Can direct coordinates be nagative? CONTCAR shows it

[simha]

Dear Vasp Users

I am relaxing internal coordinates of a system. My POSCAR file contains the DIRECT coordinates (between 0 and 1). After certain steps of ionic relaxation the CONTCAR is written. This CONTCAR shows negative coordinates as well like this

Staring coordinates

0.00106166 0.00106166 0.35832663 T T T
0.99893834 0.99893834 0.64167337 T T T

relaxed ones
0.0005738272305676 0.0005738272305676 0.3566378879904804 T T T
-0.0005738272305677 -0.0005738272305677 0.6433621120095198 T T T


I did not expect the negative sign, as direct coordinates are always between 0 and 1.

But then I continued without editing my CONTCAR file , I reached the structural minima.

My question is am I doing something wrong by not editing the CONTCAR file or is it acceptable?

Ofcourse I notice the fact that 0.001 and 0.999 can be related by translational symmetry.

I would be thankful for any suggestion or experiences in this matter.

Regards
Narasimham

[admin]

of course they can, if the atom moves beyond the respective boundary of the cell and enters on the other side (the positions are equivalent because of the translational symmetry, you are correct).