lammps/doc/pair_hybrid.html

294 lines
13 KiB
HTML
Raw Normal View History

<HTML>
<CENTER><A HREF = "http://lammps.sandia.gov">LAMMPS WWW Site</A> - <A HREF = "Manual.html">LAMMPS Documentation</A> - <A HREF = "Section_commands.html#comm">LAMMPS Commands</A>
</CENTER>
<HR>
<H3>pair_style hybrid command
</H3>
<H3>pair_style hybrid/omp command
</H3>
<H3>pair_style hybrid/overlay command
</H3>
<H3>pair_style hybrid/overlay/omp command
</H3>
<P><B>Syntax:</B>
</P>
<PRE>pair_style hybrid style1 args style2 args ...
pair_style hybrid/overlay style1 args style2 args ...
</PRE>
<UL><LI>style1,style2 = list of one or more pair styles and their arguments
</UL>
<P><B>Examples:</B>
</P>
<PRE>pair_style hybrid lj/cut/coul/cut 10.0 eam lj/cut 5.0
pair_coeff 1*2 1*2 eam niu3
pair_coeff 3 3 lj/cut/coul/cut 1.0 1.0
pair_coeff 1*2 3 lj/cut 0.5 1.2
</PRE>
<PRE>pair_style hybrid/overlay lj/cut 2.5 coul/long 2.0
pair_coeff * * lj/cut 1.0 1.0
pair_coeff * * coul/long
</PRE>
<P><B>Description:</B>
</P>
<P>The <I>hybrid</I> and <I>hybrid/overlay</I> styles enable the use of multiple
pair styles in one simulation. With the <I>hybrid</I> style, exactly one
pair style is assigned to each pair of atom types. With the
<I>hybrid/overlay</I> style, one or more pair styles can be assigned to
each pair of atom types. The assignment of pair styles to type pairs
is made via the <A HREF = "pair_coeff.html">pair_coeff</A> command.
</P>
<P>Here are two examples of hybrid simulations. The <I>hybrid</I> style could
be used for a simulation of a metal droplet on a LJ surface. The
metal atoms interact with each other via an <I>eam</I> potential, the
surface atoms interact with each other via a <I>lj/cut</I> potential, and
the metal/surface interaction is also computed via a <I>lj/cut</I>
potential. The <I>hybrid/overlay</I> style could be used as in the 2nd
example above, where multiple potentials are superposed in an additive
fashion to compute the interaction between atoms. In this example,
using <I>lj/cut</I> and <I>coul/long</I> together gives the same result as if
the <I>lj/cut/coul/long</I> potential were used by itself. In this case,
it would be more efficient to use the single combined potential, but
in general any combination of pair potentials can be used together in
to produce an interaction that is not encoded in any single pair_style
file, e.g. adding Coulombic forces between granular particles.
</P>
<P>All pair styles that will be used are listed as "sub-styles" following
the <I>hybrid</I> or <I>hybrid/overlay</I> keyword, in any order. Each
sub-style's name is followed by its usual arguments, as illustrated in
the example above. See the doc pages of individual pair styles for a
listing and explanation of the appropriate arguments.
</P>
<P>In the pair_coeff commands, the name of a pair style must be added
after the I,J type specification, with the remaining coefficients
being those appropriate to that style. For example, consider a
simulation with 3 atom types: types 1 and 2 are Ni atoms, type 3 are
LJ atoms with charges. The following commands would set up a hybrid
simulation:
</P>
<PRE>pair_style hybrid eam/alloy lj/cut/coul/cut 10.0 lj/cut 8.0
pair_coeff * * eam/alloy nialhjea Ni Ni NULL
pair_coeff 3 3 lj/cut/coul/cut 1.0 1.0
pair_coeff 1*2 3 lj/cut 0.8 1.3
</PRE>
<P>If pair coefficients are specified in the data file read via the
<A HREF = "read_data.html">read_data</A> command, then the same rule applies.
E.g. "eam/alloy" or "lj/cut" must be added after the atom type, for
each line in the "Pair Coeffs" section, e.g.
</P>
<PRE>Pair Coeffs
</PRE>
<PRE>1 lj/cut/coul/cut 1.0 1.0
...
</PRE>
<P>Note that the pair_coeff command for some potentials such as
<A HREF = "pair_eam.html">pair_style eam/alloy</A> includes a mapping specification
of elements to all atom types, which in the hybrid case, can include
atom types not assigned to the <I>eam/alloy</I> potential. The NULL
keyword is used by many such potentials (eam/alloy, Tersoff, AIREBO,
etc), to denote an atom type that will be assigned to a different
sub-style.
