L117: Whether to delete points which are too close together and how close to get to the iso surface in search.

0	No; Approx. 1.0 D-6 (N=20)
1	Yes, using a default criterion (0.05 Angstroms).
-N	Yes, using a (10-N Angstroms) criteria: 2.0-N.
N	2.0-N

L101, L106, L108, L109, L110: Input units.

0	Angstroms degrees.
1	Bohrs degrees.
2	Angstroms Radians.
3	Bohrs Radians.

Whether to round tetrahedral angles.

0	Default (Yes).
1	Yes, round angles within 0.001 degree.
2	No.

IOp(1/30)

L103: Are the read-write files to be updated? This option is set for the last call to 103 in frequency calculations in order to preserve the values of the variables for archiving. It also suppresses error termination on large gradients.

0	Yes.
1	No, print internal coordinate step but don’t set up for microiterations and don’t update the RWF.
2	Set up for step but don’t update coordinates; for QM/MM iterative frequencies.

IOp(1/32)

Title card punch control.

0	Don’t punch.
1	Punch.

L116:  Whether to read initial E-field.

0	Start with 0.0.
1	Read from checkpoint file.
2	Read from input stream.

L123: Truncation error threshold for the modified BS integration routine (EPS).

0	Default (10-8 Bohr).
N	10-N Bohr.

Order of multipoles in numerical SCRF.

0	Dipole.
1	Quadrupole.
2	Octapole.
3	Hexadecapole.

IOp(1/53)

L120: Action of each invocation of L120.

0	Do nothing.
1	Set up point MM on RWF from initial data.
2	Set up point MM on RWF from initial data and restore point MM on checkpoint file if ONIOM data is present there.
3	Restore point M from data on the RWF.
4	Integrate energy.
5	Integrate energy and gradient.
6	Integrate energy, gradient, and Hessian.
7	Restore point MM from RWF but do not create a new model system.
NN0	Save necessary information (some RWF’s, energy, gradients, Hessian) of point NN of the ONIOM grid. NN = MaxLev2 + 1 (currently 17) to restore real system.
MM000	Next point to do is MM.
	Calc Level High 4—7—9 |    |   | Mid 2—5—8 |    |   | Low 1—3—6   S  M  L system size

---

IOp(1/54)

---

IOp(1/55)

L103: Options for GDIIS. ICos*1000+IChkC*100+IMix*10+Method form.

L115: IRC optimization.

0	Default, use gradients to find the next geometry.
1	Use displacements to find the next geometry.

---

IOp(1/56)

Set of atom type names to parse.

0	Accept any.
1	Dreiding/UFF.
2	Amber.
3	Amber allowing any symbol, for use with parameters in input stream.

---

IOp(1/57)

Whether to produce connectivity.

0	Default (4 if reading geometry from checkpoint file and connectivity is there, otherwise 3).
1	No.
2	Yes, read from input stream.
3	Yes, generate connectivity.
4	Yes, read from checkpoint file.
5	Yes, read from RWF file.
10	Read modifications.
100	Connectivity input is in terms of z-matrix entries, including dummy atoms.

---

IOp(1/58)

---

IOp(1/59)

L103: Update of coordinates.

0	Default (1).
1	New versions (RedCar/RedQ2X); fall back to ORedCr/RedQX1 if RedCar fails.
2	Old version (ORedCr/RedQX).
3	Old version (ORedCr/RedQX1).
4	New version (ORedCr/RedQ2X) but fall back to ORedCr/RedQX if RedCar fails.
MMMM0	IAprBG in Red2BG.
10	Re-use eigenvectors of G only if exact.
20	Re-use eigenvectors of G if they are linearly independent.
30	Test old eigenvectors of G but don’t re-use them.
40	Don’t look at old eigenvectors.
50	Re-use eigenvectors of G if the RMS of the elements of the new G in the old null space is less than the threshold.
NN00	RMS < 10-N, default 4.
M0000	Default (1).
10000	Form G-inverse from the B eigen-values/vectors.
20000	Form G-inverse directly from G.
30000	Do G- via diagonalization of G (NYI).
40000	Do G- via SVD on B, returning only the eigenvectors with nontrivial eigenvalues.

---

IOp(1/60)

Interpret extra integer and fp values in z-matrix as scan information.

0	Default (No).
1	Yes.
2	No.

---

IOp(1/61)

How ONIOM should leave the RWF at the end of each geometry.

0	Default (1).
1	Normal: leave the RWF set up for the low-level calculation on the real system.
2	MOMM: leave the RWF set up for the real system, but with NBasis and NBsUse for the high-level calculation on the model system. Useful for treating the full system as having electrons only on the QM atoms. This is really a hack for two layer QM:MM ONIOM ADMP and should probably be generalized to behave like an ONIOM-PCM-A case.
3	Lowest level is MO, but normal setup at end.
00	Default (10, but 20 if doing EE microiterations).
10	Leave the charges file (605) from the best calculation that produced one.
20	Leave 605 in its normal state (present if from real, low-level).

---

---

IOp(1/63)

Step in counterpoise calculation.

LMM	L: order of derivatives: 1 = Energy, 2 = Gradient,…
	MM: 0 = Supermolecule, 1-NFrag = Fragments with ghost atoms,
	NFrag+1-2*NFrag = lone fragments.

---

IOp(1/64)

Molecular mechanics force field selection.

0	None.
1	Dreiding.
2	UFF.
3	AMBER.
000	Use only hard-wired.
100	Use soft and hard-wired, hard-wired has priority.
200	Use soft and hard-wired, soft has priority.
300	Use only soft. Lowest 2 digits then have no meaning.
0000	Do not read modifications to parameter set.
1000	Read modifications to parameter set.
00000	With soft parameters, abort when different parameters match to the same degree.
10000	Use the first when there are equivalent matches.
20000	Use the last when there are equivalent matches.

If IOp(67) = 3, then the default is to apply soft parameters with higher priority.

---

IOp(1/65)

Control of which terms are included in MM, corresponding to the ‘classes’ in FncInf.

0	Do all (default).
1	Non-bonded.
10	Stretching.
100	Bending.
1000	Torsion.
10000	Out-of-plane.
100000	Stretch-bend.

---

IOp(1/66)

Whether to generate QEQ charges, over-write the values in AtChMM, or to use the values already there.

0	Default (2, 1 ⇒ 221).
1	Do QEq here in GenChg.
2	Don’t do QEq.
3	Generate charges here using QEqN.
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.
MMMMM000	IType passed to QEq.

---

IOp(1/67)

Source of MM parameters.

0	Default: 2 if reading geometry from checkpoint file, else 1.
1	Generate here, reading from input if requested by IOp(64).
2	Copy from checkpoint file.
3	Pick up non-standard parameters from checkpoint file.

---

IOp(1/70)

L118: Type of sampling (Nact).

0	Default (same as 3).
1	Orthant sampling.
2	Classical microcanonical normal mode sampling.
3	Fixed normal mode energy.
4	Local mode sampling (thermal sampling based on RTemp).
Currently 0, 2, 3 and 4 are working.
10	Read in Hessian from checkpoint for initial sampling.

L124: SCRF flag.

IJKLLMM	See comments in overlay 3.
00000000	Default (same as 2).
10000000	Do the 1st iteration in gas-phase.
20000000	Do the 1st iteration in solution.

---

IOp(1/71)

L103: How often to calculate analyic 2nd derivatives.

0	Default (every cycle if IOp(10)=4).
N	Every Nth geometry, starting with the initial one.

L115: Whether to print out input files for each structure along an IRC.

0	No.
1	Yes.

L118: Hessian update.

-1	Gradient only.
-2	Gradient+Hessian, but never calculate full H (only updates).
0	Full Hessian at every step.
NN	Try to do NN updates between full Hessians.
000	Default updating (same as 300).
100	SR1 Hessian updating algorithm.
200	PSB Hessian updating algorithm.
300	Bofill’s Hessian updating algorithm.
400	Sqrt(Bofill) Hessian updating algorithm.
500	No update.
0000	Default (same as 1000).
1000	Reintegrated updated steps.
2000	Suppress reintegration.

L123: Hessian calculation.

-1	Gradient only.
0	Either only update or CalcAll as determined by IOp(10).
N	Recalculate analytic Hessian every Nth calculation.

---

IOp(1/72)

L121: Lagrangian constrain method for ADMP (ICType).

Half*Gamma*Tr[(P*P-P)2] + Lambda*[Tr(P)-Ne] + Eta*Tr(P*P-P)

0	Default same as 7 if no mass-weighting (IOp(76) < 0). Same as 10 if mass-weighting (IOp(76) > 0).
1	Use Lambda and Eta only. (Gamma=0).
2	Use Lambda, Eta, Gamma. Gamma = .2.
3	Use Lambda, Eta, Gamma. Gamma = 1. Constraints for scalar mass case.
4	Use exact constraint Sum(ij)[Vij*(P2-P)ij].
5-7	Iterative Scheme same as 4. Different initial guesses. 7 is default for scalar mass case. Constraints for tensorial mass.
8-11	Mass-weighting constraints. Documentation may be found in DVelV1. 10 is default.

L124: Solvent type for external iteration PCM. (ISolv).

---

IOp(1/73)

L123: Whether to compute projected frequencies: 0/1/2 Default (No)/No/Yes.

L118 and L121: Initial Kinetic energy of the Nuclei (EStrtC).

0	Default (.1 Hartree).
N>0	N*micro-Hartree.
N<0	0.0 Hartree

---

IOp(1/74)

Charge scaling for charge embedding in ONIOM. IJKLMN 6th through 1st nearest neighbors of current layer scaled by I*0.2, J*0.2, etc. 0 Þ 5³ IAtTyp=6 (no scaling); all layers are scaled by at least as much as ones farther out. The default is 500.

M	Factor for charges one bond away from link atom.
K00	Factor for charges three bonds away from link atom. IJ etc.
	The actual factors used for each value of IAtTyp are:
	1: 0.0   2: 0.2 3: 0.4   4: 0.6 5: 0.8   6: 1.0

---

IOp(1/75)

ADMP control flag (ICntrl).

0	Standard ADMP.
1	Read converged density at every step.
2	Fix the nuclear coordinates.
3	Test time reversibility (MaxStp must be even).
00	Default (20).
10	Read stopping parameters from input.
20	Do not read stopping parameters.

---

IOp(1/76)

Fictitious electron mass (EMass). Format of input: +/- XXXXZYYYY.

YYYY	Default (1000)
	IOp(1/76) > 0  YYYY*.0001 amu. MW core functions more than valence functions.
	IOp(1/76) < 0  YYYY*.0001 amu. Use uniform scaling for all basis functions.
	(Note YYYY > 9999 makes no sense).
Z	Mass-weighting option. If IOp(76) < 0, Z is meaningless.
XXXX	If PBC: Mass of Box Coordinates (BoxMas) = XXXX*.0001 AMU BoxMas=0
	Box coordinates not propagated (default).

---

IOp(1/77)

(For UHF, Alpha and Beta each get half this energy).

0	Default (0.0 Hartree).
N>0	N*micro-Hartree IWType is used to figure out how the initial velocity is computed (in gnvelp). If XXYYYY < 0: Initial velocity = 0.0 Hartree (i.e., currently same as N=0 above).

---

IOp(1/78)

L121: Sparse.

-N	Sparse here with cutoff 10-N, full elsewhere.
0	Use full matrices or sparse based on standard settings.
1	Use sparse fixed form.

---

IOp(1/79)

L115:  IRCMax convergence.

L118, L121: Stopping Criteria.

0	Default, do not analyze pure rotation and vibration for polyatomic molecules.
1	Do pure rotational and harmonic normal mode analysis for polyatomic molecule; EBK theory for diatomic vibrational analysis (require equilibrium information for each of the polyatomic molecules from saved checkpoint files and Morse parameters for diatomic molecule).
2	Do pure rotational and harmonic normal mode analysis for polyatomic molecule; local harmonic vibrational analysis for diatomic molecule (require equilibrium information for each of the fragments from saved checkpoint files).

---

IOp(1/80)

L106: 0/1 Cartesian/Normal mode/ Internal coordinate differentiation. 2 is NYI.

L118: = 1 to suppress the 5th order correction after surface hop has been made in Trajectory Surface Hopping calculations. Needs also IOp(10/80 = 1)

L121: Nuclear Kinetic Energy Thermostat Option (currently only velocity scaling is implemented).

0	No Thermostat.
11XXXXX	Velocity scaling, but only for the first XXXXX simulation steps. (This option is useful, if thermostating is only required during equilibration.)
1000000	Velocity scaling, all the way through the simulation.

---

IOp(1/81)

Nuclear KE thermostat in ADMP — temperature is checked and scaled every IOp(81) steps.

---

IOp(1/82)

L121: Temperature for nuclear KE thermostat.

---

IOp(1/83)

Whether to read in frequencies for electric and magnetic perturbations.

0	Default (No).
1	Yes unless geom=allcheck.
2	No.
3	Yes, even if geom=allcheck.
00	Default (10).
10	Ensure that the static case is done, by putting a zero frequency at the beginning.
20	Do not put a zero frequency at the beginning.
000	Default (100).
100	Sort the frequencies into increasing order.
200	Do not sort the frequencies (for debugging).

---

IOp(1/84)

Differentiation of frequency-dependent properties.

0	No.
N	Mask for which properties on file 721 will be differentiated.

---

IOp(1/85)

---

---

IOp(1/87)

ONIOM integration of density.

