Three-dimensional integrated circuits including multiple semiconductor dies and methods thereof

CN122991216APending Publication Date: 2026-10-27SAMSUNG ELECTRONICS CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202610539620.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2025-09-15
Filing Date
2026-04-22
Publication Date
2026-10-27

AI Technical Summary

Technical Problem

[0006]另外,在三维集成电路中,用于连接半导体管芯的工艺缺陷率可能显著高于用于单个硅元件内的金属布线的工艺缺陷率,因此半导体管芯之间的互连的缺陷率可能很高

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122991216A_ABST
    Figure CN122991216A_ABST
Patent Text Reader

Abstract

A method performed by at least one processor can include designing a three-dimensional integrated circuit including a first semiconductor die including a first logic circuit and a first interface circuit, a second semiconductor die including a second logic circuit and a second interface circuit, and a plurality of interconnects configured to connect the first interface circuit to the second interface circuit, and separately performing static timing analysis (STA) for a first set of STA targets, a second set of STA targets, and a third set of STA targets of the three-dimensional integrated circuit.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Cross-references to related applications

[0002] This application is based on and claims priority to Korean Patent Application No. 10-2025-0054112 filed with the Korean Intellectual Property Office on April 24, 2025 and Korean Patent Application No. 10-2025-0132053 filed with the Korean Intellectual Property Office on September 15, 2025, the disclosures of which are incorporated herein by reference in their entirety. Technical Field

[0003] Some embodiments of this disclosure relate to a three-dimensional integrated circuit, and more specifically, to a three-dimensional integrated circuit comprising a plurality of semiconductor dies stacked on top of each other. Background Technology

[0004] A three-dimensional integrated circuit may include multiple semiconductor dies (or integrated circuits) stacked on top of each other. The semiconductor dies included in a three-dimensional integrated circuit may be electrically connected to each other to perform communication. Specifically, the first semiconductor die and the second semiconductor die may perform communication when the first flip-flop of the first semiconductor die sends a signal to the second flip-flop of the second semiconductor die or when the third flip-flop of the first semiconductor die receives a signal from the fourth flip-flop of the second semiconductor die.

[0005] When designing 3D integrated circuits, the process and operating voltage of each of multiple semiconductor dies should be configured in various combinations to perform timing signatures, such as static timing analysis. Because a single semiconductor die can include a very large number of active devices and multiple dies are manufactured separately, the turnaround time for static timing analysis when designing 3D integrated circuits can increase exponentially.

[0006] Furthermore, in three-dimensional integrated circuits, the process defect rate for connecting semiconductor dies can be significantly higher than the process defect rate for metal wiring within a single silicon element, resulting in a potentially high defect rate for interconnects between semiconductor dies. To address this issue, three-dimensional integrated circuits can include repair circuitry for replacing defective interconnects with normal interconnects.

[0007] Since defective interconnects can cause overcurrent or leakage current in three-dimensional integrated circuits, a circuit structure is needed to isolate defective interconnects. Summary of the Invention

[0008] According to some embodiments of this disclosure, a 3D integrated circuit can be provided that effectively reduces the turnaround time for static timing analysis of the 3D integrated circuit by designing the interface circuit of the semiconductor die in the 3D integrated circuit to have a symmetrical and simple structure.

[0009] According to some embodiments of this disclosure, a three-dimensional integrated circuit may be provided, which includes circuitry configured to effectively isolate defective interconnections therein.

[0010] According to some embodiments of this disclosure, a method executed by at least one processor may include: designing a three-dimensional integrated circuit, the three-dimensional integrated circuit including: a first semiconductor die, the first semiconductor die including a first logic circuit and a first interface circuit; a second semiconductor die, the second semiconductor die including a second logic circuit and a second interface circuit; and a plurality of interconnects configured to connect the first interface circuit to the second interface circuit; and performing STA separately for a first static timing analysis (STA) target group, a second STA target group, and a third STA target group of the three-dimensional integrated circuit, wherein a first portion of the first logic circuit and the first interface circuit is included in the first STA target group, wherein a first portion of the second logic circuit and the second interface circuit is included in the second STA target group, and wherein a second portion of the first interface circuit, the plurality of interconnects, and the second portion of the second interface circuit are included in the third STA target group.

[0011] According to some embodiments of this disclosure, a method executed by at least one processor may include: designing a three-dimensional integrated circuit, the three-dimensional integrated circuit comprising: a first semiconductor die, the first semiconductor die including a first logic circuit configured to perform a first data processing operation and a first interface circuit including a first unit circuit; a second semiconductor die, the second semiconductor die including a second logic circuit configured to perform a second data processing operation and a second interface circuit including a second unit circuit; and a plurality of interconnects configured to connect the first unit circuit to the second unit circuit; and performing static timing analysis (STA) relative to the three-dimensional integrated circuit, wherein the first unit circuit includes a plurality of first flip-flops and a plurality of transmit buffers, the plurality of first flip-flops and the plurality of transmit buffers being configured to transmit a plurality of first data signals and a first clock signal through the plurality of interconnects, wherein the second unit circuit includes a plurality of second flip-flops and a plurality of receive buffers, the plurality of second flip-flops and the plurality of receive buffers being configured to receive a plurality of first data signals and a first clock signal through the plurality of interconnects, and wherein the components of the first unit circuit are arranged symmetrically relative to the components of the second unit circuit.

[0012] According to some embodiments of this disclosure, a method executed by at least one processor may include: designing a three-dimensional integrated circuit, the three-dimensional integrated circuit comprising: a first semiconductor die, the first semiconductor die including a first interface circuit, the first interface circuit including a plurality of first unit circuits; a second semiconductor die, the second semiconductor die including a second interface circuit, the second interface circuit including a plurality of second unit circuits; and a plurality of interconnects configured to connect the first interface circuits to the second interface circuits; and performing STA for at least one static timing analysis (STA) target group of the three-dimensional integrated circuit based on the process conditions of the first semiconductor die and the second semiconductor die, wherein each of the plurality of first unit circuits is configured to output a clock signal and a plurality of data signals through some of the plurality of interconnects, wherein each of the plurality of second unit circuits is configured to receive a clock signal and a plurality of data signals through some of the plurality of interconnects, and wherein the plurality of first unit circuits, the plurality of second unit circuits and the plurality of interconnects are included in at least one STA target group. Attached Figure Description

[0013] The embodiments will become clearer from the following detailed description taken in conjunction with the accompanying drawings, in which:

[0014] Figure 1 This is a block diagram of a three-dimensional integrated circuit according to an embodiment;

[0015] Figure 2A , 2B 2C is an interconnection diagram according to an embodiment;

[0016] Figure 3 This is a flowchart of a three-dimensional integrated circuit design method according to an embodiment;

[0017] Figure 4 yes Figure 3 A flowchart of a specific embodiment of the operation in the diagram;

[0018] Figure 5 This is a detailed block diagram of a three-dimensional integrated circuit according to an embodiment;

[0019] Figure 6 This is a detailed block diagram of a three-dimensional integrated circuit according to an embodiment;

[0020] Figure 7 This is a detailed block diagram of a three-dimensional integrated circuit according to an embodiment;

[0021] Figure 8 This is a detailed block diagram of a three-dimensional integrated circuit according to an embodiment;

[0022] Figure 9 This is a detailed block diagram of a three-dimensional integrated circuit according to an embodiment;

[0023] Figure 10 This is a detailed block diagram of a three-dimensional integrated circuit according to an embodiment;

[0024] Figure 11A and Figure 11B This is a diagram illustrating an embodiment of the drive circuit according to an example;

[0025] Figure 12A and Figure 12B This is a diagram illustrating an embodiment of the drive circuit according to an example;

[0026] Figure 13 This is a detailed block diagram of a three-dimensional integrated circuit according to an embodiment;

[0027] Figure 14 This is a detailed block diagram of a three-dimensional integrated circuit according to an embodiment;

[0028] Figure 15 This is a flowchart of a method for obtaining defect information about interconnects according to an embodiment;

[0029] Figure 16 This is a flowchart of an operation method for a three-dimensional integrated circuit according to an embodiment; and

[0030] Figure 17 This is a structural diagram of an example electronic system including a high-bandwidth memory (HBM) according to an embodiment. Detailed Implementation

[0031] In the following, non-limiting exemplary embodiments of the present disclosure are described in detail with reference to the accompanying drawings.

[0032] It will be understood that when a component or layer is referred to as being "on" another component or layer, "connected to", or "coupled to" another component or layer, the component or layer may be directly on, directly connected to, or directly coupled to the other component or layer, or there may be intermediate components or layers. Conversely, when a component or layer is referred to as being "directly on" another component or layer, "directly connected to", or "directly coupled to" another component or layer, there are no intermediate components or layers.

[0033] Figure 1 This is a block diagram of a three-dimensional integrated circuit 1 according to an embodiment. Hereinafter, the embodiment is described focusing on the three-dimensional integrated circuit 1, which includes a first semiconductor die 10 and a second semiconductor die 20, but this is for ease of explanation. The three-dimensional integrated circuit 1 may include more than two semiconductor dies.

