Apparatus and memory device including spiral TSV connections
By adopting spiral through-substrate through-hole connection technology in 3D memory devices, multiple core chips or memory dies are stacked face to face and shared functional circuits, the signal delay and power consumption problems in the prior art are solved, and more efficient memory access and smaller chip size.
Patent Information
- Application Number
- CN202410688328.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-06-06
- Filing Date
- 2024-05-30
- Publication Date
- 2024-12-06
AI Technical Summary
While improving data reliability, high-speed memory access and reducing power consumption, existing 3D memory devices are difficult to effectively reduce chip size, and there are still problems with signal delay and power consumption.
Using helical through-substrate through-hole (TSV) connection technology, multiple core chips or memory dies are stacked face-to-face, and command signals are provided through spiral structure TSVs, so that the same functional circuit is shared between functional blocks at different levels. , reduces the number of circuits and power supply resistance, and improves the flexibility and efficiency of signal transmission.
Higher memory access speed, lower power consumption and smaller chip size are achieved, while reducing the number of circuits and power network requirements, improving chip robustness and overall device compactness.
Smart Images

Figure CN119095385A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to an apparatus and a memory device including a spiral TSV connection. Background Art
[0002] High data reliability, high-speed memory access, lower power consumption, and reduced chip size are features required of semiconductor memory. Some three-dimensional (3D) memory devices can be formed by vertically stacking memory chips (or memory dies) and interconnecting the stacked memory chips using through-substrate (or through-silicon) vias (TSVs). The benefits of 3D memory devices include: shorter interconnects, which reduce signal delays and power consumption; a larger number of vertical vias between layers, which allow wide bandwidth buses between functional blocks in different layers; and a relatively small footprint. Therefore, 3D memory devices contribute to higher memory access speeds, lower power consumption, and reduced chip size. Example 3D memory devices include high bandwidth memory (HBM) and hybrid memory cube (HMC). HBM is a type of memory that includes a high-performance dynamic random access memory (DRAM) interface chip and vertically stacked DRAM chips. Summary of the invention
[0003] On the one hand, the present disclosure provides a device comprising a plurality of core chips, the plurality of core chips including a plurality of spiral through-substrate vias (TSVs), wherein the plurality of core chips are stacked on each other in a face-to-face manner to define a common channel in a first and a second core chip facing each other among the plurality of core chips, the first and the second core chips respectively including a first functional circuit and a second functional circuit coupled to the first functional circuit, and the first functional circuit and the second functional circuit provide the same function.
[0004] On the other hand, the present disclosure further provides a device, which includes multiple core chips, which are stacked on each other in a face-to-face manner to provide shared functions between a first and a second core chip, each of the core chips including: multiple wiring layers on a semiconductor substrate, the topmost wiring layers of the first and second core chips facing each other; multiple channels, the channels of the first and second core chips corresponding to each other; and multiple spiral through-substrate vias (TSVs) configured to provide command signals to the channels of the first and second core chips in a spiral manner, wherein the first functional circuit of the first core chip and the second functional circuit of the second core chip provide the shared functions in response to the command signals provided by the spiral TSVs, and the shared functions are identical to each other.
[0005] On the other hand, the present disclosure further provides a memory device, comprising a plurality of memory dies, the plurality of memory dies including a plurality of spiral through-substrate vias (TSVs), wherein the plurality of memory dies are stacked on each other in a face-to-face manner to define a common channel in a first and a second memory die, the first and the second memory dies respectively including a first analog circuit and a second analog circuit, and the first analog circuit and the second analog circuit provide the same function through the common channel in response to a command signal provided by the spiral TSVs. BRIEF DESCRIPTION OF THE DRAWINGS
[0006] Figure 1A is a schematic illustration of at least a portion of a device including a plurality of core chips CC and an IF chip having a plurality of TSVs according to an embodiment of the present disclosure.
[0007] Figure 1B is a schematic illustration of a portion of a channel layout of a core chip layer according to an embodiment of the present disclosure.
[0008] Figure 1C is a schematic illustration of a portion of a core chip layer in a cross-sectional view according to an embodiment of the present disclosure.
[0009] Figure 1D is a schematic illustration of a portion of a core chip layer in a cross-sectional view according to an embodiment of the present disclosure.
[0010] Figure 2A is a schematic illustration of at least a portion of a device including a plurality of core chips CC and an IF chip having a plurality of TSVs according to an embodiment of the present disclosure.
[0011] Figure 2B is a schematic illustration of a portion of a channel layout of a core chip layer according to an embodiment of the present disclosure.
[0012] Figure 2C is a schematic illustration of a portion of a core chip layer in a cross-sectional view according to an embodiment of the present disclosure.
[0013] Figure 2D is a schematic illustration of a portion of a core chip layer in a cross-sectional view according to an embodiment of the present disclosure.
[0014] Figure 3A is a schematic illustration of at least a portion of a device 300 including a plurality of core chips CC and an IF chip having a plurality of TSVs according to an embodiment of the present disclosure.
[0015] Figure 3B is a schematic illustration of a portion of a channel layout of a core chip layer according to an embodiment of the present disclosure.
[0016] Figure 4 is a schematic diagram of an apparatus including a memory device and a processor according to an embodiment of the present disclosure. DETAILED DESCRIPTION
[0017] Various example embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. The following detailed description refers to the accompanying drawings, which show, by way of illustration, specific aspects of the embodiments in which the present disclosure may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the present disclosure. Other embodiments may be utilized, and structural changes, logical changes, and electrical changes may be made without departing from the scope of the present disclosure. The various embodiments disclosed herein are not necessarily mutually exclusive, as some disclosed embodiments may be combined with one or more other disclosed embodiments to form new embodiments.
[0018] In the description, common or related elements and substantially the same elements are denoted by the same symbols, and the description thereof may be reduced or omitted. In the drawings, the size and size ratio of each unit do not necessarily match the actual size and size ratio in the embodiments.
