阅读说明：本技术 半导体存储装置及其控制方法 (Semiconductor memory device and control method thereof ) 是由 石田贵士 菅野裕士 于 2021-01-25 设计创作，主要内容包括：实施方式提供一种能够提高对存储单元晶体管的擦除动作的可靠性的半导体存储装置及其控制方法。根据一实施方式,半导体存储装置具备与多个存储单元晶体管电连接的第一布线。所述装置还具备对所述第一布线施加第一电压而擦除存储于所述存储单元晶体管的数据的擦除部。而且,所述擦除部以所述第一电压上升到第一值、从所述第一值下降到第二值、维持在所述第二值的方式施加所述第一电压。(Embodiments provide a semiconductor memory device and a control method thereof capable of improving reliability of an erase operation of a memory cell transistor. According to one embodiment, a semiconductor memory device includes a first wiring electrically connected to a plurality of memory cell transistors. The device further includes an erasing unit configured to erase data stored in the memory cell transistor by applying a first voltage to the first wiring. The erase unit applies the first voltage to raise the first voltage to a first value, lower the first voltage from the first value to a second value, and maintain the first voltage at the second value.)

1. A semiconductor memory device includes:

a first wiring electrically connected to the plurality of memory cell transistors; and

an erasing section for erasing data stored in the memory cell transistor by applying a first voltage to the first wiring,

the erasing unit applies the first voltage to the first element such that the first voltage rises to a first value, falls from the first value to a second value, and is maintained at the second value.

2. The semiconductor memory device according to claim 1,

the first wiring is a source line.

3. The semiconductor memory device according to claim 1,

the first wiring is a bit line.

4. The semiconductor storage device according to any one of claims 1 to 3,

the semiconductor memory device further includes:

a selection transistor provided between the first wiring and the memory cell transistor; and

a second wiring electrically connected to the selection transistor,

the erasing unit erases data stored in the memory cell transistor by using a current generated by applying the first voltage to the first wiring and applying a second voltage different from the first voltage to the second wiring.

5. The semiconductor memory device according to claim 4,

the first wiring is a source line, and the second wiring is a source side selection line.

6. The semiconductor memory device according to claim 4,

the first wiring is a bit line, and the second wiring is a drain side select line.

7. The semiconductor memory device according to claim 1,

the plurality of memory cell transistors includes:

a plurality of word lines stacked above the substrate so as to be separated from each other;

a charge accumulation layer provided in the word line; and

a semiconductor layer provided in the word line with the charge storage layer interposed therebetween,

the first wiring is electrically connected to the semiconductor layer.

8. The semiconductor memory device according to claim 7,

the first wiring is provided between the substrate and the word line.

9. The semiconductor memory device according to claim 7,

the first wiring is provided above the word line.

10. A control method of a semiconductor memory apparatus, comprising the steps of:

the first voltage is generated by the erase portion,

erasing data stored in the memory cell transistor by applying the first voltage to a first wiring electrically connected to the memory cell transistors through the erasing portion,

the erasing unit applies the first voltage to the first element such that the first voltage rises to a first value, falls from the first value to a second value, and is maintained at the second value.

Technical Field

Embodiments of the invention relate to a semiconductor memory device and a control method thereof.

Background

In a semiconductor memory device, it is desired to improve reliability of an erasing operation of a memory cell transistor.

Disclosure of Invention

Embodiments provide a semiconductor memory device and a control method thereof capable of improving reliability of an erase operation of a memory cell transistor.

According to one embodiment, a semiconductor memory device includes a first wiring electrically connected to a plurality of memory cell transistors. The device further includes an erasing unit configured to erase data stored in the memory cell transistor by applying a first voltage to the first wiring. The erase unit applies the first voltage to raise the first voltage to a first value, lower the first voltage from the first value to a second value, and maintain the first voltage at the second value.

Drawings

Fig. 1 is a block diagram showing a configuration of a semiconductor memory device according to a first embodiment.

Fig. 2 is a circuit diagram showing the configuration of the NAND string of the first embodiment.

Fig. 3 is a graph for explaining the erasing operation of the first embodiment.

Fig. 4 is a sectional view showing the structure of the semiconductor memory device according to the first embodiment.

