JP2004069651A - Film thickness measuring device - Google Patents

Abstract

An object of the present invention is to provide a film thickness measuring apparatus using a VAR method suitable for in-line measurement in a semiconductor manufacturing process, an FPD manufacturing process, or the like.
A light emitting side optical system that irradiates a measurement medium light including various irradiation angle components to a film thickness measurement point of a sample, and a large number of photoelectric conversion units are arranged in an array on a light receiving surface. Receiving side optics including a photoelectric conversion unit array means, and a lens for guiding reflected light of the measurement medium light for each irradiation angle coming from the film thickness measurement point of the sample onto a light receiving surface of the photoelectric conversion unit array means And a calculating means for calculating a film thickness to be measured based on a series of received light amount data obtained from each photoelectric conversion unit of the photoelectric conversion unit array means, wherein the light receiving side optical system includes The distance between the lens and the light receiving surface of the photoelectric conversion unit array means is set so as to substantially coincide with the focal length of the lens.
[Selection diagram] FIG.

Description

【００４７】
ここで、Ｒは反射率、ｄは薄膜５ａの膜厚、ｎは薄膜５ａの屈折率、θ０は入射角、λは入射光の波長、ｒ０，ｒ１は基板の屈折率ｎ０，薄膜の屈折率ｎ，入射角θ０に関する量である。
【００４８】
よって、基板５の屈折率ｎと、薄膜５ａの屈折率ｎ０と、入射光の波長λとがわかっていれば、式１は膜厚ｄと入射角θ０との関数となり、任意の膜厚ｄにおける入射角θ０に応じた反射率分布が求められる。
【００４９】
例えば、基板５の材質がＳｉ、薄膜５ａの材質がＳｉＯ２のとき、薄膜５ａの膜厚ｄが１０００ｎｍの場合と１５００ｎｍの場合との２つの場合を考える。このとき、式１にて規定される入射角θ０に応じた反射率分布曲線は、図６のグラフに表される。すなわち、同図のグラフにおいて、ａは膜厚１０００ｎｍのときの反射率分布曲線、ｂは膜厚１５００ｎｍのときの反射率分布曲線である。このグラフに示されるように、膜厚が変われば反射率分布波形も異なるものとなることから、入射角に応じた反射率分布を知ることで、逆に、膜厚を測定できることが理解されるであろう。また、入射角に対する反射率の組み合わせを多数求めることができれば、薄膜の膜厚以外に薄膜の屈折率、基板の屈折率が未知数でも、式１からこれらを求めることが可能である。
【００５０】
なお、図５では単層膜に限定して式１を導出しているが（図７及び図８参照）、多層膜に対応した理論式も導出可能であり（図７及び図９参照）、これにより多層膜の膜厚測定も可能である。
【００５１】
演算処理部３のＣＰＵ３１における膜厚の計算処理法としては、カーブフィッティング法を利用することができる。カーブフィッティング法とは、予め計算してテーブルとして記憶しておいた各膜厚に対する波形データ（テーブルデータ）と測定した受光量データとを比較し、最小自乗法により受光量データと最も誤差（すなわち図１０のハッチング部分）の少ないデータを抽出し、その波形データの膜厚を測定対象となっている薄膜の膜厚とする方法である。膜圧の計算処理法としては、他にも極値探索法のような膜厚の計算方法を利用することも可能である。
【００５２】
膜厚の計算処理法として、カーブフィッティングを利用する場合の詳細を以下に説明する。測定対象となる薄膜の屈折率ｎ及びｒ０，ｒ１がキーボード等から入出力部３６を介して入力されると、ＣＰＵ３１では、膜厚ｄ及び入射角θ０の各値に対する反射率の値を式１により演算し、これらをＣＰＵ３１内のメモリにテーブルとして保持する。このようなテーブルの一例が図１１に示されている。同図に示されるように、このテーブルによれば、例えば、膜厚ｄｘ，入射角θｐのときの反射率はＲｄｘ（θｐ）、膜厚ｄｙ，入射角θｐ＋Δθのときの反射率はＲｄｙ（θｐ＋Δθ）であることが判る。
【００５３】
その後、図４のフローチャートに示される処理手順で、カーブフィッティングが実行される。すなわち、先ず、ＣＰＵ３１は、デジタル化された測定データ（反射率の測定データ）Ｍ（θ）をＡ／Ｄ変換部３３から取得する（ステップ４０１）。この測定データＭ（θ）は、言うまでもないが、一次元ＣＣＤ２３から得られたものである。その後、測定データＭ（θ）の取得に続いて、膜厚算出処理が開始される。
【００５４】
この膜厚算出処理では、膜厚ｄを最小膜厚ｄｘに初期設定したのち（ステップ４０２）、図１１の理論テーブルを用いて、膜厚ｄ＝ｄｘにおける反射率の理論データＲｄｘ（θ）と測定データＭ（θ）との差の自乗［Ｒｄｘ（θ）−Ｍ（θ）］^２を入射角範囲θｐからθｑまで、Δθ刻みで計算し、その和Ｐ（ｄ）＝■（Ｒｄｘ（θ）−Ｍ（θ））^２を求めて、メモリ内に記憶する一連の処理が実行される（ステップ４０３）。この理論データと測定データとの差の自乗和を求めてメモリ内に記憶する処理（ステップ４０３）は、膜厚ｄの値をΔｄ刻みで増加させつつ（ステップ４０４）、膜厚ｄがその最大膜厚ｄｚに達するまで（ステップ４０５ＮＯ）、繰り返し実行される（ステップ４０３）。
【００５５】
最大膜厚ｄｚまで自乗和の計算が終了すると（ステップ４０５ＹＥＳ）、メモリに記憶しておいた膜厚範囲ｄｘ〜ｄｚにおける自乗和Ｐ（ｄｘ）〜　Ｐ（ｄｚ）の中から最小の値をとる自乗和Ｐ（ｄｙ）が抽出され（ステップ４０６）、このときの膜厚ｄｙが測定膜厚として決定される（ステップ４０７）。
［第１実施形態］
【００５６】
次に、距離バタツキが組み込まれた膜厚測定装置の実施形態を図１２〜図１９を参照しつつ詳細に説明する。
【００５７】
距離バタツキ対策が組み込まれたセンサヘッド部２の光学的構成が図１２に示されている。この膜厚測定装置は、試料の膜厚測定点に対して、様々な照射角度成分（図示例ではθ０〜θ１）を含む測定媒体光を照射する投光側光学系（図示例では、光源２１と集光レンズ２２とが相当）と、多数の光電変換部を受光面上にアレイ状に配置してなる光電変換部アレイ手段（図示例では一次元ＣＣＤ２３が相当）を含むと共に、試料の膜厚計測点から到来する各照射角度毎の計測媒体光の反射光を光電変換部アレイ手段の受光面上に導くレンズ（図示例ではレンズ２４が相当）を含む受光側光学系と、光電変換部アレイ手段の各光電変換部から得られる一連の受光量データに基づいて測定対象となる膜厚を求める演算手段（図示例では演算処理部３が相当）と、を含んでいる。加えて、受光側光学系に含まれるレンズ（図示例ではレンズ２４が相当）と光電変換部アレイ手段（図示例では一次元ＣＣＤ２３が相当）の受光面との距離は、当該レンズの焦点距離（ｆ）とほぼ一致するように設定されている。
【００５８】
膜厚測定のための基本動作は次の通りである。光源２１から発せられた測定媒体光は、集光レンズ２２の作用で測定対象である基板５上の薄膜５ａに集光して照射される。試料の膜厚測定点は、入射光のほぼ集光位置に置かれる。このとき、θ０〜θ１の範囲の連続した入射角成分を有する測定媒体光が試料へと入射されることになる。
集光レンズ２２を介して入射された測定媒体光は試料で反射される。試料の膜厚測定点から到来する測定媒体光の反射光は、受光レンズ２４の作用で一次元ＣＣＤ２３の受光面に導かれる。これにより、一次元ＣＣＤ２３からは各受光素子（画素）２３ａの受光量データをシリアルに並べたものに相当する一次元ＣＣＤ出力信号ｓ２が送出される。この一次元ＣＣＤ出力信号ｓ２に基づいて、入射角（θ０〜θ１）のそれぞれに応じた反射光強度分布が観測される。このとき観測される入射角のそれぞれに応じた反射光強度分布は、入射角に応じた反射率分布に対応した量であり、これを演算処理部３で理論反射率分布と対比することにより目的とする膜厚が求められる。なお、演算処理部３における膜厚測定のための処理については、先に図１〜図１１を参照して説明した通りである。
【００５９】
図１２に示されるセンサヘッド部２の大きな特徴は、（１）受光レンズ２４と一次元ＣＣＤ２３の受光面とが平行であること、及び（２）受光レンズ２４と一次元ＣＣＤ２３の受光面との距離が、受光レンズ２４の焦点距離ｆとほぼ一致していること、にある。換言すれば、受光レンズ２４に対して試料（基板５）側を前方、一次元ＣＣＤ２３側を後方と定義すれば、一次元ＣＣＤ２３の受光面の位置は受光レンズ２４のほぼ後方焦点位置であると表現することができる。そして、このような配置によれば、距離バタツキの影響を受けない光学系を実現することができる。
【００６０】
距離バタツキ対策の作用説明図が図１３に示されている。今仮に、集光レンズ２２からの光束のうちで、入射角が最も離隔した２本の光線をＬ１，Ｌ２と定義する。また、基準高さＨｒｅｆにあるときの基板を符号５（薄膜を５ａ）、また距離バタツキにより下降したときの基板を符号５´（薄膜を５ａ´）とする。また、光線Ｌ１が基板５で反射された反射光線を符号Ｌ１１、基板５´で反射された反射光線をＬ１２とする。また、光線Ｌ２が基板５で反射された反射光線をＬ２１、基板５´で反射された反射光線をＬ２２とする。さらに、一次元ＣＣＤ２３の受光面上における反射光線Ｌ１１，Ｌ１２の入射点をＰ１、反射光線Ｌ２１，Ｌ２２の入射点をＰ２とする。すると、図から明らかなように、同一の入射光線Ｌ１，Ｌ２に関しては、距離バタツキにより基板が上下移動したとしても、対応する反射光（Ｌ１１，Ｌ１２），（Ｌ２１，Ｌ２２）については、一次元ＣＣＤ２３の受光面上の同一の入射点Ｐ１，Ｐ２に入射することが理解されるであろう。
【００６１】
上述の距離バタツキによる影響を解消する作用は、以下の原理に基づくものである。スネルの法則によれば、入射光の角度と試料の法線方向とが決定されれば、反射角は一意に決定されることが知られている。図１４に示されるように、入射角θで試料に光線Ｌ１が入射している状況を考えると、このときに距離バタツキにより試料の上下変動ΔＬが生じたとしても、入射光線Ｌ１の角度θと試料の法線Ｌ０１，Ｌ０２，Ｌ０３，Ｌ０４の方向は変化しないから、反射角も変化しないことが判る。しかし、距離バタツキが生ずると、反射面の平行移動に伴い、反射する点がＰ１１，Ｐ１２，Ｐ１３，Ｐ１４の如く移動するため、反射光線Ｌ１１，Ｌ１２，Ｌ１３，Ｌ１４も平行に移動する。
ここで、それぞれ異なる入射角θ１，θ２，θ３を有する３本の入射光線Ｌ１，Ｌ３，Ｌ２を想定する。このとき、距離バタツキ（基準高さの基板５，下降位置の基板５′、上昇位置の基板５″）が生ずると、図１５〜図１７に示されるように、各入射光線Ｌ１，Ｌ２，Ｌ３のそれぞれについて、平行な３本の反射光線（Ｌ１０，Ｌ１１，Ｌ１２），（Ｌ２０，Ｌ２１，Ｌ２２），（Ｌ３０，Ｌ３１，Ｌ３２）が生ずる。
【００６２】
先に述べたように、受光レンズ２４と一次元ＣＣＤ２３の受光面との距離は、受光レンズ２４の焦点距離ｆとほぼ一致させてあるため、図１８に示されるように、それら３組の平行光線（Ｌ１０，Ｌ１１，Ｌ１２），（Ｌ２０，Ｌ２１，Ｌ２２），（Ｌ３０，Ｌ３１，Ｌ３２）は、一次元ＣＣＤ２３の受光面上の３点であるＰ１，Ｐ２，Ｐ３に収束することになる。つまり、距離バタツキが生じたとしても、一次元ＣＣＤ２３の出力信号ｓ２を介して観測される波形の入射角（θ１〜θ３）に応じた反射光強度分布は変化しないので、正常な膜厚測定が可能になることが理解されるであろう。
ただ、受光レンズ２４には収差があるのが普通であるから、仮に受光レンズ２４と一次元ＣＣＤ２３の受光面２３ａが平行で、かつ両者の距離が受光レンズ２４の焦点距離ｆと完全に一致したとしても、図１９に示されるように、各組の平行光線の集光点は、厳密な意味では受光面２３ｂ上の一点に収束しない。『〜焦点距離ｆとほぼ一致〜』と表現したのは、このことを意識したためである。
［第２実施形態］
【００６３】
次に、角度バタツキが組み込まれた膜厚測定装置の実施形態を図２０〜図２３を参照しつつ詳細に説明する。
【００６４】
角度バタツキ対策が組み込まれたセンサヘッド部２の光学的構成が図２０に示されている。この膜厚測定装置２は、試料の膜厚測定点に対して、様々な照射角度成分（図示例ではθ０〜θ１）を含む測定媒体光を照射する投光側光学系（図示例では、光源２１と集光レンズ２２とが相当）と、多数の光電変換部を受光面上にアレイ状に配置してなる光電変換部アレイ手段（図示例では一次元ＣＣＤ２３が相当）を含むと共に、試料の膜厚計測点から到来する各照射角度毎の計測媒体光の反射光を光電変換部アレイ手段の受光面上に導く受光光学系（図示省略）と、光電変換部アレイ手段の各光電変換部から得られる一連の受光量データに基づいて測定対象となる膜厚を求める演算手段（図示例では演算処理部３が相当）と、を含んでいる。
【００６５】
さらに、投光側光学系には、測定媒体光に対して基準光軸に相当する特徴付けを行う特徴化手段（図示例ではエッジ整形手段として機能するスリット板２５のスリット２５ａが相当）が含まれており、また受光側光学系には、試料の膜厚計測点から到来する測定媒体光の反射光を受光してそれに含まれる基準光軸に相当する特徴を検出するための第２の光電変換手段（図示例では一次元ＣＣＤ２３そのものが相当）が含まれている。さらに、演算手段には、第２の光電変換手段により検出された基準光軸に相当する特徴に基づいて、各光電変換部から得られる一連の受光量データに含まれる試料の角度バタツキによる誤差成分を修正する受光量データ修正手段が含まれている。
【００６６】
