# Overlay 4

IOp(4/5)

Type of guess.

0 | Default. This uses the Harris functional except for semi-empirical, for which the modified core Hamiltonian is diagonalized. |

-1 | Skip out and leave all files as left over on the rwf from whatever was done previously. |

1 | Read guess from the checkpoint file. |

2 | Guessfrom model Hamiltonian, chosen via IOp(4/11). |

3 | Huckel guess (only valid for NDDO-type methods). |

4 | Projected ZDO guess. |

5 | Renormalize and orthogonalize the coefficients which are on the read-write files. |

6 | Renormalize and orthogonalize intermediate SCF results which are on the RWF. |

7 | Read intermediate SCF results which are on the checkpoint file. |

8 | Read the generalized density specified by IOp(4/38) from the checkpoint file and generate natural orbitals from it. |

9 | Read the generalized density specified by IOp(4/38) from the RWF file and generate natural orbitals from it. |

10-14 | Generated internally and correspond to 0 and 5-8 for sparse. |

16 | Use the orthonormal set provided by L302 as MOs, avoiding any diagonalization. |

17 | Store unit matrices for a dummy guess. |

18 | Copy orbitals and densities that are in the chk file without checking or alteration. |

100 | Convert Guess=Check to Guess=Restart or to generating guess depending on what if anything is on the checkpoint file. |

1000 | Use the simultaneous optimization recipe: S^{-0.5}* V. |

00000 | Default (1 for PBC without alter, otherwise 2). |

10000 | Re-use Fock matrices instead of orbitals. |

20000 | Re-use orbitals not Fock matrices. |

100000 | Read the name of a checkpoint file from the input stream and read guess MOs from it, or read an option for how to generate the guess. |

Note that variable IGuess here has 4,3,2,1 corresponding to 1,2,3,4 above. IGuess values of 10-14 are generatedinternally and are the sparse versions of 0 and 5-8.

IOp(4/6)

L401: Projection, orthogonalization, and checking of initial guess.

0 | Default (1 except 3 for IOp(129)=1). |

1 | Force projected read-in guess, even when bases are identical. |

2 | Suppress projection. |

3 | Project only if basis sets are different. |

00 | Default orthogonalization (perform if Guess=Cards). |

10 | Schmidt orthogonalize guess orbitals. |

20 | Suppress orthogonalization. |

000 | Default MO checking (check if Guess=Cards or Guess=Mix). |

100 | Check MOs for othornormality. |

200 | Don’t check MOs for othornormality. |

100000000 | Default all 3 to on |

200000000 | Default all 3 to off. |

IOp(4/8)

L401: Alteration of configuration.

0 | Default (3). |

1 | Read in pairs of integers in free format indicating which pairs of MOâ€™s are to be interchanged. Pairs are read until a blank card is encountered. |

2 | Read in a permutation of the orbitals. |

3 | Do not alter configuration. |

10 | Read alteration information from the read-write file. |

100 | Use alpha orbitals for guess for both alpha and beta. |

1000 | Biorthogonalize UHF MOs. |

Note: If the configuration is altered on an open shell system, two sets of data as described above will be expected, first for alpha, second for beta.

IOp(4/9)

L401: SCF symmetry control.

0 | Default, same as 104 except 4 for IGuess=16, and 204 if C1 symmetry. |

1 | Read groups of irreducible representations to combine in the SCF. These are read before any orbitals and before alteration commands. |

2 | Use no symmetry in the SCF. |

3 | Pick up the symmetry mixing information from the alteration read-write file. |

4 | Use the full Abelian point group, as represented by the symmetry adapted basis functions produced by link 301. Initial guess orbital symmetries are assigned. |

5 | (Use symmetry in SCF if possible, but do not assign initial guess Abelian symmetries). |

10 | Localize all occupied orbitals together and all virtual orbitals together. |

20 | Localize the orbitals within the selected or defaulted symmetry. |

30 | Localize all occupied and virtual orbitals together. |

40 | Do not localize. |

100 | Assign orbital symmetries for printing in full symmetry. |

200 | Do not assign orbital symmetries in full symmetry. |

1000 | Force the guess orbitals to have the Abelian symmetry. |

NN0000 | Use localization method NN-1 (see LocMO). |

This option can cause the symmetry adapted basis function common blocks to be modified.

IOp(4/11)

L401: Type of Guess.

For iterative ZDO Guess:

-1 | Force old path using old Huckel. |

0 | Best available (8,4 in order of preference). |

1 | Old Huckel. |

2 | CNDO. |

3 | INDO. |

4 | New Huckel. |

5 | Iterative extended Huckel. |

6 | Harris, converted to IGuess=3 and IZDO=3 here. |

7 | Harris with interpolated QEq atomic charges, converted to IGuess=3 IZDO=5 here. |

8 | Harris with new densities. |

9 | Iterated Harris with QEq guess, converted to IGuess=3 IZDO=7. |

10 | Unused. |

11 | NYI? Harris using charges from previous SCF, converted to IGuess=3 IZDO=9. |

For unprojected single diagonalization guess:

0 | Default(1 for DFTB, 2 for AM1/PM6, 3 for ab initio). |

1 | Use bare core matrix. |

2 | Dress core Hamiltonian with QEq-based density. |

3 | Use Harris Functional with old densities. |

4 | Neutral atom AM1/PMx guess. |

5 | Harris functional with interpolated QEq charges. |

6 | Harris functional with iterated charges. |

7 | Harris functional with iterated charges starting from QEq. |

8 | Use Harris Functional with new densities. |

9 | Harris using charges from previous SCF |

000 | Default, same as 2. |

100 | Use at least SG1 in Harris guess. |

200 | Use at least FineGrid in Harris guess. |

300 | Use at least UltraFine in Harris guess. |

400 | Use an unpruned (199,590) or (399,590) grid depending on the range of primitive exponents.. |

500 | Use(399,974) and 10^{-12} in Harris functional. |

1000 | Save energy in Gen(43) for Harris functional. |

MMMM00000 | Use functional MMMM. |

IOp(4/13)

L401: Mixing of orbitals.

-2 | No mixing. |

-1 | Mix HOMO and LUMO (skipping beta high-spin orbitals for GHF). |

0 | Default: Mix HOMO and LUMO to make complex guess for CRHF and CUHF if generating RUHF guess, otherwise do nothing. |

>0 | Bits request actions as follows: |

0: Mix HOMO and LUMO (skipping beta high-spin virtuals for GHF), done after complex/spin mixings. | |

1: Do complex mixing, changing spin direction for GHF. | |

2: Use real rather than imaginary coefficients. | |

3: Flip sign of complex mixing. | |

4: Read in a spin-vector and rotate to align spins in this direction instead of Z. GHF only. | |

5: Read in two spin-vectors and use them for alternate orbitals. | |

6: Reverse rotation direction applied to spin. | |

Note that this will usually destroy both spatial and alpha/beta symmetry. The mixing is done after any alterations. Bits 1-3 are only relevant for complex wfns. |

IOp(4/14)

L401: Reading of specific orbitals.

0 | No. |

1 | Yes. For alpha orbitals, read one card with the format for the orbitals, followed by zero or more sets of IVec (I5): vector to replace. If IVec is -1, all NBasis vectors follow.(Vector(I), I=1, NBasis): vector in the specified format. Input is terminated by IVec=0. For b orbitals, the same format as for a is used. Note that if Alter is also specified, the replacements are read before the corr. alterations (thus the order is a orbitals, a alterations, b orbitals, b alterations). |

2 | Yes. Read using the format described in Routine RdMO2. Here a range of MOs is indicated by two integers followed by an integer giving the number of basis functions. Then a list of MO energies are given. Lastly, the MO coefficients are read in sequence. All of the reading is carried out in free format. |

10 | Orbitals are assumed to have mixed normalization for Cartesian d and higher functions (equivalent to having AdjMO applied to them). |

100 | Reorder d and f coefficients from the order used in NWChem (as of January, 2013) to the conventional order used in Gaussian. |

900 | Read permutation arrays for p and higher functions for use in reordering read-in MO coefficients. (NYI) |

IOp(4/15)

L401: Spin-state for initial guess.

0 | Use multiplicity in /Mol/. |

N | Use multiplicity N. Useful for generating guesses for open-shell singlets or unusual spin states involving orthogonal orbs by treating them as high-spin in the guess (which only does UHF). |

IOp(4/16)

L401: Whether to translate basis functions of read in guess.

0 | Default (same as 3). |

1 | Use the basis functions as is. |

2 | Translate to the current atomic coordinates. |

3 | Translate to the current atomic coordinates, and determine an overall rotation to provide to the read-in orbitals. |

IOp(4/17)

L402: Number of open-shell orbitals (not electrons).

0 | Number of open electrons. |

N | N. |

L405: Number of electrons in the CAS space.

IOp(4/18)

L402: Number of orbitals in CI. Default is number of open shells.

Number of orbitals in the CAS space.

IOp(4/19)

L402: Spin change in CI (default based on multiplicity).

L405: Truncation level for excitations — default full CAS.

IOp(4/20)

L402: Type of model. (This is also tested in L401 to see whether atomic numbers greater than 102 are special flags).

0 | Default (AM1). |

1 | CNDO. |

2 | INDO. |

3 | MINDO/3. |

4 | MNDO. |

5 | AM1. |

6 | Unused. |

7 | PM3. |

8 | PM3 with mechanics correction. |

9 | Dreiding mechanics. |

10 | UFF mechanics. |

11 | AMBER mechanics. |

12 | MM2 mechanics. |

13 | MM3 mechanics. |

14 | Extended Huckel, Hoffmann parameters. |

15 | Extended Huckel, Muller parameters. |

16 | Extended Huckel, Initial guess parameters. |

17 | External program. |

18 | MMFF. |

19 | QFF. |

IOp(4/21)

L402: SCF type.

0 | Default (no Pulay, no Camp-King, 3/4 point on unless Pulay or Camp-King, use pseudo-diagonalization). |

1 | 3/4. |

2 | No 3/4. |

10 | No Pulay (DIIS). |

20 | Pulay. |

100 | No Camp-King. |

200 | Camp-King. |

1000 | Use pseudo-diagonalization. |

2000 | No pseudo-diagonalization. |

L405: Flags for MCSCF.

1 | Read options from input stream. |

10 | Use Slater determinants. |

100 | Just list configurations. |

1000 | Use determinant basis with Sz=b/2. |

10000 | Write unformatted file (NDATA) of symbolic matrix elements. |

100000 | Write formatted file of symbolic matrix elements. |

IOp(4/22)

L402: Derivatives to do:

0 | None. |

1 | 1st derivatives. |

2 | 2nd derivatives. |

12 | Restart 2nd derivatives. |

100 | Do 1st derivatives analytically if possible. |

IOp(4/23)

L402: Number of iterations.

0 | Default. |

N | N. |

L405: NDiag.

IOp(4/24)

L402: Whether to update orbitals, eigenvalues, /Mol/, and ILSW on the RWF.

0 | Default (don’t update). |

1 | Update, multiplying by S^{-1/2}. |

2 | Don’t update. (For Opt=MNDOFC). |

3 | Update, but don’t convert from Lowdin orbitals. |

10 | Update second force array instead of first. (For Opt=MNDOFC). |

L405: NRow.

IOp(4/25)

L402: Wavefunction.

0 | Default (Same as 1). |

1 | Single determinant, RHF/UHF from IOp(4/5). |

2 | ROHF (NYI). |

3 | Bi-radical 1/2 CI (only for MINDO3, MNDO, AM1). |

4 | Closed-shell 1/3 CI (only for MINDO3, MNDO, AM1). |

5 | General CI, using specified orbitals. |

-N | General CI, with N microstates read in. |

L405: 10 binary switches.

IOp(4/26)

Whether to mix orbitals in generated guess density.

0 | No. |

-3 | Yes, mix valence occupieds with 0.05 au (according to ZDO) of the HOMO and virtuals within 0.15 au. |

-2 | Yes, mix valence orbitals and an equal number of virtuals. |

-1 | Yes, mix all equally. |

N | Equal occupations of the lowest N virtuals and high N occupieds. |

IOp(4/28)

L402: SCF Convergence (10^{-N}, default 10^{-7}).

IOp(4/29)

L405: Number of core orbitals.

IOp(4/33)

Printing of guess.

0 | No printing. |

1 | Print the MO coefficients. |

2 | Print everything. |

IOp(4/34)

Dump option.

0 | No dump. |

1 | Turn on all possible printing. |

IOp(4/35)

Overlap matrix.

0 | Default (copy on disk is used). |

1 | Overlap assumed to be unity. |

2 | Copy on disk is used. |

IOp(4/36)

ZIndo reformatting.

0 | No. |

1 | Yes, reformat ZIndo integrals and wavefunction into RWF. |

IOp(4/37)

L402: Selection of old MNDO parameters.

0 | Defaults. |

1 | Old Si parameters. |

2 | Old S parameters. |

IOp(4/38)

Generalized density to use for natural orbitals.

0 | Default (-1, current for method on chk). |

N | Density number N. |

IOp(4/39)

Angle for mixing during Guess=Mix.

0 | Default (Pi/4). |

N | Pi/N. |

IOp(4/43)

L402: Handling of background charge distribution.

00 | Same as 21 for MM, 22 for everything else. |

1 | Consider external charges. |

2 | Do not consider external charges. |

10 | Consider self-consistent solvent charges. |

20 | Do not consider self-consistent solvent charges. |

L405: = IDiEij: = switch for direct matrix element calculation. | |

0 | For normal route, with all matrix elements calculated here and stored on disk. Configs printed as normal. |

1 | For direct route. Eij’s calculated here and stored on disk. A flag is automatically sent to L510 to tell it to compute the remaining matrix elements directly. This type of computation can only be done in a CAS comp. Also L510 must use Lanczos. |

2 | Like option 1, but all configurations are printed. This will be the only way to print configs in a direct matrix element calc, since there can be many thousands in a large CAS. |

IOp(4/44)

L405: Prepare input for CAS-MPZ when set to 1.

IOp(4/45)

Ipairs= number of GVB pairs in GVBCAS.

0 | Default. No pairs, normal CAS calculation. |

N | There are N pairs: 2*n extra orbitals and electrons will be added into the active space later. L405 performs a CAS on the inner space, and sets up L510 to compute extra matrix elements etc. implicitly. This is a normal GVBCAS calculation. |

-N | There are N pairs: 2*n orbitals and electrons of the specified CAS are to be considered to be GVB type orbitals when generating configs/matrix elements. L510 will execute normally. This occupies as such space as a full CAS in this link, but is smaller subsequently. This is the GVBCAS test mode. |

IOp(4/46)

CI basis in CASSCF.

1 | Hartree-Waller functions for singlets. |

2 | Hartree-Waller functions for triplets. |

3 | Slater determinants. |

10 | Write SME on disk. |

IOp(4/47)

Convert to sparse storage after generating guess.

-3 | Save sparse storage Fock matrix for guess. |

-2 | Save full storage Fock matrix for guess. |

-1 | No, use the Lewis dot structure to generate a sparse guess directly. |

0 | Default (-1 if sparse is turned on). |

1 | Yes. |

IOp(4/48)

L402: Whether to do (sparse) conjugate gradient methods.

0 | No. |

1 | Yes. Use Lewis dot structure guess density. |

2 | Yes. Use diagonal guess density. |

IOp(4/60)

Override standard values of IRadAn.

IOp(4/61)

Override standard values of IRanWt.

IOp(4/62)

Override standard values of IRanGd.

IOp(4/63)

Flags for which terms to include in MM energy.

0 | Default (111111). |

1 | Turn on all terms, r^{-1} Coulomb. |

2 | Turn on all terms, r^{-2} Coulomb. |

10 | Turn on non-bonded terms. |

100 | Turn on inversions/improper torsions. |

1000 | Turn on torsions. |

10000 | Turn on angle bending. |

100000 | Turn on bond stretches. |

IOp(4/65)

Tighten the zero thresholds as the SCF calculation proceeds.

0 | Default: Yes, initial threshold 5×10-5. |

1 | No variable thresholds. |

N | Yes, initial threshold 10^{-N}. |

N<-100 | Yes, initial threshold 5 x 10 ^{N+100}. |

IOp(4/66)

Dielectric constant to be used in MM calculations.

0 | Eps = 1.0. |

N | Eps = N / 1000. |

IOp(4/67)

Whether to use QEq to assign MM charges.

0 | Default (211 if UFF, 2 otherwise, 1⇒ 221). |

1 | Do QEq. |

2 | Don’t do QEq. |

00 | Default (20). |

10 | Do for atoms which were not explicitly typed. |

20 | Do for all atoms regardless of typing. |

000 | Default (200). |

100 | Do for atoms which have charge specified or defaulted to 0. |

200 | Do for all atoms regardless of initial charge. |

IOp(4/68)

L402: Convergencecriterion for micro-iterations.

0 | Default. |

N | 10^{-N}. |

IOp(4/69)

Whether to do a new additional guess in addition to reading orbitals from the RWF.

0 | Default (2). |

1 | Yes if no Guess=Alter, Harris guess, and not a small geometry step. |

2 | Do not do the extra guess. |

3 | Do the extra guess and store as the initial Fock matrix. |

4 | Do the extra guess regardless. |

5 | Store the normal guess as the alternative (for SimOpt). |

00 | Default (10 for PBC, 20 otherwise). |

10 | Save the Harris guess as an initial Fock matrix. |

20 | Just generate orbitals from the Harris guess. |

IOp(4/71)

L402: Write out AM1 integrals.

0 | No |

1 | Yes |

IOp(4/72)

Irreps to keep in MCSCF CI-wavefunction.

0 | All |

IJKLMNOP | List of up to 8 irreducible representation numbers to include. |

IOp(4/80)

The maximum conjugate gradient step size (MMNN).

0000 | No maximum step size. |

MMNN | Step size of MM.NN. |

IOp(4/81)

Sparse SCF Parameters.

MM | Maximum number of SCF DIIS cycles. (MM=00 defaults to 20 cycles, MM=01 turns DIIS off). |

NN00 | F(Mu,Nu) atom–atom cutoff criterion (angstroms) Mu, Nu are basis functions on the same atom.(defaults to no F(Mu,Nu) cutoff). |

PP0000 | F(Mu,Lambda) atom–atom cutoff criterion (angstroms) Mu, Lambda are basis functions on different atoms. (defaults to 15 angstroms). |

IOp(4/82)

Conjugate-Gradient Parameters.

MM | Maximum number of CG cycles per SCF iteration. (defaults to 4 CG cycles). |

NN00 | Maximum number of purification cycles per CG iteration. (defaults to 3 cycles). |

00000 | Don’t use CG DIIS. |

10000 | Use CG DIIS. |

000000 | Polak-Ribiere CG minimization. |

100000 | Fletcher-Reeves CG minimization. |

0000000 | Use diagonal preconditioning in Conjugate-Gradient. |

1000000 | No preconditioning. |

IOp(4/90)

L402: Step size in dynamics (see IOp(4/8) in L118).

0 | Default (0.025 femtosec). |

N | N*0.0001 femtosec. |

IOp(4/91)

L402: Trajectory type and initial velocity (see IOp(4/9) in L118).

0 | Default (same as 4). |

3 | Read in initial Cartesian velocity. |

4 | Read in initial mass weighted Cartesian velocity. |

IOp(4/92)

L402: Maximum points in one trajectory (see IOp(4/42) in L118).

0 | Default (100). |

N | N points in trajectory. |

IOp(4/93)

L402: Read isotopes for trajectory (see IOp(4/45) in L118).

0 | Do not read isotopes. |

1 | Read isotopes. |

IOp(4/110)

L402: Scaling of rigid fragment steps during micro-iterations.

1 | Scale by (# fragatoms)^{-1}. |

2 | Scale by 1/SQRT (# fragatoms). |

N | Scale by N/1000. |

IOp(4/111)

IDoV in Harris guess. See HarFok for details.

0 | Default (2). |

IOp(4/112)

Compression for ONIOM.

