Nonbonded interactions are the most expensive part of MM calculations. Incorporating such interactions implicitly within the bonded forces (harmonic angle stretching, generally) is more efficient and a reasonable representation of the physics.
Nonbonded interactions are the most expensive part of MM calculations. Incorporating such interactions implicitly within the bonded forces (harmonic angle stretching, generally) is more efficient and a reasonable representation of the physics.
Also, the distances between 1-3 neighbours may be shorter than the sum of the vdW radii (e.g. for two carbons attached to a central sp3-hybridised carbon). Including vdW interactions for this pair would lead to an artificial repulsive vdW contribution, which one would need to compensate somehow.
1) in organic chemistry, using bonded potentials are common due to limited type of bonds that one needs to define and parametrize.
2) in material (non-organic) field people are mostly using non-bonded potentials (pair potentials) as atoms can change their neighbors by jumps (without any chemical reaction).
I suggest you watch first couple of lectures from this course and you will find the answer to you question by an expert:
MIT 3.320 Atomistic Computer Modeling of Materials (available on youtube)
“Why in today's force fields do we not consider …” In the simulations I am familar with we always do. Otherwise molecular conformation would not reproduced correctly.
@Jason: A common force field for chain molecules is the “tangent Lennard-jones chain”, where one assumes a Lennard-Jones pair potential between all sites—inter- and intramolecular—plus rigid bonds or a stiff harmonic potential between adjacent sites. The intramolecular vdW forces make long chain molecules coil up at low pressure. Furthermore, you have to include the intramolecular vdW contributions when you want to calculate the pressure from the virial theorem. This, however, is probably not an issue in biomolecular simulations.