[0034] Reference Figure 1The three-dimensional integrated circuit 1 may include a first semiconductor die 10, a second semiconductor die 20, and a plurality of interconnects 30_1 to 30_n (e.g., interconnects 30_1, 30_2, ..., 30_n).

[0035] Each of the first semiconductor die 10 and the second semiconductor die 20 may include an integrated circuit performing a certain function, and may be referred to as a logic die, memory die, and analog die depending on its function. The first semiconductor die 10 and the second semiconductor die 20 may perform the same or different functions. The first semiconductor die 10 and the second semiconductor die 20 may be formed on a silicon substrate, a silicon-on-insulator substrate, or a compound semiconductor substrate. The first semiconductor die 10 and the second semiconductor die 20 may include a plurality of active devices and a plurality of passive devices, wherein each device can be wired through a metal layer. Depending on the process scheme for the three-dimensional integrated circuit 1, the process used to form the first semiconductor die 10 may be the same as or different from the process used to form the second semiconductor die 20. When designing the three-dimensional integrated circuit 1, it is advantageous to consider that the process used to form the first semiconductor die 10 may be different from the process used to form the second semiconductor die 20. This may increase the turnaround time for static timing analysis during the design process and thus reduce the efficiency of resources used in the design of the comparative embodiments. The characteristic structure of the three-dimensional integrated circuit 1 according to the embodiments can effectively reduce the turnaround time for static timing analysis.

[0036] Multiple interconnects 30_1 to 30_n, which can be structures for transmitting electrical signals between the first semiconductor die 10 and the second semiconductor die 20, can be implemented as through-silicon vias (TSVs), microbumps, or hybrid junctions.

[0037] In addition, the first semiconductor die 10 and the second semiconductor die 20 can be stacked perpendicularly to each other. The first semiconductor die 10 and the second semiconductor die 20 can be stacked using die-to-die stacking, die-to-wafer stacking, or wafer-to-wafer stacking methods.

[0038] In this embodiment, the first semiconductor die 10 may include a first interface circuit 100 and a first logic circuit 120. The second semiconductor die 20 may include a second interface circuit 200 and a second logic circuit 220.

[0039] A first logic circuit 120 capable of performing data processing operations (such as arithmetic, control, or decision operations) corresponding to certain functions supported by the first semiconductor die 10 may include at least one of arithmetic logic circuitry, control circuitry, registers, multiplexers, sequential logic circuitry, and combinational logic circuitry. A second logic circuit 220 capable of performing data processing operations (such as arithmetic, control, or decision operations) corresponding to certain functions supported by the second semiconductor die 20 may include at least one of arithmetic logic circuitry, control circuitry, registers, multiplexers, sequential logic circuitry, and combinational logic circuitry. In some embodiments, the first logic circuit 120 and the second logic circuit 220 may be defined differently, respectively, as all or part of the regions of the first semiconductor die 10 and the second semiconductor die 200, excluding the first interface circuit 100 and the second interface circuit 200.

[0040] In an embodiment, the first interface circuit 100 and the second interface circuit 200, as circuits responsible for transmitting and receiving various signals between the first semiconductor die 10 and the second semiconductor die 20, can be implemented with a symmetrical structure. In an embodiment, the first interface circuit 100 and the second interface circuit 200 can be implemented with a simple structure by including multiple flip-flops and multiple buffers. Furthermore, in an embodiment, the first interface circuit 100 can be arranged as close as possible to the boundaries of the first semiconductor die 10 adjacent to the multiple interconnects 30_1 to 30_n, and the second interface circuit 200 can be arranged as close as possible to the boundaries of the second semiconductor die 20 adjacent to the multiple interconnects 30_1 to 30_n.

[0041] In this embodiment, the multiple flip-flops and multiple buffers of the first interface circuit 100 and the second interface circuit 200 can be divided into multiple unit circuits. In this specification, a unit circuit can be defined as a circuit for transmitting or receiving a clock signal and a certain number of data signals.

[0042] The following describes a static timing analysis of a three-dimensional integrated circuit 1 during the design phase. In this specification, static timing analysis may include techniques for analyzing timing by statically calculating the delay time of each path in the target circuit to verify that all signals are correctly transmitted in the target circuit within a given clock cycle. In this specification, static timing analysis may also be referred to as STA.

[0043] In an embodiment, the first part of the first logic circuit 120 and the first part of the first interface circuit 100 may be included in the first STA target group STA_TG1, the first part of the second logic circuit 220 and the second part of the second interface circuit 200 may be included in the second STA target group STA_TG2, and the second part of the first interface circuit 100, the plurality of interconnects 30_1 to 30_n and the second part of the second interface circuit 200 may be included in the third STA target group STA_TG3.

[0044] In the embodiments, static timing analysis for the first STA target group STA_TG1, static timing analysis for the second STA target group STA_TG2, and static timing analysis for the third STA target group STA_TG3 can be performed separately.

[0045] In this embodiment, after completing the static timing analysis for the first STA target group STA_TG1 and the second target group STA_TG2, a static timing analysis for the third STA target group STA_TG3 can be performed. However, this is merely an example, and the embodiments of this disclosure are not limited thereto. Static timing analysis can be performed sequentially for the first to third STA target groups STA_TG1, STA_TG2, and STA_TG3 in various orders.

[0046] In an embodiment, a static timing analysis for a first STA target group STA_TG1 can be performed based on the process conditions of the first semiconductor die 10. In this specification, process conditions can be defined based on changes in the characteristics of the active device caused by process variations during the manufacturing process. For example, when assuming the active device is manufactured at a reference speed using a normal process, the process conditions can be defined as a typical-typical (TT) angle; when assuming the active device is manufactured more slowly due to process variations, the process conditions can be defined as a slow-slow (SS) angle; and when assuming the active device is manufactured more quickly due to process variations, the process conditions can be defined as a fast-fast (FF) angle. In some embodiments, process conditions can be defined as a greater number of angles than those described above. Since the first STA target group STA_TG1 belongs to the first semiconductor die 10, it is expected that the first STA target group STA_TG1 will have the same process variations. Therefore, a static timing analysis for the first STA target group STA_TG1 can be performed considering only the process conditions of the first semiconductor die 10. Therefore, the turnaround time for the static timing analysis may be short.

[0047] In this embodiment, a static timing analysis for the second STA target group STA_TG2 can be performed based on the process conditions of the second semiconductor die 20. That is, since the second STA target group STA_TG2 belongs to the second semiconductor die 20, it is expected that the second STA target group STA_TG2 will have the same process variations. Therefore, the static timing analysis for the second STA target group STA_TG2 can be performed considering only the process conditions of the second semiconductor die 20. Consequently, the turnaround time for the static timing analysis may also be short.

[0048] Furthermore, in this embodiment, a static timing analysis for the third STA target group STA_TG3 can be performed based on the process conditions of the first semiconductor die 10 and the second semiconductor die 20. That is, since a portion of the third STA target group STA_TG3 belongs to the first semiconductor die 10 and another portion belongs to the second semiconductor die 20, it is expected that the components of the third STA target group STA_TG3 will have different process variations from each other. Therefore, a static timing analysis for the third STA target group STA_TG3 can be performed considering all process conditions of the first semiconductor die 10 and the second semiconductor die 20. However, since the first interface circuit 100 and the second interface circuit 200 have symmetrical and simple structures, the turnaround time may not be long even when considering all process conditions of the first semiconductor die 10 and the second semiconductor die 20.

[0049] As a result, the overall turnaround time can be effectively reduced when designing the three-dimensional integrated circuit 1. Voltage and temperature conditions can be further considered in the static timing analysis of the first to third STA target groups STA_TG1, STA_TG2, and STA_TG3.

[0050] Figures 2A to 2C This is a diagram of interconnections 30A, 30B, and 30C according to an embodiment.

[0051] Reference Figure 2A The first semiconductor die 10A may include a substrate SUB1 and a layer L1 formed by a back-end process (BEOL), and the layer L1 may include an active device AC1 formed by a front-end process (FEOL). The second semiconductor die 20A may include a substrate SUB2 and a layer L2 formed by BEOL, and the layer L2 may include an active device AC2 formed by FEOL.

[0052] The first semiconductor die 10A and the second semiconductor die 20A can be stacked face-to-face (F2F). Specifically, a bonding pad 31A connected to the active device AC1 can be formed on layer L1 of the first semiconductor die 10A, and a bonding pad 32A connected to the active device AC2 can be formed on layer L2 of the second semiconductor die 20A, wherein the bonding pad 31A can be bonded to the bonding pad 32A. Therefore, the interconnect 30A can include bonding pads 31A and 32A.

[0053] Further reference Figure 2B The first semiconductor die 10B may include a substrate SUB1 on which a TSV 33B is formed, and a layer L1 on which an active device AC1 is formed and connected to the TSV 33B. The second semiconductor die 20B may include a substrate SUB2 on which a TSV 34B is formed, and a layer L2 on which an active device AC2 is formed and connected to the TSV 34B.

