Dipole data with i-PI to analyze IR spectra

[M. X. L (M. X.)]

Dear MBX users and developers,

I want analyze IR spectra of centroid trajectories by TRPMD simulations. I use MBX interfaced with i-PI. But the output of i-PI software is only xyz trajectories. Can I output the dipole moment for the trajectory? I saw similar questions on the group last year. Would we have a new release to fix it?

Please forgive my presumption and urgency, this is very important to me.

Thank you very much for your time. 

Regards,
L. M. X.

[Richa Rashmi]

Dear L.M.X,

Currently, there isn't a straightforward method to print dipole moments using MBX+i-PI. However, you can utilize MBX+LAMMPS to obtain the dipole moments for your trajectory. You can extract the centroid positions from your i-PI TRPMD simulation (likely printed in simulation.xc.xyz), convert the .xyz file into a LAMMPS trajectory file format, and then do a LAMMPS rerun on those coordinates to print the dipole moments.

We hope this is helpful. Please feel free to reach out if you require assistance with the MBX+LAMMPS rerun to print dipole moments.

Thank you.
Best regards, 

MBX team

[M. X. L (M. X.)]

Thank you! It's very hepful for me!!!!

[M. X. L]

I am very grateful for all your selfless help.

I am trying to use the MBX1.1.0+LAMMPS to calculate the dipole moment of simulation.xc.xyz by reruning. I need to calculate the dipole moment of each atom to analyze the IR.

I have no experience with LAMMPS. I checked the input in /MBX-1.1.0/examples/lammps/256h2o/ and the previous posts (Dipoles and polarization (google.com) and fix mbx print/dipoles in lammps (google.com) ). 

I have converted the trajectory file (simulation.xc.xyz) to a dump file and the initial structure to a data file.

I have the following error:

Finding 1-2 1-3 1-4 neighbors ...
  special bond factors lj:    0        0        0      
  special bond factors coul:  0        0        0      
     2 = max # of 1-2 neighbors
     1 = max # of 1-3 neighbors
     1 = max # of 1-4 neighbors
     2 = max # of special neighbors
  special bonds CPU = 0.014 seconds
  read_data CPU = 0.050 seconds
[g02r1n08:808  :0:808] Caught signal 11 (Segmentation fault: address not mapped to object at address 0x194264f0)
==== backtrace (tid:    808) ====
 0 0x000000000004ed85 ucs_debug_print_backtrace()  ???:0
 1 0x0000000000942cc4 LAMMPS_NS::FixMBX::mbx_fill_system_information_from_atom()  /work/home/acqz25xa41/software/lammps-stable/src/Obj_mpi_mbx/../fix_mbx.cpp:507
 2 0x00000000009414ab LAMMPS_NS::FixMBX::FixMBX()  /work/home/acqz25xa41/software/lammps-stable/src/Obj_mpi_mbx/../fix_mbx.cpp:202
 3 0x000000000080512f _INTERNAL688e006b::style_creator<LAMMPS_NS::Fix, LAMMPS_NS::FixMBX>()  /work/home/acqz25xa41/software/lammps-stable/src/Obj_mpi_mbx/../modify.cpp:44
 4 0x000000000081646e LAMMPS_NS::Modify::add_fix()  /work/home/acqz25xa41/software/lammps-stable/src/Obj_mpi_mbx/../modify.cpp:914
 5 0x0000000000416494 LAMMPS_NS::Input::fix()  /work/home/acqz25xa41/software/lammps-stable/src/Obj_mpi_mbx/../input.cpp:1528
 6 0x0000000000416494 LAMMPS_NS::Input::execute_command()  /work/home/acqz25xa41/software/lammps-stable/src/Obj_mpi_mbx/../input.cpp:787
 7 0x0000000000414cb8 LAMMPS_NS::Input::file()  /work/home/acqz25xa41/software/lammps-stable/src/Obj_mpi_mbx/../input.cpp:302
 8 0x0000000000410aeb main()  /work/home/acqz25xa41/software/lammps-stable/src/Obj_mpi_mbx/../main.cpp:105
 9 0x00000000000223d5 __libc_start_main()  ???:0
10 0x00000000004109a9 _start()  ???:0
=================================
srun: error: g02r1n08: task 14: Segmentation fault
srun: launch/slurm: _step_signal: Terminating StepId=11578238.0
slurmstepd: error: *** STEP 11578238.0 ON g02r1n08 CANCELLED AT 2024-05-27T10:28:12 ***
srun: error: g02r1n08: tasks 0-13,15-29: Terminated
srun: Force Terminated StepId=11578238.0

It looks like it got stuck in "fix 2 all mbx 1 5 h2o print/settings json "mbx.json" print/dipoles".

I apologize for any trouble or inconvenience I may have caused, and I am immensely grateful for any advice you provide.

在2024年3月29日星期五 UTC+8 03:16:01<Richa Rashmi> 写道：

[M. X. L]

I solved it! I write my solution here in the hope that it will help users like me who have no experience with lammps

1. The pair_style in lammps.in should be consistent with twobody_cutoff in mbx.json.

2. It should be "fix              3 all mbx 1 h2o 1 2 3 2 1 1 json mbx.json print/settings print/dipoles", and that corresponds to O=1, H=2 in initial.data.

3. A lot of attention needs to be paid to the consistency of initial.data and trajectories. For non-periodic systems it may to add a large enough box first.

This is a rerun of lammps.in, and I hope can it help sombody.