MEMORY DEVICE COMPRISING A TILE ARRAY OF INDEPENDENTLY AND SIMULTANEOUSLY OPERABLE MEMORY TRANSISTORS - Patent application

Abstract

The memory device of the present disclosure comprises a plurality of tiles of a memory circuit, each tile including a physically separated array of storage transistors ("memory array") electrically connected to and operated by a module control circuit, the memory array having a three dimensional array of storage transistors arranged into multiple memory pages of the storage transistors, each storage transistor being accessed by a word line and a bit line, the module control circuit in communication with the memory array to perform memory operations on one or more memory pages of the storage transistors, each tile being individually addressed by an associated module control circuit and configurable to operate independently of one another to perform memory operations on units of memory pages of the storage transistors in the memory array in response to memory access commands specified for the tile, and two or more randomly addressed tiles of the memory circuit being configurable to simultaneously perform overlapping memory operations.

Description

[Technical field]

[0001] The present invention relates to memory systems having large capacity and high bandwidth, and in particular to memory devices including tiled arrays of memory transistors capable of operating independently and simultaneously, and to memory systems having multi-channel configurations for high bandwidth access. [Background technology]

[0002] High density memory arrays such as a three-dimensional array of NOR memory strings ("3D NOR memory arrays") are disclosed, for example, in U.S. Patent No. 10,121,553, entitled "Capacitively Coupled Nonvolatile Thin Film Transistor NOR Strings in a Three-Dimensional Array," filed August 26, 2016, and issued November 6, 2018. The entire disclosure of U.S. Patent No. 10,121,553 is incorporated herein by reference for all purposes. The storage transistors or memory transistors of U.S. Patent No. 10,121,553 are configured as a three-dimensional array of NOR memory strings formed on a plane of a semiconductor substrate. In addition to providing high memory density and capacity, such three-dimensional NOR memory arrays can provide highly desirable speeds of memory circuits that are comparable to conventional memory circuits, such as dynamic random access memories ("DRAMs"), which have much lower circuit density and consume significantly more power.

[0003] Furthermore, the memory circuit of Patent Document 1 is also referred to as a "quasi-volatile memory" or "QV memory". Each memory cell of the QV memory stores a data bit as an electric charge in a charge storage material (e.g., ONO) similar to a memory cell of a non-volatile memory (NVM). Due to the nature of the charge storage layer, a typical QV memory cell has a much longer data retention time than a DRAM cell, and therefore has a lower refresh rate than a DRAM cell. For example, a typical DRAM system is designed to refresh every 64 milliseconds, while a QV memory with a comparable effective access performance may refresh every 10 minutes. Such a reduction in refresh rate provides the QV memory with the significant advantages of reduced power requirements, reduced heat dissipation, and improved memory availability for better host performance. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] U.S. Pat. No. 10,121,553 [Patent Document 2] U.S. Patent Application Serial No. 17 / 812,375 [Patent Document 3] U.S. Patent Application Serial No. 17 / 936,320 [Patent Document 4] U.S. Patent Application Serial No. 17 / 525,712 Summary of the Invention [Means for solving the problem]

[0005] In accordance with the present disclosure, there is provided a memory device including an array of tiles of memory transistors operable independently and simultaneously from one another. In accordance with the present disclosure, there is also provided a memory system having a multi-channel configuration for high bandwidth access. The present disclosure is substantially as hereinbefore shown and / or described, for example with reference to at least one of the drawings, and as more fully set forth in the claims.

[0006] In one embodiment, a memory device of the present disclosure comprises a plurality of tiles of a memory circuit, each tile including a physically separated array of storage transistors ("memory array") electrically connected to and operated by a module control circuit, the memory array having a three dimensional array of storage transistors arranged into multiple memory pages of the storage transistors, each storage transistor being accessed by a word line and a bit line, the module control circuit in communication with the memory array to perform memory operations on one or more memory pages of the storage transistors, each tile being individually addressed by an associated module control circuit and configurable to operate independently of one another to perform memory operations on units of memory pages of the storage transistors in the memory array in response to memory access commands specified for the tile, and two or more randomly addressed tiles of the memory circuit being configurable to simultaneously perform overlapping memory operations.

[0007] In another embodiment, a memory module of the present disclosure comprises a plurality of semiconductor memory dies, each semiconductor memory die including a three dimensional array of storage transistors divided into a plurality of partitions, corresponding partitions across the plurality of semiconductor memory dies forming a memory channel, the plurality of partitions across the plurality of semiconductor memory dies forming a first number of memory channels accessible independently of one another; and a memory controller die including memory control circuitry for accessing and operating the plurality of semiconductor memory dies, the plurality of semiconductor memory dies being connected to the memory controller die via a first set of interconnect structures, the memory control circuitry including a first number of channel controllers, each channel controller being connected to operate one of the memory channels to perform memory operations on storage transistors associated with each memory channel independently of and in parallel with memory operations performed on the storage transistors of the other memory channels.

[0008] In some embodiments, another memory module of the present disclosure comprises a plurality of semiconductor memory dies, each semiconductor memory die including a plurality of memory arrays, each memory array including a three dimensional array of storage transistors, the plurality of semiconductor memory dies including a first number of semiconductor memory dies providing a designated memory capacity of the memory module and at least one spare semiconductor memory die providing redundant memory capacity; and a memory controller die including memory control circuitry that accesses and operates the plurality of semiconductor memory dies to perform memory operations, the plurality of semiconductor memory dies being connected to the memory controller die via a first set of interconnect structures, wherein the memory controller receives input requests for memory operations from a host processor addressed to a first memory address space spanning the first number of semiconductor memory dies and excluding the memory space of the at least one spare semiconductor memory die.

[0009] These and other advantages, aspects and novel features of the present invention, as well as details of illustrated embodiments thereof, will become more fully understood by reference to the following description and accompanying drawings. [Brief description of the drawings]

[0010] Various embodiments of the present invention are disclosed in the following detailed description and in the accompanying drawings, which depict various embodiments of the present invention, but which are not intended to be limited to the depicted embodiments. It should be understood that like reference numerals indicate like structural elements in the drawings, and that the depictions in the drawings are not necessarily drawn to scale.

[0011] [Figure 1A] FIG. 1A illustrates a memory module according to an embodiment of the present disclosure. [Figure 1B] FIG. 1B illustrates a memory module structure formed as stacked memory dies according to an embodiment of the present disclosure. [Figure 2A] FIG. 2A illustrates a memory module with multiple memory channels that can be accessed independently of one another, according to an embodiment of the present disclosure. [Figure 2B] FIG. 2B illustrates a memory module having a stacked memory die configuration, according to some embodiments. [Diagram 3] FIG. 3 is a top view of a semiconductor memory die illustrating the configuration of a storage transistor according to an embodiment of the present disclosure. [Figure 4A] FIG. 4A illustrates a memory structure including a three-dimensional array of NOR-type memory strings, according to some embodiments. [Figure 4B] FIG. 4B illustrates a memory structure including a three-dimensional array of NOR-type memory strings, according to some embodiments. [Diagram 5] FIG. 5 is a circuit diagram illustrating a memory array of NOR type memory strings according to an embodiment of the present disclosure. [Figure 6] FIG. 6 is a circuit diagram illustrating exemplary support circuitry coupled to bitlines of a NOR memory string according to an embodiment of the present disclosure. [Figure 7] FIG. 7 is a schematic diagram of a memory controller according to an embodiment of the present disclosure. [Figure 8] FIG. 8 is a schematic diagram of a channel controller implemented in the memory controller of FIG. 7 according to an embodiment of the present disclosure. [Figure 9] FIG. 9 is a cross-sectional view in the YZ plane of a tile in a memory device according to an embodiment of the invention. [Figure 10] FIG. 10 shows a two-dimensional array of tiles forming part of a memory device according to an embodiment of the present invention. [Figure 11] FIG. 11 is a block diagram illustrating modular control circuitry that may be incorporated into each tile of a memory device to provide intelligent, semi-autonomous memory operation control in accordance with some embodiments of the present invention. [Figure 12]FIG. 12 is a block diagram illustrating modular control circuitry that may be incorporated into each tile of a memory device according to another embodiment of the present invention. [Figure 13] FIG. 13 is a block diagram illustrating bit line selection and sense amplifier configuration in a module control circuit according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] In an embodiment of the present invention, a memory module includes a plurality of semiconductor memory dies connected to a memory controller, the semiconductor memory dies being divided into a plurality of independently accessible memory channels, each memory channel being formed across the plurality of semiconductor memory dies. That is, a memory channel is formed from each memory partition of the plurality of semiconductor memory dies. The memory controller includes a channel controller corresponding to each memory partition to control memory operations in each memory partition independently from memory operations in other partitions. With such a configuration, the memory module realizes a large capacity memory with high bandwidth access enabled by the plurality of parallel and independently accessible memory channels.

[0013] In another embodiment of the present invention, a memory system includes a memory device of storage transistors (or memory transistors) arranged into multiple memory arrays or "tiles," which interact with a controller device to perform read and write operations. In one embodiment, each tile is a unit of operation for memory access, also referred to herein as a "memory bank." In some embodiments, the controller device is configured to issue write and end write commands to the memory device that specify the memory bank. The write command instructs the memory device to initiate a write operation to a specified memory bank in the memory device, and the end write command instructs the memory device to end the memory operation in the memory bank. In some embodiments, the controller device issues a write suspend (abort) command as a write end command to end an ongoing write operation in one memory bank of the memory device in order to issue a read command to the memory device to read data from the same memory bank. In some examples, the read command is directed to a memory page in a different memory bank than the memory page in which the write operation was in progress. The ended write operation can be resumed after the read operation is completed.

[0014] As used herein, the terms "semiconductor memory die," "memory die," "semiconductor memory device," or "memory device" are used interchangeably to refer to a memory circuit of memory transistors or storage transistors formed on a semiconductor substrate. In embodiments of the present disclosure, the semiconductor memory device includes a three-dimensional array of storage transistors. In some embodiments, the semiconductor memory device is configured with a three-dimensional array of NOR memory strings formed on a semiconductor substrate as described in U.S. Pat. No. 6,399,363. In embodiments of the present disclosure, the semiconductor memory device includes a memory array of quasi-volatile storage transistors, also referred to as "quasi-volatile memory" or "QV memory." Because the quasi-volatile storage transistors have a much longer retention time than typical DRAM memory cells, the quasi-volatile memory device is refreshed much less frequently than typical DRAM memory devices. For example, a DRAM memory device needs to refresh its DRAM memory cells every 64 milliseconds, while a quasi-volatile memory device need only refresh its quasi-volatile storage transistors at intervals of 10 minutes or more. As used herein, a NOR memory string includes storage transistors formed on a plane of a semiconductor substrate sharing a common source region and a common drain region, and each storage transistor can be individually addressed and accessed. In some examples, a three-dimensional array can be formed by providing NOR memory strings in multiple planes (e.g., 8 or 16 planes) on a semiconductor substrate, with the NOR memory strings on each plane arranged in columns. As used herein, the term "memory device" can refer to a single memory die or a set of multiple memory dies connected to a memory controller.

[0015] As used herein, the term "storage transistor" is used interchangeably with "memory transistor" to refer to data storage structures formed in the memory die described herein. In some examples, the disclosed semiconductor memory devices including NOR-type memory strings of randomly accessible storage transistors (or memory transistors) can have applications in computer systems, e.g., as main memory, where data storage locations are directly accessible by the processor of the computer system, in the role previously played in the art by conventional random access memories (RAMs), such as dynamic RAMs (DRAMS) and static RAMs (SRAMs). For example, the disclosed memory structures can be applied in computer systems to function as random access memories to support the operation of microprocessors, graphical processors, and artificial intelligence processors. In other examples, the disclosed memory structures can also be applied to provide long-term data storage in computer systems, to form storage systems such as solid-state drives, or to replace hard drives.

[0016] In some embodiments, the semiconductor memory device is formed with thin-film storage transistors that implement charge trapping as a data storage mechanism, and data is stored in a charge storage film of each storage transistor. For example, the charge storage film may include a tunnel dielectric layer, a charge trapping layer, and a blocking layer, which may be realized as a multilayer of silicon oxide, silicon nitride, and silicon oxide, in that order, also referred to as an ONO layer. An electric field applied to the charge storage film changes the threshold voltage of the storage transistor by adding or removing charge from the charge trapping layer of the charge storage film, thereby encoding a given logic state into the storage transistor.

[0017] In another embodiment, the semiconductor memory device is formed using a ferroelectric field effect transistor as the storage transistor. More specifically, a ferroelectric field effect transistor (also referred to herein as a ferroelectric transistor or FeFET) is formed by using a ferroelectric material as a gate dielectric layer between the gate conductor and the channel of the field effect transistor. The ferroelectric transistor achieves memory function by storing data as polarization states in a ferroelectric gate dielectric layer (also referred to as a ferroelectric dielectric layer). Specifically, a voltage applied to the gate conductor induces an electric polarization in the ferroelectric dielectric layer, which can be reversed by applying a voltage of the opposite polarity. The induced polarization state in the ferroelectric gate dielectric layer changes the threshold voltage of the ferroelectric storage transistor. The change or shift in the threshold voltage of the ferroelectric storage transistor due to the different polarization states can be used to represent different logic states of data. For example, two logic states (e.g., "0" and "1") can be represented by a high and a low threshold voltage of the ferroelectric transistor as a result of the two induced electric polarization states in the ferroelectric dielectric layer. A three-dimensional array of NOR-type memory strings of thin-film ferroelectric transistors is disclosed, for example, in U.S. patent application Ser. No. 17 / 812,375, entitled "THREE-DIMENSIONAL MEMORY STRING ARRAY OF THIN-FILM FERROELECTRIC TRANSISTORS," filed on July 13, 2022 (Patent Document 2) (the entire disclosure of Patent Document 2 is incorporated herein by reference).

[0018] In some implementations, the ferroelectric dielectric layer is a doped hafnium oxide layer. In some examples, the doped hafnium oxide layer includes one or more of zirconium doped hafnium oxide (HZO), silicon doped hafnium oxide (HSO), aluminum zirconium doped hafnium oxide (HfZrAlO), aluminum doped hafnium oxide (HfO2:Al), lanthanum doped hafnium oxide (HfO2:La), hafnium zirconium oxynitride (HfZrON), hafnium zirconium aluminum oxide (HfZrAlO), and any hafnium oxide with zirconium impurities.

[0019] In yet another embodiment, the three-dimensional array of NOR type memory strings is formed using junctionless ferroelectric storage transistors. That is, the ferroelectric storage transistors do not include p / n junctions as drain or source regions in the channel. Instead, the drain and source regions are formed by conductive layers such as metal layers, and the semiconductor channel region is formed by amorphous oxide semiconductor materials such as indium gallium zinc oxide (IGZO). In some examples, the source / drain conductive layers can be formed from metal layers or low-resistance metal conductive materials, such as molybdenum (Mo), tungsten (W), tungsten nitride (WN), ruthenium, or titanium tungsten alloy (TiW). In some examples, the semiconductor channel region may be formed from other oxide semiconductor materials, such as indium zinc oxide (IZO), indium tungsten oxide (IWO), or indium tin oxide (ITO). A three-dimensional array of NOR-type memory strings of junctionless thin-film ferroelectric transistors is disclosed, for example, in U.S. patent application Ser. No. 17 / 936,320, entitled “Memory Structure Including Three-Dimensional NOR-Type Memory Strings of Junctionless Ferroelectric Memory Transistors and Method for Manufacturing the Same,” filed on September 28, 2022 (Patent Document 3) (the entire disclosure of Patent Document 3 is incorporated herein by reference).

