TWI302741B - Nand type non-volatile memory and manufacturing method and operstion method thereof - Google Patents

Nand type non-volatile memory and manufacturing method and operstion method thereof Download PDF

Info

Publication number
TWI302741B
TWI302741B TW95120523A TW95120523A TWI302741B TW I302741 B TWI302741 B TW I302741B TW 95120523 A TW95120523 A TW 95120523A TW 95120523 A TW95120523 A TW 95120523A TW I302741 B TWI302741 B TW I302741B
Authority
TW
Taiwan
Prior art keywords
voltage
gate
memory
memory cell
layer
Prior art date
Application number
TW95120523A
Other languages
Chinese (zh)
Other versions
TW200746399A (en
Inventor
Chao Wei Kuo
Chih Ming Chao
Hann Ping Hwang
Houng Chi Wei
Pittikoun Saysamone
Original Assignee
Powerchip Semiconductor 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 Powerchip Semiconductor Corp filed Critical Powerchip Semiconductor Corp
Priority to TW95120523A priority Critical patent/TWI302741B/en
Publication of TW200746399A publication Critical patent/TW200746399A/en
Application granted granted Critical
Publication of TWI302741B publication Critical patent/TWI302741B/en

Links

Landscapes

  • Non-Volatile Memory (AREA)
  • Semiconductor Memories (AREA)

Description

1302741 19618twf.doc/g 九、發明說明: 【發明所屬之技術領域】 本發明是有關於-種半導體元件,且特別是有關於一 種反及ffl(NAND)型非揮魏記憶體及其製打法與操作 方法。 ” 【先前技術】 *非揮發性記憶體元件由於具有可進行多次資料存入、 續取、抹除等動作,且存入之資料在斷電後也不會消失之 優點,因此已成為個人電腦和電子設備所廣泛採用的一種 記憶體元件。 典型的非揮發性記憶體元件,一般是被設計成具有堆 ®式閘極(Stacked-Gate)結構,其中包括以摻雜多晶石夕製作 的洋置閘極(Floating Gate)與控制閘極(c〇ntr〇1恤)。浮置 閘極位於控制極和基底之間,且歧浮置㈣,沒有和 任何電路,連接,而控制_職字元線(Wwd Une)相 接,此外還包括穿隨氧化層(TunneHng 〇xide)和問間介電 層(I齡Gate Dielectric Layer)分別位於基底和浮置間極之 間以及浮置閘極和控制閘極之間。 、另彳面’目别業界較常使用的快閃記憶體陣列包括 反或間(NOR)型陣列結構與反及閑(na肋)型陣列結構。由 =反及閘(NAND)型陣列的非揮發性記憶體結構是使各記 串接在—起,其積集度與面積率較反㈣(NOR) ㈣列的非揮發性記憶體佳,已經廣泛地應用在多種電子 產品中。 1302741 19618twf.doc/g 圖1所繪示為習知的反及閘型非揮發性記憶體的結構 剖面圖。 如圖1所示,在基底100上設置有多數個記憶胞M1 〜M8與兩個選擇電晶體ST1、ST2。記憶胞M1〜M8設置 於兩個選擇電晶體ST卜ST2之間。在記憶胞M1〜M8i 間的基底100中、記憶胞Ml與選擇電晶體ST1之間的基 底100中以及記憶胞M8與選擇電晶體ST2之間的基底1〇〇 中形成有摻雜區102。這些記憶胞Ml〜]VI8及選擇電晶體 ST1、ST2經由摻雜區102串接在一起而構成記憶胞行。 在記憶胞行兩侧設置有源極區104與汲極區1〇6。源極線 SL與源極區1〇4電性連接。位元線BL透過插塞1〇8與汲 極區106電性連接。 在上述的NAND型非揮發性記憶體中,由於各記憶胞 Ml〜M8之間藉由摻雜區102連接在一起。在元件尺寸持 續縮小的情況下,記憶胞寬度越來越小,相鄰的捧雜區1〇2 之間會有短通道效應、汲極引發的能帶降低(Drain hduad Barrier Lowering,DIBL)效應等問題,而影響記情體可 靠度。 〜 、此外,對於上述的NAND型非揮發性記憶體而言,在 私式化選定記憶胞時,在同一記憶胞列中的其他非選定記 憶胞都是作為傳輸閘極。因此,在進行程式化時,需 選定記憶胞處於完全敝之狀態。使麵定記憶胞:於* 全開啟狀態的偏壓將會限制記憶胞啟始電壓範圍的嗖二= 舉例來說,當此記憶胞為單階記憶胞時,藉由基準讀取電 1302741 19618twf.doc/g 來判別兩種不同啟始電壓(Vthl 壓 Vref vthi<vrefi<Vth2。Vthl 為小於 〇 伏特;Vrefl 為 〇 伏特左 右’ Itth2為大於0伏特左右。當非選定記憶胞處於完全開 啟狀態的偏壓為5伏特時,則vth2只能設定在〇伏特與5 伏特之严曰 1,而使得Vth2的範圍較小。#此記憶胞為多階記 憶胞時,藉由基準讀取電壓Vi*efl、Vl*ef2、Vref3,來_ 四種不同啟始電壓(VtM、Vth2、νω、_4), νώι<νππ<νώ2<νΓεί2<νώ3<νΓεβ<·4。同樣的,當非 選定記憶胞處於完全開啟狀態的偏壓為5伏特時, 小於〇伏特;Vrefl為〇伏特左右;遍、、靴3 : Vref3、Vth4需設定在ο伏特與5伏特 …特、v㈣為u伏特、vth3為 :為 =為2.4伏特、Vth4為2.8〜3魏1302741 19618twf.doc/g IX. Description of the Invention: [Technical Field] The present invention relates to a semiconductor element, and more particularly to a non-swing (NAND) type non-swing memory and its manufacturing Method and method of operation. [Prior Art] * Non-volatile memory components have become an individual because they have the ability to perform multiple data deposits, continuation, erasure, etc., and the stored data does not disappear after power-off. A memory component widely used in computers and electronic devices. Typical non-volatile memory components are typically designed to have a Stacked-Gate structure, including doped polycrystalline spine The floating gate and the control gate (c〇ntr〇1 shirt). The floating gate is located between the control pole and the base, and the floating (4) is not connected with any circuit, and the control _ The Wwd Une is connected, and the TunneHng 〇xide and the Age Dielectric Layer are located between the substrate and the floating interpole and the floating gate. Between the pole and the control gate. Another flash memory array that is commonly used in the industry includes an anti- or inter-NOR array structure and an anti-free (na rib) array structure. The non-volatile memory structure of the NAND type array is to make each record From the beginning, the non-volatile memory with better integration and area ratio than the (N) (N) column has been widely used in many electronic products. 1302741 19618twf.doc/g Figure 1 is shown as A cross-sectional view of a conventional non-volatile memory structure is shown in Fig. 1. As shown in Fig. 1, a plurality of memory cells M1 to M8 and two selection transistors ST1 and ST2 are disposed on a substrate 100. Memory cells M1~ M8 is disposed between the two selection transistors STb, ST2, in the substrate 100 between the memory cells M1 to M8i, in the substrate 100 between the memory cell M1 and the selection transistor ST1, and between the memory cell M8 and the selection transistor ST2. A doped region 102 is formed in the substrate 1 . The memory cells M1 to VI8 and the selective transistors ST1 and ST2 are connected in series via the doping region 102 to form a memory cell row. The source region 104 and the drain region 1〇6. The source line SL is electrically connected to the source region 1〇4. The bit line BL is electrically connected to the drain region 106 through the plug 1〇8. In the NAND type non-volatile memory, since the memory cells M1 to M8 are connected together by the doping region 102. In the case of continuous shrinking, the memory cell width is getting smaller and smaller, and there are problems such as short channel effect and Drain hduad Barrier Lowering (DIBL) effect between 1捧2 in the adjacent holding area. In addition, for the NAND type non-volatile memory described above, when the selected memory cell is privately selected, other unselected memory cells in the same memory cell are transmitted. Gate. Therefore, when programming, you need to select the memory cell to be completely paralyzed. To make the surface of the memory cell: the bias voltage in the * fully on state will limit the starting voltage range of the memory cell. For example, when the memory cell is a single-order memory cell, the reference reading power 1302741 19618twf .doc/g to discriminate between two different starting voltages (Vthl voltage Vref vthi < vrefi < Vth2. Vthl is less than 〇Vat; Vrefl is around 〇Vt' Itth2 is greater than 0 volts. When unselected memory cells are fully open When the bias voltage is 5 volts, then vth2 can only be set to 〇1 and 5 volts, and the range of Vth2 is small. ## When the memory cell is a multi-order memory cell, the voltage is read by the reference Vi. *efl, Vl*ef2, Vref3, come _ four different starting voltages (VtM, Vth2, νω, _4), νώι<νππ<νώ2<νΓεί2<νώ3<νΓεβ<·4. Similarly, when non-selected memory cells When the bias voltage is 5 volts in the fully open state, it is less than 〇VV; Vrefl is about 〇VV; and the boots 3, Vref3, Vth4 need to be set at οV and 5 volts, especially, v(4) is uV, and vth3 is: ==2.4 volts, Vth4 is 2.8~3 Wei

It、I:的範圍較*。如此,在習知的職D型非揮The range of It, I: is better than *. So, in the conventional job type D non-swing

