JP3591282B2 - Planar processing apparatus and planar processing method using the same - Google Patents

Description

【００２１】
ただし、α ：円盤体の線膨張係数（１／℃）
λ ：円盤体の熱伝導度（ｃａｌ／ｍｍ・秒・℃）
Ｄ ：円盤体の直径（ｍｍ）
Ｑ ：円盤体の伝熱密度（ｃａｌ／ｍｍ^２ ）
Ｈ ：円盤体の厚さ（ｍｍ）
λｃ：研磨布の熱伝導度（ｃａｌ／ｍｍ・秒・℃）
Ｈｃ：研磨布の厚さ（ｍｍ）
Ｔｃ：研磨布の表面温度（℃）
λｓ：スケールの熱伝導度（ｃａｌ／ｍｍ・秒・℃）
Ｈｓ：スケールの厚さ（ｍｍ）
Ｔｓ：スケールの表面温度（℃）
ｆ ：境界層の伝熱係数（ｃａｌ／ｍｍ^２ ・秒・℃）
Ｔｗ：冷却水の温度（℃）
なお、ΔＦの符号は円盤体１が観察面側で凸平面となる場合をプラス（＋）とし、Ｑの付号は観察面側から反対面側へ熱が流れる場合をプラス（＋）とする。
【００２２】
式（Ｉ）より、変形した円盤体１の表面の形状は放物面で近似されることがわかる。
ここで、αとλは円盤体１の構成材料に依存し、ＤとＨは円盤体１の寸法そのものであるから、定数である。これに対し、ＴｗとＴｓは加工条件、Ｈｃとλｃは研磨布２の目詰まり状況や構成材料、Ｈｓとλｓはスケール層４の付着状況、ｆは冷却水６の流速によりそれぞれ変動する。したがって、人為的な制御を比較的容易にΔＦに反映できると考えられる項はＱとＴｗであるが、中でも変動要因の無いＱを制御する方がΔＦを精度良く制御できることは明らかである。
【００２３】
以上の考察は、円盤体１が冷却水流路３と接触しているモデルに関するものであるが、冷却水流路３と接触しない円盤体１については、上式（Ｉ）からスケール層４と境界層５と冷却水６に関連する項、すなわちλｓ，Ｈｓ，Ｔｓ，ｆ，Ｔｗを削除すればよく、基本的な考え方は同じである。
また、円盤体１の表面に研摩布２が張設されていない場合には、研摩布２に関連する項、すなわちλｃ，Ｈｃ，Ｔｃを削除すればよい。
さらに、上記円盤体１は、加工盤、被加工体、あるいは被加工体を保持する研磨ヘッドのいずれと考えてもよい。
本発明は、かかる考察にもとづき、すなわち加工盤の作業面の表面プロファイル、被加工体の表面プロファイル、あるいは被加工体を保持する研磨ヘッドの被加工体保持面の表面プロファイルの少なくともいずれかを熱的に制御するのである。この熱的な制御は、上記加工盤、上記被加工体の裏面、もしくは上記研磨ヘッドの少なくともいずれかに設けられた加熱手段への通電量の調節を通じて行うことができる。
【００２４】
図２ないし図４に、これらの制御を実現するための構成を概念的に示す。ここでは、より実用的な構成を考慮して、加熱手段としてのヒータをいずれも断熱材層で被覆している。
図２は、加工盤１１の作業面１１ａとは反対側の面に加熱手段としてのヒータ１２が当接され、このヒータ１２がさらに断熱材層１３で被覆された状態を示している。ヒータ１２が断熱材層１３で被覆されることにより、熱流の方向を該ヒータ１２から作業面１１ａへと規定することができ、加熱効率を高めることができる。
【００２５】
図３は、被加工体１４の被加工面１４ａとは反対側の面にヒータ１２が当接され、このヒータ１２がさらに断熱材層１３で被覆された構成を示している。かかる構成は、被加工体１４が肉厚板体である場合に適しており、たとえば次に述べる被加工体保持部材の被加工体保持面（以下、保持面と称する。）の加工を行う場合に適用することができる。
なお、本明細書における「肉厚板体」とは、仮に熱変形を起こす隣接部材が一方の主面に当接されていても、その変形による自身の表面プロファイル変化が無視し得る程度にしか生じない板体を指すものとする。このような肉厚板体の表面プロファイルを変化させるためには、隣接部材ではなく、肉厚板体自身を直接に加熱する必要がある。
【００２６】
図４は、被加工体１４が被加工体保持部材１５の保持面１５ａ上に保持されており、該保持面１５ａと反対側の面にヒータ１２が当接され、このヒータ１２がさらに断熱材層１３で被覆された構成を示している。かかる構成は、被加工体１４が肉薄板体である場合に適しており、たとえば半導体ウェーハの加工を行う場合に適用することができる。
なお、本明細書における「肉薄板体」とは、熱変形を起こす隣接部材が一方の主面に当接されている場合に、その変形の影響を受けて自身の表面プロファイルが変化するような板体を指すものとする。
【００２７】
上記ヒータ１２は、平面ヒータとすることが好適である。具体的には、たとえばニクロム線等の抵抗発熱体をシリコーンゴム等のフレキシブルな絶縁体で被包したシリコーンラバーヒータを用いることができる。平面ヒータは、面内の発熱密度分布ができるだけ均等なものがよい。また、平面ヒータと電源との間に変圧器を介在させれば、この変圧器で該平面ヒータへの通電量を制御して発熱量を調節することができる。この発熱量の調節により加工盤の作業面、被加工体、もしくは被加工体を保持する研磨ヘッドの被加工体保持部材の表面プロファイルを迅速、容易、かつ可逆的に変更することができる。特に、加工盤や研磨ヘッドに関しては、従来のような加工工具を用いた表面プロファイルの調整作業が不要となり、コストや時間の観点から極めて有利となる。
【００２８】
図５および図６に、前掲の図２の加工盤１１の作業面１１ａの表面プロファイルが加熱により変化し、これに伴って被加工体１４の最終的な加工形状が変化する様子を模式的に示す。ここで、加工盤１１と被加工体１４は、それぞれの中心線の回りに矢印Ａおよび矢印Ｂで示されるごとく回転している。
図５（ａ）はヒータ１２への通電が無く、作業面１１ａが平坦な初期状態、図５（ｂ）は平面ヒータ１２への通電により作業面１１ａが中凹に変形した状態、図５（ｃ）は通電量の増加により作業面１１ａの変形量がさらに大きくなった状態をそれぞれ示す。これに伴って、被加工体１４の表面プロファイルも図５（ａ）の平坦から、図５（ｂ）の中凸、さらに図５（ｃ）の大きな中凸へと変化する。
あるいは、図６（ａ）に示されるように、作業面１１ａの初期状態を中凸としておき、この形状を通電による凹方向の変形と相殺して図６（ｂ）図に示されるように平坦なプロファイルを得ることもできる。これに伴って、被加工体１４の表面プロファイルは図６（ａ）の中凹から図６（ｂ）の平坦へと変化する。
【００２９】
なお、図５および図６には、加工盤１１のみに表面プロファイルの変化が与えられる場合を図示したが、同様の変化を前掲の図３のように被加工体１４に与えても、あるいは前掲の図４のように被加工体保持部材１５の保持面１５ａに与えても、さらにあるいはこれらの変化を組み合わせてもよい。
【００３０】
これらの平面加工装置を用いて所望の被加工体の加工形状を得ようとする場合、まずある条件で平面加工を行い、得られた被加工体の形状を測定する。このときの測定形状と目的形状との差を求め、その差が補正できるような熱変形を加工盤１１、被加工体１４、あるいは被加工体保持部材１５の少なくともいずれかに与えて再び平面加工を行う。
本発明の平面加工方法では、これら一連の工程を所望の形状が得られるまで繰り返すことになるが、熱変形は通電量の調節により自在に起こすことが可能であるため、たとえば従来のごとく専用の加工工具を用いて加工盤や被加工体保持部材を作製し直すといった作業は一切不要であり、コスト的にも時間的にも極めて有利である。
【００３１】
上記加工盤は、作業面を有する上半部と平面ヒータが当接される下半部との接合体より構成されてもよい。さらに、この上半部は、作業面のほぼ全域にわたる部分が互いに弾性材料層を介して隣接されかつ個々に前記下半部と接合される剛性ブロックの集合体であってもよい。これらの構成は、後述する冷却水流通用の空洞の形成し易さにも関係している。
ところで、本発明の平面加工方法では、表面プロファイルの制御に冷却手段による冷却を併用してもよい。冷却を併用すると、前述の加熱手段により発生した熱の一部がこの冷却で奪われることになるため、熱の有効利用の観点からは不利となるが、冷却に関連する制御パラメータが増えるため、より高精度な表面プロファイル制御が可能となる。
【００３２】
かかる冷却手段の一例としては冷媒循環系統が挙げられ、その典型例は冷却水を循環させるものである。
図７に、内部に冷却水を循環させるようになされた加工盤１１の構成例を示す。この加工盤１１は、作業面１１ａを有する上半部１１ｔと、ヒータ１２が当接される下半部１１ｂとが接合一体化されたものであり、両者の接合面の一部に冷却水流通用の空洞１６が形成されている。上記空洞１６はたとえば渦巻き状に形成されており、その両端部には、上記加工盤１１の回転軸である脚部１７の内部に挿通された冷却水配管１８が接続されている。
【００３３】
図８には、加工盤１１の上半部１１ｔを複数の剛性ブロック２０の集合体より構成した例を示す。個々の剛性ブロック２０は下端部が脚部２０ａ、上端部がフランジ部２０ｂとされており、該フランジ部２０ｂの側面においてシール材２１を介して多数が隣接配置されることにより、全体としてひとつの作業面１１ａを構成している。また、上記複数の剛性ブロック２０は、脚部２０ａの部分で個々に下半部１１ｂと接合されている。上記作業面１１ａ内における上記フランジ部２０ｂの平面形状は特に限定されるものではないが、作業面１１ａをできるだけ少ない隙間で充填する観点から、矩形あるいは六角形が好適である。
上記構成においては、隣接する剛性ブロック２０と弾性を有するシール材２１と下半部１１ｂとで囲まれる空洞２２が、冷却水の流路となる。これらの空洞２２は互いに適当な順路で連結されている必要があり、上記脚部２０ａに設けられた貫通孔２０ｃがこの連結の役目を果している。
【００３４】
かかるブロック一体化構成によれば、たとえ平面加工中の摩擦熱により個々の剛性ブロック２０に熱変形が生じても、これが弾性を有するシール材２１で吸収されるため、作業面１１ａ全体としては最初に設定した表面プロファイルをそのまま維持することができる。つまり、下半部１１ｂの熱変形が、妨害されることなく作業面１１ａの表面プロファイルに直接反映されることになる。
また、個々の剛性ブロック２０の隣接部に空洞２２が設けられて冷却総面積を大きく確保しているため、冷却能力も十分である。
なお、上述のブロック一体化構成は必ずしも空洞を伴うものでなくてもよく、たとえば、個々の剛性ブロックが単純な柱状をなすものであってもよい。このような構成であっても、下半部１１ｂｇの熱変形を作業面１１ａの表面プロファイルに忠実に反映させる効果は期待することができる。
【００３５】
ところで、本発明では上述の冷媒循環系統の他に、冷却手段として熱電冷却素子を使用することもできる。熱電冷却素子は、異種金属の接点に電流を流した時に吸熱が生ずるペルチェ効果を利用する素子である。平面ヒータのような加熱手段を用いている場合に、冷却手段としてかかる熱電冷却素子を利用すれば、加熱も冷却も共に電気的に行うことができ、装置構成を単純化することができる。
なお、熱電冷却素子は電流を流す方向を逆とすれば発熱体としても利用することができるので、これを平面ヒータの代わりに加熱手段として用いたり、あるいは加熱・冷却の双方に用いることも可能である。後者の場合には、たとえば熱電冷却素子の電気回路を２系統設け、一方の回路には発熱が生ずる方向、他方の回路には吸熱が生ずる方向に電流を流す構成が可能である。ただし、熱電冷却素子は元来が局所的な冷却／加熱に適する素子であるため、大面積の加熱には平面ヒータを使用した方が熱効率は良い。
【００３６】
本発明では、加工盤の作業面に被加工体を摺接させる際に、該作業面上に研磨剤を供給し、該被加工体を研磨してもよい。研磨剤は、従来公知のものをいずれも使用することができ、典型的にはコロイダルシリカが用いられる。
特に、加工盤の作業面上に研磨布を張設し、かつ該研磨布の表面に研磨剤を供給する研磨剤供給手段を設けることにより、被加工体の化学機械研磨を行う平面加工装置を提供することができる。かかる装置は、被加工体が半導体ウェーハである場合に特に有用である。
【００３７】
【実施例】
以下、本発明の具体的な実施例について説明する。
【００３８】
実施例１
本実施例では、加工盤の裏面と研磨ヘッドの被加工体保持部材の裏面の双方にそれぞれヒータと断熱材層とを備えた化学機械研磨装置の構成例について、図９を参照しながら説明する。
この装置は、作業面１１ａ上に研磨布１９が張設された円盤状の加工盤１１と、被加工体１４としてたとえば半導体ウェーハを保持する研磨ヘッド３５とを平行に対向配置させたものであり、加工盤１１の矢印Ａ方向の回転運動と研磨ヘッド３５の矢印Ｂ方向の回転運動とを組み合わせて、研磨スラリーの存在下で半導体ウェーハの表面を化学機械研磨するようになされたものである。
【００３９】
上記加工盤１１は、厚さ３０ｍｍのＳＵＳ鋼板からなる上半部１１ｔと、同じく厚さ３０ｍｍのＳＵＳ鋼板からなる下半部１１ｂとがボルト等の固定手段を用いて堅固に一体化されてなる直径７００ｍｍの円盤体である。上半部１１ｔと下半部１１ｂとの接合面の一部は空洞１６とされ、冷却水流路とされている。
加工盤１１は矢印Ａで示されるごとく、下半部１１ｂの中心に接続される脚部１７の回りに図示されない駆動手段を用いて回転される。この脚部１７は、上記空洞１６に冷却水を導入するための冷却水配管１８の通路を兼ねており、該冷却水配管１８は該脚部１７の中途部に設けられたロータリージョイント３４を経由して外部へ導出される。上記空洞１６の冷却面積は、作業面１１ａの面積の約
７０％であり、ここに温度２０℃、流量５リットル／分で冷却水を流した。
【００４０】
下半部１１ｂの裏面側にはヒータ１２ａを当接させ、該ヒータ１２ａはさらに断熱材層１３ａで被覆した。 ＾
上記ヒータ１２ａとしては、たとえばニクロム線をシリコンゴム層で被覆し、１００Ｖの印加電圧で０．６Ｗ／ｃｍ^２ の発熱密度が得られるシリコンラバーヒータを複数枚並列に接続して使用した。ヒータ１２の配線は、脚部１７の内部を通り、該脚部１７の中途部に設けられたロータリーコネクタ３１を経由して外部へ導出され、変圧器３２を介して電源３３に接続されている。この変圧器３２を操作することにより、ヒータ１２ａへの通電量、すなわち発熱量を容易かつ迅速に変化させることができる。
上記断熱材層１３としては、一例として厚さ６ｍｍのシリコンラバー発泡体を使用した。
【００４１】
上記作業面１１ａの上に張設される研磨布１９としては、ウレタン不織布を使用した。
上記加工盤１１の上方中央には、研磨スラリーを供給するためのノズル４５が開口されている。研磨スラリーとしては、たとえばコロイダルシリカが研磨布１９の表面へ供給される。上記加工盤１１のエッジ部から溢れた研磨スラリーを回収するためのドレン受け、回収された研磨スラリーを貯蔵するためのタンク、研磨スラリーをノズル４５へ圧送するためのポンプの図示はいずれも省略したが、前掲の図１５を参照しながら説明したとおりである。
【００４２】
一方、上記研磨ヘッド３５は、加圧室３６の底面中央に被加工体保持部材として被加工体保持部材１５と吸着治具４０とを備えたものであり、被加工体１４を真空排気により被加工体保持部材１５上に吸着保持しながら、上記研磨布１９に対しては該被加工体１４を空気加圧により所定の圧力で押圧付勢するものである。
上記加圧室３６の壁は研磨ヘッド３５の骨格をなす部材であり、矢印Ｂで示されるごとく、その天井部中央に接続される脚部３７の回りに図示されない駆動手段を用いて回転される。この脚部３７は、上記加圧室３６に空気を導入するための空気配管３９、後述の吸着治具４０に接続される真空配管４１、後述のヒータ１２ｂの配線の通路を兼ねている。
【００４３】
上記被加工体保持部材１５は、たとえば直径２１０ｍｍ、厚さ３０ｍｍの多孔質アルミナセラミック板であり、その一方の面が被加工体１４を保持するための保持面１５ａ、該保持面１５ａと反対側の面が吸着治具４０の装着面とされる。上記吸着治具４０は、たとえば直径２１０ｍｍ、厚さ２ｍｍのＳＵＳ鋼板であり、空気の吸引路となる溝を多数有している。この溝は、上述の真空配管４１に接続されており、この真空配管４１は上記脚部３７の内部を通り、その中途部に設けられたロータリージョイント３８を経由して外部へ導出され、図示されない真空ポンプに接続されている。
【００４４】
吸着治具４０の背面には、ヒータ１２ｂと断熱材層１３ｂが設けられている。これらは、加工盤１１の下半部１１ｂに装着したものと同じである。上記ヒータ１２ｂの配線もやはり上記脚部３７の内部を通り、その中途部に設けられたロータリーコネクタ４２を経由して外部へ導出され、変圧器４３を介して電源４４に接続されている。この変圧器４３を操作することにより、ヒータ１２ｂへの通電量、すなわち発熱量を容易かつ迅速に変化させることができる。
かかる構成により、本装置は加工盤１１の単独熱変形、被加工体１４の単独熱変形、被加工体保持部材１５の単独熱変形のすべての制御を可能とするものである。
【００４５】
実施例２
ここでは、実施例１で上述した装置の研磨ヘッド３５側のヒータ１２ｂのみを作動させて、該ヒータ１２ｂへの印加電圧を変化させながら実際にシリコンウェーハの化学機械研磨を行った場合の、ウェーハ表面の変形量を調べた。
研磨サンプルとしたシリコンウェーハは、インゴットから切り出された後にラッピングを施された直径２００ｍｍのウェーハである。研磨条件は、一例として研磨圧力＝３００ｇ／ｃｍ^２ 、摺接速度＝４５ｍｍ／分とし、研磨スラリーとしてはコロイダルシリカを用いた。ヒータ１２ｂへの印加電圧は、０〜４０Ｖの範囲で変化させた。研磨されたシリコンウェーハの変形量は、レーザ式形状測定機を用いて非接触式に測定した。
なお、加工盤１１の作業面１１ａの表面プロファイル、およびヒータ１２ｂへ通電していない場合の被加工体保持部材１５の保持面１５ａの表面プロファイルは、いずれも平坦である。
【００４６】
結果を図１０に示す。この図は、被加工体保持部材１５の裏面側のヒータ１２ｂへの印加電圧（Ｖ）に対してウェーハ表面の変形量（μｍ）をプロットしたものである。ここで変形量とは、平面を基準として半導体ウェーハの端部がどの程度上昇したか、あるいは低下したかを示す数値であり、端部が中央部より高ければ中凹、端部が中央部より低ければ中凸、同じなら平坦である。
ヒータ１２ｂへ通電しない場合には、半導体ウェーハは０．２μｍの中凹形状となった。このことは、たとえ加工盤１１の作業面１１ａと被加工体保持部材１５の保持面１５ａの表面プロファイルが共に平坦であっても、半導体ウェーハの表面を必ずしも平坦に加工できないことを示している。しかし、ヒータ１２ｂへの印加電圧を高めるにしたがい、被加工体保持部材１５の保持面１５ａの表面プロファイルが平坦→中凹へと変化し、これに伴って半導体ウェーハの表面プロファイルが中凹→平坦→中凸へと変化することが明らかとなった。図１０によれば、半導体ウェーハの表面プロファイルをちょうど平坦とできるような印加電圧は２０Ｖ近傍にある。ちなみに、この時の被加工体保持部材１５の保持面１５ａの表面プロファイルはわずかな中凹であった。
【００４７】
実施例３
本実施例では、加工盤１１の上半部１１ｔをブロック一体化構成とし、その内部に形成された冷却水流路を併用して精度の高い作業面１１ａの表面プロファイル制御を可能とした化学機械研磨装置を用い、半導体ウェーハを研磨した。
まず、使用した装置の構成を図１１に示す。この装置は、前掲の図８に示した加工盤１１を、前掲の図９に示した化学機械研磨装置に組み込んだものに相当し、符号も共通である。
【００４８】
ここで、上半部１１ｔを構成する個々の剛性ブロック２０は、縦５０ｍｍ、横５０ｍｍ、厚さ３０ｍｍのＳＵＳ鋼板の一方の面を加工することにより、２０ｍｍ四方の四角柱状の脚部２０ａと、５０ｍｍ四方の正方形のフランジ部２０ｂとが形成され、さらに該脚部２０ａに横方向の貫通孔２０ｃが穿設されたものである。この剛性ブロック２０を直径７００ｍｍ、厚さ３０ｍｍのＳＵＳ鋼板からなる下半部１１ｂの上に密に多数隣接して配置し、個々にボルトを用いて固定した。これら多数の剛性ブロック２０により構成された作業面１１ａを適当な加工工具を用いて平坦に仕上げ、さらに、シリコーンゴム系のシール材２１でブロック間の隙間を埋め込んだ。
【００４９】
上記構成によると、隣接する剛性ブロック２０の脚部２０ａの隙間が空洞２２となり、これが冷却水流路として使用されるが、冷却水による冷却面積は、作業面１１ａの面積の約１７０％となった。この空洞２２に接続される冷却水配管１８は、加工盤１１の脚部１７の内部を通って、その中途部に設けられたロータリージョイント３４を経由して外部へ導出されている。
また、下半部１１ｂに当接されるヒータ１２ｂの配線も加工盤１１の脚部１７の内部を通り、その中途部に設けられたロータリーコネクタ３１を経由して外部へ導出され、変圧器３２を介して電源４４に接続されている。
この他の部材については、図８および図１１と符号も共通なので、詳しい説明は省略する。
【００５０】
次に、上述のように構成された上半部１１ｔの表面にウレタン不織布からなる研磨布１９を張設し、研磨ヘッド３５側のヒータ１２ｂと加工盤１１側のヒータ１２ａとを共にオフとし、空洞２２には、２０℃に温調された冷却水を５リットル／分の流速で循環させた。
この状態で、直径２００ｍｍのシリコンウェーハを研磨圧力＝３００ｇ／ｃｍ^２ 、摺接速度＝４５ｍ／分にて連続して研磨した。研磨後のウェーハ全面の厚さ分布を静電容量式平坦度測定器を用いて測定したところ、球面状の中凹形状となり、ウェーハの中央部は端部よりも０．３５μｍ薄かった。
【００５１】
そこで、ウェーハの厚さ分布を均等とするため、上記加工盤１１のヒータ１２ａに１０Ｖの電圧を印加し、直径２００ｍｍのシリコンウェーハの摺接領域において作業面１１ａを凹方向に０．３５μｍ変形させた。なお、この印加電圧は、加工盤１１の線膨張係数α、同じく加工盤１１の熱伝導度Ｑ、およびヒータの特性より理論的に算出した値である。印加電圧以外の条件は同じとして、再びシリコンウェーハを研磨した。研磨後のウェーハの表面を平坦度測定器を用いて測定したところ、中凹形状や中凸形状は発生せず、高度に平坦化されていることが確認され、本発明の効果が実証された。
【００５２】
実施例４
本実施例では、加工盤の裏面と研磨ヘッドの被加工体保持部材の裏面の双方にそれぞれヒータと断熱材層とを備えたラップ機の構成例について、図１２を参照しながら説明する。
この機械は、太陽ギヤ５１と内周ギヤ５４との双方に噛合されることにより遊星運動を行うキャリア６４と、上記太陽ギヤ５１と同軸的に配され独自の回転運動を行う加工盤５７とを近接対向させ、上記キャリア６４に装着された被加工体１４を上記定盤５７の表面に摺接させることにより、該被加工体１４のラッピングを行うようになされたものである。
【００５３】
上記太陽ギヤ５１は、モータ５３に駆動される回転軸５２の上端部に一体的に取り付けられている。上記内周ギヤ５４は、上記加工盤５７の外周を包囲する枠体５５の上端部に形成されており、この枠体５５は上記モータ５３とは独立に制御されるモータ５６により回転駆動される。
加工盤５７は、直径１４００ｍｍ、厚さ７０ｍｍの球状黒鉛鋳鉄からなる円盤体であり、その裏面側中央には脚部５８が上記回転軸５２を包囲するごとく接続され、該脚部５８の末端にてモータ５３により独立に回転駆動されるようになされている。また、加工盤５７の表面には余分な研磨スラリーを周縁部へ誘導するための多数の溝５９が刻設されている。研磨スラリーの循環再利用系統については従来と同じなので、図示は省略する。
さらに、加工盤５７の裏面にはヒータ１２ａを当接させ、該ヒータ１２ａはさらに断熱材層１３ａで被覆した。これらヒータ１２ａと断熱材層１３ａは、前述の実施例１や実施例３で使用したものと同じである。ヒータ１２ａからの配線は、加工盤５７の脚部５８の内部を通過し、その中途部に設けられたロータリーコネクタ６１を経由して外部へ導出され、変圧器６２を経て電源６３に接続されている。
【００５４】
被加工体１４は、キャリア６４の底面の開口６５に嵌め込まれて保持されている。被加工体１４の背面側、すなわち被加工面と反対側の面にはヒータ１２ｂが当接され、このヒータ１２ｂはさらに断熱材層１３ｂで被覆されている。断熱剤層１３ｂの裏面側には支持部材６６が配され、この内部に上記ヒータ１２ｂからの配線を通過させ、ロータリーコネクタ６７を経由して外部へ導出させるようになされている。外部へ導出された配線は、変圧器６８を経て電源６９に接続される。
かかる構成により、本ラップ機は加工盤５７の単独熱変形、被加工体１４の単独熱変形、あるいは加工盤５７と被加工体１４の同時熱変形の３種類の制御を可能とするものである。
【００５５】
実施例５
ここでは、実施例４で上述したラップ機の加工盤５７側のヒータ１２ａのみを作動させて、該ヒータ１２ａへの印加電圧を変化させながら加工盤５７の変形に関する基礎的な検討を行った。
ここでは、ヒータ１２ａへの印加電圧０〜１０Ｖの範囲で変化させた場合の加工盤５７の変形量を、レーザ式形状測定機を用いて非接触式に測定した。なお、ヒータ１２ａに通電していない場合の加工盤５７の作業面の初期表面プロファイルは、直径１４００ｍｍの加工盤５７全体で１０μｍの中凸である。
【００５６】
結果を図１３に示す。この図は、加工盤５７の裏面側のヒータ１２ａへの印加電圧（Ｖ）に対して加工盤５７の作業面の変形量（μｍ）をプロットしたものである。ここで、変形量とは、平面を基準として加工盤５７の端部がどの程度上昇したか、あるいは低下したかを示す数値であり、端部が中央部より高ければ中凹、端部が中央部より低ければ中凸、同じなら平坦である。
図より、ヒータ１２ａへの印加電圧を高めるにしたがい、加工盤５７の表面プロファイルが中凸→平坦→中凹へと変化することが明らかとなった。この変化の傾向は、前述の理論式（Ｉ）に示した２次関数によく一致していることが確認された。
【００５７】
次に、上述のラップ機を用いて実際に加工を行った。
被加工体１４は、直径６００ｍｍ、厚さ２０ｍｍのアルミナセラミックからなる肉厚板体であり、その片方の表面プロファイルを５μｍの凸球面状に仕上げることを目標とする。
研磨スラリーＬとしては、ＳｉＣ（シリコンカーバイド）砥粒を防錆剤、界面活性剤等の添加剤と共に水に分散させたものを用いた。
研磨荷重は、被加工体１４、ヒータ１２ｂ、断熱剤層１３ｂ、支持部材６６およびその付属品の合計自重のみとした。
【００５８】
加工盤５７の裏面側のヒータ１２ａへの印加電圧は、該加工盤５７の表面が被加工体１４の目標形状と同一曲率を持つ凹球面状となるように、図１３から求めた。すなわち、一般に円盤体の変形量は直径の二乗に比例するので、直径６００ｍｍの被加工体１４が５μｍ変形した時の曲率は、直径１４００ｍｍの加工盤５７が２７〔＝５×（１４００／６００）^２ 〕μｍだけ変形した時の曲率に等しい。しかも、被加工体１４の目標プロファイルが凸球面状であれば、これに対応する加工盤５７の表面プロファイルは凹球面状であるから、変形量の符号はマイナスである。図１３より、加工盤５７の変形量が−２７μｍとなる時の印加電圧は、１０Ｖである。
【００５９】
以上の条件により被加工体１４をラッピングしたところ、表面プロファイルは７μｍの凸球面状となり、目標を２μｍ上回っていた。この差異の原因としては、加工盤５７の表面プロファイル以外に、溝５９に沿って流れる研磨スラリーＬやその他の要因による蓄熱効果が考えられた。したがって、実際に必要な加工盤５７の変形量は、被加工体１４の目標変形量を便宜的に３（＝５−２）μｍと考えて算出できることになり、−１６μｍ〔＝−３×（１４００／６００）^２ 〕である。この変形量は、図１３より、印加電圧８Ｖで達成できる。
そこで、印加電圧を８Ｖに低下させた他は同じ条件で再度ラッピングを行ったところ、今度は被加工体１４について目標どおり５μｍの凸球面状の表面プロファイルを得ることができた。
このように、本発明では熱的制御により加工盤の表面プロファイルを極めて容易に変更することができ、従来のように加工工具を用いた煩雑な調整作業を繰り返す必要が全くない。
【００６０】
以上、本発明を５例の実施例について説明したが、本発明はこれらの実施例に何ら限定されるものではない。
たとえば、上述の実施例では加工盤側にも、また研磨ヘッドやキャリアの側にもヒータと断熱材層を備え、いずれか一方の側のヒータに通電した場合の加工例について述べたが、もちろん双方のヒータに通電して目的とする被加工体の加工形状に最適な表面プロファイルを得るようにしてもよい。さらに、ヒータは前掲の図９，図１１，図１２のように必ずしも加工盤側と研磨ヘッドまたはキャリア側の双方に備えられている必要はなく、いずれか一方に備えられていてもよい。本発明の平面加工装置で一度に加工できる被加工体の数は１枚に限られず、研磨ヘッドやキャリアに複数枚の被加工体を装着できるものであってもよい。
【００６１】
さらに、本発明の平面加工装置は、必ずしも上述のような相対的に大型の加工盤と小型の研磨ヘッドとを組み合わせたものである必要はない。たとえば、２枚の大型の加工盤の間に被加工体を保持して両面研磨を行うような装置についても、これらの加工盤の少なくとも一方にヒータと断熱材層を設置し、同様の表面プロファイルの最適化を行うことができる。
この他、装置構成、各部の寸法や形状や仕様、被加工体の種類、加工条件については、適宜選択、変更、組合せが可能である。
【００６２】
【発明の効果】
以上の説明からも明らかなように、本発明の平面加工方法および平面加工装置では加工盤の作業面、被加工体、または被加工体を保持する研磨ヘッドの被加工体保持部材の表面プロファイルの少なくともいずれかを熱的に制御するので、従来のように被加工体の加工結果に応じてたとえば加工盤の表面プロファイルを専用の加工工具を用いて創り直す等の手間や時間を省くことができ、加工精度と生産性の向上に大きく寄与するものである。
上記の熱的な制御は、加熱手段への通電量の調節を通じて迅速かつ簡便に行うことができる。被加工体については、それが肉厚板体であれば直接に、肉薄板体であれば被加工体保持部材を通じて間接に加熱を行ってその表面プロファイルを変化させることができる。薄い被加工体としては、半導体ウェーハを用いるプロセスが実用価値の高いものである。
【００６３】
冷却手段による冷却を併用した場合には、制御パラメータを増やしてより精密な制御を行うことが可能となる。装置構成上は、加工盤を上半部と下半部との接合一体化構成とし、これらの接合面の一部に空洞を設けてこれを冷却水流路として使用するものが好適である。
また、上半部を互いに弾性材料層を介して隣接される多数の剛性ブロックの集合体とし、かつこれらの剛性ブロックを個別に下半部に固定しておけば、加工時に発生する摩擦熱による上半部の熱変形を最小限に抑え、下半部に意図的に与えられた熱変形のみをほぼ上半部に反映させることができ、制御の精度を向上させる上で有効である。また、ブロック間の隙間を冷却水流路として利用すれば、冷却面積を増大させることも可能である。
被加工体を作業面に摺接させる際に研磨剤を供給すれば、研磨効率を高めることができる。被加工体と研磨剤の組み合わせによっては、メカノケミカル作用による化学機械研磨も可能となり、半導体ウェーハの研磨を行う場合に特に好適である。
【００６４】
本発明は、被加工体について中凸、中凹といった所望の表面プロファイルを得る加工はもちろん、半導体ウェーハの研磨のように高度な平坦性が要求される平面加工にも、既存の製造設備や作業手順を大きく変更することなく対応することが可能であり、その産業上の価値は極めて高いものである。
【図面の簡単な説明】
【図１】円盤体の熱変形モデルを示す模式的断面図である。
【図２】加工盤の作業面とは反対側の面にヒータと断熱材層を配した構成を示す模式的断面図である。
【図３】被加工体の被加工面とは反対側の面にヒータと断熱材層を配した構成を示す模式的断面図である。
【図４】被加工体を保持する被加工体保持部材の裏面にヒータと断熱材層を配した構成を示す模式的断面図である。
【図５】初期状態が平坦な加工盤の作業面の表面プロファイルの変化に伴う被加工体の加工形状の変化を示す模式的断面図であり、（ａ）はヒータの発熱がない場合、（ｂ）はヒータの発熱が小さい場合、（ｃ）はヒータの発熱が大きい場合をそれぞれ表す。
【図６】初期状態が中凸である加工盤の作業面の表面プロファイルの変化に伴う被加工体の加工形状の変化を示す模式的断面図であり、（ａ）はヒータの発熱がない場合、（ｂ）はヒータの発熱がある場合をそれぞれ表す。
【図７】上半部と下半部との接合面の一部に冷却水流通用の空洞を備える加工盤の構成例を示す模式的断面図である。
【図８】上半部が複数の剛性ブロックの集合体より構成される加工盤の構成例を示す模式的断面図である。
【図９】本発明の平面加工装置の一構成例として、加工盤の裏面と被加工体保持部材の裏面にヒータと断熱材層とがそれぞれ配された化学機械研磨装置を示す模式的断面図である。
【図１０】図９の化学機械研磨装置を用いたシリコンウェーハの研磨において、被加工体保持部材の保持板裏のヒータへの印加電圧と、ウェーハの表面の変形量との関係を示すグラフである。
【図１１】本発明の平面加工装置の他の構成例として、加工盤の上半部が複数の剛性ブロックの集合体からなる化学機械研磨装置を示す模式的断面図である。
【図１２】本発明の平面加工装置のさらに他の構成例として、加工盤の裏面と被加工体の裏面にヒータと断熱材層とがそれぞれ配されたラップ機を示す模式的断面図である。
【図１３】図１２のラップ機の加工盤裏のヒータへの印加電圧と、加工盤の変形量との関係を示すグラフである。
【図１４】従来の典型的なラップ機の構成例を示す図であり、（ａ）は上面図、（ｂ）はそのＸ−Ｘ線断面図である。
【図１５】従来の典型的な化学機械研磨装置の構成例を示す模式的断面図である。
【符号の説明】
１１，５７ 加工盤
１１ａ 作業面
１１ｔ 上半部
１１ｂ 下半部
１２，１２ａ，１２ｂ ヒータ
１３，１３ａ，１３ｂ 断熱材層
１４ 被加工体
１５ 被加工体保持部材
１５ａ 保持面
１６，２２ 空洞
１８ 冷却水配管
１９ 研磨布
２０ 剛性ブロック
２１ シール材
３１，４２，６１，６７ ロータリーコネクタ
３２，４３，６０，６８ 変圧器
３３，４４，６３，６９ 電源
３４，３８ ロータリージョイント
３５ 研磨ヘッド
６４ キャリア[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention makes it possible to quickly and easily set and maintain optimum processing conditions in accordance with a target processing shape of a workpiece and its temporal variation.Planar processing apparatus and planar processing method using the apparatusAbout.
[0002]
[Prior art]
Many of the mechanical parts have a certain plane, and plane processing for processing this plane with high accuracy is one of the basic techniques of machining.
A typical device for performing planar processing is a lapping machine. This is a machine that performs more accurate polishing on a workpiece that has been subjected to rough processing such as lathe processing and grinding processing.
FIG. 14 shows a typical configuration example of a lapping machine. FIG. 14A is a top view, and FIG. 14B is a sectional view taken along line XX. This machine includes a carrier 95 that performs planetary motion by being engaged with both the sun gear 81 and the inner peripheral gear 84, and a surface plate 87 that is arranged coaxially with the sun gear 81 and performs its own rotational motion. The workpiece W (workpiece) mounted on the carrier 95 is slid in contact with the surface of the surface plate 87 so that the workpiece W is wrapped.
[0003]
The sun gear 81 is integrally attached to an upper end of a rotating shaft 82 driven to rotate by a motor 83.
The inner peripheral gear 84 is formed at the upper end of a frame 85 surrounding the outer periphery of the surface plate 87, and the frame 85 is driven to rotate by a motor 86 controlled independently of the motor 83. .
The platen 87 is disposed such that its legs 88 surround the rotation shaft 82, and is independently rotated and driven by a motor 90 at the ends of the legs 88.
The workpiece W is held so as to be fitted into the opening 96 of the carrier 95.
When the workpiece W is brought into sliding contact with the surface of the surface plate 87, that is, the work surface, the polishing slurry L is supplied from a nozzle 94 opened above the work surface to increase polishing efficiency and remove frictional heat. It has been done. A large number of grooves 89 are formed on the surface of the surface plate 87 to guide the excess polishing slurry L to the peripheral edge. The polishing slurry L falling from the periphery of the surface plate 87 is collected by a drain receiver 91, stored in a slurry tank 92, pumped by a pump 93, and circulated and reused.
