Reaction with the hydroxyl (OH) radical is the dominant removal
process for volatile organic compounds (VOCs) in the atmosphere. Rate
coefficients for reactions of OH with VOCs are therefore essential parameters
for chemical mechanisms used in chemistry transport models, and are required
more generally for impact assessments involving the estimation of atmospheric
lifetimes or oxidation rates for VOCs. Updated and extended
structure–activity relationship (SAR) methods are presented for the reactions
of OH with aliphatic organic compounds, with the reactions of aromatic
organic compounds considered in a companion paper. The methods are optimized
using a preferred set of data including reactions of OH with 489 aliphatic
hydrocarbons and oxygenated organic compounds. In each case, the rate
coefficient is defined in terms of a summation of partial rate coefficients
for H abstraction or OH addition at each relevant site in the given organic
compound, so that the attack distribution is defined. The information can
therefore guide the representation of the OH reactions in the next generation
of explicit detailed chemical mechanisms. Rules governing the representation
of the subsequent reactions of the product radicals under tropospheric
conditions are also summarized, specifically their reactions with
It is well documented that volatile organic compounds (VOCs) are emitted into
the atmosphere in substantial quantities from both anthropogenic and biogenic
sources (e.g. Guenther et al., 2012; Huang et al., 2017). The degradation of
VOCs has a major influence on the chemistry of the troposphere, contributing
to the formation of ozone (
The complete gas-phase oxidation of emitted hydrocarbons and oxygenated organic compounds into carbon dioxide and water proceeds via highly detailed mechanisms, and produces a wide variety of intermediate oxidized organic products (e.g. Saunders et al., 2003; Aumont et al., 2005). As a result of the complexity of the emitted speciation, and of the degradation chemistry, the atmosphere contains an extremely large number of structurally different organic compounds, which possess a wide range of reactivities. For the majority of these, reaction with the hydroxyl (OH) radical is the dominant or exclusive removal process, such that it plays an important role in determining the atmospheric lifetime and impact of a given organic compound. As a result, quantified rate coefficients for the reactions of OH with organic compounds are essential parameters for chemical mechanisms used in chemistry transport models, and are invariably required more generally for environmental assessments of their impacts, e.g. to estimate the kinetic component of ozone formation potentials (Bufalini et al., 1976; Carter, 1994; Derwent et al., 1998; Jenkin et al., 2017) or atmospheric lifetimes for the calculation of global warming potentials (e.g. Kurylo and Orkin, 2003). In addition to the total rate coefficient, quantification of the branching ratio for attack of OH at each site within a given compound is required for explicit representation of the subsequent oxidation pathways in chemical mechanisms.
As part of the present work, a set of preferred kinetic data has been assembled for the reactions of OH with 556 organic compounds, based on reported experimental studies, of which 489 are for aliphatics (see Sect. 2 for further details). Previous assessments using explicit organic degradation mechanisms have demonstrated that the atmosphere contains an almost limitless number of organic compounds (e.g. Aumont et al., 2005), for which it is impractical to carry out experimental kinetics studies. This has resulted in the development of estimation methods for OH rate coefficients (e.g. see Calvert et al., 2015; and references therein), which have been applied widely in chemical mechanisms and impact assessments.
In the present paper, updated structure–activity relationship (SAR) methods are presented for the reactions of OH with aliphatic organic compounds, with the reactions of aromatic organic compounds considered in a companion paper (Jenkin et al., 2018a). In each case, the rate coefficient is defined in terms of a summation of partial rate coefficients for H-atom abstraction or OH addition at each relevant site in the given organic compound, so that the attack distribution is also defined. Particular use is made of the methods reported by Kwok and Atkinson (1995) and Peeters et al. (2007), which are updated and extended on the basis of the current preferred data. These approaches are also supplemented by newly developed methods for some compound classes (e.g. cumulative dienes and alkynes), and application of the methods is illustrated with examples in the Supplement.