St確卿:以使程式化記憶胞準確二 口包巧’如此將化費較長的時間。而且,在對 讀取操作時,在同—記憶胞列中的其他非選定記,i i亦谷易對於選定記憶胞造成讀取干擾。 〜 【發明内容】 本發明的目的就是在提供一種nand 憶體;錢造料鱗財法,何 本务明的再一目的是提供—種nand 憶體及其製造方法與操作方法,可容易的與=== 1302741 19618twf.doc/g 在一起,而可以增加製程裕度。 本發明的又一目的是提供一種NAND型非揮發性記 憶體及其製造方法與操作方法,可以加寬記憶胞的設定啟 始電壓範圍,縮短記憶胞的程式化時間。 本發明提出一種反及閘型非揮發性記憶體,此反及閘 型非揮發性記憶體包括多數個記憶單元行。各記憶單元行 包括源極區與汲極區、多數個記憶單元、多數個傳輸閘極、 第一遙擇電晶體與第二選擇電晶體。源極區與汲極區設置 ;基底中夕數個s己憶單元設置於源極區與沒極區之間的 ,,上、’,各個記憶單元包括記憶胞與電晶體,且記憶胞與 电晶,5聯連接在一起。多_傳輸閘極分別設置於相鄰 兩記!!單元之間的基底上,而使記憶單元串聯連接在一 ^第π麵電晶體與第二選擇電晶體分別與最外侧之兩 固記憶^元連接,且分別與源極區與秘區相鄰。 憶 憶 層 依照本發明的較佳實施例所述之反及閘型非揮發性記 ‘ j述傳輪閘極填滿相鄰兩記憶單元之間的間隙。 …、本lx明的較佳實施例所述之反及閘型非揮發性記 雷ίίίΓ些記憶胞由該基底起至少包括穿随介電 ft層1間介電層與控制閘極。 層之材質包括氧化發。貝為減多㈣。上述穿隨介電 本么月的較佳實施例所述之反及閘型非揮發性記 1302741 196l8twf.doc/g 體,更包括夕數條元件隔離結構。多數條元件隔離纟士構 平行設置於基底中,各記憶單元行設置於相鄰兩元件結構 之間。上述元件隔離結構的表面低於電荷儲存層與基底間 之介面而形成凹陷部,且控制閘極填滿上述凹陷部。一 依照本發明的較佳實施例所述之反及閘型非揮發性記 憶體,更包括閘介電層。此閘介電層設置於控制閘^與基 底之間,各電晶體係由控制閘極、閘介電層及該基 依照本發明的較佳實施例所述之反及閘型非揮^發性記 憶體’上述的記憶單元行,呈二維配置,而成記憶胞陣列。 反及閘型轉發性記憶體更包括多數條字元線、多數條位 元線、多數條源極線、多數條選擇閘極線與多數條傳輸閘 極線。多數條字元線在列方向平行排列,且連接同一列之 記憶胞之㈣閘極及f㈣之閘極。多數條位元線在行方 向平行排列,分卿接同—行之記憶單元行的汲極區。多 數條源極線在列方向平行排列,分別連接同—列之記憶單 元行的源極區。多數條選擇閘择線在列方向平行排列,分 別,接同一列之記憶單元行的第一選擇電晶體之閘極與第 二選擇電晶體之閘極。多數條傳輸閘極線在列方向平行排 列’分^連接同—列之記憶單元行的傳輸閘極。 立―依,、?、本發明的較佳實施例所述之反及閘型非揮發性記 更包括多數條元件隔離結構。元件隔離結構設置於 ^氐杜社且於仃方向平行排列,各記憶單元行設置於相鄰 構之間。上述元件隔離結構的表面低於電荷儲存 層人基f間之介面而形成凹陷部,且控制閘極填滿凹陷部。 依…、本务明的較佳實施例所述之反及閘型非揮發性記 !3〇2741 19618twf.doc/g 思—入0付旧"_…百▽ >U/r"丨电置於控制閘極盥基 底之間,各電晶體是由控制閘極、閘介電層與基底所構成。 在上述之反及閘型非揮發性記憶體中,由於元件隔離 結構之表面低於電荷儲存層與基底間之介面,而形成:陷 部,並在此凹陷部設置與記憶胞並聯連接的電晶體。此電 晶體的設置將有助於記憶單元的操作,可以縮短程式化操 作時間,並避免讀取干擾。 ” “,且,在上述之反及閘型非揮發性記憶體中,由於在 纪憶單兀之間設置傳輸閘極,而無須設置摻雜區,因此可 避免短通道效應、汲極引發的能帶降低效應等所造 =漏電流等。此外,於記憶單元之間設置的傳輪間極為 ¥體’可以遮蔽相鄰兩記憶胞的電荷儲存層 記憶胞對記憶胞之間_合干擾。 此別牛低 i另外’在上述之反及閘型非揮發性記憶體中,於 ^之間設置的傳輸閘極。在對本發明之非揮發性_: J行抹除操作時,可使電子從電荷儲存層穿過閘間介:: 減子移除。由於,此種抹除方式4 層之壽命,並二 =靠因此可以提高穿随介電 t發明提出-種反及閘型非揮 去,包括下列步驟。首先, 7衣造方 成有第-介雷Μ Μ先k基底’此基底上已依序形 導㈣、電層、弟—導體層與第二介電層。圖案化第1 苐二2成平行排列的多數個第-條狀導體層,之二 條狀導體層往第-方向延伸。於這些第-條狀導: 1302741 19618twf.doc/g 向延伸的多數條溝渠。接著, 於基底中的溝木内形成多數個隔離結構 表面低:第-條狀導體層與基底間之介面而开 部,。於暴露出的部分基底表面介 電層後,於基底上形成填滿凹陷部 成 第二導體層、第二介電層及第-條狀導體層 其中第二導體層經圖案化之後= 弟-方向《纽平列的多數個第二條狀導體層。之 後,的堆疊閘極結構之間及最外側之兩個堆疊閘極 結構之侧土形成多數個第三條狀導體層。 t立體例所述之反及閉型非揮發性記 ^體^方上條狀導體層_案化後形成多 數個浮置=。上述浮置閘極之材質包括摻雜多晶石夕。 明的較佳實施例所述之反及間型非揮發 上述第—介電層之材質包括氧化石夕。i 述第二¥體層之材質包括摻雜多晶㈣多砂化金屬之1 中之^上述第二介電層包括氧切/氮化石观化石夕, 實施例所述之反及閘型非揮發性記 LI U =方法,於基底中的溝渠内形成隔離結構之步驟 絕緣上形成—層絕緣層’然後移除部分絕 緣層使、、、巴緣層之表面低於基底表面。 氧化製程。 上这弟一,|电層之形成方法包括進行熱 12 1302741 19618twf.doc/g 憶體:施例所述之反及閑型非揮發性記 層之間形成絕緣間^堆豐間極結構與第三條狀導體 憶體議述之反及閘型非揮發性記 構成一電晶體迷弟三介電層與部分第二條狀導體層 憶體列所述之反及閉型非揮發性記 條狀導體居相鄰兩堆疊閑極結構之間的第三 構之側辟^第輪間極,·形成於最外侧之兩堆疊閘極結 X# t L 導體層作為選湖H條狀導體 層之材貝包括摻雜多晶矽。 于版 的較佳實施例所述之反及閑型非揮發性記 •,衣以方法,更包括於基底中形成源極區及汲極區。 離社摄夕Ϊ之非挥發性記憶體之製造方法中,由於元件隔 邻面低於基底表面,而形成凹陷部,並在此凹陷 體電層與部分第二條狀導體層構成的電晶 曰體與記憶胞並聯連接在一起。此電晶體的形成 名s於魏體的操作,可以縮短程式化操作 避 免讀取干擾。 而且,在上述之非揮發性記憶體之製造方法中,由於 在堆疊閑極結構之間形成第三條狀導體層(傳輸閑極),因 ^可避免短通道效應、沒極引發的能帶降低效應等所遠成 的記憶胞漏電流等。 此外,於堆疊閘極結構之間形成第三條狀導體層(傳輸 13 1302741 19618twf.doc/g =^^_層(傳輸_可以遮蔽相鄰 ==的子置閘極,而能夠降低記憶胞對記憶胞之間的 而且’田在上述之反及閘型非揮發性記憶體之 中,於堆豐閘極結構之間开彡#裳- ^ 極)。在嘛恢 ==介電層注入第三條狀導體層置 人J之由於’此種抹除方式可減少電子穿越穿隨 數’因此可以提高穿隨介電層之壽命,並增加 。上述之反及閘型非揮發性記憶體之製造方 法可=的與-般製程整合在一起,而可以增加製程裕度。 本㈣提出—種反及_非揮發性記憶體之操作方 適用於包括多數個記憶單元行的記憶體陣列。各記憶 早以丁設置於基底上,具有:多數個記憶單元,設置於源 ^區與沒極區之間,各記憶單元包括並聯連接在—起的記 憶胞與電晶體;多數個傳輸閘極,設置於記憶單元之間的 基底上,而使§己’丨思單元串聯連接在一起;第一選擇電晶體 f第二選擇電晶體’分別與最外側之兩記憶單元連接曰,曰且 第:選擇電晶體與汲極區相鄰,第二選擇電晶體與源極區 相郯,多數字元線在列方向平行排列,且分別連接同一列 之記憶胞之控制閘極及電晶體之閘極;多數條源極線分別 連接同一列之源極區;多數條位元線在行方向平行排列, 且分別連接同一行之汲極區;多數條第一選擇閘極線,在 列方向平行排列,分別連接同一列之記憶單元行的第一選 14 1302741 擇電晶體之閘極;多數條第一選擇閘極線,在列方向平行 排列,分別連接同一列之記憶單元行的第二選擇電晶體之 閘極;多數條傳輸閘極線,在列方向平行排列,分別連接 同一列之記憶單元行的傳輸閘極。反及閘型非揮發性記情 體之操作方法包括:對選定記憶單元的記憶胞進行程式化 、 操作時,於選定之記憶單元所耦接之位元線施加第一電 壓,於非選定之位元線施加第二電壓,於第一選擇閘極線 _ 施加第三電壓,於選定之記憶單元所耦接之字元線上施加 第四電壓,非選定字元線上施加第五電壓,於所有的傳輪 閘極線施加第六電壓,以利用通道Fowier-Nordheim (F-N) 牙隨效應私式化运疋之$己^思胞,其中第四電壓與第^一電厚 的電壓差可引發F-N穿隧效應,第三電壓大於或等於第一 選擇電晶體的啟始電壓,第二電壓可抑制非選定記憶單元 行的第-選擇電晶體開啟,第五電壓大於或等於該電晶體 的啟始電壓,且第六電壓可使傳輸閘極下方的通道導通。St. Qing: In order to make the stylized memory cell accurate, it will take a long time. Moreover, in the case of a read operation, other non-selected records in the same-memory cell column cause read disturb to the selected memory cell. ~ [Description of the Invention] The object of the present invention is to provide a nand memory; the money-making method, another purpose of the present invention is to provide a kind of nand memory and its manufacturing method and operation method, which can be easily Together with === 1302741 19618twf.doc/g, you can increase the process margin. It is still another object of the present invention to provide a NAND type non-volatile memory, a method of fabricating the same, and a method of operating the same, which can widen the set start voltage range of the memory cell and shorten the stylized time of the memory cell. The present invention provides an anti-gate type non-volatile memory, and the inverse type non-volatile memory includes a plurality of memory cell rows. Each memory cell row includes a source region and a drain region, a plurality of memory cells, a plurality of transmission gates, a first remote selection transistor, and a second selection transistor. The source region and the bungee region are arranged; the plurality of suffix units in the base are disposed between the source region and the non-polar region, and upper, ', each memory unit includes a memory cell and a transistor, and the memory cell and Electro-crystal, 5 connected together. Multiple _ transmission gates are set in adjacent two! On the substrate between the cells, the memory cells are connected in series, and the second π-plane transistor and the second selection transistor are respectively connected to the outermost two solid memories, and are respectively connected to the source region and the secret region. adjacent. Recalling layer The anti-gate type non-volatile memory described in the preferred embodiment of the present invention fills the gap between adjacent memory cells. The anti-gate type non-volatile memory described in the preferred embodiment of the present invention includes at least a dielectric layer and a control gate from the dielectric ft layer. The material of the layer includes oxidized hair. Bay is reduced by more (four). The above-mentioned wear-through dielectric is preferably a reverse-type non-volatile type 1302741 196l8 twf.doc/g body as described in the preferred embodiment of the present invention, and further includes a plurality of element isolation structures. A plurality of element isolation gentleman structures are arranged in parallel in the substrate, and each memory cell row is disposed between adjacent two element structures. The surface of the element isolation structure is lower than the interface between the charge storage layer and the substrate to form a depressed portion, and the control gate fills the recess. An anti-gate type non-volatile memory body according to a preferred embodiment of the present invention further includes a gate dielectric layer. The gate dielectric layer is disposed between the control gate and the substrate, and each of the electro-crystalline systems is controlled by a gate, a gate dielectric layer, and the substrate is reversed and gated according to a preferred embodiment of the present invention. The memory unit of the above-mentioned memory unit is arranged in two dimensions to form a memory cell array. The gate-type forward memory further includes a plurality of word lines, a plurality of bit lines, a plurality of source lines, a plurality of selected gate lines, and a plurality of transmission gate lines. Most of the word lines are arranged in parallel in the column direction, and are connected to the (4) gate of the memory cell of the same column and the gate of f (4). Most of the bit lines are arranged in parallel in the row direction, and the branching line is connected to the bungee area of the row of memory cells. A plurality of source lines are arranged in parallel in the column direction, and are respectively connected to the source regions of the memory cell rows of the same column. The plurality of selected gate lines are arranged in parallel in the column direction, and the gates of the first selected transistor and the gate of the second selected transistor are connected to the memory cell row of the same column. A plurality of transmission gate lines are arranged in parallel in the column direction to divide the transmission gates of the memory cell rows of the same column. The anti-gate type non-volatile memory described in the preferred embodiment of the present invention includes a plurality of element isolation structures. The component isolation structures are disposed in the 氐Dushe and are arranged in parallel in the 仃 direction, and the memory cell rows are disposed between the adjacent structures. The surface of the element isolation structure is lower than the interface between the person bases f of the charge storage layer to form a depressed portion, and the control gate fills the depressed portion. According to the preferred embodiment of the present invention, the anti-valve type non-volatile record! 3〇2741 19618twf.doc/g think - enter 0 pay old "_... hundred ▽ > U / r " The electricity is placed between the control gates and the substrate, and each transistor is composed of a control gate, a gate dielectric layer and a substrate. In the above-mentioned anti-gate type non-volatile memory, since the surface of the element isolation structure is lower than the interface between the charge storage layer and the substrate, a trap portion is formed, and electricity connected in parallel with the memory cell is disposed in the recess portion. Crystal. The setting of this transistor will help the operation of the memory unit, shorten the stylized operation time and avoid reading interference. ", and, in the above-mentioned anti-gate type non-volatile memory, since the transmission gate is provided between the records, and the doping region is not required, the short channel effect and the bungee are avoided. The energy can be reduced by the effect, etc. = leakage current. In addition, the volume between the transfer units disposed between the memory units can shield the charge storage layer of the adjacent two memory cells from interfering with the memory cells. This is not the same as the transmission gate set between ^ and the gate type non-volatile memory. In the non-volatile _: J line erase operation of the present invention, electrons can be removed from the charge storage layer through the inter-gate:: subtractor. Because of this, the life of the 4 layers of the erase method, and the second = can therefore improve the wear and tear dielectric t - the invention and the non-swipe, including the following steps. First, the 7-layer fabric has a first-channel Μ Μ k k-substrate'. The substrate has been sequentially guided (4), the electric layer, the di-conductor layer and the second dielectric layer. A plurality of strip-shaped conductor layers arranged in parallel in the first and second parallel patterns are patterned, and the two strip-shaped conductor layers extend in the first direction. In these first-strip guides: 1302741 19618twf.doc/g extends to a number of ditches. Then, a plurality of isolation structures are formed in the trenches in the substrate, and the surface is low: the interface between the first strip conductor layer and the substrate is opened. After exposing a portion of the surface dielectric layer of the substrate, a filled recess is formed on the substrate to form a second conductor layer, a second dielectric layer, and a strip-shaped conductor layer, wherein the second conductor layer is patterned. Directions "Many second strip conductor layers of Newpin. Thereafter, a plurality of third strip conductor layers are formed on the side soils between the stacked gate structures and the outermost two stacked gate structures. The inverse and closed type non-volatile memory described in the t-dimensional example forms a plurality of floating = after the strip-shaped conductor layer. The material of the floating gate includes doped polysilicon. The anti- and non-volatile non-volatile materials described in the preferred embodiment include the oxidized stone. i The material of the second body layer comprises doped polycrystalline (tetra) multi-sanded metal 1 of the above second dielectric layer including oxygen cut/nitrite stone, the anti-gate type non-volatile as described in the embodiment The character LI U = method, the step of forming an isolation structure in the trench in the substrate, forming an insulating layer on the insulating layer and then removing a portion of the insulating layer so that the surface of the edge layer is lower than the surface of the substrate. Oxidation process. On the first brother, the formation method of the electric layer includes heat 12 1302741 19618twf.doc/g memory: the anti-free and non-volatile layer described in the example forms an inter-insulation between the non-volatile layers and The third strip conductor recalls the opposite and the gate type non-volatile record constitutes an anti-and closed-type non-volatile record as described in the three-dielectric layer of the transistor and the second strip-shaped conductor layer. The strip conductor is located on the side of the third structure between the adjacent two stacked idler structures, and the second stack gate junction X# t L conductor layer is formed as the selected lake H strip conductor layer The material shell includes doped polysilicon. The anti-free and non-volatile recording method described in the preferred embodiment of the invention further comprises forming a source region and a drain region in the substrate. In the method for manufacturing a non-volatile memory, the recessed portion is formed because the element adjacent surface is lower than the surface of the substrate, and the electric layer formed by the recessed electric layer and a portion of the second strip-shaped conductor layer The crystal body is connected in parallel with the memory cell. The formation of this transistor is the operation of the Wei body, which can shorten the stylized operation and avoid reading interference. Moreover, in the above-described method for manufacturing a non-volatile memory, since a third strip conductor layer (transmission idler) is formed between the stacked idler structures, the short channel effect and the infinitely induced energy band can be avoided. Reduce the memory leakage current and so on. In addition, a third strip conductor layer is formed between the stacked gate structures (transmission 13 1302741 19618twf.doc/g = ^^_ layer (transmission_ can shield adjacent sub-gates of ==, and can reduce memory cells) Between the memory cells and the 'field in the above-mentioned reverse-type non-volatile memory, between the stacking gate structure is open #彡-^ pole. In the recovery == dielectric layer injection The third strip of conductor layer is placed in the J because of 'this erasing method can reduce the electron crossing the number of wear and tear', so the life of the dielectric layer can be increased and increased. The above-mentioned anti-gate type non-volatile memory The manufacturing method can be integrated with the general process, and can increase the process margin. This (4) proposes that the operation of the non-volatile memory is applicable to a memory array including a plurality of memory cell rows. Each memory is disposed on the substrate as long as it has a plurality of memory cells disposed between the source region and the non-polar region, and each of the memory cells includes a memory cell and a transistor connected in parallel; and a plurality of transmission gates , set on the substrate between the memory cells, so that The 丨 单元 cells are connected in series; the first selection transistor f is the second selection transistor ′ respectively connected to the outermost two memory cells, and the first: the selection transistor is adjacent to the drain region, and the second selection transistor Contrary to the source region, the plurality of digital elements are arranged in parallel in the column direction, and are respectively connected to the control gates of the memory cells of the same column and the gates of the transistors; the plurality of source lines are respectively connected to the source regions of the same column; The plurality of bit lines are arranged in parallel in the row direction and are respectively connected to the drain regions of the same row; the plurality of first selected gate lines are arranged in parallel in the column direction, respectively connecting the first row of the memory cells of the same column 14 1302741 The gate of the electrification crystal; the plurality of first selection gate lines are arranged in parallel in the column direction, respectively connected to the gates of the second selection transistor of the memory cell row of the same column; the plurality of transmission gate lines are in the column direction Parallel arrangement, respectively connecting the transmission gates of the memory cells of the same column. The operation method of the non-volatile grammatical body of the gate type includes: programming and operating the memory cells of the selected memory unit Applying a first voltage to a bit line coupled to the selected memory cell, applying a second voltage to the unselected bit line, applying a third voltage to the first selected gate line _, coupled to the selected memory cell A fourth voltage is applied to the word line, a fifth voltage is applied to the unselected word line, and a sixth voltage is applied to all of the pass gate lines to utilize the channel Fowier-Nordheim (FN) tooth to effect the private operation. The voltage of the fourth voltage and the first electrical thickness can induce a FN tunneling effect, the third voltage is greater than or equal to the starting voltage of the first selected transistor, and the second voltage can suppress the non-selected memory. The first-select transistor of the cell row is turned on, the fifth voltage is greater than or equal to the start voltage of the transistor, and the sixth voltage is such that the channel under the transfer gate is turned on.

