JP2012253107A - Laminated module and interposer used for the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a laminated module that can suppress temperature rises accompanying heat generation of semiconductor devices and can cause the semiconductor devices to stably operate even when the laminated module has a structure formed by laminating the semiconductor devices with large power consumption.SOLUTION: A laminated module comprises: an interposer 70 that has a channel 71 through which a fluid flows; one or more first semiconductor devices disposed on one side of the interposer 70; and one or more second semiconductor devices disposed on a side of the interposer 70 opposite to the one or more first semiconductor devices. The channel 71 of the interposer 70 is configured such that the fluid is adiabatically expanded during moving from a first region to a second region while having the first region with a relatively large cross-sectional area and the second region with a relatively small cross-sectional area.

Description

  The present invention relates to a stacked module formed by stacking a plurality of semiconductor devices (hereinafter, also referred to as a stacked semiconductor device), and more specifically, can suppress a rise in operating temperature of a semiconductor device with high power consumption and operate stably. The present invention relates to a laminated module that can be used and an interposer used therefor.

  In recent years, technological progress in the semiconductor industry represented by silicon has been significant, and it has greatly contributed to downsizing, weight reduction, price reduction, high functionality, etc. of equipment and systems regardless of industrial or consumer use. . On the other hand, the demand for semiconductor devices is not limited, and further higher integration, higher speed, and higher level are expected, and miniaturization is also expected.

  As a measure to meet these requirements, there is a case where the size of unit elements (for example, transistors) constituting a semiconductor device is reduced and the number of mounted unit elements is increased. The advantage of this measure is an increase in operation speed (acceleration) accompanying miniaturization and an increase in functions (or reduction in the number of required semiconductor devices) due to high integration. However, with the increase in speed and integration, the power consumption inside the semiconductor device increases, and the risk of unstable operation or destruction of the semiconductor device itself increases. In order to reduce these risks, a heat dissipation technology (or cooling technology) for semiconductor devices is essential.

  Many techniques have been developed for lowering the operating temperature of semiconductor devices. For example, there is a technique for cooling a semiconductor device by attaching a heat radiating fin (often made of aluminum alloy) to a high-power semiconductor device and blowing a flow of air onto the fin. When the power consumption is relatively low (for example, several watts), this technique can solve the problem. However, the power consumption of the latest semiconductor devices is even greater, and the CPU of a computer or the like can reach 100 watts or more. For this reason, in such a semiconductor device with high power consumption, if the heat radiation is not sufficient, the temperature of the semiconductor device rises, and thermal runaway or thermal destruction may occur. Therefore, it can be said that the upper limit of the operation of the semiconductor device is dominated by the heat dissipation technology.

  A “laminated module (laminated semiconductor device)” formed by laminating a plurality of semiconductor devices has an advantage that high integration can be realized relatively easily. In such a configuration, power consumption in the semiconductor device disposed in the lower layer not only increases the temperature of the semiconductor device but also increases the temperature of the semiconductor device disposed in the upper layer. For this reason, when a semiconductor device whose characteristics are sensitive to the operating temperature is arranged in the upper layer of the stacked module, the operation of the entire stacked module may become unstable. For this reason, in the case of a stacked module, heat generation from a semiconductor device with high power consumption is stacked by being released to the outside of the stacked module before being transmitted to a semiconductor device in a higher layer. A mounting structure in which heat transfer does not occur between the semiconductor devices is preferable.

  Conventionally, a mounting structure shown in FIG. 31 has been proposed as a cooling technique for stacked semiconductor devices (laminated modules). This figure is published as FIG. 1A of Patent Document 1. FIG.

  In FIG. 31, the chip stack 110 is composed of a stack of three semiconductor chips denoted by reference numerals 110a, 110b, and 110c. Each chip 110a, 110b, 110c is provided with a plurality of channels 175 formed by etching (in FIG. 31, reference numeral 175 indicates a representative channel). The chip stack 110 is cooled by flowing a fluid (refrigerant) through the channel 175. This fluid flows in a narrow channel 175 formed between the stacked chips 110a, 110b, 110c. When the chips 110a, 110b, and 110c are made of a semiconductor substrate, the thickness is usually several hundred micrometers (μm) or less.

  In the vertical direction of each of the semiconductor chips 110a, 110b, and 110c, the chips 110a, 110b, and 110c are electrically connected to each other by a TSV (Through Silicon Via) denoted by reference numeral 123.

US Patent Application Publication No. 2009/031186

  In the conventional stacked semiconductor device (stacked module) mounting structure shown in FIG. 31, the channel 175 is formed as if “a corridor in which a number of pillars stand” and has a height of several hundreds of micrometers. Since it is below a meter (micrometer), in order to flow a fluid (refrigerant) in the channel 175, it seems that a big pressure is required.

  Further, the flow direction of the fluid is indicated by a downward arrow and a right arrow below the reference numeral 111 in FIG. 31. The fluid flowing into the periphery of the chip stack 110 along the downward arrow is indicated by the right arrow. Along the channel 175 between the chips 110a, 110b, 110c along the flow path, and also flows around the chips 110a, 110b, 110c. As described above, considering that the height of the channel 175 is low and that there is a wide space around the chips 110a, 110b, and 110c, it is difficult to flow a fluid only into the channel 175, and many It appears that the fluid will flow along the periphery of the chips 110a, 110b, 110c. In addition, since the shape of the channel 175 formed in the chips 110a, 110b, and 110c (the shape along the direction in which the fluid flows) differs for each of the chips 110a, 110b, and 110c, it is uniform in all the channels 175 formed in each layer. It would also be difficult to achieve a smooth fluid flow.

  For these reasons, it is not always easy to achieve sufficient cooling of the chips 110a, 110b, and 110c (or heat dissipation from the chips 110a, 110b, and 110c) in the conventional mounting structure of FIG.

  Further, since the power consumption in the chip 110c arranged in the lower layer of the chip stack 110 causes not only the temperature rise of the chip 110c but also the temperature rise of the chips 110a and 110b arranged in the higher layer, the chip stack It is desirable not only to cool 110 itself, but also to prevent the heat generated in the lower layer chip 110c from being transmitted to the upper layer chips 110a and 110b. However, since it is difficult to realize a uniform fluid flow to the channel 175 formed in each layer, it is difficult to suppress heat conduction between the chips 110a, 110b, and 110c.

  The present invention has been made in consideration of the above circumstances, and the object of the present invention is to generate heat of the semiconductor devices even when the semiconductor devices having a large power consumption are stacked. An object of the present invention is to provide a laminated module that can be stably operated while suppressing the accompanying temperature rise, and an interposer used therefor.

  Another object of the present invention is to provide a stacked module capable of suppressing heat conduction between stacked semiconductor devices and suppressing an increase in temperature thereof, and an interposer used therefor.

  Other objects of the present invention which are not specified here will become apparent from the following description and the accompanying drawings.

(1) The laminated module of the present invention is
An interposer with a channel through which fluid flows;
One or more first semiconductor devices disposed on one side of the interposer;
One or more second semiconductor devices arranged on the opposite side of the interposer from the first semiconductor device,
The channel of the interposer has a first region having a relatively small cross-sectional area and a second region having a relatively small cross-sectional area, and moves from the first region to the second region. The fluid is configured to cause adiabatic expansion along the way.

  Since the laminated module of the present invention has the above-described configuration, heat generated from the first and second semiconductor devices is effectively transferred to the outside by the fluid by flowing a fluid through the channel of the interposer. Released. Moreover, since the fluid passing through the channel causes adiabatic expansion while passing from the first region to the second region, the temperature of the fluid decreases, and as a result, The temperature drops. Thus, since the interposer itself has a cooling (endothermic) action, the cooling action of the first and second semiconductor devices is higher than that in the case where adiabatic expansion is not used.

  As a result, since heat generated from the first and second semiconductor devices can be effectively discharged to the outside, even a stacked module having a configuration in which semiconductor devices with large power consumption are stacked can be used. Stable operation can be performed while suppressing a temperature rise caused by heat generation.

  In addition, since the interposer configured so that the fluid passing through the channel causes adiabatic expansion is interposed between the first semiconductor device and the second semiconductor device, the semiconductor device constituting the stacked module It is possible to suppress the heat conduction between them and suppress the temperature rise of the laminated module.

  (2) In a preferred example of the laminated module of the present invention, the height of the channel is constant, and the width of the second region is set larger than the width of the first region.

(3) The interposer of the present invention
A channel through which fluid flows,
The channel includes a first region having a relatively small cross-sectional area and a second region having a relatively small cross-sectional area, and a fluid that passes through the channel from the first region to the second region. Is configured to cause adiabatic expansion.

  Since the interposer of the present invention has the above-described configuration, the heat generated from the semiconductor device arranged in the vicinity of the interposer is effectively released to the outside by the fluid by flowing the fluid through the channel. The Moreover, since the fluid passing through the channel causes adiabatic expansion while passing from the first region to the second region, the temperature of the fluid decreases, and as a result, The temperature drops. Thus, since the interposer itself has a cooling (endothermic) action, the cooling action of the semiconductor device is higher than the case where adiabatic expansion is not used.

  As a result, the heat generated from the semiconductor device can be effectively discharged to the outside, so that even in a stacked module having a configuration in which semiconductor devices with large power consumption are stacked, the temperature rise due to heat generation of those semiconductor devices Stable operation can be suppressed.

  Further, by interposing the interposer between the semiconductor devices, it is possible to suppress the heat conduction between the semiconductor devices constituting the stacked module and to suppress the temperature rise of the stacked module.

  (4) In a preferred example of the interposer of the present invention, the height of the channel is constant, and the width of the second region is set larger than the width of the first region.

  (5) In another preferred example of the interposer of the present invention, a plurality of the channels are provided, and the inlets and outlets of the channels are formed independently of each other.

  (6) In still another preferred example of the interposer of the present invention, a plurality of the channels are provided, the inlets of the channels are connected to each other, and the outlets of the channels are formed independently of each other.

  (7) In this specification, related terms are defined as follows.

  -Substrate: As long as the substrate has rigidity capable of supporting the laminated module, its configuration and material are arbitrary.

  Semiconductor device: Refers to all semiconductor devices including the following (i) and (ii).

  (I) A semiconductor chip (bare chip) cut out from a semiconductor wafer after the completion of the wafer process. The semiconductor chip includes a so-called integrated circuit chip in which a semiconductor element such as at least one transistor or diode is arranged.

  (Ii) The packaged semiconductor chip. These are packaged in various packages called ball grid array (BGA), chip size package (CSP) and the like.

  Electronic components: Components that are also called passive elements, such as resistors, capacitors, and inductors (coils). There is a configuration (for example, module resistance) in which a plurality of single elements (individual parts) are combined. Further, sensors and actuators having specific functions are also included in the electronic component. Furthermore, the sensor and the actuator in which a signal processing circuit, a drive circuit, and the like are integrated are also included in the electronic component.

  Stacked module: has a structure in which two or more semiconductor devices are stacked. There are wire bonding, a through electrode (TSV) and the like as a method of interconnection between the layers constituting the laminated structure, but the method is not limited. For example, when the interposer is composed of one or more first semiconductor devices disposed on one side thereof and one or more second semiconductor devices disposed on the opposite side of the interposer, It becomes a “stacked module”. When more semiconductor devices are stacked, a “multi-layer stacked module” is obtained.

  Mounting structure: A structure including a substrate and the stacked module mounted thereon. However, a cover mounted on the substrate may be included.

  Interposer: It is arranged between the first semiconductor device and the second semiconductor device, and has a channel that penetrates (extends) from one end to the other end. Electrical connection points connected to electrical connection points (also referred to as pads) provided in the semiconductor device may be formed on the front and back surfaces of the interposer, respectively. Furthermore, a wiring pattern called a “rewiring layer” may be provided on the front and back surfaces of the interposer. In the above-described laminated module, a “wiring board” for “forming an electric conductive path between the semiconductor devices arranged above and below” may be inserted between the semiconductor devices constituting the laminated module. This “wiring board” may also be referred to as an “interposer”. However, in this specification, the “wiring board” is not included in the “interposer”.

  -Fluid: A gas or liquid that absorbs heat by heat conduction and has a heat dissipation or exhaust heat effect. A fluid having such a function is also referred to as a “refrigerant”. Specific examples include (i) chlorofluorocarbons and non-fluorocarbons (used frequently and many types), (ii) organic compounds such as butane and isobutane, and (iii) inorganic compounds such as hydrogen, helium, ammonia, and water. And carbon dioxide.

  (8) You may further provide the board | substrate which mounts the said lamination | stacking module, and the cover which adhere | attaches and fixes on the said board | substrate so that the said lamination | stacking module may be included, and forms internal space with the said board | substrate. In this case, it is preferable to provide a weir (partition member) disposed between the cover and the substrate for partitioning the internal space. The shape of the cover depends on the external shape of the laminated module. Although it is preferable that it is a substantially rectangular parallelepiped (a substantially cube is included), it is not limited to this. The vertices and ridges of the rectangular parallelepiped (the line segment where the surfaces intersect) may be smooth.

  Regarding the positions of the inlet and outlet formed in the cover, it is necessary that the fluid flowing into the internal space from the inlet passes through the channel of the interposer and flows out of the outlet. That is, the inlet is disposed on the upstream side of the weir, and the outlet is disposed on the downstream side of the weir. In other words, the inlet communicates with the upstream space, and the outlet is disposed on the downstream space. It is necessary to communicate. As long as this positional relationship is satisfied, there is no restriction on the positions of the inlet and the outlet. For example, (a) the inlet may be disposed on a specified surface of the cover, and the outlet may be disposed on a surface opposite to the surface of the cover, or (b) the inlet and the outlet Both may be disposed on the same surface (for example, the upper surface) of the cover, or (c) the inlet may be disposed on a designated surface of the cover, and the outlet may be disposed on the upper surface of the cover.

  As the material of the cover, metal, resin, or the like can be used. If it is desired to increase the cooling (heat dissipation) effect, the cover is preferably formed of a metal material, but the present invention is not limited to this. When the cover is formed of a resin material, a metal layer may be provided on the front side or back side of the cover, or on the front side and back side surfaces, in order to increase the cooling (heat dissipation) effect.

  The cover may be formed as an integral structure and fixed in close contact with the surface of the substrate. In order to adhere and fix the cover to the surface of the substrate, an adhesive (a gas generated during solidification preferably does not adversely affect the characteristics of the semiconductor device) may be used. When the cover is made of a metal material, the cover may be adhered and fixed to the metal layer provided on the surface of the substrate by metal / metal bonding (for example, welding, soldering, etc.).

  The cover may be composed of a plurality of components, and these structural components may be combined (assembled) to become the cover. For example, the cover may be configured by combining a “lid” (which has a flat plate shape) on which the inlet and the outlet are arranged and a “frame” forming a side surface portion of the cover. In this configuration example, the lower surface of the frame is in close contact with the surface of the substrate, and the upper surface of the frame is in close contact with the lower surface of the lid. The material of the frame may not be the same as that of the cover. For example, the lid may be made of a metal material and the frame may be made of resin or glass.

  Adhesive for adhesion / bonding of the lid and the frame, and adhesion / fixation of the frame and the surface of the substrate (preferably the gas generated during solidification does not adversely affect the characteristics of the semiconductor device) May be used. When both the lid and the frame are made of a metal material, metal / metal bonding (for example, welding or soldering) may be performed. When the lid is a metal (for example, aluminum) and the frame is glass, electrostatic bonding (a metal-glass bonding method) may be used. If the frame is a metal material, metal / metal bonding (for example, welding or soldering) may be performed between the frame and a metal layer provided on the surface of the substrate. When the frame is glass and the surface of the interposer is metal (or the reverse combination), electrostatic bonding may be used.

