A non-volatile memory and an operating method thereof
By configuring multiple blocks and sub-blocks in high-density nonvolatile memory and adjusting the threshold voltage of the memory cell using read setting operations, the threshold voltage offset problem caused by the resistivity change of the memory cell channel is solved, and the memory is improved operating window and stability.
Patent Information
- Application Number
- CN202110508760.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2021-01-20
- Filing Date
- 2021-05-10
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2041-05-10
AI Technical Summary
In high-density nonvolatile memory, the channel resistivity of the memory cell transistor changes over time, causing the threshold voltage to shift, reducing the memory operation window.
By configuring multiple blocks and sub-blocks in the memory, the read setting bias is applied to multiple memory cells simultaneously using the read setting operation, the threshold voltage of the memory cell is adjusted to match the voltage state during the programming period, and the channel stress state is maintained to reduce the impact of resistivity changes.
The memory operation window is effectively improved, the data loss of the memory unit is reduced, and the memory operation efficiency and stability are improved.
Smart Images

Figure CN114512167B_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to integrated circuit memories including non-volatile memories and methods of operating the same. Background Art
[0002] In high-density non-volatile memories (e.g., 3D NAND memories), the channels of memory cell transistors may have different resistivities before and after stress is applied. For example, the resistivity of thin-film polysilicon in the vertical channel structure of 3D NAND memories may change over time. The different resistances resulting from this characteristic may cause the threshold voltage to shift (e.g., 10 mV to 30 mV), which in turn may reduce the operating windows of the memory.
[0003] The present disclosure contemplates providing techniques that can improve the operating window for memory under these conditions.
[0004] Public Content
[0005] The present disclosure provides a non-volatile memory, including a memory array having a plurality of blocks, wherein a plurality of blocks in the plurality of blocks are configured to be enabled for memory operations in response to a plurality of word lines of a corresponding plurality of groups. A plurality of control circuits are part of the memory or coupled to the memory, and include logic for performing a read set operation, the read set operation including simultaneously applying a read set bias to a plurality of memory cells of a selected block in the plurality of blocks. The non-volatile memory includes logic for covering a plurality of blocks in the plurality of blocks to apply the read set operation to the plurality of blocks. A plurality of blocks in the plurality of blocks may respectively include a plurality of sub-blocks, wherein a plurality of sub-blocks in the plurality of sub-blocks are configured to be connected to the plurality of global bit line sets for a plurality of memory operations in response to a corresponding plurality of sub-block string select lines, and the sub-blocks include a plurality of NAND strings. The read setup operation may include covering multiple sub-blocks in the selected block to simultaneously apply a read setup bias to multiple memory cells corresponding to individual multiple sub-blocks in the selected block, including embodiments where one or more NAND strings connected to a single global bit line simultaneously receive the setup bias. For purposes of this description, the term "simultaneously" refers to biasing devices that at least partially overlap in time, such that all simultaneous biasing configurations consume current to increase the combined current consumption of the memory array.
[0006] Embodiments may include maintaining a block status table identifying a plurality of obsolete blocks, and read set operations may be applied to the plurality of obsolete blocks identified in the block status table. Additionally, a plurality of blocks in the array may be addressed sequentially or in a systematic sequence as a background operation or independently of a plurality of read instructions addressing the plurality of blocks to maintain a prepared state of a plurality of memory cells for a plurality of subsequent read operations.
[0007] The present disclosure provides an operation method of a non-volatile memory, wherein the non-volatile memory may include a memory plane, the memory plane includes a plurality of blocks of a plurality of memory cells and a plurality of bit lines, each block includes a plurality of NAND string sets, which have a plurality of string select lines for connecting to a corresponding plurality of bit lines of the plurality of bit lines, wherein each NAND string in the plurality of NAND string sets of the block is connected to a plurality of word line sets for the block. The method includes: covering a plurality of blocks to apply a plurality of read setting bias configurations to a plurality of read operations to condition a plurality of stale blocks in a plurality of stale blocks. The plurality of read setting biases include applying a read setting bias simultaneously to a plurality of memory cells of a selected block of the plurality of blocks. The method can be applied in a memory, wherein each of a plurality of blocks of a plurality of memory cells includes a plurality of sub-blocks, each sub-block includes a distinct subset of the set of a plurality of NAND strings of the block, a plurality of NAND strings of different subsets in each sub-block are operably connected to respective sub-block string select lines, a plurality of gate voltages are applied to a plurality of string select gates of the plurality of NAND strings of different subsets of the sub-blocks through the select lines, and a plurality of read set bias operations are simultaneously applied to one or more sub-blocks. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] Other aspects and advantages of the present disclosure will become apparent from a review of the following drawings, detailed description, and claims.
[0009] Figure 1 A perspective view of a three-dimensional vertical NAND structure with block and sub-block architecture.
[0010] Figure 2 is a circuit diagram of a NAND flash memory cell block, which can be used similar to Figure 1 architecture to achieve this.
[0011] Figure 3is a simplified schematic diagram of a sense amplifier and bit line bias circuit that may be used in the devices described in the present disclosure.
[0012] Figure 4 is a block diagram of segmentation of a memory plane according to an embodiment described in the present disclosure.
[0013] Figure 5 is a block diagram of a segment of a memory array including multiple planes according to an embodiment described in the present disclosure.
[0014] Figure 6 A timing diagram for a read setup bias configuration for an embodiment described in the present disclosure.
[0015] Figure 7 A timing diagram illustrating variations in the read setup bias configuration for embodiments described herein.
[0016] Figure 8 Timing diagrams for more variations of the read setup bias configurations for the embodiments described in the present disclosure.
[0017] Fig. 9 A timing diagram for an alternative read setup bias configuration for embodiments described herein.
[0018] Fig.10 Timing diagrams for further alternative read setup bias configurations for embodiments described herein.
[0019] Fig.11 Timing diagrams for further alternative read setup bias configurations for embodiments described herein.
