ArticlePDF Available

Abstract and Figures

We present a study of the formation and distribution of benzene (C6H6) on Titan. Analysis of the Cassini Mass Spectrometer (INMS) measurements of benzene densities on 12 Titan passes shows that the benzene signal exhibits an unusual time dependence, peaking ~20 s after closest approach, rather than at closest approach. We show that this behavior can be explained by recombination of phenyl radicals (C6H5) with H atoms on the walls of the instrument and that the measured signal is a combination of (1) C6H6 from the atmosphere and (2) C6H6 formed within the instrument. In parallel, we investigate Titan benzene chemistry with a set of photochemical models. A model for the ionosphere predicts that the globally averaged production rate of benzene by ion-molecule reactions is ~107 cm-2 s-1, of the same order of magnitude as the production rate by neutral reactions of ~4 × 106 cm-2 s-1. We show that benzene is quickly photolyzed in the thermosphere, and that C6H5 radicals, the main photodissociation products, are ~3 times as abundant as benzene. This result is consistent with the phenyl/benzene ratio required to match the INMS observations. Loss of benzene occurs primarily through reaction of phenyl with other radicals, leading to the formation of complex aromatic species. These species, along with benzene, diffuse downward, eventually condensing near the tropopause. We find a total production rate of solid aromatics of ~10-15 g cm-2 s-1, corresponding to an accumulated surface layer of ~3 m.
Content may be subject to copyright.
Formation and distribution of benzene on Titan
V. Vuitton,
1
R. V. Yelle,
1
and J. Cui
1
Received 29 August 2007; revised 17 December 2007; accepted 30 January 2008; published 30 May 2008.
[1] We present a study of the formation and distribution of benzene (C
6
H
6
) on Titan.
Analysis of the Cassini Mass Spectrometer (INMS) measurements of benzene densities on
12 Titan passes shows that the benzene signal exhibits an unusual time dependence,
peaking 20 s after closest approach, rather than at closest approach. We show that
this behavior can be explained by recombination of phenyl radicals (C
6
H
5
) with H atoms
on the walls of the instrument and that the measured signal is a combination of (1) C
6
H
6
from the atmosphere and (2) C
6
H
6
formed within the instrument. In parallel, we
investigate Titan benzene chemistry with a set of photochemical models. A model for the
ionosphere predicts that the globally averaged production rate of benzene by ion-molecule
reactions is 10
7
cm
2
s
1
, of the same order of magnitude as the production rate by
neutral reactions of 4 10
6
cm
2
s
1
. We show that benzene is quickly photolyzed in
the thermosphere, and that C
6
H
5
radicals, the main photodissociation products, are
3 times as abundant as benzene. This result is consistent with the phenyl/benzene ratio
required to match the INMS observations. Loss of benzene occurs primarily through
reaction of phenyl with other radicals, leading to the formation of complex aromatic
species. These species, along with benzene, diffuse downward, eventually condensing near
the tropopause. We find a total production rate of solid aromatics of 10
15
gcm
2
s
1
,
corresponding to an accumulated surface layer of 3m.
Citation: Vuitton, V., R. V. Yelle, and J. Cui (2008), Formation and distribution of benzene on Titan, J. Geophys. Res., 113, E05007,
doi:10.1029/2007JE002997.
1. Introduction
[2] One of the more surprising results from the Cassini
mission is the existence of remarkably complex molecules
in the upper atmosphere of Titan [Vuitton et al., 2006a;
Waite et al., 2007]. Although the existence of a robust
chemistry on Titan was established by results from the
Voyager mission [Hunten et al., 1984], subsequent inves-
tigations tended to focus on the chemistry of the strato-
sphere rather than the thermosphere and ionosphere. For
example, the seminal study by Yung et al. [1984], which
served as a model for all subsequent photochemical models,
considered the ionosphere as only a source of N atoms. The
synthesis of complex molecules in the Yung et al. [1984]
and subsequent models occurs primarily (sequentially) in
the stratosphere. Some later work attempted to incl ude
coupling between the chemistry of the ionosphere and
neutral atmosphere, but only in a simplified way and
without the benefit of strong observational constraints
[Banaszkiewicz et al., 2000; Wilson and Atreya, 2004].
