TW457712B - Flash memory cell structure for improving the data preservation problem and the programming speed and the formation method thereof - Google Patents
Flash memory cell structure for improving the data preservation problem and the programming speed and the formation method thereof Download PDFInfo
- Publication number
- TW457712B TW457712B TW89120822A TW89120822A TW457712B TW 457712 B TW457712 B TW 457712B TW 89120822 A TW89120822 A TW 89120822A TW 89120822 A TW89120822 A TW 89120822A TW 457712 B TW457712 B TW 457712B
- Authority
- TW
- Taiwan
- Prior art keywords
- layer
- gate
- oxide layer
- oxide
- silicon nitride
- Prior art date
Links
Landscapes
- Non-Volatile Memory (AREA)
- Semiconductor Memories (AREA)
Abstract
Description
457 712 五'發明說明(1) 發明領域: 一種改 程所造 構及其 本發明係有關於非揮發性記憶元件’特別是指 善分閘快閃記憶胞資料流失問題,同時改善氧化過 成之微笑效應而不利於程式化速度之快閃記憶胞結 形成方法。 發明背景: 内建記 當相機 至電腦 當數位相機最初進入市場時,它們被製作成有 憶體的系統,讓您可以每次拍攝大約三十張照片。 的記憶體滿載時,使用者便必須先將照片影像下載 中,才能拍攝更多影像。 今天一種低耗電、高存取速度,及防震、耐移 高穩定性等安全資料存取條件方面的全新的儲存系 快閃記憶卡的使用已使上述問題迎刃而解。快閃記 需非常少的電源。資訊可以較有效率的記憶區段.( )方式來記錄(或消除),而不像按位元組依序紀 緩慢。此外,資料一旦存到磁片之後’就不再需要 源用來保留資料。一般而言,以目前之技術即使電 關掉後仍可保留儲存的資料至少十年以上。這種優 得其他可攜式儲存系統黯然失色,因此和其它的儲 相比較, 快閃 記憶體具有 十足 的 競爭力 。十足明日 動性、 統,即 憶體僅 blocks 錄那麼 任何電 源是在 勢已使 存媒體 之星457 712 Five 'Description of the Invention (1) Field of the Invention: A modified structure and the present invention are related to the non-volatile memory element, especially the problem of data loss of flash memory cells, and the improvement of oxidation. The smile effect is not conducive to stylized speed flash memory cell formation methods. Background of the Invention: Built-in Cameras When Cameras to Computers When digital cameras first entered the market, they were made into memory systems, allowing you to take about thirty photos at a time. When the memory is full, the user must download the photo images before taking more images. Today, the use of a new type of storage system with low power consumption, high access speeds, and shock-resistant, mobile-resistant, high-stability, and other secure data access conditions has solved the above problems. Flash memory requires very little power. Information can be recorded (or erased) in a more efficient memory segment. () Method, instead of being slower in chronological order than bytes. In addition, once the data has been stored on the disk, the source is no longer needed to hold the data. Generally speaking, current technology can retain stored data for at least ten years even after the power is turned off. This is overshadowing other portable storage systems, so compared to other storage, flash memory is extremely competitive. Tomorrow is dynamic and systematic, that is, the memory is only recorded in blocks, so any power source is in the media star
&始接著去除光阻圖案後,以經定義之氮化矽層為罩幕, 丹施以高溫的熱氧化製程使第一複晶矽層1〇的上表面長一& After removing the photoresist pattern, using a defined silicon nitride layer as a mask, Dan Shi made a high temperature thermal oxidation process to make the upper surface of the first polycrystalline silicon layer 10 grow by one.
第6頁 4¾ 7 712 五、發明說明(3) 氧化區2 5,氣化區2 5的中央厚約1 2 0 nm ’值兩邊則很薄, 這樣的結果將使得第一複晶矽層1 〇的上表面呈現出兩終端 向上尖的結構。接著’再以熱磷酸鹽溶液去除其餘之氮化 矽層罩幕,然後以電漿蝕刻劑去除未被氧化區2 5覆蓋之第 一複晶石夕層1 〇後’此夾於閘極氧化層8及氧化區2 5之第一 複晶矽層1 0係做為懸浮閘極》 請參考圖二’再以熱氧化的方法形成另一氧化層3〇。 為減少不利於程式化速度的微笑效應現象,其形成步驟包 含:先以熱氧化法再長40埃的氧化層,接著,在較高溫的 f5 環境下再沉積HTO氧化層約1 30埃’最後,再以熱氧化法成 長7 0埃的氧化層’氧化層3 0以較閘極氧化層8厚一些的目 的係用以降低懸浮閘極對控制閘極的輕合比。然而上述的-氧化步驟已使得懸浮閘極底部邊緣產生某種程度的微笑效 應現象’如圖二標號3 5所示。 ' 接著’請參考圖三,再形成第二複晶矽層4〇。此第二 複晶妙層40厚度約為150_25 0 nm典型值約20 0 0埃。再以光 阻圖案及敍刻技術定義字線(word 1 ine),在光阻圖案剝 除後’即以另一曝露出源極區及部分氧化區25的光阻圖案 7 0形成於所有區域β源極佈植後,光阻圖案5 〇剝除,再進 行含氧氣氛下的退火,以擴大源極區6〇,以增加源極區60 對懸浮閘極1 〇的耦合比。然而上述的退火將再次惡化微笑 效應現象’而使得程式化速度變差(因部分閘極氧化層變Page 6 4¾ 7 712 V. Description of the invention (3) The center of the oxidized region 25 and the vaporized region 25 is about 120 nm. The values are thin on both sides. This result will make the first polycrystalline silicon layer 1 The upper surface of 〇 shows a structure in which both terminals point upward. Then 'remove the remaining silicon nitride layer mask with a hot phosphate solution, and then use plasma etchant to remove the first polycrystalline stone layer 10 which is not covered by the oxidized area 25'. This is sandwiched by the gate oxide The first polycrystalline silicon layer 10 of the layer 8 and the oxidized region 25 is used as a floating gate. Please refer to FIG. 2 'and form