JP3910647B2 - Charged beam drawing method and drawing apparatus - Google Patents

Description

【００４２】
次いで、ＣＰ候補セルから、アパーチャマスクに形成できる所定数の候補セルを選んで、可能な組み合わせを作成する（Ｓ４）。例えば、図２に示したように、ＣＰ用アパーチャが４個アパーチャマスク上に可能である場合には、図７の例では、 ₈Ｃ₄ ＝７０通りの組み合わせを作成する。
【００４３】
次いで、図８のフローチャートに従い、ＣＰ候補セルの全ての組み合わせについて、各組み合わせに属するＣＰ候補セルをキャラクタセルとして露光した場合の描画時間を算出する（Ｓ５）。以下、この算出方法を詳細に説明する。
【００４４】
まず、チップ全体をフレーム及びサンプルエリアに分割し、各サンプルエリアを可変成形ビームのみで描画した場合の必要描画時間Ａを求める。同時に、候補セルｃをＣＰ露光した場合の候補セル毎の各サンプリルエリアの描画短縮時間ΔＡｃを求める。ここで、フレームとは描画領域を偏向系の偏向幅で決まる所定幅のストライプに分割した領域であり、サンプルエリアとはフレームをステージ連続移動方向に仮想的に小領域に分割した領域である。
【００４５】
以下、ｉ番目のフレームのｊ番目のサンプルエリアを可変成形ビームのみを用いて描画した場合の描画時間Ａを、新たにｔ^ijVSB と書くことにすると、前述の描画装置の場合には、
ｔ^ijVSB ＝Ｎsub ×ｔs1＋（Ｎshot−Ｎsub)×ｔs2＋Ｎshot×ｔd … (1)
で与えられる。但し、
ｔs1：主偏向セトリング時間
ｔs2：副偏向セトリング時間
ｔd ：照射時間
Ｎsub ：ｉ番目のフレームのｊ番目のサンプルエリアのサブフィールド数
Ｎshot：ｉ番目のフレームのｊ番目のサンプルエリアのショット数
サブフィルード：副偏向の切り換えのみで露光できる領域
である。
【００４６】
また、ｉ番目のフレームのｊ番目のサンプルエリアにおいて、候補セルｃのみをＣＰ露光した場合の、可変成形ビームのみで露光した場合と比較した短縮描画時間ΔＡｃを新たにΔｔ^ijｃと書くと、
Δｔ^ijｃ＝（Ｎshot,c(VSB) −Ｎshot,c(CP)）×（ｔs2＋ｔd) … (2)
となる。但し、
Ｎshot,c(VSB) ：ｉ番目のフレームのｊ番目のサンプルエリアにある、全てのセルＣをＶＳＢのみで露光した際の総ショット数
Ｎshot,c(CP)：ｉ番目のフレームのｊ番目のサンプルエリアにある、全てのセルＣをＣＰ露光した場合の総ショット数
である。同様にして、全てのサンプルエリアについて、ｔ^ijVSB とΔｔ^ijｃを求める。また、Δｔ^ijｃについては、全てのＣＰ候補セルについて求めるものとする。
【００４７】
次に、上述したＣＰ候補セルの組み合わせの一つ一つについて、チップ全体を描画するのに必要な描画時間を上記ｔ^ijVSB 及びΔｔ^ijｃから求める方法について説明する。図７中の候補セルについて、組み合わせ（Ａ，Ｂ１，Ｃ１，Ｄ）を例にとって、算出方法を説明する。まず、Ａ，Ｂ１，Ｃ１，ＤをＣＰ方式で、チップの残りの部分を可変成形ビームで露光するとして、各フレームのそれぞれのサンプルエリアにおける必要描画時間を以下の式で算出する。
【００４８】
ｔ^ij＝ｔ^ijVSB −Δｔ^ijA −Δｔ^ijB1−Δｔ^ijC1−Δｔ^ijD … (3)
即ち、各サンプルエリアにおける必要描画時間は、図９に示すように、可変成形ビームのみを用いて描画するのに必要な時間から、キャラクタ化による描画時間の短縮分を引けばよい。なお、図９（ａ）は各サンプルエリアにおける必要描画時間、図９（ｂ）は対応するフレームの模式図を示している。
【００４９】
次に、この値をもとに、各フレームにおけるステージ速度を算出する。まず、フレーム毎に必要描画時間の最も長いサンプルエリアを求め、このサンプルエリアを描画できる最高の速度をステージ速度とする。即ち、ｉ番目のフレーム上において必要描画時間の大きいサンプルエリアがｍ番目のサンプルエリアであるとすると、ステージ速度ｖ^imは
ｖ^im＝ｌ／ｔ^im … (4)
となる。但し、ｌはサンプルエリアの幅である。
【００５０】
従って、ｉ番目のフレームの実際の描画時間ｔⁱ は、
ｔⁱ ＝Ｌ／ｖ^im … (5)
と求まる。
【００５１】
この処理を全てのフレームについて行えば、全描画時間Ｔａｌｌは、
Ｔａｌｌ＝Σｔⁱ … (6)
ⁱ
で、求まる。
【００５２】
以上の処理を、候補セルの全ての組み合わせについて行い、それぞれの組み合わせに対してチップ全体の描画時間を求め、描画時間が最短となった組み合わせのＣＰ候補セルをアパーチャマスクのＣＰ用アパーチャ図形として選択する（Ｓ６）。即ち、
Ｔｏｐｔ＝Ｍｉｎ（Ｔ¹ _all ，Ｔ² _all ，…） … (7)
として、Ｔｏｐｔを与える組合わせに属するＣＰ候補セルをアパーチャマスクのＣＰアパーチャ図形として作成する。但し、Ｔ^k _all はｋ番目のＣＰ候補セルの組み合わせをＣＰ露光した場合のチップ全体の描画時間である。
【００５３】
このように本実施例によれば、ステージ連続移動方式の電子ビーム露光装置において、描画時間が最も短くなるＣＰ候補セルの組み合わせをアパーチャマスクに形成しておくことにより、露光ショット数を少なくして描画スループットの向上をはかることができる。
【００５４】
そしてこの場合、各サンプルエリアの可変成形ビームのみを用いた場合の描画時間Ａと各ＣＰ候補セルｃをキャラクタ化した場合に短縮される描画時間ΔＡｃとを全サンプルエリアについて一旦求めておき、ＡとΔＡｃからＣＰ候補セルの各組み合わせのステージ速度、及びこれから導かれる描画時間を算出することにより、描画時間が最短である候補セルの組み合わせを容易に選択することができる。即ち、描画時間を最も短縮する効果のある繰り返しパターンを求める際に、無駄な手順を省き、効率的な選択をすることができる。
（実施例２）
図１０は、本発明の第２の実施例を示すブロック図であり、特に設計データから描画データを作成する描画データ作成手段の構成を示している。
【００５５】
本実施例の描画装置は、描画データ作成手段５０と描画手段６０からなる。描画データ作成手段５０は、階層構造組み替え処理部５１、階層的図形データ処理部５２、キャラクタ化セル決定処理部５３、キャラクタ制御コード置換処理部５４、描画領域別描画データ出力部５５から構成されている。描画手段６０は描画装置本体に相当するものであり、前記図１に示したものと同様である。
【００５６】
階層構造組み替え処理部５１は、これに続く階層的図形データ処理の前処理であり、設計データを読み出してその階層構造を再構成し、階層的図形データ処理によってセル境界付近で図形重複或いは間隙などのエラー箇所が発生することを防止する。階層的図形データ処理部５２は、設計データの階層構造を維持したまま、図形重複除去，レイヤー間図形論理演算，リサイズ，倍率補正などの図形データ処理をセル単位に行う。
【００５７】
キャラクタ化セル決定処理部５３は、図１１に示すように規則的パターン領域をＣＰ描画領域として、それ以外の基本形状ビーム描画領域と分離する処理であり、階層的図形データ処理の処理結果を基に、セル単位にキャラクタ化効果を評価してキャラクタ化セルの集合を決定する。回路パターン設計時に設計者が繰り返し単位として定義したセルをキャラクタ化する図形群の単位として採用することにより、回路パターンから繰り返し単位を抽出する手間を大幅に軽減することができる。
【００５８】
キャラクタ制御コード置換処理部５４は、キャラクタ化が決まったセルの図形データを抜き出して、キャラクタ制御コードに置き換える処理である。キャラクタ制御コードは、描画手段がキャラクタビームを発生する際に、キャラクタ形状を識別する制御コードとしている。
【００５９】
最後の描画領域別描画データ出力部５５は、ＶＳＢ描画データとキャラクタ制御コードを描画位置に応じて、描画領域単位（フレーム領域など）に振り分けて出力するものである。以上の処理により、ＣＰ描画方式に対応した描画データが作成されるようになっている。
【００６０】
次に、図１２に示したキャラクタ化セル決定処理部５３のブロック図に基づいてキャラクタ化セル決定処理の詳細を説明する。キャラクタ化セル決定処理部５３は、セル情報解析部５３ａ，第１次候補セル選択部５３ｂ，候補セル再構成部５３ｃ，第２次候補セル選択部５３ｄ，キャラクタ化効果評価部５３ｅから構成される。
【００６１】
セル情報解析部５３ａは、階層的図形データ処理の処理結果であるセル内図形データとセル配置データを読み込み、セル情報テーブルを作成する。このセル情報テーブルは、各セルに割り振られたセル識別番号、セル内に定義された図形の外接矩形のサイズであるセルサイズ、そのセルのチップ内の総参照数、基本形状ビームで描画した場合のショット数、キャラクタ化した場合のショット数といった項目から構成される。
【００６２】
このセル情報テーブルを基に、キャラクタ化効果の比較的高いセルを、第１次候補セル選択部５３ｂにより第１次キャラクタ化候補セルとして選択し、以後の処理の対象となるセルの数を絞り、処理負荷を軽減しておく。
【００６３】
セルを単位にキャラクタ化することにより、回路パターンから繰り返し単位を抽出することが大幅に容易となる。しかし、セルは設計時の定義された図形群のため、そのサイズ及び種類数は任意である。一方、キャラクタは装置仕様の制約から、サイズ及び種類数が制限されている。そのため、第１次キャラクタ候補セルをそのままキャラクタ化することができない場合がある。そこで、適切なセルサイズにすると共に、限定されたキャラクタ種類数を最大限活用することができるように、候補セルの定義内容を再構成する必要がある。候補セル再構成部５３ｃでは、次のような処理を行う。
（１）階層構造の部分的展開
（２）パターンの部分抽出
（３）パターン認識
（４）アレイ再構成
（５）セル分割
以下に各処理内容の詳細を説明する。
（１）階層構造の部分的展開
図１３（ａ）に示すように、セルＡ内でセルＢがアレイ配置されており、セルＢ内でセルＦが４種類の配置方向で参照配置されているものとする（Ｆの添字が配置方向を示す）。この参照構造を木表記法で表すと、図１３（ｂ）となる。セルＦのセルサイズは、最大ビーム寸法の１／２以下であり、セルＢを単位にキャラクタ化すれば、４個のセルＦの領域を一括描画することができる。しかし、セルＢ内ではセルＦを参照配置してはいるが、図形は定義されておらず、単純にセルＢをキャラクタ化すると中身が空のキャラクタとなってしまう。
【００６４】
そこで、指定したセルより下位の階層構造を展開し、下位セルで定義されていた図形を指定したセルに与える。この例では、セルＢ以下で参照されているセルＦ１〜Ｆ４内に定義されている図形をセルＢに与える。これにより、セルＢのキャラクタ化効果が高まる。