Fig. 5 is a cross-sectional view showing the structure of a semiconductor memory device according to a modification of the first embodiment.

Fig. 6 is a graph for explaining the erasing operation of the first embodiment.

Fig. 7 is another graph for explaining the erasing operation of the first embodiment.

Fig. 8 is another graph for explaining the erasing operation of the first embodiment.

Description of the reference numerals

1: memory cell array, 2: sense amplifier, 3: row decoder, 4: drive circuit, 5: BL/SL driver, 6: voltage generation circuit, 7: detection circuit, 8: control circuit, 11: substrate, 12: interlayer insulating film, 13: interlayer insulating film, 14: insulating layer, 15: interlayer insulating film, 21: barrier insulating film, 22: charge storage layer, 23: tunnel insulating film, 24: channel semiconductor layer, 25: core insulating film, 26: buried semiconductor layer, 27: contact plug-in, BLK: block, GP: memory bank, S: NAND string, MT: memory cell transistor, ST, selection transistor, SL: source line, GL: gate line, SGS: source side select line, WL: word line, BL: bit line, SGD: drain side select line

Detailed Description

Hereinafter, embodiments of the present invention will be described with reference to the drawings. In fig. 1 to 8, the same components are denoted by the same reference numerals, and redundant description is omitted.

(first embodiment)

Fig. 1 is a block diagram showing a configuration of a semiconductor memory device according to a first embodiment. The semiconductor memory device of this embodiment is, for example, a three-dimensional NAND memory.

The semiconductor memory device of the present embodiment includes a memory cell array 1, a sense amplifier 2, a plurality of row decoders 3, a drive circuit 4, a BL/SL driver 5, a voltage generation circuit 6, a detection circuit 7, and a control circuit 8. The drive circuit 4, the BL/SL driver 5, the voltage generation circuit 6, the detection circuit 7, and the like are examples of the erasing section.

The memory cell array 1 includes a plurality of (here, four) banks BLK0 to BLK3, and the banks BLK0 to BLK3 include a plurality of (here, four) memory groups GP0 to GP3, respectively. These memory groups GP0 to GP3 each include a plurality of NAND strings S. Each NAND string S includes a plurality of memory cell transistors (memory cells) connected in series with each other, and two selection transistors (selection gates) arranged so as to sandwich the memory cell transistors. Hereinafter, the blocks BLK0 to BLK3 are respectively denoted as "blocks BLK", and the memory groups GP0 to GP3 are respectively denoted as "memory groups GP".

The memory cell array 1 can store data in the memory cells of each NAND string S. The data stored in the memory cells are erased together for each block BLK. The number of blocks BLK in the memory cell array 1 is four in this embodiment, but may be other than four. Similarly, the number of memory groups GP in each block BLK is four in the present embodiment, but may be other than four.

The sense amplifier 2 detects and amplifies data read out from the memory cell array 1. In addition, when writing data into the memory cell array 1, the sense amplifier 2 transfers the write data to the memory cell array 1.

Each column decoder 3 corresponds to one block BLK. The semiconductor memory device of the present embodiment includes four column decoders 3 corresponding to four blocks BLK. Each row decoder 3 selects a row in the corresponding block BLK at the time of reading or writing data.

The drive circuit 4 supplies voltages necessary for reading, writing, and erasing data to the row decoders 3. This voltage is applied to the memory cells by the row decoders 3.

The BL/SL driver 5 applies voltages necessary for reading, writing, and erasing data to bit lines and/or source lines described later. The BL/SL driver 5 applies an erase voltage V to the bit lines and/or the source lines, for example, when erasing data stored in the memory cell array 1_ERA。

The voltage generation circuit 6 generates voltages necessary for reading, writing, and erasing of data, and supplies the voltages to the drive circuit 4 and the BL/SL driver 5. The voltage generation circuit 6 generates an erase voltage V at the time of erasing data, for example_ERA。

The detection circuit 7 monitors the erase voltage V generated from the voltage generation circuit 6 when data is erased_ERA. The detection circuit 7 is also based on the erase voltage V_ERAGenerates a flag signal and controls the drive circuit 4.

