform coolant temperature distribution, and reduced cooling pressure at the coolant inlet and outlet ports.

[0009] The additional layer also helps to account for the mismatch in coefficients of thermal expansion (CTE) between the silicon MCM 204 and the CDM 208 made of thermoplastic materials. For example, the CTE of silicon is about 3 ppm / °C, while the CTE of thermoplastic materials can range from 25 ppm / °C to 50 ppm / °C. An additional increase in the temperature of the CDM 208 can cause stress in the bonding interface between the silicon MCM 204 and the CDM 208 (e.g., in the adhesive layer 210). The adhesive layer 210 can be made thicker so that the adhesive can deform without causing cracks or applying additional stress to the silicon MCM 204 due to the difference in thermal expansion between the silicon MCM 204 and the CDM 208. The layer of silicon MCM 204 closest to SOC 102 experiences the highest temperature variation due to its proximity to SOC 102, while each subsequent layer closer to the coolant distribution manifold 208 experiences a gradually decreasing temperature variation, effectively isolating CDM 208 from SOC 102. This reduces the temperature variation of CDM 208, thereby minimizing the effects of CTE mismatch between silicon MCM 204 and CDM 208. The additional silicon layer 205 can thus reduce stress at the bonding interface. However, silicon is a relatively expensive material, and the additional layer 205 increases the height and overall size of the cooling device 202, which is undesirable in space-constrained electronics.

[0010] See now Figure 2 Based on some examples, SOC component 300 is shown. SOC component 300 can be used with... Figure 1 The SOC component 100 is substantially similar; the differences shown and described herein are for cost reduction and to reduce the overall height of the cooling device 202. One or more layers 205 of the silicon MCM 204 (e.g., top layer 205c) can be replaced by a centering element 309. The centering element 309 is positioned between the CDM 208 and the silicon MCM 204. The centering element 309 can be bonded to the CDM 208 by a first adhesive layer 312 and to the silicon MCM 204 by a second adhesive layer 313. The centering element 309 may include fluid channels (schematically shown as channel 348) that fluidly connect the fluid channels of the CDM 208 (schematically shown as channel 247) to the fluid channels of the silicon MCM 204 (schematically shown as channel 249) and the movement from the silicon MCM 204 to the CDM 208 gradually increases. For example, the cross-sectional area of ​​the upper channel of the silicon MCM 204 can be about 10 to 15 times that of the fluid channels (e.g., microchannels) in the lower layer 205a. The cross-sectional area of ​​the channel of the centering member 309 can be increased to about 5 to 10 times that of the upper channel of the silicon MCM 204, and the cross-sectional area of ​​the channel of the CDM 208 can be increased to about 10 to 15 times that of the channel of the centering member 309. The centering member 309 may also include electrical traces extending from the top to the bottom of the centering member 309, allowing electrical connections between the silicon MCM 204 and the CDM 208 to be made through the centering member 309.

[0011] The centering element 309 can be made of epoxy molding compound (EMC) and can be assembled using interconnect substrate technology. For example, the EMC can be an Ajinomoto deposited film (ABF), and its CTE can be approximately 7 ppm / °C (e.g., between approximately 5 ppm / °C and approximately 9 ppm / °C, or between approximately 4 ppm / °C and approximately 10 ppm / °C). Because the CTE of the centering element 309 can be between the CTE of silicon MCM 204 (e.g., 3 ppm / °C) and the CTE of CDM 208 (e.g., 25 ppm / °C to 50 ppm / °C), the centering element 309 can be extended to be larger than silicon MCM 204 and smaller than CDM 208. Therefore, the centering element 309 can provide a transition layer compared to directly connecting silicon MCM 204 and CDM 208 together (e.g., Figure 1The cooling device 202 (shown) includes a transition layer that reduces stress at the interface (e.g., at adhesive layers 312, 313). Because adhesive layer 313 is applied as a very thin layer, the adhesive is not squeezed into the narrow fluid channels between the centering element 309 and the silicon MSM 204 during assembly, so a similar CTE match may be expected between the centering element 309 and the silicon MSM 204. Because the channels at the interface between the centering element 309 and the CDM 208 are large, a thicker adhesive layer 312 (thicker than adhesive layer 313) can be used, which flexes and deforms under stress to compensate for the CTE difference between the centering element 309 and the CDM 208. In some examples, the centering element 309 may include multiple layers 310, each of which may include different formulations of EMC with different CTEs. For example, layer 310a, which is closely adjacent to silicon MCM 204, can be fabricated with EMC having a CTE of 7 ppm / °C, intermediate layer 310b can be fabricated with EMC having a CTE of 9 ppm / °C, and top layer 310c can be fabricated with EMC having a CTE of 11 ppm / °C. Therefore, the intermediate component 309 can provide progressively increasing CTEs from silicon MCM 204 to CDM 208.

