TW312042B - Multi-level high density mask read only memory - Google Patents

Multi-level high density mask read only memory Download PDF

Info

Publication number
TW312042B
TW312042B TW85111373A TW85111373A TW312042B TW 312042 B TW312042 B TW 312042B TW 85111373 A TW85111373 A TW 85111373A TW 85111373 A TW85111373 A TW 85111373A TW 312042 B TW312042 B TW 312042B
Authority
TW
Taiwan
Prior art keywords
memory
transistor
type
gate
thickness
Prior art date
Application number
TW85111373A
Other languages
Chinese (zh)
Inventor
Jeng-Jyh Gong
Shing-Ren Sheu
Original Assignee
United Microelectronics Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by United Microelectronics Corp filed Critical United Microelectronics Corp
Priority to TW85111373A priority Critical patent/TW312042B/en
Application granted granted Critical
Publication of TW312042B publication Critical patent/TW312042B/en

Links

Landscapes

  • Semiconductor Memories (AREA)

Abstract

A nonvolatile memory device comprises of: (1) multiple memory cells, which includes one memory transistor which consist of one source/drain diffusion region, one channel region for separating source/drain diffusion region, and one gate oxide, one gate in sequence located on the channel region; (2) in which each memory transistor characteristic is determined by one selected among at least four threshold voltages, and different memory transistor has different threshold voltage; (3) in which memory transistor array consists of first-type memory transistor, whose gate is formed by first material with first function, the first material determines threshold voltage of the first-type transistor; and second-type memory transistor, whose gate is formed by second material with second function, the second material determines threshold voltage of the second-type memory transistor.

Description

3120 絡 WF.DOC/002 Μ 3120 絡 WF.DOC/002 Μ 經濟部中央標準局員工消費合作社印製 Β7 五、發明説明(/ ) 本發明是有關於資料儲存的記憶電晶體的排列,且特 別是有關於一非揮發性記憶體,如唯讀記憶體(ROM)。 現代記憶體的設計受到半導體製程技術上的限制,而 要降低記憶體元件之尺寸則需更加改良製程上的技術。習 知的唯讀記憶體(ROM)由一陣列的場效電晶體所組成,每 一記憶單元(memory cell)均包含一單一場效元件;而每一場 效電晶體形成以提供電晶體特性之兩預定値其中之一,此 選擇性的電晶體特性例如是電晶體之啓始電壓(threshold voltage)。佈植雜質進入電晶體之通道區會使得電晶體具有 較低的啓始電壓,而電晶體就可施加到閘極之電壓Vcc所 打開;而沒有佈植雜質進入通道區所形成的電晶體具有較 高的啓始電壓,電晶體不會被施加到閘極之電壓Vcc打開。 二位元資料(binary data)可藉由選擇性地佈植雜質進入電晶 體之通道區來儲存於記憶體內;電晶體之通道區有佈植雜 質的可儲存邏輯上的”〇”,而通道區內無佈植雜質的則儲存 邏輯上的”1”。 現在施行的記憶電晶體尺寸設計受限於製程上的技 術,例如:可程式化唯讀記憶體,其藉由選擇性地佈植雜 質進入經佈植罩幕定義過之電晶體的通道區內;以0.5 的設計尺寸而言,對準佈植罩幕的程序不僅耗時且易產生 錯誤,增加唯讀記憶體的製造成本,而得到產品的尺寸也 非原先所設計的。元件設計上的考量也會限制到習知的唯 讀記憶體所能增加儲存資料的程度;例如,習知的唯讀記 憶體以埋入式N+型線連接一列電晶體之源/汲極擴散區,這 〆 -· ' 4 ' C讀先閱讀背面之注意事項%,填寫本頁) -裝- 訂 線 本纸張尺度適用中國國家標準(CNS ) A4規格(2丨OX 2W公釐) 0294TWF.DOC/002 A7 B7 __ 五、發明説明(> ) 些連接用的N+型線隨設計尺寸的降低而越做越小’而使得 N+型線之阻値增加也不可避免,因爲較高濃度的雜質佈植 以形成N+型線時,會伴隨使得雜質擴散進入通道區內以及 鄰近的元件區;當埋入式N+型線之阻値增加後’電阻電容 時間常數(re si stance-capacitor time constant)增加’使得唯讀 記憶體之資料讀出需要更多的時間。 要增加唯讀記憶體儲存資料的密度,必須克服許多難 題,其中一被提出用以改善唯_記憶體之資料儲存密度的 方法是在每一儲存單元裡儲存超過二位元的資料;此方法 已藉由形成一具有三組以上的預定値其中之一特性之電晶 體來達成,經由具有不同特性之電晶體’多位元的資料可 被儲存在單一記憶單元裡;雖然比起第一位元的唯讀記憶 體而言,多位元唯讀記憶體必須提供炱複雜的線路以讀出 資料,但理論上可藉由儲存多位元資料於單一記憶單元 裡,改善唯讀記憶體之儲存密度。然而,此多位元唯讀記 憶體之儲存策略在實際施行時卻遇到許多的困難於形成具 有預定可程式化性質之多位元唯讀記懷體,以準確地寫入 或讀出特定的資料。 一多位兀唯讚g己憶體之實施例已由Stark於美國專利 第4,287,57〇號中提出,在該專利中每/記憶電晶體之通道 寬度是記憶體之可程式化特性,改變通道寬度則會改變電 晶體之阻抗或導電性,其可藉由施一參考電壓於電晶體之 閘極’然後測量電晶體間之電位差而得知,故製作不同的 通道寬度可比較記憶電晶體與〜或更爹個參考電晶體間的 '5 ϋ浪尺度^用中國國家榡準(CNS〉Μ規格(--〆 --—- C請先閲讀背面之注意事項-Μ'填寫本頁) .裝. *1Τ 經濟部中央標準局員工消費合作社印製 3l20M WF.DOC/002 A7 濟 部 t 央 標 準 貝 工 消 費 合 作 社 印 製 ------B7__ 五、發明説明($ ) 電位差來完成。具有不同通道寬度之電晶體,其可藉選擇 性形成不同寬度的閘極於基底上之主動區,然後再形成自 動對準於鬧極之源/汲極擴散區,完成具有不同通道寬度之 電晶體製h。然而’此設計面臨了許多的難題;設計不同 通道寬度用以儲存資料’必須製作不闻尺寸的電晶體,故 需要足夠的空間以容納一具有最大通道寬度之電晶體在每 —記憶體所在的位置,以形成一規則的電晶體排列。因此, 如上所述’最小尺寸之兀件將不適用於此方法,且根據Stark 的專利所製作出來之唯讀記憶體浪費了許多空間。另一關 於Stark專利提到的記憶電晶體在製作上的難題是記憶電 晶體容易發生罩幕上的錯誤(例如焦距上的誤差),使得形成 通道過程上其寬度會有明顯的誤差,故必須要有足夠的容 忍度以允許此類的誤差;而提供這些容忍度更是限制了 Stark所設計用以降低元件尺寸原理的適合度。 第二個施行多位元記憶體的是由Takizawa等人於美國 專利第5,386,281號中所提出,其主要是形成不同厚度之閘 極氧化層於預備形成的記憶電晶體之通道區上,然後再選 擇性地形成電晶體於其上;藉此,電晶體閘極氧化層厚度 可一致,或一電晶體通道區上之閘電極厚度一半是160A, 另一半是320A。根據Takizawa等人提出的專利,可以進行 四種不同程度的雜質佈植,分別是:未摻雜;經由厚度一 致爲320A之閘極氧化層佈植;經由一分隔的閘極氧化層佈 植,其中一半之氧化層厚度爲320A,另一半之氧化層厚度 約爲160A。經由較厚氧化層的雜質佈植,可降低佈植到通 〆 〆 6 "巾關家標率(CNS ) A4規格(210X 297公釐〉3120 線 WF.DOC / 002 Μ 3120 網 WF.DOC / 002 Μ Printed by the Consumer Standardization Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs Β5. Description of the invention (/) The present invention relates to the arrangement of memory transistors for data storage, and it is special It is about a non-volatile memory, such as read-only memory (ROM). Modern memory design is limited by semiconductor process technology, and to reduce the size of memory devices requires more improved process technology. The conventional read-only memory (ROM) is composed of an array of field effect transistors. Each memory cell includes a single field effect element; and each field effect transistor is formed to provide transistor characteristics. One of the two predetermined values, the selective transistor characteristic is, for example, the threshold voltage of the transistor. The implantation of impurities into the channel region of the transistor will make the transistor have a lower starting voltage, and the transistor can be opened by the voltage Vcc applied to the gate; and the transistor formed without implantation of impurities into the channel region has The higher the starting voltage, the transistor will not be opened by the voltage Vcc applied to the gate. Binary data can be stored in the memory by selectively implanting impurities into the channel area of the transistor; the channel area of the transistor has a logical "〇" that can store the impurities and the channel If no impurity is planted in the area, the logical "1" is stored. The size design of memory transistors currently implemented is limited