0	Do not integrate.
1	Integrate current densities.
2	Integrate densities specified by following digits:
K0	Density to use from gridpoint 1
L00	Density to use from gridpoint 2
M000 etc.
Values for K, L, M, etc.:
	0: SCF
	1: MP first order
	2: MP2
	3: MP3
	4: MP4
	5: CI one-particle
	6: CI
	7: QCI/CC
	8: Correct to second order

---

IOp(1/88)

---

IOp(1/89)

Maximum allowed deviation from average nuclear KE during ADMP, in Kelvin.

---

IOp(1/90)

To read in the velocity in Cartesian coordinates.

---

IOp(1/91)

Nuclear Kinetic Energy Thermostat Option. Average energy (in micro-Hartree) to be maintained during simulation, as required by IOp(80).

---

IOp(1/92)

Thermostat Option. Maximum allowed deviation from average nuclear KE specified in IOp(81). Also in micro-Hartree.

---

IOp(1/93)

QM/MM TS vector guess.

000	Default (112)
1	Retrieve from checkpoint file if available, otherwise diagonalize QM Hessian or read from input.
2	Do not try to retrieve from checkpoint file.
10	Diagonalize QM contribution to Hessian.
20	Read from input.
N00	How to deal with ‘suspicious RFO solutions’ (default is 1).
	1. Just take the step.
	2. Check if there is an eigenvector with wrong curvature. If there is, flip its eigenvalue.
	3. Check if there is an eigenvector with wrong curvature. If there is, take a small step into this direction, followed by a linear search. This should step out (or stay in) the wrong region, and fix the eigenvalue.

---

---

IOp(1/95)

RFO/Davidson control for quadratic micro-iterations.

MM	Convergence (7).
LLL00	<0: Regular Davidson.
	0: Only check convergence on first vector, and iterate the lowest root only. Use all the intermediate vectors.
	>0: Only check convergence on first vector, and iterate the new vectors LLL times with the explicit last row/column. This is specifically appropriate for RFO. The last row/column of the Hessian comes after the diagonal elements.

---

IOp(1/96)

Options for generating initial guess vectors for RFO/Davidson diagonalization in coupled QM/MM macro steps. Note that other RFO/Davidson diagonalization controls for coupled QM/MM macro steps are available in IOp(97). Format of input: GHIJKLLMM.

MM	Number of initial guess vectors to get from CG steps. The default is 0.
LL	Number of initial guess vectors from the diagonal of the QM block (4). The default is 4.
K	Add 0,…,0,1 guess vector?
0	Default: K = 1;
1	Yes
2	No
J	Add the gradient vector to the guesses?
	0 Default: J = 1 1 Yes 2 No
I	Pre-diagonalize a Hessian/RFO matrix without non-bonding contributions? Note that this control is only valid for IOp(98) > 3; otherwise, I is ignored.
	0 Default: I = 1 1 Yes 2 No
H	Scale factor for the size of the Davidson sub-space in early iterations.
	0 Default: H = 4 1 Same as no scaling. H Use a sub-space in early iterations that is H times the number of requested vectors.
G	Number of vectors to solve using Davidson diagonization.
	0 Default: G = 1 G Solve for G vectors.

---

IOp(1/97)

RFO/Davidson control for coupled QM/MM macro step. Note that other RFO/Davidson diagonalization conrols for coupled QM/MM macro steps are available in IOp(96). Format of input: GHIJKLMM.

MM	Convergence in Davidson iterations. Convergence is set to 10-MM. The default value is MM=5.
L	What is being diagonalized? This option is set explicitly in subroutines before calling the Davidson diagonalization code. Therefore, the value set in this IOp is ignored and serves only as a place holder.
0	the Hessian
1	the augmented-Hessian/RFO matrix
K	Check convergence on which roots?
0	Default: For L=0, K=2; For L=1, K=1
1	Check convergence on lowest root only.
2	Check convergence on all roots.
J	Appears to be unused.
I	Number of Davidson iterations to store.
0	Default: Keep all iterations.
1	Keep only the last iteration.
H	Number of new vectors to create in each Davidson iteration.
0	Default: For L=0, H=1; For L=1, H=2.
1	Iterate all roots/vectors.
2	Iterate lowest root/vector only up to the maximum number of iterations that are default in the Davidson code (ignores J above) and keeping vectors from all iteratations (ignores I above).
3	Iterate lowest root/vector only. Note that this option is essentially the same as H=2, though J and I option settings are honored.
G	Initial approximation to use for Davidson diagonization.
0	Default: G=2.
1	Use a diagonal Hessian approximation together with the gradient vector. This is best used in RFO applications
2	Use the inverted Hessian for the QM block
3	Use a diagonal Hessian approximation.

---

IOp(1/98)

Control of quadratic micro-iterations and coupled QM/MM quadratic macro step.

