Full Parameter List

This page is a comprehensive reference for TeraChem input-file (.in) keywords, grouped by topic. In a value list, the character | should be read as "or": e.g. yes|no means enter one of yes or no.

Authoritative source

The dedicated topic pages (SCF, DFT, AIMD, QM/MM, PCM, etc.) are the authoritative reference for their keywords. Where this catch-all table and a topic page disagree, follow the topic page (and ultimately the parser, params.cpp).

General & molecule setup

Parameter	Description	Default
jobname	Name of the job, used to name output files. If unset, deduced from the basename of the coordinates file (e.g. caffeine.xyz → caffeine).	autodetected
scrdir	Scratch directory (most output files go here).	./scr
keep_scr	yes|no — keep old scratch directories from a previous run.	no
genscrdata	yes|no / true|false — generate data in the scratch folder (molden, c0, etc.).	true
save_sao	yes|no / true|false — save the AO overlap matrix to $scrdir/sao.	false
run	Calculation type: energy (single-point energy), gradient (energy + gradient), project (wavefunction projection for a better guess), minimize (geometry optimization), ts (transition-state search), md (Born–Oppenheimer MD), initcond (generate MD initial conditions), plumed (metadynamics with PLUMED), numgradient (numerical gradients), frequencies (numerical frequencies), tdci (time-dependent CI), exciton (exciton model), ci_vec_overlap (CI vector overlaps), ciopt (conical-intersection search).	energy
gpus	Number of GPUs to use in parallel, optionally with explicit ordering. gpus 3 uses the first three (order 0 1 2); gpus 3 1 3 2 selects specific devices (numbering from 0) — useful when device 0 is reserved for display.	all
coordinates	Name of the file containing atomic coordinates (.xyz or .pdb).	not set
charge	Total molecular charge (integer). May instead be protein to determine the charge from PDB protonation states (requires standard 3-letter residue codes).	not set
spinmult	Spin multiplicity (2S+1) of the wavefunction (integer).	1
units	Units of the input coordinates: angstrom|bohr.	angstrom
basis	Basis set. Built-in sets live in instdir/basis; custom sets are supported. Examples: sto-3g, 6-31++G**, 6-311++G(3df,3pd), aug-cc-pvdz, mini(s), midi!. On parsing, * becomes s and parentheses become brackets (so 6-31G** and 6-31Gss are equivalent).	not set
auxbasis	Auxiliary basis set; if set, two-electron integrals use the density-fitting approximation.	not set
sphericalbasis	yes|no — use a spherical-harmonic basis representation. If off, Cartesian is used.	no
max_l_in_basis	Highest angular momentum to keep in the basis (0|1|2|... for s, p, d, …) — lets you neglect high-angular-momentum polarization functions.	not set
projectbasis	In project jobs, the converged wavefunction is projected onto this basis and stored in scr/prjct (scr/prjcta/scr/prjctb for unrestricted).	not set

QM/MM (point charges & water)

Parameter	Description	Default
qmmm	File of MM water coordinates; its presence triggers a QM/MM calculation.	not set
water	Water model for QM/MM: tip3p|spc|spc/e|tip4p.	tip3p
rigidbonds	yes|no / true|false — rigid O–H bonds for MM water in QM/MM.	true
pointcharges	filename|no — add point charges from a file. The file lists q x y z per charge (charges in a.u., coordinates in Å).	no
pointcharges_self_interaction	yes|no / true|false — include the self-interaction of MM point charges in the energy and forces.	true
direct_reaction_field	yes|no — IEDRF polarizable embedding (OpenMM driver only; needs the DRF2EPOL build). See DRF page.	no
same_site_exact	Evaluate same-site two-electron DRF integrals exactly (yes) or via resolution of identity (no).	yes
cutoff_exponent	Damping exponent \(\alpha\) (bohr\(^{-2}\), > 0) in \(C(r) = [1-\exp(-\alpha r^2)]^q\).	4.0
cutoff_power	Damping power \(q\) (\(\geq 2\)); GPU evaluation restricts to compiled-in values.	2
cpp_integrals	DRF polarization integrals on gpu (fast) or cpu (any cutoff_power).	gpu
save_induced_dipoles	Write per-state induced dipoles (Debye) to scr/induced_dipoles_I.dat.	no
mm_ecp	Basis file whose ECP block is read and placed on the MM atoms (e.g. mm_ecp_marefat2020). Recommended with DRF to suppress electron spill-out.	not set
ecps_on_mm_link_atoms	Keep MM-ECPs on MM atoms bonded to the QM region.	no

