Fund Project:Project supported by the Fundamental Research Fund for the Central Universities, China (Grant No. NS2017055).
Received Date:23 September 2018
Accepted Date:28 November 2018
Available Online:01 January 2019
Published Online:20 January 2019
Abstract:Zigzag graphene nanoribbon (ZGNR) is important for novel carbon-based spintronic applications. Currently, most of ZGNR spintronic studies focus on the collinear magnetism where the up-spin and down-spin are separated clearly. But in some cases, e.g. doping and adsorption, the magnetization profile can be modulated and thus noncollinear magnetism can occur. In order to shed light on possible noncollinear magnetism in ZGNR, we study non-collinear magnetism and electronic transport of boron or nitrogen-doped zigzag graphene nanoribbon based on noncollinear density functional theory and non-equilibrium Green's function method. For pristine ZGNR, our results show that the ZGNR presents helical magnetization distribution due to noncollinear magnetization in left and right lead. As the ZGNR is doped with boron and nitrogen atoms, the ZGNR shows a characteristic two-zone feature in the magnetization distribution. Near the dopant site, the magnetic moment of carbon atom is small. However, the magnetic moments of carbon atoms in the left (right) region of dopant are close to those of the left (right) lead. Such a feature provides the possibility of constructing domain walls with various widths on the edge of ZGNR. Moreover, the transmission at the Fermi level (E = 0 eV) decreases with the increase of relative angle between magnetizations of left and right lead, indicating that the spin-flip scattering dominates the electronic transport. However, at E = ±0.65 eV, there is a transmission dip with low transmission, which implies that the dopant induces the strong backscattering. To understand the origin of this dip, we calculate the density of states (DOS) and project the DOS onto each atom of doped ZGNR. The projected DOS shows a large and broad peak at E = ?0.65 eV for N-doped ZGNR but at E = +0.65 eV for B-doped ZGNR. The consistency between the position of dip in transmission and the position of peak in DOS indicates that the transmission dip mentioned above is attributed to strong backscattering from the dopant-induced bound state. Our theoretical results are expected to be useful for understanding the noncollinear magnetism and spin scattering in the doped ZGNR-based devices. Also, our work provides a considerable insight into the design of ZGNR-based nanoelectronic devices, such as the transistor based on spin transfer torque effect. Keywords:graphene nanoribbon/ electronic transport/ density functional/ non-equilibrium Green’s function
全文HTML
--> --> -->
3.计算结果与讨论掺杂石墨烯纳米带的非共线磁序与电子输运包含了两类效应: 非共线磁序和掺杂效应. 首先来探讨未掺杂石墨烯纳米带的结果, 以此来分析非共线磁序的影响. 图2(a)给出了左右电极磁化偏转角$\theta=0$° 时弛豫得到的磁化分布, 这一分布属于共线磁序, 所有磁化的方向都沿着z轴, 这一结果与其他研究者的计算结果是一致的[32]. 当右电极磁化方向逐渐偏离z轴时(图2(b)—(f)), 中心区变成了螺旋式排列的非共线磁序. $\theta$越大, 每一个磁化相对于z轴的偏转就越大. 如前所述, 左右电极磁化的大小和方向是保持固定的. 在这一边界条件下, 为了获得中心区磁化分布的最稳定的状态, 磁交换作用使得中心区右侧的磁化向右电极靠拢, 中心区左侧的磁化向左电极靠拢, 而中间的磁化介于二者之间, 这就使得螺旋式的磁化分布变成了较为有利的状态. 图 2 (a)—(f) 不同角度下的未掺杂石墨烯纳米带的磁化分布; (g) 透射系数; (h) 能带结构(红色与黑色分别代表自旋向上和自旋向下) Figure2. (a)?(f) Magnetization distribution; (g) transmission; (h) band structure of undoped ZGNR (red and black lines in (h) denote up-spin and down-spin).