The information is currently being used to guide the representation of the
OH-initiation reactions in the next generation of explicit detailed chemical
mechanisms, based on the Generator for Explicit Chemistry and Kinetics of
Organics in the Atmosphere (GECKO-A; Aumont et al., 2005), and the Master
Chemical Mechanism (MCM; Saunders et al., 2003). It therefore contributes to
a revised and updated set of rules that can be used in automated mechanism
construction, and provides formal documentation of the methods. To facilitate
this, rules governing the representation of the initial rapid reactions of
the product radicals under tropospheric conditions are also summarized,
specifically their reactions with
A set of preferred kinetic data has been assembled from which to develop and
validate the estimation methods for the OH rate coefficients. The complete
set includes data for 172 hydrocarbons and 384 oxygenated organic compounds.
The subset relevant to the present paper includes 298 K data for a total of
489 organic compounds, comprising alkanes (49 reactions), alkenes/polyalkenes
(92 reactions), alkynes (6 reactions), saturated oxygenated organic compounds
(259 reactions) and unsaturated aliphatic oxygenated organic compounds
(83 reactions), with temperature dependences also defined for a subset of
153 organic compounds. In two cases, the preferred rate coefficient is an
upper-limit value, and in one case a lower-limit value. The information is provided
as a part of the Supplement (spreadsheets SI_1 to SI_5). As described
in more detail in Sects. 3.2 and 4.2, the oxygenates
include both monofunctional and multifunctional compounds containing a
variety of functional groups that are prevalent in both emitted VOCs and
their degradation products, namely -OH, -OOH, -C(
The reactions of OH with saturated organic compounds almost exclusively
result in the abstraction of an H atom from a C-H or O-H bond. The
representation of H-atom abstraction reactions in the current methodology is
an update and extension to the widely applied SAR method of Kwok and Atkinson
(1995), for which selected updated parameters for 298 K have also been
reported in some other more recent studies (e.g. Atkinson, 2000; Bethel et
al., 2001; Calvert et al., 2008, 2011). The estimated rate coefficients are
thus based on a summation of rate coefficients for H-atom abstraction from
the primary (-
Arrhenius parameters (
A number of studies, including Kwok and Atkinson (1995), have defined rate
coefficients for reaction at other specific oxygenated groups, with these
also being assumed to be independent of the identity of neighbouring
substituent groups. These include abstraction of the H atom in carboxyl
(-C(
The values of
The substituent factors,
Temperature-dependent recommendations are available for 17 acyclic
(non-methane) alkanes in Arrhenius format (
Ring factors,
Notes:
Arrhenius parameters (
Notes:
The parameter values determined above can also be applied to calculate rate
coefficients for the reactions of OH with cyclic alkanes. As discussed
previously by Kwok and Atkinson (1995), ring strain has an impact on the
H-atom abstraction kinetics in cyclic systems. The data for 22 cyclic alkanes
were therefore used to optimize empirical ring-strain factors,
For polycyclic alkanes, a value of
Similarly to the values derived (or assumed) by Kwok and Atkinson (1995), the
optimized
Substituent factors,
Notes:
Temperature-dependent parameters are recommended for the series of
unsubstituted monocyclic alkanes, cyclopropane through to cyclooctane, in
Arrhenius format (see spreadsheet SI_1). The recommended
Consistent with the approach adopted by Kwok and Atkinson (1995), the value
of the rate coefficient for H-atom abstraction from a hydroxy group,
In the present work, the performance of the method is significantly improved
by defining a set of rate coefficients for H-atom abstraction from formyl
groups in RC(
The parameters in Table 4 were optimized in conjunction with the substituent
factors listed in Table 5, which relate to the general influence of hydroxyl
and carbonyl groups on H abstraction from sites other than formyl groups in
these compounds. The parameter values were initially optimized for 298 K,
using a global fit to the preferred kinetic data indicated above, using the
values of
The estimation method reproduces the observed 298 K values to within a factor of 2 for almost all of the compounds considered, with particularly good descriptions for aldehydes (within 30 %) and hydroxyaldehydes (within 10 %) due in part to the adjusted methodology described above (see Fig. 2). Similarly to the results of Bethel et al. (2001) and Mason et al. (2010), the method systematically underestimates the rate coefficients for 1,3- and 1,4-di-alcohols, by factors in the range 1.7–2.5. As also discussed previously (e.g. Mellouki et al., 2003; Calvert et al., 2011), this is likely due to longer-range influences of hydroxy substituents that are difficult to include in a practical SAR method.
