1291721 (1) 玖、發明說明 相關申請案 此申請案申告2 002年6月23日提出申請的United States Provisional Application Serial Number 60/391,01] 之權利,茲將其中所述者全數列入參考。 【發明所屬之技術領域】 本發明一般係關於半導體範圍。更特定言之,本發明 係關於半導體裝置和晶圓上的膜之原子層移除和原子層交 換。 【先前技術】 新一代半導體裝置需要較薄的介電膜用於MOS電晶 體閘極,和電容器介電物。隨著氧化物減少,隧道漏電顯 著並將閘極氧化物的有用範圍限制至約1 · 8 n m或以下。 高介電常數氧化物(如:Hf〇2 ( k = 20 ) 、Zr02 ( k = 20)及矽酸Hf和Zr )被視爲氧化矽(k = 3 ·9 )的替代 材料,用以提供高電容且無漏電問題的閘極介電物。但是 ’以前技術澱積技巧(如:化學蒸鍍(C V D ))越來越難 滿足先進薄膜的要求。CVD法可藉逐步覆蓋地調整以提 供適應膜,C V D法通常須要高加工溫度,導致高雜質濃 度’且先質或反應物的利用效能欠佳。例如,製造高k閘 極介電物的阻礙之一是在CVD法期間內形成界面氧化矽 層。另一阻礙是以前技術的CVD法對用於高k閘極介電 物的矽底質上之超薄膜澱積造成限制。 -4- 1291721 (2) 使用此技術習知的技巧移除或蝕刻製造半導體裝置中 所用的膜。這些技巧包括溼蝕刻、電漿蝕刻之類。這些技 巧的機構無法將材料移除控制至原子尺寸。通常,不同材 料於相同條件下的蝕刻率不同。因此,典型蝕刻技巧用於 如通過不同材料之多層層合物之孔、溝渠或通道之類的部 件時,部件的側壁無法維持直或均勻形式。此導致後續材 料以連續模式澱積於部件中。因此,這些技巧無法用以一 次準確地移除一個原子層膜。據此,非常須要發展出控制 以原子層尺寸蝕刻材料的技巧。 附圖1和2說明使用以前技術的方法處理底質層之多 層層合物的問題。附圖1中,包含Si02、Si3N4和Al2〇3 層位於底質層上的層合物以氫氟酸(HF )浸液處理。因 爲HF以不同速率蝕刻不同材料,所得層合物具有所不欲 的鋸齒狀邊緣。類似地,附圖2中,H F蝕刻用以移除接 觸孔底部的氧化物層,穿透含有高溫氧化物(Η Τ 0 )、 ShN4、Si〇2、高溫氧化物(ΗΤΟ )和硼磷酸鹽砂酸鹽玻璃 (BPSG )層位於矽底質上的膜層合物。因爲HF與各層材 料的反應速率不周,所得接觸孔會具有,,鋸齒狀邊緣,,壁結 構。接觸孔覆以屏障層(如:氮化鉅(TaN )時,此不均 句的構造造成實行困難。增添TaN屏障層以防止金屬離 子(如:銅)擴散進入晶圓的介電(Si02 )層。TaN覆於 表面上,表面不一致會使得塗層不均勻。這些不均勻度提 供金屬滲透進入經塗覆的層的途徑,此如附圖3所示者。 非均勻接觸孔側壁的影響在晶圓尺寸收縮時更爲顯著。 1291721 (3) 【發明內容】 本發明提出一種藉原子層移除(ALR )和/或原子層 父換(A L E X )用以修飾底質(如:半導體裝置或晶圓) 表面的方法和系統。 藉ALR和ALEx法及此處所述的系統(其中,使用 方法自底質表面均勻地移除第一種固態化合物分子層)可 獲致本發明之優點。底質暴於第一種反應性氣體,與表面 上的第一個分子層反應,在表面上形成中間產物固態化合 物的單層分子。移除第一種反應性氣體而不再與底質接觸 之後’引入第二種反應性氣體。第二種反應性氣體與中間 產物固態化合物反應,形成揮發性或可揮發的產物。之後 ,移除此第二種反應性氣體和揮發性或可揮發產物,使它 們不再與底質接觸’使得留在其表面上的第一種固態化合 物分子層比原來少一層。 本發明的另一實施例利用原子層交換法。於底質上形 成具多個原子層的膜,至少一個第一種固態化合物(至少 具有第一和第二種固態化學物種)的分子層暴於第一種反 應性氣體下。第一種反應性氣體至少具有第一和第二種氣 態化學物種。發生表面反應,其中,第一種固態化學物種 與第一種氣態化學物種交換,形成中間產物固態化合物。 本發明的此特點的另一實施例中,底質是第二種反應性氣 體,其與中間產物固態化合物反應而製得一或多種揮發性 產物,留下第一種固態化合物膜,其分子層比初時少了一 層0 1291721 (4) 【實施方式】 本發明提出一種原子層移除及原子層交換法和系統。 參考此方法的移除觀點,通常使用縮寫”ALR”代表原子層 移除。此外,也可以使用原子層交換”ALEx”描述此方法 。此兩種方法相關,各者用以自底質移除固態化合物表層 。一般而言,本發明提出原子層移除和原子層交換法及系 統,此處,有膜澱積或生長於底質表面上的底質置於反應 器或槽中。第一種反應性氣體被引至反應器或槽中以與膜 的第一層反應,以將第一層轉化成單層固體化合物。反應 器或槽經滌氣以移除未反應的第一種反應物氣體和/或任 何氣相產物。之後將第二種反應物氣體引至反應器或槽中 ,以與單層固態化合物反應而形成氣態化合物,其藉第二 個滌氣步驟自反應器滌除。 本發明的一個實施例中,第一種反應物氣體可以另包 含能源以活化或有助於第一種反應物氣體與膜的第一層之 反應。此能源可爲電磁波形式或一些其他能量傳輸。此電 磁波包括,但不限於,可見光、紅外光、紫外光、微波射 線、無線電頻率射線之類。此射線可以連續形式自設備( 如:雷射)或以非連續形式(如:非同相)形式自設備( 如:燈)供應。使用能源有助於第一種反應物氣體與第一 層膜之反應。 本發明可用以自晶圓表面移除單一原子層。此外,原 子層移除法是自身終結的連續原子層移除法。此方法可用 以在後續的澱積法之前’自表面移除多種化合物。移除的 化合物可以包含故意或意外澱積的氧化物覆層及固態底質 -7- 1291721 (5) 材料和/或藉化學蒸鍍法澱積的層中之化合物。本發明可 用以減低澱積的導電或介電膜厚度,以達到所欲最終膜厚 度。這些僅爲本發明應用的一小部分。 通常,第一種反應性氣體與膜的頂層反應,將頂層轉 化成單層固態化合物。此第一種反應性氣體是化學劑,用 以將表面膜中的第一種固態化合物原子層或底質的較上層 轉化成單一層中間產物固態化合物,後者可以進一步與第 二種反應性氣體反應。第一種反應性氣體之選擇視膜的分 子組成和第一種固態化合物的化性而定。單層中間產物固 態化合物之後與第二種反應性氣體反應,形成揮發性或可 揮發的產物。藉此技術習知的任何適當方式(如:使用惰 性滌氣氣體、使用真空幫浦或使用此兩種技巧之組合)移 除此揮發性或可揮發的產物。 本ALR法的此實施例的前述步驟可節錄成下歹!J式: A (固體)+ B (氣體)一♦ D (固體) (1 ) D (固體)+ C (氣體)一 E (氣體) 个 (2) 其中 A (固體)是底質表面上的第一種固態化合物,B ( 氣體)是第一種反應性氣體,D ( @體)是中間產物固態化合 物,C (氣1 )是第二種反應性氣體,而E (氣if )是揮發性或 可揮發的產物。可重複反應步驟(1 )和(2 )直到自底質 移除所欲的A ( [1體)分子層數。 根據本發明之方法的一個較詳細實施例示於附圖4, 具有膜1 2澱積或生長於底質表面上的底質1 〇置於可控制 反應性氣體暴露順序的反應器、槽或一些其他環境中。此 膜基本上由多個原子層構成,此實例中,由第一種固態化 -8- (6) 1291721 合物A ( ®體_ )構成。此實例中,a (丨固體)膜厚度爲3分子。 但任何厚度的膜和/或其他固體或澱積的表面化合物可根 據本發明處理。存在於底質表面上的第一種固態化合物的 分子或原子層數可視特定應用而有廣泛的變化。被移除的 第一種固態化合物A ( ®體)可以包括半導體加工中所用的 任何類型的膜,如:半導體裝置中常見的任何閘極介電物 或陶瓷、金屬氧化物、矽氧化物、金屬鋁酸鹽、矽鋁酸鹽 、金屬矽酸鹽、金屬氮化物、矽氮化物、純金屬或其他膜 。理論上,只要選用適當反應性氣體,任何表面可以本發 明之方法處理。審視圖(I )所示底.質1 0和膜1 2暴於第 一種氣體 B )。B 以將膜或底質表面上的第一種 固態化合物 A (固體―)的原子和/或分子層轉化成單層中間 產物固態化合物D (固體)的化學劑爲佳。較佳情況中,選 擇與第一種固態化合物A (㈤μ )反應的第一種反應物氣體 B ( m ^ ),以於底質表面上形成單層中間產物固態化合物 D ( ® r,)。第一種反應物氣體B (氣體)與第一種固態化合物 A (固體)之反應以相對於單層中間產物固態化合物D (固體) 反應較爲熱力趨向和/或動力趨向的反應爲佳。換言之, 第一種反應物氣體B (氣體」與乾淨的A (固體)表面之反應應 該會形成經單層中間產物固態化合物D ( ®㈣)塗覆的均勻 表面。前文所謂的熱力趨向是指根據下列式的B (㉟)與 D (丨固)之反應 B (氣體)+D (固體)—X ( 3 ) 放熱情況實質上比反應(1 )來得低。前文所謂動力 趨向是指反應(3 )的反應速率實質上比反應(1 )來得慢 -9- 1291721 (7) 。亦佳的情況中,中間產物固態化合物D (㈤體)形成阻礙 額外的桌一種反應物氣體B ( g㈣)之氣相分子與位於下方 的第一種固態化合物A ( s 層接觸的層,使得僅形成單 層。作爲根據本發明之第一種反應物氣體B (氣體) 的化合物例包括,但不限於,臭氧、氫、醇、水、氨之類1291721 (1) 玖 发明 发明 相关 相关 相关 相关 相关 相关 相关 相关 相关 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 此 United . TECHNICAL FIELD OF THE INVENTION The present invention generally relates to the semiconductor range. More particularly, the present invention relates to atomic layer removal and atomic layer exchange of films on semiconductor devices and wafers. [Prior Art] A new generation of semiconductor devices requires a thin dielectric film for MOS transistor gates and capacitor dielectrics. As the oxide is reduced, tunnel leakage is significant and limits the useful range of gate oxide to about 1 · 8 n m or less. High dielectric constant oxides (eg, Hf〇2 (k = 20), Zr02 (k = 20), and tannic acid Hf and Zr) are considered as alternative materials for yttrium oxide (k = 3 · 9 ) to provide Gate dielectric with high capacitance and no leakage problems. However, previous technological deposition techniques (such as chemical vapor deposition (C V D)) have become increasingly difficult to meet the requirements of advanced films. The CVD method can be adjusted to provide an accommodating film by stepwise coverage. The C V D method usually requires a high processing temperature, resulting in a high impurity concentration and a poor utilization of the precursor or reactant. For example, one of the barriers to making high-k gate dielectrics is the formation of an interfacial yttrium oxide layer during the CVD process. Another impediment is that prior art CVD methods impose limitations on ultra-thin film deposition on the ruthenium substrate for high-k gate dielectrics. -4- 1291721 (2) The film used in the fabrication of a semiconductor device is removed or etched using techniques known in the art. These techniques include wet etching, plasma etching, and the like. These skilled mechanisms are unable to control material removal to atomic size. Generally, different materials have different etch rates under the same conditions. Thus, when typical etching techniques are used for components such as