图 4 石墨烯纳米带投影到几个原子上的电子态密度 (a) 未掺杂; (b) 硼掺杂; (c) 氮掺杂. C1代表中心区上边缘的最左边的碳原子, C8代表掺杂原子左侧的最近邻碳原子, B/N代表掺杂原子, C9代表未掺杂时上边缘最中心的碳原子 Figure4. Projected density of states: (a) Undoped; (b) B-doped; (c) N-doped ZGNR. C1 denotes the leftmost carbon atom on the upper edge. C8 denotes the left nearest neighboring carbon atom of the dopant. B/N is boron/nitrogen atom. C9 denotes the central carbon atom in the upper edge of undoped ZGNR.
图7给出了硼或氮原子掺杂的石墨烯纳米带的透射系数. 由于考虑的石墨烯纳米带包含了非共线磁序和掺杂, 因而电子输运将会受到这两类效应的影响. 第一类效应是非共线磁序效应, 即(1)式, 这一效应主要反映了当磁化方向非共线时的自旋翻转散射. 第二类效应是杂质所产生的束缚态, 这一束缚态对传导电子形成背散射, 降低透射系数. 对于氮掺杂的透射系数(图7(a)), 我们先与未掺杂的结果比较. 当$\theta=0$° 时, 相比于未掺杂的结果(图2(g)), 掺杂引起了 ±0.2 eV附近的透射系数的显著降低, 但在其他能量, 如费米面(E = 0 eV)处, 透射系数依然为2, 与未掺杂的情况是相同的. 我们知道, ±0.2 eV附近的电子态对应于边缘电子态, 因而掺杂实际上是通过改变石墨烯纳米带的边缘原子结构, 从而影响边缘电子态及其透射系数. 我们再看不同磁化偏转角的情况, 当$\theta$变化时, 透射系数的变化主要表现在两个区域, 一个区域是E = ?0.65 eV附近, 另一个是费米面附近. 在第一个区域, 即E = ?0.65 eV附近, 存在一个较宽的dip且透射系数基本不随偏转角的变化而改变. 在另一个区域, 即费米面附近, 透射系数随着偏转角增大而逐渐减小. 这一特征表明: 在第一个区域中, 杂质散射效应占据主导地位, 引起透射系数降低的自旋翻转散射并不明显; 而在第二个区域, 自旋翻转散射效应更具优势, 杂质散射效应并不太突出. 为了解释这一现象, 我们在图7(b)中给出了投影在氮原子上和其他碳原子上的投影电子态密度. 通过与未掺杂石墨烯纳米带的比较, 可以看到碳原子的投影态密度主要由边缘电子态贡献, 而氮原子的投影态密度则主要由位于E = ?0.65 eV附近的新峰来贡献, 这个新峰对应着氮原子贡献的束缚态, 束缚态的能量位置与透射系数中dip的位置一致, 表明图7(a)中的dip结构代表着氮原子束缚态引起的传导电子的强的背散射. 而在其他能量位置, 如费米面附近, 氮原子的束缚态贡献不大, 因此电子输运由自旋翻转散射来决定. 图7(c)给出了硼原子掺杂的石墨烯纳米带的透射系数, 其主要特征与氮原子掺杂的结果类似, 只是由于硼原子的电荷转移方向与氮原子相反, dip结构出现在E = +0.65 eV处, 与图7(d)中给出的硼原子的束缚态的能量位置是一致的. 图 6$\theta=30$°时未掺杂与硼/氮掺杂的石墨烯纳米带的上边缘碳原子的(a)磁矩大小和(b)方向与碳原子位置之间的关系 Figure6. (a) Size and (b) direction of magnetic moment of atoms in upper edge of ZGNR as function of atom position at $\theta=30$°.
图 7 掺杂石墨烯纳米带的透射系数和投影态密度与能量之间的关系 (a), (b)为氮掺杂; (c), (d)为硼掺杂 Figure7. Transmission (top panel) and projected density of states (bottom panel) of doped ZGNRs as function of energy: (a), (b) N-doping; (c), (d) B-doping.