The control circuit 8 controls the operation of the semiconductor memory device of the present embodiment. The control circuit 8 controls the operation of the voltage generation circuit 6 to generate voltages necessary for reading, writing, and erasing data, for example, at the time of reading, writing, and erasing data.

Fig. 2 is a circuit diagram showing the configuration of the NAND string S of the first embodiment.

Fig. 2 shows one NAND string S in the memory cell array 1. The NAND string S shown in fig. 2 includes a plurality of (here, 48) memory cell transistors MT0 to MT47 connected in series with each other, and two select transistors ST1 and ST2 arranged so as to sandwich the memory cell transistors MT0 to MT 47. One main terminal of the selection transistor ST1 is electrically connected to the memory cell transistor MT47, and the other main terminal of the selection transistor ST1 is electrically connected to the bit line BL. One main terminal of the selection transistor ST2 is electrically connected to the memory cell transistor MT0, and the other main terminal of the selection transistor ST2 is electrically connected to the source line SL. Hereinafter, the memory cell transistors MT0 to MT47 are referred to as "memory cell transistors MT", and the selection transistors ST1 and ST2 are referred to as "selection transistors ST".

As shown in fig. 2, the memory cell array 1 further includes a plurality of (here, 48) word lines WL0 to WL47 electrically connected to the control terminals (gate terminals) of the memory cell transistors MT0 to MT47, a drain side select line SGD electrically connected to the control terminal of the select transistor ST1, and a source side select line SGS electrically connected to the control terminal of the select transistor ST 2. The memory cell array 1 further includes the bit lines BL and the source lines SL described above. The bit line BL and the source line SL are examples of the first wiring. The source side selection line SGS and the drain side selection line SGD are examples of the second wiring. Hereinafter, word lines WL0 to WL47 are respectively referred to as "word lines WL".

The word line WL is electrically connected to the corresponding memory cell transistor MT, and is provided to control the operation of the corresponding memory cell transistor MT. The drain side select line SGD is provided to control the operation of the select transistor ST 1. The source side selection line SGS is provided to control the operation of the selection transistor ST 2.

The bit line BL is electrically connected to the memory cell transistors MT0 to MT47 via the selection transistor ST 1. The source line SL is electrically connected to the memory cell transistors MT0 to MT47 via the selection transistor ST 2. The bit lines BL and the source lines SL are used for reading, writing, and erasing data for the memory cell transistors MT0 to MT47, for example.

FIG. 2 shows memory cell transistors MT0, MT1,. cndot.. MTm,. cndot.. MT23, MT24,. cndot.. MTn,. cndot.. MT46, MT 47. m is any integer satisfying 1 < m < 23, and n is any integer satisfying 24 < n < 46. FIG. 2 also shows word lines WL0, WL1, · · WLm, · · WL23, WL24, · · WLn, · · WL46, WL 47. m and n are used in the description of fig. 3 described later.

Fig. 3 is a graph for explaining the erasing operation of the first embodiment.

Fig. 3 shows various voltages used in the erasing operation of the semiconductor memory device of the present embodiment. Specifically, fig. 3 shows a voltage (erase voltage) V applied to the source line SL_ERAVoltage V applied to source side selection line SGS_{ERA_GIDL}And voltages VSWL0 to VSWL47 applied to word lines WL0 to WL 47. Erase voltage V_ERAIs an example of the first voltage. Voltage V_{ERA_GIDL}Is an example of the second voltage. Hereinafter, voltages VSWL0 to VSWL47 are also referred to as "voltages VSWL", respectively.

Erase voltage V_ERAThe voltage is generated by a voltage generation circuit 6 (see fig. 1. the same applies hereinafter) and applied to the source lines SL via a BL/SL driver 5. Voltage V_{ERA_GIDL}The voltage is generated by the voltage generation circuit 6 and applied to the source side selection line SGS via the drive circuit 4 and the row decoder 3. Voltages VSWL0 to VSWL47 are generated by the voltage generation circuit 6 and applied to the word lines WL0 to WL47 via the drive circuit 4 and the row decoder 3. As shown in FIG. 3, the values of these voltages are set so that V_ERA＞V_{ERA_GIDL}The relationship > VSWL0 ═ VSWL24 > VSWL1 ═ VSWL25 · · ═ VSWLm ═ VSWLn · · > VSWL22 ═ VSWL46 > VSWL23 ═ VSWL47 holds.

