[Wien] local rotation matrix - followup question on triclinic systems

[Peter Blaha]

Have a look into SRC_nn/dirlat.f
BR2 defines the bravais matrix and gives the relation between carthesian 
coordinates and crystal axes. For P1 it is:
       BR2(1,1)=1.0d0*sin(gamma0)*sin(beta1)
       BR2(1,2)=1.0d0*cos(gamma0)*sin(beta1)
       BR2(1,3)=1.0d0*cos(beta1)
       BR2(2,1)=0.0d0
       BR2(2,2)=1.0d0*sin(alpha1)
       BR2(2,3)=1.0d0*cos(alpha1)
       BR2(3,1)=0.0d0
       BR2(3,2)=0.0d0
       BR2(3,3)=1.0d0
So the c-axis points into z.
Yes, this coordinate system is the same for all atoms.
Yes, for P1 symmetry you can use any local rotation matrix you like.

Cormac McGuinness schrieb:
> Dear all
> 
> I have a followup question regarding the local rotation matrix and global and
> local coordinates as applied to triclinic unit cells, for instance many organic
> molecular solids (e.g pentacene), though the same question could apply to
> complex metal oxides. (note the question is not specifally about pentacene)
> 
> In the case of a triclinic crystal where every inequivalent site has a
> pointgroup symmetry of 1, the calculation then gives the identity matrix as the
> local rotation matrix for each and every inequivalent site. Should one interpret
> the direction of the x, y, and z projected axes of the orbitals as being the
> same for all of the sites ?
> 
> If so then is the x direction parallel to the crystal axis a ?
> What then determines how y and z are oriented as the crystal axes a,b,c are not
> orthonormal ? How should these directions be interpreted in this case.
> 
> I ask as for planar pi-conjugated organic molecules the most important axis to
> distinguish is that perpendicular to the plane of the molecule. How can I know
> for sure which direction in terms of the physical crystal structure is
> associated with x, y and z in a triclinic system ?
> 
> 
> If I want to replace this definition of x,y,z, can I simply rerun with my own
> rotation matrix in place of each instance of the identity matrix, say to ensure
> that one of x,y,z is perpendicular to the plane of the molecule, in the manner
> described previously ?
> 
> Yours gratefully
> 
> Cormac
> 
> 
> Quoting Igor Djerdj <Igor.Djerdj at mat.ethz.ch>:
> 
>> Dear all,
>> I have a question concerning local rotation matrix. I have done and saved scf
>> calculation using default values for local rotation matrix. Now, I want to
>> change from global to local coordinate system to get dz2 and dx2y2 as bonding
>> orbitals pointed with their lobes into oxygen. I changed in struct file local
>> rotation matrix around V atom, and now the question is what should I do the
>> next in order to calculate DOS, should I run the programs according the order
>> in w2web interface or should I rerun again scf?
>> Regards,
>> Igor
>>
> 
> 
> --
> Dr. Cormac McGuinness, Lecturer, School of Physics, Trinity College Dublin.
> Phone: +353 1 8963547    Fax: +353 1 6711759    Office: SNIAM 3.17
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien