Ramp-down sampling between program voltage and test voltage in a storage device
By repeatedly sampling threshold voltages and adjusting programming settings, the solution optimizes memory device performance by reducing overprogramming and enhancing read/verification accuracy in semiconductor memory devices.
Patent Information
- Application Number
- DE112019000159
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2018-04-13
- Filing Date
- 2019-02-07
- Publication Date
- 2024-12-05
- Estimated Expiration
- 2039-02-07
AI Technical Summary
Existing semiconductor memory devices face challenges in optimizing programming performance due to changes over time, leading to issues like overprogramming and read/verification errors, particularly in NAND chains where unselected memory cells cannot be fully turned on during scanning.
The solution involves evaluating the threshold voltage of memory cells by repeatedly sampling between programming and verify voltages, adjusting programming settings based on the detected tails of the threshold voltage distribution, and storing these settings in spare NAND strings for future optimizations.
This approach minimizes additional sampling time, reduces overprogramming, and enhances programming efficiency by adapting settings to the changing performance of memory cells, thereby improving read and verification accuracy.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
BACKGROUND
[0001] The present technology relates to the operation of storage devices.
[0002] Semiconductor memory devices have become increasingly popular for use in various electronic devices. For example, non-volatile semiconductor memory is used in mobile phones, digital cameras, personal digital assistants, mobile computing devices, non-mobile computing devices, and other devices.
[0003] A charge storage material, such as a floating gate or a charge trapping material, can be used in such memory devices to store a charge representing a data state. A charge trapping material can be arranged vertically in a three-dimensional (3D) stacked memory structure or horizontally in a two-dimensional (2D) memory structure. An example of a 3D memory structure is the Bit-Cost Scalable (BiCS) architecture, which features a stack of alternating conductive and dielectric layers.
[0004] A memory device includes memory cells that may be arranged in series, for example, in NAND strings (e.g., NAND strings), with select gate transistors provided at the ends of a NAND string to selectively connect a channel of the NAND string to a source line or a bit line. However, several challenges arise in the operation of such memory devices.
[0005] US 2008 / 0 316 820 A1 relates to a method for programming a memory device in which multiple verification operations are performed sequentially after applying a programming voltage. US 6 671 207 B1 relates to a method for programming memory cells, which comprises generating a programming-verify pulse with stepped sections. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a block diagram of an exemplary memory device. Fig. Figure 2 is a block diagram illustrating one embodiment of a sampling block 51 of Fig. 1 represents. Fig. 3 illustrates an exemplary implementation of the power control module 116 of Fig. 1 for providing voltages to blocks of memory cells. Fig. Figure 4A shows an example of map 119 of Fig. 1. Fig. Figure 4B illustrates an example of programming setting(s) stored in a memory location referenced in card 119 of Fig. 4A is referred to. Fig. 5 is a perspective view of a memory device 500 including a set of blocks in an exemplary 3D device of the memory structure 126 of Fig. 1 includes. Fig. Figure 6A illustrates an exemplary cross-sectional view of a portion of block BLKO of Fig. 5. Fig. Figure 6B illustrates an example transistor 650. Fig. Figure 6C shows a detailed view of region 622 of the stack of Fig. 6A. Fig. Figure 7 shows an exemplary view of NAND chains in a block BLKO, which are Fig. 5 and Fig. 6A are consistent. Fig. Figure 8A shows the control gate layers in the block BLKO, which are Fig. 7 are consistent. Fig. Figure 8B shows an additional detail of the block BLKO of Fig. 7. Fig. 8C represents a subdivision of the block BLKO of Fig. 8B into a primary set of NAND chains 850 and a reserve set of NAND chains 851. Fig. 8D provides additional details of SB0 from Fig. 8B, which is Fig. 8C, including a primary set of NAND chains SB0p and a reserve set of NAND chains SB0s. Fig. 9A illustrates a threshold voltage (Vth) distribution of SLC memory cells in an erased (Er) and a programmed (P) state, including an example of overprogramming and sense voltages used to determine an amount of overprogramming. Fig. Figure 9B illustrates an example of the Vth distribution of a set of memory cells in eight data states. Fig. Figure 9C shows the threshold voltage distribution of memory cells programmed from the Er state to the A state, in accordance with Fig. 9B, and a sensing voltage V13 used to determine the amount of programming of the memory cells. Fig. Figure 9D shows the threshold voltage distribution of memory cells programmed from the Er state to the A state, in accordance with Fig. 9B, including sensing voltages VvA-V16, which is used to determine the amount of programming of the memory cells. Fig. Figure 10A illustrates a voltage signal used in a series of programming loops in an exemplary programming procedure leading to the Vth distribution of Fig. 9A leads. Fig. Figure 10B illustrates a voltage signal used in a series of programming loops in an exemplary programming operation resulting in the Vth distribution of Fig. 9B leads. Fig. 11A illustrates an exemplary programming process with an option to scan a set of memory cells while decreasing a voltage of a selected word line from a program voltage to a verify voltage. Fig. 11B illustrates an exemplary process for scanning a set of memory cells while decreasing the voltage of a selected word line, as in step 1104a of Fig. 11A. Fig. 11C illustrates an exemplary process for determining the updated programming setting(s) based on sensor data, as in step 1104b of Fig. 11A. Fig. 11D illustrates a graphical representation of an initial Vpgm versus the detected end of a Vth distribution of a set of memory cells, in an exemplary implementation of step 1112 in Fig. 11C. Fig. 11E illustrates a graphical representation of an initial Vpgm versus a loop count at which a set of memory cells reaches a programming milestone, in an example implementation of step 1114 of Fig. 11C. Fig. 11F illustrates a graphical representation of an intermediate voltage (Vint) versus a lower end of a Vth distribution of a set of memory cells, in an exemplary implementation of step 1119 of Fig. 11B. Fig. 12 illustrates an exemplary process for storing updated programming setting(s) in one or more reserve sets of NAND strings, beginning with step 1115 of Fig. 11C and with Fig. 11D and Fig. 11E matches. Fig. 13A illustrates exemplary voltage signals applied to word lines during a program loop, wherein the voltages on the selected and unselected word lines are reduced to 0 V between the program voltage and the verify voltages, as a comparison example to steps 1104 and 1104a of the programming process of Fig. 11A. Fig. Figure 13B illustrates exemplary voltage signals generated during a programming loop, in accordance with Fig. 11A and step 1104, to word lines, wherein the selected word line voltage is reduced from the programming voltage to the lowest verify voltage of a plurality of verify voltages without performing a ramp-down scan, in accordance with Fig. 9B. Fig. Figure 13C illustrates exemplary voltage signals applied to word lines during a program loop, in accordance with Fig. 11A and step 1104a, wherein the selected wordline voltage is decreased from the program voltage to a verify voltage while performing a ramp-down scan to detect a lower tail of a Vth distribution, in accordance with Fig. 9A, voltages VvP-V5, or an upper tail of a Vth distribution, in accordance with Fig. 9D, voltages VvA-V16. Fig. Figure 13D illustrates exemplary voltage signals applied to word lines during a program loop, in accordance with Fig. 11A and step 1104a, wherein the selected wordline voltage is decreased from the programming voltage while performing a ramp-down scan to detect a lower tail of a Vth distribution or an upper tail of a Vth distribution, in accordance with Fig. 9D, voltages VvA-V16, and no verification test is performed or the final sampling result is used as the verification result, in accordance with Fig. 9A, voltages VvP-V5. Fig. 13E illustrates exemplary voltage signals applied to word lines during a program loop, in accordance with Fig. 11A and step 1104a, wherein the selected wordline voltage is decreased from the program voltage to a verify voltage while performing a ramp-down scan to detect a lower tail of a Vth distribution, in accordance with Fig. 9A, voltages VvP-V5, or an upper tail of a Vth distribution, in accordance with Fig. 9D, voltages VvA-V6-V11. Fig. Figure 13F illustrates exemplary voltage signals applied to word lines during a program loop, in accordance with Fig. 11A and step 1104a, wherein the selected wordline voltage is reduced from the programming voltage to a verify voltage while performing a first ramp-down sampling sequence to detect a lower tail of a Vth distribution, in accordance with Fig. 9A, voltages V6-V11, and a second sequence of ramp-down sampling is performed to detect an upper tail of a Vth distribution, in accordance with Fig. 9A, voltages VvP-V5. DETAILED DESCRIPTION
[0006] According to the invention, an apparatus and a method having the features of the independent claims are provided; dependent claims relate to preferred embodiments. Apparatus and techniques for optimizing a programming operation in a memory device are described.
[0007] In some memory devices, memory cells are connected together in a block or sub-block, as in NAND strings. Each NAND string comprises a series of memory cells connected in series between one or more drain-end select gate transistors (called SGD transistors) on a drain end of the NAND string, which is connected to a bit line, and one or more source-end select gate transistors (called SGS transistors) on a source end of the NAND string, which is connected to a source line. Furthermore, the memory cells may be arranged with a common control gate line (e.g., word line) acting on a control gate. A set of word lines extends from the source side of a block to the drain side of a block. Memory cells may also be connected in other types of strings, as well as in other ways.
[0008] In a 3D memory structure, the memory cells can be arranged in vertical NAND strings in a stack, with the stack comprising alternating conductive and dielectric layers. The conductive layers act as word lines connected to the memory cells. Each NAND string can be shaped like a column that intersects with the word lines to form the memory cells.
[0009] The memory cells may include data memory cells suitable for storing user data, and dummy or non-data memory cells that are not suitable for storing user data. A dummy memory cell may have the same construction as a data memory cell, but is deemed by the controller to be unsuitable for storing any type of data, including user data. A dummy word line is connected to a dummy memory cell. One or more dummy memory cells may be provided at the drain and / or source end of a chain of memory cells to provide a gradual transition in the channel voltage gradient.
[0010] During a programming operation, the memory cells are programmed according to a wordline programming sequence. For example, programming may begin at the wordline on the source side of the block and continue at the wordline on the drain side of the block. In one approach, each wordline is fully programmed before a next wordline is programmed. For example, a first wordline, WL0, is programmed using one or more programming passes until programming is complete. Next, a second wordline, WL1, is programmed using one or more programming passes until programming is complete, and so on. A programming pass may include a set of increasing voltages applied to the wordline in respective programming loops or programming verify passes, as in Fig. 10B. Verification operations may be performed after each program voltage to determine whether the memory cells have completed programming. Once programming for a memory cell is complete, it may be excluded from further programming while programming for other memory cells continues in subsequent program loops.
[0011] The memory cells can also be programmed according to a sub-block programming sequence, where memory cells connected to a word line are programmed in one sub-block, then in a next sub-block, and so on.
[0012] Each memory cell can be assigned a data state according to write data in a programming command. Based on its data state, a memory cell either remains in the erased state or is programmed to a programmed state. For example, in a one-bit-per-cell memory cell device, also referred to as an SLC or single-level memory cell, there are two data states, including the erased state and the programmed state. See Fig. 9A. MLC or multi-level memory cells store two or more bits per cell. For example, in a memory device with two bits per cell, there are four data states, including the erased state and three higher data states called the A, B, and C data states. In a memory device with three bits per cell, there are eight data states, including the erased state and seven higher data states called the A, B, C, D, E, F, and G data states (see Fig. 9B). In a memory device with four bits per cell, there are sixteen data states, including the erased state, and fifteen higher data states. The data states can be referred to as the S0-S15 data states, where S0 is the erased state.
[0013] SLC programming is fast and has high endurance, although the storage density is lower than when multiple bits are stored in each memory cell. In terms of programming speed, SLC programming can often be performed with just a single programming pulse. The read speed is also high because only one read voltage is required, and error correction can be simplified by a large margin between the two data states. In terms of endurance, data programmed with SLC programming can be accurately read back after many read cycles because there is a larger margin between the erased and programmed data states compared to MLC programming. SLC programming can therefore be used for applications where programming and read speed, as well as endurance, are of utmost importance.For example, a group of SLC memory cells can be used as a cache between DRAM and a group of MLC memory cells. An external controller can quickly program data into a block of SLC memory cells, and this data can then be transferred to a block of MLC memory cells while the external controller performs another task. High endurance is important when data is read many times in repeated read cycles, such as data stored in a server.
[0014] After the memory cells are programmed, the data can be read back in a read operation. A read operation may involve applying a read voltage, or a series of read voltages, to a word line while a sensing circuit determines whether cells connected to the word line are in a conductive (on) or non-conductive (off) state. When a cell is in a non-conductive state, the Vth of the memory cell exceeds the read voltage. The read voltages are set to levels expected to be between the threshold voltages of neighboring data states. During the read operation, the voltages of the unselected word lines are driven to a read pass level sufficiently high to place the unselected memory cells in a highly conductive state to avoid interfering with the sensing of the selected memory cells.
[0015] However, with SLC or MLC, programming performance changes over time as the memory device is used. For example, the gate oxide of the memory cells can degrade, making them easier to program. This can lead to overprogramming if the programming voltages are not adjusted. If the programming voltages are too high, this will result in read and verify errors in a NAND chain because the unselected memory cells cannot be fully turned on to perform the scan for a selected memory cell.
