I am performing an Unrestrticted Kohn Sham (UKS) DFT simulations using cp2k
In my system of 289 atoms, I have 1 Fe atom.
I would like to perform the DFT calculations for different spin states of Fe by varying the multiplicity and find out the ground state spin
Fe has 4s23d6 and for a +2 oxidation state, Multiplicity=2S+1=2*(4/2)+1 =5
Fe can exists in +2,+3,+4,+5,+6 oxidation states and so corresponding multiplicity is (5, 6, 5,4,1) respectively
But when I input 5 for multiplicity, the program throws an error
" LSD: try to use a different multiplicity "
When UKS is enabled, it requests a spin polarised calculation using alpha and beta orbitals.
When I did not give any multiplicity, the code provides
Spin 1
Number of electrons: 773
Spin 2
Number of electrons: 772
which indicates multiplicity =2. Is it the ground state spin state?
Can somebody help me in circumventing this error?
Thank you
Is there an equivalent in CP2K for the "MAGMOM" input parameter in VASP? We generally assign metal atoms like Fe to the highest magnetic moment (or number of unpaired electrons) possible, and then let VASP figure out what the final spin state is. That generally works. In some cases, we verify the spin multiplicity of the ground state using G16 by single point energy calculations in the vicinity of the VASP-predicted value to see which one gives us the lowest energy. Fore example, for a cluster of 16 Co atoms (3 unpaired electrons each), we specified Magmom (= 2S) = 48 (16 x 3), and VASP converged to 2S = 34, i.e., 2S+1 = 35. We did a Gaussian calculation for 2S+1 = 33, 35, and 37 and confirmed that 35 is the lowest energy state. Of course, that does not rule out that a smaller 2S+1 may be even lower in energy but that seems unlikely from our experience.
Thank you prof. B. R. Ramachandran for your insights.
I had only 1 Fe atom and so I did something similar like what you have mentioned for different oxidation states of Fe as Fe is coordinated with other elements in the system ( In my system in addition to Fe, I also have Si, O, Al). But whenever I mentioned values like 1,5, the program threw the above mentioned error.
This is definition of multiplicity in cp2k manual
MULTIPLICITY {Integer}Two times the total spin plus one. Specify 3 for a triplet, 4 for a quartet, and so on. Default is 1 (singlet) for an even number and 2 (doublet) for an odd number of electrons.
So cp2k is counting the total number of electrons rather than the number of unpaired electrons only on the transition metal. So In my case for 1545 total number of electrons, even multiplicity (2,4,6) is not throwing errors.
Does that mean oxidation states corresponding to odd multiplicity is not possible for Fe in my system? For eg. when Fe is in +2 oxidation state, unpaired electron = 4, multiplicity=2s+1=5. this state is not possible for my system.
Usually, even number of electrons leads to an odd number spin multiplicity (2S+1) and odd number of electrons to an even value. Since you are dealing with 1545 electrons in the neutral, the +2 state will have 1543 electrons. Knowing nothing else about the system, even numbers for 2S+1 seem to be OK.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Atomic, Molecular and Optical Physics
- Atoms