Method & DFT

Parameter	Description	Default
method	SCF wavefunction / DFT functional. Restricted examples: rhf, svwn, bhandhlyp, blyp, b3lyp, b3lyp1, b3lyp5, b3p86, b3pw91, pbe, pw91, revpbe, pbe0, revpbe0, wpbe, wpbeh, bop, mubop, camb3lyp, b97, wb97, wb97x, wb97xd3. b3lyp/b3lyp1 use VWN1 correlation (Gaussian's B3LYP); b3lyp5 uses VWN5 (GAMESS's B3LYP). Add a u prefix for unrestricted (uhf, ublyp, …) or ro for restricted-open-shell (rohf, rob3lyp, …).	rhf
hfx	Fraction of HF exchange in global hybrids (e.g. hfx 0.10). Usable with b3lyp, b3lyp1, b3lyp5, bhandhlyp, revpbe0, pbe0, b3p86, b3pw91.	functional-dependent
rc_w	Range-correction scaling parameter \(\omega\) for long-range-corrected functionals (wpbe, wpbeh, mubop, wb97, wb97x, wb97xd3, camb3lyp, rcamb3lyp).	functional-dependent
c_ex	Short-range exact-exchange contribution; usable with the wpbeh functional.	functional-dependent
libxc	Comma-separated list of libxc functional names, e.g. libxc xc_gga_x_b88,xc_gga_c_lyp. If set, the libxc functional(s) are used and method is ignored for the XC part.	not set
libxc_fractions	Comma-separated weighting factors for the libxc functionals (same order); assumed 1 if unspecified.	not set
dftgrid	DFT grid density, integer 0–5 (larger = denser/more accurate). Grid 0 ≈ 800 pts/atom; grid 5 ≈ 80,000 pts/atom; default grid 1 ≈ 3,000 pts/atom. Also disables dynamic grids unless dynamicgrid yes is given. Construction documented under Integration Grids.	1
dynamicgrid	yes|no — use dynamic DFT grids: grid 0 converges until the DIIS error reaches gridthre, then grid dftgrid finalizes convergence.	yes (no if dftgrid is set)
gridthre	DIIS-error threshold for switching dynamic DFT grids.	0.01
dispersion / dftd (alias)	yes|no|d3|d2 — empirical dispersion correction. D21 and D32 are Grimme parameterizations.	no

Integral thresholds & precision

Parameter	Description	Default
threall	Two-electron integral threshold; integrals below it are neglected.	1.0e-11
xtol	Basis-set linear-dependence threshold. With diffuse functions, xtol and convthre may need raising to ~1.0e-3.	1.0e-4
precision	single|mixed|dynamic|double. dynamic (default) computes integrals above an automatically determined threshold in double precision and the rest in single. mixed uses a static threshold (threspdp). Accumulation on the GPU is always double precision.	dynamic
threspdp	Threshold (a.u.) splitting single/double precision work in mixed precision.	0.001
threcl	Coulomb integral threshold.	1.0e-11
pqthre	PQ threshold.	1.0e-15
threoe	One-electron integral threshold.	1.0e-12
thregr	Gradient integral threshold.	1.0e-11
threex	Exchange integral threshold.	1.0e-11
kscreen	Exchange screening method: none|dist|schwarz.	none
kscreendist	For kscreen dist: distance (Å) beyond which separated exchange elements are neglected.	not set
kscreenthre	For kscreen schwarz: neglect exchange elements with Schwarz bound below this value (a.u.).	not set

Properties & output

Parameter	Description	Default
nbo	NBO analysis: no|yes|npa|full|advanced|$nbo. yes/npa = natural population analysis only; full = full NPA/NBO; advanced/$nbo = analysis with a $nbo input group. Uses NBO v6.0.	no
polarizability	yes|no — compute the polarizability tensor (restricted wavefunctions only).	no
poptype	Population-analysis method: mulliken|vdd|vanalsenoy|camm_mulliken|none. Output appears in scr/charge_xxx.xls (xxx = mull|vdd|vanalsenoy) or scr/camm_mulliken.	mulliken
chkfile	Checkfile name. No checkfile is created if unset.	not set
memcheck	yes|no — perform a memory check.	yes
edensity	yes|no — compute the electron-density map (output in scr/density.dx).	no
bond_order_mat	yes|no — print the Mayer bond-order matrix3 to scr/bond_order.mat.	no
bond_order_list	yes|no — print Mayer bond orders as a list to scr/bond_order.list.	no
bond_order_thresh	Only bond orders above this threshold are printed in the list.	0.001
ml_prop	yes|no — generate machine-learning properties. If yes, properties (e.g. Mulliken populations) are written to scr/mullpop after each optimization cycle rather than only for the final structure.	no

Performance & hardware

Parameter	Description	Default
gpumem	GPU memory (MWords) reserved for host/device buffers (128 MWords = 1 GB).	128
timings	verbose|yes|no — print timings.	no

Initial guess

Parameter	Description	Default
guess	generate|path/to/WFfile|sad|hcore|sadlp|frag. generate builds the guess from scratch by maximum orbital overlap4; it can also be loaded from a WF file (dumped to scr/c0, or scr/ca/scr/cb for unrestricted — to load these pass two files, e.g. guess scr/ca scr/cb). hcore = core-Hamiltonian guess; sad = superposition of atomic densities; sadlp = SAD with UHF; frag = guess from converged UHF fragment densities.	generate
guess sad path/to/file	Optional file giving net charge, spin multiplicity, and spin up/down (a|b) per atom, in coordinate-file order. Without it, atoms are neutral, spin up, with NIST ground-state spins (up to \(Z=56\)).	no default
guess frag path/to/file	Required file: number of fragments on line 1, then atoms-per-fragment, net charge, multiplicity, and spin (a|b) per fragment, with atoms assigned in coordinate-file order.	no default
sad_spherical	yes|no / true|false — spherically symmetrize SAD densities.	true
sad_fxnl / frag_fxnl	hf|blyp — functional used to generate SAD or fragment densities.	hf
sad_avgspin / frag_avgspin	yes|no / true|false — average alpha/beta spins of the SAD or fragment guess.	true
frag_guess	hcore|generate — initial guess for each fragment.	generate
frag_dblprec	yes|no / true|false — use double precision in fragment convergence.	false
frag_threshscale	Scale factor for the fragment convergence threshold.	1.0
frag_maxit	Maximum SCF iterations for fragment calculations.	50
frag_minspin	yes|no / true|false — ignore the fragment-file multiplicities and minimize each fragment's multiplicity automatically (NIST ground-state spins).	false

SCF convergence

Parameter	Description	Default
purify	Density-matrix purification: guess|scf|no. guess purifies the initial guess before SCF; scf uses purification5 instead of diagonalization (no FON).	guess
scf	diis (Pulay's DIIS6) or diis+a (hybrid DIIS/A-DIIS7) for SCF convergence.	diis
diismaxvecs	Maximum number of DIIS vectors for diis+a.	10
fock	Fock-build algorithm: incremental|conventional|auto (auto = conventional if diffuse functions are detected, else incremental).	auto
maxit	Maximum number of SCF iterations.	100
convthre	Wavefunction convergence threshold (largest component of the DIIS error vector).	3.0e-5
levelshift	yes|no — apply level shifting.	no
levelshiftvala	If levelshift yes, the shift (hartree) for the virtual orbitals.	no default
levelshiftvalb	If levelshift yes and UHF/UKS or ROHF/ROKS, the shift (hartree) for the beta-spin / open-shell orbitals.	no default
watcheindiis	yes|no — switch from DIIS to ADIIS if DIIS predicts an energy increase for two consecutive iterations.	no

Fractional occupation (FOMO-SCF / FON)

Parameter	Description	Default
fon	yes|no — use FOMO-SCF (thermal smearing).	no
fon_method	Smearing type: gaussian|fermi|constant.	gaussian
fon_temperature	\(kT\) for FON (a.u., max 1.0).	0.25
fon_print	Print level for FON calculations (0|1; 1 is more verbose).	not set
fon_mix	yes|no / true|false — mix alpha and beta electrons in UHF.	no
fon_guess	File name to read FON occupation numbers from.	none
fon_anneal	yes|no|failsafe — anneal the FOMO temperature from high to low. Useful in normal DFT (final temperature 0) and when FOMO-SCF struggles to converge. Use the marzari fon_method for failsafe mode.	no
fon_anneal_iter	Number of SCF iterations over which annealing completes (should be < total SCF iterations).	100
fon_target	Target annealing temperature.	0
fon_converger	If yes, FON accelerates convergence and SCF finishes when DIIS converges (temperature is re-annealed on failure). If no, FON explores the SCF solution space, saving each converged solution.	yes
fon_tests	Number of annealing attempts to try.	3
fon_steer	Magnitude of the steering term (only in exploration mode); pushes the wavefunction away from previously converged densities.	6
fon_coldstart	If yes, start FON directly at the target temperature, only raising/annealing if SCF fails there.	value of fon_converger
closed	Number of doubly occupied orbitals for FOMO.	none
active	Number of orbitals with non-integer occupation for FOMO.	none

Geometry optimization & TS search (DL-Find)

Parameter	Description	Default
new_minimizer	yes|no — use the geomeTRIC minimizer (yes) or DL-Find (no).8	no
min_print	Output verbosity: something|verbose|debug.	verbose
nstep	Maximum number of optimization / TS-search steps.	100
min_tolerance	Convergence criterion on the maximum energy-gradient component.	4.5e-4
min_tolerance_e	Convergence criterion on the SCF energy change.	1.0e-6
min_coordinates	Coordinate system: cartesian|dlc|dlc_tc|hdlc|hdlc_tc.	dlc
min_method	Method: sd (steepest descent), cg1/cg2 (conjugate gradient), lbfgs (L-BFGS), bfgs (BFGS; needs min_init_hess).	lbfgs
gp_c3, gp_c4	Weights on the gradient terms in the gradient-projection conical-intersection search.9	1.0, 0.9
mdci_coordinates	Reference coordinate file whose distance is minimized for a minimum-distance conical intersection (MDCI).	value of coordinates
min_hess_update	Hessian-update algorithm: never|powell|bofill|bfgs. With never, the Hessian is recomputed by finite differences each step.	bfgs
min_init_hess	Initial Hessian (only for min_method bfgs): fischer-almlof|one-point|two-point|diagonal|identity. one-/two-point = exact finite-difference Hessian; diagonal = diagonal-only; identity = identity matrix; fischer-almlof = chemically-motivated guess (usually best).	fischer-almlof
min_delta	Atomic displacement in finite-difference calculations.	0.003
min_max_step	Maximum step size in internal coordinates.	0.5
min_restart	yes|no — start from checkpoint files (yes) or from scratch (no).	no
ts_method	TS-search method: neb_free|neb_restricted|neb_frozen|neb_free_cart|neb_restricted_cart|neb_frozen_cart. neb_free = NEB with free endpoints; neb_restricted = endpoints move perpendicular to their tangent; neb_frozen = frozen endpoints; *_cart = initialize in min_coordinates but search in Cartesians.	neb_free
min_image	Number of NEB images (> 1), listed in the coordinates file. Missing images are interpolated; at least two endpoints must be given. The last (min_image-th) is the climbing image.	10
min_nebk	Minimum/standard spring constant between NEB images.	0.01
max_nebk	Maximum spring constant for variable springs. If max_nebk ≤ min_nebk, the standard approach is used.	0.01
min_maxstephess	Update the Hessian every this many steps.	50
min_dump	Dump the optimization checkpoint every this many cycles.	10
min_scale_step	Scale each step by this factor.	1.0
num_deriv_formula	Numerical-derivative formula: central|forward|backward.	central
num_deriv_points	Number of points for numerical derivatives.	3
num_deriv_displacement	Displacement value for numerical derivatives.	0.01
lbfgs_mem	Number of steps kept in the L-BFGS memory.	#dof
lbfgs_reset	Number of steps after which to reset the L-BFGS minimizer.	never