（２）パターンの部分抽出
図１４に示すように、セルＥ及びセルＰはセルＦに類似のパターンであり、類似部分を切り出して同一のキャラクタで置き換えることができれば、キャラクタ描画の効率が高まる。そこで、セル内のパターンの一部を切り出して新たな候補セルとする機能が必要となる。
（３）パターン認識
セル識別番号は異なるが、同一のパターンが内部に定義されている場合が存在する。これらを別々のキャラクタとして抽出したり、一方をＶＳＢ描画してしまうと、ＣＰ描画の効率が低下する。そこで、パターン認識を行ってパターンを検査し、同一内容のセルについてセル識別番号を統一する。前項（２）のパターン部分抽出と組み合わせれば、類似図形を含むセルの一部をキャラクタ化することが可能となる。
（４）アレイ再構成
最大ビーム寸法より小さい候補セルがアレイ配置されている例を、図１５に示す。最大ビーム寸法が３μｍ、アレイ配置ピッチが２μｍの場合、最大ビーム寸法とアレイ配置ピッチの最小公倍数は６μｍである。この領域内に候補セルを展開すると、図１５（ａ）のように３×３＝９個のセルが入る。これを最大ビーム寸法で分割すると図１５（ｂ）のように４個のセルに分割され、アレイ配置構造は再構成される。このアレイ再構成により作成されたセルをそれぞれキャラクタ化すると、キャラクタ種類数は増加するが、総ショット数は低減する。ここで示した例では、キャラクタ数は１個から４個に増加し、総ショット数は４／９に減少している。
【００６５】
以上の候補セル再構成を行った後、第２次候補セル選択部５３ｄにより所定数の候補セルを第２次候補セルとして選択し、キャラクタ化効果評価部５３ｅによりキャラクタ化効果を評価する。ここで、キャラクタ化効果の評価規範として、選択した第２次候補セルをＶＳＢ描画した場合と、キャラクタ化した場合の描画時間との差を算出して用いると、より現実に近い評価を行うことができ、適切なキャラクタ化効果を評価することが可能となる。
【００６６】
さらに、第２次選択する候補セルの組み合わせを種々変更してキャラクタ化効果を評価すると共に、場合によっては再度、候補セル再構成，第２次候補セル選択及びキャラクタ化効果評価を繰り返すことによって、より適切なキャラクタ集合を求めることができる。
【００６７】
このように本実施例によれば、描画データ作成手段５０を図１０のように構成し、ＶＳＢ描画パターンとＣＰ描画パターンとを分離する前に階層的図形データ処理を行うことによって、設計パターンデータからキャラクタビームにより一括描画する図形群を容易に抽出することができ、かつＶＳＢ描画するパターンとの整合性も保たれた描画データを作成することができる。そして、抽出されたキャラクタ図形の集合が適切なものとなるため、描画データ作成に要する時間を短縮化して描画スループットを向上させることができる。その結果、電子ビーム描画装置の稼働率を高めると共に、ＬＳＩの生産性を高めることができる。
【００６８】
なお、本発明は上述した各実施例に限定されるものではない。第１の実施例では、アパーチャマスクに形成するキャラクタセルの数を４個として候補セルの組み合わせを求めたが、キャラクタセルの数はこれに限定されるものではなく適宜変更可能である。さらに、キャラクタセルの数は必ずしも一定にする必要はなく、キャラクタセルの数を候補セルの大きさにより可変にしてもよい。例えば、候補セルの大きさが異なる場合には、各々の候補セルの面積の和が所定範囲内に入るようにして候補セルの組み合わせを求めるようにすればよい。
【００６９】
また、本発明を実施するための装置の構成は図１に何等限定されるものではなく、基本図形用アパーチャとキャラクタ用アパーチャが形成されたアパーチャマスクを用い、試料を載置したステージを連続移動しながら、所定幅に分割されたフレームを順次描画するものであればよい。さらに、実施例では電子ビーム露光装置について説明したが、イオンビーム露光装置においても同様に適用できるのは勿論のことである。
【００７０】
また、第２の実施例で説明した描画データ作成処理と第１の実施例で説明したアパーチャパターン決定処理とを組み合わせて用いることにより、描画スループットのより一層の向上をはかることも可能である。その他、本発明の要旨を逸脱しない範囲で、種々変形して実施することができる。
【００７１】
【発明の効果】
以上詳述したように本発明（請求項１）によれば、各サンプルエリアの可変成形ビームのみを用いた場合の描画時間Ａと各ＣＰ候補セルｃをキャラクタ化した場合に短縮される描画時間ΔＡｃとを全サンプルエリアについて一旦求めておき、ＡとΔＡｃからＣＰ候補セルの各組み合わせのステージ速度及びこれから導かれる描画時間を算出することにより、ＣＰ描画方式の効果を十分に発揮してスループットの極めて高い描画を行うことができ、しかもアパーチャマスクに形成すべきＣＰセルの抽出を容易に行い得る荷電ビーム描画方法を実現することが可能となる。
【００７２】
また、本発明（請求項２〜４）によれば、ＶＳＢ描画されるパターンとＣＰ描画されるパターンとを分離する処理に先立ち、階層的図形データ処理の処理単位となった図形群（セル）を単位としてキャラクタビームの形状を決定し、キャラクタ化する図形群をＣＰ制御コードと置換し、ＶＳＢ描画パターン及びＣＰ制御コードを描画領域単位に描画データとして出力することにより、設計パターンデータからキャラクタビームにより一括描画する図形群を容易に抽出することができ、かつＶＳＢ描画するパターンとの整合性も保って描画データを作成することができ、描画データ作成に要する時間を短縮して描画スループットの向上をはかり得る荷電ビーム描画装置を実現することが可能となる。
【図面の簡単な説明】
【図１】第１の実施例に使用した電子ビーム描画装置を示す概略構成図。
【図２】ビーム成形アパーチャの構造を示す模式図。
【図３】成形ビームの形状を示す模式図。
【図４】描画データの体系を示す模式図。
【図５】実施例方法における処理概要を示すフローチャート。
【図６】実施例で描画する繰り返しパターンの例を示す模式図。
【図７】実施例で使用するＣＰ候補セルの例を示す模式図。
【図８】実施例における描画時間算出の処理を説明するためのフローチャート。
【図９】実施例におけるステージ速度算出の方法を示す模式図。
【図１０】第２の実施例を説明するためのもので、描画データ作成手段の構成を示すブロック図。
【図１１】キャラクタビーム描画領域と基本形状ビーム描画領域のパターン例を示す図。
【図１２】キャラクタ化セル決定処理部の構成を示すブロック図。
【図１３】階層構造の部分的展開処理を説明するため図。
【図１４】パターン部分抽出処理を説明するための図。
【図１５】アレイ再構成処理を説明するための図。
【図１６】従来のステージ連続移動方式の描画方法を説明するための模式図。
【符号の説明】
１０…試料室 １１…試料
１２…テーブル １３…テーブル駆動回路
１４…位置回路 ２０…電子光学系
２１…電子銃 ２２〜２６…レンズ
３１〜３４…偏向器 ３５，３６…ビーム成形アパーチャ
４０…制御計算機 ４１…磁気ディスク（記憶媒体）
４２…パターンメモリ ４３…パターンデータデコーダ
４４…描画データデコーダ ４５…ブランキング回路
４６…ビーム成形器ドライバ ４７…主偏向回路ドライバ
４８…副偏向器ドライバ ５０…描画データ作成手段
５１…階層構造組み替え処理部 ５２…階層的図形データ処理部
５３…キャラクタ化セル決定処理部 ５４…キャラクタ制御コード置換処理部
５５…描画領域別描画データ出力部 ６０…描画手段[0001]
[Industrial application fields]
The present invention relates to a charged beam drawing technique for drawing a pattern of a semiconductor integrated circuit such as an LSI on a sample such as a mask or a wafer with high speed and high accuracy, and in particular, exposes a repeated pattern portion typified by a memory device collectively. The present invention relates to a charged beam drawing method and a drawing apparatus.
[0002]
[Prior art]
The electron beam exposure system has the advantage of high accuracy because it can draw extremely fine patterns corresponding to the beam diameter and can electrically correct deflection and distortion and aberration of the beam. Widely used as a tool.
[0003]
Conventionally, as this electron beam exposure apparatus, an apparatus that draws in a manner of one-stroke writing with a Gaussian beam or a variable shaped beam has been the mainstream, but as the pattern becomes finer and more complex, the number of shots increases and throughput increases. There was a drawback of lowering. Therefore, recently, a semiconductor integrated circuit has been focused on the fact that many of the basic circuits are repetitions of the same pattern, and has a mask (hereinafter referred to as an aperture mask) in which a large number of transmission holes corresponding to the respective repeated pattern shapes are formed. A batch exposure (character projection, hereinafter abbreviated as CP) type electron beam exposure apparatus has been developed.
