wrong fixed cell

[elena]

 Dear USPEX users,

 I am trying to perform a fixed cell calculation.
The lattice vectors I want to fix are,

%Latticevalues
3.2 -6.5 0
3.2 6.5 0
0 0 13.4
%Endvalues

however, the initial generated POSCARS do not have these lattice vectors, even if the volume of the cells are right.

Since the cell is very flat, i modify the min_Angle to 45, but it does not solve the problem.

Thank you in advance for your answer,

 Elena Redondo
 PhD. student.

[huy pham]

Hi Elena,
From the Latticevalues in the INPUT file, USPEX read it as 6 parameter for the cell:

a, b, c and angles: alpha, beta, gamma.

From these 6 parameter, USPEX generate random structures by "random_cell"

in which a, b, c are permuted and alpha, beta, gamma are permuted, too.

For this reason, you rarely get exactly the cell you gave at the beginning.

Now, question is: Do you really need the exactly lattice vectors at the beginning?

I don't think you need! With one crystal structure there any many ways to choose the lattice vectors.

So I think that you can continue your calculation.

Hope this helps,
Huy

--
You received this message because you are subscribed to the Google Groups "USPEX" group.
To unsubscribe from this group and stop receiving emails from it, send an email to uspex+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Pham Cong Huy
PhD Candidate,
Condensed Matter Section
SISSA International School for Advanced Studies
Via Bonomea 265, I- 34136, Trieste, Italy
E-mail:
pchu...@gmail.com
cp...@sissa.it

[elena]

Hi Huy,

I really want to fix these lattice parameters, since they are the experimental ones. I know that USPEX generates random cells, but in theory, it is also possible to fix
experimental lattice vectors. I have already done it with previous versions without any problem.
the cell is quite big, and i want to restrict the calculations......

thank you,

elena redondo

[Qiang Zhu]

Dear Elena,

You know in monoclinic systems, there are two ways to define the special axis, either b or c. In USPEX, we always define the axis along b. So the best way is to put a,b,c, alpha, beta, gamma here (instead of 3*3 matrix) if you want to do fixed lattice;

http://han.ess.sunysb.edu/uspex_manual/sect0016.html

You can also specify unit cell parameters just by listing a, b, c, , , and  values:

% Latticevalues 
10.1 8.4 12.5 90.0 101.3 90.0 
% Endvalues

And be sure you always put non-90 degree angle in beta.

Best regards,

Qiang

--
You received this message because you are subscribed to the Google Groups "USPEX" group.
To unsubscribe from this group and stop receiving emails from it, send an email to uspex+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Qiang Zhu (Research Assistant Professor)
ESS 365
Department of Geosciences
Stony Brook University
NY 11794-2100

http://uspex.stonybrook.edu/qzhu.html

[Elena Redondo]

Dear Qiang,

I have redefined the structure with the axis along b.

I have done both ways, with the 3x3 matrix or

specifying the lattice parameters a,b,c $\alpha$,$\beta$,$\gamma$, with a non-90 degree angle in beta.

But, the generated POSCARS even if they have the right volume, they do not have the right lattice parameters. For instance, there are no angles different from 90 or 120 degrees.

I have  also tried with the lattice parameters you were sending as an example,

% Latticevalues 
10.1 8.4 12.5 90.0 101.3 90.0 
% Endvalues

and there is the same problem, the generated POSCARS do not contain the right $\beta$ angle.

Best regards,

Elena Redondo

PhD. student

[Qiang Zhu]

It shouldn't be. Can you check the results1/non_optimized_structures? Are the initial lattice correct?

Did you allow the change of cell in the structure relaxation? 

Qiang

[Elena Redondo]

Dear Quiang,

I was already looking at the first generated POSCARS in the CalcFold1.....folders. Also, I am not allowing any relaxation of cell parameters, ISIF=2 in VASP.  The non-optimized structures correspond to the first generated POSCARS, so, i have checked,

but unfortunately, the cell parameters do not correspond to the input.

However, if  the %Latticevalues correspond to an orthorhombic or cubic cell (with all the angles =90), the generated POSCARs

are O.K. 

It seems to be related to  fixing a cell with an non-90(or 120) angle.

Best regards,

Elena Redondo

[Qiang Zhu]

How did you specify the space group?

Sent from my iPhone

[Elena Redondo]

Dear Quiang,

I was using

% symmetries
1-230
% endSymmetries

or

% symmetries
2-230
% endSymmetries

or

thanks,

Elena Redondo

[Qiang Zhu]

If you are targeting the monoclinic cell, you need to specify it within those monoclinic space groups.

Qiang