膜厚測定のための基本動作は次の通りである。光源２１から発せられた測定媒体光は、集光レンズ２２の作用で測定対象である基板５上の薄膜５ａに集光して照射される。試料の膜厚測定点は、入射光のほぼ集光位置に置かれる。このとき、θ０〜θ１の範囲の連続した入射角成分を有する測定媒体光が試料へと入射されることになる。
集光レンズ２２を介して入射された測定媒体光は試料で反射される。試料の膜厚測定点から到来する測定媒体光の反射光は、図示しない受光光学系（図１２に示した受光レンズ２４と同等）の作用で一次元ＣＣＤ２３の受光面に導かれる。これにより、一次元ＣＣＤ２３からは各受光素子（画素）２３ａの受光量データをシリアルに並べたものに相当する一次元ＣＣＤ出力信号ｓ２が送出される。この一次元ＣＣＤ出力信号ｓ２に基づいて、入射角（θ０〜θ１）のそれぞれに応じた反射光強度分布が観測される。このとき観測される入射角のそれぞれに応じた反射光強度分布は、入射角に応じた反射率分布に対応した量であり、これを演算処理部３で理論反射率分布と対比することにより目的とする膜厚が求められる。なお、演算処理部３における膜厚測定のための処理については、先に図１〜図１１を参照して説明した通りである。
【００６７】
図２０に示されるセンサヘッド部２の大きな特徴は、（１）集光レンズ２２の入射側にエッジ整形手段として機能するスリット板２５（スリット２５ａ）が配置されて、入射光の断面がエッジ整形されて、基準光軸Ｌｒｅｆ０．Ｌｒｅｆ１が特徴づけられていること、（２）反射光の基準光軸Ｌｒｅｆ０′，Ｌｒｅｆ１′の到達点Ｐｒｅｆ０．Ｐｒｅｆ１が一次元ＣＣＤ２３を介して検出されること、（３）検出された入射点Ｐｒｅｆ０，Ｐｒｅｆ１に基づいて、一次元ＣＣＤ２３からの出力信号ｓ２に含まれる試料の角度バタツキによる誤差成分が修正されること、にある。
【００６８】
すなわち、投光光学系に配置されたエッジ整形手段としてのスリット板２５には、図２１（ａ）に示されるようにその中央部にスリット２５ａが形成されており、このスリット２５ａの長手方向の両端縁により入射角範囲（θ０〜θ１）が決定され、換言すれば、測定媒体光中に２本の基準光軸Ｌｒｅｆ０，Ｌｒｅｆ１が特徴付けされる。これらの２本の基準光軸Ｌｒｅｆ０，Ｌｒｅｆ１は、試料で反射されて反射光の基準光軸Ｌｒｅｆ０′，Ｌｒｅｆ１′となり、一次元ＣＣＤ２３の受光面上の画素位置Ｐｒｅｆ０，Ｐｒｅｆ１に入射される。これら入射点の画素座標Ｒｒｅｆ０，Ｒｒｅｆ１は、一次元ＣＣＤ２３の出力信号ｓ２を例えば暗レベル相当のしきい値で二値化することで、容易に検出することができる。したがって、入射角範囲の両端（θ０とθ１）を一次元ＣＣＤ２３の各画素２３ａを介して観測し、基準位置からの画素位置のずれを見ることで、角度バタツキを検出し、かつ角度バタツキのない状態に戻すことができる。ここで基準位置とは、図２０においてＰｒｅｆ０，Ｐｒｅｆ１に示されるように、角度バタツキが存在しない状態で測定した画素位置のことである。
【００６９】
なお、この例では、エッジ整形手段（スリット板２５）は集光用レンズ２２近傍の光源２１側に配置されているが、レンズ２２近傍の基板５側に配置しても同様の作用を得ることができる。
【００７０】
角度バタツキの検出方法及び補正方法についてさらに詳細に説明する。先に説明したように、この種の膜厚測定装置において角度バタツキが生ずると、一次元ＣＣＤを介して観測される波形の画素列方向両端が基準位置からずれる。この状態では演算処理部で正常に波形処理を行うことができないため、これを膜厚測定のための演算処理に先立って補正する必要がある。
【００７１】
本発明者等の鋭意研究によれば、角度バタツキとそれに伴う測定条件の変動との間には、図２２に示される因果関係があることが知見された。角度バタツキが生ずると、それに伴って試料の傾きもΔθだけ傾き、その結果、試料の法線はＬ０からＬ０′となる。すると、入射角もΔθずれることとなり、入射角範囲はθ０＋Δθ〜θ１＋Δθとなる。入射角がΔθずれると、一次元ＣＣＤ２３で観測される光強度分布波形は、基準位置（Ｐｒｅｆ０〜Ｐｒｅｆ１）に対して２Δθだけずれた画素位置（Ｐｘ０〜Ｐｘ１）へと移動することとなり、一次元ＣＣＤにおいては、この入射角範囲θ０＋２Δθ〜θ１＋２Δθの画素位置（Ｐｘ０〜Ｐｘ１）にて光強度分布波形が観測される。したがって、一次元ＣＣＤ２３で観測された波形から、基準位置に対して波形が何度ずれているかを見ることで、試料が基準位置（Ｈｒｅｆ）に対して何度傾いているかを測定することができる。例えば、観測された波形が、基準位置から２Δθずれていれば、試料はΔθだけ傾いていることになる。
【００７２】
次に、検出した波形を、角度バタツキのない状態に戻すには、観測された波形の入射角範囲を検出されたずれ角度２Δθに応じてΔθだけ戻す操作を演算処理部３で行えばよい。例えば、入射角度範囲θ０＋２Δθ〜θ１＋２Δθの画素位置（Ｐｘ０〜Ｐｘ１）で観測された波形は、入射角度範囲θ０＋Δθ〜θ１＋Δθの範囲（Ｐｒｅｆ０〜Ｐｒｅｆ１）に戻される。この方法で、一次元ＣＣＤで観測される光強度分布波形を角度バタツキのない状態に戻すことができる。この結果、角度バタツキがあっても、正常な膜厚測定が可能になる。
【００７３】
演算処理部３にて実行される角度修正処理を示すフローチャートが図２３に示されている。同図に示されるように、演算処理部３では、先ず、Ａ／Ｄ変換部３３より測定データＭ（θ）を取得し（ステップ２３０１）、次いで測定データＭ（θ）に基づいて前述のアルゴリズムにより波形ずれ量２Δθの算出を行い（ステップ２３０２）、最後に、観測された入射角範囲をΔθだけ戻す処理を実行することにより、角度修正を完了する。その後、先に図４に示した処理を実行することで、正確な膜厚測定が可能となる。
【００７４】
なお、本実施形態においては、基準光軸特徴付けのためのエッジ整形手段としてスリット板２５を使用したが、これに代えて、図２１（ｂ）に示されるアパーチャ板２５′、図２１（ｃ）に示されるナイフエッジ板２５″等の他のエッジ整形手段を採用しても良い。なお、２５ａ′はアパーチャ、２５ａ″はナイフエッジである。
［第３実施形態］
【００７５】
距離バタツキ及び角度バタツキの双方に対する対策が組み込まれたセンサヘッド部の光学的構成が図２４に示されている。同図に示されるように、この膜厚測定装置２は、試料の膜厚測定点に対して、様々な照射角度成分を含む測定媒体光を照射する投光側光学系（図では、光源２１、コリメータレンズ２６、集光レンズ２２が相当）と、多数の光電変換部を受光面上にアレイ状に配置してなる光電変換部アレイ手段（図では、一次元ＣＣＤ２３が相当）を含むと共に、試料の膜厚計測点から到来する各照射角度毎の計測媒体光の反射光を光電変換部アレイ手段の受光面上に導くレンズ（図では、レンズ２４が相当）を含む受光側光学系と、光電変換部アレイ手段の各光電変換部から得られる一連の受光量データに基づいて測定対象となる膜厚を求める演算手段（図では、演算処理部３が相当）と、を含んで構成される。加えて、受光側光学系に含まれるレンズ（レンズ２４）と光電変換部アレイ手段（一次元ＣＣＤ２３）の受光面との距離は、当該レンズの焦点距離（ｆ）とほぼ一致するように設定されている。
【００７６】
また、投光側光学系には、測定媒体光に対して基準光軸に相当する特徴付けを行う特徴化手段（図ではエッジ整形手段として機能するスリット板２５が相当）が含まれており、受光側光学系には、試料の膜厚計測点から到来する測定媒体光の反射光を受光してそれに含まれる基準光軸に相当する特徴を検出するための第２の光電変換手段（図では一次元ＣＣＤ２３自体が相当）が含まれており、さらに演算手段には、第２の光電変換手段により検出された基準光軸に相当する特徴に基づいて、各光電変換部から得られる一連の受光量データに含まれる試料の角度バタツキによる誤差成分を修正する受光量データ修正手段が含まれている。
【００７７】
膜厚測定のための基本動作は次の通りである。光源２１から発せられた測定媒体光は、コリメータレンズ２６にて平行光線とされたのち、スリット板２５のスリット２５ａを通過し、集光レンズ２２の作用で測定対象である基板５上の薄膜５ａに集光して照射される。試料の膜厚測定点は、入射光のほぼ集光位置に置かれる。このとき、θ０〜θ１の範囲の連続した入射角成分を有する測定媒体光が試料へと入射されることになる。
集光レンズ２２を介して入射された測定媒体光は試料で反射される。試料の膜厚測定点から到来する測定媒体光の反射光は、受光レンズ２４の作用で一次元ＣＣＤ２３の受光面に導かれる。これにより、一次元ＣＣＤ２３からは各受光素子（画素）２３ａの受光量データをシリアルに並べたものに相当する一次元ＣＣＤ出力信号ｓ２が送出される。この一次元ＣＣＤ出力信号ｓ２に基づいて、入射角（θ０〜θ１）のそれぞれに応じた反射光強度分布が観測される。このとき観測される入射角のそれぞれに応じた反射光強度分布は、入射角に応じた反射率分布に対応した量であり、これを演算処理部３で理論反射率分布と対比することにより目的とする膜厚が求められる。なお、演算処理部３における膜厚測定のための処理については、先に図１〜図１１を参照して説明した通りである。
【００７８】
図２４に示されるセンサヘッド部２の大きな特徴は、（１）受光レンズ２４と一次元ＣＣＤ２３の受光面とが平行であること、（２）受光レンズ２４と一次元ＣＣＤ２３の受光面との距離が、受光レンズ２４の焦点距離ｆとほぼ一致していること、（３）集光レンズ２２の入射側にエッジ整形手段として機能するスリット板２５（スリット２５ａ）が配置されて、入射光の断面がエッジ整形されて、基準光軸Ｌｒｅｆ０．Ｌｒｅｆ１が特徴づけられていること、（４）反射光の基準光軸Ｌｒｅｆ０′，Ｌｒｅｆ１′の到達点が一次元ＣＣＤ２３を介して検出されること、（５）検出された反射光の基準光軸Ｌｒｅｆ０′，Ｌｒｅｆ１′の入射点と基準入射点Ｐｒｅｆ０，Ｐｒｅｆ１とに基づいて、一次元ＣＣＤ２３からの出力信号ｓ２に含まれる試料の角度バタツキによる誤差成分が修正されること、にある。
【００７９】
距離及び角度バタツキ対策を併用する場合の作用説明図（その１〜その３）が図２５〜図２７に示されている。先に、図１４〜図１８を参照して説明したように、（１）受光光学系に含まれる受光レンズ２４と一次元ＣＣＤ２３の受光面とが平行であること、及び（２）受光レンズ２４と一次元ＣＣＤ２３の受光面との距離が、受光レンズ２４の焦点距離ｆとほぼ一致すること、なる条件が満たされる限り、受光レンズ２４に入射される平行光線は一次元ＣＣＤ２３の受光面上の一点に集光して入射される。ここで、図２５〜図２７を参照して明らかなように、ズレ角Δθが一定である限り距離バタツキの有無に拘わらず、各高さ位置（上昇位置、基準位置、下降位置）にある基板５″，５，５′からの３組の反射光（Ｌ１０，Ｌ１１，Ｌ１２），（Ｌ２０，Ｌ２１，Ｌ２２），（Ｌ３０，Ｌ３１，Ｌ３３）は、いずれの組においても互いに平行な関係を維持している。そのため、この実施形態にあっても、入射光Ｌ１，Ｌ２，Ｌ３に対する反射光（Ｌ１０，Ｌ１１，Ｌ１２），（Ｌ２０，Ｌ２１，Ｌ２２），（Ｌ３０，Ｌ３１，Ｌ３３）は、一次元ＣＣＤ２３の受光面上の３点Ｐ１，Ｐ２，Ｐ３に必ず集光されるから、距離バタツキによる影響の受けない光学系を実現することができる。
【００８０】
一方、図２５〜図２７において、ズレ角Δθが変動した場合には、先に、図２２を参照して説明したように、一次元ＣＣＤ２３の受光面に入射する反射光の入射点は、法線Ｌ０から法線Ｌ０′へのズレ角（Δθ）に応じて画素列方向へと所定角度分（２Δθ）移動することとなる。先の例で説明したように、投光光学系に配置されたスリット板２５のスリット２５ａは、エッジ整形手段として機能して入射角範囲（θ０〜θ１）を規定しており、換言すれば、入射光中に基準光軸Ｌｒｅｆ０，Ｌｒｅｆ１を特徴づけている。これら基準光軸Ｌｒｅｆ０，Ｌｒｅｆ１の反射光の一次元ＣＣＤ２３への入射点Ｐｘ０，Ｐｘ１は、一次元ＣＣＤ２３の出力信号ｓ２から検出することができる。演算処理部３では、検出された入射点Ｐｘ０，Ｐｘ１と基準となる入射点Ｐｒｅｆ０，Ｐｒｅｆ１とのズレ角２θに基づいて、試料の実際のズレ角Δθを求め、観測波形を角度バタツキのない状態に戻す処理を実行する。
【００８１】
したがって、この実施形態によれば、試料の距離バタツキ並びに角度バタツキに拘わらず、試料の単層薄膜又は多層薄膜の膜厚を高精度に測定することが可能となる。殊に、この実施形態によれば、距離バタツキに関する補正は一次元ＣＣＤ２３で観測された時点で既に完了しており、距離バタツキによる観測波形の変化は解消されている。したがって、観測波形が変化する要因は角度バタツキのみであり、演算処理部では角度バタツキに起因する誤差成分を補正する処理を実行するだけで済む。このように、距離バタツキ及び角度バタツキは互いに独立な過程で補正されるため、これら２種類のバタツキに起因する観測波形の変化を演算処理部で同一の過程で補正する場合のように、２つの補正処理が互いに競合して演算処理が収束しないと言った不都合は生じない。
その結果、この実施形態によれば、設置条件（距離／角度バタツキ）が緩和され、オートフォーカス機能が不要となる。同時に、装置の小型化、高速演算処理も実現され、インライン計測に適した装置を提供することが可能となる。
［第４実施形態］
【００８２】
２種類の光源並びに距離、角度バタツキ対策を施したセンサヘッドの光学的構成が図２８に示されている。同図に示されるように、この実施形態における光学系には、２種類の光源２１Ａ，２１Ｂが含まれている。これらの光源２１Ａ，２１Ｂからの光はハーフミラー２１Ｃを経て同軸整合されたのち、コリメータレンズ２６を経て平行光に整形される。平行光の経路部分には、スリット２５ａと偏光子２７とが配置されており、これにより入射角範囲（θ０〜θ１）と、光の偏光方向が決定される。偏光子２７を通過した光は集光レンズ２２によって集められ、測定対象である基板５上の薄膜５ａに集光して照射される。試料の膜厚測定点は入射光のほぼ集光位置に置かれる。ここで、偏光子２７は、入射光の偏光状態を入射面に対して４５度傾けるように設定する。これにより、連続した入射角θ０〜θ１を有すると共に、偏光子２７によって決定された偏光方向を有する測定媒体光が試料に入射されることになる。
【００８３】
入射された測定媒体光は試料で反射され、受光レンズ２４を通ってから、偏光ビームスプリッタ２８を経てＳ偏光とＰ偏光に分離される。分離された反射光は、２つの一次元ＣＣＤ２３Ａ，２３Ｂの一連の画素を介して入射角（θ０〜θ１）に応じたＳ偏光とＰ偏光の反射光強度分布として観測される。観測された反射強度分布波形は、入射角に応じた反射率分布に対応した量であり、これを演算処理部３で、理論反射率分布と対比して膜厚が求められる。
【００８４】
なお、受光レンズ２４と各一次元ＣＣＤ２３Ａ，２３Ｂとの距離は受光レンズ２４の焦点距離ｆとほぼ一致させてあり、又演算処理部３では前述の角度バタツキに対応したデータ修正処理が実行されることは言うまでもない。
【００８５】
なお、先に説明した反射率の理論式（式１）においては、波長も変数（λ）として扱われているため、２つの光源２１Ａ，２１Ｂが使用されている第４実施形態にあっては、それら２つの光源から照射される測定媒体光の波長を異なるものとすることで（使用する波長の数を増やせば）、測定できる反射率の情報が増えるため測定精度を上げることが可能である。
【数２】

