200949932 六、發明說明： 【發明所屬之技術領域】 本發明係關於蝕刻量算出方法、記憶媒體以及蝕刻量 算出裝置，尤其關於使用光罩膜在晶圓上形成溝渠或孔等 之凹部時的蝕刻量算出方法。 【先前技術】 © 在半導體裝置之製造工程中，在晶圓執行使用光罩膜 在被蝕刻層形成溝渠或孔之蝕刻。雖然在蝕刻中不被光罩 膜所覆蓋之部分的被蝕刻層藉由電漿以物理性、化學性之 方式被削除，但是在溝渠之形成中，必須控制該溝渠之深 度。因此，必須於蝕刻中算出溝渠深度，即是蝕刻量，以 往蝕刻量之算出方法係廣泛使用利用光干涉的方法。 第22圖爲用以說明蝕刻中之光干涉的圖式。 在第22圖中，於具有被形成在被蝕刻層130上之光 ® 罩膜131的晶圓W，藉由蝕刻形成有溝渠132，但是當對 該晶圓W照射雷射光L1之時，則產生來自光罩膜131表 面之反射光L2、來自光罩膜131及被蝕刻層130之境界 ^ 面之反射光L3以及來自溝渠132之底部的反射光L4。 於以檢測器接受反射光L2〜L4之時，則如第22圖所 示般，因各反射光之光路長僅有光罩膜131之厚度或溝渠 132之深度部分不同，故在檢測器之受光面各反射光之相 位爲不同，產生干涉光（例如反射光L2及反射光L4之干 涉光（以下稱爲「溝渠干涉光」）或反射光L2及反射光 -5- 200949932 L3之干涉光（以下稱爲「光罩膜干涉光」）。 然後，在鈾刻中，因溝渠1 32之深度時時刻刻在變化 ，故反射光L2和反射光L4之光路長差也時時刻刻在變化 而干涉光之強度變化，即是自反射光L2和反射光L4產生 干涉波（以下稱爲「溝渠干涉波」）。干涉波之週期因藉 由溝渠132之深度之變化速度（蝕刻率）而決定，故自干 ’BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an etching amount calculation method, a memory medium, and an etching amount calculation device, and more particularly to etching when a concave portion such as a trench or a hole is formed on a wafer using a photomask film. The method of calculation. [Prior Art] © In the manufacturing process of a semiconductor device, etching using a photomask film to form a trench or a hole in an etched layer is performed on a wafer. Although the portion of the etched layer that is not covered by the photomask during etching is physically and chemically removed by the plasma, the depth of the trench must be controlled in the formation of the trench. Therefore, it is necessary to calculate the depth of the trench in the etching, that is, the etching amount, and the method of calculating the etching amount is widely used by the method of using light interference. Figure 22 is a diagram for explaining the interference of light in etching. In Fig. 22, a trench 132 is formed by etching on a wafer W having a light-shielding film 131 formed on the layer to be etched 130, but when the wafer W is irradiated with the laser light L1, The reflected light L2 from the surface of the photomask film 131, the reflected light L3 from the boundary surface of the photomask film 131 and the layer to be etched 130, and the reflected light L4 from the bottom of the trench 132 are generated. When the detector receives the reflected light L2 to L4, as shown in FIG. 22, since the optical path length of each reflected light is different only by the thickness of the photomask film 131 or the depth of the trench 132, the detector is Interfering light (for example, interference light of reflected light L2 and reflected light L4 (hereinafter referred to as "ditch interference light") or reflected light L2 and reflected light - 5, 200949932 L3 interference light is generated when the phases of the reflected light are different. (hereinafter referred to as "mask film interference light"). Then, in the uranium engraving, since the depth of the trench 1 32 changes moment by moment, the optical path length difference between the reflected light L2 and the reflected light L4 is also constantly changing. The intensity of the interference light changes, that is, the interference wave (hereinafter referred to as "ditch interference wave") is generated from the reflected light L2 and the reflected light L4. The period of the interference wave is due to the change speed (etching rate) of the depth of the trench 132. Decide, so do it yourself'
涉波之週期算出蝕刻率，並且自所算出之蝕刻率和蝕刻時 間算出蝕刻量（溝渠132之深度）。 Q 再者，在蝕刻中因光罩膜131也以微量逐漸被蝕刻而 厚度產生變化，故也自反射光“及反射光L3產生干涉波 (以下稱爲「光罩膜干涉波」）。因各干涉波被相同檢測 器檢測出，故該檢測器所檢測出之干涉波成爲重疊具有不 同週期之多數干涉波者（以下稱爲「重疊干涉波」）（參 照第2 3圖）。 爲了自第23圖所示之重疊干涉波算出溝渠132之深 度（被蝕刻層1 3 0之蝕刻量），必須自重疊干涉波分離溝 ❹ 渠干涉波。 在第23圖之重疊干涉波中，短週期之干涉波和長週 期之干涉波能夠比較明確地分離。在此，蝕刻中之溝渠 132之深度之變化速度因大於光罩膜131之厚度之變化速 度，故溝渠干涉波之週期較光罩膜干涉波之週期短。因此 ，第23圖之重疊干涉波中之短週期的干涉波爲溝渠干涉 波，可以自短週期之干涉波中之極値間之時間（圖中之「 △ t」容易算出溝渠干涉波之週期。 -6- 200949932 自重疊干涉波讀取極値間之時間的方法中，因重疊干 涉波中之短週期的干涉波和長週期之干涉波必須能夠比較 明確地分離，故短週期之干涉波和長週期之干涉波難以分 離之重疊干涉波係無法算出溝渠干涉波之週期。再者，在 短週期之干涉波中之極値間因溝渠干涉波之週期視爲一定 ’故所算出之溝渠干涉波之週期（被触刻層130之蝕刻率 )則如第24圖所示般，成爲台階狀。即是，自重疊干涉 ® 波讀取極値間之時間的方法爲分解能低。 因此，近年開發有不自重疊干涉波讀取極値間之時間 藉由頻率解析算出溝渠干涉波之週期的方法。該方法係自 重疊干涉波藉由頻率解析（例如高速傅立葉變換法）取得 頻率分佈（參照第26圖（A))，自該頻率分佈檢測出 溝渠干涉波之週期（例如，參照專利文獻1)。 [專利文獻1]日本特開平2-71517號公報 【發明內容】 (發明所欲解決之課題） 但是，如第25圖所示般，在重疊千涉波有因雷射光 源或檢測器之異常（圖中之「I」）或光罩膜干涉波和溝 渠干涉波之干涉所產生外表上之週期變化（圖中之「II」 )等的干擾加注於重疊干涉波之情形。上述使用頻率解析 之方法中，因從解析蝕刻中經過所有時間的重疊干涉波而 所取得之頻率分佈僅檢測出溝渠干涉波之週期，故對重疊 干涉波給予干擾之時，則在原本不存在之干涉週期產生峰 200949932 値等，頻率分佈成爲不正確（參照第26圖（B))，其結 果無法正確安定算出蝕刻量。 本發明之目的係提供即使給予干擾亦可以正確安定算 出蝕刻量之蝕刻量算出方法、記憶媒體及蝕刻量算出裝置 (用以解決課題之手段） 爲了達成上述目的，申請專利範圍第1項所記載之蝕 © 刻量算出方法，係在使用光罩膜形成凹部的基板蝕刻中算 出上述凹部之蝕刻量，其特徵爲：具有對上述基板照射光 之照射步驟；接受至少來自上述光罩膜之反射光及來自上 述凹部之底部的反射光之干涉光被重疊於其他干涉光之重 疊干涉光的受光步驟；自上述被接受之重叠干涉光算出重 疊干涉波之干涉波算出步驟；自上述重疊干涉波抽出特定 期間之波形的波形抽出步驟；對上述被抽出之波形施予頻 率解析的頻率解析步驟；自藉由上述頻率解析所取得之頻 〇 率分佈檢測出來自上述光罩膜之反射光及來自上述凹部之 底部的反射光之干涉波之週期的干涉週期檢測步驟；一邊 使上述特定期間僅偏移特定時間一邊重複上述干涉波算出 步驟、上述波形抽出步驟、上述頻率解析步驟及上述干涉 週期檢測步驟，於每次重複累積平均上述被檢測出之干涉 波之週期的累積平均步驟；和根據上述被累積平均之干涉 波之週期算出上述凹部之蝕刻量的蝕刻量算出步驟。 申請專利範圍第2項所記載之蝕刻量算出方法係如申 -8 - 200949932 請專利範圍第1項所記載之蝕刻量算出方法中，上述特定 期間係大於比來自上述光罩膜之反射光及來自上述凹部之 底部之反射光的干涉波之週期長的上述其他干涉光之波形 之1週期。 申請專利範圍第3項所記載之蝕刻量算出方法係如申 請專利範圍第1或2項所記載之蝕刻量算出方法中，又具 有於上述其他干涉光之波形之週期比來自上述光罩膜之反 〇 射光及來自上述凹部之底部之反射光之干涉波之週期長時 ，於上述頻率解析步驟之前，由自上述重疊干涉波所抽出 之特定期間之波形除去上述其他干涉光之波形所佔有之大 部分的解析前處理步驟，上述頻率解析步驟係對除去上述 其他干涉光之波形所佔有之大部分的波形施予頻率解析。 如申請專利範圍第4項所記載之蝕刻量算出方法係如 申請專利範圍第3項所記載之蝕刻量算出方法中’在上述 解析前處理步驟中，自上述被抽出之波形除去以二次多項 ® 式使該被抽出之波形近似之波形。 申請專利範圍第5項所記載之蝕刻量算出方法係如申 請專利範圍第3或4項所記載之飩刻量算出方法中’上述 特定期間爲上述其他干涉光之波形之1/4週期以下。 申請專利範圍第6項所記載之蝕刻量算出方法係如申 請專利範圍第3至5項中之任一項所記載之蝕刻量算出方 法中，上述基板之表面中的上述凹部之開口率爲0·5%以 下或是上述凹部爲深溝渠。 申請專利範圍第7項所記載之蝕刻量算出方法係如申 -9- 200949932 請專利範圍第1至6項中之任一項所記載之蝕刻量算出方 法中，上述頻率解析係使用最大熵法（maximum entropy method )。 申請專利範圍第8項所記載之蝕刻量算出方法係如申 請專利範圍第1至7項中之任一項所記載之蝕刻量算出方 法中，又具有於自上述頻率分佈所檢測出之上述干涉波之 週期相當於異常値之時，除去該干涉波之週期的干涉週期 修正步驟。 © 申請專利範圍第9項所記載之蝕刻量算出方法係如申 請專利範圍第8項所記載之蝕刻量算出方法中，在上述干 涉週期修正步驟中，將自求得相當於上述異常値之上述干 渉波之週期的上述特定期間之前的上述特定期間或是之後 的上述特定期間所求出之上述干涉波之週期，視爲取得相 當於上述異常値之上述干涉波之週期的上述特定期間之干 涉波之週期。 申請專利範圍第1 〇項所記載之蝕刻量算出方法係如 〇 申請專利範圍第1至9項中之任一項所記載之蝕刻量算出 方法中，事先預測來自上述光罩膜之反射光及來自上述凹 部之底部之反射光的干涉波之週期，在上述干涉週期檢測 步驟中，在藉由上述頻率解析所取得之頻率分佈中，自上 述預測之週期附近檢測出來自上述光罩膜之反射光及來自 上述凹部之底部之反射光的干涉波之週期。 申請專利範圍第1 1項之飩刻量算出方法係如申請專 利範圍第1至10項中之任一項所記載之蝕刻量算出方法 -10- 200949932 中’上述其他之干涉光爲來自上述光罩膜表面之反射光， 以及來自上述光罩膜及上述基板表面之境界面之反射光的 干涉光。 .爲了達成上述目的，申請專利範圍第12項所記載之 記憶媒體，屬於儲存有使電腦實行蝕刻量算出方法之程式 的電腦可讀取的記憶媒體，上述蝕刻量算出方法係在使用 光罩膜形成凹部之基板蝕刻中算出上述凹部之蝕刻量，其 ® 特徵爲：上述蝕刻量算出方法具有對上述基板照射光之照 射步驟；接受至少來自上述光罩膜之反射光及來自上述凹 部之底部的反射光之干涉光被重疊於其他干涉光之重叠干 涉光的受光步驟；自上述被接受之重疊干涉光算出重疊干 涉波之干涉波算出步驟；自上述重疊干涉光抽出特定期間 之波形的波形抽出步驟；對上述被抽出之波形施予頻率解 析的頻率解析步驟；自藉由上述頻率解析所取得之頻率分 佈檢測出來自上述光罩膜之反射光及來自上述凹部之底部 ^ 的反射光之干涉波之週期的干涉週期檢測步驟；一邊使上 述特定期間僅偏移特定時間一邊重複上述干涉波算出步驟 . 、上述波形抽出步驟、上述頻率解析步驟及上述干涉週期 . 檢測步驟，於每次重複累積平均上述,被檢測出之干涉波之 週期的累積平均步驟；和根據上述被累積平均之干涉波算 出上述凹部之蝕刻量的蝕刻量算出步驟。 爲了達成上述目的’申請專利範圍第13項所記載之 蝕刻量算出裝置’係在使用光罩膜形成凹部的基板蝕刻中 算出上述凹部之蝕刻量，其特徵爲：具有對上述基板照射 -11 - 200949932 光之照射部；接受至少來自上述光罩膜之反射光及來自上 述凹部之底部的反射光之干涉光被重疊於其他干涉光之重 疊干涉光的受光部；自上述被接受之重疊干涉光算出重疊 干涉波之干涉波算出部；自上述重疊干涉光抽出特定期間 之波形的波形抽出部；對上述被抽出之波形施予頻率解析 的頻率解析部；自藉由上述頻率解析所取得之頻率分佈檢 測出來自上述光罩膜之反射光及來自上述凹部之底部的反 射光之干涉波之週期的干涉週期檢測部；一邊使上述特定 期間僅偏移特定時間一邊重複上述重叠干涉波之算出、上 述特定期間之波形抽出、上述頻率解析及上述干涉波之週 期檢測，於每次重複累積平均上述被檢測出之干涉波之週 期的累積平均部；和根據上述被累積平均之干涉波之週期 算出上述凹部之蝕刻量的鈾刻量算出部。 [發明效果] 若藉由申請專利範圍第1項所記載之蝕刻量算出方法 、申請專利範圍第12項所記載之記憶媒體及申請專利範 圍第13項所記載之蝕刻量算出裝置時，一邊使特定期間 僅偏移特定時間一邊重複重疊干涉波之算出、特定期間之 波形抽出、頻率解析以及檢測出來自光罩膜之反射光及來 自凹部底部之反射光的干涉波之週期，於每次重覆時累積 平均被檢測出之干涉波之週期，根據被累積平均之干涉波 之週期算出凹部之蝕刻量。因此，即使例如對被抽出之某 特定期間之波形給予干擾之時，根據該某特定期間之波形 -12- 200949932 而所檢測出之干涉波之週期，因與根據其他特定期間之波 形而所檢測出之干涉波之週期累積平均，故可以縮小根據 給予干擾的特定期間之波形而所檢測出之干涉波之週期， 對累積平均之干涉波之週期所造成之影響，因此即使給予 干擾也可以正確安定執行算出蝕刻量。 若藉由申請專利範圍第2項所記載之触刻量算出方法 時，上述特定時間因大於較來自光罩膜之反射光及來自凹 © 部之底部之反射光的干涉波之週期長的其他干涉光之波形 的1週期大，故可以提高特定期間之波形之頻率解析的信 賴性，因此可以更正確執行算出蝕刻量。 若藉由申請專利範圍第3項所記載之蝕刻量算出方法 ，因於其他之干渉光之波形的週期較來自光罩膜之反射光 及來自凹部之底部之反射光的干涉波之週期長之時，於頻 率解析之前，由自重疊干涉波所抽出之特定期間之波形除 去其他干涉光之波形所佔有之大部分，故即使來自凹部之 ^ 底部之反射光之光量少時，其他干涉光之波形佔據重疊干 涉波之大部分之時，亦可以在除去後之波形增大來自光罩 . 膜之反射光及來自凹部之底部之反射光的干涉光所佔有部 . 分的比率，因此在頻率解析中可以正確算出來自光罩膜之 反射光及來自凹部之底部之反射光的干涉波之週期。 若藉由申請專利範圍第4項所記載之蝕刻量算出方法 時，自於頻率解析前抽出之波形，除去以二次多項式近似 該被抽出之波形的波形。於來自凹部之底部之反射光之光 量少時，其他干涉光之波形佔據重疊干涉波之大部分之時 -13- 200949932 ，因重疊干涉波與其他干涉光之波形幾乎相等，故以二次 多項式近似被抽出之重疊干涉波之波形也與其他干涉光之 波形幾乎相等。因此，可以自所抽出之波形確實除去其他 干涉光之波形所佔有之大部分。 若藉由申請專利範圍第5項所記載之蝕刻量算出方法 _ ，上述特定期間爲其他干涉光之波形之1/4週期以下。因 佔據重疊干涉波之大部分的其他千涉光之波形接近於正弦 波，故若抽出其他干涉光之波形之1/4週期以下之部分， @ 該被抽出之波形藉由二次多項式可以正確近似。依此，可 以自所抽出之波形正確除去其他干涉光之波形所佔有之大 部分。 若藉由申請專利範圍第7項所記載之蝕刻量算出方法 時，因頻率解析使用最大熵法，故即使特定期間之波形之 數量少，亦可以提高頻率解析之信賴性，因此可以更正確 執行蝕刻量之算出。 若藉由申請專利範圍第8項所記載之鈾刻量算出方法 〇 時，因於自頻率分佈所檢測出之干涉波之週期相當於異常 値之時，除去該干涉波之週期，故可以除去根據受到干擾 的特定期間之波形而所檢測出之干涉波之週期，對累積平 均之干涉波之週期所造成之影響，因此即使受到干擾亦可 以更安定正確算出蝕刻量。 若藉由申請專利範圍第9項所記載之鈾刻量算出方法 時，因將自求得相當於異常値之干涉波之週期的特定期間 之前的特定期間或是之後的特定期間所求出之千涉波之週 -14- 200949932 期，視爲求得相當於異常値之干涉波之週期的特定期間之 干涉波之週期，故可以確實除去根據受到干擾之特定期間 之波形所檢測出之干涉光之影響。 若藉由申請專利範圍第10項所記載之蝕刻量算出方 法時，在事先預測來自光罩膜之反射光及來自凹部底部之 反射光之干涉波之週期，並藉由特定期間之波形之頻率解 析所取得之頻率分佈中，因自所預測之週期附近檢測出干 涉波之週期，故可以迅速執行干涉波之週期的檢測，並且 可以抑制檢測出異常値。 【實施方式】 以下，針對本發明之實施型態，參照圖面予以說明。 首先，針對適用本發明之第1實施型態所涉及之蝕刻 量算出方法及蝕刻量算出方法的基板處理裝置予以說明。 該基板處理裝置係構成對當作基板之半導體晶圓（以下單 稱爲「晶圓」）W施予使用電漿之蝕刻。並且，晶圓W 係如上述第22圖所示般，具有被蝕刻層130和以特定圖 案被形成在該被蝕刻層130上之光罩膜131。 第1圖爲槪略性表示適用本實施型態所涉及之蝕刻量 算出方法之基板處理裝置之構成的剖面圖。 在第1圖中，基板處理裝置10具備有由例如鋁等之 導電性材料所構成之處理室11，和當作載置晶圓W之載 置台而被配設在處理室11內之底面之下部電極12，和隔 著特定間隔被配設在該下部電極12之上方的噴灑頭13。 -15- 200949932 在處理室11之下部連接有真空排氣裝置（無圖 所連接之排氣部14，在下部電極12經整合器15連 高頻電源16，在噴灑頭13之內部之緩衝室17連接 氣體導入管18，在該處理氣體導入管18連接有處理 供給裝置19。噴灑頭13係在下部，具有使緩衝室Π 灑頭13以及屬於下部電極12間之空間的處理空間S 之多數氣體穴20。噴灑頭13係將從處理氣體導入1 被導入至緩衝室17之處理氣體經多數氣體穴20供給 理空間S。 該基板處理裝置1〇係藉由排氣部14將處理室 減壓至特定真空度之後，在自下部電極12將高頻電 加至處理空間S之狀態下，從噴灑頭13將處理氣體 至處理空間S，在處理空間S自處理氣體產生電漿。 生的電漿係在晶圓W衝突、接觸於不被光罩膜131 之被蝕刻層130而蝕刻該被蝕刻層130，在該被蝕 130形成溝渠132 (凹部）。 在處理室11內之噴灑頭13配設有自上方觀測被 於下部電極12之監視裝置21。監視裝置21係由圓 之構件所構成，貫通噴灑頭13。在監視裝置21之上 置有由石英玻璃等之透明體所構成之窗構件22。再 在處理室11之上方配置經監視裝置21之上端和聚光 23而對向之光纖24。 光纖24係連接於算出被蝕刻層130之蝕刻量的 量算出裝置25。蝕刻量算出裝置25係具備各連接於 式） 接有 處理 氣體 、噴 連通 f 1 8 至處 1內 壓施 供給 該發 覆蓋 刻層 載置 筒狀 端設 者， 透鏡 蝕刻 光纖 -16- 200949932 24之雷射光源26(照射部）及檢測器27(受光部），和 連接於檢測器27之運算部28 (干涉波算出部、波形抽出 部、頻率解析部、干涉頻率檢測部、累積平均部、蝕刻量 算出部），在基板處理裝置10之控制器29之控制下動作 。作爲雷射光源26使用例如半導體雷射。再者，作爲檢 測器 27使用例如雷射光源 27使用光電倍增管（ Photomultiplier)或發光二極體。並且，控制器29不僅 運算部28’也連接於基板處理裝置1〇之各構成要素，例 如高頻電源1 6，控制各構成要素之動作。 蝕刻量算出裝置25係將來自雷射光源26之雷射光經 光纖24、聚光透鏡23及監視裝置21朝向下部電極12上 之晶圓W照射，並且經光纖24等藉由檢測器27接受重 疊來自晶圓W之反射光，即是溝渠干涉光（來自光罩膜 之反射光及來自凹部之底部之反射光的干涉光）或光罩膜 干涉光（其他干涉光）所重叠之重叠干涉光。藉由檢測器 27所接受到之重疊干涉光被變換成電性訊號而傳送至運 算部28。 運算部28係根據所接收之電性訊號而從重疊干涉光 算出重疊干涉波。再者，運算部28係根據所算出之重疊 干涉波，實行後述之第6圖之蝕刻量算出方法而算出溝渠 1 3 2之蝕刻量。 再者，在蝕刻中蝕刻率並非一定’由於各種因素（處 理空間S之壓力變化或高頻電壓之干擾）而產生變化。尤 其，以蝕刻形成縱橫比大之溝渠（例如深溝渠）之時’當 -17- 200949932 溝渠132之蝕刻量（蝕刻深度）變大之時，因在溝渠132 之入口具有附著物而抑制電漿進入至溝渠132內’故蝕刻 率下降（參照第2圖）。再者，如第2圖所示般’因鈾刻 率重複微小變化，故爲了正確算出溝渠1 3 2之蝕刻量’必 須逐漸地算出蝕刻率。 在蝕刻中爲了逐漸地算出蝕刻率，通常若在各時序將 來自晶圓W之反射光之波形予以微分即可，但是如上述 般，因來自晶圓W之反射光爲重疊溝渠干涉光或光罩膜 © 之重疊干涉光，故即使單純在各時序將反射光之波形予以 微分亦無法正確取得溝渠132之蝕刻率。 在此，在本實施型態中執行頻率解析，自重疊干涉波 算出溝渠干涉波之週期（以下稱爲「溝渠干涉週期」。」 再者，在頻率解析中由於使用解析對象之波形之1週期以 上之資料長，具體上有助於提高頻率解析之信賴性，故爲 了求出某時序之蝕刻率，在本實施型態中，自重疊干涉波 抽出特定期間之波形，將該被抽出之波形予以頻率解析。 〇 再者，如上述般，鈾刻率因與溝渠干涉頻率關聯，故在本 實施型態中，自被抽出之波形求出溝渠干涉週期，並自該 溝渠干涉波算出蝕刻率。 第3圖爲用以說明本實施型態所涉及之蝕刻量算出方 法中自重疊干涉波抽出特定期間之波形的圖式。 在第3圖中，因重疊干涉波3 0以大約3 0秒週期振動 ，故將上述特定期間設定成30秒間。在此，爲了求取時 序A之溝渠干涉週期，抽出從時序A往前30秒之期間31 -18- 200949932 中之重疊干涉波30的波形（以圖中之四角形所包圍之部 份之波形）。並且在本實施型態中將上述特定期間稱爲「 窗」。窗31具有起點32及終點33，起點32相當於從時 序A往前30秒，終點33相當於時序A。 然後，在本實施型態中，對所抽出之窗31之波形施 予頻率解析。在此，因所抽出之波形最多爲1週期份，故 使用最大熵法作爲解析方法。最大熵法爲自極短之測量時 © 間之測量結果以高分解能算出觀測現象之頻率分佈之方法 (「科學計算用之波形資料處理」（CQ出版社，昭和61 年4月30日初版發行），因並不那樣需要解析對象之波 形數，故比起需要幾個週期份之波形的高速傅立葉變換法 ，較適用於本實施型態所涉及之蝕刻量算出方法。 第4圖爲表示藉由使用最大熵法之頻率解析自第3圖 中之窗之波形所取得之頻率分佈之圖式。 在重疊干涉波30因主要含有溝渠干涉波和光罩膜干 ^ 涉波之兩個，故在窗31之波形所得之頻率分佈，如第4 圖所示般，主要存在兩個表示峰値之頻率（干涉波之週期 )(在第4圖中爲大約0·012Ηζ和大約〇.〇37Hz)。