[Wien] [second time query] overestimated band gap by PBE

[Peter Blaha]

The total energy is not variational with respect to the number of 
k-points, i.e. a lower total energy does not mean a better calculation 
when you compare 2 k-meshes. Clearly the 1000k calculation should be better.

PS: The size of the basis set is variational: a larger RKmax has to give 
you a lower energy.

PPS: As mentioned before, the GAP without spin-orbit is a meaningless 
number and you should NEVER say that the calculation which gives for the 
wrong reason the experimental gap is the "better" calculation.

Am 23.11.2016 um 13:15 schrieb Dr. K. C. Bhamu:
> Dear Prof Peter,
>
> Please comment of my query when you get time:This is in the queue on my
> previous query regarding CH3NH3PbI3:
>
>
>     I used two strategy and got different results:
>
>
>     The experimental band gap is 1.67 eV for orthorhombic lead halide
>     perovskite.
>
>     I ran two cases (with PBE);
>     1. with 1000 k, div: ( 11  7 11) and "min -j 'run_lapw -p -I -NI -i
>     120 *-ec *0.00001 *-ec * 0.0001 -fc 1
>     2. with 400 k, div: (  8  5  8)   and   min -j 'run_lapw -p -I -NI
>     -i 120 *-ec * 0.0001 *-ec * 0.0005 -fc 2'
>
> -ec switch is used twice instead of ec and cc (it is by mistake)
> In case 2 with less k-points and normal scf criteria (-ec 0.0001 -ec
> 0.0005 -fc 2), I got minimized ENE than case 1. But the band gap is good
> from case 1.
>
>
> Results:
>
> 1. GAP 1.687 eV,    FER: 0.1030934295, ENE: -339059.11079128
> 2. GAP  1.777 eV,   FER: 0.1004906191, ENE: -339059.12432403
>
> You see that band gap is in the error of /+- 0.01eV.
>
> *[* next i tried with
>
> 1. with 1000 k, div: ( 11  7 11) and "min -j 'run_lapw -p -I -NI -i 120
> -ec 0.00001 _-cc_ 0.0001 -fc 1   >> second -ec is replaced by -cc
> 2. with 400 k, div: (  8  5  8)   and   min -j 'run_lapw -p -I -NI -i
> 120 -ec 0.0001 _-cc _ 0.0005 -fc 2'      ***>> second -ec is replaced by
> -cc*
> *
> *and nothing happened scf immediately got exit and the structure was
> already optimized]*
>
>
>
>     I did nothing special, just reduced rmt (5% and further rmt of "I"
>     was reduced by 0.02).
>     /_Please correct me what is wrong here. Because usually PBE always
>     underestimates the experimental band gap._/
>
>     Kind regards
>     Bhamu
>
>
>
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:  http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>

-- 
Peter Blaha
Inst.Materials Chemistry
TU Vienna
Getreidemarkt 9
A-1060 Vienna
Austria
+43-1-5880115671