TW200907975A - Non-volatile memory and method for predictive programming - Google Patents

Non-volatile memory and method for predictive programming Download PDF

Info

Publication number
TW200907975A
TW200907975A TW097113094A TW97113094A TW200907975A TW 200907975 A TW200907975 A TW 200907975A TW 097113094 A TW097113094 A TW 097113094A TW 97113094 A TW97113094 A TW 97113094A TW 200907975 A TW200907975 A TW 200907975A
Authority
TW
Taiwan
Prior art keywords
memory
stylized
volatile memory
checkpoint
threshold voltage
Prior art date
Application number
TW097113094A
Other languages
Chinese (zh)
Other versions
TWI371041B (en
Inventor
Raul-Adrian Cernea
Original Assignee
Sandisk 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
Priority claimed from US11/733,706 external-priority patent/US7551483B2/en
Priority claimed from US11/733,694 external-priority patent/US7643348B2/en
Application filed by Sandisk Corp filed Critical Sandisk Corp
Publication of TW200907975A publication Critical patent/TW200907975A/en
Application granted granted Critical
Publication of TWI371041B publication Critical patent/TWI371041B/en

Links

Classifications

    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C16/00Erasable programmable read-only memories
    • G11C16/02Erasable programmable read-only memories electrically programmable
    • G11C16/06Auxiliary circuits, e.g. for writing into memory
    • G11C16/10Programming or data input circuits
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C11/00Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
    • G11C11/56Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using storage elements with more than two stable states represented by steps, e.g. of voltage, current, phase, frequency
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C11/00Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
    • G11C11/56Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using storage elements with more than two stable states represented by steps, e.g. of voltage, current, phase, frequency
    • G11C11/5621Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using storage elements with more than two stable states represented by steps, e.g. of voltage, current, phase, frequency using charge storage in a floating gate
    • G11C11/5628Programming or writing circuits; Data input circuits
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C16/00Erasable programmable read-only memories
    • G11C16/02Erasable programmable read-only memories electrically programmable
    • G11C16/04Erasable programmable read-only memories electrically programmable using variable threshold transistors, e.g. FAMOS
    • G11C16/0483Erasable programmable read-only memories electrically programmable using variable threshold transistors, e.g. FAMOS comprising cells having several storage transistors connected in series
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C16/00Erasable programmable read-only memories
    • G11C16/02Erasable programmable read-only memories electrically programmable
    • G11C16/06Auxiliary circuits, e.g. for writing into memory
    • G11C16/10Programming or data input circuits
    • G11C16/12Programming voltage switching circuits
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C16/00Erasable programmable read-only memories
    • G11C16/02Erasable programmable read-only memories electrically programmable
    • G11C16/06Auxiliary circuits, e.g. for writing into memory
    • G11C16/34Determination of programming status, e.g. threshold voltage, overprogramming or underprogramming, retention
    • G11C16/3436Arrangements for verifying correct programming or erasure
    • G11C16/3454Arrangements for verifying correct programming or for detecting overprogrammed cells
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C2211/00Indexing scheme relating to digital stores characterized by the use of particular electric or magnetic storage elements; Storage elements therefor
    • G11C2211/56Indexing scheme relating to G11C11/56 and sub-groups for features not covered by these groups
    • G11C2211/562Multilevel memory programming aspects
    • G11C2211/5621Multilevel programming verification

Abstract

In a nonvolatile memory having an array of memory cells, wherein the memory cells are individually programmable to one of a range of threshold voltage levels, there is provided a predictive programming mode in which a predetermined function predicts what programming voltage level needs to be applied in order to program a given memory cell to a given target threshold voltage level. In this way, no verify operation needs to be performed, thereby greatly improving the performance of the programming operation. In a preferred embodiment, the predetermined function is linear and is calibrated for each memory cell under programming by one or more checkpoints. The checkpoint is an actual programming voltage that programs the memory cell in question to a verified designated threshold voltage level.

Description

200907975 九、發明說明: 【發明所屬之技術領域】 本發明大體而言係關於諸如電可抹除可程式化唯讀記憶 體(EEPR〇M)及快閃EEPROM之非揮發性半導體記憶體’ 且特定言之係關於記憶體及程式化操作,其中最小化程式 驗證操作之數目。 【先前技術】 能夠進行電荷之非揮發性料,特定言之採取封褒為較 小形狀因數之卡的EEPR0M及快閃EEpR〇M之形式之固態 記憶體近來已成為多種行動及掌上型裝置(尤其為資訊設 備及消費型電子產品)中所選之儲存器。不同於亦為固態 記憶體之RAM(隨機存取記憶體),快閃記憶體為非揮發性 的,且即使在斷開電源之後仍保持其儲存之資料。儘管成 本較高’但快閃記憶體愈來愈多地使用於大量儲存應用 中。基於諸如硬碟機及軟性磁碟之旋轉磁性媒體之習知大 量儲存器不適於行動及掌上型環境。此係由於硬碟機傾向 於體積較大,易於產生機械故障且具有較高潛時及較高功 率要求。此等不合需要之屬性使得基於碟片之儲存器在大 多數行動及攜帶型應用設備中不實用。另一方面,嵌埋式 及採取抽取式卡之形式之快閃記憶體由於其較小尺寸、較 低力率消耗、較两迷度及較高可靠性之特徵而理想地適於 行動及掌上型環境中。 、 EEPR〇M及電可程式化唯讀記憶體(EPROM)為可經抹除 且將新的資料寫人或|,程式化,,人記憶體單元之非揮發性記 I30505.doc 200907975 憶體。兩者均利用場效電晶體結構中的定位於半導體基板 中源極與汲極區域之間的通道區域上之浮動(未連接博導 閘極。接著在浮動閘極上提供控制閘極。由保持於浮動間 極上之電荷之量來控制電晶體之臨限電壓特徵。亦即,對 於浮動閘極上之給定量之電荷,存在必須在"接通"電晶體 以允許其源極與沒極區域之間導通之前施加至控制^之 相應電壓(臨限)。 浮動閘極可固持一定範圍之電荷且因此可經程式化為臨 限電麼窗内之任何臨限電壓位準。由裝置之最小及最大臨 限位準而對臨限電麼窗之大小定界]亥等臨限位準又對應 於可程式化至浮動閘極上之電荷之範圍。臨限窗一般視記 憶體袭置之特徵、操作條件及歷史而^。窗内之每一獨 特、可解析臨限電壓位準範圍可(原則上)用以表示單元之 明確記憶體狀態。當將臨限電壓分割為兩個不同區域時之 每s己憶體單凡將能夠儲存一位元資料。類似地,當將臨 :電壓窗分割為兩個以上不同區域時’每一記憶體單元將 能夠儲存一位元以上資料。 、在常見雙態EEPR0M單元中,建立至少—電流斷點位準 以將導通窗(conduction window)分割為兩個區域。a 田稭由 施加預定、固定電壓而讀取單元時,藉由與斷點位準(或 參考電流IREF)比較而將其源極/汲極電流解析為記憶體狀 態。若電流讀數高於斷點位準之讀數,則判定單元處於— 邏輯狀態(例如,,,0,,狀態)。另一方面,若電流小於斷點: 準之電流,則判定單元處於另一邏輯狀態(例如, 130505.doc 200907975 心)因在匕,該冑態單凡儲;^ 一位元之數位二資訊。常 可於外部程式化之參考電流源作為記憶體系統之部 生斷點位準電流。 α刀產 為了增大記憶體容量,隨半導體技術之狀態進步而製造 具有愈來愈高之密度的快閃EEPR〇M裝置。增大儲存容= 之另一方法為使得每一記憶體單元儲存兩個以上狀能。 對於多狀態或多位準EEPR0]VU£憶體單元,冑由—個以 上斷點將導通窗分割為兩個以上區域以使得每—單元能夠 儲存一個以上位元之資料。給定eepr〇m陣列可儲存之資 訊因此隨每一單元可儲存之狀態之數目而增加。美國專: 第5,172,338號中已描述具有多狀㈣多位準記憶體單元之 EEPROM 或快閃 EEPROM。200907975 IX. Description of the Invention: [Technical Field of the Invention] The present invention generally relates to non-volatile semiconductor memories such as electrically erasable programmable read only memory (EEPR〇M) and flash EEPROM In particular, it relates to memory and stylized operations, where the number of program verification operations is minimized. [Prior Art] A solid-state memory in the form of a EEPR0M and a flash EEpR〇M capable of carrying a charge, in particular, a card that is sealed as a small form factor has recently become a multi-action and palm-type device ( Especially for storage devices selected for information equipment and consumer electronics. Unlike RAM (random access memory), which is also a solid-state memory, flash memory is non-volatile and retains its stored data even after the power is turned off. Despite the higher cost, flash memory is increasingly used in mass storage applications. Conventional mass storage devices based on rotating magnetic media such as hard disk drives and flexible disks are not suitable for mobile and palm-sized environments. This is because the hard disk drive tends to be bulky, prone to mechanical failure and has higher latency and higher power requirements. These undesirable attributes make disk-based storage impractical in most mobile and portable applications. On the other hand, flash memory in the form of embedded and removable cards is ideally suited for action and palm due to its small size, low power consumption, more obscurity and higher reliability. In a type environment. , EEPR〇M and electro-programmable read-only memory (EPROM) are erasable and can write new data or |, stylized, non-volatile memory of human memory unit I30505.doc 200907975 . Both utilize floating in the field region of the field effect transistor structure that is positioned between the source and drain regions of the semiconductor substrate (the gate is not connected. The gate is then provided on the floating gate. The amount of charge on the pole of the floating gate controls the threshold voltage characteristic of the transistor. That is, for a given amount of charge on the floating gate, there must be a "turn-on" transistor to allow its source and immersion regions The voltage applied to the control before the conduction is turned on. The floating gate can hold a certain range of charge and can therefore be programmed to be any threshold voltage level within the threshold window. And the maximum threshold level and the size limit of the threshold power window] Hai and other threshold levels correspond to the range of charge that can be programmed to the floating gate. The threshold window generally depends on the characteristics of the memory attack. , operating conditions and history. ^ Each unique, resolvable threshold voltage level within the window can (in principle) be used to represent the unit's explicit memory state. When dividing the threshold voltage into two different regions Each suffix will be able to store one meta-data. Similarly, when the voltage window is divided into two or more different regions, each memory cell will be able to store more than one bit of data. In a common two-state EEPR0M unit, at least a current breakpoint level is established to divide a conduction window into two regions. a field straw is read by a predetermined, fixed voltage, and a breakpoint is used. The reference source (or reference current IREF) is compared to resolve its source/drain current to the memory state. If the current reading is higher than the breakpoint level reading, then the decision cell is in the - logic state (eg,, 0,, State). On the other hand, if the current is less than the breakpoint: the quasi-current, the decision unit is in another logic state (for example, 130505.doc 200907975 heart), because the embarrassing state, the embarrassing state is stored; ^ one-dimensional Digital two-information. The externally stylized reference current source can be used as the breakpoint level current of the memory system. In order to increase the memory capacity, α-knife manufacture is becoming more and more high with the advancement of semiconductor technology. It Flash EEPR〇M device. Another way to increase storage capacity is to store more than two states in each memory cell. For multi-state or multi-bit quasi-EEPR0] VU memory cells, More than one breakpoint divides the conduction window into more than two regions so that each cell can store more than one bit of information. The information that can be stored for a given eepr〇m array is therefore increased with the number of states that each cell can store. U.S. Patent No. 5,172,338 describes an EEPROM or flash EEPROM having a multi-shaped (four) multi-bit memory cell.

U 通常藉由兩個機制中之-者將用作記憶體單元之電曰體 程式化為|,程式化,,狀態。纟”熱電子注入,,中,施加至:極 之高電壓將電子加速穿越基板通道區域。同時,施加至控 制閑極之高電壓將熱電子經由較薄閘極介電質拉至浮動閘 極上。在"穿隨注入"中,相對於基板向控制極施加較高 電壓。以此方式,將電子 卞目基板拉至插入之浮動閘極。 可藉由許多機制而抹除記憶體裝置。對於贿⑽而 言,可藉由以紫外輕射將電荷自浮動閑極移除而整體抹除 記憶體。對於耐臟而言,可藉由相對於控制閉極施加 車父南之電壓至基板以誘發浮動閘極中之電子經由較薄氧化 物而穿隨至基板通道區域(亦即’福勒_諾德海姆穿隧)而電 抹除記憶體單元。通常,可逐位元組地抹除卿函 130505.doc 200907975 於快閃EEPROM而言,可一次全部或每次一或多個區塊地 電抹除S己憶體,其中一區塊可由記憶體之5丨2或5丨2以上之 位元組組成。 記憶體裝置通常包含可安裝於一卡上之一或多個記憶體 a曰片。每一記憶體晶片包含由諸如解碼器及抹除、寫入及 讀取電路之周邊電路支援的記憶體單元之陣列。較為複雜 之記憶體裝置亦與執行智慧型及較高層級記憶體操作及介 面連接之外部記憶體控制器一同操作。 現今正使用許多商業成功之非揮發性固態記憶體裝置。 此專§己憶體裝置可為快閃EEPR0M或可使用其他類型之非 揮發性記憶體單元。美國專利第5,〇7〇,〇32號、第5,〇95,344 號、第 5,315,541 號、第 5,343,063 及第 5,661,053 號、第 5,313,421號及第6,222,762中給出快閃記憶體及系統之實例 以及製造其之方法。特定言之,美國專利第5,57〇,315號、U usually stylizes the electrical body used as a memory unit into |, stylized, and state by two mechanisms.热"Hot-electron injection,", applied to: a very high voltage accelerates electrons across the substrate channel region. At the same time, a high voltage applied to the control idler pulls the hot electrons through the thinner gate dielectric to the floating gate In "wearing &", a higher voltage is applied to the gate relative to the substrate. In this way, the electronic substrate is pulled to the inserted floating gate. The memory device can be erased by a number of mechanisms For bribe (10), the memory can be erased as a whole by removing the charge from the floating idler by ultraviolet light. For dirt resistance, the voltage of the south of the vehicle can be applied by controlling the closed pole to The substrate electrically insulates the memory cells by inducing electrons in the floating gate to pass through the thinner oxide to the substrate channel region (ie, 'Fowler_Nordheim tunneling'). Typically, the bitwise tuple In the case of a flash EEPROM, the S memory can be erased all at once or one or more blocks at a time, one of which can be 5丨2 or 5 of the memory.丨2 or more of the byte group. Memory device usually One or more memory cells are mounted on a card. Each memory chip includes an array of memory cells supported by peripheral circuits such as decoders and erase, write and read circuits. Complex memory devices are also operated with external memory controllers that perform intelligent and higher-level memory operations and interface connections. Many commercially successful non-volatile solid-state memory devices are being used today. The device may be flash EEPROM or other types of non-volatile memory cells may be used. U.S. Patent Nos. 5, 〇 7 〇, 〇 32, 5, 〇 95, 344, 5, 315, 541, 5, 343, 063 and 5, 661, 053, Examples of flash memory and systems and methods of making the same are given in U.S. Patent Nos. 5,313,421 and 6,222,762. In particular, U.S. Patent No. 5,57,315,

