Molcas Forum
Support and discussions for Molcas and OpenMolcas users and developers
You are not logged in.
- Topics: Active | Unanswered
Announcement
#1 2022-12-15 23:23:09
- kaufold.b
- Member
- Registered: 2022-12-15
- Posts: 3
[SOLVED] Differences in Results When Z-Matrix Is Used
Hello all,
I would like to report that I have encountered significant differences in multireference results when I use a Z-matrix to specify the molecular geometry vs. when I use an XYZ file, my usual procedure. The results when I use an XYZ file are more accurate.
Here I have tried to run a calculation on hydrogen sulfide, using the same geometry in an XYZ file and a Z-matrix. All other parameters, including the active space and (lack of) symmetry are identical as well. I have verified that the geometries are identical for the calculation using an XYZ file...
*******************************************
* InterNuclear Distances / Angstrom *
*******************************************
1 S1 2 H2 3 H3
1 S1 0.000000
2 H2 1.336584 0.000000
3 H3 1.336584 1.928263 0.000000
**************************************
* Valence Bond Angles / Degree *
**************************************
Atom centers Phi
2 H2 1 S1 3 H3 92.33...and a Z-matrix...
*******************************************
* InterNuclear Distances / Angstrom *
*******************************************
1 S1 2 H2 3 H3
1 S1 0.000000
2 H2 1.336584 0.000000
3 H3 1.336584 1.928256 0.000000
**************************************
* Valence Bond Angles / Degree *
**************************************
Atom centers Phi
2 H2 1 S1 3 H3 92.33...however, the resulting orbital energies are radically different. Here are the orbitals from a CASSCF(8,8)/ANO-RCC-VTZP calculation using the XYZ file and M06-2X starting orbitals as summarized by &GRID_IT.
GridName= 1 62 4.5869 (0.00) s
GridName= 1 61 4.5060 (0.00) s
GridName= 1 60 4.0589 (0.00) s
GridName= 1 59 3.8408 (0.00) s
GridName= 1 58 3.8005 (0.00) s
GridName= 1 57 3.7505 (0.00) s
GridName= 1 56 3.5957 (0.00) s
GridName= 1 55 3.5941 (0.00) s
GridName= 1 54 2.9934 (0.00) s
GridName= 1 53 2.9133 (0.00) s
GridName= 1 52 2.7481 (0.00) s
GridName= 1 51 2.6436 (0.00) s
GridName= 1 50 2.5663 (0.00) s
GridName= 1 49 2.4501 (0.00) s
GridName= 1 48 2.3797 (0.00) s
GridName= 1 47 2.3719 (0.00) s
GridName= 1 46 2.3547 (0.00) s
GridName= 1 45 2.3157 (0.00) s
GridName= 1 44 2.1147 (0.00) s
GridName= 1 43 2.1128 (0.00) s
GridName= 1 42 1.9500 (0.00) s
GridName= 1 41 1.7716 (0.00) s
GridName= 1 39 1.7654 (0.00) s
GridName= 1 40 1.7654 (0.00) s
GridName= 1 38 1.6512 (0.00) s
GridName= 1 37 1.4249 (0.00) s
GridName= 1 36 1.4166 (0.00) s
GridName= 1 35 1.3235 (0.00) s
GridName= 1 34 1.3021 (0.00) s
GridName= 1 33 1.2689 (0.00) s
GridName= 1 32 1.2496 (0.00) s
GridName= 1 31 1.2255 (0.00) s
GridName= 1 30 1.1325 (0.00) s
GridName= 1 29 1.1205 (0.00) s
GridName= 1 28 0.8930 (0.00) s
GridName= 1 27 0.7500 (0.00) s
GridName= 1 26 0.6250 (0.00) s
GridName= 1 25 0.6186 (0.00) s
GridName= 1 24 0.6071 (0.00) s
GridName= 1 23 0.5649 (0.00) s
GridName= 1 22 0.5602 (0.00) s
GridName= 1 21 0.4562 (0.00) s
GridName= 1 20 0.3612 (0.00) s
GridName= 1 19 0.2898 (0.00) s
GridName= 1 18 0.2867 (0.00) s
GridName= 1 17 0.2682 (0.00) s
GridName= 1 16 0.2402 (0.00) s
GridName= 1 15 0.1752 (0.00) s
GridName= 1 14 0.1581 (0.00) s
GridName= 1 11 0.0000 (0.17) 2
GridName= 1 13 0.0000 (0.18) 2
GridName= 1 12 0.0000 (0.20) 2
GridName= 1 10 0.0000 (0.34) 2
GridName= 1 9 0.0000 (1.33) 2
GridName= 1 8 0.0000 (1.82) 2
GridName= 1 7 0.0000 (1.97) 2
GridName= 1 6 0.0000 (1.99) 2
GridName= 1 5 -6.7618 (2.00) i
GridName= 1 4 -6.7619 (2.00) i
GridName= 1 3 -6.7623 (2.00) i
GridName= 1 2 -9.1323 (2.00) i
GridName= 1 1 -92.3720 (2.00) iHere is the same summary when a Z-matrix is used..