[0020] FIG. 1A illustrates a memory module according to an embodiment of the present disclosure. With reference to FIG. 1A, the memory module 10 is constructed as a multi-die structure with one or more semiconductor memory dies 12 having memory arrays formed thereon and a memory controller die 14 (memory controller) having control circuitry formed thereon. In the example illustrated in FIG. 1A, the memory module 10 includes two semiconductor memory dies 12, namely memory die A (memory device A) and memory die B (memory device B), connected to each other by interconnect structures 13, such as through silicon vias (TSVs). The memory dies 12 communicate with the memory controller 14 (chiplets) through a memory array interface 15. In some embodiments, the memory array interface 15 is a high-bandwidth data interface implemented on an interconnect structure (e.g., TSV or hybrid bond) that connects the memory dies 12 to the controller die 14 (memory controller). The memory controller 14 also includes one or more external interfaces, such as a memory interface for host access or other system functions. For example, the memory controller 14 includes a host interface 16 for communicating with a host processor. The host interface 16 communicates with the host processor, for example to receive requests from the host to read data from or write data to the memory module 10, and to send responses to the host, such as write completion responses and read data.

[0021] As used herein, a "memory module" refers to one or more semiconductor memory dies connected to an associated memory controller die to form a high density and large capacity memory system. Each semiconductor memory die (also referred to as a memory die or memory device) includes multiple three-dimensional arrays of storage transistors (also referred to as memory transistors or memory cells) for storing memory data. As used herein, a memory controller die (also referred to as a "memory controller", "controller die", "controller device", or "chiplet") includes control circuitry for accessing and operating the memory devices and performing other memory control functions such as data routing and error correction. The control circuitry may also include one or more external interfaces, such as a memory interface for host access. In this embodiment, the memory module is constructed as a multi-die structure with the memory devices formed on one semiconductor die and the memory controller formed on another semiconductor die. The memory dies and memory controller die can be integrated using various integration technologies such as TSVs, hybrid bonds, exposed contacts, interposers, printed circuit boards, and other suitable interconnect technologies, particularly those for high density interconnects.

[0022] With such a configuration, the memory module 10 of the present disclosure can integrate one or more semiconductor memory dies and one memory controller die, thereby realizing a large capacity memory system with faster memory controller operation and faster memory performance. In addition, the memory dies and the controller die can be manufactured separately using dedicated manufacturing processes to optimize the performance of each integrated circuit. More specifically, the memory module 10 can be manufactured using a manufacturing process optimized for the memory circuits and a manufacturing process optimized for the memory controller. For example, the memory controller can be manufactured using a manufacturing process optimized for forming low-voltage and high-speed logic circuits. In this manner, the performance of the memory circuits and the performance of the memory controller can be individually optimized, resulting in a memory module 10 with large capacity, high bandwidth, and high-speed memory operation.

[0023] In one embodiment, the memory controller is formed as a separate semiconductor die or integrated circuit customized for use as a memory controller, e.g., an application specific integrated circuit. In another embodiment of the invention, the memory controller is realized by a general purpose integrated circuit (e.g., a central processing unit (CPU), a graphics processing unit (GPU), a communication chip, or a field programmable gate array). The functional blocks forming the memory controller are embedded in the general purpose integrated circuit, and the memory array interface of the memory controller is electrically and physically connected to the memory device using one of the techniques described above. With such a configuration, the embedded memory controller does not include host interface circuitry, but can directly communicate with logic circuitry via interconnect lines formed in or on the general purpose integrated circuit. Such a configuration is also referred to as "in-memory computing." In-memory computing is particularly desirable in data-intensive artificial intelligence and machine learning applications that require large amounts of memory in close proximity to the CPU or GPU core processor embedded in the memory controller functional block.

[0024] FIG. 1B illustrates a memory module structure formed as stacked memory dies according to an embodiment of the present disclosure. For example, the memory module 10 of FIG. 1A is formed by stacking two or more memory dies on one another and integrating the stacked memory dies with a memory controller die, as shown in FIG. 1B. Referring to FIG. 1B, a memory module 20 includes a plurality of memory devices 22 formed by stacking them on one another. For example, the memory module 20 includes a memory device 22-1 and a memory device 22-2 stacked on one another. The stacked memory devices 22 (or "memory stack") are then integrated with a controller die 24. In some embodiments, the memory devices 22 (memory stack) are interconnected to the controller die 24 via through silicon vias 23 (TSVs) formed through the memory devices 22 and connected to contact pads on the controller die 24. In other embodiments, other methods of interconnecting the stacked memory devices 22 to the controller die 24 can be used, such as hybrid bonds, copper studs, interposers, or other suitable interconnect methods. By integrating multiple memory devices (memory dies), memory module 20 can realize a large memory circuit capacity in a small footprint, and the resulting advantages are especially evident when multiple memory devices (memory dies) are stacked to provide high memory capacity while sharing a memory controller between multiple memory devices (memory dies) to reduce the cost per unit.

[0025] Memory Channel Configuration

[0026] FIG. 2A illustrates a memory module with multiple memory channels that can be accessed independently of each other, according to an embodiment of the present disclosure. Similar elements in FIG. 1A and FIG. 2A are labeled with similar reference numerals for ease of description. With reference to FIG. 2A, a memory module 30 includes multiple semiconductor memory dies 12 (memory devices) connected to each other via interconnect structures 13, such as through silicon vias (TSVs). In this example, the memory module 30 includes four memory dies 12, designated memory die 0 through memory die 3. The memory dies 12 communicate with a memory controller 14 via a memory array interface 32. The memory controller 14 includes one or more external interfaces, such as a memory interface 16 for communicating with a host or a host processor.

[0027] In an embodiment of the present disclosure, each memory die 12 is divided into N partitions, and corresponding partitions in all memory dies 12 are grouped to form N memory channels that can be accessed independently of each other (N is an integer equal to or greater than 1). In this embodiment, four memory channels Ch0 to Ch3 are illustrated. In other embodiments, the memory module may be divided into an appropriate number of memory channels based on factors such as bandwidth and / or reliability and quality of service requirements. With this configuration, each memory channel operates independently of each other to provide memory functions using storage transistors in the channel partitions. The memory controller 14 operates each memory channel independently of each other and in parallel with the other channels to perform memory operations such as read operations or write operations. The memory array interface 32 provides a separate memory channel interface for each memory channel. That is, the memory channels Ch0 to Ch3 are accessible independently of each other through their respective memory channel interfaces 32-0 to 32-3. In an embodiment of the present invention, the memory controller 14 has channel controllers 0-3, designated 17-0-17-3, for accessing respective memory channels Chn through respective memory array interfaces 32-n. The memory controller 14 has control logic circuitry 18, including control circuitry for controlling the channel controllers 17-n, host interface circuitry for communicating with a host through the memory interface 16, and other circuitry for controlling memory operations. The memory controller 14 routes read or write requests received from the host to respective channel controllers 17-n to store or retrieve memory data from the corresponding memory channels. By dividing and operating the memory die 12 as individually accessible memory channels, the memory module 30 can provide high bandwidth data transfers to the host.

[0028] A notable feature of the memory channel configuration in the memory module 30 is that each memory channel is formed across multiple semiconductor memory dies 12, and each memory channel is individually controlled by a respective channel controller of the memory controller 14. That is, the memory channel Ch0 is formed from each partition of the memory die 0 to memory die 3. Similarly, the memory channels Ch1 to Ch3 are formed from each partition of the memory die 0 to memory die 3, respectively. The memory channel configuration of the present invention is particularly advantageous when applied to a memory module having a stacked memory die configuration. FIG. 2B illustrates a memory module 30 having a stacked memory die configuration according to some embodiments. The stacked memory die configuration is formed by stacking the memory dies 0 to 3 on one another and stacking them on the memory controller 14. The stacked memory dies 0-3 are connected to the memory controller 14 by an interconnect structure. In an embodiment of the present invention, each memory die 12 is divided into N memory partitions. The corresponding memory partitions in the multiple memory dies 12 stacked on one another form a memory channel. In other words, each memory channel Ch-n includes the memory partitions of the memory die 0 to memory die 3. In the diagram shown in Figure 2B, memory channel Ch-n is formed longitudinally across multiple memory dies stacked on top of each other. N memory channels are formed parallel to each other across multiple memory dies stacked on top of each other. Each memory channel has its own channel controller so that each memory channel is individually controlled and accessed in parallel to receive input memory data and provide memory output data.

[0029] The inventive memory channel configuration implemented by the memory module 30 has several advantages. First, the memory channel configuration can be adapted to any number of memory dies 12 used to form a memory module. For N memory channels, the memory controller 14 can be easily modified to address any number of memory dies included in a memory stack by simply configuring the memory address bits designated to select a memory die from a memory channel. Thus, the inventive memory channel configuration allows for a scalable design of the memory module. Second, the inventive memory channel configuration allows the memory controller to exploit parallelism in memory accesses, resulting in more efficient utilization of the storage transistors on each memory die. In operation, the memory controller 14 minimizes access contention by distributing memory accesses among the N memory channels, thereby increasing utilization of the large number of storage transistors formed on each memory die 12. The inventive memory channel configuration allows for multiple storage transistors to be accessed simultaneously and in parallel across the N channels, thereby achieving high bandwidth memory accesses.

[0030] FIG. 3 is a top view of a semiconductor memory die illustrating the configuration of storage transistors according to an embodiment of the present disclosure. Referring to FIG. 3, a semiconductor memory die 40 (memory device) includes a plurality of three-dimensional arrays (memory arrays) of thin-film storage transistors. The memory array is configured as a two-dimensional array of tiles 42 (Tiles) formed on or in a semiconductor substrate (i.e., the tiles are arranged in rows and columns). Each tile 42 (also referred to as a "memory tile") includes a three-dimensional array of thin-film storage transistors formed on a plane of the semiconductor substrate. As used herein, tiles 42 in the memory die 40 refer to physically isolated arrays of memory cells with localized modular control circuits where tiles can operate simultaneously with other tiles and each tile performs memory operations based on an access unit of memory data, e.g., a page of memory data. Thus, tiles 42 in the memory die 40 refer to a regular array in an addressable modular structure of regularly arranged memory cells. In some embodiments, each tile 42 includes a memory array of semi-volatile storage transistors configured as a three-dimensional array of NOR type memory strings. The memory array in memory die 40 is also referred to herein as a semi-volatile memory circuit.

[0031] In the memory die 40, each tile 42 can be configured to be individually and independently addressable. In an embodiment of the present invention, each tile 42 is used as an operational unit for memory access and is also referred to as a "memory bank" or "bank". Thus, a memory bank is composed of one tile as an operational unit of memory access, and each tile or memory bank operates on one access unit of memory data (e.g., a page or "memory page") for each memory operation. That is, each memory access from the host is based on an access unit of memory data. An access unit is also referred to as a page or memory page of memory data. In the memory die 40, each memory bank includes one tile, and one tile alone provides an entire access unit of memory data or an entire memory page of memory data. For example, each memory page includes 512 bits of memory data. This is in contrast to conventional memory devices where a memory die includes multiple memory banks, each memory bank provides only a portion of an access unit of memory data, and the entire access unit of memory data must be obtained by combining memory data from multiple memory banks.

[0032] More specifically, in some embodiments, each tile is made up of multiple memory pages of storage transistors, with each memory page including a subset of the storage transistors in the tile. For example, a memory page may include 512 storage transistors, and a tile may include over 120K memory pages of storage transistors. With this configuration, the storage transistors in the memory device are accessed in units of memory pages having a predetermined byte size. That is, each read or write operation to the memory device is performed in units of memory pages. In one example, the memory device is accessed with a memory page size of 64 bytes or 512 bits.

[0033] In the embodiments described herein, a memory bank is described as being composed of one tile. In other embodiments, one or more tiles may be configured to form a memory access operation unit or memory bank. For example, a row of tiles or a two-dimensional block of tiles may be configured to be addressed together as a memory bank. In other words, in another embodiment, a memory bank may include a single tile 42, or a block of tiles, such as a row or section of tiles. Such an arrangement allows the tile 42 to be a building block that allows flexibility in configuring a memory module to fit the requirements of an application.

[0034] As used herein, a "tile" refers to a physically isolated memory array of memory cells with local modular control circuitry, and a "memory bank" refers to an operational or logical unit of memory access. As used herein, a memory bank comprises one tile, and the terms "memory bank" or "bank" are used interchangeably with the term "tile" to refer to a memory access operational unit consisting of a single tile or a single physically isolated memory array. It should be noted that, in general, a "tile" refers to a physical memory array, and a "memory bank" refers to a logical unit of memory access operation.

[0035] On the memory die 40, support circuits for operating the thin-film storage transistors of each tile are formed locally on or in the semiconductor substrate beneath the memory array. In some embodiments, the support circuits for each tile are formed locally and provided for modularization in a portion of the semiconductor substrate beneath each memory array. The tile-based support circuits, also referred to as modular control circuits or "under-array circuits" ("CuA"), may include various voltage sources for power supply, ground, programming, erase, or read voltages, sense amplifiers, various data latches or registers, various logic circuits, various analog circuits, and other circuits used in memory operations. Examples of logic circuits include timing control circuits, address decoder circuits, redundant logic circuits, and control circuits. Examples of analog circuits include data drivers, word line and bit line drivers and select transistors, and bias control transistors. Additionally, in embodiments of the present disclosure, each CuA includes a state machine or sequencer for executing instructions (commands) executed in the associated tile. Each CuA incorporates a sequencer that functions as the CuA's local processor, enabling the CuA to form intelligent control circuitry that allows each tile to operate independently and self-contained, and that allows multiple tiles to operate simultaneously and independently.

[0036] During operation, the sequencer in each CuA associated with each tile receives commands from the controller circuitry targeted to the associated tile and decodes the commands to perform memory operations in the associated tile. For example, the sequencer performs read and write operations in the associated tile in response to commands received from the controller device. The sequencer also issues instructions (commands) to perform entire read or write operation sequences on memory pages in the tile specified by memory addresses associated with the memory operations. Importantly, in the memory die 40, the complete circuitry of the local module control circuit (CuA) is replicated for each tile such that each tile operates independently to perform memory operations in response to commands from the controller device addressed to the particular tile.

[0037] In an embodiment of the present invention, a first manufacturing process forms a tile-based support circuit on a semiconductor substrate, and then the semiconductor substrate with the tile-based support circuit formed thereon is subjected to a second manufacturing process to form thin film storage transistors.

[0038] With this configuration, each tile 42 in the memory die 40 operates as a semi-autonomous mini-array of memory cells in the memory die 40. The intelligent CuA support circuitry allows each tile 42 to operate semi-autonomously, independent of other tiles in the memory die 40. Each tile 42 has its own associated CuA, which allows for simultaneous memory access to a large number of storage transistors in the memory die 40, resulting in increased memory throughput and reduced latency. In some embodiments, two tiles adjacent to each other can share certain support circuitry in the CuA. For example, a tile can include a set of sense amplifiers that is shared with its adjacent tile. When selected for access, each tile uses its own set of sense amplifiers and the set of sense amplifiers of the tile it is adjacent to. In this case, the tile whose set of sense amplifiers is borrowed is marked as inaccessible until the operation of the borrowed sense amplifiers is completed.

[0039] In the embodiment shown in Figure 3, memory die 40 is illustrated as including tiles arranged in 8 rows and 8 columns. The embodiment shown in Figure 3 is illustrative only and is not intended to be limiting. In one example, a memory die may include 1024 tiles arranged in 32 rows and 32 columns, or 2048 tiles arranged in 64 rows and 32 columns. The number of tiles and the arrangement of the tiles within memory die 40 may be selected based on various design factors, such as the size and dimensions of the memory die or the arrangement of interconnect structures.

[0040] In an embodiment of the present disclosure, the memory die 40 is divided into multiple partitions to form separate memory channels by dividing a two-dimensional array of tiles. In this embodiment, the memory die 40 is divided into four partitions by tile columns to form four memory channels Ch0-Ch3. In this example, each memory channel (Ch0-Ch3) includes two columns of eight tiles, for a total of 16 tiles per memory channel. In another example, in a memory die having 1024 tiles, the memory die can be divided into eight partitions by tile columns to form eight memory channels, with each memory channel including four columns of 32 tiles, for a total of 128 tiles per memory channel.

