TWI682388B - Semiconductor device - Google Patents

Semiconductor device Download PDF

Info

Publication number
TWI682388B
TWI682388B TW108101265A TW108101265A TWI682388B TW I682388 B TWI682388 B TW I682388B TW 108101265 A TW108101265 A TW 108101265A TW 108101265 A TW108101265 A TW 108101265A TW I682388 B TWI682388 B TW I682388B
Authority
TW
Taiwan
Prior art keywords
variable resistance
array
line
transistor
string
Prior art date
Application number
TW108101265A
Other languages
Chinese (zh)
Other versions
TW201933352A (en
Inventor
李峰旻
林昱佑
Original Assignee
旺宏電子股份有限公司
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 US15/873,369 external-priority patent/US10719296B2/en
Priority claimed from US15/887,166 external-priority patent/US20190244662A1/en
Priority claimed from US16/233,404 external-priority patent/US10957392B2/en
Application filed by 旺宏電子股份有限公司 filed Critical 旺宏電子股份有限公司
Publication of TW201933352A publication Critical patent/TW201933352A/en
Application granted granted Critical
Publication of TWI682388B publication Critical patent/TWI682388B/en

Links

Images

Classifications

    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C13/00Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00
    • G11C13/0002Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00 using resistive RAM [RRAM] elements
    • G11C13/0021Auxiliary circuits
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C13/00Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00
    • G11C13/0002Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00 using resistive RAM [RRAM] elements
    • G11C13/0021Auxiliary circuits
    • G11C13/004Reading or sensing circuits or methods
    • 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
    • G11C16/26Sensing or reading circuits; Data output circuits
    • HELECTRICITY
    • H10SEMICONDUCTOR DEVICES; ELECTRIC SOLID-STATE DEVICES NOT OTHERWISE PROVIDED FOR
    • H10BELECTRONIC MEMORY DEVICES
    • H10B63/00Resistance change memory devices, e.g. resistive RAM [ReRAM] devices
    • H10B63/30Resistance change memory devices, e.g. resistive RAM [ReRAM] devices comprising selection components having three or more electrodes, e.g. transistors

Abstract

An array of variable resistance cells based on a programmable threshold transistor and a resistor connected in parallel is described, including 3D and split gate variations. An input voltage applied to the transistor, and the programmable threshold of the transistor, can represent variables of sum-of-products operations. Programmable threshold transistors in the variable resistance cells comprise charge trapping memory transistors, such as floating gate transistors or dielectric charge trapping transistors. The resistor in the variable resistance cells can comprise a buried implant resistor connecting the current-carrying terminals (e.g. source and drain) of the programmable threshold transistor. A voltage sensing sense amplifier is configured to sense the voltage generated by the variable resistance cells as a function of an applied current and the resistance of the variable resistance cells.

Description

半導體元件 Semiconductor components

本發明涉及可用於執行或支持乘積和運算的電路。 The present invention relates to a circuit that can be used to perform or support a product-sum operation.

在神經形態運算系統,機器學習系統和用於依據線性代數的一些計算類型的電路中,乘積和函數可以是重要的元件。函數可以表示如下:

Figure 108101265-A0305-02-0003-1
In neuromorphic computing systems, machine learning systems, and some types of computational circuits based on linear algebra, product sum functions can be important components. The function can be expressed as follows:
Figure 108101265-A0305-02-0003-1

在該表達式中,每個乘積項是可變輸入Xi和權重Wi的乘積。權重Wi可以在術語之間變化,對應於例如可變輸入Xi的係數。 In this expression, the variable input of each product term is the product of X i and the weighting W i. The weight W i may vary between terms, corresponding to, for example, the coefficient of the variable input X i .

乘積和函數可以實現於使用交叉點陣列架構的電路操作,其中陣列單元的電特性招致該功能。 The product-sum function can be implemented in circuit operations using a cross-point array architecture, where the electrical characteristics of the array unit incur this function.

對於高速實現,期望具有非常大的陣列,使得許多操作可以並行執行,或者非常大的乘積和系列可以執行。在系統中,有非常大數量的輸入和輸出,使總電流消耗可能非常大。 For high-speed implementations, it is desirable to have a very large array, so that many operations can be performed in parallel, or a very large product and series can be performed. In the system, there are a very large number of inputs and outputs, so that the total current consumption may be very large.

本發明期望提供適用於在大陣列中實現的乘積和函數的結構,並且可以更加有效節能。 The present invention is expected to provide a structure suitable for products and functions implemented in a large array, and can be more effective in energy saving.

本發明描述了一種包括可變電阻單元陣列的裝置,其中在陣列中一可變電阻單元包括一可編程閾值電晶體和並聯電阻。該裝置可以操作,使得施加到電晶體的輸入電壓和電晶體的可編程閾值可以表示乘積和運算的變化。在本文描述的實施例中,每個可變電阻單元的可變電阻是施加到單元中可編程閾值電晶體的控制閘極的電壓和電阻的函數。 The present invention describes an apparatus including an array of variable resistance cells, in which a variable resistance cell includes a programmable threshold transistor and a parallel resistor. The device is operable so that the input voltage applied to the transistor and the programmable threshold of the transistor can represent changes in the product and operation. In the embodiments described herein, the variable resistance of each variable resistance cell is a function of the voltage and resistance applied to the control gate of a programmable threshold transistor in the cell.

在一些實施例中,該裝置包括電壓感測感測放大器,用以感測藉由可變電阻單元而產生的電壓作為施加電流和可變電阻單元的電阻之函數。以這種方式,電流產生的大小來產出的乘積和可以被限制或固定,並降低功耗。 In some embodiments, the device includes a voltage sensing sense amplifier for sensing the voltage generated by the variable resistance unit as a function of the applied current and the resistance of the variable resistance unit. In this way, the product sum of the magnitude of current generation can be limited or fixed, and reduce power consumption.

利用由一個電晶體和一個電阻(1T-1R)組成的單元之陣列可以實現於二維(2D)和三維(3D)陣列中。此外,本文描述的實施例可以將電阻實現為具有單一可變閾值電晶體的佈局覆蓋區的埋藏式植入電阻,實際上製造一個電晶體(1T)單元的陣列,以配置於具有電壓感測的乘積和運算之非常緊湊的佈局。 The array of cells composed of a transistor and a resistor (1T-1R) can be implemented in two-dimensional (2D) and three-dimensional (3D) arrays. In addition, the embodiments described herein can implement the resistor as a buried implant resistor with a layout coverage of a single variable threshold transistor, in effect fabricating an array of transistor (1T) cells to be configured with voltage sensing The very compact layout of the product sum operation.

本發明描述了陣列中可變電阻單元配置於串聯可變電阻單元的多個串線路(String)之實施例。多個字元線可耦接到串聯可變電阻單元串線路的該些串線路。字元線驅動器連接到多個字元線,以施加可變閘極電壓至可變電阻單元中的可編程閾值電晶體。 The present invention describes an embodiment in which the variable resistance unit in the array is arranged on a plurality of strings in series with the variable resistance unit. A plurality of word lines may be coupled to the string lines of the serial variable resistance unit string lines. The word line driver is connected to multiple word lines to apply a variable gate voltage to a programmable threshold transistor in the variable resistance unit.

本發明實施例描述了可變電阻單元中的可編程閾值電晶體包括電荷捕獲儲存電晶體,例如是浮動閘極電晶體或介電電荷捕獲電晶體。 The embodiments of the present invention describe that the programmable threshold transistor in the variable resistance unit includes a charge trapping storage transistor, such as a floating gate transistor or a dielectric charge trapping transistor.

本發明實施例描述了可變電阻單元中的電阻包括串聯的可編程閾值電晶體的載流端子的埋藏式植入電阻;載流端子例如是源極和汲極。 Embodiments of the present invention describe a buried resistance of a resistor in a variable resistance unit including a current-carrying terminal of a programmable threshold transistor connected in series; the current-carrying terminal is, for example, a source electrode and a drain electrode.

本發明提供一種用於產生乘積和數據的裝置,該裝置包括可變電阻單元陣列,陣列中的各可變電阻單元,每個單元包括可編程閾值電晶體和並聯之電阻,該陣列包括具有串聯的單元串線路之n行單元及m列單元。控制和偏壓電路耦接到該陣列,該控制和偏壓電路包括用於陣列中利用對應於相應單元的權重因子數值Wmn的閾值來編程可編程閾值電晶體的邏輯。輸入驅動器耦接至m個列單元之對應列,輸入驅動器選擇性地施加輸入Xm至m列。行驅動器被配置於施加電流In至n行單元之一對應行。電壓感測電路可操作地耦接到行單元。 The present invention provides an apparatus for generating product and data. The apparatus includes an array of variable resistance units, each variable resistance unit in the array, each unit includes a programmable threshold transistor and a resistor in parallel, the array includes a series The n-row unit and m-column unit of the unit string circuit. A control and bias circuit is coupled to the array. The control and bias circuit includes logic for programming a programmable threshold transistor in the array with a threshold corresponding to the weighting factor value W mn of the corresponding cell. The input driver is coupled to the corresponding column of m column units, and the input driver selectively applies input X m to m columns. The row driver is configured to apply a current I n to a corresponding row of n row units. The voltage sensing circuit is operatively coupled to the row unit.

本發明描述了一種包括記憶體陣列和使用數據路徑控制器互連的乘積和加速器陣列的系統。乘積和加速器陣列包括可編程電阻單元陣列。記憶體陣列可以與乘積和加速器陣列協調使用,以用於配置與乘積和函數的運算。 The present invention describes a system including a memory array and a product and accelerator array interconnected using a data path controller. The product and accelerator array includes a programmable resistance cell array. Memory arrays can be used in coordination with product and accelerator arrays for configuration and product and function operations.

一種用於操作可變電阻單元陣列以產生乘積數據總和的方法,包括在陣列中利用對應於相應單元的權重因子數值的閾值來編程可編程閾值電晶體;選擇性地施加輸入至陣列中的列單元,施加電流至陣列中行單元的相應行;以及在陣列中行單元的一或多行單元上感測電壓。 A method for operating a variable resistance cell array to produce a sum of product data, including programming a programmable threshold transistor in the array with a threshold value corresponding to the weighting factor value of the corresponding cell; selectively applying input to a column in the array Cells, applying current to corresponding rows of row cells in the array; and sensing voltages on one or more row cells of the row cells in the array.

為了對本發明之上述及其他方面有更佳的瞭解,下文特舉實施例,並配合所附圖式詳細說明如下: In order to have a better understanding of the above and other aspects of the present invention, the following examples are specifically described in conjunction with the accompanying drawings as follows:

12、35、3635‧‧‧可編程閾值電晶體 12,35,3635‧‧‧Programmable threshold transistor

14、36、3636‧‧‧電阻 14, 36, 3636‧‧‧ resistance

21、22、23、24‧‧‧電流源 21, 22, 23, 24 ‧‧‧ current source

26‧‧‧接地線 26‧‧‧Ground

30、3630‧‧‧第一載流節點 30、3630‧‧‧First current carrying node

31、3631‧‧‧第二載流節點 31、3631‧‧‧second current carrying node

32、3632‧‧‧控制端子 32、3632‧‧‧Control terminal

100、200、300、3105‧‧‧基板 100, 200, 300, 3105‧‧‧ substrate

101、201‧‧‧源極端子 101、201‧‧‧Source terminal

102、202‧‧‧汲極端子 102、202‧‧‧ Jiji

103‧‧‧浮動閘極多晶矽層 103‧‧‧ Floating gate polysilicon layer

104、204‧‧‧控制閘極多晶矽層 104、204‧‧‧Control gate polysilicon layer

105、205‧‧‧閘極介電層 105, 205‧‧‧ Gate dielectric layer

106‧‧‧內多晶矽介電質 106‧‧‧Inner polysilicon dielectric

107、108、207、208‧‧‧接點 107, 108, 207, 208

109、209‧‧‧接觸層 109, 209‧‧‧ contact layer

110、210、304、451‧‧‧埋藏式植入電阻 110, 210, 304, 451

112、114、212、214‧‧‧電流路徑 112, 114, 212, 214‧‧‧ current path

113、213‧‧‧p型通道區域 113, 213‧‧‧ p-type channel area

203‧‧‧介電電荷捕獲層 203‧‧‧dielectric charge trapping layer

206‧‧‧閉塞介電層 206‧‧‧Block dielectric layer

301、302‧‧‧淺溝槽隔離結構 301,302‧‧‧Shallow trench isolation structure

303‧‧‧邊界 303‧‧‧Border

310‧‧‧源極區域 310‧‧‧Source area

311‧‧‧汲極區域 311‧‧‧ Drainage area

315‧‧‧浮動閘極 315‧‧‧floating gate

316‧‧‧控制閘極 316‧‧‧Control gate

320‧‧‧側壁 320‧‧‧Side wall

321‧‧‧蝕刻停止層 321‧‧‧Etching stop layer

322‧‧‧內連接介電質 322‧‧‧Interconnect dielectric

325、326‧‧‧內連接接點 325, 326‧‧‧ Internal connection contacts

400‧‧‧可變電阻單元的串聯串線路 400‧‧‧series connection of variable resistance unit

401、402‧‧‧串線路選定字元線 401, 402‧‧‧ string selected character line

410、411、412、413、414、415‧‧‧字元線閘極堆 410, 411, 412, 413, 414, 415‧‧‧‧ character line gate stack

420、421、422、423、424、425、426、427‧‧‧n型植入物 420,421,422,423,424,425,426,427 n-type implants

450‧‧‧p型保護層 450‧‧‧p-type protective layer

502‧‧‧位元線接點 502‧‧‧bit line contact

503‧‧‧源極線接點 503‧‧‧Source line contact

504、505‧‧‧主動區 504, 505‧‧‧ Active area

600、601‧‧‧載流節點 600, 601‧‧‧ current-carrying node

602‧‧‧輸入節點 602‧‧‧ input node

650‧‧‧感測放大器 650‧‧‧Sense amplifier

651‧‧‧緩衝器 651‧‧‧Buffer

652、656、965、975‧‧‧線 652, 656, 965, 975‧‧‧

655‧‧‧參考電壓電路 655‧‧‧ Reference voltage circuit

660‧‧‧暫存器 660‧‧‧register

661‧‧‧算術邏輯單元 661‧‧‧ arithmetic logic unit

665‧‧‧參考行 665‧‧‧ Reference line

666‧‧‧電阻分壓器 666‧‧‧Resistance voltage divider

667‧‧‧選擇器 667‧‧‧selector

680、690、700‧‧‧操作串線路 680, 690, 700 ‧‧‧ operation string circuit

681‧‧‧行參考串線路 681‧‧‧ line reference string line

691、701、702‧‧‧參考串線路 691, 701, 702

692‧‧‧參考行上的區域 692‧‧‧ Reference area

703‧‧‧陣列區域 703‧‧‧Array area

901‧‧‧積體電路 901‧‧‧Integrated circuit

905‧‧‧數據匯流排 905‧‧‧Data bus

910‧‧‧控制邏輯 910‧‧‧Control logic

920、3410‧‧‧方塊 920, 3410‧‧‧ block

930‧‧‧匯流排 930‧‧‧Bus

940‧‧‧字元線驅動器 940‧‧‧ character line driver

945、3006、3108、3205、3405a、3405b‧‧‧字元線 945, 3006, 3108, 3205, 3405a, 3405b ‧‧‧ character line

960‧‧‧神經形態的記憶體陣列 960‧‧‧ neuromorphic memory array

970‧‧‧行解碼器 970‧‧‧line decoder

980‧‧‧緩衝電路 980‧‧‧buffer circuit

985‧‧‧第二數據線 985‧‧‧Second data cable

990‧‧‧數據緩衝器 990‧‧‧Data buffer

991‧‧‧輸入/輸出電路 991‧‧‧ input/output circuit

993‧‧‧數據路徑 993‧‧‧Data path

1000‧‧‧系統 1000‧‧‧System

1001‧‧‧乘積和加速器陣列 1001‧‧‧ Product and accelerator array

1002‧‧‧記憶體陣列 1002‧‧‧Memory array

1003‧‧‧數據路徑控制器 1003‧‧‧Data path controller

2700‧‧‧U形垂直NAND類串線路 2700‧‧‧U-shaped vertical NAND string circuit

2800‧‧‧第一NAND類串線路 2800‧‧‧The first NAND string circuit

2801‧‧‧第二NAND類串線路 2801‧‧‧Second NAND string circuit

2901、3201‧‧‧第一圖案金屬層 2901, 3201‧‧‧ First pattern metal layer

2902、3202‧‧‧第二圖案金屬層 2902, 3202‧‧‧Second pattern metal layer

2903、3106、3208‧‧‧串線路選定線 2903, 3106, 3208

2904、3107、3203‧‧‧接地選定線 2904, 3107, 3203‧‧‧Ground selected line

2905、3005a、3005b、3103‧‧‧垂直通道結構 2905、3005a、3005b、3103‧‧‧Vertical channel structure

2906‧‧‧輔助閘極結構 2906‧‧‧ auxiliary gate structure

3001、3101、3401‧‧‧位元線 3001, 3101, 3401‧‧‧bit line

3002、3402‧‧‧源極線 3002, 3402‧‧‧ source line

3003、3004、3403、3404‧‧‧選定閘極 3003, 3004, 3403, 3404 ‧‧‧ selected gate

3007‧‧‧後閘極 3007‧‧‧ Rear gate

3102‧‧‧內連接 3102‧‧‧ Internal connection

3104‧‧‧共源極線擴散 3104‧‧‧Common source line diffusion

3109‧‧‧介電層 3109‧‧‧dielectric layer

3110、3206‧‧‧電荷捕獲層 3110, 3206‧‧‧ charge trapping layer

3111‧‧‧半導體材料薄膜 3111‧‧‧semiconductor material film

3112‧‧‧摻雜多晶矽 3112‧‧‧doped polysilicon

3204‧‧‧輔助閘極線 3204‧‧‧Auxiliary gate line

3207‧‧‧薄膜通道層 3207‧‧‧thin film channel layer

3209‧‧‧絕緣層 3209‧‧‧Insulation

3406‧‧‧共用浮動閘極結構 3406‧‧‧ shared floating gate structure

3407‧‧‧多晶矽間介電層 3407‧‧‧polysilicon dielectric layer

3408‧‧‧隧穿介電層 3408‧‧‧Tunnel dielectric layer

3409‧‧‧摻雜半導體核心 3409‧‧‧doped semiconductor core

第1圖係繪示乘積和運算的功能圖,乘積和運算可以是先前技術中已知的神經形態運算系統的基本元件。 Figure 1 is a functional diagram showing the product-sum operation. The product-sum operation may be a basic element of a neuromorphic operation system known in the prior art.