holes, trenches or channels of multilayer laminates of different materials, the sidewalls of the component cannot maintain a straight or uniform form. This causes the subsequent material to be deposited in the part in a continuous mode. Therefore, these techniques cannot be used to accurately remove an atomic layer film at a time. Accordingly, it is highly desirable to develop techniques for controlling the etching of materials at atomic layer sizes. Figures 1 and 2 illustrate the problem of treating multiple layers of the underlying layer using prior art methods. In Fig. 1, a laminate comprising SiO 2 , Si 3 N 4 and Al 2 〇 3 layers on a substrate layer is treated with a hydrofluoric acid (HF) immersion liquid. Since HF etches different materials at different rates, the resulting laminate has unwanted jagged edges. Similarly, in Figure 2, HF etching is used to remove the oxide layer at the bottom of the contact hole, which penetrates high temperature oxides (Η Τ 0 ), ShN 4 , Si 〇 2, high temperature oxides (ΗΤΟ), and borophosphates. A film layer of a sulphate glass (BPSG) layer on the sputum substrate. Since the reaction rate of HF with each layer material is not uniform, the resulting contact holes will have, jagged edges, and wall structures. When the contact hole is covered with a barrier layer (for example, TaN), the construction of the uneven sentence is difficult to implement. A TaN barrier layer is added to prevent metal ions (such as copper) from diffusing into the dielectric of the wafer (Si02). Layer. TaN overlies the surface, and surface inconsistencies can make the coating non-uniform. These unevenness provides a way for the metal to penetrate into the coated layer, as shown in Figure 3. The effect of the non-uniform contact hole sidewalls is The wafer size shrinks more significantly. 1291721 (3) SUMMARY OF THE INVENTION The present invention proposes an atomic layer removal (ALR) and/or atomic layer father exchange (ALEX) to modify the substrate (eg, a semiconductor device or Wafer) Surface Methods and Systems The advantages of the present invention are obtained by the ALR and ALEX methods and the systems described herein in which the first solid compound molecular layer is uniformly removed from the substrate surface using methods. a single reactive gas that reacts with the first molecular layer on the surface to form a monolayer of an intermediate solid compound on the surface. After removing the first reactive gas and no longer contacting the substrate 'lead Into a second reactive gas, the second reactive gas reacts with the intermediate solid compound to form a volatile or volatile product. Thereafter, the second reactive gas and volatile or volatile products are removed. They are no longer in contact with the substrate' such that the first solid compound molecular layer remaining on its surface is one less layer than the original. Another embodiment of the invention utilizes an atomic layer exchange method to form a plurality of atomic layers on the substrate. a membrane in which at least one of the first solid compounds (having at least the first and second solid chemical species) is exposed to a first reactive gas. The first reactive gas has at least a first and a second Gaseous chemical species. A surface reaction occurs in which a first solid chemical species is exchanged with a first gaseous chemical species to form an intermediate solid compound. In another embodiment of this feature of the invention, the substrate is a second reaction a gas that reacts with an intermediate solid compound to produce one or more volatile products, leaving a first solid compound film with a molecular layer that is less than the initial layer One layer 0 1291721 (4) [Embodiment] The present invention proposes an atomic layer removal and atomic layer exchange method and system. Referring to the removal viewpoint of this method, the abbreviation "ALR" is usually used to represent atomic layer removal. This method is described using atomic layer exchange "ALEx", which is used to remove the solid compound surface layer from the substrate. In general, the present invention proposes an atomic layer removal and atomic layer exchange method and system. Wherein the substrate deposited or grown on the surface of the substrate is placed in a reactor or tank. The first reactive gas is introduced into the reactor or tank to react with the first layer of the membrane to One layer is converted to a single layer of solid compound. The reactor or tank is scrubbed to remove unreacted first reactant gas and/or any gas phase product. The second reactant gas is then directed to the reactor or tank. The reaction is carried out with a single layer of solid compound to form a gaseous compound which is removed from the reactor by a second scrubbing step. In one embodiment of the invention, the first reactant gas may additionally comprise an energy source to activate or facilitate the reaction of the first reactant gas with the first layer of the membrane. This energy source can be in the form of electromagnetic waves or some other energy transfer. Such electromagnetic waves include, but are not limited to, visible light, infrared light, ultraviolet light, microwave radiation, radio frequency rays, and the like. This ray can be supplied from a device (eg laser) in a continuous form or from a device (eg a lamp) in a non-continuous form (eg non-in-phase). The use of energy assists in the reaction of the first reactant gas with the first membrane. The invention can be used to remove a single atomic layer from the surface of the wafer. In addition, the atomic layer removal method is a continuous atomic layer removal method of its own termination. This method can be used to remove multiple compounds from the surface prior to subsequent deposition methods. The removed compound may comprise an intentionally or accidentally deposited oxide