8.16.1.2 Molecular structure: coordinates, symmetry and basis setsThere are three different ways to specify the molecular structure, symmetry and the basis sets in Gateway:
The three different modes will be described below.
8.16.1.2.1 Native inputIf the geometry is specified in a native MOLCAS format, only symmetry inequivalent atoms should be specified. The default units are atomic units. By default, symmetry is not used in the calculation.
Example of an input in native MOLCAS format:
8.16.1.2.2 Zmatrix and XYZ inputSome times it is more convenient to set up information about coordinates in a standard form of Zmatrix or Cartesian coordinates. In this case, the basis set for the atoms should be specified after the XBAS keyword. After that either ZMAT or XYZ should appear to specify the coordinates. Note that coordinates in these formats use ångström as units.
8.16.1.2.3 Advanced XYZ inputIf the geometry is specified in XYZ format, all atoms should be specified. The default units are Ångstroms. By default, maximum possible symmetry is used.'Molcas XYZ' file format is an extension of plain XYZ format.
If XYZ input has been used in gateway, a file with native MOLCAS input will be produced and stored in working directory under the name findsym.std. Note that choosing XYZ input you are expecting that the coordinates might be transformed. It can be shown by the following example:
&gateway coord 3 O 0 0 0 H 1.0000001 0 0 H 0 1 0.0000001 *nomove *group=c1 The geometry of the molecule is slightly distorted, but within a threshold it is C_{2v}. Thus by default (keywords nomove and group are not active), the coordinates will be transformed to maintain the highest possible symmetry. If keyword nomove is active, the molecule is not allowed to rotate, and a mirror plane XY is the only symmetry element. Since the third hydrogen atom is slightly out of this plane, it will be corrected. Only activation of the keyword group=C1 will ensure that the geometry is unchanged. Advanced keywords:
or, in short EMIL notation:
Coordinate file may contain variables, as demonstrated in an example:
The atom name in XYZ file can contain an orbitrary label (to simplify assigning of different basis sets). To indicate the label, use _: e.g. . The same label should be defined in the basis section: . The basis set label can be also added into the name of an element:
XYZ file can also contain information about point charges. There are three possibilities to setup atomic charges in XYZ file. The main option is to use Q as an element name, and in this case the forth number, the charge, should be specified. Another possibility is to use element names ended with minus sign. In this case, a formal charge for the element will be used. E.g. H, Li, Na, K defines +1 charge located in the corresponding location. Mg, Ca  defines charge +2, Al  +3, C, Si +4, for anions, F, Cl, Br, I defines 1, O, S  2. Finally, a label at the comment line of XYZ file  CLUSTER followed by an integer number can specify how many atoms are 'real', so the rest will be treated as charges with default values for this element.
8.16.1.3 ConstraintsIn case of optimizations with constraints these are defined in the GATEWAY input. For a complete description of this keyword see the section .
8.16.1.4 Explicit auxiliary basis setsThe socalled Resolution of Identity (RI) technique (also called Density Fitting, DF) is implemented in the MOLCAS package. This option involves the use of an auxiliary basis set in the effective computation of the 2electron integrals. MOLCAS incorporates both the use of conventionally computed, externally provided, auxiliary basis sets (RIJ, RIJK, and RIC types), and onthefly generated auxiliary basis sets. The latter are atomic CD (aCD) or the atomic compact CD (acCD) basis sets, based on the Cholesky decomposition method. The externally provided auxiliary basis sets are very compact, since they are tailored for special wave function methods. However, they are not provided for all available valence basis sets. The aCD or acCD RI auxiliary basis sets are a more general option and provides auxiliary basis sets for any wave function model and valence basis set.

File  Contents 
ONEINT  Oneelectron integral file used to store the Pauli repulsion integrals 
RUNFILE  Communications file. The last computed selfconsistent reaction field (SCF or RASSCF) will be stored here to be used by following programs 
GV.off  Input file for the external program ``geomview'' (see Tutorial section ``Solvent models''), for the visualization of PCM cavities 
Compulsory keywords
Keyword  Meaning 
RFInput  Activate reaction field options.