[0041] In an embodiment of the present disclosure, the memory die 40 includes a data interface region 43 for forming interconnect structures 45 for connecting to a memory controller. In this embodiment, the data interface region is provided in the center of the memory die 40 and is located in the middle of two tile columns. Furthermore, to support memory channel configurations, the interconnect structures 45 are divided according to the channel configuration, with interconnect structures 45-0 to 45-3 being provided for memory channels Ch0 to Ch3. For example, the interconnect structures 45 are through silicon vias (TSVs), with each memory channel comprising a dedicated set of interconnect structures or TSVs for transferring data of that memory channel to and from a memory controller, and more specifically, to and from an associated channel controller in the memory controller. As shown in FIG. 3, the memory die 40 may include a set of additional interconnect structures 44 for providing power and ground connections to the memory array formed in the tile 42. In some embodiments, as shown in FIG. 2B, multiple memory dies 40 are used to form a memory stack, and the formed memory stack is stacked on a memory controller to form a memory module. In the memory stack, the memory channels are formed by tiles in the same channel partition across all memory dies in the memory stack.

[0042] FIG. 4A illustrates a memory structure 50 including a three-dimensional array of NOR type memory strings according to some embodiments. In an embodiment of the present disclosure, the memory structure 50 of FIG. 4A is used to form a three-dimensional array of storage transistors in the tiles of the memory die 40 of FIG. 3. Various methods for forming three-dimensional NOR type memory strings are described in the above-mentioned US Pat. No. 6,399,943. Referring to FIG. 4A, the memory structure 50 includes thin-film storage transistors 60 formed as NOR type memory strings along a horizontal direction (Y direction) in multiple planes, each plane being formed by a set of active layers 56. Specifically, the memory structure 50 includes a number of active layers 56 formed on a plane of a semiconductor substrate 52. A buffer oxide layer 54 is provided between the semiconductor substrate 52 and the active layers 56. The active layers 56 are formed stacked on top of each other along the Z direction (i.e., a direction perpendicular to the plane of the semiconductor substrate 52) and are separated from each other by an insulating dielectric layer 55, such as a silicon carbide oxide (SiOC) layer. The active layer 56 is divided into narrow strips in the X-direction ("active strips") 57. The active strips 57 are stacked together to form a stack of active strips ("active stack") extending in the Y-direction.

[0043] Each active layer 56 includes a first doped semiconductor layer 62 and a second doped semiconductor layer 64 (e.g., n+ polysilicon or heavily doped n-type polysilicon) separated from each other by a dielectric layer 63 (e.g., silicon oxide). The first doped semiconductor layer 62 and the second doped semiconductor layer 64 form the drain and source regions of the storage transistor 60. Each active layer 56 includes one or more conductive layers 61, 65 (e.g., tungsten (W) lined with titanium nitride (TiN)). Each conductive layer 61, 65 is formed adjacent to one of the doped semiconductor layers 62, 64 to reduce the resistivity of the doped semiconductor layer that contacts it. During intermediate processing steps, the active layer includes a sacrificial layer (e.g., silicon nitride or carbon) that is later replaced with a conductive layer. Subsequent processing steps form a channel region 66 (e.g., p-polysilicon or lightly doped p-type polysilicon), a charge storage film 67, and a gate conductor or gate electrode 68 (e.g., W lined with TiN) in the narrow trench between the isolated active stacks. The gate electrode 68 and the charge storage film 67 are formed as a columnar structure extending in the Z direction. In this embodiment, the charge storage film 67 surrounds the gate electrode 68 to form a columnar structure. In this specification, the gate electrode 68 is also referred to as a "local word line," and the gate electrode 68 and the surrounding charge storage film 67 are collectively referred to as a local word line (LWL) structure 58.

[0044] In the embodiment shown in FIG. 4A, the storage transistor in the three-dimensional memory array is a charge trapping type storage transistor, and the charge storage film 67 includes a tunnel dielectric layer, a charge trapping layer, and a blocking layer. The tunnel dielectric layer can be any silicon oxide (SiOx), silicon nitride (SixNy), silicon oxide nitride (SiON), any aluminum oxide (AlOx), any hafnium oxide (HfOx), zirconium oxide (ZrOx), any hafnium silicon oxide (HfSixOy), any hafnium zirconium oxide (HfZrO), or any combination thereof. The charge trapping layer can be multi-layered and can include any silicon nitride (SixNy), hafnium oxide (HfO2), or hafnium silicon nitride (HfSiON). The blocking layer can be any silicon oxide (SiOx), aluminum oxide, or both. In one example, the charge storage film 67 is also referred to as an ONOA file, and includes a silicon oxide layer (SiOx), a silicon nitride layer (SiN), a silicon oxide layer (SiOx), and an aluminum oxide layer (Al2O3), which are stacked in this order from the channel layer side toward the gate conductor layer.

[0045] The first and second doped semiconductor layers of each active strip form a source region 64 ("common source line") and a drain region 62 ("common bit line") of a storage transistor formed along one or both sides of the active strip 57. Specifically, a storage transistor 60 is formed at the junction of the active strip 57 with a channel region 66 and an LWL structure 58. In the example of FIG. 4A, the LWL structures 58 are staggered (staggered) in trenches adjacent to the active strip 57 such that the storage transistors formed on either side of the active strip 57 are offset from each other in the Y direction along the memory string. When the storage transistor is a ferroelectric storage transistor, as will be described in more detail below, the first and second doped semiconductor layers are omitted and the conductive layers 61, 65 function as source and drain terminals.

[0046] With this configuration, the storage transistor 60 is composed of a first doped semiconductor layer 62 forming a drain region (common bit line), a second doped semiconductor layer 64 forming a source region (common source line), a channel region 66 in contact with both the drain region 62 and the source region 64, a gate electrode 68, and a charge storage film 67 located between the gate electrode 68 and the channel region 66. Each storage transistor 60 is insulated from adjacent storage transistors along the active stack (in the Z direction) by an insulating dielectric layer 55. With this configuration, the storage transistors sharing a common source line and a common bit line along each active strip (in the Y direction) form a NOR type memory string (also referred to herein as a "horizontal NOR type memory string" or "HNOR type memory string").

[0047] In another embodiment, the storage transistors in the three-dimensional memory array are junction-less ferroelectric storage transistors. In some implementations, the ferroelectric storage transistors are constructed in a similar manner to the charge trapping storage transistors by replacing the charge storage film with a ferroelectric gate dielectric layer incorporating a ferroelectric material. FIG. 4B illustrates a memory structure including a three-dimensional array of NOR memory strings, according to some embodiments. More specifically, FIG. 4B illustrates an exemplary three-dimensional memory array of junction-less ferroelectric storage transistors. Similar elements in FIG. 4A and FIG. 4B are labeled with similar reference numerals and will not be described. Referring to FIG. 4B, the memory structure 50B includes junction-less thin-film ferroelectric storage transistors 60B formed as NOR memory strings along the horizontal direction (Y direction) in multiple planes, each plane being formed by a set of active layers 56. Each active layer 56 includes a first conductive layer 61 and a second conductive layer 65 (e.g., tungsten (W) lined with titanium nitride (TiN)) separated from each other by a dielectric layer 63 (e.g., silicon oxide). The first conductive layer 61 and the second conductive layer 65 of the ferroelectric storage transistor function as drain and source terminals. Subsequent processing steps form a channel region 66B (e.g., an oxide semiconductor material such as IGZO), a ferroelectric dielectric layer 67B, and a gate conductor or gate electrode 68 (e.g., W lined with TiN) in the narrow trench between the isolated active stacks. The gate electrode 68 and the ferroelectric dielectric layer 67B are formed as columnar structures extending in the Z direction. The gate electrode 68 is also referred to herein as a "local word line," and the gate electrode 68 and the surrounding ferroelectric dielectric layer 67B are collectively referred to as the local word line structure 58.

[0048] In the embodiment shown in FIG. 4B, the storage transistors in the three-dimensional memory array are junctionless ferroelectric storage transistors. In some implementations, the ferroelectric gate dielectric layer 67B is a doped hafnium oxide (HfO2) layer. In one example, the hafnium oxide is doped with zirconium oxide (ZrO2) to form a hafnium zirconium oxide layer (HZO). In another example, the hafnium oxide is doped with silicon (Si), iridium (Ir), and lanthanum (La). In some embodiments, the gate dielectric layer may further include an interfacial layer between the channel region and the gate dielectric layer, such as a material having a high dielectric constant.

[0049] The first and second conductive layers of each active strip form the source line ("common source line") and drain line ("common bit line") of a storage transistor formed along one or both sides of the active strip 57. In the example shown in FIG. 4B, the storage transistor is formed only on one side of the active strip 57, while the other side of the active strip is adjacent to an auxiliary trench 59. The auxiliary trench 59 does not include any active transistor element. A ferroelectric storage transistor 60B is formed at the junction of the active strip 57 with the channel region 66B and the LWL structure 58. In this configuration, the storage transistors sharing a common source line and a common bit line along each active strip (in the Y direction) form a NOR type memory string or a HNOR type memory string. In some examples, a three-dimensional array of NOR type memory strings of junctionless thin-film ferroelectric transistors is disclosed in U.S. Pat. No. 6,399,323, the disclosure of which is incorporated herein by reference in its entirety. The memory structure 50B of FIG. 4B can be configured based on the memory structure described in U.S. Pat. No. 6,399,323.

[0050] 4A and 4B, to complete the memory circuit, various types of support circuits that support the operation of the NOR memory string are formed in or on the semiconductor substrate 52. As described above, the support circuits for the storage transistors in the tiles are formed locally under each tile and are referred to as "circuit under array" or CuA. The circuits in the CuA may include power supplies, sense amplifiers, data latches, logic circuits, and analog circuits, as described above. In an embodiment of the present disclosure, the CuA includes a state machine or sequencer to control and execute memory operations on the storage transistors in each tile. By incorporating a state machine in each CuA, multiple tiles in a memory die can be accessed for simultaneous memory operations.

[0051] In one exemplary embodiment, each tile in the memory die includes 8 active layers, i.e., 8 layers of storage transistors, each active layer includes 2k bitlines (or 2k active stacks) with storage transistors formed on either side of the bitlines, and each tile includes 4k wordlines (gate electrodes), for a total of 8M bits per layer, and 64M bits or 64M storage transistors in a tile.

[0052] In some embodiments, memory operations on a memory die are performed in units of memory pages of memory data, also referred to as access units of memory data. Each memory access, in this embodiment, is to one memory page in a memory bank (or tile). More specifically, the bit lines of storage transistors in a memory page are selected simultaneously, and the storage transistors are simultaneously identified to provide read data or are driven in the same write operation to store write data. In one example, a memory page of memory data is 512 bits. Thus, each time a read or write operation is performed, 512 storage transistors are accessed in a tile. In one embodiment, each memory access activates one word line (WL) associated with the 32 memory pages, and one of the 32 memory pages is selected for the memory operation by selecting the bit line associated with the storage transistor in the selected memory page.

[0053] In embodiments of the present disclosure, the support circuitry (CuA) of each tile includes as many sense amplifiers as there are data bits in the memory page. Thus, in each read operation, all sense amplifiers are used to read the stored data from the selected bit line, so no additional address bits are required to select a subset of the sense amplifiers. Similarly, in a write operation, all sense amplifiers are used to drive the write data to the selected bit line, so no additional address bits are required to select the sense amplifiers. This is in contrast to conventional memory devices, where one or more column address bits are required to select a subset of the sense amplifiers to provide the read data. By providing as many sense amplifiers in the CuA as there are data bits in the memory page, the size of the CuA can be kept small, which allows the CuA to be formed under each tile. Also, memory operations are simplified because no address bits are required to select a subset of the sense amplifiers. In some embodiments, the support circuitry may include additional sense amplifiers to identify or drive additional bit lines associated with other data stored in the memory array, such as refresh pointer data, metadata, or memory health indicator bits.

[0054] In one embodiment, the interconnect structures 45 (FIG. 3) for each memory channel include connection structures for at least as many data bits in a memory page. In one example, each set of interconnect structures 45 includes 300 through silicon vias (TSVs) for each memory channel to output 512 data bits in a memory page over two clock cycles (256 bits each clock cycle), plus additional data bits for error correction and control signals.

[0055] In this description, the memory structure 50 of FIG. 4A includes storage transistors implementing charge trapping as a data storage mechanism, and the memory structure 50B of FIG. 4B includes storage transistors implementing ferroelectricity as a data storage mechanism. The exact nature of the data storage mechanism implemented in the memory structure of the present invention is not critical to the practice of the present invention. As will be appreciated by those skilled in the art, the bias conditions applied to the storage transistors for read and write operations are a function of the data storage mechanism implemented. For purposes of illustration, bias conditions for ferroelectric storage transistors are used in the following description.

[0056] FIG. 5 is a circuit diagram illustrating a memory array of NOR type memory strings according to an embodiment of the present disclosure. FIG. 5 illustrates a memory circuit of NOR type memory strings that can be formed by the memory structure of FIG. 4A or FIG. 4B. Referring to FIG. 5, a memory array 70 of storage transistors 72 is illustrated, which represents a portion of the storage transistors in the three-dimensional memory array of FIG. 4A or FIG. 4B. The memory array 70 includes a plurality of memory strings 75 formed on each active layer, each memory string 75 including a series of storage transistors 72 connected in parallel to each other between a common bit line 74 and a common source line 76. Because the storage transistors are connected in parallel to each other in a NOR configuration, the memory strings 75 are also referred to as NOR type memory strings. The NOR type memory strings 75 of storage transistors form the basic building blocks from which a two-dimensional array or a three-dimensional array of storage transistors can be formed. That is, multiple strings of storage transistors can be used to form a two-dimensional array of storage transistors, or a plane of storage transistors. Also, a three-dimensional array of storage transistors can be formed by stacking multiple planes of two-dimensional arrays of storage transistors. In this description, the semiconductor memory device is implemented with an array or multiple arrays of strings of storage transistors, and the exact configuration or arrangement of the strings of storage transistors is not critical to the practice of the invention.

[0057] The storage transistors 72 are thin-film storage transistors having drain terminals connected to bit lines 74, source terminals connected to source lines 76, gate or control terminals connected to word lines 78, and a data storage film for storing data of the storage transistors. For example, the data storage film can be a set of charge storage films or a ferroelectric dielectric layer. More specifically, the gate terminals of the storage transistors 72 are driven by word lines (WLx) 78, and each word line WLx activates one storage transistor 72 in a NOR memory string 75 while simultaneously activating other storage transistors in other NOR memory strings. With this configuration, when a word line WL is selected, all storage transistors 72 connected to that word line (e.g., WLn) are activated. In operation, a selected word line activates P memory pages. Each memory page includes Q storage transistors associated with Q NOR memory strings. A bit line belonging to the selected memory page is selected for a memory operation. Thus, a memory page is selected by the selected word line and the selected bit line, thereby allowing access to the Q storage transistors in the selected memory page.

[0058] For example, the bit lines of each memory string are connected to a sense amplifier circuit for sensing the stored data in a memory read operation. The sense amplifier circuit and other circuit elements, as well as control signals for facilitating the operation of the sense amplifier and memory array, are not shown in FIG. 5. For example, the bit lines may include discharge transistors for discharging the bit lines after a read or write operation. In another example, the sense amplifier circuit may include transistors and devices for implementing a reset function for the sense amplifier. Additionally, the sense amplifier circuit may include a latch circuit for latching the output of the sense amplifier. Details of an exemplary sense amplifier circuit are described below with reference to FIG. 6.

[0059] In an embodiment of the present disclosure, a memory device includes storage transistors (or "memory cells") that can be read, programmed, or erased. Program and erase operations are collectively referred to as write operations. A memory device performs memory operations including a read operation to read data from a storage transistor and a write operation to write data to a storage transistor. A memory device may perform other operations, such as a refresh operation, which are not described in this description. In this description, a write operation includes two operations or two phases: an erase operation or an erase phase, and a program operation or a program phase. In this embodiment, an erase operation is associated with writing a first logic state (e.g., logic "1") to a memory cell, and a program operation is associated with writing a second logic state (e.g., logic "0") to a memory cell. It should be noted that the particular logic states assigned to an erase operation or a program operation are arbitrary and not critical to the practice of the present invention. In other embodiments, an erase step is associated with writing a logic "0" to a memory cell, and a program step is associated with writing a logic "1" to a memory cell. In this embodiment, the erase operation is also referred to as a Set 1 operation, and the program operation is also referred to as a Set 0 operation.