coating and a solid substrate -7-1291721 (5) material and/or a compound deposited in a layer deposited by chemical vapor deposition. The present invention can be used to reduce the thickness of the deposited conductive or dielectric film to achieve the desired final film thickness. These are only a small part of the application of the invention. Typically, the first reactive gas reacts with the top layer of the membrane to convert the top layer to a single layer solid compound. The first reactive gas is a chemical agent for converting the atomic layer of the first solid compound in the surface film or the upper layer of the substrate into a single layer of the intermediate product solid compound, which may further be combined with the second reactive gas reaction. The choice of the first reactive gas depends on the molecular composition of the membrane and the chemistry of the first solid compound. The single layer of the intermediate solid compound is then reacted with a second reactive gas to form a volatile or volatile product. This volatile or volatile product is removed by any suitable means known in the art (e.g., using an inert scrubber, using a vacuum pump, or using a combination of these two techniques). The foregoing steps of this embodiment of the ALR method can be described as a lower jaw! J: A (solid) + B (gas) - ♦ D (solid) (1) D (solid) + C (gas) - E (gas (2) where A (solid) is the first solid compound on the surface of the substrate, B (gas) is the first reactive gas, D (@body) is the intermediate solid compound, C (gas 1) It is the second reactive gas, and E (gas if) is a volatile or volatile product. Reaction steps (1) and (2) can be repeated until the desired number of A ([1 body) molecular layers is removed from the substrate. A more detailed embodiment of the method according to the invention is shown in Figure 4, a substrate 1 having a film 12 deposited or grown on a substrate surface, a reactor, a tank or some of which can control the order of exposure of the reactive gases In other environments. The film consists essentially of a plurality of atomic layers, in this example consisting of the first solidified -8-(6) 1291721 compound A (TM body _). In this example, the a (丨 solid) film thickness was 3 molecules. However, any thickness of film and/or other solid or deposited surface compound can be treated in accordance with the present invention. The number of molecules or atomic layers of the first solid compound present on the surface of the substrate can vary widely depending on the particular application. The first solid compound A (TM body) to be removed may include any type of film used in semiconductor processing, such as any gate dielectric or ceramic, metal oxide, tantalum oxide, commonly used in semiconductor devices, Metal aluminate, yttrium aluminate, metal silicate, metal nitride, niobium nitride, pure metal or other film. In theory, any surface can be treated by the method of the present invention as long as a suitable reactive gas is selected. The bottom view (I) shows that the mass 10 and the film 1 2 are in the first gas B). B is preferably a chemical agent for converting the atomic and/or molecular layer of the first solid compound A (solid) on the surface of the film or substrate into a single layer intermediate product solid compound D (solid). Preferably, the first reactant gas B ( m ^ ) reacted with the first solid compound A ((5) μ) is selected to form a single layer of intermediate solid compound D ( ® r,) on the surface of the substrate. The reaction of the first reactant gas B (gas) with the first solid compound A (solid) is preferably a reaction which is more thermally and/or motivated than the reaction of the monolayer intermediate solid compound D (solid). In other words, the reaction of the first reactant gas B (gas) with a clean A (solid) surface should form a uniform surface coated with a single layer of intermediate solid compound D (® (iv)). The so-called thermal tendency is According to the following formula, B (35) and D (tamping) reaction B (gas) + D (solid) - X (3) exotherm is substantially lower than reaction (1). The so-called power tendency refers to the reaction ( 3) The reaction rate is substantially slower than the reaction (1) -9-1291721 (7). Also in the case, the intermediate solid compound D ((5) body forms an obstacle to the extra table of a reactant gas B (g(4)) The gas phase molecule is in contact with the first solid compound A (the s layer underlying layer) such that only a single layer is formed. Examples of the compound as the first reactant gas B (gas) according to the present invention include, but are not limited to , ozone, hydrogen, alcohol, water, ammonia, etc.
如審視圖(II )和前面的討論所示,暴於第一種反應 物氣體B (氣體」將第一種固態化合物A (固體)分子的表層轉 化成單層中間產物固態化合物D ( ®㉟)。使用此技術標準 技巧(如:使用惰性滌氣氣體,使用真空幫浦或之類,或 使用一或多種這樣的技巧之組合),自反應器或槽抽除過 量的第一種反應物氣體B (氣體)。之後,底質暴於第二種 反應性氣體C (氣體),其將表面單層中間產物固態化合物 D ( ®體)轉化成揮發性或可揮發產物E ^㈢,此如審視圖 (III )所示者。以選擇第二種反應性氣體C (㉟使得其 與單層中間產物固態化合物D ( θ體)之反應實質上比C (氣 體〕與膜化合物Α (固體)之反應更爲放熱(熱力趨向)或以 實質上較快的反應速率進行(動力趨向)爲佳。換言之, 較佳情況中 C(氣體)+A(固體)—Y (4) 其中 Y是一些所不欲產物,實質上比反應(2 )放熱 來得少和/或反應(3 )的反應速率實質上比反應(4 )來 得快。此方式中,經單層中間產物固態化合物D ( ©體)塗 覆的表面以第二種反應性氣體c 處理,僅移除單層 中間產物固態化合物D ( ®)而不會有第二種反應性化合 -10- 1291721 (8) 物C (氣體)對位於下方的第一種固態化合物A (固體〗 外的攻擊。一個實例中,與經轉化的單層中間產物 合物D ( ®)反應的第二種反應性氣體C (氣體〕可以 國素的來源。可用於此方法之含有鹵素的化合物的 括,但不限於C1F3、NF3、HF和氯。 揮發性或可揮發的產物E (㉟M ,揮發或自表面 留下膜12,此處,在示於附圖4的實例中,含有 態化合物A ( @ 的分子層,比處理之前少了一層 以與前述用以移除第一種反應性氣體相同的方式自 移除/滌除揮發的產物E (氣體)和任何過量的第二 性氣體C (氣體)。重複前述步驟,可移除第一種固 物A ( ® fi )的第二、第三和額外層,直到有所欲層 底質上爲止。通常,第一和第二種反應性氣體分別 應器或槽中達足以與膜上的一個單一,或原子/分 反應的時間。第一和第二種反應性氣體的暴露時間 0· 5至60秒鐘,更特別是約1至2秒鐘。 本發明的另一實施例示於附圖5 A至5 D。此實 ,膜或晶圓表面由第一和第二種固體化合物(分 和B )構成,此如附圖5 A所示者。晶圓表面可含 或可僅含有矽及形成於其上的原有氧化物。此實例 生原子層交換,晶圓暴於氣態先質CD,此如附圖 不者。弟二個步驟(附圖5C所示者)中,表面反 種交換發生於晶圓上的頂層處。此實例中,表面反 層轉化成單層固態化合物AD ( ®體)。亦形成,,廢,,氣 物CB (氣體),其藉前述並示於附圖5]〇的方式自槽 層作額 固態化 是含有 例子包 揮發, 兩層固 。