在此 . ，如上述般，因溝渠干涉波之週期較光罩膜干涉波之週期 短（頻率高），故大約0.037 Hz之頻率相當於溝渠干涉頻 率。在此，溝渠干涉週期檢測出大約0.03 7 Hz之頻率。如 此一來，在本實施型態中，於藉由頻率解析所取得之頻率 分佈預見存在兩個峰値，於頻率解析之前事先預測溝渠干 涉週期，自在所取得之頻率分佈中所預測之溝渠千涉週期 -19- 200949932 之附近檢測出溝渠週期爲佳。 窗31因包含從時序A往前30秒之重疊干涉波30之 波形，故第4圖所示之頻率分佈成爲自時序A往前30秒 之重叠干涉波30中之頻率分佈。因此，自第4圖所示之 頻率分佈所檢測出之溝渠干涉週期，雖然成爲自時序A 往前30秒之重疊干涉波30中之溝渠干涉光之平均週期， 但是在本實施形態中，爲了方便將自第4圖所示之頻率分 部所檢測出之溝渠干涉週期視爲時序A中之溝渠干涉週 期。並且，在本實施型態中，如後述般對重疊干涉波30 設定多數窗，累積平均自各窗之波形之頻率分佈所取得之 溝渠干涉週期，而算出溝渠干涉週期之全體平均値，故解 除將窗31中之溝渠干涉光之平均週期視爲時序A中之溝 渠干涉週期之弊害。 再者，在本實施型態所涉及之鈾刻量算出方法中，算 出從經過最初之特定期間（起點3 2相當於蝕刻開始時且 終點3 3對應於從蝕刻開始時往後3 0秒之窗3 1 )時至算 出蝕刻量之時序的所有期間中之蝕刻率的平均値，自該鈾 刻率之平均値算出蝕刻量。 第5圖爲用以說明本實施型態所涉及之蝕刻量算出方 法中蝕刻率之平均値之算出方法的圖式，表示蝕刻開始後 經過8 0秒之情形。 第5圖中，相對於重叠干涉波50設定有僅偏移At( 特定時間）之11個窗^^=1〜11，11爲自然數），在各 窗Wk求取頻率分佈，從各頻率分佈檢測出n個之溝渠干 -20- 200949932 渉週期€“1{：=1〜11，11爲自然數）。 接著，根據下述式（1)累積平均η個溝渠干涉週期 fk， [式1] fave= 2]fk/n …（1) k»l 溝渠干涉週期fave係當作蝕刻開始後8 0秒之溝渠干涉波 之平均値而被算出。並且，當將測量波長（來自雷射光源 26之雷射光之波長）設爲λ之時，則從下述式（2)算出 至蝕刻開始後80秒之蝕刻率之平均値。 蝕刻率之平均値=faveX λ /2 …（2) ❹ 接著，算出從下述式（3)至蝕刻開始後80秒之鈾刻 量。 蝕刻量=蝕刻率之平均値X蝕刻時間 …（3) 在本實施型態所涉及之蝕刻量算出方法中，當某窗 Wt(t爲1〜η中之任一者的自然數）中之重疊干涉波50 受到干擾時，從該窗wtm求出之溝渠干涉頻率ft成爲異 常値，但是從該溝渠干涉頻率ft與其他窗WU(U爲1〜η 中之任一者，t以外之自然數）所求出之溝渠干涉週期fu -21 - 200949932 累積平均，故對累積平均溝渠干涉週期ft之溝渠干涉週 期fave之影響爲小。 接著，針對本實施型態所涉及之蝕刻量算出方法予以 說明。 第6圖爲表示本實施型態所涉及之蝕刻量算出方法之 流程圖。 在第6圖中，首先基板處理裝置10開始蝕刻晶圓W 之被蝕刻層130之後，雷射光源26將雷射光L1經光纖 ❹ 24、聚光透鏡23及監視裝置21朝向晶圓W照射（步驟 S61 )(照射步驟）。檢測器27經光纖24等接受屬於來 自晶圓W之反射光的重疊干涉光（步驟S62)(受光步驟 )。 接著，在步驟S63，運算部28判別現時序T是否超 過事先所設定之蝕刻結束時間，於超過蝕刻結束時間之時 (步驟S6 3中YES )，結束本處理，於未超過鈾刻之結束 時間時（步驟S63中NO )，運算部28則根據藉由檢測器 ❹ 27所接受之重疊干涉光，算出（更新）自蝕刻開始時至 現時序T爲止之重疊干涉波（步驟s64)(干涉波算出步 驟）。 接著，運算部28抽出將現時序τ設爲終點之窗之波 形（步驟S65)(波形抽出步驟），使用最大熵法對該被 抽出之窗之波形施予頻率解析（步驟S66)(頻率解析步 驟）。在此’從窗之起點至終點之時間設定成較光罩膜干 涉波之1週期長。 -22- 200949932 之後，運算部28係在藉由頻率解析所取得之頻率分 佈中’將表示事先所預測之溝渠干涉週期附近之峰値的頻 率’當作現時序T中之溝渠干涉週期予以檢測出（步驟 S67)(干涉週期檢測步驟）。 接著，運算部28使用上述式（1)累積平均此次所檢 測出之現時序T中之溝渠干涉週期，和從經過最初之特定 時間時至現時序T爲至之期間被檢測出之各時序中之溝渠 © 干涉週期（步驟S68 )(累積平均步驟），根據從蝕刻開 始時至現時序T爲止之期間的測量波長及蝕刻時間（蝕刻 開始時至現時序T爲止之時間），使用上述式（2) 、（3 )將被累積平均之溝渠干涉週期換算成溝渠132之蝕刻量 (步驟S69)(蝕刻量算出步驟）。The etching rate was calculated for the period of the wave, and the etching amount (depth of the trench 132) was calculated from the calculated etching rate and etching time. In addition, since the thickness of the mask film 131 is gradually etched and the thickness is changed during the etching, an interference wave (hereinafter referred to as "mask film interference wave") is generated from the reflected light "and the reflected light L3". Since each interference wave is detected by the same detector, the interference wave detected by the detector overlaps a plurality of interference waves having different periods (hereinafter referred to as "overlapping interference waves") (refer to Fig. 2). In order to calculate the depth of the trench 132 (the amount of etching of the layer to be etched 130) from the superimposed interference wave shown in Fig. 23, it is necessary to separate the trench interference wave from the superimposed interference wave. In the overlapping interference wave of Fig. 23, the short-period interference wave and the long-period interference wave can be relatively clearly separated. Here, since the change speed of the depth of the trench 132 during etching is greater than the change speed of the thickness of the photomask film 131, the period of the trench interference wave is shorter than the period of the interference film of the photomask film. Therefore, the short-period interference wave in the superimposed interference wave of Fig. 23 is a trench interference wave, and it is possible to easily calculate the period of the trench interference wave from the time between the peaks in the short-period interference wave ("Δt" in the figure) -6- 200949932 In the method of reading the time between the overlapping interference waves, the short-period interference wave and the long-period interference wave in the overlapping interference wave must be able to be separated relatively clearly, so the short-period interference wave The overlapping interference wave system, which is difficult to separate from the long-period interference wave, cannot calculate the period of the trench interference wave. Further, in the short-period interference wave, the period of the trench interference wave is regarded as a certain ditch. The period of the interference wave (the etching rate of the etched layer 130) is stepped as shown in Fig. 24. That is, the method of reading the time between the electrodes of the superimposed interference® wave is low in decomposition energy. In recent years, it has been developed to calculate the period of the trench interference wave by frequency analysis without self-interstitial interference wave reading. The method is based on frequency analysis (for example, fast Fourier transform). The method of obtaining a frequency distribution (see FIG. 26(A)), and detecting the period of the trench interference wave from the frequency distribution (for example, see Patent Document 1). [Patent Document 1] Japanese Patent Laid-Open No. Hei 2-71517 Contents] (Problems to be Solved by the Invention) However, as shown in Fig. 25, in the case of overlapping thousands of waves, there is an abnormality of the laser light source or detector ("I" in the figure) or a mask film interference wave. The interference such as the periodic change ("II" in the figure) generated by the interference of the trench interference wave is added to the overlapped interference wave. In the above method of using the frequency analysis, the overlap is performed all the time from the analysis etching. The frequency distribution obtained by the interference wave detects only the period of the trench interference wave. Therefore, when the interference is caused to the overlapping interference wave, the peak is generated in the interference period which does not exist originally, and the frequency distribution becomes incorrect (see 26th). (B)), the result is that the amount of etching cannot be accurately calculated. The object of the present invention is to provide an etching amount calculation method, a memory medium, and a method for calculating the etching amount while accurately correcting the interference amount. In order to achieve the above object, the etching method according to the first aspect of the invention is to calculate the etching of the concave portion in the substrate etching using the mask film forming concave portion. The method of irradiating light to the substrate, and receiving light that receives interference light of at least the reflected light from the photomask film and the reflected light from the bottom of the concave portion is superimposed on the overlapping interference light of the other interference light a step of calculating an interference wave of the superimposed interference wave from the received superimposed interference light; a waveform extraction step of extracting a waveform of a specific period from the superimposed interference wave; and a frequency analysis step of applying a frequency analysis to the extracted waveform; An interference period detecting step of detecting a period of an interference wave from the reflected light of the photomask film and the reflected light from the bottom of the concave portion by the frequency frequency distribution obtained by the frequency analysis; and the specific period is only biased The interference wave calculation step, the waveform extraction step, and the waveform extraction step are repeated while shifting the specific time a frequency analysis step and the interference period detecting step, wherein a cumulative average step of cumulatively averaging the periods of the detected interference waves is repeated each time; and etching for calculating an etching amount of the concave portion based on a period of the cumulative average interference wave The calculation step. In the etching amount calculation method according to the first aspect of the invention, the specific period is greater than the reflected light from the photomask film, and the method for calculating the etching amount according to the second aspect of the invention. The waveform of the other interference light having a long period of the interference wave of the reflected light from the bottom of the concave portion is one cycle. The method for calculating an etching amount according to the third aspect of the invention is the method for calculating an etching amount according to the first or second aspect of the invention, wherein the cycle of the other interference light has a period ratio from the photomask film. When the period of the interference light from the reflected light and the reflected light from the bottom of the concave portion is long, the waveform of the specific period extracted from the overlapping interference wave is removed from the waveform of the other interference light before the frequency analysis step. In most pre-analytical processing steps, the frequency analysis step applies frequency analysis to a waveform in which most of the waveforms of the other interference light are removed. The method for calculating an etching amount according to the fourth aspect of the invention is as follows: in the method for calculating an etching amount according to the third aspect of the patent application, in the pre-analytical processing step, the waveform extracted from the above is removed by a plurality of times. ® mode approximates the waveform of the extracted waveform. In the method of calculating the etching amount described in the fifth or fourth aspect of the patent application, the predetermined period of time is 1/4 cycle or less of the waveform of the other interference light. The method of calculating an etching amount according to any one of claims 3 to 5, wherein the opening ratio of the concave portion in the surface of the substrate is 0. • 5% or less or the above recess is a deep trench. The method of calculating the etching amount according to any one of the first to sixth aspects of the present invention, wherein the frequency analysis system uses the maximum entropy method, in the method of calculating the etching amount according to any one of the first to sixth aspects of the invention. (maximum entropy method ). The method of calculating an etching amount according to any one of the first to seventh aspects of the invention, wherein the method of calculating the etching amount described in any one of the first to seventh aspects of the invention, further comprising the interference detected from the frequency distribution When the period of the wave corresponds to an abnormal 値, the interference period correction step of removing the period of the interference wave is performed. The method for calculating an etching amount according to the ninth aspect of the invention is the method for calculating an etching amount according to the eighth aspect of the invention, wherein in the interference period correction step, the self-determination corresponding to the abnormality The period of the interference wave obtained by the specific period before or after the specific period of the period of the dry wave period is determined as the interference of the specific period of the period of the interference