Nothing Special   »   [go: up one dir, main page]

TW200849325A - Plasma processing method and plasma processing apparatus - Google Patents

Plasma processing method and plasma processing apparatus Download PDF

Info

Publication number
TW200849325A
TW200849325A TW97104310A TW97104310A TW200849325A TW 200849325 A TW200849325 A TW 200849325A TW 97104310 A TW97104310 A TW 97104310A TW 97104310 A TW97104310 A TW 97104310A TW 200849325 A TW200849325 A TW 200849325A
Authority
TW
Taiwan
Prior art keywords
plasma
bias
wafer
frequencies
frequency
Prior art date
Application number
TW97104310A
Other languages
Chinese (zh)
Other versions
TWI366861B (en
Inventor
Masahito Mori
Naoyuki Kofuji
Naoshi Itabashi
Original Assignee
Hitachi High Tech Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from JP2008002709A external-priority patent/JP5014166B2/en
Application filed by Hitachi High Tech Corp filed Critical Hitachi High Tech Corp
Priority to TW97104310A priority Critical patent/TW200849325A/en
Publication of TW200849325A publication Critical patent/TW200849325A/en
Application granted granted Critical
Publication of TWI366861B publication Critical patent/TWI366861B/zh

Links

Landscapes

  • Drying Of Semiconductors (AREA)

Abstract

The invention provides a plasma processing apparatus and a dry etching method for etching a multilayered film structure having steps with high accuracy. The plasma processing apparatus comprises a vacuum reactor 107, a lower electrode 113 placed within a processing chamber of the vacuum reactor and having a wafer 112 to be etched mounted on the upper surface thereof, bias supplying units 118 and 120 for supplying high frequency power for forming a bias potential to the lower electrode 113, a gas supply means 111 for feeding reactive gas into the processing chamber, an electric field supplying means 101 through 103 for supplying a magnetic field for generating plasma in the processing chamber, and a control unit 127 for controlling the distribution of ion energy in the plasma being incident on the wafer 112 via the high frequency power.

Description

200849325 九、發明說明 【發明所屬之技術領域】 本發明關於半導體製造方法,特別關於具有金屬閘極 /high-k構造與段差構造、立體構造,對底層膜與遮罩層 要求高的選擇性,而且要求垂直形狀的閘極形成時使用之 電漿處理方法及電漿處理裝置。 【先前技術】 數位家電、個人電腦、行動電話等使用之MOSFET ( 金氧半場效電晶體)裝置被要求局集積化、局速化、局機 能化。對應於此要求,髓行動電話多晶矽/Si02構造閘極 之微細化,新材料、新構造之閘極被檢討。 具有此種金屬閘極/high_k構造之平面型MOSFET或 F IN-FET之閘極形成方法所使用的乾蝕刻加工原理,係以 電磁波電漿化反應性氣體,使用產生之離子與自由基進行 之離子促進反應。因此,具體化此方法之電漿處理裝置, 係由電漿產生機構,反應性氣體導入機構,壓力控制機構 ,Si晶圓設置用之下部電極機構,Si晶圓搬送機構,及彼 等之動作時序之控制機構等構成。其中,下部電極機構係 由:Si晶圓固定用之靜電吸附機構,Si晶圓溫度控制機構 ,及偏壓施加機構等構成。 使用具彼等機構構成之電漿處理裝置,而控制離子能 量(IEDF,Ion Energy Distribution Function)分布的方 法,習知者有藉由偏壓施加之波形或頻率加以影響。例如 200849325 藉由施加脈衝狀偏壓之方法及施加具有25kHz以下低 2MHz以上局頻的2頻率偏壓之方法,而提升絕緣膜 時對Si之選擇性之技術被提案(例如專利文獻1 )。 ,偏壓機構輸出之頻率會受到電漿通過電漿鞘之時間 ,而具有IEDF,此被報告於例如非專利文獻1。 另外,關於電漿狀態檢測監控,藉由監控高頻之 、電流、相位,而檢測出腔室內壁或下部電極之絕緣 膜之絕緣劣化等異常的技術被提案於例如專利文獻2。 專利文獻1 :特開2002 — 1 4 1 34 1號公報 專利文獻2:特開2007— 250755號公報 非專利文獻 1 : Journal of Vacuum Science Technology A Volume 20 ρ·1759 【發明內容】 (發明所欲解決之課題) 如圖3 ( a)所示,針對具有STI段差310,具有 膜構造的金屬閘極/high-k閘極進行乾蝕刻而形成時 用習知電漿處理裝置時,係如圖3 ( b )所示,難以兼 能防止high-k閘極絕緣膜之底層貫穿312,防止下擺 形狀3 1 4之產生,且能獲得垂直之下擺形狀。同樣之 選擇性與垂直加工性,基於基板上設置之配線疏密引 疏密形狀差之產生等問題,對於FIN-FET之閘極蝕刻 更加嚴苛。因爲,50nm程度之FIN之段差上部之底 穿、閘極長度部分之上部之側壁鈾刻(side etching ) 頻與 鈾刻 另外 影響 電壓 覆蓋 and 多層 ,使 顧既 下拉 底層 起之 變爲 層貫 、以 200849325 及下部之裙襬殘留等之垂直方向之形狀差,會成爲特彳生_ 動之主要原因。 本發明目的在於提供’針對具有STI構造或立體閘極 構造(FIN-FET等)、由包含金屬材料與high-k材料的多 數層之膜所構成之閘極材料,進行乾蝕刻時,可實丨見胃的 底層膜選擇性、且穩定之垂直加工的電漿處理方法及電槳 處理裝置。 r (用以解決課題的手段) 達成上述目的之電漿處理裝置,係具備:真空容器; 下部電極,配置於該真空容器之處理室內,其上面載置電 漿處理對象之晶圓;偏壓施加機構,供給多數頻率之偏壓 電力,而於該下部電極形成偏壓電位;氣體供給機構,將 反應性氣體導入上述處理室內;調整機構,調整上述處理 室內之氣體壓力;及電磁波供給機構,於上述處理室內產 生電漿;具備:IEDF控制機構,其使射入上述晶圓之離 子之能量及IEDF獨立變化;及檢測機構,其檢測出相對 於偏壓頻率的電漿狀態。 此時,IEDF控制機構,係由振盪產生多數頻率的電 源部及各個匹配器構成,另外,檢測電漿狀態的機構,係 具備檢測相對於偏壓施加機構所供給之各個頻率的電漿阻 抗之機構。 達成上述目的之電漿處理方法,係具備:載置工程’ 將具備膜構造之晶圓載置於真空容器內部之處理室內之下 200849325 部電極上,該膜構造爲,表面由在high-k材料上包含有金 屬材料的多數層之膜所構成、具有段差構造者;導入工程 ’將蝕刻氣體導入該處理室內;調整工程,調整處理壓力 ;產生工程,於上述處理室內產生電漿;及供給工程,供 給1或多數個頻率之偏壓電力而於上述晶圓上形成偏壓電 位;藉由變化上述偏壓電力之輸出而進行上述膜構造之倉虫 刻者;其特徵爲具有:檢測工程,由上述偏壓施加機構側 檢測出電漿阻抗之時間變化;終點判斷工程,依據該檢測 結果來判斷電漿處理之終點;及控制工程,於上述終點判 斷之後,獨立控制射入上述晶圓之離子能量及其分布。 獨立控制射入上述晶圓之離子能量及其分布的控制工 程,係包含變化多數頻率之偏壓電力之輸出及彼等之混合 比’或者於檢測上述電漿之阻抗之時間變化的工程之後, 分離成爲壁面狀態成份與晶圓正上方成份之阻抗的工程, 具有:比較工程,將分離出之資料和資料庫或變動模型加 以比較;及依據該比較結果進行壁面潔淨的工程,或變化 次回之晶圓處理條件的工程。 【實施方式】 以下參照圖面說明本發明實施形態。首先,參照圖3 說明設爲本發明實施形態對象之膜構造之1例。圖3 ( a ) 爲具備金屬鬧極/high-k構造之平面(planar)型CMOS之 飩刻處理前之樣本斷面圖,圖3 ( b )爲習知方法之蝕刻處 理後之斷面圖,(c )爲本發明之蝕刻處理後之斷面圖。 -8- 200849325 於圖3(a),晶圓係由以下構成:形成於S i基板 309之上的STI308;形成於Si基板309與STI308之上的 HfSiON ( high-k )絕緣膜3 07 ;形成於High-k絕緣膜307 上的金屬閘極層3 0 6 ;於其上依序被積層形成的閘極C ap 層305;下層遮罩304;中間層遮罩303; BARC (抗反射 膜)層3 02 ;及於其上被形成的阻劑3 01。如圖所示,針 對STI3 0 8之形成引起之具有STI段差310的平面型積層 金屬/hi gh-k閘極,使用具有400kHz之單一偏壓頻率的習 知機構之閘極蝕刻裝置,使用Cl2/HBr氣體進行蝕刻時, 係如圖3 ( b )所示,在成爲MOSFET之主動部分之閘極 下部3 1 1會產生HfSiON ( high-k )膜之底層貫穿3 12。另 外,於STI段差310之角部會以裙襬(skirt)擴大狀產生 STI段差部閘極材料之殘留部分3 1 3。另外,於配線部分 之場(field)部分之閘極下部會以無底層貫穿形狀方式而 產生成爲裙襬擴大狀的下擺下拉形狀314,或於STI表面 產生STI上之閘極材料之殘留部分315。 亦即,在積層金屬/high-k閘極之蝕刻加工中,難以 兼顧閘極附近之底層high-k選擇性之提升及防止下擺下拉 形狀之產生。 如圖3 ( b )所示,400kHz偏壓下之底層選擇性之提 升及下擺下拉形狀產生之防止困難之原因在於,IEDF (離 子之能量分布)以平均能量爲中心具有擴散,薄膜上之光 學透明材料之蝕刻終點判斷變慢。 圖2爲離子之能量分布以同一時間平均能量爲中心呈 - 9 - 200849325 現不同分布之例。圖2爲’在電子溫度3eV、射入離子之 質量數79.9、電漿密度1 X ΙΟ^ηΓ1之電漿下,於低頻 4 0 0kHz與高頻13.56MHz之晶圓上施加Vpp二200 V之RF 時之多數IEDF之例之模式圖。於400kHz之IEDF203 ’相 對於時間平均之100eV ’分布寬度204成爲約200eV ’在 0V附近與200V附近具有2個峰値°另外’於高頻之 13.56MHz之IEDF201,分布寬度202成爲較窄之約50eV 。於低頻與高頻2個頻率各以100Vpp混合之IEDF205 ’ 則具有中間之分布。此乃因爲,越是高頻偏壓,加速離子 之鞘電壓變化之週期變爲越快,鞘振動時間相對於離子通 過鞘之時間變小,離子無法追隨鞘變動而以電場之平均値 被加速之故。 如上述說明,於400kHz之頻率進行蝕刻時,時間平 均之離子能量之約2倍之高能量離子使選擇性劣化之同時 ,約OeV之低能量離子成爲下擺下拉形狀之原因,兩者問 題之兼顧變爲困難。因此,欲兼顧選擇性及下擺下拉形狀 時,較好是以高頻窄化離子能量分布。另外,低頻時具有 較大擴散之分布,係例如被使用於以高能量離子以物理方 式濺鍍表面之變質層時,或需要高能量離子之絕緣膜蝕刻 時。另外,藉由使用低頻,在不影響電漿解離或分布情況 下,可以活用相對於平均能量具有擴散分布之高能量離子200849325 IX. Description of the Invention [Technical Field] The present invention relates to a semiconductor manufacturing method, and more particularly to a metal gate/high-k structure and a step structure, a three-dimensional structure, and a high selectivity to an underlying film and a mask layer, Further, a plasma processing method and a plasma processing apparatus which are used when forming a gate of a vertical shape are required. [Prior Art] A MOSFET (Gold Oxygen Half Field Effect Transistor) device used for digital home appliances, personal computers, and mobile phones is required to be integrated, speeded, and functionalized. Corresponding to this requirement, the miniaturization of the polysilicon/SiO2 structure gate of the pith mobile phone, the gate of new materials and new structures was reviewed. The principle of dry etching used in the gate forming method of a planar MOSFET or F IN-FET having such a metal gate/high_k structure is to electrolyze a reactive gas by electromagnetic waves, using the generated ions and radicals. Ions promote the reaction. Therefore, the plasma processing apparatus embodying this method is a plasma generating mechanism, a reactive gas introducing mechanism, a pressure control mechanism, a lower electrode mechanism for Si wafer mounting, a Si wafer transfer mechanism, and the like. The timing control mechanism and the like are constituted. The lower electrode mechanism is composed of an electrostatic adsorption mechanism for fixing a Si wafer, a Si wafer temperature control mechanism, and a bias application mechanism. The method of controlling the distribution of Ion Energy Distribution Function (IEDF) using a plasma processing apparatus having such a mechanism is known to have a waveform or frequency applied by a bias voltage. For example, in 200849325, a technique of applying a pulse-like bias and a method of applying a two-frequency bias having a local frequency of 2 MHz or more and a low frequency of 2 MHz or more are proposed, and a technique for improving the selectivity of Si in an insulating film is proposed (for example, Patent Document 1). The frequency at which the output of the biasing mechanism is subjected to the passage of the plasma through the plasma sheath has the IEDF, which is reported, for example, in Non-Patent Document 1. In addition, a technique for detecting an abnormality such as insulation degradation of an insulating film in a chamber wall or a lower electrode by monitoring a high frequency, a current, and a phase is proposed, for example, in Patent Document 2. Patent Document 1: JP-A-2002- 1 4 1 34 No. Patent Document 2: JP-A-2007-250755 Non-Patent Document 1: Journal of Vacuum Science Technology A Volume 20 ρ·1759 [Summary of the Invention] Problem to be solved) As shown in FIG. 3( a ), when a conventional plasma processing apparatus is formed by dry etching using a metal gate/high-k gate having a film structure having an STI step difference 310, it is as shown in FIG. As shown in Fig. 3(b), it is difficult to prevent the bottom layer of the high-k gate insulating film from penetrating 312, preventing the hem shape 3 1 4 from being generated, and obtaining a vertical pendulum shape. In the same way, the selectivity and vertical processability are more severe for the gate etching of the FIN-FET based on the problem that the wiring is densely distributed on the substrate and the density difference is generated. Because the bottom of the FIN segment of 50nm is the bottom of the bottom, the side etching of the upper part of the gate length portion and the uranium engraving additionally affect the voltage coverage and multiple layers, so that the bottom layer is turned into a layer, The difference in the vertical direction of the 200849325 and the skirt of the lower part of the skirt will become the main reason for the special _. An object of the present invention is to provide a gate material composed of a film having a STI structure or a three-dimensional gate structure (FIN-FET or the like) and a plurality of layers including a metal material and a high-k material, which can be used for dry etching. A plasma processing method and an electric paddle processing apparatus for selecting a stable and stable vertical processing of the underlying film of the stomach. r (Means for Solving the Problem) The plasma processing apparatus for achieving the above object includes a vacuum container, and a lower electrode disposed in a processing chamber of the vacuum container, on which a wafer to be subjected to plasma processing is placed; An application mechanism supplies a bias power of a plurality of frequencies to form a bias potential at the lower electrode; a gas supply mechanism that introduces a reactive gas into the processing chamber; an adjustment mechanism that adjusts a gas pressure in the processing chamber; and an electromagnetic wave supply mechanism And generating plasma in the processing chamber; comprising: an IEDF control mechanism that independently changes energy of ions entering the wafer and the IEDF; and a detecting mechanism that detects a plasma state with respect to a bias frequency. In this case, the IEDF control unit is composed of a power supply unit that generates a plurality of frequencies by oscillation, and each of the matching units, and the mechanism for detecting the state of the plasma includes detecting the impedance of the plasma with respect to each frequency supplied from the bias applying unit. mechanism. A plasma processing method for achieving the above object includes: mounting a project: a wafer having a film structure is placed on a 200849325 electrode under a processing chamber inside a vacuum vessel, the membrane being constructed such that the surface is made of a high-k material a film comprising a plurality of layers of a metal material and having a step structure; introducing a project to introduce an etching gas into the processing chamber; adjusting the process to adjust the processing pressure; generating a project to generate a plasma in the processing chamber; and supplying the project Providing a bias potential of 1 or a plurality of frequencies to form a bias potential on the wafer; and performing the film structure by changing the output of the bias power; and having the following features: The time change of the plasma impedance is detected by the bias applying mechanism side; the end point determining project determines the end point of the plasma processing according to the detection result; and the control project, after the end point judgment, independently controls the injection into the wafer The ion energy and its distribution. Control engineering for independently controlling the ion energy and its distribution of the above-mentioned wafers, including the output of the bias power that changes most frequencies and their mixing ratios or after the process of detecting the time variation of the impedance of the plasma, Separating into the impedance of the wall state component and the component directly above the wafer, having: comparing engineering, comparing the separated data with the database or the variation model; and performing wall cleaning according to the comparison result, or changing the second time Engineering of wafer processing conditions. [Embodiment] Hereinafter, embodiments of the present invention will be described with reference to the drawings. First, an example of a film structure which is an object of the embodiment of the present invention will be described with reference to Fig. 3 . Figure 3 (a) is a cross-sectional view of the sample before the engraving process of a planar CMOS with a metal pole/high-k structure, and Figure 3 (b) is a cross-sectional view of the conventional method after etching treatment. (c) is a cross-sectional view of the etching process of the present invention. -8- 200849325 In FIG. 3(a), the wafer is composed of: STI 308 formed on the S i substrate 309; HfSiON (high-k) insulating film 3 07 formed on the Si substrate 309 and the STI 308; a metal gate layer 3 0 6 formed on the High-k insulating film 307; a gate C ap layer 305 formed thereon in this order; a lower mask 304; an intermediate layer mask 303; BARC (anti-reflection film) a layer 3 02; and a resist 3 01 formed thereon. As shown in the figure, for a planar laminated metal/hi gh-k gate having an STI step difference 310 caused by the formation of STI 308, a gate etching apparatus using a conventional mechanism having a single bias frequency of 400 kHz is used, and Cl2 is used. When the /HBr gas is etched, as shown in FIG. 3(b), the underlayer 311 of the HfSiON (high-k) film is formed in the lower portion of the gate which becomes the active portion of the MOSFET. Further, at the corner of the STI step 310, the residual portion 3 1 3 of the STI step gate material is generated by a skirt enlargement. In addition, a lower portion of the gate portion of the field portion of the wiring portion may have a lower sag pull-down shape 314 that becomes a skirt-like shape, or a residual portion 315 of the gate material on the STI surface. . That is, in the etching process of the laminated metal/high-k gate, it is difficult to balance the improvement of the underlying high-k selectivity near the gate and the prevention of the sag pull-down shape. As shown in Fig. 3(b), the reason why the improvement of the underlying selectivity at 400 kHz bias and the prevention of the sag pull-down shape are caused by the fact that the IEDF (ion energy distribution) has diffusion centered on the average energy, and the optical on the film The etching end point of the transparent material is judged to be slow. Fig. 2 shows an example in which the energy distribution of ions is centered on the average energy of the same time as - 9 - 200849325. Figure 2 is a sample of Vpp II 200 V applied to a low frequency 400 kHz and a high frequency 13.56 MHz wafer at a plasma temperature of 3 eV, mass of injected ions of 79.9, and plasma density of 1 X ΙΟ Γ Γ Γ1. A pattern diagram of most of the IEDF examples at RF. The IEDF 203 'with respect to the time average of 100 eV 'distribution width 204 becomes about 200 eV at 400 kHz. 'There are two peaks at around 0 V and 200 V near the vicinity of 200 V. In addition, the IEDF 201 at 13.56 MHz of high frequency has a narrower distribution width 202. 50eV. The IEDF205' with a mixture of low frequency and high frequency at 100Vpp each has an intermediate distribution. This is because the higher the frequency of the high-frequency bias, the faster the period of the sheath voltage change of the accelerated ion becomes, the shorter the sheath vibration time is relative to the time the ion passes through the sheath, and the ion cannot follow the sheath change and is accelerated by the average value of the electric field. The reason. As described above, when the etching is performed at a frequency of 400 kHz, the high-energy ion of about twice the time-averaged ion energy deteriorates the selectivity, and the low-energy ion of about OeV becomes the hem shape of the hem. It becomes difficult. Therefore, in order to achieve both the selectivity and the hem pendulum shape, it is preferred to narrow the ion energy distribution at a high frequency. Further, a distribution having a large diffusion at a low frequency is used, for example, when a metamorphic layer of a surface is physically sputtered with high energy ions, or when an insulating film of high energy ions is required to be etched. In addition, by using low frequencies, high energy ions with a diffusion distribution relative to the average energy can be utilized without affecting the dissociation or distribution of the plasma.