之後 反應器 種反應 態化合 數留在 引至反 子,層 通常約 施例中 和是 A 有膜, 中,發 5B所 應和物 應將頂 態化合 移除或 -11 - 1291721 (9) 滌除。此實施例中’步驟可以節錄成下列式: A B (固體)+ C! D (氣體)—A D (固體)+ C B (氣體) (5 ) 許多類型的氣態先質或反應物可用於本發明之方法, 並部分基於膜的化學組成而選擇。氣態先質的其他例子包 括,但不限於,臭氧、氨、水、氫、聯氨、醇、鹵素之類 。如前述者,原子層交換發生於氣相中的自由基或分子和 晶圓或膜或底質表面之間。可藉由數個參數(包括溫度、 脈衝時間、槽壓力、分子尺寸和反應性)控制這些氣態先 質通過晶圓表面之擴散,以避免多層原子交換。 本發明之方法的另一實施例更特定地示於附圖6 A至 6 I,其詳細列出後續步驟。此實例中進行原子層交換和原 子層移除。首先,根據下列式,實施原子層交換以修飾膜 表面的化性: ( CB (固體)+DE (氣體)一(固體)+DB (氣體)个 (6) 其中CB (固體)是第一種固態化合物。第一種反應 性氣體D E μ )被運送至反應器或槽,此如附圖6 B所示 者。之後,第一種反應性氣體DE ( a 被視情況活化而形 成一或多個第一種氣相自由基物種,此如附圖6C所示者 。如前述者,活化可藉各式各樣方式進行,如:藉溫度、 能量脈衝、電磁波之類。就一些第一種反應性氣體DE (氣 體)物種和反應條件而言,不須活化便能進行反應。原子 層與膜的頂層交換。之後,所得膜的表層是中間產物固態 化合物CE (固體),此如附圖6D所示者。交換之後,槽或 反應器經滌氣,自槽移除形成的第一種氣態副產物D B (氣 體)和任何未反應的第一種反應性氣體DE ( μ )或自由基 -12- 1291721 (10) 〇膜的頂層現被轉化成中間產物固態化合物C E (固體) 如附圖6 E所示者。 之後,進行原子層移除以移除膜的頂部中間產物 化合物(C E (丨固體))層’此如附圖6 F至61所示者。 例中,選擇第二種反應性氣體X Y (氣體),使得中間 固態化合物C E ( ®體)的表面膜層與第二種反應性氣體 (氣體)反應,且額外的第一種固態化合物C B ( ®)層 膜上之位於中間產物固態化合物c E (固體)單層下方的 不會與XY (氣體)反應。第二種反應性氣體XY (氣體) 情況地被活化(如··藉能量脈衝、溫度或電磁波或其 式)而形成一或多種第二種氣相自由基物種’此如 6 G所示者。此實例中,氣相中的反應產物是: C E (固體)+ X Y (氣體)—C X (氣體)+ E Y (氣體)八 ( 其中 CX (氣體)和EY (氣體)分別是第二和第三種 副產物。發生此反應,氣態副產物CX (氣體 > 和EY ( 形成及自槽滌除,此如附圖6H和61所示者。因此, 種固態化合物C B (固體)的膜被移除。可以視須要地多 複這些步驟以進一步移除原子層。 由自接觸孔移除原有氧化物會更瞭解本發明之方 於以前技術方法之處。如附圖2中所示者,藉反應性 蝕刻或一些相仿方法,於經澱積的底質材料之多層層 形成接觸孔,原有氧化物污染可藉由使底質和多層層 暴於氫氟酸溶液(HF浸液)而移除。如前文的討論 圖2中所示者,在反應性蝕刻浸液中,自外露的層表 除材料時,反應速率受限,蝕刻劑與不同層材料的反 ,此 固態 此實 產物 XY (在 層) 可視 他方 附圖 7) 氣態 氣體) 第一 次重 法優 離子 合物 合物 和附 面移 應速 -13- 1291721 (11) 率不同,因此,慣用的濕蝕刻之後,會得到不平整的側壁 。本發明的一個實例中,穿透多層層合材料形成的接觸孔 底物的氧化物層被移除且不會形成非平整孔洞。本發明用 以移除接觸孔底部的氧化物並維持側壁的平面性。因爲本 發明之方法,均勻地移除單一單層,無論各反應循環中的 表面化合物反應性如何皆然,此方法避免動力驅動的濕蝕 刻問題(其中,反應性較高的材料被蝕刻的速率比應性較 低的材料來得高)。此實例中,使用水作爲第一種反應物 氣體以在所有外露氧化物層表面上形成-0H終端的表面。 矽氧化物和金屬氧化物的介電膜具親水性且此反應容易進 行。之後引入HF作爲第二種反應物氣體及移除膜頂層( 水和SiF4或金屬氟化物形式)。通常,HF與矽氧化物或 金屬氧化物於無水存在時之反應非常緩慢。因此,移除第 一層時,反應會自身終結。可重複此程序直到移除接觸孔 底部處的氧化物爲止。因爲無論氧化物的化學本質如何, 各循環僅移除一層,所以能夠維持側壁的平面性。 如附圖7所示者,根據本發明的一個實施例的步驟是 (I )暴於氬氣及以UV活化而使非氧化物表面鈍化,(II )引入水蒸汽(第一種反應性氣體)或一些其他羥基離子 和/或羥基自由基來源(如:醇),以使介電物/氧化物 層的表面狀態改變和標準化,(III )槽/反應器經滌氣 以移除水蒸汽,(IV )引入HF蒸汽以移除用於介電物和 氧化物的分子表面層,及(V )視情況需要地重複步驟( 11 ) - ( V )以移除額外層。如前述者,相較於以前技術 HF蝕刻法(其中,底質和介電層浸於HF水浴中),此 1291721 (12) 技巧較有利。本發明之方法防止所得接觸孔和由數種不同 介電或氧化物層構成的其他表面具非均勻表面,這是因爲 HF蒸汽與羥基化的表面之反應僅會移除羥基(水蒸汽形 式)及其相關原子(如:就、si〇2和ai2o3而言,分別是 SiF4或A1F3 )之故。一旦發生氧化物-HF反應,留下的非 羥基化表面受到HF攻擊的反應性不及具附著羥基的表面 〇 疏水性表面(如:S i )之鈍化,實質上阻礙這些區域 的HF攻擊,這是因爲水蒸汽實質上不會吸附於鈍化區域 (無氫鍵)並因此,HF蒸汽反應不強烈之故。 附圖8節錄根據本發明之ALR和/或ALEx方法的 步驟。含有一或多個膜和/或固體化合物層的底質置於提 供控制大氣壓至表面的反應器、槽或其他系統冲3 00底 質上的任何疏水表面可以視情況地藉,如:底質暴於氫氣 或一些會與表面氧化物澱積物反應的其他化合物,而鈍化 ,提供不會吸附水蒸汽的表面3 02。此鈍化步驟3 02可藉 由使氣體混合物於反應性環境中對底質上方照射足以破壞 氣相氫之分子鍵的紫外光而活化,此處可以使用任何鈍化 劑。鈍化步驟3 02亦包括對反應器或槽滌氣以移除鈍化用 氣體。之後,第一種反應性氣體引至底質上方。如前述者 ,第一種反應性氣體可以是氧來源,如:臭氧、水或醇。 或者,第一種反應性化合物可以是氫或氨或一些類似的化 合物。作爲第一種反應性氣體的化合物之確實選擇以所欲 反應中間產物和其與下文討論之選用的第二種反應性氣體 之反應性爲基礎。嫻於此技術者已知的氣體-和表面·相化 •15- (13) 1291721 學特徵及慣用實驗應足以視欲處理的底質地決定第一和第 二種反應性氣體的適當組合。 第一種反應性化合物可視情況地經前文所討論的電磁 波或一些其他能量輸入(如:射線、傳導或對流加熱)而 活化。視反應條件(包括,但不限於,底質和膜化學、溫 度之類)而定,一些反應性氣體可以”自身活化’’或自發地 與膜表面或感興趣的標的物反應。無論反應是否被活化, 第一種反應性氣體與底質表面的表面單層反應,形成中間 產物固態化合物的分子層3 1 0。較佳情況中,選擇第一種 反應性氣體,使得第一種反應性氣體與底質或膜表面之反 應完全且起始固態化合物的分子表層轉化成中間產物固態 化合物。此第一種反應性氣體未進一步與中間產物固態化 合物反應,第一種反應性氣體也不會擴散通過中間產物固 態化合物的表面單層而與位於氣體一固體表面下方的起始 固態化合物的額外層反應。如所示者,這些P艮制可爲熱力 或動力限制。例如,第一種反應性氣體與中間產物固態化 合物之反應的動態趨向程度不及第一種反應性氣體與起始 固態化合物之反應。或者,第一種反應性氣體與起始固態 化合物的反應速率比第一種反應性氣體與中間產物固態化 合物之反應或第一種反應性氣體與起始固態化合物(位於 中間產物固態化合物表面單層下方)之反應快得多。第一 種反應性氣體與起始固態化合物反應足夠時間之後,槽或 反應器經滌氣以自與底質接觸者潔淨任何殘留氣體3 1 2。 底質暴於第一種反應性氣體的時間長短視反應動力和熱力 而定,但通常是1至60秒鐘。極具反應性的氣體的反應 -16- (14) 1291721 時間較短,可爲0.5秒鐘至1秒鐘。但通常第一種反應性 氣體的暴露時間約1至2秒鐘。 之後將第二種反應性氣體引至槽或反應器中3 1 4。選 擇第二種反應性氣體,使得其與步驟3 ] 0中形成的中間產 物固態化合物之反應有效完成,此藉由將中間產物固態化 合物轉化成蒸汽壓較高的化合物或可自底質表面蒸發或容 易藉由一些能量輸入(如:加熱)而蒸發的化合物。-一個 實施例中,第二種反應性氣體是含鹵素的氧化劑化合物。 選用的化合物包括,但不限於,c 1F 3 ' N F 3、H F和氯氣。 但嫻於此技術者有能力基於此處所述者地選擇適用於本發 明的其他化合物。 至於用以添加第一種反應性氣體,第二種反應性氣體 可視情況地經一些能量輸入(如前述者)而活化_3 1 6。視 選用的第一種反應性氣體、中間產物固態化合物和第二種 反應性氣體而定,須要或不須活化。第二種反應性氣體暴 於底質的反應時間約0.5至60秒鐘,典型反應時間約1 至2秒鐘。第二種反應性氣體與中間產物固態化合物反應 ,將中間產物固態化合物轉化成揮發性或可揮發的產物, 此產物於形成之後立刻進入氣相或者自表面蒸發3 2 0。較 佳情況中,選擇第二種反應性氣體,使得實質上僅與在步 驟3 1 0中形成單槽的中間產物固態化合物反應。第二種反 應性氣體與起始固態化合物的反應限制可爲動力或熱力限 制,此如前述關於第一種反應性氣體中所述者。較佳情況 中,於低能量情況下,第二種反應性氣體與起始固態化合 物之反應實質上比第二種反應性氣體與中間產物固態化合 -17- 1291721 (15) 物之反應較爲有利。或者,第二種反應性氣體與中間產物 固態化合物的反應速率實質上高於第二種反應性氣體與起 始固態化合物。 如果揮發性或可揮發產物於形成之後,實質上未自底 質表面蒸發,由射線、熱或一些其他已知技巧提供能量或 一些蒸發誘導,使得揮發性或可揮發產物釋放至氣相322 。之後自槽滌除氣相蒸發的產物和任何過量第二種反應性 氣體324。用以自底質移除額外分子層,可多次重複步驟 3 04-324。通常,選用的鈍化步驟僅用於程序之初。可於 未再度鈍化的情況下移除後續層。 另一實施例中,本發明之方法可用以使超薄膜澱積於 底質上並嚴密控制尺寸。例如,以前技術澱積技巧在澱積 超薄(如:厚.'度3埃)介電膜時有限制.。使用本發明之方 法,厚度如1 〇埃的介電膜可先澱積於底質表面上。之後 ,可以使用前述原子層移除法,自底質移除介電膜層。對 於欲移除的層數沒有限制。因此,如果希望介電膜厚3埃 ,則可重複前述步驟地移除7埃厚的介電膜,在底質表面 上留下3埃膜。 本發明的原子層移除和原子層交換法應用廣泛。例如 ,本發明可用以蝕刻金屬和介電物,形成攝影刻印遮蔽物 ,改善液晶顯示器的解析度,及其他應用。此外,本發明 的‘原子層移除可用以在形成閘極之前,減低最終膜厚度和 /或移除所不欲表面糙度。可藉本發明的原子層交換和低 溫ALD高k介電法調整矽-高k介電介面。 1291721 (16) 理論例 下列理論例說明本發明之方法和系統。這些實例僅作 說明之用,不欲對本發明之範圍造成任何限制。 實例1 一個說明例中,使用氮化鈦(TiN )層,藉本發明之 方法和系統移除至所欲厚度。氮化鈦(TiN )是一種用於 閘極的較佳材料。欲使TiN薄膜澱積於底質上,可以先使 相當厚的TiN膜澱積於閘極介電物表面上。根據本發明的 此實施例,引入臭氧氣體,以使TiN膜原子頂層轉化成單 層一氧化欽(Ti〇2) 。之後’此固態Ti〇2層進一*步與氯 化氫(HF )蒸汽反應,形成氣態氟化鈦(TiF4 )和水,其 自反應器移出。方法的各次循環可自底質移除一個TiN原 子層。藉由重複此方法,得到位於矽底質上之具有所欲厚 度的TiN膜。對於藉本發明之原子層移除法可移除的TiN 層數沒有限制。 實例2 另一實例中’乙醇可以作爲第一種反應物氣體,與已 灑積於底質上的二氧化矽(s丨〇 2 )膜反應。於底質頂面上 形成單層氫氧化矽(SiOH )。氟化氫(HF )可以作爲第 一種反應物氣體’與固態氫氧化矽反應而形成氣態氟化矽 (S1F4 )和水,自反應器將其移除。 已經以本發明的特定實施例和實例的前述描述作說明 和fe述’雖然以某些實例說明本發明,本發明不因此受限 -19 - (17) 1291721 。不欲將本發明限於所提出的確實形式,由前文所述者, 顯然可作出許多修飾、實施例和變化。希望本發明之範圍 涵蓋此處所提出的一般範圍和所附申請專利範圍及它們的 對等物。 【圖式簡單說明】 由本發明的詳細描述和下文所附申請專利範圍及參考 附圖,會明瞭本發明的其他目的和優點,附圖中: 附圖1所示者是以前技術的氫氟酸蝕刻處理對於不同 化學品層合物之影響。 附圖2所示者是以前技術的氫氟酸蝕刻處理對於穿透 不同固體材料的接觸孔壁之影響。 附圖3所示者是以聯技術的酸触技巧得'到的不均勻接 觸孔壁對於TaN屏障層之影響及所得金屬離子擴散進入 Si02和Si3N4層對於半導體晶圓之影響。 附圖4所不者是根據本發明的一個實施例之原子層移 除法的步驟。 附圖5A至5D所示者是根據本發明的另一實施例之 原子層移除法的步驟。 附圖6A至61所示者是根據本發明的另一實施例之原 子層交換和之後的原子層移除之步驟。 附圖7所示者是根據本發明的原子層移除技巧術的步 驟,其中,藉由在暴於HF蒸汽之前,形成H·終端的疏水 表面,使得矽底質被鈍化至不受HF影響,以均勻地自晶 圓層合物中的各層移除表面單層分子。 -20- (18) 1291721 附圖8所不者是流程圖,草述根據本發明的一個實施 例之原子層移除法的步驟。 