United States
Environmental Protection
Agency
Atmospheric Sciences
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S3-86/031 Aug. 1986
&EPA Project Summary
Development and Testing of a
Surrogate Species Chemical
Reaction Mechanism
William P. L Carter, Frederick W. Lurmann, Roger Atkinson,
and Alan C. Lloyd
During the first year of a two-year
program, a photochemical reaction
mechanism was updated and exten-
sively evaluated. The testing and refine-
ment of this surrogate species mecha-
nism was performed in order to create
an improved chemical mechanism for
the atmospheric simulation models
that are used to develop ozone control
strategies.
The updated mechanism was tested
against over 400 environmental cham-
ber experiments carried out in four dif-
ferent chambers. Tests were performed
to assess the accuracy of the chamber
characterization procedures, of reac-
tions for single organic compounds,
and of the overall mechanism for
complex organic mixtures, including
mixtures obtained from automobile ex-
haust. The results indicate the mecha-
nism's predictions are generally consis-
tent with the experimental data.
This Project Summary was devel-
oped by EPA's Atmospheric Sciences
Research Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in two separate volumes of the
same title (see Project Report ordering
information at back).
Introduction
Atmospheric simulation models are
essential planning tools for the develop-
ment of emission control strategies for
regions that presently exceed the ambi-
ent air quality standard for ozone. The
models are used to estimate the emis-
sions control requirements needed to
prevent exceedances of the air quality
standards in the future. One of the most
important components in atmospheric
simulation models is the chemical
mechanism that describes the forma-
tion of ozone from volatile organic com-
pounds (VOC) and oxides of nitrogen
(NOX). As part of a coordinated research
program to develop reliable ozone con-
trol strategies, the U.S. Environmental
Protection Agency (EPA) sponsored this
research study to develop and test an
improved chemical mechanism.
Many chemical mechanisms have
been developed in the last ten years for
simulating ozone formation from VOC
and NOX. All of the mechanisms in-
tended for atmospheric applications
incorporate approximations and con-
densations of species and reactions be-
cause it is currently impossible to ex-
plicitly include reactions for the
hundreds of organic compounds
present in ambient air. Several ap-
proaches are available for lumping the
organic species into a manageable
number of chemical classes in the
mechanisms. The approach adopted in
this study was to use the surrogate spe-
cies approximation where the explicit
chemistry of selected compounds is
used to represent the chemistry of all
similar compounds. For example, the
explicit reactions for propene and trans-
2-butene are used as surrogates for the
reactions of terminally double-bonded
and internally double-bonded alkenes.
Provided the mechanism is formulated
with a sufficient number of surrogate
species (10 or more), the surrogate spe-
cies approach is quite capable of repre-
senting the majority of organic species
present in urban air.
-------
The surrogate species approach has
several advantages over other schemes
such as the carbon bond lumping ap-
proach. Surrogate species mechanisms
can easily be updated and expanded
since the reactions for each surrogate
species are independent of other parts
of the mechanism. Also, since whole
molecules rather than lumped bond
groups are used in the surrogate spe-
cies mechanisms, they are not reliant on
the assumption that different parts of
the molecules react independently.
Prior to this study, surrogate species
chemical mechanisms had been sub-
jected to only limited testing against en-
vironmental chamber data. A key pur-
pose of this research program was to
extensively test the predictive ability of
an updated surrogate species mecha-
nism against chamber data for a broad
range of conditions.
Formulation of the Chemical
Mechanism
The mechanism of Atkinson et al.*
was extensively updated and expanded
to incorporate the most recent kinetic
and mechanistic data obtained in labo-
ratory studies. The new mechanism in-
cludes reactions for the following spe-
cies:
• Inorganic: The reactions of the in-
organic species NO, N02, 03, CO,
N03, N205, HNO2, HONO2, H02N02,
OH, H02, and H202 are represented
explicitly.
• Alkanes: A lumped alkane species
"C4-C5" is used to represent the re-
actions of the C4 and C5 alkanes. The
reactions of the C6 and higher al-
kanes are represented using a
"lumped reaction" approach which
represents the net effect of all C6+
alkane reactions. Ethane and
propane are assumed to be unreac-
tive.
• Alkenes: The reactions of the al-
kenes ethene, propene, and trans-2-
butene are represented explicitly.
Other terminal olefins are treated as
propene; other internal olefins are
represented as trans-2-butene.
