JP2004184194A

JP2004184194A - Device and method for measuring shape

Info

Publication number: JP2004184194A
Application number: JP2002350599A
Authority: JP
Inventors: Tsutomu Morimoto; 勉森本; Hiroyuki Takamatsu; 弘行高松
Original assignee: Kobe Steel Ltd
Current assignee: Kobe Steel Ltd
Priority date: 2002-12-03
Filing date: 2002-12-03
Publication date: 2004-07-02

Abstract

<P>PROBLEM TO BE SOLVED: To remove an influence of surface roughness in the direction perpendicular to an end side of a measurement object in a predetermined region of the measurement object such as the shear drop shape in the vicinity of the end side of a wafer back face and accurately measure the shape. <P>SOLUTION: Two-dimensional surface displacement in the predetermined measurement region A of the measurement object is measured. The one-dimensional profile in the direction (the radial direction) perpendicular to the end side 2a is calculated by processing the measured surface displacement so as to be averaged in the direction parallel to the side end 2a of the measurement object. A width in the direction parallel to the end side 2a of the measurement region A is larger than a spatial period of unevenness due to the surface roughness. The two-dimensional surface displacement of the measurement region A is calculated based on an interference image detected by an interferometer on a surface of the measurement region A. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

【０００１】
【発明の属する技術分野】
本発明は，被測定物の所定の測定領域の表面形状（平坦度等）を測定する形状測定装置であって，特に，半導体ウェーハ裏面の端辺付近の表面形状を測定するのに好適な形状測定装置及び形状測定方法に関するものである。
【０００２】
【従来の技術】
一般に，半導体ウェーハ（以下，ウェーハという）においては，その中央部での平坦度は十分確保されるが，その端辺（エッジ）付近では，端辺に向かうに従って平坦度が悪化するいわゆるダレ（ロールオフ）が形成される傾向がある。
一方，近年，１枚のウェーハから得られる半導体チップの数を極力増やすため，ウェーハの端辺により近い部分まで高い平坦度が要求される。このため，ウェーハ全範囲のうち，端辺近辺の領域の表面形状（即ち，ダレ）を測定することが重要となる。
例えば，非特許文献１や特許文献１には，ウェーハのエッジ先端にブロック部材を当接させ，該ブロック部材のウェーハに当接する面の位置を基準とし，触針式の形状計をウェーハの端辺の直角方向（ウェーハの端辺が円弧状である場合にはその円弧の半径方向）に移動させながらウェーハの表面形状を測定することによってウェーハの端辺付近のロールオフ値（ダレ）を測定する技術が示されている。
ところで，ウェーハを取り扱うプロセスでは，ウェーハが載置されるステージによりウェーハの裏面を吸着するステッパーやウェーハをその裏面から研磨手段に押し付けるＣＭＰ（化学機械研磨）プロセス等においてウェーハの裏面形状がデバイスを作成するウェーハのおもて面の形状に影響を及ぼす。このように，ウェーハのダレ（ロールオフ）は裏面側でも問題となるため，デバイスの歩留まり向上のためにはウェーハの裏面側の端辺付近の形状（即ち，ダレ形状）を把握する必要がある。
【０００３】
【特許文献１】
特開２０００−１４６５６９号公報
【非特許文献１】
「Ｍ．Ｋｉｍｕｒａｅｔａｌ．，Ｊｐｎ．Ｊ．Ａｐｐｌ．Ｐｈｙｓ．Ｖｏｌ．３８（１９９９）」
【０００４】
【発明が解決しようとする課題】
しかしながら，一般に，ウェーハの裏面はポリッシュされていない粗面であるため，特許文献１及び非特許文献１に示されるような従来の方法でウェーハ裏面の表面変位を測定すると，粗面の凹凸によって測定値（表面変位）が大きく変化する。ところが，実際に把握したい裏面形状は，その表面粗さによる凹凸の空間周期よりも緩慢に形状が変化するダレ形状であり，従来の方法では正確なダレ形状を測定できないという問題点があった。
また，短周期成分と長周期成分とを有するデータから長周期成分のみを抽出する場合，一般に，ＦＦＴによるローパスフィルタ処理や荷重平均フィルタ処理の１つであるガウシアンフィルタリング処理等を行うことが考えられるが，ＦＦＴでは，入力波形に一般的に適用される窓関数の影響により，また，荷重平均フィルタの場合は被測定物の端部への適用が困難（例えば，２ｍｍ程度の移動平均では端部の１ｍｍ程度の範囲の計算ができない）なことから，これらフィルタリング処理によりダレ形状を抽出することは困難である。
従って，本発明は上記事情に鑑みてなされたものであり，その目的とするところは，ウェーハ裏面の端辺付近のダレ形状のように被測定物の所定領域について被測定物の端辺に略直角の方向における表面粗さの影響を除去した形状を正確に測定することができる形状測定装置及び形状測定方法を提供することにある。
【０００５】
【課題を解決するための手段】
上記目的を達成するために本発明は，被測定物の所定の測定領域における前記被測定物の端辺に略直角の方向の表面形状を測定する形状測定装置において，前記測定領域の２次元の表面変位を測定する表面変位測定手段と，測定された前記表面変位について前記端辺に略平行な方向に所定の平均化処理を行うことにより前記端辺に略直角の方向の１次元プロフィール（即ち，１次元の表面変位分布）を算出する１次元プロフィール算出手段と，を具備してなることを特徴とする形状測定装置として構成されるものである。
