JP2004500481A - Void / pillar network structure - Google Patents

Abstract

A novel porous membrane comprising a network of silicon pillars in continuous voids and formed at low temperatures of about 250 ° C. or less using high-density plasma has been disclosed. This silicon film is a two-dimensional array of nano-sized rod-shaped columns. The morphology of the voids / columns is controlled by the deposition conditions and the porosity can be varied up to 90%. Simultaneous deposition and etching at a low temperature in a plasma approach can simultaneously obtain columnar structures, continuous voids, and polycrystalline columnar compositions. Unique devices can be made using porous continuous films by plasma deposition of films on glass, metal foil, insulator or plastic substrates.

Description

これらの変数は、以下で詳細に説明する。
【００３５】
プラズマと基板間の電圧（ＲＦ基板バイアス）
プラズマと基板間の加速電位（即ち、基板バイアス）は、高密度プラズマ（ＨＤＰ）ツールにより独立して制御することができる。このファクタは、ＨＤＰ堆積ツールにおいて独立して利用可能であり、また可変である。即ち、このクラスのツールでは、電圧は、従来のＲＦ又はＤＣプラズマ堆積システムとは異なり、プラズマ・パワーと無関係に調節することができる。ＨＤＰで堆積した多孔性膜を作製するために使用した電子サイクロトロン共鳴（ＥＣＲ）高密度プラズマ装置では、この電圧はプラズマ及び基板間にＲＦバイアスを印加することによりプラズマパワーと無関係に制御される。
【００３６】
薄膜シリコンの堆積 について議論する論文（Ｓ． Ｂａｅ． Ａ．Ｋ． Ｋａｌｋａｎ， Ｓ． Ｃｈｅｎｇ． ａｎｄ Ｓ．Ｉ． Ｆｏｎａｓｈ． Ｊ． Ｖａｃ． Ｓｃｉ． ＆ Ｔｅｃｈ． Ａ１６， ｐ．ｌ９１２ （１９９８））にあるように、プラズマと基板間にＲＦバイアスが存在すると、柱状体の形態が破壊し気孔構造のない連続膜が得られる。実際、ＪＶＳＴ論文の目的は多孔性膜の生成を回避する方法を提示することであった。即ち、膜に起こりえる柱状体の形態を取り除く方法を提示することであった。その論文の際には、膜の柱状体がテーパー状でなく、以前の柱状体シリコン膜で通常見られたカリフラワー形状でもなく、完全に制御可能な空隙率を有することは、知られていなかった。また、本件のボイド構造が、特異な連続ボイドであることも知られていなかった。また、論文時には、柱状体膜は、厚さとともに発展する形態及びテーパー状の柱状体を有する典型的なゾーン１構造を持つものと考えられていた。従って、ＪＶＳＴ論文の目的はシリコン堆積膜の柱状体生成を回避する方法を提示することであった。ＪＶＳＴ論文には、柱状体の形態は、プラズマと基板間にＲＦバイアスを印加することにより避けることができ、このＲＦバイアスアプローチの有効性が示されている。成長する膜表面に到着する種のエネルギーを増加させることにより（ＲＦ基板バイアスを増加させることにより）、ゾーン１の柱状体形成が回避できることは、ストラクチュア・ゾーン・モデルから予想される。堆積種は、ＲＦ基板バイアスの増加により、より大きなエネルギーを持つであろうし、その結果として生じる表面移動度の増加は、気孔フリー、柱状体フリーな均質膜をもたらすことになる。これが、ＪＶＳＴ論文で予期され、議論された振る舞いである。
【００３７】
本件薄膜のボイド・柱状体網目構造物を作製するとき、プラズマと基板間の電圧を最小（即ち、ＲＦ基板バイアスがゼロ）としたときに、最良の膜形態（非テーパ状でカリフラワ状でない柱状体）が得られることが発見された。本件のＥＣＲシステムでは、これは、任意のプラズマパワーレベルに対し、ＲＦバイアス電圧がゼロの場合に得ることができる。
【００３８】
基板（堆積）温度
堆積中の基板温度を増加させると表面移動度は増加する。従って、ストラクチュア・ゾーン・モデルから予測されるように、空隙率は基板温度の増加とともに減少する。しかしながら、我々が検討した温度範囲（５０℃＜Ｔ＜２５０℃）において柱状体は継続して均一であり（即ち、テーパ状にならず）、カリフラワー状タイプの成長傾向を示すことなく配列するという事実は、ＪＶＳＴ論文では予期されず議論もされなかった。膜は、膜厚により発達する形態を示すものではない。注目すべき点は次の通りである。即ち、定められた範囲の温度は、空隙率を調節するためのツールとなり得ることが分かり、これは予期されるかも知れないが、驚くことに、連続的なボイド（気孔）網目構造にわたって突出し、テーパー状でなくしかも均一な柱状体を失うことなく達成できることである。この点は、本技術分野では予期されず、以前には議論されていなかった。
【００３９】
プラズマ（マイクロ波）パワー／プロセス圧力
ストラクチュア・ゾーン・モデルでは、プラズママイクロ波パワーを増加させるとより高エネルギの表面種が生成し、従って、空隙率は小さくなると予期される。しかしながら、実際に観察されたのは、以下の通りである。即ち、プラズマ・パワー及びプロセス圧力以外のファクタをすべて固定したとき、空隙率はパワーとともに減少するが、驚くことに、連続的ボイド網目構造体中で突出する、配向、非テーパ状の均一な柱状体が存続し続け、また膜厚により発達することがないのである。もし柱状体が存続すれば、プラズマ・パワーを増加すると柱状体形成が緩和するか、又は少なくともテーパ状、カリフラワー状の成長が促進されると予測される。しかし、現実に起こるのは、ボイド・サイズの系統的な減少及び同じ均一な配向柱状体の存続である。
【００４０】
システムのプラズマ・パワーを変えたとき、圧力は必ずしも独立ではないことが発見された。所定のプロセス圧力及びガス流量に対し、入射するマイクロ波パワーをあるしきい値レベル以上とすると、このデモンストレーションで用いられた水素とシランのＥＣＲプラズマは安定する。しきい値レベル又はその少し上では、全ての入射パワーはプラズマに吸収される。もしパワーレベルをしきい値レベル以上にさらに大きくすると、吸収されるパワーはほとんど変化せず、過剰のパワーは反射される。即ち、吸収される正味のマイクロ波パワーとプロセス圧力とは強く結合する。図３は、特別な高密度プラズマシステムを用い、水素とシランの流量をそれぞれ４０ｓｃｃｍ及び２ｓｃｃｍとし、圧力を変化させたときに吸収されるパワーレベルの軌跡を示す。
【００４１】
図４は、１００℃で堆積された１組の膜について測定された光学的反射率スペクトルを示す。ここで、また、種々のプロセス圧力は、図３に与えられるように、あるパワーレベルに対応する。約２７５ｎｍのピークは、膜の結晶性を特徴づけるピークである。スペクトルには、約４００ｎｍの上で、光学的吸収の増加に伴って、短波長側向かって減衰する基板からの反射光に起因する干渉が示されている。従って、誘電体混合（ｄｉｅｌｅｃｔｒｉｃ ｍｉｘｉｎｇ）理論に基づいて、図４に見られる高空隙率膜の干渉フリーな反射率から、次のように推測することができる。即ち、他の堆積パラメーターを一定とした場合、プロセス圧力とともに（それ故、図３に従いパワーを減少させると）空隙率は単調に増加する。図４の挿入図は、３８０ｎｍの反射率に基づく空隙率の計算値を示すグラフである。
【００４２】
図５に、１００℃及び種々の圧力（６、８及び１０ｍＴｏｒｒ）で堆積した３つの膜のＡＦＭ表面像を示した。ここで、突出する特徴は、柱状体のクラスタを示している。ＡＦＭのチップは、そのサイズ及び形状によって、クラスタ間の急激な降下を追従することができない。図５から、クラスタ密度は、パワーとともに（即ち、圧力の減少により）系統的に増加することが分かる。図６に示し又以下で議論するように、ホトルミネセンス（ＰＬ）は図５の全ての試料について観測された。従って、これら全ての試料の柱状体サイズは、数１０ｎｍの範囲にあるはずである。また、図５及び６から、圧力を下げると、クラスタ及び柱状体間の平均空隙率が減少することが分かる。
【００４３】
従来の電気化学的にエッチングされた多孔性Ｓｉ膜がそうであるように、我々の配列したボイド・柱状体膜はホトルミネセンスを示す。１００℃及び６、８及び１０ｍＴｏｒｒで堆積した３つの膜のホトルミネセンス・スペクトルを図６に示す。ここで、〜０．３ｅＶの半値幅を有する〜１．８ｅＶのバンドピークは、多孔性Ｓｉで観察されたものと類似するものである（Ｌ． Ｔｓｙｂｅｓｋｏｖ， ＭＲＳ Ｂｕｌｌｅｔｉｎ， Ａｐｒｉｌ １９９８． ３３ （１９９８））。
【００４４】
基板近傍での磁界
ＨＤＰシステムは、基板位置又はその近くに磁界（ＭＦ）を形成することができる。これは本件のＥＣＲ高密度プラズマ堆積システムもそうである。基板近傍での磁界強度が膜空隙率に影響することが示された。膜形態に対する磁界印加の研究から、他の全ての堆積ファクタを一定に保持した場合、堆積表面での磁界制御は、膜空隙率を増加させるのに有用であることが分かった。また、堆積表面に垂直な磁界の制御が膜空隙率を増加させるのにより好ましく有用である。図７の走査型電子顕微鏡写真から、堆積面及びその直上に磁場を加えないと、より高密度に充填された構造となり、即ち、空隙率は小さくなることが分かった。この磁界の効果は、膜形態のストラクチュア・ゾーン・モデルでは考慮すらされてない。
【００４５】
堆積ガス及び流速
表１に示したように、水素及びシランの前駆体ガスについて流速範囲を検討した。柱状体膜の堆積に用いるガス組成及び流速が形態に影響することが分かった。もちろん、例えばシリコンの堆積に用いるガスは、シリコン含有分子を含むものでなければならないが、さらに、膜のユニークな柱状体及び連続的なボイド（気孔）構造を得るためには、水素もまた重要であることが分かった。例えば、文献からは、プラズマパワーを増加させると、より多くの水素ラジカルが得られ、従って、プラズマパワーとともに空隙率が増加することが予想される。しかしながら、本発明者は、プラズマパワーにより空隙率が現実には減少することを発見し、複雑な相互作用が起こることを示した。多孔性堆積膜の形態に及ぼす水素やパワーのこのような役割は、観察されたことはなくまた文献にも報告されていない。
【００４６】
水素ラジカルは、遷移層に垂直で、連続的なボイド中に成長する柱状体の位置を決定する核生成プロセスにおいて、重要な役割を担っていると考えられる。水素ラジカル／衝撃種の相互作用は、文献で議論されたものからは予期し得ない核生成の振る舞いに結びつくものである。図１に示されるナノ構造は、約６×１０^１１ｃｍ^−２の柱状体形成サイトの核密度に相当する。この密度は、水素希釈シランを用いＲＦ−ＰＥＣＶＤにより堆積したナノ結晶シリコンに見られる密度（即ち、３×１０^１０ｃｍ^−２）よりはるかに大きいものである（Ｈ．Ｖ． Ｎｇｕｙｅｎ ａｎｄ Ｒ．Ｗ． Ｃｏｌｌｉｎｓ． Ｐｈｙｓ． Ｒｅｖ． Ｂ４７． １９１１ （１９９３））。空隙率に直接影響し、高いパワーの結果として生じる高い核密度は、通常、形態理解のガイドとなるストラクチュア・ゾーン・モデルでは全く考慮されていなかった。ここで提案されたモデルでは、これは水素ラジカル形成効果を通して起こるということである。柱状体成長核密度は１０^１１〜１０^１２ｃｍ^−２である。
