Electronic device and method for producing same
By integrating rigid members with high rigidity and thermal conductivity into the bridge layer, the thermal expansion mismatch and heat dissipation issues in electronic devices are addressed, enhancing the mechanical stability and performance of semiconductor devices.
Patent Information
- Application Number
- PCT/JP2024/016867
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-12
- Filing Date
- 2024-05-02
- Publication Date
- 2024-11-21
AI Technical Summary
Electronic devices face warpage issues due to significant differences in thermal expansion coefficients between chip and bridge layers, and inadequate heat dissipation, particularly as semiconductor devices increase in size.
Incorporating rigid members with higher rigidity and thermal conductivity into the bridge layer to adjust the thermal expansion coefficient and enhance heat dissipation, while maintaining low manufacturing costs and preventing electrical leakage.
The solution effectively reduces warpage and improves heat dissipation in electronic devices by aligning the thermal expansion coefficients of the chip and bridge layers, ensuring mechanical stability and performance.
Smart Images

Figure JP2024016867_21112024_PF_FP_ABST
Abstract
Description
Electronic device and manufacturing method thereof
[0001] The present invention relates to an electronic device such as a semiconductor device and a manufacturing method thereof.
[0002] Conventionally, electronic devices have been known in which two chips are arranged on one side of a wiring layer, and another chip connected to the two chips is arranged on the other side of the wiring layer to electrically connect the two chips (see, for example, Patent Document 1). In this specification, a chip for connecting two chips like this is referred to as a "bridge chip." Connecting two chips using a bridge chip allows the number of wires between the chips to be increased while miniaturizing each wire, which has the advantage of improving the performance of both chips, reducing power consumption, and reducing costs, for example.
[0003] Patent No. 6912423
[0004] Incidentally, with reference to the invention described in Patent Document 1, it is considered to adopt, as the structure of the electronic device, a structure such as that shown in Fig. 13. Note that the electronic device 100' shown in Fig. 13 was created on a trial basis and has the problem that the present invention is intended to solve, but the applicant does not acknowledge that an electronic device with such a structure was publicly known at the time of filing of the present application.
[0005] 13, the electronic device 100' includes a first wiring layer 10, a chip layer 20 provided on one side of the first wiring layer 10, and a bridge layer 30 provided on the other side. The bridge layer 30 is a layer located between the first wiring layer 10 and a second wiring layer 40, and a plurality of solder balls 50 are attached to the second wiring layer 40. The chip layer 20 includes a first chip 21, a second chip 22, and a third chip 23, each having a logic element or a memory element, a conductive member 24 connecting these chips to the first wiring layer 10, and a first resin part 25 sealing them. The bridge layer 30 also includes a first bridge chip 31 for electrically connecting the first chip 21 and the second chip 22, a second bridge chip 32 for electrically connecting the first chip 21 and the third chip 23, a conductive member 36 for connecting these bridge chips to the first wiring layer 10, a through via 37 that directly connects the first wiring layer 10 and the second wiring layer 40, and a second resin part 38 that seals these.
[0006] In a structure such as the electronic device 100′ shown in FIG. 13 , the chip layer 20 is mostly occupied by the chips 21-23, with the first resin portion 25 accounting for a relatively small proportion. In contrast, the bridge layer 30 is mostly occupied by the second resin portion 38, with the bridge chips 31 and 32 accounting for a relatively small proportion. When the proportions of resin in the chip layer 20 and the bridge layer 30 differ significantly, a large difference occurs between the coefficient of thermal expansion (CTE) of the entire chip layer 20 and the coefficient of thermal expansion of the entire bridge layer 30. Such a large difference in the coefficients of thermal expansion between the chip layer 20 and the bridge layer 30 may cause warping of the electronic device 100′, for example, during thermal processes such as reflow and baking during the formation and assembly of the electronic device 100′. In particular, semiconductor devices have tended to become larger in recent years, and as packages become larger, warping due to differences in the coefficients of thermal expansion of the layers becomes more pronounced.
[0007] 13, it is effective to densely arrange multiple chips 21 to 23 on the chip layer 20. However, in this case, the amount of heat generated from each of the chips 21 to 23 increases, and therefore, it becomes a challenge to dissipate the heat appropriately.
[0008] Therefore, a first object of the present invention is to appropriately adjust the coefficient of thermal expansion of the bridge layer in order to suppress warpage in an electronic device. A second object of the present invention is to provide an electronic device with excellent heat dissipation. An object of the present invention is to provide a means for solving both or at least one of these objects.
[0009] After extensive research into ways to achieve the above-mentioned objectives, the inventors of the present invention came up with the idea of arranging a rigid and / or thermally conductive member in the bridge layer. By arranging a rigid member in the bridge layer, the thermal expansion coefficient of the entire bridge layer can be appropriately adjusted to approach that of the chip layer, for example. Furthermore, by arranging a thermally conductive member in the bridge layer, heat generated from the chip in the chip layer can be appropriately dissipated. Based on the above findings, the inventors then came up with the idea of solving the problems of the prior art and completed the present invention. The following describes in detail the configuration and steps of the present invention.
[0010] A first aspect of the present invention relates to an electronic device 100 such as a semiconductor device. The electronic device 100 according to the present invention includes a wiring layer 10 including a conductive member, a chip layer 20 provided on one side of the wiring layer 10, and a bridge layer 30 provided on the other side of the wiring layer 10. The chip layer 20 includes at least two chips 21-23 and a first resin portion 25 that encapsulates the chips 21-23. The bridge layer 30 includes multiple bridge chips 31 and 32 electrically connected to the chips 21-23 via the conductive member of the wiring layer 10, one or more rigid members 33-35, and a second resin portion 38 that encapsulates the bridge chips 31 and 32 and the rigid members 33-35. The "bridge chip" is essentially configured to electrically connect two chips in the chip layer via the wiring layer. Furthermore, the "rigid member" refers to a member having a higher rigidity than the second resin portion. Alternatively, the "rigid member" refers to a member having a lower coefficient of thermal expansion (CTE) than the second resin portion. Here, the one or more rigid members 33-35 include one (inter-chip rigid member) disposed between the bridge chips 31, 32. In this way, by disposing the rigid members 33-35 on the chip layer 20, it is possible to adjust the thermal expansion coefficient of the entire bridge layer 30. Furthermore, by using these rigid members 33-35 to adjust the difference in thermal expansion coefficient between the chip layer 20 and the bridge layer 30 so that it is within a predetermined range, it is possible to suppress deformation occurring in the electronic device 100. Furthermore, by disposing the rigid member 33 between the bridge chips 31, 32, it is possible to effectively control the difference in thermal expansion coefficient between the chip layer 20 and the bridge layer 30.
[0011] In the electronic device 100 according to the present invention, one or more of the rigid members 33 to 35 may include a non-connected rigid member that is not electrically connected to the chips 21 to 23. The rigid members 33 to 35 that do not require electrical connection can be manufactured inexpensively, and therefore, even when the rigid members 33 to 35 are added to the bridge layer 30, an increase in manufacturing costs can be suppressed.
[0012] In the electronic device 100 according to the present invention, it is preferable that one or more of the rigid members 33 to 35 have thermal conductivity. By arranging the thermally conductive rigid members 33 to 35 in the bridge layer 30, it is possible to improve the heat dissipation performance of the electronic device 100. In particular, the area between the bridge chips 31 and 32 is a region where heat is likely to accumulate, and by arranging the thermally conductive rigid member 33 there, it is possible to effectively dissipate the heat generated from each of the chips 21 to 23.
[0013] In the electronic device 100 according to the present invention, the rigid members 33 to 35 may be insulating. By forming the rigid members 33 to 35 from an insulating material, it is possible to prevent leakage of electricity between the bridge chips 31 and 32, for example.
[0014] In the electronic device 100 according to the present invention, the rigid members 33 to 35 are preferably formed from one or more materials selected from the group consisting of silicon, metal, ceramic, and glass. By forming the rigid members 33 to 35 from these materials, the thermal expansion coefficient of the bridge layer 30 can be appropriately reduced, and the heat dissipation performance of the bridge layer 30 can also be improved. Note that the rigid members 33 to 35 may have a structure in which, for example, a conductive member is embedded in the above-mentioned material.