Ring factors,
Notes:
Temperature-dependent recommendations are available for 32 compounds
containing combinations of carbonyl and hydroxy groups (in addition to
formaldehyde and glyoxal). These were used to provide representative values
of the temperature coefficient
The preferred data also include rate coefficients for seven cycloketones
(specifically defined as compounds where the >C
The preferred data for the reactions of OH with saturated hydroperoxides are
limited to recommended values for methyl hydroperoxide and
The limited data available suggest that a neighbouring hydroperoxy group has
a significant activating effect on OH reactivity, as discussed previously
(Jenkin et al., 1997; Saunders et al., 2003). In the present work, the value
of
Arrhenius parameters (
Notes:
Substituent factors,
Notes:
The values of a number of parameters relevant to the oxidation of ethers are
shown in Tables 7 and 8. These were optimized using the preferred data for 14
acyclic mono-ethers, 13 acyclic di-ethers and 8 acyclic hydroxyethers. The
original method of Kwok and Atkinson (1995) used the substituent factor
In the present work, the performance of the method is improved by defining a
set of three rate coefficients for H-atom abstraction from carbon atoms
adjacent to ether linkages (see Table 7), which are applied independently of
neighbouring group substituent factors. Similarly to Calvert et al. (2011), a
substituent factor for
Temperature-dependent recommendations are available for 22 of the above
acyclic compounds. Of these, the data for 11 acyclic mono-ethers were used to
provide optimized values of the temperature coefficients and pre-exponential
factors for the group rate coefficients,
The preferred data also include rate coefficients for seven cyclic
mono-ethers and five cyclic di-ethers, which were not included in the
optimization procedure described above. The limited dataset was used to
define a set of
Tables 7 and 8 also show the values of a number of parameters relevant to the
oxidation of esters. These were optimized using the preferred data for
6 formates, 10 acetates, 12 higher esters, 5 dibasic esters, 2 hydroxy esters
(lactates) and 1 carbonate. The original method of Kwok and Atkinson (1995)
used the substituent factors
Temperature-dependent recommendations are available for 18 of the above
compounds. In contrast to most of the preferred data, the preferred
temperature dependences are described by a modified Arrhenius expression of
the form
The preferred data also include rate coefficients for six carboxylic acids,
which include all the C
Temperature-dependent recommendations are available for five of the above
compounds. These were used to optimize the values of
Substituent factors,
Notes:
The preferred data include rate coefficients for sets of compounds containing
nitrate (or nitro-oxy) groups (-O
Temperature-dependent recommendations are available for methyl nitrate, ethyl
nitrate and 2-propyl nitrate. These data were used to provide optimized
values of the temperature coefficients and pre-exponential factors for the
nitrate group substituent factors, as shown in Table 9. The resultant
A log–log correlation of
The preferred data for compounds containing nitro groups include rate
coefficients for a series of five nitroalkanes, based on the atmospheric
pressure study of Nielsen et al. (1989). As discussed previously (e.g.
Calvert et al., 2011), these rate coefficients are systematically higher than
those reported at low pressure (e.g. by Liu et al., 1990), particularly for
nitromethane. This has been interpreted in terms of the reaction proceeding
by partial addition of OH to the -
The preferred data also include an upper-limit rate coefficient for
peroxyacetyl nitrate (PAN), based on the study of Talukdar et al. (1995). The
value of the substituent factor
The reaction of OH with a given unsaturated organic compound can occur by
both addition of OH to either side of each C
The estimation of rate coefficients for H-atom abstraction (
Arrhenius parameters (
Notes:
For isolated C
The values of
Substituent factors,
Notes:
The resultant values of the optimized parameters are given in Tables 10, 11
and 12. The values of
Temperature-dependent recommendations are available for eight of the acyclic
monoalkenes in Arrhenius format, as given in the preferred data in the
Supplement (spreadsheet SI_3). These were used to provide optimized values of the
temperature coefficient
The parameter values determined above were also applied to calculate rate
coefficients for the reactions of OH with six acyclic unconjugated (isolated)
dienes (i.e. with remote C
Substituent factors
Notes:
The optimized parameter values were also used to estimate rate coefficients
for the reactions of OH with 22 cyclic alkenes and cyclic unconjugated dienes
for which preferred kinetic data are available in the database. For these
calculations, no adjustments were made for possible impacts of ring strain or
steric effects on the OH addition rate coefficients, although the empirical
ring-strain factors,
A correlation of the optimized values of
Temperature-dependent parameters are recommended for limonene,
The estimation of rate coefficients for OH addition to conjugated diene
systems is also based on the method described by Peeters et al. (2007).
