Dataset for "Deep learning enabled strategies for modelling of complex aperiodic plasmonic metasurfaces of arbitrary size" DOI:10.5258/SOTON/D2063 Readme author: Peter Wiecha, University of Southampton / LAAS-CNRS Toulouse, France Supports the paper Clément Majorel, Christian Girard, Arnaud Arbouet, Otto L. Muskens, Peter R. Wiecha, "Deep learning enabled strategies for modelling of complex aperiodic plasmonic metasurfaces of arbitrary size". ACS Photonics, https://doi.org/10.1021/acsphotonics.1c01556 (2022) DOI:10.1021/acsphotonics.1c01556 ---------------------------- "dataset.zip" containing folders "figX" with X from 1 to 8 raw data used for producing the main paper figures. Each figure's dataset is bundled in an according subfolder. The data is labeled in an understandable form, if applicable headers have been added to raw-data textfiles. ---------------------------- full training datasets: dataset_1000x1000nm2_wavelength635nm.zip - dataset_1000x1000nm2_wavelength635nm.h5 Dataset for fixed wavelength training (at plasmon resonance of a single goldrod on glass: 635nm). Contains around 65,000 simulations of random gold-rod assemblies in an area of 1000x1000 nm2, in maps of 100x100 pixels. Each pixel corresponds to a step of 10nm. 'geo_2d': top view images of the structures. 'dressed_alpha_2d': 18 maps corresponding to real (first 9) and imaginary parts (last 9) of the 9 components of the dressed polarizability tensors. Component order is: xx,xy,xz,yx,yy,yz,zx,zy,zz. dataset_600x600nm2_spectral.zip - dataset_600x600nm2_spectral.h5 Dataset for spectral wavelength training in a range from 600nm to 800nm. Contains around 150,000 simulations of random gold-rod assemblies in an area of 600x600 nm2, in maps of 60x60 pixels. Each pixel corresponds to a step of 10nm. 'geo_2d': top view images of the structures. 'dressed_alpha_2d': 18 maps corresponding to real (first 9) and imaginary parts (last 9) of the 9 components of the dressed polarizability tensors. Component order is: xx,xy,xz,yx,yy,yz,zx,zy,zz. 'wl': wavelength of the dressed polarizability tensor of each sample.