[0015] In the electronic device 100 according to the present invention, one or more of the rigid members 33 to 35 may include a connecting type rigid member electrically connected to the chips 21 to 23. By allowing power supply wiring and signal wiring to be formed on one or more of the rigid members 33 to 35, the degree of freedom of wiring in the bridge layer 30 can be maintained even when the rigid members 33 to 35 are arranged on the bridge layer 30.
[0016] In the electronic device 100 according to the present invention, the region of the bridge layer 30 that overlaps with the chips 21 to 23 of the chip layer 20 is referred to as the "chip overlapping region 30'." This chip overlapping region 30' is further divided into a "bridge chip placement region 30a" where the bridge chips 31 and 32 are placed, and a "bridge chip non-placement region 30b" where the bridge chips 31 and 32 are not placed. In this case, the rigid members 33 to 35 are preferably placed so as to occupy 30% or more of the area of the bridge chip non-placement region 30b. The bridge chip non-placement region 30b of the bridge layer 30 is the main cause of increasing the difference in thermal expansion coefficients between the bridge layer 30 and the chip layer 20. Therefore, by filling 30% or more of this bridge chip non-placement region 30b with the rigid members 33 to 35, the difference in thermal expansion coefficients between the bridge layer 30 and the chip layer 20 can be appropriately reduced.
[0017] In the electronic device 100 according to the present invention, the bridge layer 30 may include an inter-chip rigid member 33 arranged between the bridge chips 31 and 32, and non-inter-chip rigid members 34 and 35 arranged at positions other than between the bridge chips 31 and 32. In this case, it is preferable that the inter-chip rigid member 33 overlaps entirely with the chips 21 to 23 of the chip layer 20. It is also preferable that the non-inter-chip rigid members 34 and 35 overlap partially with the chips 21 to 23 of the chip layer 20, and that the remaining portions do not overlap with the chips 21 to 23 of the chip layer 20. In this way, by arranging the non-inter-chip rigid members 34 and 35 so as to overlap the edges (boundaries) of the chips 21 to 23 of the chip layer 20, it is possible to improve the mechanical strength of the electronic device 100 while suppressing warpage of the electronic device 100.
[0018] In the electronic device 100 according to the present invention, the coefficient of thermal expansion (CTE) of the bridge layer 30 is preferably in the range of 80 to 120%, assuming that the coefficient of thermal expansion of the chip layer 20 is 100%. By setting the difference in the coefficients of thermal expansion between the chip layer 20 and the bridge layer 30 within a predetermined range, warping of the electronic device 100 can be effectively suppressed. The coefficient of thermal expansion of a composite member such as the chip layer 20 or the bridge layer 30 can be calculated using the well-known Turner's theoretical formula.
[0019] In the electronic device 100 according to the present invention, the one or more rigid members preferably include a capacitor 70 electrically connected to at least one of the chips 21-23 via the wiring layer 10. In other words, the capacitor 70 functions not only to store electric charge but also as a type of rigid member. The capacitor 70 is preferably a deep trench capacitor (DTC). By providing the capacitor 70 in the bridge layer 30, the capacitor 70 is positioned in close proximity to the chips 21-23, shortening the wiring length from the wiring layer 10 to the capacitor 70 and reducing wiring resistance and parasitic capacitance. As a result, high-speed charge transfer between the chips 21-23 and the capacitor 70 is possible, improving the operating speed and signal waveform shaping performance of each chip 21-23. Furthermore, integrating the capacitor 70 into the bridge layer 30 allows the electronic device 100 to be miniaturized as a whole.
[0020] In the electronic device 100 according to the present invention, the capacitor 70 is preferably disposed between the bridge chips 31 and 32 in a position overlapping one of the chips 21 to 23. By disposing the capacitor 70 in a position overlapping the chips 21 to 23 in this way, the wiring can be further shortened, further reducing the effects of noise associated with high-speed operation. Furthermore, by overlapping the capacitor 70 with the chips 21 to 23, the mounting area of the electronic device 100 can be effectively utilized, enabling high-density mounting in a small occupied area. Furthermore, when the chips 21 to 23 generate heat, the overlapping capacitor 70 serves as a heat dissipation path, thereby contributing to facilitating cooling of the chips 21 to 23.
[0021] In the electronic device 100 according to the present invention, the capacitor 70 is preferably formed from a material containing ceramic as a dielectric. Capacitors 70 containing ceramic dielectrics have superior rigidity, heat dissipation, and insulation properties compared to other capacitors whose dielectrics are molded from metal or resin, for example. Therefore, the ceramic capacitor 70 exerts the same effect as the rigid members 33 to 35 described above, and is useful for adjusting the thermal expansion coefficient of the entire bridge layer 30. Furthermore, disposing the ceramic capacitor 70 on the bridge layer 30 also contributes to promoting cooling of the chips 21 to 23.
[0022] In the electronic device 100 according to the present invention, at least one of the bridge chips 31, 32 may include a capacitor 70 electrically connected to at least one of the chips 21 to 23 via the wiring layer 10. In this manner, the capacitor 70 may be incorporated into the bridge chips 31, 32.
[0023] In the electronic device 100 according to the present invention, the connection-type rigid member may include a capacitor 70 electrically connected to at least one of the chips 21 to 23 via the wiring layer 10. In this manner, the capacitor 70 may be incorporated into the connection-type rigid member.
[0024] A second aspect of the present invention relates to a method for manufacturing an electronic device 100. In this manufacturing method, first, at least two chips 21-23 are arranged on one side of a wiring layer 10 including conductive members, and then the chips 21-23 are sealed with resin to form a chip layer 20 (a chip layer forming process). Then, multiple bridge chips 31 and 32 electrically connected to the chips 21-23 via the conductive members of the wiring layer 10, and one or more rigid members 33-35, are arranged on the other side of the wiring layer 10, and then these are sealed with resin to form a bridge layer 30 (a bridge layer forming process). Note that the chip layer forming process and the bridge layer forming process may be performed in either order, or these processes may be performed simultaneously. Preferably, the one or more rigid members 33-35 include those arranged between the bridge chips 31 and 32.
[0025] According to the present invention, it is possible to appropriately adjust the coefficient of thermal expansion of the bridge layer, and it is also possible to provide an electronic device with excellent heat dissipation properties.
[0026] FIG. 1 is a cross-sectional view of an electronic device according to a first embodiment. FIG. 2 is a cross-sectional view of an electronic device according to a second embodiment, particularly showing a structure in which multiple chipsets are connected. FIG. 3 is a plan view of an electronic device including multiple chipsets, particularly showing the connection state of each chipset. FIG. 4 is a plan view of an electronic device including multiple chipsets, particularly showing an example of the arrangement of rigid members. FIG. 5 is a plan view of a chipset, showing the range in which rigid members are arranged in the bridge layer. FIG. 6 is a plan view of an electronic device including multiple chipsets, particularly showing another example of the arrangement of rigid members. FIG. 7 shows a first modified example of the electronic device according to the first embodiment shown in FIG. 1. FIG. 8 shows a modified example of the electronic device according to the second embodiment shown in FIG. 2. FIG. 9 shows a second modified example of the electronic device according to the first embodiment shown in FIG. 1. FIG. 10 shows a third modified example of the electronic device according to the first embodiment shown in FIG. 1. FIG. 11 shows a fourth modified example of the electronic device according to the first embodiment shown in FIG. 1. FIG. 12 shows a fifth modified example of the electronic device according to the first embodiment shown in FIG. 1. FIG. 13 is a cross-sectional view showing an electronic device that has a problem in that differences in the thermal expansion coefficients of the layers are likely to occur.
[0027] Hereinafter, embodiments of the present invention will be described with reference to the drawings. The present invention is not limited to the embodiments described below, and includes appropriate modifications of the embodiments described below within the scope obvious to those skilled in the art.
[0028] The drawings in this application show an orthogonal coordinate system with X, Y, and Z axes. The XY plane is horizontal and corresponds to the planar direction of the electronic device, and the Z axis is vertical and corresponds to the thickness direction of the electronic device.
[0029] 1 is a cross-sectional view showing an example of an electronic device 100 according to a first embodiment of the present invention. As shown in Fig. 1, the electronic device 100 according to this embodiment includes a first wiring layer 10, a chip layer 20, a bridge layer 30, and a second wiring layer 40. These layers are stacked in the order of chip layer 20, bridge layer 30, and second wiring layer 40 from the top.
