Is it possible to specifiy the multiplicity of a metallic (or bimetallic) system AND include the Fermi keyword?

[Natalie Austin]

Hello,

I added FIXED_MAGNETIC_MOMENT -1  to the SMEAR section for the multiplicity 
4 calculation but the alpha and beta orbitals were still fractional values: 
303.548262   301.451738

In the Non-integer total number of Alpha/Beta Electrons. 
<https://groups.google.com/forum/#%21searchin/cp2k/fixed_magnetic_moment/cp2k/zXcQmb2xkks/E1Qhvsmne_UJ> 
post someone stated that "One uses the smearing for metallic systems, and 
the Fermi energy of the two spin channels has to be the same.
This is not the case if the electrons are not free to fill the lowest 
energy states, irrespective of the spin"

Does this mean by restricting the alpha and beta to integer numbers my 
system could be (or is always) higher in energy than a system where 
fractional occupations are allowed?

Generally, I'm a bit unsure about the standard for performing spin 
polarized calculations in CP2K, do you only need to set UKS TRUE? is 
setting the multiplicity just as important? 

Thanks,

Natalie


On Monday, November 2, 2015 at 5:06:53 PM UTC-5, Matt W wrote:
>
> Hi Natalie,
>
> sounds like FIXED_MAGNETIC_MOMENT in the SMEAR section is what you want.
>
> Matt
>
> On Monday, November 2, 2015 at 7:30:12 PM UTC, Natalie Austin wrote:
>>
>> Hello,
>>
>> I've been comparing the output files for the geometry optimization of 
>> Cu55 at different multiplicities (M= 2, 4, 6,8). What I noticed is that 
>> although I specify multiplicity the atomic population from Mulliken 
>> Population Analysis is not representative of the multiplicity that I 
>> included. 
>>
>> Specifically,
>>  
>> for the doublet case i specified 303 for alpha and 302 for beta
>> and in the muliken analysis the values were 302.499984   302.500016
>> ive included the values for the multiplicity 4, 6, and 8 also below
>>
>> quartet
>> 304   301
>>
>> 303.546324   301.453676
>>
>> sextet
>>
>> 305  300
>> 303.548877   301.451123
>>
>> octet
>> 306    299
>>
>> 303.550009   301.449991
>>
>> So it seems that the multiplicity isn't reflected in the final result. I'm 
>> starting to believe that this might be attributed to the fermi keyword 
>> which I read in another thread does not constrain the multiplicity. If this 
>> is the case, is it not possible to set the multiplicity in metallic systems 
>> or are there other ways to converge metallic systems without using the 
>> fermi keyword? 
>>
>> I've included my input and output files for the doublet case below.
>>
>> Thanks, 
>>
>> Natalie
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20151103/25dd91ec/attachment.htm>