[0012] The center element 309 can replace one or more layers 205 of the silicon MCM 204 with a lower-cost EMC while providing the same or more efficient cooling. For example, as Figure 2 As shown, the cooling device 302 includes a silicon MCM 204 with two layers 205. Compared to Figure 1 In the cooling device 202, the top layer 205c of the silicon MCM 204 has been eliminated and replaced by a center element 309 having four layers (e.g., a bottom layer 310a, a top layer 310c, and two intermediate layers 310b, 310b'). The complete cooling device 202 with two silicon layers 205 and a center element 309 provides equivalent or improved cooling compared to a cooling device with three silicon layers 205. The center element 309 can also be thinner than the top silicon layer 205c, resulting in a lower overall height of the cooling device 202. For example, each layer 205 of the silicon MCM can be approximately 420 micrometers thick, while each layer 310 of the EMC center element can be approximately 70 micrometers thick (e.g., 30 micrometers to 140 micrometers thick). Therefore, if the layers 205 of the silicon MCM 204 can be replaced by three-layer center elements 309, the overall height of the cooling device 202 can be reduced by approximately 210 micrometers. In some examples, the central element 309 can replace multiple layers 205 of a typical cooling unit 202. For example, as Figure 3As shown, the cooling device 302 may include a silicon MCM 204 having a silicon layer 205, wherein a centering element 309 replaces the intermediate and top layers 205 of the silicon MCM 204. Eliminating the silicon layer 205 also eliminates the additional manufacturing steps required to bond multiple layers 205 together. It should be understood that “replacement” as used herein refers to including a centering element 309 in the cooling device 202 to replace one or more additional silicon layers in the silicon MCM 204, which will provide the same or similar cooling. Therefore, “replacement” does not mean removing layers of the silicon MCM 204 after manufacturing the silicon MCM 204. The centering element 309 may include several thin layers 310 (e.g., thinner than layer 205 of the silicon MCM 204) to allow fluid channels to fan out from the lower layer 310a adjacent to the silicon MCM 204 to the layer 310c adjacent to the CDM 208. Figures 5 to 10 An example of a centering element 309 incorporated into a silicon MCM 204 is shown in further detail.

[0013] See now Figure 4An example method for manufacturing the center piece 309 is shown. At operation 402 of method 400, a carrier 452 is provided. At operations 404 and 406 of method 400, a dry film layer 454 and a second material 456 are applied (e.g., plated) to the carrier 452. The second material 456 can be a metallic material such as copper or a copper alloy, which forms metallic traces or trace portions on the carrier. The dry film 454 can allow the second material 456 layer to be supported as it accumulates from the carrier 452. At operation 408 of method 400, the dry film 454 can be removed, thereby leaving the second material 456 on the carrier 452. At operation 410 of method 400, the remaining second material 456 can be molded into a molding compound 458 to form a layer surrounding the second material 456, which can be an epoxy molding compound (EMC) or other non-metallic material. At operation 412 of method 400, an additional layer of dry film 454 and second material 456 may be applied over the EMC and the lower layer of second material 456 to form additional traces or trace portions of second material 456. A portion of the additional metallic material may contact the metallic material applied in operations 404 and 406, such that the trace is formed by multiple trace portions. After completing method 400, these traces may extend from the carrier 452 to the top of the center member 309. At operation 414 of method 400, dry film 454 may be removed. At operation 416 of method 400, an additional layer of second material 456 may be molded into another layer of molding compound 458. These operations may be repeated as needed to add additional layers of second material 456 and molding compound 458. At operation 418 of method 400, carrier 452 may be removed. At operation 420 of method 400, a portion or all of the second material 456 may be removed, thereby retaining the center member 309 with channels 460 that penetrate the layer of molding compound 458. For example, if the molding compound 458 is EMC and the second material forming the traces is copper, selective chemical etching can be used to remove the copper and form the channels 460. Method 400 allows the layer of molding compound 458 to be very thin, with precisely and densely arranged channels 460.

[0014] Method 400 also allows a portion of the second material to remain within the centering member 309 without being removed, for example, by masking these portions of the second material during an electrochemical etching process (e.g., in operation 420). For example, copper traces may extend from the upper surface of the centering member 309 to the lower surface of the centering member 309 to conduct electrical signals through the centering member 309. The copper traces may be electrically connected to electrical contacts in the silicon MCM 204 and CDM 208 to conduct electrical signals between the silicon MCM 204 and CDM 208. The electrical contacts may be electrically connected to electrical components in the silicon MCM 204, such as flow sensors, temperature sensors, pressure sensors, or other sensors or electrical components. For example, CDM 208 may include or be coupled to a controller that is communicatively coupled to the electrical components via the traces. The controller may, for example, receive sensor data from the electrical components and use the sensor data to control various aspects of the cooling device 302. For example, when an increase in temperature is detected in the silicon MCM 204, the controller can control the fluid pump to increase the speed at which cooling fluid flows through the channels.