[0004]
By the way, in the CP type electron beam exposure apparatus, it is impossible to incorporate all the repeatedly appearing figures and groups of figures into the aperture mask, so the rest need to be drawn with combinations of basic figures. Therefore, the drawing time varies depending on which figure or group of figures is incorporated in the aperture mask. In other words, selecting an appropriate pattern to be incorporated into the aperture mask is extremely important for increasing the drawing throughput.
[0005]
As a first conventional example, in Japanese Patent Application Laid-Open No. 5-13313, pattern data that is repeatedly used is extracted from pattern data, and the pattern data is divided into predetermined block sizes to obtain candidate block data. A block pattern extracting unit is provided that sets a high effectiveness for the candidate block data in order of decreasing the total number of exposure shots, and selectively extracts a predetermined number of block patterns from the candidate block data. ing.
[0006]
As a second conventional example, in Japanese Patent Laid-Open No. 4-320021, a pattern to be incorporated in an aperture mask corresponds to the basic figure aperture from among the feature figures or feature figure groups repeatedly appearing in the pattern to be drawn. It is described that a characteristic figure or a characteristic figure group that requires a larger number of shots is preferentially selected and formed to reduce the total number of exposure shots when it is assumed that a shot is made with a combination of shaped beams.
[0007]
By the way, an electron beam exposure apparatus includes a so-called stage continuous movement drawing method that performs drawing while continuously moving a sample stage (stage) on which a wafer to be drawn is placed. As shown in FIG. 16, this method is a method in which the entire drawing area is divided by a deflectable width (frame) of the electron beam and drawing is performed in units of frames while continuously moving the stage. The stage speed is assumed to be constant for each frame.
[0008]
In such an electron beam drawing method of continuous stage movement, the drawing time of each frame is given by (frame length / stage speed), and the drawing time of the entire drawing area is the sum of the drawing time of the frame and the folding at the stage end. It is given as the sum of overhead time such as. Therefore, in this method, the drawing time is determined by the stage speed, and in order to improve the drawing throughput, it is necessary to optimize the stage speed as much as possible to increase the speed.
[0009]
As a third conventional example, Japanese Patent Laid-Open No. 4-61221 proposes a charged beam drawing method capable of easily obtaining an optimum moving speed of a stage without actually moving the stage. In this method, the required drawing time required for drawing is obtained for each sample area, the maximum drawing time of the small area at this speed is obtained on the assumption that the stage is moved at a certain speed, and the small area in the frame is taken from the end. A value A obtained by sequentially evaluating and adding a necessary time for each small area is compared with a sum B of maximum drawing times corresponding to the number of evaluated small areas. Then, based on whether or not these differences A-B exceed the sum C of the maximum drawing times included in the maximum beam deflection width, a substantially maximum stage moving speed at which A-B does not exceed C is determined. ing.
[0010]
As described above, in the electron beam drawing method adopting the CP method, a high degree of effectiveness is set in order of decreasing the total number of exposure shots for the extracted candidate block data, and a predetermined number of block patterns are selectively selected from the candidate blocks. Is extracted and determined as a block subjected to CP exposure.
[0011]
However, in the case of the stage continuous moving drawing method, the drawing time is finally determined in the form of the stage speed. The stage speed of each frame is determined not only by the total number of shots in the frame, but also by the density distribution of the number of shots and the arrangement mode of the cell data. For this reason, in the charged beam drawing method with continuous movement of the stage, even if it is adopted as the CP cell in the order of decreasing the total number of exposure shots, the stage speed may vary depending on the distribution of the number of shots and the cell data layout. Therefore, there is a problem that the drawing throughput is not improved.