第2圖係繪示配置於乘積和運算的可變電阻單元陣列的一部分圖。 FIG. 2 is a partial diagram of a variable resistance cell array disposed in a product-sum operation.

第3圖繪示根據本文描述的實施例的可變電阻單元的示意圖。 FIG. 3 is a schematic diagram of a variable resistance unit according to embodiments described herein.

第4圖繪示包括浮動閘極儲存電晶體和埋藏式植入電阻的可變電阻單元的簡化剖面圖。 FIG. 4 shows a simplified cross-sectional view of a variable resistance unit including floating gate storage transistors and buried implant resistors.

第5圖繪示包括介電電荷捕獲儲存電晶體和埋藏式植入電阻的可變電阻單元的簡化剖面圖。 FIG. 5 shows a simplified cross-sectional view of a variable resistance unit including dielectric charge trapping storage transistors and buried implant resistors.

第6-9圖繪示根據本文描述的實施例的可變電阻單元的製造流程圖。 FIGS. 6-9 illustrate a manufacturing flowchart of the variable resistance unit according to the embodiments described herein.

第10A圖和第10B圖係繪示可變電阻單元串聯排列於NAND類(非及類,NAND-Like)結構的剖面圖及佈局圖。 FIGS. 10A and 10B are a cross-sectional view and a layout view showing that variable resistance units are arranged in series in a NAND-like (NAND-Like) structure.

第11A圖和第11B圖係繪示可變電阻單元的例示操作圖。 11A and 11B are diagrams illustrating exemplary operation of the variable resistance unit.

第12圖係繪示配置實施於乘積和運算的可變電阻單元之串線路(String)。 FIG. 12 illustrates a string line (String) of variable resistance units configured for product-sum operation.

第13圖係繪示可與可變電阻單元陣列一起使用於乘積和運算的感測電路之簡化方塊圖。 FIG. 13 is a simplified block diagram of a sensing circuit that can be used in a product-sum operation together with a variable resistance cell array.

第13A圖係繪示利用第13圖電路的感測操作目的之啟示圖。 FIG. 13A is an enlightening diagram showing the purpose of the sensing operation using the circuit of FIG. 13.

第14圖繪示可與如第13圖的感測電路一起使用的參考電壓電路之簡化圖。 Figure 14 shows a simplified diagram of a reference voltage circuit that can be used with the sensing circuit as shown in Figure 13.

第15圖繪示包括參考串(String)的可變電阻單元陣列的配置。 FIG. 15 illustrates the configuration of a variable resistance cell array including a reference string (String).

第16圖係繪示包括參考串和未使用單元的可變電阻單元陣列的另一種配置圖。 FIG. 16 shows another configuration diagram of a variable resistance cell array including a reference string and unused cells.

第17圖係繪示包括兩個參考串和未使用單元的可變電阻單元陣列的另一種配置圖。 FIG. 17 shows another configuration diagram of a variable resistance cell array including two reference strings and unused cells.

第18-22圖係繪示配置於利用多個位元權重來實施乘積和運算項之可變電阻單元的功能組圖。 Figures 18-22 are functional diagrams of variable resistance units configured to use multiple bit weights to implement product and operation terms.

第23圖是包括可變電阻單元陣列應用的裝置之簡化方塊圖,例如是應用於神經形態記憶體。 Figure 23 is a simplified block diagram of a device including variable resistance cell array applications, for example, for neuromorphic memory.

第24-26圖係繪示包括乘積加速器陣列及其不同操作的系統。 Figures 24-26 show the system including the product accelerator array and its different operations.

第27圖是可變電阻單元的垂直U形NAND類串線路的示意電路圖。 Fig. 27 is a schematic circuit diagram of a vertical U-shaped NAND string circuit of a variable resistance unit.

第28圖是可變電阻單元的替代垂直NAND類串線路的示意電路圖。 Fig. 28 is a schematic circuit diagram of an alternative vertical NAND-like string circuit of a variable resistance unit.

第29圖是可變電阻單元的垂直U形NAND類串線路的3D陣列透視圖。 FIG. 29 is a perspective view of a 3D array of vertical U-shaped NAND string circuits of variable resistance units.

第30圖是可變電阻單元的單一垂直U形NAND類串線路可變電阻單元串線路的剖面圖,其可實施於如第29圖的陣列。 FIG. 30 is a cross-sectional view of a single vertical U-shaped NAND series string variable resistance cell string line of the variable resistance cell, which can be implemented in the array as shown in FIG. 29.

第31圖還繪示如本文描述的可變電阻單元的垂直U形NAND類串線路的3D陣列之另一種類型。 FIG. 31 also illustrates another type of 3D array of vertical U-shaped NAND-like string lines of variable resistance cells as described herein.

第32圖係繪示包含另一實施例的可變電阻單元的垂直U形NAND類串線路的垂直通道單閘極結構。 FIG. 32 shows a vertical channel single gate structure of a vertical U-shaped NAND string circuit including a variable resistance unit of another embodiment.

第33圖是第32圖陣列的可變電阻單元之特寫圖。 Figure 33 is a close-up view of the variable resistance unit of the array of Figure 32.

第34圖是共享包括如本文所述的可變電阻單元的浮動閘極類型NAND類結構的垂直剖面圖。 FIG. 34 is a vertical cross-sectional view of a floating gate type NAND-like structure including variable resistance cells as described herein.

第35圖是第34圖共享浮動閘極類型NAND類可變電阻單元串線路的特寫圖。 Figure 35 is a close-up view of the shared floating gate type NAND variable resistance cell string circuit of Figure 34.

第36圖是根據本文描述實施例的分離閘極可變電阻單元的示意圖。 FIG. 36 is a schematic diagram of a split gate variable resistance unit according to embodiments described herein.

第1-36圖提供了對本發明實施例的詳細描述。 Figures 1-36 provide a detailed description of embodiments of the present invention.

第1圖是乘積和運算圖,其中總和項是輸入Xi與權重Wi的乘積,在該範例中,其中i從1至7。權量Wi可以不同於總和項。在操作中,權重可以被指定為一組係數,然後應用於計算總和變化的輸入作為可變輸入。此外,在執行學習過程的演算法中,隨著時間變化的權重作為學習過程以改變係數,學習得到可利用結果的總和。 Figure 1 is a product sum operation diagram, where the sum term is the product of the input X i and the weight W i , in this example, where i is from 1 to 7. The weight W i may be different from the sum term. In operation, the weights can be specified as a set of coefficients and then applied to the input that calculates the change in the sum as a variable input. In addition, in the algorithm that performs the learning process, the weights that change over time are used as the learning process to change the coefficients, and the learning results in the sum of the available results.

在繪示的範例中,總和的輸出被應用於S形函數,以利用非線性方法產生介於最小值與最大值之範圍的輸出,例如介於0和1之間。這是用於神經形態運算中突觸層的常見模型。其他激活功能(activation function)也能被利用,例如是邏輯運算。乘積和運算也可以應用於非神經形態的配置或者神經系統模型的考量。 In the illustrated example, the output of the summation is applied to the sigmoid function to generate a non-linear method to produce an output between the minimum and maximum values, such as between 0 and 1. This is a common model used for synaptic layers in neuromorphic operations. Other activation functions can also be used, such as logical operations. The product-sum operation can also be applied to non-neuromorphic configurations or neurological model considerations.

第2圖是可變電阻單元陣列的示意圖,其中陣列中的每個單 元包括可編程閾值電晶體12與並聯之電阻14。在該圖式中,陣列包括四個串線路(String)可變電阻單元,其中各串線路包括在總和節點SUM1至SUM4之間四個串聯的可變電阻單元與參考線,參考線例如是接地線26。四條字元線WL1至WL4耦接到各串線路中可變電阻單元的控制端子。如圖所示,可以有任意數量的行與總和節點接至SUMn,以及任何數量的字元線WLm。n行與m列的可變電阻單元具有的權重Wnm當作單元的可編程閾值Vt、單元的電阻Rnm以及行中的電流In之函數。 Figure 2 is a schematic diagram of a variable resistance cell array, where each cell in the array includes a programmable threshold transistor 12 and a resistor 14 in parallel. In this diagram, the array includes four string variable resistance units, where each string includes four variable resistance units connected in series between the sum nodes SUM 1 to SUM 4 and a reference line, for example是 ground wire 26. The four word lines WL1 to WL4 are coupled to the control terminals of the variable resistance units in each string. As shown, there can be any number of row and sum nodes connected to SUM n and any number of word lines WL m . variable resistance unit n rows and m columns with weight W nm as a unit of programmable threshold Vt, the resistor R nm and a function of the current I n of the row unit.

施加於字元線的電壓對應於可變輸入X1到X4,...Xm。在這個方法中,串線路中的每個可變電阻單元的可變電阻是在字元線上的電壓施加於單元的控制閘極、單元中可編程閾值電晶體的閾值,單元中的電流和電阻器的函數。 The voltage applied to the word line corresponds to the variable inputs X 1 to X 4 ,... X m . In this method, the variable resistance of each variable resistance cell in the string line is the voltage applied on the word line to the control gate of the cell, the threshold of the programmable threshold transistor in the cell, the current and resistance in the cell Function.

總和節點SUM1至SUM4,...SUMn耦接到電壓感測感測放大器,以產生表示各串線路的乘積和輸出之信號。在代表範例中,在感測操作期間,電流源21-24耦接到每個串線路,以施加固定電流至各串線路。 Sum nodes SUM 1 to SUM 4 ,... SUM n are coupled to a voltage sensing sense amplifier to generate signals representing the product and output of each string line. In a representative example, during the sensing operation, current sources 21-24 are coupled to each string line to apply a fixed current to each string line.

第3圖是一個可變電阻單元的示意圖,例如用於第2圖的陣列。可變電阻單元包括第一載流節點30、第二載流節點31和控制端子32。可編程閾值電晶體35和電阻36並聯連接到第一載流節點和第二載流節點。可編程閾值電晶體具有連接至控制端子32的閘極。 Figure 3 is a schematic diagram of a variable resistance unit, for example for the array of Figure 2. The variable resistance unit includes a first current carrying node 30, a second current carrying node 31, and a control terminal 32. The programmable threshold transistor 35 and the resistor 36 are connected in parallel to the first current-carrying node and the second current-carrying node. The programmable threshold transistor has a gate connected to the control terminal 32.

控制端子32上的電壓VG可以視為可編程閾值電晶體35的閘極電壓。控制端子32可以對應於圖2中陣列的字元線。第一載流節點30上的電壓VS可視為單元的源極電壓。第二載流節點31上的電壓VD可視為單元的汲極電壓。 The voltage VG on the control terminal 32 can be regarded as the gate voltage of the programmable threshold transistor 35. The control terminal 32 may correspond to the word line of the array in FIG. 2. The voltage V S on the first current-carrying node 30 can be regarded as the source voltage of the cell. The voltage V D on the second current-carrying node 31 can be regarded as the drain voltage of the cell.

在此範例中,將單元電流IC施加到具有設計或可調節電流振幅的第二載流節點31,以在單元中依據電阻36的單元中電壓感測放大器的電壓範圍與電阻值建立電壓降。電流振幅可依據特定陣列的實施例來調整,使一個有用的電壓範圍能被產生於串線路上以供應於總和節點。此外,電阻器的電阻大小和可編程閾值電晶體的配置可以設計成具有選定的電流級別和指定的感測範圍來操作。 In this example, the cell current I C applied to the carrier having a second node designed or adjustable amplitude of the current 31, to establish a voltage drop depending on the voltage range of the resistance value of the resistor element 36 in the voltage of the sense amplifier in the unit . The current amplitude can be adjusted according to the specific array embodiment, so that a useful voltage range can be generated on the string line to be supplied to the sum node. In addition, the resistance size of the resistor and the configuration of the programmable threshold transistor can be designed to operate with a selected current level and a specified sensing range.

可編程閾值電晶體35可使用浮動閘極記憶胞、分離閘極浮動閘極記憶胞、介電電荷捕獲記憶胞,例如SONOS元件或其它已知類型的介電電荷捕獲記憶胞如BE-SONOS和TANOS,以及分離閘極。其他可編程記憶胞技術也能被使用,例如相變記憶體、金屬氧化物記憶體等。 The programmable threshold transistor 35 can use floating gate memory cells, split gate floating gate memory cells, and dielectric charge trapping memory cells, such as SONOS devices or other known types of dielectric charge trapping memory cells such as BE-SONOS and TANOS, and separation gate. Other programmable memory cell technologies can also be used, such as phase change memory, metal oxide memory, etc.

此外,在本技術的實施例中,電阻36可以實施於可編程閾值電晶體35的源極和汲極端子之間的埋藏式植入電阻。 Furthermore, in embodiments of the present technology, the resistor 36 may be implemented as a buried implant resistor between the source and drain terminals of the programmable threshold transistor 35.

第36圖是一個分離閘極可變電阻單元的示意圖,例如可以用於第2圖的陣列的修改版本中。可變電阻單元包括第一載流節點3630、第二載流節點3631和控制端子3632。控制端子藉由電荷捕獲層涵蓋部分通道長度,以及透過沒有電荷捕獲的閘極介電質從該通道被分離,以用於在通道長度平衡中的數據儲存。可編程閾值電晶體3635和電阻3636並聯連接至第一載流節點和第二載流節點。可編程閾值電晶體具有連接到控制端子3632的閘極。 Figure 36 is a schematic diagram of a separate gate variable resistance unit, which can be used in a modified version of the array of Figure 2, for example. The variable resistance unit includes a first current carrying node 3630, a second current carrying node 3631, and a control terminal 3632. The control terminal covers part of the channel length by the charge trapping layer and is separated from the channel by the gate dielectric without charge trapping for data storage in channel length balancing. The programmable threshold transistor 3635 and the resistor 3636 are connected in parallel to the first current-carrying node and the second current-carrying node. The programmable threshold transistor has a gate connected to the control terminal 3632.

控制端子3632上的電壓VG可視為可編程閾值電晶體3635的閘極電壓。控制端子3632可以對應於第2圖所示的陣列中的字元線。第一載流節點3630上的電壓可視為單元的電壓源。第二載流節點3631上的 電壓VD可視為單元的汲極電壓。 The voltage V G at the control terminal 3632 can be regarded as the gate voltage of the programmable threshold transistor 3635. The control terminal 3632 may correspond to the word line in the array shown in FIG. 2. The voltage on the first current carrying node 3630 can be regarded as the voltage source of the cell. The voltage V D on the second current-carrying node 3631 can be regarded as the drain voltage of the cell.

第4圖是具有並聯連接至通道的電阻,以及該電阻被實施於利用離子植入而產生埋藏式植入電阻110的浮動閘極裝置的簡化剖面圖。 FIG. 4 is a simplified cross-sectional view of a floating gate device having a resistor connected in parallel to a channel, and the resistor is implemented using ion implantation to create a buried implant resistor 110.