[0054] Specifically, the first semiconductor die 10B and the second semiconductor die 20B can be stacked back-to-back (B2B). Specifically, a bonding pad 31B connecting to TSV 33B can be formed on the substrate SUB1 of the first semiconductor die 10B, and a bonding pad 32B connecting to TSV 34B can be formed on the substrate SUB2 of the second semiconductor die 20B, wherein the bonding pad 31B can be bonded to the bonding pad 32B. Therefore, the interconnect 30B can include bonding pads 31B and 32B, and TSVs 33B and 34B.

[0055] Further reference Figure 2C The first semiconductor die 10C may include a substrate SUB1 and a layer L1 on which an active device AC1 is formed. The second semiconductor die 20C may include a substrate SUB2 on which a TSV 34C is formed and a layer L2 on which an active device AC2 is formed and connected to the TSV 34C.

[0056] Specifically, the first semiconductor die 10C and the second semiconductor die 20C can be stacked in a face-to-back (F2B) configuration. Specifically, a bonding pad 31C connected to the active device AC1 can be formed on layer L1 of the first semiconductor die 10C, and a bonding pad 32C connected to the TSV 34C can be formed on the substrate SUB2 of the second semiconductor die 20C, wherein the bonding pad 31C can be bonded to the bonding pad 32C. Therefore, the interconnect 30C can include bonding pads 31C and 32C and the TSV 34C.

[0057] However, Figures 2A to 2C The implementations of interconnects 30A, 30B, and 30C are merely examples, and the embodiments of this disclosure are not limited thereto. Interconnects can be implemented in various other forms within the spirit and scope of this disclosure.

[0058] Figure 3 This is a flowchart of a three-dimensional integrated circuit design method according to an embodiment.

[0059] Reference Figure 3 In operation S100, at least one processor can design a three-dimensional integrated circuit based on a circuit design tool. For example, the circuit design tool can correspond to an electronic design automation (EDA) tool.

[0060] In operation S110, at least one processor can perform static timing analysis for each STA target group. For example, at least one processor can perform static timing analysis on each STA target group. Figure 1 The first to third STA target groups, STA_TG1, STA_TG2, and STA_TG3, are individually subjected to static timing analysis. For example, after defining the specifications of the 3D integrated circuit in operation S100 and completing the conversion to gate-level circuitry through the register-transfer stage (RTL) designed by logic synthesis, operation S110 can be executed as a pre-placement STA. In another example, after completing the placement and routing of the 3D integrated circuit in operation S100, operation S110 can be executed as a post-placement STA. In yet another example, operation S110 can be executed as the final STA immediately preceding tape-out.

[0061] In operation S120, at least one processor can determine whether there are timing violations in the results of the static timing analysis that has been performed. For example, timing violations may include at least one of setup time violation, hold time violation, clock skew violation, pulse width violation, minimum / maximum delay violation, and recovery / removal violation.

[0062] When the result of operation S120 is "yes", operation S130 can be subsequently executed, causing at least one processor to change the design of the three-dimensional integrated circuit.

[0063] When the result of operation S120 is "no", operation S140 can then be executed, causing at least one processor to perform tape-out to send the final design of the three-dimensional integrated circuit to a semiconductor foundry.

[0064] Figure 4 yes Figure 3 A flowchart of a specific embodiment of operation S110 is provided below. For ease of understanding, reference is also made to... Figure 1 .

[0065] refer to Figure 4 Operation S110 may include operations S111 to S113.

[0066] In operation S111, at least one processor can perform static timing analysis for a first STA target group STA_TG1 based on the process conditions of the first semiconductor die 10. Specifically, at least one processor can perform static timing analysis by uniformly applying varying process conditions to the design circuit corresponding to the first portion of the first logic circuit 120 and the first interface circuit 100. For example, when there are x process conditions (where x is an integer of 1 or greater), static timing analysis in operation S111 can be performed for x combinations.

[0067] In operation S112, at least one processor can perform static timing analysis for the second STA target group STA_TG2 based on the process conditions of the second semiconductor die 20. Specifically, at least one processor can perform static timing analysis by uniformly applying varying process conditions to the design circuit corresponding to the first portion of the second logic circuit 220 and the second interface circuit 200. For example, when x process conditions exist, static timing analysis in operation S112 can be performed for x combinations.

[0068] In operation S113, at least one processor can perform static timing analysis for a third STA target group STA_TG3 based on the process conditions of the first semiconductor die 10 and the second semiconductor die 20. Specifically, at least one processor can perform static timing analysis by uniformly or non-uniformly applying varying process conditions to the design circuit corresponding to the second portion of the first interface circuit 100, the plurality of interconnects 30_1 to 30_n, and the second portion of the second interface circuit 200. For example, when x process conditions exist, static timing analysis can be performed for x... 2 Static timing analysis in the combined execution operation S113.

[0069] At least one processor can change voltage and temperature conditions as well as process conditions, and perform static timing analysis for the first to third STA target groups STA_TG1, STA_TG2 and STA_TG3.

[0070] Figure 5 This is a detailed block diagram of the three-dimensional integrated circuit 1A according to an embodiment.

[0071] Reference Figure 5 The three-dimensional integrated circuit 1A may include a first semiconductor die 10, a second semiconductor die 20, first interconnects 30_11 to 30_k1 (e.g., first interconnects 30_11, 30_21, ..., 30_k1) and 30_131, and second interconnects 30_12 to 30_m2 and 30_132.

[0072] The first semiconductor die 10 may include a first interface circuit 100 and a first logic circuit 120. The first interface circuit 100 may include a first unit circuit 110_1 and a third unit circuit 110_2.

[0073] The second semiconductor die 20 may include a second interface circuit 200 and a second logic circuit 220. The second interface circuit 200 may include a second unit circuit 210_1 and a fourth unit circuit 210_2.

[0074] For example, the first unit circuit 110_1 and the third unit circuit 110_2 can be referred to as unit transmitting circuits, and the second unit circuit 210_1 and the fourth unit circuit 210_2 can be referred to as unit receiving circuits. The first interface circuit 100 may also include a unit receiving circuit, and the second interface circuit 200 may also include a unit transmitting circuit. Specific embodiments are described below.

[0075] In the embodiment, since the first unit circuit 110_1 can correspond to the second unit circuit 210_1, the components of the first unit circuit 110_1 and the components of the second unit circuit 210_1 can be arranged symmetrically relative to each other.

[0076] In an embodiment, the first unit circuit 110_1 can transmit the first data signals D_11a to D_k1a (e.g., first data signals D_11a, D_21a, ..., D_k1a) and the first clock signal CLKaa received from the first logic circuit 120 to the second unit circuit 210_1 via the first interconnects 30_11 to 30_k1 and 30_131. The first unit circuit 110_1 may include first flip-flops 111_1 to 111_k (e.g., first flip-flops 111_1, 111_2, ..., 111_k) and first transmit buffers 112_1 to 112_k (e.g., first transmit buffers 112_1, 112_2, ..., 112_k) and 113. In this specification, a transmit buffer may be defined as a buffer configured to enhance the electrical characteristics of the signal to be transmitted. The first flip-flops 111_1 to 111_k can output the received first data signals D_11a to D_k1a, synchronized with the first clock signal CLKaa, to the first transmit buffers 112_1 to 112_k. The first clock signal CLKaa, which can be output by the first clock buffer 123_1 of the first logic circuit 120, can be based on the global clock signal CLK. The first transmit buffers 112_1 to 112_k can output the first data signals D_11a to D_k1a through the first interconnects 30_11 to 30_k1. In addition, the first transmit buffer 113 can output the received first clock signal CLKaa through the first interconnect 30_131.

[0077] In an embodiment, the second unit circuit 210_1 can transmit the first data signals D_11b to D_k1b (e.g., first data signals D_11b, D_21b, ..., D_k1b) and the first clock signal CLKab received through the first interconnects 30_11 to 30_k1 and 30_131 to the second logic circuit 220. The second unit circuit 210_1 may include second flip-flops 211_1 to 211_k (e.g., second flip-flops 211_1, 211_2, ..., 211_k) and first receive buffers 212_1 to 212_k (e.g., first receive buffers 212_1, 212_2, ..., 212_k) and 213. In this specification, a receive buffer can be defined as a buffer configured to enhance the electrical characteristics of the received signal. The second flip-flops 211_1 to 211_k can output the first data signals D_11b to D_k1b, which are synchronized with the first clock signal CLKab and received through the first interconnects 30_11 to 30_k1, to the second logic circuit 220. Additionally, the first receive buffer 213 can output the first clock signal CLKab received through the first interconnect 30_131 to the second flip-flops 211_1 to 211_k and the second logic circuit 220.

[0078] In the embodiment, since the third unit circuit 110_2 can correspond to the fourth unit circuit 210_2, the components of the third unit circuit 110_2 and the components of the fourth unit circuit 210_2 can be arranged symmetrically relative to each other.

