Organometal halide perovskites have demonstrated optoelectronic properties which make them a promising potential emission layer in charge separating devices such as light-emitting diodes (LEDs). Excellent charge carrier mobilities and low exciton binding energies are factors that contribute to low non-radiative recombination rates, which is imperative for efficient LED operation. Additionally, high crystallinity and low defect density bolster perovskite’s efficiencies as a light-emitting material. In order to further outperform trapping and the forthcoming non-radiative recombination, quasi-2D perovskites can be used to incentivize additional radiative recombination. Here we report a method for 2D/Quasi-2D layered perovskite light-emitting diodes.

Biological materials present an abundance of structures that can serve as an inspiration for designs of new synthetic materials for various technological applications. In particular, the horse hoof yields outstanding mechanical properties with excellent fracture control mechanisms. Thus, the hoof was studied for inspiration in designing high impact resistant, compressive, and tensile strength materials. The horse hoof structure consists of a hierarchical assembly of helical, layered, tubular and cellular microstructures. In order to deepen our fundamental understanding of these microstructures, we identified the most prominent structures through prior research into the hoof’s energy absorbent properties. Then, numerous models were designed with varying metrics of the structures such as tubule reinforcement and tubule shape, size, and density gradients. Models were fabricated using multi-material additive manufacturing and their characterization provided a comprehensive understanding of how tubular and gradient features affect fracture. The bioinspired models were tested using drop tower, compact tension, and hopkinson bar, exhibiting behavior similar to the observed phenomena in the hoof. Findings regarding these bioinspired models provide insight into the complexity of the hoof structure and offer guidance for future bioinspired technology. This work was supported by the National Science Foundation Mechanics of Materials and Structures Program (Grant Numbers 1926353 and 1926361).

In the introductory physics curriculum, induction is taught by describing the behavior of coils and solenoids. It is then natural to introduce inductance by having students imagine what happens when a solenoid is part of an electrical circuit. However, it is perhaps useful to ask students early on, “How could you build a micro-scale inductor?” In industry, micro-and nanoelectronics are constructed using various thin film deposition and lithography techniques, and an inductor can be created simply by depositing a metallic film of copper etched into the shape of a spiral [2]. These planar inductors are core components for high-frequency antennae [ref] and can be realized by students in a hands-on way without the need for the expensive equipment that resides in a cleanroom. All the instructor needs are some paper, copper tape, and solder.

Entity linking is an uncommon term but is relevant to daily life for many people. Entity linking is the process of using recognizable real-world objects, or named entities, in text and linking them to a corresponding entry in a database with more information, such as Wikidata. This can be used for other machine learning techniques such as text retrieval, which is most associated with search engine queries. The two main approaches to entity linking are end-to-end and disambiguation-only approaches. End-to-end is a two-step process involving finding the named entities through Named Entity Recognition and disambiguating those entities to a database. Disambiguation-only takes gold-standard entities which have been manually selected and only disambiguates them from a database like the name suggests. Previous research has recognized that disambiguation-only technique is more accurate than end-to-end, however any model that involves manual manipulation is not ideal. Therefore, this research will compare both models with the use of the revolutionary transformer model for the process of Named Entity Recognition in the end-to-end model and determine if the model’s efficiency can improve the end-to-end approach. For the purposes of this research, the models will use a novel knowledge base for entity recognition and the Wikidata database will be used. Accuracy will be determined using InKb F1 scores and the end-to-end model will be processed through a General Entity Annotator Benchmark.