Error in thermo estimation for phenoxy radical

[ssommari]

I downloaded recently (less than a mont ago) RMG to get thermodynamics data.
Phenoxy radical is wrong.
I am not sure if your method uses Benson group additivity or Bozzelli corrections for radicals.

(SPECIES, Method, H, S , Cp 300, 400, 500, 600, 1000, 1500)
PHENOXY THERM 12.10, 76.47, 22.82, 29.76, 35.65, 40.40, 47.41, 52.15, 59.35
PHENOXY RMG 29.66, 73.71, 23.42, 30.70, 36.71, 41.44,48.24, 52.76, 59.57
PHENOL BOTH -22.30, 75.17, 24.40, 32.00, 38.32, 43.33, 50.62, 55.56, 63.16
H 52.11, 27.42, 3.52, 3.53, 3.55, 3.58, 3.61, 3.72, 3.90

If you take the delta H @298 for PHENOXY + H = PHENOL
RMG bond dissociation energy = ~104
THERM bond dissociation energy = ~86.5

In the common tables for B.D.E., 104 is the typical number for an ALKOXY radical (CCOJ), while VINOXY radycal (C=COJ) is on the order of 85. (which is the typical number for resonant system)

It seems to me that RMG cannot find the proper number of resonant radical: the RMG.log states "The number of resonance isomers is 0"

I attach my input and my output to see if you can reproduce my issue.

Further informations:
The adjacency list are taken from your online smiles to adjacency.
Using an adjacency list which use single-double bonds in the aromatic ring doesn't change anything.

Database: RMG_database

//thermo strategy? BensonOnly/QMforCyclics/Hybrid
BensonOnly
//QM method: both/gaussian03/mopac/mm4/mm4hr 
both
//maxradnumforQM?
0
//CheckConnectivity? off/check/confirm
check

MaxCarbonNumberPerSpecies:     20
MaxOxygenNumberPerSpecies:     10
MaxRadicalNumberPerSpecies:    01
MaxSulfurNumberPerSpecies:     10
MaxSiliconNumberPerSpecies:    10
MaxHeavyAtomNumberPerSpecies: 100
MaxCycleNumberPerSpecies:      10
END

PrimaryThermoLibrary:
Name: RMG-minimal
Location: primaryThermoLibrary
END

PHENOXY
1 C 0 {2,B} {6,B}
2 C 0 {1,B} {3,B}
3 C 0 {2,B} {4,B}
4 C 0 {3,B} {5,B}
5 C 0 {4,B} {6,B}
6 C 0 {1,B} {5,B} {7,S}
7 O 1 {6,S}

PHENOL
1 C 0 {2,B} {6,B}
2 C 0 {1,B} {3,B}
3 C 0 {2,B} {4,B}
4 C 0 {3,B} {5,B}
5 C 0 {4,B} {6,B}
6 C 0 {1,B} {5,B} {7,S}
7 O 0 {6,S}

~~~~~~~~~~~~~~~~~~~~~CONSOLE OUTPUT~~~~~~~~~~~~~~~~~~~~~
Running ThermoDataEstimator...
Warning: Problem in database: 2-norbornyl is not actually a child of (Cs)2CsJ
Warning: Problem in database: bicyclo[3.1.0]hexane-C5-3 is not actually a child
of (Cs)2CsJ
Warning: Could not find a non trivial ring correction!It is advised to review the thermochemistry data of this species.
Warning: Transport group not found: 1
Warning: Transport group not found: 2
Warning: Transport group not found: 3
Warning: Transport group not found: 4
Warning: Transport group not found: 5
Warning: Transport group not found: 6
Warning: Could not find a non trivial ring correction!It is advised to review the thermochemistry data of this species.
Warning: Transport group not found: 1
Warning: Transport group not found: 2
Warning: Transport group not found: 3
Warning: Transport group not found: 4
Warning: Transport group not found: 5
Warning: Transport group not found: 6

~~~~~~~~~~~~~~~~~~~~~~RMG.LOG~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 Environment Variables:
RMG = C:\RMG-Java
RMG_DATABASES = $RMG/databases (default)
RMG_JOB_SCRATCH = . (default)
RMG_JOB_OUTPUT = . (default)
RMG_QM_LIBRARY = .\QMThermoLibrary (default - relative to RMG_JOB_OUTPUT)
RMG_QM_CALCS = .\QMfiles (default - relative to RMG_JOB_OUTPUT)
Note: Overriding default MAX_CARBON_NUM with user-defined value: 20
Note: Overriding default MAX_OXYGEN_NUM with user-defined value: 10
Note: Overriding default MAX_RADICAL_NUM with user-defined value: 1
Note: Overriding default MAX_SULFUR_NUM with user-defined value: 10
Note: Overriding default MAX_SILICON_NUM with user-defined value: 10
Note: Overriding default MAX_HEAVYATOM_NUM with user-defined value: 100
Note: Overriding default MAX_CYCLE_NUM with user-defined value: 10
Reading Primary Thermo Library: RMG-minimal

Reading thermo database from C:\RMG-Java/databases/RMG_database/thermo_groups\
Reading thermochemistry groups
Reading radical correction groups
Reading ring correction groups
Reading other correction groups
Reading gauche and 1/5 correction groups
Reading polycyclic groups

Reading Lennard-Jones database from C:\RMG-Java/databases/RMG_database/transport_groups\