。如上述說明,貫穿多層膜進行蝕刻時必須能控制IEDF 〇 另外,如圖3 ( b )所示形狀異常之另一原因、亦即終 -10- 200849325 點判斷變慢,係爲對應於閘極之薄膜化(約1 〇〜3 Onm ) 而必要降低蝕刻速度所引起者。亦即,於習知使用發光分 光之方法中,蝕刻速度變慢時,反應生成物之電漿中之組 成比變小,電漿中之發光強度減弱,其變化亦變小。 另外,以習知膜厚干涉方式進行終點判斷時,被蝕刻 材料之閘極厚度對於檢測用光源波長( 200〜800nm)之 1 /4波長變小而爲難以發現週期性之區域,而且閘極材料 表面粗糙度存在,或底層high-k膜薄至約2nm而呈現非 均質之情況等存在有大多數,因此干涉波形之強度受到干 涉光之視角內被蝕刻面積影響,導致附加有圖案之晶圓之 終點判斷更爲困難。 另外,本發明使用之檢測電漿之電氣特性方式,晶圓 正上方之反應生成物密度對於高電漿鞘部分之變化較敏感 ,即使初期膜厚較薄時,粗糙度存在時亦可檢測出。另外 ,關於裝置之穩定稼動,藉由檢測低頻與高頻之多數頻率 於晶圓正上方之阻抗,將電漿壁面狀態之變化與晶圓正上 方之變化加以分離而可以實施終點檢測。 通常、被施加於電漿中之偏壓以等效電路表示時,不 僅可以高頻之角速度· H、低頻之角速度· L、晶圓正上方 之鞘之阻抗(Zly、y=H、L),亦可以腔室壁面形成之 鞘之阻抗(Z 2 y、y = Η、L )所構成電氣體供給部,而以 以下之式(1 )及式(2 )表示。亦即’檢測不同頻率下之 電壓(Vly、y=H、L)、電流壓(Ily、y=H、L)、相 位之監控値,以高頻側之式(1 )、低頻側之式(2 )、阻 -11 - 200849325 抗與頻率之關係、亦即式(3 )之連立方程式爲實部’解 出虛部,則可以分離出晶圓正上方與腔室壁面之阻抗資訊. As described above, it is necessary to be able to control the IEDF when etching through the multilayer film. In addition, another cause of the shape abnormality as shown in Fig. 3(b), that is, the end point of the -10-200849325 is slow, which corresponds to the gate. Thin film formation (about 1 〇 ~ 3 Onm) and it is necessary to reduce the etching speed. That is, in the conventional method of using luminescence, when the etching rate is slow, the composition ratio in the plasma of the reaction product becomes small, and the luminescence intensity in the plasma is weakened, and the change is also small. In addition, when the end point is judged by the conventional film thickness interference method, the gate thickness of the material to be etched becomes smaller than the 1/4 wavelength of the wavelength of the detection light source (200 to 800 nm), and it is difficult to find a periodic region, and the gate is There is a large amount of surface roughness of the material, or a case where the underlying high-k film is as thin as about 2 nm and is heterogeneous. Therefore, the intensity of the interference waveform is affected by the etching area in the viewing angle of the interference light, resulting in the addition of a pattern crystal. The judgment of the end of the circle is more difficult. In addition, in the method for detecting the electrical characteristics of the plasma used in the present invention, the density of the reaction product directly above the wafer is sensitive to changes in the portion of the high-plasma sheath, and even when the initial film thickness is thin, the roughness can be detected. . In addition, with regard to the stable operation of the device, the end point detection can be performed by detecting the impedance of the low frequency and the high frequency at a frequency directly above the wafer, and separating the change in the state of the plasma wall from the change directly above the wafer. Usually, when the bias voltage applied to the plasma is represented by an equivalent circuit, not only the angular velocity of the high frequency, H, the angular velocity of the low frequency, L, and the impedance of the sheath directly above the wafer (Zly, y = H, L) The electric gas supply portion may be formed by the impedance (Z 2 y, y = Η, L) of the sheath formed by the wall surface of the chamber, and is expressed by the following formulas (1) and (2). That is, 'detect the voltage at different frequencies (Vly, y = H, L), current pressure (Ily, y = H, L), phase monitoring 値, on the high frequency side (1), low frequency side (2), resistance-11 - 200849325 The relationship between the resistance and the frequency, that is, the equation of (3) is the real part of the solution. If the imaginary part is solved, the impedance information directly above the wafer and the wall of the chamber can be separated.

=Kr"⑴=Kr"(1)

AlIl^Z2LIl^Vl -(2) 式(2)AlIl^Z2LIl^Vl -(2) Equation (2)