wave corresponding to the abnormality 取得. The cycle of waves. The method for calculating an etching amount according to any one of the first to ninth aspects of the present invention, in the method of calculating an etching amount according to any one of claims 1 to 9, wherein the reflected light from the photomask film is predicted in advance The period of the interference wave from the reflected light at the bottom of the concave portion, in the interference period detecting step, the reflection from the photomask film is detected from the vicinity of the predicted period in the frequency distribution obtained by the frequency analysis The period of the interference wave of the light and the reflected light from the bottom of the concave portion. The method for calculating the amount of etching according to any one of the first to tenth aspects of the invention is the method of calculating the etching amount according to any one of claims 1 to 10, wherein the other interference light is from the light. The reflected light on the surface of the cover film and the interference light from the reflected light from the interface between the photomask film and the surface of the substrate. In order to achieve the above object, the memory medium described in claim 12 belongs to a computer-readable memory medium storing a program for causing a computer to perform an etching amount calculation method, and the etching amount calculation method is to use a photomask film. The etching amount of the concave portion is calculated in the substrate etching in which the concave portion is formed, and the etch amount calculation method includes a step of irradiating the substrate with light, and receiving at least the reflected light from the photomask film and the bottom portion from the concave portion. a step of receiving the interference light of the reflected light superimposed on the interference light of the other interference light; an interference wave calculation step of calculating the superimposed interference wave from the received superimposed interference light; and extracting a waveform of the waveform of the specific period from the superimposed interference light extraction a frequency analysis step of applying frequency analysis to the extracted waveform; detecting interference of reflected light from the photomask film and reflected light from a bottom portion of the concave portion from a frequency distribution obtained by the frequency analysis Interference period detecting step of the wave period; while making the above specific period The interference wave calculation step is repeated while shifting the specific time. The waveform extraction step, the frequency analysis step, and the interference period. The detection step is performed by repeating the cumulative average of the above-described cumulative average steps of the detected interference wave period. And an etching amount calculation step of calculating the etching amount of the concave portion based on the interference wave which is cumulatively averaged as described above. In order to achieve the above object, the etching amount calculation device described in the thirteenth aspect of the patent application method calculates the etching amount of the concave portion in the substrate etching using the mask film forming concave portion, and is characterized in that the substrate is irradiated with -11 - 200949932 a light irradiation unit; a light receiving unit that receives interference light from at least the reflected light from the photomask film and reflected light from a bottom portion of the concave portion, and superimposes the interference light of the other interference light; and the received overlapping interference light An interference wave calculation unit that calculates a superimposed interference wave; a waveform extraction unit that extracts a waveform of a specific period from the superimposed interference light; a frequency analysis unit that performs frequency analysis on the extracted waveform; and a frequency obtained by the frequency analysis An interference period detecting unit that detects a period of an interference wave from the reflected light of the photomask film and the reflected light from the bottom of the concave portion, and repeats the calculation of the superimposed interference wave while shifting the specific period by only a specific time Waveform extraction during the above specific period, frequency analysis, and periodic detection of the interference wave The cumulative average portion that cumulatively averages the period of the detected interference wave is accumulated for each repetition, and the uranium engraving amount calculation unit that calculates the etching amount of the concave portion based on the period of the cumulative average interference wave. [Effect of the invention] The etching amount calculation method described in the first aspect of the patent application, the memory medium described in claim 12, and the etching amount calculation device described in claim 13 In the specific period, the calculation of the overlapping interference wave is repeated, the waveform extraction in a specific period, the frequency analysis, and the period of the interference wave from the reflected light from the photomask film and the reflected light from the bottom of the concave portion are repeated for a specific period of time. The period of the interference wave detected by the average is accumulated, and the etching amount of the concave portion is calculated from the period of the interference wave of the cumulative average. Therefore, even if, for example, the waveform of a certain period of time to be extracted is disturbed, the period of the interference wave detected based on the waveform -12-200949932 of the certain period is detected by the waveform according to other specific periods. Since the cumulative average of the interference wave period is reduced, the period of the interference wave detected based on the waveform of the specific period in which the interference is given can be reduced, and the influence of the period of the cumulative average interference wave can be corrected, so that even if the interference is given, the correctness can be made. The stability calculation is performed to calculate the amount of etching. When the method of calculating the lithography amount described in the second paragraph of the patent application is applied, the specific time is longer than the period of the interference wave of the reflected light from the photomask film and the reflected light from the bottom of the concave portion. Since the waveform of the interference light has a large period of one cycle, the reliability of the frequency analysis of the waveform in a specific period can be improved, so that the calculation of the etching amount can be performed more accurately. According to the etching amount calculation method described in the third paragraph of the patent application, the period of the waveform of the other dry light is longer than the period of the interference light from the light of the photomask film and the reflected light from the bottom of the concave portion. At the time of frequency analysis, the waveform of the specific period extracted from the superimposed interference wave removes most of the waveform of the other interference light, so that even if the amount of reflected light from the bottom of the concave portion is small, other interference light When the waveform occupies most of the overlapping interference waves, the waveform after the removal can increase the ratio of the interference light from the reticle film and the reflected light from the bottom of the concave portion. In the frequency analysis, the period of the interference wave from the reflected light of the photomask film and the reflected light from the bottom of the concave portion can be accurately calculated. When the etching amount calculation method described in the fourth paragraph of the patent application is applied, the waveform extracted from the waveform extracted before the frequency analysis is removed, and the waveform of the extracted waveform is approximated by the quadratic polynomial. When the amount of reflected light from the bottom of the concave portion is small, the waveform of the other interference light occupies most of the overlapping interference wave -13-200949932, since the waveform of the overlapping interference wave and the other interference light are almost equal, so that the waveform is twice The waveform of the overlapping interference wave extracted by the polynomial approximation is also almost equal to the waveform of the other interference light. Therefore, it is possible to surely remove most of the waveforms of other interference light from the extracted waveform. The specific period is 1/4 cycle or less of the waveform of the other interference light by the etching amount calculation method _ described in the fifth paragraph of the patent application. Since the waveforms of other thousands of light rays occupying most of the overlapping interference waves are close to the sine wave, if the portion of the waveform of the other interference light is extracted less than 1/4 cycle, @the extracted waveform can be correctly corrected by the quadratic polynomial approximate. Accordingly, most of the waveforms of other interference light can be removed from the extracted waveform. When the etching amount calculation method described in the seventh paragraph of the patent application is applied, the maximum entropy method is used for frequency analysis. Therefore, even if the number of waveforms in a specific period is small, the reliability of frequency analysis can be improved, so that it can be executed more correctly. Calculation of the amount of etching. When the uranium engraving calculation method described in the eighth paragraph of the patent application is applied, since the period of the interference wave detected from the frequency distribution corresponds to the abnormal 値, the period of the interference wave is removed, so that it can be removed. The period of the interference wave detected based on the waveform of the specific period of the disturbance is affected by the period of the cumulative average interference wave, so that the amount of etching can be accurately calculated even if it is disturbed. When the uranium engraving calculation method described in the ninth application of the patent application is applied, the specific period before or after the specific period of the period of the interference wave corresponding to the abnormal 値 is obtained. In the period of the period of the wave-waves-14-200949932, it is regarded as the period of the interference wave in a specific period corresponding to the period of the interference wave of the abnormal ,, so that the interference detected by the waveform of the specific period subjected to the disturbance can be surely removed. The influence of light. When the etching amount calculation method described in the tenth paragraph of the patent application is applied, the period of the interference wave of the reflected light