The reaction of anthracene with OH radicals: An experimental study of the kinetics between 58 and 470 K

Article abstract of DOI:10.1063/1.1857474

The first direct measurement of the reaction rate constant of a polycyclic aromatic hydrocarbon in the gas phase in the temperature range 58-470 K is reported. The reaction is OH+ anthracene and the experiment has been performed in a continuous flow Cintique de Raction en Ecoulement Supersonique Uniforme apparatus, which had to be modified for this purpose. Pulsed laser photolysis of H2 O2 has been used to generate OH radicals and laser-induced fluorescence to observe the kinetic decay of the radicals and hence determine the rate coefficients. The reaction is found to be fast, and the rate constant increases monotonically as the temperature is lowered. The rate coefficients match the expression k (cm3 molecules-1 s-1) =1.12× 10-10 (T300) -0.46.

Full text of DOI:10.1063/1.1857474

The reaction of anthracene with OH radicals: An experimental study of the kinetics
between 58 and 470 K
F. Goulay, C. Rebrion-Rowe, J. L. Le Garrec, S. D. Le Picard, A. Canosa, and B. R. Rowe
Citation: The Journal of Chemical Physics 122, 104308 (2005); doi: 10.1063/1.1857474
View online: http://dx.doi.org/10.1063/1.1857474
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/122/10?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Reaction mechanisms and kinetics of the iminovinylidene radical with NO: Ab initio study
J. Chem. Phys. 140, 204316 (2014); 10.1063/1.4876015
Theoretical study on the gas phase reaction of allyl chloride with hydroxyl radical
J. Chem. Phys. 140, 084309 (2014); 10.1063/1.4865937
Development of a technique for high-temperature chemical kinetics: Shock tube/pulsed laser-induced
fluorescence imaging method
Rev. Sci. Instrum. 76, 064103 (2005); 10.1063/1.1938767
Dual-level direct dynamics studies for the reactions of CH 3 OCH 3 and CF 3 OCH 3 with the OH radical
J. Chem. Phys. 118, 10986 (2003); 10.1063/1.1575197
The reaction of C 2 H with H 2 : Absolute rate coefficient measurements and ab initio study
J. Chem. Phys. 116, 3700 (2002); 10.1063/1.1436481
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
138.251.14.35 On: Fri, 19 Dec 2014 21:17:10

THE JOURNAL OF CHEMICAL PHYSICS 122, 104308 ͑2005͒
The reaction of anthracene with OH radicals: An experimental study
of the kinetics between 58 and 470 K
F. Goulay, C. Rebrion-Rowe, J. L. Le Garrec, S. D. Le Picard, A. Canosa, and B. R. Rowe
Laboratoire Physique des Atomes, Lasers, Molécules et Surfaces, UMR 6627 CNRS Université Rennes
I, Campus de Beaulieu, 35042 Rennes Cedex, France
͑Received 18 November 2004; accepted 17 December 2004; published online 10 March 2005͒
The first direct measurement of the reaction rate constant of a polycyclic aromatic hydrocarbon in
the gas phase in the temperature range 58–470 K is reported. The reaction is OH+ anthracene and
the experiment has been performed in a continuous flow Cinétique de Réaction en Ecoulement
Supersonique Uniforme apparatus, which had to be modified for this purpose. Pulsed laser
photolysis of H2O2 has been used to generate OH radicals and laser-induced fluorescence to observe
the kinetic decay of the radicals and hence determine the rate coefficients. The reaction is found to
be fast, and the rate constant increases monotonically as the temperature is lowered. The rate
coefficients match the expression k͑cm3 molecules−1 s−1͒=1.12ϫ10−10͑T/300͒−0.46. © 2005
American Institute of Physics. ͓DOI: 10.1063/1.1857474͔
as naphthalene,14,17,20,23 biphenyl,16 or alkylnaphthalenes.18
I. INTRODUCTION
When the PAH becomes larger, the data become even
scarcer11,19,24–26 and to date, only three measurements of the
In their neutral or ionic forms, polycyclic aromatic hy-
reaction kinetics of anthracene with OH are available.15,24,25
Laboratory measurement of rate constants can be placed
in one of the two measurement categories: absolute or rela-
tive. The relative rate constant is obtained by monitoring the
change in the concentration of the compound of interest to-
gether with that of a reference compound due to their reac-
tions with OH radicals under conditions of continuous pro-
duction of OH. The ratio of the consumption rates is then the
same as the ratio of the rate constants. All three cited mea-
surements involving anthracene have been made in this way,
the reference compound being propene,24 acenaphthene15
͑the rate constant for this latter compound being itself ob-
tained by a relative method͒, or butene.25 In addition, these
drocarbons ͑PAHs͒ are among the most widespread chemical
species in the universe. They have been detected in extrater-
restrial media such as meteorites1 and thought to be present
in interstellar clouds, where PAH cations could be good can-
didates as the carriers of the diffuse interstellar bands.2,3
Closer to us, PAHs are ubiquitous in our environment and are
found in the atmosphere, water, soil sediments, and food.4–6
Many PAHs are mutagens and carcinogens,7 such as anthra-
cene or pyrene, and are therefore a focus of great attention by
the scientific community due to their negative impact on hu-
man health and surroundings.