[0030] The first wiring layer 10 is a layer provided with wiring that connects elements or conductive members arranged on one side to other elements or conductive members arranged on the other side. In the first wiring layer 10, conductive members for connecting elements on one side and the other side are embedded in an insulating material. The first wiring layer 10 can freely connect elements on each side by adjusting the layout of the conductive members. The conductive members of the first wiring layer 10 may be made of a known metal material such as copper (Cu) or aluminum (Al).
[0031] The chip layer 20 is located above the first wiring layer 10 and is primarily provided with one or more chips. In the embodiment shown in FIG. 1 , the chip layer 20 includes a first chip 21, a second chip 22, and a third chip 23. As described below, the first chip 21 is electrically connected to each of the second chip 22 and the third chip 23 in a one-to-one relationship. Note that in this embodiment, direct electrical connection between the second chip 22 and the third chip 23 is not anticipated. The chip layer 20 also includes conductive members 24 for connecting each of the chips 21 to 23 to the first wiring layer 10, and a first resin portion 25 for sealing the chips 21 to 23 and the conductive members 24. The conductive members 24 can be made of known metal materials such as copper (Cu) or aluminum (Al), and the first resin portion 25 can be made of known insulating resins such as molded resin (thermosetting resin).
[0032] The first chip 21 is disposed between the second chip 22 and the third chip 23 when viewed in the X-axis direction shown in FIG. 1 . The first chip 21 may employ an integrated circuit that performs digital signal processing in cooperation with each of the second chip 22 and the third chip 23. Specifically, the first chip 21 includes a logic element for controlling the second chip 22 and the third chip 23. Examples of the logic element include a CPU (central processing unit), a GPU (graphics processing unit), a GPGPU (general purpose graphics processing unit), and a DSP (digital signal processor).
[0033] 1, the second chip 22 is disposed on the left side of the first chip 21, and the third chip 23 is disposed on the right side of the first chip 21. The second chip 22 and the third chip 23 may each employ an integrated circuit that is auxiliary used for the digital signal processing by the first chip 21. Specifically, the second chip 22 and the third chip 23 include a memory element for storing information read and written by the first chip 21. Examples of the memory element are a DRAM (dynamic random access memory) and an SRAM (static random access memory).
[0034] The bridge layer 30 is located below the first wiring layer 10 and is mainly provided with one or more bridge chips 31, 32 and one or more rigid members 33-35. The bridge layer 30 also includes a conductive member 36 for connecting the bridge chips 31, 32 to the first wiring layer 10, and a second resin portion 38 for sealing the bridge chips 31, 32, the rigid members 33-35, and the conductive member 36. The second resin portion 38 may also include a through via 37 for directly electrically connecting the first wiring layer 10 and the second wiring layer 40. The conductive member 36 and the through via 37 may be made of a known metal material such as copper (Cu) or aluminum (Al). The second resin portion 38 may be made of a known insulating resin, such as a mold resin (thermosetting resin).
[0035] In the embodiment shown in FIG. 1 , the bridge layer 30 includes a first bridge chip 31 and a second bridge chip 32. Each bridge chip 31, 32 is provided for the purpose of electrically connecting two of the multiple chips 21 to 23 provided on the first wiring layer 10. Specifically, the first bridge chip 31 is connected to the first chip 21 and the second chip 22 in the chip layer 20 via the first wiring layer 10. The first bridge chip 31 includes wiring 31a for connecting the first chip 21 and the second chip 22. The first chip 21 and the second chip 22 are not connected to each other via the first wiring layer 10, but are connected only via the first bridge chip 31. This enables high-speed, high-capacity communication between the first chip 21 and the second chip 22 with low power consumption. Similarly, the second bridge chip 32 is connected to the first chip 21 and the third chip 23 in the chip layer 20 via the first wiring layer 10. The second bridge chip 32 is configured to include wiring 32a for connecting the first chip 21 and the third chip 23. The first chip 21 and the third chip 23 are not connected to each other via the first wiring layer 10, but are connected only via this first bridge chip 31. This enables high-speed, large-capacity communication to be achieved with low power consumption even between the first chip 21 and the third chip 23.
[0036] The first bridge chip 31 may also include an integrated circuit that is used auxiliary for signal processing by the first chip 21 and the second chip 22. For example, the first bridge chip 31 may incorporate a memory element connected to the logic element of the first chip 21. This allows the logic element of the first chip 21 to read and write information from and to the memory element of the first bridge chip 31. For example, the first bridge chip 31 may incorporate a memory controller connected to the memory element of the second chip 22. This allows the memory controller of the first bridge chip 31 to control the memory element of the second chip 22. Note that when both a memory element and a memory controller are incorporated in the first bridge chip 31, the memory controller can also control the memory element in the first bridge chip 31. Furthermore, the memory controller in the first bridge chip 31 may be connected to the logic element of the first chip 21, and the control of each memory element that would normally be performed by the logic element may be shared by the memory controller in the first bridge chip 31. In this way, the first bridge chip 31 may be provided with an integrated circuit that not only connects the first chip 21 and the second chip 22 but also supports these chips, thereby achieving further performance improvements.
[0037] Similarly, the second bridge chip 32 may also include an integrated circuit that is used auxiliary to the signal processing by the first chip 21 and the third chip 23. For example, the second bridge chip 32 may incorporate a memory element connected to the logic element of the first chip 21. This allows the logic element of the first chip 21 to read and write information from and to the memory element of the second bridge chip 32. Furthermore, for example, the second bridge chip 32 may incorporate a memory controller connected to the memory element of the third chip 23. This allows the memory controller of the second bridge chip 32 to control the memory element of the third chip 23. Note that if both a memory element and a memory controller are incorporated in the second bridge chip 32, the memory controller can also control the memory element in the second bridge chip 32. Furthermore, the memory controller in the second bridge chip 32 may be connected to the logic element of the first chip 21, and the control of each memory element that would normally be performed by the logic element may be shared by the memory controller in the second bridge chip 32. In this way, the second bridge chip 32 may be provided with an integrated circuit that not only connects the first chip 21 and the third chip 23 but also supports these chips, thereby achieving further performance improvements.
[0038] As shown in FIG. 1 , the bridge layer 30 further includes rigid members 33 to 35. These rigid members 33 to 35 are arranged in appropriate locations in the bridge layer 30 primarily for the purposes of adjusting the coefficient of thermal expansion (CTE) of the bridge layer 30 and dissipating heat generated in the chip layer 20. Specifically, in the embodiment shown in FIG. 1 , the first rigid member 33 is arranged between the first bridge chip 31 and the second bridge chip 32 in the x-axis direction. In this specification, a rigid member arranged between the bridge chips in this manner is referred to as an "inter-chip rigid member." In particular, this first rigid member 33 (inter-chip rigid member) is arranged to fill the gap between the first bridge chip 31 and the second bridge chip 32. For this reason, for example, if the distance between the first bridge chip 31 and the second bridge chip 32 in the X-axis direction is taken as 100%, the length of the first rigid member 33 in the X-axis direction is preferably 50 to 100%, and particularly preferably 70 to 98% or 80 to 95%.
[0039] Furthermore, when viewed in the Z-axis direction, it can be seen that the first rigid member 33 is located directly below the first chip 21 (logic element) of the chip layer 20. In particular, the first bridge chip 31 completely overlaps the first chip 21. As a result, by forming the first rigid member 33 from a material with high thermal conductivity, heat generated from the first chip 21 can be efficiently dissipated to the outside.
[0040] 1 , the second rigid member 34 is disposed to the left of the first bridge chip 31 in the X-axis direction, and the third rigid member 35 is disposed to the right of the second bridge chip 32 in the X-axis direction. Therefore, in other words, the first bridge chip 31 is provided between the second rigid member 34 and the first rigid member 33, and the second bridge chip 32 is provided between the third rigid member 35 and the first rigid member 33. However, the second rigid member 34 and the third rigid member 35 are not located between the two bridge chips 31 and 32 in the X-axis direction. In this specification, such rigid members that are not disposed between the bridge chips are referred to as "non-inter-chip rigid members."