x=l、2 ; y =H 或 L = Rx+jXx>r-(3) 本發明乃基於,可藉由多數頻率分離出多數種阻抗加 以檢測而成者。以下說明之實施形態中,針對具有段差構 造之平面型CMSOFET之多層構造閘極進行蝕刻加工時產 生之各種形狀異常(例如疏密形狀差、缺口、側壁飩刻、 底層膜損傷、裙襬、底層貫穿、推拔形狀等)問題,藉由 使用具備偏壓施加機構的電漿處理裝置加以解決,該偏壓 施加機構具備:具有多數頻率之離子能量分布(IEDF )控 制機構;及多數之電漿阻抗檢測器。 (第1實施形態) 首先,說明具提實現本發明之裝置之實施形態。圖1 -12- 200849325 爲本發明實施形態之電漿處理裝置之構成槪略之縱斷面圖 。於圖1所示電漿處理裝置,係於真空容器內部設置之處 理室內側形成電漿。使用該電漿對該處理室內配置之半導 體晶圓等被飽刻材料、亦即基板狀試料進行處理的電漿處 理裝置。該電漿處理裝置之電漿產生機構,係具備: 4 5 0MHz之UHF電源101,附加阻抗檢測器的高速響應 UHF匹配器102,天線103,及電磁鐵104。 對構成真空容器之腔室107內放出UHF波的天線103 ,係較維持真空的石英板1 05更靠近大氣側被配置。蝕刻 氣體通過由流量控制器與止流閥構成之以數秒變化氣體流 量的高速響應反應性氣體導入機構1 1 1,混合而成蝕刻氣 體之後,由噴氣頭106導入腔室107內,蝕刻氣體中之氣 體壓力,可由位於高真空泵109上方之高速響應壓力控制 機構1 1 0加以控制。 配置有被蝕刻材料、亦即S i晶圓1 1 2的下部電極1 1 3 ,係具備覆蓋其上面、亦即Si晶圓112被載置之載置面 之外周側及側壁而配置之大略環形狀之承受器1 1 4,以2 個以上同心圓形狀、成爲不同之熱區域而構成,使用溫度 控制機構1 1 5可於各個區域控制下部電極之溫度成爲特定 。鈾刻處理中,使用直流電源1 16產生之- 2000 V〜+2000 V 直流電壓以靜電吸附Si晶圓1 1 2,於Si晶圓1 1 2與下部 電極1 1 3之間隙塡充He進行壓力控制。藉由上述靜電吸 附技術之使用而調節蝕刻中之Si晶圓1 1 2之溫度 於下部電極1 1 3連接偏壓施加機構1 1 7,用於由電漿 -13- 200849325 中將離子引入至Si晶圓1 1 2,控制其之離子能量分布。偏 壓施加機構1 1 7係由··控制射入之離子能量分布的IEDF 控制機構 1 2 7,及電榮狀態檢測器(p 1 a s m a i m p e d a n c e monitor,以下稱爲PIM) 126構成。本實施形態中,IEDF 控制機構1 27,係使用具備發送、供給多數頻率之電力的 電源者、亦即使用由低頻偏壓電源部120,低頻匹配器 1 2 1 ’高頻偏壓電源部n 8,高頻匹配器丨丨9構成者。低頻 偏壓電源部1 20之頻率使用400kHz,高頻偏壓電源部1 1 8 之頻率使用13.56MHz,分別可以輸出相當於最低約1W至 最大電力約150W (連續正弦波)/12英吋直徑,於500Hz 〜3kHz範圍內進行0n-0ff調變,使用具備時間調變( Time Modulate,以下稱TM)功能者(調變時最大電力約 1 50 W ° 此時,高頻匹配器1 1 9、低頻匹配器1 2 1需要設爲可 於最低輸出例如最大輸出之0 · 5 %以上之輸出,可由1 w之 輸出加以匹配之感度。另外,電漿狀態檢測器1 26,係藉 由將其配置於IEDF控制機構127與下部電極1 13之間, 而檢測出相對於各個頻率之電壓、電流、相位之時間變化 、亦即包含電子密度、電子溫度之時間變化的阻抗變化。 又,於本實施形態之電漿處理裝置中,接受蝕刻處理 中之電漿發光信號的發光受光部1 22,係被配置於例如構 成腔室107之容器底部。該發光受光部122之輸出,被傳 送至發光分光器123。 又,具備和下部電極1 1 3上之S i晶圓1 1 2呈對向, -14- 200849325 藉由接受電漿中、或外部光源之干涉光,而檢測Si 表面之膜厚的膜厚干涉監控器124。.於腔室107側壁 表面配置,覆蓋其而被配置、使該側壁之溫度於電漿 中、處理前後分別調節爲適當値的加熱器1 0 8。 又,於本實施形態之電漿處理裝置中具備控制 1 25,可依據事前設定之値而控制其時序,或接收來 測各部之動作的感測器之輸出,使用彼等之接收信號 果進行運算,或由記憶裝置讀出資料發出各部之動作 。控制裝置125構成爲,可以和UHF電源101、附加 檢測器的高速響應UHF匹配器102、加熱器108、高 泵1 〇9、高速響應壓力控制機構1 1 0、高速響應反應 體導入機構111、溫度控制機構115、IEDF控制機構 及配置於其內部的高頻偏壓電源部1 1 8等,以及發光 器123、膜厚干涉監控器124、電漿狀態檢測器126 通信,接受彼等之輸出,對彼等各部發出動作信號。 接受發光分光器123、膜厚千涉監控器124、電漿狀 測器126之輸出,檢測Si晶圓之膜厚或處理之終點 理狀態,依據檢測結果對1 27、UHF電源1 01、附加 檢測器的高速響應UHF匹配器102、加熱器108、高 泵1 09、高速響應壓力控制機構1 1 〇、高速響應反應 體導入機構1 1 1、溫度控制機構1 1 5算出動作信號, 信號調節其動作。附加阻抗檢測器的高速響應UHF 器102係作爲對UHF電源101記錄多數以上之匹配 選擇其匹配路徑進行匹配的匹配裝置之功能。 晶圓 外周 處理 裝置 自檢 之結 指令 阻抗 真空 性氣 127 分光 進行 例如 態檢 等處 阻抗 真空 性氣 發出 匹配 點, -15- 200849325 以下依據圖4之時序說明,使用圖1之電漿處理裝置 ,針對圖3 ( a )所示,具有 STI段差310及阻劑 301/BARC層3 02/中間層遮罩3 03/下層遮罩3 04/閘極Cap 層3 0 5 /金屬閘極層3 06/HfSiON絕緣膜3 07構造的平面型 MOS之,10〜30nm之薄的金屬閘極層3 06 ( TiN)部分, 使用本發明進行蝕刻之例。 自BARC層302至閘極Cap層305係於同一腔室內進 行蝕刻處理之後,實施TiN (金屬閘極層)3 06之貫穿( break turough,以下稱BT )步驟。該BT步驟目的在於除 去,BARC層302之過蝕刻(over etching,以下有可能以 0E表示)時和氧氣體結合產生之阻礙TiN蝕刻的Ti0表 面層。BT步驟之條件,係使用 ArlOO〜200cc、壓力iPa 、UHF施加電力500W、400kHz、偏壓低頻施加電力5〇〜 100W。藉由400kHz低頻偏壓之使用,容易獲得Ti0除去 必要之高能量。 如圖4 ( a )所示爲,藉由電漿狀態檢測器126內之 400KHz用之電漿阻抗監控,檢測出之電壓、電流、相位 結果,藉由阻抗之時間變化而進行之終點判斷(EPD )。 隨TiN之BT步驟進行,表面之TiO被除去,Ti、N等被 放出至電漿鞘附近,引起電子密度、鞘厚度等之電氣特性 變化。在阻抗變大之時刻40 1,移至主蝕刻(以下有可能 以ME表示)。 在ΊΊΝ306之ME步驟,係於以Ch或HC1爲基礎添 加 HBr、NF3、CF4、SF6,壓力 〇.2Pa、UHF 施加電力 -16 - 200849325 500W條件下進行。TiN3 06之ME步驟之偏壓 400kHz低頻偏壓混合50%之13_56MHz高頻偏壓 4 ( b ) ,TiN之ME時之2頻率混合比403 )。 ,可獲得高頻偏壓施加優點之於窄分布寬度下抑 離子引起之裙襬硬塊之同時,可獲得低頻偏壓施 高能量離子引起之異方性,以及可將腔室壁面形 晶圓正上方形成之鞘之阻抗加以分離,可藉由晶 之阻抗進行高精確度之終點判斷。ME步驟之終 係於電漿狀態檢測器1 2 6檢測出之電漿阻抗開始 402,切換爲TiN之OE1步驟,設定IEDF控制 之輸出404爲高頻偏壓電力1〇〇 %之處理。此乃 次一 OE1步驟,需要統合STI段差、膜厚、電路 密形狀差或被蝕刻材料差、晶圓面內差引起之不 量’因此在到達底層膜之圖案,能維持其和底層 之選擇性之同時,需要蝕刻未到達底層之部分。 此時,藉由使用電漿狀態檢測器1 26之信號 圖4 ( c )所示藉由習知檢測反應生成物或触刻劑 値之時間變化方式而獲得的終點406,約快1〜5 得ME之終點。 之後,於電漿狀態檢測器1 26之信號穩定之 OE2步驟,〇E2步驟之電漿條件係使用,HBr/02 、全流量200〜400cc、壓力3〜10Pa、UHF電源 出 5 00〜700W。使用高頻偏壓電力 80%、低頻 2 0%之混合偏壓電力(於圖4 ( c) ,TiN之OE2 係使用, 者(於圖 其理由爲 制低能量 加優點之 成之鞘與 圓正上方 點判斷, 變化時刻 機構1 2 7 因爲’於 圖案之疏 同之蝕刻 high-k 間 ,相較於 之發光峰 秒可以取 時刻移至 、Ar稀釋 1 〇 1之輸 偏壓電力 時之2頻 -17- 200849325 率混合比4 0 5 )。其理由爲,以減少底層膜選擇性 段差部分或非開口部分之下擺下拉形狀之同時’可 子陰影損失(electron shading damage)。亦即, 合些許平均能量之接近倍數的高能量離子之比率, 顧下擺下拉形狀之減輕之同時,可維持選擇性,可 子衝撞引起之物理損傷、亦即可減輕源極/汲極之 另外,使遮罩上部帶電之電子和低能量離子中和, ,可以迴避電子陰影損失引起之缺口。 以上,藉由本發明之適用,於如圖3 ( c )所示 極加工,可以在不產生底層貫穿情況下,進行垂直 除本實施形態之閘極構造以外,其他之hi gh-k Zr02 、 Υ2〇3 、 La203 、 LaA10x 、 LaSiOx 、 A1203 、 HfAlO(N)等絕緣材料,金屬閘極層除TiN以外, TaN、TaSiN、TaC、Ru、HfN、MoN 等金屬材料之 亦可獲得同樣效果。 上述實施形態中,IEDF控制機構127係使用 產生2個不同頻率之電源部118、120及匹配器11 構成之機構。和其他之IEDF控制例如陷波偏i bias )比較,此方式具有振盪器及匹配器構成簡單 。偏壓施加機構1 1 7使用陷波偏壓電源時,和連續 頻率之重疊相同之故,雖價位高,但可以具備具有 頻率範圍的電漿狀態檢測器1 26加以對應。 又,混合比率之控制,本實施形態中係以成爲 子的電壓側量用之VPP設爲指標,但不僅Vpp,亦 與STI 減少電 藉由混 可以兼 抑制離 凹陷。 如此則 金屬閘 加工。 材料如 Hf02、 其他如 情況下 由振盪 [9 、 121 I ( Clip 之優點 之多數 連續之 加速離 可以輸 -18- 200849325 出電力之比率進行。此時,輸出電力成爲電流與電壓之乘 算,Vpp會因接地之配置或面積等而不同,因此需要考慮 此點。或者亦可依據如圖2所示計算模型,控制各電源之 輸出而使成爲所要之能量分布。此時,可以具備能量分布 計算用之電漿密度、電子溫度等之檢測機構而提升控制精 確度。 又,IEDF控制之偏壓電源,高頻側係使用13.56MHz ,低頻側係使用400kHz,基本上2個不同頻率之差較大 者IEDF之控制範圍較廣,另外,就腔室壁面與晶圓正上 方之阻抗之分離觀點而言亦較好。另外,較好是互相不成 爲整數倍,俾能活用個別頻率之高諧波。此時,爲維持和 電漿產生之間之獨立性或良好之面內分布,較好是高頻側 之頻率低於電漿產生機構之頻率。例如ECR時,100 MHz 以上時離子能量與電漿密度之獨立控制性變爲困難,因此 較好是4MHz以上100 MHz以下。另外,低頻側之頻率低 於100kHz時Si上之絕緣層容易產生充電上升(charge up )較爲不佳。因此較好是低頻側之頻率爲100kHz以上、 小於4MHz,高頻側之頻率爲 1 MHz以上、小於 100MHz 之頻率之中,儘可能組合頻率差較大者。另外,混合之頻 帶亦受電漿產生機構之影響。例如如上述說明,在分布控 制使用磁場的電漿產生機構之中,考慮E X B之交叉阻抗影 響而使用13.56MHz之高頻。ICP、CCP等則調整和電漿來 源頻率之間之平衡,可以使用27.60MHz。 -19- 200849325 (第2實施形態) 於弟1貫施形知假設壁面狀態之阻抗不變,以晶圓單 體處理之阻抗之絕値來獲得終點,但於晶圓之量產加工 之中存在腔室壁面狀態引起之隨時間變化。以下說明分離 該腔室壁面狀態而獲得終點之實施形態。於第1實施形態 ’以T i N之Μ E中之2頻率混合比5 0 %進行終點判斷時, 係如圖5之終點判斷流程所示,於檢測多數頻率對應之阻 抗之檢測工程5 0 1之後’進行分離工程5 0 2、亦即依據上 述式(1)、式(2)、式(3)將壁面與晶圓正上方之個 別之阻抗加以分離,進行終點判斷工程5 0 3,進行變化工 程504、亦即變化IEDF控制機構127、UHF電源101、高 速響應壓力控制機構1 1 0、高速響應反應性氣體導入機構 1 1 1、溫度控制機構1 1 5之輸出。此時使用之終點判斷模 式圖如圖6所示。 圖6 ( a )爲,除分離工程後之壁面部分之阻抗變化以 外,將晶圓正上方之阻抗變化,描繪於多數阻抗而成之導 納(immitance )上之圖。例如考慮終點判斷前之阻抗602 移至飩刻膜被完全除去時之阻抗6 0 1時,藉由依存於匹配 器之調諧用電容器及調諧用線圈,採用匹配路徑603時, 係於通過設定之多數阻抗範圍6 0 4之點6 0 5之時刻實施終 點判斷。 另外’圖6 ( b )爲,將習知以單一偏壓頻率檢測出之 阻抗變化,描繪於導納(immitance )上之模式圖。存在 時間變化時,隨蝕刻處理時間增加,腔室壁面部分之阻抗 -20- 200849325 變化606,在終點判斷以前會超過設定之阻抗範圍6〇4, 終點判斷前變爲不存在。針對此現象,依據本發明之多數 頻率之阻抗檢測,藉由實施壁面狀態成份與晶圓正上方成 份之分離工程的方法,可於不影響壁面狀態下,穩定取得 晶圓之終點,藉由和IEDF控制之組合,可於穩定獲得無 底層貫穿、亦不會產生下擺下拉形狀之金屬閘極/high-k 構造之鈾刻形狀。 圖5之流程,係依據控制裝置1 25 .內之控制程式而實 施,或記載於電漿狀態檢測器1 26之控制軟體而實施。本 實施形態中,雖針對如圖3 ( a )所示斷面構造樣本之各層 之中、對裝置特性之衝擊大的TiN (金屬閘極層)3 06加 以記述,但亦可使用於各BARC層3 02、中間層遮罩303 、下層遮罩3 04、閘極Cap層3 0 5等各層之終點判斷。 (第3實施形態) 以下爲使用分離出的腔室壁面與晶圓正上方之阻抗變 化之實施形態。實施多數片圖3 ( a )之金屬閘極/high-k 閘極之飩刻處理時之形成之ME終點判斷步驟之流程圖示 於圖7。於第2實施形態之要領下,於檢測多數頻率對應 之阻抗之檢測工程5 0 1之後,進行分離工程5 0 2、亦即依 據上述式(1)、式(2)、式(3)將壁面與晶圓正上方 之個別之阻抗加以分離,進行比較工程70 1、亦即將分離 出之腔室壁面之阻抗與晶圓正上方之阻抗,和過去之同一 步驟之狀態(過去之阻抗變化等之資料、變動模型)加以 -21 - 200849325 比較。 於比較工程70 1,係將現在進行中之阻抗或電流、電 壓、相位記錄於資料庫,和過去之資料加以比較,分類爲 每一片晶圓之時間變化。依據該分類而判斷腔室壁面狀態 ,進行調整工程702、亦即進行壁面狀態重置處理或以正 變化的方式適當調整各裝置輸出之値,如此則,分離出之 阻抗資訊可以活用於時間變化之抑制。 例如於該分類時,僅高頻側之電壓、電流、相位或阻 抗極端變化時,判斷爲晶圓面內之偏壓均勻性引起變化, 於次一晶圓處理,藉由調整偏壓之面內分布而使偏壓之面 內分布成爲均勻,如此則,可回授控制偏壓之均勻性,可 抑制良品率之降低。偏壓之均勻性補正方法,可藉由調整 2頻率偏壓之混合比而實現。例如低頻(400kHz : VppL) 與高頻(13·56ΜΗζ : VppH )之混合比(VppH/ ( VppH + VppL ))變化爲 0%、20%、1 00%時,多晶砂触刻 速率之分布變化如圖8 ( a ) 、( b ) 、( c )所示。鈾刻條 件爲··處理氣體使用HBr/02、壓力爲3Pa、UHF施加電力 爲 500W。 作爲偏壓電力設定1 3 · 5 6 M H z之混合比爲0 %時,多晶 矽飩刻速率之中高分布爲1 1 % (圖8 ( a ))。作爲偏壓電 力設定13.56MHz之混合比爲20%時,多晶矽蝕刻速率之 中高分布爲〇% (圖8 ( b ))。作爲偏壓電力設定 1 3.5 6 Μ Η z之混合比爲1 〇 〇 %時,多晶砂飩刻速率之中高分 布爲 12% (圖 8(c))。亦即,隨高頻偏壓電力 -22- 200849325 13.56MHz之混合比之增加,晶圓端部之分布上升,可以 混合比97%控制端部高度之1 2%及以偏壓混合比控制晶圓 之分布。 圖8(d))爲,高頻13·56ΜΗζ與低頻400kHz之監 控電壓VH、VL之混合比,以VH + VL成爲大略一定的方 式而變化時,測定200mm晶圓上之多晶矽蝕刻速率801、 Si02蝕刻速率8 02之結果。Si02蝕刻速率802係以10倍 表示。隨高頻偏壓電力13.56MHz之混合比之增加,多晶 矽飩刻速率大約以一定之2成減少,相對於此,Si02蝕刻 速率則急速減少,對氧化膜之選擇性(選擇比)803變爲 增加。如上述說明,變化偏壓之混合比而控制晶圓面內速 率時,以混合比3 0%以上控制時,於OE成爲耐性夠之高 選擇性區域(選擇比200之線804以上),而較好。其他 之鈾刻速率或形狀之面內分布之修正手段,可藉由組合電 磁鐵104之輸出或氣體流量分布、晶圓平台之面內分布等 加以控制。 又,隨晶圓處理片數增加,於同一飩刻步驟僅腔室壁 面之低頻阻抗與高頻阻抗變化時,若腔室壁面之狀態變化 ,則由觀察者判斷,或以控制裝置1 25判斷,自動或手動 進行電漿處理裝置之潔淨處理,促使元件之交換,依據變 動之補正模型而變化IEDF控制機構127、UHF電源101、 高速響應壓力控制機構1 1 〇、高速響應反應性氣體導入機 構1 1 1、溫度控制機構1 1 5之輸出。 以上之例係以腔室壁面之阻抗或晶圓正上方之阻抗, -23- 200849325 或低頻偏壓之阻抗或高頻偏壓之阻抗,而分類裝置狀態或 晶圓處理狀態。另外亦可考慮變化方向(例如電感側或電 導側)之資訊,進行更詳細分類及資料庫建構及變動要因 抽出,可依據該變動要因實施適當之回授控制。 如上述說明,在對形狀加工帶來影響之飩刻步驟( ME步驟或終點判斷步驟),使用2頻率混合之IEDF控制 偏壓,計測其阻抗、加以記錄,和過去之資料庫或變動模 型比較,如此則,可對蝕刻特性變化自動補正。 具體實現此方法之機構如圖94所示,控制裝置1 25 連接資料庫901,對控制裝置125之輸入信號,不局限於 第1實施形態、第2實施形態之偏壓施加機構1 1 7內之電 漿狀態檢測器1 26之輸出。亦即,對控制裝置1 25之輸入 信號,可使用習知電漿發光分光器123之輸出(發光光譜 )或附加阻抗檢測器的高速響應UHF匹配器1 02 (由電漿 頻率看到之電漿之阻抗、電壓、電流、相位)或膜厚干涉 監控器124之輸出,以及其他機構例如、UHF電源101、 電磁鐵1 04、加熱器1 0 8、高速響應壓力控制機構1 1 〇、高 速響應反應性氣體導入機構1 1 1、溫度控制機構1 1 5、直 流電源1 1 6之各監控値。控制裝置1 25記憶各個輸入資料 庫,控制裝置1 25係比較、參照資料庫90 1內記錄之過去 資料與新輸入之資料,依據參照資料庫901內或控制裝置 125內之變動模型及比對結果,而輸出IEDF控制機構127 或各控制裝置之信號。如上述說明,不僅本發明多數頻率 對應之電獎狀態檢測器1 2 6之資訊,亦參照發光光譜等之 -24- 200849325 各控制機構之監控値,因此可實施更詳細之裝置狀態、晶 圓處理狀態之分類可實施適當之處置。 (第4實施形態) 以下參照圖1 〇說明,針對具有圖1 〇 ( a )所示平面型 ArF阻劑301/BARC層3 02/TEOS (硬質遮罩)1001/多晶 矽 1 002/HfSi023 07被積層於 Si基板 3 09上之構造的 CMOS閘極,使用本發明進行加工時之實施形態。 圖10 (a)所示爲具有包含硬質遮罩之多層膜遮罩構 造的蝕刻前之斷面形狀。圖1 〇 ( b )爲習知方法之高頻偏 壓進行硬質遮罩蝕刻後之斷面形狀,圖1 〇 ( c )爲習知方 法之高頻偏壓進行閘極蝕刻後之斷面形狀,圖1 〇 ( d )爲 本發明進行触刻處理後之斷面形狀。 針對圖10 ( a)所示具有BARC層3 02之下方配置硬 質遮罩1001,及其下方配置多晶矽閘極層1 002之斷面構 造的Si晶圓,使用26MHz或13.56MHz之較高頻習知記 述(單一偏壓頻率電源),蝕刻硬質遮罩1 001時之例, 係被圖示於圖10(b)。如該圖所示,於硬質遮罩1〇〇1之 部分附著側壁保護膜1 004,成爲推拔形狀1 003。另外, 如圖 1〇 ( c )所示,於多晶矽閘極層 1 002之下部與 HfSi ON絕緣膜3 07之境界,於密圖案側產生缺口 1〇〇5, 且底層之Hf Si ON絕緣膜3 07之一部分絕緣被破壞。 以下說明,使用本發明之電漿狀態檢測器126及 IEDF控制偏壓,針對具有圖1〇 ( a)所示斷面構造的12 -25- 200849325 英吋晶圓進行蝕刻之例。於圖1之蝕刻裝置之 1 13設置具有圖10 ( a)所示斷面構造的Si晶圓 微調及蝕刻處理。 使用圖1 1說明微調或蝕刻處理時之施加2 偏壓之動作。圖1 1爲圖1之實施形態之電漿處 行之處理動作伴隨之時間變化圖。該分布圖係以 軸、時間爲橫軸之時序圖。又,本圖中,圖1之 裝置之動作促發,係利用使用第1實施形態、第 態之本發明的終點判斷方法,或使用來自電漿之 器123之輸出波形,圖11之發光強度之縱軸成 〇 圖11(b)分別表示2頻率偏壓之ON/OFF 頻側之偏壓電力之混合比。 於BARC層3 02之ME步驟,係於02/Ar氣 素系(CF4、CHF3、CH2F2、CH2C12、Cl2、HBr、 體,全體氣體流量調整爲 100〜400 cc、壓力爲 l〇Pa,使用5 00W〜800W之UHF電源輸出產生之 12英吋晶圓之面內分布藉由多數電磁鐵104 控制。電漿著火後,如圖1 1 ( b )所示,400kHz 壓電源部120施加約30W〜50W,微調阻劑301 蝕刻BARC層3 02 ( BARC之ME處理)。此時 之高頻未被施加,高頻偏壓之混合比爲〇%。 藉由使用400kHz之低頻IEDF,低離子亦容 壁部分,可以有效實施等方微調,另外,藉由高 下部電極 1 2,進行 頻率混合 理裝置進 動作爲縱 電漿處理 2實施形 發光分光 爲相對値 狀態及筒 體添加鹵 HC1 )氣 0 · 8 P a 〜 漿。 之電流値 之低頻偏 之同時, 13.56MHz 易射入側 能量離子 -26- 200849325 可削去側壁之線邊緣粗縫度(L E R : L i n e E d g e R 〇 u g h n e s s )’可兼顧減少之效果。此時之微調量,可以藉由氣體之 混合比、或壓力、電漿電源電力(UHF電源101之輸出) 、下部電極之溫度、DE時間加以適當控制。BARC之ME 處理之終點1 101,係使用電漿中之CN3 87nm之發光強度 之變極點。 檢測出BARC之ME處理之終點後,進行BARC之過 鈾刻(OE )。於OE中,爲提升選擇性,降低偏壓電力約 10W,另外,切換爲13.56MHz之高頻偏壓,低頻偏壓〇, 係設定高頻偏壓之混合比爲100% ( BARC之OE處理)。 之後,於硬質遮罩1〇〇1之ME步驟,由SF6、CF4、 CHF3、CH2F2、02、Ar、He之氣體之中混合適當之氣體, 全體氣體流量設爲100〜400cc、壓力爲〇.4Pa〜1.5Pa,使 用5 00W〜800W之UHF輸出來產生電漿,下部電極偏壓 電力設爲80W〜150W而進行蝕刻(硬質遮罩1〇〇1之ME 處理)。偏壓電力之頻率,係以具有高能量峰値之成爲廣 範圍之IEDF的頻率偏壓電源部之輸出,以混合比1〇〇%下 使用。其理由爲,即使於TEOS蝕刻必要之高電力情況下 於晶圓附近抑制反應生成物之再度解離。藉由抑制該再度 解離,可加工成爲垂直(圖10(d)),而非如圖10(b )所示具有推拔形狀1 003、存在疏密形狀差之硬質遮罩。 使用該低頻偏壓、高電力時,晶圓上同樣於接地部分會產 生離子衝擊,被削去之含接地材料的反應生成物會附著於 石英表面,而引起蝕刻特性之時間變化或異物之產生,此 -27- 200849325 情況下,可於不產生過度解離、維持面內加工尺寸之範圍 內,增加高頻混合比而減輕接地之濺鍍。 於硬質遮罩蝕刻之終點1102,切換爲OE,於OE步 驟,以提升和底層多晶矽1 002間之選擇性爲目的而切換 爲高頻100%,處理時間設定成爲和STI段差相當之部分 (硬質遮罩之OE處理)。 之後,開始多晶矽1 002之貫穿(BT)步驟,多晶矽 1 002之BT處理之電漿條件設爲:氣體係使用Cl2、HBr、 02、Ar或He等之氣體單體或彼等之混合氣體,全流量設 爲 200〜300cc,壓力 0.4〜0.8Pa、UHF輸出爲 500W〜 700W。IEDF控制機構127之條件設爲:400kHz低頻偏壓 混合100 %施加(多晶矽1 002之BT處理)。此乃爲,藉 由高能量離子除去圖案底面碳系物質、氧化物質,藉由低 能量離子亦容易除去側壁保護膜。於多晶矽1 002之蝕刻 處理,其加工形狀對裝置特性有大影響,因此藉由第1實 施形態、第2實施形態記載之本發明方法進行終點判斷。 多晶矽1 002之ME係使用13.56MHz之高頻比率50% 之IEDF (多晶矽1 002之ME處理)。多晶矽1〇〇2之ME 之終點11 04,係依據第2實施形態之方法取得電漿狀態檢 測器126之信號。 檢測出多晶矽1 〇 〇 2之Μ E處理之終點後,進行〇 E處 理。該〇 Ε處理,係於Ο Ε1及Ο Ε 2之2步驟進行。於〇 ε 1 步驟,爲兼顧底層選擇性與蝕刻形狀(垂直加工性、無裙 襬形狀),較好是低壓條件。於低壓條件下欲得底層選擇 -28 - 200849325 性’通常需要降低離子能量,但於400kHz之低頻,因設 爲低偏壓而存在低能量離子,離子指向性更降低,裙襬形 狀容易殘留。因此,於OE1步驟,使用和ME相同之處理 氣體,使用13.56MHz之高頻偏壓100%混合之IEDF (多 晶矽1002之 OE1處理)。偏壓電力爲10W〜50W( Vpp250V 以下)。 之後,移至OE2步驟,除去STI段差部分或面內、 p/n閘極差、疏密形狀差引起之多晶矽。OE2步驟之電漿 條件爲,處理氣體以Ar稀釋HBr/02,全流量設爲200〜 400cc,壓力 3 〜10Pa、UHF 輸出 500W 〜700W。使用 2 0% 之低頻偏壓混合80 %之高頻偏壓之2頻率混合偏壓(多晶 矽1 002之OE2處理)。此乃爲兼顧底層選擇性與電子陰 影損失(electron shading damage)之減少。亦即,藉由 減少平均離子能量之近倍之高離子能量之比率,可維持底 層選擇性,可抑制離子衝擊引起之物理損傷。又,以低能 量之離子中和遮罩上部帶電之電子,可迴避電子陰影引起 之缺口。 以上說明多層膜構造之樣本,被蝕刻材料改變時,以 電漿狀態檢測器126、發光分光器123或膜厚干涉監控器 124之信號作爲促發移至次一步驟時之IEDF控制機構127 之設定例。圖爲步驟切換時,IEDF控制電源之輸出電 壓(Vpp)產生上沖(overshoot)使過剩電壓施加於被倉虫 刻材料,不會降低選擇性之控制裝置構成之槪略圖。 圖1 5爲圖1所示實施形態之電漿處理裝置之控制裝 -29- 200849325 置構成之槪略模式圖。依據事先設定之蝕刻處理條件( recipe),藉由電漿電源(UHF電源101)、高速響應壓 力控制機構1 1 0、高速響應反應性氣體導入機構1 1 1之輸 出値或時序控制的控制裝置125,進行IEDF控制機構127 之輸出電壓與時序控制。於IEDF控制機構127,至遲於 弟X -1編5虎触刻步驟’將第X編號之局頻偏壓電源部1 1 8 之輸出設定値SVHx與低頻偏壓電源部120之輸出設定値 SVLx,或第X編號步驟之高頻匹配器119之穩定點設定 値SPHx與低頻匹配器121之穩定點設定値SPLx,依據控 制裝置1 2 5之信號事先設定。 此時,亦適當設定多數個自第X-1編號結束至第X編 號之高頻匹配器1 19之穩定點的過渡設定値SPHy與低頻 匹配器1 2 1之穩定點的過渡設定値SPLy,如此則,移至 第X編號步驟時,不會產生過剩電壓,可圓滑響應。此時 ,依據移行中之監控値之MPHx、MPLx (來自負荷之射入 電力或反射電力、匹配狀態等),輸出過渡設定値SPHy 與SPLy或電源輸出値SVHx與SVLx,則可進行回授控制 。另外,不僅監控高頻匹配器119、低頻匹配器121之信 號,亦可監控電漿狀態檢測器1 26之電流、電壓、相位、 阻抗信號而進行回授控制,亦可參照資料庫90 1之過去値 。如上述說明,係對次一目標點,於最短時間圓滑、單調 地變化時間之機構及方法,但和IEDF控制機構127之過 渡現象控制同樣,亦可對UHF電源1 0 1、高速響應壓力控 制機構1 1 0、高速響應反應性氣體導入機構11 1、溫度控 -30- 200849325 制機構1 1 5同樣實施。 (第5實施形態) 以下說明使用圖1之電漿處理裝置,形成具有圖1 2 之高段差立體構造FIN-FET時之處理方法之實施形態。圖 中(a )爲蝕刻前之膜構造斜視圖,(b )爲使用本實施形 態蝕刻後之膜構造斜視圖,(c )爲圖(a )之A-A’線切斷 的縱斷面圖。 於圖12(a)及(c),具有高段差立體構造FIN-FET 之蝕刻前之膜構造,係由形成於Si基板3 09上的Si02層 1203、Si 層 1202、TiN 層 1204、BARC 層 302 及阻劑 30 1 構成。於Si02層1 203之上形成FIN部分1201,再於 Si02層1 203及FIN部分1201之上形成high-k絕緣膜 1 205。1 20 8表示FIN部分1201與閘極之境界的部分。本 實施形態中,係由圖1 2 ( a )之膜構造以阻劑3 0 1爲遮罩 開始鈾刻,而獲得圖1 2 ( b )之膜構造的乾蝕刻。於圖12 (b ),1 205 爲 high-k 絕緣膜,1 206 爲蝕刻 TiN 層 1204 而形成之TiN閘極。Si層1 202具備位於TiN閘極1206 兩側、朝其延伸的多數FIN部分1201。圖12 ( c)爲具備 F IN部分1 2 0 1之圖1 2 ( a )之膜構造的縱斷面圖。 具備圖12(a)之膜構造的晶圓,其之FIN部分1201 之段差之上部與下部之〇E量差較大,對裝置特性帶來大 影響的閘極長度部分成爲垂直方向,因此不僅第1實施形 態一第4實施形態之IEDF控制,和氣體、壓力、匹配器 -31 - 200849325 、電極溫度連動而和反應生成物或蝕刻劑或離子之磁通量 關連之步驟間之過渡現象之控制亦重要。 和該蝕刻處理之過渡現象從動的機構,較好是於UUP 匹配器102使用,電漿射入電力無上沖(overshoot )或下 沖(undershoot),具備約於1秒可使單調遞增或單調遞 減變化穩定性能者。欲實現該性能時,可藉由例如設定著 火及穩定點之多數匹配參數及其匹配路徑之最佳値等功能 而實現。定義單調遞增之響應曲線之時間微分常時爲正, 單調遞減之響應曲線之時間微分常時爲負。 另外,欲和反應生成物或蝕刻劑之過渡現象從動時, 較好是於高速響應反應性氣體導入機構1 1 1使用,於電漿 放電中添加/減少新氣體時,流量不會有溢流(overshoot )或不足(undershoot )之機構,例如可使用氣體不致於 滯留而於等待時間繼續流入氣體之機構。 另外,欲和氣體流量變化伴隨之電漿壓力變動從動時 ,較好是具備約2秒可穩定壓力的高速響應壓力控制機構 1 1 〇。此功能可藉由設爲儘可能不產生壓力差之氣體配管 構造,或使壓力控制之運算法則最佳化而實現。另外,欲 和反應生成物之Si晶圓面內分布之變化從動時,下部電 極1 1 3較好是具備儘可能以高速(1 °C /秒以上)於步驟間 升降溫度,而且於分割內側、外側、或其以上之部分具有 可獨立控制之功能。此功能可藉由例如具備溫度控制機構 1 1 5而實現,該溫度控制機構1 1 5係於下部電極1 1 3內部 具備加熱器、溫度感測器、He氣體壓力控制等。 -32- 200849325x = l, 2; y = H or L = Rx + jXx > r - (3) The present invention is based on the fact that a plurality of impedances can be separated by a plurality of frequencies. In the embodiment described below, various shape abnormalities (for example, poorly formed shape, notch, sidewall engraving, underlying film damage, skirt, bottom layer) generated when etching a multilayered gate of a planar CMSOFET having a stepped structure are performed. The problem of penetrating, pushing out, etc. is solved by using a plasma processing apparatus having a bias applying mechanism having: an ion energy distribution (IEDF) control mechanism having a plurality of frequencies; and a majority of plasma Impedance detector. (First Embodiment) First, an embodiment of an apparatus for carrying out the present invention will be described. Fig. 1 -12 - 200849325 is a schematic longitudinal sectional view showing a configuration of a plasma processing apparatus according to an embodiment of the present invention. In the plasma processing apparatus shown in Fig. 1, a plasma is formed on the inside of the inside of the vacuum vessel. A plasma processing apparatus for treating a saturated material, that is, a substrate-like sample, such as a semiconductor wafer disposed in the processing chamber, using the plasma. The plasma generating mechanism of the plasma processing apparatus includes a UHF power supply 101 of 450 MHz, a high-speed response UHF matching unit 102 with an impedance detector, an antenna 103, and an electromagnet 104. The antenna 103 that emits UHF waves in the chamber 107 constituting the vacuum vessel is disposed closer to the atmosphere than the quartz plate 105 that maintains the vacuum. The etching gas is mixed with the high-speed reactive gas introduction mechanism 1 1 1 composed of a flow rate controller and a check valve by a gas flow rate of several seconds to form an etching gas, and then introduced into the chamber 107 by the gas jet head 106, and is etched into the gas. The gas pressure can be controlled by a high speed response pressure control mechanism 110 located above the high vacuum pump 109. The lower electrode 1 1 3 in which the material to be etched, that is, the S i wafer 1 1 2 is disposed, is provided with a rough surface covering the upper surface side and the side wall of the mounting surface on which the Si wafer 112 is placed. The ring-shaped susceptor 1 14 is composed of two or more concentric shapes and different heat regions, and the temperature control mechanism 115 can control the temperature of the lower electrode in each region to be specific. In the uranium engraving process, a DC voltage of 2,000 V to +2000 V DC voltage is used to electrostatically adsorb the Si wafer 1 1 2, and the He wafer is filled with a gap between the Si wafer 1 1 2 and the lower electrode 1 1 3 . Pressure control. Adjusting the temperature of the etched Si wafer 1 1 2 by the use of the above electrostatic adsorption technique to the lower electrode 1 1 3 is connected to the bias applying mechanism 1 1 7 for introducing ions into the plasma-13-200849325 The Si wafer 1 1 2 controls the ion energy distribution thereof. The biasing mechanism 1 1 7 is composed of an IEDF control unit 1 2 7 that controls the ion energy distribution of the incident, and a gyro state detector (hereinafter referred to as PIM) 126. In the present embodiment, the IEDF control unit 127 uses a power source that transmits and supplies power of a plurality of frequencies, that is, a low-frequency bias power supply unit 120, and a low-frequency matching unit 1 2 1 'high-frequency bias power supply unit n. 8, the high frequency matcher 丨丨 9 constitutes. The frequency of the low-frequency bias power supply unit 20 is 400 kHz, and the frequency of the high-frequency bias power supply unit 1 18 is used. 56MHz, respectively, can output the equivalent of about 1W to the maximum power of about 150W (continuous sine wave) / 12 inches diameter, 0n-0ff modulation in the range of 500Hz ~ 3kHz, using time modulation (Time Modulate, hereinafter TM) function (maximum power of about 1 50 W ° during modulation) At this time, the high frequency matcher 1 19 and the low frequency matcher 1 2 1 need to be set to output at the lowest output, for example, the maximum output of 0 · 5 % or more The sensitivity can be matched by the output of 1 w. In addition, the plasma state detector 1 26 detects the voltage with respect to each frequency by disposing it between the IEDF control mechanism 127 and the lower electrode 1 13 , The time change of the current and the phase, that is, the change in the impedance including the change in the electron density and the electron temperature. In the plasma processing apparatus of the present embodiment, the light-emitting and light-receiving unit that receives the plasma light-emitting signal in the etching process is used. It is disposed, for example, at the bottom of the container constituting the chamber 107. The output of the illuminating light receiving unit 122 is transmitted to the illuminating beam splitter 123. Further, it is provided with the S i wafer 1 1 2 on the lower electrode 1 1 3 To, -1 4- 200849325 A film thickness interference monitor 124 for detecting the film thickness of the Si surface by receiving interference light in the plasma or an external light source. It is disposed on the surface of the side wall of the chamber 107, and is disposed so as to be disposed so that the temperature of the side wall is adjusted to a suitable temperature in the plasma before and after the treatment. Further, in the plasma processing apparatus according to the present embodiment, the control unit 125 is provided, and the timing can be controlled according to the preset setting, or the output of the sensor for measuring the operation of each unit can be received, and the received signal can be used. The operation of the various parts is performed by the operation or reading of the data by the memory device. The control device 125 is configured to be connected to the UHF power source 101, the high-speed response UHF matcher 102 of the additional detector, the heater 108, the high pump 1 〇9, the high-speed response pressure control mechanism 110, the high-speed response reactant introduction mechanism 111, The temperature control unit 115, the IEDF control unit, and the high-frequency bias power supply unit 1 1 8 disposed therein, and the illuminator 123, the film thickness interference monitor 124, and the plasma state detector 126 communicate and receive their outputs. , issued an action signal to each of them. Receiving the output of the illuminating beam splitter 123, the film thickness monitor 124, and the plasma detector 126, detecting the film thickness of the Si wafer or the end state of the processing, according to the detection result, 1 27, UHF power supply 01, additional High-speed response of the detector UHF matching device 102, heater 108, high pump 109, high-speed response pressure control mechanism 1 1 〇, high-speed response reaction body introduction mechanism 1 1 1 , temperature control mechanism 1 1 5 calculation of operation signal, signal adjustment Its action. The high-speed response UHF 102 of the additional impedance detector functions as a matching device that records a majority of the matching of the UHF power source 101 and selects its matching path to match. Wafer peripheral processing device self-test knot command impedance vacuum gas 127 splitting, for example, state detection, etc., impedance vacuum gas emission matching point, -15- 200849325 Hereinafter, according to the timing diagram of FIG. 4, the plasma processing apparatus of FIG. 1 is used. , as shown in Figure 3 (a), with STI step difference 310 and resist 301 / BARC layer 032 / intermediate layer mask 303 / lower layer mask 3 04 / gate Cap layer 3 0 5 / metal gate layer 3 A portion of a planar MOS of a 06/HfSiON insulating film 307, a thin metal gate layer 306 (TiN) of 10 to 30 nm, which is etched using the present invention. After the BARC layer 302 to the gate Cap layer 305 are etched in the same chamber, a TiN (metal gate layer) 3 06 break turough (hereinafter referred to as BT) step is performed. The BT step is intended to eliminate the Ti0 surface layer that hinders TiN etching when combined with oxygen gas when the BARC layer 302 is overetched (represented by 0E). The conditions of the BT step are that ArlOO~200cc, pressure iPa, UHF applied power 500W, 400kHz, bias low frequency application power 5〇~100W. With the use of a 400 kHz low frequency bias, it is easy to obtain the necessary high energy for Ti0 removal. As shown in FIG. 4( a ), the voltage, current, and phase results detected by the plasma impedance monitoring of the 400 KHz in the plasma state detector 126 are judged by the time change of the impedance ( EPD). With the BT step of TiN, the surface TiO is removed, and Ti, N, etc. are released to the vicinity of the plasma sheath, causing changes in electrical characteristics such as electron density and sheath thickness. At the time when the impedance becomes large, it is moved to the main etch (hereinafter referred to as ME). In the ME step of ΊΊΝ306, HBr, NF3, CF4, SF6, pressure 〇 are added on the basis of Ch or HC1. 2Pa, UHF applied power -16 - 200849325 500W conditions. The bias of the ME step of TiN3 06 400kHz low frequency bias mixing 50% of the 13_56MHz high frequency bias 4 (b), the 2 frequency mixing ratio of the TiN ME is 403). The advantage of high-frequency bias can be obtained at the same time as the narrow block width suppresses the slab hard block caused by the ion, and the anisotropy caused by the high-frequency bias applied by the high-energy ion can be obtained, and the chamber wall-shaped wafer can be positively The impedance of the sheath formed above is separated, and the end point of high precision can be judged by the impedance of the crystal. The ME step is terminated by the plasma impedance detector 402 detected by the plasma state detector 1 26, switched to the OE1 step of TiN, and the output 404 of the IEDF control is set to be a high frequency bias power of 1%. This is the next step of OE1, which needs to integrate the STI step difference, the film thickness, the poor circuit shape or the difference in the material to be etched, and the in-wafer difference caused by the wafer. Therefore, the pattern of the underlying film can be maintained and the choice of the underlying layer can be maintained. At the same time, it is necessary to etch the portion that does not reach the bottom layer. At this time, by using the signal of the plasma state detector 1 26 as shown in FIG. 4(c), the end point 406 obtained by conventionally detecting the time change pattern of the reaction product or the etchant , is about 1 to 5 Get the end of ME. Thereafter, in the OE2 step of the signal stabilization of the plasma state detector 126, the plasma condition of the 〇E2 step is used, HBr/02, full flow rate of 200 to 400 cc, pressure of 3 to 10 Pa, and UHF power supply of 500 to 700 W. Use high-frequency bias power 80%, low-frequency 20% mixed bias power (in Figure 4 (c), TiN OE2 system used (the reason is to make low energy plus advantages of the sheath and circle) The point above the point is judged, the changing time mechanism 1 2 7 is because of the etching of the pattern between the high-k, compared with the illuminating peak seconds, it can be taken to the time when the Ar is diluted by 1 〇1. 2 frequency -17- 200849325 rate mixing ratio 4 0 5 ). The reason is to reduce the electron shading damage while reducing the shape of the underlying film selective step or non-opening portion. A ratio of high energy ions close to a multiple of the average energy, while reducing the shape of the hem pendulum, while maintaining selectivity, physical damage caused by sub-collision, and also reducing the source/drainage Neutralization of the charged electrons and low-energy ions in the upper part of the cover can avoid the gap caused by the loss of the electron shadow. The above, by the application of the present invention, the pole processing is as shown in Fig. 3 (c), and the underlayer can be prevented from being generated. In addition, in addition to the gate structure of the present embodiment, other insulating materials such as hi gh-k Zr02 , Υ 2 〇 3 , La 203 , La A 10 x , La SiO x , A 120 3 , and HfAlO (N) are used, and the metal gate layer is other than TiN. The same effect can be obtained by metal materials such as TaN, TaSiN, TaC, Ru, HfN, and MoN. In the above embodiment, the IEDF control unit 127 uses a mechanism that generates power supply units 118 and 120 and two matching units 11 having different frequencies. Compared with other IEDF controls such as the notch bias i bias, this method has a simple configuration of the oscillator and the matcher. When the bias applying means 1 17 uses the notch bias power supply, it is the same as the overlap of the continuous frequencies, and although it has a high price, it can be provided with a plasma state detector 1 26 having a frequency range. Further, in the present embodiment, the VPP for the voltage side amount of the sub-portion is used as an index. However, not only Vpp but also STI is reduced, and the dispersion is suppressed. In this case, the metal gate is processed. Materials such as Hf02, other cases are oscillating [9, 121 I (the advantage of Clip, the majority of continuous acceleration can be transmitted -18-200849325 output power ratio. At this time, the output power becomes the multiplication of current and voltage, Vpp will vary depending on the configuration or area of the grounding, so you need to consider this point. Alternatively, you can calculate the model as shown in Figure 2, and control the output of each power supply to make it the desired energy distribution. In this case, you can have the energy distribution. Calculate the detection mechanism of the plasma density, electron temperature, etc. to improve the control accuracy. Also, the IEDF controlled bias power supply, high frequency side system is used. 56MHz, the low-frequency side uses 400kHz, basically the difference between the two different frequencies is larger. The IEDF has a wider control range, and the separation between the wall of the chamber and the impedance above the wafer is also good. In addition, it is preferable that they do not become integer multiples of each other, and it is possible to use high harmonics of individual frequencies. At this time, in order to maintain the independence between the plasma generation and the good in-plane distribution, it is preferred that the frequency on the high frequency side is lower than the frequency of the plasma generating mechanism. For example, in the case of ECR, independent control of ion energy and plasma density at 100 MHz or more becomes difficult, so it is preferably 4 MHz or more and 100 MHz or less. Further, when the frequency on the low frequency side is lower than 100 kHz, the insulating layer on Si is liable to cause a charge up. Therefore, it is preferable that the frequency on the low frequency side is 100 kHz or more and less than 4 MHz, and the frequency on the high frequency side is 1 MHz or more and less than 100 MHz, and the frequency difference is larger as much as possible. In addition, the mixed frequency band is also affected by the plasma generating mechanism. For example, as described above, in the plasma generation mechanism in which the distributed control uses a magnetic field, the use of the cross-impedance effect of E X B is used. High frequency of 56MHz. The balance between ICP, CCP, etc. and the frequency of the plasma source can be used. 60MHz. -19- 200849325 (Second Embodiment) Yu Di 1 knows that the impedance of the wall state is constant, and the end point is obtained by the absolute resistance of the wafer processing, but in the mass production processing of the wafer There is a change in the state of the wall of the chamber caused by time. Next, an embodiment in which the state of the wall surface of the chamber is separated to obtain an end point will be described. In the first embodiment, when the end point determination is performed at a frequency mixing ratio of 50% in T i N, the detection process of the impedance corresponding to the majority of frequencies is detected as shown in the end point determination flow of FIG. 5 . After 1 'the separation process 5 0 2, that is, according to the above formula (1), formula (2), formula (3), the wall surface and the individual impedance directly above the wafer are separated, and the end point judgment project 5 0 3, The change engineering 504, that is, the output of the IEDF control unit 127, the UHF power source 101, the high-speed response pressure control unit 110, the high-speed response reactive gas introduction unit 1 1 1 , and the temperature control unit 1 1 5 are changed. The endpoint judgment mode used at this time is shown in Fig. 6. Fig. 6(a) is a diagram showing the impedance change directly above the wafer, except for the impedance change of the wall portion after the separation process, on the immitance of the majority impedance. For example, when the impedance 602 before the end point determination is moved to the impedance 6 0 1 when the etch film is completely removed, the matching path 603 is adopted by the tuning capacitor and the tuning coil depending on the matching unit, and the setting is The end point judgment is performed at the time of the majority of the impedance range of 6 0 4 and 6 0 5 . Further, Fig. 6(b) is a schematic diagram showing an impedance change detected at a single bias frequency and plotted on an immitance. When there is a change in time, as the etching processing time increases, the impedance of the wall portion of the chamber -20- 200849325 changes 606, and before the end point judgment, the set impedance range is exceeded 6〇4, and the end point becomes non-existent before the judgment. In view of this phenomenon, according to the impedance detection of most frequencies of the present invention, by performing the separation process between the wall state component and the component directly above the wafer, the end point of the wafer can be stably obtained without affecting the wall state, by The combination of IEDF control can stably obtain the uranium engraved shape of the metal gate/high-k structure without underlayer penetration and without the shape of the hem. The flow of Figure 5 is based on the control device 1 25 . The internal control program is implemented or described in the control software of the plasma state detector 126. In the present embodiment, TiN (metal gate layer) 306 having a large impact on device characteristics among the layers of the cross-sectional structure sample shown in Fig. 3(a) is described, but it can also be used for each BARC. The end points of each layer such as layer 3 02, intermediate layer mask 303, lower layer mask 3 04, and gate Cap layer 3 0 5 are judged. (Third Embodiment) The following is an embodiment in which the impedance of the separated chamber wall surface and the wafer directly above the impedance are changed. A flowchart for the ME end point judging step which is formed when the majority of the metal gate/high-k gate of Fig. 3(a) is performed is shown in Fig. 7. In the method of the second embodiment, after the detection of the impedance corresponding to the plurality of frequencies is detected, the separation process 5 0 1 is performed, that is, according to the above formulas (1), (2), and (3). The wall is separated from the individual impedance directly above the wafer, and the comparison is performed. 70 1. The impedance of the wall of the chamber to be separated and the impedance directly above the wafer, and the state of the same step in the past (the past impedance change, etc.) The data, the change model) is compared with --2149325. In Comparative Engineering 70 1, the current impedance or current, voltage, and phase are recorded in a database and compared with past data to be classified as time variation for each wafer. According to the classification, the state of the wall surface of the chamber is determined, and the adjustment process 702 is performed, that is, the wall state resetting process is performed, or the output of each device is appropriately adjusted in a positively changing manner, so that the separated impedance information can be used for time variation. Suppression. For example, in this classification, when only the voltage, current, phase, or impedance on the high frequency side changes extremely, it is determined that the uniformity of the bias in the wafer surface changes, and the surface of the bias is adjusted by the second wafer. The inner distribution makes the in-plane distribution of the bias voltage uniform, and thus, the uniformity of the control bias can be fed back, and the decrease in the yield can be suppressed. The uniformity correction method of the bias voltage can be realized by adjusting the mixing ratio of the two frequency biases. For example, when the mixing ratio of low frequency (400 kHz: VppL) to high frequency (13.56 ΜΗζ: VppH) (VppH / (VppH + VppL)) is 0%, 20%, 100%, the distribution of polycrystalline sand etch rate The changes are shown in Figures 8 (a), (b), and (c). The uranium engraving conditions are: HBr/02 for the treatment gas, 3 Pa for the pressure, and 500 W for the UHF. When the mixing ratio of the bias power is set to 1 3 · 5 6 M H z is 0%, the high distribution of the polycrystalline engraving rate is 11% (Fig. 8(a)). As a bias voltage setting 13. When the mixing ratio of 56 MHz is 20%, the medium-high distribution of the polysilicon etch rate is 〇% (Fig. 8(b)). As the bias power setting 1 3. When the mixing ratio of 5 6 Μ Η z is 1 〇 〇 %, the high distribution of the polycrystalline sand engraving rate is 12% (Fig. 8(c)). That is, with high frequency bias power -22- 200849325 13. As the mixing ratio of 56MHz increases, the distribution of the ends of the wafer rises, and the mixing ratio can be controlled by 12% of the height of the 97% control end and the distribution of the wafer with the bias mixing ratio. Fig. 8(d)) shows a mixture ratio of the high-frequency 13·56 ΜΗζ and the low-frequency 400 kHz monitor voltages VH and VL, and when the VH + VL is changed to a substantially constant manner, the polysilicon etch rate 801 on the 200 mm wafer is measured. The result of the etch rate of SiO 2 822. The SiO2 etch rate 802 is expressed in 10 times. With high frequency bias power 13. When the mixing ratio of 56 MHz is increased, the polycrystalline engraving rate is reduced by about 20%. In contrast, the SiO 2 etching rate is rapidly decreased, and the selectivity (selection ratio) 803 to the oxide film is increased. As described above, when the mixing ratio of the bias voltage is changed and the in-plane rate of the wafer is controlled, when the mixing ratio is controlled to 30% or more, the OE becomes a highly selective region with sufficient resistance (selection ratio 804 or more of 200). better. Other means of correcting the in-plane distribution of the uranium engraving rate or shape can be controlled by combining the output of the electromagnet 104 or the gas flow distribution, the in-plane distribution of the wafer platform, and the like. Moreover, as the number of wafer processing sheets increases, when only the low-frequency impedance and the high-frequency impedance of the chamber wall surface change in the same engraving step, if the state of the chamber wall surface changes, it is judged by the observer or judged by the control device 150 Automatically or manually performing clean processing of the plasma processing apparatus to promote the exchange of components, and changing the IEDF control mechanism 127, the UHF power source 101, the high-speed response pressure control mechanism 1 1 , and the high-speed response reactive gas introduction mechanism according to the modified correction model 1 1 1. The output of the temperature control mechanism 1 1 5 . The above example classifies the device state or wafer processing state by the impedance of the chamber wall or the impedance directly above the wafer, -23-200849325 or the impedance of the low frequency bias or the impedance of the high frequency bias. In addition, information on the direction of change (such as the inductive side or the conducting side) can be considered. For more detailed classification and database construction and change factors, appropriate feedback control can be implemented based on the change factor. As described above, in the engraving step (ME step or end point judging step) that affects the shape processing, the 2-frequency mixed IEDF control bias is used, the impedance is measured, recorded, and compared with the past database or variation model. In this way, the change in etching characteristics can be automatically corrected. The mechanism for realizing this method is as shown in Fig. 94. The control device 185 is connected to the database 901, and the input signal to the control device 125 is not limited to the bias applying mechanism 1 1 7 of the first embodiment or the second embodiment. The output of the plasma state detector 1 26 . That is, for the input signal of the control device 125, the output of the conventional plasma light-emitting beam splitter 123 (luminous spectrum) or the high-speed response of the additional impedance detector UHF matcher 102 (the power seen by the plasma frequency) can be used. Output of plasma impedance, voltage, current, phase) or film thickness interference monitor 124, and other mechanisms such as UHF power supply 101, electromagnet 104, heater 108, high-speed response pressure control mechanism 1 1 , high speed In response to each of the monitoring gases of the reactive gas introduction mechanism 1 1 1 , the temperature control mechanism 1 15 , and the DC power source 1 16 . The control device 185 memorizes each input database, and the control device 25 compares and refers to the past data and the newly input data recorded in the database 90 1 according to the variation model and the comparison in the reference database 901 or the control device 125. As a result, the signals of the IEDF control unit 127 or the respective control devices are output. As described above, not only the information of the lottery state detector 1 26 corresponding to the majority of the frequency of the present invention, but also the monitoring information of the control mechanisms of the -24-200849325 of the illuminating spectrum, etc., can implement more detailed device states and wafers. The classification of the treatment status can be appropriately disposed. (Fourth Embodiment) Hereinafter, with reference to Fig. 1A, a planar ArF resist 301/BARC layer 301/TEOS (hard mask) 1001/polycrystalline 002/HfSi023 07 having the shape shown in Fig. 1(a) is used. The CMOS gate electrode having a structure laminated on the Si substrate 3 09 is an embodiment in which the present invention is used for processing. Fig. 10 (a) shows the cross-sectional shape before etching having a multilayer film mask structure including a hard mask. Fig. 1 (b) is a cross-sectional shape of a conventional method of high-frequency bias after hard mask etching, and Fig. 1 (c) is a cross-sectional shape of a conventional method of high-frequency bias for gate etching Fig. 1 〇(d) is the cross-sectional shape of the present invention after the etch processing. For the Si wafer having the hard mask 1001 disposed under the BARC layer 302 and the cross-layer structure of the polysilicon gate layer 002 disposed as shown in Fig. 10(a), 26 MHz or 13. The upper frequency of 56 MHz (single bias frequency power supply) and the case of etching the hard mask 1 001 are shown in Fig. 10(b). As shown in the figure, the side wall protective film 1 004 is adhered to the portion of the hard mask 1〇〇1, and the push-out shape 1 003 is obtained. In addition, as shown in FIG. 1A(c), a gap between the lower portion of the polysilicon gate layer 1002 and the HfSiON insulating film 307 is formed on the dense pattern side, and the underlying Hf SiON insulating film is formed. One of the 3 07 insulation was destroyed. Hereinafter, an example in which a 12-25-200849325 inch wafer having the sectional structure shown in Fig. 1(a) is etched using the plasma state detector 126 of the present invention and the IEDF control bias is used. The Si wafer trimming and etching process having the cross-sectional structure shown in Fig. 10(a) is provided in the etching apparatus of Fig. 1. The action of applying a bias voltage during the trimming or etching process will be described using FIG. Fig. 11 is a timing chart accompanying the processing operation of the plasma in the embodiment of Fig. 1. The distribution map is a time chart with the axis and time as the horizontal axis. Further, in the figure, the operation of the apparatus of Fig. 1 is based on the end point determination method of the present invention using the first embodiment and the first aspect, or the output waveform from the plasma device 123, and the luminous intensity of Fig. 11 is used. The vertical axis is shown in Fig. 11(b), which shows the mixing ratio of the bias power on the ON/OFF frequency side of the 2 frequency bias. The ME step of the BARC layer 312 is based on the 02/Ar gas system (CF4, CHF3, CH2F2, CH2C12, Cl2, HBr, and the whole gas flow rate is adjusted to 100 to 400 cc, the pressure is l〇Pa, and the pressure is 5 The in-plane distribution of the 12-inch wafer generated by the UHF power output of 00W to 800W is controlled by a plurality of electromagnets 104. After the plasma is ignited, as shown in Fig. 11 (b), the 400 kHz voltage supply unit 120 applies about 30 W~ 50W, fine-tuning resist 301 etched BARC layer 3 02 (ME treatment of BARC). At this time, the high frequency is not applied, and the mixing ratio of high-frequency bias is 〇%. By using the low frequency IEDF of 400 kHz, the low ion is also acceptable. In the wall portion, the equal-side fine adjustment can be effectively performed, and the high-frequency electrode 1 2 is used to perform the frequency mixing device to perform the longitudinal plasma processing. 2 The shape-emitting luminescence is performed in a relative state and the halogen is added to the cylinder. · 8 P a ~ pulp. At the same time as the low frequency of the current 値, 13. 56MHz easy to enter side Energy ion -26- 200849325 The edge of the line can be cut off (L E R : L i n e E d g e R 〇 u g h n e s s )' can reduce the effect. The amount of fine adjustment at this time can be appropriately controlled by the mixing ratio of the gas, or the pressure, the plasma power source (the output of the UHF power source 101), the temperature of the lower electrode, and the DE time. The end point of the ME treatment of BARC 1 101 is the use of the pole of the luminescence intensity of CN3 87 nm in the plasma. After the end of the ME treatment of the BARC was detected, the uranium engraving (OE) of the BARC was performed. In OE, in order to improve selectivity, the bias power is reduced by about 10W, and in addition, it is switched to 13. The high frequency bias of 56MHz, the low frequency bias 〇, sets the mixing ratio of the high frequency bias to 100% (BAC OE processing). Then, in the ME step of the hard mask 1〇〇1, a suitable gas is mixed among the gases of SF6, CF4, CHF3, CH2F2, 02, Ar, He, and the total gas flow rate is set to 100 to 400 cc, and the pressure is 〇. 4Pa~1. 5Pa, UHF output of 500 rpm to 800 W is used to generate plasma, and the lower electrode bias power is set to 80 W to 150 W for etching (hard mask 1 〇〇 ME treatment). The frequency of the bias power is used as the output of the frequency bias power supply unit having a wide energy range of IEDF having a high energy peak, and is used at a mixing ratio of 1%. The reason for this is that the re-dissociation of the reaction product is suppressed in the vicinity of the wafer even in the case of high power required for TEOS etching. By suppressing the re-dissociation, it is possible to process vertically (Fig. 10(d)) instead of having a push-out shape 1 003 as shown in Fig. 10(b) and a hard mask having a poorly-formed shape. When the low frequency bias voltage and high power are used, an ion impact is generated on the wafer at the same ground portion, and the removed reaction product containing the ground material adheres to the quartz surface, causing temporal changes in etching characteristics or generation of foreign matter. In the case of -27-200849325, it is possible to reduce the grounding sputtering by increasing the high frequency mixing ratio without excessive dissociation and maintaining the in-plane processing size. At the end point 1102 of the hard mask etch, switching to OE, in the OE step, switching to a high frequency of 100% for the purpose of improving the selectivity between the underlying polysilicon 矽1 002, and the processing time is set to be equivalent to the STI step (hard) Mask OE processing). Thereafter, the penetration (BT) step of the polycrystalline germanium 1 002 is started, and the plasma condition of the BT treatment of the polycrystalline germanium 1 002 is set as follows: the gas system uses a gas monomer such as Cl2, HBr, 02, Ar or He or a mixed gas thereof, The total flow rate is set to 200~300cc, and the pressure is 0. 4~0. 8Pa, UHF output is 500W~ 700W. The conditions of the IEDF control mechanism 127 are set to: 400 kHz low frequency bias mixed with 100% application (polysilicon 002 BT treatment). This is because the carbonaceous material and the oxidizing substance on the bottom surface of the pattern are removed by high-energy ions, and the sidewall protective film is easily removed by low-energy ions. In the etching treatment of the polycrystalline silicon 002, the processed shape greatly affects the device characteristics. Therefore, the end point determination is performed by the method of the present invention described in the first embodiment and the second embodiment. The ME system of polycrystalline 矽1 002 is used. IEDF (polycrystalline 矽1 002 ME treatment) with a high frequency ratio of 56 MHz. The end point 11 04 of the ME of the polysilicon 矽1〇〇2 is obtained by the method of the second embodiment to obtain the signal of the plasma state detector 126. After the detection of polycrystalline 矽1 〇 〇 2, the end of the E treatment, 〇 E treatment. The 〇 Ε processing is performed in the steps of Ο Ε 1 and Ο Ε 2 . In the step ε 1 , in order to achieve both the underlying selectivity and the etched shape (vertical workability, no skirt shape), it is preferably a low pressure condition. Under low-pressure conditions, the underlying selection is required. -28 - 200849325 The property usually needs to reduce the ion energy, but at a low frequency of 400 kHz, low-energy ions are present due to the low bias voltage, the ion directivity is further reduced, and the shape of the skirt is liable to remain. Therefore, in the OE1 step, the same process gas as the ME is used, 13. The high frequency bias of 56 MHz is 100% mixed IEDF (Olympus 1002 OE1 processing). The bias power is 10W to 50W (Vpp250V or less). Thereafter, the process proceeds to the OE2 step to remove the polysilicon caused by the difference in the STI step portion or the in-plane, the p/n gate difference, and the sparse shape difference. The plasma condition of the OE2 step is that the treatment gas is diluted with HBr/02 in Ar, the total flow rate is set to 200 to 400 cc, the pressure is 3 to 10 Pa, and the UHF output is 500 W to 700 W. Use a 20% low frequency bias to mix 80% of the high frequency biased 2 frequency mixing bias (poly 矽1 002 OE2 treatment). This is a reduction in the underlying selectivity and electron shading damage. That is, by reducing the ratio of the high ion energy which is nearly double the average ion energy, the bottom layer selectivity can be maintained, and the physical damage caused by the ion impact can be suppressed. In addition, the low-energy ions neutralize the electrons charged in the upper portion of the mask to avoid gaps caused by the shadow of the electrons. The above description of the sample of the multilayer film structure, when the material to be etched is changed, the signal of the plasma state detector 126, the illuminating beam splitter 123 or the film thickness interference monitor 124 is used as the IEDF control mechanism 127 when the stimuli are moved to the next step. Setting example. The figure shows that when the step is switched, the output voltage (Vpp) of the IEDF control power supply generates an overshoot that causes excess voltage to be applied to the material to be etched, without degrading the selectivity of the control device. Fig. 15 is a schematic view showing the configuration of the control device -29-200849325 of the plasma processing apparatus of the embodiment shown in Fig. 1. According to a predetermined etching processing condition, a plasma power source (UHF power source 101), a high-speed response pressure control mechanism 110, a high-speed response reactive gas introduction mechanism 1 1 1 output 时序 or a timing control device 125. Perform output voltage and timing control of the IEDF control unit 127. The IEDF control unit 127 sets the output of the X-th local frequency bias power supply unit 1 1 8 and the output of the low-frequency bias power supply unit 120 to the next step X '1. The stable point setting 値SPHx of the SVLx, or the high frequency matcher 119 of the Xth step, and the stable point setting 値SPLx of the low frequency matcher 121 are set in advance according to the signal of the control device 152. At this time, the transition setting 値SPLy of the transition point 値SPHy of the stable point of the high frequency matching unit 1 19 from the end of the X-1 number to the X number and the stable point of the low frequency matching unit 1 2 1 is also appropriately set, In this way, when moving to the Xth step, no excessive voltage is generated and the response is smooth. At this time, feedback control can be performed according to the MPHx, MPLx (injected power or reflected power from the load, matching state, etc.) in the transition, output transition setting 値SPHy and SPLy or power output 値SVHx and SVLx. . In addition, not only the signals of the high frequency matcher 119 and the low frequency matcher 121 but also the current, voltage, phase, and impedance signals of the plasma state detector 126 can be monitored for feedback control, and reference can also be made to the database 90 1 In the past. As described above, it is a mechanism and method for smoothing and monotonously changing the time in the shortest time for the next target point, but similar to the transient phenomenon control of the IEDF control unit 127, it is also possible to control the UHF power supply 110, high-speed response pressure. Mechanism 110, high-speed response reactive gas introduction mechanism 11 1 and temperature control -30-200849325 The mechanism 1 1 5 is similarly implemented. (Fifth Embodiment) Hereinafter, an embodiment of a processing method for forming a high-step differential three-dimensional FIN-FET of Fig. 12 using the plasma processing apparatus of Fig. 1 will be described. In the figure, (a) is a perspective view of the film structure before etching, (b) is a perspective view of the film structure after etching using this embodiment, and (c) is a longitudinal section cut by the line A-A' of the figure (a). Figure. 12(a) and (c), the film structure before etching having a high-step three-dimensional structure FIN-FET is formed by a SiO 2 layer 1203, a Si layer 1202, a TiN layer 1204, and a BARC layer formed on the Si substrate 3 09. 302 and resist 30 1 constitute. A FIN portion 1201 is formed over the SiO 2 layer 1 203, and a high-k insulating film 1 205 is formed over the SiO 2 layer 1 203 and the FIN portion 1201. 1 20 8 represents a portion of the boundary between the FIN portion 1201 and the gate. In the present embodiment, the uranium engraving is started by the film structure of Fig. 12 (a) with the resist 310 as a mask, and dry etching of the film structure of Fig. 12 (b) is obtained. In Fig. 12(b), 1205 is a high-k insulating film, and 1206 is a TiN gate formed by etching the TiN layer 1204. The Si layer 1 202 has a plurality of FIN portions 1201 extending on both sides of the TiN gate 1206 and extending therethrough. Fig. 12 (c) is a longitudinal sectional view showing the film structure of Fig. 12 (a) having the F IN portion 1 2 0 1 . In the wafer having the film structure of Fig. 12(a), the difference between the upper portion and the lower portion of the FIN portion 1201 is large, and the gate length portion having a large influence on the device characteristics is perpendicular, so that The IEDF control according to the first embodiment and the fourth embodiment, and the control of the transition phenomenon between the steps of the gas, the pressure, and the matching device - 31 - 200849325 and the electrode temperature and the magnetic flux of the reaction product or the etchant or the ion are also controlled. important. And the mechanism of the transition phenomenon of the etching process is preferably used in the UUP matcher 102, and the plasma injection power has no overshoot or undershoot, and has a monotonous increase of about 1 second or Monotonically decreasing performance stability. To achieve this performance, for example, by setting the most matching parameters of the ignition and stabilization points and the best function of the matching path. The time differential of the monotonically increasing response curve is always positive, and the time differential of the monotonically decreasing response curve is always negative. Further, when the transition phenomenon of the reaction product or the etchant is to be driven, it is preferably used in the high-speed response reactive gas introduction mechanism 1 1 1 , and when the new gas is added/reduced in the plasma discharge, the flow rate does not overflow. A mechanism for overshooting or undershooting, for example, a mechanism in which gas does not remain and continues to flow into the gas during waiting time. Further, in order to follow the fluctuation of the plasma pressure accompanying the change in the gas flow rate, it is preferable to have a high-speed response pressure control mechanism 1 1 可 which can stabilize the pressure for about 2 seconds. This function can be realized by setting the gas piping structure that does not cause a pressure difference as much as possible, or by optimizing the pressure control algorithm. Further, when the change in the in-plane distribution of the Si wafer to be reacted with the reaction product is driven, the lower electrode 1 1 3 preferably has a high-speed (1 ° C /sec or more) step-and-step temperature rise and fall. The inner, outer, or more portions have independently controllable functions. This function can be realized, for example, by providing a temperature control mechanism 115 for providing a heater, a temperature sensor, a He gas pressure control, and the like inside the lower electrode 1 1 3 . -32- 200849325