PAHs can have their origin in anthropogenic and in bio-
genic sources, but most of the atmospheric PAHs are gener-
ated and emitted in the combustion processes of fossil and
nonfossil fuels.8 PAHs can either be present in fuels as a
component, such as in diesel, or being produced during the
combustion process,9,10 where they are precursors of larger
particles such as soot and fullerenes. Once released into the
atmosphere, the fate of the PAHs depends on their physico-
chemical properties.
measurements have been done either at
a
single
temperature24 or on a 20 K temperature range.15
The measurements presented in this paper are absolute
measurements, obtained by monitoring the real-time decay
of OH due to its reaction with an excess of anthracene. The
reaction rates have been determined at four temperatures,
ranging from 58 to 470 K.
Anthracene is one of the smallest PAHs and therefore is
one of the most volatile. Under typical tropospheric condi-
tions, anthracene exists mainly in the gas phase and can thus
be removed from the troposphere by gas-phase reactions
with the usual oxidants, namely, nitrate radicals, ozone, and
hydroxyl radicals, or by photolysis.11–13 In this frame, the
addition of the hydroxyl radical on the aromatic ring is the
first step of a complex mechanism leading to the formation
of oxygenated and nitro compounds14–20 and the determina-
tion of the atmospheric lifetime of PAHs rely on the deter-
mination of their reaction rate with OH.21
II. EXPERIMENT
A. Technique and apparatus
The basis of the current technique and some of the pro-
cedures used in the present experiment have been described
in detail elsewhere.27 In the Cínétique de Réaction en
Ecoulement Supersonique Uniforme ͑CRESU͒ technique,
the low temperatures are achieved by the isentropic expan-
sion of a gas mixture through a Laval nozzle. Before expan-
sion, the mixture is prepared in a reservoir, which until now
was kept at room temperature or cooled with liquid nitrogen.
Therefore, the use of the CRESU technique was until now
limited to compounds whose vapor pressure was such that
To date, OH radical reaction rate constants have been
measured with more than 900 organic chemicals,22 but very
few data exist concerning gas-phase reaction rates with
PAHs. Most of these data concern bicyclic compounds, such
0021-9606/2005/122͑10͒/104308/7/$22.50
122, 104308-1
© 2005 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
138.251.14.35 On: Fri, 19 Dec 2014 21:17:10

104308-2
Goulay et al.
J. Chem. Phys. 122, 104308 ͑2005͒
FIG. 1. Schematic view of the CRESU apparatus.
they do not condense on the walls of the reservoir. In order to
use low vapor pressure compounds, an apparatus has been
developed, in which the reservoir is heated. Here, we shall
give a relatively brief description of the method, emphasiz-
ing those aspects of the experiment which are peculiar to this
heated apparatus. A general sketch of the apparatus is given
in Fig. 1.
flow could be derived. The flow characteristics are listed in
Table I. As an example, nozzle No. 2 gave a flow which was
uniform on a 30 cm length ͑Fig. 2͒.
