JP3694662B2 - Method and apparatus for measuring film throughput in semiconductor element manufacturing process, method and apparatus for processing material using the same, and method for determining end point of process using the same - Google Patents

Description

ここで、係数ｂ，ａは、サンプリング周波数及びカットオフ周波数により数値が異なる。また、このデジタルフィルタの係数値は段差に関する波長域（第１波長帯域）、例えば、２７５ｎｍから５００ｎｍとマスク材の削れ（残膜厚）に関する波長域（第２波長帯域）、例えば、５２５ｎｍから７５０ｎｍとで異なっていても良い。例えば、段差に関する波長域の場合、ａ２＝−１.１４３ ，ａ３＝０.４１２８，ｂ１＝０.０６７４５５，ｂ２＝０.１３４９１，ｂ３＝0.067455（サンプリング周波数１０Ｈｚ，カットオフ周波数１Ｈｚ）であり、マスク材の残膜厚に関する波長域の場合、ａ２＝−０.０００７３６１２，ａ３＝0.17157,ｂ１＝０.２９２７１，ｂ２＝０.５８５４２，ｂ３＝０.２９２７１（カットオフ周波数０.２５Ｈｚ）である。
【００４７】
２次微係数値の時系列データｄｉ，ｄ′ｉは、微分器１３，２３それぞれにより５点の時系列データＹｉの多項式適合平滑化微分法を用いて式（２）から以下のように算出される。

ここで、重みｗに関して、ｗ−２＝２，ｗ−１＝−１，ｗ０＝−２，ｗ１＝−１，ｗ２＝２、である。
なお、微分器１３，２３の間で、ｊは同一又は異なって良い。
【００４８】
前記微係数値の時系列データｄｉ，ｄ′ｉを用いて、平滑化微係数時系列データＤｉ，Ｄ′ｉはデジタルフィルタ回路１４，２４（２次バタワース型のローパスフィルタ）により式（３），（４）により求められる。但し、デジタルフィルタ回路１４，２４の間で係数ａ，ｂの値は異なっても良い。

【００４９】
このようにして、図１のエッチング量測定装置によれば、図３（ａ），図３（ｂ）にＰＡ，ＰＢ，ＰＣ，Ｐａ，Ｐｂ，Ｐｃとして示したような、複数波長についての微分値の標準パターンを少くとも１つ設定し、被処理材の干渉光の複数波長の強度をそれぞれ測定し、該測定された干渉光強度の各波長の微分値の実パターンを求め、標準パターンと微分値の実パターンとを比較することにより、段差とマスク材の残膜厚を求めることができる。例えば、シリコンエッチング深さ３３５ｎｍすなわち図２のＢの位置を検出したい場合には、予め、エッチング量（段差，マスク材の残膜厚）Ｂ，ｂに対応する複数波長についての微分値の標準パターンＰＢ，Ｐｂを設定し、複数の波長において実パターンのこれら標準パターンに対する一致率が、それぞれ判定値σ_S0，σ_M0以内に達したことにより、マスク材表面からの段差４４が４００ｎｍで、マスク材の残膜厚４２が６５ｎｍ（被処理材のシリコン深さ４３が３３５ｎｍ）になったことを検出できる。標準パターンとしては、一次微分値パターン，二次微分値パターンのいずれか一方あるいは両方を用いればよい。この実施例によれば、被処理材のエッチング量（段差，マスク材残膜厚）を求めることによりシリコンエッチング深さが例えば335ｎｍであることを正確に測定することができる。
【００５０】
次に、計測される干渉光強度にノイズ成分を多く含む場合に、測定するエッチング量の精度を向上するようにした第１実施例の変形例を、該変形例の構成を示す図５のブロック図と図６のフローチャートにより説明する。この変形例は、例えば、ウェハ毎にそのパターンが異なり、従ってウェハ毎にエッチング条件（例えば、放電条件）が異なることにより干渉波がウェハ毎に異なる場合に適用される。最初に、被処理材（シリコン，マスク材）の目標加工深さ（ここでは、目標エッチング深さ：図２（ａ）の４３）を設定する（ステップ５５０）。次に、予め保存されている段差，マスク残膜厚に関する微分波形パターンと収束判定値（段差：Ｐ_Sｊ，σ_S0、マスク残膜厚：Ｐ_Mｊ，σ_M0）とを微分波形パターンデータベース１６，２６から読出して微分波形比較器１５，２５にそれぞれ設定する。エッチング処理開始に伴い、干渉光のサンプリングを開始する(ステップ502）。次いで、図４のステップ４０４，４１０，４２４−４３０と同様にして、ステップ５０４−５１０，５２４−５３０を実行する。分光器１１からの短波長域と長波長域の光はそれぞれ第１デジタルフィルタ回路１２，２２と微分器１３，２３と第２デジタルフィルタ回路１４，２４を介し平滑化微係数時系列データＤｉ，ｊ、Ｄ′ｉ，ｊを求める。これらの平滑化微係数時系列データＤｉ,ｊ、Ｄ′ｉ,ｊと予め微分波形比較器１５，２５に設定された微分パターンＰ_Sｊ，Ｐ_Mｊをそれぞれ比較し、その時刻での段差値Ｓｉとマスク材削れ量（残膜厚）Ｍｉを算出する（ステップ５１５，５３５）。ここで、σ＞σ_S0(又はσ＞σ_M0)の時は、ステップ５１５（またはステップ５３５）の時刻で得られたＳｉ（またはＭｉ）は収納せず、回帰分析器１９での処理ではこの時刻での残膜厚さデータは除外する。
【００５１】
ステップ５１５，５３５で得られた段差値とマスク材削れ量（残膜厚）は時系列データＳｉ，Ｍｉとしそれぞれデータ記録器１８，２８に収納される。収納された過去の時系列データＳｊ，Ｍｊを用いて、回帰分析器１９，２９により１次回帰直線Ｙ＝Ｘａ＊ｔ＋Ｘｂ（Ｙ：エッチング量（段差値，マスク材残膜厚），ｔ：エッチング時間，Ｘａ：Ｘａの絶対値がエッチング速度，Ｘｂ：初期膜厚）を求め、この回帰直線より現時点のエッチング量（段差：Ｓ，マスク材残膜厚：Ｍ）を算出する（ステップ５１６，５３６）。ここで、エッチング時間，エッチング速度，初期膜厚，残膜厚等はプロセス量（ここではエッチング量）である。
【００５２】
次に、終点判定器３０において、これらエッチング量Ｓ，Ｍにより加工深さ（＝Ｓ−Ｍ）（図２（ａ）の４３）を求め、この値と目標加工深さとを比較し、目標加工深さ以上であれば被処理材のエッチング量が所定値になったものとしてエッチング処理を終了し、その結果を表示器１７に表示する。目標加工深さ以下である場合、ステップ５０４，５２４に戻る。最後に、サンプリング終了の設定を行う（ステップ５１４）。
【００５３】
図７に、上記実施例により行ったシリコン深さの測定結果（ステップ５１６，５３６での算出値）を示す。図には、シリコン加工深さ、及びマスク材からの段差の時間的変化を示しており、ＳＴＩ加工時おけるエッチングの様子が明確にわかる。なお、この実施例では、シリコン深さ５０６ｎｍ（段差５２９ｎｍ）までエッチング処理を行った。
【００５４】
以上述べた本実施例のエッチング量測定装置によれば、半導体デバイスの製造プロセス等における被処理材のエッチング量を正確に測定することができる。従って、このシステムを利用して、被処理材のエッチングを高精度に実施する方法を提供することができる。また、変形例によれば、第１実施例と異なり、標準パターンで設定した被処理材のエッチング量以外の任意のエッチング量を測定できる。
【００５５】
なお、図５の構成では、第１デジタルフィルタ回路，微分器，第２デジタルフィルタ回路，微分波形パターンデータベース，微分波形比較器，回帰分析器からなる構成を、それぞれ第１，第２の波長帯域について別々に設けた。しかし、そのような構成を第１，第２の波長帯域について共通に１つだけ設け、所定時間毎に微係数等を切り換えて、交互に第１，第２の波長帯域について実パターンを得るようにしても良い。
【００５６】
次に、本発明の第２実施例を図８，図９，図１０，図１１を用いて説明する。図８に示す本実施例の構成は、図５の一方の波長帯域についての構成１１−１９と同一であり、終点判定器１３０の動作が図５の終点判定器３０の動作と異なる。エッチング加工する被処理材の構造を図９に示す。このエッチング処理において加工されるポリシリコン５０の領域はマスク材５１（例えば、窒化膜やホトレジスト）の無い部分であり、観測される干渉光はポリシリコン５０表面の反射光９０Ａと下地酸化膜５２からの反射光９０Ｂとの干渉により起こる。この干渉光の計測によりポリシリコン５０のエッチング量（残膜厚５３：下地酸化膜からのポリシリコンの厚み）を測定する方法を図１０のフローチャートにしたがって説明する。
【００５７】
最初に、被処理材（ポリシリコン）の目標残膜値と、微分波形パターンデータベース１６に予め保存されているポリシリコンの膜厚に関する全ての標準微分パターン（Ｐ_Zｊ）と収束判定値σ_Z0を微分波形比較器１５に設定する（ステップ６００）。エッチング処理開始に伴い、干渉光のサンプリングを開始する（ステップ６０２）。分光器１１からの多波長の光はそれぞれ第１デジタルフィルタ１２と微分器１３と第２デジタルフィルタ回路１４を介し、第１実施例のステップ４０４−４１０と同様にして、平滑化微係数時系列データＤｉ，ｊを求める。これらの平滑化微係数時系列データＤｉ，ｊと予め微分波形比較器１５に設定された微分パターンＰ_Zｊ を比較し、その時刻での残膜値Ｚｉを算出する（ステップ６１５）。ここで、σ＞σ_z0の時は、ステップ６１５でこの時刻で得られたＺｉ値は収納せず、回帰分析器１９での処理ではこの時刻での残膜厚さデータは除外する。
【００５８】
ステップ６１５で得られた残膜値は時系列データＺｉとしてデータ記録器１８に収納される。収納された過去の時系列データＺｉを用いて、回帰分析器１９により１次回帰直線Ｙ＝Ｘａ＊ｔ＋Ｘｂ（Ｙ：残膜量，ｔ：エッチング時間，Ｘａ：Ｘａの絶対値がエッチング速度，Ｘｂ：初期膜厚）を求め、この回帰直線より現時点の残膜量Ｚを算出する（ステップ６１６）。
【００５９】
次に、終点判定器１３０において、残膜量Ｚと目標残膜値を比較し、目標残膜以下であれば被処理材のエッチング量が所定値になったものとしてその結果を表示器１７に表示する。目標残膜以上である場合、ステップ６０４に戻る。最後に、サンプリング終了の設定を行う（ステップ６１４）。
【００６０】
ところで、目標残膜値が予め保存されているポリシリコンの膜厚に関する微分波形パターンデータベースＰ_Zｊ より少ないとステップ６１８で判定された場合は、以下のような処理を行いエッチング処理を終了する。残膜量Ｚが保存されているポリシリコンの膜厚に関する微分波形パターンデータベースの最小膜厚Ｙｍと等しくなった時に、上記１次回帰直線Ｙ＝Ｘａ＊ｔ＋Ｘｂより、目標残膜値Ｙ_T になるエッチング時刻（ｔ_T＝［Ｙ_T−Ｘｂ］／Ｘａ）を算出し、この時刻ｔ_T までエッチング処理を継続し行う。
【００６１】
図１１に、本実施例により行ったポリシリコンの残膜測定結果を示す。これは、ポリシリコン膜厚に関する微分波形パターンデータベースには最小膜厚Ｙｍ＝４５ｎｍまでのデータベースを用い目標残膜値Ｙｔ＝２０ｎｍを予測したものであり、１次回帰直線により目標残膜値が２０ｎｍとなるエッチング時刻９６秒が明確にわかる。これにより微分波形パターンデータベースが無い残膜量の終点判定が可能になる。
【００６２】
