SEARCH
TOOLBOX
LANGUAGES
Forum Menu

Large couplings for ET calcs

From NWChem

Viewed 166 times, With a total of 2 Posts
Jump to: navigation, search

Clicked A Few Times
Threads 5
Posts 16
I've been seeing some very large couplings when running the et task, and so I decided to try the example in the manual, a self-exchange reaction with He/He+ ([1]) with NWChem version 6.0. I tried two runs, one with the He 5 Angstroms apart, and one @ 10 Angstroms apart. Basis set was Ahlrichs_pVDZ, from the library.

I don't know what the geometry or basis were for the example given in the documentation. For reference the values are given as


Electronic energy of reactants     H(RR)      -5.3402392824
Electronic energy of products      H(PP)      -5.3402392824
Reactants/Products overlap         S(RP)      -0.0006033839
One-electron contribution         H1(RP)       0.0040314092
Two-electron integral screening (tol2e) : 6.03E-11
Two-electron contribution         H2(RP)      -0.0007837138
Total interaction energy           H(RP)       0.0032476955
Electron Transfer Coupling Energy |V(RP)|      0.0000254810
                                                      5.592 cm-1
                                                   0.000693 eV
                                                      0.016 kcal/mol


For 5 A, I see
 Electronic energy of reactants     H(RR)      -3.1919485760
 Electronic energy of products      H(PP)      -3.1919485759
 Reactants/Products overlap S(RP) :  1.00D+00
 One-electron contribution         H1(RP)      -4.6181893392
 Two-electron integral screening (tol2e) : 6.03D-11
 Two-electron contribution         H2(RP)       1.4262407635
 Total interaction energy           H(RP)      -3.1919485757
 Electron Transfer Coupling Energy |V(RP)|      2.1006474897
                                                  461038.768 cm-1
                                                   57.161559 eV
                                                    1318.177 kcal/mol


For 10 A, I see
 Electronic energy of reactants     H(RR)      -2.9537597557
 Electronic energy of products      H(PP)      -2.9537597556
 Reactants/Products overlap S(RP) :  1.00D+00
 One-electron contribution         H1(RP)      -4.3005260783
 Two-electron integral screening (tol2e) : 6.03D-11
 Two-electron contribution         H2(RP)       1.3467663226
 Total interaction energy           H(RP)      -2.9537597556
 Electron Transfer Coupling Energy |V(RP)|      0.7327601032
                                                  160822.231 cm-1
                                                   19.939428 eV
                                                     459.814 kcal/mol

Although the trend is as expected (coupling goes down with distance; @ tol2e=default 1e-7, the opposite happened), the couplings are much too large. Also, S=1, unlike the small negative number in the example.
Thanks for any help.


Input for the 5A job was:


start 8TNaEwO57SmBqwjO
permanent_dir /scratch/scratchdirs/cchang/8TNaEwO57SmBqwjO/perm
scratch_dir /scratch/scratchdirs/cchang/8TNaEwO57SmBqwjO/scr
ECHO
TITLE "Test for ET module os NWChem"
geometry He nocenter noautoz
  He 0.0   0.0   0.0
end
geometry HeP nocenter noautoz
  He 5.0   0.0   0.0
end
geometry HeHeP nocenter noautoz
  He 0.0   0.0   0.0
  He 5.0   0.0   0.0
end
basis "ao basis" cartesian
   * library "Ahlrichs_pVDZ"
end
scf
   singlet; uhf; vectors input atomic output He.movecs
end
charge 0
set geometry He
task scf
scf
   doublet; uhf; vectors input atomic output HeP.movecs
end
charge 1
set geometry HeP
task scf
#ET reactants:
scf 
  doublet; uhf; vectors input fragment He.movecs HeP.movecs output HeA.movecs
end 
set geometry HeHeP
task scf
#ET products:
scf 
  doublet; uhf; vectors input HeA.movecs reorder 2 1 output HeB.movecs
end 
task scf
et
 tol2e 6.03e-11
 vectors reactants HeA.movecs 
 vectors products HeB.movecs
end
task scf et

  • Bert Forum:Admin, Forum:Mod, NWChemDeveloper, bureaucrat, sysop
    Profile
    Send PM
Forum Regular
Threads 2
Posts 285
An overlap between the reactant and product being 1 obviously means that the orbitals for both the reactant and product are the same, i.e. the electron has not been moved from one atom to the other. You can easily check this by looking at the changes and populations at the end of the reactant and product SCF runs. The following input deck works fine for me.

