TWI601152B - Semiconductor memory device - Google Patents

Semiconductor memory device Download PDF

Info

Publication number
TWI601152B
TWI601152B TW104143444A TW104143444A TWI601152B TW I601152 B TWI601152 B TW I601152B TW 104143444 A TW104143444 A TW 104143444A TW 104143444 A TW104143444 A TW 104143444A TW I601152 B TWI601152 B TW I601152B
Authority
TW
Taiwan
Prior art keywords
memory cell
data
word line
level
page
Prior art date
Application number
TW104143444A
Other languages
Chinese (zh)
Other versions
TW201627998A (en
Inventor
Yoshikazu Harada
Original Assignee
Toshiba Memory Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Toshiba Memory Corp filed Critical Toshiba Memory Corp
Priority to TW104143444A priority Critical patent/TWI601152B/en
Publication of TW201627998A publication Critical patent/TW201627998A/en
Application granted granted Critical
Publication of TWI601152B publication Critical patent/TWI601152B/en

Links

Landscapes

  • Read Only Memory (AREA)

Description

半導體記憶裝置 Semiconductor memory device

本發明之實施形態係關於一種半導體記憶裝置。 Embodiments of the present invention relate to a semiconductor memory device.

NAND(Not AND,反及)型快閃記憶體等半導體記憶裝置已廣為人知。 Semiconductor memory devices such as NAND (Not AND) type flash memory are well known.

本發明之實施形態提供一種可準確地檢測資料之半導體記憶裝置。 Embodiments of the present invention provide a semiconductor memory device that can accurately detect data.

本實施形態之半導體記憶裝置包含具有複數個記憶胞之記憶胞陣列。複數條字元線係連接於複數個記憶胞。複數條位元線係連接於上述複數個記憶胞之電流路徑之一端。感測放大器部於檢測連接於字元線中之字元線WLn(n為整數)之記憶胞之資料時,重複進行複數次檢測動作。控制部根據連接於字元線WLn-1之記憶胞之資料及連接於字元線WLn+1之記憶胞之資料,而選擇由複數次檢測動作所取得之複數個檢測結果中之任一個。 The semiconductor memory device of this embodiment includes a memory cell array having a plurality of memory cells. A plurality of character line lines are connected to a plurality of memory cells. A plurality of bit lines are connected to one of the current paths of the plurality of memory cells. The sense amplifier section repeats the plurality of detection operations when detecting the data of the memory cell connected to the word line WLn (n is an integer) in the word line. The control unit selects any one of a plurality of detection results obtained by the plurality of detection operations based on the data of the memory cell connected to the word line WLn-1 and the data of the memory cell connected to the word line WLn+1.

1‧‧‧記憶胞陣列 1‧‧‧ memory cell array

2a‧‧‧列解碼器/字元線驅動器 2a‧‧‧ Column Decoder/Word Line Driver

2b‧‧‧行解碼器 2b‧‧‧ line decoder

3‧‧‧頁面緩衝器 3‧‧‧Page Buffer

4‧‧‧NAND單位單元 4‧‧‧NAND unit unit

5a‧‧‧列位址暫存器 5a‧‧‧ column address register

5b‧‧‧行位址暫存器 5b‧‧‧ row address register

6‧‧‧邏輯控制器 6‧‧‧Logic Controller

7‧‧‧序列控制電路 7‧‧‧Sequence control circuit

8‧‧‧內部電壓產生電路 8‧‧‧Internal voltage generation circuit

9‧‧‧I/O緩衝器 9‧‧‧I/O buffer

10‧‧‧NAND晶片 10‧‧‧NAND chip

11‧‧‧控制器 11‧‧‧ Controller

ALE‧‧‧位址閂鎖啟動信號 ALE‧‧‧ address latch activation signal

Bce‧‧‧晶片驅動信號 Bce‧‧‧ chip drive signal

bRE‧‧‧讀取啟動信號 bRE‧‧‧Read start signal

bWE‧‧‧寫入啟動信號 bWE‧‧‧Write start signal

BL‧‧‧位元線 BL‧‧‧ bit line

BL0‧‧‧位元線 BL0‧‧‧ bit line

BL1‧‧‧位元線 BL1‧‧‧ bit line

BLi-1‧‧‧位元線 BLi-1‧‧‧ bit line

BLi-2‧‧‧位元線 BLi-2‧‧‧ bit line

BLkm-1‧‧‧單元區塊 BLkm-1‧‧‧Unit Block

BLK1‧‧‧單元區塊 BLK1‧‧‧ unit block

BLK0‧‧‧單元區塊 BLK0‧‧‧ unit block

CELSRC‧‧‧共用源極線 CELSRC‧‧‧Shared source line

CLE‧‧‧指令啟動信號 CLE‧‧‧ command start signal

E1‧‧‧位準 E1‧‧‧

E2‧‧‧位準 E2‧‧‧

E3‧‧‧位準 E3‧‧‧

E4‧‧‧位準 E4‧‧‧

LM‧‧‧位準 LM‧‧‧

LM1‧‧‧位準 LM1‧‧‧

LM2‧‧‧位準 LM2‧‧‧

LM3‧‧‧位準 LM3‧‧‧

MC0‧‧‧記憶胞 MC0‧‧‧ memory cell

MC1‧‧‧記憶胞 MC1‧‧‧ memory cell

MC2‧‧‧記憶胞 MC2‧‧‧ memory cell

MC63‧‧‧記憶胞 MC63‧‧‧ memory cell

MCn-1‧‧‧記憶胞 MCn-1‧‧‧ memory cell

MCn+1‧‧‧記憶胞 MCn+1‧‧‧ memory cell

S1‧‧‧電晶體 S1‧‧‧O crystal

S2‧‧‧電晶體 S2‧‧‧O crystal

SA‧‧‧感測放大器電路 SA‧‧‧Sense Amplifier Circuit

SGD‧‧‧選擇閘極線 SGD‧‧‧Selected gate line

SGS‧‧‧選擇閘極線 SGS‧‧‧Selected gate line

VA‧‧‧電壓位準 VA‧‧‧voltage level

VB‧‧‧電壓位準 VB‧‧‧ voltage level

VC‧‧‧電壓位準 VC‧‧‧ voltage level

VLM‧‧‧位準 VLM‧‧‧

VLM1‧‧‧位準 VLM1‧‧‧

VLM2‧‧‧位準 VLM2‧‧‧

VLM3‧‧‧位準 VLM3‧‧‧

VREAD‧‧‧高位準電壓 VREAD‧‧‧ high level voltage

WL0‧‧‧字元線 WL0‧‧‧ character line

WL1‧‧‧字元線 WL1‧‧‧ character line

WL2‧‧‧字元線 WL2‧‧‧ character line

WL62‧‧‧字元線 WL62‧‧‧ character line

WL63‧‧‧字元線 WL63‧‧‧ character line

圖1係表示第1實施形態之NAND型快閃記憶體系統之構成的方塊圖。 Fig. 1 is a block diagram showing the configuration of a NAND flash memory system according to the first embodiment.

圖2係表示單元陣列1之具體構成之圖。 Fig. 2 is a view showing a concrete configuration of the cell array 1.

圖3係表示執行2位元之資料之寫入時之記憶胞MC之閾值分佈的圖。 Fig. 3 is a view showing a threshold distribution of memory cells MC when writing of 2-bit data is performed.

圖4係表示記憶區塊內之頁面寫入順序之圖。 Fig. 4 is a view showing a page writing sequence in a memory block.

圖5(A)~圖5(D)係表示連接於字元線WLn之複數個記憶胞MCn之下級頁面受到之鄰接干擾效應的圖。 5(A) to 5(D) are diagrams showing the adjacent interference effects of the lower-order pages of the plurality of memory cells MCn connected to the word line WLn.

圖6(A)~圖6(C)係表示第1實施形態之記憶體之IDL動作之時序圖。 6(A) to 6(C) are timing charts showing the IDL operation of the memory of the first embodiment.

圖7(A)~圖7(C)係表示第1實施形態之變化例1之記憶體之IDL動作的時序圖。 7(A) to 7(C) are timing charts showing the IDL operation of the memory of the first modification of the first embodiment.

圖8(A)~圖8(C)係表示第1實施形態之變化例2之記憶體之IDL動作的時序圖。 8(A) to 8(C) are timing charts showing the IDL operation of the memory of the second modification of the first embodiment.

圖9(A)~圖9(C)係表示第2實施形態之記憶體之IDL動作之時序圖。 9(A) to 9(C) are timing charts showing the IDL operation of the memory of the second embodiment.

圖10(A)~圖10(C)係表示第3實施形態之記憶體之IDL動作之時序圖。 10(A) to 10(C) are timing charts showing the IDL operation of the memory of the third embodiment.

圖11(A)~圖11(C)係表示第4實施形態之記憶體之IDL動作之時序圖。 11(A) to 11(C) are timing charts showing the IDL operation of the memory of the fourth embodiment.

圖12(A)係表示自記憶胞MCn之三個檢測結果選擇之選擇結果之表。 Fig. 12(A) is a table showing selection results of selection of three detection results from the memory cell MCn.

圖12(B)係表示自記憶胞MCn之兩個檢測結果選擇之選擇結果之表。 Fig. 12(B) is a table showing selection results of selection of two detection results from the memory cell MCn.

以下,參照圖式說明本發明之實施形態。本實施形態並不限定本發明。於說明時,針對所有圖式,對共用之部分附加共用之參照符號。但,應注意的是圖式為示意圖,厚度與平面尺寸之關係、各層之厚度比率等與實物有所不同。 Hereinafter, embodiments of the present invention will be described with reference to the drawings. This embodiment does not limit the present invention. For the sake of explanation, the common reference numerals are added to the shared portions for all the drawings. However, it should be noted that the drawing is a schematic diagram, the relationship between the thickness and the plane size, the thickness ratio of each layer, and the like are different from the actual ones.

因此,具體之厚度或尺寸應參酌以下說明而加以判斷。又,當然,圖式相互間亦包含彼此之尺寸關係或比例等不同之部分。 Therefore, the specific thickness or size should be judged by considering the following instructions. Further, of course, the drawings also include mutually different dimensional relationships or ratios.

(第1實施形態) (First embodiment)

圖1係表示第1實施形態之NAND型快閃記憶體之構成之一例的方塊圖。本實施形態之NAND型快閃記憶體包含NAND晶片10、及控制NAND晶片10之控制器11。NAND晶片10及控制器11可作為多晶片封裝體(MCP,Multi-chip Package)而被樹脂密封於一個封裝體。 Fig. 1 is a block diagram showing an example of a configuration of a NAND flash memory according to the first embodiment. The NAND flash memory of the present embodiment includes a NAND wafer 10 and a controller 11 that controls the NAND wafer 10. The NAND wafer 10 and the controller 11 can be resin-sealed to one package as a multi-chip package (MCP).

構成NAND晶片10之記憶胞陣列1係以將複數個記憶胞MC二維排列成矩陣狀而構成。該記憶胞MC具備電荷蓄積層。該記憶胞MC並非限定於FG(Floating Gate,浮動閘極)型之記憶胞,例如亦可為MONOS(Metal Oxide Nitride Oxide Semiconductor,金屬-氧化物-氮化物-氧化物半導體)型之記憶胞。列解碼器/字元線驅動器2a、行解碼器2b、頁面緩衝器3及內部電壓產生電路8構成以頁為單位對記憶胞陣列1進行資料之寫入及讀取之資料寫入/讀取電路。列解碼器/字元線驅動器2a係選擇性地驅動記憶胞陣列1之字元線。頁面緩衝器3具備一頁量之感測放大器電路與資料保持電路,且以頁為單位對記憶胞陣列1進行資料之讀取及寫入。 The memory cell array 1 constituting the NAND wafer 10 is configured by two-dimensionally arranging a plurality of memory cells MC in a matrix. This memory cell MC has a charge accumulation layer. The memory cell MC is not limited to a FG (Floating Gate) type memory cell, and may be, for example, a MONOS (Metal Oxide Nitride Oxide Semiconductor) type memory cell. The column decoder/word line driver 2a, the row decoder 2b, the page buffer 3, and the internal voltage generating circuit 8 constitute data writing/reading of data writing and reading to the memory cell array 1 in units of pages. Circuit. The column decoder/word line driver 2a selectively drives the word lines of the memory cell array 1. The page buffer 3 has a one-page sense amplifier circuit and a data hold circuit, and reads and writes data to the memory cell array 1 in units of pages.

頁面緩衝器3之一頁量之讀取資料由行解碼器2b依序進行行選擇,且經由I/O(Input/Out,輸入/輸出)緩衝器9而輸出至外部I/O端子。自I/O端子供給之寫入資料由行解碼器2b選擇並載入至頁面緩衝器3。對頁面緩衝器3載入一頁量之寫入資料。列位址信號及行位址信號係經由I/O緩衝器9而輸入,並分別傳送至列解碼器2a及行解碼器2b。列位址暫存器5a於刪除動作中保存刪除區塊位址,於寫入或讀取動作中保存頁位址。於開始寫入動作之前,對行位址暫存器5b輸入用於載入寫入資料之標頭行位址、或用於寫入動作之標頭行位址。於以特定之條件切換寫入啟動信號bWE或讀取啟動信號bRE之前,行位址暫存器5b將保存所輸入之行位址。 The read data of one page of the page buffer 3 is sequentially selected by the row decoder 2b, and output to the external I/O terminal via the I/O (Input/Out) buffer 9. The write data supplied from the I/O terminal is selected by the row decoder 2b and loaded into the page buffer 3. A page amount of write data is loaded to the page buffer 3. The column address signal and the row address signal are input via the I/O buffer 9, and are transmitted to the column decoder 2a and the row decoder 2b, respectively. The column address register 5a saves the deleted block address in the delete operation, and saves the page address in the write or read operation. Before the start of the write operation, the row address register 5b is input with a header row address for loading the write data or a header row address for the write operation. The row address register 5b will save the input row address before switching the write enable signal bWE or the read enable signal bRE under certain conditions.

邏輯控制電路6基於晶片驅動信號bCE、指令啟動信號CLE、位 址閂鎖啟動信號ALE、寫入啟動信號bWE、讀取啟動信號bRE等控制信號,控制指令或位址之輸入,及控制資料之輸入輸出。讀取動作或寫入動作係按照指令執行。序列控制電路7接收指令而進行讀取、寫入或刪除之序列控制。內部電壓產生電路8接收外部電源電壓VCC,被控制電路7控制而產生各種動作所需之特定電壓。下述之感測放大器用之內部電源電壓VDC係於內部電壓產生電路8中產生。 The logic control circuit 6 is based on the wafer drive signal bCE, the command enable signal CLE, the bit Address latch enable signal ALE, write enable signal bWE, read enable signal bRE and other control signals, control commands or address input, and control data input and output. The read action or the write action is performed in accordance with the instruction. The sequence control circuit 7 receives an instruction to perform sequence control of reading, writing or deleting. The internal voltage generating circuit 8 receives the external power supply voltage VCC and is controlled by the control circuit 7 to generate a specific voltage required for various operations. The internal power supply voltage VDC for the sense amplifier described below is generated in the internal voltage generating circuit 8.

控制器11係以適於NAND晶片10之當前之寫入狀態之條件執行資料之寫入及讀取控制。再者,當然亦可於NAND晶片10側進行讀取控制之一部分。 The controller 11 performs writing and reading control of data in a condition suitable for the current write state of the NAND wafer 10. Furthermore, it is of course also possible to perform a part of the read control on the NAND wafer 10 side.

圖2係表示記憶胞陣列1之具體構成之一例的圖。於該例中,由串列連接之64個記憶胞MC0~MC63與連接於其兩端之選擇閘極電晶體S1及S2構成NAND單位單元(NAND串)4。選擇閘極電晶體S1之源極連接於共用源極線CELSRC,選擇閘極電晶體S2之汲極連接於位元線BL(BL0~BLi-1)。即,位元線BL連接於記憶胞MC之電流路徑之一端。記憶胞MC0~MC63之控制閘極分別連接於字元線WL(WL0~WL63),選擇閘極電晶體S1、S2之閘極連接於選擇閘極線SGS、SGD。 FIG. 2 is a view showing an example of a specific configuration of the memory cell array 1. In this example, 64 memory cells MC0 to MC63 connected in series and selected gate transistors S1 and S2 connected to both ends thereof constitute a NAND unit cell (NAND string) 4. The source of the gate transistor S1 is connected to the common source line CELSRC, and the gate of the gate transistor S2 is connected to the bit line BL (BL0~BLi-1). That is, the bit line BL is connected to one end of the current path of the memory cell MC. The control gates of the memory cells MC0~MC63 are respectively connected to the word line WL (WL0~WL63), and the gates of the gate transistors S1 and S2 are connected to the selection gate lines SGS and SGD.

沿著一條字元線之複數個記憶胞之範圍成為作為統一之資料讀取及資料寫入之單位之頁面。又,沿著字元線方向排列之複數個NAND單位單元之範圍構成成為資料統一刪除之單位之單元區塊BLK。圖2中,於位元線BL方向上排列共用位元線BL之複數個單元區塊BLK0~BLKm-1,而構成記憶胞陣列1。字元線WL及選擇閘極線SGS、SGD被列解碼器2a驅動。各位元線BL連接於頁面緩衝器3之感測放大器電路SA。感測放大器電路SA檢測藉由位元線BL及資源線WL選擇之記憶胞MC之資料。 The range of a plurality of memory cells along a word line becomes a page for a unified unit of data reading and data writing. Further, the range of the plurality of NAND unit cells arranged along the direction of the word line constitutes a unit block BLK which is a unit for collectively deleting data. In FIG. 2, a plurality of unit blocks BLK0 to BLKm-1 sharing the bit line BL are arranged in the direction of the bit line BL to constitute the memory cell array 1. The word line WL and the selection gate lines SGS, SGD are driven by the column decoder 2a. The bit lines BL are connected to the sense amplifier circuit SA of the page buffer 3. The sense amplifier circuit SA detects the data of the memory cell MC selected by the bit line BL and the resource line WL.

圖3係表示執行2位元之資料寫入時之記憶胞MC之閾值分佈的 圖。於本實施形態中,於一個記憶胞MC記憶2位元之資料。區塊內之所有記憶胞MC之閾值電壓Vt因區塊刪除而成為最低之「E(刪除)」位準。其後,於寫入下級頁面時,針對下級頁面資料「0」之記憶胞進行使閾值電壓上升至「LM」位準之寫入。「E」位準及「LM」位準於其後受到進行寫入之鄰接之記憶胞之影響而產生變動,閾值分佈幅度擴展。於寫入下一個上級頁面時,藉由根據上級頁面資料使閾值分佈進一步變動,而產生分別與資料「11」、「01」、「00」、「10」對應之四個較窄之閾值分佈「E」、「A」、「B」、「C」。該情形時,最低之刪除E位準以E位準直接使用。第二低之A位準藉由自E位準位移而產生。B、C位準藉由自LM位準位移而產生。 Figure 3 is a diagram showing the threshold distribution of the memory cell MC when the data of 2 bits is written. Figure. In the present embodiment, the data of two bits is memorized in one memory cell MC. The threshold voltage Vt of all memory cells MC in the block becomes the lowest "E (deletion)" level due to block deletion. Thereafter, when the lower-level page is written, the writing of the threshold voltage to the "LM" level is performed for the memory cell of the lower-level page data "0". The "E" level and the "LM" level are subject to change due to the influence of the adjacent memory cells being written, and the threshold distribution amplitude is expanded. When the next upper page is written, the threshold distribution is further changed according to the upper page data, and four narrow threshold distributions corresponding to the data "11", "01", "00", and "10" are generated. "E", "A", "B", "C". In this case, the lowest deletion E level is used directly at the E level. The second lowest A level is produced by displacement from the E level. The B and C levels are generated by displacement from the LM level.

圖4係表示記憶體區塊內之頁面寫入順序之圖。將以「圓圈」表示之下級頁面之寫入與以「四角」表示之上級頁面分別示為不同之頁面。交替寫入互不相同之字元線WL之下級頁面與上級頁面。下級頁面之「圓圈」及上級頁面之「四角」內所記載之數字係表示寫入順序。再者,字元線WL之數目及位元線BL之數目並無特別限定。 Figure 4 is a diagram showing the page writing sequence in a memory block. The "circle" is used to indicate that the lower level page is written as the "four corners" to indicate that the upper level page is different from the upper level page. The sub-pages and the upper-level pages of the character line WL which are different from each other are alternately written. The "circles" of the lower-level page and the numbers in the "four corners" of the parent page indicate the writing order. Furthermore, the number of word lines WL and the number of bit lines BL are not particularly limited.

例如,於對字元線WL0之下級頁面寫入資料後,對字元線WL1之下級頁面寫入資料。接著,對字元線WL0之上級頁面寫入資料。然後,對字元線WL2之下級頁面寫入資料,對字元線WL1之上級頁面寫入資料。其後,反覆進行字元線WLn(n為3以上之整數)之下級頁面之寫入與字元線WLn-1之上級頁面之寫入。 For example, after the data is written to the lower level page of the word line WL0, the data is written to the lower level page of the word line WL1. Next, data is written to the upper page of the word line WL0. Then, data is written to the lower level page of the word line WL2, and data is written to the upper level page of the word line WL1. Thereafter, the writing of the lower-level page of the word line WLn (n is an integer of 3 or more) and the writing of the upper-level page of the word line WLn-1 are repeatedly performed.

如此,資料被交替地寫入至互不相同之字元線WL之下級頁面與上級頁面。藉此,連接於字元線WLn之記憶胞之上級頁面之寫入係於連接於與其前後鄰接之字元線WLn+1之記憶胞之下級頁面及連接於字元線WLn-1之記憶胞之上級頁面之寫入之後執行。藉此,可於某種程度上抑制記憶胞之閾值分佈因鄰接干擾效應而擴展。 Thus, the data is alternately written to the lower level page and the upper level page of the character line WL which are different from each other. Thereby, the writing of the upper cell page of the memory cell connected to the word line WLn is connected to the memory cell lower level page connected to the word line WLn+1 adjacent thereto and the memory cell connected to the word line WLn-1. Executed after the writing of the upper level page. Thereby, the threshold distribution of the memory cell can be suppressed to some extent due to the adjacent interference effect.