<0	Do not use dynamic convergence criteria for the micro-iterations.
0	Default (11).
1	Regular non-coupled macro step.
2	Coupled macro step, full diagonalization.
3	Coupled macro step, direct /w full Hessian in core.
4	Coupled macro step, direct /w MM Hessian in core.
5	Coupled macro step, fully direct.
6	Go through the QuadMac machinery, but use the fully integrated ONIOM Hessian to calculate the Hessian-vector products. Switch to regular micro-iterations at points without analytic second derivatives.
7	Fully quadratic at 2nd derivative points (1st in CalcFC, all in CalcAll), QuadMac with integrated Hessian at non-2nd derivative points.
10	Regular micro-iterations.
20	Quadratic micro-iterations, full diagonalization.
30	Quadratic micro-iterations, direct with prepared Hessian in core.
40	Quadratic micro-iterations, direct with raw MM Hessian in core.
50	Quadratic micro-iterations, fully direct.
60	No micro-iterations.

---

IOp(1/99)

Accuracy used for the non-bonded interactions in the Hessian-vector product calculations in the fully direct algorithms. Setting this IOp does not affect the MM energy and gradient calculations; only the direct evaluation in the QuadMac optimization step. When IOp(99) < 0, the full accuracy is used.

MM	Maximum multipole level (8)
LLL00	Box size in FMM (12)
KKK00000	Cutoff in van der Waals (30)

---

IOp(1/101-104)

L115, L118: Phase control in N1, N2, N3, N4.

---

IOp(1/105)

Reaction direction.

Damped-Velocity Verlet (DVV) options for Dynamic Reaction Path following.

00	Default (Same as 10).
10	Forward direction.
20	Reverse direction
0	Default (Same as 2).
1	Use DVV.
2	Do not use DVV.
00	Default (Same as 10).
10	Follow the rxn path in the forward direction
20	Follow the rxn path in the reverse direction.
000	Default (Same as 200).
100	Time step correction not used.
200	Time step correction used but not to recalculate current DVV step.
300	Time step correction used and current DVV step recalculated.
0000	Default (Same as 1000).
1000	Use DVV stopping criteria.
2000	Do NOT use DVV stopping criteria

---

IOp(1/106)

Damping constant for DVV Dynamic Reaction Path following (v0).

0	Default v0 = 0.04 (N=400).
N	v0 is set to N*0.0001.

---

IOp(1/107)

Error tolerance for DVV time step correction (Error).

0	Default Error = 0.003 (N=30).
N	Error = N*0.0001.

---

IOp(1/108)

Gradient magnitude for DVV stopping criteria (Crit1).

0	Default (N = 15).
<0	Turn off this test.
N	N*0.0001

---

IOp(1/109)

Force-velocity angle for DVV stopping criteria (Crit2).

0	Default (90 Degrees).
<0	Turn off this test.
N	Use N Degrees.

---

IOp(1/110)

Scaling of rigid fragment steps during microiterations.

0	Do not scale.
1	Scale with 1/NRA (NRA = number of atoms in fragment).
2	Scale with 1/Sqrt(NRA).
-n	Scale with 1/n.

---

IOp(1/111)

---

IOp(1/112)

L101: Temperature for thermochemistry.

0	Default (standard temperature, unless read in).
N	N/1000 degrees.
1	Default.
-N<1	N/1000000 degrees.
-N	N/1000000 degrees.

---

IOp(1/113)

Pressure for thermochemistry.

0	Default (1 atmosphere, unless read in).
N	N/1000 atmospheres.
-1	Default.
-N<1	N/1000000 atmospheres.

---

IOp(1/114)

Scale factor for harmonic frequencies for use in thermochemistry and harmonic vibration-rotation analysis.

0	Default (1 unless specified by IOp in overlay 7 or read in).
-1	Force to 0 (default).
N	N/1000000.

---

IOp(1/115)

Compression for MOMM quadratic steps.

4	Second derivatives generated over active atoms, with real system terms done iteratively during micro-iterations.
N¹4	Full second derivative matrices are used.

---

IOp(1/116)

Options for ONIOM Conical intersections: which calculations have adiabatic couplings done.

0	Default (111 all component calculations).

---

---

IOp(1/119)

Control Initial Lanczos Vector (ILzVec).

-1	Read guess by card in input file.
-2	Use the largest elements of H as a guess.
-3	Use the five largest contributions of H as a guess.
0x	For Opt, IRC, dynamics read guess from previous cycle.
1x	For Opt, IRC, dynamics generate a fresh guess for each cycle

---

IOp(1/120)

Flags for QM:QM embedding. NOTE: The standard embedding flags must also be set in the same way as necessary for QM:MM embedding calculations.

