(no subject)

Dear Sir!

I speak in the name of a research group which is focused on solving problems involving collisions of gas particles from surfaces using numerical simulations. Our goal is to determine the effects of asymmetrical scattering of particles from a general surface on gas dynamics, and to apply our analysis to a specific case, e.g. Argon scattering from mono-crystalline Lithium Flouride (LiF).

We approached the problem iteratively, refining the simulation ever so more (at least that is the plan) to achieve correlation with measurements obtained by the fellow researchers - mainly we are considering the work of Salvador Miret-Artés and Eli Pollak, who wrote an article titled Classical theory of atom–surface scattering: The rainbow effect (published by Elsevier in 2011 - link to the article: https://www.sciencedirect.com/science/article/abs/pii/S0167572912000179).

Not to get too much into the details of it - it is well known that pure elastic collisions of particles with an ideally flat surface are symmetrical. The article mentioned above however presents a more complete model, which describes such collisions as asymmetrical (which means that it isn’t elastic), as well as bi-modal with a known probability distribution function, with the velocity and inbound angle as the inputs to their model and the outbound velocity probability and outbound angle probability as the output of their model.

We would like to implement this kind of model into LAMMPS in order to improve our simulation. One of the possibilities is to redefine and modify one of the existing functions (for example fix wall/reflect) to get the model working, but currently we’re having difficulties with the complexity of the code (none of our team members are great programmers, to be honest) and we’re searching for an easier approach. Could you please give us some guidance on implementing such a function in LAMMPS or perhaps offer your own observations?

We would like to express our sincere gratitude for any help you may offer us and wish to thank you for your time.

Your sincerely,

Uros Radosavljevic

University of Ljubljana
Faculty of Mechanical Engineering
Askerceva 6
1000 Ljubljana

Slovenia

two comments:

  • if modifying fix wall/reflect already gives you headaches and concerns, then you need to find a collaborator with programming experience, since this is one of the rather simple pieces of code in LAMMPS
  • i don’t quite get why you want to do an advanced wall model, if you can have the real thing right away. i.e. you can build a wall of (classical) particles and use fix deposit to hurl atoms at it and then you can study the distribution of reflected particles as it would be for real. well, for as real as the (classical) force model (i.e. pair style) is. you can even model damage/sputtering with the right velocity of the deposited particles. so why add a more complex wall model into LAMMPS? what would be the benefit to other LAMMPS users or for you?

axel