eFF force field questions

Hi everyone,

We are recently trying to self-taught how to use eFF on LAMMPS through the LAMMPS document and example scripts, and we have some questions, and were wondering if any experienced eFF users or members of Prof Jaramillo-Botero’s group could help us with these:

  1. We found in the examples that, in the input files, all the atom information (for ions and electrons) are stored in a separate file, and then loaded by read_data command, rather than using the create_atoms command. Can we alternatively create lattices and create atoms on lattices, and finally use set command to set for electron spins and eradius?

  2. As we found in the LAMMPS documents, the atom section for electron atom styles require the following items:
    atom-ID atom-type q spin eradius x y z
    Comparing with the example input scripts, we found that q for electrons are zero. Why is it not -1? Does the eFF code recognise electron by this (e.g. if q == 0 then treat it as an electron)?

  3. We found in the example input scripts (e.g. in hydrogen simulations) that eradius is not zero for electrons. If we wish to perform the all-electron calculation (i.e. without using ECP), can we put an arbitrary value, or zeros there?

Many thanks in advance.

Best regards,

I am not a member of Andres’ group, but will try to answer your questions from information in the documentation and the source code.

No. All available/usable features in LAMMPS are documented. There are no references to the eFF per-atom properties in the set command — LAMMPS documentation, so it is not supported. With a little experience in C++ programming, it can probably be added, if desired.

q is the (full or effective) charge of the (classical) atom cores and electrons don’t have one. Electrons can also be recognized by their spin and their eradius value.

No, eradius is the radius of the gaussian wavepacket representing an electron. It should be initialized to a reasonable value close to the converged result.


Hi Axel,

Many thanks for your help; your answer is very useful on our confusions. For the initial eradius value, we were wondering if there is a method to obtain a reasonable initial guess? If you have no much clue about this, we can try to contact Prof Jaramillo-Botero asking about this. Do you know if he or his group members is active in matsci.org? If not, I’ll drop an email to ask and post the reply (with their approval).

Best regards,

I don’t know. The LAMMPS category in the MatSci Discourse has been established only rather recently. So you are probably well advised to e-mail Andres directly. Questions about eFF are very infrequent in the forum or mailing list.