3.22 Mechanical properties of materials Review: Molecular Dynamics xxx Markus J. Buehler Outline: 4 Lectures on Molecular Dynamics (=MD) Lecture 1: Basic Classical Molecular Dynamics General concepts, difference to MC methods, challenges, potential and implementation Lecture 2: Introduction to Interatomic Potentials Discuss empirical atomic interaction laws, often derived from quantum mechanics or experiment Lecture 3: Modeling of Metals Application of MD to describe deformation of metals, concepts: dislocations, fracture Lecture 4: Reactive Potentials New frontier in research: Modeling chemistry with molecular dynamics using reactive potentials © 2006 Markus J. Buehler, CEE/MIT Elasticity and atomistic bonding “atomistic” discrete “continuum” © 2006 Markus J. Buehler, CEE/MIT Molecular dynamics versus Monte Carlo MD is an alternative approach to MC by sampling phase and state space, but obtaining actual deterministic trajectories; thus: Full dynamical information In long time limit - for equilibrium properties - the results of MC correspond to results obtained by MD MD can model processes that are characterized by extreme driving forces and that are non-equilibrium processes Example: Fracture © 2006 Markus J. Buehler, CEE/MIT Motivation: Fracture Materials under high load are known to fracture MD modeling provides an excellent physical description of the fracture processes, as it can naturally describe the atomic bond breaking processes Other modeling approaches, such as the finite element method, are based on empirical relations between load and crack formation and/or propagation; MD does not require such input What “is” fracture? © 2006 Markus J. Buehler, CEE/MIT Ductile versus brittle materials BRITTLE DUCTILE Glass Polymers Ice... Copper, Gold Shear load Figure by MIT OCW. (a) (b) © 2006 Markus J. Buehler, CEE/MIT Molecular dynamics Total energy of system E = K +U 1 N 2 K = m∑ v j 2 j =1 U = U (rj ) m d 2 rj dt 2 = −∇ r j U (rj ) j = 1..N Coupled system N-body problem, no exact solution for N>2 System of coupled 2nd order nonlinear differential equations Solve by discretizing in time (spatial discretization given by “atom size”) © 2006 Markus J. Buehler, CEE/MIT Solving the equations + 1 2 ri (t0 + Δt ) = ri (t0 ) + vi (t0 )Δt + ai (t0 )(Δt ) + ... 2 1 2 ( ) ri (t0 − Δt ) = ri (t0 ) − vi (t0 )Δt + ai (t0 ) Δt + ... 2 ri (t0 + Δt ) = −ri (t0 − Δt ) + 2ri (t0 )Δt + ai (t0 )(Δt ) + ... 2 Positions at t0-Δt Positions at t0 Accelerations at t0 “Verlet central difference method” How to obtain accelerations? f i = mai ai = f i / m Need forces on atoms! © 2006 Markus J. Buehler, CEE/MIT Typical modeling procedure Set particle positions Calculate force on each particle Move particles by timestep Δt Stop simulation Assign particle velocities Save current positions and velocities Reached max. number of timesteps? Analyze data print results © 2006 Markus J. Buehler, CEE/MIT Modeling vs. simulation Modeling: Building a mathematical or theoretical description of a physical situation; maybe result in a set of partial differential equations For MD: Choice of potential, choice of crystal structure, write down F=ma… Simulation: Numerical solution of the problem at hand (code, infrastructure..) Solve the equations – e.g. Verlet method, parallelization (later) Simulation is usually followed by analysis methods – post­ processing (RDF, temperature…) © 2006 Markus J. Buehler, CEE/MIT 5 5 4 4 Solid Argon 3 g(r) g(r) Radial distribution function: Solid versus liquid versus gas 2 0 0 0.2 Liquid Ar (90 K) 3 Gaseous Ar (300 K) 2 Liquid Argon 1 Gaseous Ar (90 K) 1 0.4 distance/nm 0.6 0.8 0 0 0.2 0.4 distance/nm 0.6 0.8 Figure by MIT OCW. Note: The first peak corresponds to the nearest neighbor shell, the second peak to the second nearest neighbor shell, etc. In FCC: 12, 6, 24, and 12 in first four shells © 2006 Markus J. Buehler, CEE/MIT Mean square displacement (MSD) function Liquid 1 < Δr >= N 2 Crystal ∑ (r (t ) − r (t = 0)) 2 i i i Position of atom i at time t Relation to diffusion constant: d lim < Δr 2 >= 2dD t→∞ dt d=2 2D d=3 3D Position of atom i at time t=0 d lim < Δr 2 > t→∞ dt =D 2d J. Buehler, CEE/MIT © 2006 Markus Atomic scale Atoms are composed of electrons, protons, and neutrons. Electron and protons are negative and positive charges of the same magnitude, 1.6 × 10-19 Coulombs Chemical bonds between atoms by interactions of the electrons of different atoms (see QM part later in IM/S!) “Point” representation e- ep o p+ no +n p+ p o no no n+ no + p p V(t) + e- e- r(t) a(t) y x eFigure by MIT OCW. eFigure by MIT OCW. © 2006 Markus J. Buehler, CEE/MIT Pair interaction approximation U total = 1 5 2 1 2 ∑ ∑ | U (rij ) i=1.. N j=1.. N i≠ j Any function that expresses U (rij ) energy for atomic distance.. 3 All pair interactions of atom 1 with neighboring atoms 2..5 4 1 5 2 All pair interactions of atom 2 with neighboring atoms 1, 3..5 3 4 Double count bond 1-2 therefore factor © 2006 Markus J. Buehler, CEE/MIT Lennard-Jones potential Attractive Units: Energy ⎛ ⎡σ ⎤12 ⎡σ ⎤ 6 ⎞ φweak (r) = 4ε ⎜ .⎢ ⎥ − ⎢ ⎥ ⎟ ⎜ ⎣r⎦ ⎟ r ⎣ ⎦ ⎝ ⎠ Repulsive Units: Energy/length=force dV (r) F =− dr xi Fi = F r r x2 F x1 © 2006 Markus J. Buehler, CEE/MIT MD updating scheme: Complete (1) Updating method (integration scheme) 2 ri (t0 + Δt ) = −ri (t0 − Δt ) + 2ri (t0 )Δt + ai (t0 )(Δt ) + ... Positions at t0-Δt Positions at t0 (2) Obtain accelerations from forces f i = mai ai = Fi / m Accelerations at t0 “Verlet central difference method” (5) Crystal (initial conditions) Positions at t0 (3) Obtain forces from potential dV (r ) F =− dr xi Fi = F r (4) Potential ⎛ ⎡σ ⎤12 ⎡σ ⎤ 6 ⎞ φweak (r ) = 4ε ⎜ .⎢ ⎥ − ⎢ ⎥ ⎟ ⎜ ⎣r⎦ ⎟ r ⎣ ⎦ ⎝ ⎠ © 2006 Markus J. Buehler, CEE/MIT Deformation of crystals Deformation of a crystal is similar to pushing a sticky tape across a surface: F~ τ ⋅ L “homogeneous shear” F ≈ Fripple “localized slip (ripple)” Lcrit ≈ Fripple τ Beyond critical length L it is easer to have a localized © 2006 Markus ripple… J. Buehler, CEE/MIT Ductile versus brittle materials BRITTLE Glass Polymers Ice... DUCTILE Copper, Gold Shear load Figure by MIT OCW. © 2006 Markus J. Buehler, CEE/MIT