another oxide layer 30 by thermal oxidation. In order to reduce the smile effect which is not conducive to the stylization speed, the formation step includes: firstly thermally oxidizing and then growing an oxide layer of 40 angstroms, and then depositing an HTO oxide layer of about 130 angstroms under a higher temperature f5 environment. In order to reduce the light-to-light ratio of the suspended gate to the control gate, the purpose of growing an oxide layer of 70 angstroms by the thermal oxidation method to make the oxide layer 30 thicker than the gate oxide layer 8 is to reduce the light gate ratio of the floating gate to the control gate. However, the above-oxidation step has caused a certain degree of smile effect at the bottom edge of the suspended gate, as shown by the reference numeral 35 in FIG. 'Next', please refer to FIG. 3, and then form a second polycrystalline silicon layer 40. The thickness of this second polycrystalline layer 40 is about 150-25 nm, and the typical value is about 200 angstroms. Then use the photoresist pattern and engraving technology to define the word line (word 1 ine). After the photoresist pattern is stripped, another photoresist pattern 70 that exposes the source region and part of the oxidized region 25 is formed in all regions. After the β source is implanted, the photoresist pattern is stripped 50 and then annealed in an oxygen-containing atmosphere to expand the source region 60 and increase the coupling ratio of the source region 60 to the floating gate 10. However, the above annealing will worsen the smile effect phenomenon again and make the programming speed worse (because part of the gate oxide layer changes
麟Lin
第7頁 «57712 五、發明說明(4) 厚了)。 請參考圖四A,再形成間隙壁6 5、6 8於字線側壁, (步驟包含沉積約1 5 0 0埃的氮化矽層,再以一非等向性蝕 刻方法在閘極(控制閘極4 0及懸浮閘極1 0 )的側壁上分別形 成間隙壁6 5、6 8。最後為進行汲極區的佈植,再形成一光 阻圖案7 0在記憶胞上,以防止佈植離子傷害到懸浮閘極結 構(特別是懸浮閘極上方的氧化層2 5 ),然而由於疊對誤差 2限制’被植入字線之離子便有某種程度的不均勻而使得 子線的阻值變得不穩定。圖四B的俯視圖同時顯示由於光 阻圖案的疊對誤差,以致於不對稱。 請參 所有區域 術的能量 面,以產 欽金屬碎 圖案,否 物的特性 促使欽金 板反應, 再去除未 等,最後 轉成具有 考圖五’在全面形成厚度約300埃的鈦金屬層於 ’再進行石夕離子混層(ion mixing)佈植,這一技, 約以調整至複晶矽層或矽半導體基板和鈦金屬界 生一非晶碎(amorphous silicon)即可,以提高 化層的品質,請注意,這一步驟應避免使用光阻 則去光阻的步驟中,鈦將因氧化而劣化金屬矽化 ,最後,先進行第一次較低溫的快速熱退火,以 屬層與該第二複晶矽層及源、汲極區之半導體基 形成低溫金屬矽化物80例如C49 Ti Si 2,接著, 反應之金屬層,例如間隙壁,隔離區上的金屬層 再施以第二次退火,以促使低溫金屬矽化物層8 0 較低阻質之金屬矽化物層例如C 5 4 T i S i 2。Page 7 «57712 V. Description of the invention (4) Thick). Please refer to FIG. 4A, and then form the spacers 65, 68 on the sidewalls of the word line. (The steps include depositing a silicon nitride layer of about 150 angstroms, and then anisotropically etching the gate (control The gate electrode 40 and the floating gate electrode 10) are respectively formed with gap walls 65 and 68. Finally, for the implantation of the drain region, a photoresist pattern 70 is formed on the memory cell to prevent the cloth The implanted ions damage the suspended gate structure (especially the oxide layer 2 5 above the suspended gate). However, due to the overlap error 2 limitation, the ions implanted into the word line have a certain degree of non-uniformity, which makes the strands The resistance value becomes unstable. The top view of Figure 4B also shows the asymmetry due to the stacking error of the photoresist pattern. Please refer to the energy surface of all areas to produce a broken metal pattern. The gold plate was reacted, and then removed, and finally converted into a titanium metal layer having a thickness of about 300 angstroms in the whole figure, and then the ion mixing layer was implanted. This technique is about to adjust Amorphous to the polycrystalline silicon layer or silicon semiconductor substrate and titanium (Amorphous silicon) can improve the quality of the chemical layer. Please note that in this step, photoresist should be avoided. In the step of photoresist removal, titanium will degrade the silicon due to oxidation. Finally, the first comparison The rapid thermal annealing at low temperature forms a low-temperature metal silicide 80 such as C49 Ti Si 2 with the metal layer and the second polycrystalline silicon layer and the semiconductor base of the source and drain regions. Then, a reactive metal layer, such as a spacer, is isolated. The metal layer on the region is subjected to a second annealing to promote the low-temperature metal silicide layer 80 and the lower-resistance metal silicide layer such as C 5 4 T i S i 2.
第8頁Page 8
五 '發明說明(5) 發明目的及概述 本發明之一目的係提供一種改善八 保存問題與程式化速度之快閃二:閘快閃記憶體資料 體製造方法。 本發明之另一目的係提供—種分 用以改善分閘快閃記憶體資料保存門快閃記憶胞結構, 于问嗵與程式化速度。 本發明再一目的係提供一種分閘快 氪化矽覆蓋層將源極區靠近懸浮鬧J S己憶胞結構, ,氧化層形成過程中『特:極,邊緣覆蓋, 用 溫氧化導致懸滚明把疋介複晶石夕氧*介 於有 此吁用从1高溫氧化導致懸浮:別是介複晶矽氧价 廣的形成)此外也因為控制間極之閑極氧化層餐 尊的問題二上層的氧化層氮石夕氧化層覆 光陴微影且士過程限制導致控制閘極的阻植不 制蘭極因金屬發化物形成前的佈植對懸浮閘極 損傷° 了整 植 定 時由於 膺 τ整:ί光陴微 程限制導致控制閘極的阻植不 焱明揭霧一種分閘快閃記憶胞結構及形成方法,用 本分閘快閃記憶罈資料保存問題與程式化速度^其中 以改善1胞必含•具有一閣極氧化層/第一複晶石夕晶層/第 快閃纪ί咭構之懸浮閘極結構’形成於一半導體基板上’ ^ 457712 • ·*-«5 'Explanation of the invention (5) Purpose and summary of the invention One object of the present invention is to provide a flash memory which can improve the storage problem and programming speed 2: the method of manufacturing flash memory data. Another object of the present invention is to provide a method for improving the flash memory cell structure of the flash memory data storage gate for opening and closing, and the speed of programming and programming. Another object of the present invention is to provide a fast-breaking siliconized silicon coating layer that closes the source region to the suspended JS memory cell structure. During the formation of the oxide layer, "specially, the electrode is covered by the edge. The interstitial complex spar oxygen * is caused by this call to oxidize from 1 high temperature to cause suspension: not to mention the formation of metasilicon silicon with a wide range of oxygen values) In addition, it is also due to the problem of controlling the leisure time of the oxide layer. The oxide layer on the upper layer is covered with light. The lithography and photolithography are limited, and the control of the gate is blocked by the control of the gate. The blue pole is damaged by the planting before the formation of metal compounds. τ Integrity: The limitation of the micro-traveling limit leads to the blocking of the control gate. It is not clear how to open the fog. A flash memory cell structure and method of forming the flash gate. Use this flash memory field to save data and program speed. In order to improve 1 cell must contain • Suspended gate structure 'formed on a semiconductor substrate' with a polar oxide layer / first polycrystalline spar crystal layer / Earth-Earth stage flash structure ^ 457712 • · *-«
Ά .V 五、發明說明(6) 一源極區形成於半導體基板内,由懸浮閘極向懸浮閘極結 構外之半導體基板延伸’一控制閘極則形成於半導體基板 上’並在懸浮閛極結構侧壁及該第一氧化層的一部分上 方,一汲極區形成於半導體基板内,自控制閘極邊向離開 懸浮閘極結構之半導體基板延伸,以使得汲極、控制閘 極、懸浮閘極及源極區延一方向依序呈現,此外一第二氧 化層形成於半導體基板上’及該懸浮閘極結構上。 本發明之快閃記憶胞結構特徵為:具有一氮矽氧化覆 蓋層形成於第二氧化層上,該氮矽氧化覆蓋層覆蓋源極 區’並向懸浮閘極結構側壁覆蓋,再覆蓋第一氧化層的上 方表面的一部分’因此’懸浮閘極上方的第一氧化層完全 被控制閘極及氮矽氧化覆蓋層共同覆蓋,控制閘極並經由 部分之該第二氧化層’相鄰接於該半導體基板及該懸浮閘 極。其中該氮矽氧化覆蓋層的形成方法係以沉積氮化矽 層’再以光阻圖案定義位置,最後在源極雜質離子佈植後 的退火步驟時以含氧氣氛下進行而轉變成氮矽氧化層,當 然此.氮矽氧化層也可以直接以控制氧和氮的比例直接沉 積0 由於此氮矽氧化覆蓋層的存在,可以防止離子植入時 損傷第一氧化層’導致第一氧化層補捉(trap)電子,以致 於儲存資料不穩定的問題。V.V V. Description of the invention (6) A source region is formed in the semiconductor substrate, and the suspended gate extends to the semiconductor substrate outside the suspended gate structure. 'A control gate is formed on the semiconductor substrate.' Above the sidewall of the electrode structure and a portion of the first oxide layer, a drain region is formed in the semiconductor substrate, extending from the edge of the control gate to the semiconductor substrate leaving the suspended gate structure, so that the drain, control gate, and suspension The gate and source regions are sequentially presented in one direction, and a second oxide layer is formed on the semiconductor substrate and the floating gate structure. The flash memory cell structure of the present invention is characterized in that a nitrogen silicon oxide coating is formed on the second oxide layer, the nitrogen silicon oxide coating covers the source region and covers the side wall of the suspended gate structure, and then covers the first A portion of the upper surface of the oxide layer 'so' the first oxide layer above the suspended gate is completely covered by the control gate and the nitrogen-silicon oxide cover layer, and the control gate is adjacently connected via a portion of the second oxide layer The semiconductor substrate and the floating gate. The method for forming the nitrogen-silicon oxide coating layer is to deposit a silicon nitride layer, and then define a position with a photoresist pattern. Finally, the annealing step after the implantation of source impurity ions is performed in an oxygen-containing atmosphere to convert the silicon nitride layer Oxide layer, of course this. Nitrogen silicon oxide layer can also be deposited directly to control the ratio of oxygen and nitrogen. Because of the existence of this nitrogen silicon oxide coating, it can prevent the first oxide layer from being damaged during ion implantation and cause the first oxide layer. Trap electrons so that the stored data is unstable.
