TWI300933B - Method of reducing settling time in flash memories and improved flash memory - Google Patents
Method of reducing settling time in flash memories and improved flash memory Download PDFInfo
- Publication number
- TWI300933B TWI300933B TW95101686A TW95101686A TWI300933B TW I300933 B TWI300933 B TW I300933B TW 95101686 A TW95101686 A TW 95101686A TW 95101686 A TW95101686 A TW 95101686A TW I300933 B TWI300933 B TW I300933B
- Authority
- TW
- Taiwan
- Prior art keywords
- word line
- voltage
- biasing
- selected word
- line
- Prior art date
Links
- 230000015654 memory Effects 0.000 title claims description 55
- 238000000034 method Methods 0.000 title claims description 34
- 230000008878 coupling Effects 0.000 claims description 14
- 238000010168 coupling process Methods 0.000 claims description 14
- 238000005859 coupling reaction Methods 0.000 claims description 14
- 239000000463 material Substances 0.000 claims description 4
- 230000004044 response Effects 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims 1
- 230000035622 drinking Effects 0.000 claims 1
- 239000010902 straw Substances 0.000 claims 1
- 210000004027 cell Anatomy 0.000 description 44
- 230000003071 parasitic effect Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000004088 simulation Methods 0.000 description 4
- 238000007796 conventional method Methods 0.000 description 3
- 239000013078 crystal Substances 0.000 description 3
- 239000002184 metal Substances 0.000 description 3
- 229910052751 metal Inorganic materials 0.000 description 3
- 230000002238 attenuated effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 229910021420 polycrystalline silicon Inorganic materials 0.000 description 2
- 229920005591 polysilicon Polymers 0.000 description 2
- 239000004575 stone Substances 0.000 description 2
- 239000000758 substrate Substances 0.000 description 2
- 241000255925 Diptera Species 0.000 description 1
- 241000282376 Panthera tigris Species 0.000 description 1
- 241000239226 Scorpiones Species 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000003990 capacitor Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 230000000550 effect on aging Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 210000004692 intercellular junction Anatomy 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 239000011295 pitch Substances 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 125000006850 spacer group Chemical group 0.000 description 1
- 210000000352 storage cell Anatomy 0.000 description 1
- 235000013618 yogurt Nutrition 0.000 description 1
Landscapes
- Read Only Memory (AREA)
Description
1300933 18451twf.doc/y 九、發明說明: 【發明所屬之技術領域】 本發明一般性的涉及快閃記憶體裝置以及從其讀取 資料的方法。 【先前技術】 NAND型EEPROM(電可抹除可編褒唯讀記憶體)或快 閃記憶體已被開發以用於可檇式音樂播放器、移動電話、 數位相機等的固態大容量存儲應用,且其已被認為是硬 驅動器(HDD)的替代品。 ,、 、快閃記憶體陣列的狹長多曰*石夕字元線具有長的RC延 ?是與其相關聯的,當從所述陣列讀取資料時必須考慮到 這一點。寄生電容以及字元線到字元線電容提供Rc延 因選定字元線最她合卿近未選定字元線的較高電 f ’此=味我進行精確讀取直到紋字元線耗散了此輛 二的電壓。在間卩肖小的字元線之間齡效應變得尤其強 ^。例如’在Chen的美國專利第6,291,297號(美國‘297) 人=描述’選定字元線的選擇閘極與鄰近字元線之間的輕 二提t、刀_閘快閃記憶體晶胞中的此 引用的方式全部併入本文中。 寻冽 Μ ίΐΐΐ ’這些RC延遲提供從快閃記憶體存儲晶胞 的總存取時間。通常,勤於低於512 _的 胞:度而言,存取時間約為10叩,且對於高於1Gb 可二二^、d勺為25 Μ。儘管至少概念上講,金屬分流器 如此狹窄‘門遲’但因為與其相關聯的額外成本及在 狹乍的間距巾提供分流金屬(shunt metal)的困難,使用 1300933 18451twf.doc/y 金屬分流器一般不是優選的。 因此,仍需要一種具有經改進讀取 閃記憶體記憶體,以及從—快閃記_=:==快 【發明内容】 本發明提供―種在—快 的方法’其中—敎字元線被敎在資料子元線 讀取操作,所述方法包括以下步驟:以用於- 段之後,以—奴字元線電壓ί戶延遲時 所述讀取操作。 1 、疋子兀線以執行 …夸發明另外提供一種快閃記憶體 體裝置包括:至少—記㈣塊,其 以快,憶 及偏壓電路,其用於以一用於夫〜if‘予兀線;以 線電壓偏壓所述字元線,並以一 iii二;!=字元 定字元線以執行—讀取操作,其中,在—讀壓一選 定字元線電壓偏壓所述未選定字元i起:延 遲岐之後,以選定字元線電壓偏壓所述選定字元。i 為讓本發明之上述和其他目的、特徵和優 董下下文特舉較佳實施例,並配合所附圖式’二羊細說 【實施方式】 參看圖1,顯示包含一存儲晶胞陣列的電 程唯?記憶趙,其形成於一晶片基板上。如所== 術人貝將輯到,圖1是一 NAND快間記憶體陣列的1300933 18451twf.doc/y IX. Description of the Invention: TECHNICAL FIELD OF THE INVENTION The present invention generally relates to a flash memory device and a method of reading data therefrom. [Prior Art] NAND-type EEPROM (Electrically Erasable and Editable Read-Only Memory) or Flash Memory has been developed for solid-state mass storage applications such as portable music players, mobile phones, and digital cameras. And it has been considered as a replacement for hard drives (HDDs). , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , The parasitic capacitance and the word line to word line capacitance provide the Rc delay due to the selected word line. The highest power f of the unselected word line is the same. I am able to read accurately until the line of characters is dissipated. The voltage of this two cars. The age effect between the small word lines is particularly strong. For example, 'Chen's US Patent No. 6,291,297 (US '297) ==Describes the choice between the selected gate of the selected word line and the adjacent word line, the knife-gate flash memory crystal The manner in which this reference is made in the cell is fully incorporated herein. These RC delays provide the total access time from the flash memory storage unit. In general, the access time is about 10 勤 for a cell of less than 512 _, and 25 Μ for a value higher than 1 Gb. Although at least conceptually, metal shunts are so narrow 'doors', but because of the additional cost associated with them and the difficulty of providing shunt metal in narrow pitches, use 1300933 18451twf.doc/y metal shunts Generally not preferred. Therefore, there is still a need for an improved read flash memory memory, and from - flashing _ =: = = fast [invention] The present invention provides a "plant-on-fast method" in which - the word line is smashed In the data sub-line read operation, the method includes the following steps: after the - segment, the read operation is delayed by the - slave line voltage. 1 , 疋 兀 以 以 ... 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸 夸兀 ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; The unselected character i is: after the delay 岐, the selected character is biased with the selected word line voltage. BRIEF DESCRIPTION OF THE DRAWINGS The above and other objects, features and advantages of the present invention are set forth in the preferred embodiments of the present invention, and in conjunction with the accompanying drawings. FIG. 1 shows a memory cell array. The circuit is only memory Zhao, which is formed on a wafer substrate. If the == operator will be compiled, Figure 1 is a NAND fast memory array
1300933 18451twf.doc/y 分的電路圖。各種元件,如列以及行解碼器、感應電路以 及其他控制電路,均未圖示,以避免使本發明的揭示内容 變得模糊不清。然而,這些組件是所屬領域的技術人員所 熟知的。 將一例示性記憶體陣列劃分為很多記憶體“塊,,。每 一塊具有若干“頁,,。一頁具有很多記憶體“晶胞,,。例 如’ 1 Gb的存儲器具有1〇24個塊,且一個塊具有64頁。 每一頁具有2 K位元組(即,16 κ位元)。一字元線含有一 頁或多頁。在位元線方向上為考^一塊提供一個晶胞串或兩 個曰曰胞串。一個晶胞串具有16位元、32位或64位。 ^圖· 1所說明的記憶體陣列在所說明的實施例中包括複 數條平行位元線BL0、BL1、· · ·到BL16383,其耦合到 各個存儲晶胞’此存儲晶胞由在記憶體陣列中的個別位元 線與字元線之交叉所定義。在一優選實施例中,儘管涵蓋 如SONOS或分離閘快閃記憶體晶胞的其他晶胞結構,但 個別晶胞是浮置閘極快閃記憶體晶胞。平行字元線WL〇、 WL1、WL2 · · · WL15絕緣形成於基板上,以為個別快閃 記憶體存儲晶胞形成控制閘極。選擇電晶體與每一位元線 相關聯,並耦合到信號SS以及GS。 塊選擇信號BK用於選擇一存儲晶胞塊。cs表示共有 源極線。一晶胞串通過一側上的ss選擇電晶體連接到一 位元線,並通過GS電晶體連接到cs。在資料存取期間, 共用(global)字元線信號GWL0、GWL1、· ·.到GWL15通 1300933 18451twf.doc/y 過^自的字元線驅動電晶體’分別對字元線WL〇到WL15 不同類型的存儲晶胞具有不同的用 料的條件。-般地說,—敎的字元_卩,從其 的被偏壓在—電壓,此電壓在小於用於偏壓鄰近的 未選定字7〇_電壓。但鄰近未選定字元_較高偏壓可 耦合到此敎字H轉近未蚊偏壓必須在 可從選定字元線進行精確讀取之前將其耗散。 圖2顯示1快閃記憶體f列的字元線電阻-電容模 型。具體而言’ @ 2顯示字元線到字元㈣合電容 (Ccoup)、其他寄生電容(Cpar)以及字元線電阻(幻。對於一 0·13 μιη電晶體產生而言,即對於具有〇13 _的通道長度 的電晶體而言,可存在以下條件: ⑴子元線到字元線耦合電容(Ccoup) : 〇·〇96ίΡ/晶胞; (11)一字元線中的寄生電容(Cpar) : 0.064 fF/晶胞; • (111)總計的字元線到字元線耦合電容(16384個晶胞): 1.57 pF ; (iv) 總計的寄生電容·· 1.05 pF ; (v) 字元線寬度:013 μπι ; (vi) 字元線長度(16 X 1024 χ 2 X 013 μηι) · 4259 卿; (vii) 字元線電阻(4259 μιη/0·13 μιη X 60) : = 2 X 1〇6 Ω,假設字元線由η+多晶矽製成,且其薄層電阻為6〇 ohm ;以及 1300933 1845 ltwf.doc/y (viii)字元線 RC(1.57 pF + 1·〇5 pF) x 2 χ ι〇6 Ω ·· 5 叫。 在一種讀取常規浮置閘極快閃記憶體晶胞的方法 中,將上述選定字元線的字元線電壓設定為〇 V,以區分 編程狀態(臨界電壓高於1 V)或抹除狀態(臨界電壓低於“ V)。將未選定字元線設定為一較高的電壓,如4 V。來自 較高偏壓的未選定字元線的電壓耦合到選定字元線,並必 須在可進行讀取之前將其耗散。因此,在讀取週期期間#必 • 須提供足夠的容限以慮及這一設置時間。 - 同樣地,在讀取如美國‘29^的分離閘晶胞的其他晶胞 結構期間也須提供適當的容限。·在讀取這些分離閘晶^期 間,/巧有存儲晶胞的選擇閘極在約4·5 V偏壓以充當傳輸 、 電晶體(pass transistor)。在1·5 V偏壓選定字元線,並在9 - 偏壓未選定字元線。每—字元線在其_具#^選擇= 極。選擇閘極上的4.5 V以及鄰近未選定字元線上的9 ; 搞合到選定字元線。在晶胞結構中,字元線與選 • 間的空間很窄,例如-具有14〇 A有效氧化物厚度的_ 層。為此,選定字元線與選擇閘極之間的柄合尤发強、丄 合電容導致設置時間增加,且讀取所需ς取‘ 的审.0 圖3 -般性的說明此習知技術讀取操作信號時、, 沿-敎塊記憶體塊的字元線標注不同節點以 = 考,例如WL1B録接近轉f晶_ ^料元1參 的開始部分,WL1M指多晶料元線wu的Circuit diagram of 1300933 18451twf.doc/y. Various components, such as column and row decoders, inductive circuits, and other control circuits, are not shown to avoid obscuring the disclosure of the present invention. However, these components are well known to those skilled in the art. An exemplary memory array is divided into a number of memory "blocks." Each block has a number of "pages,". A page has a lot of memory "cells," for example, a 1 Gb memory has 1 24 blocks, and one block has 64 pages. Each page has 2 K bytes (ie, 16 κ bits). A word line contains one or more pages. In the direction of the bit line, a cell string or two cell strings are provided for the block. A cell string has 16 bits, 32 bits or 64 bits. The memory array illustrated in Fig. 1 includes a plurality of parallel bit lines BL0, BL1, ···· to BL16383 in the illustrated embodiment, which are coupled to respective memory cells 'this memory cell is used in the memory array The individual bit lines in the middle are defined by the intersection of the word lines. In a preferred embodiment, although the other unit cell structures such as SONOS or split gate flash memory cells are covered, the individual unit cells are floating gates. Flash memory cell. Parallel word lines WL〇, WL1, WL2 · · · WL15 is formed on the substrate to form a control gate for individual flash memory cells. Select transistor and each bit line Associated with and coupled to signals SS and GS. Block selection signal BK is used to select a store The cell block, cs, represents the common source line. A cell string is connected to a bit line through the ss selection transistor on one side and connected to the cs through the GS transistor. During the data access, the