Introduction to Atomic Potentials for Molecular Dynamics

[@mohamedadel9295]

I liked your videos and logic much. It helps me a lot in the start of LAMMPS. Thanks.

[@mpmeran]

Thank you, Brad. Your video tutorials are very helpful. I really appreciate it. I am wondering if it is possible to prepare a video for coarse-grained simulation of polymers.

[@fionagwynethvandermerwe6817]

Great Video and significantly helpful. Please, can you make videos on how to conduct tests like tensile, hardness, thermal and impact on polymer/(graphene/any nanotube/any nanocluster) composites? I've been on this for years now. Thanks

[@ernestobarraza1469]

Thank you! Could you recommend a book on molecular dynamics?

[@bishwasadhikari4837]

Computational materials science by Jume Gunn Lee

[@martinelizalde3998]

6:05, has anyone tried to skip the for loops of the double summation with a matrix multiplication? maybe it is not possible

[@Jere5120]

That is a great question. Whether it is possible or not I think would definitely depend on the potential. Now, in practice I don't think that it comes down to a matrix multiply very often. At least the many body potentials I am familiar with all have limits to avoid doing the calculations for as many atoms as possible (for example, distance cutoffs or shielding). This makes it a bit more conducive to do a loop.

[@anjalishankar]

how to create sw potential file? i don't understand which parameters to substitute for different different elements. where can i find all those parameters?

[@movaffaqkateb7894]

Thanks. Could you please make a tutorial video on using potfit for making meam or any other potential?

[@gustavo_2298]

When atoms move apparently more than they should at ~room temperate is the problem in the potential? I'm using a ReaxFF potential from papers to simulate ZnO, and the erro below appears.. ERROR: Pair reax/c requires use of fix qeq/reax (../pair_reaxc.cpp:366) Any hint is very useful, thank you!

[@marufmorshed3612]

Hello, I want to do a MD simulation of a Binary structure, but the required potential file is not present in Lammps Software, I found a Journal Paper that contained the necessary potentials . Can you tell me how i can build a potential file from the given potentials in the journal paper so that i can use it in Lammps

[@narogen3431]

Great vid, thanks a bunch.

[@mohamedadel9295]

It is nice to me to see LJ for metal material. I have tried both EAM and LJ potentials within the metal micro-cutting process, but I believe LJ is better. However, the evidence is required in the case of LJ as it is little applied in metal as in literature, Is their references or evidence for what you report for LJ in metal? Thanks in advance.

[@youssefouhti4922]

Thank you so much 😊

[@stefanhiemer751]

One question: How do you acchieve to the high/low fidelity classification? I ask it, because I have serious doubts about putting a potential like REAX as high fidelity for two reasons: REAX is incredibly expensive, so you will not be able to relax your structures in any way and catching even the start of proper dynamics in a polymer will be close to impossible. Additionally REAX has a lot of parameters, whose impact on the final simulation result is not trivial and the parameterization is a challenge. To me this high/low fidelity classification actually sounds more like a tunability or computational cost classification. PS: Thanks for your videos.

[@Jere5120]

This is a great point! First answer: these are my classifications and certainly not everyone agrees with me! Second, I am not trying to say that the "high fidelity" potentials are necessarily better, certainly not for all use cases. You are definitely correct that the REAX potential (as well as MEAM(-BO) and others) are very expensive - many of my colleagues use potentials like a Dreiding potential to relax the structures prior to doing other analyses. However, we have seen that the simpler potentials (for polymers, for example) don't necessarily capture correct free volume characteristics or glass transition temperatures. All that to say: the ones I labelled "high fidelity" have the potential to better represent a variety of energy states. The caveat, that you mentioned, is that these usually have many parameters - so they are only as reliable as the states that you validate them in.

[@stefanhiemer751]

@@Jere5120 Have you tried refitting these generic potentials to said states of interest? For example COMPASS/PCFF force fields are also generic, have more degrees of freedom but unfortunately do not allow bonds to break. However those potentials are still easy to assess and comparably fast. Since you mention Dreiding, I would advise against using that one. Dreiding was made by Goddard to be transferable to unkown molecules, but precision was not the goal of this potential. At that time, this was already a success, since apart from UFF there was no transferable force field for organic molecules. Maybe one should go for more often used and better parameterized potentials like AMBER to get things like glass transition closer to the preferred states.

[@mymotivation_amz]

hello sir, which potential we can use for PEO+LiTFSI solid state electrolyte system ? Also i want to know how can we create data file by using mpich2 ? I have been studying the Li ion transport in solid state electrolyte in the molecular level with the help of lammps. The molecular amorphous structure was modeled in the Materials Studio with PCFF forcefield and saved the CAR file which then read by the Lammps tool called “msi2lmp” to give data file for lammps, But when i run BATCH FILE now there appeared some error information, such as“Unable to find bondbond data for cp c2 cp, Unable to find bondangle data for cp c2 cp”, etc. Could anyone tell me how I can get the complete pcff forcefield datafile? I'd really appreciate it if you give me any advice or tutorials. error looks like this unable to find bond data for sf n unable to find angle data for o= sf n etc please reply I really need help

[@ankitanayak951]

what are the steps necessary to add a new potential file to LAMMPS?

[@anshulgour4668]

please make a video on TIP4P/2005 water model

[@Jere5120]

Added to my list for a potential future video!

[@shreyanghosh9381]

Hello sir can you please let me know how can we make out own potential file , I mean if I am having two different atoms which have own cohesive energies bit for adhesion how can we make our own potential system.

[@IbrahimMohamed-gn3wv]

Hello Sir, i have a question. can i simulate water molecules under the effect of magnetic field using LAMMPS?

[@Jere5120]

Definitely you can! You will have to carefully choose your potential. Unless there is a water potential that takes into account magnetism naturally (I don't think so? But check!), you will probably want to use some kind of hybrid potential that combines a magnetic force with the normal force field and apply the known equilibrium magnetic charges on the water molecules. That's my first thought. Thanks for the question!

[@mahesh16095]

Hello Sir Thank You for this information Can you tell me how to choose potential for nanomaterials (Ni-Ti) and Intermetallic Compounds (Mg2Si)

[@Jere5120]

Hello! You have some options for those two systems, but it definitely depends on what types of simulations you want to do. If you are looking at NiTi as a shape memory material, you probably want something that can capture mechanical properties - potentially MEAM or ADP (there are always others that I don't know about as well). The same type of advice goes for Mg2Si - if you are looking for mechanical properties, you probably have to go with MEAM or ADP or similar because Mg's reference crystal structure is HCP. I know an Mg-Si potential is out there for MEAM, and I think a Ni-Ti potential has been published - but you always need to check on how good the published potential is for your purposes.

[@hamdanisoumia2777]

How should I choose the energy minimization ?????

[@Jere5120]

Are you referring to the energy minimization of the system in LAMMPS? Typically I have found that, as long as your initial system is pretty near an energy minimum, the parameters or type of energy minimization is not very important. There may be some differences for some systems though. What are you working on?

[@hamdanisoumia2777]

@@Jere5120 First of all I want to thank you for the helpful information you give on this domain.I work on nanoindentation of metallic bilayer systeme, my algorithm of simulation is divided on two step ,within the system is relaxed under npt ensemble and box/relax using energy configuration # Energy Minimization min_style cg minimize 1e-16 1e-16 100000 100000 NB.this system comprised 254352 atoms afterwards the indentation step'll applied on the free surface Under NVT ensemble using data file obtained by the relaxation step at this stage I didn't use Energy Minimization . the problem that 1-lammps couldn't assign all atoms from data file 2-the relaxed structure wasn't totally homogenous Please ,Could you help me in this matter ? How could we relax the bilayer system correctely, How could we confirm that we have chosen the right tolerance for energy minimization???? and thank you in advance

[@ricardocastillo6656]

You'll find the lines for minimization in here: lammps.sandia.gov/doc/minimize.html

[@allandiayedieng5224]

thank you !!!

[@anjalishankar]

I want to create the sw potential file for binary system...