在此範例中,該裝置實現於基板100上,該基板100可以是p型基板。源極端子101和汲極端子102透過在基板100中n型離子植入來實施。源極端子101和汲極端子102具有其形成的接點107、108,耦接到具有電壓VS的源極節點和具有電壓VD的汲極節點。p型通道區域113設置在埋藏式植入電阻110和閘極介電層105(通道氧化物)之間,以覆蓋在源極端子101和汲極端子102之間的基板100上。浮動閘極多晶矽層103設置於閘極介電層105上。設置於浮動閘極多晶矽層103上的內多晶矽介電質106於一些實施例中使用包含氧化矽,氮化矽和氧化矽層(ONO)的多層結構。控制閘極多晶矽層104設置在內多晶矽介電質106上。接觸層109形成在控制閘極多晶矽層104上。側壁結構(未編號)沿著閘極堆的側壁形成。 In this example, the device is implemented on a substrate 100, which may be a p-type substrate. The source terminal 101 and the drain terminal 102 are implemented by n-type ion implantation in the substrate 100. The source terminal 101 and the drain terminal 102 have their formed contacts 107, 108, coupled to a source node having a voltage V S and a drain node having a voltage V D. The p-type channel region 113 is provided between the buried implant resistor 110 and the gate dielectric layer 105 (channel oxide) to cover the substrate 100 between the source terminal 101 and the drain terminal 102. The floating gate polysilicon layer 103 is disposed on the gate dielectric layer 105. In some embodiments, the inner polysilicon dielectric 106 disposed on the floating gate polysilicon layer 103 uses a multilayer structure including silicon oxide, silicon nitride, and silicon oxide layer (ONO). The control gate polysilicon layer 104 is disposed on the inner polysilicon dielectric 106. The contact layer 109 is formed on the control gate polysilicon layer 104. The sidewall structure (not numbered) is formed along the sidewall of the gate stack.

第4圖繪示的結構可使用浮動閘極單元製造技術來實現,並透過額外摻雜的步驟來修改,以形成埋藏式植入電阻110。埋藏式植入電阻110連接源極端子101和汲極端子102,以作為被動電阻。在此方式中,浮動閘極裝置和埋藏式植入電阻110提供可編程閾值電晶體和並聯電阻於第一載流端子-源極端子101、和第二載流端子-汲極端子102之間。 The structure shown in FIG. 4 can be implemented using floating gate cell manufacturing technology and modified by additional doping steps to form buried implant resistor 110. The buried implant resistor 110 connects the source terminal 101 and the drain terminal 102 as a passive resistor. In this way, the floating gate device and the buried implant resistor 110 provide a programmable threshold transistor and a parallel resistor between the first current carrying terminal-source terminal 101 and the second current carrying terminal-drain terminal 102 .

在第4圖中繪示了電流路徑112,電流路徑112通過源極端子101和汲極端子102之間的埋藏式植入電阻110。在第4圖中也繪示了電流路徑114,當在浮動閘極中閘極電壓與捕獲電荷和源極電壓VS結合時, 電流路徑114被激活,以使電流流經電晶體的通道。 In FIG. 4, a current path 112 is shown. The current path 112 passes through a buried implant resistor 110 between the source terminal 101 and the drain terminal 102. The current path 114 is also shown in FIG. 4. When the gate voltage is combined with the trapped charge and the source voltage V S in the floating gate, the current path 114 is activated to allow current to flow through the channel of the transistor.

因此,該裝置具有可變電阻(或可變電導),可變電阻是埋藏式植入電阻110的電阻和浮動閘極裝置通道的電阻之函數。浮動閘極裝置通道的電阻是閘極電壓和在浮動閘極中捕獲電荷的函數。 Therefore, the device has a variable resistance (or variable conductance), which is a function of the resistance of the buried implant resistor 110 and the resistance of the floating gate device channel. The resistance of the floating gate device channel is a function of the gate voltage and the charge trapped in the floating gate.

第5圖是具有電阻並聯連接至通道的介電電荷捕獲裝置的簡化剖面圖,以及實施於使用離子植入方法,以產生埋藏式植入電阻210。 FIG. 5 is a simplified cross-sectional view of a dielectric charge trapping device having a resistor connected in parallel to a channel, and is implemented using an ion implantation method to produce a buried implant resistor 210.

在此範例中,該裝置實施於基板200上,基板200可以是p型基板。源極端子201和汲極端子202通過在基板200中n型離子植入來實施。源極端子201和汲極端子202具有其形成的接點207、208,並耦接到具有電壓VS的源極節點和具有電壓VD的汲極節點。p型通道區域213設置在埋藏式植入電阻210和閘極介電層205(通道氧化物)之間,以覆蓋在源極端子201和汲極端子202之間的基板200上。介電電荷捕獲層203設置在通道介電層205上。閉塞介電層206設置在介電電荷捕獲層203上,控制閘極多晶矽層204設置在閉塞介電層206上。接觸層209形成在控制閘極多晶矽層204上。側壁結構(未編號)沿著閘極堆的側壁形成。 In this example, the device is implemented on a substrate 200, which may be a p-type substrate. The source terminal 201 and the drain terminal 202 are implemented by n-type ion implantation in the substrate 200. The source terminal 201 and the drain terminal 202 have contacts 207, 208 formed by them, and are coupled to a source node having a voltage V S and a drain node having a voltage V D. The p-type channel region 213 is provided between the buried implant resistor 210 and the gate dielectric layer 205 (channel oxide) to cover the substrate 200 between the source terminal 201 and the drain terminal 202. The dielectric charge trap layer 203 is provided on the channel dielectric layer 205. The blocking dielectric layer 206 is disposed on the dielectric charge trapping layer 203, and the control gate polysilicon layer 204 is disposed on the blocking dielectric layer 206. The contact layer 209 is formed on the control gate polysilicon layer 204. The sidewall structure (not numbered) is formed along the sidewall of the gate stack.

第5圖繪示的結構可以使用介電電荷捕獲記憶胞製造技術來實現,可透過額外的摻雜步驟來修改,以形成埋藏式植入電阻210。埋藏式植入電阻210連接源極端子201和汲極端子202,以作為被動電阻。在這種方式中,介電電荷捕獲裝置和埋藏式植入電阻210提供可編程閾值電晶體和在源極端子201和汲極端子202之間的並聯電阻。 The structure shown in FIG. 5 can be implemented using a dielectric charge trapping memory cell manufacturing technique, which can be modified through additional doping steps to form a buried implant resistor 210. The buried implant resistor 210 connects the source terminal 201 and the drain terminal 202 as a passive resistance. In this manner, the dielectric charge trapping device and the buried implant resistor 210 provide a programmable threshold transistor and a parallel resistance between the source terminal 201 and the drain terminal 202.

在第5圖中繪示了電流路徑212,電流路徑212通過源極端子201和汲極端子202之間的埋藏式植入電阻210。第5圖中也繪示了電流 路徑214,當在介電電荷捕獲層中閘極電壓與捕獲電荷結合時,電流路徑214被激活,以使電流流經裝置通道。 In FIG. 5, a current path 212 is shown, and the current path 212 passes through the buried implant resistor 210 between the source terminal 201 and the drain terminal 202. Figure 5 also shows the current The path 214, when the gate voltage and the trapped charge are combined in the dielectric charge trapping layer, the current path 214 is activated to allow current to flow through the device channel.

因此,該裝置具有可變電阻(或電導),可變電阻是埋藏式植入電阻110的電阻和介電電荷捕獲裝置通道的電阻之函數。介電電荷捕獲裝置通道的電阻是閘極電壓和在介電電荷捕獲閘極中捕獲電荷的函數。 Therefore, the device has a variable resistance (or conductance), which is a function of the resistance of the buried implant resistor 110 and the resistance of the channel of the dielectric charge trapping device. The resistance of the channel of the dielectric charge trapping device is a function of the gate voltage and the charge trapped in the dielectric charge trapping gate.

在第4圖和第5圖的兩個實施例中繪示了由一個電晶體和一個電阻(1T-1R)組成的單元。此外,第4圖和第5圖的實施例可以將電阻器實施於單一可變閾值電晶體佈局覆蓋區內的埋藏式植入電阻,並有效地製造一個電晶體(1T)單元的陣列,以配置於具有電壓感測的乘積和運算之非常緊湊的佈局。 In the two embodiments of FIG. 4 and FIG. 5, a unit composed of a transistor and a resistor (1T-1R) is shown. In addition, the embodiments of FIGS. 4 and 5 can implement the resistor in a buried variable resistor within the coverage area of a single variable threshold transistor layout, and effectively manufacture an array of transistor (1T) cells to Configured in a very compact layout with product and operation of voltage sensing.

在操作中,在第4圖和第5圖中所繪示的單元可表示如下。 In operation, the units depicted in Figures 4 and 5 can be represented as follows.

當閘極至源極電壓VGS小於閾值電壓Vt時,電流可以流入埋藏式植入電阻,卻沒有形成於電晶體通道(“表面通道”),僅允許電流IB於埋藏式植入電阻。因此,單元中的電流等於IB,而單元的電阻等於汲極至源極電壓VDS除以電流IBWhen the gate-to-source voltage V GS is less than the threshold voltage V t , current can flow into the buried implant resistor, but it is not formed in the transistor channel (“surface channel”), and only the current I B is allowed to be in the buried implant resistor . Therefore, the current in the cell is equal to I B , and the resistance of the cell is equal to the drain-to-source voltage V DS divided by the current I B.

當閘極至源極電壓VGS大於閾值電壓Vt時,表面通道電流IS和隱藏式電阻電流IB都被誘發。通道電阻可以遠小於隱藏式電阻的電阻,而當電晶體導通時,Is可以控制。因此,行中的電流In被分成單元,使得電流In等於IS+IB的總和,並且單元電阻等於汲極至源極電壓VDS除以電流InWhen the gate-to-source voltage V GS is greater than the threshold voltage V t , both the surface channel current IS and the hidden resistance current I B are induced. Channel resistance can be much smaller than the resistance of resistor hidden, and when the transistor is turned on, I s can be controlled. Accordingly, the current row is divided into unit I n, I n is equal to the sum of such currents I S + I B, and the cell resistance is equal to the drain to source voltage V DS divided by the current I n.

由於浮動閘極閾值或介電電荷捕獲單元是可編程的,該單元電阻可模擬由閘極電壓所表示的參數X(i)、以及由單元中電荷捕獲、電 阻器的電阻和單元電流所表示的參數W(i)之乘積。參數W(i)可以是二進位數,其中單元操作於兩種狀態之一(IB僅較高電阻狀態和IB+IS較低電阻狀態)。如果單元操作於場效電晶體行為的線性區域中,則參數W(i)可以是類比,並且根據單元中電荷捕獲的範圍內變化。 Since the floating gate threshold or dielectric charge trapping unit is programmable, the cell resistance can simulate the parameter X(i) represented by the gate voltage, as well as the charge trapping in the cell, the resistance of the resistor and the cell current The product of the parameter W(i). The parameter W(i) can be a binary number, where the unit operates in one of two states (I B only higher resistance state and I B + IS lower resistance state). If the cell operates in a linear region of field-effect transistor behavior, the parameter W(i) may be an analog and vary within the range according to charge trapping in the cell.

第6-9圖繪示了可用於實現類似第4圖單元的製造流程圖。在第6圖中,繪示了提供單元的介電邊界之淺溝槽隔離結構301和302之後,基板300被顯示出來。並植入已經被用來形成由提供該單元形成的基板300中之區域的邊界303所表示之p型井。在陣列中單元的相異方塊可以在分離方塊中實施,以允許井的獨立偏壓用於分離方塊。 Figures 6-9 show a manufacturing flow chart that can be used to implement a unit similar to Figure 4. In FIG. 6, after the shallow trench isolation structures 301 and 302 providing the dielectric boundary of the cell are depicted, the substrate 300 is shown. And implanted a p-type well that has been used to form the boundary 303 of the area in the substrate 300 formed by providing the unit. The different blocks of the cells in the array can be implemented in separate blocks to allow the independent bias of the well to be used to separate the blocks.

第7圖繪示了n型摻雜劑(例如磷和砷)被用於形成在淺溝槽隔離結構301和302之間的埋藏式植入電阻304之後的階段。 FIG. 7 illustrates the stage after n-type dopants (such as phosphorus and arsenic) are used to form the buried implant resistor 304 between the shallow trench isolation structures 301 and 302.

第8圖繪示了閘極堆結構(浮動閘極315、控制閘極316、通道介電和沿著側壁320的多晶矽介電質)的形成以及使用n型摻雜劑植入的源極區域和汲極區域310和311形成之後的階段。 Figure 8 shows the formation of the gate stack structure (floating gate 315, control gate 316, channel dielectric and polysilicon dielectric along the sidewall 320) and the source region implanted using n-type dopants The stage after the formation of the drain regions 310 and 311.

第9圖繪示了在內連接介電質322和內連接接點325和326的形成之後的製造階段。該結構使用在源極區域和汲極區域上形成矽化物接點的方法而形成於繪示的範例中,接著是在閘極堆以及源極區域與汲極區域310和311上方的薄介電質和蝕刻停止層321。內連接介電質322被沉積,並且孔被蝕刻以形成開口,其中內連接接點325和326由鎢沉積或其他技術所形成。 FIG. 9 illustrates the manufacturing stage after the formation of the interconnect dielectric 322 and the interconnect contacts 325 and 326. This structure is formed in the illustrated example using the method of forming silicide contacts on the source and drain regions, followed by the thin dielectric above the gate stack and the source and drain regions 310 and 311质和刻止层321。 Quality and etching stop layer 321. The interconnect dielectric 322 is deposited, and the hole is etched to form an opening, where the interconnect contacts 325 and 326 are formed by tungsten deposition or other techniques.

可以看出如第4圖所示的可變電阻單元是依據這些程序來製造。這些相同程序可以被修改為如第5圖所示的製造單元的目的,這些相 同程序透過包括閘極介電質、電荷捕獲層、阻擋層和控制閘極的閘極堆疊來修改。 It can be seen that the variable resistance unit shown in FIG. 4 is manufactured according to these procedures. These same procedures can be modified for the purpose of manufacturing cells as shown in Figure 5, these phases The same procedure is modified by the gate stack including the gate dielectric, charge trapping layer, barrier layer and control gate.

具有如第4圖和第5圖中所示結構的可變電阻單元可以使用連接於內連接接點325、326的圖案導體層來串聯排列。 The variable resistance units having the structure shown in FIGS. 4 and 5 can be arranged in series using the pattern conductor layer connected to the interconnection contacts 325, 326.

第10A圖和第10B圖是2D NAND類(NAND-Like)結構中串聯排列的可變電阻單元之剖面和佈局視圖。 10A and 10B are cross-sectional and layout views of variable resistance cells arranged in series in a 2D NAND-like (NAND-Like) structure.

第10A圖繪示了基板的簡化剖面,其中可變電阻單元的串聯串線路400被形成。包括電荷捕獲層(浮動閘極或介電質)和字元線的字元線閘極堆410-415疊置在基板,且延伸至圖頁方向垂直的字元線元件。在代表性實施中,可能存在例如32或64個主動字元線。在一些實施例中,串聯串線路可包括較少數量的主動字元線或較大數量的主動字元線,以適用於特定實現。在一些情況下,可能存在一個或多個虛擬字元線,其可以在串線路的相對端上,典型例子如高密度NAND快閃記憶體。虛擬字元線可以實現於製造品質或偏置目的,但不用於串線路的乘積和運算。 FIG. 10A shows a simplified cross section of the substrate, in which the series string line 400 of the variable resistance unit is formed. A word line gate stack 410-415 including a charge trapping layer (floating gate or dielectric) and a word line is stacked on the substrate and extends to a word line element perpendicular to the page direction. In a representative implementation, there may be, for example, 32 or 64 active character lines. In some embodiments, the series string line may include a smaller number of active word lines or a larger number of active word lines to suit a particular implementation. In some cases, there may be one or more dummy word lines, which may be on opposite ends of the string line, typical examples are high-density NAND flash memory. The virtual word line can be used for manufacturing quality or offset purposes, but it is not used for the product and operation of serial lines.

在此範例中,該基板是p型基板,並且可變電阻單元的載流端子(即源極/汲極端子)由n型植入物420-427來實施。在一些高密度實施例中,植入物不用於單元之間的載流端子,因此載流端子依賴於通道區域中電荷載體的反轉。如圖繪示的NAND類的實施例中,沒有接點直接形成於所有單元之間的載流端子。 In this example, the substrate is a p-type substrate, and the current-carrying terminals (ie, source/drain terminals) of the variable resistance unit are implemented by n-type implants 420-427. In some high-density embodiments, implants are not used for current-carrying terminals between cells, so the current-carrying terminals rely on the inversion of charge carriers in the channel region. In the illustrated NAND-like embodiment, no contacts are directly formed between current-carrying terminals between all cells.

串線路選定字元線401和402設置在串聯串線路的相對端上。在基板中的主動區504和505包括用於串聯串線路的位元線與共源極線連接之n型植入物。主動區504和505相較於可變電阻單元的載流端子可以 是更深植入物或更高導電率的植入物,位元線接點502連接主動區504至覆蓋圖案導體層中的位元線。源極線接點503連接主動區505至覆蓋圖案導體層中的源極線。 The string line selected character lines 401 and 402 are provided at opposite ends of the series string line. The active regions 504 and 505 in the substrate include n-type implants for connecting bit lines and common source lines of a serial string. The active regions 504 and 505 can be compared with the current-carrying terminals of the variable resistance unit It is a deeper implant or a higher conductivity implant. The bit line contact 502 connects the active area 504 to the bit line in the pattern conductor layer. The source line contact 503 connects the active area 505 to the source line in the pattern conductor layer.

