玖、發明說明： 【與相關之專利申請案交互為蒼考資料】 本專利申請案主張於2002年6月28日提申的美國臨時 申請案序列編號60/392,057以及於2002年6月14日提申的英 5 國專利申請案編號0213708.1與於2002年6月18日提申的英 國專利申請案編號0213888.1之優先權。 I：發明所屬之技術領域3 發明領域 本發明係關於具有一介電常數(k)低於大約2.5之孔隙 10 性介電薄膜。過去多年來持續在研發製造該等供用於(特別 是）半導體裝置之具有低介電常數介電材料，俾與日益縮小 尺寸之裝置架構相容。依目前所認定使一具通用絕緣體達 到k值低於大約2.5，該等材料會無可避免地具有某種程度 之孔隙性。此孔隙性於整合上會呈現重大問題，特別是以 15 導孔或互連模組形成之介電層，其係由於在蝕刻時，若孔 隙實際互連或貫穿表面，則該以蝕刻形成之面板將會粗糙 且可能被穿透。 於目前典型之架構中，銅被沉積入這些被蝕刻之溝槽 及導孔中，且由於銅將容易地擴散至介電材料，因此介電 20 材料必須包裝一隔離擴散物。一個理想是一種具有隔離特 性之絕緣體，但目前解決方案則是倚賴分隔沉積層。 【先前技術3 發明背景 沉積這些隔離層傳統上是使用物理氣相沉積技術，但 5 1223331 這些技術無法提供足夠之隔離覆層均一度’因而使用化學 氣相沉積(CVD)技術，例如：金屬有機CVD、金屬_化物 CVD及原子層CVD。由於CVD技術能夠產生接近1 〇〇%均一 度，前導體及反應物皆能穿透孔隙介電物。此效應被顯示 5 於第 1 圖（資料來源：IMEC at ARMM 2001Conference)及第2 圖（資料來源：Passemard et al; “intergration issue of low k and ULK materials and damascene structure” at ARMM 2001Conference)。於二者之情形中，可觀察到面板顯現“模 糊（fuzzy)”，此係意指CVD前導物已被吸附至孔隙性介電 10層，使介電及隔離層之間產生一個無分界隔離。 EP-A-1195801描述之方法，據信是可增加面板的孔隙 度，且其係藉由提供一保護層或密封層來產生面板封孔。 該方法暗示這種密封層可藉由一種包含氧氣及氮氣之電漿 來形成，但其並沒有實際描述該方法。添加額外之材料至 15高的高寬比之導孔中是不為所欲的，且可能會增加導孔中 銅的電阻。是否如該申請案所暗示會於密封層表面維持局 部低k值，這並不清楚。 【發明内容3 發明概要 20 由本發明之一方面而言，本發明係包含一種具有一介 電常數(k)低於大約2 · 5及一碳含量不低於丨〇 %之孔隙性介 電薄膜，其包含一導孔或其他居於導孔内的蝕刻形成物， 其中e亥居於導孔或形成物之外露表面或由薄膜所構成之表 面是實質為非孔隙性。 6 可被瞭解的是此方法與EP-A-1195801完全相反，該案 之介電製程會增高居於表面層之局部孔隙度，且此困擾僅 可藉由添加一額外的密封層來予以克服。 尚需注意的是密封居於結構頂部及底部之孔隙性表 5 面，較諸密封該等與送入反應物流平行之側面而言，被認 為是較為容易的。 於一較佳具體例中，其外露表面或諸等表面是以一層 較諸整片薄膜而言為碳空乏薄膜製成。額外地或可選擇性 地，該外露表面或諸等表面可以一密度高於整片薄膜材料 10 之層體製成。於一特佳具體例中，該外露表面或諸等表面 可以一層較諸整片薄膜而言為碳空乏薄膜製成。額外地或 可選擇性地，該外露表面或諸等表面可以一實質以矽-矽鍵 結構成之層體製成，且這些鍵結可以被形成於三價矽分子 之間。其他的機構目前尚不明瞭，但其係存在於該外露表 15 面或諸等表面。 於每一個此等情形中可被瞭解的是，構成外露表面或 諸等表面之層體的製成是藉由改質蝕刻介電材料，而非藉 由進一步沉積。 整片薄膜較佳地是以一種SiCOH材料製成。 20 表面形成層體或諸等層體可以含氮或氳之電漿處理蝕 刻表面或諸等表面來予以製成，此製程與另一製程(例如： 光阻去除）可能會至少部分地同步進行。 本發明進一步還包含一覆蓋外露表面或諸等表面之阻 障層，於此種情形之阻障層沒有穿透外露表面或諸等表 7 面。該阻障層較佳是藉由化學氣相沉積(CVD)來予以沉積。 於本發明更進一步之方面係包含一種於一半導體裝置 中製造一種互連模組層體之方法，該方法包含： a:於一基材上沉積一低K值孔隙性介電薄膜； 5 b:沉積一光阻； c:於光阻上形成圖型，俾以界定蝕刻孔洞； d··於介電層上，經由孔洞來蝕刻導孔或蝕刻形成物；及 e:去除光阻，其中去除該光阻是以一種氮氣或一種惰 性氣體或此等之組合，及氫氣電漿或氮氣或一種惰性氣體 10 或此等之組合，及氧氣電漿，且導孔或蝕刻形成物所在之 外露表面亦於同時曝露電漿，使該等構成外露表面之表層 敏密。 該方法亦包含於該緻密外露表面沉積一阻障層。此阻 障層可藉由化學氣相沉積法來予以沉積。 15 較佳地，氮氣或惰性氣體多於氫氣或氧氣。因此，較 佳之氮氣(N2) ··氫氣(H2)比例是5 : 1。 該基材於光阻去除期間可以是射頻偏壓。 於可選擇性之方法中，可於蝕刻步驟進行緻密，且其 可藉由非氧化電漿製程（例如：當低k值材料之特性為有機 20 物時）來予以製造。 雖然本發明已被界定如上文所述，可被瞭解的是本發 明係包含由上文提出或下文所描述之特點所構成之任何一 種創作組合。 圖式簡單說明 8 本發明可以各種不同的方式實施，在此藉由示例之方 式並參照下列圖式來描述一特定具體例，其中： 第1圖顯示穿透式電子顯微鏡(TEM)剖視圖（AL)，其顯 示CVD氮化鈦(TiN)阻障層與孔隙性低k值金屬層間介電層 (IMD)之間的交互作用； 第2圖顯示掃瞄式電子顯微鏡(SEM)剖面圖（M0)，其顯 示CVD氮化鈦(TiN)阻障層的交互作用； 第3圖是闡釋本發明特徵之一種製有溝槽之介電材料 的穿透式電子顯微鏡(TEM)剖視圖； 第4圖是第3圖之部分放大圖；及 第5a-d圖是以圖表顯示延第4圖所指示之a線之電子能 量耗失光譜(EELS)。 第6a圖是本發明介電薄膜之穿透式電子顯微鏡(TEM) 明視野剖面圖。 第6b及6c圖是闡釋前置技藝之前驅物擴散之穿透式電 子顯微鏡明視野剖面圖。 第7圖是闡釋本發明特徵之一個〇18微米結構的明視 野穿透式電子顯微鏡(TEM)圖（其相當於第3圖，但其結構 更微小）。 弟8圖是第7®結構之鈦以a)是氮氣及氧氣光阻去除/處 氮氣及氧氣光阻去除/處理之電子能量耗失光譜(eels) 掃猫。 第9圖是糊結構之碳叫是聽及氧氣轨去除/處 _氮氣及氫氣光阻絲/處理之電子能量耗失切(EELS) 9¾。 第10圖是闡釋本發明特徵之一種製有溝槽之介電材料 的明視野穿透式電子顯微鏡(TEM)圖，其中a)是全覽圖， b)、c)及d)是高倍放大圖。 第11圖顯示經由以本發明製成呈〇.18微米溝槽結構漏 電流。 第12圖顯示經由以本發明製成呈〇. 2 5微米溝槽結構漏 電流。 第13圖顯示本發明製造結構之凹陷式通道(Rc)產品。 第14圖顯示金屬有機化學氣相沈積(M〇CVD)與物理 氣相沈積(PVD)阻障層對本發明介電薄膜之比較。 t實施方式3 較佳實施例之詳細說明 參見第3圖所闡釋之一組測試疊層，其係顯示嵌刻導線 製程架構中常見之一種結構。該疊層10配置於一基底介電 層11之上，且係由一餘刻終止層12、一介電層13 (具有餘 刻於其中之溝槽14)、一居於介電層13頂面之碳化矽帽蓋 15、以及一阻卩早層16所構成。添加一氧化石夕層η及一偏光 層18則僅供TEM樣品製備之目的。 介電層11及13是由一具有超過1〇%碳之低]^iSic〇HM 料構成’由本案申明人給予之商品名為〇ri〇n。此材料具孔 隙性且具有一介電常數k為大約2.2。偏光層18是一種由本 案申請人給予之商品名為Flowfill的材料。 沉積介電層11及13是使用如WO-A-01/01472所述之发明 Description of the invention: [Interaction with related patent applications is Cangkao materials] This patent application claims the US provisional application serial number 60 / 392,057 filed on June 28, 2002 and on June 14, 2002 Priority is claimed for UK Patent Application No. 0213708.1 filed on June 18, 2002 and UK Patent Application Number 0213888.1 filed on June 18, 2002. I: Technical Field to which the Invention belongs 3. Field of the Invention The present invention relates to a porous 10-type dielectric film having a dielectric constant (k) of less than about 2.5. The development and manufacture of these low-k dielectric materials for (particularly) semiconductor devices has been ongoing for many years, and is compatible with increasingly smaller device architectures. According to what is currently determined to achieve a k value below about 2.5 for a general-purpose insulator, these