五、發明說明(7) 並d:除丄提供單一分閘*决閃記憶胞之具有結構外 圾供nor型快閃記憶胞陣列之形成方法。 發明詳細說明: 以 列幾個 法之快閃記憶胞製程中有下 上發明背景所述傳統方 問題需要加以克服:V. Description of the invention (7) and d: In addition to providing a structure of a single-opening * flash memory cell, a method for forming a nor-type flash memory cell array. Detailed description of the invention: In the flash memory cell manufacturing process of the following methods, the traditional method described in the background of the invention needs to be overcome:
上1)由於氧化層形成過程’ #別是介複晶梦氧化層的 A閘快閃記憶體閘極氧化層因氧化而產生懸浮閘 極邊緣閘極氧化層變厚的微笑效果(sraiHng effect),將 不利於源極對懸浮閘極的耦合比,而使得程式化速度變 差0 (2) 由於淡極區離子佈植時,通常要加一光罩在記憶 胞上’以’防止佈植離子傷害到懸浮閘極結構(特別是懸浮 閘極上方的氧化層)’然而這卻會使得因疊對誤差造成控 制閘極導電性雜質分佈不均導致阻值不穩定之問題。 (3) 此外,由於在形成源没極及閘極之自對準鈦金屬 矽化層時,通常會在鈦金屬沉積後先進行離子混層(i on m i x i ng)以提高鈦金屬矽化層的品質,(複晶矽層和鈦金屬 界面離子植入以產生一非晶矽(amorphous silicon),在 這一離子混層步驟最好不要再形成光阻圖案,以免不利鈦Above 1) As a result of the oxide layer formation process, #other than the A-gate flash memory gate oxide layer of the mesomorphic dream oxide layer, the sraiHng effect of the thickened gate oxide layer due to oxidation is generated due to oxidation. , Which is not conducive to the coupling ratio of the source to the floating gate, and makes the stylization speed worse. 0 (2) When ion implantation in the light pole region, a photomask is usually added to the memory cell to prevent the implantation. The ions hurt the suspended gate structure (especially the oxide layer above the suspended gate). However, this will cause the resistance of the gate to be unstable due to uneven distribution of the conductive impurities in the control gate due to stacking errors. (3) In addition, when forming a self-aligned titanium silicide layer of a source electrode and a gate electrode, an i on mixi ng is usually performed after the titanium metal is deposited to improve the quality of the titanium silicide layer. (Ion implantation at the interface between the polycrystalline silicon layer and the titanium metal to produce an amorphous silicon. It is best not to form a photoresist pattern in this ion mixing step, so as not to adversely affect titanium.
第11頁 45>?71 2 五、發明說明(8) ' -- 金属石夕化物之品質’而這不使用光阻圖案的結果又將使r 懸浮閘極上方的氧化層曝露而造成離子損傷,包含電子ρ 入(electron trapping) ’而使得資料的保持能力變差,a 因此傳統方法中這一步驟會造成兩難。 為了解決上述的幾個問題’本發明將提供一種分間快 閃記憶胞結構,利用一氮矽氧化覆蓋層覆蓋源極區,並向 鄰接懸浮閘極結構側壁覆蓋,再覆蓋第一氧化層的上方表 面的一部分’如此氮化妙覆蓋層及控制閘極的^晶矽層可 以完整覆蓋懸浮閘極上方的氧化層’因此對汲極區進行離 子佈植時可全面佈植,不需光阻圖案’且鈦金屬沉積後離 子混層的問題可以解決’另一好處是不只間複晶矽氧化層 可防止微笑現象產生,同時,源極佈植後使源極區擴大的 退火’也可防止微笑現象再次出現β 以下發明的細節可參考圖示來加以詳細說明並描繪本 發明之分閘快閃記憶體元件的製作方法。 圖六Α及Β分別顯示閘極氧化層丨〇 8 /第一複晶矽晶層 1 1 0/第一氧化層1 2 5之懸浮閘極堆疊結構形成於一半導體 基板1 0 5的主動區1 〇 4上的橫截面示意圖及俯視圖。懸浮閘 極堆營結構形成方法如下:首先,以傳統方法在半導體基 板10 5定義淺溝渠隔離區2及主動區1〇4,再以熱氧化 法,溫度約8 0 0 - 9 5 0 °C形成一閘極氧化層1 〇 8於一 < 1 〇 〇 >晶Page 11 45 >? 71 2 V. Description of the invention (8) '--Quality of metal oxides' and the result of not using a photoresist pattern will expose the oxide layer above the levitation gate and cause ion damage Including the electron trapping ('electron trapping') makes the data retention ability worse, a. Therefore, this step in the traditional method will cause a dilemma. In order to solve the above-mentioned problems, the present invention will provide a partitioned flash memory cell structure. The source region is covered with a silicon nitride oxide coating layer, and the side wall adjacent to the suspended gate structure is covered, and then covered above the first oxide layer. A part of the surface 'so that the nitrided cover layer and the crystalline silicon layer controlling the gate can completely cover the oxide layer above the suspended gate', so the ion region of the drain region can be fully implanted without the need for a photoresist pattern 'And the problem of ion mixing layer after titanium metal deposition can be solved.' Another benefit is that not only the inter-crystalline silicon oxide layer can prevent the occurrence of smiles, but also the annealing of the source area after the source is implanted can prevent smiles. Reappear β. The details of the following invention can be described in detail with reference to the drawings and describe the method of making the flash memory element of the present invention. Figures 6A and B respectively show the gate oxide layer. The floating gate stack structure of the gate oxide layer 1/10 / first oxide layer 1 10 / first oxide layer 25 is formed on the active area of a semiconductor substrate 105. Cross-section schematic and top view on 104. The method of forming the floating gate stack structure is as follows: First, the shallow trench isolation region 2 and the active region 104 are defined on the semiconductor substrate 105 by the conventional method, and then the temperature is about 80 0-9 50 ° C by the thermal oxidation method. Forming a gate oxide layer 108 on a < 100 > crystal
第12頁 15771 2 五、發明說明(9) 體方向之單晶半導體基板i 05閘極氧化層i 〇8的厚度約為7. 0-10.0⑽,典型值為80 nffi。接著以低壓化學氣相沉積法 (LPCVD)全面沉積一第一複晶矽層11〇,沉積溫度約 5 5 0 - 6 5 0。C,此第一複晶矽11 0厚度約為8 〇 _ i 2 〇 ηιη,典 型值為100 nmePage 12 15771 2 V. Description of the invention (9) The thickness of the single-crystal semiconductor substrate i 05 gate oxide layer i 〇8 in the bulk direction is approximately 7.0-10.0 ⑽, with a typical value of 80 nffi. A low-pressure chemical vapor deposition (LPCVD) method is then used to deposit a first polycrystalline silicon layer 110 at a deposition temperature of about 550-650. C, the thickness of this first polycrystalline silicon 11 0 is about 8 〇 _ i 2 〇 ηη, the typical value is 100 nme
