1.Hunan Provincial Key Laboratory of Intelligent Information Processing and Applications, College of Physics and Electronic Engineering, Hengyang Normal University, Hengyang 421002, China 2.School of Physics and Electronics, Hunan Normal University, Changsha 410081, China 3.School of Physics and Electronics, Hunan University, Changsha 410082, China
Abstract:The spin-orbit interaction (SOI) of light refers to the mutual conversion and coupling between the spin angular momentum and orbital angular momentum. It is a fundamental effect in optics, and has been widely found in many basic optical processes, such as reflection, refraction, scattering, focusing, and imaging. So it plays an important role in the fields of optics, nanophotonics, and plasmonics, and has great potential applications in precision measurement and detection, information storage and processing, particle manipulation, and various functional photonic devices. Recently, it has been found that a circularly polarized light beam normally passing through an isotropic sharp interface can undergo an SOI process, that is, part of the incident beam experiences a spin-flip and acquires a spin-dependent vortex phase with a topological charge of $ \pm2 $. However, the physical origin of this phase and the role of the interface played in the SOI process are still unclear at present. In this work, a Fresnel Jones matrix is first established to describe the relationship between the incident beam and the transmitted beam, based on which we unveil that the vortex phase is in fact a spin-redirection Berry geometric phase, originating from the topological structure of the beam itself. The properties of the interface affect the conversion efficiency of the SOI. This kind of SOI is very similar to that in the azimuthal Pancharatnam-Berry phase elements. The difference lies in the fact that the Pancharatnam-Berry phase originates from the external anisotropy of the composite material. Generally, the efficiency of this SOI is extremely low, which limits its applications. The existing method of enhancing this SOI employs an isotropic epsilon-near-zero slab, whose maximum efficiency can reach only about 20%. Since the anisotropic medium (such as birefringent uniaxial crystals) has more degrees of freedom, we further point out that the weak SOI can be greatly enhanced by an optically thin uniaxial slab whose optical axis is parallel to the normal direction of the interface. And under certain conditions, the conversion efficiency can reach 100%. Our study not only establishes a simple and convenient full-wave theory for this SOI, but also reveals the relevant underlying physics, and further provides a possible scheme to significantly enhance the SOI. Keywords:spin-orbit interaction of light/ vortex phase/ geometric phase/ angular momentum of light
全文HTML