[0016] Techniques provided herein address the above and other problems. In one aspect, the threshold voltage (Vth) of a set of memory cells is evaluated by repeatedly sampling the memory cells during a program loop between the program voltage and a verify voltage. The repeated sampling may occur, for example, when the voltage of the selected wordline is decreased from the program voltage to the verify voltage. This minimizes the additional time that may be used to perform the repeated sampling. Further, the repeated sampling may, for example, identify an upper or lower end of a Vth distribution.The position of the end may indicate a programming performance of the memory cells, for example, with regard to a degree of overprogramming, and this information may be used to adjust the programming setting(s), such as an initial programming voltage, in a subsequent programming operation. The programming setting(s) may therefore be adjusted to optimize the performance of the memory device. The programming setting(s) may be stored in a memory location for use in a later programming operation. In some cases, the programming setting(s) may be used to adjust the remainder of an ongoing programming operation.
[0017] In one approach, a predetermined number of sampling times is set during the decreasing voltage during which the sampling occurs. In another approach, the sampling stops based on the results of the previous sampling during the decreasing voltage. For example, the sampling can stop when the tail is detected. This saves time by avoiding additional sampling operations and instead driving the voltage directly to the verify voltage or another end voltage. Additionally, it is possible to replace the repeated sampling with a single sampling during the decreasing of the selected wordline voltage.
[0018] The selected wordline voltage may be lowered at a first rate to an intermediate level between the programming and verify voltages before being ramped down at a controlled second rate smaller than the first rate. Repeated sampling may occur during ramp-down at the second rate, but not during ramp-down at the first rate. In some cases, the intermediate level may be adaptively adjusted depending on the programming setting(s).
[0019] In another aspect, no verify voltage is applied and no verification test is performed in a program loop, so that the selected word line voltage is reduced to a final level such as 0 V after scanning, rather than to a verify voltage.
[0020] In another aspect, the programming setting(s) for a set of memory cells may be stored in a memory location comprising transistors, including select-gate transistors and / or dummy memory cells in spare NAND chains of a block. These are transistors that are not subject to erasure in a block erase operation. Furthermore, unlike other memory components, such as registers, the programming setting(s) may be retained even when a memory device is powered off. The programming setting(s) may be read using a card that connects, for example, a memory location to a wordline. The data may be stored for each wordline, a group of wordlines, each subblock, each block, or for other groups of memory cells.The programming setting(s) is / are read before a programming operation and used to set an optimal programming parameter, such as an initial programming voltage for the programming operation.
[0021] Furthermore, the repeated sampling to obtain updated programming setting(s) may occur periodically, such as once every N programming operations, where N is a number such as 256, 512, or 1028. In one approach, a random number generator is used at the beginning of a programming operation to determine whether the sampling is performed. In some cases, the sampling is performed in one or more programming loops of a programming operation.
[0022] In another aspect, the voltage on the unselected word lines remains at a power-up voltage throughout the program loop, including the program voltage, the decrease of the selected word line voltage, the repeated sampling, and the verify voltage. This avoids a delay by decreasing the voltage on the unselected word lines from the program voltage to 0 V and then increasing the voltage on the unselected word lines from 0 V to the verify voltage.
[0023] In another aspect, the evaluation of the memory cells and the optimization of the programming setting(s) can be performed using an on-chip control circuit without the assistance of an external controller. In another aspect, the evaluation and optimization can be performed for memory cells of a block, a sub-block, a word line, or a section of a word line.
[0024] These and other features are further explained below.
[0025] Fig. 1 is a block diagram of an exemplary memory device. The memory device 100, such as a non-volatile memory system, may include one or more memory elements 108. The memory chip 108 includes a memory structure 126 of memory cells, such as an array of memory cells, a control circuit 110, and read / write circuits 128. The memory structure 126 is addressable by word lines via a row decoder 124 and by bit lines via a column decoder 132. The read / write circuits 128 include a plurality of sense blocks 51, 52, ... 53 (sense circuitry) and enable parallel reading or programming of a page of memory cells. Typically, a controller 122 is included in the same memory device 100 (e.g., a removable memory card) as the one or more memory dies 108. The controller may be separate from the memory element.Commands and data are transferred between the host 140 and the controller 122 via a data bus 120 and between the controller and the one or more memory elements 108 via lines 118.
[0026] The memory structure may be 2D or 3D. The memory structure may include one or more arrays of memory cells, including a 3D array. The memory structure may include a monolithic three-dimensional memory structure in which multiple memory levels are formed above (rather than within) a single substrate, such as a wafer, with no intervening substrates. The memory structure may include any type of non-volatile memory monolithically formed in one or more physical levels of memory cell arrays having an active area disposed above a silicon substrate. The memory structure may be located in a non-volatile memory device that includes circuitry associated with the operation of the memory cells, whether the associated circuitry is located above or within the substrate.
[0027] The control circuit 110 cooperates with the read / write circuits 128 to perform store operations on the memory structure 126 and includes a state machine 112, a column redundancy circuit 111, an on-chip address decoder 114, a power control module 116 (power control circuit), a random number generator 117, a map 119 of memory locations for the programming setting(s), and a circuit 121 for determining the programming setting(s). The state machine 112 provides control of store operations at the memory level. A memory region 113 may be provided, e.g., for operating parameters, and software / code. In one embodiment, the state machine is programmable by software. In other embodiments, the state machine does not use software and is implemented entirely in hardware (e.g., electrical circuits).
[0028] The column redundancy circuit provides an association between spare NAND chains that replace defective primary NAND chains. See also Fig. 8C and Fig. 8D. The on-chip address decoder 114 provides an address interface between that used by the host or a memory controller and the hardware address used by decoders 124 and 132. The power control module 116 controls the power and voltages supplied to the word lines, select gate lines, bit lines, and source lines during memory operations. It may include drivers for word lines, SGS and SGO transistors, and source lines. See also Fig. 3. The sensing blocks may include bitline drivers in one approach. The random number generator 117 may be used to generate a random number to decide whether the performance of a set of memory cells should be evaluated during a programming operation or during a programming loop. See also Fig. 11A, step 1100a. The random number generator 117 may alternatively be provided in the controller 122. The map 119 may be used to store data identifying locations in the memory structure 126 where the programming setting(s) for optimizing the performance of a set of memory cells is / are stored. For example, the map may associate a set of memory cells connected to a word line with an initial programming voltage for use in a programming operation. The storage location may be one or more specific spare NAND chains in a spare area of a block.
[0029] The circuit 121 for determining the programming setting(s) may be configured with hardware, firmware and / or software for determining the updated programming setting(s) based on measurement data, such as by implementing the correlations in Fig. 11D and Fig. 11E.
[0030] Examples of programming setting(s) include the initial program voltage, the program voltage step size, the program voltage duration, the voltage settling time, the maximum allowable program voltage, and the bit line and source line voltages to be used during the program voltages. Furthermore, in addition to the programming setting(s), other types of data may be stored to optimize the performance of a set of memory cells, such as optimal read voltages and optimal turn-on voltages for unselected word lines. Card 119 may therefore provide a cross-reference between different sets of memory cells and different respective memory locations storing the respective programming setting(s) or other corresponding control parameters.
[0031] In some implementations, some of the components may be combined. In various embodiments, one or more of the components (alone or in combination) other than memory structure 126 may be considered to be at least one control circuit configured to perform the techniques described herein, including the steps of the methods described herein. For example, a control circuit may include any one of the components or a combination of a control circuit 110, a state machine 112, decoders 114 and 132, a power control module 116, sampling blocks 51, 52, ..., 53, read / write circuits 128, a controller 122, and so on.
[0032] The off-chip controller 122 (which in one embodiment is an electrical circuit) may include a processor 122c, storage devices (memory) such as ROM 122a and RAM 122b, and an error correction code (ECC) engine 245. The ECC engine may correct a number of read errors.
[0033] The controller 122 or control circuit 110 may be configured with hardware, firmware and / or software to implement the graphical representations of Fig. 11D-11F and other processes, including the processes in the flowcharts described herein.
[0034] A memory interface 122d may also be provided. The memory interface, which communicates with the ROM, RAM, and processor, is an electrical circuit that provides an electrical interface between the controller and the memory element. For example, the memory interface may modify the format or timing signals, provide a buffer, provide spike isolation, latch an I / O, and so on. The processor may issue commands to the control circuit 110 (or any other component of the memory element) via the memory interface 122d.
[0035] The memory device includes code, such as a set of instructions, and the processor is operable to execute the set of instructions to provide the functionality described herein. Alternatively or additionally, the processor may access code from a storage device 126a of the memory structure, such as a reserved area of memory cells in one or more wordlines.
[0036] For example, the code may be used by the controller to access the memory structure, such as for program, read, and erase operations. The code may include boot code and control code (e.g., a set of instructions). Boot code is software that initializes the controller during a boot or startup process and allows the controller to access the memory structure. The code may be used by the controller to control one or more memory structures. During bootup, processor 122c retrieves the boot code from ROM 122a or storage device 126a for execution, and the boot code initializes the system components and loads the control code into RAM 122b. Once the control code is loaded into RAM, it is executed by the processor.The control code contains drivers to perform basic tasks such as controlling and allocating memory, prioritizing the processing of instructions, and controlling input and output ports.
[0037] In general, the control code may include instructions for performing the functions described herein, including the flowchart steps discussed below, and for providing the voltage waveforms, including those discussed below. A control circuit may be configured to execute the instructions to perform the functions described herein.
[0038] In one embodiment, the host is a computing device (e.g., a laptop, a desktop, a smartphone, a tablet, a digital camera) that includes one or more processors, one or more processor-readable storage devices (RAM, ROM, flash memory, hard disk drive, solid-state memory) storing processor-readable code (e.g., software) for programming the one or more processors to perform the methods described herein. The host may also include additional system memory, one or more input / output interfaces, and / or one or more input / output devices in communication with the one or more processors.
[0039] Other types of non-volatile memory in addition to NAND flash memory can also be used.
[0040] Semiconductor memory devices include volatile memory devices, such as dynamic random access memory ("DRAM") or static random access memory ("SRAM"), non-volatile memory devices, such as resistive random access memory ("ReRAM"), electrically erasable programmable read-only memory ("EEPROM"), flash memory (which can also be considered a subset of EEPROM), ferroelectric random access memory ("FRAM"), and magnetoresistive random access memory ("MRAM"), and other semiconductor elements capable of storing information. Each type of memory device may have different devices. For example, flash memory devices may be implemented in a NAND or NOR device.
[0041] The memory devices can be formed from passive and / or active elements in any combination. As a non-limiting example, passive semiconductor memory elements include ReRAM device elements, which in some embodiments include a resistive switching memory element such as an anti-fuse or phase-change material, and optionally a steering element such as a diode or transistor. Furthermore, as a non-limiting example, active semiconductor memory elements include EEPROM and Flash memory device elements, which in some embodiments include elements containing a charge storage region, such as a floating gate, conductive nanoparticles, or a dielectric charge storage material.
[0042] Multiple storage elements can be configured to be connected in series, or such that each element is individually accessible. As a non-limiting example, flash memory devices typically include storage elements connected in series in a NAND device (NAND memory). A NAND chain is an example of a set of series-connected transistors comprising memory cells and SG transistors.
[0043] A NAND memory array can be configured such that the array is composed of multiple memory chains, each chain being composed of multiple memory elements that share a single bit line and are accessed as a group. Alternatively, memory elements can be configured such that each element is individually accessible, e.g., a NOR memory array. NAND and NOR memory configurations are examples, and memory elements can be configured differently.
[0044] The semiconductor memory elements located within and / or above a substrate may be arranged in two or three dimensions, such as a two-dimensional memory structure or a three-dimensional memory structure. In a two-dimensional memory structure, the semiconductor memory elements are arranged in a single plane or a single memory device level. Typically, in a 2D memory structure, memory elements are arranged in a plane (e.g., in an xy-direction plane) that extends substantially parallel to a major surface of a substrate supporting the memory elements. The substrate may be a wafer over or in which the layer of memory elements is formed, or it may be a carrier substrate that is attached to the memory elements after they are formed. As a non-limiting example, the substrate may include a semiconductor, such as silicon.
[0045] The memory elements may be arranged in an ordered array at the individual memory device level, such as in a plurality of rows and / or columns. However, the memory elements may be arranged in non-regular or non-orthogonal configurations. The memory elements may each have two or more electrodes or contact lines, such as bit lines and word lines.
[0046] A 3D memory array is arranged so that memory elements occupy multiple planes or multiple memory device levels, thereby forming a structure in three dimensions (i.e., in the x, y, and z directions, with the z direction being substantially perpendicular and the x and y directions being substantially parallel to the main surface of the substrate).
[0047] As a non-limiting example, a 3D memory structure may be arranged vertically as a stack of multiple 2D memory device levels. As another non-limiting example, a 3D memory array may be arranged as multiple vertical columns (e.g., columns extending substantially perpendicular to the main surface of the substrate, i.e., in the y-direction), with each column having multiple memory elements. The columns may be arranged in a 2D device, e.g., in an xy-plane, resulting in a 3D array of memory elements with elements on multiple vertically stacked memory levels. Other devices of memory elements in three dimensions may also form a 3D memory array.