[0079] In an embodiment, the third unit circuit 110_2 can transmit the second data signals D_12a to D_m2a and the second clock signal CLKba received from the first logic circuit 120 to the fourth unit circuit 210_2 via the second interconnects 30_12 to 30_m2 and 30_132. For example, the third unit circuit 110_2 may include the same or a different number of flip-flops and transmit buffers as the first unit circuit 110_1. The second clock signal CLKba, which may be output by the second clock buffer 123_2 of the first logic circuit 120, may be based on the global clock signal CLK.

[0080] In an embodiment, the fourth unit circuit 210_2 can transmit the second data signals D_12b to D_m2b and the second clock signal CLKbb received through the second interconnects 30_12 to 30_m2 and 30_132 to the second logic circuit 220. For example, the fourth unit circuit 210_2 may include the same or a different number of flip-flops and receive buffers as the second unit circuit 210_1.

[0081] In an embodiment, the length of the first data path in the first semiconductor die 10 from the first logic circuit 120 to the first flip-flops 111_1 to 111_k can differ from the length of the second data path in the second semiconductor die 20 from the second flip-flops 211_1 to 211_k to the second logic circuit 220 by less than a threshold. The threshold can be determined based on the error range in the actual process. During the design phase of the three-dimensional integrated circuit 1A, the first data path can be designed to be the same as the second data path.

[0082] In an embodiment, the length of the third data path from the starting point of the first transmit buffers 112_1 to 112_k and 113 in the first semiconductor die 10 to the first interconnects 30_11 to 30_k1 and 30_131 can differ from the length of the fourth data path from the ending point of the first interconnects 30_11 to 30_k1 and 30_131 to the first receive buffers 212_1 to 212_k and 213 in the second semiconductor die 20 by less than a threshold. The threshold can be determined based on the error range in the actual process. During the design phase of the three-dimensional integrated circuit 1A, the third data path can be designed to be the same as the fourth data path.

[0083] Through the above design, the first unit circuit 110_1 and the second unit circuit 210_1 can have a symmetrical structure.

[0084] In an embodiment, in the first unit circuit 110_1, the length difference between the first clock signal paths used to supply the first clock signal CLKaa to the first flip-flops 111_1 to 111_k can be less than a threshold. The threshold can be determined based on the error range in the actual process. During the design phase of the three-dimensional integrated circuit 1A, the first clock signal paths can be designed to have the same length as each other. As a result, clock-to-data skew within the first unit circuit 110_1 can be minimized.

[0085] Furthermore, in the embodiment, the length difference between the second clock signal paths used to supply the first clock signal CLKab to the second flip-flops 211_1 to 211_k in the second unit circuit 210_1 can be less than a threshold. The threshold can be determined based on the error range in the actual process. During the design phase of the three-dimensional integrated circuit 1A, the second clock signal paths can be designed to have the same length as each other. As a result, clock-to-data skew within the second unit circuit 210_1 can be minimized.

[0086] In the embodiments, reference is made to Figure 1The first STA target group STA_TG1 described may include first flip-flops 111_1 to 111_k of the first unit circuit 110_1, the second STA target group STA_TG2 may include second flip-flops 211_1 to 211_k of the second unit circuit 210_1, and the third STA target group STA_TG3 may include first flip-flops 111_1 to 111_k, first transmit buffers 112_1 to 112_k and 113, first interconnects 30_11 to 30_k1 and 30_131, first receive buffers 212_1 to 212_k and 213, and second flip-flops 211_1 to 211_k.

[0087] In some embodiments, refer to Figure 1 The first STA target group STA_TG1 described may include portions of the first flip-flops 111_1 to 111_k of the first unit circuit 110_1 (e.g., input portions), the second STA target group STA_TG2 may include portions of the second flip-flops 211_1 to 211_k of the second unit circuit 210_1 (e.g., output portions), and the third STA target group STA_TG3 may include other portions of the first flip-flops 111_1 to 111_k (e.g., output portions), the first transmit buffers 112_1 to 112_k and 113, the first interconnects 30_11 to 30_k1 and 30_131, the first receive buffers 212_1 to 212_k and 213, and other portions of the second flip-flops 211_1 to 211_k (e.g., input portions).

[0088] The aforementioned design method can also be applied to the third unit circuit 110_2 and the fourth unit circuit 210_2.

[0089] The first to fourth unit circuits 110_1, 110_2, 210_1, and 210_2 according to the embodiment can be implemented with a simple structure. For example, since the first transmit buffer 112_1 is arranged between the first flip-flop 111_1 and the first interconnect 30_11, the first flip-flop 111_1 and the first transmit buffer 112_1 can form a signal path for the first data signal D_11a in the first unit circuit 110_1.

[0090] Figure 6 This is a detailed block diagram of the three-dimensional integrated circuit 1B according to an embodiment. In the following text, references may be omitted for ease of explanation. Figure 5 The given descriptions are basically the same.

[0091] Reference Figure 6The three-dimensional integrated circuit 1B may include a first semiconductor die 10, a second semiconductor die 20, first interconnects 30_11 to 30_k1 and 30_131, and second interconnects 30_12 to 30_m2 and 30_132.

[0092] For example, when the data signal transmission speed is greater than the clock signal transmission speed due to variations in the first interconnects 30_11 to 30_k1 and 30_131 and the second interconnects 30_12 to 30_m2 and 30_132, a hold-time violation may occur. Therefore, additional components can be provided to compensate for this violation.

[0093] In an embodiment, the first unit circuit 110_1 may further include a first delay buffer 114 for delaying the first clock signal CLKaa provided to the first flip-flops 111_1 to 111_k by a certain time. In this specification, a delay buffer can be defined as a buffer configured to intentionally add a certain delay to satisfy a certain timing.

[0094] In an embodiment, the second unit circuit 210_1 may further include a second delay buffer 214_1 to 214_k (e.g., second delay buffers 214_1, 214_2, ..., 214_k) for delaying the first data signals D_11b to D_k1b provided to the second flip-flops 211_1 to 211_k for a certain time.

[0095] However, Figure 6 The embodiments described are merely examples, and the embodiments disclosed herein are not limited thereto. Delay buffers may be arranged differently within the cell circuit to prevent timing violations.

[0096] As another example, when the transmission speed of the clock signal is greater than the transmission speed of the data signal due to variations in the first interconnects 30_11 to 30_k1 and 30_131 and the second interconnects 30_12 to 30_m2 and 30_132, additional components for compensation may be arranged differently from the embodiments described above.

[0097] Figure 7 This is a detailed block diagram of a three-dimensional integrated circuit 1C according to an embodiment. In the following text, references may be omitted for ease of explanation. Figure 5 The given descriptions are basically the same.

[0098] Reference Figure 7 The three-dimensional integrated circuit 1C may include a first semiconductor die 10, a second semiconductor die 20, first interconnects 30_11 to 30_k1 and 30_131, and second interconnects 30_12 to 30_m2 and 30_132.

[0099] In an embodiment, the first edge (e.g., rising edge) of the first clock signal CLKaa used to operate the first flip-flops 111_1 to 111_k may be different from the second edge (e.g., falling edge) of the first clock signal CLKab used to operate the second flip-flops 211_1' to 211_k'.

[0100] Therefore, in the static timing analysis of the three-dimensional integrated circuit 1C, the holding time violation can be greatly reduced, and the signal propagation delay through the first interconnects 30_11 to 30_k1 and 30_131 and the second interconnects 30_12 to 30_m2 and 30_132 can be reduced.

[0101] However, Figure 7 The embodiments of the three-dimensional integrated circuit 1C are merely examples, and the embodiments disclosed herein are not limited thereto.

[0102] Figure 8 This is a detailed block diagram of a three-dimensional integrated circuit 1D according to an embodiment. In the following text, references may be omitted for ease of explanation. Figure 5 The given descriptions are basically the same.

[0103] Reference Figure 8 The three-dimensional integrated circuit 1D may include a first semiconductor die 10, a second semiconductor die 20, first interconnects 30_11 to 30_k1 and 30_131, and redundant interconnects 30_711, 30_721 and 30_181.

[0104] In an embodiment, the first interface circuit 100 of the first semiconductor die 10 may include a first unit circuit 110_1, and the first unit circuit 110_1 may include first flip-flops 111_1 to 111_k, first transmit buffers 112_1 to 112_k and 113, a first repair circuit 115_1, a second repair circuit 115_2, first redundant flip-flops 116_1 and 116_2, and first redundant transmit buffers 117_1, 117_2, and 118. For example, the first redundant flip-flops 116_1 and 116_2 may replace some of the first flip-flops 111_1 to 111_k, and the first redundant transmit buffers 117_1, 117_2, and 118 may replace some of the first transmit buffers 112_1 to 112_k and 113.