【００８６】
理論式（式１）は単一波長を想定しているため、１つの光源のみを使用して上記したように複数の波長を使用する際には、受光部で何らかの分光装置を用いて複数の波長に分離し、各波長に対する反射率を測定する必要がある。第４実施形態では、２つの光源２１Ａ，２１Ｂを使用しているため、それら各光源を異なる波長のものとして時分割的に動作（例えば均等に交互に光らせる等）させ、それぞれの波長に対する反射率を交互に測定するといった測定方法を採用することができる。
【００８７】
使用する光源の波長は、任意に変えることができ、より薄い膜を測定したいときには波長の短いものを用いる等、用途によって使い分けることができる。具体的には、紫外光、可視光、赤外光等の半導体レーザを用いることができる。また、複数の波長の光源を使用する結果、屈折率の分光特性を得ることも可能である。
【００８８】
また、通常のレンズには色収差があるため、集光レンズによって試料に集光して照射する際に、集光点が波長によって異なる等の問題がある。しかし、装置に使用する集光レンズに、アクロマティックレンズのような、色収差の補正されたレンズを用いれば、光源を取り替えるだけで様々な波長の光源に対応することが可能である。
【００８９】
この実施形態にあっては、図２８に示されるように、偏光子２７を使用することにより、試料への入射光の偏光状態を入射面に対して４５度傾けるように設定している。このような偏光特性を有する光を試料に入射し、受光部の偏光ビームスプリッタ２８でＳ偏光とＰ偏光に分離して、反射率の入射角分布を測定している。前述のように、反射率の入射角分布は式１で表される。しかし、式１に含まれるｒ０，ｒ１の値は、
【外１】

