|
|
potentype (integer)
The different settings for potentype require a different set of variables afterwards. For each
potentype I list a description of the nonbonded potential and the set of variables that must be
specified below the potentype. This documentation is only kept up to date for the most current
release version. Last updated for version 4.0.0.
- potentype = 0: 12-6 Lennard-Jones potential.
- If the two atoms are separated by more than 3 bonds, or are on different molecules then
- Unonbond = 4 * nbcoeff(2) * [ (nbcoeff(1)/r)12 - (nbcoeff(1)/r)6 ]
- else if the two atoms are separated by exactly 3 bonds then
- Unonbond = 4 * nbcoeff(4) * [ (nbcoeff(3)/r)12 - (nbcoeff(3)/r)6 ]
- mixrule (integer)
- mixrule = 0: Lorentz-Berthelot (arithmetic mean of sigma, geometric
mean of epsilon) mixing rules.
- mixrule = 1: Geometric (geometric mean of sigma and epsilon)
mixing rules.
- mixrule = 3: Gromos (geometric mean of sigma and epsilon with
some special cases) mixing rules.
- mixrule = 4: Explicit (defined in towhee_ff files) mixing rules.
- lshift (logical)
- .true. if you want the nonbonded potential to be
shifted so that it is zero at the cutoff.
- .false. if you do not want to shift the nonbonded potential.
- ltailc (logical)
- .true. if you want to apply analytical tail corrections for the
portion of the potential that is past the cutoff.
Note that you cannot have a shifted potential and tail corrections
at the same time.
- .false. if you do not want analytical tail corrections.
- rmin (double precision)
- A hard inner cutoff that can speed computation for Lennard-Jones
systems, and is required to avoid the potential hitting infinity
for exponential repulsion systems which also contain point charges.
This should be set smaller than the smallest radius of any atom.
Generally I set this to 0.5 or 1.0 Angstroms.
- rcut (double precision)
- The potential cutoff in Angstroms.
- rcutin (double precision)
- The inner nonbonded cutoff used in configurational-bias Monte
Carlo moves. This dual-cutoff method can speed configurational-bias
computations by at least a factor of 2, without affecting the acceptance
rate. The inner cutoff is used during the growth procedure, and
the full potential is calculated at the end of the move and everything
is fixed up in the acceptance criteria. I typically set this to
5 Angstroms for noncoulombic simulations, and to 10 Angstroms for
coulombic simulations.
- potentype = 1: 9-6 Lennard-Jones potential
- If the two atoms are separated 3 or more bonds, or are on different molecules then
- Unonbond = nbcoeff(2) * [ 2*(nbcoeff(1)/r)9 - 3*(nbcoeff(1)/r)6 ]
- mixrule (integer)
- mixrule = 2: Compass (sixth order combination of sigma and epsilon) mixing rules.
- lshift (logical)
- .true. if you want the nonbonded potential to be
shifted so that it is zero at the cutoff.
- .false. if you do not want to shift the nonbonded potential.
- ltailc (logical)
- .true. if you want to apply analytical tail corrections for the
portion of the potential that is past the cutoff.
Note that you cannot have a shifted potential and tail corrections
at the same time.
- .false. if you do not want analytical tail corrections.
- rmin (double precision)
- A hard inner cutoff that can speed computation for Lennard-Jones
systems, and is required to avoid the potential hitting infinity
for exponential repulsion systems which also contain point charges.
This should be set smaller than the smallest radius of any atom.
Generally I set this to 0.5 or 1.0 Angstroms.
- rcut (double precision)
- The potential cutoff in Angstroms.
- rcutin (double precision)
- The inner nonbonded cutoff used in configurational-bias Monte
Carlo moves. This dual-cutoff method can speed configurational-bias
computations by at least a factor of 2, without affecting the acceptance
rate. The inner cutoff is used during the growth procedure, and
the full potential is calculated at the end of the move and everything
is fixed up in the acceptance criteria. I typically set this to
5 Angstroms for noncoulombic simulations, and to 10 Angstroms for
coulombic simulations.