Geometry optimization (geomeTRIC)

Parameter	Description	Default
new_minimizer	yes|no — use geomeTRIC.	no
min_coordinates	Coordinate system: prim|dlc|hdlc|tric|cart.	tric
min_maxiter	Maximum number of optimization / TS-search steps.	500
min_converge_gmax	Convergence criterion on the maximum energy-gradient component.	4.5e-4
min_converge_grms	Convergence criterion on the RMS energy gradient.	3.0e-4
min_converge_dmax	Convergence criterion on the maximum atom-displacement component.	1.8e-3
min_converge_drms	Convergence criterion on the RMS atom displacement.	1.2e-3
min_converge_e	Convergence criterion on the SCF energy change.	1.0e-6

Frequencies & initial conditions

Parameter	Description	Default
nderivpoints	Number of points in the numerical derivative: 3|5|7.	3
displacement	Step size in the finite difference (bohr).	0.005
initcondtemp	Temperature (K) for initial conditions.	0
husimi	true|false — sample from the Husimi distribution (default is Wigner).	false
maxgradnorm	The run stops if the maximum gradient component exceeds this (the initial geometry is assumed to be a stationary point; override with mincheck).	0.0001
mincheck	true|false — check that the initial geometry is a stationary point.	true
initcondtssign	Force displacement along imaginary modes in a given direction (-1|0|1); 0 does not displace. Run both signs to find which leads to reactants vs products.	0
cross_sections	Raman output: no (intensities), yes (cross-sections), or ti (temperature-independent part only).	no
thermo_temp	Temperature (K) for the vibrational/thermochemical analysis.	300

Molecular dynamics

Parameter	Description	Default
nstep	Total number of MD steps (set 0 for a single-point energy).	10^6
integrator	How the initial WF guess is built in successive MD steps: reversible_d (time-reversible with dissipation10), reversible (without dissipation11), regular (guess = converged WF from previous step), reset (regenerated from scratch each step). Also works with scfintegrator.	reversible_d
cisintegrator	As integrator, but for the excited state.	—
zintegrator	As integrator, but for the Z-vector.	—
md_binaryoutput	yes|no — write MD output in binary (DCD).	no
md_outputfreq	Frequency of MD output.	1
seed	Random-number-generator seed. Use different integers for independent runs, or a known seed to reproduce a run.	1351351
timestep	MD integration time step (fs).	1.0
thermostat	Temperature control: rescale (velocity rescaling), langevin, or bp (Bussi–Parrinello Langevin).	rescale
rescalefreq	With velocity rescaling, how often velocities are rescaled (e.g. 1000 = every 1000th step). For NVE, set larger than nstep.	2109
tinit	Initial temperature (K); velocities sampled from a Boltzmann distribution at tinit.	300.0
t0	Thermostat temperature (K).	300.0
lnvtime	Langevin damping time (fs); used only for thermostat langevin or bp.	1000.0
molden	Path to the output file of canonical MOs in Molden format.	molden.molf
orbitalswrtfrq	How often orbitals are written (integer). Orbitals can be large (100 MB+), so writing every step is rarely sensible.	2109
velocities	File of initial velocities, or random to draw from a Boltzmann distribution.	random
initialtime	Initial simulation time.	0.0
restartmd	Path to an MD restart binary file; if set, the run starts from its data (not from coordinates/velocities/temperature).	not set
restartmdfreq	How often restart data is dumped to scr/restart.md.	100