[0041] Furthermore, when viewed in the Z-axis direction, it can be seen that the second rigid member 34 is located below the second chip 22 (memory element) of the chip layer 20, and the third rigid member 35 is located below the third chip 23 (memory element) of the chip layer 20. However, the second rigid member 34 and the third rigid member 35 do not completely overlap the second chip 22 and the third chip 23, respectively, in the Z-axis direction; rather, they are arranged so that some portions overlap and other portions do not overlap. In other words, the second rigid member 34 has a portion that protrudes from the second chip 22 toward the left in the X-axis direction, and the third rigid member 35 has a portion that protrudes from the third chip 23 toward the right in the X-axis direction. In order to enhance the mechanical strength of the electronic device 100, the second rigid member 34 and the third rigid member 35 are arranged to overlap the edges (boundaries) of the second chip 22 and the third chip 23, respectively.
[0042] The rigid members 33-35 are essentially arranged for the purpose of bringing the thermal expansion coefficient of the bridge layer 30 closer to that of the chip layer 20. More specifically, by providing the rigid members 33-35 within the bridge layer 30, a portion of the second resin portion 38 within the bridge layer 30 can be replaced with the rigid members 33-35, thereby reducing the thermal expansion coefficient of the bridge layer 30. Therefore, the rigid members 33-35 are preferably formed from a material that is more rigid than the second resin portion 38 but has a lower thermal expansion coefficient. Examples of such materials include silicon, glass, metal, and ceramic. Metals that can be used include copper, aluminum, stainless steel, titanium, magnesium, iron, steel, lead, and zinc. Ceramics that can be used include zirconia, alumina, and silicon nitride, and these may be used alone or in combination depending on the application.
[0043] Furthermore, in order to suppress warping and deformation that may occur during thermal processes in the formation and assembly of the electronic device 100, it is preferable that the difference in the coefficient of thermal expansion (CTE) between the chip layer 20 and the bridge layer 30 be within a predetermined range. Since the chips to be placed on the chip layer 20 are basically fixed and adjusting the coefficient of thermal expansion is difficult, the coefficient of thermal expansion of the bridge layer 30 can be made closer to that of the chip layer 20 by adjusting the number and materials of the rigid members 33 to 35 to be placed on the bridge layer 30. Specifically, when the coefficient of thermal expansion of the entire chip layer 20 is taken as 100%, the coefficient of thermal expansion of the entire bridge layer 30 is preferably in the range of 80 to 120%, and particularly preferably in the range of 85 to 115% or 90 to 110%.
[0044] The chip layer 20 and the bridge layer 30 are composite members containing resin, metal, etc., and the thermal expansion coefficient of such composite members can be calculated using the following formula based on the well-known Turner theoretical formula (Turner, P.S., J. Research of the National Bureau of Standards, 36, 239 (1946)). [Thermal expansion coefficient of composite member] α = (α 1 E 1 V 1 +α 2 E 2 V 2 +...) / (E 1 V 1 +E 2 V 2 +...) In the above formula, α is the coefficient of thermal expansion, E is the elastic modulus, V is the volume, and V 1 +V 2 +=1. The subscript ( 1 , 2, ...) refer to the first material, second material, ... constituting the composite member (here, each layer of the chip layer and the bridge layer). Here, chips 21 to 23 included in chip layer 20 and bridge chips 31 and 32 included in bridge layer 30 are formed by combining a silicon substrate with other fine conductive materials and insulating materials. However, most of these chips are silicon substrates, and their thermal expansion coefficients are most significantly affected by the silicon substrate. Therefore, the thermal expansion coefficient α, elastic modulus E, and volume V of these chips can be considered to be the thermal expansion coefficient, elastic modulus, and volume of the silicon substrate. Specifically, in the embodiment shown in FIG. 1 , the thermal expansion coefficient of chip layer 20 can be calculated using Turner's theoretical formula, where the first material is the silicon substrate of first chip 21, the second material is the silicon substrate of second chip 22, the third material is the silicon substrate of third chip 23, the fourth material is the conductive member 24, and the fifth material is the first resin portion 25. Similarly, the thermal expansion coefficient of the bridge layer 30 can be calculated using Turner's theoretical formula, where the first material is the silicon substrate of the first bridge chip 31, the second material is the silicon substrate of the second bridge chip 32, the third material is the first rigid member 33, the fourth material is the second rigid member 34, the fifth material is the third rigid member 35, the sixth material is the conductive member 36, the seventh material is the through via 37, and the eighth material is the second resin part 38.
[0045] Furthermore, the rigid members 33-35, particularly the first rigid member 33 located directly below the first chip 21, are arranged for the purpose of improving the heat dissipation performance of the electronic device 100. More specifically, by replacing the second resin portion 38 within the bridge layer 30 with the rigid members 33-35, the heat dissipation performance of the bridge layer 30 is improved. For this reason, the rigid members 33-35, particularly the first rigid member 33, are preferably formed from a material with high thermal conductivity. Examples of such materials include silicon, metal, and ceramic. Metals that can be used include copper and aluminum, as well as stainless steel, titanium, magnesium, iron, steel, lead, and zinc. Ceramics that can be used include, for example, zirconia, alumina, and silicon nitride. These may be used alone or in combination depending on the application.
[0046] 1, the bridge chips 31, 32 and the rigid members 33 to 35 are preferably arranged so that their lower surfaces are in close contact with the second wiring layer 40. Furthermore, conductive members 36 may be provided between the bridge chips 31, 32 and the first wiring layer 10, and between the rigid members 33 to 35 and the first wiring layer 10. In this embodiment, the rigid members 33 to 35 do not incorporate conductive members or semiconductor elements, and are not electrically connected to the first wiring layer 10. However, for convenience of manufacturing, conductive members 36 may be provided between the rigid members 33 to 35 and the first wiring layer 10, similar to the conductive members 36 provided between the bridge chips 31, 32 and the first wiring layer 10.
[0047] The rigid members 33 to 35 differ from the bridge chips 31 and 32 in that they do not include semiconductor elements such as logic elements or memory elements. As will be described in detail later, the rigid members 33 to 35 may include conductive wiring.
[0048] The second wiring layer 40 is a layer for connecting elements or conductive members arranged on one side to solder balls 50 arranged on the other side. Conductive members for connecting elements and the solder balls 50 to each other are embedded in an insulating material in the second wiring layer. The conductive members of the second wiring layer 40 may be made of known metal materials such as copper (Cu) or aluminum (Al). The second wiring layer 40 is a so-called fan-out wiring layer, and can connect one chip to multiple solder balls 50. Furthermore, power supply wiring and / or signal wiring can be provided to the first chip 21 (logic element), the second chip 22 (memory element), and the third chip 23 (memory element) in the chip layer 20 via, for example, the second wiring layer 40, the through vias 37 in the bridge layer 30, the first wiring layer 10, and the conductive members 24 in the chip layer 20. Power supply wiring and / or signal wiring can also be provided directly to the first bridge chip 31 and the second bridge chip 32 via the second wiring layer 40.
[0049] The solder balls 50 are terminals for connecting the electronic device 100 to a substrate (printed circuit board 200: see FIG. 3 , etc.). Metal pads or metal bumps may be used instead of the solder balls 50. In addition to connecting the electronic device 100 to the substrate, the solder balls 50 also have other functions, such as dissipating heat generated between the electronic device 100 and the substrate and adjusting the electrical characteristics (impedance, etc.) between the electronic device 100 and the substrate.
[0050] Next, with reference to FIGS. 2 and 3, an embodiment will be described in which the device shown in FIG. 1 is used as a single chipset, and multiple such chipsets are connected to improve the overall performance of an electronic device. In this embodiment, as shown in FIG. 3, the chipsets shown in FIG. 1 are connected in two directions along the X axis and two directions along the Y axis, for a total of four chipsets 100(a)-(d). FIG. 2 shows a cross-sectional view of the XZ plane, illustrating the cross-sectional structure of two of the four chipsets 100(a)-(d), 100(a) and (b). As shown in FIG. 2, the two chipsets 100(a) and (b) have basically the same configuration as the electronic device 100 (chipset) described in FIG. 1. Therefore, the same elements as those in the electronic device 100 of FIG. 1 are designated by the same reference numerals, and their description will be omitted. Note that Figure 2 only illustrates the cross-sectional structures of two chip sets 100(a) and (b) out of the four chip sets 100(a) to (d) shown in Figure 3, but the remaining two chip sets 100(c) and (d) basically have the same cross-sectional structure.