[0048] As a non-limiting example, the storage elements in a 3D NAND memory array may be coupled together to form a NAND chain within a single horizontal (e.g., xy) memory device level. Alternatively, the storage elements may be coupled together to form a vertical NAND chain spanning multiple horizontal memory device levels. Other 3D devices may be contemplated, with some NAND chains containing storage elements in a single memory level, while other chains containing storage elements spanning multiple memory levels. 3D memory arrays may also be configured as a NOR device and as a ReRAM device.
[0049] Typically, in a monolithic 3D memory array, one or more memory device levels are formed above a single substrate. Optionally, the monolithic 3D memory array may also have one or more memory layers at least partially within the single substrate. As a non-limiting example, the substrate may include a semiconductor, such as silicon. In a monolithic 3D array, the layers forming each memory device level of the array are typically formed on top of the layers of the array's underlying memory device levels. However, layers of adjacent memory device levels of a monolithic 3D memory array may be shared or have intermediate layers between memory device levels.
[0050] 2D arrays can be formed separately and then packaged together to form a non-monolithic memory device with multiple memory layers. For example, non-monolithic stacked memories can be constructed by forming memory levels on separate substrates and then stacking the memory levels on top of each other. The substrates can be thinned or removed from the memory device levels before stacking, but because the memory device levels are initially formed over separate substrates, the resulting memory arrays are not monolithic 3D memory arrays. Furthermore, multiple 2D memory arrays or 3D memory arrays (monolithic or non-monolithic) can be formed on separate chips and then packaged together to form a stacked-chip memory device.
[0051] Associated circuitry is typically required for the operation of the memory elements and for communication with the memory elements. As non-limiting examples, memory devices may include circuitry used to control and drive memory elements to perform functions such as programming and reading. This associated circuitry may be located on the same substrate as the memory elements and / or on a separate substrate. For example, a controller for memory write / read operations may be located on a separate control chip and / or on the same substrate as the memory elements.
[0052] One of ordinary skill in the art will recognize that this technology is not limited to the described exemplary 2D and 3D structures, but covers all relevant memory structures within the spirit and scope of the technology as described herein and as understood by one of ordinary skill in the art.
[0053] Fig. Figure 2 is a block diagram illustrating one embodiment of a sampling block 51 of Fig. 1. A single sense block 51 may be divided into one or more core sections, referred to as sense circuits 180-183 or sense amplifiers, and a common section referred to as management circuit 190. In one embodiment, there is a separate sense circuit for each bitline / NAND chain and a common management circuit 190 for a set of multiple, e.g., four or eight, sense circuits. Each of the sense circuits in a group communicates with the associated management circuit via data bus 172. Thus, there is one or more management circuits that communicate with the sense circuits of a set of memory elements.
[0054] Sense circuitry 180 includes, as one example, sensing circuitry 170 that performs sensing by determining whether a line current in a connected bitline is above or below a predetermined threshold level. For example, during the ramp-down sensing described herein, the sensing circuitry detects a voltage on the bitlines of the sensed NAND strings at each sensing time. In one approach, if a memory cell is conductive, the bitline voltage discharges below a trigger voltage at a specified time after the sensing time. If the bitline voltage does not discharge below the trigger voltage, the memory cell is considered non-conductive. See also Fig. 11B regarding sampling of the lower or upper end of the Vth distribution.
[0055] Sense circuit 180 also includes a bitline latch 184, which is used to set a voltage state on the connected bitline. For example, a predetermined state latched in a bitline latch results in the connected bitline being pulled to a state indicating a program disable (e.g., 1.5 to 3 V). As an example, a flag = 0 may prevent programming, while a flag = 1 does not prevent programming.
[0056] Management circuitry 190 includes a processor 192, four exemplary sets of data latches 194-197, and an I / O interface 196 coupled between the set of data latches 194 and data bus 120. A set of data latches, such as LDL and UDL, may be provided for each sensing circuit. In some cases, additional data latches may be used. LDL stores one bit for a lower page of data, and UDL stores one bit for an upper page of data. This applies to a memory device with four levels or two bits per memory element. One additional data latch per bit line may be provided for each additional data bit per memory element.
[0057] Processor 192 performs calculations to, for example, determine the data stored in the scanned memory element and store the determined data in the set of data latches. Each set of data latches 194-197 is used to store data bits determined by processor 192 during a read operation and to store data bits imported from data bus 120 during a program operation, representing write data to be programmed into memory. I / O interface 196 provides an interface between data latches 194-197 and data bus 120.
[0058] During a read, the operation of the system is under the control of state machine 112, which controls the supply of different control gate voltages to the addressed memory element. Because it switches the various predefined control gate voltages corresponding to the different memory states supported by the memory, sampling circuit 180 may trigger at one of these voltages, and a corresponding output is provided from sampling circuit 180 via bus 172 to processor 192. At this point, processor 192 determines the resulting memory state by considering the sampling circuit trigger event(s) and information about the applied control gate voltage from the state machine via input lines 193. It then computes a binary encoding for the memory state and stores the resulting data bits in data latches 194-197.In another embodiment of the management circuit 190, the bit line latch serves both as a latch for latching the output of the sensing circuit 180 and as a bit line latch, as described above.
[0059] Some implementations may include multiple processors 192. In one embodiment, each processor 192 includes an output line (not shown) such that each of the output lines is wired together via an OR line. In some embodiments, the output lines are inverted before being connected to the wired OR line. This feature allows for a quick determination during the programming verification process when the programming process is complete because the state machine receiving the wired OR can determine when all programmed bits have reached the desired level. For example, when each bit has reached its desired level, a logic zero for that bit is sent to the wired OR line (or a data one is inverted).When all bits output a data 0 (or an inverted data 1), the state machine knows to end the programming procedure. Since each processor communicates with eight sampling circuits, the state machine must read the wired-OR line eight times, or logic is added to processor 192 to accumulate the results of the associated bit lines so that the state machine only needs to read the wired-OR line once. Similarly, by correctly selecting logic levels, the global state machine can detect when the first bit changes state and modify the algorithms accordingly.
[0060] During memory cell programming or verification operations, the data to be programmed (write data) is stored in the set of data latches 194-197 from the data bus 120 in the LDL and UDL latches in a two-bit per memory element implementation. In a three-bit per memory element implementation, an additional data latch may be used. The programming operation, under the control of the state machine, comprises a series of programming voltage pulses applied to the control gates of the addressed memory elements. Each programming voltage pulse is followed by a readback (verify) to determine whether the memory element has been programmed to the desired memory state. In some cases, processor 192 monitors the state of the readback memory relative to the desired memory state.If the two match, processor 192 sets the bitline latch to cause the bitline to be pulled into a state indicating a program lock. This prevents the memory element coupled to the bitline from being further programmed, even if programming pulses appear on its control gate. In other embodiments, the processor initially loads the bitline latch, and the sensing circuit sets it to a lock value during the verification process.
[0061] Each set of data latches 194-197 may be implemented as a stack of data latches for each sampling circuit. In one embodiment, there are three data latches per sampling circuit 180. In some implementations, the data latches are implemented as shift registers so that the parallel data stored therein is converted into serial data for the data bus 120 and vice versa. All of the data latches corresponding to the read / write block of memory elements may be connected together to form a block shift register so that a block of data can be input or output via serial transfer. In particular, the bank of read / write circuits is adapted so that each of its set of data latches shifts data sequentially into or out of the data bus, as if they were part of a shift register for the entire read / write block.
[0062] Data latches identify when a corresponding memory element has reached certain portions of a program operation. For example, latches can identify that a memory element's Vth is below a certain verify level. Data latches indicate whether a memory element is currently storing one or more bits from a page of data. For example, LDL latches can be used to store a lower page of data. An LDL latch is flipped (e.g., from 0 to 1) when a lower page bit is stored in a corresponding memory element. A UDL latch is flipped when an upper page bit is stored in a corresponding memory element. This occurs when a corresponding memory element completes programming, for example, when its Vth exceeds a target verify level such as VvA, VvB, or VvC.
[0063] Fig. 3 illustrates an exemplary implementation of the power control module 116 of Fig. 1 for providing voltages to blocks of memory cells. In this example, the memory structure 126 includes a set 410 of four associated blocks, BLK_0 through BLK_3, and another set 411 of four associated blocks, BLK_4 through BLK_7. The blocks may be located in one or more levels. The row decoder 124 of Fig. 1 provides voltages to the word lines and selects the gates of each block via pass transistors 422. The row decoder provides a control signal to pass transistors connecting the blocks to the row decoder. In one approach, the pass transistors of each set of blocks are controlled by a common control gate voltage. Thus, the pass transistors for a set of blocks are either all on or off at a given time. When the pass transistors are on, a voltage is provided from the row decoder to the corresponding control gate lines or word lines. When the pass transistors are off, the row decoder is disconnected from the corresponding control gate lines or word lines, so that the voltage on the corresponding control gate lines or word lines floats.
[0064] For example, a control gate line 412 is connected to sets of pass transistors 413, 414, 415, and 416, which in turn are connected to control gate lines of BLK_4, BLK_5, BLK_6, and BLK_7, respectively. A control gate line 417 is connected to sets of pass transistors 418, 419, 420, and 421, which in turn are connected to control gate lines of BLK_0, BLK_1, BLK_2, and BLK_3, respectively.
[0065] Typically, program or read operations are performed on a selected block simultaneously and on a selected sub-block of the block. An erase operation can be performed on a selected block or sub-block. The row decoder can connect global control lines 402 to local control lines 403. The control lines represent conductive traces. Voltages are provided on the global control lines by a series of voltage drivers. Some of the voltage drivers can provide voltages to switches 450 connected to the global control lines. Pass transistors 424 are controlled to pass voltages from the voltage drivers to the switches 450.
[0066] The voltage drivers may include a selected data word line driver 447 (WL), which provides a voltage on a data word line selected during a program or read operation, a driver 448 for unselected data word lines, and dummy word line drivers 449-449c, which provide voltages on dummy word lines. For example, the dummy word line drivers 449, 449a, 449b, and 449c may provide voltages on the control gate layers or word line layers WLDD0, WLDD1, WLDS1, and WLDS0, respectively. Fig. 6A during an update process as described herein.
[0067] The voltage drivers may also include separate SGS and SGD drivers for each sub-block. For example, SGS drivers 445, 445a, 445b, and 445c, and SGD drivers 446, 446a, 446b, and 446c may be provided for SB0, SB1, SB2, and SB3, respectively, as shown in Fig. 7, Fig. 8A and Fig. 8B. In another option, an SGS driver is common to the different sub-blocks in a block. Furthermore, in some cases, as in Fig. 7, Fig. 8B and Fig. 8D, there can be multiple SGD transistors, multiple SGS transistors, multiple drain-side dummy memory cells, and / or multiple source-side dummy memory cells in each NAND chain. To provide maximum flexibility in programming these memory cells and transistors with programming voltages or other wordline-specific parameters, a separate driver can be present for each select-gate transistor and each dummy memory cell in a NAND chain in one approach.
[0068] The various components, including the row decoder, may receive instructions from a controller such as state machine 112 or controller 122 to perform the functions described herein.
[0069] In one approach, the well region 433 is common to the blocks and can be driven by a voltage driver 430 via a path 432. A set of bit lines 442 is also shared by the blocks. A bit line voltage driver 440 provides voltages to the bit lines. In a stacked memory device, as in Fig. As shown in Figures 5 to 8B, sets of connected memory cells may be arranged in NAND strings extending vertically upward from the substrate. The bottom (or source end) of each NAND string is in contact with the well region, and the top (or drain end) of each NAND string is connected in a boss to a corresponding bit line.
[0070] Fig. Figure 4A shows an example of map 119 of Fig. 1. As mentioned, the card can store data linking different sets of memory cells to different corresponding memory locations that store the corresponding programming setting(s). In one approach, each word line, such as WL0, WL1, WL2, WL3, ..., is mapped to a spare NAND chain (NS), such as NS 704n, 705n, 706n, 707n, ... in Fig. 8D. The storage location for a word line could be in one or more spare NAND chains and, for example, in one or more transistors of the spare NAND chains. The transistors used to store the programming setting(s) may include dummy memory cells and selected gate transistors that are not normally erased during a block erase operation. This allows the programming setting(s) to be stored across multiple program and erase cycles of a set. In contrast, data storage cells are typically erased in a block erase operation. Furthermore, by using spare NAND chains to store the programming setting(s), there is no capacitance loss in the primary NAND chains. The number of spare NAND chains used may be a small fraction of all spare NAND chains, so the overall cost is acceptable.Once the dummy memory cells and / or select gate transistors of a spare NAND chain are used to store programming settings, the remaining data memory cells of the spare NAND chain may not be available for storing data in one approach. The remaining spare NAND chains can be used to replace primary NAND chains that prove to be defective. The column redundancy circuit 111 (. Fig. 1) can be used to map the defective primary NAND chains to available replacement NAND chains.
[0071] By programming the programming setting(s) into multiple dummy memory cells and / or select gate transistors of a spare NAND chain, the number of spare NAND chains used can be minimized. To further reduce overall cost, multiple bits of the programming setting(s) can be stored in each dummy memory cell and / or select gate transistor. Or, for greater reliability and write / read speed, a single bit of the programming setting(s) can be stored in each dummy memory cell and / or select gate transistor. See also Fig. 12.