global word is shared. The meta-line signals GWL0, GWL1, ···. to GWL15 pass 1300933 18451twf.doc/y The word line drive transistor 'transceives the word line WL〇 to WL15 respectively. The different types of memory cells have different materials. The condition of the 敎 敎 - - - - - - - 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从 从The higher bias voltage can be coupled to this H word H. The near-free mosquito bias must be dissipated before it can be accurately read from the selected word line. Figure 2 shows the word line of the 1 flash memory f column. Resistor-capacitor model. Specifically, ' @ 2 displays word line to character (four) combined capacitance (Ccoup), other parasitic capacitance (Cpar), and word line resistance (magic. For a 0·13 μιη transistor generation That is, for a transistor having a channel length of 〇13 _, the following conditions may exist: (1) Sub-line to word Line Coupling Capacitance (Ccoup): 〇·〇96ίΡ/cell; (11) Parasitic capacitance in a word line (Cpar): 0.064 fF/cell; • (111) total word line to word line coupling Capacitance (16384 cells): 1.57 pF ; (iv) Total parasitic capacitance · 1.05 pF ; (v) Word line width: 013 μπι ; (vi) Word line length (16 X 1024 χ 2 X 013 μηι · 4259 卿; (vii) Word line resistance (4259 μιη/0·13 μιη X 60) : = 2 X 1〇6 Ω, assuming that the word line is made of η+ polysilicon and its sheet resistance is 6 〇ohm ; and 1300933 1845 ltwf.doc/y (viii) character line RC (1.57 pF + 1·〇5 pF) x 2 χ ι〇6 Ω ·· 5 In a method of reading a conventional floating gate flash memory cell, the word line voltage of the selected word line is set to 〇V to distinguish between a programming state (a threshold voltage higher than 1 V) or an erase State (threshold voltage is below "V." Set the unselected word line to a higher voltage, such as 4 V. The voltage from the unbiased unselected word line of the higher bias is coupled to the selected word line and must It is dissipated before it can be read. Therefore, during the read cycle, # must provide sufficient tolerance to account for this settling time. - Similarly, when reading a separate gate such as the US '29^ Appropriate tolerances must also be provided during the other cell structure of the unit cell. During the reading of these split gates, the select gate of the memory cell is biased at approximately 4.6 V to serve as a transmission, electricity. Pass transistor. Select the word line at 1. 5 V bias and unselect the word line at 9 - bias. Each word line is in its _ with #^ select = pole. Select 4.5 on the gate V and 9 adjacent to the unselected word line; fit to the selected word line. In the unit cell structure, the word line and the selection The space is very narrow, for example - a layer having an effective oxide thickness of 14 〇 A. For this reason, the shank between the selected word line and the selected gate is particularly strong, and the coupling capacitance causes an increase in the settling time, and the reading is performed. It is necessary to take a review of '0'. Figure 3 - General description When reading the operation signal by this conventional technique, mark the different nodes along the word line of the - block memory block to test, for example, WL1B records close to f Crystal _ ^ material element 1 the beginning of the reference, WL1M refers to the polycrystalline element line wu
1300933 18451twf.doc/y WL1E指多晶石夕字元線wu的末八 WL2字元線的開始部分,L2B指多晶石夕 t ^ WL2E „ , WL: ^ 况明方便起見’僅^條字元_說明。在料 =開’假設在讀取期間未選定字元“二 偏£’並且叙予元線在li5V偏壓。在圖3巾,選 用於讀取,並未選定WL〇、WL1以及wu到肌5。姐由 其各自字兀線選擇電晶體塊選擇信號BK同時打開每一字 元線。在此串的未選定字元線yL〇、WL1以及wl3到 WL5被偏壓在高壓’紗6 ¥。丨這6 v偏壓從鄰近的未選 定字“ $線WL1以及WL 3耦合到選定字元線WL 2,從而導 致沿選定字元線WL2的電壓的峰值。在讀取週期中分配 額外時間以在試圖讀取之前等待WL2設置為其15 v偏 壓。此額外時間提供足夠的容限來允許耦合電壓的耗散, 以允許精確確定(g卩,讀取)讀取晶胞是處於編程狀態或是 抹除狀態。 模擬說明了此不當耦合對選定字元線的設置時間之 影響。圖4是一顯示耦合電壓對選定字元線WL2的設置 時間之影響的曲線。假設多晶矽字元缚以狹窄的200ASiO2 間隔片彼此間隔。將字元線厚度設定為0.2 μιη並將閘極長 度設定為0·13 μιη。每一字元線具有2048位元組晶胞或 32Κ F長度,其中F是在製造記憶體陣列的製造過程中可 以光刻界定的最小尺寸,如0.13 μιη。位元線間距(即,位 1300933 18451twf.doc/y 元線寬度加上位元線到位元線間隔)是〇·26 μπι(即,2 F), 且因此字元線長度是4259 μιη。字元線的電阻為約2 ΜΩ。 字元線到字元線耦合電容是每晶胞每側0.096 fp。另一寄 生電容是約每晶胞0.064 fF。 如從圖4的曲線部分A可見,將未選定字元線(WL1) 偏壓到高電壓,如為類比的目的為6伏特。節點WL1B快 速上升到此高電壓,而WL1的末端節點(WL1E)以及中間 節點(WL1M)在達到高偏壓電壓過程中有一定延遲。如圖4 的曲線部分B所示,將選定字午線WL2偏壓到一較低電 壓’此處為1.5伏特,同時將未—定WL1偏壓到高電壓。 WL2的開始節點WL2B的電壓快速穩定在1·5 V,因為其 W 蠡 最接近字元線驅動器電晶體。然而,字元線的中間節點 (WL2M)處的電壓最初跳躍到約3 V,這歸因於從鄰近未選 定字元線(例如,WL1)的高電壓的電壓耦合,且其緩慢衰 減,經過約16 ps到達選定字元線的1.5 V偏壓電壓。同樣 地,選定字元線的末端節點(WL2E)處的電壓最初跳躍到約 3·6 V,並花費類似的時間量來設置為1·5 V選定字元線偏 壓。 提出一種新的讀取快閃記憶體晶胞的方法,其減少選 定字元線的設置時間。讀取方法適用於不同快閃記憶體晶 胞,如浮置閘極、分離閘、SONOS等,其經歷增加從晶胞 讀取資料所需的時間容限的設置時間。圖5說明新的快閃 記憶體的一部分及其時序。 