これにより，前記測定領域における被測定物の端辺に略直角の方向の各位置について，前記表面変位が平均化処理された変位量が算出され，該変位量が前記端辺に略直角の方向の１次元プロフィールを表すことになる。この１次元プロフィールを構成する各変位量は，複数の前記表面変位が平均化処理されたものであるので，被測定物の表面粗さによる凹凸の影響が除去されている。しかも，その変位量は，被測定物の前記端辺に略平行な方向に平均化処理されたものであるので，前記端辺に略直角の方向に移動平均等の平均化処理を行う場合のように，前記端辺に略直角の方向における空間分解能は低下しない。従って，ウェーハ裏面の端辺付近のダレ形状のように被測定物の所定領域について被測定物の端辺に略直角の方向における表面粗さの影響を除去した形状を正確に測定することができる。
ここで，前記被測定物の端辺とは，前記被測定物の端部の一部をなす辺であり，直線状の端辺に限るものでなく，例えば円弧状の端辺等の曲線状のものも含まれる。また，前記端辺が曲線状である場合，前記端辺に略直角の方向とは，前記端辺の接線方向に対して略直角の方向のことをいうものとする。例えば，前記端辺が円弧状である場合，前記端辺の略直角の方向とは，その円弧の半径方向又はそれに近い方向を表す。
【０００６】
また，前記平均化処理により被測定物の表面粗さの凹凸の影響を十分除去するためには，前記測定領域の前記端辺に略平行な方向の幅が，前記測定領域の表面粗さによる凹凸の空間周期よりも十分長いものであることが望ましい。
また，前記表面変位測定手段としては，例えば，前記測定領域表面における干渉像を検出する干渉計と，検出された前記干渉像に基づいて前記２次元の表面変位を算出する表面変位算出手段とを具備するものが考えられる。
これにより，光学的に精度の高い測定を短時間で行うことが可能となる。
【０００７】
また，本発明は，前記形状測定装置に対応する形状測定方法として捉えてもよい。即ち，被測定物の所定の測定領域における前記被測定物の端辺に略直角の方向の表面形状を測定する形状測定方法において，前記測定領域の２次元の表面変位を測定する表面変位測定手順と，測定された前記表面変位について前記端辺に略平行な方向に所定の平均化処理を行うことにより前記端辺に略直角の方向の１次元プロフィールを算出する１次元プロフィール算出手順と，を有してなることを特徴とする形状測定方法である。
【０００８】
【発明の実施の形態】
以下添付図面を参照しながら，本発明の実施の形態について説明し，本発明の理解に供する。尚，以下の実施の形態は，本発明を具体化した一例であって，本発明の技術的範囲を限定する性格のものではない。
ここに，図１は本発明の実施の形態に係る形状測定装置Ｘの概略構成を表す図，図２は本発明の実施の形態に係る形状測定装置Ｘで形状測定するウェーハの測定領域を設定するための校正用ウェーハがステージ上に載置された状態を表す図，図３は本発明の実施の形態に係る形状測定装置Ｘのステージ上にウェーハを載置した状態の平面及び側面を表す図，図４は本発明の実施の形態に係る形状測定装置Ｘによる形状測定と従来法による形状測定との各測定結果であるウェーハ端辺付近の一次元プロフィールを表すグラフ，図５は測定面の表面粗さを空間周期が０．１ｍｍ及び振幅が１μｍの正弦波で表した場合における５ｍｍの範囲の平均化処理を行った結果を表すグラフである。
【０００９】
まず，図１を用いて，本発明の実施の形態に係る形状測定装置Ｘの構成について説明する。
本形状測定装置Ｘは，被測定物の一例であるウェーハ１の載置台であり該ウェーハ１を位置決めするためのピン１８が設けられたステージ１７と，該ステージ１７上にその測定面１ａを上にして載置された前記ウェーハ１の上方に設けられた斜入射干渉計１０と，該斜入射干渉計１０から得られる干渉像（干渉縞画像）を入力する演算器３１とを具備している。ここで，前記ウェーハ１の測定面１ａは，ポリッシュ研磨されていない粗面である裏面であるとする。
【００１０】
前記斜入射干渉計１０は，前記ウェーハ１に対して測定光１２を出射する発光器１１，その測定光１２を前記ウェーハ１の測定面１ａに対して斜め方向の平行ビームとするコリメータレンズ１４，その平行ビーム１２の一部を透過し，一部を反射する基準平面１５ａを有する三角プリズム１５，前記ウェーハ１の測定面１ａで反射された測定光が前記基準平面１５ａ（前記三角プリズム１５の一面）及びコリメータレンズ１４’を経た光を受光するＣＣＤカメラ等の受光器１６，及び前記三角プリズム１５を前記ウェーハ１の面に略垂直な方向に移動させるピエゾアクチュエータ１９を備えている。このような構成により，前記基準平面１５ａでの反射光と，前記ウェーハ１の測定面１ａでの反射光とには，前記基準平面１５ａと前記測定面１ａとの距離に対応する光路差があるため，この光路差が，前記受光器１６において，両反射光で形成される干渉縞として観測される。前記三角プリズム１５は，前記基準平面１５ａが前記ウェーハ１の測定面１ａに近接するよう（例えば，０．１ｍｍ程度の距離）前記ピアゾアクチュエータ１９により位置決めされる。前記測定光１２としては，例えば，ＨｅＮｅレーザ（λ＝６３３ｎｍ）や，半導体レーザ等のコヒーレント光が用いられる。また，前記三角プリズム１５は，フィゾー干渉で用いられるようなウェッジプリズム等の参照板を用いてもかまわない。
【００１１】
前記受光器１６で得られる干渉縞画像は，画像入力手段を有するパーソナルコンピュータ等である前記演算器３１に入力され，入力された干渉縞画像に基づいて前記ウェーハ１の測定面１ａの表面形状（２次元の表面変位分布）が演算される。該演算は，前記演算器３１に記憶された所定のプログラムが実行されることにより行われる。
ここで，前記三角プリズム１５の前記基準平面１５ａは，前記ピアゾアクチュエータ１９によって複数の位置に位置決めされ，それぞれの位置での干渉縞画像が輝度データとして前記演算器３１に取り込まれる。前記基準平面１５ａの位置決めは，前記測定光１２の前記ウェーハ１の測定面１ａへの入射角θ（不図示）を考慮した位置（前記測定面１ａとの距離）に位置決めされる。例えば，前記基準平面１５ａと前記測定面１ａとの距離を，所定の基準距離を０として，０，λ／８×１／ｃｏｎθ，２λ／８×１／ｃｏｓθ，３λ／８×１／ｃｏｓθ（λ＝λ’／ｃｏｓψ （λ’は前記測定光１２の波長，ψは前記測定光１２の前記ウェーハ１への入射角））の各位置に移動させ，それぞれの位置での干渉縞画像を輝度データとして前記演算器３１に取り込めば，それぞれ位相が０°，９０°，１８０°，２７０°に変化した干渉縞画像が得られる。それぞれの位置での干渉縞画像の座標（ｘ，ｙ）における画像データ（輝度）をＩ_０（ｘ，ｙ），Ｉ_９０（ｘ，ｙ），Ｉ_１８０（ｘ，ｙ），Ｉ_２７０（ｘ，ｙ）とすると，干渉縞画像の位相φ（ｘ，ｙ）は，次の（１）式で求められる。
φ（ｘ，ｙ）＝ａｒｃｔａｎ（Ｉ_０−Ｉ_１８０）／（Ｉ_９０−Ｉ_２７０） …（１）