- potentype = 2: Exponential-6 potential
- If the two atoms are separated by more than 3 bonds, or are on different molecules then
- Unonbond = nbcoeff(1)/r^6 + nbcoeff(2) * exp[nbcoeff(3)*r]
- mixrule (integer)
- mixrule = 4: Explicit (defined in towhee_ff files) mixing rules.
- lshift (logical)
- .true. if you want the nonbonded potential to be
shifted so that it is zero at the cutoff.
- .false. if you do not want to shift the nonbonded potential.
- ltailc (logical)
- .true. if you want to apply analytical tail corrections for the
portion of the potential that is past the cutoff.
Note that you cannot have a shifted potential and tail corrections
at the same time.
- .false. if you do not want analytical tail corrections.
- rcut (double precision)
- The potential cutoff in Angstroms.
- rcutin (double precision)
- The inner nonbonded cutoff used in configurational-bias Monte
Carlo moves. This dual-cutoff method can speed configurational-bias
computations by at least a factor of 2, without affecting the acceptance
rate. The inner cutoff is used during the growth procedure, and
the full potential is calculated at the end of the move and everything
is fixed up in the acceptance criteria. I typically set this to
5 Angstroms for noncoulombic simulations, and to 10 Angstroms for
coulombic simulations.
- potentype = 3: Hard Sphere potential
- If the two atoms are separated by more than 3 bonds, or are on different molecules then
- Unonbond = Infinity if r <= nbcoeff(1), or 0 otherwise
- mixrule (integer)
- mixrule = 5: Hard sphere (arithmetic mean of sigmas) mixing rules.
- rcutin (double precision)
- The inner nonbonded cutoff used in configurational-bias Monte
Carlo moves. This dual-cutoff method can speed configurational-bias
computations by at least a factor of 2, without affecting the acceptance
rate. The inner cutoff is used during the growth procedure, and
the full potential is calculated at the end of the move and everything
is fixed up in the acceptance criteria. If you are using the Hard Sphere potential
without coulombic interations then just set this to something large (like 100), if you are
using coulombic interactions then I would suggest a value of 5 sigma.
- potentype = -3: Repulsive Sphere potential. Added to towhee in version 1.4.6. This
is used to help setup and equilibrate a hard sphere system where it is sometimes challenging to create an
initial conformation with no overlaps. Use the -3 option to equilibrate until the nonbonded potential energy
is 0.0, and then switch back to the normal hard sphere potential.
- If the two atoms are separated by more than 3 bonds, or are on different molecules then
- Unonbond = 1d5 + 1d5 * (nbcoeff(1)^2 - r^2) if r <= nbcoeff(1), or 0 otherwise
- mixrule (integer)
- mixrule = 5: Hard sphere (arithmetic mean of sigmas) mixing rules.
- rcutin (double precision)
- The inner nonbonded cutoff used in configurational-bias Monte
Carlo moves. This dual-cutoff method can speed configurational-bias
computations by at least a factor of 2, without affecting the acceptance
rate. The inner cutoff is used during the growth procedure, and
the full potential is calculated at the end of the move and everything
is fixed up in the acceptance criteria. If you are using the Hard Sphere potential
without coulombic interations then just set this to something large (like 100), if you are
using coulombic interactions then I would suggest a value of 5 sigma.
- potentype = 4: Exponential plus 12-6 Lennard-Jones potential
- If the two atoms are separated by more than 3 bonds, or are on different molecules then
- Unonbond = nbcoeff(1)/r^6 + nbcoeff(2)/r^12 + nbcoeff(3)*exp[nbcoeff(4)*r]
- mixrule (integer)
- mixrule = 4: Explicit (defined in towhee_ff files) mixing rules.
- lshift (logical)
- .true. if you want the nonbonded potential to be
shifted so that it is zero at the cutoff.
- .false. if you do not want to shift the nonbonded potential.
- ltailc (logical)
- .true. if you want to apply analytical tail corrections for the
portion of the potential that is past the cutoff.
Note that you cannot have a shifted potential and tail corrections
at the same time.
- .false. if you do not want analytical tail corrections.
