BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Teaching oxidation states to neural networks - Dr Cristiano Malica
 \, University of Bremen\, Germany
DTSTART:20251124T140000Z
DTEND:20251124T143000Z
UID:TALK239755@talks.cam.ac.uk
CONTACT:Dr Fabian Berger
DESCRIPTION:While the accurate description of redox reactions remains a ch
 allenge for first-principles calculations\, it has been shown that extende
 d Hubbard functionals (DFT+U+V) can provide a reliable approach\, mitigati
 ng self-interaction errors\, in materials with strongly localized d or f e
 lectrons. Here\, we first show that DFT+U+V molecular dynamics is capable 
 of following the adiabatic evolution of oxidation states over time\, using
  representative Li-ion cathode materials. In turn\, this allows to develop
  redox-aware machine-learning potentials. We show that considering atoms w
 ith different oxidation states (as accurately predicted by DFT+U+V) as dis
 tinct species in the training leads to potentials that are able to identif
 y the correct ground state and pattern of oxidation states for redox eleme
 nts present. This can be achieved\, e.g.\, through a systematic combinator
 ial search for the lowest-energy configuration or with stochastic methods.
  This brings the advantages of machine-learning potentials to key technolo
 gical applications (e.g.\, rechargeable batteries)\, which require an accu
 rate description of the evolution of redox states.\n\nC. Malica & N. Marza
 ri\, npj Computational Materials 11\, 212 (2025)
LOCATION:https://zoom.us/j/92447982065?pwd=RkhaYkM5VTZPZ3pYSHptUXlRSkppQT0
 9
END:VEVENT
END:VCALENDAR