Tutorials

A series of representative example input and .pqr files exist in the tutorials_and_examples folder in the MPMC root directory. A brief description of each folder follows, it is encouraged to cross reference the input files with the Commands page and to run each example yourself.

BSS H₂ + MOF-5 Simulation (uVT)

This is a basic electrostatic/Lennard-Jones-based simulation of H₂ sorption in MOF-5 (using the BSS H₂ model and UFF parameters on MOF-5). The simulation may be started by simply invoking mpmc MOF5+BSS.inp. Every corrtime steps averages are output to stdout, including the average potential energy, number of molecules sorbed, Qst, timing information, etc.

Several different output files are created in the simulation directory by default:

• *.traj.pqr - this is the entire trajectory of the system at every corrtime

• *.restart.pqr - this is the most recent snapshot of the system

• *.final.pqr - output upon completion of the simulation, this is the last snapshot of the system

• histogram.dat - this is a histogram of sorbate positions in the .dx format

• frozen.dx - this is the frozen (MOF-5) atoms in the .dx format

BSSP H₂ + MOF-5 Simulation (uVT)

This is a very similar simulation to the first tutorial with the main difference arising from the use of the BSSP polarizable H₂ model. The polarizability options in this input file should be robust for most situations, it is advised to double check the polar_max_iter produces a converged polarization energy when simulating in a novel system however.

PHAHST H₂ + HKUST-1 Simulation (uVT)

This is a similar simulation to the second tutorial with the Lennard-Jones repulsion/dispersion potential replaced by physically ground exponential repulsion/damped dispersion potential:

\[\begin{split}U_{rd} &= \sum_{i \neq j} \frac{F_0}{\beta_{ij}}e^{\beta_{ij}(r_{ij}-\rho_{ij})}+\sum_{n=3}^5 f_{2n}(\beta r_{ij} ) \frac{C_{2n,ij}}{r_{ij}^{2n}} \\ f_{2n}( \beta r_{ij} ) &= 1 - e^{-\beta r_{ij}} \sum_{k=0}^{2n} \frac{(\beta r_{ij})^k}{k!} \\ \rho_{ij} &= \frac{1}{2}(\rho_{ii} + \rho_{jj}) \\ \beta_{ij} &= 2 \frac{\beta_{ii} \beta_{jj}}{\beta_{ii}+\beta_{jj}}\\ C_{2n,ij} &= \sqrt{C_{2n,ii} C_{2n,jj}}\end{split}\]

The use of the cavity_autoreject_repulsion and polar_precision options ensure that nuclear fusion doesn’t happen due to a polarization catastrophe and finite exponential repulsion potential energy. For more information on the PHAHST force field please see: https://pubs.acs.org/doi/abs/10.1021/acs.jctc.0c00837.

Simulated Annealing

This is a sample simulated annealing script for H₂ in the MOF NOTT-112. Simulated annealing is typically used to identify minimums in the potential energy surface, i.e. binding sites in material simulations. The input pqr contains the MOF with a single H₂ molecule about 2.5 angstrom from the CuC open metal site. The temperature starts at 40 K and is decreased by a factor of 0.99999 each MC step. At the end of the simulation the H₂ molecule will be in its lowest energy configuration around the open metal site. Small values for move_factor and rot_factor prevent the H₂ from straying away too far from its initial position.

Replaying a Trajectory

The “replay” ensemble takes in a *.traj.pqr file, either produced from a previous run or created artificially, and recalculates averages without running a simulation. This is useful in case some quantity needs to be recalculated without running a full simulation. The provided MOF5+BSS.traj-00000.pqr file was produced from the first tutorial.

Single Point Energy and Quantum Rotation Calculations

This folder contains the PQR files with the BSS model located at all four H₂ sorption sites in MOF-5 (alpha, beta, gamma, and delta). An input file to run the quantum rotation calculations is also included in the folder. Note that running quantum rotations calculations is predicated upon turning the quantum rotations part of the code on when compiling MPMC. By executing quantum rotation calculations on the four H₂ sorption sites (by varying pqr_input), you should obtain rotational levels that are very close to those shown in Table 1 of Ivana’s JCP 2012 paper (see link) for the respective sites. Note that the total_energy ensemble here calculates the potential energy (and rotational eigenspectrum) but does not perform any Monte Carlo steps.

Paper: http://scitation.aip.org/content/aip/journal/jcp/137/1/10.1063/1.4730906

Potential Energy Surface Fitting (surf_fit)

These are sample inputs for fitting an Argon potential energy surface in MPMC using the default fitting code. An additional file (or files) is needed to specify the ab initio surface to fit to. The files are as follows:

• *.input (MPMC input)

• *.pqr (initial parameters/coordinates)

• *.dat (PES input data)

Potential Energy Surface Fitting (surf_multi_fit)

These are sample inputs for fitting an argon potential using the surf_fit fitting code. A slightly different (and more accurate) ab initio surface is used. The files are as follows:

• *.input (MPMC input)

• *.pqr (initial parameters/)

• configs.out (PES input data)

Multisorbate uVT

These input files will run a multi-sorbate simulation in an rht-MOF (NOTT-112). The two sorbates are H2 (BSSP) and CO2 (PHAST). In multi-sorbate simulations a separate PQR file containing the molecules to be inserted/deleted needs to provided, insert.pqr here. Note this system is rather large and treated with explicit polarization so averages will take a relatively long time to converge.

BSSP H₂ + MOF5 Simulation (uVT) with NVIDIA CUDA

This is the second tutorial with the sole exception of the switch cuda on. MPMC needs to be compiled with cuda and a cuda capable GPU must be present. Performance may be compared with the second tutorial.

1D Chain Replay

Here we use ensemble replay to read in a series of increasingly larger 1D-chain samples (starting from 2-atoms and up to 512). Each component of the energy is re-calculated for each sample, for various potential options (the different input files). One may take the energy output files and process them (via scale.sh) to check for the small-size scalability and accuracy of our calculations.

3D Chain Replay

Here we use ‘ensemble replay’ to read in a series of increasingly larger crystal samples (starting from 2-atoms and up to 1024).

Each component of the energy is re-calculated for each sample, for various potential options (the different input files). One may take the energy output files and process them (via scale.sh) to check for the small-size scalability and accuracy of our calculations.