[0020] Fig.12 Another timing diagram for yet another alternative read setup bias configuration for embodiments described herein.
[0021] Fig.13 is a block diagram of a memory system described in the present disclosure.
[0022] Fig.14 A flow chart of a read setup operation according to the present disclosure.
[0023] Fig.15 Flowchart for a read operation with a read set operation for a stale block.
[0024] [Description of Reference Numerals]
[0025] 10: Substrate
[0026] 11: Word line layer
[0027] 12: Multiple serial selection lines
[0028] 13: Lower selection line
[0029] 15: Column
[0030] 18: Common source conductor
[0031] 20: Global bit line conductor
[0032] 21: First stack
[0033] 22: Second stack
[0034] 201: Dashed Line
[0035] 202: Sub-block serial selection gate
[0036] 203: Bit line
[0037] 318: Transistor
[0038] 319: Line
[0039] 320: Global bit line
[0040] 321: Transistor
[0041] 322: BLC1 line
[0042] 323: Connecting Nodes
[0043] 327: Transistor
[0044] 332: Sensing Node
[0045] 335: Transfer transistor
[0046] 336: Transistor
[0047] 337: Capacitor
[0048] 338: Sensing signal node
[0049] 339: Sensing transistor
[0050] 400: Flat
[0051] 405: Bit line bias circuit
[0052] 410, 411, 412, 413, 414: word line WL driver
[0053] 500: Memory
[0054] 600, 601, 602, 700, 701, 702: falling edge
[0055] 610, 611, 612, 710, 711, 712: Time
[0056] 800, 900, 1000, 1201: voltage
[0057] 1101, 1102: Line segment
[0058] 1110, 1111, 1112: Time
[0059] 1202: Host
[0060] 1208: Memory device
[0061] 1214, 1216, 1244: Line
[0062] 1218: Input / Output Circuit
[0063] 1226: Bus system
[0064] 1228: Memory Circuit
[0065] 1234: Control circuit
[0066] 1235: Status register
[0067] 1245, 1246, 1265: Arrow
[0068] 1248: Bit line driver circuit
[0069] 1264: Block
[0070] 1266: Bit line
[0071] 1276A: Word Line Decoder
[0072] 1276B: Word line driver circuit
[0073] 1277: Word Line
[0074] 1278: Memory Array
[0075] 1278A: SSL / GSL Decoder
[0076] 1278B: SSL / GSL driver circuit
[0077] 1279: GSL line
[0078] 1300, 1301, 1302, 1303, 1304, 1310, 1311, 1312, 1313, 1314, 1315, 1400, 1401, 1402, 1403, 1404, 1405: Blocks
[0079] 1350, 1450: Block status table
[0080] 1351, 1451: Status register DETAILED DESCRIPTION
[0081] Please refer to Figures 1 to 15 , provides a detailed description of the embodiments of the present disclosure.
[0082] Figure 1 The perspective view of a three-dimensional semiconductor device includes multiple blocks and sub-blocks of memory cells in multiple vertical NAND strings. It includes multiple word line layers 11 of multiple stacks, which are configured as multiple word line layers of a first stack 21 and multiple word line layers of a second stack 22, each stack being parallel to a substrate 10, and multiple pillars 15, which are oriented vertically to the substrate in the Z direction as shown in the figure and extend through the corresponding multiple word line layers of the multiple stacks. The pillars include respective semiconductor bodies, which can be thin film channels with a thickness of less than 10 nm in multiple series-connected memory cells at multiple intersections between the multiple pillars and the multiple word lines in the NAND string configuration. As shown in the figure, multiple series select lines (SSL) 12 are oriented parallel to the substrate in the Y direction and are located above the multiple word line layers 11. In this example, a plurality of first and second blocks of a plurality of memory cells are formed in the first stack 21 and the second stack 22, respectively, each block being coupled to a different set of a plurality of NAND strings. Each of the plurality of string select lines intersects a respective different subset (e.g., one or more rows) of a plurality of sets of pillars in a corresponding block, wherein each sub-block of the plurality of memory cells in the corresponding block is formed in a subset of the plurality of pillars coupled to the corresponding string select line.
[0083] The structure also includes a plurality of parallel global bit line conductors 20, which are arranged parallel to the substrate extending in the X direction shown in the figure and are located above the string select lines. Each of the global bit line conductors is located above a corresponding column of a plurality of pillars in an array spanning a plurality of blocks, each column including one pillar in each subset of the plurality of pillars in each string select line.
[0084] Each intersection of a pillar and a string select line may define a select gate of the pillar for connection to a corresponding bit line. Each pillar is located under one of a plurality of bit line conductors through the select gate of the pillar, and each pillar is coupled to one of a plurality of bit line conductors.
[0085] A lower select line (lower SG) 13 is formed below the word line layer 11 to couple the pillars to one or more source lines, such as a common source conductor 18. Alternatively, the common source conductor may be connected to a bias circuit through multiple vertical connections between multiple blocks.
[0086] Figure 1 The structure of is an example of a non-volatile memory, which includes multiple blocks of multiple storage cells and multiple bit lines, each block including multiple word line sets (i.e. Figure 1 The invention relates to a memory device of the present invention, wherein each of the plurality of NAND strings in the plurality of blocks of memory cells has a plurality of sub-blocks, each sub-block including a different subset of the set of NAND strings of the block. In this example, different subsets of the plurality of NAND strings in each sub-block are operably connected to a respective sub-block string select line, and a plurality of gate voltages are applied to the plurality of string select gates of the plurality of NAND strings in the different subsets in the sub-block via the select line.
[0087] In such Figure 1 In the structure shown, an operation can be applied to limit the effect of the time-varying resistivity of the channel semiconductor material. For example, when programming the memory cell in some memory architectures, the channel polysilicon is stressed so that the threshold set by the program is based on the resistivity in the stressed state (stressed condition resistivities). To address this problem, a stress read bias can be applied to the stressed memory cell to be read. After the stress is read, the channel can maintain the stressed state for a period of time, such as about 10 minutes. Therefore, the memory cell can be read during this interval without the need for another stress read bias.