[0060] In the memory array 70, each storage transistor of a NOR memory string is read, programmed, or erased by appropriately biasing its associated word line 78 (WLx) and a common bit line 74 (BLy) that it shares with other storage transistors in the NOR memory string 75. The associated word line of a storage transistor is shared with a storage transistor of a NOR memory string on another plane that is aligned with the storage transistor along a direction perpendicular to the plane of the semiconductor substrate (the "orthogonal direction"). Also, each word line is shared between two storage transistors of adjacent NOR memory strings on the same plane (see Figures 4A and 4B). In some embodiments, the common source line is typically electrically floating. That is, the common source line is not connected to any potential. During a read, program, or erase operation, the common source line of the NOR memory string is typically supplied with a relatively constant voltage maintained by a voltage source or a charge in an associated capacitor ("virtual ground"), such as the parasitic capacitance of the common source line. For example, the common source line of a NOR memory string can be biased to a given voltage by a precharge operation in which a desired voltage is provided on the common bit line, and the common source line is charged to the voltage on the bit line through one or more precharge transistors. To program or erase the storage transistors, for example, a significant voltage difference (e.g., 8V for charge storage type storage transistors, 3V for ferroelectric storage transistors) is imposed between the common bit line and the word line. To mitigate disturbance to unselected storage transistors, undesired erasure or programming of unselected storage transistors can be suppressed by imposing a predetermined voltage difference between the associated word line of the unselected storage transistors and its common bit line that is significantly lower than the voltage required for programming or erasing.

[0061] FIG. 6 is a circuit diagram illustrating an exemplary support circuit connected to the bit lines of a NOR type memory string according to an embodiment of the present disclosure. Specifically, FIG. 6 illustrates sense amplifier circuitry and associated circuit elements for performing read and write memory operations. For ease of illustration, additional circuit elements and control signals are omitted. With reference to FIG. 6, each bit line 74 is connected to a sense amplifier 80 via a bit line selector (not shown). In practice, each bit line selector is connected to P bit lines and selects one of the P bit lines for sensing by an associated sense amplifier. In other words, each bit line selector is connected to the bit lines of the same data bit across P memory pages. If an access unit includes 512 bits of memory data, 512 bit line selectors are provided to select the bit lines of the selected memory page for access. With such an arrangement, a selected word line activates P memory pages (e.g., 32 memory pages), and a set of bit line selectors in each sense amplifier selects the bit lines associated with the selected memory page for access. FIG. 6 illustrates a selected bit line 74 connected to a sense amplifier 80. In FIG. 6, for simplicity, the bit line selectors and other bit lines that share the same sense amplifiers are omitted.

[0062] In a read operation, the sense amplifier 80 senses a voltage signal indicative of the bit line current on the selected bit line to determine the logic state of the selected storage transistor and generates a sense amplifier output SAOUT (node ​​82) in response to the sensing. In this embodiment, the sense amplifier output SAOUT (node ​​82) is connected to a pair of data latches 84, 86. In this embodiment, a first data latch 84 (DL1) can be used to store data locally in the memory tile, such as read data for a refresh operation. Also, a second data latch 86 (DL2) can be used to store data to be exchanged with the memory controller, such as read data read from the selected storage transistor or write data received from the memory controller. It should be noted that the sense amplifier circuit configuration including two data latches 84, 86 is illustrative only and is not intended to be limiting. Other sense amplifier circuit configurations are possible. In one embodiment, the sense amplifier 80 may be configured to incorporate a data latch function itself, such that the sense amplifier itself functions as a data latch to store the read data. In this case, the support circuitry of each tile includes three data latches: a first data latch DL1, a second data latch DL2, and a third data latch DL3 as a sense amplifier. In this case, the third data latch DL3 (sense amplifier) ​​is used to store read data read from a selected storage transistor, and the second data latch DL2 is used to store write data received from the memory controller.

[0063] In a read operation, the sense amplifier 80 senses a voltage signal on the selected bit line 74 indicative of a bit line current associated with the erased or programmed state of the selected storage transistor. The sense amplifier 80 generates a sense amplifier output signal SAOUT having a logic state indicative of the sensed bit line voltage signal. In one embodiment, in a read operation, the read data is stored in the data latch DL2. The read data is sent to the data bus 90 by the data driver 88 and provided to the memory controller. In practice, the data driver 88 is controlled by a clock signal and can send the read data to the data bus 90 synchronously with the clock signal.

[0064] In a write operation, write data from the memory controller is sent to the data bus 90, and the write driver 92 sends the write data to the data latch DL2. The write driver 92 is also controlled by a clock signal and can send the write data to the data latch DL2 in response to the clock signal. To perform an erase or program operation, a bit line (BL) bias control circuit 94, under the control of a state machine in CuA, applies a program voltage or an erase voltage to the bit line 74 depending on the logic state of the write data to be written to the storage transistor.

[0065] Memory Controller Architecture

[0066] FIG. 7 is a schematic diagram of a memory controller according to an embodiment of the present disclosure. In some examples, the memory controller 100 of FIG. 7 can be used to implement the memory controller 14 of the memory module 30 of FIG. 2A. Specifically, the memory controller 100 is configured to operate a memory device configured with multiple memory channels that are accessed independently of one another. With reference to FIG. 7, the memory controller 100 includes a host interface circuit 106 for interfacing with a host, a memory control circuit 110 for interfacing with the memory device 101, and a processor 108 for controlling the operation of the host interface circuit 106 and the memory control circuit 110. The memory controller 100 operates based on one or more clock signals. For example, the host interface circuit 106 can use a clock signal with a first clock frequency, and the memory control circuit 110 and the processor 108 can use a clock signal with a second clock frequency that is different from the first clock frequency. Although the clock signals for operating the memory controller are not shown in FIG. 7 for ease of explanation, it should be understood that the memory controller 100 operates based on one or more clock signals.

[0067] The host interface circuit 106 is connected to a host interface bus 102 for communicating with a host, such as a host processor. The memory controller 100 receives requests from the host processor and sends responses to the host processor via the host interface circuit 106. For example, the memory controller 100 receives read requests and write requests with write data from the host via the host interface bus 102. The memory controller 100 provides read data and write completion responses to the host via the host interface bus 102. In an exemplary embodiment, the host interface circuit 106 communicates with the host via a PCIe 5.0 serial bus using a Compute Express Link™ (CXL) protocol. In the CXL protocol, the host processor issues a request with no data (REQ) as a read request and a request with data (RwD) as a write request with write data. Also, in the CXL protocol, the memory controller 100 issues a response with data (DRS) as read data and a response with no data (NDR) as a write completion response.

[0068] The memory control circuitry 110 is configured to operate on the memory channels of the memory device 101 and communicate with the memory device via the memory array interface 103. In an embodiment of the present disclosure, the memory device 101 is a memory stack including multiple memory dies stacked on top of each other. In this example, the memory stack 101 includes four memory dies, namely, die 0, die 1, die 2, and die 3. The memory dies in the memory stack 101 are divided into N memory channels that can be accessed independently of each other according to the memory channel configuration scheme described above. In this example, the four memory dies, namely, die 0, die 1, die 2, and die 3, are divided into four memory channels, namely, Ch0, Ch1, Ch2, and Ch3. In other examples, the memory dies may be divided into four channels or 16 channels. Under the memory channel configuration scheme in the exemplary embodiment of the present disclosure, each memory channel Chn is formed across all semiconductor dies in the memory stack 101. That is, each memory channel Ch0-Ch3 includes memory partitions from memory dies die 0-die 3. The memory control circuit 110 communicates with the memory stack 101 via a memory array interface 103 that includes individual memory channel interfaces 104-0 to 104-3 for each memory channel Ch0 to Ch3. Specifically, each memory channel interface 104-n includes a data interface and a command interface for each memory channel.

[0069] The configuration of the memory device 101 in FIG. 7 is illustrative only and is not intended to be limiting. The memory control circuitry 110 is configured to operate a memory device of any configuration, such as a single semiconductor memory die or multiple semiconductor dies. The memory control circuitry 110 is configured to operate on the memory channels of the memory device, where the memory channels can be variously configured. The exact channel configuration of the memory device is not important to the practice of the invention. In FIG. 7, the memory device is a memory stack, and the channel configuration spans the memory dies in the stack. In another example, the memory device includes multiple memory dies (e.g., k memory dies) arranged in a planar fashion adjacent to a memory controller, and each memory die includes multiple memory channels (e.g., n memory channels). The memory control circuitry 110 of the memory controller 100 can be configured to operate k×n memory channels across the k memory dies.

[0070] The memory control circuit 110 includes a channel controller 116 that is instantiated N times to provide one channel controller 116 for each memory channel. In this embodiment, four instances of the four channel controllers 116, including channel controller 0 to channel controller 3, are provided for the four memory channels of the memory stack 101. Each channel controller 116 communicates with a memory channel via a respective memory channel interface 104-n. In this manner, each memory channel Ch0 to Ch3 of the memory stack 101 can be accessed independently of one another, and high bandwidth memory access can be achieved by accessing the memory channels Ch0 to Ch3 in parallel.

[0071] The memory control circuit 110 further includes an address translation circuit 114 and a channel arbiter 112, both of which communicate with all the channel controllers 116. The host interface 106 receives input requests from the host, such as a read request (REQ) or a write request (RwD), which are provided to the address translation circuit 114 of the memory control circuit 110. The address translation circuit 114 decodes the logical address in each request to determine which memory channel the request should be sent to, and sends the request to the channel controller 116 of the specified memory channel. Upon completion of the memory operation, each channel controller 116 provides a response, such as read data or a write completion response, to the channel arbiter 112. The channel arbiter 112, under the control of the channel arbitration logic, selects which response to send to the host interface 106. For example, the channel arbitration logic returns the responses in the same order as the corresponding requests received. In another example, the channel arbitration logic implements a credit management scheme that determines which responses from which memory channels are returned to the host processor. In some examples, the channel arbitration logic is implemented within the processor 108.

[0072] FIG. 8 is a schematic diagram of a channel controller implemented in the memory controller of FIG. 7 according to an embodiment of the present disclosure. As described above with reference to FIG. 7, the memory controller 100 receives an input request on the host interface bus 102, and the received request is provided to the address translation circuit 114 of the memory control circuit 110. The address translation circuit 114 decodes the logical address in the request to generate a physical address of the memory device 101 corresponding to the logical address. The physical address is also referred to as a memory address in this specification. Based on the channel address indicated in the decoded physical address, the request is provided to each channel control unit 116. In some embodiments, the address bits for selecting the memory channel are the lower address bits of the physical address. In this manner, the input requests are distributed to different memory channels, reducing channel contention and, as a result, improving memory utilization. As a result of the address translation, each channel controller 116 receives an input request for a read or write operation designated for that memory channel.

[0073] It should be noted that the operation of the channel controller 116 is based on one or more clock signals. Typically, the channel controller 116 operates based on a controller clock signal at a given clock frequency. Signals sent to and from the channel controller may cross over different clock domains (i.e., different clock frequencies). In that case, a buffer or clock crossing FIFO circuit may be used for signals crossing over at two different clock frequencies. For simplicity of explanation herein, the clock signals and associated clock crossing circuit are not shown in FIG. 8. It should be understood that the operation of the channel controller is synchronous to the controller clock signal. For example, the channel controller 116 sends a command to the memory device 101 at each clock cycle of the controller clock signal. In one example, the controller clock signal has a frequency of 500 MHz, and the channel controller sends a command to the memory device for which there is an available command every clock cycle or every two clock cycles, i.e., every 2 ns or every 4 ns.

[0074] In some embodiments, the address translation circuit 114 decodes the logical address in the request into a physical address including a memory channel address, a memory bank address, a memory die address, a word line address, and a memory page address. The decoded physical address thus selects a memory channel from the N memory channels and selects a memory die from the K memory dies in the selected memory channel. The physical address further selects a memory bank (or memory tile) in the selected memory die of the selected memory channel. In the selected memory bank, the physical address selects a word line, which activates P memory pages in the memory tile. The physical address ultimately selects one memory page from the P memory pages associated with the selected word line. As described above, each memory page includes Q data bits, for example, 512 bits.

[0075] 8, the channel controller 116 receives incoming read or write requests from the address translation circuitry 114 and stores the requests in respective ingress buffers. Specifically, incoming read requests are stored in the ingress read buffer 120 and incoming write requests are stored in the ingress write buffer 122. The incoming write requests are then provided to and stored in the write staging buffer 128 as pending write requests. The ingress channel arbiter 124 regulates the flow of pending read and write requests to the read queue 130 and the write queue 132 based on one or more predefined priority rules provided by the read / write arbitration logic circuitry 125. Specifically, the ingress channel arbiter 124 arbitrates between pending read requests from the read ingress buffer 120 and pending write requests evicted from the write staging buffer 128 based on the predefined priority rules to select requests to send to the read queue 130 and the write queue 132. For example, the ingress channel arbiter 124 may prioritize read requests as long as the write staging buffers are not nearly full for one or more memory banks. The ingress channel arbiter 124 may further consider attributes such as the age of the request and the state of the memory bank in determining the priority. In another example, the ingress channel arbiter 124 may forward read requests to a read queue or write requests to a write queue in a manner that avoids excessive bank contention.

[0076] The read queue 130 and the write queue 132 store read and write requests and generate commands for the memory device 101 based on the stored requests. The read queue 130 and the write queue 132 generate commands to bid for access to the memory device 101 through a command selector 135, also referred to as a global arbitration circuit or global scheduler. The command selector 135 selects a command to be sent to an associated memory channel of the memory device 101 to perform a memory operation at a destination memory address associated with each request. The command selected by the command selector 135 is also referred to as a "winning command."

[0077] In an embodiment of the present disclosure, the channel controller 116 of FIG. 8 is adapted to a memory device in which the write latency is much longer than the read latency. A salient feature of the channel controller of the present invention is that the channel controller is configured to mask the potentially long write latency from the host. In some embodiments, the channel controller 116 stores the incoming write request and write data in the write staging buffer 128 and returns a write completion response to the host before the write operation is actually executed in the memory device, i.e., before the write data is stored in the memory device. At the same time, the channel controller 116 manages the write request and performs the write operation to the memory device in the background of the host operation, so that the extended write latency at the memory device is hidden from the host system, and the host system can operate as if it only had the nominal write latency at the memory device.

[0078] In some embodiments, in the channel controller 116, each write request stored in the ingress write buffer 122 is first flushed to the write staging buffer 128, and then the pending write request entries in the write staging buffer 128 are provided to the ingress channel arbiter 124 for arbitration with read requests from the ingress read buffer 120. In one example, the write staging buffer 128 can be used to prevent too many write requests to the same memory bank of the memory channel from being sent to the write queue 132, filling the write queue and causing backpressure on the ingress write buffer 122. In some embodiments, the write staging buffer 128 is sized to store a large number of write requests sufficient to handle the write latency of the memory device that is masked from the host. In this embodiment, the flushing of the write requests from the ingress write buffer 122 to the write staging buffer 128 notifies the host of a write completion response. That is, each time a write request is evicted from ingress write buffer 122 and forwarded to write staging buffer 128, a write completion response for that write request is sent to the host. From the host's perspective, that particular write request is completed while channel controller 116 maintains and processes write requests to be completed in memory device 101. In this embodiment, the write completion response (e.g., NDR under the CXL protocol) for each memory channel n is provided to channel arbiter 112b, which receives the write completion responses from all memory channels (e.g., Chn, x to z). Channel arbiter 112b selects the write completion response from the memory channels that are provided to the host via the host interface based on predetermined arbitration rules.