P 隨後’再以低壓化學氣相沉積法再全面沉積一厚度約 為8 Onm之氮化矽層(未圖示)隨後,形成一光阻圖案(未圖 示)於氮化破層(未圖示)上以定義懸浮閘極的位置。接 著,以一非等向性蝕刻方法蝕刻氮化矽層(未圖示)以轉移 光阻圖案至氮化矽層上》在去除光阻圖案後,施以高溫的 熱乳化製程’使第一複晶石夕層110的上表面成長第一氧化 層125 ’第一氧化層125的甲央厚約12〇 nm,但兩邊則很 薄’這樣的結果將使得第一複晶矽層11 〇的上表面呈現出 兩終端向上尖的結構。接著,再以熱磷酸鹽溶液去除其餘 之氮化矽層罩幕。然後再以電漿蝕刻去除未被第一氧化層 12 5覆蓋之第一複晶矽層110後’最後以稀釋的氫氟酸或 B 0 E溶液潤濕以去除半導體基板1 0 5上的閘極氧化層1 〇 8 » 隨後’請參考圖七所示的橫截面示意圖在懸浮閘極堆 叠結構形成後再形成第二氧化層1 3 0。第二氧化層1 3 0的形 成方法如下’首先施以兩溫的熱氧化製程以形成薄氧化 層,約40埃即可’此步驟也可以省略,以全面防止氧化過 程中閘極氧化層邊緣變厚的微笑現象。接著以LPCVD法全P Then, a low-pressure chemical vapor deposition method is used to fully deposit a silicon nitride layer (not shown) with a thickness of about 8 Onm. Subsequently, a photoresist pattern (not shown) is formed on the nitride breakdown layer (not shown). (Shown) to define the position of the floating gate. Next, an anisotropic etching method is used to etch the silicon nitride layer (not shown) to transfer the photoresist pattern onto the silicon nitride layer. After removing the photoresist pattern, a high temperature thermal emulsification process is performed to make the first The first oxide layer 125 is grown on the top surface of the polycrystalline stone layer 110. The thickness of the first oxide layer 125 is about 120 nm, but both sides are very thin. This result will make the first polycrystalline silicon layer 11 The upper surface shows a structure with the two terminals pointing upward. Then, the remaining silicon nitride mask is removed with a hot phosphate solution. Then, plasma etching is used to remove the first polycrystalline silicon layer 110 that is not covered by the first oxide layer 12 5 'and finally wetting with a diluted hydrofluoric acid or B 0 E solution to remove the gate on the semiconductor substrate 105 Polar oxide layer 1 0 8 »Subsequently, please refer to the schematic cross-sectional view shown in FIG. 7 to form a second oxide layer 130 after the suspension gate stack structure is formed. The method for forming the second oxide layer 130 is as follows: 'First, a two-temperature thermal oxidation process is performed to form a thin oxide layer, which is about 40 angstroms.' This step can also be omitted to prevent the gate oxide layer edge during the oxidation process. Thickening smile phenomenon. LPCVD
第13頁 457712 五、發明說明(ίο) 面沉積一高溫氧化層(以下稱HTO氧化層)1 3 〇在懸浮閘極結 構側壁及半導體基板10 5之其餘位置上。以一較佳的實施 例而言’ ΗT0氧化層1 30係在800-900。 C沉積至1 〇-2 0 nm °Page 13 457712 V. Description of the invention (ίο) A high-temperature oxide layer (hereinafter referred to as HTO oxide layer) is deposited on the side surface of the floating gate structure and the rest of the semiconductor substrate 105. In a preferred embodiment, the 'ΗT0 oxide layer 1 30 is in the range of 800-900. C deposited to 10--20 nm
再以LPCVD法約65 0。C沉積一氮化矽覆蓋層140於所 有之表面,氮化矽層厚度約10-40 nm (典型值約200埃), 隨後,形成一光阻圖案1 4 5於氮化矽層上,再以光阻圖案 1 4 5為罩幕,施以银刻技術,以形成氮化石夕覆蓋層1 4 〇。以 —較佳的實施例而言’懸浮閘極相鄰接之源極區及懸浮閘 極上的第一氧化層12 5大約一半區域都需以連續的氮化石夕 覆蓋層14 0覆蓋。此外’氮化妙覆蓋層140覆蓋的程度至少 要能與之後形成之控制閘極有部分重疊以完整覆蓋第一氧 化層1 2 5。 氮化矽覆蓋層14 0可以避免微笑效應的發生,間接因 此增加源極區對懸浮閘極的耦合比,而因此增加程式化速 度。因此’在氮化石夕覆蓋層1 4 0形成後,再繼續進行高温 熱氧化製程可以再形成一約厚70埃的氧化層,請注意在這 一步驟中只有靠近汲極區的一側如圖中指標14 8所示,會 有微笑效應現象’靠源極區的一侧如指標1 4 9所示則無, 因此,源極對懸浮閘極的耦合比就不會降低,請注意此時 氮化矽覆蓋層1 40將因此轉成氮矽氧化覆蓋層 (oxyni tride)140 〇About 650 by LPCVD. C deposit a silicon nitride cover layer 140 on all surfaces, the thickness of the silicon nitride layer is about 10-40 nm (typically about 200 angstroms), and then a photoresist pattern 1 4 5 is formed on the silicon nitride layer, and then The photoresist pattern 1 4 5 is used as a mask, and a silver engraving technique is applied to form a nitride stone cover layer 14. In the preferred embodiment, the source region adjacent to the floating gate and the first oxide layer 12 5 on the floating gate need to be covered with a continuous nitride coating layer 140. In addition, the coverage of the nitride coating layer 140 must be at least partially overlapped with the control gate formed later to completely cover the first oxide layer 1 2 5. The silicon nitride covering layer 140 can avoid the occurrence of the smile effect, thereby indirectly increasing the coupling ratio of the source region to the floating gate, and thus increasing the programming speed. Therefore, after the formation of the nitrided stone overcoat layer 140, the high temperature thermal oxidation process can be continued to form an oxide layer with a thickness of about 70 Angstroms. Please note that only the side close to the drain region is shown in this step as shown in the figure. As shown in the medium index 14 8, there will be a smile effect phenomenon. The side relying on the source region is not shown as the index 1 4 9. Therefore, the coupling ratio of the source to the floating gate will not decrease. Please note that at this time The silicon nitride coating 1 40 will thus be converted to an oxyni tride 140.