[0088] Figure 2Schematic diagram of a block of a plurality of memory cells in a three-dimensional NAND device that may include a plurality of blocks, wherein one block includes a plurality of sub-blocks. In the schematic diagram, a plurality of global bit lines MBL0 to MBLn overlay an array of a plurality of NAND strings arranged in a plurality of rows and a plurality of columns. Each NAND string includes a series-connected string of a plurality of memory cells (e.g., dielectric charge trapping memory cells) between corresponding bit lines and auxiliary lines (e.g., shared source lines CSL). In some embodiments, the shared source lines for the blocks may be implemented as one or more auxiliary lines and may be coupled to bias circuits through which a plurality of operating voltages are applied during various operations of the memory.
[0089] In a three-dimensional NAND configuration, such as Figure 2 The plurality of NAND series of blocks shown correspond to, for example, Figure 1 In this example, among all NAND strings in a block, multiple NAND strings in the multiple NAND strings are coupled to corresponding stacked multiple word lines WL0 to WLn+k, wherein each word line is coupled to multiple memory cells at its layer. At word line WLn, the planar structure of each word line layer is represented by dotted line 201. Therefore, all memory cells in the block at a given word line (e.g., WLn) level in the block are coupled to the given word line (e.g., WLn), so that they can be activated by the voltage applied to the given word line.
[0090] In addition, each NAND string includes a corresponding sub-block string select gate (eg, sub-block string select gate 202) configured to connect the NAND string to a specific bit line (eg, bit line 203) among the plurality of bit lines.
[0091] Multiple sub-block string selection lines SSL0 to SSLn are operably coupled to multiple string selection gates of multiple NAND strings of respective different subsets, wherein each subset of multiple NAND strings includes a sub-block of a block of multiple memory cells, so as to apply multiple gate voltages to the multiple sub-block string selection gates.
[0092] In addition, each NAND string includes a corresponding lower select gate configured to connect the NAND string to a shared source line or one of one or more auxiliary lines for implementing a shared source line. In this example, the lower select gate layer GSL is coupled to all lower select gates for multiple NAND strings in a block. In another example, multiple lower select gate lines can be configured to connect to the lower select gates in a block.
[0093] In this example, the lower dummy word line DWLG is located between the lower selection gate layer GSL and the bottommost word line layer WL0, and the upper dummy word line layer DWLS is located between the series selection lines SSL0 to SSLn and the topmost word line layer WLn+k.
[0094] exist Figure 2 In a circuit of the type described above, in order to select a specific memory cell in a block, a sub-block is activated by a sub-block string select line, wherein the sub-block string select line connects each NAND string in the selected sub-block to a corresponding bit line among the plurality of bit lines, and the word line layer selects a memory cell at the level of the selected word line on each NAND string in the selected sub-block. The selected memory cell is activated by selecting a bit line corresponding to the NAND string in which the selected memory cell is located. This configuration is capable of activating a plurality of memory cells, one NAND string in each NAND string in the selected sub-block in parallel via their corresponding bit line and word line layers.
[0095] As used in this disclosure, the term "activate" refers to applying a specific bias voltage to act on multiple connected memory cells or multiple switches. Depending on the operation and memory design, the bias voltage can be high or low. For descriptive purposes, the term "charging" refers to both driving a node to a higher voltage and driving a node to a lower voltage, including ground voltage and negative voltage in some embodiments.
[0096] The NAND blocks described in the present disclosure may be implemented using three-dimensional NAND memory technology. It may also be implemented using two-dimensional NAND technology, where a NAND block is logically defined across multiple two-dimensional NAND arrays.
[0097] Figure 3 The structure of a sense amplifier and a bit line bias circuit is shown, which can be used to apply multiple bias voltages to each bit line of a plurality of bit lines. The page buffer can include a sense amplifier and a bit line bias circuit, each bit line is coupled to a selected block of the array.
[0098] Figure 3The circuit in is connected to the global bit line 320. The bit line select transistor 318 has a first source / drain terminal (connected to the global bit line 320) and a second source / drain terminal. The gate of the bit line select transistor 318 is connected to the control signal BLS on the line 319. The bit line clamp transistor 321 has a first source / drain terminal connected to the transistor 318 and a second source / drain terminal connected to the connection node 323. The gate of the bit line clamp transistor 321 is connected to the BLC1 line 322, and during the uncharged circuit and other operations, the circuit not shown applies multiple bias voltages on this BLC1 line 322 to control the voltage level of MBL. Transistor 327 is used to connect node 323 to the BLC2 line, and multiple bias voltages are applied to the BLC2 line through the circuit not shown. The transmission transistor 335 is connected between the connection node 323 and the sensing node 332.
[0099] The transfer transistor 335 is controlled by a control signal BLC3, which controls the connection and disconnection of the connection node 323 and the sense node 332. The transistor 336 is connected between the sense node 332 and the bias voltage VGW2 and is controlled by a signal BLC4. The capacitor 337 (capacitor) is coupled from the sense node 332 to the sense signal node 338. The sense transistor 339 has a gate connected to the sense node 332, a first current carrying terminal connected to the sense pulse node 338, and a second current carrying terminal providing a sense amplifier output, wherein the sense amplifier output is connectable to a plurality of latches of the page buffer.
[0100] During read operations and other operations, transistors 318, 327, and 321 are operable to set a bias voltage level on a selected bit line to suit a particular operation.
[0101] Figure 4 and Figure 5 Segmentation of a memory array on a memory device at a plane, block, and sub-block level is shown according to which the techniques described in the present disclosure may be applied. The described techniques include applying multiple read setup operations that include applying multiple bias voltages simultaneously to multiple memory cells to condition the multiple memory cells for a subsequent read operation. This conditioning may condition the memory cells to match or approximate the threshold voltage established during the programming operation described above.