[0079] As described above, the ingress channel arbiter 124 arbitrates between read requests from the ingress read buffer 120 and write requests from the write staging buffer 128 based on predefined priority rules to select requests to send to the read and write queues. In another example, the ingress channel arbiter 124 forwards read requests to the read queue or write requests to the write queue to avoid excessive bank contention. For example, if there is a pending write request for a given memory bank of a given memory die in the write queue, the ingress channel arbiter 124 does not forward additional write requests for the same memory bank of the same memory die to the write queue, but instead forwards write requests destined for other memory banks to the write queue. In this way, the ingress channel arbiter 124 avoids sending an excessive number of write requests for the same memory bank of the same memory die to the write queue 132, which would cause the write queue to fill up if the memory device has a long write latency.

[0080] In an embodiment of the present disclosure, the ingress channel arbiter 124 further evaluates the incoming read request to determine whether the read request is for the same memory address as a pending write request in the write staging buffer 128. In an embodiment of the present disclosure, the ingress read buffer 120 transfers the memory address of the pending read request ("read address") to the write staging buffer 128. The write staging buffer 128 determines whether the read request is for the same memory address as a pending write request in the write staging buffer 128. If the write staging buffer 128 determines that the read request has a memory address that matches the destination memory address of the pending write request in the write staging buffer, the channel controller 116 implements various techniques to provide the write data of the matching pending write request as the read data of the read request. In one example, the read request is satisfied by the write data from the write staging buffer 128. In another example, the channel controller 116 forwards matching pending write requests to the write queue 132, and read requests are satisfied with write data from the write queue 132 (more specifically, the write data storage 134 associated with the write queue).

[0081] In this embodiment, the read queue 130 is a data structure that stores pending read requests for an associated memory channel. The read queue 130 stores pending read requests along with a destination memory address, such as a memory bank address or a memory page address, for the read request. The read queue 130 further stores other attributes associated with each read request, such as the state (status) of the memory bank indicated by the destination memory address and the command being processed in the memory bank. The state of the memory bank ("bank state") is used to indicate whether the read request is eligible to be sent to the memory device. In this embodiment, each entry in the read queue 130 can request the sending of an activate command, a read command, or an abort command. The activate command instructs the memory device to start sensing data stored in a specified memory page of a specified memory bank using a sense amplifier circuit and to store the sensed data in a data latch (e.g., data latch DL2 of FIG. 6) connected to the sense amplifier circuit. The read command outputs the sensed data stored in the data latch to the channel controller 116. For example, data read from the memory device 100 is stored in the read data storage 142. The interrupt command will be described in detail later.

[0082] In this embodiment, write queue 132 is a data structure that stores active write requests for an associated memory channel. Write queue 132 stores active write requests along with a destination memory address (e.g., memory bank address, memory page address) for each write request. In this embodiment, write data for each write request is stored in write data store 134, which communicates with write queue 132. Write queue 132 further stores other attributes associated with each write request, such as the state of the memory bank indicated by the destination memory address and the command being processed in the memory bank for each write request. The memory bank state ("bank state") is used to indicate whether the write request is eligible to be sent to the memory device. In this embodiment, each entry in write queue 131 can request the sending of a write command. The write command is sent to the memory device along with associated write data stored in write data storage 134, instructing the memory device to write data to a specified memory page of a specified memory bank. In some embodiments, the write data is first stored in a data latch (eg, data latch DL2 of FIG. 6), and the bit line bias control circuitry passes the write data to the storage transistor.

[0083] In this embodiment, the channel controller 116 includes an issued write queue 133. The write queue 132 allocates an active write request entry to the issued write queue 133 when the active write request is a winning command in the command selector 135. Specifically, the write queue 132 transmits a write command together with write data to the memory device 101 in response to the active write request selected by the command selector 135. For example, the write data is sent to the memory device 101 and stored in a data latch (e.g., data latch DL2) of the specified memory page. The write queue 132 then transfers the write request to the issued write queue 133. The issued write queue 133 processes the remaining write operation flow. For example, each entry in the issued write queue 133 can transmit a commit command to instruct the memory device 101 to initiate a bias voltage sequence for writing the previously transmitted write data to the storage transistors of the destination memory address. For example, a commit command is sent to activate a program, erase, or refresh operation flow in the support circuitry (CuA) of the specified memory tile, which applies appropriate voltages to the bit lines of the storage transistors associated with the selected memory page. The issued write queue 133 deallocates the write request when the write flow ends, i.e., when the erase and program operations on the storage transistors are completed. Additionally, in this embodiment, each entry of the issued write queue 133 can also send a write resume command, which will be described in more detail below.

[0084] In some embodiments, issued write queue 133 stores issued write requests until they are completed. Note that the use of issued write queue 133 is optional and may be omitted in other embodiments. In other embodiments, issued write requests may be stored in write queue 132 itself until the write request is completed.

[0085] In this embodiment, an active read or write request in the respective read / write queue is eligible for bid access if its destination memory bank does not currently have a memory operation in progress. For example, a read or write request is eligible if its destination memory bank is not currently being read or written. Eligible read or write requests bid for access to memory device 101 via command selector 135. Command selector 135 arbitrates commands received from eligible read or write requests and determines a command (the "winning command") to send to memory device 101 every clock cycle (or every x clock cycles).

[0086] In an embodiment of the present disclosure, the command selector 135 selects commands to send to the memory device 101 based on predefined priority rules. In this embodiment, the command selector 135 sends commands to the respective memory channels of the memory device over two command buses: (1) a memory command bus 144 for commands that do not include data, such as activate commands, suspend commands, commit commands, and write resume commands, and (2) a data command bus 146 for commands that include data, such as read commands and write commands. In some embodiments, the command selector 135 includes a memory command selector 136 and a data command selector 138 for selecting from eligible commands provided by the read queue 130 and the write queue 132. The memory command selector 136 selects commands to be sent to the memory device on the memory command bus 144, and the data command selector 138 selects commands to be sent to the memory device on the data command bus 146. In this embodiment, the commands provided to the memory command selector 136 include, for example, an activate command to set up a read operation. The commands sent on the memory command bus 144 are not associated with read or write data. Meanwhile, commands provided to data command selector 138 may include, for example, a read command or a write command. A write command is provided on data command bus 146 synchronously with associated write data provided from write data storage 134 on data bus 148. In response to a read command, memory device 101 provides read data on data bus 148 and stores the read data in read data storage 142. The read data is provided as a read data response to the memory channel.

[0087] In this embodiment, a read data response (e.g., DRS) for each memory channel n is provided to a channel arbiter 112a, which receives read data responses from all memory channels (e.g., Chn, x-z). The channel arbiter 112a selects a read data response from a memory channel to provide to the host via the host interface based on a predefined arbitration rule.

[0088] In an embodiment of the present disclosure, the channel controller 116 is configured to operate to maximize the performance of the memory device. For example, in one embodiment, the channel controller is configured to operate in a greedy mode where it always attempts to send commands to the memory device for execution as long as there are pending requests. Thus, the read queue 130 and the write queue 132 always send eligible commands to the command selector 135 to bid for access to the memory device. Meanwhile, the command selector 135 operates based on a predefined priority rule and selects a winning command to send to the memory device every x clock cycles (x is 1 or more). In one example, the command selector 135 is configured in a read-first mode where commands for write operations are sent in the free clock cycles between commands for read operations. In another example, the command selector 135 is configured in a write-first mode where commands for read operations are sent in the free clock cycles between commands for write operations. In some embodiments, the predefined priority rule includes a fixed priority rule or a dynamic priority rule.

[0089] In operation, the read queue 130 and the write queue 132 store active read and write requests to the memory device, with each read / write request associated with a destination memory page in a destination memory bank of a destination memory die of the memory channel. The read queue 130 and the write queue 132 determine which of the pending requests stored therein are eligible to be executed by the memory device. The read queue 130 and the write queue 132 always attempt to send all eligible commands to the command selector 135, which arbitrates the commands using predefined priority rules and selects a winning command every clock cycle (or x clock cycles). The winning command is then sent to the memory device 101 (along with the write data on the data bus 148) via the respective memory command bus 144 or data command bus 146. To keep track of memory operations being executed on the memory device, the winning command is provided to a command tracker 140, which operates to track the progress of each command sent to the memory device. The winning command is also provided to a bank tracker 145, which tracks the state of each memory bank in the memory channel. Read queue 130 and write queue 132, as well as issued write queue 133, use the information in bank tracker 145 and command tracker 140 to determine the status of each memory bank that has a request pending and, based on the determined bank state, determine which commands are eligible to bid for access in command selector 135.

[0090] In some embodiments, bank tracker 145 is a data structure that stores bank status data for each memory bank in a memory channel, where the bank status indicates whether a given memory bank is inactive (i.e., eligible to receive commands) or busy (i.e., not eligible to receive commands). The bank tracker may further store data indicating memory operations being performed on busy memory banks. Read queue 130 and write queue 132 obtain the status of each memory bank associated with their respective pending requests from bank tracker 145.

[0091] In some embodiments, the command tracker 140 tracks each command being sent and executed in each memory bank of the memory device based on clock timing. In one embodiment, each command executed in the memory device 101 is assigned a predetermined command execution time. The command tracker 140 tracks commands issued for execution in the memory device and indicates a given command is completed upon expiration of the predetermined command execution time assigned to that command. In some examples, the progress of each command being executed is tracked using clock cycles. In one example, commands sent to the memory device 101 share the same time base, e.g., a command is sent every 4 ns. In some embodiments, the command tracker 140 is implemented as a shift register. In one example, the winning command is assigned to the top of the shift register and is shifted every clock cycle to track the progress of commands being executed in the memory device. In some embodiments, configurable tap points are provided in the shift register to indicate associated timing constraints. Commands progressing through the shift register are compared to each tap point. A match at a tap point indicates that time or clock cycles have elapsed from the command being issued to the given tap point. In this manner, the command tracker tracks the progress of each command issued to the memory device. A command that has progressed beyond the tap point associated with its allotted command execution time is marked as a completed command and the associated memory bank is marked as free.

[0092] In this manner, the read queue 130 and the write queue 132 determine the bank status and ongoing memory operations for each memory bank for which a request is pending. The bank status and memory operation progress information allows the read queue and the write queue to determine whether the pending request is associated with a busy memory bank or an inactive or empty memory bank. For pending requests associated with an inactive memory bank, the read queue and the write queue issue commands that are eligible to be bid for access. The memory command selector 136 and the data command selector 138 of the command selector 135 select a winning command from among the eligible commands every clock cycle and send it to the memory device 101. The command tracker 140 and the bank tracker 145 update their respective statuses every clock cycle in response to the winning command. In this manner, the winning command is identified for the read queue 130 and the write queue 132, and the bank status associated with the memory bank of the winning command is updated to busy. This makes pending requests to the same memory bank ineligible for bidding.

[0093] The structure and operation of the channel controller has been described herein to explain the process flow from receiving a request to generating a command for a memory device. It should be understood that the channel controller may include other circuit elements not shown or described to support memory operations. For example, the channel controller may perform error detection and correction. The channel controller may also include an ECC encoder for performing error correction encoding and an ECC decoder for detecting and correcting bit errors, thereby preventing data loss. The ECC circuitry and other support circuitry included in the channel controller have been omitted from FIG. 8 for ease of illustration.

[0094] In the above-described embodiment, the memory device is formed by a memory stack of K memory dies with memory channel partitions across the memory dies in the stack. In another embodiment of the present invention, the K memory dies of the memory device are arranged side-by-side on an interconnect structure, such as an interposer, and adjacent to a memory controller die formed on the same interconnect structure. Such a packaging structure is also referred to as 2.5-dimensional packaging. In some examples, when formed on an interposer, the interposer provides high density die-to-die connections between the K memory dies and the memory controller die. In some embodiments, each memory die is divided into S memory channels and forms S×K memory channels with the K memory dies. With such a configuration, the memory device provides a large number of memory channels, which has the beneficial effect of reducing access contention and increasing the availability of memory banks in the K memory dies.

[0095] In the above-described embodiments, the memory device includes a memory stack of K memory dies stacked together to provide the intended or designated memory capacity of the memory module. In embodiments of the present invention, the memory stack includes at least one additional spare memory die that is used to provide redundant memory capacity within the memory module. In some embodiments, the spare memory die is formed as part of the memory stack and is connected to the K memory dies via the same interconnect structures, such as TSVs, that connect the K memory dies.

[0096] Thus, in some embodiments, a memory module includes a memory device integrated with a memory controller, the memory device being formed by a memory stack including K+1 memory dies. With such a configuration, preferably, the K memory dies provide the intended or designated memory capacity of the memory module, and one additional memory die (the "spare memory die") provides additional memory capacity for redundancy. With such a configuration, the memory address space of the spare memory die is unknown to the host processor, and is known only to the memory controller. That is, the host processor is not aware of the physical presence of the spare memory die, nor of the memory address space associated with the spare memory die. The host processor is only aware of the memory address space of the K memory dies (the "host address space"). That is, requests sent from the host processor to the memory module are directed only to the memory address space across the K memory dies, which is known to the host processor. Meanwhile, the memory controller controls the physical memory address space across the K+1 memory dies, and manages the mapping of logical memory addresses to physical memory addresses in the host address space, including the mapping to the spare memory die when a redundancy replacement is applied. In other words, the memory controller communicates with the host processor over a memory address space spanning the K memory dies and manages memory operations in a memory address space spanning the K+1 memory dies. In some embodiments, the memory controller controls and directs access to a spare memory die via the address translation circuitry, for example, when the memory capacity of the spare memory die is applied to replace a non-functioning or failed memory tile within the known address space of the K memory dies.

[0097] In another embodiment, a memory module includes a memory device integrated with a memory controller, the memory device being formed by K+1 memory dies arranged side-by-side on an interconnect structure with the memory controller in a packaging structure also referred to as 2.5D packaging. In some examples, the interconnect structure is an interposer. In other examples, the K+1 memory dies are arranged side-by-side to surround the memory controller on the interconnect structure. Importantly, the K+1 memory dies of the memory device of the present invention have a variety of packaging structures, either stacked (3D) or 2.5D packaging. The particular packaging structure of the K+1 memory dies and thus the memory controller die is not critical to the practice of the present invention.

[0098] In some embodiments, the memory controller is configured to monitor the operational or health status of the memory arrays (or tiles) in each memory die. For example, the memory controller may obtain health information of the storage transistors while performing a refresh operation. The memory controller may also perform error correction on memory data read in a read operation and determine the health or operational status of the memory arrays and / or storage transistors therein. The memory controller may determine that one or more memory arrays are malfunctioning, failing, failing, or degraded. As used herein, a memory array (tile) is considered malfunctioning or failing when one or more storage transistors in the memory array are malfunctioning or failing, or other transistors or circuit elements in the memory array are malfunctioning or failing. For example, a malfunctioning or degraded storage transistor refers to a storage transistor that is still functional but has degraded electrical characteristics, such as a smaller than expected memory window, a smaller than expected transistor "on" current (Ion), or a significant deviation from the performance specifications of the storage transistor. In another example, a memory controller may detect that one or more memory arrays are malfunctioning or failing by detecting a significant deviation from a memory device's performance specifications.

[0099] In some embodiments, the memory controller is configured to determine whether to replace one or more malfunctioning or failed memory arrays (tiles). If the memory controller determines that a redundancy replacement should be applied to tiles in the K memory dies, the memory controller replaces the failed memory tiles with redundant memory tiles in the spare memory die. The memory controller manages address mapping information such that future requests received from the host processor that address the replaced memory tiles in the host address space are directed to the replacement memory tiles in the spare memory die.

[0100] For example, if during the operational life of the memory module, a tile of one of the K memory dies is detected to be failed or failing, the memory controller uses a functioning tile from the spare memory die to replace the suspect tile. In some embodiments, the redundancy replacement can be performed on individual tiles or groups of tiles. In some embodiments, the address translation circuitry of the memory controller manages the mapping of logical memory addresses in the host address space to physical memory addresses, such that requests input from the host processor to the replacement memory banks of the K memory dies are redirected to the replacement memory banks in the spare memory die. The tile replacement operation using the physical memory addresses of the spare memory die is completely transparent to the host processor and is managed internally within the memory module by the memory controller. The host processor can continue to operate using the logical memory addresses of the host address space. Meanwhile, the memory controller controls the physical memory address space of the K+1 memory dies and manages the mapping of logical memory addresses to physical memory addresses, including the mapping to the spare memory die when a redundancy replacement is applied. In this manner, the memory module continues to support its full specified memory capacity even if some tiles within the K memory dies or some storage transistors within a tile malfunction or fail. Additionally, the memory controller predicts the impending failure or degraded health of one or more tiles and initiates redundancy replacement, thereby improving the operational reliability of the memory module.

