WO2011030467A1 - 半導体装置 - Google Patents
半導体装置 Download PDFInfo
- Publication number
- WO2011030467A1 WO2011030467A1 PCT/JP2009/066040 JP2009066040W WO2011030467A1 WO 2011030467 A1 WO2011030467 A1 WO 2011030467A1 JP 2009066040 W JP2009066040 W JP 2009066040W WO 2011030467 A1 WO2011030467 A1 WO 2011030467A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- power supply
- vias
- lsi
- power
- tsv
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L22/00—Testing or measuring during manufacture or treatment; Reliability measurements, i.e. testing of parts without further processing to modify the parts as such; Structural arrangements therefor
- H01L22/30—Structural arrangements specially adapted for testing or measuring during manufacture or treatment, or specially adapted for reliability measurements
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L23/00—Details of semiconductor or other solid state devices
- H01L23/52—Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames
- H01L23/522—Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames including external interconnections consisting of a multilayer structure of conductive and insulating layers inseparably formed on the semiconductor body
- H01L23/528—Geometry or layout of the interconnection structure
- H01L23/5286—Arrangements of power or ground buses
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L25/00—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof
- H01L25/03—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof all the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N, e.g. assemblies of rectifier diodes
- H01L25/04—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof all the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N, e.g. assemblies of rectifier diodes the devices not having separate containers
- H01L25/065—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof all the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N, e.g. assemblies of rectifier diodes the devices not having separate containers the devices being of a type provided for in group H01L27/00
- H01L25/0652—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof all the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N, e.g. assemblies of rectifier diodes the devices not having separate containers the devices being of a type provided for in group H01L27/00 the devices being arranged next and on each other, i.e. mixed assemblies
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L25/00—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof
- H01L25/03—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof all the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N, e.g. assemblies of rectifier diodes
- H01L25/04—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof all the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N, e.g. assemblies of rectifier diodes the devices not having separate containers
- H01L25/065—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof all the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N, e.g. assemblies of rectifier diodes the devices not having separate containers the devices being of a type provided for in group H01L27/00
- H01L25/0657—Stacked arrangements of devices
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L25/00—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof
- H01L25/18—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof the devices being of types provided for in two or more different subgroups of the same main group of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2225/00—Details relating to assemblies covered by the group H01L25/00 but not provided for in its subgroups
- H01L2225/03—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00
- H01L2225/04—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00 the devices not having separate containers
- H01L2225/065—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00 the devices not having separate containers the devices being of a type provided for in group H01L27/00
- H01L2225/06503—Stacked arrangements of devices
- H01L2225/06513—Bump or bump-like direct electrical connections between devices, e.g. flip-chip connection, solder bumps
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2225/00—Details relating to assemblies covered by the group H01L25/00 but not provided for in its subgroups
- H01L2225/03—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00
- H01L2225/04—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00 the devices not having separate containers
- H01L2225/065—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00 the devices not having separate containers the devices being of a type provided for in group H01L27/00
- H01L2225/06503—Stacked arrangements of devices
- H01L2225/06541—Conductive via connections through the device, e.g. vertical interconnects, through silicon via [TSV]
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L23/00—Details of semiconductor or other solid state devices
- H01L23/48—Arrangements for conducting electric current to or from the solid state body in operation, e.g. leads, terminal arrangements ; Selection of materials therefor
- H01L23/50—Arrangements for conducting electric current to or from the solid state body in operation, e.g. leads, terminal arrangements ; Selection of materials therefor for integrated circuit devices, e.g. power bus, number of leads
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2924/00—Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
- H01L2924/0001—Technical content checked by a classifier
- H01L2924/0002—Not covered by any one of groups H01L24/00, H01L24/00 and H01L2224/00
Definitions
- the present invention relates to a semiconductor device in which a plurality of semiconductor integrated circuits are three-dimensionally integrated using a through silicon via (TSV).
- TSV through silicon via
- Patent Document 1 a technique in which a plurality of chips are three-dimensionally stacked and each chip is connected using through-silicon vias to perform inter-chip communication is disclosed in Patent Document 1. Is disclosed. Further, Patent Document 2 introduces a technique for finding a short-circuit defect of a through electrode used for a signal.
- the communication distance between chips can be shortened, and deterioration in performance due to communication between chips can be suppressed. It becomes.
- the circuits in the chip operate at high speed, the power consumed by each chip is relatively large, and it is necessary to supply sufficient power to each chip.
- the chips are not integrated three-dimensionally as in the past, the power consumed by one chip is limited, so power is supplied by wiring to the package substrate on which the chips are arranged. I was able to do it.
- the power to be supplied increases as the number of stacked layers increases. Accordingly, it is necessary to prepare a large number of TSVs for supplying power.
- TSV when used, it is necessary to form a hole penetrating the LSI silicon substrate and form a metal wiring penetrating the chip. Therefore, a relatively large circuit such as a CPU or the like is formed around the through via. Difficult to place on-chip memory. Therefore, it is necessary to examine how it is efficient to arrange a large number of power supply TSVs.
- one type of chip may be stacked, or multiple types, for example, different chips such as a processor chip and a memory chip must be stacked. In this case, it is difficult to adopt the arrangement of through vias optimized for the circuit mounted on each chip.
- the stacked LSIs do not operate at all or abnormal operation occurs. Further, when the number of defects is small relative to the total number of electrodes, abnormal operation may occur only in a specific operation state of the LSI.
- a conventional LSI power supply test method a method of inspecting by applying a probe to an electrode on the wafer surface at the time of manufacture has been adopted.
- a test method for stacked LSIs connected by through silicon vias in particular, a test method for the connection state of through silicon vias, is not assumed in the conventional test methods.
- the test is performed by applying a probe to the electrode on the wafer surface. I can't test.
- the electrodes of the LSI in the lower layer are hidden by the LSI in the upper layer, and the probe cannot be applied for inspection.
- the method introduced in Patent Document 2 for discovering a short-circuit defect between a through electrode used for a signal and an LSI substrate is a defect in the disconnection of a through electrode or a poor contact between the through electrodes that occurs after stacking a plurality of LSIs. Etc. cannot be detected. When disconnection failure or contact failure occurs, necessary and sufficient power is not supplied, and operation failure occurs. Therefore, it is necessary to inspect for these failures.
- the number of stacked layers can be changed from a chip having the same configuration to develop various products. Therefore, when a test is performed from outside the apparatus, it is necessary to change the test procedure for each number of stacked sheets, and there is a problem that the inspection process becomes complicated.
- an object of the present invention is to supply sufficient power to three-dimensionally stacked LSI chips and to arrange common through vias in different types of chips. It is also to propose a new test method for through silicon vias for power supplies.
- a typical semiconductor device of the present invention is as follows. First, regarding the arrangement of the power supply through vias, a first semiconductor having first and second sides extending in a first direction and third and fourth sides extending in a second direction intersecting the first direction. A first circuit block formed on the substrate and a plurality of signals connected to the first semiconductor substrate and transmitting a signal output from the first circuit block to a second circuit block formed on another second semiconductor substrate A plurality of power supply through vias for supplying power to the first circuit block, the plurality of power supply through vias extending along the third side and the fourth side.
- One semiconductor substrate is formed at the end and formed in a plurality of rows in the first direction.
- the first processing unit having the normal mode and the power consumption mode, the first LSI having the first power supply test circuit, and the first LSI are stacked and have the normal mode and the power consumption mode.
- the power consumption of the first processing unit in the mode is greater than the power consumption of the first processing unit in the normal mode, and the power consumption of the second processing unit in the power consumption mode is the power consumption of the second processing unit in the normal mode.
- the first power supply And the second power supply test circuit measures the voltage value of the operating voltage supplied through the power supply through-via when the first processing unit is in the power consumption mode. When the unit is in the power consumption mode, the voltage value of the operating voltage supplied through the power supply through via is measured.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- 1 is a cross-sectional view of a laminated LSI to which the present invention is applied.
- FIG. 2 is a schematic diagram of a layout of TSVs and circuits of a laminated LSI to which the present invention is applied.
- FIG. 2 is a schematic diagram of a layout of TSVs and circuits of a laminated LSI to which the present invention is applied.
- 1 is a cross-sectional view of a laminated LSI to which the present invention is applied. It is the schematic of the power supply wiring of the laminated LSI to which this invention was applied.
- FIG. 1 is a cross-sectional view of a laminated LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- 1 is a cross-sectional view of a laminated LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- 1 is a cross-sectional view of a laminated LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- 1 is a cross-sectional view of a laminated LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied. It is the schematic of the power supply wiring of the laminated LSI to which this invention was applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- 1 is a cross-sectional view of a laminated LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- 1 is a cross-sectional view of a laminated LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- 1 is a cross-sectional view of a laminated LSI to which the present invention is applied.
- FIG. 3 is a schematic diagram of an arrangement of TSVs of a stacked LSI to which the present invention is applied.
- 1 is a cross-sectional view of a laminated LSI to which the present invention is applied. It is a conceptual diagram of the cross-sectional direction of the semiconductor device which concerns on this invention. It is a circuit diagram which shows the structure of the principal part of the semiconductor device which concerns on this invention.
- FIG. 5 is a schematic diagram showing a part of a circuit structure for simulating the maximum power in a functional block to be mounted in a semiconductor device according to the present invention.
- 1 is a schematic diagram showing an embodiment of a functional block configuration for performing a power supply test in a semiconductor device according to the present invention.
- 2 shows a procedure of a power supply test in a semiconductor device according to the present invention.
- 2 shows a table configuration for recording a power test result in a semiconductor device according to the present invention.
- It is a circuit diagram which shows the structure of the power supply of the semiconductor device which concerns on this invention.
- It is a circuit diagram which shows the structure of the principal part of the semiconductor device which concerns on this invention.
- 1 is a schematic diagram showing an embodiment of a functional block configuration for performing a power supply test in a semiconductor device according to the present invention.
- 2 shows a procedure of a power supply test in a semiconductor device according to the present invention.
- FIG. 1 shows a schematic arrangement of through silicon vias (TSV) of LSIs that are three-dimensionally stacked using the present invention.
- TSV through silicon vias
- black circles are power supply TSVs for supplying power to the stacked chips
- white circles are signal line TSVs for transmitting signals between the stacked chips.
- a plurality of power supply TSVs are arranged on the outer periphery of the chip, and signal line TSVs are arranged on the inner side.
- FIG. 2 shows a cross-sectional structure when two CPU chips each having a plurality of CPUs (Central Processing Units) and two memory chips each having a memory array are stacked using this TSV arrangement configuration.
- FIG. 2 is a cross section of the chip of FIG.
- the power supply TSV and the signal line TSV connecting the respective chips are arranged at the same place in the vertical direction.
- Each chip has a metal wiring layer stacked downward, and TSV is connected to the metal wiring layer of the chip stacked above, and then penetrates through the silicon layer of the chip stacked below. Are connected to the metal wiring layers of the chips stacked on the substrate. This is shown in FIG. Here, the lowermost chip is connected to the package substrate. Here, the lowermost chip has a metal wiring layer facing downward in order to shorten the distance of the signal line to the package substrate, and all chips are laminated downward accordingly.
- FIG. 3 shows an example of the arrangement of TSVs and CPUs on a CPU chip using the present invention.
- this LSI chip eight CPU cores CPU1 to CPU8 are integrated.
- TSVs for power supply and signal lines are arranged around the chip, and a CPU is arranged at the center of the chip.
- circuit blocks such as a CPU are arranged in an area where no TSV is placed.
- FIG. 4 shows an example of TSV arrangement and memory arrangement of a memory chip using the present invention.
- TSVs for power supply and signal lines are arranged around the chip, and RAM1 and RAM2 which are memory arrays are arranged at the center of the chip. In this way, circuit blocks such as memory are arranged in areas where TSVs are not placed.
- a plurality of power supply TSVs are arranged on the outermost periphery of the chip.
- the circuit block and the signal line TSV are arranged on the outermost periphery, the circuit block and the signal line TSV can be freely arranged and wired inside.
- the power supply TSV is arranged on the outermost periphery of the chip, and the signal line TSV is arranged between the power supply TSV and the circuit block.
- the reason why the power supply TSV and the signal line TSV are arranged on the outer periphery of the circuit block is to secure a large area where the circuit block can be arranged, and the signal line TSV is arranged inside the power supply TSV.
- the distance between the circuit block and the signal line TSV is shortened, and stacking between different types of chips is taken into consideration as will be described later.
- FIG. 5 shows a cross section of two CPU chips and two memory chips, and an interface chip and a package substrate for connecting an LSI to the package substrate.
- the LSI chip and the package substrate are connected by metal bumps.