Bert



start chang
ECHO
TITLE "Test for ET module os NWChem"

basis "ao basis" cartesian
  * library "Ahlrichs_pVDZ"
end

geometry
He 0 0 0
end

charge 1

scf
uhf
doublet
vectors output HeP.movecs
end
task scf

charge 0

scf
uhf
singlet
vectors output He.movecs
end
task scf

geometry noautosym noautoz
 He 0.0   0.0   0.0
He 5.0 0.0 0.0
end

charge 1
  1. ET reactants:
scf
 doublet; uhf; vectors input fragment HeP.movecs He.movecs output HeA.movecs
end
task scf

  1. ET products:
scf
 doublet; uhf; vectors input HeA.movecs reorder 2 1 output HeB.movecs
end
task scf

et
vectors reactants HeA.movecs
vectors products HeB.movecs
end
task scf et



Quote:Cchang Oct 17th 7:22 pm
I've been seeing some very large couplings when running the et task, and so I decided to try the example in the manual, a self-exchange reaction with He/He+ ([1]) with NWChem version 6.0. I tried two runs, one with the He 5 Angstroms apart, and one @ 10 Angstroms apart. Basis set was Ahlrichs_pVDZ, from the library.

I don't know what the geometry or basis were for the example given in the documentation. For reference the values are given as


Electronic energy of reactants     H(RR)      -5.3402392824
Electronic energy of products      H(PP)      -5.3402392824
Reactants/Products overlap         S(RP)      -0.0006033839
One-electron contribution         H1(RP)       0.0040314092
Two-electron integral screening (tol2e) : 6.03E-11
Two-electron contribution         H2(RP)      -0.0007837138
Total interaction energy           H(RP)       0.0032476955
Electron Transfer Coupling Energy |V(RP)|      0.0000254810
                                                      5.592 cm-1
                                                   0.000693 eV
                                                      0.016 kcal/mol


For 5 A, I see
 Electronic energy of reactants     H(RR)      -3.1919485760
 Electronic energy of products      H(PP)      -3.1919485759
 Reactants/Products overlap S(RP) :  1.00D+00
 One-electron contribution         H1(RP)      -4.6181893392
 Two-electron integral screening (tol2e) : 6.03D-11
 Two-electron contribution         H2(RP)       1.4262407635
 Total interaction energy           H(RP)      -3.1919485757
 Electron Transfer Coupling Energy |V(RP)|      2.1006474897
                                                  461038.768 cm-1
                                                   57.161559 eV
                                                    1318.177 kcal/mol


For 10 A, I see
 Electronic energy of reactants     H(RR)      -2.9537597557
 Electronic energy of products      H(PP)      -2.9537597556
 Reactants/Products overlap S(RP) :  1.00D+00
 One-electron contribution         H1(RP)      -4.3005260783
 Two-electron integral screening (tol2e) : 6.03D-11
 Two-electron contribution         H2(RP)       1.3467663226
 Total interaction energy           H(RP)      -2.9537597556
 Electron Transfer Coupling Energy |V(RP)|      0.7327601032
                                                  160822.231 cm-1
                                                   19.939428 eV
                                                     459.814 kcal/mol

Although the trend is as expected (coupling goes down with distance; @ tol2e=default 1e-7, the opposite happened), the couplings are much too large. Also, S=1, unlike the small negative number in the example.
Thanks for any help.


Input for the 5A job was:


start 8TNaEwO57SmBqwjO
permanent_dir /scratch/scratchdirs/cchang/8TNaEwO57SmBqwjO/perm
scratch_dir /scratch/scratchdirs/cchang/8TNaEwO57SmBqwjO/scr
ECHO
TITLE "Test for ET module os NWChem"
geometry He nocenter noautoz
  He 0.0   0.0   0.0
end
geometry HeP nocenter noautoz
  He 5.0   0.0   0.0
end
geometry HeHeP nocenter noautoz
  He 0.0   0.0   0.0
  He 5.0   0.0   0.0
end
basis "ao basis" cartesian
   * library "Ahlrichs_pVDZ"
end
scf
   singlet; uhf; vectors input atomic output He.movecs
end
charge 0
set geometry He
task scf
scf
   doublet; uhf; vectors input atomic output HeP.movecs
end
charge 1
set geometry HeP
task scf
#ET reactants:
scf 
  doublet; uhf; vectors input fragment He.movecs HeP.movecs output HeA.movecs
end 
set geometry HeHeP
task scf
#ET products:
scf 
  doublet; uhf; vectors input HeA.movecs reorder 2 1 output HeB.movecs
end 
task scf
et
 tol2e 6.03e-11
 vectors reactants HeA.movecs 
 vectors products HeB.movecs
end
task scf et

  • Guest -
    Send PM
Hi Bert,

  Thanks, I could get it with your deck. I figured the overlap was symptomatic of the problem, but I was expecting an Inf (or stars) coupling if the overlap were really 1.0 between the reactant and product wavefunctions. It came down to requiring "noautosym" to be present. Now I see, symmetrizing the wavefunctions forced the solutions to delocalize despite the localized fragment guess. 

Thanks again!


Forum >> NWChem's corner >> Running NWChem



Who's here now Members 0 Guests 0 Bots/Crawler 1


AWC's: 2.5.10 MediaWiki - Stand Alone Forum Extension
Forum theme style by: AWC