Adding in new surface families #385
Conversation
mazeau
force-pushed
the
new-families-rebase
branch
from
14dea2d to
b5ce288
Compare
mazeau
force-pushed
the
new-families-rebase
branch
2 times, most recently
from
e2bed43 to
3f6bde3
Compare
|
I tried this branch for methane catalytic combustion, running with the RMG-Py official master. Five minutes into the run it throws an error as RMG tries to add a CH3 radical adsorbate to the core: |
|
Hmm interesting. I didn't get that while running for methanol synthesis, but I also specified maximum radical electrons to be 2, so maybe some crazy species are being proposed? It also doesn't seem like that PR is going to be merged in anytime soon, so I'm not sure what the next steps should be. |
|
I have for example just a radical oxygen adsorbate with a single bond to the surface, and a hydroxyl radical with a vdW bond to the surface in the edge... like it is considering unsensible bond orders to the surface. I wonder why I get that since you didn't get it for methanol synthesis. You were also running on rmg-py master? |
|
input.py.txt ...sorry github is being weird for me lol jk I'm also getting a crap ton of surface radicals |
mazeau
force-pushed
the
new-families-rebase
branch
from
1020f42 to
653eded
Compare
|
It looks like this molecule is the reason I'm getting most of the radicals on the surface when I run the methanol synthesis. I've added forbidden species in the families that this could match with this to forbid any labeled atom with a charge, and I no longer generate any surface radicals. But for now, this only sort of addresses the radicals proposed by these families:
So I just added a bunch of forbidden charged species, let me know if that helps alleviate the problem at all/what your input file you're running is so I can try it as well (when running with ReactionMechanismGenerator/RMG-Py#1828) |
mazeau
force-pushed
the
new-families-rebase
branch
3 times, most recently
from
9b166b1 to
f6bb63e
Compare
|
I pulled your changes and tried it again, and now I don‘t have the adsorbed CO with the charges (which was in the edge before), but I still have radical adsorbates in the edge. I added my input and output.html files here. Note: I haven't updated the surface site density value for Pt111 in the input file. |
|
Thanks, I'll take a look!
|
mazeau
force-pushed
the
new-families-rebase
branch
2 times, most recently
from
a4fca03 to
85cbd05
Compare
|
@kblondal I think I've got it so that there can be no charges or radicals on the surface for your input file, but feel free to double check! I've been running on surface_resonance branch but I'm going to try it out on the current master branch as well (and hopefully it works there, so this won't be dependent on ReactionMechanismGenerator/RMG-Py#1828) |
|
I let this run for about 8 minutes on the current master branch and I was able to find all new surface reaction families in the edge output_edge.html.txt I think it's ready to be merged in! |
mazeau
force-pushed
the
new-families-rebase
branch
2 times, most recently
from
0298aa6 to
deb062d
Compare
|
This PR is about the surface families. Can you remove the commits with thermo groups libraries and radical corrections? |
mazeau
force-pushed
the
new-families-rebase
branch
from
deb062d to
8bd4efd
Compare
mazeau
force-pushed
the
new-families-rebase
branch
4 times, most recently
from
66e7a6f to
f92c296
Compare
mazeau
force-pushed
the
new-families-rebase
branch
from
f92c296 to
ccd0bbb
Compare
This family is is for a gas phase molecule to react directly to a surface bound molecule. Most kinetic parameters were taken from Ethylene Dimerization and Oligomerization to 1-Butene and Higher Olefins with Chromium-Promoted Cobalt on Carbon Catalyst Doi:10.1021/acscatal.7b03205 and Theoretical Investigation of the Mechanisms for Olefinic Hydrogenation on Pt(110) and Pt(111) Surfaces
This is for a surface bonded molecule to "migrate" where it is bonded to the surface. This was family found in Ethylene Dimerization and Oligomerization to 1-Butene and Higher Olefins with Chromium-Promoted Cobalt on Carbon Catalyst Doi:10.1021/acscatal.7b03205 Fixed typo where I accidentally used the reverse rates instead of the forward ones
I am moving this from development to normal surface because I have expanded the tree and added training reactions from "Mechanism of Methanol Synthesis on Cu through CO2 and CO Hydrogenation", Grabow and Mavrikakis et. al. (doi:10.1021/cs200055d) and "Micro-kinetic modeling of NH3 decomposition on Ni and its application to solid oxide fuel cells", Deutschmann et. al. (doi: 10.1016/j.ces.2011.07.007)