1300933 18451twf.doc/y1300933 18451twf.doc/y WL1E refers to the beginning of the last eight WL2 character lines of the polycrystalline stone character line wu, L2B refers to the polycrystalline stone eve t ^ WL2E „ , WL: ^ It is convenient to see 'only ^ Character_Description. In material = On' Assume that the character "two offsets" is not selected during reading and that the meta-line is biased at li5V. In Figure 3, selected for reading, WL〇, WL1, and wu to muscle 5 were not selected. The sister selects the transistor block selection signal BK from its respective word line to simultaneously turn on each word line. The unselected word lines yL 〇, WL1 and wl3 to WL5 of this string are biased at the high voltage 'yarn 6 ¥. The 6 v bias is coupled from the adjacent unselected word "$ line WL1 and WL 3 to the selected word line WL 2, resulting in a peak of the voltage along the selected word line WL2. Additional time is allocated during the read cycle to Wait for WL2 to be set to its 15 V bias before attempting to read. This extra time provides sufficient tolerance to allow dissipation of the coupling voltage to allow accurate determination (g卩, read) of the read cell is in the programmed state Or erase the state. The simulation illustrates the effect of this improper coupling on the settling time of the selected word line. Figure 4 is a graph showing the effect of the coupling voltage on the settling time of the selected word line WL2. The narrow 200ASiO2 spacers are spaced apart from each other. The word line thickness is set to 0.2 μηη and the gate length is set to 0·13 μηη. Each word line has a 2048-bit cell or 32Κ F length, where F is The minimum size that can be photolithographically defined during the fabrication of the memory array, such as 0.13 μm. The bit line spacing (ie, bit 1300933 18451 twf.doc/y line width plus bit line to bit line spacing) is 〇·26 Ππι( , 2 F), and thus the word line length is 4259 μιη. The resistance of the word line is about 2 Μ Ω. The word line to word line coupling capacitance is 0.096 fp per side of the unit cell. The other parasitic capacitance is about every The cell is 0.064 fF. As can be seen from curve A of Figure 4, the unselected word line (WL1) is biased to a high voltage, for an analogy of 6 volts. Node WL1B rises rapidly to this high voltage, while WL1 The end node (WL1E) and the intermediate node (WL1M) have a certain delay in reaching a high bias voltage. As shown in the curve section B of Figure 4, the selected meridian WL2 is biased to a lower voltage 'here At 1.5 volts, the undetermined WL1 is simultaneously biased to a high voltage. The voltage at the start node WL2B of WL2 is quickly stabilized at 1.5 V because its W 蠡 is closest to the word line driver transistor. However, the word line The voltage at the intermediate node (WL2M) initially jumps to about 3 V due to the high voltage voltage coupling from the adjacent unselected word line (eg, WL1), and it slowly decays, reaching the selected word after about 16 ps. The 1.5 V bias voltage of the line. Similarly, the end node of the selected word line (WL2E) The voltage at the first jumps to about 3.6 V and takes a similar amount of time to set the selected word line bias to 1.25 V. A new method of reading the flash memory cell is proposed, which reduces The set time of the selected word line. The reading method is applicable to different flash memory cells, such as floating gates, split gates, SONOS, etc., which experience the setting of the time tolerance required to read data from the unit cell. Time. Figure 5 illustrates a portion of the new flash memory and its timing. 1300933 18451twf.doc/y
新的快閃記憶體陣列包括一以相 存儲晶胞塊,其具有複數條參考字元線,、衣w的爹考 列塊具有相同特徵(例如,輕合電容、寄生带2此與日日胞陣 儘管在晶胞陣列塊中僅顯示了五條字元線^令以及電阻)。 所示一例示性塊包括十六條字元線。同样,瞭解圖1 示為包括五條字元線,但應包括足夠數目的t 4考塊顯 反映晶胞陣列塊的耦合以及寄生電容。此、予711線以精確 陣列塊中所提供的字元線數目相同,而是==與晶, 一例示性實施例中在九條與十_^条字元線之間。又v,如在 如在下文所描述,參考塊用丨於提供—電壓^ 以追,-選定字元線(即,選定用於讀取)中人^The new flash memory array includes a phase storage cell block having a plurality of reference word lines, and the reference block of the clothing w has the same characteristics (for example, a light capacitance, a parasitic band 2 and the day) The cell array shows only five word lines and resistors in the cell array block. An exemplary block shown includes sixteen character lines. Similarly, it is understood that Figure 1 is shown to include five word lines, but should include a sufficient number of t 4 blocks to reflect the coupling of the cell array blocks and the parasitic capacitance. Thus, the number of word lines provided in the 711 lines in the exact array block is the same, but == and crystal, in the exemplary embodiment between nine lines and ten_^ word lines. Again v, as described below, the reference block is used to provide - voltage ^ to chase, - select the word line (ie, selected for reading) ^
