Interatomic

Poten-als in LAMMPS Aidan P. Thompson Sandia Na1onal Laboratories

2nd LAMMPS Users Workshop August 2011, Albuquerque, New Mexico

Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy s National Nuclear Security Administration under contract DE-AC04-94AL85000."

Recently Added Poten-als

COMB poten-al (Genera-on 2)
Tzu-Ray Shan (U Florida), talk on Tues PM metal and semiconductors and their oxides Xiaowang Zhou (Sandia), ionic compounds combo of Li, Na, K, Rb, Cs, F, Cl, Br, and I Me1n Aktulga (LBNL), talk on Wed PM can be faster than Fortran version by 2-3x Andres Jaramillo (Caltech), explicit electron dynamics in extreme condi1ons

Embedded ion method poten-als (EIM) C++ version of ReaxFF

Electron force eld (eFF)

Recently Added Poten-als

AIREBO poten-al bug xes Mishin ADP poten-al Dreiding poten-al
Marcel Fallet & Steve Stuart (Clemson), one more bug- x, upgrade is imminent

Chris Weinberger (Sandia) & Chandra Veer Singh (Cornell) angular-dependent EAM for metals and alloys Tod Pascal (Caltech) hydrogen bonding for solvated bio- molecules Mike Parks & Stuart Silling (Sandia), talk on Wed PM fracture at the meso and con1nuum scales

New Peridynamics poten-als

Upcoming Poten-als
Core/shell poten-al Mike Chandross (Sandia), zero-order model for polariza1on, uranium and other nuclear fuel materials COMB poten-al (Genera-on 3) Tzu-Ray Shan (U Florida), more materials with polariza1on eects MGPT poten-al from John Moriarty & Jaime Marian (LLNL), tantalum and other transi1on metals BOP poten-al Xiaowang Zhou and Don Ward (Sandia) High-accuracy poten1al for semiconductors

LAMMPS Poten-al benchmarks

Compute Cost of Interatomic Poten-als Growing Exponen-ally

Compute cost of LAMMPS poten1als versus publica1on date
Cost [core-sec/atom-timestep]

10

-2

GAP (Estim.)

10 10

-3
ReaxFF eFF COMB

-4
MEAM CHARMm

ReaxFF/C AIREBO

Drivers Cycles are cheap Availability of quantum calcula1ons (N < 100) Qualita1ve accuracy no longer enough

10 10

-5

EIM Stillinger-Weber EAM Tersoff

-6

1980

1990 2000 Year Published

2010

h`p://lammps.sandia.gov/bench.html#poten1als

Challenges with Complex Poten-als

How to Implement in LAMMPS?
Rewrite code from scratch (REBO) Integrate exis1ng serial code (ReaxFF) Access via general API (KIM)

How to Validate LAMMPS Version? How to Handle New Versions?

How to Fit Poten-als to New Materials?

Automated eng procedures exist for certain classes of materials and poten1als (EAM) More commonly, good ts can be obtained only by gurus (Baskes, van Duin) Combina1on of nonlinear op1miza1on and physical intui1on Increasing interest in automated machine-learning approaches Splines Gene1c programs Mul1-objec1ve op1miza1on Neural networks Series expansions

GAP Approach for Interatomic Poten-als

GAP: A systema-c, informa-cs approach rcut j $ j j Based on QM and mathema1cs rather than r ! ( r) = # # c m!,mY m!,m (" , #, $ ) empiricism. j=0 m!,m=" j Local density around each atom expanded in 4D hyperspherical harmonics, analogous to Fourier series Atomic congura1ons described by bispectrum of lowest-order coecients in series Preserves universal physical symmetries: invariance w.r.t. rota1on, transla1on, permuta1on Gaussian process (GP) regression used to interpolate energy of QM congura1ons 100-1000x more expensive than MEAM Diamond: Force errors for GAP `ed to DFT. Far cheaper than QM, linear scaling Adding higher-order GAP coecients Can trade performance and accuracy

systema1cally increases accuracy Bartok et al., PRL 104 136403 (2010)

GAP Poten-al for Beryllium

Beryllium Working with GAP developers, eng directly to forces and energies from high-temperature DFT MD simula1ons of small systems (from Mike Desjarlais, 1640)
6000 5000
Temperature [K]
Dr09 a-axis Ba09i a-axis DFT DFT Melt Line EOS Benedict et al.

GAP Forces versus DFT Forces for Liquid Be at 5,000 K, 250 atoms
!" #$

#"

"

!$

!#"

!#$ !!"

!#$

!#"

!$

"

#"

#$

!"

4000 3000 2000 1000

Dr09 a-axis Ba09i a-axis DFT DFT Melt Line 100 EOS 50 Benedict et al.

Ini1al t to:

6000 5000
Temperature [K]

ambient HCP high-pressure BCC and liquid

0 0

4000 3000 2000

150 200 Pzz [GPa]

250

300

Accurately reproduced that data Problems with:

ambient elas1c constants high-pressure HCP

Need to ret with more data