と、
【外２】

では、式２に表されるように異なる理論式となる。
【数３】

【００９０】
よって、Ｓ偏光とＰ偏光の両方を測定することで、測定できる反射率の情報が増えるため測定精度を向上することが可能である。
［第５実施形態］
【００９１】
基準光軸を専用の光源にて設定するようにしたセンサヘッド部の光学的構成が図２９に示されている。先に説明した第２乃至第４実施形態にあっては、角度補正に必要な入射光の基準軸特徴付けを、スリット等のエッジ整形手段を使用して行っているが、この第５実施形態にあっては、別途専用の光源を付加することで、基準軸の特徴付けを行っている。また、先に説明した第２乃至第４実施形態にあっては、角度補正に必要な基準光軸の位置検出を膜厚測定用ＣＣＤそのものを使用して行っていたが、この第５実施形態にあっては、別途専用の光電センサ（ＰＳＤ）を使用することで、基準光軸の位置検出を行っている。
【００９２】
すなわち、図２９において、光源２１から出射された光は、ビームスプリッタ２９ｂを通り、集光レンズ２２によって集められ、測定対象である基板５上の薄膜５ａに集光して照射される。試料膜厚測定点は、入射光（測定媒体光）のほぼ集光位置に置かれる。入射光は試料で反射され、受光レンズ２４を通ってから、偏光ビームスプリッタ２９ｃを通り、一次元ＣＣＤ２３の各画素２３ａで、入射角に応じた反射光強度分布として観測される。観測された波形は、入射角に応じた反射率分布に対応した量であり、これを演算処理部３で、理論反射率分布と対比して膜厚が求められる。受光レンズ２４は一次元ＣＣＤ２３と平行に、かつほぼ後方焦点位置に配置される。この配置により、距離バタツキの影響を受けない光学系を実現することができる。
【００９３】
一方、光源２９ａから出射された細いビーム状の光は、ビームスプリッタ２９ｂで向きを変えられ、集光レンズ２２を通過して、測定対象である基板５上の薄膜５ａに基準光軸として照射される。入射された基準光軸に相当する光は試料で反射され、受光用レンズを通過してから、偏光偏光ビームスプリッタ２９ｃで向きを変えられ、ＰＳＤ（Ｐｏｓｉｔｉｏｎ　Ｓｅｎｓｉｎｇ　Ｄｅｖｉｃｅ）２９ｄに照射される。このとき、光軸設定用の光源２９ａからの光は、膜厚測定用の光源２１からの光とは異なる偏光状態の光とし、偏光ビームスプリッタ２９ｃにより光源２１からの光と分離されている。ＰＳＤ２９ｄにおける観測位置Ｐｘに対して、基準位置Ｐｒｅｆからのずれを検出することで、角度バタツキを検出し、かつ角度バタツキのない状態に戻すことができる。ここで、基準位置Ｐｒｅｆとは、図３０に示されるように、角度バタツキが存在しない状態で測定した基準光軸の照射位置のことである。この配置により、角度バタツキの影響を受けない光学系を実現することができる。なお、角度バタツキの修正に必要な演算処理については、先に、第２実施形態にて詳細に説明したので省略する。
【００９４】
なお、上記した第５実施形態においては、光源２１と光源２９ａを時分割的に動作させるとともに、それに同期させて強度分布波形を習得することで、２つの波長の測定媒体光に基づく膜厚測定を行うことも可能である。或いは、光源２１と光源２９ａの波長を異なったものとし、ビームスプリッタ２９ｂ、２９ｃに波長分離機能を有するダイクロイックミラー等を用いることで、２つの波長の測定媒体光に基づく膜厚測定を行うようにしてもよい。
【００９５】
以上説明した第１乃至第５実施形態からも明らかなように、本発明の膜厚測定装置は、試料の膜厚測定点に対して、様々な照射角度成分を含む測定媒体光を照射する投光側光学系と、多数の光電変換部を受光面上にアレイ状に配置してなる光電変換部アレイ手段を含むと共に、前記試料の膜厚計測点から到来する各照射角度毎の計測媒体光の反射光を前記光電変換部アレイ手段の受光面上に導くレンズを含む受光側光学系と、前記光電変換部アレイ手段の各光電変換部から得られる一連の受光量データに基づいて測定対象となる膜厚を求める演算手段と、を具備し、前記受光側光学系に含まれる前記レンズと前記光電変換部アレイ手段の前記受光面との距離は、当該レンズの焦点距離とほぼ一致するように設定されている、ことを特徴とするものである。
【００９６】
そのため、試料と受光光学系との距離が変動したとしても、同一の入射光軸に対応する反射光軸は単に平行移動するだけであるから、受光側光学系に含まれるレンズと光電変換部アレイ手段の受光面との距離が当該レンズの焦点距離とほぼ一致するように設定されていれば、各入射光軸に対応する反射光は光電変換部アレイ手段上の決められたアレイ位置に必ず到達することとなるから、距離バタツキに起因する反射光の強度分布波形の変動が抑制されて、距離バタツキに対する耐性が向上する。
【００９７】
また、前記投光側光学系には、前記測定媒体光に対して基準光軸に相当する特徴付けを行う特徴化手段が含まれており、前記受光側光学系には、前記試料の膜厚計測点から到来する測定媒体光の反射光を受光してそれに含まれる基準光軸に相当する特徴を検出するための第２の光電変換手段が含まれており、さらに前記演算手段には、前記第２の光電変換手段により検出された基準光軸に相当する特徴に基づいて、各光電変換部から得られる一連の受光量データに含まれる試料の角度バタツキによる誤差成分を修正する受光量データ修正手段が含まれている、ものである。
【００９８】
そのため、試料に対する入射角度が変動すると、第２の光電変換手段にて検出される基準光軸に相当する特徴に基づいて、各光電変換部から得られる一連の受光量データに含まれる試料の角度バタツキによる誤差成分が修正されて、角度バタツキに起因する反射光の強度分布波形の変動が抑制されることとなるから、角度バタツキに対する耐性も向上する。しかも、このような膜厚測定装置にあっては、距離バタツキに対する修正機能と角度バタツキに対する修正機能とは互いに影響を及ぼすことなく独立に作用するため、両者の修正機能を同時に実現しつつも、演算処理を複雑化することがないと言う利点を有する。すなわち、距離バタツキに関する修正機能は光学系の構成（『焦点距離にほぼ一致』）で実現される一方、角度バタツキに関する修正機能は電気系の構成（『一連の受光量データの誤差修正演算』）で実現されるため、両者が相互に影響し合って演算結果の収束に長時間を要すると言った不都合が回避される。
【００９９】
なお、以上の第１乃至第５実施形態にあっては、投光用レンズ（集光レンズ２２）と受光用レンズ（受光レンズ２４）とを分けて、試料に斜入射する方式を採用したが、図３３を参照して説明した従来例（特開平９−１９６６３０）のように、投光用と受光用に同一のレンズ（中心入射角が０°）を用いた同軸型でも距離／角度バタツキに強い光学系を実現することができる。このように同軸型で構成した本発明に係るセンサヘッドの一例を図３１に示す。
【０１００】
ただし、同軸型では集光レンズ（受光レンズ）の開口数Ｎ．Ａ．が０．４５以上であると、図３２に示されるように、受光レンズ２２Ａの後方焦点距離に一次元ＣＣＤ２３を配置することができない。よって、同軸型では開口数Ｎ．Ａ．を０．４５以下（入射角範囲は２７°以下）にしないと、距離／角度バタツキに強い光学系を実現することは困難である。反射率の理論式である式１は入射角の関数であり、入射角範囲は広く取れることが望ましい。斜入射型は、投光用と受光用が分離しているため、上記のような制限がなく、入射角範囲を広く取れる利点がある。
【０１０１】
【発明の効果】
以上の説明で明らかなように、本発明によれば、小型で高速演算処理が可能であり、しかも距離バタツキや角度バタツキに対する耐性を有するＶＡＲ法利用の膜厚測定装置を提供することができ、この種の膜厚測定装置を例えば半導体製造プロセスやＦＰＤ製造プロセス等におけるインライン計測に適用するとが可能となる。
【図面の簡単な説明】
【図１】ＶＡＲ方向決め用いた膜厚測定装置の全体を示す図である。
【図２】センサヘッド部の光学的構成を示す図である。
【図３】演算処理部の電気的構成を示す図である。
【図４】膜厚測定処理を示すフローチャートである。
【図５】薄膜と入射光、反射光との関係を示す説明図である。
【図６】入射角に応じた反射率分布曲線を示すグラフである。
【図７】多層薄膜と各層の特性係数との関係を示す説明図である。
【図８】単層膜の光学特性を示す説明図である。
【図９】多層膜の光学特性を示す説明図である。
【図１０】最小自乗法での採用データを示す説明図である。
【図１１】演算部のメモリに格納されたテーブルの内容を示す図である。
【図１２】距離バタツキ対策の組み込まれたセンサヘッド部の光学的構成を示す図である。
【図１３】距離バタツキ対策の作用説明図である。
【図１４】距離バタツキ対策の原理説明図である。
【図１５】距離バタツキ対策の作用を入射光路別に示す説明図（その１）である。
【図１６】距離バタツキ対策の作用を入射光路別に示す説明図（その２）である。
【図１７】距離バタツキ対策の作用を入射光路別に示す説明図（その３）である。
【図１８】距離バタツキ対策の作用を受光用レンズの後段について示す説明図である。
【図１９】一次元ＣＣＤ受光面付近の集光状態を拡大して示す説明図である。
【図２０】角度バタツキ対策の組み込まれたセンサヘッド部の光学的構成を示す図である。
【図２１】エッジ整形手段の説明図である。
【図２２】角度バタツキ対策の作用説明図である。
【図２３】角度修正処理を示すフローチャートである。
【図２４】距離バタツキ及び角度バタツキの双方に対する対策が組み込まれたセンサヘッド部の構成を示す図である。
【図２５】距離及び角度バタツキ対策を併用する場合の作用説明図（その１）である。
【図２６】距離及び角度バタツキ対策を併用する場合の作用説明図（その２）である。
【図２７】距離及び角度バタツキ対策を併用する場合の作用説明図（その３）である。
【図２８】２つの光源並びに距離、角度バタツキ対策を施したセンサヘッドの光学的構成を示す図である。
【図２９】基準光軸を専用の光源にて設定する場合の光学的構成を示す図である。
【図３０】基準光軸を専用の光源にて設定する場合の作用説明図である。
【図３１】同軸型で構成したセンサヘッド部の説明図（その１）である。
【図３２】同軸型で構成したセンサヘッド部の説明図（その２）である。
【図３３】ＶＡＲ法を用いた従来の膜厚測定装置の説明図である。
【図３４】距離バタツキの説明図である。
【図３５】角度バタツキの説明図である。
【符号の説明】
１　　　　　膜厚測定装置
２　　　　　センサヘッド部
３　　　　　演算処理部
４　　　　　演算処理部
５　　　　　基板
５ａ　　　　薄膜
２１，２１Ａ，２１Ｂ　　　　光源
２１Ｃ　　　ハーフミラー
２２　　　　集光用レンズ
２３，２３Ａ，２３Ｂ　　　　一次元ＣＣＤ
２３ａ　　　画素
２３ｂ　　　受光面
２４　　　　受光用レンズ
２５　　　　スリット板
２５ａ　　　スリット
２５′　　　アパーチャ板
２５ａ′　　アパーチャ
２５″　　　ナイフエッジ板
２５ａ″　　ナイフエッジ
２６　　　　コリメータレンズ
２７　　　　偏光子
２８　　　　偏光ビームスプリッタ
２９ａ　　　基準光軸用の光源
２９ｂ　　　偏光ビームスプリッタ
２９ｃ　　　偏光ビームスプリッタ
２９ｄ　　　ＰＳＤ
Ｈｒｅｆ　　試料の基準位置
Ｌｒｅｆ０，Ｌｒｅｆ１　　　　基準光軸
ｆ　　　　　焦点距離
Ｐ１，Ｐ２，Ｐ３　　　　入射点
θ０〜θ１　　　　入射角[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a film thickness measurement apparatus suitable for performing in-line film thickness measurement in, for example, a semiconductor manufacturing process or an FPD (Flat Panel Display) manufacturing process, and more particularly, to a sample distance flutter that tends to occur during in-line measurement. The present invention relates to a film thickness measurement apparatus using a VAR (Variable Angle Reflectometry) method that enables high-precision in-line measurement by improving resistance to angle and flutter.
[0002]
[Prior art]
In recent years, in the semiconductor manufacturing process, with the enlargement of semiconductor substrates and the miniaturization of design rules, the possibility of enormous damage to defects and the need to manage subtle abnormalities have arisen. Sex is growing more and more. Also, in the process of manufacturing an FPD (Flat Panel Display) typified by an LCD (Liquid Crystal Display) or a PDP (Plasma Display Panel), as the size of a glass substrate increases, a large screen, high definition, and high quality are achieved. Inspection is rapidly progressing, and the importance of inspection is increasing in order to produce high quality products with high yield.
[0003]
Conventionally, product inspection in this type of manufacturing process has been performed by off-line measurement using a large and expensive film thickness measuring device. This off-line measurement is performed through a series of procedures of extracting a product from the manufacturing process, transporting the product to a remote film thickness measuring device, and performing measurement and confirmation. In such off-line measurement, if the measurement result is out of the control standard, it takes time to feed back the information and reflect / correct it in the process. It is not possible to judge whether it is off or not, and there is a problem that the yield is reduced.
[0004]
Therefore, for example, in-line measurement is realized by incorporating a film thickness measuring device during a film forming process (in-situ) or in a manufacturing line immediately after the film forming process, and 100% measurement is performed without removing a product from the manufacturing process. As a result, there is an increasing need to improve product yield.
[0005]
A film thickness measurement apparatus applicable to such in-line measurement has (1) the same performance as a film thickness measurement apparatus used for conventional off-line measurement, and (2) is small and capable of high-speed arithmetic processing. (3) It is required to have various conditions such as resistance to distance and angle flutter described later.
[0006]
By the way, as a film thickness measuring method capable of measuring the film thickness of a thin film with high accuracy, a VAR (Variable Angle Reflectometry) method is conventionally known. In the VAR method, the thickness of a thin film is measured using a reflectance distribution according to an incident angle of light.
[0007]
FIG. 33 is an explanatory diagram of a conventional film thickness measuring apparatus using the VAR method. FIG. 1A shows an incident angle scanning type (tentative name) film thickness measuring device described in Japanese Patent Application Laid-Open No. 04-184104, and FIG. What is described is a multi-angle simultaneous incidence type (tentative name) film thickness measuring device described in Japanese Patent Application Laid-Open No. 9-196630.
[0008]
First, an incident angle scanning type film thickness measuring apparatus will be described with reference to FIG. The laser light sources 101 and 104 are output stabilized HeNe lasers having a wavelength of 6328 °, and output lights of the lasers 101 and 104 are converted into S and P polarized lights by polarizers 103 and 106, respectively. These S and P polarized lights are radiated to the measurement sample O via the polarizing beam splitter 107 having a high extinction ratio. The polarization state of the light applied to the measurement sample O can be selected from S and P by the shutters 102 and 105. The laser light sources 101 and 104, the shutters 102 and 105, the polarizers 103 and 106, and the polarization beam splitter 107 constitute a light source device. The measurement sample O is supported by the turntable 111. The rotation arm 110 is coaxially attached to the turntable 111. When the rotation arm 110 rotates by 2θ, the turntable 111 rotates by θ. The turntable 111 and the rotary arm 110 constitute support means. In addition, 108 is a calculating means, and 109 is a light receiving element.
[0009]
Next, a multi-angle simultaneous incidence type film thickness measuring apparatus will be described with reference to FIG. This film thickness measuring apparatus includes an irradiation optical system 215 (mainly composed of a laser light source 214, a half mirror 213, and an objective lens 216) for irradiating convergent light or divergent light 215 onto the surface of a sample 217, and the convergent light. Alternatively, a light receiving optical system 214 (mainly an objective lens 216, a half mirror 213, a half mirror 210, a condenser lens 209, a pinhole 208, a beam splitter 219, and array sensors 207 and 211 for guiding divergent light reflected from the sample) And an optical constant calculation device for calculating the optical constant of the sample or the optical constant of a thin film formed on the sample surface based on the light intensity distribution in the light beam of the reflected light guided by the light receiving optical system. 206. Reference numeral 218 denotes a movable stage, 201 denotes a CCD camera, 202 denotes a relay lens, 203 denotes a white light source, 204 denotes a condenser lens, and 205 denotes a half mirror.
[0010]
[Problems to be solved by the invention]
However, in the incident angle scanning type film thickness measuring device shown in FIG. 33A, the measurement sample O or the optical system is rotated in order to measure the reflectance distribution according to the incident angle. This method has disadvantages such as an increase in the size of the entire apparatus and a long time required for measurement due to the presence of a rotating mechanism, and is not suitable for in-line measurement.
[0011]
On the other hand, according to the multi-angle simultaneous incidence type film thickness measuring apparatus shown in FIG. 33 (b), by using an objective lens having a large aperture diameter, measurement medium light having various incident angle components can be simultaneously transmitted to the sample. Since the irradiation is performed, a mechanism for rotating the sample or the optical system is not required, and the entire apparatus can be reduced in size accordingly. However, in this multi-angle simultaneous incidence type film thickness measuring apparatus, even if the relative positional relationship between the optical system and the sample slightly changes, the intensity of the reflected light observed through the photoelectric element array is small. Since the distribution waveform fluctuates greatly, the installation conditions of the apparatus are greatly limited, and are still unsuitable for in-line measurement. Here, the variation in the relative positional relationship between the optical system (mainly the objective lens) and the sample is assumed to be a distance flap (tentative name) shown in FIG. 34 and an angle flap (tentative name) shown in FIG. You.
[0012]
With reference to FIG. 34, distance flutter will be described. In the drawing, 301 is an incident light beam, 302 is a reflected light beam, 311 is an intensity distribution waveform of reflected light when the sample is in the reference height state, and 312 is an intensity distribution waveform of reflected light when the sample is in the up state. Reference numeral 313 denotes an intensity distribution waveform of reflected light when the sample is in a lowered state, 321 denotes a light receiving element array such as a one-dimensional CCD or LED array, 322 denotes a reference height, 323 denotes an objective lens, 324 denotes a semiconductor product or FDP, etc. The sample 324a is a thin film on the sample surface.
[0013]
Distance flapping is a phenomenon in which the distance between the optical system (for example, the objective lens 323) and the sample 324 fluctuates. When this distance flutter occurs, the width of the intensity distribution waveform 311 of the reflected light observed through the light receiving element array 321 in the array column direction changes, so that the optical constant of the thin film calculated based on the intensity distribution waveform is It will be wrong.
[0014]
As shown in FIG. 3B, in the reference height state where the upper surface of the sample 324 coincides with the reference height 322, the converging point of the incident light beam 301 coincides with the upper surface of the sample 324, and the incident light beam 301 Since the reflected light flux 302 and the reflected light beam 302 completely overlap each other, an intensity distribution waveform of reflected light having a reference array column direction width 311 is observed on the output side of the light receiving element array 321.
[0015]
As shown in FIG. 11A, in the ascending state in which the upper surface of the sample 324 has risen above the reference height 322, the focal point of the incident light beam 301 is located below the upper surface of the sample 324, and Since the contour of the reflected light beam 302 is smaller than the contour of the light beam 301, an intensity distribution waveform 312 of the reflected light narrower than the reference intensity distribution waveform 311 is observed on the output side of the light receiving element array 321. You.
[0016]
As shown in FIG. 3C, in the descending state in which the upper surface of the sample 324 is lower than the reference height 322, the condensing point of the incident light beam 301 is located above the upper surface of the sample 324, and Since the contour of the reflected light beam 302 is larger than the contour of the light beam 301, an intensity distribution waveform 313 of the reflected light wider than the reference intensity distribution waveform 311 is observed on the output side of the light receiving element array 321. .
[0017]
The angle flutter will be described with reference to FIG. In the drawing, 401 is an incident light beam, 402 is a reflected light beam, 411 is an intensity distribution waveform of reflected light when the sample is in a reference angle (horizontal) state, and 412 is reflected light when the sample is in a downwardly inclined state. 413, an intensity distribution waveform of reflected light when the sample is inclined downward to the right, 421, a light receiving element array such as a one-dimensional CCD or LED array, 422, a reference angle (horizontal), 423, an objective lens. Reference numeral 424 denotes a sample such as a semiconductor product or FDP, and 424a denotes a thin film on the surface of the sample.
[0018]
Angle flapping is a phenomenon in which the angle formed between the optical axis of an optical system (for example, the objective lens 423) and the sample 424 fluctuates. When this angular flapping occurs, the position of the intensity distribution waveform 411 of the reflected light observed via the light receiving element array 421 in the array column direction changes (shifts), so that the thin film calculated based on the intensity distribution waveform is calculated. The optical constants will be wrong.
[0019]