[0105] In an embodiment, the first repair circuit 115_1 can route the first data signal D_11a to D_k1a based on defect information regarding the first interconnects 30_11 to 30_k1. For example, the first repair circuit 115_1 can route the first data signal D_11a to D_k1a to replace at least one of the first interconnects 30_11 to 30_k1 identified as defective with at least one of redundant interconnects 30_711 and 30_721. First redundant triggers 116_1 and 116_2 and first redundant transmit buffers 117_1 and 117_2 can form alternative data paths in the first semiconductor die 10.

[0106] In an embodiment, the second repair circuit 115_2 can route the first clock signal CLKaa based on defect information regarding the first interconnect 30_131. For example, the second repair circuit 115_2 can route the first clock signal CLKaa to replace the defective first interconnect 30_131 with a redundant interconnect 30_181. The first redundant transmit buffer 118 can form an alternative clock path in the first semiconductor die 10.

[0107] In an embodiment, the second interface circuit 200 of the second semiconductor die 20 may include a second unit circuit 210_1, and the second unit circuit 210_1 may include second flip-flops 211_1 to 211_k, first receive buffers 212_1 to 212_k and 213, a third repair circuit 215_1, a fourth repair circuit 215_2, second redundant flip-flops 216_1 and 216_2, and first redundant receive buffers 217_1, 217_2, and 218. For example, the second redundant flip-flops 216_1 and 216_2 may replace some of the second flip-flops 211_1 to 211_k, and the first redundant receive buffers 217_1, 217_2, and 218 may replace some of the first receive buffers 212_1 to 212_k and 213.

[0108] In an embodiment, the third repair circuit 215_1 can route the first data signal D_11b to D_k1b based on defect information regarding the first interconnects 30_11 to 30_k1. For example, the third repair circuit 215_1 can route the first data signal D_11b to D_k1b to replace at least one of the first interconnects 30_11 to 30_k1 identified as defective with at least one of redundant interconnects 30_711 and 30_721. The second redundant triggers 216_1 and 216_2 and the first redundant receive buffers 217_1 and 217_2 can form alternative data paths in the second semiconductor die 20.

[0109] In an embodiment, the fourth repair circuit 215_2 can route the first clock signal CLKab based on defect information regarding the first interconnect 30_131. For example, the fourth repair circuit 215_2 can route the first clock signal CLKab to replace the defective first interconnect 30_131 with a redundant interconnect 30_181. The first redundant receive buffer 218 can form an alternative clock path in the second semiconductor die 20.

[0110] Furthermore, in the embodiments, the components of the first unit circuit 110_1 and the components of the second unit circuit 210_1 can be arranged symmetrically relative to each other. That is, in Figure 8 In the first unit circuit 110_1, the first repair circuit 115_1, the second repair circuit 115_2, the first redundant triggers 116_1 and 116_2, and the first redundant transmit buffers 117_1, 117_2, and 118 can be arranged symmetrically with the third repair circuit 215_1, the fourth repair circuit 215_2, the second redundant triggers 216_1 and 216_2, and the first redundant receive buffers 217_1, 217_2, and 218 of the second unit circuit 210_1, respectively.

[0111] However, Figure 8 The embodiments of the three-dimensional integrated circuit 1D are merely examples, and the embodiments disclosed herein are not limited thereto.

[0112] Figure 9 This is a detailed block diagram of a three-dimensional integrated circuit 1E according to an embodiment. In the following text, references may be omitted for ease of explanation. Figure 5 The given descriptions are basically the same.

[0113] Reference Figure 9 The three-dimensional integrated circuit 1E may include a first semiconductor die 10, a second semiconductor die 20, third interconnects 30_111 to 30_g11 (e.g., third interconnects 30_111, 30_211, ..., 30_g11) and 30_1311, and fourth interconnects 30_121 to 30_h21 and 30_1321.

[0114] The second semiconductor die 20 may include a second interface circuit 200 and a second logic circuit 220. The second interface circuit 200 may include a fifth unit circuit 210_3 and a seventh unit circuit 210_4.

[0115] The first semiconductor die 10 may include a first interface circuit 100 and a first logic circuit 120. The first interface circuit 100 may include a sixth unit circuit 110_3 and an eighth unit circuit 110_4.

[0116] For example, the fifth unit circuit 210_3 and the seventh unit circuit 210_4 can be referred to as unit transmitting circuits, and the sixth unit circuit 110_3 and the eighth unit circuit 110_4 can be referred to as unit receiving circuits.

[0117] In the embodiment, since the fifth unit circuit 210_3 can correspond to the sixth unit circuit 110_3, the components of the fifth unit circuit 210_3 and the components of the sixth unit circuit 110_3 can be arranged symmetrically relative to each other.

[0118] In an embodiment, the fifth unit circuit 210_3 can transmit the third data signals D_11a1 to D_g1a1 (e.g., third data signals D_11a1, D_21a1, ..., D_g1a1) and the third clock signal CLKaa1 received from the second logic circuit 220 to the sixth unit circuit 110_3 via the third interconnects 30_111 to 30_g11 and 30_1311. The fifth unit circuit 210_3 may include third flip-flops 111_13 to 111_g3 (e.g., third flip-flops 111_13, 111_23, ..., 111_g3) and second transmit buffers 112_13 to 112_g3 (e.g., second transmit buffers 112_13, 112_23, ..., 112_g3) and 1133. The third flip-flops 111_13 to 111_g3 can output the received third data signals D_11a1 to D_g1a1, synchronized with the third clock signal CLKaa1, to the second transmit buffers 112_13 to 112_g3. The third clock signal CLKaa1, which can be output by the third clock buffer 223_1 of the second logic circuit 220, can be based on the global clock signal CLK. The second transmit buffers 112_13 to 112_g3 can output the third data signals D_11a1 to D_g1a1 through the third interconnects 30_111 to 30_g11. In addition, the second transmit buffer 1133 can output the received third clock signal CLKaa1 through the third interconnect 30_1311.

[0119] In an embodiment, the sixth unit circuit 110_3 can transmit the third data signals D_11b1 to D_g1b1 (e.g., third data signals D_11b1, D_21b1, ..., D_g1b1) and the third clock signal CLKab1 received through the third interconnects 30_111 to 30_g11 and 30_1311 to the first logic circuit 120. The sixth unit circuit 110_3 may include fourth flip-flops 211_13 to 211_g3 (e.g., fourth flip-flops 211_13, 211_23, ..., 211_g3) and second receive buffers 212_13 to 212_g3 (e.g., second receive buffers 212_13, 212_23, ..., 212_g3) and 2133. The fourth flip-flops 211_13 to 211_g3 can output the third data signals D_11b1 to D_g1b1, which are synchronized with the third clock signal CLKab1 and received through the third interconnects 30_111 to 30_g11, to the first logic circuit 120. Additionally, the second receive buffer 2133 can output the third clock signal CLKab1 received through the third interconnect 30_1311 to the first logic circuit 120.

[0120] In the embodiment, since the seventh unit circuit 210_4 can correspond to the eighth unit circuit 110_4, the components of the seventh unit circuit 210_4 and the components of the eighth unit circuit 110_4 can be arranged symmetrically relative to each other.

[0121] In an embodiment, the seventh unit circuit 210_4 can transmit the fourth data signals D_12a1 to D_h2a1 and the fourth clock signal CLKba1 received from the second logic circuit 220 to the eighth unit circuit 110_4 via the fourth interconnects 30_121 to 30_h21 and 30_1321. For example, the seventh unit circuit 210_4 may include the same or a different number of flip-flops and transmit buffers as the fifth unit circuit 210_3. The fourth clock signal CLKba1, which can be output by the fourth clock buffer 223_2 of the second logic circuit 220, may be based on the global clock signal CLK.

[0122] In an embodiment, the eighth unit circuit 110_4 can transmit the fourth data signals D_12b1 to D_h2b1 and the fourth clock signal CLKbb1 received through the fourth interconnects 30_121 to 30_h21 and 30_1321 to the first logic circuit 120. For example, the eighth unit circuit 110_4 may include the same or a different number of flip-flops and receive buffers as the sixth unit circuit 110_3.

[0123] Figure 10This is a detailed block diagram of a three-dimensional integrated circuit 300 according to an embodiment. (Refer to...) Figures 10 to 15 The described embodiments can be applied to Figures 1 to 9 For example, the configuration of the transmit driver circuit or the receive driver circuit described below can be applied to... Figure 5 The configuration of the first transmit buffers 112_1 to 112_k and 113 or the configuration of the first receive buffers 212_1 to 212_k and 213.

[0124] Reference Figure 10 The three-dimensional integrated circuit 300 may include a first semiconductor die 310, a second semiconductor die 320, a first interconnect 331, a second interconnect 332, and a third interconnect 333.

[0125] In an embodiment, the first semiconductor die 310 may include a first repair circuit 311, a first transmit drive circuit 312_1, a second transmit drive circuit 312_2, and a third transmit drive circuit 312_3. The second semiconductor die 320 may include a second repair circuit 321, a first receive drive circuit 322_1, a second receive drive circuit 322_2, and a third receive drive circuit 322_3.