  (9) The dam prevents the fluid flowing into the internal space from the inlet from being guided to the outlet through the side surface (outer periphery) of the stacked module without flowing into the channel. This is because the desired cooling effect cannot be obtained if the fluid flowing in from the inlet flows out of the outlet directly without passing through the channel. The dam having such a function includes (a) a gap provided in advance in a position where the dam of the cover is to be disposed, (b) the cover is closely attached to the substrate, and (c) from the gap. It can be formed by a procedure of injecting a material (synthetic resin or the like) that becomes the weir. The synthetic resin used here is thermosetting, has high adhesion to the laminated module and the cover, and has a thermal expansion coefficient close to that of the laminated module and the cover (preferably Preferably match))). However, the present invention is not limited to this.

  Due to the presence of the weir, part of the fluid that has flowed into the internal space from the inlet enters the upstream space (the space on the inlet side formed between the stacked module, the cover, and the weir). And there is a tendency to stay there. That is, the fluid tends to hardly flow in the upstream space. When such stagnation occurs, the heat dissipation effect by the fluid is reduced. In order to prevent this reduction in heat dissipation effect, it is preferable that the entire amount of the fluid always flows from the inlet to the outlet. For this purpose, the weir needs to extend to the vicinity of the inlet side end of the laminated module to reduce the volume of the upstream space. For example, the width of the weir (flow of the fluid) The length along the direction) may be increased. Such a wide weir can be easily realized by increasing the gap for injecting the synthetic resin or the like forming the weir.

  As described above, when the width of the weir is increased to reduce the volume of the upstream space, the “gap” increases. When forming the weir, the resin is injected from the gap provided in the cover. However, when using a manufacturing process in which the resin is extruded from a syringe by pneumatic pressure or the like, the resin is exposed in the gap. The resin is also applied to the outer surface of the laminated module. However, the resin may not be applied over the entire outer surface of the exposed laminated module. That is, the resin is applied only to the upstream end and the downstream end of the gap, and the resin is applied to the central portion of the gap (the portion between the upstream end and the downstream end). It may not be done.

  In this way, when the resin is applied only to the upstream end and the downstream end of the gap, it is necessary to make the flow of the resin from the syringe narrow, so that the syringe has a small diameter. Is used. In this case, even after application of the resin, the exposed portion remains on the outer surface of the laminated module, and heat is radiated from the exposed portion. The heat radiation from the exposed portion is the heat radiated directly from the gap of the cover to the outside, and is not the heat radiation utilizing the fluid.

  (10) About the said weir, the structure which developed further the structure mentioned above is also considered. For example, the heat radiation from the exposed portion of the outer surface of the laminated module is not performed by thermal radiation, but a second fluid (a fluid different from the cooling fluid passing through the channel) is used. can do. In this configuration, a second cover is placed on the substrate so as to cover the laminated module having the exposed portion on the outer surface (which is between the upstream end and the downstream end) and the cover. By closely contacting and fixing, a second internal space is formed between the cover and the second cover together with the substrate. Then, the second fluid is poured into the second internal space.

  In this configuration, it is desirable to apply a pressure to the fluid supplied to the internal space so that the fluid smoothly flows into the channel having a small cross-sectional area. On the other hand, the second fluid supplied to the second internal space does not need to be pressurized like the fluid. This is because it is not necessary to allow the fluid to flow into the channel having a small opening cross-sectional area. For this reason, the second cover is provided with a single port through which the second fluid flows in and out (that is, functions as an inlet and an outlet), and a heat pump (compressor) mounted on a notebook computer. It is also possible to adopt a configuration similar to that of the one not using. Employing such a double cover configuration has the advantage that the heat dissipation effect of the heat generated in the laminated module is further enhanced.

  (11) The interposer may include at least one of a heat radiation layer and a heat reflection layer disposed in a designated region of the inner wall of the channel. For example, a heat radiation layer may be disposed in a first designated region of the inner wall of the channel of the interposer, and a heat reflecting layer may be disposed in a second designated region of the inner wall. In this case, for example, (a) a stacked module composed of at least two semiconductor devices and an interposer sandwiched between the semiconductor devices vertically adjacent to each other; and (b) a substrate on which the stacked module is mounted. (C) a channel through which a fluid flows is formed inside the interposer, and (d) a mounting structure in which a heat radiation layer is disposed in a first designated region of the inner wall of the channel. Or (a) at least two semiconductor devices, and a laminated module composed of an interposer sandwiched between the semiconductor devices vertically adjacent to each other; and (b) a substrate on which the laminated module is mounted. (C) A mounting structure is provided in which a channel through which a fluid flows is formed inside the interposer, and (d) a heat reflecting layer is disposed in a second designated region of the inner wall of the channel.

  The heat reflection layer and the heat radiation layer may be disposed over the entire inner wall or a part thereof. The heat reflection layer and the heat radiation layer may be arranged only in a specified region of the inner wall. For example, the first designated area of the inner wall is the “ceiling” of the channel, the heat reflecting layer is disposed on the ceiling, the second designated area of the inner wall is the “floor” of the channel, and the heat is applied to the floor. A radiation layer may be disposed, and the heat reflection layer and the heat radiation layer may not be disposed on the “side wall” of the channel. Further, the heat reflecting layer is not disposed in the peripheral part of the “ceiling” (region close to the “side wall”), or the heat radiation is provided in the peripheral part of the “floor” (region close to the “side wall”). It is good also as a structure which does not arrange | position a layer.

  The “ceiling”, “floor” and “side wall” constituting the channel may be made of the same material. For example, when these are all formed of single crystal silicon, an electronic circuit or an electric wiring layer is formed on the front and back surfaces of the interposer, or on the “ceiling” and the “floor” by using a well-known integrated circuit manufacturing technique. It can be formed easily.

  The “ceiling”, “floor” and “side wall” constituting the channel may be made of different materials. For example, the “ceiling” and “floor” may be made of single crystal silicon, and the “side wall” may be made of glass. In this configuration example, the single crystal silicon and the glass can be tightly adhered to each other in an airtight manner (so that fluid does not leak) by electrostatic bonding. The “ceiling” and “floor” are made of a resin material (for example, a printed circuit board material), and the “side wall” is made of a highly adhesive resin material (for example, “SU-8” which is a photoresist material). Also good. The “ceiling”, “floor”, and “side wall” are preferably formed of a material having high thermal conductivity, but it is not always necessary.

  (12) Electronic circuits or electrical wiring layers may be disposed on the front and back surfaces of the interposer (surfaces facing the first and second semiconductor devices, respectively). Electronic circuits or electrical wiring layers may also be arranged on the inner walls (“ceiling”, “floor”, and “side wall”) of the channel. In this case, since the heat reflection layer and the heat radiation layer do not hinder the operation of the electronic circuit (for example, short circuit of the electronic circuit), between the heat reflection layer and the electronic circuit and the heat radiation It is necessary to dispose an insulating layer between the layer and the electronic circuit.

  A penetration electrode may be formed in the interposer. By forming a through electrode, an electrical signal from the second semiconductor device arranged above the interposer (on the side opposite to the substrate) is arranged below the interposer (on the substrate side). Transmission to the first semiconductor device, the substrate, and the like becomes possible.

  The channel may not have a “ceiling”. In this case, the facing surface of the second semiconductor device disposed on the surface of the interposer (the surface opposite to the substrate) is a “ceiling”. The channel may not have a “floor”. In this case, the facing surface of the first semiconductor device disposed on the back surface (surface on the substrate side) of the interposer becomes a “floor”. In these configuration examples, the “ceiling” or “floor” of the interposer is covered with the front surface or the back surface of the first semiconductor device.

  (13) There are many variations on the shape of the channel.

  The cross-sectional shape of the channel (cross-sectional shape in a plane perpendicular to the fluid flow direction) is preferably a substantially rectangular shape including a substantially square shape (for example, a rectangular shape having a large width direction and a small height direction). The shape of the channel in a plane parallel to the substrate is preferably a substantially rectangular shape including a substantially square shape, but is not limited thereto. For example, both the vicinity region of the inlet (upstream side) where the fluid flows into the channel and the vicinity region of the outlet (downstream side) where the fluid flows out, or either the vicinity region of the inlet or the vicinity of the outlet narrows. It can be made into the shape which is. That is, the shape is “at least one end of the channel is constricted, and the center is thick”. The narrowed region is formed by bending at least one of the side walls of the channel inward, and a through electrode may be disposed at the bent portion. The position where the inlet and the outlet are arranged is preferably the center in the width direction of the channel, but is not limited thereto. For example, the inlet may be disposed to the left in the channel width direction, and the outlet may be disposed to the right in the channel width direction.

  (14) The number of the first and second semiconductor devices sandwiching the interposer may be two or more. For example, a plurality of semiconductor devices may be arranged in the same plane on the front or back surface of the interposer.

  Two or more interposers may be used, and the configuration of the stacked module may include a multi-stage semiconductor device. For example, the interposer is mounted on the surface of the first semiconductor device, the second semiconductor device is mounted on the surface of the interposer, an additional interposer is mounted on the surface of the second semiconductor device, and the additional interposer It is good also as a structure which mounts a 3rd semiconductor device on the surface. In this case, the channel may be formed in each of the interposer and the additional interposer, or the channel may be formed only in the interposer or the additional interposer.

  (15) The interposer may make a cross-sectional area of the channel specified in the first region smaller than a cross-sectional area of the channel specified in the second region. In this case, the first region having a relatively small cross-sectional area is located on the upstream side, and the second region having a relatively large cross-sectional area is located on the downstream side. In other words, the first region is close to the inlet of the channel and the second region is close to the outlet of the channel. More specifically, the first region is located upstream of the second region (side closer to the channel inlet). The cross-sectional area is the area of the channel in a cross section perpendicular to the fluid flow direction.

  The first region of the channel is a region where the fluid flows into the channel, and the second region is a region where the fluid flows out of the channel. In this configuration, the fluid first flows into the first region having a small cross-sectional area, and is guided to the second region having a large cross-sectional area before flowing out of the channel. As a result, “adiabatic expansion” occurs in the fluid and an endothermic action occurs. Due to this endothermic action, the heat generated in the first and second semiconductor devices arranged on both sides of the interposer is absorbed by the fluid, so that the operating temperature of the first and second semiconductor devices is lowered. Can be made. The heat absorbed by the fluid is released to the outside by the fluid.

  The height of the channel (the distance between the “ceiling” and the “floor”) is preferably constant in terms of manufacturing technology, but is not limited thereto. For example, in the first region where the fluid flows into the channel, the height of the channel is relatively small, and in the second region where the fluid flows out of the channel, the height of the channel is relatively large. Also good. In order to increase the endothermic effect, it is preferable to increase the ratio of the cross-sectional area in the first region to the cross-sectional area in the second region. In order to increase the ratio of the cross-sectional areas, it is preferable to increase both the ratio of the channel width (the distance between the “side walls” of the channel) and the ratio of the height.

  The channel may not have a “ceiling”. In a configuration without a “ceiling”, the lower surface of the second semiconductor device disposed on the upper side of the interposer (the side opposite to the substrate) serves as a “ceiling”. The channel may not have a “floor”. In a configuration without a “floor”, the upper surface of the first semiconductor device disposed below the interposer (on the side of the substrate) serves as a “floor”. In these configurations, the upper surface of the second semiconductor device or the lower surface of the first semiconductor device covers the opening of the channel.

  (16) There are many options for the planar shape of the channel that causes “adiabatic expansion” in the fluid. For example, (a) a region with a constant cross-sectional area and a relatively small cross-sectional area → a region with an increased cross-sectional area → a region with a constant cross-sectional area and a relatively large cross-sectional area, and (b) a constant cross-sectional area. There is no region, and the cross-sectional area increases sequentially (continuously) in the direction in which the fluid flows. In addition, when the region having a relatively small cross-sectional area is changed to a region having a relatively large cross-sectional area, this change is preferably steep, but this is not restrictive. Furthermore, the position where the fluid flows in and the position where it flows out are not particularly limited.

  (17) The number of the channels may be one or more. For example, the two channels may be arranged inside the interposer, and the regions where the cross-sectional areas thereof change may be arranged at different positions. More specifically, an area where the cross-sectional area of the first channel changes is arranged at the lower left of the interposer (when the interposer is viewed from above), and the cross-sectional area of the second channel is A changing area is located in the upper right of the interposer. In such a configuration, the region where the cross-sectional area changes is brought close to a region (hot spot) where the power consumption of the first or second semiconductor device is large.

  In the case of the above-described configuration having two or more channels, it is possible to combine the inlets of the channels into one. For example, the fluid may flow into the plurality of channels at one place, and the fluid may be branched into each of the two or more channels inside the interposer.

  The heat reflecting layer and the heat radiation layer may be disposed on the ceiling and the floor of the channel, respectively. Moreover, you may arrange | position the said heat | fever reflection layer or the said heat radiation layer to the said ceiling or the said floor of the said channel, respectively. The heat reflection layer and the heat radiation layer may be disposed only on designated areas of the ceiling and the floor, respectively, or disposed on the entire surface of the ceiling and the entire surface of the floor, respectively. May be. Furthermore, the heat reflecting layer or the heat radiation layer may be selectively disposed also on the side wall of the channel.

  (18) The signal extraction method (electrical wiring) from the laminated module is not limited at all. For example, there are a configuration using a bonding wire (bonding module) or a configuration using a through electrode (through electrode module). In the configuration using the through electrode, the first and second semiconductor devices disposed on both sides (up and down) of the interposer are electrically connected to each other by the through electrode formed in the interposer. Furthermore, an electronic circuit or an electric wiring layer may be arranged on the front surface or the back surface of the interposer, or both the front surface and the back surface to facilitate the electrical connection.

  (19) In the stacked module, the number of the first and second semiconductor devices sandwiching the interposer may be two or more. For example, a plurality of semiconductor devices may be arranged in the same plane on the front or back surface of the interposer. Two or more interposers may be used to form a stacked module composed of multistage semiconductor devices. For example, the interposer is mounted on the surface of the first semiconductor device, the second semiconductor device is mounted on the surface of the interposer, an additional interposer is mounted on the surface of the second semiconductor device, and the additional interposer The structure which mounts a 3rd semiconductor device on the surface may be sufficient. In this configuration, the channel may be arranged in each interposer, or the channel may be arranged only in a designated interposer.

  The laminated module is usually the bonding module or the through electrode module described above. However, in the configuration in which more semiconductor devices are laminated, the electrical connection using bonding and the electrical connection using the through electrode are mixed. Also good. An example of such a configuration is a configuration in which the bonding module and the through electrode module are “mixed”.

  According to the laminated module and the interposer of the present invention, (a) even when the semiconductor device having a large power consumption is laminated, it is possible to stably operate by suppressing a temperature rise caused by heat generation of the semiconductor device. It is possible to obtain the effect that (b) the heat conduction between the stacked semiconductor devices can be suppressed and the temperature rise of the stacked module can be suppressed.