このようにして求められた位相φ（ｘ，ｙ）に周知のアンラップ処理を施し，アンラップ処理後の位相をφ’（ｘ，ｙ）とすると，例えば次の（２）式により，前記測定面１ａの表面変位Ｌａ（ｘ，ｙ）（表面形状）を求めることができる。
Ｌａ（ｘ，ｙ）＝（φ’（ｘ，ｙ）／２π）×（λ／２） …（２）
このような処理（演算）を前記演算期３１で行うことにより，前記基準平面１５ａでカバーされる領域の一部又は全部について前記測定面１ａと前記基準平面１５ａとの距離の分布，即ち，前記測定面１ａの２次元の表面変位を算出できる。もちろん，表面変位の算出方法としては，周知の他の位相シフト法を用いてもかまわない。ここで，前記干渉計１０及び前記演算器３１が，前記表面変位測定手段の一例である。
【００１２】
次に，図２を用いて，前記演算器３１による前記ウェーハ１の表面形状（表面変位）の測定領域について説明する。
本形状測定装置Ｘでは，被測定物である前記ウェーハ１を測定する前に，予め，前記ウェーハ１と平面形状が略等しい所定の校正用ウェーハを用いて測定領域を設定する。
図２は，前記ウェーハ１の測定領域を設定するための前記校正用ウェーハ２が前記ステージ１７上に載置された状態を表す図である。図２に示すように，円盤状の前記校正用ウェーハ２の表面には，その中心Ｏからその端辺２ａ付近まで２本の直線▲３▼，▲４▼がけがかれている（以下，この直線をケガキ線▲３▼，▲４▼という）。該ケガキ線▲３▼，▲４▼は所定の角度（例えば２°）をなしている。さらに，前記端辺２ａ付近には，円弧状の前記端辺２ａに平行な円弧▲１▼，▲２▼がそれぞれ前記ケガキ線▲３▼，▲４▼の両方に交差するようけがかれている（以下，この円弧をケガキ線▲１▼，▲２▼という）。例えば，前記ケガキ線▲１▼は前記中心Ｏから半径ｒ＝１４０ｍｍの位置に，前記ケガキ線▲２▼は半径ｒ＝１４９ｍｍの位置にそれぞれけがかれている。これらケガキ線▲１▼〜▲４▼で囲まれる領域を測定領域Ａとする。
このような校正用ウェーハ２をその端部が２つの前記ピン１８に当接するとともに，前記測定領域Ａが前記三角プリズム１５の前記基準平面１５ａの範囲内（破線枠内）に入るように配置した状態で，前記受光器１６及び前記演算器３１により画像を取り込むことにより，前記演算器３１によって前記ケガキ線▲１▼〜▲４▼を検出することができる。さらに，検出された前記ケガキ線▲１▼〜▲４▼の位置（座標）から，前記測定領域Ａと，前記ケガキ線▲１▼，▲２▼の方向，即ち，円弧の中心及び半径とが求められる。このようにして求められる前記ケガキ線▲１▼，▲２▼の方向は，前記校正用ウェーハ２の前記端辺２ａと平行な方向（以下，端辺平行方向という）である。該端辺平行方向は，前記校正用ウェーハ２が円盤状であるのでその周方向（に平行な方向）である。従って，図４に示すように，前記校正用ウェーハ２をこれと同等の平面形状を有する被測定物である前記ウェーハ１に置き換えて同様に配置した場合には，前記ケガキ線▲１▼，▲２▼により求めた前記端辺平行方向は，前記ウェーハ１の前記三角プリズム１５（の前記基準平面１５ａ）でカバーされる領域における端辺１ｂと略平行な方向（周方向に平行な方向）であるといえる。
【００１３】
次に，前記演算器３１による前記測定領域Ａの表面形状の測定処理について説明する。
前記校正用ウェーハ２を用いて前記測定領域Ａ及び前記端辺平行方向を予め求めた後，まず，図４に示すように被測定物である前記ウェーハ１をその端部が２つの前記ピン１８に当接するように前記ステージ１７上に載置した状態で，干渉縞画像（干渉像）が取り込まれる。
次に，所定のアンラップ処理等により，前記測定領域Ａにおける２次元の表面変位が測定される（前記表面変位測定手段の処理及び前記表面変位測定手順の一例）。
さらに，測定された前記測定領域Ａの表面変位のデータについて，前記端辺平行方向に積算して該積算を行ったデータ数で割るいわゆる単純平均を求める（前記１次元プロフィール算出手段の処理及び前記１次元プロフィール算出手順の一例）。このようにして求められた前記表面変位データの単純平均は，前記端辺平行方向に直角な方向（即ち，前記ウェーハ１の端辺１ｂの略直角方向）の１次元プロフィール（１次元の表面変位分布）を表すことになる。
ここで，前記単純平均を求める際に，前記端辺平行方向に対応する角度による座標系と前記受光器１６（即ち，カメラ）のＸＹ座標系（画素の座標系）との変換が必要となり，該変換によって前記受光器１６の画素と一致しない座標のデータが必要となる場合も生じるが，そのような場合は，一般的な補間法を用いて補間したデータを用いるようにすればよい。
【００１４】
図４（ａ），（ｂ）は，前記端辺平行方向（周方向）の平均化処理を行わずに単純に前記ウェーハ１の前記測定領域Ａにおいて，半径方向（前記端辺平行方向に直角の方向）に表面変位（形状）を測定した１次元プロフィールの一例を表すグラフである。ここで，半径方向の位置は，前記ウェーハ１の端辺（エッジ）からの距離で表す。図４（ａ），（ｂ）に示されるように，前記ウェーハ１の裏面（前記測定面１ａ）は粗面であるため，粗面の凹凸によって測定値（表面変位）が大きく変化している。このような表面変位データに基づいてＲＯＡ（Ｒｏｌｌ−ｏｆｆＡｍｏｕｎｔ）を評価しても，表面粗さによる凹凸の空間周期よりも緩慢な形状変化であるダレ形状を表すものではないため，そのＲＯＡは前記ウェーハ１の加工制御に用いることができない。
【００１５】
一方，図４（ｃ），（ｄ）は，図４（ａ），（ｂ）と同じ前記ウェーハ１の前記測定領域Ａについて，本形状測定装置Ｘにより測定した半径方向（前記端辺平行方向に直角の方向）の１次元プロフィールを表すグラフである。
図４（ｃ），（ｄ）に示されるように，本形状測定装置Ｘで計測された１次元プロフィールを構成する各変位量（表面変位の平均値）は，前記端辺平行方向における複数の表面変位データが平均化処理されたものであるので，前記測定面１ａの表面粗さによる凹凸の影響が除去されている。しかも，その変位量は，前記ウェーハ１の端辺１ｂに略平行な方向に平均化処理されたものであるので，前記端辺１ｂに略直角の方向に移動平均等の平均化処理を行う場合のように，前記端辺１ｂに略直角の方向（半径方向）における空間分解能は低下しない。従って，前記ウェーハ１ｂ裏面の端辺付近のダレ形状を正確に測定することができる。このような変位量を用いてＲＯＡを求めれば，そのＲＯＡは，前記ウェーハ１の端辺付近（端部付近）の加工制御に用いることが可能となる。
【００１６】
また，図３に示した例では，前記測定領域Ａの前記端辺平行方向の幅は，前記ウェーハ１の最も内側（前記ケガキ線▲１▼の位置）で約４．９ｍｍ，最も外側（前記ケガキ線▲２▼の位置）で約５．２ｍｍである。ここで，５ｍｍの範囲を平均化処理した場合，測定データの空間周期が５ｍｍ程度以下である成分を減衰するローパスフィルタとして作用する。前記ウェーハ１では，前記測定面１ａの表面粗さによる凹凸の空間周期が０．１ｍｍ〜０．０５ｍｍ程度であることを考慮すると，５ｍｍ程度の範囲について平均化処理を行うことにより，前記測定面１ａの表面粗さによる前記凹凸の影響を十分抑える（遮断する）ことができる。
例えば，表面粗さを空間周期λ_ｎ，振幅Ｂの正弦波（平均化処理を行う方向の座標をｘとする）として表すと，次の（３）式で表される。
【数１】

