Phonon Calculations in Materials Science using VASP & phonopy
ฝัง
- เผยแพร่เมื่อ 1 มิ.ย. 2024
- Kindly Click Here: bit.ly/2UtvbHE
Phonon Calculations in Materials Science using VASP & phonopy.
In this unit, I talk about the phonon calculations in materials science using VASP and phonopy.
Phonons play a major role in many of the physical properties of condensed matter, such as thermal conductivity and electrical conductivity. The study of phonons is an important part of condensed matter physics.
By listening to this unit, we will learn how to perform phonon calculations and then obtain thermal properties for our target systems using VASP and phonopy. - วิทยาศาสตร์และเทคโนโลยี
Thank you for this detailed video. As asked by others, please make a video on HSE06 PDOS, Band Calculations. Thank you again :)
Respected
Rasoul Khaledialidusti kindly make a tutorial on magnetic properties and thermoelctric properties. Thanks a million for your efforts
Dear Dr. R. Khaledialidusti, I want to thank you for your very useful resource !
Thank you very much. This is the most important video on TH-cam, thank you for democratizing knowledge. You are an inspiration!
Sir simply outstanding
You effort will make a huge difference for our fast learning.
Please keep it up.
Excellent description. Thank you.
Thanks for your helpful video. Wish to learn more about the DFPT phonopy calculation.
very informative. I am waiting for next vedio
Please explain phonon calculations using DFPT, for getting thermodynamic properties as a function like Cp or Cv
It is really a very nice video, thanks for it and waiting for many more to come!!
I love your tutorials , thank you
Thanks for the video. I am also interested in the way to calculate the thermal property of molecules. I hope you make the video in the future.
Great work,
Thank you
Hi, thank you for your efforts and helpful videos. I use quantum espresso, and I need to calculate a phonon band for a hetterostructure include Pt atom. First I used dim 3 3 1 and the result included a negative frequency. I tried a bigger suplercell but in need a lot of RAM so
I was wondering if I could use the gamma point only and run the calculation for dim 6 61 for example.
Thanks in advance for your help
Thank you very much for this work. I like your channel
At 11:00 what is the run command you use for the self-consistent calculations [scf] (or when ibrion = -1). That is, we used phonopy -d --dim='4 4 1' for the relaxation, but what do we use for the scf? Also, for the force constants did you use an input file for the vasprun.xml or is the vasprun.xml already implemented in phonopy?
Very informative video, thank you a lot. What about electron - phonon coupling constant calculations with VASP? Is it possible to obtain superconducting transition temperature using these phonon calculations?
hey Rasoul
your tutorials are very informative. am currently using phonopy interfaced with qe.
when are we expecting video on how to interpret phonon frequencies
Thanks for your efforts. Kindly make a tutorial about Boltztrap for calculation analysis
This video is very helpful. Thank you. Can you please show how to calculate thermal conductivity?
Dear @Resoul Khaledialidusti, thank you very much its very helpful but i have problem ..plz can you help me to find lattice thermal properties its not mentioned in the video.
Thank you for your good summary! i want to ask you in situation of 'absorbates on support'. I fixed all support atoms but only absorbate atoms were relaxed using finite difference method using 1 1 1 kpoints. In this situation, how much can i set the kpoints? Should i set it only considering absorbates atoms?
Great work, Please make a video on process after this i.e., "how to calculate volumetric thermal expansion coefficient?"
Thanks for information
Sir for graphene monolayer(2dimension) how many Poscar (as in this case six files are generated) files will generate.?
Thank you for your detailed videos
can you please show in detail how to study the thermal conductivity of a compound material optimizing the structure in VASP going through all the different codes in sequence? :)
Thank you very much, it is really helpful
could you explain how to do the calculation of mode Grüneisen parameters also vasp and phono3py please
a very helpful information. But sir I am currently using MedeA version of Vasp and i dont have the output file with .xml extension. Can anyone help me with this issue. Thanks!