from the photomask film and the reflected light from the bottom of the concave portion is predicted in advance, and the frequency of the waveform by the specific period is used. In the frequency distribution obtained by the analysis, since the period of the interference wave is detected from the vicinity of the predicted period, the detection of the period of the interference wave can be quickly performed, and the detection of the abnormal 値 can be suppressed. [Embodiment] Hereinafter, embodiments of the present invention will be described with reference to the drawings. First, a substrate processing apparatus to which the etching amount calculation method and the etching amount calculation method according to the first embodiment of the present invention are applied will be described. This substrate processing apparatus is configured to etch a semiconductor wafer (hereinafter simply referred to as "wafer") W as a substrate by using plasma. Further, as shown in Fig. 22, the wafer W has an etched layer 130 and a mask film 131 formed on the etched layer 130 in a specific pattern. Fig. 1 is a cross-sectional view showing a schematic configuration of a substrate processing apparatus to which an etching amount calculation method according to the present embodiment is applied. In the first embodiment, the substrate processing apparatus 10 is provided with a processing chamber 11 made of a conductive material such as aluminum, and a bottom surface disposed in the processing chamber 11 as a mounting table on which the wafer W is placed. The lower electrode 12 and the shower head 13 disposed above the lower electrode 12 with a predetermined interval therebetween. -15- 200949932 A vacuum exhaust device is connected to the lower portion of the processing chamber 11 (the exhaust portion 14 is not connected to the figure, and the high-frequency power source 16 is connected to the lower electrode 12 via the integrator 15 in the buffer chamber inside the shower head 13 17 is connected to the gas introduction pipe 18, and the processing gas supply pipe 18 is connected to the processing supply device 19. The shower head 13 is provided at the lower portion, and has a majority of the processing space S for the buffer chamber shower head 13 and the space between the lower electrodes 12. The gas nozzle 20. The shower head 13 supplies the processing gas introduced into the buffer chamber 17 from the processing gas into the processing space S through the plurality of gas pockets 20. The substrate processing apparatus 1 reduces the processing chamber by the exhaust portion 14. After being pressed to a specific degree of vacuum, in a state where high frequency electricity is applied from the lower electrode 12 to the processing space S, the processing gas is supplied from the shower head 13 to the processing space S, and plasma is generated from the processing gas in the processing space S. The plasma etches the etched layer 130 when the wafer W collides with the etched layer 130 that is not exposed to the mask film 131, and forms a trench 132 (recess) in the etched 130. The shower head in the processing chamber 11 13 is equipped with observation from above The monitoring device 21 of the lower electrode 12. The monitoring device 21 is composed of a round member and penetrates the shower head 13. On the monitoring device 21, a window member 22 made of a transparent body such as quartz glass is placed. An optical fiber 24 that is opposed to the upper end of the monitoring device 21 and the condensed light 23 is disposed above the chamber 11. The optical fiber 24 is connected to the amount calculating device 25 for calculating the etching amount of the etched layer 130. The etching amount calculating device 25 is provided with each connection. In the formula), the processing gas, the spray communication f 1 8 is applied to the inner pressure, and the inner surface of the cover layer is placed, and the laser light source 26 (illumination unit) is etched by the lens to cut the optical fiber-16-200949932 24 And a detector 27 (light receiving unit) and a computing unit 28 (an interference wave calculating unit, a waveform extracting unit, a frequency analyzing unit, an interference frequency detecting unit, a cumulative averaging unit, and an etching amount calculating unit) connected to the detector 27 on the substrate The controller 29 of the processing device 10 operates under the control of the controller 29. As the laser light source 26, for example, a semiconductor laser is used. Further, as the detector 27, for example, a photomultiplier or a light-emitting diode is used using the laser light source 27. Further, the controller 29 is connected not only to the calculation unit 28' but also to each component of the substrate processing apparatus 1, for example, the high-frequency power source 16 to control the operation of each component. The etching amount calculation device 25 irradiates the laser light from the laser light source 26 toward the wafer W on the lower electrode 12 via the optical fiber 24, the condensing lens 23, and the monitoring device 21, and receives the overlap by the detector 27 via the optical fiber 24 or the like. The reflected light from the wafer W is the overlapping interference light in which the trench interference light (reflected light from the photomask film and the reflected light from the bottom of the concave portion) or the mask film interference light (other interference light) overlaps. . The superimposed interference light received by the detector 27 is converted into an electrical signal and transmitted to the arithmetic unit 28. The calculation unit 28 calculates the superimposed interference wave from the superimposed interference light based on the received electrical signal. Further, the calculation unit 28 calculates the etching amount of the trench 1 3 2 by performing the etching amount calculation method of Fig. 6 which will be described later based on the calculated superimposed interference wave. Further, the etching rate is not constant during etching, and changes occur due to various factors (pressure variations in the processing space S or interference of high-frequency voltage). In particular, when etching is performed to form a trench having a large aspect ratio (for example, a deep trench), when the etching amount (etching depth) of the trench 136 is increased, the plasma is suppressed at the entrance of the trench 132. Entering into the trench 132, the etching rate is lowered (see Fig. 2). Further, as shown in Fig. 2, since the uranium engraving rate is slightly changed, the etching rate must be gradually calculated in order to accurately calculate the etching amount of the trench 133. In order to gradually calculate the etching rate during etching, it is generally necessary to differentiate the waveform of the reflected light from the wafer W at each timing. However, as described above, the reflected light from the wafer W is overlapped trench interference light or light. Since the cover film © overlaps the interference light, the etching rate of the trench 132 cannot be accurately obtained even if the waveform of the reflected light is differentiated at each timing. Here, in the present embodiment, the frequency analysis is performed, and the period of the trench interference wave is calculated from the superimposed interference wave (hereinafter referred to as "ditch interference period"." Further, in the frequency analysis, the waveform of the analysis target is used for one cycle. The above information is long and, in particular, contributes to the improvement of the reliability of the frequency analysis. Therefore, in order to obtain the etching rate at a certain timing, in the present embodiment, the waveform of the specific period is extracted from the superimposed interference wave, and the extracted waveform is extracted. Further, as described above, the uranium engraving rate is related to the frequency of the trench interference. Therefore, in the present embodiment, the trench interference period is obtained from the extracted waveform, and the etching rate is calculated from the trench interference wave. Fig. 3 is a view for explaining a waveform of a specific period from the superimposed interference wave extraction in the etching amount calculation method according to the present embodiment. In Fig. 3, the interference wave 30 is overlapped by about 30 seconds. Since the periodic vibration is set, the specific period is set to 30 seconds. Here, in order to obtain the trench interference period of the timing A, the weight from the timing A to the first 30 seconds is extracted from the period 31 -18 to 200949932. The waveform of the interference wave 30 (the waveform of the portion surrounded by the square in the figure). In the present embodiment, the specific period is referred to as a "window." The window 31 has a start point 32 and an end point 33, and the start point 32 is equivalent to 30 seconds from the timing A, the end point 33 corresponds to the timing A. Then, in the present embodiment, the waveform of the extracted window 31 is subjected to frequency analysis. Here, the waveform extracted is at most one cycle. Therefore, the maximum entropy method is used as the analytical method. The maximum entropy method is a method of calculating the frequency distribution of the observed phenomenon with high decomposition energy from the measurement result between the very short measurement ("Current Data Processing for Scientific Calculation" (CQ Press) , the first edition issued on April 30, 2011, because there is no need to analyze the number of waveforms of the object, so it is more suitable for the etching involved in this embodiment than the fast Fourier transform method which requires waveforms of several cycles. Fig. 4 is a diagram showing the frequency distribution obtained by analyzing the waveform from the window in Fig. 3 by using the frequency of the maximum entropy method. The overlapping interference wave 30 mainly contains the trench interference wave and The mask film is dry and two waves are involved, so the frequency distribution obtained by the waveform of the window 31, as shown in Fig. 4, mainly has two frequencies representing the peaks (the period of the interference wave) (in Figure 4). The middle is about 0·012 Ηζ and about 〇.