また、１次回帰直線の傾きの絶対値（＝｜Ｘａ｜）はエッチング速度を示しており、このエッチング速度を量産管理することによりエッチング装置の状態管理が行える。すなわち、エッチング速度がある許容値以内であれば、エッチング装置が正常稼動していることを、また許容値以外である場合は異常であることを監視できる。
【００６３】
さらに、１次回帰直線の切片（＝Ｘｂ）は被処理材の初期膜厚を示しており、この初期膜厚を量産管理することによりエッチング処理前の成膜状態の管理が行える。すなわち、初期膜厚がある許容値以内であれば、成膜装置が正常稼動していることを、また許容値以外である場合は異常であることを監視しフィードバックできる。
【００６４】
次に、本発明の第３実施例を図１２，図１３を用いて説明する。本実施例は、エッチング深さを管理する時、ウェハ毎に有機膜の膜厚に誤差があるため、初期膜厚と測定した残膜厚とからエッチング深さを測定するようにするものである。図１２に示す本実施例の構成は、図８の構成１１−１９と同一であり、終点判定器２３０の動作が図８の終点判定器１３０の動作と異なる。エッチング加工する被処理材の構造を図１３に示す。このエッチング処理において加工される有機膜６０の領域（溝構造）はマスク材６１（例えば、窒化膜やホトレジスト）の無い部分であり、観測される干渉光は有機膜６０表面の反射光と配線材６２（例えば、Ｃｕ）からの反射光との干渉により起こる。
【００６５】
この干渉光の計測により有機膜６０のエッチング量（溝深さ：ＤとＥとの距離６５）を測定する方法を図１４のフローチャートにしたがって説明する。最初に、被処理材（有機膜）の目標深さ値と、微分波形パターンデータベース１６に予め保存されている有機膜の膜厚に関する全ての標準微分パターン（Ｐ_Fｊ ）と収束判定値（σ_F0）を微分波形比較器１５に設定する（ステップ７００）。エッチング処理開始に伴い、干渉光のサンプリングを開始する（ステップ７０２）。分光器１１からの多波長の光はそれぞれ第１デジタルフィルタ回路１２と微分器１３と第２デジタルフィルタ回路１４を介し、第２実施例のステップ６０４−６１０と同様にして、平滑化微係数時系列データＤｉ，ｊを求める。これらの平滑化微係数時系列データＤｉ，ｊと予め微分波形比較器１５に設定された微分パターンＰ_Fｊ を比較し、その時刻での残膜値Ｆｉを算出する（ステップ７１５）。ここで、σ＞σ_F0の時は、この時刻での残膜厚さ値は収納せず、回帰分析器１９での処理ではこの時刻での残膜厚さデータは除外する。
【００６６】
ステップ７１５で得られた残膜値は時系列データＦｉとしてデータ記録器１８に収納される。収納された過去の時系列データＦｊを用いて、回帰分析器１９により１次回帰直線Ｙ＝Ｘａ＊ｔ＋Ｘｂ（Ｙ：残膜量，ｔ：エッチング時間，Ｘａ：絶対値がエッチング速度，Ｘｂ：初期膜厚）を求め、この回帰直線より現時点の残膜量Ｆと初期膜厚Ｘｂを算出する（ステップ７１６）。
【００６７】
次に、終点判定器２３０において、現時点の溝深さ（＝Ｘｂ−Ｆ）（図１３の６５）を残膜量Ｆ（図１３の６４）と初期膜厚Ｘｂ（図１３の６３）より求め、この溝深さと目標深さ値と比較し、目標深さ以上であれば被処理材のエッチング量が所定値になったものとしてその結果を表示器１７に表示する。目標深さ以下である場合、ステップ７０４に戻る。最後に、サンプリング終了の設定を行う（ステップ７１４）。このように、溝加工時のエッチング深さは残膜量Ｆと初期膜厚Ｘｂを回帰分析により求めることにより測定が可能である。
【００６８】
次に、本発明の第４実施例を図１５，図１６，図１７を用いて説明する。図１５に示す本実施例の構成は、図１２の構成１１−１９と同一であり、終点判定器３３０の動作が図１２の終点判定器２３０の動作と異なり、更に制御装置1000が設けられている。同一のエッチング装置１を用いて、種々膜質の異なる被エッチング材をエッチング処理することがしばしばある。この場合、制御装置1000に予め設定されたエッチング条件（例えば、エッチングガス条件やプラズマ発生電力条件やバイアス条件など）に基づき、制御装置１０００によりガス供給装置１００１やプラズマ発生装置１００２やウェハバイアス電源１００３を時間制御しエッチング処理が行われる。しかしながら、エッチング加工する被処理材が半導体素子構造により図１６に示すような積層構造になっている場合、エッチング処理が複雑となり、単純な時間制御によるエッチング処理ではダメージのない素子加工が困難になる。図１６を用いて、このような積層構造素子のエッチング加工について説明する。このエッチング処理において加工されるポリシリコン膜７０の上には、ＢＡＲＣ７３（Back Anti-Reflection Coating）やマスク材７１（例えば、窒化膜やホトレジスト）が、またポリシリコン膜７０の下には、下地酸化膜７２が形成されている。さらに、下地酸化膜７２はトランジスタのゲート電極部７８の厚み（例えば、約２ｎｍ）とトランジスタ素子を分離するための溝部７９（ＳＴＩ）の厚み（例えば、約３００ｎｍ）とが大きく異なっている構造となっている。この構造の加工においては、まず、ＢＡＲＣ７３のエッチング処理を行い、続いてポリシリコン膜７０のエッチング処理を同一のエッチング装置で一貫処理する。このようなエッチング処理では、それぞれの膜を適切にエッチングできないとゲート電極部７８の下地酸化膜７２を削り過ぎてしまい、素子にダメージを与える。そのため、まずＢＡＲＣ７３エッチング処理において、ポリシリコン膜７０をできる限り削らない様にエッチングを制御する必要がある。このことは、ＢＡＲＣ７３エッチング処理においてＢＡＲＣの残膜量を測定し、僅かな残膜量７５（例えば、２０ｎｍ）になった時点で、エッチング条件をポリシリコンが削れ難い条件に変更し残っているＢＡＲＣ材をエッチングすることが重要である。次に、ポリシリコン膜７０のエッチング処理においては、ポリシリコンの残膜量（残膜厚）を測定し、僅かな残膜量（残膜厚）７７（例えば、２０ｎｍ）になった時点で、エッチング条件を下地酸化膜が削れ難い条件に変更し残っているポリシリコン材をエッチングすることが重要である。
【００６９】
ＢＡＲＣの残膜厚さ量測定に用いる光は、ＢＡＲＣ表面からの反射光とポリシリコン境界面からの反射光との干渉光９６を利用する。また、ポリシリコンの残膜厚さ量測定に用いる光は、ポリシリコン表面からの反射光と下地酸化膜境界面からの反射光との干渉光９５Ａ又は９５Ｂを利用する。この際、下地酸化膜の厚みがゲート電極部７８の厚み７７とトランジスタ素子を分離するための溝部７９の厚み７６とが異なっているため、それぞれの部分からの干渉光強度９５Ａ，９５Ｂは異なっており、素子分離部７９からの干渉光強度９５Ｂの方がゲート電極部７８からの干渉光強度９５Ａに比べ大きい。従って、ポリシリコンの残膜厚さ測定においては、素子分離部７９上のポリシリコンを対象として行う。つまり、ポリシリコンのエッチング処理ではこの点を考慮し干渉光強度９５Ｂを用いて、ポリシリコンを残膜厚さが厚み７６になるまでエッチング処理を行い、その後、下地酸化膜が削れ難いエッチング条件でのポリシリコン材のエッチング処理を行う。
【００７０】
このエッチング処理を行う手順を図１７のフローチャートにしたがって説明する。最初に、制御装置１０００に積層膜（例えば、ＢＡＲＣ７３とポリシリコン膜７０）に対するエッチング条件(例えば、ガス条件，放電条件，圧力条件など)と各膜７３，７０の目標残膜厚さ値７５，７６と収束判定値が設定される（ステップ８００）。次に、それぞれの膜種に応じて、微分波形パターンデータベース１６に予め保存されている各膜７３，７０の膜厚に関する全ての標準微分パター（Ｐ_Zｊ ）と、収束判定値（σ_Z0）を微分波形比較器１５に設定される（ステップ８０１）。次のステップにおいてエッチング処理の開始と干渉光のサンプリングを開始する（ステップ８０２）。次に、制御装置１０００からの指示により第１に処理される被エッチング材（例えば、ＢＡＲＣ材）に関する標準微分パターンＰ_Zｊ と収束判定値σ_Z0が微分波形パターンデータベース１６から微分波形比較器１５に設定される（ステップ８０３）。分光器１１からの多波長の光はそれぞれ第１デジタルフィルタ回路１２より平滑化時系列データＹｉ，ｊを求める（ステップ８０４）。さらに、微分器１３と第２デジタルフィルタ回路１４を介し、第３実施例のステップ７０４−７１０と同様にして、平滑化微係数時系列データＤｉ，ｊを求める（ステップ８０６，８０８）。これらの平滑化微係数時系列データＤｉ，ｊと予め微分波形比較器１５に設定された微分パターンＰ_Fｊ を比較し、収束値σ＝Σ(Ｄｉ，ｊ−Ｐ_Zｊ)²が最も小さい収束値に対する残膜厚さを求める。この時、σ≦σ_Z0である場合は求めた残膜厚さをこの時刻での残膜厚さ値Ｚｉとしデータ記録器１８に収納する（ステップ８１０，８１５）。σ＞σ_Z0である場合、この時刻での残膜厚さ値は収納せず、回帰分析器１９での処理ではこの時刻での残膜厚さデータは除外する(ステップ８１５)。収納された過去の時系列データＺｉを用いて、回帰分析器１９により１次回帰直線Ｙ＝Ｘａ＊ｔ＋Ｘｂを求め、この回帰直線より現時点の残膜厚さ量Ｚを算出する（ステップ８１６）。
【００７１】
次に、終点判定器３３０において、現時点の残膜厚さ量Ｚと制御装置１０００からの目標残膜厚さ値７５と比較し（例えば、ＢＡＲＣ残膜厚さ２０ｎｍ）、目標残膜値７５以下であれば被処理材のエッチング量が所定値になったものとしてその結果を表示器１７に表示すると、同時にエッチング条件を切り換えて次のエッチング処理（ポリシリコンが削れにくいＢＡＲＣエッチング条件）を行う（ステップ８１８）。目標残膜値以上である場合、ステップ８０４に戻る。切り換えた次のエッチング処理（ポリシリコンが削れにくいＢＡＲＣエッチング条件）は例えば、制御装置１０００によりウェハバイアス電源１００３を制御し、残膜部をエッチングし、予め設定された時間が経過後に、エッチング条件の切り換えを行う、すなわち、次の膜種のエッチングのためにエッチング条件を切り換える必要があるかどうかを判定する（ステップ８１９）。
【００７２】
次の膜種（例えば、ポリシリコン膜７０）のエッチングでは、予め設定されたエッチング条件にガス供給装置１００１やプラズマ発生装置１００２やウェハバイアス電源１００３を制御すると共に、次の膜種（例えば、ポリシリコン）の残膜厚さ７６に関する標準微分パターン(Ｐ_Zｊ，σ_Z0）を微分波形比較器１５に設定し（ステップ８０３）、再びステップ８０４から８１８を実行する。終点判定器３３０において、ポリシリコンの残膜厚さ量Ｚが目標残膜厚さ値７６となった（例えば、ポリシリコン残膜厚さ２０ｎｍ）ことを判定し（ステップ８１８）、その結果を表示器１７に表示すると、同時にエッチング条件を切り換えて次のエッチング処理（下地酸化膜が削れにくいポリシリコンのエッチング条件）を行う（ステップ８１８）。このエッチング処理（下地酸化膜が削れにくいポリシリコンのエッチング条件）は時間制御によりポリシリコンの残膜部をエッチングし、予め設定された時間が経過後に、エッチング条件切り換えの判定を行う(ステップ８１９)。次のエッチング処理が無い場合は、エッチング処理の終了及びサンプリング終了の設定を行う（ステップ８１４）。
【００７３】
以上のように、第２実施例による上記の方法では、被処理材の残膜厚さに関する微分パターンデータベースが十分になくとも、残膜厚さを予測することができる。さらに、残膜厚さに関する微分パターンデータベースの作成において、サンプルウェハを完全にエッチングする必要が無くなり、サンプルウェハを削減することができる。
【００７４】
また、本発明の回帰分析によりエッチング速度が求められるので、エッチング速度のウェハ毎の監視により量産管理が行え、製品ウェハの不良発生を未然に防止できる。
【００７５】
また、本発明の回帰分析により被処理材の初期膜厚が算出できるので、その結果を成膜装置にフィードバックすることにより、成膜装置とエッチング装置との一貫処理の量産管理が行える。
【００７６】
第３実施例による方法によれば、被処理材の初期膜厚にバラツキがあっても、精度良くエッチング深さを算出でき、目標とするエッチング深さを正確に判定するエッチング処理が行える。
【００７７】
また、第４実施例の方法によれば、積層膜構造を持った被処理材のエッチング処理において、それぞれの膜をエッチングする時に残膜厚さ測定を行い、所定の残膜厚さにおいて、エッチング条件を切り換えることにより、下地膜のエッチング削れを僅かにすることができる。
【００７８】
また、ゲート電極のポリシリコンエッチングにおいては、素子分離部からの干渉光によりポリシリコン残膜厚さを測定することにより、正確に所定残膜厚さを判定でき、ゲート電極部の下地酸化膜を過剰にエッチングすることが無く、ウェハの不良処理枚数を最小限に抑えることができる。
【００７９】
なお、上記の各実施例においては、光源を有する分光器から多波長の放射光を放出し、被処理材からの反射光の干渉光を利用して測定を行う。一方、光源を有しない分光器を用い、プラズマにより放出される多波長の放射光を光源として利用するようにしても良い。
【００８０】
なお、本発明は、さらに次の特徴を有する。
（１）試料のエッチング途中に測定したデータを蓄積し、該蓄積された過去データを用いて前記エッチングにおけるエッチング深さを予測し、所定のエッチング深さで前記エッチングを止めることを特徴とするエッチング方法。
（２）下地膜を有さない被エッチング材のエッチング方法において、前記被エッチング材の上に形成されたマスクの削れ量と前記マスク上面からの前記被エッチング材のエッチング底面までのエッチング段差を測定し、前記被エッチング材のエッチング深さを管理することを特徴とするエッチング方法。
（３）試料のエッチング途中の干渉光データを用いて、前記試料の被エッチング部の残膜厚さを予測し、所定の残膜厚さで前記エッチングを止めることを特徴とするエッチング方法。
（４）ダマシンエッチングにおいて、前記エッチング途中の干渉光データを用いて、初期膜厚を予測し溝深さを決め、所定の溝深さで前記エッチングを止めることを特徴とするエッチング方法。
（５）面内厚さが異なる下地膜上に形成されたゲート材をエッチングする方法において、前記下地膜の厚さの厚い部分で干渉光を測定し、前記厚さの厚い下地膜の上に形成されたゲート材の残膜厚さを測定し、前記ゲート材の膜厚を管理することを特徴とするエッチング方法。
（６）複数の膜が積層された被エッチング材のエッチングにおいて、前記被エッチング材からの干渉光を測定し、前記膜毎にデジタルフィルタを切り換えて干渉光のデータを処理し、前記膜毎に膜厚を管理することを特徴とするエッチング方法。
（７）ＢＡＲＣエッチングにおいて、エッチング中の被処理材から測定した干渉光を用いて前記ＢＡＲＣの膜厚を管理し、下地の被エッチング材の削れを防止することを特徴とするエッチング方法。
（８）ＳＴＩ部を有する半導体素子の製造において、前記半導体素子の一部を形成するＰｏｌｙ−Ｓｉのエッチングを行う際に、前記ＳＴＩ部上に形成されたＰｏｌｙ−Ｓｉの残膜厚さで、前記Ｐｏｌｙ−Ｓｉのエッチングを管理することを特徴とする半導体装置の製造方法。
【００８１】
【発明の効果】
本発明によれば、エッチング深さや残膜量といった処理の量をより精密に検出し、あるいは制御できる半導体ウエハの処理方法および処理装置を提供できる。
【図面の簡単な説明】
【図１】本発明の第１実施例によるエッチング量測定装置を備えた半導体ウェハのエッチング装置の全体構成を示すブロック図である。
【図２】（ａ）は、エッチング処理途中の被処理材の縦断面形状を示す図であり、(ｂ)，（ｃ）はそれぞれ異なる波長の干渉光の波長実パターンの例を示す図である。
【図３】（ａ），（ｂ）は、図２（ａ），図２（ｂ）のＡ，Ｂ，Ｃに示す各段差（マスク材表面からの距離）及びａ，ｂ，ｃに示すマスクの各残膜厚さに対応する、干渉光の微係数値時系列データの波長をパラメータとする図である。
【図４】図１のエッチング量測定装置でエッチング処理を行う際に、被処理材の段差及びマスク残膜厚さを求める手順を示すフローチャートである。
【図５】本発明の第１の実施例の変形例によるエッチング深さ測定装置を備えた半導体ウェハのエッチング装置の全体構成を示すブロック図である。
【図６】図５の実施例の動作を示すフローチャートである。
【図７】図５の実施例のエッチング深さ測定結果を示す図である。
【図８】本発明の第２の実施例による残膜厚さ測定装置を備えた半導体ウェハのエッチング装置の全体構成を示すブロック図である。
【図９】エッチング処理途中の被処理材の縦断面形状を示す図である。
【図１０】図８の実施例の動作を示すフローチャートである。
【図１１】図８の実施例のポリシリコン残膜厚さの測定結果及び回帰直線を示す図である。
【図１２】本発明の第３の実施例によるエッチング深さ測定装置を備えた半導体ウェハのエッチング装置の全体構成を示すブロック図である。
【図１３】エッチング処理途中の被処理材の縦断面形状を示す図である。
【図１４】図１２の実施例の動作を示すフローチャートである。
【図１５】本発明の第４の実施例によるエッチング残膜測定装置を備えた半導体ウェハのエッチング装置の全体構成を示すブロック図である。
【図１６】エッチング処理途中の被処理材の縦断面形状を示す図である。
【図１７】図１５の実施例の動作を示すフローチャートである。
【符号の説明】
１…エッチング装置、２…真空容器、３…プラズマ、４…被処理材、５…試料台、８…光ファイバー、９…放射光、１０…測定装置、１１…分光器、１２，２２…第１デジタルフィルタ回路、１３，２３…微分器、１４，２４…第２デジタルフィルタ回路、１５，２５…微分波形比較器、１６，２６…微分波形パターンデータベース、１７…表示器、１８，２８…時系列データ記録器、１９，２９…回帰分析器、３０，１３０，２３０，３３０…終点判定器、４０…シリコン、４１…窒化膜、５０…ポリシリコン、５１，６１，７１…マスク材、５２…下地酸化膜、６０…有機膜、６２…配線材、７０…ポリシリコン膜、７２…下地酸化膜、７３…ＢＡＲＣ、７８…ゲート電極部、７９…溝部、１０００…制御装置、１００１…ガス供給装置、１００２…プラズマ発生装置、１００３…ウェハバイアス電源。[0001]
BACKGROUND OF THE INVENTION
  The present invention relates to a semiconductor wafer for manufacturing a semiconductor device.In particular, the amount of processing performed on the film of the material to be processed formed on the surface of the semiconductor wafer to be processed is analyzed by analyzing the light from the semiconductor wafer. The present invention relates to a semiconductor wafer processing method and a processing apparatus for detecting a processing state.
[0002]
[Prior art]
In the manufacture of semiconductor wafers, dry etching is widely used to remove or pattern various layers of material formed on the surface of the wafer, and in particular, layers of dielectric material. The most important for process parameter control is to accurately determine the etch endpoint to stop etching at the desired etch depth and thickness during processing of such a layer.
[0003]
During the dry etching process of the semiconductor wafer, the emission intensity of the specific wavelength in the plasma light changes with the progress of the etching of the specific film. Therefore, as one of the methods for detecting the etching end point of a semiconductor wafer, conventionally, a change in emission intensity of a specific wavelength from plasma is detected during the dry etching process, and the etching end point of a specific film is detected based on the detection result. There is a way to do it. At that time, it is necessary to prevent erroneous detection based on the fluctuation of the detected waveform due to noise. As a method for accurately detecting a change in emission intensity, for example, Japanese Patent Application Laid-Open Nos. 61-53728 and 63-200533 are known. In Japanese Patent Laid-Open No. 61-53728, noise is reduced by a moving average method, and in Japanese Patent Laid-Open No. 63-200533, noise is reduced by approximation processing by a first-order least square method.
[0004]
Along with the recent miniaturization and high integration of semiconductors, the aperture ratio (area to be etched of a semiconductor wafer) has decreased, and the emission intensity of specific wavelengths of reaction products taken into the photodetector from the optical sensor has become weak. ing. As a result, the level of the sampling signal from the photodetector is reduced, and it is difficult for the end point determination unit to reliably detect the etching end point based on the sampling signal from the photodetector.
[0005]
In addition, when detecting the end point of etching and stopping the process, it is actually important that the remaining thickness of the dielectric layer is equal to a predetermined value. Conventional processes monitor the entire process using time thickness control techniques based on the assumption that the etch rate of each layer is constant. The value of the etching rate is obtained by processing a sample wafer in advance, for example. In this method, the etching process is stopped simultaneously with the elapse of time corresponding to a predetermined etching film thickness by the time monitoring method.
[0006]
However, actual films such as SiO formed by LPCVD (Iow Pressure Chemical Vapor Deposition) technique₂ Layers are known to have low thickness reproducibility. Thickness tolerance due to process variations during LPCVD is SiO₂ This corresponds to about 10% of the initial thickness of the layer. Therefore, the method by time monitoring is the SiO remaining on the silicon substrate.₂ The actual final thickness of the layer cannot be measured accurately. The actual thickness of the remaining layer is then measured by a standard spectroscopic interferometer technique, and if it is found to be over-etched, the wafer is discarded as rejected. become.
[0007]
In addition, in an insulating film etching apparatus, a change over time is known such that the etching rate decreases as etching is repeated. In some cases, etching may stop in the middle, and the solution is essential. In addition, it is important for the stable operation of the process to monitor the variation of the etching rate over time. However, the conventional method only monitors the time for determining the end point. There was no appropriate way to deal with major changes and fluctuations. In addition, when the etching time is as short as about 10 seconds, the end point determination must be an end point determination method that shortens the determination preparation time, and the step of the determination time needs to be sufficiently shortened, but is not necessarily sufficient. Furthermore, in the insulating film, the area to be etched is often 1% or less, and the plasma emission intensity change from the reaction product generated by etching is small. Therefore, an end point determination system that can detect even a slight change is required, but no practical and inexpensive system is found.
[0008]
On the other hand, other methods for detecting the etching end point of a semiconductor wafer were disclosed in JP-A-5-179467, JP-A-8-274082, JP-A-2000-97648, JP-A-2000-106356, and the like. A method using an interferometer is also known. In this interferometer usage, monochromatic radiation emitted from a laser is applied at a normal incidence angle to a wafer containing a laminated structure of dissimilar materials. For example, Si_ThreeN_FourSiO on the layer₂ In the case where the layer stack is stacked, SiO₂ Synchrotron radiation reflected from the top surface of the layer and SiO₂ Layer and Si_ThreeN_FourInterference fringes are formed by the radiation reflected from the interface formed between the layers. The reflected radiation is applied to a suitable detector, which is the SiO during etching.₂ Generates a signal whose strength varies with layer thickness. During the etching process, SiO₂ As soon as the top surface of the layer is exposed, the etch rate and current etch thickness can be monitored continuously and accurately. There is also known a method of measuring a predetermined radiation emitted by plasma instead of a laser by a spectrometer.
[0009]
[Problems to be solved by the invention]
The contents of the above-mentioned conventional documents are summarized as follows.
In Japanese Patent Laid-Open No. 5-179467, interference light (plasma light) is detected using three types of color filters of red, green, and blue, and etching end point detection is performed.
[0010]
In Japanese Patent Laid-Open No. 8-274082 (US Pat. No. 5,658,418), the extreme values of the interference waveform (maximum and minimum of the waveform: zero pass point of the differential waveform) are obtained by using the temporal change of the interference waveform of two wavelengths and the differential waveform. ). The etching rate is calculated by measuring the time until the count reaches a predetermined value, the remaining etching time until the predetermined film thickness is reached is obtained based on the calculated etching rate, and the etching process is stopped based on the etching time.
[0011]
Japanese Patent Laid-Open No. 2000-97648 discloses a waveform (wavelength as a parameter) of the difference between the light intensity pattern of interference light before processing (with wavelength as a parameter) and the light intensity pattern of interference light after processing or during processing. And the step (film thickness) is measured by comparing the difference waveform with the difference waveform stored in the database.
[0012]
Japanese Patent Laid-Open No. 2000-106356 relates to a spin coater and determines the film thickness by measuring the temporal change of interference light over multiple wavelengths.
[0013]
In US Pat. No. 6,081,334, characteristic behavior of interfering light with respect to time is obtained by measurement and stored in a database, and the end of etching is determined by comparing the database with the measured interference waveform. This determination prompts a change in the etching process conditions.
[0014]
In the above known example, the following problems occur.
(1) When etching using a mask material (for example, resist, nitride film, oxide film) is performed, interference light from the mask material is superimposed on interference light from the material to be etched.
(2) In the etching process of the material to be processed (for example, silicon and the mask material provided thereon) in the etching process, the mask material is etched together with the silicon. There is a possibility that the etching amount of silicon cannot be measured accurately only by aiming at the depth.
(3) In processing wafers for mass production processes, the initial thickness of the mask material and the initial thickness of the material to be etched have a distribution within the wafer surface due to the device structure, so interference light from different film thicknesses is superimposed. The
[0015]
From the above points, it is possible to accurately measure and control the layer to be processed (layer to be subjected to semiconductor process processing), in particular, the etching depth and the amount of remaining film in the plasma etching process with the required measurement accuracy. It was difficult.
[0016]
  The object of the present invention is to overcome the problems of the prior art.In view of the above, it is an object of the present invention to provide a semiconductor wafer processing method and a processing apparatus capable of more accurately detecting or controlling the amount of processing such as etching depth and amount of remaining film.
[0020]
[Means for Solving the Problems]
  In order to achieve the above object, the inventors obtain a time derivative of the interference waveform for each of a plurality of wavelengths of the interference light obtained from the surface of the semiconductor wafer, obtain a differential value pattern of the interference waveform based thereon, The amount of processing is measured using a pattern.
[0021]
  In the present invention, the interference waveformThe reason for using the time differential value of is as follows.Since the measurement is based on in-situ (real time) measurement during etching, the film thickness of the film to be processed changes every moment. Therefore, time differentiation processing of the interference waveform is possible. Further, the noise of the interference waveform can be removed by this differentiation process. In addition, the refractive index of the material to be etched (for example, silicon and the mask material nitride film) is different with respect to the wavelength. Therefore, characteristic changes of each substance (depending on the film thickness) by interference light measurement over multiple wavelengths. Can be detected.
[0022]
  More specifically, the object is a semiconductor wafer processing method for processing a film, which is a material to be processed, formed on the surface of the semiconductor wafer, and a plurality of methods obtained from the semiconductor wafer surface during the processing. Detecting a differential value of each intensity of the interference light of the wavelength, and the intensity of the interference light of the plurality of wavelengths with respect to a predetermined amount of the processing obtained in advance for a semiconductor wafer having substantially the same configuration as the semiconductor wafer. By a method for processing a semiconductor wafer, comprising: detecting a processing amount for the film being processed based on differential value data and a set of differential values of the detected intensity for the plurality of wavelengths. Achieved.
[0023]
  Further, there is provided a semiconductor wafer processing apparatus for performing a predetermined processing on a semiconductor wafer to be processed disposed on a sample stage in a vacuum vessel using plasma formed in the vacuum vessel, wherein the processing is performed during the processing. A measuring device that receives interference light of a plurality of wavelengths from the surface of the semiconductor wafer, detects a differential value of each intensity of the interference light of the plurality of wavelengths received by the measuring device, and is substantially the same as the semiconductor wafer; Using the differential value data of the interference light intensity of the plurality of wavelengths with respect to the predetermined amount of the processing obtained in advance for the semiconductor wafer having the configuration, and the set of the differential value of the intensity related to the detected plurality of wavelengths This is achieved by a semiconductor wafer processing apparatus having a function of detecting the amount of processing for the film being processed.
[0024]
  Further, this is achieved by processing the semiconductor wafer by switching the processing conditions when it is determined that the detected processing amount for the film during the processing reaches a predetermined amount. Furthermore, this is achieved by detecting the remaining thickness of the film as the amount of processing to be detected. Furthermore, the time at which the amount of processing reaches a predetermined amount is determined by regression analysis using the amount of processing obtained before the detection of the differential value sets of the interference light intensities of the plurality of wavelengths. This is achieved by prediction. Furthermore, this is achieved by obtaining the speed of the processing by the regression analysis.
[0025]
DETAILED DESCRIPTION OF THE INVENTION
Examples of the present invention will be described below. In the following embodiments, those having the same functions as those of the first embodiment are denoted by the same reference numerals as those of the first embodiment, and detailed description thereof is omitted. In the following examples, a method for measuring an etching amount (etching depth and film thickness) in an etching process of a material to be processed will be described as a method for determining an end point of a semiconductor element manufacturing process according to the present invention. However, the present invention is not limited to this, and can also be applied to a method of measuring a film formation amount (film formation thickness) or the like in a film formation process such as plasma CVD or sputtering.
[0026]
The first embodiment of the present invention will be described below with reference to FIGS.
In this embodiment, when plasma etching is performed on a material to be processed such as a semiconductor wafer, the differential value of interference light with respect to each etching amount of the sample material to be processed (sample wafer) and the mask material of the processing material. Standard pattern P showing the wavelength dependence of the wavelength (with wavelength as parameter)_S And P_M Set each. Next, the intensity of a plurality of wavelengths of the interference light in the actual processing of the actual processing material (actual wafer) having the same configuration as the sample processing material is measured, and a differential value of the measured interference light intensity. The actual pattern showing the wavelength dependence (with the wavelength as a parameter) is obtained, the standard pattern of the differential value is compared with the actual pattern, and the actual etching amount of the material to be processed (end point of the process) is obtained. .