Thanks a million sir
Please when you do the hse06, kindly don't forget the bandgap plotting. Since I am always getting zigzag band stucture.
H guys, I am struggling to plot phonon band structure. In the video Primitive_AXIS is set to AUTO that DOES NOT work in my case (Number of elements in PRIMITIVE_AXIS has to be 9.) So, pls, tell me whta should be these axes ? Say, I use dim 2 2 2 should I specify single primitive lattice vecors or should double the values ? Thanks !
Is it relevant for molecular crystal? Thank you
Hi Dr Rasol
thanks 10000* times for the video.
i have 2D material for which i add vacuum above the top surface. so if we make kpoints [X X 1] is generally i do. is that ok for phonon calculations? also for super cell 1-1.5 nm thumb rule also applies for z-direction?
Dear sir.
Can you please make a video on how to install boltztrip in vasp and how to perform thermoelectric properties for Half Heusler.
Please
For the phonon DOS, what are the units of the x-axis frequency?
Hello, thank you for the video. I have few questions:
- Did i need to perform calculations with the kpoints in the symmetry path like for electronic band structure in order to calculate the phonon band structure?
- Can you do the same tutorial but with phono3py to calculate lattice thermal conductivity?
Best regards.
In the "STEP2: Phonon", can you show us the INCAR file ?
Sir can you make a video on IR and Raman spectroscopic simulations using Phonopy-SpectroscoPy with installation tutorial. Thank you.
Thanks for the detailed video. Could you please explain how to add the x and y labels in graphs generated via the pre-process like band.conf ?
I am also looking for the same. Did you get the answer?
In the case of a structure that has symmetric 40 atoms, how can we consider symmetry to decrease the phonon lines?
where is the next video containing dfpt calculation. please share the link here if anyone has already found it
For structure relaxation you used Gamma centered Kpoints is there any specific reason?
Sir how can convert units in phonopy , I want band structure in cm-1
Hey, I want to know how to control QPOINTS during phonon computations
Thank you for your effort. Please upload the next video as soon as possible explaining the physical significance of phonon spectrum. My Professor is eating me for not knowing the explanation :) .
Have you gotten the explanation now? If yes kindly let me know
Thank you sir 🙏
Hey. Can you please make a video on calculating elastic constants using IBRION 5 or 6 with NSW 1 .. I am having problems in doing that. Thank you in advance.
how to save data of thermal properties, to plot in other ploting softwares like origin. Please if someone could help?
Great video, I was wondering what do you use to plot/visualize the supercell structure you create with phonopy?
Thank you for you feedback. I have used VESTA for the visualization of the structure.
Why should the lattice constant be in range 1-1.5nm ?
Can you give me help for kappa calculation (l attice thermal conductivity)
Thank you for the video,. I have one question. I already have a few 4x4 supercells that I run VASP on for adsorption simulations. I would like to run phononpy on them to determine the gibbs energy etc. So the question is can I just use phonopy -d --dim="1 1 1" to create the neccesary displacementfiles without changing my surface size?
Nevermind
Did you get an answer? It is a very good question?
Actually, I did that using dim=111, but the problem is it creat 300 new structures and my model contains 71 atoms, so you see I can not apply the same procedure because it will take forever to calculate all the structures. In the adsorption study, it will be enough if we get the phonon for the adsorbed molecule only, no need to calculate the phonon for the slab, so there must be a way to consider only the adsorbed molecule like we do the frequency calculation of the adsorbate only in VASP.
thanks
Thank you very much for the explicit presentation. I am interested in Phonon calculation using DFPT. Please Share the link with me if you have done it. Many Thanks!
I will make a tutorial to explain the DFPT methodology.
@@RasoulKha still waiting.
Hi,
1. I have created FORCE_SETS and now am trying to produce FORCE_CCONSTANTS using "phonopy --fc dis-001/vasprun.xml dis-002/vasprun.xml dis-003/vasprun.xml...." with FDM. But it is showing that "'dis-001/vasprun.xml' does not contain the necessary information" message.