以供與-參考電壓作比較。此比較結果⑽=選 疋子4的讀取電壓的啟動。儘管衰減可從選定字元線直 =監控’但與提供-參考塊相比,其並非最佳選擇,因為 j而要在每-記紐塊中,每—字元線與比較電路之間 連接,而非僅僅來自參考塊的字元線連接。 、,擇存儲晶胞陣列的字元線,如晶胞陣列塊的WL2 =供讀取。也從參考塊選擇或指定-字元線,如RWL2。 間,與上述習知技術的選定字孚線的h5 V(或其他 子元線碩取電壓)相比,以一更低電壓0 V偏壓選定參 、1| 6 ,且與以上結合圖4描述的習知技術的未 元線一樣,剩餘的未選定參考字元線被偏壓在6 $他未選定字元線電壓)。以未選定參考字元線相同的 12 1300933 18451twf.doc/y 電壓(即,6 V)偏壓晶胞陣列塊的未選定字元線(即,WL()、 WL1、WL3、WL4專)。不疋隶初以選定字元線讀取電壓(例 如,1·5 V)作偏壓,而是最初以一低於選定字元線讀取電 壓的笔壓(如,0 V)以偏壓上述晶胞陣列塊的選定字元線 WL2在一最初延遲時段(表示為“D”)。這時,通過如上 所述被拉到6 V的鄰近未選定參考字元線將參考字元線 RWL2的電壓耦合起來。同樣地,通過耦合到鄰近字元線 • 的高電壓耦合至選定字元線WL2。 選定參考字元線RWL2處气電壓達到峰值之後,其將 耦合電壓衰減向0 V,即,施加1 到RWL2的接地電壓。在 選定元線WL2中發生類似衰減,因為字元線陣列塊被 〜 類似構造以及偏壓。由於在此期間選定字元線WL2的功 率為〇 V(與習知技術的1·5 V相對),因此與參考字元線的 最遠節點(RWL2E)—樣,其最遠節點(WL2E)僅耦合到2.9 V的峰值(如圖6所見),其低於以上結合圖4而描述的3 6 φ V。 此種新的讀取方法,被晶胞結構的模擬方法所測試, 如圖4。上述的測试結果繪示於圖6。圖6的曲線部分(B) 顯示耦合到約2·9 V的峰值並接著向,〇 v衰減的RWL2e 處的電壓。當RWL2E電壓接近或達到此實施例中一選定 字疋線的電壓偏壓1·5 V時,產生一控制信號ACT(曲線部 为(A)所示)來觸發以選定字元線讀取電壓作偏壓以選定字 元線WL2,即,將電壓偏壓從〇 v切換到15 v。圖6的 13 1300933 18451twf.doc/y 曲線部分(c)顯示了回應控制信號ACT的WL2處的電壓 切換。 實際上,麻ACT錢仙作㈣錢來以所屬領 域的技術人員所熟知的方式控制gwl電壓,即將gwL 選定字元線讀取電壓從(例如)〇v切換到15 v。 在此讀取方法中,在-第一延遲時段D期間(在圖5 ^間T1直到時間T2),將—較低電 叙子元線减電_,丨如L5 V)胁偏麵定字 此幫助加速沿選定字元線,且帘在其最遠節點WL2E ‘ 電減。在所說明I實施例中將瞭解,節點 •化費农長的時間來耗餘合電壓,因 不應進行讀取的動作。這一 网明 Ο V 1 ^ Λ7 ^ L遲守丰又之後,開關的電源從 點WL2B ·#、ΓΓ在選定字元線上,將選定字元線的開始節 二 •、速拉到所需電壓。當節點WL2E處的電壓設晋For comparison with the - reference voltage. This comparison result (10) = the start of the read voltage of the selected dice 4. Although the attenuation can be directly from the selected word line = monitoring 'but it is not the best choice compared to the provided - reference block, because j is in each block, the connection between each word line and the comparison circuit Instead of just the word line connections from the reference block. Select the word line of the memory cell array, such as WL2 of the cell array block for reading. Also select or specify a -word line from the reference block, such as RWL2. In comparison with the h5 V of the selected word line of the above-mentioned prior art (or other sub-line voltage), the reference parameter, 1|6 is biased with a lower voltage of 0 V, and combined with FIG. 4 above. As with the conventional techniques described for the un-elements, the remaining unselected reference character lines are biased at 6 $he is not selected for the word line voltage). The unselected word lines of the cell array block (ie, WL(), WL1, WL3, WL4) are biased with the same 12 1300933 18451 twf.doc/y voltage (ie, 6 V) that is not selected for the reference word line. The bias voltage (eg, 1·5 V) is initially biased at the selected word line, but initially biased with a pen voltage (eg, 0 V) that reads the voltage below the selected word line. The selected word line WL2 of the above-described unit cell array block is in an initial delay period (denoted as "D"). At this time, the voltage of the reference word line RWL2 is coupled by the adjacent unselected reference word line pulled to 6 V as described above. Similarly, the selected word line WL2 is coupled by a high voltage coupled to the adjacent word line. After the gas voltage at the selected reference word line RWL2 reaches a peak value, it attenuates the coupling voltage to 0 V, i.e., applies a ground voltage of 1 to RWL2. A similar attenuation occurs in the selected element line WL2 because the word line array block is ~ similarly constructed and biased. Since the power of the selected word line WL2 during this period is 〇V (relative to the 1. 5 V of the prior art), it is the farthest node (WL2E) of the farthest node (RWL2E) of the reference word line. Only coupled to a peak of 2.9 V (as seen in Figure 6), which is lower than the 3 6 φ V described above in connection with Figure 4. This new reading method is tested by the simulation method of the unit cell structure, as shown in Fig. 4. The above test results are shown in Figure 6. The curve portion (B) of Figure 6 shows the voltage at RWL2e coupled to a peak of about 2·9 V and then attenuated to 〇v. When the voltage of RWL2E approaches or reaches the voltage bias of 1·5 V of a selected word line in this embodiment, a control signal ACT (shown as curve (A)) is generated to trigger the voltage reading with the selected word line. Bias is applied to select word line WL2, i.e., the voltage bias is switched from 〇v to 15v. The 13 1300933 18451 twf.doc/y curve portion (c) of Fig. 6 shows the voltage switching at WL2 in response to the control signal ACT. In fact, the ACT money is used to control the gwl voltage in a manner well known to those skilled in the art, i.e., the gwL selected word line read voltage is switched from, for example, 〇v to 15 v. In this reading method, during the -first delay period D (between T1 and T2 in FIG. 5), the lower-lower sub-line is reduced by _, such as L5 V) This helps speed up along the selected word line, and the curtain is electrically subtracted at its farthest node WL2E'. In the described embodiment of the invention, it will be understood that the time spent by the farmer is to consume the voltage and voltage, since the reading operation should not be performed. After the net ΟV 1 ^ Λ7 ^ L is late and the switch power is switched from the point WL2B ·#, ΓΓ on the selected word line, the starting point of the selected word line is twice, and the speed is pulled to the required voltage. . When the voltage at node WL2E is set