[0072] The map 119 may also compare a set of memory cells to a portion of a NAND chain or to multiple NAND chains. For example, a map entry could contain a cross-reference WL0 to the drain-side dummy memory cells and / or selected gate transistors of NS 704n (see, e.g., dummy memory cells 885 and 886 and select gate transistors 887 and 888 of Fig. 8D), and another entry of the map could contain a cross-reference WL1 to the source side dummy memory cells and / or the select gate transistors of NS 704n (see, e.g., the dummy memory cells 882 and 883 and the select gate transistors 880 and 881 of Fig. 8D). In another example, one entry of the map could contain a cross-reference WL0 to the drain-side dummy memory cells of NS 704n and 705n (see, e.g., dummy memory cells 885 and 886 in NS 704n and dummy memory cells 891 and 892 in NS 705n). Another entry of the map could contain a cross-reference WL1 to the drain-side select gate transistors of NS 704n and 705n (see, e.g., SGD transistors 887 and 887 in NS 704n and SGD transistors 893 and 894 in NS 705n).
[0073] Fig. Figure 4B illustrates an example of programming setting(s) stored in a memory location referenced in card 119 of Fig. 4A. In this example, the memory locations are NAND chains (NS), such as NS 704n, 705n, 706n, 707n,..., which store an initial programming voltage, such as 15 V, 15.5 V, 16 V, and 15 V, respectively. The initial programming voltages are an example of one or more programming settings and can be represented by different sets of bits. Each set of bits can be correlated to an initial programming voltage by the on-chip circuitry. A set of bits can include, for example, 4 bits or 8 bits (one byte).
[0074] The number of different initial programming voltages depends on the number of bits associated with the programming setting(s). In one approach, the programming setting(s) identify more than one data type. For example, a byte of the programming setting(s) may be stored, with the first and second halves of the byte representing the first and second types of programming setting(s), respectively. In one example, the first and second types of programming setting(s) are an initial programming voltage and a step size, or an initial programming voltage and a programming voltage duration, respectively.
[0075] Fig. 5 is a perspective view of a memory device 500 including a set of blocks in an exemplary 3D device of the memory structure 126 of Fig. 1. On the substrate 501 are example blocks BLKO, BLK1, BLK2, and BLK3 of memory cells (memory elements) and peripheral regions with circuit logic for use by the blocks. The peripheral region 504 runs along one edge of each block, while the peripheral region 505 is located at one end of the set of blocks. The circuitry may include voltage drivers that can be connected to control gate layers, bit lines, and source lines of the blocks. In one approach, control gate layers at a common height in the blocks are commonly driven. The substrate 501 may also support circuitry beneath the blocks and one or more lower metal layers patterned into conductive traces to carry signals of the circuitry. The blocks are formed in an intermediate region 502 of the memory device.In an upper region 503 of the memory device, one or more upper metal layers are patterned into conductive traces to carry logic signals. Each block comprises a stacked region of memory cells, with alternating levels of the stack representing word lines. In one possible approach, each block has opposing stepped sides from which vertical contacts extend upward to an upper metal layer to establish connections to conductive traces. While four blocks are shown by way of example, two or more blocks extending in the x- and / or y-direction may be used.
[0076] In one possible approach, the blocks are arranged in a plane, and the length of the plane in the x-direction represents a direction in which signal paths extend to word lines in the one or more upper metal layers (a word line or SGD line direction), and the width of the plane in the y-direction represents a direction in which signal paths extend to bit lines in the one or more upper metal layers (a bit line direction). The z-direction represents a height of the memory device. The blocks could also be arranged in multiple planes.
[0077] Fig. Figure 6A illustrates an exemplary cross-sectional view of a portion of block BLKO of Fig. 5. The block comprises a stack 610 of alternating conductive and dielectric layers. In this example, the conductive layers comprise two SGD layers, one SGS layer, two source-side dummy wordline layers (or wordlines) WLDS1 and WLDS0, two drain-side dummy wordline layers WLDD1 and WLDD0, and eleven data wordline layers (or data wordlines) WLL0-WLL10. WLL0 is a source-side data wordline, and WLDS1 is a dummy wordline layer adjacent to the source-side data wordline. WLDS0 is another dummy wordline layer adjacent to WLDS1. WLL10 is a drain-side data wordline, and WLDD1 is a dummy wordline layer adjacent to the drain-side data wordline. WLDD0 is another dummy wordline layer adjacent to WLDD1. The dielectric layers are designated DL1-1L19. Also shown are regions of the stack containing NAND chains 700n and 710n.Each NAND chain encloses a storage hole 618 or 619 filled with materials that form memory cells adjacent to the word lines. Region 622 of the stack is shown in . Fig. 6C shown in more detail.
[0078] The stack includes a substrate 611. In one approach, a portion of the source line SL has a well region 611a as an n-type source diffusion layer or recess in the substrate. The well region is in contact with a source end of each chain of memory cells in a block. An erase pulse can be applied to this layer in an erase operation. In one possible implementation, the n-type well region 611a is formed in a p-type well region 611b, which in turn is formed in an n-type well region 611c, which in turn is formed in a p-type semiconductor substrate 611d. The n-type source diffusion layer can be shared by all blocks in a plane in one approach.
[0079] The NAND chain 700n has a source end 613 at a bottom 616b of the stack 616 and a drain end 615 at a top 616a of the stack. Metal-filled slots 617 and 620 may be provided periodically above the stack as connections extending through the stack, for example, to connect the source line to a line above the stack. The slots may be used during word line fabrication and subsequently filled with metal. A portion of a bit line BL0 is also shown. A conductive path 621 connects the drain end 615 to BL0.
[0080] In one approach, the block of memory cells comprises a stack of alternating control gate and dielectric layers, and the memory cells are arranged in vertically extending memory holes in the stack.
[0081] In one approach, each block has a terraced edge in which vertical connections connect to each layer, including the SGS, WL, and SGD layers, and extend upward to horizontal paths to voltage drivers.
[0082] Fig. 6B illustrates an exemplary transistor 650. The transistor includes a control gate CG, a drain D, a source S, and a channel CH and may represent, for example, a memory cell or a select gate transistor.
[0083] Fig. Figure 6C shows a detailed view of region 622 of the stack of Fig. 6A. Memory cells are formed at the various levels of the stack at the intersection of a wordline layer and a memory hole. In this example, SGD transistors 718 and 717 are provided above dummy memory cells 716 and 715 and a data memory cell 714. These SGD transistors are located at the drain end of the NAND string.
[0084] A series of layers may be deposited along the sidewall (SW) of the memory hole 630 and / or within each wordline layer, e.g., using atomic layer deposition. For example, each pillar 685 or column formed by the materials within a memory hole may include a blocking oxide layer 663, a charge-trapping layer 664 or a foil such as silicon nitride (Si3N4) or another nitride, a tunneling layer 665 (e.g., comprising a gate oxide that may degrade over time), a channel 660 (e.g., comprising polysilicon), and a dielectric core 666 (e.g., comprising silicon dioxide). A wordline layer may include a metal barrier 661 and a conductive metal 662, such as tungsten, as a control gate. For example, control gates 690-694 are provided. In this example, all layers except the metal are provided in the storage hole.In other approaches, some of the layers may be in the control gate layer. Additional columns are similarly formed in the various storage holes. A column may form a column-shaped active area (AA) of a NAND chain.
[0085] Each NAND chain or set of connected transistors has a channel that extends continuously from one or more source-end select gate transistors to one or more drain-end select gate transistors. For example, channels 700a, 710a, 720a, and 730a extend continuously in NAND chains 700n, 710n, 720n, and 730n, respectively. Channel 700a extends continuously in NAND chains 700n from SGS transistor 701 to SGD transistors 717 and 718. Channel 700a is continuous in that it is uninterrupted and can therefore provide a continuous conduction path in the NAND chain.
[0086] When a memory cell is programmed, electrons are stored in a portion of the charge-trapping layer associated with the memory cell. These electrons are drawn into the charge-trapping layer from the channel and through the tunnel layer. The Vth of a memory cell increases proportionally to the amount of stored charge. During an erase operation, the electrons return to the channel.
[0087] Each of the memory holes may be filled with a plurality of annular layers including a blocking oxide layer, a charge trapping layer, a tunneling layer, and a channel layer. A core region of each of the memory holes is filled with a body material, and the plurality of annular layers are located between the core region and the word line in each of the memory holes.
[0088] The NAND chain can be considered to have a floating-body channel, since the entire length of the channel is not formed on a substrate. Furthermore, the NAND chain is formed by a plurality of wordline layers arranged one above the other in a stack and separated by dielectric layers.
[0089] Fig. Figure 7 shows an exemplary view of NAND chains in the block BLKO, which are Fig. 5 and Fig. 6A. The NAND chains are arranged in sub-blocks of the block in a 3D device. Each sub-block includes a plurality of NAND chains, with an exemplary NAND chain being illustrated. For example, SB0, SB1, SB2, and SB3 include exemplary NAND chains 700n, 710n, 720n, and 730n, respectively. The NAND chains have data word lines, dummy word lines, and select gate lines connected to Fig. 6A. Each sub-block includes a set of NAND strings extending in the x-direction that share a common SGD line or control gate layer. NAND strings 700n, 710n, 720n, and 730n are located in sub-blocks SB0, SB1, SB2, and SB3, respectively. Programming of the block may be performed based on a wordline programming command. One possibility is to program the memory cells in different sections of a wordline located in the different sub-blocks, one sub-block at a time, before programming the memory cells of the next wordline. Another possibility is to program all of the memory cells in a sub-block, one wordline at a time, before programming the memory cells of the next sub-block.For example, the wordline programming sequence can start at WL0, the source-end wordline, and end at WLL10, the drain-end wordline.
[0090] The NAND chains 700n, 710n, 720n and 730n have channels 700a, 710a, 720a and 730a respectively.
[0091] In addition, NAND chain 700n includes SGS transistor 701, dummy memory cells 702 and 703, data memory cells 704, 705, 706, 707, 708, 709, 710, 711, 712, 713 and 714, dummy memory cells 715 and 716, and SGD transistors 717 and 718.
[0092] NAND chain 710n includes SGS transistor 721, dummy memory cells 722 and 723, data memory cells 724, 725, 726, 727, 728, 729, 730, 731, 732, 733 and 734, dummy memory cells 735 and 736, and SGD transistors 737 and 738.
[0093] The NAND chain 720n includes the SGS transistor 741, dummy memory cells 742 and 743, data memory cells 744, 745, 746, 747, 748, 749, 750, 751, 752, 753 and 754, dummy memory cells 755 and 756, and SGD transistors 757 and 758.
[0094] The NAND chain 730n includes the SGS transistor 761, dummy memory cells 762 and 763, data memory cells 764, 765, 766, 767, 768, 769, 770, 771, 772, 773 and 774, dummy memory cells 775 and 776, and SGD transistors 777 and 778.
[0095] One or more SGD transistors are provided at the drain end of each NAND chain, and one or more SGS transistors are provided at the source end of each NAND chain. The SGD transistors in SB0, SB1, SB2, and SB3 can each be controlled via separate control lines SGD0(0) and SGD1(0), SGD0(1) and SGD1(1), SGD0(2) and SGD1(2), and SGD0(3) and SGD1(3) in one approach. In another approach, all SGD transistors in a sub-block are connected together and controlled together. The SGS transistors in SB0, SB1, SB2, and SB3 can each be controlled via separate control lines SGS(0), SGS(1), SGS(2), and SGS(3). In another approach, all SGS transistors in a block are connected together and controlled together.
[0096] Fig. Figure 8A shows the control gate layers in the block BLKO, which are Fig. 7 are consistent. The control gate layers are arranged in a stack 800 and include dummy wordline layers or control gate layers WLDS0, WLDS1, WLDD0, and WLDD1, and data wordline layers or control gate layers WLL0-WLL10 shared by the various sub-blocks SB0-SB3. The control gate layers include a separate SGS control gate layer for each sub-block, e.g., SGS(0)-SGS(3), and separate SGD control gate layers for each sub-block. For example, SB0 includes SGD0(0) and SGD1(0), SB1 includes SGD0(1) and SGD1(1), SB2 includes SGD0(2) and SGD1(2), and SB3 includes SGD0(3) and SGD1(3). Additionally, four example memory holes are shown in each sub-block. SGD transistors 718, 718a, 718b, and 718c are shown in SGD0(0), SGD transistor 717 is shown in SGD1(0), and dummy memory cell 716 is shown in WLDD0.
[0097] Fig. Figure 8B shows an additional detail of the block BLKO of Fig. 7. Example memory cells are shown extending in the x-direction along word lines in each sub-block. Each memory cell is represented as a cube for simplicity. SB0 includes NAND chains 700n, 701n, 702n, and 703n. SB1 includes NAND chains 710n, 711n, 712n, and 713n. SB2 includes NAND chains 720n, 721n, 722n, and 723n. SB3 includes NAND chains 730n, 731n, 732n, and 733n. Bit lines are connected to sets of NAND chains. For example, a bit line BL0 is connected to NAND chains 700n, 710n, 720n, and 730n, a bit line BL1 is connected to NAND chains 701n, 711n, 721n, and 731n, a bit line BL2 is connected to NAND chains 702n, 712n, 722n, and 732n, and a bit line BL3 is connected to NAND chains 703n, 713n, 723n, and 733n. A sensing circuit may be connected to each bit line. For example, the sensing circuits 180, 181, 182, and 183 of Fig. 2 connected to the bit lines BL0, BL1, BL2 and BL3.