END Of RFInput  This marks the end of the input to the reaction field utility. 
Optional keywords for the Kirkwood Model
Keyword  Meaning 
REACtion Field  This command is exclusive to the Kirkwood model. It indicates the beginning of the specification of the reaction field parameters. The subsequent line will contain the dielectric constant of the medium, the radius of the cavity in Bohrs (the cavity is always centered around the origin), and the angular quantum number of the highest multipole moment used in the expansion of the change distribution of the molecule (only charge is specified as 0, charge and dipole moments as 1, etc.). The input specified below specifies that a dielectric permitivity of 80.0 is used, that the cavity radius is 14.00 a.u., and that the expansion of the charge distribution is truncated after l=4, i.e. hexadecapole moments are the last moments included in the expansion. Optionally a fourth argument can be added giving the value of the dielectric constant of the fast component of the solvent (default value 1.0). 
RFInput
Reaction field
80.0 14.00 4
End Of RFInput
Sample input for a complete reaction field calculation using the Kirkwood model. The SCF computes the reaction field in a self consistent manner while the MRCI program adds the effect as a constant perturbation.
&GATEWAY
Title = HF molecule
Symmetry
X Y
Basis set
F.ANOS...3S2P.
F 0.00000 0.00000 1.73300
End of basis
Basis set
H.ANOS...2S.
H 0.00000 0.00000 0.00000
End of basis
Well integrals
4
1.0 5.0 6.75
1.0 3.5 7.75
1.0 2.0 9.75
1.0 1.4 11.75
RFInput
Reaction field
80.0 4.75 4
End of RFInput
&SEWARD
&SCF
Occupied = 3 1 1 0
&MOTRA
LumOrb
Frozen = 1 0 0 0
RFPert
&GUGA
Electrons = 8
Spin = 1
Inactive = 2 1 1 0
Active = 0 0 0 0
CiAll = 1
&MRCI
SDCI
Optional keywords for the PCM Model
Keyword  Meaning 
PCMmodel  If no other keywords are specified, the program will execute a standard PCM calculation with water as solvent. The solvent reaction field will be included in all the programs (SCF, RASSCF, CASPT2, etc) invoked after SEWARD: note that in some cases additional keywords are required in the corresponding program sections. Some PCM parameters can be changed through the following keywords. 
SOLVent  Used to indicate which solvent is to be simulated. The name of the requested solvent must be written in the line below this keyword. Find implemented solvents in the PCM model below this section. 
DIELectric constant  Defines a different dielectric constant for the selected solvent; useful to describe the system at temperatures other that 298 K, or to mimic solvent mixtures. The value is read in the line below the keyword. An optional second value might be added on the same line which defines a different value for the infinite frequency dielectric constant for the selected solvent (this is used in nonequilibrium calculations; by default it is defined for each solvent at 298 K). 
CONDuctor version  It requires a PCM calculation where the solvent is represented as a polarized conductor: this is an approximation to the dielectric model which works very well for polar solvents (i.e. dielectric constant greater than about 5), and it has some computational advantages being based on simpler equations. It can be useful in cases when the dielectric model shows some convergence problems. 
AAREa  It is used to define the average area (in Å^{2}) of the small elements on the cavity surface where solvation charges are placed; when larger elements are chosen, less charges are defined, what speeds up the calculation but risks to worsen the results. The default value is 0.4 Å^{2} (i.e. 60 charges on a sphere of radius 2 Å). The value is read in the line below the keyword. 
RMIn  It sets the minimum radius (in Å) of the spheres that the program adds to the atomic spheres in order to smooth the cavity surface (default 0.2 Å). For large solute, if the programs complains that too many sphere are being created, or if computational times become too high, it can be useful to enlarge this value (for example to 1 or 1.5 Å), thus reducing the number of added spheres. The value is read in the line below the keyword. 
PAULing  It invokes the use of Pauling's radii to build the solute cavity: in this case, hydrogens get their own sphere (radius 1.2 Å). 
SPHEre radius  It is used to provide sphere radii from input: for each sphere given
explicitly by the user, the keyword ``Sphere radius'' is required,
followed by a line containing two numbers: an integer indicating the
atom where the sphere has to be centered, and a real indicating its
radius (in Å). For example, ``Sphere radius'' followed by ``3 1.5''
indicates that a sphere of radius 1.5 Å is placed around atom #3;
``Sphere radius'' followed by ``4 2.0'' indicates that another sphere of
radius 2 Å is placed around atom #4 and so on.