- a TSV is connected to the upper layer metal wiring of the chip stacked above, and the TSV is formed through the silicon of the chip stacked below and connected to the upper layer metal wiring of the lower chip. Is done. Therefore, the metal wiring on the upper layer of each chip is designed to be connected to the TSV. That is, in the example of FIG. 5, when two memory chips are made in the same shape, the upper metal wiring of the lower memory chip is configured to be connected to the TSV.
- the pitch between TSVs is narrower than the wiring pitch of the package substrate because a large number of TSVs are arranged. Therefore, it is difficult to directly connect metal bumps for connecting to the package substrate. Therefore, the interface chip is inserted between the lowermost memory chip and the package substrate, and the metal wiring is designed in a shape connectable to the TSV and the metal bump, thereby connecting the stacked CPU chip and memory chip to the package substrate. It becomes possible. That is, in the interface chip, the wiring for bump connection can be formed by extending the wiring from the TSV provided on the outer periphery to the inside by the uppermost wiring.
- FIG. 6 shows an arrangement example of power supply wirings connected to the TSV of the LSI chip using the present invention.
- the power supply wiring is a power supply wiring extending in the first direction (left-right direction in the figure) by the uppermost wiring layer and a power supply wiring extending in the second direction (up-down direction in the figure) by the lower layer wiring. It is composed of wiring, and upper and lower power supply wirings are connected at each intersection or arbitrary intersection, and wired in a mesh shape.
- the power supply wiring passing through the central portion of the chip is configured to pass through at intervals, such as one out of two.
- the signal wiring of the signal line TSV provided inside the power supply TSV can be easily passed.
- the metal wiring layer connected to the TSV is necessary for the upper layer wiring of the chip in order to connect to the upper side of the TSV, and is also necessary for the lower layer wiring of the chip to connect to the lower side of the TSV penetrating silicon. . Since the lower layer wiring of the chip is also used for normal wiring, there is a demand for reducing the occupancy ratio of the area where the operation circuit is placed as much as possible.
- the density of the power supply wiring is increased for the portion connected to the TSV, and the density of the power supply wiring is increased for the portion where the operation circuit such as the CPU circuit is arranged. It is possible to lower the signal wiring.
- FIG. 7 is a cross-sectional view showing the connection of the LSI chip, package substrate, TSV, and metal bump using the present invention.
- the TSV is connected to the metal wiring in the upper layer of the chip stacked above, and is connected to the metal wiring in the lower layer of the chip through the silicon portion of the chip stacked below.
- Each bump connects the metal wiring of the upper layer of the chip laminated on the lowermost layer and the package substrate.
- FIG. 8A shows an allocation diagram of LSI power supply TSVs using the present invention.
- LSI power supplies include a ground potential power supply VSS and a high potential power supply VDD.
- VSS ground potential power supply
- VDD high potential power supply
- the TSV density is actually low, and the wiring can be easily passed.
- a TSV having a diameter of 10 ⁇ m can be arranged at a 50 ⁇ m pitch.
- a space of 40 ⁇ m can be provided between the TSVs, and a plurality of power supply wirings can be easily arranged between the TSVs. is there.
- the power supply line is made thick and convex. As a result, the power supply line and the power supply TSV can be reliably connected.
- the present invention when used, it is possible to supply power to LSI chips stacked three-dimensionally. Further, by arranging a plurality of power supply TSVs, a large current can be supplied to a stacked LSI in which a plurality of chips are stacked and a large amount of current is consumed.
- FIG. 9 is a schematic view of TSV arrangement when chips of different sizes are stacked three-dimensionally using the present invention.
- the chip A is, for example, a CPU chip on which a plurality of CPUs are mounted
- the chip B is, for example, a memory chip on which a memory is mounted. Since memory chips and the like have relatively low power consumption, the number of TSVs for power supply can be reduced. Therefore, the number of TSVs for power supply is larger in the chip A than in the chip B. Further, the number of signal line TSVs necessary for communication with another chip differs between chip A and chip B.
- FIG. 10 is a cross section when the chip of FIG. 9 is cut along BB ′.
- the power supply TSV and the signal line TSV that connect each chip are arranged in the same place in the vertical direction, and the TSVs on the right and lower half of the chip exist for the connection of the chip B. Not done.
- the power supply TSV is arranged on the outermost periphery, and the signal TSV is arranged on the inner side thereof. It becomes possible to stack chips of different sizes. In addition, if one side is aligned in the same way, it is possible to stack smaller chips.
- the present invention when used, it is possible to supply power to LSI chips stacked three-dimensionally and to stack different types of LSI chips.
- the power supply TSVs on the outer periphery of the chip, it is possible to easily arrange the TSVs that supply power when LSIs having different sizes are stacked.
- FIG. 11 is a schematic view of TSV arrangement when chips of different sizes are stacked three-dimensionally using the present invention.
- the chip A is, for example, a CPU chip on which a plurality of CPUs are mounted
- the chip C is, for example, a memory chip on which a memory is mounted. Since memory chips and the like have relatively low power consumption, the number of TSVs for power supply can be reduced, and the number of TSVs for power supply is larger in chip A than in chip B. In the present embodiment, unlike the embodiment 1-2, the number of signal TSVs can be made equal.
- the power TSVs are arranged in a plurality of rows on the outermost periphery, so that it is possible to match the positions of the signal TSVs by not connecting a part of the power TSVs of the large chip to the small chips. . Therefore, in this case, it is necessary to align the positions of the signal line TSVs in the chip A and the chip C.
- FIG. 12 is a cross section when the chip of FIG. 11 is cut along C-C ′.
- the signal line TSVs connecting the respective chips are arranged at the same place in the vertical direction.
- the TSV for power supply exists on the chip A that needs to supply a large amount of power, but there is a portion that does not exist on the chip C.
- chips having different sizes and the same number of signal lines can be stacked.
- the present invention when used, it is possible to supply power to LSI chips stacked three-dimensionally and to stack different types of LSI chips. Further, by arranging power supply TSVs on the outer periphery of the chip, when LSIs of different sizes are stacked, TSVs for supplying power can be easily arranged and the same number of communication TSVs can be arranged between different chips. Is possible.
- FIG. 13 is a schematic view showing the arrangement of TSVs when chips of different sizes are three-dimensionally stacked using the present invention.
- Chip A is a CPU chip on which a plurality of CPUs are mounted, for example
- chip D is a memory chip on which a memory is mounted, for example.
- a power supply wiring and a signal wiring are provided between the chip A and the chip D. It is necessary to insert an interface chip in order to connect.
- FIG. 14 is a cross-sectional view of the chip of FIG. 13 including the interface chip when cut by D-D ′.
- the power supply and signal line TSVs are not arranged at the same position in the vertical direction. Therefore, an interface chip is inserted between them, and the power supply wirings and signal lines of the upper and lower chips A and D are connected at that portion. In this way, when stacking chips having different sizes and having different TSVs of power supply lines and signal lines, the interface power supply and signal lines can be connected by inserting and stacking interface chips. It becomes possible.
- the present invention when used, it is possible to supply power to LSI chips stacked three-dimensionally and to stack different types of LSI chips. Further, by arranging the power supply TSV on the outer periphery of the chip, when LSIs having different sizes are stacked, it is possible to supply substantially the same current to all the chips.
- FIG. 15A shows an allocation diagram of LSI power supply TSVs using the present invention. Except for assigning VDD to the outermost TSV of the chip and assigning VSS to the inner TSV, this is the same as in Example 1-1.
- VDD voltage division duplex
- VSS voltage division multiplexing
- FIG. 15 (b) either one of VDD or VSS power supply wiring may be provided between the power supply TSVs, and the number of power supply lines is reduced as compared with FIG. Can do. Therefore, it is possible to increase the thickness of each power supply line, and to reduce the resistance value.
- the present invention when used, it is possible to supply power to LSI chips stacked three-dimensionally and to stack different types of LSI chips.
- the allocation of TSVs to VDD and VSS from the outer periphery, it is possible to pass one power supply line for one TSV column, and it is possible to increase the width of the power supply line. .
- FIG. 16A shows an allocation diagram of LSI power supply TSVs using the present invention.
- power supply line TSVs are assigned to four columns on the outer periphery of the chip.
- the same kind of power sources that is, VDD and VSS are allocated to four neighboring TSVs, and they are arranged in a staggered pattern.
- Other configurations are the same as those in Example 1-1.
- the number of TSVs for power supply can be made equal to VDD and VSS, and for example, when one TSV is destroyed during manufacturing, the system does not operate for some reason.
- either one of the power lines VDD or VSS may be provided between the power supply TSVs (FIG. 16B). Can be made thicker.
- the present invention when used, it is possible to supply power to LSI chips stacked three-dimensionally and to stack different types of LSI chips. Further, by allocating TSVs to VDD and VSS by a plurality of TSVs, it is possible to form a power TSV that is resistant to destruction during manufacturing, and one power line for one TSV column. Therefore, it is possible to increase the width of the power supply line.
- FIG. 17 shows an allocation diagram of LSI power supply TSVs using the present invention.
- power supply line TSVs are assigned to four columns on the outer periphery of the chip.
- the same type of power supplies that is, VDD and VSS are assigned to TSVs in the same row arranged in the horizontal direction.
- Other configurations are the same as those in Example 1-1.
- the number of TSVs for power supply can be made equal to VDD and VSS, and the power supply wiring layer to which the TSV is connected can be arranged in the horizontal direction of the chip along the TSV. It becomes.
- FIG. 18 shows the configuration of the power supply wiring layer connected to the TSV.
- the power supply lines are wired between the power supply TSVs.
- the same type of power supplies are arranged in the left-right direction, they are not wired between the power supply TSVs.
- wiring can be performed on the power TSV, and the power line can be made thicker.
- in order to connect the power supply wiring and the power supply TSV it is necessary to form a convex shape with a part of the power supply line thickened.
- This wiring layer is a wiring layer on the upper layer of the chip connected to the upper surface of the TSV, and shows a state of the lower wiring layer of the chip connected to the TSV penetrating silicon.
- the wiring layer in the lower layer of the chip has a shape through which the TSV penetrates. Therefore, in another embodiment, by providing the above-described convex shape, the lower layer can be connected to the power supply TSV.
- the power supply line is thicker than TSV, a convex shape is unnecessary, and any part of the wiring may be in contact with the power supply line and power supply in the lower wiring layer. Connection of TSV becomes easy.
- the present invention when used, it is possible to supply power to LSI chips stacked three-dimensionally and to stack different types of LSI chips. Further, by arranging the power supply types supplied by the TSVs arranged in the horizontal direction, it is possible to arrange the power supply lines for supplying power on the TSVs.
- FIG. 19A shows an allocation diagram of LSI power supply TSVs using the present invention.
- power supply line TSVs are assigned to four columns on the outer periphery of the chip.
- this chip shows a configuration having a high power supply voltage VDDH, a low power supply voltage VDDL, and a ground potential power supply VSS.
- power supplies of the same type that is, VDDH, VDDL, and VSS are assigned to four neighboring TSVs, and these are arranged in a staggered pattern.
- the number of TSVs is equal to VSS, which is the sum of the number of TSVs for VDDH and the number of TSVs for VDDL.
- the number of TSVs for power supply can be made equal to the number obtained by adding VDDH and VDDL and the number of VSS, and for example, one TSV is destroyed at the time of manufacturing. Even in the case where the system does not operate at the same time, current concentration does not occur because the same type of TSVs for power supply are arranged around. That is, by adopting this configuration, it is possible to arrange a power supply TSV that is resistant to destruction during manufacturing.
- VDDH or VDDL TSVs are arranged in the same row (column) in the horizontal direction or the vertical direction. Therefore, as shown in FIG. 19B, only one of the VDDH, VDDL, and VSS power lines needs to be provided between the power TSVs, and the power lines can be made thicker. It becomes possible.
- the present invention when used, it is possible to supply power to LSI chips stacked three-dimensionally and to stack different types of LSI chips.
- the three types of power supply in a solid manner, it is possible to pass one power supply line through one TSV column, and the width of the power supply line can be increased.
- FIG. 20A shows an allocation diagram of LSI power supply TSVs using the present invention.
- power supply line TSVs are assigned to four columns on the outer periphery of the chip.
- this chip shows a configuration having a high power supply voltage VDDH, a low power supply voltage VDDL, and a ground potential power supply VSS.
- power sources on the VDD side and the VSS side are assigned to four neighboring TSVs, and these are arranged in a staggered pattern.
- the number of TSVs is equal to VSS, which is the sum of the number of TSVs for VDDH and the number of TSVs for VDDL.