</P>
<P>For the <I>hybrid</I> style, each atom type pair I,J is assigned to exactly
one sub-style. Just as with a simulation using a single pair style,
if you specify the same atom type pair in a second pair_coeff command,
the previous assignment will be overwritten.
</P>
<P>For the <I>hybrid/overlay</I> style, each atom type pair I,J can be
assigned to one or more sub-styles. If you specify the same atom type
pair in a second pair_coeff command with a new sub-style, then the
second sub-style is added to the list of potentials that will be
calculated for two interacting atoms of those types. If you specify
the same atom type pair in a second pair_coeff command with a
sub-style that has already been defined for that pair of atoms, then
the new pair coefficients simply override the previous ones, as in the
normal usage of the pair_coeff command. E.g. these two sets of
commands are the same:
</P>
<PRE>pair_style lj/cut 2.5
pair_coeff * * 1.0 1.0
pair_coeff 2 2 1.5 0.8
</PRE>
<PRE>pair_style hybrid/overlay lj/cut 2.5
pair_coeff * * lj/cut 1.0 1.0
pair_coeff 2 2 lj/cut 1.5 0.8
</PRE>
<P>Coefficients must be defined for each pair of atoms types via the
<A HREF = "pair_coeff.html">pair_coeff</A> command as described above, or in the
data file or restart files read by the <A HREF = "read_data.html">read_data</A> or
<A HREF = "read_restart.html">read_restart</A> commands, or by mixing as described
below.
</P>
<P>For both the <I>hybrid</I> and <I>hybrid/overlay</I> styles, every atom type
pair I,J (where I <= J) must be assigned to at least one sub-style via
the <A HREF = "pair_coeff.html">pair_coeff</A> command as in the examples above, or
in the data file read by the <A HREF = "read_data.html">read_data</A>, or by mixing
as described below.
</P>
<P>If you want there to be no interactions between a particular pair of
atom types, you have 3 choices. You can assign the type pair to some
sub-style and use the <A HREF = "neigh_modify.html">neigh_modify exclude type</A>
command. You can assign it to some sub-style and set the coefficients
so that there is effectively no interaction (e.g. epsilon = 0.0 in a
LJ potential). Or, for <I>hybrid</I> and <I>hybrid/overlay</I> simulations, you
can use this form of the pair_coeff command in your input script:
</P>
<PRE>pair_coeff 2 3 none
</PRE>
<P>or this form in the "Pair Coeffs" section of the data file:
</P>
<PRE>3 none
</PRE>
<P>If an assignment to <I>none</I> is made in a simulation with the
<I>hybrid/overlay</I> pair style, it wipes out all previous assignments of
that atom type pair to sub-styles.
</P>
<P>Note that you may need to use an <A HREF = "atom_style.html">atom_style</A> hybrid
command in your input script, if atoms in the simulation will need
attributes from several atom styles, due to using multiple pair
potentials.
</P>
<HR>
<P>The potential energy contribution to the overall system due to an
individual sub-style can be accessed and output via the <A HREF = "compute_pair.html">compute
pair</A> command.
</P>
<HR>
<P>IMPORTANT: Several of the potentials defined via the pair_style
command in LAMMPS are really many-body potentials, such as Tersoff,
AIREBO, MEAM, ReaxFF, etc. The way to think about using these
potentials in a hybrid setting is as follows.
</P>
<P>A subset of atom types is assigned to the many-body potential with a
single <A HREF = "pair_coeff.html">pair_coeff</A> command, using "* *" to include
all types and the NULL keywords described above to exclude specific
types not assigned to that potential. If types 1,3,4 were assigned in
that way (but not type 2), this means that all many-body interactions
between all atoms of types 1,3,4 will be computed by that potential.
Pair_style hybrid allows interactions between type pairs 2-2, 1-2,
2-3, 2-4 to be specified for computation by other pair styles. You
could even add a second interaction for 1-1 to be computed by another
pair style, assuming pair_style hybrid/overlay is used.
</P>
<P>But you should not, as a general rule, attempt to exclude the
many-body interactions for some subset of the type pairs within the
set of 1,3,4 interactions, e.g. exclude 1-1 or 1-3 interactions. That
is not conceptually well-defined for many-body interactions, since the
potential will typically calculate energies and foces for small groups
of atoms, e.g. 3 or 4 atoms, using the neighbor lists of the atoms to
find the additional atoms in the group. It is typically non-physical
to think of excluding an interaction between a particular pair of
atoms when the potential computes 3-body or 4-body interactions.
