Jump-Start Guide

ORCA 4.1

March 2018

Frank Neese & Christoph Riplinger

With contributions by

Alexander A Auer, Tran Van Anh, Dipayan Datta, Bernardo de Souza, Robert Izsak, Lucas Lang, Quing Lu, Maurice
van Gastel, Kantharuban Sivalingam, Georgi Stoychev, Frank Wennmohs

1
 The ORCA Quantum Chemistry Suite
Jump-Start Guide

This document is intended to provide a “jump start” to the use of ORCA. It is not intended to replace the manual.
There are many more options and features in the program that can be controlled by the user.

1 GENERAL ......................................................................................................................................4

2 INPUT FILE....................................................................................................................................4

3 COORDINATES ..............................................................................................................................4

4 BASIS SETS ...................................................................................................................................5

5 AUXILIARY BASIS SETS ..................................................................................................................5

6 DENSITY FUNCTIONALS.................................................................................................................6

7 COMPOSITE METHODS FOR OPTIMIZATION ..................................................................................7

8 SELF-CONSISTENT FIELD CALCULATIONS ........................................................................................7

9 COMPLETE ACTIVE SPACE SELF-CONSISTENT FIELD CALCULATIONS ................................................8

10 SOLVATION ................................................................................................................................8

11 MØLLER-PLESSET PERTURBATION THEORY ..................................................................................8

12 COUPLED CLUSTER CALCULATIONS..............................................................................................8

13 RELATIVITY .................................................................................................................................9

14 GEOMETRY OPTIMIZATION .........................................................................................................9

15 FREQUENCY CALCULATIONS...................................................................................................... 10

16 PARALLEL RUNS ........................................................................................................................ 10

17 EXCITED STATE CALCULATIONS ................................................................................................. 10

18 SPECTROSCOPIC AND OTHER PROPERTIES ................................................................................. 13

19 ANALYSIS TOOLS AND INTERFACES ........................................................................................... 20

2
20 VISUALIZATION TOOLS AND INTERFACES................................................................................... 22

21 ORCA FILE TYPES....................................................................................................................... 23

3
1 General
 Make sure ORCA is visible in the system in your $PATH variable
 Provide an ASCII input file, e.g. MyJob.inp
 Run orca MyJob.inp >& MyJob.out &. This produces at least MyJob.out and
MyJob.gbw. The latter contains a binary summary of geometry, basis sets and orbitals. It is
necessary for restarting jobs or using one job as input for another job

2 Input File
The input file has the general structure:

# comments – write whatever you want

! Keywords (control methods, basis sets, job type,…)
! moread (optional: start from existing orbitals; also
requires %moinp, see next line)
%moinp “MyInp.gbw” (optional: Input MOs to start from)
%maxcore 4096 (optional: scratch array size in MB)
%base “MyBaseName” (optional: name of the generated scratch files)
* xyz Charge Mult (required: coordinate type, charge, multiplicity)
C 1.23 16.481 -9.87 (Cartesian coordinates Angström or Bohr)
* (end of coordinate input)
$new_job
(now you can add another job – as many as you want)
Keywords can be given in any order, do not need to be on one line and they are not case sensitive. The input
molecular orbitals do not need to be in the same basis set as the calculation to be performed and /or do not need
to be calculated at the same geometry as given in the input file (but the atom types and their sequence must
match).

3 Coordinates
! Angs or Bohrs (optional: coordinate type, default = Angström)

* xyz charge multiplicity

C 0.0 0.0 0.0 (Cartesian coordinates for this atom)
O 0.0 0.0 1.128
*

or

* int charge multiplicity

AT NA NB NC R A D (AT= atom type H,C,N,O,…)
(NA= atom for distance NA-NB)
(NB= atom for angle- this atom-NA-NB)
(NC= dihedral angle- this atom-NA-NB-NC)
(R = actual distance (Angström or Bohr)
(A = Angle in degrees)
(D = dihedral angle in degree)

or

*xyzfile charge multiplicity myCoordinateFile.xyz

4
charge is the total charge of the molecule and multiplicity is the spin multiplicity (=2S+1 where S is the
total spin of the desired state. NOTE – for non-closed shell systems, typically spin-unrestricted calculations are
performed. These do NOT conserve the actual multiplicity. One obtains an eigenfunction of the z-component of
2 1
the total spin operator, not of S !).

You can also input and run a series of structures from “multiple XYZ” files (see manual).

4 Basis Sets
! BasisSet

BasisSet= def2-SVP, def2-TZVP, def2-TZVPP, def2-QZVPP,

ma-def2-SVP, ma-def2-TZVP, ma-def2-TZVPP, ma-def2-QZVPP,

cc-pVnZ, aug-cc-pVnZ, cc-pCVnZ, aug-cc-pCVnZ (n= D,T,Q,5,6)

We recommend the ‘def2’ basis sets of the Karlsruhe group for most HF/DFT calculations.

By default, basis sets make use of ECPs. The def2 basis sets automatically load Stuttgart-Dresden effective core
potentials for elements Rb-Rn.