[0051] 2 and 3, similarly to the electronic device 100 of FIG. 1, adjacent chips 21 to 23 in the chip layer 20 are electrically connected by bridge chips 31 and 32 of the bridge layer 30. That is, as shown in FIG. 2, in each chip set 100(a) and (b), the first chip 21 and the second chip 22 are connected by the first bridge chip 31, and the first chip 21 and the third chip 23 are connected by the second bridge chip 32.
[0052] As shown in the plan view of FIG. 3 , in each chip set 100(a)-(d), multiple second chips 22(a)-(c) (memory elements) can be connected to one first chip 21 (logic element), and similarly, multiple third chips 23(a)-(c) (memory elements) can be connected to one first chip 21 (logic element). A first bridge chip 31(a)-(c) is used to connect one first chip 21 to multiple second chips 22(a)-(c), respectively. In this way, multiple first bridge chips 31(a)-(c) can be used to connect multiple second chips 22(a)-(c) to one first chip 21. In this case, the number of first bridge chips 31(a)-(c) should be the same as the number of second chips 22(a)-(c). Similarly, second bridge chips 32(a)-(c) are used to connect one first chip 21 to multiple third chips 23(a)-(c), respectively. In this way, multiple second bridge chips 32(a)-(c) can be used to connect multiple third chips 23(a)-(c) to one first chip 21. In this case, the number of second bridge chips 32(a)-(c) provided should be the same as the number of third chips 23(a)-(c).
[0053] 3, with the X-axis direction defined as the left-right direction and the Y-axis direction defined as the up-down direction in a plan view, the second chipset 100(b) is disposed adjacent to the first chipset 100(a) on the right side in the X-axis direction. The third chipset 100(c) is disposed adjacent to the first chipset 100(a) on the lower side in the Y-axis direction. The fourth chipset 100(d) is disposed below the second chipset 100(b) in the Y-axis direction and adjacent to the third chipset 100(c) on the right side in the X-axis direction.
[0054] In this case, the first chip 21 (logic element) of the first chip set 100(a) and the first chip 21 (also logic element) of the third chip set 100(c) are positioned adjacent to each other in the Y-axis direction. Therefore, in this embodiment, a third bridge chip 39(a) is further provided in the bridge layer 30, and this third bridge chip 39(a) electrically connects the first chip 21 of the first chip set 100(a) to the first chip 21 of the third chip set 100(c) in a one-to-one relationship. By connecting the first chips 21 functioning as logic elements in this manner, for example, parallel processing can be performed using two logic elements. For example, one first chip 21 can be used as a CPU and the other first chip 21 as a GPU, with the CPU processing a small amount of complex calculations and the GPU processing a large amount of simple data calculations, such as image processing or machine learning.
[0055] Similarly, the first chip 21 (logic element) of the second chip set 100(b) and the first chip 21 (also a logic element) of the fourth chip set 100(d) are positioned adjacent to each other in the Y-axis direction. Therefore, in this embodiment, a third bridge chip 39(b) is further provided in the bridge layer 30, and this third bridge chip 39(b) electrically connects the first chip 21 of the second chip set 100(b) to the first chip 21 of the fourth chip set 100(d) in a one-to-one relationship. In this way, the third bridge chips 39(a) and 39(b) of the bridge layer 30 can be used to connect the first chips 21 of adjacent chip sets 100(a) to 100(d).
[0056] On the other hand, the first chips 21 (logic elements) of each chipset 100(a) to (d) are not arranged adjacent to each other in the X-axis direction. This is because, within each chipset 100(a) to (d), second chips 22(a) to (c) and third chips 23(a) to (c), which function as memory elements, are provided on both the left and right sides of the first chip 21 in the X-axis direction. That is, the second chip 22 and the third chip 23, which function as memory elements, are interposed between the first chip 21 of the first chipset 100(a) and the first chip 21 of the second chipset 100(b). The same is true between the first chip 21 of the third chipset 100(c) and the first chip 21 of the fourth chipset 100(d). For this reason, the bridge chips in the bridge layer 30 described above cannot electrically connect the first chips 21 of each chipset 100(a) to (d) in the X-axis direction.
[0057] Therefore, in this embodiment, logic elements arranged on either side of a memory element are electrically connected to each other not by a bridge chip but by the conductive member of the first wiring layer 10. Specifically, as shown in FIG. 2 , the first chip 21 (logic element) of the first chip set 100(a) and the first chip 21 (logic element) of the second chip set 100(b) are connected to each other by the conductive member of the first wiring layer 10. This allows the first chips 21 of the first chip set 100(a) and the second chip set 100(b) to be electrically connected to each other in the X-axis direction, straddling the second chip 22 and the third chip 23 that function as memory elements. Note that, although not shown in FIG. 2 , the first chips 21 of the third chip set 100(c) and the fourth chip set 100(d) shown in FIG. 3 are also electrically connected to each other by the conductive member of the first wiring layer 10. In this way, it is preferable to use both a connection method that connects adjacent chips using the bridge chip of the bridge layer 30 and a connection method that connects chips across one or more other chips using the conductive members of the first wiring layer 10. This makes it possible to realize an electronic device with excellent performance in a compact configuration.
[0058] Next, with reference to FIGS. 2 and 4 to 6, a preferred arrangement of the aforementioned rigid members in an embodiment in which multiple chipsets 100(a) to (d) are connected will be described. As shown in FIGS. 2 and 4, each of the chipsets 100(a) to (d) has a first rigid member 33, a second rigid member 34, and a third rigid member 35 provided in the bridge layer 30, each spaced apart in the X-axis direction, essentially similar to the electronic device 100 of FIG. 1. That is, in each of the chipsets 100(a) to (d), the first rigid member 33 is arranged directly below the first chip 21 functioning as a logic element, the second rigid member 34 is arranged below the second chip 22 functioning as a memory element, and the third rigid member 35 is arranged below the third chip 23 functioning as a memory element. However, as shown in FIG. 4, one rigid member is shared between the first chipset 100(a) and the second chipset 100(b), which are adjacent in the X-axis direction. In other words, the third rigid member 35 for the first chipset 100(a) and the second rigid member 34 for the second chipset 100(b) are located in the same position, so one rigid member fulfills the roles of these two rigid members. In this way, one rigid member that functions as both the second rigid member 34 and the third rigid member 35 is located on the boundary between the first chipset 100(a) and the second chipset 100(b). This relationship of sharing one rigid member also applies to the third chipset 100(c) and the fourth chipset 100(d).
[0059] 4, in addition to the first to third rigid members 33 to 35 described above, a fourth rigid member 61 and a fifth rigid member 62 may also be provided. The fourth rigid member 61 is located in the bridge layer 30 and above the first bridge chip 31 and the second bridge chip 32 in the Y-axis direction. The fourth rigid member 61 partially overlaps with the chips 21 to 23 in the chip layer 20, but other portions do not overlap with the chips 21 to 23. On the other hand, the fifth rigid member 62 is located in the bridge layer 30 and below the first bridge chip 31 and the second bridge chip 32 in the Y-axis direction. The fifth rigid member 62 also partially overlaps with the chips 21 to 23 in the chip layer 20, but other portions do not overlap with the chips 21 to 23. In the embodiment shown in Figure 4, a fourth rigid member 61 is disposed above the first chipset 100(a) and the second chipset 100(b) in the Y-axis direction, and a fifth rigid member 62 is disposed below the third chipset 100(c) and the fourth chipset 100(d) in the Y-axis direction. Therefore, when viewed from above in a plan view, the entire electronic device consisting of the four chipsets 100(a) to 100(d) has rigid members 34, 35, 61, and 62 disposed along each of the four sides of the electronic device. By disposing the rigid members 34, 35, 61, and 62 along the outer edges of the entire electronic device in this manner, these rigid members function as a frame (lid structure) that improves the mechanical strength of the electronic device.