- the current supplied from VDDH and VDDL needs to flow through VSS, and the number of TSVs is determined by the current capacity. Furthermore, since VDDH supplies a higher potential than VDDL, it is considered that a large amount of current needs to be supplied. Therefore, more TSVs are assigned to VDDH than VDDL in order to equalize the current capacity. As shown in FIG. 20B, the uppermost power supply wiring adjacent to the TSV is assigned to VDDH and VSS, and connected from each TSV by wiring. Further, the power supply wiring adjacent to the TSV one layer below the uppermost layer is assigned to VDDH, VDDL, and VSS, and is connected from each TSV by wiring.
- FIG. 21 is a schematic view of TSV arrangement when chips of different sizes are stacked three-dimensionally using the present invention.
- the chip A is, for example, a CPU chip on which a plurality of CPUs are mounted
- the chip E is, for example, a memory chip on which a memory is mounted. Basically, it is based on the same concept as in the embodiment 1-2, but in this example, an example in which the size of the chip E is about 1/4 of the chip A is shown. In this case, it is possible to stack chips having different areas by disposing the TSV for power supply on the outer peripheral portion of the chip A and disposing and stacking the TSV of the chip E at the same position above the chip A.
- FIG. 22 is a cross-sectional view of the chip of FIG. 21 taken along E-E ′.
- the power supply and signal line TSVs connecting the respective chips are arranged at the same location in the vertical direction. Thus, if the positions of the signal line TSVs are aligned, chips having different sizes and the same number of signal lines can be stacked.
- FIG. 23 is a schematic diagram of TSV arrangement when a plurality of types of chips are stacked using the present invention.
- the power supply TSVs are arranged at the outer periphery of the chip as in the first embodiment, and the signal line TSVs are arranged at the center of the chip, unlike the first embodiment.
- FIG. 24 is a cross-sectional view when the chip of FIG. 23 is cut by F-F ′.
- This example shows a configuration in which two CPU chips each having a plurality of CPUs and two memory chips each having a memory are stacked.
- the power supply and signal line TSVs connecting the respective chips are arranged at the same location in the vertical direction.
- FIG. 25 shows an arrangement configuration example of each CPU circuit in the CPU chip. Eight CPUs are mounted on this chip, and power supply TSVs are arranged on the outer periphery. Further, a signal line TSV is disposed in a portion where the CPU in the center of the chip is not disposed.
- FIG. 26 shows an arrangement configuration example of each memory in the memory chip. In the memory chip, like the CPU chip, a memory circuit is arranged at the center of the chip, and a signal line TSV is arranged in a portion where the memory circuit is not arranged. In this configuration, necessary power supply capability is ensured by arranging the power supply wiring TSVs on the outer periphery of the chip, and the signal line TSVs are effectively utilized in the gaps of the internal circuits. It becomes possible to arrange the signal line TSV. Furthermore, since it is possible to arrange a TSV for a signal line in the vicinity of the internal operation circuit, the performance when performing communication through the signal line is improved.
- FIG. 27 is a schematic diagram of TSV arrangement when a plurality of types of chips are stacked using the present invention.
- a plurality of power supply TSVs are arranged on the left and right outer peripheral portions of the chip, and a plurality of signal line TSVs are arranged inside the TSVs. All the stacked chips have the same TSV arrangement.
- TSVs are not disposed on the upper and lower peripheral portions of the chip.
- FIG. 27 shows the case where chips having substantially the same area are stacked, but the configuration of this embodiment may be advantageous when a plurality of chips having a small area are stacked.
- FIG. 28 shows an arrangement example of TSVs in this case and a schematic diagram of each chip.
- the chip G is a chip having a large area
- the chip H is a chip having an area of about 1/8 of the area of the chip G.
- TSVs are arranged on the outer periphery of the chip in the left-right direction.
- a plurality of chips having a small area are stacked at the same position as the TSV.
- the arrangement efficiency is higher than that of the configuration of Embodiment 1-2.
- FIG. 29 shows a schematic arrangement of coils used for communication between a TSV and chips when a plurality of types of chips are stacked using the present invention.
- TSVs are used for signal communication between chips when the chips are three-dimensionally stacked.
- a configuration example in the case of using inductively coupled communication using a coil for communication between chips will be described.
- a plurality of power supply TSVs are arranged on the outer periphery of the chip.
- an operation circuit of each chip such as a CPU and a memory, is disposed at the center of the chip, and a coil for chip-through credit is formed by using the upper layer wiring.
- FIG. 30 is a cross-sectional view of the chip of FIG. 29 taken along the line I-I ′.
- This example shows a configuration in which two CPU chips each having a plurality of CPUs and two memory chips each having a memory are stacked.
- the TSVs for power supply connecting the respective chips are arranged at the same place in the vertical direction, and further the communication coils for performing communication between the chips are arranged at the same place in the vertical direction.
- the TSV for power supply is arranged on the outer periphery of the chip, when the TSV is not used for communication between chips, it is possible to arrange an operation circuit and a circuit used for inter-chip communication at the center of the chip. Therefore, it is possible to arrange a circuit for a laminated chip by effectively using the area in the chip. Furthermore, when the present invention is used, it is possible to supply power to LSI chips stacked three-dimensionally and to stack different types of LSI chips.
- FIG. 31 is a schematic view of TSV arrangement when chips of different sizes are stacked three-dimensionally using the present invention.
- the chip J is a CPU chip on which a plurality of CPUs are mounted, for example
- the chip A is a memory chip on which a memory is mounted, for example.
- the number of TSVs for power supply can be reduced.
- the number of TSVs for power supply is large. Necessary. Further, the number of signal line TSVs required for chip A and chip J is equal.
- FIG. 32 is a cross-sectional view of the chip of FIG. 31 taken along J-J ′.
- the signal line TSVs connecting the respective chips are arranged at the same place in the vertical direction.
- the TSV for power supply exists on the chip J that needs to supply a large amount of power, but there is a portion that does not exist on the chip A.
- chips having different sizes and the same number of signal lines can be stacked.
- the large chip J is arranged on the small chip A, the region where the chip A under the chip J does not exist needs to be filled with an insulating layer or the like.
- a chip with large power is stacked on the upper side of the stacked chip, it is necessary to supply power from the chip side with large power consumption by wire bonding or the like.
- test method described below can be applied to configurations other than the invention described in the first embodiment.
- the entire stacked LSI may be able to operate sufficiently.
- the invention described below is particularly useful because such a situation can be taken into consideration.
- FIG. 33 shows a sectional view of a semiconductor device according to the second embodiment of the present invention and the appearance of defects.
- This embodiment shows a configuration in which semiconductor integrated circuits LSI_A, LSI_B, and LSI_C are stacked above the package substrate PKCB in this order. Each of the three LSIs is stacked with the circuit mounting surface facing downward (face-down).
- the lowermost LSI_A is electrically connected to the package substrate PKCB via the solder bump BMP, and the LSI_B and LSI_C are electrically connected to the lower LSI via the micro bump MBMP and the through silicon via TSV.
- TSV_VSS is a through silicon via for supplying a ground potential to each chip
- TSV_VDD is a through silicon via for supplying a power supply potential to each chip.
- TSV_VSS and TSV_VDD are arranged at the same position in each LSI, and the upper and lower chips are electrically connected via the micro bump MBMP. That is, LSI_B is supplied with power via LSI_A, and LSI_C is supplied with power via LSI_B. LSI_A is supplied with power from the outside via the package substrate PKCB.
- the laminated LSI provides a power supply test means for detecting a power supply failure due to such disconnection or contact failure.
- a power supply test control circuit PTEST_CTRL is a functional block for controlling the power supply test sequence of the stacked LSI package, and is mounted on the lowermost LSI_A.
- the power test circuit PTEST is a functional block for executing a power test of the mounted LSI alone, and is mounted on LSI_B and LSI_C. The PTEST executes a power test for each LSI in response to a control signal from the PTEST_CTRL.
- TSV_SIG is a signal through-silicon via for data communication with LSI_A, LSI_B, and LSI_C.
- TSV_TREQ, TSV_SCK, TSV_TDI (TSV_TDI_A, TSV_TDI_B, TSV_TDI_C) and TSV_TDO (TSV_TDO_B, TSV_TDO_C) are through-silicon vias for signals used for power supply tests.
- TSV_TREQ and TSV_SCK are through-silicon vias for transmitting a signal output from PTEST_CTRL to PTEST, TSV_TREQ is used to control a power test performed by PTEST, and TSV_SCK is a clock signal for reading a power test result. This is a through-silicon via for PTEST.
- TSV_TDI and TSV_TDO are electrically separated inside the LSI. That is, TSV_TDO_C is electrically connected to TSV_TDI_B, and TSV_TDO_B is electrically connected to TSV_TDI_A.
- FIG. 34 shows an embodiment of a logical configuration and connection of functional blocks mounted on LSI_A and LSI_B in FIG.
- the logical configuration and connection of LSI_C are not shown, but are the same as LSI_B.
- TSV_SIG, TSV_TREQ, TSV_SCK, TSV_TDI, and TSV_TDO indicate the same electrodes as those indicated by the same reference numerals in FIG. 33, and the broken lines indicate that there is an electrical connection.
- LSI_A includes a processing unit PU_A0 that is an execution processing unit for arithmetic instructions, 3DI that is a communication interface between stacked LSIs, a test access port TAP that can be connected to an external debugger, an external tester, and the like, A local bus controller LBSC for accessing the device bus, PTEST_CTRL for controlling the power supply test sequence of the stacked LSI package shown in FIG. 33, and an on-chip interconnect OCI_A for connecting the functional blocks in the LSI_A Prepare.
- PTEST_CTRL includes a test result table RSLT_TBL for storing a power test result.
- LSI_A is connected to an external nonvolatile memory NVMEM via the LBSC.
- the nonvolatile memory NVMEM stores a power supply test program PESTPROG describing a power supply test sequence.
- LSI_B has four processing units (PU_B0, PU_B1, PU_B2, PU_B3) as execution processing units for arithmetic instructions, functional modules IP_B0 and IP_B1 for performing specific arithmetic processing and control processing, and a communication interface between stacked LSIs.
- a read-only memory ROM composed of an array of fixed storage elements
- an on-chip interconnect OCI_B that connects functional blocks in LSI_B
- MD_CTRL includes a register MDREG for setting an operation mode.
- the read-only memory ROM stores a maximum power consumption program MAXPROG.
- LSI_B simulates an operation that consumes maximum power by activating all processing units and functional modules mounted on LSI_B for detecting defects in the power supply through silicon via. It is characterized by having a mechanism to perform. Hereinafter, this mechanism is called a maximum power consumption mechanism, which will be described in detail. Note that the “maximum power consumption” in this specification does not mean that the processing unit or functional module consumes the largest amount of power that can be logically considered, and the average power consumption in a predetermined period is This is a state in which it becomes larger than in normal operation (in normal mode, during normal program processing).
- the operation start / end of the maximum power consumption mechanism is controlled by PTEST.
- PTEST asserts the test request signal TREQ
- MD_CTRL writes 1 to a specific bit (power consumption mode transition bit) in the internal register MDREG.
- PTEST negates the test request signal TREQ MD_CTRL writes 0 to the power consumption mode transition bit in MDREG.
- MD_CTRL switches assertion / negation of the power consumption mode transition request signal MAXPREQ according to the content of the power consumption mode transition bit in the MDREG.
- MD_CTRL asserts the power consumption mode transition request signal MAXPREQ
- the processing units PU_B0 to PU_B3 and the functional modules IP_B0 and IP_B1 shift to the power consumption mode.
- MD_CTRL negates MAXPREQ it returns to the state before the power consumption mode transition.
- the maximum power consumption program MAXPPROG is loaded from the read-only memory ROM and executed via the on-chip interconnect OCI_B.
- the maximum power consumption program processing for activating all or most of the circuit blocks included in the processing unit is defined. This can be realized, for example, by executing a specific instruction sequence in a loop.
- the processing unit performs the maximum power consumption operation. In this maximum power consumption operation, a larger number of circuit blocks are constantly activated than in the normal mode, so that a larger amount of power than that in the normal mode is consumed in a predetermined period.
- the function module When the function module (IP_B0 / IP_B1) shifts to the power consumption mode, the function module performs a maximum power consumption operation preset in advance. This can be realized by executing the processing with the heaviest processing load among the processes that can be realized by the functional module. However, when communication with other LSIs is required for such processing, a specific operation mode as described below is provided inside, and the maximum power consumption operation is performed by transitioning to the operation mode.
- FIG. 34 in the case of a functional block in which IP_B0 / IP_B1 needs to communicate with another LSI, for example, an Ethernet interface block that communicates with the outside of the stacked LSI, the receiving-side Ethernet interface is receivable. It is unclear whether it is in a state, and it is unclear whether the communication path to the external input / output pin via the communication interface 3DIC etc. is also operating. Can not be operated to consume. In this case, for example, an operation that consumes the maximum power is simulated by activating all or most of the included flip-flops. This is realized by the following method.