by process technology, such as programmable read-only memory, which selectively implants impurities into the channel area of the transistor defined by the implantation mask In terms of the design size of 0.5, the process of aligning the implantation screen is not only time-consuming and error-prone, but also increases the manufacturing cost of the read-only memory, and the size of the product obtained is not originally designed. Component design considerations will also limit the extent to which conventional read-only memory can increase the amount of stored data; for example, conventional read-only memory connects the source / drain diffusion of a row of transistors with an embedded N + type line Area, this 〆- · '4' C reading first read the precautions on the back%, fill out this page) -installation- binding paper size is applicable to China National Standard (CNS) A4 specification (2 丨 OX 2W mm) 0294TWF .DOC / 002 A7 B7 __ 5. Description of the invention (>) The N + type wires used for connection become smaller and smaller as the design size decreases, so that the resistance of the N + type wires increases inevitably because of the higher concentration When the impurities are implanted to form the N + type line, it will be accompanied by diffusion of impurities into the channel area and the adjacent element area; when the resistance value of the buried N + type line increases, the resistance capacitance time constant (re si stance-capacitor time constant) increase 'makes it more time to read the data in the read-only memory. To increase the density of read-only memory storage data, many difficulties must be overcome. One of the methods proposed to improve the data storage density of _memory is to store more than two bits of data in each storage unit; this method It has been achieved by forming a transistor with one of more than three sets of predetermined values. Through transistors with different characteristics, multi-bit data can be stored in a single memory unit; although compared to the first As far as read-only memory is concerned, multi-bit read-only memory must provide a complicated circuit to read data, but in theory, by storing multi-bit data in a single memory unit, the read-only memory can be improved. Storage density. However, the storage strategy of this multi-bit read-only memory encountered many difficulties in the actual implementation of forming a multi-bit read-only memory with predetermined programmable properties to accurately write or read specific data of. An example of more than one embodiment of the memory has been proposed by Stark in US Patent No. 4,287,57. In this patent, the channel width of each memory transistor is a programmable characteristic of the memory, which changes The channel width will change the impedance or conductivity of the transistor. It can be known by applying a reference voltage to the gate of the transistor and then measuring the potential difference between the transistors, so making different channel widths can compare memory transistors. With ~ or more reference transistors, the '5 ϋ wave scale ^ use the Chinese National Standard (CNS> Μ specifications (--〆 ---- C please read the notes on the back-Μ' to fill in this page) .Installed. * 1Τ Printed by the Employee Consumer Cooperative of the Central Bureau of Standards of the Ministry of Economic Affairs 3l20M WF.DOC / 002 A7 Printed by the Ministry of Economic Affairs, Central Standard Beige Consumer Cooperatives ------ B7__ V. Invention description ($) Potential difference to complete .Transistors with different channel widths can be formed by selectively forming gates of different widths on the active area on the substrate, and then forming source / drain diffusion areas that are automatically aligned to the gate to complete the different channel widths. Transistor made h. However 'this The design faces many difficulties; designing different channel widths to store data 'must produce transistors of insignificant size, so sufficient space is needed to accommodate a transistor with the largest channel width at each memory location, to A regular transistor arrangement is formed. Therefore, as mentioned above, the smallest size element will not be suitable for this method, and the read-only memory made according to the Stark patent wastes a lot of space. The difficulty in the production of the memory transistor is that the memory transistor is prone to errors on the cover screen (such as errors in the focal length), so that the width of the channel formation process will have obvious errors, so it must have sufficient tolerance In order to allow such errors; providing these tolerances limits the suitability of Stark's design to reduce the size of the device. The second implementation of multi-bit memory is U.S. Patent No. 5,386,281 by Takizawa et al. It is proposed that it is mainly to form gate oxide layers with different thicknesses on the channel region of the memory transistor to be formed, and then Selectively form a transistor on it; thereby, the thickness of the gate oxide layer of the transistor can be uniform, or the thickness of the gate electrode on one transistor channel area is half 160A and the other half is 320A. According to the patent proposed by Takizawa et al. , Can be implanted in four different degrees of impurities, namely: undoped; implanted through a gate oxide layer with a uniform thickness of 320A; implanted through a separate gate oxide layer, of which half of the oxide layer thickness is 320A, the thickness of the other half of the oxide layer is about 160A. Impurity implantation through a thicker oxide layer can reduce the implantation rate to the standard of the "Shenzhen Pass 6" (CNS) A4 specification (210X 297mm>