[0019] According to some embodiments of the present disclosure, a semiconductor device (e.g., a memory device) and another semiconductor device (e.g., a processor) may be disposed on a package substrate. In some embodiments, the memory device may be a dynamic random access memory (DRAM), a high bandwidth memory (HBM), or a hybrid memory cube (HMC). In some embodiments, the processor may be a central processing unit, a graphics processor, a memory controller, or the like. In some embodiments, an interposer IP may be disposed between the semiconductor device and the package substrate. The semiconductor device may be coupled to the interposer IP through external terminals such as bumps or microbumps. The external terminals may form electrodes between the semiconductor device and the interposer IP. The interposer IP may be stacked on the package substrate and coupled to the package substrate through, for example, solder balls.
[0020] According to some embodiments of the present disclosure, a semiconductor device may have a layered structure in which a plurality of core chips CC (or core dies) are stacked on each other on an interface (IF) chip (or IF die). The IF chip may be coupled to an interposer IP via an external terminal. In some embodiments, the core chip CC and the IF chip may be semiconductor chips, each semiconductor chip including a semiconductor substrate, such as a silicon (Si) substrate, and further including circuits, such as command circuits, control circuits, and buffer circuits. In some embodiments, each single chip may also be referred to as a semiconductor device, and a plurality of chips form a layered or stacked structure of a plurality of semiconductor devices. In some embodiments, each core chip CC includes a plurality of wiring layers (e.g., metal layers) provided by, for example, a back-end (BEOL) process and / or a middle-end (MOL) process on a semiconductor substrate. The wiring layers may be stacked and connected to each other via, for example, conductive contacts. In some embodiments, the core chip CC may be arranged in a face-up manner, wherein the uppermost wiring layer in the wiring layer of each chip faces up. The face-up manner may also be referred to as a forward manner or simply as a forward manner. In some embodiments, the core chip CC may be arranged in a face-down manner, wherein the uppermost wiring layer in the wiring layer of each chip faces down. The face-down method may also be referred to as the reverse method or simply reverse.
[0021] In the case of a memory device, a plurality of core chips CC may include a plurality of memory chips (or memory dies) stacked on an IF chip. The number of the plurality of memory chips may be four or eight, but is not limited thereto. Each memory chip may include a memory array for storing data, and further include a circuit for performing memory operations (such as read and write operations). Each memory chip may include only a single channel or multiple channels. In some embodiments, a channel may be a circuit block that can be operated as a single memory device. In some embodiments, a channel may represent a separable addressable memory space that can be accessed to read data from a channel and write data to a channel. A channel may be referred to or represented as channel A, B, C, D, E, F, G, H, etc. Each chip / die may include a command circuit for each channel. Command signals may be independently allocated and transmitted for each channel. The command signal may include a control signal, a clock signal, and a data signal.
[0022] In some embodiments, multiple core chips CC and IF chips can be coupled to each other through multiple conductive vias. The IF chip can provide multiple interfaces, which provide signals to the core chip CC or the channel of the core chip CC and / or receive signals from the core chip CC or the channel of the core chip CC. The signal may be an external signal transmitted by the IF chip. The conductive vias may provide input / output lines (IO) between the channel of the core chip CC and the IF chip. In each chip, internal signals may also be provided to or from the corresponding channel through the conductive vias and other conductive lines (such as wiring and terminals) provided to the chip. The conductive vias may be through-substrate or through-silicon vias (TSVs). TSVs may be provided to the core chip CC and the IF chip, vertically penetrating the corresponding chip, including its semiconductor substrate in the corresponding layer. Bumps may be arranged between adjacent chips in the upper and lower layers. Bumps may be arranged in alignment with TSVs on a horizontal plane. Bumps may electrically connect or couple adjacent chips.
[0023] The TSVs of the corresponding chips may be aligned with each other and form a vertically extending TSV row as a terminal providing connection between the chips in the upper and lower layers. The TSV row may provide IO between the channels of the chip. In some embodiments, in the case of a horizontal plane having an XY coordinate plane with an X-axis and a Y-axis perpendicular to each other, a TSV of one chip in the upper layer may have the same X-axis coordinate and Y-axis coordinate as another TSV of another chip in the adjacent lower layer. Therefore, the two TSVs penetrate the front and rear surfaces of the two chips along a vertical plane perpendicular to (or substantially perpendicular to within a reasonable tolerance of manufacturing, measurement, etc.) the horizontal plane at the same XY coordinate.
[0024] In some embodiments, the TSVs may be coupled or connected in a spiral structure. In a spiral structure, the TSVs or terminal connections provided by the TSVs deviate from one chip in the upper layer to another chip in the adjacent lower layer in one direction in a horizontal plane. For example, in the case of a horizontal plane along the XY axis, the XY coordinates of the TSVs in the upper layer are offset from the XY coordinates of the TSVs in the adjacent lower layer in at least one of the XY directions. Such TSV or terminal connections may be referred to as spiral TSVs or spiral connections.
[0025] Figure 1A An example of at least a portion of an apparatus 100 including a plurality of core chips CC and an IF chip 101 having a plurality of spiral TSVs 102 a and 102 b according to an embodiment of the present disclosure is schematically illustrated. Figure 1B An example of a portion of a channel layout of a core chip layer according to an embodiment of the present disclosure is schematically illustrated. Figure 1C An example of a portion of a core chip layer in a cross-sectional view according to an embodiment of the present disclosure is schematically illustrated. Figure 1DAnother example of a portion of a core chip layer in a cross-sectional view according to an embodiment of the present disclosure is schematically illustrated. Device 100 may be a semiconductor device. The semiconductor device may provide a memory device. Multiple core chips CC may include memory chips or memory dies each having a memory array of multiple memory cells. TSVs 102a and 102b may also be collectively referred to as TSVs 102. In the accompanying drawings, TSVs 102a and TSVs 102b and signals passing through and / or corresponding to TSVs 102a and TSVs 102b are illustrated with dashed lines and solid lines, respectively, to distinguish one from the other. Figure 1B In the figure, for the convenience of description, TSV 102 and other components, such as IF chip 101 and bump 103, are omitted.