You mentioned in one comment to write "FORCE_CCONSTANTS = WRITE" in the bash file. I did not get the point. Can you please clarify/specify in which file I need to write that?
2. How can I get the Group_Velocity ?
"phonopy --fc dis-001/vasprun.xml dis-002/vasprun.xml dis-003/vasprun.xml...." creates FORCE_SETS. If you already made FORCE_SETS, you need to add "FORCE_CCONSTANTS = WRITE" in the band.conf file and then type "phonopy -p -s band.conf" to get the FORCE_CCONSTANTS and phonon dispersion. If everything is performed right, you should not get any error message.
@@RasoulKha Thank you very much. I was able to create FORCE_CCONSTANTS by phonopy --writefc. Can yo please suggest me the way to create GROUP_VELOCITY file?
Thank you so much! I've just begun trying Phonopy.
Some questions!
The FORCE_CONSTANTS can be obtained from DFPT calculation. What are the differences between FORCE_SETS and FORCE_CONSTANTS?
I now know that FORCE_SETS is an ensemble result of all displacements. Are they the same? Which one is more accurate?
Are you planning to upload the 'phonopy-qha'?
When you follow the procedure to phonon dispersion calculations and set FORCE_CCONSTANTS = WRITE in your bash file, then, the force constants will be produced. Generally, both DFPT and FDM are the same with the same accuracy. I have plan to talk about DFPT as well in the near future.
Sure, Quasi-harmonic Approximation (QHA) is also in my plan list and will talk about that. Meanwhile, if you have any questions, please feel free to contact me.
@@RasoulKha Thank you for the reply. One more I have... I worry about the qualitative y-axis in the graph of the thermal properties. Let say I use conventional FCC Al structure as POSCAR (4 atoms in the 1x1x1 cell). Then, each quantity is the value per 4 atoms? Here, the POSCAR is 1x1x1 unit cell, and POSCAR in dis-001 is displaced 2x2x2 supercell.
@@nokilee489 Each quantity is per mole. You need to change the units if you need them per atom.
Thank you very much for all your explanations. One question, is it possible to apply the same procedure to a slab structure with a defined vacuum space in the z-direction? or should it be bulk structure?
Thanks in advanced for your time
The same procedure can be applied to the slab structures.
@@RasoulKha Hi Mr. Khaledialidusti. After installing phonopy using conda in a Windows pc (I tried to install it in Linux environment in a specific cluster but the lack of admin permissions did not allow it). However, the installation of version 2.4.2 was successful. The problem now is that when I run the command phonopy -d --dim "1 1 1" it generates 402 images. My system is a slab with dimensions of 0.93x1.27x1.82 nm with 0.8 nm vacuum space in the z-direction. Is there any way to obtain a lower number of images (e.g. with your system you got 6 images)?
Thank you very much in advance for your collaboration.
@Rasoul Khaledialidusti Mr. Khaledialidusti, just to be more concise, I am checking both vasp and phonopy documentation and found some interesting insights about phonon calculations using DFPT method. I know you mentioned that it is a whole different method which is going to be addressed in another video. But taking into account the number of images I am getting for the finite-difference method, just want to ask for any advice regarding the tags each site propose to use ( www.vasp.at/wiki/index.php/Phonons_from_finite_differences phonopy.github.io/phonopy/vasp-dfpt.html?highlight=dfpt )
@@franciscoospinaacevedo6923 Yes. There is. You need to take the structure with the highest possible symmetry. By --symmetry option in phonopy and the chosen tolerance value, the structure with the highest possible symmetry will be obtained. If you take that structure and optimize with ISYM = 1, then, the single atomic displacements would be less.
Sir, how we can decide dimension --dim tag for phonon calculation? .My system contain 3 atom in a unit cell with 4.23 lattice constant
It should be between 1-1.5 nm.
4.23 Å is not enough for the phonon calculations and you need to make it at least 3 times bigger.
Thankyou so much sir
Great!!!! could you pls make a video for HSE calculation? Thanks in advance.
For the next one, I will do that.
Vasp default ISYM=2 which is memory conserve charge symmetry method (same as ISYM=1). Hence if we don't take ISYM tag vasp should take care of my symmetry.
kindly enlight me on this.
Nice illustration of phonopy!!!
agree, in vasp specifications it is recommended to use with USPP, which is not my case for example, so ISYM=2 seems to be more universal way, video is very useful, thx!
Rasoul, please! How do you get the MP = 24 24 48 information? I mean, how do I know such numbers from my input/output files? Thank you very much! Very nice video, already a subscriber =)
@Rasoul I have the same question as Guilherme above. My first thought is that MP should be the same as what I used for the calculations but you don't do that. Are they independent of each other? All I can tell is that in your example they are the inverse of the dim where dim was 2 2 1 and for MESH.conf they are in inverted proportions (makes sense for the reciprocal space) but I'm not seeing a rhyme or reason to 24 24 48.
Respected Rasoul Khaledialidusti, I am dealing a mono layer contains 32 atoms. Is it possible to do calculation by keeping the same super cell. Mean if I do not increase the number of atom by running the command phonopy -d --dim="1 1 1"......
Yeah. It would be possible. However, generally, the lattice parameters should be almost 1.5 nm. If the negative phonon branches were observed in the phonon calculations, you need to apply the supercell.
At 15:55, you show the command "phonopy -f dis-001/vasprun.xm dis-002/vasprun.xm ..." to include all the files.
Is there a quicker way to do this on windows? I know on linux, you can do something along the lines of "phonopy -f dis-{1..6}/vasprun.xml", but I can't find a way to do this in windows... :(
Does "phonopy -f dis-*/vasprun.xml" work for you?
@@chocoberrymuffins No it does not
@@dylanevans6855 Did you happy to copy and paste my comment into your command line? I had to use - instead of - because youtube was reformatting my comment weirdly when I used regular dashes.
@@chocoberrymuffins I tried both. Also, that was a recommended way to input the code on the phonopy website, and none of their suggestions worked either.
Can you please share the link for phonon calculation using DFPT Method video.
have you got any idea how to use DFPT? if yes please share it.
@@apphysics163 share link if you have good resource or video like this one.
Thank you very much for the video.
Could you please explain a bit about the K path needed for the phonon band structure?
Thanks for the feedback. K-path for the band structure is dependent on the structure type. If you are not familiar with the proper K-path of your structure, I would definitely suggest using VASPKIT, which gives you the best path.
@@RasoulKha
I am able to get the high symmetry k path from Materials cloud/seek path. But I am not able to understand how to write it in the band.conf file.
Suppose for an example the high symmetry k path be
Γ-X-P|G-M
and the coordinates are
G 0.5000000000 0.5000000000 -0.0762743321
Γ 0.0000000000 0.0000000000 0.0000000000
M 0.5000000000 0.5000000000 -0.5000000000
P 0.2500000000 0.2500000000 0.2500000000
X 0.0000000000 0.0000000000 0.5000000000
How can I represent it in the band.conf file?
Please suggest. Thank you.
@@nirajamoharana1622 you need to only write them in a row: BAND = 0.5000000000 0.5000000000 -0.0762743321 0.0000000000 0.0000000000 0.0000000000 0.5000000000 0.5000000000 -0.5000000000 0.2500000000 0.2500000000 0.2500000000 0.0000000000 0.0000000000 0.5000000000
@@RasoulKha
Shall I have to follow the order of the k path ?
@@nirajamoharana1622 sure. You need to follow the path that you desire to have.
Great effort dear,
Can you please help us in using hse06 bandgap calculations earlier. I need it to submit one pending task.
I couldn't find any clear tutorials as yours.
Anyway, I will be waiting for your tutorial.
Regards
Sure. I will make it as the next one. :)
@@RasoulKha Thank you very much, I am very excited for that.
I want to extract force constants using DFPT using VASP and create a fc2.hdf5 file. Is this possible using phonopy? I want to avoid extracting forces from multiple files(vasprun.xml) created by each displacement in the video you have shown
Yes. It is possible. You can generate force constants using both methods of DFPT and FDM. Generally, both methods are the same, and Force constants can be produced.
@@RasoulKha Then how do I go about doing that?? There is no proper procedure in phonopy site to create the fc2 file which I need. Different question : any good books or papers to study the computational procedure for a beginner??
@@dwaipayanroychowdhury5279 When you follow the procedure to phonon dispersion calculations and set FORCE_CCONSTANTS = WRITE in your bash file, then, the force constants will be produced.
On Ubuntu 20.04 the command phonopy -d --dim "2 2 2" is correct WITHOUT '=' sign
Many thanks for your reply
Hello, I already have sqrt(3)xsqrt(3)x3 structure, containing 78 atoms. If I increase the dimension, the number of atoms will be large enough for VASP calculation. How do I start the phonon calculation? **************phonopy -d --dim=" ? ? ?" ************************** to create the necessary displacement files
The dimension would be enough depending on your system. In most cases, it would be fair enough.
Thank you so much for this useful video.
I have installed the phonopy on linux which seems to be ok. But when I wanted to make super cell using phonopy -d --dim = "3 3 1" , I have just received the below error
ValueError: invalid literal for int() with base 10: '='
Could you please let me know what wrong is?
Hi Tahereh Mohammadi
, I just installed version 2.4.2 and got the same error. It was fixed by removing the "=" sign from the command line. Hope it works for you too
if you get an error, please try : phonopy -d --dim="3 3 1" -c POSCAR
@@franciscoospinaacevedo6923 Thank you :)
@@RasoulKha Thank you :)
While creating the supercell for nanowires (vacuum in x and y directions), i used phonopy -d --dim=113. But it does not create any output files. Could you or anyone please guide me, what am i doing wrong. I get the following statements after typing this command.
Number of elements of DIM tag has to be 3 or 9.
Please check the setting tags and options.
Many thanks
Hello Dear Ghulam,
I think your problem might just be a simple case of a syntax error. It's precisely what the error message states. The DIM tag in your statement expects 3 - 9 elements. You have entered "dim=113", which translates to a single element. You would have to enter "dim = 1 1 3" instead. The spaces differentiate the elements from each other.
I am repeating the same procedure for graphene/mos2 heterostructures as it has about 99 atoms so after processing with phonopy I am getting about 200 POSCAR as there are six in your case. So for my case it would be difficult to repeat the same case by 200 times. Please enlighten something. Thanks in the advance.
Your structure has low symmetry and that is why you got 200 single-atomic displacements. I would suggest you finding the symmetrize of your structure and then relax it with ISIF = 1. Then, obtain the supercell with phonopy. You would get less sing-atomic displacements.
@@RasoulKha But in my case I have taken already 5*5*1 supercell of graphene and 4*4*1 supercell of mos2 so lattice constant are already 12.435 angstrom means they are nm range so if I will make supercell of this again then no of atoms will increase further. Also as my case is in two dimension, there is vacumm of 20 angstrom along z axis so what should be the appropriate KPOINTS. I kept 4*4*1
@@RasoulKha Also thanks for the early and valuable reply.. Your work is really incredible
@@renusingla2402 4*4*1 is a proper choice for the KPOINTS.
@@RasoulKha ok and so how should I move further as I have already a large supercell?
Hi sir,
can you upload a video regarding the elastic constant of 2d materials.
Yes, sure
Thanks sir fir your response
Im trying to calculate dielectric constant but facing some problem
@@RasoulKha can i claculate elastic constant with respect to time by phono?
@@atulpandey8727 try using the vaspkit code in order to calculate elastic constants 2d or 3d materials it works fine