The ABINIT Users Mailing List ( CLOSED )

[matthieu verstraete <mjv500@york.ac.uk>]

Hmmm,

for qpt 28 everything is 0

apart from that your values look more or less ok. Lambda is quite small, 
and the linewidths too. The value of mustar, however, is fit for Al. 
Most people take 0.12 or so, but in your case it's probably very different:

The McMillan formula blows up when mu and lambda are comparable (check 
the denominator), which is what is happening here. Try the formula for 
lambda=1 (stronger coupling!) and you get Tc=10e-5 Ha. Try it for mu=0.5 
and you get Tc=1.e-3.

Your system is exotic with low frequencies and high pressures, so you 
should be more careful when applying all these approximations (well, you 
shouldn't apply them at all).

Try reading the papers by EKU Gross and Andrea Floris on Li under 
pressure. They do not use an external mu and may have some insight to 
the strength of the Coulomb interaction here.

Otherwise, start at 0 pressure and fit mu to get the correct Tc (if 
there is one at 0GPa). This still won't give you a value for high 
pressure, but it's a start. mu may be either weaker or (probably) 
stronger at high pressure, and in either case the McMillan formula will 
give more realistic figures, but its applicability still remains very 
dubious. I think lambda should be fine (and omegalog too).



Matthieu