1.Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China 2.School of Physics, University of Chinese Academy of Sciences, Beijing 100049, China 3.Songshan Lake Materials Laboratory, Dongguan 523808, China 4.Department of Physics, Cornell University, Ithaca 14853, USA 5.Center of Materials Science and Optoelectronics Engineering, University of Chinese Academy of Sciences, Beijing 100049, China
Fund Project:Project supported by the National Key R&D Program of China (Grant Nos. 2016YFA0300602, 2016YFA0302400, 2017YFA0302903), the National Natural Science Foundation of China (Grant No. 11227903), the Beijing Municipal Science and Technology Commission, China (Grant Nos. Z181100004218007, Z191100007219011), the National Basic Research Program of China (Grant No. 2015CB921304), and the Strategic Priority Research Program of Chinese Academy of Sciences (Grant Nos. XDB07000000, XDB28000000, XDB33000000)
Received Date:23 April 2021
Accepted Date:10 May 2021
Available Online:07 June 2021
Published Online:20 September 2021
Abstract:The interplay between non-trivial band topology and magnetic order can induce exotic quantum phenomena, such as the quantum anomalous Hall effect and axion insulator state. A prevalent approach to realizing such topological states is either by magnetic doping or through heterostructure engineering, while the former will bring in inhomogeneity and the latter requires complex procedures. Intrinsic magnetic topological insulators are expected to avoid the aforementioned disadvantages, which is of great significance in both studying and practically using these exotic quantum phenomena. Recently, a Zintl compound EuIn2As2 is predicted to be an intrinsic antiferromagnetic axion insulator. The bulk magnetic order of EuIn2As2 has been reported in a lot of experiments, while the topological nature has not yet been confirmed. The surface properties of intrinsic magnetic topological insulators play an important role in the interplay between magnetic order and non-trivial surface state. Here in this work, we study the surface structure and electronic property of EuIn2As2 single crystal by using scanning tunneling microscopy/spectroscopy (STM/S) and non-contact atomic force microscopy (NC-AFM). Considering the strength of bonds, the easy cleavage plane of the crystals possibly lies between In-In layers or between Eu-As layers. The STM topographies show that the cleaved surface is dominated by a striped pattern. And the dominated step height is an integer multiple