Many other and specialized basis sets are documented in the manual. Basis sets can also be read from a file or
manually input or they can be printed using (! PrintBasis) and modified basis sets can be input into the
program using:

%basis GTOName “MyGTOBasis.bas” end

The format of the basis set input file is essentially that from the EMSL library for Gamess-US. ORCA will accept
basis sets that are not normalized and will normalize them. Normalized basis sets are printed by ORCA using the
! PrintBasis keyword. Individual basis sets for specific atoms or atom types can also be given in the input as
described in the manual.

5 Auxiliary Basis sets

Methods that make use of the “RI” (‘resolution of the identity’ or density fitting) approximation require an
auxiliary basis set. For example, pure DFT calculations are done by default in the RI mode. For this an auxiliary
basis set is necessary.

! AuxBasisSet

AuxBasisSet=def2/J, SARC/J

For Coulomb fitting in RI-DFT (GGA, meta-GGA functional). SARC/J is for scalar relativistic all electron
calculations.

AuxBasisSet= def2-SVP/C, def2-TZVP/C, def2-TZVPP/C, def2-QZVPP/C, cc-pVnZ/C,

aug-cc-pVnZ/C (n=D,T,Q,5,6)

For Correlation fitting (MP2 and Coupled Cluster)

1
If you want to have a spin-eigenfunction for an open-shell system, you can resort to open-shell spin restricted
calculations (for DFT only available for all unpaired spins aligned in parallel), or you proceed to a
multiconfigurational wavefunction such as CASSCF.

5
AuxBasisSet= def2/JK, cc-pVnZ/JK, aug-cc-pVnZ/JK (n=T,Q,5)

For Coulomb and exchange fitting (HF and hybrid DFT)

If you do fitting for both, the SCF and a correlation calculation, you can provide two auxiliary basis sets.

! AutoAux

Automatic construction of a general-purpose auxiliary basis for simultaneously fitting Coulomb, exchange and
correlation calculations. This will be accurate, but will be larger than the optimized auxiliary basis sets. Safe choice
in case of doubt or non-standard orbital bases.

6 Density Functionals
! Functional GridN NoFinalGrid VDW

Functional=

LDA : LSD, HFS, …

GGA : BP=BP86, PBE, PW91, OLYP, OPBE, BLYP, PWP, …

Note: LDAs/GGAs in RI mode require an auxiliary basis set (‘/J’)

Meta-GGA : TPSS, revTPSS, M06L, …

Hybrid : B3LYP, PBE0, X3LYP, BHandHLYP, B3P, B3PW, …

Range-Sep. Hyb. : wB97, wB97X, CAM-B3LYP, LC-BLYP

Hybrid Meta-GGA : TPSSh, TPSS0, M06, M062X, …

Double-Hybrid : RI-B2PLYP, RI-MPW2PLYP, RI-B2T-PYLP, RI-B2K-PYLP,

RI-B2GP-PLYP

All also available without RI approximation. Double hybrids in RI mode require an auxiliary basis set (‘/C’)

There are many more functionals documented in the manual.

GridN= Grid1, Grid2, …, Grid7

NoFinalGrid= Do not use a larger grid for the final energy, alternatively use
FinalGridN (N = 0-7)

The default is to use a combination of Grid2 (110 angular points) and Grid4 (302 angular points)

VDW=

D3BJ Grimme’s atom-pairwise dispersion correction with Becke Johnson

damping (recommended)

D3Zero atom-pairwise dispersion correction with zero damping

D2 Grimme’s dispersion correction from 2006

6
7 Composite Methods for Optimization
! HF-3c, PBEh-3c, B97-3c

These are low-cost, composite methods from the Grimme group that are designed for optimization using a small
basis, and corrections for dispersion, BSSE and basis set incompleteness.

8 Self-Consistent Field Calculations

! SCF-Keywords

SCF-Keywords= SP (Single point calculation)

NormalSCF (Normal SCF convergence criteria)

TightSCF (Tight SCF convergence criteria)
VeryTightSCF (Very Tight SCF convergence criteria)

SlowConv (Slow convergence expected)

LShift (Turn on level shifting)
SOSCF (Turn on approx. second order SCF)
NRSCF (Turn on Newton-Raphson SCF)
DIIS (Turn on DIIS)
Direct (Integral direct mode)
Conv (Integral conventional mode)

RHF (closed shell calculation)

UHF (spin unrestricted calculation)
ROHF (restricted open shell calculation; more
Input may be required)
The default is a combination of DIIS and SOSCF, unless transition metals are involved. The default is to run singlets
closed shell and higher multiplicities spin unrestricted. The combination KDIIS SOSCF is often a good choice for
rapid convergence but may not be as robust as the default for electronically difficult species.

Powerful approximations

! RI

For non-hybrid DFT (automatically invoked. Turn Off=NoRI)

! RIJCOSX

Efficient approximation to HF and hybrid DFT. Does the Coulomb Term with RI and the exchange by
seminumerical integration

Requires an auxiliary basis set (‘/J’) and an X-Grid for the seminumerical approximation to the exact
exchange

GridXN = N = 1-9 (e.g. GridX6) Grids for the COSX approximation.

! RI-JK

Fully RI based treatment of Coulomb and exchange

Requires an auxiliary basis set (‘/JK’). Will not scale well with system size!