壓時,節點WL2M處的電壓也快速增加到二V 快速S ==分(Q顯示選定字元線憶的所有節點 A/.而的I.5 V選定字元線偏壓電壓,所有節% 即開始節點士扣〜 "’即點 WL2E。在此极」中間郎‘點W^2M以及末端節點 在約7叩後所古中,與圖4的整個模擬所示的16叩相比, 在、力7叩後所有節點已 選定字元線的偏[m 通過延遲 尤豆是#定〜- 奴犄間,直到沿所述字元線的電壓, 子元線的末端節點鄰近的耦合電壓可衰減或基 1300933 18451twf.doc/y 本衰減到選定字元線偏壓電壓與節點電壓之間的 (或為0V)時,所述設置時間減少。此較快的設置時間大大 減少了讀取一選定字元線的存儲晶胞所需的存取視窗。 圖7顯示一用於產生控制信號ACT的例示性電路1〇 的一實施例。圖7也顯示所述電路的一時序圖。ACT —般 通過耦合到PMOS電晶體Ml的重設信號RST設定為低電 位,所述PMOS電晶體搞合於電源電壓vcc與nm〇S電 晶體M2之間。此電路包括一比較器12,其輸入耦合到參 考電壓VR並耦合到節點處的電壓。可將VR設 定為1·5 V,或兩個偏壓電壓之^的另一觸發電壓(即,選 定字/^線偏壓電壓與未經選定字樣偏壓電壓之間),且更 優選為小於所述選定字元線上的翻峰值電壓,且更優選 地设定為約所述選定字元線偏壓電壓。開始時,犯 處的電^ 0V。在這雜件下,騎 (ουτ)為高電位。比較器12的輸_ 一具有一= ==:序邊圖Tr4々~ 選定來考字亓綠、序圖所不,RWL2E通過與鄰近的未 於VR。MWue或古所述陣列/存在的其他電壓搞合而高 當out低時,所述3二=雷輸_虎⑽變為低。 變高。〇UT=i5sv放電到參彻对號〇υτ即;At voltage, the voltage at node WL2M also rapidly increases to two V fast S == minutes (Q shows all nodes A/. of the selected word line recall. I.5 V selected word line bias voltage, all sections % That is, start the node deduction ~ "' point WL2E. In this pole "the middle Lang" point W^2M and the end node in about 7叩, compared with the 16模拟 shown in the entire simulation of Figure 4, After the force of 7 叩, all the nodes have selected the bias of the word line [m by delaying the yoghurt is #定~- 犄 between the slaves until the voltage along the word line, the coupling voltage of the end node of the sub-element Attenuation or base 1300933 18451twf.doc/y This attenuation is reduced (or 0V) between the selected word line bias voltage and the node voltage. This faster setup time greatly reduces read time. An access window required to store a cell of a selected word line. Figure 7 shows an embodiment of an exemplary circuit 1 for generating a control signal ACT. Figure 7 also shows a timing diagram of the circuit. Generally, the reset signal RST coupled to the PMOS transistor M1 is set to a low potential, and the PMOS transistor is engaged with Between the source voltage vcc and the nm〇S transistor M2. The circuit includes a comparator 12 having an input coupled to a reference voltage VR and coupled to a voltage at the node. The VR can be set to 1·5 V, or two biases Another trigger voltage of the voltage (ie, between the selected word line voltage and the unselected word bias voltage), and more preferably less than the peak value of the selected word line, and more Preferably, it is set to about the selected word line bias voltage. At the beginning, the electric charge is 0. In this miscellaneous piece, the ride (ουτ) is high. The output of the comparator 12 has a = ==: The edge map Tr4々~ The selected word is green, the sequence is not, and the RWL2E is high when the out is low, and the other voltages that are not in the VR.MWue or the ancient array/existence are high. The 3 2 = lightning loss _ tiger (10) becomes low. It becomes high. 〇 UT = i5sv discharges to the reference sign 〇υ τ;
脈衝。 切換觸發—將ACT設定為高的pUL 1300933 18451twf.doc/y 在一貫施例中,正邊緣觸發電路14包括一延遲鏈16, 其包括如可數個φ接反㈣18的複數個延遲元件。正邊緣 觸發電路14也包括一 NAND閘2〇,其具有第一輸入,耦 合以接收延遲鏈16的輸出,以及第二輸入,搞合以接收信 號OUT的。在運行中,除當兩個輸入都高時,nand 2〇 產生一低輸出信號。因此,來自反相證22的信號僅 當^AND 20產生一低輸出時高。僅當信號〇υτ從低切換 φ 到鬲之後,且低k號QUT穿過延遲元件16花費的時間 ^ 内,NAND20的輸入均高。在一y例示性實施例中,所述延 遲為約2 ns。 * N^ND閘20的輸出耦合到一反相器22,此反相器提 - 供信號PUL至NM0S電晶體M2,當電晶體M2為導通時,pulse. Switching Trigger - Setting ACT to High pUL 1300933 18451 twf.doc / y In a consistent embodiment, positive edge trigger circuit 14 includes a delay chain 16 that includes a plurality of delay elements such as a countable number of φ inverted (four) 18. The positive edge trigger circuit 14 also includes a NAND gate 2A having a first input coupled to receive the output of the delay chain 16, and a second input coupled to receive the signal OUT. In operation, nand 2〇 produces a low output signal except when both inputs are high. Therefore, the signal from the inverted certificate 22 is high only when ^AND 20 produces a low output. The input of NAND 20 is high only when the signal 〇υτ switches from low to φ to 鬲 and the time that the low k number QUT passes through the delay element 16 is within ^. In a y exemplary embodiment, the delay is about 2 ns. * The output of the N^ND gate 20 is coupled to an inverter 22 which supplies a signal PUL to the NM0S transistor M2, when the transistor M2 is turned on,