2. Injection of anthracene
Anthracene ͑C14H10͒ was provided by Sigma Aldrich at
99% purity and was used without further purification. At
room temperature, anthracene is a white solid. It was vapor-
ized upstream of the reservoir in an oven ͑Fig. 3͒, whose
temperature was maintained constant ͑±2.5 K͒ by a PID
regulator in the 500–550 K range. The vapor pressure of
C14H10, Pvap ͑in bar unit͒, at temperature T ͑in K͒, is mod-
eled by Antoine’s law,
1. Characterization of the nozzles
A gas mixture consisting of a carrier gas—here helium
or molecular nitrogen ͑Air Liquide, U-grade͒—reactant
gases, and the precursor of the radical ͑H2O2͒ is prepared in
the reservoir, formerly kept at room temperature and heated
in this experiment at 470 K. The expansion of this gas mix-
ture through a suitable Laval nozzle generates a supersonic
flow in which the temperature, the density of the gas, and the
mole fraction of the reagent in excess ͑here, anthracene͒ are
uniform along the flow. The density of the flow ͑typically
1016–1018 cm–3͒ is high enough to ensure thermal equilib-
rium by collisions. Several aluminum nozzles have been de-
signed for this experiment, each providing a particular tem-
perature and density for the selected carrier gas. It is also
possible to generate a subsonic flow, whose temperature is
the same as the temperature of the reservoir. All nozzles were
characterized by impact-pressure ͑Pitot͒ measurements from
which the Mach number, temperature, and density of the
B
2692.622
log10͑Pvap͒ = A −
= 4.688 976 −
,
͑T + C͒
͑T − 40.092͒
͑1͒
where A, B, and C have been obtained by fitting experimen-
tal data28 in the 418–615 K range. The anthracene vapor was
flushed into the nozzle reservoir by a carrier gas flow, at a
typical rate of 500 SCCM ͑SCCM—cubic centimeter per
minute at STP͒, this carrier gas being the same as the main
buffer gas. The carrier gas flow could be regulated by means
of mass flow controllers and adjusted, as could the total pres-
sure in the oven. The anthracene flow rate qant is determined
from the carrier flow rate qcarrier, the anthracene vapor pres-
TABLE I. Hydrodynamic conditions provided by the CRESU apparatus, with the heated reservoir.
Nozzle
No.
Buffer
gas
TReservoir
͑K͒
RReservoir
͑mbar͒
Tflow
͑K͒
Pflow
͑mbar͒
Total density
͑1016 molecules cm−3
Hydrodynamic
time ͑␮s͒
͒
1
2
3
4
He
N2
N2
N2
470
470
470
470
52.25
38.2
7.33
1.4
58
120
235
470
0.321
0.317
0.670
1.39
4.1
112
529
387
¯
1.96
2.06
2.1
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
138.251.14.35 On: Fri, 19 Dec 2014 21:17:10

104308-3
Reaction of anthracene with OH radicals
J. Chem. Phys. 122, 104308 ͑2005͒
FIG. 3. Schematic view of the oven used to vaporize anthracene, situated
upstream of the reservoir.
FIG. 2. Temperature of the flow generated by nozzle No. 2 as a function of
the distance from the exit of the nozzle, calculated from impact pressure
͑Pitot͒ measurements.
in a conical hollow aluminum support mounted on the reser-
voir and heated by thermal conduction. The reservoir and
nozzle temperatures are controlled by two K-type thermo-
couples and are always set at a value that ensures no rede-
position of anthracene before or during expansion.
sure Pvap, and the total pressure in the oven Poven by the
relation,
Pvap͑T͒
Poven − Pvap͑T͒
qant = qcarrier
.
͑2͒
4. Protection of the main chamber
To avoid blocking the oven exit and the line by large size
solid particles of anthracene, the gaseous mixture is filtered
by a Swagelock 0.5 ␮m filter ͑Swagelok SS-4FWS-05͒ in-
side the oven and to avoid condensation downstream of the
oven, the gas line between the oven and the reservoir is
heated at a higher temperature than the oven, via a 600 W
heating wire wound around the tube.
The remaining parts of the experimental apparatus are at
room temperature. When the cold supersonic jet meets the
surface of the chamber which is in front of the nozzle, it
becomes subsonic and thermodynamic considerations tell us
that its temperature recovers its initial value, that is, the tem-
perature of the reservoir. Therefore, the surface is cooled by
chilled water ͑15 °C͒ to maintain a temperature low enough
to protect the Viton o-rings. In addition, to avoid too much
anthracene penetrating into the pumps, a big cooled collector
has been placed between the chamber and the pumps. The
apparatus has, however, to be cleaned regularly, as anthra-
cene is deposited on the inner walls.