[0004]
On the other hand, there is a chemical mechanical polishing (CMP) apparatus as an apparatus for performing finishing processing such as mirror polishing with higher accuracy than the above lapping.
FIG. 15 shows a typical configuration example of a chemical mechanical polishing apparatus. In this apparatus, a substantially disk-shaped surface plate 101 having a polishing cloth stretched on its surface and a polishing head 110 holding a workpiece W are arranged in parallel and opposed to each other. The surface of the workpiece W is polished by combining the rotation of the polishing head 110 with the rotational movement of the polishing head 110. The polishing mechanism is based on a complicated mechanochemical action in which a mechanical process and a chemical process using the polishing slurry L are combined.
[0005]
The platen 101 has an upper half portion 101t and a lower half portion 101b firmly integrated by using fixing means such as bolts. A cavity 102 is provided to serve as a cooling water channel. By circulating the cooling water inside the cavity 102, deformation of the platen 101 due to frictional heat during polishing is prevented, and thereby the flatness of the workpiece W is improved. The contact area between the upper half portion 101t and the lower half portion 101b, in other words, the cooling area of the surface plate 101 by the cooling water is about half the area of the work surface of the surface plate 101, and the deformation of the surface plate 101 due to the polishing load. It is devised to suppress as much as possible.
The center of the bottom surface of the lower half 101b is connected to the rotating shaft 103, and is rotated at a constant speed by a driving unit (not shown).
[0006]
The polishing cloth 105 stretched on the upper surface of the surface plate 101 is a viscoelastic body having appropriate stickiness and elasticity, and a urethane nonwoven fabric is typically used.
At the center of the upper surface of the polishing cloth 105, for example, colloidal silica is pumped as a polishing slurry L by a pump 108 and supplied from an upper nozzle 109. Further, a drain receiver 106 for collecting the polishing slurry L overflowing from the edge of the surface plate 101 is provided on the outer peripheral portion of the surface plate 101. The polishing slurry L collected through the drain receiver 106 is once stored in a slurry tank 107, and is then circulated and reused by being pumped again by a pump 108.
[0007]
On the other hand, the head 110 has a workpiece holding plate 112 on its lower surface, and holds one or more workpieces W on the wafer holding plate 112 by a method such as vacuum suction or wax bonding. Has been done. The polishing head 110 is also connected to the rotating shaft 113, is driven to rotate at a predetermined rotation speed by a drive mechanism (not shown) attached to the rotating shaft 113, and is pressed against the polishing cloth 105 by a predetermined force. Have been.
The workpiece W is, for example, a silicon wafer sliced from a silicon ingot and subjected to lapping, or a substrate on which an insulating film or a conductive film having surface irregularities is formed. Polishing of a silicon wafer is mirror polishing performed at the final stage of wafer finishing. Flattening of an insulating film by polishing is an indispensable step for improving the reliability of forming a circuit pattern in a multilayer wiring process. Polishing of a conductive film is a useful process for filling a connection hole with a high aspect ratio opened in an insulating film with a metal plug.
[0008]
When the workpiece W is polished using the chemical mechanical polishing apparatus having such a configuration, the workpiece W is held between the workpiece holding plate 112 and the platen 101 which are parallel to each other and has elasticity. Is pressed by the abrasive cloth. For this reason, the region where the plate thickness of the workpiece W is large is removed preferentially by receiving a higher contact pressure than the region where the plate thickness is small, and the planarization of the workpiece W proceeds.
[0009]
[Problems to be solved by the invention]
By the way, the plane of the workpiece W obtained by the above-described lapping machine is not limited to a highly flat surface, and a shape in which the central portion is slightly convex (hereinafter, referred to as a middle convex shape) depending on the use, or the like. The central portion may be slightly concave (hereinafter, referred to as a middle concave shape). Therefore, the surface profile of the surface plate 87 also needs to be formed in a shape corresponding to this. That is, when the workpiece W is required to have a convex shape, a concave surface profile is required, and when the workpiece W is required to have a concave shape, a convex surface profile is required. Will be.
[0010]
These surface profiles of the surface plate are produced using special processing tools while adjusting the operating conditions of the lapping machine. Whether or not a desired processing plane can be obtained by using the surface plate 87 is confirmed by actually processing the workpiece W. If the desired processing surface cannot be obtained, the surface profile of the surface plate 87 is recreated. Repeat for adjustment. For this reason, the adjustment work becomes extremely complicated. In addition, since the surface profile of the surface plate 87 is deformed due to abrasion even when the processing of the workpiece W is repeated, it is very disadvantageous in terms of cost and time to perform the adjustment each time.
[0011]
The situation is more complicated when the workpiece is a semiconductor wafer. In the case of a semiconductor wafer, it is necessary to perform processes such as lithography, etching, and thin film formation with high precision in accordance with a fine design rule from submicron to quarter micron, and even less.
In a recent plasma apparatus or reduction projection exposure apparatus compatible with microfabrication, since the process is performed in a state where the wafer is vacuum-adsorbed to the stage, a slight warp or distortion is corrected. Therefore, the flatness, which is a problem in the present specification, indicates a degree of variation in the thickness of the wafer.
[0012]
The mirror polishing of a semiconductor wafer is performed based on a mechanochemical action using a chemical mechanical polishing apparatus as described above. However, such polishing has many factors that affect flatness. For example, the surface roughness of the polishing pad 105, clogging of the polishing pad 105 by the material removed by polishing, the thickness and uniformity of the polishing pad 105, the viscoelasticity of the polishing pad 105, the workpiece W and the polishing pad 105 The surface temperature of the polishing cloth 105 and the degree of the polishing slurry L reaching the surface of the workpiece are closely related to the frictional heat. However, it is difficult to say that the degree of influence of all these factors on flatness has been elucidated and properly controlled.
When a silicon wafer is polished, for example, using the above-described conventional apparatus, the silicon wafer generally assumes a shape that is spherically concave toward the center, and the height difference between the periphery and the center is about 200 mm (8 inches) in diameter. Fluctuates over time between about 0.1 and 0.6 μm.
[0013]
One of the major causes of the occurrence of the concave shape is the influence of thermal deformation of the surface plate 101. The heat for deforming the platen 101 includes frictional heat generated by sliding between the workpiece W and the polishing cloth 105 and frictional heat generated by a polishing load applied to the polishing cloth 105 from the workpiece W. .
Part of the frictional heat generated by the rubbing between the workpiece W and the polishing pad 105 is part of the surface of the polishing pad 105 → the main body of the polishing pad 105 → the surface of the platen 101 → the back surface of the platen 101 → the surface of the platen 101. The scale is adhered to the cavity 102 constituting the back surface → the boundary layer of the cooling water → the main body of the cooling water, and is removed along the route. This heat value varies depending on factors related to the type of the polishing cloth 105 such as surface roughness, dynamic friction coefficient, and clogging state, or process conditions such as polishing time, polishing pressure, and polishing rate.
On the other hand, a part of the frictional heat generated by the polishing load applied to the polishing pad 105 from the workpiece W is unevenly distributed on the polishing pad 105. In addition, the thermal conductivity of the polishing pad 105, which serves as a conduction path of the frictional heat to the cooling water, changes depending on the clogging condition, and the thickness of the scale attached to the inner wall of the cavity 102 and the thermal conductivity of the boundary layer are reduced. It fluctuates depending on the flow rate of water. Due to these factors, the surface of the platen 101 actually has an extremely complicated shape.
[0014]
Further, since the upper half portion 101t and the lower half portion 101b of the base plate 101 are firmly connected mechanically, the respective thermal deformations affect each other. Further, since the cooling area of the upper half 101t of the surface plate 101 by the cooling water is as small as about half of the total area of the working surface, the ability to remove frictional heat is limited. These factors combine to increase the amount of thermal deformation of the upper half portion 101t of the platen 101, promote a rise in the surface temperature and uneven distribution of the polishing pad 105, and finally deteriorate the flatness of the workpiece W. Has been invited.
[0015]
Therefore, the present invention provides a method for obtaining a desired shape of a workpiece even when a desired planar processing shape of the workpiece is diverse or when the shape of the workpiece changes over time due to deformation of the apparatus itself. It is an object of the present invention to provide a plane processing apparatus which can immediately correct the surface profile of the apparatus required for the above, and a method capable of always performing stable plane processing using the apparatus.
[0016]
[Means for Solving the Problems]
The planar processing apparatus of the present invention (see FIG. 9) faces a disk-shaped processing board 11 having a polishing cloth 19 stretched on a work surface 11a and a polishing head 35 holding a workpiece 14 in parallel. In a planar processing apparatus that arranges and rotates the processing board 11 and the polishing head 35 in directions opposite to each other and chemically and mechanically polishes the surface of the workpiece 14 in the presence of the polishing slurry,
The processing board 11 is a disc body in which an upper half 11t and a lower half 11b, each of which is a rigid plate, are integrated, and a part of a joint surface between the upper half 11t and the lower half 11b is cooled. A first planar heater 12a is provided on the back side of the lower half 11b, and a heat insulating material layer 13a is provided on the back side of the heater 12a. The planar heater 12a is connected to a power supply 33 via a transformer 32, and the amount of heat generation can be controlled by operating the transformer 32.
The polishing head 35 includes a workpiece holding member 15 and a suction jig 40 as a rigid plate at the center of the bottom surface of the pressurizing chamber 36, and vacuum-adsorbs the workpiece 14 onto the workpiece holding member 15. While holding, the workpiece 14 is pressed against the polishing cloth 19 by air pressure.
The workpiece holding member 15 has one surface serving as a holding surface 15a for holding the workpiece 14 and the other surface serving as a mounting surface of the suction jig 40,
The suction jig 40 has a number of grooves connected to the vacuum pipe 41 and serving as an air suction path. A second planar heater 12b is provided on the back surface of the suction jig 40, and a back surface of the heater 12b. Is provided with a heat insulating material layer 13b, and the second planar heater 12b is connected to a power supply 44 via a transformer 43 so that a calorific value can be controlled by operating the transformer 43,
The planar processing apparatus is characterized in that all of the thermal deformation of the processing board 11, the thermal deformation of the workpiece 14, and the thermal deformation of the workpiece holding member 15 can be controlled by the configuration. .
[0017]
A planar processing method according to the present invention is a planar processing method (chemical mechanical polishing method) of a semiconductor wafer (the workpiece 14) by the planar processing apparatus according to claim 1, wherein the first processing is performed during the planar processing. At least one of a calorific value of the planar heater 12a, a calorific value of the second planar heater 12b, a temperature of cooling water supplied to the cavity 16 formed in the processing board 11, and / or a circulation amount. Controlling at least one of the surface profile of the work surface of the processing board 11, the surface profile of the semiconductor wafer, and the surface profile of the workpiece holding member 15 of the polishing head 35. This is a planar processing method characterized by performing the following.
[0018]
BEST MODE FOR CARRYING OUT THE INVENTION
In order to propose the present invention, the present inventors have made basic considerations on thermal deformation of a disk.
The model used for this consideration is shown in FIG. Here, a polishing cloth 2 is stretched on the front surface of the disk 1, and a cooling water channel 3 is formed on the back surface. On this back surface, a scale layer 4 is adhered and formed. On the surface of the scale layer 4, there is a thin boundary layer 5 which is a transition region from laminar flow to turbulent flow, and the outside of the boundary layer 5 is free. There is cooling water 6 flowing in the air.
[0019]
Now, assuming that the disk 1 is uniformly heated from the observation surface side (here, the polishing cloth 2 side), the thermal deformation ΔF from the reference surface S is expressed by the following equation (I).
[0020]
(Equation 1)

[0021]
Where α is the coefficient of linear expansion of the disk (1 / ° C)
λ: thermal conductivity of the disc (cal / mm · sec · ° C)
D: diameter of disk (mm)
Q: Heat transfer density of the disk (cal / mm²)
H: Thickness of disk (mm)
λc: Thermal conductivity of the polishing cloth (cal / mm · sec · ° C)
Hc: thickness of polishing cloth (mm)
Tc: Surface temperature of polishing cloth (° C)
λs: thermal conductivity of the scale (cal / mm · sec · ° C)
Hs: scale thickness (mm)
Ts: Scale surface temperature (° C)
f: heat transfer coefficient of the boundary layer (cal / mm²・ Seconds ・ ℃)
Tw: Temperature of cooling water (° C)
The sign of ΔF is plus (+) when the disc 1 is convex on the observation surface side, and the sign of Q is plus (+) when heat flows from the observation surface side to the opposite surface side. .
[0022]
From the formula (I), it can be seen that the shape of the surface of the deformed disk 1 is approximated by a paraboloid.
Here, α and λ depend on the constituent material of the disk 1, and D and H are constants because they are the dimensions of the disk 1 themselves. On the other hand, Tw and Ts vary depending on the processing conditions, Hc and λc vary depending on the clogging state and constituent materials of the polishing pad 2, Hs and λs vary depending on the adhesion state of the scale layer 4, and f varies depending on the flow rate of the cooling water 6. Therefore, the terms that are considered to be able to reflect artificial control on ΔF relatively easily are Q and Tw. However, it is clear that ΔF can be controlled more accurately by controlling Q having no fluctuation factors.
[0023]
The above considerations relate to the model in which the disk 1 is in contact with the cooling water flow path 3. However, for the disk 1 not in contact with the cooling water flow path 3, the scale layer 4 and the boundary layer The terms related to 5 and the cooling water 6, that is, λs, Hs, Ts, f, and Tw, may be deleted, and the basic concept is the same.
When the polishing cloth 2 is not stretched on the surface of the disk 1, the terms related to the polishing cloth 2, that is, λc, Hc, and Tc may be deleted.
Further, the disk body 1 may be considered as any of a processing board, a workpiece, and a polishing head for holding the workpiece.
The present invention is based on such considerations, that is, at least one of the surface profile of the work surface of the processing board, the surface profile of the workpiece, and the surface profile of the workpiece holding surface of the polishing head holding the workpiece is heated. It controls it. This thermal control can be performed by adjusting the amount of electricity to a heating means provided on at least one of the processing board, the back surface of the workpiece, and the polishing head.
[0024]
2 to 4 conceptually show a configuration for realizing these controls. Here, in consideration of a more practical configuration, each of the heaters as the heating means is covered with a heat insulating material layer.
FIG. 2 shows a state in which a heater 12 as a heating means is brought into contact with a surface of the processing board 11 opposite to the work surface 11 a, and the heater 12 is further covered with a heat insulating material layer 13. By covering the heater 12 with the heat insulating material layer 13, the direction of the heat flow can be defined from the heater 12 to the work surface 11a, and the heating efficiency can be increased.
[0025]
FIG. 3 shows a configuration in which the heater 12 is in contact with the surface of the workpiece 14 opposite to the processing surface 14 a, and the heater 12 is further covered with a heat insulating material layer 13. Such a configuration is suitable when the workpiece 14 is a thick plate body. For example, when processing a workpiece holding surface (hereinafter, referred to as a holding surface) of a workpiece holding member described below. Can be applied to
In this specification, the term "thick plate" means that even if an adjacent member that causes thermal deformation is in contact with one main surface, the change in its own surface profile due to the deformation is negligible. It refers to a plate that does not occur. In order to change the surface profile of such a thick plate, it is necessary to directly heat the thick plate itself, not the adjacent members.
[0026]
FIG. 4 shows that a workpiece 14 is held on a holding surface 15a of a workpiece holding member 15, and a heater 12 is brought into contact with a surface opposite to the holding surface 15a. The configuration covered with the layer 13 is shown. Such a configuration is suitable when the workpiece 14 is a thin plate, and can be applied, for example, when processing a semiconductor wafer.
In the present specification, “thin plate body” means that when an adjacent member that causes thermal deformation is in contact with one main surface, its own surface profile changes under the influence of the deformation. It refers to a plate.
[0027]
The heater 12 is preferably a flat heater. Specifically, for example, a silicone rubber heater in which a resistance heating element such as a nichrome wire is covered with a flexible insulator such as silicone rubber can be used. It is preferable that the flat heater has as uniform a heat generation density distribution as possible. Further, if a transformer is interposed between the flat heater and the power supply, the amount of heat can be adjusted by controlling the amount of current supplied to the flat heater by the transformer. By adjusting the calorific value, the surface profile of the working surface of the processing board, the workpiece, or the workpiece holding member of the polishing head that holds the workpiece can be changed quickly, easily, and reversibly. In particular, with regard to a processing board and a polishing head, the work of adjusting the surface profile using a conventional processing tool is unnecessary, which is extremely advantageous from the viewpoint of cost and time.
[0028]
FIGS. 5 and 6 schematically show how the surface profile of the work surface 11a of the processing board 11 shown in FIG. 2 changes due to heating, and the final processing shape of the workpiece 14 changes accordingly. Show. Here, the processing board 11 and the workpiece 14 are rotating around their respective center lines as indicated by arrows A and B.
FIG. 5A shows an initial state in which the heater 12 is not energized and the work surface 11a is flat, and FIG. 5B shows a state in which the work surface 11a is deformed into a concave by energizing the flat heater 12. FIG. (c) shows a state in which the amount of deformation of the work surface 11a is further increased due to an increase in the amount of electricity. Along with this, the surface profile of the workpiece 14 also changes from flat in FIG. 5 (a) to convex in FIG. 5 (b) and further to large convex in FIG. 5 (c).
Alternatively, as shown in FIG. 6 (a), the initial state of the work surface 11a is set to be convex in the middle, and this shape is canceled out by the deformation in the concave direction due to energization and flattened as shown in FIG. 6 (b). Profile can be obtained. Along with this, the surface profile of the workpiece 14 changes from the middle concave in FIG. 6A to the flat in FIG. 6B.
[0029]
FIGS. 5 and 6 show the case where the surface profile is changed only on the processing board 11, but the same change may be applied to the workpiece 14 as shown in FIG. 4 may be applied to the holding surface 15a of the workpiece holding member 15, or these changes may be combined.
[0030]
When trying to obtain a desired processed shape of a workpiece using these planar processing devices, first, planar processing is performed under certain conditions, and the shape of the obtained workpiece is measured. At this time, a difference between the measured shape and the target shape is obtained, and thermal deformation is applied to at least one of the processing board 11, the workpiece 14, or the workpiece holding member 15 so that the difference can be corrected. I do.
In the plane processing method of the present invention, these series of steps are repeated until a desired shape is obtained. However, since thermal deformation can be freely caused by adjusting the amount of current, for example, a special There is no need for any operation of re-fabricating the processing board or the workpiece holding member using the processing tool, which is extremely advantageous in terms of cost and time.
[0031]
The processing board may be constituted by a joined body of an upper half having a work surface and a lower half to which the flat heater is brought into contact. Further, the upper half may be an aggregate of rigid blocks in which portions substantially all over the working surface are adjacent to each other via the elastic material layer and are individually joined to the lower half. These configurations also relate to the ease of forming a cavity for cooling water circulation described below.
By the way, in the planar processing method of the present invention, the cooling by the cooling means may be used together with the control of the surface profile. If cooling is used in combination, a part of the heat generated by the above-mentioned heating means will be taken away by this cooling, which is disadvantageous from the viewpoint of effective use of heat.However, since control parameters related to cooling increase, More accurate surface profile control becomes possible.
[0032]
An example of such a cooling means is a refrigerant circulation system, a typical example of which is to circulate cooling water.
FIG. 7 shows a configuration example of a processing board 11 in which cooling water is circulated. The processing board 11 is formed by joining and integrating an upper half 11t having a working surface 11a and a lower half 11b to which the heater 12 is brought into contact. Cavity 16 is formed. The cavity 16 is formed, for example, in a spiral shape, and a cooling water pipe 18 inserted into a leg 17 which is a rotation shaft of the processing board 11 is connected to both ends thereof.
[0033]
FIG. 8 shows an example in which the upper half 11t of the processing board 11 is formed of an aggregate of a plurality of rigid blocks 20. Each rigid block 20 has a leg portion 20a at the lower end and a flange portion 20b at the upper end, and a large number of the rigid blocks 20 are arranged adjacent to each other via a sealing material 21 on the side surface of the flange portion 20b. The work surface 11a is constituted. The plurality of rigid blocks 20 are individually joined to the lower half 11b at the legs 20a. The planar shape of the flange portion 20b in the work surface 11a is not particularly limited, but is preferably a rectangle or a hexagon from the viewpoint of filling the work surface 11a with a gap as small as possible.
In the above configuration, the cavity 22 surrounded by the adjacent rigid block 20, the elastic sealing material 21, and the lower half portion 11b serves as a flow path of the cooling water. These cavities 22 need to be connected to each other by an appropriate route, and the through holes 20c provided in the leg portions 20a serve for this connection.
[0034]
According to such a block integrated configuration, even if the individual rigid blocks 20 undergo thermal deformation due to frictional heat during planar processing, they are absorbed by the elastic sealing material 21, so that the entire work surface 11a is initially Can be maintained as it is. That is, the thermal deformation of the lower half portion 11b is directly reflected on the surface profile of the work surface 11a without being disturbed.
Further, since the cavities 22 are provided adjacent to the individual rigid blocks 20 to secure a large total cooling area, the cooling capacity is sufficient.
Note that the above-described block integrated structure does not necessarily have to have a cavity, and for example, each rigid block may have a simple columnar shape. Even with such a configuration, an effect of faithfully reflecting the thermal deformation of the lower half portion 11bg on the surface profile of the work surface 11a can be expected.
[0035]
Incidentally, in the present invention, in addition to the above-described refrigerant circulation system, a thermoelectric cooling element may be used as a cooling means. A thermoelectric cooling element is an element that utilizes the Peltier effect in which heat is absorbed when a current flows through a contact made of dissimilar metals. In the case where a heating means such as a flat heater is used, if such a thermoelectric cooling element is used as the cooling means, both heating and cooling can be performed electrically, and the apparatus configuration can be simplified.
The thermoelectric cooling element can also be used as a heating element if the direction of current flow is reversed, so it can be used as heating means instead of a flat heater, or used for both heating and cooling It is. In the latter case, for example, a configuration is possible in which two electric circuits of thermoelectric cooling elements are provided, and current flows in one circuit in a direction in which heat is generated and in the other circuit in a direction in which heat is absorbed. However, since a thermoelectric cooling element is originally an element suitable for local cooling / heating, it is better to use a plane heater for heating a large area.
[0036]
In the present invention, when the workpiece is slid on the working surface of the processing board, an abrasive may be supplied onto the working surface to polish the workpiece. As the abrasive, any conventionally known abrasive can be used, and typically, colloidal silica is used.
In particular, a planar processing apparatus that performs chemical mechanical polishing of a workpiece by providing a polishing cloth on a work surface of a processing board and providing abrasive supply means for supplying an abrasive to the surface of the polishing cloth. Can be provided. Such an apparatus is particularly useful when the workpiece is a semiconductor wafer.
[0037]
【Example】
Hereinafter, specific examples of the present invention will be described.
[0038]
Example 1
In this embodiment, a configuration example of a chemical mechanical polishing apparatus including a heater and a heat insulating material layer on both the back surface of a processing board and the back surface of a workpiece holding member of a polishing head will be described with reference to FIG. .
In this apparatus, a disk-shaped processing board 11 in which a polishing cloth 19 is stretched on a work surface 11a and a polishing head 35 for holding, for example, a semiconductor wafer as a workpiece 14 are arranged in parallel and opposed to each other. The rotation of the processing board 11 in the direction of arrow A and the rotation of the polishing head 35 in the direction of arrow B are combined to chemically and mechanically polish the surface of the semiconductor wafer in the presence of the polishing slurry.
[0039]
The processing board 11 is configured such that an upper half 11t made of a SUS steel plate having a thickness of 30 mm and a lower half 11b also made of a SUS steel plate having a thickness of 30 mm are firmly integrated using fixing means such as bolts. It is a disc having a diameter of 700 mm. A part of the joint surface between the upper half 11t and the lower half 11b is formed as a cavity 16 and serves as a cooling water flow path.
As shown by the arrow A, the processing board 11 is rotated around a leg 17 connected to the center of the lower half portion 11b by using a driving means (not shown). The leg 17 also serves as a passage of a cooling water pipe 18 for introducing cooling water into the cavity 16, and the cooling water pipe 18 passes through a rotary joint 34 provided at an intermediate portion of the leg 17. And is led out. The cooling area of the cavity 16 is about the area of the work surface 11a.
The cooling water was supplied at a temperature of 20 ° C. and at a flow rate of 5 L / min.
[0040]
A heater 12a was brought into contact with the back surface of the lower half 11b, and the heater 12a was further covered with a heat insulating material layer 13a. ＾
As the heater 12a, for example, a nichrome wire is covered with a silicon rubber layer, and an applied voltage of 100 V is applied to 0.6 W / cm.²A plurality of silicon rubber heaters capable of obtaining the heat generation density described above were connected in parallel and used. The wiring of the heater 12 passes through the inside of the leg 17, is led out through a rotary connector 31 provided in the middle of the leg 17, and is connected to a power supply 33 through a transformer 32. . By operating the transformer 32, the amount of electricity supplied to the heater 12a, that is, the amount of heat generated can be changed easily and quickly.
As the heat insulating material layer 13, a 6 mm thick silicone rubber foam was used as an example.
[0041]
A urethane nonwoven fabric was used as the polishing cloth 19 stretched on the work surface 11a.
In the upper center of the processing board 11, a nozzle 45 for supplying the polishing slurry is opened. As the polishing slurry, for example, colloidal silica is supplied to the surface of the polishing cloth 19. A drain receiver for collecting the polishing slurry overflowing from the edge of the processing board 11, a tank for storing the collected polishing slurry, and a pump for pumping the polishing slurry to the nozzle 45 are not shown. However, this is as described with reference to FIG.
[0042]
On the other hand, the polishing head 35 is provided with a workpiece holding member 15 and a suction jig 40 as a workpiece holding member at the center of the bottom surface of the pressurizing chamber 36, and the workpiece 14 is vacuum-evacuated. The workpiece 14 is pressed against the polishing pad 19 with a predetermined pressure by air pressure while being suction-held on the workpiece holding member 15.
The wall of the pressurizing chamber 36 is a member forming a skeleton of the polishing head 35, and is rotated around a leg 37 connected to the center of the ceiling thereof by a driving means (not shown) as shown by an arrow B. . The leg 37 also serves as an air pipe 39 for introducing air into the pressurizing chamber 36, a vacuum pipe 41 connected to a suction jig 40 described later, and a wiring path for a heater 12b described later.
[0043]
The workpiece holding member 15 is, for example, a porous alumina ceramic plate having a diameter of 210 mm and a thickness of 30 mm, and one surface thereof has a holding surface 15 a for holding the workpiece 14, and a side opposite to the holding surface 15 a. Is the surface on which the suction jig 40 is mounted. The suction jig 40 is, for example, a SUS steel plate having a diameter of 210 mm and a thickness of 2 mm, and has many grooves serving as air suction paths. This groove is connected to the above-described vacuum pipe 41, which passes through the inside of the leg portion 37, is led out to the outside via a rotary joint 38 provided in the middle of the leg portion 37, and is not shown. Connected to vacuum pump.
[0044]
On the back surface of the suction jig 40, a heater 12b and a heat insulating material layer 13b are provided. These are the same as those attached to the lower half 11b of the processing board 11. The wiring of the heater 12 b also passes through the inside of the leg portion 37, is led out to the outside via a rotary connector 42 provided in the middle of the leg portion 37, and is connected to a power supply 44 via a transformer 43. By operating the transformer 43, the amount of electricity supplied to the heater 12b, that is, the amount of generated heat can be easily and quickly changed.
With this configuration, the present apparatus can perform single thermal deformation of the processing board 11, single thermal deformation of the workpiece 14,All of the independent thermal deformation of the workpiece holding member 15This enables control.
[0045]
Example 2
In this case, only the heater 12b on the polishing head 35 side of the apparatus described in the first embodiment is operated, and the chemical mechanical polishing of the silicon wafer is actually performed while changing the voltage applied to the heater 12b. The amount of surface deformation was investigated.
The silicon wafer used as the polished sample was a 200 mm diameter wafer that was cut out from an ingot and wrapped. The polishing conditions are, for example, polishing pressure = 300 g / cm.²The sliding speed was 45 mm / min, and colloidal silica was used as the polishing slurry. The voltage applied to the heater 12b was changed in the range of 0 to 40V. The amount of deformation of the polished silicon wafer was measured in a non-contact manner using a laser type shape measuring instrument.
The surface profile of the working surface 11a of the processing board 11 and the surface profile of the holding surface 15a of the workpiece holding member 15 when the heater 12b is not energized are both flat.
[0046]
The results are shown in FIG. In this figure, the deformation amount (μm) of the wafer surface is plotted against the voltage (V) applied to the heater 12 b on the back surface side of the workpiece holding member 15. Here, the amount of deformation is a numerical value indicating how much the edge of the semiconductor wafer has risen or decreased with respect to the plane, if the edge is higher than the center, the concave is concave, and the end is more than the center. If it is low, it is convex, and if it is the same, it is flat.
When the heater 12b was not energized, the semiconductor wafer had a 0.2 μm middle concave shape. This indicates that the surface of the semiconductor wafer cannot always be processed to be flat even if both the work surface 11a of the processing board 11 and the surface profile of the holding surface 15a of the workpiece holding member 15 are flat. However, as the voltage applied to the heater 12b is increased, the surface profile of the holding surface 15a of the workpiece holding member 15 changes from flat to concave, and accordingly, the surface profile of the semiconductor wafer becomes concave to flat. → It became clear that it changed to a convex center. According to FIG. 10, the applied voltage that can make the surface profile of the semiconductor wafer just flat is around 20V. Incidentally, the surface profile of the holding surface 15a of the workpiece holding member 15 at this time was slightly concave.
[0047]
Example 3
In the present embodiment, the upper half portion 11t of the processing board 11 is made into a block integrated structure, and a chemical mechanical polishing that enables highly accurate control of the surface profile of the work surface 11a by using a cooling water flow passage formed therein. The semiconductor wafer was polished using the apparatus.
First, the configuration of the used apparatus is shown in FIG. This apparatus corresponds to an apparatus in which the processing board 11 shown in FIG. 8 described above is incorporated in the chemical mechanical polishing apparatus shown in FIG. 9 and has the same reference numerals.
[0048]
Here, the individual rigid blocks 20 constituting the upper half portion 11t are formed by processing one surface of a SUS steel plate having a length of 50 mm, a width of 50 mm, and a thickness of 30 mm, thereby forming a square pillar leg 20 a of 20 mm square, A 50 mm square square flange portion 20b is formed, and a lateral through hole 20c is formed in the leg portion 20a. A number of the rigid blocks 20 were densely arranged adjacently on the lower half 11b made of a SUS steel plate having a diameter of 700 mm and a thickness of 30 mm, and were individually fixed using bolts. The working surface 11a constituted by the plurality of rigid blocks 20 was finished flat using an appropriate processing tool, and a gap between the blocks was filled with a silicone rubber-based sealing material 21.
[0049]
According to the above configuration, the gap between the leg portions 20a of the adjacent rigid blocks 20 becomes the cavity 22, which is used as the cooling water flow path. The cooling area by the cooling water is about 170% of the area of the working surface 11a. . The cooling water pipe 18 connected to the cavity 22 passes through the inside of the leg portion 17 of the processing board 11, and is led to the outside via a rotary joint 34 provided in the middle thereof.
Further, the wiring of the heater 12b abutting on the lower half portion 11b also passes through the inside of the leg portion 17 of the processing board 11, is led out to the outside via a rotary connector 31 provided in the middle thereof, and is connected to the transformer 32. Is connected to the power supply 44 via the.
Since other members have the same reference numerals as those in FIGS. 8 and 11, detailed description is omitted.
[0050]
Next, a polishing cloth 19 made of urethane non-woven fabric is stretched on the surface of the upper half 11t configured as described above, and both the heater 12b on the polishing head 35 and the heater 12a on the processing board 11 are turned off. Cooling water adjusted to 20 ° C. was circulated through the cavity 22 at a flow rate of 5 liter / min.
In this state, a silicon wafer having a diameter of 200 mm was polished at a polishing pressure of 300 g / cm.²The polishing was continuously performed at a sliding contact speed of 45 m / min. When the thickness distribution of the entire surface of the polished wafer was measured using a capacitance type flatness measuring instrument, it was found to be spherical and concave, and the central part of the wafer was 0.35 μm thinner than the end part.
[0051]
Therefore, in order to make the thickness distribution of the wafer uniform, a voltage of 10 V is applied to the heater 12a of the processing board 11 to deform the working surface 11a in the concave direction by 0.35 μm in the sliding contact area of the silicon wafer having a diameter of 200 mm. Was. The applied voltage is a value theoretically calculated from the linear expansion coefficient α of the processing board 11, the thermal conductivity Q of the processing board 11, and the characteristics of the heater. The silicon wafer was polished again under the same conditions except for the applied voltage. When the surface of the polished wafer was measured using a flatness measuring device, no concave or convex shape was generated, and it was confirmed that the surface was highly flat, and the effect of the present invention was proved. .
[0052]
Example 4
In the present embodiment, an example of the configuration of a lapping machine provided with a heater and a heat insulating material layer on both the back surface of the processing board and the back surface of the workpiece holding member of the polishing head will be described with reference to FIG.
This machine includes a carrier 64 that performs planetary motion by being engaged with both the sun gear 51 and the inner peripheral gear 54, and a processing board 57 that is arranged coaxially with the sun gear 51 and performs a unique rotational motion. The workpiece 14 mounted on the carrier 64 is brought into close contact with the surface of the surface plate 57 so that the workpiece 14 is wrapped.
[0053]
The sun gear 51 is integrally attached to an upper end of a rotating shaft 52 driven by a motor 53. The inner peripheral gear 54 is formed at the upper end of a frame 55 surrounding the outer periphery of the processing board 57, and the frame 55 is driven to rotate by a motor 56 controlled independently of the motor 53. .
The processing board 57 is a disc body made of spheroidal graphite cast iron having a diameter of 1400 mm and a thickness of 70 mm, and a leg 58 is connected to the center of the back side so as to surround the rotation shaft 52. The motor 53 is independently driven to rotate. In addition, a large number of grooves 59 are formed on the surface of the processing board 57 for guiding excess polishing slurry to the periphery. The illustration of the circulation and reuse system of the polishing slurry is omitted because it is the same as the conventional system.
Further, a heater 12a was brought into contact with the back surface of the processing board 57, and the heater 12a was further covered with a heat insulating material layer 13a. The heater 12a and the heat insulating material layer 13a are the same as those used in the first and third embodiments. The wiring from the heater 12 a passes through the inside of the leg 58 of the processing board 57, is led out through a rotary connector 61 provided in the middle thereof, and is connected to a power supply 63 via a transformer 62. I have.
[0054]
The workpiece 14 is fitted and held in an opening 65 on the bottom surface of the carrier 64. A heater 12b is in contact with the back side of the workpiece 14, that is, a surface opposite to the workpiece surface, and the heater 12b is further covered with a heat insulating material layer 13b. A support member 66 is disposed on the back side of the heat insulating agent layer 13b, and the wiring from the heater 12b passes through the support member 66 and is led to the outside via the rotary connector 67. The wiring led out is connected to a power supply 69 via a transformer 68.
With this configuration, the present lapping machine enables three types of control, namely, independent thermal deformation of the processing board 57, independent thermal deformation of the workpiece 14, or simultaneous thermal deformation of the processing board 57 and the workpiece 14. .
[0055]
Example 5
Here, a basic study on the deformation of the processing board 57 was performed while operating only the heater 12a on the processing board 57 side of the lapping machine described in the fourth embodiment and changing the voltage applied to the heater 12a.
Here, the amount of deformation of the processing board 57 when the voltage applied to the heater 12a was changed in the range of 0 to 10 V was measured in a non-contact manner using a laser shape measuring instrument. Note that the initial surface profile of the working surface of the processing board 57 when the heater 12a is not energized has a central convexity of 10 μm over the entire processing board 57 having a diameter of 1400 mm.
[0056]
FIG. 13 shows the results. In this figure, the deformation amount (μm) of the working surface of the processing board 57 is plotted against the applied voltage (V) to the heater 12 a on the back side of the processing board 57. Here, the amount of deformation is a numerical value indicating how much the end of the processing board 57 has risen or decreased with respect to the plane. If the end is higher than the center, the concave portion is located. If it is lower than the part, it is convex, otherwise it is flat.
It is clear from the figure that as the voltage applied to the heater 12a is increased, the surface profile of the processing board 57 changes from convex to concave to flat to concave. It was confirmed that the tendency of this change was in good agreement with the quadratic function shown in the above-mentioned theoretical formula (I).
[0057]
Next, actual processing was performed using the above-described lapping machine.
The workpiece 14 is a thick plate made of alumina ceramic having a diameter of 600 mm and a thickness of 20 mm, and is intended to finish one surface profile into a convex spherical shape of 5 μm.
As the polishing slurry L, a slurry in which SiC (silicon carbide) abrasive grains were dispersed in water together with additives such as a rust inhibitor and a surfactant was used.
The polishing load was only the total weight of the workpiece 14, the heater 12b, the heat-insulating agent layer 13b, the support member 66, and its accessories.
[0058]
The voltage applied to the heater 12a on the back side of the processing board 57 was determined from FIG. 13 so that the surface of the processing board 57 had a concave spherical shape having the same curvature as the target shape of the workpiece 14. That is, since the amount of deformation of the disk is generally proportional to the square of the diameter, the curvature when the workpiece 14 having a diameter of 600 mm is deformed by 5 μm is 27 [= 5 × (1400/600)²] It is equal to the curvature when deformed by μm. In addition, if the target profile of the workpiece 14 is a convex spherical surface, the surface profile of the processing board 57 corresponding to the target profile is a concave spherical surface, and the sign of the deformation amount is minus. According to FIG. 13, the applied voltage when the amount of deformation of the processing board 57 becomes −27 μm is 10V.
[0059]
When the workpiece 14 was wrapped under the above conditions, the surface profile became a convex spherical shape of 7 μm, which exceeded the target by 2 μm. The cause of this difference was considered to be a heat storage effect due to the polishing slurry L flowing along the groove 59 and other factors, in addition to the surface profile of the processing board 57. Therefore, the actually required deformation amount of the processing board 57 can be calculated by considering the target deformation amount of the workpiece 14 as 3 (= 5-2) μm for convenience, and −16 μm [= −3 × ( 1400/600)²]. According to FIG. 13, this amount of deformation can be achieved with an applied voltage of 8V.
Then, lapping was performed again under the same conditions except that the applied voltage was reduced to 8 V. As a result, a 5 μm convex spherical surface profile of the workpiece 14 could be obtained as desired for the workpiece 14.
As described above, according to the present invention, the surface profile of the processing board can be changed very easily by thermal control, and there is no need to repeat a complicated adjustment operation using a processing tool as in the related art.
[0060]
As described above, the present invention has been described with respect to the five embodiments, but the present invention is not limited to these embodiments.
For example, in the above-described embodiment, a processing example in which a heater and a heat insulating material layer are provided on the processing board side and also on the side of the polishing head and the carrier, and power is supplied to the heater on either side has been described. Both heaters may be energized to obtain a surface profile that is optimal for the target processing shape of the workpiece. Further, the heater does not necessarily need to be provided on both the processing board side and the polishing head or the carrier side as shown in FIGS. 9, 11 and 12, and may be provided on any one of them. The number of workpieces that can be processed at once by the planar processing apparatus of the present invention is not limited to one, and a workpiece that can mount a plurality of workpieces on a polishing head or a carrier may be used.
[0061]
Further, the planar processing apparatus of the present invention does not necessarily need to combine a relatively large processing board and a small polishing head as described above. For example, for an apparatus in which a workpiece is held between two large processing boards to perform double-side polishing, a heater and a heat insulating material layer are provided on at least one of these processing boards, and a similar surface profile is provided. Can be optimized.
In addition, the apparatus configuration, dimensions, shapes and specifications of each part, types of workpieces, and processing conditions can be appropriately selected, changed, and combined.
[0062]
【The invention's effect】
As is clear from the above description, in the planar processing method and the planar processing apparatus of the present invention, the work surface of the processing board, the workpiece, or the surface profile of the workpiece holding member of the polishing head that holds the workpiece is described. Since at least one of them is thermally controlled, it is possible to save time and effort such as recreating a surface profile of a processing board using a dedicated processing tool according to a processing result of a workpiece as in the related art. It greatly contributes to the improvement of processing accuracy and productivity.
The above-described thermal control can be performed quickly and easily by adjusting the amount of electricity supplied to the heating means. The workpiece can be directly heated if it is a thick plate, or indirectly through a workpiece holding member if it is a thin plate, to change its surface profile. As a thin workpiece, a process using a semiconductor wafer has high practical value.
[0063]
When the cooling by the cooling means is used, more precise control can be performed by increasing the control parameters. In terms of the apparatus configuration, it is preferable that the processing board is formed by joining and integrating the upper half and the lower half, and a cavity is provided in a part of the joining surface and the cavity is used as a cooling water flow path.
In addition, if the upper half is formed as an aggregate of a large number of rigid blocks adjacent to each other via an elastic material layer, and these rigid blocks are individually fixed to the lower half, frictional heat generated at the time of processing can be used. The thermal deformation of the upper half can be minimized, and only the thermal deformation intentionally applied to the lower half can be substantially reflected in the upper half, which is effective in improving control accuracy. If the gap between the blocks is used as a cooling water flow path, the cooling area can be increased.
If the abrasive is supplied when the workpiece is brought into sliding contact with the work surface, the polishing efficiency can be increased. Depending on the combination of the workpiece and the abrasive, chemical mechanical polishing by mechanochemical action is also possible, which is particularly suitable for polishing a semiconductor wafer.
[0064]
The present invention is applicable not only to a process for obtaining a desired surface profile such as a center convex and a concave for a workpiece, but also to a flat process that requires a high degree of flatness such as polishing of a semiconductor wafer. It can be handled without major changes to the procedure, and its industrial value is extremely high.
[Brief description of the drawings]
FIG. 1 is a schematic sectional view showing a thermal deformation model of a disk.
FIG. 2 is a schematic cross-sectional view showing a configuration in which a heater and a heat insulating material layer are arranged on a surface opposite to a working surface of a processing board.
FIG. 3 is a schematic cross-sectional view showing a configuration in which a heater and a heat insulating material layer are arranged on a surface of a workpiece opposite to a workpiece surface.
FIG. 4 is a schematic cross-sectional view showing a configuration in which a heater and a heat insulating material layer are arranged on the back surface of a workpiece holding member that holds a workpiece.
5A and 5B are schematic cross-sectional views illustrating a change in the processing shape of a workpiece according to a change in a surface profile of a work surface of a processing board whose initial state is flat, and FIG. (b) shows the case where the heat generation of the heater is small, and (c) shows the case where the heat generation of the heater is large.
FIG. 6 is a schematic cross-sectional view showing a change in a processing shape of a workpiece according to a change in a surface profile of a work surface of a processing board having a center convex in an initial state, and FIG. , (B) show the case where the heater generates heat.
FIG. 7 is a schematic cross-sectional view showing a configuration example of a processing board provided with a cavity for flowing cooling water at a part of a joint surface between an upper half and a lower half.
FIG. 8 is a schematic cross-sectional view showing a configuration example of a processing board whose upper half is composed of an aggregate of a plurality of rigid blocks.
FIG. 9 is a schematic cross-sectional view showing a chemical mechanical polishing apparatus in which a heater and a heat insulating material layer are respectively disposed on the back surface of a processing board and the back surface of a workpiece holding member as one configuration example of the planar processing device of the present invention. It is.
10 is a graph showing a relationship between a voltage applied to a heater behind a holding plate of a workpiece holding member and a deformation amount of a surface of a wafer in polishing of a silicon wafer using the chemical mechanical polishing apparatus of FIG. is there.
FIG. 11 is a schematic cross-sectional view showing a chemical mechanical polishing apparatus in which an upper half of a processing board is formed of an aggregate of a plurality of rigid blocks, as another configuration example of the planar processing apparatus of the present invention.
FIG. 12 is a schematic cross-sectional view showing a lapping machine in which a heater and a heat insulating material layer are respectively disposed on the back surface of a processing board and the back surface of a workpiece as still another configuration example of the planar processing device of the present invention. .
13 is a graph showing a relationship between a voltage applied to a heater on the back of a processing board of the lapping machine of FIG. 12 and an amount of deformation of the processing board.
14A and 14B are diagrams illustrating a configuration example of a conventional typical lapping machine, wherein FIG. 14A is a top view and FIG. 14B is a cross-sectional view taken along line XX.
FIG. 15 is a schematic cross-sectional view showing a configuration example of a conventional typical chemical mechanical polishing apparatus.
[Explanation of symbols]
11,57 processing board
11a Work surface
11t Upper half
11b Lower half
12, 12a, 12b heater
13, 13a, 13b Insulation layer
14 Workpiece
15 Workpiece holding member
15a Holding surface
16,22 cavity
18 Cooling water piping
19 polishing cloth
20 Rigid block
21 Sealing material
31, 42, 61, 67 Rotary connector
32,43,60,68 transformer
33,44,63,69 Power supply
34, 38 Rotary joint
35 polishing head
64 career

Claims (2)

A disk-shaped processing board on which a polishing cloth is stretched on a work surface, and a polishing head for holding a workpiece to be processed are arranged in parallel and opposed to each other. In the presence of a, in a plane processing apparatus for chemically mechanical polishing the surface of the workpiece,
  The processing board is a disc body in which the upper half and the lower half are both rigid plate bodies, and a part of the joint surface between the upper half and the lower half is circulated and supplied with cooling water. A first flat heater on the back side of the lower half, and a heat insulating material layer on the back side of the flat heater, and the first flat heater includes a transformer. Connected to a power supply and the amount of heat generated is controllable by operating the transformer,
  The polishing head includes a workpiece holding member in the center of the bottom surface of the pressurizing chamber, and a suction jig that is a rigid plate body, and performs the polishing while holding the workpiece by vacuum suction on the workpiece holding member. The workpiece is pressed against the cloth by air pressure.
  The workpiece holding member, one surface is a holding surface for holding the workpiece, the opposite surface is a mounting surface of the suction jig,
  The suction jig has a number of grooves connected to a vacuum pipe and serving as an air suction path. A second planar heater is provided on a back surface of the suction jig, and a heat insulating material is provided on a back surface of the planar heater. Layer, and the second planar heater is connected to a power supply via a transformer so that the amount of heat generated can be controlled by operating the transformer.
  The planar processing apparatus, wherein the configuration enables control of all of thermal deformation of the processing board, thermal deformation of the workpiece holding member, and thermal deformation of the workpiece. 2. The method for processing a semiconductor wafer by the planar processing apparatus according to claim 1, wherein a calorific value of the first planar heater, a calorific value of the second planar heater, and the processing board during the planar processing. 3. By controlling at least one of the temperature and / or the amount of circulation of the cooling water supplied to the cavity formed in the cavity, the surface profile of the working surface of the processing board, the surface profile of the workpiece, A planar processing method, wherein at least one of the surface profiles of the workpiece holding member of the polishing head is thermally controlled.

Images