- FIG. 35 is a schematic diagram showing a part of a circuit structure for simulating the maximum power in such a functional block.
- FF is a flip-flop
- CLOGIC is a circuit block configured by combinational logic
- RNDGEN is a pseudo-random number generator
- SEL is a selector that selects an output from two inputs according to a control signal.
- MAXPREQ is the above-described power consumption mode transition request signal
- CLK is an operation clock signal.
- the power consumption of the entire functional module IP_B0 does not become maximum.
- MAXPREQ is asserted, that is, in the power consumption mode
- the output of each flip-flop FF is directly connected to the input of the neighboring flip-flop FF, that is, the entire flip-flop FF has a single chain configuration.
- the starting point of this flip-flop chain is connected to the pseudo random number generator RNDGEN.
- RNDGEN pseudo random number generator
- the power consumption in the functional module is mostly the power required for clock supply for flip-flop operation and the power required for signal transition of the combinational logic circuit CLOGIC.
- the power consumption becomes the maximum value.
- the power consumption in the state in which most of the flip-flop FF is activated is often larger than the power consumption in the normal operation of the function block, but the function block is forcibly brought into the power consumption operation state. Therefore, a highly reliable test can be performed from the viewpoint of inspection of the power supply.
- the configuration in which the flip-flops are connected in a chain shape is often incorporated by a conventional LSI test design, and may be configured by using it. Further, an operation of simulating the maximum power consumption using this flip-flop chain may be applied to the processing units PU_B0 to PU_B3.
- the state in which the LSI_B to which power is supplied via the through silicon via consumes the maximum power in the operation realized by the LSI_B is characterized by comprising an operation mechanism that can be realized without accessing the system.
- the power consumed by LSI_B at the time of design is estimated, and a large number of TSV_VDD are provided so that the drop in power supply potential is sufficiently small. Therefore, the power supply potential of LSI_B during operation rises to a predetermined value.
- the parasitic resistance value of the power supply path rises and falls below a predetermined power supply potential estimated at the time of design. That is, if the power supply potential exceeds a predetermined value during the operation of LSI_B, it is determined that disconnection and contact failure have not occurred in the power supply through silicon via, or even if LSI_B does not affect the operation of LSI_B. it can.
- the power supply potential on the LSI_B is predetermined in a state where the power of the LSI_B is maximized by the maximum power consumption mechanism, that is, in a state where the maximum current flows in the power supply path to the LSI_B.
- the operation of a specific part of the circuit may cause a temporary or local voltage drop, and the LSI_B may malfunction.
- the delay monitor may be provided as necessary, and need not be provided if an operation failure can be detected only by the voltage monitor. On the contrary, the delay monitor may be used to test only that the critical path is operating correctly. In this case, only the delay monitor may be provided.
- LSI_C also has similar characteristics.
- FIG. 36 shows an embodiment of a PTEST configuration for performing a power test of LSI_B in FIG.
- PTEST includes a power supply test control block TESTCTRL, a delay monitor DELAYMON, a voltage monitor VMON, and a shift register SHIFTREG that takes in internal values of DELAYMON and VMON.
- TSV_TREQ, TSV_SCK, TSV_TDI, and TSV_TDO are the same as the signals shown in FIG.
- the power test control block TESTCTRL is a functional block that controls the start / end of the power test.
- the power supply test requires checking the power supply potential and circuit operation as described above.
- TEST_CTRL asserts a test request signal TREQ when TSV_TREQ is asserted.
- the TEST_CTRL activates the delay monitor DELAYMON and the voltage monitor VMON to measure the operation state of the LSI_B.
- the shift register SHIFTREG records a value input from each connected monitor block (delay monitor DELAYMON or voltage monitor VMON) in an internal register when no periodic cycle signal is input from TSV_SCK.
- delay monitor DELAYMON or voltage monitor VMON
- the value of the internal register is shifted upward by 1 bit, the most significant bit is output to the OUT terminal to the outside, and the value input from the IN terminal to the least significant bit Is granted.
- the voltage monitor VMON is a block that measures the voltage supplied to the LSI_B. This is configured by combining, for example, a ring oscillator and a counter circuit.
- the voltage monitor VMON operates as follows. First, the maximum voltage value that can be measured by the voltage monitor VMON is recorded in advance in the shift register SHIFTREG connected to the voltage monitor VMON. The voltage monitor VMON during the measurement operation periodically measures the voltage value of the LSI. Further, the measured value is compared with the value (recorded value) recorded in the shift register SHIFTREG connected to the voltage monitor VMON. When the measured value falls below the recorded value, the measured value is shifted to the shift register SHIFTREG. To record. That is, when the measurement operation is completed, the minimum voltage value during the measurement operation period is recorded in the shift register SHIFTREG.
- the delay monitor DELAYMON is a block for determining whether a logical path (critical path) having the most severe timing constraint among the circuits mounted in the LSI_B operates correctly. This is constituted by, for example, a circuit configuration that imitates a critical path and a circuit that determines whether the operation result is correct or not.
- the delay monitor DELAYMON records the occurrence of the fraudulent delay in the connected shift register SHIFTREG when the fraud occurs in the calculation result of the critical path imitation circuit even once during the measurement operation period.
- the supply voltage value of the mounted LSI decreases and the LSI malfunctions, most of the causes are critical path malfunctions. That is, by determining whether the delay of the critical path imitation circuit is incorrect, it is possible to inspect a malfunction due to power failure of the LSI_B.
- FIG. 37 shows one mode of a power supply test procedure for the stacked LSI in FIG.
- This embodiment is a procedure for executing a power supply test at the time of starting up the stacked LSI.
- the test procedure is divided into three stages: (1) reading the power test program by PU_A0, (2) executing the power test program by PU_A0, and (3) processing the power test result.
- the processor units (PU_B0 to PU_B3, PU_C0 to PU_C3) and functional modules (IP_B0, IP_B1, IP_C0, IP_C1) mounted on the LSI_B and LSI_C ) Enters a standby state, and only the PU_A0 mounted on the LSI_A starts to operate.
- the standby state of LSI_B and LSI_C here refers to a state that does not hinder the operation of LSI_A and the operation of PTEST mounted in each of LSI_B and LSI_C. Specifically, a state where the actual operation is stopped, such as a clock supply stop or a power cut-off, is desirable. This is because these are not particularly dependent on the quality of the through silicon via for supplying power.
- the power supply test program PESTPROG is read from the external nonvolatile memory NVRAM via the local bus controller LBSC on the LSI_A.
- Reading here is not particularly limited to loading a program into the internal memory or main memory, but means that the address pointed to by the program counter of PU_A0 is set in PESTPROG.
- PU_A0 executes the power supply test program.
- PU_A0 controls PTEST_CTRL via the on-chip interconnect OCI_A, and starts the power supply test sequence shown in the following (2-1) and (2-2).
- PTEST_CTRL asserts the TSV_TREQ signal when the power supply test sequence starts.
- the TSV_TREQ signal propagates to the PTEST provided in the LSI_B and LSI_C stacked through the through silicon via.
- the processing units PU_B0 to PU_B3 or PU_C0 to PU_C3
- functional modules IP_B0 / IP_B1 or IP_C0 / IP_C1 mounted on the LSI_B and LSI_C are shifted to the power consumption mode, and at the same time, In the upper PTEST, measurement of power supply potential and malfunction inspection are started.
- This state is a power test, and the power test is continuously performed for a predetermined period determined by the power test program.
- PU_A0 reads the power test result.
- PU_A0 controls PTEST_CTRL and negates the TSV_TREQ signal.
- the processing units PU_B0 to PU_B3 or PU_C0 to PU_C3
- the functional modules IP_B0 / IP_B1 or IP_C0 / IP_C1 mounted on the LSI_B and LSI_C are restored from the power consumption mode to the original state (standby state).
- PU_A0 controls PTEST_CTRL to generate a periodic cycle signal in TSV_SCK.
- TSV_SCK is propagated to PTEST provided in LSI_B and LSI_C stacked via through silicon vias. Then, as shown in FIG. 36, the value recorded in the shift register SHIFTREG in each PTEST is output to the PTEST_CTRL of the LSI_A serially bit by bit through the TSV_TDI / TSV_TDO. That is, the PU_A0 continues the periodic cycle signal for a predetermined period, thereby transferring the lowest power supply potential and the incorrect lazy determination result of the stacked LSI_B and LSI_C inspected by the PTEST to the PTEST_CTRL in order from the lower LSI. It is possible to read.
- the number of through-silicon vias used for the power supply test can be reduced. Further, since the test results can be read out in the stacking order, the results of LSI_B can be taken out even when LSI_C is not operating due to a power failure of the through silicon via as shown in FIG. Further, in the serial bit string to be read, the location where the test result in each LSI is recorded is naturally determined by the order of stacking. Thereby, although details will be described later in (3), the process of associating the test result with each LSI can be performed by LSI_A alone.
- PU_A0 performs an analysis process on the test result read out to PTEST_CTRL.
- the power supply measurement value in each LSI and the critical path delay fraud determination result are stored in the stacking order.
- PU_A0 analyzes the case where the power supply measurement value is less than or equal to a predetermined threshold value and the case where the delay fraud occurs in each LSI, and if either one occurs, it is mounted on that LSI.
- the processing unit and function module to be used are determined to be unusable.
- the result of such analysis processing is recorded in the test result table RSLT_TBL in PTEST_CTRL.
- test result table RSLT_TBL usable processing units and functional modules mounted on the stacked LSI are recorded in the test result table RSLT_TBL.
- An example of the recording form of the test result table RSLT_TBL is shown in FIG.
- the test result table RSLT_TBL has a corresponding bit “0” when a usable processing unit and a functional module are usable, and a corresponding bit “0” when a usable unit is not usable. Is recorded.
- PU_A0 indicates a process of managing tasks of the entire stacked LSI and assigning tasks to usable processing units and functional modules while referring to the test result table RSLT_TBL.
- task here refers to the overall processing executed by the processing unit or functional module.
- LSIs with different configurations are used for LSI_A and LSI_B, but it is also possible to stack LSIs having the same configuration.
- PTEST_CTRL having an input pad is mounted in advance on each LSI.
- the PTEST_CTRL is configured to be in an inactive state when nothing is connected to the input pad, and to be in an active state only when an installation potential is applied.
- the PTEST_CTRL After the LSIs are stacked, only the lowermost LSI, that is, the LSI connected to the package substrate PKCB, activates the PTEST_CTRL by applying a ground potential to the input pad connected to the PTEST_CTRL by wire bonding or the like.
- the input pad connected to PTEST_CTRL is released and deactivated.
- FIG. 36 it is shown that one voltage monitor VMON and one delay monitor DELAYMON are mounted on PTEST, but a plurality of voltage monitors VMON and delay monitors DELAYMON are mounted on PTEST. Also good. In general, power supply fluctuations are different in each part of an LSI. Therefore, by mounting multiple voltage monitors VMON and delay monitors DELAYMON, and measuring the power supply potential and delay fraud in many places on the mounted LSI, the effects of through silicon vias for power supply can be more accurately affected. It becomes possible to test. As a matter of course, a plurality of shift registers SHIFTREG connected to the voltage monitor VMON and the delay monitor DELAYMON are also mounted. Also in this case, the power supply test can be performed without changing the test procedure shown in FIG. 37 by connecting the shift registers SHIFTREG in a chain shape.
- the PU_A0 on the LSI_A executes the power test program to perform the power test of the stacked LSI, but the present invention is not limited to this.
- the power supply test program can be executed by directly controlling PTEST_CTRL via TAP in FIG. 34 using an external debugger or an external tester. Thereby, it is not particularly necessary to connect the nonvolatile memory NVRAM to the stacked LSI. Therefore, it is possible to perform a power supply test at a timing such as after packaging in manufacturing, and the test cost can be reduced.
- Example 2-2 Next, a case where a defect due to a short circuit occurs in the through silicon via for supplying power in FIG. 33 will be described. Specifically, this corresponds to a failure in which electrical connection occurs between TSV_VDD or TSV_VSS in FIG.
- FIG. 39 is a schematic configuration diagram schematically showing an embodiment of a power supply configuration in the stacked LSI according to the second embodiment of the present invention.
- the portion described as “**” indicates all the symbols indicated by A0, B0, and the like.
- VDD_SW_A 0, VDD_SW_A 1, VSS_SW_A 0, VSS_SW_A 1, VDD_SW_B 0, VDD_SW_B 1, VSS_SW_B 0, VSS_SW_B 1 are each a conduction switch that controls conduction / shutoff between two terminals, and is composed of, for example, a combination of a MOS transistor switch and a control circuit.