[0015] The centering element 309 manufactured according to method 400 may be superior to other potential alternatives for several reasons. For example, as discussed, the additional silicon layer 205 in the silicon MCM 204 can increase cost and the overall size of the cooling device 202. The centering element 309, even a centering element with multiple layers 310, can be thinner than a single layer 205 of the silicon MCM 204. The centering element 309 manufactured according to method 400 can also allow for fine fluid spacing between channels and allows for narrower slots than alternative manifolds that include injection-molded thermoplastic manifolds. For example, prototypes have been manufactured to exhibit fluid channel spacing of less than 150 micrometers, with channels 73 micrometers wide spaced only 76 micrometers apart. The slots have also been formed to be as narrow as 43 micrometers. Injection-molded thermoplastic manifolds have much higher minimum fluid channel spacing and minimum channel size (e.g., width closer to 1 millimeter). Injection-molded thermoplastic manifolds also feature relatively high CTE, significant draft angles for demolding, low glass transition temperature, and rough surface finish, and are typically not manufactured with multiple layers. In contrast, the center piece 309 can be made of chemically stable ABF material, which can be ground with a grinding wheel to achieve a smooth surface finish, thus allowing for strong adhesion to the silicone MCM.

[0016] See now Figures 5 to 7 The side sectional view, top view, and bottom view of the example component 309 are shown respectively. Figure 5In the side sectional view, the centering element 309 is shown connected to the silicon MCM 204. The silicon MCM 204 includes two layers 205, and the centering element 309 includes three layers 310. The lower layer 205a of the silicon MCM 204 includes microchannels 252 extending through it, these microchannels being near the lower surface 254 of the silicon MCM 204, which is bonded to the SOC 102. The microchannels 252 may be approximately 50 micrometers wide and 600 micrometers deep. The upper layer 205 of the silicon MCM 204 includes vertically extending channels 256, these vertically extending channels fluidly connecting the microchannels 252 to the lower channel 352 of the lower layer 310a of the centering element 309. Figure 5 As shown, the lateral lower channel 352 extends from below the center member 309 along the upper surface of the silicon MCM 204 and is fluidly connected to a vertically extending channel 354 in the second layer 310b of the center member 309 (e.g., the middle layer of the center member 309). The vertically extending channel 354 fluidly connects the lower channel 352 to a distribution channel 356 in the upper layer 310c of the center member 309. For example, as Figure 6 As shown, multiple vertically extending channels 354 can be connected to a larger fluid distribution channel 356. The larger fluid distribution channel 356 extends above the central member 309 and can be fluidly connected to the fluid channels of the CDM 208. The average cross-sectional area of ​​the lower channel 352 (e.g., along the fluid flow direction) can be smaller than the average cross-sectional area of ​​the larger fluid distribution channel 356. Figure 7 As shown in the bottom view, lower channels 352 lead to the bottom of the center member 309, and these lower channels can be fluidly connected at the bottom to the vertically extending channels 256 of the silicon MCM 204. Figure 5 As shown, when the centering member 309 is connected to the silicon MCM 204, the upper surface of the silicon MCM 204 forms the lower surface of the fluid channel 352, the sealing opening channel 352, and the vertically extending channel 256 of the silicon MCM 204 is fluidly connected to the vertically extending channel 354 of the centering member 309.

[0017] Figures 8 to 10 A side sectional view, top view, and bottom view of component 309 in another example are shown respectively. Figures 8 to 10 The central component 309 can be combined with Figures 5 to 7 The center component 309 is basically similar, but the differences are shown and described in this article. Figures 8 to 10 The centering element comprises four layers 310 and is shown connected to a silicon MCM 204 comprising one layer 205. (See also...) Figures 5 to 7The silicon MCM 204 includes a layer 205 comprising microchannels 252 extending therethrough, these microchannels being located near a lower surface 254 of the silicon MCM 204, which is bonded to the SOC 102. However, the silicon MCM 204 does not include a second upper layer 205 having vertically extending channels 256, as... Figures 5 to 7 As shown, instead, the centering member 309 includes a fourth layer at its lower end with a vertically extending channel 358 that fluidly connects the microchannel 252 to the lateral lower channel 353. Therefore, the lateral lower channel 353 does not lead to the bottom of the centering member 309, as... Figures 5 to 7 As shown. Figure 10 The diagram shows lower channels 358 leading to the bottom of the center element 309, and these channels can be fluidly connected at the bottom to microchannels 252 of the monolayer silicon MCM 204. (See diagram for reference.) Figure 9 The top view shown is... Figure 6 The top views shown are essentially the same (as are the top layer 310c of each center piece 309 in these figures), the difference being in... Figure 8 In the middle, the upper surface of the lower layer 310a is visible through the vertically extending channel 354.