7
9 Complete Active Space Self-Consistent Field Calculations
CASSCF calculations are not for the faint-hearted! They always require the user to look carefully at the orbitals and
frequently require some manual adjustments to achieve self-consistent field convergence. Since the matter is
complex, we have prepared a CASSCF tutorial in a separate document that walks the user through some typical
calculations. The manual also provides a list of additional information

On the most elementary level, just use:

%casscf nel n (number of electrons in the active space)

norb m (number of orbitals in the active space)
nroots N (number of states averaged over)
nevpt2 true (calculate a second order correction for the
end dynamical correlation energy – easy!)

10 Solvation
! CPCM(Solvent) or CPCM

CPCM alone uses an infinite dielectric

Solvent = Water,Acetonitrile, Acetone, Ammonia, Ethanol, Methanol, CH2Cl2,

CCl4, DMF, DMSO, Pyridine, THF, Chloroform, Hexane, Toluene

!CPCM
%cpcm smd true # default false
solvent “Name”
end

For the SMD solvation model there are at present 179 solvents in the library. If required, the user can manually
specify the descriptors. For more information see manual.

11 Møller-Plesset Perturbation Theory

! MP2, SCS-MP2, F12-MP2, RI-MP2, RI-SCS-MP2, F12-RI-MP2, DLPNO-MP2, F12-
DLPNO-MP2, OO-RI-MP2, OO-RI-SCS-MP2, MP3, RI-MP3, SCS-MP3, RI-SCS-MP3

%maxcore 2048 # higher maxcore required

RI methods require an auxiliary correlation fitting basis set (‘/C’). Additionally, the RIJCOSX or RI-JK methods can be
used to accelerate the SCF calculation. A corresponding auxiliary basis set is required. Analytic gradients and
Hessians are available for MP2 and RI-MP2.

12 Coupled Cluster Calculations

! CC-Variant Extrapolate(n/m,bas)

Extrapolate allows you to automatically extrapolate the result to the basis set limit using standard basis sets.

%maxcore 2048 # higher maxcore required

CC-Variant=CCSD, CCSD(T), QCISD, QCISD(T)

8
 CCSD-F12, CCSD(T)-F12, QCISD-F12, QCISD(T)-F12

F12 methods require F12-specific basis sets (‘-F12’ and ‘-F12/CABS’)

CCSD-F12/RI, CCSD(T)-F12/RI, QCISD-F12/RI, QCISD(T)-F12/RI

Besides the F12-specific basis sets a regular auxiliary basis (‘/C’) must be provided. It is
recommended to step one level up compared to the F12-basis (e.g. ! cc-pVDZ-F12 cc-
pVDZ-F12-CABS cc-pVTZ/C)

CPF/1, NCPF/1, CEPA/1, NCEPA/1

DLPNO-CCSD, DLPNO-CCSD(T), DLPNO-QCISD, DLPNO-QCISD(T)

Local correlation methods. For DLPNO calculations an auxiliary basis (‘/C’) must be
provided

Extrapolate(n/m,bas)=

n/m: cardinal number combinations, n,m = 2-5; n < m

Extrapolation of SCF and CC-Variant energies to the basis set limit. More extrapolation
methods are described in the manual.

bas: cc, aug-cc, cc-core, ano, saug-ano, aug-ano, def2

Coupled cluster densities are available for closed and open shell systems (also in the DLPNO case), but coupled
cluster gradients and Hessians are not available.

13 Relativity
! RelMethod BasisSet AuxBasisSet

RelMethod= ZORA, DKH, DKH2, ZORA/RI

BasisSet = ZORA-def2-XVP, DKH-def2-XVP, ma-ZORA-def2-XVP, ma-DKH-def2-XVP

(where XVP stands for the whole range of def2 basis sets)

Adapted versions of the def2 basis sets. For elements H-Kr.

SARC-DKH-TZVPP, SARC-ZORA-TZVPP

For use with DKH2 and ZORA. For elements beyond Xe.

AuxBasisSet= SARC/J

More available basis sets are described in the manual. Spin-orbit coupling is treated by quasi-degenerate
perturbation or linear-response theory in a variety of modules. What ORCA does not do are two- or four-
component calculations with relativistic spinors (we simply don’t like these methods).

14 Geometry Optimization
! OptMethod TightOpt

TightOpt= increases convergence criteria; default NormalOpt

9
OptMethod= Opt Standard optimization in redundant internals
Copt Optimization in Cartesian coordinates
OptTS Transition State optimization
ScanTS relaxed surface scan followed by automated TS opt

Many other options for constraint optimization, surface scans, minimum energy crossing point optimization,
QM/MM optimizations etc. are documented in the manual.

15 Frequency Calculations
! Freq

Performs an analytic frequency calculation.

! NumFreq

Performs a numerical frequency calculation.

16 Parallel Runs
! palN

N= 2 – 8, 16, or use

%pal nprocs N end # N=any integer

Performs a parallel ORCA run using N cores. Installation of openmpi is required. PATH and LD_LIBRARY_PATH
variables need to be set for openmpi.

17 Excited State Calculations

A number of ORCA modules calculate excited states. They all are capable of generating absorption and CD spectra
at the various levels of theory. Only TD-DFT/CIS features analytic gradients.