請 先 閱Please read first

II

II

裝 訂 線 0294TWF.DOC/002 A7 0294TWF.DOC/002 A7 經濟部中央標準局貝工消費合作社印製 B7 -----------------——— 五、發明説明(士) 道區內之雜質濃度,故此四種雜質佈植可在記億體內至作 出具有四種雜質濃度之通道區。而具有不同雜質濃度之通 道區可製作出具有不同啓始電壓之電晶體;啓始電壓之偵 測可先施以不同的閘極電壓於記憶體上,當電晶體開啓時 便可偵測的到。 然而,Takizawa之唯讀記憶體在施行上也有許多的困 難,最重要的是需利用精確的罩幕自動對準程序以形成 Takizawa唯讀記憶體,而精確的罩幕自動對準程序則需施 以兩化層結構唯讀記憶體之通道區上,且此結構之尺 寸^^之一半,故罩幕自動對準精確度之要求比製作 記憶電晶體上的製程還來的嚴格。若以此設計原則來製作 記憶體,製作出如Takizawa專利中所提的具有精確對準於 分隔的通道氧化層罩幕將是不可行的。因此,根據此原理 的Takizawa唯讀記憶體,其並未善用到所有小尺寸元件之 優點。 有鑒於此,本發明提出一更容易大量製造之多位元唯 讀記憶體元件,該元件是非揮發性的且每一記憶元件包含 有多重記憶單元,每一記憶單元含有一記憶體,其具有一 源/汲極擴散區,一分隔源/汲極擴散區之通道區,一覆蓋通 道區之閘極氧化層及閘電極。每一記憶體之性質由啓始電 壓來定義,其是選自於至少三種預定値其中之一,不同的 電晶體具有不同的啓始電壓。記憶體陣列包含有第一型記 憶體,其包含具有第一功能之第一材質組成的閘電極,而 第一材質部份決定第一電晶體的啓始電壓大小;以及第二 ' 7 ' ϋ張尺度適用中國國家標準(CNS ) Λ4規格(2丨0X297公釐) IU---------裝L--^---訂------線 .C請先閲讀背面之注意事項弄填寫本頁} 0294TWF.DOC/002 A7 0294TWF.DOC/002 A7 經濟部中央標準局員工消費合作社印裝/ B7 五、發明説明(夂) 型記憶體,其包含有第二功能之第二材質組成的閘電極, 而第二材質部份決定第二電晶體之啓始電壓大小。 本發明之另一特點是提供一包含複數個記憶單元之非 揮發性記憶元件,每一記憶單元含有一記憶電晶體,這些 記憶電晶體包含有一源/汲極擴散區,一分隔源/汲極擴散區 之通道區,一覆蓋於通道區上的閘極氧化層及閘電極。每 一記憶電晶體之性質由選自於至少四種啓始電壓値其中之 一來定義,不同的電晶體具有一相異的啓始電壓値。記憶 體陣列包含有第一型記憶體,其包含具有第一功能之第一 材質組成的閘電極,而第一材質部份決定第一電晶體的啓 始電壓大小;以及第二型記憶體,其包含有第二功能之第 二材質組成的閘電極,而第二材質部份決定第二電晶體之 啓始電壓大小。此外,此陣列包含有一具第一閘極氧化層 厚度之第三記憶電晶體,及一具第二閘極氧化層厚度之第 四記憶電晶體。本發明之另一特點是提供一含有複數個第 一型電晶體之第一列(column)結構,每一第一型電晶體均包 含一第一源/汲極擴散區,及一第一閘極;以及一含有複數 個第二型電晶體之第二列結構,每一第二電晶體均包含有 一第二源/汲極擴散區,及一第二閘極。第一電晶體之第一 閘極與第二電晶體之第二閘極更包含一連續的導線,其中 第一閘極其中之一含有N型複晶矽且其相鄰的第二電晶體 之第二閘極含有P型複晶矽;第一電晶體中至少有一含有 第二厚度之閘極氧化層。 爲讓本發明之上述和其他目的、特徵、和優點能更明 I.----^-----^-------ΪΤ------m ·(請先閱讀背面之注意事項Η'填寫本頁) 本紙張尺度適用中國國家標準(CNS ) Λ4規格(210Χ 297公釐) S12042 94TWF.DOC/002 A7 B7 五、發明説明(冷) 顯易懂,下文特舉一較佳實施例,並配合所附圖式,作詳 細說明如下: 圖示之簡單說明: 第1圖是一根據本發明之罩幕式唯讀記憶電晶體之部 份剖面示意圖。 第2圖是一根據本發明之罩幕式唯讀記憶電晶體之另 一部份剖面示意圖。 第3〜12圖是根據本發明以製造罩幕式唯讀記憶體之部 份剖面示意圖。 實施例 經濟部中央標準局員工消費合作社印製 本發明之一特點是提供一罩幕式唯讀記憶電晶體,其 資料之編碼是藉選擇性程式化陣列中的每一電晶體之啓始 電壓,其大小爲三個或三個以上預定値之其中之一,更好 地是有四種不同的啓始電壓可供陣列中的每一電晶體選 擇,以使至少2位元的資料可被儲存在每一記憶單元中。 多位元資料可被儲存在不同的電晶體中,例如可藉不同性 質之材質選擇性形成具有不同啓始電壓之閘電極;另一技 術,例如改變不周電晶體閘極氧化層之厚度,也許可改變 啓始電壓之放置,而其他結合另類技術以調整啓始電壓之 技術亦可被應用。根據本發明之較佳實施例其特點是提供 電晶體,其閘電極之材料是選自於兩不同功能性質材質其 中之一,且其閘極氧化層厚度是選自於兩厚度其中之一; 結合兩種閘電極材料與兩種閘電極氧層厚I,使得記辑 體可選擇性提供四種不同的啓始電壓之一,用以在記憶體 9 本5張尺度適用中國國家標準(CNs i^格(210 X 297公釐) A7 B7 0294TWF.DOC/002 五、發明説明(Q ) 中儲存資料。例如第一、最高啓始電壓値可儲存二位元資 料,,〇〇,’,第二、較低啓始電壓値可儲存二位元資料”01” ’ 第三起始電壓可儲存二位元資料,,】〇,,,最低啓始電壓値則 可儲存二位元資料”ιι”。 根據本發明之較佳實施例,其藉選擇電晶體之啓始電 壓將記憶體之資料予以程式化;而選擇記憶體之啓始電壓 例如可以選擇用以形成閘電極之材料性質’或改變通道區 之雜質濃度,或改變閘極氧化層之厚度’或改變在閘極氧 化層內之固定電荷,或是其他所熟知的方法來完成;然而’ 這些方法很多在程式化記憶體並不可行。比較可行的是以 不同材料形成電晶體之閘電極來製作具有不同啓始電壓之 電晶體;不過,將會有更多的困難及花費用以形成此種記 憶體陣列。除非多位元唯讀記憶體如同現在所使用的單一 位元唯讀記億體般容易程式化,否則多位元唯讀記憶體將 不具有商業可實行性。 以下將描述根據本發明之較佳實施例,其提供1S[沏或P _型龙複_晶矽以形成電晶體之Μ有不词啓始電壓之, 這些記憶電晶體可被便宜地大量製造,其使用的方法插述 如下:本較佳實施例利用由兩種不同功能性質材料(如Ν 型、ρ型複晶矽)組成的電晶體之閘電極,連接其他程式化 的佈局(strategy),已定義出至少三種不同的啓始電壓。 某些根據本發明之實施例,P型複晶矽與N型複晶砂 均被用以當作閘電極材料,其具有不同的功能性質,故可 製作出具不同啓始電壓之電晶體;電晶體之閘電極可選擇p ---- - 10 本紙張尺度適用中國國^^( CNS ) A4規格--'s................. „ ^ L-- 1請先閱讀背面之注意事填寫本頁) 訂 線 經濟部中央標隼局員工消費合作社印製 五 煩 Wh 、年 .· G Λ —- -7J、· f 7 ! if· k \h 日 所 提 之 0294TWF1.DOC/002 第85〗11373號中文說明書修正頁 A7 B7 修正日期86.4.24 經濟部中央標準局員工消費合作社印製 發明説明(Y ) 型或N型複晶矽爲材料’而形成一具有高或低啓始電壓之 電晶體’簡易地完成程式化的工作。複晶矽閘電極之材料 例如可用第一型導電雜質作全面性地佈植,然後形成一罩 幕於預定程式化區域上,再用第二型導電雜質佈植到複晶 矽閘電極內’形成具有第二型導電層雜質部份摻雜之複晶 矽區域。 本較佳實施例利用兩種不同閘電極材料,第二型記憶 體之性質可能也只是再至少兩可選的啓始電壓値中改 變;例如電晶體可形成具有兩種不同厚度之閘極氧化層其 中之一:其中較厚的閘極氧化層厚度約〗較薄的閘極 氧化層其厚度約13 5 Α。改變介於閘電極與矽基底間的閘極 氧化層厚度’可產生兩種不同厚度之啓始電壓,例如啓始 電壓3.6V(P型矽基底,閘極氧化層厚度約300A),26v(b 型砂基底’閘極氧化層厚度約30〇A),1.95V(l型砂基底, 閘極氧化層厚度約135A) ’ 〇_95V(N型矽基底,閘極氧化 層厚度約135A),其已由場效電晶體佈局之組合而獲得。 利用選擇性摻雜複晶矽閘電極以編碼之唯讀記憶電晶 體,簡單描述於第1圖’其所顯示的是唯讀記憶體之中間 製程階段,且其爲一堆位元唯讀記憶體,僅藉選擇性慘雜 g己憶體之閘電極以完成程式化工作;第1圖中之唯讀記憶 體的單一位元實施例是藉由選擇性摻雜閘電極來程式化唯 讀記憶體。第1圖之唯讀記憶體的製造方法是:提供一 P型 砂基底10,並形成有埋入式N+型位兀線陣列12_於基底之 單元區內,及週邊電路區所需要的N型井區14及用以定義 Γ ¥ (請先閲讀背面之注$項再填寫本頁)Binding line 0294TWF.DOC / 002 A7 0294TWF.DOC / 002 A7 Printed by the Beigong Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs -----------------———— V. Description of the invention (Shi) The concentration of impurities in the road area, so the implantation of the four types of impurities can be made into the channel area with the four types of concentration of impurities in the body. The channel regions with different impurity concentrations can be made with transistors with different starting voltages; the starting voltage can be detected by applying different gate voltages to the memory, which can be detected when the transistor is turned on To. However, the implementation of Takizawa's read-only memory also has many difficulties. The most important thing is to use the accurate mask automatic alignment process to form Takizawa's read-only memory, and the accurate mask automatic alignment process needs to be implemented. The two-layer structure only reads the channel area of the memory, and the size of this structure is half of ^^, so the accuracy of the automatic alignment of the mask is stricter than the process of making the memory transistor. If the memory is manufactured based on this design principle, it would be infeasible to produce a channel oxide layer mask with precise alignment to the separation as mentioned in the Takizawa patent. Therefore, Takizawa read-only memory based on this principle does not make full use of the advantages of all small-sized components. In view of this, the present invention proposes a multi-bit read-only memory element that is easier to mass-produce. The element is non-volatile and each memory element includes multiple memory cells, and each memory cell includes a memory, which has A source / drain diffusion region, a channel region separating the source / drain diffusion region, and a gate oxide layer and gate electrode covering the channel region. The nature of each memory is defined by the starting voltage, which is selected from one of at least three predetermined values, and different transistors have different starting voltages. The memory array includes a first type of memory, which includes a gate electrode composed of a first material with a first function, and the first material portion determines the starting voltage of the first transistor; and the second '7' ϋ The Zhang scale is applicable to the Chinese National Standard (CNS) Λ4 specification (2 丨 0X297mm) IU --------- installed L-^ --- set ----- line. C please read the back first Note to fill in this page} 0294TWF.DOC / 002 A7 0294TWF.DOC / 002 A7 Printed by the Employees ’Consumer Cooperative of the Central Bureau of Standards of the Ministry of Economic Affairs / B7 V. Invention description (夂) type memory, which contains the second function The gate electrode composed of the second material, and the second material part determines the starting voltage of the second transistor. Another feature of the present invention is to provide a non-volatile memory element including a plurality of memory cells, each memory cell includes a memory transistor, the memory transistor includes a source / drain diffusion region, and a separate source / drain The channel area of the diffusion area includes a gate oxide layer and a gate electrode covering the channel area. The nature of each memory transistor is defined by one selected from at least four starting voltage values. Different transistors have a different starting voltage value. The memory array includes a first-type memory, which includes a gate electrode composed of a first material having a first function, and the first material portion determines the starting voltage of the first transistor; and the second-type memory, It includes a gate electrode composed of a second material with a second function, and the second material portion determines the starting voltage of the second transistor. In addition, the array includes a third memory transistor with a first gate oxide thickness and a fourth memory transistor with a second gate oxide thickness. Another feature of the present invention is to provide a first column structure including a plurality of first-type transistors, each of which includes a first source / drain diffusion region and a first gate Electrodes; and a second column