How to specify the molecule and basis set using xyz files

The actual xyz file only contains the number of centers, the atoms, and the corresponding coordinates (by default in angstroms). It could look like this one:

6
Methanol, identifying the alcoholic proton as H1
C      -0.000000010000       0.138569980000       0.355570700000
O      -0.000000010000       0.187935770000      -1.074466460000
H       0.882876920000      -0.383123830000       0.697839450000
H      -0.882876940000      -0.383123830000       0.697839450000
H      -0.000000010000       1.145042790000       0.750208830000
H1     -0.000000010000      -0.705300580000      -1.426986340000

The first line is the number of atoms. The second line can be blank or an arbitrary comment. It is not read by the program. If you ommit this line, then your molecule will be wrong, one atom will be missing! The following lines contain atom types and coordinates. The first entry on each line is the name of the atom. Any string up to eight characters will do, but if the name is present in the periodic table DIRAC will recognize the charge and can find the corresponding basis set.

From the xyz file DIRAC cannot know what basis set you want to use. For this reason, basis set and symmetry specification is then done in the input file using the keywords given below. This is in contrast to the traditional mol files. This means that the keywords below do not make sense when combined with a traditional mol file.

*BASIS

Here the basis set information is specified. .DEFAULT specifies the default large component basis set for all atoms. This can be modified for specific atom by the .SPECIAL keyword. For example, if we use the methanol.xyz file listed above, the following input specifies a cc-pVDZ on carbon and the 3 hydrogens labeled H, while oxygen and H1 will be treated with a cc-pVTZ basis:

**MOLECULE
*BASIS
.DEFAULT
 cc-pVDZ
.SPECIAL
 O BASIS cc-pVTZ
.SPECIAL
 H1 BASIS cc-pVTZ

*CENTERS

Here the physical properties of the centers in our molecule can be specified. This is useful if the name of the center differs from the IUPAC element name or in case we want to use nuclei with fractional charges. This is controlled by the keyword nucleus that can be repeated as often as is needed. For the geometry file methanol.xyz we need to specify:

*CENTERS
.NUCLEUS
 H1 1.0

Another reason to change the charge of the nucleus is to be able to run counterpoise calculations in which a ghost basis set is to be placed without adding a charge. This can be achieved by defining the ghost center as (for instance) O.Gh and then putting the nuclear charge to zero using:

*CENTERS
.NUCLEUS
 O.Gh 0.0 8.0

the first number is the charge to be used, the second corresponds to the element number and is used to search the basis set library. In this example you can therefore simply specify the basis set name using the DEFAULT or SPECIAL keywords and the program will then find the corresponding oxygen basis.

*COORDINATES

Here you can specify the units to be used when reading the coordinates. If nothing is specified, angstrom is the default (this is then a real xyz file). If we specify ‘AU’ we use atomic units instead:

*COORDINATES
.UNITS
AU

We can also specify our own unit, by typing the name and the conversion factor when going from this new unit to atomic units:

*COORDINATES
.UNITS
nm 18.8972

Now all coordinates are entered in nanometer. If we add ‘A’ at the end, we can enter the new unit in angstrom:

*COORDINATES
.UNITS
nm 10.0 A

*SYMMETRY

With this keyword we can control symmetry recognition. To turn symmetry detection off and to force C1 symmetry use:

*SYMMETRY
.NOSYM

Including ghost centers and point charges

In some situations you may want to add a ghost center, for instance in counterpoise calculations (see Basis set superposition error in the DFT) or to lower the symmetry detected by DIRAC (see Electronic structure of CmF). An example of the latter case is to take out inversion symmetry for an atom:

2

Ne    0.0 0.0 0.0
foo   0.0 0.0 10.0

We give a non-standard name to the ghost center, so that DIRAC does not find a nuclear charge. This has to be accompanied by telling DIRAC not to add a basis to the ghost center:

**MOLECULE
*BASIS
.DEFAULT
dyall.2zp
.SPECIAL
foo NOBASIS

In a counterpoise calculation, on the other hand, you want the center to carry a basis set. For a ghost neon atom you would then specify:

**MOLECULE
*CENTERS
.NUCLEUS
foo 0.0 10.0

As already explained, the first number on the last line is the charge of the center, and the second the nuclear charge used to find the basis. If you actually want to add a point charge, let us say of charge -2.3, you can use the xyz-file above, but now you want charge, but not basis, and write:

**MOLECULE
*BASIS
.DEFAULT
dyall.2zp
.SPECIAL
foo NOBASIS
*CENTERS
.NUCLEUS
foo -2.3

Examples of xyz-based symmetry detection

We provide (in test/xyz_symmetry_recognition) collection of geometries aimed for the DIRAC automatic symmetry recognition of molecular structures given in the xyz-coordinates format.

Sources of xyz-inputs

Molecules were handled by a suitable software and their geometries (coordinates), reflecting imposed symmetry, were exported, preferably in the xyz-format.

If the GUI software does not save/export the geometry coordinates in xyz-format, try to save (export) them in another common format, and read it in an other software, capable to export the desired xyz file. Note that sometimes the resulting xyz-geometry file had to be manually controlled and coordinates corrected, if necessary, to keep the symmetry of the system.

Very good choice for getting symmetry imposed xyz coordinates is the GUI of the commercial ADF software (www.scm.com) .

List of symmetries

The DIRAC’s hersym.F/FIND_PGROUP subroutine can handle the following list of point (full) groups:

  • O(h)
  • I(h)
  • T(d)
  • D(oo,h)
  • D(nh)
  • D(nd)
  • D(n)
  • C(oo,v)
  • C(nv)
  • C(nh)
  • C(n)
  • C(s)
  • C(i)
  • C(1)
  • S(n)

The detected full group is then turned into lower computational group, represented as the point group.

Examples

The following table contains examples of the symmetry detection from the xyz coordinates.

xyz input molecule; imposed symmetry full group represented as
molecule01 heterocycycle; D(2h) D(2h) D2h
molecule02 Re2(CO)10; D(4d) C(4v) (!) C2v
molecule03 Os(CO)5; D(3h) C(2v) (!) C2v
molecule04 acetylene; D(oo,h) D(oo,h) D2h
molecule05 SF6; O(h) O(h) D2h
molecule06 allene; S(4) D(2d) (!) D2
molecule07 NH3; C(3v) C(3v) Cs
molecule08 PF5; D(3h) D(3h) C2v
molecule09 Hs(CO)4; {close to D(2d)} D(2) (!) D2
molecule10 ethane; D(3d) C(2h) (!) C2h
molecule11 dodecahedrane; I(h) C(2h) (!) C2h
molecule12 dodecahedrane; D(5d) D(5d) C2h
molecule13 benzene; D(6h) D(6h) D2h
molecule14 methane; T(d) T(d) D2
molecule15 H2O2; C(2h) C(2h) C2h
molecule16 ferrocene; D(5h) C(s) (!) C2v
molecule17 ferrocene-stagg; D(5d) C(2h) (!) C2h
molecule18 Hs(CO)4; D(2d) D(2d) D2

The (!) symbol means that the (ADF’s determined) symmetry does not correspond with the DIRAC’s detected full group.

More examples have to be be added here to demostrate (and test) the full variety of symmetry detections in DIRAC.