Carlos R. Salazar

IFM (Theoretical Physics)

Linköping University, Campus Valla

581 83 Linköping

I am a Guatemalan PhD student in Theoretical Physics at Linköping University, Sweden, working with Prof. Davide Sangiovanni. I study how hard-coating ceramics and metals behave at high temperature, combining ab initio molecular dynamics with machine-learning interatomic potentials. My current work looks at how Al and Si substitution balances strength and toughness in Ti–Al–Si–N coatings, at anomalous diffusion in bcc metals close to melting, and at the Hard-coating Alloy DataBase (HADB), which pairs high-throughput simulations of nitride alloys with machine-learning predictions.

Before coming to Linköping, I studied physics at Jacobs University Bremen (BSc) and Aix-Marseille Université (MSc), and did research internships at CEMES and UMET (CNRS) with Dr. Julien Lam. That work on machine-learning force fields and local structure classification led to papers on how zinc oxide nanocrystals nucleate and grow, and on identifying crystal phases and defects in complex oxides.

news

Sep 01, 2026	Two new preprints from my PhD work. In the first, ab initio molecular dynamics shows how adding Si to Ti–Al–N coatings activates transformation plasticity and delays crack growth, with the best balance of strength and toughness at comparable Al and Si contents. The second presents the extended Hard-coating Alloy DataBase (HADB), which combines high-throughput elastic and deformation data for nitride alloys with machine-learning predictions.
Jun 15, 2026	Our work on growth-driven phase transitions in zinc oxide nanoparticles is published in Nanoscale. Atom-by-atom deposition simulations show how growing particles switch to the wurtzite structure.
Apr 02, 2026	New preprint: we show that concerted, collective atomic motion drives the anomalous diffusion seen in anharmonic bcc metals such as titanium close to melting.
Nov 01, 2025	New paper in Advanced Theory and Simulations on excitation energy transfer between porphyrin dyes on a clay surface, studied with multifidelity machine learning.
Jan 15, 2025	Our paper on local structure classification is out in Computer Physics Communications. We use supervised learning to tell crystal phases and dislocations apart in complex oxides such as alumina and olivine.

selected publications

Competing nucleation pathways in nanocrystal formation

Carlos R Salazar, Akshay Krishna Ammothum Kandy, Jean Furstoss, and 3 more authors

npj Computational Materials, 2024
Bonding adaptability and transformation plasticity in Ti–Al–Si–N alloys

Carlos R Salazar, Sheuly Ghosh, Ferenc Tasnádi, and 4 more authors

Available at SSRN 7038127, 2026