元件對照表 1 〇 :底質 12 :膜 1 4 :第一種固態化合物A (固體)的表層 3 00 :含有一或多個膜和/或固態化合物層的底質表 面 3 02 :鈍化疏水表面(視情況) 3 04 :引入第一種反應性氣體 3 06 :活化第一種反應性氣體(視情況) 3 1 0 :第一種反應性氣體和/或其經活化組份反應, 在表面上形成中間產物固態化合物單層 3 1 2 :滌除第一種反應性氣體 3 14 :引入第二種反應性氣體 3 1 6 :活化第二種反應性氣體(視情況) 3 20 :第二種反應性氣體和/或其經活化的組份與表 面上的中間產物固態化合物反應,形成可揮發產物 322 :使可揮發產物揮發 3 24 :滌除第二種反應性氣體和可揮發產物As shown in the review view (II) and the previous discussion, the first reactant gas B (gas) converts the surface layer of the first solid compound A (solid) molecule into a single layer intermediate product solid compound D (®35). Use this technique to standardize techniques (eg, using an inert scrubbing gas, using a vacuum pump or the like, or using a combination of one or more of these techniques) to remove excess first reactant from the reactor or tank Gas B (gas). Thereafter, the substrate is exposed to a second reactive gas C (gas), which converts the surface monolayer intermediate product D (TM) into a volatile or volatile product E ^ (3). As shown in the review (III), the second reactive gas C (35 is selected such that it reacts with the monolayer intermediate solid compound D (the θ body) substantially more than C (gas) and the membrane compound Α (solid The reaction is more exothermic (thermal tendency) or at a substantially faster reaction rate (power tendency). In other words, preferably C (gas) + A (solid) - Y (4) where Y is Some unwanted products, substantially more than the reaction (2) The exotherm is less and/or the reaction rate of the reaction (3) is substantially faster than the reaction (4). In this manner, the surface coated with the monolayer intermediate solid compound D (© body) is in the second Reactive gas c treatment, removing only the single layer intermediate product solid compound D (®) without a second reactive compound -10- 1291721 (8) C (gas) pair below the first solid compound A (solid) attack. In one example, the second reactive gas C (gas) that reacts with the converted monolayer intermediate D ( ® ) can be a source of a national element. Halogen compounds include, but are not limited to, C1F3, NF3, HF, and chlorine. Volatile or volatile product E (35M, volatilized or leaving film 12 from the surface, here, in the example shown in Figure 4, The molecular layer containing the compound A ( @, one layer less than before the treatment, the self-removed/depleted volatile product E (gas) and any excess second in the same manner as described above for the removal of the first reactive gas Sex gas C (gas). Repeat the above steps to remove the first The second, third and additional layers of the solid A ( ® fi ) until the desired layer is on the substrate. Typically, the first and second reactive gases are respectively sufficient to form on the membrane a single, or atomic/minute reaction time. The exposure time of the first and second reactive gases is from 0.5 to 60 seconds, more particularly from about 1 to 2 seconds. Another embodiment of the invention is shown in Figures 5A to 5D. In this case, the film or wafer surface is composed of first and second solid compounds (minutes and B), as shown in Figure 5 A. The wafer surface may or may only Containing niobium and the original oxide formed on it. This example is atomic layer exchange, and the wafer is exposed to a gaseous precursor CD, as shown in the drawing. In two steps (shown in Figure 5C), surface inversion exchange occurs at the top layer on the wafer. In this example, the surface reverse layer is converted into a single layer solid compound AD (TM). Also formed, waste, gas CB (gas), which is solidified from the trough layer by the means described above and shown in Figure 5], contains an example package volatilization, two layers solid. After that, the reactive state of the reactor species is left to the emitter, and the layer is usually about the same as the A film. In the case of the 5B, the top state should be removed or -11 - 1291721 (9) eliminate. The 'steps in this example can be summarized as follows: AB (solid) + C! D (gas) - AD (solid) + CB (gas) (5) Many types of gaseous precursors or reactants can be used in the present invention. The method is selected based in part on the chemical composition of the membrane. Other examples of gaseous precursors include, but are not limited to, ozone, ammonia, water, hydrogen, hydrazine, alcohols, halogens, and the like. As mentioned above, atomic layer exchange occurs between free radicals or molecules in the gas phase and the surface of the wafer or film or substrate. The diffusion of these gaseous precursors through the wafer surface can be controlled by several parameters including temperature, pulse time, bath pressure, molecular size and reactivity to avoid multilayer atom exchange. Another embodiment of the method of the present invention is more particularly shown in Figures 6A through 6I, which detail the subsequent steps. Atomic layer swapping and atomic layer removal are performed in this example. First, atomic layer exchange is performed according to the following formula to modify the chemical properties of the membrane surface: (CB (solid) + DE (gas) - (solid) + DB (gas) (6) where CB (solid) is the first The solid compound. The first reactive gas DE μ ) is transported to the reactor or tank as shown in Figure 6B. Thereafter, the first reactive gas DE (a is activated as appropriate to form one or more first gas phase radical species, as shown in Figure 6C. As described above, activation can be varied. The method is carried out, such as by temperature, energy pulse, electromagnetic wave, etc. For some of the first reactive gas DE (gas) species and reaction conditions, the reaction can be carried out without activation. The atomic layer is exchanged with the top layer of the membrane. Thereafter, the surface layer of the resulting film is the intermediate solid compound CE (solid) as shown in Fig. 6D. After the exchange, the tank or reactor is scrubbed and the first gaseous by-product DB formed by the removal from the tank ( Gas) and any unreacted first reactive gas DE (μ) or free radical -12-1291721 (10) The top layer of the ruthenium film is now converted to an intermediate solid compound CE (solid) as shown in Figure 6 E Thereafter, atomic layer removal is performed to remove the top intermediate compound (CE (tank) layer) of the film as shown in Figures 6F to 61. In the example, the second reactive gas XY is selected. (gas), making the intermediate solid compound CE ( The surface layer of the ® body reacts with the second reactive gas (gas), and the additional first solid compound CB ( ® ) film is located below the monolayer solid compound c E (solid) single layer Will react with XY (gas). The second reactive gas XY (gas) is activated (eg, by energy pulse, temperature or electromagnetic wave or its formula) to form one or more second gas phase radical species. 