In this embodiment, data stored in the memory cell transistors MT are erased in a block by block BLK. Specifically, in the semiconductor memory device of the present embodiment, when data is collectively erased from all the memory cell transistors MT of a certain NAND string S, the erase voltage V is applied to the source line SL connected to the NAND string S_ERA。

At this time, the semiconductor memory device of the present embodiment applies the voltage V to the source side selection line SGS connected to the NAND string S_{ERA_GIDL}. As a result, GIDL (Gate Induced Drain Leakage) current is generated in the channel region of the selection transistor ST 2. The GIDL current flows from the source line SL side to the memory cell transistor MT side. The data stored in each memory cell transistor MT is erased by the GIDL current. In the erase operation, voltages VSWL0 to VSWL47 are applied to word lines WL0 to WL 47.

This embodimentIn the semiconductor memory device of the formula, at the time of the erase operation, the erase verify operation and the erase voltage applying operation are alternately repeated one or more times. In the erasing voltage applying operation, the erasing voltage V is applied to the source line SL_ERAApplying a voltage V to a source side selection line SGS_{ERA_GIDL}A voltage VSWL is applied to each word line WL. The number of cycles shown in fig. 3 indicates the number of times of performing the erase voltage applying operation. In this embodiment, the erase voltage V_ERAAnd a voltage V_{ERA_GIDL}Linearly increasing according to the number of cycles. On the other hand, the voltage VSWL is constant regardless of the number of cycles.

Fig. 4 is a sectional view showing the structure of the semiconductor memory device according to the first embodiment.

The semiconductor memory device of the present embodiment includes a substrate 11, an interlayer insulating film 12, an interlayer insulating film 13, a plurality of insulating layers 14, and a plurality of memory holes MH. Fig. 4 shows two of these memory holes MH. The semiconductor memory device of the present embodiment further includes a barrier insulating film 21, a charge storage layer 22, a tunnel insulating film 23, a channel semiconductor layer 24, and a core insulating film 25, which are sequentially formed in each memory hole MH. The channel semiconductor layer 24 is an example of a semiconductor layer.

The semiconductor memory device of the present embodiment further includes a source line SL, a gate line GL, a source side selection line SGS, and a plurality of word lines WL as a plurality of wiring layers formed on the substrate 11. The source line SL includes a metal layer SLa, a lower semiconductor layer SLb, an intermediate semiconductor layer SLc, and an upper semiconductor layer SLd. The source side selection line SGS includes three wiring layers SGSa, SGSb, and SGSc.

The substrate 11 is a semiconductor substrate such as a silicon substrate. Fig. 4 shows the X direction and the Y direction parallel to and perpendicular to the surface of the substrate 11, and the Z direction perpendicular to the surface of the substrate 11. In this specification, the + Z direction is treated as an upward direction, and the-Z direction is treated as a downward direction. the-Z direction may or may not coincide with the direction of gravity.

An interlayer insulating film 12 is formed on the substrate 11. The interlayer insulating film 12 is, for example, a silicon oxide film. The interlayer insulating film 12 may be formed directly on the substrate 11, or may be formed on the substrate 11 with another layer interposed therebetween.

The source line SL includes a metal layer SLa, a lower semiconductor layer SLb, an intermediate semiconductor layer SLc, and an upper semiconductor layer SLd, which are sequentially formed on the interlayer insulating film 12. The metal layer SLa is, for example, a tungsten layer. The lower semiconductor layer SLb, the intermediate semiconductor layer SLc, and the upper semiconductor layer SLd are, for example, polysilicon layers. The intermediate semiconductor layer SLc of this embodiment is in contact with the channel semiconductor layer 24 in each memory hole MH. Thereby, the source line SL is electrically connected to the channel semiconductor layer 24 in each memory hole MH.

The interlayer insulating film 13 is formed on the source line SL. The interlayer insulating film 13 is, for example, a silicon oxide film.

The gate line GL is formed on the interlayer insulating film 13. The gate line GL is, for example, a polysilicon layer.