materials will inevitably have some degree of porosity. This porosity presents a major problem in integration, especially the dielectric layer formed with 15 vias or interconnect modules, because during the etching, if the pores are actually interconnected or penetrate the surface, it should be formed by etching. The panel will be rough and may be penetrated. In the current typical architecture, copper is deposited into these etched trenches and vias, and since copper will easily diffuse into the dielectric material, the dielectric 20 material must be packaged with an isolated diffuser. An ideal is an insulator with isolation properties, but the current solution is to rely on a separate deposit. [PRIOR ART 3 BACKGROUND OF THE INVENTION The deposition of these isolation layers has traditionally been done using physical vapor deposition techniques, but 5 1223331 these techniques do not provide sufficient isolation coating uniformity 'and therefore use chemical vapor deposition (CVD) techniques such as: metal organic CVD, metal oxide CVD and atomic layer CVD. Because the CVD technology can produce nearly 100% homogeneity, both the front conductor and the reactant can penetrate the porous dielectric. This effect is shown in Figure 1 (Source: IMEC at ARMM 2001Conference) and Figure 2 (Source: Passemard et al; "intergration issue of low k and ULK materials and damascene structure" at ARMM 2001Conference). In both cases, it can be observed that the panel appears "fuzzy", which means that the CVD precursor has been adsorbed to the porous dielectric 10 layer, which creates an unbounded isolation between the dielectric and the isolation layer. . The method described in EP-A-1195801 is believed to increase the porosity of a panel, and it creates a panel seal by providing a protective or sealing layer. The method implies that the sealing layer can be formed by a plasma containing oxygen and nitrogen, but it does not actually describe the method. Adding additional material to the 15-height aspect ratio vias is undesirable and may increase the resistance of the copper in the vias. It is unclear whether the local low k value will be maintained on the surface of the sealing layer as suggested by the application. [Summary of the Invention 3 Summary of the Invention 20] According to one aspect of the present invention, the present invention includes a porous dielectric film having a dielectric constant (k) of less than about 2.5 and a carbon content of not less than 0%. It includes a guide hole or other etched formations dwelling in the guide hole, wherein ehai resides on the exposed surface of the guide hole or the formation or the surface formed by the film is substantially non-porous. 6 It can be understood that this method is completely opposite to EP-A-1195801. The dielectric process in this case will increase the local porosity of the surface layer, and this problem can only be overcome by adding an additional sealing layer. It should also be noted that sealing the porosity surfaces on the top and bottom of the structure is considered easier than sealing the sides parallel to the incoming reaction stream. In a preferred embodiment, the exposed surface or surfaces are made of a carbon-depleted film rather than the entire film. Additionally or alternatively, the exposed surface or surfaces may be made of a layer having a density higher than that of the entire film material 10. In a particularly preferred embodiment, the exposed surface or surfaces may be made of a carbon-depleted film as compared to the entire film. Additionally or alternatively, the exposed surface or surfaces may be made of a layer consisting essentially of silicon-silicon bonds, and these bonds may be formed between trivalent silicon molecules. Other institutions are currently unknown, but they exist on the exposed surface or surfaces. It can be understood in each of these cases that the layers constituting the exposed surface or surfaces are made by modifying the etched dielectric material rather than by further deposition. The entire film is preferably made of a SiCOH material. 