Registered User Area
Username

Password


Crystal structure prediction using the USPEX code

October 22, 2012 to October 26, 2012

Location : CECAM-HQ-EPFL, Lausanne, Switzerland
   EPFL on iPhone

Details
Participants
Program
Lectures
Report
 

Organisers

  • Andriy Lyakhov (SUNY Stony Brook, USA)
  • Gilles Frapper (University of Poitiers, France)
  • Artem Oganov (SUNY Stony Brook, USA)
  • Mario Valle (Swiss National Supercomputing Centre, Switzerland)

Supports

   CECAM

International Union of Crystallography

Swiss Crystallographic Society

Description

 

Crystal structure prediction has long remained a major unsolved problem in physical sciences [1]. A number of approaches have been formulated over years - most of these are summarized in the recent book [2]. Crystal structure prediction is a powerful tool for designing new materials "in silico", thus replacing the traditional Edisonian trial-and-error approach with design by artificial intelligence. It is also a major instrument for discovering new phenomena at extreme conditions. Crystal structure prediction should thus be an everyday tool at the hands of nearly every computational materials scientist. 
A major advance in this field happened with the development of the evolutionary algorithm USPEX for crystal structure prediction [3], which has led to a number of important findings [4-6]. This proved to be a very efficient and reliable method, and the USPEX code, based on it and freely distributed to academic scientists, is currently used by over 500 researchers worldwide and this number grows rapidly. We should note that in addition to the evolutionary structure prediction, USPEX code features many other techniques, such as random sampling, metadynamics, minima hopping, particle swarm optimization - all of which can be used in real applications or tested against one another. Thus, familiarity with this code will imply a solid background in other structure prediction techniques. USPEX is the widest used crystal structure prediction code and there is a need to train new users through hands-on tutorials and workshops. One such workshop was organized in Poitiers, France (June 2011) and proved to be a great success. Another workshop will take place in Xi'an, China (August 2011). Through a regular series of such pedagogical events we want to train a new generation of users and developers of crystal structure prediction techniques. 

Crystal structure prediction has long remained a major unsolved problem in physical sciences [1]. A number of approaches have been formulated over years - most of these are summarized in the recent book [2]. Crystal structure prediction is a powerful tool for designing new materials "in silico", thus replacing the traditional Edisonian trial-and-error approach with design by artificial intelligence. It is also a major instrument for discovering new phenomena at extreme conditions. Crystal structure prediction should thus be an everyday tool at the hands of nearly every computational materials scientist. 
A major advance in this field happened with the development of the evolutionary algorithm USPEX for crystal structure prediction [3], which has led to a number of important findings [4-6]. This proved to be a very efficient and reliable method, and the USPEX code, based on it and freely distributed to academic scientists, is currently used by over 500 researchers worldwide and this number grows rapidly. We should note that in addition to the evolutionary structure prediction, USPEX code features many other techniques, such as random sampling, metadynamics, minima hopping, particle swarm optimization - all of which can be used in real applications or tested against one another. Thus, familiarity with this code will imply a solid background in other structure prediction techniques. USPEX is the widest used crystal structure prediction code and there is a need to train new users through hands-on tutorials and workshops. One such workshop was organized in Poitiers, France (June 2011) and proved to be a great success. Another workshop will take place in Xi'an, China (August 2011). Through a regular series of such pedagogical events we want to train a new generation of users and developers of crystal structure prediction techniques.

Our program will consist of theory lectures, hands-on tutorial sessions, and round table discussions. Morning theory lectures will focus on the nature of the crystal structure prediction problem, various ways to address it (in particular evolutionary algorithms), theory of energy landscapes, applications of crystal structure prediction, problems related to low-dimensional systems. Afternoon tutorials will give each participant to tackle a real research project, and special attention will be paid to the tools for data analysis and visualization.

References

[1] Maddox J. (1988). Crystals from first principles. Nature 335, 201.
[2] Oganov A.R. (Ed.) (2010). Modern methods of crystal structure prediction. Berlin: Wiley-VCH, ISBN 978-3-527-40939-6.
[3] Oganov A.R., Glass C.W. (2006). Crystal structure prediction using ab initio evolutionary techniques: principles and applications. J. Chem. Phys. 124, 244704.
[4] Ma Y., Eremets M.I., Oganov A.R., Xie Y., Trojan I., Medvedev S., Lyakhov A.O., Valle M., Prakapenka V. (2009). Transparent dense sodium. Nature 458, 182-185.
[5] Oganov A.R., Chen J., Gatti C., Ma Y.-M., Yu T., Liu Z., Glass C.W., Ma Y.-Z., Kurakevych O.O., Solozhenko V.L. (2009). Ionic high-pressure form of elemental boron. Nature 457, 863-867.
[6] Oganov A.R., Lyakhov A.O., Valle M. (2011). How evolutionary crystal structure prediction works - and why. Acc. Chem. Res. 44, 227-237.


CECAM - Centre Européen de Calcul Atomique et Moléculaire
Ecole Polytechnique Fédérale de Lausanne, Batochime (BCH), 1015 Lausanne, Switzerland