使用具備該功能機構的圖1之電漿處理裝置,進 SOI ( Silicon on Insulator )基板上作成之 FIN-FET 時之蝕刻處理之時序圖如圖1 3所示,圖1 3 ( a )爲例 用發光強度的EPD用信號Η時間變化,圖13 ( b )爲 中之下部電極之溫度,圖1 3 ( c )爲供給至蝕刻腔室 之添加氣體之流量,圖1 3 ( d )爲蝕刻腔室1 0 7內之 壓力,圖13 ( e )爲由天線103導入鈾刻腔室107 UHF電力,圖13 ( f)爲IEDF控制機構127進行之 頻率之偏壓分配比。 控制下部電極1 1 3之電極溫度的溫度控制機構1 初期設定値,係如圖1 3 ( b )所示,使內周部1 3 1 1及 部1 3 1 2分別由40 °C開始。依據第4實施形態之製程 ,使用〇2/Cl2/Ar/CF4進行BARC層302之ME步驟, 漿發光強度開始減少之時刻1 3 02移至BARC層3 02 ;; 步驟。於OE中,需要提升和TiN層1 204間之選擇 進行斜波(ramp)控制而使成爲TiN之蝕刻劑之C1; 發光強度減弱之時間1 3 0 1以下逐次呈現單調遞減( )。於習知氣體流量控制波形1 304,僅進行Cl2氣體 氣操作閥之開關控制之故,成爲瞬間減少。另外,於 電漿中之壓力變動1307,瞬間關閉Cl2氣體之制動閥 ,急速減少之後,需要花費約5秒而回復設定壓力。 相對於此,本實施形態中,藉由使用高速響應反 氣體導入機構1 1 1及高速響應壓力控制機構1 1 〇,配 應生成物之減少漸漸減少蝕刻量,依此而補正壓力變 行於 飩刻 如使 處理 107 氣體 內之 多數 15之 外周 順序 於電 匕OE 性, 丨,於 1303 之空 習知 之故 應性 合反 動使 -33- 200849325 成爲一定,又,壓力成爲一定之故,可以抑制習知於UHF 匹配器產生之電漿射入電力之變動1 3 09或Vpp之變動’ 可以抑制形狀異常。 又,BARC層302之OE中之IEDF控制機構127之頻 率,和第1實施形態同樣,由400kHz之低頻切換爲 13.56MHz之高頻,偏壓輸出則使用30W〜50W之範圍。 僅於蝕刻FIN段差1 207之時間進行OE處理後,移至TiN 層1 204之BT步驟。此時,於切換氣體時將UHF波輸出 及偏壓電力設定開關1 3 0成爲OFF狀態。於中斷放電進行 切換氣體之1 〇數秒間,使TiN蝕刻中之內周部分之電極 溫度131 1上升20 °C。此爲抑制TiN蝕刻中之再度射入之 反應生成物之再度附著。又,下部電極1 1 3之外周部分之 溫度1312,可考慮排氣效率引起之反應生成物之分布差異 而調節爲低10〜20°C。The timing chart of the etching process when the FIN-FET formed on the SOI (Silicon on Insulator) substrate is used in the plasma processing apparatus of Fig. 1 having the functional mechanism is shown in Fig. 13 and Fig. 13 (a) is taken as an example. The EPD of the luminous intensity is time-varying with a signal, and Fig. 13 (b) shows the temperature of the middle and lower electrodes, and Fig. 13 (c) shows the flow rate of the additive gas supplied to the etching chamber, and Fig. 13 (d) is etching. The pressure in the chamber 1 0 7 , Fig. 13 ( e ) is the UHF power introduced into the uranium engraving chamber 107 by the antenna 103, and Fig. 13 (f) is the bias distribution ratio of the frequency performed by the IEDF control unit 127. The temperature control means 1 for controlling the electrode temperature of the lower electrode 1 1 3 is initially set to start at 40 ° C as shown in Fig. 13 (b), and the inner peripheral portion 1 3 1 1 and the portion 1 3 1 2 are respectively started. According to the process of the fourth embodiment, the ME step of the BARC layer 302 is performed using 〇2/Cl2/Ar/CF4, and the time at which the pulverization intensity starts to decrease is moved to the BARC layer 301; In OE, it is necessary to upgrade and select between TiN layers 1 204 to perform ramp control to make C1 which is an etchant for TiN; and the time when the luminescence intensity is weakened is 1⁄0 1 or less, and monotonically decreasing ( ) is successively presented. In the conventional gas flow control waveform 1 304, only the switching control of the Cl2 gas-operated valve is performed, which is instantaneously reduced. Further, when the pressure in the plasma fluctuates by 1307, the brake valve of the Cl2 gas is instantaneously turned off, and after the rapid decrease, it takes about 5 seconds to return to the set pressure. On the other hand, in the present embodiment, by using the high-speed response back gas introduction mechanism 1 1 1 and the high-speed response pressure control mechanism 1 1 〇, the reduction of the matching product gradually reduces the amount of etching, and accordingly, the pressure is corrected. If the engraving of the majority of the gas in the processing of the gas is in the order of the electricity 匕 OE, 丨, in the 1303, the customary reaction and the reaction are made -33- 200849325 becomes a certain, and the pressure becomes a certain reason, It is known that the variation of the plasma injection power generated by the UHF matcher 1 3 09 or the variation of Vpp can suppress the shape abnormality. Further, the frequency of the IEDF control unit 127 in the OE of the BARC layer 302 is switched from a low frequency of 400 kHz to a high frequency of 13.56 MHz as in the first embodiment, and a bias output of 30 W to 50 W is used. After the OE process is performed only at the time when the FIN step difference of 1 207 is etched, the BT step of the TiN layer 1 204 is moved. At this time, the UHF wave output and the bias power setting switch 130 are turned OFF when the gas is switched. The electrode temperature 131 1 of the inner peripheral portion of the TiN etching was raised by 20 ° C for 1 sec. of the switching of the gas during the interruption of the discharge. This is a re-attachment of the reaction product which suppresses re-injection in the TiN etching. Further, the temperature 1312 of the outer peripheral portion of the lower electrode 1 1 3 can be adjusted to be 10 to 20 ° C lower in consideration of the difference in distribution of the reaction product due to the exhaust efficiency.