It is sectional explanatory drawing which shows the mounting structure which mounted the laminated module (bonding module) which electrically connected using the metal fine wire on the board | substrate. It is a section explanatory view showing a mounting structure which mounts a lamination module (penetration electrode module) which electrically connected using a penetration electrode on a substrate. FIG. 1B is a cross-sectional explanatory view showing a mounting structure when a channel-attached interposer is added to the stacked module (bonding module) of FIG. 1A. It is sectional drawing along the AA line of the interposer shown to FIG. 2A. It is a section explanatory view showing a mounting structure at the time of adding an interposer with a channel to a lamination module (penetration electrode module) of Drawing 1B. It is sectional drawing along the BB line of the interposer shown to FIG. 3A. FIG. 3B is a cross-sectional explanatory view showing the mounting structure in which a cover that covers the laminated module (through electrode module) is added to the mounting structure of FIG. 3A and fluid is supplied to the inside. It is sectional drawing along CC line of the mounting structure of FIG. 4A. FIG. 4B is a cross-sectional explanatory view illustrating a first example of a mounting structure in which a weir for partitioning a space between a cover and a laminated module (through electrode module) is added to the mounting structure of FIG. 4A. It is sectional drawing along CC line of the mounting structure of FIG. 5A. FIG. 4B is a perspective view showing a state before the cover is attached to the substrate in the first example of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. 4A. FIG. 4B is a perspective view showing a process of forming a weir in the second example of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. 4A. FIG. 7B is a continuation of FIG. 7A, showing a process for forming the weir in the second example of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. 4A. FIG. 4B is an explanatory cross-sectional view showing a process of forming a weir in the second example of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. 4A. FIG. 8B is a cross-sectional explanatory diagram illustrating a process of forming the weir in the second example of the mounting structure in which the weir for partitioning the space between the cover and the stacked module is added to the mounting structure of FIG. 4B is a perspective view showing a third example of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. 4A, showing a state before the cover is mounted on the substrate. FIG. It is a perspective view which shows the manufacturing method of the 4th example of the mounting structure which added the weir which partitions the space between a cover and a lamination | stacking module to the mounting structure of FIG. 4A. FIG. 10B is a perspective view illustrating a manufacturing method of the fourth example of the mounting structure in which a weir for partitioning the space between the cover and the stacked module is added to the mounting structure of FIG. FIG. 4B is a cross-sectional explanatory view showing a fourth example manufacturing method of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. 4A. FIG. 11B is a cross-sectional explanatory diagram illustrating a manufacturing method of the fourth example of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. It is a perspective view which shows the manufacturing method of the 5th example of the mounting structure which added the weir which partitions the space between a cover and a lamination | stacking module to the mounting structure of FIG. 4A. FIG. 12B is a perspective view showing a manufacturing method of the fifth example of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. FIG. 4B is an explanatory cross-sectional view illustrating a fifth example manufacturing method of a mounting structure in which a weir for partitioning a space between a cover and a laminated module is added to the mounting structure of FIG. 4A. FIG. 14B is a cross-sectional explanatory view showing the manufacturing method of the fifth example of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. FIG. 14B is a cross-sectional explanatory view showing the manufacturing method of the fifth example of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. FIG. 14C is a cross-sectional explanatory view showing the manufacturing method of the fifth example of the mounting structure in which a weir for partitioning the space between the cover and the laminated module is added to the mounting structure of FIG. It is sectional explanatory drawing which shows the 1st example of a structure of the interposer in which a channel has a radiation | emission and reflection structure. FIG. 15 is a cross-sectional explanatory view showing a stacked module (a mounting structure thereof) using the interposer of FIG. 14. It is sectional drawing along the DD line of the interposer shown to FIG. 15A. It is sectional explanatory drawing which shows the 2nd example of a structure of the interposer in which a channel has a radiation | emission and reflection structure. It is a section explanatory view showing the 3rd example of composition of an interposer in which a channel has radiation and reflection structure. It is sectional explanatory drawing which shows the 1st example of the manufacturing method of the interposer in which a channel has a radiation | emission and reflection structure. It is sectional explanatory drawing which shows the 2nd example of the manufacturing method of the interposer in which a channel has a radiation | emission and reflection structure. It is sectional explanatory drawing which shows the 3rd example of the manufacturing method of the interposer in which a channel has a radiation | emission and reflection structure. It is sectional explanatory drawing which shows the 4th example of the manufacturing method of the interposer in which a channel has a radiation | emission and reflection structure. It is a disassembled perspective view of the interposer used for the lamination | stacking module which concerns on 1st Embodiment of this invention which increases the cooling effect using the adiabatic expansion of the fluid. FIG. 21B is a perspective view of the interposer of FIG. 21A. (A) is sectional explanatory drawing which shows the change of the flow-path cross-sectional area of the interposer of FIG. 21A, (b) is the plane explanatory drawing. It is a section explanatory view showing a lamination module concerning a 1st embodiment of the present invention using an interposer (refer to Drawing 21A) which increases a cooling effect using adiabatic expansion of fluid. It is sectional drawing along the EE line of FIG. 22A. It is sectional explanatory drawing which shows the lamination | stacking module which concerns on 2nd Embodiment of this invention using the interposer (refer FIG. 21A) which increases the cooling effect using the adiabatic expansion of a fluid. It is sectional drawing which followed the FF line | wire of FIG. 23A. It is plane explanatory drawing which shows the modification of the interposer (refer FIG. 21A-FIG. 21C) which concerns on 1st and 2nd embodiment of this invention. It is plane explanatory drawing which shows the other modification of the interposer (refer FIG. 21A-FIG. 21C) which concerns on 1st and 2nd embodiment of this invention. FIG. 22 is an explanatory plan view showing still another modified example of the interposer (see FIGS. 21A to 21C) according to the first and second embodiments of the present invention. FIG. 22 is an explanatory plan view showing still another modified example of the interposer (see FIGS. 21A to 21C) according to the first and second embodiments of the present invention. It is sectional explanatory drawing which shows the 1st example of the manufacturing method of the interposer (refer FIG. 21A-FIG. 21C) which concerns on 1st and 2nd embodiment of this invention. It is a perspective view which shows the 1st example of the manufacturing method of the interposer (refer FIG. 21A-FIG. 21C) which concerns on 1st and 2nd embodiment of this invention. It is sectional explanatory drawing which shows the 2nd example of the manufacturing method of the interposer (refer FIG. 21A-FIG. 21C) which concerns on 1st and 2nd embodiment of this invention. It is a perspective view which shows the 2nd example of the manufacturing method of the interposer (refer FIG. 21A-FIG. 21C) which concerns on 1st and 2nd embodiment of this invention. It is a section explanatory view showing the 3rd example of the manufacturing method of the interposer (refer to Drawing 21A-Drawing 21C) concerning the 1st and 2nd embodiments of the present invention. It is a perspective view which shows the 3rd example of the manufacturing method of the interposer (refer FIG. 21A-FIG. 21C) which concerns on 1st and 2nd embodiment of this invention. FIG. 24 is a perspective view showing a state in which a cover is attached to a substrate in the laminated module (see FIG. 23A) according to the second embodiment of the present invention. It is a systematic diagram which shows the fluid supply example in the mounting structure which combines the interposer (refer FIG. 21A-FIG. 21C) which concerns on 1st and 2nd embodiment of this invention with a lamination | stacking module. It is sectional explanatory drawing which shows the mounting structure (and cooling method) of the conventional laminated module.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of a laminated module (laminated semiconductor device) and an interposer according to the invention will be described with reference to the accompanying drawings.

(Configuration example of laminated module)
1A and 1B show the configuration of the laminated module.

  In FIG. 1A, a first semiconductor device 11a and a second semiconductor device 12a are stacked in this order on a substrate 10 made of resin, ceramic, semiconductor, and the like, and bonding wires (metal thin wires) 13 are bonded to connect them. This is an example of the configuration performed. Hereinafter, the laminated module having this configuration is referred to as a “bonding module”. Here, the lower first semiconductor device 11a and the upper second semiconductor device 12a are both formed in a chip shape, but are not limited to a chip shape. The first semiconductor device 11a is fixed to the surface (upper surface) of the substrate 10 with an adhesive 17 or the like. The second semiconductor device 12a is fixed to the upper surface of the first semiconductor device 11a with an adhesive 17 or the like. Electrical connection between these semiconductor devices 11 a and 12 a and the substrate 10 is performed using bonding wires 13.

  In the configuration example of FIG. 1B, as in the bonding module of FIG. 1A, the first semiconductor device 11b and the second semiconductor device 12b are stacked in this order on the substrate 10, but the electrical connection method is different. That is, the front surface and the back surface of the first semiconductor device 11b are electrically interconnected by the through electrode 14 that penetrates the device 11b in the thickness direction, and the electrical connection between the substrate 10 and the first semiconductor device 11b. (And mechanical connection) and electrical connection (and mechanical connection) between the first semiconductor device 11 b and the second semiconductor device 12 b are both performed using the conductive balls 15. Thus, the substrate 10, the first semiconductor device 11b, and the second semiconductor device 12b are electrically connected to each other and mechanically connected (fixed) to each other. Underfillers 16 are filled in the gaps between the substrate 10 and the first semiconductor device 11b and between the first semiconductor device 11b and the second semiconductor device 12b, respectively. This is to increase the mechanical connection strength between the substrate 10 using the balls 15 and the devices 11b and 12b. Hereinafter, the laminated module having the configuration of FIG. 1B is referred to as a “penetrating electrode module”.

  Each of the bonding module of FIG. 1A and the laminated module of FIG. 1B has advantages and disadvantages. That is, since the bonding module simply laminates the two semiconductor devices 11a and 12a on the substrate 10 and performs their electrical wiring with the bonding wires 13, no new technical development is required. However, it is necessary to satisfy the conditions such that the first semiconductor device 11b is larger than the second semiconductor device 12a and that wire bonding from a position with a different height can be accurately performed. There is also a bonding module in which three or more semiconductor devices are stacked to have three or more layers by this method. In this configuration example, when a cooling fluid (refrigerant) is flowed around the semiconductor devices 11a and 12a, it is necessary to protect the bonding wire 13 in some form so that the bonding wire 13 is not cut. .

  In the stacked module of FIG. 1B, a technique for embedding the through electrode 14 in the lower first semiconductor device 11b, a manufacturing technique for stably mounting, melting, and re-solidifying a number of conductive balls 15, the substrate 10 and the first semiconductor device A technique for pouring the underfiller 16 into a narrow gap between the first semiconductor device 11b and the second semiconductor device 12b is required. In this configuration, since all of the areas related to the wiring are inside the semiconductor chips 11b and 12b or are covered with the underfiller 16, the stacked module is destroyed even if the cooling fluid described above is flowed around. There is little possibility of being.

(Configuration example of bonding module using channel interposer)
2A is a cross-sectional view showing a configuration example (stacked module) in which the interposer 20 with a channel is mounted on the bonding module shown in FIG. 1A, and FIG. 2B is a cross-sectional view of the interposer 20 along the line AA. is there.

  As shown in FIG. 2A, in this configuration example, the channel-equipped interposer 20 is disposed on the surface (upper surface) of the first semiconductor device 11a disposed on the surface (upper surface) of the substrate 10, and the surface of the interposer 20 is disposed. The second semiconductor device 12a is disposed on the (upper surface). In other words, the second semiconductor device 12a and the first semiconductor device 11a are respectively disposed above and below the interposer 20, and both the devices 12a and 11a sandwich the interposer 20.

  The first semiconductor device 11a is fixed to the surface (upper surface) of the substrate 10 with an adhesive 17 or the like. The interposer 20 is fixed to the surface (upper surface) of the first semiconductor device 11a with an adhesive 17 or the like. The second semiconductor device 12a is fixed to the upper surface of the interposer 20 with an adhesive 17 or the like. The adhesive 17 is preferably a material having a high thermal conductivity (for example, a resin in which a metal filler is kneaded), but is not limited thereto.

  The first semiconductor device 11 a is smaller than the substrate 10. The interposer 20 is smaller than the first semiconductor device 11a. The second semiconductor device 12a is smaller than the interposer 20. Electrical connection between these semiconductor devices 11 a and 12 a and the substrate 10 is performed using bonding wires 13.

  The channel 21 formed inside the interposer 20 is supported by an upper wall 23, a lower wall 24, and side walls 22 on both left and right sides. In other words, the channel 21 is defined by the upper wall 23, the lower wall 24 and the side wall 22. The front end portion (upstream end portion) and the rear end portion (downstream end portion) of the channel 21 are open, and therefore the front wall and the rear wall of the channel 21 do not exist. Cooling fluid flows in the channel 21 in the direction indicated by the arrow 25 in FIG. 2B and in FIG. 2A in the direction perpendicular to the paper surface (from front to back, from the front end to the rear end). Yes. In this configuration, the heat generated in the first and second semiconductor devices 11a and 12a is absorbed by the fluid flowing through the channel 21 and discharged to the outside of the stacked module. Such a flow of fluid can be easily realized by a pump, a compressor, and piping (both not shown), as will be described later.

  The fluid for cooling is also called “refrigerant”, and has a characteristic of absorbing heat from a heat generating object and transferring it to the outside. Examples of the fluid include (1) chlorofluorocarbons and non-fluorocarbons (used frequently and many types), (2) organic compounds such as butane and isobutane, (3) hydrogen, helium, ammonia, water, carbon dioxide Inorganic compounds such as can be used.

  It is sufficient that the channel 21 is supported by at least the side walls 22 on both sides, and at least one of the upper wall 23 and the lower wall 24 may be omitted.

  The upper wall 23 and the lower wall 24 of the channel 21 are preferably formed of a material having a high thermal conductivity, but this is not necessarily limited thereto. Further, the upper wall 23 and the lower wall 24 are made of single crystal silicon, and electronic components and transistors are formed on the outer surfaces of the upper wall 23 and the lower wall 24 (that is, the lower surface and the upper surface of the upper wall 23 and the upper surface and the lower surface of the lower wall 24). An electronic circuit composed of the above or an electric wiring layer may be disposed. When electronic circuits or electrical wiring layers are formed on the surfaces exposed to the channels 21 (that is, the lower surface of the upper wall 23 and the upper surface of the lower wall 24), an insulating layer (not shown) is formed on the outermost layer of these surfaces. It is preferable to protect the electronic circuit or electrical wiring layer by preventing erosion or contamination of the electronic circuit or electrical wiring layer by the cooling fluid.

(Configuration example of through electrode module using channel interposer)
3A is a cross-sectional view showing a configuration example (laminated module) in which the interposer 30 with a channel is mounted on the through electrode module shown in FIG. 1B, and FIG. 3B is a cross-sectional view of the interposer 30 along the line BB. It is.

  As shown in FIG. 3A, in this configuration example, the interposer 30 with the channel is disposed on the surface (upper surface) of the first semiconductor device 11b disposed on the surface (upper surface) of the substrate 10, and the surface of the interposer 30 is disposed. The second semiconductor device 12b is disposed on the (upper surface). In other words, the second semiconductor device 12b and the first semiconductor device 11b are respectively disposed above and below the interposer 30, and both the devices 12b and 11b sandwich the interposer 30.

  A through electrode 36 is embedded in the left and right side walls 32 of the interposer 30 so as to penetrate the side wall 32 in the thickness direction. These through electrodes 36 are electrically connected (and mechanically connected) to the electronic circuit of the second semiconductor device 12b via the conductive balls 37, and are connected to the electronic circuit of the first semiconductor device 11b via the conductive balls 38. Electrical connection (and machine connection). That is, the electronic circuits of the first semiconductor device 11b and the second semiconductor device 12b are electrically connected to each other using the through electrode 36 and mechanically connected (fixed) to each other. Thus, the substrate 10, the first semiconductor device 11b, and the second semiconductor device 12b are electrically connected to each other and mechanically connected (fixed) to each other. In addition, the electrical connection and the mechanical connection between the board | substrate 10 and the 1st semiconductor device 11b are performed using the conductive ball 15 similarly to the case of FIG. 1B.

  Underfillers 16 are filled in gaps between the substrate 10 and the first semiconductor device 11b, between the first semiconductor device 11b and the interposer 30, and between the interposer 30 and the second semiconductor device 12b, respectively. This is to increase the mechanical connection strength between the substrate 10, the interposer 30, and the devices 11b and 12b using the balls 15, 37, and 38.

  The through electrode 36 of the interposer 30 cannot be formed on the upper wall 33 and the lower wall 34, but can be formed only on the left and right side walls 32. In other words, the electrical connection points (corresponding to bonding pads) arranged in the semiconductor devices 11 b and 12 b need to be arranged in a portion facing either the left or right side wall 32 of the interposer 30.

  Unlike the configuration example of the bonding module, the first semiconductor device 11b, the interposer 30, and the second semiconductor device 12b are almost the same size and are smaller than the substrate 10. Here, the sizes (areas) of the first semiconductor device 11b, the interposer 30, and the second semiconductor device 12b constituting the stacked module 40 are all equal, but the present invention is not limited to this. For example, the interposer 30 may be smaller than the first semiconductor device 11 b and the second semiconductor device 12 b may be smaller than the interposer 30.

