No abstract

No abstract

We report magnetism of tetragonal β-Fe3Se4 nanoplates controllably synthesized by thermal decomposition at 603 K of inorganic–organic (β-Fe2Se3)4[Fe(tepa)] hybrid nanoplates (tepa = tetraethylenepentamine). (β-Fe2Se3)4[Fe(tepa)] hybrid precursor and β-Fe3Se4 nanoplates are in single crystal features as characterized by selected area electron diffraction. Rietveld refinements reveal that ordered inorganic–organic (β-Fe2Se3)4[Fe(tepa)] hybrid nanoplates are in a tetragonal layered crystal structure with a space group of I4cm (108) and room-temperature lattice parameters are a = 8.642(0) Å and c = 19.40(3) Å, while the as-synthetic tetragonal β-Fe3Se4 nanoplates have a layered crystal structure with the P4/nmm space group, and room-temperature lattice parameters are a = 3.775(8) Å and c = 5.514(5) Å. Magnetic measurements show the weak ferrimagnetism for (β-Fe2Se3)4[Fe(tepa)] hybrid nanoplates at room temperature, while the as-synthetic β-Fe3Se4 nanoplates are antiferromagnetic in a temperature range between 120 and 420 K but in a ferrimagnetic feature below ~120 K. The as-synthetic β-Fe3Se4 nanoplates are thermally instable, which are transformed to ferrimagnetic β-Fe3Se4 nanoplates by annealing at 623 K (a little higher than the synthetic temperature). There is an irreversible change from antiferromagnetism of the as-synthetic β-Fe3Se4 phase to the ferrimagnetism of the as-annealed β-Fe3Se4 phase in a temperature between 420 and 470 K. Above 470 K, the tetragonal β-Fe3Se4 phase transforms to monoclinic Fe3Se4 phase with a Curie temperature (TC) of ~330 K. This discovery highlights that crystal structure and magnetism of Fe-Se binary compounds are highly dependent on both their phase compositions and synthesis procedures.

Recently, deep learning (DL) based automatic modulation classification (AMC) has received much attention. Various network structures with higher complexity are utilized to boost the performance of classification model. We divide the issue of AMC into two objectives and propose a disentangled approach with a signal processing module. Unlike popular end-to-end training strategy, we first consider a simple model with much fewer trainable parameters to learn accurate modulation features for classification. Then a U-net based signal processing module using a specially designed function is introduced to transfer the knowledge stored in classification module. We compare the performance of the proposed method with several baseline models on two well known datasets. Experimental results demonstrate that the proposed method gives superior performance with lower computational complexity compared with other methods. Furthermore, we also verify the feasibility and huge potential of the knowledge transferring in the field of wireless communications. INDEX TERMS Automatic modulation classification, deep learning, knowledge transfer.

No abstract

No abstract

No abstract

No abstract