TiN 層係由 BT、ME、0E1、0E2 構成,BT、ME 係和 第1實施形態同樣適用本發明之終點判斷方法及蝕刻條件 〇 習知UHF匹配器之情況下,移至BT步驟S時,點火 時電漿射入電力之變動1 3 08會出現,但藉由UHF匹配器 102之使用可以圓滑移行而不會產生上沖或下沖。此乃因 爲,UHF匹配器102在點火時及穩定時移行至不同之匹配 參數時,可以適當選擇其匹配路徑。 BT步驟’係導入BT處理氣體之同時,施加UHF電 力設爲低頻側偏壓而進行,檢測出B T步驟之終點丨3 〇 i時 -34- 200849325 ’停止B T處理氣體之供給,停止UHF電力及低頻側電力 之供給。 於ME步驟,係供給ME處理氣體之同時,開始對天 線進行UHF電力供給,偏壓電力設爲低頻側與高頻側爲i ;1。於ME步驟,於電漿狀態檢測器1 26之阻抗開始減 少之時刻13 14,切換爲TiN OE1步驟。 OE1步驟之目的爲統合段差或疏密、面內差引起之蝕 刻量、亦即快速到達底層膜之圖案與乃未到達底層膜之圖 案混合。因此,於到達底層膜之圖案維持與hi gh-k間之選 擇性,而且於未到達底層膜之圖案部分,特別是垂直方向 之閘極長度部分,推拔形狀與裙襬形狀之減少成爲必要。 欲提升選擇性時,如圖1 3 ( c )所示,配合以單調遞增而 呈逐次減少的反應生成物量加以控制而添加F系氣體。添 加氣體,可考慮底層材料與閘極材料而使用NF3、SF6、 CF4、02、N2、CH2C12氣體等。閘極/閘極絕緣膜構造圍多 晶矽/Si02時,氧或氮具有同樣功能。此時,藉由高速響 應反應性氣體導入機構1 1 1、高速響應壓力控制機構1 1 0 、UHF匹配器102,以不會產生氣體流量溢流1 3 05、壓力 變動1 3 06、射入UHF波之變動1 3 08或Vpp之變動而添加 彼等氣體。OE1中,IEDF控制機構127,爲提升選擇性而 使具有窄能量分布而以高頻100%處理,如此則,不存在 開口部形狀異常(側壁蝕刻、底層貫穿)。可以減少疏密 形狀差或P/N閘極部之差。 又,本實施形態中,藉由和TiN之ME之終點13 14 -35 - 200849325 同時下降電極溫度20 °C,使變少之反應生成物之吸附機率 增加,而增加出現選擇性之沈積物之附著量。此乃藉由反 應生成物之變少,蝕刻劑比率上升,可抑制進入側壁鈾刻 〇 另外,於OE2,需要除去FIN部分與閘極之境界部分 1208殘留之TiN,因爲成爲閘極長度之故需要高精確度控 制。因此,使用2頻率偏壓1313之高頻13.56MHz設爲 8 0%之Vpp,使全電源輸出之IEDF平均成爲50V以下之 値(lOOVpp以下)。其理由爲,對於FIN上部之high-k 材料絕緣膜需要高選擇性。鈾刻閘極1 206與FIN部分 1201之境界部1 208時,以低能量離子由境界之上部加以 削落。偏壓之全輸出抑制於lOOVpp以下,可兼顧選擇性 與鈾刻。此時之偏壓電力大約1 W/1 2英吋。 另外,爲確保OE時間而使用,和OE2中途逐次變低 之境界部之高度連動而使混合比朝高選擇側增加的鋸齒狀 控制波形13 16爲有效。另外,上述BARC蝕刻、TiN鈾 刻時之2頻率混合比,需要依據圖案密度或FIN高度(段 差高度)適當調整。 另外,IEDF控制機構127內之高頻匹配器1 19、低頻 匹配器121,亦和UHF匹配器102同樣,較好是具有適當 設定多數匹配點與其之匹配路徑之工程者,如此則,可防 止離子能量、或Vpp、或輸出電力之震動、或上沖、下沖 。另外,以可以圓滑移至蝕刻之離子促進反應,使對飩刻 有影響的電漿內部參數(自由基種、密度、離子密度、射 -36- 200849325 入之離子能量)可以圓 '滑(單調遞減或單調遞增)轉移的 方式,藉由控制各機構之運算法則來實施。電漿內部參數 係指表示電漿特徵之量’彼等之量之中,自由基種、密度 、離子密度係使用發光分光器1 23或新的密度檢測探針, 關於射入之離子能量可藉由電漿狀態檢測器1 26之信號檢 測進行回授控制,或事前與資料庫準備多數個轉移點(移 行點)加以控制。此時’步驟轉移間之各個竟之設定値需 要同時變化,因此控制需以收斂方式進行。 (第6實施形態) 以下依據圖14說明於使用μ波-ECR電漿處理裝置中 ,處理閘極以外之膜構造而形成之例。此例中說明於圖1 4 (a )之矽基板3 09形成深孔之情況。此情況下之飩刻步 驟,係由BARC 3 02、硬質遮罩1001、與矽309之蝕刻步 驟構成,BARC層3 02與硬質遮罩1001之鈾刻步驟,係依 據第4實施形態之要領進行ME、OE製程。之後,混合 SF6、CF4、CHF3、CH2F2、SiCl4、SiF4 等含氟氣體或氧, 於100〜3 00cc氣體流量、〇.4Pa-1.5Pa壓力,電漿產生機 構之μ波輸出500W〜800W之狀態下產生電漿,電極溫度 內/外差約爲5°C〜2(TC,內側設爲較高而進行Si層之蝕 刻。圖1 4 ( d )所示1 40 1爲硬質遮罩鈾刻後之溝槽。 本實施形態中,IEDF控制機構127係使用離子能量 分布變窄之高頻13·56ΜΗζ之100%成份。此乃爲抑制,低 頻IEDF存在之低能量之離子能量之助長圖14 ( 〇 )所示 -37- 200849325The TiN layer is composed of BT, ME, 0E1, and 0E2, and the BT and ME systems and the first embodiment are similarly applied to the end point determination method and the etching condition of the present invention. When the UHF matching device is used, when moving to the BT step S, Variations in plasma injection power during ignition will occur 1 3 08, but can be smoothly moved by UHF matcher 102 without overshoot or undershoot. This is because the UHF matcher 102 can appropriately select its matching path when it is ignited and stabilized to different matching parameters. In the BT step, when the BT process gas is introduced, the UHF power is applied as the low-frequency side bias, and when the end point of the BT step is detected, 343 〇i -34-200849325 'stops the supply of the BT process gas, and stops the UHF power and The supply of low frequency side power. In the ME step, the UHF power supply to the antenna is started while the ME process gas is supplied, and the bias power is set to the low frequency side and the high frequency side as i; In the ME step, the step 1414 is switched to the TiN OE1 step at the time when the impedance of the plasma state detector 126 begins to decrease. The purpose of the OE1 step is to integrate the etched amount caused by the step or the density, the in-plane difference, that is, the pattern that quickly reaches the underlying film and the pattern that does not reach the underlying film. Therefore, the pattern reaching the underlying film maintains selectivity with hi gh-k, and it is necessary to reduce the shape of the push-pull and the shape of the skirt without reaching the pattern portion of the underlying film, particularly the gate length portion in the vertical direction. . When the selectivity is to be increased, as shown in Fig. 13 (c), the F-type gas is added by controlling the amount of the reaction product which is successively decreased in a monotonous manner. When adding a gas, NF3, SF6, CF4, 02, N2, CH2C12 gas, etc. may be used in consideration of the underlying material and the gate material. When the gate/gate insulating film structure is surrounded by polysilicon/SiO 2 , oxygen or nitrogen has the same function. At this time, by the high-speed response reactive gas introduction mechanism 1 1 1 , the high-speed response pressure control mechanism 1 10 , and the UHF matcher 102, gas flow overflow 1 3 05, pressure fluctuation 1 3 06, and injection are not generated. The UHF wave changes by adding a gas to the change of 1 3 08 or Vpp. In OE1, the IEDF control unit 127 has a narrow energy distribution and a high frequency of 100% in order to improve the selectivity. Thus, there is no abnormality in the shape of the opening (sidewall etching, underlayer penetration). It is possible to reduce the difference in the shape of the denseness or the difference in the P/N gate. Further, in the present embodiment, by lowering the electrode temperature by 20 ° C at the same time as the end point of the ME of TiN 13 14 -35 - 200849325, the adsorption probability of the reaction product which is reduced is increased, and the deposit having selectivity is increased. The amount of adhesion. This is because the reaction product is less, the etchant ratio is increased, and the entry into the sidewall uranium can be suppressed. In addition, in the OE2, it is necessary to remove the TiN remaining in the FIN portion and the gate boundary portion 1208 because it is the gate length. High precision control is required. Therefore, the high frequency 13.56 MHz of the two-frequency bias 1313 is set to 80% Vpp, so that the IEDF of the full power supply output becomes an average of 50 V or less (100 Vpp or less). The reason is that high selectivity of the high-k material insulating film on the upper portion of the FIN is required. When the uranium gate 1 206 and the FIN portion 1201 are at the boundary 1 208, the low energy ions are cut off from the upper part of the boundary. The full output of the bias voltage is suppressed below 100Vpp, which can balance selectivity with uranium engraving. The bias current at this time is approximately 1 W/1 2 inches. Further, in order to secure the OE time, it is effective to increase the zigzag control waveform 13 16 in which the mixing ratio is increased toward the high selection side in conjunction with the height of the boundary portion where the OE 2 is gradually lowered. Further, the frequency mixing ratio of the above BARC etching and TiN uranium engraving needs to be appropriately adjusted depending on the pattern density or the FIN height (step height). In addition, the high frequency matching unit 19 and the low frequency matching unit 121 in the IEDF control unit 127 are also similar to the UHF matching unit 102, and it is preferable to have an engineer who appropriately sets a matching point with a plurality of matching points, thereby preventing Ion energy, or Vpp, or vibration of output power, or overshoot, undershoot. In addition, the ions can be smoothly moved to the etched ions to promote the reaction, so that the internal parameters of the plasma (radical species, density, ion density, and ion energy of the injection -36-200849325) can be rounded and slipped. The mode of monotonically decreasing or monotonically increasing the transfer is implemented by controlling the algorithms of each institution. The internal parameters of the plasma refer to the amount of the characteristics of the plasma. Among the quantities, the radical species, density, and ion density are determined by using the illuminating beam splitter 1 23 or a new density detecting probe. The feedback control is performed by the signal detection of the plasma state detector 126, or it is controlled in advance with the database to prepare a plurality of transition points (transition points). At this time, the setting of each step in the step transfer needs to be changed at the same time, so the control needs to be performed in a convergent manner. (Sixth Embodiment) An example in which a membrane structure other than a gate is formed in a μ wave-ECR plasma processing apparatus will be described below with reference to Fig. 14 . In this example, the case where the substrate 309 of Fig. 14 (a) forms a deep hole is described. In this case, the engraving step is performed by the etching steps of BARC 301, hard mask 1001, and 矽309, and the uranium engraving step of BARC layer 312 and hard mask 1001 is performed according to the method of the fourth embodiment. ME, OE process. Thereafter, a fluorine-containing gas or oxygen such as SF6, CF4, CHF3, CH2F2, SiCl4, or SiF4 is mixed, and a gas flow rate of 100 to 300 cc, a pressure of 44 Pa-1.5 Pa, and a state of a pulse output of 500 W to 800 W of the plasma generating mechanism. The plasma is generated, and the inner/outer difference of the electrode temperature is about 5 ° C to 2 (TC, the inner side is set to be higher and the Si layer is etched. Figure 1 4 (d) shows that 1 40 1 is a hard mask uranium engraving. In the present embodiment, the IEDF control unit 127 uses a 100% component of a high frequency 13·56 变 whose ion energy distribution is narrowed. This is a suppression of low-energy ion energy in the low-frequency IEDF. (〇) as shown -37- 200849325