實施n型埋藏式植入電阻451,此範例中來自由位元線側串線路選定字元線401控制的選定閘極的通道邊緣延伸到由源極線側的串線路選定字元線402控制的選定閘極的通道邊緣。在這種方式中,選定閘極操作於將埋藏式植入電阻451連接和不連接至主動區504、505。 An n-type buried implant resistor 451 is implemented. In this example, the channel edge from the selected gate controlled by the selected word line 401 on the bit line side string line selection extends to the controlled by the string line selected word line 402 on the source line side. The channel edge of the selected gate. In this manner, the selected gate is operated to connect and not connect the buried implant resistor 451 to the active regions 504, 505.

在此範例中,具有比可變電阻單元的通道區域更高p型雜質濃度之p型保護層450設置於通道和埋藏式植入電阻451之間。p型保護層450有助於屏蔽埋藏式植入電阻451受於閘極電壓,並保持並聯電阻值的穩定性。 In this example, a p-type protective layer 450 having a higher p-type impurity concentration than the channel region of the variable resistance unit is disposed between the channel and the buried implant resistor 451. The p-type protective layer 450 helps shield the buried implant resistor 451 from the gate voltage and maintain the stability of the parallel resistance value.

第10B圖是實現如第10A圖所示的可變電阻單元的串聯串線路之平面圖。公共參考標號(401、402、410-415)被給予至閘極堆(包括字元線)和選定線。同樣地,公共參考標號502、503被給予至位元線和源極線接點。 Fig. 10B is a plan view of a series string line implementing the variable resistance unit shown in Fig. 10A. Common reference numbers (401, 402, 410-415) are given to the gate stack (including the word line) and the selected line. Likewise, common reference numbers 502, 503 are given to the bit line and source line contacts.

第10B圖繪示了兩個串聯串線路被編排於並聯位元線500、501,其實施於字元線閘極堆410-415上覆蓋的圖案化之導體層。 FIG. 10B shows that the two series strings are arranged on the parallel bit lines 500 and 501, which are implemented on the patterned conductor layer covered on the word line gate stacks 410-415.

第27圖給示如本文描述之U形垂直NAND類串線路2700包含具有電阻(例如,2701)並聯至可編程閾值、可操作電荷捕獲記憶胞(例如,2702)的可變電阻單元,作為乘積和加速器的示意圖。在第27圖中,NAND類串線路包括耦接到位元線BL之一U形串線路,其包括串聯一串線路選擇開關SSL、多個第一單元耦接到右側上各自的字元線WL、結 構底部的輔助閘極AG。多個第二單元耦接到左側上各自的字元線WL,以及接地選擇開關GSL耦接到源極參考線CS。操作的輸入被施加於如本文所述的字元線WL並進行配置如說明。使用電壓和電流感測技術或其他感測技術,以感測位元線或參考線上的輸出。 Figure 27 shows that a U-shaped vertical NAND-like string line 2700 as described herein includes a variable resistance unit with a resistance (eg, 2701) in parallel to a programmable threshold, an operable charge trapping memory cell (eg, 2702) as a product And accelerator schematics. In FIG. 27, the NAND-type string line includes one U-shaped string line coupled to the bit line BL, which includes a series of line selection switches SSL connected in series, and a plurality of first cells coupled to respective word lines WL on the right side Knot The auxiliary gate AG at the bottom of the structure. The plurality of second cells are coupled to respective word lines WL on the left side, and the ground selection switch GSL is coupled to the source reference line CS. The input of the operation is applied to the word line WL as described herein and configured as explained. Use voltage and current sensing techniques or other sensing techniques to sense the output on the bit line or reference line.

第28圖是包括第一NAND類串線路2800和第二NAND類串線路2801的替代垂直結構之示意圖,其包括具有並聯至可編程閾值、電荷捕獲記憶胞的電阻之可變電阻單元,可如本文所述的乘積和加速器之動作。在第29圖中,第一NAND類串線路延伸於第一位元線BL和源極參考線CS之間,並且包括串線路選定閘極SSL、多個字元線WL和串聯的接地選定閘極GSL。同樣地,第二NAND類串線路延伸於第二位元線BL和源極參考線CS之間,並包括串線路選定閘極SSL、多個字元線WL和串聯的接地選定閘極GSL。操作輸入被施加到如本文所述的字元線WL和配置,使用電流感測技術或其他感測技術,可感測位元線或參考線上的輸出。 FIG. 28 is a schematic diagram of an alternative vertical structure including a first NAND-type string line 2800 and a second NAND-type string line 2801, which includes a variable resistance unit having a resistance in parallel to a programmable threshold, charge trapping memory cell, such as The product and accelerator actions described in this article. In FIG. 29, the first NAND-type string line extends between the first bit line BL and the source reference line CS, and includes a string line selection gate SSL, a plurality of word lines WL, and a series ground selection gate Very GSL. Similarly, the second NAND-type string line extends between the second bit line BL and the source reference line CS, and includes a string line selected gate SSL, a plurality of word lines WL, and a series-connected ground selected gate GSL. The operation input is applied to the word line WL and configuration as described herein, and the output on the bit line or the reference line can be sensed using current sensing technology or other sensing technology.

第29-35圖繪示了可用於實現垂直NAND類實施例的各種結構。在第29圖中,類似於第27圖的U形NAND類串線路實施於密集的3D陣列中。該結構包括被位元線實施的第一圖案金屬層2901,第二圖案金屬層2902包括參考線。選定閘極層包括串線路選定線2903和接地選定線2904。垂直通道結構2905透過字元線延伸至多個層。垂直通道結構連接至由輔助閘極結構2906支撐的U形圖案。垂直通道結構包括具有自然電阻的數據儲存層,例如電荷捕獲結構,以及N型多晶矽摻雜垂直通道。垂直通道結構被結構的各層中的字元線所環繞,從而形成所謂的全閘極可變電阻單元。因此,當字元線上的輸入沒有導通至電荷捕獲單元時,則電阻由垂直通道 的電阻率來決定。當字元線上的輸入導通至電荷捕獲單元時,電阻由電荷捕獲單元的並聯電阻和垂直通道的電阻率來確定。請參見Tanaka,H.,et al.,"Bit Cost Scalable Technology with Punch and Plug Process for Ultra High Density Flash Memory," 2007 Symp.VLSI Tech.,Digest of Tech.Papers,pp 14-15.。 Figures 29-35 illustrate various structures that can be used to implement vertical NAND-like embodiments. In FIG. 29, a U-shaped NAND string circuit similar to FIG. 27 is implemented in a dense 3D array. The structure includes a first pattern metal layer 2901 implemented by bit lines, and a second pattern metal layer 2902 including reference lines. The selected gate layer includes a string line selection line 2903 and a ground selection line 2904. The vertical channel structure 2905 extends through the word lines to multiple layers. The vertical channel structure is connected to the U-shaped pattern supported by the auxiliary gate structure 2906. The vertical channel structure includes a data storage layer with natural resistance, such as a charge trapping structure, and an N-type polysilicon doped vertical channel. The vertical channel structure is surrounded by word lines in each layer of the structure, thereby forming a so-called full gate variable resistance cell. Therefore, when the input on the word line is not conducted to the charge trap unit, the resistance To determine the resistivity. When the input on the word line is conducted to the charge trap unit, the resistance is determined by the parallel resistance of the charge trap unit and the resistivity of the vertical channel. See Tanaka, H., et al., "Bit Cost Scalable Technology with Punch and Plug Process for Ultra High Density Flash Memory," 2007 Symp.VLSI Tech., Digest of Tech.Papers, pp 14-15.

第30圖繪示單一U形NAND類串線路,其可實施於類似第29圖的陣列。位元線3001經由選定閘極3004連接到垂直通道結構3005a左側上一系列的可變電阻單元,並且透過垂直通道結構的水平延伸部分和後閘極3007連接到相對垂直通道結構3005b上可變電阻單元的第二垂直串線路。可變電阻單元的第二垂直串線路透過選定閘極3003耦接到源極線3002。各串線路包括多個字元線(例如3006),以如本文所述的配置進行操作輸入。在U形NAND類串線路的兩側上的垂直通道結構3005a、3005b可以包括n型摻雜多晶矽,以提供隱藏電阻,該隱藏電組由原位摻雜或由植入所形成。 Figure 30 shows a single U-shaped NAND-like string circuit, which can be implemented in an array similar to Figure 29. The bit line 3001 is connected to a series of variable resistance units on the left side of the vertical channel structure 3005a via the selected gate 3004, and is connected to the variable resistance of the opposite vertical channel structure 3005b through the horizontal extension of the vertical channel structure and the rear gate 3007 The second vertical string line of the unit. The second vertical string line of the variable resistance unit is coupled to the source line 3002 through the selected gate 3003. Each string includes a plurality of word lines (for example, 3006), and is operated and input in the configuration as described herein. The vertical channel structures 3005a, 3005b on both sides of the U-shaped NAND-like string circuit may include n-type doped polysilicon to provide hidden resistance, which is formed by in-situ doping or implantation.

第31圖繪示了實現類似第28圖電路的替代垂直NAND類串線路。在第31圖的結構中,位元線3101實施於第一圖案金屬層中。位元線3101透過內連接3102連接到垂直通道結構3103。垂直通道結構3103包括沿著垂直通道結構側面延伸的半導體材料薄膜3111。在此實施例中垂直通道結構3103的內部包括摻雜多晶矽3112,以提供隱藏電阻。垂直通道結構3103在多個導體堆之間延伸,該些導體堆被配置於串線路的相對端上與字元線3108之間的串線路選定線3106和接地選定線3107。垂直通道結構3103耦接到共源極線擴散3104和基板3105。電荷捕獲層3110設置在垂直通道結構3103中作為字元線(例如3108)和半導體材料薄膜3111的導體 之間。介電層3109分離導電材料串線路,以作為字元線和串線路選定線。由於垂直通道結構3103相對側上的串線路選定線3106是分開的,所以這些垂直NAND類串線路可以共享單一位元線3101。請參見Jang,et al.,“Vertical Cell Array using TCAT(Terabit Cell Array Transistor)Technology for Ultra High Density NAND Flash Memory”,2009 Symposium on VLSI Technology Digest of Technical Papers。 Figure 31 shows an alternative vertical NAND-like string circuit that implements the circuit of Figure 28. In the structure of FIG. 31, the bit line 3101 is implemented in the first pattern metal layer. The bit line 3101 is connected to the vertical channel structure 3103 through the interconnection 3102. The vertical channel structure 3103 includes a thin film of semiconductor material 3111 extending along the side of the vertical channel structure. In this embodiment, the interior of the vertical channel structure 3103 includes doped polysilicon 3112 to provide hidden resistance. The vertical channel structure 3103 extends between a plurality of conductor stacks, which are disposed on the string line selection line 3106 and the ground selection line 3107 between the opposite ends of the string line and the character line 3108. The vertical channel structure 3103 is coupled to the common source line diffusion 3104 and the substrate 3105. The charge trapping layer 3110 is provided in the vertical channel structure 3103 as a conductor of a word line (for example, 3108) and a thin film of semiconductor material 3111 between. The dielectric layer 3109 separates the string line of conductive material to serve as a character line and a string line selection line. Since the string line selection lines 3106 on opposite sides of the vertical channel structure 3103 are separated, these vertical NAND-like string lines can share a single bit line 3101. See Jang, et al., "Vertical Cell Array using TCAT (Terabit Cell Array Transistor) Technology for Ultra High Density NAND Flash Memory", 2009 Symposium on VLSI Technology Digest of Technical Papers.

第32圖繪示了密集3D陣列,密集3D陣列包括如本文所述的單閘極、包含與電荷捕獲記憶胞並聯的電阻之可變電阻單元的U形垂直NAND類串線路。在此結構中,位元線實施於第一圖案金屬層3201,源極線實施於第二圖案金屬層3202。位元線和源極線透過內連接連接器連接到包含半導體材料的U形薄膜之垂直通道結構,垂直通道結構用於如本文實施例所述之通道和隱藏式電阻。垂直通道結構延伸於多個導體之間,垂直通道結構被配置為串線路選定線3208或接地選定線3203、字元線(例如3205)和輔助閘極線3204。這些導體由絕緣材料來分開。垂直通道結構包括電荷捕獲層3206以及延伸於U形圖案的薄膜通道層3207。該結構覆蓋於絕緣層3209。薄膜通道層3207可包括摻雜多晶矽,薄膜通道層3207透過多晶矽膜的原位摻雜或是摻雜多晶矽的沉積來形成。參見美國專利No.US 9,698,156和美國專利No.9,524,980。 FIG. 32 shows a dense 3D array including a single gate as described herein, a U-shaped vertical NAND-like string circuit including variable resistance cells in parallel with a charge trapping memory cell. In this structure, the bit line is implemented on the first pattern metal layer 3201, and the source line is implemented on the second pattern metal layer 3202. The bit line and the source line are connected to the vertical channel structure of the U-shaped film containing semiconductor material through the interconnection connector. The vertical channel structure is used for the channel and the hidden resistor as described in the embodiments of the present invention. The vertical channel structure extends between the plurality of conductors, and the vertical channel structure is configured as a string line selected line 3208 or a ground selected line 3203, a character line (eg, 3205), and an auxiliary gate line 3204. These conductors are separated by insulating materials. The vertical channel structure includes a charge trap layer 3206 and a thin film channel layer 3207 extending in a U-shaped pattern. This structure covers the insulating layer 3209. The thin film channel layer 3207 may include doped polysilicon. The thin film channel layer 3207 is formed by in-situ doping of the polysilicon film or deposition of doped polysilicon. See US Patent No. US 9,698,156 and US Patent No. 9,524,980.

第33圖繪示了耦接到字元線3205的兩個垂直通道可變電阻單元。每個可變電阻單元包括電荷捕獲層3206和摻雜的薄膜通道層3207。 FIG. 33 shows two vertical channel variable resistance units coupled to the word line 3205. Each variable resistance unit includes a charge trap layer 3206 and a doped thin film channel layer 3207.

堆疊中的導體被配置為字元線,字元線可用於施加輸入至NAND類串線路陣列的操作以作為乘積加速器結構。 The conductors in the stack are configured as word lines, which can be used to apply input to NAND-like string line array operations as a product accelerator structure.

第34圖還繪示了包括如本文所述的可變電阻單元之另一種垂直NAND類結構。在此範例中,這一系列單元耦接於位元線3401和源極線3402之間。一堆導體所包含導體被配置為選定閘極3403和3404的導體,以及導體被配置為字元線(例如3405a和3405b)。共用浮動閘極結構(例如3406)設置在被配置為字元線(例如3405a,3405b)的導體之間。垂直通道結構包括隧穿介電層3408和摻雜半導體核心3409,例如n型摻雜多晶矽,垂直通道結構摻雜表示為可變電阻單元的隱藏電阻。多晶矽間介電層3407將配置為字元線的導體從垂直通道結構和浮動閘極來分開。 Figure 34 also illustrates another vertical NAND-type structure including variable resistance cells as described herein. In this example, the series of cells is coupled between the bit line 3401 and the source line 3402. The pile of conductors includes conductors configured as conductors of selected gates 3403 and 3404, and conductors configured as word lines (eg, 3405a and 3405b). A common floating gate structure (eg 3406) is provided between conductors configured as word lines (eg 3405a, 3405b). The vertical channel structure includes a tunneling dielectric layer 3408 and a doped semiconductor core 3409, such as n-type doped polysilicon. The vertical channel structure doping is expressed as the hidden resistance of the variable resistance unit. The inter-polysilicon dielectric layer 3407 separates the conductors configured as word lines from the vertical channel structure and the floating gate.

第35圖繪示了來自第34圖方塊3410的結構。可以看出,垂直通道結構由包含摻雜半導體核心3409和隧穿介電層3408所設置與環繞。導體配置為字元線3405a和3405b,以及共用浮動閘極結構3406。多晶矽間介電層3407將配置為字元線3405a和3405b的導體從垂直通道結構(包括摻雜半導體核心3409)和共用浮動閘極結構3406來分離,參見Seo,et al.,“A Novel 3-D Vertical FG NAND Flash Memory Cell Arrays Using the Separated Sidewall Control Gate(S-SCG)for Highly Reliable MLC Operation”,2011 IEEE。 Figure 35 shows the structure of block 3410 from Figure 34. It can be seen that the vertical channel structure is arranged and surrounded by the doped semiconductor core 3409 and the tunneling dielectric layer 3408. The conductors are configured as word lines 3405a and 3405b, and a common floating gate structure 3406. The polysilicon dielectric layer 3407 separates the conductors configured as word lines 3405a and 3405b from the vertical channel structure (including the doped semiconductor core 3409) and the common floating gate structure 3406, see Seo, et al., "A Novel 3 -D Vertical FG NAND Flash Memory Cell Arrays Using the Separated Sidewall Control Gate (S-SCG) for Highly Reliable MLC Operation", 2011 IEEE.