[0018] Figures 5 to 10 In the example shown, centering element 309 allows the number of layers 205 in the silicon MCM 204 to be reduced from three or more to two or even one. The multiple layers 310 of the centering element can be thinner than the layers 205 of the silicon MCM 204, such that the centering element 209 with multiple layers 310 can be thinner than even a single layer 205 of the silicon MCM 204. Therefore, replacing the layers 205 of the silicon MCM 204 with centering element 309 can reduce the overall height of the cooling device 302. Replacing two layers 205 of the silicon MCM 204 makes the silicon MCM comprise only a single layer (such as...). Figures 8 to 10 (As shown), this reduces the overall height of the cooling device 302 by even more. This allows the cooling device 302 to be more easily packaged into an electrical device without increasing the size of the device housing. Since multiple layers will not need to be bonded together in additional manufacturing steps, a single-layer silicon MCM 204 can also be easily manufactured. Due to the cost involved in forming fluid channels in multiple layers 205 of the silicon MCM 204 and bonding these layers 205 together, the cost of manufacturing a cooling device 302 with a center element 309 and a silicon MCM 204 having one or more layers 205 can be lower than the cost of manufacturing a cooling device 202 without a center element but with a silicon MCM 204 having three or more layers 205.

[0019] Figure 11 A side sectional view of a portion of an example fixture 309 coupled to an example silicon MCM 204 is shown (e.g., with). Figure 5(Similar to the central portion of the centering element 309). As described above, the centering element 309 can be manufactured by molding a metal material (e.g., copper) into an epoxy resin film plastic (e.g., ABF). The metal material can then be removed (e.g., using selective etching) to form fluid channels 352, 354, 356. However, a portion of the metal material can be retained in the centering element 309 to serve as electrical traces 360. The electrical traces 360 can extend from the upper part 362 through the centering element 309 to the lower part 364, where they can be electrically connected to electrical contacts 260 in the silicon MCM 204. The electrical contacts 260 can be electrically connected to electrical components 262, such as flow sensors, temperature sensors, pressure sensors, or other sensors or electrical components. The upper ends of the electrical traces 360 can be electrically connected to additional components in the CDM 208 to establish an electrical connection between the CDM 208 and the silicon MCM 204. For example, CDM 208 may include or be coupled to a controller that is communicatively coupled to electrical component 262 via trace 360 ​​and may, for example, receive sensor data from electrical component 262. Any of the centering devices 309 discussed herein may include electrical trace 360, such as... Figure 11 As shown.

[0020] See now Figure 12According to some examples, a SOC assembly 500 with a cooling device 502 is shown. Due to manufacturing defects, the substrate 110 and SOC 102 as a whole can exhibit bow-shaped bending as the temperature of SOC 102 increases. This bow-shaped bending can become more pronounced in larger SOC 102s. A single large silicon MCM 204 will resist the bending of SOC 102, which causes stress in the TIM 206 and in the interface where SOC 102 and silicon MCM 204 are bonded with TIM 206. To address this issue, the SOC assembly 500 includes multiple silicon MCMs 204, each with an independent TIM 206 bonded to SOC 102. Each of these silicon MCMs 204 can be spaced apart, thereby allowing each silicon MCM 204 to move relative to the others. Because the silicon MCMs 204 are spaced apart, as the SOC 102, silicon MCMs 204, and TIM 206 are heated, the silicon MCMs 204 can expand toward each other and / or move vertically relative to each other (as shown in the figure) without applying additional stress to the SOC 102, TIM 206, and silicon MCMs 204. Because the spacing between the silicon MCMs 204 alleviates stress caused by thermal expansion, rigid materials, including metal-based TIMs (such as sintering paste), can be used for TIM 206, rather than more compliant TIMs. Therefore, the SOC assembly 500 allows TIMs with very high thermal conductivity and very low contact resistance for use, despite their rigidity. In some examples, TIM 206 can be continuous, while the silicon MCMs 204 are spaced apart. For example, the SOC component 500 may include a single segment of TIM 206 bonded to two or more silicon MCMs 204, without including separate segments of TIM 206 for each silicon MCM 204. The separation of the silicon MCMs 204 may be sufficient to reduce stress without simultaneously separating the TIM 206, as... Figure 12 As shown.

