This discussion on the cluster expnasion of metallic sysytems using ATAT along with quantum espresso. I am new to ATAT as well as DFT. I am trying to do cluster expansion of a metallic system, say Mo-V sysytem. I am following the instructions given the manual to start the expansion and there is a script code to generate the input files for energy calculations. The code is attached to this.
I am trying to do cluster expansion for BCC system with lattice parameter 3.51. The script is generating the input file with both ibrav & celldm along with CELL_PARAMETERS. With ibran not equal to 0, this is reulting error, "redundant data for cell parameters". With a little investigation I found that error is due to the info given in CELL_PARAMETERS. By removing this info, The calculations could continue.
Now, my question is,
Dear Rajendra Prasad Gorrey,
Quantum Espresso has different ways of specifying the unit cell parameters and the atomic positions for various reasons. In principle, all of them are equivalent. Therefore, using two of them at the same time introduces ambiguities and the code stops. (QE does not check if the structures are the same, it simply stops because of the redundant pieces of information.) As you said correctly, by removing the ambiguity QE runs.
Setting the ibrav to the appropriate value and then using celldm (remember they are in Bohr) usually provides QE with a better symmetry analysis that might reduce the computational cost. If you set ibrav=0 and provide the CELL_PARAMETERS for the BCC structure, QE will produce the same results as in setting ibrav and celldm, within the numerical error. Indeed, the developers of QE are discouraging the use of ibrav=0.
Hope this helps,
Roberto
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory