General parameters file¶
The general_db file ($ADFHOME/atomicdata/UFF/general_db) contains all the parameters used to calculate the forces and energies. The format is:
MMAtomType, ri, phi, xi, di, psi, zmm, vsp3, vsp2, chi, nc.
The items in the list are:
- MMAtomType: name, max 5 characters
- ri: Valence Bond [Å]
- phi: Valence Angle [Degree]
- xi: Nonbond Distance [Å]
- di: Nonbond Energy [kcal/mol]
- psi: Nonbond scale [Number]
- zmm: Effective Charge [Charge]
- vsp3: sp3 Torsional Barrier [kcal/mol]
- vsp2: sp2 Torsional Barrier [kcal/mol]
- chi: Electronegativity
- nc: Number of directly attached atoms, aka coordination number. This is required for counting the number of possible dihedrals, and is defined only for the sp2 and sp3 centers (types 2, R, and 3)
The current set of parameters comes from the deMonNano program, and is a combination of published parameters and fitted data to fill in the gaps. The deMonNano documentation says the following about the parameters:
Implementation of the Universal Force Field (UFF) in deMonNano
--------------------------------------------------------------
As far as possible, UFF molecular mechanics forcefield in deMon
follows the published forcefield definition in [1]. In several
cases, the definitions and expressions in [1] are not consistent
with the published applications of the forcefield [1,5,6].
In those cases, an attept was made to correct the errors and omissions,
using information from [2].
The following changed were made, compared to the published UFF
forcefield description (all equation and page numbers refer to [1]).
1. Sign error in Eq. 2 (equilibrium bond length) was corrected
- electronegativity correction must be negative!
2. Equilibrium valence angle for O_3_z was corrected from 146.0
degree to 145.45 degree.
3. Bending periodicity (Eq. 10) for linear coordination was
corrected from 1 to 2.
4. Sign errors were corrected in eqs. 13 and and unnumbered equation
for the beta parameter (between eqs. 13 and 14).
5. The reference value of the UFF amide force constant, of 105.5
kcal/mol/rad**2 (p. 10028) is wrong. The results are consistent
with the force constant of 211.0 kcal/mol/rad**2.
6. Equilibrium torsional angle for a bond between a group-6A atom
(oxygen ...) and an sp2 atom (90 degree) is wrong (p. 10028).
It should be 0 degree.
7. The conditional for the special-case sp2-sp3 torsion (p. 10029)
is wrong, and should be inverted - see [4].
8. The overall shape of the UFF torsional potential degenerates to
a Heavyside function when one of the bond angles becomes linear,
leading to failures in geometry optimization and force constant
evaluation. The UFF torsional term was augmented with a smooth
masking function, to avoid this. See "uff_4centre.f90" for
details.
9. UFF inversion potential is not defined in [1] for group 5A
elements (from phosphorus down). Taking the equilibrium inversion
coordinate of 87 degree, and the suitable expressions for the
cosine weights (see uff_get_inversion_shape in "uff_database.f90")
appears to reproduce published UFF structures and energetics.
The following atom types have been fully tested, and are believed
to reproduce published UFF forcefield results exactly. The examples
refer to the $deMon/examples/test.mm directory.
Atom type Example Description
--------- ------- -----------
Al3 alme3tma Trivalent aluminum
As3+3 asf3 Trivalent arsenic
B_2 bcl3 Planar (sp2) boron
B_3 b2h5nme2 Tetrahedral (sp2) boron, including
charge transfer adducts and borohydrates
Br bbr3 Univalent bromine
C_1 c2h2, co Linear (sp) carbon
C_2 acetone Planar tricoordinated (sp2) carbon
C_3 c2h6 Tetrahedral (sp3) carbon
C_R c4h6 Resonant, variable bond order (sp2) carbon.
Cl socl2 Univalent chlorine
F_ sof2ncl Univalent fluorine
Ge3 geh3ogeh3 Tetrahedral (sp3) germanium
H_ h2o Normal, non-bridging hydrogen
H_b b2h5nme2 Bridging hydrogen, for use in boranes
(NOT SUITABLE FOR H-BONDS!)
I_ bi3 Univalent iodine
N_1 ch3cn Monocoordinated (sp) nitrogen, triple bond
N_2 ch3n2ch3 Dicoordinated (sp2) nitrogen, single-double bond
N_3 ch3nh2 Amine (sp3) nitrogen, three single bonds
N_3+4 b2h5nme2 Charged amine (sp3) nitrogen, four single bonds
(THIS IS NOT A STANDARD UFF TYPE!)
N_R c5h5n Resonant planar (sp2) nitrogen, for use in
aromatics and amides. For amides, use 1.41 bond
order!
O_1 co Special "co" type, one triple bond.
O_2 acetone One-coordinated (sp2) oxygen, one double bond.
O_3 h2o Two-coordinated (sp3) oxygen, two single bonds
O_3_z sih3osih3 Special two-coordinated oxygen, for use in
Si-O bonds
O_R c4h4o Resonant planar (sp2) oxygen, also for use in
nitro groups and such.
P_3+3 ph3 Pyramidal (sp3) phosphorus, three single bonds
P_3+5 p4o7 Tetrahedral hypervalent phosphorus
P_3+q bh3ph3 Dative tetrahedral (sp3) phosphorus, watch for
the bond order!
S_3+2 ch3sch3 Bent two-coordinated sulfur (sp3), two single bonds
S_3+4 socl2 Pyramidal three-coordinated hypervalent sulfur
S_3+6 so2cl2 Tetrahedral four-coordinated hypervalent sulfur
Se3+2 h2se Bent two-coordinated (sp3) selenium
Si3 si4o4h8 Tetrahedral silicon
Additionally, parameter sets for the following atom types are believed
be complete, and may be expected produce results identical to the published
UFF data: Li, Na, K_, Rb, Cs (Note that UFF does not specify atomic charges
- it is your responsibility to assign those, if charges are needed!)
For the remaining atom types, UFF definition [1] relies on an unpublished
set of electronegativities [2]. In deMon, these values were replaced by
Pauling electronegativities, scaled to fit published UFF electronegativities.
This can be expected to produce small deviations in bond lengths and bond
angles, compared to published UFF results.
If you wish to use other parameters, you should copy the general_db file, and rename it. This new file can also be placed outside of $ADFHOME/atomicdata/UFF. To use this new parameter file in UFF, specify the following keyword in your input:
UFF
database /path/to/the/new/file/mygeneral_db
End