- VDD_A is a wiring that supplies a power supply potential in LSI_A
- VSS_A is a wiring that supplies a ground potential in LSI_A
- VDD_STACK is a wiring for supplying a power supply potential to the LSI stacked on LSI_A and above LSI_A
- VSS_STACK is a power supply potential applied to the LSI stacked on LSI_A and above LSI_A. Wiring.
- VDD_A and VDD_STACK, VSS_A and VSS_STACK are not shown in the drawing, a power supply potential and a ground potential are respectively supplied from the outside via the package substrate PKCB.
- VDD_B0 is a wiring that applies a power supply potential to PU_B0 in LSI_B
- VSS_B0 is a wiring that supplies a ground potential to PU_B0.
- VDD _ ** is a wiring that supplies a power supply potential to PU _ **
- VSS _ ** is a wiring that supplies a ground potential to PU _ **.
- TSV_VDD_C0 is a wiring formed by through silicon vias and micro bumps that electrically connect VDD_C0 and VDD_SW_B0.
- VDD_SW_B0 is electrically connected to VDD_B0.
- TSV_VDD_C1 is electrically connected to VDD_C1 and VDD_SW_B1
- TSV_VSS_C0 is electrically connected to VSS_C0 and VSS_SW_B0
- TSV_VSS_C1 is electrically connected to VSS_C1 and VSS_SW_B1
- _SS is electrically connected to VDD_SW_B1 and VSS_B1 to VSS. Connected.
- TSV_VDD_B0 is a wiring formed by through silicon vias and micro bumps that electrically connect VDD_B0 and VDD_SW_A0. Further, VDD_SW_A0 is electrically connected to VDD_STACK. Similarly, TSV_VDD_B1 is electrically connected to VDD_B1 and VDD_SW_A1, TSV_VSS_B0 is electrically connected to VSS_B0 and VSS_SW_A0, TSV_VSS_B1 is electrically connected to VSS_B1 and VSS_SW_A1, and ___ is electrically connected to VDD_SW_A1 and VDD_STACK. Has been.
- TSV_VDD_B0 and TSV_VSS_B0 are electrically short-circuited due to defective formation of through silicon vias.
- the conduction switches VDD_SW _ **, VSS_SW _ **
- the power supply wiring is short-circuited in the LSI stacked above the LSI_A, and the power supply potential Does not rise to a predetermined value, or excessive current flows into the stacked LSI. Therefore, all of PU_B0, PU_B1, PU_C0, and PU_C1 cannot be operated or excessive power is consumed.
- VDD_SW_A0 and VSS_SW_A0 are shut off, even if TSV_VDD_B0 and TSV_VSS_B0 are short-circuited, there is no path for supplying power, so that a power supply short-circuit does not occur. Accordingly, since no potential is supplied to VDD_B0, VSS_B0, VDD_C0, and VSS_C0, PU_B0 and PU_C0 cannot be operated, but the other PU_A0, PU_B1, and PU_C1 can be operated.
- VDD_SW_B1 and VSS_SW_B1 are short-circuited.
- VDD_SW_C0 and VSS_SW_C0 are short-circuited.
- VSS_SW_C1 are cut off from each other, so that some of the processing units of the stacked LSI can be operated.
- the multilayer LSI according to the present embodiment is characterized in that, when a short circuit occurs in the through silicon via group for supplying power, it is remedied by turning it off.
- each conduction switch (VDD_SW _ **, VSS_SW _ **) is in a conduction state in the initial state, if there is a short circuit in the through silicon via for supplying power, the power supply is short-circuited when the stacked LSI is turned on. Will be activated. Therefore, it is desirable that the initial state of these conduction switches is a cut-off state. After the through-silicon via short-circuit test is completed, the conduction / cutoff state of each conduction switch is set. The procedure of this short circuit test will be described later. In addition, in FIG.
- TSV_VDD _ ** and TSV_VSS _ ** are each shown as one wiring, but in the case where this is formed by a plurality of through silicon vias and micro bumps, they are similarly short-circuited. Needless to say, the situation can be remedied.
- FIG. 40 shows an embodiment of a logical configuration and connection of functional blocks mounted on LSI_A and LSI_B in FIG. 33 in the stacked LSI according to the second embodiment of the present invention. It is a figure contrasted with FIG.
- LSI_B has four processing units PU_B0 to PU_B3, but PU_B2 and PU_B3 not shown in FIG. 39 have the same power supply path as PU_B0 and PU_B1 in FIG. Shall.
- FIG. 40 portions corresponding to those in FIG. 34 are denoted by the same reference numerals, and detailed description thereof will not be repeated. 40 differs from the configuration shown in FIG. 34 as follows.
- LSI_A includes an interrupt controller INTC that receives an interrupt signal from a regulator REGULATOR that supplies power to the stacked LSI from the outside.
- the regulator REGULATOR inputs an interrupt signal to the interrupt controller INTC when an excessive current exceeding a predetermined value flows into the stacked LSI.
- PU_A0 of LSI_A can detect that the power supply is short-circuited.
- LSI_B has four functional blocks that perform power test of PTEST_B0, PTEST_B1, PTEST_B2, and PTEST_B3 instead of PTEST and MD_CTRL. Yes.
- the present invention is not limited to this.
- the power supply test function block may be provided for each shared group.
- FIG. 41 shows a connection configuration among PU_B0, PU_B1, PTEST_B0, and PTEST_B1 in FIG.
- a difference between PTEST shown in FIG. 36 and PTEST_B0 and PTEST_B1 is that the TSV_TREQ_I signal is input to the shift register SHIFTFT in PTEST_B0, and the value input to the shift register SHIFTFTREG is input to TESTCTRL.
- the output of the shift register SHIFTREG of PTEST_B0 is input to the shift register of PTEST_B1.
- the shift register SHIFTREG for recording values of the voltage monitor VMON and the delay monitor DELAYMON is also configured in a chain shape. Further, although not shown in FIG.
- PTEST_B2 connected to PU_B2 and PTEST_B3 connected to PU_B3 have the same configuration, and the value of the shift register SHIFTREG in PTEST_B3 is set via TSV_TREQ_O by the periodic cycle signal of TSV_TREQ_SCK. Then, it is propagated to LSI_C stacked above or to LSI_A stacked below via TSV_TDO_B.
- TESTCTRL in PTEST_B0 switches assertion / negation of the power consumption mode transition request signal MAXPREQ_B0 to the connected processing unit PU_B0 according to the value of the connected shift register SHIFTREG. That is, when the value in the shift register SHIFTREG is “1”, the power consumption mode transition request signal MAXPREQ_B0 is asserted to the processing unit PU_B0. Thereby, PU_B0 transits to the power consumption mode described in the first embodiment. This configuration is the same in PTEST_B1.
- any combination of processors can be operated in the power consumption mode in the processing units PU_B0 to PU_B3 mounted on LSI_B. It becomes possible to make it.
- FIG. 42 shows an example of a procedure for performing a short-circuit test of a through silicon via for power supply in LSI_B in the present embodiment.
- the value in the shift register SHIFTREG in PTEST_B0 to PTEST_B3 is set to “0”, and the power supply switches VDD_SW _ ** and VSS_SW _ ** in FIG. 39 are all set to the cutoff state. For this reason, power is not supplied to PU_B0 to PU_B3 on LSI_B, and it starts in a stopped state.
- PU_A0 on LSI_A is activated in an operable state because power is supplied via VDD_A and VSS_A.
- PU_B0 is activated, and only this PU_B0 is in a state where a power test can be executed.
- “1” is input from TSV_TREQ_I for only one cycle.
- a signal “1” is written to SHIFTREG in PTEST_B0, and a continuity test for PU_B0 is started.
- PTEST_B0 asserts the power consumption mode transition request signal MAXPREQ_B0 to PU_B0, and PU_B0 operates in the power consumption mode.
- FIG. 42 shows a state in which an interrupt signal from the regulator REGULATOR is input during the continuity test period of PU_B1.
- an interrupt signal from the regulator REGULATOR is input only when an excessive current flows into the stacked LSI. That is, it is possible to determine that a short circuit has occurred on the through silicon via that supplies power to PU_B1 and the wiring connected thereto.
- the PU_A0 records that the PU_B1 is unusable in the RSLT_TBL in the PTEST_CTRL, and as described in FIG. 39, the through silicon via that supplies power to the PU_B1
- the conduction switches VDD_SW_A1 and VSS_SW_A0 connected to are cut off.
- the semiconductor device provides a short circuit test for the power supply through silicon via and a relief means when a short circuit occurs.
- CPU1 to CPU8 ... Processor, RAM1 / RAM2 ... Memory, PKCB ... Package substrate, LSI_A, LSI_B, LSI_C ... LSI for stacking, BMP ... Solder bump, MBMP ... Micro bump, TSV_VDD, TSV_VSS ... Through silicon via for power supply, TSV_SIG: Through silicon via for signal, TSV_SCK, TSV_TREQ, TSV_TDI_A, TSV_TDI_B, TSV_TDI_C, TSV_TDO_A, TSV_TDO_B, TSV_TDO_C ... Through silicon via for power test signal, PTEST ...
- Power test circuit PTEST_CTRL: power supply test control circuit, PU_A0, PU_B0, PU_B1, PU_B2, PU_B3 ... processing unit, P_B0, IP_B1 ... functional block, OCI_A, OCI_B ... on-chip interconnect, 3DI ... communication interface between stacked LSIs, TAP ... test access port, LBSC ... local bus controller, NVMEM ... Non-volatile memory, PESTPROG ... Power supply test program, RSLT_TBL ... Test result table, TREQ ... Test request signal, MAXPREQ ... Power consumption mode transition request signal, MD_CTRL ... Operation mode control unit, MDREG ...
- Operation mode register ROM ... Read only memory, MAXPROG ... Maximum power consumption program, SEL ... Selector, FF ... Flip flop, CLK ... Clock signal, TESTCTRL ... Source test control block, DELAYMON ... delay monitor, VMON ... voltage monitor, SHIFTFT ...
- VDD_SW_B2 VSS_SW_A0 , VSS_SW_A1, VSS_SW_B0, VSS_SW_B1... Power switch, PTEST_B0, PTEST_B1, PTEST_B2, PTEST_B3.
Landscapes
- Engineering & Computer Science (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Power Engineering (AREA)
- Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Condensed Matter Physics & Semiconductors (AREA)
- General Physics & Mathematics (AREA)
- Manufacturing & Machinery (AREA)
- Geometry (AREA)
- Semiconductor Integrated Circuits (AREA)
- Design And Manufacture Of Integrated Circuits (AREA)
- Tests Of Electronic Circuits (AREA)
Abstract
Description
図1に本発明を用いた3次元に積層するLSIのシリコン貫通ビア(TSV)の配置概略図を示す。図1において、黒丸は電源を積層したチップに供給するための電源用TSV、白丸は積層したチップ間で信号を伝送する信号線用TSVである。チップの外周部に電源用のTSVが複数列配置され、その内側に信号線用のTSVが配置されている。このTSVの配置構成を用いて、CPU(Central Processing Unit)が複数搭載されているCPUチップとメモリアレイが搭載されたメモリチップを2枚ずつ積層した場合の断面構造を図2に示す。図2は、図1のチップをA-A’で切断した場合の断面である。それぞれのチップ間を接続する電源用TSVと信号線用TSVは上下方向に同じ場所に配置されている。ぞれぞれのチップはメタル配線層が下向きに積層され、TSVは上に積層されているチップの金属配線層に接続され、そこから下に積層されているチップのシリコン層を貫通して下に積層されているチップの金属配線層に接続される。この様子を図7に示す。ここで、一番下のチップは、パッケージ基板と接続する。ここで、一番下のチップはパッケージ基板との間の信号線の距離を短くするためメタル配線層を下向きとし、また、それに合わせてすべてのチップを下向きに積層している。
図9に、本発明を用いて、異なる大きさのチップを3次元に積層する場合のTSVの配置概略図を示す。チップAは例えばCPUが複数搭載されたCPUチップ、チップBは例えばメモリが搭載されたメモリチップである。メモリチップ等は比較的消費電力が低いため、電源用のTSVの数は少なくすることが可能である。よって、電源用のTSVはチップAの方がチップBよりも多くなっている。また、チップAとチップBでは、別チップとの通信に必要な信号線用のTSV数は異なっている。図10は、図9のチップをB-B’で切断した場合の断面である。チップの左および上半面では、それぞれのチップ間を接続する電源用TSVと信号線用TSVは上下方向に同じ場所に配置されており、さらに右および下半面のTSVはチップBの接続に関しては存在していない。このように、最外周に電源用TSV、その内側に信号用TSVを配置しているため、チップの2辺の位置をそろえて配置し、電源用および信号線用TSVの位置を揃えることで、異なる大きさのチップを積層することが可能となる。また、同様の考え方で、1辺を揃えれば、さらに規模の小さいチップを積層することも可能となる。
図11に、本発明を用いて、異なる大きさのチップを3次元に積層する場合のTSVの配置概略図を示す。チップAは例えばCPUが複数搭載されたCPUチップ、チップCは例えばメモリが搭載されたメモリチップである。メモリチップ等は比較的消費電力が低いため、電源用のTSVの数は少なくすることが可能であり、電源用のTSVはチップAの方がチップBよりも多くなっている。また、本実施例では、実施例1-2とは異なり、信号用TSVの数を等しくできる。これは、電源用TSVを最外周に複数列配置しているため、大きなチップの電源用TSVの一部を小さいチップに接続しないことで、信号用TSVの位置を併せることが可能となっている。よって、この場合には、チップAとチップCで信号線用のTSVの位置を揃える必要がある。
図13に、本発明を用いて、異なる大きさのチップを3次元に積層する場合のTSVの配置概略図を示す。チップAは例えばCPUが複数搭載されたCPUチップ、チップDは例えばメモリが搭載されたメモリチップである。チップAおよびチップDともに同様に外周部分に電源用のTSVおよび信号線用のTSVを配置する必要があるチップを積層する必要がある場合には、チップAとチップD間で電源配線および信号配線を接続するためにインタフェースチップを挿入する必要がある。
図15(a)には、本発明を用いたLSIの電源用TSVの割り当て図を示す。チップ最外周のTSVにVDDを割り当て、その内側のTSVにVSSが割り当てられている以外は、実施例1-1と同等である。このように、同じ役割を持つ電源を隣り合って配置することにより、電源配線とTSVを接続するパターンを容易に作ることが可能となる。即ち、図15(b)に示すように、各電源用TSVの間にVDD若しくはVSSの電源配線のいずれか一方を配線すればよく、図8と比較して、電源線の本数を少なくすることができる。よって、1本あたりの電源線を太くすることが可能となり、抵抗値を下げることが可能となる。
図16(a)には、本発明を用いたLSIの電源用TSVの割り当て図を示す。この例では、チップ外周の4列に電源線用のTSVが割り当てられている例を示している。このチップでは、近傍の4つのTSVに同じ種類の電源、つまりVDDおよびVSSが割り当てられ、それらが千鳥格子状に配置されている。また、それ以外の構成に関しては、実施例1-1と同様である。この構成をとった場合には、電源用のTSVの数をVDDとVSSで揃えることが可能になるとともに、例えば一つのTSVが製造時に破壊されるなどのなんらかの理由で動作しない状態となった場合にも周りに同種類の電源用のTSVが配置されているため、電流の集中が起きない。つまり、この構成をとることで、製造時の破壊等に強い電源用TSV配置が可能となる。更に本実施例では、実施例1-5と同様に、各電源用TSVの間にVDD若しくはVSSのいずれか一方の電源線を配線すればよく(図16(b))、1本の電源線を太くすることが可能となる。
図17には、本発明を用いたLSIの電源用TSVの割り当て図を示す。この例では、チップ外周の4列に電源線用のTSVが割り当てられている例を示している。このチップでは、横方向に並ぶ同一の行のTSVには同じ種類の電源、つまりVDDおよびVSSが割り当てられている。また、それ以外の構成に関しては、実施例1-1と同様である。この構成をとった場合には、電源用のTSVの数をVDDとVSSで揃えることが可能となるとともに、TSVが接続される電源配線層をTSVに沿ってチップの横方向に並べることが可能となる。TSVに接続される電源配線層の構成を図18に示す。図に示すとおり、同じ種類の電源、VDDとVSSが同一行状に並ぶため、電源用の金属配線をストライプ状に配置することが容易となる。特に、今までの実施例では、電源用TSVの間に電源線を配線していたが、本実施例では、同種の電源が左右方向に並ぶため、電源用TSVの間に配線するのではなく、電源用TSV上に、配線することが可能となり、更に電源線を太くすることが可能となる。言い換えれば、上述の実施例では、電源配線と電源用TSVとを接続するために、電源線の一部を太くした凸型形状とする必要があるが、本実施例では、電源用TSVのすべてを包含する(電源用TSVより太い)配線とすることが可能となり、凸型形状を形成する必要がない。また、この配線層は、TSVの上面に接続するチップ上層の配線層であるとともに、シリコンを貫通したTSVが接続するチップの下層の配線層の様子を示している。実際には、チップの下層の配線層は、TSVが突き抜ける形なるため、他の実施例においては、上述の凸型形状を設けることにより、下層においても電源用TSVとの接続が可能となる。しかしながら、凸型形状とTSVの位置を合わせる必要があるため、位置ずれ等を考慮する必要がある。一方、本実施例では、電源線をTSVより太くしているため、凸型形状が不要で、配線のどの部分と接触してもよいことになり、下層の配線層での電源線と電源用TSVの接続が容易となる。
図19(a)には、本発明を用いたLSIの電源用TSVの割り当て図を示す。この例では、チップ外周の4列に電源線用のTSVが割り当てられている例を示している。また、このチップでは、高い電源電圧VDDHと低い電源電圧VDDLおよび接地電位電源VSSを持つ構成を示している。このチップでは、近傍の4つのTSVに同じ種類の電源、つまりVDDH、VDDLおよびVSSが割り当てられ、それらが千鳥格子状に配置されている。また、TSVの数は、VDDH用のTSVの個数とVDDL用のTSVの個数を加えた数がVSSと等しくなっている。これは、VDDHおよびVDDLから供給された電流はVSSを経由して流れる必要があるため、電流容量によってTSVの個数が決められているためである。この構成をとった場合には、電源用のTSVの数をVDDHとVDDLを加えた数とVSSの数で揃えることが可能となるとともに、例えば一つのTSVが製造時に破壊されるなどのなんらかの理由で動作しない状態となった場合にもまわりに同種類の電源用のTSVが配置されているため、電流の集中が起きない。つまり、この構成をとることで、製造時の破壊等に強い電源用TSV配置が可能となる。また、この実施例では、電源の種類が3種類の場合の例を示しているが、同様の考え方で4種類以上の電源を持つ場合にも対応可能である。さらに、本実施例では4つのTSVをまとめて千鳥格子状に並べる例を示したが、2x3の6つや3x3の9つのTSVごとにまとめて千鳥格子状に並べる構成も可能となる。
図20(a)には、本発明を用いたLSIの電源用TSVの割り当て図を示す。この例では、チップ外周の4列に電源線用のTSVが割り当てられている例を示している。また、このチップでは、高い電源電圧VDDHと低い電源電圧VDDLおよび接地電位電源VSSを持つ構成を示している。このチップでは、近傍の4つのTSVにVDD側およびVSS側の電源が割り当てられ、それらが千鳥格子状に配置されている。また、TSVの数は、VDDH用のTSVの個数とVDDL用のTSVの個数を加えた数がVSSと等しくなっている。これは、VDDHおよびVDDLから供給された電流はVSSを経由して流れる必要があるため、電流容量によってTSVの個数が決められているためである。さらに、VDDHはVDDLよりも高い電位を供給しているため多くの電流を供給する必要があると考えられるため、電流容量を揃えるために、VDDHにVDDLよりも多くのTSVが割り当てられている。また、図20(b)に示されるように、TSVに隣接する最上層の電源配線はVDDHおよびVSSに割り当てられ、それぞれのTSVから配線で接続される。また、最上層の1層下のTSVに隣接する電源配線はVDDH、VDDLおよびVSSに割り当てられ、それぞれのTSVから配線で接続される。この構造を持つことによって、3種類以上ある電源用のすべてのTSVから各電源配線に接続することが可能となる。本図では、TSVに隣接する電源配線のみ図示したが、実際にはTSV間には十分のスペースがあるため、追加の配線層が配置されている。たとえば、最上層にはVDDHおよびVSSの配線層のみ図示しているが、その間にVDDLの配線や、VDDHおよびVSSの余分な配線を配置することで、電源配線の抵抗を下げることは可能である。また、本実施例では、電源の種類が3種類の場合の例を示しているが、同様の考え方で4種類以上の電源を持つ場合にも対応可能である。さらに、本実施例では4つのTSVをまとめて千鳥格子状に並べる例を示したが、2x3の6つや3x3の9つのTSVごとにまとめて千鳥格子状に並べる構成も可能となる。その場合には、さらに細かくVDDHとVDDLの比率を設定することも可能となる。
図21に、本発明を用いて、異なる大きさのチップを3次元に積層する場合のTSVの配置概略図を示す。チップAは例えばCPUが複数搭載されたCPUチップ、チップEは例えばメモリが搭載されたメモリチップである。基本的には、実施例1-2と同じ考え方に基づくが、この例では、チップEの大きさがチップAの約1/4となっている場合の例を示す。この場合、チップAの外周部に電源用のTSVを配置し、その上部の同じ位置にチップEのTSVを配置して積層することによって、異なる面積のチップを積層することが可能となる。
図23に、本発明を用いて、複数の種類のチップを積層する場合のTSVの配置概略図を示す。この構成では、実施例1同様電源用TSVをチップ外周部に配置するとともに、信号線用TSVは実施例1と異なり、チップ中央部に配置している。
図27に、本発明を用いて、複数の種類のチップを積層する場合のTSVの配置概略図を示す。この構成では、実施例1-1と異なり、チップの左右の外周部に電源用のTSVを複数列配置し、その内側に信号線用のTSVを複数列配置している。積層されたチップはすべて同様のTSV配置の構成をしている。この構成では、実施例1-1と異なり、チップの上下側の外周部にはTSVが配置されていない。内部の回路の形状が縦に長く、外周部に近い場所まで内部回路を置く必要がある場合には、本実施例の構成をとることで、チップ面積を縮小することが可能となる。
図29に、本発明を用いて、複数の種類のチップを積層する場合のTSVとチップ間の通信に用いられるコイルの配置概略図を示す。実施例1-1~1-12では、チップを3次元に積層した場合に、チップ間の信号通信にTSVを用いる例を示した。本実施例では、チップ間の通信に、コイルによる誘導結合通信を用いる場合の構成例を示す。この構成では、チップの外周部に電源用のTSVを複数列配置している。さらに、チップの中央部にはそれぞれのチップの動作回路、たとえばCPUやメモリが配置されるが、その上層の配線を用いて、チップ貫通信用のコイルが形成される。
図31に、本発明を用いて、異なる大きさのチップを3次元に積層する場合のTSVの配置概略図を示す。チップJは例えばCPUが複数搭載されたCPUチップ、チップAは例えばメモリが搭載されたメモリチップである。メモリチップ等の比較的消費電力が低いチップに関しては、電源用のTSVの数は少なくすることが可能であり、逆にCPUチップ等の電力が大きいチップに関しては、電源用のTSVの数が多く必要となる。また、チップAとチップJで必要な信号線用TSVの数は等しい。
図33は、本発明の実施例2に係る半導体装置の積層断面図と不良の発生の様相を示したものである。
次に、図33における電源供給用のシリコン貫通ビアで短絡による不良が発生している場合について言及する。具体的には、図33においてTSV_VDDもしくはTSV_VSSの間で電気的な接続が発生してしまう不良に相当する。
Claims (20)
- 第1の方向に延びる第1及び第2の辺と、前記第1の方向と交差する第2の方向に延びる第3及び第4の辺とを有する第1半導体基板上に形成された第1回路ブロックと、
前記第1半導体基板と接続され、前記第1回路ブロックから出力される信号を他の第2半導体基板に形成された第2回路ブロックに伝達する複数の信号線用貫通ビアと、
前記第1回路ブロックに電源を供給するための複数の電源用貫通ビアとを具備し、
前記複数の電源用貫通ビアは、前記第3の辺及び第4の辺に沿って、前記第1半導体基板の端に形成され、かつ、前記第1の方向に複数列形成されることを特徴とする半導体装置。 - 請求項1において、
前記複数の信号線用貫通ビアは、前記第1回路ブロックと前記複数の電源用貫通ビアの間に形成されることを特徴とする半導体装置。 - 請求項1において、
前記複数の電源用貫通ビアは、更に、前記第1の辺及び前記第2の辺に沿って、前記第2の方向に複数列形成されることを特徴とする半導体装置。 - 請求項1において、
前記複数の電源用貫通ビアは、前記第1半導体基板の前記第3の辺に沿った第1の領域と前記第4の辺に沿った第2の領域に形成され、前記第1回路ブロック及び前記複数の信号線用貫通ビアは、前記第1の領域と前記第2の領域との間の第3の領域に形成され、
前記第1の領域に形成された前記複数の電源用貫通ビアを接続する第1電源線は、前記第1の方向に延在すると共に、前記第2の方向に並んだ前記複数の電源用貫通ビアの間のすべてに形成され、
前記第2の領域に形成された前記複数の電源用貫通ビアを接続する第2電源線は、前記第1の方向に延在すると共に、前記第2の方向に並んだ前記複数の電源用貫通ビアの間のすべてに形成され、
前記第1の領域に形成された前記複数の電源用貫通ビアと前記第2の領域に形成された前記複数の電源用貫通ビアとを接続する第3電源線の数は、前記第1の電源線及び前記第2の電源線の数より少ないことを特徴とする半導体装置。 - 請求項1において、
前記複数の電源用貫通ビアは、第1の電圧を供給する複数の第1電源用貫通ビアと、前記第1の電圧とは異なる電圧を供給する複数の第2電源用貫通ビアとを含み、
前記第2の方向に一列に並んで形成された前記複数の電源用貫通ビアは、前記第1電源用貫通ビアと、前記第2電源用貫通ビアとが交互に設けられ、
前記第1の方向に一列に並んで形成された複数の電源用貫通ビアは、前記第1電源用貫通ビアと前記第2電源用貫通ビアとが交互に設けられることを特徴とする半導体装置。 - 請求項1において、