3. The heated reservoir
The use of low vapor pressure compounds led us to build
a heated reservoir. This reservoir is a stainless steel cylinder
͑see Fig. 4͒ which has triple walls. The external layer is
occupied by a circulating oil bath, which keeps the reservoir
walls at 478 K. The oil is heated and its temperature regu-
lated by a Lauda CS20CS bath. The goal of this oil circula-
tion is to warm up the buffer gas and the anthracene vapors,
which are introduced in the inner layer made of the second
and third walls. They penetrate into the reservoir through
little holes distributed all around the third wall, which is the
reservoir surface. The aluminum Laval nozzle is embedded
B. Formation and detection of the OH radicals
Hydrogen peroxide ͑Solvay Interox, 85% pure, used
without further purification͒ is our precursor of OH radicals.
Liquid H2O2 is placed in a glass vessel outside the CRESU
chamber, and its vapor is entrained by bubbling a small con-
trolled flow of He through it. Usually, the precursor of the
FIG. 4. Schematic view of the heated reservoir maintained at 470 K by a circulating oil bath.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
138.251.14.35 On: Fri, 19 Dec 2014 21:17:10

104308-4
Goulay et al.
J. Chem. Phys. 122, 104308 ͑2005͒
͑PTFE͒ tubes to a glass tube which penetrates into the reser-
voir, as close as possible to the nozzle entrance. Photolysis of
hydrogen peroxide was achieved using the fourth harmonic
͑266 nm͒ of a yttrium aluminum garnet ͑YAG͒ pulsed laser
beam ͑Spectra Physics, GCR 190, 10 Hz, 50 mJ͒ propagated
along the flow, producing a stream tube filled with a uniform
density of OH radicals. In the detection region, a tunable
pulsed laser ͑Spectra Physics, MOPO 730 with frequency
doubling option, 10 Hz, ϳ10␮J͒ propagated perpendicularly
to the flow and excited the ͑1,0͒ band of the OH͑A 2⌺
−X 2⌸͒ system at 282 nm.
The off-resonant fluorescence laser-induced fluorescence
͑LIF͒ signal from the ͑1,1͒ band at 310 nm was collected
using a UV enhanced, optically fast telescope-mirror combi-
nation mounted inside the vacuum chamber and set at right
angles to both the flow axis and the probe laser beam. It was
then imaged through a slit, onto the photocathode of a UV-
sensitive photomultiplier tube ͑Thorn EMI, 9813QSB͒ after
passing through a narrow band interference filter centered at
310 nm ͑bandpass 10 nm FWHM; Corion͒ to reduce the
scattered light from the photolysis laser. PAH deposition on
the mirror or the telescope hinder the detection, as they gen-
erate a very large fluorescence, induced by the 266 nm pho-
tons, even if the deposit is small. Therefore, the telescope-
mirror combination is flushed with a helium flow to avoid
deposition. Both lasers enter the CRESU apparatus through
Brewster angle windows. It is essential to avoid anthracene
condensation on the photolysis laser entrance window be-
cause anthracene turns to carbon black when interacting di-
rectly with the energetic 266 nm laser. Thus, the window is
protected against anthracene deposition by a small aerody-
namic window, which has a 1 cm hole on the axis of the
laser and in which helium is flown and generates a sonic
expansion.
FIG. 5. ͑a͒ First-order decay of LIF signal from OH in the presence of
19.3ϫ1012 molecules cm−3 of anthracene at 119 K in nitrogen, fit to a
single exponential decay, with residuals shown above. ͑b͒ First-order decay
constants for OH at 119 K in nitrogen, plotted against the concentration of
anthracene.