[0101] In some embodiments, the address translation circuitry of the memory controller (e.g., such as address translation circuitry 114 of FIG. 8 ) is configured to process incoming requests from a host processor and be received by the memory module, each request including a logical memory address in the host address space of the K semiconductor memory dies. Meanwhile, the address translation circuitry of the memory controller converts the logical memory addresses of the incoming requests into physical memory addresses that identify storage transistors in the K+1 semiconductor memory dies. Specifically, the address translation circuitry of the memory controller manages the mapping of logical memory addresses in the host address space that are targeted to the malfunctioning or failed memory array and replaced with physical memory addresses in the memory address space of the spare memory die. The host processor continues to send requests to the memory module using the logical memory addresses of the host address space of the K memory dies, while the memory controller, via the address translation circuitry, manages the mapping of the logical memory addresses to the physical memory addresses of the K+1 memory dies, including mapping the logical memory addresses to physical memory addresses on the spare memory die when a redundancy replacement is applied.

[0102] Memory Device Architecture

[0103] In embodiments of the present invention, the memory devices described herein implement a tile-based architecture that includes an arrangement of arrays or tiles of memory transistors that can operate independently and simultaneously, with each tile including memory transistors arranged in a three-dimensional array and localized modular control circuitry that operates the memory transistors in the tile. The tile-based architecture of the present invention allows for simultaneous memory access to multiple tiles in a memory device, thereby allowing independent and simultaneous memory operations to be performed across multiple tiles. The tile-based simultaneous access to a memory device has the advantage of increasing memory bandwidth and lowering tail latency of the memory device by ensuring high availability of storage transistors.

[0104] As used herein, memory access concurrency refers to the simultaneous execution of multiple memory operations in randomly addressed tiles of a memory device, with each tile performing memory operations on an access unit or memory page of memory data. In some examples, the memory operations include read, write, and refresh operations. In other words, the memory device operates with overlapping memory operations being performed across multiple tiles, with the memory device continuing to receive commands input from the controller device while executing previously received commands, with each command being executed by or directed to a different tile within the memory device. When the memory device is divided into multiple memory channels, memory access concurrency refers to the simultaneous execution of multiple memory operations in multiple tiles in each channel of the memory device. Also, in this description, the multiple memory operations are performed asynchronously with each other. That is, the multiple memory operations are not initiated in the same clock cycle, but rather in different clock cycles when the memory device (or memory channel) receives a command.

[0105] Memory access concurrency is particularly beneficial in memory devices with asymmetric read / write latencies. In an exemplary embodiment of the present invention, a memory device may have a write latency that is much longer than the read latency. For example, the write latency may be 300ns to 1μs and the read latency may be 70 to 90ns. In this case, memory performance is improved by overlapping read and write operations occurring simultaneously in different tiles, continuing to provide high memory bandwidth and high memory availability despite the memory device's long write latency. Conventional memory devices (e.g., DRAMs) that do not have asymmetric read and write latencies or have short read and write latencies do not need to implement memory access concurrency because they typically operate by completing each memory operation before the next operation. In memory devices that have asymmetric read / write latencies, or write latencies that are longer than the read latencies, such as the memory devices in the embodiments of the present disclosure, memory access concurrency can be implemented to perform overlapping memory operations to mask what would be a long write latency, such that the memory device behaves as if it had short read and write latencies.

[0106] In an embodiment of the present disclosure, a memory structure for forming a memory device of the present invention includes a memory array portion configured as described above with reference to FIGS. 4A and 4B to form a three-dimensional array of randomly accessible NOR memory strings of storage transistors. To complete the memory device, the memory structure includes a staircase portion at the end (Y direction) of the memory strings, as shown in FIG. 9. The thin-film storage transistors of the NOR memory strings are formed in the memory array portion, and the staircase portion on the opposite side of the array portion includes a staircase structure that provides connection to a common bit line and optionally a common source line of the NOR memory strings through conductive vias. In some embodiments, the common source line is precharged to act as a virtual voltage reference source during programming, reading, and erasing operations, thereby eliminating the need for continuous electrical connection to support circuitry during such operations. The common source line is described herein as being electrically floating to refer to the lack of continuous electrical connection to the common source line. Various processing steps can be used to form the staircase structure in the memory structure in embodiments of the present disclosure. The processing steps for forming the staircase structure may occur before, after, or between the processing steps for forming the memory array portion.

[0107] The memory structure described above with reference to Figures 4A and 4B illustrates the configuration of a memory array including a three-dimensional array of NOR type memory strings. This memory structure can be used as a building block to form a large capacity, high density memory device. In an embodiment of the present disclosure, the memory structure described above can be used as a building block to form a tile including a memory array as a modular memory unit and a local modular control circuit formed below the memory array. The memory device of the present disclosure is formed using an array of tiles. In an exemplary embodiment, the memory device is configured as a two-dimensional array of tiles arranged along the X and Y directions, each tile including a three-dimensional array of storage transistors with support and control circuitry for each tile formed below the respective tile. More specifically, the memory device includes a plurality of memory arrays of thin film storage transistors configured as a two-dimensional array of "tiles" (i.e., the tiles are arranged in rows and columns) formed on a plane of a semiconductor substrate. Each tile is configured to be individually addressed. This configuration makes the tiles modular units that allow flexibility in configuring memory modules to fit application requirements.

[0108] 9 is a cross-sectional view in the YZ plane of a tile in a memory device according to an embodiment of the present invention. Referring to FIG. 9, a tile 201 is formed on a semiconductor substrate 200. The memory structures of the tile 201 are formed in an insulating film 211, and a protective film 212 (passivation film) is formed on the insulating film 211. In some implementations, the insulating film 211 is made of silicon oxide (SiO x) and protective film 212 is formed from polyimide. In this embodiment, the memory structure includes a three-dimensional array of storage transistors ("memory array") configured as described above with reference to the memory structure of Figure 4A or Figure 4B. More specifically, the memory structure may be a three-dimensional array of charge trapping type storage transistors (Figure 4A) or a three-dimensional array of junctionless type ferroelectric storage transistors (Figure 4B).

[0109] A P-type or N-type diffusion region 221 is formed on the top surface of the semiconductor substrate 200. Other structures (not shown in FIG. 9), such as isolation structures or shallow trench isolation (STI) structures, may also be formed in the semiconductor substrate 200. A gate electrode 222 is formed on the semiconductor substrate 200 and is insulated from the semiconductor substrate by a gate dielectric layer. For example, the gate dielectric layer may be a thin silicon oxide layer. The gate electrode 222 together with the P-type and N-type diffusion regions 221 form a transistor in the semiconductor substrate 200. The transistor may be used to form a circuit element. For example, the transistor may be used to form support circuitry for operating storage transistors in a 3D NOR type memory array formed in the tile 201. The circuit elements are interconnected by interconnects 224 formed in the insulating film 211 of the lower interconnect portion 232 and contacts 223 that connect to one or more layers of vias 225 to form the support circuitry. In some embodiments, the support circuitry for the storage transistor is provided in the circuit element portion 231 and the lower interconnect portion 232. For example, support circuitry forming modular control circuitry for an array of storage transistors is formed in circuit element portion 231 and lower interconnect portion 232 .

[0110] In the tile 201, a three-dimensional NOR type memory array 210 is formed in a memory array portion 233. An upper interconnect portion 234 is formed on the upper side of the memory array portion 233. An insulating film 211 in the upper interconnect portion 234 is provided with interconnects 226 and vias 227 for forming further electrical connections. In some embodiments, a conductive pad 228 is provided in the upper interconnect portion 234 for connecting to a circuit element outside the semiconductor memory device. For example, a protective film 212 is formed on the upper interconnect portion 234, seals the upper interconnect portion 234, and has an opening that exposes at least a part of the conductive pad 228.

[0111] In the memory array portion 233, thin-film storage transistors are arranged as a three-dimensional array of NOR type memory strings in the memory array section 202. The memory array portion 202 is disposed between the staircase portion 203a and the staircase portion 203b. The staircase portion 203a and the staircase portion 203b are formed on opposite sides of the tile 201. The staircase portion 203a and the staircase portion 203b are provided with connections to a common bit line and optionally a common source line of the NOR type memory strings through conductive vias. In some embodiments, the common source line is precharged and then maintained at a relatively constant voltage to act as a virtual voltage reference during programming, erasing, and reading operations, thereby eliminating the need for continuous electrical connections to support circuitry during such operations. In FIG. 9, the array portion 202, the staircase portion 203a, and the staircase portion 203b are not drawn to scale. For example, the array portion 202 may be much larger in area than either the staircase portion 203a or the staircase portion 203b.

[0112] In the memory array portion 202, thin film storage transistors are formed at the intersections of common drain and source lines (collectively designated 204) and local word lines 205. A gate dielectric layer 206 is formed between the conductive local word lines and a channel layer (not shown in FIG. 9). The common drain and source lines are arranged in multiple planes extending in the Y direction, and the local word lines 205 are arranged in the Y direction as columnar structures extending in the Z direction, forming storage transistors in a three-dimensional array arranged in multiple rows in the X direction on multiple planes in the Z direction along each memory string in the Y direction. In FIG. 9, global word line conductors 208 provide electrical connection between circuitry 222 underneath the memory array 210 and the local word lines 205 associated with the three-dimensional memory stack.

[0113] In the above embodiment, the support circuits are described as being formed under the memory array portion 233. Such a configuration is merely exemplary and is not intended to be limiting. For example, in other embodiments, both the memory array portion and the support circuits may be formed directly on the semiconductor substrate 200. In this case, for example, the support circuits may be disposed around the memory array portion. In other embodiments, the support circuits may be formed on a separate semiconductor substrate. In this case, for example, the semiconductor substrate on which the memory array portion is formed and the semiconductor substrate on which the support circuits are formed are bonded to each other after forming the respective memory elements and circuit elements.

[0114] FIG. 9 illustrates one exemplary embodiment of a tile of storage transistors, or a physically separated memory array. The depiction of a tile 201 in FIG. 9 is for illustration only and is not intended to be limiting. FIG. 9 is provided to illustrate the incorporation of the memory structure of FIG. 4B or FIG. 4A to form a tile that includes modular memory units (memory arrays) and modular control circuitry. The tile can then be used as a building block to form a memory device that includes multiple arrays of three-dimensional storage transistors, such as junctionless ferroelectric storage transistors, to provide a desired memory capacity at a high density level.

[0115] In an embodiment of the present invention, the tile 201 implements a horizontal NOR type (HNOR type) memory architecture in which the bit lines are stacked in multiple planes and the word lines are formed as columnar structures. The bit lines are connected to bit line driver circuits (including bit line select transistors) to select and drive a memory page of the bit lines for memory operations. The word lines are connected to word line driver circuits (including word line select transistors) to select one of the word lines in the tile 201 for memory operations. With this configuration, in the support circuitry formed below the memory array, the bit line driver circuits are disposed below the staircase portions 203a, 203b, and the word line driver circuits are disposed below the memory array portion 202. With this configuration, a compact (small) under-array circuitry can be formed because the bit line drive / select transistors are generally smaller in size than the word line drive / select transistors. In memory devices with thin film storage transistors, such as charge trapping storage transistors and ferroelectric storage transistors, the word line drive / select transistors are typically configured to have a larger on / off voltage difference than the bit line drive / select transistors. Therefore, the word line driver circuits are typically much larger than the bit line driver circuits.

[0116] The HNOR memory architecture of the present disclosure allows for compact under-array circuitry to be formed since the bit line drive / select transistors require a smaller footprint and can be pulled under the staircase portion, while the larger word line drive / select transistors are placed under the memory array portion where there is more space to accommodate the larger transistors. Thus, the tile 201 in this embodiment is scalable because the size of the tile 201 is not limited by the size of the staircase structure or the driver circuit. This is in contrast to conventional memory architectures (e.g., NAND flash memory) in which the word lines are stacked and the bit lines are formed in columnar structures. In these conventional memory architectures, the larger word line drive / select transistors must be placed under the staircase portion, so the large size of the word line drive / select transistors determines the size of the tile or limits the size of the memory tile to a minimum size. Conventional memory architectures with stacked word lines cannot support memory devices with a large number of tiles because each tile must have a large tile size to accommodate the word line drive / select transistors that must be placed under the staircase portion. The HNOR memory architecture implemented in the memory device of the present invention allows for compact tile sizes to be formed so that the memory device can be formed to include a large number of tiles. A memory device having a large number of tiles allows for concurrency of memory operations, increasing memory bandwidth and memory availability of the memory device.

[0117] FIG. 10 illustrates a two-dimensional array of tiles forming part of a memory device according to an embodiment of the present invention. Referring to FIG. 10, in an embodiment of the present disclosure, a memory device 280 includes a two-dimensional array of tiles 201, each including a memory array 210, which is a three-dimensional array of storage transistors, and a module control circuit 250. The tiles 201 may be configured as described above with reference to FIGS. 3, 4A, 4B, and 9. The memory array 210 of the tile 201 is formed on a semiconductor substrate portion 240. An insulating layer 253 is provided between the semiconductor substrate portion 240 and the memory array 210 formed thereon. A module control circuit (CuA) 250 is formed on the semiconductor substrate portion 240, which implements support circuitry for operating the storage transistors in each tile. In particular, each tile 201 includes its own module control circuit (CuA) 250 formed below each memory array 210. As used herein, semiconductor substrate portion 240 refers to semiconductor substrate 241 and interconnect structures 242 formed on semiconductor substrate 241. Additionally, in an embodiment of the present invention, each module control circuit 250 has approximately the same planar dimensions (in the XY plane) as memory array 210.

[0118] In the memory device 280, each tile is controlled by a respective module control circuit 250 to perform memory operations on the storage transistors in the memory array. The memory device also includes additional peripheral control circuits in the areas between the tiles or around the memory device. In some examples, the peripheral control circuits include input / output circuits for communicating with a memory controller, protection circuits (e.g., against electrostatic discharge), data path circuits, interface circuits, and other control logic circuits. The peripheral control circuits may further include analog circuits (e.g., regulators, voltage reference circuits, temperature sensors) and non-volatile memory (e.g., electronic fuse memory, one-time programmable memory) for storing trimming data. The peripheral control circuit circuits support the operation of each module control circuit 250 but do not directly control memory operations on the storage transistors, such as reading from or writing to the storage transistors.

[0119] With such a configuration, the memory device 280 implementing the tile-based architecture of the present invention includes multiple tiles, each tile formed as a modular memory unit by a physically separated memory cell array, and each modular memory unit operated by a local modular control circuit (CuA). In other words, the memory array includes multiple tiles, each tile operated independently of each other by a modular control circuit formed in each tile. Each modular memory unit is a physically separated memory cell array, each modular memory unit includes its own three-dimensional array of memory cells and its own staircase structure for connecting the memory cells to the modular control circuit. Each modular memory unit is an individual and independently operable memory array, and the multiple modular memory units are multiple instances of the same modular memory unit structure, and each individual modular memory unit is physically separated from the other modular memory units. Each modular memory unit operates by its own local modular control circuit and performs semi-autonomous memory operations (e.g., read operations, write operations).

[0120] In an embodiment of the invention, the memory device interacts with a memory controller, such as those described above with reference to Figures 1A, 1B, 2A, 2B, 7, and 8, to receive input commands having addresses for performing memory operations specified by the commands in the respective memory tiles specified by the addresses. Each tile having a module control circuit is described herein as operating semi-autonomously in that the module control circuit receives commands having addresses that specify its associated memory array and operates independently to perform memory operations specified by the commands in its associated memory array. The module control circuit does not receive control signals for memory operations from the peripheral control circuit, although it may receive regulated voltage levels or clock signals from the peripheral control circuit. Rather, the module control circuit is itself the processing unit for its associated module memory unit and generates all control signals for controlling its associated module memory unit (or memory array) to perform memory operations.