As described above, the source side selection line SGS includes the three wiring layers SGSa, SGSb, and SGSc. These wiring layers SGSa to SGSc are stacked above the gate line GL so as to be separated from each other. A plurality of insulating layers 14 are provided between these wiring layers SGSa to SGSc. Similarly, the word lines WL are stacked above the source side selection line SGS so as to be separated from each other. A plurality of insulating layers 14 are provided between the word lines WL. Each insulating layer 14 includes, for example, a silicon oxide film. Each of the wiring layers SGSa to SGSc includes, for example, a tungsten layer. Each word line WL includes, for example, a tungsten layer.

Each memory hole MH penetrates the insulating layers 14, the word lines WL, the wiring layers SGSa to SGSc of the source side selection line SGS, the gate line GL, and the interlayer insulating film 13, and a part of each memory hole MH is formed in the source line SL. The barrier insulating film 21, the charge storage layer 22, the tunnel insulating film 23, the channel semiconductor layer 24, and the core insulating film 25 are formed in this order in each memory hole MH. The barrier insulating film 21 is, for example, a silicon oxide film. The charge storage layer 22 is, for example, a silicon nitride film. The charge storage layer 22 may be a polysilicon layer. The tunnel insulating film 23 is, for example, a silicon oxide film. The channel semiconductor layer 24 is, for example, a polysilicon layer. The channel semiconductor layer 24 of the present embodiment is electrically connected to the source line SL, and the gate line GL, the source side selection line SGS, and each word line WL are electrically insulated from each other. The core insulating film 25 is, for example, a silicon oxide film.

The block insulating film 21, the charge storage layer 22, the tunnel insulating film 23, the channel semiconductor layer 24, and the core insulating film 25 in each memory hole MH constitute a NAND string S (see fig. 2) together with the plurality of word lines WL. Fig. 4 shows one memory cell transistor MT included in the NAND string S by a dotted line.

As shown in fig. 4, the channel semiconductor layer 24 of the present embodiment includes an n-type diffusion layer on the side of the source line SL and the gate line GL. The n-type diffusion layer contains n-type impurities at a high concentration. As shown in fig. 4, the channel semiconductor layer 24 of the present embodiment further includes n on the side of the wiring layer SGSa in the source side selection line SGS^－And a type diffusion layer. N is^－The n-type diffusion layer is located above the n-type diffusion layer, and contains an n-type impurity so that the concentration of the n-type impurity changes rapidly in the Z direction. In the present embodiment, n can be represented by^－The action of the type diffusion layer generates a GIDL current.

The semiconductor memory device of the present embodiment applies an erase voltage V to the source line SL_ERAApplying a voltage V to a source side selection line SGS_{ERA_GIDL}Thereby generating GIDL current in the channel region of the selection transistor ST 2. N in FIG. 4^－The type diffusion layer is provided in the channel region of the selection transistor ST 2. The data stored in the memory cell transistor MT shown in fig. 4 is erased by the GIDL current. In fig. 4, the source line SL is an example of a first wiring, and the source side selection line SGS is an example of a second wiring.

In this case, if n^－The concentration distribution of the n-type impurity in the type diffusion layer greatly differs for each memory hole MH, and the amount of GIDL current generated greatly differs for each memory hole MH. This may reduce the reliability of the erase operation in the present embodiment. Therefore, the semiconductor memory device of the present embodiment adopts the erasing operation described with reference to fig. 6 to 8.

Fig. 5 is a cross-sectional view showing the structure of a semiconductor memory device according to a modification of the first embodiment.

The semiconductor memory device according to the present modification includes, in addition to the components shown in fig. 4, a drain-side select line SGD, an interlayer insulating film 15, a buried semiconductor layer 26 and a contact plug 27 provided for each memory hole MH, and a bit line BL. The drain-side select line SGD includes four wiring layers SGDa, SGDb, SGDc, SGDd.

These wiring layers SGDa to SGDd are stacked above the word line WL, as in the case of the word line WL, so as to be separated from each other. A plurality of insulating layers 14 are provided between these wiring layers SGDa to SGDd. Each of the wiring layers SGDa to SGDd includes, for example, a tungsten layer. The drain-side select line SGD of the present modification is electrically insulated from the channel semiconductor layer 24 in each memory hole MH.