[0027]
First, the overall configuration of a semiconductor wafer etching apparatus provided with an etching amount (here, actual etching depth and film thickness of processing material) measuring apparatus according to the present invention will be described with reference to FIG. The etching apparatus 1 includes a vacuum container 2. The etching gas introduced into the inside of the etching apparatus 1 is decomposed by microwave power or the like to form plasma, and the plasma 3 causes a material 4 to be processed such as a semiconductor wafer on the sample stage 5 to be processed. Etched. Multi-wavelength radiated light from a measurement light source (for example, a halogen light source) included in the spectroscope 11 of the etching amount (for example, etching depth and film thickness) measuring apparatus 10 is guided into the vacuum vessel 2 by the optical fiber 8 and is covered. It is applied perpendicularly to the treatment material 4 at an incident angle. As shown in FIG. 2A, the material to be processed 4 includes a silicon 40 as a material to be etched and a nitride film 41 as a mask material, and the emitted light is reflected by the mask material. Interference light is formed by the light 9A and the radiated light reflected by the silicon surface without the mask material. That is, the interference light is an interference component due to a step between the mask material and silicon. Reflected light 9A from the mask material is radiated light 9a reflected on the nitride film surface and radiated light 9b reflected on the interface between the nitride film and silicon, and interference light is formed by these radiated light. That is, the interference light is an interference component due to the mask material being scraped. These interference lights are guided to the spectroscope 11 of the etching amount measuring device 10 through the optical fiber 8, and based on the state, the silicon etching depth and the thickness of the mask material are measured, and the end point of the process (here, etching) is determined. Process.
[0028]
The etching amount measuring apparatus 10 includes a spectrometer 11, first digital filter circuits 12 and 22, differentiators 13 and 23, second digital filter circuits 14 and 24, differential waveform pattern databases 16 and 26, and differential waveform comparators 15 and 25. The end point determination unit 30 for determining the end point of etching based on the results of these comparators, and the display unit 17 for displaying the determination result of the end point determination unit 30 are provided.
[0029]
FIG. 1 shows a functional configuration of the etching amount measuring apparatus 10, and the actual configuration of the etching amount measuring apparatus 10 excluding the display unit 17 and the spectroscope 11 includes a CPU, an etching depth, and the like. Consists of a ROM that holds various data such as a film thickness measurement processing program and interference light differential waveform pattern database, a RAM that stores measurement data, and an external storage device, a data input / output device, and a communication control device can do. The same applies to other embodiments such as FIG.
[0030]
The multi-wavelength emission intensities relating to the material to be processed taken in by the spectroscope 11 are converted into voltage signals as current detection signals corresponding to the respective emission intensities. Signals of a plurality of specific wavelengths output as sampling signals by the spectroscope 11 are stored in a storage device such as a RAM (not shown) as time series data yij, y′ij. The first and second wavelength bands of the time series data yij and y′ij are then smoothed by the first digital filter circuits 12 and 22, respectively, and smoothed as time series data Yij and Y′ij. It is stored in a storage device such as. Based on the smoothed time series data Yij and Y′ij, the differentiators 13 and 23 calculate the time series data dij and d′ ij of the differential coefficient values (primary differential value or secondary differential value), respectively. It is stored in a storage device such as. The time series data dij and d'ij of the differential coefficient values are smoothed by the second digital filter circuits 14 and 24, respectively, and stored in a storage device such as a RAM as the smoothed differential coefficient time series data Dij and D'ij. The Then, an actual pattern indicating the wavelength dependence of the differential value of the interference light intensity (using the wavelength as a parameter) is obtained from the smoothed differential coefficient time series data Dij and D′ ij.
[0031]
In order to obtain different actual patterns for the first and second wavelength bands, the apparatus of FIG. 1 is configured as follows. That is, when the first and second wavelength bands are the same, the differential coefficients of the first digital filter circuits 12 and 22 are set to different values. In this case, the values of the differentiators 13 and 23 are set to be the same or different. The differential coefficients of the second digital filter circuits 14 and 24 may be the same or different. On the other hand, when the first and second wavelength bands are different, the differential coefficients of the first digital filter circuits 12 and 22 are set to the same or different values, and the values of the differentiators 13 and 23 are set to be the same or different. The differential coefficients of the digital filter circuits 14 and 24 may be the same or different.
[0032]
In the configuration of FIG. 1, a configuration including a first digital filter circuit, a differentiator, a second digital filter circuit, a differential waveform pattern database, and a differential waveform comparator is provided separately for the first and second wavelength bands, respectively. It was. However, only one such configuration is provided in common for the first and second wavelength bands, and the differential coefficient and the like are switched every predetermined time so that an actual pattern is obtained alternately for the first and second wavelength bands. Anyway.
[0033]
On the other hand, in the differential waveform pattern database 16, the differential waveform pattern data value P of the interference light intensity for the first wavelength band corresponding to the step between the silicon to be processed and the mask material to be measured for the etching amount is stored._Sj is preset. In the differential waveform pattern database 26, the differential waveform pattern data value P of the interference light intensity for the second wavelength band corresponding to the film thickness of the mask material of the material to be processed is stored._Mj is preset. In the differential waveform comparator 15, the actual pattern corresponding to the step and the differential waveform pattern data value P_Sj is compared to determine the step (depth from the surface of the mask material to the bottom of the groove etched with silicon). In the differential waveform comparator 25, the actual pattern corresponding to the film thickness of the mask material and the differential waveform pattern data value P_Mj is compared to obtain the film thickness (residual film thickness) of the mask material. As a result, the etching amount of the material to be processed, that is, the etching depth is obtained and displayed on the display 17.
[0034]
In the present embodiment and each of the following embodiments, only one spectrometer 11 is shown. However, when it is desired to measure and control the in-plane of the material to be processed widely, a plurality of spectrometers 11 are provided. It may be provided.
[0035]
FIG. 2A shows an example of the actual wavelength pattern of interference light, and FIG. 2B and FIG. In FIG. 2A, a processing material (wafer) 4 has a mask material 41 laminated on a silicon substrate 40. In this etching process, the silicon substrate is the material to be etched, and such processing is called, for example, STI (Shallow Trench Isolation) etching for element isolation.
[0036]
The multi-wavelength light emitted from the spectroscope 11 is incident on the material to be processed 4 including the stacked structure of the material to be etched and the mask material at an incident angle. The radiated light 9 guided to the mask material 41 interferes with the radiated light 9 a reflected from the upper surface of the mask material 41 and the radiated light 9 b reflected from the boundary surface formed between the mask material 41 and the silicon substrate 40. Light is formed. The radiated light 9 guided to the etched portion without the mask material 41 includes the radiated light 9B reflected from the upper surface of the material to be etched 40 and the radiated light 9A reflected from the mask material 41 (9a and 9b). Interference light is formed. The reflected positions of the radiated light 9B and 9a change as A (a), B (b), and C (c) as the etching process proceeds. The reflected light is guided to the spectroscope 11 and generates a signal whose intensity varies depending on the thickness of the layer to be etched 40 and the mask material being etched.
[0037]
As shown in FIG. 2B, the raw waveform of the interference light in the long wavelength region (second wavelength band: 700 nm, for example) changes slowly as the etching process proceeds. On the other hand, as shown in FIG. 2C, the raw waveform of the interference light in the short wavelength region (first wavelength band: 300 nm, for example) is changed by superposing a short period wave on a long period wave. This indicates that interference light in the long wavelength region changes in interference components due to the mask material being scraped (a, b, and c surfaces in FIG. 2A), and the material to be etched is not used for interference light in the short wavelength region (for example, 300 nm). This is because a change in the interference component appears due to the level difference 44 between the silicon substrate and the mask material (differences between the a, b, and c surfaces in FIG. 2A and the A, B, and C surfaces). Based on the smoothed time series data Yij and Y′ij of the interference light over multiple wavelengths, the differential value time series data dij and d′ ij of the primary differential value or the secondary differential value are calculated, respectively. FIG. 2 (b) shows the first and second differential values of interference light having a wavelength of 700 nm, and FIG. 2 (c) shows the first and second differential values of interference light having a wavelength of 300 nm. Yes.
[0038]
As is clear from FIGS. 2B and 2C, these differential processes clarify the change in the interference component due to the mask material scraping and the change in the interference component due to the step between the silicon substrate and the mask material. . This is because the refractive index of the material to be etched (for example, the refractive index of the nitride film of silicon and the mask material and the refractive index of the vacuum in the groove portion) differs with respect to the wavelength. In the present invention, paying attention to this fact, the step 44 between the silicon substrate and the mask material is obtained by the interference light in the short wavelength region, and the mask material is scraped by the interference light in the long wavelength region (residual film thickness 41 of the mask material). ) Is accurately measured.
[0039]
FIG. 3A shows the primary differential waveform pattern data of the interference light, and FIG. 3B also shows the secondary differential waveform pattern data.
[0040]
The PA, PB, and PC shown in FIGS. 3A and 3B are etched by A (step 44 = 300 nm), B (step = 400 nm), and C (step = 500 nm) in FIG. The differential waveform pattern data in quantity is shown. Similarly, Pa, Pb, and Pc in FIGS. 3A and 3B are respectively a (residual film thickness of mask material 42 = 95 nm) and b (residual film thickness of mask material = 65 nm) and differential waveform pattern data for each etching amount of c (mask material remaining film thickness = 35 nm) are shown. The silicon etching depth 43 at this time is 205 nm at the position A in FIG. 2A, 335 nm at the position B, and 465 nm at the position C.
[0041]
As is clear from FIGS. 3A and 3B, the primary differential waveform pattern and the secondary differential waveform pattern of the interference light are characteristic patterns for each etching amount of the material to be processed. . Since these patterns change depending on the material of the material to be processed, the data for the various materials required for processing and the range of etching amounts are obtained in advance through experiments, etc., and the primary differential waveform pattern and the secondary differential waveform pattern are obtained. It is good to hold | maintain in a recording device (16, 26) as a standard pattern.
[0042]
Next, a procedure for obtaining the etching amount of the material to be processed when performing the etching process with the etching amount measuring apparatus 10 of FIG. 1 will be described with reference to the flowchart of FIG.
First, the target etching amount (that is, the target step and the target mask remaining film thickness) is set, and the differential pattern P corresponding to the target step and the target mask remaining film thickness is obtained from the pattern databases 16 and 26._Sj, P_Mj and judgment value σ_S0, Σ_M0Is set (steps 400 and 420). That is, the target step is selected from the standard patterns PA, PB, and PC of differential values for a plurality of wavelengths as shown in FIGS. One corresponding differential pattern is set. Similarly, a target mask is selected from standard patterns Pa, Pb, and Pc of differential values for a plurality of wavelengths as shown in FIGS. 3A and 3B, which are stored in the differential waveform pattern database 26 in advance. One differential pattern corresponding to the remaining film thickness is set.
[0043]
In the next step, sampling of interference light (for example, every 0.25 to 0.4 seconds) is started (step 402). That is, a sampling start command is issued with the start of the etching process. The multi-wavelength emission intensity that changes with the progress of etching is detected by a photodetector in the spectroscope 11 as a light detection signal having a voltage corresponding to the emission intensity. The photodetection signal of the spectroscope 11 is digitally converted to obtain sampling signals yi, j, y′i, j.
[0044]
Next, the multi-wavelength output signals yi, j, y′i, j from the spectroscope are filtered by the first-stage digital filter circuits 12 and 22, and time-series data Yi, j, Y′i. , J are calculated respectively (steps 404 and 424). That is, noise is reduced by the first-stage digital filter to obtain smoothed time series data yi.