17.1 Time dependent Density Functional Theory

%tddft NRoots 10 (number of roots to determine)
MaxDim 100 (max. size of the expansion space)
TDA true (Tamm-Dancoff approximation used or not)
Triplets true (calc. triplets for closed shell references?)
end

Note that this input is also valid to do CIS or RPA calculations in the Hartree-Fock framework. Many other options
to speed up TD-DFT calculations, (D) corrections, analytic gradients and other topics are discussed in the manual.
The module works together with RIJCOSX but not with RI-JK (but there are RI-JK like options).

17.2 Equation of Motion Coupled Cluster Theory

! EOM-CCSD

or

%mdci doeom true end

and the number of roots must be specified as well under

10
%mdci NRoots 10 (number of roots to determine)
NDav 20 (defines max. size of the exp. space as NDav*NRoots)
DoRootwise true (solve one root at a time, often more stable)
CCSD2 (EOM with MP2 amplitudes, for RHF reference only)
FollowCIS true (follow CIS roots instead of energetic ordering)
DoTDM true (request transition properties)
end

along with other options. The EOM module is also capable of computing ionization energies, and it can handle RHF
and UHF reference functions, the latter being more expensive. With a UHF reference, DOMO to virtual excitations
must be handled with care since they are strongly spin contaminated. The COSX approximation is also available for
certain terms, as described in the manual. Unlike in TDDFT, the transition properties require additional work,
although a cheap approximation is printed at the end of the calculation.

For a cheaper but accurate option for computing excitation energies, the similarity transformed version of EOM
(STEOM) can be invoked as

! STEOM-CCSD

or

%mdci dosteom true end

The keywords NRoots, NDav and DoRootwise function similarly as in the EOM case. A few of the STEOM
specific keywords include

%mdci DoTriplet true (calculate triplet excited states)

OThresh 0.001 (CIS cutoff for the occupied block)
VThresh 0.001 (CIS cutoff for the virtual block)
end

The triplet option is only available for closed shell calculations. STEOM requires the selection of an active space to
define the similarity transformation. While this is done automatically, the parameters can also be manually set,
among them the most important ones are the occupied and virtual space CIS cutoff parameters described above.
For closed shell calculations, both EOM and STEOM can be carried out in conjunction with DLPNO-CCSD
calculations for the ground state, which are then back-transformed into the canonical basis, see the bt-PNO
options in the manual.

17.3 Restricted open shell CIS (ROCIS) and ROCIS/DFT

%rocis NRoots 10 (number of roots to determine)
MaxDim 100 (max. size of the expansion space)
end

Performs CIS calculations on top of a RHF or a high-spin ROHF reference function

! B3LYP/BHLYP

%rocis NRoots 10 (number of roots to determine)

MaxDim 100 (max. size of the expansion space)
DoDFTCIS true (request the ROCIS/DFT calculation)
DFTCIS_c = 0.20, 0.40, 0.30 (parameters for B3LYP functional)
end

Performs a ROCIS/DFT calculation.

11
In addition it is possible to treat relativistic effects or request spin-orbit coupling (SOC) treatment in the framework
of the quasi-degenerate perturbation theory (QDPT). Many other options are available to speed-up the ROCIS
calculations and to control the calculation output. Other topics are discussed in the manual.

17.4 N-Electron Valence Perturbation Theory (NEVPT2)

Strongly Contracted and Fully Internally Contracted NEVPT2 can easily be set up on top of any working CASSCF
input adding the simple input keywords.

!NEVPT2, RI-NEVPT2 (strongly contracted)

!FIC-NEVPT2, DLPNO-NEVPT2 (fully internally contracted)

Refined settings (e.g. F12 correction, RI Approximation) are set in the %CASSCF block and the NEVPT sub-block –
see the manual for details. A number of properties and spectra are accessible with NEVPT2. The CASSCF tutorial
(downloadable from our website) covers many examples on the subject.

Note that the RI approximation is available for SC- and FIC-NEVPT2. The RI and DLPNO versions require auxiliary
basis sets (/C or /JK) for the integral transformation. The approaches can be combined with !RIJK and !RIJCOSX for
additional speed-up.

17.5 Multireference Configuration Interaction (MRCI)

A traditional (uncontracted) MRCI with a plethora of truncation options resides in the orca_mrci module and
allows the computation of various spectra and properties (SOC, SSC,MCD, g-Tensor, D-Tensor,...). Examples can be
found in the CASSCF tutorial (downloadable from our website) and the manual. The minimal input requires a set of
orbitals and a multiplicity block:

!MORead Allowrhf NoIter (reading arbitrary orbitals e.g. from CASSCF,HF)

%moinp “start.gbw”

%mrci
citype MRCI (MRCI, SORCI, MRDDCI3, MRDDCI2,MRDDCI1,...)
newblock M * (multiplicity block, * can be replaced by IRREP)
ref CAS(n,m) end (reference WF, can also be RAS or a list of CFGs)
NRoots N (number of roots in the multiplicity/IRREP block)
end
end

ORCA also features a fully internally contracted MRCI, which is implemented in the orca_autoci module. The input
mimics a standard CASSCF input mask. The reference wavefunction is restricted of the CAS-CI type.

!MORead CASSCF NoIter (reading CASSCF orbitals)

%moinp “start.gbw”

%autoci
citype FICMRCI (FICMRCI, FICDDCI3, FICMRCEPA0)
nel n (number of active electrons)
norb m (number of active orbitals)
Mult M (multiplicity)
NRoots N (number of roots in the multiplicity)
end

The contracted and uncontracted MRCI approaches can be combined with the RI approximation. In general, MRCI
calculations are very resource demanding and thus should be carefully designed.

