Water-order : A fast C++ code to calculate order parameters for water droplets

(Currently an alpha version, the code is being tested)

This software calculates the radial distribution of several order parameters, including (1) oriental tetrahedral order parameter (q_T) (2) fifth nearest neighbour (d₅) parameter, (3) Translational tetrahedral order parameter (S_k) and (4) Voronoi cell based density (ρ_V) for water droplets. It is mainly aimed at analysis of molecular dynamics simulations of water droplets with various molecules/ions in it. The code assumes that there is no periodic boundary condition.

It calculates the requested order parameter for each oxygen, then takes the average of all oxygens in a certain distance range (--bin-width) from the centre-of-mass of the droplet, to get the radial distribution. For a trajectory, the distributions are also averaged over the whole trajectory.

Depends on the chemfiles library: https://github.com/chemfiles/chemfiles/, TCLAP library http://tclap.sourceforge.net/ and Voro++ library https://github.com/chr1shr/voro. All of source files for the libraries that were used to compile the latest version are already included in the repository. This has been done to avoid incompatibility between different versions of those libraries, and to avoid the need to hunt down dependencies.

Compiles successfully on Windows with MSVC++ 2019, on Linux with GCC 10, and it should also compile on Mac OS (Instructions for compiling are provided at the bottom of the page).

Is this software suitable for you?

Water-order is mainly aimed at calculations on water droplets. For droplets, the main feature of interest is the radial distribution of the order parameters because it reveals the radial structure of the droplets (as they are mostly spherical). The radius is always calculated from the centre of mass of the droplet.

This is why water-order calculates the radial distribution of various parameters, by averaging over a certain range. It is also assumed that there is no periodic boundary condition (which is true for droplets). Water-order is not for periodic systems, neither does it output data for individual water molecules, although it is possible to make the program output them by modifying the "Water_order.cpp" source file.

Detailed description

Tetrahedral order parameter (q_T) is a parameter that is used to characterise the degree of order, particularly for water. The q_T for an oxygen can be calculated by considering the 4 nearest oxygens of that oxygen, using this expression:

Where ψ_jk indicates the angle formed by the oxygens j and k and the central oxygen. For ice, q_T is close to 1 (0.8-0.9), whereas for liquid water q_T is lower, around 0.5. In the gaseous state, q_T should be close to 0. At intermediate temperatures, there are intermediate values

The fifth nearest neighbour parameter (d₅) is simply the distance to the 5th nearest oxygen from the central oxygen. The value is lower for ice and is higher for liquid water. Both of these parameters are routinely used in calculations involving water, supercooled water and especially on supercooled water droplets.

The translational tetrahedral order parameter (S_k) is similar to the oriental tetrahedral order parameter in that it considers the four nearest neighbours. However, it considers the distances instead of angles.

Where r_k is the distance between central oxygen and the k-th nearest neighbour and is the arithmetic mean of the distances to the 4 nearest neighbours.

The Voronoi cell based density (ρ_V) constructs Voronoi cells for each heavy atom in the system (i.e. excluding hydrogen). The volume of the Voronoi cell associated with each water molecule (because it has one oxygen - which is a heavy atom), is calculated. The density of that molcule is then, the mass of one water molecule divided by the volume of the cell containing it.

This program is calculates the radial distribution of order parameters for droplets/nanodroplets of water. The calculation procedure is as follows:

1. First the centre of mass is evaluated for one frame
2. The order parameter is calculated for each water molecule (or rather, each oxygen) in the frame
3. The program considers spherical shells from the centre of mass (r=0) to the value entered through --rmax, by increasing the radius as --bin-width. (In other words, it considers radius ranges of size --bin-width upto --rmax.) Then the average of q_T or d₅ values of all oxygens in that range is taken. This gives the radial distribution.
4. Steps 1-3 are iterated for each frame in the trajectory.
5. The radial distributions from all of the frames are summed and then divided by the total number of frames, to get the average radial distribution for the trajectory

The droplet of water can contain ions or other molecules, it only affects the centre of mass, it doesn't affect the rest of the calculation in any way.

How to run

The executable can be run with command line arguments (also without):

USAGE:

   Water_order  [-f <string>] [-s <string>] [-c <string>] [-t <OTO|d5|Sk
                |rhoV>] [-o <string>] [--rmax <float (Angstrom)>]
                [--bin-width <float (Angstrom)>] [--start <positive
                integer>] [--stop <positive integer>] [--] [--version]
                [-h]


Where:

   -f <string>,  --trajectory <string>
     Name of the trajectory file

   -s <string>,  --atom-info <string>
     Name of the file containing the atom information (topology)

   -c <string>,  --oxygen-name <string>
     Name of oxgyen atoms in the atom-info file

   -t <OTO|d5|Sk|rhoV>,  --task <OTO|d5|Sk|rhoV>
     Task requested to run: OTO = Oriental Tetrahedral Order parameter, d5
     = d5 parameter, Sk = Translational Tetrahedral Order parameter, rhoV =
     Voronoi cell based water density; default is OTO

   -o <string>,  --output-file <string>
     Base name of output file; default is 'Water_order'

   --rmax <float (Angstrom)>
     Maximum distance from centre-of-mass considered for distribution (rmin
     = 0 always)

   --bin-width <float (Angstrom)>
     Size of each range considered for averaging

   --start <positive integer>
     Frame number to start calculation from; default is first frame
     (counting starts from 0)

   --stop <positive integer>
     Frame number to end calculation at; default is last frame (counting
     starts from 0)

   --,  --ignore_rest
     Ignores the rest of the labeled arguments following this flag.