[0026] A plurality of core chips CC0 to CC7 are stacked on the IF chip 101 in a sequential order of CC0, CC1, CC2, etc. The example shows eight core chips CC0 to CC7. In the case where each core chip is a memory chip, there are eight memory chips stacked on each other. The number of core chips in the stack is not limited to this example, and may be determined based on specifications, design, etc. The core chips CC0 to CC7 may also be referred to as core dies. Each of the core chips CC0 to CC7 may form an individual semiconductor device, and a group of core chips CC0 to CC7 may form a stacked semiconductor device.
[0027] Each core chip CC0 to CC7 includes a plurality of channels Ch. In the example, two adjacent core chips (eg, CC0 and CC1, CC2 and CC3, etc.) form a pair or a group (in Figure 1A ), and a plurality of channels ChA to ChE are allocated to each pair. For example, in the pair of core chips CC0 and CC1, the lower core chip CC0 includes four channels ChA to ChD, and the upper core chip CC1 includes four channels ChE to ChH. The four channels are arranged in a matrix in a plan view or a horizontal plane (for example, it is the XY axis plane illustrated in the figure) in the corresponding layer. In the case of a memory device, in one example, the core chip CC0 may be a 4GB memory chip / die, the core chip CC1 may be another 4GB memory chip / die, and the pair may provide a unit of 8GB memory. In another example, the pair may provide a unit of 16GB memory, wherein each of the two core chips of the pair is an 8GB memory chip / die. In the latter example, each core chip may include eight channels. The memory size is not limited thereto, and may be determined based on the specifications, design, etc. of the memory device.
[0028] Each core chip CC0 to CC7 may include a command circuit for an assigned channel. Command signals may be independently assigned and transmitted for each channel. Command signals may include control signals, clock signals, and data signals. Each core chip CC0 to CC7 includes a plurality of wiring layers (e.g., metal layers) 104a and 104b disposed above a semiconductor substrate 105. Wiring layers 104a and 104b are stacked and connected to each other via conductive contacts. Core chips CC0 to CC7 are stacked on IF chip 101 in a face-up manner, wherein the top wiring layer 104a of each chip (e.g., the top metal layer provided by a BEOL process) faces up. Between the respective chips are a plurality of bumps 103. Bumps 103 are aligned with corresponding TSVs 102. Bumps 103 include conductive materials. Bumps 103 electrically connect or couple upper and lower chips. Bumps 103 may be micro bumps. The bumps 103 facing each other in a vertical direction (eg, which is a direction along the Z-axis illustrated in the drawings) may be connected via the conductive pads 106 .
[0029] The IF chip 101 at the bottom of the chip stack structure can be coupled to the interposer IP (not depicted separately) via external terminals or balls. The IF chip 101 provides multiple interfaces for signals (including external signals) to and from the stacked core chips CC0 to CC7 or their channels. Signals are transmitted between the core chips CC0 to CC7 and the IF chip 101 through the TSVs 102 of the corresponding chips (and the bumps 103 between the corresponding chips).
[0030] The TSVs 102a and 102b or the terminal connections provided by the TSVs 102a and 102b are in a spiral structure, wherein the terminal connections deviate or shift in one direction from one chip of the upper layer to another chip of the lower layer in a horizontal plane. In the example, the TSVs 102a can be used for channels ChE to ChH of the corresponding core chip CC to provide individually assigned command signals, and the TSVs 102b can be used for channels ChA to ChD of the corresponding core chip CC to provide individually assigned command signals. From the top layer of the stacked structure, the TSVs 102a in the first column that form a pair with the adjacent TSVs 102b are first coupled or connected to the corresponding terminals of the channels ChE to ChH at the core chip CC7 layer of the first pair. The TSVs 102a then extend to the core chip CC6 layer of the same first pair, skipping channels ChA to ChD of the core chip CC6, and shifting in the X direction (e.g., by coupling or connecting to at least one of the bumps 103 or at least one of the wiring layers 104b at adjacent positions in the XY axis plane). TSV 102a extends to the core chip CC5 layer in the next pair and is offset back in the opposite X direction (e.g., by coupling or connecting to one of the bumps 103 or one of the wiring layers 104b at the original position in the XY axis plane) to couple or connect to the corresponding terminals of the channels ChE to ChH of the core chip CC5. The same spiral path continues until it reaches the IF chip 101. At an adjacent position in the first column, TSV 102b extends from the core chip CC7 layer to the core chip CC6 layer and is offset in the X direction (e.g., by coupling or connecting to one of the bumps 103 or one of the wiring layers 104b at an adjacent position in the XY axis plane) to couple or connect to the corresponding terminals of the channels ChA to ChD of the core chip CC6. TSV 102b then extends to the core chip CC5 layer, skipping the channels ChE to ChH of the core chip CC5, and is offset back in the opposite X direction (e.g., by coupling or connecting to one of the bumps 103 or one of the wiring layers 104b at the original position in the XY axis plane). TSV 102b further extends to the core chip CC4 layer and is offset again in the X direction to couple or connect to the corresponding terminals of channels ChA to ChD of the core chip CC4. The same spiral path continues until it reaches the IF chip 101. More specifically, referring to Figure 1CAs in the illustrated example, TSV 102a provides a spiral path between core chip CC0 (ChA to ChD) and core chip CC1 (ChE to ChE) with appropriate wiring forming a spiral structure, and provides, for example, clock signal E to logic circuit CC1-LC in core chip CC1 in a spiral manner. Similarly, TSV 102b provides another spiral path between core chip CC0 (ChA to ChD) and core chip CC1 (ChE to ChE) with appropriate wiring forming a spiral structure, and provides, for example, clock signal A to logic circuit CC0-LC in core chip CC0 in a spiral manner. In the illustrated example, spiral TSVs 102a and 102b have wiring that is arranged to connect at least one of wiring layers 104b in the XY plane and form a spiral structure so that the signal travels from TSV 102a to the adjacent TSV 102b, and vice versa. This spiral structure or spiral connection may also be referred to as a staggered structure or staggered connection. Also, the same spiral or staggered path is arranged for other rows forming pairs with other adjacent TSVs 102a and 102b. Since such a path is clearly illustrated in the drawings and its details can be known and understood from the above detailed description of the first row, its description is omitted herein.