Solvents implemented in the PCM model are
Name  Dielectric  Name  Dielectric  Name  Dielectric  
constant  constant  constant  
water  78.39  dichloroethane  10.36  toluene  2.38  
dimethylsulfoxide  46.70  quinoline  9.03  benzene  2.25  
nitromethane  38.20  methylenchloride  8.93  carbontetrachloride  2.23  
acetonitrile  36.64  tetrahydrofuran  7.58  cyclohexane  2.02  
methanol  32.63  aniline  6.89  heptane  1.92  
ethanol  24.55  chlorobenzene  5.62  xenon  1.71  
acetone  20.70  chloroform  4.90  krypton  1.52  
isoquinoline  10.43  ethylether  4.34  argon  1.43 
Sample input for the reaction field part (PCM model): the solvent is water, a surface element average area of 0.2 Å^{2} is requested.
RFinput
PCMmodel
Solvent
water
AAre
0.2
End of RFinput
Sample input for a standard PCM calculation in water. The SCF and RASSCF programs compute the reaction field self consistently and add its contribution to the Hamiltonian. The RASSCF is repeated twice: first the ground state is determined, then a nonequilibrium calculation on the first excited state is performed.
&GATEWAY
Coord
4
formaldehyde
O 0.000000 0.000000 1.241209
C 0.000000 0.000000 0.000000
H 0.000000 0.949585 0.584974
H 0.000000 0.949585 0.584974
Basis = STO3G
Group = C1
RFinput
PCMmodel
solvent = water
End of RFinput
&SEWARD ; &SCF
&RASSCF
nActEl = 4 0 0
Symmetry = 1
Inactive = 6
Ras2 = 3
CiRoot
1 1
1
LumOrb
&RASSCF
nActEl = 4 0 0
Symmetry = 1
Inactive = 6
Ras2 = 3
CiRoot
2 2
1 2
0 1
JOBIPH
NonEq
RFRoot = 2
Again the user is recommended to read section of the
examples manual for further details.
Keyword  Meaning 
FNMC  Request that the socalled Finite Nuclear Mass Correction exclude, by the BornOppenheimer approximation, be added to the oneelectron Hamiltonian. 
WELL integrals  Request computation of Pauli repulsion integrals for dielectric cavity reaction field calculations. The first line specifies the total number of primitive well integrals in the repulsion integral. Then follows a number of lines, one for each well integral, specifying the coefficient of the well integral in the linear combination of the well integrals which defines the repulsion integral, the exponent of the well integral, and the distance of the center of the Gaussian from the origin. In total three entries on each line. All entries in atomic units. If zero or a negative number is specified for the number of well integrals a standard set of 3 integrals with their position adjusted for the radius of the cavity will be used. If the distance of the center of the Gaussian from the origin is negative displacements relative to the cavity radius is assumed. 
XFIEld integrals  Request the presence of an external electric field represented by a number of partial charges and dipoles. Optionally, polarisabilities may be specified whose induced dipoles are determined selfconsistently during the SCF iteration. The first line may contain, apart from the first integer [nXF] (number of centers), up to four additional integers. The second integer [nOrd] specifies the maximum multipole order, or 1 signifying no permanent multipoles. Default is 1 (charges and dipoles). The third integer [p] specifies the type of external polarisabilities: 0 (default) no polarisabilities, 1 (isotropic), or 2 (anisotropic). The fourth integer [nFrag] specifies the number of fragments one multipole may contribute to (relevant only if polarisabilities are present). The default is 0, meaning that each permanent multipole is only excluded in the calculation of the field at its own polarisability, 1 means that one gives a fragment number to each multipole and that the static multipoles do not contribute to the polarising field within the same fragment, whereas 2 can be used in more complex situations, e.g. polymers, allowing you to specify a second fragment number so that junction atoms does not contribute to either of the neighbouring fragments. Finally, the fifth and last integer [nRead] (relevant only if Langevin dipoles are used) may be 0 or 1 (where 0 is default), specifying wheather an element number (e.g. 8 for oxygen) should be read for each multipole. In that case the default radius for that element is used to determine which Langevin grid points should be annihilated. A negative element number signifies that a particular radius should be used for that multipole, in thousands of a Bohr (1400 meaning 1.4 Bohr). Then follows nXF lines, one for each center. On each line is first nFrag+nRead (which may equal 0) integers, specifying the fragments that the multipole should not contribute to (the first fragment is taken as the fragment that the polarisability belongs to) and the element number. Then follows the three coordinates of the center, followed by the multipoles and polarisabilities. The number of multipole entries is 0 for nOrd=1, 1 for nOrd=0, 4 for nOrd=1, and 10 for nOrd=2. The number of polarisability entries are 0 for p=0, 1 for p=1, and 6 for p=2. The order of quadrupole moment and anisotropic polarisability entries is xx, xy, xz, yy, yz, zz. If default is used, i.e. only specifying the number of centers on the first line, each of these lines will contain 7 entries (coordinates, charge, and dipole vector). All entries are in atomic units, if not otherwise requested by the Angstrom keyword that must be placed between nXF and nOrd. All these data can be stored in a separate file whose name must be passed as an argument of the XField keyword. 
ANGM  Supplement ONEINT for transition angular momentum calculations. Entry which specifies the angular momentum origin (in au). 
OMQI  Supplement ONEINT for transition orbital magnetic quadrupole calculations. Entry which specifies the orbital magnetic quadrupole origin (in au). 
AMPR  Request the computation of angular momentum product integrals. The keyword is followed by values which specifies the angular momentum origin (in au). 
DSHD  Requests the computation of diamagnetic shielding integrals. The first
entry specifies the gauge origin. Then follows an integer
specifying the number of points at which the diamagnetic
shielding will be computed. If this entry is zero, the diamagnetic
shielding will be computed at each nucleus. If nonzero, then the
coordinates (in au) for each origin has to be supplied, one entry for each
origin.