[0102] Figure 4The configuration of a single plane 400 in a memory array is shown. Plane 400 includes a plurality of blocks, block 0, block 1, ... block (b-1), and block (b). Each block includes a plurality of sub-blocks. Thus, block 0 includes sub-blocks 00 to 0n, block 1 includes sub-blocks 10 to 1n, block (b-1) includes sub-blocks (b-1)0 to (b-1)n, and block b includes sub-blocks (b)0 to (b)n.
[0103] A plurality of global bit lines 412 (MBL) are overlaid on all blocks in the plane and are shared with all blocks in the plane. A set of multiple sense amplifiers and multiple bit line bias circuits 405 (e.g., FIG. 3 ) (which may be part of a page buffer circuit) are coupled to the multiple global bit lines 412 to support multiple read set operations by applying multiple bias voltages to the multiple global bit lines 412. The set of multiple sense amplifiers and multiple bit line bias circuits 405 are shared by all blocks in the plane. Each block includes a corresponding string select line SSL and word line WL driver 410, 411, 413, 414, through which multiple bias voltages can be applied to support multiple read set operations. Similarly, a shared source line driver can be applied to each block.
[0104] In some embodiments, a read-set operation may be applied to only one block in a given plane. In other embodiments, a read-set operation may be applied to multiple blocks in a given plane simultaneously. In other embodiments, a read-set operation for a block having multiple sub-blocks numbered "n" may be applied to more than one and less than multiple sub-blocks numbered "n" simultaneously. In other embodiments, a read-set operation may be applied to one or more sub-blocks in one block of a plane and one or more sub-blocks in another block simultaneously.
[0105] Figure 5 A memory 500 is shown that includes multiple planes (Plane 0, Plane 1, Plane 2, and Plane 3) in this example. Each plane includes different page buffer circuits, including Page Buffer 0, Page Buffer 1, Page Buffer 2, Page Buffer 3. The page buffers are coupled to input / output circuits not shown, thereby supporting multiple high throughput memory operations on multiple planes. As shown, each plane includes multiple blocks. Plane 0 includes Block 00, Block 01, Block 02, Block 03... Plane 1 includes Block 10, Block 11, Block 12, Block 13... Plane 2 includes Block 20, Block 21, Block 22, Block 23... Plane 3 includes Block 30, Block 31, Block 32, Block 33...
[0106] Please refer to Figure 5 , a read-set operation may be applied to a block or blocks in a single plane. Likewise, in some embodiments, a read-set operation may be applied to a block or blocks in a plane and simultaneously to a block or blocks in another plane. Likewise, a read-set operation may be applied to one or more sub-blocks in a block in a plane and simultaneously to one or more sub-blocks in a block in another plane. Likewise, multiple read-set operations may be applied to multiple other read-set units, in addition to sub-block, block, and plane units, to suit a particular memory configuration.
[0107] Figures 6 to 11 A read setup bias configuration in an alternative embodiment is shown that will be applied to multiple parallel memory cells simultaneously during multiple read setup operations described in the present disclosure. The bias configuration applied during the read setup operation can be determined by trial and error or by simulation based on the memory structure of the memory array. Generally speaking, the bias configuration applied during the read setup operation should be set to prevent the amount of data loss from being disturbed in the memory cell. Generally speaking, such a bias device will have a voltage on the order of the voltage applied during a normal read operation.
[0108] exist Figures 6 to 11 In all figures, refer to Figure 2 and Figure 3 , describing the global bit line MBL (VMBL), the string select lines SSLs (VSSL), the word lines WLs (VWL), the bottommost select gate lines GSLs (VGSL) and the common source line CSL (VCSL) applied to a selected block. The voltage on the global bit line MBL (VMBL) can be controlled in response to the control signals VBL (VVBL), BCL1 (VBCL1) and BCL2 (VBCL2) described above.
[0109] exist Figure 6In the example shown, all bias voltages are raised from ground to corresponding read set bias levels (not shown to scale) at approximately the same time (610). The plurality of voltage pulses applied to the plurality of global bit lines MBL (VMBL), the plurality of word lines WLs (VWL), and the shared source line CSL (VCSL) terminate at time 611. The plurality of voltage pulses applied to the plurality of string select lines SSLs (VSSL) and the bottommost select gate line GSL (VGSL) terminate shortly thereafter at time 612. Therefore, the falling edge 600 of the voltage on the plurality of word lines WLs (VWL) precedes the falling edges 601 and 602 of the plurality of voltages on the aligned plurality of string select lines SSLs (VSSL) and the plurality of bottom select gate lines GSL (VGSL).
[0110] In this example, VSSL and VGSL are set to levels that turn on the string select gates to allow current to flow between the global bit lines and the shared source line. In other embodiments, the levels are set to V'SSL and V'GSL to turn off the string select gates. In some embodiments, one of the VSSL and VGSL voltages is set to turn on the string select gates, the other of the VSSL and VGSL voltages is set to turn off the string select gates, and only the WL voltage is applied to stress the channel during a read set operation. Similarly, in some embodiments, the shared source line can be set to V'CSL (which can be ground or another low reference voltage) to establish a voltage drop across the corresponding multiple NAND strings to encourage or prevent current from flowing through the channel during a read set operation.
[0111] Furthermore, in some embodiments, the applied voltage pulses may not begin simultaneously (need not all begin at time 610 ).