'This is shown as 6 G. In this example, the reaction product in the gas phase is: CE (solid) + XY (gas) - CX (gas) + EY (gas) eight (where CX (gas) and EY ( The gases are the second and third by-products respectively. This reaction occurs, gaseous by-products CX (gas > and EY (formed and removed from the tank, as shown in Figures 6H and 61. Therefore, a solid compound) The CB (solid) film is removed. These steps can be repeated as needed to further remove the atomic layer. The removal of the original oxide from the contact hole will give a better understanding of the present invention in the prior art methods. Figure 2, by reactive etching or some similar The method forms contact holes in the multilayered layer of the deposited substrate material, and the original oxide contamination can be removed by causing the substrate and the multilayer layer to be exposed to the hydrofluoric acid solution (HF immersion liquid). As discussed above. As shown in Fig. 2, in the reactive etching immersion liquid, when the material is removed from the exposed layer, the reaction rate is limited, and the etchant is opposite to the material of the different layer, and the solid product XY (in the layer) can be seen elsewhere. Fig. 7) Gaseous gas) The first heavy-grained ionic compound and the surface-shifting rate of -13 to 1291721 (11) are different, and therefore, after conventional wet etching, uneven sidewalls are obtained. In one embodiment of the invention, the oxide layer of the contact hole substrate formed through the multilayer laminate is removed and does not form non-flat holes. The present invention is used to remove oxides at the bottom of the contact holes and maintain the planarity of the sidewalls. Because of the method of the present invention, the uniform removal of a single monolayer, regardless of the reactivity of the surface compounds in each reaction cycle, avoids the problem of power driven wet etching (where the more reactive material is etched) Higher than the lower material). In this example, water was used as the first reactant gas to form a -0H terminal surface on the surface of all exposed oxide layers. The dielectric film of cerium oxide and metal oxide is hydrophilic and the reaction is easy to carry out. HF is then introduced as the second reactant gas and the top layer of the membrane (water and SiF4 or metal fluoride form) is removed. Generally, the reaction of HF with cerium oxide or metal oxide in the absence of water is very slow. Therefore, when the first layer is removed, the reaction ends itself. This procedure can be repeated until the oxide at the bottom of the contact hole is removed. Since each cycle removes only one layer regardless of the chemical nature of the oxide, the planarity of the sidewall can be maintained. As shown in Figure 7, the steps according to one embodiment of the present invention are (I) argon gas and UV activation to passivate the non-oxide surface, (II) introduction of water vapor (the first reactive gas) Or some other source of hydroxyl ions and/or hydroxyl radicals (eg alcohol) to change and normalize the surface state of the dielectric/oxide layer, (III) tank/reactor scrubbed to remove water vapor (IV) introducing HF vapor to remove molecular surface layers for dielectrics and oxides, and (V) repeating steps (11)-(V) as needed to remove additional layers. As previously mentioned, this 1291721 (12) technique is advantageous over prior art HF etching methods in which the substrate and dielectric layers are immersed in an HF water bath. The method of the present invention prevents the resulting contact holes and other surfaces composed of several different dielectric or oxide layers from having a non-uniform surface because the reaction of the HF vapor with the hydroxylated surface only removes the hydroxyl groups (in the form of water vapor) And its related atoms (such as: Si, Si2 and A1F3, respectively, si〇2 and Ai2o3). Once the oxide-HF reaction occurs, the remaining non-hydroxylated surface is less reactive by HF attack than the surface of the hydrophobic surface (eg, S i ) with attached hydroxyl groups, which substantially hinders HF attack in these areas. This is because the water vapor is not substantially adsorbed in the passivation region (no hydrogen bonding) and, therefore, the HF vapor reaction is not strong. Figure 8 illustrates the steps of the ALR and/or ALEX method in accordance with the present invention. The substrate containing one or more membrane and/or solid compound layers is placed on any hydrophobic surface that provides a reactor, tank or other system for controlling atmospheric pressure to the surface. Optionally, borrowing, such as: substrate It is violently exposed to hydrogen or some other compound that will react with the surface oxide deposits, and is passivated to provide a surface that does not adsorb water vapor. This passivation step 312 can be activated by irradiating the gas mixture in a reactive environment above the substrate with ultraviolet light sufficient to destroy the molecular bonds of the gas phase hydrogen, any