[0021] like Figure 12 As shown, all of the silicon MCM 204 are connected to a central element 309. In other examples, fewer than all of the silicon MCM 204 may be connected to a central element 309. For example, the cooling device 502 may include multiple central elements 309, each connected to a subset of these silicon MCM 204. Figure 14As shown, the SOC assembly 600 includes a cooling device 602, which includes a first center member 309a having a flexible joint 311 and two joint portions 314, each connected to a silicon MCM 204. The cooling device also includes a second center member 309b, which does not include the flexible joint 311 and is connected to a single silicon MCM. In some examples, there are separate center members 309 connected to each cooling unit. However, it may be desirable to use a single center member 309 to connect multiple silicon MCMs 204 together, allowing for tight control of the spacing between these silicon MCMs 204. For example, it may be desirable to bring these silicon MCMs 204 very close to each other, such that the entire surface of the SOC 102 (e.g., the upper surface of the SOC 102, such as...) is... Figure 3 (As shown) is essentially covered by silicon MCM 204, with gaps located on insignificant portions of the surface. For example, the gap between each silicon MCM 204 can be approximately 200 micrometers wide or less. Such small gaps could be difficult to maintain if each silicon MCM 204 were individually bonded to the SOC 102. Therefore, each silicon MCM 204 can be initially coupled to a centering element 309, thereby maintaining very small gaps between the silicon MCM 204s, and the complete cooling device 202 can be integrated into the SOC 102. This can also allow the cooling device 202 to be pre-assembled for placement on an existing SOC 102, for example, to replace an existing heatsink. The cooling device 202 can be manufactured and / or assembled using precision mechanics to maintain very small gaps between the silicon MCM 204s. However, if a rigid centering element 309 is coupled to multiple silicon MCM 204s, the centering element 309 can resist the ability of these silicon MCM 204s to move relative to each other. If TIM 206 is also rigid, this can cause stress at the interface between the silicon MCM 204 and SOC102.

[0022] like Figure 12As shown, the cooling device 502 includes a center member 309 having flexible joints 311 in the region between the silicon MCMs 204. Various portions between the flexible joints 311 can be coupled to the silicon MCMs 204; these portions may be referred to as joint portions 314 (e.g., center member wafers, wafer portions, etc.). When the SOC 102 deforms, the flexible joints 311 allow the silicon MCMs 204 to move relative to each other. Therefore, the center member 309 can provide minimal resistance to movement of the silicon MCMs 204 relative to each other. However, because the center member 309 still connects multiple silicon MCMs 204 together, these silicon MCMs 204 can be precisely positioned and coupled to the center member 309 to maintain a tightly controlled gap between the silicon MCMs 204. The flexible joints 311 may, for example, be in the form of partial or full cutouts or a combination of partial and full cutouts in the center member 309. Therefore, the cross-sectional area of ​​the flexible joint 311 (e.g., on the plane extending longitudinally through the flexible joint 311) can be smaller than the cross-sectional area of ​​the adjacent joint portion 314 on the parallel plane. Figure 13 An example of a centering member 309 is shown, which has ten bonding portions 314 to be coupled to ten silicon MCMs 204s. Flexible bonding portions 311 are positioned between each adjacent bonding portion 314 of the ten bonding portions 314. The flexible bonding portions 311 include both a through-cut portion 315 and a partial cut portion 317. The through-cut portion 315 does not provide resistance to movement of the silicon MCMs 204s relative to each other. The partial cut portion 317 provides relatively minor resistance to movement of the silicon MCMs 204s relative to each other while connecting the individual bonding portions 314 together.

[0023] See now Figure 15 An example method 430 for manufacturing a cooling device (e.g., cooling device 202) is shown. At operation 432 of method 430, a centering member 309 is formed according to method 400. At operation 434 of method 400, the centering member 309 may be cut to form a flexible joint between two adjacent sheet portions. The flexible joint may be a slot, a through cut, or a combination of both. The following description is in contrast to... Figure 13 and Figure 14The centering member 309 with a flexible joint is further discussed. In other examples, the flexible joint can be formed in the centering member 309 during operation 432. Instead of cutting the centering member, a metallic trace can be formed on the underside or top of the centering member 309, and a non-metallic material can be formed around the trace. The metallic trace can be removed (e.g., by etching) to leave a recess or channel on the surface of the centering member 309, which can serve as a flexible joint alone or in conjunction with additional recesses or channels. In other examples, the centering member 309 can be formed and assembled into the cooling device 202 without any flexible joint. At operation 436 of method 400, a first side of the centering member 309 is coupled to a first manifold (e.g., silicon MCM 204) to fluidly couple the fluid channels (e.g., microchannels) of the first manifold to the fluid channels of the centering member 309. An adhesive can be used to couple the centering member 309 to the first manifold. Connecting the first manifold to the centering member 309 may include electrically connecting electrical traces formed in the centering member 309 to electrical contacts in the first manifold to allow electrical signals to be sent to and from the first manifold via the centering member 309.