• 即當PUL為高且RST為高時,NMOS電晶體又將信號PUL 傳輸到栓鎖電路24。 在一實施例中,栓鎖電路24包含一對反相器26、28, ⑩ 反相器26的輸出耦合到反相器28的輸入,且反相器28 的輸出耦合到反相器26的輸入。栓鎖電路24經配置以操 取並保持正邊緣觸發電路14的啟動輸出,以供如上所描述 般隨後使用。 讀取之後,將輸入新的命令(即,讀取、寫入或抹除命 令)。當輸入新的命令時,啟動RST以重定act信發。 上述讀取方法在僅利用單個驅動器電晶體的快閃記 憶體陣列中最為有效,驅動器電晶體通常置於最接近字元 1300933 18451twf.doc/y 線的開始節點WLB處。如果存在一個驅動電晶體,花費 最長時間來耗散所述搞合電壓的節點處的電壓可優先被監 控。通常,這一節點將是離驅動電晶體最遠的節點。如= 利用兩個驅動器電晶體,每一個位於各自字元線的每一 端,那麼所述字元線的電壓將更快地穩定。然而,出於大 小、成本、複雜性以及/或功率考慮,使用一個以上驅動器 電晶體通常是不希望的。然而,即使快閃記憶體陣列每條 • 字元線利用兩個驅動器電晶體,可利用本文中描述的方法 • 來進一步改進設置時間。在此實_例中,優選將選定字元 線的中間節點用作所述控制信號k生電路10的比較器12 的輸入。 uyr · - 儘管以上已結合一讀取協議描述了所述經改進的讀 一 取方法,其中在選定字元線處將所述偏壓設定為一正電 壓,如1·5 V,並在未選定字元線處將其設定為一更高電 壓,如6 V,但所述方法並非如此限制。例如,對於一標 肇準浮置閘極快閃記憶體晶胞而言,其中所述讀取條件在選 定字元線處為0 V,且在未選定字元線處為4 ν,可最初將 選定字元線偏壓到某一負電壓直到觸發信號ACT,如當— 參考字元線衰減到〇 V時。相應地設零圖7電路的參考電 壓VR 〇 儘管以上已參考用於特定類变快閃$己彳思體晶胞的讀 取條件描述了所述新的讀取方法,但所屬領域的技術人員 將瞭解所述方法一般可用於以下任何快閃ό己憶體晶胞結 17 1300933 18451twf.doc/y 構·其展不㈣的字元相字元線耦合,且其巾所述選6 字兀線具有-與未選定字元線偏壓電壓不同的偏壓電壓, 儘管已根據例示性實施例描述了本發明,但其並不限 於此。相反,應將隨附權利要求書廣泛地解釋為包括所^ 領域的技術人員可在沒有背離本發明的均等物的範疇以及 範圍的情況下做出的本發明的其他變體以及實施例了 【圖式簡單說明】 附圖說明本發明的優選實施例及有關本揭示内 其他資訊,其中: 1 圖1是一具有複數個NANDf存儲晶胞的快 憶體的•一部分的電路圖。 ^體5己 明。圖2是圖1快閃記憶體中固有的電容以及電阻的說 明。圖3是—快閃記憶體裝置的習知技術讀取條件的說 圖4疋使用圖3的讀取條件 不同節點的設置時_模擬圖。 Μ子I線的 ,5,明根據本發明的_快閃 條件的一實施例。 1 貝取 以及條件,顯示沿-選定字元線 圖二===同雷節點的設置時間的模擬圖。 ' 電路的—實施例的電路圖。 【主要元件符號說明】 1300933 18451twf.doc/y BLO、BL1、…· BL16383 :位元線 WLO、WL1、WL2 …· WL15 :字元線 SS、GS :信號 BK :塊選擇信號 CS :共有源極線 GWLO、GWL1、· . .GWL15 :共滑字元線信號 Ccoup :搞合電容 φ Cpar :寄生電容 WL1B:最接近驅動電晶體$多晶矽字元線WL1的開 始部分 WL1M :指多晶矽字元線WL1的中間部分 蠡 、 WL1E ··指多晶矽字元線WL1的末端部分 WL2B:最接近驅動電晶體的多晶矽字元線WL2的開 始部分 WL2M :指多晶矽字元線WL2的中間部分 φ WL2E :指多晶矽字元線WL2的末端部分 T1 :讀取操作的開始之時間 RWL0〜RWL4 :參考塊字元線 D:最初延遲時段 _ RWL1B :與參考字元線RWL1起點的節點 RWL1M :與參考字元線RWL1中間的節點 RWL1E :與參考字元線RWL1最遠的節點 RWL2B :與參考字元線RWL2起點的節點 19 1300933 18451twf.doc/y RWL2M :與參考字元線RWL2中間的節點 RWL2E ··與參考字元線RWL2最遠的節點 ACT :控制信號 10 :用於產生控制信號ACT的例示性電路• That is, when PUL is high and RST is high, the NMOS transistor in turn transmits signal PUL to latch circuit 24. In an embodiment, the latch circuit 24 includes a pair of inverters 26, 28, the output of which is coupled to the input of the inverter 28, and the output of the inverter 28 is coupled to the inverter 26. Input. The latch circuit 24 is configured to operate and maintain the enable output of the positive edge trigger circuit 14 for subsequent use as described above. After reading, a new command (ie, read, write, or erase command) will be entered. When a new command is entered, the RST is started to reset the act. The above described readout method is most effective in flash memory arrays that utilize only a single driver transistor, which is typically placed at the start node WLB closest to the character 1300933 18451 twf.doc/y line. If a drive transistor is present, the voltage at the node that takes the longest to dissipate the engagement voltage can be prioritized. Typically, this node will be the node farthest from the drive transistor. If = two driver transistors are used, each at each end of the respective word line, then the voltage of the word line will be more stable. However, the use of more than one driver transistor is generally undesirable for size, cost, complexity, and/or power considerations. However, even if the flash memory array utilizes two driver transistors per word line, the method described in this article can be used to further improve setup time. In this embodiment, the intermediate node of the selected word line is preferably used as the input to the comparator 12 of the control signal k-generation circuit 10. Uyr - although the above described read-and-take method has been described in connection with a read protocol, wherein the bias voltage is set to a positive voltage, such as 1·5 V, at the selected word line, and Set it to a higher voltage, such as 6 V, at the selected word line, but the method is not so limited. For example, for a standard floating gate flash memory cell, wherein the read condition is 0 V at the selected word line and 4 ν at the unselected word line, initially The selected word line is biased to a negative voltage until the trigger signal ACT, such as when the reference word line is attenuated to 〇V. Correspondingly, the reference voltage VR of the circuit of FIG. 7 is set. Although the new reading method has been described above with reference to the reading conditions for a particular class of flashing cells, those skilled in the art. It will be appreciated that the method can be generally applied to any of the following flash phase memory cell junctions 17 1300933 18451 twf. doc / y construction, which is not (4) character phase word line coupling, and the towel is selected as 6 words 兀The line has a bias voltage different from the unselected word line bias voltage, although the invention has been described in terms of illustrative embodiments, it is not limited thereto. Rather, the appended claims are to be construed as broadly construed and construed in the claims of the invention BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings illustrate preferred embodiments of the present invention and other aspects of the present disclosure, wherein: FIG. 1 is a circuit diagram of a portion of a fast memory having a plurality of NANDf memory cells. ^体5己明. Figure 2 is a diagram showing the capacitance and resistance inherent in the flash memory of Figure 1. Fig. 3 is a view showing a conventional technique reading condition of a flash memory device. Fig. 4 is a view showing a setting condition of a different node using the reading condition of Fig. 3. An example of the _flash condition according to the present invention is shown in the scorpion I line. 1 Beck and condition, display along the selected