MD boundary conditions & nanoreactor

Parameter	Description	Default
mdbc	spherical (spherical boundary conditions) or spherical+rf (with an Onsager reaction field).	not set
epsilon	Solvent dielectric constant (for spherical+rf or pcm).	80.0
md_density	If set, \(R_1\) is adjusted automatically to this density (g/mL).	1.0
md_r1, md_r2	\(R_1\) and \(R_2\) parameters for spherical boundary conditions (Å).	0.0, 0.0
md_k1, md_k2	\(k_1\) and \(k_2\) force constants for spherical boundary conditions (kcal mol⁻¹ Å⁻²).	10.0, 0.0
mdbc_hydrogen	yes|no — include hydrogen atoms in the spherical boundary condition.	no
mdbc_mass_scaled	yes|no — mass-weight the spherical boundary condition.	no
mdbc_no_centering	yes|no — fix the boundary origin at (0,0,0); if no, it follows the system's center of mass.	no
mdbc_t1, mdbc_t2	Nanoreactor cycling: boundary md_r1 is active for mdbc_t1 steps, then md_r2 for mdbc_t2 steps, repeating.	not set

Interactive MD & steering

Parameter	Description	Default
imdport	Invoke interactive MD by specifying the VMD port.	not set
imdtime	Milliseconds per MD step (a lower bound: if SCF is slower the display freezes; if faster, MD waits).	not set
imdsubsteps	Substeps per MD step for integrating the haptic force.	imdtime/50 + 1
imddrawfreq	Drawing rate (frames per second) for the displayed system.	30
steering	adaptive|name_of_file|no. adaptive needs steeratom1, steeratom2, steerforce; a file name reads steering from that file.	no

Excited states (CIS / TDDFT)

Parameter	Description	Default
cis	yes|no — request a CIS (HF) or TDDFT/TDA (DFT) calculation.	no
rpa	yes|no — request an RPA (HF) or full TDDFT (DFT) calculation.	no
cisnumstates	Number of excited states.	1
cismult	Multiplicity of the excited states.	same as ground state
cistarget	Target state for properties or gradient.	0
cisalgorithm	Diagonalization algorithm: davidson|diis|debug.	davidson
ciss2	true|false — compute \(\langle S^2\rangle\).	true
cistransdipole	yes|no — compute transition dipole moments.	yes
cistransdipolederiv	yes|no — compute transition-dipole derivatives.	no
cisunrelaxdipole	yes|no — compute unrelaxed dipole moments.	yes
cisrelaxdipole	yes|no — compute relaxed dipole moments.	no
cisdipolederiv	yes|no — compute dipole-moment derivatives.	no
cischarges	yes|no — compute excited-state charges.	no
cisprintcivec	yes|no — print CI vectors.	yes
cisprintthresh	Threshold (absolute CI coefficient) above which contributions are printed.	0.1
cisnos	yes|no — print natural orbitals.	no
cisattachdetach	yes|no — print attachment/detachment densities.	no
cisntos	yes|no — print natural transition orbitals.	no
cistransdensity	yes|no — print the transition density.	no
cisdiffdensity	yes|no — print the difference of densities.	no
cisguessvecs	Number of guess vectors.	= cisnumstates
cissubspace	Number of vectors to store.	= cisnumstates
cismaxiter	Maximum number of Davidson iterations.	20
cismax	Maximum subspace size.	= cisnumstates*10
ciscollapse	Collapse vectors per root.	2
cisconvtol	Convergence tolerance on the residual.	1.0e-4
cisspdpconvtol	Convergence tolerance before switching single → double precision.	1.0e-4
cisstol	Orthogonality tolerance.	1.0e-12
cisuserguess	User guess file for CIS: lines label occ virt coeff (label blank for restricted, alpha/beta for unrestricted, x/y for RPA), each guess wrapped by vector/end.	not set
cisrestart	Name of a CIS restart file.	not set

Davidson-solver keyword prefixes

The Davidson-solver controls (cisguessvecs, cissubspace, cismaxiter, cismax, ciscollapse, cisconvtol, cisspdpconvtol, cisstol, cisuserguess, cisrestart) are built internally by prefixing a shared base name. The same options exist for other methods under the prefixes rocis, rci, dci, and ah (e.g. rocismaxiter).

CP-CIS

Parameter	Description	Default
cpcisalgorithm	CP-CIS algorithm: diis|inc_diis.	diis
cpcisiter	Maximum CP-CIS iterations.	50
cpcistol	CP-CIS convergence threshold.	1.0e-4

hh-TDA

See hh-TDA for full coverage.

Parameter	Description	Default
hhtda	yes|no — run an hh-TDA calculation (starts from an \(N+2\)-electron reference and annihilates two electrons).1213 Restricted wavefunctions only.	no
hhtdasinglets	Number of singlet states (the ground state is included).	0
hhtdatriplets	Number of triplet states.	0
cisnumstates	Must equal hhtdasinglets + hhtdatriplets.	0
cistarget	Target state for the gradient.	0
cismult	Target-state multiplicity.	1
hhtdaexactexchange	yes|no — use the functional-independent, HF-like response kernel (Yang and coworkers); default uses the functional-dependent linear-response kernel.	no
nacstate1	State 1 in nonadiabatic-coupling calculations.	—
nacstate2	State 2 in nonadiabatic-coupling calculations.	—
charge	If fon false, set to the actual charge minus 2 (generating orbitals for the \(N+2\) system; final states correspond to charge + 2). If fon true, charge is the SCF electron count and the final charge is set by closed/active.	—
fon	yes|no — fractional orbital occupation in the SCF (requires a fon_method and an active space). Final charge \(= \sum_A Z_A - 2(n_\text{active} + n_\text{closed})\).	—
closed	Number of closed orbitals.	0
active	Number of active orbitals (electrons from which the double annihilation occurs).	0

Recommended hh-TDA setting

A benchmark-based setting for vertical excitation energies: method bhandhlyp, charge 0, hhtda yes, fon yes, fon_method constant, closed 0, active \(= N_\text{el}/2 + 1\) (where \(N_\text{el}\) is the number of electrons).

