Multiwfn forum

taha55

Member

Registered: 2019-09-19

Posts: 38

Fukui f+, f- calculating formula

Hello Dear

The formulas for f+,f- which is many articles and multiwfn manual;

f+= q(N+1)− qN and f-= q(N)− q(N-1)   (page 387 in multiwfn manual) (1. Eq.) (N=charge 0, N+1=charge-1, N-1= charge+1)

Fuki calculation is given below in section ''Calculating global and atomic indices ''(page 679) in multiwfn manual

Units used below are "e" (elementary charge)
Atom q(N)        q(N+1) q(N-1)    f-         f+       f0       CDD
1(C ) -0.0587 -0.1185 0.0852 0.1439 0.0598 0.1018  -0.0841
2(C ) -0.0390 -0.1674 0.0268 0.0658 0.1284 0.0971   0.0626
3(C ) -0.0597 -0.1873 0.0319 0.0916 0.1276 0.1096   0.0360

If f+and f- are calculated for 1(C) according to equation 1;

f+= q(N+1)− qN=-0.1185-( -0.0587)=-0.0598
f-=q(N)− q(N-1)=-0.0587-0.0852=-0.1439

But obove the table are calculated as f+= 0.0598 and f-=0.1439  so  Do these formulas must arranged as follows?

f+= qN(r)-q(N+1)  (2. Eq.)
f- =q(N-1)-qN(r)

Which formula is correct for f+and f- (1. or 2. Eq.)?

Best regards

Last edited by taha55 (2020-03-03 11:43:53)

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,830

Website

Re: Fukui f+, f- calculating formula

“f+= q(N+1)− qN and f-= q(N)− q(N-1) ”  is wrong, please carefully check Section 4.7.3 of latest version of Multiwfn manual to understand the reason.

taha55

Member

Registered: 2019-09-19

Posts: 38

Re: Fukui f+, f- calculating formula

Thank you. Can you suggest an article or yours containing the correct formulas in Section 4.7.3 for me to cite?
Best Regards

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,830

Website

Re: Fukui f+, f- calculating formula

Eq. 5 of Acta Phys. -Chim. Sin. 2014, 30 (4), 628-639（http://www.whxb.pku.edu.cn/CN/abstract/ … 8694.shtml）
or
Eq. 3 of DOI: 10.1007/s11426-015-5494-7

taha55

Member

Registered: 2019-09-19

Posts: 38

Re: Fukui f+, f- calculating formula

Thank you very much for help
Best Regards

behnam12

Member

Registered: 2020-02-26

Posts: 11

Re: Fukui f+, f- calculating formula

Hi Dr Tian Lu
I want to calculate the fukui function.
I have 10 samples.
Unfortunately, several output files are generated, neutral, anion, and cation, but for some, anion output is not generated, and some cation samples are not.
Error 502 is found in Gaussian.
Please help me.
Best regards
Behnam dehbandi
*******

No special actions if energy rises.
EnCoef did   100 forward-backward iterations
Restarting incremental Fock formation.
Restarting incremental Fock formation.
Restarting incremental Fock formation.
Restarting incremental Fock formation.
Restarting incremental Fock formation.
Restarting incremental Fock formation.
>>>>>>>>>> Convergence criterion not met.
SCF Done:  E(UM062X) =  -4021.89078842     A.U. after  129 cycles
             Convg  =    0.5974D-05             -V/T =  2.0070
<Sx>= 0.0000 <Sy>= 0.0000 <Sz>= 0.5000 <S**2>= 0.7948 S= 0.5222
<L.S>= 0.000000000000E+00
Annihilation of the first spin contaminant:
S**2 before annihilation     0.7948,   after     0.7518
DiagDN has N=  1317 LTot=  3985 but NE2=     3 cannot use DSYEVD.
Convergence failure -- run terminated.

Last edited by behnam12 (2020-04-09 18:46:00)

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,830

Website

Re: Fukui f+, f- calculating formula

Evidently, the SCF convergence is failed. You need to add some special keywords to facilitate SCF convergence. I have a blog article to show how to solve this problem (written in Chinese, it is not difficult to understand via Google translator): http://sobereva.com/61

It is worth to note that if the basis set is large and especially when diffuse functions are heavily added, anion system is always somewhat difficult to converge. If there is no special reason, I suggest using other DFT functional instead of M06-2X, because it is well known that M06-2X is usually more difficult to make SCF converge. In addition, using "scf=conver=6", which has been mentioned in my blog article, is safe for present purpose, with this keyword the SCF convergence criterion is much easier to be reached.

behnam12

Member

Registered: 2020-02-26

Posts: 11

Re: Fukui f+, f- calculating formula

Hi Dr Tian Lu
Thank you very much for help
Follow the previous question about the error of convergence of Gamess
software?
Best regards
BehnamDehbandi

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,830

Website

Re: Fukui f+, f- calculating formula

Do you mean you need methods to solve SCF convergence problem in GAMESS-US?

behnam12

Member

Registered: 2020-02-26

Posts: 11

Re: Fukui f+, f- calculating formula

Hi dr
Yes

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,830

Website

Re: Fukui f+, f- calculating formula

$SCF DIIS=.T. ETHRSH=2.0 SWDIIS=0.005 FDIFF=.F.
$DFT SWITCH=1.0 SWOFF=1.0 $END
Increase MAXIT in $CONTRL
and so on...

By the way, if SCF can converge in Gaussian, you can use Multiwfn to load .fch file, then use subfunction 2 of main function 100 to export GAMESS-US input file, the loaded wavefunction can be written as initial guess wavefunction in the file. In this case, if you use exactly identical calculation level in GAMESS-US as in Gaussian, the SCF convergence will be very easy to reach since the initial guess is perfect.