  The channel 31 formed inside the interposer 30 is supported by an upper wall 33, a lower wall 34, and side walls 32 on both the left and right sides. In other words, the channel 31 is defined by the upper wall 33, the lower wall 34, and the side wall 32. The front end portion (upstream end portion) and the rear end portion (downstream end portion) of the channel 31 are open, and therefore the front wall and the rear wall of the channel 31 do not exist. A cooling fluid (refrigerant) flows in the channel 31 in the direction indicated by the arrow 35 in FIG. 3B and in FIG. 3A in the direction perpendicular to the paper surface (from front to back, from the front end to the rear end). It has become. In this configuration, the heat generated in the first and second semiconductor devices 11b and 12b is absorbed by the fluid flowing through the channel 31 and discharged to the outside of the stacked module. Such a fluid flow can be easily realized by a pump, a compressor, and piping (both not shown). As the fluid, those described above in the configuration example of the bonding module using the interposer with the channel can be used.

  It is sufficient that the channel 31 is supported by at least the side walls 32 on both sides, and at least one of the upper wall 33 and the lower wall 34 may be omitted.

  The upper wall 33 and the lower wall 34 of the channel 31 are preferably formed of a material having a high thermal conductivity, but this is not necessarily limited thereto. Further, the upper wall 33 and the lower wall 34 are made of single crystal silicon, and electronic components and transistors are formed on the outer surfaces of the upper wall 33 and the lower wall 34 (that is, the lower surface and the upper surface of the upper wall 33 and the upper surface and the lower surface of the lower wall 34). An electronic circuit composed of the above or an electric wiring layer may be disposed. When electronic circuits or electrical wiring layers are formed on the surfaces exposed to the channel 31 (that is, the lower surface of the upper wall 33 and the upper surface of the lower wall 34), an insulating layer (not shown) is formed on the outermost layer of these surfaces. It is preferable to protect the electronic circuit or electrical wiring layer by preventing erosion or contamination of the electronic circuit or electrical wiring layer by the cooling fluid.

  The underfiller 16 filled in the gaps between the substrate 10, the first semiconductor device 11b, the interposer 30, and the second semiconductor device 12b is made of a material having a high thermal conductivity (for example, a resin kneaded with a metal filler). Although it is preferable, it is not this limitation.

(Configuration example for cooling a through electrode module using an interposer with a channel)
4A is a cross-sectional view showing a configuration example (mounting structure) in the case where the through electrode module 40 (having the interposer 30 with a channel) 40 shown in FIG. 3B is cooled using a fluid, and FIG. It is sectional drawing along C line.

  As described above, the stacked module 40 includes the first semiconductor device 11b arranged on the surface of the substrate 10, the interposer 30 arranged on the surface of the first semiconductor device 11b, and the second semiconductor device arranged on the surface of the interposer 30. The semiconductor device 12b is a main component, and a channel 31 is formed inside the interposer 30.

  On the substrate 10, a cover 42 that encloses the entire laminated module 40 is closely attached and fixed, and an internal space 50 is formed between the substrate 10 and the cover 42. The cover 42 is provided with an inlet 43 for introducing the cooling fluid L into the internal space 50 and an outlet 44 for discharging the fluid L from the internal space 50. A leg 45 is formed at the lower end of the cover 42, and a mounting portion 46 is formed at a position corresponding to the leg 45 on the surface of the substrate 10. The cover 42 is closely attached to the mounting portion 46. It is fixed to the substrate 10 by being fixed.

  The fluid L flows from the inlet 43 into the internal space 50 in the direction indicated by the arrow 47a by the pump P (or compressor) provided outside the cover 42, flows out in the direction indicated by the arrow 47b from the outlet 44, and flows into the pump P. It comes to return. T1 is a pipe that connects the inlet 43 and the pump P, and T2 is a pipe that connects the outlet 44 and the pump P.

  The fluid L introduced into the internal space 50 through the inlet 43 passes through the channel 31 of the interposer 30 from the upstream end to the downstream end along the path indicated by the arrow 48, while the first semiconductor is in the meantime. It is expected to absorb the heat generated in the device 11b and the second semiconductor device 12b. However, in the internal space 50, in addition to such an expected flow of the fluid L, an unexpected flow of the fluid L indicated by an arrow 49 along the periphery of the stacked module 40 also occurs. The cross-sectional area of the channel 30 (particularly the height of the channel 30) is small (for example, about several hundred micrometers), while the gap (cross-sectional area) between the stacked module 40 and the cover 42 is considerably larger ( For example, the flow rate of the flow along the arrow 48 is small, and the flow rate of the flow along the arrow 49 is larger. If there is such a magnitude relationship between the flow rates, there is a high possibility that a desired endothermic effect through the fluid L flowing through the channel 31 cannot be obtained. Therefore, in order to reliably obtain a desired cooling effect for the laminated module 40, a configuration (see FIGS. 4A and 4B) in which the cover 42 covering the entire laminated module 40 is mounted on the substrate 10 is not sufficient. It is necessary to take some measures.

(First example of a mounting structure in which weirs are added to the configuration example for cooling the through electrode module)
5A, 5B, and 6 show a first example of a stacked module mounting structure in which a weir is added to the mounting structure of FIGS. 4A and 4B. 5A is a longitudinal sectional view along the flow direction of the cooling fluid L of this mounting structure, and FIG. 5B is a sectional view taken along the line CC of FIG. 5A. FIG. 6 is a perspective view showing a state in which the cover 42 of this mounting structure is removed from the substrate 10.

  The first example of the stacked module mounting structure is the same as the configuration example shown in FIGS. 4A and 4B (using a through electrode module), and the internal space 50 is located between the upstream space and the downstream between the stacked module 40 and the cover 42. A weir 51 that divides the side space into two is provided. 5A and 5B, the same reference numerals as those shown in FIGS. 4A and 4B indicate the same components.

  The overall shape of the weir 51 is substantially an inverted U-shape, and the path passing through the periphery of the laminated module 40 is blocked by covering the outside of the laminated module 40 in a band shape inside the cover 42. In other words, the weir 51 divides the internal space 50 into two parts: an upstream space on the inlet 43 side and a downstream space on the outlet 44 side. Thus, the generation of an undesired flow of the fluid L passing around the laminated module 40 indicated by the arrow 49 is prevented.

  Since the weir 51 is provided, the upstream space and the downstream space of the internal space 50 are connected to each other only by the channel 31 of the interposer 30, and therefore all of the fluid L flowing into the upstream space from the inlet 43 is It passes through the channel 31 along the path indicated by 48 and reaches the downstream space. Thereafter, it is discharged from the outlet 44. For this reason, an efficient endothermic effect is realized, and a desired cooling effect for the laminated module 40 can be reliably obtained.

  About the position where the weir 51 is arranged and the thickness of the weir 51 (the length along the flow direction of the fluid L), the inlet (upstream end) and the outlet (downstream end) of the channel 41 are blocked. There is no particular limitation except that it is not.

(Second example of a mounting structure in which a weir is added to the configuration example for cooling the through electrode module)
FIGS. 7A and 7B and FIGS. 8A and 8B show a manufacturing method of a second example of a stacked module mounting structure in which weirs are added to the mounting structures of FIGS. 4A and 4B. 7A and 7B are perspective views showing the manufacturing method, and FIGS. 8A and 8B are sectional views thereof. In FIG. 8A and FIG. 8B, the vertical cross section of the mounting structure is drawn on the upper side and the horizontal cross section is drawn on the lower side.

  In the second example of the stacked module mounting structure, as shown in FIGS. 8B (d1) and (d2), the cover 42 of the first example described above is composed of two cover halves 42a and 42b. At the same time, the cover halves 42a and 42b are fixed on the substrate 10 with a predetermined gap G therebetween. As in the first example, the weir 51 is substantially U-shaped and covers the outside of the laminated module 40 in a band shape, but is disposed in the gap G between the cover halves 42a and 42b. Is exposed to the outside from the cover 42 (cover halves 42a and 42b). Other configurations are the same as those in the first example.

  Thus, the weir 51 may be exposed to the outside from the cover 42, and the entire weir 51 may not be disposed inside the cover 42.

  Next, a manufacturing method of the second example of this mounting structure will be described.

  First, as shown in FIG. 7A (a) and FIGS. 8A (a1) and (a2), the laminated module 40 is mounted on the surface of the substrate 10. Since the configuration of the laminated module 40 is the same as the through electrode module 40 (having the interposer 30 with a channel) shown in FIGS. 3A and 3B, the description thereof is omitted. Here, the sizes (areas) of the first semiconductor device 11b, the interposer 30 and the second semiconductor device 12b constituting the stacked module 40 are all equal, but the present invention is not limited to this, May be different. For example, the interposer 30 may be smaller than the first semiconductor device 11 b and the second semiconductor device 12 b may be smaller than the interposer 30.

  Next, as shown in FIGS. 7A (b) and 8A (b1) and (b2), an adhesive is applied to the peripheral area of the bottom of the laminated module 40 on the surface of the substrate 10, and the adhesive layer 55 is formed on the substrate. 10 is formed. The adhesive used for the adhesive layer 55 is, for example, an epoxy resin, and its viscosity is set to a value such that the shape is maintained even after application (that is, the thickness does not decrease due to flow).

  Next, as shown in FIGS. 7B (c) and 8B (c1) and (c2), the cover half 42a and the cover half 42b are disposed on the surface of the substrate 10. The cover halves 42a and 42b have a shape obtained by dividing the cover 42 in FIG. 6 into two parts, and an inlet 43 and an outlet 44 are provided in each of them. In a state where the cover halves 42a and 42b are arranged on the surface of the substrate 10, a gap G is formed between the cover halves 42a and 42b. The size (interval) of the gap G is usually several millimeters (mm), but is not limited thereto.

  The leg portions 45 respectively provided on the cover halves 42a and 42b and the attachment portions 46 provided at positions corresponding to the leg portions 45 on the surface of the substrate 10 are joined to each other using an adhesive. Other than an adhesive can be used for this joining. For example, soldering (when the cover halves 42a and 42b and the surface of the mounting portion 46 are both metal), electrostatic bonding (when the cover halves 42a and 42b are metal and the surface of the mounting portion 46 is glass), intermolecular A well-known method such as direct bonding (when the surfaces of the cover halves 42a and 42b and the mounting portion 46 are both silicon crystals) can be used.

  The leg portions 45 of the cover halves 42a and 42b preferably overlap with the end portions of the adhesive layer 55, but this is not restrictive. In the case of overlapping, by pressing the leg 45 against the surface of the substrate 10, the end of the adhesive layer 55 that overlaps the leg 45 is deformed.

  Subsequently, the adhesive of the adhesive layer 55 and the adhesive between the leg portion 45 and the attachment portion 46 are solidified to ensure mechanical strength in the subsequent processes. When the leg portion 45 and the attachment portion 46 are joined by a material or technique other than the adhesive, only the adhesive of the adhesive layer 55 may be solidified.

  Next, as shown in FIG. 7B (d) and FIGS. 8B (d1) and (d2), the resin 56 is poured from the gap G between the cover halves 42a and 42b, and is almost opposite to the outer peripheral surface of the through electrode module 40. A U-shaped weir 51 is formed. An epoxy resin or the like can be used as the resin 56. The gap between the cover halves 42 a and 42 b and the laminated module 40 is completely filled with the resin 56 using the fluidity of the resin 56. By appropriately selecting the viscosity of the resin 56, the resin 56 can remain only in the peripheral region of the gap G so as not to block the inlet and outlet of the channel 31 of the interposer 30.

  Thus, since the weir 51 is formed by pouring the resin 56 into the gap G between the cover halves 42a and 42b, there is an advantage that the formation of the weir 51 is easier than in the case of the first example. .

  Since the pressure of the fluid L flowing from the inlet 43 is applied to the weir 51, if there is a possibility that the resin 56 forming the weir 51 may be deformed in the long term, the resin 56 is sufficiently solidified. It is necessary to make it. As described above, since the weir 51 has a role of insulating (separating) the upstream space on the inlet 43 side and the downstream space on the outlet 44 side, the front side (outside) of the weir 51 is the cover half 42a. And 42 b need to be firmly attached to the inner wall surfaces, and the lower end of the weir 51 must be firmly attached to the adhesive layer 55.

  In the manufacturing process of the weir 51 in the second example of the stacked module mounting structure, the gap G defines the region where the weir 51 is disposed. In general, as shown in FIGS. 7B (d) and 8B (d1) and (d2), the lengths of the cover halves 42a and 42b (lengths along the flow direction of the fluid L) are made equal to each other. The gap G, that is, the weir 51 is disposed at the center of the laminated module 40. However, this arrangement is not necessarily required. For example, the length of the cover half 42a may be shorter than the length of the cover half 42b. In this case, the gap G, that is, the weir 51 is arranged so as to be shifted to the inlet 43 side from the central portion of the laminated module 40.

  7B and 8B only show an example of the arrangement of the inlet 43 and outlet 44 in the cover halves 42a and 42b, and other arrangements are possible. For example, (i) the inlet 43 may be disposed on the upper surface of the cover half body 42a, the outlet 44 may be disposed on the upper surface of the cover half body 42b, and (ii) the inlet 43 may be disposed on the side surface of the cover half body 42a. The outlet 44 may be disposed on the upper surface of the cover half 42b. The area where the inlet 43 and the outlet 44 are arranged is not particularly limited, and is appropriately determined according to the environment where the mounting structure of the second example is arranged (for example, the arrangement state of components on the printed circuit board). Is done.

  Also in the second example of the stacked module mounting structure, the weir 51 blocks the path passing around the stacked module 40. In other words, the weir 51 divides the internal space 50 into two parts, an upstream space on the inlet 43 side and a downstream space on the outlet 44 side. For this reason, the flow of the fluid L that is not expected to pass around the laminated module 40 indicated by the arrow 49 does not occur, and all of the fluid L that flows into the upstream space from the inlet 43 follows the path indicated by the arrow 48. After passing through the channel 31 and reaching the downstream space, it is discharged from the outlet 44. For this reason, as in the case of the first example described above, an efficient endothermic effect is realized, and a desired cooling effect on the laminated module 40 can be reliably obtained.

(Third example of a mounting structure in which a weir is added to the configuration example for cooling the through electrode module)
FIG. 9 shows a third example of the stacked module mounting structure in which weirs are added to the configuration example for cooling the through electrode module.

  This is a stack module 40a having a five-stage configuration (three-stage configuration excluding the interposer) composed of three semiconductor devices and two interposers. The other configurations are the same as those in the first example described above. The same.

  The stacked module 40a includes a first semiconductor device 11b disposed on the surface (upper surface) of the substrate 10, an interposer 30 with a channel 31 disposed on the surface (upper surface) of the first semiconductor device 11b, and the surface of the interposer 30 ( The second semiconductor device 12b disposed on the upper surface), the interposer 30 with another channel 31 disposed on the surface (upper surface) of the second semiconductor device 12b, and the surface (upper surface) of the interposer 30. The third semiconductor device 12b '. In other words, the second semiconductor device 12b and the first semiconductor device 11b are respectively disposed above and below the lower interposer 30, both the devices 12b and 11b sandwich the lower interposer 30, and the upper interposer 30. The third semiconductor device 12b ′ and the second semiconductor device 12b are respectively disposed above and below the structure 30, and both the devices 12b ′ and 12b sandwich the upper interposer 30.

  Since the substantially inverted U-shaped weir 51 is disposed on the outer peripheral surface of the laminated module 40 a, the fluid L can pass only through the channels 31 of the two interposers 30. The fluid L flowing into the internal space 50 from the inlet 43 of the cover 42 flows out of the outlet 44 of the cover 42 to the outside after passing through the channel 31. The weir 51 is entirely inside the cover 42 and is not exposed from the cover 42.

  Note that the channel 31 may not be formed on both of the two interposers 30. For example, the channel 31 may be formed only in the lower interposer 30 or only in the upper interposer 30.

(Fourth example of a mounting structure in which a weir is added to a configuration example for cooling a through electrode module)
FIGS. 10A and 10B and FIGS. 11A and 11B show a manufacturing method of a fourth example of the stacked module mounting structure in which weirs are added to the configuration example for cooling the through electrode module. 10A and 10B are perspective views showing this manufacturing method, and FIGS. 11A and 11B are cross-sectional views thereof. In FIG. 11A and FIG. 11B, the vertical cross section of the mounting structure is drawn on the upper side, and the horizontal cross section is drawn on the lower side.