12
18 Spectroscopic and other Properties

18.1 Electric Moments and Polarizabilities

%elprop dipole true (compute the dipole moment)
quadrupole true (compute the quadrupole moment)
polar 1 (compute the polarizability using analytical
end derivatives)

This input block will yield the dipole moment vector (in a.u. and Debye) including some informations about
rotational parameters, the quadrupole moment (as full tensor and diagonalized) and run a CP-SCF calculation to
obtain the polarizability tensor (raw Cartesian and diagonalized). Note that the analytical derivative for the
polarizability is only available at the HF and DFT level of theory. For other methods, the polarizability can be
computed as finite differences.

%elprop polar 3 (compute the polarizability using fully numerical

efield 1e-5 derivatives, efield is used in the finite difference
end scheme and should be experimented with)

For this purpose - like for all numerical derivative schemes – very tight SCF criteria (VeryTightSCF) are highly
recommended. Polarizabilities with methods, for which analytical dipole moments are implemented (like MP2), a
mixed analytical / numerical scheme is available. Use the same input as above but specify polar 3 instead.

18.2 NMR Parameters

! NMR

Nuclear magnetic resonance chemical shielding constants are available for HF and DFT (excluding range-separated
hybrid functionals) levels of theory, as well as for RI-MP2 and double-hybrid DFT functionals (with RI). TPSS and
M06L are usually a good choice. Implicit solvation (CPCM) is also supported. It is recommended to use the pcSseg-n
basis sets in combination with the def2/JK fitting basis (even for non-hybrid functionals). For RIJCOSX a larger grid
should be used, e.g.

! GridX6 NoFinalGridX

%eprnmr GIAO_2el GIAO_2el_RIJCOSX end (also use RIJCOSX for GIAO integrals
GIAO_2el_RIJK is default)

Remember to run the calculation on the reference molecule (e.g. TMS) with the same method and basis set. The
chemical shift is then calculated as δmol = σref - σmol. Many more options are available and described in the manual

Indirect nuclear spin-spin coupling (J-coupling) constants are also available for HF and DFT:

%eprnmr Nuclei = all H {ssall, ist = 1 } (calculate J for all H nuclei

assuming they are all 1H)

SpinSpinRThresh 6.0 (couple nuclei up to 6 Angstroms apart)

end

18.3 EPR Parameters

13
The calculations of the EPR parameters (g-tensor, D-tensor, hyperfine coupling constant (HFC) etc.) can be
calculated at HF, DFT and hybrid DFT level of theory. The calculations for these parameters can be invoked with the
%eprnmr module in the input file.

%eprnmr

#calculation of g-tensors
gtensor 1
#calculation of D-tensor
dtensor so spin-orbit part
ss spin-spin part
ssandso both parts

#method for spin-orbit and spin-spin part

DSOC cp coupled-perturbed method (default)
DSS uno use of spin density from UNOs

#choice of gauge-origin
Ori centerofelcharge centre of the electric charge (default)
n Number of the atom user defines as the origin
x,y,z Origin of your choice

#method of solving CP-SCF equations

solver pople Pople solver (default)
cg conjugate gradient

#calculation of the hyperfine coupling constants

Nuclei = all types {FLAGS}

”all types” refers to the atom(s) of interest. For e.g., “all N {FLAGS}”
calculates HFC for all Nitrogen atoms or HFC parameters can also be
calculated for individual atoms of interest, for e.g. “Nuclei = 1, 3, 4
{FLAGS}”. IMP NOTE: Counting of atoms starts from 1 in this section!

FLAGS: aiso (Calculates isotropic part of the hyperfine couplings)

adip (Calculates dipolar term of the hyperfine coupling)

PrintLevel N amount of output required by the user, default is 2

end

NOTE: The correct choice of DFT functional and basis-set is important when calculating these properties
(especially using DFT or hybrid DFT methods), for e.g. hybrid DFT functionals (B3LYP, TPSSh, PBE0) give reliable
results. At even higher cost, double hybrids like B2PLYP give even better results. Basis-sets, such as “EPR-II, EPR-III,
IGLO-II and IGLO-III” add flexibility in the core region, thus performs well for the organic radicals.

In many instances, the calculation of EPR parameters on the basis of CASSCF/NEVPT2 wavefunctions is more
reliable than DFT (in particular for transition metal complexes). For details please see the CASSCF tutorial and main
manual.

14
18.4 Mößbauer Parameters
Requested through the EPRNMR module.

%eprnmr
nuclei = all Fe {fgrad, rho} (rho: electron density at Fe nuclei)
(fgrad: electric-field gradients and quadrupole splitting parameters)
origin CenterOfMass
CenterOfNucCharge
CenterOfElCharge (default)
PrintLevel n (controls the amount of output, default 2)
end

The use of core property basis CP(PPP) basis for iron is recommended for these calculations and this may be
requested via

%basis NewGTO Fe CP(PPP) end

The output file should contain the following information:

Mößbauer quadrupole splitting parameter (proper coordinate system)

e**2qQ = ---- MHz = ---- mm/s (quadrupole splitting)
eta = --- (asymmetry parameter 0≤eta≤1)
Delta-EQ=(1/2{e**2qQ}*sqrt(1+1/3*eta**2) = --- MHz = --- mm/s
RHO(0) = ----- a.u.**-3 (electron density at the Fe nucleus)

Calculation of Mößbauer parameters at the DLPNO-CCSD level of theory may be requested via

! RHF DLPNO-CCSD (for S=0 systems)

! UHF/UKS DLPNO-CCSD (for S≠0 systems)

and requesting in addition

%mdci DenMat UnRelaxed end

18.5 Vibrational spectra (IR, Raman, resonance Raman, NRVS)

! Freq

Automatically performs an analytical calculation of the IR spectrum.

