This keyword can be used to calculate the heat current autocorrelation (HAC) and running thermal conductivity (RTC) using the Green-Kubo method. The results will be written to the hac.out output file.


This keyword has 3 parameters:

compute_hac <sampling_interval> <correlation_steps> <output_interval>

The first parameter is the sampling interval for the heat current data. The second parameter is the maximum correlations steps. The third parameter for is the output interval of the HAC and RTC data.


Example 1

time_step 1
compute_hac 10 100000 1
run 10000000

This means that

  • You want to calculate the thermal conductivity using the Green-Kubo method (the EMD method) in this run, which contains 10 milillion steps with a time step of 1 fs.

  • The heat current data will be recorded every 10 steps. Therefore, there will be 1 million heat current data in each direction.

  • The maximum number of correlation steps is \(10^5\), which is one tenth of the number of heat current data. This is a very sound choice. The maximum correlation time will be \(10^5 \times 10=10^6\) time steps, i.e., 1 ns.

  • The HAC/RTC data will not be averaged before outputting, generating \(10^5\) rows of data in the output file.

Example 2

compute_hac 10 100000 10

This is similar to the above example but with one expection: The HAC/RTC data will be averaged for every 10 data before outputing, generating \(10^4\) rows of data in the output file.