另一方面,多值記憶體(MLC(Multi-Level Cell:多位階記憶體 胞)之情形時,於即將寫入某一字元線WL之上級頁面之前,檢測該字元線WL之下級頁面。該動作被稱作IDL(Internal Data Load:內部數據加載)。於藉由IDL參照下級頁面之狀態(E或LM)後,可寫入上級頁面之資料。但,記憶胞MC之下級頁面有時會受到與該記憶胞MC之兩側鄰接之記憶胞之上級頁面之寫入或下級頁面之寫入之影響。 On the other hand, multi-level memory (MLC (Multi-Level Cell) In the case of a cell, the page below the word line WL is detected before the upper page of the word line WL is to be written. This action is called IDL (Internal Data Load). After referring to the status (E or LM) of the lower page by IDL, the data of the upper page can be written. However, the lower-level pages of the memory cell MC are sometimes affected by the writing of the upper-level pages of the memory cells adjacent to the two sides of the memory cell MC or the writing of the lower-level pages.

例如,著眼於圖4所示之連接於字元線WL1之記憶胞之下級頁面。於連接於字元線WL1之記憶胞之下級頁面之寫入(第2個動作)與上級頁面之寫入(第5個動作)之間,執行連接於字元線WL0之記憶胞之上級頁面之寫入(第3個動作)及連接於字元線WL2之記憶胞之下級頁面之寫入(第4個動作)。因此,於即將寫入連接於字元線WL1之記憶胞之上級頁面之前,連接於字元線WL1之記憶胞之下級頁面會受到連接於字元線WL0之記憶胞之上級頁面之寫入及連接於字元線WL2之記憶胞之下級頁面之寫入之影響。 For example, attention is paid to the memory cell sub-level page connected to the word line WL1 shown in FIG. Between the writing of the lower-level page of the memory cell connected to the word line WL1 (the second action) and the writing of the upper-level page (the fifth action), the memory cell upper-level page connected to the word line WL0 is executed. The writing (the third action) and the writing of the lower-level page of the memory cell connected to the word line WL2 (the fourth action). Therefore, before the upper stage page of the memory cell connected to the word line WL1 is written, the lower level page of the memory cell connected to the word line WL1 is written by the upper page of the memory cell connected to the word line WL0 and The effect of writing to the lower level page of the memory cell connected to word line WL2.

如此之情形時,連接於字元線WL1之記憶胞之下級頁面之閾值分佈擴展。如圖3所示,下級頁面包含E位準及LM位準之資料。因此,若E位準及LM位準之閾值分佈拓展,則感測放大器電路SA無法於IDL中準確地檢測E位準與LM位準。 In this case, the threshold distribution of the lower-level pages of the memory cells connected to the word line WL1 is expanded. As shown in Figure 3, the lower level page contains information on the E level and the LM level. Therefore, if the threshold distribution of the E level and the LM level is expanded, the sense amplifier circuit SA cannot accurately detect the E level and the LM level in the IDL.

圖5(A)~圖5(D)係表示連接於字元線WLn之複數個記憶胞MCn之下級頁面受到之鄰接干擾效應的圖。於寫入前,記憶胞MCn係如圖5(A)所示般為刪除狀態(E1)。即,尚未對記憶胞MCn之下級頁面寫入資料。於該階段,對連接於與字元線WLn鄰接之字元線WLn-1之記憶胞MCn-1之下級頁面寫入資料。因此,記憶胞MCn之閾值分佈受到因記憶胞MCn-1之下級頁面之寫入導致之鄰接干擾效應,而自E1之狀態變化為E2之狀態。然而,於該階段,因記憶胞MCn之下級頁面尚未寫入,故記憶胞MCn之閾值分佈自E1變化為E2不構成問題。 5(A) to 5(D) are diagrams showing the adjacent interference effects of the lower-order pages of the plurality of memory cells MCn connected to the word line WLn. Before writing, the memory cell MCn is in the deleted state (E1) as shown in Fig. 5(A). That is, the data has not been written to the lower level page of the memory cell MCn. At this stage, data is written to the lower level page of the memory cell MCn-1 connected to the word line WLn-1 adjacent to the word line WLn. Therefore, the threshold distribution of the memory cell MCn is affected by the adjacent interference effect caused by the writing of the lower cell page of the memory cell MCn-1, and the state changes from the state of E1 to the state of E2. However, at this stage, since the lower level page of the memory cell MCn has not been written, the change of the threshold distribution of the memory cell MCn from E1 to E2 does not pose a problem.

接著,如圖5(B)所示,寫入記憶胞MCn之下級頁面。藉此,記憶 胞MCn之閾值分佈被劃分為E2位準之記憶胞與LM1位準之記憶胞。於該階段,記憶胞MCn之下級頁面(E2及LM1)並未受到鄰接干擾效應。 Next, as shown in FIG. 5(B), the lower level page of the memory cell MCn is written. With this, memory The threshold distribution of the cell MCn is divided into the memory cell of the E2 level and the memory cell of the LM1 level. At this stage, the sub-pages (E2 and LM1) of the memory cell MCn are not affected by the adjacent interference.

接著,寫入連接於字元線WLn-1之記憶胞MCn-1之上級頁面。如參照圖3所說明般,上級頁面之寫入包含自E位準朝E或A位準之寫入、及自LM位準朝B或C位準之寫入。該等寫入動作中,自E位準朝E位準之寫入及自LM位準朝B位準之寫入時,記憶胞之閾值電壓幾乎未產生變化。另一方面,自E位準朝A位準之寫入、及自LM位準朝C位準之寫入時,記憶胞之閾值電壓大幅度變化。因此,記憶胞MCn係自複數個記憶胞MCn-1中之自E寫入至A之鄰接記憶胞、及自LM寫入至C之鄰接記憶胞受到鄰接干擾效應。因此,如圖5(C)所示,記憶胞MCn之下級頁面之閾值分佈係自E2變化為E3、及自LM1變化為LM2。 Next, the upper page of the memory cell MCn-1 connected to the word line WLn-1 is written. As illustrated with reference to FIG. 3, the writing of the upper page includes writing from the E level to the E or A level, and writing from the LM level to the B or C level. In these write operations, the threshold voltage of the memory cell hardly changes when writing from the E level to the E level and writing from the LM level to the B level. On the other hand, when writing from the E level to the A level and writing from the LM level to the C level, the threshold voltage of the memory cell largely changes. Therefore, the memory cell MCn is subjected to the adjacent interference effect from the adjacent memory cells in which E is written to A in the plurality of memory cells MCn-1, and the adjacent memory cells written from C to C. Therefore, as shown in FIG. 5(C), the threshold distribution of the lower page of the memory cell MCn changes from E2 to E3, and from LM1 to LM2.

接著,寫入連接於字元線WLn+1之記憶胞MCn+1之下級頁面。下級頁面之寫入係自E位準朝E或LM位準之寫入。該等寫入動作中,自E位準朝E位準之寫入時,記憶胞之閾值電壓幾乎未產生變化。另一方面,自E位準朝LM位準之寫入時,記憶胞之閾值電壓大幅度變化。因此,記憶胞MCn係自複數個記憶胞MCn+1中之自E位準寫入至LM位準之鄰接記憶胞受到鄰接干擾效應。因此,如圖5(D)所示,記憶胞MCn之下級頁面之閾值分佈係自E3變化為E4、及自LM2變化為LM3。如此,若記憶胞MCn自鄰接記憶胞MCn-1、MCn+1受到鄰接干擾效應,則記憶胞MCn之下級頁面之閾值分佈如圖5(D)般擴展。 Next, the lower level page of the memory cell MCn+1 connected to the word line WLn+1 is written. The writing of the lower level page is written from the E level to the E or LM level. In the write operation, the threshold voltage of the memory cell hardly changes when writing from the E level to the E level. On the other hand, when writing from the E level to the LM level, the threshold voltage of the memory cell changes greatly. Therefore, the memory cell MCn is subjected to the adjacent interference effect from the contiguous memory cells of the plurality of memory cells MCn+1 written from the E level to the LM level. Therefore, as shown in FIG. 5(D), the threshold distribution of the lower page of the memory cell MCn changes from E3 to E4, and from LM2 to LM3. Thus, if the memory cell MCn is subjected to the adjacent interference effect from the adjacent memory cells MCn-1, MCn+1, the threshold distribution of the lower level page of the memory cell MCn is expanded as shown in FIG. 5(D).

接著,對記憶胞MCn執行IDL。於該階段,如圖5(D)所示,記憶胞MCn之閾值分佈已大幅度擴展。閾值電壓存在區域Rvt之記憶胞MC根據自鄰接記憶胞受到之鄰接緩衝效應之程度,可成為具有E位準之閾值電壓之記憶胞或具有LM位準之閾值電壓之記憶胞。 Next, IDL is performed on the memory cell MCn. At this stage, as shown in FIG. 5(D), the threshold distribution of the memory cell MCn has been greatly expanded. The memory cell MC of the threshold voltage existence region Rvt can be a memory cell having a threshold voltage of the E level or a memory cell having a threshold voltage of the LM level according to the degree of the adjacent buffer effect received by the adjacent memory cell.

如此,即使於下級資料之閾值分佈已擴展之狀態下執行IDL,亦有感測放大器電路SA無法準確地檢測下級資料之情形。IDL中之下級 資料之錯誤與通常之資料讀取動作中之錯誤不同,不執行使用ECC(Error Correcting Code:錯誤修正碼)之修正。因此,IDL中之下級資料之錯誤成為硬位元錯誤,而有難以進行資料修正之情形。該情形時,導致讀取錯誤。 Thus, even if the IDL is executed in a state where the threshold distribution of the lower-level data has been expanded, there is a case where the sense amplifier circuit SA cannot accurately detect the lower-level data. Under the IDL The error of the data is different from the error in the normal data reading operation, and the correction using the ECC (Error Correcting Code) is not performed. Therefore, the error of the subordinate data in the IDL becomes a hard bit error, and there is a case where it is difficult to perform data correction. In this case, a read error is caused.

然而,本發明者判定,記憶胞MCn之下級資料之位移亦可藉由記憶胞MCn-1之上級頁面之寫入(自E朝A之寫入、自LM朝C之寫入)、及記憶胞MCn+1之下級頁面之寫入(自E朝LM之寫入)而產生。因此,於記憶胞MCn之IDL時,可檢測記憶胞MCn-1及記憶胞MCn+1中之任一者或兩者之資料,使用記憶胞MCn-1及/或記憶胞MCn+1之檢測資料,而修正字元線WLn之下級頁面之檢測資料。 However, the inventors have determined that the displacement of the subordinate data of the memory cell MCn can also be written by the upper page of the memory cell MCn-1 (writing from E to A, writing from LM to C), and memory. The write of the lower level page of cell MCn+1 (generated from E to LM). Therefore, when the IDL of the memory cell MCn is used, the data of either or both of the memory cell MCn-1 and the memory cell MCn+1 can be detected, and the memory cell MCn-1 and/or the memory cell MCn+1 are detected. The data is corrected, and the detection data of the lower level page of the character line WLn is corrected.

例如,於IDL中,藉由將字元線WLn之電壓設定為VLM1,可使感測放大器SA識別E2位準之記憶胞與LM1~LM3位準之記憶胞。藉由將字元線WLn之電壓設定為VLM2,可使感測放大器SA識別E3位準之記憶胞與LM2、LM3位準之記憶胞。進而,藉由將字元線WLn之電壓設定為VLM3,可使感測放大器SA識別E4位準之記憶胞與LM3位準之記憶胞。 For example, in IDL, by setting the voltage of the word line WLn to VLM1, the sense amplifier SA can identify the memory cell of the E2 level and the memory cell of the LM1~LM3 level. By setting the voltage of the word line WLn to VLM2, the sense amplifier SA can identify the memory cell of the E3 level and the memory cell of the LM2, LM3 level. Further, by setting the voltage of the word line WLn to VLM3, the sense amplifier SA can identify the memory cell of the E4 level and the memory cell of the LM3 level.

因此,於本實施形態之記憶體中,感測放大器電路SA於寫入動作之IDL中,於檢測連接於字元線WLn之記憶胞MCn之資料時,重複進行複數次檢測動作,閂鎖複數個檢測結果。而且,作為控制部之序列控制電路7(參照圖1)係根據連接於與字元線WLn之兩側鄰接之字元線WLn-1、WLn+1之記憶胞MCn-1、MCn+1之資料,選擇該複數個檢測結果中之任一個。 Therefore, in the memory of the present embodiment, the sense amplifier circuit SA repeats the plurality of detection operations when the data of the memory cell MCn connected to the word line WLn is detected in the IDL of the write operation, and the latch is plural. Test results. Further, the sequence control circuit 7 (see Fig. 1) as the control unit is based on the memory cells MCn-1, MCn+1 connected to the word lines WLn-1, WLn+1 adjacent to both sides of the word line WLn. Data, select any one of the plurality of test results.

接著,參照圖6(A)~圖6(C),更具體地說明IDL時之記憶胞MCn之資料檢測動作。 Next, the data detecting operation of the memory cell MCn at the time of IDL will be described more specifically with reference to FIGS. 6(A) to 6(C).

圖6(A)~圖6(C)係表示第1實施形態之記憶體之IDL動作之時序圖。圖6(A)~圖6(C)表示即將寫入字元線WLn之上級頁面之前所進行 之IDL動作的時序圖。 6(A) to 6(C) are timing charts showing the IDL operation of the memory of the first embodiment. 6(A) to 6(C) show that it is performed immediately before the upper page of the word line WLn is written. Timing diagram of the IDL action.

(連接於WLn-1之記憶胞MCn-1之資料檢測) (Detection of data connected to WLn-1 memory cell MCn-1)

於時點t0~t1,檢測連接於字元線WLn-1之複數個記憶胞MCn-1之資料。於即將寫入記憶胞MCn之上級頁面之前,記憶胞MCn-1之上級頁面亦被寫入。因此,記憶胞MCn-1之閾值電壓可獲得E、A、B及C中之任一位準。為識別E、A、B及C位準,一面將字元線WLn-1之電壓位準變更為VA、VB及VC一面執行記憶胞MCn-1之資料檢測。如圖3所示,VA係處於E位準與A位準之間之電壓位準,VB係處於A位準與B位準之間之電壓位準,VC係處於B位準與C位準之間之電壓位準。 At time t0~t1, the data of the plurality of memory cells MCn-1 connected to the word line WLn-1 is detected. The memory cell MCn-1 upper level page is also written before the upper stage page of the memory cell MCn is to be written. Therefore, the threshold voltage of the memory cell MCn-1 can obtain any of E, A, B, and C. In order to identify the E, A, B, and C levels, the data level of the memory cell MCn-1 is performed while changing the voltage level of the word line WLn-1 to VA, VB, and VC. As shown in Figure 3, the VA is at the voltage level between the E level and the A level, the VB is at the voltage level between the A level and the B level, and the VC is at the B level and the C level. The voltage level between them.

例如,於字元線WLn-1之電壓為VA之情形時,E位準之記憶胞MCn-1成為導通狀態,但其以外之A、B及C位準之記憶胞MCn-1維持斷開狀態。因此,於字元線WLn-1之電壓為VA之情形時,可對複數個記憶胞MCn-1中之E位準之記憶胞與A、B及C位準之記憶胞加以區分。同樣地,於字元線WLn-1之電壓為VB之情形時,可對A、B及C位準之記憶胞MCn-1中之A位準之記憶胞與B及C位準之記憶胞加以區分。進而,於字元線WLn-1之電壓為VC之情形時,可對B及C位準之記憶胞MCn-1中之B位準之記憶胞與C位準之記憶胞加以區分。 For example, when the voltage of the word line WLn-1 is VA, the memory cell MCn-1 of the E level is turned on, but the memory cell MCn-1 of the other A, B, and C levels remains disconnected. status. Therefore, when the voltage of the word line WLn-1 is VA, the memory cells of the E level in the plurality of memory cells MCn-1 can be distinguished from the memory cells of the A, B, and C levels. Similarly, when the voltage of the word line WLn-1 is VB, the memory cells of the A level in the memory cell MCn-1 of the A, B, and C levels and the memory cells of the B and C levels can be used. Make a distinction. Further, when the voltage of the word line WLn-1 is VC, the memory cells of the B level in the memory cell MCn-1 of the B and C levels can be distinguished from the memory cells of the C level.

藉此,檢測各記憶胞MCn-1之閾值電壓為E、A、B及C中之何種位準。即,感測放大器SA可檢測複數個記憶胞MCn-1之各自之2位元資料。記憶胞MCn-1之資料被保存(閂鎖)於感測放大器SA或頁面緩衝器PB。 Thereby, it is detected which threshold level of each of the memory cells MCn-1 is E, A, B, and C. That is, the sense amplifier SA can detect the respective 2-bit data of the plurality of memory cells MCn-1. The data of the memory cell MCn-1 is saved (latched) to the sense amplifier SA or the page buffer PB.

再者,如上所述,於鄰接記憶胞MCn-1為A、C位準之情形,及鄰接之記憶胞MCn+1為LM位準之情形時,記憶胞MCn受到鄰接干擾效應。因此,只要可判斷鄰接記憶胞MCn-1及MCn+1為A、C位準 或E、B位準中之何種位準,則判定記憶胞MCn是否受到鄰接干擾效應。因此,為降低鄰接緩衝效應,序列控制電路7亦可將其壓縮為表示為A、C位準或E、B位準中之何種位準之1位元資料。藉此,可減少感測放大器SA或頁面緩衝器PB應保存之資料容量。 Furthermore, as described above, when the adjacent memory cell MCn-1 is at the A and C levels, and the adjacent memory cell MCn+1 is the LM level, the memory cell MCn is subjected to the adjacent interference effect. Therefore, as long as the adjacent memory cells MCn-1 and MCn+1 can be judged to be A and C levels Or which of the E and B levels is used to determine whether the memory cell MCn is subject to the adjacent interference effect. Therefore, in order to reduce the adjacent buffering effect, the sequence control circuit 7 can also compress it into a 1-bit data indicating which of the A, C, or E, B levels. Thereby, the data capacity that the sense amplifier SA or the page buffer PB should hold can be reduced.

再者,此時,因其他字元線WL0~WLn-2及WLn~WL63為非選擇狀態,故維持為高位準電壓VREAD。藉此,以低電阻將記憶胞MCn-1連接於位元線BL與單元源極CELSRC之間。 Furthermore, at this time, since the other word lines WL0 to WLn-2 and WLn to WL63 are in the non-selected state, the high level voltage VREAD is maintained. Thereby, the memory cell MCn-1 is connected between the bit line BL and the cell source CELSRC with a low resistance.

(連接於WLn+1之記憶胞MCn+1之資料檢測) (Detection of data connected to WLn+1 memory cell MCn+1)

於時點t1~t2,檢測連接於字元線WLn+1之複數個記憶胞MCn+1之資料。於即將寫入記憶胞MCn之上級頁面之前,記憶胞MCn+1之下級頁面亦被寫入,記憶胞MCn+1之上級頁面尚未被寫入。因此,記憶胞MCn+1之閾值電壓可獲得E及LM中之任一位準。為識別E及LM之閾值位準,記憶胞MCn+1之資料檢測係將字元線WLn+1之電壓位準設為VLM而執行。如圖3所示,VLM係處於E位準與LM位準之間之電壓位準,且為E位準與LM位準之間之讀取電壓。再者,圖5之VLM1~3為VLM之一例。 At time points t1 to t2, data of a plurality of memory cells MCn+1 connected to the word line WLn+1 are detected. Before the upper stage page of the memory cell MCn is written, the memory cell MCn+1 lower level page is also written, and the memory cell MCn+1 upper level page has not been written. Therefore, the threshold voltage of the memory cell MCn+1 can obtain any of E and LM. To identify the threshold levels of E and LM, the data detection of the memory cell MCn+1 is performed by setting the voltage level of the word line WLn+1 to VLM. As shown in FIG. 3, the VLM is at a voltage level between the E level and the LM level, and is a read voltage between the E level and the LM level. Furthermore, VLM1~3 of FIG. 5 is an example of VLM.

例如,於字元線WLn+1之電壓為VLM之情形時,E位準之記憶胞MCn+1成為導通狀態,但其以外之LM位準之記憶胞MCn+1維持斷開狀態。因此,於字元線WLn+1之電壓為VLM之情形時,對記憶胞MCn+1中之E位準之記憶胞與LM位準之記憶胞加以區分。 For example, when the voltage of the word line WLn+1 is VLM, the memory cell MCn+1 of the E level is turned on, but the memory cell MCn+1 of the other LM level remains in the off state. Therefore, when the voltage of the word line WLn+1 is VLM, the memory cell of the E level in the memory cell MCn+1 is distinguished from the memory cell of the LM level.

藉此,檢測各記憶胞MCn+1之閾值電壓為E及LM中之何種位準。即,感測放大器SA可檢測複數個記憶胞MCn+1之各自之下級頁面之資料。記憶胞MCn+1之資料亦被保存(閂鎖)於感測放大器SA或頁面緩衝器PB。 Thereby, it is detected which threshold voltage of each memory cell MCn+1 is E and LM. That is, the sense amplifier SA can detect the data of the respective lower level pages of the plurality of memory cells MCn+1. The data of the memory cell MCn+1 is also saved (latched) to the sense amplifier SA or the page buffer PB.

再者,此時,因其他字元線WL0~WLn及WLn+2~WL63為非選擇狀態,故維持為高位準電壓VREAD。藉此,以低電阻將記憶胞 MCn+1連接於位元線BL與單元源極CELSRC之間。 Furthermore, at this time, since the other word lines WL0 to WLn and WLn+2 to WL63 are in the non-selected state, they are maintained at the high level voltage VREAD. Thereby, the memory cell is low in resistance MCn+1 is connected between the bit line BL and the cell source CELSRC.

(連接於WLn之記憶胞MCn之資料檢測(IDL)) (Data detection (IDL) of memory cell MCn connected to WLn)

於時點t2~t3,檢測連接於字元線WLn之複數個記憶胞MCn之資料。此時,由於係即將寫入記憶胞MCn之上級頁面之前,故連記憶胞MCn之下級頁面亦被寫入。因此,記憶胞MCn之閾值電壓可獲得E及LM中之任一位準。 At time t2 to t3, the data of the plurality of memory cells MCn connected to the word line WLn is detected. At this time, since the system is about to be written before the upper stage page of the memory cell MCn, the lower level page of the memory cell MCn is also written. Therefore, the threshold voltage of the memory cell MCn can obtain any of E and LM.

然而,如上所述,記憶胞MCn之下級頁面因記憶胞MCn-1之上級頁面之寫入(自E朝A之寫入(以下,亦稱作「WRITE(E-A)」)、自LM朝C之寫入(以下,亦稱作「WRITE(LM-C)」)、及記憶胞MCn+1之下級頁面之寫入(自E朝LM之寫入(以下,亦稱作「WRITE(E-LM)」)而受到鄰接干擾效應。因此,如圖5(D)所示,推斷出記憶胞MCn之下級頁面之閾值分佈擴展。 However, as described above, the lower level page of the memory cell MCn is written by the upper page of the memory cell MCn-1 (writing from E to A (hereinafter, also referred to as "WRITE (EA)"), from LM to C Write (hereinafter, also referred to as "WRITE (LM-C)"), and write of the lower-level page of the memory cell MCn+1 (write from E to LM (hereinafter, also referred to as "WRITE (E-) LM)") is subject to the adjacent interference effect. Therefore, as shown in Fig. 5(D), the threshold distribution spread of the lower level page of the memory cell MCn is estimated.