EPOT  An integer follows which represents the number of points for which the electric potential will be computed. If this number is zero, the electric field acting on each nucleus will be computed. If nonzero, then the coordinates (in au) for each point have to be supplied, one entry for each point. This keyword is mutually exclusive with EFLD and FLDG. 
EFLD  An integer follows which represents the number of points for which the electric potential and electric field will be computed. If this number is zero, the electric field acting on each nucleus will be computed. If nonzero, then the coordinates (in au) for each point have to be supplied, one entry for each point. This keyword is mutually exclusive with EPOT and FLDG. 
FLDG  An integer required which represents the number of points for which the electric potential, electric field and electric field gradient will be computed. If this number is zero, the electric field gradient acting on each nucleus will be computed. If nonzero, then either the coordinates (in au) for each point or labels for each atom center have to be supplied, one entry for each point. In case a label is supplied it must match one of those given previous in the input during specification of the coordinates of the atom centers. Using a label instead of a coordinate can e.g. be useful in something like a geometry optimization where the coordinate isn't known when the input is written. This keyword is mutually exclusive with EPOT and EFLD. 
EMPC  Use point charges specified by the keyword XField when calculating the OrbitalFree Embedding potential. 
RFInput  Specification of reaction field parameters, consult the reaction field section of this manual. 
Keyword  Meaning 
FINIte  Request a finite center representation of the nuclei by a single exponent stype Gaussian. 
MGAUSsian  Request a finite center representation of the nuclei by a modified Gaussian. 
Keyword  Meaning 
RPCoordinates  This activates the Saddle method for TS geometry optimization.
The line is followed by an integer specifying the number of symmetry unique coordinates to be specified. This
is followed by two sets of input  one line with the energy and then the Cartesian coordinates in bohr  for
each of the two starting structures of the Saddle method. Note that the order of the coordinates must always
match the order specified with the conventional input of the coordinates of the molecular system.
Alternatively, two lines with the filenames containing the coordinates of reactants and products, respectively,
(in XYZ format) can be given.