Site-specific rate coefficients for addition of OH to the internal carbon
atoms of the diene system can be estimated using the parameters optimized
above for monoalkenes. Addition of OH to the outer carbon atoms of the diene
system generates resonance-stabilized hydroxy-substituted radicals, for which
a further set of site-specific parameters is defined (see Peeters et al.,
2007):
The values of the group rate coefficients were initially optimized for 298 K,
using the preferred kinetic data for the 11 acyclic conjugated dienes in the
database, with 2 of these (
The value of one of the group rate coefficients (
Temperature-dependent recommendations are available for buta-1,3-diene and
isoprene in Arrhenius format. These were used to optimize values of the
temperature coefficient
The optimized parameter values were also used to estimate rate coefficients
for the reactions of OH with five cyclic conjugated dienes for which preferred
kinetic data are available in the database. As above, no adjustments were
made for the possible impacts of ring strain or steric effects on the OH
addition rate coefficients, but the empirical ring-strain factors,
A correlation of the optimized values of
Preferred kinetic data are available for the reactions of OH with four
cumulative dienes, namely propadiene, buta-1,2-diene, penta-1,2-diene and
3-methyl-buta-1,2-diene. Addition of OH to these structures cannot be
described by the parameters defined above, so a further set of site-specific
parameters is defined here, as summarized in Table 10. The rate
The values of the group rate coefficients were optimized for 298 K, using
the preferred kinetic data for the four cumulative dienes. Abstraction of an
H atom at a carbon atom adjacent to the diene system potentially generates a
resonant radical. However, because of the vinyl character of one of the
resonant forms, the corresponding substituent factor,
The values of
A temperature-dependent recommendation is available for propadiene in
Arrhenius format. A corresponding rounded value of
As indicated above, addition of OH to the central carbon atom of a cumulative
diene system leads to the radical centre being on either of the two outer
carbon atoms. In the absence of data, the formation ratio of the two possible
radical products in asymmetric systems is also based on the relative values
of the relevant rate coefficients,
Comparison of estimated and observed total branching ratios for
H-atom abstraction,
Comments: observed values reported in the following studies:
The site-specific partial rate coefficients estimated by the above methods
also define the branching ratios for both OH addition and H-atom abstraction
for the reaction of OH with a given alkene. The total 298 K branching ratios
for H-atom abstraction,
The preferred 298 K data include rate coefficients for reactions of OH with
81 unsaturated oxygenated compounds containing C
Table 11 presents substituent factors,
Parameters calculated in this way currently only apply to a limited number of
unsaturated oxygenates for which kinetic data are available, and the
corresponding abstraction routes generally make relatively minor
contributions to the overall calculated rate coefficient. As a result, this
approach must be regarded as provisional, with further information required
for its full validation. In the specific case of H-atom abstraction from a
formyl group adjacent to a C
Temperature-dependent recommendations are available for a subset of 22
unsaturated organic oxygenates. Where possible, these were used to provide
representative values of the temperature coefficients (
The site-specific partial rate coefficients estimated by the above SAR methods can also be used to define the branching ratios for both OH addition and H-atom abstraction for the reaction of OH with a given unsaturated oxygenate. Where available, the present methods appear to provide a reasonable representation of reported product yields and mechanistic information (e.g. see examples given in the Supplement).
Representative rate coefficients for rapid decomposition or
ring-opening reactions of thermalized organic radicals (
Comments:
A log–log correlation of
Reported kinetic data for the reactions of OH with alkynes are available for
ethyne (acetylene), propyne, but-1-yne, but-2-yne, pent-1-yne and hex-1-yne.