of c/2, which implies that only one kind of cleavage plane is preferred. Atomic-resolved surface topographies show that the striped pattern is the $ 1\times 2 $ surface reconstruction with 50% coverage. Hence an In-terminated surface which will be 100% coverage is ruled out. The spatial evolution of STS near vacancies on the striped pattern shows a hole-doping feature. All of these results reveal that the striped pattern is the $ 1\times 2 $ surface reconstruction of the Eu terminated surface with 50% coverage. Using the STS, we measure the local densities of states on the striped surface at various temperatures. We find that there is an asymmetric valley-peak feature in the density of states near the Fermi energy at 4 K, which is gradually weakened with increasing temperature, and disappears above the antiferromagnetic Néel temperature, indicating that the asymmetric valley-peak feature is closely related to the antiferromagnetic order. Besides, a maze-like pattern is observed occasionally near some step edges. The STM topographies show atoms both on bright and dark stripes of the maze-like pattern, which form a whole hexagonal lattice. And the NC-AFM images show that the maze-like pattern is about 1 ? higher than the Eu terminated striped pattern. Based on these results, the maze-like pattern can be explained as the buckled Eu surface with 100% coverage. These results provide important information for understanding the surface electronic band structure and topological nature of EuIn2As2. Keywords:scanning tunneling microscopy/ EuIn2As2/ magnetic topological insulator/ surface reconstruction
全文HTML
--> --> --> 1.引 言磁性与拓扑结合可以产生新奇的量子现象[1], 比如量子反常霍尔效应[2]、轴子绝缘体态[3]、手性马约拉纳费米子[4]等. 实现这些拓扑相有两种广为人知的途径, 分别是磁性掺杂和构造异质结(拓扑绝缘体/磁性薄膜)[5,6]. 然而前者会给材料引入无序; 后者需要复杂的制备流程, 进而影响对材料性质的研究和应用. 规避上述问题的一个直接办法, 是寻找具有确定化学计量比, 且具有内禀磁序的拓扑绝缘体, 即内禀磁性拓扑绝缘体. 过去几年, 理论预言了一些内禀磁性拓扑绝缘体, 包括MnBi2Te4家族系列[7-9]以及EuIn2As2, EuSn2As2, EuSn2P2等含Eu系列材料[10-12]. 其中, EuIn2As2具有内禀的反铁磁序, 早先的磁测量结果显示它的磁结构可能是A型反铁磁, 即Eu原子在层内是铁磁耦合, 相邻Eu层之间是反铁磁耦合, 材料的反铁磁转变温度TN约为18 K[13-16]. 理论预言在反铁磁相下它是轴子绝缘体, 磁性原子的自旋取向决定了它所处的拓扑态: 当自旋沿着ab面内时, 它是具有无能隙表面态的拓扑晶体绝缘体; 当沿着c方向时, 它是具有一维棱态(hinge state)的高阶拓扑绝缘体[10]. 最新的中子衍射实验认为, EuIn2As2可能具有更复杂的螺旋形反铁磁序, 具有$ {C}_{2}\times {\cal{T}} $的对称性, 计算表明它将具有受此对称性保护的无能隙表面态[17]. 实验上, 霍尔测量显示载流子为空穴[13,18], 角分辨光电子能谱(angle resolved photoemission spectroscopy, ARPES) 观测到以布里渊区Γ点为中心的空穴型费米口袋[13,18,19]以及重空穴掺杂的表面态[18], 表明EuIn2As2拓扑表面态位于费米能以上的未占据态. 在变温ARPES中, 不同小组在反铁磁相变前后观测到了不同的能带结构[18,19]. 而表面电子态的测量结果与解理面的结构直接相关, 理论上对表面能带结构的计算和预言也依赖于对表面结构的认识[20]. 