forbidden group N_birad_singlet_2singleBonds looks shady

[nickvandewiele]

e791c0a shows a test in which the forbidden group (
https://github.com/ReactionMechanismGenerator/RMG-database/blob/master/input/forbiddenStructures.py#L55)

N_birad_singlet_2singleBonds

1 N u0 p1 {2,S} {3,S}
2 R ux {1,S}
3 R ux {1,S}

is successfully discovered in the "normal" looking molecule [CH2]-NH2

multiplicity 2
1 N u0 p1 c0 {2,S} {5,S} {6,S}
2 C u1 p0 c0 {1,S} {3,S} {4,S}
3 H u0 p0 c0 {2,S}
4 H u0 p0 c0 {2,S}
5 H u0 p0 c0 {1,S}
6 H u0 p0 c0 {1,S}

Although the name of the group suggests the forbidden group should hit singlet biradical nitrogen, I am unsure whether this molecule qualifies as such.

My recommendation would be to remove that forbidden structure from the list, as it would prevent a lot of normal nitrogen containing compounds from being formed.

[nickvandewiele]

@alongd would you agree?

[alongd]

I agree that N_birad_singlet_2singleBonds is problematic, I had to remove this structure (among others) from my forbiddenStructures.py file to be able to run N-chem in RMG.

Still, shouldn't this forbidden structure restrict N p1 with 2 single bonds? The N atom in [CH2]-NH2 has three.

[rwest]

It is difficult (impossible?) to represent absence of a bond in RMG's adjacency list format for functional groups: a group that insists on there being [at least] 2 bonds will match a molecule that has 3 bonds. We have previously resorted to work-arounds like specifying number of unpaired electrons, or whatever, to limit the overall valence; it might be time to think of a syntax for absence of bonds (would help in some cyclic transition state templates too, I think?)

If you're able to share what else you had to do to get N-chem working, please do so :)

[alongd]

I had to remove 5 structures from the forbiddenStructures.py file, only one directly involves nitrogen (the one Nick pointed to).

N_birad_singlet_2singleBonds restricts NH3 as well... I agree we need a no-bond syntax.
Carbene_D_singlet restricts the H2CC species from the ERC-FoundationFuelv0.9 lib.
Carbene_D_triplet restricts the C2O species from the Glarborg/C3 lib.
Carbene_S_singlet restricts the HCOH species from the Nitrogen_Dean_and_Bozzelli lib.
C_quintet restricts the C species from a new library I work with: AramcoMech1.3.

I wonder how come no one else had this problem when using, for example, ERC-FoundationFuelv0.9 or Glarborg/C3 in RMG.

[alongd]

Sorry, one more which IS N-related:

N_monorad_3singleBonds restricts the H2NO species from the Nitrogen_Glarborg_Gimenez_et_al lib.

It is indeed an N-radical with 3 single bonds, yet should be allowed:
multiplicity 2
1 N u1 p0 c+1 {2,S} {3,S} {4,S}
2 H u0 p0 c0 {1,S}
3 H u0 p0 c0 {1,S}
4 O u0 p3 c-1 {1,S}

[alongd]

See RMG-database PR 120 (ReactionMechanismGenerator/RMG-database#120)