Stretching a full periodic box embedding rough walls

Dear LAMMPS community,

I have a query regarding the deformation of an inorganic-organic material. The structure of this material can be visualized as a sequence of alternating layers of inorganic and organic materials, with a repeated pattern of organic-inorganic-organic-inorganic… in one direction only.

My objective is to extract the stress-strain curve of this material system using the “fix deform” command in a fully periodic cell, which is a straightforward process. However, for certain reasons, I need to immobilize the inorganic layers by freezing their atoms within the same fully periodic cell and perform the same deformation. Unfortunately, during the remapping process, the frozen atoms can experience stretching, which can interfere with the mechanical response of the bulk phase of the material.

I am seeking suggestions or ideas from anyone who may have encountered a similar situation. Specifically, I would like to know how to deform the full periodic box while keeping the inorganic layers fixed as if they were rough walls.

Thank you for your time and consideration in advance.

Regards,

Reza.

You can try to fix deform the “fluid” group instead of group all, and you can further refuse to integrate the “solid” group.

If that doesn’t work, you can always make a group of particles obey any equations of motion you want using atom-style variables and a combination of velocity (for initial velocities), fix move or fix setforce commands.

Having said that:

I’m going off-topic (for the LAMMPS forums) but unless your reasons are very compelling, I’m immediately very doubtful about this way of doing things. If you are imposing shear on a macroscopic sample of the material you should not expect one set of layers to be immobilised (to angstrom precision!), and so you should deform the whole box. On the other hand, if you are using nanofluidic control to move an organic fluid relative to an inorganic framework, you should customise the fluid’s motion (using fix addforce for the most customisability), but leave the box unmodified to reflect the relative immobility of the framework.

Not that my opinion matters – I am just an Internet stranger, really – but I hope you have a good justification and do not end up wasting a lot of time on simulations that do not conform to reality … or, worse still, a reviewer’s expectations. :wink:

1 Like