[0126] In an embodiment, the first repair circuit 311 can route the first data signal D_1a and the second data signal D_2a based on defect information regarding any one of the first to third interconnects 331, 332, and 333, to send them to the second semiconductor die 320. That is, the first repair circuit 311 can route the first data signal D_1a and the second data signal D_2a based on defect information to block signal paths including defective interconnects. The second repair circuit 321 can route the first data signal D_1b and the second data signal D_2b based on defect information regarding any one of the first to third interconnects 331, 332, and 333, to send them to the logic circuitry of the second semiconductor die 320. That is, the second repair circuit 321 can route the first data signal D_1b and the second data signal D_2b based on defect information to block signal paths including defective interconnects.

[0127] In this embodiment, when the first interconnect 331 is functioning normally, the first transmitting drive circuit 312_1 can be enabled based on the first enable control signal EN11 to receive either the first data signal D_1a or the second data signal D_2a from the first repair circuit 311, and transmit the received data signal to the second semiconductor die 320 via the first interconnect 331. Additionally, when the first interconnect 331 is functioning normally, the first receiving drive circuit 322_1 can be enabled based on the second enable control signal EN12 to transmit the data signal received via the first interconnect 331 to the second repair circuit 321.

[0128] In an embodiment, when the first interconnect 331 is defective, the first transmitting drive circuit 312_1 can be disabled based on the first enable control signal EN11, and the first receiving drive circuit 322_1 can also be disabled based on the second enable control signal EN12. By disabling both the first transmitting drive circuit 312_1 and the first receiving drive circuit 322_1, the defective first interconnect 331 can be completely isolated. As a result, overcurrent or leakage current caused by the defective first interconnect 331 can be prevented.

[0129] Similarly, depending on whether the second interconnect 332 is defective, the second transmit drive circuit 312_2 and the second receive drive circuit 322_2 can be enabled or disabled by the third enable control signal EN21 and the fourth enable control signal EN22, respectively. Additionally, depending on whether the third interconnect 333 is defective, the third transmit drive circuit 312_3 and the third receive drive circuit 322_3 can be enabled or disabled by the fifth enable control signal EN31 and the sixth enable control signal EN32, respectively.

[0130] Figure 10 The three-dimensional integrated circuit 300 is merely an example, and the embodiments disclosed herein are not limited thereto. The three-dimensional integrated circuit 300 may include a greater number of semiconductor dies or interconnects, or each of the first semiconductor die 310 and the second semiconductor die 320 of the three-dimensional integrated circuit 300 may include a greater number of transmit drive circuits or receive drive circuits.

[0131] Figure 11A and Figure 11B This is a diagram illustrating one embodiment of the first transmitting drive circuit 312_1 and the first receiving drive circuit 322_1 according to an embodiment.

[0132] Reference Figure 11A The first transmitting drive circuit 312_1 can be connected to the first receiving drive circuit 322_1 via the first interconnection 331.

[0133] In an embodiment, the first transmitting drive circuit 312_1 may include a first tri-state buffer 312_11a, a first data input terminal 312_12, a first control input terminal 312_13, and a first data output terminal 312_14.

[0134] In this embodiment, the first tri-state buffer 312_11a can be enabled or disabled by receiving a first enable control signal EN11 via the first control input terminal 312_13. For example, when the first tri-state buffer 312_11a is enabled based on the first enable control signal EN11 being at a high level, data signals at low or high levels can be output through the first data output terminal 312_14, and when the first tri-state buffer 312_11a is disabled based on the first enable control signal EN11 being at a low level, the first data output terminal 312_14 can be in a high impedance (Hi-Z) state.

[0135] In an embodiment, the first receiving drive circuit 322_1 may include a first logic gate buffer 322_11a, a second data input terminal 322_12, a third data input terminal 322_13, and a second data output terminal 322_4. For example, the first logic gate buffer 322_11a may be an AND gate buffer.

[0136] In this embodiment, the first logic gate buffer 322_11a can receive the second enable control signal EN12 through the third data input terminal 322_13 to control the output of the second data output terminal 322_4. For example, based on the second enable control signal EN12 being at a high level, the first logic gate buffer 322_11a can output a signal corresponding to the level of the data signal received through the second data input terminal 322_12 through the second data output terminal 322_4. This can be defined as the first logic gate buffer 322_11a being enabled. Conversely, based on the second enable control signal EN12 being at a low level, the first logic gate buffer 322_11a can fix the second data output terminal 322_4 to a low level state, regardless of the state of the second data input terminal 322_12. This can be defined as the first logic gate buffer 322_11a being disabled.

[0137] In some embodiments, the first logic gate buffer 322_11a can be implemented as an OR gate buffer, and the structure of the first receive drive circuit 322_1 can be modified so that the OR gate buffer performs the same or similar operation as the AND gate buffer.

[0138] Further reference Figure 11B An example is described where a defective first interconnect 331 is isolated according to an embodiment.

[0139] First, an example involving the first transmit drive circuit 312_1 and the defective first interconnect 331 is described. When the first tri-state buffer 312_11a is disabled by the first enable control signal EN11 at a low level and the first data output terminal 312_14 is in the Hi-Z state, the current path from the power supply voltage VDD to ground of the first tri-state buffer 312_11a can be blocked, which would otherwise be formed due to a defect in the first interconnect 331. In some embodiments, when the first interconnect 331 is connected to ground due to a defect in the first interconnect 331, the current path to ground formed through the first interconnect 331 can be blocked by the first tri-state buffer 312_11a.

[0140] An example involving the first receive drive circuit 322_1 and the defective first interconnect 331 is also described. When the second data input 322_12 is disconnected from the first receive drive circuit 322_1 and floated due to the defect of the first interconnect 331, leakage current that may occur from the power supply voltage VDD of the first receive drive circuit 322_1 to ground can be blocked by disabling the first logic gate buffer 322_11a by the second enable control signal EN12 which is at a low level.

[0141] As a result, the defective first interconnect 331 can be completely isolated.

[0142] Figure 12A and Figure 12B This is a diagram illustrating another embodiment of the first transmitting drive circuit 312_1 and the first receiving drive circuit 322_1 according to the embodiments.

[0143] Reference Figure 12A The first transmitting drive circuit 312_1 can be connected to the first receiving drive circuit 322_1 via the first interconnection 331.

[0144] In an embodiment, the first transmitting drive circuit 312_1 may include a first tri-state buffer 312_11b, a first data input terminal 312_12, a first control input terminal 312_13, and a first data output terminal 312_14.

[0145] In this embodiment, the first tri-state buffer 312_11b can be enabled or disabled by receiving the first inverting enable control signal ENB11 via the first control input terminal 312_13. For example, when the first tri-state buffer 312_11b is enabled based on the first inverting enable control signal ENB11 being at a low level, a low-level or high-level data signal can be output via the first data output terminal 312_14, and when the first tri-state buffer 312_11b is disabled based on the first inverting enable control signal ENB11 being at a high level, the first data output terminal 312_14 can be in the Hi-Z state.

[0146] In an embodiment, the first receiving drive circuit 322_1 may include a first logic gate buffer 322_11b, a second data input terminal 322_12, a third data input terminal 322_13, and a second data output terminal 322_4. For example, the first logic gate buffer 322_11b may be an OR gate buffer.

[0147] In this embodiment, the first logic gate buffer 322_11b can receive the second inverting enable control signal ENB12 through the third data input terminal 322_13 to control the output of the second data output terminal 322_4. For example, based on the second inverting enable control signal ENB12 being at a low level, the first logic gate buffer 322_11b can output a signal corresponding to the level of the data signal received through the second data input terminal 322_12 through the second data output terminal 322_4. This can be defined as the first logic gate buffer 322_11b being enabled. Alternatively, based on the second inverting enable control signal ENB12 being at a high level, the first logic gate buffer 322_11b can fix the second data output terminal 322_4 to a high level state, regardless of the state of the second data input terminal 322_12. This can be defined as the first logic gate buffer 322_11b being disabled.

[0148] In some embodiments, the first logic gate buffer 322_11b can be implemented as an AND gate buffer, and the structure of the first receive drive circuit 322_1 can be modified so that the AND gate buffer performs the same or similar operation as the OR gate buffer.

[0149] Further reference Figure 12B An example of a defective first interconnect 331 being isolated according to an embodiment is described.

[0150] First, an example involving the first transmit drive circuit 312_1 and the defective first interconnect 331 is described. When the first tri-state buffer 312_11b is disabled by the first inverting enable control signal ENB11 at a high level and the first data output terminal 312_14 is in the Hi-Z state, the current path from the power supply voltage VDD to ground of the first tri-state buffer 312_11b can be blocked, which would otherwise be formed due to a defect in the first interconnect 331. In some embodiments, when the first interconnect 331 is connected to ground due to a defect in the first interconnect 331, the current path to ground formed through the first interconnect 331 can be blocked by the first tri-state buffer 312_11b.

