[Wien] MgO Ghostbands

[thomas.mazet]

> For your problem: Usually you can forget a QTL-B value of 1. These are
> NOT ghostbands, but slightly enhanced values for the u-dot contribution.
> I would start worrying for values of 5 and above.
> 
> It is most likely because of some shallow semicore states (Mg 2p or O 2s)
> and it depends on RMT whether or not an E-TOP is found or not. If not, it
> keeps the linearization energy at the defaul value and this might not be
> optimal.
> 
> You may try:   run_lapw -in1new 1
> 
> But please note: after that ALWAYS use the -in1new switch (or change back
> explicitly with -in1old

Sorry I'm not sure to well understand.
 What do you mean precisely by "ALWAYS use the -in1new switch" ?
Regards

Thomas

-- 
Thomas MAZET
Laboratoire de Chimie du Solide Minéral (UMR 7555)
Université Henri Poincaré-Nancy I
BP 239 54506 Vandoeuvre les Nancy Cedex
FRANCE
Tel.: +33-(0)3-83-68-46-77
Fax : +33-(0)3-83-68-46-11