Please do keep making these tutorials, especially for polymers. Can you help showing a tutorial on how to generate a little bit more complex material using a mol file, making polystyrene or polynorbornene manually is quite painful...
Today I had the satisfaction of having the first contact with your videos. They are excellent. A question: do you have a video teaching how to compute melting temperature using two-phase coexistence method? Thanks a lot.
kindly consider providing the txt files in the video description to download. a polite request. please. And thanx a million. This video series is the most helpful one available in youtube
Hi EnCodeVentor. I think you are missing some dihedrals and angles. It looks like the data file you showed at 21:22 only has dihedral interactions between the carbon atoms (C-C-C-C) and angle interactions for (H-C-H and C-C-C). Most force fields would also create dihedrals for the H-C-C-H, and H-C-C-C, and angles for H-C-C. Both OPLSAA and GAFF generate these kinds of interactions. COMPASS does also. But in addition to angles and dihedrals, COMPASS also creates improper interactions for each carbon atom as well. So I suggest that, at some point in the video, you mention that the number of bonded interactions (angles, dihedrals, impropers) for a given molecule varies depending upon the force field being used. ---- Long version (and some ranting) ---- For a traditional force field like OPLSAA, I would expect 10 C-C-C-C dihedrals, 10x4=40 H-C-C-H dihedrals, 10x2=20 H-C-C-C dihedrals (70 dihedrals total), ...and 10 C-C-C, 10 H-C-H, and 40 H-C-C angles (60 angles total). However, I could be wrong. I often make mistakes and omit or miscount interactions myself when attempting it manually. Things can get messy. (As mentioned above, COMPASS also creates improper interactions for polyethylene. OPLSAA doesn't. To add to the confusion, I've noticed that the number of dihedral and improper interactions generated by COMPASS and OPLSAA do not always agree, although it'spossible this could be a software artifact.) The process of building the files needed to run an MD simulation is so laborious and byzantine, people often choose which simulation software they want to use (LAMMPS, OpenMM, GROMACS, NAMD, AMBER, ESPResSo, MedeA, MaterialsStudio, Desmond ...) depending on which tool comes with with the easiest and most reliable molecule builder which supports the force field they want (which usually chosen according to the type of molecule they are simulating). I'd love to say that using a molecule builder program will take care of everything, but this would be ridiculous. There are no fully automated molecule builder tools, and I'm not sure it is possible to write one. (Disclaimer: I wrote one of these programs.) In the long run, I am hoping that efforts like OpenForceField (eg the SMIRNOFF force field) and/or the ATB (atb.uq.edu.au/) will gradually make it easier for users to build arbitrary molecules and have the topology and force field parameters generated reasonably well automatically (for, at least, a large fraction of the molecules people typically care about). Again, I don't think you necessarily need to correct the data file in your video. (There is no "correct" data file, since the number of dihedral and improper interactions varies with different force fields.) My only suggestion is that you warn users that the number and type of bonded interactions is force-field dependent. (This will hopefully give people some idea about the difficulty of the task ahead.) (Incidentally, I suspect the crooked shape of the molecule at 22:40 happens for a different reason. I speculate that perhaps you don't have angles and dihedrals connecting atoms at opposite ends of the polymer across the periodic boundary.)
Thanks for the great and detailed comment! I very much appreciate that you mention the other dihedrals - polymers is not my area of expertise and as you mentioned, there are tons of combinations of dihedrals and angles. I tried to include at least most of them that had parameters from the original Drieding paper, but I know I missed some. I think we traded emails a few years ago talking about your molecular generator?
very helpful comment. How do you calculate the number of angles and dihedrals ? is there any specific formula ? so i can check the generated data file w.r.t what is expected I've generated this data file with Topotools , and i got 60 angles (exactly what you've expected), but the dihedrals are 85 !
Awesome tutorial. One suggestion though, we can create polymers without considering the bond angles. We can just create an unrealistic chain configuration and then do pair_style soft and then energy minimization in LAMMPS. Skipping that step is much less painful haha, esp for complex structures.
Thanks for the tutorial! Very useful! You're missing an H on each end of the polymer: each C should have 4 bonds and the structure should be H-CH2-CH2...CH2-CH2-H Also, I read the Dreiding paper (okay, more like skimmed) and I'm not sure where you got the force field values from. It would really useful if you made a video on how/where to get force field values.
Hi, Is there any way by which we can generate data files for randomly placed or different orientation coarse-grained chains (about 50-75 chains) with 8-10 beads on each chain?
I always recieve an incorrect data format error, Can you help me please? The only way it works is when I add another row in the whole file. Can you explain why this is happening?
Hello EnCodeVentor. I hope you are doing great. I am silently following your tutorials for a few days. Those are really helpful. Can you make a video on how I can link Lapack and Blas with Lammps? It will be really helpful.