C. Deriving the rate coefficients
At a time taken as the origin, OH radicals are created
along the flow by photolysis of H2O2 at 266 nm. These OH
radicals are vibrationally cold, but their initial rotational tem-
perature is in the range 1500–1700 K. As discussed by Sims
et al.,29 10–20 ␮s are necessary to achieve rotational relax-
ation. For a fixed hydrocarbon density ͑the hydrocarbon is
always in excess͒, the LIF signal of the remaining OH radi-
radical is introduced in the reservoir together with the carrier
and reactant gases. This was not possible in this experiment,
as hydrogen peroxide decomposes at high temperature and
when in contact with stainless steel. The H2O2/He mixture
was therefore carried through PolyTetraFluoroEthylene
TABLE II. Measurements of the rate constant of OH+ butene with the CRESU apparatus. The reservoir is either
heated ͑this work͒ or at room temperature ͑Ref. 29͒.
Reservoir
temperature
͓Butene͔
Rate constant
Carrier
gas
Total density
͑1016 molec cm−3
͑1012 molecules
͑10−11 cm3
Reference
T͑K͒
͑K͒
͒
cm−3
͒
mol−1 s−1
͒
29
23
44
He
Ar
He
N2
N2
N2
N2
Ar
295
295
470
295
350
295
470
295
4.73
2.90
4.1
3–30
20–80
45.2
40.3
28.8
31.7
31.2
16.9
9.2
29
This work
29
58
30–90
75
1.67
1.85
0.57
2.06
45.6
90–290
6–25
This work
29
89.6
170
235
295
30–110
16.5–164
123–489
This work
29
6.83
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
138.251.14.35 On: Fri, 19 Dec 2014 21:17:10

104308-5
Reaction of anthracene with OH radicals
J. Chem. Phys. 122, 104308 ͑2005͒
FIG. 7. Rate constants for the reaction of OH with anthracene at different
temperatures. The filled circles show the results of the present work, and the
FIG. 6. Rate constants for the reaction of OH with ͑Z͒-but-2-ene at different
temperatures. The circles show the results of Ref. 29 in the CRESU appa-
ratus with a room temperature reservoir and the triangles show the results of
the present measurements in the CRESU apparatus with a heated reservoir.
line is a fit to our experimental data, yielding k͑cm3 molecules−1 s−1
͒
=1.12ϫ10−10͑T/298͒−0.46. Results of other authors are as indicated on the
figure.
cals is recorded and averaged over several hundreds of laser
shots, as a function of the time delay ⌬t between the pho-
tolysis and the probe laser. During this time, OH radicals
react with the hydrocarbon with a reaction rate k. After the
OH relaxation is completed, the decrease of the LIF signal as
a function of time delay was well fitted by the exponential
function exp͑−k1st⌬t͒ using a nonlinear least-squares fitting
program employing the Levenburg–Marquadt algorithm. Ex-
amination of the residuals confirmed that the decays were
truly exponential. A typical example of the trace of the LIF
signal decay from OH of the kind used to extract values of
k1st is displayed in Fig. 5͑a͒. The measurement was repeated
for different hydrocarbon flow rates. Since k1st=kϫ ͓hydro-
carbon͔, the plot of k1st versus ͓hydrocarbon͔ is a straight line
whose slope is the rate constant k, at the temperature and
density of the nozzle used ͓Fig. 5͑b͔͒. When butene was
used, it flowed directly from the cylinder and its flow was
adjusted by mass flow controllers. The anthracene flow is
varied by changing the pressure in the oven. Knowledge of
the hydrocarbon and buffer gas flows and of the total gas
density from aerodynamic ͑Pitot͒ measurements, enabled the
calculation of the hydrocarbon density in the flow.
The present results have been obtained with nozzle Nos. 1
and 3, the reservoir temperature being fixed at 470 K for
nozzle No. 1 and 470 K or 350 K for nozzle No. 3. Our
results are given in Table II, together with the results of Sims
et al.29 The graph of the rate constant versus temperature
͑Fig. 6͒ shows that the agreement is entirely satisfactory.
III. RESULTS AND DISCUSSION
A. Rate constants of OH + anthracene
Our experimental results are summarized in Table III
and displayed together with measurements and calculations
of other groups in Fig. 7.