[0031] The channels ChA to ChD of the core chips CC0, CC2, CC4 and CC6 may not be in an active state at the same time. Similarly, the channels ChE to ChH of the core chips CC1, CC3, CC5 and CC7 may not be in an active state at the same time. For example, the channel ChA of the core chip CC0 and the core chip CC2 in different pairs may not be active at the same time, and the channel ChE of the core chips CC1 and CC3 in different pairs may not be active at the same time. On the other hand, the channel ChA of the core chip CC0 and the channel E of the core chip CC1 in the same pair may be activated at the same time. Reference Figure 1C and 1D , each of the core chips CC0 and CC1 may include a logic circuit CC0-LC / CC1-LC and / or an analog circuit (e.g., a power circuit CC0-PC / CC1-PC), and all circuits may be in an active state. The same is true for the corresponding chip pair. In the case where the core chip CC is a memory chip such as a DRAM chip or an HBM chip, a logic circuit and an analog circuit may be provided for the corresponding memory bank.
[0032] Figure 2A An example of at least a portion of an apparatus 200 including a plurality of core chips CC and an IF chip 201 having a plurality of spiral TSVs 202 a and 202 b according to an embodiment of the present disclosure is schematically illustrated. Figure 2B An example of a portion of a channel layout of a core chip layer according to an embodiment of the present disclosure is schematically illustrated. Figure 2C An example of a portion of a core chip layer in a cross-sectional view according to an embodiment of the present disclosure is schematically illustrated. Figure 2D Another example of a portion of a core chip layer in a cross-sectional view according to an embodiment of the present disclosure is schematically illustrated. Device 200 may be a semiconductor device. The semiconductor device may provide a memory device. Multiple core chips CC may include memory chips or memory dies each having a memory array of multiple memory cells. TSVs 202a and 202b may also be collectively referred to as TSVs 202. In the accompanying drawings, TSVs 202a and TSVs 202b and signals passing through and / or corresponding to TSVs 202a and TSVs 202b are illustrated with dashed lines and solid lines, respectively, to distinguish them from each other. Figure 2B In the figure, for the convenience of description, TSV 202 and other components, such as IF chip 201 and bump 203, are omitted.
[0033] and Figure 1A and 1B The sequential stacking order of the core chips CC0 to CC7 in the example is different, for example, Figure 2A and 2B The stacking order of core chip CC1 and core chip CC2 in the example is opposite, so that core chip CC2 is adjacent to core chip CC0 in the vertical direction (that is, as illustrated in the figure, along the direction of the Z axis which is perpendicular to or substantially perpendicular to the X and Y axes). Channels ChA to ChD of core chip CC0 and channels ChA to ChD of core chip CC2 are arranged in the same matrix pattern. Channels ChE to ChH of core chip CC1 and channels ChE to ChH of core chip CC3 are arranged in the same matrix pattern. Similarly, the layer positions of core chip CC5 and core chip CC6 are exchanged so that in the vertical direction, core chip CC6 is adjacent to core chip CC4, and core chip CC5 is adjacent to core chip CC7. Channels ChA to ChD of core chips CC4 and CC6 are in the same matrix pattern. Channels ChE to ChH of core chips CC5 and CC7 are in the same matrix pattern. The matrix arrangement of channels Ch in the corresponding core chips CC is not limited to this and may be different from each other. For example, channels ChA to ChD of core chip CC2 may be arranged in a different position or in a different order than channels ChA to ChD of core chip CC0 within the matrix of core chip CC2, such that, for example, ChA of CC2 faces ChD of CC0. Various arrangements are applicable depending on specifications, design, etc.
[0034] In addition, adjacent core chips CC are arranged face to face. In the example, the core chips CC are stacked alternately face up and face down. Every other core chip of the first group is face up, and every other core chip of the second group is face down. For example, the core chip CC0 at the bottom of the stack is arranged face up on the IF chip 201, wherein its uppermost wiring layer 204a (at Figure 2C The top wiring layer 204a of the core chip CC2 is located at the bottom of the inner layer structure and faces the semiconductor substrate 205 of the core chip CC1 in a face-up manner. The top wiring layer of the core chip CC1 faces the top wiring layer of the core chip CC3 in a face-down manner. The core chips CC4 (face up) and CC6 (face down) and the core chips CC5 (face up) and CC7 (face down) are also arranged in a face-to-face manner. The face-up and face-down manners can be referred to as forward and reverse manners, respectively, or simply as forward and reverse. The top wiring layers of the corresponding core chips CC facing each other are connected by TSV 202 (and bumps 203, such as micro bumps, arranged in alignment with the corresponding TSV 202 between the adjacent layers). By stacking the core chips CC in a face-to-face manner or stacking the core chips CC in a face-up and face-down manner or alternately in forward and reverse directions, the degree of freedom in placing the TSV 202 is further increased, and the degree of freedom in exchanging signals between the stacked core chips CC is further improved. The uppermost wiring layer may be a top metal wiring layer formed by, for example, a back-end BEOL process.