0	Default – Same as 1.
1	Use Mulliken charges.
-1	Use the nuclear charge stored in array AtmChg.
-2	Set the charges to zero.
00	Default (Same as 20).
10	Just use the charges that already reside in AtChMM.
20	Update AtChMM using current atomic charges on the RWF.

This option is only employed immediately following low-level real-system sub-calculations.

---

IOp(1/121)

---

---

IOp(1/123)

Version of /Mol/ to save on disk.

0	Default (current, version 2).
1	Version 1 (flag -12345).
2	Version 2 (flag -12346).
N<0	Flag value N.

---

IOp(1/124)

Flavor of ONIOM-PCM to use.

1	“A,” reaction field from the ONIOM integrated density;
2	“B,” reaction field from the real system low level;
3	“C,” reaction field in the real system low level only;
4	“X,” reaction field in all subcalculations using the real system cavity.
0x	Default (same as 1 unless a semiempirical method is involved);
1x	Integrate the density for ONIOM-PCM-A;
2x	Integrate the potential for ONIOM-PCM-A;
1xx	Flag to indicate ONIOM-PCM-X as first iteration of ONIOM-PCM-A.

---

IOp(1/125)

---

IOp(1/126)

Whether to read an input section with atom opt/freeze information.

0	Default (2).
1	Yes.
2	No.

---

IOp(1/127)

Use of MM coordinates and forces in GEDIIS:

0	Default (-3 for ME, -3 for EE).
-3	Don’t pass MM info.
-2	Zero MM forces and MM part of step, equivalent to not passing MM info.
-1	Zero MM forces but interpolate step in MM coords.
N	Use MM forces scaled by 1/N.

---

IOp(1/128)

Initial micro-iterations in L120 before first QM step, and micro-iterations in L120 during numerical differentiation in L103.

0	Default (No).
1	Yes.
2	No.

---

IOp(1/129)

---

IOp(1/130)

Eigenvector number to be followed during coordinate driving jobs (IOp(1/19 = 10)).

0	Default (1).
N	Follow input structure Hessian eigenvector number N.

---

IOp(1/131)

Options for corrector integration in predictor-corrector IRC calulcations (Link 123).

00	Should the corrector integration scheme be run in an (macro-cycle) iterative fashion? Default = 2.
-01	After each corrector integration, evaluate the actual energy and derivatives, but do not actually use these. The final IRC will be the same as 1.
01	Never check convergence of the corrector integration. Always do one corrector integration for each predictor integration.
02	Always check for convergence of the corrector integration end point. Convergence is acheived when the change in corrector integration end point geometry is less than the convergence criteria indicated by IOp(7).
03	Same as 2, but never accept convergence after the first corrector integration at a point.
10	This flag iteratively refines the DWI fitted surface if multiple corrector integration macro-cycles are taken by adding the largest standard deviation point from the previous BS cycle.
-11	This flag forces a PES evaluation step after each corrector integration. This is similar option -1, except that the actual energy and derivatives are used for the next predictor step rather than the values on the DWI fitted surface at the corrector end-point.
0xx	Should DWI surfaces employ numerical thrid-order terms in Taylor series? Default = 1.
1xx	Use DWI surface with Taylor series expansions truncated at second-order.
2xx	Use DWI surface with Taylor series expansions truncated at third-order using numerical third-derivatives.
Nxxx	What power should be used for DWI weights that include delta-x raised to an even power. The value of this option setsthat power to 2*N. Default = N = 1.
0xxxx	How should the first step off of the transition structure point be handled in corrector integration cycles? Note, in all cases, the transition vector is used to define the tangent at the transition structure. Default = 2.
1xxxx	Run the requested number of Euler steps in the standard way. Only the first Euler step taken uses the transition vector.
2xxxx	Take a large step off of the transition structure point along the transition vector. This step is taken to be half of the total requested step size given by TotStp.
3xxxx	This the same as option 2 in concept. The only difference is that the first step off of the transition structure is taken as one-third the total requested step size given by TotStp.
4xxxx	This the same as option 2 in concept. The only difference is that the first step off of the transition structure is taken as one-fourth the total requested step size given by TotStp.
0xxxxx	Should update vectors be used in DWI fits if possible? Default = 1.
1xxxxx	Yes, when possible.
2xxxxx	Never.

---

IOp(1/132)

Whether to check for divalent link atoms in ONIOM input:

0	Default (yes).
1	Yes.
2	No.

---

IOp(1/133)

---

IOp(1/135)

MM non-bonded switching function.

---

IOp(1/136)

MM van der Waals outer cutoff in Angstroms.

