[Wien] Mini NEWT and/or PORT - what precision can one expect?

[Torsten Andersen]

Dear Peter, Laurence, and all the other experts on this list,

I have for some time been trying to stabilize the 4th digit of the 
interlayer distance in simple metal films, using min_lapw, first with 
NEWT and then with PORT, without success.

I take films of growing thickness, and expect that for a certain 
thickness the contractions or expansions of surface layers should be 
constant (the difference between centre-layers in a 15ML film is roughly 
the same as in a 17ML film, with variations of less than 0.1% from the 
bulk value).

It is easy in both cases to get the 3rd digit reasonably stabilized, 
that is, whether the distance between the two outermost layers is 1.81 
or 1.82 Aangströms. This I can achieve with both "NEWT 1.0" and with 
"PORT 1.0" in case.inM. However, I would have liked to get an answer to 
whether 1.82 is 1.816 or 1.824.

I thought this would be straight-forward by changing to "NEWT 0.1" or 
"PORT 0.1" in case.inM. However, in NEWT I end up with the 
"stp.le.tolx", although there is plenty of space between the MT's, and 
in PORT I run into the "curvature condition failed".

I have tried many different tricks, but nothing seems to work, and the 
convergence never get better (after exiting with one of the two errors 
above) than what I get with "NEWT 1.0" or "PORT 1.0". When I show the 
numbers to some colleagues, they say that this is the convergence I can 
expect.

Am I expecting too much?

Best regards,
Torsten Andersen.


-- 
Dr. Torsten Andersen        TA-web: http://deep.at/myspace/
AG Hübner, Department of Physics, Kaiserslautern University
http://cmt.physik.uni-kl.de    http://www.physik.uni-kl.de/