〇37 Hz. Here, as described above, since the period of the trench interference wave is shorter than the period of the interference film of the photomask film (high frequency), the frequency of about 0.037 Hz is equivalent to The frequency of the trench interference. Here, the frequency of the trench interference period is about 0.03 7 Hz. Thus, in the present embodiment, the frequency distribution obtained by the frequency analysis predicts that there are two peaks in the frequency analysis. Before the trench interference period is predicted in advance, the trench period is preferably detected in the vicinity of the trench cycle -19-200949932 predicted in the obtained frequency distribution. Since the window 31 includes the waveform of the superimposed interference wave 30 30 seconds from the timing A, the frequency distribution shown in Fig. 4 becomes the frequency distribution in the superimposed interference wave 30 30 seconds before the timing A. Therefore, the trench interference period detected from the frequency distribution shown in Fig. 4 is the average period of the trench interference light in the superimposed interference wave 30 30 seconds before the timing A, but in the present embodiment, It is convenient to regard the trench interference period detected from the frequency division shown in Fig. 4 as the trench interference period in the timing A. Further, in the present embodiment, as shown later, a plurality of windows are set for the superimposed interference wave 30, and the ditch interference period obtained by averaging the frequency distribution of the waveforms of the respective windows is accumulated, and the overall average 値 of the groove interference period is calculated, so that the cancellation is performed. The average period of the trench interference light in the window 31 is regarded as a disadvantage of the trench interference period in the timing A. Further, in the uranium engraving calculation method according to the present embodiment, the first predetermined period is calculated (the starting point 3 2 corresponds to the start of etching and the end point 3 3 corresponds to 30 seconds after the start of etching) At the time of the window 3 1 ), the average 値 of the etching rate in all the periods of the timing of the etching amount is calculated, and the etching amount is calculated from the average 値 of the uranium engraving rate. Fig. 5 is a view for explaining a method of calculating the average 値 of the etching rate in the etching amount calculation method according to the present embodiment, and shows a case where 80 seconds have elapsed after the start of etching. In Fig. 5, with respect to the superimposed interference wave 50, 11 windows ^^=1 to 11 with offset only At (specific time) are set, and 11 is a natural number), and a frequency distribution is obtained in each window Wk, from each frequency. The distribution detects n ditches -20- 200949932 渉 period € “1{:=1~11, 11 is a natural number.) Then, the average η ditch interference period fk is accumulated according to the following formula (1), 1] fave= 2]fk/n (1) k»l The trench interference period fave is calculated as the average 値 of the trench interference wave 80 seconds after the start of etching. And, when the wavelength is measured (from the laser) When the wavelength of the laser light of the light source 26 is λ, the average 値 of the etching rate is 80 seconds after the start of etching from the following formula (2). The average etch rate = faveX λ /2 (2) ❹ Next, the amount of uranium engraved from the following formula (3) to 80 seconds after the start of etching is calculated. Etching amount = average etch rate of etch rate X (3) In the method of calculating the etching amount according to the present embodiment When the overlapping interference wave 50 in a window Wt (t is a natural number of 1 to η) is disturbed, the channel interference frequency f obtained from the window wtm t becomes abnormal 値, but the cumulative interference average of the trench interference period fu -21 - 200949932 obtained from the trench interference frequency ft and other windows WU (U is a natural number other than 1 to η, other than t) The influence of the interference amount fave of the cumulative average trench interference period ft is small. Next, the etching amount calculation method according to the present embodiment will be described. Fig. 6 is a view showing the etching amount calculation method according to the present embodiment. In Fig. 6, after the substrate processing apparatus 10 starts etching the etched layer 130 of the wafer W, the laser light source 26 directs the laser light L1 toward the crystal via the fiber ❹ 24, the condensing lens 23, and the monitoring device 21. The circle W is irradiated (step S61) (irradiation step). The detector 27 receives the superimposed interference light belonging to the reflected light from the wafer W via the optical fiber 24 or the like (step S62) (light receiving step). Next, in step S63, the arithmetic unit 28 It is determined whether or not the current timing T exceeds the etching end time set in advance, and when the etching end time is exceeded (YES in step S63), the processing is terminated, and when the end time of the uranium engraving is not exceeded (NO in step S63), The portion 28 calculates (updates) the superimposed interference wave from the start of the etching to the current time T based on the superimposed interference light received by the detector ( 27 (step s64) (interference wave calculation step). Next, the arithmetic unit 28 The waveform of the window whose end time τ is set as the end point is extracted (step S65) (waveform extraction step), and the waveform of the extracted window is subjected to frequency analysis using the maximum entropy method (step S66) (frequency analysis step). 'The time from the start point to the end point of the window is set to be longer than one period of the interference wave of the photomask film. -22-200949932, the calculation unit 28 detects the frequency of the peak 附近 indicating the vicinity of the previously observed trench interference period as the channel interference period in the current time series T in the frequency distribution obtained by the frequency analysis. (Step S67) (interference period detecting step). Next, the calculation unit 28 accumulates the channel interference period in the current timing T detected by the above-described equation (1), and the timings detected from the time when the first specific time elapses until the current timing T is reached. In the middle of the interference period (step S68) (cumulative averaging step), the above-described formula is used based on the measurement wavelength and the etching time (the time from the start of etching to the current time T) from the start of etching to the current time series T. (2) and (3) Converting the cumulative average trench interference period into the etching amount of the trench 132 (step S69) (etching amount calculation step).
之後，運算部28在現時序T加上At而更新現時序T ，即是使窗之終點僅偏移△ t (步驟S70 )，返回步驟S63 〇 胃若藉由本實施型態所渉及之蝕刻量算出方法時，一邊 使窗之終點僅偏移At —邊重複算出重疊干涉波、自重疊 干涉波抽出窗之波形、頻率解析及檢測出現時序T中之溝 渠干涉週期，於每次重覆時累積平均所檢測出之現時序T 中之溝渠干涉週期，和在經過最初之特定期間至現時序T 爲止之期間被檢測出之各時序中之溝渠干涉週期，將該累 積平均之溝渠干涉週期換算成溝渠132之蝕刻量。因此， 即使例如被抽出之某特定期間之窗的波形受到干擾之時， 自該某窗所求出之異常値的溝渠干涉週期，因與自其他窗 -23- 200949932 所求出之溝渠干涉週期累積平均，故可以縮小對累積平均 屬於異常値之溝渠干涉週期的溝渠干涉週期所造成之影響 ，因此即使受到干擾亦可以正確安定執行溝渠132之蝕刻 量算出。 再者，在本實施型態所涉及之蝕刻量算出方法中，因 從上述窗之起點至終點之時間大於在步驟S64所算出之重 疊干涉波之1週期，故可以提高窗中之重疊干涉波之頻率 解析之信賴性。 並且，在本實施型態所涉及之蝕刻量算出方法中，因 頻率解析使用最大熵法，故即使窗之波形之數量少亦可以 提高頻率解析之信賴性。 再者，在本實施型態所涉及之蝕刻量算出方法中，事 先預測溝渠干涉週期，在藉由頻率解析所取得之頻率分佈 中，因自預測之溝渠干涉週期附近檢測出現時序T中之溝 渠干涉週期，故可以迅速執行溝渠干涉週期之檢測’並且 可以抑制溝渠干擾週期檢測出異常値。 接著，針對本發明之第2實施型態所涉及之鈾刻量算 出方法予以說明。 本實施型態因基本上其構成、作用與上述第1實施型 態相同，故針對重複構成、作用省略說明’以下針對不同 之構成、作用進行說明。 第7圖爲表示本實施型態所涉及之蝕刻量算出方法之 流程圖。 在第7圖中，首先實行步驟S61至S67，接著在步驟 200949932 S71中，判別運算部28在步驟S67作爲現時序T中之溝 渠干涉週期被檢測出之溝渠干涉週期是否相當於異常値（ 例如，在步驟S66所取得之頻率分佈中之最大頻率或最小 頻率）。 步驟S71之判別之結果，作爲現時序Τ中之溝渠干涉 週期被檢測出之溝渠干涉週期不相當於異常値之時（在步 驟S71爲NO )，則直接前進於步驟S68，於被檢測出之 ❹ 干涉週期相當於異常値之時（在步驟S爲YES )，則除去 作爲現時序T中之溝渠干涉週期被檢測出之溝渠干涉週期 ，並且藉由將自對應於比現時序T早一個時序之窗所檢測 出之溝渠干渉週期，當作現時序T中之溝渠干涉週期予以 設定，依此修正溝渠干涉週期（步驟S72)(干涉週期修 正步驟）。 接著，運算部28係實行步驟S68至S70。 若藉由本實施型態所涉及之蝕刻量算出方法時，於作 ® 爲現時序T中之溝渠干涉週期被檢測出之溝渠干涉週期相 當於異常値之時，除去該被檢測出之溝渠干涉週期，並且 . 因將自對應於較現時序T早一個時序之窗所檢測出之溝渠 干涉週期，當作現時序T中之溝渠干涉週期而加以設定， 故可以除去相當於異常値之溝渠干涉週期對被累積平均之 溝渠干涉週期所造成之影響，因此即使受到干擾亦可以更 安定正確算出溝渠132之鈾刻量。 在上述本實施型態所涉及之蝕刻量算出方法中，雖然 於被檢測出之溝渠干涉週期相當於異常値之時，將自對應 -25- 200949932 於較現時序T早一個時序之窗所檢測出之溝渠干涉週期， 當作現時序Τ中之溝渠干涉週期而加以設定，但是即使將 自對應於較現時序Τ晚一個時序之窗所檢測出之溝渠干涉 週期，當作現時序Τ中之溝渠干涉週期而加以設定亦可。 接著，針對本發明之第3實施型態所涉及之蝕刻量算 出方法予以說明。 本實施型態因基本上其構成、作用與上述第1實施型 態相同，因僅有於對被抽出之窗之波形施予頻率解析之前 ❹ ，對該窗之波形施予前處理之點不同，故針對重複構成、 作用省略說明，以下針對不同之構成、作用進行說明。 在晶圓W之表面中溝渠132之開口部所佔有之比率 的百分比（以下稱爲「開口率」）爲小之時，例如低於 0.5%之時，因來自溝渠132之底部之反射光L4之絕對光 量變少，故在檢測器2 7所接受之重疊干涉光中溝渠干涉 光所占有之部分的比率變小。 第9圖爲表示開口率變化之時重疊干涉波之一部分變 〇 化的圖式。 如第9圖所示般，於開口率爲5%之時，重疊干涉波 明顯重疊兩種千涉波（光罩膜干涉波、溝渠干涉波），但 是於開口率爲0.5%之時，反射光L4之絕對光量變少，在 重疊干涉波幾乎不出現溝渠干涉波之波形。如此之現象， 因爲即使在晶圓W之表面，孔之開口部所占有之比率爲 小之時或溝渠（或是孔）之縱橫比爲大之時（例如，溝渠 132爲深溝渠之時），來自溝渠或孔之底部之反射光之絕 -26- 200949932 對光量也變少，故得以發生。 於重疊干涉波幾乎不出現溝渠干涉波之波形之時，（ 開口率爲0.5%之時），自該重疊干涉波抽出窗31之波形 ，當直接對所抽出之波形施予頻率解析之時，則在所取得 之頻率分佈中，溝渠干涉波之週期（溝渠干涉週期）之峰 値變小。 第10圖爲表示開口率變化之時窗之波形的頻率分佈 © 變化的圖式。 如第10圖所示般，於開口率爲5%之時，頻率分佈 明顯出現兩個峰値（溝渠干涉波之週期（大約〇.8Hz)、 光罩膜波之週期（大約〇·1Ηζ))，但是於開口率爲0.5 %之時，頻率分佈明顯僅出現一個峰値（光罩膜干涉波之 週期），幾乎不出現溝渠干涉波之週期。其結果，無法正 確檢測出溝渠干涉週期，無法正確算出蝕刻率。 在此，在本實施型態中，將從重疊干涉波藉由窗31 ® 被抽出之波形予以頻率解析之前，自重疊干涉波除去光罩 膜干涉波所佔有之大部分。 .第11圖爲表示開口率爲0.5%之時的重疊干涉波之圖 式。 如第11圖所示般，於開口率爲0.5%之時，如第1 1 圖所示般，因於重疊干涉波幾乎不出現週期短之溝渠干涉 波，故重疊干涉波幾乎由光罩膜干涉波所佔據。因此，近 似重疊干涉波之波形幾乎與光罩膜干涉波之波形相等。在 此，在本實施型態中，自重疊干涉波除去近似該重疊干涉 -27- 200949932 波之波形。依此，可以自重疊干涉波除去光罩膜干涉波所 佔有之大部分。 再者’如第11圖所示般，幾乎被光罩膜干涉所佔據 之重疊干涉波接近正弦波，正弦波之1/4週期以下之部分 可以藉由二次多項式正確近似。在此，在本實施型態中， 將從重疊干涉波抽出窗31之波形之時，抽出光罩膜干涉 波之1/4週期以下的部分。 即是，在本實施型態中，如第12圖所示般，各使相 當於光罩膜干涉波之1/4週期以下之η個窗Wk(k=l〜η ，η爲自然數）偏移At而予以設定，抽出各窗Wk之波形 ，自抽出之波形除去藉由二次多項式近似該抽出之波形的 波形，取得除去光罩干涉膜所占有之大部分的波形（第 13圖），並將除去後之波形予以頻率解析。依此，藉由 第14圖所示般，可以取得溝渠干涉週期（大約0.8Hz) 之峰値明顯出現之頻率分佈。並且，在第14圖之頻率分 佈中不出現光罩膜干涉波之週期（大約0.1Hz)之峰値， 係因爲從窗Wk之波形除去光罩膜干涉波所佔有之大部分 〇 第15圖爲表示本實施型態所涉及之蝕刻量算出方法 之流程圖。並且，本實施型態所涉及之蝕刻量算出方法係 於開口率小之時，例如於該開口率低於0 · 5 %之時實行。 在第15圖中，首先實行步驟S61至S64，接著運算 部28在重疊干涉波抽出將現時序T設爲終點之1/4週期 以下之部分當作窗之波形（步驟S65)(波形抽出步驟） -28- 200949932 接著，運算部28係算出藉由二次多項式近似所抽出 之窗之波形的波形（以下單稱爲「近似波形」）（步驟 S151)，自被抽出之窗之波形除去該被算出之近似波形（ 步驟S152)(解析前處理步驟），取得除去光罩膜干渉 波所佔有之大部分的波形，並將近似波形除去後之波形予 以頻率解析（步驟S153)。 © 接著，運算部28係實行步驟S67至S70。 若藉由本實施型態所涉及之蝕刻量算出方法時，因於 頻率解析之前自被抽出之窗之波形除去光罩膜干涉波所占 有之大部分，故即使開口率小之時，例如開口率低於〇. 5 %之時，在近似波形除去後之波形亦可以增大溝渠干涉波 所佔有之部份之比率，依此可以藉由頻率解析取得明顯出 現溝渠干涉週期之峰値的頻率分佈。其結果，可以正確算 出溝渠干涉波之週期。並且，上述之本實施型態所涉及之 W 蝕刻量算出方法可以僅使用於光罩膜干涉光之波形之週期 比溝渠干涉光之週期長之時。 本實施型態所涉及之蝕刻量算出方法，係自於頻率解 析之前被抽出之窗的波形，除去藉由二次多項式近似該被 抽出之窗之波形的波形（近似波形）。於開口率小之時， 重疊干涉波之波形因與光罩膜干涉波幾乎相等，故近似波 形也與光罩膜幾乎相等。因此，可以自所抽出之窗的波形 確實除去光罩膜干涉波所佔有之大部分。 再者，在本實施型態所涉及之蝕刻量算出方法，係將 -29- 200949932 光罩膜干涉波之1/4週期以下之部分當作窗之波形予以抽 出。因佔據重叠干涉波之大部分的光罩膜干涉波接近於正 弦波，故若抽出光罩膜干涉波之1/4週期以下之部分，該 被抽出之窗之波形藉由二次多項式可以正確近似。依此， 可以正確自所抽出之窗的波形除去光罩膜干涉波所佔有之 大部分。 並且，於本實施型態中之開口率小之時，不僅係在晶 圓W之表面溝渠132之開口部所佔之比率小之時，也相 @ 當於在晶圓W表面孔之開口部所佔之比率小之時，或溝 渠（或是孔）之縱橫比大之時。 上述各實施型態中雖然頻率解析使用最大熵法，但是 即使於各窗內之干涉波形之數量爲多之時使用高速傅立葉 變換法亦可。高速傅立葉變換法因較最大熵法所需要之計 算次數少，故可以更快算出溝渠132之蝕刻量。 再者，使用上述各實施型態所涉及之蝕刻量算出方法 算出某溝渠之蝕刻量（蝕刻深度）之時，則有光罩膜131 〇 爲使雷射透過之膜時，如第8圖所示般，於所算出之蝕刻 量（圖中「監視深度」）和實際測量之蝕刻量（圖中「蝕 刻深度」）之間產生誤差之情形。該應爲重疊干涉光主要 不係反射光L2及反射光L4之干涉光，包含反射光L3及 反射光L4之干涉光，反射光L3之光路長變化不僅對光罩 膜131之厚度變化，也對光罩膜131之折射率產生影響之 故。 於在所算出之蝕刻量和實際測量之蝕刻量之間產生誤 -30- 200949932 差之時，於蝕刻量算出之前，使用試驗用之晶圓W實際 測量溝渠1 3 2之蝕刻量（蝕刻率），並且藉由上述各實施 型態所涉及之蝕刻量算出方法算出溝渠132之蝕刻量’並 求出實際測量之蝕刻量和所算出之鈾刻量之迴歸式等爲佳 。然後在之後的蝕刻，若藉由上述各實施型態所涉及之蝕 刻量算出方法，算出溝渠132之蝕刻量之後，以回歸式補 正該被算出之蝕刻量即可。 Ο 在上述各實施型態中，雖然算出溝渠132之蝕刻量， 但是即使實行第6圖、第7圖或第15圖之蝕刻量算出方 法，算出孔之蝕刻量亦可。 再者，本發明之目的也藉由將記錄有軟體之程式碼的 記憶媒體供給至電腦（例如控制器29 )，該軟體係用以 實現上述各實施型態之功能，並且電腦之CPU讀出並實 行儲存於記憶媒體之程式碼而達成。 此時，自記憶媒體被讀出之程式碼本身實現上述各實 ❹ W 施型態之機能，構成程式碼及記憶有其程式碼之記憶媒體 構成本發明。 再者，作爲用以供給程式碼之記憶媒體，若爲例如 RAM、NV-RAM、軟碟（註冊商標）、硬碟、光磁碟、 CD-ROM ' CD-R ' CD-RW ' DVD ( DVD-ROM > DVD-RAM 、DVD-RW、DVD + RW )等之光碟、磁帶、非揮發性之記 億卡、其他之ROM等之可以記憶上述程式碼者即可。或 是上述程式碼即使藉由自連接於網際網路、商用網路或是 區域網路等之無圖式之其他電腦或資料庫等下載，而被供 -31 - 200949932 給至電腦亦可。 再者，藉由實行電腦讀出之程式碼，不僅實現上 實施型態之功能，也包含根據其程式碼之指示，CPU 轉之os (操作系統）等執行實際處理之一部分或全 藉由其處理，實現上述各實施型態之機能的情形。 