NOALign  By default, the two starting structures are aligned to minimize the root mean square distance (RMSD) between them,
in particular, the first structure is moved and the second structure remains fixed.
If this keyword is given, the starting structures are used as given.

ALIGn only  The two starting structures are aligned, but nothing more is done.
An input block for seward is still needed, but no integrals are computed.

WEIGhts  Relative weights of each atom to use for the alignment and for the calculations of the
``distance'' between structures. The possibilities are:
MASS. This is the default. Each atom is given a weight proportional to its mass. Equivalent to massweighted coordinates. EQUAL. All atoms have an equal weight. HEAVY. Only heavy atoms are considered, with equal weights. Hydrogens are given zero weight. A list of N numbers can also be provided, and they will be used as weights for the N symmetryunique atoms. For example:
will align only atoms 712 out of 16. Note that, in any case, weights of 0 are likely to cause problems with constraints, and they will be increased automatically.

SADDle  Step size reduction for each macro iteration of the saddle method.
The value is given in weighted coordinates, divided by the square root of the total weight
(see the WEIGHTS keyword).
Default value is 0.1 au.

H
O 0.982011 0 1
H 0.982013 0 104.959565 0 2 1
H 1.933697 1 107.655494 1 114.496053 1 2 3 1
O 0.988177 0 173.057942 1 56.200750 1 4 2 3
H 0.979890 0 104.714572 0 179.879745 1 5 4 2
where the three columns of real numbers are internal coordinates, and the last three columns of integers indicate which other atoms that are used to define the coordinate. The type of coordinates from left to right are bond distances, bond angles and dihedral angels, both for the coordinates and the link. The column of integers just to the right of each coordinate indicate if this coordinate should be optimized or not (1 = optimize, 0 = do not optimize).
There are also two utilitykeywords used to create a zmatrix or to write out a constraintdefinition for slapaf and keywords to rotate and translate fragments. (See documentation for GEO for more details)
Keyword  Meaning 
HYPER  This keyword is used to specify that a geometry optimization with constrained
internal coordinates shall be performed later, a zmatrix and a set of
displaced geometries are therefore constructed. The keyword should be followed by three
real numbers defining the maximum displacement for each coordinate type.
The order from left to right is bond distances, bond angles and dihedral angles.
To use default values for the parameters the mutually exclusive keyword
geo should be entered instead.

GEO  This keyword is used to specify that a geometry optimization with constrained
internal coordinates shall be performed later, a zmatrix and a set of displaced
geometries are therefore constructed. Default values of 0.15 Å, 2.5 degrees,
and 2.5 degrees are used for the maximum displacement of bond distances, bond
angles and dihedral angles respectively. To enter other values for the parameters
the mutually exclusive keyword hyper should be used.

OPTH  This keyword is used to define the specific details of the optimization algorithm used
for the geometry optimization in constrained internal coordinates.
This keyword should be followed by two to three lines of parameter. The first line should
contain an integer indicating optimization type (1 = steepest descent, 2 = a mix of
steepest descent and Newton's method, and 3 = Newton's method). The second line
should contain a real number defining a step factor.
This number is multiplied with the gradient to obtain the step length.
For optimization type 2 a third line containing a real number that defines a gradient limit
should be entered. This limit determines how large the gradient must be for the steepest
descent algorithm to be used. When the gradient is smaller than this limit Newton's method
is used instead.

OLDZ  This keyword is used both to start a new calculation from a userdefined zmatrix and
to restart calculations. When using the keyword for a new calculation a directory
$Project.GEO must exist and contain a file called $Project.zmt with a zmatrix in
the format defined above. The directory must not contain any files with the suffix .info
when performing a fresh calculation since these files contain restart information.