---

IOp(1/137)

MM Coulomb outer cutoff in Angstroms.

---

IOp(1/138)

MM soft cutoff range (applied to both vdW and Coulomb, in Angstroms.

---

IOp(1/139)

Number of MM microiterations allowed.

0	Default, based on NAtoms but at least 5000.
N	N

---

IOp(1/140)

Whether to restore the real system from chk file:

0	Default (yes if ONIOM).
1	Yes.
2	No.

---

---

IOp(1/142)

Whether to copy MM charges to link atoms:

0	Default (3 if QEq is done; otherwise 1).
1	Copy if link atom charge is zero.
2	Do not copy.
3	Always copy.

---

IOp(1/143)

Hessian during IRC restarts.

0	No change in when Hessian is done.
1	Do Hessian at first new point in each direction.

---

IOp(1/144)

Whether to analyze residue geometry.

0	Default (Yes, if secondary structure present and N Atoms ≤ 10000).
1	Yes.
2	No.

---

IOp(1/145)

Controls for the new internal coordinate data structure.

0	Should the new internal coordinate data structure be set-up for use? (Default = 1).
1	No.
2	Yes. The coordinates are either generated by default (e.g., Option 9×2) or obtained from the user-provided input (e.g., Option 6×2). Alternatively, they can be taken from a chk file if they exist there and if geom=check without any GIC keyword (i.e., IOp(1/145=0) is used.
3	Yes, but do two calls to CrdDef when building the coordinate definitions. The first call adds the coordinates in the same way that Option 2 does. The second call adds coordinates from a list of user-provided input coordinates in the same way as Option 6×2 does. In other words, there will be two sets of coordinates to be merged. The first set is generated by default (e.g., in the case of Option 9×3 without geom=check) or it comes from a chk file (Option 9×3 with geom=check). The second set is provided by the user as a separate input section, and it contains some additional coordinates or modifications to the first set. A blank line can be used instead of the second set’s input section. It is also possible to make the user-provided input’s reading to be the first step and the construction of complementary coordinates by CrdCon to be the second step. The latter is Option 6×3.
0x	Which coordinate derivative terms should be included in the data structure? (Default = 3).
1x	Include only values in the definitions list.
6xx	Read the coordinate definitions from the input file using the symbolic form, (See Routine CrInp1), and then automatically differentiate the active coordinate derivatives as needed.
7xx	Define (the full set of) distance matrix coordinates. This option can be combined with 2xxx to use inverse distance matrix coordinates.
8xx	Read the coordinate definitions from the input file using the OLD symbolic form (See Routine CrDfSy), and then automatically differentiate the active coordinate derivatives as needed.
9xx	Use the new coordinate generation code to create an intermediate RP structure based on the molecular connectivity (See Routine CrdCon), then have CrdDef build the data structure using the intermediate RP data from CrdCon and build derivatives wrt Cartesian coordinates.
0xxx	When building the internal coordinate definitions, should any systematic modifications be done? (Default = 1).
1xxx	No. Simply define the coordinates to be the same as they would have been using the old coordinate code. 2xxx … Invert all bond stretch coordinates.
2xxx	Invert all bond stretch coordinates.
0000000	Default for default is 1 (GIC).
1000000	Default to 1 if not specified.
2000000	Default to 2 if not specified.

---

IOp(1/146)

Control of SCVS:

0	Default (03045011).
1	Include forces in virial ratio.
2	Do not include forces in virial ratio.
1x	Use Murdoch’s extrapolation.
2x	Do not use Murdoch’s extrapolation.
Nxx	Apply SCVS when max force on nuclei is below 10-N.
Mxxx	The convergence on the virial Eta. Default is 5.
Lxxxx	The convergence threshhold on |E| * (2*Eta – 2), where |E| is the magnitude of the kinetic energy is 10-M. Default is 3.
Kxxxxx	The initial order of extrapolation is 10L
JJJxxxxxx	Maximum of KKK SCVS iterations.

---

IOp(1/148)

L106: Storage of derivatives.

0	Default (221).
1	Normal storage of numerical first derivatives.
2	Store numerical first and diagonal second derivatives.
10	Print differences between numerical and analytic quantities.
20	Do not print differences.
N00	Threshold for printing differences is -6-N.

---

IOp(1/150)

L112: Initial scale factor.

0	Default (1.0)
N	1 + N/100000000

---

IOp(1/151)

How many vectors with negative curvature to use in downhill step during minimization:

0	Default (3)
-1	None, do regular RFO step.
N	Up to N vectors.

---

IOp(1/152)

L103: Control of MaxStp (allocated max number of steps in L103).