利用可變電阻單元的乘積和陣列實施例可以具有非常大的陣列,其包括具有數千或數百萬個可變電阻單元陣列。被用於大規模2D和3D NAND裝置的製造技術可以被應用,製造技術所具有的附加步驟用於埋藏式植入電阻或其他電阻結構,如第10A圖、第10B圖和第27-36圖所示的NAND類結構中實施於大量乘積和陣列的製造。用以應用寫入(編程和抹除)權重至可編程電阻單元的操作技術,類似於那些使用大規模NAND 裝置的操作技術。如上所述,可編程電阻單元可以被操作於模擬模式。在模擬模式中,用於感測電路和訊號路由的外圍電路可能很複雜。 Product and array embodiments utilizing variable resistance cells can have very large arrays, including arrays with thousands or millions of variable resistance cells. The manufacturing technology used for large-scale 2D and 3D NAND devices can be applied. The manufacturing technology has additional steps for buried implant resistors or other resistive structures, such as Figures 10A, 10B, and 27-36 The illustrated NAND structure is implemented in the manufacture of a large number of products and arrays. Operating techniques used to apply write (program and erase) weights to programmable resistance cells, similar to those using large-scale NAND The operating technology of the device. As mentioned above, the programmable resistance unit can be operated in an analog mode. In analog mode, the peripheral circuits used for sensing circuits and signal routing can be complex.

外圍電路可由可編程電阻單元配置於單元陣列中而被簡化,以用“二進制”模式來操作。可編程閾值電晶體可以儲存二進制狀態。施加於行的電流可以是常數,或是施加固定數量的二進制級。可編程電阻單元中的電阻在整個陣列中可以是常數,或者實施於固定數量的二進制級電阻。 Peripheral circuits can be simplified by configuring programmable resistance cells in the cell array to operate in "binary" mode. Programmable threshold transistors can store binary states. The current applied to the row can be constant or a fixed number of binary levels can be applied. The resistance in the programmable resistance unit can be constant throughout the array, or implemented in a fixed number of binary-level resistances.

二進制模式操作可透過減少需要編程於單元中閾值的可編程演算法之複雜性來允許外圍電路的簡化,電流源用於施加電流至陣列中的行,以及感測電路用於產生輸出值。 Binary mode operation can allow the simplification of peripheral circuits by reducing the complexity of programmable algorithms that need to be programmed into thresholds in cells, current sources for applying current to rows in the array, and sensing circuits for generating output values.

第11A圖表示出了單一可編程電阻單元的電路示意圖。第11B圖可透過每單位一位元中的單元操作、二進制模式來提供了IV曲線(電流-電壓曲線)而被理解。單元作為載流節點600和601。輸入節點602連接到如上所述的可編程電晶體閘極。單元中並聯電阻的電阻設置為一數值Rmn,其中m對應於單元的列,並且n對應於單元的行。 Fig. 11A shows a schematic circuit diagram of a single programmable resistance unit. Figure 11B can be understood by providing the IV curve (current-voltage curve) through the unit operation in one bit per unit, the binary mode. Units act as current-carrying nodes 600 and 601. The input node 602 is connected to the programmable transistor gate as described above. The resistance of the parallel resistors in the cell is set to a value R mn , where m corresponds to the column of the cell and n corresponds to the row of the cell.

第11B圖表示了兩個電壓對電流曲線圖。第一電壓對電流曲線對應於“1”單元權重wmn,其中單元具有低閾值Vt。第二曲線對應於“0”單元權重wmn,其中單元具有高閾值Vt。當輸入值低,因此低Vt大於輸入電壓,單元中的電晶體關閉,並且對於單元的二進位權重傳導非常低的電流。當輸入值高時,低Vt小於輸入電壓,輸入電壓小於高Vt,若對應到低Vt單元的權重為“1”,則單元中的電晶體導通,以及若對應到高Vt單元的權重為“0”,則單元中的電晶體截止。 Figure 11B shows two voltage vs. current curves. The first voltage versus current curve corresponds to the "1" cell weight w mn , where the cell has a low threshold V t . The second curve corresponds to the "0" cell weight w mn , where the cell has a high threshold V t . When the input value is low and therefore the low V t is greater than the input voltage, the transistor in the cell is turned off and conducts very low current to the cell's binary weight. When the input value is high, the low V t is less than the input voltage, and the input voltage is less than the high V t . If the weight corresponding to the low V t cell is “1”, the transistor in the cell is turned on, and if the high V t cell The weight of is "0", the transistor in the cell is cut off.

當晶體管截止時,由流經電阻I * Rmn的電流造成的電壓降所支配的較大電壓降VdLg被誘發。當晶體管導通時,由流經電晶體通道的電流造成的電壓降所支配的較小的電壓降VdSm被誘發,可視為接近於0V的較小的電壓降VdSm由電流通過晶體管溝道引起的電壓降所支配。這種關係由下面的表格1來說明。 When the transistor is turned off, a larger voltage drop V dLg dominated by the voltage drop caused by the current flowing through the resistance I*R mn is induced. When the transistor is turned on, the smaller voltage drop V dSm dominated by the voltage drop caused by the current flowing through the transistor channel is induced, which can be regarded as a smaller voltage drop V dSm close to 0V caused by the current through the transistor channel Is governed by the voltage drop. This relationship is illustrated by Table 1 below.

Figure 108101265-A0305-02-0022-2
Figure 108101265-A0305-02-0022-2

二進制操作可以擴展到如第12圖所示的可變電阻單元串線路。在第12圖中,三個單元被繪示在n行陣列中的單一串線路上。該行接收固定電流In和各列上的輸入值X1至X3。行中的電壓降取決於行中各單元的權重W1n、W2n和W3n以及輸入值。此範例使用三個獨立單元來實現乘積和運算的i從1到3之三個項XiWi,以產生代表總和的電壓VnThe binary operation can be extended to the variable resistance unit string circuit shown in FIG. In Figure 12, three cells are depicted on a single string in an n-row array. This row receives the fixed current I n and the input values X 1 to X 3 on each column. The voltage drop in the row depends on the weights W 1n , W 2n and W 3n of the cells in the row and the input values. In this example, three independent units are used to implement the three terms X i W i of the product sum operation i from 1 to 3 to generate the voltage V n representing the sum.

在第一行中表示出的三個輸入變數,以及在表格的第二列中表示出的三個潛在權重,並且假設每個單元的固定電流和固定電阻值,在行中電壓降Vn的變化可在表格2中看到(假設VdSm接近(“~”)於零)。 The three input variables shown in the first row and the three potential weights shown in the second column of the table, and assuming a fixed current and fixed resistance value for each cell, the voltage drop V n in the row The change can be seen in Table 2 (assuming that V dSm is close to (“~”) at zero).

Figure 108101265-A0305-02-0022-3
Figure 108101265-A0305-02-0022-3
Figure 108101265-A0305-02-0023-4
Figure 108101265-A0305-02-0023-4

藉由依據這四個Vn位準來設置感測參考電壓,可以將遍布行的電壓轉換為0和3之間的數字輸出,如下面的表3所示。 By setting the sensing reference voltage based on these four V n levels, the voltage across the row can be converted to a digital output between 0 and 3, as shown in Table 3 below.

Figure 108101265-A0305-02-0023-5
Figure 108101265-A0305-02-0023-5

隨著提供唯一輸入的列數和單元行數的增加,陣列可以產生複合乘積和而依賴於單個可編程電阻單元的“二進制”操作(即,將電晶體編程為低閾值或高閾值)。 As the number of columns and cell rows providing unique inputs increases, the array can produce a composite product sum that depends on the "binary" operation of a single programmable resistance cell (ie, programming the transistor to a low or high threshold).

在一些實施例中,多位元二進制權重可以儲存在陣列中的一些或所有單元中,以增加單元的可編程權重之進一步解析度。 In some embodiments, multi-bit binary weights may be stored in some or all cells in the array to increase the further resolution of the programmable weights of the cells.

第13圖是可以與配置於乘積和運算的可變電阻單元陣列,與如上所述之電壓感測一起使用的感測電路方塊圖。該範例中的感測電路包括實施於使用運算放大器的範例或其他類型比較器的感測放大器650。感測放大器650的輸入包括在線652上的電壓vn和參考電壓Vref。電壓vn在選定行上產生,並且可以透過緩衝器651來傳送。緩衝器651可實施於配置在運算放大器的單位增益之範例或電壓放大器的電壓。線652上的參考電壓Vref由參考電壓電路655提供,參考電壓電路655被配置於透過對應於由感測放大器650響應於線656上序列信號所區分的各電壓位準之一組參考電壓的序列。參考電壓電路655可以接收輸入電壓Vmax和Vmin,參考電壓電路655可以決定在線652上產生的最小和最大電壓為參考電壓VrefFIG. 13 is a block diagram of a sensing circuit that can be used with voltage sensing as described above with a variable resistance cell array arranged in a product-sum operation. The sensing circuit in this example includes a sense amplifier 650 implemented in an example using an operational amplifier or other types of comparators. The input of the sense amplifier 650 includes the voltage v n on line 652 and the reference voltage V ref . The voltage v n is generated on the selected line and can be transmitted through the buffer 651. The buffer 651 may be implemented in an example of unity gain of an operational amplifier or a voltage of a voltage amplifier. The reference voltage V ref on line 652 is provided by a reference voltage circuit 655 configured to pass a set of reference voltages corresponding to the voltage levels distinguished by the sense amplifier 650 in response to the sequence signal on line 656 sequence. The reference voltage circuit 655 may receive the input voltages V max and V min , and the reference voltage circuit 655 may decide that the minimum and maximum voltages generated on the line 652 are the reference voltage V ref .

第13A圖是感測操作的啟示圖。第14圖電路的特定電壓Vmax和Vmin在多級中可以產生具有如圖表所示的感測範圍之參考電壓。在陣列中選定行上產生的電壓Vn可以落在感測範圍內的位準,電壓Vn具有不同於上述電壓Vmin的電壓。感測電路決定電壓Vn的位準。在這種情況下,電壓Vn高於各參考電壓V1至V5而低於參考電壓V6。因此,可以指定對應於參考電壓V6的電壓數值VnFig. 13A is a revelation diagram of the sensing operation. The specific voltages V max and V min of the circuit of FIG. 14 can generate a reference voltage having a sensing range as shown in the graph in multiple stages. In the array generated on the selected row voltage level Vn may fall within the sensing range, the voltage V n has a voltage different from the voltage V min of. The sensing circuit determines the level of the voltage V n . In this case, the voltage V n is higher than the respective reference voltages V 1 to V 5 and lower than the reference voltage V 6 . Therefore, the voltage value V n corresponding to the reference voltage V 6 can be specified.

感測放大器650的輸出包括與輸入參考電壓位準相對應的信號排序。這些信號可以儲存於暫存器660,暫存器660被提供至算術邏輯單元661或其他類似的通用處理器,其中算術運算可以進一步執行乘積和運算。例如,依據如下所討論之可編程電阻單元陣列是如何配置,以使產生 在陣列的多個行上的輸出可以組合於產生乘積和運算的單一項之目的。 The output of the sense amplifier 650 includes signal sequencing corresponding to the input reference voltage level. These signals may be stored in a temporary register 660, which is provided to an arithmetic logic unit 661 or other similar general-purpose processor, where the arithmetic operation may further perform a product-sum operation. For example, based on how the programmable resistance cell array discussed below is configured to produce The output on multiple rows of the array can be combined for the purpose of generating a single item of the product sum operation.

第14圖是可以與如第13圖所安排的感測放大器裝置一起使用的參考電壓電路的方塊圖。在第14圖中,可編程電阻單元陣列中一參考行或多個參考行665,或是使用如那些應用於陣列中的單元結構,可以安排至提供電壓Vmax和Vmin之一或兩者。此範例的電壓Vmax和Vmin被施加到電阻分壓器666,電阻分壓器666在電阻分壓器666中電阻之間的節點處產生多個參考電壓位準。響應於參考電壓位準的節點耦接到選擇器667。耦接到第13圖配置中的感測放大器650之選擇器667響應於線656上的序列信號,以在線652上提供一序列的參考電壓VrefFig. 14 is a block diagram of a reference voltage circuit that can be used with the sense amplifier device arranged as in Fig. 13. In FIG. 14, a reference row or a plurality of reference rows 665 in the programmable resistance cell array, or using a cell structure such as those applied in the array, can be arranged to provide one or both of the voltages V max and V min . The voltages V max and V min of this example are applied to a resistive voltage divider 666 that generates multiple reference voltage levels at the nodes between the resistors in the resistive voltage divider 666. The node responsive to the reference voltage level is coupled to the selector 667. A selector 667 coupled to the sense amplifier 650 in the configuration of FIG. 13 is responsive to the sequence signal on line 656 to provide a sequence of reference voltages V ref on line 652.

第15圖繪示了配置於產生電壓Vmin目的之參考行的一種配置,該配置可用於如第13圖和第14圖的參考所描述在應用於感測電路中產生參考電壓的目的。在此範例中,n行上的三個單元操作串線路680被配置於乘積和運算,其中該單元具有輸入X1到X3以及權重W1n,W2n和W3n。該權重依據乘積和運算項被編程到操作串線路680中,以被執行。由流經串線路的電流In所產生的電壓表示為VnFIG. 15 shows a configuration of the reference line configured for the purpose of generating the voltage V min , which can be used for the purpose of generating the reference voltage in the sensing circuit as described in the reference of FIGS. 13 and 14. In this example, three unit operation string lines 680 on n rows are configured for product-sum operation, where the unit has inputs X 1 to X 3 and weights W 1n , W 2n and W 3n . The weight is programmed into the operation string circuit 680 according to the product sum operation term to be executed. The voltage generated by the current I n flowing through the string line is expressed as V n .

行參考串線路681實施於陣列中,行參考串線路681使用三個單元,該三個單元可具有匹配於操作串線路680中所使用的那三個單元之電特性。為了產生電壓Vmin,在行參考串線路681表示為W1ref,W2ref和W3ref的單元權重都被設定成對應於低閾值狀態的數值(在這種情況下為“1”)。行參考串線路681中單元的輸入被綁在一起,並在操作期間耦接到電壓VON,以使行參考串線路681中的所有單元都導通,並產生小電壓降VdSm。因此,在此範例中電壓Vmin約等於3 * VdSm,或是在操作串線路 680中使用的單位單元的小電壓降的三倍。下面的表格4繪示了用於特定輸入和操作串線路的權重配置(計算行)和參考行之操作範例。 The row reference string line 681 is implemented in an array. The row reference string line 681 uses three cells, which can have electrical characteristics matching those of the three cells used in the operation string line 680. In order to generate the voltage V min , the line reference string line 681 is represented as W 1ref , and the cell weights of W 2ref and W 3ref are both set to values corresponding to the low threshold state (“1” in this case). The inputs of the cells in the row reference string line 681 are tied together and are coupled to the voltage V ON during operation, so that all the cells in the row reference string line 681 are turned on, and a small voltage drop V dSm is generated . Therefore, in this example, the voltage V min is approximately equal to 3*V dSm , or three times the small voltage drop of the unit cell used in the operation string circuit 680. Table 4 below shows an example of the operation of the weight configuration (calculation row) and reference row for specific input and operation string lines.

Figure 108101265-A0305-02-0026-6
Figure 108101265-A0305-02-0026-6

在參考串線路僅用於產生Vmin的實施例中,用於參考電壓電路的數值Vmax可以被設定為足夠高的數值,以提供良好的操作幅度於裝置。第15圖所繪示的範例是依據包括三個可變電阻單元的串線路。 In an embodiment where the reference string circuit is only used to generate V min , the value V max for the reference voltage circuit can be set to a sufficiently high value to provide a good operating range for the device. The example shown in FIG. 15 is based on a string circuit including three variable resistance units.

在本技術的實施例中,可變電阻單元可實施於使用NAND類技術的大規模陣列。因此,耦接到串線路中的任何特定的行單元可具有例如16、32、64或更多個單元。在任何特定乘積和運算的配置中,少於特定行中的所有單元可被使用。 In the embodiments of the present technology, the variable resistance unit may be implemented in a large-scale array using NAND-like technology. Therefore, any particular row unit coupled into the string line may have, for example, 16, 32, 64 or more units. In any particular product-sum operation configuration, fewer than all cells in a particular row can be used.

第16圖繪示了包括在操作行和參考行上的區域692中具有多個未使用單元的操作行中操作行n和參考串線路691的操作串線路690之示例配置。此範例中參考行被配置於產生電壓Vmin的目的,其可以被用作如第13圖和第14圖的參考所描述於應用在感測電路中產生參考電壓的目的。 FIG. 16 shows an example configuration of the operation string line 690 including the operation line n and the reference string line 691 in the operation line having a plurality of unused cells in the area 692 on the operation line and the reference line. In this example, the reference line is configured for the purpose of generating the voltage V min , which can be used for the purpose of generating the reference voltage in the sensing circuit as described in the reference of FIGS. 13 and 14.