[0098] Programming and reading can occur for selected cells in one word line and one sub-block at a time. This allows each selected cell to be controlled by a respective bit line and / or source line. For example, a set of memory cells including an exemplary memory cell 714 is connected to WLL10 in SB0. This is the drain-end data word line. WLL0 is the source-end data word line. A set of memory cells can be programmed or read simultaneously. An additional set of memory cells is connected to WLL10 in each of the other sub-blocks SB1-SB3. For example, a set of memory cells including an exemplary memory cell 734 is connected to WLL10 in SB1.
[0099] In this example, the source line SL or the source region (well region 611a) is driven with a voltage Vsource.
[0100] Each NAND chain includes one or more drain-end SGD transistors and one or more source-end SGS transistors. In this case, there are two SGD transistors and one SGS transistor per chain. Each SGD transistor can be connected to a separate control line layer, as shown in Fig. 8A so that it can be controlled separately, or the two or more SGD transistors in a chain can be connected to their control gates and controlled together.
[0101] The NAND chain 700n includes SGD transistors 788 and 787 connected to selected gate lines SGD0(0) and SGD1(0), respectively, dummy memory cells 786 and 785 connected to WLDD0 and WLDD1, respectively, and a data memory cell 784 connected to WLL10.
[0102] Fig. 8C represents a subdivision of the block BLKO of Fig. 8B into a primary set of NAND chains 850 and a reserve set of NAND chains 851. As mentioned, spare NAND chains can be used to replace defective primary NAND chains according to a diagram of the column redundancy circuit ( Fig. 1). Defects, such as short circuits, can be detected, for example, at the time of manufacture or during use of the memory device by the end user. Some of the techniques described herein may use the spare NAND chains to store programming settings for sets of memory cells.
[0103] Fig. 8D provides additional details of SB0 from Fig. 8B, which is Fig. 8C. Similar to the block in Fig. 8C, SB0 can be divided into a primary set of NAND chains SB0p and a reserve set of NAND chains SB0s. The primary set of NAND chains in SB0 includes the Fig. 8A. The spare set of NAND chains in SB0 includes NAND chains 700n-703n. The spare NAND chains include drain-side transistors 890 including dummy memory cells and SGD transistors, source-side transistors 889 including dummy memory cells and SGS transistors, and data storage cells 895. In one approach, as mentioned, the drain-side transistors and the source-side transistors can be used to store programming setting(s) for optimizing the programming of various sets of memory cells.
[0104] The NAND chain 704n includes exemplary data memory cells 883 and 884 (connected to WLL0 and WLL10, respectively), dummy memory cells 881, 882, 885, and 886 (connected to WLDS0, WLDS1, WLDD1, and WLDD0, respectively), SGS transistor 880 (connected to SGS(0) and SGD transistors 887 and 888 (connected to SGD1(0) and SGD0(0), respectively). The NAND chain 705n includes exemplary dummy memory cells 891 and 892 (connected to WLDD1 and WLDD0, respectively) and exemplary SGD transistors 893 and 894 (connected to SGD1(0) and SGD0(0), respectively).
[0105] Fig. 9A illustrates a threshold voltage (Vth) distribution of SLC memory cells in an erased (Er) state (Vth distribution 900) and a programmed (P) state (Vth distributions 901 and 902), including an example of overprogramming (Vth distribution 902) and sense voltages used to determine an amount of overprogramming. In Fig. In Figures 9A-9D, the vertical axis represents a row of memory cells on a logarithmic scale, and the horizontal axis represents Vth on a linear scale. Initially, a block of memory cells is erased, and the memory cells are in the Er state. The erase operation may use the erase verify voltage VvEr to establish the Vth distribution 900. Subsequently, when a program operation is performed, some of the memory cells remain in the Er state, while others are programmed to the P state in one or more program loops. The program operation may use the program verify voltage VvP to establish the Vth distribution 901 for a new memory device. However, in a cyclic memory device, the memory cells are easier to program, so the Vth distribution 902 is achieved. This Vth distribution is shifted upward and expanded compared to the Vth distribution 901.
[0106] In this example, the memory cells are SLC cells that are programmed in one or two programming loops, e.g., using one or two programming voltages or pulses, such as using the voltage signal from Fig. 10A. For example, assume that Vth distributions 901 and 902 are both obtained after a programming pulse. After the first programming pulse, a verification test can be performed to ensure that Vth is above VvP. If Vth does not exceed VvP, a second programming pulse can be applied. The SLC cells can thus be programmed very quickly. As mentioned above, the upper or lower tail of a Vth distribution can be sampled in a ramp-down sampling process that takes place between a programming voltage and a verify voltage. For example, the memory cells may be sampled using voltages VvP-V5 (VvP, V1, V2, V3, V4, and V5) to detect the lower tail of the Vth distributions 901 or 902, and using voltages V6-V11 (V6, V7, V8, V9, V10, and V11) to detect the upper tail of the Vth distributions 901 or 902.The tail of a Vth distribution can be classified according to the voltage range it falls within. For example, the lower tails of Vth distributions 901 and 902 fall within the voltage ranges from VvP to V1 and V2 to V3, respectively, and the upper tails of Vth distributions 901 and 902 fall within the voltage ranges from V6 to V7 and V9 to V10, respectively.
[0107] A sequence of bits b0 to b6 that can be used to encode a sample result that identifies, for example, the lower tail of the Vth distribution 902. Each bit is associated with a voltage range and can be a 1 or 0 to indicate whether the tail lies above or below the associated range. For example, b0 to b2 can be 1 to indicate that the tail of the Vth distribution 902 lies above the associated ranges of <VvP, VvPbis V1 bzw. V1 bis V2 liegt, und b3-b6 kann 0 sein, um anzugeben, dass der Schwanz nicht über den assoziierten Bereichen von V2 bis V3, V3 bis V4, V4 bis V5 bzw. > V5. The tail is therefore located in the lowest of these associated regions (V2 to V3). Another possibility is that a bit is set to 1 when the tail falls into the associated region, and the other bits are set to 0, for example. For example, b0, b1, b2, b4, b5, and b6 = 0, and b3 = 1.The tail is therefore in the associated area of b3 (e.g. V2 TO V3).
[0108] By analyzing the Vth distribution, the programming setting(s) for the associated memory cells can be adjusted and adapted as the performance of the memory device changes. For example, if overprogramming increases upon power-on of the memory device, the initial programming voltage and / or the programming pulse duration can be reduced to reduce overprogramming in a subsequent programming operation. In some cases, the sampling may show that the Vth distribution has shifted downward over time, and the programming setting(s) can be adjusted accordingly, for example, by increasing the initial programming voltage and / or the programming pulse duration in a subsequent programming operation.
[0109] The number of sampling voltages (e.g., control gate or wordline voltages used during sampling) can be decided based on various criteria, including the desired number of possible adjustments of the programming setting(s). If a higher adjustment granularity is desired, a larger number of sampling voltages can be used. In one approach, the sampling voltages are equally spaced, although alternatively, unequal increments could be used. The time increment between consecutive sampling operations can also be equal or unequal. If both the upper and lower tails are sampled (see Fig. 13F), the number of sampling voltages and the distance between them can be the same or different for the upper and lower tail.
[0110] In Fig. 9A, the voltages VvP to V5 and V6 to V11 are examples of evaluation voltages.
[0111] Fig. Figure 9B illustrates an example of the Vth distribution of a set of memory cells in eight data states. In one approach, at the beginning of a program operation, all memory cells are initially in the erased state, as represented by Vth distribution 910. After the program operation is successfully completed, the memory cells assigned to the Er state are represented by Vth distribution 910a. The Vth distribution is shifted upward due to program disturbances.
[0112] The memory cells programmed to the A, B, C, D, E, F, and G states using verify voltages of VvA, VvB, VvC, VvD, VvE, VvF, and VvG, respectively, are represented by Vth distributions 911, 912, 913, 914, 915, 916, and 917, respectively. The Vth distributions are obtained immediately after the completion of the programming operation. The verify voltages are used in programming verification tests for the memory cells. Read voltages VrA, VrB, VrC, VrD, VrE, VrF, and VrG can be used to read the states of the memory cells in a read operation. These verify voltages and read voltages are examples of control gate read levels of the selected word line voltage.
[0113] During an erase operation, the data storage cells transition from the Vth distributions of the programmed data states, e.g., states A through G, to the erased state. The erase operation includes an erase phase in which the memory cells are biased for erasure, followed by an erase verification test. The erase verification test may use an applied erase verification voltage, VvEr.
[0114] Fig. Figure 9C shows the threshold voltage distribution of memory cells programmed from the Er state to the A state, in accordance with Fig. 9B, and a sampling voltage V13, which is used to determine the amount of programming of the memory cells. Instead of the multiple sampling voltages of Fig. 9A, a single sense voltage V13 can be used to evaluate the performance of a set of memory cells and characterize the Vth distribution. In this example, the memory cells are programmed from the Er-state Vth distribution 910 to the Vth distributions 920 and 921 after the first and second programming pulses of a programming operation, respectively. The Vth distribution 911 can be reached after additional programming pulses.
[0115] After each programming pulse, a sampling operation can be performed using V13. A number of memory cells having Vth>V13 can be counted, and the count can be compared to a threshold. If the count exceeds the threshold, the memory cells are considered to have reached a program milestone. The associated program loop number (e.g., 2) can be recorded as a form of programming setting(s) used to optimize future programming of the memory cells. The memory cells subject to counting can include, for example, all memory cells being programmed or the memory cells being programmed to the A state (the lowest programmed state).In the case of programming MLC memory cells, it is useful to focus on the beginning of the programming process when all or most of the memory cells have not yet completed programming, so that the Vth distribution represents the full range of programming speeds of the memory cells.
[0116] In Fig. In Figure 9C, voltage V13 is an example of an evaluation voltage.
[0117] Fig. Figure 9D shows the threshold voltage distribution of memory cells programmed from the Er state to the A state, in accordance with Fig. 9B, including sense voltages VvA-V16, which is used to determine the amount of programming of the memory cells. This example enables detection of the upper tail of the Vth distribution with a higher granularity. Multiple sense operations may be performed between the program voltage and the verify voltage as described herein to detect the upper tail of the Vth distribution. An example of such an approach is detecting the upper tail after a predetermined number of program pulses or loops (e.g., 2). In this example, the Vth distribution 921 is reached after the second program pulse, and the upper tail is detected as being between V14 and V15.
[0118] As an example, assume that the Vth distributions are reached in one, two, and four program loops, respectively. These loops may be a plurality of program loops of a program operation, wherein in each loop, a voltage driver is configured to apply a program voltage followed by a decreasing voltage to a selected word line, and sense circuits are configured to perform a repeated scan of the set of memory cells during the decreasing voltage. The repeated scan indicates that the set of memory cells reaches a program milestone in a program loop (e.g., loop 4 of the plurality of program loops). A control circuit may be configured to adjust the program setting(s) of a subsequent program operation based on the one program loop.Furthermore, the programming operation may include one or more additional programming loops in which the Vth tail detection and repeated sampling are not performed. Generally, read circuits are configured to perform repeated sampling of the set of memory cells during a decreasing voltage in one programming loop, but not in all programming loops of a plurality of programming loops of a programming operation.
[0119] In Fig. 9D, the voltages VvA to V16 are examples of evaluation voltages.
[0120] Fig. Figure 10A illustrates a voltage signal used in a series of programming loops in an exemplary programming procedure leading to the Vth distribution of Fig. 9A. As mentioned, SLC programming can be fast and involve one or sometimes two programming pulses in one approach. In Fig. 10A and Fig. In FIG. 10B, the horizontal axis represents a program loop number, and the vertical axis represents a voltage. The voltage signal 1050 includes a program pulse 1051 and the program verify voltage VvP 1052 in a first program loop, and a program pulse 1053 and the program verify voltage VvP 1054 in a second and final program loop. The voltage signal is applied to a word line selected for programming.
[0121] The initial programming voltage and step size can be adjusted for SLC programming of Fig. 10A compared to the MLC programming of Fig. 10B may be different.
[0122] Fig. Figure 10B illustrates a voltage signal used in a series of programming loops in an exemplary programming operation resulting in the Vth distribution of Fig. 9B. During a programming operation, programming loops are performed for a selected word line in a selected block. A programming loop includes a programming section in which a programming voltage or pulse is applied to the selected word line, followed by a verifying section in which a verifying signal is applied to the selected word line while one or more verification tests, referred to as programming verification tests, are performed for the associated memory cells. With the exception of the erased state, each assigned state has a verifying voltage used in the verification test for the state in a programming operation. In addition, as described herein, e.g., in Fig. 13C to 13F, repeated scans occur when the selected word line voltage is ramped down between the program voltage and the verify voltage.