この周期の表面粗さを有する前記測定面１ａの測定データを，５ｍｍの範囲（区間）で平均化処理した結果は，次の（４）式で表される。
【数２】

この（４）式を，Ｂ＝１μｍ，λ_ｎ＝０．１ｍｍ近郷でグラフ化したものが図５である。図５に示されるように，平均化処理を行うことによって前記測定面１ａの表面粗さ（振幅Ｂ＝１μｍ）の影響（変動）を約０．００３倍以下と十分小さく抑えられることがわかる。従って，この測定領域Ａの前記端辺平行方向の幅は前記ウェーハ１のダレ形状測定としては十分な長さである。
また，前記ケガキ線▲３▼，▲４▼を平行な線として前記測定領域Ａの前記端辺平行方向の幅を等幅とすること等，前記測定領域Ａの形状は他の形状であってもかまわない。
【００１７】
本形状測定装置Ｘでは，図４（ｃ），（ｄ）に示されるような１次元プロフィールが前記演算器３１の記憶部に記憶されるとともに，ＣＲＴ等の表示装置に表示される。
なお，本実施の形態では，前記表面変位のデータを前記端辺平行方向に単純平均する例を示したが，これに限るものでなく，例えば，メディアンや中央値，或いはガウシアンやダブルガウシアンのような重み付け平均等の他の平均化処理を行うものであってもよい。特に，ガウシアンやダブルガウシアンによるフィルタリングは，前述した単純平均よりも，より高周波成分の減衰化特性に優れるという特徴を有しているため好適である。
また，２次元の表面変位の測定は，斜入射干渉計を用いることに限らず，フィゾー干渉計やマイケルソン干渉計を用いること等，他の測定装置及び方法を用いるものであってもかまわない。
【００１８】
【発明の効果】
以上説明したように，本発明によれば，被測定物の所定の測定領域における端辺に略直角の方向の各位置について，表面変位が平均化処理された変位量が算出され，該変位量が前記端辺に略直角の方向の１次元プロフィールを表すことになるので，この１次元プロフィールにより表される形状は，測定領域の表面粗さによる凹凸の影響が除去されている。しかも，その変位量は，被測定物の前記端辺に略平行な方向に平均化処理されたものであるので，前記端辺に略直角の方向における空間分解能は低下しない。従って，ウェーハ裏面の端辺付近のダレ形状等の表面粗さの影響を除去した形状を正確に測定することができる。
また，測定領域の２次元の表面変位を干渉計による干渉像から算出するよう構成することにより，光学的に精度の高い測定を短時間で行うことが可能となる。
【図面の簡単な説明】
【図１】本発明の実施の形態に係る形状測定装置Ｘの概略構成を表す図。
【図２】本発明の実施の形態に係る形状測定装置Ｘで形状測定するウェーハの測定領域を設定するための校正用ウェーハがステージ上に載置された状態を表す図。
【図３】本発明の実施の形態に係る形状測定装置Ｘのステージ上にウェーハを載置した状態の平面及び側面を表す図。
【図４】本発明の実施の形態に係る形状測定装置Ｘによる形状測定と従来法による形状測定との各測定結果であるウェーハ端辺付近の一次元プロフィールを表すグラフ。
【図５】測定面の表面粗さを空間周期が０．１ｍｍ及び振幅が１μｍの正弦波で表した場合における５ｍｍの範囲の平均化処理を行った結果を表すグラフ。
【符号の説明】
１…ウェーハ（被測定物）
１ａ…測定面（粗面）
１ｂ…ウェーハの端辺
２…校正用ウェーハ
２ａ…校正用ウェーハの端辺
１０…斜入射干渉計
１１…発光器
１２…測定光
１４，１４’…コリメータレンズ
１５…三角プリズム
１５ａ…基準平面
１６…受光器
１７…ステージ
１８…ピン
１９…ピエゾアクチュエータ
３１…演算器（１次プロフィール算出手段）
Ａ…測定領域[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a shape measuring apparatus for measuring a surface shape (flatness or the like) of a predetermined measurement area of a device under test, and particularly to a shape suitable for measuring a surface shape near an edge of a back surface of a semiconductor wafer. The present invention relates to a measuring device and a shape measuring method.
[0002]
[Prior art]
Generally, in a semiconductor wafer (hereinafter, referred to as a wafer), the flatness at the center is sufficiently ensured, but near the edge, the so-called sagging (roll) in which the flatness deteriorates toward the edge. Off) tends to form.
On the other hand, in recent years, in order to increase the number of semiconductor chips obtained from one wafer as much as possible, high flatness is required to a portion closer to an edge of the wafer. For this reason, it is important to measure the surface shape (that is, sag) of the region near the edge in the entire wafer range.
For example, in Non-Patent Document 1 and Patent Document 1, a block member is brought into contact with the tip of the edge of a wafer, and a stylus-type shape meter is used as a reference for the position of the surface of the block member that comes into contact with the wafer. The roll-off value (sag) near the edge of the wafer is measured by measuring the surface shape of the wafer while moving in the direction perpendicular to the side (in the case where the edge of the wafer is an arc, in the radial direction of the arc). The technique to do is shown.
By the way, in the process of handling wafers, the shape of the backside of the wafer is created by a stepper that attracts the backside of the wafer by the stage on which the wafer is mounted, or a CMP (chemical mechanical polishing) process that presses the wafer against the polishing means from the backside. Affects the shape of the front surface of the wafer. As described above, since the sagging (roll-off) of the wafer is also a problem on the back side, it is necessary to grasp the shape near the back side of the wafer (ie, sagging shape) in order to improve the yield of the device. .
[0003]
[Patent Document 1]
Japanese Patent Application Laid-Open No. 2000-146569 [Non-Patent Document 1]
"M. Kimura et al., Jpn. J. Appl. Phys. Vol. 38 (1999)".
[0004]
[Problems to be solved by the invention]
However, since the back surface of the wafer is generally a rough surface that has not been polished, when the surface displacement of the back surface of the wafer is measured by a conventional method as shown in Patent Literature 1 and Non-Patent Literature 1, the measurement is based on the roughness of the rough surface. The value (surface displacement) changes greatly. However, the shape of the back surface that one actually wants to grasp is a sag shape in which the shape changes more slowly than the spatial period of the irregularities due to the surface roughness, and there has been a problem that the conventional sag shape cannot be measured accurately.
When only a long-period component is extracted from data having a short-period component and a long-period component, it is generally considered that a low-pass filter process by FFT or a Gaussian filtering process which is one of the weighted average filter processes is performed. However, in FFT, it is difficult to apply to the end of the DUT due to the influence of the window function generally applied to the input waveform (for example, the moving average of about 2 mm Cannot be calculated in the range of about 1 mm), it is difficult to extract the sag shape by these filtering processes.
Accordingly, the present invention has been made in view of the above circumstances, and an object thereof is to substantially match a predetermined region of an object to be measured with an edge of the object to be measured such as a sagging shape near the edge of the back surface of the wafer. It is an object of the present invention to provide a shape measuring device and a shape measuring method capable of accurately measuring a shape in which the influence of surface roughness in a right angle direction is removed.
[0005]
[Means for Solving the Problems]
