[Pw_forum] NEB: Huge error for first image
Maxim Skripnik
maxim.skripnik at uni-konstanz.de
Fri Nov 3 15:07:24 CET 2017
Dear developers,
I perform NEB calculations with an electric field (eamp != 0). The first
and last images are relaxed with pw.x with a specific value for eamp.
The relaxed xyz coordinates are then used for the neb calculation.
However, the error for the first image is really high (above 25 eV/A).
In some cases it even doesn't fit in the reserved decimal places (see
below). Is this an issue which has to be dealt with or can one neglect
this? I don't understand why the error is so big while the structure has
been properly relaxed.
The NEB calculation converges to reasonable activation energies, even
though the error of the first images stays at a high value. I just want
to ensure that this is not an issue. Maybe one of the developers can
explain why the error is so big?
Best regrads
Maxim Skripnik
Univ. Konstanz
------------------------------ iteration 2
------------------------------
tcpu = 3835.6 self-consistency for image 2
Info: using nr1, nr2, nr3 values from input
Info: using nr1, nr2, nr3 values from input
tcpu = 4152.5 self-consistency for image 3
Info: using nr1, nr2, nr3 values from input
Info: using nr1, nr2, nr3 values from input
tcpu = 4510.8 self-consistency for image 4
Info: using nr1, nr2, nr3 values from input
Info: using nr1, nr2, nr3 values from input
tcpu = 4856.1 self-consistency for image 5
Info: using nr1, nr2, nr3 values from input
Info: using nr1, nr2, nr3 values from input
tcpu = 5176.3 self-consistency for image 6
Info: using nr1, nr2, nr3 values from input
Info: using nr1, nr2, nr3 values from input
activation energy (->) = 0.776966 eV
activation energy (<-) = 0.435476 eV
image energy (eV) error (eV/A) frozen
1 -27262.4585336 ********** T
2 -27262.3382765 2.780244 F
3 -27262.1422089 1.372929 F
4 -27261.8032661 1.653251 F
5 -27261.6815672 1.336730 F
6 -27261.9419108 0.967285 F
7 -27262.1170431 0.002386 T
path length = 4.553 bohr
inter-image distance = 0.759 bohr
More information about the users
mailing list