磊晶鋁薄膜的成長與傳輸特性研究

Abstract

本篇論文主要研究使用分子束磊晶技術 (molecular beam epitaxy, MBE) 成長鋁薄膜。我們將砷化鎵基板在腔體中轉換成鎵表面型 (Ga-rich GaAs) 做為磊晶的樣板。即使鋁與砷化鎵存在超過25%的晶格不匹配，在使用了低溫及高沉積率的成長方式後，我們成長出(110)面向的鋁薄膜，在僅僅成長40奈米的厚度之後，表面就達到次奈米級 (sub-nm) 的平整度 (約0.354奈米)。這樣的平整度可以比擬剛磊晶完成的砷化鎵樣品。我們成長的樣品可以在紫外光波段 (200 ~ 350奈米) 達到0.85的反射率，這在其他的金屬 (譬如金、銀等等) 上是很難達成的。這樣的樣品對於應用在紫外光波段的電漿光學元件具有相當大的發展潛力，同時也可以作為一個新的樣板來沉積其他新穎的材料。除此之外，在厚度小於10奈米的樣品上，樣品表面會形成多孔隙狀 (percolative)，我們可以在3、6、8奈米的樣品上觀測到清楚的弱反侷限現像 (weak antilocalization, WAL)。我們發現在最薄的3奈米樣品上，電子同調性的破壞 (electron dephasing, τi-1) 來自於單純的電子-電子交互作用 (electron-electron dephasing, τee-1)。這在一個由控制掺雜 (modulation doped) 的半導體異質接面形成的理想二維系統中很容易被觀測到，但在處於混合維度的超導金屬材料卻很難被觀測到。為了進一步淬取在3奈米的樣品中，WAL中無法淬取到的電子-聲子能量釋放率 (electron-phonon energy relaxation, τe-p-1)，我們量測了載子的熱效應。為了確定這個方式的可靠度，我們先比較了分別用熱效應與WAL在6與8奈米的樣品上淬取出來的值。我們第一次實驗上證明了兩種分別立基於古典物理及量子物理的量測可以得到一致的結果，不管是溫度的相依性或是數值本身。我們驗證了淬取出τe-p-1數據的可信度。於是我們使用熱效應的量測，來淬取3奈米樣品的τe-p-1。我們觀測到，與前人的量測類似，τe-p-1具有2次方的溫度相依性，然而數值卻小了兩個量級。熱效應的分析除了可以得到無法直接在WAL中被淬取的τe-p-1以外，我們也證明了在WAL存在的溫度區間內，τe-p-1遠小於τee-1。使得在3奈米的樣品中τi-1完全由τee-1主導 (τe-p-1很小) 的這件事，在WAL與熱效應的分析中，都能夠得到驗證。

This dissertation mainly studies the deposition of aluminum thin films using molecular beam epitaxy (MBE) as the deposition technique. A flat Ga-rich GaAs surface transferred in situ will serve as the epi-template. Even there exists over 25% of strain caused from lattice mismatch in this Al-GaAs system. By using high deposition rate and low process temperature, we can get a sub-nm roughness Al(110) thin film with only 40 nm aluminum deposition (roughness = 0.354 nm). The flatness can reach the same level as an as-grown GaAs sample. The reflectivity of the deposit aluminum film can reach 0.85 in the ultraviolet range (200 ~ 350 nm) that is difficult to achieve by using other noble metals (gold, silver, etc). The deposited aluminum film is suitable for the development of plasmonic devices in the ultraviolet range. Additionally, the deposited aluminum film can serve as a new epi-template to deposit new materials. In contrast, the surface of the deposited aluminum film is percolative if the film thickness is less than 10 nm. We have observed clear weak antilocalization (WAL) in 3, 6, and 8 nm samples. We observe a pure electron-electron dephasing in the thinnest 3 nm sample. A pure electron-electron dephasing is generally observed in many ideal two-dimensional systems formed by modulation doped semiconductor heterostructures, but it would be difficult observed in mix dimensional superconducting metallic systems. For further extraction of the electron-phonon energy relaxation rate (τe-p-1) in the 3 nm sample, we have measured the carrier heating effect in the 3 nm sample. We initially compare the extracted τe-p-1 by WAL and heating effect to clarify its reliability in the 6 and 8 nm samples that τe-p-1 could be directly extracted by WAL. We have for the first time experimentally obtained a coherent result of both experimental extractions in its temperature dependence and the value of τe-p-1. Hence the extracted τe-p-1 is convincing. Furthermore, we have extracted τe-p-1 in the 3 nm sample by heating effect. The extracted τe-p-1 has a T2 dependence but two-order smaller than previous experimental observation. The value of extracted τe-p-1 is much small compared to τee-1 and is negligible at all the temperature which WAL exist as we have observed in WAL measurement.