在所示的範例中,操作行n上的三個單元操作串線路690被配置於乘積和運算,其中操作串線路690上的單元具有輸入X1到X3以及 權重W1n,W2n和W3n。該些權重依據乘積和運算項,被編程到操作串線路690中,以被執行。操作行n上未使用的單元為特定輸入Y1和Y2以及權重W4n和W5n。由流經串線路的電流In所產生的電壓表示為Vn。輸入Y1和Y2以及權重W4n和W5n被配置,以便在乘積和運算期間操作行n中的未使用單元為導通。 In the example shown, the three unit operation string lines 690 on operation line n are configured for product-sum operation, where the units on operation string line 690 have inputs X 1 to X 3 and weights W 1n , W 2n and W 3n . These weights are programmed into the operation string circuit 690 according to the product sum operation term to be executed. The unused cells on operation line n are specific inputs Y 1 and Y 2 and weights W 4n and W 5n . The voltage generated by the current I n flowing through the string line is expressed as V n . The inputs Y 1 and Y 2 and the weights W 4n and W 5n are configured so that the unused cells in the operation row n are turned on during the product-sum operation.

參考串線路691實施於陣列中或參考陣列中,在參考行中使用三個單元的參考串線路691可具有匹配於用在操作串線路690中那三個單元的電特性。參考行上包括參考串線路691的未使用單元具有權重W4ref和W5ref。為了產生電壓Vmin,參考串線路691中的單元權重表示為W1ref、W2ref和W3ref,以及在行的未使用部分中具有權重W4ref和W5ref的單元權重都被設定為對應於低閾值狀態的數值(在這種情況下“1”)。參考行上的區域692中行的未使用部分中參考串線路691的單元中的電阻可以具有一固定值R。該電阻匹配數值R於操作串線路690中的單元以及在同行中參考行上的區域692的單元數值R是操作串線路690。在操作期間參考串線路691中包括未使用單元的單元輸入被綁在一起並且耦接到電壓VON,使得行中包括參考串線路691的所有單元都導通,並產生小電壓降VdSm。因此,在該範例的串線路中具有五個單元的電壓Vmin約等於5 * VdSm,或是操作串線路680中使用的單位單元之小電壓降的五倍。隨著串線路中有更多單元,Vmin值將相應地轉移。 The reference string line 691 is implemented in an array or a reference array, and the reference string line 691 using three cells in the reference row may have electrical characteristics matching those of the three cells used in the operation string line 690. Unused cells including reference string line 691 on the reference line have weights W 4ref and W 5ref . In order to generate the voltage V min , the unit weights in the reference string line 691 are expressed as W 1ref , W 2ref and W 3ref , and the unit weights with weights W 4ref and W 5ref in the unused portion of the row are set to correspond to low The value of the threshold state ("1" in this case). The resistance in the cell of the reference string line 691 in the unused portion of the row in the region 692 on the reference row may have a fixed value R. The resistance matching value R is for the cell in the operation string line 690 and the cell value R for the region 692 on the reference line in the same line is the operation string line 690. During operation, the cell inputs of the reference string line 691 including unused cells are tied together and coupled to the voltage V ON , so that all cells in the row including the reference string line 691 are turned on, and a small voltage drop V dSm is generated . Therefore, the voltage V min with five cells in the string line of this example is approximately equal to 5*V dSm , or five times the small voltage drop of the unit cell used in the string line 680. As there are more units in the string line, the V min value will shift accordingly.

下面的表格5繪示了對於第16圖配置的特定輸入和操作串線路的權重配置(計算行)和參考行之操作範例。 Table 5 below shows an example of the operation of the weight configuration (calculation line) and reference line for the specific input and operation string lines configured in FIG. 16.

表格5

Figure 108101265-A0305-02-0028-7
Form 5
Figure 108101265-A0305-02-0028-7

第17圖繪示了電壓Vmin和Vmax兩者被產生的示例配置。在此配置中,操作行n中的操作串線路700包括如第15圖和第16圖範例中的三個單元。因此,操作行n被配置於乘積和運算,其中操作串線路700中的單元具有輸入X1至X3和權重W1n,W2n和W3n。該權重依據乘積和運算項被編程到操作串線路700中,以被執行。操作行n上未使用單元被表示為輸入Y1和Y2以及權重W4n和W5n。由流經串線路的電流In所產生的電壓表示為Vn。輸入Y1和Y2以及權重W4n和W5n被配置,以在乘積和運算期間操作行n中的未使用單元為導通。 FIG. 17 shows an example configuration in which both voltages V min and V max are generated. In this configuration, the operation string line 700 in the operation row n includes three units as in the examples of FIGS. 15 and 16. Therefore, the operation line n is configured in a product-sum operation, in which the cells in the operation string line 700 have inputs X1 to X3 and weights W 1n , W 2n and W 3n . The weight is programmed into the operation string circuit 700 according to the product sum operation term to be executed. Unused cells on operation line n are represented as inputs Y1 and Y2 and weights W 4n and W 5n . The voltage generated by the current I n flowing through the string line is expressed as V n . The inputs Y1 and Y2 and the weights W 4n and W 5n are configured to turn on the unused cells in the operation row n during the product-sum operation.

在陣列區域703中Vmin參考行包括參考串線路701和未使用的單元。參考串線路701包含可具有匹配於用在操作串線路700中那三個單元的電特性。Vmin參考行上的未使用單元包括參考串線路691權重

Figure 108101265-A0305-02-0028-15
Figure 108101265-A0305-02-0028-16
。為了產生電壓Vmin,參考串線路701中的單元權重表示為
Figure 108101265-A0305-02-0028-17
Figure 108101265-A0305-02-0028-18
Figure 108101265-A0305-02-0028-19
,以及在行的未使用部分中具有權重
Figure 108101265-A0305-02-0028-20
Figure 108101265-A0305-02-0028-21
的未使用單元的權重都被設定為對應於低閾值狀態的數值(在這種情況下“1”)。在參考串線路701的單元和行的未使用部分中的電阻可以具有一固定值R。該電阻匹配數值R於操作串線路700中的單元以及在陣列區域703的單元 數值R是操作行n。在操作期間的Vmin參考行中包含參考串線路701和未使用單元的單元輸入被綁在一起並且耦接到電壓VON,使得在Vmin參考行中包括參考串線路701的所有單元都導通,並假設Iref等於In,以產生小電壓降VdSm。因此,在該範例的串線路中具有五個單元的電壓Vmin約等於5 * VdSm,或是操作串線路700中使用的單位單元之小電壓降的五倍。隨著串線路中有更多單元,Vmin值將相應地轉移。 The V min reference line in the array area 703 includes a reference string line 701 and unused cells. The reference string line 701 contains electrical characteristics that can be matched to those three units used in operating the string line 700. Unused cells on V min reference line include reference string line 691 weight
Figure 108101265-A0305-02-0028-15
with
Figure 108101265-A0305-02-0028-16
. To generate the voltage V min , the unit weight in the reference string line 701 is expressed as
Figure 108101265-A0305-02-0028-17
,
Figure 108101265-A0305-02-0028-18
with
Figure 108101265-A0305-02-0028-19
, And have weight in the unused portion of the row
Figure 108101265-A0305-02-0028-20
with
Figure 108101265-A0305-02-0028-21
The weights of unused cells are all set to the value corresponding to the low threshold state ("1" in this case). The resistance in the unused portions of the cells and rows of the reference string line 701 may have a fixed value R. The resistance matching value R is for the cell in the operation string line 700 and the cell value R in the array area 703 is the operation line n. The cell inputs containing the reference string line 701 and unused cells in the V min reference line during operation are tied together and coupled to the voltage V ON so that all cells including the reference string line 701 in the V min reference line are turned on , and assuming that I ref equal to I n, to generate a small voltage drop V dSm. Therefore, the voltage V min with five cells in the string circuit of this example is approximately equal to 5*V dSm , or five times the small voltage drop of the unit cells used in the string circuit 700. As there are more units in the string line, the V min value will shift accordingly.

在陣列區域703中Vmax參考行包括參考串線路702和未使用的單元。參考串線路702包含可具有匹配於用在操作串線路700中那三個單元的電特性。Vmax參考行上包括參考串線路702的未使用單元權重

Figure 108101265-A0305-02-0029-10
Figure 108101265-A0305-02-0029-11
。為了產生電壓Vmax,參考串線路702中的單元權重表示為
Figure 108101265-A0305-02-0029-12
Figure 108101265-A0305-02-0029-13
Figure 108101265-A0305-02-0029-14
,以及在Vmax參考行的未使用部分中具有權重W4Href和W5Href的未使用單元的權重在Vmin行中都被設定為對應於高閾值狀態的數值(在這種情況下“0”)。在參考串線路701的單元和行的未使用部分中的電阻可以具有一固定值R。該電阻匹配數值R於操作串線路700中的單元以及在陣列區域703的單元數值R是操作行n。在操作期間Vmax參考行中包含參考串線路702的單元輸入被綁在一起並且耦接到電壓VOFF,以及未使用部分耦接到電壓VON,使得在行的三個單元中包括Vmax參考串線路701的電晶體都導通,並假設Iref等於In,以產生大電壓降VdLg。因此,在此範例的串線路中具有五個單元的電壓Vmax約等於3 * VdLg,或是操作串線路700中使用的單位單元之大電壓降的三倍。隨著串線路中有更多單元,Vmax值將相應地轉移。 The V max reference line in the array area 703 includes the reference string line 702 and unused cells. The reference string line 702 includes electrical characteristics that can be matched to those three units used in operating the string line 700. The unused unit weights of the reference string line 702 on the V max reference line
Figure 108101265-A0305-02-0029-10
with
Figure 108101265-A0305-02-0029-11
. To generate the voltage V max , the unit weight in the reference string line 702 is expressed as
Figure 108101265-A0305-02-0029-12
,
Figure 108101265-A0305-02-0029-13
with
Figure 108101265-A0305-02-0029-14
, And the weights of unused cells with weights W4 Href and W5 Href in the unused portion of the V max reference line are set to values corresponding to the high threshold state in the V min line (in this case "0" ). The resistance in the unused portions of the cells and rows of the reference string line 701 may have a fixed value R. The resistance matching value R is for the cell in the operation string line 700 and the cell value R in the array area 703 is the operation line n. During operation, the inputs of the cells containing the reference string line 702 in the V max reference row are tied together and coupled to the voltage V OFF , and the unused part is coupled to the voltage V ON , so that V max is included in the three cells of the row reference crystal string line 701 are turned on, and assuming that I ref equal to I n, to generate a large voltage drop V dLg. Therefore, the voltage V max with five cells in the string circuit in this example is approximately equal to 3*V dLg , or three times the large voltage drop of the unit cell used in the string circuit 700. As there are more cells in the string, the V max value will shift accordingly.

下面的表格6繪示了對於第17圖配置的特定輸入和操作串 線路的權重配置(計算行)和參考行之操作範例。 Table 6 below shows the specific input and operation strings for the configuration in Figure 17 Example of operation of line weight configuration (calculation line) and reference line.

Figure 108101265-A0305-02-0030-8
Figure 108101265-A0305-02-0030-8

在參考第12圖和第15-17圖描述的實施例中,可編程電阻單元陣列被配置成具有一輸入Xi的功能組,並且包括一個成員單元,操作行中各單元實施於具有一位元二進制權重Wi的項XiWi,XiWi由單元中的可編程閾值電晶體的閾值來決定。單元中電阻器的電阻R和串線路中的電流In是常數。 In the embodiments described with reference to FIGS. 12 and 15-17, the programmable resistance cell array is configured as a functional group with an input X i and includes a member cell, and each cell in the operation row is implemented with one bit binary metadata item weights W i X i W i, X i W i is determined by the threshold of the programmable threshold transistor unit. The resistance R of the resistor in the cell and the current In in the string line are constant.

在一些實施例中,可編程電阻單元陣列可被配置成具有一個輸入和多個成員的功能組,以實施乘積和運算的項XiWi,其中使用編程於單元中可編程電晶體中的單一位元值之權重Wi可以是除了一二進位元“0”或“1”之外的值,例如多位元二進制值。 In some embodiments, the programmable resistance cell array may be configured as a functional group with one input and multiple members to implement the product sum operation X i W i , where the The weight W i of a single bit value may be a value other than a binary “0” or “1”, for example, a multi-bit binary value.

第18-22圖繪示了實現多位元二進制值的功能組之一些範例 配置。 Figures 18-22 illustrate some examples of functional groups that implement multi-bit binary values Configuration.

第18圖繪示了在陣列的n行中的單一串線路上包括三個成員單元功能組。該行接收固定電流In。輸入值Xm連接到所有三列中的單元中的電晶體閘極。在此範例中,組中三個單元中的電阻器R1,mn、R2,mn和R3,mn的電阻是不同的。因此,電阻器R3具有電阻R,電阻器R2具有電阻2 * R,以及電阻器R1具有電阻4 * R。因此,依據從0 * R(功能組中的所有電晶體導通)至7 * R(功能組中的所有電晶體截止)不等的有效電阻所產生的組合之功能組權重具有一個三位元二進制值,範圍從0到7。使用第18圖功能組來實現的乘積和運算項可以表示為Xm(W1 * 4R+W2 * 2R+W3 * R)。在其他實施例中,如第18圖單元中的單元功能組可以具有多於三個成員,三個成員與共同輸入Xm連接於行中。 Figure 18 illustrates the inclusion of three member cell functional groups on a single string in n rows of the array. This row receives a fixed current I n . The input value X m is connected to the transistor gates in the cells in all three columns. In this example, the resistances of the resistors R 1,mn , R 2,mn, and R 3,mn in the three cells in the group are different. Therefore, the resistor R3 has the resistance R, the resistor R2 has the resistance 2*R, and the resistor R1 has the resistance 4*R. Therefore, the weight of the combined function group based on the effective resistance ranging from 0*R (all transistors in the function group turned on) to 7*R (all transistors in the function group turned off) has a three-bit binary Value, ranging from 0 to 7. The product and operation terms implemented using the function group in Figure 18 can be expressed as X m (W1 * 4R+W2 * 2R+W3 * R). In other embodiments, the unit function group as in the unit of FIG. 18 may have more than three members, and the three members are connected in a row with the common input X m .

如上所討論的單元陣列可以使用邏輯電路來配置於實現乘積和運算,如上所討論的單元陣列可以使用多個功能組來配置於運算項的形成。 The cell array as discussed above can be configured using logic circuits to implement the product-sum operation, and the cell array as discussed above can be configured using multiple functional groups to form the calculation terms.

第19圖繪示了陣列的三個不同行n1,n2和n3上一列陣列中包括三個成員單元的功能組。三行中的每一行接收固定電流In。輸入值Xm連接到列中單元中的電晶體閘極。在此範例中,組中的三個單元中電阻器的電阻R1 mn、R2,mn和R3,mn是不同的。因此,電阻器R3具有電阻R,電阻器R2具有電阻2 * R,以及電阻器R1具有電阻4 * R。在各行中產生的電壓Vn1,Vn2和Vn3總和於外圍電路,以提供總和輸出項。 Figure 19 shows a functional group of three member cells in an array of three different rows n1, n2, and n3 of the array. Each of the three rows receives a fixed current In. The input value X m is connected to the transistor gate in the cell in the column. In this example, the resistances R 1 mn , R 2,mn, and R 3,mn of the resistors in the three cells in the group are different. Therefore, the resistor R3 has the resistance R, the resistor R2 has the resistance 2*R, and the resistor R1 has the resistance 4*R. The voltages V n1 , V n2 and V n3 generated in each row are summed in the peripheral circuit to provide a sum output term.

使用第19圖的功能組所實施的乘積和運算項可以表示為Xm(W1 * I4R+W2 * I2R+W3 * IR),其具有產生代表項的一部分電壓之各 行。因此,依據從0 * IR(功能組中的所有電晶體導通)至7 * IR(功能組中的所有電晶體截止)不等的電壓所產生的組合之功能組權重具有一個三位元二進制值,範圍從0到7。 The product and operation terms implemented using the functional group of FIG. 19 can be expressed as X m (W1 * I4R+W2 * I2R+W3 * IR), which has rows that generate a part of the voltage of the representative term. Therefore, the weight of the combined function group based on a voltage ranging from 0*IR (all transistors in the function group turned on) to 7*IR (all transistors in the function group turned off) has a three-bit binary value , Ranging from 0 to 7.

配置於執行總和的周圍電路可包括類比總和放大器或數位邏輯。在一個範例中,各行上的電壓可以依次被感測,並且添加於算術邏輯中的每個感測步驟結果如第13圖所示的範例。 The peripheral circuit configured to perform the sum may include an analog sum amplifier or digital logic. In one example, the voltage on each row can be sensed in sequence, and the result of each sensing step added to the arithmetic logic is shown in the example shown in FIG. 13.

