Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft...
Transcript of Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft...
![Page 1: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/1.jpg)
Introduction to
Computer Simulations of Soft Matter
Methodologies and Applications Boulder July, 19-20, 2012
K. Kremer
Max Planck Institute for Polymer Research, Mainz
![Page 2: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/2.jpg)
Overview • Simulations, general considerations • Monte Carlo (MC)
• Basics • Application to Polymers (single chains, many chains systems • methods
• Molecular Dynamics(MD) • Basics • Ensembles • Application to Polymers • Example: Melt of linear and ring polymers • Simple membrane models
• (DPD and Lattice Boltzmann) • Multiscale Techniques
![Page 3: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/3.jpg)
Time and length scales
↔ ↔
↔ ↔
(Sub)atomic electronic structure chemical reactions excited states
Microscopic L ≅ 1Å - 3Å T ≅ 10-13 sec Energy dominates
Mesoscopic L ≅ 10Å - 50Å T ≅ 10-8 - 10-4 sec Entropy dominates
Mesoscopic L ≅ 10Å - 50Å T ≅ 10-8 - 10-4 sec Entropy dominates
Semi macroscopic L ≅ 100Å - 1000Å T ≅ 0 (1 sec)
Macroscopic domains etc.
generic/universal *** chemistry specific
Properties
![Page 4: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/4.jpg)
Time and length scales
↔ ↔
↔ ↔
(Sub)atomic electronic structure chemical reactions excited states
Microscopic L ≅ 1Å - 3Å T ≅ 10-13 sec Energy dominates
Mesoscopic L ≅ 10Å - 50Å T ≅ 10-8 - 10-4 sec Entropy dominates
Mesoscopic L ≅ 10Å - 50Å T ≅ 10-8 - 10-4 sec Entropy dominates
Semi macroscopic L ≅ 100Å - 1000Å T ≅ 0 (1 sec)
Macroscopic domains etc.
generic/universal *** chemistry specific
Properties Ansatz: integrate equations of motion of a classical atomistic model timestep ≈ 10-15 sec global relaxation time O(1s) still “small” system ≈ 106 atoms ⇒ At least 1021 integration time steps
In most cases neither useful, nor possible Use alternative options, employ universality, focus on question to be solved
![Page 5: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/5.jpg)
Time and length scales
↔ ↔
↔ ↔
(Sub)atomic electronic structure chemical reactions excited states
Microscopic L ≅ 1Å - 3Å T ≅ 10-13 sec Energy dominates
Mesoscopic L ≅ 10Å - 50Å T ≅ 10-8 - 10-4 sec Entropy dominates
Mesoscopic L ≅ 10Å - 50Å T ≅ 10-8 - 10-4 sec Entropy dominates
Semi macroscopic L ≅ 100Å - 1000Å T ≅ 0 (1 sec)
Macroscopic domains etc.
generic/universal *** chemistry specific
Properties General Advise: - Models should be as simple as possible, taken the question one asks into account - Use theory information as much as possible
- Avoid conserved extensive quantities, when possible
(causes transport issues, slow) - Try to beat natural slow dynamics for faster averaging
(cannot be used to study dynamics)
![Page 6: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/6.jpg)
Time
Atomistic
Local Chemical Properties ⇔ Scaling Behavior of Nanostructures Energy Dominance ⇔ Entropy Dominance of Properties
Finite Elements, Macrosc. Theory
Length
bilayer buckles Molecular
GROMACS www.gromacs.org
ESPResSo www.espresso.mpg.de
www.cpmd.org
Quantum
Soft Matter Theory: Comprehensive Understanding of Physical and Chemical Properties
Analytic Theory
Soft Fluid
![Page 7: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/7.jpg)
Simulations, general considerations
Pure MD Deterministic dynamics (Newton’s eq., Liouville Eq.) MD coupled to Noise (Fokker Planck Eq.) Brownian Dynamics (Smoluchowski Eq.) Force Biased MC Pure MC Stochastic dynamics
![Page 8: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/8.jpg)
PEP CαCβ
Alanine-rich regions in silk proteins
Simulations, general considerations
![Page 9: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/9.jpg)