structure containing a plurality of second-type transistors, each second transistor includes a second source / drain diffusion region, and a second gate. The first gate of the first transistor and the second gate of the second transistor further comprise a continuous wire, wherein one of the first gate contains N-type polycrystalline silicon and the adjacent second transistor The second gate contains P-type polycrystalline silicon; at least one of the first transistors contains a gate oxide layer with a second thickness. In order to make the above and other objects, features, and advantages of the present invention clearer I. ---- ^ ----- ^ ------- ΪΤ ------ m Note on the back Η 'fill in this page) This paper scale is applicable to the Chinese National Standard (CNS) Λ4 specifications (210Χ 297 mm) S12042 94TWF.DOC / 002 A7 B7 5. Description of the invention (cold) It is easy to understand, especially below A preferred embodiment, in conjunction with the attached drawings, will be described in detail as follows: Brief description of the drawings: FIG. 1 is a partial cross-sectional schematic diagram of a mask-type read-only memory transistor according to the present invention. Fig. 2 is another schematic sectional view of a mask-type read-only memory transistor according to the present invention. Figures 3 to 12 are schematic cross-sectional views of parts of a mask-based read-only memory manufactured according to the present invention. Embodiments Printed by the Employee Consumer Cooperative of the Central Bureau of Standards of the Ministry of Economy One of the features of the present invention is to provide a screen-type read-only memory transistor whose data is encoded by selectively programming the starting voltage of each transistor in the array , Whose size is one of three or more than three predetermined values, preferably four different starting voltages for each transistor in the array to select, so that at least 2 bits of data can be Stored in each memory unit. Multi-bit data can be stored in different transistors, for example, gate electrodes with different starting voltages can be selectively formed by materials with different properties; another technique, such as changing the thickness of the gate oxide layer of the imperfect transistor, It is possible to change the placement of the starting voltage, and other techniques that combine alternative technologies to adjust the starting voltage can also be applied. According to a preferred embodiment of the present invention, a transistor is provided, the gate electrode material is selected from one of two materials with different functional properties, and the thickness of the gate oxide layer is selected from one of the two thicknesses; Combining two kinds of gate electrode materials and two kinds of gate electrode oxygen layer thickness I, the recorder can selectively provide one of four different starting voltages, which are used to apply Chinese national standards (CNs i ^ Grid (210 X 297 mm) A7 B7 0294TWF.DOC / 002 5. Store the data in the description of invention (Q). For example, the first and highest starting voltage values can store two-bit data, 〇〇, ', Second, the lower starting voltage value can store two-bit data "01" 'The third starting voltage can store two-bit data,】 〇 ,, and the lowest starting voltage value can store two-bit data " ιι ". According to a preferred embodiment of the present invention, the data of the memory is programmed by selecting the starting voltage of the transistor; and selecting the starting voltage of the memory can, for example, select the material properties used to form the gate electrode. Or change the impurity concentration in the channel area Or change the thickness of the gate oxide layer or change the fixed charge in the gate oxide layer, or other well-known methods; however, many of these methods are not feasible in programmed memory. It is more feasible to Different materials form gate electrodes of transistors to produce transistors with different starting voltages; however, there will be more difficulties and expense to form such a memory array. Unless the multi-bit read-only memory is as it is now The single-bit read-only memory used is easy to program, otherwise the multi-bit read-only memory will not be commercially viable. The following will describe a preferred embodiment according to the present invention, which provides 1S [Bab or P _ 型 龙 复 _ crystalline silicon to form transistors has a starting voltage, these memory transistors can be mass-produced inexpensively, the method of use is as follows: The preferred embodiment uses two different The gate electrodes of transistors composed of functional materials (such as N-type and p-type polycrystalline silicon), connected to other programmed layouts, have defined at least three different starting voltages. Some roots In the embodiments of the present invention, both P-type polycrystalline silicon and N-type polycrystalline sand are used as gate electrode materials, which have different functional properties, so transistors with different starting voltages can be fabricated; gates of transistors The electrode can be selected p -----10 The paper size is suitable for China ^^ (CNS) A4 specification-'s ................. „^ L-- 1 Please read the precautions on the back first and fill out this page) The line of the Ministry of Economic Affairs, Central Standard Falcon Bureau Employee Consumer Cooperative printed five troubles Wh, N. G Λ —- -7J, f 7! If · k \ h No. 0294TWF1.DOC / 002 No. 85〗 11373 Chinese Manual Amendment Page A7 B7 Amendment Date 86.4.24 Description of Invention (Y) or N-type polycrystalline silicon printed by the Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs as a material Transistors with high or low starting voltages can easily complete the programmed work. For example, the material of the polycrystalline silicon gate electrode can be fully implanted with the first-type conductive impurities, and then a mask is formed on the predetermined programming area, and then the second-type conductive impurities are implanted into the polycrystalline silicon gate electrode. A polysilicon region partially doped with impurities of the second type conductive layer is formed. In this preferred embodiment, two different gate electrode materials are used, and the properties of the second type memory may only be changed by at least two optional starting voltage values; for example, transistors may form gate oxides with two different thicknesses One of the layers: the thicker gate oxide layer has a thickness of approximately 135 Å. Changing the thickness of the gate oxide layer between the gate electrode and the silicon substrate can generate two different starting voltages, for example, the starting voltage is 3.6V (P-type silicon substrate, the thickness of the gate oxide layer is about 300A), 26v ( b-type sand substrate 'gate oxide layer thickness of about 30〇A), 1.95V (l-type sand substrate, gate oxide layer thickness of about 135A)' 〇_95V (N-type silicon substrate, gate oxide layer thickness of about 135A), which Has been obtained from the combination of field effect transistor layout. The use of selectively doped polycrystalline silicon gate electrodes to encode read-only memory transistors is briefly described in Figure 1 'which shows the intermediate process stage of read-only memory, and it is a pile of bit read-only memory Body, only the gate electrode of the selective memory is used to complete the programming work; the single bit embodiment of the read-only memory in FIG. 1 is to program the read-only by selectively doping the gate electrode Memory. The manufacturing method of the read-only memory in FIG. 1 is: providing a P-type sand substrate 10, and forming an embedded N + type bit line array 12_ in the unit area of the substrate and the N type required by the peripheral circuit area Well area 14 and used to define Γ ¥ (Please read the note $ item on the back before filling this page)

,1T 11 本紙張尺度適用中國國家榡準(CNS 規格(210X297公釐) 經濟部中央標準局員工消費合作社印製 0294TWF.DOC/002 _Β7 五、發明説明(彳) 單元區及隔離週邊電路區之場氧化區15 ;閘極氧化層16 覆蓋於記憶體之通道區及埋入式Ν+型位元線陣列12上;Ν 型複晶矽線20形成於閘極氧化層16上,以選擇作爲部份記 憶體之閘電極,而Ρ型複晶矽線22形成用以作爲剩餘的記 憶體之閘電極。具有Ν型複晶矽電極20之電晶體的啓始電 壓約介於0.9〜1.0V,而具有Ρ型複晶矽電極22之電晶體的 啓始電壓約爲1.9V ;將近IV的啓始電壓差異,可以習知 的摻雜技術控制所形成的Ρ型或Ν型複晶矽內之雜質量而 達到。程式化差異之程度可藉由偵測記憶體之阻抗,或偵 測施加一參考電壓於記憶體之閘電極時,電晶體啓動之情 況。後述之方法應用於如圖1所示,設計尺寸小於0.5 之唯讀記憶體時是容易的,其操作電壓可調到更低,約 1.5V。P/Ν接合面24是形成於N型與P型複晶矽閘電極線 之界面,其使得閘電極不具導電性。此外,形成一導體層 26於N型複晶矽閘電極20及P型複晶矽閘電極22上,使 得N型複晶矽、P型複晶矽間產生歐姆式接觸(ohmic contact);此N型複晶矽、P型複晶矽及導電層所形成的層 狀結構,允許記憶體之閘電極在此類唯讀記憶體中作爲字 元線。週邊電路基本上包括NMOS電晶體30及PM0S電晶 體32兩者,其可以偶合用以形成反相電路及緩衝電路區。 如第1圖所示之唯讀記憶體,其可以此方法大量製造 及程式化,此唯讀記憶體成本便宜且有足夠的再現性,在 定義不同的啓始電壓之狀態時可控制良好。統一薄度之閘 極氧化層16式沈積於埋入式N+型位元線12上;完成後, ,:12 :; 本紙張尺度適用中國國家ϋ ( CNS ) A4规格(210X 297公釐) ~ --- 請先閲讀背面之注意事&孑填寫本頁) 裝· .*1' A7 0294TWF.DOC/002 B7 五、發明説明(川) 沈積一經統一摻雜一種導電雜質(例如N型雜質)之複晶矽 層,再形成一罩幕於複晶矽層上,並裸露出電位電晶體 (potential transistor)之通道區上的複晶砂部份;然後,以相 反的導電雜質(例如P型雜質)佈植或擴散進入裸露的複晶 矽區域。定義用的罩幕除去後,沈積一耐熱金屬或一耐熱 金屬矽化物層於複晶矽層上,然後,再形成一罩幕以定義 出字元線,較上層之導電層及複晶矽層被蝕刻,然後便依 習知的方法完成唯讀記憶體之製造。以此方法定義之唯讀 記憶體,其比需要的提供更大的容忍度,且其若完全採用 不同之材料以形成記億體之閘電極,其可有更不同的功 能;此外,自動對準的容忍度比以習知通道摻雜定義出的 記憶體還來的大。特別的是,本較佳實施例利用到選擇性 摻雜技述。 選擇性提供一種或另一種不同導電性質之複晶矽作爲 閘電極,便可簡易地提供兩種不同的啓始電壓;而多位元 唯讀記憶體其設計必須提供二個以上不同記憶電晶體之起 始電壓値,其是以交替程式化技術所定義出來的,可以許 多方法完成,例如可以選擇性離子佈植進入記憶電晶體之 通道區,並結合含有相異雜質特性之閘電極,可提供至少 四種不同的啓始電壓値;不過此步驟並非完全可靠,因爲 提供確實固定量之雜質以多重佈植是困難的。此外,結合 更多可大量製造之程式化技術以摻雜複晶矽閘極,提供不 同的啓始電壓以形成多位元記憶體是需要的。 一提供至少兩額外啓始電壓値之技術,其所選擇不同 _ . 13 ' ϋ张尺度適用中國國^準(CNS ) Λ4規格(210X 297公釐) - — i----------裝 I ί . .、請先閲讀背面之注意事^4填寫本頁) 訂 經濟部中央標準局員工消費合作社印裝, 1T 11 This paper standard is applicable to China National Standard (CNS specification (210X297mm) Printed by the Employee Consumer Cooperative of the Central Bureau of Standards of the Ministry of Economic Affairs 0294TWF.DOC / 002 _Β7 Fifth, the invention description (彳) of the unit area and the isolated peripheral circuit area Field oxide region 15; gate oxide layer 16 covers the channel region of the memory and the buried N + type bit line array 12; N type polycrystalline silicon line 20 is formed on the gate oxide layer 16 to select as Part of the gate electrode of the memory, and the P-type polycrystalline silicon wire 22 is formed as the gate electrode of the remaining memory. The starting voltage of the transistor with the N-type polycrystalline silicon electrode 20 is about 0.9 ~ 1.0V , And the starting voltage of the transistor with the P-type polycrystalline silicon electrode 22 is about 1.9V; the difference of the starting voltage is close to IV, you can control the formation of P-type or N-type polycrystalline silicon by conventional doping techniques The degree of miscellaneous quality can be achieved. The degree of programmatic difference can be detected by detecting the impedance of the memory, or detecting the situation where the transistor is activated when a reference voltage is applied to the gate electrode of the memory. As shown, when the design size is less than 0.5 read-only memory It is easy, and its operating voltage can be adjusted to a lower level, about 1.5 V. The P / N junction 24 is formed at the interface between the N-type and P-type polycrystalline silicon gate electrode lines, which makes the gate electrode non-conductive. In addition, Forming a conductor layer 26 on the N-type polycrystalline silicon gate electrode 20 and the P-type polycrystalline silicon gate electrode 22, so that an ohmic contact occurs between the N-type polycrystalline silicon and the P-type polycrystalline silicon; this N-type The layered structure formed by polycrystalline silicon, P-type polycrystalline silicon and conductive layer allows the gate electrode of the memory to be used as a word line in this type of read-only memory. The peripheral circuit basically includes NMOS transistor 30 and PMOS Both crystals 32 can be coupled to form an inverter circuit and a buffer circuit area. As shown in Figure 1, the read-only memory can be manufactured and programmed in large quantities by this method. This read-only memory is inexpensive and has Sufficient reproducibility, it can be controlled well when defining different starting voltage states. The gate oxide layer of uniform thinness is deposited on the buried N + type bit line 12 by 16; after completion,,: 12 :; The paper size is applicable to China National (CNS) A4 (210X 297mm) ~ --- please Read the notes on the back & fill in this page) Install ·. * 1 'A7 0294TWF.DOC / 002 B7 5. Description of the invention (Sichuan) Deposited polycrystalline silicon once uniformly doped with a conductive impurity (such as N-type impurities) Layer, and then form a mask on the polycrystalline silicon layer, and expose the part of the polycrystalline sand on the channel region of the potential transistor; then, implant with opposite conductive impurities (such as P-type impurities) Or diffuse into the exposed polycrystalline silicon area. After the defined mask is removed, deposit a heat-resistant metal or a heat-resistant metal silicide layer on the polycrystalline silicon layer, and then form a mask to define the character line, the upper conductive layer and the polycrystalline silicon layer It is etched, and then the manufacturing of read-only memory is completed according to the conventional method. The read-only memory defined in this way provides greater tolerance than necessary, and if it uses completely different materials to form the gate electrode of the memory, it can have more different functions; The quasi-tolerance is greater than the memory defined by conventional channel doping. In particular, this preferred embodiment utilizes selective doping techniques. Selectively provide one or another type of polycrystalline silicon with different conductive properties as the gate electrode, which can easily provide two different starting voltages; and the design of multi-bit read-only memory must provide more than two different memory transistors The initial voltage value, which is defined by alternating programming techniques, can be accomplished in many ways. For example, it can be selectively ion implanted into the channel region of the memory transistor and combined with a gate electrode containing different impurity characteristics. Provide at least four different starting voltage values; however, this step is not completely reliable, because it is difficult to provide a fixed amount of impurities for multiple implants. In addition, it is necessary to incorporate more programming techniques that can be mass-produced to dope polysilicon gates to provide different starting voltages to form a multi-bit memory. 1. Provide at least two additional starting voltage values, the choices are different _. 13 'ϋ Zhang scale is applicable to China National Standard (CNS) Λ4 specifications (210X 297 mm)-— i -------- -Install I ί .. Please read the notes on the back ^ 4 fill in this page) Order the printed version of the employee consumer cooperative of the Central Bureau of Standards of the Ministry of Economic Affairs

