Tsukasa Kobayashi

Epitaxial Growth of Al on Si by Gas-Temperature-Controlled Chemical Vapor Deposition

Epitaxial Al(111) film was deposited on Si(111) by low-pressure chemical vapor deposition with the use of tri-isobutyl aluminum (TIBA) at the substrate temperature of 400°C with the deposition rete of 0.9 µm/min. It was necessary for epitaxy to preheat the TIBA gas just before the deposition on the substrate (gas-temperature controlling). The film surface was very smooth; reflectance was higher than 90%. Streaks were observed in RHEED patterns. The rocking curve measured by X-ray diffraction was very narrow. Analysis by SIMS showed the film contained about 0.1% of Si and 20 ppm of O, C, and H. No hillock appeared on the film after 430°C annealing for 40 min.

Gas-Temperature-Controlled (GTC) CVD of Aluminum and Aluminum-Silicon Alloy Film for VLSI Processing

In order to investigate the feasibility of CVD metallization for IC manufacturing processes, Al and Al-Si film depositions on silicon (100) wafers were studied using a precursor gas of triisobutylaluminum with and without disilane. A new mechanism for the gas temperature controller (GTC) was provided in the process chamber so as to activate the gas thermally just prior to its arrival at a heated substrate. The typical deposition rate was 700 nm/min. Without post-annealing, an alloy of Al-0.4%Si was obtained. Specific resistivity was 3.0 µΩcm, and specularity was about 60% at 550 nm for the GTC-CVD films of 1 µm thickness.

Growth of Al films by gas‐temperature‐controlled chemical vapor deposition

Epitaxial single‐crystal aluminum (Al) films were deposited on 4‐in.‐diam silicon (Si) substrates by gas‐temperature‐controlled chemical vapor deposition using tri‐isobutyl aluminum (TIBA). The TIBA was preheated just prior to the deposition by a gas‐temperature controller (GTC). The crystallographic orientation of Al films was determined by the gas flow rate and the degree of thermal decomposition of TIBA. The GTC was thought to work in such a manner as to efficiently generate a desirable intermediate of unknown composition. Four kinds of Al single‐crystalline epitaxial films on Si substrates were obtained: Al(001)/Si(001), Al(001)/Si(111), Al(111)/Si(111), and Al(110)/Si(115). The common epitaxial relationship in these films was Al[110]∥Si[110] in the interfacial plane. Along this direction, four lattice spacings of Al matched three lattice spacings of Si at the interface.

Aluminum Metallization Using a Combination of Chemical Vapor Deposition and Sputtering

An aluminum metallization process that combines blanket chemical vapor deposition and sputtering was developed for use in fabrication of future ultralarge scale integration interconnections. Blanket chemical vapor deposition of aluminum using dimethylaluminum hydride on titanium nitride, which provides superior step coverage and a smooth surface morphology for films of less than approximately 0.15 μm thickness, was only used for hole-filling. Subsequent aluminum alloy sputtering, which has a high deposition rate and provides smooth surface films, was used for the thickening of the aluminum films. This combination process draws on the respective advantages of both chemical vapor deposition and sputtering, which mutually compensate for each others drawbacks. As a result, via holes with a diameter of 0.3 μm and an aspect ratio of 2.7 were successfully filled. The resistance of contact holes fabricated by the combination process was slightly lower than that obtained in the conventional tungsten plug process due to low film resistivity of chemically vapor deposited aluminum. The contact resistivity for contacts to p- and n-type Si were 1.0 x 10 -7 and 2.9 x 10 -8 Ω cm 2 , respectively. Via hole resistance for 0.45 μm diameter holes was less than 1 Ω, which corresponds to a contact resistivity of less than 1.6 x 10 -9 Ω cm 2 between chemically vapor deposited aluminum and the underlayer titanium nitride.

Substrate temperature dependence of hillock, grain, and crystal orientation in sputtered Al–alloy films

The effects of substrate temperature during deposition on hillock, grain and crystal orientation were investigated for Al, Al–Ti, Al–Si, Al–Si–Ti, and Al–Cu films. The films were deposited by magnetron sputtering in the substrate temperature range from room temperature to 350 °C and followed by vacuum annealing at 400 °C for 5 min. Hillock density was varied as a function of the substrate temperature in a different manner for each investigated material. The substrate temperature during deposition is a very important factor for suppressing the hillock formation. A close relationship between hillock density and grain size was observed. This was modified when the substrate temperature was changed. This is probably related to the change of grain texture morphology determined by the x‐ray diffraction measurements.

Epitaxial growth of Al(100) on Si(100) by gas‐temperature‐controlled chemical vapor deposition

Epitaxial Al(100) film was deposited on Si(100) by low‐pressure chemical vapor deposition with the use of tri‐isobutylaluminum (TIBA) at the substrate temperature of 380–400 °C with the deposition rate of 400 nm/min. For the epitaxial film growth, it was concluded that preheating the TIBA gas just before the incidence onto the substrate was very important (gas‐temperature‐controlling). The specular reflectance of the 1‐μm‐thick aluminum film on silicon substrate was the same level as that of the sputtered film. The x‐ray diffraction peak was very narrow and only Al(100) orientation was observed. The transmission electron diffraction (TED) patterns of Al(100) films showed clear spots. Using disilane, we could suppress the diffusion of substrate silicon into the aluminum film.

Epitaxial Growth of Al on Si By Gas-Temperature-Controlled CVD

Epitaxial Al(111) film was deposited on Si(lll) by low-pressure chemical vapor deposition with the use of tri-isobutyl aluminum (TIBA) at the substrate temperature of 400 °C with the deposition rate of 0.9 μm/min. It was necessary for epitaxy to preheat the TIBA gas just before the deposition on the substrate. The preheat was made by gas-temperature-controller provided in the chamber. The film surface was so smooth that reflectance was higher than 90 % of Si and 20 ppm of O, C, and H. No hillock appeared on the film after 430 °C annealing for 40 min. The interface of Al and Si was rather stable so that no alloy penetration occured. The possibility of epitaxial growth of Al(100) on Si(100) was also shown.

Simulation of Erosion Rate Distribution on a Target used in a Magnetron Sputtering System

半 導 体製 造 用,液 晶表 示 デバ イ ス 用 の マ グネ トロン・ スパ ッタ リン グ装 置 は 近年 ます ます 大 型 化す る傾 向に あ る. そ こで は,経 済 性 の 観 点 か ら タ ー ゲ ッ トの利 用 効 率 が常 に 注 目 され て い る. タ ー ゲ ッ トの エ ロー ジ ョン形 状 は, タ ー ゲ ッ トの利 用 効 率 や 膜 厚分 布 等 に直 接 影 響す る た め, これ を 磁 界 分布 等 か ら予 測 で きれ ぽ 装 置 設 計 上 非 常 に有 用 と考 え られ る. マ グネ トロン ・ス パ ッ タ リ ングの タ ー ゲ ッ ト ・エ ロ ー ジ ョン分 布 の 本 格 的 シ ミュ レー シ ョンは, Wendt ら 1) に よ り初 め て報 告 され た. そ こで は,磁 界 中 の 電子 の運 動 を解 析 す る こ とに よ り, 二 次 元 的 な エ ロー ジ ョン分布 が 計 算 され て い る.Sheridanら2)は, 磁 界 中 の 多数 の 電子 の 運動,イ オ ン化 を モ ンテ カ ル ロ法 に よ り扱 い,円 形 カ ソー ドの エ ロー ジ ョンを解 析 した.以 後 同様 な方 法 に よ る解 析 が 種 々報 告 され て い るが3,4,5), そ れ らは 二 次 元 あ る い は 円形 カ ソー ドに限 られ て いた. 最 近, 吉 川 ら6)は Sheridan と同 様 な 方 法 で 矩 形 カ ソ ー ドの 三 次 元 シ ミ ュ レー シ ョ ンを 行 な った. モ ンテ カル ロ法 では, 出 発 電子 の分布 と最終的に得 られるイオンフラックスの分布が矛 盾 しない ように, 繰 り返 し計算を行ない収束解 を求め る 必要がある.こ の様な計算 には多大な計算時間が必要 と な り,近 年手軽に利用出来る ようにな った ワー クステー シ ョン上 での実行が難 しい.今 回,Wendtら1)の 方法を 三次元に拡張 し,収 束解を求めるため の繰 り返 し計算 が 不 要 な方法 を検討 した. また,こ の方法 を15.7×55.88 cm サイズの矩形 プ レーナ ー型マ グネ トロンカ ソー ドに 適用 した ところ,実 測 との良い一致がみ られた.