[Pw_forum] symmetries

[Stefano de Gironcoli]

Elwira wrote:

> But I'm afraid I still don't quite understand  how it works.
>
> In example 2 when bulk Si is introduced there is no ltaucry=.true. but
> the atom positions are in lattice vectors. Why is it so?
>

NO
in example2 the positions are in cartesian coordinates in units of celldm(1)
(also called a_0 in the output) ...

It happens by chance (or misfortune if you prefer) that 1/4 1/4 1/4 are also
the
crystal coordinates for that particular atom, but without specifing
ltaucry=.true.
the code assumes that the coordinates are cartesian in unit of celldm(1)

> This time I used ibrav=0, ltaucry=.true. and lattice vectors and atomic
> postions as follows:
> 1.0 0.0 0.0
> 0.0 0.8660254 0.0
> 0.0 0.0 1.0
> hexagonal
> 0.00 0.00 0.00 1
> 0.33333333333333333 0.33333333333333333 0.50 1
>
> In the output file the crystal is described by 12 symmetry operations
> which is ok, but the atomic positions are as follows:
>
> site n.     atom                  positions (a_0 units)
>           1        Cu     tau( 1) = (  0.0000000  0.0000000  0.0000000  )
>           2        Cu     tau( 2) = (  0.3333333  0.2886751  0.5000000  )
>
> I don't understand why the second coordinate of second atom was changed?

the program has scaled the coordinates in the format used internally that is
the cartesian one ... if you had used a c/a ratio not equal to one also the
third
component would have changed.

Iif you specify in the namelist iprint=1 the output is more verbose and show
both
cartesian and crystal coordinates in the output

regards, stefano