trouble with M06-2X
Clicked A Few Times
3:22:32 AM PDT - Mon, Aug 1st 2016
Hi,

I'm trying to calculate raman intensities using XC M06-2X. Optimization works fine, but when it starts calculating the polarizability for raman calculations i get this error:

  xm06: d2 not coded                    0

This is my input file:

echo

start G1_m06-2x

title "G1_m06-2x"
charge 0

geometry units angstroms print xyz autosym
   C        1.15429        1.60549        0.33593
   C        1.33587        0.26389        0.00855
   C        0.29764       -0.50396       -0.53404
   C       -0.93893        0.11684       -0.74543
   C       -1.14265        1.45874       -0.42483
   C       -0.08826        2.17724        0.11269
   H        2.30821       -0.18965        0.17290
   H        1.95639        2.20417        0.74989
   H       -1.75642       -0.45627       -1.16903
   H       -2.09691        1.94373       -0.59027
   F       -0.27459        3.48482        0.42565
   C        0.50201       -1.97015       -0.85071
   C        0.33956       -2.89194        0.37051
   H        1.50643       -2.11782       -1.26124
   H       -0.21283       -2.28534       -1.61570
   N       -1.03310       -2.85502        0.88302
   H        1.09740       -2.62243        1.12364
   H        0.55104       -3.91908        0.05779
   H       -1.24620       -1.94618        1.28220
   H       -1.16528       -3.54347        1.61572
end

basis
  * library 6-311++G**
end

driver
  tight
  maxiter 100
end

vib
  temp 1 50.0
end

property
 response 1 0.0856335
end

dft
  xc m06-2x
  grid xfine
  iterations 200
  direct
  mult 1
end

raman
  normal
  width 0.5
end

task dft optimize
task dft raman

Forum Regular
6:02:33 AM PDT - Mon, Aug 1st 2016
Calculating the polarizability requires second derivatives of the exchange-correlation functional, and as stated in the documentation, second derivatives of the Minnesota functionals have not been implemented. The new (and still listed as experimental) NWXC module (http://nwchemgit.github.io/index.php/Release66:Density_Functional_Theory_for_Molecules#NWXC_...) potentially should be able to calculate those second derivatives through its automatic differentiation. Though as I stated, that module is still listed as experimental so check results with extra care (if it works at all).

Clicked A Few Times
6:44:08 AM PDT - Mon, Aug 1st 2016
Quote:Sean Aug 1st 6:02 am
Calculating the polarizability requires second derivatives of the exchange-correlation functional, and as stated in the documentation, second derivatives of the Minnesota functionals have not been implemented. The new (and still listed as experimental) NWXC module (http://nwchemgit.github.io/index.php/Release66:Density_Functional_Theory_for_Molecules#NWXC_...) potentially should be able to calculate those second derivatives through its automatic differentiation. Though as I stated, that module is still listed as experimental so check results with extra care (if it works at all).


thanks for your answer, I thought it was something like that.

when I use:

 XC new m06-2x

it says i dont have nwxc_input in my build, how do I install that module? I thought it's supposed to be in release 6.6 already (am running 6.6).

Forum Regular
11:04:34 AM PDT - Mon, Aug 1st 2016
You should be able to just add nwxc to your NWCHEM_MODULES variable and then recompile (including make nwchem_config step). For example, NWCHEM_MODULES="all nwxc"

Forum Vet
5:54:07 PM PDT - Mon, Aug 1st 2016
Unfortunately,the second derivative part of the nwxc module might not be ready for use in NWChem

Clicked A Few Times
11:49:24 PM PDT - Mon, Aug 1st 2016
Quote:Sean Aug 1st 11:04 am
You should be able to just add nwxc to your NWCHEM_MODULES variable and then recompile (including make nwchem_config step). For example, NWCHEM_MODULES="all nwxc"

thank you, Sean

I'll try this and see if it works

Quote:Edoapra Aug 1st 5:54 pm
Unfortunately,the second derivative part of the nwxc module might not be ready for use in NWChem

what does might not be ready mean? unstable? unreliable result?

In any case, I will try this and see what happens.

Thank you for responding,
Afik

Forum Vet
8:57:41 AM PDT - Tue, Aug 2nd 2016
Quote:Faklerus Aug 1st 10:49 pm

Quote:Edoapra Aug 1st 5:54 pm
Unfortunately,the second derivative part of the nwxc module might not be ready for use in NWChem

what does might not be ready mean? unstable? unreliable result?

In any case, I will try this and see what happens.

Thank you for responding,
Afik


I meant to say: you might not be able to get the code producing the 2nd derivatives for m06-2x working (i.e. not even being able to produce a single result)


Forum >> NWChem's corner >> Running NWChem