In order to achieve the above object, the present invention relates to a shape measuring apparatus for measuring a surface shape in a direction substantially perpendicular to an edge of the measured object in a predetermined measuring region of the measured object. A one-dimensional profile in a direction substantially perpendicular to the edge by performing a predetermined averaging process on the measured surface displacement in a direction substantially parallel to the edge; , A one-dimensional profile calculation means for calculating a one-dimensional surface displacement distribution).
Thereby, for each position in the direction substantially perpendicular to the edge of the measured object in the measurement area, the displacement amount obtained by averaging the surface displacement is calculated, and the displacement amount is calculated in the direction substantially perpendicular to the edge. Will be represented. Since each of the displacements constituting the one-dimensional profile is obtained by averaging a plurality of the surface displacements, the influence of the unevenness due to the surface roughness of the measured object is eliminated. In addition, since the displacement amount is averaged in a direction substantially parallel to the end side of the object to be measured, it is necessary to perform averaging processing such as moving average in a direction substantially perpendicular to the end side. As described above, the spatial resolution in the direction substantially perpendicular to the edge does not decrease. Therefore, it is possible to accurately measure a shape such as a sagging shape near the edge of the back surface of the wafer, in which the influence of surface roughness in a direction substantially perpendicular to the edge of the object is removed for a predetermined region of the object. .
Here, the edge of the object to be measured is a side that forms a part of the edge of the object to be measured, and is not limited to a linear edge, but may be a curved edge such as an arc edge. Also included. Further, when the edge is curved, a direction substantially perpendicular to the edge means a direction substantially perpendicular to a tangent direction of the edge. For example, when the edge is arc-shaped, the direction substantially perpendicular to the edge indicates the radial direction of the arc or a direction close thereto.
[0006]
In order to sufficiently remove the influence of the unevenness of the surface roughness of the object to be measured by the averaging process, the width of the measurement region in a direction substantially parallel to the edge is determined by the surface roughness of the measurement region. It is desirable that the length be sufficiently longer than the spatial period of the irregularities.
The surface displacement measuring means includes, for example, an interferometer for detecting an interference image on the surface of the measurement area, and a surface displacement calculating means for calculating the two-dimensional surface displacement based on the detected interference image. What is provided is conceivable.
This makes it possible to perform optically accurate measurement in a short time.
[0007]
Further, the present invention may be considered as a shape measuring method corresponding to the shape measuring device. That is, in a shape measuring method for measuring a surface shape in a direction substantially perpendicular to an edge of the measured object in a predetermined measuring region of the measured object, a surface displacement measuring step of measuring a two-dimensional surface displacement of the measured region is performed. And a one-dimensional profile calculation procedure of calculating a one-dimensional profile in a direction substantially perpendicular to the edge by performing a predetermined averaging process on the measured surface displacement in a direction substantially parallel to the edge. This is a shape measuring method characterized by having a shape.
[0008]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings to facilitate understanding of the present invention. The following embodiments are examples embodying the present invention, and do not limit the technical scope of the present invention.
Here, FIG. 1 is a view showing a schematic configuration of a shape measuring apparatus X according to an embodiment of the present invention, and FIG. FIG. 3 is a view showing a state in which a calibration wafer is mounted on a stage, and FIG. 3 is a plan view and a side view showing a state in which the wafer is mounted on a stage of the shape measuring apparatus X according to the embodiment of the present invention. FIGS. 4 and 5 are graphs showing one-dimensional profiles near the edge of the wafer, which are the measurement results of the shape measurement by the shape measuring apparatus X according to the embodiment of the present invention and the shape measurement by the conventional method. FIG. 5 is a graph showing a result of performing an averaging process in a range of 5 mm when the surface roughness is represented by a sine wave having a spatial period of 0.1 mm and an amplitude of 1 μm.
[0009]
First, the configuration of a shape measuring apparatus X according to an embodiment of the present invention will be described with reference to FIG.
The shape measuring apparatus X is a stage 17 on which a pin 18 for positioning the wafer 1 is provided, which is a mounting table for the wafer 1 which is an example of an object to be measured. An oblique incidence interferometer 10 provided above the wafer 1 placed in the above-described manner, and a calculator 31 for inputting an interference image (interference fringe image) obtained from the oblique incidence interferometer 10. . Here, it is assumed that the measurement surface 1a of the wafer 1 is a rough rear surface that has not been polished.
[0010]