經濟部中央標隼局員工消費合作社印裝 摻雜的閘電極材料與不同厚度之閘極氧化層要一致;這些 年來已有不少的文獻發表如何製造具有良好定義及再現 性的閘極氧化層,其中包括厚度。改變基底與閘電極間的 間格即提供不同厚度之閘極氧化層,可改變記憶體之啓始 電壓。此外,本較佳實施例特別的是選擇性提供兩種不同 厚度之閘極氧化層以及兩種不同功能之閘電極材料。 第2圖所顯示的是如第1圖所示之唯讀記憶體,其提 供一 P型矽基底10,其上並形成有一埋入式N+型位元線陣 列及位於週邊電路區之N型井區14,及分隔單元區與週邊 電路區之場氧化區15 ;薄閘極氧化層40形成於選擇作爲 記憶電晶體之通道區上,而較厚的閘極氧化層42則形成於 其他記憶電晶體之通道區上。薄閘電極氧化層40之厚度例 如約小於150A,而較厚之閘極氧化層42其厚度約大於 250A。閘極氧化層厚度之差異主要是決定於可大量製造性 及不同記憶體間可被有效偵測之啓始電壓差異程度。 N型複晶矽電極44級P型複晶矽電極46是選擇性形成 於要將特殊資料儲存於唯讀記憶體內有關的記憶體之通 道區上:導線48則形成於N型及P型複晶矽電極上,以提 供一沿閘電極的連續導電途徑,作爲唯讀記憶體之字元 線。週邊電路區同時包括丽OS電晶體50及PMOS電晶體 52,以形成反相電路及緩衝電路區,如傳統唯讀記憶體之 設計。依此製造方法便可選擇性地形成一具有四種啓始電 壓其中之一的記憶電晶體;例如可提供啓始電壓値約爲 3.6V(P型矽基底,閘極氧化層厚度約爲300A),2.6ν(ϋ型 _14_ (請先閲讀背面之注意事項再填寫本頁) 訂 本紙張尺度適用中國國家標準(CNS ) Α4規格(210Χ297公釐) A7 B7 0294TWF1.DOC/002 312042 五、發明説明(丨1) 矽基底,閘極氧化層厚度約爲300A),1.95V(已型矽基底, 閘極氧化層厚度約爲135A),〇.95V({ijy矽基底,閘極氧 化層厚度約爲135A)。最高的啓始電壓(3.6V)可儲存二位 元資料”00”,次高的啓始電壓(2.6V)可儲存二位元資 料”01”,更低的啓始電壓(1.95V)可儲存二位元資料”10”, 最低的啓始電壓(0.95V)可儲存二位元資料”11”。這些不同 的啓始電壓値可由記憶體之感應放大器(sense amplifiers) 讀出,且資料之區分圖解如由Stark獲得的美國專利第 4,287,570號及由Takizawa等人獲得的美國專利第 5,386,381號所描述,這些專利合適的部份均合倂於參考 資料中。 如第2圖所示之唯讀記憶體之製造方法,於第3〜12圖 描述之。首先,請先參照第3圖,提供一P型矽基底60 , 其上並形成有PMOS及NMOS之N型井區70及場氧化層72 , 平行排列之埋入式N+型區74形成於基底以作爲位元線及 場效電晶體之源/汲極擴散區,然後再形成一相當厚的氧 化層76於基底上,再形成罩幕層78於氧化層76上,並覆 蓋於預定具有較厚的閘極氧化層之電晶體的通道區上,使 其將具有四個資料儲存系統中之兩較高的啓始電壓其中 之一;然後,蝕刻氧化層76,留下選擇的通道區上的氧化 層77,如第4圖所示。 其次,請參照第5圖及第6圖’除去罩幕78後,以熱 氧化法形成一較薄之閘極氧化層80於剩餘之基底表面,其 也形成於週邊電路區之電晶體通道區上的基底表面:此時 基底之記憶單元區具有較厚之閘極氧化層77及較薄之閘極 (咚先聞讀背面之注意事項再填寫本頁)The printed and doped gate electrode materials of the Central Standard Falcon Bureau Employee Consumer Cooperative must be the same as the gate oxide layers of different thicknesses; in recent years, many documents have published how to manufacture gate oxide layers with good definition and reproducibility , Including thickness. Changing the interval between the substrate and the gate electrode provides gate oxide layers of different thicknesses, which can change the starting voltage of the memory. In addition, this preferred embodiment is particularly selective in providing two gate oxide layers with different thicknesses and two gate electrode materials with different functions. Figure 2 shows the read-only memory as shown in Figure 1, which provides a P-type silicon substrate 10 on which a buried N + type bit line array and N-type located in the peripheral circuit area are formed The well area 14, and the field oxide area 15 separating the cell area and the peripheral circuit area; a thin gate oxide layer 40 is formed on the channel area selected as the memory transistor, and a thick gate oxide layer 42 is formed on other memories On the channel area of the transistor. The thickness of the thin gate oxide layer 40 is, for example, less than about 150A, and the thickness of the thicker gate oxide layer 42 is more than about 250A. The difference in the thickness of the gate oxide layer is mainly determined by the manufacturability and the difference in the starting voltage that can be effectively detected between different memories. N-type polycrystalline silicon electrode 44-level P-type polycrystalline silicon electrode 46 is selectively formed on the channel area of the memory to store special data in the read-only memory: the wire 48 is formed on the N-type and P-type complex On the crystalline silicon electrode, a continuous conductive path along the gate electrode is provided as the word line of the read-only memory. The peripheral circuit area includes both the OS transistor 50 and the PMOS transistor 52 to form an inverter circuit and a buffer circuit area, such as the design of a conventional read-only memory. According to this manufacturing method, a memory transistor with one of four starting voltages can be selectively formed; for example, a starting voltage value of about 3.6V can be provided (P-type silicon substrate, the thickness of the gate oxide layer is about 300A ), 2.6ν (ϋ type_14_ (please read the precautions on the back before filling in this page). The size of the paper is applicable to the Chinese National Standard (CNS) Α4 specification (210Χ297 mm) A7 B7 0294TWF1.DOC / 002 312042 V. DESCRIPTION OF THE INVENTION (丨 1) Silicon substrate, gate oxide layer thickness is about 300A), 1.95V (formed silicon substrate, gate oxide layer thickness is about 135A), 0.95V ({ijy silicon substrate, gate oxide layer) Thickness is about 135A). The highest starting voltage (3.6V) can store binary data "00", the next highest starting voltage (2.6V) can store binary data "01", and the lower starting voltage (1.95V) can Store binary data "10", the lowest starting voltage (0.95V) can store binary data "11". These different starting voltage values can be read by the sense amplifiers of the memory, and the distinguishing diagrams of the data are as described in US Patent No. 4,287,570 obtained by Stark and US Patent No. 5,386,381 obtained by Takizawa et al., The appropriate parts of these patents are incorporated in the references. The manufacturing method of the read-only memory as shown in Figure 2 is described in Figures 3-12. First, please refer to FIG. 3 to provide a P-type silicon substrate 60 on which an N-type well region 70 and a field oxide layer 72 of PMOS and NMOS are formed, and a buried N + region 74 arranged in parallel is formed on the substrate To serve as the source / drain diffusion area of the bit line and field effect transistor, and then form a relatively thick oxide layer 76 on the substrate, and then form a mask layer 78 on the oxide layer 76, and cover the predetermined Thick gate oxide layer on the channel area of the transistor, so that it will have one of two higher starting voltages in the four data storage systems; then, the oxide layer 76 is etched, leaving the selected channel area The oxide layer 77 is shown in Figure 4. Next, please refer to Figure 5 and Figure 6 'After removing the mask 78, a thin gate oxide layer 80 is formed on the remaining substrate surface by thermal oxidation, which is also formed in the transistor channel area of the peripheral circuit area The upper surface of the substrate: At this time, the memory cell area of the substrate has a thick gate oxide layer 77 and a thin gate electrode (Don first read the precautions on the back side and then fill out this page)