Overview • Simulations, general considerations • Monte Carlo (MC)
• Basics • Application to Polymers (single chains, many chains systems • methods
• Molecular Dynamics(MD) • Basics • Ensembles • Application to Polymers • Example: Melt of linear and ring polymers • Simple membrane models
• (DPD and Lattice Boltzmann) • Multiscale Techniques
![Page 10: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/10.jpg)
![Page 11: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/11.jpg)
Simulations, general considerations MC models/moves for Rouse Dynamics
Generate “new” bonds inside chain: Mimics Rouse coupling to heat bath
![Page 12: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/12.jpg)
Simulations, general considerations MC models/moves for Rouse Dynamics
Diamond lattice data
![Page 13: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/13.jpg)
Simulations, general considerations MC models/moves for Rouse Dynamics
Bond fluctuation model
JCP 1991
![Page 14: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/14.jpg)
Simulations, general considerations MC models/moves for Rouse Dynamics
Bond fluctuation model
early state equilibrated
![Page 15: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/15.jpg)
Simulations, general considerations MC models/moves for Rouse Dynamics
Bond fluctuation model
early state equilibrated
- Lattice MC methods still used by many groups - Fast algorithms - Problems at high densities - No simple shear etc possible - Switch to alternative methods, continuum
![Page 16: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/16.jpg)
M. Müller, K. Daoulas et al: Coupling SCF calculations to particle based Monte Carlo
Hybrid methods: SCF + MC (Mueller, Daoulas, de Pablo, Schmid)
![Page 17: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/17.jpg)
Overview • Simulations, general considerations • Monte Carlo (MC)
• Basics • Application to Polymers (single chains, many chains systems • methods
• Molecular Dynamics(MD) • Basics • Ensembles • Application to Polymers • Example: Melt of linear and ring polymers • Simple membrane models
• (DPD and Lattice Boltzmann) • Multiscale Techniques
![Page 18: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/18.jpg)
Basic Idea: Integrate Newton’s equations of motion for a collection of N classical particles
Integrate equations of motion: Microcanonical, “NVE Ensemble”
Interaction potential e.g. LJ
𝑈𝐿𝐿 𝑟𝑖𝑖 = {( 𝜎𝑟𝑖𝑖
)12 – ( 𝜎𝑟𝑖𝑖
)6 }
![Page 19: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/19.jpg)
Basic Idea: Integrate Newton’s equations of motion for a collection of N classical particles
Integrate equations of motion: Velocity Verlet, symplectic
Interaction potential e.g. LJ
𝑈𝐿𝐿 𝑟𝑖𝑖 = {( 𝜎𝑟𝑖𝑖
)12 – ( 𝜎𝑟𝑖𝑖
)6 }
Variants of integration scheme, cf books by Frenkel and Smit, Allen and Tildesley
![Page 20: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/20.jpg)
Simulations, general considerations MD of single, isolated chain in space??
Bead-Bead interaction Plus FENE spring for bonds
![Page 21: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/21.jpg)
Simulations, general considerations MD of single, isolated chain in space??
Bead-Bead interaction Plus FENE spring for bonds
NVE Ensemble integration: Never equilibrates, Rouse modes do not couple strongly enough Need noise term needed!
Fermi-Pasta-Ulam Problem
![Page 22: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/22.jpg)
FPU Problem weakly anharmonic chain in d=1, with periodic boundary conditions
E/N =0.7, r=3, α=0.1
Short time
Long time No equilibration! Applies also to harmonic crystal
![Page 23: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/23.jpg)
FPU Problem weakly anharmonic chain in d=1, with periodic boundary conditions
E/N =0.7, r=3, α=0.1 E/N =1.2, r=3, α=0.1
non ergodic ergodic
For ergodicity need (strongly) mixing modes!