These data suggest that the rate coefficient for OH addition to C
The values of
The addition of OH can potentially occur at the carbon atoms on either side
of the triple bond. Product yields reported for propyne in some experimental
studies (Hatakeyama et al., 1986; Lockhart et al., 2013), and a theoretical
appraisal of the propyne system (Zádor and Miller, 2015), suggest that
formation of the more substituted product radical is strongly favoured, but
with evidence for addition to both sides of the C
At present, data for compounds containing both a C
Carbon-centred organic radicals (R) formed from the reactions that initiate
VOC degradation (or from other routes, such as decomposition of larger oxy
radicals) can react with molecular oxygen (
Table 14 summarizes the instances where the thermalized organic radical, R,
is represented to undergo a rapid decomposition or rearrangement that is
either its exclusive fate under atmospheric conditions or is competitive with
Reaction (R1). Organic radicals with -OOH, -OOR
In some cases, organic radicals formed specifically from the reactions of OH
with VOCs are formed chemically activated, [R] Abstraction of the formyl H atom in methylglyoxal
( For thermalized R The addition of OH to unsaturated VOCs generates chemically activated
Prompt rearrangements of chemically activated [ROO]
Comments:
Table 15 summarizes the instances where chemically activated [ROO] The reactions of As will be discussed in more detail elsewhere (Jenkin et al., 2018b), the
thermalized >C(OH)
Reported fractional formation of thermalized
Comments:
The reactions of The chemically activated acyl radical, [R The reactions of The assigned product ratios are based primarily on the OH yields reported by
Lockhart et al. (2013) for ethyne, propyne and but-2-yne, but are also
informed by the observations of The reactions of
If an organic radical possesses an allyl resonance, there are two possible
addition sites for
The reversible addition of
Root mean square error (RMSE), mean absolute error (MAE), mean bias
error (MBE) and box plot for the error distribution in the estimated log
Root mean square error, mean absolute error, mean bias error and box
plot for the error distribution in the estimated log
Root mean square error, mean absolute error, mean bias error and box
plot for the error distribution in the estimated log
Reported experimental kinetic and thermodynamic data are limited to
information on the reactions of
Partial rate coefficients for the addition of
Comments:
Arrhenius parameters (
Comments:
Peeters et al. (2014) have estimated parameters for a set of hydroxy-substituted allyl radicals formed from the addition of OH to isoprene, using a combination of DFT and ab initio methods. Suggestions for refinements were subsequently made by Peeters (2015), taking account of provisional laboratory results reported by Crounse et al. (2014). Those recommendations (given in Table S3) were previously adopted for use in MCM v3.3.1 (Jenkin et al., 2015), and remain the preferred values for the hydroxyalkyl-substituted allyl and allyl peroxy radicals formed specifically from the addition of OH to isoprene. Because the addition of OH to conjugated dienes represents an important source of allyl radicals, the information has also been used to define approximate rate coefficients for a generic set of hydroxyalkyl-substituted allyl and allyl peroxy radicals for provisional application to other systems, which are also summarized in Tables 17 and 18.
The treatment of allyl radicals containing a number of oxygenated
substituents is significantly simplified. Addition of
Updated and extended structure–activity relationship (SAR) methods have been
developed to estimate rate coefficients for the reactions of the OH radical
with aliphatic organic species. The group contribution methods were optimized
using a database including a set preferred rate coefficients for 489 species.
The overall performance of the SARs in determining log
The distribution of errors (log
The calculated log
This work has focused on the reactions of OH radicals with hydrocarbons and oxygenated organic compounds, which play a central role in tropospheric chemistry. Although outside the scope of the present study, it is noted that development of SAR methods for reactions with emitted organic compounds containing halogens, sulfur and nitrogen would also be of value.
All relevant data and supporting information have been provided in the Supplement.
All authors defined the scope of the work. MEJ developed and extended the SAR methods and drafted the manuscript, which were both reviewed by all co-authors. RV and BA tested the SAR methods in GECKO-A and carried out the statistical analysis in Sect. 7.
The authors declare that they have no conflict of interest.
This work received funding from the Alliance of Automobile Manufacturers, and as part of the MAGNIFY project, with funding from the French National Research Agency (ANR) under project ANR-14-CE01-0010, and the UK Natural Environment Research Council (NERC) via grant NE/M013448/1. It was also partially funded by the UK National Centre for Atmospheric Sciences (NCAS) Composition Directorate. Marie Camredon (LISA, Paris) and Luc Vereecken (Forschungszentrum Jülich) are gratefully acknowledged for helpful discussions on this work. We also thank Geoff Tyndall (NCAR, Boulder) and an anonymous referee for review comments and suggestions that helped to improve the manuscript. Edited by: James B. Burkholder Reviewed by: Geoffrey Tyndall and one anonymous referee