前記複数の電源用貫通ビアは、第1の電圧を供給する複数の第1電源用貫通ビアと、前記第1の電圧とは異なる電圧を供給する複数の第2電源用貫通ビアとを含み、
前記複数の第1電源用貫通ビアは、前記第2の方向に一列に形成され、
前記複数の第2電源用貫通ビアは、前記複数の第1電源用貫通ビアと前記複数の信号線用貫通ビアの間に、前記第2の方向に一列に形成されることを特徴とする半導体装置。 - 請求項1において、
前記複数の電源用貫通ビアは、第1の電圧を供給する複数の第1電源用貫通ビアと、前記第1の電圧とは異なる電圧を供給する複数の第2電源用貫通ビアとを含み、
前記第2の方向に一列に並んで形成された前記複数の電源用貫通ビアは、前記第1電源用貫通ビアと、前記第2電源用貫通ビアとが所定数毎に交互に設けられ、
前記第1の方向に一列に並んで形成された複数の電源用貫通ビアは、前記第1電源用貫通ビアと前記第2電源用貫通ビアとが前記所定数毎に交互に設けられることを特徴とする半導体装置。 - 請求項1において、
前記複数の電源用貫通ビアは、第1の電圧を供給する複数の第1電源用貫通ビアと、前記第1の電圧とは異なる電圧を供給する複数の第2電源用貫通ビアとを含むと共に、前記第1半導体基板の前記第3の辺に沿った第1の領域と前記第4の辺に沿った第2の領域に形成され、
前記第1及び第2の領域に形成された複数の第1電源用貫通ビアは、前記第1の方向に一列に配置され、
前記第1及び第2の領域に形成された複数の第2電源用貫通ビアは、前記第1の方向に一列に配置され、
前記第1及び第2の領域に形成された複数の電源用貫通ビアを接続する電源線は、前記第1の方向に延在することを特徴とする半導体装置。 - 請求項1において、
前記複数の電源用貫通ビアは、第1の電圧を供給する複数の第1電源用貫通ビアと、前記第1の電圧とは異なる電圧を供給する複数の第2電源用貫通ビアと、前記第1及び第2の電圧とは異なる第3の電圧を供給する複数の第3電源用貫通ビアとを含み、
前記複数の第1電源用貫通ビアは、前記複数の第2及び第3電源用貫通ビアから前記第1回路ブロックを介して電流が流れ、
前記複数の第1電源用貫通ビアの数は、前記複数の第2電源用貫通ビアの数、及び、前記複数の第3電源用貫通ビアの数の夫々より多いことを特徴とする半導体装置。 - 請求項9において、
前記複数の第1電源用貫通ビアの数は、前記複数の第2電源用貫通ビア及び前記複数の第3電源用貫通ビアの数の合計と同数であることを特徴とする半導体装置。 - 請求項1において、
前記第1回路ブロックは、プロセッシング・ユニットを有し、
前記プロセッシング・ユニットは、通常動作を行うノーマルモードにおける電力消費より大きい電力を消費する電力消費モードを有することを特徴とする半導体装置。 - 請求項11において、
前記第1回路ブロックは、機能モジュールを更に有し、
前記機能モジュールは、スキャン・チェーン部と、前記スキャン・チェーン部に接続された乱数発生部と、を有し、
前記プロセッシング・ユニットが前記電力消費モードとなった場合に、前記乱数発生部から前記スキャン・チェーン部に乱数を送出することを特徴とする半導体装置。 - 請求項11において、
前記第1回路ブロックは、前記電力消費モードにおいて、前記複数の電源用貫通ビアより供給された電圧を計測する電圧モニタを有することを特徴とする半導体装置。 - 請求項13において、
前記第1回路ブロックは、前記電力消費モードにおいて、クリティカル・パスのディレイを計測するディレイ・モニタを有することを特徴とする半導体装置。 - ノーマルモードと電力消費モードを有する第1プロセッシング・ユニットと、第1電源テスト回路を有する第1LSIと、
前記第1LSIと積層され、前記ノーマルモードと前記電力消費モードを有する第2プロセッシング・ユニットと第2電源テスト回路とを有する第2LSIと、
前記第1LSIと前記第2LSIとを接続し、前記第1LSI及び前記第2LSIに動作電圧を供給する電源用貫通ビアとを具備し、
前記電力消費モードにおける前記第1プロセッシング・ユニットの消費電力は、前記ノーマルモードにおける前記第1プロセッシング・ユニットの消費電力より大きく、
前記電力消費モードにおける前記第2プロセッシング・ユニットの消費電力は、前記ノーマルモードにおける前記第2プロセッシング・ユニットの消費電力より大きく、
前記第1電源テスト回路は、前記第1プロセッシング・ユニットが前記電力消費モードとなっている場合に、前記電源用貫通ビアを介して供給される前記動作電圧の電圧値を測定し、
前記第2電源テスト回路は、前記第2プロセッシング・ユニットが前記電力消費モードとなっている場合に、前記電源用貫通ビアを介して供給される前記動作電圧の電圧値を測定することを特徴とする半導体装置。 - 請求項15において、
前記第1プロセッシング・ユニットが前記電力消費モードで動作する期間は、前記第2プロセッシング・ユニットが前記電力消費モードで動作する期間と並行し、
前記第1電源テスト回路が前記動作電圧の電圧値を測定する期間及び前記第2電源テスト回路が前記動作電圧の電圧値を測定する期間は、前記第1プロセッシング・ユニットと前記第2プロセッシング・ユニットが前記電力消費モードにて動作している期間と並行していることを特徴とする半導体装置。 - 請求項15において
前記第1LSIは、前記電力消費モードにおいて、前記第1プロセッシング・ユニットが実行する最大電力消費プログラムを格納する第1メモリを更に具備し、
前記第2LSIは、前記電力消費モードにおいて、前記第2プロセッシング・ユニットが実行する最大電力消費プログラムを格納する第2メモリを更に具備することを特徴とする半導体装置。 - 請求項15において、
前記第1LSIは、動作モードが設定される第1レジスタを有し、
前記第2LSIは、動作モードが設定される第2レジスタを有し、
前記半導体装置は、外部から前記第1レジスタに動作モードを設定するためにデータを転送する第1テスト用貫通ビアと、前記第1LSIから前記第2レジスタに動作モードを設定するためにデータを転送する第2テスト用貫通ビアを更に具備することを特徴とする半導体装置。 - 請求項15において、
前記第1電源テスト回路は、前記第1プロセッシング・ユニットのクリティカル・パスを模擬した第1ディレイ・モニタを有し、
前記第2電源テスト回路は、前記第2プロセッシング・ユニットのクリティカル・パスを模擬した第2ディレイ・モニタを有し、
前記第1及び第2ディレイ・モニタは、前記電力消費モードにおいて、前記クリティカル・パスを正しく動作しているか否かを判定することを特徴とする半導体装置。 - 請求項15において、
前記第1LSIは、第1スキャン・チェーンと前記第1スキャン・チェーンに接続される第1乱数発生部とを有する第1機能モジュールを更に有し、
前記第2LSIは、第2スキャン・チェーンと前記第2スキャン・チェーンに接続される第2乱数発生部とを有する第2機能モジュールを更に有し、
前記第1及び第2乱数発生部は、前記電力消費モードにおいて、前記第1及び第2スキャン・チェーンに乱数を送出することを特徴とする半導体装置。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2011530722A JP5420671B2 (ja) | 2009-09-14 | 2009-09-14 | 半導体装置 |
PCT/JP2009/066040 WO2011030467A1 (ja) | 2009-09-14 | 2009-09-14 | 半導体装置 |
US13/388,990 US8653645B2 (en) | 2009-09-14 | 2009-09-14 | Semiconductor device comprising stacked LSI having circuit blocks connected by power supply and signal line through vias |
US14/095,859 US9318397B2 (en) | 2009-09-14 | 2013-12-03 | Stacked semiconductor chips including test circuitry |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2009/066040 WO2011030467A1 (ja) | 2009-09-14 | 2009-09-14 | 半導体装置 |
Related Child Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US13/388,990 A-371-Of-International US8653645B2 (en) | 2009-09-14 | 2009-09-14 | Semiconductor device comprising stacked LSI having circuit blocks connected by power supply and signal line through vias |
US14/095,859 Division US9318397B2 (en) | 2009-09-14 | 2013-12-03 | Stacked semiconductor chips including test circuitry |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2011030467A1 true WO2011030467A1 (ja) | 2011-03-17 |
Family
ID=43732152
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2009/066040 WO2011030467A1 (ja) | 2009-09-14 | 2009-09-14 | 半導体装置 |
Country Status (3)
Country | Link |
---|---|
US (2) | US8653645B2 (ja) |
JP (1) | JP5420671B2 (ja) |
WO (1) | WO2011030467A1 (ja) |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012156186A (ja) * | 2011-01-24 | 2012-08-16 | Keio Gijuku | 積層型半導体集積回路装置 |
JP2013084697A (ja) * | 2011-10-07 | 2013-05-09 | Hitachi Ltd | 半導体装置 |
CN103650136A (zh) * | 2012-05-10 | 2014-03-19 | 松下电器产业株式会社 | 具有电源电压的稳定化结构的三维集成电路及其制造方法 |
JP2014072289A (ja) * | 2012-09-28 | 2014-04-21 | Renesas Electronics Corp | 半導体装置 |
JP2014072499A (ja) * | 2012-10-02 | 2014-04-21 | Hitachi Ltd | 半導体装置 |
JP2014220473A (ja) * | 2013-05-10 | 2014-11-20 | キヤノン株式会社 | 集積回路装置およびその構成方法 |
JP2016025103A (ja) * | 2014-07-16 | 2016-02-08 | 富士通株式会社 | 電子部品、電子装置及び電子装置の製造方法 |
JP2018160490A (ja) * | 2017-03-22 | 2018-10-11 | 富士通株式会社 | 電子装置及びその製造方法、電子部品 |
US10103124B2 (en) | 2015-08-17 | 2018-10-16 | Fujitsu Limited | Semiconductor device |
JP2019079862A (ja) * | 2017-10-20 | 2019-05-23 | 電子商取引安全技術研究組合 | 半導体装置 |
WO2019146724A1 (ja) * | 2018-01-25 | 2019-08-01 | ソフトバンク株式会社 | 中央演算装置、gpgpu、メモリーを含む三次元積層集積回路の冷媒による冷却方式と、それを用いた三次元積層集積回路と、給電方式 |
US10734360B2 (en) | 2018-03-22 | 2020-08-04 | Toshiba Memory Corporation | Semiconductor device |
JP2021090010A (ja) * | 2019-12-05 | 2021-06-10 | ソフトバンク株式会社 | 半導体パッケージならびに穴開きインターポーザーによる液浸冷却方式を用いた三次元積層集積回路 |
WO2021171785A1 (ja) * | 2020-02-28 | 2021-09-02 | ソニーセミコンダクタソリューションズ株式会社 | 半導体装置、および、テストシステム |
JP2021141240A (ja) * | 2020-03-06 | 2021-09-16 | 本田技研工業株式会社 | 半導体装置とその製造方法 |
CN115842013A (zh) * | 2023-02-13 | 2023-03-24 | 浙江力积存储科技有限公司 | 一种三维堆叠存储器及其数据处理方法 |
WO2024127812A1 (ja) * | 2022-12-16 | 2024-06-20 | 先端システム技術研究組合 | 半導体モジュール |
US12140624B2 (en) | 2020-02-28 | 2024-11-12 | Sony Semiconductor Solutions Corporation | Semiconductor device and test system |
Families Citing this family (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012243910A (ja) * | 2011-05-18 | 2012-12-10 | Elpida Memory Inc | 半導体チップのクラックのチェックテスト構造を有する半導体装置 |
WO2013101243A1 (en) * | 2011-12-31 | 2013-07-04 | Intel Corporation | High density package interconnects |
JP2013258271A (ja) * | 2012-06-12 | 2013-12-26 | Ps4 Luxco S A R L | 半導体装置 |
KR101960496B1 (ko) * | 2012-08-29 | 2019-03-20 | 에스케이하이닉스 주식회사 | 반도체 장치 |
KR102032887B1 (ko) * | 2012-12-10 | 2019-10-16 | 삼성전자 주식회사 | 반도체 패키지 및 반도체 패키지의 라우팅 방법 |
US9482720B2 (en) * | 2013-02-14 | 2016-11-01 | Duke University | Non-invasive pre-bond TSV test using ring oscillators and multiple voltage levels |
US9030025B2 (en) * | 2013-03-15 | 2015-05-12 | IPEnval Consultant Inc. | Integrated circuit layout |
KR20140148179A (ko) * | 2013-06-21 | 2014-12-31 | 삼성전자주식회사 | 반도체 패키지 장치 및 이의 제조 방법 |
US9748002B2 (en) | 2013-10-23 | 2017-08-29 | Etron Technology, Inc. | System-in-package module with memory |
US9726691B2 (en) * | 2014-01-07 | 2017-08-08 | International Business Machines Corporation | 3D chip testing through micro-C4 interface |
US8929169B1 (en) * | 2014-05-13 | 2015-01-06 | Sandisk Technologies Inc. | Power management for nonvolatile memory array |
US10147658B2 (en) * | 2014-06-09 | 2018-12-04 | SK Hynix Inc. | Stacked semiconductor apparatus being electrically connected through through-via and monitoring method |
JP6235423B2 (ja) * | 2014-06-30 | 2017-11-22 | 東芝メモリ株式会社 | 半導体装置 |
JP6495692B2 (ja) | 2015-03-11 | 2019-04-03 | 東芝メモリ株式会社 | 半導体装置及びその製造方法 |
