inpaired electrons but no spin density in the properties

[jbaltrus]

I am performing ground state spin unpaired triplet ridft calculations and plotting properties with $pointval fmt=cub. I get td and sd files, but sd doesn't contain any plot (that is the file is populated but no net spin gets visualized), but in my calculations I clearly see that there are two unpaired electrons, so I should see some spin density since they are not paired. These two things contradict each other but I can't quite understand what's going on

 alpha:

    irrep                422a        423a        424a        425a        426a   
 eigenvalues H         -0.21785    -0.21165    -0.20815    -0.20067    -0.19771
            eV          -5.9282     -5.7594     -5.6640     -5.4606     -5.3800
 occupation              1.0000      1.0000      1.0000      1.0000      1.0000

    irrep                427a        428a        429a        430a        431a   
 eigenvalues H         -0.19570    -0.18398    -0.18340    -0.18204    -0.17661
            eV          -5.3253     -5.0064     -4.9905     -4.9536     -4.8057
 
 beta: 

    irrep                424a        425a        426a        427a        428a   
 eigenvalues H         -0.20942    -0.20175    -0.19905    -0.19844    -0.18620
            eV          -5.6987     -5.4899     -5.4165     -5.3999     -5.0667
 occupation              1.0000      1.0000      1.0000      1.0000      1.0000

    irrep                429a        430a        431a        432a        433a   
 eigenvalues H         -0.18499    -0.18310    -0.17785    -0.17710    -0.16507
            eV          -5.0339     -4.9824     -4.8395     -4.8192     -4.4918
 

[antti_karttunen]

What does dscf/ridft report if you carry out a -proper run with
$pointval integrate?
And what isovalue did you use for the visualization?

Antti

[jbaltrus]

I used any value, no luck. Total density is plotted ok. I will check the keyword you recommended and get back

[jbaltrus]

well, it generates exactly the same size sd.cub file as previously which doesn't display any spin density at any isovalue

[antti_karttunen]

OK, but what does the output say in the very ending? It should integrate the spin density over the whole box and report it.

Antti

[jbaltrus]

not sure what that means

 integrated electron number(s):
                   -853.4042877      1.9766234
absolute values:   -853.4042877     -2.0566246

[antti_karttunen]

The number on the left is the total number of electrons (probably a bit off from the correct since the default pointval-grid is not accurate enough). The number on the right is the number of the unpaired electrons. So, there are about two unpaired electrons. But why you are not seeing the density? I have no idea...

Antti

[jbaltrus]

*panic mode*

[antti_karttunen]

Panic? Did you also try negative isovalues for plotting the spin density (since you have more beta than alpha electrons=? There is one strange thing here, though: Why is the integrated electron number negative? For me it's always positive (at least for the systems I've tested). The integrated spin density (first row) is negative for my tests, if b > a. But for you the absolute value is negative? What Turbomole version are you using?

Antti

[jbaltrus]

6.3

I also noticed that in my SCF output electron number is -2

ENERGY CONVERGED !


 ITERATION  ENERGY          1e-ENERGY        2e-ENERGY     NORM[dD(SAO)]  TOL
   3  -6124.1363485995    -78383.317633     38285.901959    0.000D+00 0.212D-10
                            Exc = -607.709306855     Coul =  38893.6112657   
                         Ntotal = 854.00078377      Nspin =-2.0000048134   
                             Na = 426.00038948         Nb = 428.00039429   
                            current damping =19.800
 
          Norm of current diis error: 0.10859E-01
          max. resid. norm for Fia-block=  6.748D-04 for orbital      2a    beta
          max. resid. fock norm         =  7.022D-02 for orbital   1432a    beta

 End of SCF iterations

[antti_karttunen]

And what about plotting spin with negative isovalue? Still nothing?

Antti

[jbaltrus]

that did the trick. not sure why I would have -2 unpaired electrons or how that might have switched to negative, but it worked, thanks

[antti_karttunen]

OK, I'm glad to hear it worked. The integrated numbers are indeed a bit funny, but everything else now makes sense since this a "spin_b > spin_a" case.

Antti