GridName= 1 13 0.0000 (0.00) 2
GridName= 1 12 0.0000 (0.00) 2
GridName= 1 11 0.0000 (0.00) 2
GridName= 1 9 0.0000 (1.33) 2
GridName= 1 10 0.0000 (1.33) 2
GridName= 1 8 0.0000 (1.33) 2
GridName= 1 7 0.0000 (2.00) 2
GridName= 1 6 0.0000 (2.00) 2
GridName= 1 62 -0.2763 (0.00) s
GridName= 1 61 -0.3773 (0.00) s
GridName= 1 60 -1.3833 (0.00) s
GridName= 1 59 -1.4289 (0.00) s
GridName= 1 58 -2.2859 (0.00) s
GridName= 1 57 -2.5722 (0.00) s
GridName= 1 56 -3.1976 (0.00) s
GridName= 1 55 -4.0712 (0.00) s
GridName= 1 54 -4.6751 (0.00) s
GridName= 1 53 -4.7442 (0.00) s
GridName= 1 52 -4.7579 (0.00) s
GridName= 1 51 -6.6757 (0.00) s
GridName= 1 50 -9.6529 (0.00) s
GridName= 1 49 -10.1530 (0.00) s
GridName= 1 48 -10.7130 (0.00) s
GridName= 1 47 -11.3990 (0.00) s
GridName= 1 46 -13.0240 (0.00) s
GridName= 1 45 -13.3810 (0.00) s
GridName= 1 44 -14.2330 (0.00) s
GridName= 1 43 -16.3500 (0.00) s
GridName= 1 42 -20.1930 (0.00) s
GridName= 1 41 -26.7600 (0.00) s
GridName= 1 40 -30.2870 (0.00) s
GridName= 1 39 -32.1590 (0.00) s
GridName= 1 38 -33.2930 (0.00) s
GridName= 1 37 -34.0950 (0.00) s
GridName= 1 36 -34.5390 (0.00) s
GridName= 1 35 -34.8840 (0.00) s
GridName= 1 34 -36.6150 (0.00) s
GridName= 1 33 -39.1750 (0.00) s
GridName= 1 32 -39.2400 (0.00) s
GridName= 1 31 -39.6650 (0.00) s
GridName= 1 30 -39.8500 (0.00) s
GridName= 1 29 -46.2200 (0.00) s
GridName= 1 28 -46.7980 (0.00) s
GridName= 1 27 -49.3350 (0.00) s
GridName= 1 26 -51.0420 (0.00) s
GridName= 1 25 -52.5470 (0.00) s
GridName= 1 24 -53.9760 (0.00) s
GridName= 1 23 -59.8550 (0.00) s
GridName= 1 22 -69.9500 (0.00) s
GridName= 1 21 -72.0440 (0.00) s
GridName= 1 20 -73.7240 (0.00) s
GridName= 1 19 -77.2610 (0.00) s
GridName= 1 18 -77.7820 (0.00) s
GridName= 1 17 -78.0200 (0.00) s
GridName= 1 16 -81.2340 (0.00) s
GridName= 1 15 -81.2960 (0.00) s
GridName= 1 14 -84.9280 (0.00) s
GridName= 1 5 -740.5100 (2.00) iSome orbitals are missing, and secondary orbitals have negative energies when they shouldn't. Is this a bug? Or is there just an additional procedure I need to take when I want to use a Z-matrix for my geometry?
For completeness, here is my XYZ file.
3
S 0.000000000000 0.000000000000 0.783794900443
H 0.964131503231 0.000000000000 -0.141897450221
H -0.964131503231 0.000000000000 -0.141897450221Additionally, my input file using this geometry.
>>> Export MOLCAS_NEW_DEFAULTS=NO
>>> Export MOLCAS_TRAP=OFF
&GATEWAY
Title = Hydrogen Sulfide MR, XYZ Coordinates
Coord = optimized.xyz
Basis = ANO-RCC-VTZP
Group = NoSym
&SEWARD
grid input
grid=ultrafine
end of grid input
&SCF
KSDFT = M062X
&RASSCF
nActEl = 8
Ras2 = 8
Charge = 0
Spin = 1
CIRoot = 6 6 1
OutOrbital = Natural; 6
LevShft = 1.0
&CASPT2
MaxIter = 10000
Prop
&MCPDFT
KSDFT = T:PBE
&RASSI
&GRID_IT
FileOrb = $Project.RasOrb
Name = RasOrb
AllFinally, my input file using a Z-matrix.
>>> Export MOLCAS_NEW_DEFAULTS=NO
>>> Export MOLCAS_TRAP=OFF
&SEWARD
XBAS
S.ANO-RCC-VTZP
H.ANO-RCC-VTZP
End of Basis
ZMAT
S1
H2 1 1.336584
H3 1 1.336584 2 92.33
End of Z-Matrix
grid input
grid=ultrafine
end of grid input
&SCF
KSDFT = M062X
&RASSCF
nActEl = 8
Ras2 = 8
Charge = 0
Spin = 1
CIRoot = 6 6 1
OutOrbital = Natural; 6
LevShft = 1.0
&CASPT2
MaxIter = 10000
Prop
&MCPDFT
KSDFT = T:PBE
&RASSI
&GRID_IT
FileOrb = $Project.RasOrb
Name = RasOrb
AllI realize it is not recommended to use &SEWARD to build the molecular geometry. However, I experienced a different error when I used the &GATEWAY module; following modules did not get the proper information about the basis functions. I built the molecule in &SEWARD directly to ensure the proper information was passed down to other modules. Thank you for any input!
Offline
#2 2022-12-16 09:56:08
- Ignacio
- Administrator
- From: Uppsala
- Registered: 2015-11-03
- Posts: 1,204
Re: [SOLVED] Differences in Results When Z-Matrix Is Used
It seems the XBAS keyword does not trigger the automatic switching of relativistic Hamiltonians according to the basis sets. Add "Relativistic = R02O02" to the &SEWARD input with Z-matrix and see what happens.
Offline
#3 2023-01-05 21:16:12
- kaufold.b
- Member
- Registered: 2022-12-15
- Posts: 3
Re: [SOLVED] Differences in Results When Z-Matrix Is Used
It seems that did the trick! It also works when I specify other parameterizations (for better comparison with my calculation using an XYZ file I also tried R02E00). Thank you!
Offline