本實施形態之記憶體為檢測記憶胞MCn之下級資料而重複進行複數次檢測動作。此時,字元線WLn之電壓維持為固定電壓VLM。例如,於時點t2-1,感測放大器SA檢測記憶胞MCn之資料(第1檢測)。接著,於時點t2-2,感測放大器SA再次檢測記憶胞MCn之資料(第2檢測)。接著,於時點t2-3,感測放大器SA再次檢測記憶胞MCn之資料(第3檢測)。第1至第3檢測之結果被保存於感測放大器SA或頁面緩衝器PB。 The memory of the present embodiment repeats the plurality of detection operations for detecting the lower level data of the memory cell MCn. At this time, the voltage of the word line WLn is maintained at the fixed voltage VLM. For example, at time t2-1, the sense amplifier SA detects the data of the memory cell MCn (first detection). Next, at time t2-2, the sense amplifier SA detects the data of the memory cell MCn again (second detection). Next, at time t2-3, the sense amplifier SA detects the data of the memory cell MCn again (third detection). The results of the first to third detections are stored in the sense amplifier SA or the page buffer PB.

於NAND型快閃記憶體中,感測放大器SA藉由自感測節點經由位元線BL向記憶胞MC流通電流(電荷),而檢測記憶胞MC之資料。記憶胞MCn中流通與資料之邏輯及字元線WLn之固定電壓VLM相應之固定之單位電流。因此,感測放大器SA藉由檢測感測節點之電壓變化而可識別記憶胞MC之資料之邏輯。 In the NAND type flash memory, the sense amplifier SA detects the data of the memory cell MC by flowing a current (charge) from the sensing node to the memory cell MC via the bit line BL. The fixed unit current corresponding to the logic of the memory cell MCn and the fixed voltage VLM of the word line WLn. Therefore, the sense amplifier SA can recognize the logic of the data of the memory cell MC by detecting the voltage change of the sensing node.

如本實施形態般,藉由以不同時序(t2-1~t2-3)檢測記憶胞MCn之資料,感測節點之電壓於第2檢測時t2-2較第1檢測時t2-1變低,於 第3檢測時t2-3較第2檢測時t2-2變低。例如,於t2~t2-1期間,自感測節點釋放電荷Id1。於t2~t2-2期間,自感測節點釋放電荷Id2(>Id1)。於t2~t2-3期間,自感測節點釋放電荷Id3(>Id2)。因此,感測節點之電壓於第1至第3檢測時不同。該情形實質上等價於變更字元線WLn之電壓而檢測記憶胞MCn之資料者。例如,可使第1檢測與使用VLM1之檢測對應,使第2檢測與使用VLM2之檢測對應,使第3檢測與使用VLM3之檢測對應。即,本實施形態係藉由變更資料檢測時序以取代變更字元線WLn之設定電壓,而執行考慮到鄰接干涉效應之IDL。 As in the present embodiment, by detecting the data of the memory cell MCn at different timings (t2-1 to t2-3), the voltage of the sensing node becomes lower at the second detection time t2-2 than at the first detection time t2-1. ,to At the third detection, t2-3 is lower than t2-2 at the second detection. For example, during t2~t2-1, the self-sensing node releases the charge Id1. During t2~t2-2, the self-sensing node releases the charge Id2 (>Id1). During t2~t2-3, the self-sensing node releases the charge Id3 (>Id2). Therefore, the voltage of the sensing node is different at the first to third detections. This case is substantially equivalent to changing the voltage of the word line WLn to detect the data of the memory cell MCn. For example, the first detection may be associated with the detection using VLM1, the second detection may be associated with the detection using VLM2, and the third detection may be associated with the detection using VLM3. That is, in the present embodiment, the IDL in consideration of the adjacent interference effect is performed by changing the data detection timing instead of changing the set voltage of the word line WLn.

如上所述,記憶胞MCn之下級頁面受到之鄰接干擾效應係因記憶胞MCn-1之上級資料之寫入(WRITE(E-A)、WRITE(LM-C))、及記憶胞MCn+1之下級頁面之寫入(WRITE(E-LM)而產生。該等鄰接干擾效應在某種程度上可預測,如參照圖5(D)所說明般,藉由將字元線WLn之設定電壓變更為VLM1~VLM3,可考慮到鄰接干擾效應而識別E位準之記憶胞與LM位準之記憶胞。同樣地,藉由以與VLM1~VLM3對應之方式變更感測放大器SA之檢測時序(t2-1~t2-3),可考慮到鄰接干擾效應而識別E位準之記憶胞與LM位準之記憶胞。 As described above, the adjacent interference effect of the lower level page of the memory cell MCn is due to the writing of the upper level data of the memory cell MCn-1 (WRITE (EA), WRITE (LM-C)), and the lower level of the memory cell MCn+1. The page is written (WRITE (E-LM). These adjacent interference effects are predictable to some extent, as explained with reference to FIG. 5(D), by changing the set voltage of the word line WLn to VLM1~VLM3 can recognize the memory cell of the E level and the memory cell of the LM level in consideration of the adjacent interference effect. Similarly, the detection timing of the sense amplifier SA is changed in a manner corresponding to VLM1 to VLM3 (t2- 1~t2-3), the memory cell of the E level and the memory cell of the LM level can be identified by considering the adjacent interference effect.

序列控制電路7根據鄰接記憶胞MCn-1、MCn+1之資料而選擇保存於感測放大器SA或頁面緩衝器PB之第1及第2檢測結果中之任一個。 The sequence control circuit 7 selects one of the first and second detection results stored in the sense amplifier SA or the page buffer PB based on the data of the adjacent memory cells MCn-1 and MCn+1.

例如,於某一位元線BL(行)中,記憶胞MCn-1儲存A(或C)位準之資料,且記憶胞MCn+1儲存LM位準之資料之情形時,記憶胞MCn自記憶胞MCn-1、MCn+1之兩者受到鄰接干擾效應,可判定為屬於圖5(D)之E4或LM3之閾值分佈。因此,序列控制電路7選擇並採用與VLM3對應之第1檢測結果。 For example, in a bit line BL (row), when the memory cell MCn-1 stores the data of the A (or C) level, and the memory cell MCn+1 stores the data of the LM level, the memory cell MCn is self- Both of the memory cells MCn-1 and MCn+1 are subjected to the adjacent interference effect, and can be determined to belong to the threshold distribution of E4 or LM3 of FIG. 5(D). Therefore, the sequence control circuit 7 selects and adopts the first detection result corresponding to the VLM 3.

例如,於另一位元線BL(行)中,記憶胞MCn-1儲存A(或C)位準 之資料,且記憶胞MCn+1儲存E位準之資料之情形時,記憶胞MCn自記憶胞MCn-1受到鄰接干擾效應,可判定為屬於圖5(D)之E3或LM2之閾值分佈。因此,序列控制電路7選擇並採用與VLM2對應之第2檢測結果。 For example, in another bit line BL (row), the memory cell MCn-1 stores the A (or C) level. When the memory cell MCn+1 stores the E-level data, the memory cell MCn is subjected to the adjacent interference effect from the memory cell MCn-1, and can be determined as belonging to the threshold distribution of E3 or LM2 of FIG. 5(D). Therefore, the sequence control circuit 7 selects and adopts the second detection result corresponding to the VLM 2.

例如,於又一位元線BL(行)中,記憶胞MCn-1儲存E(或B)位準之資料,且記憶胞MCn+1儲存LM位準之資料之情形時,記憶胞MCn自記憶胞MCn+1受到鄰接干擾效應,可判定為屬於圖5(D)之E3或LM2之閾值分佈。因此,序列控制電路7選擇並採用與VLM2對應之第2檢測結果。 For example, in another bit line BL (row), when the memory cell MCn-1 stores the E (or B) level data, and the memory cell MCn+1 stores the LM level data, the memory cell MCn is self-contained. The memory cell MCn+1 is subjected to the adjacent interference effect and can be determined to belong to the threshold distribution of E3 or LM2 of FIG. 5(D). Therefore, the sequence control circuit 7 selects and adopts the second detection result corresponding to the VLM 2.

例如,於進而又一位元線BL(行)中,記憶胞MCn-1儲存E(或B)位準之資料,且記憶胞MCn+1儲存E位準之資料之情形時,記憶胞MCn未自記憶胞MCn-1、MCn+1受到鄰接干擾效應,可判定為屬於圖5(D)之E2或LM1之閾值分佈。因此,序列控制電路7選擇並採用與VLM1對應之第3檢測結果。 For example, in a further line BL (row), when the memory cell MCn-1 stores the E (or B) level data, and the memory cell MCn+1 stores the E level data, the memory cell MCn The memory cells MCn-1 and MCn+1 are not subjected to the adjacent interference effect, and can be determined as belonging to the threshold distribution of E2 or LM1 of FIG. 5(D). Therefore, the sequence control circuit 7 selects and adopts the third detection result corresponding to the VLM 1.

於NAND快閃記憶體中,為降低位元成本而廣泛使用在一個記憶胞記憶2位元以上之資料之多值技術。又,為縮小晶片尺寸,記憶胞之微細化不斷進展。於如此之狀況下,可忽略鄰接單元間干擾所造成之閾值分佈之擴展(鄰接干擾效應)。鄰接干涉效應係已寫入資料之記憶胞之閾值電壓因對鄰接記憶胞之資料寫入動作而產生位移之現象。藉由鄰接干擾效應,記憶胞之閾值電壓分佈擴大,而有感測放大器中檢測到之資料之可靠性降低之虞。 In NAND flash memory, a multi-value technique that uses data in a memory cell memory of more than two bits is widely used to reduce bit cost. Moreover, in order to reduce the size of the wafer, the miniaturization of the memory cells is progressing. Under such conditions, the extension of the threshold distribution caused by inter-cell interference (adjacent interference effects) can be ignored. The adjacent interference effect is a phenomenon in which the threshold voltage of the memory cell in which the data has been written is displaced due to the writing operation to the data of the adjacent memory cell. By the adjacent interference effect, the threshold voltage distribution of the memory cell is expanded, and the reliability of the data detected in the sense amplifier is reduced.

與此相對,根據本實施形態,於IDL時,感測放大器SA檢測連接於與作為IDL之對象之字元線WLn之兩側鄰接之字元線WLn-1及WLn+1之記憶胞之資料,進而以不同時序重複檢測複數次連接於字元線WLn之記憶胞MCn之資料。而且,序列控制電路7根據鄰接記憶胞MCn-1、MCn+1之資料,而選擇記憶胞MCn之複數個檢測結果中之 任一個。序列控制電路7針對各行執行該選擇動作。藉此,序列控制電路7可採用考慮到來自鄰接記憶胞MCn-1、MCn+1之鄰接干擾效應之檢測結果。藉此,本實施形態之記憶體可於IDL中準確地檢測資料。 On the other hand, according to the present embodiment, in the IDL, the sense amplifier SA detects the data of the memory cell connected to the word lines WLn-1 and WLn+1 adjacent to both sides of the word line WLn which is the target of the IDL. Further, the data of the memory cell MCn connected to the word line WLn is repeatedly detected at different timings. Moreover, the sequence control circuit 7 selects a plurality of detection results of the memory cell MCn based on the data of the adjacent memory cells MCn-1, MCn+1. Any one. The sequence control circuit 7 performs this selection action for each line. Thereby, the sequence control circuit 7 can take the detection result in consideration of the adjacent interference effect from the adjacent memory cells MCn-1, MCn+1. Thereby, the memory of this embodiment can accurately detect data in the IDL.