- rmin (double precision)
- A hard inner cutoff that can speed computation for Lennard-Jones
systems, and is required to avoid the potential hitting infinity
for exponential repulsion systems which also contain point charges.
This should be set smaller than the smallest radius of any atom.
Generally I set this to 0.5 or 1.0 Angstroms.
- rcut (double precision)
- The potential cutoff in Angstroms.
- rcutin (double precision)
- The inner nonbonded cutoff used in configurational-bias Monte
Carlo moves. This dual-cutoff method can speed configurational-bias
computations by at least a factor of 2, without affecting the acceptance
rate. The inner cutoff is used during the growth procedure, and
the full potential is calculated at the end of the move and everything
is fixed up in the acceptance criteria. I typically set this to
5 Angstroms for noncoulombic simulations, and to 10 Angstroms for
coulombic simulations.
- potentype = 5: Stillinger-Weber potential
(see Stillinger and Weber 1985)
- This is an atomic potential and can only be used with monatomic molecules in Towhee
- U = nbcoeff(1)*[nbcoeff(2)*Sum u2(rij) + nbcoeff(7)*Sum u3(rij,rjk)]
- u2(rij) = [nbcoeff(3)*{rij/nbcoeff(4)}-nbcoeff(5) - 1]
* exp{1/[(rij/nbcoeff(4) - nbcoeff(6))]}
* Heaviside(nbcoeff(6) - [rij/nbcoeff(4)])
- u3(rij,rjk) =
exp[ nbcoeff(8)/{rij/nbcoeff(4) - nbcoeff(6)} + nbcoeff(8)/{rjk/nbcoeff(4) - nbcoeff(6)}]
* (cos(thetaijk)-nbcoeff(9))^2 * Heaviside(nbcoeff(6)
- rij/nbcoeff(4)) * Heaviside(nbcoeff(6) - rjk/nbcoeff(4))
- mixrule (integer)
- mixrule = 4: Explicit (defined in towhee_ff files) mixing rules.
- potentype = 6: Embedded Atom Method
(see Daw and Baskes 1983)
- This is an atomic potential and can only be used
with monatomic molecules in Towhee. Historically, the
Embedded Atom Method (EAM) used a lookup table for
computing intermolecular interactions and this is an
option in Towhee. When tabular data is used the
interpolatestyle determines the method for
interpolating between the specified values. Other
functional forms are allowed as described in the
Towhee Force Field Documentation.
EAM is a short-ranged potential
that captures the many-body effects by computing a local
density about each atom. This so-called density is
actually a distance dependent function. The sum of the
local densities is then fed into the embedding function to
yield the embedding energy. Additionally, there is a
pair potential term.
- interpolatestyle
(character*20)
- 'cubicspline': Uses a cubic spline to interpolate between the tabulated force field data points
provided in the force field files.
- 'linear': Linear interpolation between the data points of the tabulated force field
- rcut (double precision)
- The distance (in Angstroms) beyond which the density and pair potential are set to zero.
Please note that this will be automatically adjusted up to the maximum entry in the case of a EAM potential
using tabular potential forms.
- potentype = 7: 12-6 Lennard-Jones with implicit solvent
- This uses the standard Lennard-Jones 12-6 plus an extra implicit solvation term. See
Lazaridis and Karplus 1999 for more information.
- mixrule (integer)
- mixrule = 1: Geometric (geometric mean of the nbcoeffs) mixing rules.
- mixrule = 4: Explicit (defined in towhee_ff files) mixing rules.
- lshift (logical)
- .true. if you want the nonbonded potential to be
shifted so that it is zero at the cutoff.
- .false. if you do not want to shift the nonbonded potential.
- ltailc (logical)
- .true. if you want to apply analytical tail corrections for the
portion of the potential that is past the cutoff.
Note that you cannot have a shifted potential and tail corrections
at the same time.
- .false. if you do not want analytical tail corrections.
- rmin (double precision)
- A hard inner cutoff that can speed computation for Lennard-Jones
systems, and is required to avoid the potential hitting infinity
for exponential repulsion systems which also contain point charges.