[0151] An example involving the first receive drive circuit 322_1 and the defective first interconnect 331 is also described. When the second data input 322_12 is disconnected from the first receive drive circuit 322_1 and floated due to the defect of the first interconnect 331, leakage current that may occur from the power supply voltage VDD of the first receive drive circuit 322_1 to ground can be blocked by disabling the first logic gate buffer 322_11b by the second inverted enable control signal ENB12 which is at a high level.

[0152] As a result, the defective first interconnect 331 can be completely isolated.

[0153] Figure 13 This is a detailed block diagram of a three-dimensional integrated circuit 300 according to an embodiment. In the following text, references may be omitted for ease of explanation. Figure 10 The given descriptions are basically the same.

[0154] Reference Figure 13 The three-dimensional integrated circuit 300 may include a first semiconductor die 310, a second semiconductor die 320, and first to third interconnects 331, 332 and 333.

[0155] In the embodiment, with Figure 10 Compared to other embodiments, the first semiconductor die 310 may further include a first enable circuit 313 and a first memory 314. The first enable circuit 313 may generate a first enable control signal EN11, a third enable control signal EN21, and a fifth enable control signal EN31 based on first defect information DEFECT_INFO1 stored in the first memory 314 for the first to third interconnects 331, 332, and 333, and provide the first enable control signal EN11, the third enable control signal EN21, and the fifth enable control signal EN31 to the first to third transmit drive circuits 312_1, 312_2, and 312_3, respectively. In some embodiments, the first defect information DEFECT_INFO1 may include information that directly or indirectly indicates the transmit drive circuit to be enabled and the transmit drive circuit to be disabled.

[0156] In the embodiment, with Figure 10Compared to other embodiments, the second semiconductor die 320 may further include a second enable circuit 323 and a second memory 324. The second enable circuit 323 may generate a second enable control signal EN12, a fourth enable control signal EN22, and a sixth enable control signal EN32 based on the second defect information DEFECT_INFO2 stored in the second memory 324 for the first to third interconnects 331, 332, and 333, and provide the second enable control signal EN12, the fourth enable control signal EN22, and the sixth enable control signal EN32 to the first to third receive drive circuits 322_1, 322_2, and 322_3, respectively. In some embodiments, the second defect information DEFECT_INFO2 may include information that directly or indirectly indicates the receive drive circuit to be enabled and the receive drive circuit to be disabled.

[0157] In the embodiments, the first memory 314 and the second memory 324 can be implemented as non-volatile memory. For example, the non-volatile memory can be implemented as read-only memory (ROM), programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), flash memory, NAND flash memory, NOR flash memory, magnetoresistive RAM (MRAM), resistive RAM (ReRAM), or ferroelectric RAM (FRAM).

[0158] In some embodiments, the three-dimensional integrated circuit 300 may further include a first non-volatile memory disposed outside and connected to the first semiconductor die 310, and a second non-volatile memory disposed outside and connected to the second semiconductor die 320. For example, the first non-volatile memory may store first defect information DEFECT_INFO1, and the second non-volatile memory may store second defect information DEFECT_INFO2. When the chip including the three-dimensional integrated circuit 300 is powered on, the first semiconductor die 310 may read the first defect information DEFECT_INFO1 from the first non-volatile memory and store it in the first memory 314, while the second semiconductor die 320 may read the second defect information DEFECT_INFO2 from the second non-volatile memory and store it in the second memory 324. The first and second non-volatile memories may be integrated into a single non-volatile memory.

[0159] In some embodiments, the first memory 314 may be disposed outside the first semiconductor die 310, and the second memory 324 may be disposed outside the second semiconductor die 320.

[0160] Figure 14This is a detailed block diagram of a three-dimensional integrated circuit 300 according to an embodiment. In the following text, references may be omitted for ease of explanation. Figure 13 The given descriptions are basically the same.

[0161] Reference Figure 14 The three-dimensional integrated circuit 300 may include a first semiconductor die 310, a second semiconductor die 320, and first to third interconnects 331, 332 and 333.

[0162] In the embodiment, with Figure 13 Depending on the specific implementation, the second memory 324 can be implemented as volatile memory. For example, the volatile memory can be implemented as static RAM (SRAM), dynamic RAM (DRAM), synchronous RAM (SDRAM), or double data rate (DDR) SDRAM.

[0163] When the chip including the three-dimensional integrated circuit 300 is powered off, the second defect information DEFECT_INFO2 stored in the second memory 324 can be deleted. Therefore, the second defect information DEFECT_INFO2 can be further stored in the first memory 314, which can be a non-volatile memory.

[0164] In this embodiment, when the chip is powered on, the second semiconductor die 320 can receive the second defect information DEFECT_INFO2 from the first semiconductor die 310 and store it in the second memory 324.

[0165] Figure 15 This is a flowchart of a method for obtaining defect information about interconnects according to an embodiment.

[0166] Reference Figure 15 This allows for the generation of defect information regarding interconnections included in the 3D integrated circuit during the mass production phase. The mass production phase may include operations S200 to S220.

[0167] In operation S200, semiconductor dies included in a three-dimensional integrated circuit can be stacked.

[0168] In operation S210, tests can be performed on the interconnects connecting the semiconductor dies.

[0169] Additionally, during operation S220, defect information regarding interconnections can be obtained. The obtained defect information can be stored in at least one memory of the semiconductor die.

[0170] Figure 16 This is a flowchart of an operation method for a three-dimensional integrated circuit according to an embodiment.

[0171] Reference Figure 16During operation of S300, the chip, including the three-dimensional integrated circuit, can be powered on or can exit low-power mode.

[0172] In operation S310, the enable circuit included in each semiconductor die of the three-dimensional integrated circuit can set the enable control signal based on defect information.

[0173] In addition, during operation of S320, the semiconductor die of the three-dimensional integrated circuit can perform die-to-die (D2D) communication based on the set enable control signal.

[0174] Figure 17 This is a structural diagram of an example electronic system 1500 including a high-bandwidth memory (HBM) 1510 according to an embodiment.

[0175] Reference Figure 17 The electronic system 1500 may include an HBM 1510 and a host 1520. The HBM 1510 may be a three-dimensional integrated circuit according to the foregoing embodiments. The HBM 1510 may include a plurality of semiconductor dies 1511 to 1515 (e.g., semiconductor dies 1511, 1512, 1513, 1514, and 1515), and the plurality of semiconductor dies 1511 to 1515 may be electrically interconnected with each other via a plurality of interconnects. According to an embodiment, the interface circuitry included in each of the plurality of semiconductor dies 1511 to 1515 may be configured to perform static timing analysis during the design phase, and the interface circuitry may include drive circuitry for completely isolating defective interconnects.

[0176] The HBM 1510 and host 1520 can be mounted on an interposer 1530, and the interposer 1530 on which the HBM 1510 and host 1520 are mounted can be mounted on a package substrate 1540. Conductive devices (such as solder balls 1550) can be mounted on the bottom of the package substrate 1540. The host 1520 can be any semiconductor device requesting memory access.

[0177] Intermediate layer 1530 can be implemented in various forms, including TSV, organic printed circuit board (PCB) or non-TSV embedded multi-die interconnect bridge (EMIB).

[0178] In some embodiments, the host 1520 may also be implemented as a three-dimensional integrated circuit, wherein the host 1520 may include a plurality of stacked semiconductor dies, and the plurality of semiconductor dies may be electrically interconnected through a plurality of interconnects. According to an embodiment, the interface circuitry included in each semiconductor die of the host 1520 may be configured to effectively perform static timing analysis during the design phase, and the interface circuitry may include drive circuitry for completely isolating defective interconnects.

[0179] According to some embodiments of this disclosure, a three-dimensional integrated circuit may include: a first semiconductor die, the first semiconductor die including: a first logic circuit configured to perform a first data processing operation; and a first interface circuit including a first unit circuit; a second semiconductor die, the second semiconductor die including: a second logic circuit configured to perform a second data processing operation; and a second interface circuit including a second unit circuit; and a plurality of interconnects configured to connect the first unit circuit to the second unit circuit, wherein the first unit circuit includes a plurality of first flip-flops and a plurality of transmit buffers, the plurality of first flip-flops and the plurality of transmit buffers being configured to transmit a plurality of first data signals and a first clock signal through the plurality of interconnects, wherein the second unit circuit includes a plurality of second flip-flops and a plurality of receive buffers, the plurality of second flip-flops and the plurality of receive buffers being configured to receive a plurality of first data signals and a first clock signal through the plurality of interconnects, and wherein the components of the first unit circuit are arranged symmetrically with respect to the components of the second unit circuit.

[0180] According to some embodiments of this disclosure, a three-dimensional integrated circuit may include: a first semiconductor die including a first transmit drive circuit; a second semiconductor die including a first receive drive circuit; and a first interconnect configured to connect the first transmit drive circuit to the first receive drive circuit, wherein the first transmit drive circuit includes a first tri-state buffer configured to receive a first enable control signal via a control input terminal, and the first receive drive circuit includes a first logic gate buffer configured to receive a second enable control signal via a data input terminal.

