JP2004128112A - Manufacturing method of semiconductor device - Google Patents

Manufacturing method of semiconductor device Download PDF

Info

Publication number
JP2004128112A
JP2004128112A JP2002288356A JP2002288356A JP2004128112A JP 2004128112 A JP2004128112 A JP 2004128112A JP 2002288356 A JP2002288356 A JP 2002288356A JP 2002288356 A JP2002288356 A JP 2002288356A JP 2004128112 A JP2004128112 A JP 2004128112A
Authority
JP
Japan
Prior art keywords
polishing
conductive film
polishing tool
copper
processing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
JP2002288356A
Other languages
Japanese (ja)
Inventor
Takatada Yamaguchi
山口 宇唯
▲片▼桐 創一
Souichi Katagiri
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Renesas Technology Corp
Original Assignee
Renesas Technology Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Renesas Technology Corp filed Critical Renesas Technology Corp
Priority to JP2002288356A priority Critical patent/JP2004128112A/en
Publication of JP2004128112A publication Critical patent/JP2004128112A/en
Pending legal-status Critical Current

Links

Images

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem of production yield degradation caused by dishing or eroding at flattening of a copper wiring, resulting in increase in variation of wiring resistances after flattening or in generation of shorting caused by the polishing residual of copper. <P>SOLUTION: With an oxidant, organic acid, anticorrosive, surfactants, polishing liquid of pure water, a polishing pad, a fixed abrasive table, and a dresser for dressing the surface of the polishing tool provided, the polishing liquid for fast polishing copper is supplied to the polishing pad for polishing in an initial process stage. In an intermediate process stage, a polishing liquid of high protrusion/recess selecting ratio is supplied to the fixed abrasive table for polishing to eliminate a residual step of the copper. In a later process stage directly below or after exposion of a barrier film, the polishing liquids are switched so that the copper and barrier films are polished almost at the same speed, and the dresser is driven for in-process conditioning. <P>COPYRIGHT: (C)2004,JPO