[0024] In an example where a flexible connection has been formed in the centering member 309, a first manifold can be coupled to a first portion of the centering member 309, and method 400 further includes operation 438. At operation 438, a second manifold (e.g., a silicon MCM 204) can be coupled to a second portion of the centering member on the first side of the centering member to connect the fluid channels (e.g., microchannels) of the second manifold to the fluid channels of the centering member 309. An adhesive can be used to join the centering member 309 to the second manifold. The first and second manifolds can be spaced apart from each other, having a flexible connection therebetween. At operation 440 of method 400, a second side of the centering member 309 is coupled to a CDM 208 to fluidly connect the fluid channels of the CDM 208 to the fluid channels of the centering member 309, and thereby fluidly connect the fluid channels of the CDM to the fluid channels of the first manifold, and in some examples, to the fluid channels of the second manifold. An adhesive or a gasket can be used to join the centering member 309 to the CDM 208. Method 400 may include additional operations not listed. For example, an additional flexible junction may be formed in the center member 309 to form an additional wafer portion, and an additional silicon MCM 204 may be coupled to the additional wafer portion. In some examples, method 430 may not include some of the operations not listed. For example, as described above, the flexible junction may not be formed in the center member 309, and the plurality of silicon MCMs 204 may not be coupled to the center member 309.

[0025] In one aspect, a method for assembling a cooling device is provided. The method includes forming a central component by the following steps: plating a metallic material onto a carrier to form a first metallic trace; forming a first layer of non-metallic material around the first metallic trace to form a first central component layer; and removing the first metallic trace to form a first fluid channel. The method further includes: connecting a first side of the central component to a first manifold including a second fluid channel to fluidly connect the first fluid channel to the second fluid channel.

[0026] In some examples, the method further includes connecting a second side of the central component to a coolant distribution manifold that includes a third fluid passage to fluidly connect the first fluid passage to the third fluid passage.

[0027] In some examples, the non-metallic material is epoxy molding compound, and the metallic material includes copper, and removing the first metallic trace includes selectively etching the copper.

[0028] In some examples, the method further includes cutting the center piece to form a flexible joint between a first portion and a second portion adjacent to the first portion, and connecting a second manifold including a fourth fluid channel to the second portion on a first side of the center piece, wherein connecting the first side of the center piece to the first manifold includes connecting the first portion to the first manifold.

[0029] On the other hand, a cooling device for a system-on-a-chip is provided. The cooling device includes a central member comprising: a first layer having a first plurality of fluid channels extending from a first surface of the central member and having a first average cross-sectional area; and a second layer coupled to the first layer and including a second plurality of fluid channels extending from a second surface of the central member, fluidly coupled to the first plurality of fluid channels, and having a second average cross-sectional area larger than the first average cross-sectional area. The cooling device also includes a microchannel manifold comprising a first side to be coupled to the system-on-a-chip, a second side coupled to the first surface of the central member, and a third plurality of fluid channels fluidly coupled to the first plurality of fluid channels; and In some examples, the cooling device also includes a coolant distribution manifold connected to a second side of the central member and includes a fourth plurality of fluid channels connected to a second plurality of fluid channels.

[0030] In some examples, the cooling device also includes a first adhesive layer for connecting the central component to the microchannel manifold and a second adhesive layer for connecting the central component to the coolant distribution manifold, wherein the second adhesive layer is thicker than the first adhesive layer.

[0031] In some examples, the first layer of the centering member includes a first epoxy molding compound having a first coefficient of thermal expansion, and the second layer of the centering member includes a second epoxy molding compound having a second coefficient of thermal expansion, which is higher than the first coefficient of thermal expansion.

[0032] In some examples, the center piece includes metal traces molded into the epoxy molding compound.

[0033] In some examples, metal traces extend from the first side of the center unit to the second side of the center unit to enable electrical connection between the microchannel manifold connected to the first side and the coolant distribution manifold connected to the second side.

[0034] In some examples, the centerpiece includes multiple sheet portions to be connected to multiple manifolds and a flexible joint between adjacent sheet portions, the flexible joint having a smaller cross-sectional area in a plane extending longitudinally through the flexible joint than the cross-sectional area of ​​the adjacent sheet portions in a parallel plane.

[0035] In some examples, the intermediate component includes an intermediate layer between the first and second layers, which includes a fifth plurality of fluid channels that fluidly connect the first plurality of fluid channels to the second plurality of fluid channels.

[0036] In another aspect, a method for assembling a cooling device is provided. The method includes forming a centering component by the following steps: plating copper on a carrier to form a first copper trace portion; forming a first layer of epoxy molding compound around the first copper trace portion to form a first centering component layer; plating copper on the first centering component layer to form a second copper trace portion in contact with the first copper trace portion; forming a second layer of epoxy molding compound around the second copper trace portion to form a second centering component layer; and removing the first and second copper trace portions to form a first plurality of fluid channels. The method further includes: connecting a first side of the centering component to a first silicon microchannel manifold including a plurality of microchannels to fluidly connect the first plurality of fluid channels to the plurality of microchannels; and connecting a second side of the centering component to a coolant distribution manifold including a second plurality of fluid channels to fluidly connect the first plurality of fluid channels to the second plurality of fluid channels.

