[Pw_forum] the procedure of calculating the Raman tensor using the finite electric field method

[li niu]

Dear PaoloU,
   
  Thanks for your help. I now still have some puzzlements about the procedure of 
  calculating vibrationl spectrum by finite electric field method.
   
  1.I consider a model structure of tetrahedral amorphous carbon obtained by my schoolmate using a Car-Parrinello molecular dynamics approach.    
   
  2.When running a self-consistent calculation for this structure (using different exchange-crrelation energy and pseudopotential with in CPMD), I found the forces and stress are big.
In this case, I should use pw.x to compute the equilibrium geometry 
before running finite electric field method to calculate Raman tensor. Is it right?
 
   Total force =     0.633437     Total SCF correction =     0.000080
     total   stress  (Ry/bohr**3)                   (kbar)     P=  -69.10
  -0.00087420   0.00095392  -0.00017791       -128.60    140.33    -26.17
   0.00095392   0.00033110   0.00005680        140.33     48.71      8.35
  -0.00017791   0.00005680  -0.00086609        -26.17      8.35   -127.41
   
  3.I run a vc-relax calculation and input and output files are as follows.
It is obvious the cell angle parameters have been changed and the structure isn't the cubic structure. 
   
  (1)ATOMIC_POSITIONS (angstrom)
&CONTROL
  calculation  = "vc-relax",
  prefix       = "ta-c",
  pseudo_dir   = "/home/niuli/espresso-3.2/pseudo",
  outdir       = "/home/niuli/tmp",
  etot_conv_thr=1.0D-4,
  forc_conv_thr=1.0D-5
  tstress = .true. ,
  tprnfor = .true. ,
                       nstep = 65 ,
                       dt = 50 ,
/
&SYSTEM
    ibrav = 1, 
    celldm(1) = 13.9103, 
    nat= 64, ntyp= 1,
    ecutwfc =55    
/
&ELECTRONS
  conv_thr    = 1.D-9,
  mixing_beta = 0.7D0,
/
&IONS
  
/
&CELL
               cell_dynamics = 'damp-w' ,
                      press = 0.0 ,
/
ATOMIC_SPECIES
 C   12.0107  C.pz-mt.UPF
  ATOMIC_POSITIONS (angstrom)
 C      0.335930     -0.353760     -0.911160
 C      1.410315     -0.756669      3.705328
 C     -0.462432      3.049465     -0.168292
  ...
K_POINTS automatic
1 1 1 0 0 0
----------------------------------------------------------
final unit-cell volume =   2649.6086 (a.u.)^3
  input alat =      13.9103 (a.u.)
  CELL_PARAMETERS (alat)
   0.988103212   0.021498496  -0.002655708
   0.021493430   1.013103454  -0.000556700
  -0.002655491  -0.000556401   0.983829644
  ATOMIC_POSITIONS (angstrom)
C        0.292559767  -0.288781798  -0.891844186
C        1.355532741  -0.743068572   3.628414295
C       -0.355484435   3.081913445  -0.167170533
C        0.095087835   3.109571795   3.835737135
C        4.156057825   0.859970231  -0.191362362
...
-------------------------------------------------------------
(2)ATOMIC_POSITIONS (crystal)
  
CELL_PARAMETERS (alat)
   0.997839246   0.002690608  -0.000417664
   0.002690608   1.001016967   0.000136231
  -0.000417664   0.000136231   0.997967842
 
ATOMIC_POSITIONS (crystal)
C        0.040549747  -0.039126226  -0.123246079
C        0.194302645  -0.104600063   0.504317492
C       -0.056833511   0.414314558  -0.024646663
C        0.005446848   0.417400176   0.531096601
C        0.569734741   0.101952019  -0.026079258
C        0.849788292   0.163379218   0.703668015
C        0.572774545   0.583287732   0.029202254
...
------------------------------------------------------------
  4.If I set ibrav=0 and use the obtained new cell parameter and atomic positions in the next scf calculation, the results on forces and stress are the same with that in vc-relax.
But if I set ibrav=1 in the next scf calculation, how to read the new lattice constant?
 
5. Then run ph.x code to obtain the phonon frequencies and eigenmodes. The phonon density of states is also obtained by dynmat.x code.
In your papers,the dynamical matrix is calculated by taking finite differences of atomic forces for atomic displacements of a constant bohr.What is the difference between two methods?
 Why don't you use the linear response method(ph.x code) to calculate the vibrational properties? 
   
  6. Supposing I obtain the equilibrium geometry of ibrav= 1, I then can calculate Raman tensor by considering second derivatives of the atomic forces(cp.x code).
 
Is my thought right on the procedure of calculating?
 
Best wishes,
   
  Niu Li
Harbin Institute of Technology
China


       
---------------------------------
抢注雅虎免费邮箱3.5G容量，20M附件！ 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20070531/5d07a2e3/attachment.html>