ZONLY  This keyword is used to construct a zmatrix from a set of xyzfiles (fragments)
and store it in the directory $Project.GEO. The optimization parameters
of the resulting zmatrix are set so that only coordinates linking fragments are
set to 1 (= optimize coordinate).

ZCONS  This keyword is used to define constraints from a set of xyzfiles (fragments)
on a form that could be supplied to the
slapaf in order to keep the fragments rigid. The resulting constraintsfile
is named $Project.cns and stored in the directory $Project.GEO. The
atomnumbers in this constraintfile will not match those of your original xyzfile and
should not be used together with this. Instead a new xyzfile named cons.xyz is created
and placed into the directory $Project.GEO, this has the proper numbering to use together with the constraints.

ORIGIN  This keyword is used to translate and rotate a set of xyzfiles. The keyword must be entered
before the xyzfiles is entered with coord. The keyword should be followed by one
line with the number of fragments and then one line for each fragment that should be translated.
This row should contain 13 numbers. One integer defining which fragment that should be moved, (the fragments are numbered based on order of appearance in the inputfile from top to bottom), 3
real numbers defining a translation (x, y, z) and 9 numbers defining a rotation (row1, row2, row3 of
rotation matrix). The keyword origin is mutually exclusive with the keyword frgm
which is an alternative way to express the same rotations and translations.

FRGM  This keyword is used together with the keywords rot and trans to define
rotation and translation of a specific fragment. Frgm defines an active fragment (each xyzfile is considered a fragment, the files are numbered based on
order of appearance in the input from top to bottom). The keyword must be entered before the xyzfile it is supposed to modify is
entered with coord. Each occurence of
frgm should be followed by either one of or both of the keywords rot and trans
to define rotation and translation. This keyword is mutually exclusive with the keyword orgin

ROT  This keyword should be followed by nine real numbers defining the rotation for the fragment defined by
the preceeding frgm. The numbers should be the nine elements of a rotation matrix
listed with one full row at the time.

TRANS  This keyword should be followed by three real numbers defining the translation for the fragment defined
by the preceeding frgm. The numbers should be the x, y and z coordinates of the translation
in that order.

Example of an input:
&GATEWAY
Title
Water Dimer
frgm=2
trans=3.0 0.0 0.0
Coord=water_monomer.xyz
Coord=water_monomer.xyz
Group=c1
basis=ccpVTZ
hyper
0.2 3.0 3.0
opth
3
15.0d0
In this example a water dimer is constructed from a single monomer by translating it 3.0 Åwith the keyword trans. An optimization in constrained internal coordinates using newtons method with a stepfactor of 15.0d0 are prepared for. For more details on these optimization see the manual entry for the module geo.
The following keywords apply to QM/MM calculations performed with the MOLCAS/GROMACS interface (see section for more details).
Keyword  Meaning 
GROMacs  Requests that the definition of the full QM+MM system should be imported from GROMACS. The keyword should be followed by one of the options SIMPLE or CASTMM on the next line. In the case of SIMPLE, all MM atoms defined in the GROMACS input will be treated as outer MM atoms in MOLCAS. This means, for example, that in a geometry optimization, their positions will be updated using microiterations rather than the conventional optimization scheme. Conversely, CASTMM requests that certain MM atoms should be treated as inner MM atoms in MOLCAS. Their positions will be updated with the same scheme as used for the QM atoms. The CASTMM option should be followed by two additional input lines, the first one containing the number of MM atoms to convert from outer to inner type, and the second containing a list of those atoms (using their corresponding GROMACS indices).

LINKatoms  Defines link atoms for use with the Morokuma updating scheme. The desired number of link atoms should be given as an integer on the next line. This should be followed by additional input lines, one for each link atom to be defined. Each definition should be of the form ILA, IQM, IMM, SCALE, where ILA, IQM and IMM are the GROMACS indices of the link atom and the corresponding QM and MM frontier atoms, respectively. SCALE is the scaling factor to be used in the Morokuma scheme. Note that each link atom must be defined as a QM atom in the GROMACS input. In addition, the frontier MM atom must be an inner MM atom specified as discussed above.
