phono3py-get-kappa.md

phono3py-get-kappa

phono3py-get-kappa is a simple command-line script to extract the thermal-conductivity tensors (κ_latt) from the kappa-m*.hdf5 files produced by Phono3py.

The program writes the six unique elements of κ_latt at each calculation temperature to a comma-separated values (CSV)-format text file. It also calculates and outputs the isotropic average κ_iso = (κ_xx + κ_yy + κ_zz) / 3.

The CSV format is very simple and easily read by most analysis and visualisation software. (For an example of how to read and plot the data in Python, see below.)

Usage examples

To extract data from a Phono3py HDF5 file, run the script with the file path - e.g.:

$ phono3py-get-kappa kappa-m161616.hdf5

By default, phono3py-get-kappa writes output files to the current working directory named <root>.csv, so the command above will output the data to a file named kappa-m161616.csv. The CSV format is one row per line with headers/values separated by commas:

T [K],k_xx [W/m.K],k_yy [W/m.K],k_zz [W/m.K],k_yz [W/m.K],k_xz [W/m.K],k_xy [W/m.K],k_iso [W/m.K]
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
10.0,55.07339604014415,15.693212221821625,82.24785666053268,1.6022468809599445e-17,-7.671283605723772e-15,-5.898830900512673e-34,51.00482164083282
20.0,21.4688557567765,6.16560852088868,33.95198398725003,2.4801540741230823e-18,-3.1109728142044685e-15,-3.7708283987080234e-35,20.52881608830507
30.0,12.829446029165446,3.8886640431286335,19.874474063726712,-1.0245685468343207e-18,-1.8433243372704187e-15,1.5010928438204353e-35,12.197528045340263
40.0,9.33311620472965,2.8759553824371076,13.998472815827432,-1.4416382736688985e-18,-1.319249003270521e-15,2.343224246251316e-35,8.735848134331397
50.0,7.497596082528465,2.3535572873071198,10.911024851520345,-1.3378251274424403e-18,-1.0434576243120703e-15,2.428404832022923e-35,6.9207260737853105
60.0,6.350972715989032,2.0379983032786004,9.021988516551945,-1.1694887651441092e-18,-8.729406642763391e-16,2.3500289482360983e-35,5.803653178606527
70.0,5.544742745783705,1.8177554501382969,7.73450748243381,-1.0181459436311814e-18,-7.54995533492234e-16,2.227519947979205e-35,5.0323352261186045
80.0,4.933802590537467,1.6479669635208298,6.789558283080802,-8.940712462477584e-19,-6.671238157723123e-16,2.0944009912124933e-35,4.4571092790463664
90.0,4.448623617065409,1.5092008348755863,6.059985228460045,-7.938734251588838e-19,-5.983770116811075e-16,1.9633114805814556e-35,4.00593656013368
100.0,4.051303739267106,1.3919940640409125,5.476318212926482,-7.124858258576519e-19,-5.42770676491479e-16,1.839562132785157e-35,3.6398720054114997

The default output file name can be overridden using the -o/--output-file command-line option:

$ phono3py-get-kappa kappa-m161616.hdf5 -o kappa.csv

phono3py-get-kappa can accept multiple input files. Using the automatic naming, this makes it easy to process a directory with multiple files, as might for example be generated from a mesh convergence test:

$ phono3py-get-kappa kappa-m*.hdf5

(Note that the -o/--output-file option does not work when more than one input file is specified.)

Working with CSV files in Python

The following function can be used to read the CSV files produced by phono3py-get-kappa into a Python script:

import csv

import numpy as np

...

def ReadKappaCSV(file_path):
    # 8 columns: T, k_xx, k_yy, k_zz, k_xz, k_yz, k_xy, k_ave.
    
    data_cols = [
        [] for _ in range(0, 8)
        ]
    
    # Read input file.
    
    with open(file_path, 'r') as input_reader:
        input_reader_csv = csv.reader(input_reader)
        
        # Skip header row.
        
        next(input_reader_csv)
        
        for row in input_reader_csv:
            for i, val in enumerate(float(item) for item in row):
                data_cols[i].append(val)
    
    # Convert lists to NumPy arrays and return.
    
    return [
        np.array(data_col, dtype = np.float64)
            for data_col in data_cols
        ]

The following example plotting script reads the temperature, the diagonal κ_xx, κ_yy and κ_zz components of κ_latt and the average κ_iso from a CSV file and plots them using Matplotlib:

import matplotlib.pyplot as plt

...

t, k_xx, k_yy, k_zz, _, _, _, k_ave = ReadKappaCSV(r"kappa-m161616.csv")

plt.figure(
    figsize = (8.6 / 2.54, 7.0 / 2.54)
    )

plot_data = [
    (k_xx , r"$\kappa_{xx}$" , 'b', '^'),
    (k_yy , r"$\kappa_{yy}$" , 'r', 'o'),
    (k_zz , r"$\kappa_{zz}$" , 'g', 's')
    ]

for y_data, label, colour, marker in plot_data:
    plt.plot(
        t, y_data, label = label, color = colour,
        marker = marker, markersize = 4.0, markerfacecolor = 'none', markeredgecolor = colour
        )

plt.plot(
    t, k_ave, label = r"$\bar{\kappa}$", color = 'k'
    )

plt.yscale('log')

plt.xlim(0.0, 1000.0)
plt.ylim(1.0e-1, 1.0e2)

plt.legend(
    loc = 'upper right', frameon = False
    )

plt.xlabel(r"$T$ [K]")
plt.ylabel(r"$\kappa$ [W m$^{-1}$ K$^{-1}$]")

plt.tight_layout()

plt.savefig(
    r"Kappa.png", dpi = 300
    )