[0123] The voltage signal 1000 includes a series of programming voltages, including an initial programming voltage 1001 applied to a word line selected for programming. In this example, the voltage signal includes programming voltages whose amplitude increases incrementally in one or more programming loops of a programming pass using a fixed or varying step size. This is referred to as incremental step-pulse programming, where the programming voltage starts at an initial level (see programming voltage 1001) and increases in each successive programming loop, for example, until the completion of the programming operation. The operation is successfully completed when the threshold voltages of the selected memory cells reach the verify voltages of the assigned data states.
[0124] A programming operation may include a single programming pass or multiple programming passes, with each pass using, for example, incremental step pulse programming.
[0125] The verification signal in each program loop, including example verification signal 1002, may include one or more verify voltages based on the assigned data states being verified for the program loop. The verification tests may include lower assigned data states, then intermediate assigned data states, and then higher assigned data states as the programming operations progress. For simplicity, the example verification signals represent three verify voltages.
[0126] For example, all memory cells may be in the erased state at the beginning of the programming process. After the programming process is completed, a Vth distribution similar to that in Fig. 9B, and data can be read from the memory cells using read voltages that are between the Vth distributions. At the same time, a read pass voltage, Vpass (e.g., 8 to 10 V), also called the pass voltage or turn-on voltage, is applied to the remaining (unselected) word lines. By testing whether the Vth of a given memory cell is above or below one or more of the read reference voltages, the system can determine the data state represented by a memory cell. These voltages are demarcation voltages because they demarcate between Vth ranges of different data states.
[0127] Additionally, the data being programmed or read can be arranged in pages. For example, with four data states or two bits per cell, two pages of data can be stored. An example encoding of bits for the Er, A, B, and C states is 11, 10, 00, and 01, respectively, in the high page bit (UP) / low page bit (LP) format. For a low page being read, VrA and VrC can be used, and for a high page being read, VrB can be used.
[0128] With eight data states, or three bits per cell, three pages of data can be stored. An example encoding of bits for the A, B, C, D, E, F, and G states is 111, 110, 100, 000, 010, 011, 001, and 101, respectively. The low page data can be determined by reading the memory cells using read voltages of VrA and VrE. The middle page data can be determined by reading the memory cells using read voltages of VrB, VrD, and VrF. The high page data can be determined by reading the memory cells using read voltages of VrC and VrG.
[0129] Fig. 11A illustrates an exemplary programming process with an option to scan a set of memory cells while decreasing a voltage of a selected wordline from a programming voltage to a verify voltage. In a program loop, the selected wordline voltage decreases from the programming voltage to the verify voltage during a transition period. During this transition, repeated scans may be performed to evaluate a programming voltage-induced programming effort of the memory cells. For example, as mentioned, the scans may identify an upper and / or lower tail of a Vth distribution.
[0130] At step 1100, a command is received to perform a programming operation for a set of memory cells connected to a selected word line of a block. For example, the command may be received by the controller 122 from the host 140 ( Fig. 1). Step 1100a includes generating a random number to decide whether to evaluate the performance of the set of memory cells in the programming operation. For example, an eight-bit random number may be generated and compared to a predetermined number, such as 11111111. If there is a match, the performance is evaluated. If there is no match, the performance is not evaluated. For an n-bit random number, the evaluation occurs on average once every 2^n times the wordline is programmed. For example, for n=8, the evaluation occurs on average every 256 times the wordline is programmed. This approach is useful because it avoids the overall cost of tracking the number of programming operations for each block and / or wordline.Additionally, a random number generator may be used to decide whether a word line, block, or other unit of memory cells in a memory device should be evaluated.
[0131] When an evaluation is performed, it may occur in one or more programming loops. In an approach such as SLC programming, where programming is usually completed in a programming loop, in accordance with Fig. 9A, the evaluation may occur during the first programming loop.
[0132] In other approaches, such as MLC programming, where the evaluation consists in determining when a set of memory cells reaches a program milestone, the evaluation can be performed in each programming loop, starting with the first programming loop until the program milestone is reached, in accordance with Fig. 9C and Fig. 9D. This can be done, for example, in 2 to 4 programming loops, as in the example of Fig. 9C and Fig. 9D. Evaluation may be omitted in subsequent programming loops within an approach. In some cases, the decision whether to perform evaluation in a programming loop may be based on the result of evaluation in a previous programming loop. For example, if a milestone is reached as a result of evaluation in one programming loop, evaluation may be omitted for subsequent programming loops of the program process.
[0133] Step 1101 includes reading the programming setting(s) from a memory location associated with the selected word line. The card 119 and the table of Fig. 4A or 4B can be used in this step. A program loop begins at step 1102. Step 1103 involves applying Vpgm to the selected wordline and applying a turn-on voltage to unselected wordlines of the block. The turn-on voltage can be applied to data wordlines, dummy wordlines, and selected gate transistors in one approach. Vpgm can be set based on the programming setting(s) discussed. After step 1103, step 1104 is reached if no evaluation is to be performed, or step 1104a is reached if evaluation is to be performed.
[0134] Step 1104 involves decreasing the selected wordline voltage from Vpgm to a verify voltage without performing a scan, while maintaining the turn-on voltage on the unselected wordlines. The turn-on voltage, e.g., 8-10 V, of a wordline ensures that the memory cells connected to the wordline are in a conductive state. The transition from the program voltage to the verify voltage can be direct in that the voltage is decreased but not increased. See, for example, Fig. 13B and graphical representation 1320 at t3 to t4. Step 1105 includes applying a verify voltage to the selected wordline while scanning the selected memory cells and maintaining the turn-on voltage on the unselected wordlines. See, for example, Fig. 13B, plots 1320 and 1321 at t4 to t9. A decision step 1106 determines whether one or more verification tests have passed, i.e., whether all or almost all memory cells subjected to a corresponding verification test pass the corresponding verification test by being in a non-conductive state when scanned, indicating that their Vth is greater than the verify voltage. If decision step 1106 is false, Vpgm is raised at step 1108, and a next program loop begins at step 1102. If decision step 1106 is true, the program operation occurs at step 1107.
[0135] Step 1104a involves applying a decreasing voltage to the selected wordline while repeatedly sampling the memory cells and maintaining the turn-on voltage on the unselected wordlines. The transition from the program voltage to the verify voltage may be direct in that the voltage is decreased but not increased. The sampling may be referred to as ramp-down sampling or decreasing ramp sampling and allows many sampling points to be reached quickly because the wordline voltage is not stabilized for each sampling point. This provides sampling with a coarse accuracy that is acceptable when evaluating the Vth distribution. In practice, good accuracy is achieved because the sampling voltages are relatively close to each other and the wordline voltage is ramped down at a controlled rate. See, for example, Fig. 13C, graphic representations 1340 and 1341 at t4 to t5, Fig. 13D, graphic representations 1360 and 1361 at t4-t5, Fig. 13E, plots 1380 and 1381 at t4 to t5, and Fig. 13F, plots 1480 and 1481 at t4 to t5 and t6 to t7. Step 1104b includes determining the updated programming setting(s) for the selected wordline based on the results of the scan. See, for example, Fig. 11C to 11E. Then, step 1105 is reached, in which the verification voltage is applied. See, for example, Fig. 13C, graphic representation 1340 at t5 to t7, Fig. 13E, graphic representation 1380 at t6 to t8, and Fig. 13F, graphic representation 1480 at t7 to t8.
[0136] One option is to stabilize the wordline voltage for one or more sample points during the ramp-down sampling so that the ramp-down voltage on the selected wordline resembles, for example, a decreasing staircase waveform.
[0137] In one approach, the programming setting(s) previously determined for a given set of memory cells are used in the subsequent programming of the same set of memory cells. However, it is also possible for one or more programming settings previously determined for a first set of memory cells to be shared and used in a subsequent programming of a different, second set of memory cells. This can be done if the first and second sets of memory cells are expected to have similar programming performance. For example, a programming operation for a block can program the memory cells connected to a word line in a first sub-block (e.g., WLL0 in SB0, Fig. 7), but not the memory cells connected to the word line in a second sub-block (e.g. WLL0 in SB1, Fig. 7). In this case, the programming setting(s) obtained for the memory cells of the first sub-block can be used for the subsequent programming of these memory cells as well as the memory cells of the second sub-block.
[0138] Furthermore, in one approach, the second sub-block can be omitted from the evaluation process when it is programmed. Or, an evaluation of the second sub-block can be used to update the programming setting(s), which are subsequently used to guide the programming of the first or second sub-block.
[0139] Fig. 11B illustrates an exemplary process for scanning a set of memory cells while decreasing the voltage of a selected word line, as in step 1104a of Fig. 11A. Typically, the scanning of the set of memory cells includes a voltage range that is substantially below the program voltage. Accordingly, the selected wordline voltage may be reduced from the program voltage to an intermediate voltage that is used as the first voltage for evaluation. The selected wordline voltage may then be reduced from the intermediate voltage to a final voltage that may be equal to or higher than the verify voltage (or the initial verify voltage if more than one is present). For example, in Fig. 9A, the intermediate voltage can be V5 or V11 if the evaluation includes the lower or upper tail, respectively, and the final voltage can be VvP or V6. In Fig. 9D, the intermediate voltage can be V16 and the final voltage VvA. In one approach, the intermediate voltage is fixed and independent of the programming setting(s).
[0140] In another approach, an intermediate voltage is determined in step 1119 based on the programming setting(s) in the memory location for the set of memory cells. This approach adaptively adjusts the intermediate voltage to potentially reduce the total number of samples that occur and thus reduce the sampling time between the program voltage and the verify voltage. For example, if it is known that the tail of the Vth distribution was considered relatively low during the last evaluation, the intermediate voltage may be set to a relatively low value, since the tail is not expected to have shifted significantly higher since the last evaluation.
[0141] Sampling circuits can be configured to start the repeated sampling of the set of memory cells when the decreasing voltage is at a specific voltage based on one or more programming settings that indicate the degree of degradation of the set of memory cells.
[0142] Step 1120 includes reducing the voltage of the selected word lines from the programming voltage to the intermediate voltage at a first rate. This may be the highest rate available to the voltage driver 447 ( Fig. 3) and may include the request for a step change in the voltage. Due to limitations of the voltage driver and the RC time constant of the word lines, there is a finite delay between the requested voltage and the actual word line voltages. See, for example, Fig. 13C, graphic representation 1340 at t3 to t4, Fig. 13D, graphic representation 1360 at t3 to t4, Fig. 13E, graphic representation 1380 at t3 to t4 and Fig. 13F, plot 1480 at t3 to t4. Step 1121 includes beginning a controlled ramp-down of the selected wordline voltage at a second rate, the second rate being an order of magnitude lower than the first rate. See, for example, Fig. 13C, graphic representation 1340 at t4 to t5, Fig. 13D, graphic representation 1360 at t4 to t5, Fig. 13E, graphic representation 1380 at t4 to t5 and Fig. 13F, graphic representation 1480 at t4 to t5 and t6 to t7.
[0143] Step 1122 includes sampling the selected memory cells at sampling times during the controlled shutdown. For example, the memory cells may be repeatedly sampled at specific time steps. See, for example, Fig. 13C, graphic representation 1340 at t4 to t5, Fig. 13D, graphic representation 1360 at t4 to t5, Fig. 13E, graphic representation 1380 at t4 to t5 and Fig. 13F, graphical representation 1480 at t4 to t5 and t6 to t7, where the circles at the specified time intervals indicate the times at which the sampling occurs and the corresponding word line voltages.
[0144] Based on the sampling, step 1123 includes counting the number of memory cells in a conductive or non-conductive state. Step 1124 includes comparing the number to a threshold. For example, if a lower tail of a Vth distribution is detected during the evaluation, the transition of the count of conductive memory cells below a threshold (e.g., 1 to 3%) while the selected wordline voltage is decreased, in one approach, indicates the presence of the lower tail of the Vth distribution. Referring to Fig. For example, in Figure 9A, it is assumed that the lower tail of the Vth distribution 902 is sensed and that the wordline voltages VvP-V5 are applied in decreasing order, with V5 first. The number of memory cells sensed as conductive is gradually decreased as V5, V4, and V3 are applied. The number will be zero to or near zero, e.g., perhaps less than 1 to 3% of the total number of memory cells in the selected wordline, when V2 is applied. This indicates that the lower tail of the Vth distribution 902 falls within the voltage range V2 to V3.
[0145] Or suppose the upper tail of the Vth distribution 902 is sensed, and the wordline voltages V6 to V11 are applied in decreasing order, with V11 first. The number of memory cells sensed as non-conductive will initially be zero or close to zero, e.g., perhaps less than 1 to 3% of the total number of memory cells in the selected wordline when V11 and V10 are applied. This indicates that the upper tail of the Vth distribution 902 falls within the voltage range V10 to V11. The transition of the number of non-conductive memory cells above a threshold (e.g., 1 to 3%) as the selected wordline voltage decreases, in one approach, signifies the presence of the upper tail of the Vth distribution.
