Ground state of NiI2 inconsistent with DFT

[Loris Naudin]

Hello,

I'm trying to obtain the ground state of monolayer NiI2 without electric field, and I consistently obtain spin textures different than the spiral state. It's incoherent because DFT predicts that the ground state is spin spiral and by other Monte-Carlo simulations we obtain spin spiral as well. This is in order to clear out a debate about ground state of NiI2 because some papers say it's spin spiral, others skyrmion lattice etc. The picture (spins2) looks like a random structure, it doesn't even look like more rare phases lire stripes domain wall or labyrinthine phases.

I've used a customized ucf file to model a hexagonal lattice, and I've used all the suitable parameters. I would like to have more informations about time-series program too. What I know is that there is an equilibration time followed by integration, and that it is made for equilibrium calculations.

I give all files attached

Loris Naudin - Intern at CEMES - CNRS (France)

[gabo...@gmail.com]

Not sure if it helps, but I have seen that DFT (ab initio) and atomistic can have different length scales:

Figure from https://psi-k.net/download/highlights/Highlight_123.pdf

There was a post in the WIEN2k mailing list that could be of interest as I remember it talking about how DFT might not account for magnetic domain physics which could cause magnetic results to differ between DFT calculations and atomistic/micromagnetic calculations.  That post should be here:

https://www.mail-archive.com/wi...@zeus.theochem.tuwien.ac.at/msg17489.html

Kind Regards,

Gavin

VAMPIRE user

[Loris Naudin]

Yeah DFT and Atomistic Spin Dynamics have different length scale but with another Monte-Carlo code my supervisor obtain spin spiral so something must be wrong with my simulation, but I've verified everything, interactions, the geometry etc. (everything seems fine) and I've always obtained these strange structures and no spin spiral whatsoever

Thank you for the help

Best regards,

Loris Naudin

[idomeneo-spinuser]

I think it might be something wrong with the unit cell file. Here's a file from my previous work for a C3v frustrated monolayer (similar to NiI2). You can try to subtitute the values to your parameters . 

By the way, from previous works, the 2nd neighbor exchange in NiI2 monolayer is smaller than the 1st and 3rd neighbors. Also, for more precise simulations, the exchange anisotropy and Kitaev terms 

should be considered. 

[Loris Naudin]

Thank you for the help, yeah here the J2 value is smaller than 1st and 3st neighbors, but maybe something is wrong in the ucf yeah, because I've checked the parameters carefully. I've included single-ion anisotropy as well, and plan also to add the biquadratic interaction after that I get good results here

[Loris Naudin]

So I've tried using the new ucf made by idomeneo but I still don't obtain anything relevant, I don't understand, it's really strange that I don't obtain physically meaningful results while another MC code gives the expected answer

[Richard Evans]

The VAMPIRE MC solver is quite aggressive, but also the default equilibration temperature is 300K so that might introduce a lot of disorder in the system. However, what you seem to have is maze domains so with a more gentle equilibration, eg with zero-field cooling or from an angle away from the easy axis you might find a better groundstate. Also, the LLG solver will also relax more quickly for these kinds of systems.

Cheers,

Richard

-- 
You received this message because you are subscribed to the Google Groups "Vampire Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to vampire-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/vampire-users/4156ab25-f4e5-4b13-a112-b07d776a6c8an%40googlegroups.com.

[Loris Naudin]

Thanks for the tip, I will try that too, and update here if I find something relevant (or not)

Best regards,

Loris Naudin

[Loris Naudin]

So I've succeeded to obtain the spin spiral state (in-plane). I used the following parameters :

sim:equilibration-time-steps=200000
sim:time-steps-increment=1000
sim:total-time-steps=400000
sim:cooling-time=100 !ps
sim:cooling-function = gaussian
sim:minimum-temperature = 0.0
sim:maximum-temperature = 4.0
sim:equilibration-temperature = 4.0

#------------------------------------------
sim:program = field-cool
sim:integrator = monte-carlo

#===================================================
# Sample vampire material file V5
#===================================================
#---------------------------------------------------
material:num-materials=1
#---------------------------------------------------
# Material 1 Nickel Generic
# Interaction d'échange + anisotropie
#---------------------------------------------------
material[1]:material-name=Ni
material[1]:damping-constant=1
material[1]:exchange-matrix[1]= ...
material[1]:exchange-matrix[2]= ...
material[1]:exchange-matrix[3]= ...
material[1]:atomic-spin-moment=1.5 !muB
material[1]:material-element=Ni
material[1]:uniaxial-anisotropy-constant= ...
material[1]:initial-spin-direction=0,0,1

In the "..." you have to put DFT calculated exchange interactions. I've also used an ucf file to construct all the interactions using two non-equivalent atoms.

[Loris Naudin]

I have some troubles for getting the right Curie Temperature, everything goes down to almost 0 (in mean-magnetization) after 0K, it does not makes sense ! I used many steps in order to have a well-converged result