The oblique incidence interferometer 10 includes a light emitter 11 that emits measurement light 12 to the wafer 1, a collimator lens 14 that converts the measurement light 12 into a parallel beam oblique to the measurement surface 1 a of the wafer 1, A triangular prism 15 having a reference plane 15a that transmits a part of the parallel beam 12 and reflects a part thereof, and the measurement light reflected on the measurement surface 1a of the wafer 1 is reflected by the reference plane 15a (one surface of the triangular prism 15). ) And a light receiver 16 such as a CCD camera for receiving light passing through the collimator lens 14 ′, and a piezo actuator 19 for moving the triangular prism 15 in a direction substantially perpendicular to the surface of the wafer 1. With such a configuration, the reflected light on the reference plane 15a and the reflected light on the measurement surface 1a of the wafer 1 have an optical path difference corresponding to the distance between the reference plane 15a and the measurement surface 1a. Therefore, this optical path difference is observed in the light receiver 16 as an interference fringe formed by both reflected lights. The triangular prism 15 is positioned by the Piazo actuator 19 so that the reference plane 15a is close to the measurement surface 1a of the wafer 1 (for example, a distance of about 0.1 mm). As the measurement light 12, for example, a coherent light such as a HeNe laser (λ = 633 nm) or a semiconductor laser is used. Further, the triangular prism 15 may use a reference plate such as a wedge prism used for Fizeau interference.
[0011]
The interference fringe image obtained by the light receiver 16 is input to the computing unit 31 which is a personal computer or the like having an image input unit, and the surface shape of the measurement surface 1a of the wafer 1 is determined based on the input interference fringe image. (Two-dimensional surface displacement distribution) is calculated. The calculation is performed by executing a predetermined program stored in the calculator 31.
Here, the reference plane 15a of the triangular prism 15 is positioned at a plurality of positions by the Piazo actuator 19, and the interference fringe images at the respective positions are taken into the arithmetic unit 31 as luminance data. The reference plane 15a is positioned at a position (distance from the measurement surface 1a) in consideration of an incident angle θ (not shown) of the measurement light 12 to the measurement surface 1a of the wafer 1. For example, assuming that the distance between the reference plane 15a and the measurement surface 1a is a predetermined reference distance of 0, 0, λ / 8 × 1 / con θ, 2λ / 8 × 1 / cos θ, 3λ / 8 × 1 / cos θ ( λ = λ ′ / cosψ (λ ′ is the wavelength of the measurement light 12, ψ is the angle of incidence of the measurement light 12 on the wafer 1)), and the interference fringe image at each position is illuminated. If the data is taken into the computing unit 31, interference fringe images whose phases have changed to 0 °, 90 °, 180 °, and 270 ° can be obtained. The image data (luminance) at the coordinates (x, y) of the interference fringe image at each position is represented by I ₀ (x, y), I ₉₀ (x, y), I ₁₈₀ (x, y), I ₂₇₀ (x , Y), the phase φ (x, y) of the interference fringe image is obtained by the following equation (1).
φ (x, y) = arctan (I ₀ −I ₁₈₀ ) / (I ₉₀ −I ₂₇₀ ) (1)
If the phase φ (x, y) thus obtained is subjected to a well-known unwrapping process and the phase after the unwrapping process is assumed to be φ ′ (x, y), for example, the measurement surface is obtained by the following equation (2). The surface displacement La (x, y) (surface shape) of 1a can be obtained.
La (x, y) = (φ ′ (x, y) / 2π) × (λ / 2) (2)
By performing such processing (calculation) in the calculation period 31, the distribution of the distance between the measurement plane 1a and the reference plane 15a for a part or all of the area covered by the reference plane 15a, A two-dimensional surface displacement of the measurement surface 1a can be calculated. Of course, any other well-known phase shift method may be used for calculating the surface displacement. Here, the interferometer 10 and the calculator 31 are an example of the surface displacement measuring means.
[0012]
Next, the measurement area of the surface shape (surface displacement) of the wafer 1 by the arithmetic unit 31 will be described with reference to FIG.
In the present shape measuring apparatus X, before measuring the wafer 1 as an object to be measured, a measurement area is set in advance by using a predetermined calibration wafer having a plane shape substantially equal to the wafer 1.
FIG. 2 is a diagram illustrating a state in which the calibration wafer 2 for setting a measurement area of the wafer 1 is mounted on the stage 17. As shown in FIG. 2, on the surface of the disk-shaped calibration wafer 2, two straight lines (3) and (4) are drawn from the center O to the vicinity of the end side 2a (hereinafter, this straight line is referred to as a straight line). Straight lines are called marking lines (3) and (4)). The marking lines (3) and (4) form a predetermined angle (for example, 2 °). Further, in the vicinity of the end side 2a, arcs (1) and (2), which are parallel to the end side 2a, are marked so as to intersect both the marking lines (3) and (4), respectively. (Hereinafter, these arcs are referred to as marking lines (1) and (2)). For example, the marking line (1) is scribed at a radius r = 140 mm from the center O, and the marking line (2) is scribed at a radius r = 149 mm. The area surrounded by the marking lines (1) to (4) is referred to as a measurement area A.