作为研究表面结构和电子结构的重要实验手段, 扫描隧道显微镜(scanning tunneling microscope, STM) 对 EuIn2As2的研究还未见报道. 本文利用STM, 研究了EuIn2As2单晶解理表面的结构. STM形貌图表明, 材料的解理面以条纹状的有序结构为主, 结合原子分辨的STM形貌图、晶格对称性分析以及扫描隧道谱(scanning tunneling spectroscopy, STS) 信息, 认为条纹面起源于Eu截止面50%覆盖度形成的表面重构. 通过对条纹面态密度的测量, 发现费米能附近的态密度随温度的演化与反铁磁相变有很强的关联特征. 此外, 在台阶边缘附近偶尔观测到少量迷宫状结构, 通过分析原子像与台阶高度, 认为它可能是Eu原子形成的翘曲结构. 作为表面能带计算与实验观测的桥梁, 对表面结构的测定为理解EuIn2As2的表面电子结构以及拓扑性质提供了重要线索. 2.实验方法EuIn2As2单晶采用自助熔方法生长, 选择In作为助熔剂, 以摩尔比Eu∶As∶In = 1∶3∶12进行配比, 将原材料放入氧化铝坩埚, 然后装进石英管中, 此操作在充满氩气的手套箱中进行. 将封装好的石英管经过12 h升到1100 ℃, 保持12 h, 以1 K/h的速率降温至700 ℃, 然后迅速离心甩出多余的助熔剂In, 得到六方片状单晶. 测量所用仪器是一台自主搭建的低温扫描探针显微镜, 兼具扫描隧道显微镜与原子力显微镜(atomic force microscope, AFM) 的功能. 样品在低温超高真空环境下解理, 解理温度约20 K. 如不加特殊说明, 扫描头的基础温度是4.2 K. 实验使用电化学腐蚀的W针尖, 并在Au靶上做场发射处理. STM形貌图在恒流模式下获取. STS采用标准的锁相放大技术获取, 调制电压频率为781.1 Hz. AFM测量基于qPlus传感器, 采用非接触调频模式. 3.结果与讨论EuIn2As2具有P63/mmc空间群, Eu2+阳离子层与[In2As2]2–结构沿着c方向交替堆叠, 如图1(a)所示. 晶格常数为a = 4.2 ?, c = 17.9 ?. 最近邻Eu层的间距为c方向晶格常数的一半, c/2 = 8.95 ?. 在[In2As2]2–结构内部, In—In和In—As为共价结合, Eu2+与上下层的As之间是离子结合[14]. 这种结构与CaKFe4As4类似, Eu类似于Ca和K, 可以被看成是插层原子[21]. 从能量角度看, EuIn2As2存在两个易断键位置: In—In键, 断裂后形成In的六角格子表面; 或者Eu—As键, 断裂后形成Eu或者As的截止面. 图1(b)是单晶X射线衍射的结果, 可以看到一系列清晰的(00L) 晶面衍射峰, 说明实验使用的晶体拥有很高的质量. 图 1 EuIn2As2的晶格结构和表面形貌 (a) EuIn2As2的晶格结构示意图, 图中所示Eu—Eu, Eu—In, In—As原子层间距分别是8.95, 3.07, 3.96 ?; (b) EuIn2As2的X射线衍射图案, 插图是典型EuIn2As2单晶的照片; (c) 包含台阶的大范围STM形貌图(Vs = 1 V, I = 100 pA), 插图是沿着图中黄线所标示位置的高度轮廓, 台阶高度8.95 ?, 与晶格常数的一半 c/2 = 8.95 ? 一致; (d) 典型的条纹面STM形貌(Vs = 1 V, I = 1 nA); (e) 包含迷宫状结构的STM形貌图 (Vs = –200 mV, I = 20 pA), 图中左下角为上层台阶, 右侧为条纹面, 插图是沿着图中黄线所标示位置的高度轮廓, 迷宫面与条纹面高度差约1 ? Figure1. Crystal structure and surface morphologies of EuIn2As2. (a) Schematic crystal structure of EuIn2As2, where the interlayer distance Eu—Eu, Eu—In, and In—As is 8.95, 3.07, and 3.96 ?, respectively. (b) X-ray diffraction pattern of EuIn2As2 crystal. The inset is a picture of typical EuIn2As2 single crystal. (c) Large scale STM topography showing step edges (Vs = 1 V, I = 100 pA). Inset is the height profile along yellow line, the step height is 8.95?, which is consistent with half unit cell c/2 = 8.95 ?. (d) STM topography of typical stripe surface (Vs = 1 V, I = 1 nA). (e) STM topography containing a maze-like structure (Vs = –200 mV, I = 20 pA), which is located between a higher step (left bottom) and a stripe surface (right part). The inset is the height profile along yellow line. The step height between maze-like surface and stripe surface is about 1 ?.