[
3] It came as a surprise then, if not a shock, that Titan’s
upper thermosphere and ionosphere contained some very
heavy molecules [Vuitton et al., 2006a; Waite et al., 2007].
The general rule in planetary atmospheres is that lighter
species become more abundant at high altitudes because of
diffusive separation and photo- and electron-dissociation
processes, which operate primarily in the upper atmosphere.
Titan goes against this trend with ions as heavy as several
hundred amu [Coates et al., 2007] and neutrals up to 91 amu
[Vuitton et al., 2007]. It is likely that even heavier neutrals
are present but are beyo nd the mass range of Cassini
instruments.
[
4] The most abundant heavy molecule in Titan’s ther-
mosphere is the aromatic ring benzene (C
6
H
6
)[Waite et al.,
2005]. The first evidence for the presence of benzene in the
atmosphere of Titan was found with the Infrared Space
Observatory (ISO). A column density of 2.4 10
15
cm
2
between 0.2 and 20 mbar and a mole fraction of 6.1 10
10
at 110 km averaged over all latitudes was retrieved [Coustenis
et al., 2003]. The Composite InfraRed Spectrometer (CIRS)
onboard Cassini recently provided a firm detection of
benzene in Titan’s atmosphere and discovered strong ben-
zene latitude variations in the stratosphere, with an increase
of about an order of magnitude from south to north, reach-
ing a maximum abundance of about 3.5 10
9
around
70°N[Coustenis et al., 2007]. Limb scans reveal that the
mixing ratio profile is mainly constant with altitude from
168 to 307 km at 80°N[Vinatier et al., 2007]. These mid-IR
observations are not sensitive to the high altitude benzene
discovered by the INMS. Solid benzene has been tentatively
identified on the surface by the Gas Chromatograph Mass
Spectrometer (GCMS) [Niemann et al., 2005]. Upon the
impact of the Huygens probe, the GCMS inlet line heater
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 113, E05007, doi:10.1029/2007JE002997, 2008
1
Lunar and Planetary Laboratory, University of Arizona, Tucson,
Arizona, USA.
Copyright 2008 by the American Geophysical Union.
0148-0227/08/2007JE002997$09.00
E05007 1of18
warmed the Titan surface, leading to the evaporation of
liquid/solid materials and rendering thei r detection possible.
The Visual Infrared Mapping Spectrometer (VIMS) onboard
Cassini also reported a tentative detection of solid benzene
on the surface (R. N. Clark, manuscript in preparation,
2008). These observations suggest that a substantial quan-
tity of benzene is either produced by heterogeneous chem-
istry on the surface or through gas-phase chemistry in the
atmosphere followed by condensation.
[
5] The abundance of benzene detected by the Ion Neu-
tral Mass Spectrometer (INMS) on the Cassini spacecraft
[Waite et al., 200 7] is far greater than fou nd in the
stratosphere and far greater than predicted by pre-Cassini
models of the composition of the upper atmosphere. The
data from the closed ionization source indicate a mole
fraction of a few ppm at an altitude of 1000 km, whereas
photochemical models predicted a mole fraction of a few
ppb. This suggests that the source of benzene is in the upper
atmosphere.
[
6] In order to investigate the complex chemistry in
Titan’s thermosphere and ionosphere further and understand
its larger role in chemical cycles on Titan we present here an
in-depth study of the formation and distribution of benzene
on Titan. In section 2 we describe the INMS observations
and data reduction in more detail than presented previously.