[0112] Figure 7 An alternative embodiment of a read and set operation is shown for use with Figure 6 The same signal bias configuration. Figure 7In an embodiment of the present invention, the same bias voltages may be applied as described above. All bias voltages are raised from ground to a read set bias level (710) at approximately the same time. However, in this embodiment, the multiple voltage pulses applied to the multiple global bit lines MBL (VMBL), the multiple string select lines SSLs (VSSL), the multiple bottom select gate lines GSLs (VGSL), and the shared source line CSL (VCSL) terminate at time 711. The multiple voltage pulses applied to the multiple word lines WLs (VWL) terminate shortly after time 712. Therefore, the falling edge 700 of the voltage on the multiple word lines WLs (VWL) is later than the falling edges 701 and 702 of the multiple voltages on the aligned multiple string select lines SSLs (VSSL) and the multiple bottom select gate lines GSLs (VGSL).
[0113] Figure 8 For similar Figure 6 Timing diagram of . Figure 8 and Figure 6 The bias configuration differs in that the voltage (800) applied to the plurality of bit lines MBL (VMBL) is reduced to the voltage applied during precharge for a read operation on the NAND string, thereby reducing the current generated during the read setup relative to the average read current.
[0114] Fig. 9 For similar Figure 6 Timing diagram of . Fig. 9 and Figure 6 The bias configuration of FIG. 8 is different in that the voltage 900 applied to the plurality of word lines WLs (VWL) is reduced compared to a pass voltage applied to a word line during a read operation.
[0115] Fig.10 For similar Figure 6 Timing diagram of . Fig.10 and Figure 6 The bias configuration differs in that the voltage on the shared source line (as shown by voltage 1000) can be adjusted to set a magnitude of the voltage drop across the NAND string relative to the corresponding global bit line so that it is different from the voltage drop applied during a read operation, and in some embodiments is reduced relative to the voltage drop applied during a read operation.
[0116] Fig.11 For similar Figure 6 Timing diagram of . Fig.11 and Figure 6The bias configuration of FIG. 1 is different in that the voltages applied to the plurality of bit lines MBL (VMBL) and the voltages applied to the plurality of word lines WL (VWL) have reduced durations as shown by line segments 1101 and 1102. In other aspects, the control signals have Figure 6 The time points 1110, 1111, 1112 discussed show a number of timing characteristics.
[0117] Fig.12 For similar Figure 6 Timing diagram of . Fig.12 and Figure 6 The difference between the bias configuration of FIG. 1 and FIG. 2 is that the voltage 1201 applied to the plurality of bit lines MBL (VMBL) is maintained at the ground voltage. In other aspects, the control signal has the following characteristics: Figure 6 Multiple time features are shown.
[0118] Although in some techniques, a read setup operation may be performed by performing a normal read operation of multiple memory cells within a specific period of time prior to an immediate read operation, in the techniques described in the present disclosure, a setup read operation may be applied in a systematic or periodic manner to cover the entire memory array including one or more planes on a single chip. In the techniques applied in the present disclosure, multiple NAND strings coupled to a single bit line may be turned on simultaneously, thereby facilitating an increase in current on multiple bit lines. As described in the present disclosure, during the read setup operation, a bias configuration may be applied to reduce the peak current on multiple bit lines. Thus, the read setup operation may have a lower drain side bias compared to a normal read. Similarly, the read setup operation may have a lower word line bias compared to a normal read. Similarly, the read setup bias operation may have a higher common source line bias compared to a normal read. Similarly, the read setup operation may have a shorter word line and global bit line setup time to speed up the read setup operation.
[0119] Fig.13is a simplified diagram of a memory system including a flash memory device 1208 implemented on an integrated circuit and a host 1202 including logic for issuing commands, such as read commands, and programming instructions with addresses and data to be programmed. In some embodiments, the host may issue a read set command to initiate a read set operation on the memory device 1208. The memory device 1208 may be implemented on a single integrated circuit chip, a polysilicon module, or multiple chips to be configured to suit a particular need.
[0120] The memory device 1208 in this example includes a memory array 1278 that includes multiple blocks as described above on an integrated circuit substrate, each block having multiple sub-blocks. The memory array 1278 may be a NAND flash memory implemented using two-dimensional or three-dimensional array technology.
[0121] In various embodiments, the memory device 1208 may have a plurality of single-level cells (SLC) or a plurality of multiple-level cells (eg, MLC, TLC, or XLC) storing more than one bit per memory cell.
[0122] Memory device 1208 includes a memory array 1278, which may be NAND flash memory implemented using three-dimensional array technology having one or more planes, each plane having multiple blocks, and each block having multiple sub-blocks.
[0123] The word line decoder 1276A is coupled to a plurality of word lines 1277 in a memory array 1278 through a plurality of word line driver circuits 1276B. The SSL / GSL decoder 1278A is coupled to the bit line side (SSL) and shared source side (GSL) series select gates in the array through SSL / GSL driver circuits 1278B through SSL and GSL lines 1279. A plurality of page buffer circuits 1238 are coupled to a plurality of bit lines 1266 in the memory array 1278 through a plurality of bit line driver circuits 1248. In some embodiments, a plurality of column decoder circuits may be included to route data from a plurality of bit line drivers to a plurality of selected bit lines. The plurality of page buffer circuits 1238 may store a plurality of data pages that define a data pattern for a page programming operation and may include a plurality of sensing circuits used in read and verify operations.
[0124] The plurality of bit lines for the plurality of memory arrays may include a plurality of global bit lines (GBL) and a plurality of local bit lines (LBL). The plurality of bit lines typically include a plurality of metal conductors in a plurality of higher patterned layers that cover a plurality of memory cell blocks in an array. The plurality of global bit lines are connected to a plurality of NAND strings to allow current to flow into and out of the plurality of bit lines, which in turn are connected to a plurality of bit line driver circuits 1248 and a plurality of page buffer circuits 1238. Similarly, the plurality of word lines may include a plurality of global word lines and a plurality of local word lines with corresponding plurality of word line driver circuits 1276B in a plurality of word line drivers.
[0125] In a sensing operation, sensing data from the plurality of page buffer circuits 1238 is provided via a second data line in the bus system 1226 to the plurality of cache memory circuits 1228, which in turn are coupled to the plurality of input / output circuits 1218 via a plurality of data path lines 1216. In this example, input data is applied to the plurality of cache memory circuits 1228 on the plurality of lines 1216 and the plurality of page buffer circuits 1238 on the bus system 1226 for supporting program operations.