第5,903,495號、第6,046,935號中描述了具有NAND串結構 之快閃記憶體裝置。又,非揮發性記憶體裝置亦由具有用 於儲存電荷之介電層的記憶體單元製造。替代早先描述之 傳導浮動閘極元件,使用介電層。利用介電質儲存元件之 該等記憶體裝置已由Eitan等人於',NR〇M: A Novel Localized Trapping, 2-Bit Nonvolatile Memory Cell·' » IEEEA flash memory device having a NAND string structure is described in U.S. Patent Nos. 5,903,495 and 6,046,935. Also, the non-volatile memory device is also fabricated from a memory cell having a dielectric layer for storing charge. Instead of the conductive floating gate elements described earlier, a dielectric layer is used. These memory devices utilizing dielectric storage elements have been used by Eitan et al., 'NR 〇M: A Novel Localized Trapping, 2-Bit Nonvolatile Memory Cell·' » IEEE

Electron Device Letters,第 21卷,第 11 號,200(^U 月, 第543-545頁中描述。ΟΝΟ介電層延伸跨越源極與汲極擴 散區之間的通道。將用於一資料位元之電荷定位於介電層 中鄰近於汲極處’且將用於另一資料位元之電荷定位於介 130505.doc 200907975 電層中鄰近於源極處。舉例而言,美國專利第5,768,192及 第6,011,725號揭示具有夾於兩個二氧化矽層之間的捕集介 電質之非揮發性記憶體單元。藉由分別讀取介電質内之空 間分離電荷儲存區域之二元狀態來實施多狀態資料儲存。 為了改良讀取及程式化效能,並行地讀取或程式化陣列 中之多個電荷儲存元件或記憶體電晶體。因此,一同讀取 或程式化記憶體元件之”頁在現有記憶體架構中,—列Electron Device Letters, Vol. 21, No. 11, 200 (^U, pp. 543-545. The ΟΝΟ dielectric layer extends across the channel between the source and drain diffusion regions. It will be used for a data bit. The charge of the element is positioned adjacent to the drain in the dielectric layer and the charge for the other data bit is located adjacent to the source in the electrical layer 130505.doc 200907975. For example, US Patent No. 5 No. 6, 011, 725 discloses a non-volatile memory cell having a dielectric sandwiched between two ruthenium dioxide layers. The memory storage region is separated by reading the space within the dielectric, respectively. Dual state to implement multi-state data storage. To improve read and program performance, read or program multiple charge storage elements or memory transistors in an array in parallel. Therefore, read or program memory together The "page of the component" in the existing memory architecture, - column

通常含有交錯頁或者其可組成―胃。將—同讀取或程 式化一頁之所有記憶體元件。 使用一系列交替程式化/驗證循環之習知程式化技術係 用來處理程式化過程中之不確定性,纟中單元之臨限電麼 最初回應於VPGM之相對較大的改變而快速增長。然而,增 長減慢且最後停止,因為程式化至浮動閘極中之電荷心 屏蔽而減小用於進-步將電子穿㈣浮動㈣中的有效^ 場。該過程呈現為高度非線性的,且因此使用試誤法 (trial-and-error approach)。 径式化/驗證程式 几日艰牯用日寻間 且衫響效能。該問題由於能夠儲存 1几之s己憶體單元 的實轭而加劇。本質上,需對 態中之之可能多個狀 母者執订驗扭。對於具有16個可能 記憶體,此音喟备一舲續半跟1 體狀也之 在… 驗*步驟會引起至少16個感應操作。 二八他機制中,可能甚至還 少仏、杜 ^方丁-人因此,喃芸 夕位準記憶體單元("MLC")中可 增大,铲"狀匕、位準之數目不斷 a式化/驗證機制之驗證循環變得愈加耗時。 130505.doc 200907975 由Loc Tu等人於2006年9月12日申請之題為,,Meth〇d f〇rUsually contains staggered pages or it can form a stomach. All memory elements of a page will be read or programmed. The conventional stylization technique using a series of alternating stylization/verification loops is used to deal with the uncertainty in the stylization process. The current limit of the unit is initially increased in response to the relatively large changes in VPGM. However, the growth slows down and eventually stops because the stabilizing of the charge core in the floating gate reduces the effective field used to further (4) float (4). This process appears to be highly non-linear and therefore uses a trial-and-error approach. The calibre/verification program is difficult to use in a few days. This problem is exacerbated by the ability to store the yoke of a few of the suffix units. In essence, it is necessary to perform a check on the possible multiple mothers in the state. For a 16-possible memory, this tone is also a half-length with a body shape. The test step will cause at least 16 sensing operations. In the mechanism of the 28th, there may be even less 仏, Du 方方丁-人, therefore, the 芸 芸 位 quasi-memory unit ("MLC") can be increased, the number of shovel " The verification loop of the a-style/verification mechanism becomes more time consuming. 130505.doc 200907975 Applied by Loc Tu et al. on September 12, 2006, Meth〇d f〇r

Non-volatile Memory with Linear Estimation of Initial Programming Voltage”的美國專利申請案序列號第 11/53 1,227號揭*藉由線性估計來估計初始程式化電壓之 方法。為了達成非揮發性記憶體之良好程式化效能,必須 於工廠最佳地選擇初始程式化電壓VpGMQ及步長。此藉由 測試記憶體單元之每-頁而完成。藉由具有階梯波形的一 系列電壓脈衝來連續地程式化耦接至選定頁之字線,其中 在脈衝之間進行驗證直至該頁經驗證為指定模式。在對頁 進行程式化驗證時之程式化電壓將用以藉由線性按比例縮 減至該頁之起始程式化電壓的初始值而進行估 第二次通過中使用得自帛一次通過的估相進—步改進估 計。因此,使用習知的交替程式化與驗證來產生用於成功 地程式化-頁之最終程式化電壓。接著,線性按比例縮減 最終程式化電壓以達到該頁之所估計的初始程式化電壓。 此類型之按比例縮放係在頁級別之總規模上,且不解決習 知的在逐單元基礎上程式化及驗證現場之記憶體之劣勢。 十因此’#线於高容量及高效能非揮發性記憶體之普遍 萬要特疋口之,知要具有一種具有改良的程式化效能之 高容量非揮發性記憶體,在其中最小化前述劣勢。 【發明内容】 根據本發明之一一般態樣,在具有記憶體單元之陣列的 非揮發性記憶體中(其中記憶體單元可個 限電壓位準之範圍中之一者),提供預測需施:=: 130505.doc -10- 200907975 電壓位準以將給定記憶體單元程式化為給定目標臨限電壓 位準之預疋函數。以此方式,無需執行驗證操作,藉此大 幅改良程式化操作之效能。 在實施例中,由按比例產生給定目標臨限電壓位準之 程式化電壓位準的線性函數來近似得出預定函數。該線性 函數具有由適用於記憶體陣列之一定數目單元的預定平均 值給出之斜率。藉由針對給定記憶體單元預定線性函數上 的檢查點來針對該給定記憶體單元唯一地判定線性函數。 檢查點係基於將記憶體單元程式化至指定臨限電壓位準之 實際程式化電壓。檢杏 檢查點較佳地對應於記憶體單元之最低 程式化狀悲、中之—者。最初藉由(例如)習知程式化/驗證程 式化技術來將記憶體單元程式化至檢查點且對其加以驗 證。以此方式,判定對於將記憶體單元程式化至指定記情 體狀態為必要之實際程式化電壓的檢查點值。因此預定I、 數在被用以判定用於將記憶體單元程式化至目標臨限電壓 :準之程式化電厂堅值之前經校正以產生在以檢查點臨限電 壓位準5平估時的檢查點程式化電壓值。 在:-實施例中,可在記憶體單元所支援之可 壓位準的範圍中指定多個檢查點。每一檢查點將用以校: 母一檢查點附近之局部歡函數。使用局部預定函數來預 測用:程式化至相關聯之檢查點附近的目 之程式化電壓位準。 + 要=Γ=為有利的,因為程式化至目標狀態不需 要驗《作。僅需要驗證操作來列定―般而言在數目上^ 130505.doc 200907975 點。 誤結果之變化,但此等 "、可藉由適當錯誤校正 小於可能記憶體狀態之數目的檢查 將存在將使得預測程式化產生錯 誤差將為在統計上可預測之誤差, 碼("ECC")而得到處理。 明之額 自對本發明之較佳實施例的以下描述將瞭解 外特徵及優勢,應結合隨附圖式來理解該描述。 【實施方式】 -$ 記憶體系統Non-volatile Memory with Linear Estimation of Initial Programming Voltage, US Patent Application Serial No. 11/53, 227, discloses a method for estimating an initial stylized voltage by linear estimation. For good stylized performance, the initial programmed voltage VpGMQ and the step size must be optimally selected at the factory. This is done by testing each page of the memory cell. Continuously stylized by a series of voltage pulses with a staircase waveform. a word line coupled to the selected page, wherein verification is performed between pulses until the page is verified to be in a specified mode. The stylized voltage at the time of programmatic verification of the page is used to linearly scale down to the page Estimating the initial value of the stylized voltage and estimating the phase-by-step improvement estimate using the one-pass pass in the second pass. Therefore, using conventional alternate stylization and verification to generate a successful stylization - The final stylized voltage of the page. Next, the final stylized voltage is linearly scaled down to achieve the estimated initial programmed voltage for the page. The scaling of the type is on the total scale of the page level, and does not solve the conventional disadvantage of staging on the unit-by-unit basis and verifying the memory of the site. Ten therefore the ## line is high-capacity and high-efficiency non-volatile The memory is generally versatile, and it is known to have a high-capacity non-volatile memory with improved stylized performance, in which the aforementioned disadvantages are minimized. [Invention] According to one aspect of the present invention, In non-volatile memory with an array of memory cells (where the memory cell can be one of a range of voltage levels), predictive needs are provided: =: 130505.doc -10- 200907975 Voltage level To pre-program a given memory unit to a predetermined target threshold voltage level. In this way, no verification operations need to be performed, thereby greatly improving the performance of the stylized operation. In an embodiment, by proportional Generating a linear function of a programmed voltage level for a given target threshold voltage level to approximate a predetermined function having a predetermined number of units suitable for the memory array The slope given by the mean. The linear function is uniquely determined for the given memory unit by a checkpoint on a predetermined linear function for a given memory unit. The checkpoint is based on stylizing the memory unit to a specified threshold voltage. The actual stylized voltage of the level. The checkpoint of the check april preferably corresponds to the lowest stylized sorrow of the memory unit. The memory is initially stored by, for example, conventional stylization/verification stylization techniques. The body unit is stylized to the checkpoint and verified. In this way, the checkpoint value for the actual stylized voltage necessary to program the memory unit to the specified statistic state is determined. Therefore, the predetermined I and number are Used to determine the stylized voltage value used to program the memory cell to the target threshold voltage: the quasi-programmed power plant is calibrated to generate a checkpoint at the checkpoint threshold voltage level 5 . In the:-embodiment, a plurality of checkpoints can be specified in the range of compressible levels supported by the memory unit. Each checkpoint will be used to: a local joy function near the checkpoint. Use a local predetermined function to predict: programmatically level to the programmed voltage level near the associated checkpoint. + ==Γ= is advantageous because stylized to the target state does not need to be tested. Only verify operations are required to set the number in general - 130505.doc 200907975 points. A change in the result of the error, but such a check that the number of possible memory states can be corrected by appropriate error correction will cause the prediction to be stylized to produce a false error that will be a statistically predictable error, code ("ECC") is handled. The description will be understood from the following description of the preferred embodiments of the invention, and the description will be understood in conjunction with the accompanying drawings. [Embodiment] -$ Memory System