第14頁 J57T1 2 .. 五、發明說明(11) 接著請參考圖八的橫截面示意圖,再全面形成第二複 晶石夕層。此第二複晶碎層厚度約為1 5 0 - 2 5 0 n m典型值約. 2 0 0 0埃。再以光阻圖案(未圖示)及蝕刻技術定義字線 (w 〇 r d 1 i n e ) 1 5 0 ’請注意字線至少需能覆蓋部分的氮矽氧 化覆蓋層1 4 0以防止懸浮閘極上的第一氧化層1 2 5 (或第二 氧化層1 3 0 )外露。在光阻圖案剝除後,即以另一曝露氮矽 氧化覆蓋層1 4 0的光阻圖案1 6 0形成於所有區域,用以進行 源極區1 7 0離子佈植。 接著,光阻圖案1 6 (3剝除,再進行含氧氣氛下的退 火,以擴大源極區1 7 0,以增加源極區1 7 0對懸浮閘極1 1 0 的搞合比。以一較佳的實施例而言,佈植的離子係η型導 電雜質如鱗等,佈植的能量和劑量分別約為2 0 - 6 0 k e V和1 X 1 0 14-lx 1 0 16 / cm2。由於有氮矽氧化覆蓋層1 40的覆蓋, 退火不致於產生微笑效應現象,因此源極區對懸浮閘極的 耦合面積不會被減少,自然可提昇程式化速度。 請參考圖九A,以低壓化學氣相沉積法全面沉積一約 1 5 0 0埃的氮化矽層於所有區域,再施以一非等向性姓刻方 法以分別形成大、小間隙壁1 6 5、1 6 8於字線側壁,及懸浮 閘極堆疊結構之側壁上(大、小間隙壁1 6 5、1 6 8係由於深 寬比的關係),由於先前沉積之氮化石夕覆蓋層已氧化為氮 矽氧化覆蓋層1 4 0。而氮化矽層與氮矽氧化層具有蝕刻選Page 14 J57T1 2 .. V. Description of the invention (11) Next, please refer to the schematic diagram of the cross section in Fig. 8 to form the second polycrystalline stone layer. The thickness of this second multicrystalline fragment is about 15 0-2 5 0 n m, with a typical value of about 2 0 0 0 Angstroms. Then use the photoresist pattern (not shown) and etching technology to define the word line (w 〇rd 1 ine) 1 5 0 'Please note that the word line must be able to cover at least part of the nitrogen silicon oxide coating 1 4 0 to prevent the floating gate The first oxide layer 1 2 5 (or the second oxide layer 1 3 0) is exposed. After the photoresist pattern is peeled off, another photoresist pattern 160, which exposes the silicon nitride oxide overcoat layer 140, is formed in all regions for the ion implantation of 170 in the source region. Next, the photoresist pattern 16 is stripped and then annealed in an oxygen-containing atmosphere to expand the source region 170 and increase the ratio of the source region 170 to the floating gate 1 110. In a preferred embodiment, the implanted ionic n-type conductive impurities, such as scales, have an energy and dose of approximately 2 0-6 0 ke V and 1 X 1 0 14-1x 1 0 16 / cm2. Because it is covered by the nitrogen-silicon oxide cover layer 1 40, the annealing will not cause a smile effect, so the coupling area of the source region to the floating gate will not be reduced, which can naturally increase the programming speed. Please refer to Figure 9 A, a low-pressure chemical vapor deposition method is used to fully deposit a silicon nitride layer of about 1 500 angstroms in all areas, and then an anisotropic method is used to form large and small gaps, respectively. 1 6 8 on the side wall of the word line and the side wall of the floating gate stack structure (large and small gaps 1 6 5 and 1 6 are due to the aspect ratio), due to the previously deposited nitrided oxide coating layer has been oxidized The silicon nitride oxide cover layer 1 40. The silicon nitride layer and the silicon nitride oxide layer have an etching option.