[0045]
Next, the differentiators 13 and 23 calculate the differential coefficients di, j, d'i, j by the S-G method (Savitzky-Golay method), respectively (steps 406 and 426). That is, the coefficient (primary or secondary) di of the signal waveform is obtained by differential processing (SG method). Further, the smoothing differential coefficient time series data Di, j, D′ i, j are respectively calculated by the second stage digital filter circuits 14 and 24 (steps 408 and 428). Next, in the differential waveform comparator 15, σ = Σ (Di, j−P) regarding the step._Sj)² A value is calculated (step 410). Similarly, in the differential waveform comparator 25, σ ′ = Σ (D′ i, j−P) with respect to the shaving (remaining film thickness) of the mask material._Mj)² A value is calculated (step 430). Further, in the end point determination unit 20, σ ≦ σ_S0And σ ′ ≦ σ_M0Each of these determinations is made (step 412). σ ≦ σ_S0And σ ′ ≦ σ_M0In this case, it is determined that the step and the remaining film thickness of the mask material have reached predetermined values, the etching process is terminated, and the result is displayed on the display unit 17. σ ≦ σ_S0, Σ ′ ≦ σ_M0If any of the above is not satisfied, the process returns to steps 404 and 424. Finally, the end of sampling is set (step 414).
[0046]
Here, calculation of the smoothed differential coefficient time series data Di and D′ i will be described. As the digital filter circuits 12, 22, 14, and 24, for example, a second-order Butterworth low-pass filter is used. The digital filter circuits 12 and 22 have the same configuration, and the coefficients b and a may be the same or different between the digital filter circuits 12 and 22. Only the digital filter circuit 12 will be described here. The smoothed time series data Yi is obtained by the equation (1) using a secondary Butterworth low-pass filter.

Here, the coefficients b and a have different numerical values depending on the sampling frequency and the cutoff frequency. Further, the coefficient value of this digital filter is a wavelength range (first wavelength band) relating to a step, for example, 275 nm to 500 nm, and a wavelength range (second wavelength band) relating to shaving of the mask material (residual film thickness), eg, 525 nm to 750 nm. And may be different. For example, in the case of the wavelength range related to the step, a2 = −1.143, a3 = 0.4128, b1 = 0.067455, b2 = 0.13491, b3 = 0.067455 (sampling frequency 10 Hz, cutoff frequency 1 Hz), In the case of the wavelength range relating to the remaining film thickness of the mask material, a2 = −0.0000073612, a3 = 0.17157, b1 = 0.29271, b2 = 0.58542, b3 = 0.29271 (cutoff frequency 0.25 Hz). .
[0047]
The time-series data di and d′ i of the second derivative values are calculated from the equation (2) as follows using the polynomial-adapted smoothing differentiation method of the 5-point time-series data Yi by the differentiators 13 and 23 respectively. Is done.

Here, with respect to the weight w, w−2 = 2, w−1 = −1, w0 = −2, w1 = −1, and w2 = 2.
Note that j may be the same or different between the differentiators 13 and 23.
[0048]
Using the time-series data di and d′ i of the differential coefficient values, the smoothed differential coefficient time-series data Di and D′ i are expressed by the equation (3) by the digital filter circuits 14 and 24 (second-order Butterworth low-pass filter). , (4). However, the values of the coefficients a and b may be different between the digital filter circuits 14 and 24.

[0049]
In this way, according to the etching amount measuring apparatus of FIG. 1, differentials for a plurality of wavelengths as shown as PA, PB, PC, Pa, Pb, and Pc in FIGS. 3 (a) and 3 (b). Set at least one standard pattern of values, measure the intensities of multiple wavelengths of interference light of the material to be processed, determine the actual pattern of the differential value of each wavelength of the measured interference light intensity, By comparing the actual pattern of the differential value, the step and the remaining film thickness of the mask material can be obtained. For example, when it is desired to detect the silicon etching depth 335 nm, that is, the position B in FIG. 2, a standard pattern of differential values for a plurality of wavelengths corresponding to etching amounts (steps, remaining film thickness of mask material) B and b in advance. PB and Pb are set, and the coincidence rate of the actual pattern with respect to these standard patterns at a plurality of wavelengths is determined by the determination value σ_S0, Σ_M0Therefore, it can be detected that the step 44 from the surface of the mask material is 400 nm and the remaining film thickness 42 of the mask material is 65 nm (the silicon depth 43 of the material to be processed is 335 nm). As the standard pattern, either one or both of the primary differential value pattern and the secondary differential value pattern may be used. According to this embodiment, it is possible to accurately measure that the silicon etching depth is, for example, 335 nm by obtaining the etching amount (step difference, mask material remaining film thickness) of the material to be processed.
[0050]
Next, when the interference light intensity to be measured includes a lot of noise components, a modification of the first embodiment in which the accuracy of the etching amount to be measured is improved, and the block of FIG. 5 showing the configuration of the modification. This will be described with reference to the flowchart of FIG. This modification is applied, for example, when the pattern is different for each wafer, and therefore the interference wave is different for each wafer due to different etching conditions (for example, discharge conditions) for each wafer. First, a target processing depth (here, target etching depth: 43 in FIG. 2A) of a material to be processed (silicon, mask material) is set (step 550). Next, a differential waveform pattern and a convergence determination value (step: P_Sj, σ_S0Mask residual film thickness: P_Mj, σ_M0Are read from the differential waveform pattern databases 16 and 26 and set in the differential waveform comparators 15 and 25, respectively. As the etching process starts, sampling of interference light is started (step 502). Next, steps 504-510 and 524-530 are executed in the same manner as steps 404, 410, and 424-430 in FIG. Light in the short wavelength range and long wavelength range from the spectroscope 11 passes through the first digital filter circuits 12 and 22, the differentiators 13 and 23, and the second digital filter circuits 14 and 24, respectively, and the smoothed differential coefficient time series data Di, Find j, D'i, j. These smoothed differential coefficient time series data Di, j, D′ i, j and the differential pattern P set in advance in the differential waveform comparators 15 and 25._Sj, P_MEach j is compared, and the step value Si and the mask material shaving amount (remaining film thickness) Mi at that time are calculated (steps 515 and 535). Where σ> σ_S0(Or σ> σ_M0), The Si (or Mi) obtained at the time of step 515 (or step 535) is not stored, and the residual film thickness data at this time is excluded in the processing by the regression analyzer 19.
[0051]
The step value and the mask material shaving amount (remaining film thickness) obtained in steps 515 and 535 are stored in the data recorders 18 and 28 as time series data Si and Mi, respectively. Using the stored past time series data Sj, Mj, the regression analyzers 19 and 29 use the linear regression line Y = Xa * t + Xb (Y: etching amount (step value, mask material residual film thickness), t: etching. Time, Xa: Absolute value of Xa is etching rate, Xb: Initial film thickness), and the current etching amount (step: S, mask material remaining film thickness: M) is calculated from this regression line (steps 516, 536) ). Here, the etching time, the etching rate, the initial film thickness, the remaining film thickness, and the like are process amounts (here, the etching amount).
[0052]
Next, in the end point determination device 30, a processing depth (= SM) (43 in FIG. 2A) is obtained from these etching amounts S and M, and this value is compared with the target processing depth, and the target processing is performed. If the depth is greater than or equal to the depth, the etching process is terminated assuming that the etching amount of the material to be processed has reached a predetermined value, and the result is displayed on the display 17. When it is below the target machining depth, the process returns to steps 504 and 524. Finally, the end of sampling is set (step 514).
[0053]
FIG. 7 shows the silicon depth measurement results (calculated values in steps 516 and 536) performed in the above example. In the figure, the silicon processing depth and the temporal change of the step from the mask material are shown, and the state of etching in the STI processing can be clearly seen. In this example, etching was performed to a silicon depth of 506 nm (step difference 529 nm).
[0054]
According to the etching amount measuring apparatus of the present embodiment described above, it is possible to accurately measure the etching amount of the material to be processed in the semiconductor device manufacturing process or the like. Therefore, it is possible to provide a method for performing etching of a material to be processed with high accuracy using this system. Moreover, according to the modification, unlike the first embodiment, it is possible to measure an arbitrary etching amount other than the etching amount of the material to be processed set by the standard pattern.
[0055]
In the configuration of FIG. 5, the configuration including the first digital filter circuit, the differentiator, the second digital filter circuit, the differential waveform pattern database, the differential waveform comparator, and the regression analyzer, respectively, has the first and second wavelength bands. Was provided separately. However, only one such configuration is provided in common for the first and second wavelength bands, and the differential coefficient and the like are switched every predetermined time so that an actual pattern is obtained alternately for the first and second wavelength bands. Anyway.
[0056]
Next, a second embodiment of the present invention will be described with reference to FIGS. 8, 9, 10, and 11. FIG. The configuration of this embodiment shown in FIG. 8 is the same as the configuration 11-19 for one wavelength band in FIG. 5, and the operation of the end point determiner 130 is different from the operation of the end point determiner 30 in FIG. The structure of the material to be processed for etching is shown in FIG. The region of the polysilicon 50 processed in this etching process is a portion without the mask material 51 (for example, a nitride film or a photoresist), and the observed interference light is reflected from the reflected light 90A on the surface of the polysilicon 50 and the base oxide film 52. Caused by interference with the reflected light 90B. A method of measuring the etching amount of the polysilicon 50 (remaining film thickness 53: thickness of polysilicon from the base oxide film) by measuring the interference light will be described with reference to the flowchart of FIG.
[0057]
First, all the standard differential patterns (P) relating to the target remaining film value of the material to be processed (polysilicon) and the polysilicon film thickness stored in advance in the differential waveform pattern database 16_Zj) and convergence judgment value σ_Z0Is set in the differential waveform comparator 15 (step 600). As the etching process starts, sampling of interference light is started (step 602). The multi-wavelength light from the spectroscope 11 passes through the first digital filter 12, the differentiator 13, and the second digital filter circuit 14, respectively, in the same manner as in Steps 404-410 of the first embodiment. Data Di, j is obtained. These smoothed differential coefficient time series data Di, j and the differential pattern P set in the differential waveform comparator 15 in advance._Zj are compared, and the remaining film value Zi at that time is calculated (step 615). Where σ> σ_z0In this case, the Zi value obtained at this time in step 615 is not stored, and the residual film thickness data at this time is excluded in the processing by the regression analyzer 19.
[0058]
The remaining film value obtained in step 615 is stored in the data recorder 18 as time series data Zi. Using the stored past time series data Zi, the regression analyzer 19 uses the linear regression line Y = Xa * t + Xb (Y: amount of remaining film, t: etching time, Xa: absolute value of Xa is etching rate, Xb : Initial film thickness) is obtained, and the current remaining film amount Z is calculated from this regression line (step 616).
[0059]
Next, the end point determination unit 130 compares the remaining film amount Z with the target remaining film value, and if it is equal to or less than the target remaining film, the etching amount of the material to be processed becomes a predetermined value and the result is displayed on the display unit 17. indicate. If it is equal to or greater than the target remaining film, the process returns to step 604. Finally, the end of sampling is set (step 614).