Description

【0001】
【発明の属する技術分野】
本発明は、基板表面を平坦化する研磨や研削技術に係り、特に半導体基板上に形成された薄膜が研磨や研削される半導体集積回路の製造方法に関する。
【0002】
【従来の技術】
近年、半導体集積回路装置(以下LSIと記す)の高集積化、高性能化に伴って新たな微細加工技術が開発されている。 化学機械研磨法(ChemicalMechanical Polishing;CMPと記す)もそのひとつであり、LSI製造工程、特に多層配線形成工程における層間絶縁膜の平坦化、金属プラグ形成、埋め込み配線形成において頻繁に利用される技術である。この平坦化技術は、例えば特許文献1に開示されている。
【0003】
また、最近はLSIの高速性能化を達成するために、配線材料を従来のアルミニウム合金から低電気抵抗の銅合金を利用しようとすることが試みられている。しかし、銅合金はアルミニウム合金配線の形成で頻繁に用いられたドライエッチング法による微細加工が困難である。そこで、加工して溝の形成された絶縁膜上に、絶縁膜中への銅拡散防止のためにバリア膜として、タンタルやタンタル合金及び窒化タンタルやその他のタンタル化合物等を堆積し、更にバリア膜の上に銅或いは銅合金薄膜を堆積し、絶縁膜に形成された溝内に埋め込まれた部分以外の銅或いは銅合金薄膜をCMPにより除去して埋め込み配線を形成する、いわゆるダマシン法が主に採用されている。このダマシン法による配線技術は、例えば特許文献2に開示されている。
【0004】
金属のCMPの一般的な方法は、円形の研磨定盤(プラテン)上に研磨パッドを貼り付け、研磨パッド表面を金属用研磨液で浸し、基体の金属膜を形成した面を押し付けて、基体の裏面から所定の圧力(以下研磨圧力と記す)を加えた状態で研磨定盤を回し、研磨液と金属膜の凸部との機械的摩擦によって凸部の金属膜を除去するものである。
【0005】
配線に用いられる銅合金等の金属のCMPに用いられる研磨剤は、固体砥粒と酸化性物質を主成分とするものが一般的である。酸化性物質の酸化作用で金属表面を酸化しながら、固体砥粒によってその酸化物を機械的に除去するのが基本的なCMPのメカニズムである。CMPのメカニズムに関しては、非特許文献1の第299頁に開示されている。固体砥粒としては、数10から数100nmの粒子径を持つアルミナ砥粒やシリカ砥粒が知られているが、一般に市販されている金属研磨用の固体砥粒のほとんどは前者である。酸化性物質としては、過酸化水素(H)、硝酸第二鉄(Fe(NO)、過ヨウ素酸カリウム(KIO)が一般に広く用いられており、これらは例えば、非特許文献1の第299頁から第300頁に開示されている。
また、固定砥粒を用いたSTI層平坦化に関しては、例えば、特許文献3に記載がある。
【0006】
なお、複数のポリシングパッドを用いる従来技術としては、特許文献4、特許文献5、特許文献6、特許文献7、特許文献8、特許文献9などに開示されている。
【0007】
【特許文献1】
米国特許No.4944836
【特許文献2】
特開平2−278822号公報
【特許文献3】
特開2000―49122号公報
【特許文献4】
特開2000−164595号公報
【特許文献5】
特開平9−326392号公報
【特許文献6】
特開2001−135605号公報
【特許文献7】
特開2001−170858号公報
【特許文献8】
特開2001−332517号公報
【特許文献9】
特表2002−512894号公報
【非特許文献1】
柏木正弘編集、「CMPのサイエンス」1997年8月20日(株式会社サイエンスフォーラム発行)、第299頁および300頁
【非特許文献2】
「ジャーナルオブ・エレクトロケミカル・ソサエテイ(J.Electrochem.Soc.)」第141巻第10号、1994年10月、p.2842−2848
【非特許文献3】
「次世代ULSI多層配線の新材料・プロセス技術」技術情報協会、p.242−246
【0008】
【発明が解決しようとする課題】
しかしながら、上記従来技術では研磨速度は高いものの、平坦性不足であったり、平坦性は高いものの研磨速度不足であるという、高研磨速度と高平坦性を両立できない問題があった。更に、堆積されたCu膜の凸部の表面と凹部の底面までの段差をCu残段差または残段差と呼ぶことにすると、により変化すること及び研磨液についての配慮がされておらず、以下に挙げる(1)平坦性不足、(2)コスト高という問題があった。
(1)研磨パッドが軟質(例えば、ポリウレタンからなる研磨パッドで、圧縮弾性率が100MPa以下)であることに起因して、基板表面を押しつける力により研磨パッドが凹凸に変形してしまい、絶縁膜に形成された溝の内部に埋め込まれる金属配線の表面中央部分が周辺部分よりも過剰に研磨されて凹む現象(以後ディッシングと記す)や配線部周囲の絶縁膜表面が過剰に研磨される現象(以後エロージョンと記す)が発生する問題があった(図11)。また、バリア膜として用いられるタンタルやタンタル合金及び窒化タンタルやその他のタンタル化合物は、化学的に安定でエッチングが難しく、硬度が高いために機械的な研磨も銅及び銅合金ほど容易ではない。そこで、砥粒の硬度を上げた場合には、銅或いは銅合金に研磨キズが発生して電気特性不良の原因になったり、砥粒の粒子濃度を高くした場合には、絶縁膜の研磨速度が大きくなり、エロージョンが発生しやすくなる問題があった。
【0009】
更に、配線パターンとなる溝が形成された絶縁膜上に金属膜が堆積されたウエハのCMPでは、局所的に過剰に研磨される箇所が発生することが知られている。これは、CMP前のウエハ表面には配線パターンとなる溝を反映した凹凸が金属膜の表面に生じており、CMPを行う場合にパターン密度に応じて局所的に高い圧力がかかり、その部分の研磨速度が速くなるためである。従って、ディッシングやエロージョンは金属部分の面積が広いパッド(0.1mm角程度の面積)や密集配線パターンで顕著な問題となる。これらは、ジャーナルオブ・エレクトロケミカル・ソサエテイ第141巻第10号、1994年10月(非特許文献2参照)に記載されている。配線表面が凹むディッシングやエロージョンを生じると、その結果として配線抵抗値のばらつき幅が増大する。特に図10に示すような多層配線構造をとるシステムLSIと呼ばれるロジックデバイスでは重要な課題となる。つまり、図11に示すように下層の平坦性が低いとCMPの性能以上に平坦化性が損なわれ、研磨残りによる配線間のショートや断線が発生しやすくなるという致命的な不良を生じる原因となる。この内容については「次世代ULSI多層配線の新材料・プロセス技術」技術情報協会(非特許文献3参照)に開示されている。
(2)CMPに用いる消耗品のコストが高い問題がある。これは研磨剤で用いる砥粒の製造コストが高く、粒子サイズを揃えたり、分粒性を維持したり、不純物を除去したりするために極めて注意を要するからである。特にアルミナ砥粒はシリカ砥粒に比べて数倍高価である。また、研磨布は一般的に発泡ポリウレタンが用いられている。CMPを行うと、この研磨布に研磨砥粒が付着して、いわゆる”目詰まり”現象を起こして研磨速度が低下する。これを防止するために適宜ダイヤモンド粒子を固着させた砥石(以下コンディショナと記す)で研磨布表面を削る必要があった。そのため研磨布の寿命は短くなり、研磨砥粒に次ぐ高コストの消耗品となっていた。CMPプロセスのコストに関しては、例えばリアライズ社最新技術講座1996年5月「CMP装置と関連材料の最新動向とその問題点」に記載されている。
【0010】
本発明は係る点に鑑みてなされたものであり、(1)高研磨速度を維持しつつ、埋め込み配線形成時のディッシングやエロージョンの発生を抑制する、(2)研磨液と研磨布のコストを低減できる半導体装置の製造方法を提供することを目的とする。
【0011】
【課題を解決するための手段】
上記の(1)、(2)の目的を達成するための手段について図5を用いて詳細に説明する。研磨は回転定盤に固定した研磨工具上に研磨薬液を供給しつつ、ウエハ基板に圧力をかけながら研磨工具に対してウエハ基板を相対運動することによりなされる。研磨工程をCu高速研磨(工程I)、Cu残段差解消(工程II)、バリア膜研磨(工程III)の3段階に分けて考える。まず工程Iでは研磨工具として研磨パッドを用い、研磨パッド上にCu研磨液1を供給して高研磨速度でCuを研磨する。次に工程IIでは研磨工具を固定砥粒盤に切り換え、研磨液1よりも界面活性剤の濃度を増やしたCu研磨液2を供給し、バリア膜露出前にCu残段差を解消する。更にバリア膜露出直前又は直後の工程IIIの段階でバリア膜研磨液に切り換え、加工中に固定砥粒盤表面をドレスすることによって効率良く銅配線のダマシン平坦化が達成される。
【0012】
次に、この理由について説明する。Cu凸部の研磨速度をR1、Cu凹部の研磨速度をR2とすると、凸部と凹部の研磨速度の比R1/R2(以下、凸凹選択比と言う)は残段差により変化する。すなわち残段差依存性を有する。更に、凸凹選択比の残段差依存性は、研磨工具と研磨液の組み合わせによっても異なる(図7A)。研磨パッドと研磨液のみを用いる従来技術では、凸凹選択比は研磨開始直後に最大値をとり、研磨が進行しCu段差が解消されるにつれてほぼ1に漸近する。これに対して、本発明では固定砥粒盤を用いることで、従来技術よりも大きな凸凹選択比を維持するので、加工後形状の高平坦化が達成される。凸凹選択比がCu残段差が解消する前に1に達する場合、銅膜研磨完了(バリア膜露出)までに凸凹選択比が1になった時点の残段差を維持し続けることを意味する。つまり、従来技術では残段差を解消することが出来ないことがわかる(図7B)。Cu段差を解消し、平坦化するのに最低限必要とされるCuの凸部と凹部の研磨速度の比をn(以下、必要選択比と言う)とすると、必要選択比nは残膜厚と残段差との比の関数で表される。つまり必要選択比nは残膜厚と残段差の幾何学形状から一意に決定する(図12)。一般的なCuダマシン配線の平坦化では、加工前の銅膜の膜厚と残段差の関係は残膜厚/残段差≒2であり、必要選択比n≒2である。加工が進行し、残段差が解消するにつれて、残膜厚/残段差は2よりも小さくなる傾向がある。つまり、残膜厚/残段差の値が2よりも小さくなる(1に近づく)と、図12から明らかなように、必要選択比nは急激に増大する。従って、必要選択比nを出来るだけ小さくするように加工することが望ましい。
【0013】
ここで、凸凹選択比R1/R2が必要選択比nよりも常に大ならば、(R1/R2≧nであれば)銅膜の残膜厚がゼロに達するまで(バリア膜が露出する前)に銅膜に存在する段差を解消できることを意味する。バリア膜が露出する前に銅膜の段差を解消できれば、銅膜研磨に続くバリア膜研磨の際に発生し得るディッシングやエロージョンを抑制する効果が飛躍的に向上するので望ましい。
【0014】
研磨工具1及びCu研磨液1を研磨工具2とCu研磨液2に切り換えるタイミングは、工程Iにおける研磨工具1とCu研磨液1によるCuの凸凹選択比R1/R2が、切り換える時点でのCu残膜厚とCu残段差とから算出される必要選択比nを下回る前であることが望ましく、Cu残膜厚/Cu残段差>2であると更に望ましい。
【0015】
上記のCu研磨液1の凸凹選択比R1/R2≧2、Cu研磨液2の凸凹選択比R1/R2≧3であると好ましく、Cu研磨液1の凸凹選択比R1/R2≧3、Cu研磨液2の凸凹選択比R1/R2≧5であると更に好ましい。
【0016】
上記のCu研磨液1及び2は酸化剤、有機酸、防食剤、界面活性剤、純水からなっており、加工中に所定量が供給される。
【0017】
上記のCu研磨液1及び2の酸化剤としてはとしては過酸化水素が金属成分を含まず、かつ強酸ではないため特に望ましい。
【0018】
上記のCu研磨液1及び2の有機酸としては、リンゴ酸が高研磨速度、低エッチング速度の観点から、本発明の研磨液に使用する酸として望ましい。リンゴ酸は食品添加物としても一般に使用されており、毒性が低く、廃液としての害も低く、臭いもなく、水への溶解度も高いために本発明の研磨液に使用する酸として特に望ましい。
【0019】
上記のCu研磨液1及び2の有機酸として用いるリンゴ酸の添加量は,0.1重量%以上が,Cu研磨速度の観点から望ましい。
【0020】
上記のCu研磨液1及び2の防食剤としては、研磨液中に混合することでエッチングが抑制され、かつ十分な研磨速度が得られる物質であればよい。特に銅合金の防食性物質としてはベンゾトリアゾール(以下BTAと記す)は効果が大きく望ましい。
【0021】
上記のCu研磨液1及び2の防食剤として用いるBTAの添加量は,0.1重量%以上が研磨速度及び防食効果の観点から望ましい。
【0022】
上記のCu研磨液1及び2の防食剤としてBTAの替わりにイミダゾール及びその誘導体を用いることが出来ることは言うまでもない。
【0023】
上記のCu研磨液1及び2の界面活性剤としては、アクリル酸重合体及びそのアンモニウム塩等が高研磨速度、低エッチング速度の観点から本発明の研磨液に使用する界面活性剤として望ましい。
【0024】
上記のCu研磨液1及び2の界面活性剤の分子量としては、分子量5000以上がディッシング抑制効果が大きくて望ましい。
【0025】
研磨液の各材料に対する研磨速度特性は図6のとおりであり、酸化剤の濃度により各材料の研磨速度比(材料選択比)を制御できる。
【0026】
上記のCu研磨液1及びCu研磨液2の過酸化水素水濃度は0.05体積%以上添加することが好ましく、30体積%がCu研磨速度が高くとれて特に好ましい。
【0027】
上記のバリア膜研磨液の過酸化水素水濃度は10体積%以下にすることが好ましく、5体積%がバリア膜の研磨速度がCuの研磨速度と同等にとれて、特に好ましい。
【0028】
上記の(2)研磨液と研磨布のコストを低減する目的は、研磨液に固体砥粒を含まない研磨液を用いることよって達成される。研磨液中に固体砥粒を含まないと、研磨パッドの目詰まりが飛躍的に低減し、ダイヤモンドドレッサによるコンディショニング回数も減らせるので、研磨パッドの長寿命化が可能となり、コスト低減するのに望ましい。
【0029】
【発明の実施の形態】
(実施例1)
以下、本実施例の装置構成を図1から図3を用いて説明する。本発明はこれらの実施例により限定されるものではない。なお、それぞれ図中において、同一の機能の部分には同一の番号を付した。装置は、ウエハを平坦化する研磨工具1及び研磨工具2、ウエハを保持するヘッド3、ヘッドを揺動するスイングアーム4、研磨工具1及び研磨工具2が接着され回転運動する回転定盤10、研磨工具1及び研磨工具2上に研磨液を供給する研磨液供給系6及び7及び8、研磨工具の研磨能力を回復させるためにドレッシングをするドレッサ5を有する。ヘッド3はウエハ9を保持するとともに回転定盤10と同方向に自転できる他、ウエハ表面を研磨工具へ押し付けて所定の力で荷重する機能も有している。更に、CMP中にウエハがはずれないように図示しないリテーナリングが設けられている。研磨液は供給系から100から300cc/minの速度で滴下される。本発明における研磨液供給系は図1に示すとおり3系統ある。これら3系統の研磨液は各液の切替えや流量の制御が行なえる。ドレッサ5は所定の圧力でドレスが可能である。このドレッサにはダイヤモンド砥粒を固着したリング型ドレッサが装着され、毎分200回転程度まで任意の回転駆動が出来る。このドレッサは研磨工具全面をコンディションニングできる構成となっていることはいうまでもない。
【0030】
本実施例では、これら一連の加工プロセスを研磨工具1としてロデール社製の発泡ポリウレタン製研磨パッドIC1000を研磨工具2として独自に開発した固定砥粒盤を用いた。この固定砥粒盤は砥石と弾性パッドの積層構造となっており(以下、積層式固定砥粒盤と呼ぶ)、IC000と同様に、一定の圧力でドレスする定圧ドレス方式によってドレッシングできる。研磨中のウエハ研磨荷重は210g/cm2、回転定盤とヘッドの回転数はともに46rpm(線速度換算で53cm/sec)とした。なお、研磨荷重や回転数はこれに限られるものではない。一般に、荷重や定盤回転数を増やすことにより研磨速度が速くなるが、スクラッチが入りやすくなる。但し、本発明では研磨液中の遊離砥粒濃度が低いので、又は含まないので荷重に対する研磨傷の発生は少ない。ウエハを加工する前に研磨パッド及び積層式固定砥粒盤の表面はドレッサ5によりドレッシングされている。この研磨パッドにCu研磨液1を供給し研磨するが、この時ドレッサ5は研磨パッドの表面からは離れた位置で停止している。
【0031】
次に,研磨液について説明する。まず、Cu研磨液及びバリア膜研磨液の酸化剤である過酸化水素水の調合割合を決定するため、配線パターンが形成されていないウエハを用いて予備実験をした。試料はシリコンウエハ上に厚さ200nmのシリコン酸化膜を形成した後、接着層として厚さ50nmのTaN膜と厚さ800nmの銅膜をスパッタリング法によって真空中で連続成膜したものである。ウエハ直径は8インチである。研磨前後に四探針式抵抗測定機によりシート抵抗を測定し、シート抵抗値の変化を換算して研磨速度を求めた。過酸化水素水の濃度と各材料の研磨速度特性は図6にあるとおり、研磨パッド、固定砥粒盤ともに、過酸化水素水濃度が30体積%で銅の研磨速度が最大となる特性を得た。つまり、この過酸化水素濃度が30体積%のとき、ウエハ一枚の加工時間を最も短縮できてスループットが向上するという望ましい効果が生じる。
【0032】
次に被加工ウエハ9の断面構造について図5を用いて説明する。本図ではウエハ基板のシリコン部分は省略してあり、酸化膜15、バリア膜16と銅膜17の構成のみが表されている。この酸化膜15が二酸化珪素膜以外の低誘電率材料であってもよいし、酸化膜と低誘電率材料との層対であってもよいことは言うまでもない。酸化膜15には配線溝が形成されており、その上から酸化膜中への銅拡散防止のためのバリア膜16が成膜され,さらに銅膜17が成膜されている。
【0033】
次に,本実施例での研磨工程を図5を用いてCu高速研磨(工程I)、Cu残段差解消(工程II)、バリア膜研磨(工程III)の3段階に分けて考える。
【0034】
工程Iでは、研磨工具として研磨パッドを用い、研磨パッド上にCu研磨液1を供給して高研磨速度でCuを研磨する。Cuの高研磨速度化の観点から、予備実験の結果を参考に、酸化剤として過酸化水素水(市販品30体積%濃度)を30体積%、有機酸としてリンゴ酸を0.15重量%、防食剤としてベンゾトリアゾール(BTA)0.2重量%、凸凹選択比R1/R2が約2程度となるように界面活性剤を調整したCu研磨液1を用いた。
【0035】
工程IIでは、研磨工具を固定砥粒盤に切り換え、Cu研磨液2を供給し、バリア膜露出前にCu残段差を解消する。高平坦性の観点から、予備実験の結果を参考に、酸化剤として過酸化水素水(市販品30体積%濃度)を30体積%、有機酸としてリンゴ酸を0.15重量%、防食剤としてベンゾトリアゾール(BTA)0.2重量%、凸凹選択比R1/R2が約3程度となるように界面活性剤を調整したCu研磨液2を用いた。
【0036】
工程IIIでは、バリア膜露出直前又は直後の段階でバリア膜研磨液に切り換え、加工中に固定砥粒盤表面をドレスすることによって効率良く銅配線のダマシン平坦化を達成する。ディッシングやエロージョンといった銅や酸化膜の削れ過ぎを防止する観点から、予備実験の結果も参考に、加工中ドレッシング使用時における銅とバリア膜の研磨速度比が1:1となるように酸化剤濃度を調整したバリア膜研磨液を用いた。
【0037】
本実施例では、図5の工程Iで銅の除去が進み、銅膜の残膜厚が320nm、残段差が150nm程度になった時点で研磨工具をIC1000から積層式固定砥粒盤に替え、研磨液もCu研磨液1からCu研磨液2に切り換えた(工程II)。銅膜の残膜厚が320nm、残段差が150nm程度になった時点での必要選択比nは1.88であり、積層式固定砥粒盤とCu研磨液2による凸凹選択比R1/R2は2.33で必要選択比nを上回っており、ディッシングを抑制しつつ平坦化した。更に銅の除去が進み、バリア膜が露出する直前の状態になった時点でCu研磨液2をバリア膜研磨液に切り換えた(工程III)。本実施例においては研磨液供給系2を止めて研磨液供給系3からバリア膜研磨液の供給に変更すればよい。
【0038】
工程IIIにおける加工中ドレッシングとは、ドレッサ5を研磨工具2である積層式固定砥粒盤の表面に接するまで下降させ、ダイヤモンド粒を積層式固定砥粒盤の表面に接触し、目つぶれした樹脂を除去するとともに新しい砥粒面の露出と遊離砥粒を発生させる機能のことである。ドレッサ5のドレス圧は、加工前に研磨工具1であるIC1000をドレッシングした場合と同一でもよいし、5kg重程度更に増やしてもよい。ドレス圧を増やすことにより遊離砥粒19がより多数発生し、バリア膜16の研磨速度をより向上することができて望ましい。
【0039】
バリア膜16を除去し、酸化膜15が露出すると加工を終了することになる。バリア膜除去の際、バリア膜研磨速度がウエハ面内で均一でない場合、予備実験等から求めた加工時間よりも長めに設定して研磨する(オーバー研磨する)ことがある。酸化膜15の研磨速度がバリア膜16や銅膜17に比べて非常に遅いと、オーバー研磨することにより銅膜やバリア膜の研磨残りを防止することが可能となり、望ましい。更に、酸化膜15の研磨速度がバリア膜16や銅膜17に比べて非常に遅いと、オーバー研磨の際に膜減りを防止することも可能となり、望ましい。例えば、被加工面上のほとんどのバリア膜が除去されるまでの研磨時間を「ジャスト研磨時間」とした場合、研磨残りを防止するには、総研磨時間をジャスト研磨時間の1.3から1.5倍程度にしてオーバー研磨するとよい。
【0040】
上記の平坦化方法を適用してパターン付き8インチ径ウエハを加工した結果、図7Bの結果を得た。図7B及びAにおける相対残段差とは、図7Bで使用したウエハの研磨開始時点での銅膜の初期段差を100として相対化して表したものである。図7Bで使用したウエハの銅膜の初期段差350nm、銅膜の初期膜厚750nm、用いたパターンは銅配線が周期的に並ぶ、いわゆるライン&スペースで1つのライン及びスペースの幅はともに100μmである。図7Bの横軸はこのパターン付きウエハの銅膜の相対残膜厚、縦軸は銅膜の残段差である。樹脂製研磨パッドのみ用いる従来のCMPでは、銅膜の残膜厚がゼロに達しても残段差を解消することが出来ないのに対して、本発明では、銅膜の残膜厚がゼロに達する前に、すなわちバリア膜が露出する前に残段差をゼロとすることが出来、従来技術と比較して極めて高い平坦化性能を実証出来ていることがわかる。この高い平坦化性能は銅の凸部と凹部の加工選択比の残段差依存性を考慮して、より高い選択比を得ることが出来る研磨工具と研磨液に切り換えることに起因し,特に従来のCMPパッドよりも3から10倍高い圧縮弾性率を有する積層式固定砥粒盤をCu研磨の2段目に用いたことに起因する銅の凸部と凹部の高い加工選択比による効果である(図7A)。
【0041】
本実施例では、2プラテン構成のCMP装置を用いたが、前記の諸機能を有するCMP装置であれば、本発明を実施することが出来、例えばアプライド・マテリアルズ社のCMP装置MIRRA MESAを使用することが出来る。その場合、第一プラテンに研磨工具1としてCMPパッドを設置し、第二プラテンに研磨工具2として固定砥粒盤を設置することで、本発明を実施することが出来る。また、本実施例では、回転定盤上に設置した研磨工具を使用したが、研磨工具1として、ベルト式研磨工具を使用し、研磨工具2として砥粒を固定した硬質研磨パッドを使用してもよいことは言うまでもない。更に、本実施例では、研磨工具として回転定盤上に設置した研磨工具を使用したが、ウエハ径よりも小さい、小径研磨工具を使用してもよいことは言うまでもない。また、本実施例では、研磨工具面上に研磨液を滴下したが、研磨工具から研磨液が涌き出るような構成でもよいことは言うまでもない。また,防食剤として,BTAの替わりにイミダゾールを用いることができることは言うまでもない。
【0042】
また本発明は、Cu研磨工程を2段研磨することで実施されることから,Cuの凸部と凹部の選択比が十分高く取れるのであれば、研磨液のみを替えて研磨工具を替えなくてもよいことは言うまでもない。
【0043】
(実施例2)
本実施例では、研磨工具2として砥石の下に弾性パッドがない、つまり積層式ではない通常の固定砥粒盤を用いてCuダマシン配線平坦化を実現する方法を説明する。図4に示すように、研磨工具2に積層式ではない通常の固定砥粒盤を用いる場合は、研磨工具2のドレッシングとして、定寸切り込みによる方法(定寸ドレス方式)を用いることが出来る。この定寸ドレス方式は、ダイヤモンド等の硬質の砥粒を埋め込んだ直径30から70mmのリング又はディスクを、毎分3000から10000回転の高速で回転させながら、ダイヤモンド工具と砥石の距離を一定に保ち砥石面内を相対的に移動させて切り込む方法で、固定砥粒盤表面を精度よくドレッシングすることが出来る。具体的には、固定砥粒盤の表面を深さ1μm程度で極表面層のみを除去するように、絶対高さ位置制御をかけながら固定砥粒盤表面上を走査することで実現できる。このような定寸切り込み加工では、工具高さの位置決め精度を高めれば、原理的により高い平坦度を得ることが出来る。ドレッシングによる固定砥粒盤の研削量は通常10μm以下であるから、固定砥粒盤の厚さが10mmであれば、ウエハを1000枚加工した時点で固定砥粒盤は使用限界に達すると予想できる。ドレッシングによる固定砥粒盤の研削量を減じれば、固定砥粒盤の寿命が更に延びることは自明である。
【0044】
本実施例においても、銅の除去が進み、銅膜の残段差が150nm程度になった時点で研磨工具をIC1000から固定砥粒盤に替え、研磨液もCu研磨液1からCu研磨液2に切り換える。本実施例においても実施例1で既に説明したように、Cu研磨液2の成分構成はCu研磨液1と同様で、界面活性剤の添加濃度が異なる。この添加濃度は銅の凸部と凹部の研磨速度の比が大きくなり、かつ銅の研磨速度が可能な限り大きくなるようになるように調整する。更に銅の除去が進み工程IIIの状態になった時点でCu研磨液2をバリア膜研磨液に切り換える。本実施例においても研磨液供給系2を止めて研磨液供給系3からバリア膜研磨液の供給に変更すればよい。バリア膜研磨液の構成成分は研磨液2と同じで、成分割合の調整方法は酸化剤濃度の調整と後述する加工中のドレッシングの併用で行い、銅とバリア膜の研磨速度比を1:1となるようにした。こうすることによりディッシングやエロージョンといった銅や酸化膜の削れ過ぎを防止できるので、望ましい。
【0045】
工程IIIにおけるドレッシングとは、ドレッサ5を研磨工具2である固定砥粒盤の表面に接するまで下降させ、ダイヤモンド粒が固定砥粒盤面に接触させて目つぶれした樹脂を除去するとともに新しい砥粒面の露出と遊離砥粒を発生させる機能のことである。このドレッサの位置決め高さ位置は、工程IIの平坦化加工前にドレッシングした位置と同一位置でもよいし、1μm程度更に切り込んでもよい。切り込むことにより遊離砥粒19がより多数発生し、バリア膜16の研磨速度をより向上することができて望ましい。
【0046】
次に、バリア膜16を除去し、酸化膜15が露出すると加工を終了することになる。バリア膜除去の際、バリア膜の研磨速度がウエハ面内で均一でない場合、予備実験等から求めた加工時間よりも長めに設定して研磨する(オーバー研磨する)ことがある。酸化膜15の研磨速度がバリア膜16や銅膜17に比べて非常に遅いと、オーバー研磨することにより銅膜やバリア膜の研磨残りを防止することが可能となり、望ましい。更に、酸化膜15の研磨速度がバリア膜16や銅膜17に比べて非常に遅いと、オーバー研磨の際に膜減りを防止することも可能となり、望ましい。例えば、被加工面上のほとんどのバリア膜が除去されるまでの研磨時間を「ジャスト研磨時間」とした場合、研磨残りを防止するには、総研磨時間をジャスト研磨時間の1.3から1.5倍程度にするとよい。
【0047】
上記の平坦化方法を適用してパターン付きウエハを加工した結果、ウエハ全面に渡って図8の結果を得た。用いたパターンは銅配線が周期的に並ぶ、いわゆるライン&スペースで、1つのラインの幅は20μm、ラインの長さは1900μmである。このパターンとパターンの間の距離を横軸とし、ディッシングとエロージョンの和を縦軸とした。樹脂製研磨パッドのみ用いる従来のCMPと比べ、高い平坦化性能を実証出来ていることがわかる。この高い平坦化性能は砥粒を含まない研磨液と従来のCMPのパッドよりも3から10倍高い圧縮弾性率を有する固定砥粒盤をCu研磨の2段目に用いたことに起因する凸部と凹部の高い加工選択比による効果である(図7A)。
【0048】
硬質な固定砥粒盤で平坦化加工する場合の課題のひとつにスクラッチがある。本発明の固定砥粒盤は高純度なシリカ微細砥粒を均質に分散固定してある。よって、スクラッチを生じる原因となる異物や大径粒子を含まないのでスクラッチが生じにくい。図9は銅の120μm四方のパターンを加工した後の表面形状を測定した結果である。表面の凹凸の幅は10nm以下であり、半導体の配線として十分な鏡面となっていることを確認した。
【0049】
(実施例3)
実施例1及び2では平坦化加工前に研磨工具2である固定砥粒盤の全面をドレッサ5によりドレッシングして固定砥粒盤表面に固定砥粒18を露出させたが、この工程を省くことも可能である。銅膜17は砥粒が存在しなくても加工できることが既に知られているので、必ずしも加工前のドレッシングが不可欠ではないからである。また、固定砥粒盤の寿命はドレッシングによる厚さの低減により決定されるものであることから、極力ドレッシングの頻度は下げるべきであり、この観点からも望ましいといえる。従って、実施例1及び実施例2における工程のうち、加工前のドレッシングを省略した半導体配線構造の平坦化加工が可能である。本実施例によれば銅を用いたダマシン配線構造の平坦化が低コスト、高スループット、高平坦化であることに加え、固定砥粒盤の長寿命化も両立できるという大きな効果が見込める。
【0050】
(実施例4)
実際に本発明を半導体デバイスに適用した実施例について図10を用いて説明する。本構造は6層の多層配線ロジックデバイスの断面である。シリコン基板23の表面に浅溝素子分離溝(STI)を酸化膜平坦化加工技術により形成した後、ゲートパターン22等を形成しトランジスタを形成する。その後、上部配線層とのW製コンタクトプラグ21をW平坦化加工技術により形成する。Wプラグも銅配線構造と同様に絶縁膜との界面にバリア膜20が成膜される。このWプラグから上層がすべて銅配線層であり、6層とも本発明を用いて形成した。下地層が平坦であるので、図11で説明した研磨残りやディッシング、エロージョン、スクラッチによる電気的短絡の不良問題が全く発生しない良好な平坦化が行なえた。また、STI層やWプラグ平坦化も固定砥粒を用いた平坦化を適用することによってより平坦な半導体装置を製造することが可能になることは言うまでもない。
【0051】
【発明の効果】
本発明によれば、高研磨速度を維持しつつ、埋め込み配線形成時のディッシングおよびエロージョンの和を、ウエハ全面に渡って40nm未満にまで半減できる。これにより配線抵抗値のばらつき低減や断線不良の低減といった歩留りの向上ができる。また、研磨液と研磨布のコスト低減が図れる。
【図面の簡単な説明】
【図1】本実施例の装置構成を説明する正面図である。
【図2】本実施例の装置構成を説明する側面図である。
【図3】本実施例の装置構成を説明する側面図である。
【図4】本実施例の装置構成を説明する側面図である。
【図5】本実施例の加工工程を説明する図である。
【図6】本実施例の加工薬液酸化剤濃度と研磨速度を説明する図である。
【図7】本実施例の平坦化性能を説明する図である。
【図8】本実施例の平坦化性能を説明する図である。
【図9】本実施例の平坦化後の銅表面形状を説明する図である。
【図10】多層配線構造を有する半導体装置の断面を説明する図である。
【図11】平坦化性能不足に伴う課題を説明する図である。
【図12】本実施例の必要選択比の残段差と残膜厚の依存特性を説明する図である。
【符号の説明】
1.研磨工具1、2.研磨工具2、3.ヘッド、4.スイングアーム、5.ドレッサ、6.研磨液供給系1、7.研磨液供給系2、8.研磨液供給系3、9.ウエハ、10.回転定盤、11.基準面、12.Cu研磨液1、13.Cu研磨液2、14.バリア膜研磨液、15.酸化膜、16.バリア膜、17.銅膜、18.固定砥粒、19.遊離砥粒、20.Wプラグのバリア膜、21.Wプラグ、22.ゲート、23.シリコン基板、24.基板。
[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a polishing or grinding technique for flattening a substrate surface, and more particularly to a method for manufacturing a semiconductor integrated circuit in which a thin film formed on a semiconductor substrate is polished or ground.
[0002]
[Prior art]
2. Description of the Related Art In recent years, a new fine processing technology has been developed in accordance with high integration and high performance of a semiconductor integrated circuit device (hereinafter, referred to as LSI). A chemical mechanical polishing method (Chemical Mechanical Polishing; also referred to as CMP) is one of the techniques, and is a technique frequently used in the LSI manufacturing process, particularly in the flattening of the interlayer insulating film, the formation of metal plugs, and the formation of embedded wiring in a multilayer wiring formation process. is there. This flattening technique is disclosed in, for example, Patent Document 1.
[0003]
In addition, recently, in order to achieve high-speed performance of LSI, it has been attempted to use a copper alloy having a low electric resistance from a conventional aluminum alloy as a wiring material. However, it is difficult to finely process a copper alloy by a dry etching method frequently used for forming an aluminum alloy wiring. Therefore, a tantalum, a tantalum alloy, a tantalum nitride, another tantalum compound, or the like is deposited as a barrier film on the processed insulating film having a groove to prevent copper diffusion into the insulating film. The so-called damascene method of depositing a copper or copper alloy thin film on the substrate and removing the copper or copper alloy thin film other than the portion buried in the groove formed in the insulating film by CMP to form a buried wiring is mainly used. Has been adopted. The wiring technique based on the damascene method is disclosed in, for example, Patent Document 2.
[0004]
A general method of metal CMP is to attach a polishing pad to a circular polishing platen (platen), immerse the polishing pad surface with a metal polishing solution, and press the surface of the substrate on which a metal film is formed, to form a substrate. The polishing platen is rotated while applying a predetermined pressure (hereinafter, referred to as a polishing pressure) from the back surface of the substrate, and the metal film on the convex portion is removed by mechanical friction between the polishing liquid and the convex portion of the metal film.
[0005]
Generally, an abrasive used for CMP of a metal such as a copper alloy used for wiring mainly contains solid abrasive grains and an oxidizing substance. The basic mechanism of CMP is to mechanically remove the oxide by solid abrasive grains while oxidizing the metal surface by the oxidizing action of the oxidizing substance. The mechanism of CMP is disclosed on page 299 of Non-Patent Document 1. As the solid abrasive grains, alumina abrasive grains and silica abrasive grains having a particle diameter of several tens to several hundreds of nm are known, but most of the commercially available solid abrasive grains for metal polishing are the former. Oxidizing substances include hydrogen peroxide (H 2 O 2 ), Ferric nitrate (Fe (NO 3 ) 3 ), Potassium periodate (KIO 3 ) Are widely used in general, and these are disclosed, for example, in Non-Patent Document 1 at pages 299 to 300.
Further, the flattening of the STI layer using fixed abrasive grains is described in, for example, Patent Document 3.
[0006]
Conventional techniques using a plurality of polishing pads are disclosed in Patent Literature 4, Patent Literature 5, Patent Literature 6, Patent Literature 7, Patent Literature 8, Patent Literature 9, and the like.
[0007]
[Patent Document 1]
U.S. Pat. 4944836
[Patent Document 2]
JP-A-2-278822
[Patent Document 3]
JP 2000-49122 A
[Patent Document 4]
JP-A-2000-164595
[Patent Document 5]
JP-A-9-326392
[Patent Document 6]
JP 2001-135605 A
[Patent Document 7]
JP 2001-170858 A
[Patent Document 8]
JP 2001-332517 A
[Patent Document 9]
JP-T-2002-512894
[Non-patent document 1]
Edited by Masahiro Kashiwagi, "Science of CMP", August 20, 1997 (published by Science Forum Inc.), pp. 299 and 300.
[Non-patent document 2]
"Journal of Electrochemical Society (J. Electrochem. Soc.)", Vol. 141, No. 10, October 1994, p. 2842-2848
[Non-Patent Document 3]
"New Materials and Process Technologies for Next-Generation ULSI Multilayer Wiring", Technical Information Association, p. 242-246
[0008]
[Problems to be solved by the invention]
However, in the above-described prior art, there is a problem that the polishing rate is high but the flatness is insufficient, or the polishing rate is insufficient even though the flatness is high. Furthermore, if the step between the surface of the convex portion of the deposited Cu film and the bottom surface of the concave portion is referred to as a Cu residual step or a residual step, no consideration is given to the change and the polishing solution, and There are (1) insufficient flatness and (2) high cost.
(1) Since the polishing pad is soft (for example, a polishing pad made of polyurethane and has a compression modulus of 100 MPa or less), the polishing pad is deformed into irregularities by a force pressing the substrate surface, and the insulating film is formed. The phenomenon that the central part of the surface of the metal wiring embedded in the groove formed in the trench is excessively polished and dented (hereinafter referred to as dishing) from the peripheral part and the phenomenon that the surface of the insulating film around the wiring part is excessively polished ( (Hereinafter referred to as erosion) (FIG. 11). Further, tantalum, tantalum alloys, tantalum nitride, and other tantalum compounds used as barrier films are chemically stable and difficult to etch, and have high hardness, so that mechanical polishing is not as easy as copper and copper alloys. Therefore, when the hardness of the abrasive grains is increased, polishing scratches occur in copper or a copper alloy, which causes electrical characteristics failure. When the particle concentration of the abrasive grains is increased, the polishing rate of the insulating film is increased. And erosion is likely to occur.
[0009]
Further, it is known that in a CMP of a wafer in which a metal film is deposited on an insulating film in which a groove serving as a wiring pattern is formed, a portion which is locally excessively polished occurs. This is because unevenness reflecting grooves serving as wiring patterns is formed on the surface of the metal film on the wafer surface before CMP, and when CMP is performed, locally high pressure is applied in accordance with the pattern density, and the This is because the polishing rate increases. Therefore, dishing or erosion is a significant problem in a pad having a large metal area (area of about 0.1 mm square) or a dense wiring pattern. These are described in Journal of Electrochemical Society, Vol. 141, No. 10, October 1994 (see Non-Patent Document 2). When dishing or erosion in which the wiring surface is depressed occurs, as a result, the variation width of the wiring resistance value increases. In particular, this is an important issue in a logic device called a system LSI having a multilayer wiring structure as shown in FIG. That is, as shown in FIG. 11, when the flatness of the lower layer is low, the flatness is impaired more than the performance of CMP, and a short circuit or disconnection between wirings due to unpolished residue easily occurs, which causes a fatal defect. Become. This content is disclosed in "New Material and Process Technology for Next-Generation ULSI Multilayer Wiring" Technical Information Association (see Non-Patent Document 3).
(2) There is a problem that the cost of consumables used for CMP is high. This is because the production cost of the abrasive used as the abrasive is high, and extreme care is required to make the particle size uniform, maintain the particle size, and remove impurities. In particular, alumina abrasive grains are several times more expensive than silica abrasive grains. Further, foamed polyurethane is generally used for the polishing cloth. When the CMP is performed, abrasive grains adhere to the polishing cloth, causing a so-called "clogging" phenomenon, and the polishing rate is reduced. In order to prevent this, it was necessary to sharpen the polishing cloth surface with a grindstone (hereinafter, referred to as a conditioner) to which diamond particles were appropriately fixed. As a result, the life of the polishing cloth has been shortened, and it has become a high-cost consumable item next to abrasive grains. The cost of the CMP process is described in, for example, “Latest Trends in CMP Equipment and Related Materials and Their Problems” in May 1996, the latest technical course by Realize.
[0010]
The present invention has been made in view of the above points, and (1) suppresses the occurrence of dishing and erosion when forming embedded wiring while maintaining a high polishing rate. (2) Reduces the cost of a polishing liquid and a polishing cloth. It is an object to provide a method for manufacturing a semiconductor device which can be reduced.
[0011]
[Means for Solving the Problems]
Means for achieving the above objects (1) and (2) will be described in detail with reference to FIG. Polishing is performed by supplying a polishing agent solution onto a polishing tool fixed to a rotary platen, and moving the wafer substrate relative to the polishing tool while applying pressure to the wafer substrate. The polishing process is divided into three stages: high-speed polishing of Cu (process I), elimination of residual steps of Cu (process II), and polishing of a barrier film (process III). First, in step I, a polishing pad is used as a polishing tool, and a Cu polishing liquid 1 is supplied onto the polishing pad to polish Cu at a high polishing rate. Next, in the step II, the polishing tool is switched to the fixed abrasive disk, and the Cu polishing liquid 2 having a higher concentration of the surfactant than the polishing liquid 1 is supplied to eliminate the Cu residual step before the barrier film is exposed. Further, by switching to the barrier film polishing liquid at the stage of the process III immediately before or immediately after the exposure of the barrier film and dressing the surface of the fixed abrasive disk during the processing, the damascene flattening of the copper wiring is efficiently achieved.
[0012]
Next, the reason will be described. Assuming that the polishing rate of the Cu convex portion is R1 and the polishing speed of the Cu concave portion is R2, the ratio R1 / R2 of the polishing speed of the convex portion and the concave portion (hereinafter, referred to as the convex / concave selectivity) changes depending on the remaining step. That is, it has dependency on the remaining step. Furthermore, the residual step difference dependence of the unevenness selectivity differs depending on the combination of the polishing tool and the polishing liquid (FIG. 7A). In the conventional technique using only a polishing pad and a polishing liquid, the unevenness selectivity has a maximum value immediately after the start of polishing, and gradually approaches 1 as polishing progresses and the Cu step is eliminated. On the other hand, in the present invention, by using a fixed abrasive disc, a higher unevenness selectivity is maintained than in the prior art, so that a highly flattened shape after processing is achieved. When the unevenness selectivity reaches 1 before the Cu residual step is eliminated, it means that the residual step at the time when the unevenness selectivity becomes 1 is continuously maintained until the polishing of the copper film is completed (barrier film exposure). That is, it is understood that the remaining step cannot be eliminated by the conventional technique (FIG. 7B). Assuming that the ratio of the polishing rate of the convex portion to the concave portion, which is the minimum required for eliminating the Cu step and flattening, is n (hereinafter, referred to as a required selection ratio), the required selection ratio n is the remaining film thickness. It is expressed as a function of the ratio between the remaining step and the remaining step. That is, the necessary selection ratio n is uniquely determined from the geometrical shape of the remaining film thickness and the remaining step (FIG. 12). In general planarization of Cu damascene wiring, the relationship between the film thickness of the copper film before processing and the remaining step is the remaining film thickness / the remaining step ≒ 2, and the necessary selection ratio n ≒ 2. As the processing proceeds and the remaining step is eliminated, the ratio of remaining film thickness / remaining step tends to be smaller than 2. That is, when the value of remaining film thickness / remaining step becomes smaller than 2 (approaches 1), the necessary selection ratio n sharply increases as is apparent from FIG. Therefore, it is desirable to perform processing so that the necessary selection ratio n is made as small as possible.
[0013]
Here, if the unevenness selectivity R1 / R2 is always higher than the required selectivity n, the remaining film thickness of the copper film reaches zero (before the barrier film is exposed) (if R1 / R2 ≧ n). Means that the step existing in the copper film can be eliminated. It is desirable that the step of the copper film can be eliminated before the barrier film is exposed, because the effect of suppressing dishing and erosion that may occur during polishing of the barrier film subsequent to polishing of the copper film is remarkably improved.
[0014]
The timing at which the polishing tool 1 and the Cu polishing liquid 1 are switched to the polishing tool 2 and the Cu polishing liquid 2 is determined by the selection ratio R1 / R2 of Cu between the polishing tool 1 and the Cu polishing liquid 1 in the process I. It is desirable that the ratio be less than the required selection ratio n calculated from the film thickness and the Cu residual step, and it is more desirable that Cu residual film thickness / Cu residual step be greater than 2.
[0015]
It is preferable that the above-mentioned unevenness selectivity R1 / R2 ≧ 2 of the Cu polishing liquid 1 and the unevenness selectivity R1 / R2 ≧ 3 of the Cu polishing liquid 2 be satisfied. It is further preferable that the unevenness selectivity R1 / R2 ≧ 5 of the liquid 2 is satisfied.
[0016]
The Cu polishing liquids 1 and 2 are composed of an oxidizing agent, an organic acid, an anticorrosive, a surfactant, and pure water, and are supplied in a predetermined amount during processing.
[0017]
As an oxidizing agent for the Cu polishing liquids 1 and 2, hydrogen peroxide is particularly preferable because it does not contain a metal component and is not a strong acid.
[0018]
As the organic acid of the Cu polishing liquids 1 and 2, malic acid is desirable as the acid used in the polishing liquid of the present invention from the viewpoint of a high polishing rate and a low etching rate. Malic acid is also commonly used as a food additive, and is particularly desirable as an acid used in the polishing liquid of the present invention because of its low toxicity, low harm as a waste liquid, no odor, and high solubility in water.
[0019]
The amount of malic acid used as the organic acid in the Cu polishing liquids 1 and 2 is preferably 0.1% by weight or more from the viewpoint of the Cu polishing rate.
[0020]
As the anticorrosive for the Cu polishing liquids 1 and 2, any substance can be used as long as mixing with the polishing liquid suppresses etching and provides a sufficient polishing rate. In particular, benzotriazole (hereinafter referred to as BTA) is highly effective and desirable as an anticorrosive substance for copper alloys.
[0021]
The addition amount of BTA used as an anticorrosive for the Cu polishing liquids 1 and 2 is preferably 0.1% by weight or more from the viewpoint of the polishing rate and the anticorrosion effect.
[0022]
It goes without saying that imidazole and its derivatives can be used instead of BTA as the anticorrosive for the Cu polishing liquids 1 and 2.
[0023]
As the surfactant of the above Cu polishing liquids 1 and 2, an acrylic acid polymer and its ammonium salt are desirable as the surfactant used in the polishing liquid of the present invention from the viewpoint of a high polishing rate and a low etching rate.
[0024]
As the molecular weight of the surfactant of the Cu polishing liquids 1 and 2, a molecular weight of 5,000 or more is preferable because the dishing suppressing effect is large.
[0025]
The polishing rate characteristics of the polishing liquid for each material are as shown in FIG. 6, and the polishing rate ratio (material selection ratio) of each material can be controlled by the concentration of the oxidizing agent.
[0026]
The concentration of hydrogen peroxide in the Cu polishing liquid 1 and the Cu polishing liquid 2 is preferably 0.05 vol% or more, and 30 vol% is particularly preferable because the Cu polishing rate can be increased.
[0027]
The aqueous hydrogen peroxide concentration of the above-mentioned barrier film polishing solution is preferably 10% by volume or less, and 5% by volume is particularly preferable because the polishing rate of the barrier film can be made equal to the polishing rate of Cu.
[0028]
The object of (2) reducing the cost of the polishing liquid and the polishing cloth is achieved by using a polishing liquid containing no solid abrasive grains. If solid abrasive grains are not contained in the polishing liquid, clogging of the polishing pad is dramatically reduced, and the number of times of conditioning by the diamond dresser can be reduced, so that the life of the polishing pad can be extended, which is desirable for cost reduction. .
[0029]
BEST MODE FOR CARRYING OUT THE INVENTION
(Example 1)
Hereinafter, the device configuration of the present embodiment will be described with reference to FIGS. The present invention is not limited by these examples. In the drawings, the same function is denoted by the same reference numeral. The apparatus includes a polishing tool 1 and a polishing tool 2 for flattening a wafer, a head 3 for holding a wafer, a swing arm 4 for swinging the head, a rotating platen 10 on which the polishing tool 1 and the polishing tool 2 are bonded and rotated, It has polishing liquid supply systems 6, 7 and 8 for supplying a polishing liquid onto the polishing tool 1 and the polishing tool 2, and a dresser 5 for dressing in order to recover the polishing capability of the polishing tool. The head 3 can hold the wafer 9 and rotate in the same direction as the rotary platen 10, and also has a function of pressing the wafer surface against a polishing tool and applying a load with a predetermined force. Further, a retainer ring (not shown) is provided so that the wafer does not come off during the CMP. The polishing liquid is dropped from the supply system at a rate of 100 to 300 cc / min. As shown in FIG. 1, there are three polishing liquid supply systems in the present invention. These three types of polishing liquids can switch the respective liquids and control the flow rate. The dresser 5 can be dressed with a predetermined pressure. The dresser is equipped with a ring-type dresser to which diamond abrasive grains are fixed, and can be arbitrarily rotated up to about 200 revolutions per minute. Needless to say, this dresser is configured so that the entire surface of the polishing tool can be conditioned.
[0030]
In the present embodiment, a fixed abrasive disc, which was originally developed as a polishing tool 2, was a polishing pad IC1000 made of foamed polyurethane made by Rodale Co., Ltd. as a polishing tool 1 for these series of processing processes. This fixed abrasive disk has a laminated structure of a grindstone and an elastic pad (hereinafter, referred to as a laminated fixed abrasive disk), and can be dressed by a constant pressure dressing method of dressing at a constant pressure, similarly to the IC000. The wafer polishing load during polishing was 210 g / cm 2, and the rotation speeds of the rotating platen and the head were both 46 rpm (53 cm / sec in terms of linear velocity). The polishing load and the number of rotations are not limited to these. Generally, the polishing rate is increased by increasing the load and the number of revolutions of the platen, but scratches are likely to occur. However, in the present invention, since the concentration of free abrasive grains in the polishing liquid is low or not contained, the occurrence of polishing scratches with respect to the load is small. Before processing the wafer, the surfaces of the polishing pad and the lamination-type fixed abrasive disc are dressed by a dresser 5. The polishing pad 1 is supplied with the Cu polishing liquid 1 for polishing. At this time, the dresser 5 is stopped at a position away from the surface of the polishing pad.
[0031]
Next, the polishing liquid will be described. First, a preliminary experiment was performed using a wafer on which a wiring pattern was not formed in order to determine the proportion of hydrogen peroxide solution that is an oxidizing agent for the Cu polishing liquid and the barrier film polishing liquid. The sample is obtained by forming a 200-nm-thick silicon oxide film on a silicon wafer and then continuously forming a 50-nm-thick TaN film and an 800-nm-thick copper film as an adhesive layer in a vacuum by a sputtering method. The wafer diameter is 8 inches. Before and after the polishing, the sheet resistance was measured with a four-probe resistance meter, and the change in the sheet resistance value was converted to obtain the polishing rate. As shown in FIG. 6, the concentration of the hydrogen peroxide solution and the polishing rate characteristics of each material are as shown in FIG. 6. For both the polishing pad and the fixed abrasive disk, the characteristic that the polishing rate of copper becomes maximum when the hydrogen peroxide solution concentration is 30% by volume is obtained. Was. That is, when the hydrogen peroxide concentration is 30% by volume, a desirable effect that the processing time of one wafer can be shortened most and the throughput is improved is produced.
[0032]
Next, a sectional structure of the wafer 9 to be processed will be described with reference to FIG. In this drawing, the silicon portion of the wafer substrate is omitted, and only the configuration of the oxide film 15, the barrier film 16, and the copper film 17 is shown. It goes without saying that oxide film 15 may be a low dielectric constant material other than the silicon dioxide film, or may be a layer pair of an oxide film and a low dielectric constant material. A wiring groove is formed in the oxide film 15, a barrier film 16 for preventing copper diffusion into the oxide film is formed thereon, and a copper film 17 is further formed.
[0033]
Next, the polishing process in the present embodiment will be considered with reference to FIG. 5 by dividing into three stages of high-speed polishing of Cu (step I), elimination of residual steps of Cu (step II), and polishing of a barrier film (step III).
[0034]
In step I, a polishing pad is used as a polishing tool, and a Cu polishing liquid 1 is supplied onto the polishing pad to polish Cu at a high polishing rate. From the viewpoint of increasing the polishing rate of Cu, referring to the results of preliminary experiments, 30% by volume of hydrogen peroxide solution (commercially available 30% by volume concentration) as an oxidizing agent, 0.15% by weight of malic acid as an organic acid, As the anticorrosive, benzotriazole (BTA) was used in an amount of 0.2% by weight, and a Cu polishing liquid 1 in which a surfactant was adjusted so that the roughness selectivity R1 / R2 was about 2 was used.
[0035]
In the step II, the polishing tool is switched to the fixed abrasive disk, the Cu polishing liquid 2 is supplied, and the residual Cu step is eliminated before the barrier film is exposed. From the viewpoint of high flatness, referring to the results of the preliminary experiment, 30% by volume of aqueous hydrogen peroxide (commercially available 30% by volume concentration) as an oxidizing agent, 0.15% by weight of malic acid as an organic acid, and as an anticorrosive A Cu polishing liquid 2 in which benzotriazole (BTA) was 0.2% by weight and a surfactant was adjusted so that the roughness selectivity R1 / R2 was about 3 was used.
[0036]
In the process III, the barrier film polishing liquid is switched immediately before or immediately after the barrier film is exposed, and the damascene flattening of the copper wiring is efficiently achieved by dressing the surface of the fixed abrasive disk during the processing. From the viewpoint of preventing excessive removal of copper and oxide film such as dishing and erosion, refer to the results of preliminary experiments and adjust the oxidizing agent concentration so that the polishing rate ratio between copper and barrier film when using dressing during processing is 1: 1. Was used.
[0037]
In this embodiment, the removal of copper proceeds in step I of FIG. 5, and when the remaining film thickness of the copper film becomes about 320 nm and the remaining step becomes about 150 nm, the polishing tool is changed from the IC 1000 to a laminated fixed abrasive disc, The polishing liquid was also switched from Cu polishing liquid 1 to Cu polishing liquid 2 (Step II). When the remaining film thickness of the copper film is 320 nm and the remaining step is about 150 nm, the necessary selection ratio n is 1.88, and the unevenness selection ratio R1 / R2 between the laminated fixed abrasive disk and the Cu polishing liquid 2 is At 2.33, the required selection ratio n was exceeded, and flattening was performed while suppressing dishing. When the removal of copper further progressed and the state immediately before the barrier film was exposed was reached, the Cu polishing liquid 2 was switched to the barrier film polishing liquid (step III). In this embodiment, the polishing liquid supply system 2 may be stopped and the supply of the polishing liquid may be changed from the polishing liquid supply system 3 to the supply of the barrier film polishing liquid.
[0038]
In-process dressing in the process III means that the dresser 5 is lowered until the dresser 5 comes into contact with the surface of the lamination-type fixed abrasive disc, which is the polishing tool 2, and the diamond particles are brought into contact with the surface of the lamination-fixed abrasive disc, and the crushed resin is removed. This is a function that removes abrasive grains and exposes new abrasive grains and generates loose abrasive grains. The dressing pressure of the dresser 5 may be the same as that when the IC 1000 as the polishing tool 1 is dressed before processing, or may be further increased by about 5 kg. By increasing the dress pressure, more free abrasive grains 19 are generated, and the polishing rate of the barrier film 16 can be further improved, which is desirable.
[0039]
When the barrier film 16 is removed and the oxide film 15 is exposed, the processing ends. When removing the barrier film, if the barrier film polishing rate is not uniform in the wafer plane, the polishing may be performed with a longer time set than the processing time obtained from preliminary experiments (over polishing). If the polishing rate of the oxide film 15 is much lower than that of the barrier film 16 or the copper film 17, over-polishing can prevent the copper film or the barrier film from being left unpolished, which is desirable. Further, when the polishing rate of the oxide film 15 is much lower than that of the barrier film 16 or the copper film 17, it is possible to prevent the film from being reduced during overpolishing, which is desirable. For example, when the polishing time until most of the barrier film on the surface to be processed is removed is referred to as “just polishing time”, the total polishing time is reduced from 1.3 to 1 of the just polishing time in order to prevent remaining polishing. It is preferable to over-polish by about 0.5 times.
[0040]
As a result of processing an 8-inch diameter wafer with a pattern by applying the above flattening method, the result of FIG. 7B was obtained. The relative residual steps in FIGS. 7B and 7A are expressed as relative values with the initial step of the copper film at the start of polishing of the wafer used in FIG. The initial step height of the copper film of the wafer used in FIG. 7B is 350 nm, the initial thickness of the copper film is 750 nm, and the pattern used is a so-called line & space in which copper wiring is periodically arranged. is there. The horizontal axis in FIG. 7B is the relative remaining film thickness of the copper film of this patterned wafer, and the vertical axis is the remaining step difference of the copper film. In the conventional CMP using only the resin polishing pad, the residual step cannot be eliminated even if the residual film thickness of the copper film reaches zero, whereas in the present invention, the residual film thickness of the copper film becomes zero. Before reaching, that is, before the barrier film is exposed, the remaining step can be made zero, and it can be seen that extremely high flattening performance has been demonstrated as compared with the conventional technology. This high planarization performance is caused by switching to a polishing tool and a polishing solution that can obtain a higher selectivity in consideration of the dependency of the processing selectivity of the copper convex and concave portions on the residual step difference. This is the effect of the high processing selectivity of the copper projections and depressions due to the use of the laminated fixed abrasive disc having a compression elastic modulus of 3 to 10 times higher than that of the CMP pad in the second stage of Cu polishing ( (FIG. 7A).
[0041]
In this embodiment, a CMP apparatus having a two-platen structure is used. However, the present invention can be implemented as long as the CMP apparatus has the above-described functions. For example, a CMP apparatus MIRRRA MESA manufactured by Applied Materials is used. You can do it. In this case, the present invention can be implemented by installing a CMP pad as the polishing tool 1 on the first platen and installing a fixed abrasive disk as the polishing tool 2 on the second platen. In this embodiment, a polishing tool installed on a rotary platen was used. However, a belt-type polishing tool was used as the polishing tool 1, and a hard polishing pad on which abrasive grains were fixed was used as the polishing tool 2. Needless to say, it is good. Further, in this embodiment, a polishing tool installed on a rotary platen is used as a polishing tool, but it goes without saying that a small-diameter polishing tool smaller than the wafer diameter may be used. Further, in the present embodiment, the polishing liquid is dropped on the polishing tool surface, but it is needless to say that the polishing liquid may flow out of the polishing tool. It goes without saying that imidazole can be used instead of BTA as an anticorrosive.
[0042]
Further, since the present invention is carried out by performing a two-stage polishing in the Cu polishing step, if the selectivity between the convex portions and the concave portions of Cu can be sufficiently high, the polishing tool can be changed by changing only the polishing liquid without changing the polishing tool. Needless to say, it is good.
[0043]
(Example 2)
In the present embodiment, a method of realizing Cu damascene wiring flattening by using an ordinary fixed abrasive disc that does not have an elastic pad under a grindstone as a polishing tool 2, that is, is not a laminated type will be described. As shown in FIG. 4, when an ordinary fixed abrasive disk that is not a lamination type is used for the polishing tool 2, a method using a fixed size cut (a fixed size dressing method) can be used as the dressing of the polishing tool 2. In this fixed-size dressing system, the distance between the diamond tool and the grinding wheel is kept constant while rotating a ring or disc with a diameter of 30 to 70 mm embedded with hard abrasive grains such as diamond at a high speed of 3000 to 10000 revolutions per minute. The fixed abrasive grain surface can be dressed accurately by a method of cutting by relatively moving the inside of the grindstone surface. Specifically, it can be realized by scanning the surface of the fixed abrasive disc while controlling the absolute height position so that the surface of the fixed abrasive disc is only about 1 μm deep and only the very surface layer is removed. In such a fixed size cutting process, if the positioning accuracy of the tool height is increased, a higher flatness can be obtained in principle. Since the grinding amount of the fixed abrasive disc by dressing is usually 10 μm or less, if the thickness of the fixed abrasive disc is 10 mm, it can be expected that the fixed abrasive disc will reach the use limit at the time of processing 1000 wafers. . It is obvious that reducing the grinding amount of the fixed abrasive disc by dressing further extends the life of the fixed abrasive disc.
[0044]
Also in this example, when the removal of copper progressed and the residual step of the copper film became about 150 nm, the polishing tool was changed from the IC 1000 to a fixed abrasive disk, and the polishing liquid was changed from the Cu polishing liquid 1 to the Cu polishing liquid 2. Switch. Also in this embodiment, as already described in the first embodiment, the component composition of the Cu polishing liquid 2 is the same as that of the Cu polishing liquid 1, and the added concentration of the surfactant is different. This addition concentration is adjusted so that the ratio of the polishing rate of the convex portion to the concave portion of the copper becomes large and the polishing rate of the copper becomes as high as possible. Further, when the removal of copper proceeds and the state of step III is reached, the Cu polishing liquid 2 is switched to the barrier film polishing liquid. Also in this embodiment, the polishing liquid supply system 2 may be stopped and the supply of the polishing liquid from the polishing liquid supply system 3 may be changed. The constituent components of the barrier film polishing liquid are the same as those of the polishing liquid 2. The component ratio is adjusted by adjusting the concentration of the oxidizing agent and using dressing during processing, which will be described later, to reduce the polishing rate ratio between copper and the barrier film to 1: 1. It was made to become. This is preferable because it is possible to prevent excessive removal of copper or an oxide film such as dishing or erosion.
[0045]
The dressing in the process III means that the dresser 5 is lowered until the dresser 5 comes into contact with the surface of the fixed abrasive disc, which is the polishing tool 2, and the diamond grains are brought into contact with the fixed abrasive disc face to remove the crushed resin and to form a new abrasive grain face. And the function of generating loose abrasive grains. The positioning height position of the dresser may be the same position as the position dressed before the flattening process in step II, or may be further cut by about 1 μm. By cutting, more free abrasive grains 19 are generated, and the polishing rate of the barrier film 16 can be further improved, which is desirable.
[0046]
Next, when the barrier film 16 is removed and the oxide film 15 is exposed, the processing ends. When removing the barrier film, if the polishing rate of the barrier film is not uniform in the wafer plane, the polishing may be performed with a longer setting than the processing time obtained from preliminary experiments (over polishing). If the polishing rate of the oxide film 15 is much lower than that of the barrier film 16 or the copper film 17, over-polishing can prevent the copper film or the barrier film from being left unpolished, which is desirable. Further, when the polishing rate of the oxide film 15 is much lower than that of the barrier film 16 or the copper film 17, it is possible to prevent the film from being reduced during overpolishing, which is desirable. For example, when the polishing time until most of the barrier film on the surface to be processed is removed is referred to as “just polishing time”, the total polishing time is reduced from 1.3 to 1 of the just polishing time in order to prevent remaining polishing. It is good to make it about 5 times.
[0047]
As a result of processing the patterned wafer by applying the above-mentioned flattening method, the result of FIG. 8 was obtained over the entire surface of the wafer. The pattern used is a so-called line & space in which copper wirings are arranged periodically, and the width of one line is 20 μm and the length of the line is 1900 μm. The horizontal axis represents the distance between the patterns, and the vertical axis represents the sum of dishing and erosion. It can be seen that higher planarization performance can be demonstrated as compared to conventional CMP using only a resin polishing pad. This high flattening performance is caused by the use of a polishing solution containing no abrasive grains and a second-stage Cu polishing using a fixed abrasive disc having a compression modulus of 3 to 10 times higher than that of a conventional CMP pad. This is the effect of the high processing selection ratio between the part and the concave part (FIG. 7A).
[0048]
One of the problems when flattening with a hard fixed abrasive disc is scratching. The fixed abrasive disk of the present invention has high-purity silica fine abrasive particles dispersed and fixed uniformly. Therefore, scratches are unlikely to occur because they do not contain foreign matter or large-diameter particles that cause scratches. FIG. 9 shows the result of measuring the surface shape after processing a copper pattern of 120 μm square. The width of the irregularities on the surface was 10 nm or less, and it was confirmed that the surface had a mirror surface sufficient for semiconductor wiring.
[0049]
(Example 3)
In the first and second embodiments, the entire surface of the fixed abrasive disc serving as the polishing tool 2 was dressed by the dresser 5 before the flattening process, so that the fixed abrasive grains 18 were exposed on the surface of the fixed abrasive disc. Is also possible. This is because it is already known that the copper film 17 can be processed even without abrasive grains, so dressing before processing is not necessarily indispensable. In addition, since the life of the fixed abrasive disc is determined by the reduction in thickness due to dressing, the frequency of dressing should be reduced as much as possible, which is preferable from this viewpoint. Therefore, it is possible to perform the flattening process of the semiconductor wiring structure in which the dressing before the processing is omitted in the steps in the first and second embodiments. According to the present embodiment, in addition to flattening the damascene wiring structure using copper at low cost, high throughput, and high flattening, a great effect that a longer life of the fixed abrasive disk can be achieved can be expected.
[0050]
(Example 4)
An embodiment in which the present invention is actually applied to a semiconductor device will be described with reference to FIG. This structure is a cross section of a six-layer multilayer wiring logic device. After a shallow trench isolation trench (STI) is formed on the surface of the silicon substrate 23 by an oxide film flattening technique, a gate pattern 22 and the like are formed to form a transistor. Thereafter, a W contact plug 21 with the upper wiring layer is formed by W flattening processing technology. The barrier film 20 is formed on the interface between the W plug and the insulating film, similarly to the copper wiring structure. All layers above the W plug are copper wiring layers, and all six layers are formed using the present invention. Since the underlayer is flat, good flattening without the problem of unsatisfactory electrical shorting due to polishing residue, dishing, erosion, or scratching described with reference to FIG. 11 can be performed. In addition, it goes without saying that a flattened semiconductor device can be manufactured by applying the flattening using the fixed abrasive to the STI layer and the W plug.
[0051]
【The invention's effect】
According to the present invention, the sum of dishing and erosion during the formation of embedded wiring can be halved to less than 40 nm over the entire surface of the wafer while maintaining a high polishing rate. As a result, the yield can be improved, such as reduction in variation in wiring resistance value and reduction in disconnection failure. Further, the cost of the polishing liquid and the polishing cloth can be reduced.
[Brief description of the drawings]
FIG. 1 is a front view illustrating the configuration of an apparatus according to an embodiment.
FIG. 2 is a side view for explaining the device configuration of the present embodiment.
FIG. 3 is a side view illustrating an apparatus configuration of the present embodiment.
FIG. 4 is a side view illustrating an apparatus configuration of the present embodiment.
FIG. 5 is a diagram illustrating a processing step of the present embodiment.
FIG. 6 is a diagram illustrating the concentration of a processing chemical oxidant and a polishing rate according to the present embodiment.
FIG. 7 is a diagram illustrating the flattening performance of the present embodiment.
FIG. 8 is a diagram illustrating the flattening performance of the present embodiment.
FIG. 9 is a diagram for explaining a copper surface shape after flattening in this example.
FIG. 10 is a diagram illustrating a cross section of a semiconductor device having a multilayer wiring structure.
FIG. 11 is a diagram illustrating a problem associated with insufficient flattening performance.
FIG. 12 is a diagram illustrating the dependence of the required selectivity on the remaining step and the remaining film thickness in this embodiment.
[Explanation of symbols]
1. Polishing tools 1, 2,. Polishing tools 2, 3, 2. Head, 4. Swing arm, 5. Dresser, 6. Polishing liquid supply system 1,7. 7. Polishing liquid supply system 2,8. Polishing liquid supply system 3,9. Wafer, 10. Rotating platen, 11. Reference plane, 12. Cu polishing liquid 1, 13. 13. Cu polishing liquid 2, 14 14. barrier film polishing liquid; Oxide film, 16. Barrier film, 17. Copper film, 18. Fixed abrasive, 19. Loose abrasives, 20. 21. W plug barrier film W plug, 22. Gate, 23. Silicon substrate, 24. substrate.

Claims (18)

基板上に開口部を有する絶縁膜に第1の導電膜と第2の導電膜を形成する工程と、
前記第1の導電膜と前記第2の導電膜の一部を研磨工具を用いて加工することにより除去し、前記開口部内に前記第1の導電膜および前記第2の導電膜を残存せしめる工程とを備え、
前記第2の導電膜の加工には第1の研磨工具を使用し、前記第2の導電膜の加工途中で第2の研磨工具に切り替える工程を含むことを特徴とする半導体装置の製造方法。
Forming a first conductive film and a second conductive film on an insulating film having an opening on a substrate;
Removing a part of the first conductive film and the second conductive film by processing with a polishing tool, and leaving the first conductive film and the second conductive film in the opening; With
A method for manufacturing a semiconductor device, comprising a step of using a first polishing tool for processing the second conductive film and switching to a second polishing tool during the processing of the second conductive film.
基板上に開口部を有する絶縁膜に第1の導電膜と第2の導電膜を形成する工程と、
前記第1の導電膜と前記第2の導電膜の一部を研磨工具を用いて加工することにより除去し、前記開口部内に前記第1の導電膜および前記第2の導電膜を残存せしめる工程とを備え、
前記第2の導電膜の加工には第1の研磨工具を使用し、前記第1の導電膜の加工には第2の研磨工具を使用することを特徴とする半導体装置の製造方法。
Forming a first conductive film and a second conductive film on an insulating film having an opening on a substrate;
Removing a part of the first conductive film and the second conductive film by processing with a polishing tool, and leaving the first conductive film and the second conductive film in the opening; With
A method for manufacturing a semiconductor device, comprising: using a first polishing tool for processing the second conductive film; and using a second polishing tool for processing the first conductive film.
前記第1の研磨工具には、第1の圧縮弾性率を有する材料が設けられ、前記第2の研磨工具には、前記第1の圧縮弾性率より高い圧縮弾性率を有する材料が設けられていることを特徴とする請求項1に記載の半導体装置の製造方法。The first polishing tool is provided with a material having a first compression modulus, and the second polishing tool is provided with a material having a compression modulus higher than the first compression modulus. 2. The method for manufacturing a semiconductor device according to claim 1, wherein: 前記第1の研磨工具を研磨パッドとし、前記第2の研磨工具を固定砥粒盤とすることを特徴とする請求項1に記載の半導体装置の製造方法。2. The method according to claim 1, wherein the first polishing tool is a polishing pad, and the second polishing tool is a fixed abrasive disk. 前記第2の導電膜の凸部の研磨速度をR1、凹部の研磨速度をR2とすると、前記第2の研磨工具における前記第2の導電膜の凸部と凹部の研磨速度の比R1/R2が、前記第1の研磨工具における前記第2の導電膜の凸部と凹部の研磨速度の比よりも大きいことを特徴とする請求項1に記載の半導体装置の製造方法。Assuming that the polishing rate of the convex portion of the second conductive film is R1 and the polishing speed of the concave portion is R2, a ratio R1 / R2 of the polishing speed of the convex portion and the concave portion of the second conductive film in the second polishing tool 2. The method of manufacturing a semiconductor device according to claim 1, wherein the ratio is larger than a ratio of a polishing rate of the convex portion to a concave portion of the second conductive film in the first polishing tool. 前記第2の導電膜の凸部と凹部の研磨速度の比R1/R2が、R1/R2>1であることを特徴とする請求項1に記載の半導体装置の製造方法。2. The method according to claim 1, wherein a ratio R1 / R2 of a polishing rate between the convex portion and the concave portion of the second conductive film satisfies R1 / R2> 1. 3. 前記第1の研磨工具と前記第2の研磨工具に供給する研磨液が異なることを特徴とする請求項1に記載の半導体装置の製造方法。2. The method according to claim 1, wherein a polishing liquid supplied to the first polishing tool and a polishing liquid supplied to the second polishing tool are different. 基板上に絶縁膜を形成し、前記絶縁膜に開口部を形成する工程と、
前記開口部と前記絶縁膜上に第1の導電膜を形成する工程と、
前記第1の導電膜上に第2の導電膜を形成する工程と、
前記第1の導電膜と前記第2の導電膜の一部を研磨工具を用いて加工し、前記開口部内に前記第1の導電膜および前記第2の導電膜を形成する工程とを備え、
前記第2の導電膜の加工には第1の研磨工具および第1の研磨液をを使用し、前記第2の導電膜の加工途中で第2の研磨工具および第2の研磨液に切り替え、さらに、前記第1の導電膜が露出する前に前記第2の研磨工具の表面を加工中に目立てする工程と、第3の研磨液に切り替える工程とを含むことを特徴とする半導体装置の製造方法。
Forming an insulating film on the substrate, forming an opening in the insulating film,
Forming a first conductive film over the opening and the insulating film;
Forming a second conductive film on the first conductive film;
Processing a part of the first conductive film and the second conductive film using a polishing tool to form the first conductive film and the second conductive film in the opening;
A first polishing tool and a first polishing liquid are used for processing the second conductive film, and the processing is switched to a second polishing tool and a second polishing liquid during the processing of the second conductive film; The method of manufacturing a semiconductor device, further comprising: a step of dressing the surface of the second polishing tool during processing before the first conductive film is exposed; and a step of switching to a third polishing liquid. Method.
前記第1の研磨工具を研磨パッドとし、前記第2の研磨工具を固定砥粒盤とすることを特徴とする請求項8に記載の半導体装置の製造方法。The method according to claim 8, wherein the first polishing tool is a polishing pad, and the second polishing tool is a fixed abrasive disk. 前記第1および前記第2の研磨液は酸化剤、有機酸、防食剤もしくは界面活性剤、純水で構成されていることを特徴とする請求項1に記載の半導体装置の製造方法。2. The method according to claim 1, wherein the first and second polishing liquids include an oxidizing agent, an organic acid, an anticorrosive or a surfactant, and pure water. 前記第1、前記第2および前記第3の研磨液は界面活性剤の添加量がお互いに異なることを特徴とする請求項1に記載の半導体装置の製造方法。2. The method according to claim 1, wherein the first, second, and third polishing liquids have different amounts of surfactant added thereto. 3. 前記第1および前記第2の研磨液は界面活性剤の分子量が異なることを特徴とする請求項1に記載の半導体装置の製造方法。2. The method according to claim 1, wherein the first and second polishing liquids have different surfactant molecular weights. 3. 前記第1、前記第2および前記第3の研磨液は防食剤の添加量がお互いに異なることを特徴とする請求項1に記載の半導体装置の製造方法。2. The method according to claim 1, wherein the first, second, and third polishing liquids have different amounts of anticorrosive added to each other. 3. 前記第1および前記第2の研磨液は防食剤が異なることを特徴とする請求項1に記載の半導体装置の製造方法。The method according to claim 1, wherein the first and second polishing liquids have different anticorrosive agents. 前記第2および前記第3の研磨液は酸化剤の添加量が異なることを特徴とする請求項1に記載の半導体装置の製造方法。2. The method according to claim 1, wherein the second and third polishing liquids have different amounts of an oxidizing agent. 前記第2および前記第3の研磨液は有機酸の添加量が異なることを特徴とする請求項1に記載の半導体装置の製造方法。2. The method according to claim 1, wherein the second and third polishing liquids have different amounts of organic acid added. 3. 前記第1および前記第2の研磨液の酸化剤として過酸化水素水、有機酸としてリンゴ酸、防食剤としてベンゾトリアゾール又はその誘導体、界面活性剤としてアクリル酸重合体およびそのアンモニア塩、溶媒として純水を含むことを特徴とする半導体装置の製造方法。Hydrogen peroxide water as an oxidizing agent of the first and second polishing liquids, malic acid as an organic acid, benzotriazole or a derivative thereof as an anticorrosive, an acrylic acid polymer and its ammonium salt as a surfactant, and pure solvent as a solvent A method for manufacturing a semiconductor device, comprising water. 前記第1の導電膜はタンタルあるいはタンタル合金もしくはタンタル化合物を含み、前記第2の導電膜は銅あるいは銅を主成分とする合金もしくは銅化合物を含むことを特徴とする請求項1または8に記載の半導体装置の製造方法。9. The method according to claim 1, wherein the first conductive film contains tantalum, a tantalum alloy, or a tantalum compound, and the second conductive film contains copper, an alloy containing copper as a main component, or a copper compound. Manufacturing method of a semiconductor device.
JP2002288356A 2002-10-01 2002-10-01 Manufacturing method of semiconductor device Pending JP2004128112A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2002288356A JP2004128112A (en) 2002-10-01 2002-10-01 Manufacturing method of semiconductor device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2002288356A JP2004128112A (en) 2002-10-01 2002-10-01 Manufacturing method of semiconductor device

Publications (1)

Publication Number Publication Date
JP2004128112A true JP2004128112A (en) 2004-04-22

Family

ID=32280872

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2002288356A Pending JP2004128112A (en) 2002-10-01 2002-10-01 Manufacturing method of semiconductor device

Country Status (1)

Country Link
JP (1) JP2004128112A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007129115A (en) * 2005-11-07 2007-05-24 Fuji Electric Holdings Co Ltd Manufacturing method for semiconductor device
JP2008507855A (en) * 2004-07-26 2008-03-13 インテル・コーポレーション Method and apparatus for conditioning a polishing pad
US20090056102A1 (en) * 2007-08-31 2009-03-05 Fujitsu Microelectronics Limited Method for fabricating semiconductor device
JP2009516928A (en) * 2005-11-22 2009-04-23 キャボット マイクロエレクトロニクス コーポレイション Friction reduction aid for CMP
CN103646866A (en) * 2013-11-29 2014-03-19 上海华力微电子有限公司 A chemical-mechanical polishing apparatus and a chemical-mechanical polishing method
WO2023212981A1 (en) * 2022-05-06 2023-11-09 长鑫存储技术有限公司 Chemical mechanical polish process method and device

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008507855A (en) * 2004-07-26 2008-03-13 インテル・コーポレーション Method and apparatus for conditioning a polishing pad
JP2007129115A (en) * 2005-11-07 2007-05-24 Fuji Electric Holdings Co Ltd Manufacturing method for semiconductor device
JP2009516928A (en) * 2005-11-22 2009-04-23 キャボット マイクロエレクトロニクス コーポレイション Friction reduction aid for CMP
US20090056102A1 (en) * 2007-08-31 2009-03-05 Fujitsu Microelectronics Limited Method for fabricating semiconductor device
US8286344B2 (en) 2007-08-31 2012-10-16 Fujitsu Semiconductor Limited Method for fabricating semiconductor device
US20130012019A1 (en) * 2007-08-31 2013-01-10 Fujitsu Semiconductor Limited Method for fabricating semiconductor device
US8991042B2 (en) 2007-08-31 2015-03-31 Fujitsu Semiconductor Limited Method for fabricating semiconductor device
CN103646866A (en) * 2013-11-29 2014-03-19 上海华力微电子有限公司 A chemical-mechanical polishing apparatus and a chemical-mechanical polishing method
WO2023212981A1 (en) * 2022-05-06 2023-11-09 长鑫存储技术有限公司 Chemical mechanical polish process method and device

Similar Documents

Publication Publication Date Title
EP0946979B1 (en) Method for chemical-mechanical planarization of a substrate on a fixed-abrasive polishing pad
JP4095731B2 (en) Semiconductor device manufacturing method and semiconductor device
US6790768B2 (en) Methods and apparatus for polishing substrates comprising conductive and dielectric materials with reduced topographical defects
US7104869B2 (en) Barrier removal at low polish pressure
Kondo et al. Abrasive‐Free Polishing for Copper Damascene Interconnection
US6368190B1 (en) Electrochemical mechanical planarization apparatus and method
US20040248412A1 (en) Method and composition for fine copper slurry for low dishing in ECMP
US6720265B2 (en) Composition compatible with aluminum planarization and methods therefore
JP2001308040A (en) High throughput copper cmp with reduced erosion and dishing
KR20030078002A (en) Manufacturing method of semiconductor device
EP1147546A1 (en) Method to decrease dishing rate during cmp in metal semiconductor structures
TWI294456B (en)
US6734103B2 (en) Method of polishing a semiconductor device
EP1345734B1 (en) Crosslinked polyethylene polishing pad for chemical-mechnical polishing and polishing apparatus
US7037174B2 (en) Methods for reducing delamination during chemical mechanical polishing
US7121926B2 (en) Methods for planarization of group VIII metal-containing surfaces using a fixed abrasive article
US20020173221A1 (en) Method and apparatus for two-step polishing
JP4719204B2 (en) Chemical mechanical polishing slurry and semiconductor device manufacturing method
JP2004128112A (en) Manufacturing method of semiconductor device
JP2001291684A (en) Chemical-mechanical polishing composition and metal layer polishing method using the same
US20100096360A1 (en) Compositions and methods for barrier layer polishing
US6887131B2 (en) Polishing pad design
JP2006319052A (en) Method of manufacturing semiconductor device, and polishing device
US7141502B1 (en) Slurry-less polishing for removal of excess interconnect material during fabrication of a silicon integrated circuit
JP2005260185A (en) Polishing pad