CPHF / CPSCF

Parameter	Description	Default
cphfdiismin	DIIS minimum vectors for CPSCF equations.	1
cphfdiismax	DIIS maximum vectors for CPSCF equations.	10
cphfdiisstart	DIIS start for CPSCF equations.	0
cphfiter	Maximum CPSCF iterations.	20
cphftol	CPSCF convergence threshold.	1.0e-4
cphfalgorithm	CPSCF solver: diis|inc_diis|cg|precon_cg|jacobi|inc_jacobi.	diis

PCM solvation

See PCM Solvation for full coverage.

Parameter	Description	Default
pcm	Polarizable continuum model: cosmo (C-PCM) or xppcm (XP-PCM extreme-pressure model1415).	no PCM
epsilon	Solvent dielectric constant (default = bulk water).	80.0
pcm_scale	PCM charge scaling: 0 uses \(f=(\epsilon-1)/\epsilon\) (G-COSMO16, C-PCM17); 1 uses \(f=(\epsilon-1)/(\epsilon+0.5)\) (original COSMO18).	0
solvent_radius	Solvent probe radius (Å).	0.0
pcm_grid	PCM sphere grid: polyhedron (original COSMO cavity by Klamt), lebedev (Lebedev grid, point charge per point), iswig (switching Gaussian by Lange & Herbert), swig (switching Gaussian by York & Karplus), sphere (single fixed spherical cavity for mdbc=1).	iswig
nspa	Gathered tessellation order (NSPA), 0–5; only for pcm_grid polyhedron, must satisfy nspa ≤ nppa.	2
nppa	Basic tessellation order (NPPA), 0–5; only for pcm_grid polyhedron.	3
pcmgrid_h	Lebedev grid order \(\ell\) for hydrogen atoms: 5|7|11|17|23|29|35|41|47|53|59 (14 → 1202 points/atom respectively).	17
pcmgrid_heavy	Lebedev grid order \(\ell\) for heavy atoms (same options as pcmgrid_h).	17
pcmgrid_h_mm	Lebedev grid order \(\ell\) for MM hydrogen atoms (same options).	17
pcmgrid_heavy_mm	Lebedev grid order \(\ell\) for MM heavy atoms (same options).	17
pcmgrid_ptchrg	Lebedev grid order \(\ell\) for point charges in point-charge QM/MM (same options).	17
pcm_radii	PCM radii method: klamt (Klamt paper19), ff, read (from pcm_radii_file), bondi (Bondi radii2021 with a new H definition22). Relevant for pcm_grid ≠ sphere.	bondi
pcm_rad_scale	Scaling factor for the solute atom radius in COSMO.	1.2
pcm_ptchrg_radius	Universal radius for all point charges in QM/MM COSMO.	1.5
cosmodsc	\(\delta^{SC}\) for the original-COSMO cavity (Klamt); only for pcm_grid polyhedron.	0.0
pcm_radii_file	File of PCM radii when pcm_radii read.	not specified
pcm_global_radius	Radius of the single spherical cavity when pcm_grid sphere and mdbc=1.	10
pcm_matrix	Storage of matrix \(\mathbf{A}\): full, packed (lower triangle), no (build \(Ax\) on the fly, double precision), no_sp (on the fly, single precision).	full
pcm_pair_driven	yes|no — use the primitive-pair-driven algorithm for the grid-point electrostatic potential.	no
pcm_threoe	Screening threshold for PCM one-electron integrals.	1.0e-12
dynamiccg	Dynamic CG mode: 0 (off — use cgtoltight throughout23), 1 (2-\(\delta\) switching), 2 (empirical-formula threshold).	2
cgtoltight	If dynamiccg 0, the fixed CG threshold (\(\|Ax-b\| <\) this). If dynamiccg 1, the tight threshold \(\delta_2\).	1.0e-6
cgtolscale	Loose-threshold scaling for the 2-\(\delta\) scheme (\(\delta_1 = \delta_2 \cdot\) cgtolscale).	10000
dynamiccgtol	For dynamiccg 1: start using the tight threshold once the DIIS error drops below this.	1.0e-3
cgprecond	CG preconditioner: jacobi (diagonal) or blockjacobi (random block; set cgblocksize).	jacobi
cgblocksize	CG block size when cgprecond blockjacobi.	100
print_ms	Print the molecular-surface coordinates (sas.xyz): 1=yes|0=no.	0
ss_pcm_solvation	State-specific24 solvation for CIS-type calculations: eq (equilibrium), neq (nonequilibrium), ground_neq (ground-state nonequilibrium). Only with run energy.	none
sspcm_maxit	Maximum SS-PCM iterations.	20
sspcm_convthre	SS-PCM convergence threshold.	1.0e-5
lr_pcm_solvation	Linear-response25 solvation for CIS-type calculations: eq or neq. Requires cis yes and the restricted-CIS parameters.	none
pcm_write	File name to write the final PCM field to.	no default
pcm_read	File name to read a PCM field from (for SS nonequilibrium calculations with ss_pcm_solvation neq|ground_neq).	no default
fast_epsilon	Fast dielectric constant for nonequilibrium solvation (ss_pcm_solvation neq|ground_neq, lr_pcm_solvation neq, or split_pcm_energy 1).	2.0
split_pcm_energy	0|1 — split the ground-state PCM solvation energy into fast and slow parts and print them (final energy unchanged).	0
xppcm_nb	XP-PCM only: number of valence electrons of the solvent.	0
xppcm_mb	XP-PCM only: molecular weight of the solvent.	18.0
xppcm_rhob	XP-PCM only: number density of the solvent.	0