於本實施形態中,感測放大器SA檢測所有鄰接記憶胞MCn-1及鄰接記憶胞MCn+1之資料。因此,序列控制電路7可於考慮到記憶胞MCn之下級頁面受到之鄰接干擾效應(WRITE(E-A)、WRITE(LM-C)及WRITE(E-LM)之鄰接干擾效應)之基礎上,選擇記憶胞MCn之檢測結果。 In the present embodiment, the sense amplifier SA detects data of all adjacent memory cells MCn-1 and adjacent memory cells MCn+1. Therefore, the sequence control circuit 7 can select based on the adjacent interference effects (the adjacent interference effects of WRITE (EA), WRITE (LM-C), and WRITE (E-LM)) in the lower page of the memory cell MCn. The detection result of memory cell MCn.

再者,於本實施形態中,感測放大器SA以不同時序分三次檢測記憶胞MCn之資料。但,感測放大器SA亦可以不同時序分兩次檢測記憶胞MCn之資料。該情形時,記憶體可考慮鄰接記憶胞MCn-1或MCn+1中之任一者或兩者之鄰接干擾效應而選擇並採用檢測結果。 Furthermore, in the present embodiment, the sense amplifier SA detects the data of the memory cell MCn three times at different timings. However, the sense amplifier SA can also detect the data of the memory cell MCn twice in different timings. In this case, the memory may select and employ the detection result in consideration of the adjacent interference effect of either or both of the memory cells MCn-1 or MCn+1.

例如,圖12(A)係表示自記憶胞MCn之三個檢測結果選擇之選擇結果之表。圖12(A)所示之實例1係鄰接記憶胞MCn-1為E或B位準、鄰接記憶胞MCn+1為E位準之實例。該情形時,記憶胞MCn並未自鄰接記憶胞MCn-1、MCn+1受到鄰接干擾效應。因此,記憶體選擇在t2-1檢測出之Sense1。實例2係鄰接記憶胞MCn-1為E或B位準、鄰接記憶胞MCn+1為LM位準之實例。該情形時,記憶胞MCn雖未自鄰接記憶胞MCn-1受到鄰接干擾效應,但自鄰接記憶胞MCn+1受到鄰接干擾效應。因此,記憶體選擇在t2-2檢測出之Sense2。實例3係鄰接記憶胞MCn-1為A或C位準、鄰接記憶胞MCn+1為E位準之實例。該情形時,記憶胞MCn雖未自鄰接記憶胞MCn+1受到鄰接干擾效應,但自鄰接記憶胞MCn-1受到鄰接干擾效應。因此,記憶體選擇在t2-2檢測出之Sense2。進而,實例4係鄰接記憶胞MCn-1為A或C位準、鄰接記憶胞MCn+1為LM位準之實例。該情形時,記憶胞MCn自鄰接 記憶胞MCn-1及MCn+1之兩者受到鄰接干擾效應。因此,記憶體選擇在t2-3檢測出之Sense3。 For example, Fig. 12(A) shows a table of selection results of selection of three detection results from the memory cell MCn. Example 1 shown in Fig. 12(A) is an example in which the adjacent memory cell MCn-1 is at the E or B level and the adjacent memory cell MCn+1 is at the E level. In this case, the memory cell MCn is not subjected to the adjacent interference effect from the adjacent memory cells MCn-1, MCn+1. Therefore, the memory selects Sense1 detected at t2-1. Example 2 is an example in which the adjacent memory cell MCn-1 is at the E or B level, and the adjacent memory cell MCn+1 is the LM level. In this case, although the memory cell MCn is not subjected to the adjacent interference effect from the adjacent memory cell MCn-1, it is subject to the adjacent interference effect from the adjacent memory cell MCn+1. Therefore, the memory selects Sense2 detected at t2-2. Example 3 is an example in which the adjacent memory cell MCn-1 is at the A or C level and the adjacent memory cell MCn+1 is at the E level. In this case, although the memory cell MCn is not subjected to the adjacent interference effect from the adjacent memory cell MCn+1, the adjacent memory cell MCn-1 is subject to the adjacent interference effect. Therefore, the memory selects Sense2 detected at t2-2. Further, Example 4 is an example in which the adjacent memory cell MCn-1 is at the A or C level and the adjacent memory cell MCn+1 is the LM level. In this case, the memory cell MCn is self-contiguous Both memory cells MCn-1 and MCn+1 are subject to adjacent interference effects. Therefore, the memory selects Sense3 detected at t2-3.

圖12(B)係表示自記憶胞MCn之兩個檢測結果選擇之選擇結果之表。檢測結果為兩個之情形時(例如Sense1及Sense2之情形時),記憶體根據鄰接干擾效應而選擇Sense1及Sense2中之任一個。 Fig. 12(B) is a table showing selection results of selection of two detection results from the memory cell MCn. When the detection result is two (for example, in the case of Sense1 and Sense2), the memory selects either one of Sense1 and Sense2 depending on the adjacent interference effect.

圖12(B)所示之實例11中,與實例1同樣地,記憶胞MCn未自鄰接記憶胞MCn-1、MCn+1受到鄰接干擾效應。因此,記憶體選擇在t2-1檢測出之Sense1。實例14係與實例4同樣地,記憶胞MCn自鄰接記憶胞MCn-1及MCn+1之兩者受到鄰接干擾效應。因此,記憶體選擇在t2-2檢測出之Sense2。 In the example 11 shown in Fig. 12(B), as in the case of the example 1, the memory cell MCn is not subjected to the adjacent interference effect from the adjacent memory cells MCn-1, MCn+1. Therefore, the memory selects Sense1 detected at t2-1. Example 14 In the same manner as in Example 4, the memory cell MCn was subjected to the adjacent interference effect from both adjacent memory cells MCn-1 and MCn+1. Therefore, the memory selects Sense2 detected at t2-2.

此處,實例12係與實例2同樣地,記憶胞MCn雖未自鄰接記憶胞MCn-1受到鄰接干擾效應,但自鄰接記憶胞MCn+1受到鄰接干擾效應。實例13係與實例3同樣地,記憶胞MCn雖未自鄰接記憶胞MCn+1受到鄰接干擾效應,但自鄰接記憶胞MCn-1受到鄰接干擾效應。因此,於實例2、3中,記憶胞MCn僅受到一次鄰接干擾效應。因此,於該情形時,亦可如選擇結果Result1般不考慮鄰接干擾效應而選擇Sense1。或者,亦可如選擇結果Result2般考慮到鄰接干擾效應而選擇Sense2。 Here, in the example 12, as in the case of the example 2, the memory cell MCn is not subjected to the adjacent interference effect from the adjacent memory cell MCn-1, but is subjected to the adjacent interference effect from the adjacent memory cell MCn+1. Example 13 In the same manner as in Example 3, although the memory cell MCn was not subjected to the adjacent interference effect from the adjacent memory cell MCn+1, the adjacent memory cell MCn-1 was subjected to the adjacent interference effect. Therefore, in Examples 2 and 3, the memory cell MCn is only subjected to a contiguous interference effect. Therefore, in this case, Sense1 can be selected without considering the adjacent interference effect as in the selection result Result1. Alternatively, Sense2 may be selected in consideration of the adjacent interference effect as in the selection Result2.

(變化例1) (Variation 1)

圖7(A)~圖7(C)係表示第1實施形態之變化例1之記憶體之IDL動作的時序圖。根據變化例1,於檢測時序t2-1~t2-3之各者,使字元線WLn之電壓升高(step-up)至VWLn1、VWLn2、VWLn3。變化例1之其他動作可與第1實施形態之對應之動作相同。 7(A) to 7(C) are timing charts showing the IDL operation of the memory of the first modification of the first embodiment. According to the first modification, the voltage of the word line WLn is step-up-up to VWLn1, VWLn2, and VWLn3 at each of the detection timings t2-1 to t2-3. The other operations of the first modification can be the same as the operations corresponding to the first embodiment.

於變化例1中,IDL時感測放大器SA檢測連接於與作為IDL之對象之字元線WLn之兩側鄰接之字元線WLn-1及WLn+1之記憶胞之資料,進而一面變更字元線WLn之電壓,一面複數次檢測記憶胞MCn之 資料。只要t2-1~t2-3中之複數次檢測與將字元線WLn之電壓設定為VWLn1~VWLn3時獲得之檢測對應,則亦可如變化例1般使字元線WLn之電壓以與檢測時序t2-1~t2-3對應之方式變化。 In the first modification, the sense amplifier SA at the IDL detects the data of the memory cell connected to the word lines WLn-1 and WLn+1 adjacent to both sides of the word line WLn which is the target of the IDL, and further changes the word. The voltage of the WLn of the element line detects the memory cell MCn multiple times data. As long as the plurality of detections in t2-1 to t2-3 correspond to the detection obtained when the voltage of the word line WLn is set to VWLn1 to VWLn3, the voltage of the word line WLn can be detected and detected as in the first modification. The timing t2-1~t2-3 changes in a corresponding manner.

(變化例2) (Variation 2)

圖8(A)~圖8(C)係表示第1實施形態之變化例2之記憶體之IDL動作的時序圖。變更字元線WLn之電壓之情形時,亦可如圖8(A)~圖8(C)之t2~t3所示般變更鄰接位元線WLn+1、WLn-1之電壓。若利用鄰接干擾效應,則字元線WLn之電壓伴隨鄰接字元線WLn+1、WLn-1之電壓之變更而變動。如此,即使以與檢測時序t2-1~t2-3對應之方式變更鄰接字元線WLn+1、WLn-1之電壓,亦可與上述變化例1同樣地,使字元線WLn之電壓升高至VWLn1、VWLn2、VWLn3。再者,於t2~t3,因對字元線WLn例如施加有VLM,故於圖8之t2~t3,字元線WLn之電壓固定顯示為VLM。然而,實際上字元線WLn之電壓會受到來自鄰接字元線WLn+1、WLn-1之電壓之鄰接干擾效應而與圖7之t2~t3之字元線WLn同樣地動作。 8(A) to 8(C) are timing charts showing the IDL operation of the memory of the second modification of the first embodiment. When the voltage of the word line WLn is changed, the voltages of the adjacent bit lines WLn+1 and WLn-1 may be changed as shown by t2 to t3 in FIGS. 8(A) to 8(C). When the adjacent interference effect is utilized, the voltage of the word line WLn fluctuates with the change of the voltage of the adjacent word lines WLn+1 and WLn-1. In this manner, even if the voltages of the adjacent word lines WLn+1 and WLn-1 are changed so as to correspond to the detection timings t2-1 to t2-3, the voltage of the word line WLn can be raised in the same manner as in the first modification. Up to VWLn1, VWLn2, VWLn3. Further, since t3 to t3, for example, VLM is applied to the word line WLn, the voltage of the word line WLn is fixedly displayed as VLM in t2 to t3 of FIG. However, in actuality, the voltage of the word line WLn is subjected to the adjacent interference effect from the voltages of the adjacent word lines WLn+1 and WLn-1, and operates in the same manner as the word line WLn of t2 to t3 of Fig. 7 .