0	Default: compute based on number of variables, NStep, etc.
N>0	Make MaxStp at least N.
N<0	Make MaxStp at least -N.

---

IOp(1/153)

L103: Whether to fill OT file.

0	Default: do so for Cartesian and redundant internal coordinate optimizations.
1	Yes
2	No

---

IOp(1/154)

Linear angle test during internal coordinate generation.

0	Default (15 degrees, applied to all 3 tests).
-N	Threshold N degrees, applied to the angle only.
N	Threshold N degrees, applied to all 3 tests.

---

IOp(1/155)

Number of steps to take in guessing a TS during QST2:

0	Default (10).
N	Divide the overall step into N increments.

---

---

IOp(1/157)

Maximum step when going down an eigenvector:

0	Default (0.6)
N	N/1000

---

IOp(1/158)

L124: PCM defaults; L301: Details.

---

---

IOp(1/160)

Type of fragment calculation:

0	Default (11).
1	Force use of new ghost atoms in counterpoise.
2	Force use of old ghost atoms in counterpoise.
10	Counterpoise.
20	Fragment guess.
80	Excitation energy transfer.
0xxx	Default (same as 1).
1xxx	Use full-system PCM cavity for fragments.
2xxx	Use fragment PCM cavity for fragments.

---

IOp(1/161)

L123: Hack to save results at each IRC point.

0	Default (No).
1	Yes.

---

IOp(1/162)

Frequency of analytic Hessians during IRC corrector cycles.

---

IOp(1/163)

Copy of external input section from chk file:

0	Default; copy if Geom=AllCheck.
1	Copy regardless.
2	Do not copy.

---

Each pair is given on one line of the form: IA,IB,R0,A,B,ROn,ROff,C(n),C(n-1),…,C(0).

The potential is of the form P(X)*Exp(A*X)*Exp(B*X*X)*Sw(Rab,ROn,ROff).

If A is zero, the exponential is omitted. If B is zero, the Gaussian is omitted, and if ROn is zero, the Stratmann-Scuseria switching function Sw is omitted. P(X) is the polynomial of degree n given by the specified coefficients. n is determined from the # of C’s provided and can be 0 for a constant scale factor.

---

IOp(1/165)

Convergence of MM microiterations.

0	Default (10x tighter than macro, except 10x for FOSimult with MM included).
N	Nx tighter.

---

---

IOp(1/167)

Version of GEDIIS.

0	Default (2).
1	Old.
2	New.

---

IOp(1/168)

---

IOp(1/170)

Control for selecting the initial conditions.

1	Initial Cartesian coordinates and velocities are read-in from the input file. This data is read in a free-format fashion.
2	Initial Cartesian coordinates and mass-weighted velocities are read-in from the input file. This data is read in a free-format fashion.
3	Initial coordinates are given by the user-input geometry; initial velocities are determined by selecting a random velocity which gives an kinetic energy equal to the value set in IOp(73).

---

IOp(1/172)

L101: Whether to print internal coordinates in L101.

0	Default (No).
1	Print if present.

---

IOp(1/173)

L106: Extra files to differentiate.

-11	Differentiate the files related to ground-state post-SCF 2nd derivatives (Lagrangian, P, and W).
-10	Differentiate the files related to CIS/TD second derivatives (Lagrangian, P, W, and T).
-2	Read two lists, each terminated by a blank line, the first the files to differentiate and the second the files where the numerical derivatives will be stored (0 for items which will not be saved).
-1	Read a list from an input section, terminated by a blank link. The numerical derivatives will be printed but not stored on disk.
N>0	Do file number N, storing derivatives in file 795.

---

IOp(1/174)

Reading parameters for classical electrostatics and dispersion.

0	Default, 3 if Geom=AllCheck else 2.
1	Read selected parameters for elements from input.
2	Use default parameters.
3	Read user-specified parameters from chk file if present.
4	Read both user-specified and default parameters from chk file if present.

---

IOp(1/175)

Electronic embedding type:

0	Default (1)
1	ONIOM style.
2	QM/MM style.

---

---

IOp(1/179)

Use of subprocesses for energy+derivatives, for debugging:

0	Default (1002).
1	Yes.
2	No.
3	Yes, using loop over steps in L106.
4	Yes, using loop over calls to utility in L106.
5	Yes, using Linda.
1x	Debug output and save of subprocess files.
1xx	Full copy of rwf back and forth.
1xxx	Iterate purturbations forward.
2xxx	Iterate purturbations backward.

---

---

IOp(1/181)

Maximum number of middle iterations during ONIOM-EE:

0	Default (40).
N	N.

---

Last updated on: 21 October 2016. [G16 Rev. C.01]