並且，也包含自記憶媒體被讀出之程式碼，被寫 插入至電腦之機能擴充埠或連接於電腦之機能擴充單 具備之記憶體後，根據其程式碼之指示，其機能擴充 機能擴充單元所具備之CPU等執行實際處理之一部 全部，並且藉由其處理實現上述各實施型態之機能的 0 上述程式碼之型態即使由目標碼、藉由編譯器所 之程式碼、被供給至0S之腳本資料（script data) 型態構成亦可。 [實施例] 接著，針對本發明之實施例予以說明。 實施例1 首先，準備在由矽所構成之被蝕刻層130上形成 化膜所構成之使雷射光L1透過之光罩膜131的晶圓 並且在基板處理裝置10藉由飽刻對被触刻層130形 溝渠1 3 2。此時之蝕刻條件如同下述般。 實際之蝕刻率 1 200nm/分 述各 上運 部， 入至 元所 埠或 份或 情形 實行 等之 由氧 W - 成深 -32- 200949932 選擇比 10對1(被蝕刻層130對光罩膜131) 開口比 0.05 測量波長（雷射光L1之波長） 3 0 0nm 取樣率 10Hz 在深溝渠132之蝕刻中，將從窗之起點至終點爲止之 時間設定爲30秒，實行第6圖之蝕刻量算出方法，求取 ® 各時序中被累積平均之溝渠干涉週期，自該被累積平均之 溝渠干涉週期求出各時序中之深溝渠132之蝕刻率。然後 ，將所求出之蝕刻率表示於曲線圖（參照第16圖）。 比較例1 再者，在上述深溝渠132之蝕刻中，藉由檢測器27 觀測來自晶圓W之重叠干涉波，讀取該重叠干涉波中之 短週期之干涉波，自該短週期之干涉波中之各極値間之時 ® 間求出溝渠千涉週期，並自該溝渠干涉週期求出各極値間 中之深溝渠132之蝕刻率。然後，將所求出之鈾刻率表示 於曲線圖（參照第16圖）。 由第1 6圖之曲線圖可知實施例1之蝕刻率較比較例 1之蝕刻率變動小爲安定。 並且，將實施例1之蝕刻量中之誤差和比較例1之飩 刻量中之誤差時間序列性表示於曲線圖（參照第1 7圖） 〇 由第17圖之曲線圖可知實施例1之蝕刻量較比較例 -33- 200949932 1之蝕刻量誤差小。依此，可知第6圖之蝕刻量算出方法 可以正確執行蝕刻量之算出。 實施例2 接著，準備在由矽所構成之被蝕刻層130上形成由氧 化膜所構成光罩膜131的晶圓W，並且利用基板處理裝置 10藉由蝕刻對被蝕刻層130形成淺溝渠132。此時之蝕刻 條件如同下述般。 @ 實際之蝕刻率 3 60nm/分 選擇比 1 〇對1 (被蝕刻層1 3 0對光罩膜1 3 1) 開口比 〇·2 測量波長（雷射光L1之波長） 30〇nm 取樣率 10Hz 在淺溝渠132之蝕刻中，將從窗之起點至終點爲止之 時間設定爲25秒，實行第6圖之蝕刻量算出方法，算出 Ο 各時序中淺溝渠132之蝕刻量（蝕刻深度）。然後，將所 算出之蝕刻量表示於曲線圖（參照第18圖）。 比較例2 再者，在上述淺溝渠132之蝕刻中，藉由檢測器27 觀測來自晶圓W之重疊干涉波，並且藉由頻率解析自蝕 刻開始時至各時序爲止之所有之重疊干涉波取得頻率分佈 ，根據該頻率分佈求出從蝕刻開始時至各時序之間的溝渠 -34- 200949932 干涉週期，自該溝渠干涉週期算出各時 之蝕刻量（蝕刻深度）。即是，不使用 重疊干涉波算出蝕刻量。然後’將所算 曲線圖（參照第18圖）。 在第18圖之曲線圖中，比較例2 亂，該係重疊干涉波受到干擾之故。另 刻量之資料並無混亂。因實施例2和比 © 擾之相同重疊干涉波所求出之蝕刻量， 之鈾刻量算出方法即使成受干擾亦可以 量之算出。 實施例3 上述實施例1、2係在其他晶圓W 刻中，實行第6圖之蝕刻量算出方法而 渠132之蝕刻率。然後，將所求出之飩 (參照第19圖）。 比較例3 再者，在與實施例3相同之蝕刻中 使用高速傅立葉變換法之外，實行與第 方法相同條件之鈾刻量算出方法而求1 1 3 2之蝕刻率。然後，將所求出之蝕刻 參照第19圖）。 由第19圖之曲線圖可知實施例3 序中之淺溝渠132 第3圖所示之窗自 出之鈾刻量表示於 之蝕刻量之資料混 外，實施例2之蝕 較例2係由承受干 故依此可知第6圖 安定正確執行蝕刻 中之溝渠132之蝕 求出各時序中之溝 刻率表示於曲線圖 ，不是最大熵法， 6圖之蝕刻量算出 ϋ各時序中之溝渠 率表示於曲線圖（ 之蝕刻率較比較例 -35- 200949932 3之蝕刻率變動小爲安定。依此，可知當使用最大熵法時 可以安定執行蝕刻量之算出。 實施例4 首先，準備開口率爲5%之晶圓W和開口率爲0.5% 之晶圓W，蝕刻各晶圓W之被蝕刻層130之時，實行第 15圖之蝕刻量算出方法而求出各蝕刻率。然後’將所求 出之蝕刻率表示於曲線圖（參照第20圖）。 比較例4 與實施例4相同，準備開口率爲5%之晶圓W和開口 率爲0.5%之晶圓W，蝕刻各晶圓W之被飩刻層130之時 ，實行第6圖之蝕刻量算出方法而求出各蝕刻率。然後’ 將所求出之鈾刻率表示於曲線圖（參照第21圖）。 當比較第20圖及第21圖之曲線圖時’可知於實行第 6圖之蝕刻量算出方法之時，開口率爲5%之蝕刻率爲安 定，但是開口率爲0.5%之鈾刻率則不安定’對此於實行 第15圖之蝕刻量算出方法之時，開口率爲5%之蝕刻率 及開口率爲0.5 %之蝕刻率中之任一者皆爲安定。依此， 可知當對抽出之窗的波形執行頻率解析之前’自該被抽出 之窗的波形除去該窗之波形的近似波形時’即使開口率小 時，亦可以正確求出蝕刻率。 【圖式簡單說明】 -36- 200949932 第1圖爲槪略性表示適用本實施型態所涉及之蝕刻量 算出方法之基板處理裝置之構成的剖面圖。 第2圖爲用以說明溝渠之鈾刻中蝕刻率下降之圖式。 第3圖爲用以說明本實施型態所涉及之蝕刻量算出方 法中自重疊干涉波抽出特定期間之波形的圖式。 第4圖爲表示藉由使用最大熵法之頻率解析自第3圖 中之窗之波形所取得之頻率分佈之圖式。 ® 第5圖爲用以說明本實施型態所涉及之蝕刻量算出方 法中蝕刻率之平均値之算出方法的圖式。 第6圖爲表示本實施型態所涉及之蝕刻量算出方法之 流程圖。 第7圖爲表示本發明之第2實施型態所涉及之鈾刻量 算出方法之流程圖。 第8圖爲表示所算出之蝕刻量和實際測量之蝕刻量之 間的誤差之圖式。 ® 第9圖爲表示開口率變化之時重疊干涉波之一部分變 化的圖式。 .第10圖爲表示開口率變化之時，窗之波形的頻率分 佈變化之圖式。 第11圖爲表示開口率爲0.5%之時的重疊干涉波之圖 式。 第12圖爲用以說明本發明之第3實施型態所涉及之 蝕刻量算出方法中抽出窗之波形的圖式。 第13圖爲表示自重疊干涉波除去光罩膜干涉波所佔 -37- 200949932 有之大部分之波形的圖式。 第14圖爲表示由重疊干涉波除去光罩膜干涉波所佔 有之大部分之波形藉由頻率解析而所取得的頻率分佈之圖 式。 第15圖爲表示本實施型態所涉及之蝕刻量算出方法 之流程圖。 第16圖爲藉由第6圖之蝕刻量算出方法所算出之蝕 刻率和自干涉波中之各極値間之時間所求出之蝕刻率之比 較圖。 第17圖爲藉由第6圖之蝕刻量算出方法所算出之蝕 刻量及實際之蝕刻量之誤差，和自干涉波中之各極値間之 時間所求出之蝕刻量及實際之蝕刻量之誤差的比較圖。 第18圖爲藉由第6圖之蝕刻量算出方法所算出之蝕 刻量，和藉由從蝕刻開始至各時序爲止之重疊干涉波之頻 率解析而求出之蝕刻量的比較圖。 第19圖爲使用最大熵法所算出之鈾刻率和使用高速 傅立葉變換法所算出之蝕刻率之比較圖。 第20圖爲表示使用第15圖之蝕刻量算出方法而取得 之開口率不同之各晶圓之蝕刻率的圖式。 第21圖爲表示使用第6圖之蝕刻量算出方法而取得 之開口率不同之各晶圓之蝕刻率的圖式。 第22圖爲用以說明蝕刻中之光干涉的圖式。 第23圖爲表示重疊干涉波之圖式。 第24圖爲表示自干涉波中各極値間之時間所求出之 -38- 200949932 蝕刻率之圖式。 第25圖爲表示受到干擾之重疊干涉波之圖式。 第26圖爲藉由重疊干涉波之頻率解析所取得之頻率 分佈，第26圖（A)爲重叠干涉波不受到干擾之情形， 第26圖（b)爲重疊干涉波受到干擾之情形。 【主要元件符號說明】 β L1 :雷射光 L2、l3、L4 :反射光 w ::晶圓 1〇 :基板處理裝置 25 :蝕刻量算出裝置 26 :雷射光源 27 :檢測器 28 :運算部 • 30、50 :重叠干涉波 31 :窗 1 3 0 :被蝕刻層 1 3 1 :光罩膜 132 :溝渠 -39-Thereafter, the arithmetic unit 28 updates the current timing T by adding At at the current timing T, that is, shifting the end point of the window by only Δt (step S70), and returning to step S63, if the stomach is etched by the present embodiment. In the calculation method, the overlapped interference wave, the waveform of the self-overlapping interference wave extraction window, the frequency analysis, and the detection of the channel interference period in the detection timing T are repeated while the end point of the window is shifted by At-, at each repetition. The trench interference period in the current timing T detected by the cumulative average, and the trench interference period in each of the timings detected during the period from the initial specific period to the current timing T, and the cumulative average trench interference period is converted. The amount of etching into the trench 132. Therefore, even if, for example, the waveform of the window of a certain period that is extracted is disturbed, the abnormal interference channel period determined from the certain window is due to the trench interference period determined from the other window -23-200949932. Since the cumulative average is used, it is possible to reduce the influence of the channel interference period on the cumulative average of the channel interference period which is abnormal, so that the etching amount of the trench 132 can be correctly settled even if it is disturbed. Further, in the etching amount calculation method according to the present embodiment, since the time from the start point to the end point of the window is larger than one cycle of the superimposed interference wave calculated in step S64, the overlapping interference wave in the window can be improved. The reliability of frequency analysis. Further, in the etching amount calculation method according to the present embodiment, since the maximum entropy method is used for the frequency analysis, the reliability of the frequency analysis can be improved even if the number of waveforms of the window is small. Further, in the etching amount calculation method according to the present embodiment, the trench interference period is predicted in advance, and in the frequency distribution obtained by the frequency analysis, the trench in the timing T is detected in the vicinity of the self-predicted trench interference period. Interference period, so the detection of the trench interference period can be quickly performed' and the abnormal channel can be detected by suppressing the channel interference period. Next, a method of calculating the uranium engraving according to the second embodiment of the present invention will be described. Since the configuration and operation of the present embodiment are basically the same as those of the above-described first embodiment, the description of the configuration and operation will be omitted for the following description. Fig. 7 is a flow chart showing a method of calculating an etching amount according to this embodiment. In Fig. 7, first, steps S61 to S67 are carried out, and then in step 200949932 to S71, it is determined whether or not the channel interference period detected by the calculation unit 28 as the channel interference period in the current timing T in step S67 corresponds to an abnormality (for example, The maximum frequency or the minimum frequency in the frequency distribution obtained in step S66). As a result of the determination in step S71, when the channel interference period detected as the channel interference period in the current time series is not equal to the abnormal time (NO in step S71), the process proceeds directly to step S68, and is detected. ❹ When the interference period is equivalent to the abnormal 値 (YES in step S), the trench interference period detected as the trench interference period in the current timing T is removed, and by timing one time from the corresponding current timing T The trench dry period detected by the window is set as the trench interference period in the current timing T, and the trench interference period is corrected accordingly (step S72) (interference period correction step). Next, the arithmetic unit 28 executes steps S68 to S70. According to the etching amount calculation method according to the present embodiment, when the trench interference period detected by the trench interference period in the current timing T is equivalent to the abnormal 値, the detected trench interference period is removed. And the channel interference period detected by the window corresponding to the timing of the timing T is set as the channel interference period in the current timing T, so that the channel interference period equivalent to the abnormal 値 can be removed. The effect of the cumulative average trench interference period is such that even if it is disturbed, the uranium engraving of the trench 132 can be calculated more accurately. In the above-described etching amount calculation method according to the present embodiment, when the detected interference period of the trench corresponds to an abnormal 値, it is detected from a window corresponding to the current timing T of the corresponding time period T-25-200949932. The interference period of the trench is set as the interference period of the trench in the current timing, but even if the trench interference period detected from the window corresponding to the timing of the later timing is regarded as the current timing It is also possible to set the trench interference period. Next, a method of calculating the etching amount according to the third embodiment of the present invention will be described. Since the configuration and operation of the present embodiment are basically the same as those of the first embodiment described above, since the waveform is applied to the waveform of the extracted window, the waveform is pre-processed before the waveform is applied. Therefore, the description of the repetitive configuration and the operation will be omitted, and the different configurations and operations will be described below. When the percentage of the ratio occupied by the opening of the trench 132 in the surface of the wafer W (hereinafter referred to as "opening ratio") is small, for example, less than 0.5%, the reflected light L4 from the bottom of the trench 132 Since the absolute amount of light is small, the ratio of the portion occupied by the trench interference light in the superimposed interference light received by the detector 27 is small. Fig. 9 is a view showing a state in which one of the overlapping interference waves is changed when the aperture ratio is changed. As shown in Fig. 9, when the aperture ratio is 5%, the overlapping interference waves clearly overlap the two kinds of thousands of waves (the mask interference wave and the trench interference wave), but when the aperture ratio is 0.5%, the reflection The absolute light amount of the light L4 is small, and the waveform of the trench interference wave hardly occurs in the overlapping interference wave. Such a phenomenon is because even when the ratio of the opening of the hole is small or the aspect ratio of the trench (or the hole) is large on the surface of the wafer W (for example, when the trench 132 is a deep trench) , the reflection of light from the bottom of the ditch or the hole -26- 200949932 The amount of light is also reduced, so it can happen. When the waveform of the trench interference wave hardly occurs in the overlapping interference wave (when the aperture ratio is 0.5%), when the waveform of the overlapping interference wave extraction window 31 is directly applied to the extracted waveform, the frequency analysis is performed. Then, in the obtained frequency distribution, the peak of the period of the trench