ESP / RESP charges

See RESP for full coverage.

Parameter	Description	Default
resp	yes|no — compute ESP/RESP charges.	no
esp_grid_scale	Cavity radius of the first Connolly-surface layer (float, > 1.0).	1.4
esp_grid_incr	Radius increment between Connolly-surface layers (float, > 0). Layer \(i\) radius: esp_grid_scale \(+\, i\times\) esp_grid_incr \(\times\) vdW radius.	0.2
esp_grid_layers	Number of Connolly-surface layers (integer).	4
probe_radius	Radius of the rolling-ball probe for the Connolly surface (Å).	0
esp_grid_dens	Grid-point density (points/Ų). Increase until the charge converges.	1.0
esp_restraint_a	Restraint strength \(a\) in \(a\sum_k\big((q_k^2+b^2)^{1/2}-b\big)\) (a.u.).	0.0005
esp_restraint_b	Restraint tightness \(b\) near the bottom of the restraint (\(e^-\)).	0.1

Uniform electric field

Parameter	Description	Default
finite_field	yes|no — apply a uniform electric field.	no
efx	Field magnitude along the lab-frame x-direction (units of \(e\)).	0.0
efy	Field magnitude along the lab-frame y-direction (units of \(e\)).	0.0
efz	Field magnitude along the lab-frame z-direction (units of \(e\)).	0.0

Exciton model

See Exciton Model overview for theory, input format, and worked examples.

Parameter	Description	Default
exmod_uscf	Run monomer SCFs unrestricted.	no
exmod_read_scr	Read cached intermediates from the exciton scratch directory.	yes
exmod_scr_dir	Subdirectory of scrdir for exciton intermediates.	exmod_rst
exmod_debug	Print extra diagnostics during the exciton build.	no
exmod_tdinfo	Print monomer TDDFT/CIS auxiliary information.	no
exmod_ct	Number of orbital pairs in the CT block; total CT states \(= 2M(M-1)N^2\).	0
exmod_tda_ct	Apply Tamm-Dancoff approximation in the CT block.	no
exmod_tt	Include the triplet-triplet (multiexcitonic) state. Requires exmod_ct 1.	no
exmod_tda_tt	Apply Tamm-Dancoff approximation in the TT block.	no
exmod_ct_tt_class	Classification analysis on the mixed CT/TT block.	no
exmod_tuning	Reserved; not yet implemented (aborts if set).	no
exmod_approx	Enable additional dimer-approximation strategies.	no
exmod_min_dist	Pair cutoff (Å); above this distance pairs are treated only electrostatically.	\(10^9\)
exmod_gs_grad	Compute the ground-state gradient.	no
exmod_es_grad	Compute the excited-state gradient (state given by exmod_state).	no
exmod_num_grad	Numerical gradients (verification only).	no
exmod_gen_cube	Generate cube files of monomer/exciton orbitals and densities.	no
exmod_md	Run exciton-Hamiltonian Born-Oppenheimer MD.	no
exmod_md_rst	Read an existing exciton MD restart file.	yes
exmod_dt	MD timestep (fs).	0.1
exmod_nsteps	Number of MD steps.	0
exmod_state	Eigenstate of the exciton Hamiltonian to propagate (0 = lowest).	0
exmod_nvt	Berendsen thermostat (NVT).	no
exmod_ref_T	Berendsen target temperature (K).	300
exmod_tau_T	Berendsen relaxation time (fs).	20
exmod_esp_chg	Run the per-monomer ESP fit during exciton-model setup.	no
exmod_esp_test	Run ESP self-consistency tests during setup.	no
exmod_esp_read	Read pre-computed ESP charges (disables iterative ESP).	no
exmod_esp_file	File of ESP charges to read or write.	final.chg
exmod_mm_0	Include MM atoms at the GS-only level.	no
exmod_mm_iter	Iterate the MM-induced ESP charges to self-consistency.	no
exmod_mm_1	Include MM atoms in monomer Fock matrices.	no
exmod_mm_2	Include MM atoms in dimer Fock matrices.	no
exmod_use_pauli	Add dimer-SCF Pauli repulsion to inter-monomer interaction.	no
exmod_use_lj	Augment inter-monomer interaction with a 6-12 Lennard-Jones term.	no
exmod_excl_qmmm	Exclude the QM/MM energy/gradient from the reported totals.	no
exmod_mm_freeze	Freeze MM atoms during exciton-MD or FMS.	no
exmod_scan	Scan the first dimer's centre-of-mass separation.	no
exmod_scan_rmin	Scan starting separation (Å).	1.80
exmod_scan_dr	Scan step size (Å).	0.15
exmod_scan_num	Number of scan points.	40

