HOWTO change the Hartree-Fock exchange in hybrid ...

[kany]

Hi there,

Using google (search for post topic) I found a .pdf file on your server which explains how to modify the amnt of HF exchange in b3lyp

T

he keyword to change the amount of HF exchange is called b3-lyp_own.
In the control file the syntax looks like this:
$dft
functional b3-lyp_own 0.15 1.0
gridsize m3
In the output the formulars with the scaling factors are given then.
------------------
density functional
------------------
B3-LPY with your own HF weights
exchange: A * HF + (1-A) * LDA + B * 0.72*B88
original B3-LYP: A=0.2, B=1.0
correlation: 0.19*LDA(VWN) + 0.81*LYP
Found B3-LYP_own functional with :
Scaling factor for HF: A= 0.150000000000000
Scaling factor for Becke: B= 1.00000000000000
The above settings are for the 15% exchange.

Is this feature still implemented? My attempts (using Turbomole 6.5) were not successful so far, but turbmole seems to recognize "b3-lyp_own" and correctly interpret the new scaling factors.

           ------------------
           density functional
           ------------------
B3-LPY with your own HF weights
exchange:   A * HF + (1-A) * LDA + B * 0.72*B88
    original B3-LYP: A=0.2, B=1.0
correlation: 0.19*LDA(VWN) + 0.81*LYP
Found B3-LYP_own functional with :
  Scaling factor for HF: A=  0.150000000000000     
  Scaling factor for Becke: B=   1.00000000000000     
iterations will be done with small grid
 
spherical integration : Lebedev's spherical grid
spherical gridsize    :           3
    i.e. gridpoints    :         302
value for diffuse not defined
radial integration    : Chebyshev 2nd kind (scaling 3)
radial gridsize       :           3
integration cells     :          18
partition function    : becke
partition sharpness   :           3
 

          ------------------------
          nuclear repulsion energy  :   721.306886796   
          ------------------------

program stopped due to: functional name unknown

Does anyone know something about this feature?

Thanks
Mauro Schilling

[christof.haettig]

yes.