As shown in FIG. 13B, in a reference angle (horizontal) state where the upper surface of the sample 424 coincides with the reference angle (horizontal) 422, the focal point of the incident light beam 401 coincides with the upper surface of the sample 424. Since the incident light beam 401 and the reflected light beam 402 completely and coaxially overlap each other, on the output side of the light receiving element array 421, the intensity distribution waveform of the reflected light having the width and position in the array column direction indicated by reference numeral 411 is observed. You.
[0020]
As shown in FIG. 9A, in the downwardly inclined state where the upper surface of the sample 424 is inclined to the left from the reference angle (horizontal level), the optical axis of the reflected light beam 402 is more than the optical axis of the incident light beam 401. Is also inclined to the left, an intensity distribution waveform 412 shifted leftward from the reference intensity distribution waveform 411 is observed on the output side of the light receiving element array 421.
[0021]
As shown in FIG. 13C, in a downward-sloping state in which the upper surface of the sample 424 is inclined downward to the right from the reference angle (horizontal level), the optical axis of the reflected light beam 402 is smaller than the optical axis of the incident light beam 401. Is also tilted to the right, an intensity distribution waveform 413 shifted to the right from the reference intensity distribution waveform 411 is observed on the output side of the light receiving element array 421.
[0022]
The multi-angle simultaneous incidence type film thickness measuring apparatus described in the above-mentioned Japanese Patent Application Laid-Open No. 9-196630 has an auto-focus function and a stage moving function in order to deal with distance flutter and angle flutter. ing. However, a film thickness measuring apparatus equipped with these functions is still unsuitable for in-line measurement because the whole apparatus becomes large and the processing time becomes long.
[0023]
The present invention has been made in view of the above-mentioned problems in the conventional film thickness measuring apparatus, and has as its object the use of the VAR method suitable for in-line measurement in, for example, a semiconductor manufacturing process or an FPD manufacturing process. To provide a film thickness measuring device.
[0024]
A more specific object of the present invention is to provide a film thickness measuring apparatus using the VAR method, which is compact, capable of high-speed arithmetic processing, and has resistance to distance and angle flutter.
[0025]
Still other objects and operational effects of the present invention will be easily understood by those skilled in the art by referring to the following description in the specification.
[0026]
[Means for Solving the Problems]
The film thickness measuring apparatus according to the present invention includes a light projecting side optical system that irradiates a measurement medium light including various irradiation angle components to a film thickness measuring point of a sample, and an array of a large number of photoelectric conversion units on a light receiving surface. A lens including a photoelectric conversion unit array means arranged in a shape, and guiding reflected light of measurement medium light for each irradiation angle coming from the film thickness measurement point of the sample onto a light receiving surface of the photoelectric conversion unit array means A light receiving side optical system comprising: a light receiving side optical system; and a calculating means for obtaining a film thickness to be measured based on a series of light receiving amount data obtained from each photoelectric conversion unit of the photoelectric conversion unit array means. The distance between the lens included in the system and the light receiving surface of the photoelectric conversion unit array means is set so as to substantially coincide with the focal length of the lens.
[0027]
According to such a configuration, even if the distance between the sample and the light receiving optical system fluctuates, the reflected light axis corresponding to the same incident optical axis simply moves in parallel, and is included in the light receiving side optical system. If the distance between the lens and the light receiving surface of the photoelectric conversion unit array means is set to substantially match the focal length of the lens, the reflected light corresponding to each incident optical axis is determined on the photoelectric conversion unit array means. Since the light always reaches the array position, the fluctuation of the intensity distribution waveform of the reflected light due to the distance flutter is suppressed, and the resistance to the distance flutter is improved.
[0028]
Here, the expression "substantially coincides with the focal length" with respect to the distance between the lens and the light receiving surface is based on the consideration that the lens may have a slight deviation from the focal length due to the presence of aberration. is there.
[0029]
In the film thickness measuring apparatus of the present invention, the light-emitting side optical system includes characterization means for performing characterization corresponding to a reference optical axis with respect to the measurement medium light, and the light-receiving side The optical system includes second photoelectric conversion means for receiving reflected light of the measurement medium light coming from the film thickness measurement point of the sample and detecting a feature corresponding to a reference optical axis included therein. Further, the arithmetic means includes an angle flutter of a sample included in a series of received light amount data obtained from each photoelectric conversion unit based on a feature corresponding to a reference optical axis detected by the second photoelectric conversion means. And a light-receiving-amount data correcting means for correcting an error component due to
[0030]
According to such a configuration, when the incident angle with respect to the sample changes, a series of received light amount data obtained from each photoelectric conversion unit based on a feature corresponding to the reference optical axis detected by the second photoelectric conversion unit Since the error component due to the angle flutter of the sample included in the above is corrected and the fluctuation of the intensity distribution waveform of the reflected light caused by the angle flutter is suppressed, the resistance to the angle flutter is also improved. Moreover, in such a film thickness measuring device, the correction function for the distance flutter and the correction function for the angle flutter operate independently without affecting each other. There is an advantage that the arithmetic processing is not complicated. That is, the correction function relating to the distance flutter is realized by the configuration of the optical system ("substantially coincides with the focal length"), while the correction function relating to the angular flutter is realized by the configuration of the electric system ("error correction calculation of a series of received light amount data"). Therefore, it is possible to avoid such a disadvantage that the two influence each other and it takes a long time to converge the operation result.
[0031]
As the characterizing means of the present invention, it is possible to adopt various configurations such as providing a dedicated light source separately from the measurement medium light to form a beam indicating the reference optical axis. However, if the characterization means is a cross-sectional contour shaping means for edge-shaping a portion corresponding to the reference optical axis in the cross-sectional contour of the measurement medium light, such a dedicated light source is not required, which is advantageous in cost. It becomes. As such a section contour shaping means, a slit, an aperture (including a case where a lens itself used for light projection functions as an aperture), a knife edge, or the like can be used. Further, in the film thickness measuring apparatus of the present invention, if the photoelectric conversion unit array means also serves as the second photoelectric conversion means, the second photoelectric conversion means itself becomes unnecessary, which is more advantageous in cost. Become.
[0032]
In the film thickness measuring apparatus according to the present invention, the received light amount data correcting means may calculate an angle deviation amount of the sample from a light receiving position on the second photoelectric conversion means in a portion corresponding to a reference optical axis of the measurement medium light. A function of shifting a series of received light amount data obtained from each of the photoelectric conversion units of the photoelectric conversion unit array means in a direction in which the angle shift amount is corrected, in accordance with the obtained angle shift amount. It may have. According to such a configuration, the arithmetic processing is simplified.
[0033]
In the film thickness measuring device of the present invention, the measurement medium light emitted from the light emitting side optical system includes two or more wavelength components, and the light receiving side optical system includes the respective wavelength components. Two or more photoelectric conversion unit array means corresponding to the components, and wavelength-specific spectroscopy for separating the reflected light coming from the film thickness measurement point on the sample into each wavelength component and guiding the reflected light to each of the corresponding photoelectric conversion unit array means. Means may be included.
[0034]
In the film thickness measuring device of the present invention, the measurement medium light emitted from the light emitting side optical system contains two or more polarized light components, and the light receiving side optical system contains the polarized light components. Polarization-based spectroscopy for separating two or more photoelectric conversion unit array means corresponding to the components and reflected light coming from the film thickness measurement point on the sample into respective polarized light components and leading the polarized light components to the corresponding photoelectric conversion unit array means. Means may be included.
[0035]
In the film thickness measuring apparatus of the present invention, the light projecting optical system includes means for irradiating two or more kinds of measurement medium lights having different wavelength components or polarization components in a time-division manner, The means, by synchronizing with the irradiation timing of the measurement medium light, obtains received light amount data for each wavelength component or each polarized light component via each photoelectric conversion unit of the photoelectric conversion unit array means, Is also good.
[0036]
According to such a configuration, by providing two or more wavelengths and polarizations, the amount of detection information for each incident angle component is increased, and more accurate film thickness measurement becomes possible.
[0037]
In each of the above film thickness measuring apparatuses, a one-dimensional or two-dimensional CCD or PD array can be used as the photoelectric conversion unit array means. However, if a one-dimensional CCD is used, data processing can be facilitated and the cost of the apparatus can be reduced.
[0038]
Such a film thickness measuring apparatus of the present invention can be arranged in a production line of a product with a film forming process such as a semiconductor product or an FPD and used for in-line film thickness measurement of the product. The yield can be improved.
[0039]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, a preferred embodiment of a film thickness measuring apparatus according to the present invention will be described in detail with reference to the accompanying drawings.
[Basic principle of film thickness measuring device]
[0040]
Before describing the present invention, the basic principle of a film thickness measuring apparatus using a VAR method, which is a premise of the present invention, will be described.
[0041]
FIG. 1 is a configuration diagram showing the entire thickness measuring apparatus using the VAR method. As shown in FIG. 1, the film thickness measuring apparatus 1 includes a sensor head unit 2, an arithmetic processing unit 3, and an HMI (Human Machine Interface) unit 4 such as a monitor, a keyboard, and a mouse. In the drawing, reference numeral 5 denotes a substrate constituting a sample (for example, a semiconductor or an FPD), and reference numeral 5a denotes a thin film to be measured existing on the surface of the substrate 5.
[0042]
The optical configuration of the sensor head unit 2 is shown in FIG. As shown in the figure, a light source 21 whose light emission state is controlled by a light projecting unit control signal s1 and a thin film 5a of a substrate (sample) 5 by condensing light from the light source 21 are provided in the sensor head unit 2. And a one-dimensional CCD 23 in which reflected light of the measurement medium light coming from the film thickness measurement point is guided through a light receiving optical system (not shown). Reference numeral 23a denotes a photoelectric conversion element (pixel) constituting the one-dimensional CCD 23. The photographing operation of the one-dimensional CCD 23 is controlled by a one-dimensional CCD control signal s3. A series of received light amount data obtained from each pixel 23a of the one-dimensional CCD 23 is output to the outside as a one-dimensional CCD output signal s2.
[0043]
FIG. 3 shows an electrical configuration of the arithmetic processing unit 3. As shown in the figure, the arithmetic processing unit 3 includes a light emitting unit drive circuit 32 that generates and outputs a light emitting unit control signal s1, an AD conversion unit 33 that converts the one-dimensional CCD output signal s2 into a digital signal, A CCD drive circuit 34 for generating and outputting a one-dimensional CCD control signal s3, a ROM 35 storing various system programs, an input / output unit 36 functioning as an interface with a keyboard and mouse constituting the HMI unit, and an HMI unit A display unit 37 that functions as an interface with a display that performs processing, and a CPU 31 that performs overall control of the constituent elements 32 to 37 and executes a film thickness measurement calculation, an angle flapping calculation, and the like, which will be described later. .
[0044]
Next, the operation of the film thickness measuring device will be described. In FIG. 2, light emitted from a light source 21 is collected by a condenser lens 22, and is condensed and irradiated on a thin film 5a on a substrate 5 to be measured. The measurement point of the film thickness of the sample is located at a position where the incident light is focused. At this time, measurement medium light having a continuous incident angle component from the angle θ0 to the angle θ1 is incident on the film thickness measurement point of the sample.
[0045]
The incident measurement medium light is reflected by the sample. The reflected light from the sample thickness measurement point is guided to the light receiving surface of the one-dimensional CCD 23 via a light receiving optical system (not shown). On the light receiving surface of the one-dimensional CCD 23, there is formed a pixel row in which a large number of light receiving elements (pixels) 23a are arranged in one row. The received light amount data of each light receiving element 23a constituting the pixel row is serially output to the outside as a one-dimensional CCD output signal s2. Through this one-dimensional CCD output signal s2, a reflected light intensity distribution waveform corresponding to the incident angle (θ0 to θ1) is observed. Since the observed reflected light intensity distribution waveform is an amount corresponding to the reflectance distribution according to the incident angle, the arithmetic processing unit 3 can compare this with the theoretical reflectance distribution to determine the film thickness.
[0046]
The theoretical reflectance distribution can be obtained as follows. As shown in FIG. 5, when the measurement medium light (incident light Ii) is incident on the thin film 5a on the substrate 5, the measurement medium light is reflected on the front and back surfaces of the thin film 5a and interferes with each other. At this time, it is known that the theoretical expression of the reflectance R is represented by Expression 1.
(Equation 1)