F-SAPT

See F-SAPT for theory, the fA.dat/fB.dat fragment-file format, and worked examples.

Parameter	Description	Default
fsapt	yes|no — run the functional-group SAPT (F-SAPT) interaction-energy decomposition.	no
fsapt_na	Number of leading atoms forming monomer A (the remaining atoms are monomer B).	not set
fsapt_chargea	Total charge of monomer A.	0
fsapt_chargeb	Total charge of monomer B.	0
fsapt_fraga	Path to the monomer-A fragment file (fA.dat); if unset, monomer A is a single fragment.	not set
fsapt_fragb	Path to the monomer-B fragment file (fB.dat); if unset, monomer B is a single fragment.	not set

Orbital localization (IAO / IBO)

See Localization (and IAO charges).

Parameter	Description	Default
localization	yes|no — localize the orbitals / compute IAO charges.	no
loctype	Localization method: pm (Pipek–Mezey), fb (Foster–Boys), iao (Knizia IAO charges), ibo (Knizia intrinsic bond orbitals).	pm
locblock	Orbital block to localize: occ, virt, closed, active, ca, ov, all.	occ
locmaxiter	Maximum localization iterations.	1000
iao_dump	yes|no — dump the IAO matrix and IBO coefficients to the scratch directory.	no

---

1. S. Grimme, J. Comp. Chem. 27, 1787 (2006). ↩

2. S. Grimme, J. Antony, S. Ehrlich, and H. Krieg, J. Chem. Phys. 132, 154104 (2010). ↩

3. I. Mayer, J. Comp. Chem. 28, 204–221 (2007). ↩

4. J.-M. Langlois et al., J. Phys. Chem. 98, 13498 (1994). ↩

5. A. M. N. Niklasson, Phys. Rev. B 66, 155115 (2002). ↩

6. P. Pulay, J. Comp. Chem. 3, 556 (1982). ↩

7. X. Hu and W. Yang, J. Chem. Phys. 132, 054109 (2010). ↩

8. L.-P. Wang and C. Song, J. Chem. Phys. 144, 214108 (2016). ↩

9. T. W. Keal, A. Koslowski, and W. Thiel, Theor. Chem. Acc. 118, 837 (2007). ↩

10. A. M. N. Niklasson et al., J. Chem. Phys. 130, 214109 (2009). ↩

11. A. M. N. Niklasson et al., Phys. Rev. Lett. 97, 123001 (2006). ↩

12. C. Bannwarth, J. K. Yu, E. G. Hohenstein, T. J. Martínez, J. Chem. Phys. (2020); ChemRxiv DOI: 10.26434/chemrxiv.11828256. ↩

13. J. K. Yu, C. Bannwarth, E. G. Hohenstein, T. J. Martínez, in preparation. ↩

14. R. Cammi, V. Verdolino, B. Mennucci, and J. Tomasi, Chem. Phys. 344, 135 (2008). ↩

15. A. Gale, E. Hruska, F. Liu, J. Chem. Phys. 154, 244103 (2021). ↩

16. T. N. Truong and E. V. Stefanovich, Chem. Phys. Lett. 240, 253–260 (1995). ↩

17. V. Barone and M. Cossi, J. Phys. Chem. A 102, 1995–2001 (1998). ↩

18. A. Klamt and G. Schüürmann, J. Chem. Soc. Perkin Trans. 2 799–805 (1993). ↩

19. A. Klamt and G. Schüürmann, J. Chem. Soc. Perkin Trans. 2 799–805 (1993). ↩

20. A. Bondi, J. Phys. Chem. 68, 441–451 (1964). ↩

21. R. S. Rowland and R. Taylor, J. Phys. Chem. 100, 7384–7391. ↩

22. M. Mantina, A. C. Chamberlin, R. Valero, C. J. Cramer, D. G. Truhlar, J. Phys. Chem. A 113, 5806–5812. ↩

23. F. Liu, N. Luehr, J. K. Kulik, T. J. Martinez, J. Chem. Theory Comput. DOI: 10.1021/acs.jctc.5b00370. ↩

24. R. Cammi, S. Corni, B. Mennucci, J. Tomasi, J. Chem. Phys. 122, 104513 (2005). ↩

25. R. Cammi, S. Corni, B. Mennucci, J. Tomasi, J. Chem. Phys. 122, 104513 (2005). ↩