Resistance switching in oxides with inhomogeneous conductivity

Electric-field-induced resistance switching （RS） phenomena have been studied for over 60 years in metal/dielectrics/metal structures. In these experiments a wide range of dielectrics have been studied including binary transition metal oxides, perovskite oxides, chalcogenides, carbon- and silicon-based materials, as well as organic materials. RS phenomena can be used to store information and offer an attractive performance, which encompasses fast switching speeds, high scalability, and the desirable compatibility with Si-based complementary metal-oxide-semiconductor fabrication. This is promising for nonvolatile memory technology, i.e., resistance random access memory （RRAM）. However, a comprehensive understanding of the underlying mechanism is still lacking. This impedes faster product development as well as accurate assessment of the device performance potential. Generally speaking, RS occurs not in the entire dielectric but only in a small, confined region, which results from the local variation of conductivity in dielectrics. In this review, we focus on the RS in oxides with such an inhomogeneous conductivity. According to the origin of the conductivity inhomogeneity, the RS phenomena and their working mechanism are reviewed by dividing them into two aspects： interface RS, based on the change of contact resistance at metal/oxide interface due to the change of Schottky barrier and interface chemical layer, and bulk RS, realized by the formation, connection, and disconnection of conductive channels in the oxides. Finally the current challenges of RS investigation and the potential improvement of the RS performance for the nonvolatile memories are discussed.     

基于忆耦器实现神经突触可塑性和神经网络模拟

正现代计算机自问世以来一直采用冯·诺依曼结构,即运算器与存储器分离,这种结构使得运算器与存储器之间的数据传输成为影响系统性能的瓶颈(称为冯·诺依曼瓶颈[1]),大大限制了计算机性能的提高;同时,由于现代计算机中的运算器和主存储器都是易失性器件,不仅在断电后信息立即消失,而且具有较高的能耗。作为对比,人类的大脑是一个高效的信息存储与计算系统,而且具有非常低的功耗(约20W)。主要原因在于     

基于忆阻器阵列的下一代储池计算

储池计算是类脑计算范式的一种,具有结构简单、训练参数少等特点,在时序信号处理、混沌动力学系统预测等方面有着巨大的应用潜力.本文提出了一种基于存内计算范式的储池计算硬件实现方法,利用忆阻器阵列完成非线性向量自回归过程中的矩阵向量乘法操作,有望进一步提升储池计算的能效.通过忆阻器阵列仿真实验,在Lorenz63时间序列预测任务中验证了该方法的可行性,以及该方法在噪声条件下预测结果的鲁棒性,并探究忆阻器阵列阻值精度对预测结果的影响.这一结果为储池计算的硬件实现提供了一种新的途径.     

亚音速火焰喷涂Y2O3-HA生物涂层的研究

采用亚音速火焰喷涂方法，通过合理选用打底层粉和工作层粉，探讨喷涂羟基磷灰石生物涂层的可行性。试验证明：以纯钛粉作为打底层粉，在喷涂过程中靠粉末燃烧自放热效应，在打底层中形成了TiO2，Ti2O3和TiN相，造成部分质点达到微冶金结合。该相与Ti6AhV基材线膨胀系数相近，喷涂中产生的热应力较小，获得了较高的结合强度。在喷涂过程中，HA粉发生了一定量的分解，最终涂层形成HA相、TTCP和TCP相。而加入5％Y2O3的HA粉所形成的涂层，HA含量高，形成的TTCP相和TCP相少，并有Y2O3相产生，且涂层贯穿裂纹减少，孔隙增多，利于人工种植，提高结合强度。     

Realization of a flux-driven memtranstor at room temperature

The memtranstor has been proposed to be the fourth fundamental circuit memelement in addition to the memristor,memcapacitor, and meminductor. Here, we demonstrate the memtranstor behavior at room temperature in a device made of the magnetoelectric hexaferrite(Ba_(0.5)Sr_(1.5)Co_2Fe_(11) AlO_(22)) where the electric polarization is tunable by external magnetic field. This device shows a nonlinear q–р relationship with a butterfly-shaped hysteresis loop, in agreement with the anticipated memtranstor behavior. The memtranstor, like other memelements, has a great potential in developing more advanced circuit functionalities.     

基于磁电耦合效应的基本电路元件和非易失性存储器

磁电耦合效应是指磁场控制电极化或者电场控制磁性的物理现象,它们为开发新型电子器件提供了额外的物理状态自由度,具有巨大的应用潜力.磁电耦合系数作为磁电耦合材料的重要参量,体现了材料磁化和电极化的耦合性能,其随外加物理场的变化可以表现出非线性回滞行为,具备作为非易失存储的物理状态特征.本文讨论了基于磁电耦合效应如何建立起电荷-磁通之间的直接关联,继而实现了第四种基本电路元件并构建了完整的电路元件关系图.在此基础上,研究了多铁性异质结中的非线性磁电耦合效应,并利用其独特的电荷-磁通关联特性,开发了基于磁电耦合系数的电写-磁读型非易失性信息存储、逻辑计算与类神经突触记忆等一系列新型信息功能器件.     

Magnetocaloric effect in the layered organic–inorganic hybrid(CH_3NH_3)_2CuCl_4

We present a study of magnetocaloric effect of the quasi-two-dimensional(2D) ferromagnet(CH_3NH_3)_2CuCl_4 in ab plane(easy-plane). From the measurements of magnetic field dependence of magnetization at various temperatures,we have discovered a large magnetic entropy change associated with the ferromagnetic–paramagnetic transition. The heat capacity measurements reveal an abnormal adiabatic change below the Curie temperature T_c~8.9 K, which is caused by the nature of quasi-2D layered crystal structure. These results suggest that perovskite organic–inorganic hybrids with a layered structure are suitable candidates as working substances in magnetic refrigeration technology.