Using Spin-Polarized Carriers in Semiconductor Lasers for Optical Interconnects

自旋电子学试图操纵和利用电子和光子所拥有的自旋来存储和处理信息。在操纵电子自旋和相关磁矩方面取得的进展获得了2007年诺贝尔物理学奖的认可，这对于使用自旋阀大幅提高计算机硬盘容量至关重要。这种自旋阀依赖于磁电阻的概念，其中装置的磁性结构决定了电流是允许的还是限制的。虽然自旋阀在磁存储和传感信息方面仍然很有价值，但它们在信号处理和数字逻辑方面的用途有限；考虑其他基于自旋的器件方案的实际途径将是至关重要的。该提案提供了一个这样的机会，即重新研究如何改进广泛应用于dvd、光通信和医学的半导体激光器。在商用半导体激光器中，产生足够数量的电子和空穴，它们的电荷相反，多余的载流子重新组合，发射出具有相同波长和相位的相干光子。虽然光子、电子和空穴都有自旋，但传统半导体激光器的工作原理完全不受它们的影响，因为指向不同方向（“上”和“下”）的自旋之间没有净不平衡。提议的工作将提供一个紧密结合的教育和推广工作，为高中生组织一个为期一周的关于激光、实验和理论的夏季研讨会，并在一年一度的SPIE光学和光子学会议上组织专题讨论会，以弥合自旋电子学和光学界之间的差距。提出的研究建立在最近的实验进展的基础上，表明激光的操作可以通过光学或电注入自旋极化载流子来强烈地改变，从而产生自旋不平衡。在稳态和低频操作中，改变这种自旋激光器中注入载流子的偏振已经实现了：(a)与传统的（自旋非极化）对应物相比，激光阈值降低和发射强度增强；(b)发射光的强调制，即使在固定的注入强度下。自旋激光器最重要的应用在很大程度上仍未被探索，并且与超快操作和其优越的动力学性能有关。传统激光器的带宽通常受到弛豫振荡频率的限制。然而，自旋激光器应该能够大幅增加振荡频率，并且)一个新的和更高的频率尺度来控制发射光的偏振振荡。PI将开发自旋激光器的详细建模，并探索如何：(1)定制谐振腔各向异性以实现100 GHz的动态带宽，(2)减少寄生频率调制啁啾，(3)改善开关特性和数字操作；并用实验数据进行验证。PI将研究替代器件几何形状，重点关注GaN自旋纳米激光器的最新进展。传统的金属互连被认为是摩尔定律缩放的瓶颈和功耗的主要来源。由于以激光器为关键元件的光互连可以解决潜在的限制，因此自旋激光器的这项工作可能具有变革性并实现新的互连。

Spintronics seeks to manipulate and use the spin that electrons and photons possess to store and process information. Advances in manipulating the spin of an electron and the associated magnetic moment, recognized by the 2007 Nobel Prize in Physics, were crucial for a dramatic increase in the capacity of computer hard drives using spin-valves. Such spin-valves rely on the concept of magnetoresistance in which the magnetic configuration of a device determines if the current flow is permitted or restricted. While spin-valves remain valuable for magnetically storing and sensing information, they are of limited use for signal processing and digital logic; it would be crucial to consider practical paths to other spin-based device schemes. One such opportunity is afforded in this proposal by reexamining how to improve semiconductor lasers, widely used in DVDs, optical communication, and medicine. In commercial semiconductor lasers, a sufficiently large number of electrons and holes, having an opposite charge, are generated and the excess charge carriers recombine to emit coherent photons that possess the same wavelength and phase. Although photons, electrons and holes all have spin, the working principle of conventional semiconductor lasers is completely unaffected by them, because there is no net imbalance between spins pointing in different directions ('up' and 'down'). The proposed work will provide a closely integrated educational and outreach efforts by organizing a week-long Summer Workshop on Lasers, experiment and theory for high school students and organizing symposia to bridge the gap between the spintronics and optics communities at the annual SPIE Optics and Photonics Conference.The proposed research builds on recent experimental advances demonstrating that the operation of lasers can be strongly modified by optically or electrically injecting spin-polarized carriers, having thus a spin imbalance. In the steady-state and low-frequency operation changing the polarization of injected carriers in such spin-lasers has already enabled: (a) lasing threshold reduction and enhanced emission intensity as compared to their conventional (spin-unpolarized) counterparts; (b) strong modulation of the emitted light, even at a fixed injection intensity. The most important applications of spin-lasers are still largely unexplored and pertain to an ultrafast operation and their superior dynamical performance. The bandwidth in conventional lasers is typically limited by the relaxation oscillation frequency. However, spin-lasers should enable a large increase in oscillation frequency , and) a novel and much higher frequency scale governing the polarization oscillation of the emitted light. The PI will develop a detailed modeling of spin-lasers and explore how to: (1) tailor the resonant cavity anisotropy to achieve dynamical bandwidths 100 GHz, (2) reduce parasitic frequency modulation-chirp, and (3) improve switching properties and digital operation; and validate with the experimental data. The PI will study alternative device geometries focused on the recent advances in GaN spin-nanolasers. Conventional metallic interconnects are recognized as the bottleneck in Moore's law scaling and the main source of power dissipation. Since optical interconnects, having lasers as their key element, could address the underlying limitations, this work on spin-lasers may have transformative character and enable novel interconnects.