當然,亦可交換鄰接字元線WLn+1之動作與鄰接字元線WLn-1之動作,而使字元線WLn之電壓變動。 Of course, the operation of the adjacent word line WLn+1 and the operation of the adjacent word line WLn-1 may be exchanged to change the voltage of the word line WLn.

(第2實施形態) (Second embodiment)

圖9(A)~圖9(C)係表示第2實施形態之記憶體之IDL動作的時序圖。於第2實施形態中,在t0~t1,記憶胞MCn-1之資料檢測係將字元線WLn-1之電壓位準設定為VC而執行。未執行字元線WLn-1之電壓位準為VA及VB時之資料檢測。因此,於第2實施形態中,對記憶胞MCn-1中之處於E、A及B位準之記憶胞與處於C位準之記憶胞加以區分。即,記憶胞MCn-1中之處於E、A及B位準之記憶胞未被區分。第2實施形態之其他動作可與第1實施形態之對應之動作相同。 9(A) to 9(C) are timing charts showing the IDL operation of the memory of the second embodiment. In the second embodiment, at t0 to t1, the data detection of the memory cell MCn-1 is performed by setting the voltage level of the word line WLn-1 to VC. The data detection when the voltage level of the word line WLn-1 is VA and VB is not performed. Therefore, in the second embodiment, the memory cells at the E, A, and B levels in the memory cell MCn-1 are distinguished from the memory cells at the C level. That is, the memory cells at the E, A, and B levels in the memory cell MCn-1 are not distinguished. The other operations of the second embodiment can be the same as the operations corresponding to the first embodiment.

如此,因記憶胞MCn-1中之處於E、A及B位準之記憶胞未被區 分,故不考慮記憶胞MCn之下級頁面受到之鄰接干擾效應(WRITE(E-A)、WRITE(LM-C)及WRITE(E-LM)造成之鄰接干擾效應)中之記憶胞MCn-1之上級頁面之WRITE(E-A)造成之鄰接干擾效應。即,序列控制電路7不考慮WRITE(E-A)造成之鄰接干擾效應而選擇記憶胞MCn之檢測結果。 Thus, due to memory cell MCn-1, the memory cell is not in the E, A and B levels. Therefore, the memory cell MCn-1 in the upper level of the memory cell MCn is not considered to be affected by the adjacent interference effect (the adjacent interference effect caused by WRITE (EA), WRITE (LM-C) and WRITE (E-LM)). The adjacent interference effect caused by the WRITE (EA) of the page. That is, the sequence control circuit 7 selects the detection result of the memory cell MCn regardless of the adjacent interference effect caused by WRITE (E-A).

然而,序列控制電路7可於考慮到其他WRITE(LM-C)及WRITE(E-LM)造成之鄰接干擾效應之基礎上,選擇記憶胞MCn之檢測結果。因此,即使記憶胞MCn之閾值分佈因鄰接干擾效應而重複,第2實施形態之記憶體亦可於IDL中某種程度上準確地檢測資料。 However, the sequence control circuit 7 can select the detection result of the memory cell MCn on the basis of the adjacent interference effects caused by other WRITE (LM-C) and WRITE (E-LM). Therefore, even if the threshold distribution of the memory cell MCn is repeated due to the adjacent interference effect, the memory of the second embodiment can detect the data to some extent accurately in the IDL.

(第3實施形態) (Third embodiment)

圖10(A)~圖10(C)係表示第3實施形態之記憶體之IDL動作的時序圖。於第3實施形態中,在t0~t1,未執行記憶胞MCn-1之資料檢測。即,記憶胞MCn-1之資料未被區分。 10(A) to 10(C) are timing charts showing the IDL operation of the memory of the third embodiment. In the third embodiment, data detection of the memory cell MCn-1 is not performed at t0 to t1. That is, the data of the memory cell MCn-1 is not distinguished.

因記憶胞MCn-1之資料未被區分,故不考慮記憶胞MCn之下級頁面受到之鄰接干擾效應(WRITE(E-A)、WRITE(LM-C)及WRITE(E-LM)造成之鄰接干擾效應)中之記憶胞MCn-1之上級頁面之WRITE(E-A)及WRITE(LM-C)造成之鄰接干擾效應。即,序列控制電路7不考慮WRITE(E-A)及WRITE(LM-C)造成之鄰接干擾效應而選擇記憶胞MCn之檢測結果。 Since the data of the memory cell MCn-1 is not distinguished, the adjacent interference effects (WRITE (EA), WRITE (LM-C) and WRITE (E-LM) caused by the adjacent pages of the memory cell MCn are not considered. The adjacent interference effects caused by WRITE (EA) and WRITE (LM-C) of the memory cell MCn-1 upper page. That is, the sequence control circuit 7 selects the detection result of the memory cell MCn regardless of the adjacent interference effect caused by WRITE (E-A) and WRITE (LM-C).

如此,於第3實施形態中,因不考慮記憶胞MCn-1之上級頁面之寫入所造成之鄰接干擾效應,故t2~t3之檢測動作可分t2-1及t2-2兩次進行。藉由將檢測動作設為兩次,可縮短驗證之讀入時間。第3實施形態之其他動作可基本上與第1實施形態之對應之動作相同。 As described above, in the third embodiment, since the adjacent interference effect caused by the writing of the upper page of the memory cell MCn-1 is not considered, the detection operation of t2 to t3 can be performed twice by t2-1 and t2-2. By setting the detection action twice, the verification read time can be shortened. The other operations of the third embodiment can be basically the same as the operations corresponding to the first embodiment.

於第3實施形態中,序列控制電路7可於考慮到記憶胞MCn+1之WRITE(E-LM)造成之鄰接干擾效應之基礎上,選擇記憶胞MCn之檢測結果。因此,即使記憶胞MCn之閾值分佈因鄰接干擾效應而重複, 第3實施形態之記憶體亦可於IDL中某種程度上準確地檢測資料。 In the third embodiment, the sequence control circuit 7 can select the detection result of the memory cell MCn in consideration of the adjacent interference effect caused by the WRITE (E-LM) of the memory cell MCn+1. Therefore, even if the threshold distribution of the memory cell MCn is repeated due to the adjacent interference effect, The memory of the third embodiment can also detect data to some extent accurately in the IDL.

(第4實施形態) (Fourth embodiment)

圖11(A)~圖11(C)係表示第4實施形態之記憶體之IDL動作的時序圖。於第4實施形態中,在t1~t2,未執行記憶胞MCn+1之資料檢測。即,記憶胞MCn+1之資料未被檢測。 11(A) to 11(C) are timing charts showing the IDL operation of the memory of the fourth embodiment. In the fourth embodiment, data detection of the memory cell MCn+1 is not performed at t1 to t2. That is, the data of the memory cell MCn+1 is not detected.

因記憶胞MCn+1之資料未被檢測,故不考慮記憶胞MCn之下級頁面受到之鄰接干擾效應(WRITE(E-A)、WRITE(LM-C)及WRITE(E-LM)造成之鄰接干擾效應)中之記憶胞MCn+1之下級頁面之WRITE(E-LM)造成之鄰接干擾效應。即,序列控制電路7不考慮WRITE(E-LM)造成之鄰接干擾效應而選擇記憶胞MCn之檢測結果。 Since the data of the memory cell MCn+1 is not detected, the adjacent interference effects (WRITE (EA), WRITE (LM-C) and WRITE (E-LM) caused by the adjacent pages of the memory cell MCn are not considered. The adjacent interference effect caused by WRITE (E-LM) of the memory cell MCn+1 lower level page. That is, the sequence control circuit 7 selects the detection result of the memory cell MCn regardless of the adjacent interference effect caused by WRITE (E-LM).

如此,於第4實施形態中,因不考慮記憶胞MCn+1之上級頁面之寫入所造成之鄰接干擾效應,故t2~t3之檢測動作可分t2-1及t2-2兩次進行。藉由將檢測動作設為兩次,可縮短驗證之讀入時間。第4實施形態之其他動作可基本上與第1實施形態之對應之動作相同。 As described above, in the fourth embodiment, since the adjacent interference effect caused by the writing of the upper page of the memory cell MCn+1 is not considered, the detection operation of t2 to t3 can be performed twice by t2-1 and t2-2. By setting the detection action twice, the verification read time can be shortened. The other operations of the fourth embodiment can be basically the same as the operations corresponding to the first embodiment.

於第4實施形態中,序列控制電路7可於考慮到記憶胞MCn-1之WRITE(E-A)及WRITE(LM-C)造成之鄰接干擾效應之基礎上選擇記憶胞MCn之檢測結果。因此,即使記憶胞MCn之閾值分佈因鄰接干擾效應而重複,第4實施形態之記憶體亦可於IDL中某種程度上準確地檢測資料。 In the fourth embodiment, the sequence control circuit 7 can select the detection result of the memory cell MCn in consideration of the adjacent interference effect caused by the WRITE (E-A) and the WRITE (LM-C) of the memory cell MCn-1. Therefore, even if the threshold distribution of the memory cell MCn is repeated due to the adjacent interference effect, the memory of the fourth embodiment can detect the data to some extent accurately in the IDL.

第3及第4實施形態可與上述變化例1或變化例2進行組合。 The third and fourth embodiments can be combined with the above-described modification 1 or modification 2.

關於記憶胞陣列111之構成,例如被記載於「三維積層非揮發性半導體記憶體」之2009年3月19日申請之美國專利申請案12/407,403號中。又,記載於「三維積層非揮發性半導體記憶體」之2009年3月18日申請的美國專利申請案12/406,524號、「非揮發性半導體記憶裝置及其製造方法」之2010年3月25日申請的美國專利申請案12/679,991號、及「半導體記憶體及其製造方法」之2009年3月23日申請的美國專利 申請案12/532,030號中。該等專利申請案之全部內容以參照之方式而引用於本說明書中。 The configuration of the memory cell array 111 is described in, for example, U.S. Patent Application Serial No. 12/407,403, filed on Jan. 19, 2009. In addition, U.S. Patent Application Serial No. 12/406,524, filed on Mar. U.S. Patent Application Serial No. 12/679,991, filed on Jan. 23, 2009, and U.S. Patent Application Serial No. Application No. 12/532,030. The entire contents of these patent applications are hereby incorporated by reference.

雖已說明本發明之若干實施形態,但該等實施形態係作為示例而提出者,並非意欲限制本發明之範圍。該等實施形態可以其他多種形態實施,可於不脫離本發明之主旨之範圍內進行各種省略、置換及變更。該等實施形態及其變形包含於發明之範圍或主旨,同樣地包含於專利申請範圍所記載之發明及其均等之範圍內。 The embodiments of the present invention have been described, but are not intended to limit the scope of the present invention. The embodiments may be embodied in a variety of other forms, and various omissions, substitutions and changes may be made without departing from the scope of the invention. The scope of the invention and the scope of the invention are intended to be included within the scope of the invention and the scope of the invention.

