JP4102966B2

JP4102966B2 - Pulling method of silicon single crystal

Info

Publication number: JP4102966B2
Application number: JP2001192162A
Authority: JP
Inventors: 森林符; 洋二鈴木; 和浩原田; 久降屋; 朗樋口
Original assignee: Sumco Corp
Current assignee: Sumco Corp
Priority date: 2001-06-26
Filing date: 2001-06-26
Publication date: 2008-06-18
Anticipated expiration: 2021-06-26
Also published as: JP2003002784A

Description

【０００１】
【発明の属する技術分野】
本発明は、シリコン融液にカスプ（ＣＵＳＰ）磁場を印加しながら、シリコン単結晶のインゴットをシリコン融液から引上げる方法に関するものである。
【０００２】
【従来の技術】
従来、シリコン単結晶の製造方法として、シリコン単結晶のインゴットをチョクラルスキー法（以下、ＣＺ法という）により引上げる方法が知られている。このＣＺ法は、石英るつぼに貯留されたシリコン融液に種結晶を接触させ、石英るつぼ及び種結晶を回転させながら種結晶を引上げることにより、円柱状のシリコン単結晶のインゴットを製造する方法である。
【０００３】
一方、半導体集積回路を製造する工程において、歩留りを低下させる原因として酸化誘起積層欠陥（Oxidation Induced Stacking Fault、以下、ＯＳＦという。）の核となる酸素析出物の微小欠陥や、結晶に起因したパーティクル（Crystal Originated Particle、以下、ＣＯＰという。）や、或いは侵入型転位（Interstitial-type Large Dislocation、以下、ＬＤという。）の存在が挙げられている。ＯＳＦは、結晶成長時にその核となる微小欠陥が導入され、半導体デバイスを製造する際の熱酸化工程等で顕在化し、作製したデバイスのリーク電流の増加等の不良原因になる。またＣＯＰは、鏡面研磨後のシリコンウェーハをアンモニアと過酸化水素の混合液で洗浄したときにウェーハ表面に出現する結晶に起因したピットである。このウェーハをパーティクルカウンタで測定すると、このピットも本来のパーティクルとともに光散乱欠陥として検出される。
【０００４】
このＣＯＰは電気的特性、例えば酸化膜の経時絶縁破壊特性（Time Dependent dielectric Breakdown、ＴＤＤＢ）、酸化膜耐圧特性（Time Zero Dielectric Breakdown、ＴＺＤＢ）等を劣化させる原因となる。またＣＯＰがウェーハ表面に存在するとデバイスの配線工程において段差を生じ、断線の原因となり得る。そして素子分離部分においてもリーク等の原因となり、製品の歩留りを低くする。更にＬＤは、転位クラスタとも呼ばれたり、或いはこの欠陥を生じたシリコンウェーハをフッ酸を主成分とする選択エッチング液に浸漬するとピットを生じることから転位ピットとも呼ばれる。このＬＤも、電気的特性、例えばリーク特性、アイソレーション特性等を劣化させる原因となる。この結果、半導体集積回路を製造するために用いられるシリコンウェーハからＯＳＦ、ＣＯＰ及びＬＤを減少させることが必要となっている。
【０００５】
このＯＳＦ、ＣＯＰ及びＬＤを有しない無欠陥のシリコンウェーハを切出すためのシリコン単結晶インゴットの製造方法が特開平１１−１３９３号公報に開示されている。一般に、シリコン単結晶のインゴットを速い速度で引上げると、インゴット内部に空孔型点欠陥の凝集体が支配的に存在する領域［Ｖ］が形成され、インゴットを遅い速度で引上げると、インゴット内部に格子間シリコン型点欠陥の凝集体が支配的に存在する領域［Ｉ］が形成される。このため上記製造方法では、インゴットを最適な引上げ速度で引上げることにより、上記点欠陥の凝集体が存在しないパーフェクト領域［Ｐ］からなるシリコン単結晶を製造できるようになっている。
【０００６】
【発明が解決しようとする課題】
しかし、上記従来の特開平１１−１３９３号公報に示されたシリコン単結晶インゴットの製造方法では、シリコン単結晶のインゴットとシリコン融液との固液界面近傍での鉛直方向の温度勾配が均一になるように制御する必要があり、この制御はシリコン融液の残量の変化や対流の変化による影響を受けるため、インゴットの直胴部全長にわたって、無欠陥のシリコン単結晶を製造することは困難であった。
本発明の目的は、比較的少ない電力及び比較的狭い空間で、無欠陥であって、しかも酸素濃度が制御されたシリコン単結晶のインゴットを比較的容易に製造できる、シリコン単結晶の引上げ方法を提供することにある。
【０００７】
【課題を解決するための手段】
請求項１に係る発明は、図１に示すように、シリコン融液１３を貯留する石英るつぼ１４を所定の回転速度で回転させ、シリコン融液１３から引上げられるシリコン単結晶のインゴット１６を所定の回転速度で回転させ、石英るつぼ１４の外径より大きなコイル直径を有する第１及び第２コイル１１，１２を石英るつぼ１４の回転軸をそれぞれコイル中心としかつ鉛直方向に所定の間隔をあけて配設し、第１及び第２コイルに互いに逆向きの電流を流すことにより第１及び第２コイルの各コイル中心から第１及び第２コイル間の中立面１７ａを通るカスプ磁場１７を発生させ、インゴット１６内が格子間シリコン型点欠陥の凝集体及び空孔型点欠陥の凝集体の存在しないパーフェクト領域となる引上げ速度でインゴット１６を引上げるシリコン単結晶の引上げ方法の改良である。
その特徴ある構成は、カスプ磁場１７の中立面１７ａとシリコン融液１３の表面との距離をＨとするとき、２８０ｍｍ≦Ｈ≦７００ｍｍを満たすように、中立面１７ａをシリコン融液１３の表面の下方に制御するか、或いは１４０ｍｍ≦Ｈ≦６００ｍｍを満たすように、中立面１７ａをシリコン融液１３の表面の上方に制御し、インゴット１６の直径が３００ｍｍでありかつ石英るつぼ１４の内径が６５０ｍｍであるとき、カスプ磁場１７の水平方向の強度を０．０２テスラに制御し、石英るつぼ１４を４ｒｐｍの回転速度で回転させかつインゴット１６を石英るつぼ１４とは反対方向に１２〜８ｒｐｍの回転速度で回転させながらインゴット１６を０．２〜０．５ｍｍ／分の引上げ速度で引上げるところにある。
