search

cfgoldsmith / RMG-Py

Python version of the amazing Reaction Mechanism Generator (RMG).
http://reactionmechanismgenerator.github.io/RMG-Py/
Other
1 stars 4 forks source link

New Eley-Rideal families for RMG-database #63

Open mazeau opened 5 years ago

mazeau commented 5 years ago

I can't figure out how to open up an issue on the RMG-database so I just put it here instead

Now that Eley-Rideal is a little added, we came across some other families to potentially add to help us deal with bidentate/multidentate species.

Have just added:

1 EleyRideal_H_addition_multiple_bond

2   3=4        3-4-2
|        <=>   |      
X              X

where 4=3 can be a double, triple, or quadruple bond, and 2 is an H atom.

Some potential families to add are:

2

1  + 2     1-2
||     <=> | 
X          X

where 2 is a radical, probably an H.

Would be good to have

3

1=2   + 3      3-1-2
|          <=>   | |
X X              X X

where 3 is a radical, probably a hydrogen or something?

Probably less likely than number 2, above. If you had that starting point, it would probably just go like this instead:

1=2            1-2
|        <=>  || |
X X            X X

4

                 4
                 |
1-2=3  + 4     1-2-3
|          <=> |   |
X   X          X   X

where 2=3 can be a double or triple bond. 4 is probably an H

Probably a similar story to number 3. It's overall 3rd order (two surface sites and a gas phase species)

5

                 4
                 |
1-2=3  + 4     1-2-3
|          <=> | |
X X            X X

where 2=3 can be a double or triple bond. 4 is probably an H

Probably a similar story to number 3. It's overall 3rd order (two surface sites and a gas phase species)

6

2
|
1 + 3• <=>  1 + 3-2
|          ||
X           X

3 is a radical, abstracts 2. eg. 1 is CH2, 3 is CH3, 2 is H, you get methane and =CH2

Seems reasonable. Not unlike H abstraction in gas phase.

Would be good to have

7

2
||
1 + 3• <=>  1  + 3-2•
|          |||
X           X

Not going to work. because you have to break a double bond (1=2), and 2 ends up unhappy.

cfgoldsmith commented 5 years ago

Great! My thoughts are: 1st rxn, inserting 3-4 between 1-2: probably going to have a high barrier? It should happen, but probably not my first choice. 2cd rxn, dangling double falls flat: very important! I would do this one first. It touches on the other thread we have on pi- vs di-sigma. 3rd + 4th rxns: I think I'm missing something here, but it looks like it won't satisfy the octet rule for 2 in rxn3 or 3 in rxn4? Let's say that 1-2=3 are allyl, so that we have surface propene (X-CH2-CH=CH2). in rxn3, we have X-CH2-CH-CH2-X, so C2 is short one bond. Ditto C3 in the other case. fifth rxn: if 2 is H, then yes, could be important. If 2 is some other functional group, probably less so. I would probably make this my number 2 choice. However, I would make one big difference. Radicals on surfaces aren't really much of a thing, definitely not for the bonding atom. I would remove the radical from X-1* in the product and instead increase the bond order, X=1. 6th rxn: seems less likely. I guess the idea is to have addition/isomerization/elimination, and not a single abstract step? probably hard to find decent data to parameterize that one.

mazeau commented 5 years ago

for the dangling double falls flat, I don't think that follows the octet rule either. I think it would have to be something like this:

1=2            3-1-2
|      + 3 <=>   | |
X X              X X

where 3 would be a hydrogen or something?

rwest commented 5 years ago

I edited your original post above, to add some numbers, as I was getting confused by the counting. Now maybe the other comments need editing so they correspond . Also, reading Franklin's first comment on reaction 1 had me (or him) confused if 1 is itself an X (as I now drew) or was adsorbed to an X.

cfgoldsmith commented 5 years ago

Thanks, that helps. For Reaction 2, I was thinking of two different cases:

1=2                1-2
|    +        <=> || |
X    X             X X

which, with the bond order increase, does preserve the octet rule. Then, additionally, the pi- to -di-sigma would be:

(1=2)            1-2
  :    +  <=>    | |
  X    X         X X

where (1=2) indicates a vdW structure

mazeau commented 5 years ago

I don't like those as much. My original meaning behind that one was to get benzene fully adsorbed to the surface and then hydrogenated and desorbed.

rwest commented 5 years ago

Look for things by one of Norskov's students. An oral talk at AIChE a year or two ago. Probably published by now.