! NumFreq

Automatically performs a numerical calculation of the IR spectrum.

%elprop polar 1
end
Performs a Raman calculation, including activities and depolarization ratios.

Raman/Polarizability calculations only work in combination with numfreq!

For NRVS calculations: run the following utility programs on the hessian file (extension: .hess) after completion of
the frequency calculation:

orca_vib MyJob.hess > MyJob.vib.out

orca_mapspc MyJob.vib.out NRVS

15
This produces a stick NRVS spectrum MyJob.vib.out.stk and a dressed NRVS spectrum
MyJob.vib.out.dat. These ASCII files can subsequently be imported in any plotting program (e.g. Origin).

NRVS calculations only work for iron containing systems.

For Resonance Raman calculations a three-step procedure is required:

1. Optimize molecule and create hessian file

! Opt Freq

2. Calculate electronic spectrum, normal mode gradients and dimensionless nuclear displacements

! NMGrad

%cis nroots 10 (number of roots to determine)

maxdim 100 (dimension of expansion space, maxdim*roots)
end

%rr
HessName "mycalc.hess" (name of hessian file from step 1)
ASAInput true (create input for step 3)
Tdnc 0.005 (cutoff threshold for dimensionless nuclear
displacements)
States 1,2,3,4,5,6,7,8,9,10 (list of excited states for which numerical
gradients and displacements will be
calculated)
end

3. Calculate Resonance Raman spectrum

Inspect and analyze the calculated electronic spectrum and determine which calculated bands correspond to the
experimentally observed bands. Shifts of up to 100 nm are not uncommon! Modify the newly created file
-1
mycalc.asa.inp to include the excitation energies in wavenumbers (e.g. 63500, 63800 and 64000 cm ) by
adding

RRSE 63500, 63800, 64000

The excitation energies have to be chosen in order to mimic experimental excitation into one of the bands of the
electronic spectrum. The energies are thus only identical to the energy of the laserlight in cases where perfect
agreement between the experimental and calculated electronic spectrum is reached.

Also adjust linewidth parameters of the electronic transitions as well as the linewidth parameter of the Resonance
Raman bands in the asa input file to match the experimentally observed linewidths. Then run

orca_asa mycalc.asa.inp > mycalc.asa.out

to obtain files

mycalc.asa.rrs.63500.stk
mycalc.asa.rrs.63500.dat
which contain the Resonance Raman spectra as a stick spectrum (.stk) and as a dressed spectrum (.dat).

16
18.6 Optical spectra
Ultraviolet and visible absorption spectroscopy (UV/vis), (electronic) circular dichroism spectroscopy (CD) and
magnetic circular dichroism spectroscopy (MCD) are methods to probe electronically excited states of molecules.

ORCA provides a multitude of different methods by which electronically excited states are accessible, most notably
TDDFT, CASSCF/NEVPT2, MRCI, ROCIS, as well as EOM-CC and STEOM-CC methods.

Transition dipole moments and rotational strengths (needed respectively for UV/Vis and CD spectra) are
automatically printed at the end of every TDDFT, CASSCF, NEVPT2 and MRCI calculation. For EOM-CC and STEOM-
CC methods, transition dipole moments are also automatically printed at the end of a calculation, but rotational
strengths are currently not available. In ROCIS, transition dipole moments are automatically printed, but
rotational strengths have to be explicitly requested by

%rocis
docd true
doquad true
end

MCD results can be generated by CASSCF (including NEVPT2), MRCI and ROCIS modules. The respective parts in the
input file are (please refer to the dedicated sections of the different methods for information on other keywords)

%casscf
rel
dosoc true (turn on spin-orbit coupling)
mcd true (turn on MCD)
B 43500 (static magnetic field in Gauss)
Temperature 299.0 (the experimental temperature)
end
end

%mrci
soc
dosoc true
mcd true
B 43500
Temperature 299.0
end
end

%rocis
soc true
domcd true
B 43500
SOCTemp 299.0
end

18.7 Absorption Fluorescence and Phosphorescence Lifetimes and Bandshapes

Emission rates and spectra, as well as absorption spectra can be calculated if one has at least a ground state
geometry and Hessian. First, optimize the ground state and get its Hessian using

! Opt Freq

Then call the ORCA_ESD module using ABS or FLUOR and give the Hessian file name, as

17
! ESD(ABS) or ESD(FLUOR)

%tddft nroots 10
end
%esd GSHessian “Basename.hess”
states 1,2,3,4...
end
The spectrum will be saved on a Basename.spectrum file and the rates will be printed on the output.