This should be set smaller than the smallest radius of any atom.
Generally I set this to 0.5 or 1.0 Angstroms.
- rcut (double precision)
- The potential cutoff in Angstroms.
- rcutin (double precision)
- The inner nonbonded cutoff used in configurational-bias Monte
Carlo moves. This dual-cutoff method can speed configurational-bias
computations by at least a factor of 2, without affecting the acceptance
rate. The inner cutoff is used during the growth procedure, and
the full potential is calculated at the end of the move and everything
is fixed up in the acceptance criteria. I typically set this to
5 Angstroms for noncoulombic simulations, and to 10 Angstroms for
coulombic simulations.
- potentype = 8: 12-9-6 potential
- This is a variant of the Lennard-Jones family of potentials. It was implemented for force fields
that use a combination of LJ 12-6 and LJ 9-6.
- v(r) = nbcoeff(1) * (1/r)12 + nbcoeff(2) * (1/r)9 + nbcoeff(3) * (1/r)6
- mixrule (integer)
- mixrule = 1: Geometric (geometric mean of the nbcoeffs) mixing rules.
- mixrule = 4: Explicit (defined in towhee_ff files) mixing rules.
- lshift (logical)
- .true. if you want the nonbonded potential to be
shifted so that it is zero at the cutoff.
- .false. if you do not want to shift the nonbonded potential.
- ltailc (logical)
- .true. if you want to apply analytical tail corrections for the
portion of the potential that is past the cutoff.
Note that you cannot have a shifted potential and tail corrections
at the same time.
- .false. if you do not want analytical tail corrections.
- rmin (double precision)
- A hard inner cutoff that can speed computation for Lennard-Jones
systems, and is required to avoid the potential hitting infinity
for exponential repulsion systems which also contain point charges.
This should be set smaller than the smallest radius of any atom.
Generally I set this to 0.5 or 1.0 Angstroms.
- rcut (double precision)
- The potential cutoff in Angstroms.
- rcutin (double precision)
- The inner nonbonded cutoff used in configurational-bias Monte
Carlo moves. This dual-cutoff method can speed configurational-bias
computations by at least a factor of 2, without affecting the acceptance
rate. The inner cutoff is used during the growth procedure, and
the full potential is calculated at the end of the move and everything
is fixed up in the acceptance criteria. I typically set this to
5 Angstroms for noncoulombic simulations, and to 10 Angstroms for
coulombic simulations.
- potentype = 9: Square Well potential
- Unonbond = Infinity if r <= nbcoeff(1)
- Unonbond = -nbcoeff(3) if nbcoeff(1) < r <= nbcoeff(2)
- Unonbond = 0 if nbcoeff(2) < r
- mixrule (integer)
- mixrule = 4: Explicit (defined in towhee_ff files) mixing rules.
- mixrule = 7: Square Well mixing rules. Arithmetic mean
for nbcoeff(1) and nbcoeff(2). Geometric mean for
nbcoeff(3)
- rcutin (double precision)
- The inner nonbonded cutoff used in configurational-bias Monte
Carlo moves. This dual-cutoff method can speed configurational-bias
computations by at least a factor of 2, without affecting the acceptance
rate. The inner cutoff is used during the growth procedure, and
the full potential is calculated at the end of the move and everything
is fixed up in the acceptance criteria. If you are using the
Square Well potential without coulombic interations then
this isn't going to help you at all as the potential is
already very short ranged so just set this to something larger
than the largest nbcoeff(2),
if you are
using coulombic interactions then I would suggest a value of
around 5 times the largest nbcoeff(1).
- potentype = 10: Tabulated Pair Potential
- This potential uses a table to describe the interactions between atoms.
The interpolatestyle determines the method for interpolating between the specified values.
pair potential term. Cross terms between unlike atoms are described explicitly. The potential is
listed with the distance and corresponding energy on each line.
- interpolatestyle (character*20)
- 'cubicspline': Uses a cubic spline to interpolate between the tabulated force field data points
provided in the force field files.
- 'linear': Linear interpolation between the data points of the tabulated force field
Return to the main towhee web page
|