依照本發明的較佳實施例所述之反及閘型非揮發性記 憶體之操作方法, 時,於選定之記II 於選定之記憶单元所知& ^ ^ ^ 對遠定記憶單元的記憶胞進行讀取操作In accordance with a preferred embodiment of the present invention, the operation method of the non-volatile memory of the gate type is selected, and the memory of the remote memory unit is selected and selected by the selected memory unit. Cell read operation

疋之义fe單元所耦接之字元線上施加第十電 1302741 19618twf.doc/g ,非运疋予元線上施加第^ 電壓,於所有的傳輪閘 線施加第十二電壓,以讀取選定之記憶胞,其中第八電^ 大於或等於第一選擇電晶體的啟始電壓,第九電壓大於^ . 等於第二選擇電晶體的啟始電壓,第十一電壓大於或等二 電晶體的啟始電壓,且第十二電壓可使傳輸閘極下方/ 道導通。 3通 依照本發明的較佳實施例所述之反及閘型非揮發性— • 憶體之操作方法,第七電壓為1.5伏特左右;第八電壓^ 5伏特左右;第九電壓為5伏特左右;第十電壓為〇伏特 左右;第十一電壓為5伏特左右;第十二電壓為5伏特^ 右。 二 依照本發明的較佳實施例所述之反及閘型非揮發性纪 fe體之刼作方法,對記憶單元的記憶胞進行抹除操作時, 於所有的傳輸閘極線施加第十三電壓,使基底浮置,以利 用F-N穿隧效應抹除該些記憶胞,其中第十三電壓與基 的電壓差可引發F_N穿隧效應。 〃 土一 眷 依照本發明的較佳實施例所述之反及閘型非揮發性記 憶體之操作方法,第十三電壓為15伏特左右。 在上述之反及閘型非揮發性記憶體之製造方法中,由 於各個記憶單元分別是由並聯設置的一個電晶體與一個記 ’ 憶胞所構成,因此即使未選定字元線上施加的電壓無法打 • 開記憶胞的通道,但是只要此電壓可打開電晶體的通道, 即可使電流通過,並到達選定記憶單元。於是,在進行程 式化操作時,由於可藉由電晶體導通電流,因此記憶胞啟 16 1302741 19618twf.doc/g 始電廢範圍的設定不t受到使非選定記憶胞處於完全開啟 狀態的偏壓的限制,使得記憶胞的啟始電壓範圍較廣,因 此可以減少程式化的確認次數與步驟,而可以縮短程式化 操作的時間。 而且在17買取运疋S己憶胞時,並不會受到共用同一條 位兀線之記憶胞的干擾。於記憶單元之間設置傳輸閘極, 可避免短通道效應、汲極引發的能帶降低效應等所造成的 記憶胞漏電流等。 此外,上述之反及閘型非揮發性記憶體之製造方法 中,利用通道F-N穿隧效應(F-N Tunneling)使電子經由通 道穿過穿隧介電層注入電荷儲存層中,以進行記憶胞之程 式化操作;並利用F-N穿隧效應使電子從電荷儲存層穿過 閘間介電層注入傳輸閘極中,以進行記憶胞之抹除操作。 由於’此種操作方式減少了電子穿越穿隧介電層之次數, 因此可以提高穿隧介電層之壽命,並增加元件的可靠度。 為讓本發明之上述和其他目的、特徵和優點能更明顯 易懂,下文特舉較佳實施例,並配合所附圖式,作詳細說 明如下。 '' σ 【實施方式】 圖2為繪示一種本發明之NAND(反及閘)型非揮發性 記憶體陣列之電路簡圖。在本實施例中係以8個記憶單元 為一行、且共三行之NAND記憶單元行為例做說明。 請參照圖2,NAND(反及閘)型非揮發性記憶體陣列包 括多數個選擇電晶體ST11〜ST31與ST12〜ST32、多數個 17 1302741 19618twf.doc/g 記憶單元Qll〜Q38、多數條字元線WL1〜WL8、選擇閘 極線SG1與SG2、位元線BL1〜BL3與傳輸閘極線(pass gate line)PLl 〜PL7 〇 • 記憶單元Q11〜Q18串聯連接於選擇電晶體stii與選 , 擇電晶體ST12之間,而在行之方向形成記憶單元行 MR卜記憶胞Q21〜Q28串聯連接於選擇電晶體阳丨與 選擇電晶體ST22之間,而在行之方向形成記憶單元;于 • MR2。記憶胞Q31〜Q38串聯連接於選擇電晶體ST31與 選擇電晶體ST32之間,而在行之方向形成記憶單元^ MR3 〇 吕己憶單元Q11〜Q38分別是由記憶胞Mn〜M38與電 晶體T11〜T38所構成,記憶胞M11〜M38與電晶體τη 〜T38並聯連接在一起。舉例來說,記憶單元Qn是由記 怳胞Mil與電晶體T11所構成,且記憶胞M11與電晶體 τιι並聯連接在一起;記憶單元Q12是由記憶胞組2與電 日日日體T12所構成,且記憶胞M12與電晶體T12並聯連接 在一起;…;依此類推,記憶單元Q38是由記憶胞Μ38 與電晶體Τ38所構成,且記憶胞顧與電晶體Τ3 連接在一起。 多數字元線WL1〜WL8在列方向平行排列,且連接 同一列之纪憶胞之控制閘極與電晶體之閘極。舉例來說, • 記憶胞M11〜Μ31之控制閘極與電晶體T11〜T31之閘極 則耦接至所對應之字元線WL1;記憶胞M12〜M32之^制 閘極與電晶體T12〜T32之閘極則耦接至所對應之字元線 18 1302741 19618twf.doc/g WL2 ·. ’依此類推,記憶胞龍〜M38之控制間極 晶體T18〜T38之閘極則耦接至所對應之字元線w y 選擇電晶體 ST11〜ST31 ^ ,綱晶體一如::;^ BL1〜犯。選擇電晶體ST12〜ST32之閘極物技= 閉極線SG2。選擇電晶體ST12〜ST32之源 接 極線SL。在同:行之相鄰兩記憶單元之間設置 極,亦即在ΐ憶皁7° Q11〜Q18彼此之間分別形成有二二 問極」記,單元Q21〜Q28彼此之間分別形成有傳幹二 極’記憶單7G Q31〜Q38彼此之間分別形成有傳門= 多數傳輸閘極線PL1〜PL7在列方向平行排列,且二 列之傳輸閘極。亦即,記憶單元Q11〜Q31與記憶⑽ 〜Q32之間_輸閘極姑至所對應之傳輸閉 記憶單元⑽〜Q32與記憶衫Q13〜Q33之間的=’ 極減至所對應之傳輸卩雜線PL2 ;依此錄 :: QH〜Q输憶單元Q18〜Q38之間的傳輪閑極:: 所對應之傳輸閘極線PL7。 馬接至 一在上述實施例中,係以使八個記憶單元串接在 貫例做說明。當然,在本發明中串接的記憶單 了 可以視實際需要串接適當的數目,舉例來說,同 ^ 線可以串接32至64個記憶單元。 /卞70 圖3A所緣示為對記憶胞進行程式化操作之 不意圖。圖3B所繪示為記憶胞進行讀取操作之—實 不意圖。@ 3C所繪示為對所有記憶胞進行抹除摔作之一 19 1302741 19618twf.doc/g 實例的示意圖。 接著,說明本發明之NAND(反及閘)型非揮發性記憶 W歹⑷_模式,其係包括程式化、抹除與資料讀取等 刼作模^。就本發明之非揮發性記憶體之操作方法而言, 以下僅提供一較佳實施例作為說明。但本發明之非揮發性 、脰的知作方法,並不限定於這些方法。在下述說明中 係以圖2所f之記憶單元Q22為實例做說明。 Γ It麥照圖2及圖3A,當對選定記憶單元行MR2 中的記憶單元Q22進行程式化操作時,於選定之位元線 BL2知加電壓Vpl。於非選定之位元線BL1、BL3施加電 壓Vp2。於選擇閘極線SG1施加電壓Vp3。於選擇閘極線 SG2施加電壓Vp4。於選定之記憶單$⑽所減之字元 線yL2上^加電壓Vp5。非選定字元線wli、〜 士鉍加电壓Vp6。於所有的傳輸閘極線pu〜pL7施加電 壓Vp7’以利用通道F-N穿隧效應程式化選定記憶單元 Q22之記憶胞M22。 ★、,由於電壓Vp5與電壓Vpl的電壓差需足以引發F_N 牙隧效應’因此電| Vp5與電壓Vpl白勺電壓差需為12〜 20伏特左右。在本實例中,電壓Vp5例如是2〇伏特左右, 電壓Vpl例如是〇伏特左右。 由於選擇電晶體ST21需處於開啟狀態,因此電壓Vp3 而^於或等於逵擇電晶體s丁21的啟始電壓。在本實例中, 電壓Vp3例如疋5伏特左右。由於,選擇電晶體ST22需 處於關閉狀怨,因此電壓Vp4需小於選擇電晶體ST22的 20 1302741 19618twf.doc/g 啟始電壓。在本實例中,電壓Vp4例如是〇伏特左右。 而且,為了避免共用字元線WL2的其他非選定記憶 單元Q12〜Q32受到程式化干擾,也可以在其他非選定位 • 元線施加電壓VP2。電壓Vp2需抑制非選定記憶單元行 MR1、MR3的選擇電晶體STn、ST31開啟,因此電壓 • VP2需大於等於選擇電晶體ST1卜ST31的啟始電壓。$ 本實例中,電壓VP2例如是5伏特左右。當然,亦^以使 # 電壓Vp5與電壓VP2的電壓差不足以引發F-N穿隧效應, 此時電壓Vp2例如是1〇伏特左右。 由於品要使έ己憶單元行MR2中的其他非選定的記憶 單元Q2卜Q23〜Q28(包括記憶胞Μ2][、Μ23〜Μ28或電 晶體Τ21、Τ23〜Τ28)的通道都為開啟狀態(非選定的記憶 單元Q21、Q23〜Q28皆作為通過閘)。因此,電壓Vp6至 J需大於或等於電晶體T11〜T38的啟始電壓,甚至大於 或等於記憶胞M21、M23〜M28的啟始電壓。在本實例中, • 電壓VP6例如是10伏特左右。電壓Vp7可使傳輸閘極下 方的通道導通。在本實例中,電壓Vp7例如是5伏特左右。 在上述偏壓情況下,即可在選定記憶胞M22之浮置閘 極與基底之間建立一個大的電場,而得以利用通道F-N穿 随效應(Channel F-N Tunneling)使電子由通道注入電荷儲 存層中。The tenth electric 1302741 19618twf.doc/g is applied to the word line coupled to the fe unit, and the second voltage is applied to the non-transmission line, and the twelfth voltage is applied to all the transmission lines to read a selected memory cell, wherein the eighth voltage is greater than or equal to a starting voltage of the first selected transistor, the ninth voltage is greater than ^. equal to the starting voltage of the second selected transistor, and the eleventh voltage is greater than or equal to the second transistor The starting voltage, and the twelfth voltage can make the channel below the transmission gate/channel. In accordance with a preferred embodiment of the present invention, the anti-gate type non-volatile - • memory method of operation, the seventh voltage is about 1.5 volts; the eighth voltage is about 5 volts; the ninth voltage is 5 volts. Left and right; the tenth voltage is about volts; the eleventh voltage is about 5 volts; the twelfth voltage is 5 volts ^ right. According to a preferred embodiment of the preferred embodiment of the present invention, a method for fabricating a non-volatile gate type is applied to a memory cell of a memory cell, and a thirteenth is applied to all of the transmission gate lines. The voltage causes the substrate to float to erase the memory cells by the FN tunneling effect, wherein the voltage difference between the thirteenth voltage and the base can induce an F_N tunneling effect. According to a preferred embodiment of the present invention, the thirteenth voltage is about 15 volts in accordance with the operation method of the gate type non-volatile memory. In the above-described method for manufacturing a gate-type non-volatile memory, since each memory cell is composed of one transistor arranged in parallel and one memory cell, even if the voltage applied to the word line is not selected, • Open the channel of the memory cell, but as long as this voltage can open the channel of the transistor, the current can pass and reach the selected memory unit. Therefore, during the stylization operation, since the current can be turned on by the transistor, the setting of the memory cell start 16 1302741 19618 twf.doc/g is not biased by the non-selected memory cell in the fully-on state. The limitation of the memory cell is that the starting voltage range of the memory cell is wide, so that the number of steps and steps of the stylization can be reduced, and the time for the stylized operation can be shortened. Moreover, when 17 is purchased, it is not interfered by the memory cells sharing the same line. The transmission gate is arranged between the memory cells to avoid the memory leakage current caused by the short channel effect and the band-reduction effect caused by the drain. In addition, in the above-described method for manufacturing a gate-type non-volatile memory, channel FN tunneling is used to cause electrons to be injected into the charge storage layer through the tunnel through the tunnel dielectric layer to perform memory cell. Stylized operation; and FN tunneling effect is used to inject electrons from the charge storage layer through the inter-gate dielectric layer into the transfer gate for memory cell erase operation. Since this mode of operation reduces the number of times electrons pass through the tunneling dielectric layer, the lifetime of the tunneling dielectric layer can be increased and the reliability of the component can be increased. The above and other objects, features and advantages of the present invention will become more <RTIgt; '' σ [Embodiment] FIG. 2 is a circuit diagram showing a NAND (anti-gate) type non-volatile memory array of the present invention. In the present embodiment, a description will be given of an example of the behavior of a NAND memory cell in which eight memory cells are one line and three lines in total. Referring to FIG. 2, the NAND (non-gate) type non-volatile memory array includes a plurality of selection transistors ST11 to ST31 and ST12 to ST32, a plurality of 17 1302741 19618 twf.doc/g memory cells Q11 to Q38, and a plurality of words. The source lines WL1 WL WL8, the selection gate lines SG1 and SG2, the bit lines BL1 BLBL3 and the pass gate lines PL1 ~ PL7 〇 • the memory cells Q11 ~ Q18 are connected in series to the selection transistor stii and selected. Selecting between the transistors ST12 and forming the memory cells in the row direction MR memory cells Q21~Q28 are connected in series between the selective transistor anode and the selection transistor ST22, and form a memory cell in the row direction; MR2. The memory cells Q31~Q38 are connected in series between the selection transistor ST31 and the selection transistor ST32, and form a memory cell in the direction of the line. MR3 〇 Lu yiyi units Q11~Q38 are respectively composed of memory cells Mn~M38 and transistor T11 The structure of ~T38, the memory cells M11~M38 are connected in parallel with the transistors τη~T38. For example, the memory cell Qn is composed of the cell M and the transistor T11, and the memory cell M11 and the transistor τιι are connected in parallel; the memory cell Q12 is composed of the memory cell group 2 and the electric day and day body T12. And the memory cell M12 is connected in parallel with the transistor T12; and so on, the memory cell Q38 is composed of the memory cell 38 and the transistor Τ38, and the memory cell is connected to the transistor Τ3. The plurality of digital element lines WL1 WL WL8 are arranged in parallel in the column direction, and are connected to the gate of the control column of the same column and the gate of the transistor. For example, the control gates of the memory cells M11~Μ31 and the gates of the transistors T11~T31 are coupled to the corresponding word line WL1; the gates of the memory cells M12~M32 and the transistor T12~ The gate of T32 is coupled to the corresponding word line 18 1302741 19618twf.doc/g WL2 ·. 'And so on, the gate of the control cell crystal T18~T38 of the memory cell ~ M38 is coupled to the gate Corresponding character line wy selects the transistor ST11~ST31^, and the crystal is as follows: :;^ BL1~. Select the gate technique of the transistors ST12 to ST32 = the closed line SG2. The source line SL of the transistors ST12 to ST32 is selected. In the same line: the poles are arranged between the adjacent two memory cells, that is, the two-two-question poles are formed between the memory cells 7° Q11 and Q18, and the cells Q21 to Q28 are respectively formed with each other. The dry two-pole 'memory single 7G Q31 to Q38 are respectively formed with gates = the majority of the transmission gate lines PL1 to PL7 are arranged in parallel in the column direction, and the transmission gates of the two columns are arranged. That is, between the memory units Q11 to Q31 and the memory (10) to Q32, the transmission between the closed memory units (10) to Q32 and the memory sheets Q13 to Q33 is reduced to the corresponding transmission. Miscellaneous line PL2; according to this record:: QH~Q transmission unit Q18~Q38 between the transmission idle pole:: corresponding transmission gate line PL7. In the above embodiment, the eight memory cells are connected in series to illustrate. Of course, in the present invention, the serially connected memory sheets can be connected in an appropriate number according to actual needs. For example, the same line can be connected in series with 32 to 64 memory units. / 卞 70 Figure 3A shows the intention of stylizing memory cells. Figure 3B illustrates the memory cell for a read operation - not intended. @ 3C is a schematic diagram of an example of the erasure of all memory cells 19 1302741 19618twf.doc/g. Next, the NAND (reverse and gate) type non-volatile memory W歹(4)_ mode of the present invention will be described, which includes stylization, erasing, and data reading. In the case of the method of operation of the non-volatile memory of the present invention, only a preferred embodiment is provided below as an illustration. However, the method for knowing non-volatile and antimony of the present invention is not limited to these methods. In the following description, the memory unit Q22 of Fig. 2 will be described as an example. Γ It Photo 2 and FIG. 3A, when the memory cell Q22 in the selected memory cell row MR2 is programmed, the voltage Vpl is applied to the selected bit line BL2. A voltage Vp2 is applied to the unselected bit lines BL1, BL3. A voltage Vp3 is applied to the selection gate line SG1. A voltage Vp4 is applied to the selection gate line SG2. The voltage Vp5 is applied to the word line yL2 subtracted from the selected memory list $(10). The unselected word line wli, ~ 士 铋 voltage Vp6. A voltage Vp7' is applied to all of the transmission gate lines pu to pL7 to program the memory cell M22 of the selected memory cell Q22 by the channel F-N tunneling effect. ★,, because the voltage difference between voltage Vp5 and voltage Vpl needs to be sufficient to induce F_N tunneling effect, so the voltage difference between Vp5 and voltage Vpl needs to be about 12~20 volts. In the present example, the voltage Vp5 is, for example, about 2 volts, and the voltage Vpl is, for example, about volts. Since the selection transistor ST21 needs to be in an on state, the voltage Vp3 is equal to or equal to the starting voltage of the selection transistor 21. In the present example, the voltage Vp3 is, for example, about 5 volts. Since the selection transistor ST22 needs to be turned off, the voltage Vp4 needs to be smaller than the 20 1302741 19618 twf.doc/g starting voltage of the selection transistor ST22. In the present example, the voltage Vp4 is, for example, about 〇Vot. Moreover, in order to avoid stylized interference of other unselected memory cells Q12 to Q32 sharing the word line WL2, voltage VP2 may be applied to other non-selected positioning cells. The voltage Vp2 needs to suppress the non-selected memory cell rows MR1, MR3 select transistors STn, ST31 are turned on, so the voltage • VP2 needs to be greater than or equal to the start voltage of the selected transistor ST1 and ST31. In this example, the voltage VP2 is, for example, about 5 volts. Of course, the voltage difference between the # voltage Vp5 and the voltage VP2 is not sufficient to induce the F-N tunneling effect, and the voltage Vp2 is, for example, about 1 volt. Because the product is to make the channel of other non-selected memory cells Q2 Q23~Q28 (including memory cell 2] [, Μ23~Μ28 or transistor Τ21, Τ23~Τ28) in the cell line MR2 is turned on ( The unselected memory cells Q21, Q23~Q28 are all used as pass gates. Therefore, the voltages Vp6 to J need to be greater than or equal to the starting voltages of the transistors T11 to T38, and even greater than or equal to the starting voltages of the memory cells M21, M23 to M28. In this example, • the voltage VP6 is, for example, about 10 volts. Voltage Vp7 turns on the channel below the transmission gate. In the present example, the voltage Vp7 is, for example, about 5 volts. Under the above bias condition, a large electric field can be established between the floating gate of the selected memory cell M22 and the substrate, and the channel FN tunneling can be used to inject electrons from the channel into the charge storage layer. in.

