This is available for download by clicking the following link. mdtep-4-4.tar.gz (releases 4.4 or later. For for compatibility with earlier releases use mdtep-old.tar.gz).

The included makefile should be easily modified for any system. MDTEP has been tested only on Intel Linux and Sun Solaris platforms, but no problems on other systems are anticipated. Note that GNUMake must be used with this makefile.

Once compiled, an MDTEP input file should be placed in the same directory as the .md and .castep file from which information is to be extracted. The program should then be invoked with:

mdtep myinputfile.input

The structure of the input file uses FORTRAN name-lists. An example is shown below.

`&RunName `

`SeedName = 'mymdrun' `

Seedname of CASTEP calculation

`/ `

`&Positions `

MD step at which to begin analysis

`Start = 1 `

MD step at which to begin analysis

`Final = 800 `

MD step at which to end analysis

`/ `

`&CalcFlags `

`CalcRDF = .TRUE. `

Calculate radial distribution function

`CalcVAC = .TRUE. `

Calculate velocity autocorrelation function

`CalcMSD = .TRUE. `

Calculate mean square displacement

`CalcHeatCap = .TRUE. `

Use fluctuations to calculate heat capacity

`CalcAlphaP = .TRUE. `

Use fluctuations to calculate expansion coefficient

`CalcBetaT = .TRUE. `

Use fluctuations to calculate bulk modulus

`CalcTempDist = .TRUE. `

Calculate distribution of temperature samples

`CalcVolDist = .TRUE. `

Calculate distribution of volume samples

`XmolFile = .TRUE. `

Generate <seedname>.xmol for visualisation

`XFSFile = .TRUE. `

Generate <seedname>.axsf for visualisation

`/`

`&VACParams`

`VACLength = 128 `

Length in MD steps over which VACF calculated

`VACInterval = 1 `

Steps between VACF functions to be averaged over

`/`

`&MSDParams`

`MSDLength = 128 `

Length in steps over which MSD is calculated

`MSDInterval = 1 `

Steps between MSDs to be averaged over

`/`

Notes:

- Quantities calculated from fluctuations require many thousands of MD steps to produce accurate quantities.
- Heat capacity calculations are only valid for simulations using a thermostat.
- Expansion coefficient and bulk modulus calculations should only be performed for variable cell MD runs.
- All calculations based on fluctuations invoke a blocking analysis to estimate the error in the calculated quantity - see Haile [4] for details.
- The distribution of temperature samples is written to file along with the ideal case for large N.
- The distribution of volume samples is written to file along with the ideal case calculated assuming the estimate of obtained is accurate.

Each calculated quantity is written to a separate output file, prefixed with the calculation seedname. The names of these files should be self explanatory.

As well as computing the quantities specified in the input file, MDTEP writes the following files, the first three of which are suitable for direct plotting in e.g. gnu-plot or xmgrace.

This file contains four columns of data, the first being the elapsed simulation time in picoseconds. The subsequent columns list the total ionic kinetic energy, the Hamiltonian energy for the appropriate ensemble, and the value of the Born-Oppenheimer Hamiltonian.

This file lists for each time-step (in picoseconds) the nine components of the matrix of cell vectors in atom units of length.

This file contains four columns, listing time in picoseconds, temperature in Kelvin, pressure in mega-pascals, and volume in atomic units.

This file lists the three components of the force of each atom. The inner loop is over the number of atoms, and the outer loop over the number of time-steps.

This is an intermediate format used when creating .xmol and .xsf files.