VA‧‧‧電壓位準 VA‧‧‧voltage level

VB‧‧‧電壓位準 VB‧‧‧ voltage level

VC‧‧‧電壓位準 VC‧‧‧ voltage level

VLM‧‧‧電壓位準 VLM‧‧‧ voltage level

VREAD‧‧‧高位準電壓 VREAD‧‧‧ high level voltage

Claims (10)

一種半導體記憶裝置,其包含:記憶胞陣列,其具有複數個記憶胞,該等記憶胞可記憶具有下級頁面及上級頁面之複數個位元之資料;複數條字元線,其連接於上述複數個記憶胞;複數條位元線,其連接於上述複數個記憶胞之電流路徑之一端;及感測放大器部,其檢測對連接於上述複數條字元線中之上述字元線WLn-1(n為整數)之記憶胞寫入之上述上級頁面之資料,接著於檢測對連接於上述字元線WLn+1之記憶胞寫入之上述下級頁面之資料後,檢測對連接於上述字元線WLn之記憶胞寫入之上述下級頁面之資料;且於檢測連接於上述字元線WLn之記憶胞之資料之際,重複進行複數次檢測,取得複數次檢測結果。 A semiconductor memory device comprising: a memory cell array having a plurality of memory cells, wherein the memory cells can memorize data having a plurality of bits of a lower page and a higher page; and a plurality of word lines connected to the plurality of bits a memory cell; a plurality of bit lines connected to one end of the current path of the plurality of memory cells; and a sense amplifier portion detecting the pair of word lines WLn-1 connected to the plurality of word lines (n is an integer) the data of the upper page of the memory cell is written, and then after detecting the data of the lower-level page written to the memory cell connected to the word line WLn+1, the detection pair is connected to the character The memory of the line WLn is written to the data of the lower-level page; and when the data of the memory cell connected to the word line WLn is detected, the plurality of detections are repeated to obtain a plurality of detection results. 如請求項1之半導體記憶裝置,其進而包含序列控制電路,其複數次檢測連接於上述字元線WLn之記憶胞之資料,選擇該複數次檢測結果中之任一個。 The semiconductor memory device of claim 1, further comprising a sequence control circuit that detects the data of the memory cells connected to the word line WLn at a plurality of times and selects any one of the plurality of detection results. 如請求項2之半導體記憶裝置,其中上述序列控制電路所進行之對連接於上述字元線WLn之記憶胞之資料之該複數次檢測結果中之任一個之選擇,係依據連接於上述字元線WLn-1及上述WLn+1之相鄰記憶胞之檢測資料而進行。 The semiconductor memory device of claim 2, wherein the selection of any one of the plurality of detection results of the data of the memory cell connected to the word line WLn by the sequence control circuit is based on the connection to the character The detection data of the adjacent memory cells of the line WLn-1 and the above WLn+1 are performed. 如請求項3之半導體記憶裝置,其中上述序列控制電路所進行之對上述資料之該複數次檢測結果中之任一個之選擇,係於每個上述位元線進行。 The semiconductor memory device of claim 3, wherein the selection of any one of the plurality of detection results of the data by the sequence control circuit is performed on each of the bit lines. 如請求項1之半導體記憶裝置,其中上述複數次檢測係不同時序 之第1檢測(t2-1)、第2檢測(t2-2>t2-1)及第3檢測(t2-3>t2-2),且分別以上述第1、第2及第3檢測來檢測感測放大器之感測節點之電壓變化。 The semiconductor memory device of claim 1, wherein the plurality of detections are different timings The first detection (t2-1), the second detection (t2-2>t2-1), and the third detection (t2-3>t2-2) are performed by the first, second, and third detections, respectively. Detecting the voltage change of the sense node of the sense amplifier. 如請求項1之半導體記憶裝置,其中於檢測連接於上述字元線WLn之上述記憶胞之上述下級頁面之資料時,上述字元線WLn之電壓係設定為固定。 The semiconductor memory device of claim 1, wherein when the data of the lower-level page of the memory cell connected to the word line WLn is detected, the voltage of the word line WLn is set to be fixed. 如請求項6之半導體記憶裝置,其中於檢測對連接於上述字元線WLn+1之記憶胞寫入之上述下級頁面之資料之際對上述字元線WLn+1施加之電壓,與接著於檢測對連接於上述字元線WLn之記憶胞寫入之上述下級頁面之資料之際對上述字元線WLn施加之電壓係相同。 The semiconductor memory device of claim 6, wherein the voltage applied to the word line WLn+1 is detected when detecting data of the lower-level page written to the memory cell connected to the word line WLn+1, and then The voltage applied to the word line WLn is the same when the data of the lower-level page written to the memory cell of the word line WLn is detected. 如請求項1之半導體記憶裝置,其中於檢測連接於上述字元線WLn之上述記憶胞之上述下級頁面時,上述字元線WLn之電壓變更為與複數次檢測動作之各者對應之電壓。 The semiconductor memory device of claim 1, wherein when the lower page of the memory cell connected to the word line WLn is detected, the voltage of the word line WLn is changed to a voltage corresponding to each of the plurality of detection operations. 如請求項1之半導體記憶裝置,其中於檢測連接於上述字元線WLn之上述記憶胞之上述下級頁面時,上述字元線WLn+1或WLn-1之電壓係設定為與複數次檢測動作之各者對應之電壓。 The semiconductor memory device of claim 1, wherein when detecting the lower-level page of the memory cell connected to the word line WLn, the voltage of the word line WLn+1 or WLn-1 is set to a plurality of detection operations. The voltage corresponding to each. 一種半導體記憶裝置,其包含:記憶胞陣列,其具有複數個記憶胞,該等記憶胞可記憶具有下級頁面及上級頁面之複數個位元之資料;複數條字元線,其連接於上述複數個記憶胞;複數條位元線,其連接於上述複數個記憶胞之電流路徑之一端;及感測放大器部,其於檢測對連接於上述複數條字元線中之上述字元線WLn+1之記憶胞寫入之上述下級頁面之資料後,檢測對連接於上述字元線WLn之記憶胞寫入之上述下級頁面之資料;且 於檢測連接於上述字元線WLn之記憶胞之資料之際,重複進行複數次檢測;檢測連接於上述字元線WLn之上述記憶胞之上述下級頁面之電壓,與接著檢測對連接於上述字元線WLn之記憶胞寫入之上述下級頁面之資料之電壓,係設定為相同之固定電壓。 A semiconductor memory device comprising: a memory cell array having a plurality of memory cells, wherein the memory cells can memorize data having a plurality of bits of a lower page and a higher page; and a plurality of word lines connected to the plurality of bits a memory cell; a plurality of bit lines connected to one of the current paths of the plurality of memory cells; and a sense amplifier portion for detecting the pair of word lines WLn+ connected to the plurality of word lines After the data of the above-mentioned lower page is written by the memory cell, the data of the lower-level page written to the memory cell connected to the word line WLn is detected; And detecting a plurality of times of detecting the data of the memory cell connected to the word line WLn; detecting a voltage of the lower page connected to the memory cell of the word line WLn, and subsequently detecting the pair connected to the word The voltage of the data of the above-mentioned lower page written by the memory cell of the source line WLn is set to the same fixed voltage.
TW104143444A 2013-08-19 2013-08-19 Semiconductor memory device TWI601152B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW104143444A TWI601152B (en) 2013-08-19 2013-08-19 Semiconductor memory device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW104143444A TWI601152B (en) 2013-08-19 2013-08-19 Semiconductor memory device

Publications (2)

Publication Number Publication Date
TW201627998A TW201627998A (en) 2016-08-01
TWI601152B true TWI601152B (en) 2017-10-01

Family

ID=57181837

Family Applications (1)

Application Number Title Priority Date Filing Date
TW104143444A TWI601152B (en) 2013-08-19 2013-08-19 Semiconductor memory device

Country Status (1)

Country Link
TW (1) TWI601152B (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090115442A1 (en) * 1999-12-27 2009-05-07 Elpida Memory, Inc. Semiconductor integrated circuit and electronic device
US20110051530A1 (en) * 2009-08-26 2011-03-03 Kabushiki Kaisha Toshiba Semiconductor memory device and method of updating data stored in the semiconductor memory device
US20120069666A1 (en) * 2010-09-22 2012-03-22 Kabushiki Kaisha Toshiba Memory system
US20120236636A1 (en) * 2011-03-17 2012-09-20 Kabushiki Kaisha Toshiba Non-volatile semiconductor storage device
US20130003454A1 (en) * 2011-06-29 2013-01-03 Kabushiki Kaisha Toshiba Non-volatile semiconductor memory device
US8385126B2 (en) * 2011-04-06 2013-02-26 Kabushiki Kaisha Toshiba Nonvolatile semiconductor memory device
US20130077404A1 (en) * 2011-09-26 2013-03-28 Kabushiki Kaisha Toshiba Nonvolatile semiconductor memory device
US20130148435A1 (en) * 2011-12-09 2013-06-13 Kabushiki Kaisha Toshiba Non-volatile semiconductor storage device

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090115442A1 (en) * 1999-12-27 2009-05-07 Elpida Memory, Inc. Semiconductor integrated circuit and electronic device
US20110051530A1 (en) * 2009-08-26 2011-03-03 Kabushiki Kaisha Toshiba Semiconductor memory device and method of updating data stored in the semiconductor memory device
US20120069666A1 (en) * 2010-09-22 2012-03-22 Kabushiki Kaisha Toshiba Memory system
US20120236636A1 (en) * 2011-03-17 2012-09-20 Kabushiki Kaisha Toshiba Non-volatile semiconductor storage device
US8385126B2 (en) * 2011-04-06 2013-02-26 Kabushiki Kaisha Toshiba Nonvolatile semiconductor memory device
US20130003454A1 (en) * 2011-06-29 2013-01-03 Kabushiki Kaisha Toshiba Non-volatile semiconductor memory device
US20130077404A1 (en) * 2011-09-26 2013-03-28 Kabushiki Kaisha Toshiba Nonvolatile semiconductor memory device
US20130148435A1 (en) * 2011-12-09 2013-06-13 Kabushiki Kaisha Toshiba Non-volatile semiconductor storage device

Also Published As

Publication number Publication date
TW201627998A (en) 2016-08-01

Similar Documents

Publication Publication Date Title
US10163518B2 (en) Non-volatile memory device for reading data with optimized read voltage
CN110085272B (en) Semiconductor memory device with a memory cell having a memory cell with a memory cell having a memory cell
KR102355580B1 (en) Nonvolatile memory device, storage device having the same, and operation method thereof
CN108573728B (en) Semiconductor memory device and memory system
US20180151237A1 (en) Operation method of a nonvolatile memory device for controlling a resume operation
TWI496152B (en) Non-volatile semiconductor memory device and write-in method thereof
US9466381B2 (en) Semiconductor device
US9818477B2 (en) Methods of programming memory cells in non-volatile memory devices
TW201640649A (en) Semiconductor memory device including dummy memory cells and method of operating the same
US9318203B2 (en) Semiconductor device being capable of improving program speed and program disturbance characteristics
JP2016170837A (en) Semiconductor storage
KR20150091667A (en) Nonvolatile memory device and wordline driving method thereof
US10109360B2 (en) Semiconductor memory device and method of operating the same
JP2012069192A (en) Memory system
KR102645731B1 (en) Semiconductor memory device and operating method thereof
TWI614760B (en) Semiconductor memory device
JP2018163727A (en) Semiconductor memory device
CN107103933B (en) NAND flash memory and programming method thereof
JP2008091011A (en) Flash memory element and program method thereof
JP6539608B2 (en) Semiconductor memory device
JP5784788B2 (en) Nonvolatile semiconductor memory device and writing method thereof
US10409499B2 (en) NAND flash memory device and system including SLC and MLC write modes
US9558815B2 (en) Semiconductor storage device
TW201601156A (en) Semiconductor device
TWI601152B (en) Semiconductor memory device