中可實施本發明之各 圖1至圖5說明實例記憶體系統, 種態樣。 圖ό說明習知程式化技術。 圖7至圖16說明本發明之各種態樣及實施例。 圖1示意地說明本發明可實施於之非揮發性記憶體晶片 之功能區塊。記憶體晶片1〇〇包括記憶體單元之二維陣列 ⑽、控制電路21G及諸如解碼器、讀取寫人電路及多工器 之周邊電路。 藉由字線經由列解碼器23〇(拆分為23〇α、23〇β)及藉由 位元線經由行解碼器260(拆分為26〇Α、26〇Β)而可定址記 It體陣列200(亦見圖4及圖5)。讀取寫入電路27〇(拆分為 270A、270B)允許並行讀取或程式化記憶體單元之頁。資 料I/O匯流排231耦接至讀取寫入電路27〇β 在較佳實施例中,一頁由共用同一字線之一列鄰近記憶 體單7G組成。在另一實施例中,將一列記憶體單元分割為 多個頁,提供區塊多工器250(拆分為25〇a及250B)以將讀 130505.doc 200907975 取/寫入電路270多工至個別頁。舉例而言,將分別由奇數 與偶數行記憶體單元形成之兩頁多工至讀取寫入電路。 圖1說明一較佳配置’其中以在陣列之相反兩側上對稱 之方式實施各種周邊電路對記憶體陣列200之存取,以使 知·母側上之存取線及電路的密度減半。因此,將列解碼器 拆分為列解碼器230A及230B且將行解碼器拆分為行解碼 器260A及260B。在將一列記憶體單元分割為多個頁之實 也例中將頁多工益250拆分為頁多工器25 〇A及25 0B。類 似地,將讀取寫入電路27〇拆分為自陣列2〇〇之底部連接至 位元線之讀取寫入電路27〇八及自陣列2〇〇之頂部連接至位DETAILED DESCRIPTION OF THE INVENTION Each of Figures 1 through 5 illustrates an example memory system, aspects of which. The figure illustrates the conventional stylization technique. 7 through 16 illustrate various aspects and embodiments of the invention. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a schematic illustration of functional blocks of a non-volatile memory chip that can be implemented in accordance with the present invention. The memory chip 1 includes a two-dimensional array of memory cells (10), a control circuit 21G, and peripheral circuits such as a decoder, a read writer circuit, and a multiplexer. It can be addressed by the word line via the column decoder 23 (divided into 23〇α, 23〇β) and by the bit line via the row decoder 260 (split into 26〇Α, 26〇Β) Body array 200 (see also Figures 4 and 5). The read write circuit 27 (divided into 270A, 270B) allows pages of memory cells to be read or programmed in parallel. The data I/O bus 231 is coupled to the read write circuit 27 〇 β. In the preferred embodiment, a page consists of a column adjacent to the memory bank 7G sharing one of the same word lines. In another embodiment, a column of memory cells is divided into a plurality of pages, and a block multiplexer 250 is provided (split into 25A and 250B) to read 130505.doc 200907975 to/write circuit 270. To individual pages. For example, two-page multiplexes formed by odd and even row memory cells, respectively, are applied to the read write circuit. Figure 1 illustrates a preferred configuration 'where access to the memory array 200 is performed by various peripheral circuits in a manner symmetrical on opposite sides of the array to halve the density of the access lines and circuitry on the female side . Therefore, the column decoder is split into column decoders 230A and 230B and the row decoder is split into row decoders 260A and 260B. In the case of dividing a column of memory cells into a plurality of pages, the page multi-worker 250 is also divided into page multiplexers 25 〇 A and 25 0B. Similarly, the read write circuit 27 is split into a read write circuit 27〇 connected to the bit line from the bottom of the array 2〇〇 and connected to the top of the array 2〇〇.

:線之讀取寫入電路27〇B。以此方式,讀取/寫入模組之 费度及因此感應模組38〇之密度基本上減半。 控制電路110為與讀取寫入電路27〇合作以執行記憶體陣 列2〇0上之記憶體操作的晶片上控制器。控制電路110通常 包括狀態機112及諸如晶片上位址解碼器及功率控制模組 崔展丁)的其他電路。狀態機112提供記憶體操作之晶 片級控制1制電路經由外部記憶體控制器與主機通传。 通常將記憶體陣列200組織為記憶體單元之二維陣列, =記=元排列為多列與多行且可藉由字線及 -址。可根據職型或_型架構而形成陣列。 圖2示意地說明非揮^ ^ ^ ^ ^ 一 °己隐體早70。記憶體單元10可 由八有诸如净動閘極 晶體實施。記,體單存单元20的場效電 極30。 Uh10亦包括源極H、没極16及控制閉 130505.doc -13· 200907975 現今正使用許多商業成功之非揮發性固態記憶體裝置。 此等記憶體裝置可使用不同類型之記憶體單元,每一類型 具有一或多個電荷儲存元件。 典型非揮發性記憶體單元包括EEPROM及快閃 EEPROM。EEPROM單元之實例及製造其之方法給出於美 國專利第5,595,924號中。快閃EEPROM單元之實例、其在 記憶體系統中之使用及製造其之方法給出於美國專利第 5,070,032 號、第 5,095,344 號、第 5,315,541 號、第 5,343,063 號、第 5,661,053 號、第 5,313,421 號及第 6,222,762號中。特定言之,具有NAND單元結構之記憶體 裝置的實例描述於美國專利第5,570,3 15號、第5,903,495 號、第6,046,935號中。又,利用介電質儲存元件之記憶體 裝置的實例已由 Eitan %人於"NROM: A Novel Localized Trapping, 2-Bit Nonvolatile Memory Cell" > IEEE Electron: Line read write circuit 27〇B. In this way, the cost of the read/write module and thus the density of the sensing module 38 is substantially halved. Control circuit 110 is an on-wafer controller that cooperates with read write circuit 27 to perform memory operations on memory array 2〇0. Control circuit 110 typically includes state machine 112 and other circuitry such as on-wafer address decoder and power control module Cui Zhanding. The state machine 112 provides a wafer level control of the memory operation. The circuit is communicated to the host via the external memory controller. The memory array 200 is typically organized into a two-dimensional array of memory cells, with the =signs being arranged in multiple columns and rows and by word lines and addresses. The array can be formed according to the job type or the _ type architecture. Fig. 2 schematically illustrates that the non-volatility ^ ^ ^ ^ ^ ° has been hidden as early as 70. The memory cell 10 can be implemented by eight crystals such as a net moving gate. The field effect electrode 30 of the unit memory unit 20 is recorded. Uh10 also includes source H, immersion 16 and control closure 130505.doc -13· 200907975 Many commercially successful non-volatile solid state memory devices are being used today. These memory devices can use different types of memory cells, each type having one or more charge storage elements. Typical non-volatile memory cells include EEPROM and flash EEPROM. An example of an EEPROM cell and a method of making the same are given in U.S. Patent No. 5,595,924. Examples of flash EEPROM cells, their use in memory systems, and methods of making them are given in U.S. Patent Nos. 5,070,032, 5,095,344, 5,315,541, 5,343,063, 5,661,053, 5,313,421. And in No. 6,222,762. In particular, an example of a memory device having a NAND cell structure is described in U.S. Patent Nos. 5,570,315, 5,903,495, 6,046,935. Further, an example of a memory device using a dielectric storage element has been used by Eitan % "NROM: A Novel Localized Trapping, 2-Bit Nonvolatile Memory Cell"> IEEE Electron