[0037] In some examples, the method further includes: cutting a slot in the centering member to form a flexible joint between a first portion and a second portion adjacent to the first portion, and connecting a first side of the centering member to a first silicon microchannel manifold includes connecting the first portion to the first silicon microchannel manifold. The method also includes: connecting a second silicon microchannel manifold to a second portion, the second portion being spaced apart from the first portion and having a flexible joint positioned between the first portion and the second portion.

[0038] In some examples, the method further includes: plating copper on a carrier to form a third copper trace portion, wherein a first layer of epoxy molding compound is also formed around the third copper trace portion; plating copper on a first intermediate layer to form a fourth copper trace portion in contact with the third copper trace portion to form an electrical trace, wherein a second layer of epoxy molding compound is also formed around the fourth copper trace portion; and electrically connecting the electrical trace to electrical contacts of a first silicon microchannel manifold.

[0039] It should be noted that certain paragraphs of this disclosure may refer to terms such as “first” and “second”, which are combined with means of apparatus, surface or side of apparatus, mode of operation, transmission chain, antenna, etc., to identify or distinguish it from another apparatus. These terms are not merely intended to describe the temporal or sequential association of entities (e.g., first apparatus and second apparatus), although in some cases such a relationship may exist. These terms also do not limit the number of entities (e.g., operations, apparatuses) that may operate within the system or environment.

[0040] While this specification contains details of specific implementations, these should not be construed as limiting the scope of what may be claimed, but rather as descriptions of features specific to particular implementations. Certain features described herein, within the context of individual implementations, may also be implemented in combination in a single implementation. Conversely, various features described within the context of a single implementation may also be implemented individually or in any suitable sub-combination in multiple implementations. Furthermore, although features may be described as functioning in certain combinations and even initially claimed in this way, in some cases features of the claimed combination may be removed from the combination, and the claimed combination may involve sub-combinations or variations thereof.

[0041] As used herein with respect to structural features (e.g., describing shape, size, orientation, direction, relative position, etc.), the terms “about,” “approximately,” “substantially,” and similar terms are intended to cover minor variations in structure that may arise, for example, from manufacturing or assembly processes, and are intended to have a broad meaning, conforming to common and accepted usage by one of ordinary skill in the art to which this disclosure pertains. Unless otherwise stated, when the words “about,” “approximately,” “probably,” “substantially,” etc., are used to indicate numerical values, they mean plus or minus ten percent.

[0042] As used herein, the term "connection" and its variations mean joining two components directly or indirectly to each other. This connection can be fixed (e.g., permanent or fixed) or movable (e.g., removable or releasable). This connection can be achieved by directly joining the two components to each other, joining the two components to each other using a separate intermediary member and any additional intermediate members joined together, or joining the two components to each other using an intermediary member integrally formed with one of the two components as a single monolithic body. If "connection" or its variations are modified by an additional term (e.g., direct connection), the general definition of "connection" described above is modified by the general meaning of the additional term (e.g., "direct connection" means joining two components without any separate intermediary member), resulting in a narrower definition than the general definition of "connection" described above. Such a connection can be a mechanical connection, an electrical connection, or a fluid connection.

[0043] The positions of elements mentioned herein (e.g., "top", "bottom", "upper", "middle", "lower", "above", "below", "vertical", "horizontal", etc.) are used only to describe the orientation of the various elements in the accompanying drawings. It should be noted that the orientation of the various elements may vary depending on other examples, and such variations are intended to be covered by this disclosure.

[0044] Regarding the use of virtually all plural and / or singular terms in this document, those skilled in the art may, depending on the context and / or application, appropriately convert plural forms to singular forms and / or singular forms to plural forms. For clarity, various singular / plural forms may be explicitly listed herein.

[0045] For illustrative and descriptive purposes, the foregoing description of illustrative examples has been presented. This description is not intended to be exhaustive or restrictive of the precise forms disclosed, and modifications and variations may be made in consideration of the foregoing teachings, or based on the examples disclosed in practice. The scope of the subject matter intended to be protected is defined by the claims appended herein and their equivalents.

Claims

1. A method for assembling a cooling device, the method comprising: The central component is formed through the following steps: A metal material is plated onto the carrier to form a first metal trace; A first layer of non-metallic material is formed around the first metallic trace to form a first central component layer; as well as Remove the first metal trace to form a first fluid channel; as well as The first side of the central component is connected to a first manifold including a second fluid channel to fluidly connect the first fluid channel to the second fluid channel.