[0035] In a manner similar to TSVs 102a and 102b, TSVs 202a and 202b may be used for channels ChE to ChH and channels ChA to ChD of the corresponding core chip CC, respectively. However, in order to correspond to the core chip CC and channel Ch arrangement of the present example, TSVs 202a and 202b provide a spiral connection different from the spiral connection provided by TSVs 102a and 102b. From the top layer of the stacked structure, TSVs 202a in the first column that form a pair with adjacent TSVs 202b are first coupled or connected to the corresponding terminals of channels ChE to ChH (which may be in an active state) of the core chip CC7 and CC5 layers. TSVs 202a extend to the core chip CC6 layer, offset in the X direction (e.g., by coupling or connecting to at least one of the bumps 203 or at least one of the wiring layers 204b in an adjacent position in the XY plane), and further extend to the core chip CC4 layer, while skipping channels ChA to ChD of the core chips CC6 and CC4 (for which TSVs 202a are not responsible). TSV 202a then extends to the core chip CC3 layer, deflects back in the opposite X direction (e.g., by coupling or connecting to at least one of bumps 203 or at least one of wiring layers 204b at the original position in the XY plane), and reaches CC1 while coupling or connecting to corresponding terminals of channels ChE to ChH (which may be in an active state) of core chips CC3 and CC1. TSV 202a further extends to the core chip CC2 layer, deflects again in the X direction, and reaches the core chip CC0 layer while skipping coupling and connection with channels ChA to ChD of core chips CC2 and CC0. TSV 202a then finally reaches IF chip 201. At an adjacent position in the same first column, TSV 202b extends from the core chip CC7 layer and passes through the core chip CC5 layer, skipping channels ChE to ChH of core chips CC7 and CC5 (for which TSV 202b is not responsible). TSV 202b reaches the core chip CC6 layer, deflects in the X direction, and extends to the core chip CC4 layer, while coupling or connecting to the corresponding terminals of channels ChA to ChD (which may be in an active state) of core chips CC6 and CC4. TSV 202b then extends and reaches the core chip CC3 layer, where it deflects back in the opposite X direction, further extends through the core chip CC1 layer to the core chip CC2 layer, where it deflects again in the X direction, and continues to the core chip CC0 layer. Until it reaches the core chip CC2 layer, TSV 202b skips coupling or connection with channels ChE to ChH (for which TSV 202b is not responsible) of core chips CC3 and CC1.At the core chip CC2 and CC0 layers, TSV 202b is coupled or connected to corresponding terminals of channels ChA to ChD (which may be in an active state) of core chips CC2 and CC0. TSV 202b then eventually reaches IF chip 201. The same or similar spiral (or staggered) paths are arranged for other rows that form pairs with other adjacent TSVs 202a and 202b. Since such paths are clearly illustrated in the drawings and their details can be easily understood and understood from the above detailed description of the first row, their description is omitted herein.
[0036] In the example core chip layer structure, in which the spiral TSV 202 connects the corresponding chip layers CC to provide the assigned or corresponding command signals thereto in a spiral manner, the channels ChA to ChD of one layer (e.g., the core chip CC0 layer or the core chip CC4 layer) and the channels ChA to ChD of another adjacent layer (e.g., the core chip CC2 layer or the core chip CC6 layer) may not be in an active state or an inactive state at the same time. Figure 2C, while the channels ChA to ChD (e.g., logic circuits CC0-LC and power analog circuits CC0-PC) of the core chip CC0 in a face-up manner may be in an inactive state, the channels ChA to ChD (e.g., logic circuits CC2-LC and power analog circuits CC2-PC) of the core chip CC2 in a face-down manner may be in an active state. Similarly, when the channels ChE to ChH of one layer (e.g., core chip CC1 layer or core chip CC5 layer) may be in an inactive state, the channels ChE to ChH of another adjacent layer (e.g., core chip CC3 layer or core chip CC7 layer) may be in an active state. In this way, the functional circuits or functions provided by the channels ChA to ChD or channels ChE to ChH of the adjacent core chips in the lower and upper layers (e.g., CC0 and CC2, CC1 and CC3, CC4 and CC6, and CC5 and CC7) may be shared through the corresponding spiral TSV 202. Therefore, at least one common channel or at least one common function is defined in the core chips facing each other. For example, the channel ChA of the core chip CC0 may be in an inactive state, in which the functional circuit provided thereto does not function or the function assigned thereto is not executed, while the channel ChA of the core chip CC2 may be in an active state, in which the functional circuit provided thereto functions or the function assigned thereto is executed in response to the assigned command signal provided by the corresponding spiral TSV 202. The functional circuit or function of the channel ChA of the core chip CC2 is the same as the functional circuit or function of the channel ChA of the core chip CC0. Similarly, the channel ChE of the core chip CC1 may be in an inactive state, in which the functional circuit or its function does not function or is executed, while the channel ChE of the core chip CC3 may be in an active state, in which the functional circuit or its function identical to the channel ChE of the core chip CC1 functions or is executed in response to the assigned command signal provided by the corresponding spiral TSV 202. Still in one example, channels ChA to ChD of a core chip (e.g., CC2) and channels ChA to ChD of another layer (e.g., CC6) may be activated at the same time, and channels ChE to ChH of a core chip (e.g., CC3) and channels ChE to ChH of another core chip (e.g., CC7) may be activated at the same time. The action and inaction states of the channels of the corresponding core chips are not limited to the described examples. Such states may be determined and switched as appropriate. Examples of functional circuits or functions include, but are not limited to, logic circuits and analog circuits. Examples of analog circuits include, but are not limited to, power supply circuits, power generation circuits, and compensation capacitor circuits. The command signal may come from a processor, such as a memory controller, and is transmitted in a spiral manner by the IF chip 201 via TSV 202 to the corresponding core chip CC and channel Ch. The command signal may include a control signal, a clock signal, a data signal, or the like.
[0037] According to this embodiment and example, a circuit having half of the full functionality may be required in each of two adjacent core chips CC to implement the full functionality between the adjacent core chips. Figure 2C , as an example, with Figure 1D Compared to the core chips CC0 and CC1 in the example, each of the adjacent core chips CC0 and CC2 requires half the number of power supply circuits CC0-PC / CC2-PC. This reduces the space required for the power supply circuit while suppressing or mitigating resistance value fluctuations. Thus, further chip miniaturization is achieved and the overall size of the device 200 is reduced.
