Coverage dependence sign(s) incorrect in surface kinetics library?

[rwest]

In /input/kinetics/libraries/Surface/CPOX_Pt/Deutschmann2006_adjusted

we have

entry(
    index = 30,
    label = "H2OX + OX <=> HOX + HOX",
    kinetics = SurfaceArrhenius(
        A=(1.0E20, 'cm^2/(mol*s)'),
        n = 0,
        Ea=(90500, 'J/mol'),
        Tmin = (200, 'K'),
        Tmax = (3000, 'K'),
        coverage_dependence = {'OX': {'a': 0.0, 'm': 0.0, 'E': (-240580, 'J/mol')}},
    ),
    shortDesc = u"""Default""",
    longDesc = u"""R30. H2OX is vdW H2O. Ea raised from 43.1 to 90.5 kJ/mol to
    match endothermicity of reaction.""",
	metal = "Pt",
)

RMG-database/input/kinetics/libraries/Surface/CPOX_Pt/Deutschmann2006_adjusted/reactions.py

Line 513 in 99cd172

coverage_dependence = {'OX': {'a': 0.0, 'm': 0.0, 'E': (-240580, 'J/mol')}},

note the coverage dependence term which (unless I'm mistaken?) REDUCES the activation energy as the coverage of O increases.

Compare that with the source
https://www.detchem.com/public/files/mechanisms/11_CH4_O2_ReducedGas_Quiceno2006/sm_CH4_O2_Ptwire_2006_chemkin.txt
which has

O_Pt + H2O_Pt => OH_Pt + OH_Pt                1.000E+20  0.000      43.100 
    COV / O_Pt                                0.000E+00  0.000      240.580 / 

which INCREASES the activation energy as the coverage of O increases.

(This increasing the E with O coverage may also reduce the need for your ad hoc adjustment of +47.4 kJ/mol to the barrier, to match the endothermicity. It would balance it as soon as you reached 20% O coverage. Perhaps we should keep a "non-adjusted" version of Deutschmann's model too? )

@mazeau please could you double-check all the other values for their signs.

[mazeau]

just double checked it, and that's the only one with the inverse sign. thanks for catching this. I will open a PR

[rwest]

Please could you comment on this bit:

(This increasing the E with O coverage may also reduce the need for your ad hoc adjustment of +47.4 kJ/mol to the barrier, to match the endothermicity. It would balance it as soon as you reached 20% O coverage. Perhaps we should keep a "non-adjusted" version of Deutschmann's model too? )

and then perhaps close this issue if it is fully addressed by #520?