An interlayer insulating film 15 is formed above the drain-side select line SGD. The interlayer insulating film 15 is, for example, a silicon oxide film. In the present modification, each memory hole MH penetrates the insulating layers 14, the word lines WL, and the wiring layers SGDa to SGDd of the drain-side selection line SGD, and a part of each memory hole MH is formed in the interlayer insulating film 15.

In each memory hole MH, an embedded semiconductor layer 26 is formed on the core insulating film 25 and is in contact with the channel semiconductor layer 24. On each memory hole MH, a contact plug 27 is formed on the embedded semiconductor layer 26 and contacts the embedded semiconductor layer 26. As a result, the two contact plugs 27 shown in fig. 5 are electrically connected to the channel semiconductor layer 24 in the corresponding memory hole MH.

The bit line BL is formed on these contact plugs 27 in the interlayer insulating film 15. As a result, the bit line BL of the present modification is electrically connected to the channel semiconductor layer 24 in each memory hole MH.

As shown in fig. 5, the channel semiconductor layer 24 of the present modification includes an n-type diffusion layer on the side of the interlayer insulating film 15. The n-type diffusion layer contains n-type impurities at a high concentration. As shown in fig. 5, the channel semiconductor layer 24 of the present modification further includes n on the side of the wiring layer SGDd in the drain-side selection line SGD^－And a type diffusion layer. N is^－The n-type diffusion layer is located below the n-type diffusion layer, and contains an n-type impurity so that the concentration of the n-type impurity changes sharply in the-Z direction. In the present modification, n can be used^－The action of the type diffusion layer generates a GIDL current.

Semiconductor memory device of this modificationThe memory device is formed by applying an erase voltage V to a bit line BL_ERAA voltage V is applied to a drain side selection line SGD_{ERA_GIDL}Thereby generating GIDL current in the channel region of the selection transistor ST 1. N in FIG. 5^－The type diffusion layer is provided in the channel region of the selection transistor ST 1. The data stored in the memory cell transistor MT shown in fig. 5 is erased by the GIDL current. In fig. 5, a bit line BL is an example of a first wiring, and a drain-side select line SGD is an example of a second wiring.

In this way, the semiconductor memory device of the present modification applies the erase voltage V to the bit line BL instead of the source line SL_ERAA voltage V is applied to the drain side selection line SGD instead of the source side selection line SGS_{ERA_GIDL}. The erasing operation described with reference to fig. 3 and 4 can also be realized by the method of the present modification.

The problem described with reference to fig. 4 may also occur in the present modification. I.e. if n^－The concentration distribution of the n-type impurity in the type diffusion layer greatly differs for each memory hole MH, and the amount of GIDL current generated greatly differs for each memory hole MH. This may reduce the reliability of the erasing operation in the present modification. Therefore, the semiconductor memory device according to the present modification employs the erase operation described with reference to fig. 6 to 8.

The erasing operation of the first embodiment will be described in further detail below with reference to fig. 6 to 8. Note that the contents described with reference to fig. 6 to 8 can also be applied to the modification of fig. 5 by replacing the source line SL with the bit line BL and replacing the source side selection line SGS with the drain side selection line SGD.

Fig. 6 is a graph for explaining the erasing operation of the first embodiment.

A curve C1 shows a temporal change in the source line voltage during the erase operation of the semiconductor memory device according to the comparative example of the present embodiment. A curve C1' shows a temporal change in the source line voltage in the erasing operation of the semiconductor memory device according to the present embodiment. The semiconductor memory device of the present embodiment has substantially the same configuration as the semiconductor memory device of the comparative example, but performs different erasing operations as shown by curves C1 and C1'.

The source line voltage in the erase operation is an erase voltage V applied to the source line SL in the erase operation_ERA. In the following description, not only the "source line SL" but also the "erase voltage V" are used in the description of the semiconductor memory device of the present embodiment_ERA", and other words are also used in the description of the semiconductor memory device of the comparative example. For details of these terms, please refer to the description part of fig. 1 to 5.