-*1T 線< 經濟部中央標準局貝工消费合作社印裝 本紙張尺度適用中國國家揉準(CNS ) Α4規格(21〇'〆297公釐) A7 B7 0294TWF.DOC/002 五、發明説明() 氧化層80。然後,沈積一厚度約介於1000〜250〇A之複晶矽 層82於厚、薄閘極氧化層上,其較佳厚度約是1500A,其 可以習知的低壓化學氣相沈積法於溫度約620 °C時形成。 N型及P型複晶矽閘電極之定義可以單一罩幕製程完 成,可先以第一型雜質摻雜整個複晶矽層,例如利用擴散 或是全面性離子佈植,形成一第一導電層,然後再形成— 罩幕,裸露出預定形成相反雜質摻雜過的複晶矽閘極區_ 面,再以相反導電性之雜質佈植進入裸露之複晶矽表面, 形成一具有相反導電性之雜質之複晶矽電極。如第6 _[戶斤 示,複晶矽層82先以氟化硼作全面性佈植,佈植劑量約爲 2X1015atoms/cm2且佈植能量約爲50KeV,形成一p裂複曰 矽,此摻雜量儘量適合用以製造具有薄及厚閘極氧化騰_ 記憶體的啓始電壓’其可允許儲存相關的資料被可靠之 測。換具話說’佈植之雜質濃度必須足以製作出一電晶偵 其啓始電壓足夠相異於具有其他啓始電壓値之電晶赠 得程式化後之資料可被可靠地寫入及讀出。 ^ 請參照第7圖及第8圖,晶全面性佈植後,沈積〜、/ 經濟部中央標準局員工消費合作社印裝 % 層84於複晶矽層82上,然後以微影成像及蝕刻技術舷光电 罩幕86於預備形成的N型複晶矽閘極區域上,暴露战\ 形成具有較低啓始電壓之場效電晶體之閘電極,且致靖條 與P型複晶矽閘電極相同厚度之閘極氧化層;而周 經常採用N型複晶砂閘極,故罩幕%經常暴露預定%略 電路形成場效電晶體之閘極氧化層的複晶矽層82。浐發壤 再將氯化氧磷(POCh)擴散或將砷、磷以 16 本紙張尺度適用中國國家標準(CNS ) Λ4規格(21〇X:297公釐) 0294TWF.DOC/002 A7 B7 經濟部中央標準局員工消費合作社印聚 五、發明説明(丨T) 4Xl〇15atoms/cm2且能量爲60KeV佈植進入未被罩幕86覆 蓋之區域,形成N型複晶矽區域之雜質濃度必須夠濃’使 得所製作出之記憶電晶體的啓始電壓可以與其他的記憶電 晶體的啓始電壓區分,形成一讀及寫容易之多位元唯讀記 憶體。若氯化氧鱗之擴散是用以局部性地摻雜複晶矽層’ 則在擴散過程中必須形成一氧化罩幕於光阻層上’若使用 砷或磷則不需此氧化罩幕。 接著,請參照第9圖,完成N型及P型雜質之佈植後, 除去罩幕86,複晶矽層包括有局部摻雜之N型複晶矽層90 及局部摻雜之P型複晶矽層92,P/N接面91則是位於相鄰 之P型及N型複晶矽層之介面;當閘電極形成於複晶矽層 時,P/N接面間幾乎是無法導通的。 然後,請參照第10圖,形成一導電層94於複晶矽層上, 其將部份形成閘電極(字元線)及連接訊號越過唯讀記憶體 到陣列中之場效電晶體閘極;且導電層94最好具有高導電 性,使N型複晶矽及P型複晶矽間產生歐姆接觸’矽化鎢 是一種較佳之導電層材料,而其他耐熱金屬及耐熱金屬矽 化物亦可提供此特性之材料。 接著,請參照第11圖,導電層94及複晶矽層90及92 經蝕刻後,形成字元線及閘電極,同時在週邊電路區形成 NMOS元件之閘極96及PMOS元件之閘極98。 最後,請參照第12圖,週邊電路區之NMOS元件及 PMos元件之閘極形成後,利用習知的淡摻雜技術(丨ightly doping drain)分別形成NMOS元件及PMOS元件之源/汲極 17 . I II 裝 I 訂 !—線 為 - (請先閱讀背面之注意事項名填寫本頁) 本紙張尺度適用中國國家標準(CNS ) Λ4規格(210X29*7公釐) A7 0294TWF.DOC/002 B7 五、發明説明(β) 擴散區。最後,再沈積一硼磷矽玻璃lOO(BPSG),封裝唯讀 記憶體,然後再以習知的製程形成記憶體之接觸窗口,完 成記憶體之製造,其詳細製程非本發明之重點,在此不再 贅述。 雖然本發明已以一較佳實施例揭露如上,然其並非用 以限定本發明,任何熟習此技藝者,在不脫離本發明之精 神和範圍內,當可作些許之更動與潤飾,因此本發明之保 護範圍當視後附之申請專利範圍所界定者爲準。 請先閱讀背面之注意事填寫本頁) -裝. 訂 線 經濟部中央標準局員工消費合作社印製 本紙張尺度適用中國國家標準(CNS ) Λ4規格(210X 297公釐)-* 1T line < The paper standard printed by Beigong Consumer Cooperative of Central Bureau of Standards of Ministry of Economic Affairs is applicable to China National Standard (CNS) A4 specification (21〇'〆297mm) A7 B7 0294TWF.DOC / 002 V. Description of invention () Oxide layer 80. Then, deposit a polycrystalline silicon layer 82 with a thickness of about 1000 ~ 250 OA on the thick and thin gate oxide layer, the preferred thickness of which is about 1500A, which can be known at low temperature by chemical vapor deposition Formed at about 620 ° C. The definition of N-type and P-type polycrystalline silicon gate electrodes can be completed in a single mask process. The entire polycrystalline silicon layer can be doped with first-type impurities, for example, by diffusion or comprehensive ion implantation to form a first conductivity Layer, and then form a mask to expose the surface of the polysilicon gate region doped with the opposite impurity, and then implant the impurity with opposite conductivity into the surface of the exposed polysilicon to form a surface with opposite conductivity Polycrystalline silicon electrode with sexual impurities. As shown in Section 6_ [Fu Jin, the polycrystalline silicon layer 82 is first implanted with boron fluoride, the implantation dose is about 2X1015atoms / cm2 and the implantation energy is about 50KeV, forming a p-fracture complex silicon, this The amount of doping is as suitable as possible to produce thin and thick gate oxides_the starting voltage of the memory ', which allows storage-related data to be reliably measured. To put it another way, the implanted impurity concentration must be sufficient to produce a transistor whose starting voltage is sufficiently different from that of other transistors with other starting voltage values. The programmed data can be reliably written and read out . ^ Please refer to Figure 7 and Figure 8, after the crystals are fully implanted, deposit ~, / The Ministry of Economic Affairs Central Standards Bureau employee consumer cooperative prints% layer 84 on the polycrystalline silicon layer 82, and then uses lithography imaging and etching The technology side photomask 86 is exposed on the N-type polycrystalline silicon gate area to be formed, forming the gate electrode of the field effect transistor with a lower starting voltage, and causing the strip and the P-type polycrystalline silicon gate The gate oxide layer with the same thickness of the electrode; and the N-type polycrystalline sand gate is often used, so the mask is often exposed to a predetermined percentage to form the polycrystalline silicon layer 82 of the gate oxide layer of the field effect transistor. In the soil, diffuse oxychloride phosphorus (POCh) or apply arsenic and phosphorus on 16 paper scales to the Chinese National Standard (CNS) Λ4 specification (21〇X: 297 mm) 0294TWF.DOC / 002 A7 B7 Ministry of Economic Affairs Printed by the Central Bureau of Standards 'Staff Consumer Cooperative V. Invention description (丨 T) 4X1015atoms / cm2 with an energy of 60KeV implanted into the area not covered by the mask 86, the impurity concentration forming the N-type polycrystalline silicon area must be sufficiently concentrated' The starting voltage of the manufactured memory transistor can be distinguished from the starting voltage of other memory transistors to form a multi-bit read-only memory that is easy to read and write. If the diffusion of oxygen chloride scales is used to locally dope the polycrystalline silicon layer, then an oxidation mask must be formed on the photoresist layer during the diffusion process. If arsenic or phosphorous is used, this oxidation mask is not required. Next, referring to FIG. 9, after the implantation of N-type and P-type impurities is completed, the mask 86 is removed, and the polycrystalline silicon layer includes a locally-doped N-type polycrystalline silicon layer 90 and a locally-doped P-type complex The crystalline silicon layer 92 and the P / N junction 91 are the interfaces between the adjacent P-type and N-type polycrystalline silicon layers; when the gate electrode is formed on the polycrystalline silicon layer, the P / N junction is almost impossible to conduct of. Then, referring to FIG. 10, a conductive layer 94 is formed on the polycrystalline silicon layer, which partially forms a gate electrode (character line) and connects the signal over the read-only memory to the field effect transistor gate in the array ; And the conductive layer 94 preferably has high conductivity, so that ohmic contact between N-type polycrystalline silicon and P-type polycrystalline silicon 'tungsten silicide is a preferred material for the conductive layer, and other heat-resistant metals and heat-resistant metal silicide can also Provide materials with this characteristic. Next, referring to FIG. 11, after the conductive layer 94 and the polycrystalline silicon layers 90 and 92 are etched, word lines and gate electrodes are formed, and at the same time, gates 96 of NMOS elements and gates 98 of PMOS elements are formed in the peripheral circuit area . Finally, please refer to FIG. 12, after the gates of the NMOS device and the PMos device in the peripheral circuit area are formed, the source / drain of the NMOS device and the PMOS device are formed using the conventional lightly doping drain technology . I II Pack I order! —The line is- (please read the note name on the back to fill in this page) This paper size is applicable to Chinese National Standard (CNS) Λ4 specification (210X29 * 7mm) A7 0294TWF.DOC / 002 B7 5. Description of the invention (β) Diffusion zone. Finally, a borophosphosilicate glass 100 (BPSG) is deposited to encapsulate the read-only memory, and then the contact window of the memory is formed by a conventional process to complete the manufacture of the memory. The detailed process is not the focus of the present invention. This will not be repeated here. Although the present invention has been disclosed as above in a preferred embodiment, it is not intended to limit the present invention. Anyone who is familiar with this skill can make some changes and modifications within the spirit and scope of the present invention. The scope of protection of an invention shall be deemed as defined by the scope of the attached patent application. Please read the precautions on the back first and fill out this page)-Packing. Threading Printed by the Employee Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs. This paper scale is applicable to the Chinese National Standard (CNS) Λ4 specification (210X 297 mm)