--> --> --> 1.引 言光既可以具有自旋角动量, 又可携带轨道角动量. 自旋角动量与光的偏振有关, 如左右旋圆偏振光子分别携带$ \pm \hbar $的自旋角动量. 轨道角动量有两类[1,2], 一类为内禀(intrinsic)的轨道角动量, 与涡旋光场有关, 每个光子携带$l\hbar $的轨道角动量, 其中l为涡旋相位的拓扑荷数; 另一类为外禀(extrinsic)的轨道角动量, 和光束传播的轨迹有关, 定义为坐标原点到光束中心的距离与线动量的叉乘, 与经典粒子的机械角动量类似. 光的自旋角动量和轨道角动量之间的相互转换和耦合被称为自旋-轨道相互作用(spin-orbit interaction, SOI)或耦合[1,2]. 它是光学中的一种基本效应, 广泛存在于界面的反射和折射、非均匀各向异性介质、强聚焦、粒子散射、表面波和消逝波等体系中, 在光学、纳米光子学和等离子光学等领域扮演越来越重要的角色, 并在精密测量与探测、信息存储与处理、微粒操纵以及各种功能光子器件设计等方面显示出巨大的应用潜力[1-9]. 在旋转对称的系统中, 光的SOI表现为自旋可控的涡旋相位的产生(内禀轨道角动量); 在旋转对称性破缺的系统中, 它表现为自旋霍尔效应(外禀轨道角动量)[1,2]. 光的自旋霍尔效应存在于很多体系中, 如光束在突变界面的斜入射[1,10-17]、一维的潘查拉特南-贝里(Pancharatnam-Berry, PB)相位元件[18-21]、各向同性的非均匀材料[21-23]等; 自旋可控的涡旋相位也在方位变化的PB相位元件[24-27]、强聚焦[28,29]、单轴晶体中的传输[30,31]等体系中出现. 然而, 有趣的是, 当光束正入射至均匀的、各向同性的突变界面时, 也能产生自旋相关的涡旋相位[32-34]. 光束正入射时, 极小的一部分入射光束发生自旋反转(左旋变为右旋或者右旋变为左旋), 并获得拓扑荷数为±2的涡旋相位(图1(a)). 其内在机制被认为是SOI, 但这种相位的物理来源、为什么拓扑荷数为±2以及界面在其中究竟扮演何种角色等一系列的问题, 目前并不清楚. 另外, 该SOI与光束通过方位变化的各向异性PB相位元件[24-27]时产生涡旋相位的过程极为相似. 光束入射到方位PB相位元件时, 一部分入射光束发生自旋反转并获得2倍于元件拓扑荷数(q)的涡旋相位因子2q?, 其中?是PB相位元件的局部的光轴方向, 是坐标位置的函数. 也就是说这种相位因子来源于PB相位元件的非均匀的各向异性. 而前文所提到的界面是各向同性且均匀的, 这与PB相位元件的情况又有何联系和区别? 图 1 光束正入射至各向同性的突变界面时SOI的示意图 (a) 左旋圆偏振光束正入射至界面后, 部分光束发生自旋反转变成右旋光, 并获得拓扑荷数为2的涡旋相位(两个小图分别表示一种典型的涡旋光束的强度和相位分布); 注意, 未发生SOI的那部分光束并没有在图中画出; $\left| + \right\rangle $和$\left| - \right\rangle $分别表示左、右旋圆偏振; (b) 光束中各平面波分量的自旋与局部坐标的旋转耦合的示意图, 其中圆锥代表光束的角谱, 绿色的箭头线代表任意的两支平面波的波矢, 橙色带箭头的小圆圈表示各平面的偏振矢量在实验室坐标上的投影(均为圆偏振), ${\varOmega _\xi }$为坐标旋转的空间旋转 Figure1. Schematic illustration of the SOI for a light beam normally impinging onto a sharp isotropic interface. (a) When a left-circularly polarized beam normally passes through the interface, part of the incident beam converts into a right-circularly polarized beam, and carries a vortex phase with a topological charge of 2. Note that the spin-maintained portion is not shown in the picture. $\left| + \right\rangle $ and $\left| - \right\rangle $ denotes the left- and right-handed polarization, respectively. (b) Schematic illustration of rotational coupling between the local coordinates and the spin of the plane wave components within the beam spectra. The cone represents the angular spectrum of the beam. The two green arrows represent the wave vectors of arbitrary two plane waves. The orange circles with arrows indicate the projection of polarization vectors of each plane wave on the laboratory coordinates (all circularly polarized). ${\varOmega _\xi }$ is the spatial coordinate rotation.
式中透射光束分为两部分, 一部分与入射光束相同, 是寻常模式; 另一部分表现出自旋反转现象并携带拓扑荷数为2的涡旋相位, 是反常模式. 图2给出了寻常模式和反常模式的光强和相位分布. 反常模式(图2(a))光斑中心是光强为0的空心区域, 相位在方位方向变化4π, 即拓扑荷数为2的涡旋相位; 寻常模式(图2(b)) 光斑中心是实心区域, 不携带方位方向的涡旋相位. 当寻常模式强度为0, 即只有反常模式时, 该SOI过程中的转换效率为100%. 然而一般情况下, 对于传统材料构成的界面, $\left| {{t_{\rm{TM}}} - {t_{\rm{TE}}}} \right|$实际上是一个非常小的值, 尤其在入射角较小时. 这种情况意味着SOI的转换效率是极低的, 常规的实验精度难以被观察到. 这也是这种效应迄今为止没有在实验上被观察到的原因之一. 图 2 左旋圆偏振贝塞尔光束正入射至一个界面时, 透射光束的反常模式(a)和寻常模式(b)的归一化光强分布, 其中两个小图分别表示为对应的相位分布, 在计算中, 取入射光束的波长$\lambda = 1$且${w_0} = 20\lambda $ Figure2. Normalized intensity distribution of the abnormal mode (a) and normal mode (b) of transmitted light beam under the normal incidence of a left-handed circularly polarized Bessel beam at a sharp interface. The insets represent the phase distribution of corresponding modes. Here, we take the working wavelength as $\lambda = 1$ and ${w_0} = 20\lambda $.