【０００８】
この請求項１に記載されたシリコン単結晶の引上げ方法では、カスプ磁場１７の中立面の位置と、カスプ磁場の強度と、石英るつぼ１４の回転速度と、インゴット１６の回転速度とを制御しながら、インゴットを引上げると、シリコン融液１３に所定の対流２１〜２３が発生し、これらの対流２１〜２３により固液界面１９形状が上側に凸となる。この結果、固液界面の中心がシリコン融液１３表面の延長面上より上方に位置するため、固液界面１９の中心における鉛直方向の温度勾配が大きくなり、固液界面の中心における鉛直方向の温度勾配と、固液界面の周縁における鉛直方向の温度勾配との差が小さくなる。従って、略全長にわたって無欠陥で高品質のシリコン単結晶のインゴット１６を比較的容易に製造できる。
【０００９】
請求項２に係る発明は、図１に示すように、シリコン融液１３を貯留する石英るつぼ１４を所定の回転速度で回転させ、シリコン融液１３から引上げられるシリコン単結晶のインゴット１６を所定の回転速度で回転させ、石英るつぼ１４の外径より大きなコイル直径を有する第１及び第２コイル１１，１２を石英るつぼ１４の回転軸をそれぞれコイル中心としかつ鉛直方向に所定の間隔をあけて配設し、第１及び第２コイルに互いに逆向きの電流を流すことにより第１及び第２コイルの各コイル中心から第１及び第２コイル間の中立面１７ａを通るカスプ磁場１７を発生させ、インゴット１６内が格子間シリコン型点欠陥の凝集体及び空孔型点欠陥の凝集体の存在しないパーフェクト領域となる引上げ速度でインゴット１６を引上げるシリコン単結晶の引上げ方法の改良である。
その特徴ある構成は、カスプ磁場１７の中立面１７ａとシリコン融液１３の表面との距離をＨとするとき、Ｈ＝５００ｍｍを満たすように、中立面１７ａをシリコン融液１３の表面の下方に制御し、インゴット１６の直径が２００ｍｍでありかつ石英るつぼ１４の内径が６００ｍｍであるとき、カスプ磁場１７の水平方向の強度を０．０１８テスラに制御し、石英るつぼ１４を３〜３．５ｒｐｍの回転速度で回転させかつインゴット１６を石英るつぼ１４とは反対方向に２０〜１６ｒｐｍの回転速度で回転させながらインゴット１６を０．３〜０．７ｍｍ／分の引上げ速度で引上げるところにある。
【００１１】
【発明の実施の形態】
次に本発明の実施の形態を図面に基づいて説明する。
図１に示すように、本発明のシリコン単結晶の引上げ方法は、シリコン融液１３を貯留する石英るつぼ１４を所定の回転速度Ｒ₁で回転させ、シリコン融液１３から引上げられるシリコン単結晶のインゴット１６を所定の回転速度Ｒ₂で回転させ、かつシリコン融液１３に第１及び第２コイル１１，１２を用いてカスプ磁場１７を印加しながら、上記シリコン融液１３から上記インゴット１６を引上げる方法である。上記第１及び第２コイル１１，１２は、石英るつぼ１４の外径より大きなコイル直径を有し、石英るつぼの回転軸をそれぞれコイル中心としかつ鉛直方向に所定の間隔をあけて配設される。また第１及び第２コイル１１，１２には互いに逆向きの電流が流され、これにより第１及び第２コイルの各コイル中心から第１及び第２コイル間の中立面１７ａを通るカスプ磁場１７が発生するようになっている。なお、上記中立面１７ａは、第１及び第２コイル１１，１２間における、鉛直方向の磁場強度がゼロとなる水平面である。また、図１の符号１８は石英るつぼ１４の外周面を包囲するヒータである。
【００１２】
一方、上記インゴット１６は、このインゴット内が格子間シリコン型点欠陥の凝集体及び空孔型点欠陥の凝集体の存在しないパーフェクト領域となる引上げ速度で引上げられる。即ち、インゴットは、ＣＺ法によりホットゾーン炉内のシリコン融液１３からボロンコフ（Voronkov）の理論に基づいた所定の引上げ速度プロファイルで引上げられる。
【００１３】
一般的に、ＣＺ法によりホットゾーン炉内のシリコン融液１３からシリコン単結晶のインゴット１６を引上げると、インゴット内には、点欠陥（point defect）と点欠陥の凝集体（agglomerates：三次元欠陥）が発生する。点欠陥は空孔型点欠陥と格子間シリコン型点欠陥という二つの一般的な形態がある。空孔型点欠陥は一つのシリコン原子がシリコン結晶格子で正常的な位置の一つから離脱したものである。このような空孔が空孔型点欠陥になる。一方、原子がシリコン結晶の格子点以外の位置（インタースチシャルサイト）で発見されるとこれが格子間シリコン点欠陥になる。
【００１４】
点欠陥は一般的にシリコン融液１３とインゴット１６の間の接触面、即ち固液界面１９で形成される。しかし、インゴット１６を継続的に引上げることによって固液界面１９であった部分は引上げとともに冷却し始める。冷却の間、空孔型点欠陥又は格子間シリコン型点欠陥は拡散により互いに合併して、空孔型点欠陥の凝集体（vacancy agglomerates）又は格子間シリコン型点欠陥の凝集体（interstitial agglomerates）が形成される。言い換えれば、凝集体は点欠陥の合併に起因して発生する三次元構造となる。
【００１５】
空孔型点欠陥の凝集体は、前述したＣＯＰの他に、ＬＳＴＤ（Laser Scattering Tomograph Defects）又はＦＰＤ（Flow Pattern Defects）と呼ばれる欠陥を含み、格子間シリコン型点欠陥の凝集体は前述したＬＤと呼ばれる欠陥を含む。ＦＰＤとは、インゴットをスライスして作製されたシリコンウェーハを３０分間セコエッチング（Secco etching、ＨＦ：Ｋ₂Ｃｒ₂Ｏ₇(0.15mol/l)＝２：１の混合液によるエッチング）したときに現れる特異なフローパターンを呈する痕跡の源であり、ＬＳＴＤとは、シリコン単結晶内に赤外線を照射したときにシリコンとは異なる屈折率を有し散乱光を発生する源である。
【００１６】