This is for a vdW bonded species that is double bonded to dissociate and each become double bonded to the surface. This is from Mechanism of Methanol Synthesis on Cu through CO2 and CO Hydrogenation", Grabow and Mavrikakis et. al. (doi:10.1021/cs200055d)
This is for a single bonded surface species adding to a vdW double, triple, or quadruple bonded species and adsorbing to a surface. This was taken from "Mechanism of Methanol Synthesis on Cu through CO2 and CO Hydrogenation" by Grabow and Mavrikakis doi:10.1021/cs200055d
This is for a vdW species splitting, adsorbing to the surface, and transferring a functional group to a double, triple, or quadruple bonded surface species. This was taken from "Mechanism of Methanol Synthesis on Cu through CO2 and CO Hydrogenation" by Grabow and Mavrikakis doi:10.1021/cs200055d
This is for two vdW species reacting together and abstracting a functional group and then forming a single bond to the surface. This was taken from "Mechanism of Methanol Synthesis on Cu through CO2 and CO Hydrogenation" by Grabow and Mavrikakis doi:10.1021/cs200055d
This is for one surface species adding to the double/triple bond of another surface species causing the second to become double bonded to the surface. This is from "Mechanism of Methanol Synthesis on Cu through CO2 and CO Hydrogenation" by Grabow and Mavrikakis doi:10.1021/cs200055d
This is for the adsorption of a double/triple bonded vdW species to the surface while abstracting a functional group from a single bonded surface species causing it to become double bonded to the surface. This is from "Mechanism of Methanol Synthesis on Cu through CO2 and CO Hydrogenation" by Grabow and Mavrikakis doi:10.1021/cs200055d
Making the tree better and adding in training reactions from "Mechanism of Methanol Synthesis on Cu through CO2 and CO Hydrogenation", Grabow and Mavrikakis et. al. (doi:10.1021/cs200055d), "Micro-kinetic modeling of NH3 decomposition on Ni and its application to solid oxide fuel cells", Deutschmann et. al. (doi: 10.1016/j.ces.2011.07.007), "Experimental and microkinetic modeling of steady-state NO reduction by H2 on Pt/BaO/Al2O3 monolith catalysts", Xu, Clayton, Balakotaiah, Harold et al. (doi: 10.1016.j.apcatb.2007.08.008), and "Surface Reaction Kinetics of Steam- and CO2-Reforming as well as Oxidation of Methane over Nickel-Based Catalysts", Delgado et al, Catalysts, 2015, 5, 871-904.
Adding in a training reaction from "Mechanism of Methanol Synthesis on Cu through CO2 and CO Hydrogenation" by Grabow and Mavrikakis doi:10.1021/cs200055d
…e families Adding in training reactions from "Experimental and microkinetic modeling of steady-state NO reduction by H2 on Pt/BaO/Al2O3 monolith catalysts". from Xu, Clayton, Balakotaiah, Harold et al. doi: 10.1016.j.apcatb.2007.08.008 and "Surface Reaction Kinetics of Steam- and CO2-Reforming as well as Oxidation of Methane over Nickel-Based Catalysts", Delgado et al, Catalysts, 2015, 5, 871-904. and "Detailed surface reaction mechanism in a three-way catalyst", Chatterjee, Deutschmann, Warnatz et al. doi: 10.1039/b101968f
mazeau
force-pushed
the
new-families-rebase
branch
from
ccd0bbb to
b8afcb9
Compare
ndharms
left a comment
ndharms
left a comment
There was a problem hiding this comment.
All of these changes look good to me. I'm concerned the PR might be a bit too big is the only issue. Possibly break it up into smaller PRs so other reviewers can get through it faster, but the content is good. @davidfarinajr made some good comments too.
amarkpayne
left a comment
amarkpayne
left a comment
There was a problem hiding this comment.
I think this looks fine from our perspective since it has been reviewed by the RMG-Cat team, and does not affect anything else.
There were a couple of old review comments that I couldn't determine if they had been resolved already. @mazeau let me know if these comments have already been addressed. If so I will merge in this PR.
6 participants
This is v2 of #368 and also includes Katrin's #371 and includes new nitrogen data from @rwest's HAN branch. Once this ReactionMechanismGenerator/RMG-Py#1884 gets merged in, all tests should pass and the training reactions should be able to be used and averaged.