As shown by the curve C1', the erase voltage V of the present embodiment_ERAFrom zero to a value of V1, from a value of V1 to a value of V2, and thereafter maintained at a value of V2(0 < V2 < V1). The value V1 is the erase voltage V_ERAIs the erase voltage V, the value V2_ERAA stable value of (c). Reference numeral t1 denotes an erase voltage V of the present embodiment_ERAThe time until the value V1 is reached, and reference numeral t2 denotes the erasing voltage V of the present embodiment_ERATime to converge to a value V2. Thus, the erase voltage V of the present embodiment_ERATemporarily reaches a maximum value V1 higher than the steady value V2, and thereafter, falls from the maximum value V1 to a steady value V2. The value V1 is an example of a first value and the value V2 is an example of a second value. In addition, the erase voltage V shown in FIG. 3 is described above_ERAMore specifically, the steady value V2.

On the other hand, as shown by the curve C1, the erasing voltage V of the comparative example_ERARises from zero to a value of V2 and then remains at a value of V2. Erase Voltage V of comparative example_ERAErase voltage V of the present embodiment_ERAIs indicated by arrow P1. Erase voltage V of the present embodiment_ERAWaveform of (2) was passed through the erase voltage V of the comparative example_ERAThe waveform of (2) is realized by applying a kick component that brings about a voltage rise as shown by an arrow P1.

As described with reference to fig. 4 and 5, if n is n^－The concentration distribution of the n-type impurity in the type diffusion layer greatly differs for each memory hole MH, and the amount of GIDL current generated greatly differs for each memory hole MH. This may reduce the reliability of the erase operation in the present embodiment. Therefore, the erasing operation in the present embodimentUsing an erase voltage V that varies as curve C1_ERA. Thus, even n^－The concentration distribution of the n-type impurity in the type diffusion layer is greatly different for each memory hole MH, and the difference in erasing operation between different memory holes MH can be reduced. This is explained in detail with reference to fig. 7.

Fig. 7 is another graph for explaining the erasing operation of the first embodiment.

Fig. 7 shows curves C2, C3, C3 'in addition to the curve C1' described above. A curve C2 shows a temporal change in channel voltage (channel potential) during the erase operation of the semiconductor memory device according to the comparative example. The curve C3 also shows the temporal change in the channel voltage during the erase operation of the semiconductor memory device of the comparative example. However, a curve C2 shows a temporal change in the channel voltage in the memory hole MH where the amount of GIDL current generated is large, and a curve C3 shows a temporal change in the channel voltage in the memory hole MH where the amount of GIDL current generated is small.

Further, a curve C3' shows a temporal change in channel voltage during an erasing operation of the semiconductor memory device of the present embodiment. Note that, the curve C3' shows the temporal change in the channel voltage in the memory hole MH where the GIDL current generation amount is small, similarly to the curve C3. The channel voltage in the erase operation is a voltage applied to the channel semiconductor layer 24 in the erase operation.

Here, the present embodiment is compared with a comparative example.

The channel voltage of the comparative example rises to a high steady value in the memory cell MH where the amount of GIDL current generated is large (curve C2), but rises only to a low steady value in the memory cell MH where the amount of GIDL current generated is small (curve C3). Reference numeral D denotes a difference between the steady value of the curve C2 and the steady value of the curve C3. As described above, in the erasing operation of the comparative example, the stable value of the channel voltage varies among the memory holes MH. In the comparative example, n is the same as^－Since the concentration distribution of the n-type impurity in the type diffusion layer greatly differs for each memory hole MH, the amount of GIDL current generated greatly differs for each memory hole MH. Thus, it is possible to provideThe above-described variation in channel voltage (see difference D) occurs.

On the other hand, the channel voltage GIDL of the present embodiment changes as the curve C2 in the memory cell MH in which the amount of generated current is large, as in the comparative example, but changes as the curve C3' in the memory cell MH in which the amount of generated GIDL current is small, as in the comparative example. That is, the channel voltage GIDL according to the present embodiment rises to a high stable value not only in the memory cell MH where the amount of generated current is large (curve C2), but also in the memory cell MH where the amount of generated GIDL current is small (curve C3'). Reference numeral D 'denotes a difference between the stable value of the curve C2 and the stable value of the curve C3'. The difference D 'is smaller than the above-mentioned difference D (D' < D). As described above, according to this embodiment, it is possible to suppress variations in the stable value of the channel voltage between memory holes MH during the erase operation.

The curve C3 of the comparative example is different from the curve C3' of the present embodiment by an arrow P2. In this embodiment, the erase voltage V_ERATemporarily rising to a maximum value V1 higher than the steady value V2. Erase voltage V_ERAThe temporary rise of (3) has an effect of assisting the rise of the channel voltage in the memory hole MH where the generation amount of the GIDL current is small. The arrow P2 shows the case of changing to the curve C3' by this action curve C3. Thus, in the present embodiment, variations in the stable value of the channel voltage are suppressed.

As described with reference to fig. 4 and 5, if n is n^－The concentration distribution of the n-type impurity in the type diffusion layer greatly differs for each memory hole MH, and the amount of GIDL current generated greatly differs for each memory hole MH. This may reduce the reliability of the erase operation in the present embodiment. Therefore, in the erasing operation of the present embodiment, the erasing voltage V which changes as shown by the curve C1' is used_ERA. This can help increase the channel voltage in memory hole MH with a small amount of generated GIDL current, and suppress variation in the stable value of the channel voltage. Thus, according to the present embodiment, even n^－The concentration distribution of n-type impurity in the diffusion layer is greatly different for each memory hole MH, and can be reducedThe difference in erase action between different memory holes MH. This can improve the reliability of the erasing operation of the present embodiment.

Fig. 8 is another graph for explaining the erasing operation of the first embodiment.

A curve C4 shows the relationship between the threshold voltage Vth and the Bit count (Bit count) of the memory cell transistor MT in the above comparative example. A curve C4' shows the relationship between the threshold voltage Vth and the bit count of the memory cell transistor MT in the present embodiment. According to this embodiment, as shown in fig. 8, the distribution of the threshold voltages Vth can be improved.

Hereinafter, referring again to fig. 6, the erasing operation of the present embodiment will be described in further detail.

As described above, the erase voltage V of the present embodiment_ERAFrom zero to a value V1, from a value V1 to a value V2, and then maintained at a value V2 (curve C1'). The value V1 may be set to any value as long as the ratio V2 is large. However, since the value V1 is too large, the waste of the erase operation becomes large, the value V1 is preferably set to 2 times or less the value V2 (V1. ltoreq. 2 XV 2), for example, 1.1 to 1.3 times the value V2 (1.1 XV 2. ltoreq. V1. ltoreq. 1.3 XV 2). In addition, an erase voltage V_ERAThe value of (b) can be measured by, for example, measuring the potential of a wiring which penetrates the insulating layer 14 and the word line WL and is electrically connected to the source line SL, similarly to the channel semiconductor layer 24.

In addition, the erase voltage V of the present embodiment_ERARises to a value V2 before time t1, rises to a value V1 at time t1, and falls to a value V2 after time t1 (specifically, time t 2). From the erasing voltage V_ERAThe time from the rise to the value V1 to the fall to the value V1 may be adjusted to any value, but if it is too long, the erase operation is wasted. Therefore, the time is preferably adjusted to 1 μ second to 100 μ second, for example, to several μ seconds to several tens of μ seconds.

As described above, the erase voltage V of the present embodiment_ERAFor example by the voltage generating circuit 6. The voltage generation circuit 6 includes, for example, a pump circuit capable of switching an output voltage between a value V1 and a value V2. In this case, the present embodimentThe semiconductor memory device can switch the output voltage from the voltage generation circuit 6 between the value V1 and the value V2, thereby applying the erase voltage V as shown by the curve C1_ERAIs applied to the source lines SL.

As described above, the semiconductor memory device of the present embodiment applies the erase voltage V to the source line SL_ERASo that the erase voltage V is_ERARises to a value of V1, falls from a value of V1 to a value of V2, and remains at a value of V2. This can improve the reliability of the erase operation of the memory cell transistor MT.

Although several embodiments have been described above, these embodiments are presented as examples and are not intended to limit the scope of the invention. The novel apparatus and methods described herein can be implemented in other various ways. Various omissions, substitutions, and changes in the form of the devices and methods described herein may be made without departing from the spirit of the invention. The appended claims and their equivalents are intended to include such forms or modifications as are included in the scope or spirit of the invention.