ボロンコフの理論は、欠陥の数が少ない高純度インゴット１６を成長させるために、インゴットの引上げ速度をＶ（ｍｍ／分）、インゴットとシリコン融液１３の界面近傍のインゴット中の温度勾配をＧ（℃／ｍｍ）とするときに、Ｖ／Ｇ（ｍｍ²／分・℃）を制御することである。この理論では、図２に示すように、Ｖ／Ｇを横軸にとり、空孔型点欠陥濃度と格子間シリコン型点欠陥濃度を同一の縦軸にとって、Ｖ／Ｇと点欠陥濃度との関係を図式的に表現し、空孔領域と格子間シリコン領域の境界がＶ／Ｇによって決定されることを説明している。より詳しくは、Ｖ／Ｇ比が臨界点以上では空孔型点欠陥濃度が優勢なインゴットが形成される反面、Ｖ／Ｇ比が臨界点以下では格子間シリコン型点欠陥濃度が優勢なインゴットが形成される。図２において、［Ｉ］は格子間シリコン型点欠陥が支配的であって、格子間シリコン型点欠陥の凝集体が存在する領域（(Ｖ／Ｇ)₁以下）を示し、［Ｖ］はインゴット内での空孔型点欠陥が支配的であって、空孔型点欠陥の凝集体が存在する領域（(Ｖ／Ｇ)₂以上）を示し、［Ｐ］は空孔型点欠陥の凝集体及び格子間シリコン型点欠陥の凝集体が存在しないパーフェクト領域（(Ｖ／Ｇ)₁〜(Ｖ／Ｇ)₂）を示す。領域［Ｐ］に隣接する領域［Ｖ］にはＯＳＦ核を形成する領域［ＯＳＦ］（(Ｖ／Ｇ)₂〜(Ｖ／Ｇ)₃）が存在する。
【００１７】
このパーフェクト領域［Ｐ］は更に領域［Ｐ_I］と領域［Ｐ_V］に分類される。［Ｐ_I］はＶ／Ｇ比が上記(Ｖ／Ｇ)₁から臨界点までの領域であり、［Ｐ_V］はＶ／Ｇ比が臨界点から上記(Ｖ／Ｇ)₂までの領域である。即ち、［Ｐ_I］は領域［Ｉ］に隣接し、かつ侵入型転位を形成し得る最低の格子間シリコン型点欠陥濃度未満の格子間シリコン型点欠陥濃度を有する領域であり、［Ｐ_V］は領域［Ｖ］に隣接し、かつＯＳＦを形成し得る最低の空孔型点欠陥濃度未満の空孔型点欠陥濃度を有する領域である。なお、上記ＯＳＦは、結晶成長時にその核となる微小欠陥が導入され、半導体デバイスを製造する際の熱酸化工程等で顕在化し、作製したデバイスのリーク電流の増加等の不良原因になる。
【００１８】
図１に戻って、シリコン融液１３とインゴット１６との固液界面１９形状が上側に凸となるように、カスプ磁場１７の中立面１７ａの位置と、カスプ磁場１７の強度と、石英るつぼ１４の回転速度と、インゴット１６の回転速度とを制御する。具体的には、カスプ磁場１７の中立面１７ａとシリコン融液１３の表面との距離をＨとするとき、２８０ｍｍ≦Ｈ≦７００ｍｍを満たすように、上記中立面１７ａをシリコン融液１３の表面の下方に制御するか、或いは１４０ｍｍ≦Ｈ≦６００ｍｍを満たすように、上記中立面１７ａをシリコン融液１３の表面の上方に制御する。
【００１９】
また石英るつぼ１４の直径が大きくなるに従ってカスプ磁場１７の強度が強くなるように、第１及び第２コイル１１，１２に流す電流を制御する。このように第１及び第２コイルの電流を制御するのは、固液界面１９が上側に凸となるようにシリコン融液１３に対流を発生させるローレンツ力を、石英るつぼ１４の直径が大きくなるに従って大きくする必要があるためである。例えば、直径が２００ｍｍのインゴット１６を引上げるために、内径が６００ｍｍの石英るつぼ１４を用いた場合には、カスプ磁場１７の強度を０．０１８テスラに制御し、直径が３００ｍｍのインゴット１６を引上げるために、内径が６５０ｍｍの石英るつぼ１４を用いた場合には、カスプ磁場１７の強度を０．０２テスラに制御する。
【００２１】
上述のように、カスプ磁場１７の中立面１７ａの位置と、カスプ磁場１７の強度と、石英るつぼ１４の回転速度と、インゴット１６の回転速度とを制御しながら、シリコン単結晶のインゴット１６を引上げると、石英るつぼ１４の底部中央から固液界面１９の中央に向って上昇した後に、固液界面の外周縁近傍から石英るつぼ１４の底部中央に流下する第１対流２１が発生し、石英るつぼ１４の底部外周縁から周縁に沿って上昇した後に、上記第１対流２１に沿って流下する第２対流２２が発生し、更に第１対流２１の外方かつ第２対流２２の上方であってシリコン融液１３の表面近傍を循環する第３対流２３が発生する。上記第１対流２１は固液界面１９を押上げるので、固液界面形状は上側に凸となる。
【００２２】
この結果、固液界面１９の中心がシリコン融液１３表面の延長面上より上方に位置するため、固液界面１９の中心における鉛直方向の温度勾配が大きくなり、固液界面の中心における鉛直方向の温度勾配と、固液界面の周縁における鉛直方向の温度勾配との差が小さくなる。従って、略全長にわたって無欠陥で高品質のシリコン単結晶のインゴット１６を比較的容易に製造できる。またカスプ磁場１７の強度は極めて低いので、少ない消費電力でインゴット１６を引上げることができるとともに、第１及び第２コイル１１，１２を小型化できるので、比較的狭い空間でインゴット１６を引上げることができる。なお、第２対流２２は、場合によっては存在しないこともあり、第３対流２３は、シリコン融液１３から引上げられたインゴット１６内の酸素濃度に対して影響する可能性がある。
【００２３】
【実施例】
次に本発明の実施例を詳しく説明する。
＜実施例１＞
図１に示すように、シリコン単結晶のインゴット１６を引上げるときのシリコン融液１３に発生する対流２１〜２３を、カスプ磁場１７の中立面１７ａの位置を変えて数値解析により求めた後に、固液界面１９の周縁及び中心における鉛直方向の温度勾配の差の変化（図３）と、固液界面の周縁に対する中心の鉛直方向への距離（図４）とを算出した。このときの具体的な条件は次の通りである。石英るつぼ１４の内径は６５０ｍｍであり、石英るつぼの外周面をヒータ１８により包囲した。またコイル直径が１４５０ｍｍの第１及び第２コイル１１，１２を、石英るつぼ１４の回転軸をコイル中心としかつ鉛直方向に４００ｍｍあけて配設した。これらのコイル１１，１２の互いに逆向きの電流を流すことにより、各コイル中心から第１及び第２コイルの中立面を通るカスプ磁場１７を発生させた。このとき中立面１７ａと石英るつぼ１４の内周面の延長面との交線上におけるカスプ磁場１７の水平方向の強度が０．０２テスラとなるように各コイルに流す電流を制御した。この状態で石英るつぼ１４を＋４ｒｐｍの回転速度で回転させ、かつインゴット１６を−１２〜−８ｒｐｍの回転速度で回転させながら、直径３００ｍｍのインゴットを０．２〜０．５ｍｍ／分の引上げ速度で上記シリコン融液１３から引上げた。
【００２４】
その結果、図３から明らかなように、カスプ磁場の中立面がシリコン融液表面から下方に２８０〜７００ｍｍの範囲と、シリコン融液表面から上方に１４０〜６００ｍｍの範囲で、固液界面の周縁と中心で鉛直方向の温度勾配の差が殆ど無くなった。また図４から明らかなように、カスプ磁場の中立面がシリコン融液表面から下方に２８０〜７００ｍｍの範囲と、シリコン融液表面から上方に１４０〜６００ｍｍの範囲で、固液界面形状が上側に凸となった。この結果、カスプ磁場の中立面の位置が固液界面形状に大きな影響を与えることが判った。
【００２５】
＜実施例２＞
図１に示すように、シリコン単結晶のインゴット１６を引上げるときのシリコン融液１３に発生する対流２１〜２３を、石英るつぼ１４の回転速度を変えて数値解析により求めた後に、固液界面１９の周縁及び中心における鉛直方向の温度勾配の差の変化（図５）と、固液界面の周縁に対する中心の鉛直方向への距離（図６）とを算出した。このときの具体的な条件は次の通りである。石英るつぼ１４の内径は６００ｍｍであり、石英るつぼの外周面をヒータ１８により包囲した。またコイル直径が１４５０ｍｍの第１及び第２コイル１１，１２を、石英るつぼ１４の回転軸をコイル中心としかつ鉛直方向に４００ｍｍあけて配設した。これらのコイル１１，１２の互いに逆向きの電流を流すことにより、各コイル中心から第１及び第２コイルの中立面を通るカスプ磁場１７を発生させた。このとき上記中立面の位置はシリコン融液１３表面から下方に５００ｍｍとし、中立面１７ａと石英るつぼ１４の内周面の延長面との交線上におけるカスプ磁場１７の水平方向の強度が０．０１８テスラとなるように各コイルに流す電流を制御した。この状態で直径２００ｍｍのインゴット１６を０．３〜０．７ｍｍ／分の引上げ速度で上記シリコン融液１３から引上げた。なお、インゴット１６の回転速度Ｒ₂を−２０〜−１６ｒｐｍに設定し、石英るつぼ１４の回転速度Ｒ₁を−１〜＋７ｒｐｍの範囲内の１３種類の速度に変えて上記対流などをそれぞれ算出した。
【００２６】
その結果、図５から明らかなように、石英るつぼの回転速度が３〜３．５ｒｐｍの範囲で、固液界面の周縁と中心で鉛直方向の温度勾配の差が最も小さくなった。また図６からも明らかなように、石英るつぼの回転速度が３〜３．５ｒｐｍの範囲で、固液界面形状が上側に最も突出する凸形状となった。
【００２７】
【発明の効果】
以上述べたように、本発明によれば、シリコン融液とインゴットとの固液界面形状が上側に凸となるように、カスプ磁場の中立面の位置と、カスプ磁場の強度と、石英るつぼの回転速度と、インゴットの回転速度とを制御し、シリコン単結晶のインゴット内がパーフェクト領域となるような引上げ速度でインゴットを引上げるので、シリコン融液に所定の対流が発生し、これらの対流により固液界面形状が上側に凸となる。この結果、固液界面の中心がシリコン融液表面の延長面上より上方に位置するという理由から、固液界面の中心における鉛直方向の温度勾配が大きくなるので、この温度勾配と、固液界面の周縁における鉛直方向の温度勾配との差が小さくなる。従って、略全長にわたって無欠陥で高品質のシリコン単結晶のインゴットを比較的容易に製造できる。またカスプ磁場の強度は極めて低いので、少ない消費電力でインゴットを引上げることができるとともに、第１及び第２コイルを小型化できるので、比較的狭い空間でインゴットを引上げることができる。
【図面の簡単な説明】
【図１】本発明実施形態のシリコン単結晶のインゴットを引上げている状態を示す断面構成図。
【図２】ボロンコフの理論を基づいた、Ｖ／Ｇ比が臨界点以上では空孔型点欠陥濃度が優勢なインゴットが形成され、Ｖ／Ｇ比が臨界点以下では格子間シリコン型点欠陥濃度が優勢なインゴットが形成されることを示す図。
【図３】カスプ磁場の中立面とシリコン融液の表面との距離を変化させたときの、固液界面の周縁及び中心における鉛直方向の温度勾配の差の変化を示す図。
【図４】カスプ磁場の中立面とシリコン融液の表面との距離を変化させたときの、固液界面の周縁に対する中心の鉛直方向への距離の変化を示す図。
【図５】石英るつぼの回転速度を変化させたときの、固液界面の周縁及び中心における鉛直方向の温度勾配の差の変化を示す図。
【図６】石英るつぼの回転速度を変化させたときの、固液界面の周縁及び中心における鉛直方向の温度勾配の差の変化を示す図。
【符号の説明】
１１第１コイル
１２第２コイル
１３シリコン融液
１４石英るつぼ
１６シリコン単結晶のインゴット
１７カスプ磁場
１７ａカスプ磁場の中立面
１９固液界面[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method for pulling a silicon single crystal ingot from a silicon melt while applying a cusp magnetic field to the silicon melt.
[0002]
[Prior art]
Conventionally, as a method for producing a silicon single crystal, a method of pulling an ingot of a silicon single crystal by the Czochralski method (hereinafter referred to as CZ method) is known. This CZ method is a method of manufacturing a cylindrical silicon single crystal ingot by bringing a seed crystal into contact with a silicon melt stored in a quartz crucible and pulling the seed crystal while rotating the quartz crucible and the seed crystal. It is.
[0003]
On the other hand, in the process of manufacturing a semiconductor integrated circuit, as a cause of lowering the yield, micro defects of oxygen precipitates that are the core of an oxidation-induced stacking fault (hereinafter referred to as OSF) and particles caused by crystals (Crystal Originated Particles, hereinafter referred to as COP) or the presence of interstitial-type large dislocation (hereinafter referred to as LD). OSF is introduced with a micro defect that becomes a nucleus during crystal growth, and becomes apparent in a thermal oxidation process or the like when manufacturing a semiconductor device, and causes a defect such as an increase in leakage current of the manufactured device. COPs are pits caused by crystals that appear on the wafer surface when the mirror-polished silicon wafer is washed with a mixture of ammonia and hydrogen peroxide. When this wafer is measured with a particle counter, this pit is also detected as a light scattering defect together with the original particles.
[0004]
This COP causes deterioration of electrical characteristics, for example, dielectric breakdown characteristics (Time Dependent dielectric Breakdown, TDDB) of oxide films, oxide breakdown voltage characteristics (Time Zero Dielectric Breakdown, TZDB), and the like. Further, if COP exists on the wafer surface, a step is generated in the device wiring process, which may cause disconnection. In addition, the element isolation portion also causes leakage and the like, thereby reducing the product yield. Furthermore, LD is also called a dislocation cluster, or a pit is formed when a silicon wafer having such a defect is immersed in a selective etching solution containing hydrofluoric acid as a main component. This LD also causes deterioration of electrical characteristics such as leakage characteristics and isolation characteristics. As a result, it is necessary to reduce OSF, COP, and LD from a silicon wafer used for manufacturing a semiconductor integrated circuit.
[0005]
Japanese Patent Application Laid-Open No. 11-1393 discloses a method of manufacturing a silicon single crystal ingot for cutting out a defect-free silicon wafer having no OSF, COP, and LD. Generally, when a silicon single crystal ingot is pulled up at a high speed, a region [V] in which agglomerates of vacancy-type point defects exist predominantly is formed inside the ingot, and when the ingot is pulled up at a low speed, the ingot A region [I] in which agglomerates of interstitial silicon type point defects exist predominantly is formed inside. For this reason, in the manufacturing method described above, a silicon single crystal consisting of a perfect region [P] in which no agglomerates of point defects are present can be manufactured by pulling up the ingot at an optimal pulling rate.
[0006]
[Problems to be solved by the invention]
However, in the conventional method for producing a silicon single crystal ingot disclosed in Japanese Patent Application Laid-Open No. 11-1393, the vertical temperature gradient in the vicinity of the solid-liquid interface between the silicon single crystal ingot and the silicon melt is uniform. It is difficult to produce a defect-free silicon single crystal over the entire length of the straight body of the ingot because this control is affected by changes in the remaining amount of silicon melt and changes in convection. Met.
An object of the present invention is to provide a silicon single crystal pulling method capable of relatively easily producing an ingot of a silicon single crystal having a relatively low power and a relatively narrow space, defect-free and controlled oxygen concentration. It is to provide.
[0007]
[Means for Solving the Problems]
In the invention according to claim 1, as shown in FIG. 1, the quartz crucible 14 storing the silicon melt 13 is rotated at a predetermined rotational speed, and the silicon single crystal ingot 16 pulled up from the silicon melt 13 is The first and second coils 11 and 12 having a coil diameter larger than the outer diameter of the quartz crucible 14 are rotated at a rotational speed, and the rotation axis of the quartz crucible 14 is set as the coil center and spaced apart at a predetermined interval in the vertical direction. The cusp magnetic field 17 passing through the neutral surface 17a between the first and second coils is generated from the center of each coil of the first and second coils by flowing currents in opposite directions to the first and second coils. Silico pulling up the ingot 16 at a pulling speed at which the inside of the ingot 16 becomes a perfect region in which no interstitial silicon type point defect aggregates and no vacancy type point defect aggregates exist. It is an improvement of the pulling process of the single crystal.
The characteristic configuration is that when the distance between the neutral surface 17a of the cusp magnetic field 17 and the surface of the silicon melt 13 is H, the neutral surface 17a is made of the silicon melt 13 so as to satisfy 280 mm ≦ H ≦ 700 mm. The neutral surface 17a is controlled above the surface of the silicon melt 13 so that 140 mm ≦ H ≦ 600 mm is satisfied, or the ingot 16 has a diameter of 300 mm and the quartz crucible 14 has an inner diameter. Is 650 mm, the horizontal strength of the cusp magnetic field 17 is controlled to 0.02 Tesla, the quartz crucible 14 is rotated at a rotational speed of 4 rpm, and the ingot 16 is rotated in the direction opposite to the quartz crucible 14 at 12 to 8 rpm. The ingot 16 is pulled up at a pulling speed of 0.2 to 0.5 mm / min while rotating at a rotating speed .
[0008]
In the pulling method of the silicon single crystal described in claim 1, the position of the neutral surface of the cusp magnetic field 17, the strength of the cusp magnetic field, the rotational speed of the quartz crucible 14, and the rotational speed of the ingot 16 are controlled. However, when the ingot is pulled up, predetermined convections 21 to 23 are generated in the silicon melt 13, and the shape of the solid-liquid interface 19 is convex upward by these convections 21 to 23. As a result, since the center of the solid-liquid interface is located above the extended surface of the surface of the silicon melt 13, the temperature gradient in the vertical direction at the center of the solid-liquid interface 19 increases, and the vertical direction at the center of the solid-liquid interface increases. The difference between the temperature gradient and the temperature gradient in the vertical direction at the periphery of the solid-liquid interface is reduced. Therefore, a high-quality silicon single crystal ingot 16 having substantially no defect can be manufactured relatively easily.
[0009]
In the invention according to claim 2, as shown in FIG. 1, the quartz crucible 14 storing the silicon melt 13 is rotated at a predetermined rotational speed, and the silicon single crystal ingot 16 pulled up from the silicon melt 13 is The first and second coils 11 and 12 having a coil diameter larger than the outer diameter of the quartz crucible 14 are rotated at a rotational speed, and the rotation axis of the quartz crucible 14 is set as the coil center and spaced apart at a predetermined interval in the vertical direction. The cusp magnetic field 17 passing through the neutral surface 17a between the first and second coils is generated from the center of each coil of the first and second coils by flowing currents in opposite directions to the first and second coils. Silico pulling up the ingot 16 at a pulling speed at which the inside of the ingot 16 becomes a perfect region in which no interstitial silicon type point defect aggregates and no vacancy type point defect aggregates exist. It is an improvement of the pulling process of the single crystal.
The characteristic configuration is that when the distance between the neutral surface 17a of the cusp magnetic field 17 and the surface of the silicon melt 13 is H, the neutral surface 17a is placed on the surface of the silicon melt 13 so as to satisfy H = 500 mm. When the diameter of the ingot 16 is 200 mm and the inner diameter of the quartz crucible 14 is 600 mm, the horizontal strength of the cusp magnetic field 17 is controlled to 0.018 Tesla, and the quartz crucible 14 is moved to 3 to 3. The ingot 16 is pulled at a pulling speed of 0.3 to 0.7 mm / min while rotating at a rotating speed of 5 rpm and rotating the ingot 16 at a rotating speed of 20 to 16 rpm in the opposite direction to the quartz crucible 14. .
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Next, embodiments of the present invention will be described with reference to the drawings.
As shown in FIG. 1, the silicon single crystal pulling method of the present invention is such that a quartz crucible 14 storing a silicon melt 13 is rotated at a predetermined rotation speed R ₁ and the silicon single crystal pulled from the silicon melt 13 is pulled. The ingot 16 is pulled from the silicon melt 13 while rotating the ingot 16 at a predetermined rotational speed R ₂ and applying the cusp magnetic field 17 to the silicon melt 13 using the first and second coils 11 and 12. It is a way to raise. The first and second coils 11, 12 have a coil diameter larger than the outer diameter of the quartz crucible 14, and are arranged at predetermined intervals in the vertical direction with the rotation axis of the quartz crucible as the coil center. . Also, currents flowing in opposite directions are passed through the first and second coils 11 and 12, whereby a cusp magnetic field passes through the neutral surface 17a between the first and second coils from the center of each of the first and second coils. 17 is generated. The neutral surface 17a is a horizontal surface between the first and second coils 11 and 12 where the vertical magnetic field strength is zero. Further, reference numeral 18 in FIG. 1 denotes a heater that surrounds the outer peripheral surface of the quartz crucible 14.
[0012]
On the other hand, the ingot 16 is pulled up at a pulling speed at which the inside of the ingot becomes a perfect region in which no interstitial silicon type point defect aggregates and no vacancy type point defect aggregates exist. That is, the ingot is pulled up from the silicon melt 13 in the hot zone furnace with a predetermined pulling speed profile based on the Boronkov theory by the CZ method.
[0013]
Generally, when a silicon single crystal ingot 16 is pulled up from a silicon melt 13 in a hot zone furnace by the CZ method, point defects and agglomerates (agglomerates: three-dimensional) are formed in the ingot. Defect) occurs. There are two general forms of point defects: vacancy-type point defects and interstitial silicon-type point defects. A vacancy-type point defect is one in which one silicon atom leaves one of the normal positions in the silicon crystal lattice. Such holes become hole-type point defects. On the other hand, when an atom is found at a position (interstitial site) other than the lattice point of the silicon crystal, this becomes an interstitial silicon point defect.
[0014]
Point defects are generally formed at the contact surface between the silicon melt 13 and the ingot 16, that is, at the solid-liquid interface 19. However, by continuously pulling up the ingot 16, the portion that was the solid-liquid interface 19 begins to cool as it is pulled up. During cooling, vacancy point defects or interstitial silicon point defects merge with each other by diffusion to form vacancy agglomerates or interstitial agglomerates. Is formed. In other words, the aggregate has a three-dimensional structure generated due to the merge of point defects.
[0015]
The agglomerates of vacancy-type point defects include defects called LSTD (Laser Scattering Tomograph Defects) or FPD (Flow Pattern Defects) in addition to the above-mentioned COP. Includes defects called. FPD means when a silicon wafer produced by slicing an ingot is subjected to secco etching (Secco etching, etching with a mixed solution of HF: K ₂ Cr ₂ O ₇ (0.15 mol / l) = 2: 1). LSTD is a source that generates a scattered light having a refractive index different from that of silicon when an infrared ray is irradiated into a silicon single crystal.
[0016]
Boronkov's theory is that in order to grow a high-purity ingot 16 having a small number of defects, the pulling rate of the ingot is V (mm / min), and the temperature gradient in the ingot near the interface between the ingot and the silicon melt 13 is G ( V / G (mm ² / min · ° C.) when controlling the temperature. In this theory, as shown in FIG. 2, the relationship between V / G and point defect concentration is shown with V / G on the horizontal axis and vacancy-type point defect concentration and interstitial silicon type point defect concentration on the same vertical axis. Is described schematically, and it is explained that the boundary between the void region and the interstitial silicon region is determined by V / G. More specifically, when the V / G ratio is equal to or higher than the critical point, an ingot having a dominant vacancy-type point defect concentration is formed. On the other hand, when the V / G ratio is lower than the critical point, an ingot having a dominant interstitial silicon-type point defect concentration is formed. It is formed. In FIG. 2, [I] indicates a region where an interstitial silicon type point defect is dominant and an aggregate of interstitial silicon type point defects exists ((V / G) ₁ or less), and [V] indicates The vacancy-type point defect in the ingot is dominant, and indicates a region where an aggregate of the vacancy-type point defect exists ((V / G) ₂ or more), and [P] indicates the vacancy-type point defect. A perfect region ((V / G) ₁ to (V / G) ₂ ) where no aggregates and aggregates of interstitial silicon type point defects exist is shown. A region [OSF] ((V / G) ₂ to (V / G) ₃ ) that forms an OSF nucleus exists in the region [V] adjacent to the region [P].
[0017]
The perfect region [P] is further classified into a region [P _I ] and a region [P _V ]. [P _I ] is a region where the V / G ratio is from the above (V / G) ₁ to the critical point, and [P _V ] is a region where the V / G ratio is from the critical point to the above (V / G) _2. is there. That is, [P _I ] is a region adjacent to the region [I] and having an interstitial silicon type point defect concentration lower than the lowest interstitial silicon type point defect concentration capable of forming interstitial dislocations, and [P _V]. ] Is a region adjacent to the region [V] and having a vacancy-type point defect concentration lower than the lowest vacancy-type point defect concentration capable of forming an OSF. The OSF is introduced with a micro defect serving as a nucleus during crystal growth, and is manifested in a thermal oxidation process or the like when manufacturing a semiconductor device, and causes a defect such as an increase in leakage current of the manufactured device.
[0018]
Returning to FIG. 1, the position of the neutral surface 17a of the cusp magnetic field 17, the strength of the cusp magnetic field 17, and the quartz crucible so that the shape of the solid-liquid interface 19 between the silicon melt 13 and the ingot 16 is convex upward. 14 and the rotation speed of the ingot 16 are controlled. Specifically, when the distance between the neutral surface 17a of the cusp magnetic field 17 and the surface of the silicon melt 13 is H, the neutral surface 17a is formed so as to satisfy 280 mm ≦ H ≦ 700 mm. or controlled under side of the surface of the liquid 13, or to meet the 140 mm ≦ H ≦ 600 mm, for controlling the neutral plane 17a above the surface of the silicon melt 13.
[0019]
Further, the currents flowing through the first and second coils 11 and 12 are controlled so that the strength of the cusp magnetic field 17 increases as the diameter of the quartz crucible 14 increases. In this way, the currents in the first and second coils are controlled by increasing the diameter of the quartz crucible 14 using the Lorentz force that generates convection in the silicon melt 13 so that the solid-liquid interface 19 protrudes upward. This is because it is necessary to make it larger. For example, when a quartz crucible 14 having an inner diameter of 600 mm is used to pull up the ingot 16 having a diameter of 200 mm, the strength of the cusp magnetic field 17 is controlled to 0.018 Tesla, and the ingot 16 having a diameter of 300 mm is adjusted. When the quartz crucible 14 having an inner diameter of 650 mm is used for pulling up, the strength of the cusp magnetic field 17 is controlled to 0.02 Tesla .
[0021]
As described above, while controlling the position of the neutral surface 17 a of the cusp magnetic field 17, the strength of the cusp magnetic field 17, the rotational speed of the quartz crucible 14, and the rotational speed of the ingot 16, the silicon single crystal ingot 16 is When pulled up, a first convection 21 is generated that rises from the center of the bottom of the quartz crucible 14 toward the center of the solid-liquid interface 19 and then flows down from the vicinity of the outer peripheral edge of the solid-liquid interface to the center of the bottom of the quartz crucible 14. After rising from the outer peripheral edge of the bottom of the crucible 14 along the peripheral edge, a second convection 22 flowing down along the first convection 21 is generated, and further outside the first convection 21 and above the second convection 22. Thus, the third convection 23 circulating around the surface of the silicon melt 13 is generated. Since the first convection 21 pushes up the solid-liquid interface 19, the shape of the solid-liquid interface becomes convex upward.
[0022]
As a result, since the center of the solid-liquid interface 19 is located above the extended surface of the surface of the silicon melt 13, the temperature gradient in the vertical direction at the center of the solid-liquid interface 19 becomes large, and the vertical direction at the center of the solid-liquid interface. And the difference between the temperature gradient in the vertical direction at the periphery of the solid-liquid interface is reduced. Therefore, a high-quality silicon single crystal ingot 16 having substantially no defect can be manufactured relatively easily. Further, since the strength of the cusp magnetic field 17 is extremely low, the ingot 16 can be pulled up with low power consumption, and the first and second coils 11 and 12 can be miniaturized, so that the ingot 16 can be pulled up in a relatively narrow space. be able to. Note that the second convection 22 may not exist in some cases, and the third convection 23 may affect the oxygen concentration in the ingot 16 pulled up from the silicon melt 13.
[0023]
【Example】
Next, embodiments of the present invention will be described in detail.
<Example 1>
As shown in FIG. 1, the convections 21 to 23 generated in the silicon melt 13 when pulling up the silicon single crystal ingot 16 are obtained by numerical analysis by changing the position of the neutral surface 17a of the cusp magnetic field 17. The change in the difference in the temperature gradient in the vertical direction at the periphery and center of the solid-liquid interface 19 (FIG. 3) and the distance in the vertical direction of the center with respect to the periphery of the solid-liquid interface (FIG. 4) were calculated. Specific conditions at this time are as follows. The inner diameter of the quartz crucible 14 was 650 mm, and the outer peripheral surface of the quartz crucible was surrounded by the heater 18. Further, the first and second coils 11 and 12 having a coil diameter of 1450 mm were arranged with the rotation axis of the quartz crucible 14 as the coil center and 400 mm apart in the vertical direction. By flowing currents in the coils 11 and 12 in opposite directions, a cusp magnetic field 17 passing from the center of each coil through the neutral surfaces of the first and second coils was generated. At this time, the current passed through each coil was controlled so that the horizontal strength of the cusp magnetic field 17 on the intersection line between the neutral surface 17a and the extended surface of the inner peripheral surface of the quartz crucible 14 was 0.02 Tesla . In this state, the quartz crucible 14 is rotated at a rotational speed of +4 rpm, and the ingot 16 is rotated at a rotational speed of −12 to −8 rpm, while an ingot having a diameter of 300 mm is pulled up at a speed of 0.2 to 0.5 mm / min. The silicon melt 13 was pulled up.
[0024]
As a result, as is clear from FIG. 3, the neutral surface of the cusp magnetic field is in the range of 280 to 700 mm downward from the silicon melt surface and in the range of 140 to 600 mm upward from the silicon melt surface. There was almost no difference in temperature gradient in the vertical direction between the periphery and the center. As is clear from FIG. 4, the neutral surface of the cusp magnetic field is in the range of 280 to 700 mm downward from the silicon melt surface and in the range of 140 to 600 mm upward from the silicon melt surface, and the solid-liquid interface shape is on the upper side. It became convex. As a result, it was found that the position of the neutral surface of the cusp magnetic field greatly affects the solid-liquid interface shape.
[0025]
<Example 2>
As shown in FIG. 1, the convections 21 to 23 generated in the silicon melt 13 when the silicon single crystal ingot 16 is pulled up are obtained by numerical analysis by changing the rotation speed of the quartz crucible 14, and then the solid-liquid interface. A change in the difference in temperature gradient in the vertical direction at the periphery and center of 19 (FIG. 5) and the distance in the center in the vertical direction with respect to the periphery of the solid-liquid interface (FIG. 6) were calculated. Specific conditions at this time are as follows. The inner diameter of the quartz crucible 14 was 600 mm, and the outer peripheral surface of the quartz crucible was surrounded by the heater 18. Further, the first and second coils 11 and 12 having a coil diameter of 1450 mm were arranged with the rotation axis of the quartz crucible 14 as the coil center and 400 mm apart in the vertical direction. By flowing currents in the coils 11 and 12 in opposite directions, a cusp magnetic field 17 passing from the center of each coil through the neutral surfaces of the first and second coils was generated. At this time, the position of the neutral surface is 500 mm downward from the surface of the silicon melt 13, and the horizontal strength of the cusp magnetic field 17 on the intersection line between the neutral surface 17 a and the extended surface of the inner peripheral surface of the quartz crucible 14 is 0. The current flowing through each coil was controlled to be .018 Tesla . In this state, the ingot 16 having a diameter of 200 mm was pulled up from the silicon melt 13 at a pulling rate of 0.3 to 0.7 mm / min. Incidentally, the rotation speed R ₂ of the ingot 16 is set to -20 to-16 rpm, was calculated the convection etc., respectively by changing the rotational speed R ₁ of the quartz crucible 14 to 13 different speeds in the range of -1 to + 7 rpm .
