BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:CMSP Seminar (Atomistic Simulation Seminar Series): Atomistic Simu
 lations under Anisotropic Stress: From CAIMD to DeepMD-Driven Machine Lear
 ning
DTSTART;VALUE=DATE-TIME:20260520T090000Z
DTEND;VALUE=DATE-TIME:20260520T100000Z
DTSTAMP;VALUE=DATE-TIME:20260524T033032Z
UID:indico-event-11354@ictp.it
DESCRIPTION:\n	Mengqi Cheng \n	(ICTP)\n\n	\n	Abstract:\n	In this talk\, I
  present our work on atomistic simulations of materials under extreme temp
 erature\, pressure\, and anisotropic stress. I first introduce a constrain
 ed ab initio molecular dynamics (CAIMD) framework that enables simulatio
 ns under non-uniform stress conditions\, including uniaxial compression an
 d shear. Building on this\, I incorporate an on-the-fly machine learning s
 cheme into CAIMD\, which significantly improves computational efficiency w
 hile maintaining first-principles accuracy and enables continuous generati
 on of high-quality training data. I also outline ongoing efforts toward a 
 DeepMD-based ML-CAIMD framework for larger-scale simulations.\n	Using this
  approach\, I investigate the high-temperature stability and melting behav
 ior of cubic boron arsenide\, revealing a reentrant melting behavior linke
 d to structural changes in the liquid phase. I further extend the method t
 o disordered systems and perform large-scale simulations of glassy carbon 
 under anisotropic stress\, identifying new transformation pathways toward 
 quenchable amorphous diamond and highlighting the key role of shear in sta
 bilizing sp\\textsuperscript{3} bonding.\n	Finally\, I discuss its applica
 tion to superionic water. Recent experiments have reported a mixed close-p
 acked structure in the superionic regime\, but it remains unclear whether 
 the observed stacking disorder originates from strong shock-driven uniaxia
 l\, high-strain-rate compression or represents an intrinsic feature of sup
 erionic ice. This open question motivates simulations under combined shear
  and compression\, which are directly accessible within the present framew
 ork and relevant to realistic planetary conditions.\n	Ref.\n	1. Mengqi Che
 ng and Hong Sun\, Physical Review Materials 8\, 113604 (2024)\, “Stabi
 lity of c-BAs under extreme conditions: An ab initio molecular dynamics st
 udy.”\n	2. Mengqi Cheng\, Weidong Luo\, and Hong Sun\, Physical Review 
 B 112\, 094207 (2025)\, “On-the-fly machine learning augmented constrai
 ned ab initio molecular dynamics to design routes from glassy carbon to qu
 enchable amorphous diamond.”\n	3. L. Andriambariarijaona et al.\, Natur
 e Communications 17\, 374 (2026) “Observation of a mixed close-packed st
 ructure in superionic water.”\n\n	Zoom registration link:https://zoom.us
 /meeting/register/SHbwboMjRqOBISPZF0dcEA\n	 \n\n//indico.ictp.it/event/11
 354/
LOCATION:Euler Lecture Hall (Leonardo Building\, terrace level) and via Zo
 om
URL://indico.ictp.it/event/11354/
END:VEVENT
END:VCALENDAR