Si層之空洞1402。另外,藉由能量分布之整合,亦可抑 制阻劑之晶面(facet )角之變大,深孔之尺寸不會擴大, 可實現不會產生如圖14(d)所示空洞之孔(hole)之高 精確度加工。如此則,進行和氧、SiCl4、SiF4等沈積氣體 之時序控制,可實施更微細、更高深寬比之加工。 藉由上述方法、機構,在具備包含段差、金屬材料、 high-k材料的多層構造之平面型CMOSFET、立體構造( FIN-FET)等之閘極触刻中,可實現穩定之飩刻加工,不 會有形狀異常(疏密形狀差、缺口、側壁蝕刻、底層膜損 傷、裙襬形狀、底層貫穿、推拔形狀等)之產生。 彼等機構、實施形態隨於S i晶圓之半導體加工被實 施’但是亦可對應於下部電極1 1 3之形狀,而適用於電漿 顯示器、液晶、MEMS製造等之電漿蝕刻全體。 【圖式簡單說明】 圖1爲本發明實施形態之電漿處理裝置之斷面圖。A cavity 1402 of the Si layer. In addition, by integrating the energy distribution, it is also possible to suppress the increase of the facet angle of the resist, and the size of the deep hole is not enlarged, so that a hole which does not have a void as shown in Fig. 14(d) can be realized ( High precision machining of holes). In this way, timing control with deposition gases such as oxygen, SiCl4, and SiF4 can be performed to perform finer and higher aspect ratio processing. According to the above-described method and mechanism, stable etching can be realized in a gate-type CMOSFET having a multilayer structure including a step, a metal material, and a high-k material, and a gate structure such as a three-dimensional structure (FIN-FET). There are no abnormalities in shape (poor shape, notch, sidewall etching, underlying film damage, skirt shape, underlayer penetration, push-out shape, etc.). These mechanisms and embodiments are implemented in accordance with the semiconductor processing of the S i wafer, but may be applied to the plasma etching of plasma display, liquid crystal, MEMS manufacturing, etc., depending on the shape of the lower electrode 1 1 3 . BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a cross-sectional view showing a plasma processing apparatus according to an embodiment of the present invention.