Y 在進行上述程式化操作時,共用同一條字元線WL2 之記憶單元Q12、Q32並不會程式化。這是因為未選定位 元線BL1、BL3上施加5伏特之電壓,故選擇電晶體ST11、 21 1302741 19618twf.doc/g ST3^ 二I 隧現象,當然就不會程式化記憶胞 Q12、Q32 〇 而且,由於未選定字元線WL1、WL3〜㈣上施加 10伏特之電壓’此電壓只是用於打開記憶單元之通道,而 不足以引發通道F-N雜現象,因此非選定字元線、 WL3〜WL8所連接的記憶胞qU〜⑼、即 % 4Y When the above stylized operation is performed, the memory cells Q12 and Q32 sharing the same word line WL2 are not programmed. This is because the voltage of 5 volts is applied to the unselected locating elements BL1 and BL3, so the transistor ST11, 21 1302741 19618 twf.doc/g ST3^I tunneling phenomenon is selected, and of course, the memory cells Q12 and Q32 are not programmed. Moreover, since a voltage of 10 volts is applied to the unselected word lines WL1, WL3 to (4) 'this voltage is only used to open the channel of the memory cell, and is not sufficient to cause the channel FN phenomenon, the unselected word line, WL3 WL WL8 Connected memory cells qU~(9), ie % 4

〜Q38不會被程式化。 Q~Q38 will not be stylized. Q

此外,由於本發明之各個記憶單元Qn〜Q38分別是 由並聯設置的-個電晶體T11〜T38與—個記憶胞Mu〜 M38所構成,而電晶體T11〜T38的啟始電壓低於記憶胞 Mil〜Μ38的啟始電壓,因此即使未選定字元線WL1、 WL3〜WL8上施加的電壓無法打開記憶胞的通道,但是只 要此電壓可打開電晶體的通道,即可使電流通過,並到達 選定記憶單元。於是,在程式化操作時,由於可藉由電晶 體T11〜T38導通電流,因此,記憶胞啟始電壓範圍的設 定不會受到使非選定記憶胞處於完全開啟狀態的偏壓的限 制,使得記憶胞Mil〜M38的啟始電壓範圍較廣,因此可 以減少程式化的確認次數與步驟,而可以縮短程式化操作 的時間。 舉例來說,當此記憶胞為單階記憶胞時,藉由基準讀 取電壓Vref,來判別兩種不同啟始電壓(Threshold Voltage)(Vthl、Vth2),Vthl&lt;Vrefl&lt;Vth2。Vthl 為小於 0 伏特;Vrefl為〇伏特左右;Vth2為大於〇伏特左右。當 22 1302741 19618twf.doc/g 非選定記憶胞處於完全開啟狀態的偏壓為5伏特時,V也2 也可以設定為大於5伏特,而使得Vth2的範陳廣。舍此 記憶胞為多階記憶胞時,藉由基準讀取電 V-,來判別四種…始電二: V〇ltage)(Vthl 、Vth2 、Vth3 、VtM), Vthl&lt;Vrefl&lt;Vth2&lt;Vref2&lt;Vth3&lt;Vref3&lt;Vth4。同樣的,當非 選定記憶胞處於完全開啟狀態的偏壓為5伏 7 小於〇伏f為〇伏特左右;购為〇^2伏特為 Vref2為2.2伏特、Vth3為2.4〜4 2伏特、Vref3為4 4伏 特、Vth4為4.6〜6.4伏特;而使得vth2、Vth3、m 範圍較大。 时一而且在上述說明中,雖係以記憶元件陣列中單一記憶 早凡為單位進行程式化U本發明之NAND(反及閘)型 非揮發性記題_之程式化切藉由各字元線、選擇問 極線、位TL線的控制,而以位元組、節區,或是區塊為單 位進行程式化。 ^同時麥照圖2及圖3B,當對選定記憶單元行MR2 的。己匕、單元Q22進行讀取操作日夺,於選定之位元線 紅加私壓Vrl。於非選定之位元、線BU、BL3施加電壓 。於廷擇閘極、線SG1施加電壓Vr3。於選擇閘極線SG2 :力电[Vr4。於選定之記憶單a⑽所弟馬接之字元線 上知加電M Vr5。非選定字元線WL1、WL3〜WL8 ^加包壓Vr6 °於所有的傳輸閘極線PL1〜PL7施加電 【Vl7 ’以碩取選定記憶單元Q22之記憶胞M22。 23 1302741 19618twf.d〇c/g 電壓Vrl為施加於選定位元線BL2的讀取偏壓。在本 實例中,電壓Vrl例如是L5伏特左右。電壓Vr2則例如 是〇伏特左右。 由於選擇電晶體ST21及選擇電晶體ST22需處於開啟 狀態,因此電壓Vr3及電壓Vr4需大於或等於選擇電晶體 ST21及選擇電晶體ST22的啟始電壓。在本實例中,電壓 Vr3及電壓vr4例如是5伏特左右。 由於需要使記憶單元行MR2中的其他非選定的記憶 單元Q2卜Q23〜Q28(包括記憶胞M21、M23〜M28或電 晶體T21、T23〜T28)的通道都為開啟狀態(非選定的記憶 單元Q21、Q23〜Q28皆作為通過閘)。因此,電壓vr6需 大於或等於電晶體T11〜丁38的啟始電壓。在本實例中, 電壓Vr6例如是5伏特左右。電壓Vr7可使傳輸閘極下方 的通道導通。在本實例中,電壓Vr7例如是5伏特左右。 在上述偏壓情況下,可藉由偵測記憶胞之通道電流大 小來判斷儲存於此記憶胞中的數位資訊。 而且,由於本發明之各個記憶單元Q11〜Q38分別是 由並聯設置的一個電晶體T11〜T38與一個記憶胞Mil〜 M38所構成,而電晶體T11〜T38的啟始電壓低於記憶胞 Mil〜M38的啟始電壓,因此即使未選定字元線WL1、 WL3〜WL8上施加的電壓無法打開記憶胞的通道,但是只 要此電壓可打開電晶體的通道,即可使電流通過。因此, 在讀取選定記憶單元Q22時,並不會受到共用同一條位元 線BL2之記憶胞Q21、Q23〜Q28的干擾。 24 1302741 19618twf.doc/g 此外,在本發明之記憶單元之間,以傳輸閘極取代習 知的摻雜區,因此可避免短通道效應、汲極引發的能帶降 低(Drain Induced· Barrier Lowering,DIBL)效應等所造成的 記憶胞漏電流等。 而且在上述說明中,雖係以記憶元件陣列中單一記憶 元件為單位進行讀取操作,然而本發明之NAND(反及閘) 型快閃記憶胞陣列之讀取操作也可藉由各字元線、選擇閘 極線、位元線的控制,而讀取以位元組、節㊣,或是區塊 為早位之貧料。 接著說明本發明NAND(反及閘)型非揮發性記憶體陣 列之抹除料。本發明之抹除方法係為對整個nand(反及 閘)型非揮發性記憶體陣列作抹除為例作說明。 請同時參照圖2及圖3C,當對記憶單元陣列進行抹除 :,於所有傳輸閘極線PL1至PL7上施加偏壓%卜源極 、字元線WL1〜WL8、位元線阳〜bu及選擇閉 ;、,、SG1〜SG2及基底為浮置。於是施加於傳輸閉極 紅間的足以在傳輸閘極與基底之間建立—個大㈣ :::得以利用F-N穿隧效離N Tu 電荷刪之間: and (反及閘)型非揮發性記憶: 由傳輸_㈣彳,响==== 25 1302741 19618twf.doc/g 除。舉例來說,若尸選遮 右建擇於傳輸閘極線PL1施加偏壓Vel, 貝z、。己_Mll〜M3卜記憶胞M12〜M32中的資料會 f抹除。亦即’共用一傳輪閘極線的兩列記憶胞中的資料 會被抹除。 此外’。本發明於進行Ναν〇(反及問)型非揮發性記憶 • ^列之#作時’係利用通道F_N穿隧效應(F-N Tunneling) ,電由通道穿過穿隨介電層注人電荷儲存層中,以進 1丁°己十思胞之程式化操作;並利用F_N穿隧效應(F-N Tunneling)使電子從電荷儲存層穿電層注入傳輸 1極中以進行胞之抹除操作。由於,本發明之操作 方=減4 了電子穿越穿隧介電層之次數,因此可以提高穿 電層之哥命,並增加元件的可靠度。而且,由於在進 行%式化刼作時,係利用電子注入效率較高的通道f_n穿 隨效應,故可以降低記憶胞電流,並且能夠提高操作速度。 另外由於程式化及抹除之動作均利用F-N f隧效應,電流 消耗小,可有效降低整個記憶體元件之功率損耗。 圖4A為繪示本發明之較佳實施例的反及閘型非揮發 陡记憶體之上視圖。圖4B為繪示圖4A中沿A_A,線之結 構剖面圖。圖4C為繪示圖4A中沿B_B,線之結構剖面圖。 九請參照圖4A至圖4C,本發明之反及閘型非揮發性記 ,體包括多數個記憶單元行]y[Rl〜MR氕Mem()ry R〇w)、 夕數條字元線WL1〜WL8、多數條位元線BL1〜BL4、多 數條源極線SL1、多數條傳輸閘極線pL1〜pL7、多數條選 擇閘極線SG卜SG2。 26 1302741 19618twf.doc/g 多數個記憶單元行MR1〜MR4例如是排列成行/列陣 列A。在本實施例中,只繪示由四個記憶單元行MRl〜 MR4構成的記憶胞陣列。但是,本發明之反及閘型非揮發 性纪fe體例如是由多數個陣列A所構成,在行方向(X方向) 上,相鄰的兩陣列A例如是以鏡向對稱的方式配置。相鄰 的兩陣列A共用源極線SL與位元線BL1〜BL4。 多數個記憶單元行MR 1〜MR4例如是設置於基底2 0 0 上。在基底200例如是N型矽基底或p型矽基底。在基底 200中例如是設置有深N型井區2〇la與位於深N型井區 201a上的P型井區201b。在基底2〇〇中例如是設置有多數 個元件隔離結構202,以定義出主動區。這些元件隔離結 構202例如是位於p型井區2〇ib中,而且這些元件隔離 結構202之表面低於基底200表面,而在基底2〇〇中形成 凹陷部204。元件隔離結構202在行方向(X方向)平行排列。 接著’說明記憶單元行之結構。由於各記憶單元行之 結構均相同,因此在下述說明中只以記憶單元行MR2為 例作說明。記憶單元行MR2包括源極區206、汲極區208、 多數個記憶單元Q、傳輸閘極210、選擇閘極212a與選擇 閘極212b。 源極區206與汲極區208例如是設置於基底200中, 且源極區206與汲極區208例如是相距一段距離。 多數個記憶單元Q例如是設置於源極區206與汲極區 2〇8之間的基底200上。在相鄰的兩個記憶單元q彼此之 間具有間隙214。 27 1302741 19618twf.doc/g 接著,說明記憶單元Q之結構。各記憶單元Q包括記 憶胞Μ及電晶體Τ。記憶胞M及電晶體τ並聯設置在一 起。 §己憶胞Μ由基底200起依序為穿隧介電層216、電荷 儲存層218、閘間介電層220、控制閘極222。 控制閉極222例如是設置於基底細上,且填滿凹陷 部204。控制間極222的材質例如是摻雜多晶石夕、金屬或 金屬矽化物等導體材料。 電何儲存層川例如是設置於控制閉極從與基底 200 ^間’包荷儲存層218的材質包括導體材料(如換雜多 晶矽等)或電荷陷入材料(如氮化矽等)。 穿隨介電層216例如是設置於基底2〇〇與電荷儲存層 ⑽之間’其材質例如是氧化石夕。間間介電層從例如是 設置於控制閑極222與電荷儲存層218 =㈣質例如是氧切'氮切'氮物或= 材料如乳切化切、氧切/氮切/氧化石夕等。 创二上亦可選擇性的設置頂蓋層 員皿曰224之材貝包括絕緣材料,例如是氧化石夕、氮 入1 [此外’在由穿随介電層216、電荷儲存層218、閘^ =層220、控制閘極222所構成之堆疊間極結構的側^ ^可以设置有絕緣間隙壁226。絕緣間隙壁226之^ 括絕緣材料,例如是氧化矽、氮化矽等。 才貝匕 電晶體Τ包括閘介電層22 8及閘極。電晶體 為控制閘極222的一部分。閑介電層挪設置於凹陷 28 1302741 19618twf.doc/g 所暴露的基底2G()表面,且位於 之間。電晶體τ例如是由吃㈣/鎌222與基底200 的導體層延伸至電冇儲存厚ΓΙ ^白勺作為控制閘極222 曰、丨甲王甩何儲存層(汙置閘極)旁 202的凹陷部204處,而和電符儲m 、、、口 ^ 200 ^ -V 〇啫存層(洋置閘極)下方的基 底200所軸之寄生電㈣所構成。由於 222做為電晶體T的閑極,因此可使她 Τ並聯設置在一起。 及屯日日體In addition, since the memory cells Qn to Q38 of the present invention are respectively composed of a plurality of transistors T11 to T38 arranged in parallel and one memory cell Mu~M38, the starting voltages of the transistors T11 to T38 are lower than the memory cells. The starting voltage of Mil~Μ38, so even if the voltage applied on the unselected word lines WL1, WL3~WL8 cannot open the channel of the memory cell, as long as this voltage can open the channel of the transistor, the current can pass and reach Select the memory unit. Therefore, during the stylization operation, since the current can be turned on by the transistors T11 to T38, the setting of the memory cell starting voltage range is not limited by the bias voltage for the unselected memory cells to be fully turned on, so that the memory Since the cell Mil~M38 has a wide starting voltage range, it can reduce the number of steps and steps of stylization confirmation, and can shorten the time of stylized operation. For example, when the memory cell is a single-order memory cell, two different threshold voltages (Vth1, Vth2), Vthl &lt; Vrefl &lt; Vth2, are discriminated by the reference reading voltage Vref. Vthl is less than 0 volts; Vrefl is about volts; Vth2 is greater than about volts. When 22 1302741 19618twf.doc/g is not fully selected, the bias voltage of the non-selected memory cell is 5 volts, and V 2 can also be set to be greater than 5 volts, making Vth2 wide. When the memory cell is a multi-order memory cell, the four readings are made by reading the electric V-, and the first two are: V〇ltage) (Vthl, Vth2, Vth3, VtM), Vthl&lt;Vrefl&lt;Vth2&lt;Vref2&lt;;Vth3&lt;Vref3&lt;Vth4. Similarly, when the unselected memory cell is in the fully open state, the bias voltage is 5 volts, 7 is less than the 〇f, which is about 〇VV; the 〇^2 volt is Vref2 is 2.2 volts, Vth3 is 2.4 to 4 volts, and Vref3 is 4 4 volts, Vth4 is 4.6 ~ 6.4 volts; and the range of vth2, Vth3, m is larger. In the above description, although the single memory of the memory device array is programmed in a unit of memory, the NAND (anti-gate) type non-volatile character _ is programmed by each character. The line, the selection of the polar line, and the control of the bit TL line are programmed in units of bytes, sections, or blocks. ^ Simultaneously, as shown in Figure 2 and Figure 3B, when the selected memory cell row is MR2. The unit Q22 performs a read operation, and the selected bit line is red and privately pressed Vrl. Voltage is applied to unselected bits, lines BU, and BL3. A voltage Vr3 is applied to the gate and the line SG1. Select gate line SG2: force [Vr4. It is known to power up M Vr5 on the character line of the selected memory list a (10). The unselected word lines WL1, WL3 WL WL8 ^ apply voltage Vr6 ° to apply power [Vl7 ' to all of the transfer gate lines PL1 to PL7 to obtain the memory cell M22 of the selected memory cell Q22. 23 1302741 19618twf.d〇c/g The voltage Vrl is the read bias applied to the selected bit line BL2. In the present example, the voltage Vrl is, for example, about L5 volts. The voltage Vr2 is, for example, about 〇Vot. Since the selection transistor ST21 and the selection transistor ST22 need to be in an on state, the voltages Vr3 and Vr4 need to be greater than or equal to the starting voltages of the selection transistor ST21 and the selection transistor ST22. In the present example, the voltage Vr3 and the voltage vr4 are, for example, about 5 volts. Because the channels of other unselected memory cells Q2 Q23~Q28 (including memory cells M21, M23~M28 or transistors T21, T23~T28) in the memory cell row MR2 need to be turned on (non-selected memory cells) Q21, Q23~Q28 are all used as gates. Therefore, the voltage vr6 needs to be greater than or equal to the starting voltage of the transistors T11 to D38. In the present example, the voltage Vr6 is, for example, about 5 volts. Voltage Vr7 turns the channel below the transmission gate on. In the present example, the voltage Vr7 is, for example, about 5 volts. In the above bias condition, the digital information stored in the memory cell can be judged by detecting the channel current of the memory cell. Moreover, since the memory cells Q11 to Q38 of the present invention are respectively composed of one transistor T11 to T38 arranged in parallel and one memory cell Mil to M38, the starting voltages of the transistors T11 to T38 are lower than the memory cell Mil~ The starting voltage of M38, therefore, even if the voltage applied to the unselected word lines WL1, WL3 WL WL8 cannot open the channel of the memory cell, as long as this voltage can open the channel of the transistor, the current can be passed. Therefore, when the selected memory cell Q22 is read, it is not interfered by the memory cells Q21, Q23 to Q28 sharing the same bit line BL2. 24 1302741 19618twf.doc/g In addition, between the memory cells of the present invention, the conventional doped regions are replaced by transmission gates, thereby avoiding short channel effects and buckoo-induced energy band reduction (Drain Induced Barrier Lowering) , DIBL) effect, etc. caused by memory leakage current. Moreover, in the above description, although the reading operation is performed in units of a single memory element in the memory element array, the reading operation of the NAND (inverse gate) type flash memory cell array of the present invention can also be performed by each character. The control of the line, the selection gate line, and the bit line, and reading the bit group, the section positive, or the block is the poor material of the early position. Next, the eraser of the NAND (anti-gate) type non-volatile memory array of the present invention will be described. The erasing method of the present invention is an example of erasing the entire nand (non-gate) type non-volatile memory array. Referring to FIG. 2 and FIG. 3C simultaneously, when the memory cell array is erased, a bias voltage source source, a word line WL1 WL WL8, and a bit line yang ~bu are applied to all of the transmission gate lines PL1 to PL7. And select closed;,,, SG1 ~ SG2 and the base is floating. Therefore, it is sufficient to establish between the transmission gate and the substrate - a large (four) ::: can be used to separate the N Tu charge by the FN tunneling effect: and (anti-gate) type non-volatile Memory: Divide by transmission _ (four) 彳, ring ==== 25 1302741 19618twf.doc/g. For example, if the corpse is chosen to be applied to the transmission gate line PL1, a bias voltage Vel, Bay z, is applied. The data in _Mll~M3 memory cells M12~M32 will be erased. That is, the data in the two columns of memory cells sharing the gate line of one pass will be erased. In addition. The present invention uses the channel F_N tunneling effect (FN Tunneling) to perform the Ναν〇 (reverse and ask) type non-volatile memory. In the layer, the stylized operation of the cell is carried out; and the F_N tunneling effect is used to inject electrons from the charge storage layer through the transmission layer to perform the cell erase operation. Since the operation of the present invention = 4 reduces the number of times the electrons cross the tunneling dielectric layer, the life of the electrical layer can be increased and the reliability of the component can be increased. Further, since the channel f_n wear-through effect with high electron injection efficiency is utilized in the %-formation operation, the memory cell current can be lowered and the operation speed can be improved. In addition, since the stylization and erasing actions utilize the F-N f tunneling effect, the current consumption is small, which can effectively reduce the power loss of the entire memory component. 4A is a top plan view of a reverse gate non-volatile steep memory in accordance with a preferred embodiment of the present invention. Fig. 4B is a cross-sectional view showing the structure taken along line A_A of Fig. 4A. 4C is a cross-sectional view showing the structure taken along line B_B of FIG. 4A. Referring to FIG. 4A to FIG. 4C, the anti-gate type non-volatile memory of the present invention includes a plurality of memory cell rows]y[R1~MR氕Mem()ry R〇w), and a plurality of word lines. WL1 to WL8, a plurality of bit lines BL1 to BL4, a plurality of source lines SL1, a plurality of transmission gate lines pL1 to pL7, and a plurality of strip selection gate lines SG and SG2. 26 1302741 19618twf.doc/g A plurality of memory cell rows MR1 to MR4 are arranged, for example, in row/column array A. In the present embodiment, only the memory cell array composed of four memory cell rows MR1 to MR4 is shown. However, the anti-gate type non-volatile body of the present invention is composed of, for example, a plurality of arrays A, and adjacent arrays A are arranged, for example, in a mirror-direction manner in the row direction (X direction). The adjacent two arrays A share the source line SL and the bit lines BL1 BLBL4. A plurality of memory cell rows MR 1 to MR 4 are, for example, disposed on the substrate 200. The substrate 200 is, for example, an N-type germanium substrate or a p-type germanium substrate. In the substrate 200, for example, a deep N-type well region 2〇la and a P-type well region 201b located on the deep N-type well region 201a are disposed. In the substrate 2, for example, a plurality of element isolation structures 202 are provided to define an active area. These element isolation structures 202 are, for example, located in p-type well regions 2〇ib, and the surface of these element isolation structures 202 is lower than the surface of the substrate 200, and recesses 204 are formed in the substrate 2''. The element isolation structures 202 are arranged in parallel in the row direction (X direction). Next, the structure of the memory cell row will be described. Since the structure of each memory cell row is the same, only the memory cell row MR2 will be described as an example in the following description. The memory cell row MR2 includes a source region 206, a drain region 208, a plurality of memory cells Q, a transfer gate 210, a selection gate 212a, and a selection gate 212b. The source region 206 and the drain region 208 are, for example, disposed in the substrate 200, and the source region 206 and the drain region 208 are, for example, at a distance. A plurality of memory cells Q are, for example, disposed on the substrate 200 between the source region 206 and the drain region 2〇8. There are gaps 214 between the adjacent two memory cells q. 27 1302741 19618twf.doc/g Next, the structure of the memory unit Q will be described. Each memory unit Q includes a memory cell and a transistor. The memory cell M and the transistor τ are arranged in parallel. The memory cell layer 216, the charge storage layer 218, the inter-gate dielectric layer 220, and the control gate 222 are sequentially formed by the substrate 200. The control closing electrode 222 is, for example, disposed on the base thinner and fills the recessed portion 204. The material of the control interpole 222 is, for example, a conductive material such as doped polycrystalline stone, metal or metal halide. The electrical storage layer is, for example, disposed of a material that controls the closed-pole from the substrate 200 to include a conductive material (such as a polysilicon or the like) or a charge trapping material (such as tantalum nitride). The dielectric layer 216 is disposed, for example, between the substrate 2 and the charge storage layer (10). The material thereof is, for example, oxidized oxide. The intervening dielectric layer is, for example, disposed on the control idler 222 and the charge storage layer 218 = (4), such as oxygen cut 'nitrogen cut' nitrogen or = material such as milk cutting, oxygen cutting / nitrogen cutting / oxidized stone Wait. On the second generation, the top cover layer member 224 can also be selectively provided with an insulating material, such as oxidized oxide, nitrogen, and nitrogen [1] in addition to the dielectric layer 216, the charge storage layer 218, and the gate. The side of the stacking pole structure formed by the layer 220 and the control gate 222 may be provided with an insulating spacer 226. The insulating spacer 226 is made of an insulating material such as tantalum oxide, tantalum nitride or the like.匕贝匕 The transistor Τ includes a gate dielectric layer 22 8 and a gate. The transistor is part of the control gate 222. The free dielectric layer is placed in the recess 21301301 19618twf.doc/g exposed surface 2G() surface and located between. The transistor τ is, for example, extended from the conductor layer of the (4)/镰222 and the substrate 200 to the thickness of the electric storage layer as the control gate 222 曰, the armor king 甩 storage layer (soil gate) 202 The recessed portion 204 is formed by the parasitic electric power (four) of the axis of the substrate 200 under the electrical storage m, , and the port ^ 200 ^ - V storage layer (the ocean gate). Since 222 acts as the idle pole of the transistor T, it is possible to arrange her in parallel. Japanese and Japanese