interference wave (the channel interference period) becomes small. Fig. 10 is a diagram showing the frequency distribution of the waveform of the time window in which the aperture ratio is changed. As shown in Fig. 10, when the aperture ratio is 5%, two peaks appear in the frequency distribution (the period of the trench interference wave (about 〇8 Hz), and the period of the mask film wave (about 〇·1Ηζ). However, when the aperture ratio is 0.5%, the frequency distribution obviously shows only one peak 値 (the period of the mask film interference wave), and the period of the trench interference wave hardly occurs. As a result, the trench interference period cannot be correctly detected, and the etching rate cannot be accurately calculated. Here, in the present embodiment, most of the interference film interference is removed from the superimposed interference wave before the frequency is analyzed by the waveform of the superimposed interference wave extracted by the window 31 ® . Fig. 11 is a view showing a pattern of overlapping interference waves when the aperture ratio is 0.5%. As shown in Fig. 11, when the aperture ratio is 0.5%, as shown in Fig. 1, since the overlapping interference waves hardly appear as short-distance trench interference waves, the overlapping interference waves are almost covered by the photomask film. Occupied by interference waves. Therefore, the waveform of the near-overlapping interference wave is almost equal to the waveform of the interference film of the photomask film. Here, in the present embodiment, the waveform of the overlapping interference -27-200949932 wave is removed from the superimposed interference wave. Accordingly, most of the interference waves of the photomask film can be removed from the overlapping interference waves. Further, as shown in Fig. 11, the overlapping interference wave occupied by the interference of the photomask film is close to a sine wave, and the portion below the 1/4 cycle of the sine wave can be correctly approximated by the quadratic polynomial. Here, in the present embodiment, when the waveform of the window 31 is extracted from the overlapping interference wave, a portion of the photomask film interference wave which is equal to or less than 1/4 cycle is extracted. In other words, in the present embodiment, as shown in Fig. 12, each of the n windows Wk (k = 1 to η, η is a natural number) corresponding to 1/4 cycle or less of the interference film of the photomask film is used. The offset is set to At, and the waveform of each window Wk is extracted, and the waveform of the extracted waveform is approximated by the quadratic polynomial from the extracted waveform, and the waveform occupied by the mask interference film is removed (Fig. 13). And the waveform after the removal is subjected to frequency resolution. Accordingly, as shown in Fig. 14, the frequency distribution in which the peak of the trench interference period (about 0.8 Hz) is apparent can be obtained. Further, in the frequency distribution of Fig. 14, the peak of the period (about 0.1 Hz) of the interference film of the photomask film does not occur, because most of the interference of the mask film is removed from the waveform of the window Wk. It is a flowchart which shows the method of calculating the etching amount concerning this embodiment. Further, the etching amount calculation method according to the present embodiment is carried out when the aperture ratio is small, for example, when the aperture ratio is less than 0.5%. In the fifteenth diagram, the steps S61 to S64 are first executed, and the calculation unit 28 takes the portion of the superimposed interference wave extraction which is equal to or less than the quarter of the end of the current sequence T as the window waveform (step S65) (wave extraction step) -28-200949932 Next, the calculation unit 28 calculates a waveform of a waveform extracted by a quadratic polynomial approximation (hereinafter simply referred to as an "approximate waveform") (step S151), and removes the waveform from the extracted window. The calculated approximate waveform (step S152) (pre-analytical processing step) acquires a waveform in which most of the dry film wave is removed, and the waveform after the approximate waveform is removed is subjected to frequency analysis (step S153). © Next, the arithmetic unit 28 executes steps S67 to S70. According to the etching amount calculation method according to the present embodiment, since most of the interference of the photomask interference wave is removed from the waveform of the window to be extracted before the frequency analysis, even when the aperture ratio is small, for example, the aperture ratio Below 〇. 5 %, the waveform after the approximate waveform is removed can also increase the ratio of the portion occupied by the trench interference wave, so that the frequency distribution of the peak 値 of the trench interference period can be obtained by frequency analysis. . As a result, the period of the trench interference wave can be correctly calculated. Further, the W etching amount calculation method according to the above-described embodiment can be used only when the period of the waveform of the mask film interference light is longer than the period of the trench interference light. The etching amount calculation method according to this embodiment is a waveform of a window extracted from the frequency before the frequency analysis, and a waveform (approximate waveform) of the waveform of the extracted window is approximated by a quadratic polynomial. When the aperture ratio is small, the waveform of the superimposed interference wave is almost equal to the interference wave of the photomask film, so that the approximate waveform is almost equal to the mask film. Therefore, it is possible to surely remove most of the interference of the mask film interference wave from the waveform of the extracted window. Further, in the etching amount calculation method according to the present embodiment, a portion of the 1/4-cycle of the -29-200949932 mask film interference wave is extracted as a window waveform. Since the interference film of the photomask film occupying most of the overlapping interference waves is close to the sine wave, if the portion of the interference film of the photomask film is extracted less than 1/4 cycle, the waveform of the extracted window can be correct by the quadratic polynomial. approximate. Accordingly, most of the interference of the mask film interference wave can be removed from the waveform of the window to be extracted. Further, when the aperture ratio in the present embodiment is small, not only when the ratio of the opening portion of the surface trench 132 of the wafer W is small, but also when the opening portion of the surface of the wafer W is small. When the ratio is small, or when the aspect ratio of the ditch (or hole) is large. In each of the above embodiments, the maximum entropy method is used for frequency analysis, but the fast Fourier transform method may be used even when the number of interference waveforms in each window is large. Since the fast Fourier transform method requires less calculations than the maximum entropy method, the etching amount of the trench 132 can be calculated more quickly. Further, when the etching amount (etching depth) of a certain trench is calculated by the etching amount calculation method according to each of the above embodiments, when the mask film 131 is a film for transmitting laser light, as shown in FIG. As a general rule, an error occurs between the calculated etching amount ("monitoring depth" in the drawing) and the actually measured etching amount ("etching depth" in the drawing). This should be that the superimposed interference light is mainly the interference light of the reflected light L2 and the reflected light L4, and includes the interference light of the reflected light L3 and the reflected light L4. The change in the optical path length of the reflected light L3 not only changes the thickness of the photomask film 131 but also The refractive index of the photomask film 131 is affected. When the difference between the calculated etching amount and the actually measured etching amount is -30-200949932, the etching amount of the trench 13 2 is actually measured using the wafer W for the test before the etching amount is calculated (etching rate) It is preferable to calculate the etching amount of the trench 132 by the etching amount calculation method according to each of the above-described embodiments, and to obtain the regression amount of the actually measured etching amount and the calculated uranium amount. Then, in the subsequent etching, the etching amount of the trench 132 is calculated by the etching amount calculation method according to each of the above embodiments, and then the calculated etching amount is corrected by regression. Ο In each of the above embodiments, the etching amount of the trench 132 is calculated. However, even if the etching amount calculation method of Fig. 6, Fig. 7, or Fig. 15 is performed, the etching amount of the hole may be calculated. Furthermore, the object of the present invention is also to provide a memory medium on which a software code is recorded to a computer (for example, controller 29) for implementing the functions of the above embodiments, and reading the CPU of the computer. And implemented by storing the code stored in the memory medium. At this time, the program code itself read from the memory medium realizes the functions of the above-described various embodiments, and the memory code constituting the program code and the memory code thereof constitutes the present invention. Further, as the memory medium for supplying the code, for example, RAM, NV-RAM, floppy disk (registered trademark), hard disk, optical disk, CD-ROM 'CD-R 'CD-RW 'DVD ( CD-ROM > DVD-RAM, DVD-RW, DVD + RW, etc., such as CD-ROM, magnetic tape, non-volatile card, other ROM, etc. can memorize the above code. Or the above code is available to the computer even if it is downloaded from other computers or databases connected to the Internet, commercial network or local area network, etc. Furthermore, by implementing the program code read by the computer, not only the function of the implementation mode but also the execution of the CPU according to the instruction of the program code, the CPU to the os (operating system), etc. Processing, the realization of the functions of the above embodiments. Moreover, the program code read from the memory medium is also inserted into the memory of the computer or connected to the memory of the computer, and the function expansion unit is activated according to the instruction of the program code. The CPU or the like is provided to perform all of the actual processing, and the type of the above-mentioned code that realizes the functions of the above-described embodiments is processed by the target code and the code of the compiler. The script data type configuration to 0S is also possible. [Embodiment] Next, an embodiment of the present invention will be described. Embodiment 1 First, a wafer formed by forming a film on the etched layer 130 made of ruthenium to allow the laser