[0181] While non-limiting exemplary embodiments of the present disclosure have been described above with reference to the accompanying drawings, it will be understood that various changes in form and detail may be made therein without departing from the spirit and scope of the present disclosure.

Claims

1. A method executed by at least one processor, the method comprising: Design a three-dimensional integrated circuit, the three-dimensional integrated circuit comprising: The first semiconductor die includes a first logic circuit and a first interface circuit; The second semiconductor die includes a second logic circuit and a second interface circuit; and Multiple interconnects, the multiple interconnects being configured to connect the first interface circuit to the second interface circuit; and STA is performed separately for the first, second, and third STA target groups of the three-dimensional integrated circuit. The first logic circuit and the first part of the first interface circuit are included in the first STA target group. The second logic circuit and the first part of the second interface circuit are included in the second STA target group, and The second part of the first interface circuit, the plurality of interconnects, and the second part of the second interface circuit are included in the third STA target group.

2. The method according to claim 1, wherein, The execution includes: The static timing analysis for the first STA target group is performed based on the first process conditions of the first semiconductor die. The static timing analysis for the second STA target group is performed based on the second process conditions of the second semiconductor die; and The static timing analysis for the third STA target group is performed based on the first process conditions and the second process conditions.

3. The method according to claim 1, wherein, The execution includes: after the static timing analysis for the first STA target group and the static timing analysis for the second STA target group are completed, performing the static timing analysis for the third STA target group.

4. The method according to claim 1, wherein, The first interface circuit includes a first unit circuit, and the first unit circuit includes: A plurality of first flip-flops, the plurality of first flip-flops being configured to output a plurality of first data signals, the plurality of first data signals being received by the plurality of first flip-flops from the first logic circuit and synchronized with a first clock signal received from the first logic circuit; and A plurality of first transmit buffers are configured to output the first clock signal and the plurality of first data signals through a first interconnect of the plurality of interconnects.

5. The method according to claim 4, wherein, The first part of the first interface circuit includes the plurality of first flip-flops, and the second part of the first interface circuit includes the plurality of first flip-flops and the plurality of first transmit buffers.

6. The method according to claim 4, wherein, The second interface circuit includes a second unit circuit, and the second unit circuit includes: A plurality of first receive buffers, the plurality of first receive buffers being configured to receive the plurality of first data signals and the first clock signal via the first interconnect; and A plurality of second flip-flops are configured to output a plurality of first data signals to the second logic circuit, the plurality of first data signals being received from some of the plurality of first receive buffers and synchronized with a first clock signal received from another of the plurality of first receive buffers.

7. The method according to claim 6, wherein, The second part of the first interface circuit includes the plurality of first flip-flops and the plurality of first transmit buffers, and the second part of the second interface circuit includes the plurality of first receive buffers and the plurality of second flip-flops.

8. The method according to claim 6, wherein, The length of the first data path from the first logic circuit in the first semiconductor die to the plurality of first flip-flops differs from the length of the second data path from the plurality of second flip-flops in the second semiconductor die to the second logic circuit by less than a predetermined first threshold.

9. The method according to claim 7, wherein, The length of the third data path from the plurality of first transmit buffers in the first semiconductor die to the first point of the first interconnect differs from the length of the fourth data path from the second point of the first interconnect to the plurality of first receive buffers in the second semiconductor die by less than a predetermined second threshold.

10. The method according to claim 6, wherein, The first unit circuit further includes a first delay buffer, which is configured to delay the first clock signal received from the plurality of first flip-flops, and The second unit circuit further includes a plurality of second delay buffers, which are configured to delay the plurality of first data signals received from the plurality of second triggers.

11. The method according to claim 6, wherein, The first edge of the first clock signal used to operate the plurality of first flip-flops is different from the second edge of the first clock signal used to operate the plurality of second flip-flops.

12. The method according to claim 4, wherein, The first unit circuit also includes: At least one first redundant trigger, the at least one first redundant trigger being configured to replace at least one of the plurality of first triggers; At least one first redundant transmit buffer, the at least one first redundant transmit buffer being configured to receive the output of the at least one first redundant trigger; A second redundant transmit buffer is configured to replace a first transmit buffer among the plurality of first transmit buffers, wherein the first transmit buffer is configured to output the first clock signal; A first repair circuit, configured to route the plurality of first data signals based on defect information regarding the first interconnect; and A second repair circuit is configured to route the first clock signal based on the defect information.

13. The method according to claim 4, wherein, The first interface circuit further includes a third unit circuit, and the third unit circuit includes: A plurality of third flip-flops, configured to output a plurality of second data signals, the plurality of second data signals being synchronized with and received from the second logic circuit, a second clock signal received from the first logic circuit; and A plurality of second transmit buffers are configured to output the second clock signal and the plurality of second data signals from the plurality of third flip-flops via a second interconnect of a plurality of interconnects.

14. The method according to claim 13, wherein, The second interface circuit includes a fourth unit circuit, and the fourth unit circuit includes: A plurality of second receive buffers, the plurality of second receive buffers being configured to receive the plurality of second data signals and the second clock signal via the second interconnect; and A plurality of fourth flip-flops are configured to output a plurality of second data signals to the second logic circuit, the plurality of second data signals being received from some of the plurality of second receive buffers and synchronized with a second clock signal received from another of the plurality of second receive buffers.

15. The three-dimensional integrated circuit according to claim 13, wherein, The number of the plurality of first flip-flops in the first unit circuit is different from the number of the plurality of third flip-flops in the third unit circuit, and the number of the plurality of first transmit buffers in the first unit circuit configured to output the first clock signal is the same as the number of the plurality of third transmit buffers in the third unit circuit configured to output the second clock signal.

16. The method according to claim 1, wherein, The first interface circuit includes a transmit driver circuit, which is configured to output a data signal or a clock signal through a first interconnect of the plurality of interconnects. The second interface circuit includes a receive driver circuit, which is configured to receive the data signal or the clock signal through the first interconnect. The transmitting drive circuit includes a tri-state buffer, which is configured to receive a first enable control signal via a control input terminal. The receiving drive circuit includes a logic gate buffer, which is configured to receive a second enable control signal through a data input terminal.

17. The three-dimensional integrated circuit according to claim 16, wherein, The logic gate buffer includes an AND gate buffer or an OR gate buffer.

18. A method executed by at least one processor, the method comprising: Design a three-dimensional integrated circuit, the three-dimensional integrated circuit comprising: A first semiconductor die, the first semiconductor die comprising: A first logic circuit, configured to perform a first data processing operation; and A first interface circuit, the first interface circuit including a first unit circuit; The second semiconductor die includes: A second logic circuit, configured to perform a second data processing operation; and A second interface circuit, the second interface circuit including a second unit circuit; and Multiple interconnects, the multiple interconnects being configured to connect the first unit circuit to the second unit circuit; and Static timing analysis (STA) is performed relative to the three-dimensional integrated circuit. The first unit circuit includes multiple first flip-flops and multiple transmit buffers, wherein the multiple first flip-flops and multiple transmit buffers are configured to transmit multiple first data signals and a first clock signal through the multiple interconnections. The second unit circuit includes multiple second flip-flops and multiple receive buffers, wherein the multiple second flip-flops and multiple receive buffers are configured to receive the multiple first data signals and the first clock signal through the multiple interconnections. The components of the first unit circuit are arranged symmetrically with respect to the components of the second unit circuit.

19. The three-dimensional integrated circuit according to claim 18, wherein, The execution includes performing the static timing analysis separately for the first STA target group, the second STA target group, and the third STA target group of the three-dimensional integrated circuit. The first logic circuit and the first part of the first unit circuit are included in the first STA target group. The second logic circuit and the first part of the second unit circuit are included in the second STA target group, and The second portion of the first unit circuit, the plurality of interconnections, and the second portion of the second unit circuit are included in the third STA target group.

20. A method executed by at least one processor, the method comprising: Design a three-dimensional integrated circuit, the three-dimensional integrated circuit comprising: A first semiconductor die, the first semiconductor die including a first interface circuit, the first interface circuit including a plurality of first unit circuits; A second semiconductor die, the second semiconductor die including a second interface circuit, the second interface circuit including a plurality of second unit circuits; and Multiple interconnects, the multiple interconnects being configured to connect the first interface circuit to the second interface circuit; and Based on the process conditions of the first semiconductor die and the second semiconductor die, STA is performed for at least one static timing analysis (STA) target group of the three-dimensional integrated circuit. Each of the plurality of first unit circuits is configured to output clock signals and multiple data signals through the plurality of interconnections. Each of the plurality of second unit circuits is configured to receive the clock signal and the plurality of data signals through some of the plurality of interconnects, and The plurality of first unit circuits, the plurality of second unit circuits, and the plurality of interconnections are included in the at least one STA target group.

Citation Information

Patent Citations

  • Aerosol generating devices with use restrictions

    KR1020250054112A

  • Light emitting device and display device having the same

    KR1020250132053A