Device Letters,第 21卷,第 11 號,2000 年 η 月,第 543- 5斗5頁中及在美國專利第5,768,192號與第6,〇11,725號中描 述。 實務上通吊精由在向控制閘極施加參考電壓時感應跨 越單元之源極與汲極電極之傳導電流而讀取單元之記憶體 狀態。因此,對於單元之浮動閘極上之每一給定電荷,可 偵測到關於固定參考控制閘極電壓之相應傳導電流。類似 地,可釭式化至浮動閘極上之電荷之範圍界定相應臨限電 壓®或相應傳導電流窗。 或者,替代偵測經分割之電流窗中之傳導電流,在控制 130505.doc •14. 200907975 閘極處可能對於受測之給定記憶體狀態㈣臨限電壓且债 測傳導電流是否低於或高於臨限電流。在一實施中,# 、 檢查傳導電流經由位元線之電容而放電之速率來完成= 於臨限電流對傳導電流之镇測。 圖3說明對於浮動開極於任—時間可選擇性地料^ 個不同電荷Q1-Q4的源極-汲極電流1〇與控制閘極電壓να 之間的關係。四個實線Id*Vcg之關係曲線表示可經程: 化至記憶體單7L之浮動閘極上的四個可能電荷位準,其分 別對應於四個可能記憶體狀態。作為一實例,一定數目之 單元之臨限電壓窗可在0.5 V至3·5 ¥之範圍中。可藉由以 各為0·5 V之間隔將臨限窗分割為五個區域而劃分分別表 示一個抹除狀態與六個程式化狀態之七個可能記憶體狀態 t、"Ρ ' "2”、"3”、”4"、”5"、"6”。舉例而言,若如圖 示而使用2 μΑ之參考電流IREF,則可將以…而程式化之 單元視作處於記憶體狀態"Γ,中,因為其曲線與Iref相交於 臨限窗之藉由VCG=0.5 乂與^ v所劃分之區域中。類似 地,Q4處於記憶體狀態”5"中。 如自上文之描述可見,使記憶體單元儲存愈多狀態,其 私限1¾受到愈精細之劃分。舉例而言,記憶體裝置可具有 3有在-1.5 V至5 V之範圍内變動之臨限窗的記憶體單 元。此提供6·5 V之最大寬度。若記憶體單元待儲存16個 狀態,則每一狀態可佔據臨限窗中之2〇〇爪¥至3〇〇 mV。 此將需要程式化及讀取操作中之較高精度以便能夠達成所 需解析度。 130505.doc -15- 200907975 圖4說明記憶體單元之職陣列之—實例。在記憶體陣 列200中’每-列記憶體單元藉由其源極14及沒極μ以菊 式鏈接(daisy-chain)之方式而連接。有時將此設計稱作虛 擬接地設計…列中之單㈣之控制閘極3()連接至諸如字 線42之字線。-行中之單元之源極及汲極分別連接至諸如 位元線34及36之選定位元線。 圖5A示意地說明經組織為财仙串之—串記憶體單元。 NAND串50包含一系列記憶體電晶體M1、M2...Mn(例如, η 4 8 1 6或16以上)’該等記憶體電晶體藉由其源極及 汲極而經菊式鏈接。一對選擇電晶體S1、S2分別控制記憶 體電晶體鏈經由NAND串之源極端子54及汲極端子%與外 部之連接。在§己憶體陣列中,當接通源極選擇電晶體s ^ 時,源極端子耦接至源極線(見圖5B)<>類似地,當接通汲 極選擇電晶體S2時’ NAND串之沒極端子純至記憶體陣 列之位兀線。鏈中之每一記憶體電晶體丨〇充當記憶體單 兀。其具有一電荷儲存元件2〇以儲存給定量之電荷以便表 不所欲之記憶體狀態。每一記憶體電晶體之控制閘極3〇允 s午對讀取及寫入操作之控制。如將於圖5B中所見,—列 NAND串之相應記憶體電晶體的控制閘極3 〇均連接至同— 字線。類似地,選擇電晶體S1、82中之每一者之控制閘極 32分別經由NAND串之源極端子54及汲極端子56而提供對 NAND串之控制存取。同樣,一列NAND串之相應選擇電 晶體的控制閘極3 2均連接至同一選擇線。 當在程式化期間讀取或驗證NAND串内之經定址之記情 130505.doc -16- 200907975 體電晶體ίο時,向其控制閘極30供應一適當電壓。同時, 藉由向NAND串5 0中未經定址之記憶體電晶體的剩餘部分 的控制閘極施加充足電壓而完全接通NAND串5〇中未經定 址之5己憶體電晶體的剩餘部分。以此方式,自個別記憶體 電晶體之源極至NAND串之源極端子54及同樣地自個別記 憶體電晶體之汲極至單元之汲極端子56有效產生傳導路 t °美國專利第5,57〇,315號、第5,9〇3,495號、第 6,〇46,935號中描述具有該等NAND串結構之記憶體裝置。 圖5B說明由諸如圖5A中所示2NAND串的NAND串5〇組 成之汜憶體單元之NAND陣列200的實例。沿NAND串之每 行,諸如位元線3 6之位元線耦接至每一 NAND串之汲極 端子56。沿NAND串之每一組,諸如源極線34之源極線耦 接至每一 NAND串之源極端子54。沿NAND串之一組中的 列§己憶體單元之控制閘極亦連接至諸如字線42之字線。 沿NAND串之一組中的一列選擇電晶體之控制閘極連接至 諸如選擇線44之選擇線。可藉由NAND串之組之字線及選 擇線上的適當電壓來定址NAND串之組中的一整列記憶體 單元。在讀取NAND串内之一記憶體電晶體時,串中之剩 餘έ己憶體電晶體經由其相關聯之字線而被硬接通以使得流 過該串之電流基本上視儲存於所讀取之單元中之電荷的位 準而定。 圖6說明用於將記憶體單元程式化至目標記憶體狀態之 習知技術。經由耦接字線向記憶體單元之控制閘極施加程 式化電壓vPGM。vPGM為自初始電壓位準VpGMQ開始之階梯 130505.doc •17· 200907975 波形之形式的-系列程式化電壓脈衝。經程式化之單元經 受此系列程式化電屋脈衝,其中每—次進行向浮動問極添 加遞增電荷的嘗試。在程式仆 枉式化脈衝之間,回讀或驗證單元 以判定其相對於斷點位準之源極_汲極電流。回讀過程可 涉及一或多個感應操作。當單元經驗證為達到目標狀態時 停止針對其之程式化。所佶用 、 所使用之私式化脈衝串可具有遞增 之週期或振幅以抵消程式化至記憶體單元之電荷儲存單元 中的累積電子。程式化電路一般而言向選定字線施加一系 列程式化脈衝。以此方式,控制開_接至字線的記憶體 單元之頁可-同經程式化。只要該頁之一記憶體單元被程 式化至其目標狀態,該單元將受到程式化禁止,而其他單 元繼續經受程式化直至該頁之所有單元均經程式化驗證。 使用-系列交替程式化/驗證循環之習知程式化技術係 用來處理程式化過程中之不確定性,其中單元之臨限電壓 最初回應於VPGMi相對較大的改變而快速增長。然而,增 長減慢最後停止,因輕式化至浮動閘極巾之電荷充^ 屏蔽以減小用於進一步將電子穿隧至浮動閘極中的有效電 場。該過程呈現為高度非線性,且因此使用試誤法。 程式化/驗證程式化技術之劣勢在於驗證循環佔用時間 且影響效能。該問題由於能夠儲存多個位元之記憶體單元 的實施而加劇。基本上’需對於記憶體單元之可能多個狀 態中之每-者執行驗證。對於具有16個可能記憶體狀態之 記憶體,此意謂每-驗證步驟會W起至少16個感應操作。 在一些其他機制中,可能甚至還要多若干次。因此,隨著 130505.doc 18 200907975 多位準記憶體單元(”MLC”)中 化/驗證機制之驗證循《得愈加肖耗時Γ。 最二前=有改良程式化效能之記憶體裝置,在” 預測程式化技術 根據本發明之一 _般態樣,在 — 非揮發性記憶體中,_早兀之p歹、 BP „ τ °己隐體早兀可個別地程式化至臨Device Letters, Vol. 21, No. 11, 2000, November, pp. 543-5, 5 pages, and U.S. Patent Nos. 5,768,192 and 6, 〇 11,725. In practice, the memory state of the cell is read by sensing the conduction current across the source and drain electrodes of the cell when a reference voltage is applied to the control gate. Thus, for each given charge on the floating gate of the cell, a corresponding conduction current for the fixed reference control gate voltage can be detected. Similarly, the range of charge that can be ramped to the floating gate defines the corresponding threshold voltage® or corresponding conduction current window. Alternatively, instead of detecting the conduction current in the split current window, at the gate of control 130505.doc •14.200907975, the threshold voltage may be measured for the given memory state (4) and the measured conduction current is lower than or Above the threshold current. In one implementation, #, checking that the conduction current is discharged through the capacitance of the bit line is completed = the lifetime of the threshold current versus the conduction current. Figure 3 illustrates the relationship between the source-drain current 1 〇 and the control gate voltage να for a floating open-pole, optionally, different charge Q1-Q4. The relationship between the four solid lines Id*Vcg indicates the four possible charge levels that can be converted to the floating gate of the memory bank 7L, which correspond to the four possible memory states, respectively. As an example, a threshold voltage window for a certain number of cells can range from 0.5 V to 3·5 ¥. The seven possible memory states t, "Ρ ' " respectively, can be divided into five erased states and six stylized states by dividing the threshold window into five regions at intervals of 0·5 V. 2", "3", "4", "5", "6". For example, if a reference current IREF of 2 μΑ is used as shown, the unit programmed with ... can be regarded as In the memory state "Γ, because its curve intersects Iref in the region of the threshold window by VCG = 0.5 乂 and ^ v. Similarly, Q4 is in the memory state "5". As can be seen from the above description, the more the memory cells are stored, the more finely divided the privacy limits are. For example, the memory device can have a memory cell with a threshold window that varies from -1.5 V to 5 V. This provides a maximum width of 6.5 V. If the memory unit is to be stored in 16 states, each state can occupy 2 paws from the threshold window to 3 〇〇 mV. This will require a higher degree of precision in the stylization and read operations in order to achieve the desired resolution. 130505.doc -15- 200907975 Figure 4 illustrates an example of an array of memory cells. In the memory array 200, the 'per-column memory cell' is connected by a source 14 and a gateless daisy-chain. This design is sometimes referred to as a virtual grounding design... The control gate 3() of the single (4) column is connected to a word line such as word line 42. The source and drain of the cell in the row are connected to selected bit lines such as bit lines 34 and 36, respectively. Figure 5A schematically illustrates a string memory unit organized as a string of wealth. NAND string 50 includes a series of memory transistors M1, M2 ... Mn (e.g., η 4 8 16 or 16). The memory transistors are daisy chained by their source and drain. A pair of select transistors S1, S2 respectively control the connection of the memory transistor chain via the source terminal 54 of the NAND string and the % of the drain terminal to the outside. In the § memory array, when the source selection transistor s ^ is turned on, the source terminal is coupled to the source line (see FIG. 5B) <> similarly, when the drain select transistor S2 is turned on When the 'NAND string is not extremely pure to the bit line of the memory array. Each memory transistor in the chain acts as a memory unit. It has a charge storage element 2 to store a given amount of charge in order to express an undesired state of the memory. The control gate of each memory transistor allows control of read and write operations. As will be seen in Figure 5B, the control gates 3 相应 of the respective memory transistors of the column NAND strings are all connected to the same-word line. Similarly, control gate 32 of each of select transistors S1, 82 provides controlled access to the NAND string via source terminal 54 and gate terminal 56 of the NAND string, respectively. Similarly, the control gates 3 2 of the respective select transistors of a column of NAND strings are connected to the same select line. When the address symmetry in the NAND string is read or verified during the stylization, an appropriate voltage is supplied to its control gate 30. At the same time, the remaining portion of the unresolved 5 mnoreic transistors in the NAND string 5 完全 is fully turned on by applying sufficient voltage to the control gate of the remaining portion of the unaddressed memory transistor in the NAND string 50. . In this way, from the source of the individual memory transistor to the source terminal 54 of the NAND string and likewise from the drain of the individual memory transistor to the terminal of the cell 56, the conduction path is effectively generated. A memory device having the NAND string structure is described in , No. 5, No. 5, No. 5, No. 5, 495, No. 6, No. 46,935. Figure 5B illustrates an example of a NAND array 200 of memory cells composed of NAND strings 5A such as the 2 NAND strings shown in Figure 5A. Along each row of the NAND string, a bit line such as bit line 36 is coupled to the drain terminal 56 of each NAND string. Along each group of NAND strings, a source line such as source line 34 is coupled to source terminal 54 of each NAND string. The control gates of the column § memory cells in a group of NAND strings are also connected to word lines such as word lines 42. The control gates of the selected transistor along a column of one of the NAND strings are coupled to a select line such as select line 44. An entire column of memory cells in the group of NAND strings can be addressed by the word lines of the set of NAND strings and the appropriate voltage on the select lines. When reading one of the memory transistors in the NAND string, the remaining memory cells in the string are hard-on via their associated word lines such that the current flowing through the string is substantially stored in the The level of charge in the cell being read depends on the level of charge. Figure 6 illustrates a prior art technique for staging memory cells to a target memory state. The programmed voltage vPGM is applied to the control gate of the memory cell via the coupled word line. vPGM is the step from the initial voltage level VpGMQ 130505.doc •17· 200907975 The form of the waveform - series of stylized voltage pulses. The stylized unit is subjected to this series of stylized electric house pulses, each of which attempts to add an incremental charge to the floating pole. Between the program sputum pulses, the unit is read back or verified to determine its source _ drain current relative to the breakpoint level. The readback process can involve one or more sensing operations. Stylize for the unit when it is verified that it has reached the target state. The private burst used, which is used, may have an increasing period or amplitude to cancel the accumulated electrons in the charge storage unit that is programmed into the memory unit. A stylized circuit typically applies a series of stylized pulses to a selected word line. In this way, the page of the memory unit that controls the open-to-word line can be programmed. As long as one of the memory cells of the page is programmed to its target state, the cell will be stabilized and the other cells will continue to be stylized until all cells of the page have been programmatically verified. The conventional stylization technique using the -series alternate stylization/verification loop is used to deal with the uncertainty in the stylization process, where the threshold voltage of the unit initially grows rapidly in response to a relatively large change in VPGMi. However, the increase slows down and eventually stops because the charge is reduced to the charge of the floating gate wiper to reduce the effective electric field used to further tunnel electrons into the floating gate. This process appears to be highly non-linear and therefore uses trial and error. The disadvantage of stylized/verified stylization techniques is that it verifies loop time and affects performance. This problem is exacerbated by the implementation of memory cells capable of storing multiple bits. Basically, verification is required for each of a plurality of possible states of the memory unit. For a memory with 16 possible memory states, this means that each verification step will have at least 16 sensing operations. In some other mechanisms, it may even be several more times. Therefore, with the 130505.doc 18 200907975 multi-level memory unit ("MLC") verification/verification mechanism verification, it is more and more time-consuming. The second best = memory device with improved stylized performance, in "predictive stylization technology according to one aspect of the invention, in - non-volatile memory, _ early p歹, BP „ τ ° The invisible body can be programmed to individual

J广立準之範圍中之一者),提供預測需施加何程式化 塗位準以將給定記憶體單S程式化為給定目標臨限電壓 t準之預定函數。以此方式,無需執行驗證操作,藉此大 中田改良程式化操作之效能。 ;在一實施例中’由按比例產生用於給定目標臨限電壓位 準之私式化電壓位準的線性函數來近似得出預定函數。該 線性函數具有由適用於記憶體陣列之—定數目單S的預定 平均值給出之斜率。藉由針對給定記憶體單元預定線性函 數上的檢查點來針對該給定記憶體單元唯一地判定線性函 數。檢查點係基於將記憶體單元程式化至指定臨限電壓位 準之實際程式化電壓。檢查點較佳地對應於記憶體單元之 最低程式化狀態中之一者。最初藉由(例如)習知程式化/驗 證程式化技術來將記憶體單元程式化至檢查點且對其加以 驗證。以此方式,判定對於將記憶體單元程式化至指定記 憶體狀態為必要之實際程式化電壓的檢查點值。因此預定 函數在被用以判定用於將記憶體單元程式化至目標臨限電 壓位準之程式化電壓值之前經校正以產生當在檢查點臨限 130505.doc -19- 200907975 電麼位準處評估時的檢查點程式化電壓值。 f施例中’可在記憶體單元所支援之可能臨限 壓位準的範圍中指金夕乂 ^ 中扎疋夕個檢查點。每一檢查點將用以校正One of J's scopes provides a predetermined function for predicting the application of a stylized coating level to program a given memory single S to a given target threshold voltage. In this way, there is no need to perform verification operations, thereby improving the performance of Dazhongtian's stylized operations. In one embodiment, the predetermined function is approximated by a linear function that proportionally produces a private voltage level for a given target threshold voltage level. The linear function has a slope given by a predetermined average of a fixed number of single Ss for the memory array. The linear function is uniquely determined for the given memory unit by a checkpoint on a predetermined linear function for a given memory unit. The checkpoint is based on the actual stylized voltage that stylizes the memory cells to the specified threshold voltage level. The checkpoint preferably corresponds to one of the lowest stylized states of the memory unit. Memory cells are initially stylized to checkpoints and verified by, for example, conventional stylization/verification stylization techniques. In this way, a checkpoint value is determined for the actual stylized voltage necessary to program the memory unit to the specified memory state. Therefore, the predetermined function is corrected to be used to determine the programmed voltage value for staging the memory unit to the target threshold voltage level to produce a level of accuracy at the checkpoint threshold of 130505.doc -19-200907975 The checkpoint stylized voltage value at the time of evaluation. f In the example, the range of possible thresholds supported by the memory unit may refer to the checkpoint of Jin Xizhen. Each checkpoint will be used to correct

每一檢查點附近之片却hA 局邛預疋函數。使用局部預定函數來 測用於程式化至相關5样 相關聯之檢查點附近的目標臨限電壓位 之程式化電壓位準。The slice near each checkpoint has an hA pre-function. A local predetermined function is used to measure the stylized voltage level used to program the target threshold voltage level near the associated checkpoint.

預測程式化技術為有利的,因為程式化至目標狀態不需 要驗也操作。僅需要驗證操作以判定—般而言在數目上遠 小於可此§己憶體狀態之數目的檢查點。 、將存在使侍預測程式化產生錯誤結果之變化,但此等誤 差將為在統計上可預測之誤差,其可藉由適當錯誤校正: ('E C C")而得到處理。 、圖7說明程式化給定記憶體裝置$之各個記憶體單元的 樣本及所知加程式化電廢與臨限電塵位準之間的觀測得的 ,係。=察到個別記憶體單^之程式化行為甚至在臨限 自大:範圍上亦驚人地線性。陣列内之個別單元或多或 v以相同方式行動,其中程式化臨限電壓之改變大體上與 式化電壓位準之改變成比例。冑因於個別單元受到多深 又抹除及其他® f,個W單元可相差該數目之程式化 脈衝以達到指定檢查點,但個別單元中之每—者之預定函 八有大體上類似之斜率。圖7展示緩慢程式化單元、快 速紅式化單兀中之每—者之實例以及中速程式化單元之- 些實例(實線)。可觀察到不同實例之間斜率上的變化為微 小的。 130505.doc -20- 200907975 圖8不意地說明針 媒士 & 、σ疋°己t體裴置中之記憶體單元的 樣本的圖7所示之斜率八 ^ Q 〇 K . ^ S亥分布基本上顯示出具有 〇.9左右之平均斜率及 你^ 彳1之“準差的正態分布。實際 程式化使用兩個不同程式 * 1 式化電壓步長。已觀測到兩個步長 產生類似的分布及平均斜率。 圖9說明用以提供將 _ _ 、 、f °己隐體早兀程式化至目標臨限電壓 位準所需之程式化電愿的箱中 &的預疋函數之較佳實施例。預定函 數由線性函數近似得出,立φPredictive stylization techniques are advantageous because stylized to target states do not require inspection. Only verification operations are required to determine, in general, checkpoints that are much smaller in number than the number of states that can be used. There will be changes to stagger the wait predictions to produce erroneous results, but such errors will be statistically predictable errors that can be handled by appropriate error correction: ('E C C"). Figure 7 illustrates the observations between the samples of the memory cells of a given memory device $ and the known additions of the stylized electrical waste and the threshold dust level. = The stylized behavior of individual memory singles is observed. Even in the marginal arrogance: the range is also surprisingly linear. Individual cells within the array, either multi or v, act in the same manner, with the change in the programmed threshold voltage being substantially proportional to the change in the voltage level.胄Because the individual units are subjected to deep and erased and other ® f, the W units can differ by the number of stylized pulses to reach the specified checkpoint, but each of the individual units has a substantially similar Slope. Figure 7 shows some examples of slow stylized units, fast redformed units, and medium-speed stylized units (solid lines). It can be observed that the change in slope between different instances is small. 130505.doc -20- 200907975 Figure 8 is not intended to illustrate the slope of the sample of the memory cell in the needle media & σ 疋 己 t 八 八 八 八 ^ ^ ^ . . . . . . 分布Basically, it shows an average slope of about 9.9 and a normal distribution of the “quasi-difference of your ^1. The actual stylization uses two different programs*. The voltage step size has been observed. Two steps have been observed. Similar distribution and average slope. Figure 9 illustrates the pre-complex function of the boxed & required to provide the _ _ , , f ° 隐 隐 兀 兀 至 至 至 至 至 目标 目标 目标Preferred embodiment. The predetermined function is approximated by a linear function, and the vertical φ

