NCCR MARVEL staff wins award for ground-break

picture: The analysis was featured on the quilt of Nature Nanotechnology when it was first revealed.
view Moree

Credit score: Nature Nanotechnology

Investigations into an interesting class of novel supplies that might underpin next-generation digital and optoelectronic purposes and quantum computing has received a staff led by Nicola Marzari the inaugural PRACE (Partnership for Superior Computing in Europe) HPC Excellence Award. It was established to honor “essentially the most distinguished tasks and researchers within the area of high-performance computing.”

The award-winning analysis was initially described within the 2018 Nature Nanotechnology article “Two-dimensional supplies from high-throughput computational exfoliation of experimentally identified compounds,” whose objective was to speed up supplies discovery: whereas Nobel-prize successful work producing and characterizing graphene had moved 2D supplies from concept to actuality some 15 years prior, progress in figuring out novel 2D supplies was sluggish – only some had been recognized experimentally.

Whereas it was clear {that a} computational method may assist deal with the problem of figuring out which of the a whole bunch of 1000’s of identified compounds may really be exfoliated or be secure as 2D monolayers, elementary points remained. How may they set up whether or not or not a crystal construction may very well be exfoliated? How ought to they deal with and ultimately carry out simulations over an in depth set of candidate supplies? How may they streamline predictions in order that they’d extra reliably and robotically go from construction to property?

Marzari and his staff answered all these questions with a knowledge and computing intensive method, managing to slender a down a area of 108,423 distinctive, experimentally identified inorganic compounds to only below 2000 that may very well be exfoliated into novel 2D supplies. They then investigated the vibrational, digital, magnetic and topological properties of a subset of essentially the most attention-grabbing 258 compounds, figuring out novel magnetic supplies, half-metals, and half-semiconductors. The richness of this portfolio was then explored in additional than a dozen extra research, screening for optimum efficiency in numerous purposes, from digital transport, to superconductivity, to band topology. This led to the invention of the primary ever Kane-Mele topological insulator, to the fabric with largest superconducting temperature in 2D, and essentially the most performant spin-FET transistor.

None of this is able to have been potential and not using a novel method to computing and information. The unique high-throughput research required the staff to carry out half one million calculations on 1000’s of various supplies, typically combining a number of codes to focus on advanced properties. Doing this by hand not solely would have required an unlimited effort and appreciable human energy, however would have made it vulnerable to errors and never reproducible.

This problem drove the staff to method HPC in a radically new method, determining methods to cope with these calculations robotically, robustly, and reliably, whereas preserving a completely reproducible report of all the calculation protocols and workflows. The event of the core informatics platform of MARVEL was key: AiiDA (https://www.aiida.internet) because the infrastructure to automate, handle, persist, share and reproduce all of the advanced workflows and information, and the Supplies Cloud ( to disseminate these to the neighborhood at giant. AiiDA is ready to deal with concurrently 1000’s of calculations, automating the submission and management course of in addition to the retrieval and storage of the outcomes, and one can then expose all the workflow and the uncooked and curated information on the Supplies Cloud.

Notably, the staff that may share the award – all of whom have been at EPFL when the work was carried out – has now moved to unbiased positions worldwide: Nicolas Mounet, analysis scientist at CERN; Marco Gibertini, assistant professor on the College of Modena and Reggio Emilia; Philippe Schwaller, assistant professor at EPFL, Davide Campi, assistant professor on the College of Milan-Bicocca; Andrius Merkys, researcher at Vilnius College; Antimo Marrazzo, junior assistant professor on the College of Trieste; Thibault Sohier, researcher at CNRS, Laboratoire Charles Coulomb; Ivano E. Castelli, affiliate professor on the Technical College of Denmark; Andrea Cepellotti, analysis scientist at Harvard College; Giovanni Pizzi, senior scientist at EPFL and group chief at PSI, and Nicola Marzari at EPFL. And naturally they’re all very grateful to the Swiss Nationwide Supercomputing Middle and to PRACE for offering the help and computing assets that underpinned all of the analysis.

The prize can be introduced at PASC22held from the 27 to the 29 of June 2022 in Basel, Switzerland.

Disclaimer: AAAS and EurekAlert! will not be accountable for the accuracy of reports releases posted to EurekAlert! by contributing establishments or for using any data by the EurekAlert system.

Leave a Comment