light L1 to pass through is prepared, and the substrate processing apparatus 10 is touched by a saturated pair. Layer 130 shaped trenches 1 3 2 . The etching conditions at this time are as follows. The actual etching rate is 1 200 nm / the respective upper moving parts, the input into the element or the part or the case is performed by the oxygen W - deep - 32 - 200949932, the selection ratio is 10 to 1 (the etched layer 130 is applied to the mask film) 131) Opening ratio 0.05 Measurement wavelength (wavelength of laser light L1) 3 0 0nm sampling rate 10 Hz In the etching of deep trench 132, the time from the start point to the end point of the window is set to 30 seconds, and the etching amount of Fig. 6 is performed. The calculation method is performed to obtain the cumulative interference mean channel period in each of the timings, and the etching rate of the deep trench 132 in each timing is obtained from the cumulative average trench interference period. Then, the obtained etching rate is shown in a graph (refer to Fig. 16). Comparative Example 1 Further, in the etching of the deep trench 132, the detector 27 observes the superimposed interference wave from the wafer W, and reads the short-period interference wave of the superimposed interference wave, and interferes from the short period. The time interval of the trench is determined between the times of each pole in the wave, and the etching rate of the deep trench 132 in each pole is obtained from the interference period of the trench. Then, the obtained uranium engraving rate is shown in a graph (refer to Fig. 16). As is clear from the graph of Fig. 6, the etching rate of Example 1 was smaller than that of Comparative Example 1, and the stability was small. Further, the error in the etching amount of the first embodiment and the error time series in the etching amount of the comparative example 1 are shown in the graph (see FIG. 7). From the graph of FIG. 17, the example 1 is known. The etching amount is smaller than that of the comparative example -33-200949932. Accordingly, it is understood that the etching amount calculation method in Fig. 6 can accurately calculate the etching amount. Second Embodiment Next, a wafer W in which a photomask film 131 made of an oxide film is formed on an etched layer 130 made of ruthenium is prepared, and a shallow trench 132 is formed on the etched layer 130 by etching using the substrate processing apparatus 10. . The etching conditions at this time are as follows. @ Actual etching rate 3 60nm/min Selection ratio 1 〇 to 1 (etched layer 1 3 0 to mask film 1 3 1) Opening ratio 〇·2 Measurement wavelength (wavelength of laser light L1) 30〇nm sampling rate 10Hz In the etching of the shallow trench 132, the time from the start point to the end point of the window is set to 25 seconds, and the etching amount calculation method in Fig. 6 is carried out to calculate the etching amount (etching depth) of the shallow trench 132 in each of the timings. Then, the calculated etching amount is shown in a graph (refer to Fig. 18). Comparative Example 2 Further, in the etching of the shallow trench 132, the superimposed interference wave from the wafer W is observed by the detector 27, and all the superimposed interference waves from the start of the etching to the respective timings are obtained by the frequency analysis. The frequency distribution is obtained from the trench-34-200949932 interference period from the start of the etching to the respective timings based on the frequency distribution, and the etching amount (etching depth) at each time is calculated from the trench interference period. That is, the amount of etching is calculated without using the overlapping interference wave. Then 'the calculated graph (see Figure 18). In the graph of Fig. 18, the comparative example 2 is disordered, and the overlapping interference waves of the system are disturbed. There is no confusion about the amount of information. The uranium engraving method can be calculated by the method of calculating the amount of etching obtained by the same overlapping interference wave in the second embodiment and the ratio of the interference. [Embodiment 3] In the above-described first and second embodiments, the etching rate of the channel 132 was carried out in the etching amount calculation method of Fig. 6 in the other wafers. Then, the obtained 饨 (see Fig. 19). Comparative Example 3 Further, in the same etching as in Example 3, an etch rate of 1 1 3 2 was obtained by performing a uranium engraving method similar to the first method, except for using the fast Fourier transform method. Then, the obtained etching is referred to Fig. 19). It can be seen from the graph of Fig. 19 that the uranium engraving of the shallow trench 132 in the third embodiment is shown in Fig. 3, and the etched amount of the etched amount is shown in Fig. 19. According to this, it can be seen that the etching of the trench 132 in the etching is performed correctly. The groove engraving rate in each sequence is shown in the graph, not the maximum entropy method, and the etching amount of the graph is calculated in the trenches in each timing. The rate is shown in the graph (the etching rate is smaller than the etching rate variation of Comparative Example -35-200949932 3). Accordingly, it can be understood that the calculation of the etching amount can be performed stably when the maximum entropy method is used. Example 4 First, prepare the opening. When the wafer W having a rate of 5% and the wafer W having an aperture ratio of 0.5% are etched into the etched layer 130 of each wafer W, the etching amount calculation method of Fig. 15 is performed to obtain each etching rate. The obtained etching rate is shown in a graph (see Fig. 20). Comparative Example 4 A wafer W having an opening ratio of 5% and a wafer W having an aperture ratio of 0.5% were prepared in the same manner as in Example 4, and etching was performed. When the wafer W is etched by the layer 130, the etching amount calculation method of FIG. 6 is performed. The etching rate is then expressed. Then, the obtained uranium engraving rate is shown in the graph (refer to Fig. 21). When comparing the graphs of Fig. 20 and Fig. 21, it is known that the etching amount of Fig. 6 is calculated. At the time of the method, the etch rate of the opening ratio of 5% is stable, but the uranium engraving rate of the opening ratio of 0.5% is unstable. 'When the etching amount calculation method of Fig. 15 is carried out, the aperture ratio is 5%. Any one of the etching rate and the etch rate of the aperture ratio of 0.5% is stable. Accordingly, it is understood that the waveform of the window is removed from the waveform of the extracted window before performing frequency analysis on the waveform of the extracted window. When the waveform is approximated, the etching rate can be accurately obtained even when the aperture ratio is small. [Simplified description of the drawing] -36- 200949932 Fig. 1 is a schematic view showing the substrate processing to which the etching amount calculation method according to the present embodiment is applied. Fig. 2 is a view for explaining a decrease in the etching rate in the uranium engraving of the trench. Fig. 3 is a view for explaining the self-overlapping interference wave extraction in the etching amount calculation method according to the present embodiment. A pattern of waveforms for a specific period. 4 is a diagram showing the frequency distribution obtained by analyzing the waveform of the window in FIG. 3 by using the frequency of the maximum entropy method. FIG. 5 is a diagram for explaining the etching amount calculation method according to the present embodiment. Fig. 6 is a flow chart showing a method of calculating an etching amount according to the present embodiment. Fig. 7 is a view showing a uranium according to a second embodiment of the present invention. Fig. 8 is a diagram showing the error between the calculated etching amount and the actually measured etching amount. ® Fig. 9 is a diagram showing a change in one of the overlapping interference waves when the aperture ratio is changed. Fig. 10 is a diagram showing a change in the frequency distribution of the waveform of the window when the aperture ratio is changed. Fig. 11 is a view showing a pattern of overlapping interference waves when the aperture ratio is 0.5%. Fig. 12 is a view for explaining the waveform of the extraction window in the etching amount calculation method according to the third embodiment of the present invention. Figure 13 is a diagram showing the waveforms of most of the interference waves occupied by the interference film from the overlapped interference wave -37-200949932. Fig. 14 is a view showing a frequency distribution obtained by frequency analysis of a waveform of a majority of the interference wave of the photomask film removed by the superimposed interference wave. Fig. 15 is a flow chart showing a method of calculating an etching amount according to this embodiment. Fig. 16 is a graph showing the ratio of the etching rate calculated by the etching amount calculating method in Fig. 6 and the etching rate obtained from the time between the respective poles in the interference wave. Fig. 17 is an etch amount and an actual etching amount obtained by the etching amount calculated by the etching amount calculation method in Fig. 6 and the actual etching amount, and the time between the respective turns in the interference wave. A comparison chart of the errors. Fig. 18 is a comparison diagram of the etching amount calculated by the etching amount calculation method of Fig. 6 and the etching amount obtained by analyzing the frequency of the superimposed interference wave from the start of etching to each timing. Fig. 19 is a comparison diagram of the uranium engraving rate calculated by the maximum entropy method and the etching rate calculated by the high-speed Fourier transform method. Fig. 20 is a view showing the etching rate of each wafer having different aperture ratios obtained by the etching amount calculation method of Fig. 15. Fig. 21 is a view showing the etching rate of each wafer having different aperture ratios obtained by the etching amount calculation method of Fig. 6. Figure 22 is a diagram for explaining the interference of light in etching. Figure 23 is a diagram showing overlapping interference waves. Fig. 24 is a graph showing the etching rate of -38 to 200949932 obtained from the time between each pole in the interference wave. Fig. 25 is a diagram showing overlapping interference waves subjected to interference. Fig. 26 is a frequency distribution obtained by frequency analysis of overlapping interference waves, Fig. 26(A) shows a case where overlapping interference waves are not interfered, and Fig. 26(b) shows a case where overlapping interference waves are interfered. [Description of main component symbols] β L1 : laser light L2 , l3 , L4 : reflected light w :: wafer 1 〇 : substrate processing apparatus 25 : etching amount calculation device 26 : laser light source 27 : detector 28 : calculation unit 30, 50: overlapping interference wave 31: window 1 3 0 : layer to be etched 1 3 1 : mask film 132: trench -39-