   --version
     Displays version information and exits.

   -h,  --help
     Displays usage information and exits.


   Water order analysis

This message is printed by running Water_order -h. For the -f argument, any trajectory file supported by chemfiles can be used (e.g. DCD, XTC, TRR, etc.). For the -s argument (i.e. atom information file), you should enter a PSF file (CHARMM/NAMD/XPLOR), or any valid file format containing mass and name information that can be read by chemfiles. A list of formats read by chemfiles can be found here. (Right now it does not mention PSF because that functionality has been implemented very recently.)

If the program is run without command line arguments, or some required command line arguments are missing, it will prompt for keyboard input automatically.

Example usage:

Water_order -f trajectory.dcd -s topology.psf -c OH2 -t d5 --rmax 20 --bin-width 0.5

This calculates the radial distribution of d5 parameter of trajectory.dcd in the distance of 20 Angstroms from the centre of mass of droplet. The d5 values are averaged in ranges of 0.5 Angstroms. The topology file can be a psf file, it can also be any file containing mass information that can be read by chemfiles.

The output file

The program outputs the radial distribution as a CSV file, containing two columns. The first column is r (i.e. radius) and the second column is the qT or d5 or Sk or rhoV, depending on which calculation is requested (i.e. 2nd column contains values of the parameter). The Voronoi cell based densities(ρ_V) have the units of g/cm³; the other parameters are all unitless. The data can be easily plotted with python, or xmgrace, or whatever visualisation software you prefer.

The base name of the output file is determined by the -o argument in the command line. When running OTO calculation, _oto.csv is appended to the base name. Similarly for the other calculations, _d5.csv, _Sk.csv, _rhoV.csv is appended to the base name. The default base name is Water_order, so the default output files are named Water_order_oto.csv or Water_order_d5.csv etc.

Example

First run the tests given in the tests folder.

An example of plotting the output data (from the test) with Python (using pandas and matplotlib):

import pandas as pd
import matplotlib.pyplot as plt

# read the CSV file
data = pd.read_csv("Water_order_oto.csv")
# Then plot the x and y axes
plt.plot(data["r"],data["qT"])
# Label the axes
plt.xlabel("r")
plt.ylabel("$q_\mathrm{T}$(r)")
# Show the plot
plt.show()

Current issues

1. If you attempt to run Water_order on a large DCD file (from NAMD, CHARMM or LAMMPS) it will run out of memory. This is due to a problem in the VMD molfile plugin, which is used by chemfiles for reading DCD files. VMD molfile reads the whole trajectory into memory. (This is mentioned in chemfiles issues here and here.) A workaround is to convert the DCD file to gromacs TRR or XTC format, which do not have this problem. The conversion can be done by catdcd of VMD or by mdconvert tool of MDTraj, and many other tools.
2. For histogramming, the sum is accumulated in a double type variable. If the trajectory is very large, this can cause overflow. The program checks for the overflow, and will automatically stop. If this happens, please open an issue, and I will try to implement a workaround.

How to compile

What is required:

1. C/C++ compiler that supports C++11 standard or higher.
2. CMake

Simple compilation:

1. First open a terminal or command prompt where cmake and the C/C++ compilers are in PATH. On Windows, with Visual Studio, this can be done by opening "x64 Native Tools Command Prompt for VS 2019" or something similar to that from the start menu. On Linux, the compilers are usually already in the PATH; if they are not then some script that sets the environment variables has to be run, it depends on which compiler you are using.
2. Set the environment variable CC and CXX if you are not using the default compiler for CMake (For Linux default is GNU compiler, for Windows it is Visual C/C++).
3. Run the install script. For Linux, Mac OS X or other POSIX systems, use the install.sh script. For Windows, run the install.bat batch file. The scripts should pick up the compiler from the CXX environment variable, if it is set. Otherwise it uses the default compiler.
4. After compiling, there should be an executable named "Water_order" in the folder.

For more details about installation, please check the INSTALL file.

Note that the program can also be compiled without OpenMP, it will give the same results. Compiling with OpenMP means that the frames are processed parallely in each thread, which improves performance.

---

Please acknowledge if you use this software.

Feel free to open an issue if there are any problems.

License and Contributing

The software is released under MIT license. Water-order also contains code from tclap-v1.2.4, released under MIT license and chemfiles-v0.11.0, released under BSD-3 clause license. Please read the the LICENSE file for the full terms and conditions of the license.

Any contribution is greatly appreciated. Please open a pull request/issue if you wish to contribute to the code.