在其他實施例中,如第19圖陣列中的單元功能組可以具有多於三個成員,三個成員與共同輸入Xm連接於行中。 In other embodiments, the unit function group as in the array of FIG. 19 may have more than three members, and the three members are connected in a row with the common input Xm .

第20圖繪示了陣列的三個不同行n1、n2和n3上一列陣列中包括三個成員單元的功能組。輸入值Xm連接到列中單元中的電晶體閘極。在此範例中,組中的三個單元中電阻器的電阻R1 mn、R2,mn和R3,mn是相同的,三行中的每一行接收不同的固定電流In。因此,電流源提供具有電流I的I3至行3,電流源提供具有電流2 * I的I2至行2,以及電流源提供具有電流4 * I的I1至行1。在功能組的行中產生的電壓Vn1、Vn2和Vn3被總和至外圍電路,以提供總和輸出項。因此,依據從0 *IR(功能組中的所有電晶體導通)至7 *IR(功能組中的所有電晶體截止)不等的輸出所產生的組合之功能組權重具有一個三位元二進制值,範圍從0到7。 Figure 20 shows a functional group including three member cells in an array on three different rows n1, n2, and n3 of the array. The input value X m is connected to the transistor gate in the cell in the column. In this example, the resistances R 1 mn , R 2,mn, and R 3,mn of the resistors in the three cells in the group are the same, and each of the three rows receives a different fixed current I n . Therefore, the current source provides I 3 to row 3 with current I, the current source provides I 2 to row 2 with current 2*I, and the current source provides I 1 to row 1 with current 4*I. The voltages V n1 , V n2 and V n3 generated in the rows of the function group are summed to the peripheral circuit to provide a sum output term. Therefore, the combined function group weights generated based on outputs ranging from 0*IR (all transistors in the function group turned on) to 7*IR (all transistors in the function group turned off) have a three-bit binary value , Ranging from 0 to 7.

使用第20圖的功能組所實施的乘積和運算項可以表示為Xm(W1 * 4IR+W2 * 2IR+W3 * IR),其具有產生代表項的一部分電壓之各行。 The product sum operation term implemented using the function group of FIG. 20 can be expressed as X m (W1 * 4IR+W2 * 2IR+W3 * IR), which has rows that generate a part of the voltage of the representative term.

被配置於執行總和的外圍電路可包括類比總和放大器或數位邏輯。在一個範例中,各行上的電壓可以依次被感測,並且添加於算術 邏輯中的每個感測步驟結果如第13圖所示的範例。 The peripheral circuit configured to perform the sum may include an analog sum amplifier or digital logic. In one example, the voltage on each row can be sensed in sequence and added to the arithmetic The result of each sensing step in the logic is the example shown in Figure 13.

在其他實施例中,如第20圖陣列中的單元功能組可以具有多於三個成員,三個成員與共同輸入Xm連接於行中。 In other embodiments, the unit function group as in the array of FIG. 20 may have more than three members, and the three members are connected in a row with the common input Xm .

第21圖繪示了陣列的三個不同行n1、n2和n3上一列陣列中包括三個成員單元的功能組。輸入值Xm連接到列中單元中的電晶體閘極。在此範例中,組中的三個單元中的電阻器的電阻R1 mn、R2,mn和R3,mn是相同的,三行中的每一行接收不同的固定電流In,在功能組的行中產生的電壓Vn1、Vn2和Vn3分別被除以4、2和1,然後總和於外圍電路,以提供總和輸出項。因此,依據從0 *IR(功能組中的所有電晶體導通)至7 *IR(功能組中的所有電晶體截止)不等的輸出所產生的組合之功能組權重具有一個三位元二進制值,範圍從0到7。 Figure 21 shows a functional group including three member cells in an array on three different rows n1, n2, and n3 of the array. The input value X m is connected to the transistor gate in the cell in the column. In this example, the resistances R 1 mn , R 2,mn and R 3,mn of the resistors in the three units in the group are the same, each of the three rows receives a different fixed current I n , in the function The voltages Vn1 , Vn2, and Vn3 generated in the rows of the group are divided by 4, 2, and 1, respectively, and then summed to the peripheral circuit to provide a summed output term. Therefore, the combined function group weights generated based on outputs ranging from 0*IR (all transistors in the function group turned on) to 7*IR (all transistors in the function group turned off) have a three-bit binary value , Ranging from 0 to 7.

使用第21圖的功能組所實施的乘積和運算項可以表示為Xm(W1 * 4IR+W2 * 2IR+W3 * IR),其具有產生被分成外圍電路的電壓之各行以代表項的一部分。 The product and operation terms implemented using the functional group of FIG. 21 can be expressed as X m (W1 * 4IR+W2 * 2IR+W3 * IR), which has rows that generate voltages divided into peripheral circuits to represent a part of the terms.

配置於執行總和的周圍電路可包括類比總和放大器或數位邏輯。在一個範例中,各行上的電壓可以依次被感測,並且添加於算術邏輯中的每個感測步驟結果如第13圖所示的範例。 The peripheral circuit configured to perform the sum may include an analog sum amplifier or digital logic. In one example, the voltage on each row can be sensed in sequence, and the result of each sensing step added to the arithmetic logic is shown in the example shown in FIG. 13.

在其他實施例中,如第21圖陣列中的單元功能組可以具有多於三個成員,三個成員與共同輸入Xm連接於行中。 In other embodiments, the unit function group as in the array of FIG. 21 may have more than three members, and the three members are connected in a row with the common input Xm .

第22圖繪示了陣列的兩個不同行n1和n2上包含於一列陣列中的兩個單元以及陣列第二列的兩個單元之四個成員單元的功能組。輸入值Xm連接至功能組的兩列之所有單元中的電晶體閘極。在此範例中,組 中的四個單元中的電阻器的電阻R1 mn、R2,mn、R3,mn和R4,mn是不同的。因此,電阻器R3和R4具有電阻R,電阻器R1和R2具有電阻4 * R。在兩行的每一個接收不同的固定電流In。因此,電流源提供具有電流I的I2至行2,電流源提供具有電流2 * I的I1至行1。在功能組的兩行中產生的電壓Vn1、Vn2總和於外圍電路,以提供總和輸出項。 FIG. 22 shows the functional group of two member cells contained in one array of two different rows n1 and n2 of the array and two member cells of the second column of the array. The input value X m is connected to the transistor gates in all units of the two columns of the function group. In this example, the resistances R 1 mn , R 2,mn , R 3,mn and R 4,mn of the resistors in the four cells in the group are different. Therefore, resistors R3 and R4 have resistance R, and resistors R1 and R2 have resistance 4*R. Each of the two rows receives a different fixed current I n . Therefore, the current source provides I 2 to row 2 with current I, and the current source provides I 1 to row 1 with current 2*I. The voltages V n1 and V n2 generated in the two rows of the function group are summed in the peripheral circuit to provide a sum output item.

使用第22圖的功能組所實施的乘積和運算項可以表示為Xm(W1*2I*4R+W2*I*4R+W3*2I*R+W

Figure 108101265-A0305-02-0034-9
4*I*R),其具有產生代表項的一部分電壓之各行。因此,依據從0 *IR(功能組中的所有電晶體導通)至15 *IR(功能組中的所有電晶體截止)不等的輸出所產生的組合之功能組權重具有一個四位元二進制值,範圍從0到15。 The product and operation terms implemented using the function group in Figure 22 can be expressed as X m (W1*2I*4R+W2*I*4R+W3*2I*R+W
Figure 108101265-A0305-02-0034-9
4*I*R), which has rows that generate a portion of the voltage representing the term. Therefore, the combined function group weights generated from outputs ranging from 0*IR (all transistors in the function group turned on) to 15*IR (all transistors in the function group turned off) have a four-bit binary value , Ranging from 0 to 15.

配置於執行總和的周圍電路可包括類比總和放大器或數位邏輯。在一個範例中,各行上的電壓可以依次被感測,並且添加於算術邏輯中的每個感測步驟結果如第13圖所示的範例。 The peripheral circuit configured to perform the sum may include an analog sum amplifier or digital logic. In one example, the voltage on each row can be sensed in sequence, and the result of each sensing step added to the arithmetic logic is shown in the example shown in FIG. 13.

在其他實施例中,如第22圖陣列中單元的功能組可具有多於三個成員,單元的功能組與共同輸入Xm連接於行中。 In other embodiments, as shown in FIG. 22, the functional group of the cells in the array may have more than three members, and the functional group of the cells and the common input X m are connected in a row.

其他功能組配置也可以被使用。 Other functional group configurations can also be used.

巨大的可編程電阻單元陣列可以配置在操作之間,以執行具有總和項的各種函數的複合乘積和運算,如同各計算執行的所需。此外,總和項的係數(即權重)能以非揮發性的形式設置於單元的電晶體中,並且總和項的係數藉由如各計算執行所需的編程和抹除操作來改變。 A huge array of programmable resistance cells can be configured between operations to perform complex product sum operations of various functions with sum terms, as required for the execution of each calculation. In addition, the coefficient of the sum term (that is, the weight) can be set in the transistor of the cell in a non-volatile form, and the coefficient of the sum term is changed by performing the required programming and erasing operations as in each calculation.

第23圖是積體電路901的簡化晶片方塊圖,積體電路901包括具有電壓感測的乘積和陣列,以及如第5圖和第6圖與第10A/10B圖 所繪示的隱藏通道單元,隱藏通道單元被配置為神經形態的記憶體陣列960。 FIG. 23 is a simplified chip block diagram of an integrated circuit 901, which includes a product and an array with voltage sensing, and as shown in FIGS. 5 and 6 and FIGS. 10A/10B The hidden channel unit shown is configured as a neuromorphic memory array 960.

字元線驅動器940耦接到多個字元線945。驅動器包括如在一些實施例中數位類比轉換器產生每個選定線的輸入變數x(i),或者在替代方案中,二進制字元線驅動器可以應用於二進制輸入。行解碼器970經由線965耦接到神經形態的記憶體陣列960中的行而排列的一或多層之串聯單元串線路,以用於選擇讀取乘積和數據串線路與將參數數據寫入至神經形態的記憶體陣列960。在匯流排930上來自控制邏輯(控制器)910的位址被提供至行解碼器970和字元線驅動器940。電壓感測感測放大器經由線975耦接到行解碼器,並且電壓感測放大器依次耦接到緩衝電路980。施加負載電流In的電流源與感測電路耦接。編程緩衝器可以包含於緩衝電路980中的感測放大器中,以儲存編程數據於單元中的編程閾值電晶體的兩級或多級編程。此外,控制邏輯910可以包括用於選擇性地使用編程以及禁止陣列中的串線路電壓以響應於編程緩衝器中的編程數據值之電路。 The word line driver 940 is coupled to a plurality of word lines 945. The driver includes a digital analog converter that generates an input variable x(i) for each selected line as in some embodiments, or in the alternative, a binary word line driver can be applied to the binary input. The row decoder 970 is coupled to one or more serial cell string lines arranged in a row in the neuromorphic memory array 960 via line 965 for selecting the reading product and data string lines and writing the parameter data to Neuromorphic memory array 960. The address from the control logic (controller) 910 on the bus 930 is provided to the row decoder 970 and the word line driver 940. The voltage sense sense amplifier is coupled to the row decoder via line 975, and the voltage sense amplifier is sequentially coupled to the buffer circuit 980. Applying a load current I n current source is coupled to the sensing circuit. The programming buffer may be included in the sense amplifier in the buffer circuit 980 to store the programming data in the two-level or multi-level programming of the programming threshold transistor in the cell. In addition, the control logic 910 may include circuitry for selectively using programming and disabling string line voltages in the array in response to programming data values in the programming buffer.

來自感測放大器的感測數據經由第二數據線985被提供至數據緩衝器990,數據緩衝器990經由數據路徑993依次耦接到輸入/輸出電路991。感測放大器可包括被配置於施加單位增益或期望增益水平的運算放大器,以及提供類比訊號至數位類比轉換器或其他信號處理或信號路由電路。附加的算術單元和路由電路可以被包含,以提供於神經形態電路中多層單元串線路的排列。 The sensed data from the sense amplifier is provided to the data buffer 990 via the second data line 985, and the data buffer 990 is sequentially coupled to the input/output circuit 991 via the data path 993. The sense amplifier may include an operational amplifier configured to apply unity gain or a desired gain level, and provide an analog signal to digital analog converter or other signal processing or signal routing circuit. Additional arithmetic units and routing circuits can be included to provide the arrangement of multi-layer cell strings in neuromorphic circuits.

此外,算術單元和路由電路可以被包含,以提供於矩陣乘法單元中的串線路層排列。 In addition, an arithmetic unit and a routing circuit may be included to provide a serial line layer arrangement in the matrix multiplication unit.

輸入/輸出電路991將數據驅動到積體電路901的外部目的地。輸入/輸出數據和控制信號經由數據匯流排905移動於輸入/輸出電路991、控制邏輯910和積體電路901上的輸入/輸出端口或其他內部數據源或積體電路901外部,例如通用處理器或專用應用電路,或提供由神經形態的記憶體陣列960所支持的系統晶片功能的模組之間。 The input/output circuit 991 drives data to an external destination of the integrated circuit 901. Input/output data and control signals are moved via the data bus 905 to the input/output circuit 991, control logic 910, and input/output ports on the integrated circuit 901 or other internal data sources or external to the integrated circuit 901, such as a general-purpose processor Or a dedicated application circuit, or a module that provides the system chip function supported by the neuromorphic memory array 960.

在第23圖所示的範例中,使用偏壓配置狀態機的控制邏輯910經由電壓源來控制透過電壓供應所產生或提供的供應電壓之應用或在方塊920的供應,以用於乘積和讀取操作,以及設定參數的參數寫入操作,例如由包括電荷捕獲單元和浮動閘極單元、抹除、驗證和編程偏壓的電荷捕獲位準所表示的單元權重。控制邏輯910耦接到數據緩衝器990和神經形態的記憶體陣列960。 In the example shown in FIG. 23, the control logic 910 using the bias configuration state machine controls the application of the supply voltage generated or provided by the voltage supply or the supply at block 920 via the voltage source for product and read Fetch operations, as well as parameter writing operations that set parameters, such as the cell weights represented by charge trap levels including charge trapping cells and floating gate cells, erase, verify, and program bias. The control logic 910 is coupled to the data buffer 990 and the neuromorphic memory array 960.

使用本領域已知的專用邏輯電路可以實現控制邏輯910。在替代實施例中,控制邏輯包括通用處理器,控制邏輯可以被實施於同一積體電路上,該積體電路執行計算機程序以控制設備操作。在其他實施例中,專用邏輯電路和通用處理器的組合可用於控制邏輯的實施。 The control logic 910 can be implemented using dedicated logic circuits known in the art. In alternative embodiments, the control logic includes a general-purpose processor, and the control logic may be implemented on the same integrated circuit that executes a computer program to control the operation of the device. In other embodiments, a combination of dedicated logic circuits and general-purpose processors may be used for the implementation of control logic.

第24-26圖繪示了系統1000的配置,系統1000包括使用數據路徑控制器1003來互連的記憶體陣列1002和乘積和加速器陣列1001。乘積和加速器陣列1001包括依據上述任何實施例的可編程電阻單元陣列。記憶體陣列可包括NAND快閃陣列、SRAM陣列、DRAM陣列、NOR快閃陣列或可與乘積和加速器陣列1001協調使用的其他類型記憶體。 24-26 show the configuration of the system 1000, which includes a memory array 1002 and a product and accelerator array 1001 interconnected using a data path controller 1003. The product and accelerator array 1001 includes a programmable resistance cell array according to any of the embodiments described above. The memory array may include a NAND flash array, an SRAM array, a DRAM array, a NOR flash array, or other types of memory that can be used in coordination with the product and accelerator array 1001.

系統可以從如第24圖所示的系統外接收輸入/輸出數據,並將數據路由至記憶體陣列。數據可以包括用於配置為實現一個或多個乘積 和運算項的單元功能組之配置數據、在陣列中用於操作的的功能組權重、以及用於乘積和運算的輸入值。 The system can receive input/output data from outside the system as shown in Figure 24 and route the data to the memory array. The data may include information for configuring one or more products The configuration data of the unit function group of the sum operation item, the weight of the function group used for operation in the array, and the input value used for the product and operation.

如第25圖所示,來自記憶體陣列1002的數據可以被傳送到乘積和加速器陣列1001,其利用數據路徑控制器1003所使用的直接數據路徑可被控制。或者,透過數據路徑控制器1003的數據路徑可用於將來自記憶體陣列1002的數據傳送到乘積和加速器陣列1001,如適用於特定實現。 As shown in FIG. 25, data from the memory array 1002 can be transferred to the product and accelerator array 1001, which can be controlled using the direct data path used by the data path controller 1003. Alternatively, the data path through the data path controller 1003 can be used to transfer data from the memory array 1002 to the product and accelerator array 1001, as applicable to a particular implementation.