US9439330B1 (en) * | 2015-03-29 | 2016-09-06 | Banqiu Wu | 3D IC computer system |
CN107305861B (zh) * | 2016-04-25 | 2019-09-03 | 晟碟信息科技(上海)有限公司 | 半导体装置及其制造方法 |
US10170448B2 (en) * | 2016-12-07 | 2019-01-01 | Micron Technology, Inc. | Apparatus and method of power transmission sensing for stacked devices |
JP2019046951A (ja) * | 2017-09-01 | 2019-03-22 | ルネサスエレクトロニクス株式会社 | 半導体装置、半導体システム、及び、半導体装置の製造方法 |
FR3077925B1 (fr) * | 2018-02-14 | 2021-06-18 | Commissariat Energie Atomique | Circuit integre tridimensionnel face a face de structure simplifiee |
TWI750467B (zh) | 2018-05-15 | 2021-12-21 | 南韓商三星電子股份有限公司 | 半導體封裝 |
US11107771B2 (en) * | 2019-12-26 | 2021-08-31 | Taiwan Semiconductor Manufacturing Company, Ltd. | Segregated power and ground design for yield improvement |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2007150208A (ja) * | 2005-11-30 | 2007-06-14 | System Fabrication Technologies Inc | 半導体装置 |
JP2008004853A (ja) * | 2006-06-26 | 2008-01-10 | Hitachi Ltd | 積層半導体装置およびモジュール |
JP2008096312A (ja) * | 2006-10-12 | 2008-04-24 | Elpida Memory Inc | 積層型半導体装置及びそのテスト方法 |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6566848B2 (en) * | 2000-12-26 | 2003-05-20 | Intel Corporation | Auto-calibrating voltage regulator with dynamic set-point capability |
JP4272968B2 (ja) * | 2003-10-16 | 2009-06-03 | エルピーダメモリ株式会社 | 半導体装置および半導体チップ制御方法 |
US8274792B2 (en) * | 2005-09-06 | 2012-09-25 | Beyond Blades Ltd. | 3-dimensional multi-layered modular computer architecture |
US8106520B2 (en) * | 2008-09-11 | 2012-01-31 | Micron Technology, Inc. | Signal delivery in stacked device |
US7835207B2 (en) * | 2008-10-07 | 2010-11-16 | Micron Technology, Inc. | Stacked device remapping and repair |
-
2009
- 2009-09-14 WO PCT/JP2009/066040 patent/WO2011030467A1/ja active Application Filing
- 2009-09-14 US US13/388,990 patent/US8653645B2/en not_active Expired - Fee Related
- 2009-09-14 JP JP2011530722A patent/JP5420671B2/ja not_active Expired - Fee Related
-
2013
- 2013-12-03 US US14/095,859 patent/US9318397B2/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2007150208A (ja) * | 2005-11-30 | 2007-06-14 | System Fabrication Technologies Inc | 半導体装置 |
JP2008004853A (ja) * | 2006-06-26 | 2008-01-10 | Hitachi Ltd | 積層半導体装置およびモジュール |
JP2008096312A (ja) * | 2006-10-12 | 2008-04-24 | Elpida Memory Inc | 積層型半導体装置及びそのテスト方法 |
Cited By (33)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012156186A (ja) * | 2011-01-24 | 2012-08-16 | Keio Gijuku | 積層型半導体集積回路装置 |
JP2013084697A (ja) * | 2011-10-07 | 2013-05-09 | Hitachi Ltd | 半導体装置 |
CN103650136A (zh) * | 2012-05-10 | 2014-03-19 | 松下电器产业株式会社 | 具有电源电压的稳定化结构的三维集成电路及其制造方法 |
CN103650136B (zh) * | 2012-05-10 | 2017-05-24 | 松下知识产权经营株式会社 | 具有电源电压的稳定化结构的三维集成电路及其制造方法 |
JP2014072289A (ja) * | 2012-09-28 | 2014-04-21 | Renesas Electronics Corp | 半導体装置 |
JP2014072499A (ja) * | 2012-10-02 | 2014-04-21 | Hitachi Ltd | 半導体装置 |
JP2014220473A (ja) * | 2013-05-10 | 2014-11-20 | キヤノン株式会社 | 集積回路装置およびその構成方法 |
US9824954B2 (en) | 2013-05-10 | 2017-11-21 | Canon Kabushiki Kaisha | Semiconductor package comprising stacked integrated circuit chips having connection terminals and through electrodes symmetrically arranged |
JP2016025103A (ja) * | 2014-07-16 | 2016-02-08 | 富士通株式会社 | 電子部品、電子装置及び電子装置の製造方法 |
US10103124B2 (en) | 2015-08-17 | 2018-10-16 | Fujitsu Limited | Semiconductor device |
JP2018160490A (ja) * | 2017-03-22 | 2018-10-11 | 富士通株式会社 | 電子装置及びその製造方法、電子部品 |
JP7010428B2 (ja) | 2017-10-20 | 2022-01-26 | 電子商取引安全技術研究組合 | 半導体装置 |
JP2019079862A (ja) * | 2017-10-20 | 2019-05-23 | 電子商取引安全技術研究組合 | 半導体装置 |
JPWO2019146724A1 (ja) * | 2018-01-25 | 2020-02-06 | ソフトバンク株式会社 | 三次元積層集積回路 |
JP6998986B2 (ja) | 2018-01-25 | 2022-01-18 | ソフトバンク株式会社 | 三次元積層集積回路 |
WO2019146180A1 (ja) * | 2018-01-25 | 2019-08-01 | ソフトバンク株式会社 | 三次元積層集積回路の冷媒による冷却方式と、それを用いた三次元積層集積回路 |
US11244887B2 (en) | 2018-01-25 | 2022-02-08 | Softbank Corp. | Three-dimensional laminated integrated circuit |
JP2020107909A (ja) * | 2018-01-25 | 2020-07-09 | ソフトバンク株式会社 | 三次元積層集積回路 |
WO2019146039A1 (ja) * | 2018-01-25 | 2019-08-01 | ソフトバンク株式会社 | 三次元積層集積回路の冷媒による冷却方式と、それを用いた三次元積層集積回路 |
WO2019146212A1 (ja) * | 2018-01-25 | 2019-08-01 | ソフトバンク株式会社 | 三次元積層集積回路の冷媒による冷却方式と、それを用いた三次元積層集積回路 |
WO2019146724A1 (ja) * | 2018-01-25 | 2019-08-01 | ソフトバンク株式会社 | 中央演算装置、gpgpu、メモリーを含む三次元積層集積回路の冷媒による冷却方式と、それを用いた三次元積層集積回路と、給電方式 |
US10734360B2 (en) | 2018-03-22 | 2020-08-04 | Toshiba Memory Corporation | Semiconductor device |
JP2021090010A (ja) * | 2019-12-05 | 2021-06-10 | ソフトバンク株式会社 | 半導体パッケージならびに穴開きインターポーザーによる液浸冷却方式を用いた三次元積層集積回路 |
WO2021111960A1 (ja) * | 2019-12-05 | 2021-06-10 | ソフトバンク株式会社 | 半導体パッケージならびに穴開きインターポーザーによる液浸冷却方式を用いた三次元積層集積回路 |
GB2604267B (en) * | 2019-12-05 | 2024-09-25 | Softbank Corp | Semiconductor package and three-dimensional stacked integrated circuit using liquid immersion cooling system by perforated interposer |
GB2604267A (en) * | 2019-12-05 | 2022-08-31 | Softbank Corp | Three-dimensional laminated integrated circuit in which is used liquid immersion cooling method performed using semiconductor package and perforated |
WO2021171785A1 (ja) * | 2020-02-28 | 2021-09-02 | ソニーセミコンダクタソリューションズ株式会社 | 半導体装置、および、テストシステム |
US12140624B2 (en) | 2020-02-28 | 2024-11-12 | Sony Semiconductor Solutions Corporation | Semiconductor device and test system |
JP2021141240A (ja) * | 2020-03-06 | 2021-09-16 | 本田技研工業株式会社 | 半導体装置とその製造方法 |
JP7424580B2 (ja) | 2020-03-06 | 2024-01-30 | 本田技研工業株式会社 | 半導体装置とその製造方法 |
WO2024127812A1 (ja) * | 2022-12-16 | 2024-06-20 | 先端システム技術研究組合 | 半導体モジュール |
CN115842013B (zh) * | 2023-02-13 | 2023-06-09 | 浙江力积存储科技有限公司 | 一种三维堆叠存储器及其数据处理方法 |
CN115842013A (zh) * | 2023-02-13 | 2023-03-24 | 浙江力积存储科技有限公司 | 一种三维堆叠存储器及其数据处理方法 |
Also Published As
Publication number | Publication date |
---|---|
US20140159041A1 (en) | 2014-06-12 |
US9318397B2 (en) | 2016-04-19 |
JP5420671B2 (ja) | 2014-02-19 |
JPWO2011030467A1 (ja) | 2013-02-04 |
US8653645B2 (en) | 2014-02-18 |
US20120136596A1 (en) | 2012-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5420671B2 (ja) | 半導体装置 | |
US6734549B2 (en) | Semiconductor device having a device for testing the semiconductor | |
US9134959B2 (en) | Integrated circuit die stack | |
JP5165404B2 (ja) | 半導体装置と半導体装置の製造方法及びテスト方法 | |
US8242589B2 (en) | Semiconductor device | |
US7816154B2 (en) | Semiconductor device, a method of manufacturing a semiconductor device and a testing method of the same | |
US20130159587A1 (en) | Interconnect Redundancy for Multi-Interconnect Device | |
TW201507087A (zh) | 半導體裝置 | |
KR20120045331A (ko) | 반도체 장치와 그 테스트 방법 및 제조방법 | |
US11842985B2 (en) | Semiconductor devices having through-stack interconnects for facilitating connectivity testing | |
JP2013197576A (ja) | 半導体装置 | |
JP2008140530A (ja) | 半導体装置およびそのテスト方法 | |
Taouil et al. | Interconnect test for 3D stacked memory-on-logic | |
KR20170034178A (ko) | 반도체 패키지 장치 | |
KR20130070249A (ko) | 시스템 집적회로 | |
Yang et al. | A TSV repair scheme using enhanced test access architecture for 3-D ICs | |
Taouil et al. | Exploring test opportunities for memory and interconnects in 3D ICs | |
JP2011211113A (ja) | 半導体装置の製造方法 | |
Garita-Rodríguez et al. | Challenges for High Volume Testing of Embedded IO Interfaces in Disaggregated Microprocessor Products | |
Cheng | On 3-d ic design for test | |
JP2005156479A (ja) | 半導体集積回路装置およびそのテスト方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 09849244 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2011530722 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13388990 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 09849244 Country of ref document: EP Kind code of ref document: A1 |