【００４７】
チャンバ・コンディショニング
チャンバの清浄度は、明らかに膜収率に影響する。例えば、チャンバ内のパーティクルは膜質に有害である。また、チャンバ壁の化学状態（即ち、膜コーティング）も膜に影響する。多孔性膜の堆積に高密度プラズマを用いる場合に特有な、又はこれに起因する基本的なチャンバ・プラズマ種の相互作用が存在する。これらの壁・種相互作用は膜成長中の動力学及び種バランスに影響を与え、それにより空隙率に影響を与える。例えば、チャンバ壁の窒化シリコンコーティングは、各堆積後の膜再現性や膜質に有益である。ボイド・柱状体網目膜の堆積前に行うＨ_２またはＯ_２のプラズマ・チャンバ処理も、コンディショニングの他の例である。この場合、プラズマはチャンバ壁上の種を揮発させ（Ｈ_２またはＯ_２）又は酸化させ（Ｏ_２）、その後の多孔性シリコンの堆積に安定で再現性のある環境とすることができる。
【００４８】
基板表面
多孔性膜が堆積される表面の状態は、形成される膜に影響する。例えば、シリコンウエハ、インジウム・スズ酸化物（ＩＴＯ）、未処理ガラス、金属コートガラス、及び窒化シリコンコートガラスを比較すると、堆積膜に空隙率の違いが現れる。ＩＴＯ及び未処理ガラス上の膜は非常に類似し、窒化物コートガラスの場合と比べてわずかに小さな空隙率となる。即ち、より高い空隙率は、膜及び基板間のより大きな構造的ミスマッチに関連し、より低い核密度となる。
【００４９】
本件の膜は、湿式電気化学的に作製された多孔性シリコンについてのみ今まで報告されてきた多くの特性を示す。例えば、電気化学エッチングされた多孔性シリコンについて報告されているガス及び蒸気に対する感度を示す。（Ｉ． Ｓｃｈｅｃｔｅｒ ｅｔ ａ１．， Ａｎａｌ． Ｃｈｅｍ． ６７， ３７２７ （ｌ９９５））。予備的検討として、我々の膜の水蒸気に対する感度を研究した。図８は、ＲＨの変化に応答する膜の典型的な挙動を示すものである。気孔サイズに依存して、弱い応答がしきい値湿度レベルまで見られ、それ以上では急激な増加（一般に指数関数的）が観察された。この急激な変化の開始は、気孔サイズ及び安定化・増感処理によって制御される。図８の場合は、多孔性堆積Ｓｉの上にＰｄ層を堆積して、６００℃で１時間のアニール処理を行った。
【００５０】
伝導率の増加は、ボイド内部の水の毛管凝縮に帰結する。セラミックスと同様の挙動が観察され、電荷輸送はプロトン（Ｈ^＋）、ヒドロニウムイオン（Ｈ_３Ｏ^＋）、又は水酸イオン（ＯＨ⁻）を介するものと考えられる。（Ｔ． Ｈｕｂｅｒｔ， ＭＲＳ Ｂｕｌｌｅｔｉｎ， ４９ （Ｊｕｎｅ １９９９）； Ｂ．Ｍ． Ｋｕｌｗｉｃｋｉ， Ｉ． Ａｍ． Ｃｅｒａｍ． Ｓｏｃ． ７４， ６９７ （１９９１））。一旦シリコン表面が湿度と相互作用を起こすと、化学的に結合した水酸イオンの形成が大きな電荷密度及び強い静電場を引き起こす。それ故、この化学吸着層の上に物理吸着する水分子は、この高い静電場により容易に解離してイオンとなる（２Ｈ_２Ｏ←→Ｈ_３Ｏ^＋＋ＯＨ⁻）。（Ｔ． Ｈｕｂｅｒｔ． ＭＲＳ Ｂｕｌｌｅｔｉｎ， ４９ （Ｊｕｎｅ １９９９）； Ｂ．Ｍ． Ｋｕｌｗｉｃｋｉ， Ｊ． Ａｍ． Ｃｅｒａｍ． Ｓｏｃ． ７４． ６９７ （１９９１））。プロトン輸送は、Ｈ_３Ｏ^＋が、近隣のＨ_２Ｏにプロトンを放出しＨ_３Ｏ^＋に変換させる等により行われる（Ｇｒｏｔｔｈｕｓｓ連鎖反応）（（Ｔ．Ｈｕｂｅｒｔ， ＭＲＳ Ｂｕｌｌｅｔｉｎ， ４９ （Ｊｕｎｅ １９９９）， Ｂ．Ｍ． Ｋｕｌｗｉｃｋｉ， Ｊ． Ａｍ． Ｃｅｒａｍ． Ｓｏｃ． ７４， ６９７ （１９９１））。伝導率の急激な上昇は、ＲＨの増加に伴う水分子層の成長に関係していると考えられている。（Ｊ．Ｊ． Ｍａｒｅｓ ｅｔ ａ１．， Ｔｈｉｎ Ｓｏｌｉｄ Ｆｉｌｍｓ ２５５， ２７２ （１９９５））。ＲＨを増加させると伝導率の急激な増加に続いて飽和が起こる。これは、ボイドの容積が凝縮水により完全に充填される時に起こると考えられる。従って、与えられた表面張力又は吸着係数に対し、ケルビンの毛管凝縮に関する関係から示されるように、より大きなボイド又は気孔サイズに対して、飽和はより大きなＲＨで起こると考えられる（Ｔ． Ｈｕｂｅｒｔ． ＭＲＳ Ｂｕｌｌｅｔｉｎ． ４９ （Ｊｕｎｅ １９９９）； Ｊ．Ｊ． Ｍａｒｅｓ ｅｔ ａｌ． Ｔｈｉｎ Ｓｏｌｉｄ Ｆｉｌｍｓ ２５５， ２７２ （１９９５））。
【００５１】
図９は、１００℃及び種々の圧力（６、８及び１０ｍＴｏｒｒ）で堆積した３つの膜の安定化した湿度レスポンスを示す。これらの膜の中で、１０ｍＴｏｒｒで堆積した膜（最も高い空隙率膜）が最も規則的な挙動を示している。図１０は、この高空隙率膜を用い、分、時間及び日をおいて繰り返し測定した結果を示すグラフである。再現性は明白に重要である。図８〜１０のデータは安定化・増感処理した膜のものであるが、このような処理は必ずしも必要としない。
【００５２】
Ｍａｒｅｓらは、電気化学的に作製した多孔性Ｓｉを用い０から１００％までＲＨを変化させ、図８と同様のＳ字カーブを示し、３桁増加する伝導率を報告した。（Ｊ．Ｊ． Ｍａｒｅｓ ｅｔ ａ１．， Ｔｈｉｎ Ｓｏｌｉｄ Ｆｉｌｍｓ， ２５５． ２７２（１９９５））。しかしながら、それらの測定は、非常に不安定であり、安定した読みを得るには数日を要する。この理由は、電気化学的に作製した多孔性シリコン層に固有なシリコン材料の存在により、電流は多孔性シリコン層の厚さ方向にのみ輸送することができるからである。もし横方向の電界をこの試料に適用すると、これは電流をシリコンウエハを通してシャントさせることになり、電気的性質の横方向のモニタリングは不可能となる。従って、従来の多孔性シリコンを２つの電極間に挟んだサンドイッチモニタリング構造が必要となる。これは、水蒸気が多孔性半導体層に達する前に上部の電極（接点）を通過しなければならないので大きな欠点となる。この結果、応答時間とともに感度も大きく低下することになる。
【００５３】
一方、堆積したボイド・柱状体網目構造物では容易に遂行されるが、多孔性膜を絶縁体上に容易に形成でき、これにより電気的性質は横方向コンタクト構成で容易に横方向でモニターすることができる。この構造により、接点間の領域で、雰囲気と計測される膜との間の直接的な相互作用が可能となる。その結果、ＲＨが０から１００％に急に変化したとき又はその逆の場合、膜は０．１秒台以下の応答時間を示すことになる。更に、本件の形態では、ボイドは相互に連結しているため、低空隙率の場合であっても柱状体は分離しており、膜全体にわたって膜・雰囲気界面に垂直に配列した連続的なボイド網目構造となる。これは、膜中へ又は膜外への均一で迅速な物質移動を促進する。電気化学的に作製した多孔性シリコン膜は、ボイドが必ずしも相互に連結しないので、このような特徴を持ち得ない。
【００５４】
図９には、飽和が起こるＲＨが３つの膜の異なるボイドサイズに応じてシフトする様子が示されている。上述した毛管凝縮に関する議論によると、ボイド・サイズの増加は、プロセス圧力の増加（パワーの減少）に関連すると推測される。これは、ＡＦＭ及びＰＬのデータによる我々の結論と全面的に一致する。それ故、ＲＨに対する多孔性膜の伝導度の挙動から、形態について貴重な情報が得られることが分かる。ボイド容積がすべて凝縮水で充填された場合、空隙率が最も大きくかつ厚いという理由から、１０ｍＴｏｒｒ試料が最大の電流を示すと予想される。しかしながら、現実に起こったのはこれと逆であり、図９（６ｍＴｏｒｒから１０ｍＴｏｒｒまで）から見られるように、ボイド容積（空隙率と膜厚の積に比例）とともに飽和電流は減少する。この理由は、本件膜では、支配的な電荷輸送は、バルク液体中では起こらず、表面電荷により水がもっとも効果的に解離するシリコン表面の近くで起こるためと考えられる。表面でのイオン化比はわずかに１％程度と見積もられるが、これは液体の水の中での値より６桁大きな値である。（Ｂ．Ｍ． Ｋｕｌｗｉｃｋｉ． Ｊ． Ａｍ． Ｃｅｒａｍ． Ｓｏｃ． ７４， ６９７ （ｌ９９１））。このため、最も高い柱状体密度の６ｍＴｏｒｒ試料が、最大の内表面積により、最大の飽和電流を示す。一方、１０ｍＴｏｒｒ膜は、内表面積が最小となるだけでなく、柱状体間の平均間隔は最大となる。間隔が大きくなるほど、イオン濃度は、表面での値から間隔の中心に向かって一層大きく減少し、より大きな直列抵抗が加わることになる。
【００５５】
その場又は外部での堆積後処理は、形態、感度及び特性を調整するために用いることができる。例えば、図１１は、堆積・エッチング後に水素プラズマ・エッチングを行うことにより、空隙率が増加する様子を示している。プラズマ露出時間に伴う空隙率の増加は、シリコン柱状体、クラスタ又はその両方の直径の減少に帰結される。上述したように、膜は、高温処理、化学的相互作用、その後の堆積、湿式処理、及び／又はプラズマ処理により改変することができる。さらなる具体例として、劣化を回避するために多孔性膜の表面をＦ又はＣｌで終端させることもできる。化学的、生物学的又はプラズマ処理により、膜表面を機能化させることができ、水蒸気以外の特定の検出ターゲットに対する感度を向上させることができる。
【００５６】
これら制御された空隙率を有する多孔性のボイド・柱状体網目構造シリコン膜は、迅速で再現性がよく、制御可能にエッチングできるという能力により、アイソレーション、分離層、ボイド形成及び閉込め（ｃｏｎｆｉｎｅｍｅｎｔ）など種々の用途に用いることができる。それらは、化学的、電気的、物理的、及び機械的アイソレーションに用いることができる。我々はこれらの用途の全てのタイプを、エアギャップ、アイソレーション又はリリース層用途という。例えば電気的アイソレーションの場合、これらの膜を酸化し、ＳｉＯ_２柱状体の２次元配列（図１２ａの層３）を形成する。得られる空気・ＳｉＯ_２材料は非常に小さな比誘電率を持つことになり、容量結合を最小にしなければならない場合に要求される。特別な応用の１つに、マイクロエレクトロニクスの配線がある。また、２つの２−Ｄ配列層（層３）を示す図１３に見られるように、キャップ層を二次元配列上に堆積することもできる。電気接続は、層１、ベース層、及び／または基板内に存在させることができる。これらの層間の相互接続は、ドーピング又は他の導電体により多孔層を通して達成することができる。
【００５７】
このような電気的接続又は回路は、二次元（２−Ｄ）の空気・ＳｉＯ_２配列の優れた電気的絶縁特性により、即ち、多孔性絶縁層の低い誘電率により、基板又は他の層の上又は中の伝導体又は回路に対し、最小の電気的容量結合を持つことになる。例えば、酸化シリコンブリッジ構造の空隙率が７０％である場合、有効誘電率は１．８４となる。
【００５８】
機械的、物理的及び化学的アイソレーションは、ボイド・柱状体網目構造により与えられる制御された構造の堆積膜を用い、選択領域の材料を除去することにより得ることができる。例えば、キャップ層１は、シリコンのまま又は化学反応によりＳｉＯ_２、金属ケイ化物又は他のシリコン化合物となる膜３の上に直接堆積する。キャップ層は真空を破り、又は破ることなく形成することができ、図１２に示すように、窒化物、酸化物、金属又は別の多孔性でない（即ち、連続的）シリコン層（キャップ１）が、基板５（例えば半導体、ガラス又はプラスチック）上に形成された２−Ｄボイド・柱状体網目構造（膜３）を覆うように形成される。図１２（ｂ）及び１２（ｃ）に示すように、キャップの堆積前又は後に、リソグラフィのような従来アプローチが施され、エッチング・アクセス孔１０が形成される。これらの孔１０により、エッチング液が多孔性のボイド・柱状体網目構造シリコン（又はＳｉＯ_２など）に入り、図１２（ｂ）に示すように、予め決められた領域での除去を可能にする。結果として、アイソレーション又は閉込め領域が、多孔性材料の除去により定められる。多孔性のシリコンのボイド・柱状体網目構造材料はこのエッチングによって容易に除去することができる。これは、エッチング液のアクセス及び反応生成物の除去を可能とする柱状体間のボイド又は気孔の連続的な網目構造により、多孔性の材料シリコンのすべての領域がアクセス可能だからである。図１４は、このようなボイド・柱状体網目構造Ｓｉ領域を除去することによって作製した、実際の「エアギャップ」構造を示す。この処理により、絶縁体層間にキャビティが作製されることが分かる。この具体例においては、キャビティは２０分のエッチングで形成された。多結晶シリコンを用いてキャビティを形成する従来のアプローチは、１桁以上の長い時間を必要とし、又はエッチングを加速するには複雑な電子ビームリソグラフィを用いる必要があった（Ｐ．Ｊ． Ｆｒｅｎｃｈ． Ｊ． Ｍｉｃｒｏｍｅｃｈ． Ｍｉｃｒｏｅｎｇ．， ６， １９７ （１９９６））。
【００５９】
図１４で示される除去領域（キャビティ又はエアギャップ）は、ダイアフラム、加速度計、ボロメーター、及び低減された容量結合に用いられる以外に、クロマトグラフィー、ＤＮＡ選別、光学的用途及び流体工学等の応用におけるチューブとして用いることができる。
【００６０】
図１５は他の用途における多孔性堆積膜の使用例（質量脱着分光用サポート基板）を示す。ここで、多孔性堆積シリコンは、その後に検出される分子を付着する媒体として働く。検出には、多孔性シリコンに吸着する分子を脱着させるとともに、イオン化させるためのレーザーが用いられる。脱着した分子から生じるイオンは、その後、質量分析により同定される。多孔性堆積シリコン膜上に意図的に吸着されたいくつかの分子が、図１５の質量分析スペクトルに見られように、同定される。基板から分子を脱着させるためにレーザーを用い続いて質量分光法を用いることは以前にも行われているが、それには、検出する分子を結合させる有機媒体が必要であった。これは、いわゆる、レーザー脱着及びイオン化による質量分析（ｍａｓｓ ａｎａｌｙｓｉｓ ｂｙ ｌａｓｅｒ ｄｅｓｏｒｐｔｉｏｎ ａｎｄ ｉｏｎｉｚａｔｉｏｎｍＡＬＤＩ）技術である。我々の技術では、有機媒体結合材料は不要である。明らかなように、レーザー照射により脱着及びイオン化が起こり、分子の同定が可能となる。電気化学的にエッチングされた多孔性シリコン基板もまた限られた質量範囲でこのレーザー脱着分光に有効であることが示された。また、結合材料は不要となる（Ｎａｔｕｒｅ． ｖｏｌ． ３９９， ｐ．２４３−２４６， （ｌ９９９））。しかしながら、このエッチングされた多孔性シリコンのアプローチには、上述した問題が付随する。即ち、電気化学的エッチングのために裏面に電気接点が必要なこと、及び検出される分子の結合に必要なナノ構造を得るためには厚い出発シリコンが必要なことである。一方、我々の膜は、堆積によるナノ構造を有し、エッチングステップを要求せず、しかもプラスチックを含むあらゆる基板上に堆積することができる。
【００６１】
これら全て用途及び他の用途において、ここで示された多孔性堆積シリコンを製造するアプローチは、電気化学的湿式エッチ材料に比べ明白に利点を有するものである。これらの利点は主に次の４つのグループに分類される：（１）基板の多様性、（２）薄膜に対してさえ空隙率及び膜相を調整できること、（３）厚さと空隙率の一貫性、及び（４）プラズマ処理の柔軟性。
【００６２】
本発明の低温直接堆積アプローチにより、多孔性半導体膜は、絶縁体、金属箔及びプラスチックを含む任意の基板上に形成することができる。本件膜はドープしてもしなくてもよく、約１０ｎｍより大きな任意の膜厚が可能であるが、１０〜２０ｎｍが好ましい。電気化学的アプローチでは、多孔性半導体層は伝導性の基板又は層上でのみ得ることができる。出発材料はドープする必要があり、また、要求される形態を得るには十分な厚さが必要となる。また、従来の堆積アプローチでは、膜厚とともに変化する形態となり、別個のエッチングプロセスを行った後でのみ我々の膜と同程度の空隙率を得ることができる。そして、常に非晶質となる。一方、プラズマを基礎とするプロセス並びに我々が採用するパラメーター及びコンディショニングを用いたエッチング及び堆積の同時処理は、任意の基板上に堆積される薄膜の空隙率及び相を制御するドライプロセスを可能とする。
【００６３】
上述したように、これらの膜が絶縁構造上に堆積可能という事実は、センサ構造における用途を促進する。更に、膜表面に対し垂直に配列し、連続かつ均一なボイド網目構造及びボイド容積を有する本発明のナノ構造により、膜への又は膜からの物質移動が促進され、質量脱着への応用のみならず電気的センサ、電気泳動及びクロマトグラフィーのような応用にも用いられる。非テーパ状柱状体及び相対的に均一で連続的なボイドを有する配列ボイド・柱状体網目構造形態の特徴は、本発明の多孔性シリコンでのみ見られるものである。このため、本件の膜は、吸着、脱着、輸送、及び不動化の用途（例えば、ＤＮＡのような種を同定、選別、固定並びにモレキュラーエレクトロニクスの分子の接触等のために不動化及び固定化すること）に特に優れている。直接堆積アプローチにより、多孔性半導体は透明性、分極性等の種々の光学材料基板上にも作製することができる。これは、反射防止、光学的キャビティ、及び光トラッピングコーティングのような光学的応用における多孔性半導体の使用を促進し、拡大することになる。また、多孔性半導体を柔軟性のあるポリマ基板上に堆積すれば、非平面の屈曲した、多孔性半導体層も利用可能になる。
【００６４】
本アプローチは、プラズマを基礎とするプロセス技術の利点を活用するものである。例えば、本件の高空隙率シリコン膜の化学組成は、同時堆積・エッチングプロセス中にプラズマに組み入れられる前駆体ガスにより制御することができる。これにより、イントリンシック、ｎ型又はｐ型の多孔性Ｓｉだけでなく、多孔性のＧｅ、ＧｅＣ、ＳｉＣ、ＳｉＯ、ＳｉＦ、ＳｉＣｌ等を作製することができる。更に、膜の化学組成は、堆積中に変化させることができ、それ故、膜厚全体にわたって変化させることができる。対照的に、前もって成長させた材料の電気化学的エッチングの場合、最終的に表面となるところ以外では多孔性シリコン層の化学組成を決定することはできない。また、その表面組成の制御は使用される電解液により制限される。さらに、本発明の膜は高真空中で成長させるため、大気露出による汚染を防止することができるが、従来の多孔性シリコン膜は、湿式化学薬品中で処理されるため、簡単に汚染されてしまう。
【００６５】
電気化学的エッチングアプローチ又は本発明のプラズマ堆積アプローチのいずれを採用するにしろ、得られる多孔性シリコン材料は、作製完了時には水素原子で終端されている。しかしながら、雰囲気の状態で、水素終端表面はかなり室温酸化及び水素ロスをこうむることになる。これらのメカニズムの両方とも、多孔性シリコンの物理的性質を劣化させる。従って、安定した表面パッシベーションを直ちに行えば、多孔性シリコン膜の劣化は抑えることができる。これは、真空処理により、マルチチャンバシステムを用い、コーティング、ディッピング及び蒸気露出等を別個のチャンバで行うことで、又は膜堆積・エッチングの後にチャンバ雰囲気を切り替えその場（ｉｎ ｓｉｔｕ）で行うことで、達成することができる。この観点から、より小さな活性化エネルギーでどのような種類の反応も促進させ、プラズマ表面処理を容易に行えるので、本発明の高密度プラズマアプローチは非常に柔軟性が高い。一般に、表面の化学状態は堆積中に又は堆積後に制御することができる。一例として、多孔性膜のＦ−又はＣｌ−終端表面はその場で形成することができ、このような表面は劣化に対して高い耐性があると考えられる。もちろん、多孔性膜を外部で（ｅｘ ｓｉｔｕ）で処理することも可能であり、表面処理により疎水性又は親水性にすることもできる。また、本発明の膜は、表面処理により機能化することも可能である。
【００６６】
以上、本発明に関連していくつかの実施例を述べてきたが、当技術分野の熟練者に明白な変更を加えてもよいことは明らかに理解できるものである。従って、本発明は詳細に記載された範囲に限定されるものではなく、特許請求の範囲のスコープ内で変更及び修正を含むものである。
【図面の簡単な説明】
【図１】
ガラス上のボイド・柱状体網目構造シリコン薄膜の断面のＴＥＭ写真である。
【図２】
６、８及び１０ｍＴｏｒｒ、１００℃で堆積したボイド・柱状体膜のＸＲＤパターンである。
【図３】
水素及びシラン流速がそれぞれ４０ｓｃｃｍ及び２ｓｃｃｍである場合に、吸収されるマイクロ波パワーとプロセス圧力の関係を示すグラフである。
【図４】
５〜１２ｍＴｏｒｒの種々のプロセス圧力及び１００℃で堆積した膜について、硫酸バリウムレフレクタによりモニタした反射率を示すグラフである。挿入図は空隙率の計算値を示す。
【図５】
（ａ）は１００℃、６ｍＴｏｒｒで堆積したボイド・柱状体膜のＡＦＭ表面像である。（ｂ）は１００℃、８ｍＴｏｒｒで堆積したボイド・柱状体膜のＡＦＭ表面像である。（ｃ）は１０ｍＴｏｒｒ、１００℃で堆積したボイド・柱状体膜のＡＦＭ表面像である。
【図６】
１００℃、６、８及び１０ｍＴｏｒｒで堆積したボイド・柱状体シリコン膜のホトルミネセンススペクトルである。
【図７】
基板近傍の磁界が構造に及ぼす効果を示すＳＥＭ写真である。（ａ）は１００℃、１０ｍＴｏｒｒで基板に垂直な１００Ｇの磁界を加えてＩＴＯ（インジウム・スズ酸化物）上に堆積した膜、（ｂ）は基板で磁場をかけずに堆積した膜のＳＥＭ写真である。反射率データに基づく空隙率の計算値はそれぞれ０．６２，０．２２である。
【図８】
相対湿度（ＲＨ）の増加に対応するボイド・柱状体シリコン膜の伝導率（１０Ｖの固定電圧で流れる電流）の増加を示すグラフである。膜は、８ｍＴｏｒｒ、１００℃で堆積し、使用に先立ちＰｄ安定化処理を行った。再現性改善のために、１００Å厚のＰｄ層の堆積及び６００℃、１時間のアニール後に測定した。印加電圧は１０Ｖである。
【図９】
異なる空隙率を有するボイド・柱状体シリコン膜の伝導率が相対湿度（ＲＨ）の変化に応じて変化する様子を示すグラフである。膜は、６、８及び１０ｍＴｏｒｒの圧力で堆積した。印加電圧は５０Ｖである。測定はルイス酸処理後に行った。これらの膜は、使用に先立ち、ルイス酸に曝される。
【図１０】
再現性及び安定性を示す電流ｖｓ相対湿度のグラフであり、１００℃及び１０ｍＴｏｒｒで堆積したボイド・柱状体シリコン膜を用いて５０Ｖの印加電圧で測定されたものである。この膜は、使用に先立ちルイス酸に曝された。
【図１１】
５分及び１５分の水素プラズマ堆積後処理により、反射率スペクトルから見られるボイド・柱状体シリコン膜の空隙率増加を示すグラフである。
【図１２】
（ａ）は、多孔性材料の除去により、アイソレーション又は閉込め領域の作製に有用な基本的構造を示す。この構造は、キャップ層、ナノ構造の柱状体、ベース層及び基板が含まれる。（ｂ）は、（ａ）の材料から作製されるアイソレーション領域の平面図である。（ｃ）は、（ａ）のＣ−Ｃ断面図である。
【図１３】
多孔性の材料を用いたアイソレーション又は閉込め領域の作製に有用な基本的な構造におけるいくつかの層を示す。２つの多孔層、キャップ層、ベース層及び基板が示されている。
【図１４】
ボイド・柱状体網目構造の多孔性シリコンを除去することにより作製されるチューブ又はエアギャップを示す。
【図１５】
レーザー脱着スペクトルである。試料滴は多孔性シリコン膜に直接適用され、レーザー照射に先立ち乾燥した。試料中の分子が同定されている。[0001]
This application is based on US Provisional Application No. 60 / 137,385 on June 3, 1999, US Provisional Application No. 60 / 139,608 on June 17, 1999 and October 27, 1999. Claims the benefit of US Provisional Application No. 60 / 161,848, the contents of which are incorporated herein by reference.
[0002]
【Technical field】
The present invention relates to the production of a semiconductor thin film having a high porosity by a plasma CVD system. More particularly, it relates to a unique high-density plasma approach using plasma deposition and etching simultaneously to obtain high porosity semiconductor thin films.
[0003]
[Related technology]
Porous silicon structures are of much interest. The main reasons are the following two. First, these porous membranes can be used for various applications such as microelectromechanical devices (MEMS), interlayer insulators, microsensors, immobilization of cells and molecules, and microfluids. Second, the material is very compatible with Si microelectronics. There are various approaches to producing porous silicon materials. Today, the most prominent technologies in the production of porous silicon are based on the use of wet chemical solutions and anodizing electrochemical technology (RC Anderson, RC Muller, and C.S. W. Tobias, Journal of Microelectromechanical System.vol.3 (1994) .10). Until now, this technology has allowed the highest porosity among known approaches. The starting material for this wet-etched material is either a silicon wafer or a thin film Si formed by a deposition process such as low pressure CVD (LPCVD) or plasma CVD (PECVD). In an electrochemical wet etching process, the sample is immersed in a wet solution and current is applied to the contacts of the etching sample, the etching sample, the solution (eg, a mixture of hydrofluoric acid, water and ethanol), and the electrodes (cathodes, eg, (Platinum). This current causes pit formation and etching in Si, and a porous network structure is produced.
[0004]
In the electrochemical (anodic) etching process, the structure of pores, such as pore size and spacing, and the thickness of porous Si are determined by the resistivity (size and type) of silicon itself, current density, applied voltage, electrolyte composition, Can be controlled by irradiation, temperature, processing time and the like. With a sufficiently long process and a sufficiently thick starting material, the electrochemical etching process can be continued to the point where a nanoscale structure (ie, characterized on the order of nanometers) is obtained. If the sample is etched from a single crystal wafer, the silicon feature will be a continuous single crystal, and if the sample is etched from a deposited film, it will be large granular polycrystalline silicon. All of these conventional (electrochemically etched) porous silicon are (1) the result of a wet electrochemical etching process, and (2) this wet etching requires contacts on the sample. (3) having a discrete pore region, which can be subsequently connected by wet etching, but usually (4) being the result of a continuous process of silicon formation and subsequent wet etching; It is characterized by
[0005]
In 1990, he was stimulated by the discovery of room-temperature visible light emission of porous Si electrochemically formed by Canham (LT Canham, Appl. Phys. Lett. 57, 1046 (1990)). Research activities on porous semiconductors have been more active in the last decade. Shortly after the discovery of Canham, more interesting properties of porous semiconductors, such as gas sensitivity, biocompatibility, micromachining, etc., were realized (I. Scheter et al., Anal. Chem. 67, 3727). J. Wei et al. Nature 399.243 (1999); LT Canham et al., Thin Sold Films, 297, 304 (1997); P. Steiner et al., Thin sold 25 52 (1995)). All applications shown to date are based on porous silicon made by electrochemically etching silicon wafers or deposited films.
[0006]
The present approach to porous silicon employs a deposition method that grows as a deposited porous film. In particular, porous silicon is a deposited columnar material, the pores of which are voids between columns and clusters of columns. In the present invention, the pore (void) region is fairly uniform in the film thickness and the in-plane direction. This process is unique as follows. That is, the process is performed at a low temperature. The inventor has also demonstrated that the present invention can be used to control void size and void ratio. The network form of the voids / columns does not change with the practical thickness. Further, the columns can be polycrystalline and a plasma approach can be employed to control the interaction between deposition and etching during growth. The process can produce controlled pore size materials with high porosity (up to about 90%) without the use of backside contacts or anodic wet treatment. Unlike other deposition processes, the process is based on the interaction of high-density plasma deposition and etching, and therefore has a controllable high porosity (up to 90%), a thickness-invariant morphology, A doped or non-doped polycrystalline column can be formed. Further, a feature unique to the present invention is that the above porous silicon can be formed on various types of substrates such as glass, metal foil, insulator, plastic, and semiconductor-containing materials.
[0007]
In the demonstration of the present invention, an electron cyclotron resonance plasma apparatus was used as a high-density plasma (HDP) deposition tool. In particular, our porous silicon has hydrogen-diluted silane (H2O) at substrate deposition temperatures of 250 ° C or lower. ₂ : SiH ₄ ) Was used as a precursor gas, and a high-density plasma apparatus (for example, an electron cyclotron resonance plasma CVD (ECR-PECVD) apparatus (PlasmaTherm SLR770)) was used. This device performs etching and deposition of silicon to form a two-dimensional silicon array, and as a result of analysis, the size of silicon pillars can be controlled, and the spacing between pillars can be controlled. And that the morphology does not change with thickness. Unlike other columnar deposited silicon, the columnar spacing can be maintained as the film grows and thickens, and the columnar phase can be controllably changed from polycrystalline to amorphous. The resulting void / columnar network structure has a nanoscale feature size and develops completely after a film thickness of 10-20 nm is formed. This allows for the direct deposition of high porosity polycrystalline or amorphous silicon on any substrate at any thickness greater than about 10 nm (preferably 10-20 nm). The porous semiconductor film manufactured according to the present invention can be converted into an insulator or a metal compound by in-situ or ex-situ treatment.
[0008]
There are two approaches to conventional porous silicon: (1) deposited silicon to produce porous silicon having a coral-like morphology of polycrystalline or single crystal silicon "fingers". Or wet electrochemical etching of a silicon wafer, (2) deposition of silicon to produce a porous material consisting of a tapered amorphous silicon rod having a morphology that varies with thickness. The former has the advantage of controllable morphology and porosity and has been the subject of much research and development, but its formation requires wet chemical etching. The latter is effective only for the amorphous phase, and has the disadvantage that the morphology changes with the film thickness and that a subsequent wet etching is required to control the porosity. On the other hand, the porous silicon of the present invention does not require wet processing, has a completely controllable morphology and porosity, and can be polycrystalline or amorphous as required.
[0009]
The present invention relates to low temperature deposition of porous materials. These materials are particularly suitable for deposition on substrates requiring low temperature processing, such as glass, plastic substrates, or substrates on which sensors, electronic or optoelectronic devices, circuits, etc. are formed. Due to the wide porosity range demonstrated, the materials of the present invention provide sensing, air gaps (light mixing, microfluidics, molecular sorting, low dielectric structures, etc.), molecular and cell fixation and electrical contact, and mass desorption ( Numerous applications are possible, such as mass desorption. The present invention is described using porous deposited silicon.
[0010]
DISCLOSURE OF THE INVENTION
The present invention provides a porous deposited film having a porosity of up to about 90% and composed of a plurality of perturbations extending into the void. The plurality of perturbations are arranged substantially perpendicular to the substrate or base layer. The plurality of perturbations are rod-shaped columns and semiconducting. Also, it is a polycrystal such as a silicon material. The porosity is the result of connecting voids. The perturbation is adjustable in height by thickness and has a diameter of about 1 nm to about 50 nm. More preferably, the columns have a diameter of about 3 nm to about 7 nm. In addition, perturbations are found in clusters having a diameter of about 50-500 nm.
[0011]
The present invention provides a composite structure comprising a substrate and a porous film having a porosity of up to 90% and comprising a plurality of perturbations extending into a void, wherein the porous film is formed on the substrate. provide. The composite structure further has a coating layer, and the porous membrane is formed on the coating layer. The coating layer is characterized by being at least one active material of an organic insulator, an insulating material such as silicon nitride and silicon oxide, a piezoelectric, a ferroelectric, a metal and a semiconductor. The composite structure further includes a cap layer, wherein the porous film is formed between the cap layer and the substrate. Here, the substrate may be coated. The cap layer is characterized by being an organic insulator, an insulating material such as silicon nitride and silicon oxide, or at least one active material of piezoelectric, ferroelectric, metal and semiconductor. The thickness of the porous membrane is greater than about 10 nm, where the porous membrane can be changed from polycrystalline to amorphous. The porous deposited film is characterized in that it has a structure in which rod-shaped perturbations are periodically arranged two-dimensionally. The porous membrane is, for example, oxidized to SiO ₂ To react with metal to form silicide, ₃ N ₄ Can be converted to The film may also be doped or non-doped. The substrate of the composite structure is selected from the group consisting of glass, metal, including foil, insulating materials, plastics, and semiconductor-containing materials. The substrate may optionally be coated with an atomic motion barrier, a thermal barrier, an electrically insulating, or stress controlling film. As a specific film, there is a silicon nitride barrier layer, and the thickness of the silicon nitride barrier layer can be changed from several hundreds of square meters or less to about 5,000 nm. The film can be varied during deposition and may vary throughout the film thickness.
[0012]
The present invention provides a method for producing a composite structure comprising a substrate and a porous film. The fabrication method includes depositing a porous film on a substrate by high density plasma deposition. The porous membrane of the method is characterized by having a porosity of up to about 90% and consisting of a plurality of perturbations extending into the void. The method is further characterized by the step of etching the porous film, wherein the deposition step and the etching step are preferably performed simultaneously. The etching is characterized in that it is performed in a plasma in the presence of corrosive agents such as hydrogen, chlorine, fluorine, HC1, HF and their induced radicals. The high-density plasma deposition is performed in the presence of at least one precursor gas selected from the group consisting of hydrogen and a silicon-containing gas. More specifically, the silicon-containing gas is silane. Also, the deposition step is performed at a temperature of about 250 ° C. or lower with a magnetic field in the substrate region of 800-600 gauss and a microwave excitation power of about 100-1200 W. Also, the deposition step is performed without an applied voltage between the plasma and the substrate. An additional step is to remove the porous layer in selected areas of the porous membrane by etching, thereby forming an air gap, release or isolation structure.
[0013]
More particularly, the present invention provides for preparing a substrate, conditioning a surface of a deposition tool, using a plasma deposition tool to generate ions, radicals and other excited species, and introducing power to a plasma chamber. Providing a precursor gas to a plasma deposition tool to generate a plasma, using controls to further adjust the deposition kinetics, and depositing a film on the substrate by plasma deposition.
[0014]
Depending on the type of substrate, the substrate may be coated before the deposition of the porous film. The coating uses a material selected from those useful for functionalization, such as electrical isolation, planarization, atom transfer barriers, stress control and thermal coupling. Also, the substrate may be subjected to a surface texturing treatment such as liquid chemical etching or plasma chemical etching.
[0015]
The plasma deposition tool used in the above method is a high-density plasma deposition tool, and in particular, the high-density plasma deposition tool is an electron cyclotron resonance device.
[0016]
Porosity is controlled by substrate processing such as coating, plasma power, magnetic field in substrate area, gas composition, deposition temperature, plasma-substrate bias, chamber conditioning, processing pressure, deposition gas and flow rate. Can be. The energy of the plasma species impacting the film to be deposited and etched is important and must be kept low. This control is ensured by the high density plasma system. For example, in an ECR device, the kinetic energy of the impact species is expected to be 45 eV or less.
[0017]
The present invention also considers the columnar phase, for example, for silicon, at high deposition pressures (eg, 20 mTorr), the columns become amorphous instead of polycrystalline.