[0012]
Further, as easily inferred from the third conventional example, all combinations that can be made by selecting a predetermined number of candidate cells from the CP candidate cells are created, and the stage speed and the drawing time are calculated for each of the combinations, and the drawing time is calculated. A method of selecting the combination that gives the shortest is expected. However, in this method, since each drawing time is obtained by repeatedly applying the method of the third conventional example for each of all combinations, a problem that the processing time is enormous is expected.
[0013]
On the other hand, when drawing a circuit pattern using an electron beam drawing apparatus employing the CP method, design pattern data created using a circuit pattern design tool such as CAD (Computer Aided Design) is input to the electron beam drawing apparatus. It is necessary to convert to a possible data format, that is, an electron beam drawing data format that satisfies conditions such as the drawing method and the specifications of the drawing apparatus. Such “conversion processing from design data to drawing data” is generally called “electron beam drawing data conversion”. The processing contents of the electron beam drawing data conversion necessary for the conventional VSB type drawing apparatus are summarized as follows.
(1) Inter-layer graphic logic (AND, OR, NOT, etc.)
(2) Pattern overlap removal (for the purpose of preventing multiple exposure)
(3) Resize (thick / thin)
(4) Magnification correction (enlargement / reduction)
(5) Rotation correction (rotation / mirror image)
(6) Divide by drawing area (divide by boundary line such as subfield area / frame area)
(7) Basic figure division (Division / approximation into basic figures that can be input to the drawing device)
(8) Data format conversion (converted into a rendering device-specific representation format and output)
When designing a circuit pattern, a designer first defines a graphic group called a cell, calls it in another cell, arranges it, and combines many cells to form the entire pattern. By performing the graphic processing while maintaining the hierarchical cell reference structure of such design data, the drawing data conversion processing time is shortened and the drawing data amount is reduced. Such a processing technique is called hierarchical graphic data processing.
[0014]
In the drawing apparatus adopting the CP method, the regular pattern is drawn by the CP method, and the irregular pattern is drawn by the VSB method. For this reason, in the drawing data conversion process corresponding to the CP method, in addition to the above function, a data process for separating a VSB drawing pattern by the basic shape beam and a CP drawing pattern by the character beam is required.
[0015]
In this separation process, it is necessary to obtain regularity from a large number of figures constituting a circuit pattern and to extract the repeated unit figure group as a character shape. In that case, there are two points: not causing an error in drawing data such as an overlap or a gap between the VSB drawing pattern and the CP drawing pattern, and efficiently extracting a set of character figures that optimize the drawing throughput. is important. In particular, in the latter case, it is necessary to obtain a set of character figures that provide the best drawing throughput under the limitations on the number of character types and the maximum beam size based on the device specifications, and the processing is extremely complicated. For this reason, there is a problem that the data processing time becomes long and the throughput of the entire drawing system is lowered.
[0016]
[Problems to be solved by the invention]
As described above, in the conventional stage-consecutive charged beam drawing method using the CP method, the drawing time is finally determined in the form of the stage speed, so the CP cell is adopted in the order of decreasing the total number of exposure shots. However, there is a problem in that the stage speed does not change depending on the density distribution of the number of shots and the arrangement pattern of the cell data, and therefore the drawing throughput does not improve. Further, in the method inferred from the third conventional example, the drawing time is obtained by repeatedly applying the method of the third conventional example for each of all the combinations, so that the processing time is enormous. The problem of becoming is expected.
[0017]
In addition, in a charged beam drawing apparatus adopting the CP method, in order to create drawing data from design data, graphic data processing is required to separate a VSB drawing pattern by a basic shape beam and a CP drawing pattern by a character beam. As a result, the data processing time for the separation processing becomes longer, and the throughput of the entire drawing system is also reduced.
[0018]
The present invention has been made in consideration of the above-described circumstances, and the object of the present invention is to perform the drawing of extremely high throughput by fully exhibiting the effect of the CP drawing method, and to form on the aperture mask. It is an object of the present invention to provide a charged beam writing method capable of easily determining a CP cell to be used.
[0019]
Another object of the present invention is to easily extract a group of figures to be drawn collectively by a character beam from design pattern data and to create drawing data while maintaining consistency with a pattern to be drawn by VSB. An object of the present invention is to provide a charged beam drawing apparatus capable of shortening the time required for creating drawing data and improving the drawing throughput.
[0020]
[Means for Solving the Problems]
The essence of the present invention is that the drawing time A when only the variable shaped beam of each sample area is used and the drawing time ΔAc that is shortened when each CP candidate cell c is characterized are obtained for all the sample areas. , A and ΔAc are used to calculate the stage speed of each combination of CP candidate cells and the drawing time derived therefrom.
[0021]
That is, according to the present invention (claim 1), a sample is placed using an aperture mask for forming a basic figure aperture for forming a shaped beam corresponding to a basic figure and a character aperture corresponding to a repetitive pattern. In a charged beam drawing method in which drawing areas (frames) divided into predetermined widths are sequentially drawn while continuously moving the stage, each pattern to be repeatedly used is extracted from desired data to be drawn, and the character cell is extracted. As a candidate cell to be used, the frame is virtually divided into small areas (sample areas) in the direction of continuous movement of the stage, and the drawing time A and each candidate necessary for drawing the sample area using only the basic figure aperture For each sample area, the reduced time ΔAc when the character is drawn with respect to the cell is drawn. After obtaining, by selecting a predetermined number to be formed on the aperture mask from each candidate cell, creating all combinations, and subtracting the sum of the corresponding shortening times ΔAc from the drawing time A for each combination, The drawing time in each sample area is calculated, the drawing time T in the sample area having the longest drawing time in each frame is obtained, and the drawing time in each frame is obtained by (frame length / sample area width) × T. Calculate the drawing time of the entire drawing area given by the sum of these frame drawing times, find the combination of candidate cells with the shortest drawing time of the entire drawing area, and form the aperture pattern on the aperture mask with this combination It is characterized by.
[0022]
Here, preferred embodiments of the present invention include the following.
(1) Set the stage movement speed optimally for each frame. More specifically, T is the drawing time of the sample area that takes the longest drawing time in the frame, and the stage moving speed is set by the width / T of the sample area.
(2) A candidate cell combination is obtained with a fixed number of character cells formed in the aperture mask.
(3) The number of character cells formed in the aperture mask is made variable according to the size of the candidate cell, and a combination of candidate cells is obtained. More specifically, the combination of candidate cells is obtained so that the sum of the areas of the candidate cells falls within a predetermined range.
[0023]
Another essence of the present invention is that, prior to the process of separating the pattern drawn by VSB and the pattern drawn by CP, the character beam is processed in units of graphic groups (cells) which are processing units of hierarchical graphic data processing. The graphic data conversion process is configured to determine the shape, replace the graphic group to be characterized with the CP control code, and output the VSB drawing pattern and the CP control code as drawing data for each drawing area.
[0024]
  That is, the present invention (Claim 2) includes a character beam formed in the same shape as a repeating unit that repeatedly appears in a pattern to be drawn, and a basic shape beam formed into a basic shape such as a rectangle and a right-angled isosceles triangle. In a charged beam drawing apparatus that draws a desired pattern in combination, a drawing data creation means for creating drawing pattern data from LSI design pattern data is stored in a predetermined figure while maintaining the hierarchical reference structure of the design pattern data. Hierarchical graphic data processing unit that performs data processing, and hierarchical structure recombination processing unit that reconfigures the hierarchical structure of design pattern data in advance to prevent occurrence of an error near the cell boundary due to hierarchical graphic data processing And a pattern for drawing the pattern data after hierarchical graphic data processing using a character beam. A character determination processing unit that separates the pattern group to be drawn by a combination of a basic group beam and a pattern group to be drawn, a character control code replacement processing unit to replace the pattern group to be drawn by the character beam with a control code indicating the type of beam shape, A drawing data output unit for each drawing area that allocates graphic data of a pattern drawn with a basic shape beam and a control code indicating the type of character beam to a drawable unit area.Further, the character determination processing unit analyzes the result of the graphic data processing in units of pattern groups (cells) which are processing units in the hierarchical graphic data processing, and obtains a cell identification number and a cell reference number. A cell information analysis process for calculating the cell size, the basic shape beam conversion shot number and the character beam conversion shot number, and a primary characterization candidate cell selection process for selecting a cell with a high characterization effect as a characterization candidate cell; Characterized candidate cell reconstruction process for processing pattern definition contents of cells selected as candidate cells, secondary characterization candidate cell selection process for selecting a predetermined number from the reconstructed candidate cells, and selected cells Characterization effect evaluation processing for evaluating the characterization effect for a combination of Formation process is repeated a second-order character of the candidate cell selection process and the character effect evaluation process to determine the appropriate combination of characters of the cellIt is what I did.
[0025]
  Here, preferred embodiments of the present invention include the following.
(1) In the characterization candidate cell reconstruction process, the pattern defined in the cell whose circumscribed rectangle of the pattern existence area is a predetermined size or less, or the pattern defined in the specified cell is used as the cell that refers to the cell. Partial hierarchical expansion processing to expand, pattern recognition processing that considers cells with the same pattern shape but different cell names as the same cell, and the reference pitch and maximum beam size of the cells that are array-referenced Combines the array reconstruction process that calculates the least common multiple and expands the cell array within the range and then divides the cell array by the maximum beam size, and the partial extraction process that partially extracts the pattern in the cell and creates a new cell. To process.
(2) As graphic processing in the character determination processing section, inter-graphic duplication removal, inter-layer graphic logical operation, graphic division at the unit deflection area boundary, resizing processing, and the like are performed.
  Further, the present invention (Claim 4) includes a character beam formed in the same shape as a repeating unit that repeatedly appears in a pattern to be drawn, and a basic shape beam formed into a basic shape such as a rectangle and a right-angled isosceles triangle. In a charged beam drawing apparatus that draws a desired pattern in combination, a drawing data creation means for creating drawing pattern data from LSI design pattern data is a predetermined figure while maintaining the hierarchical reference structure of the design pattern data. Hierarchical graphic data processing unit for performing data processing, and hierarchical structure recombination processing for reconfiguring the hierarchical structure of design pattern data in advance in order to prevent an error location from occurring near a cell boundary due to the hierarchical graphic data processing And the pattern data after processing the hierarchical graphic data is drawn with a character beam. A character determination processing unit for separating a pattern group to be drawn by a combination of a basic shape beam and a pattern group to be drawn, and a character control code replacement processing unit to replace the pattern group to be drawn by the character beam with a control code indicating the type of beam shape And a drawing data output unit for each drawing area that allocates graphic data of a pattern drawn by the basic shape beam and a control code indicating the type of the character beam to a drawable unit area. As an evaluation standard for evaluating the characterizing effect in the part, the difference between the drawing time when drawing with the basic shape beam and the drawing time when drawing with the character beam is calculated, and a predetermined number of cells with high characterizing effect are obtained. It is characterized by characterizing only.
[0026]
[Action]
According to the method of the present invention (Claim 1), the drawing time A necessary for drawing the sample area using only the basic figure aperture and the shortened time ΔAc when the candidate cell is drawn by characterizing can be obtained. For example, the drawing times of all combinations of CP candidate cells can be easily obtained from this. As a result, a combination of candidate cells with the shortest drawing time can be easily selected. That is, extraction of CP candidate cells for performing drawing with extremely high throughput by fully exhibiting the effects of the CP drawing method can be obtained by efficient processing in a short time.
[0027]
Further, according to the present invention (claims 2 to 4), since the hierarchical graphic data processing is performed before the VSB drawing pattern and the CP drawing pattern are separated, there is no desired error portion such as an overlap or a gap between the two. Drawing data can be created in a short time. For this reason, the data processing time for the separation process is shortened, and the throughput of the entire drawing system can be improved.
[0028]
In addition, a two-stage selection method in which the same cell as the processing unit of hierarchical graphic processing is first extracted as a primary character candidate, and candidate cell reconstruction processing such as division / fusion is performed, and then the characterization effect is evaluated again. Therefore, an appropriate combination of character figures can be determined in a short time. Furthermore, since the characterization effect is evaluated by the time difference between the VSB drawing and the CP drawing, it is possible to determine an appropriate character set that is more realistic.
[0029]
In view of the recent trend toward miniaturization and high integration of LSIs, it is expected that the processing performance of the charged beam drawing apparatus will be a serious problem. On the other hand, the drawing apparatus provided with the drawing data creating means as in the present invention can not only shorten the processing time required for drawing data conversion corresponding to the CP drawing method but also simultaneously improve the drawing processing performance, Its usefulness is great.
[0030]
【Example】
Embodiments of the present invention will be described below with reference to the drawings.
Example 1
FIG. 1 is a schematic block diagram showing an electron beam exposure apparatus used in the first embodiment of the present invention. In the figure, reference numeral 10 denotes a sample chamber, and a table (stage) 12 on which a sample 11 such as a semiconductor wafer or a glass mask is placed is accommodated in the sample chamber 10. The table 12 is driven by the table drive circuit 13 in the X direction (left and right direction on the paper surface) and in the Y direction (front and back direction on the paper surface). The position of the table 12 is measured by a position circuit 14 using a laser length meter or the like.
[0031]
An electron beam optical system 20 is disposed above the sample chamber 10. The optical system 20 includes an electron gun 21, various lenses 22 to 26, a blanking deflector 31, a beam size varying deflector 32, a beam scanning sub deflector 34, beam shaping aperture masks 35 and 36, and the like. It is configured.
[0032]
Then, the main deflector 33 positions in a predetermined unit drawing area (subfield), and the subdeflector 34 performs figure drawing positioning in the subfield, and also includes a beam shape control deflector 32 and a beam shaping aperture mask. The beam shape and beam dimensions are controlled by means of 35 and 36, and the drawing process is performed in units of drawing stripes in which the frame regions obtained by dividing the LSI chip into strips according to the beam deflection width while continuously moving the table 12 in the X or Y direction. To do. Further, the table 12 is moved to the start position of the next drawing stripe, and the above processing is repeated to sequentially draw each drawing stripe.
[0033]
On the other hand, a storage medium 41 such as a magnetic disk is connected to the control computer 40, and drawing data relating to the LSI chip to be drawn is stored on the magnetic disk 41. The drawing data read from the magnetic disk 41 is temporarily stored in the pattern memory 42 as drawing data for each drawing stripe. The drawing data for each drawing stripe stored in the pattern memory 42, that is, the drawing stripe data composed of the drawing position and the basic graphic data is decoded by the pattern data decoder 43 and the drawing data decoder 44, which are data decoding units, It is sent to the ranking circuit 45, the beam shaping driver 46, the main deflection driver 47 and the sub deflection driver 48.
[0034]
That is, in the pattern decoder 43, the drawing stripe data is input, and the basic figure data defined as the drawing stripe data is converted into a CP cell figure, rectangle or triangle that can be formed by the combination of the beam shaping aperture masks 35 and 36. The drawing unit graphic group is divided into graphic groups and sent to a beam forming driver 46 created based on the drawing graphic information. A predetermined deflection signal is applied from the beam shaping driver 46 to the beam shape control deflector 32 of the electron optical system 20 to control the beam shape and beam size of the electron beam.
[0035]
Next, a beam generation method in the above-described apparatus will be described in detail. FIG. 2A shows the arrangement of the beam passage holes of the first beam shaping aperture mask 35, in which one beam passage hole 35a is formed. FIG. 2B shows the arrangement of the beam passage holes of the second beam shaping aperture mask 36, and first to sixth passage holes 36a to 36f are formed.
[0036]
FIG. 3 shows a desired beam by electro-optical superposition of an aperture image (first shaping aperture image) of the first beam shaping aperture mask 35 and a beam passage hole formed in the second beam shaping aperture mask 36. The appearance of generating shapes and beam dimensions is shown. FIG. 3A shows a state in which a rectangular beam is generated by a combination of the first shaped aperture image and the first beam passage hole 36 a of the second shaped aperture mask 36. FIG. 3B shows a state in which a triangular beam is generated by a combination of the first shaped aperture image and the second beam passage hole 36 b of the second shaped aperture mask 36. The rectangular beam and the triangular beam become unit drawing figures for drawing a pattern portion that cannot be collectively transferred by the aperture projection image of the second shaping aperture mask 36.