In the experiments with anthracene, the uncertainty on
the rate constant is greater than in previous studies.27,29 This
is because the anthracene density in the flow is more difficult
to evaluate than the density of species that are gaseous at
room temperature. The regulation of the temperature of the
oven plays a crucial role in this evaluation as the vapor pres-
sure depends exponentially on the temperature of the oven: a
variation of temperature of 5 K ͑1.25%͒ induces an error of
15% on the anthracene vapor pressure. Combined with the
uncertainty of the pressure in the oven, the carrier gas flow
rate and the statistical uncertainty of the determination of
k1st, this led us to evaluate the uncertainty on the rate con-
stant to be 30%.
D. Validation of the new apparatus
To test the heated device, we have used the reaction of
͑Z͒-but-2-ene with OH. The rate constants for this reaction
have been measured in the 23–295 K range29 with the “room
temperature” CRESU PLP-LIF ͑PLP—pulsed laser photoly-
sis͒ method. The ͑Z͒-but-2-ene gas was obtained from union
carbide ͑95% purity͒ and used without further purification.
The rate constant increases monotonically as the tem-
perature is lowered and this dependence of the rate constant
on the temperature can be fitted to the expression k͑T͒
−0.46
=1.12ϫ10−10͑T/298͒
.
TABLE III. Rate constants for OH + anthracene obtained with the CRESU apparatus.
Total density
͑1016 molecules cm−3
͓Anthracene͔ Range used
Rate constant
͑10−11 cm3 mol−1 s−1
͒
T͑K͒
Carrier gas
͒
͑1012 molecules cm−3
͒
58
He
N2
N2
N2
4.01
1.96
2.06
2.14
60.5–161.4
4.6–48
25.2
12.7
16.2
8
119
235
470
9.1–59.8
9–35
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
138.251.14.35 On: Fri, 19 Dec 2014 21:17:10

104308-6
Goulay et al.
TABLE IV. Rate constants for OH + anthracene from other authors.
J. Chem. Phys. 122, 104308 ͑2005͒
Reference
compound
Rate constant
͑10−11 cm3 molecules−1 s−1
͒
Reference
T ͑K͒
P ͑mbar͒
24
25
15
15
30
Propene
325
296
346
365
300
987
987
1013
1013
/
11
Phenanthrene
Acenaphthene
Acenaphthene
Theoretical
1.3
18.3
18.0
20
Up to now, the measurements of the rate constant of
OH + anthracene were indirect measurements. They have
been obtained by reference to a compound whose reaction
rate with OH was known, this compound being propene,24
butene,25 or acenaphthene.15 Klamt30 has performed molecu-
lar orbital based calculations and proposed a room tempera-
ture value for this rate constant. All available data are sum-
marized in Table IV. The 325 K temperature value calculated
from our experimental fit is in very good agreement with the
result of Biermann et al.24 and if one considers the uncer-
tainty on our measurements, the agreement with the results
of Brubaker and Hites15 and the theoretical prediction of
Klamt30 is fair as the rate constants differ by a factor of 2. On
the contrary, the measurement of Kwok et al.25 is a factor 8
lower than our value at the same temperature, and one order
of magnitude lower than other data. This measurement, how-
ever, was performed on anthracene present in phenanthrene
as an isomeric impurity, and the authors themselves present
their measurement as “approximate.” Our measurements
confirm that the order of magnitude of the rate constant is
10−10 cm3 s−1 at room temperature, and that the measure-
ment of Kwok et al. is erroneously low.
All channels do not have the same importance, and their
importance depends on temperature and pressure. For ex-
ample, it is known that for benzene, the H-abstraction reac-
tion is slow at room temperature,34 and becomes an impor-
tant channel only at temperatures greater than 500–600 K
͑Refs. 31 and 32͒ where the adduct decomposes very rapidly
back to reactants. The same has been observed for the reac-
tion OH + naphthalene:14,23 the abstraction channel becomes
important for TϾ410 K and has a rate which is one order of
magnitude smaller than the rate of formation of the adduct.
Theoretical calculations on OH + naphthalene35 indicate that
the formation of the adduct is barrierless, and that further
reaction of the adduct to form the alcohol has a barrier rang-
ing from 105 to 174 kJ/mol, depending on what isomer is
formed.