  In the fourth example of the mounting structure, as shown in FIGS. 11B (d1) and (d2), the cover 52 includes two cover halves 52a and 52b, and the cover halves 52a and 52b The substrate is adhered and fixed on the substrate 10 with a predetermined gap G. This configuration of the cover 52 is similar to the case of the second example described above, but the length of the cover halves 52a and 52b is shorter than that of the second example, so that the cover halves 52a and 52b. Is different from that of the second example in that the gap G is set to a value close to the total length of the laminated module 40. Further, the weir 51 has a substantially inverted U-shape, as in the second example, and covers almost the entire outer surface of the laminated module 40 in a band shape, but the weir 51 is large between the cover halves 52a and 52b. Since it is arranged in the gap G, its length (length in the flow direction of the fluid L) is larger than in the case of the second example described above, and is a value close to the total length of the stacked module 40. Other configurations are the same as those in the first example.

  In the second example described above (see FIGS. 8B (d1) and (d2)), the internal space 50 is divided into an upstream space and a downstream space by a weir 51. In the upstream space, there is an almost inverted U-shaped narrow area between the inner surface of the cover half body 42a and the outer peripheral surface of the laminated module 40, but when the fluid L enters this area, it stagnates there (fluid) L hardly flows). When such stagnation of the fluid L occurs in the internal space 50, the exhaust heat action by the fluid L is hindered, and a desired cooling action of the laminated module 40 cannot be obtained. The same applies to the downstream space. In the fourth example, the difficulty is removed.

  That is, in this fourth example, the gap G between the cover halves 52a and 52b is set to a value close to the total length of the laminated module 40, so that the length of the weir 51 (the length of the fluid L in the flow direction). ) Is also close to the total length of the laminated module 40. For this reason, an almost inverted U-shaped narrow region is hardly formed between the inner surfaces of the cover halves 52 a and 52 b and the outer peripheral surface of the laminated module 40. As a result, the exhaust heat action is not hindered due to the stagnation of the fluid L, and a desired cooling action of the laminated module 40 is obtained.

  Next, a manufacturing method of the fourth example of this mounting structure will be described.

  First, the mounting process of the laminated module 40 shown in FIGS. 10A (a) and 11A (a1) and (a2), and the formation process of the adhesive layer 55 shown in FIGS. 10A (b), 11A (b1) and (b2). Is the same as the second example described above. Therefore, those descriptions are omitted.

  Next, as shown in FIG. 10B (c) and FIGS. 11B (c1) and (c2), the cover halves 52a and 52b are arranged on the surface of the substrate 10. Similarly to the cover halves 42a and 42b in FIG. 7B, the cover halves 52a and 52b have a shape obtained by dividing the cover 52 into two parts, and are provided with an inlet 43 and an outlet 44, respectively. In a state where the cover halves 52a and 52b are in close contact with and fixed to the surface of the substrate 10, a gap G is formed between the cover halves 52a and 52b. The size (interval) of the gap G is set to a value substantially equal to the total length of the laminated module 40.

  The leg portions 45 of the cover halves 52a and 52b and the attachment portion 46 on the surface of the substrate 10 are joined to each other using an adhesive or the like as in the second example.

  The inside of the leg portion 45 preferably overlaps with the end portion of the adhesive layer 55 described above, but this is not restrictive. In the case of overlapping, by pressing the leg 45 against the surface of the substrate 10, the end of the adhesive layer 55 that overlaps the leg 45 is deformed.

  Subsequently, the adhesive of the adhesive layer 55 and the adhesive between the leg portion 45 and the attachment portion 46 are solidified to ensure mechanical strength in the subsequent processes. When the leg portion 45 and the attachment portion 46 are joined by a material or technique other than the adhesive, only the adhesive of the adhesive layer 55 may be solidified. This is the same as in the second example.

  Next, as shown in FIG. 10B (d), FIG. 11B (d1), and (d2), the resin 56 is poured from the gap G between the cover halves 52a and 52b, and is substantially inverted U to the outer peripheral surface of the through electrode module 40. A character-shaped weir 51 is formed. An epoxy resin or the like can be used as the resin 56. The two gaps between the cover halves 52 a and 52 b and the laminated module 40 are completely filled with the resin 56 using the fluidity of the resin 56. The gap G is larger than that in the mounting structure of the second embodiment. However, by appropriately selecting the viscosity of the resin 56, the gap and the outlet of the channel 31 of the interposer 30 are not blocked. The resin 56 can be poured and retained so as to embed the entire G.

  At this time, since the weir 51 covers almost the entire outer peripheral surface of the through electrode module 40, the outer peripheral surface of the through electrode module 40 is not exposed from the cover 52. The outer peripheral surface of the weir 51 is exposed from the gap G of the cover 52.

  As in the second example, since the pressure of the fluid L flowing from the inlet 43 is applied to the weir 51, there is a possibility that the resin 56 forming the weir 51 may be deformed in the long term. It is necessary to solidify the resin 56 sufficiently. As described above, since the weir 51 has a role of insulating (separating) the upstream space on the inlet 43 side and the downstream space on the outlet 44 side, the front side (outside) of the weir 51 is the cover half 42a. And 42 b need to be firmly attached to the inner wall surfaces, and the lower end of the weir 51 must be firmly attached to the adhesive layer 55.

  Also in the manufacturing process of the weir 51 in the fourth example, the gap G defines the region where the weir 51 is disposed.

  10B and 11B show only an example of the arrangement of the inlet 43 and the outlet 44 in the cover halves 52a and 52b, and other arrangements are possible. This is the same as in the case of the second example.

  As described above, in the fourth example of the stacked module mounting structure, the exhaust heat action is not hindered due to the stagnation of the fluid L as in the third example. There is an effect that the cooling effect is higher than in the case of the second example.

  10A and 10B and FIGS. 11A and 11B show a structure in which the weir 51 fills the entire gap G, but the fourth example is not limited to this. For example, by pouring the resin 56 only in the vicinity of both ends of the gap G (that is, the downstream end of the cover half 52a and the upstream end of the cover half 52b), FIG. 13C (d1) As shown in (d2), weirs 51a and 51b may be dispersed and formed only at both ends of the gap G. In this case, most of the outer peripheral surface of the laminated module 40 is exposed to the outside of the cover 52 from between the weirs 51a and 51b.

(Fifth example of a mounting structure in which a weir is added to a configuration example for cooling a through electrode module)
12A and 12B and FIGS. 13A to 13D show a manufacturing method of the fifth example of the stacked module mounting structure in which weirs are added to the configuration example for cooling the through electrode module. 12A and 12B are perspective views showing this manufacturing method, and FIGS. 13A to 13D are sectional views thereof. 13A to 13D, the vertical cross section of the mounting structure is drawn on the upper side, and the horizontal cross section is drawn on the lower side.

  The fifth example of the mounting structure is characterized in that a second cover (outer cover) 57 is attached in addition to the cover 52 that covers the laminated module 40 to form a double cover. That is, as shown in FIG. 13D (e1) and (e2), the cover 52 is composed of two cover halves 52a and 52b as in the fourth example, and the cover halves 52a and 52b The gap G between them is set to a value close to the total length of the laminated module 40. However, the weirs 51a and 51b are formed (dispersed and arranged) only at both ends of the gap G, and most of the outer peripheral surface of the laminated module 40 is outside the cover 52 from the gap G between the weirs 51a and 51b. It differs from the fourth example in that it is exposed.

  A second cover (outer cover) 57 is further mounted on the surface of the substrate 10 so as to cover the cover 52. The second cover 57 includes an inlet 58 on the upstream side and an outlet 59 on the downstream side. In the space between the cover 52 and the second cover 57, the second cooling fluid L2 flows from the inlet 58 toward the outlet 59. The inlet 43 and outlet 44 of the cover 52 pass through the side wall of the second cover 57 and exit to the outside, and the through portions are sealed with resin or the like so that the fluid L and the second fluid L2 do not leak. Has been.

  The arrangement of the inlet 43 and outlet 44 on the cover halves 52a and 52b and the arrangement of the inlet 58 and outlet 59 of the second cover 57 can be arbitrarily changed as necessary.

  Next, the manufacturing method of the fifth example will be described.

  First, the mounting process of the laminated module 40 shown in FIGS. 12A (a) and 13A (a1) and (a2), and the forming process of the adhesive layer 55 shown in FIGS. 12A (b), 13A (b1) and (b2). Since the mounting process of the cover 52 (cover half bodies 52a and 52b) shown in FIGS. 12A (c) and 13B (c1) and (c2) is the same as the fourth example described above, the description thereof is omitted. To do.

  Next, as shown in FIGS. 12B (d) and 13C (d1) and (d2), in the vicinity of the downstream end of the cover half 52a and in the vicinity of the upstream end of the cover half 52b, The resin 56 is poured separately, and two substantially U-shaped weirs 51 a and 51 b are formed on the outer peripheral surface of the through electrode module 40. An epoxy resin or the like can be used as the resin 56. The gap between the cover halves 52a and 52b and the laminated module 40 in the vicinity of the downstream end of the cover half 52a and the upstream end of the cover half 52b uses the fluidity of the resin 56. As a result, the resin 56 completely closes it. Both ends of the gap G are covered with the resin 56 (the weirs 51a and 51b), but the central part remains exposed.

  Finally, as shown in FIGS. 12B (e) and 13D (e1) and (e2), the second cover (outer cover) 57 is attached to the substrate 10 so as to cover the cover 52 (cover halves 52a and 52b). Adhere and fix to the surface. Thus, the fifth example of this mounting structure is completed.

  As described above, in the fifth example of the stacked module mounting structure, the double cover structure is used, and the inner space 50 of the inner cover 52 is divided into an upstream space and a downstream space by the weirs 51a and 51b. While being divided, the cooling fluid L is introduced and discharged from the inlet 43 toward the outlet 44. In addition, the cooling second fluid L <b> 2 is allowed to flow in the space between the cover 52 and the second cover 57 from the inlet 58 toward the outlet 59. There is no weir in this space. For this reason, in addition to the exhaust heat by the cooling fluid L, exhaust heat by the second fluid L2 is simultaneously performed. Therefore, the structure is a little complicated compared to the third and fourth examples described above, but the cooling effect on the stacked module 40 (second semiconductor device 12b and first semiconductor device 11b) is the same as that of the third and fourth examples. There is an effect that it becomes higher than the case.

  The fluid L and the second fluid L2 need not be the same type of fluid. For example, the fluid L may be a liquid and the second fluid may be a gas. As described above, since the fluid L passes through the channel 31 having a small (narrow) cross-sectional area, it is preferable to supply the fluid whose pressure has been increased to the inlet 43 using a compressor, a pump, or the like. On the other hand, since the second fluid L2 passes through a region having a much larger cross-sectional area than the channel 31, a high pressure is not required. For this reason, it is also possible to employ a simpler configuration for the second fluid. For example, the second cover (outer cover) 57 is provided with a single port through which the second fluid L2 flows in and out (this port functions as an inlet / outlet), and a heat pump mounted on a notebook computer. A configuration similar to that (which does not use a compressor) is also possible.

(First example of an interposer configuration in which the channel has a radiation / reflection structure)
FIG. 14 shows a first example of an interposer configuration in which the channel has a radiation / reflection structure. 15A and 15B show a first example of a stacked module (mounting structure thereof) using the interposer of FIG. In addition, although it is set as the penetration electrode module here, it cannot be overemphasized that it is good also as a bonding module. Moreover, when mounting this module on a board | substrate, all of the mounting structure (those which have a board | substrate and a cover) mentioned above are applicable.

  This laminated module is characterized in that the interposer 30a used therein incorporates a heat reflection layer 61a and a heat radiation layer 61b exposed in the channel 31.

  As shown in FIG. 14, a heat reflecting layer 61a is formed on the inner surface of the upper wall 33 of the interposer 30a via an insulating layer 62a, and a heat reflecting layer 61a is formed on the inner surface of the lower wall 34 via an insulating layer 62b. A radiation layer 61b is formed. The heat reflecting layer 61 a covers the entire inner surface of the upper wall 33. The heat radiation layer 61 b covers the entire inner surface of the lower wall 34. Thus, in the channel 31 of the interposer 30a, the upper heat reflection layer 61a and the lower heat radiation layer 61b face each other. For this reason, the upper wall and the lower wall of the channel 31 are defined by the heat reflection layer 61a and the heat radiation layer 61b, respectively.

  A heat source 63 (which corresponds to the second semiconductor device 12b) is brought into contact with the outer surface of the upper wall 33. A heat source 64 (which corresponds to the first semiconductor device 11b) is brought into contact with the outer surface of the lower wall 34.

  The structure of the interposer 30a shown in FIG. 14 is also referred to as a “radiation / reflection structure”.

  The heat reflecting layer 61a is often formed from a metal thin film or the like, and has a function of reflecting heat incident from both above and below. The heat reflecting layer 61a is formed of, for example, a thin film of gold or aluminum, and the surface thereof is preferably a mirror surface, but this is not restrictive. Moreover, although it is preferable that the heat | fever reflection layer 61a is arrange | positioned over the whole inner surface of the upper wall 33, it is not this limitation. For example, the heat reflecting layer 61a may be disposed only in a specified region on the inner surface of the upper wall 33.

  The heat radiation layer 61b is formed of, for example, a heat radiation layer also called “golden black”. “Gold black” is obtained by evaporating gold in an atmosphere having a relatively low degree of vacuum. The surface of “Gold Black” has minute irregularities and looks black when observed with visible light. “Gold black” is known to absorb heat when its surface temperature is lower than ambient temperature, but to radiate heat when its surface temperature is higher than ambient temperature. As an alternative to “golden black”, other materials such as a resin colored in black may be used. Further, the heat radiation layer 61b is preferably disposed over the entire inner surface of the lower wall 34, but this is not restrictive. For example, the heat radiation layer 61b may be disposed only in a specified region on the inner surface of the lower wall 34.

  The upper wall 33 and the lower wall 34 are preferably made of a material having a high thermal conductivity, but this is not necessarily the case. Both the upper wall 33 and the lower wall 34 are made of single crystal silicon, and each surface (the inner surface of the upper wall 33, the inner surface of the lower wall 34) is an electronic circuit or an electric wiring layer composed of electronic components or transistors. Is preferably disposed with insulating layers 62a and 62b interposed therebetween. These insulating layers 62a and 62b are short-circuited by the heat reflection layer 61a (which is conductive because of a metal thin film) or the heat radiation layer 61b (which is conductive when gold is black). It has a role to prevent.

  In FIG. 14, arrows 66a and 66b indicate the direction in which the cooling fluid L flows. The heat radiated from the heat radiation layer 61b into the channel 31 by the fluid L flowing through the channel 31 and the heat reflected into the channel 31 by the heat reflection layer 61a are discharged to the outside of the interposer 30a. Such a flow of the fluid L can be easily realized by using a pump, a compressor, piping, and the like (both not shown). The fluid L is also called “refrigerant” and has a characteristic that it can absorb heat from a heat-generating object and transfer it to the outside. For example, (1) chlorofluorocarbons and non-fluorocarbons (used frequently and many types), (2) organic compounds such as butane and isobutane, (3) inorganic compounds such as hydrogen, helium, ammonia, water, carbon dioxide, etc. There is. This point is the same as described above.

  Here, it is assumed that the lower heat source 64 generates a larger amount of heat than the upper heat source 63. Then, the heat generated by the lower heat source 64 passes through the lower wall 34 and the insulating layer 62b to reach the surface of the heat radiation layer 61b (the surface on the channel 31 side), and the fluid L flowing through the channel 31 on this surface. Radiated towards Most of the heat thus radiated is absorbed by the fluid L. The heat that has reached the heat reflecting layer 61a without being absorbed by the fluid L is reflected toward the fluid L by the surface of the heat reflecting layer 61a (the surface on the channel 31 side), and thus is also absorbed by the fluid L.