[0121] A salient feature of the memory device of the present invention is that each tile is used as an operational unit of memory access, i.e., each tile operates on an access unit of memory data, such as a page or memory page of memory data. Thus, each command from the memory controller is addressed to a single tile, and each tile operates on the command to read or write to a memory page of memory data, independent of commands addressed to other tiles. As mentioned above, as used herein, an access unit or memory page of memory data refers to the number of data bits of memory data in each memory access request from the host device to the memory controller. In some examples, the access unit is 512 bits of memory data. It is useful to note that the present description only discusses memory data in an access unit, with the understanding that each memory access may include additional data bits, such as metadata bits, error flags, or other data bits that are used by the host but are not strictly memory data.

[0122] When connected and operated by a memory controller, the memory device 280 provides memory access concurrency, where multiple memory operations are performed simultaneously in multiple tiles, each memory operation is performed independently of the other, and overlapping memory operations are performed in different memory tiles. The memory controller receives memory access requests from a host device and issues commands (with memory addresses) to the memory device in response to the received memory access requests. For example, the memory controller issues commands to each tile of the memory device every 4 ns or 5 ns. In an embodiment of the present disclosure, the memory controller implements a tile or bank conflict avoidance scheme to avoid sending new commands to a tile that is currently processing an existing command. Tile conflict avoidance is particularly beneficial when the memory device has asymmetric latencies, such as a very long write latency relative to the read latency. For example, when the write latency (e.g., 1 μs) is much longer than the read latency (e.g., 90 ns), the memory controller performs a write operation in one tile and overlaps the read and write operations performed in other tiles to enable concurrency. In this way, long write latencies do not hold up read operations while waiting for a write operation to complete. An example of a tile or bank conflict avoidance scheme was described above with reference to FIG.

[0123] With such an arrangement, the memory device receives commands from the memory controller, each command being addressed to a tile, and no command is issued to a tile that is currently executing another command. Thus, the memory device receives commands to activate other tiles that are not currently executing commands. Furthermore, each command is issued for an entire access unit of memory data, and each tile operates independently of the others to perform memory operations on an entire access unit or memory page of memory data. In this manner, tiles within the memory device can be addressed individually and are operated individually and simultaneously, such that each tile operates on an entire memory page of memory data for each memory access request. The memory device operated by the memory controller ensures high availability of tiles, thereby increasing memory bandwidth and reducing tail latency of memory access requests. Importantly, by dividing the memory address space of the memory device into multiple tiles, the probability of memory access conflicts in which incoming memory accesses are addressed to the same tile is significantly reduced, thereby contributing to reduced tail latency of the memory device and improved quality of service levels. In particular, the tail latency of the memory devices of the present disclosure is improved by having a large number of tiles available for host memory accesses, thereby avoiding situations where host memory accesses must be stalled in order to activate all or most of the memory banks.

[0124] In an embodiment of the present invention, each tile of the memory device is operated by a local module control circuit (CuA). With continued reference to FIG. 10, each module control circuit 250 is a complete control circuit replicated for each tile such that memory operations within each tile are self-contained. The module control circuit 250 includes a sequencer that functions as an intelligent processing unit that executes commands addressed to each tile received from the memory controller and triggers memory operations in the associated tile in response to the received commands. Exemplary embodiments of the module control circuit are described below with reference to FIGS. 11 and 12. The structure and configuration of the module control circuit described herein are exemplary only and are not intended to be limiting. Other configurations and design schemes can be used in other embodiments.

[0125] FIG. 11 is a block diagram illustrating a module control circuit that may be incorporated into each tile of a memory device to provide intelligent, semi-autonomous memory operation control, according to some embodiments of the present invention. In some examples, the module control circuit 300 of FIG. 11 may be used to implement the module control circuit 250 of FIG. 10, or the support circuit or CuA described in the above embodiments. With reference to FIG. 11, the module control circuit (CuA) 300 implements the support circuitry for the module memory unit and is formed locally under the respective memory array. The module control circuit 300 includes a circuit portion formed under the staircase portion of the tile and a circuit portion formed under the memory cell array portion of the tile. In general, the module control circuit 300 includes bit line (BL) and word line (WL) selection circuitry for selecting a memory page of storage transistors from the memory array in response to a received address, bias control circuitry for controlling the voltage bias applied to the selected bit line and word line to perform a memory operation triggered by a received command, and control circuitry.

[0126] The following exemplary memory device configurations are used to facilitate the description of the present invention. In some examples, each memory array in a memory device or in a memory channel of a memory device includes M word lines. In each memory operation, one of the M word lines is activated to select P memory pages. Each memory page includes Q data bits. Thus, each memory array includes P×Q bit lines formed on multiple planes of the memory array, each bit line being associated with a NOR type memory string including M storage transistors. A word line is activated to select P memory pages, and Q bit lines are selected to select a memory page of storage transistors. Note that the staircase portion of the tile is also referred to herein as an upper staircase portion and a lower staircase portion. The terms "upper" and "lower" are used to facilitate reference to the figures and are not intended to indicate a particular orientation or a particular arrangement of circuit elements.

[0127] In an embodiment of the present invention, the module control circuit 300 includes a circuit portion 252 formed in a central portion of the module control circuit. The circuit portion 252 includes a control circuit, also referred to as a tile logic circuit. In some embodiments, the tile logic circuit 252 includes a sequencer or processor unit that executes a sequence of memory operations on the memory array. Specifically, the sequencer receives input commands from a memory controller addressed to an associated tile along with associated memory addresses and write data (if applicable). The sequencer decodes each command received, such as by decoding microcode in the command, and executes instructions according to the command to perform memory operations on the memory array, each memory operation including a sequence or steps of operations operating on storage transistors of the memory array. For example, the command may be a read command to read data from a particular memory page in the memory array. In another example, the command may be a write command having write data to store data in a particular memory page in the memory array. In yet another example, the command may be a refresh command to refresh the stored data of a particular page in the memory array. Specifically, the sequencer generates control signals to execute the sequence of operations in the memory array and the timing of each memory operation. For example, the sequencer generates control signals that dictate timing sequences for performing various operational sequences, such as selecting word lines and bit lines, applying bias voltages, and sensing, programming, and erasing selected storage transistors. Importantly, the sequencers in each module control circuit operate independently of one another without requiring control from circuit elements external to the module control circuit. In this manner, the sequencers in the tile logic circuit 252 provide intelligence to the module control circuits, enabling the module control circuits 300 to perform semi-autonomous memory operations.

[0128] In an embodiment of the invention, the module control circuit 300 includes circuit portions 254A and 254B that implement row circuits for selecting and activating word lines in the memory array. In this embodiment, the row circuit is divided into two circuit portions 254A, 254B formed adjacent to the tile logic portion 252 in the center of the module control circuit. In other embodiments, the row circuit may be formed in a single circuit portion 254 and provided to one side of the tile logic portion 252. The row circuit includes a row decoder circuit 255, a row bias control circuit 256, and a word line selection circuit 257, also referred to as a word line (WL) driver circuit. In this embodiment, the circuit portions 254A and 254B may each be configured to control half of the word lines in the memory array. For example, the circuit portion 254A may be configured to control the upper half of the word lines in the memory array, and the circuit portion 254B may be configured to control the lower half of the word lines in the memory array.

[0129] The row decoder circuit 255 is configured to decode memory addresses associated with commands received by the tile logic circuit 252 to determine which word lines should be activated for the memory operation. The row bias control circuit 256 generates and controls word line voltage values ​​for biasing selected word lines as a function of the memory operation being performed. The row bias control circuit 256 also generates and controls word line voltage values ​​for unselected word lines, for example, to minimize disturbance of unselected memory cells. In one example, the selected word lines are applied with a first voltage value for a read operation and a second voltage value different from the first voltage value for a write operation. In another example, the unselected word lines are applied with an inhibit voltage value to minimize disturbance of data stored in unselected memory cells. Finally, the word line driver circuit 257 includes word line drive / select transistors connected to the M word lines in the memory array. In response to a decoded address from row decoder 255 specifying a word line for selection, word line driver circuit 257 selects the specified word line and applies to the selected word line an appropriate bias voltage value associated with the memory operation to be performed. Word line driver circuit 257 also applies an inhibit voltage to the other M-1 unselected word lines. The row circuit (circuit portions 254A, 254B) operates in this manner to select and drive one word line out of the M word lines for a memory operation.

[0130] The module control circuit 300 includes circuit portions 258A and 258B that implement column circuits for selecting and activating bit lines in the memory array. In this embodiment, the column circuit is divided into two circuit portions 258A, 258B formed adjacent to the row circuit portions 254A, 254B. In other embodiments, the column circuit may be formed in a single circuit portion 258 and provided on one side of the tile logic circuit portion 252. The column circuit includes column decoder circuitry and column bias control circuitry. In this embodiment, the circuit portions 258A and 258B may each be configured to control half of the bit lines in the memory array. For example, the circuit portion 258A may be configured to control the upper half of the bit lines in the memory array, and the circuit portion 258B may be configured to control the lower half of the bit lines in the memory array.

[0131] The column decoder circuitry in the column circuitry 258A or 258B is configured to decode a memory address associated with a command received by the tile logic circuitry 252 to determine the memory page of the bit line to be activated for the memory operation. The column bias control circuitry in the column circuitry 258A or 258B generates and controls the bit line voltage values ​​for driving the selected bit lines as a function of the memory operation to be performed. The column bias control circuitry also generates and controls the bit line voltage values ​​for the unselected bit lines to minimize disturbance of the unselected memory cells. In one example, the selected bit lines are applied with a third voltage value for a read operation and a fourth different voltage value different from the third voltage value for a write operation. In another example, the unselected bit lines are applied with an inhibit voltage value to minimize disturbance of data stored in the unselected memory cells.

[0132] The module control circuit 300 includes circuit portions 260A, 260B that realize sense circuits and latch circuits. In this embodiment, the sense circuits and latch circuits are divided into two circuit portions 260A, 260B formed adjacent to the column circuit portions 258A, 258B, respectively. The arrangement and division of the sense circuits and latch circuits depends on the configuration of the staircase structure of the tile that provides the bit lines of the NOR type memory strings of the memory array for connection to the module control circuit. When the staircase structure includes two portions formed at both ends of the tile, it is beneficial for better sensing and driving performance to place the corresponding sense circuits and latch circuits near the respective bit line staircase portions.

[0133] In the circuit portions 260A and 260B, the sense amplifier circuit includes Q sense amplifiers corresponding to the number of data bits in the memory page. The Q sense amplifiers are connected to the Q selected bit lines. Thus, each memory operation is performed on the memory page of data bits and no sense amplifier selection is required. By providing the same number of sense amplifiers in the module control circuit 300 as the number of data bits in the memory page, the size of the module control circuit can be kept compact and the module control circuit can be formed under each memory array. In this embodiment, Q selected bit lines are provided through the upper staircase portion and the lower staircase portion. Thus, half of the sense amplifiers are provided in the circuit portion 260A and connected to the Q / 2 selected bit lines of the upper staircase portion, and the other half of the sense amplifiers are provided in the circuit portion 260B and connected to the Q / 2 selected bit lines of the lower staircase portion.

[0134] The data latch circuitry may include two or more data latches or registers for storing output read data or input write data. The sense amplifier control circuitry directs operation of the sense amplifiers and data latches to sense read data from selected bit lines or provide write data to selected bit lines. In some examples, the sense amplifiers, data latches, and bias control circuitry may be configured as described above with reference to FIG.

[0135] In an embodiment of the present invention, the module control circuit 300 includes circuit portions 262A and 262B that implement a bit line (BL) driver circuit for selecting and driving the bit lines (e.g., Q bit lines) of a selected memory page in the memory array. The bit line driver circuit includes a set of bit line select transistors that select and drive the selected bit line. In this embodiment, the bit line select circuit is split into two circuit portions 262A, 262B formed on either side of the module control circuit. Also, the two circuit portions of the bit line select circuit are formed under the respective staircase portions on either side of the tile.

[0136] More specifically, the bit line selection circuit includes a first set of bit line selection transistors formed under the upper staircase portion and a second set of bit line selection transistors formed under the lower staircase portion. As described above, in this embodiment, P×Q bit lines are provided through the upper staircase portion and the lower staircase portion. That is, half of the bit lines of the memory array are connected to the staircase structure formed at the first end (e.g., the top end) of the tile and connected to the bit line selection transistors formed under the staircase structure. The other half of the bit lines of the memory array are connected to the staircase structure formed at the second end (e.g., the bottom end) opposite to the first end of the tile and connected to the bit line selection transistors formed under the staircase structure. With this configuration, the first set of bit line selection transistors (circuit portion 262A) are connected to (P×Q) / 2 bit lines and are formed under the upper staircase portion to select and drive Q / 2 bit lines for memory operations. A second set of bitline select transistors (circuit portion 262B) is connected to the (P×Q) / 2 bitlines and is formed below the lower staircase portion to select and drive the Q / 2 bitlines for memory operations. Thus, the bitline select circuitry selects Q bitlines from the P×Q bitlines in the memory array to perform memory operations.

[0137] In this configuration, the first and second sets of bitline select transistors are connected to the bitlines of the NOR memory string at the upper and lower staircases to select the bitlines and apply appropriate voltage levels to the selected bitlines for read and write operations. Specifically, the bitline select transistors select a memory page of Q bitlines from all P×Q bitlines in response to a column decoder that decodes a memory address associated with a command received by the tile logic circuit 252. The bitline select transistors apply bias voltages generated by the column bias control circuitry to the selected bitlines via the sense amplifier circuitry to perform the specified memory operation. The unselected bitlines are left electrically floating or are not actively biased to a particular voltage potential.

[0138] In operation, in the module control circuit 300, the tile logic circuit 252 receives an input command and a memory address. The row decoder 255 decodes the memory address and determines which word line to select. The row bias control circuit 256 generates bias voltage levels for the selected and unselected word lines. The word line driver circuit 257 selects a designated one of the M word lines for a memory operation. Meanwhile, the column decoder (circuit portions 258A, 258B) decodes the memory address and determines which memory page to select. The column bias control circuit generates bias voltage levels for the selected bit line. The bit line driver circuit (circuit portions 262A, 262B) selects the Q bit lines of the selected memory page that are connected to the sense amplifiers. A sequencer in the tile logic circuit 252 controls the sequence and timing of the various steps to perform a memory operation. For example, for a read operation, the sense amplifier first precharges the selected bit line and then senses a current or voltage value on the selected bit line as a result of the storage transistor in the selected memory page being activated by the selected word line. The sensed data value of the Q data bit is stored in the data latch (circuit portions 260A and 260B) for reading by the memory controller on the I / O bus of the memory device. For a write operation, the sense amplifier precharges the selected bit line and applies a bit line bias voltage to the selected bit line according to the write data value (stored in the data latch), which is stored in the storage transistor in the selected memory page activated by the selected word line. In some embodiments, the unselected bit lines are biased to a forbidden voltage value by virtual capacitive coupling with the unselected word lines. The sequencer performs read and write operations, each operation including various stages with different bias voltage values ​​and timing. The sequencer manages the sequence of stages in each memory operation as well as the timing of each other stage.

[0139] In this embodiment, the module control circuitry 300 is configured with the tile logic circuitry 252 in the center, flanked by decoder circuits, bias control circuits, selection circuits, sense circuits, and latch circuits. In this configuration, control signals from the tile logic circuitry 252 to other circuits can be routed across the tile area, such as using conductive interconnects and vias in the interconnect portion 232 (FIG. 9).

