About negative value in Structure Factor calculation, it matters my defense.

[Hu Li]

To all developers and users who are familiar with Structure factor: 

In  (Silicon Liquid Structure and Crystal Nucleation from Ab Initio Deep Metadynamics  )

, it refers to a code to calculate structure factor.

I had used this code to identify different carbon phases, it worked well in small system, just had one significant question here. When I use it calculate diamond structre, 216 atoms, I got negative value. 

LOCAL_DEBYE_STRUCTURE_FACTOR ...
  LABEL=cv1
  CUTOFF=4.0
  ACTIVE_Q=1.90
  MEAN
  MORE_THAN={RATIONAL R_0=0.60 NN=12 MM=24}
  SPECIES=1-216
... LOCAL_DEBYE_STRUCTURE_FACTOR 

it indicates I calculate the number of atoms which is larger that 0.60.

All my work has done, but this question was left aside initially. This is a fatal to my defense. Hope any suggestion will save me. 

Hope you all well.

iota here.

[Michele Invernizzi]

Hi,

I am not sure I understand your question, but yes, you can indeed get negative values for that CV, depending on the CUTOFF and other parameters. This shouldn't be a problem if your goal is to use the CV to drive sampling and not to compare with experimental results. The Debye structure factor assumes isotropy (see here), which is typically not correct for the small systems used in MD.
You can get a better estimate of the experimental xrd with STRUCTURE_FACTOR_SPHERICALLY_AVERAGED, see https://github.com/invemichele/plumed2/blob/structure_factor/src/structurefactor/README.md or https://docs.lammps.org/compute_xrd.html

Regards,

Michele

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/plumed-users/fa984634-9756-493b-bfa5-bdbfd7a675c2n%40googlegroups.com.

[Luigi Bonati]

Hi,
in addition to Michele's answer, I think a cutoff of 4 Ang. is too short to measure the presence of order through the structure factor (in the paper cited for the diamond structure of Silicon 8 Ang. was used).

Best, 
Luigi

To view this discussion visit https://groups.google.com/d/msgid/plumed-users/CAGBhpicdiOvwT8gbd%3D-r6hU%3D1MjfEWCZscS1oAaOU%2B_UJq6%3DeQ%40mail.gmail.com.

[Hu Li]

Many thanks to  Michele and  Luigi for your precious comments.

Before sampling, I tested on cutoff 4 and 8, it all worked well and all got negative value for diamond, but it was more than enough to distinguish interested phases. For saving  computational resources, I chose 4 in the end. After sampling,  I checked many times on the configurations in results by OVITO as a double check. 

 I assume you are right, it is because the system was too small, because I tested bigger diamond system, the valuse is positive. However,  CVs drives the system slowly and it cannot work basically(Or maybe it takes time because of big system(about thousands of atoms).? I only let it run one day ). 

 Btw, I learned tons of things from your work, as well as the tutorials online. Thanks! Hope I would make some contributions as you made. 

Best regards,

Weihu Li (iota)