20 The surface forming layer or layers can be made by plasma treatment of etched or other surfaces containing nitrogen or hafnium. This process may be performed at least partially simultaneously with another process (such as photoresist removal). . The present invention further includes a barrier layer covering the exposed surface or surfaces, in which case the barrier layer does not penetrate the exposed surface or surfaces. The barrier layer is preferably deposited by chemical vapor deposition (CVD). A further aspect of the present invention includes a method for manufacturing an interconnect module layer body in a semiconductor device, the method comprising: a: depositing a low-K porosity dielectric film on a substrate; 5 b : Depositing a photoresist; c: forming a pattern on the photoresist to define an etching hole; d ·· on the dielectric layer, etching a via hole or an etching formation through the hole; and e: removing the photoresist, wherein The photoresist is removed by a nitrogen or an inert gas or a combination thereof, and a hydrogen plasma or a nitrogen or an inert gas 10 or a combination thereof, and an oxygen plasma, and the via hole or the etching formation is exposed. The surface is also exposed to plasma at the same time, so that the surface layers constituting the exposed surface are dense. The method also includes depositing a barrier layer on the dense exposed surface. This barrier layer can be deposited by a chemical vapor deposition method. 15 Preferably, there is more nitrogen or inert gas than hydrogen or oxygen. Therefore, a better nitrogen (N2) ·· hydrogen (H2) ratio is 5: 1. The substrate may be RF biased during photoresist removal. In the alternative method, the densification can be performed in the etching step, and it can be manufactured by a non-oxidizing plasma process (for example, when the characteristics of a low-k material are organic). Although the present invention has been defined as described above, it is understood that the present invention includes any type of creative combination of the features proposed above or described below. Brief Description of the Drawings 8 The present invention can be implemented in various different ways. Here, a specific specific example is described by way of example and with reference to the following drawings, in which: Figure 1 shows a transmission electron microscope (TEM) cross-sectional view (AL ), Which shows the interaction between a CVD titanium nitride (TiN) barrier layer and a low-k porosity metal interlayer dielectric layer (IMD); Figure 2 shows a scanning electron microscope (SEM) section (M0 ), Which shows the interaction of CVD titanium nitride (TiN) barrier layers; FIG. 3 is a cross-sectional view of a transmission electron microscope (TEM) illustrating a trench-made dielectric material illustrating the features of the present invention; FIG. 4 It is a partially enlarged view of FIG. 3; and FIGS. 5a-d are graphs showing electron energy loss spectra (EELS) extending along line a indicated by FIG. 4. Figure 6a is a cross-sectional view of the bright field of a transmission electron microscope (TEM) of a dielectric film of the present invention. Figures 6b and 6c are cross-sectional views of the bright field of a transmission electron microscope illustrating the diffusion of precursors before the pre-processing technique. Fig. 7 is a bright-field transmission electron microscope (TEM) image of an 018 micron structure (which is equivalent to Fig. 3, but with a smaller structure) explaining the features of the present invention. Figure 8 shows the 7th structure of titanium. A) is the nitrogen and oxygen photoresist removal / treatment. The electron energy loss spectrum (eels) of the nitrogen and oxygen photoresist removal / treatment. Figure 9 shows the carbon structure of the paste structure. It is listened to the oxygen track removal / placement. _ Nitrogen and hydrogen photoresistance wire / processing electron energy loss cut (EELS) 9¾. Figure 10 is a bright-field transmission electron microscope (TEM) image of a trenched dielectric material illustrating the features of the present invention, where a) is a full view, b), c), and d) are high magnifications Illustration. Figure 11 shows the leakage current through a trench structure of 0.18 micrometers made in the present