passivating agent can be used herein. Passivation step 302 also includes scrubbing the reactor or tank to remove the passivating gas. Thereafter, the first reactive gas is introduced above the substrate. As mentioned above, the first reactive gas may be a source of oxygen such as ozone, water or alcohol. Alternatively, the first reactive compound can be hydrogen or ammonia or some similar compound. The choice of the compound as the first reactive gas is based on the reactivity of the desired intermediate and its reactivity with the second reactive gas selected below. Gas- and surface-phase correlations known to those skilled in the art • 15-(13) 1291721 The characteristics and routine experimentation should be sufficient to determine the appropriate combination of the first and second reactive gases depending on the nature of the substrate to be treated. The first reactive compound may optionally be activated by electromagnetic waves or some other energy input (e.g., ray, conduction or convection heating) as discussed above. Depending on the reaction conditions (including, but not limited to, substrate and membrane chemistry, temperature, etc.), some reactive gases may "self-activate" or spontaneously react with the surface of the membrane or the subject of interest. Activated, the first reactive gas reacts with the surface monolayer of the surface of the substrate to form a molecular layer 310 of the intermediate solid compound. Preferably, the first reactive gas is selected such that the first reactivity The reaction of the gas with the substrate or the surface of the membrane is complete and the molecular surface layer of the starting solid compound is converted into an intermediate solid compound. This first reactive gas is not further reacted with the intermediate solid compound, and the first reactive gas is not Diffusion by the surface monolayer of the intermediate solid compound reacts with an additional layer of the starting solid compound located below the surface of the gas-solid. As shown, these P-systems can be thermally or mechanically limited. For example, the first reaction The reaction between the gas and the solid compound of the intermediate product is less dynamic than the first reactive gas and the starting solid compound. Or the reaction rate of the first reactive gas with the starting solid compound is higher than the reaction of the first reactive gas with the intermediate solid compound or the first reactive gas with the starting solid compound (in the intermediate solid state The reaction of the compound surface below the monolayer is much faster. After the first reactive gas has reacted with the starting solid compound for a sufficient period of time, the tank or reactor is scrubbed to clean any residual gas 3 1 2 from contact with the substrate. The length of time during which the substrate is exposed to the first reactive gas depends on the reaction power and heat, but is usually 1 to 60 seconds. The reaction of a highly reactive gas - 16 - (14) 1291721 The time is short. It is from 0.5 seconds to 1 second, but usually the exposure time of the first reactive gas is about 1 to 2 seconds. Then the second reactive gas is introduced into the tank or reactor 3 1 4 . Reactive gas, such that it is efficiently reacted with the intermediate solid compound formed in step 3] 0, by converting the intermediate solid compound into a compound having a higher vapor pressure or steaming from the surface of the substrate A compound that evaporates by some energy input (eg, heating). - In one embodiment, the second reactive gas is a halogen-containing oxidant compound. Selected compounds include, but are not limited to, c 1F 3 ' NF 3, HF and chlorine. However, those skilled in the art have the ability to select other compounds suitable for use in the present invention based on the ones described herein. As for the addition of the first reactive gas, the second reactive gas may be used as appropriate. The ground is activated by some energy input (such as the above) to activate _3 16 . Depending on the first reactive gas selected, the intermediate solid compound and the second reactive gas, activation or activation is required. The reaction time of the reactive gas to the substrate is about 0.5 to 60 seconds, and the typical reaction time is about 1 to 2 seconds. The second reactive gas reacts with the intermediate solid compound to convert the intermediate solid compound into a volatile or A volatile product which enters the gas phase immediately after formation or evaporates from the surface. In the preferred case, the second reactive gas is selected such that it substantially reacts only with the intermediate solid compound which forms a single tank in step 310. The reaction limit of the second reactive gas with the starting solid compound may be a power or thermal limitation as described above for the first reactive gas. Preferably, in the case of low energy, the reaction of the second reactive gas with the starting solid compound is substantially more than the reaction of the second reactive gas with the intermediate solid-state compound -17-1291721 (15). advantageous. Alternatively, the rate of reaction of the second reactive gas with the intermediate solid compound is substantially higher than the second reactive gas and the starting solid compound. If the volatile or volatile product, after formation, does not substantially evaporate from the surface of the substrate, energy or some evaporation induction is provided by radiation, heat or some other known technique to cause the volatile or volatile product to be released to the gas phase 322. The vapor phase vaporized product and any excess second reactive gas 324 are then removed from the tank. To remove additional molecular layers from the substrate, repeat steps 3 04-324 as many times. Typically, the passivation step chosen is for the beginning