@@Jere5120 i am facing another problem. It will be helpful if you can give any suggestion. I would like to build a curved graphene stake with pore in it using python. Can you make a tutorial about it. It could be a lifesaver!
This is a great question - and actually the reason I started making these videos. I couldn't find one really good resource. I haven't put together a comprehensive resource for writing LAMMPS codes, but you should check out the "LAMMPS Input Script" videos on my channel. Besides that, there are some examples distributed with LAMMPS (lammps.sandia.gov/doc/Examples.html) and there are some other websites with some helpful content (icme.hpc.msstate.edu/mediawiki/index.php/LAMMPS_Tutorial_1, www.u.arizona.edu/~stefanb/Files/HPCTutorials/MD-lammps-Final.pdf) I would like to create a comprehensive LAMMPS course.
Thanks for watching! Absolutely it is possible. You can create rigid volumes in any shape supported by the "region" commands in LAMMPS (lammps.sandia.gov/doc/region.html, lammps.sandia.gov/doc/fix_wall_region.html). You can also fix a set of atoms to be a rigid body such that they do not translate relative to each other (lammps.sandia.gov/doc/fix_rigid.html).
Hello! That is definitely a simulation you can do. The first question (as always) is what are you trying to simulate about it? That will begin to define what type of potential you should use.
The tutorials are super helpful, could you possibly upload your py files and lammps files to github? I also see you have grown a quarantine beard much like me lol
Thanks for watching! I do plan on updating my git repo with my most recent video's content. Sorry I haven't done it yet. If you haven't been there, my git repo for the channel is: github.com/bdhuddleston/encodeventor And +1 for quarantine beards!
Yes. Maybe. It depends on the size of bundle you are looking at - you might be limited by the number of atoms you are able to use. But, if you can find a potential that represents it well, you can model it in lammps.
Your tutorials are really helpful. Thank you!
just starting with lammps... this is really helpful. Thanks!!!
Please do keep making these tutorials, especially for polymers. Can you help showing a tutorial on how to generate a little bit more complex material using a mol file, making polystyrene or polynorbornene manually is quite painful...
Today I had the satisfaction of having the first contact with your videos. They are excellent. A question: do you have a video teaching how to compute melting temperature using two-phase coexistence method? Thanks a lot.
thank you.. i m learning..
kindly consider providing the txt files in the video description to download. a polite request. please.
And thanx a million. This video series is the most helpful one available in youtube
Hi EnCodeVentor. I think you are missing some dihedrals and angles. It looks like the data file you showed at 21:22 only has dihedral interactions between the carbon atoms (C-C-C-C) and angle interactions for (H-C-H and C-C-C). Most force fields would also create dihedrals for the H-C-C-H, and H-C-C-C, and angles for H-C-C. Both OPLSAA and GAFF generate these kinds of interactions. COMPASS does also. But in addition to angles and dihedrals, COMPASS also creates improper interactions for each carbon atom as well. So I suggest that, at some point in the video, you mention that the number of bonded interactions (angles, dihedrals, impropers) for a given molecule varies depending upon the force field being used.
---- Long version (and some ranting) ----
For a traditional force field like OPLSAA, I would expect 10 C-C-C-C dihedrals, 10x4=40 H-C-C-H dihedrals, 10x2=20 H-C-C-C dihedrals (70 dihedrals total), ...and 10 C-C-C, 10 H-C-H, and 40 H-C-C angles (60 angles total). However, I could be wrong. I often make mistakes and omit or miscount interactions myself when attempting it manually. Things can get messy. (As mentioned above, COMPASS also creates improper interactions for polyethylene. OPLSAA doesn't. To add to the confusion, I've noticed that the number of dihedral and improper interactions generated by COMPASS and OPLSAA do not always agree, although it'spossible this could be a software artifact.) The process of building the files needed to run an MD simulation is so laborious and byzantine, people often choose which simulation software they want to use (LAMMPS, OpenMM, GROMACS, NAMD, AMBER, ESPResSo, MedeA, MaterialsStudio, Desmond ...) depending on which tool comes with with the easiest and most reliable molecule builder which supports the force field they want (which usually chosen according to the type of molecule they are simulating).
I'd love to say that using a molecule builder program will take care of everything, but this would be ridiculous. There are no fully automated molecule builder tools, and I'm not sure it is possible to write one. (Disclaimer: I wrote one of these programs.) In the long run, I am hoping that efforts like OpenForceField (eg the SMIRNOFF force field) and/or the ATB (atb.uq.edu.au/) will gradually make it easier for users to build arbitrary molecules and have the topology and force field parameters generated reasonably well automatically (for, at least, a large fraction of the molecules people typically care about).
Again, I don't think you necessarily need to correct the data file in your video. (There is no "correct" data file, since the number of dihedral and improper interactions varies with different force fields.) My only suggestion is that you warn users that the number and type of bonded interactions is force-field dependent. (This will hopefully give people some idea about the difficulty of the task ahead.)
(Incidentally, I suspect the crooked shape of the molecule at 22:40 happens for a different reason. I speculate that perhaps you don't have angles and dihedrals connecting atoms at opposite ends of the polymer across the periodic boundary.)
Thanks for the great and detailed comment! I very much appreciate that you mention the other dihedrals - polymers is not my area of expertise and as you mentioned, there are tons of combinations of dihedrals and angles. I tried to include at least most of them that had parameters from the original Drieding paper, but I know I missed some.
I think we traded emails a few years ago talking about your molecular generator?
very helpful comment.
How do you calculate the number of angles and dihedrals ? is there any specific formula ? so i can check the generated data file w.r.t what is expected
I've generated this data file with Topotools , and i got 60 angles (exactly what you've expected), but the dihedrals are 85 !
very helpful, thanks!
Awesome tutorial. One suggestion though, we can create polymers without considering the bond angles. We can just create an unrealistic chain configuration and then do pair_style soft and then energy minimization in LAMMPS. Skipping that step is much less painful haha, esp for complex structures.
Absolutely right! (as long as your potential is good enough!)
Thanks for the tutorial! Very useful!
You're missing an H on each end of the polymer: each C should have 4 bonds and the structure should be H-CH2-CH2...CH2-CH2-H
Also, I read the Dreiding paper (okay, more like skimmed) and I'm not sure where you got the force field values from. It would really useful if you made a video on how/where to get force field values.
Hi, Is there any way by which we can generate data files for randomly placed or different orientation coarse-grained chains (about 50-75 chains) with 8-10 beads on each chain?
Hi, Can you do a video on creating bilayer graphene or h-BN system and running it using drip potential?
I always recieve an incorrect data format error, Can you help me please? The only way it works is when I add another row in the whole file. Can you explain why this is happening?
Thank you very much
so useful! thank you so much
Hello EnCodeVentor. I hope you are doing great. I am silently following your tutorials for a few days. Those are really helpful. Can you make a video on how I can link Lapack and Blas with Lammps? It will be really helpful.
This is a great question. I have build LAMMPS many times, but I can't recall if I ever linked it with Lapack and BLAS. I will look into it!
@@Jere5120 Thanks for your reply. I already did it using user-atc package.
@@Jere5120 i am facing another problem. It will be helpful if you can give any suggestion. I would like to build a curved graphene stake with pore in it using python. Can you make a tutorial about it. It could be a lifesaver!
Can you suggest me a book or any platform from where I can learn writing lammps codes from scratch?
This is a great question - and actually the reason I started making these videos. I couldn't find one really good resource. I haven't put together a comprehensive resource for writing LAMMPS codes, but you should check out the "LAMMPS Input Script" videos on my channel. Besides that, there are some examples distributed with LAMMPS (lammps.sandia.gov/doc/Examples.html) and there are some other websites with some helpful content (icme.hpc.msstate.edu/mediawiki/index.php/LAMMPS_Tutorial_1, www.u.arizona.edu/~stefanb/Files/HPCTutorials/MD-lammps-Final.pdf)
I would like to create a comprehensive LAMMPS course.
@@Jere5120 Thanks for sharing these useful resources...Can I have your email-id ?I want to talk with you regarding this?
Sure! You can email me at my channel email: encodeventor@gmail.com
Nice video. Sir, is it possible to create rigid sphere or rod like patchy particle(Janus)?
Thanks for watching! Absolutely it is possible. You can create rigid volumes in any shape supported by the "region" commands in LAMMPS (lammps.sandia.gov/doc/region.html, lammps.sandia.gov/doc/fix_wall_region.html).
You can also fix a set of atoms to be a rigid body such that they do not translate relative to each other (lammps.sandia.gov/doc/fix_rigid.html).
Hello please i have some problem on my simulation on nanoscale of C-S-H can you please give me some tips on that
Hello! That is definitely a simulation you can do. The first question (as always) is what are you trying to simulate about it? That will begin to define what type of potential you should use.
The tutorials are super helpful, could you possibly upload your py files and lammps files to github? I also see you have grown a quarantine beard much like me lol
Thanks for watching! I do plan on updating my git repo with my most recent video's content. Sorry I haven't done it yet. If you haven't been there, my git repo for the channel is: github.com/bdhuddleston/encodeventor
And +1 for quarantine beards!
wow, amazing！
Can we create carbon fiber bundle in laamps
Yes. Maybe. It depends on the size of bundle you are looking at - you might be limited by the number of atoms you are able to use. But, if you can find a potential that represents it well, you can model it in lammps.