You have to select one excited state method in order to compute the gradient and transition dipoles. ORCA_ESD is
currently optimal for TDDFT, but it also works with ROCIS, (ST)EOM and CASSCF (without analytic gradients).

It is necessary that the geometry on the second input is the same as in the .hess file. It can be obtained from
Basename.xyz after optimization or copied from the .hess (in Bohrs then). To include vibronic coupling (Herzberg-
Teller effect) in weakly allow transition, set DOHT TRUE under %ESD.

If you want a higher quality spectra/rates, set Hessflag AHAS, with more methods discussed on the manual.

For phosphorscence, you have to also optimize the triplet using a UHF/UKS method and input both Hessians, as
well as the energy difference (in wavenumbers!) between them on each geometry

!ESD(PHOSP)

%tddft nroots 10
doSOC TRUE
end
%esd GSHessian “Basename.hess”
TSHessian “Basename_T.hess”
delE 20000
states 1
end
In this case, you also have to set the flag for including SOC on each module.

OPTIONAL:

Resonant Raman spectra can also be computed by using the same module, the main difference is that you have to
set a laser energy, as in

! ESD(RR)

%tddft nroots 10
end
%esd GSHessian “Basename.hess”
states 1,2,3,4…
laserE 15000
end

18.8 X-Ray absorption and emission spectra

ORCA features several approaches to X-ray absorption and emission spectra.

TD-DFT can be used to calculate the pre-edge K-edge X-ray absorption spectra of transition metal complexes.

%tddft
NRoots 80 (number of roots to determine)
MaxDim 500 (max. size of the expansion space)
OrbWin[0] = 0,0,-1,-1 (for an RKS reference function choose e.g.

18
 excitations from the 1s orbital (0,0) to the
entire (-1,-1) acceptor space)
Doquad true (request for Quadrupole and magnetic dipole
contributions
end

ROCIS or ROCIS/DFT can also be used to calculate X-ray absorption spectra. In fact these methods are especially
designed to treat the metal L/M-edge problems in transition metal complexes.

%rocis
NRoots 80
SOC true (request for SOC treatment)
DoRI true (request for RI treatment)
PrintLevel 3 (Control the amount of printing)
DoLowerMult true (Invokes a CI calculation with S’=S-1)
DoHigherMult true (Invokes a CI calculation with S’=S+1)
OrbWin = 6,8,0,2000 (request for excitations from the 2p orbital
(6,8) to an upper limit (default value is 2000)
(0,2000) acceptor space)
end

This approach can also be used to calculate resonant inelastic X-ray scattering (RIXS) and resonant X-ray emission
(RXES) spectra while the X-ray absorption spectra of large molecules can be calculated by employing the PNO
version of these methods (core PNO-ROCIS, core PNO-ROCIS/DFT). Further details are described in the manual

CASCI/NEVPT2 can also provide access to X-ray spectra. This is a multistep protocol and requires significant insight
from the user. In a nutshell:
1) The orbitals from a previous SA-CASSCF valence electron calculation are used as input
2) The frozen core treatment is explicitly deactivated
3) The core orbitals in request are rotated into the active space
4) The CASCI/NEVPT2 problem is then solved in the space of singly core-excited electronic configurations
using the previously optimized set of orbitals

!MORead

%moinp "MOs.gbw"

%method
FrozenCore FC_NONE
end

%scf
rotate {6,39,90} {7,40,90} {8,41,90} end
end

%casscf
nel 11 (number of electrons, core + valence)
norb 8 (number of orbitals, core + valence)
mult 6,4 (requested multiplicities)
nroots 16,173 (requested roots to be determined for mults 6 and 4)
nevpt2 true (request the NRVPT2 calculation)
maxiter 1 (number of iterations, need to set to 1 for the
CASCI/NEVPT2 calculation)
end

19
This approach can also be used to calculate RIXS and RXES spectra. Further details are provided in the manual.

An alternative and far more simplistic approach to X-ray absorption, and X-ray emission intensities is provided by
the one-electron approach based on the DFT orbitals and orbital energies.

! UKS BP86

%xes
coreorbs 0,0 (range of core orbitals to be treated)
orbop 0,1 (include spin-up and spin-down MOs)
end

This approach can also be used to calculate valence to core resonant X-ray emission spectra (VtC RXES). Further
details are provided in the manual.

19 Analysis tools and interfaces

19.1 Population analysis

By default, the Mulliken Population Analysis, Loewdin Population Analysis and Mayer Population Analysis
are automatically performed after SCF calculations. To turn off a certain population analysis, use keyword:

! NoMulliken

Turns off Mulliken population analysis

! NoLoewdin

Turns off Loewdin population analysis

! NoMayer

Turns off Mayer population analysis

To turn them on simply type “Mulliken”, “Loewdin” or “Mayer” in the input. Other useful keywords include:

! Allpop

Turns on all population analysis

! NoPop

Turns off all population analysis

19.2 Local energy decomposition (LED)

The local energy decomposition is a way to analyze the DLPNO coupled cluster energy in physical terms (such as
electrostatic interactions and dispersion energy). All that is required is that the calculation is a DLPNO-CCSD or
DLPNO-CCSD(T) calculation and that the atoms are assigned to fragments as in the example below. The LED is
requested by:

20
! LED
%mdci printlevel 3
end
* xyz 0 1
O(1) -0.03927172712553 -0.05314132591150 0.08218758558907
H(1) 0.93067536168966 -0.05970332184924 0.07804455467818
H(1) -0.28396499015190 0.88544835339923 0.07926277816645
O(2) -0.90315523402475 -1.18937566538863 2.59796265057253
H(2) -0.64227331879731 -0.83788044760509 1.72189785393425
H(2) -1.47240409159018 -1.94579659264477 2.39478357705952
*
Turns on the LED calculation (! LED) and prints additional information (printlevel 3). Each water in the
dimer in this example is assigned to an own fragment in the geometry block.