、 于® 其中目軲臨限位準VT藉由以下關 係被給定為程式化電壓VpGM之函數: vt (Vpgm)=〈斜率〉VPGM+VT(0) 等式(1) (其中〈斜率 &gt;=avtmvpqm) 相反地, 等式(2), in which the threshold VT is given as a function of the stylized voltage VpGM by the following relationship: vt (Vpgm) = <slope> VPGM + VT(0) Equation (1) (where <slope> ;=avtmvpqm) Conversely, equation (2)

Vpgm(Vt)=1/&lt;斜率 &gt;[ντ_ντ(〇)]; 在較佳實_中’可藉由測試來自類似生產批次之工廢 樣本來預定平均&lt;斜率 &gt;。舉例而言,該測試可得到&lt;斜率&gt; 為〇.9。VT(〇)依賴於單元且在每_單元之預測程式化之前 藉由來自每一記憶體單元之檢查點而加以預定。一旦已知 &lt;斜率 &gt;及VT(0),則界定記憶體單元之預定函數,且等式 (2)可用以獲得程式化至目標臨限電壓位準所需之程式化電 壓位準。 圖10為說明根據本發明之一般實施例而進行之預測程气 化的流程圖。 步驟300 :提供針對在程式化中之記憶體單元的預定函 數’該函數產生依據記憶體單元要被程式化至的臨限電整 130505.doc -21 - 200907975 位準的程式化電壓值。 步驟3 I 〇 ··以且古 程式化至…遞S振幅之程式化電壓將記憶體單元 目^臨限電m位準。經㈣接字線 施加至記情_苗_ , a I。电咬 U體早几之控制閘極。 步驟320:在程式化雷题普傲^ 壓位準處 差實質上已達到由在目標臨限電 元1= 定函數所判定之值之後停止對記憶體單 之頁進f =通常’同時對轉接至同一字線的記憶體單元 討論中之記憶體單元已接收到由在目 隹0限電壓位準處評估之預定函數所判定之值時,呈被林Vpgm(Vt) = 1 / &lt;slope &gt;[ντ_ντ(〇)]; In the preferred real_, the average &lt;slope &gt; can be predetermined by testing the work waste samples from similar production lots. For example, the test yields &lt;slope&gt; is 〇.9. VT(〇) depends on the unit and is predetermined by the checkpoint from each memory unit before the prediction of each unit. Once &lt;slope &gt; and VT(0) are known, the predetermined function of the memory cell is defined, and equation (2) can be used to obtain the programmed voltage level required to program to the target threshold voltage level. Figure 10 is a flow chart illustrating predictive gasification in accordance with a general embodiment of the present invention. Step 300: Providing a predetermined function for the memory unit in the stylization' This function generates a stylized voltage value according to the threshold 130505.doc -21 - 200907975 level to which the memory unit is to be programmed. Step 3 I 〇 ······················································· The (4) word line is applied to the _ _ seedling_, a I. Electric biting U body early control gate. Step 320: In the stylized thunder, the pressure level difference has substantially reached the end of the page of the memory list after the value determined by the target threshold element 1 = the fixed function f = usually 'simultaneously When the memory unit in the memory unit discussion transferred to the same word line has received the value determined by the predetermined function evaluated at the threshold voltage level of the target,

止進行進—步料化,㈣在字線 T 廿,, 于深上可旎存在用於該頁之 Μ記憶體單元之額外程式化脈衝。 Κ …般而’’無需以線性函數來近似得出預定函數。若預 疋函數係用以準確地覆甚 、 — 復琉大範圍之臨限電壓位準,則I可 猎由在工廠測試生產批次而 八 型化。 j疋且籍由某一合適函數來模 一般而言,正被程式化 乏。己隐體皁兀為同時處於程式化 t之一頁類似記憶體單元中一 之者。將存在針對該頁之每 一記憶體單元提供之預定函 貝之母 元共用同一字線,所…:於該頁之所有記憶體單 —已糟由預測程式化電壓程式化 戎頁之一記憶體單元,兑钟 一就被禁止進行進一步程式化。 圖1 〇中所說明之預測崧4 〈預測%式化模式較佳地實施於 體陣列200之記憶體操作的 心 圖&quot;中。 路U。中之狀態機112(見 圖11說明隨圖9所示之預定 疋函數之权正繼之以其在預測 130505.doc -22- 200907975 程式化模式中之應用時的程式化電壓。 在最初階段中,記憶體單元之檢查點(〇)經指定處於稍 间於被視為與抹除狀態相關聯之電壓位準的臨限電壓位準 (私查點臨限電壓位準)處。施加一系列遞增之程式化電壓 脈衝以朝向檢查點臨限電壓位準程式化記憶體單元。程式 化杈式可為父替程式化與驗證直至程式化驗證了檢查點臨 限電屢位準之習知模式。一旦已知針對檢查點⑼之座標系 [VPGM,VT]檢u⑼,以等式(2)之形式之預定函數(見圖9)即 可針對VT(〇)得到求解且經完全規定。 在規定等式(2)之形式之預定函數之後,隨後可使用預 定函數在預測模式中程式化記憶體單元以提供針對目標臨 限電Μ位準或針對目標記憶體狀態之估計程式化電壓位 準。在較佳實施例中,調整程式化電壓步長以使得每一額 外脈衝將程式化記憶體單元至下一記憶體狀態。以具㈣ 個可能記憶體狀態之記憶體單元為例,脈衝大小可為則 ‘以此方式’ _額外脈衝將程式化記憶體至狀態⑴, 另—額外脈衝將程式化記憶體至狀態⑺等等。因此,程式 化至給定記憶體狀態可經簡化為自狀態(〇)計數狀態之數目 且供應相同數目之脈衝。舉例而言,可在狀態⑼下設定一 次旗標,且其後可以與目標狀態距狀態(〇)的狀態數目相同 之數目的脈衝來程式化記憶體單元。 其他程式化脈衝大小為可能的。舉例而言,對於具有Μ 個可能記憶體狀態之記憶體單元,脈衝大小可為&quot;Ο 心。在彼情況下’將花費兩個脈衝來自—記憶體狀態程 130505.doc -23- 200907975 式化至下一鄰接記憶體狀態。此將提供程式化 之解析度,其在使用自目庐碎w 用的目&amp;限之餘裕的-些實施中為有 圖12示意地說明在圖u所示 …玄 所不之程式化過程期間,記憶體 早几之-頁之各種記憶體狀態的分布。記憶體單元之頁以 ;有記憶體單元處於抹除狀態中開始,該等記憶體單元可 一 在取初程式化階段期 二 驗證循環(例如,X個程式化脈衝加上 ^個驗證步驟之總數)將記憶體單元自抹除狀態程式化至 態(〇)° 一般而言’每—記憶體單元之碰此獨立。一旦 2憶體單元處於狀_中,預測程式化模式即開始,且每 一額外脈衝將程式化記憶體單元至下―記憶體狀態。 圖13為將圖1〇所示之步驟句日日泛— τ ^ _ 〇兒月為包括使用檢查點校 針對S己憶體單元之預定函數的流程圖。 化f、指定函數之檢查點作為可由相應檢查點程式 電墊值知式化的指定檢查點臨限電壓位準。 步驟304 :藉由交替栽十/ ^ ^ 式化…驗si直至程式化驗證了檢 查點臨限《位準而判定相應檢查點程式化電屢值。 步驟306 :在預定函數被 ,^ 被用以判疋用於將記憶體單元程 式化至目標臨限電壓位準 叙_ #式化電壓值之前校正預定函 數以產生當在檢查點臨限 化電塵值。. Μ處评估時的檢查點程式 圖14說明使用—個以 墙电 彳欢查2之預測程式化之實施例的 例在第—實例中,待由預測模式程式化之每一記 130505.doc •24- 200907975 隐體狀恶之前為在檢查點模 在圖&quot;中於步驟302及步驟;。4中之記憶體狀態。 漸將記憶體單元程式化為狀態(〇)、狀離=模式^若可逐 則可將偶數狀態狀態(〇)、狀〜、)、狀癌⑺…’ 點。以程式化/驗證程式化 等狀態。自每一檢查、:(見圖6)自先前狀態到達此 測方切切了 — 數可得到校正且用以以預 態(3)、狀態⑺…。狀-例如’奇數的狀態⑴、狀 =中所說明之實施例允許以預測模式 由”二: 以如此短之間隔顯示出良好線性。 =:查點模式僅程式化每隔—個記憶 免^知程式化(見圖6)中之至少一半驗證操作。 圖15說明圖14中所說 之各種相關聯預定函數的方;:: = :檢查點附近 數為線性的,且具有相同預定平::丰察到^ ^ ^ ^ ίΐ - Τ 々斜率。檢查點(0)用以設 函數400之邊界條件。類似地,檢 »又疋局部預定函數420之邊界條 — 部預定函數44G之邊界條件等等。-點㈣以設定局 -旦局部預定函數經設定,其即可 憶體狀態之程式化雷廢办、、隹m 用於下一圮 以產U 準。因此,局部職函數用 於將單元程式化至狀_之程式化電 用 :=:°用以產生用於將單元程式化至狀態心 \且局部預定函數440用以產生用於將單元 I30505.doc -25- 200907975 程式化至狀態(5)之程式化電壓位準等等。 圖I6說明使用一個以上檢查點之預測程式化之實施例的 第二實例。與圖14及圖丨5中所示之第一實例形成對比,第 一實例具有母四個記憶體狀態地指定之檢查點。因此,於 狀態(0)、於狀態(4)、於狀態(8)等等指定檢查點(〇)。由檢 查點(0)設定之局部預定函數將用以分別針對下三個記憶體 狀態(即,狀態(1)、狀態(2)及狀態(3))預測程式化電壓。 類似地,由檢查點(4)設定之局部預定函數將用以分別針對The advancement is performed, and (4) at the word line T 廿, the extra stylized pulses for the memory cells of the page are present in the depth. Κ ...like '' does not need to approximate the predetermined function by a linear function. If the pre-function is used to accurately cover and reproduce a wide range of threshold voltage levels, then I can be modeled by testing the production batch at the factory. j疋 and by a suitable function, in general, it is being stylized. The cryptic saponin is one of the memory cells of the one page of the stylized t. The parent elements of the predetermined letter provided for each memory unit of the page share the same word line, ...: all memory sheets on the page - one of the memories of the stylized page of the predicted stylized voltage The body unit is forbidden for further stylization. The prediction 嵩4 described in Fig. 1 (predictive % mode is preferably implemented in the heartbeat of the memory operation of the volume array 200). Road U. The state machine 112 (see Figure 11 illustrates that the weight of the predetermined 疋 function shown in Figure 9 is followed by the stylized voltage when it is applied in the predicted 130505.doc -22-200907975 stylized mode. In the initial phase The memory cell checkpoint (〇) is specified at a threshold voltage level (private checkpoint threshold voltage level) that is slightly considered to be the voltage level associated with the erased state. The incremental stylized voltage pulse stylizes the memory unit towards the checkpoint threshold voltage level. The stylized style can be used for parental stylization and verification until the stylization verifies the checkpoint threshold. Once the coordinate system [VPGM, VT] is detected for the checkpoint (9), u(9), the predetermined function in the form of equation (2) (see Figure 9) can be solved for VT(〇) and is fully specified. After specifying a predetermined function in the form of equation (2), the memory unit can then be programmed in the prediction mode using a predetermined function to provide an estimated programmed voltage level for the target threshold level or for the target memory state. In the preferred embodiment Adjusting the stylized voltage step size so that each extra pulse will stylize the memory cell to the next memory state. For example, in a memory cell with (four) possible memory states, the pulse size can be 'in this way' ' _ extra pulse will program the memory to state (1), another - extra pulse will be stylized memory to state (7), etc. Therefore, stylized to a given memory state can be simplified to self-state (〇) count state The number and supply of the same number of pulses. For example, the flag can be set once in state (9), and thereafter the memory cell can be programmed with the same number of pulses as the number of states of the target state state (〇). Stylized pulse size is possible. For example, for a memory cell with one possible memory state, the pulse size can be &quot;Ο. In that case, 'will take two pulses from the memory state. 130505.doc -23- 200907975 Modification to the next adjacent memory state. This will provide a stylized resolution, which is used in the case of using the target &amp; In the implementation, FIG. 12 schematically illustrates the distribution of various memory states of the memory-pages during the stylization process shown in Fig. u. The pages of the memory cells are; Starting in the erase state, the memory cells can be self-erased by the memory cell unit during the initial programming phase of the second verification cycle (eg, X stylized pulses plus the total number of verification steps). To the state (〇) ° In general, the 'per-memory cell's collision is independent. Once the 2 memory cell is in the state, the predictive stylization mode begins, and each extra pulse will program the memory cell to The following is a memory state. Fig. 13 is a flow chart showing the step-by-day _ _ ^ _ 〇 月 月 〇 包括 包括 包括 包括 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 The checkpoint of the specified function is used as the specified checkpoint threshold voltage level that can be learned by the corresponding checkpoint program pad value. Step 304: Determine the stylized electric power value of the corresponding checkpoint by alternately planting the decimal/testing method until the stylization verifies the checkpoint threshold. Step 306: After the predetermined function is used, the predetermined function is corrected before the program is used to program the memory unit to the target threshold voltage level to generate a power limit at the checkpoint. Dust value. Checkpoints at the time of assessment Figure 14 illustrates an example of an example of using a predictive stylization of wall-walled check 2 In the first instance, each record to be stylized by the predictive mode 130505.doc • 24-200907975 Before the hidden form is in the checkpoint in the figure &quot; in step 302 and step; Memory state in 4. Gradually, the memory unit is programmed into a state (〇), a form of separation = mode ^, if it can be, the even state state (〇), shape ~,), cancer (7) ... point. Stylized/verified stylized, etc. From each check, : (see Figure 6), the test is cut from the previous state - the number can be corrected and used to predict (3), state (7).... The embodiment described in the example - the odd state (1), the shape = allows the prediction mode to be "two: show good linearity at such short intervals. =: the enumeration mode is only stylized every other memory free ^ Know at least half of the verification operations in the stylization (see Figure 6). Figure 15 illustrates the sides of the various associated predetermined functions described in Figure 14;:: = : The number near the checkpoint is linear and has the same predetermined flat: : 丰察到^ ^ ^ ^ ίΐ - Τ 々 slope. Checkpoint (0) is used to set the boundary condition of function 400. Similarly, check the boundary condition of local predetermined function 420 - boundary condition of predetermined function 44G Etc. - Point (4) is set by setting the local predetermined function, which can be used to recall the stylized state of the body, and 隹m is used for the next level to produce U. Therefore, the local job function is used. Stylize the unit to a stylized version: =: ° is used to generate the unit for stylization to the state heart \ and a local predetermined function 440 is used to generate the unit I30505.doc -25 - 200907975 Stylized voltage level to state (5), etc. Figure I6 illustrates the use of a A second example of the embodiment of the predictive stylization of the above checkpoints. In contrast to the first example shown in Figures 14 and 5, the first instance has a checkpoint designated by the parent four memory states. The checkpoint (〇) is specified in state (0), state (4), state (8), etc. The local predetermined function set by checkpoint (0) will be used to respectively target the next three memory states (ie , state (1), state (2), and state (3)) predict the stylized voltage. Similarly, the local predetermined function set by checkpoint (4) will be used to target