[0047]
Here, R is the reflectance, d is the thickness of the thin film 5a, n is the refractive index of the thin film 5a, θ0 is the incident angle, λ is the wavelength of the incident light, r0 and r1 are the refractive index n0 of the substrate, and the refractive index of the thin film. n is a quantity related to the incident angle θ0.
[0048]
Therefore, if the refractive index n of the substrate 5, the refractive index n0 of the thin film 5a, and the wavelength λ of the incident light are known, Equation 1 becomes a function of the film thickness d and the incident angle θ0, and an arbitrary film thickness d , A reflectance distribution according to the incident angle θ0 is obtained.
[0049]
For example, when the material of the substrate 5 is Si and the material of the thin film 5a is SiO2, two cases are considered: a case where the thickness d of the thin film 5a is 1000 nm and a case where it is 1500 nm. At this time, the reflectance distribution curve according to the incident angle θ0 defined by Expression 1 is shown in the graph of FIG. That is, in the graph of FIG. 3, a is a reflectance distribution curve when the film thickness is 1000 nm, and b is a reflectance distribution curve when the film thickness is 1500 nm. As shown in this graph, if the film thickness changes, the reflectance distribution waveform also changes, so it is understood that the film thickness can be measured by knowing the reflectance distribution according to the incident angle. Will. In addition, if a large number of combinations of the reflectance with respect to the incident angle can be obtained, even if the refractive index of the thin film and the refractive index of the substrate other than the film thickness of the thin film are unknown, they can be obtained from Expression 1.
[0050]
In FIG. 5, Equation 1 is derived only for a single-layer film (see FIGS. 7 and 8), but a theoretical equation corresponding to a multilayer film can also be derived (see FIGS. 7 and 9). As a result, the thickness of the multilayer film can be measured.
[0051]
As a method for calculating the film thickness in the CPU 31 of the arithmetic processing unit 3, a curve fitting method can be used. The curve fitting method compares waveform data (table data) for each film thickness calculated in advance and stored as a table with measured light receiving amount data, and calculates the error (ie, the least error) from the received light amount data by the least square method. This is a method in which data with a small amount (hatched portion in FIG. 10) is extracted, and the thickness of the waveform data is used as the thickness of the thin film to be measured. As a method for calculating the film pressure, it is also possible to use a film thickness calculation method such as an extreme value search method.
[0052]
Details of the case where curve fitting is used as a method of calculating the film thickness will be described below. When the refractive index n and r0, r1 of the thin film to be measured are input from a keyboard or the like via the input / output unit 36, the CPU 31 calculates the reflectance value with respect to each value of the film thickness d and the incident angle θ0 according to Equation 1. And these are stored as a table in a memory in the CPU 31. An example of such a table is shown in FIG. As shown in the figure, according to this table, for example, the reflectance at the film thickness dx and the incident angle θp is Rdx (θp), and the reflectance at the film thickness dy and the incident angle θp + Δθ is Rdy (θp + Δθ). ).
[0053]
Thereafter, the curve fitting is executed according to the processing procedure shown in the flowchart of FIG. That is, first, the CPU 31 acquires digitized measurement data (reflectance measurement data) M (θ) from the A / D converter 33 (step 401). Needless to say, this measurement data M (θ) is obtained from the one-dimensional CCD 23. Then, following acquisition of the measurement data M (θ), the film thickness calculation processing is started.
[0054]
In this film thickness calculation processing, after initially setting the film thickness d to the minimum film thickness dx (Step 402), the theoretical data Rdx (θ) of the reflectance at the film thickness d = dx is obtained using the theoretical table of FIG. Square of difference from measurement data M (θ) [Rdx (θ) -M (θ)] ² From the incident angle range θp to θq in increments of Δθ, and the sum P (d) = ■ (Rdx (θ) −M (θ)) ² , A series of processes for storing in the memory is executed (step 403). In the process of calculating the sum of squares of the difference between the theoretical data and the measured data and storing the sum in a memory (step 403), while the value of the film thickness d is increased in steps of Δd (step 404), the film thickness d reaches its maximum value. The process is repeatedly executed until the film thickness dz is reached (step 405 NO) (step 403).
[0055]
When the calculation of the sum of squares is completed up to the maximum film thickness dz (step 405 YES), the minimum value is taken from the sum of squares P (dx) to P (dz) in the film thickness range dx to dz stored in the memory. The sum of squares P (dy) is extracted (step 406), and the film thickness dy at this time is determined as a measured film thickness (step 407).
[First Embodiment]
[0056]
Next, an embodiment of a film thickness measuring apparatus in which distance flutter is incorporated will be described in detail with reference to FIGS.
[0057]
FIG. 12 shows an optical configuration of the sensor head unit 2 in which a countermeasure against distance flapping is incorporated. This film thickness measuring apparatus is a light projecting side optical system (in the illustrated example, a light source 21) that irradiates a measurement medium light including various irradiation angle components (θ0 to θ1 in the illustrated example) to the thickness measuring point of the sample. And a condenser lens 22), and a photoelectric conversion unit array means (corresponding to a one-dimensional CCD 23 in the illustrated example) in which a large number of photoelectric conversion units are arranged in an array on a light receiving surface. A light receiving side optical system including a lens (corresponding to lens 24 in the illustrated example) for guiding the reflected light of the measurement medium light for each irradiation angle coming from the thickness measurement point onto the light receiving surface of the photoelectric conversion unit array means, and a photoelectric conversion unit Calculating means for calculating a film thickness to be measured based on a series of received light amount data obtained from each photoelectric conversion unit of the array means (corresponding to the processing unit 3 in the illustrated example). In addition, the distance between the lens (corresponding to the lens 24 in the illustrated example) included in the light receiving side optical system and the light receiving surface of the photoelectric conversion unit array means (corresponding to the one-dimensional CCD 23 in the illustrated example) is determined by the focal length of the lens ( f) is set so as to substantially match.
[0058]
The basic operation for measuring the film thickness is as follows. The measurement medium light emitted from the light source 21 is condensed and irradiated on the thin film 5a on the substrate 5 to be measured by the action of the condenser lens 22. The measurement point of the film thickness of the sample is located at a position where the incident light is focused. At this time, measurement medium light having a continuous incident angle component in the range of θ0 to θ1 is incident on the sample.
The measurement medium light incident via the condenser lens 22 is reflected by the sample. The reflected light of the measurement medium light coming from the sample thickness measurement point is guided to the light receiving surface of the one-dimensional CCD 23 by the action of the light receiving lens 24. As a result, the one-dimensional CCD 23 sends out a one-dimensional CCD output signal s2 corresponding to a serial arrangement of the light receiving amount data of each light receiving element (pixel) 23a. Based on the one-dimensional CCD output signal s2, a reflected light intensity distribution corresponding to each of the incident angles (θ0 to θ1) is observed. The reflected light intensity distribution corresponding to each of the incident angles observed at this time is an amount corresponding to the reflectance distribution corresponding to the incident angle, and is compared with the theoretical reflectance distribution by the arithmetic processing unit 3 to obtain the objective. Is required. Note that the processing for measuring the film thickness in the arithmetic processing unit 3 is as described above with reference to FIGS.
[0059]
The major features of the sensor head unit 2 shown in FIG. 12 are that (1) the light receiving lens 24 and the light receiving surface of the one-dimensional CCD 23 are parallel, and (2) the light receiving lens 24 and the light receiving surface of the one-dimensional CCD 23 The distance is substantially equal to the focal length f of the light receiving lens 24. In other words, if the sample (substrate 5) side is defined as the front side and the one-dimensional CCD 23 side is defined as the rear side with respect to the light receiving lens 24, the position of the light receiving surface of the one-dimensional CCD 23 is substantially the rear focal position of the light receiving lens 24. Can be expressed. According to such an arrangement, it is possible to realize an optical system that is not affected by distance flutter.
[0060]
FIG. 13 is an explanatory diagram of the operation of the distance flapping countermeasure. Now, let us suppose that, of the light beams from the condenser lens 22, the two light beams whose incident angles are the most separated are defined as L1 and L2. The substrate at the reference height Href is denoted by reference numeral 5 (the thin film is 5a), and the substrate when the substrate is lowered due to the flutter is denoted by reference numeral 5 '(the thin film is 5a'). Further, the light beam L1 reflected by the substrate 5 is denoted by symbol L11, and the light beam reflected by the substrate 5 'is denoted by L12. The reflected light beam L2 reflected by the substrate 5 is denoted by L21, and the reflected light beam reflected by the substrate 5 'is denoted by L22. Further, the incident point of the reflected light beams L11 and L12 on the light receiving surface of the one-dimensional CCD 23 is P1, and the incident point of the reflected light beams L21 and L22 is P2. Then, as is clear from the drawing, as for the same incident light beams L1 and L2, even if the substrate moves up and down due to the distance flutter, the corresponding reflected lights (L11, L12) and (L21, L22) are one-dimensional. It will be understood that the light enters the same incident points P1 and P2 on the light receiving surface of the CCD 23.
[0061]
The effect of eliminating the influence of the distance flutter described above is based on the following principle. According to Snell's law, it is known that the reflection angle is uniquely determined if the angle of the incident light and the normal direction of the sample are determined. As shown in FIG. 14, considering that the light beam L1 is incident on the sample at the incident angle θ, even if the vertical fluctuation ΔL of the sample occurs due to the distance flapping at this time, the angle θ of the incident light beam L1 Since the directions of the normal lines L01, L02, L03, and L04 of the sample do not change, it can be seen that the reflection angle does not change. However, when the fluttering of the distance occurs, the reflecting point moves as indicated by P11, P12, P13, and P14 along with the parallel movement of the reflecting surface, so that the reflected light beams L11, L12, L13, and L14 also move in parallel.
Here, three incident light beams L1, L3, and L2 having different incident angles θ1, θ2, and θ3 are assumed. At this time, when the distance flapping (the substrate 5 at the reference height, the substrate 5 'at the descending position, and the substrate 5 "at the ascending position) occurs, as shown in FIGS. 15 to 17, the incident light beams L1, L2, L3 , Three parallel reflected light beams (L10, L11, L12), (L20, L21, L22), and (L30, L31, L32) are generated.
[0062]
As described above, since the distance between the light receiving lens 24 and the light receiving surface of the one-dimensional CCD 23 is almost equal to the focal length f of the light receiving lens 24, as shown in FIG. The light beams (L10, L11, L12), (L20, L21, L22) and (L30, L31, L32) converge on three points P1, P2, and P3 on the light receiving surface of the one-dimensional CCD 23. In other words, even if the distance fluctuates, the reflected light intensity distribution according to the incident angles (θ1 to θ3) of the waveform observed via the output signal s2 of the one-dimensional CCD 23 does not change. It will be appreciated that it is possible.
However, since the light receiving lens 24 usually has aberration, it is assumed that the light receiving lens 24 and the light receiving surface 23a of the one-dimensional CCD 23 are parallel and the distance between the two completely matches the focal length f of the light receiving lens 24. However, as shown in FIG. 19, the focal point of each set of parallel rays does not converge to one point on the light receiving surface 23b in a strict sense. The expression "~ substantially equal to the focal length f ~" is due to this fact.
[Second embodiment]
[0063]
Next, an embodiment of a film thickness measuring device incorporating angle flapping will be described in detail with reference to FIGS.
[0064]
FIG. 20 shows an optical configuration of the sensor head unit 2 in which a countermeasure against angle flapping is incorporated. The film thickness measuring device 2 is a light emitting side optical system (in the illustrated example, a light source in the illustrated example) that irradiates a measurement medium light including various irradiation angle components (θ0 to θ1 in the illustrated example) to the thickness measuring point of the sample. 21 and a condenser lens 22), and a photoelectric conversion unit array means (a one-dimensional CCD 23 in the illustrated example) in which a large number of photoelectric conversion units are arranged in an array on a light receiving surface. A light receiving optical system (not shown) for guiding the reflected light of the measurement medium light for each irradiation angle coming from the film thickness measurement point onto the light receiving surface of the photoelectric conversion unit array means, and from each photoelectric conversion unit of the photoelectric conversion unit array means Calculating means for calculating the film thickness to be measured based on the obtained series of received light amount data (the processing unit 3 corresponds in the illustrated example).
[0065]
Further, the light projecting side optical system includes characterization means (characterized as a slit 25a of the slit plate 25 functioning as an edge shaping means in the illustrated example) for performing characterization corresponding to the reference optical axis with respect to the measurement medium light. In addition, the light receiving side optical system receives a reflected light of the measurement medium light coming from the film thickness measurement point of the sample and detects a characteristic corresponding to the reference optical axis included in the reflected light. A conversion means (corresponding to the one-dimensional CCD 23 itself in the illustrated example) is included. Further, the calculating means includes an error component due to angular flutter of the sample included in a series of received light amount data obtained from each photoelectric conversion unit based on a feature corresponding to the reference optical axis detected by the second photoelectric conversion means. Is included.
[0066]
The basic operation for measuring the film thickness is as follows. The measurement medium light emitted from the light source 21 is condensed and irradiated on the thin film 5a on the substrate 5 to be measured by the action of the condenser lens 22. The measurement point of the film thickness of the sample is located at a position where the incident light is focused. At this time, measurement medium light having a continuous incident angle component in the range of θ0 to θ1 is incident on the sample.
The measurement medium light incident via the condenser lens 22 is reflected by the sample. The reflected light of the measurement medium light coming from the sample thickness measurement point is guided to the light receiving surface of the one-dimensional CCD 23 by the operation of a light receiving optical system (not shown) (equivalent to the light receiving lens 24 shown in FIG. 12). As a result, the one-dimensional CCD 23 sends out a one-dimensional CCD output signal s2 corresponding to a serial arrangement of the light receiving amount data of each light receiving element (pixel) 23a. Based on the one-dimensional CCD output signal s2, a reflected light intensity distribution corresponding to each of the incident angles (θ0 to θ1) is observed. The reflected light intensity distribution corresponding to each of the incident angles observed at this time is an amount corresponding to the reflectance distribution corresponding to the incident angle, and is compared with the theoretical reflectance distribution by the arithmetic processing unit 3 to obtain the objective. Is required. Note that the processing for measuring the film thickness in the arithmetic processing unit 3 is as described above with reference to FIGS.
[0067]
The major feature of the sensor head unit 2 shown in FIG. 20 is that (1) a slit plate 25 (slit 25a) functioning as an edge shaping unit is arranged on the incident side of the condenser lens 22 so that the cross section of the incident light is edge-shaped. Then, the reference optical axes Lref0. Lref1 is characterized, and (2) the reaching points Preref0. Lref1 'of the reference optical axes Lref0' and Lref1 'of the reflected light. Pref1 is detected via the one-dimensional CCD 23, and (3) an error component due to angular flutter of the sample included in the output signal s2 from the one-dimensional CCD 23 is corrected based on the detected incident points Pref0 and Pref1. That is in.
[0068]
That is, as shown in FIG. 21A, a slit 25a is formed at the center of the slit plate 25 as an edge shaping means arranged in the light projecting optical system, and the slit 25a extends in the longitudinal direction of the slit 25a. The incident angle range (θ0 to θ1) is determined by the both edges, in other words, two reference optical axes Lref0 and Lref1 are characterized in the measurement medium light. These two reference optical axes Lref0 and Lref1 are reflected by the sample to become the reference optical axes Lref0 'and Lref1' of the reflected light, and are incident on the pixel positions Preref0 and Preref1 on the light receiving surface of the one-dimensional CCD 23. The pixel coordinates Rref0 and Rref1 of these incident points can be easily detected by binarizing the output signal s2 of the one-dimensional CCD 23 with, for example, a threshold value corresponding to a dark level. Therefore, both ends (θ0 and θ1) of the incident angle range are observed through the pixels 23a of the one-dimensional CCD 23, and the deviation of the pixel position from the reference position is detected, so that the angular flutter is detected and there is no angular flutter. You can return to the state. Here, the reference position is a pixel position measured in a state where there is no angular flapping, as indicated by Pref0 and Pref1 in FIG.
[0069]
In this example, the edge shaping means (slit plate 25) is arranged on the light source 21 side near the condenser lens 22, but the same effect can be obtained by disposing it on the substrate 5 side near the lens 22. Can be.
[0070]
The method for detecting and correcting the angle flutter will be described in more detail. As described above, when an angle flutter occurs in this type of film thickness measuring apparatus, both ends in the pixel column direction of the waveform observed via the one-dimensional CCD deviate from the reference position. In this state, waveform processing cannot be performed normally by the arithmetic processing unit, and it is necessary to correct this before performing arithmetic processing for film thickness measurement.
[0071]
According to the earnest research by the present inventors, it has been found that there is a causal relationship shown in FIG. 22 between the angular flapping and the accompanying fluctuation of the measurement condition. When the angular flapping occurs, the inclination of the sample is also inclined by Δθ, and as a result, the normal line of the sample changes from L0 to L0 ′. Then, the incident angle is also shifted by Δθ, and the incident angle range is θ0 + Δθ to θ1 + Δθ. If the incident angle shifts by Δθ, the light intensity distribution waveform observed by the one-dimensional CCD 23 moves to a pixel position (Px0 to Px1) shifted by 2Δθ with respect to the reference position (Pref0 to Pref1). In the CCD, a light intensity distribution waveform is observed at pixel positions (Px0 to Px1) in the incident angle range θ0 + 2Δθ to θ1 + 2Δθ. Therefore, by observing how many times the waveform deviates from the reference position from the waveform observed by the one-dimensional CCD 23, it is possible to measure how many times the sample is inclined with respect to the reference position (Href). . For example, if the observed waveform deviates from the reference position by 2Δθ, the sample is tilted by Δθ.
[0072]
Next, in order to return the detected waveform to a state in which there is no angle flapping, the arithmetic processing unit 3 may perform an operation of returning the range of incident angles of the observed waveform by Δθ in accordance with the detected deviation angle 2Δθ. For example, the waveform observed at the pixel positions (Px0 to Px1) in the incident angle range θ0 + 2Δθ to θ1 + 2Δθ is returned to the incident angle range θ0 + Δθ to θ1 + Δθ (Pref0 to Pref1). With this method, the light intensity distribution waveform observed by the one-dimensional CCD can be returned to a state without angular flutter. As a result, it is possible to measure the film thickness normally even if there is angle flutter.
[0073]
FIG. 23 is a flowchart showing the angle correction processing executed by the arithmetic processing unit 3. As shown in the figure, the arithmetic processing unit 3 first acquires the measurement data M (θ) from the A / D conversion unit 33 (step 2301), and then, based on the measurement data M (θ), Is calculated (step 2302), and finally, the angle correction is completed by executing a process of returning the observed incident angle range by Δθ. Thereafter, by executing the processing shown in FIG. 4 earlier, accurate film thickness measurement becomes possible.
[0074]
In the present embodiment, the slit plate 25 is used as an edge shaping means for characterizing the reference optical axis. Instead, an aperture plate 25 'shown in FIG. Other edge shaping means such as a knife edge plate 25 "shown in FIG. 1) may be employed. In addition, 25a 'is an aperture, and 25a" is a knife edge.
[Third embodiment]
[0075]
FIG. 24 shows an optical configuration of the sensor head unit in which measures against both the distance flutter and the angle flutter are incorporated. As shown in the figure, the film thickness measuring device 2 is a light emitting side optical system (in the figure, a light source 21) that irradiates a measurement medium light including various irradiation angle components to a film thickness measurement point of a sample. , A collimator lens 26 and a condenser lens 22), and photoelectric conversion unit array means (a one-dimensional CCD 23 in the figure is equivalent) in which a large number of photoelectric conversion units are arranged in an array on a light receiving surface. A light receiving side optical system including a lens (in the figure, a lens 24 is equivalent) for guiding the reflected light of the measurement medium light at each irradiation angle coming from the sample film thickness measurement point onto the light receiving surface of the photoelectric conversion unit array means; Calculating means for calculating a film thickness to be measured based on a series of received light amount data obtained from each photoelectric conversion unit of the photoelectric conversion unit array means (in the figure, the processing unit 3 is equivalent); . In addition, the distance between the lens (lens 24) included in the light receiving side optical system and the light receiving surface of the photoelectric conversion unit array means (one-dimensional CCD 23) is set to substantially match the focal length (f) of the lens. ing.
[0076]
In addition, the light projecting side optical system includes characterization means (corresponding to a slit plate 25 functioning as an edge shaping means in the figure) for performing characterization corresponding to the reference optical axis with respect to the measurement medium light, The light-receiving side optical system includes a second photoelectric conversion means (in the figure, the second photoelectric conversion means) for receiving the reflected light of the measurement medium light coming from the sample film thickness measurement point and detecting a feature corresponding to the reference optical axis included therein. The one-dimensional CCD 23 itself is included), and the arithmetic means includes a series of light receiving units obtained from each photoelectric conversion unit based on a feature corresponding to the reference optical axis detected by the second photoelectric conversion unit. Included is a received light amount data correcting means for correcting an error component caused by angle flutter of the sample included in the amount data.
[0077]
The basic operation for measuring the film thickness is as follows. The measurement medium light emitted from the light source 21 is collimated by the collimator lens 26, passes through the slit 25 a of the slit plate 25, and acts on the thin film 5 a on the substrate 5 to be measured by the condensing lens 22. Is condensed and irradiated. The measurement point of the film thickness of the sample is located at a position where the incident light is focused. At this time, measurement medium light having a continuous incident angle component in the range of θ0 to θ1 is incident on the sample.
The measurement medium light incident via the condenser lens 22 is reflected by the sample. The reflected light of the measurement medium light coming from the sample thickness measurement point is guided to the light receiving surface of the one-dimensional CCD 23 by the action of the light receiving lens 24. As a result, the one-dimensional CCD 23 sends out a one-dimensional CCD output signal s2 corresponding to a serial arrangement of the light receiving amount data of each light receiving element (pixel) 23a. Based on the one-dimensional CCD output signal s2, a reflected light intensity distribution corresponding to each of the incident angles (θ0 to θ1) is observed. The reflected light intensity distribution corresponding to each of the incident angles observed at this time is an amount corresponding to the reflectance distribution corresponding to the incident angle, and is compared with the theoretical reflectance distribution by the arithmetic processing unit 3 to obtain the objective. Is required. Note that the processing for measuring the film thickness in the arithmetic processing unit 3 is as described above with reference to FIGS.
[0078]
The major features of the sensor head unit 2 shown in FIG. 24 are (1) the light receiving lens 24 and the light receiving surface of the one-dimensional CCD 23 are parallel, and (2) the distance between the light receiving lens 24 and the light receiving surface of the one-dimensional CCD 23. (3) A slit plate 25 (slit 25 a) functioning as an edge shaping means is disposed on the incident side of the condenser lens 22, so that the cross section of the incident light is Are edge-shaped, and the reference optical axes Lref0. Lref1 is characterized; (4) the arrival points of the reference optical axes Lref0 'and Lref1' of the reflected light are detected via the one-dimensional CCD 23; and (5) the reference optical axis of the detected reflected light. The error component due to the angular flutter of the sample included in the output signal s2 from the one-dimensional CCD 23 is corrected based on the incidence points of Lref0 'and Lref1' and the reference incidence points Pref0 and Pref1.
[0079]
FIGS. 25 to 27 show operation explanatory diagrams (parts 1 to 3) in the case where the measures against distance and angle flapping are used together. As described above with reference to FIGS. 14 to 18, (1) the light receiving lens 24 included in the light receiving optical system and the light receiving surface of the one-dimensional CCD 23 are parallel, and (2) the light receiving lens 24 The parallel light incident on the light-receiving lens 24 is on the light-receiving surface of the one-dimensional CCD 23 as long as the condition that the distance between the light-receiving lens 24 and the light-receiving surface of the one-dimensional CCD 23 is substantially equal to the focal length f of the light-receiving lens 24 is satisfied. Light is condensed and incident on one point. Here, as is apparent with reference to FIGS. 25 to 27, as long as the shift angle Δθ is constant, the substrate at each height position (elevated position, reference position, lowered position) regardless of the presence or absence of the distance flapping. The three sets of reflected light (L10, L11, L12), (L20, L21, L22) and (L30, L31, L33) from 5 ″, 5, 5 ′ maintain a parallel relationship with each other in any set. Therefore, even in this embodiment, the reflected lights (L10, L11, L12), (L20, L21, L22), (L30, L31, L33) with respect to the incident lights L1, L2, L3 are primary. Since the light is always condensed at three points P1, P2, and P3 on the light receiving surface of the original CCD 23, it is possible to realize an optical system that is not affected by distance flutter.
[0080]
On the other hand, in FIGS. 25 to 27, when the deviation angle Δθ fluctuates, as described above with reference to FIG. 22, the incident point of the reflected light incident on the light receiving surface of the one-dimensional CCD 23 is It moves by a predetermined angle (2Δθ) in the pixel column direction in accordance with the deviation angle (Δθ) from the line L0 to the normal L0 ′. As described in the previous example, the slit 25a of the slit plate 25 arranged in the light projecting optical system functions as an edge shaping unit and defines an incident angle range (θ0 to θ1). In other words, The reference optical axes Lref0 and Lref1 are characterized in the incident light. The incident points Px0 and Px1 of the reflected light of these reference optical axes Lref0 and Lref1 to the one-dimensional CCD 23 can be detected from the output signal s2 of the one-dimensional CCD 23. The arithmetic processing unit 3 obtains the actual deviation angle Δθ of the sample based on the deviation angle 2θ between the detected incident points Px0 and Px1 and the reference incident points Pref0 and Pref1, and converts the observed waveform to a state without angular flutter. Execute the process to return to.
[0081]
Therefore, according to this embodiment, it is possible to measure the thickness of the single-layer thin film or the multilayer thin film of the sample with high accuracy regardless of the distance flutter and the angle flutter of the sample. In particular, according to this embodiment, the correction for the distance flutter has already been completed at the time of observation by the one-dimensional CCD 23, and the change in the observed waveform due to the distance flutter has been eliminated. Therefore, the cause of the change in the observed waveform is only the angle flutter, and the arithmetic processing unit only needs to execute the process of correcting the error component caused by the angle flutter. As described above, since the distance flutter and the angle flutter are corrected in a process independent of each other, the change in the observed waveform caused by these two types of flutter is corrected by the arithmetic processing unit in the same process. There is no inconvenience that the correction processes compete with each other and the calculation process does not converge.
As a result, according to this embodiment, the installation condition (distance / angle flapping) is reduced, and the autofocus function is not required. At the same time, downsizing of the device and high-speed arithmetic processing are realized, and it is possible to provide a device suitable for in-line measurement.
[Fourth embodiment]
[0082]
FIG. 28 shows an optical configuration of a sensor head in which two types of light sources and a measure against distance and angle flutter are taken. As shown in the figure, the optical system in this embodiment includes two types of light sources 21A and 21B. The light from these light sources 21A and 21B is coaxially aligned via a half mirror 21C, and then shaped into parallel light via a collimator lens 26. A slit 25a and a polarizer 27 are arranged in the path of the parallel light, whereby the incident angle range (θ0 to θ1) and the polarization direction of the light are determined. The light that has passed through the polarizer 27 is collected by the condenser lens 22, and is condensed and irradiated on the thin film 5a on the substrate 5 to be measured. The measurement point of the film thickness of the sample is located substantially at the condensing position of the incident light. Here, the polarizer 27 is set so that the polarization state of the incident light is inclined by 45 degrees with respect to the incident plane. Thus, the measurement medium light having the continuous incident angles θ0 to θ1 and having the polarization direction determined by the polarizer 27 is incident on the sample.
[0083]
The incident measurement medium light is reflected by the sample, passes through the light receiving lens 24, and is separated into S-polarized light and P-polarized light through the polarizing beam splitter 28. The separated reflected light is observed as a reflected light intensity distribution of S-polarized light and P-polarized light corresponding to the incident angles (θ0 to θ1) through a series of pixels of the two one-dimensional CCDs 23A and 23B. The observed reflection intensity distribution waveform is an amount corresponding to the reflectance distribution according to the incident angle, and the arithmetic processing unit 3 calculates the film thickness in comparison with the theoretical reflectance distribution.
[0084]
Note that the distance between the light receiving lens 24 and each of the one-dimensional CCDs 23A and 23B is substantially equal to the focal length f of the light receiving lens 24, and the arithmetic processing unit 3 executes data correction processing corresponding to the above-mentioned angle flapping. Needless to say.
[0085]
In the above-described theoretical formula of reflectivity (formula 1), the wavelength is also treated as a variable (λ). Therefore, in the fourth embodiment in which two light sources 21A and 21B are used. By making the wavelengths of the light emitted from the two light sources different from each other (by increasing the number of wavelengths used), the information on the reflectivity that can be measured increases, so that the measurement accuracy can be improved. .
(Equation 2)