E/N =1.0, r=3, α=0.1, most probably border line non ergodic - ergodic
![Page 24: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/24.jpg)
Simulations, general considerations FPU Problem
FPU Hamiltonian for α=0 integrable, phase space is a N-dimesional manifold of the general 2N-dimensional phase space for α≠0 not integrable anymore, however there is NO analytical expression, of αmin, for which modes properly mix and make system ergodic
Fast equilibrating MD has strongly mixing modes, thus chaotic dynamics Intrinsically instable, not deterministic for longer times
![Page 25: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/25.jpg)
Simulations, general considerations FPU Problem
FPU Hamiltonian for α=0 integrable, phase space is a N-dimesional manifold of the general 2N-dimensional phase space for α≠0 not integrable anymore, however there is NO analytical expression, of αmin, for which modes properly mix and make system ergodic
Fast equilibrating MD has strongly mixing modes, thus chaotic dynamics Intrinsically instable, not deterministic for longer times Need stabilization, coupling to a thermostat -> NVT, canonical ensemble
![Page 26: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/26.jpg)
MD – Ensembles
Integrating Newton’s equations => microcanonical (NVE) stability issues for long runs… MD plus Thermostat => Canonical Ensemble, NVT two options: local vs global coupling possible consequences for dynamics MD plus Barostat => constant pressure Ensemble, NPE, usually with Thermostat, NPT
![Page 27: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/27.jpg)
Application: Polymer melts and networks - how to generate an equilibrated polymer melt? - role of topological constraints - ring polymers vs open chains
![Page 28: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/28.jpg)
Bead-spring model K.K & G.S. Grest
Dense monomeric liquid Flexible chains
bNLbclbc
LlNbcR
K
K
)1( ,97.07.1
)1(22
−====
=−≈
∞
∞
∞
σ
→
R
3-0.85Density
σρ =
![Page 29: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/29.jpg)
Equilibration of initial melt Auhl et al JCP, 2003
•Run a short chain melt to equilibrium by “brute force” and/or algorithm with global moves •This is the “reference” system for longer chain melts!
typically
NlNR K=)(2
Create Target Function
{k
k
nii
lnlnlnln
nR
rrnR
><
∝
−= +
,,
)(
)()(2
2
22
n
R /n 2
)( eNON =
![Page 30: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/30.jpg)
Equilibration of initial melt
•Create random walks with “correct” statistics by Monte Carlo procedure (e.g. NRRWs)
NlNR K=)(2
* Position walks randomly in space * Move walks around by random procedure (translation, rotation, inflection) to minimize density fluctuations * replace/exchange walks randomly to reduce density fluctuations * Slowly increase excluded volume * Control target functions permanently * Eventually complement by double-bridging moves
![Page 31: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/31.jpg)
Equilibration of initial melt
•Create random walks with “correct” statistics by Monte Carlo procedure (e.g. NRRWs)
NlNR K=)(2
* Position walks randomly in space * Move walks around by random procedure (translation, rotation, inflection) to minimize density fluctuations * replace/exchange walks randomly to reduce density fluctuations
![Page 32: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/32.jpg)
MD runs plus very slow insertions of the excluded volume
bad
good
* Slowly increase excluded volume * Control target functions permanently * Eventually complement by double-bridging moves
![Page 33: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/33.jpg)
{k
k
nii
lnlnlnln
nR
rrnR
><
∝
−= +
,,
)(
)()(2
2
22
n
R /n 2
Equilibration of initial melt
Stiff chains
Flexible chains
![Page 34: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/34.jpg)
Equilibration of initial melt: ABSOLUTELY CRUCIAL
![Page 35: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/35.jpg)
DPD: Dissipative Particle Dynamics
shear viscosity diffusion
![Page 36: Introduction to Computer Simulations of Soft MatterIntroduction to Computer Simulations of Soft Matter Methodologies and Applications Boulder July, 19-20, 2012 K. Kremer Max Planck](https://reader034.fdocuments.net/reader034/viewer/2022050105/5f43db50bd52ca760d566093/html5/thumbnails/36.jpg)
General Literature
Reviews - Adv. Polymer Science Vol. 173 (2005), 185 (2005), 221 (2009),
Springer Verlag, Berlin, New York, C. Holm, K. Kremer Eds.
- S. J. Marrink et al, Biochimica Biophysica Acta, 2008, general review on lipid models and membranes
- C. Peter, K. Kremer, Introductory Lecture for FD 144 Faraday Discuss., 144, 9 (2010)
Books: - Frenkel and Smit, “Understanding Molecular Simulations”, Academic Press, 2002 - Allen and Tildesley, “Computer Simulatiosn of Liquids”,
Clarendon Press, 1989 - G. Voth, ed., “Coarse-Graining of Condensed Phase and
Biomolecular Systems”, Taylor and Francis, 2009