character line Figure 2 === simulation diagram of the set time of the same Thunder node. Circuit diagram of the circuit-embodiment. [Description of main component symbols] 1300933 18451twf.doc/y BLO, BL1, ... BL16383: Bit lines WLO, WL1, WL2 ... WL15: Word line SS, GS: Signal BK: Block selection signal CS: Common source Lines GWLO, GWL1, . . . GWL15: Co-sliding word line signal Ccoup: Engaged capacitance φ Cpar: Parasitic capacitance WL1B: The closest starting portion of the driving transistor $polycrystalline word line WL1 WL1M: refers to the polycrystalline word line WL1 The middle portion 蠡, WL1E · refers to the end portion WL2B of the polysilicon character line WL1: the beginning portion WL2M of the polycrystalline word line WL2 closest to the driving transistor: the middle portion φ WL2E of the polycrystalline word line WL2: refers to the polycrystalline word End portion T1 of the element line WL2: time at which the start of the read operation RWL0 to RWL4: reference block word line D: initial delay period _ RWL1B: node RWL1M from the start of the reference word line RWL1: intermediate with the reference word line RWL1 Node RWL1E: node RWL2B farthest from the reference word line RWL1: node 19 with the start of the reference word line RWL2 1300933 18451twf.doc/y RWL2M: node RWL2E and reference character in the middle of the reference word line RWL2 The farthest node of line RWL2 ACT: control signal 10: an exemplary circuit for generating a control signal ACT
Ml : PMOS電晶體 M2 : NMOS電晶體 々Ml : PMOS transistor M2 : NMOS transistor 々
Vcc :電源電壓 RST :重設信號 12 ··比較器 y VR :參考電壓 OUT :比較器12的輸出 /-tA 秦 14 :觸發電路 PUL :觸發電路的輸出 16 :延遲鏈 18 :反相器 20 : NAND 閘 22 :反相器 24 :栓鎖電路 26、28 :反相器 20Vcc : power supply voltage RST : reset signal 12 · comparator y : VR : reference voltage OUT : output of comparator 12 / -tA Qin 14 : trigger circuit PUL : output of trigger circuit 16: delay chain 18 : inverter 20 : NAND gate 22: inverter 24: latch circuit 26, 28: inverter 20
Claims (1)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW95101686A TWI300933B (en) | 2006-01-17 | 2006-01-17 | Method of reducing settling time in flash memories and improved flash memory |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW95101686A TWI300933B (en) | 2006-01-17 | 2006-01-17 | Method of reducing settling time in flash memories and improved flash memory |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| TW200729219A TW200729219A (en) | 2007-08-01 |
| TWI300933B true TWI300933B (en) | 2008-09-11 |
Family
ID=45070094
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| TW95101686A TWI300933B (en) | 2006-01-17 | 2006-01-17 | Method of reducing settling time in flash memories and improved flash memory |
Country Status (1)
| Country | Link |
|---|---|
| TW (1) | TWI300933B (en) |
-
2006
- 2006-01-17 TW TW95101686A patent/TWI300933B/en active
Also Published As
| Publication number | Publication date |
|---|---|
| TW200729219A (en) | 2007-08-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI897341B (en) | semiconductor memory devices | |
| TWI322432B (en) | Efficient verification for coarse/fine programming of non-volatile memory | |
| US9922716B2 (en) | Architecture for CMOS under array | |
| US9953717B2 (en) | NAND structure with tier select gate transistors | |
| US7009881B2 (en) | Semiconductor memory device | |
| CN106531219B (en) | memory device | |
| US10120816B2 (en) | Bad column management with data shuffle in pipeline | |
| US11404122B2 (en) | Sub-block size reduction for 3D non-volatile memory | |
| TW201137882A (en) | Nonvolatile semiconductor memory | |
| US8724391B2 (en) | Semiconductor memory device | |
| US9673798B1 (en) | Digital pulse width detection based duty cycle correction | |
| JPWO2013011600A1 (en) | Method for driving semiconductor memory device | |
| US7649775B2 (en) | Flash memory device applying erase voltage | |
| US9251903B2 (en) | Nonvolatile semiconductor memory device and control method thereof | |
| JP2011054233A (en) | Semiconductor memory device | |
| US7366040B2 (en) | Method of reducing settling time in flash memories and improved flash memory | |
| JP6437421B2 (en) | Nonvolatile semiconductor memory device | |
| TWI300933B (en) | Method of reducing settling time in flash memories and improved flash memory | |
| CN100514496C (en) | Method for erasing flash memory unit and flash memory device using the same | |
| US12272419B2 (en) | Charge pump having switch circuits for blocking leakage current during sudden power-off, and flash memory including the same | |
| US8861281B2 (en) | Method of programming memory and memory apparatus utilizing the method | |
| CN101026007B (en) | Method for reducing flash memory setup time and improved flash memory | |
| CN102314940A (en) | Non-volatile memory device having transistors in parallel with resistance value switching means | |
| TW567496B (en) | Non-volatile memory and operating method thereof |