如第26圖所示,來自乘積加速器陣列的輸出數據可以透過數據路徑控制器1003施加到系統1000的輸入輸出數據路徑。系統1000的輸入輸出數據路徑可以耦接到處理單元,該處理單元被配置於計算權重,以提供輸入並利用乘積加速器陣列的輸出。 As shown in FIG. 26, the output data from the product accelerator array can be applied to the input and output data paths of the system 1000 through the data path controller 1003. The input and output data path of the system 1000 may be coupled to a processing unit configured to calculate weights to provide input and utilize the output of the product accelerator array.

此外,來自乘積加速器陣列1001的輸出數據可以透過數據路徑控制器1003路由回記憶體陣列1002,以用於迭代乘積和運算。 In addition, the output data from the product accelerator array 1001 can be routed back to the memory array 1002 through the data path controller 1003 for iterative product and operation.

在一些實施例中,包含記憶體、乘積和加速器陣列和數據路徑邏輯的系統1000可以實施於單一積體電路。此外,系統1000可以包括相同或不同的積體電路、算術邏輯單元、數位訊號處理器、通用CPU,狀態機以及在計算機處理的執行時被配置於類似乘積和加速器陣列1001的利用。 In some embodiments, the system 1000 including memory, product and accelerator arrays, and data path logic may be implemented in a single integrated circuit. In addition, the system 1000 may include the same or different integrated circuits, arithmetic logic units, digital signal processors, general-purpose CPUs, state machines, and the use of similar product and accelerator arrays 1001 configured during the execution of computer processing.

一種根據本文描述的任何實施例來使用可編程電阻單元陣列的方法可以使用如第24-26圖的系統來執行,使用邏輯實施於同一積體電路,耦接到積體電路,或是執行配置步驟的兩者組合,其中陣列中具有各自權重的單元功能組被編程,以及操作步驟,並且該陣列用於產生乘積和數據。 A method of using a programmable resistance cell array according to any of the embodiments described herein can be performed using the system as shown in Figures 24-26, using logic implemented on the same integrated circuit, coupled to the integrated circuit, or performing configuration The two of the steps are combined, in which the cell function groups with respective weights in the array are programmed, as well as the operation steps, and the array is used to generate product and data.

一種用於操作可變電阻單元陣列以產生乘積和數據的方法包括在陣列中編程具有對應於相應單元的權重因子數值的閾值之可編程閾值電晶體;選擇性地施加輸入到陣列中的列單元,施加電流至陣列中對應於行單元之一行;以及在陣列中的一行或多行單元上感測電壓。 A method for operating an array of variable resistance cells to generate product and data includes programming a programmable threshold transistor having a threshold value corresponding to a weighting factor value of the corresponding cell in the array; selectively applying input to column cells in the array , Apply current to one of the rows in the array corresponding to the row of cells; and sense the voltage on one or more rows of cells in the array.

這種方法可以包括將陣列中的單元配置成包括一個或多個功能單元組;其中功能組實施代表性的乘積和函數項。每個功能組可以接收相應的輸入項,並且可以用權重編程,該權重是功能組的一個或多個功能組的可編程閾值之函數。功能組可以以各種方式配置,例如上述第18-22圖的參考。在這種方式中,在陣列中具有被配置於利用一位元二進制模式作為個別單元的權重之可編程電阻單元可被配置為具有多位元權重的功能組,可以被配置於具有多位元權重的功能組。多位元權重可以被配置於使用具有不同電阻的各電阻器之單元功能組,在功能組中的相異行上的感測期間使用不同的電流位準,使用算術邏輯去結合在功能組中具有不同權重的各自行上所感測的電壓,以及其他如本文所描述的方法。 This method may include configuring the cells in the array to include one or more functional cell groups; wherein the functional groups implement representative products and function terms. Each functional group can receive a corresponding input item and can be programmed with weights that are a function of programmable thresholds for one or more functional groups of the functional group. The functional groups can be configured in various ways, such as the references in Figures 18-22 above. In this way, a programmable resistance unit having a weight configured to use a one-bit binary mode as an individual unit in the array can be configured as a functional group having a multi-bit weight, and can be configured to have a multi-bit weight Functional group of weights. Multi-bit weights can be configured in the unit functional group using resistors with different resistances, using different current levels during sensing on different rows in the functional group, and using arithmetic logic to combine in the functional group Voltages sensed on respective rows with different weights, and other methods as described herein.

此外,在一些實施例中,系統可以被操作於使用參考行單元來產生行參考電壓,或低行參考電壓和高行參考電壓適用於特定實施方式。該方法可以包括產生感測參考電壓以作為一個或多個行參考電壓的函數。感測操作可以包括將選定行單元上的電壓與感測參考電壓進行比較,以產生表示在選定行上電壓位準的輸出。 Furthermore, in some embodiments, the system may be operated to use a reference row unit to generate a row reference voltage, or a low row reference voltage and a high row reference voltage are suitable for a particular implementation. The method may include generating a sensed reference voltage as a function of one or more row reference voltages. The sensing operation may include comparing the voltage on the selected row cell with the sensing reference voltage to produce an output that represents the voltage level on the selected row.

儘管通過參考上文詳述的優選實施方案和實施例公開了本發明,但應理解這些實施例旨在說明而不是限制。預期本領域技術人員將容易想到修改和組合,這些修改和組合將在本發明的精神和所附權利要求 的範圍內。 Although the invention has been disclosed by reference to the preferred embodiments and examples detailed above, it should be understood that these examples are intended to be illustrative and not limiting. It is expected that those skilled in the art will easily think of modifications and combinations, which will be within the spirit of the invention and the appended claims In the range.

12‧‧‧可編程閾值電晶體 12‧‧‧Programmable threshold transistor

14‧‧‧電阻 14‧‧‧Resistance

21、22、23、24‧‧‧電流源 21, 22, 23, 24 ‧‧‧ current source

26‧‧‧接地 26‧‧‧Ground

Claims (10)

一種半導體元件,包括配置於垂直NAND類串線路(String)中的一可變電阻單元陣列,該可變電阻單元陣列具有複數個可變電阻單元,在該可變電阻單元陣列中的各該些可變電阻單元包含一可編程閾值電晶體和並聯的一第一電阻;其中各該些可變電阻單元的一可變電阻為施加於各該些可變電阻單元之一控制閘極的一電壓、該可編程閾值電晶體的一閾值、各該些可變電阻單元的一電流與該第一電阻之函數。 A semiconductor element includes a variable resistance cell array disposed in a vertical NAND-like string line (String), the variable resistance cell array has a plurality of variable resistance cells, and each of the variable resistance cell arrays The variable resistance unit includes a programmable threshold transistor and a first resistor connected in parallel; wherein a variable resistance of each of the variable resistance units is a voltage applied to a control gate of each of the variable resistance units 2. A threshold of the programmable threshold transistor, a current of each variable resistance unit and a function of the first resistance. 如申請專利範圍第1項所述之半導體元件,其中該可變電阻單元陣列包括串聯的可變電阻單元之複數個NAND類串線路。 The semiconductor device as described in item 1 of the patent application range, wherein the variable resistance cell array includes a plurality of NAND-like string circuits of the variable resistance cells connected in series. 如申請專利範圍第2項所述之半導體元件,包括:複數個字元線,耦接到串聯的可變電阻單元之該些NAND類串線路;以及複數個字元線驅動器,連接到該些字元線用以施加可變閘極電壓至該些可變電阻單元中的該些可編程閾值電晶體。 The semiconductor device as described in item 2 of the patent application scope includes: a plurality of word lines, the NAND-like string lines coupled to the variable resistance unit connected in series; and a plurality of word line drivers, connected to the The word line is used to apply a variable gate voltage to the programmable threshold transistors in the variable resistance cells. 如申請專利範圍第1項所述之半導體元件,其中各該些可變電阻單元中的該可編程閾值電晶體包括一浮動閘極電荷捕獲儲存電晶體,以及在各該些可變電阻單元中的該第一電阻在浮動閘極電荷捕獲儲存電晶體中包括一埋藏式植入電阻。 The semiconductor device as described in item 1 of the patent application range, wherein the programmable threshold transistor in each of the variable resistance units includes a floating gate charge trapping storage transistor, and in each of the variable resistance units The first resistor includes a buried implant resistor in the floating gate charge trapping storage transistor. 如申請專利範圍第1項所述之半導體元件,其中各該些可變電阻單元中之該可編程閾值電晶體包括一介電電荷捕獲儲存電晶體,以及各該些可變電阻單元中的該第一電阻在該介電電荷捕獲儲存電晶體中包括一埋藏式植入電阻。 The semiconductor device as described in item 1 of the patent application range, wherein the programmable threshold transistor in each of the variable resistance units includes a dielectric charge trapping storage transistor, and the The first resistor includes a buried implant resistor in the dielectric charge trapping storage transistor. 如申請專利範圍第1項所述之半導體元件,包括:一感測放大器,連接至該可變電阻單元陣列,以響應於由一施加電流與在該可變電阻單元陣列中該些可變電阻單元的一總和所產生之一電壓。 The semiconductor device as described in item 1 of the patent application scope includes: a sense amplifier connected to the variable resistance cell array in response to an applied current and the variable resistances in the variable resistance cell array A voltage generated by a sum of cells. 一種半導體元件,包括以垂直NAND類串線路排列的複數個可變電阻單元之一陣列,該陣列中的各該些可變電阻單元包括一可編程閾值電晶體和並聯之一電阻,其中該陣列中的各該些可變電阻單元的該可編程閾值電晶體包括一分離閘極電晶體; 其中各該些可變電阻單元的一可變電阻為施加於各該些可變電阻單元之一控制閘極的一電壓、該可編程閾值電晶體的一閾值、各該些可變電阻單元的一電流與該電阻之函數。 A semiconductor element includes an array of a plurality of variable resistance units arranged in a vertical NAND-like string line, each of the variable resistance units in the array includes a programmable threshold transistor and a resistor in parallel, wherein the array The programmable threshold transistor of each of the variable resistance units in includes a separate gate transistor; Wherein a variable resistance of each of the variable resistance units is a voltage applied to a control gate of each of the variable resistance units, a threshold of the programmable threshold transistor, and each of the variable resistance units A function of current and resistance. 一種半導體元件,包括:複數個垂直串線路,該些垂直串線路具有複數個可變電阻單元;該些垂直串線路中的各該些可變電阻單元,具有一第一載流節點、一第二載流節點和一控制端子,並且包括一可編程閾值電晶體和並聯至該第一載流節點和該第二載流節點之一電阻,該可編程閾值電晶體具有連接到該控制端子之一閘極;其中:該些垂直串線路中的各該些可變電阻單元之一可變電阻是施加於該閘極之一電壓、該可編程閾值電晶體的閾值和該電阻的函數。 A semiconductor element includes: a plurality of vertical string lines, the vertical string lines having a plurality of variable resistance units; each variable resistance unit in the vertical string lines has a first current-carrying node, a first Two current-carrying nodes and a control terminal, and includes a programmable threshold transistor and a resistor connected in parallel to the first current-carrying node and the second current-carrying node, the programmable threshold transistor has a connection to the control terminal A gate; wherein: the variable resistance of each of the variable resistance units in the vertical string lines is a function of a voltage applied to the gate, the threshold of the programmable threshold transistor, and the resistance. 如申請專利範圍第8項所述之半導體元件,其中各該些可變電阻單元中的該電阻包括一電荷捕獲儲存電晶體中的一埋藏式植入電阻。 The semiconductor device as described in item 8 of the patent application range, wherein the resistance in each of the variable resistance units includes a buried implanted resistance in a charge trapping storage transistor. 如申請專利範圍第9項所述之半導體元件,包括用以編程具有多級的該電荷捕獲儲存電晶體的閾值之電路。 The semiconductor device as described in item 9 of the patent application includes a circuit for programming the threshold of the charge trapping storage transistor having multiple stages.
TW108101265A 2018-01-17 2019-01-11 Semiconductor device TWI682388B (en)

Applications Claiming Priority (8)

Application Number Priority Date Filing Date Title
US15/873,369 2018-01-17
US15/873,369 US10719296B2 (en) 2018-01-17 2018-01-17 Sum-of-products accelerator array
US15/887,166 US20190244662A1 (en) 2018-02-02 2018-02-02 Sum-of-products array for neuromorphic computing system
US15/887,166 2018-02-02
US201862753726P 2018-10-31 2018-10-31
US62/753,726 2018-10-31
US16/233,404 US10957392B2 (en) 2018-01-17 2018-12-27 2D and 3D sum-of-products array for neuromorphic computing system
US16/233,404 2018-12-27

Publications (2)

Publication Number Publication Date
TW201933352A TW201933352A (en) 2019-08-16
TWI682388B true TWI682388B (en) 2020-01-11

Family

ID=67274173

Family Applications (1)

Application Number Title Priority Date Filing Date
TW108101265A TWI682388B (en) 2018-01-17 2019-01-11 Semiconductor device

Country Status (2)

Country Link
CN (1) CN110047542B (en)
TW (1) TWI682388B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112885830B (en) * 2019-11-29 2023-05-26 芯恩(青岛)集成电路有限公司 Stacked neuron device structure and method of making the same
US11387242B2 (en) 2020-03-03 2022-07-12 Taiwan Semiconductor Manufacturing Company, Ltd. Non-volatile memory (NVM) cell structure to increase reliability

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6525394B1 (en) * 2000-08-03 2003-02-25 Ray E. Kuhn Substrate isolation for analog/digital IC chips
TW200735378A (en) * 2006-03-06 2007-09-16 Taiwan Semiconductor Mfg Co Ltd Split gate memory cell and method for fabricating the same
US20110286258A1 (en) * 2010-05-24 2011-11-24 Macronix International Co., Ltd. Nonvolatile memory device having a transistor connected in parallel with a resistance switching device
US20120044742A1 (en) * 2010-08-20 2012-02-23 Micron Technology, Inc. Variable resistance memory array architecture
TW201618284A (en) * 2014-09-23 2016-05-16 台灣積體電路製造股份有限公司 Split gate memory cell structure and method for forming the same
TW201730935A (en) * 2015-12-14 2017-09-01 台灣積體電路製造股份有限公司 Integrated circuit

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3073645B2 (en) * 1993-12-27 2000-08-07 株式会社東芝 Nonvolatile semiconductor memory device and method of operating the same
US5541875A (en) * 1994-07-01 1996-07-30 Advanced Micro Devices, Inc. High energy buried layer implant to provide a low resistance p-well in a flash EPROM array
KR20150037120A (en) * 2013-09-30 2015-04-08 에스케이하이닉스 주식회사 Resistance variable memory apparatus and manufacturing method thereof
EP3262571B1 (en) * 2016-03-11 2022-03-02 Hewlett Packard Enterprise Development LP Hardware accelerators for calculating node values of neural networks

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6525394B1 (en) * 2000-08-03 2003-02-25 Ray E. Kuhn Substrate isolation for analog/digital IC chips
TW200735378A (en) * 2006-03-06 2007-09-16 Taiwan Semiconductor Mfg Co Ltd Split gate memory cell and method for fabricating the same
US20110286258A1 (en) * 2010-05-24 2011-11-24 Macronix International Co., Ltd. Nonvolatile memory device having a transistor connected in parallel with a resistance switching device
US20120044742A1 (en) * 2010-08-20 2012-02-23 Micron Technology, Inc. Variable resistance memory array architecture
TW201618284A (en) * 2014-09-23 2016-05-16 台灣積體電路製造股份有限公司 Split gate memory cell structure and method for forming the same
TW201730935A (en) * 2015-12-14 2017-09-01 台灣積體電路製造股份有限公司 Integrated circuit

Also Published As

Publication number Publication date
CN110047542A (en) 2019-07-23
CN110047542B (en) 2021-07-27
TW201933352A (en) 2019-08-16

Similar Documents

Publication Publication Date Title
CN110047540B (en) Product term and accelerator array
US10957392B2 (en) 2D and 3D sum-of-products array for neuromorphic computing system
US20190244662A1 (en) Sum-of-products array for neuromorphic computing system
US10777566B2 (en) 3D array arranged for memory and in-memory sum-of-products operations
US10664746B2 (en) Neural network system
US11055606B2 (en) Vertically integrated neuro-processor
US9379129B1 (en) Assist gate structures for three-dimensional (3D) vertical gate array memory structure
CN110275694A (en) For generating the device and its operating method of long-pending item sum
KR102476232B1 (en) A High-Density Neuromorphic Computing Element
TW201935325A (en) Device structure for neuromorphic computing system
CN110766148B (en) Neural network system and control method thereof
JP2012150876A (en) Architecture for three-dimensional memory array
CN109003977A (en) Integrated circuit device including different types of memory cell
TWI682388B (en) Semiconductor device
US20220328506A1 (en) Layout method by buried rail for centralized anti-fuse read current
KR20020097486A (en) Ultra-late programming ROM and method of manufacture
KR102632211B1 (en) Three-dimensional stacked and type synapse array-based neuromorphic system and manufacturing method and operating method of the same
JP2022076545A (en) Semiconductor device
KR20230164302A (en) Memory cell combining short-term memory device and long-term memory device, array using the same, and operation and fabrication methods thereof