[0060]
By the way, the differential waveform pattern database P relating to the thickness of the polysilicon in which the target remaining film value is stored in advance._ZIf it is determined in step 618 that the number is smaller than j, the following processing is performed and the etching processing is terminated. When the remaining film amount Z becomes equal to the minimum film thickness Ym of the differential waveform pattern database relating to the polysilicon film thickness stored, the target remaining film value Y is obtained from the above-mentioned linear regression line Y = Xa * t + Xb._T Etching time (t_T= [Y_T-Xb] / Xa), and this time t_T The etching process is continued until.
[0061]
FIG. 11 shows the result of measuring the remaining film of polysilicon performed in this example. In this case, the target residual film value Yt = 20 nm is predicted using a database up to the minimum film thickness Ym = 45 nm as the differential waveform pattern database relating to the polysilicon film thickness, and the target residual film value is 20 nm according to the linear regression line. The etching time of 96 seconds can be clearly seen. This makes it possible to determine the end point of the remaining film amount without the differential waveform pattern database.
[0062]
The absolute value of the slope of the linear regression line (= | Xa |) indicates the etching rate, and the state of the etching apparatus can be managed by mass-managing the etching rate. That is, if the etching rate is within a certain allowable value, it can be monitored that the etching apparatus is operating normally, and if it is other than the allowable value, it can be monitored that it is abnormal.
[0063]
Further, the intercept of the linear regression line (= Xb) indicates the initial film thickness of the material to be processed, and the film formation state before the etching process can be managed by mass production management of this initial film thickness. That is, if the initial film thickness is within a certain allowable value, it can be monitored and fed back that the film forming apparatus is operating normally, and if the initial film thickness is other than the allowable value, it can be monitored.
[0064]
Next, a third embodiment of the present invention will be described with reference to FIGS. In this embodiment, when the etching depth is managed, there is an error in the film thickness of the organic film for each wafer. Therefore, the etching depth is measured from the initial film thickness and the measured remaining film thickness. . The configuration of this embodiment shown in FIG. 12 is the same as the configuration 11-19 of FIG. 8, and the operation of the end point determination unit 230 is different from the operation of the end point determination unit 130 of FIG. The structure of the material to be processed for etching is shown in FIG. The region (groove structure) of the organic film 60 processed in this etching process is a portion without the mask material 61 (for example, a nitride film or a photoresist), and the observed interference light is reflected light on the surface of the organic film 60 and the wiring material. This is caused by interference with reflected light from 62 (for example, Cu).
[0065]
A method of measuring the etching amount (groove depth: distance 65 between D and E) of the organic film 60 by measuring the interference light will be described with reference to the flowchart of FIG. First, all the standard differential patterns (P) relating to the target depth value of the material to be processed (organic film) and the film thickness of the organic film stored in the differential waveform pattern database 16 in advance._Fj) and the convergence judgment value (σ_F0) Is set in the differential waveform comparator 15 (step 700). As the etching process starts, sampling of interference light is started (step 702). Multi-wavelength light from the spectroscope 11 passes through the first digital filter circuit 12, the differentiator 13, and the second digital filter circuit 14, respectively, in the same way as in Steps 604 to 610 of the second embodiment. The series data Di, j is obtained. These smoothed differential coefficient time series data Di, j and the differential pattern P set in the differential waveform comparator 15 in advance._Fj are compared, and the remaining film value Fi at that time is calculated (step 715). Where σ> σ_F0In this case, the remaining film thickness value at this time is not stored, and the residual film thickness data at this time is excluded in the processing by the regression analyzer 19.
[0066]
The remaining film value obtained in step 715 is stored in the data recorder 18 as time series data Fi. Using the stored past time series data Fj, the regression analyzer 19 uses the linear regression line Y = Xa * t + Xb (Y: amount of remaining film, t: etching time, Xa: absolute value is etching rate, Xb: initial stage Film thickness) is obtained, and the current remaining film amount F and initial film thickness Xb are calculated from this regression line (step 716).
[0067]
Next, the end point determination unit 230 obtains the current groove depth (= Xb−F) (65 in FIG. 13) from the remaining film amount F (64 in FIG. 13) and the initial film thickness Xb (63 in FIG. 13). The groove depth is compared with the target depth value, and if the depth is equal to or larger than the target depth, the result is displayed on the display unit 17 as the etching amount of the material to be processed becomes a predetermined value. If the depth is less than the target depth, the process returns to step 704. Finally, the end of sampling is set (step 714). Thus, the etching depth at the time of groove processing can be measured by obtaining the residual film amount F and the initial film thickness Xb by regression analysis.
[0068]
Next, a fourth embodiment of the present invention will be described with reference to FIG. 15, FIG. 16, and FIG. The configuration of this embodiment shown in FIG. 15 is the same as the configuration 11-19 of FIG. 12, the operation of the end point determination unit 330 is different from the operation of the end point determination unit 230 of FIG. 12, and a control device 1000 is further provided. Yes. The same etching apparatus 1 is often used to etch materials to be etched having different film qualities. In this case, based on the etching conditions (for example, etching gas conditions, plasma generation power conditions, bias conditions, etc.) preset in the control apparatus 1000, the control apparatus 1000 causes the gas supply apparatus 1001, plasma generation apparatus 1002, and wafer bias power supply 1003 to be used. The time is controlled to perform the etching process. However, if the material to be processed to be processed has a laminated structure as shown in FIG. 16 due to the semiconductor element structure, the etching process becomes complicated, and the element processing without damage becomes difficult by the etching process by simple time control. . The etching process of such a laminated structure element will be described with reference to FIG. A BARC 73 (Back Anti-Reflection Coating) and a mask material 71 (for example, a nitride film or a photoresist) are formed on the polysilicon film 70 processed in the etching process, and a base oxide is formed under the polysilicon film 70. A film 72 is formed. Further, the base oxide film 72 has a structure in which the thickness of the gate electrode portion 78 of the transistor (for example, about 2 nm) and the thickness of the groove portion 79 (STI) for separating the transistor elements (for example, about 300 nm) are greatly different. It has become. In the processing of this structure, first, the etching process of BARC 73 is performed, and then the etching process of the polysilicon film 70 is consistently performed by the same etching apparatus. In such an etching process, if the respective films cannot be etched appropriately, the base oxide film 72 of the gate electrode portion 78 is excessively etched, and the element is damaged. Therefore, first, in the BARC 73 etching process, it is necessary to control the etching so that the polysilicon film 70 is not cut as much as possible. This is because the BARC 73 etching process measures the BARC residual film amount, and when the residual film amount becomes 75 (for example, 20 nm), the etching condition is changed to a condition in which the polysilicon is difficult to be removed. It is important to etch the material. Next, in the etching process of the polysilicon film 70, the amount of remaining film (residual film thickness) of the polysilicon is measured, and when the amount of remaining film (residual film thickness) 77 (for example, 20 nm) is reached, It is important to etch the remaining polysilicon material by changing the etching conditions to a condition in which the underlying oxide film is difficult to be removed.
[0069]
The light used for the BARC residual film thickness measurement uses interference light 96 between the reflected light from the BARC surface and the reflected light from the polysilicon interface. The light used for the measurement of the remaining film thickness of polysilicon uses interference light 95A or 95B between the reflected light from the polysilicon surface and the reflected light from the boundary surface of the underlying oxide film. At this time, since the thickness of the base oxide film is different from the thickness 77 of the gate electrode part 78 and the thickness 76 of the groove part 79 for separating the transistor elements, the interference light intensities 95A and 95B from the respective parts are different. The interference light intensity 95B from the element isolation part 79 is larger than the interference light intensity 95A from the gate electrode part 78. Therefore, the measurement of the remaining film thickness of polysilicon is performed on the polysilicon on the element isolation portion 79. In other words, in consideration of this point, the polysilicon etching process is performed by using the interference light intensity 95B until the remaining film thickness becomes 76, and then the etching condition under which the underlying oxide film is difficult to be removed. The polysilicon material is etched.
[0070]
The procedure for performing this etching process will be described with reference to the flowchart of FIG. First, the control apparatus 1000 causes the etching conditions (for example, gas conditions, discharge conditions, pressure conditions, etc.) for the laminated films (for example, the BARC 73 and the polysilicon film 70) and the target remaining film thickness values 75, 76 and a convergence judgment value are set (step 800). Next, all standard differential patterns (P) relating to the film thicknesses of the respective films 73 and 70 stored in advance in the differential waveform pattern database 16 according to the respective film types._Zj) and the convergence judgment value (σ_Z0) Is set in the differential waveform comparator 15 (step 801). In the next step, the etching process and sampling of interference light are started (step 802). Next, the standard differential pattern P regarding the material to be etched (for example, the BARC material) to be processed first in accordance with an instruction from the control device 1000._Zj and convergence judgment value σ_Z0Is set in the differential waveform comparator 15 from the differential waveform pattern database 16 (step 803). The multi-wavelength light from the spectroscope 11 obtains the smoothed time series data Yi, j from the first digital filter circuit 12 (step 804). Further, the smoothed differential coefficient time series data Di, j is obtained through the differentiator 13 and the second digital filter circuit 14 in the same manner as in Steps 704 to 710 of the third embodiment (Steps 806 and 808). These smoothed differential coefficient time series data Di, j and the differential pattern P set in the differential waveform comparator 15 in advance._Fj and the convergence value σ = Σ (Di, j−P_Zj)²The remaining film thickness with respect to the convergence value with the smallest is obtained. At this time, σ ≦ σ_Z0If it is, the obtained remaining film thickness is set as the remaining film thickness value Zi at this time and stored in the data recorder 18 (steps 810 and 815). σ> σ_Z0In this case, the residual film thickness value at this time is not stored, and the residual film thickness data at this time is excluded in the processing by the regression analyzer 19 (step 815). Using the stored past time series data Zi, the regression analyzer 19 obtains a primary regression line Y = Xa * t + Xb, and the current residual film thickness Z is calculated from the regression line (step 816).
[0071]
Next, the end point determination unit 330 compares the current remaining film thickness amount Z with the target remaining film thickness value 75 from the control device 1000 (for example, BARC remaining film thickness 20 nm), and the target remaining film value 75 or less. If the result indicates that the etching amount of the material to be processed has reached a predetermined value and the result is displayed on the display unit 17, the etching condition is switched at the same time, and the next etching process (the BARC etching condition in which polysilicon is not easily etched) is performed ( Step 818). If it is equal to or greater than the target remaining film value, the process returns to step 804. The next etching process (the BARC etching condition in which polysilicon is hard to be removed) is controlled by, for example, controlling the wafer bias power supply 1003 by the control device 1000 and etching the remaining film portion. It is determined whether or not it is necessary to switch the etching conditions for the next film type etching (step 819).