invention. Fig. 12 shows leakage current through a trench structure of 0.25 micrometers made in the present invention. FIG. 13 shows a recessed channel (Rc) product of the manufacturing structure of the present invention. FIG. 14 shows a comparison of a metal organic chemical vapor deposition (MOCVD) and a physical vapor deposition (PVD) barrier layer to a dielectric film of the present invention. Embodiment 3 Detailed Description of the Preferred Embodiments Refer to a group of test stacks illustrated in FIG. 3, which shows a structure common in the process structure of embedded conductors. The stack 10 is disposed on a base dielectric layer 11 and consists of a stop layer 12, a dielectric layer 13 (with trenches 14 etched therein), and a top surface of the dielectric layer 13. The silicon carbide cap 15 and an early blocking layer 16 are formed. The addition of a monoxide layer η and a polarizing layer 18 is only for the purpose of TEM sample preparation. The dielectric layers 11 and 13 are composed of a low carbon content of more than 10%] iSicOHM material ', which is a trade name given by the claimant of this case. This material is porous and has a dielectric constant k of about 2.2. The polarizing layer 18 is a material named Flowfill given by the applicant of the present case. The dielectric layers 11 and 13 are deposited using the method described in WO-A-01 / 01472
Trikon/xP™工具機，該案揭露内容在此併入本案作為參考 資料。此材料是一種聚合物的冷沉積物，其後以氫氣電漿 固化。 在沒有一硬式光罩下，於一部Trikon MORI™迴旋共振 5 供應電漿蝕刻工具機中，使用（CF4/CH2F2)化學，以射頻(RF) 晶圓偏壓來蝕刻溝槽14。接續劃定供用於溝槽之蝕刻孔洞 之電漿蝕刻光阻後，於MORI™工具機，再次以迴旋共振波 型電漿供應器及射頻(RF)晶圓偏壓，使用5 : 1氮氣(N2): 氫氣(H2)化學來原位去除光阻（即於同一反應室内）。此去 10 除工阻亦同時移除聚合物殘基。如本技藝中所熟知，於去 除光阻至接續MOCVD阻障沉積阻障層16之間可進一步進 行濕式或乾式製程步驟，然而於此個例中並不施行任一 者。由於這些製程為習知本技藝人士所已知，因此不在此 詳述。 15 於一獨立系統中，使用TDEAT(四二乙胺基鈦）及氨前 驅物以及氦氣供壓來沉積MOCVD氮化鈦(矽）(TiN(Si))。迅 即於沉積之後，以氫氣電漿處理該MOCVD薄膜，其後浸泡 矽烷。於沉積之前無加熱或電漿處理。 可立即由第3圖中參見，甚至於第4圖會更清楚看見， 20 居於阻障層16與溝槽侧壁之間的交界面是平滑且連續，這 與該等顯示於第1及第2圖之前置技藝的佈置是完全相反 的。更進一步來看，該阻障層本身是平滑且連續。其顯微 照片更進一步顯示：臨近阻障層之溝槽側壁較諸該等側壁 遠端區域具有較少之孔隙（較為緻密）。該緻密區域於明視 11 野TEM影像中較暗，且於顯微照片中被標記為j及κ。 I由於_薄財㈣均雜，因麟賴侃（電浆钱 /或接、只光阻去除)之期間，該溝槽側壁產生緻密化。據 信由於在_形成期間，有大量聚合物存在於難(俾以產 5生異向㈣）’其可藉由接續去除製程予以移除，因此至少 大部分的緻密化是發生於去除光阻之期間。 旦更進—步的側壁緻密化證據是來自第5a-c圖。電子能 I耗失光譜(EELS)分析可以被使用來提供有關樣品整體厚 度及、，且成物以及個別元件分佈之資料。空間圖像可以取第4 1〇圖所確岭之軸線A作—系列一維圖像來予以產生其結果被 顯不於第5a-b圖。 比車乂諸等圖像可支持一緻密化溝槽侧壁的存在。繪製 於第5圖之訊號隨樣品厚度及樣品組成物/密度而異。居於 阻障層16雙峰之間之介電層n所產生之凹陷性訊號顯示介 15電層13靠近側壁是較為緻密或稠密的。據信該差異並非由 於厚度。第5b圖顯示鈦訊息，並證實阻障層16之界限緊密， 由薄膜13無法偵測到任何鈦訊號。第5c圖顯示溝槽側壁是 石厌空乏，而苐5d圖之氧氣曲線圖則相當平坦。 因此結論是溝槽蝕刻及/或光阻去除製程會緻密孔隙 20性低^值層體之溝槽側壁，藉此產生一平滑表面來避免阻障 層前驅物或反應物滲入。這能夠使一連續性阻障層被沉積 來防止銅滲入。雖然這些試驗僅施用於本案申請人之材 料’然而相信以至少某些其他具有超低k值孔隙性介電薄膜 (特別是諸等SiCOH家族，其係為孔隙性且含有二氧化矽 12 1223331 之氫化碳）應可得到相同之結果。此等薄膜内之碳及氫典 型是C-Η3基團，其具有效結合大量氫之c_si鍵，而此氫被 認為是使薄膜基質及所產生孔隙具有低]^值之主要因素。 緻密化之確實機制尚不清楚，但據信可能是碳由緻密 5層缺失使三價石夕原子之間形成Si_Si(石夕-石夕)鍵結。 BARC之反應性離子蝕刻製程及居於2〇〇匪晶圓上具 有一光阻遮罩之孔隙性低k值SiCOH材料為： 製程氣體 CF4及CH2F2呈4.4 : 1至6.6 : 1之比例 壓力 1_5-2亳托 _ 電漿功率 丨·25仟瓦(KW)至一感應天線 晶圓偏壓功率 400瓦特Trikon / xP ™ machine tools, the disclosure of this case is hereby incorporated into this case for reference. This material is a cold deposit of a polymer that is then cured with a hydrogen plasma. Without a hard mask, trenches 14 are etched with radio frequency (RF) wafer bias using (CF4 / CH2F2) chemistry in a Tricon MORI ™ cyclotron resonance 5 plasma plasma tool. Following the plasma etching photoresist that was destined for the etching holes of the trenches, it was again biased with a cyclotron-plasma supplier and radio frequency (RF) wafer at a MORI ™ machine tool using 5: 1 nitrogen ( N2): Hydrogen (H2) chemically removes photoresist in situ (ie, in the same reaction chamber). This removal of resistance also removes polymer residues. As is well known in the art, further wet or dry process steps may be performed between the removal of the photoresist and the subsequent MOCVD barrier deposition barrier layer 16, but neither is performed in this example. Since these processes are known to those skilled in the art, they are not described in detail here. 15 In a separate system, MOCVD titanium nitride (silicon) (TiN (Si)) was deposited using TDEAT (tetradiethylamino titanium) and ammonia precursors and helium pressure. Immediately after the deposition, the MOCVD film was treated with a