圖2爲400kHz、13·56ΜΗζ及2頻率混合偏壓之IEDF 〇Figure 2 shows the IEDF 400 of 400kHz, 13.56ΜΗζ and 2 frequency mixing bias.

圖3爲具備多層構造之平面(planar)型CM0SFET $斷面圖’(a )爲蝕刻處理前,(b )爲習知方法之鈾刻 處理’ (c )爲使用本發明之蝕刻處理,該多層構造包含 &差、金屬材料及high-k材料。 圖4爲處理圖3之斷面構造之晶圓時使用之電氣 t點判斷及2頻率偏壓之時序圖及習知終點判斷方法之時 -38- 200849325 序比較。 61 5爲本發明之終點判斷流程圖。 圖6爲終點判斷用之晶圓正上方之阻抗與導納( immitance)之圖,(a)爲本發明,(b)爲1頻率。 圖7爲本發明之穩定化、變動補正之流程。 圖8爲多晶矽速率之分布及多晶矽速率與Si02速率 、選擇性之混合比依存性之圖,(a )爲低頻與高頻之混 合比0%時,(b )爲低頻與高頻之混合比20%時,(c ) 爲低頻與高頻之混合比1 00%時,(d )爲多晶矽速率與 Si〇2速率、選擇性之混合比依存性。 圖9爲實現圖7之流程的機構圖‘。 圖10爲硬質遮罩樣本之斷面形狀模式圖,(a)爲蝕 刻處理前,(b )爲習知方法之高頻偏壓之硬質遮罩蝕刻 後’ (c )爲習知方法之高頻偏壓之閘極鈾刻後,(d )爲 本發明之多數高頻偏壓之閘極蝕刻後。 圖11爲圖10(a)之斷面構造之Si晶圓處理時使用 之EPD與2頻率偏壓之時序圖。 圖12爲FIN-FET構造之樣本之鳥瞰圖及斷面圖,(a )爲閘極蝕刻前,(b )爲使用本發明之閘極鈾刻後,(c )爲使用本發明處理後之FIN-FET構造之樣本之沿圖12 (a )之A-A’線的斷面圖。3 is a plan view type CMOS transistor having a multilayer structure, a cross-sectional view '(a) is before etching, and (b) is a conventional method of uranium engraving treatment' (c) for etching treatment using the present invention, The multilayer construction contains & poor, metallic materials and high-k materials. Fig. 4 is a timing chart of the electrical t-point judgment and the 2-frequency bias used in the processing of the wafer of the cross-sectional structure of Fig. 3 and the conventional end point judging method -38-200849325. 61 5 is a flowchart for determining the end point of the present invention. Fig. 6 is a diagram showing impedance and immitance directly above the wafer for endpoint determination, (a) is the present invention, and (b) is 1 frequency. Fig. 7 is a flow chart showing the stabilization and variation correction of the present invention. Fig. 8 is a graph showing the distribution of polysilicon rate and the ratio of polysilicon rate to the ratio of SiO2 rate and selectivity. (a) when the mixing ratio of low frequency to high frequency is 0%, and (b) is the mixture ratio of low frequency and high frequency. At 20%, (c) is a mixing ratio of low frequency to high frequency of 100%, and (d) is a mixture ratio of polysilicon rate to Si〇2 rate and selectivity. FIG. 9 is a mechanism diagram ‘implementing the flow of FIG. 7. Figure 10 is a schematic cross-sectional view of a hard mask sample, (a) before the etching process, (b) after the hard mask of the high frequency bias of the conventional method is etched, '(c) is the height of the conventional method After the frequency-biased gate uranium is engraved, (d) is the gate of most of the high-frequency bias of the present invention after etching. Fig. 11 is a timing chart showing the EPD and the 2 frequency bias used in the Si wafer processing of the cross-sectional structure of Fig. 10(a). Figure 12 is a bird's eye view and a cross-sectional view of a sample of a FIN-FET structure, (a) before gate etching, (b) after etching using the gate uranium of the present invention, and (c) after treatment using the present invention. A cross-sectional view of the sample of the FIN-FET structure taken along line A-A' of Fig. 12(a).

圖13爲處理FIN-FET時使用之時序圖,(a)爲EPD 波形,(b )爲下部電極之溫度,(c )爲添加氣體之流量 ’ (d)爲氣體壓力,(e)爲射入UHF電力,(f )爲 -39- 200849325 IEDF控制機構之分配比。 圖1 4爲樣本之斷面形狀模式圖,(a )爲鈾刻處理前 ,(b )爲使用本發明以硬質遮罩触刻樣本之後’ (c )爲 使用習知技術鈾刻處理樣本之後之深孔之斷面形狀’ (d )爲使用本發明蝕刻處理樣本之後之深孔之斷面形狀。 圖1 5爲圖1所示實施形態之電漿處理裝置之控制裝 置構成之槪略模式圖。 【主要元件符號說明】 101 : UHF 電源 102 :附加阻抗檢測器的高速響應UHF匹配器 1 0 3 :天線 1 0 4 :電磁鐵 1 0 5 :石英板 1 0 6 :噴氣板 1 0 7 :飩刻腔室 1 0 8 :加熱器 1 0 9 :高真空幫補 1 1 0 :高速響應壓力控制機構 1 1 1 :高速響應反應性氣體導入機構 1 1 2 :矽晶圓 1 1 3 :下部電極 1 1 4 :承受器 1 1 5 :溫度控制機構 -40- 200849325 1 1 6 :直流電源 1 1 7 :偏壓施加機構 1 1 8 :高頻偏壓電源部 1 1 9 :高頻匹配器 120 :低頻偏壓電源部 1 2 1 :低頻匹配器 122 :發光受光部 123 :發光分光器 124 :膜厚干涉監控器 1 2 5 :控制裝置 126 :電漿狀態檢測器 127 : IEDF控制機構Figure 13 is a timing diagram used when processing a FIN-FET, (a) is the EPD waveform, (b) is the temperature of the lower electrode, (c) is the flow rate of the added gas ' (d) is the gas pressure, and (e) is the shot Into UHF power, (f) is the allocation ratio of -39- 200849325 IEDF control agency. Figure 14 is a cross-sectional shape pattern of the sample, (a) before the uranium engraving treatment, and (b) after the hard mask is used to inspect the sample using the present invention' (c) after processing the sample using the conventional technique uranium engraving The cross-sectional shape of the deep hole '(d) is the cross-sectional shape of the deep hole after the sample is etched using the present invention. Fig. 15 is a schematic view showing the configuration of a control device of the plasma processing apparatus of the embodiment shown in Fig. 1. [Main component symbol description] 101 : UHF power supply 102 : High-speed response UHF matcher with additional impedance detector 1 0 3 : Antenna 1 0 4 : Electromagnet 1 0 5 : Quartz plate 1 0 6 : Jet plate 1 0 7 :饨Engraving chamber 1 0 8 : Heater 1 0 9 : High vacuum assist 1 1 0 : High-speed response pressure control mechanism 1 1 1 : High-speed response reactive gas introduction mechanism 1 1 2 : Silicon wafer 1 1 3 : Lower electrode 1 1 4 : susceptor 1 1 5 : temperature control mechanism -40 - 200849325 1 1 6 : DC power supply 1 1 7 : bias application mechanism 1 1 8 : high-frequency bias power supply unit 1 1 9 : high-frequency matcher 120: Low-frequency bias power supply unit 1 2 1 : Low-frequency matcher 122 : Light-emitting light-receiving unit 123 : Light-emitting beam splitter 124 : Film thickness interference monitor 1 2 5 : Control device 126 : Plasma state detector 127 : IEDF control mechanism

201 :高頻(13·56ΜΗζ)之 IEDF 202 :高頻時之分布寬度 203 :低頻(400kHz )之 V p p 2 0 0 V 之 IE D F 204:低頻時之分布寬度201 : High frequency (13·56 ΜΗζ) IEDF 202 : Distribution width at high frequency 203 : Low frequency (400 kHz) V p p 2 0 0 V IE D F 204: Distribution width at low frequencies

205 :低頻(400kHz )之 VpplOOV 與高頻(13·56ΜΗζ )之VpplOOV混合時之IEDF 3 0 1 :阻劑 3 02 : B ARC 3 0 3 :中間層遮罩 3 04 :下層遮罩 3 0 5 :閘極Cap層 3 0 6 :金屬閘極層 -41 - 200849325 3 07 : HfSiON 絕緣膜 308 : STI 3 09 :矽基板 310 : STI 段差 3 1 1 :主動部分之閘極下部 312 : HfSiON膜之底層貫穿 3 1 3 : S TI段差部之閘極材料殘留 3 1 4 :下擺下拉形狀 3 15 : STI上之閘極材料殘留 3 1 6 :本發明處理之閘極 40 1 : BT終點(阻抗變大之時刻) 402 : TiN之ME終點(阻抗開始變化之時刻 40 3 : TiN之ME時之2頻率混合比 4 04 : TiN之OE1時之2頻率混合比 40 5 : TiN之OE2時之2頻率混合比 406 :習知基於發光峰値之TiN之ME終點 ME時之 終點 407 :膜厚干涉監控器檢測出之TiN之BT : 輸出波形 408 :膜厚干涉監控器檢測出之TiN之ME $ 60 1 :蝕刻膜被完全除去時之阻抗 6 0 3 :匹配路徑 604 :設定之多數阻抗範圍 60 5 :匹配路徑通過604之點 606 :經時變化存在時之終點判斷前之阻抗 -42- 200849325 607 :經時變化存在時之終點判斷後之阻抗 8 0 1 :多晶矽蝕刻速率 802 : Si02蝕刻速率 8 03 :對氧化膜之選擇性(選擇比) 804 :選擇比200之線 901 :資料庫 1001 :硬質遮罩 1 0 0 2 :多晶矽閘極層 1 0 03 :存在推拔形狀:疏密形狀差之硬質遮罩 1 004:附著於側壁之反應生成物 1 0 5 ··缺口 110 1: B ARC之終點 1 102 :硬質遮罩鈾刻之終點 1 103 :多晶矽之BT步驟之終點 1 1 04 :多晶矽之ME步驟之終點 1 1 05 : B ARC蝕刻時之2頻率混合比 1 1 06 :硬質遮罩蝕刻時之2頻率混合比 1 1 〇7 :多晶矽鈾刻時之2頻率混合比 1 1 0 8 :多晶矽之〇 E蝕刻時之2頻率混合比 1201 : FIN 部分 1202 :矽層 1 203 : Si02 層 1204 : TiN 層 1 205 : High-k 絕緣膜 -43- 200849325 1 206 :被鈾刻之TiN閘 1 207 : FIN 段差 1 208 : FIN部分與閘極 1 3 0 1 :發光強度減少之 1 3 02 : CN發光強度之P 1 3 0 3 :斜波控制 1 3 04 :習知技術之氣體 1 3 0 5 :氣體流量之溢流 1 3 0 6 :壓力變動 1 3 07 :習知技術之電漿 1 3 0 8 :點火時電漿射入 1309:電漿射入電力之 1310:發光強度充分上 13 11: TiN蝕刻中之內 13 12 : TiN蝕刻中之外 13 13: TiN 之 ME 步驟: 13 14 : Ti發光強度開始 1 3 1 5 :於高速響應反應 調遞增之氟系添加氣體之流 1 3 1 6 :斜波控制波形 1401 :硬質遮罩鈾刻後 1402 :空洞 極 之境界部分 時間 S始減少時刻 流量控制波形 (overshoot) 中之壓力變動 電力之變動 變動 升之時刻 周部分之電極溫度 周部分之溫度 匕2頻率偏壓 減少之時刻 性氣體導入機構被控制成爲單 量 之溝槽 -44-205: Low frequency (400 kHz) VpplOOV mixed with high frequency (13·56 ΜΗζ) VpplOOV IEDF 3 0 1 : Resistor 3 02 : B ARC 3 0 3 : Intermediate layer mask 3 04 : Lower layer mask 3 0 5 : Gate Cap Layer 3 0 6 : Metal Gate Layer -41 - 200849325 3 07 : HfSiON Insulation Film 308 : STI 3 09 : 矽 Substrate 310 : STI Step Difference 3 1 1 : Active Part Gate Lower Part 312 : HfSiON Film The bottom layer runs through 3 1 3 : S TI step difference gate material residue 3 1 4 : hem pull down shape 3 15 : STI gate material residue 3 1 6 : The gate of the invention treatment 40 1 : BT end point (impedance change 402 moment: ME end point of TiN (time when impedance starts to change 40 3 : frequency of 2 times of TiN ME 4 4: Mix ratio of OE1 of TiN 2 frequency mixing ratio 40 5 : 2 frequency of OE2 of TiN Mixing ratio 406: conventionally based on the end point of the ME end point ME of the illuminating peak Ti TiN: BT of the TiN detected by the film thickness interference monitor: Output waveform 408: MEN 60 of the TiN detected by the film thickness interference monitor 1 : Impedance when the etched film is completely removed 6 0 3 : Matching path 604: Setting a majority of the impedance range 60 5 : Matching path through 604 606: Impedance before the end point of the change over time -42- 200849325 607: The impedance after the end of the time change is judged 8 0 1 : Polysilicon etch rate 802 : Si02 etching rate 8 03 : Selection of oxide film Sex (selection ratio) 804: Select line than 200 901: Database 1001: Hard mask 1 0 0 2: Polycrystalline 矽 gate layer 1 0 03 : There is a push-up shape: a hard mask with a poorly dense shape 1 004: Reaction product attached to the sidewall 1 0 5 ·· Notch 110 1: End point of B ARC 1 102 : End point of hard mask uranium engraving 1 103 : End point of BT step of polycrystalline germanium 1 1 04 : End point of ME step of polycrystalline germanium 1 1 05 : B ARC etching frequency mixing ratio 1 1 06 : 2 times mixing ratio of hard mask etching 1 1 〇7 : 2 frequency mixing ratio of polycrystalline uranium engraving 1 1 0 8 : polycrystalline germanium E etching Time 2 frequency mixing ratio 1201 : FIN part 1202 : 矽 layer 1 203 : SiO 2 layer 1204 : TiN layer 1 205 : High-k insulating film -43- 200849325 1 206 : TiN gate uranium engraved 1 207 : FIN step difference 1 208 : FIN part and gate 1 3 0 1 : luminous intensity reduction 1 3 02 : CN luminous intensity P 1 3 0 3: ramp control 1 3 04 : gas of conventional technology 1 3 0 5 : overflow of gas flow 1 3 0 6 : pressure change 1 3 07 : plasma of conventional technology 1 3 0 8 : plasma during ignition Injection 1309: Plasma injection power 1310: Luminous intensity is sufficient 13 11: Within TiN etching 13 12 : TiN etching outside 13 13: TiN ME Step: 13 14 : Ti luminous intensity starts 1 3 1 5: Fluorine-added gas flow in response to high-speed response response 1 3 1 6 : Ripple control waveform 1401: Hard mask uranium engraving 1402: Part of the boundary of the void pole S time reduction time flow control waveform (overshoot) The fluctuation of the pressure fluctuation in the power is increased. The temperature of the peripheral portion of the electrode is the temperature of the peripheral portion. The frequency of the frequency is reduced. The gas introduction mechanism is controlled to become a single groove-44-