[0140] The configuration of the module control circuit 300 in FIG. 11 is illustrative only and is not intended to be limiting. The circuits in the module control circuit 300 may be arranged in other configurations depending on other design considerations such as the size of the transistors or circuit elements and the need for signal routing. FIG. 12 is a block diagram illustrating a module control circuit that may be incorporated into each tile of a memory device according to another embodiment of the present invention. Referring to FIG. 12, the module control circuit 350 includes the same circuit elements as the module control circuit 300 in FIG. 11, and the same elements are numbered the same for ease of explanation. The module control circuit 350 is constituted by the tile logic circuit 252 arranged on the side of the tile area. The row circuit, including the row decoder 255, the row bias control circuit 256, and the word line driver circuit 257, is arranged in the center adjacent to the tile logic circuit 252. The column decoder and bias control circuits 258A and 258B are similarly arranged on either side of the row circuit, adjacent to the tile logic circuit 252. In some examples, the circuit layout of the module control circuitry 350 accommodates sense amplifier or column decoder circuitry that occupies a larger area so that the tile area can be used more efficiently by placing the tile logic to the side. In the configuration of the module control circuitry 350, the routing of control signals from the tile logic 252 to other circuit elements can include routing in two different directions.

[0141] In the module control circuit 300 of FIG. 11, control signals from the tile logic circuit 252 to other circuits can be routed in only one direction on the tile area, for example, from top to bottom, or vice versa, as shown by the arrows in FIG. 11. In this case, command and address signals can be routed to the module control circuit 350 of each tile using the routing area between tiles in the memory device, or using some area on the tile. In the module control circuit 350 of FIG. 12, control signals from the tile logic circuit 252 to other circuits can be routed both horizontally and vertically, as shown by the arrows in FIG. 12. In this case, command and address signals to the module control circuit 350 of each tile are generally routed using the routing area between tiles in the memory device.

[0142] FIG. 13 is a block diagram illustrating a bit line selection and sense amplifier configuration in a module control circuit according to an embodiment of the present invention. The circuit elements in the block diagram of FIG. 13 are illustrative of the circuit elements in the module control circuit of FIG. 11 or FIG. 12. Furthermore, for ease of explanation, assume that the memory array has 16,384 bit lines, each word line selects one memory page out of 32 memory pages, and the size of the memory page is 512 bits. Additional data bits for metadata or error correction or health indication may be provided, but are not included in this figure for simplicity. With reference to FIG. 13, the 16k bit lines of the memory array are split into two parts and connected to the module control unit via two staircase structures. In this embodiment, the upper staircase part is connected to 8,192 bit lines and the lower staircase part is connected to another 8,192 bit lines. In each staircase part, the bit lines are connected to a respective bit line driver circuit, each circuit selects the bit lines of half of the memory page, i.e., 256 bit lines. For example, in the upper staircase portion, the bit line driver circuit includes 256 selector circuits, each of which selects one of the 32 bit lines. Similarly, in the lower staircase portion, the bit line driver circuit includes 256 selector circuits, each of which selects one of the 32 bit lines. The selector circuits receive a page address and determine which bit line should be selected. The outputs of the 256×2 selector circuits corresponding to the 256×2 selected bit lines are connected to respective sense amplifier circuits. A set of 256 sense amplifiers is provided in the upper portion, and a set of 256 sense amplifiers is provided in the lower portion. The 256×2 sense amplifiers are then connected to respective data latches. For example, a set of data latches 261A is connected to sense amplifiers 260A, and a set of data latches 261B is connected to sense amplifiers 260B.

[0143] FIG. 13 illustrates the data flow of a read operation. The sensed data from the 256×2 sense amplifiers are stored in the data latches 261A, 261B for output to the memory controller. Thus, each set of data latches provides 256 bits or 32 bytes of output data. The memory controller receives and combines data from the two sets of data latches 261A and 261B to obtain a memory page of 512 bits or 64 bytes of memory data. The configuration of FIG. 13 also applies to a write operation, where 512 bits of write data are received by the module control circuitry and stored in the two sets of data latches 261A and 261B. The bit line driver circuitry selects the 512 bit lines of the memory page to be written, and the sense amplifiers provide the write data on the selected bit lines.

[0144] With this configuration, the module control circuitry includes as many sense amplifiers for sensing or driving bit lines as there are data bits in a memory page. No select signals are used to select the sense amplifiers. Rather, bit line select is used to connect only a memory page size of bit lines to the sense amplifiers. With this configuration, a compact module control circuitry can be formed, simplifying circuit design and control, and enabling high bandwidth operation.

[0145] In the above embodiment, the memory device includes an array of tiles, each tile providing an access unit of memory data, such as a memory page (e.g., 512 bits) of memory data. Note that the above description is in terms of memory operations in which each memory access from the host responds to a host request based on an access unit of memory data. In actual implementation, each tile may perform additional auxiliary memory operations that are not part of the host memory request, such as refresh and wear leveling. In some cases, each memory operation performed on a target memory page in a tile may include additional auxiliary memory operations on another related memory page in the same tile. The module control circuit (CuA) cooperates with the memory controller to control the auxiliary memory operations.

[0146] In an embodiment of the present disclosure, a write operation is performed together with a partial refresh operation, and each time a memory page is written, another memory page associated with the same word line is selected and refreshed at the same time. As described above, a destination memory address activates a selected word line, activating P memory pages. While one of the P memory pages is selected for a write operation, a partial refresh operation selects the other memory page belonging to the activated word line for refresh. The partial refresh operation may be advantageously applied to reduce disturbances experienced by unselected storage transistors associated with the activated word line. The partial refresh operation is disclosed in U.S. Patent Application No. 17 / 525,712, entitled “Methods For Reducing Disturb Errors By Refreshing Data Alongside Programming Or Erase Operations,” filed on November 12, 2021 (the disclosure of which is incorporated herein by reference in its entirety). When a partial refresh operation is performed, each write operation in a tile is performed on a specified memory page and a memory page to be refreshed, i.e., each write operation is performed on two memory pages. However, the memory access unit remains a single memory page, and only one memory page is written with the input write data, while the other memory page only has its contents refreshed. This description of memory operations performed in a tile based on an access unit of memory data does not exclude additional auxiliary memory operations being performed on other memory pages in the same tile.

[0147] Also, in the above-described embodiments, the tile-based support circuitry or module control circuitry (CuA) for each tile has been described as being formed in or on the semiconductor substrate in which the array of storage transistors is formed. In another embodiment, the module control circuitry (CuA) may be formed on a separate semiconductor substrate different from the semiconductor substrate in which the storage transistors are formed. In such a configuration, the memory array of storage transistors is formed on a first semiconductor die and the module control circuitry for each array of storage transistors is formed on a second semiconductor die. The memory device is formed by electrically and mechanically connecting the first semiconductor die to the second semiconductor die, with each memory array and associated module control circuitry forming a tile in the memory device, and the memory device includes an array of tiles formed by bonding the first semiconductor die and the second semiconductor die. The tiles, including the module control circuitry and associated memory arrays, can be individually addressed by an external memory controller.

[0148] For example, the tile 201 of Figure 9 may be formed by two semiconductor dies, with the memory array portion 210 formed on a first semiconductor die and the module control circuitry formed on a second semiconductor die. The second semiconductor die may be joined to the first semiconductor die via an interconnect structure formed on the underside (or bottom) or top side of the memory array portion 210. For example, the interconnect structure may be a hybrid bond. Forming the memory array portion and the module control circuitry on separate semiconductor dies has the beneficial effect of allowing the manufacturing process to be optimized separately for the memory circuitry and the control circuitry.

[0149] In some embodiments, the circuitry of each modular control circuit may be split between a portion of the control circuit formed below or above the memory array of storage transistors in the first semiconductor die and the remaining control circuitry formed in the second semiconductor die. For example, the bit line selector circuit / driver circuitry of each tile may be formed in or above the semiconductor substrate of the first semiconductor die below each memory array. Alternatively, in another example, the bit line selector circuit / driver circuitry of each tile may be formed as a vertical thin film transistor above (the opposite side of the semiconductor substrate) the respective memory array. In this aspect, the first semiconductor die need only provide a global bit line connection to the second semiconductor die. With such an arrangement, the memory device of the present invention may be configured in a variety of ways to provide an array of tiles of storage transistors with modular control circuitry. The memory device may be formed on a monolithic semiconductor die or in a multi-die configuration to allow for optimization of the manufacturing process between the memory circuitry and the support circuitry. Whether formed monolithically or in a multi-die configuration, the memory tiles including the modular control circuits and associated memory arrays may be individually addressable by an external memory controller.

[0150] In this detailed description, a process step described in one embodiment can be used in another embodiment even if not explicitly described in another embodiment. When referring to a method including two or more defined steps in this specification, the defined steps can be performed in any order or simultaneously, unless the context dictates or a specific indication is provided otherwise in this specification. Furthermore, the method can also include one or more other steps performed before any defined step, between two defined steps, or after all defined steps, unless the context dictates or an explicit indication is provided otherwise.

[0151] In this detailed description, various embodiments or examples of the invention may be implemented in various forms, such as processes, devices, systems, and compositions of matter. A detailed description of one or more embodiments of the invention has been provided above, along with accompanying drawings illustrating the principles of the invention. Although the invention has been described in connection with such embodiments, the invention is not limited to any embodiment. Various changes and modifications are possible within the scope of the invention. The scope of the invention is limited only by the appended claims, and the invention encompasses various alternatives, modifications, and equivalents. Numerous specific details have been described herein in order to provide a thorough understanding of the invention. These details have been provided for illustrative purposes, and the invention may be practiced according to the claims without some or all of these specific details. For the sake of clarity, technical matters known in the art related to the invention have not been described in detail so as not to unnecessarily obscure the invention. The invention is defined by the appended claims.

Claims

1. 1. A memory device, comprising: a plurality of tiles of memory circuitry, each of the tiles including a memory array, which is a physically separated array of storage transistors, electrically connected to and operated by a module control circuit, the memory array having a three dimensional array of storage transistors arranged into a plurality of memory pages of the storage transistors, each of the storage transistors being accessed by a word line and a bit line, the module control circuitry communicating with the memory array to perform memory operations on one or more memory pages of the storage transistors; each said tile being individually addressable by an associated said module control circuitry and configurable to operate independently of one another to perform memory operations on memory pages of said storage transistors in said memory array in response to memory access commands specified for said tile; A memory device, wherein two or more of the randomly addressed tiles of the memory circuitry are configurable to simultaneously perform overlapping memory operations.

2. 2. The memory device of claim 1, the plurality of tiles of the memory circuit are formed on a semiconductor substrate; a semiconductor substrate having a first tile and a second tile, the first tile including an array of physically isolated storage transistors formed above a plane of the semiconductor substrate, and a modular control circuit formed below the memory array on or within the semiconductor substrate.

3. 2. The memory device of claim 1, the tiles of memory circuitry include a plurality of the physically isolated arrays of storage transistors formed on a first semiconductor die and corresponding modular control circuitry for the arrays of storage transistors formed on a second semiconductor die; The first semiconductor die is electrically and mechanically connected to the second semiconductor die, whereby each of the memory arrays and associated modular control circuitry forms a tile of the memory device.

4. 2. The memory device of claim 1, a plurality of tiles of the memory circuit including a plurality of the physically isolated arrays of storage transistors, at least a portion of the module control circuitry formed on a first semiconductor die, and remaining corresponding module control circuitry for the arrays of storage transistors formed on a second semiconductor die; The first semiconductor die is electrically and mechanically connected to the second semiconductor die, whereby each memory array having a portion of the module control circuitry and associated module control circuitry forms a tile of the memory device.

5. 2. The memory device of claim 1, Each of the memory access commands is based on an access unit of memory data; the access unit is a memory page, The modular control circuitry of the tile performs a memory operation on a memory page of the storage transistors in response to each memory access command specified for the tile.

6. 6. The memory device of claim 5, The access unit of the memory data includes 512 bits; The memory page of storage transistors comprises 512 storage transistors.

7. 2. The memory device of claim 1, each said physically separate memory array in each said tile includes storage transistors arranged in the three dimensional array and one or more staircase structures connecting the storage transistors to the modular control circuitry.

8. 8. The memory device of claim 7, each said physically separate memory array in each said tile includes a three dimensional array of NOR type memory strings of randomly accessible storage transistors; The one or more staircase structures connect bit lines of the NOR memory strings to the module control circuitry.

9. 9. The memory device of claim 8, In each of the memory arrays, each word line activates a first number of memory pages; The module control circuitry selects bit lines associated with a first memory page of the storage transistors to perform a memory operation in response to the memory access command.

10. 2. The memory device of claim 1, A memory device, wherein each of the plurality of tiles of memory circuitry includes a semi-volatile memory circuit.

11. 11. The memory device of claim 10, The storage transistors of the semi-volatile memory circuits in each of the memory arrays are refreshed at intervals of 10 minutes or greater.

12. 11. The memory device of claim 10, The storage transistors in each of the memory arrays each include a storage material that is programmable by a direct tunneling technique.

13. 11. The memory device of claim 10, The storage transistors in each memory array each include a ferroelectric storage transistor.

14. 9. The memory device of claim 8, the three-dimensional array of NOR memory strings is arranged in parallel in a first direction and stacked in multiple layers in a second direction; Each of the NOR memory strings includes a plurality of storage transistors sharing a common source line and a common drain line; Each layer of the NOR type memory string extends along a third direction, The array of NOR type memory strings includes: (i) a plurality of local word line structures formed between adjacent stacks of the NOR memory strings; (ii) a plurality of word lines formed over the three dimensional array and extending in the first direction, each word line connected to a respective subset of the local word line structures; each said storage transistor is formed at a junction of each said NOR memory string and each said local word line structure; each said storage transistor including: (i) a channel region formed between a common source layer and a common drain layer; (ii) a gate dielectric layer; and (iii) the local word line structure as a gate conductor.

15. 15. The memory device of claim 14, Each of the module control circuits has the same dimensions in the first and third directions as its associated memory array.

16. 2. The memory device of claim 1, the module control circuits of the tiles interact with a memory controller integrated circuit formed on another semiconductor substrate; each module control circuit within the tile receives the memory access command designated for the tile along with a memory address associated with the command; A memory device, wherein the memory address specifies a memory page within the memory array for performing a memory operation specified by the memory access command.

17. 17. The memory device of claim 16, Each of the memory access commands specifies a memory operation, including read, write, and refresh operations; The memory access command provides a write operation command along with write data.

18. 17. The memory device of claim 16, each said physically separate memory array in each said tile includes a three dimensional array of NOR type memory strings of randomly accessible storage transistors; the one or more staircase structures connecting the bit lines of the NOR memory strings to the module control circuitry, and each module control circuit within the plurality of tiles is a complete control circuit replicated for each of the tiles.

19. 8. The memory device of claim 7, Each of the module control circuits is a bit line driver circuit connected to the bit lines of the memory array via the staircase structure, for selecting a bit line of a memory page designated by a memory address and a driving bias voltage in response to a memory operation designated by the memory access command; a word line driver circuit connected to select a word line of the memory array to activate the first number of memory pages of storage transistors; the bit line driver circuitry is disposed in a portion of the module control circuitry below the staircase structure of the memory array; A memory device, wherein the word line driver circuitry is disposed in a central portion of the module control circuitry.

20. 20. The memory device of claim 19, the module control circuitry of the designated tile responds to a memory address associated with the memory access command by selecting one of the word lines in the memory array via the word line driver circuitry to activate the first number of memory pages; The module control circuitry further selects bitlines of a memory page from the first number of activated memory pages to perform a memory operation indicated by the associated memory access command.

21. 20. The memory device of claim 19, each module control circuit within the tile further comprises a tile logic circuit including a processor unit for executing a sequence of memory operations on the memory array of the tile to decode the memory access command and perform the memory operation specified by the command; The tile logic circuitry performs a memory operation sequence for each of the memory access commands and generates control signals for controlling a timing sequence for performing the memory operation sequence.

22. 20. The memory device of claim 19, each module control circuit in the tile further comprises a sense amplifier circuit for the selected bit line of the bit line driver circuit for sensing or driving the selected bit line in response to the memory access command; the sense amplifier circuit includes a second number of sense amplifiers; the second number is equal to a number of data bits in an access unit of memory data associated with the memory access command; The access unit is equal to a memory page.

23. 23. The memory device of claim 22, a plurality of data latches configured to store read data retrieved from the memory array in response to a read operation and to store write data to be stored in the memory array in response to a write operation.

Images