[0060] In the example shown in FIG. 4 , the rigid member 33 indicated by the diagonal grid lines and the rigid members 34, 35, 61, and 62 indicated by the normal diagonal lines are shown separately. As mentioned above, the first rigid member 33 indicated by the diagonal grid lines is located directly below the first chip 21 functioning as a logic element and completely overlaps the first chip 21. Therefore, it is preferable that the first rigid member 33 indicated by the diagonal grid lines be formed from a material that has not only rigidity and low thermal conductivity but also heat dissipation properties. Therefore, the first rigid member 33 indicated by the diagonal grid lines functions to suppress warpage and improve heat dissipation. On the other hand, the other rigid members 34, 35, 61, and 62 indicated by the normal diagonal lines are not located so as to completely overlap the first chip 21 functioning as a logic element. Therefore, it is sufficient that they are formed from a material that has rigidity and low thermal conductivity, and they do not necessarily need to have heat dissipation properties. In this way, the materials constituting each rigid member 33, 34, 35, 61, and 62 can be different depending on their placement.
[0061] FIG. 5 illustrates an example of the range in which the rigid members 33, 34, 35, and 61 are arranged within the bridge layer 30 by extracting the first chip set 100(a) from the electronic device shown in FIG. 4 . First, the range of the bridge layer 30 that overlaps with the chips 21 to 23 of the chip layer 20 is defined as the "chip overlapping region 30'." Next, this chip overlapping region 30' is divided into a "bridge chip arrangement region 30a" and a "bridge chip non-arrangement region 30b." The bridge chip arrangement region 30a is the region of the chip overlapping region 30' where the bridge chips 31, 32, and 39 are arranged. The bridge chip non-arrangement region 30b is the region of the chip overlapping region 30' where the bridge chips 31, 32, and 39 are not arranged. In FIG. 5 , the bridge chip non-arrangement region 30b is shaded. The area of this bridge chip no-placement area 30b (shaded portion) where the rigid members 33, 34, 35, and 61 are arranged is referred to as the "rigid member arrangement area 30c." In FIG. 5, this rigid member arrangement area 30c is indicated by double diagonal lines. Note that even in areas where the rigid members 33, 34, 35, and 61 are arranged, portions that protrude from the bridge chip no-placement area 30b are not part of the rigid member arrangement area 30c. In this case, it is preferable that the rigid member arrangement area 30c (double diagonal line portion) occupies 30% or more of the area of the bridge chip no-placement area 30b. In particular, when the area of the bridge chip no-placement area 30b is taken as 100%, the area of the rigid member arrangement area 30c is preferably 30 to 80%, and particularly preferably 35 to 70%. In this way, by making 30% or more of the bridge chip non-placement area 30b of the chip overlap area 30' of the bridge layer 30 into a rigid member placement area 30c in which rigid members 33, 34, 35, and 61 are placed, the difference in thermal expansion coefficient between the bridge layer 30 and the chip layer 20 can be appropriately reduced.
[0062] FIG. 6 shows another example of the arrangement of the rigid members shown in FIG. 4. In the example shown in FIG. 4, the first rigid member 33, the second rigid member 34, the third rigid member 35, the fourth rigid member 61, and the fifth rigid member 62 are provided separately for each of the chipsets 100(a) to (d). On the other hand, in the modified example shown in FIG. 6, the second to sixth rigid members 34, 35, 61, and 62, excluding the first rigid member 33, are integrated to form a single rigid member 63. This single rigid member 63 surrounds the periphery of the electronic device consisting of the four chipsets 100(a) to (d) and extends vertically between the first chipset 100(a) and the second chipset 100(b) and between the third chipset 100(c) and the fourth chipset 100(d). In this way, except for the first rigid member 33 disposed directly below the first chip 21 (logic element), the other rigid members may be integrated to form a single rigid member 63 .
[0063] Fig. 7 shows a first modified example of the electronic device shown in Fig. 1. The electronic device in Fig. 7 will be described mainly with respect to the configuration different from that shown in Fig. 1, and a description of the same configuration will be omitted.
[0064] The electronic device 100 of FIG. 7 primarily provides wiring 33a, 34a, and 35a to each of the rigid members 33, 34, and 35 in the bridge layer 30, respectively, and utilizes these wiring 33a, 34a, and 35a as signal wiring and / or power wiring for supplying signals and power to the chips 21-23 in the chip layer 20. In the example shown in FIG. 1 , no wiring is provided within the rigid members 33, 34, and 35, and signals and power are supplied to the chips 21-23 in the chip layer 20 using, for example, through-vias 37 formed in the bridge layer 30. However, the rigid members 33, 34, and 35 become dead space, which reduces the design freedom for arranging signal wiring and power wiring within the bridge layer 30. In contrast, in the example shown in FIG. 7 , by providing wiring 33a-35a to each of the rigid members 33-35, these rigid members 33-35 can be effectively utilized as conduction paths such as signal wiring and power wiring. This reduces the dead space in the bridge layer 30, and maintains the degree of freedom in wiring design.
[0065] 7, the wiring 33a of the first rigid member 33 is electrically connected to the first chip 21 via the conductive member 36 of the bridge layer 30, the first wiring layer 10, and the conductive member 24 in the chip layer 20. The wiring 34a of the second rigid member 34 is electrically connected to the second chip 22 via the conductive member 36 of the bridge layer 30, the first wiring layer 10, and the conductive member 24 in the chip layer 20. The wiring 35a of the third rigid member 35 is electrically connected to the third chip 23 via the conductive member 36 of the bridge layer 30, the first wiring layer 10, and the conductive member 24 in the chip layer 20. In this manner, it is also possible to electrically connect the rigid members 33 to 35 to the chips 21 to 23 in the chip layer 20.
[0066] Fig. 8 shows a modification of the electronic device 100 shown in Fig. 2. Regarding the electronic device 100 in Fig. 7, the following description will focus on the configuration that is different from that shown in Fig. 1, and a description of the same configuration will be omitted.
[0067] The electronic device of FIG. 8 includes multiple chipsets 100(a) and 100(b) connected in the X-axis direction. Instead of the first chip 21 described above, each chipset 100(a) and 100(b) includes two logic elements, namely, a first logic element 21a and a second logic element 21b, in the chip layer 20. Examples of the logic elements 21a and 21b include a CPU or a GPU, as described above. Alternatively, one may be a CPU and the other a GPU. The first logic element 21a and the second logic element 21b are electrically connected by a fourth bridge chip 64 provided in the bridge layer 30. Thus, in the embodiment shown in FIG. 8, by providing two logic elements 21a and 21b in each chipset 100(a) and 100(b) and connecting these logic elements 21a and 21b via the fourth bridge chip 64, it is possible to improve the processing performance.
[0068] 8, the first chip 21 shown in FIG. 1 is changed to two logic elements 21a and 21b, and accordingly, in the embodiment of FIG. 8, the first rigid member 33 shown in FIG. 1 is changed to two logic element rigid members 33b and 33c. These logic element rigid members 33b and 33c are disposed directly below the logic elements 21a and 21b, respectively. Furthermore, these logic element rigid members 33b and 33c are disposed in the bridge layer 30 between the first bridge chip 31 and the fourth bridge chip 64, or between the second bridge chip 32 and the fourth bridge chip 64, respectively. Therefore, these logic element rigid members 33b and 33c also become "inter-chip rigid members."
[0069] 8, the first logic element 21a and the second logic element 21b of the first chipset 100(a) are electrically connected to the first logic element 21a and the second logic element 21b of the second chipset 100(b) via the conductive member of the first wiring layer 10. More specifically, the first logic element 21a of the first chipset 100(a) is connected to the first logic element 21a of the second chipset 100(b), and the second logic element 21b of the first chipset 100(a) is connected to the second logic element 21b of the second chipset 100(b). In this way, by using the first wiring layer 10, the multiple logic elements 21a and 21b included in each chipset 100(a) and (b) can be connected separately.
[0070] The electronic device described so far can be manufactured according to a known method. For example, in the embodiment shown in FIG. 1 , the first wiring layer 10 is first prepared. At this time, a support (not shown) may be prepared and the first wiring layer 10 may be placed on top of it. In this case, one side of the first wiring layer 10 is exposed and the other side is covered by the support.