[0086]
Since the theoretical equation (Equation 1) assumes a single wavelength, when using only one light source and using a plurality of wavelengths as described above, the light receiving unit uses a plurality of spectroscopic devices to generate a plurality of wavelengths. It is necessary to separate them into wavelengths and measure the reflectance for each wavelength. In the fourth embodiment, since the two light sources 21A and 21B are used, the light sources are operated at different wavelengths in a time-division manner (for example, the light sources are evenly and alternately illuminated), and the reflectance for each wavelength is changed. Can be adopted as a measurement method of alternately measuring.
[0087]
The wavelength of the light source to be used can be arbitrarily changed, and when it is desired to measure a thinner film, a light source having a shorter wavelength can be used depending on the application. Specifically, a semiconductor laser of ultraviolet light, visible light, infrared light, or the like can be used. Further, as a result of using a light source having a plurality of wavelengths, it is possible to obtain a spectral characteristic of a refractive index.
[0088]
Further, since a normal lens has chromatic aberration, there is a problem in that, when light is condensed and illuminated on a sample by a condensing lens, the converging point differs depending on the wavelength. However, if a lens with corrected chromatic aberration, such as an achromatic lens, is used as a condenser lens used in the apparatus, it is possible to cope with light sources of various wavelengths simply by replacing the light source.
[0089]
In this embodiment, as shown in FIG. 28, by using the polarizer 27, the polarization state of the light incident on the sample is set to be inclined by 45 degrees with respect to the incident plane. Light having such a polarization characteristic is incident on the sample, is separated into S-polarized light and P-polarized light by the polarization beam splitter 28 of the light receiving section, and the incident angle distribution of the reflectance is measured. As described above, the incident angle distribution of the reflectance is represented by Expression 1. However, the values of r0 and r1 included in Equation 1 are
[Outside 1]

