@@ -72,7 +72,14 @@ def modify(
7272
7373 k_soft : float, optional
7474 The force constant to use when setting angle terms involving ghost atoms
75- for non-planar triple junctions. (In kcal/mol/rad^2)
75+ for non-planar triple junctions. Should be small relative to the
76+ physical angle force constants of the force field in use, so as not to
77+ distort the physical end-state geometry, while still providing enough
78+ anchoring to prevent flapping. Boresch et al. (JCTC 2021) used 3.55
79+ kcal/mol/rad^2, derived from CGenFF ammonia H-N-H parameters. For
80+ GAFF (~41 kcal/mol/rad^2) and OpenFF (~54 kcal/mol/rad^2), the default
81+ of 5 kcal/mol/rad^2 is similarly small relative to the physical force
82+ constants and is well justified. (In kcal/mol/rad^2)
7683
7784 optimise_angles : bool, optional
7885 Whether to optimise the equilibrium value of the angle terms involving
@@ -1167,7 +1174,10 @@ def _triple(
11671174
11681175 k_soft : float, optional
11691176 The force constant to use when setting angle terms involving ghost
1170- atoms for non-planar triple junctions. (In kcal/mol/rad^2)
1177+ atoms for non-planar triple junctions. Should be small relative to
1178+ the physical angle force constants of the force field in use. For
1179+ GAFF (~41 kcal/mol/rad^2) and OpenFF (~54 kcal/mol/rad^2), the
1180+ default of 5 kcal/mol/rad^2 is appropriate. (In kcal/mol/rad^2)
11711181
11721182 optimise_angles : bool, optional
11731183 Whether to optimise the equilibrium value of the angle terms involving
@@ -1636,7 +1646,10 @@ def _higher(
16361646
16371647 k_soft : float, optional
16381648 The force constant to use when setting angle terms involving ghost
1639- atoms for non-planar triple junctions. (In kcal/mol/rad^2)
1649+ atoms for non-planar triple junctions. Should be small relative to
1650+ the physical angle force constants of the force field in use. For
1651+ GAFF (~41 kcal/mol/rad^2) and OpenFF (~54 kcal/mol/rad^2), the
1652+ default of 5 kcal/mol/rad^2 is appropriate. (In kcal/mol/rad^2)
16401653
16411654 optimise_angles : bool, optional
16421655 Whether to optimise the equilibrium value of the angle terms involving
0 commit comments