第15頁 457 71 2 五、發明說明(12) 擇比,因此非等向餘 氣矽氧化覆蓋層14Γ 成間隙壁時並不會#刻去除此 形成汲極175,此時士著’全面施以源/沒極區的佈植以 子佈植時不會對第一\於有氣碎氧化覆蓋層140的覆蓋,離 需光阻圖案,因此,i化層造成損傷,除此之外,由於不 150)的各個部位,子同時植入於字線(第二複晶石夕層 電阻不均的問題。不會如傳統方法字線150產生片 明進行汲極的離子:^示俯視圖,請和圖四Β比較,本發 對誤差的問題。 時不需光阻圖# ’因此沒有微影疊 請參考圖十’全面形成厚度約300埃的鈦金屬層於所 有區域再進行石夕離子混層(i〇n mixing)佈植,這一技術 的能量約以調整至複晶矽層或矽半導體基板和鈦金屬界 面’以產生一非晶發(amorph〇us s i 1 i con)即可,以提高 鈦金屬梦化詹的品質,以一較佳的實施例而言,佈植的能 量和劑量分別為3 0 - 5 0 keV和5x 10 u-3x l〇i5 /cm2。由於 氮矽氧化覆蓋層14 0的覆蓋離子佈植時將不會對第一氧化 層造成損傷。請注意,這一步驟仍和發明背景所述相同, 應避免使用光阻圖案,否則去光阻的步驟中,欽將氧化, 而劣化金屬碎化物的特性》 最後’再進行第一次較低溫的快速熱退火,以以促使 該金屬層與該第二複晶矽層1 5 0及源、汲極區之半導體基 板反應,形成低溫金屬矽化物19 0例如C49 TiSi2,接著,Page 15 457 71 2 V. Description of the invention (12) Select ratio, so the non-isotropic residual silicon oxide cover layer 14Γ will not be removed when the formation of the drain electrode 175 at this time. When the source / non-polar area is planted and the seed is planted, the first and the gas-breaking oxidation cover layer 140 are not covered, and the photoresist pattern is not needed. Therefore, the i-layer causes damage. In addition, Due to the different parts of 150), the sons are simultaneously implanted in the word line (the second polycrystalline stone layer has a problem of uneven resistance. The word line 150 does not generate ions that are clearly drained as in the traditional method: ^ shows a top view, Please compare with Figure 4B for the error problem. I don't need a photoresist pattern. "There is no lithography. Please refer to Figure 10." Fully form a titanium metal layer with a thickness of about 300 angstroms and perform Shi Xi ion in all areas. Ion mixing is implanted. The energy of this technology can be adjusted to the interface between the polycrystalline silicon layer or the silicon semiconductor substrate and the titanium metal to generate an amorphous hair (amorph〇us si 1 i con). In order to improve the quality of the titanium metal dream, in a preferred embodiment, the energy and dose of implantation They are 3 0-50 keV and 5x 10 u-3x l0i5 / cm2. Because the nitrogen-silicon oxide coating 14 0 is covered with ion implantation, it will not cause damage to the first oxide layer. Please note that this The steps are still the same as those described in the background of the invention. The use of photoresist patterns should be avoided. Otherwise, in the step of removing photoresist, the catalyst will be oxidized and the characteristics of the metal fragments will be deteriorated. Finally, the first low temperature rapid thermal annealing is performed. In order to promote the reaction of the metal layer with the second polycrystalline silicon layer 150 and the semiconductor substrate of the source and drain regions to form a low temperature metal silicide 19 0 such as C49 TiSi2, then,
第16頁 457712 五、發明說明(13) 再以濕式蝕刻的方式,去除未反應之金屬層,例如間隙 壁,隔離區等上的金屬層,最後再施以第二次退火,以促 使低溫金屬矽化物層轉成具有較低阻質之金屬矽化物層 1 9 0例如C 5 4 T i S i 2。以一較佳的實施例而言,第一次較低 溫的快速熱退火之退火溫度約為650-750 t,第二次較高 溫的快速熱退火之退火溫度約為750 — 9〇〇它。 以上所述僅為本發明之較佳實施例而已,並非用以限定本 發明=申請專利範圍;凡其它未脫離本發明所揭示之精神 下所7b成之等效改變或修飾,均應包含在下述之申請專利 範圍内。Page 16 457712 V. Description of the invention (13) Unreacted metal layers, such as those on the barrier wall, isolation zone, etc. are removed by wet etching, and then a second annealing is performed to promote low temperature The metal silicide layer is converted into a metal silicide layer with a lower resistance 190 such as C 5 4 T i S i 2. In a preferred embodiment, the annealing temperature of the first lower temperature rapid thermal annealing is about 650-750 t, and the annealing temperature of the second higher temperature rapid thermal annealing is about 750-900 °. The above is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention = patent application; all other equivalent changes or modifications that do not depart from the spirit disclosed by the present invention should be included in the following Within the scope of the patent application.
第17頁 4 5 7 71 g_ 圖式簡單說明 本發明的較佳實施例將於往後之說明文字中輔以下列 圖形做更詳細的闡述: .圖一顯示以傳統方法製造分閘快閃記憶胞至形成一閘 極氧化層/第一複晶矽晶層/第一氧化層之懸浮閘極堆疊結 構形成於一半導體基板上的横截面圖。 圖二顯示以傳統方法形成第二氧化層於懸浮閘極堆 疊結構及半導體基板之橫截面示意圖。 圖三顯示以傳統方法製造分閘快閃記憶胞至完成源極 區離子佈植以及退火後以擴大源極區的橫截面不意圖。 圖四A及圖四β分別為形成光阻圖案於控制閘極以進行厂· 源極區佈植但又有疊對誤差的橫截面及俯視示意圖。 圖五顯示以傳統方法製造分閘快閃記憶胞至進行離子 混層及形成金屬矽化物層之橫截面示意圖。 圖六Α及圖六Β分別顯示以本發明之方法製造分閘快閃 記憶胞至形成一閘極氧化層/第一複晶矽晶層/第一氧化層 之懸浮閘極堆疊結構形成於一半導體基板上的橫截面及俯 視示意圖。 圖七顯示.以本發明之方法形成第二氧化層、氮化矽覆 蓋層的橫截面示意圖。 圖八顯示以本發明之方法製造分閘快閃記憶胞至完成 ( 源極區離子佈植以及退火後以擴大源極區的橫截面示意 圖。 圖九A及圖九B分別顯示以本發明之方法,形成間隙壁 及施以離子佈植以形成没極區的橫截面示意圖及俯視圖。Page 17 4 5 7 71 g_ Schematic illustration of the preferred embodiment of the present invention will be described in more detail in the following explanatory text with the following graphics: Figure 1 shows the traditional method of making the flash memory opening and closing A cross-sectional view of a floating gate stack structure with a gate oxide layer / first polycrystalline silicon layer / first oxide layer formed on a semiconductor substrate. FIG. 2 is a schematic cross-sectional view of a conventional method for forming a second oxide layer on a floating gate stack structure and a semiconductor substrate. Figure 3 shows that it is not the intention to make the flash memory cell by the conventional method to complete the ion implantation in the source region and to expand the cross section of the source region after annealing. Figures 4A and 4β are schematic cross-sections and top views of forming a photoresist pattern on the control gate for plant-source region implantation but with overlapping errors. Figure 5 shows a schematic cross-sectional view of a conventional method for fabricating an open flash memory cell to perform ion mixing and form a metal silicide layer. FIG. 6A and FIG. 6B respectively show that the gate flash memory cell is manufactured by the method of the present invention to form a gate oxide layer / the first polycrystalline silicon layer / the first oxide layer and the suspended gate stack structure is formed in a A cross-section and a schematic plan view on a semiconductor substrate. FIG. 7 shows a schematic cross-sectional view of forming a second oxide layer and a silicon nitride capping layer by the method of the present invention. Figure 8 shows the cross-section of the flash memory cell manufactured by the method of the present invention to completion (source implantation in the source region and annealing to expand the source region. Figures 9A and 9B respectively show the use of the present invention. Method, forming a cross-section wall and applying ion implantation to form a schematic cross-sectional view and a top view of an electrodeless region.