[0037]
FIG. 3C shows a beam corresponding to a part of design data by a combination of the first shaping aperture image and any of the CP cell projection beam passage holes 36c to 36f formed on the second shaping aperture mask 36. FIG. An image is being generated. Using the beam generation method described above in detail, as shown in the example of FIG. 4, for example, the LSI chip is exposed using a batch transfer beam and a basic figure beam made of a triangle or rectangle.
[0038]
Next, how to select a CP aperture figure to be formed on the aperture mask when drawing using the apparatus of this embodiment will be described. FIG. 5 is a flowchart showing the flow of processing when determining a CP cell according to the present embodiment. The contents of the processing will be described in detail according to this flowchart.
[0039]
First, a repetitive pattern is extracted from the LSI pattern design (S1). Next, the extracted repetitive patterns are divided or merged, and the pattern size is optimized so as to be about the size of the CP aperture (S2). Hereinafter, the repeated pattern optimized in this way is referred to as a CP candidate cell.
[0040]
Next, the number of reference times of the CP candidate cells in the entire LSI and the number of shots when each CP candidate cell is exposed using only the variable shaped beam are obtained (S3). For example, when analyzing the design data of the LSI pattern to be drawn, it was found that there are four types of repeating patterns A to D as shown in FIG. Further, four types of repetitive patterns were optimized to a size of about the CP aperture size, and eight CP candidate cells A, B1, B2, C1, C2, C3, C4, D as shown in FIG. 7 were created. However, in this example, a simple division method is used. The results of obtaining the number of reference times and the number of shots of the eight CP candidate cells are shown in (Table 1) below.
[0041]
[Table 1]

[0042]
Next, a predetermined number of candidate cells that can be formed in the aperture mask are selected from the CP candidate cells, and possible combinations are created (S4). For example, as shown in FIG. 2, when four CP apertures are possible on the aperture mask, in the example of FIG.₈C_Four = 70 combinations are created.
[0043]
Next, according to the flowchart of FIG. 8, for all combinations of CP candidate cells, the drawing time when the CP candidate cells belonging to each combination are exposed as character cells is calculated (S5). Hereinafter, this calculation method will be described in detail.
[0044]
First, the entire chip is divided into a frame and a sample area, and a necessary drawing time A when drawing each sample area with only a variable shaped beam is obtained. At the same time, the drawing reduction time ΔAc of each sample area for each candidate cell when the candidate cell c is subjected to CP exposure is obtained. Here, the frame is an area obtained by dividing the drawing area into stripes having a predetermined width determined by the deflection width of the deflection system, and the sample area is an area obtained by virtually dividing the frame into small areas in the stage continuous movement direction.
[0045]
Hereinafter, the drawing time A when the j-th sample area of the i-th frame is drawn using only the variable shaped beam is newly set to t^ijIf we write VSB, in the case of the above-mentioned drawing device,
t^ijVSB = Nsub * ts1 + (Nshot-Nsub) * ts2 + Nshot * td (1)
Given in. However,
ts1: Main deflection settling time
ts2: Sub deflection settling time
td: Irradiation time
Nsub: number of subfields in the jth sample area of the ith frame
Nshot: Number of shots in the j-th sample area of the i-th frame
Subfield: Area that can be exposed only by switching the sub deflection
It is.
[0046]
In addition, in the j-th sample area of the i-th frame, a new shortened drawing time ΔAc compared with the case where only the candidate cell c is subjected to CP exposure compared with the case where only the variable shaped beam is exposed is Δt^ijIf you write c,
Δt^ijc = (Nshot, c (VSB) −Nshot, c (CP)) × (ts2 + td) (2)
It becomes. However,
Nshot, c (VSB): Total number of shots when all cells C in the j-th sample area of the i-th frame are exposed with only VSB
Nshot, c (CP): Total number of shots when all cells C in the j-th sample area of the i-th frame are subjected to CP exposure
It is. Similarly, for all sample areas, t^ijVSB and Δt^ijc is obtained. Δt^ijFor c, it is determined for all CP candidate cells.
[0047]
Next, for each of the CP candidate cell combinations described above, the drawing time required to draw the entire chip is set to t^ijVSB and Δt^ijA method of obtaining from c will be described. With respect to the candidate cells in FIG. 7, a calculation method will be described by taking a combination (A, B1, C1, D) as an example. First, assuming that A, B1, C1, and D are exposed by the CP method and the remaining portion of the chip is exposed by the variable shaped beam, the required drawing time in each sample area of each frame is calculated by the following equation.
[0048]
t^ij= T^ijVSB -Δt^ijA -Δt^ijB1−Δt^ijC1-Δt^ijD… (3)
That is, as shown in FIG. 9, the necessary drawing time in each sample area may be obtained by subtracting the shortened drawing time by characterization from the time required for drawing using only the variable shaped beam. FIG. 9A shows the necessary drawing time in each sample area, and FIG. 9B shows a schematic diagram of the corresponding frame.
[0049]
Next, the stage speed in each frame is calculated based on this value. First, the sample area with the longest required drawing time is obtained for each frame, and the highest speed at which this sample area can be drawn is set as the stage speed. That is, if the sample area having the long required drawing time on the i-th frame is the m-th sample area, the stage speed v^imIs
v^im= L / t^im          … (Four)
It becomes. Where l is the width of the sample area.
[0050]
Therefore, the actual drawing time t of the i-th frameⁱ Is
tⁱ = L / v^im          … (Five)
It is obtained.
[0051]
If this process is performed for all frames, the total drawing time Tall is
Tall =Σtⁱ       … (6)
ⁱ
Then you can find it.
[0052]
The above processing is performed for all combinations of candidate cells, the drawing time of the entire chip is obtained for each combination, and the CP candidate cell of the combination with the shortest drawing time is selected as the aperture figure for CP of the aperture mask (S6). That is,
Top = Min (T¹ _all , T² _all ,…)… (7)
Then, CP candidate cells belonging to the combination giving Top are created as CP aperture figures of the aperture mask. However, T^k _all Is the drawing time of the whole chip when the k-th CP candidate cell combination is subjected to CP exposure.
[0053]
As described above, according to the present embodiment, in the stage continuous movement type electron beam exposure apparatus, the combination of CP candidate cells with the shortest drawing time is formed on the aperture mask, thereby reducing the number of exposure shots. The drawing throughput can be improved.
[0054]
In this case, the drawing time A when only the variable shaped beam of each sample area is used and the drawing time ΔAc that is shortened when each CP candidate cell c is characterized are once obtained for all sample areas. And ΔAc are used to calculate the stage speed of each combination of CP candidate cells and the drawing time derived therefrom, so that the combination of candidate cells with the shortest drawing time can be easily selected. That is, when obtaining a repetitive pattern that is most effective in shortening the drawing time, it is possible to omit a useless procedure and make an efficient selection.
(Example 2)
FIG. 10 is a block diagram showing a second embodiment of the present invention, and particularly shows the configuration of drawing data creating means for creating drawing data from design data.
[0055]
The drawing apparatus of the present embodiment includes drawing data creation means 50 and drawing means 60. The drawing data creation means 50 includes a hierarchical structure rearrangement processing unit 51, a hierarchical graphic data processing unit 52, a characterization cell determination processing unit 53, a character control code replacement processing unit 54, and a drawing data output unit 55 for each drawing area. Yes. The drawing means 60 corresponds to the drawing apparatus main body and is the same as that shown in FIG.
[0056]
The hierarchical structure rearrangement processing unit 51 is a preprocessing for the subsequent hierarchical graphic data processing. The design data is read and the hierarchical structure is reconstructed, and graphic overlapping or gaps are formed near the cell boundary by the hierarchical graphic data processing. This prevents the occurrence of error. The hierarchical graphic data processing unit 52 performs graphic data processing such as graphic duplication removal, inter-layer graphic logical operation, resizing, and magnification correction for each cell while maintaining the hierarchical structure of the design data.