[0038] In addition, since the uppermost wiring layers 204a of the core chips CC0 and CC2 face each other, the flexibility of arranging terminals (such as bumps 203a) between the two chips increases. In addition to bumps 203, bumps 203a are also arranged between the adjacent upper core chip and the lower core chip. Bumps 203a may be micro bumps. The adjacent core chip or its uppermost wiring layer 204a is connected by bumps 203a. Bumps 203a may be arranged in a position corresponding to or aligned with, for example, the uppermost wiring layer 204a and / or the functional circuit. Bumps 203a facing each other in the vertical direction may be connected via conductive pads 206. Therefore, the power supply from power chips CC0-PC and CC2-PC can be further efficiently shared by such terminals arranged in appropriate positions. Therefore, the degree of freedom of arranging power supply circuits between the stacked core chips CC increases. Therefore, the number of circuits can be further reduced, or at least the increase in the number of circuits can be further suppressed. The power supply resistance or the power supply resistance becomes further uniform in the circuit. Therefore, the need for a more robust power supply network can be reduced, and the core chip CC including the power supply circuit can be further reduced.
[0039] The power supply can be shared by one or more of the bumps 203a between the adjacent core chips. Electricity can be supplied or input to the logic circuit or other functional circuit. The power supply can also be shared by one or more of the TSVs in the multiple TSVs 202 between the adjacent core chips. Some signals can also be sent and received by the bumps 203a between the functional circuits of the adjacent core chips. In the case where the power input or signal input of the functional circuit is not directly connected to the bump 203 or the uppermost wiring layer 204a of the corresponding core chip, one or more wirings can be provided in an appropriate manner so that power and / or signals are input to the corresponding functional circuit and input to the adjacent core chip from the adjacent core chip. The wiring can be a combination of vertical wiring and / or vertical and horizontal wiring arranged to provide an electrical path.
[0040] refer to Figure 2DAs another example in which the compensation capacitor circuit or compensation capacitors CC0-Cap and CC2-Cap are disposed in adjacent core chips CC0 and CC2 and connected to each other through bumps 203a (and conductive pads 206 between bumps 203a), the compensation capacitance can be shared. If the compensation capacitors of the two core chips have the same capacitance value (Cap1) and if the core chip CC0 is not functional and the core chip CC2 is functional, then for the functional core chip CC0, the total compensation capacitance is doubled (2xCap1). This indicates that, for example, if a total capacitance value Cap1 is required between adjacent core chips and even if only one of the adjacent core chips is in an active state, each of the compensation capacitors CC0-Cap and CC2-Cap connected to each other through bumps 203a only needs half the capacitance value (Cap1 / 2) and the size can be reduced by half.
[0041] Figure 3A An example of at least a portion of an apparatus 300 including a plurality of core chips CC and an IF chip 301 having a plurality of spiral TSVs 302 a and 302 b according to an embodiment of the present disclosure is schematically illustrated. Figure 3B An example of a portion of a channel layout of a core chip layer according to an embodiment of the present disclosure is schematically illustrated. Device 300 may be a semiconductor device. The semiconductor device may provide a memory device. Multiple core chips CC may include memory chips or memory dies each having a memory array of multiple memory cells. TSVs 302a and 302b may also be collectively referred to as TSVs 302. In the accompanying drawings, TSVs 302a and TSVs 302b and signals passing through and / or corresponding to TSVs 302a and TSVs 302b are illustrated with dashed lines and solid lines, respectively, to distinguish them from each other. Figure 3B In the figure, for the convenience of description, TSV 302 and other components, such as IF chip 301 and bump 303, are omitted.
[0042] and Figure 1A and 1B The sequential stacking order of the core chips CC0 to CC7 in the example is different, for example, Figure 3A and 3B The stacking order of core chip CC2 and core chip CC3 in the example is reversed, so that core chip CC3 is adjacent to the lower pair of core chip CC1, and core chip CC2 is adjacent to the upper pair of core chip CC4. Similarly, the topmost pair of core chips CC6 and core chip CC7 of the stacking structure swap their layer positions, so that core chip CC7 is adjacent to the lower pair of core chip CC5.
[0043] In addition, the core chip CC is arranged in a face-to-face manner. In the example, the core chip CC is stacked alternately in a face-up and face-down manner. For example, the core chip CC1 is arranged in a face-up manner, wherein its uppermost wiring layer faces the uppermost wiring layer of the core chip CC3 in a face-down manner. The semiconductor substrate of the core chip CC1 faces the semiconductor substrate of the core chip CC0 arranged in a face-down manner on the IF chip 301. The core chip CC2 is arranged in a face-up manner, and the core chip CC4 is arranged in a face-down manner, wherein its uppermost wiring layers face each other. The core chips CC5 (face up) and CC7 (face down) also face each other. The core chip CC6 located at the top of the stacked structure faces up, wherein its semiconductor substrate faces the semiconductor substrate of the core chip CC7. The uppermost wiring layers of the corresponding core chips CC face each other and are connected by TSV 302 (and the bumps 303 arranged in alignment with the corresponding TSV 302 between the adjacent layers). By stacking the core chips CC in a face-to-face manner or alternately stacking the core chips CC in a face-up and face-down manner (or forward and reverse), the freedom of placing the TSVs 302 is further increased, and the freedom of exchanging signals between the stacked core chips CC is further improved.