图1(c)是解理后的EuIn2As2样品表面的STM形貌图. 可以看到, 图中存在一个凸台状的台阶, 图1(c)中的插图显示了形貌图中沿黄线的高度轮廓, 可以看到台阶的高度约为8.95 ? = c/2. 经过大量观测, 绝大多数相邻台阶的高度差为c/2或其整数倍. 台阶表面是亮暗条纹构成的有序结构, 如图1(d)所示, 以下称之为条纹面. 极少数情况下, 观测到一种条纹方向多变、形似迷宫的表面, 以下称之为迷宫面. 如图1(e)所示, 左下角高亮的区域是上层台阶, 下层台面最右侧是常见的条纹面, 在条纹面与上层台阶之间是迷宫面. 两种表面沿着黄线的高度轮廓如图1(e)的插图所示. 可以看到迷宫面起伏较大, 亮暗条纹高度差约1 ?, 而条纹面起伏小, 亮暗条纹高度差约0.1 ?. 简洁起见, 以下用迷宫面亮条纹的高度指代迷宫面的高度. 迷宫面比条纹面高约1 ?, 迷宫面的暗条纹与条纹面表观高度相当. 如果材料解理时, 两种可能断键的位置(In—In键和Eu—As键) 都发生断裂, 将会出现复杂的台阶高度, 如In和Eu面的台阶高度3.07 ?, In和As面的台阶高度3.96 ?等, 如图1(a)所示, 大量实验中从未观测到这些台阶高度. 因此可推测样品只在In—In或Eu—As中一处断键. 为了探究解理面的截止情况, 本文对两种表面结构做了详细的研究. 图2(a)是条纹面上获得的STM形貌图, 可以看到亮条纹覆盖整个表面. 条纹上存在一些转折和空洞构成的断点, 这导致亮条纹表现为不连续的线段. 但条纹整体呈现一定的方向性, 这种方向性可以在微米尺度保持. 图2(b)是同一区域的原子分辨STM形貌图, 亮条纹内原子间距为晶格常数a, 亮条纹间原子间距约7.27 ?, 这个数值符合EuIn2As2面内六角格子的次近邻原子间距$ \sqrt{3}a $, 图2(b) 中标注了晶格常数为a的六角格子点阵, 可以看出条纹面具有1 × 2重构. 图2(c) 是图2(a)的快速傅里叶变换(fast Fourier transform, FFT), 图中黄色箭头标注的两个锐利的点, 源于形貌图中的平行条纹, 红圈中的6个红点是图2(b)中标注的六角格子对应的布拉格点. 二者对比, 这两个锐利的点恰好位于Γ点到布拉格点的中点位置, 再次表明亮条纹具有固定的间距, 其间距为次近邻原子的间距. FFT中锐利的点旁边存在较暗的横线, 它源于条纹的之字形转折, 沿着图2(a)中的红色虚线方向, 亮条纹转折前后错位半个周期. 图 2 条纹面的结构 (a) 条纹面的STM形貌 (Vs = 1 V, I = 200 pA), 沿图中红色虚线条纹转折一次错位半个周期, 深蓝色区域为表面缺陷; (b) 条纹面的原子分辨像 (Vs = –2 mV, I = 1 nA), 红色平行四边形表示重构后的原胞, 红色点阵代表面内晶格格点; (c) 图(a)的FFT, 黄色箭头所示为条纹的周期, 红色圆圈标注的六个红点代表图(b)所标注六角格子的布拉格点; (d)条纹面的原子排布示意图, 暗红色球代表观测到的Eu原子, 红色虚线圆圈代表亮条纹之间丢失的Eu原子, 黑色/黄色点代表Eu层以下的As/In原子; (e)跨条纹面上空位缺陷的STS谱, 插图中红线标出了测谱位置, 测谱条件为Vs = 400 mV, I = 200 pA, 调制电压幅度5 mV, 图中黄色虚线帮助示意STS曲线中的特征随空间的演变 Figure2. Structure of stripe surface. (a) STM topography of stripe surface (Vs = 1 V, I = 200 pA), the stripe shifts half unit cell along the red dashed line. Dark blue areas are surface defects. (b) Atomic resolved STM image of stripe surface (Vs = –2 mV, I = 1 nA), the red parallelogram shows the doubled unit cell, the red spots are superposed in-plane lattice structure. (c) FFT image of panel (a). The stripe modulation is marked by yellow arrows. The superposed six red dots marked by dashed red circles are the simulated Bragg points of the inserted hexagonal lattice in panel (b). (d) Schematic drawings of atomic arrangement on the stripe surface. Dark red spheres show the observed Eu atomic chain, dashed red circles show the missing Eu atoms between light stripes, black/yellow dots show the As/In atoms beneath the Eu layer. (e) STS taken across a vacancy in a stripe surface, the position is marked as the red line in the inset topographic image. Spectra are taken at Vs = 400 mV, I = 200 pA, with modulation 5 mV. The yellow dashed lines are guided for the spatial evolution of STS.
