RESUMEN
The ability to manufacture 3D metallic architectures with microscale resolution is greatly pursued because of their diverse applications in microelectromechanical systems (MEMS) including microelectronics, mechanical metamaterials, and biomedical devices. However, the well-developed photolithography and emerging metal additive manufacturing technologies have limited abilities in manufacturing micro-scaled metallic structures with freeform 3D geometries. Here, for the first time, the high-fidelity fabrication of arbitrary metallic motifs with sub-10 µm resolution is achieved by employing an embedded-writing embedded-sintering (EWES) process. A paraffin wax-based supporting matrix with high thermal stability is developed, which permits the printed silver nanoparticle ink to be pre-sintered at 175 °C to form metallic green bodies. Via carefully regulating the matrix components, the printing resolution is tuned down to ≈7 µm. The green bodies are then embedded in a supporting salt bath and further sintered to realize freeform 3D silver motifs with great structure fidelity. 3D printing of various micro-scaled silver architectures is demonstrated such as micro-spring arrays, BCC lattices, horn antenna, and rotatable windmills. This method can be extended to the high-fidelity 3D printing of other metals and metal oxides which require high-temperature sintering, providing the pathways toward the design and fabrication of 3D MEMS with complex geometries and functions.
RESUMEN
Neural networks based on memristive devices have achieved great progress recently. However, memristive synapses with nonlinearity and asymmetry seriously limit the classification accuracy. Moreover, insufficient number of training samples in many cases also have negative effect on the classification accuracy of neural networks due to overfitting. In this work, dropout neuronal units are developed based on stochastic volatile memristive devices of Ag/Ta2O5:Ag/Pt. The memristive neural network using the dropout neuronal units effectively solves the problem of overfitting and mitigates the negative effects of the nonideality of memristive synapses, eventually achieves a classification accuracy comparable to the theoretical limit. The stochastic and volatile switching performances of the Ag/Ta2O5:Ag/Pt device are attributed to the stochastical rupture of the Ag filament under high electrical stress in the Ta2O5 layer, according to the TEM observation and the kinetic Monte Carlo simulation.