4 | Compressed Hessian over active atoms. For MM calculations on the real system, this converts a second derivative calculation to just forces, since the real system 2nd derivatives are computed during micro-iterations. |

N≥4 | Full storage. (default) |

IOp(4/113)

L402: Which external method to use for ONIOM calculations using different external commands for 2 or more levels.

0 | Default (First external command). |

N | N^{th}external command (command N in file 747). |

IOp(4/114)

Which ONIOM system is being done, which is sometimes needed by external procedures.

0 | Default (1). |

1 | Real system. |

2 | Model system for 2-layer, middle for 3-layer. |

3 | Small model system for 3-layer. |

IOp(4/115)

Mixing of orbitals for GHF/Complex testing.

0 | Default (No, unless generate guess for complex). |

1 | Make MO coefficients complex. |

2 | Don’t rotate real and imaginary components of MOs. |

10 | Mix alpha and beta orbitals for GHF. |

100 | Read in S vector to apply to FC perturbation. |

200 | Read in complex-style SR, SI for GHF. |

0000 | Default FC perturbation (1). |

1000 | FC with MBS core orbitals blanked. |

2000 | Full FC. |

IOp(4/116)

Functional to use in Harris guess.

0 | Default: PBEPBE for HSE2PBE, HSE(H)1PBE and any functional involving the kinetic energy or Laplacian, the pure version of the functional for pure and hybrid GGAs, and SVWN3 for HF. |

N | Functional # (see values in 3/74). |

IOp(4/117)

Set flag for BD Guess=Read.

0 | No. |

-1 | Yes. |

IOp(4/118)

Whether to do GHF/Complex diagonalization for Harris and Core guesses.

0 | Default (1). |

1 | Yes. |

2 | No, generate UHF guess and convert. |

IOp(4/119)

Printing MM energy contributions and force field parameters.

0 | Default (print contributions if #p). | |

1 | Print contributions. | |

2 | Don’t print contributions. | |

00 | Default (20). | |

10 | Print all terms in the force field. | |

20 | Don’t print the force field. |

IOp(4/120)

L402: Number of MM microiterations allowed.

0 | Default, based on N Atoms but at least 5000. |

0 | N. |

IOp(4/121)

Convergence of iterative Harris guess.

0 | Default (0.02). |

N>0 | N/10000. |

N<0 | 10^{N}. |

IOp(4/122)

Maximum number of iterations for iterated Harris:

0 | Default, 20. |

IOp(4/123)

Control of generation QEq charges in Harris guess. See description ICntrl in GenChg.

IOp(4/124)

L402: File options for External.

IOp(4/125)

L402: Options for unformatted i/o file.

IOp(4/126)

L402: IDefCm for External.

IOp(4/127)

Whether to print atomic spin vectors, etc.

0 | Default (2). |

1 | Yes. |

2 | No. |

IOp(4/128)

Whether to print analysis of projection for read-in guesses:

0 | Default (122 if using symmetry in diagonalization, 222 otherwise). |

1 | Yes. |

2 | No. |

10 | Symmetrically orthogonalize core and valence occupieds together. |

20 | Symmetrically orthogonalize core and valence occupieds separately. |

100 | Always project virtuals. |

200 | Only project virtuals for CAS. |

IOp(4/129)

Whether to read energy from chk during Guess=Read (i.e., with SCF=Skip):

0 | Default(No). |

1 | Yes. |

IOp(4/130)

Store dispersion energy and derivatives as total?

0 | Default (No). |

1 | Yes. |

IOp(4/131)

L402: Whether to include charges in MM calculations in.

0 | Default (check ILSW for whether ONIOM or QM/MM-style). |

1 | ONIOM-style, so include. |

2 | Do not include. |

IOp(4/132)

Copy MOs from chk file to reference phase file on rwf. Reference CIS/TD amplitudes are also copied, if found on the chk file.

0 | Default(No). |

1 | Copy. |

10 | Flip sign of MOs. |

20 | Flip sign of amplitudes. |

30 | Flip sign of both MOs and amplitudes. |

Last updated on: 21 October 2016. [G16 Rev. A.03]