原子可以通过形成双聚体以减少悬挂键, 进而形成条纹结构, 比如半导体Si (100) 2 × 1的重构[22,23]. EuIn2As2的原子分辨形貌图表明亮条纹不是双聚体, 而是单原子链. 在亮条纹之间, 没有观测到轮廓清晰位置明确的原子像, 对应丢失的原子链. 可见样品表面形成1 × 2非保守重构(non-conservative reconstruction), 只有50%覆盖度, 如图2(d)所示. 如果样品在In—In处解理, 断键后In原子将形成100%覆盖度的表面, 与观测结果不符. 因此, 可以推测解理样品在Eu—As处断键, 形成了50%覆盖度且具有1 × 2重构的Eu原子表面. 谱学研究进一步确认了截止面的原子种类. 材料中的缺陷或元素替代可以带来掺杂效应, 比如阳离子的缺位导致空穴型掺杂, STS (dI/dV ) 曲线将向高能方向移动. 我们研究了EuIn2As2条纹面上广泛存在的空洞, 这些空洞在不同偏压下都表现为凹坑, 因此认为空洞对应原子的缺位. 图2(e)是条纹面上从远离缺陷位置到空洞的STS, 随着向空洞靠近, STS曲线整体向高能方向移动, 大量的实验表明, 空穴型掺杂的结论也适用于条纹面上其他尺寸的空洞, 且空洞尺寸越大能量移动越大. 这与条纹面是Eu面一致. 图3(a)给出了条纹面上大能量范围的STS, 在–1.5 V左右态密度开始明显抬升, 与ARPES实验[13,18]和理论计算[10,24]确定的Eu 4f电子构成的平带对应. 此外, 条纹面上的STS在费米能附近会随温度演化. 如图3(b)所示, 4.2 K下材料处于反铁磁相, 费米能附近态密度存在非对称的谷-峰特征, 随着温度升高态密度谷逐渐变平, 在19 K左右达到饱和, 这个温度与反铁磁相变温度TN接近, 表明其与反铁磁序密切相关. 在文献[18]报道的变温ARPES实验中观测到了类似的现象: 当温度T = 6 K时, 在Γ点费米能以下0.1 eV附近存在“M”型的能带, 并在费米能附近打开一个能隙; 当T = 35 K时, “M”能带演化为跨过费米能的大展宽的能带. 通过STS谱也观测到正偏压处的态密度峰伴随反铁磁相变而发生演化, 如图3(b)所示, 随着温度升高, 态密度峰被压制, 在18 K以上达到饱和, 表明其与反铁磁相变有关. 而根据理论预测以及ARPES的观测结果, 与拓扑转变相关的能带信息正位于费米能以上, 即正偏压的位置. 但是, 由于表面重构可以导致复杂的表面态, STS实验很难直接识别出拓扑非平庸的表面态, 需要结合理论计算等做进一步的分析. 图 3 条纹面上的STS (a) 大能量范围的STS, 采谱条件为Vs = –1.7 V, I = 1 nA, 调制电压为10 mV; (b) 不同温度的STS, 采谱条件为Vs = –250 mV, I = 200 pA, 调制电压为5 mV Figure3. STS on stripe surface: (a) Large energy range STS, where the spectrum is taken at Vs = –1.7 V, I = 1 nA with modulation 10 mV; (b) STS taken at different temperatures, where the spectra are taken at Vs = –250 mV, I = 200 pA with modulation 5 mV.
在确定条纹面的情况后, 可以进一步讨论迷宫面的结构特征. 在少数台阶内边缘附近(即靠近相邻更高台阶的区域), 可以观测到如图1(e)所示的迷宫结构, 其尺度通常小于20 nm. 图4(a)是迷宫面的原子分辨形貌图. 可以看到迷宫面没有显著的空洞缺陷, 亮条纹由单原子链组成, 原子链在nm尺度发生方向的改变, 在每个转折点亮条纹方向改变120°. 图 4 迷宫面的结构 (a) 迷宫面的原子分辨图 (Vs = 10 mV, I = 100 pA), 插图是相应的FFT图; (b) 小范围的原子分辨STM形貌图 (Vs = 10 mV, I = 100 pA), 暗条纹中可以看到清晰的原子, 原子位置图中用黄球标示; (c)迷宫面和条纹面台阶的AFM (上图)和STM (下图)形貌图, 图中左侧暗区为下层台面, 中部为条纹面, 最右侧为迷宫面, AFM/STM观测的台阶高度为1.1 ?/1.3 ?; (d), (e)迷宫面在电压脉冲下的不稳定性, 在(d)中红色圆圈内施加400 mV的电压脉冲, 表面原子排布变为(e)图所示, 其中黄圈位置的原子由暗变亮, 红圈位置的原子由亮变暗 Figure4. Structure of the maze-like surface. (a) Atomic resolved STM image of maze-like surface (Vs = 10 mV, I = 100 pA), with corresponding FFT image inserted. (b) Small scale atomic resolved STM topography (Vs = 10 mV, I = 100 pA), atoms in the dark stripes, which are marked by yellow dots, are clearly visible. (c) AFM (up panel) and STM (down panel) topography of the step coexisting of maze (right) and stripe (middle) surface. Step height from AFM (STM) is 1.1 ? (1.3 ?). (d), (e) Surface topography evolution under a voltage pulse showing the metastability of the maze-like pattern. A 400 mV voltage pulse is applied in the red circle of panel (d), leading to a rearrangement of atoms in both red and yellow circles from panel (d) to panel (e).