下三個記憶體狀態(即,狀態(5)、狀態(6)及狀態預測 程式化電壓。由此第二實例產生之所預測程式化電壓將不 如第一實例的所預測程式化電壓準確,但在許多應用中可 為足夠的。其具有進一步減少程式化驗證操作之數 勢。 指定檢查點在記憶體單元之臨限f中之臨限值的範圍令 之其他變化為可能的。可視效能與精確性之間的平衡進行 言,在—些程式化演算法中,記憶體單元之頁 於第一次通過中被程式化為接近其各別目標狀態 程式化通過將完成至各別目標狀態之程式 _人通過方法來減輕㈣記憶體單元之 化干擾。因A^ a 】《之間的程 ^ 因為卓—次通過程式化不要求如第二次 尚之精確性,所以可鋅 、過一: 次通過以節料門Γ 程式化模式來執行第. … 在—些實施中,亦預期第二 u肖L式化模式’較佳地使用適當位置處之^ 130505.doc -26 - 200907975 檢查點。 預測程式化模式不排除超出^ ^ ^ ^ ^ ^ ^ ^ α加ι限電壓位準之非常徼 小的可能性(估計為小於〇 1。/ ) 勹j m/ο。在超出目標記憶體狀態之 情況下’可藉由所實施之錯誤校正竭來校正錯誤。 、,當兩個檢查點可用時,有可能獨立地設定預定函數之斜 率。一旦如此規定預定函數,豆 ,、就可用以產生用於後續記 憶體狀態之程式化電壓位準。The next three memory states (ie, state (5), state (6), and state prediction stylized voltage. The predicted stylized voltage generated by the second instance will not be as accurate as the predicted stylized voltage of the first example, However, it can be sufficient in many applications. It has the potential to further reduce the number of stylized verification operations. It is possible to specify other checkpoints within the threshold of the memory unit's threshold f. In contrast to the accuracy, in some stylized algorithms, the pages of the memory unit are programmed to be close to their respective target states in the first pass, stylized through the completion to the respective target state. The program _ people use methods to alleviate (4) memory unit interference. Because A ^ a 】 "between the process ^ because the Zhuo - times through the stylization does not require the second time accuracy, so zinc, over one : The second pass through the throttling threshold stylized mode to perform the .... In some implementations, it is also expected that the second u-L mode will be better used at the appropriate position ^ 130505.doc -26 - 200907975 check Point. The stylized mode does not exclude the very small possibility of exceeding the ^ ^ ^ ^ ^ ^ ^ ^ α plus ι limit voltage level (estimated to be less than 〇1. / ) 勹jm/ο. Exceeding the target memory state In the case, the error can be corrected by the error correction implemented. When two checkpoints are available, it is possible to independently set the slope of the predetermined function. Once the predetermined function is specified, the bean can be used to generate Stylized voltage level for subsequent memory states.

本文_所引用之所有專利、 規範、其他公開案、文件及事 用的方式併入本文十。就所併 之任一者與本文件之本文之間 不一致性或矛盾而言,本文件 主導地位。 專利申請案、文章、書籍、 物用於所有目的而全部以引 入之公開案、文件或事物中 的術3吾之定義或使用的任何 令的術語之定義或使用應居 儘管已關於某些實施例描述了本發明之各種態樣,但應 理解’本發明有權在所㈣請專利範圍之完整料 保護。 力All patents, specifications, other publications, documents, and matters cited herein are incorporated herein by reference. This document is dominant in terms of inconsistencies or inconsistencies between any of them and the text of this document. The application or use of a patent application, article, book, or article for all purposes, all of which is defined in the disclosure, document, or thing, or the use or definition of any term used, although in relation to certain implementations The examples describe various aspects of the invention, but it should be understood that 'the invention is entitled to the complete material protection of the scope of the patents. force

【圖式簡單說明】 圖1示意地說明本發明可實施於之非揮發性記憶體晶片 之功能區塊。 圖2示意地說明非揮發性記憶體單元。 圖3說明對於浮動閘極於任一時間可選擇性地儲存之四 個不同電荷Q1_Q4的源極-汲極電流Id與控制閘極電壓Vcg 之間的關係。 圖4說明記憶體單元之;^〇11陣列之一實例。 130505.doc -27- 200907975 圖5 A示意地說明經組織化為NAND串之一串記憶體單 元0 圖5B說明由諸如圖5A中所示之NAND串的NAND串5〇組 成之記憶體單元之NAND陣列200的實例。 圖6說明用於將§己憶體單元程式化至目標記憶體狀熊之 習知技術。 圖7說明程式化給定記憶體裝置中之各個記憶體單元的BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a schematic illustration of functional blocks of a non-volatile memory chip that can be implemented in accordance with the present invention. Figure 2 schematically illustrates a non-volatile memory unit. Figure 3 illustrates the relationship between the source-drain current Id and the control gate voltage Vcg for four different charges Q1_Q4 that are selectively stored by the floating gate at any one time. Figure 4 illustrates an example of an array of memory cells; 130505.doc -27- 200907975 FIG. 5A schematically illustrates one-string memory unit 0 organized as a NAND string. FIG. 5B illustrates a memory unit composed of a NAND string 5〇 such as the NAND string shown in FIG. 5A. An example of a NAND array 200. Figure 6 illustrates a conventional technique for staging a § memory unit to a target memory bear. Figure 7 illustrates the stylization of individual memory cells in a given memory device

樣本及所施加程式化電壓與臨限電壓位準之間的觀測得的 關係。 圖8示意地說明針對給定記憶體裝置中之記憶體單元的 樣本的圖7所示之斜率的分布。 圖9說㈣以提供將記憶體單元程式化至目標臨限電壓 位準所需之程式化電壓的預定函數之較佳實施例。 圖10為說明根據本發明之一般實施例而進行之預測程式 化的流程圖。 圖11說明隨圖9所示之狀函數之校正繼之以其在預測 程式化模式中之應用時的程式化電壓。 圖12示意地說明在圖丨丨所千斗 dd _ 圚所不之程式化過程期間,記憶體 單元之一頁之各種記憶體狀態的分布。 圖13為將圖1〇所示之步驟 疋γ驟3〇〇5兒明為包括使用檢查點校 正針對記憶體單元之預定函數的流程圓。 圖1 4說明使用一個以上檢杳 ,^ 义預測私式化之實施例的 第一實例。 以校正檢查點附近 圖15說明將圖14中所說明之檢查點用 130505.doc -28- 200907975 之各種相關聯預定函數的方式。 圖1 6說明使用一個以上檢查點之預測程式化之實施例的 第二實例。 【主要元件符號說明】 10 記憶體單元 14 源極 16 汲極 20 電荷儲存單元 30 控制閘極 32 控制閘極 34 位元線/源極線 36 位元線 42 字線 44 選擇線 50 NAND 串 54 源極端子 56 汲極端子 100 記憶體晶片 110 控制電路 112 狀態機 200 NAND陣歹ij 230A 列解碼器 230B 列解碼器 231 資料I/O匯流排 130505.doc - 29 - 200907975 250A 區塊多工器/頁多工器 250B 區塊多工器/頁多工器 260A 行解碼器 260B 行解碼器 270A 讀取寫入電路 270B 讀取寫入電路 400 局部預定函數 420 局部預定函數 440 局部預定函數 Id 源極-汲極電流 Iref 參考電流 Ml、M2&quot;. Μη 記憶體電晶體 Q1-Q4 電荷 SI 源極選擇電晶體 S2 汲極選擇電晶體 VcG 控制閘極電壓 VpGM 程式化電壓 VpgMO 初始電壓位準 VT 目標臨限位準 130505.doc -30-The observed relationship between the sample and the applied stylized voltage and the threshold voltage level. Figure 8 is a schematic illustration of the distribution of the slopes shown in Figure 7 for a sample of memory cells in a given memory device. Figure 9 illustrates (4) a preferred embodiment of a predetermined function for providing a stylized voltage required to program a memory cell to a target threshold voltage level. Figure 10 is a flow chart illustrating predictive programming performed in accordance with a general embodiment of the present invention. Figure 11 illustrates the correction of the shape function as shown in Figure 9 followed by its stylized voltage when applied in predictive stylized mode. Fig. 12 is a view schematically showing the distribution of various memory states of one page of a memory cell during the stylization process of the figure dd _ 圚. Fig. 13 is a flow chart showing the step 疋 γ 〇〇 3 〇〇 5 shown in Fig. 1 为 as including a predetermined function for correcting the memory unit using the checkpoint. Figure 14 illustrates a first example of an embodiment that uses more than one check to predict privateization. Referring to the vicinity of the calibration checkpoint, Fig. 15 illustrates the manner in which the checkpoints illustrated in Fig. 14 are associated with various associated predetermined functions of 130505.doc -28-200907975. Figure 16 illustrates a second example of an embodiment of predictive stylization using more than one checkpoint. [Main component symbol description] 10 Memory unit 14 Source 16 Deuterium 20 Charge storage unit 30 Control gate 32 Control gate 34 Bit line/source line 36 Bit line 42 Word line 44 Select line 50 NAND string 54 Source terminal 56 汲 terminal 100 memory chip 110 control circuit 112 state machine 200 NAND array 歹 230A column decoder 230B column decoder 231 data I / O bus 130505.doc - 29 - 200907975 250A block multiplexer /Page multiplexer 250B Block multiplexer / page multiplexer 260A Row decoder 260B Row decoder 270A Read write circuit 270B Read write circuit 400 Local predetermined function 420 Local predetermined function 440 Local predetermined function Id source Pole-drain current Iref Reference current Ml, M2&quot;. Μη Memory transistor Q1-Q4 Charge SI Source select transistor S2 Dip select transistor VcG Control gate voltage VpGM Stylized voltage VpgMO Initial voltage level VT target The threshold level 130505.doc -30-

Claims (1)