[0071] Next, the first chip 21, second chip 22, and third chip 23, each having a conductive member 24, are fixed to one surface (exposed surface) of the first wiring layer 10. At this time, the conductive member 24 of each of the chips 21 to 23 is brought into contact with the first wiring layer 10. Thereafter, a thermosetting resin or the like is poured onto one surface of the first wiring layer 10 and cured, thereby sealing each of the chips 21 to 23 on the first wiring layer 10. This resin becomes the first resin portion 25. In this way, the chip layer 20 is formed on one surface of the first wiring layer 10.
[0072] Next, the support (not shown) is removed from the other surface of the first wiring layer 10. Then, the first bridge chip 31, the second bridge chip 32, the first rigid member 33, the second rigid member 34, and the third rigid member 35, each having a conductive member 36, are fixed to the other surface of the first wiring layer 10. At this time, the conductive members 36 of the bridge chips 31 and 32 and the conductive members 36 of the rigid members 33 to 35 are brought into contact with the first wiring layer 10. Then, a thermosetting resin or the like is poured onto the other surface of the first wiring layer 10 and cured to seal the bridge chips 31 and 32 and the rigid members 33 to 35 on the first wiring layer 10. This resin becomes the second resin portion 38. Alternatively, a hole reaching the first wiring layer 10 may be formed in the second resin portion 38 using a drill or a laser, and then a conductive material may be applied to the inner surface of the hole to form the through via 37. In this way, the bridge layer 30 is formed on the other surface side of the first wiring layer 10. It is noted that either the step of forming the chip layer 20 or the step of forming the bridge layer 30 may be performed first.
[0073] Thereafter, a second wiring layer 40 is laminated on the lower layer side of the bridge layer 30. Furthermore, a plurality of solder balls 50 are attached on the lower layer side of the second wiring layer 40. This completes the electronic device 100 of the embodiment shown in Fig. 1. Note that an electronic device including a plurality of chip sets as shown in Fig. 2 and the like may also be manufactured by referring to the manufacturing method described here.
[0074] Fig. 9 shows a second modified example of the electronic device 100 shown in Fig. 1. The electronic device 100 in Fig. 9 will be described mainly with respect to the configuration different from that shown in Fig. 1, and a description of the same configuration will be omitted.
[0075] The electronic device 100 shown in FIG. 9 further includes a capacitor 70 in the bridge layer 30 in addition to the first bridge chip 31, the second bridge chip 32, the first rigid member 33, and the second rigid member 34. The electronic device 100 shown in FIG. 9 employs the capacitor 70 instead of the third rigid member 35 in the electronic device 100 shown in FIG. 1 . The capacitor 70 also functions as a type of rigid member. The capacitor 70 is electrically connected to at least one of the first chip 21, the second chip 22, and the third chip 23 via the wiring layer 10. In particular, in the example shown in FIG. 9 , the capacitor 70 is disposed outside the second bridge chip 32, not between the first bridge chip 31 and the second bridge chip 32. Furthermore, in this example, the capacitor 70 is located below the third chip 23, and partially (or entirely) overlaps with the third chip 23 in a plan view.
[0076] Capacitor 70 is composed of two conductive plates (electrodes) and a dielectric (insulator) sandwiched between them. When a voltage is applied to the electrodes, positive charges in the dielectric are attracted to one electrode, and negative charges are attracted to the opposite electrode and stored there. This amount of stored charge becomes the capacitance of capacitor 70. This allows capacitor 70 to transfer the stored charge to chips 21 to 23, or to receive and store charge from chips 21 to 23.
[0077] For example, a capacitor 70 may be connected to a first chip 21, which is a logic element such as a CPU. The first chip 21, which is a logic element, generates large current fluctuations during operation, and the inductance components of the power supply line associated with these current fluctuations generate noise in the power supply. Connecting the capacitor 70 to the first chip 21 can effectively reduce this noise. Furthermore, the effects of electromagnetic noise can be reduced by using a filter circuit that utilizes the capacitor 70 to cut noise components passing through the input / output signal lines of the first chip 21. As shown in FIG. 9 , a capacitor 70 with such a function can be placed in the bridge layer 30, which is relatively close to the first chip 21 (CPU). This enables high-speed transfer of charge between the capacitor 70 and the first chip 21, improving the operating speed and signal waveform shaping performance of the first chip 21.
[0078] Alternatively, a capacitor 70 may be connected to the second chip 22 or the third chip 23, which are memory elements such as DRAM or SRAM. Large current fluctuations occur during operation of the second chip 22, which is a memory element, generating power supply noise. Connecting the capacitor 70 to the second chip 22 can reduce this noise. DRAM memory cells typically have built-in capacitors, and data is stored in these built-in capacitors based on the presence or absence of charge. However, this charge gradually leaks, necessitating periodic read and rewrite (refresh). Therefore, by attaching an external capacitor 70 to the second chip 22, which is a DRAM, a large amount of charge can be temporarily supplied from the capacitor 70 to the second chip 22 during refresh. As shown in FIG. 9 , a capacitor 70 with this function can be arranged in the bridge layer 30, which is relatively close to the second chip 22 and the third chip 23, which are memory elements. This enables high-speed charge transfer between the capacitor 70 and the second chip 22 or the third chip 23, ensuring stable operation of each chip 22, 23.
[0079] The capacitor 70 is preferably a deep trench capacitor (DTC). A deep trench capacitor has a trench formed in a semiconductor substrate, along which a first electrode is formed. The trench is formed perpendicular to the semiconductor substrate, such as silicon, by dry etching or other methods. The trench can have a predetermined shape, such as a cylindrical or rectangular shape, and its depth is typically in the range of 3 to 8 μm. In a deep trench capacitor, a dielectric is provided between the first electrode thus formed and a separate second electrode. Because the effective surface area of the first electrode is increased by the presence of the trench, the deep trench capacitor has a larger charge storage capacity than a typical planar capacitor.
[0080] The dielectric of the capacitor 70 is preferably made of ceramic. Examples of ceramics used as the dielectric include barium titanate (BaTiO 3 ), strontium zirconium titanate (SrZrO 3 ), calcium titanate (CaTiO 3 ), lead titanate (PbTiO 3 ), and tantalum oxide (Ta 2 O 5 ). Other examples include zirconia (ZrO 2 ), alumina (Al 2 O 3 ), and silicon nitride (Si 3 N 4Ceramic materials such as silicon dioxide, silicon dioxide, and silicon dioxide can also be used as the dielectric. These materials may be used alone or in combination as appropriate depending on the application. Ceramic materials combine high rigidity, excellent thermal conductivity, and high insulating properties. By using a ceramic material as the dielectric of the capacitor 70, its high rigidity gives the capacitor 70 a mechanically strong structure, allowing it to withstand vibrations and shocks during the assembly process as well as vibrations during actual use. This improves the reliability of the capacitor 70. Furthermore, the ceramic capacitor 70 exerts the same effect as the rigid members 33-35 described above, helping to adjust the thermal expansion coefficient of the entire bridge layer 30. Furthermore, the ceramic capacitor 70 has excellent thermal conductivity, allowing heat generated from the chips 21-23 to be efficiently dissipated through the capacitor 70. In this way, the capacitor 70 also functions as a heat sink, preventing characteristic degradation and failure due to heat generation from the chips 21-23. Furthermore, the high insulating properties inherent in ceramic materials ensure reliable insulation between the capacitor 70 and the chips 21-23 and the wiring layer 10. This prevents the occurrence of short circuits and leakage current, improving the electrical reliability of the entire electronic device 100.
[0081] Fig. 10 shows a third modified example of the electronic device 100 shown in Fig. 1. The electronic device 100 in Fig. 10 will be described mainly with respect to the configuration different from that shown in Fig. 1, and a description of the same configuration will be omitted. The electronic device 100 shown in Fig. 10 includes a capacitor 70 in the bridge layer 30, similar to the electronic device 100 shown in Fig. 9.
[0082] The electronic device 100 shown in FIG. 10 further includes a capacitor 70 in the bridge layer 30 in addition to the first bridge chip 31, the second bridge chip 32, the first rigid member 33, the second rigid member 34, and the third rigid member 35. This capacitor 70 also functions as a type of rigid member. The capacitor 70 is electrically connected to at least one of the first chip 21, the second chip 22, and the third chip 23 via the wiring layer 10. In particular, in the example shown in FIG. 10 , the capacitor 70 is disposed between the first bridge chip 31 and the second bridge chip 32 and below the first chip 21. Furthermore, in this example, the capacitor 70 entirely (or partially) overlaps the first chip 21 in a plan view. In particular, in the example shown in FIG. 10 , the capacitor 70 is disposed between the first bridge chip 31 and the first rigid member 33, and also between the second bridge chip 32 and the first rigid member 33.