[0126] The plurality of input / output circuits 1218 are connected via lines 1214 (including I / O pads) and provide a plurality of communication paths for data, addresses, and commands having destinations external to the memory device 1208 (including the host 1202 in this example). The plurality of input / output circuits 1218 provide a communication path via lines 1216 to a plurality of cache memory circuits 1228 that support memory operations. The plurality of cache memory circuits 1228 communicate data streams with a plurality of page buffer circuits 1238 (e.g., using a bus system 1226).
[0127] The plurality of control circuits 1234 are connected to the plurality of input / output circuits 1218 and include instruction decoder logic, address counters, a plurality of state machines, a plurality of timing circuits, and a plurality of other logic circuits that control various memory operations, including programming, reading, and erasing operations for the memory array 1278. The plurality of control circuit signals are distributed to the plurality of circuits in the memory device (as indicated by arrows 1245, 1246) to support the plurality of operations of the plurality of circuits. The plurality of control circuits 1234 may include a plurality of address registers, etc., for transmitting addresses to the plurality of elements of the memory device 1208 as needed, including to the cache circuit 1228 and transmitted on the line 1244 to the page buffer circuit 1238, the word line decoder 1276A, and the SSL / GSL decoder 1278A (in this example).
[0128] exist Fig.13 In the example shown, the plurality of control circuits 1234 include a plurality of control logic circuits including implementing a plurality of modules of one or more bias configuration state machines (or a plurality of bias configuration state machines) in a plurality of word line driver circuits 1276B and a plurality of bit line driver circuits 1248 for a plurality of selectable sets of programming, read setup, and read operations, the one or more bias configuration state machines controlling the application of a plurality of bias voltages generated or provided by one or more voltage supplies in block 1264, including read setup, read, erase, verify, and a plurality of programming voltages (including a plurality of precharge voltages, a plurality of pass voltages, a plurality of other bias voltages as described in the present disclosure). The plurality of bias voltages are applied to a plurality of elements of the memory device 1208 as indicated by arrows 1265 as required to support the operations.
[0129] As is known in the art, the plurality of control circuits 1234 may include a plurality of modules implemented using a dedicated logic circuit including a plurality of state machines. In an alternative embodiment, the plurality of control circuits 1234 may include a plurality of modules implemented using a general purpose processor, which may be implemented on the same integrated circuit, which executes a computer program to control the plurality of operations of the memory device 1208. In other embodiments, a combination of dedicated logic circuits and general purpose processors may be used for the implementation of the plurality of modules in the plurality of control circuits 1234.
[0130] The flash memory array 1278 may include a plurality of floating gate memory cells or a plurality of dielectric charge trapping memory cells configured to enable each memory cell to store multiple bits by establishing multiple program levels corresponding to the amount of stored charge, which in turn establishes multiple memory cell threshold voltages Vt.
[0131] This technology can be used for single-bit-per-cell flash memory, as well as other multiple-bit-per-cell and single-bit-per-cell memory technologies. In other examples, the plurality of memory cells can include a plurality of programmable resistive memory cells, a plurality of phase change memory cells, and other types of non-volatile and volatile memory cell technologies.
[0132] In the example shown, the host 1202 is coupled to the line 1214 on the memory device 1208 and multiple other control terminals not shown, such as multiple chip select terminals, etc., and can provide multiple instructions or instructions to the memory device 1208. In some examples, the host 1202 can be coupled to the memory device using serial bus technology (using shared address and data lines). The host 1202 may include a general-purpose processor, a special-purpose processor, a processor configured as a memory controller, or other processor using the memory device 1208. All or part of the host 1202 can be implemented on the same integrated circuit as the memory.
[0133] The host 1202 may include one or more file systems that store, retrieve, and update data stored in the memory based on requests from applications. In general, the host 1202 may include multiple programs that perform memory management functions (including, in some embodiments, functions to control or support multiple read and set operations as described in the present disclosure). Other memory management functions may include, for example, wear leveling, bad block recovery, power loss recovery, garbage collection, error correction, etc. Similarly, the host 1202 may include multiple applications, multiple file systems, multiple flash translation layer programs, and multiple other components (which may generate status information for data stored in the memory, including issuing instructions to program data having addresses and to be programmed).
[0134] exist Fig.13 In the example shown, the memory device includes a set of status registers 1235 for storing a plurality of parameters for a read set operation. The plurality of parameters may define a voltage level to be applied to determine whether to turn on or off a plurality of series select and ground select gates, a pulse duration, etc. Figures 6 to 11 In addition, the plurality of parameters may include a starting plane block address and a block address range (or address of other read setup units) for a particular read setup operation. The plurality of parameters may include a plurality of indicators for simultaneously enabling read setup operations for a plurality of planes, a plurality of blocks within a plurality of planes, and a plurality of sub-blocks within a plurality of blocks. Some or all of the parameters may be provided via a plurality of read setup instructions, and some or all of the parameters may be stored on-chip as configuration data.
[0135] A state machine on a memory device may access a plurality of read setup parameters and perform a read setup operation including address generation and applying a plurality of bias voltages to encompass a memory array to maintain a read ready status across the memory. The operation may include a start block and an end block. The operation may include a mode of a plurality of blocks in one or more planes that may be simultaneously subjected to a read setup operation. The operation may be configured to encompass an array or a portion of an array as a background operation without external control. The operation may be configured to operate in response to a plurality of read setup instructions and carry a plurality of read setup parameters and identify a plurality of segments of the array to be operated by the read setup operation. The plurality of read setup instructions may be generated, for example, by a memory controller in a host that monitors a plurality of block states (e.g., by identifying a plurality of cold blocks as a plurality of stale blocks in a wear leveling operation) and may send a plurality of instructions identifying stale blocks, or may send a plurality of instructions during a period when the memory array is idle or expected to be idle. The state machine may set a ready / busy pin on the memory device to send a signal to the control program on the storage controller to coordinate the read set operation.