[0057]
As shown in FIG. 11, the characterization cell determination processing unit 53 uses the regular pattern region as a CP drawing region and separates it from the other basic shape beam drawing regions, and based on the processing result of the hierarchical graphic data processing. Next, the characterizing effect is evaluated for each cell to determine a set of characterizing cells. By adopting the cell defined by the designer as the repeating unit at the time of designing the circuit pattern as the unit of the graphic group for characterizing, it is possible to greatly reduce the trouble of extracting the repeating unit from the circuit pattern.
[0058]
The character control code replacement processing unit 54 is a process of extracting graphic data of a cell that has been characterized and replacing it with a character control code. The character control code is a control code for identifying the character shape when the drawing means generates a character beam.
[0059]
The final drawing area-specific drawing data output unit 55 distributes the VSB drawing data and the character control code in units of drawing areas (frame areas, etc.) according to the drawing position and outputs them. With the above processing, drawing data corresponding to the CP drawing method is created.
[0060]
Next, the details of the characterizing cell determination process will be described based on the block diagram of the characterizing cell determination processing unit 53 shown in FIG. The characterization cell determination processing unit 53 includes a cell information analysis unit 53a, a primary candidate cell selection unit 53b, a candidate cell reconstruction unit 53c, a secondary candidate cell selection unit 53d, and a characterization effect evaluation unit 53e. .
[0061]
The cell information analysis unit 53a reads in-cell graphic data and cell arrangement data, which are processing results of hierarchical graphic data processing, and creates a cell information table. This cell information table shows the cell identification number assigned to each cell, the cell size that is the size of the circumscribed rectangle of the figure defined in the cell, the total number of references in the chip of the cell, and the basic shape beam And the number of shots when characterized.
[0062]
Based on this cell information table, cells having a relatively high characterization effect are selected as primary characterization candidate cells by the primary candidate cell selection unit 53b, and the number of cells to be processed thereafter is narrowed down. Reduce the processing load.
[0063]
By characterizing cells as units, it becomes much easier to extract repeat units from a circuit pattern. However, since the cell is a graphic group defined at the time of design, the size and the number of types are arbitrary. On the other hand, the size and the number of types of characters are limited due to restrictions on the device specifications. Therefore, the primary character candidate cell may not be characterized as it is. Therefore, it is necessary to reconfigure the definition contents of the candidate cells so that the cell size is appropriate and the limited number of character types can be utilized to the maximum extent. The candidate cell reconfiguration unit 53c performs the following processing.
(1) Partial expansion of hierarchical structure
(2) Partial pattern extraction
(3) Pattern recognition
(4) Array reconstruction
(5) Cell division
Details of each processing content will be described below.
(1) Partial expansion of hierarchical structure
As shown in FIG. 13 (a), it is assumed that cells B are arranged in an array in cell A, and cells F in cell B are arranged in a reference manner in four types of arrangement directions (F subscripts are arranged). Show directions). When this reference structure is represented by a tree notation, FIG. 13B is obtained. The cell size of the cell F is ½ or less of the maximum beam size, and if the cell B is characterized in units, the area of the four cells F can be drawn collectively. However, although the cell F is referenced and arranged in the cell B, the figure is not defined, and if the cell B is simply characterized, the content becomes an empty character.
[0064]
Therefore, the hierarchical structure below the designated cell is expanded, and the graphic defined in the lower cell is given to the designated cell. In this example, a graphic defined in the cells F1 to F4 referred to in the cells B and below is given to the cell B. Thereby, the characterization effect of the cell B increases.
(2) Partial pattern extraction
As shown in FIG. 14, the cell E and the cell P are similar patterns to the cell F. If the similar part can be cut out and replaced with the same character, the efficiency of character drawing increases. Therefore, a function for cutting out a part of the pattern in the cell and making it a new candidate cell is required.
(3) Pattern recognition
Although the cell identification numbers are different, there are cases where the same pattern is defined inside. If these are extracted as separate characters or one of them is VSB drawn, the efficiency of CP drawing is lowered. Therefore, pattern recognition is performed to inspect the pattern, and cell identification numbers are unified for cells having the same content. When combined with the pattern part extraction in the previous item (2), it becomes possible to characterize a part of the cell including the similar graphic.
(4) Array reconstruction
An example in which candidate cells smaller than the maximum beam size are arranged in an array is shown in FIG. When the maximum beam size is 3 μm and the array arrangement pitch is 2 μm, the least common multiple of the maximum beam size and the array arrangement pitch is 6 μm. When candidate cells are expanded in this area, 3 × 3 = 9 cells are entered as shown in FIG. When this is divided by the maximum beam size, it is divided into four cells as shown in FIG. 15B, and the array arrangement structure is reconfigured. When each cell created by this array reconstruction is characterized, the number of character types increases, but the total number of shots decreases. In the example shown here, the number of characters increases from 1 to 4, and the total number of shots decreases to 4/9.
[0065]
After performing the above candidate cell reconstruction, the secondary candidate cell selection unit 53d selects a predetermined number of candidate cells as secondary candidate cells, and the characterization effect evaluation unit 53e evaluates the characterization effect. Here, as a criterion for evaluating the characterization effect, if the difference between the drawing time when the selected secondary candidate cell is rendered with VSB and the rendering time when the characterization is performed is calculated and used, a more realistic evaluation is performed. Thus, it becomes possible to evaluate an appropriate characterizing effect.
[0066]
Furthermore, the combination of candidate cells to be secondarily selected is variously evaluated to evaluate the characterization effect, and in some cases, again by repeating candidate cell reconstruction, secondary candidate cell selection, and characterization effect evaluation, A more appropriate character set can be obtained.
[0067]
As described above, according to the present embodiment, the drawing data creating means 50 is configured as shown in FIG. 10, and the graphic pattern data processing is performed before the VSB drawing pattern and the CP drawing pattern are separated. Therefore, it is possible to easily extract a group of figures to be drawn together by a character beam, and to create drawing data having consistency with a pattern to be drawn by VSB. And since the set of extracted character figures becomes appropriate, the time required for drawing data creation can be shortened and the drawing throughput can be improved. As a result, the operating rate of the electron beam drawing apparatus can be increased and the productivity of the LSI can be increased.
[0068]
In addition, this invention is not limited to each Example mentioned above. In the first embodiment, the number of character cells to be formed in the aperture mask is four and the combination of candidate cells is obtained. However, the number of character cells is not limited to this and can be changed as appropriate. Furthermore, the number of character cells is not necessarily constant, and the number of character cells may be variable depending on the size of the candidate cell. For example, when the sizes of the candidate cells are different, the combination of candidate cells may be obtained so that the sum of the areas of the candidate cells falls within a predetermined range.
[0069]
The configuration of the apparatus for carrying out the present invention is not limited to that shown in FIG. 1, and the stage on which the sample is placed is continuously moved using an aperture mask in which an aperture for a basic figure and an aperture for a character are formed. However, any frame may be used as long as the frames divided into a predetermined width are sequentially drawn. Further, although the electron beam exposure apparatus has been described in the embodiments, it is needless to say that the present invention can be similarly applied to an ion beam exposure apparatus.
[0070]
Further, the drawing throughput can be further improved by using the drawing data creation process described in the second embodiment in combination with the aperture pattern determination process described in the first embodiment. In addition, various modifications can be made without departing from the scope of the present invention.
[0071]
【The invention's effect】
As described above in detail, according to the present invention (claim 1), the drawing time A when only the variable shaped beam of each sample area is used and the drawing time shortened when each CP candidate cell c is characterized. ΔAc is once obtained for all sample areas, and the stage speed of each combination of CP candidate cells and the drawing time derived therefrom are calculated from A and ΔAc, so that the effect of the CP drawing method can be fully exhibited and the throughput can be improved. It is possible to realize a charged beam drawing method capable of performing extremely high drawing and easily extracting CP cells to be formed on the aperture mask.
[0072]
Further, according to the present invention (claims 2 to 4), prior to the process of separating the pattern drawn by VSB from the pattern drawn by CP, a graphic group (cell) which is a processing unit of hierarchical graphic data processing The shape of the character beam is determined in units of characters, the graphic group to be characterized is replaced with the CP control code, and the VSB drawing pattern and the CP control code are output as drawing data for each drawing area. This makes it possible to easily extract a group of figures for batch drawing and to create drawing data while maintaining consistency with the pattern to be drawn by VSB, improving the drawing throughput by reducing the time required for drawing data creation. It is possible to realize a charged beam drawing apparatus capable of measuring the above.
[Brief description of the drawings]
FIG. 1 is a schematic configuration diagram showing an electron beam drawing apparatus used in a first embodiment.
FIG. 2 is a schematic diagram showing the structure of a beam shaping aperture.
FIG. 3 is a schematic diagram showing the shape of a shaped beam.
FIG. 4 is a schematic diagram showing a drawing data system;
FIG. 5 is a flowchart showing an outline of processing in the embodiment method.
FIG. 6 is a schematic diagram illustrating an example of a repetitive pattern drawn in the embodiment.
FIG. 7 is a schematic diagram showing an example of a CP candidate cell used in the embodiment.
FIG. 8 is a flowchart for explaining a drawing time calculation process in the embodiment;
FIG. 9 is a schematic diagram showing a method for calculating the stage speed in the embodiment.
FIG. 10 is a block diagram illustrating a configuration of drawing data creating means for explaining a second embodiment;
FIG. 11 is a diagram showing a pattern example of a character beam drawing area and a basic shape beam drawing area.
FIG. 12 is a block diagram showing a configuration of a characterization cell determination processing unit.
FIG. 13 is a diagram for explaining a partial expansion process of a hierarchical structure.
FIG. 14 is a diagram for explaining pattern part extraction processing;
FIG. 15 is a diagram for explaining array reconfiguration processing;
FIG. 16 is a schematic diagram for explaining a conventional stage continuous movement type drawing method;
[Explanation of symbols]
10 ... Sample chamber 11 ... Sample
12 ... Table 13 ... Table drive circuit
14 ... Position circuit 20 ... Electro-optical system
21 ... Electron gun 22-26 ... Lens
31-34 ... Deflector 35, 36 ... Beam shaping aperture
40 ... control computer 41 ... magnetic disk (storage medium)
42 ... Pattern memory 43 ... Pattern data decoder
44 ... Drawing data decoder 45 ... Blanking circuit
46 ... Beam shaper driver 47 ... Main deflection circuit driver
48 ... Sub deflector driver 50 ... Drawing data creation means
51 ... Hierarchical rearrangement processing unit 52 ... Hierarchical graphic data processing unit
53 ... Characterized cell determination processing unit 54 ... Character control code replacement processing unit
55 ... Drawing data output unit for each drawing area 60 ... Drawing means

Claims (4)

Using an aperture mask on which a basic figure aperture for forming a shaped beam corresponding to the basic figure and a character aperture corresponding to a repetitive pattern is formed, while the stage on which the sample is placed is continuously moved, a predetermined width is obtained. In a charged beam drawing method for sequentially drawing divided drawing regions (frames),
Extract each pattern to be used repeatedly from the desired data to be drawn and make it a candidate cell to be a character cell.
The frame is virtually divided into small areas (sample areas) in the direction of continuous movement of the stage, and the sample area is drawn for the drawing time A and each candidate cell required to draw the sample area using only the basic figure aperture. After obtaining the respective shortened times ΔAc for all the sample areas when characterizing and drawing
By selecting a predetermined number of character apertures to be formed on the aperture mask from each candidate cell, creating all combinations, and subtracting the sum of the corresponding reduction times ΔAc from the drawing time A for all combinations, Calculate the drawing time in each sample area,
The drawing time T of the sample area with the longest drawing time in each frame is obtained, the drawing time of each frame is obtained by (frame length / sample area width) × T, and the sum of these frame drawing times is obtained. Calculate the drawing time of the entire drawing area given,
A charged beam drawing method comprising: obtaining a combination of candidate cells having the shortest drawing time in the entire drawing area, and forming an aperture pattern on the aperture mask by the combination. A charged beam drawing apparatus that draws a desired pattern by combining a character beam formed in the same shape as a repeating unit that repeatedly appears in a pattern to be drawn and a basic shape beam shaped into a basic shape such as a rectangle and a right isosceles triangle In
The drawing data creation means for creating the drawing pattern data from the LSI design pattern data is
Hierarchical graphic data processing unit that performs predetermined graphic data processing while maintaining the hierarchical reference structure of design pattern data, and to prevent occurrence of an error near the cell boundary due to the hierarchical graphic data processing A hierarchical structure rearrangement processing unit for reconfiguring the hierarchical structure of the design pattern data in advance, a pattern group for drawing the pattern data after processing the hierarchical graphic data by a combination of a character beam and a basic shape beam, and A character determination processing unit that separates the pattern group, a character control code replacement processing unit that replaces a pattern group drawn by the character beam with a control code indicating a type of beam shape, graphic data of a pattern drawn by the basic shape beam, and the A control code indicating the type of character beam can be drawn. A drawing area by the drawing data output unit to allocate to the unit area, Ri Tona,
The character determination processing unit analyzes a result of the graphic data processing in units of pattern groups (cells) which are processing units in the hierarchical graphic data processing, and obtains a cell identification number, a cell reference number, and a cell size. , A cell information analysis process for calculating the basic shape beam conversion shot number and the character beam conversion shot number, a primary characterization candidate cell selection process for selecting a cell with a high characterization effect as a characterization candidate cell, and a candidate cell Combination of selected characterization candidate cell reconstruction process for processing pattern definition content of selected cell, secondary characterization candidate cell selection process for selecting a predetermined number from reconstructed candidate cells, and selected cells It consists of a characterization effect evaluation process for evaluating a characterization effect,
A charged beam drawing apparatus characterized in that an appropriate combination of characterized cells is determined by repeating characterizing candidate cell reconstruction processing, secondary characterizing candidate cell selection processing, and characterizing effect evaluation processing . In the characterization candidate cell reconstruction process, a cell in which a circumscribed rectangle of a pattern existing area is defined in a cell having a predetermined size or less, or a cell that refers to a pattern defined in a specified cell A partial hierarchy expansion process step that expands to a partial extraction process step that partially extracts a graphic group in a cell and defines it as a new cell. Pattern recognition processing step that is regarded as the same cell if it is the same, an array in which the reference pitch and the least common multiple of the maximum beam size of the cell that is referred to the array are obtained, the cell array is expanded within the range, and then divided by the maximum beam size The charged beam drawing apparatus according to claim 2 , comprising a reconstruction processing step. A charged beam drawing apparatus that draws a desired pattern by combining a character beam formed in the same shape as a repeating unit that repeatedly appears in a pattern to be drawn and a basic shape beam shaped into a basic shape such as a rectangle and a right isosceles triangle In
The drawing data creation means for creating the drawing pattern data from the LSI design pattern data is
Hierarchical graphic data processing unit that performs predetermined graphic data processing while maintaining the hierarchical reference structure of design pattern data, and to prevent occurrence of an error near the cell boundary due to the hierarchical graphic data processing A hierarchical structure rearrangement processing unit for reconfiguring the hierarchical structure of the design pattern data in advance, a pattern group for drawing the pattern data after processing the hierarchical graphic data by a combination of a character beam and a basic shape beam, and A character determination processing unit that separates the pattern group, a character control code replacement processing unit that replaces a pattern group drawn by the character beam with a control code indicating a type of beam shape, graphic data of a pattern drawn by the basic shape beam, and the A control code indicating the type of character beam can be drawn. A drawing area by the drawing data output unit to allocate to the unit area, consists,
As an evaluation standard for evaluating the characterizing effect in the character determination processing unit, the difference between the drawing time when drawing with the basic shape beam and the drawing time when drawing with the character beam is calculated, and the characterizing effect is high. A charged beam drawing apparatus characterized by characterizing a predetermined number of cells .

Images