Tailoring the degree of disorder in chalcogenide phase‐change materials (PCMs) plays an essential role in nonvolatile memory devices and neuro‐inspired computing. Upon rapid crystallization from the amorphous phase, the flagship Ge–Sb–Te PCMs form metastable rocksalt‐like structures with an unconventionally high concentration of vacancies, which results in disordered crystals exhibiting Anderson‐insulating transport behavior. Here, ab initio simulations and transport experiments are combined to extend these concepts to the parent compound of Ge–Sb–Te alloys, viz., binary Sb2Te3, in the metastable rocksalt‐type modification. Then a systematic computational screening over a wide range of homologous, binary and ternary chalcogenides, elucidating the critical factors that affect the stability of the rocksalt structure is carried out. The findings vastly expand the family of disorder‐controlled main‐group chalcogenides toward many more compositions with a tunable bandgap size for demanding phase‐change applications, as well as a varying strength of spin–orbit interaction for the exploration of potential topological Anderson insulators.

Phase change materials (PCMs) are key to the development of artificial intelligence technologies such as high‐density memories and neuromorphic computing, thanks to their ability for multi‐level data storage through stepwise resistive encoding. Individual resistance levels are realized by adjusting the crystalline and amorphous volume fraction of the memory cell. However, the amorphous phase exhibits a drift in resistance over time that has so far hindered the commercial implementation of multi‐level storage schemes. In this study, the underlying physical process of resistance drift with the goal of modeling is elucidated that will help minimize and potentially overcome drift in PCM memory devices. Clear evidence is provided that the resistance drift is dominated by glass dynamics. Resistivity convergence and drift inversion for the amorphous chalcogenide Ge15Te85 and the PCM Ge3Sb6Te5 are experimentally demonstrated and these changes are successfully predicted with a glass dynamics model. This new insight into the resistance drift process provides tools for the development of advanced PCM devices.

No abstract

A new method to describe statistical information from passive scalar fields has been proposed by Wang and Peters ͓"The length-scale distribution function of the distance between extremal points in passive scalar turbulence," J. Fluid Mech. 554, 457 ͑2006͔͒. They used direct numerical simulations ͑DNS͒ of homogeneous shear flow to introduce the innovative concept. This novel method determines the local minimum and maximum points of the fluctuating scalar field via gradient trajectories, starting from every grid point in the direction of the steepest ascending and descending scalar gradients. Relying on gradient trajectories, a dissipation element is defined as the region of all the grid points, the trajectories of which share the same pair of maximum and minimum points. The procedure has also been successfully applied to various DNS fields of homogeneous shear turbulence using the three velocity components and the kinetic energy as scalar fields ͓L. Wang and N. Peters, "Length-scale distribution functions and conditional means for various fields in turbulence," J. Fluid Mech. 608, 113 ͑2008͔͒. In this spirit, dissipation elements are, for the first time, determined from experimental data of a fully developed turbulent channel flow. The dissipation elements are deduced from the gradients of the instantaneous fluctuation of the three velocity components uЈ, vЈ, and wЈ and the instantaneous kinetic energy kЈ, respectively. The measurements are conducted at a Reynolds number of 1.7ϫ 10 4 based on the channel half-height ␦ and the bulk velocity U. The required three-dimensional velocity data are obtained investigating a 17.75ϫ 17.75ϫ 6 mm 3 ͑0.355␦ ϫ 0.355␦ ϫ 0.12␦͒ test volume using tomographic particle-image velocimetry. Detection and analysis of dissipation elements from the experimental velocity data are discussed in detail. The statistical results are compared to the DNS data from Wang and Peters ͓"The length-scale distribution function of the distance between extremal points in passive scalar turbulence," J. Fluid Mech. 554, 457 ͑2006͒; "Length-scale distribution functions and conditional means for various fields in turbulence," J. Fluid Mech. 608, 113 ͑2008͔͒. Similar characteristics have been found especially for the pdf's of the large dissipation element length regarding the exponential decay. In agreement with the DNS results, over 99% of the experimental dissipation elements possess a length that is smaller than three times the average element length.

Wiley-VCH GmbH (Pb) were employed with a wave-function cutoff of 30 Ry leading to energy convergence below 2 meV per atom in the elemental phases. For disordered systems, a 2 × 2 × 2 k-point mesh shifted away from the Γ point was used to obtain energy convergence below 1 mRy per atom and to avoid the sampling of extrema of spurious defect-level bands. For the ordered systems, a shifted 2 × 2 × 1 k-point mesh was used.Research data are not shared.