  On the other hand, the heat generated by the upper heat source 63 first passes through the upper wall 34 and the inside of the insulating layer 62a and reaches the back surface (the surface opposite to the channel 31 side) of the heat reflecting layer 61a. Reflected toward the heat source 63. The heat that has passed through the heat reflecting layer 61a without being reflected by the heat reflecting layer 61a reaches the surface of the heat reflecting layer 61a (the surface on the channel 31 side) and is absorbed by the fluid L flowing through the channel 31 on this surface.

  As a result, the heat generated by the lower heat source 64 does not reach the upper heat source 63, and the heat generated by the upper heat source 63 does not reach the lower heat source 64. In other words, the upper and lower heat sources 63 and 64 are “thermally insulated” by the interposer 30a. Therefore, for example, even if the lower heat source 64 is a semiconductor device that consumes a large amount of power and the upper heat source 63 is a semiconductor device having heat-sensitive characteristics, the interposer 30a having such a configuration is used for both. By interposing them in between, it is possible to significantly reduce the thermal influence of the lower semiconductor device on the upper semiconductor device.

  A laminated module 40b using the interposer 30a having the configuration shown in FIG. 14 is shown in FIGS. 15A and 15B.

  As shown in FIG. 15A, the stacked module 40b includes a first semiconductor device 11b mounted on the substrate 10, an interposer 30a mounted thereon, and a second semiconductor device 12b mounted thereon. ing.

  A plurality of through electrodes 36 are embedded in the side walls 32 on both sides of the interposer 30a so as to penetrate the side walls 32 in the thickness direction. These through electrodes 36 are electrically connected (and mechanically connected) to the electronic circuit of the second semiconductor device 12b via the conductive balls 37, and are connected to the electronic circuit of the first semiconductor device 11b via the conductive balls 38. Electrical connection (and machine connection). That is, the electronic circuits of the first semiconductor device 11b and the second semiconductor device 12b are electrically connected to each other using the through electrode 36 of the interposer 30a and mechanically connected (fixed) to each other. Thus, the substrate 10, the first semiconductor device 11b, and the second semiconductor device 12b are electrically connected to each other and mechanically connected (fixed) to each other. Electrical connection and mechanical connection between the substrate 10 and the first semiconductor device 11 b are performed using the conductive balls 15. Underfillers 16 are filled in gaps between the substrate 10 and the first semiconductor device 11b, between the first semiconductor device 11b and the interposer 30a, and between the interposer 30a and the second semiconductor device 12b, respectively.

  In FIG. 15A, the heat reflection layer 61a and the heat radiation layer 61b of the interposer 30a are shown, but the insulating layers 62a and 62b are omitted.

  As shown in FIG. 15B, the fluid L flows through the channel 31 of the interposer 30a as shown by the arrow 35. Therefore, in FIG. 15A, the fluid L flows from the near side of the paper toward the back. Since the heat generated in the first semiconductor device 11b is absorbed by the fluid L through the thermal radiation layer 61b, it is released to the outside by the discharge of the fluid L. Since the heat that has passed through the heat radiation layer 61b and reached the heat reflection layer 61a is also reflected and absorbed by the fluid L, it is also released to the outside by the discharge of the fluid L. Moreover, the heat generated in the second semiconductor device 12b is reflected upward by the heat reflecting layer 61a, and is radiated to the outside. Since the heat that has not been reflected by the heat reflecting layer 61a and is transmitted through it is absorbed by the fluid L flowing through the channel 31 via the heat reflecting layer 61a, it is also released to the outside by the discharge of the fluid L.

  The left and right side walls 32 of the interposer 30a define the channel 31 together with the upper wall 33 and the lower wall 34, and also serve as a heat conduction path that connects the first semiconductor device 11b and the second semiconductor device 12b. In this case, a part of the heat generated in the first semiconductor device 11b is transferred to the second semiconductor device 12b via the side wall 32. However, the width of the side wall 32 is reduced, or the side wall 32 is made of a material different from that of the upper wall 33 and the lower wall 34 (preferably a material having a low thermal conductivity). The amount of heat conduction can be reduced. Therefore, it is preferable to do so as much as possible.

  The through electrode 36 formed on the side wall 32 is usually formed of a material having a high thermal conductivity such as metal. In this case, since heat conduction through the through electrode 36 easily occurs, part of the heat generated in the first semiconductor device 11b reaches the second semiconductor device 12b, and the temperature of the second semiconductor device 12b can be increased. There is sex. However, the amount of heat conducted can be reduced by reducing the thickness of the through electrode 36 or by modifying the design of the electrical wiring layer to reduce the number of through electrodes 36. Therefore, it is preferable to do so as much as possible.

  As shown in FIGS. 15A and 15B, the heat reflecting layer 61 a disposed on the inner surface of the upper wall 33 of the channel 31 is not disposed on the entire inner surface of the upper wall 33, and is a peripheral region of the inner surface of the upper wall 33. It is arranged in the part excluding. Similarly, the heat radiation layer 61b disposed on the inner surface of the lower wall 34 of the channel 31 is not disposed on the entire inner surface of the lower wall 34, and is disposed on a portion excluding the peripheral region of the inner surface of the lower wall 34. Yes. However, the arrangement of the heat reflection layer 61a and the heat radiation layer 61b is not limited to the arrangement illustrated in FIGS. 15A and 15B.

  For example, (a) the heat radiation layer 61 b may be disposed on the entire inner surface of the lower wall 34 and the entire inner surface of the side wall 32, and the heat reflecting layer 61 a may be disposed on a designated region on the inner surface of the upper wall 33. This arrangement example is suitable in the case where only the heat generated in the semiconductor device 11b is released to the fluid L is considered. (B) The heat radiation layer 61b may be disposed in a specified region of the lower wall 34, and the heat reflection layer 61a may be disposed on the entire inner surface of the upper wall 33 and the entire inner surface of the side wall 32. (C) The heat radiation layer 61b is disposed on the entire inner surface of the lower wall 34 and in a region close to the lower wall 34 of the left side wall 32, and the heat reflecting layer 61a is disposed on the entire inner surface of the upper wall 33 and the right side wall 32. You may arrange | position to the area | region close to the upper wall 33.

  As is clear from FIG. 15B, the planar shape of the channel 31 of the interposer 30a described above is linear. However, the planar shape of the channel 31 is not limited to this. For example, a planar shape as shown in FIGS. 16A and 16B may be used.

(Second example of an interposer configuration in which the channel has a radiation / reflection structure)
FIG. 16A shows a second example of an interposer configuration in which the channel has a radiation / reflection structure.

  In the interposer 30b, the left and right side walls 32 protrude toward the inside in the upstream and downstream end regions (the upper and lower end regions in FIG. 16A) of the channel 31. That is, the width (cross-sectional area) of the channel 31 is narrower at the upstream and downstream end regions than at the center. As a result, the fluid L flowing into the channel 31 along the arrow 35a flows through the narrow inlet of the channel 31, is guided to the wide central portion, and then flows out through the narrow outlet.

  In this interposer 30b, the area on which the through electrodes 36 can be arranged is increased by pushing the left and right side walls 32 inward, which is particularly effective when the number of through electrodes 36 needs to be increased. .

  Many variations of the shape of the channel 31 are possible. For example, the longitudinal cross-sectional shape of the channel 31 (cross-sectional shape along a plane perpendicular to the substrate 10) is a substantially rectangular shape including a substantially square shape (for example, the length in the fluid flow direction is large and the length in the height direction is long). Is a small rectangle). The horizontal cross-sectional shape of the channel 31 is preferably a substantially rectangular shape including a substantially square shape, similar to the vertical cross-sectional shape, but is not limited thereto. For example, both the inlet region where the fluid L flows into the channel 31 and the outlet region where the fluid L flows out, or either the inlet region or the outlet region of the channel 31 may be narrowed. In this narrowed region, the side wall 32 is bent, but the through electrode 36 may be disposed in the bent region. This example is the configuration of the second embodiment.

  The arrangement region of the inlet and outlet of the channel 31 is preferably located at the center in the width direction of the channel 31, but is not limited thereto. For example, the entrance may be disposed on the left side of the central portion in the width direction, and the exit may be disposed on the right side. This example is the configuration of a third embodiment described later shown in FIG. 16B.

  In the interposer 30b shown in FIG. 16A, the channel 31 is constricted at the entrance end and the exit end, and the center of the channel 31 is thicker than the entrance and exit ends (the cross-sectional area is large). It has become. In such a configuration, the two requirements of increasing the area of the heat reflection layer 61a and the heat radiation layer 61b to increase the heat dissipation effect and responding to the increase in the total number of through electrodes 36 are satisfied at the same time. Can be made.

(Third example of an interposer configuration in which the channel has a radiation / reflection structure)
FIG. 16B shows a third example of the configuration of the interposer in which the channel has a radiation / reflection structure.

  In this interposer 30c, the left side wall 32 protrudes inward in the end region on the downstream side of the channel 31 (the upper end region in FIG. 16B), and the right side wall 32 is the end region on the upstream side of the channel 31. (In FIG. 16B, the area is at the lower end) and protrudes inward. That is, a bent portion that is bent stepwise is formed at the center of the channel 31. As a result, the fluid L flowing into the channel 31 along the arrow 35a flows through the narrow inlet of the channel 31, passes through the central bent portion, and then flows out through the narrow outlet. For this reason, the fluid L seems to flow through the channel 31 in an oblique direction.

  In this interposer 30c, the area on which the through electrodes 36 can be arranged is increased by pushing the left and right side walls 32 inward, which is particularly effective when the number of through electrodes 36 needs to be increased. .

  The shape of the channel 31 can be modified in the same way as described in the second embodiment.

  Also in the interposer 30c shown in FIG. 16B, the channel 31 is confined at the entrance end and the exit end, and the center of the channel 31 is thicker than the entrance and exit ends (the cross-sectional area is large). It has become. In such a configuration, the two requirements of increasing the area of the heat reflection layer 61a and the heat radiation layer 61b to increase the heat dissipation effect and responding to the increase in the total number of through electrodes 36 are satisfied at the same time. Can be made.

(First example of manufacturing method of interposer with channel having radiation / reflection structure)
FIG. 17 shows a first example of a method for manufacturing an interposer 30a having a channel having a radiation / reflection structure. Since this manufacturing method can be applied to both the above-described second and third example interposers 30b and 30c, the manufacturing method of the interposer 30a will be described here.

  First, as shown in FIG. 17A, the heat reflecting layer 61 a is formed in a predetermined region on the inner surface (lower surface in the drawing) of the plate-like upper wall 33. The upper wall 33 is preferably made of a material having a high thermal conductivity, but this need not be the case. The heat reflection layer 61a is formed by using a metal such as gold or aluminum using a method such as a vapor deposition method.

  Next, as shown in FIG. 17B, the heat radiation layer 61 b is formed in a predetermined region on the inner surface (upper surface in the drawing) of the plate-like lower wall 34. The heat radiation layer 61b is formed of, for example, a “gold-black” vapor deposition film.

  Here, the heat reflection layer 61a and the heat radiation layer 61b are selectively formed only in the central region without covering the entire inner surface of the upper wall 33 and the inner surface of the lower wall 34, respectively. It is not limited. The entire inner surface of the upper wall 33 and the entire inner surface of the lower wall may be covered.

  Next, as shown in FIG. 17C, strip-shaped side walls 32 are formed in the left and right peripheral regions of the lower wall 34 so as not to overlap the heat radiation layer 61b. The material of the side wall 32 is, for example, resin, and is fixed to the inner surface (upper surface in the drawing) of the lower wall 34 with an adhesive or the like.

  Finally, as shown in FIG. 17D, the corresponding portions of the inner surface of the upper wall 33 are fixed to the upper surfaces of the left and right side walls 32 using an adhesive or the like. Thus, the interposer 30a having the channel 31 inside is manufactured. The fluid L flows through the inside of the channel 31 from the front to the back (or from the back to the front) of the page.

  The adhesive or the like used in the first example of the manufacturing method is selected in consideration of the adhesion with the object to be bonded so that the airtightness of the channel 31 is maintained over a long period of time. For example, when the upper wall 33 and the lower wall 34 are formed of single crystal silicon and the side wall 32 is formed of glass, the bonding between the lower wall 34 and the side wall 32 and the bonding between the side wall 32 and the upper wall 33 are performed statically. It is possible to use an electrojoining technique. Since the sealing performance of the electrostatic junction is sufficiently high, the airtightness of the channel 31 over a long period is maintained. Further, since the upper wall 33 and the lower wall 34 are made of single crystal silicon, it is possible to form one or more wiring layers on the surfaces of the upper wall 33 and the lower wall 34 using a well-known technique.

  Although not shown in FIG. 17, the through electrode 36 may be formed in the peripheral region of the interposer 30a. For example, using a processing technique such as RIE (reactive ion etching), a plurality of holes penetrating the upper wall 33 (for example, single crystal silicon), the side wall 32 (for example, glass), and the lower wall 34 (for example, single crystal silicon) are formed. The through electrode 36 can be formed by forming and attaching or filling a conductive material inside the through holes with an insulating film interposed as required.

(Second example of manufacturing method of interposer with channel having radiation / reflection structure)
FIG. 18 shows a second example of the manufacturing method of the interposer 30a in which the channel has a radiation / reflection structure.

  First, as shown in FIG. 18A, a heat reflecting layer 61a is formed in a predetermined region on the inner surface (lower surface in the figure) of the plate-like upper wall 33. The upper wall 33 is preferably made of a material having a high thermal conductivity, but this need not be the case. The heat reflection layer 61a is formed using a technique such as vapor deposition of a metal such as gold or aluminum. This step is the same as the first example of the manufacturing method described above.

  Next, as shown in FIG. 18B, the lower wall 34 and the side wall 32 of the interposer 30a are integrally formed. The side walls 32 are located in the left and right peripheral regions of the lower wall 34, respectively. A recess 34 a is formed at a location surrounded by the lower wall 34 and the side wall 32. Then, the heat radiation layer 61b is formed in a predetermined region (which is in the recess 34a) on the inner surface (the upper surface in the drawing) of the lower wall 34. The heat radiation layer 61b is formed of, for example, a “gold-black” vapor deposition film. This step is different from the first example of the manufacturing method described above.

  Such a configuration can be easily realized by selectively removing the central region of the lower wall 34 to form the depression 34a and then forming the heat radiation layer 61b on the bottom surface of the depression 34a. A specific example of the processing method is as follows. That is, first, a patterned photoresist layer is formed on the upper surface of a single crystal silicon plate (the thickness of which is equal to the thickness of the side wall 32). Next, using the photoresist layer as a mask, the single crystal silicon plate is selectively removed to a predetermined depth by RIE or the like. Thus, a recess 34a is formed in the central region of the single crystal silicon plate. Finally, using a vapor deposition method or the like, a gold black layer is selectively formed on the bottom surface of the recess 34a to form the heat radiation layer 61b.

  Here, the heat radiation layer 61b is formed only in the central region without covering the entire bottom surface of the recess 34a, but is not limited thereto. For example, the heat radiation layer 61b may cover the entire bottom surface of the recess 34a, may cover the entire bottom surface of the recess 34a and part of the inner surfaces of the side walls 32, or the entire bottom surface of the recess 34a. And the entire inner surface of both side walls 32 may be covered.

  Finally, as shown in FIG. 18C, in the peripheral region of the inner surface of the upper wall 33 (which has the heat reflection layer 61a) shown in FIG. When the upper surfaces of the side walls 32 of the configuration are fixed, an interposer 30a having a channel 31 therein is manufactured. For this fixing, an epoxy adhesive 67 or the like can be used.

  When both the upper wall 33 and the lower wall 34 having the side wall 32 are formed of single crystal silicon and the heat radiation layer 61b is not formed on the upper end surface of the side wall 32 (this is the case in the second example) In other words, the upper wall 33 and the lower wall 34 having the side wall 32 can be bonded by silicon-silicon thermal bonding.