[0136] Fig.14 Flowchart of a representative method for operating a NAND memory including a plurality of blocks of a plurality of memory cells arranged in NAND strings, each block including a distinct set of NAND strings. As described above, each block may include a plurality of sub-blocks, each sub-block including a distinct subset of the plurality of NAND strings.
[0137] exist Fig.14In the embodiment of the present invention, the read setup operation begins at block 1300. The operation first identifies a stale block as a target for the read setup operation (block 1301). The operation may identify multiple stale blocks by accessing a block status table 1350, wherein the block status table 1350 indicates a status of multiple blocks in the memory as stale (which means that they are marked for a read setup operation) or ready (which means that a read setup operation can be performed on them without being read). Similarly, the operation may identify multiple stale blocks as multiple parameters carried in a read setup instruction received from an external controller. If there are no stale blocks for read setup, the algorithm loops to wait for the indication of stale blocks, or the algorithm may terminate and restart after a predetermined interval. Next, for the identified stale blocks, a read setup bias sequence is applied (block 1302). The read setup bias sequence may be configured in response to multiple read setup bias parameters stored in multiple status registers 1351. After applying the read setup bias sequence to the obsolete block, the block status table 1350 is updated to indicate that the block is ready (block 1303). Next, the operation determines whether there are more obsolete blocks (block 1304). If not, the algorithm ends at 1305. If there are more obsolete blocks at block 1304, the algorithm loops back to block 1302 to continue applying multiple read setup bias sequences to multiple obsolete blocks.
[0138] The controller managing the read set operation may also perform a state update operation (block 1310). The state update operation may monitor the block state table 1350 to identify a plurality of blocks newly set as ready (block 1311). If a newly set ready block is identified, the operation begins tracking the lifetime of the ready state of the block (block 1312). If the ready block becomes obsolete (block 1313), for example, the ready state has been set (e.g., has been read or has undergone a read set operation) and its lifetime is longer than a certain duration (e.g., 10 minutes), the controller may update the ready block state in the block state table 1350 to obsolete (block 1314). After the state is updated, the operation of the new ready state block is completed. The loop of steps (blocks) 1310 to 1315 may be performed concurrently for all blocks in the block state table. In an alternative system, all modules in the system are covered and updated to an obsolete state periodically so that the read set operation is systematically applied across the entire array without the need to monitor the read state of each module separately. In some embodiments, a block status table indicating obsolete blocks is not required, but rather the state needs to be maintained on-chip or in the host, with instructions that are independent of the order covered in the system that reads the array of instructions.
[0139] Fig.14An example of an operating method includes covering multiple blocks to apply multiple read setting bias configurations to multiple obsolete blocks among the multiple blocks, which condition multiple read operations of the multiple obsolete blocks, wherein the multiple read setting bias configurations include simultaneously applying read setting biases to multiple memory cells of a selected block among the multiple blocks.
[0140] Likewise, Fig.14 is an example of an operating method that can operate independently of a read instruction, such as in the background.
[0141] Fig.15 2 shows a read operation for a system similar to that described in the present disclosure. Fig.15 In the method, a read instruction is received at block 1400 to initiate a read operation. The algorithm determines whether the read is directed to an obsolete block by, for example, accessing a block status table 1450 (block 1401). If not, the read operation may continue to execute the instruction for the read target (block 1404). If the block is an obsolete block at step (block) 1401, the read operation causes a read set bias sequence to be performed on the obsolete block (block 1402), by setting bias parameters 1451 to apply multiple bias voltages with multiple setting parameters. After applying the read set bias sequence, the state of the block is updated in the block status table 1450 (block 1403). In addition, the read target of the instruction is executed (block 1404). After executing the target of the instruction, the operation terminates (block 1405).
[0142] Fig.14 and 5 Fig.14 As an example of operations that can be used to perform a read setup process as described in the present disclosure, multiple bias configurations are applied in parallel or simultaneously to multiple memory cells, such as to multiple memory cells coupled to a single bit line, such as to all memory cells in a sub-block, such as to all memory cells in a block, such as to all memory cells in multiple sub-blocks, or such as to all memory cells in multiple blocks.
[0143] Fig.14 and Fig.15A flow chart showing the logic executed by a memory controller or memory device. This logic can be implemented using a processor that is programmed using a computer program stored in a memory accessible to a computer system and executable by the processor, which can be executed by a combination of dedicated logic hardware including field programmable integrated circuits and dedicated logic hardware and computer programs. For all flow charts in the present disclosure, it will be recognized that many steps can be combined, executed in parallel, or executed in a different order without affecting the functions implemented. In some cases, as the reader will understand, the rearrangement of steps can only achieve the same results if certain other changes are made. In other cases, as the reader will understand, the rearrangement of steps can achieve the same results only if certain conditions are met. In addition, it should be understood that the flow charts of the present disclosure only show steps relevant to understanding the present disclosure, and it should be understood that many other steps for implementing other functions can be performed before, after, and between those shown.
[0144] The present disclosure describes a technique that can perform read and set operations at high speed and more frequently than is available in the prior art, thereby improving the memory cell operating window by keeping the memory cell under the condition of a threshold value set during the programming operation. These techniques are particularly beneficial in large, high-density memory systems. For example, if there are multiple sub-blocks in a block, the read and set operations can be performed on all sub-blocks in a block at the same time to increase the operating speed. In addition, if there are multiple blocks in a memory plane, the read and set operations can be performed on multiple blocks at the same time to increase the operating speed. In addition, if there are multiple planes in a memory device, the read and set operations can be applied to multiple blocks or multiple sub-blocks in multiple planes at the same time to increase the operating speed.