[0083] In this manner, it is preferable to arrange multiple capacitors 70 directly below the first chip 21 and electrically connect each capacitor 70 to the first chip 21 via the wiring layer 10. The first chip 21 is assumed to be a logic element such as a CPU. By arranging multiple capacitors 70 closely to the first chip 21 in this manner, the length of the wiring connecting the two can be minimized. The shorter the wiring, the lower the wiring resistance and wiring inductance. Furthermore, for the first chip 21, which performs high-speed switching, power supply noise caused by wiring inductance can be a problem. However, by distributing multiple capacitors 70 directly below the first chip 21, the impact of this high-speed switching noise can be reduced. As a result, the high-speed operation of the first chip 21 can be more stable. Furthermore, distributing multiple capacitors 70 directly below the first chip 21 ensures multiple heat dissipation paths from the first chip 21, enabling efficient heat dissipation. Furthermore, arranging the first chip 21 directly below the capacitors 70 also leads to efficient use of the mounting area. That is, by overlapping the first chip 21 and the capacitor 70, the wiring space can be effectively utilized, and the electronic device 100 as a whole can be made smaller.
[0084] FIG. 11 shows a fourth modified example of the electronic device 100 shown in FIG. 1 . The electronic device 100 shown in FIG. 11 will be described mainly with respect to the configuration that differs from that shown in FIG. 1 , and a description of the same configuration will be omitted. The electronic device 100 shown in FIG. 11 includes a capacitor 70 in the bridge layer 30, similar to the electronic device 100 shown in FIG. 9 , but this capacitor 70 is incorporated into each of the first bridge chip 31 and the second bridge chip 32. Although not shown, it is also possible to incorporate the capacitor 70 into either the first bridge chip 31 or the second bridge chip 32. By incorporating the capacitor 70 in the bridge chips 31 and 32 in this way, the bridge chips 31 and 32 can connect the two chips 21 to 23 and supply electric charge to each of the chips 21 to 23, for example.
[0085] FIG. 12 illustrates a fifth modified example of the electronic device 100 illustrated in FIG. 1 . The electronic device 100 illustrated in FIG. 12 will be described mainly with respect to the configurations different from those illustrated in FIG. 1 , and the same configurations will not be described again. The electronic device 100 illustrated in FIG. 12 includes a capacitor 70 in the bridge layer 30, similar to the electronic device 100 illustrated in FIG. 9 . However, the capacitor 70 is embedded in the first rigid member 33 (connecting rigid member) located between the two bridge chips 31 and 32. The capacitor 70 of the modified example illustrated in FIG. 12 basically achieves the same effects as the capacitor 70 of the modified example illustrated in FIG. 10 . However, in the modified example illustrated in FIG. 12 , a component in which the capacitor 70 is pre-embedded in the first rigid member 33 can be mounted on the electronic device 100. Therefore, compared to the modified example illustrated in FIG. 10 in which the capacitor 70 and the first rigid member 33 are separate components, the number of components is reduced, thereby improving the manufacturing efficiency of the electronic device 100.
[0086] In the above description of the present invention, the embodiments of the present invention have been described with reference to the drawings in order to express the contents of the present invention. However, the present invention is not limited to the above embodiments, and includes modifications and improvements that are obvious to those skilled in the art based on the matters described in the present specification.
[0087] The present invention can be suitably used in, for example, the semiconductor device manufacturing industry.
[0088] REFERENCE SIGNS LIST 10...first wiring layer 20...chip layer 21...first chip 21a...first logic element 21b...second logic element 22...second chip 23...third chip 24...conductive member 25...first resin part 30...bridge layer 30'...chip overlapping area 30a...bridge chip arrangement area 30b...bridge chip non-arrangement area 30c...rigid member arrangement area 31...first bridge chip 31a...wiring 32...second bridge chip 32a...wiring 33...first rigid member 33a...wiring 33b...rigid member for logic element 33c...rigid member for logic element 34...second rigid member 34a...wiring 35...third rigid member 35a...wiring 36...conductive member 37...through via 38...second resin part 39...third bridge chip 40...second wiring layer 50...solder ball 61...Fourth rigid member 62...Fifth rigid member 63...Single rigid member 64...Fourth bridge chip 70...Capacitor 100...Electronic device 100(a)...First chip set 100(b)...Second chip set 100(c)...Third chip set 100(d)...Fourth chip set 200...Printed circuit board
Claims
1. An electronic device having a wiring layer including a conductive member, a chip layer provided on one side of the wiring layer, and a bridge layer provided on the other side of the wiring layer, wherein the chip layer includes at least two chips and a first resin part that seals the chips, and the bridge layer includes a plurality of bridge chips electrically connected to the chips by the conductive member, one or more rigid members, and a second resin part that seals the bridge chips and the rigid members, and the one or more rigid members includes one disposed between the bridge chips.
2. The electronic device according to claim 1, wherein the one or more rigid members include a non-connected rigid member that is not electrically connected to the chip.
3. The electronic device according to claim 1, wherein the rigid member has thermal conductivity.
4. The electronic device according to claim 1, wherein the rigid member has insulating properties.
5. The electronic device according to claim 1, wherein the rigid member is made of one or more materials selected from the group consisting of silicon, metal, ceramic, and glass.
6. The electronic device of claim 1, wherein the one or more rigid members further include a connecting rigid member electrically connected to the chip.
7. The electronic device of claim 1, wherein a region of the bridge layer that overlaps with the chip of the chip layer is defined as a chip overlapping region, and when the chip overlapping region of the bridge layer is divided into a bridge chip placement region where the bridge chip is placed and a bridge chip non-placement region where the bridge chip is not placed, the rigid member is positioned so as to occupy an area of 30% or more of the bridge chip non-placement region.
8. The electronic device of claim 1, wherein the bridge layer includes an inter-chip rigid member disposed between the bridge chips and a non-inter-chip rigid member disposed at a position not between the bridge chips, the inter-chip rigid member entirely overlapping the chip in the chip layer, and the non-inter-chip rigid member has a portion overlapping the chip in the chip layer and a remainder not overlapping the chip in the chip layer.
9. The electronic device according to claim 1, wherein the bridge layer has a coefficient of thermal expansion (CTE) of 80 to 120% when the chip layer has a coefficient of thermal expansion (CTE) of 100%.
10. The electronic device according to claim 1, wherein the one or more rigid members include a capacitor electrically connected to at least one of the chips via the wiring layer.
11. The electronic device according to claim 10, wherein the capacitor is disposed between the bridge chips and at a position overlapping at least one of the chips.
12. The electronic device according to claim 10, wherein the capacitor is formed from a material containing ceramic as a dielectric.
13. The electronic device of claim 10, wherein the capacitor is a deep trench capacitor.
14. The electronic device according to claim 1, wherein at least one of the bridge chips includes a capacitor electrically connected to at least one of the chips via the wiring layer.
15. The electronic device of claim 6, wherein the connected rigid member includes a capacitor electrically connected to at least one of the chips through the wiring layer.
16. A method for manufacturing an electronic device, comprising the steps of: arranging at least two chips on one side of a wiring layer including a conductive member, and then sealing the chips with resin to form a chip layer; and arranging a plurality of bridge chips electrically connected to the chips by the conductive member, and one or more rigid members, on the other side of the wiring layer, and then sealing the bridge chip and the rigid member with resin to form a bridge layer, wherein the one or more rigid members include one arranged between the bridge chips.
Citation Information
Patent Citations
Packaging structure and forming method thereof
CN116072664A
Proximity communication package for processor, cache and memory
US20090089466A1
Embedded Organic Interposers for High Bandwidth
US20180315735A1
Heterogeneous nested interposer package for IC chips
US20200395313A1
Substrate and semiconductor device package and method for manufacturing the same
US20210327815A1
Cited By
Symmetric dummy bridge design for FLI alignment improvement
US20240186251A1