No data are available concerning the reaction with an-
thracene, but if one supposes that it behaves like benzene and
naphthalene, then we measure the reaction rate for the for-
mation of the adduct. This is supported by three facts. First,
we do not observe a sharp decrease of the reaction rate con-
stant when the temperature increases, as expected if the ab-
straction channel became dominant. Second, our decays of
OH signal are truly exponential, and this would not be the
case if the adduct decomposes back to the reactants. Third,
the half life of the OH–benzene adduct toward decomposi-
tion has been estimated to be in the range 5–10 ms.31 If the
same prevails for the anthracene–OH adduct, then the hydro-
dynamic time of our experiments is one order of magnitude
shorter. Moreover, the adduct could be collisionally stabi-
lized under our experimental conditions ͑our measurements
have been performed at typically 1 mbar͒, as our results are
very close to those of other authors15,24 who performed their
experiments at densities 1000 times larger. In this case, the
rate constant for the formation of the adduct is likely to be at
the high-pressure limit under the conditions of these experi-
ments. The magnitude of the rate constants, and the negative
dependence of the rate constants on temperature, provide
evidence for the absence of any maximum of electronic po-
tential energy along the minimum energy path leading from
separated reagents to the adduct.
B. Reaction products
It is generally understood31–35 that two reaction path-
ways compete in the reaction of OH with aromatic rings: OH
radical addition to the aromatic ring to form an OH-aromatic
adduct, e.g.,
Þ
OH + C14H10 → ͓C14H10 – OH͔
and H-atom abstraction:
OH + C14H10 → H2O + C14H9.
The initially energy-rich OH-aromatic adduct can either de-
compose back to reactants, be collisionally stabilized or un-
dergo H elimination, leading to the formation of an alcohol:
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
138.251.14.35 On: Fri, 19 Dec 2014 21:17:10

104308-7
Reaction of anthracene with OH radicals
J. Chem. Phys. 122, 104308 ͑2005͒
anthracene,15,24 and this tendency is confirmed by molecular
orbital calculations.30 One explanation could be steric hin-
drance for reaction on the carbon atoms bearing the most
positive electric charges.15 It would be interesting to check
whether this difference between the rate constants exists over
a large temperature range.
FIG. 8. Conventional numbering of the carbon atoms in anthracene.
1H. Naraoka, A. Shimoyama, and K. Harada, Earth Planet. Sci. Lett. 184,
1 ͑2000͒.
The last question to be answered concerns the isomer
that is formed through OH addition. The numbering of the
carbon atoms in anthracene according to the chemical ab-
stracts convention is displayed in Fig. 8, and three isomers
can be formed by addition on carbon 1, 2, or 9. The net
atomic charge for each carbon has been evaluated by semi-
empirical molecular orbital calculations,15 and the carbon
with the most positive net charge is considered to be the
most reactive position, provided it is not sterically hindered.
In the case of anthracene, no steric hindrance occurs, and the
most reactive carbon is number 9. We therefore expect to
form the product displayed in Fig. 9.
2A. Léger and J. L. Puget, Astron. Astrophys. 137, L5 ͑1984͒.
3A. Léger and L. d’Hendecourt, Astron. Astrophys. 146, 81 ͑1985͒.
4R. Atkinson and J. Arey, Environ. Health Perspect. 102, 117 ͑1994͒.
5M. M. C. Ferreira, Chemosphere 44, 125 ͑2001͒.
6M. D. Guillen and P. Sopelana, in Food Safety, edited by J. P. F. D’Mello
͑CAB International, Wallingford, UK, 2003͒, p. 175.
7M. A. Mehlman, M. M. Mumtaz, O. Faroon, and C. T. De Rosa, in Ad-
vances in Modern Environmental Toxicology, edited by C. T. De Rosa, J.
S. Holler, and M. A. Mehlman ͑Princeton Scientific, Princeton, NJ, 2002͒,
Vol. 26, p. 201.
8A. M. Mastral, M. S. Callen, J. M. Lopez, R. Murillo, T. Garcia, and M. V.
Navarro, Fuel Process. Technol. 80, 183 ͑2003͒.
9I. M. Kennedy, Prog. Energy Combust. Sci. 23, 95 ͑1997͒.
10H. Richter and J. B. Howard, Prog. Energy Combust. Sci. 26, 565 ͑2000͒.
11R. Atkinson, Atmos. Environ. 34, 2063 ͑2000͒.
12R. Atkinson and J. Arey, Atmos. Environ. 37, S197 ͑2003͒.