Claims (1)

200849325 十、申請專利範圍 1 . 一種電漿處理方法, 具備: 載置工程,將具備膜構造之晶圓載置於真空容器內部 之處理室內之下部電極上,該膜構造爲,表面由在high-k 材料上包含有金屬材料的多數層之膜所構成、具有段差構 造者; 導入工程,將蝕刻氣體導入該處理室內; 調整工程,調整處理壓力; 產生工程,於上述處理室內產生電漿;及 供給工程’供給多數頻率之偏壓電力而於上述晶圓上 形成偏壓電位; 藉由變化上述多數頻率之偏壓電力輸出爲不同而進行 上述晶圓之膜構造之電漿處理者; 其特徵爲: 上述電漿處理方法另具有: 檢測工程,檢測上述電漿狀態之時間變化; 判斷工程,依據該檢測結果來判斷電漿處理之終點; 及 控制工程,於上述終點判斷之後,藉由變化多數頰率 之偏壓輸出及彼等之混合比,而獨立控制射入上述晶圓之 離子能量及其分布。 2、一種電漿處理方法, 具備: -45 - 200849325 載置工程,將具備膜構造之晶圓載置於真空容器內部 之處理室內之下部電極上,該膜構造爲,表面由在high-k 材料上包含有金屬材料的多數層之膜所構成、具有段差構 者, 導入工程,將蝕刻氣體導入該處理室內; 調整工程,調整處理壓力; 產生工程,於上述處理室內產生電漿;及 供給工程,供給多數頻率之偏壓電力而於上述晶圓上 形成偏壓電位; 藉由變化上述多數頻率之偏壓電力輸出爲不同而進行 上述膜構造之電漿處理者; 其特徵爲: 上述電漿處理方法另具有: 檢測工程,檢測電漿之阻抗之時間變化; 分離工程,將檢測出之電漿之阻抗之時間變化,分離 爲壁面狀態成份與晶圓正上方成份之阻抗; 判斷工程,依據分離出的壁面狀態成份或晶圓正上方 成份之阻抗檢測結果,來判斷電漿處理之終點;及 控制工程,於終點判斷之後,藉由變化多數頻率之偏 壓輸出及彼等之混合比,而獨立控制射入上述晶圓之離子 能量及其分布。 3· —種電漿處理方法, 具備· 載置工程’將具備膜構造之晶圓載置於真空容器內部 -46- 200849325 之處理室內之下部電極上,該膜構造爲,表面由在high-k 材料上包含有金屬材料的多數層之膜所構成、具有段差構 造者; 導入工程,將蝕刻氣體導入該處理室內; 調整工程,調整處理壓力; 產生工程,於上述處理室內產生電漿;及 供給工程,供給多數頻率之偏壓電力而於上述晶圓上 形成偏壓電位; 藉由變化上述多數頻率之偏壓電力輸出爲不同而進行 上述晶圓之膜構造之電漿處理者; 其特徵爲= 上述電漿處理方法另具有: 檢測工程,檢測電漿狀態之時間變化;及 控制工程,依據檢測出之電漿狀態之時間變化,藉由 變化多數頻率之偏壓輸出及彼等之混合比,而獨立控制射 入上述晶圓之離子能量及其分布。 4.如申請專利範圍第2項之電漿處理方法,其中 上述電漿處理方法另具有: 比較工程,於分離爲壁面狀態成份與晶圓正上方成份 之阻抗的分離工程之後,將分離出之資料和資料庫或變動 模型加以比較;及 依據該比較結果進行壁面潔淨的工程,或變化次回之 晶圓處理條件的工程。 5 ·如申請專利範圍第2項之電漿處理方法,其中 一 47- 200849325 上述電漿處理方法,係將電漿處理之終點判斷工程之 中、相對於高頻側偏壓電力之全部偏壓電力的輸出比設爲 30%以上。 6.如申請專利範圍第2項之電漿處理方法,其中 另具有:在終點判斷後之移至次一鈾刻步驟的過度時 間中、使電子溫度、電漿密度、氣體種、離子能量與離子 能量分布、晶圓上之彼等之面內分布呈現單調遞減或單調 遞增而變化的工程。 7 · —種電漿處理裝置,係具備:真空容器;下部電 極’配置於該真空容器之處理室內,其上面載置電漿處理 對象之晶圓;偏壓施加機構,供給多數頻率之偏壓電力而 於該下部電極形成偏壓電位;氣體供給機構,將反應性氣 體導入上述處理室內;調整機構,調整上述處理室內之氣 體壓力;及電磁波供給機構,於上述處理室內產生電漿; 其特徵爲: 上述偏壓施加機構,係具備: 變化機構,其藉由變化上述多數頻率之偏壓電力之輸 出比’而獨立變化射入晶圓之離子能量及其分布; 檢測機構,其藉由檢測電漿之電流、電壓、相位或阻 抗’而檢測出相對於多數頻率之偏壓頻率的電漿狀態;及 終點判斷機構,其藉由檢測出相對於多數頻率之偏壓 頻率的電漿狀態,來判斷電漿處理之終點。 8·如申請專利範圍第7項之電漿處理裝置,其中 上述偏壓施加機構, -48 - 200849325 係由振盪產生多數頻率的電源部及和由多數匹配器電 源部供給至上述下部電極的上述多數頻率所對應之匹配器 構成、而且具備對多數頻率之阻抗檢測器,或檢測相對於 多數頻率之電流、電壓、相位的機構。 9.如申請專利範圍第7項之電漿處理裝置,其中 由上述偏壓施加機構施加之多數頻率之偏壓電力之中 、高頻側之偏壓電力設爲1MH z、l〇〇MH z以下或上 述電磁波供給機構之電磁波之頻率以下, 低頻側之偏壓電力設爲1 0 0 k Η z以上、未滿4 M HE ζ ο 1 0 ·如申請專利範圍第7項之電漿處理裝置,其中 具備:針對供給電力至上述電磁波供給機構的電源記 錄多數個以上之匹配點,選擇其之匹配路徑而進行匹配的 匹配裝置。 -49-200849325 X. Patent application scope 1. A plasma processing method comprising: a mounting project, wherein a wafer having a film structure is placed on a lower electrode of a processing chamber inside a vacuum vessel, the membrane being constructed such that the surface is at a high- k material comprising a plurality of layers of a metal material and having a step structure; introducing a project to introduce an etching gas into the processing chamber; adjusting the process to adjust the processing pressure; generating a process to generate a plasma in the processing chamber; a supply process of supplying a bias power of a plurality of frequencies to form a bias potential on the wafer; and a plasma processor that performs a film structure of the wafer by varying a bias power output of the plurality of frequencies; The utility model has the following features: the plasma processing method further comprises: detecting a project, detecting a time change of the plasma state; determining a project, determining an end point of the plasma processing according to the detection result; and controlling the project, after determining the end point, by Changing the bias output of most buccal rates and their mixing ratio, and independently controlling the injection into the above crystal The ion energy and its distribution. 2. A plasma processing method comprising: -45 - 200849325 a mounting project, wherein a wafer having a film structure is placed on a lower electrode of a processing chamber inside a vacuum vessel, the membrane being constructed such that the surface is made of a high-k material a film comprising a plurality of layers of a metal material, having a step structure, introducing a project, introducing an etching gas into the processing chamber; adjusting a project to adjust a processing pressure; generating a project to generate a plasma in the processing chamber; and supplying the project And supplying a bias potential to a plurality of frequencies to form a bias potential on the wafer; and a plasma processor that performs the film structure by varying a bias power output of the plurality of frequencies; wherein: The slurry processing method further has: detecting the project, detecting the time change of the impedance of the plasma; separating the project, separating the time change of the detected impedance of the plasma into the impedance of the wall state component and the component directly above the wafer; Judging plasma treatment based on the separated wall state component or the impedance detection result of the component directly above the wafer The end point; and control engineering, after the endpoint is judged, the ion energy and its distribution injected into the wafer are independently controlled by varying the bias output of the majority of frequencies and their mixing ratios. 3. A plasma treatment method, equipped with a mounting project, in which a wafer having a film structure is placed on a lower electrode of a processing chamber inside a vacuum vessel -46-200849325, the membrane is constructed such that the surface is at high-k The material comprises a film of a plurality of layers of a metal material and has a step structure; an introduction process, an etching gas is introduced into the processing chamber; an adjustment process is performed to adjust a processing pressure; and a process is generated to generate a plasma in the processing chamber; and supply Engineering for supplying a bias potential to a plurality of frequencies to form a bias potential on the wafer; and a plasma processor for performing a film structure of the wafer by varying a bias power output of the plurality of frequencies; For the above plasma treatment method, there are: a detection project to detect the time change of the plasma state; and a control project, according to the time variation of the detected plasma state, by changing the bias output of most frequencies and their mixing The ion energy and its distribution incident on the wafer are independently controlled. 4. The plasma processing method according to claim 2, wherein the plasma processing method further comprises: a comparison engineering, after separating the separation of the wall state component from the impedance of the component directly above the wafer, separating the The data is compared with a database or a variation model; and the wall cleaning process is performed based on the comparison result, or the process of changing the wafer processing conditions of the second time. 5 · The plasma processing method of claim 2, wherein one of the above-mentioned plasma processing methods is to judge the end point of the plasma treatment, and all the bias voltages with respect to the high-frequency side bias power. The output ratio of power is set to 30% or more. 6. The plasma processing method according to item 2 of the patent application, wherein the method further comprises: in the excessive time of moving to the next uranium engraving step after the end point judgment, the electron temperature, the plasma density, the gas species, the ion energy and The ion energy distribution, their in-plane distribution on the wafer, exhibits a monotonically decreasing or monotonically increasing change. 7) A plasma processing apparatus comprising: a vacuum container; a lower electrode 'disposed in a processing chamber of the vacuum container, on which a wafer for plasma processing is placed; and a bias applying mechanism for supplying a bias of a plurality of frequencies Electric power generates a bias potential at the lower electrode; a gas supply mechanism introduces a reactive gas into the processing chamber; an adjustment mechanism adjusts a gas pressure in the processing chamber; and an electromagnetic wave supply mechanism generates a plasma in the processing chamber; The bias applying mechanism includes: a changing mechanism that independently changes an ion energy of the incident wafer and a distribution thereof by changing an output ratio of the bias power of the plurality of frequencies; and a detecting mechanism Detecting the current, voltage, phase, or impedance of the plasma to detect a plasma state relative to a bias frequency of a plurality of frequencies; and an endpoint determining mechanism for detecting a plasma state with respect to a bias frequency of a plurality of frequencies To determine the end of the plasma treatment. 8. The plasma processing apparatus according to claim 7, wherein the bias applying means -48 - 200849325 is a power supply unit that generates a plurality of frequencies by oscillation and the above-mentioned supply of the plurality of matching power supply units to the lower electrode It is composed of a matching unit corresponding to a plurality of frequencies, and has an impedance detector for a plurality of frequencies or a mechanism for detecting a current, a voltage, and a phase with respect to a plurality of frequencies. 9. The plasma processing apparatus according to claim 7, wherein the bias power of the plurality of frequencies applied by the bias applying means is set to 1 MHz, and the bias power of the high frequency side is set to 1 MHz, 1 〇〇 MH z In the following or below the frequency of the electromagnetic wave of the electromagnetic wave supply mechanism, the bias power on the low frequency side is set to 1 0 0 k Η z or more, and less than 4 M HE ζ ο 1 0 · The plasma processing apparatus according to item 7 of the patent application scope Further, there is provided a matching device that performs matching by selecting a matching path of a plurality of matching points for supplying power to the power supply unit of the electromagnetic wave supply unit. -49-
TW97104310A 2007-02-13 2008-02-04 Plasma processing method and plasma processing apparatus TW200849325A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW97104310A TW200849325A (en) 2007-02-13 2008-02-04 Plasma processing method and plasma processing apparatus

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
TW96105248 2007-02-13
JP2008002709A JP5014166B2 (en) 2007-02-13 2008-01-10 Plasma processing method and plasma processing apparatus
TW97104310A TW200849325A (en) 2007-02-13 2008-02-04 Plasma processing method and plasma processing apparatus

Publications (2)

Publication Number Publication Date
TW200849325A true TW200849325A (en) 2008-12-16
TWI366861B TWI366861B (en) 2012-06-21

Family

ID=44824140

Family Applications (1)

Application Number Title Priority Date Filing Date
TW97104310A TW200849325A (en) 2007-02-13 2008-02-04 Plasma processing method and plasma processing apparatus

Country Status (1)

Country Link
TW (1) TW200849325A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI418256B (en) * 2009-10-09 2013-12-01 Hitachi High Tech Corp Plasma processing device
TWI689966B (en) * 2015-10-29 2020-04-01 美商應用材料股份有限公司 Method for atomic precision etching
TWI763223B (en) * 2020-12-31 2022-05-01 華邦電子股份有限公司 Etching apparatus and etching method thereof

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6140412B2 (en) * 2012-09-21 2017-05-31 東京エレクトロン株式会社 Gas supply method and plasma processing apparatus

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI418256B (en) * 2009-10-09 2013-12-01 Hitachi High Tech Corp Plasma processing device
US8992721B2 (en) 2009-10-09 2015-03-31 Hitachi High-Technologies Corporation Plasma processing apparatus
US10262840B2 (en) 2009-10-09 2019-04-16 Hitachi High-Technologies Corporation Plasma processing apparatus
TWI689966B (en) * 2015-10-29 2020-04-01 美商應用材料股份有限公司 Method for atomic precision etching
TWI763223B (en) * 2020-12-31 2022-05-01 華邦電子股份有限公司 Etching apparatus and etching method thereof

Also Published As

Publication number Publication date
TWI366861B (en) 2012-06-21

Similar Documents

Publication Publication Date Title
JP5568667B2 (en) Plasma processing method
US10861678B2 (en) Plasma etching apparatus and method
KR100978886B1 (en) Plasma processing method and plasma processing apparatus
KR102038174B1 (en) Semiconductor device manufacturing method
US6660127B2 (en) Apparatus for plasma etching at a constant etch rate
JP4733214B1 (en) Mask pattern forming method and semiconductor device manufacturing method
TWI559393B (en) Etch process for 3d flash structures
US7736914B2 (en) Plasma control using dual cathode frequency mixing and controlling the level of polymer formation
JP2002543613A (en) Techniques for etching low capacitance dielectric layers
JP6277004B2 (en) Dry etching method
US9418863B2 (en) Method for etching etching target layer
KR20080006457A (en) Plasma etching method and computer-readable storage medium
JP4351806B2 (en) Improved technique for etching using a photoresist mask.
TW200849325A (en) Plasma processing method and plasma processing apparatus
US20150041060A1 (en) Plasma processing apparatus
JP2917993B1 (en) Dry etching method
US11955316B2 (en) Substrate processing method, method for manufacturing semiconducor device, and plasma processing apparatus
US20230170188A1 (en) Method of Plasma Etching
JP2003068709A (en) Dry etching method
JP2003289066A (en) Semiconductor device and its manufacturing method