[0072]
In the etching of the next film type (for example, the polysilicon film 70), the gas supply device 1001, the plasma generator 1002, and the wafer bias power source 1003 are controlled to preset etching conditions, and the next film type (for example, the poly-silicon film 70) is controlled. Standard differential pattern (P) for residual film thickness 76 of silicon_Zj, σ_Z0) Is set in the differential waveform comparator 15 (step 803), and steps 804 to 818 are executed again. In the end point determination unit 330, it is determined that the remaining film thickness amount Z of the polysilicon becomes the target remaining film thickness value 76 (for example, the remaining polysilicon film thickness is 20 nm) (step 818), and the result is displayed. When displayed on the vessel 17, the etching conditions are switched at the same time to perform the next etching process (polysilicon etching conditions in which the underlying oxide film is hard to be removed) (step 818). In this etching process (polysilicon etching conditions in which the underlying oxide film is hard to be removed), the remaining film portion of the polysilicon is etched by time control, and the switching of the etching conditions is determined after a preset time has elapsed (step 819). . If there is no next etching process, the end of the etching process and the end of sampling are set (step 814).
[0073]
As described above, in the above method according to the second embodiment, the remaining film thickness can be predicted even if there is not enough differential pattern database regarding the remaining film thickness of the material to be processed. Furthermore, in creating the differential pattern database relating to the remaining film thickness, it is not necessary to completely etch the sample wafer, and the number of sample wafers can be reduced.
[0074]
Further, since the etching rate is obtained by the regression analysis of the present invention, mass production management can be performed by monitoring the etching rate for each wafer, and the occurrence of defects in the product wafer can be prevented in advance.
[0075]
In addition, since the initial film thickness of the material to be processed can be calculated by the regression analysis of the present invention, mass production management of integrated processing between the film forming apparatus and the etching apparatus can be performed by feeding back the result to the film forming apparatus.
[0076]
According to the method according to the third embodiment, even when the initial film thickness of the material to be processed varies, the etching depth can be accurately calculated, and the etching process for accurately determining the target etching depth can be performed.
[0077]
Further, according to the method of the fourth embodiment, in the etching process of the material to be processed having the laminated film structure, the remaining film thickness is measured when each film is etched, and the etching is performed at the predetermined remaining film thickness. By switching the conditions, etching etching of the base film can be reduced.
[0078]
In polysilicon etching of the gate electrode, the predetermined remaining film thickness can be accurately determined by measuring the polysilicon remaining film thickness by the interference light from the element isolation part. Etching is not excessively performed, and the number of defective wafers processed can be minimized.
[0079]
In each of the above-described embodiments, measurement is performed by emitting multi-wavelength radiation light from a spectroscope having a light source and using interference light of reflected light from a material to be processed. On the other hand, a spectroscope that does not have a light source may be used so that multi-wavelength radiation emitted by plasma is used as the light source.
[0080]
The present invention further has the following characteristics.
(1) Etching characterized by accumulating data measured during etching of a sample, predicting an etching depth in the etching using the accumulated past data, and stopping the etching at a predetermined etching depth Method.
(2) In an etching method of a material to be etched that does not have a base film, the amount of etching of the mask formed on the material to be etched and the etching step from the top surface of the mask to the etching bottom surface of the material to be etched are measured. And an etching depth of the material to be etched is controlled.
(3) An etching method characterized in that the remaining film thickness of the etched portion of the sample is predicted using interference light data during the etching of the sample, and the etching is stopped at a predetermined remaining film thickness.
(4) An etching method characterized in that in damascene etching, the initial film thickness is predicted using the interference light data during the etching, the groove depth is determined, and the etching is stopped at a predetermined groove depth.
(5) In the method of etching a gate material formed on a base film having a different in-plane thickness, interference light is measured at a thick portion of the base film, and the thick light is applied on the thick base film. An etching method characterized by measuring a remaining film thickness of a formed gate material and managing the film thickness of the gate material.
(6) In etching of a material to be etched in which a plurality of films are stacked, the interference light from the material to be etched is measured, the digital filter is switched for each film, and the data of the interference light is processed for each film. An etching method characterized by controlling a film thickness.
(7) In the BARC etching, the film thickness of the BARC is controlled using interference light measured from the material to be processed during etching, and the underlying material to be etched is prevented from being scraped.
(8) In manufacturing a semiconductor element having an STI portion, when etching Poly-Si forming a part of the semiconductor element, the remaining film thickness of Poly-Si formed on the STI portion is: A method of manufacturing a semiconductor device, wherein the etching of the poly-Si is managed.
[0081]
【The invention's effect】
  According to the present invention,It is possible to provide a semiconductor wafer processing method and processing apparatus that can detect or control the processing amount such as the etching depth and the amount of remaining film more precisely.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an overall configuration of a semiconductor wafer etching apparatus provided with an etching amount measuring apparatus according to a first embodiment of the present invention.
FIG. 2A is a diagram showing a longitudinal cross-sectional shape of a material to be processed during an etching process, and FIGS. 2B and 2C are diagrams showing examples of actual wavelength patterns of interference light having different wavelengths. is there.
FIGS. 3A and 3B are steps shown by A, B, and C in FIG. 2A and FIG. 2B (distances from the mask material surface) and a, b, and c, respectively. It is a figure which uses as a parameter the wavelength of the derivative value time series data of interference light corresponding to each remaining film thickness of a mask.
4 is a flowchart showing a procedure for obtaining a level difference of a material to be processed and a residual mask film thickness when performing an etching process with the etching amount measuring apparatus of FIG. 1;
FIG. 5 is a block diagram showing an overall configuration of a semiconductor wafer etching apparatus provided with an etching depth measuring apparatus according to a modification of the first embodiment of the present invention.
6 is a flowchart showing the operation of the embodiment of FIG.
7 is a diagram showing a result of etching depth measurement in the example of FIG. 5; FIG.
FIG. 8 is a block diagram showing an overall configuration of a semiconductor wafer etching apparatus provided with a residual film thickness measuring apparatus according to a second embodiment of the present invention.
FIG. 9 is a view showing a vertical cross-sectional shape of a material to be processed during an etching process.
FIG. 10 is a flowchart showing the operation of the embodiment of FIG.
11 is a diagram showing a measurement result and a regression line of the remaining polysilicon film thickness in the example of FIG. 8;
FIG. 12 is a block diagram showing an overall configuration of a semiconductor wafer etching apparatus provided with an etching depth measuring apparatus according to a third embodiment of the present invention.
FIG. 13 is a view showing a vertical cross-sectional shape of a material to be processed during an etching process.
14 is a flowchart showing the operation of the embodiment of FIG.
FIG. 15 is a block diagram showing an overall configuration of a semiconductor wafer etching apparatus provided with an etching residual film measuring apparatus according to a fourth embodiment of the present invention.
FIG. 16 is a view showing a vertical cross-sectional shape of a material to be processed during an etching process.
17 is a flowchart showing the operation of the embodiment of FIG.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Etching apparatus, 2 ... Vacuum container, 3 ... Plasma, 4 ... Material to be processed, 5 ... Sample stand, 8 ... Optical fiber, 9 ... Radiation light, 10 ... Measuring apparatus, 11 ... Spectroscope, 12, 22 ... 1st Digital filter circuit, 13, 23 ... Differentiator, 14, 24 ... Second digital filter circuit, 15, 25 ... Differential waveform comparator, 16, 26 ... Differential waveform pattern database, 17 ... Display, 18, 28 ... Time series Data recorder, 19, 29 ... regression analyzer, 30, 130, 230, 330 ... end point determiner, 40 ... silicon, 41 ... nitride film, 50 ... polysilicon, 51, 61, 71 ... mask material, 52 ... underlayer Oxide film, 60 ... Organic film, 62 ... Wiring material, 70 ... Polysilicon film, 72 ... Base oxide film, 73 ... BARC, 78 ... Gate electrode part, 79 ... Groove part, 1000 ... Control device, 1001 ... Gas supply device 1002 ... plasma generator, 1003 ... wafer bias power supply.

Claims (10)

A semiconductor wafer processing method for processing a film, which is a material to be processed, formed on a surface of a semiconductor wafer,
Detecting a differential value of each intensity of interference light of a plurality of wavelengths obtained from the semiconductor wafer surface during the processing;
Data of differential values of the intensity of the interference light of the plurality of wavelengths with respect to a predetermined amount of the processing obtained in advance for a semiconductor wafer having substantially the same configuration as the semiconductor wafer, and intensity of the detected plurality of wavelengths. And a method of detecting a processing amount for the film being processed based on a set of differential values. A method for processing a semiconductor wafer according to claim 1, comprising:
A method of processing a semiconductor wafer comprising a step of switching the processing conditions and processing the semiconductor wafer when it is determined that the detected amount of the film being processed reaches a predetermined amount. . A method of processing a semiconductor wafer according to claim 1 or 2,
A method of processing a semiconductor wafer, comprising: detecting a remaining thickness of the film as the amount of processing to be detected. A method for processing a semiconductor wafer according to any one of claims 1 to 3,
Predicting the time when the amount of processing reaches a predetermined amount by regression analysis using the amount of processing obtained before detection of a set of differential values of the intensity of interference light of the plurality of wavelengths A method for processing a semiconductor wafer comprising: A method for processing a semiconductor wafer according to claim 4, comprising:
A method for processing a semiconductor wafer, comprising the step of obtaining the processing speed by the regression analysis. A semiconductor wafer processing apparatus for performing a predetermined process on a semiconductor wafer to be processed disposed on a sample stage in a vacuum vessel using plasma formed in the vacuum vessel,
It has a measuring device that receives interference light of a plurality of wavelengths from the surface of the semiconductor wafer during the processing, detects a differential value of each intensity of the interference light of the plurality of wavelengths received in the measurement device, Data of the differential value of the intensity of the interference light of the plurality of wavelengths and the differential value of the intensity of the detected plurality of wavelengths with respect to a predetermined amount of the processing obtained in advance for a semiconductor wafer having substantially the same configuration as the semiconductor wafer. A wafer processing apparatus having a function of detecting the amount of processing for the film being processed. A semiconductor wafer processing apparatus according to claim 6, comprising:
A semiconductor wafer having a function of switching the processing conditions and processing the semiconductor wafer when it is determined that the amount of the processing performed on the semiconductor wafer being processed reaches a predetermined amount. Processing equipment. A semiconductor wafer processing apparatus according to claim 6 or 7,
A semiconductor wafer processing apparatus in which a remaining thickness of the film is detected as the detected amount of processing. A semiconductor wafer processing apparatus according to any one of claims 6 to 8,
The time when the amount of processing reaches a predetermined amount is predicted by regression analysis using the amount of processing obtained before the detection of a set of differential values of the intensity of interference light of the plurality of wavelengths. Semiconductor wafer processing equipment. A method for processing a semiconductor wafer according to claim 9, comprising:
A semiconductor wafer processing apparatus for obtaining the processing speed by the regression analysis.

Images