Antiferromagnetic Hall-Memristors

Abstract

Spin-memristors are a class of materials that can store memories through the control of spins, potentially leading to novel technologies that address the constraints of standard silicon electronics, thereby facilitating the advancement of more intelligent and energy-efficient computing systems. In this work, we present a spin-memristor based on antiferromagnetic materials that exhibit Hall-memresistance. Moreover, the nonlinear Edelstein effect acts as both a writer and eraser of memory registers. We provide a generic symmetry-based analysis that supports the viability of the effect. To achieve a concrete realization of these ideas, we focus on CuMnAs, which has been shown to have a controllable nonlinear Hall effect. Our results extend the two-terminal spin-memristor setting, which is customarily the standard type of device in this context, to a four-terminal device.

Figures (5)

• Figure 1: Schematic picture of the antiferromagnetic Hall-memristor. An electric field is applied in the $x$-direction through a voltage $V_L$, then a transversal current is generated and measured through a transversal voltage $V_H$. The memory effect is caused by the NLEE, which alters the Néel vector in the material when an electric field is applied, thereby changing the magnetic coupling and transport properties of the material. (b) shows our proposed circuit diagram symbol for the Hall-memristor.
• Figure 2: (a), (b), (c) and (d) show the intrinsic conductance and magnetic susceptibilities of the Titled Massive Dirac Model, in function of the chemical potential $\mu$ and $\Delta_x$. In (b) and (d) $\mu = 0.2\; eV$. (e) The current vs electric field cycle is performed using Eq.(\ref{['eq:phenomenological']}) for various values of $\omega\tau$, and the inset is the plot of the normalized area of the hysteresis loop as a function of $\omega\tau$. The parameters used are $v_F=1 \, eV$Å, $v_t=0.1 \, eV$Å, $\Delta_z = 0.2\,eV$, $J = 100\; eV$Å$^2/\hbar$ and $E_0 = 1\; eV/$Å.
• Figure 3: In (a) are the band structure of CuMnAs, it shows that at $\boldsymbol{k} = (1, 0.5)\pi$ are the avoided crossings, whose vicinity mostly contributes to the NLHE and NLEE. In (b) we can see the $k_y$ dependence of this avoided crossings.
• Figure 4: (a),(b),(c) and (d) shows the intrinsic conductance and magnetic susceptibilities for CuMnAs, as a function of the chemical potential $\mu$ and $h^x_{\rm AFM}$. For (b) and (d) $\mu=0\,eV$. (e) The current vs electric field cycle is performed for various values of $\omega\tau$, and the inset is the plot of the normalized area as a function of $\omega\tau$. The value of the parameters are $t = 0.08 \, eV$, $\tilde{t}=1\,eV$, $\alpha_R=0.8 \, eV$, $\alpha_D=0$, $h^y_{\rm AFM}=h^z_{\rm AFM}=0$, $J = 16\; eV$Å$^2/\hbar$ and $E_0 = 0.01\; eV/$Å. The full cycle, including negative $E$, is in the Supplementary Material Fig. S1 SM.
• Figure 5: Neuromorphic possibilities for the Hall-Memristor. (a) Proposed circuit for simulating a neuron. (b) output current measure in response to a sequence of pulses. This emulates the connection between neurons, which repeatedly interact, and when a certain number of interactions are made, a pulse is fired. In this memristor, the firing process is possible due to the storage of AF-magnetization that alters the conductivity. More detail in SM, Figures S2-S5.