[Pw_forum] The choice of Acoustic Sum Rule for crystal

[e kb]

Dear Baoling,
  I am not an experienced user,
  but I want to comment a little on your case.
   
  In these two position configurations, the energy difference is not as small as you think, I guess, sth like 90 meV. (why to compare in percentages??) so 0.792 is not the global minimum.
  If I were you, I would find my own equilibrium position configuration first. At least you have a good initial guess, 0.7905, which is found by using the same pseudopotential. You can search around it and find yours - which i guess would not be so far away from 0.7905. Then at your result, make a "converged" phonon calculation, using several asr schemes if you want, and later you can post your mail to the group with the title 'The choice of ASR'. But for now, I am not sure that this is your problem.
   
  Again, you dont have to take this seriously, I am just a newbie.
  Emine Kucukbenli

naivebamboo Huang <naivebamboo at hotmail.com> wrote:
  Dear Matteo,
Thanks very much. Yes, they are fractional coordinates of one Te atom 
(Te1 atom), that is, atomic position in the unit cell. I am sorry for the 
confusion. What I mean is that the energies of the two positions seems to 
change little comparing to their position change. The lattice constant is 
10.45 angstrom, and so the difference of the position is not small. I think 
this is due to the weak Te1-Te1 interaction (van der Waals interaction). I 
tried to relax the structure use vc-relaxation, but meet an error during the 
optimization. I am trying to work it out now. The configure with x=0.7905 is 
the relaxed structure that other people get use the same pseudopotential. 
The frequencies I got from PWscf are attached.
You can see mode 4 of the two configures are much different. Also, if 
ASR changes from 'simple' to 'crystal', 10% difference appears (for all the 
modes). Is this acceptable? Since the frequencies are all positive, I just 
guess they are at least local minimal configures. The configure with 
x=0.792, though energy is higher, reproduces the measured modes best. I 
don't know which result I should trust.

Thanks

Baoling

x=0.7905 total energy =-382.4400695 Ry
x=0.792 total energy =-382.43346308 Ry (however, x=0.792 reproduces 
the measured bond lengths of Te-Bi, Te-Te )

x=0.7905 ( ASR='crystal')
# mode [cm-1] [THz] IR
1 0.00 0.0000 0.0000
2 0.00 0.0000 0.0000
3 0.00 0.0000 0.0000
4 11.81 0.3541 0.0000
5 79.89 2.3950 0.0000
6 79.89 2.3950 0.0000
7 84.52 2.5339 0.0000
8 84.52 2.5339 0.0000
9 101.25 3.0353 0.0000
10 128.73 3.8593 0.0000
11 142.63 4.2760 0.0000
12 150.50 4.5118 0.0000
13 150.50 4.5118 0.0000
14 153.94 4.6148 0.0000
15 153.94 4.6148 0.0000

x=0.792 (ASR='crystal')
# mode [cm-1] [THz] IR
1 0.00 0.0000 0.0000
2 0.00 0.0000 0.0000
3 0.00 0.0000 0.0000
4 36.49 1.0941 0.0000
5 75.48 2.2627 0.0000
6 75.48 2.2627 0.0000
7 83.00 2.4884 0.0000
8 83.00 2.4884 0.0000
9 100.00 2.9978 0.0000
10 111.66 3.3476 0.0000
11 138.25 4.1446 0.0000
12 141.25 4.2346 0.0000
13 141.25 4.2346 0.0000
14 167.48 5.0210 0.0000
15 167.48 5.0210 0.0000

x=0.792 (ASR='simple')
# mode [cm-1] [THz] IR
1 0.00 0.0000 0.0000
2 0.00 0.0000 0.0000
3 0.00 0.0000 0.0000
4 53.43 1.6017 0.0000
5 56.12 1.6825 0.0000
6 56.12 1.6825 0.0000
7 69.99 2.0982 0.0000
8 69.99 2.0982 0.0000
9 98.14 2.9421 0.0000
10 98.14 2.9421 0.0000
11 113.40 3.3997 0.0000
12 113.40 3.3997 0.0000
13 114.78 3.4411 0.0000
14 131.36 3.9381 0.0000
15 145.97 4.3760 0.0000

 
---------------------------------
Access over 1 million songs - Yahoo! Music Unlimited.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/attachments/20070116/ca605b20/attachment.htm