[0018]
Further, the invention provides for preparing a substrate by steps such as barrier coating, conditioning the chamber, using a plasma deposition tool to form ions, radicals and other excited species, and introducing power into the plasma chamber. Step, H to generate plasma ₂ A precursor gas, such as a gas, is supplied to the plasma chamber and SiH ₄ Supplying gas to the deposition chamber, maintaining the deposition temperature below 250 ° C., using controls such as a magnetic field in the substrate area to further adjust the deposition kinetics, and depositing the film on the substrate Providing a method for producing a porous membrane comprising the steps of:
[0019]
The microwave power introduced into the ECR plasma chamber is between about 100W and 1200W, preferably between about 340W and 640W. The frequency of the microwave used is 2.45 GHz.
[0020]
H of the present invention ₂ The flow rate is between about 1 seem and about 500 seem, preferably between about 10 seem and 100 seem. In addition, SiH ₄ The flow rate is between about 1 sccm and 300 sccm, preferably between about 2 and 10 sccm.
[0021]
More specifically, the present invention relates to hydrogen-diluted silanes (H ₂ : SiH ₄ A) introducing a high density plasma tool using a precursor gas, such as a gas, introducing microwave power of about 100 W to 1200 W into the ECR plasma chamber through the waveguide and the fused silica window, and reflecting by adjusting the tuner; Minimizing power, setting the static magnetic flux density near the substrate to about +800 to about -600 gauss using a DC electromagnet, and applying about 1 sccm to 500 sccm of H to the ECR plasma chamber via a gas distribution ring. ₂ Supplying gas, about 1-300 sccm of SiH through a gas distribution ring about 1.3 cm above the substrate ₄ Injecting a gas into the deposition chamber, maintaining the deposition temperature below 250 ° C., and depositing the film on a substrate.
[0022]
The present invention relates to sensors used in various applications such as plasma displays, dielectrics, monitors for lateral resistivity, tubes, sorting, chromatography, optoelectronic devices for optical impedance applications, optical structures in solar cells, etc., and desorption spectral structures. , A gas detector and an air gap (void) structure. For these applications, a composite structure is typically comprised of a substrate and a porous continuous film of polycrystalline or amorphous silicon formed on the substrate and having a porosity of up to 90%. Used.
[0023]
DETAILED DESCRIPTION OF THE INVENTION
In connection with the present invention, a unique high-density plasma approach is described in which plasma deposition and etching are performed simultaneously to obtain a crystalline or amorphous semiconductor thin film having a high porosity. A dry process is used for all film formation, and no wet processing is required. With proper processing chamber conditioning, substrate preparation, and proper selection of deposition parameters, these films always have interconnected voids and columns that are perpendicular to the substrate and that are not tapered. It has a possible and adjustable void column network morphology. Unlike conventional porous silicon made by wet etching, this interconnected continuous void network can be obtained at any thickness greater than about 10-20 nm. Unlike conventional porous silicon, the films of the present invention may be doped or undoped. Further, unlike conventional porous silicon, the continuous void-column network of the present invention can be used not only on conventional substrates such as silicon wafers, but also on various substrates such as glass, metal foils and plastics. It can be manufactured on a substrate. The demonstration used silicon and ECR high-density plasma (HDP) equipment as an example, and this experiment further demonstrated visible luminescence, gas sensitivity, air gap structure formation, desorption mass spectroscopy, and the like. The unique continuous void arrangement, oriented columns and uniform nanostructures, low temperature processing, and unique process of the present invention take advantage of the advantages of plasma deposition technology, which allows for wet etching. This eliminates the need for wet processing, wet etching, and relatively thick starting materials that are required by conventional approaches to obtain smaller feature sizes, and offers a number of new possibilities. In addition, the present invention has many more technical and economic advantages as compared with the conventional porous silicon forming technology.
[0024]
Unlike other textured and columnar deposited silicon films, this process is not a sputter-damaged sputter deposition technique and can be deposited at very low temperatures, forming amorphous or polycrystalline columns. It has a high-density level porosity (within 90%), can form doped or non-doped pillars, has a shape that does not change with thickness, has high controllability, and has a void ( That is, this is a plasma etching / deposition technique for producing a film whose pore size can be adjusted. In this technique, the substrate is not constrained because the process temperature during film deposition is very low (ie, room temperature to 250 ° C.).
[0025]
A high-density plasma source was used to produce a continuous void / vertical non-tapered columnar form. In particular, in experiments, an ECR (Electron Cyclotron Resonance) high density plasma (HDP) PECVD system was used. In the HDP approach to high porosity films, an electron cyclotron resonance plasma CVD (ECR-PECVD) system (PlasmaTherm SLR-770) was used, with a hydrogen-diluted silane (H ₂ : SiH ₄ ) Was used. The deposition apparatus has a configuration in which two plasma chambers (ie, an ECR chamber and a deposition chamber) having diameters of 14.9 and 35.6 cm and heights of 13.3 and 29.2 cm are connected coaxially. To generate the plasma, microwave power at an excitation frequency of 2.45 GHz is introduced into the ECR plasma chamber through a rectangular waveguide and a fused silica window. The reflected microwave power is sent to a dummy load by a directional coupler, and the reflected power is minimized using a 3-stub tuner. The substrate bias can be independently controlled with a 13.56 MHz RF power supply to independently adjust substrate impact, but turned off to reduce impact energy. To meet electron cyclotron resonance conditions, a DC electromagnet mounted coaxially around the ECR chamber created an upper magnetic field with a quiescent magnetic flux density of 875 gauss in the ECR chamber. H ₂ Is introduced into the ECR chamber via a gas distribution ring near the quartz window. SiH ₄ Is injected into the deposition chamber through a gas introduction ring approximately 1.3 cm above the substrate location (below 27.9 cm in the ECR chamber). In the case of a void / columnar network structure Si film, other precursors containing hydrogen and silicon can be used, but in this experiment, hydrogen gas and silane were used. H ₂ (99.999%) is supplied to the ECR chamber via a gas distribution ring to generate an ECR plasma, while SiH ₄ (99.999%) was introduced via a gas distribution ring around the substrate fixture. In this study, hydrogen was supplied to the ECR chamber to form a silicon film, and ECR plasma was started. A second DC electromagnet is placed just below the substrate level to vary from about +800 to -600 Gauss and establish an adjustable magnetic flux density. This magnetic flux exists around the substrate. The vacuum chamber holding the substrate is evacuated by a turbo molecular pump, and the base pressure is 4 × 10 ^-7 Torr or less, for example, 2 × 10 ^-7 Torr. Table 1 shows the range of the studied plasma deposition parameters. The film was deposited on a Corning 1737 glass substrate coated with an 80 nm silicon nitride barrier layer (unless otherwise noted). The film thickness was measured with a profile meter (Tencor-500). The nanostructure was examined using a cross-section TEM (Hitachi HF-2000 cold field emission) and AFM (atomic force microscope) (Digital Instruments Multimode AFM system with NanoScope IIIa Controller). AFM was used in the TappingMode ™ using an Olympus cantilever and an etched silicon tetrahedral probe with a tip angle of ３５35 ° C. and a radius of 5-10 nm. An X-ray diffractometer (Phillips X'ert) equipped with a glancing angle optical system was used to examine the crystallinity. The UV reflectance of the film was determined using a Perkin Elmer Lambda 9 spectrometer, and the porosity level was evaluated. Photoluminescence was measured using a Raman spectrometer (ISA U-1000) equipped with a 488 nm argon laser (5 mW).
[0026]
The temperature of the fixture on the substrate was controlled by a heater and cooling water (T = 16 ° C.). Helium backside cooling was also used to control the deposition temperature during the plasma operation.
[0027]
When electrical contact is used for a void / columnar network structure film, it is formed in a parallel stripe shape (length 19 mm, interval 0.4 mm) on the film surface and has a conductivity that changes in response to various gas atmospheres. Monitor changes. In the measurement of relative humidity (RH), the sample was placed in a glass tube on a Teflon stage having two Au probes in contact with a stripe electrode. The RH inside the glass tube is N as the humidity carrier. ₂ Was gradually changed from about 2% to about 97% in a 20 ° C. glass tube by flowing through a heated bubbler into the tube. For reference of RH measurement, a capacitive sensor (TI-A from TOPLAS CO., Japan) equipped with an LCR meter (HP-4284A) is used. For a void / column structure, the current increase at a fixed voltage is measured with a pA meter. / DC power supply (HP-4140B).
[0028]
The observed morphology of conventional electrochemically produced porous Si is microporous (pore width <20 °), mesoporous (pore width 20-500 °), corresponding to pore size changes from nanometers to microns. , Macroporous (pore width> 500 °). (AG Cullis. LT Canham, PDJ Calcott. I. Appl. Phys. 82, 909, (l997)). Whether conventional porous silicon is microporous, mesoporous or macroporous depends on parameters such as starting material thickness, doping type and amount, crystal orientation, wet etching time. According to this pore size classification method, the membranes of the present invention typically have p ⁺ Or p ⁻ It is closest to conventional porous silicon called microporous silicon obtained from silicon. p ⁺ When silicon is anodized, long pores are generated along a direction perpendicular to the silicon surface with a width in the range of about 10 nm. However, since the pores themselves are largely branched and exhibit a fir-tree-like structure, the morphology exhibits a considerable degree of randomness (MIJ Beale, JD Benjaminm. J. Uren. G. Chew, AG Cullis, J. Cryst. Growth 73.622 (1985)). Anodized p ⁺ Microporous silicon materials exhibit an irregular, coral-like microstructure consisting of highly interconnected micropores in the size range of less than 2 nm (Beale et al.).
[0029]
The form of the film is completely different from a fir-tree or coral-like structure. The nanostructures of Si films deposited by the high-density plasma tool approach consist of rod-shaped columns located in a continuous void medium, perpendicular to the substrate surface. FIG. 1 is a TEM image showing the nanostructure of a film deposited at a temperature of 120 ° C., a process pressure of 7.8 mTorr, and a microwave power of 500 W. Here, a typical diameter of the column is 10 nm, and a typical interval is 3 nm. The columns in FIG. 1 are present in the cluster (although they cannot be identified at very high magnification). Higher porosity, greater than 90%, can be obtained, as inferred from near ultraviolet (NUV) reflectance measurements, but from FIG. 1 the porosity of the film is roughly estimated at about 60%. A careful examination of the columnar body shows that the columnar body is composed of granular crystal grains whose size is the diameter of the columnar body within the parameter range that makes it crystalline. In this case, the crystal grains constituting the columnar body are located on each other. The undulation (ie, change in void size) in the clearance between the columns can be seen to be caused by statistical variations in particle size. However, on a larger scale (e.g., on average in the plane of the membrane), the spacing and size of the pillars and the void spacing between the pillars in the pillar cluster are usually uniform. The voids are interconnected, so the columns are usually separated even at lower porosity, resulting in a continuous void network throughout the membrane. Therefore, compared to the morphology observed with electrochemically etched porous silicon, the deposited void-pillar network films show distinct nanostructures, and most importantly high uniformity. Show. Unusual morphology of the membrane occurs with all membranes about 10 nm or thicker. The XRD pattern in FIG. 2 shows that the film deposited at 100 ° C. and various powers of 6, 8, and 10 mTorr for 35 minutes is a crystal.
[0030]
Movchan and Demchisin (BA Movchan and AV Demchisin. Fiz. Met, Metauoved, 28, 653 (1969)), and Thomas (J.A. Thomasn.J.Tonton. 666 (1974), J. Vac.Sci.Technol.12, 830 (1975) and Messier (R.Messier, AP Giri, RA Roy, I.Vac.Sci.Technol.A2,500. (1984)) was developed to discuss the morphology of sputtered films, which was developed using a high-density plasma approach to deposit and etch. Sa Although as hereinbefore, according to the SDM, the structure of these porous silicon film is expected to fall in the form zone 1.
[0031]
In this zone 1, the membrane is composed of tapered columns (typically several tens of nm) separated by voids (typically several nm). Because of the tapered columns, this zone 1 structure develops with thickness. Zone 1 structures are also expected for the instant films, since the deposition temperature is very low compared to the melting point of the deposited material and the diffusion distance of molecules or atoms on the deposition surface is smaller than the average distance between physisorption sites. In addition, the kinetic energy of the colliding ions is approximately at or below the sputtering threshold (of the silicon film). Therefore, for these reasons, it is expected that the sedimentary species will become immobile when it lands. This is known as ballistic deposition. Under these conditions, two factors, statistical roughness and self-shadowing, cause void formation. However, these mechanisms result in voids and tapered (cauliflower-like) pillars, respectively, of random shape. In addition, these conditions lead to a form that evolves with the film thickness. The film is not a sputter-deposited film, but is produced by plasma deposition and etching and has voids and columns that are completely different from the expected results. In addition, the control of void size, porosity and columns of the instant membranes is quite different than expected. The uniqueness of this invention is its ability to produce controlled porosity, columns and doped or undoped polycrystalline or amorphous materials in a void-column network that does not grow in thickness. It is in. In general, it is not possible to have all of these features in a one-step process. Briefly, the uniqueness of this invention lies in the ability to utilize low surface diffusion for void generation, utilizing high density plasma reactions that promote etching and controllable crystallization despite low surface mobility. Is in its ability to. On the other hand, the conventional PECVD approach, in which there is no high density plasma chemistry, limits the structure to amorphous due to low surface mobility. Thus, simultaneously obtaining a controlled high void ratio, a controlled crystallinity, and a morphology that does not develop with thickness is considered out of the question.
[0032]
The void networks and their evolution with film thickness in amorphous silicon films deposited by RF sputtering have been studied in detail by Messier and coworkers (R. Mesier and RC Ross, I. Appl. Phys. 53, 6220 (1982)). The membrane can be controlled to polycrystalline silicon and can be up to approximately 90% porous. Messier et al. Deal with amorphous silicon, and do not disclose a method for obtaining the porosity enabled by the present technology except for a case where wet etching is performed after deposition (R. Messier. SV. Kishineswamy, LR Gilbert P. Swab. J. Appl. Phys. 51. 1611 (1980)). In high-density plasmas, the low temperature processing approach used, deposition, and hydrogen radical etching (resulting in an etching / deposition process) are assumed to play important roles in nanostructure formation. A possible mechanism is as follows: once crystal nuclei have formed in the presence of high-density hydrogen radicals, the bonds at these nucleation sites are stronger and resist etching of the hydrogen radicals. On the other hand, in other places, the bond is weak, and etching is performed by hydrogen radicals. As a result, growth occurs following the formation of crystal nuclei, which leads to the formation of protruding columns arranged in a regular shape of the network structure of voids and columns. Similarly, if any material fills the space between the columns, it will be etched away due to weak bonding. It is not yet understood why lateral growth does not occur in the process to form pillars that increase in diameter with increasing film thickness (ie, conical or cauliflower-like pillars), but this is very important. This may be due to a severe hydrogen etch that promotes directional growth. Before the growth of the columnar body starts, a void-free amorphous layer is formed at the substrate interface. From FIG. 1, the thickness of the transition zone of this sample is measured to be about 20 nm. Thereafter, a network of voids and columns of the porous membrane is formed on this transition layer.
[0033]
Characteristics of the deposited void-column network-type porous silicon (ie, non-tapered columns that do not evolve with thickness, are oriented perpendicular to the transition layer, and enter continuous voids) Can be controlled by a number of factors. These factors include (a) plasma-to-substrate voltage, (b) substrate temperature, (c) plasma power, (d) magnetic field and process pressure near the substrate, (e) deposition gas and flow rates, (f) ) Chamber conditioning, and (g) substrate surface. The effects of these many factors are not predictable. The ranges used in this study for some of the important parameters are shown in Table 1.
[0034]
[Table 1]

These variables are described in more detail below.
[0035]
Voltage between plasma and substrate (RF substrate bias)
The accelerating potential between the plasma and the substrate (ie, substrate bias) can be independently controlled by a high density plasma (HDP) tool. This factor is independently available in HDP deposition tools and is variable. That is, in this class of tools, the voltage can be adjusted independently of plasma power, unlike conventional RF or DC plasma deposition systems. In the electron cyclotron resonance (ECR) high density plasma device used to make the HDP deposited porous film, this voltage is controlled independently of plasma power by applying an RF bias between the plasma and the substrate.
[0036]
In a paper discussing the deposition of thin film silicon (S. Bae. A. K. Kalkan, S. Cheng. And SI. Fonash. J. Vac. Sci. & Tech. A16, p.l912 (1998)). As described above, when an RF bias is present between the plasma and the substrate, the form of the columnar body is broken, and a continuous film having no pore structure is obtained. In fact, the purpose of the JVST paper was to present a way to avoid the formation of porous membranes. That is to say, a method for removing the possible form of the columnar body in the film was proposed. At the time of that paper, it was not known that the columns of the film were not tapered, nor the cauliflower shape normally found in previous columnar silicon films, and had fully controllable porosity. . Also, it was not known that the void structure of the present case was a unique continuous void. Also, at the time of the paper, the columnar membrane was considered to have a typical zone 1 structure having a form that evolves with thickness and a tapered columnar body. Therefore, the purpose of the JVST paper was to present a method to avoid columnar formation of silicon deposited films. JVST papers show that columnar morphology can be avoided by applying an RF bias between the plasma and the substrate, demonstrating the effectiveness of this RF bias approach. It is expected from the structure zone model that by increasing the energy of the species arriving at the growing film surface (by increasing the RF substrate bias), zone 1 columnar formation can be avoided. The deposited species will have more energy with increasing RF substrate bias, and the resulting increase in surface mobility will result in a pore-free, column-free, homogeneous film. This is the behavior expected and discussed in the JVST paper.
[0037]
When fabricating the void-columnar network structure of the present thin film, when the voltage between the plasma and the substrate is minimized (that is, the RF substrate bias is zero), the best film form (non-tapered and not cauliflower-shaped columnar) Body) was obtained. In the present ECR system, this can be obtained for any plasma power level when the RF bias voltage is zero.
[0038]
Substrate (deposition) temperature
Increasing the substrate temperature during deposition increases the surface mobility. Therefore, the porosity decreases with increasing substrate temperature, as predicted from the structure zone model. However, in the temperature range we studied (50 ° C. <T <250 ° C.), the pillars were continuously uniform (ie, not tapered) and arranged without a tendency to grow in a cauliflower-type. The fact was not unexpectedly discussed in the JVST paper. The film does not show a form developed by the film thickness. Notable points are as follows. That is, it has been found that a defined range of temperatures can be a tool for adjusting the porosity, which might be expected, but surprisingly protrudes over a continuous void network, What can be achieved without losing a uniform columnar body that is not tapered. This is unexpected in the art and has not been previously discussed.
[0039]
Plasma (microwave) power / process pressure
In the structured zone model, increasing the plasma microwave power is expected to produce higher energy surface species and thus lower porosity. However, what was actually observed is as follows. That is, when all factors other than plasma power and process pressure are fixed, the porosity decreases with power, but surprisingly, an oriented, non-tapered, uniform columnar column that protrudes in a continuous void network. The body continues to exist and does not develop due to its thickness. If the columns persist, it is expected that increasing the plasma power will reduce column formation or at least promote tapered, cauliflower-like growth. What actually happens, however, is a systematic reduction in void size and the persistence of the same uniform orientation columns.
[0040]
When changing the plasma power of the system, it was discovered that the pressure was not necessarily independent. If the incident microwave power is above a certain threshold level for a given process pressure and gas flow rate, the hydrogen and silane ECR plasma used in this demonstration will be stable. At or just above the threshold level, all incident power is absorbed by the plasma. If the power level is further increased above the threshold level, the absorbed power changes little and the excess power is reflected. That is, the net microwave power absorbed and the process pressure are strongly coupled. FIG. 3 shows the trajectory of the power level absorbed when the pressure is changed using a special high-density plasma system with hydrogen and silane flow rates of 40 sccm and 2 sccm, respectively.
[0041]
FIG. 4 shows the measured optical reflectance spectra for a set of films deposited at 100.degree. Here again, the various process pressures correspond to certain power levels, as given in FIG. The peak at about 275 nm is a peak that characterizes the crystallinity of the film. The spectrum shows above 400 nm interference due to reflected light from the substrate, which attenuates towards shorter wavelengths with increasing optical absorption. Therefore, the following can be estimated from the interference-free reflectance of the high porosity film shown in FIG. 4 based on the dielectric mixing theory. That is, with other deposition parameters being constant, the porosity monotonically increases with process pressure (and, therefore, when power is reduced according to FIG. 3). The inset in FIG. 4 is a graph showing the calculated porosity based on the 380 nm reflectivity.
[0042]
FIG. 5 shows AFM surface images of three films deposited at 100 ° C. and various pressures (6, 8 and 10 mTorr). Here, the projecting feature indicates a columnar cluster. AFM chips cannot follow a sharp drop between clusters due to their size and shape. From FIG. 5, it can be seen that the cluster density increases systematically with power (ie, due to a decrease in pressure). As shown in FIG. 6 and discussed below, photoluminescence (PL) was observed for all samples in FIG. Therefore, the columnar size of all these samples should be in the range of tens of nm. 5 and 6 that when the pressure is reduced, the average porosity between the cluster and the columnar body is reduced.
[0043]
As with conventional electrochemically etched porous Si films, our arrayed void-columnar films exhibit photoluminescence. The photoluminescence spectra of the three films deposited at 100 ° C. and 6, 8, and 10 mTorr are shown in FIG. Here, the band peak at ~ 1.8 eV with a half width of ~ 0.3 eV is similar to that observed for porous Si (L. Tsybeskov, MRS Bulletin, April 1998. 33 (1998)). ).
[0044]
Magnetic field near the substrate
HDP systems can create a magnetic field (MF) at or near the substrate location. This is also the case for the present ECR high density plasma deposition system. It was shown that the magnetic field strength near the substrate affected the film porosity. A study of the application of a magnetic field to the film morphology has shown that controlling the magnetic field at the deposition surface is useful for increasing the film porosity when all other deposition factors are kept constant. Control of the magnetic field perpendicular to the deposition surface is also more useful for increasing film porosity. From the scanning electron micrograph of FIG. 7, it was found that when no magnetic field was applied to the deposition surface and immediately above the deposition surface, the structure became more densely packed, that is, the porosity was reduced. This effect of the magnetic field is not even taken into account in the structure zone model of the film form.
[0045]
Deposition gas and flow rate
As shown in Table 1, the flow rate range was examined for hydrogen and silane precursor gases. It was found that the gas composition and flow rate used for depositing the columnar film affected the morphology. Of course, the gas used for the deposition of silicon, for example, must contain silicon-containing molecules, but in addition, hydrogen is also important for obtaining the unique columnar and continuous void structure of the film. It turned out that. For example, from the literature, it is expected that increasing the plasma power will yield more hydrogen radicals, and thus increase the porosity with the plasma power. However, the inventor has found that the porosity is actually reduced by the plasma power and has shown that complex interactions occur. Such a role of hydrogen and power on the morphology of porous deposited films has not been observed and has not been reported in the literature.
[0046]
Hydrogen radicals are thought to play an important role in the nucleation process that determines the position of columns growing in continuous voids perpendicular to the transition layer. Hydrogen radical / bombarded species interactions lead to nucleation behavior that is unexpected from those discussed in the literature. The nanostructure shown in FIG. ¹¹ cm ^-2 Corresponds to the nucleus density of the columnar formation site. This density is the density found in nanocrystalline silicon deposited by RF-PECVD using hydrogen diluted silane (ie, 3 × 10 ¹⁰ cm ^-2 ) (HV Nguyen and RW Collins. Phys. Rev. B47. 1911 (1993)). The high nuclear density, which directly affects porosity and results from high power, was usually not considered at all in the structure zone model to guide morphological understanding. In the model proposed here, this occurs through a hydrogen radical forming effect. Column growth nucleus density is 10 ¹¹ -10 ¹² cm ^-2 It is.
[0047]
Chamber conditioning
The cleanliness of the chamber obviously affects the film yield. For example, particles in the chamber are detrimental to film quality. Also, the chemical state of the chamber walls (ie, the film coating) also affects the film. There is a fundamental chamber plasma species interaction that is unique to or results from using a high density plasma for the deposition of a porous film. These wall-species interactions affect kinetics and species balance during film growth, thereby affecting porosity. For example, a silicon nitride coating on the chamber walls is beneficial for film reproducibility and film quality after each deposition. H before deposition of void / pillar network ₂ Or O ₂ Is another example of conditioning. In this case, the plasma volatilizes the species on the chamber walls (H ₂ Or O ₂ ) Or oxidize (O ₂ ), A stable and reproducible environment for subsequent deposition of porous silicon.
[0048]
Substrate surface
The state of the surface on which the porous film is deposited affects the film formed. For example, when comparing silicon wafer, indium tin oxide (ITO), untreated glass, metal coated glass, and silicon nitride coated glass, a difference in porosity appears in the deposited film. The films on ITO and untreated glass are very similar, with slightly lower porosity compared to nitride coated glass. That is, a higher porosity is associated with a larger structural mismatch between the film and the substrate, resulting in a lower nucleus density.
[0049]
The membranes of the present case exhibit many properties that have only been reported so far for porous silicon made electrochemically. For example, the sensitivity to gases and vapors reported for electrochemically etched porous silicon is shown. (I. Schema et al., Anal. Chem. 67, 3727 (1995)). As a preliminary study, the sensitivity of our membrane to water vapor was studied. FIG. 8 shows the typical behavior of a membrane in response to a change in RH. Depending on the pore size, a weak response was seen up to the threshold humidity level, above which a sharp increase (generally exponential) was observed. The start of this rapid change is controlled by the pore size and the stabilization / sensitization process. In the case of FIG. 8, a Pd layer was deposited on the porous deposition Si, and an annealing treatment was performed at 600 ° C. for 1 hour.
[0050]
The increase in conductivity results in capillary condensation of the water inside the void. Behavior similar to that of ceramics was observed, and the charge transport was proton (H ⁺ ), Hydronium ion (H ₃ O ⁺ ) Or hydroxyl ions (OH ⁻ ). (T. Hubert, MRS Bulletin, 49 (June 1999); BM Kulwicki, I. Am. Ceram. Soc. 74, 697 (1991)). Once the silicon surface interacts with humidity, the formation of chemically bound hydroxide ions causes a large charge density and a strong electrostatic field. Therefore, water molecules physically adsorbed on the chemisorption layer are easily dissociated into ions by the high electrostatic field (2H ₂ O ← → H ₃ O ⁺ + OH ⁻ ). (T. Hubert. MRS Bulletin, 49 (June 1999); BM Kulwicki, J. Am. Ceram. Soc. 74.697 (1991)). Proton transport is H ₃ O ⁺ But H ₂ Protons are released to O and H ₃ O ⁺ (Grotthus chain reaction) (T. Hubert, MRS Bulletin, 49 (June 1999), BM Kulwicki, J. Am. Ceram. Soc. 74, 697 (1991)). The rapid increase in the rate is believed to be related to the growth of the water molecular layer with increasing RH (JJ Mares et al., Thin Solid Films 255, 272 (1995)). Increasing the saturation followed by a sharp increase in conductivity, which is believed to occur when the void volume is completely filled with condensed water, so that for a given surface tension or adsorption coefficient Larger voids or pore sizes, as shown by Kelvin's relationship to capillary condensation. Te, saturation is believed to occur at a greater RH (T. Hubert MRS Bulletin 49 (June 1999);... J.J. Mares et al Thin Solid Films 255, 272 (1995)).
[0051]
FIG. 9 shows the stabilized humidity response of three films deposited at 100 ° C. and various pressures (6, 8 and 10 mTorr). Among these films, the film deposited at 10 mTorr (the highest porosity film) shows the most regular behavior. FIG. 10 is a graph showing the results of repeated measurements using this high porosity film at intervals of minutes, hours, and days. Reproducibility is clearly important. The data in FIGS. 8 to 10 are for the film subjected to the stabilization and sensitization processing, but such processing is not necessarily required.
[0052]
Mares et al. Reported an S-shaped curve similar to that of FIG. 8 using porosity of electrochemically produced porous Si and varying the RH from 0 to 100%, showing a 3-digit increase in conductivity. (JJ Mares et al., Thin Solid Films, 255. 272 (1995)). However, their measurements are very unstable and take several days to get a stable reading. The reason is that the current can be transported only in the thickness direction of the porous silicon layer due to the presence of the silicon material unique to the porous silicon layer produced electrochemically. If a lateral electric field is applied to the sample, this will cause current to shunt through the silicon wafer, making lateral monitoring of electrical properties impossible. Therefore, a conventional sandwich monitoring structure in which porous silicon is sandwiched between two electrodes is required. This is a major drawback because the water vapor must pass through the upper electrode (contact) before reaching the porous semiconductor layer. As a result, the sensitivity is greatly reduced with the response time.
[0053]
On the other hand, a porous film can be easily formed on an insulator, which is easily accomplished with a deposited void / column network structure, so that the electrical properties can be easily monitored laterally with a lateral contact configuration. be able to. This structure allows a direct interaction between the atmosphere and the film to be measured in the region between the contacts. As a result, when RH suddenly changes from 0 to 100% or vice versa, the membrane will exhibit a response time on the order of 0.1 seconds or less. Furthermore, in the present embodiment, since the voids are interconnected, the columnar bodies are separated even in the case of a low porosity, and the continuous voids are arranged vertically to the film-atmosphere interface over the entire film. It has a mesh structure. This facilitates uniform and rapid mass transfer into or out of the membrane. Porous silicon films produced electrochemically cannot have such features because the voids are not necessarily interconnected.
[0054]
FIG. 9 shows how the RH at which saturation occurs shifts according to different void sizes of the three films. From the discussion of capillary condensation above, it is speculated that an increase in void size is associated with an increase in process pressure (decrease in power). This is entirely consistent with our conclusions from AFM and PL data. Therefore, it can be seen that the behavior of the conductivity of the porous membrane with respect to RH provides valuable information about the morphology. If the void volume is all filled with condensed water, the 10 mTorr sample is expected to show the largest current because the porosity is the largest and thickest. However, the reverse has actually occurred, as can be seen from FIG. 9 (from 6 mTorr to 10 mTorr), the saturation current decreases with void volume (proportional to the product of porosity and film thickness). This may be because in the present film, the dominant charge transport does not occur in the bulk liquid but occurs near the silicon surface where water is most effectively dissociated by the surface charge. The ionization ratio at the surface is estimated to be only about 1%, which is 6 orders of magnitude higher than the value in liquid water. (BM Kulwicki. J. Am. Ceram. Soc. 74, 697 (1991)). Thus, the 6 mTorr sample with the highest columnar density exhibits the highest saturation current due to the highest internal surface area. On the other hand, the 10 mTorr film not only has the smallest internal surface area but also has the largest average distance between the columnar bodies. As the spacing increases, the ion concentration decreases more from the value at the surface toward the center of the spacing, adding a greater series resistance.
[0055]
In-situ or external post-deposition treatment can be used to adjust morphology, sensitivity and properties. For example, FIG. 11 shows how the porosity is increased by performing hydrogen plasma etching after deposition / etching. The increase in porosity with plasma exposure time results in a decrease in diameter of silicon pillars, clusters, or both. As mentioned above, the film can be modified by high temperature treatment, chemical interaction, subsequent deposition, wet treatment, and / or plasma treatment. As a further example, the surface of the porous membrane can be terminated with F or Cl to avoid degradation. The membrane surface can be functionalized by chemical, biological or plasma treatment and the sensitivity to specific detection targets other than water vapor can be improved.
[0056]
These porous void-columnar network silicon films with controlled porosity provide isolation, separation layers, void formation and confinement due to their ability to be rapidly, reproducibly and controllably etched. ) Can be used for various applications. They can be used for chemical, electrical, physical, and mechanical isolation. We refer to all types of these applications as air gap, isolation or release layer applications. For example, in the case of electrical isolation, these films are oxidized and ₂ A two-dimensional array of pillars (layer 3 in FIG. 12a) is formed. Air / SiO obtained ₂ The material will have a very low dielectric constant, which is required if capacitive coupling must be minimized. One particular application is in microelectronic wiring. Also, the cap layer can be deposited on a two-dimensional array, as seen in FIG. 13 showing two 2-D alignment layers (layer 3). The electrical connection can be in layer 1, the base layer, and / or the substrate. Interconnection between these layers can be achieved through the porous layer by doping or other conductors.
[0057]
Such an electrical connection or circuit is a two-dimensional (2-D) air / SiO. ₂ Due to the excellent electrical insulating properties of the array, i.e. the low dielectric constant of the porous insulating layer, it will have minimal electrical capacitive coupling to conductors or circuits on or in the substrate or other layers . For example, when the porosity of the silicon oxide bridge structure is 70%, the effective dielectric constant is 1.84.
[0058]
Mechanical, physical and chemical isolation can be obtained by using a deposited film with a controlled structure provided by a void-columnar network and removing material in selected areas. For example, the cap layer 1 is made of SiO as silicon or by a chemical reaction. ₂ , Deposited directly on the film 3 which will be a metal silicide or other silicon compound. The cap layer can be formed with or without breaking the vacuum, and as shown in FIG. 12, a nitride, oxide, metal or another non-porous (ie, continuous) silicon layer (cap 1). Is formed so as to cover a 2-D void / columnar network structure (film 3) formed on a substrate 5 (for example, semiconductor, glass or plastic). As shown in FIGS. 12 (b) and 12 (c), a conventional approach, such as lithography, is performed before or after the cap is deposited to form the etch access holes 10. These holes 10 allow the etchant to be porous void / columnar network silicon (or SiO 2). ₂ , Etc.), and as shown in FIG. 12 (b), enables removal in a predetermined area. As a result, the isolation or confinement region is defined by the removal of the porous material. The porous silicon void / column network material can be easily removed by this etching. This is because all regions of porous material silicon are accessible by a continuous network of voids or pores between the columns that allows access of the etchant and removal of reaction products. FIG. 14 shows an actual “air gap” structure produced by removing such a void / column network Si region. It can be seen that this process creates a cavity between the insulator layers. In this embodiment, the cavities were formed with a 20 minute etch. Conventional approaches to forming cavities using polycrystalline silicon have required orders of magnitude or more of time, or have required the use of complex electron beam lithography to accelerate etching (PJ French. J. Micromech.Microeng., 6, 197 (1996)).
[0059]
The removal area (cavity or air gap) shown in FIG. 14 can be used for diaphragms, accelerometers, bolometers, and reduced capacitive coupling, as well as applications such as chromatography, DNA sorting, optical applications, and fluidics. Can be used as a tube.
[0060]
FIG. 15 shows an example of use of a porous deposited film in another application (support substrate for mass desorption spectroscopy). Here, the porous deposited silicon acts as a medium for attaching subsequently detected molecules. For detection, a laser for desorbing and ionizing molecules adsorbed on porous silicon is used. The ions resulting from the desorbed molecules are then identified by mass spectrometry. Some molecules intentionally adsorbed on the porous deposited silicon film are identified, as seen in the mass spectrum of FIG. The use of a laser followed by mass spectroscopy to desorb molecules from a substrate has previously been performed, but required an organic medium to bind the molecules to be detected. This is the so-called mass analysis by laser desorption and ionization technique ALDI. In our technique, no organic medium binding material is required. As is evident, desorption and ionization occur by laser irradiation, allowing the identification of molecules. Electrochemically etched porous silicon substrates have also been shown to be effective in this laser desorption spectroscopy in a limited mass range. Further, a binding material is not required (Nature. Vol. 399, p. 243-246, (l999)). However, this etched porous silicon approach is associated with the problems described above. That is, an electrical contact on the back side is required for electrochemical etching, and thick starting silicon is required to obtain the nanostructures required for binding of the molecules to be detected. On the other hand, our films have deposited nanostructures, do not require an etching step, and can be deposited on any substrate, including plastic.
[0061]
In all of these and other applications, the approach described here for producing porous deposited silicon has distinct advantages over electrochemical wet etch materials. These advantages fall into four main groups: (1) substrate diversity, (2) the ability to adjust porosity and film phase even for thin films, (3) consistency of thickness and porosity. And (4) the flexibility of the plasma treatment.
[0062]
With the low temperature direct deposition approach of the present invention, porous semiconductor films can be formed on any substrate, including insulators, metal foils and plastics. The film may or may not be doped, and may be of any thickness greater than about 10 nm, but is preferably between 10 and 20 nm. In the electrochemical approach, a porous semiconductor layer can only be obtained on a conductive substrate or layer. The starting material must be doped and must be thick enough to obtain the required morphology. Also, with conventional deposition approaches, the morphology varies with film thickness, and porosity comparable to ours can be obtained only after a separate etching process. Then, it is always amorphous. On the other hand, the simultaneous processing of etching and deposition using plasma-based processes and the parameters and conditioning we employ enables a dry process to control the porosity and phase of thin films deposited on any substrate .
[0063]
As mentioned above, the fact that these films can be deposited on insulating structures facilitates their use in sensor structures. Furthermore, the nanostructure of the present invention, which is arranged perpendicular to the membrane surface and has a continuous and uniform void network and void volume, facilitates mass transfer to and from the membrane, and is only applicable to mass desorption. It is also used in applications such as electrical sensors, electrophoresis and chromatography. The features of the non-tapered columnar and arrayed void-columnar network morphology with relatively uniform and continuous voids are only found in the porous silicon of the present invention. Thus, the membranes of the present invention are immobilized and immobilized for adsorption, desorption, transport, and immobilization applications (eg, identifying, sorting, immobilizing species such as DNA, and contacting molecular electronics molecules, etc.). ). By the direct deposition approach, porous semiconductors can also be fabricated on various optical material substrates such as transparent, polarizable, and the like. This will facilitate and expand the use of porous semiconductors in optical applications such as anti-reflection, optical cavities, and light trapping coatings. Also, if the porous semiconductor is deposited on a flexible polymer substrate, a non-planar, bent, porous semiconductor layer becomes available.
[0064]
This approach takes advantage of the advantages of plasma-based process technology. For example, the chemical composition of the present high porosity silicon film can be controlled by a precursor gas incorporated into the plasma during the co-deposition / etch process. Thereby, not only intrinsic, n-type or p-type porous Si but also porous Ge, GeC, SiC, SiO, SiF, SiCl and the like can be manufactured. In addition, the chemical composition of the film can be changed during deposition and, therefore, throughout the film thickness. In contrast, in the case of electrochemical etching of a previously grown material, it is not possible to determine the chemical composition of the porous silicon layer except where it will eventually become a surface. Also, control of the surface composition is limited by the electrolyte used. Further, the film of the present invention is grown in a high vacuum, so that contamination due to exposure to the atmosphere can be prevented.However, conventional porous silicon films are easily contaminated because they are treated in a wet chemical. I will.
[0065]
Whether employing the electrochemical etching approach or the plasma deposition approach of the present invention, the resulting porous silicon material is terminated with hydrogen atoms upon completion of fabrication. However, under ambient conditions, the hydrogen terminated surface will undergo significant room temperature oxidation and hydrogen loss. Both of these mechanisms degrade the physical properties of porous silicon. Therefore, if stable surface passivation is immediately performed, deterioration of the porous silicon film can be suppressed. This can be done in a separate chamber for coating, dipping, vapor exposure, etc., using a multi-chamber system, by vacuum processing, or by switching the chamber atmosphere after film deposition / etching, in situ. , Can be achieved. In this regard, the high-density plasma approach of the present invention is very flexible because it can facilitate any type of reaction with a lower activation energy and facilitate plasma surface treatment. In general, the surface chemistry can be controlled during or after deposition. As an example, the F- or Cl-terminated surface of the porous membrane can be formed in situ, and such a surface is considered to be highly resistant to degradation. Of course, the porous membrane can be treated externally (ex situ) and can be made hydrophobic or hydrophilic by surface treatment. Further, the film of the present invention can be functionalized by surface treatment.
[0066]
While several embodiments have been described in connection with the present invention, it will be apparent to one skilled in the art that obvious changes may be made. Accordingly, the invention is not to be limited by the details set forth but to include changes and modifications within the scope of the appended claims.
[Brief description of the drawings]
FIG.
It is a TEM photograph of the cross section of the void / columnar network structure silicon thin film on glass.
FIG. 2
5 is an XRD pattern of a void / columnar film deposited at 6, 8, and 10 mTorr and 100 ° C.
FIG. 3
5 is a graph showing the relationship between absorbed microwave power and process pressure when the hydrogen and silane flow rates are 40 sccm and 2 sccm, respectively.
FIG. 4
FIG. 4 is a graph showing the reflectance monitored by a barium sulfate reflector for films deposited at various process pressures of 5-12 mTorr and 100 ° C. FIG. The inset shows the calculated porosity.
FIG. 5
(A) is an AFM surface image of a void / columnar film deposited at 100 ° C. and 6 mTorr. (B) is an AFM surface image of the void / columnar film deposited at 100 ° C. and 8 mTorr. (C) is an AFM surface image of the void / columnar film deposited at 10 mTorr and 100 ° C.
FIG. 6
5 is a photoluminescence spectrum of a void / columnar silicon film deposited at 100 ° C., 6, 8, and 10 mTorr.
FIG. 7
5 is an SEM photograph showing an effect of a magnetic field near a substrate on a structure. (A) SEM photograph of a film deposited on ITO (indium tin oxide) by applying a magnetic field of 100 G perpendicular to the substrate at 100 ° C. and 10 mTorr, and (b) a film deposited on the substrate without applying a magnetic field. It is. The calculated values of the porosity based on the reflectance data are 0.62 and 0.22, respectively.
FIG. 8
4 is a graph showing an increase in conductivity (current flowing at a fixed voltage of 10 V) of a void / columnar silicon film corresponding to an increase in relative humidity (RH). The film was deposited at 8 mTorr and 100 ° C., and was subjected to a Pd stabilization treatment before use. In order to improve reproducibility, the measurement was performed after depositing a Pd layer having a thickness of 100 ° and annealing at 600 ° C. for 1 hour. The applied voltage is 10V.
FIG. 9
6 is a graph showing how the conductivity of a void / columnar silicon film having different porosity changes according to a change in relative humidity (RH). Films were deposited at pressures of 6, 8, and 10 mTorr. The applied voltage is 50V. The measurement was performed after the Lewis acid treatment. These films are exposed to a Lewis acid prior to use.
FIG. 10
4 is a graph of current vs. relative humidity showing reproducibility and stability, which was measured at an applied voltage of 50 V using a void / columnar silicon film deposited at 100 ° C. and 10 mTorr. The membrane was exposed to a Lewis acid prior to use.
FIG. 11
It is a graph which shows the porosity increase of a void / columnar silicon film seen from a reflectance spectrum by the hydrogen plasma post-deposition process for 5 minutes and 15 minutes.
FIG.
(A) shows a basic structure useful for creating an isolation or confinement region by removing a porous material. The structure includes a cap layer, nanostructured columns, a base layer and a substrate. (B) is a plan view of an isolation region made from the material of (a). (C) is a CC sectional view of (a).
FIG. 13
Figure 3 illustrates several layers in a basic structure useful for creating an isolation or confinement region using a porous material. Two porous layers, a cap layer, a base layer and a substrate are shown.
FIG. 14
Fig. 4 shows a tube or air gap made by removing porous silicon in a void / column network structure.
FIG.
It is a laser desorption spectrum. Sample drops were applied directly to the porous silicon membrane and dried prior to laser irradiation. The molecules in the sample have been identified.

Claims (49)

A porous membrane having a porosity of up to about 90% and comprising a plurality of perturbations extending into a void. The porous membrane according to claim 1, wherein the plurality of perturbations are arranged substantially perpendicular to a substrate. The porous membrane according to claim 1, wherein the plurality of perturbations are disposed substantially perpendicular to a base layer. The porous membrane according to claim 1, wherein the plurality of perturbations are rod-shaped columns. The porous membrane according to claim 1, wherein the perturbation is semiconducting. The porous membrane according to claim 1, wherein the perturbation is polycrystalline or amorphous. The porous membrane according to claim 6, wherein the polycrystalline or amorphous phase is a silicon material. The porous membrane according to claim 1, wherein the thickness of the membrane is greater than about 10 nm. The porous membrane according to claim 1, wherein the perturbation has a diameter of about 1 nm to about 50 nm. The porous membrane according to claim 9, wherein the perturbation has a diameter of about 3 nm to about 7 nm. The porous membrane according to claim 1, wherein the perturbation is in a cluster having a diameter of about 50-500 nm. The porosity of the porous film can be controlled by at least one of a substrate coating film, plasma power, process pressure, magnetic field, plasma-substrate bias, chamber conditioning, deposition gas, flow rate and deposition temperature. The porous membrane according to claim 1, wherein A porous film having a porosity of up to 90% and comprising a plurality of perturbations extending into voids, wherein the porous film is formed on the substrate; Composite structure. The composite structure according to claim 13, further comprising a substrate coating layer, wherein the porous film is formed on the substrate coating layer. The composite structure according to claim 14, wherein the substrate coating layer is at least one coating material selected from the group consisting of an organic insulator, silicon nitride, and silicon oxide. The composite structure according to claim 14, wherein the coating layer is at least one active material selected from the group consisting of a piezoelectric, a ferroelectric, a metal, and a semiconductor. The composite structure according to claim 13, further comprising a cap layer, wherein the porous film is formed between the cap layer and the substrate. The composite structure according to claim 17, wherein the cap layer is at least one insulating material selected from the group consisting of an organic insulator, silicon nitride, and silicon oxide. The composite structure according to claim 17, wherein the cap layer is at least one active material selected from the group consisting of a piezoelectric material, a ferroelectric material, a metal, and a semiconductor. 14. The composite structure of claim 13, wherein the thickness of the porous membrane is greater than about 10 nm. The composite structure according to claim 13, wherein the porous film is polycrystalline or amorphous. 22. The composite structure according to claim 21, wherein the polycrystalline or amorphous porous film has a structure in which rod-like perturbations are two-dimensionally and periodically arranged. 23. The composite structure according to claim 22, wherein the diameter of the rod-shaped perturbation is about 1 to 50 nm. 23. The composite structure according to claim 22, wherein the rod-shaped perturbations are present in clusters having a diameter of about 50 to 500 nm. 14. The composite structure according to claim 13, wherein the substrate is selected from the group consisting of glass, metal foil, insulating material, plastics, and semiconductor-containing material. A method for producing a composite structure comprising a substrate and a porous film, comprising a step of depositing a porous film on a substrate at a temperature lower than 250 ° C. by a high-density plasma deposition method. 27. The method of claim 26, wherein the porous membrane has a porosity of up to about 90% and is comprised of a plurality of perturbations extending into voids. 27. The method of claim 26, further comprising etching the porous membrane. 29. The method of claim 28, wherein said depositing step and said etching step are performed simultaneously. 29. The method of claim 28, wherein the etching is performed with hydrogen, chlorine, fluorine, HC1, HF and their induced radicals. 27. The method of claim 26, wherein the high density plasma deposition is performed in a precursor environment comprising a silicon-containing gas. The method of claim 31, wherein the precursor environment includes at least one gas selected from the group consisting of hydrogen and a silicon-containing gas. 33. The method of claim 32, wherein said silicon-containing gas is silane. The method of claim 26, wherein the depositing step is performed with a magnetic field near the substrate between +800 and -600 Gauss. The method of claim 26, wherein the depositing step is performed with a microwave excitation frequency of about 100-1200W. The method of claim 26, wherein the depositing step is performed without applying a voltage between the plasma and the substrate. The method of claim 26, wherein the composite structure further comprises a substrate coating layer, wherein the porous membrane is formed on the substrate coating layer. The composite structure according to claim 37, wherein the substrate coating layer is at least one coating material selected from the group consisting of an organic insulator, silicon nitride, and silicon oxide. The method of claim 37, wherein the coating layer is at least one active material selected from the group consisting of a piezoelectric, a ferroelectric, a metal, and a semiconductor. The method of claim 26, wherein the composite structure further comprises a cap layer, wherein the porous membrane is formed between the cap layer and the substrate. The method of claim 40, wherein the cap layer is at least one insulating material selected from the group consisting of an organic insulator, silicon nitride, and silicon oxide. The method of claim 40, wherein the cap layer is at least one active material selected from the group consisting of a piezoelectric, a ferroelectric, a metal, and a semiconductor. 27. The porosity can be controlled by at least one of substrate coating, plasma power, process pressure, magnetic field, plasma-substrate bias, chamber conditioning, deposition gas, flow rate and deposition temperature. the method of. 27. The method of claim 26, wherein at least a portion of the porous layer in the porous membrane is etched away, thereby forming an air gap, release or isolation structure. A porous structure comprising a substrate and a porous film having a porosity of up to 90% and comprising a plurality of perturbations extending into the voids, wherein said porous film is formed on said substrate. Features sensor. The sensor according to claim 45, wherein the sensor is capable of monitoring lateral resistivity, optical or dielectric response. A porous structure comprising a substrate and a porous film having a porosity of up to 90% and comprising a plurality of perturbations extending into the voids, wherein said porous film is formed on said substrate. Characteristic gas detector A porous structure comprising a substrate and a porous film having a porosity of up to 90% and comprising a plurality of perturbations extending into the voids, wherein said porous film is formed on said substrate. Characteristic analyzer. 49. The analyzer according to claim 48, wherein the device is capable of performing desorption mass spectroscopy.

Images