1^=.1 ^ =.
第18頁 4STT1 c 圓式簡單說明 間快閃記憶胞至進 面示意圖。 圖十顯示以本發明之方法,製造分 行離子混層及形成金屬矽化物層之橫截 圖號 :對照表: 5 、 105 半 導 體 基 板 8、 108 閑 極 氧 化 層 10、 110 第 一 複 晶 矽 層、 25 氧 化 區 30 氧 化 層 35 微 笑 效 應 現 象 40 第 二 複 晶 矽 層、 50' 70 、 145、 160 光 阻 圖 案 60 、170 源 極 區 65> 68 ' 165' 168 大 小 間 隙 壁 80 ' 190 金 屬 矽 化 物 102 隔 離 區. 104 主 動 區 125 第 一 氧 化 層 130 第 二 氧 化 層 140 .氮 化 矽 覆 蓋 層、 148 微 笑 效 應 現 象 149 無 微 笑 效 應 現象 150 字 線 (word 1 i η 1 175 汲 極 區 浮閘極 制閘極 矽氧化覆蓋層P.18 4STT1 c Simple description of the circular type. Figure 10 shows a cross-sectional view of the method of the present invention for manufacturing branch ion-mixed layers and forming a metal silicide layer: a comparison table: 5, 105 semiconductor substrates 8, 108 anode oxide layers 10, 110 first polycrystalline silicon layers, 25 Oxidation area 30 Oxidation layer 35 Smile effect phenomenon 40 Second polycrystalline silicon layer, 50 '70, 145, 160 Photoresist pattern 60, 170 Source area 65> 68' 165 '168 Size gap 80' 190 Metal silicide 102 Isolation area 104 Active area 125 First oxide layer 130 Second oxide layer 140. Silicon nitride cover layer, 148 Smile effect phenomenon 149 No smile effect phenomenon 150 Word line (word 1 i η 1 175 Drain region floating gate electrode system Gate silicon oxide coating
Claims (1)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW89120822A TW457712B (en) | 2000-10-05 | 2000-10-05 | Flash memory cell structure for improving the data preservation problem and the programming speed and the formation method thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW89120822A TW457712B (en) | 2000-10-05 | 2000-10-05 | Flash memory cell structure for improving the data preservation problem and the programming speed and the formation method thereof |
Publications (1)
Publication Number | Publication Date |
---|---|
TW457712B true TW457712B (en) | 2001-10-01 |
Family
ID=21661454
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW89120822A TW457712B (en) | 2000-10-05 | 2000-10-05 | Flash memory cell structure for improving the data preservation problem and the programming speed and the formation method thereof |
Country Status (1)
Country | Link |
---|---|
TW (1) | TW457712B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7544993B2 (en) | 2002-12-27 | 2009-06-09 | Sharp Kabushiki Kaisha | Semiconductor storage device and portable electronic equipment |
-
2000
- 2000-10-05 TW TW89120822A patent/TW457712B/en not_active IP Right Cessation
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7544993B2 (en) | 2002-12-27 | 2009-06-09 | Sharp Kabushiki Kaisha | Semiconductor storage device and portable electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2010066126A1 (en) | Method for manufacturing split gate type flash memory | |
US7488634B2 (en) | Method for fabricating flash memory device | |
JPH10163348A (en) | Manufcture of nonvolatile semiconductor storage device | |
CN103180952B (en) | Getter in memorizer charge storage structure | |
US7271059B2 (en) | Semiconductor device and method of fabricating the same | |
US6362045B1 (en) | Method to form non-volatile memory cells | |
JP2007053362A (en) | Manufacture method of nonvolatile memory device | |
TWI299193B (en) | Method for fabricating a layer arrangement, layer arrangement and memory arrangement | |
JP2001035945A (en) | Method for forming gate of stack flash type flash eeprom cell | |
TWI267942B (en) | MONOS device having buried metal silicide bit line | |
US6268624B1 (en) | Method for inhibiting tunnel oxide growth at the edges of a floating gate during semiconductor device processing | |
JPH07240478A (en) | Preparation of nonvolatile semiconductor memory device | |
JP2003197783A (en) | Method for manufacturing flash memory cell | |
TW201027624A (en) | Method for fabricating pip capacitor | |
CN101312212A (en) | Non-volatile memory utilizing high K medium and nanocrystalline floating gate and its manufacture method | |
TW457712B (en) | Flash memory cell structure for improving the data preservation problem and the programming speed and the formation method thereof | |
US6767791B1 (en) | Structure and method for suppressing oxide encroachment in a floating gate memory cell | |
CN100552921C (en) | Semiconductor device and manufacture method thereof | |
TWI277205B (en) | Flash memory structure and method for fabricating the same | |
JP3436315B2 (en) | Method of manufacturing MONOS type semiconductor nonvolatile memory device and method of manufacturing semiconductor device | |
JPH0231468A (en) | Manufacture of floating gate type semiconductor memory device | |
KR100466191B1 (en) | Method of manufacturing a flash memory device | |
JPH10189922A (en) | Manufacture of flash memory element | |
JP2004103902A (en) | Nonvolatile semiconductor storage device and its manufacturing method | |
TW514913B (en) | Manufacturing method of flash memory cell with improving programming and erase speed |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
GD4A | Issue of patent certificate for granted invention patent | ||
MK4A | Expiration of patent term of an invention patent |