Claims (1)

ABCD 312042 0294TWF.DOC/002 六、申請專利範圍 件,其中該導電層之材料是耐熱金屬或耐熱金屬矽化物。 6. 如申請專利範圍第1項所述之非揮發性記憶體元 件,其中該些第一型記憶體中有些是具有第一厚度之閘極 氧化層,其餘的是具有第二厚度之閘極氧化層。 7. 如申請專利範圍第1項所述之非揮發性記憶體元 件,其中該第一厚度約小於150A,而該第二厚度約大於 300A。 8. —非揮發性記憶體元件,包括有: 複數個記憶單元,該記憶體單元包括有一記憶電晶 體,該記憶電晶體包括一源/汲極擴散區,一分隔源/汲極擴 散區之通道區,以及一閘極氧化層、一閘電極依序位於該 通道區上, 其中每一記憶電晶體之特性由選自於至少四個啓始電 壓預定値其中之一來決定,相異之記憶電晶體具有相異之 起始電晶體, 其中記憶電晶體陣列中包含有第一型記憶電晶體,其 閘電極是由具有第一功能之第一材料所形成的,該第一材 料部份決定該第一型電晶體之啓始電壓;及第二型記憶電 晶體,其閘電極是由具有第二功能之第二材料所形成的, 該第二材料部份決定該第二型記憶電晶體之啓始電壓,以 及 其中該些記憶電晶體陣列中,包含具有第一厚度閘極 氧化層之第三電晶體及具有第二厚度之閘極氧化層之第四 電晶體。 20 本紙張尺度適用中國國家標準(CNS ) Λ4規格(210X 297公釐) :----------1---^----1T-----叫—,ii -(請先閱讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A8 0294TWF.DOC/002 B8 C8 D8 六、申請專利範圍 9. 如申請專利範圍第8項所述的非揮發性記憶體元件 元件,其中該第一記憶體具有第二厚度之閘極氧化層,該 第二記憶體具有第一厚度之閘極氧化層。 10. 如申請專利範圍第8項所述的非揮發性記憶體元 件元件,其中該第一材料是複晶矽。 11. 如申請專利範圍第10項所述的非揮發性記憶體元 件元件,其中該第一材料是N型複晶矽且其中該第二材料 是P型複晶矽。 12. —記憶體元件,包括有: 一第一列,包含有複數個第一電晶體,該第一電晶體 包含一第一源/汲極擴散區,及一第一閘極;以及 一第二列,包含有複數個第二電晶體,該第二電晶體 包含一第二源/汲極擴散區,及一第二閘極, 其中該第一電晶體之第一閘極及該第二電晶體之該第 二閘極上含有一連續導線,且至少一第一閘極含有N型複 晶矽而其相鄰之該第二閘極含有P型複晶矽,且至少一具 有第一厚度之第一電晶體及一具有第二厚度之第二電晶 J--^------、訂-----^ I ^ '(請先閲讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 21 本紙張尺度適用中國國家標準(CNS ) Λ4規格(210X297公釐) 經濟部中央標準局員工消費合作社印製 六、申請專利範圍 1. 一非揮發性記憶體元件,包括有: 複數個記憶單元,該記憶體單元包括有一記憶電晶 體,該記憶電晶體包括一源/汲極擴散區,一分隔源/汲極擴 散區之通道區,以及一閘極氧化層、一閘電極依序位於該 通道區上, 其中每一記憶電晶體之特性由選自於至少四個啓始電 壓預定値其中之一來決定,相異之記憶電晶體具有相異之 起始電晶體, 其中記憶電晶體陣列中包含有第一型記憶電晶體,其 閘電極是由具有第一功能之第一材料所形成的,該第一材 料部份決定該第一型電晶體之啓始電壓;及第二型記憶電 晶體,其閘電極是由具有第二功能之第二材料所形成的, 該第二材料部份決定該第二型記憶電晶體之啓始電壓。 2. 如申請專利範圍第1項所述之非揮發性記憶體元 件,其中該第一材料及該第二材質是複晶矽。 3. 如申請專利範圍第1項所述之非揮發性記憶體元 件,其中該第一材質是N型複晶矽且該第二材質是P型複 晶砂。 4. 如申請專利範圍第3項所述之非揮發性記憶體元 件,其中至少有一第一型記憶記憶電晶體形成與一第二型 記憶電晶體相鄰,且其中有一導電層位於組成該第一型記 憶電晶體閘極之該第二型材料上,形成一連接的導線結 構。 5. 如申請專利範圍第4項所述之非揮發性記憶體元 19 裝 訂 一 線 (請先閱讀背面之注意事項、再填寫本頁) 本紙張尺度適用中國國家標準(CNS ) Λ4規格(210X 297公釐)ABCD 312042 0294TWF.DOC / 002 VI. Patent application. The material of the conductive layer is heat-resistant metal or heat-resistant metal silicide. 6. The non-volatile memory device as described in item 1 of the patent application scope, wherein some of the first-type memories are gate oxide layers with a first thickness, and the rest are gate electrodes with a second thickness Oxide layer. 7. The non-volatile memory element as described in item 1 of the patent application, wherein the first thickness is less than 150A and the second thickness is more than 300A. 8.-Non-volatile memory device, including: a plurality of memory cells, the memory cell includes a memory transistor, the memory transistor includes a source / drain diffusion area, a separate source / drain diffusion area The channel region, and a gate oxide layer and a gate electrode are sequentially located on the channel region, wherein the characteristics of each memory transistor are determined by one of at least four predetermined threshold voltages, different The memory transistor has different starting transistors, wherein the memory transistor array includes a first type memory transistor, the gate electrode of which is formed by a first material having a first function, the first material portion Determines the starting voltage of the first type transistor; and the second type memory transistor, the gate electrode of which is formed by a second material having a second function, and the second material portion determines the second type memory transistor The starting voltage of the crystal, and among these memory transistor arrays, includes a third transistor having a first thickness gate oxide layer and a fourth transistor having a second thickness gate oxide layer. 20 The size of this paper is in accordance with the Chinese National Standard (CNS) Λ4 specification (210X 297mm): ---------- 1 --- ^ ---- 1T ----- 叫 ——, ii- (Please read the precautions on the back before filling in this page) A8 0294TWF.DOC / 002 B8 C8 D8 printed by the Employee Consumer Cooperative of the Central Bureau of Standards of the Ministry of Economic Affairs 6. Scope of patent application 9. As described in item 8 of the scope of patent application A volatile memory device element, wherein the first memory has a gate oxide layer with a second thickness, and the second memory has a gate oxide layer with a first thickness. 10. The non-volatile memory device as described in item 8 of the patent application, wherein the first material is polycrystalline silicon. 11. The non-volatile memory device as described in item 10 of the patent application range, wherein the first material is N-type polycrystalline silicon and wherein the second material is P-type polycrystalline silicon. 12. —Memory device, including: a first row, including a plurality of first transistors, the first transistors including a first source / drain diffusion region, and a first gate; and a first Two columns, including a plurality of second transistors, the second transistor includes a second source / drain diffusion region, and a second gate, wherein the first gate of the first transistor and the second transistor The second gate of the transistor contains a continuous wire, and at least one first gate contains N-type polycrystalline silicon and its adjacent second gate contains P-type polycrystalline silicon, and at least one has a first thickness The first transistor and a second transistor with a second thickness J-^ ------, order ----- ^ I ^ '(Please read the precautions on the back before filling this page) Printed by the Staff Consumer Cooperative of the Central Bureau of Standards of the Ministry of Economic Affairs 21 This paper standard is applicable to the Chinese National Standard (CNS) Λ4 specifications (210X297 mm). Body elements, including: a plurality of memory units, the memory unit includes a memory Transistor, the memory transistor includes a source / drain diffusion region, a channel region separating the source / drain diffusion region, and a gate oxide layer and a gate electrode are sequentially located on the channel region, wherein each memory The characteristics of the transistor are determined by one of at least four predetermined threshold voltages. The different memory transistors have different starting transistors. The memory transistor array includes a first type memory transistor. The gate electrode of the crystal is formed by a first material having a first function, the first material portion determines the starting voltage of the first type transistor; and the second type memory transistor, the gate electrode is formed by Formed by a second material with a second function, the second material portion determines the starting voltage of the second memory transistor. 2. The non-volatile memory element as described in item 1 of the patent scope, wherein the first material and the second material are polycrystalline silicon. 3. The non-volatile memory element as described in item 1 of the patent scope, wherein the first material is N-type polycrystalline silicon and the second material is P-type polycrystalline sand. 4. The non-volatile memory device according to item 3 of the patent application scope, wherein at least one first-type memory transistor is formed adjacent to a second-type memory transistor, and one of the conductive layers is located A connecting wire structure is formed on the second-type material of the gate electrode of the one-type memory transistor. 5. The non-volatile memory element 19 as described in item 4 of the patent application is bound to the front line (please read the precautions on the back and then fill out this page). The paper size is applicable to the Chinese National Standard (CNS) Λ4 specification (210X 297 Mm)
TW85111373A 1996-09-17 1996-09-17 Multi-level high density mask read only memory TW312042B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW85111373A TW312042B (en) 1996-09-17 1996-09-17 Multi-level high density mask read only memory

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW85111373A TW312042B (en) 1996-09-17 1996-09-17 Multi-level high density mask read only memory

Publications (1)

Publication Number Publication Date
TW312042B true TW312042B (en) 1997-08-01

Family

ID=51566480

Family Applications (1)

Application Number Title Priority Date Filing Date
TW85111373A TW312042B (en) 1996-09-17 1996-09-17 Multi-level high density mask read only memory

Country Status (1)

Country Link
TW (1) TW312042B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI747783B (en) * 2021-04-28 2021-11-21 國立陽明交通大學 Operation method of a multi-bits read only memory
US12080363B2 (en) 2021-04-23 2024-09-03 Chen-Feng Chang Operation method of multi-bits read only memory

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US12080363B2 (en) 2021-04-23 2024-09-03 Chen-Feng Chang Operation method of multi-bits read only memory
TWI747783B (en) * 2021-04-28 2021-11-21 國立陽明交通大學 Operation method of a multi-bits read only memory

Similar Documents

Publication Publication Date Title
TW503528B (en) Semiconductor device
CN1720588B (en) Flash memory cell arrays having dual control gates per memory cell charge storage element
TW475267B (en) Semiconductor memory
TW444383B (en) NAND-Type nonvolatile memory device, manufacturing method thereof and driving method thereof
US6191975B1 (en) Non-volatile NAND type semiconductor memory device with stacked gate memory cells and a stacked gate select transistor
TW454353B (en) Semiconductor integrated circuit device, the manufacturing method thereof and the operation method
TW512495B (en) Nonvolatile semiconductor memory device, process of manufacturing the same and method of operating the same
JP3695539B2 (en) Ultra-high density alternating metal virtual ground ROM, reading method thereof and manufacturing method thereof
US7391071B2 (en) Nonvolatile memory devices with trenched side-wall transistors and method of fabricating the same
JP3906177B2 (en) Nonvolatile semiconductor memory device
JPH07120719B2 (en) Semiconductor memory device
TW201362B (en)
TWI261338B (en) Memory wordline hard mask extension
US8999785B2 (en) Flash-to-ROM conversion
TW478166B (en) A semiconductor device
JP4557950B2 (en) Nonvolatile semiconductor memory device
JP2008166759A (en) Nonvolatile memory element, and manufacturing method of the same
JP2006511940A (en) Multi-level memory cell with lateral floating spacer
JP2825407B2 (en) Nonvolatile semiconductor memory device
TW312042B (en) Multi-level high density mask read only memory
TW313706B (en) Read only memory structure and manufacturing method thereof
US4460979A (en) Memory cell
EP0889520A1 (en) Method of fabrication a non-volatile semiconductor memory device with shielded single polysilicon gate memory part
TW400626B (en) The read-only memory(ROM) structure and the manufactureing method thereof
TW476981B (en) Semiconductor memory and manufacturing method of the same