  Since the channel 31 needs to maintain airtightness, the above-described adhesive or the like or the thermal bonding process is appropriately selected so that the desired airtightness of the channel 31 can be maintained.

  In the second example shown in FIG. 18, it is assumed that single crystal silicon is used for the upper wall 33 and the lower wall 34 having the side walls 32. For this reason, it is possible to form the through electrode 36 on the side wall 32 using a known processing technique, and it is also possible to form one or more electric wiring layers on the upper wall 33 and the lower wall 34, respectively. The material of the upper wall 33 and the lower wall 34 having the side walls 32 is not limited to single crystal silicon, and a resin or the like may be used.

  Note that the structure shown in FIG. 18B can be used as it is as the interposer 30a. In this case, there is no ceiling portion that defines the channel 31, but the role of this ceiling portion is performed by the lower surface of the semiconductor device 12b mounted above the interposer 30a. The heat reflection layer 61a is formed on the lower surface of the semiconductor device 12b.

  Similarly, the interposer 30a in which the floor portion defining the channel 31 does not exist can be obtained by turning the configuration of FIG. 18B upside down.

(Third example of manufacturing method of interposer with channel having radiation / reflection structure)
FIG. 19 shows a third example of the manufacturing method of the interposer 30a in which the channel has a radiation / reflection structure.

  First, as shown in FIG. 19A, a plate-shaped base material 34 ′ (here, a single crystal silicon substrate) to be the lower wall 34 is prepared.

  Next, as shown in FIG. 19B, a plurality of through electrodes 36 are formed in the left and right peripheral regions of the base material 34 ′ by a known method. More specifically, a plurality of through holes penetrating in the thickness direction are formed in the base material 34 ′ by etching by the RIE method, and the inside of these through holes is covered with an insulating layer (not shown). After that, a conductive material is coated or filled inside these insulating layers. Thus, by providing conductivity to each through hole, the through electrode 36 can be easily obtained. Thereafter, electrical connection pads 68 and 69 are formed on the front surface and the back surface of the base material 34 ′, respectively, and are electrically connected to the corresponding through electrodes 36.

  Next, as shown in FIG. 19C, the base material 34 ′ is selectively removed from the surface side by using a patterned mask (not shown; this is made of a photoresist film or the like). The depression 34a is formed. In the step of forming the recess 34a, for example, an anisotropic etching solution such as TMAH (tetramethylammonium hydroxide) or KOH (potassium hydroxide) is used. In these etching solutions, the etching rate is relatively low on a specific silicon crystal plane (that is, the (111) plane), so that the slope of the edge of the recess 34a (that is, the bottom surface of the recess 34a to the surface of the base material 34 '). ) Is the exposed (111) plane. Thus, the lower wall 34 in which the left and right side walls 32 are integrated is obtained.

  Subsequently, the heat radiation layer 61b is formed on the bottom surface of the recess 34a of the lower wall 34 by a known method. The state at this time is as shown in FIG.

  On the other hand, as shown in FIG. 19D, the upper wall 33 having a plurality of openings 33a for exposing the electrical connection pads 68 is formed. The upper wall 33 is easily formed by etching a plate-like base material (here, a single crystal silicon substrate) from above with an RIE method. Thereafter, the heat reflecting layer 61a is formed in a predetermined region on the inner surface (the lower surface in the drawing) of the upper wall 33.

  Finally, on the lower wall 34 shown in FIG. 19C (which has the through electrode 36, the electrical connection pad 68, and the heat radiation layer 61b), the upper wall 33 shown in FIG. When this is placed and fixed, the interposer 30a having the structure shown in FIG. 19E is obtained. For this fixing, an adhesive, low melting point glass or the like is used. A channel 31 is formed by the depression 34 a of the lower wall 34. The channel 31 needs to be airtight except for its inlet and outlet.

  In the third example, since both the upper wall 33 and the lower wall 34 of the interposer 30a are formed of a single crystal silicon substrate, it is possible to form one or more wiring layers on the front and back surfaces of the interposer 30a. .

  Note that the structure shown in FIG. 19C can be used as it is as the interposer 30a. In this case, there is no ceiling portion that defines the channel 31, but the role of this ceiling portion is performed by the lower surface of the semiconductor device 12b mounted above the interposer 30a. The heat reflection layer 61a is formed on the lower surface of the semiconductor device 12b.

  Similarly, by interposing the configuration of FIG. 19B upside down, the interposer 30a without the floor portion that defines the channel 31 can be obtained.

(Fourth example of manufacturing method of interposer with channel having radiation / reflection structure)
FIG. 20 shows a fourth example of the manufacturing method of the interposer 30a in which the channel has a radiation / reflection structure.

  20A to 20C, the manufacturing process of the lower wall 34 (which includes the through electrode 36, the electrical connection pads 68 and 69, and the heat radiation layer 61b) is the same as the above-described third example. Therefore, the description thereof is omitted.

  On the other hand, as shown in FIG.20 (d), the upper wall 33 made into the shape fitted in the hollow 34a of the lower wall 34 is formed. The upper wall 33 is easily formed by etching a plate-like base material (here, a single crystal silicon substrate) from the lower side thereof by the RIE method. Thereafter, the heat reflecting layer 61a is formed in a predetermined region on the inner surface (the lower surface in the drawing) of the upper wall 33.

  Finally, as shown in FIG. 20D, the lower wall 34 shown in FIG. 20C is fitted into the recess 34a of the through electrode 36, the electrical connection pads 68 and 69, and the heat radiation layer 61b. When the upper wall 33 (having the heat reflection layer 61a) shown in FIG. 20 is placed and fixed, an interposer 30a having the structure shown in FIG. 20 (e) is obtained. An adhesive or low melting point glass is used for this fixing. In this state, a gap is formed between the heat reflecting layer 61 a of the upper wall 33 and the heat radiation layer 61 b of the lower wall 34 inside the recess 34 a, which becomes the channel 31. The channel 31 needs to be airtight except for its inlet and outlet.

  Also in the fourth example, since both the upper wall 33 and the lower wall 34 of the interposer 30a are formed of a single crystal silicon substrate, it is possible to form one or more wiring layers on the front and back surfaces of the interposer 30a. .

  In the fourth example, unlike the third example shown in FIG. 19, almost the entire upper wall 33 is configured to fit into the recess 34 a of the lower wall 34, and the upper part of the left and right side walls 32. Is not covered with the upper wall 33. For this reason, it is not necessary to provide the opening 33a as in the third example on the upper wall 33. Therefore, there is an advantage that the manufacturing technique is simpler than that in the third example shown in FIG.

  Note that the structure shown in FIG. 20C can be used as it is as the interposer 30a. In this case, there is no ceiling portion that defines the channel 31, but the role of this ceiling portion is performed by the lower surface of the semiconductor device 12b mounted above the interposer 30a. The heat reflection layer 61a is formed on the lower surface of the semiconductor device 12b.

  Similarly, by interposing the configuration of FIG. 20C upside down, the interposer 30a without the floor portion that defines the channel 31 can be obtained.

(Modification example of a laminated module using an interposer whose channel has a radiation / reflection structure)
Up to this point, the interposers 30a, 30b, and 30c having the heat reflecting layer 61a and the heat radiation layer 61b and the laminated module 40b incorporating the interposers have been described. In the laminated module 40b, there are various types other than those exemplified above. Variations are possible.

  For example, when the power consumption of the upper second semiconductor device 12b is larger than the power consumption of the lower first semiconductor device 11b, the positions of the heat reflection layer 61a and the heat radiation layer 61b are switched. In other words, the heat radiation layer 61b is disposed on the second semiconductor device 12b side where the power consumption is large, and the heat reflection layer 61a is disposed on the first semiconductor device 11b side where the power consumption is small. The stacked module may have a configuration in which three or more layers of semiconductor devices are stacked, or two or more upper and lower semiconductor devices sandwiching the interposer (for example, a plurality of semiconductor devices in the same plane on the upper surface of the interposer). A configuration in which devices are arranged in parallel) may be used. In any configuration, the heat reflection layer 61a and the heat radiation layer 61b provide thermal insulation between the semiconductor devices above and below the interposer, and release heat by releasing heat from the heat radiation layer 61b to the fluid L. Getting an effect is common.

(Laminated module according to the first embodiment and interposer used therefor)
FIGS. 21A to 21C illustrate an interposer used in the stacked module according to the first embodiment of the present invention to increase the cooling effect of a semiconductor device using adiabatic expansion of a fluid. 21A is an exploded perspective view of the interposer, FIG. 21B is a perspective view thereof, and FIG. 21C is a sectional view thereof.

  In this interposer 70, it is possible to use at least one of the heat reflection layer 61a and the heat radiation layer 61b used in the interposer 30 described above.

  As shown in FIGS. 21A to 21C, the interposer 70 is configured by combining a rectangular plate-like upper wall 73, a rectangular plate-like lower wall 74, and a pair of side walls 72a and 72b. A channel 71 is formed inside by one member. The pair of side walls 72a and 72b have a mirror-symmetric shape, and are wide in a certain range from their upstream end, and narrow in their constant range from the downstream end. Between the wide portion and the narrow portion on the upstream side of the pair of side walls 72a and 72b, the width changes linearly. Accordingly, the channel 71 defined by the upper wall 73, the lower wall 74, and the side walls 72a and 72b has a funnel-shaped planar shape, and is relatively narrow in the upstream region (small in cross-sectional area) and downstream. In the side region, the width is relatively wide (the cross-sectional area is large).

  The shape of the channel 71 is as clearly shown in FIG. 21C. That is, the height of the channel 71 is constant, but its width is not constant, and is narrow in the upstream region and wide in the downstream region. An intermediate region between the upstream region and the downstream region is a transition region where the width of the channel 71 gradually increases.

  The fluid L flows from the relatively narrow upstream opening (inlet) into the channel 71 of the interposer 70 as shown by the arrow 75a, and flows out from the relatively wide downstream opening (outlet) as shown by the arrow 75b. To do. Since the width of the channel 71 suddenly expands at a location beyond a certain range from the upstream end, the fluid L causes adiabatic expansion while passing through the channel 71, and the temperature decreases. As a result, the temperature in the channel 71 decreases. That is, since the interposer 70 itself has a cooling (endothermic) action, the cooling action of the semiconductor device is improved as compared with the case where adiabatic expansion is not used.

  The structure shown in FIGS. 21A to 21C is referred to herein as a “adiabatic expansion structure”. This “adiabatic expansion structure” may be combined with the “radiation / reflection structure” of the first to third embodiments shown in FIG. 14. Then, the cooling (endothermic) action by the fluid L can be further increased. In this case, the heat reflection layer 61 a is disposed on the upper wall 73 that forms the ceiling portion of the channel 71, and the heat radiation layer 61 b is disposed on the lower wall 74 that forms the floor portion of the channel 71. What is necessary is just to make the arrangement | positioning location and arrangement | positioning range of the heat | fever reflection layer 61a and the heat radiation layer 61b the same as having mentioned above.

  FIG. 22A shows an example of a stacked module mounting structure according to the first embodiment, which is configured by incorporating the interposer 70 having the above-described configuration and operation into the bonding module shown in FIGS. 2A and 2B. FIG. 22B shows the channel 71 of the interposer 70.

  The stacked module mounting structure shown in FIG. 22A is the same as the bonding module shown in FIGS. 2A and 2B except that an interposer 70 is used in place of the interposer 20, and the description thereof will be omitted. In FIG. 22A, the fluid L flows in the direction perpendicular to the paper surface from the front to the rear of the paper.

  In the mounting structure of FIG. 22A, an interposer 70 having a “adiabatic expansion structure” is disposed between the lower first semiconductor device 11a and the upper second semiconductor device 12a. As shown in FIG. 22B, when the fluid L flows into the channel 71 from the narrow inlet of the interposer 70, the fluid L is first a first region having a wide range (relatively large cross-sectional area) from the upstream end. Enter 71a. Subsequently, the fluid L passes through a transition region where the width gradually increases, and enters the second region 71b having a narrow width (relatively small cross-sectional area) within a certain range from the downstream end, and its wide outlet. Through channel 71. When the fluid L passes through the first region 71a, the pressure rapidly decreases, and thus reaches the second region 71b while adiabatic expansion. As a result, the temperature of the fluid L decreases.

  If there is no heat generation in the first semiconductor device 11a and the second semiconductor device 12a, the temperature of the fluid L when flowing out of the channel 71 is lower than the temperature of the fluid L when flowing into the channel 71. If heat is generated in the first semiconductor device 11a and the second semiconductor device 12a, these heats are absorbed by the fluid L, so that the heat from the first semiconductor device 11a and the second semiconductor device 12a is effectively released. The

(Laminated module according to the second embodiment and interposer used therefor)
FIG. 23A is configured by incorporating an interposer 70 (see FIGS. 21A to 21C) that increases the cooling effect of a semiconductor device by utilizing adiabatic expansion of a fluid into the through electrode module illustrated in FIGS. 3A and 3B. The mounting structure of the lamination | stacking module which concerns on 2nd Embodiment is shown. FIG. 23B shows the channel 71 of the interposer 70.

  The mounting structure shown in FIG. 23A is the same as the through electrode module shown in FIGS. 3A and 3B except that an interposer 70 is used in place of the interposer 30, and the description thereof will be omitted. In FIG. 23A, the fluid L flows in the direction perpendicular to the paper surface from the front to the rear of the paper.

  Also in the mounting structure of FIG. 23A, the interposer 70 having the “adiabatic expansion structure” is disposed between the lower first semiconductor device 11b and the upper second semiconductor device 12b. As in the case of, when passing through the first region 71a, the pressure of the fluid L rapidly decreases and reaches the second region 71b while adiabatically expanding. As a result, the temperature of the fluid L decreases.

  If there is no heat generation in the first semiconductor device 11b and the second semiconductor device 12b, the temperature of the fluid L when flowing out of the channel 71 is lower than the temperature of the fluid L when flowing into the channel 71. When heat is generated in the first semiconductor device 11b and the second semiconductor device 12b, these heat is absorbed by the fluid L, so that the heat from the first semiconductor device 11b and the second semiconductor device 12b is effectively released. The

(Modification of the interposer according to the first and second embodiments)
FIGS. 24A and 24B and FIGS. 25A and 25B show a modification of the interposer 70 used in the first and second embodiments described above to increase the cooling effect by utilizing the adiabatic expansion of the fluid. 24A and 24B show an example with one channel 71, and FIGS. 25A and 25B show an example with two channels 71. FIG.

  In these drawings, reference numeral 76 is between the first region 71 a having a narrow width (relatively small cross-sectional area) and the second region 71 b having a wide width (relatively large cross-sectional area), and the width of the channel 71. (Cross-sectional area) increases (changes), that is, the transition region where the fluid L flowing into the channel 71 undergoes adiabatic expansion and the temperature of the fluid L decreases.

  In the example of FIG. 24A (a), the fluid L that has entered the channel 71 from the upstream end first enters the first region 71a having a constant width (cross-sectional area), and then gradually increases in width (cross-sectional area). The second region 71b having a constant width (cross-sectional area) is passed through the transition region 76 that increases to a predetermined value. Then, it flows out from the channel 71 through the wide outlet of the second region 71b. The fluid L suddenly drops in the transition region 76 and reaches the second region 71b while adiabatic expansion. As a result, the temperature of the fluid L decreases.

  In the example of FIG. 24A (b), the width (cross-sectional area) gradually increases at a constant rate from the upstream end of the first region 71a to the downstream end of the second region 71b. Then, adiabatic expansion occurs at the moment when it flows into the channel 71 (first region 71a).

  The example of FIG. 24A (c) is the same as the example of FIG. 24A (a), but the position of the first region 71a on the upstream side is shifted in the width direction with respect to the center line of the channel 71 (interposer 70). Since it is arranged, the fluid L appears to flow in an oblique direction (in the figure, from the lower left to the upper right).

  The example of FIG. 24B (d) is the same as the example of FIG. 24A (a), but the first region 71a and the second region 71b are arranged closer to each other than the example of FIG. In other words, the difference is that the change in the width (cross-sectional area) from the first region 71a to the second region 71b is more rapid. In this example, since the adiabatic expansion occurs more rapidly than in the example of FIG. 24A (a), the temperature drop of the fluid L becomes even greater.

  The example of FIG. 24B (e) is the same as the example of FIG. 24A (a), but the first region 71a has a width (cross-sectional area) between the first region 71a and the second region 71b, that is, in the transition region 76. The difference is that a smaller portion (throttle portion) is provided. In this example, the fluid L is adiabatically expanded after being pressurized once at the throttle portion.

  As shown in FIGS. 24A and 24B, the shape of the channel 71 is not particularly limited. Depending on the required position of the transition region 76, the shape of the channel 71 can be determined. That is, the transition region 76 can be arranged according to the heat generation state of the semiconductor devices (not shown in FIGS. 24A and 24B) arranged above and below the interposer 70 having the channel 71.

  In general, heat generation in a semiconductor device does not occur on the entire surface of a chip constituting the semiconductor device. For example, power consumption is large in an arithmetic circuit, an output circuit, etc., and heat generation is mainly generated in a region (hot spot) where these circuits are arranged. For this reason, it is preferable to arrange the transition region 76 in a region corresponding to these circuits (a portion overlapping with these circuits in the upper or lower order). In this case, there is an advantage that heat can be absorbed by the fluid L before the heat is diffused over the entire surface of the chip.

  In the example of FIG. 25A (a), two channels 71 having the same shape are provided adjacent to each other. There are two fluid L inlets and two fluid L outlets. That is, the two channels 71 are independent of each other.

  In the example of FIG. 25A (b), one (left side in the figure) channel 71 is the same as the example in FIG. 25A (a), but the other (right side in the figure) channel 71 has a transition region 76 downstream. They are shifted (upward in the figure). There are two fluid L inlets and two fluid L outlets. That is, the two channels 71 are independent of each other.

  In the example of FIG. 25B (c), the inlets of the two channels in FIG. 25A (a) are combined into one place. The channel 71 is branched into two inside the interposer 70 and then connected to the two outlets. Therefore, the two channels 71 are connected to each other inside the interposer 70.

  In the example of FIG. 25B (d), the inlets of the two channels are combined into one place as in FIG. 25B (c), but one channel 71 (the left side in the figure) has a transition region. 76 differs in that the channel 71 is arranged on the upstream side (downward in the figure), and the channel 71 on the other side (right side in the figure) differs in that the transition region 76 is arranged on the downstream side (upward in the figure). Also in this example, the two channels 71 are connected to each other inside the interposer 70.

  In FIG. 25A and FIG. 25B, an example in which there are two channels 71 and transition regions 76 is shown, but the number of channels 71 and transition regions 76 is not limited. Three or more channels 71 and transition regions 76 may be provided. The shape of the channel 71 and the number and arrangement of the transition regions 76 can be arbitrarily selected in accordance with the heat generation state of the semiconductor device (for example, how many regions with large heat generation exist).

(First example of manufacturing method of interposer according to first and second embodiments)
26A and 26B show a first example of a method for manufacturing the interposer 70 using adiabatic expansion of fluid.

  In this manufacturing method, first, as shown in FIGS. 26A (a) and 26B (a), an upper wall 73 having a rectangular plate shape is prepared. On the other hand, as shown in FIGS. 26A (b) and 26B (b), a rectangular plate-like lower wall 74 is prepared, and left and right side walls 72a and 72b are formed on the surface thereof. By the side walls 72 a and 72 b, a recess 74 a having a funnel-shaped planar shape is formed on the surface of the lower wall 74.

  After that, as shown in FIGS. 26A (c) and 26B (c), if the upper wall 73 is joined to the surfaces of the left and right side walls 72a and 72b, the interposer 70 having the adiabatic expansion structure shown in FIGS. 21A to 21C. Is manufactured. The upper surface of the recess 74 a is blocked by the upper wall 73 to form a channel 71.

  The material of the upper wall 73 and the lower wall 74 is preferably single crystal silicon or the like, and the material of the side walls 72a and 72b is preferably glass, single crystal silicon, resin, or photoresist (SU-8 or the like). However, other materials may be used. The upper wall 73, the lower wall 74, and the side walls 72a and 72b are preferably joined by an adhesive or the like, but are not limited thereto. Other known techniques can also be used.

In FIGS. 26A and 26B, the number of the channels 71 is one. However, the present invention is not limited to this, and two or more channels 71 are provided as illustrated in FIGS. 24A and 24B and FIGS. 25A and 25B. Also, the upper wall 73 or the lower wall 74 can be omitted. For example, the structure shown in FIGS. 26A (b) and 26B (b) can be used as the interposer 70 as it is. In this case, since the interposer 70 does not have a ceiling (the upper surface is opened), the lower surface of the semiconductor device disposed at the upper level is used as the ceiling. Similarly, if the structure shown in FIGS. 26A (b) and 26B (b) is turned upside down, the interposer 70 without a floor (opened at the lower surface) can be obtained. Is used as a ceiling.

(Second example of manufacturing method of interposer according to first and second embodiments)
27A and 27B show a second example of a method for manufacturing the interposer 70 using adiabatic expansion of fluid.

  In this manufacturing method, first, as shown in FIGS. 27A (a) and 27B (a), a rectangular lower wall 74 having left and right side walls 72a and 72b integrally formed on the surface is prepared. A recess 74a having a funnel-shaped planar shape is formed on the surface side of the lower wall 74 by the side walls 72a and 72b. The material of the lower wall 74 and the side walls 72a and 72b is preferably single crystal silicon, and the depression 74a is selected on the surface side of the lower wall 74 by, for example, wet etching (including anisotropic etching) or dry etching. It is formed by removing it.

  Then, as shown in FIGS. 27A (b) and 27B (b), when the upper wall 73 is joined to the surfaces of the left and right side walls 72a and 72b using an adhesive or the like, the heat insulation shown in FIGS. 21A to 21C. The interposer 70 having an expansion structure is manufactured. The upper surface of the recess 74 a is blocked by the upper wall 73 to form a channel 71. When the upper wall 73 is glass, it can be joined and integrated using electrostatic bonding or the like. When the upper wall 73 is made of single crystal silicon, it can be integrated by a silicon-silicon junction.

  27A and 27B, the number of the channels 71 is one. However, the present invention is not limited to this, and two or more channels 71 are provided as illustrated in FIGS. 24A and 24B and FIGS. 25A and 25B. Also good.

  Further, the upper wall 73 or the lower wall 74 can be omitted. For example, the structure shown in FIGS. 27A (a) and 27B (a) can be used as the interposer 70 as it is. In this case, since the interposer 70 does not have a ceiling (the upper surface is opened), the lower surface of the semiconductor device disposed at the upper level is used as the ceiling. Similarly, if the structure shown in FIGS. 27A (a) and 27B (a) is turned upside down, the interposer 70 without a floor (opened at the lower surface) can be obtained. Is used as a ceiling.

(Third example of manufacturing method of interposer according to first and second embodiments)
28A and 28B show a third example of the method for manufacturing the interposer 70 using the adiabatic expansion of fluid.

  In this manufacturing method, first, as shown in FIGS. 28A (a) and 28B (a), a rectangular plate-like lower wall 74 having pads 78 and 79 for electrical connection and a through electrode 77 is prepared. The pad 78 is formed on the surface of the lower wall 74, and the pad 79 is formed on the back surface of the lower wall 74. Each through-electrode 77 passes through the lower wall 74 and reaches the corresponding pads 78 and 79. As is clear from FIG. 28B (a), the pads 78 and 79 and the through electrode 77 are formed so as not to overlap the channel 71 only in the vicinity of the upstream end portion of the channel 71. The lower wall 74 is preferably made of single crystal silicon.

  Next, as shown in FIGS. 28A (b) and 28B (b), the surface of the lower wall 74 is selectively removed to form a recess 74a having a funnel-shaped planar shape. At this time, a pair of side walls 72a and 72b are formed by the remaining portions of the lower wall 74 on both the left and right sides of the recess 74a. The recess 74a does not overlap the pads 78 and 79, and the pads 78 and 79 and the through electrode 77 are disposed on the side walls 72a and 72b. If the lower wall 74 is made of single crystal silicon, the recess 74a can be formed by, for example, wet etching (including anisotropic etching) or dry etching.

  Thereafter, as shown in FIGS. 28A (c) and 28B (c), a rectangular plate-like upper wall 73 having a plurality of openings 73a is prepared. The opening 73 a is disposed at a position corresponding to each pad 78 of the lower wall 74. If the upper wall 73 is made of single crystal silicon, the opening 73a is formed by, for example, wet etching (including anisotropic etching) or dry etching.

  Finally, as shown in FIGS. 28A (d) and 28B (d), the upper wall 73 is joined to the lower wall 74 using an adhesive or the like. If the upper wall 73 and the lower wall 74 are both made of single crystal silicon, they can be bonded by silicon-silicon bonding. When the upper wall 73 is glass, it can be joined by electrostatic joining or the like.

  In FIGS. 28A and 28B, the number of the channels 71 is one. However, as illustrated in FIGS. 24A and 24B and FIGS. 25A and 25B, two or more channels 71 may be provided.

  Further, the upper wall 73 or the lower wall 74 can be omitted. For example, the structure shown in FIGS. 28A (b) and 28B (b) can be used as the interposer 70 as it is. In this case, since the interposer 70 does not have a ceiling (the upper surface is opened), the lower surface of the semiconductor device disposed at the upper level is used as the ceiling. Similarly, if the structure shown in FIGS. 28A (b) and 28B (b) is turned upside down, the interposer 70 without the floor (the lower surface is opened) can be obtained, so that the upper surface of the semiconductor device disposed below is provided. Is used as a ceiling.

(Configuration example of adding a substrate and a cover to the laminated module according to the first and second embodiments)
FIG. 29 shows a configuration example (mounting structure example) when the substrate 10 and the cover 82 are added to the laminated module according to the second embodiment shown in FIG. 23A. A laminated module 80 having the configuration shown in FIG. 23A is mounted on the substrate 10, and the cover 82 uses the legs 85 of the cover 82 so as to include the entire laminated module 80. Fixed on top. The cover 82 has an inlet 83 and an outlet 84. The inlet 83 is disposed on the upstream side (the side into which the fluid L flows) of the interposer 70 having the “adiabatic expansion structure”, and the outlet 84 is disposed on the downstream side (the side on which the fluid L flows out) of the interposer 70. .

  In FIG. 29, the first semiconductor device 11 disposed below the stacked module 80, the second semiconductor device 12 disposed above the stacked module 80, and the interposer 70 disposed between the semiconductor devices 11 and 12. The case where the magnitude | size of is not mutually equal is illustrated. However, there is no restriction on the relationship between these sizes. Further, two or more first semiconductor devices 11 may be arranged below the interposer 70, or two or more second semiconductor devices 12 may be arranged above the interposer 70. A configuration having a larger number of stages (the stacked module 80 in FIG. 29 has three stages) may be employed.

  The configuration example of FIG. 29 is also applicable when the substrate 10 and the cover 82 are added to the stacked module shown in FIG. 22A. In this case, the laminated module 80 only replaces the bonding module, and the other configurations are the same.

(Fluid supply to laminated module)
FIG. 30 shows an example of a fluid supply system to the laminated module according to the present invention.

  In FIG. 30, the inlet 83 of the cover 82 communicates with the upstream opening of the channel 71 of the interposer 70 built in the laminated module 80, and the outlet 84 of the cover 82 communicates with the downstream opening. Reference numeral 90 denotes a compressor, and 91 denotes a power source (usually a motor) that drives the compressor 90. The compressor 90 is connected to the inlet 83 via the pipe T3. The outlet 84 is connected to the compressor 90 via the pipe T4.

  The fluid (here, gas) L pressurized and compressed by the compressor 90 is guided to the inlet 83 via the pipe T3 and enters the upstream space inside the cover 82. Thereafter, the fluid L passes through the channel 71 of the interposer 70 and moves to the downstream space inside the cover 82. Then, it flows out from the downstream space to the outside via the outlet 84. The fluid L that has flowed out in this manner is returned to the compressor 90 via the pipe T4. Since the fluid L flowing through the pipe T4 absorbs heat generated in the laminated module 80 and its temperature is high, the temperature of the fluid L is further increased by pressurization (adiabatic compression) in the compressor 90. Therefore, the compressor 90 preferably has a function (fluid cooling function) for reducing the temperature of the fluid L. However, the fluid cooling function is not essential and can be omitted. For example, if the pipes T3 and T4 are placed in an environment where the temperature is relatively low, the fluid cooling function may be unnecessary.

  When a liquid is used as the fluid L, a pump is used instead of the compressor 90.

  The present invention can be applied to any stacked module (laminated semiconductor device) and its mounting structure that require stable operation while suppressing an increase in operating temperature of a semiconductor device with high power consumption. The present invention is not limited to semiconductor devices and modules mainly composed of electronic circuits, but other types of high power semiconductor devices and modules (for example, light emitting devices, light emitting element modules, optical communication devices, optical communication modules, etc.). Also widely applicable. Furthermore, in μTAS (micro total analysis system) aimed at miniaturization of analytical instruments, the present invention can be applied to a field in which a fluid to be measured is poured into a minute flow path.

DESCRIPTION OF SYMBOLS 10 Substrate 11, 11a, 11b 1st semiconductor device 12, 12a, 12B 2nd semiconductor device 13 Bonding wire 14 Through electrode 15 Conductive ball 16 Underfiller 17 Adhesive 20 Interposer 21 Channel 22 Side wall 23 Upper wall 24 Lower wall 25 Arrow 30, 30a, 30b, 30c Interposer 31 Channel 32 Side wall 33 Upper wall 33a Opening 34 Lower wall 34 'Base material 35, 35a Arrow 36 Through electrode 37, 38 Conductive balls 40, 40a, 40b Stacked module 41 Channel 42 Cover 42a, 42b Cover half 43 Inlet 44 Outlet 45 Leg 46 Mounting portion 47a, 47b, 48, 49 Arrow 50 Internal space 51, 51a Weir 52 Cover 52a Cover half 52b Cover half 55 Adhesive layer 56 Resin 57 Cover 58 Inlet G 59 Outlet 61a Heat reflection layer 61b Heat radiation layer 62a, 62b Insulating layer 63, 64 Heat source 66a Arrow 67 Adhesive material 68 Electrical connection pad 70 Interposer 71 Channel 71a First region 71b Second region 72a Side wall 73 Upper wall 73a Opening 74 Lower wall 75a, 75b Arrow 76 Transition region 77 Through electrode 78, 79 Pad 80 Stack module 82 Cover 83 Inlet 84 Outlet 85 Leg 90 Compressor 91 Power source G Gap L Fluid L2 Second fluid P Pumps T1, T2, T3, T4 Piping

Claims (6)

An interposer with a channel through which fluid flows;
One or more first semiconductor devices disposed on one side of the interposer;
One or more second semiconductor devices arranged on the opposite side of the interposer from the first semiconductor device,
The channel of the interposer has a first region having a relatively small cross-sectional area and a second region having a relatively small cross-sectional area, and moves from the first region to the second region. A laminated module, wherein the fluid is configured to cause adiabatic expansion along the way.   The stacked module according to claim 1, wherein a height of the channel is constant, and a width of the second region is set larger than a width of the first region. A channel through which fluid flows,
The channel includes a first region having a relatively small cross-sectional area and a second region having a relatively small cross-sectional area, and a fluid that passes through the channel from the first region to the second region. Is configured to cause adiabatic expansion.   The interposer according to claim 3, wherein a height of the channel is constant, and a width of the second region is set larger than a width of the first region.   The interposer according to claim 3 or 4, wherein a plurality of the channels are provided, and an inlet and an outlet of the channels are formed independently of each other.   The interposer according to claim 3 or 4, wherein a plurality of the channels are provided, inlets of the channels are connected to each other, and outlets of the channels are formed independently of each other.

Images