[0145] In other embodiments, this technology can also be used to place memory cells in a NOR structure with local bit lines having controllable block select gates that can control the connection or disconnection of columns of memory cells in the block to corresponding bit lines in multiple bit lines.
[0146] Although the preferred embodiments and examples described in detail above disclose the present disclosure, it should be understood that these examples are exemplary only and not restrictive. It is conceivable that those skilled in the art will easily think that modifications and combinations will fall within the scope of the spirit of the present disclosure and the scope of the appended claims.
Claims
1. A non-volatile memory comprising: a plurality of blocks, wherein a plurality of blocks of the plurality of blocks are configured to be enabled for memory operations in response to a plurality of word lines of a corresponding plurality of groups; a plurality of control circuits including logic to perform a read-set operation including simultaneously applying a read-set bias voltage to a plurality of memory cells of a selected block of the plurality of blocks; as well as the plurality of blocks of memory cells and a plurality of bit lines, each block being configured to include a plurality of sets of NAND strings having a plurality of string select gates for connection to a corresponding plurality of bit lines of the plurality of bit lines, wherein each NAND string of the plurality of sets of NAND strings of the block is connected to a plurality of word lines of the group; wherein each of the plurality of blocks of the plurality of memory cells has a plurality of sub-blocks, each sub-block comprising a different subset of the plurality of NAND string sets of the block, the plurality of NAND strings of the different subset of each sub-block being operably connected to a corresponding sub-block string select line, and a plurality of gate voltages being applied to the plurality of string select gates of the plurality of NAND strings of the different subset of the sub-block via the select line; The read setup bias includes a bias configuration including bias voltages on at least one sub-block series select line in a selected block and on the multiple word lines in the multiple word line set of the selected block; the bias configuration includes bias voltages on the multiple word lines in the set having a shorter duration than multiple pass voltages applied in a read operation.
2. The nonvolatile memory of claim 1, comprising logic for covering the blocks in the plurality of blocks to apply the read set operation to the plurality of blocks.
3. The nonvolatile memory according to claim 1, wherein: The multiple blocks in the multiple blocks include multiple sub-blocks, and the read setup operation includes covering the multiple sub-blocks in the selected block to simultaneously apply the read setup bias to the corresponding multiple memory cells in one or more sub-blocks of the multiple sub-blocks in the selected block.
4. The nonvolatile memory according to claim 1, wherein: The nonvolatile memory includes a plurality of memory planes, each of the plurality of memory planes includes a plurality of blocks, and logic for simultaneously performing the read and set operation on each selected block in more than one memory plane of the plurality of memory planes.
5. The nonvolatile memory according to claim 1, wherein: The multiple blocks in the multiple blocks respectively include multiple sub-blocks, wherein the multiple sub-blocks in the multiple sub-blocks are configured to be connected to multiple global bit line sets for multiple storage operations in response to corresponding multiple sub-block string selection lines, and a sub-block includes multiple NAND strings, and the read set operation includes covering the multiple sub-blocks in the selected block to simultaneously apply the read set bias to multiple memory cells corresponding to each of the multiple sub-blocks of the selected block. 6 . The nonvolatile memory according to claim 1 , comprising a plurality of thin film channels corresponding to the plurality of memory cells, the read setup bias voltage regulating the plurality of thin film channels for a subsequent read operation.
7. The nonvolatile memory of claim 1, comprising logic for initiating the read-set operation for a plurality of blocks having an outdated block status and updating a block status table according to the read-set operation.
8. The nonvolatile memory according to claim 1, wherein: The bias configuration includes bias voltages on the word lines in the set that are lower than pass voltages applied in a read operation.
9. The nonvolatile memory according to claim 1, wherein: The bias arrangement includes a plurality of bias voltages on at least one sub-block string select line in a selected block to close the sub-block string select gate.
10. The nonvolatile memory according to claim 1, wherein: The bias arrangement includes bias voltages on at least one sub-block string select line in a selected block having shorter durations than sub-block string voltages applied in a read operation.
11. The nonvolatile memory according to claim 1, wherein: The bias arrangement includes bias voltages on the bit lines and on a shared source line for the selected block.
12. The nonvolatile memory according to claim 11, wherein: The bias configuration includes bias voltages on the bit lines for the selected block that are lower than bit line voltages applied in a read operation.
13. The nonvolatile memory according to claim 1, wherein: The bias configuration includes a plurality of bias voltages on the plurality of bit lines and on a shared source line for the selected sub-block, wherein the voltage applied to the shared source line is configured to reduce a voltage difference between the shared source line and the corresponding plurality of bit lines relative to a voltage difference applied in a read operation in the plurality of NAND strings of the selected sub-block.
14. The nonvolatile memory according to claim 1, wherein: The biasing arrangement includes bias voltages on the bit lines and the sub-block string select line for closing the string select gate.
15. An operating method of a non-volatile memory according to any one of claims 1 to 14, wherein the non-volatile memory comprises a plurality of blocks of a plurality of memory cells and a plurality of bit lines, the method comprising: The plurality of blocks are covered to apply a plurality of read setup bias configurations to a plurality of obsolete blocks among the plurality of blocks, which regulate a plurality of read operations of the plurality of obsolete blocks, the plurality of read setup bias configurations comprising simultaneously applying a read setup bias to a plurality of memory cells of a selected block of the plurality of blocks.
16. The method of claim 15, comprising maintaining a block status table for identifying a plurality of obsolete blocks.
17. The method according to claim 15, wherein: Each of the plurality of blocks of a plurality of memory cells includes a plurality of sub-blocks, each sub-block being configured to include a different subset of the plurality of NAND string sets of the block, the plurality of NAND strings of the different subset in each sub-block being operably connected to a corresponding sub-block string select line, a plurality of gate voltages being applied to the plurality of string select gates of the plurality of NAND strings of the different subset in the sub-block via the select line, wherein the plurality of read set bias operations are simultaneously applied to one or more sub-blocks.