19.3 Ab initio ligand field theory (AILFT)

The ab initio ligand field (AILFT) analysis can be requested by:

%casscf actorbs dorbs or forbs

end

This requires that the active orbitals are just the five metal based d-orbitals in case of d-block elements or just the
seven f-orbitals in the f-block elements. Making sure that this happens is the user’s responsibility. The calculation
will automatically spit out the ligand field and Racah parameters that can then be further analyzed manually.

It is recommended to take reference of the “CASSCF-tutorial.pdf” in the ORCA forum.

19.4 Natural Bond Orbital analysis

! NPA

Turns on NPA analysis

! NBO
Turns on NBO analysis

In addition to using the HF density, one can use post-HF densities for NBO analysis

! MP2 NBO

%MP2 density relaxed

end

Turns on NBO analysis using the relaxed MP2 density.

The calculation will produce a .47 file for expert users that can be manipulated before running the NBO program
standalone.

19.5 Atoms in molecules analysis

! AIM

Generate the .wfn file for topological analysis.

Alternatively one can use orca_2aim reading a .gbw file to generate the .wfn file.

21
20 Visualization tools and interfaces
(1) Visualize canonical orbitals:

Step1: Transform the .gbw file to “Molden” format by

orca_2mkl “filename” -molden

(note: no suffix following the “filename”)

Step2: Read in the generated file by any visualization tool, such as Avogadro.

Other type of orbitals stored in the .gbw format can be visualized in the same way. Just change the suffix to gbw.
e.g. change the suffix of a localized orbital file:

mv filename.loc filename.gbw

(2) Visualize densities:

Step1: Call the interactive program orca_plot and follow the instructions.

orca_plot filename.gbw -i

Step2: Use any visualization tool to read the generated file, such as Chimera.

(3) Visualize trajectories:

Step1: Perhaps the most convenient way is to use Molden to read the .trj file:

Molden filename.trj

(4) Visualize vibrations:

Step1: Perhaps the most convenient way is to use Avogadro to read the output file:

Avogadro filename.out

(5) Plot spectra:

The spectra can be created by using “orca_mapspc” program:

orca_mapspc outputfile -(type of spectrum) -options

The type of spectra include:

ABS ABSV ABSQ CD IR Raman

NRVS VDOS MCD SOCABS XES XESV

XESQ XAS XASV XASQ XESSOC XASSOC

The options include:

-o output file

22
-cm use cm-1 (default)
-eV use eV (default cm-1)
-g use Gaussian lineshape default
-l use Lorentz lineshape
-x0 initial point of spectrum
-x1 final point of spectrum
-w line width
-kw coeffitient for the line width calculated as kw*sqrt(energy)
-n number of points

-1
Example: Plot an IR spectrum in the region 300-4000 cm , using default broadening

orca_mapspc jobname.out IR -x0300 -x14000

More examples can be found in ORCA Input Library

21 ORCA File Types

ORCA generates a wide range of files, depending on the type of calculation used.

.gbw contains information on Geometry, Basis and Wavefunction (Orbitals). Can be used as initial
guess in subsequent calculation. Can be used to plot MOs, electron density, … with orca_plot
.loc gbw type file, contains information localized orbitals (requested via %loc block)
.uno gbw type file, contains information on unrestricted natural orbitals (requested by !UNO).
.unso gbw type file, contains information on unrestricted natural spin orbitals (requested by !UNO).
.qro gbw type file, contains information on quasi-restricted orbitals (requested by !UNO).
.uco gbw type file, contains information on unrestricted corresponding orbitals (requested by !UCO).
.xyz coordinate file, contains the optimized structure after a geometry optimization. If the
optimization was not successful, it contains the geometry of the last step (requested by !Opt).
Can be visualized by standard molecular visualization programs.
.trj trajectory of geometry optimization. Contains the entire trajectory of all steps for relaxed surface
scans (requested by !Opt %geom Scan … end end).
.hess contains Hessian matrix (frequency calculation). Can be used as input to orca_vib or
orca_pltvib.
.00n.xyz coordinate file of the nth optimized structure of a relaxed surface scan.
.00n.gbw gbw file of the nth optimized structure of a relaxed surface scan.
.allxyz coordinate file, contains all optimized structures of a relaxed surface scan. Can be used as xyzfile
for SP calculations on all optimized structures of a scan.
.relaxscanscf.dat data file, contains the SCF energy for all optimized structures of the surface scan
.relaxscanact.dat data file, contains actual (SCF or post-HF) energy for all optimized structures of the surface scan
.nto gbw type file containing information on natural transition orbitals in TDDFT calculations.
.scfp contains the electron density matrix (requested by !KeepDens)
.scfr contains the spin density matrix (requested by !KeepDens)

23