Multi-physical modelling, design optimization and manufacturing of a composite dielectric solar absorber

The work presented involves the multiphysical modelling, simulation and design optimization of a key component of a Solar Selective Coatings (SSC). The investigated SSC absorber consists of a near homogeneous distribution of nanoparticles of Titanium Nitride (TiN) in a matrix of Aluminium Nitride (AlN), to form a composite dielectric. With the aim of achieving high absorbance in the visible region of the spectrum and minimum reflectance in the infrared region of the spectrum, our work highlights the numerical design, the synthesis and optical characterization of a composite dielectric of approximately 500 nm thickness. A bottom-up approach for the preparation of a stack with alternate layers, consisting of a distribution of TiN nanoparticles with a layer of AlN on top, was adopted. The TiN nanoparticles, laid on a substrate (Silicon/Glass) by wet chemical method, are coated with conformal layer of AlN, via Plasma-enhanced Atomic Layer Deposition (PE-ALD). The control of the morphology at the nanoscale is fundamental in improving the optical performance of the material. For this reason, two composites were prepared. One starting with TiN dispersions made with dry TiN powder and deionized water, and the other with ready-made TiN dispersions. In both composites, the particles were 20–30 nm in diameter. In both the cases, fewer clusters of about 0.5–1 μmof TiN particles were present however, enough steps were taken to minimize these clusters into smaller particles. Parameters, such as the size of TiN nanoparticles, the thickness of AlN thin film, were revealed by the numerical simulations, performed using Wave-Optics module in COMSOL Multiphysics. The work showcased clearly compares the two kinds of composites, using scanning electron microscope, X-ray photoelectron spectroscopy and electrical conductivity measurement. In addition, the optical performance of the two prepared composites is used as a means of validating the computational model.