Such a calibration wafer 2 is arranged such that its ends abut against the two pins 18 and that the measurement area A falls within the range of the reference plane 15a of the triangular prism 15 (within a broken-line frame). In this state, the image is captured by the light receiving device 16 and the computing unit 31, whereby the marking lines (1) to (4) can be detected by the computing unit 31. Further, from the detected positions (coordinates) of the marking lines (1) to (4), the measurement area A and the directions of the marking lines (1) and (2), that is, the center and radius of the arc, are determined. Desired. The directions of the marking lines {circle around (1)} and {circle around (2)} thus determined are directions parallel to the edge 2a of the calibration wafer 2 (hereinafter, referred to as edge parallel directions). The edge parallel direction is a circumferential direction of the calibration wafer 2 because the wafer 2 has a disk shape. Therefore, as shown in FIG. 4, when the calibration wafer 2 is replaced with the wafer 1 which is an object to be measured having an equivalent planar shape and arranged similarly, the marking lines (1), (1) The direction parallel to the edge determined in 2 ▼ is a direction substantially parallel to the edge 1b (direction parallel to the circumferential direction) in a region of the wafer 1 covered by the triangular prism 15 (the reference plane 15a). It can be said that there is.
[0013]
Next, measurement processing of the surface shape of the measurement area A by the arithmetic unit 31 will be described.
After the measurement area A and the edge side parallel direction are obtained in advance by using the calibration wafer 2, first, as shown in FIG. An interference fringe image (interference image) is taken in a state of being placed on the stage 17 so as to abut on the stage 17.
Next, a two-dimensional surface displacement in the measurement area A is measured by a predetermined unwrapping process or the like (an example of the process of the surface displacement measuring means and the surface displacement measuring procedure).
Further, a so-called simple average is obtained by integrating the measured surface displacement data of the measurement area A in the direction parallel to the edge and dividing by the number of data obtained by the integration (processing of the one-dimensional profile calculation means and the Example of one-dimensional profile calculation procedure). The simple average of the surface displacement data obtained in this way is a one-dimensional profile (one-dimensional surface displacement) in a direction perpendicular to the direction parallel to the edge (ie, substantially perpendicular to the edge 1b of the wafer 1). Distribution).
Here, when calculating the simple average, it is necessary to convert between a coordinate system based on an angle corresponding to the direction parallel to the edge and the XY coordinate system (pixel coordinate system) of the light receiver 16 (ie, camera). In some cases, data of coordinates that do not match the pixels of the light receiver 16 may be required by the conversion, but in such a case, data interpolated using a general interpolation method may be used.
[0014]
FIGS. 4A and 4B show that the averaging process is not performed in the edge parallel direction (circumferential direction), but simply in the measurement area A of the wafer 1 in the radial direction (perpendicular to the edge parallel direction). 3 is a graph showing an example of a one-dimensional profile obtained by measuring a surface displacement (shape) in a direction (a direction). Here, the position in the radial direction is represented by a distance from an edge of the wafer 1. As shown in FIGS. 4A and 4B, since the back surface of the wafer 1 (the measurement surface 1a) is a rough surface, the measured value (surface displacement) greatly changes due to the unevenness of the rough surface. . Even if ROA (Roll-offAmount) is evaluated based on such surface displacement data, it does not represent a sagging shape, which is a shape change that is slower than the spatial period of unevenness due to surface roughness. It cannot be used for processing control of the wafer 1.
[0015]
On the other hand, FIGS. 4 (c) and 4 (d) show the same measurement area A of the wafer 1 as in FIGS. 4 (a) and 4 (b) in the radial direction (the edge parallel direction) measured by the shape measuring apparatus X. 3 is a graph showing a one-dimensional profile (in a direction perpendicular to FIG. 3).
As shown in FIGS. 4 (c) and 4 (d), each displacement amount (average value of surface displacement) constituting the one-dimensional profile measured by the present shape measuring apparatus X is a plurality of displacements in the direction parallel to the edge. Since the surface displacement data has been subjected to the averaging process, the influence of unevenness due to the surface roughness of the measurement surface 1a has been removed. In addition, since the displacement amount has been averaged in a direction substantially parallel to the edge 1b of the wafer 1, when the averaging process such as moving average is performed in a direction substantially perpendicular to the edge 1b. As described above, the spatial resolution in the direction (radial direction) substantially perpendicular to the end side 1b does not decrease. Therefore, it is possible to accurately measure the sag shape near the edge of the back surface of the wafer 1b. If the ROA is obtained using such a displacement, the ROA can be used for processing control near the edge of the wafer 1 (near the edge).
[0016]
In the example shown in FIG. 3, the width of the measurement area A in the direction parallel to the edge is about 4.9 mm on the innermost side (the position of the marking line (1)) of the wafer 1, and the outermost side (the position on the marking line (1)). It is about 5.2 mm at the marking line (2). Here, when the range of 5 mm is averaged, it acts as a low-pass filter that attenuates components whose spatial period of the measurement data is about 5 mm or less. In consideration of the fact that the spatial period of the irregularities due to the surface roughness of the measurement surface 1a is about 0.1 mm to 0.05 mm, the wafer 1 is subjected to an averaging process over a range of about 5 mm, thereby obtaining the measurement surface 1a. The influence of the irregularities due to the surface roughness 1a can be sufficiently suppressed (cut off).
For example, when the surface roughness is expressed as a sine wave having a spatial period λ _n and an amplitude B (coordinate in the direction in which the averaging process is performed is x), it is expressed by the following equation (3).
(Equation 1)