[0146] A decision step 1125 determines whether another sample should be performed. As mentioned previously, in one approach, there are a predetermined number of sample times during the decreasing voltage. In another approach, the sample stops based on the results of the previous sample during the decreasing voltage. For example, if the tail is detected, it is possible to terminate any further sample of the evaluation and request that the voltage driver output the verify voltage to proceed with a verification test. Or, if no verification test is performed, the voltage driver can be requested to output a different program loop end voltage, such as 0 V. If decision step 1125 is correct, step 1122 is repeated.If decision step 1125 is false, step 1126 is optionally reached, in which the selected wordline voltage is already at a verify voltage or another terminal voltage. In this case, the verify voltage is the last voltage in the controlled ramp-down of the selected wordline voltage. See, for example, the voltage represented by a circle 1348 in FIG. Fig. 13C, which corresponds to VvP or VvA, and which is represented by a circle 1468 in Fig. 13F, which corresponds to VvP.
[0147] In another option, step 1127 is reached, which involves reducing the selected wordline voltage to a verify voltage or another final voltage. In this case, the verify voltage is lower than the last voltage in the ramp-down sample of the selected wordline voltage. The transition to the verify voltage in step 1127 may involve requesting a step change in the voltage of the voltage driver. For example, see the transition from the voltage represented by a circle 1368 to the voltage represented by a circle 1371a (e.g., 0 V) in Fig. 13D and the transition from the voltage represented by a circle 1388 to the voltage represented by a circle 1389 (e.g. VvP) in Fig. 13E.
[0148] Fig. 11C illustrates an exemplary process for determining the updated programming setting(s) based on sensor data, as in step 1104b of Fig. 11A. Step 1110 includes inputting sensor data, e.g., to the controller 122 or control circuit 110 ( Fig. 1). In one option, step 1111 includes identifying a tail of a Vth distribution of the set of memory cells, as described in connection with Fig. 9A and Fig. 9D, and step 1112 includes determining updated programming setting(s) based on the tail (e.g., based on the voltage range into which the tail falls). See also Fig. 11D. In another option, step 1113 includes determining a programming loop in which a program milestone is reached, as in connection with Fig. 9C, and step 1114 includes determining updated programming setting(s) based on the programming loop. See also Fig. 11E. In both options, step 1115 includes saving the updated programming setting(s) to the storage location.
[0149] Fig. 11D illustrates a graphical representation of an initial Vpgm versus the detected end of a Vth distribution of a set of memory cells, in an exemplary implementation of step 1112 in Fig. 11C. The detected Vth of the tail can be in one of the voltage ranges from VvP to V1, V1 to V2, V2 to V3, V3 to V4 or V4 to V5, in accordance with Fig. 9A. As already mentioned, programming settings such as an initial programming voltage (see e.g. Fig. 10A and Fig. 10B), for a set of memory cells, such as memory cells connected to a word line. The adjustment may be based on the Vth of the tail, as determined by a ramp-down sample performed between the program voltage and the verify voltage in a program loop. In this example, the initial Vpgm decreases as the Vth of the tail (e.g., the upper or lower tail) increases. If the Vth of the tail is relatively high, this indicates a relatively high overprogramming or degradation of the set of memory cells and a relatively high programming speed, so a relatively lower initial Vpgm should be used in subsequent program operations as compensation to reduce overprogramming.
[0150] The position of the tail indicates the degree of overprogramming of the memory cells. The degree is greater when the Vth of the lower tail is higher.
[0151] Fig. 11E illustrates a graphical representation of an initial Vpgm versus a loop count at which a set of memory cells reaches a programming milestone, in an example implementation of step 1114 of Fig. 11C. Similar to the approach of Fig. 11D, the programming setting(s), such as an initial programming voltage, for a set of memory cells can be adjusted based on a program loop in which a program milestone is reached during a program operation. The Vth of the tail and the program loop in which a program milestone is reached are both indicators of the programming speed of the memory cells. In this example, the initial Vpgm decreases as the program loop decreases. If the program loop to reach a certain program milestone is relatively low, this indicates a relatively high programming speed, so a relatively lower initial Vpgm should be used to reduce the programming speed in subsequent program operations.
[0152] Fig. 11F illustrates a graphical representation of an intermediate voltage (Vint) versus a lower end of a Vth distribution of a set of memory cells, in an exemplary implementation of step 1119 of Fig. 11B. Vint is the starting voltage of a decreasing ramp voltage used during ramp-down sampling. Examples of decreasing ramp voltage are shown in Fig. 13C, graphic representation 1340 at t4 to t5, Fig. 13D, graphic representation 1360 at t4 to t5, Fig. 13E, graphic representation 1380 at t4 to t5 and Fig. 13F, graphic representation 1480 shown at t4 to t5 and t6 to t7.
[0153] The detected Vth of the tail can be in one of the voltage ranges from VvP to V1, V1 to V2, V2 to V3, V3 to V4 or V4 to V5, in accordance with Fig. 9A. As mentioned earlier, the scan time can be reduced by reducing Vint, the initial scan voltage on the selected word line, during the scan that occurs between the program voltage and the verify voltage. Reducing Vint reduces the number of sample points and thus the total scan time during ramp-down scan. This can be done primarily for a new memory device with a low number of cycles. As the tail Vth increases, Vint should also increase. For example, Vint can be set to a specified voltage step higher than the tail Vth. For example, if the tail Vth is in the range of VvP to V1, V1 to V2, or V2 to V3, Vint can be set two steps higher at V3, V4, or V5, respectively. If the tail Vth is above V3, Vint is set to the maximum level of V5.
[0154] Fig. 12 illustrates an exemplary process for storing updated programming setting(s) in one or more reserve sets of NAND strings, beginning with step 1115 of Fig. 11C and with Fig. 11D and Fig. 11E. The updated programming setting(s) may, for example, change the initial Vpgm from Fig. 11D or Fig. 11E. It should be noted that Vint, which is Fig. 11F may be determined from the programming setting(s), or Vint may be stored as a separate value in addition to the programming setting(s). For example, Vint may increase as the initial programming voltage decreases. Step 1219 includes inputting the updated programming setting(s) to the controller 122 or control circuit 110, for example, based on a memory location read. Step 1220 includes identifying transistors in one or more replacement NAND chains to store the programming setting(s). As previously mentioned, for example, in connection with Fig. 8C and Fig. 8D, the transistors suitable for storing the programming setting(s) may include dummy memory cells and / or select gate transistors in replacement NAND chains.
[0155] At step 1221, a program loop begins. Step 1222 includes applying Vpgm to a selected control line (e.g., an SGD or SGS control line) or a selected dummy word line (WL) and applying a turn-on voltage to unselected control lines and unselected word lines (both unselected dummy lines and data word lines) of the block. Step 1223 includes applying a verify voltage to the selected control line or the selected dummy word line while scanning the selected transistors and while maintaining the turn-on voltage on the unselected control lines and unselected word lines. If the verification test passes at step 1224 and there is no next transistor to be programmed at decision step 1225, the program operation is performed at step 1226.If the verification test fails in step 1224, Vpgm is incremented in step 1227 and a next programming loop begins at step 1221. If a next transistor to be programmed is present in decision step 1225, step 1220 is reached.
[0156] The programming of the transistors can, as mentioned, involve SLC or MLC programming, and the programming process can be set accordingly, e.g., similar to Fig. 10A or 10B.
[0157] The programming setting(s) may be stored as a non-volatile identifier in units such as bytes or other bit sets. Furthermore, updating the programming setting(s) may involve changing only a portion of the data in a manner that avoids the need for an erase operation. For example, assume that a 0 bit is stored in the Er state and a 1 bit is stored in the P state. For example, if the programming setting(s) is / are an initial programming voltage, the bit sets representing different initial programming voltages may be configured assuming a temporal increase in the initial programming voltage.In one example, a bit sequence of 00000000 corresponds to a predefined initial programming voltage, a bit sequence of 10000000 corresponds to the predefined initial programming voltage minus a delta voltage, a bit sequence of 11000000 corresponds to the predefined initial programming voltage minus twice the delta voltage, and so on. In this case, updating the programming setting(s) involves programming transistors to change selected bits from 0 to 1, and it is not necessary to erase the transistors to update the programming setting(s). This makes saving the programming setting(s) easier.
[0158] In Fig. In Figures 13A to 13F, the vertical axis represents a wordline voltage (Vwl) and the horizontal axis represents time, with the time scales not necessarily being the same. Also, circles and squares denote a requested output voltage of the selected and unselected wordline voltage drivers 447 and 448, respectively ( Fig. 3). The requested output voltage at a given voltage is maintained until a new requested output with a different voltage is received. As mentioned earlier, there is a delay between the request and the actual output voltages. The period shown includes a programming loop.
[0159] Fig. 13A illustrates exemplary voltage signals applied to word lines during a program loop, wherein the voltages on the selected and unselected word lines are reduced to 0 V between the program voltage and the verify voltages, as a comparison example to steps 1104 and 1104a of the programming process of Fig. 11A. In this comparative example, the total time used by the program loop is relatively high due to the time used to decrease the voltages on the selected and deselected word lines to 0 V from t3 to t5, the time used to subsequently increase the voltages from t6 to t8, and the time period from t5 to t6. The program loop includes a program section from t0 to t5 and a verify section from t6 to t14.
[0160] In the programming section, graphic representation 1300 represents the selected wordline voltage. The requested voltage is 0 V at t0 (circle 1310a), Vturn on at t1 (circle 1311a), Vpgm at t2 (circle 1312), and 0 V at t3 (circle 1313a). Graphic representation 1301 represents the unselected wordline voltage. The requested voltage is 0 V at t0 (square 1310), Vturn on at t1 (square 1311), and 0 V at t3 (square 1313).
[0161] In the verification section, graphical representation 1303 represents the selected wordline voltage. The requested voltage is VvA at t7 (circle 1316), VvB at t9 (circle 1317), VvC at t11 (circle 1318), and 0 V at t13 (circle 1319a). Graphical representation 1302 represents the unselected wordline voltage. The requested voltage is Vturnon at t6 (square 1315) and 0 V at t13 (square 1319). The verification tests for states A, B, and C occur at t8, t10, and t12, respectively.
[0162] Fig. Figure 13B illustrates exemplary voltage signals generated during a programming loop, in accordance with Fig. 11A and step 1104, to word lines, wherein the selected word line voltage is reduced from the programming voltage to the lowest verify voltage of a plurality of verify voltages without performing a ramp-down scan, in accordance with Fig. 9B. In Fig. 13B to 13F, the unselected word line voltage remains at a power-up voltage during the programming and verification phase of sections of the programming loop. This approach represents a significant improvement compared to Fig. 13A provides a time saving because the selected wordline voltage is not reduced below the initial verify voltage VvA and because the unselected wordline voltage is maintained at Vturnon. Note that Vturnon may be different during the program portion compared to the verify portion of the program loop, but in this example, it is at a constant level.
[0163] The program loop includes a program section from t0 to t3 and a verify section from t4 to t9. Graphic representation 1320 represents the selected wordline voltage. The requested voltage is 0 V at t0 (circle 1324a), Vturn on at t1 (circle 1325a), Vpgm at t2 (circle 1326), and VvA at t3 (circle 1327). The program voltage may temporarily pause at a level such as Vturn on to avoid a single large transition that can cause unwanted coupling effects. The requested voltage is VvA at t3 (circle 1327), VvB at t5 (circle 1329), VvC at t7 (circle 1330), and 0 V at t9 (circle 1332a). Note that the selected wordline voltage VvA is reached and maintained at this voltage at t4 to allow an A-state verification test to be performed. Verification tests for states A, B, and C occur at t4a, t6, and t8, respectively.
[0164] Graphic representation 1321 represents the unselected word line voltage. The requested voltage is 0 V at t0 (square 1324), Vturn on at t1 (square 1325) and 0 V at t9 (square 1332).
[0165] Fig. Figure 13C illustrates exemplary voltage signals applied to word lines during a program loop, in accordance with Fig. 11A and step 1104a, wherein the selected word line voltage is decreased from the programming voltage (programming level) to a verify voltage (verify level) while performing a ramp-down scan to detect a lower tail of a Vth distribution, in accordance with Fig. 9A, voltages VvP to V5, or an upper tail of a Vth distribution, in accordance with Fig. 9D, voltages VvA-V16. The programming loop includes a programming section from t0-t3, a ramp-down sampling section from t4-t5, and a verification section from t5-t7.
[0166] The graphical representation 1340 represents the selected word line voltage. The requested voltage is 0 V at t0 (circle 1343a), Vturn on at t1 (circle 1344a), Vpgm at t2 (circle 1345) and a reduced level (Vint) of V5 or V16 at t3 (circle 1346) for the examples of Fig. 9A and 9D, respectively. The selected wordline voltage is ramped down at a controlled rate from t4 to t5 while the periodic sampling occurs as indicated by each circle. See example circles 1347, 1350, and 1348, which may represent voltages at the first, second, and sixth samples, respectively. The selected wordline voltage is maintained at VvP or VvA from t5 to t7 to allow for an optional verification test at t6. In this example, VvP and VvA are assumed to be equal, and Vv5 and Vv16 are assumed to be equal for simplicity. The requested voltage is 0 V at t7 (circle 1349a).
[0167] In one option, circuit 1348 represents the verification voltage, and an associated sample is taken at t5 as part of the ramp-down sample and Vth tail detection. An additional sample at the same verify voltage may be taken at t6 as part of a verification test, when the selected wordline voltage has time to stabilize from t5 to t6. This may result in a more accurate sample compared to the coarse sample taken while the wordline voltage is ramped down. Optionally, as in Fig. 13D, the last measurement point in the ramp-down sampling and the Vth-tail evaluation is also used as the verification result.