[0044] In a manner similar to TSVs 202a and 202b, TSVs 302a and 302b may be used for channels ChE to ChH and channels ChA to ChD of the corresponding core chip CC, respectively. However, in order to correspond to the core chip CC and channel Ch arrangement of the present example, TSVs 302a and 302b provide a spiral connection different from the spiral connection provided by TSVs 202a and 202b. From the top layer of the stacked structure, TSVs 302b in the first column that form a pair with adjacent TSVs 302a are first coupled or connected to corresponding terminals of channels ChA to ChD (which may be in an active state) at the core chip CC6 layer. TSVs 302b then extend to the core chip CC7 layer and are offset in the X direction (e.g., by coupling or connecting to at least one of the bumps 303 or at least one of the metal layers (not separately depicted) at adjacent positions in the XY plane). TSV 302b further extends through the core chip CC5 layer to the core chip CC4 layer, deflects back in the opposite X direction (e.g., by coupling or connecting to at least one of the bumps 303 or at least one of the metal layers (not separately depicted) at the original position in the XY plane), and reaches the core chip CC2 layer. Through this path, TSV 302b skips coupling or connecting to channels ChE to ChH of core chips CC7 and CC5 (which are not responsible for TSV 302b), and instead couples or connects to corresponding terminals of channels ChA to ChD of core chips CC4 and CC2 (which may be in an active state). TSV 302b then extends to the core chip CC3 layer, deflects again in the X direction, extends through the core chip CC1 layer to the bottom core chip CC0 layer, where it couples or connects to corresponding terminals of channels ChA to ChD. TSV 302b deflects back in the opposite X direction again and finally reaches IF chip 301. At an adjacent position in the same first column, TSV 302a extends from the topmost core chip CC6 layer to the core chip CC7 layer, offsets in the X direction, and further extends to the core chip CC5 layer. TSV 302a skips coupling or connection with channels ChA to ChD of core chip CC6 (for which TSV 302a is not responsible), while TSV 302a is coupled or connected to channels ChE to ChH of core chips CC7 and CC5 (which may be in an active state). TSV 302a then further extends to the core chip CC4 layer, offsets back in the opposite X direction, passes through the core chip CC2 layer to the core core layer CC3, and offsets in the X direction again, and reaches the core chip layer CC1. Through this path, TSV 302a skips coupling or connection with channels ChA to ChD of core chips CC4 and CC2 (for which TSV 302a is not responsible), but is coupled or connected to corresponding terminals of channels ChE to ChH of core chips CC3 and CC1 (which may be in an active state).TSV 302a then extends to the core chip CC0 layer, where it skips coupling or connection with channels ChA to ChD, shifts back in the opposite X direction for the last time, and reaches IF chip 301. The same or similar spiral (or staggered) path is arranged for other rows that form pairs with other adjacent TSVs 302a and 302b. Since such a path is clearly illustrated in the drawings and its details can be easily understood and understood from the above detailed description of the first row, its description is omitted herein.
[0045] In an example core chip layer structure, in which the spiral TSV 302 connects the corresponding chip layer CC to provide command signals thereto in a spiral manner, channels ChA to ChD of one layer (e.g., core chip CC2 layer) may be in an active state, and channels ChA to ChD of another adjacent layer (e.g., core chip CC4 layer) may be in an inactive state at the same time (or vice versa). Similarly, when channels ChE to ChH of one layer (e.g., core chip CC1 layer or core chip CC5 layer) may be in an active state, channels ChE to ChH of another adjacent layer (e.g., core chip CC3 layer or core chip CC7 layer) may be in an inactive state at the same time (or vice versa). The same functional circuit or function provided by channels ChA to ChD or channels ChE to ChH of adjacent core chips (e.g., CC1 and CC3, CC2 and CC4, and CC5 and CC7) in the lower and upper layers may be shared through spiral TSVs 302a and 302b. Therefore, at least one common channel or at least one common function is defined in core chips facing each other. For example, the channel ChA of the core chip CC4 may be in an inactive state, wherein the functional circuit provided thereto is inactive or the function assigned thereto is not executed, while the channel ChA of the core chip CC2 may be in an active state, wherein the functional circuit provided thereto is active or the function assigned thereto is executed in response to the assigned command signal provided by the corresponding spiral TSV 302. The functional circuits or functions of the channels ChA of the core chips CC2 and CC4 are the same. Similarly, the channel ChE of the core chip CC3 may be in an inactive state, while the channel ChE of the core chip CC1 may be in an active state. Still in one example, the channels ChA to ChD of one core chip (e.g., CC2) and the channels ChA to ChD of another layer (e.g., CC6) may be activated at the same time, and the channels ChE to ChH of one core chip (e.g., CC1) and the channels ChE to ChH of another core chip (e.g., CC5) may be activated at the same time. The active and inactive states of the channels of the corresponding core chips are not limited to the described examples. Such states may be determined and switched as appropriate. Examples of functional circuits or functions include, but are not limited to, logic circuits and analog circuits. Examples of analog circuits include, but are not limited to, power circuits, power generation circuits, and compensation capacitor circuits. Command signals may come from a processor, such as a memory controller, and are transmitted from the IF chip 301 to the corresponding core chip CC and channel Ch in a spiral manner via TSV 302. The command signal may include a control signal, a clock signal, a data signal, or the like.
[0046] Therefore, a circuit having half of the full functions may be included in each of the two lower and upper core chips CC to realize the full functions between the two core chips CC. Figure 2CIn this case, the space required for the power supply circuit is reduced, and the resistance value fluctuation is suppressed or mitigated. Therefore, further chip miniaturization is achieved, and the overall size of the device 300 is reduced. In addition, the freedom to arrange the power supply circuit between the stacked core chips CC connected in a spiral manner by TSV 302a and 302b is increased. Through any TSV between multiple TSVs 302a and 302b, the power supply is more flexible. The number of circuits can be further reduced, or the increase in the number of circuits can be further suppressed. The power supply resistance or the power supply resistance becomes further uniform in the circuit. Therefore, the need for a more robust power supply network can be reduced, and the core chip CC containing the power supply circuit can be further reduced. In the case of a shared compensation capacitor circuit, similar to Figure 2D In the case of, if the compensation capacitors of the two core chips have the same capacitance value Cap1, and if one of the adjacent core chips is inactive and the other adjacent core chip is active, the total compensation capacitance of the active core chip is doubled (2xCap1). This indicates that, for example, if a total capacitance value Cap1 is required between the adjacent core chips, and even if only one of the adjacent core chips is in an active state, each of the compensation capacitors connected to each other through the bumps 303 (and, for example, the conductive pads provided between the bumps 303 facing each other) only needs half the capacitance value and can be reduced in size by half.