</P>
<P>However, you can still use the pair_coeff none setting or the
<A HREF = "neigh_modify.html">neigh_modify exclude</A> command to exclude certain
type pairs from the neighbor list that will be passed to a manybody
sub-style. This will alter the calculations made by a many-body
potential, since it builds its list of 3-body, 4-body, etc
interactions from the pair list. You will need to think carefully as
to whether it produces a physically meaningful result for your model.
</P>
<P>For example, imagine you have two atom types in your model, type 1 for
atoms in one surface, and type 2 for atoms in the other, and you wish
to use a Tersoff potential to compute interactions within each
surface, but not between surfaces. Then either of these two command
sequences would implement that model:
</P>
<PRE>pair_style hybrid tersoff
pair_coeff * * tersoff SiC.tersoff C C
pair_coeff 1 2 none
</PRE>
<PRE>pair_style tersoff
pair_coeff * * SiC.tersoff C C
neigh_modify exclude type 1 2
</PRE>
<P>Either way, only neighbor lists with 1-1 or 2-2 interactions would be
passed to the Tersoff potential, which means it would compute no
3-body interactions containing both type 1 and 2 atoms.
</P>
<HR>
<P>Styles with a <I>cuda</I>, <I>gpu</I>, <I>omp</I>, or <I>opt</I> suffix are functionally
the same as the corresponding style without the suffix. They have
been optimized to run faster, depending on your available hardware,
as discussed in <A HREF = "Section_accelerate.html">this section</A> of the manual.
</P>
<P>Since the <I>hybrid</I> and <I>hybrid/overlay</I> styles delegate computation
to the individual sub-styles, the suffix versions of the <I>hybrid</I>
and <I>hybrid/overlay</I> styles are used to propagate the corresponding
suffix to all sub-styles, if those versions exist. Otherwise the
non-accelerated version will be used.
</P>
<P>The individual accelerated sub-styles are part of the USER-CUDA, GPU,
USER-OMP and OPT packages, respectively. They are only enabled if
LAMMPS was built with those packages. See the
<A HREF = "Section_start.html#start_3">Making LAMMPS</A> section for more info.
</P>
<P>You can specify the accelerated styles explicitly in your input script
by including their suffix, or you can use the <A HREF = "Section_start.html#start_6">-suffix command-line
switch</A> when you invoke LAMMPS, or you can
use the <A HREF = "suffix.html">suffix</A> command in your input script.
</P>
<P>See <A HREF = "Section_accelerate.html">this section</A> of the manual for more
instructions on how to use the accelerated styles effectively.
</P>
<HR>
<P><B>Mixing, shift, table, tail correction, restart, rRESPA info</B>:
</P>
<P>Any pair potential settings made via the
<A HREF = "pair_modify.html">pair_modify</A> command are passed along to all
sub-styles of the hybrid potential.
</P>
<P>For atom type pairs I,J and I != J, if the sub-style assigned to I,I
and J,J is the same, and if the sub-style allows for mixing, then the
coefficients for I,J can be mixed. This means you do not have to
specify a pair_coeff command for I,J since the I,J type pair will be
assigned automatically to the I,I sub-style and its coefficients
generated by the mixing rule used by that sub-style. For the
<I>hybrid/overlay</I> style, there is an additional requirement that both
the I,I and J,J pairs are assigned to a single sub-style. See the
"pair_modify" command for details of mixing rules. See the See the
doc page for the sub-style to see if allows for mixing.
</P>
<P>The hybrid pair styles supports the <A HREF = "pair_modify.html">pair_modify</A>
shift, table, and tail options for an I,J pair interaction, if the
associated sub-style supports it.
</P>
<P>For the hybrid pair styles, the list of sub-styles and their
respective settings are written to <A HREF = "restart.html">binary restart
files</A>, so a <A HREF = "pair_style.html">pair_style</A> command does
not need to specified in an input script that reads a restart file.
However, the coefficient information is not stored in the restart
file. Thus, pair_coeff commands need to be re-specified in the
restart input script.
</P>
<P>These pair styles support the use of the <I>inner</I>, <I>middle</I>, and
<I>outer</I> keywords of the <A HREF = "run_style.html">run_style respa</A> command, if
their sub-styles do.
</P>
<P><B>Restrictions:</B>
</P>
<P>When using a long-range Coulombic solver (via the
<A HREF = "kspace_style.html">kspace_style</A> command) with a hybrid pair_style,
one or more sub-styles will be of the "long" variety,
e.g. <I>lj/cut/coul/long</I> or <I>buck/coul/long</I>. You must insure that the
short-range Coulombic cutoff used by each of these long pair styles is
the same or else LAMMPS will generate an error.
</P>
<P><B>Related commands:</B>
</P>
<P><A HREF = "pair_coeff.html">pair_coeff</A>
</P>
<P><B>Default:</B> none
</P>
</HTML>