13R. Atkinson and J. Arey, Chem. Rev. ͑Washington, D.C.͒ 103, 4605
͑2003͒.
IV. CONCLUSION
This paper reports the first direct gas-phase measurement
of the rate constant for the reaction of the hydroxyl radical
with anthracene between 58 and 470 K. The experiments
used the pulsed laser photolysis, laser-induced fluorescence
method in the supersaturated environment provided by the
CRESU apparatus. We observe a slight negative temperature
dependence of the rate constant as expected for an exother-
mic reaction without barrier on the entrance channel, and the
rate coefficient can be fitted by k͑cm3 molecules−1 s−1͒
=1.12ϫ10−10͑T/298͒−0.46. This rate constant corresponds to
the formation of the adduct. Further investigations of the
products are needed to explore whether this adduct survives,
or decomposes to form the alcohol via H elimination.
14R. Atkinson, J. Arey, B. Zielinska, and S. M. Aschmann, Environ. Sci.
Technol. 21, 1014 ͑1987͒.
15W. W. Brubaker and R. A. Hites, J. Phys. Chem. A 102, 915 ͑1998͒.
16W. W. Brubaker and R. A. Hites, Environ. Sci. Technol. 32, 3913 ͑1998͒.
17N. J. Bunce, L. Liu, J. Zhu, and D. A. Lane, Environ. Sci. Technol. 31,
2252 ͑1997͒.
18P. T. Phousongphouang and J. Arey, Environ. Sci. Technol. 36, 1947
͑2002͒.
19F. Reisen and J. Arey, Environ. Sci. Technol. 36, 4302 ͑2002͒.
20J. S. Sasaki, S. M. Aschmann, E. C. S. Kwok, R. Atkinson, and J. Arey,
Environ. Sci. Technol. 31, 3173 ͑1997͒.
21M. J. Kurylo and V. L. Orkin, Chem. Rev. ͑Washington, D.C.͒ 103, 5049
͑2003͒.
22http://kinetics.nist.gov/, last update: 2000.
23K. Lorenz and R. Zellner, Ber. Bunsenges. Phys. Chem. 87, 629 ͑1983͒.
24H. W. Biermann, H. MacLeod, R. Atkinson, A. M. Winer, and J. N. Pitts,
Jr., Environ. Sci. Technol. 19, 244 ͑1985͒.
Another point that needs further experiments is the dif-
ference between the reactivity of phenanthrene and anthra-
cene toward OH, although both molecules are isomers. All
authors who measured the reaction rate of phenanthrene
found it considerably smaller than the reaction rate with
25E. S. C. Kwok, W. P. Harger, J. Arey, and R. Atkinson, Environ. Sci.
Technol. 28, 521 ͑1994͒.
26W. Lee, P. S. Stevens, and R. A. Hites, J. Phys. Chem. A 107, 6603
͑2003͒.
27I. R. Sims, J. L. Queffelec, A. Defrance, C. Rebrion-Rowe, D. Travers, B.
R. Rowe, and I. W. M. Smith, J. Chem. Phys. 97, 8798 ͑1992͒.
28CRC Handbook of Chemistry and Physics ͑CRC, Boca Raton, FL, 1994͒.
29I. R. Sims, I. W. M. Smith, P. Bocherel, A. Defrance, D. Travers, and B. R.
Rowe, J. Chem. Soc., Faraday Trans. 90, 1473 ͑1994͒.
30A. Klamt, Chemosphere 26, 1273 ͑1994͒.
31R. Atkinson, Chem. Rev. ͑Washington, D.C.͒ 85, 69 ͑1985͒.
32C. Barckholtz, T. A. Barckholtz, and C. M. Hadad, J. Phys. Chem. A 105,
140 ͑2001͒.
33E. C. S. Kwok and R. Atkinson, Atmos. Environ. 29, 1685 ͑1995͒.
34T. S. Lay and J. W. Bozzelli, J. Phys. Chem. 100, 6543 ͑1996͒.
35A. Ricca and C. W. Bauschlicher, Jr., Chem. Phys. Lett. 328, 396 ͑2000͒.
FIG. 9. Preferred adduct formed by addition of OH on carbon 9.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
138.251.14.35 On: Fri, 19 Dec 2014 21:17:10