Selecting Machine Learning Models to Support the Design of Al/CuO Nanothermites.

Novel properties associated with nanothermites have attracted great interest for several applications, including lead-free primers and igniters. However, the prediction of quantitative structure-energetic performance relationships is still challenging. This study investigates machine learning methods as tools to surrogate complex physical models to design novel nanothermites with optimized burning rates chosen for energetic performance. The study focuses on Al/CuO nanolaminates, for which nine supervised regressors commonly used in ML applied to materials science are investigated. For each, an ML model is built using a database containing a set of 2700 Al/CuO nanolaminate systems, specifically generated for this study. We demonstrate the superiority of the multilayer perceptron algorithm to surrogate conventional physical-based models and predict the Al/CuO nanolaminate microstructure-burn rate relationship with good efficiency: the burn rate is estimated with less than 1% error (0.07 m·s-1), which is very good for designing nano-engineered energetic materials, knowing that it typically varies from approximately 8-20 m·s-1. In addition, the optimization of the Al/CuO nanolaminate structure for burn rate maximization through machine learning takes a few milliseconds, against several days to achieve this task using a physical model, and months experimentally.

Design of mesoscopic self-collimating photonic crystals under oblique incidence.

Mesoscopic Photonic Crystals (MPhCs) are composed of alternating natural or artificial materials with compensating spatial dispersion. In their simplest form, as presented here, MPhCs are composed by the periodic repetition of a MPhC supercell made of a short slab of bulk material and a short slab of Photonic Crystal (PhCs). Therefore, MPhCs present a multiscale periodicity with a subwavelength periodicity within each PhC slab and with a few-wavelength periodicity for its supercell. Thanks to this mesoscopic structure, MPhCs allow the self-collimation of light, through a mechanism called mesoscopic self-collimation (MSC), along both directions of high symmetry and directions oblique with respect to the MPhCs slab interfaces. Here, we propose a new design method useful for conceiving MPhCs that allow MSC under oblique incidence, avoiding in-plane scattering and ensuring propagation via purely guided modes, without out-of-plane radiation losses. In addition, the proposed method allows a systematic search for optimal MSC structures, which also simultaneously satisfy the impedance matching condition at MPhC interfaces, thus reducing the effect of multiple reflections between bulk-PhC interfaces. The proposed design method has the advantage of an extreme analytical simplicity and it allows direct design of oblique-incidence MPhC structures. Its accuracy is validated through Finite Difference Time Domain simulations and the MSC performances of the designed structures are evaluated, in terms of angular direction, beam waist, overall transmittance, and through discussion of a Figure of Merit that accounts for residual beam curvature. This simple yet powerful method can pave the way for the design of advanced MSC-based photonic interconnects and circuits that are immune to crosstalk and out-of-plane losses.

General Strategy for the Design of DNA Coding Sequences Applied to Nanoparticle Assembly.

The DNA-directed assembly of nano-objects has been the subject of many recent studies as a means to construct advanced nanomaterial architectures. Although much experimental in silico work has been presented and discussed, there has been no in-depth consideration of the proper design of single-strand sticky termination of DNA sequences, noted as ssST, which is important in avoiding self-folding within one DNA strand, unwanted strand-to-strand interaction, and mismatching. In this work, a new comprehensive and computationally efficient optimization algorithm is presented for the construction of all possible DNA sequences that specifically prevents these issues. This optimization procedure is also effective when a spacer section is used, typically repeated sequences of thymine or adenine placed between the ssST and the nano-object, to address the most conventional experimental protocols. We systematically discuss the fundamental statistics of DNA sequences considering complementarities limited to two (or three) adjacent pairs to avoid self-folding and hybridization of identical strands due to unwanted complements and mismatching. The optimized DNA sequences can reach maximum lengths of 9 to 34 bases depending on the level of applied constraints. The thermodynamic properties of the allowed sequences are used to develop a ranking for each design. For instance, we show that the maximum melting temperature saturates with 14 bases under typical solvation and concentration conditions. Thus, DNA ssST with optimized sequences are developed for segments ranging from 4 to 40 bases, providing a very useful guide for all technological protocols. An experimental test is presented and discussed using the aggregation of Al and CuO nanoparticles and is shown to validate and illustrate the importance of the proposed DNA coding sequence optimization.