hydrogen plasma, and then silane was immersed. No heating or plasma treatment before deposition. It can be immediately seen from FIG. 3, and even more clearly from FIG. 4, the interface between the barrier layer 16 and the sidewall of the trench 20 is smooth and continuous, as shown in FIGS. 1 and 1. The arrangement of the prior art in Figure 2 is completely opposite. Looking further, the barrier layer itself is smooth and continuous. The photomicrographs further show that the trench sidewalls adjacent to the barrier layer have fewer pores (more dense) than the distal regions of these sidewalls. This dense area is darker in the photopic 11-field TEM image and is marked as j and κ in the photomicrograph. Due to the fact that the thin money is thin, the side wall of the trench is densified during the period of Lin Laikan (plasma money / or connection, only photoresist removal). It is believed that during the formation of _, a large number of polymers existed in difficulty (俾 producing 5 different anisotropies) ′ which can be removed by successive removal processes, so at least most of the densification occurred during photoresist removal Period. Once further—the evidence for side wall densification is from Figures 5a-c. Electronic energy loss spectroscopy (EELS) analysis can be used to provide information about the overall thickness of the sample, and the distribution of the product and individual components. The spatial image can be taken as the axis A of the range defined in Figure 4-10—a series of one-dimensional images to produce it. The results are not shown in Figures 5a-b. Images such as these can support the existence of uniformly densified trench sidewalls. The signal plotted in Figure 5 varies with sample thickness and sample composition / density. The depressed signal generated by the dielectric layer n located between the double peaks of the barrier layer 16 shows that the dielectric layer 13 is dense or dense near the side wall. It is believed that the difference is not due to thickness. Figure 5b shows the titanium message and confirms that the barrier layer 16 is tightly bounded, and no titanium signal can be detected by the film 13. Figure 5c shows that the sidewalls of the trench are exhausted, while the oxygen curve in Figure 5d is fairly flat. Therefore, it is concluded that the trench etching and / or photoresist removal processes will compact the trench sidewalls of the low-porosity layer, thereby creating a smooth surface to prevent the barrier layer precursors or reactants from infiltrating. This enables a continuous barrier layer to be deposited to prevent copper infiltration. Although these tests are only applied to the applicant's materials, it is believed that at least some other dielectric films with ultra-low-k porosity (especially the SiCOH family, which are porous and contain silicon dioxide 12 1223331) Hydrogenated carbon) should give the same results. The carbon and hydrogen in these films are typically C-Η3 groups, which have c_si bonds that effectively bind a large amount of hydrogen, and this hydrogen is considered to be the main factor that causes the film matrix and the pores to have a low value. The exact mechanism of densification is unclear, but it is believed that the absence of dense 5 layers of carbon caused Si_Si (Shi Xi-Shi Xi) bonding between the trivalent Shi Xi atoms. The reactive ion etching process of BARC and the low-k porosity SiCOH material with a photoresist mask on a 200-band wafer are as follows: The process gases CF4 and CH2F2 are at a ratio of 4.4: 1 to 6.6: 1 and pressure 1_5- 2 亳 托 _ Plasma power 丨 · 25 仟 W (KW) to an inductive antenna wafer bias power 400W