The result of averaging the measurement data of the measurement surface 1a having this period of surface roughness in a range (section) of 5 mm is expressed by the following equation (4).
(Equation 2)

FIG. 5 is a graph of the equation (4) when B = 1 μm and λ _n = 0.1 mm. As shown in FIG. 5, it can be seen that by performing the averaging process, the influence (fluctuation) of the surface roughness (amplitude B = 1 μm) of the measurement surface 1a can be sufficiently suppressed to about 0.003 times or less. Therefore, the width of the measurement region A in the direction parallel to the end side is a sufficient length for measuring the sag shape of the wafer 1.
The shape of the measurement area A is another shape, such as making the marking lines (3) and (4) parallel lines and making the width of the measurement area A in the direction parallel to the end sides equal. It doesn't matter.
[0017]
In the present shape measuring apparatus X, a one-dimensional profile as shown in FIGS. 4C and 4D is stored in the storage unit of the arithmetic unit 31 and displayed on a display device such as a CRT.
In this embodiment, an example is shown in which the data of the surface displacement is simply averaged in the direction parallel to the edge. However, the present invention is not limited to this. For example, median, median, or Gaussian or double Gaussian is used. Other averaging processing such as weighted averaging may be performed. In particular, filtering by Gaussian or double Gaussian is preferable because it has a characteristic of more excellent high frequency component attenuation characteristics than the simple averaging described above.
In addition, the measurement of the two-dimensional surface displacement is not limited to the use of the grazing incidence interferometer, and may use other measurement devices and methods such as the use of a Fizeau interferometer or a Michelson interferometer. .
[0018]
【The invention's effect】
As described above, according to the present invention, for each position in a direction substantially perpendicular to an edge in a predetermined measurement region of a measured object, a displacement amount obtained by averaging the surface displacement is calculated, and the displacement amount is calculated. Represents a one-dimensional profile in a direction substantially perpendicular to the edge, so that the shape represented by this one-dimensional profile is free from the influence of irregularities due to the surface roughness of the measurement area. In addition, since the displacement is averaged in a direction substantially parallel to the edge of the measured object, the spatial resolution in a direction substantially perpendicular to the edge is not reduced. Therefore, it is possible to accurately measure a shape in which the influence of surface roughness such as a sagging shape near the edge of the back surface of the wafer is removed.
In addition, the configuration in which the two-dimensional surface displacement of the measurement area is calculated from the interference image by the interferometer enables optically accurate measurement to be performed in a short time.
[Brief description of the drawings]
FIG. 1 is a diagram showing a schematic configuration of a shape measuring apparatus X according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating a state in which a calibration wafer for setting a measurement area of a wafer to be measured in shape by the shape measuring apparatus X according to the embodiment of the present invention is mounted on a stage.
FIG. 3 is a diagram illustrating a plane and a side surface in a state where a wafer is mounted on a stage of the shape measuring apparatus X according to the embodiment of the present invention.
FIG. 4 is a graph showing a one-dimensional profile near a wafer edge which is a result of each of the shape measurement by the shape measuring apparatus X according to the embodiment of the present invention and the shape measurement by the conventional method.
FIG. 5 is a graph showing a result of performing averaging processing in a range of 5 mm when the surface roughness of a measurement surface is represented by a sine wave having a spatial period of 0.1 mm and an amplitude of 1 μm.
[Explanation of symbols]
1 ... Wafer (object to be measured)
1a: Measurement surface (rough surface)
1b Wafer edge 2 Calibration wafer 2a Calibration wafer edge 10 Oblique incidence interferometer 11 Light emitter 12 Measurement light 14, 14 'Collimator lens 15 Triangular prism 15a Reference plane 16 Receiver 17 Stage 18 Pin 19 Piezo actuator 31 Calculator (primary profile calculating means)
A: Measurement area

Claims

In a shape measuring apparatus for measuring a surface shape in a direction substantially perpendicular to an edge of the measured object in a predetermined measurement region of the measured object,
Surface displacement measuring means for measuring a two-dimensional surface displacement of the measurement area;
One-dimensional profile calculating means for calculating a one-dimensional profile in a direction substantially perpendicular to the edge by performing a predetermined averaging process on the measured surface displacement in a direction substantially parallel to the edge;
A shape measuring device characterized by comprising:

2. The shape measuring apparatus according to claim 1, wherein a width of the measurement area in a direction substantially parallel to the end side is sufficiently longer than a spatial period of unevenness due to surface roughness of the measurement area.

The surface displacement measuring means comprises: an interferometer for detecting an interference image on the surface of the measurement area; and a surface displacement calculating means for calculating the two-dimensional surface displacement based on the detected interference image. Item 3. The shape measuring device according to any one of Items 1 and 2.

In a shape measuring method for measuring a surface shape in a direction substantially perpendicular to an edge of the measured object in a predetermined measurement region of the measured object,
A surface displacement measurement procedure for measuring a two-dimensional surface displacement of the measurement area;
A one-dimensional profile calculating step of calculating a one-dimensional profile in a direction substantially perpendicular to the edge by performing a predetermined averaging process on the measured surface displacement in a direction substantially parallel to the edge;
A shape measuring method characterized by comprising: