My Community

Upon application of the PW91 XC-functional I have noticed NaN being printed out during SCF calculation of a phosphorus atom in a box. I have pasted my INCAR file below.
SYSTEM = P atom in a box
PREC = Accurate
ENCUT = 275
ISMEAR = 0 ! Gaussian smearing
SIGMA = 0.1
GGA = 91
VOSKOWN = 1
#ISPIN = 2 ! spin polarized calculation
#NUPDOWN = 3
ALGO = Normal
ISIF = 2
IBRION = 1
POTIM = 0.1
NSW = 50
#IALGO = 3
EDIFF = 0.0001
EDIFFG = -0.01

Also I would like to provide a portion of the OUTCAR with this NaN.

Free energy of the ion-electron system (eV)
---------------------------------------------------
alpha Z PSCENC = 0.13925009
Ewald energy TEWEN = -56.25115085
-1/2 Hartree DENC = -86.97857932
-exchange EXHF = 0.00000000
-V(xc)+E(xc) XCENC = 1.54055214
PAW double counting = 139.98142999 NaN
entropy T*S EENTRO = -0.16925255
eigenvalues EBANDS = -43.04881960
atomic energy EATOM = NaN
---------------------------------------------------
free energy TOTEN = NaN eV

energy without entropy = NaN energy(sigma->0) = NaN

Could someone explain why this is occurring? I have noticed this only with the PW91 functional. I am using PBE_GGA pseudopotentials.

this error only shows up if GGA=91 is set and one of the new PPs (released in Dec 2012) is used. It will be fixed in the upcoming release of vasp.5.3.