以第一原理與量子傳輸來計算植入特殊離子的鎳鍺化物與N型鍺接面

Abstract

現今技術的發展使得元件越做越小，而傳統使用的矽基金氧半場效電晶體也逐漸接近其微縮極限。因此，尋找合適的替代材料便非常重要。其中，在元素週期表位於矽正下方的鍺，由於擁有比矽更高的載子遷移率，而被認為可以取代矽作為新的通道材料。然而，在金屬與 N 型鍺接面中，由於鍺在價電帶附近有大量的介面態產生，造成非常強的費米能階釘扎效應，使得電子蕭特基位障不容易降低，亦使得特徵接觸阻抗無法降低。因此，如何減緩費米能階釘札效應便成為首要目標。 在我們的研究中，將考慮植入不同特殊離子的鎳鍺化物與鍺接面，計算其電子結構與量子傳輸，並且鎳鍺化物與鍺的晶面方向分別選擇(112)與(001)面，考慮不同的植入離子包含了硒、鋁、錫、鈧、鈦、釩、硫、氯、鉑等離子。計算結果可分為三個部分：第一個部份，經由密度泛函理論搭配局域密度近似法，討論植入離子是否會析離於鎳鍺化物與鍺接面的介面處；第二個部分，經由密度泛函理論搭配雜化泛函法，討論植入離子在其穩定位，是否有降低電子蕭特基位障的效果，我們從結果得知，硒、釩離子較有降低蕭特基位障的效果，並分別降低 0.175、0.097 eV，且發現它們都會析離於介面並傾向聚集在半導體端，另外，實驗方面也對硒離子作了量測與分析，並得到降低蕭特基位障0.15 eV 左右的結果；第三個部分，經由密度泛函理論與非平衡格林函數理論搭配mBJ位能近似，討論植入離子在其穩定位，是否可以降低特徵接觸阻抗，我們的結果也顯示，對於硒、鋁、錫等三種植入離子，硒、錫可以降低特徵接觸阻抗大約一個數量級左右。 從以上的結果來看，經由第一原理模擬計算可以使我們知道在原子尺度下，植入離子於介面的析離作用，以及植入離子對於蕭特基位障與特徵接觸阻抗的改善效果。最後，我們也預期這些結果將對實際製程的研究方向有所幫助。

As the state-of-the-art semiconductor process techniques drive devices toward sub-10nm sizes, the traditional Si-based MOSFETs are approaching their scaling limit in near future. It is very important to find a suitable material to replace Si. The element right below Si in the periodic table, Ge, is considered a potential channel material due to its higher mobility. However, Ge suffers from its strong fermi level pinning effect because many interface states appear near its valence band. Consequently, the interfacial electronic structures exhibit a high value of Schottky barrier height, which results in an undesired specific contact resistance. Therefore, fermi level depinning is a critical step to make Ge become feasible in semiconductor devices. In our research, we calculate the electronic structures and quantum transport of the NiGe/Ge contact with various ions implanted. We choose the interface crystal orientations of NiGe and Ge to be (112) and (001), respectively. The implanted ions are Se, Al, Sn, Sc, Ti, V, S, Cl, and Pt. Our calculations can be divided into three parts. We first study, in atomic scale, how the implanted ions segregate at the interface using the local density approximation of density functional theory (DFT). Second, we calculate the Schottky barrier height under the influence of the implanted ions using the DFT’s HSE06 hybrid functional. Calculations show that Se and V ions, both most stably located at the Ge side of the interface, reduce the Schottky barrier height by 0.175 and 0.095 eV, respectively, where the former is consistent with the experiment. Finally, we calculate the specific contact resistance of the same contact by employing the nonequilibrium Green's function method with the DFT’s modified Becke-Johnson potential. Such quantum-transport calclations show that ion implantation could reduce the specific contact resistance, up to one order from our results.