Created new species: PHENOXY
ChemFormula: C6H5OJ
1  C 0 {2,B} {6,B} {8,S}
2  C 0 {1,B} {3,B} {9,S}
3  C 0 {2,B} {4,B} {10,S}
4  C 0 {3,B} {5,B} {11,S}
5  C 0 {4,B} {6,B} {12,S}
6  C 0 {1,B} {5,B} {7,S}
7  O 1 {6,S}
8  H 0 {1,S}
9  H 0 {2,S}
10  H 0 {3,S}
11  H 0 {4,S}
12  H 0 {5,S}

The number of resonance isomers is 0
The NASA data is 
!Group:Cb-H Group:Cb-H  Group:Cb-H  Group:Cb-H  Group:Cb-H  Group:Cb-Os Group:Os-CbH    Radical:COJ !Ring:Ring      
!Estimated by RMG using Group Additivity
! [_ SMILES="InChI=1/C6H5O/c7-6-4-2-1-3-5-6/h1-5H" _]
PHENOXY                 C   6H   5O   1     G   250.000  5000.000   995.043    1

 1.52919677E+01 1.59418988E-02-5.27059662E-06 7.99857227E-10-4.57664626E-14    2

 7.90156373E+03-5.87774578E+01 3.01433104E+00 1.22411598E-02 8.02885732E-05    3

-1.10109663E-07 4.12829663E-11 1.29714802E+04 1.35922625E+01                   4


ThermoData is 
29.66   73.71   23.42   30.7    36.71   41.44   48.24   52.76   59.57   
 symmetry number = 2
C6H5OJ  H = 29.660000000000004

Created new species: PHENOL
ChemFormula: C6H6O
1  C 0 {2,B} {6,B} {8,S}
2  C 0 {1,B} {3,B} {9,S}
3  C 0 {2,B} {4,B} {10,S}
4  C 0 {3,B} {5,B} {11,S}
5  C 0 {4,B} {6,B} {12,S}
6  C 0 {1,B} {5,B} {7,S}
7  O 0 {6,S} {13,S}
8  H 0 {1,S}
9  H 0 {2,S}
10  H 0 {3,S}
11  H 0 {4,S}
12  H 0 {5,S}
13  H 0 {7,S}

The number of resonance isomers is 0
The NASA data is 
!Group:Cb-H Group:Cb-H  Group:Cb-H  Group:Cb-H  Group:Cb-H  Group:Cb-Os Group:Os-CbH    !Ring:Ring  
!Estimated by RMG using Group Additivity
! [_ SMILES="InChI=1/C6H6O/c7-6-4-2-1-3-5-6/h1-5,7H" _]
PHENOL                  C   6H   6O   1     G   250.000  5000.000   995.043    1

 1.53606133E+01 1.77706838E-02-5.83246616E-06 8.81211842E-10-5.02825967E-14    2

-1.83631360E+04-5.90023800E+01 2.97793246E+00 1.38472667E-02 8.10346267E-05    3

-1.11555992E-07 4.18259325E-11-1.32403864E+04 1.40340336E+01                   4


ThermoData is 
-22.3   75.17   24.4    32.0    38.32   43.33   50.62   55.56   63.16   
 symmetry number = 2
C6H6O  H = -22.299999999999997

Done!

[rwest]

Thanks @ssommari.
@nickvandewiele is this something we're aware of or is it news?
Either way we ought to at least document it (until we fix it).

[ssommari]

@rwest: I started using RMG after @ajalan's presentation at Nascre meeting.
He suggested me to post it here as a bug, so I think it is something new.
Keep going with this great job!
Samuele

[ssommari]

Maybe it's Off Topic, but more oxygenated species (alkyl peroxy radicals) seem wrong.

Enthalpies (kcal/mol:)
NIST: -4.1
RMG: -3.9
THERM: -7.4
Reported values are fairly different (-6.8)

Blanksby et al J. Am. Chem. Soc. 2001, 123, 9585-9596
(Negative-Ion Photoelectron Spectroscopy, Gas-Phase Acidity, and Thermochemistry of the Peroxyl Radicals CH3OO and CH3CH2OO)

Just so that you know.

[avasilio]

Hey All,

I have tried to enter Thermo through the website but haven't had any luck. It I have hundreds of Thermo values from literature that have been collected over the years in the Ellison lab. ( Mostly enthalpies and fewer heat capacities) I have looked through the current RMG Thermo database and the paper recently published by Franklin and still have plenty of new good Thermo to add to RMG database. It's just a matter of the best way to get into database.

I would love to do contribute this to the RMG database.

I know Connie has been very busy and has lots of tasks but maybe we can figure out... if the website is still the best way to Thermo into RMG.

Thoughts?
AJ

Dr. AnGayle K.Vasiliou
Department of Chemical Engineering
Massachusetts Institute of Technology

[rwest]

Hi @avasilio - what format do you have the data in?

[nickvandewiele]

On the topic of the wrong enthalpy of formation for phenoxy radical:

Despite the threatening warning messages (eg Warning: Could not find a non trivial ring correction) in the log file, the real problem is not situated in terms of missing ring corrections (aromatic compounds don't need ring corrections), but rather the lack of an adequate HBI for the oxygen bound to a Cb atom.

This also has nothing to do with the aromaticity perception algorithm, which worked perfectly in this case.

I believe that RMG is only able to match the generic COJ HBI entry because nothing more accurate could be found in the radical tree that matches the atomic environment of the oxygen radical in phenoxy. In other words: RMG-Java/databases/RMG_database/thermo_groups/Radical_Library.txt is missing a CbOJ entry.

@ssommari
thanks for pointing this out! It's strange that nobody detected this missing entry, despite this probably being an important HBI. for the record: RMG uses Bozzelli's HBI methodology.

Bottom line: this is a data problem, rather than an algorithmic one. Once we have a CbOJ entry, it should be fine. Not surprisingly, checking rmg.mit.edu also shows the error.