2. The method of claim 1, further comprising connecting a second side of the central member to a coolant distribution manifold including a third fluid passage to fluidly connect the first fluid passage to the third fluid passage.

3. The method of claim 1, wherein the non-metallic material is an epoxy molding compound, and the metallic material comprises copper, wherein removing the first metallic trace comprises selectively etching the copper.

4. The method of claim 1, further comprising: Cut the center piece to form a flexible joint between the first piece and the second piece adjacent to the first piece; as well as Connecting a second manifold, including a fourth fluid channel, to the second piece portion on the first side of the centerpiece, wherein connecting the first side of the centerpiece to the first manifold includes connecting the first piece portion to the first manifold.

5. A cooling device for a system-on-a-chip, the cooling device comprising: The centering component includes: A first layer, the first layer having a first plurality of fluid channels, the first plurality of fluid channels extending from a first surface of the central member and having a first average cross-sectional area; and A second layer, connected to the first layer and including a second plurality of fluid channels, the second plurality of fluid channels extending from a second surface of the central member, fluidly connected to the first plurality of fluid channels, and having a second average cross-sectional area larger than the first average cross-sectional area; and A microchannel manifold, the microchannel manifold including a first side to be connected to a system-on-a-chip, a second side connected to the first surface of the centering member, and a third plurality of fluid channels fluidly connected to the first plurality of fluid channels.

6. The cooling device as claimed in claim 5, further comprising: A coolant distribution manifold, the coolant distribution manifold being connected to the second side of the central member and including a fourth plurality of fluid channels connected to the second plurality of fluid channels.

7. The cooling device of claim 6, further comprising a first adhesive layer for connecting the central member to the microchannel manifold and a second adhesive layer for connecting the central member to the coolant distribution manifold, wherein the second adhesive layer is thicker than the first adhesive layer.

8. The cooling device of claim 5, wherein the first layer of the central component comprises a first epoxy molding compound having a first coefficient of thermal expansion, and the second layer of the central component comprises a second epoxy molding compound having a second coefficient of thermal expansion, the second coefficient of thermal expansion being higher than the first coefficient of thermal expansion.

9. The cooling device of claim 5, wherein the central component comprises metal traces molded into an epoxy molding compound.

10. The cooling device of claim 9, wherein the metal trace extends from the first surface of the central member to the second surface of the central member to achieve an electrical connection between a microchannel manifold connected to the first surface and a coolant distribution manifold connected to the second surface.

11. The cooling device of claim 5, wherein the central member includes a plurality of plate portions to be connected to a plurality of manifolds and a flexible joint between adjacent plate portions, the flexible joint having a smaller cross-sectional area in a plane extending longitudinally through the flexible joint than the cross-sectional area of ​​the adjacent plate portions in a parallel plane.

12. The cooling device of claim 5, wherein the intermediate component includes an intermediate layer between the first layer and the second layer, the intermediate layer including a fifth plurality of fluid channels fluidly connecting the first plurality of fluid channels to the second plurality of fluid channels.

13. A method for assembling a cooling device, the method comprising: The central component is formed through the following steps: Copper is plated onto the substrate to form the first copper trace portion; A first layer of epoxy molding compound is formed around the first copper trace portion to form a first central component layer; Copper is plated on the first intermediate component layer to form a second copper trace portion that contacts the first copper trace portion; A second layer of epoxy molding compound is formed around the second copper trace portion to form a second centering layer; Remove the first copper trace portion and the second copper trace portion to form a first plurality of fluid channels; The first side of the central component is connected to a first silicon microchannel manifold comprising multiple microchannels to fluidly connect the first plurality of fluid channels to the plurality of microchannels; as well as The second side of the central component is connected to a coolant distribution manifold that includes a second plurality of fluid channels to fluidly connect the first plurality of fluid channels to the second plurality of fluid channels.

14. The method of claim 13, further comprising: Cutting a slot in the centering member to form a flexible joint between a first piece portion and a second piece portion adjacent to the first piece portion, wherein connecting the first side of the centering member to the first silicon microchannel manifold includes connecting the first piece portion to the first silicon microchannel manifold; as well as A second silicon microchannel manifold is connected to the second portion, which is spaced apart from the first portion and has the flexible joint positioned between the first portion and the second portion.

15. The method of claim 13, further comprising: Copper is plated on the carrier to form a third copper trace portion, wherein the first layer of epoxy molding compound is also formed around the third copper trace portion; Copper is plated on the first intermediate layer to form a fourth copper trace portion that contacts the third copper trace portion in order to form an electrical trace, wherein the second layer of epoxy molding compound is also formed around the fourth copper trace portion. as well as The electrical trace is electrically connected to the electrical contacts of the first silicon microchannel manifold.

Images