Salam
I am conducting a single point calculation on an optimized structure of chitosan-ZnO nanocomposite with a DFT/B3LYP/LanL2DZ method to get the energy of the anionic form of the ground state. All my calculations have failed, I used several keywords (sleazy, SCF=Restart, scf=tight, integral=ultrafine scf=noincfock). I used the chk file of the optimized structure in the route section with the keywords guess=read and geom=check, but I always got the same message:
>>>>>>>>>> Convergence criterion not met.
SCF Done: E(UB3LYP) = -3707.53491239 A.U. after 129 cycles
Convg = 0.1563D-06 -V/T = 1.8967
= 0.0000 = 0.0000 = 0.5000 = 0.8070 S= 0.5281
= 0.000000000000E+00
Annihilation of the first spin contaminant:
S**2 before annihilation 0.8070, after 0.7510
DiagDN has N= 740 LTot= 2252 but NE2= 3 cannot use DSYEVD.
Convergence failure -- run terminated.
Error termination via Lnk1e in C:\G09W\l502.exe
Is there any suggestion to accomplish this calculation without using scf=qc or scf=xqc because of its much time comception?
Here is my 9throute section :
%mem=100MW
%chk=D:\Mes calcul opto surface\Neutral surface\neutral (ground S) anion\9 essai\2 CHITO4-ZNO9 Surface.chk
# b3lyp/lanl2dz nosymm guess=read counterpoise=2 geom=check sleazy
Title Card Required
-1 2 0 1 0 1
If you calculate the total energy of the neutral at its optimized geometry and the energy of the cation with the same method/basis set at the optimized geometry of the neutral, and then subtract the two energies, you have what is called the vertical ionization potential. 'vertical' because you assumed that there was no change in geometry. If you optimize the geometry of the cation and subtract the two energies (IP = Ecation - Eneutral), then you have the adiabatic ionization potential.
Similar processes for the electron affinity. You add an electron, do the calculations and you can form the vertical electron affinity or adiabatic electron affinity.
You remove/add an electron to the molecule by changing the net charge and the spin multiplicity in the input file. In G09 for neutral formaldehyde, the line with net charge (charge = zero) and spin multiplicity (spin multiplicy = 1 for a singlet, all electrons paired) would look like
0 1
For the formaldehyde cation (charge = +1; spin multiplicity = 2 for a doublet), the line would look like
1 2
For the formaldehyde aniion (charge = -1; spin multiplicity = 2 for a doublet), the line would look like
-1 2
(The spin multiplicities in the example given above for the formaldehyde cation and anion are not correct)
The procedures I described can be followed with any method/basis set you wish (HF, DFT, MP2.....)
IFF you are using Hartree-Fock theory, then you can use the molecular orbital energies to get quick (and mostly dirty) estimates of the IP and the EA. in HF theory for a closed shell system, the IP equals the orbital energy of the HOMO (highest occupied molecular orbital) with a minus sign, IP = -epsilon(HOMO). This is known as Koopmans Approximation (Koopmans Theorem). Works out reasonably well numerically for organic molecules.
You can get EA similarly from the orbital energy of the lowest unoccupied molecular orbital (LUMO) but this does not work well numerically (and Koopmans never said it would!).
The simple approximations of relating IP/EA to HOMO/LUMO orbital energies do not work well numerically in DFT.
Conceptually, we relate IP/EA to the HOMO/LUMO orbital energies all the time and that is fine. Getting real accurate values for IP and, in particular, EA is a different story.
Mekahlia Soumia In think you should increase the maximum number of SCF cycles to 512 or 1024... Include this keyword scf=(maxcycle=512) in the route section.
Hello, I would like to thank Mr Mohamed Khedawy and Mr Melek Hajjifor their help. My calculation in completed, it takes only 2 days and 19 h by using scf =xqc.
Job cpu time: 2 days 19 hours 58 minutes 16.0 seconds.
File lengths (MBytes): RWF= 633 Int= 0 D2E= 0 Chk= 74 Scr= 1
Normal termination of Gaussian 09 at Wed Sep 29 21:57:01 2021.
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology
- Sectioning