200907975 十、申請專利範圍: I .200907975 X. Patent application scope: I. 種在—非揮發性記‘丨奋體φ沾竹 6 u體中的將一記憶體單元程式化至 一目標臨限電壓位準之方 00 万法’該非揮發性記憶體具有記 憶體早元之一陣列,装由 兵甲该專記憶體單元可個別地程式 化至臨限電壓位準 — 平(乾圍中之一者,該方法包含: 提供針對處於程式化 Λ化中之一記憶體單元的—預定函 數,3亥函數依據一記憶體置^_ 、上 早凡要被程式化至的臨限電壓 位準而產生程式化電壓值; 化電壓將該記憶體單元程式 及 達到由在該目標臨限電壓位 定之一值之後停止對該記憶 以一具有遞增振幅之程式 化至該目標臨限電壓位準; 在該程式化電壓實質上已 準處評估之該預定函數所判 體單元的程式化。A non-volatile memory of a non-volatile memory that is programmed into a target threshold voltage level in a non-volatile memory. An array of armored cells that can be individually programmed to a threshold voltage level - one of the dry circumferences, the method comprising: providing a memory for one of the stylized deuterations The unit-predetermined function, the 3H function generates a stylized voltage value according to a threshold of a memory that is programmed to the threshold voltage level; the voltage is programmed to achieve the memory unit After the target threshold voltage is set to a value, the memory is stopped to be programmed to the target threshold voltage level with an increasing amplitude; the predetermined function is determined by the predetermined voltage in the programmed voltage. Stylized. 2 ·如晴求項1之方法, 為一系列電壓脈衝。 3 ·如請求項1之方法, 數。 其中具有遞增振幅之該程式化電壓 其中該預定函數實質上為一線性函 (如之方法,其中該提供_狀函數包括: p定/函數 &lt; 檢查點作為可由—相應檢查點程式化 電壓值程式化的一指定檢查點臨限電壓位準; 藉由\ #程式化與驗證直至程式化驗證了該檢查點臨 限電壓位準而判定該相應檢查點程式化電壓值;及 在該預疋函數被用以判定用於將該記憶體單元程式化 至〇目‘ gDn限電壓位準之_程式化㈣值之前校正該預 130505.doc 200907975 數而產生當在該檢查點臨限電壓 檢查點程式化電壓值。 皁處評估時的該 ★明求項4之方法’其中該目標臨限電壓位 該檢查點臨限電壓料之—預定電壓。以於來自 6. 如a求項4之方法’其中在臨限 定:或多個檢查點臨限電壓位準。彳之錢圍中指 7. 如„月求項4之方法,其中該 估計斜率及—檢查點而界定。數為線性的,且由— f 1=:4之方法’其中該預定函數實質上為線性的, 至 &gt; 兩個檢查點而界定。 9’ 項1之方法,其中該非揮發性記憶體具有個別地 :子個以上位元之資料之記憶體單元。 10·項1之方法’其中該非揮發性記憶體具有將資料 個別地儲存為程式化至一電荷儲存元件中的量之 記憶體單元。 U·如請求項1之方法 體之—浮動閘極。 12. 如請求項1之方法 體中之~介電層。 13. 如明求項丨之方法,其中該非揮發性記憶體具有含有一 NAND結構之記憶體單元。 14. 如請求们之方法,其中該非揮發性記憶體為一快閃 EEPR〇m 〇 1 5.如凊求項i之方法,其中該非揮發性記憶體實施於一記 其中该電荷儲存元件為一場效電晶 其中該電荷儲存元件為一場效電晶 130505.doc 200907975 憶體卡中。 6. 士印求項丨之方法,其中該非揮發性記憶體嵌埋於一計 鼻裝置中。 17·如請求項1至16中任1之方法,其中該記憶體單元為 同綾程式化之一組記憶體單元中之一者。 18. —種非揮發性記憶體,其包含: α己隱體單7C之一陣列,其中該等記憶體單元可個別地 程式化至臨限電壓位準之一範圍中之一者; 。貝取寫入電路,其用於並行讀取及程式化記憶體單元 之一頁; 狀L、从,其用於控制包括該等讀取寫入電路之該非 揮發性記憶體的操作,該狀態機: 提供針對處於程式化中之一記憶體單元的一預定函 該函數依據一 5己憶體單元要被程式化至的臨限電壓 位準而產生程式化電壓值; 控制以一具有遞增振幅之程式化電壓將該記憶體單元 程式化至目標臨限電壓位準;及 在該程式化電壓實質上已達到由在該目標臨限電壓位 準處評估之.該狀函數所判定之—值之後停止對該記憶 體單元的程式化。 一 19. 如請求項18之非揮發性記憶體,其中具有遞增振幅之該 程式化電壓為一系列電壓脈衝。 20. 如請求項18之非揮發性記憶體’其中該預定函數實質上 為一線性函數。 130505.doc 200907975 η·如請求項18之非揮發性記憶體,其中提供―㈣函數之 §亥狀態機進一步包括: 指定該函數之一檢杳駄施a 笪點作為可由一相應檢查點程式化 電壓值程式化的一指定檢查點臨限電壓位準; 藉由交替程式化與驗證直至程式化驗證了該檢查點臨 限電麼位準而判定該相應檢查點程式化電麼值;及 在該預定函數被用以判定用於將該記憶體單元程式化 ^該目標臨限電μ位準之-程式化電塵值之前校正該預 疋函數而產生當在該檢查點臨限f屬位準處評估時的該 檢查點程式化電壓值。 22. 如請求項21之非揮發性記憶體,其中該目標臨限電壓位 準小於來自該檢查點臨限電壓位準之—預定電壓。 23. 如請求項21之非揮發性記憶體,纟中在臨限電壓位準之 該範圍中指定一或多個檢查點臨限電壓位準。 24. 如請求項21之非揮發性記憶體,其中該預定函數為線性 的,且由一估計斜率及一檢查點而界定。 25. 如請求項21之非揮發性記憶體’其中該預定函數實質上 為線性的,且由至少兩個檢查點而界定。 26. 如請求項18之非揮發性記憶體,其中該非揮發性記憶體 具有個別地儲存一個以上位元之資料之記憶體單元。〜 27. 如請求項18之非揮發性記憶體’其中該非揮發性記憶體 具有將資料個別地儲存為程式化至—電荷儲存元件中的 一電荷量之記憶體單元。 28. 如請求項18之非揮發性記憶體,其中該電荷儲存元件為 130505.doc 200907975 一場效電晶體之一浮動閘極。 29. 如請求項18之非揮發性記憶體,其中該 %刊仔凡件為 一场效電晶體中之一介電層。 30. 如請求項18之非揮發性記憶體,其中該非揮發性記憶體 具有含有一 NAND結構之記憶體單元。 3 1.如§青求項1 8之非揮發性兮ρ彳音體,豆φ兮t +货旺。己U骽具宁5亥非揮發性記憶體 為一快間EEPROM。 r 32. 如請求項18之非揮發性記憶體,其中該非揮發性記憶體 實施於一記憶體卡中。 &quot; 33. 如請求項18之非揮發性記憶體,其中該非揮發性記憶體 嵌埋於一計算裝置中。 34. 如請求項18至33中任一項之非揮發性記憶體,其中該記 35 憶體單元為一同經程式化之一組記憶體單元中之—者。 一種非揮發性記憶體,其包含: 記憶體單元之―陳列,让由 哔列其中该荨記憶體單元可個別地 程式化至臨限電壓位準之一範圍中之一者,· 用於提供針對處於程式化中之—記憶體單元的一預定 函數之構件’該函數依據-記憶體單元要被程式化至臨 限電壓位準而產生程式化電壓值; „用於控制以一具有遞增振幅之程式化電屢將該記憶體 早騎式化至目標臨限電壓位準之構件;及 用於在該程式化電壓眘暂^ &amp; ^ . 質上已達到由在該目標臨限電 呓_ 数所匈疋之一值之後停止對該 。己U體早几的程式化的構件。 I30505.doc 200907975 36.如請求項35之非揮發性記憶體,其進一步包含: 用於指定該函數之-檢查點作為可由一相應檢查點程 式化電壓值程式化的一指定檢查點臨限電壓位準之構 件; 用於藉由父替程式化與驗證直至程式化驗證了該檢查 點臨限電壓位準而判定該相應檢查點程式化電壓值之構 件;及 用於在該預定函數被用以判定用於將該記憶體單元程 式化至該目標臨限電壓位準之一程式化電壓值之前校正 該預定函數而產生當在該檢查點臨限電壓位準處坪估時 的該檢查點程式化電壓值之構件。 3 7.如請求項34或3 5之非揮發性記憶體,其中該記憶體_ &quot;Τ' 7Q 為一同經程式化之一組記憶體單元中之一者。 130505.doc2 · The method of claim 1, is a series of voltage pulses. 3 · As requested in item 1, the number. Wherein the stylized voltage having an increasing amplitude, wherein the predetermined function is substantially a linear function (such as the method, wherein the providing _ function includes: p-fix/function &lt; checkpoint as a usable - corresponding checkpoint stylized voltage value Stylized a specified checkpoint threshold voltage level; the programmed checkpoint voltage value is determined by \#stylization and verification until the checkpoint verifies the checkpoint threshold voltage level; and in the preview The function is used to determine the number of the stylized (four) values used to program the memory unit to the 'gDn limit voltage level' before the correction of the pre-130505.doc 200907975 number is generated at the checkpoint threshold voltage checkpoint Stylized voltage value. The method of determining the liquid at the time of soap evaluation, wherein the target threshold voltage is the predetermined voltage of the checkpoint, and the predetermined voltage is obtained from the method of 6. 'Where in the limit: or multiple checkpoints, the threshold voltage level. 彳之钱围中指7. As in the method of the monthly claim 4, where the estimated slope and - checkpoint are defined. The number is linear, and by The method of f 1=:4, wherein the predetermined function is substantially linear, and is defined by two checkpoints. The method of item 1, wherein the non-volatile memory has individual: more than one bit The memory unit of the data. The method of item 1 wherein the non-volatile memory has a memory unit that individually stores the data as an amount programmed into a charge storage element. U. The method of claim 1 The floating gate of the body. 12. The dielectric layer of the method body of claim 1. 13. The method of claim 1, wherein the non-volatile memory has a memory cell containing a NAND structure. The method of claimant, wherein the non-volatile memory is a flash EEPR 〇m 〇1 5. The method of claim i, wherein the non-volatile memory is implemented in a record in which the charge storage element is a field effect The electric crystal is in which the charge storage element is in the field of a photoelectric crystal 130505.doc 200907975. 6. The method of the invention, wherein the non-volatile memory is embedded in a nasal device. Items 1 to 16 The method of claim 1, wherein the memory unit is one of a group of memory units that are stylized. 18. a non-volatile memory comprising: an array of alpha-clone single 7C, wherein The memory cells can be individually programmed to one of a range of threshold voltage levels; a write circuit for reading and programming one page of memory cells; L, slave, It is used to control the operation of the non-volatile memory including the read write circuits, the state machine: providing a predetermined function for one of the memory cells in the stylization, the function is based on a 5 memory unit Stylized to a threshold voltage level to generate a programmed voltage value; control to program the memory unit to a target threshold voltage level with a stylized voltage having an increasing amplitude; and at the stylized voltage substantially The stylization of the memory unit is stopped after the value determined by the function is evaluated at the target threshold voltage level. 19. The non-volatile memory of claim 18, wherein the stylized voltage having an increasing amplitude is a series of voltage pulses. 20. The non-volatile memory of claim 18 wherein the predetermined function is substantially a linear function. 130505.doc 200907975 η · The non-volatile memory of claim 18, wherein the "fourth" function providing the "(four) function further comprises: specifying one of the functions to detect a point as a program can be stylized by a corresponding checkpoint The voltage value is programmed to a predetermined checkpoint threshold voltage level; the programmed value of the corresponding checkpoint is determined by alternately stylizing and verifying until the checkpoint verifies the checkpoint threshold level; and The predetermined function is used to determine that the memory unit is programmed to correct the pre-twist function before the stylized electric dust value is generated. The checkpoint stylized voltage value at the time of the assessment. 22. The non-volatile memory of claim 21, wherein the target threshold voltage level is less than a predetermined voltage from the checkpoint threshold voltage level. 23. In the non-volatile memory of claim 21, one or more checkpoint threshold voltage levels are specified in the range of threshold voltage levels. 24. The non-volatile memory of claim 21, wherein the predetermined function is linear and is defined by an estimated slope and a checkpoint. 25. The non-volatile memory of claim 21 wherein the predetermined function is substantially linear and is defined by at least two checkpoints. 26. The non-volatile memory of claim 18, wherein the non-volatile memory has a memory unit that individually stores data for more than one bit. ~ 27. The non-volatile memory of claim 18 wherein the non-volatile memory has a memory unit that individually stores the data as a charge amount programmed into the charge storage element. 28. The non-volatile memory of claim 18, wherein the charge storage element is 130505.doc 200907975 one floating gate of one effect transistor. 29. The non-volatile memory of claim 18, wherein the % is a dielectric layer of a potential transistor. 30. The non-volatile memory of claim 18, wherein the non-volatile memory has a memory cell comprising a NAND structure. 3 1. If the non-volatile 兮ρ彳 sound body of § Qing seeking item 18, beans φ兮t + goods Wang. It has a non-volatile memory for a fast EEPROM. r 32. The non-volatile memory of claim 18, wherein the non-volatile memory is implemented in a memory card. &quot; 33. The non-volatile memory of claim 18, wherein the non-volatile memory is embedded in a computing device. 34. The non-volatile memory of any one of claims 18 to 33, wherein the memory unit is one of a group of memory units that are programmed together. A non-volatile memory comprising: a display of a memory cell, wherein the memory cell can be individually programmed to one of a range of threshold voltage levels, for providing For a component of a predetermined function in the stylized memory unit, the function generates a stylized voltage value according to the memory unit to be programmed to a threshold voltage level; „for control with an increasing amplitude The stylized power repeatedly rides the memory to the target threshold voltage level component; and is used in the stylized voltage caution ^ &amp; ^ . qualitatively reached by the target _ After the value of one of the Hungarians, the stylized component of the U-body is stopped. I30505.doc 200907975 36. The non-volatile memory of claim 35, further comprising: for specifying the function a checkpoint as a component of a specified checkpoint threshold voltage level that can be programmed by a corresponding checkpoint stylized voltage value; used for stylization and verification by the parent until the stylization verifies the checkpoint a component for determining a programmed check voltage value of the corresponding checkpoint; and a programmed voltage value for determining that the predetermined function is used to program the memory cell to the target threshold voltage level The predetermined function is previously corrected to produce a component of the checkpoint stylized voltage value when the checkpoint threshold voltage level is evaluated. 3 7. Non-volatile memory as claimed in claim 34 or 35, wherein The memory _ &quot;Τ' 7Q is one of a group of memory units that are programmed together. 130505.doc
TW097113094A 2007-04-10 2008-04-10 Non-volatile memory and method for predictive programming TWI371041B (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US11/733,706 US7551483B2 (en) 2007-04-10 2007-04-10 Non-volatile memory with predictive programming
US11/733,694 US7643348B2 (en) 2007-04-10 2007-04-10 Predictive programming in non-volatile memory

Publications (2)

Publication Number Publication Date
TW200907975A true TW200907975A (en) 2009-02-16
TWI371041B TWI371041B (en) 2012-08-21

Family

ID=39712596

Family Applications (1)

Application Number Title Priority Date Filing Date
TW097113094A TWI371041B (en) 2007-04-10 2008-04-10 Non-volatile memory and method for predictive programming

Country Status (6)

Country Link
EP (1) EP2135252A2 (en)
JP (1) JP2010524147A (en)
KR (1) KR20100028019A (en)
CN (1) CN101711414B (en)
TW (1) TWI371041B (en)
WO (1) WO2008124760A2 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9105337B2 (en) 2011-01-10 2015-08-11 Micron Technology, Inc. Memories and methods of programming memories

Families Citing this family (76)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2007132456A2 (en) 2006-05-12 2007-11-22 Anobit Technologies Ltd. Memory device with adaptive capacity
WO2007132453A2 (en) 2006-05-12 2007-11-22 Anobit Technologies Ltd. Distortion estimation and cancellation in memory devices
WO2007132452A2 (en) 2006-05-12 2007-11-22 Anobit Technologies Reducing programming error in memory devices
WO2007132457A2 (en) 2006-05-12 2007-11-22 Anobit Technologies Ltd. Combined distortion estimation and error correction coding for memory devices
US8060806B2 (en) 2006-08-27 2011-11-15 Anobit Technologies Ltd. Estimation of non-linear distortion in memory devices
WO2008053472A2 (en) 2006-10-30 2008-05-08 Anobit Technologies Ltd. Reading memory cells using multiple thresholds
US7821826B2 (en) 2006-10-30 2010-10-26 Anobit Technologies, Ltd. Memory cell readout using successive approximation
US7924648B2 (en) 2006-11-28 2011-04-12 Anobit Technologies Ltd. Memory power and performance management
US7706182B2 (en) 2006-12-03 2010-04-27 Anobit Technologies Ltd. Adaptive programming of analog memory cells using statistical characteristics
WO2008068747A2 (en) 2006-12-03 2008-06-12 Anobit Technologies Ltd. Automatic defect management in memory devices
US7900102B2 (en) 2006-12-17 2011-03-01 Anobit Technologies Ltd. High-speed programming of memory devices
US8151166B2 (en) 2007-01-24 2012-04-03 Anobit Technologies Ltd. Reduction of back pattern dependency effects in memory devices
US7751240B2 (en) 2007-01-24 2010-07-06 Anobit Technologies Ltd. Memory device with negative thresholds
WO2008111058A2 (en) 2007-03-12 2008-09-18 Anobit Technologies Ltd. Adaptive estimation of memory cell read thresholds
US8001320B2 (en) 2007-04-22 2011-08-16 Anobit Technologies Ltd. Command interface for memory devices
US8234545B2 (en) 2007-05-12 2012-07-31 Apple Inc. Data storage with incremental redundancy
US8429493B2 (en) 2007-05-12 2013-04-23 Apple Inc. Memory device with internal signap processing unit
US7925936B1 (en) 2007-07-13 2011-04-12 Anobit Technologies Ltd. Memory device with non-uniform programming levels
US8259497B2 (en) 2007-08-06 2012-09-04 Apple Inc. Programming schemes for multi-level analog memory cells
US8174905B2 (en) 2007-09-19 2012-05-08 Anobit Technologies Ltd. Programming orders for reducing distortion in arrays of multi-level analog memory cells
US7773413B2 (en) 2007-10-08 2010-08-10 Anobit Technologies Ltd. Reliable data storage in analog memory cells in the presence of temperature variations
US8068360B2 (en) 2007-10-19 2011-11-29 Anobit Technologies Ltd. Reading analog memory cells using built-in multi-threshold commands
US8000141B1 (en) 2007-10-19 2011-08-16 Anobit Technologies Ltd. Compensation for voltage drifts in analog memory cells
WO2009050703A2 (en) 2007-10-19 2009-04-23 Anobit Technologies Data storage in analog memory cell arrays having erase failures
WO2009063450A2 (en) 2007-11-13 2009-05-22 Anobit Technologies Optimized selection of memory units in multi-unit memory devices
US8225181B2 (en) 2007-11-30 2012-07-17 Apple Inc. Efficient re-read operations from memory devices
US8209588B2 (en) 2007-12-12 2012-06-26 Anobit Technologies Ltd. Efficient interference cancellation in analog memory cell arrays
US8456905B2 (en) 2007-12-16 2013-06-04 Apple Inc. Efficient data storage in multi-plane memory devices
US8085586B2 (en) 2007-12-27 2011-12-27 Anobit Technologies Ltd. Wear level estimation in analog memory cells
US8156398B2 (en) 2008-02-05 2012-04-10 Anobit Technologies Ltd. Parameter estimation based on error correction code parity check equations
US7924587B2 (en) 2008-02-21 2011-04-12 Anobit Technologies Ltd. Programming of analog memory cells using a single programming pulse per state transition
US7864573B2 (en) 2008-02-24 2011-01-04 Anobit Technologies Ltd. Programming analog memory cells for reduced variance after retention
US8230300B2 (en) 2008-03-07 2012-07-24 Apple Inc. Efficient readout from analog memory cells using data compression
US8400858B2 (en) 2008-03-18 2013-03-19 Apple Inc. Memory device with reduced sense time readout
US8059457B2 (en) 2008-03-18 2011-11-15 Anobit Technologies Ltd. Memory device with multiple-accuracy read commands
WO2009151894A1 (en) * 2008-06-12 2009-12-17 Sandisk Corporation Nonvolatile memory and method with index programming and reduced verify
US7995388B1 (en) 2008-08-05 2011-08-09 Anobit Technologies Ltd. Data storage using modified voltages
US7924613B1 (en) 2008-08-05 2011-04-12 Anobit Technologies Ltd. Data storage in analog memory cells with protection against programming interruption
US8169825B1 (en) 2008-09-02 2012-05-01 Anobit Technologies Ltd. Reliable data storage in analog memory cells subjected to long retention periods
US8949684B1 (en) 2008-09-02 2015-02-03 Apple Inc. Segmented data storage
US8000135B1 (en) 2008-09-14 2011-08-16 Anobit Technologies Ltd. Estimation of memory cell read thresholds by sampling inside programming level distribution intervals
US8482978B1 (en) 2008-09-14 2013-07-09 Apple Inc. Estimation of memory cell read thresholds by sampling inside programming level distribution intervals
US8239734B1 (en) 2008-10-15 2012-08-07 Apple Inc. Efficient data storage in storage device arrays
US8261159B1 (en) 2008-10-30 2012-09-04 Apple, Inc. Data scrambling schemes for memory devices
US8208304B2 (en) 2008-11-16 2012-06-26 Anobit Technologies Ltd. Storage at M bits/cell density in N bits/cell analog memory cell devices, M>N
US8248831B2 (en) 2008-12-31 2012-08-21 Apple Inc. Rejuvenation of analog memory cells
US8397131B1 (en) 2008-12-31 2013-03-12 Apple Inc. Efficient readout schemes for analog memory cell devices
US8924661B1 (en) 2009-01-18 2014-12-30 Apple Inc. Memory system including a controller and processors associated with memory devices
US8228701B2 (en) 2009-03-01 2012-07-24 Apple Inc. Selective activation of programming schemes in analog memory cell arrays
US8832354B2 (en) 2009-03-25 2014-09-09 Apple Inc. Use of host system resources by memory controller
US8259506B1 (en) 2009-03-25 2012-09-04 Apple Inc. Database of memory read thresholds
US8238157B1 (en) 2009-04-12 2012-08-07 Apple Inc. Selective re-programming of analog memory cells
US8479080B1 (en) 2009-07-12 2013-07-02 Apple Inc. Adaptive over-provisioning in memory systems
US8495465B1 (en) 2009-10-15 2013-07-23 Apple Inc. Error correction coding over multiple memory pages
US8223556B2 (en) * 2009-11-25 2012-07-17 Sandisk Technologies Inc. Programming non-volatile memory with a reduced number of verify operations
US8677054B1 (en) 2009-12-16 2014-03-18 Apple Inc. Memory management schemes for non-volatile memory devices
US8694814B1 (en) 2010-01-10 2014-04-08 Apple Inc. Reuse of host hibernation storage space by memory controller
US8677203B1 (en) 2010-01-11 2014-03-18 Apple Inc. Redundant data storage schemes for multi-die memory systems
US8694853B1 (en) 2010-05-04 2014-04-08 Apple Inc. Read commands for reading interfering memory cells
US8572423B1 (en) 2010-06-22 2013-10-29 Apple Inc. Reducing peak current in memory systems
US8595591B1 (en) 2010-07-11 2013-11-26 Apple Inc. Interference-aware assignment of programming levels in analog memory cells
US9104580B1 (en) 2010-07-27 2015-08-11 Apple Inc. Cache memory for hybrid disk drives
US8645794B1 (en) 2010-07-31 2014-02-04 Apple Inc. Data storage in analog memory cells using a non-integer number of bits per cell
US8856475B1 (en) 2010-08-01 2014-10-07 Apple Inc. Efficient selection of memory blocks for compaction
US8493781B1 (en) 2010-08-12 2013-07-23 Apple Inc. Interference mitigation using individual word line erasure operations
US8694854B1 (en) 2010-08-17 2014-04-08 Apple Inc. Read threshold setting based on soft readout statistics
US9021181B1 (en) 2010-09-27 2015-04-28 Apple Inc. Memory management for unifying memory cell conditions by using maximum time intervals
JP2014053060A (en) 2012-09-07 2014-03-20 Toshiba Corp Semiconductor storage device and control method of the same
KR102449196B1 (en) * 2016-01-15 2022-09-29 삼성전자주식회사 Nonvolatile memory device and program method of a nonvolatile memory device
JP6539608B2 (en) * 2016-03-15 2019-07-03 東芝メモリ株式会社 Semiconductor memory device
CN110556150A (en) * 2018-06-01 2019-12-10 北京兆易创新科技股份有限公司 programming method and device of storage unit, electronic equipment and storage medium
CN110556145A (en) * 2018-06-01 2019-12-10 北京兆易创新科技股份有限公司 Programming method and device of storage unit, electronic equipment and storage medium
CN110556146A (en) * 2018-06-01 2019-12-10 北京兆易创新科技股份有限公司 programming method and device of storage unit, electronic equipment and storage medium
CN110610739B (en) * 2019-09-17 2021-06-18 珠海创飞芯科技有限公司 Threshold voltage adjusting method
US11556416B2 (en) 2021-05-05 2023-01-17 Apple Inc. Controlling memory readout reliability and throughput by adjusting distance between read thresholds
US11847342B2 (en) 2021-07-28 2023-12-19 Apple Inc. Efficient transfer of hard data and confidence levels in reading a nonvolatile memory

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0453096A (en) * 1990-06-19 1992-02-20 Toshiba Corp Analog storage device
US5729489A (en) * 1995-12-14 1998-03-17 Intel Corporation Programming flash memory using predictive learning methods
EP0913832B1 (en) * 1997-11-03 2003-07-23 STMicroelectronics S.r.l. Method for multilevel programming of a nonvolatile memory, and a multilevel nonvolatile memory
DE69927967T2 (en) * 1999-08-03 2006-07-27 Stmicroelectronics S.R.L., Agrate Brianza Programming method of a non-volatile multi-bit memory by regulating the gate voltage
US6504760B1 (en) * 2001-06-22 2003-01-07 Intel Corporation Charging a capacitance of a memory cell and charger
US7042766B1 (en) * 2004-07-22 2006-05-09 Spansion, Llc Method of programming a flash memory device using multilevel charge storage

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9105337B2 (en) 2011-01-10 2015-08-11 Micron Technology, Inc. Memories and methods of programming memories
TWI508076B (en) * 2011-01-10 2015-11-11 Micron Technology Inc Memories and methods of programming memories
US9484101B2 (en) 2011-01-10 2016-11-01 Micron Technology, Inc. Methods of programming memories

Also Published As

Publication number Publication date
JP2010524147A (en) 2010-07-15
CN101711414A (en) 2010-05-19
CN101711414B (en) 2013-06-26
WO2008124760A2 (en) 2008-10-16
WO2008124760A3 (en) 2008-11-27
TWI371041B (en) 2012-08-21
KR20100028019A (en) 2010-03-11
EP2135252A2 (en) 2009-12-23

Similar Documents

Publication Publication Date Title
TW200907975A (en) Non-volatile memory and method for predictive programming
US7965562B2 (en) Predictive programming in non-volatile memory
US7551483B2 (en) Non-volatile memory with predictive programming
KR101578812B1 (en) Nonvolatile memory and method with reduced program verify by ignoring fastest and/or slowest programming bits
US8472257B2 (en) Nonvolatile memory and method for improved programming with reduced verify
TWI417890B (en) Non-volatile memory and method with continuous scanning time-domain sensing
TWI305915B (en) Non-volatile memory and control with improved partial page program capability
US7826271B2 (en) Nonvolatile memory with index programming and reduced verify
US9361991B1 (en) Efficient scanning of nonvolatile memory blocks
US7800945B2 (en) Method for index programming and reduced verify in nonvolatile memory
US8300458B2 (en) Nonvolatile memory with correlated multiple pass programming
TW201011763A (en) Verification process for non-volatile storage
EP2289069B1 (en) Nonvolatile memory and method with index programming and reduced verify
US7796435B2 (en) Method for correlated multiple pass programming in nonvolatile memory
TW200532694A (en) Behavior based programming of non-volatile memory
US8964467B1 (en) Systems and methods for partial page programming of multi level cells
US9583207B2 (en) Adaptive data shaping in nonvolatile memory
EP2297739B1 (en) Nonvolatile memory and method for correlated multiple pass programming

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees