Simulation of Spin Transport, Diffusion and Injection into Semiconductors

公众对越来越快、越来越小的电子设备（如计算机）的需求要求在每个芯片上装入更多、更小的晶体管。这导致了纳米技术的诞生，以及最近被称为“自旋电子学”的纳米技术领域。在这里，不仅电荷，还有自旋——电子和空穴的另一个基本属性——被用来设计设备的功能。自旋电子学设备的潜在好处之一是可以使用相同的单元来计算和存储计算机，更低的功耗，小型化，更普遍的是可以设计概念上的新设备，将旧功能与全新功能混合在一起。自旋电子学相关效应的第一个商业应用是IBM在上世纪90年代为磁性硬盘驱动器设计的读磁头，它将磁盘驱动器的存储空间增加到数百亿字节，并从那时起获得了数十亿美元的市场。很明显，这一迷人领域的每一项发展对专家和公众都是重要的。自旋电子学的基础是理解自旋动力学。不幸的是，诸如如何在半导体中注入自旋电流，如何在形成器件的不同材料的界面上维持它，哪种材料/纳米结构是最好的，以及自旋电流在特定材料中可以传播的长度等关键问题仍然是悬而未决的问题。该项目旨在掌握自旋动力学的基本原理，特别关注纳米电路及其组件等应用。我的目标是完全理解自旋输运、扩散和注入到半导体中。这些特性是开发半导体和混合（金属/半导体）自旋电子学器件的基础。电子自旋注入半导体的问题（考虑电子电路时的基本问题）仍然是开放的，以及对自旋输运和动力学中多体相互作用所起作用的理解。在这种情况下，对这些现象进行计算机模拟的能力至关重要。事实上，模拟可以被认为是理想的实验，它允许系统分析系统响应，即使在条件下（例如极端温度或无缺陷或外来材料），这些条件在实验室中很难或昂贵地重现。我计划使用两种强大的技术开发计算机模拟代码，非平衡蒙特卡罗和自旋相关漂移-扩散方程技术，并将它们应用于体自旋输运和通过半导体异质结构，研究多体相互作用效应，并最终应用于器件组件。为了做到这一点，我将需要发展和调整模拟技术本身来解决具体问题，以及发展潜在的自旋输运理论，例如与多体效应有关的自旋库仑阻力，我在2000年预测到，并在2005年实验观察到。在项目的第二部分，我计划将非平衡蒙特卡罗技术与原子模型相结合：这两种技术的合并将使我能够以迄今为止从未达到的精度分析磁性杂质和纳米结构的影响。我的目标是提高我们对自旋动力学（特别是多体相互作用效应）的理解，解决电子自旋注入半导体的难题，并最终设计自旋电子学电路元件的基本元件。

Public demand for increasingly faster and smaller electronic devices, such as computers, requires that more and even smaller transistors are packed on every chip. This has led to the birth of nanotechnology and, more recently of the nanotechnology field called 'spintronics'. Here not only the charge, but also the spin -- another fundamental property of electrons and holes -- is used to design device functionalities. Among the potential benefits of spintronics devices is the possibility of computers in which the same unit is used for computation and storage, of lower power consumption, of miniaturisation, and more generally the possibility of designing conceptually new devices which mix old functionalities with completely new ones.The first commercial application of a spintronics-related effect was the read heads for magnetic hard disk drives by IBM in the 90s, which increased the storage on a disk drive to tens of gigabytes and have since then secured a market of billions of dollars. It is clear that every development in this fascinating field is important to the specialists as well as of the general public. The basis of spintronics is understanding the spin dynamics. Unfortunately key issues such as how to inject a current of spins in a semiconductor, how to sustain it across the interfaces of the different materials forming the devices, which materials/nanostructures are best and what lengths a current of spin can travel in a specific material are still open questions. This project aims to master the principles underlying the spin dynamics, with particular attention to applications such as nanocircuits and their components. My objectives are to fully understand spin transport, diffusion and injection into semiconductors. These properties are fundamental for developing semiconductor and hybrid (metal/semiconductor) spintronics devices. The problem of electrical spin injection into semiconductors (fundamental when thinking of an electronic circuit) is still open, as well as the understanding of the role played by many-body interactions in spin transport and dynamics. In this context the ability of performing computer simulations of these phenomena is of paramount importance. Simulations can in fact be thought as ideal experiments, which allow a systematic analysis of the system response even for conditions (e.g. extreme temperatures or defect-free or exotic materials) which are too difficult or expensive to reproduce in the lab. I plan to develop computer simulation codes using two powerful techniques, non-equilibrium Monte-Carlo and spin-dependent drift-diffusion equation techniques and apply them to spin-transport in bulk and through semiconductor heterostructures, to the study of many-body interaction effects, and, ultimately, to device components.In order to do so, I will need to develop and adapt the simulation techniques themselves to the specific problems as well as to develop the underlying spin-transport theory, for example in relation to many-body effects such as the spin Coulomb drag which I predicted in 2000 and was observed experimentally in 2005.In the second part of the project, I plan to combine non-equilibrium Monte-Carlo techniques with the atomistic model: the merger of these two techniques will allow me to analyse the effects of magnetic impurities and nanostructures with an accuracy never reached so far. My aims are to improve our understanding of spin-dynamics (especially many-body interaction effects), to solve the puzzle of electrical spin-injection into semiconductors and ultimately to design basic elements for spintronics circuitry components.