• Aromatics: The initial reactions of
the aromatics benzene, toluene, m-
xylene, and 1,3,5-trimethylbenzene
are represented explicitly. The de-
tails of the ring-opening processes,
and the subsequent reactions of
many of the ring-opened products
•Atkinson, R., A.C. Lloyd, and K. Winges. 1982. At-
mos. Environ. 16 (6): 1341-1355.
are not known, and these are repre-
sented by parameterized, semi-
empirical mechanisms, with the
photolysis rates of the uncharacter-
ized products adjusted to fit the en-
vironmental chamber data.
• Photooxidation Products: The re-
actions of the photooxidation prod-
ucts formaldehyde, acetaldehyde,
propionaldehyde, acetone, methyl
ethyl ketone, 3-pentanol, glyoxal,
methylglyoxal, biacetyl, peroxy-
acetyl nitrate, peroxypropionyl ni-
trate, and peroxybenzoyl nitrate and
the reactions of OH radicals with
benzaldehyde and N03 radicals with
phenolic compounds are repre-
sented explicitly. The photolysis of
benzaldehyde, the reactions of OH
radicals with phenolics, and the re-
actions of the uncharacterized aro-
matic ring-opening products are
represented by parameterized
mechanisms with parameters in
some cases adjusted to fit the cham-
ber data.
This updated mechanism incorpo-
rates the significant improvements in
our understanding of atmospheric reac-
tion mechanisms which have occurred
in recent years. Despite these improve-
ments, a number of important uncer-
tainties remain, with two of the most
important areas being in the aromatic
photooxidation mechanisms and the re-
actions of ozone with the higher al-
kenes. These uncertainties necessitate
the continued use of parameterization
in certain aspects of the mechanism, as
indicated above.
Testing of the New Mechanism
The new chemical mechanism was
tested against the data from over 400
experiments. The experiments were
carried out in four different environ-
mental chambers: the University of
North Carolina's (UNC) 150,000-1 dual
outdoor chamber, the Statewide Air
Pollution Research Center's (SAPRC)
5800-1 evacuable indoor chamber (EC),
the SAPRC's 6400-1 indoor Teflon
chamber (ITC), and the SAPRC's 50,000-
1 outdoor Teflon chamber (OTC). Proce-
dures were developed to represent the
light intensities and spectral distribu-
tions of the light sources in the different
chambers. Appropriate methods were
also developed to characterize the
major chamber effects, such as NOX off-
gassing, ozone deposition and the
chamber free radical sources, in a con-
sistent manner for all four chambers.
The testing of the chemical mecha-
nism was carried out without any run-
to-run adjustment of uncertain parame-
ters to optimize the quality of fits of the
model predictions to the experimental
results. While such adjustment would
obviously yield better fits, it was not ap-
propriate for the mechanism testing
program since it could mask possible
systematic errors in the chemistry.
The types and number of environ-
mental chamber experiments used in
the testing program are summarized in
Table 1. The N0x-air, CO-NOx-air,
n-butane-NOx-air and background air
experiments were used to test and re-
fine the chamber characterization pro-
cedures. Single organic compound-
N0x-air experiments were employed to
test the reactions for each of the organic
precursor species included in the mech-
anism. Experiments with organic mix-
tures were used to test the predictive
ability of the mechanism for conditions
representative of the real atmosphere.
These ranged from simple mixtures like
propene/n-butane to complex mixtures
including more than 15 compounds and
automobile exhaust.
In general, the performance of the
mechanism in simulating the results of
the 415 experiments modeled was rea-
sonably good considering the current
state of knowledge of chamber effects
and the N0x-air photooxidation mecha-
nisms of organic compounds. Although
a number of experiments were not well
simulated, and certain systematic dis-
crepancies were observed in the simu-
lations of some groups of experiments,
the major features of the majority of the
experiments were reasonably well sim-
ulated. Most of the discrepancies ob-
served were of a random nature, sug-
gesting that these may be due primarily
to chamber-dependent parameters
which vary from run to run, rather than
systematic errors in the chemical mech-
anism. The simulations of the outdoor
chamber experiments had greater vari-
ability in the quality of the fits than the
simulations of the indoor experiments;
this is consistent with the greater vari-
ability of conditions associated with
outdoor chamber experiments.
The accuracy of the mechanism's pre-
dictions for complex mixtures of orgari-
les and NOX is most important for its use
in atmospheric simulation models and
control strategy development. The pre-
dictions for maximum ozone concentra-
tions in 117 cases with complex alkane-
alkene-aromatic mixtures are compared
to the observed values in Figure 1.
