Classification and Prediction of Skyrmion Material Based on Machine Learning.

The discovery and study of skyrmion materials play an important role in basic frontier physics research and future information technology. The database of 196 materials, including 64 skyrmions, was established and predicted based on machine learning. A variety of intrinsic features are classified to optimize the model, and more than a dozen methods had been used to estimate the existence of skyrmion in magnetic materials, such as support vector machines, k-nearest neighbor, and ensembles of trees. It is found that magnetic materials can be more accurately divided into skyrmion and non-skyrmion classes by using the classification of electronic layer. Note that the rare earths are the key elements affecting the production of skyrmion. The accuracy and reliability of random undersampling bagged trees were 87.5% and 0.89, respectively, which have the potential to build a reliable machine learning model from small data. The existence of skyrmions in LaBaMnO is predicted by the trained model and verified by micromagnetic theory and experiments.

In situ TEM study on diversified martensitic transition behaviour in Ni50Mn35In15 alloys.

Ni-Mn-In magnetic shape-memory alloys are attractive materials due to their important functional properties relating to the martensitic transition. Understanding the complex martensitic magnetism and the transition process is of crucial importance not only from a fundamental but also from a technological point of view. Here, we demonstrate the dynamic magnetic domains and microstructures during the martensitic transition in the bulk and melt-spun ribbons of Ni50Mn35In15via in situ Lorentz transmission electron microscopy. The significant evolutionary differences in correlation with the temperature dependence of magnetization are identified between the bulk and ribbons. For a bulk alloy with L21 crystal structure at room temperature, the complete martensite with 7 M modulation in the paramagnetic state and the successive stripe magnetic domains in ferromagnetic martensite develop with a further decrease in the temperature. The stripe domains evolve into biskyrmion-like spin configurations when a perpendicular magnetic field is applied. In contrast, the partial austenitic phase always coexists with the martensitic phase in the ribbons even far below the martensitic transition temperatures and the martensitic phase presents a dominant twinning stack morphology with 5 M modulation and various magnetic domains. During the subsequent reheating-cooling cycles, the thermal hysteresis behavior and the transition reversibility in the bulk and ribbons are represented via the microstructural evolution.

Anisotropic nonvolatile magnetization controlled by electric field in amorphous SmCo thin films grown on (011)-cut PMN-PT substrates.

The tunable, nonvolatile electrical modulation of magnetization at room temperature is firstly demonstrated in a magnetically hard amorphous SmCo film grown on a (011)-cut 0.7Pb(Mg1/3Nb2/3)O3-0.3PbTiO3 (PMN-PT) substrate. Uniaxial in-plane anisotropy with hard and easy axes lying in the [100] and [01-1] directions, respectively, occurs. Bipolar electric field, E, across the thickness direction enhances the remnant magnetization, Mr, along the hard axis, while suppresses the Mr along the easy axis, and the maximal regulation is about -5.8% and +2.2%, respectively. Detailed analysis indicates that the induced effective uniaxial magnetic anisotropy field, which arises from the magnetostrictive properties of the amorphous SmCo thin film and the anisotropic strain from the PMN-PT substrate, is mainly responsible for the anisotropic tunability. The variation of the directional pair ordering of the SmCo film, which is caused by the anisotropic strain due to the electric field, also contributes to the tunability. More importantly, nonvolatile modulation and a stable two-state memory effect are demonstrated for the bipolar case, and in situ X-ray diffraction and X-ray diffraction reciprocal space mapping reveal that these phenomena originate from the electric-field-induced rhombohedral-orthorhombic phase transformation in the PMN-PT substrate. Moreover, by unipolarizing the ferroelectric substrate, a nonvolatile modulation is also observed. The anisotropic nonvolatile control of magnetization in SmCo amorphous films opens a new avenue for developing multifunctional information storage and novel spintronics devices based on hard magnetic materials.

Strong textured SmCo5 nanoflakes with ultrahigh coercivity prepared by multistep (three steps) surfactant-assisted ball milling.

The high coercivity of 26.2 kOe for SmCo5 nanoflakes are obtained by multistep (three steps) surfactant-assisted ball milling. The magnetic properties, phase structure and morphology are studied by VSM, XRD and SEM, respectively. The results demonstrate that the three step ball-milling can keep more complete crystallinity (relatively less defects) during the process of milling compared with one step high energy ball-milling, which enhances the texture degree and coercivity. In addition, the mechanism of coercivity are also studied by the temperature dependence of demagnetization curves for aligned SmCo5 nanoflakes/resin composite, the result indicates that the magnetization reversal could be controlled by co-existed mechanisms of pinning and nucleation.