多數個傳輸閘極210例如是分別設置於記憶單元〇之 間的^底2GQ上,且填滿相鄰兩記憶單元Q之間的間隙 214。藉由傳輸閘極21〇而使記憶單元Q串聯連接在一起。 在傳輸閘極21G與基底2GG之間例如是設置有閘介電層 230/閘介電層230之材質例如是氧化矽。傳輸閘極21〇 亦會填滿元件隔離結構的凹陷部204,如此當利用電晶體 τ(寄生電晶體)通過電流時,傳輸閘極21〇則可打開其旁邊 的通道。 選擇閘極212a與選擇閘極212b分別設置記憶單元q 中隶外侧之兩個s己憶早元Q之侧壁’且分別與源極區鱼汲 極區相鄰。舉例來說,選擇閘極212a與源極區204相鄰, 且選擇閘極212b與汲極區206相鄰。 多數條字元線WL1〜WL8,在列方向(Y方向)平行排 列,且分別連接同一列之記憶胞之控制閘極222。 多數條位元線BL1〜BL4,在行方向(X方向)平行排 列,分別連接同一行之記憶單元行的汲極區208。各位元 線BL1〜BL4例如是猎由插基232而與〉及極區208電性連 29 1302741 196I8twf.doc/g 接。 多數條源極線SL1,在列方向(Y方向)平行排列,分 別連接同一列之記憶單元行的源極區206。 在本發明之非揮發性記憶體中,由於元件隔離結構 202之表面低於電荷儲存層218與基底200間之介面,而 形成凹陷部204,並在此凹陷部204設置與記憶胞μ並聯 連接的電晶體Τ。此電晶體Τ的設置將有助於記憶單元的 操作,可以縮短程式化操作時間,並避免讀取干擾。 一而且,在本發明之非揮發性記憶體中,由於在記憶單 元之間,以傳輸閘極210取代習知的摻雜區,因此可避免 短通道效應、汲極引發的能帶降低(Drain Induced. Lowering ’ DIBL)效應等所造成的記憶胞漏電流等。 此外’於5己丨思單元之間設置傳輸閘極21Q,此傳輸閘 極210之材質為導體,而可以遮蔽相鄰兩記憶胞的電荷儲 存層’而降低記憶胞對記憶胞之間的耦合干擾。 、、另外,在操作本發明之非揮發性記憶體時,係利用通 ,穿隧效應(F-N Tunnding)使電子經由通道穿過穿隧 =電層注入電荷儲存層中,以進行記憶胞之程式化操作; 二利用F-N穿隧效應(F-N Tunneling)使電子從電荷儲存層 ^過閘間介電層注人傳輸閘極巾,以進行記憶胞之抹除操 ^由於,本發明之操作方式減少了電子穿越穿隧介電層 可,因此可以提高穿隧介電層之壽命,並增加元件的 接著,說明本發明之非揮發性記憶體之製造方法。 30 1302741 19618twf.doc/g 第5Aj圖至第5H圖、第6A圖至第6H圖為分別繪示 本备明一較佳貫施例的一種非揮發性記憶體之製造流程剖 面圖。第5A圖至第5H圖為對應於圖4A中a_a,線的剖面 示圖弟6A圖至弟6H圖為對應於圖4A中;B-B,線的剖 面示意圖。 凊簽知、圖5A與圖6A,首先提供基底3〇〇。此基底3〇〇 例如是矽基底。接著,於基底3〇〇上形成一層介電層3〇2。A plurality of transmission gates 210 are respectively disposed on the bottom 2GQ between the memory cells ,, and fill the gap 214 between the adjacent two memory cells Q. The memory cells Q are connected in series by transmitting the gates 21A. Between the transfer gate 21G and the substrate 2GG, for example, a material in which the gate dielectric layer 230/gate dielectric layer 230 is provided is, for example, hafnium oxide. The transfer gate 21〇 also fills the recess 204 of the element isolation structure, so that when the current is passed through the transistor τ (parasitic transistor), the transmission gate 21 turns on the channel next to it. The selection gate 212a and the selection gate 212b respectively set the sidewalls of the two outer sides of the memory cell q, which are adjacent to the source region, and are respectively adjacent to the source region. For example, select gate 212a is adjacent to source region 204 and select gate 212b is adjacent to drain region 206. The plurality of word line lines WL1 to WL8 are arranged in parallel in the column direction (Y direction), and are connected to the control gates 222 of the memory cells of the same column, respectively. The plurality of bit lines BL1 to BL4 are arranged in parallel in the row direction (X direction), and are respectively connected to the drain regions 208 of the memory cell rows of the same row. The bit lines BL1 to BL4 are connected, for example, by the interposer 232 and electrically connected to the 209 and the pole region 208. A plurality of source lines SL1 are arranged in parallel in the column direction (Y direction), and are connected to the source regions 206 of the memory cell rows of the same column, respectively. In the non-volatile memory of the present invention, since the surface of the element isolation structure 202 is lower than the interface between the charge storage layer 218 and the substrate 200, the recess 204 is formed, and the recess 204 is disposed in parallel with the memory cell μ. The transistor is crucible. The setting of this transistor will help the operation of the memory unit, shorten the stylized operation time and avoid reading interference. Moreover, in the non-volatile memory of the present invention, since the conventional doped region is replaced by the transfer gate 210 between the memory cells, the short channel effect and the buck-induced energy band reduction can be avoided (Drain Induced. Lowering 'DIBL) effect, etc. caused by memory leakage current. In addition, a transmission gate 21Q is disposed between the cells, and the material of the transmission gate 210 is a conductor, and the charge storage layer of the adjacent two memory cells can be shielded to reduce the coupling between the memory cells and the memory cells. interference. In addition, when operating the non-volatile memory of the present invention, the electron tunneling (FN Tunding) is used to pass electrons through the tunnel through the tunneling = electrical layer into the charge storage layer to perform the program of the memory cell. The operation of the present invention is reduced by using FN tunneling to cause electrons to be injected from the charge storage layer through the dielectric layer of the gate to perform the erasing operation of the memory cell. The electrons can pass through the tunneling dielectric layer, so that the lifetime of the tunneling dielectric layer can be increased and the components can be added to illustrate the method of manufacturing the non-volatile memory of the present invention. 30 1302741 19618twf.doc/g FIGS. 5A to 5H and 6A to 6H are cross-sectional views showing a manufacturing process of a non-volatile memory according to a preferred embodiment of the present invention. 5A to 5H are cross-sectional views corresponding to the line a-a in Fig. 4A, and Fig. 6A to Fig. 6H are diagrams corresponding to those in Fig. 4A; B-B. Referring to Figure 5A and Figure 6A, the substrate 3 is first provided. This substrate 3 is, for example, a crucible substrate. Next, a dielectric layer 3〇2 is formed on the substrate 3〇〇.