of the program. Subsequent layers can be removed without re-passivation. In another embodiment, the method of the present invention can be used to deposit ultra-thin films on the substrate and to tightly control the size. For example, prior art deposition techniques have limitations in depositing ultrathin (e.g., thick 'degree 3 angstrom) dielectric films. Using the method of the present invention, a dielectric film having a thickness of, for example, 1 Å can be deposited on the surface of the substrate. Thereafter, the dielectric film layer can be removed from the substrate using the aforementioned atomic layer removal method. There is no limit to the number of layers to be removed. Therefore, if it is desired that the dielectric film has a thickness of 3 angstroms, the dielectric film of 7 angstroms thick can be removed by repeating the foregoing steps, leaving a film of 3 angstroms on the surface of the substrate. The atomic layer removal and atomic layer exchange methods of the present invention are widely used. For example, the present invention can be used to etch metals and dielectrics to form photographic engraved masks, to improve the resolution of liquid crystal displays, and other applications. Furthermore, the 'atomic layer removal of the present invention can be used to reduce the final film thickness and/or remove unwanted surface roughness prior to forming the gate. The 矽-high-k dielectric interface can be adjusted by the atomic layer exchange of the present invention and the low temperature ALD high-k dielectric method. 1291721 (16) Theoretical Examples The following theoretical examples illustrate the method and system of the present invention. These examples are for illustrative purposes only and are not intended to limit the scope of the invention. EXAMPLE 1 In one illustrative example, a titanium nitride (TiN) layer was used, which was removed to the desired thickness by the method and system of the present invention. Titanium nitride (TiN) is a preferred material for the gate. To deposit a TiN film on the substrate, a relatively thick TiN film can be deposited on the surface of the gate dielectric. According to this embodiment of the invention, ozone gas is introduced to convert the top layer of the TiN film atom into a single layer of oxidized tetraoxide (Ti〇2). This solid Ti 2 layer then reacts with hydrogen chloride (HF) vapor to form gaseous titanium fluoride (TiF4) and water, which are removed from the reactor. Each cycle of the method removes a TiN atomic layer from the substrate. By repeating this method, a TiN film having a desired thickness on the base of the crucible is obtained. There is no limitation on the number of TiN layers that can be removed by the atomic layer removal method of the present invention. Example 2 In another example, 'ethanol can be used as the first reactant gas to react with a cerium oxide (s 丨〇 2 ) membrane that has been deposited on the substrate. A single layer of barium hydroxide (SiOH) is formed on the top surface of the substrate. Hydrogen fluoride (HF) can be reacted with solid ytterbium hydroxide as the first reactant gas to form gaseous cesium fluoride (S1F4) and water, which are removed from the reactor. The foregoing description of specific embodiments and examples of the invention has been described and described in the claims It is apparent that various modifications, embodiments, and changes may be made in the present invention. It is intended that the scope of the invention should be BRIEF DESCRIPTION OF THE DRAWINGS Other objects and advantages of the present invention will become apparent from the Detailed Description of the invention and the appended claims. The effect of etching treatment on different chemical laminates. Figure 2 shows the effect of prior art hydrofluoric acid etching treatments on the walls of contact holes that penetrate different solid materials. Figure 3 shows the effect of the uneven contact hole wall on the TaN barrier layer and the effect of the diffusion of the resulting metal ions into the Si02 and Si3N4 layers on the semiconductor wafer. Figure 4 is a schematic illustration of the steps of atomic layer removal in accordance with one embodiment of the present invention. 5A to 5D are steps of an atomic layer removal method according to another embodiment of the present invention. Figures 6A through 61 show the steps of atomic layer exchange and subsequent atomic layer removal in accordance with another embodiment of the present invention. Figure 7 shows the steps of the atomic layer removal technique according to the present invention, wherein the hydrophobic surface of the H. terminal is formed before being immersed in HF vapor, so that the underlying material is passivated to be unaffected by HF. To uniformly remove surface monolayer molecules from each layer in the wafer laminate. -20-(18) 1291721 Figure 8 is a flowchart showing the steps of the atomic layer removal method according to an embodiment of the present invention. Component Comparison Table 1 〇: Substrate 12: Membrane 14: Surface layer of the first solid compound A (solid) 3 00: Substrate surface containing one or more layers of film and/or solid compound 3 02 : Passivated hydrophobic surface (as appropriate) 3 04: introduction of the first reactive gas 3 06 : activation of the first reactive gas (as appropriate) 3 1 0 : reaction of the first reactive gas and / or its activated component, on the surface Forming an intermediate product solid compound monolayer 3 1 2 : purifying the first reactive gas 3 14 : introducing a second reactive gas 3 1 6 : activating the second reactive gas (as appropriate) 3 20 : second The reactive gas and/or its activated component reacts with the intermediate solid compound on the surface to form a volatile product 322: volatilizes the volatile product 3 24 : purifies the second reactive gas and the volatile product