Page 1 of 1

### Zero Unit Cell Volume in Dispersion Calculation

Posted: Mon Dec 30, 2019 4:30 pm
Hi All,
I'm following the RF2 tutorial (https://docs.abinit.org/tutorial/rf2/) to calculate the dispersion in a tetragonal unit cell.

I've used the following unit cell parameters in other calculations without issue:
acell 5 6 7
scalecart 1 1 .9640
rprim 1 0 0
0 1 0
0 0 1

But in this run, I'm now getting the error: " Input rprim and acell gives vanishing unit cell volume.
This indicates linear dependency between primitive lattice vectors
Action: correct either rprim or acell in input file."

The primitive vectors are definitely not linearly dependent. I checked the source code abinit-8.10.2/src/41_geometry/m_geometry.F90, and it seems this error message is given if the absolute value of unit cell volume is less than some tolerance, which can't be right.

Has anyone ever resolved this issue?

Thanks!!

### Re: Zero Unit Cell Volume in Dispersion Calculation

Posted: Thu Jan 16, 2020 10:20 pm
Dear jerkov,
The problem comes from the flag scalecart you used. Check its definition (https://docs.abinit.org/variables/basic/#scalecart) and you'll see that its dimension is (3) while you give only one number.
I usually directly put the right units and lengths in acell (default Bohr or adding Angstrom at the end to declare it in Angstrom), mostly when you have a regular rprim 100 010 001.
Best wishes,
Eric

### Re: Zero Unit Cell Volume in Dispersion Calculation

Posted: Thu Feb 27, 2020 3:38 pm
Hi Eric,

I'm confused by what you mean, as I think I do have three numbers: "scalecart 1 1 .9640"

I've used this exact unit cell in other calculations, and it hasn't given me zero volume.

Thanks!

### Re: Zero Unit Cell Volume in Dispersion Calculation

Posted: Sun Mar 08, 2020 10:18 pm
Dear jerkov,
OK, sorry, I read it too quickly and saw 11.9640...
Then what you report is indeed strange, mostly if it works in other cases. What about putting directly:
acell 5 6 6.748
rprim 1 0 0
0 1 0
0 0 1

or:
acell 5 6 7
rprim 1 0 0
0 1 0
0 0 0.9640

Does it works?
Best wishes,
Eric