介電層302之材質例如是氧化石夕。介電層3〇2之形成方法 例如是熱氧化法。 然後,於基底300上形成一層導體層3〇4。導體層3〇4 ,材貝例如是#雜多晶發,此導體層遍之形成方法例如 疋利用化學IU目沈積法形成-層未摻雜乡㈣層後(未緣 ’進行離子植人步驟以形成之;或者是採用臨場植I 备貝的方式以化學氣相沈積法形成之。The material of the dielectric layer 302 is, for example, oxidized stone. The method of forming the dielectric layer 3〇2 is, for example, a thermal oxidation method. Then, a conductor layer 3〇4 is formed on the substrate 300. The conductor layer 3〇4, the material shell is, for example, a #heteropolycrystalline hair, and the conductor layer is formed by a method such as 疋 using a chemical IU-deposition method to form a layer of undoped township (four) layer (the ionic implantation step is not performed) It is formed by chemical vapor deposition by means of on-site I.

底綱上形成閘間介電層3〇6,此閘間介 ^ Γ 例如是氧切/氮切/氮切,此閘間介 :二,成方法例如是先以熱氧化法形成―層底氧化 &quot;接者,再利用化學氣相沈積法形成一層氮化 其後再於氮切層上形成頂氧切層。當然 ^ :之材質也可以是氧化㈣嶋化石夕或其= 请翏照圖5B與圖6B,於基底3〇〇上形成一眉 罩幕層3〇8。此圖案化罩幕層删具有開口 sl〇 分閘間介電層3〇6。圖案化罩幕層通例如是硬罩^層。 31 1302741 19618twf.doc/g 其材質例如是氮化矽。硬罩幕層形成方法例如是先於基底 上形成一層材料層後,進行微影、蝕刻製程而形成之:當 然,圖案化罩幕層308的材質亦可以是光阻材料,其形成 f法例如是先於基底上形成一層光阻層後,對該光阻層進 行曝光、顯影而形成之。 θ 接著,以圖案化罩幕層3〇8為罩幕,移除部分閘間介 電層306、導體層3〇4、介電層3〇2與基底3〇〇,而於基底 中幵〆成夕數條溝渠312。移除部分閘間介電層306、導 體層304、介電層3〇2與基底3⑽之方法包括乾式姓刻法, 例如是反應性離子蝕刻法。溝渠312在對應圖4Α中的χ 方向上平行排列。在此步驟中,導體層304經圖案化後, 形成^條狀佈局(對應圖4Α的X方向)的導體層3〇4a。 请麥照圖5C與圖6C,於基底300上形成襯層314。 = 314之材質例如是氧化石夕。襯,314之形成方法 疋熱氧化法。 材料ΪΪμ於基底上形成一層絕緣材料層316。絕緣 ^層316之材質例如是氧化石夕。絕騎料層3!6之形成 二=如疋化學氣相沈積法。接著,移除部分絕緣材料層 I直到暴露出罩幕層规表面。移除部分絕緣材料層训 ίνΐ例如是化學機械研磨法或回1虫刻法。在此步驟中, 罩幕層规例如是作為研磨(或钱刻)終止層。 mi麥照圖5D與圖6D,移除部分絕緣材料層316與墊 &quot;14,以形成絕緣材料層316a與墊声31½。絕缘; 316a與墊層314a之#“…· 材枓層 之表面低於導體層304與基底300間之 32 1302741 19618twf.doc/g 介面’而形成凹陷部318,絕緣材料層316a與墊層M4a 即作為元件隔離結構。移除部分絕緣材料層316與墊層3i4 之方法例如是乾飯刻法。接著,移除圖案化罩幕層3〇8, 移除圖,化罩幕層3。8之方法例如是濕式侧法。 接著,於基底300上形成介電層32〇,於導體層3〇4&amp; 側壁形成介電層322。介電層32〇與介電層奶之材質例 如是氧化石夕。介電層320與介電層322之形成方法例如是 熱氧化法。在此步驟中,亦會在介電層306表面上形成一 層介電層(未緣示)。 请茶照目5E與圖6E,於基底300上形成另一層 層324 ’導體層324填滿凹陷部318。導體層似之^質例 j多晶魏金屬’由—層摻雜多晶梦層與—層金屬石夕化 物層所構成,其中摻雜多晶石夕層填滿凹 化,形成於摻雜多晶石夕層上。導體層似的形A dielectric layer 3〇6 is formed on the bottom of the gate, and the gate is, for example, an oxygen cut/a nitrogen cut/a nitrogen cut, and the gate is formed by a thermal oxidation method, for example, a layer bottom. Oxidation &quot; pick-up, then using a chemical vapor deposition method to form a layer of nitride and then forming a top oxygen layer on the nitrogen layer. Of course, the material of ^ : can also be oxidized (4) 嶋 fossil eve or its = Please refer to Figure 5B and Figure 6B to form a brow mask layer 3 〇 8 on the substrate 3 。. The patterned mask layer has an opening sl〇 dielectric layer 3〇6. The patterned mask layer is, for example, a hard mask layer. 31 1302741 19618twf.doc/g The material is, for example, tantalum nitride. The hard mask layer forming method is formed, for example, by forming a layer of material on the substrate and performing a lithography and etching process: of course, the material of the patterned mask layer 308 may also be a photoresist material, which forms the f method, for example. After forming a photoresist layer on the substrate, the photoresist layer is exposed and developed. θ Next, using the patterned mask layer 3〇8 as a mask, a part of the inter-gate dielectric layer 306, the conductor layer 3〇4, the dielectric layer 3〇2 and the substrate 3〇〇 are removed, and the substrate is 幵〆 A few ditches 312. The method of removing a portion of the inter-gate dielectric layer 306, the conductor layer 304, the dielectric layer 3〇2, and the substrate 3 (10) includes a dry-type lithography, such as reactive ion etching. The trenches 312 are arranged in parallel in the direction of the 对应 corresponding to Fig. 4A. In this step, after the conductor layer 304 is patterned, a conductor layer 3〇4a of a strip-like layout (corresponding to the X direction of FIG. 4A) is formed. A lining 314 is formed on the substrate 300 as shown in FIG. 5C and FIG. 6C. The material of =314 is, for example, oxidized stone. Lining, 314 formation method Thermal oxidation method. The material ΪΪμ forms a layer 316 of insulating material on the substrate. The material of the insulating layer 316 is, for example, oxidized stone. Formation of the rider layer 3!6 2 = Rugao chemical vapor deposition. Next, a portion of the insulating material layer I is removed until the mask gauge surface is exposed. Removing part of the insulating material layer ίνΐ is, for example, a chemical mechanical grinding method or a back-and-forth method. In this step, the mask layer gauge is, for example, a grinding (or money engraving) termination layer. Referring to FIG. 5D and FIG. 6D, a portion of the insulating material layer 316 and the pad 14 are removed to form an insulating material layer 316a and a pad 311⁄2. Insulation; 316a and the underlayer 314a #"...the surface of the material layer is lower than the 32 1302741 19618twf.doc/g interface between the conductor layer 304 and the substrate 300 to form the depressed portion 318, the insulating material layer 316a and the underlayer M4a That is, as the element isolation structure, the method of removing part of the insulating material layer 316 and the underlayer 3i4 is, for example, a dry rice carving method. Then, the patterned mask layer 3〇8 is removed, the pattern is removed, and the mask layer 3 is removed. The method is, for example, a wet side method. Next, a dielectric layer 32 is formed on the substrate 300, and a dielectric layer 322 is formed on the sidewalls of the conductor layer 3〇4 &amp; the material of the dielectric layer 32〇 and the dielectric layer milk is, for example, oxidized. The method of forming the dielectric layer 320 and the dielectric layer 322 is, for example, a thermal oxidation method. In this step, a dielectric layer (not shown) is also formed on the surface of the dielectric layer 306. 5E and FIG. 6E, another layer 324 is formed on the substrate 300. The conductor layer 324 fills the recess 318. The conductor layer is similar to the example j polycrystalline Wei metal' layer-doped polycrystalline dream layer and layer metal a layer composed of a stellite layer in which a doped polycrystalline layer is filled with a recess and formed on a doped polycrystalline layer Shaped like a conductor layer

Li,,氣相沈積法依序形成摻雜多晶發層與金屬石夕 化物層。當然,導體厚3&gt; 44· # 其他金屬材料。、θ之材貝也可以是摻雜多晶石夕或 接著於基底3〇〇上形成一層頂蓋層326。頂苗声%$ 例如是由頂蓋層326盥頂芸芦 、广曰 . 〃、項里層326b所構成。頂蓋層326a 了切。頂蓋層326b的材質例如是氮化石夕。 積法U盖層遍之形成方法例如是化學氣相沈 後’圖案化頂蓋層326、導體層324,使導體層似 &gt;、’订排列的多數個條狀導體層(對應圖4Α中,在^方 33 1302741 19618twf.doc/g 向平行排列的字元線WL1〜WL8)。圖案化頂蓋層326、導 體層324之方法例如是微影蝕刻技術。 請參照圖5F與圖6F,於頂蓋層326、導體層324之 侧壁形成絕緣間隙壁328。絕緣間隙壁328之材質例如是 氛化梦。絶緣間隙壁3 2 8之形成方法例如是先以化學氣相 沈積法形成一層絕緣材料層後,進行非等向性蝕刻製程而 形成之。Li, a vapor deposition method sequentially forms a doped polycrystalline layer and a metallized layer. Of course, the conductor thickness is 3&gt; 44· # other metal materials. The material of θ may also be a doped polycrystalline stone or a layer of capping layer 326 formed on the substrate 3〇〇. The top seedling sound %$ is composed of, for example, a top cover layer 326 dome hoist, a wide 曰 〃, and an inner layer 326b. The cap layer 326a is cut. The material of the top cover layer 326b is, for example, nitrite. The forming method of the U-coating layer is, for example, a chemical vapor deposition to "pattern the cap layer 326, the conductor layer 324, so that the conductor layer is similar", and the plurality of strip-shaped conductor layers are arranged in a row (corresponding to FIG. , in the square 33 1302741 19618twf.doc / g parallel to the word lines WL1 ~ WL8). The method of patterning the cap layer 326 and the conductor layer 324 is, for example, a lithography technique. Referring to FIGS. 5F and 6F, insulating spacers 328 are formed on the sidewalls of the cap layer 326 and the conductor layer 324. The material of the insulating spacer 328 is, for example, a dream. The insulating spacers 3 28 are formed, for example, by forming an insulating material layer by chemical vapor deposition and then performing an anisotropic etching process.

接著’以具有絕緣間隙壁328的頂蓋層326、導體層 324為罩幕’移除部分導體層304a,直到暴露出介電層3〇2。 移除部分導體層304a之後’形成彼此隔離的多數個導電塊 304b。導電塊304b是作為記憶胞的浮置閘極。介電層3〇6 是作為記憶胞的閘間介電層。導體層324是作為印情始的 控制閘極。介電層302是作為記憶胞之穿隧介電層導體 層324、介電層306、導電塊304b、介電層3〇2 ^成堆疊 閘極結構(記憶胞)。而且,導體層324填入凹陷部318之 部分及介電層320構成電晶體。藉由導體層324使得呓情 胞與電晶體並聯設置在一起。 &quot; 然後,於基底300上形成介電層33〇,於導體層3〇仆 側壁形成介電層332。介電層现與介電層332之^ 2氧化石夕。介電層330與介電層332之形成方法例士貝口是 請參照圖5G與圖6G,於基底上形成另— 層(未緣示),此導體層填滿堆疊閘極結構之間的間;,,缺 後進行非等向性侧製程,以歸部分物層,以於相^ 34 1302741 19618twf.doc/g :=:結構之間形成導體層说並於最外側之兩堆 構的側壁形成導體層334b。導體層例如是 目 '卜兀隔離結構的凹陷部318。在導體層334a盥 ^ = 2的介電層330及介電層3〇2例如是作為傳輸 甲極的閘;丨电層。在導體層334b與基底300之間的介電層 330及介電層302例如是作為選擇閘極的閘介電層。在導A portion of the conductor layer 304a is then removed by the cap layer 326 having the insulating spacers 328 and the conductor layer 324 as a mask to expose the dielectric layer 3〇2. After removing a portion of the conductor layer 304a, a plurality of conductive blocks 304b are formed which are isolated from each other. The conductive block 304b is a floating gate as a memory cell. The dielectric layer 3〇6 is a inter-gate dielectric layer as a memory cell. Conductor layer 324 is the control gate that begins as a stamp. The dielectric layer 302 is a tunneling dielectric layer conductor layer 324, a dielectric layer 306, a conductive block 304b, and a dielectric layer 3〇2 into a stacked gate structure (memory cell). Further, a portion of the conductor layer 324 filled in the recessed portion 318 and the dielectric layer 320 constitute a transistor. The conductor cells are placed in parallel with the transistor by the conductor layer 324. &quot; Then, a dielectric layer 33 is formed on the substrate 300, and a dielectric layer 332 is formed on the sidewall of the conductor layer 3. The dielectric layer is now in contact with the dielectric layer 332. The method for forming the dielectric layer 330 and the dielectric layer 332 is as follows. Referring to FIG. 5G and FIG. 6G, another layer (not shown) is formed on the substrate, and the conductor layer fills the gap between the stacked gate structures. After the absence, the anisotropic side process is performed to return to the partial layer, so that the phase is formed between the structures and the outermost two stacks are formed. The sidewall forms a conductor layer 334b. The conductor layer is, for example, a recess 318 of the object isolation structure. The dielectric layer 330 and the dielectric layer 3〇2 at the conductor layer 334a 盥 ^ = 2 are, for example, gates for transmitting the armature; The dielectric layer 330 and the dielectric layer 302 between the conductor layer 334b and the substrate 300 are, for example, gate dielectric layers that serve as gate electrodes. Guide

版層334a與導電塊304b之間及在導體層334b與導電塊 3〇4b之間的介電層332亦作為閘間介電層。 接著’於基底300中形成源極區336a與;:及極區336b。 源極區336a與汲極區336b之形成方法例如是離子植入 法。之後,於基底300上形成覆蓋層338。覆蓋層338之 材貝例如是氮化矽。覆蓋層338之形成方法例如是化學氣 相沈積法。 請芩照圖5H與圖6H,於基底300上形成層間絕緣層 340。層間絕緣層340的材質例如是磷矽玻璃、硼鱗矽玻璃 等。層間絕緣層340的形成方法例如是化學氣相沈積法。 接著,於層間絕緣層340中形成與源極區336a電性連 接的導體層342。導體層342即作為源極線。導體層342 之形成方法例如是先圖案化層間絕緣層340以形成暴露源 極區336a之開口,然後於開口中填入導體材料而形成之。 然後,於基底300上形成層間絕緣層344。層間絕緣 層344的材質例如是磷矽玻璃、硼磷矽玻璃等。層間絕緣 層344的形成方法例如是化學氣相沈積法。 35 1302741 19618twf.doc/g 接著,於層間絕緣層344、層間絕緣層34〇中形成與 汲極區336b電性連接的導體層346(導體插塞)。導體層3牝 之形成方法例如是先圖案化層間絕緣層344、層間絕緣層 • 340以形成暴露汲極區336b之開口,然後於開口中埴入^ 體材料而形成之。 ^然後,於基底3〇〇上形成導體層348。此導體層348 是作為位元線(對應圖4A中的BU〜BL4)。導體層3佔的 • 形,方法例如是於基底3〇〇上形成一層導體材料層後,進 灯微影、蝕刻製程而形成之。後續完成非揮發性記憶體之 製程為習知技術者所周知,在此不再贅述。 在本發明之非揮發性記憶體之製造方法中,由於元件 隔離結構之表面低於導體層3〇4與基底3⑻間之介面,而 形成凹陷部318,並在此凹陷部318形成與記憶胞並聯連 接的電μ體。此電晶體的形成將有助於記憶體的操作,可 以縮短程式化操作時間,並避免讀取干擾。而且,傳輸閘 極亦會填滿元件隔離結構的凹陷部31 §,如此當利用電晶 ❿ 體(寄生電晶體)通過電流時,傳輸閘極則可打開其旁邊的 通道。 而且’在本發明之非揮發性記憶體之製造方法中,由 於在記憶胞之間形成導體層334a(傳輸閘極),因此可避免 • 短通道效應、汲極引發的能帶降低(Drain Induced Barrier • LowerinS,DIBL)效應等所造成的記憶胞漏電流等。 此外,於記憶胞之間形成導體層334a(傳輸閘極),此 導體層334a(傳輸閘極)可以遮蔽相鄰兩記憶胞的導體層 36 1302741 19618twf.doc/g 304b(洋置閘極),而降低記憶胞對記憶胞之間_合干擾。 〜而=記憶胞在進行抹除操作時’可以藉由F,穿隧 =將電由:從浮置酬至導體層_ =數=種Ϊ除方式可減少了電子穿越穿隨介電 。此可以提高穿隨介電層之壽命,並增加元件 的二發性記憶體之製造方法可容易 、叙衣私I合在一起,而可以增加製程裕度。 雖然本發明已以較佳實施例揭露如上,铁 限定本發明,任何熟習此技藝者,在不脫縣發 ,圍内’當可作些許之更動與潤飾,因此本發t 範圍當視_之ΐ請專職_界定者 保達 【圖式簡單說明】 # 剖面=麟示Μ㈣反及__發性記倾的結構 記憶之練聊非揮發性 操作之一實 圖3Α所繪示為對一般記憶胞進行程 例的示意圖。 7 實 圖3Β所緣示為對源極側記憶胞進行讀取摔作之 例的示意圖。 貝斗w木TF之— 圖3C所繪示為對汲極側記憶胞進行抹除_ 例的示意®。 $仃雜作之-實 圖4Α為緣示本發明之較佳實施例的反及間型非揮發 37 1302741 19618twf.doc/g 性記憶體的上視圖。 圖4B為繪示圖4A中沿A-A’線之結構剖面圖。 圖4C為繪示圖4A中沿B-B’線之結構剖面圖。 第5A圖至第5H圖是依照本發明一較佳實施例一種非 揮發性記憶體之製造流程剖面圖。 第6A圖至第6H圖是依照本發明一較佳實施例一種非 揮發性記憶體之製造流程剖面圖。 【主要元件符號說明】 100、200、300 :基底 102 :摻雜區 104、206、336a :源極區 106、208、336b ··没極區 108 ··插塞 201a :深N型井區 201b : P型井區 202 :元件隔離結構 204 :凹陷部204 210 :傳輸閘極 212a、212b :選擇閘極 214 :間隙 216 :穿隧介電層 218 :電荷儲存層 220 :閘間介電層 222 :控制閘極 38 1302741 19618twf.doc/g 224 :頂蓋層 226、328 :絕緣間隙壁 228、230 :閘介電層 230 :插塞 • 302、320、322、330、332 :介電層 • 304、304a、324、334a、334b、342、346、348 :導體 層 304b :導體塊 ® 3G6 :賴介電層 308 ··圖案化罩幕層 310 :開口 312 :溝渠 314、314a :襯層 316、316a :絕緣材料層 326、326a、326b :頂蓋層 338 :覆蓋層 • 340、344 :層間絕緣層 BL、BL1〜BL4 :位元線The dielectric layer 332 between the layout layer 334a and the conductive block 304b and between the conductor layer 334b and the conductive bumps 3〇4b also functions as a gate dielectric layer. Next, a source region 336a and a: and a polar region 336b are formed in the substrate 300. The method of forming the source region 336a and the drain region 336b is, for example, an ion implantation method. Thereafter, a cover layer 338 is formed on the substrate 300. The material of the cover layer 338 is, for example, tantalum nitride. The formation method of the cover layer 338 is, for example, a chemical vapor deposition method. Referring to FIG. 5H and FIG. 6H, an interlayer insulating layer 340 is formed on the substrate 300. The material of the interlayer insulating layer 340 is, for example, phosphorosilicate glass, boron strontium glass, or the like. The method of forming the interlayer insulating layer 340 is, for example, a chemical vapor deposition method. Next, a conductor layer 342 electrically connected to the source region 336a is formed in the interlayer insulating layer 340. The conductor layer 342 serves as a source line. The conductor layer 342 is formed by, for example, patterning the interlayer insulating layer 340 to form an opening exposing the source region 336a, and then filling the opening with a conductor material. Then, an interlayer insulating layer 344 is formed on the substrate 300. The material of the interlayer insulating layer 344 is, for example, phosphor glass or borophosphon glass. The method of forming the interlayer insulating layer 344 is, for example, a chemical vapor deposition method. 35 1302741 19618twf.doc/g Next, a conductor layer 346 (conductor plug) electrically connected to the drain region 336b is formed in the interlayer insulating layer 344 and the interlayer insulating layer 34A. The conductor layer 3 is formed by, for example, patterning the interlayer insulating layer 344 and the interlayer insulating layer 340 to form an opening exposing the drain region 336b, and then forming the opening material into the opening. Then, a conductor layer 348 is formed on the substrate 3A. This conductor layer 348 is used as a bit line (corresponding to BU to BL4 in Fig. 4A). The shape of the conductor layer 3 is formed, for example, by forming a layer of a conductor material on the substrate 3, and then forming a lithography and etching process. The subsequent process of completing the non-volatile memory is well known to those skilled in the art and will not be described here. In the method of fabricating the non-volatile memory of the present invention, since the surface of the element isolation structure is lower than the interface between the conductor layer 3〇4 and the substrate 3(8), the depressed portion 318 is formed, and the recessed portion 318 is formed with the memory cell. Electrical μ body connected in parallel. The formation of this transistor will aid in the operation of the memory, shortening the stylized operation time and avoiding read disturb. Moreover, the transfer gate also fills the recess 31 of the element isolation structure so that when a current is passed through the transistor (parasitic transistor), the transfer gate opens the channel next to it. Further, in the method of manufacturing a non-volatile memory of the present invention, since the conductor layer 334a (transmission gate) is formed between the memory cells, the short channel effect and the energy band induced by the drain can be avoided (Drain Induced) Barrier • LowerinS, DIBL) effects such as memory leakage current. In addition, a conductor layer 334a (transmission gate) is formed between the memory cells, and the conductor layer 334a (transmission gate) can shield the conductor layers 36 1302741 19618twf.doc/g 304b (the ocean gate) of the adjacent two memory cells. , while reducing the interference between memory cells and memory cells. ~ and = memory cells in the erasing operation can be used by F, tunneling = electricity from: floating from the float to the conductor layer _ = number = type of elimination can reduce the electron crossing through the dielectric. This can improve the lifetime of the dielectric layer and increase the manufacturing process of the two-dimensional memory of the device, which can be easily combined and can increase the process margin. Although the present invention has been disclosed in the above preferred embodiments, the iron defines the present invention, and anyone skilled in the art can make some changes and refinements when not taking off the county, so the scope of the present invention is considered to be专 专 专 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Schematic diagram of the process of the cell. 7 Fig. 3 is a schematic diagram showing an example of reading and dropping the source side memory cells. The WT TF - Figure 3C is a schematic representation of the erasing of the 汲 记忆 memory cell. 。 仃 实 实 实 实 实 实 实 实 实 实 实 实 实 实 实 实 实 。 。 。 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 Fig. 4B is a cross-sectional view showing the structure taken along line A-A' in Fig. 4A. Fig. 4C is a cross-sectional view showing the structure taken along line B-B' in Fig. 4A. 5A through 5H are cross-sectional views showing a manufacturing process of a non-volatile memory in accordance with a preferred embodiment of the present invention. 6A through 6H are cross-sectional views showing a manufacturing process of a non-volatile memory in accordance with a preferred embodiment of the present invention. [Main component symbol description] 100, 200, 300: Substrate 102: Doped regions 104, 206, 336a: Source regions 106, 208, 336b · No-pole region 108 · Plug 201a: Deep N-well region 201b : P-type well region 202: element isolation structure 204: recessed portion 204 210: transmission gate 212a, 212b: selection gate 214: gap 216: tunneling dielectric layer 218: charge storage layer 220: gate dielectric layer 222 : Control gate 38 1302741 19618twf.doc/g 224 : Cover layer 226, 328: Insulation spacers 228, 230: Gate dielectric layer 230: Plugs • 302, 320, 322, 330, 332: Dielectric layer • 304, 304a, 324, 334a, 334b, 342, 346, 348: conductor layer 304b: conductor block® 3G6: dielectric layer 308 · patterned mask layer 310: opening 312: trench 314, 314a: lining 316 316a: insulating material layer 326, 326a, 326b: top cover layer 338: cover layer • 340, 344: interlayer insulating layer BL, BL1 to BL4: bit line

Ml〜M8、Mil〜M38 :記憶胞 MR1〜MR4 :記憶單元行 . PL1〜PL7 :傳輸閘極線 Q、Q11〜Q38 :記憶單元 SGI、SG2 :選擇閘極線 SL :源極線 39 1302741 19618twf.doc/g ST1、ST2、ST11 〜ST31、ST12〜ST32 :選擇電晶體 T、Til〜T38 :電晶體 WL1〜WL8 :字元線Ml~M8, Mil~M38: Memory cells MR1~MR4: Memory cell row. PL1~PL7: Transmission gate line Q, Q11~Q38: Memory unit SGI, SG2: Select gate line SL: Source line 39 1302741 19618twf .doc/g ST1, ST2, ST11 to ST31, ST12 to ST32: Select transistor T, Til~T38: transistor WL1 to WL8: word line

4040

Claims (1)

1302741 19618twf.doc/g 十、申請專利範圍: 包括多數個記憶單 1·一種反及閘型非揮發性記憶髅 元行,各該記憶單元行包括: _ 一源極區與一汲極區,設置於一基底中· 元,設置於該源‘與該汲極區之間的 纪,[咅胞些⑽早凡包括—記憶胞與—電晶體,該 °己“1兵该電晶體並聯連接在一起; 間的’㈣設置於相鄰兩該些記憶單元之 二二ϊ該些記憶單元串聯連接在-起;以及 側之該兩^單==與—第二選擇電晶體,分別與最外 鄰。己^早疋連接’且分別與該源極區與該沒極區相 憶體圍第1項所述之反及閉型非揮發性記 間隙。輪閘極填滿相鄰兩該些記憶單元之間的 層、一1儲^5己憶胞由該基底起至少包括—穿隨介電 电何鍺存層、—間間介電層與—控制閉極。 憶體==利範圍第3項所述之反及間型非揮發性記 石夕。間介電層之材質包括氧化石夕/氮化石夕/氧化 憶體3項所述之反及_揮發性記 、何储存層之材質為摻雜多晶矽。 •如申叫專利範圍第3項所述之反及卿非揮發性記 41 1302741 19618twf.doc/g 憶體,其中該穿隧介電層之材質包括氧化矽。 7.如專利範圍第3項所述之反及閘型非揮發性記 憶體’更包括多數條元件隔離結構,平行設置於該基底 各该些記憶單元行設置於相鄰兩元件結構之間。 體’其中轉讀隔離結構的表面低於該電荷儲存触 =基底間之介面而形成—凹陷部,該控制閘極填滿該二陷1302741 19618twf.doc/g X. Patent application scope: It includes a plurality of memory sheets. 1. A non-volatile memory cell line, each of which includes: _ a source region and a drain region. Set in a substrate, the element is set between the source 'and the bungee area, [the cell (10) is pre-existing - the memory cell and the - transistor, the phase has been "1 soldier" the transistor is connected in parallel Between the two (the fourth) is disposed in the adjacent two of the memory units, the two memory cells are connected in series; and the two sides of the two = single and the second selected transistor, respectively The outer neighbors have been connected to the front and the non-volatile gaps described in item 1 of the source region and the non-polar region respectively. The wheel gates fill the adjacent two. The layer between the memory cells, the memory cell from the substrate includes at least a dielectric layer, an inter-dielectric layer, and a closed-cell dielectric layer. The anti- and inter-type non-volatile memory described in the third item of the scope. The material of the inter-dielectric layer includes oxidized stone eve/nitrite eve/oxidation. The material of the reverse and _ volatile, and the storage layer of the body 3 are doped polycrystalline germanium. • The anti- and non-volatile notes described in item 3 of the patent application scope is claimed. 41 1302741 19618twf.doc/g The material of the tunneling dielectric layer includes yttrium oxide. 7. The anti-gate type non-volatile memory as described in claim 3 further includes a plurality of element isolation structures, which are disposed in parallel on the substrate. The memory cell rows are disposed between two adjacent element structures. The body 'in which the surface of the transfer isolation structure is formed below the charge storage contact-substrate interface--depression portion, the control gate fills the two depressions 9.如申請專利範圍第8項所述之反及閘型非揮 憶體’更包括-閘介電層,設置於該控制閘極與該基紅 間’各該些電晶體係由該控糊極、制介電層及該基 構成。 土 &amp; 10.如申請專利範圍第3項所述之反及閘型非揮發,丨 記憶體,其中該些記憶單元行,呈二維配置, 胞陣列,該反及閘型非揮發性記憶體更包括 性 而成一記憶9. The anti-sliding non-recallant body as described in claim 8 further includes a gate dielectric layer disposed between the control gate and the base red. The paste electrode, the dielectric layer and the base are formed. Soil &amp; 10. The non-volatile, non-volatile memory of the gate type described in claim 3, wherein the memory cells are arranged in a two-dimensional configuration, the cell array, and the gate-type non-volatile memory Body more includes sex into a memory 二多數條字元線,在列方向平行排列,且連接同一 该些記憶胞之該控制閘極及該些電晶體之閘極; 多數條位元線,在行方向平行排列,分別連 /一 之該記憶單元行的該汲極區; 订 分別連接同一列 多數條源極線,在列方向平行排列, 之該記憶單元行的該源極區; 多數條選擇閘極線,在财向平行㈣,分 二 一列之該記憶單兀行的該第一選擇電晶體之閘極鱼 選擇電晶體之閘極;以及 多數條傳輸閘極線,在列方向平行排列,分別連接门 42 1302741 19618twf.doc/g 一列之該記憶單元行的該些傳輸閘極。 二、丨1·如申请專利範圍第10項所述之反及閘型非揮發性 己L肢更包括多數條元件隔離結構,設置於該基底中, 且於行方向平行排列,各該記憶單元行設置於相鄰兩元件 結構之間。 二立12·如申請專利範圍第11項所述之反及閘型非揮發性 =憶體,其中該些元件隔離結構的表面低於該電荷儲存層 基底間之介面而形成一凹陷部,該控制閘極填滿該二 “立13.如申請專利範圍第12項所述之反及閘型非揮發性 d憶體’更包括—閘介電層,設置於該控制閘極與該基底 ’各°亥些笔日日體疋由部分該控制閘極與該閘介電層所 構成。 曰 1曰4.一種反及閘型非揮發性記憶體的製造方法,包括: μ提供一基底,該基底上已依序形成有一第一介 —第一導體層與一第二介電層; 曰 圖案化該第-導體層’以形成平行排列的多數個第一 也卞狀導體層’該些第-條狀導體層往—第—方向延伸. 士於該些第-條狀導體層之間的該基底中形成往該第一 方向延伸的多數條溝渠; 於該基底中的該些溝渠_成多數個隔離結構,該些 二離結構的表面低於該些第—條狀導體層與縣底間之介 面而形成一凹陷部,並暴露出部分該基底; 於恭露出的部分該基底表面形成一第三介電層; 於遠基底上形成-第二導體層,其中該第二導體層填 43 1302741 19618twf.doc/g 滿該凹陷部; 圖案化該第二導體層、該第二介電層及該第—條 圖二:Γίϊ個堆疊間極結構’其中該第 第二條狀導體層成::弟-方向延伸且平行排列的多數個 極結極結構之間及最外側之該兩堆疊問 稱之侧壁形成多數個第三條狀導體層。 記憶14销狀反射㈣非揮發性 ❹其中該第—條狀導體層經圖案化後形 記憶==方專,非揮發性 晶石夕。 e其中杜子置閘極之材質包括摻雜多 ihn申料鄕㈣14韻狀反及_非揮發性 ,的4造方法,其巾該第—介電層之材質包 X “ϊ=:屬::= 記憶目/ !:销敎反及_非揮發性 /氧^錢方法,其中該第二介電層包括氧切/氮化石夕 44 1302741 19618twf.doc/g 於该基底上形成一絕緣層;以及 面。移除部分該絕緣層,使該絕緣層之表面低於該基底表 如申請專·圍第14項所述之 記憶體的製造方法,豆中哕篦-八+ &amp; F ^ 行熱氧化製程。,、中為二介電層之形成方法包括進 範圍第14項所述之反及閘型非揮發性 a體的錢方法,更包括於該些堆疊閘極結構與該些第 二條狀導體層之間形成一絕緣間隙壁。 23.如申請專魏圍第14項所述之反及閘 記憶體的製造方法,其找第三介電層與料該第二條狀 導體層構成一電晶體。 24·如申凊專利範圍第14項所述之反及閘型非揮發性 記憶體的製造方:^,其巾形成於相賴麵疊閑極結構之 間的該些第三條狀導體層作為―傳輸閘極;形成於最外侧 之該兩堆疊閘極結構之側壁的該些第三條狀導體層作為一 選擇閘極。Two plurality of word lines are arranged in parallel in the column direction, and are connected to the control gates of the same memory cells and the gates of the transistors; a plurality of bit lines are arranged in parallel in the row direction, respectively a drain region of the memory cell row; a plurality of source lines connected to the same column, arranged in parallel in the column direction, the source region of the memory cell row; a plurality of strips select gate lines in the fiscal direction Parallel (four), the gate of the first selection transistor of the memory cell of the first selection transistor is divided into two columns, and the plurality of transmission gate lines are arranged in parallel in the column direction, respectively connecting the gate 42 1302741 19618 twf.doc/g A list of the transmission gates of the memory cell row. 2. 丨1· The non-volatile self-extending limb of the anti-valve type as described in claim 10 further includes a plurality of element isolation structures disposed in the substrate and arranged in parallel in the row direction, each of the memory cells The rows are placed between adjacent two component structures.立立12. The anti-valve non-volatile = memory element according to claim 11 wherein the surface of the element isolation structure is lower than the interface between the bases of the charge storage layer to form a depressed portion. The control gate fills the second "Liu 13. The anti-gate type non-volatile memory element as described in claim 12 of the patent application scope" further includes a gate dielectric layer disposed on the control gate and the substrate Each of the pens is composed of a portion of the control gate and the gate dielectric layer. 曰1曰4. A method for manufacturing a gate-type non-volatile memory, comprising: μ providing a substrate, A first dielectric layer and a second dielectric layer are sequentially formed on the substrate; the first conductive layer is patterned to form a plurality of first conductive conductor layers arranged in parallel. a first strip-shaped conductor layer extending in a - direction - a plurality of trenches extending in the first direction in the substrate between the strip-shaped conductor layers; the trenches in the substrate a plurality of isolation structures, the surfaces of the two off structures being lower than the first strips Forming a depressed portion between the body layer and the bottom of the county to expose a portion of the substrate; forming a third dielectric layer on the surface of the exposed portion; forming a second conductive layer on the far substrate, wherein the first layer The second conductor layer is filled with 43 1302741 19618 twf.doc/g to fill the recess; the second conductor layer, the second dielectric layer, and the second strip are patterned: Γ ϊ 堆叠 堆叠 堆叠 堆叠 堆叠 堆叠 其中 其中 其中The strip-shaped conductor layer is formed into: a plurality of third strip-shaped conductor layers formed between the plurality of pole-junction structures extending in parallel and parallel, and the outermost sides of the two stacks. (4) Non-volatile ❹ ❹ 该 该 条 条 条 条 条 条 条 条 条 条 = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = Shape and anti-non-volatile, 4 methods, the material of the first-dielectric layer of the towel X "ϊ =: genus:: = memory / /: 敎 敎 _ non-volatile / oxygen ^ money The method wherein the second dielectric layer comprises oxygen cut/nitridite eve 44 1302741 19618 twf.doc/g An insulating layer is formed on the substrate; and a surface. Removing a portion of the insulating layer such that the surface of the insulating layer is lower than the substrate. The method for manufacturing the memory according to the application of the above-mentioned item 14 is a thermal oxidation process of 豆-八+ &amp; F ^ . The method for forming a second dielectric layer includes the method of converting the non-volatile a-body of the gate type described in the above item 14, and further comprising the stacked gate structure and the second strip conductor An insulating spacer is formed between the layers. 23. The method for manufacturing a reverse gate memory according to claim 14, wherein the third dielectric layer and the second strip conductor form a transistor. 24. The manufacturer of the anti-gate type non-volatile memory according to claim 14 of the patent application scope, wherein the towel is formed on the third strip-shaped conductor layer between the stacked idler structures As the "transmission gate", the third strip-shaped conductor layers formed on the outermost side walls of the two stacked gate structures serve as a selection gate. 25.如申請專利範圍第14項所述之反及閘型非揮發性 記憶體的製造方法,其中該也第 摻雜多晶石夕。 26·如申請專利範圍第14項所述之反及閘型非揮發性 記憶體的製造方法,更包括於該基底巾形成源極區及没極 區。 27.—種反及閘型非揮發性記憶體之操作方法,適用於 45 1302741 19618twf.doc/g ^括多數個記憶單元行的—記憶體陣列,各該記憶單元行 &quot;又置於基底上,具有··多數個記憶單元,設置於一源極 區與二^極區之間,各該些記憶單元包括並聯連接在一起 t二己k、胞與-電·;多數個傳輸閘極,設置於該些記 單兀^間的錄底上,而使該些記憶單元串聯連接在一 ,,ϋ擇電晶體與_第二選擇電晶體,分別與最外 單元連接,且該第-選擇電晶體與該汲極區 選擇電晶體與該源極區相鄰·,多數字 且分別連接同-列之該些記憶胞之控Ϊ 電日Β體之閘極;多數條源極線分別連接同1 2:,區:多數條位元線在行方向平行排列,且‘ 方二:之5亥些汲極區;多數條第一選擇閘極線,在列 選擇電門f別Ϊ接同—列之該記憶單元行的該第— 多數條第二選擇閘極線,在列方向平 b體之二別ΐ接同一列之該記憶單元行的該第二選擇電 _傳輸閘極線,在列方向平行排列二 括:5敎觀憶單元行的該些傳輸陳,該方法= 選定記憶單元的·&quot;記憶胞進行程式化操作時,於 μ之—憶單元所耦接之該位 $於 非選定之該位元線施加一第二電壓,於亥第:二於 =7第,轉’於選定之該記憶單元所二 ^加—㈣麵,非選定該些字⑽上施加 , 屋’於所有的傳輪間極線施加一第六電麼’以利用通道= 46 1302741 19618twf.doc/g 穿隧效應程式化選定之該記憶胞,其中該第四電壓與該第 一電壓的電壓差可引發F-N穿隧效應,該第三電壓大於或 等於該第一選擇電晶體的啟始電壓,該第二電壓可抑制非 選定記憶單元行的該第一選擇電晶體開啟,該第五電壓大 於或等於該電晶體的啟始電壓,且該第六電壓可使傳輸閘 極下方的通道導通。 28. 如申請專利範圍第27項所述之反及閘型非揮發性 記憶體之操作方法,其中該第一電壓為〇伏特左右;該第 二電壓為5伏特左右;該第三電壓為5伏特左右;該第四 電壓為20伏特左右;該第五電壓為10伏特左右;該第六 電壓為5伏特左右。 29. 如申請專利範圍第27項所述之反及閘型非揮發性 記憶體之操作方法,更包括: 對一選定記憶單元的一記憶胞進行讀取操作時,於選 定之該記憶單元所耦接之該位元線施加一第七電壓,於該 第一選擇閘極線施加一第八電壓,於該第二選擇閘極線施 加一第九電壓,於選定之該記憶單元所耦接之該字元線上 施加一第十電壓,非選定該些字元線上施加一第十一電 壓,於所有的傳輸閘極線施加一第十二電壓,以讀取選定 之該記憶胞,其中該第八電壓大於或等於該第一選擇電晶 體的啟始電壓,該第九電壓大於或等於該第二選擇電晶體 的啟始電壓,該第十一電壓大於或等於該電晶體的啟始電 壓,且該第十二電壓可使傳輸閘極下方的通道導通。 30. 如申請專利範圍第29項所述之反及閘型非揮發性 47 1302741 19618twf.doc/g 記憶體之操作方法,其中該第七電壓為1.5伏特左右;該 第八電壓為5伏特左右;該第九電壓為5伏特左右;該第 十電壓為0伏特左右;該第十一電壓為5伏特左右;該第 十二電壓為5伏特左右。 31. 如申請專利範圍第27項所述之反及閘型非揮發性 記憶體之操作方法,更包括: 對記憶單元的記憶胞進行抹除操作時,於所有的傳輸 閘極線施加一第十三電壓,使該基底浮置,以利用F-N穿 隧效應抹除該些記憶胞,其中該第十三電壓與基底的電壓 差可引發F-N穿隧效應。 32. 如申請專利範圍第31項所述之反及閘型非揮發性 記憶體之操作方法,其中該第十三電壓為15伏特左右。25. The method of manufacturing a reverse-type non-volatile memory according to claim 14, wherein the doped polycrystalline silicon is also doped. 26. The method of manufacturing a reverse-type non-volatile memory according to claim 14 of the invention, further comprising forming a source region and a non-polar region in the base towel. 27. A method for operating a gate-type non-volatile memory, for 45 1302741 19618 twf.doc/g, including a memory array of a plurality of memory cell rows, each of which is placed on a substrate The upper part has a plurality of memory cells disposed between a source region and a second region, and each of the memory cells includes a parallel connection of two cells, a cell and an electricity source; and a plurality of transmission gates Provided on the recording floor between the records, and the memory cells are connected in series, the selection transistor and the second selection transistor are respectively connected to the outermost unit, and the first Selecting a transistor and the drain region selecting transistor adjacent to the source region, and multi-numbering and respectively connecting the gates of the memory cells of the same-column; the plurality of source lines respectively Connections are the same as 1 2:, area: most of the bit lines are arranged in parallel in the row direction, and 'square 2: 5 ha of some bungee areas; most of the first selection gate lines, in the column selection gate f - the first plurality of second selection gate lines of the memory cell row, in the column direction Do not connect the second selected power_transmission gate line of the memory cell row in the same column, and arrange the two lines in the column direction in parallel: the transmission of the cell line, the method = the selected memory cell &quot;When the memory cell is programmed, the bit coupled to the cell is applied to the bit line of the unselected bit line, and the second voltage is applied to the second bit: The memory cell is selected to be added to the (4) plane, and the word (10) is not selected. The house 'applies a sixth power to all the poles of the transfer wheel' to utilize the channel = 46 1302741 19618twf.doc/g The tunneling effect stylizes the selected memory cell, wherein a voltage difference between the fourth voltage and the first voltage may induce an FN tunneling effect, the third voltage being greater than or equal to a starting voltage of the first selected transistor, The second voltage can inhibit the first selection transistor of the non-selected memory cell row from being turned on, the fifth voltage being greater than or equal to the start voltage of the transistor, and the sixth voltage causing the channel under the transmission gate to be turned on. 28. The method of operating a reverse-type non-volatile memory according to claim 27, wherein the first voltage is about volts; the second voltage is about 5 volts; and the third voltage is 5. The fourth voltage is about 20 volts; the fifth voltage is about 10 volts; the sixth voltage is about 5 volts. 29. The method for operating a reverse-type non-volatile memory according to claim 27, further comprising: when a memory cell of a selected memory cell is read, the memory cell is selected Applying a seventh voltage to the bit line, applying an eighth voltage to the first select gate line, applying a ninth voltage to the second select gate line, and coupling the selected memory cell Applying a tenth voltage to the word line, applying an eleventh voltage to the unselected word lines, applying a twelfth voltage to all of the transmission gate lines to read the selected memory cell, wherein the The eighth voltage is greater than or equal to a starting voltage of the first selection transistor, the ninth voltage being greater than or equal to a starting voltage of the second selection transistor, the eleventh voltage being greater than or equal to a starting voltage of the transistor And the twelfth voltage turns on the channel below the transmission gate. 30. The method of operating a non-volatile 47 1302741 19618 twf.doc/g memory according to claim 29, wherein the seventh voltage is about 1.5 volts; the eighth voltage is about 5 volts. The ninth voltage is about 5 volts; the tenth voltage is about 0 volts; the eleventh voltage is about 5 volts; and the twelfth voltage is about 5 volts. 31. The method for operating a reverse-type non-volatile memory as described in claim 27, further comprising: applying a first to all transmission gate lines when erasing the memory cells of the memory unit The thirteen voltage causes the substrate to float to erase the memory cells by the FN tunneling effect, wherein the voltage difference between the thirteenth voltage and the substrate can induce a FN tunneling effect. 32. The method of operating a reverse-type non-volatile memory according to claim 31, wherein the thirteenth voltage is about 15 volts. 4848
TW95120523A 2006-06-09 2006-06-09 Nand type non-volatile memory and manufacturing method and operstion method thereof TWI302741B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW95120523A TWI302741B (en) 2006-06-09 2006-06-09 Nand type non-volatile memory and manufacturing method and operstion method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW95120523A TWI302741B (en) 2006-06-09 2006-06-09 Nand type non-volatile memory and manufacturing method and operstion method thereof

Publications (2)

Publication Number Publication Date
TW200746399A TW200746399A (en) 2007-12-16
TWI302741B true TWI302741B (en) 2008-11-01

Family

ID=45070552

Family Applications (1)

Application Number Title Priority Date Filing Date
TW95120523A TWI302741B (en) 2006-06-09 2006-06-09 Nand type non-volatile memory and manufacturing method and operstion method thereof

Country Status (1)

Country Link
TW (1) TWI302741B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108666316A (en) * 2017-03-31 2018-10-16 力晶科技股份有限公司 Non-volatile memory structure and method for preventing non-volatile memory structure from generating programming interference

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108666316A (en) * 2017-03-31 2018-10-16 力晶科技股份有限公司 Non-volatile memory structure and method for preventing non-volatile memory structure from generating programming interference
TWI651835B (en) * 2017-03-31 2019-02-21 力晶科技股份有限公司 Non-volatile memory structure and methods for preventing stylized interference
US10290644B2 (en) 2017-03-31 2019-05-14 Powerchip Technology Corporation Non-volatile memory structure and method for preventing non-volatile memory structure from generating program disturbance
CN108666316B (en) * 2017-03-31 2020-08-04 力晶积成电子制造股份有限公司 Non-volatile memory structure and method for preventing non-volatile memory structure from generating programming interference

Also Published As

Publication number Publication date
TW200746399A (en) 2007-12-16

Similar Documents

Publication Publication Date Title
JP4762041B2 (en) Nonvolatile semiconductor memory
US7629640B2 (en) Two bit/four bit SONOS flash memory cell
US6911690B2 (en) Flash memory cell, flash memory cell array and manufacturing method thereof
US9257446B2 (en) Semiconductor device and method of manufacturing same
US7057940B2 (en) Flash memory cell, flash memory cell array and manufacturing method thereof
US8503229B2 (en) P-/Metal floating gate non-volatile storage element
JP2002368141A (en) Non-volatile semiconductor memory device
JP2008098313A (en) Semiconductor memory device
TWI390713B (en) Non-volatile semiconductor memory device and method of manufacturing the same
TWI248087B (en) Semiconductor memory device, semiconductor device, and portable electronic apparatus
US7749855B2 (en) Capacitor structure used for flash memory
CN101118907A (en) And/not gate type non-volatility memory and manufacturing method and operation method therefor
JP2008053645A (en) Nonvolatile semiconductor storage device
US7358129B2 (en) Nonvolatile semiconductor memory device and a method of the same
JP4065671B2 (en) Nonvolatile semiconductor memory device, manufacturing method thereof, and operation method thereof
KR20050035889A (en) Non-volatile semiconductor memory element and corresponding production and operation method
US20060268607A1 (en) Operation method of non-volatile memory structure
US20190140099A1 (en) Split gate non-volatile memory (nvm) with improved programming efficiency
JP2007201244A (en) Semiconductor device
TWI302741B (en) Nand type non-volatile memory and manufacturing method and operstion method thereof
KR100990280B1 (en) Semiconductor device and method of fabricating the same
TW201537725A (en) Nonvolatile memory devices having single-layered floating gates
US20050195630A1 (en) Nonvolatile semiconductor memory device
JP2007067043A (en) Semiconductor device and its manufacturing method
JP2003332472A (en) Nonvolatile semiconductor memory and its manufacturing method