When,
[Outside 2]

Then, a different theoretical expression is obtained as shown in Expression 2.
[Equation 3]

[0090]
Therefore, by measuring both the S-polarized light and the P-polarized light, the information on the reflectivity that can be measured increases, so that the measurement accuracy can be improved.
[Fifth Embodiment]
[0091]
FIG. 29 shows an optical configuration of the sensor head unit in which the reference optical axis is set by a dedicated light source. In the second to fourth embodiments described above, the reference axis characterization of the incident light necessary for the angle correction is performed by using an edge shaping means such as a slit. In, the reference axis is characterized by adding a dedicated light source separately. In the second to fourth embodiments described above, the position of the reference optical axis required for angle correction is detected by using the film thickness measuring CCD itself. In the method, the position of the reference optical axis is detected by using a dedicated photoelectric sensor (PSD) separately.
[0092]
That is, in FIG. 29, the light emitted from the light source 21 passes through the beam splitter 29b, is collected by the condenser lens 22, and is condensed and irradiated on the thin film 5a on the substrate 5 to be measured. The sample film thickness measurement point is located at a substantially condensing position of incident light (measurement medium light). The incident light is reflected by the sample, passes through the light receiving lens 24, passes through the polarization beam splitter 29c, and is observed as a reflected light intensity distribution corresponding to the incident angle at each pixel 23a of the one-dimensional CCD 23. The observed waveform is an amount corresponding to the reflectance distribution according to the incident angle, and the arithmetic processing unit 3 calculates the film thickness in comparison with the theoretical reflectance distribution. The light receiving lens 24 is arranged in parallel with the one-dimensional CCD 23 and substantially at the rear focal position. With this arrangement, it is possible to realize an optical system that is not affected by distance flutter.
[0093]
On the other hand, the thin beam-like light emitted from the light source 29a is turned by the beam splitter 29b, passes through the condenser lens 22, and irradiates the thin film 5a on the substrate 5 to be measured as a reference optical axis. You. The incident light corresponding to the reference optical axis is reflected by the sample, passes through the light receiving lens, is changed in direction by the polarization polarizing beam splitter 29c, and is irradiated on the PSD (Position Sensing Device) 29d. At this time, the light from the light source 29a for setting the optical axis has a polarization state different from that of the light from the light source 21 for film thickness measurement, and is separated from the light from the light source 21 by the polarization beam splitter 29c. By detecting a deviation from the reference position Pref with respect to the observation position Px in the PSD 29d, it is possible to detect angular flutter and return to a state without angular flutter. Here, the reference position Pref is, as shown in FIG. 30, an irradiation position of the reference optical axis measured in a state where there is no angular flapping. With this arrangement, it is possible to realize an optical system that is not affected by angle flutter. Note that the calculation processing required for correcting the angle flutter has been described in detail in the second embodiment, and will not be described.
[0094]
In the fifth embodiment, the light source 21 and the light source 29a are operated in a time-division manner, and the intensity distribution waveform is acquired in synchronization with the light source 21 and the light source 29a. It is also possible to do. Alternatively, the wavelengths of the light source 21 and the light source 29a are different from each other, and a dichroic mirror or the like having a wavelength separating function is used for the beam splitters 29b and 29c so that the film thickness measurement based on the measurement medium light of the two wavelengths is performed. You may.
[0095]
As is clear from the above-described first to fifth embodiments, the film thickness measuring apparatus of the present invention irradiates a film thickness measuring point of a sample with a measuring medium light including various irradiation angle components. A light-side optical system, and a photoelectric conversion unit array means in which a large number of photoelectric conversion units are arranged in an array on a light-receiving surface; and a measurement medium light for each irradiation angle coming from a film thickness measurement point of the sample. A light receiving side optical system including a lens for guiding the reflected light on the light receiving surface of the photoelectric conversion unit array means, and a measurement target based on a series of received light amount data obtained from each photoelectric conversion unit of the photoelectric conversion unit array means. Calculating means for determining the thickness of the film so that the distance between the lens included in the light receiving side optical system and the light receiving surface of the photoelectric conversion unit array means is substantially equal to the focal length of the lens. Is set Than it is.
[0096]
Therefore, even if the distance between the sample and the light receiving optical system fluctuates, the reflected optical axis corresponding to the same incident optical axis simply moves in parallel, so that the lens included in the light receiving side optical system and the photoelectric conversion unit array If the distance from the light receiving surface of the means is set to be substantially equal to the focal length of the lens, the reflected light corresponding to each incident optical axis always reaches the determined array position on the photoelectric conversion unit array means. Therefore, the fluctuation of the intensity distribution waveform of the reflected light due to the distance flutter is suppressed, and the resistance to the distance flutter is improved.
[0097]
Further, the light projecting side optical system includes characterization means for performing characterization corresponding to a reference optical axis with respect to the measurement medium light, and the light receiving side optical system includes a film thickness of the sample. A second photoelectric conversion unit for receiving reflected light of the measurement medium light coming from the measurement point and detecting a feature corresponding to a reference optical axis included therein is included, and the arithmetic unit includes the second photoelectric conversion unit. Based on a feature corresponding to the reference optical axis detected by the second photoelectric conversion means, correction of an amount of received light data for correcting an error component due to angular flutter of a sample included in a series of received light amount data obtained from each photoelectric conversion unit. Means are included.
[0098]
Therefore, when the incident angle with respect to the sample changes, the angle of the sample included in a series of received light amount data obtained from each photoelectric conversion unit based on a feature corresponding to the reference optical axis detected by the second photoelectric conversion unit. Since the error component due to the flapping is corrected and the fluctuation of the intensity distribution waveform of the reflected light due to the angular flapping is suppressed, the resistance to the angular flapping is also improved. Moreover, in such a film thickness measuring device, the correction function for the distance flutter and the correction function for the angle flutter operate independently without affecting each other. There is an advantage that the arithmetic processing is not complicated. That is, the correction function relating to the distance flutter is realized by the configuration of the optical system ("substantially coincides with the focal length"), while the correction function relating to the angular flutter is realized by the configuration of the electric system ("error correction calculation of a series of received light amount data"). Therefore, it is possible to avoid such a disadvantage that the two influence each other and it takes a long time to converge the operation result.
[0099]
In the above-described first to fifth embodiments, a method in which the light projecting lens (condensing lens 22) and the light receiving lens (light receiving lens 24) are separated and the light is obliquely incident on the sample is employed. As in the conventional example described with reference to FIG. 33 (Japanese Unexamined Patent Publication No. 9-196630), the distance / angle flapping occurs even in a coaxial type using the same lens for projecting light and receiving light (center incident angle is 0 °). A strong optical system can be realized. FIG. 31 shows an example of the sensor head according to the present invention configured as a coaxial type.
[0100]
However, in the case of the coaxial type, the numerical aperture N.sub. A. Is 0.45 or more, the one-dimensional CCD 23 cannot be arranged at the rear focal length of the light receiving lens 22A as shown in FIG. Therefore, in the coaxial type, the numerical aperture N. A. Is set to 0.45 or less (the incident angle range is 27 ° or less), it is difficult to realize an optical system that is strong against distance / angle flapping. Equation 1, which is a theoretical expression of reflectivity, is a function of the incident angle, and it is desirable that the range of the incident angle be wide. The oblique incidence type is separated from the light emitting type and the light receiving type, and thus has no advantage as described above, and has an advantage that the incident angle range can be widened.
[0101]
【The invention's effect】
As apparent from the above description, according to the present invention, it is possible to provide a film thickness measuring apparatus using the VAR method, which is small, capable of high-speed arithmetic processing, and has resistance to distance and angle flapping. This type of film thickness measuring apparatus can be applied to in-line measurement in, for example, a semiconductor manufacturing process or an FPD manufacturing process.
[Brief description of the drawings]
FIG. 1 is a diagram showing an entire film thickness measuring apparatus using a VAR direction.
FIG. 2 is a diagram illustrating an optical configuration of a sensor head unit.
FIG. 3 is a diagram illustrating an electrical configuration of an arithmetic processing unit.
FIG. 4 is a flowchart illustrating a film thickness measurement process.
FIG. 5 is an explanatory diagram showing a relationship between a thin film, incident light, and reflected light.
FIG. 6 is a graph showing a reflectance distribution curve according to an incident angle.
FIG. 7 is an explanatory diagram showing a relationship between a multilayer thin film and a characteristic coefficient of each layer.
FIG. 8 is an explanatory diagram showing optical characteristics of a single-layer film.
FIG. 9 is an explanatory diagram showing optical characteristics of a multilayer film.
FIG. 10 is an explanatory diagram showing adoption data in the least squares method.
FIG. 11 is a diagram showing the contents of a table stored in a memory of a calculation unit.
FIG. 12 is a diagram showing an optical configuration of a sensor head unit in which a measure against distance flapping is incorporated.
FIG. 13 is an explanatory diagram of an action of a distance flapping countermeasure.
FIG. 14 is a diagram illustrating the principle of measures against distance flapping.
FIG. 15 is an explanatory diagram (part 1) illustrating the action of distance flapping countermeasures for each incident optical path.
FIG. 16 is an explanatory diagram (part 2) illustrating the action of the distance flapping countermeasure for each incident optical path.
FIG. 17 is an explanatory view (No. 3) illustrating the action of distance flapping countermeasures for each incident optical path.
FIG. 18 is an explanatory diagram showing the action of distance flapping countermeasures on the rear stage of the light receiving lens.
FIG. 19 is an explanatory diagram showing an enlarged light condensing state near a one-dimensional CCD light receiving surface.
FIG. 20 is a diagram illustrating an optical configuration of a sensor head unit in which measures against angle flapping are incorporated.
FIG. 21 is an explanatory diagram of an edge shaping unit.
FIG. 22 is an operation explanatory view of a measure against angle flapping.
FIG. 23 is a flowchart showing an angle correction process.
FIG. 24 is a diagram showing a configuration of a sensor head unit in which measures against both distance flutter and angle flutter are incorporated.
FIG. 25 is an operation explanatory view (part 1) in a case where measures against distance and angle flapping are used together.
FIG. 26 is an operation explanatory view (part 2) in a case where a countermeasure against distance and angle flapping is used together.
FIG. 27 is an operation explanatory view (part 3) in a case where a countermeasure against distance and angle flapping is used together;
FIG. 28 is a diagram showing an optical configuration of a sensor head in which two light sources and a countermeasure against distance and angle flapping are taken.
FIG. 29 is a diagram showing an optical configuration when a reference optical axis is set by a dedicated light source.
FIG. 30 is an explanatory diagram of the operation when the reference optical axis is set by a dedicated light source.
FIG. 31 is an explanatory diagram (part 1) of a sensor head configured as a coaxial type.
FIG. 32 is an explanatory view (part 2) of a sensor head configured as a coaxial type.
FIG. 33 is an explanatory view of a conventional film thickness measuring device using the VAR method.
FIG. 34 is an explanatory diagram of distance flapping.
FIG. 35 is an explanatory diagram of an angle flutter.
[Explanation of symbols]
1 Film thickness measuring device
2 Sensor head
3 arithmetic processing unit
4 Arithmetic processing unit
5 Substrate
5a Thin film
21, 21A, 21B Light source
21C half mirror
22 Condensing lens
23,23A, 23B One-dimensional CCD
23a pixel
23b Light receiving surface
24 Receiving lens
25 slit plate
25a slit
25 'aperture plate
25a 'aperture
25 "knife edge plate
25a "knife edge
26 Collimator lens
27 Polarizer
28 Polarizing beam splitter
29a Light source for reference optical axis
29b Polarizing beam splitter
29c polarizing beam splitter
29d PSD
Reference position of Href sample
Lref0, Lref1 Reference optical axis
f focal length
P1, P2, P3 incident point
θ0 to θ1 Incident angle

Claims (10)

A projection-side optical system that irradiates a measurement medium light including various irradiation angle components to a film thickness measurement point of a sample,
The photoelectric conversion unit includes a photoelectric conversion unit array means in which a large number of photoelectric conversion units are arranged in an array on a light receiving surface, and reflects reflected light of measurement medium light at each irradiation angle coming from a film thickness measurement point of the sample. A light receiving side optical system including a lens for guiding the light on the light receiving surface of the conversion unit array means,
Computing means for obtaining a film thickness to be measured based on a series of received light amount data obtained from each photoelectric conversion unit of the photoelectric conversion unit array means,
The film thickness measurement, wherein a distance between the lens included in the light receiving side optical system and the light receiving surface of the photoelectric conversion unit array means is set to substantially coincide with a focal length of the lens. apparatus. The projection-side optical system includes a characterization unit that performs characterization corresponding to a reference optical axis on the measurement medium light,
The light receiving side optical system includes a second photoelectric conversion unit for receiving reflected light of the measurement medium light coming from the film thickness measurement point of the sample and detecting a feature corresponding to a reference optical axis included therein. A sample included in a series of received light amount data obtained from each photoelectric conversion unit based on a feature corresponding to a reference optical axis detected by the second photoelectric conversion unit. 2. A film thickness measuring apparatus according to claim 1, further comprising means for correcting an amount of received light data for correcting an error component caused by the angle flutter. 3. The film thickness measuring apparatus according to claim 2, wherein said characterization means is a cross-sectional contour shaping means for edge-shaping a portion corresponding to a reference optical axis in a cross-sectional contour of the measurement medium light. 4. The film thickness measuring apparatus according to claim 3, wherein the cross-section contour shaping unit includes at least a slit, an aperture, or a knife edge. 3. The film thickness measuring apparatus according to claim 2, wherein the photoelectric conversion unit array means also serves as the second photoelectric conversion means. The received light amount data correction means obtains an angle shift amount of the sample from a light receiving position on the second photoelectric conversion means at a portion corresponding to a reference optical axis of the measurement medium light, and the obtained angle shift amount A function of shifting a series of received light amount data obtained from each of the photoelectric conversion units of the photoelectric conversion unit array means in a direction in which the angle shift amount is corrected in accordance with the following. 3. The film thickness measuring device according to 2. The measurement medium light emitted from the light projecting side optical system includes two or more wavelength components, and the light receiving side optical system includes two or more photoelectric conversion unit array means corresponding to each wavelength component. And a wavelength-specific spectral unit for separating reflected light arriving from the film thickness measuring point on the sample into each wavelength component and guiding the reflected light to each of the corresponding photoelectric conversion unit array units. The film thickness measuring device according to any one of the above. The measurement medium light emitted from the light projecting side optical system includes two or more polarized light components, and the light receiving side optical system includes two or more photoelectric conversion unit array means corresponding to each polarized light component. 7. A polarization-specific spectroscopic means for separating reflected light coming from a film thickness measurement point on the sample into respective polarized light components and guiding the polarized light components to respective corresponding photoelectric conversion unit array means. The film thickness measuring device according to any one of the above. The light projection optical system includes means for irradiating two or more types of measurement medium light having different wavelength components or polarization components in a time-division manner,
The arithmetic means, by synchronizing with the irradiation timing of the measurement medium light, obtains the received light amount data for each wavelength component or each polarized light component via each photoelectric conversion unit of the photoelectric conversion unit array means. The film thickness measuring device according to claim 1, wherein: The film thickness measuring apparatus according to any one of claims 1 to 9, wherein the film thickness measuring apparatus is arranged in a production line of a product including a film forming process such as a semiconductor product or an FPD and measures the film thickness of the product in-line. .

Images