模板溫度 -15°C 於同一反應室内進行200mm晶圓上反應性離子光阻去 除製程為： 氮(N£)及氫(¾)呈5 : 1之比例 5毫托 2.5仟瓦(KW)至一感應天線 200瓦特 o°c 製程氣體 壓力 電漿功率 晶圓偏壓功率 模板溫度 10 此姓刻製程是以靜電晶圓夾钳及氦氣背面壓力，且因 此使晶圓溫度接近模板溫度。低溫被使用來維持光阻完敕。 光阻去除製程不夾鉗晶圓，俾使晶圓溫度較高，藉此 改善殘基移除效率並增加去除速率.。於叱模板溫度下:曰匕 圓溫度峰值標示為⑽（工業標準加熱貼紙），而皿2 15板溫度下為l〇4^。 匕挺 13 1223331 雖然這些武驗使用氮氣及氫氣，然而設若氮氣於緻密 化製程中不具反應性時，可選擇性取代為（例如）氦、氖、氬、 氙及氪或任何一種其他適合濺鍍蝕刻之氣體。其等可選擇 性地被添加至氮氣及/或氫氣混合物中。 5 進行更進一步的試驗來闡釋本發明的效用。於第6(a) 圖顯不之明視野穿透式電子顯微鏡(TEM)影像為一個由本 發明介電薄膜所構成之完整結構，其具有一MOCVD沉積氮 化鈦阻障層以及由一濺鍍銅接種層、電鍍銅及化學機械研 磨步驟所構成之完全銅填洞。可參見沒有金屬擴散自阻障 10層或者銅擴散至介電薄膜。 更進一步於第6(a)圖顯示一個厚度5至8奈米之非結晶 層，其係被電漿處理改質，且具有一較諸整體孔隙性介電 薄膜有更高之密度。 相反地，於第6(b)及6(c)圖則顯示有前驅物擴散。於第 15 6(b)圖之影像是摘錄自 w. Besling，Proc. IITC 2002The template temperature is -15 ° C, and the reactive ion photoresist removal process on a 200mm wafer in the same reaction chamber is: nitrogen (N £) and hydrogen (¾) in a 5: 1 ratio of 5 mTorr 2.5 仟 W (KW) to An inductive antenna 200 watt o ° c process gas pressure plasma power wafer bias power template temperature 10 This engraving process is based on electrostatic wafer clamps and helium back pressure, so the wafer temperature is close to the template temperature. Low temperatures are used to maintain photoresistance. The photoresist removal process does not clamp the wafer, which increases the temperature of the wafer, thereby improving the efficiency of residue removal and increasing the removal rate. At the temperature of the 曰 template: the peak temperature of the circle is marked as ⑽ (industrial standard heating sticker), and the temperature at the plate 2 15 plate is 104 ^. Dagger 13 1223331 Although these tests use nitrogen and hydrogen, if nitrogen is not reactive in the densification process, it can be selectively replaced with, for example, helium, neon, argon, xenon and krypton or any other suitable sputtering. Etching gas. These are optionally added to a nitrogen and / or hydrogen mixture. 5 Further experiments are performed to illustrate the effectiveness of the invention. The TEM image of the unclear field of view shown in Figure 6 (a) is a complete structure composed of the dielectric film of the present invention, which has a MOCVD-deposited titanium nitride barrier layer and a sputtering layer. Complete copper filling hole formed by copper seeding layer, copper electroplating and chemical mechanical polishing steps. See self-barrier 10 layers without metal diffusion or copper diffusion into the dielectric film. Further, Fig. 6 (a) shows an amorphous layer having a thickness of 5 to 8 nm, which is modified by plasma treatment and has a higher density than the overall porous dielectric film. Conversely, precursors are shown in Figures 6 (b) and 6 (c). The image in Figure 15 6 (b) is an excerpt from w. Besling, Proc. IITC 2002
Burlingame (CA) USA，2002 pP288-291。於第 6⑷圖之影像 則是摘錄自 S. Kawamura et al· Proc· IITC 2001 San Francisco, USA，ppl95-197。由明視野TEM影像參見的是金 屬擴散入介電薄膜的一種熟知且可信賴指標。 20 第7圖是如上文所述製成之一個0.18微米結構，其參照 第3圖之TEM影像。看不到金屬擴散通過阻障層，且被更進 一步驗證於第8(a)及8(b)圖，該圖為第7圖所顯示結構之鈦 的電子能量耗失光譜(EELS)掃瞒。 第8(a)圖闡釋氮氣及氧氣混合物之電子能量耗失光譜 14 (EELS)掃瞄。第8(b)圖則使用一種由200 seem氮氣及10 seem氧氣所構成之氣體混合物（一個目前所建立之最佳比 例為40 : 1)。氧氣被熟知可移除碳，且此試驗闡釋氮氣可 減低氧氣之碳移除效應，並容許孔隙性介電薄膜能夠維持 5 住一種氣體金屬前驅物之高吸收度（雖然不如氮氣+氫氣）。 此製程與EP-A-1195801所述相反，該案使用一種氮氣/氧氣 電漿來形成一密封層。 第9(a)及9(b)圖顯示第7圖結構之碳的電子能量耗失光 譜(EELS)掃瞄。於第9(a)圖中，EELS掃瞄氮氣及氧氣顯示 10 其較諸該被闡釋於第9(b)圖之氮氣及氫氣的情形而言，介電 薄膜側壁有較多的碳耗失。 第10圖是更進一步闡釋本發明具體例之明視野TEM影 像。第10(a)圖是如上文第3圖所述製成之一個結構的全覽 圖。第10(d)圖是如上文所述之氮氣及氫氣製程的結果，而 15 第10(b)及10(c)圖則為闡釋一種氮氣與氧氣之氣體混合物 處理的影像。 第11至14圖顯示以諸等本發明具體例之介電薄膜製成 的電子測試結構結果。測試結構為單一鑲欲之線寬/間距 0.18及0.25微米溝槽/間距。指狀梳具有一周邊長44公分， 20 尺寸為100微米xl600微米。線間漏電流被檢測為0.5 MV/cm，而線間容電量則被檢測為1MV。 電漿處理/光阻去除製程如下： 15 氮氣+氫氣 (比例5 : 1) 氮氣 200 seem(每分鐘標準立方公分） 氫氣 40 seem 壓力 7毫托 模板溫度 -15°C，具有2托耳氦氣背壓之靜電吸附極 MORFM電漿供應器 感應耦合 電漿功率 2.5计瓦(kW)至一 13.56 MHz之感應天線 磁場功率 40/60安培内線圈/外線圈 模板功率（晶圓偏壓） 200 W 13.56MHz 氮氣+氧氣 (比例20 : 1) 氮氣 200 seem(每分鐘標準立方公分） 氧氣 10 seem 壓力 7毫托 模板溫度 -15°C ’具有2托耳氦氣背壓之靜電吸附極 MORJ™電漿供應器 感應轉合 電漿功率 2.5仟瓦(kW)至一 13.56 MHz之感應天線 磁場功率 60/60安培内線圈/外線圈 模板功率（晶圓偏壓） 30 W 13.56MHz 1223331 此為最佳之氮氣+氧氣緻密化製程，且其與 5 EP-A-1195801所述之密封製程相反。 應注意於這些接續的試驗中，晶圓是以靜電夾鉗，藉 此來降低其溫度，俾以接近模板溫度。已發現此等製程於 這些較低之晶圓溫度下仍具效力。 第11圖顯示該專氮氣+氫氣及氮氣+氧氣之氣體混合 10物，於0.18微米溝槽結構漏電流結果(較少為佳）。可看見較 諸氫氣而言，氧氣效能哀降。此可預期其於第9(a)及9(b)圖 之EELS結果顯示氮氣+氧氣製程會增加碳耗失。再者，一 16 種濕式清洗並不會衰降氮氣+氫氣處理介電薄膜，但卻會略 微衰降氮氣+氧氣處理介電薄膜，其更進一步指出具有某種 程度之孔隙度。此種濕式清洗被廣泛知悉且被使用於工業 界，用以移除一乾式光阻去除製程之後的任何_種殘基。 第12圖係如同第11圖，但其係於0·25微米結構上，對比 較氮氣與氫氣或氧氣之作用作更進一步闡釋。其結果與於 論同第11圖。 第13圖顯示凹陷式通道(RC)產品之測試結構（較少為 佳）。可看出一工業標準濕式清洗會略微衰降氮氣+氫氣及 氮軋+氧軋製程之RC產品，但其再次顯示氮氣+氫氣製程有 較佳之結果。 第14圖顯示以金屬有機化學氣相沈積（M〇CVD)與物 理氣相沈積(PVD)濺鍍方法沉積之阻障層對〇18及〇25微米 電子測試結構的比較結果。該比較顯示漏電流量低且相似。 【圖式1簡明】 第1圖顯示穿透式電子顯微鏡(TEM)剖視圖，其顯 示CVD氮化鈦(TiN)阻障層與孔隙性低k值金屬層間介電層 (IMD)之間的交互作用； 第2圖顯示掃瞄式電子顯微鏡(SEM)剖面圖，其顯 示CVD氮化鈦(TiN)阻障層的交互作用； 第3圖是闡釋本發明特徵之一種製有溝槽之介電材料 的穿透式電子顯微鏡(TEM)剖視圖； 第4圖是第3圖之部分放大圖；及 第5a-d圖是以圖表顯示延第4圖所指示之a線之電子处 ^月匕 量耗失光譜(EELS)。 