Example: Compound Fragments: Ni(CO)4¶
An illustration of the fragment feature of adf
A transition metal complex is built from a Nickel atom and four CO fragments. The outcomes allows for an analysis (of molecular orbitals and the Bonding energy) in terms of the fragment properties. It is a Single Point calculation. Geometry optimization would not have been possible in this set-up because an optimization requires that only single-atom fragments are used.
The three atoms are created first: C, O, and Ni. For Carbon and Oxygen a type-DZ basis set is used (double-zeta) using the Basis key, while Ni gets a type-TZP basis (triple-zeta plus polarization).
The CO molecule, to serve as a fragment template in Ni(CO)4 , is computed from the atomic fragments C and O. The coordinate values (atoms) are in bohr, rather than in Angstrom because the unit-of-length is redefined by the key units with subkey length.
The key scf is used to specify a somewhat tighter convergence criterion than the default, just to illustrate how to do this (normal settings are quite adequate).
The TAPE21 result file is renamed t21.CO.
$ADFBIN/adf <<eor title CO (as fragment for NiCO4) SCF converge 1e-8 end EPRINT SFO eig ovl END units length bohr end atoms C 0 0 0 O 0 0 2.15617844 end Basis Type DZ Core Small End endinput eor mv TAPE21 t21.CO
One needs to include the subkey SFO of the key EPRINT with arguments eig and ovl in order to get the SFO MO coefficients and SFO overlap matrix printed on standard output.
Apart from the title, the input contains comment. This does not specify computational parameters but is only echoed in the output header, similar to the title. Contrary to the title, however, such comments are not preserved, apart from their echo in output and they are not written to TAPE21 or any other result file.
The atomic coordinates (atoms) are given in bohr (Units). To supply the numerical values use is made of user-defined constants (define): xyzC and xyzOx. This is convenient and it prevents typing errors when several coordinate values are identical, in particular when they carry a lot of decimal places.
The Atoms records contain also a specification of the fragments to which the respective atoms belong: four different CO fragments. No fragment is specified for the Ni atom, which implies that it is a fragment on its own.
The numbers at the very left of the records (1 through 9, with (optionally) a period after them), have no relevance. You can set them for ease of reference or counting.
$ADFBIN/adf <<eor title Ni(CO)4, from fragments Ni and CO COMMENT No geometry optimization possible, because not all fragments are single atoms END units length bohr end EPRINT SFO eig ovl END DEFINE xyzC=2.0053211 xyzOx=3.2501913 END atoms 1. Ni 0 0 0 2. C xyzC xyzC xyzC f=CO/1 3. C -xyzC -xyzC xyzC f=CO/2 4. C xyzC -xyzC -xyzC f=CO/3 5. C -xyzC xyzC -xyzC f=CO/4 6. O xyzOx xyzOx xyzOx f=CO/1 7. O -xyzOx -xyzOx xyzOx f=CO/2 8. O xyzOx -xyzOx -xyzOx f=CO/3 9. O -xyzOx xyzOx -xyzOx f=CO/4 end fragments CO t21.CO Ni t21.Ni end endinput eor