[0047] Figure 4 is a schematic diagram of an apparatus including a memory device 1 and a processor 2 according to an embodiment of the present disclosure. The memory device 1 and the processor 2 may be disposed on an interposer 5. In some embodiments, the memory device may be a dynamic random access memory (DRAM). In some embodiments, the processor 2 may be a central processing unit, a graphics processor, or a memory controller. The memory device 1 may include a terminal coupled to the interposer 5 via a ball 3 (e.g., a microbump). The processor 2 may include a terminal coupled to the interposer 5 via a ball 4 (e.g., a microbump), and a terminal further coupled to a corresponding terminal of the memory device 1 via the interposer 5. The interposer 5 may be stacked on a package substrate (not depicted separately) via a ball 6 (e.g., a solder ball). In some embodiments, the memory device 1 may include the apparatus 100, the apparatus 200, or the apparatus 300 described above.
[0048] In some embodiments, memory devices such as static random access memory (SRAM), flash memory, erasable programmable read-only memory (EPROM), magnetoresistive random access memory (MRAM), and phase change memory may be applicable. In some embodiments, logic ICs such as microprocessors and application specific integrated circuits (ASICs) may be applicable.
[0049] Although various embodiments of the present disclosure have been described in detail, it will be understood by those skilled in the art that the embodiments of the present disclosure may extend beyond the specifically described embodiments to other alternative embodiments and / or uses and modifications and equivalents thereof. In addition, other modifications within the scope of the present disclosure will be readily apparent to those skilled in the art based on the described embodiments. Various combinations or sub-combinations of specific features and aspects of the embodiments may also be considered and still fall within the scope of the present disclosure. It should be understood that the various features and aspects of the embodiments may be combined or substituted with each other to form a variation pattern of the embodiments. Therefore, it is desired that the scope of the present disclosure should not be limited to the specific embodiments described above.
Claims
1. A device comprising a plurality of core chips, wherein the plurality of core chips comprises a plurality of spiral through substrate vias (TSVs), wherein The plurality of core chips are stacked on each other in a face-to-face manner to define a common channel in first and second core chips facing each other among the plurality of core chips, The first and second core chips respectively include a first functional circuit and a second functional circuit coupled to the first functional circuit, and The first functional circuit and the second functional circuit provide the same function. 2 . The apparatus according to claim 1 , wherein each of the core chips includes a semiconductor substrate and a plurality of wiring layers, and uppermost wiring layers of the first and second core chips face each other.
3. The apparatus of claim 1, wherein the first functional circuit and the second functional circuit provide analog circuits. 4 . The apparatus of claim 3 , the analog circuit comprising a power supply circuit and / or a compensation capacitor circuit.
5. The apparatus of claim 3, wherein the first functional circuit and the second functional circuit further provide logic circuits.
6. The apparatus of claim 1, wherein each of the core chips comprises one or more memory chips or memory dies.
7. The apparatus of claim 1, wherein the core chip is on an interface chip coupled to an interposer. 8 . The apparatus of claim 1 , further comprising a plurality of bumps aligned with the corresponding TSVs and connecting the first and second core chips. 9 . The apparatus of claim 1 , wherein the first functional circuit and the second functional circuit are simultaneously activated in response to a command signal provided by the spiral TSV.
10. The apparatus of claim 1, wherein the first core chip is in an active state and the second core chip is in an inactive state.
11. An apparatus comprising a plurality of core chips stacked on each other in a face-to-face manner to provide shared functions between a first and a second core chip, each of the core chips comprising: A plurality of wiring layers on a semiconductor substrate, the uppermost wiring layers of the first and second core chips facing each other; a plurality of channels, wherein the channels of the first and second core chips correspond to each other; and a plurality of spiral through substrate vias (TSVs) configured to provide command signals to the channels of the first and second core chips in a spiral manner, wherein The first functional circuit of the first core chip and the second functional circuit of the second core chip provide the shared function in response to the command signal provided by the spiral TSV, and the shared functions are identical to each other.
12. The apparatus of claim 11, wherein the first functional circuit and the second functional circuit provide analog circuits.
13. The apparatus of claim 12, the analog circuit providing a power supply circuit and / or a compensation capacitor circuit.
14. The apparatus of claim 11, wherein the first functional circuit and the second functional circuit further provide logic circuits.
15. The apparatus of claim 11, wherein each of the core chips comprises one or more memory chips or memory dies. 16 . The apparatus of claim 11 , further comprising a plurality of bumps aligned with the TSVs and connecting the first core chip and the second core chip.
17. A memory device comprising a plurality of memory dies, the plurality of memory dies comprising a plurality of spiral through substrate vias (TSVs), wherein The plurality of memory dies are stacked on top of each other in a face-to-face manner to define a common channel in the first and second memory dies, The first and second memory dies include first and second analog circuits, respectively, and The first analog circuit and the second analog circuit provide the same function through the common channel in response to a command signal provided by the spiral TSV.
18. The memory device of claim 17, wherein each of the first analog circuit and the second analog circuit includes a power supply circuit and / or a compensation capacitor circuit.
19. The memory device of claim 17, wherein Each of the memory dies includes a plurality of wiring layers on a semiconductor substrate, The uppermost wiring layers of the first and second memory dies face each other, and The first and second memory dies are connected by a plurality of bumps aligned with the TSVs.
20. The memory device of claim 17, wherein the memory die is on an interface chip coupled to an interposer.