第6a圖是本發明介電薄膜之穿透式電子顯微鏡(TEM) 明視野剖面圖。 第6b及6c圖是闡釋前置技藝之前驅物擴散之穿透式電 子顯微鏡明視野剖面圖。 第7圖是闡釋本發明特徵之一個0.18微米結構的明視 野牙透式電子顯微鏡(TEM)圖（其相當於第3圖，但其結構 更微小）。 第8圖疋第7圖結構之鈦以a)是氮氣及氧氣光阻去除/處 理b)氮氣及氧氣光阻去除/處理之電子能量耗失光譜(eels) 掃瞄。 第9圖是第7圖結構之碳以句是氮氣及氧氣光阻去除/處 理b)氮氣及氫氣光阻去除/處理之電子能量耗失光譜(EELS) 掃瞄。 第10圖是闡釋本發明特徵之一種製有溝槽之介電材料 的明視野穿透式電子顯微鏡(TEM)圖，其中a)是全覽圖， b)、c)及d)是高倍放大圖。 第11圖顯示經由以本發明製成呈0.18微米溝槽結構漏 電流。 第12圖顯示經由以本發明製成呈〇 · 2 5微米溝槽結構漏 電流。 第13圖顯示本發明製造結構之凹陷式通道(RC)產品。 第14圖顯示金屬有機化學氣相沈積(MOCVD)與物理 氣相沈積(PVD)阻障層對本發明介電薄膜之比較。 1223331 【圖式之主要元件代表符號表】Burlingame (CA) USA, 2002 pP288-291. The image in Figure 6 is an excerpt from S. Kawamura et al. Proc. IITC 2001 San Francisco, USA, ppl95-197. Seen from the bright-field TEM image is a well-known and reliable indicator of metal diffusion into a dielectric film. 20 Figure 7 is a 0.18 micron structure made as described above, referring to the TEM image of Figure 3. No metal diffusion can be seen through the barrier layer and it is further verified in Figures 8 (a) and 8 (b), which is the electron energy loss spectrum (EELS) of the structure shown in Figure 7 . Figure 8 (a) illustrates an electron energy loss spectrum 14 (EELS) scan of a nitrogen and oxygen mixture. Figure 8 (b) uses a gas mixture of 200 seem nitrogen and 10 seem oxygen (a currently established optimal ratio of 40: 1). Oxygen is known to remove carbon, and this test illustrates that nitrogen reduces the carbon removal effect of oxygen and allows the porous dielectric film to maintain a high absorption (although not as good as nitrogen + hydrogen) of a gas metal precursor. This process is in contrast to EP-A-1195801, which uses a nitrogen / oxygen plasma to form a sealing layer. Figures 9 (a) and 9 (b) show the electron energy loss spectroscopy (EELS) scan of carbon in the structure of Figure 7. In Figure 9 (a), the EELS scan of nitrogen and oxygen shows 10 that compared with the nitrogen and hydrogen cases that should be explained in Figure 9 (b), the dielectric film sidewall has more carbon loss. . Fig. 10 is a bright field TEM image further explaining a specific example of the present invention. Figure 10 (a) is an overview of a structure made as described in Figure 3 above. Figure 10 (d) is the result of the nitrogen and hydrogen processes described above, and 15 Figures 10 (b) and 10 (c) are images illustrating the treatment of a gas mixture of nitrogen and oxygen. Figures 11 to 14 show the results of an electronic test structure made of dielectric films of various specific examples of the present invention. The test structures are a single inset line width / space of 0.18 and a 0.25 micron trench / space. The finger comb has a circumference of 44 cm and a size of 100 micrometers x 1600 micrometers. The line-to-line leakage current was detected as 0.5 MV / cm, and the line-to-line capacitance was detected as 1 MV. The process of plasma treatment / resistance removal is as follows: 15 nitrogen + hydrogen (proportion 5: 1) nitrogen 200 seem (standard cubic centimeters per minute) hydrogen 40 seem pressure 7 mTorr template temperature -15 ° C, 2 Torr helium Back pressure of electrostatic adsorption pole MORFM plasma supplier Inductive coupling plasma power 2.5 meter watts (kW) to a 13.56 MHz induction antenna magnetic field power 40/60 amp inner coil / outer coil template power (wafer bias) 200 W 13.56MHz nitrogen + oxygen (proportion 20: 1) nitrogen 200 seem (standard cubic centimeters per minute) oxygen 10 seem pressure 7 millitorr template temperature -15 ° C 'electrostatic adsorption electrode MORJ ™ with 2 Torr helium back pressure Plasma supplier Inductive turn-on plasma power 2.5 仟 W (kW) to a 13.56 MHz inductive antenna magnetic field power 60/60 amp inner coil / outer coil template power (wafer bias) 30 W 13.56MHz 1223331 This is the best The nitrogen + oxygen densification process is the opposite of the sealing process described in 5 EP-A-1195801. It should be noted that in these successive tests, the wafer is electrostatically clamped to reduce its temperature and approach the template temperature. These processes have been found to be effective at these lower wafer temperatures. Fig. 11 shows the leakage current results of the mixed gas of nitrogen + hydrogen and nitrogen + oxygen in a 