[0026]
As a result, as apparent from FIG. 5, the difference in the temperature gradient in the vertical direction was the smallest between the periphery and the center of the solid-liquid interface when the rotation speed of the quartz crucible was in the range of 3 to 3.5 rpm. Further, as is apparent from FIG. 6, the solid-liquid interface shape is a convex shape that protrudes most upward when the rotation speed of the quartz crucible is in the range of 3 to 3.5 rpm.
[0027]
【The invention's effect】
As described above, according to the present invention, the position of the neutral surface of the cusp magnetic field, the strength of the cusp magnetic field, and the quartz crucible so that the solid-liquid interface shape between the silicon melt and the ingot is convex upward. The rotation speed of the ingot and the rotation speed of the ingot are controlled, and the ingot is pulled up at such a pulling speed that the inside of the silicon single crystal ingot becomes a perfect region, so that a predetermined convection occurs in the silicon melt, and these convection flows. As a result, the solid-liquid interface shape is convex upward. As a result, the temperature gradient in the vertical direction at the center of the solid-liquid interface increases because the center of the solid-liquid interface is located above the extended surface of the silicon melt surface. The difference with the temperature gradient in the vertical direction at the periphery of the is small. Therefore, a high-quality silicon single crystal ingot having almost no defect can be manufactured relatively easily. Further, since the strength of the cusp magnetic field is extremely low, the ingot can be pulled up with low power consumption, and the first and second coils can be miniaturized, so that the ingot can be pulled up in a relatively narrow space.
[Brief description of the drawings]
FIG. 1 is a cross-sectional configuration diagram illustrating a state where an ingot of a silicon single crystal according to an embodiment of the present invention is pulled up.
FIG. 2 shows that an ingot having a dominant vacancy-type point defect concentration is formed when the V / G ratio is equal to or higher than the critical point, and the interstitial silicon type point defect concentration is lower than the critical point based on the Boronkov theory. The figure which shows that an ingot where is dominant is formed.
FIG. 3 is a view showing a change in the difference in temperature gradient in the vertical direction at the periphery and center of the solid-liquid interface when the distance between the neutral surface of the cusp magnetic field and the surface of the silicon melt is changed.
FIG. 4 is a view showing a change in the distance in the vertical direction of the center with respect to the periphery of the solid-liquid interface when the distance between the neutral surface of the cusp magnetic field and the surface of the silicon melt is changed.
FIG. 5 is a diagram showing a change in the difference in temperature gradient in the vertical direction at the periphery and center of the solid-liquid interface when the rotation speed of the quartz crucible is changed.
FIG. 6 is a diagram showing a change in the difference in temperature gradient in the vertical direction at the periphery and center of the solid-liquid interface when the rotation speed of the quartz crucible is changed.
[Explanation of symbols]
11 First coil 12 Second coil 13 Silicon melt 14 Quartz crucible 16 Silicon single crystal ingot 17 Cusp magnetic field 17a Neutral surface 19 of cusp magnetic field Solid-liquid interface

Claims

A quartz crucible (14) for storing the silicon melt (13) is rotated at a predetermined rotation speed, and a silicon single crystal ingot (16) pulled from the silicon melt (13) is rotated at a predetermined rotation speed. The first and second coils (11, 12) having a coil diameter larger than the outer diameter of the quartz crucible (14) are centered on the rotation axis of the quartz crucible (14) and spaced apart in the vertical direction. The neutral current between the first and second coils from the center of each of the first and second coils by flowing currents in opposite directions to the first and second coils (11, 12). A cusp magnetic field (17) passing through the surface (17a) is generated, and the inside of the ingot (16) has a pulling rate that becomes a perfect region in which no interstitial silicon type point defect aggregates and no void type point defect aggregates exist. In the method of pulling up a silicon single crystal that pulls up the ingot ,
When the distance between the neutral surface (17a ) of the cusp magnetic field (17) and the surface of the silicon melt (13) is H, the neutral surface (17a) is set so as to satisfy 280 mm ≦ H ≦ 700 mm. Control below the surface of the silicon melt (13) , or control the neutral surface (17a) above the surface of the silicon melt (13) so as to satisfy 140 mm ≦ H ≦ 600 mm ,
When the diameter of the ingot (16) is 300 mm and the inner diameter of the quartz crucible (14) is 650 mm, the horizontal strength of the cusp magnetic field (17) is controlled to 0.02 Tesla;
The quartz crucible the quartz crucible (14) is rotated at the rotational speed of 4rpm and the ingot (16) (14) wherein the ingot while rotating at a rotational speed of 12~8rpm in a direction opposite to the (16) 0. A method for pulling a silicon single crystal, which is pulled at a pulling rate of 2 to 0.5 mm / min .

A quartz crucible (14) for storing the silicon melt (13) is rotated at a predetermined rotation speed, and a silicon single crystal ingot (16) pulled from the silicon melt (13) is rotated at a predetermined rotation speed. The first and second coils (11, 12) having a coil diameter larger than the outer diameter of the quartz crucible (14) are centered on the rotation axis of the quartz crucible (14) and spaced apart in the vertical direction. The neutral current between the first and second coils from the center of each of the first and second coils by flowing currents in opposite directions to the first and second coils (11, 12). A cusp magnetic field (17) passing through the surface (17a) is generated, and the inside of the ingot (16) has a pulling rate that becomes a perfect region in which no interstitial silicon type point defect aggregates and no void type point defect aggregates exist. In the method of pulling up a silicon single crystal that pulls up the ingot ,
When the distance between the neutral surface (17a ) of the cusp magnetic field (17) and the surface of the silicon melt (13) is H, the neutral surface (17a) is placed on the silicon so as to satisfy H = 500 mm. Control below the surface of the melt (13) ,
When the diameter of the ingot (16) is 200 mm and the inner diameter of the quartz crucible (14) is 600 mm, the horizontal strength of the cusp magnetic field (17) is controlled to 0.018 Tesla,
The quartz crucible (14) is rotated at a rotational speed of 3 to 3.5 rpm, and the ingot (16) is rotated while rotating the ingot (16) at a rotational speed of 20 to 16 rpm in the opposite direction to the quartz crucible (14). ) At a pulling rate of 0.3 to 0.7 mm / min .