兩個介觀物理的主題:壹： 低溫電子散射時間 貳：量子點中自旋傳輸

Abstract

在這本博士論文中，我們將討論兩個介觀物理的主題。第一個是銅鍺金合金薄膜在低溫下電子的非彈性散射時間的研究、第二個是電子在砷化鎵垂直雙量子點中自旋傳輸的研究。 電子非彈性散射時間是一個度量電子處在其基態時間長短的物理量。根據費米液體理論（Fermi-liquid theory）的預測，電子在絕對零度時，電子的非彈性散射時間會是無窮長。然而，多年來眾多的實驗結果顯示，當溫度低於某個溫度後，電子的非彈性散射時間會呈現一個不隨溫度改變的定值。這個奇異的現象吸引了許多理論和實驗學家的注意。有些理論學家認為不同於費米液體理論的預測，這個現象是一個新的本徵物理性質。一開始在這 方面有非常多的爭議，許多研究的結論漸漸地澄清了這個現象。經過長時間的研究，大部分的人都相信樣品中磁性雜質的存在將會導致一個不隨溫度變動的散射率，這等效於實驗上常常被觀測到的不隨溫度變動的非彈性散射時間。在此之後，大不分的人都會將觀測到奇異的非彈性散射時間歸因為近藤效應 （Kondo effect）。我們也做了一些銅鍺金合金 （Cu93Ge4Au3）薄膜的電子非彈性散射時間研究。我們的結果顯示了三個非常特異的現象：第一、對於不規則程度不同的樣品中，所有的樣品都在１０度（１０Ｋ）跟６度（６Ｋ）之間呈現一個不隨溫度變動的非彈性散射時間，而且對於所有的樣品在這區間非彈性散射時間都是相同的。第二、當溫度低於６度（６Ｋ）時，非彈性散射時間急速的增加而且增加的速率跟樣品的不規則程度有關。對於一個不規則程度較高的樣品，非彈性散射時間增加的速率比較快。第三、在１０度（１０Ｋ）到３０毫度（３０ｍＫ）區間，外加高達１５Ｔ的磁場依舊對電阻全無影響。所有的證據顯示動態結構缺陷效應（dynamic structure defeat effect）主宰整個系統行為。我們的結果是第一個有系統分析這個效應的研究。 這幾年，因為在量子資訊上潛在的應用，量子點中電子自旋的傳輸吸引了非常多研究上的注意。我們也做了兩個垂直量子點的題目。 第一個是有關於自旋選擇法則。我們量測了銦鎵砷（In0.05Ga0.95As）垂直雙量子點的傳輸頻譜。利用大偏壓法（large source-drain voltage），從雙電子到三電子傳輸基態跟激發態的頻譜同時可以被觀測到。在觀測到的頻譜中，從１Ｓ２單重態到１Ｓ２Ｐ三重態的基態過渡在５Ｔ被觀測到。在高於５Ｔ的磁場下，可以清楚的看到黎蔓分離（Zeeman splitting），而ｇ值（g factor）是０.３６。藉由自旋選擇法則我們可以解釋在從雙電子到三電子的傳輸中只有兩條黎蔓分離線而不是三條黎蔓分離線．對於躍遷前後的電子自旋數大於１╱２是不被允許的。因為電子在自旋雙態（doublet state）的遲逾時間（relaxation time）遠大於電子穿隧傳輸的時間，所以上自旋（spin up）跟下自旋（spin down）都可以成為傳輸的起始態。總共會有四個可能的傳輸貢獻，但只有兩個有效能量可以在傳輸頻譜上被觀測到。 第二個主題是有關於在黎蔓非吻合的量子點中的自旋傳輸。我們量測了在不同ｇ值的雙量子點的穿隧電流（tunneling currents）。結果是完全的不同於相同ｇ值的雙量子點的穿隧電流。特別的，因為兩個電子點間黎蔓不吻合的穿遂，人們預測兩個分裂的電流峰將會被觀測到。另外，聲子（phonon）的散射強烈地影響到傳隧的電流值。在弱的聲子散射只有上自旋電子可以共振（resonance）地穿遂，而下自旋電子則不行。除此之外，帶寬（bandwidth）共振穿遂電流峰在這個黎蔓不吻合的系統中被觀測到。

We report two mesoscopic topics in the thesis. First one is about the low temperature dephasing time in Cu93Ge4Au3 thin films and second one is about the spin transport in InxGa1-xAs (GaAs) vertical double quantum dots. The electron dephasing time is a time scale that how long an electron can stay at its eigenstate. Fermi-liquid theory predicts that the life-time of an electron at the Fermi surface at T = 0 is infinite. However, many experiments show that the dephasing times are always constant at low temperature. The anomalous low temperature desphasing time catches many theorists' and experimentalists' interest. Some theorists propose that the saturating low temperature dephasing time is intrinsic physics which is contrast to the Fermi-liquid theory. It makes a lot of controversy on the phenomenon. Many works were done to clarify the physics. After a lot of studies on the field, people believe that magnetic impurities would induce a constant scattering time which is equal to the observed saturating dephasing time. After that, people often refer the anomalous low temperature dephasing time to the Kondo effect. We study the low temperature dephasing time in Cu93Ge4Au3 thin films. There are three distinct features. First one is that the dephasing time is a constant value between 10 K and 6 K for all of measured films with different levels of disorder. Second one is that the dephasing time increases drastically as temperature is lower than 6 K. The increasing rates depend on the levels of disorder. For a more disordered film, the increasing rate is more drastic. Third one is that the temperature dependent resistance from 10 K down to 30 mK is insensitive to the magnetic filed up to 15 T. All of the results support that the dynamic structure defeat effect dominates the behaviors. Our experiment is the first systematic study on the dynamic structure defeat effect. The electron spin transports in quantum dots have caught considerable increasing of interest because of potential development on quantum information. We have done two subjects on spin transport in vertical double quantum dots. First subject is about the spin-selection rule. We measured electron transport state spectra of an In0.05Ga0.95As vertical double quantum dot. Both the ground and excited states of transport spectra from two electrons to three electrons are measured using a large source-drain voltage. In the obtained transition spectrum, the ground state transition from the 1S2 singlet state to the 1S2P triplet state is observed at 5 T. Zeeman splitting with a g-factor of 0.36 is clearly observed at magnetic fields higher than 5 T. The observation of two Zeeman sublevels instead of three for the triplet state is explained by the spin selection rules for the SZ components between the two-electron and three-electron spin states. Transition with the total spin difference between the initial and final states larger than 1/2 is forbidden. Because the relaxation time between doublet states is much longer than electron tunneling time, both spin up and spin down can be the initial states from spin transitions. There are four transitions contributing to tunneling processes, but only two energy differences lead to the two Zeeman sublevels in the excitation spectra. The second subject is related to spin transport in Zeeman mismatch double quantum dots. We measure tunneling currents in a vertical double quantum dot with different g factors for the two dots. The results are substantially different from those in a double quantum dot with a homogeneous g factor. In particular, two split peaks are expected to be observed due to the Zeeman mismatch of inter-dot tunneling. In contrast to the case of a homogeneous g factor, the coupling to phonons strongly affects the tunneling current in the system with an inhomogeneous g factor. For weak coupling strengths, only up-spin electrons can resonantly tunnel through the dots while the down-spin tunneling is spin blockaded. Besides, a bandwidth resonance tunneling peak also appears in the system with mismatched g factors.