0.18 micron trench structure (less preferred). It can be seen that the efficiency of oxygen is diminished compared to hydrogen. It can be expected that the EELS results in Figures 9 (a) and 9 (b) show that the nitrogen + oxygen process will increase carbon loss. Furthermore, a 16-type wet cleaning does not degrade the nitrogen + hydrogen treated dielectric film, but it does slightly degrade the nitrogen + oxygen treated dielectric film, which further indicates that it has a certain degree of porosity. This type of wet cleaning is widely known and used in industry to remove any residues after a dry photoresist removal process. Figure 12 is the same as Figure 11, but it is on a 0.25 micron structure. The effect of nitrogen and hydrogen or oxygen is further explained. The results are the same as in Figure 11. Figure 13 shows the test structure (less preferred) of recessed channel (RC) products. It can be seen that an industry standard wet cleaning will slightly reduce the RC products of nitrogen + hydrogen and nitrogen rolling + oxygen rolling process, but it again shows that the nitrogen + hydrogen process has better results. Fig. 14 shows the comparison results of the barrier layer deposited by metal organic chemical vapor deposition (MOCVD) and physical vapor deposition (PVD) sputtering methods on the electronic test structures of 018 and 025 microns. The comparison shows that the amount of leakage current is low and similar. [Simplified Figure 1] Figure 1 shows a cross-sectional view of a transmission electron microscope (TEM) showing the interaction between a CVD titanium nitride (TiN) barrier layer and a low-k porosity metal interlayer dielectric layer (IMD). Figure 2 shows a cross-sectional view of a scanning electron microscope (SEM) showing the interaction of a CVD titanium nitride (TiN) barrier layer; Figure 3 is a trenched dielectric illustrating the features of the present invention Sectional view of a transmission electron microscope (TEM) of the material; Figure 4 is an enlarged view of a portion of Figure 3; and Figures 5a-d are graphs showing the electron processing of the a line indicated by Figure 4 Lost Spectrum (EELS). Figure 6a is a cross-sectional view of the bright field of a transmission electron microscope (TEM) of a dielectric film of the present invention. Figures 6b and 6c are cross-sectional views of the bright field of a transmission electron microscope illustrating the diffusion of precursors before the pre-processing technique. Figure 7 is a 0.18 micron clear-sighted wild-type transmission electron microscope (TEM) image illustrating the features of the present invention (which is equivalent to Figure 3 but with a smaller structure). Figures 8 and 7 of the structure of titanium are a) nitrogen and oxygen photoresist removal / treatment b) electron energy loss spectrum (eels) scan of nitrogen and oxygen photoresist removal / treatment. Figure 9 shows the structure of the carbon in Figure 7 with nitrogen and oxygen photoresist removal / treatment b) Scanning of electron energy loss spectrum (EELS) of nitrogen and hydrogen photoresist removal / treatment. Figure 10 is a bright-field transmission electron microscope (TEM) image of a trenched dielectric material illustrating the features of the present invention, where a) is a full view, b), c), and d) are high magnifications Illustration. Figure 11 shows the leakage current through a 0.18 micron trench structure made in accordance with the present invention. Fig. 12 shows leakage current through a trench structure of 0.25 µm made according to the present invention. FIG. 13 shows a recessed channel (RC) product of a manufacturing structure of the present invention. FIG. 14 shows a comparison of a metal organic chemical vapor deposition (MOCVD) and a physical vapor deposition (PVD) barrier layer to a dielectric film of the present invention. 1223331 [Representative symbol table for main elements of the diagram]
10…疊層 11…基底介電層 12…餘刻終止層 13…介電層 14…溝槽 15…碳化矽帽蓋 16…阻障層 17…氧化石夕層 18…偏光層10 ... Layer 11 ... Bottom dielectric layer 12 ... End stop layer 13 ... Dielectric layer 14 ... Trench 15 ... Silicon carbide cap 16 ... Barrier layer 17 ... Stone oxide layer 18 ... Polarizing layer
1919