Teaching optics to a machine learning network.

In this Letter, we demonstrate how harmonic oscillator equations can be integrated in a neural network to improve the spectral response prediction for an optical system. We use the optical properties of a one-dimensional nanoslit array for a practical implementation of the study. This method allows to build more generalizable relations between the input parameters of the array and its optical properties, showing a 20-fold improvement for parameters outside the range used for the training. We also show how this model generates the output spectrum from phenomenological relationships between the input parameters and the output spectrum, indicating how it grasps the physical mechanisms of the optical response of the structure.

Photon-induced generation and spatial control of extreme pressure at the nanoscale with a gold bowtie nano-antenna platform.

Precise spatial and temporal control of pressure stimulation at the nanometer scale is essential for the fabrication and manipulation of nano-objects, and for exploring single-molecule behaviour of matter under extreme conditions. However, state-of-the-art nano-mechanical transducers require sophisticated driving hardware and are currently limited to moderate pressure regimes. Here we report a gold plasmonic bowtie (AuBT) nano-antennas array that can generate extreme pressure stimulus of ∼100 GPa in the ps (10-12 s) time scale with sub-wavelength resolution upon irradiation with ultra-short laser pulses. Our method leverages the non-linear interaction of photons with water molecules to excite a nano-plasma in the plasmon-enhanced near-field and induce extreme thermodynamic states. The proposed method utilizes laser pulses, which in contrast to micro- and nano-mechanical actuators offers simplicity and versatility. We present time-resolved shadowgraphic imaging, electron microscopy and simulation data that suggest that our platform can efficiently create cavitation nano-bubbles and generate intense pressure in specific patterns, which can be controlled by the selective excitation of plasmon modes of distinct polarizations. This novel platform should enable probing non-invasively the mechanical response of cells and single-molecules at time and pressure regimes that are currently difficult to reach with other methods.