[0168] Graphic representation 1341 represents the unselected word line voltage. The requested voltage is 0 V at t0 (square 1343), Vturn on at t1 (square 1344), and 0 V at t7 (square 1349).
[0169] Fig. Figure 13D illustrates exemplary voltage signals applied to word lines during a program loop, in accordance with Fig. 11A and step 1104a, wherein the selected wordline voltage is decreased from the programming voltage while performing a ramp-down scan to detect a lower tail of a Vth distribution or an upper tail of a Vth distribution, in accordance with Fig. 9D, voltages VvA to V16, and no verification test is performed or the final sampling result is used as the verification result, in accordance with Fig. 9A, voltages VvP to V5. The programming loop includes a programming section from t0 to t3 and a ramp-down sampling section from t4 to t5.
[0170] The graphical representation 1360 represents the selected word line voltage. The requested voltage is 0 V at t0 (circle 1363a), Vturn on at t1 (circle 1364a), Vpgm at t2 (circle 1365) and V5 or V16 at t3 (circle 1366) for the examples of Fig. 9A and 9D, respectively. The selected wordline voltage is ramped down at a controlled rate from t4 to t5 while the periodic sampling occurs as indicated by each circle. See example circles 1367 and 1368, which may represent voltages at the first and sixth samples, respectively. The selected wordline voltage is reduced directly to 0 V at t6 without being stabilized at a level that enables a high-accuracy verification test, as in Fig. 13C. In this example, VvP and VvA are assumed to be equal, and Vv5 and Vv16 are assumed to be equal for simplicity. The requested voltage is 0 V at t6 (circuit 1371a).
[0171] In one option, circuit 1368 represents the verification voltage, and an associated sample is taken at t5. This verification test can be sufficiently accurate even if the wordline voltage is not stabilized. This approach saves time because additional time is not spent stabilizing the wordline voltage.
[0172] Graphic representation 1361 represents the voltage on the unselected word lines. The requested voltage is 0 V at t0 (square 1363), Vturn on at t1 (square 1364), and 0 V at t6 (square 1371).
[0173] Fig. 13E illustrates exemplary voltage signals applied to word lines during a program loop, in accordance with Fig. 11A and step 1104a, wherein the selected wordline voltage is decreased from the program voltage to a verify voltage while performing a ramp-down scan to detect a lower tail of a Vth distribution, in accordance with Fig. 9A, voltages VvP to V5, or an upper tail of a Vth distribution, in accordance with Fig. 9D, voltages VvA to V6 to V11. Since the range of sense voltages for the upper tail of a Vth distribution is higher than the verify voltage, the selected wordline voltage is decreased one notch after the ramp-down sense and before the verify test, e.g., from t5 to t6.
[0174] The programming loop includes a programming section from t0 to t3, a ramp-down sampling section from t4 to t5, and a verification section from t6 to t8.
[0175] Graphic representation 1380 represents the selected wordline voltage. The requested voltage is 0 V at t0 (circle 1383a), Vturn on at t1 (circle 1384a), Vpgm at t2 (circle 1385), and V11 at t3 (circle 1386). The selected wordline voltage is ramped down at a controlled rate from t4 to t5 while periodically sampling occurs as indicated by each circle. See example circles 1387 and 1388, which may represent voltages at the first and sixth sampling times, respectively. The selected wordline voltage decreases directly to VvP at t6 and is stabilized for a verification test at t7. The requested voltage is 0 V at t8 (circle 1390a).
[0176] Graphic representation 1381 represents the voltage on the unselected word lines. The requested voltage is 0 V at t0 (square 1383), Vturn on at t1 (square 1384), and 0 V at t8 (square 1390).
[0177] Fig. Figure 13F illustrates exemplary voltage signals applied to word lines during a program loop, in accordance with Fig. 11A and step 1104a, wherein the selected wordline voltage is reduced from the programming voltage to a verify voltage while performing a first ramp-down sampling sequence (at t4 to t5) to detect a lower tail of a Vth distribution, in accordance with Fig. 9A, voltages V6 to V11, and a second sequence of ramp-down sampling (at t6 to t7) is performed to detect an upper tail of a Vth distribution, in accordance with Fig. 9A, voltages VvP to V5.
[0178] The programming loop includes a programming section from t0-t3, a ramp-down sampling section from t4 to t7, and a verification section from t7 to t8.
[0179] Graphic representation 1480 represents the selected wordline voltage. The requested voltage is 0 V at t0 (circle 1483a), Vturn on at t1 (circle 1484a), Vpgm at t2 (circle 1485), and V11 at t3 (circle 1486). The selected wordline voltage is ramped down from V11 to V6 at a controlled rate in a first sequence from t4-t5 while periodically sampling as indicated by each circle. See example circles 1487 and 1488, which may represent voltages at the first and sixth sampling events and sampling times in the first sequence, respectively. The selected wordline voltage is then requested to go down to V5 at t6 (circle 1487).
[0180] The selected wordline voltage is then ramped down from V5 to VvP at a controlled rate in a second sequence from t6 to t75, while periodically sampling occurs as indicated by each circle. See example circles 1467 and 1468, which may represent voltages at the first and sixth sampling events and sampling times in the second sequence, respectively. The selected wordline voltage is then stabilized for a verification test at t7a. The requested voltage is 0 V at t8 (circle 1490a).
[0181] Graphic representation 1481 represents the voltage on the unselected word lines. The requested voltage is 0 V at t0 (square 1483), Vturn on at t1 (square 1484), and 0 V at t8 (square 1490).
[0182] In one implementation, a device comprises: a set of memory cells connected to a selected wordline; a voltage driver configured to apply a programming voltage followed by a decreasing voltage to the selected wordline in a programming operation; and sampling circuitry configured to perform repeated sampling of the set of memory cells during the decreasing voltage.
[0183] In another implementation, a method comprises: setting a control gate voltage of a set of memory cells to a programming voltage; setting the control gate voltage of the set of memory cells to a verify voltage for a verification test; decreasing the control gate voltage at a first rate from the programming voltage to an intermediate voltage that is between the programming voltage and the verify voltage; decreasing the control gate voltage at a second rate, which is lower in magnitude than the first rate, from the intermediate voltage to the verify voltage; and scanning the set of memory cells while decreasing the control gate voltage at the second rate.
[0184] In another implementation, a device comprises: a set of memory cells in a block; means for applying a programming voltage to the set of memory cells; means for reducing a voltage of the set of memory cells from the programming voltage to a verify voltage; and means for performing a verification test for the set of memory cells when the voltage of the set of memory cells reaches the verify voltage.
[0185] The means for applying a programming voltage may comprise the power control module 116 of Fig. 1, the voltage driver 447, the row decoder 124 and the pass transistors 422 of Fig. 3 or other logic hardware and / or other executable code stored on a computer-readable storage medium or device. Other embodiments may include similar or equivalent means.
[0186] The means for reducing the voltage may also comprise the power control module 116 of Fig. 1, the voltage driver 447, the row decoder 124 and the pass transistors 422 of Fig. 3 or other logic hardware and / or other executable code stored on a computer-readable storage medium or device. Other embodiments may include similar or equivalent means.
[0187] The means for performing a verification test may comprise the power control module 116 and the sampling blocks 51 to 53 of Fig. 1, the scanning block 51 of Fig. 2 and the voltage driver 447, the row decoder 124 and the pass transistors 422 of Fig.3 or other logic hardware and / or other executable code stored on a computer-readable storage medium or device. Other embodiments may include similar or equivalent means.
[0188] In another implementation, a device comprises: a plurality of NAND chains arranged in a block, each NAND chain comprising data memory cells connected to data word lines; a voltage driver configured to apply a programming voltage, an evaluating voltage, and a verifying voltage to a selected data word line; and sensing circuits configured to detect a tail of a threshold voltage distribution of memory cells connected to the selected data word line during the evaluating voltage and to determine whether the memory cells connected to the selected data word line pass a verifying test during the verifying voltage, wherein the evaluating voltage is greater than the verifying voltage.
[0189] In another implementation, a system comprises: a controller configured to receive a programming command; and a memory chip coupled to the controller. The memory chip comprises: a set of coupled memory cells; a voltage driver configured to, in response to the programming command, apply a control gate voltage at a program level to the set of coupled memory cells, reduce the control gate voltage from the programming level to a reduced level, and reduce the control gate voltage at a controlled ramp rate from the reduced level to a verify level; and sampling circuitry configured to repeatedly sample the set of coupled memory cells during ramping down of the control gate voltage and configured to sample the set of coupled memory cells during application of the control gate voltage at the verify level.
[0190] The foregoing detailed description of the invention has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form disclosed. Many modifications and variations are possible in light of the above teachings. The described embodiments were chosen to best explain the principles of the invention and its practical application, to thereby enable others skilled in the art to best utilize the invention in various embodiments and with various modifications as are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the claims appended hereto.
Claims
[1] Device comprising: a set of memory cells (704-714, 724-734, 744-754, 764-774) connected to a selected word line (WL0-WL10); a voltage driver (447) configured to apply a programming voltage followed by a decreasing voltage to the selected word line in a programming operation; and Sampling circuits (51-53, 180-183) arranged to perform repeated sampling of the set of memory cells during the decreasing voltage; wherein the decreasing voltage is decreased from the programming voltage at a first rate and then at a second rate that is lower in magnitude than the first rate; and wherein the sampling circuits (51-53, 180-183) are arranged to perform repeated sampling of the set of memory cells as the decreasing voltage is reduced at the second rate. [2] The device according to claim 1, further comprising: a control circuit (110, 122) configured to evaluate a degree of overprogramming of the set of memory cells based on the repeated scanning of the set of memory cells and to adjust the programming setting(s) of a subsequent programming operation based on the degree of overprogramming. [3] Device according to claim 1 or 2, further comprising: a control circuit (110, 122) configured to detect a tail of a threshold voltage distribution of the set of memory cells based on the repeated scanning of the set of memory cells. [4] Device according to claim 3, wherein: the sampling circuits (51-53, 180-183) are arranged to stop the repeated sampling of the set of memory cells when the tail of the threshold voltage distribution of the set of memory cells is detected. [5] Device according to one of claims 1 to 4, wherein: in each of a plurality of programming loops of the programming process, the voltage driver (447) is arranged to apply the programming voltage followed by the decreasing voltage to the selected word line, and the sampling circuits are arranged to perform the repeated sampling of the set of memory cells during the decreasing voltage; and the repeated sampling indicates that the set of memory cells reaches a program milestone in a programming loop of the plurality of programming loops. [6] Device according to claim 5, further comprising: a control circuit (110) configured to adjust the programming setting(s) of a subsequent programming operation based on said one programming loop. [7] Device according to one of claims 1 to 6, wherein: the sampling circuits (51-53, 180-183) are arranged to start the repeated sampling of the set of memory cells when the decreasing voltage is at a voltage based on the programming setting(s) of the set of memory cells. [8] Device according to one of claims 1 to 7, wherein: the voltage driver (447) is arranged to apply a verify voltage to the selected word line in the programming process after the application of the decreasing voltage, wherein the decreasing voltage is reduced to the verify voltage; and the sampling circuits (51-53, 180-183) are arranged to sample the set of memory cells during the application of the verification voltage. [9] Apparatus according to claim 8, further comprising: a set of memory cells connected to a non-selected word line; and a voltage driver (448) configured to continuously apply a turn-on voltage (Vturn on) to the unselected word line during the program voltage, the decreasing voltage, and the verify voltage, the turn-on voltage providing the set of memory cells connected to the unselected word line in a conductive state. [10] Device according to one of claims 1 to 9, wherein: the sampling circuits are arranged to perform the repeated sampling of the set of memory cells during the decreasing voltage in one programming loop, but not in all programming loops of a plurality of programming loops of the programming process. [11] Device according to one of claims 1 to 10, further comprising: a set of bit lines (BL0-BL3) connected to the set of memory cells, wherein the sensing circuits (51-53, 180-183) detect a voltage on the bit lines during the repeated sensing. [12] Method comprising: Setting a control gate voltage of a set of memory cells (704-714, 724-734, 744-754, 764-774) to a programming voltage (Vpgm); Setting the control gate voltage of the set of memory cells to a verification voltage for a verification test; reducing (1120) the control gate voltage at a first rate from the programming voltage to an intermediate voltage (Vint) that is between the programming voltage and the verify voltage; reducing (1121) the control gate voltage at a second rate, which is lower in magnitude than the first rate, from the intermediate voltage to the verify voltage; and Sampling (1122) the set of memory cells while decreasing the control gate voltage at the second rate. [13] A method according to claim 12, wherein: the set of memory cells is not sampled while decreasing the control gate voltage at the first rate; and the set of memory cells is repeatedly scanned at the second rate while decreasing the control gate voltage. [14] The method according to claim 12 or 13, further comprising: Storing data indicating a voltage range of a tail of a threshold voltage distribution of the set of memory cells; and Adjusting the programming setting(s) of a programming operation for the set of memory cells based on the voltage range.
Citation Information
Patent Citations
Method of programming memory device
US20080316820A1
Piggyback programming with staircase verify for multi-level cell flash memory designs
US6671207B1