These results show that the mechanism
-------
Table 1. Summary of Environmental Chamber Runs Used for Chemical Mechanism
Testing
Number of Runs*
Types of Runs
Background air, NOx-air, Tracer-NOx, and CO-NOX
HCHO-air and HCHO-NOX
Acetaldehyde-air and Acetaldehyde-NOx
Propionaldehyde, Methyl ethyl ketone, and
Acetone-NOx
o-Cresol-NOx
Ethene-NOx
Propene-NOx
Dynamic propene-NOx
1-Butene-NOx
tra ns-2-Butene-NOx
lsobutene-NOx
n-Butane-NOx
C5+ Alkane-NOx
Benzene-N0x
Toluene-NOx
m-Xylene-NOx
o-Xylene-NOx
1,3,5-Trimethylbenzene-NOx
2-, 3- and 4-Component runs, no aromatics
2- and 3-Component runs, with aromatics
Dynamic propene-n-butane-toluene-NOx
Complex multi-component runs
Auto exhaust runs
UNC
27
13
9
5
6
22
5
4
1
7
6
5
4
15
9
4
25
23
EC
7
7
3
1
6
15
3
3
14
6
13
4
3
15
7
11
ITC
5
7
2
3
1
3
8
6
2
2
5
25
OTC
5
2
6
5
7
25
Totals
(Grand Total = 415 runs)
190
118
63
44
"UNO = UNC outdoor chamber, EC = SAPRC evacuable Indoor chamber, ITC = SAPRC indoor Teflon
chamber, OTC = SAPRC outdoor Teflon chamber.
0,8 .
I
"g 0.4
.§
0.2 .
D SAPRC EC
+ SAPRC ITC
$ SAPRC OTC
A UNC Chamber
0.2 0.4 0.6
Observed Maximum Ozone (ppm)
I
0.8
predicts the observed ozone maxima
within ±30% in about 90% of the cases.
On the average for these cases, the
mechanism overpredicts the ozone
maxima by 4%, or 0.02 ppm. The aver-
age error is ±18%, or ±0.088 ppm.
These predictions are considered quite
good. The distribution of the error in
maximum ozone predictions with re-
spect to the initial ratio of VOC to NOX is
shown in Figure 2 for these same cases.
The scattered nature of the distribution
shows that there are no systematic de-
pendencies of the error on this impor-
tant ratio, as should be the case for a
mechanism that will be applied for a
wide range of atmospheric conditions.
Conclusions
The Atkinson et al. surrogate species
chemical mechanism has been updated
to reflect the current state of chemical
knowledge. It has been tested over a
broad range of conditions and has been
shown to provide predictions that are
generally consistent with the data. The
generally high level of performance in
simulating complex organic mixtures
suggests that the mechanism is suitable
for use in atmospheric simulation
models.
A second phase of this research pro-
gram is being carried out to adapt the
chemical mechanism for use in atmos-
pheric simulation models. The mecha-
nism will be modified slightly so that it
is suitable for use in EPA's Empirical Ki-
netic Modeling Approach (EKMA). Also
a version of the mechanism with fewer
reactions and species will be developed
for use in multi-cell airshed simulation
models.
Figure 1. Comparison of predicted and observed maximum ozone concentrations for
irradiations of complex organic mixtures and /VO*
-------
0.3
0.2 -
| 0
a/-
u»
-0.2-
-0.3
f
•*
A A
+ +
D " ^
A +
P SAPRC EC
-f SAPRC ITC
O SAPRC ore
A UNC Chamber
I I I I I I I T I I I T 1 I 1I ( 1^
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Log(HC/NOt) (ppmC/ppm)
Figure 2. Error in the predicted maximum ozone concentrations^ a function oflogiofHC/NO,)
for irradiations of organic mixtures and A/0«.
-------
William P. L. Carter and Roger Atkinson are with the Statewide Air Pollution
Research Center, University of California, Riverside, CA 92521; Frederick W.
Lurmann andAlan C.Lloydare with EnvironmentalResearch & Technology, Inc.,
NewburyPark, CA 91320.
Marcia Dodge is the EPA Project Officer (see below).
The complete report consists of two volumes, entitled "Development and Testing of
a Surrogate Species Chemical Reaction Mechanism:"
"Volume I." (Order No. PB 86-212 404/AS; Cost: $28.95)
"Volume II."(Order No. PB 86-212 412/AS; Cost: $28.95)
The above reports will be available only from: (cost subject to change)
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Atmospheric Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use S300
EPA/600/S3-86/031
0000329 PS
U S ENVIR PROTECTION AGENCY
li81!"oi.a8Mt!T««T M
CHICAGO I*- 60604
------- |