In section 3, we analyze the response of the INMS to a gas
mixture containing benzene, phenyl (C
6
H
5
), methyl (CH
3
),
and H. We explain the signal at m/z = 77 and 78 by a
combination of atmospheric C
6
H
6
and recombination of
C
6
H
5
radicals with H atoms on the walls of the closed
source and the signal at m/z = 92 with the recombination of
C
6
H
5
with CH
3
radicals. We demonstrate that the benzene
and phenyl densities required to fit the data are consistent
with the results of the photochemical models discussed in
section 4. The models show that benzene is efficiently
produced by ion chemistry in the upper atmosphere. How-
ever, benzene is quickly photolyzed and phenyl radicals, the
main photodissociation products, are 3 times as abundant
as benzene in the upper atmosphere. Loss of benzene occurs
primarily through reaction of phenyl with other radicals,
producing more complex aromatic species. Section 5 dis-
cusses these results and section 6 presents some concluding
thoughts.
2. Observations and Data Reduction
2.1. Identification of C
6
H
6
and C
7
H
8
[7] Benzene has been observed in all of Cassini’s low-
altitude passes through Titan’s atmosphere [Waite et al.,
2007]. Here we consider only measurements obtained with
the Closed Source Neutral (CSN) channel of the INMS.
This operational mode utilizes an enhancement chamber in
front of the mass spectrometer that concentrates atmospheric
molecules resulting in higher sensitivity for inert molecules
that do not react with chamber walls [Waite et al., 2005].
Observations in the Open Source Neutral mode have lower
sensitivity and have not resulted in detection of benzene.
[
8] CSN spectra obtained at low altitudes show a clear
signal in the m/z = 7379 region that we interpret as
evidence for C
6
H
6
. Figures 1a1b show spectra from the
inbound and outbound legs of T16. The spectra represent
averages over the 960980 km region. Also shown are
models for the benzene spectrum based on the INMS
calibration (J. Cui, manuscript in preparation, 2008). The
fit is excellent, confirming the identification of benzene.
The high quality of the match between models and obser-
vations implies that C
6
H
6
is likely benzene rather than any
of the other (numerous) isomers. Figures 2a 2f show the
benzene cracking pattern from the INMS calibration and the
cracking patterns from the NIST database for benzene
(Figure 2b) and several other isomers of C
6
H
6
(Figures 2c
2f). The NIST data for benzene is quite similar to the INMS
calibration data but both differ markedly from the cracking
patterns for other isomers. In particular the signal at m/z =
6062 is strong in most isomers, but weaker in benzene and
in the INMS observations of Titan. This, along with the fact
that it is the most stable isomer, strongly suggests that the
signal is due to benzene.
[
9] The count rates at m/z = 78 in the outbound measure-
ments are complicated by a time shift discussed below;
however, the inbound data are organized quite well as a
function of altitude. Figures 3a 3b show the density at
m/z = 78 for T16 and T23. Here, and in the following
analysis, we concentrate on data from T16 because this pass
penetrated the deepest into Ti tan’s atmosphere and the
benzene signal is strongest. Also shown are simple diffusive
equilibrium models for a species of mass 78 amu assuming
a constant temperature determined from analysis of the N
2
density distribution [Mu¨ller-Wodarg et al., 2008]. These
simple models provide an excellent fit to the data implying
that the measured signal is due to an atmospheric species
with a mass close to 78 amu. The benzene density distri-
bution should only reflect the scale height for a mass of 78
amu if benzene is close to diffusive equilibrium. Analysis of
INMS argon measurements indicates that the homopause on
Titan is near 800 km [Yelle et al., 2008] and that molecular
diffusion should dominate over eddy mixing at the altitudes
under consideration here. For the benzene distribution to be
in diffusive equilibrium however, also requires that diffu-
sion be faster than chemical production or loss processes.
We will return to this in section 5.
[
10] Aside from the benzene peaks at m/z = 62 63 and
7379, there are no clear signatures at m/z > 60 for the
inbound T16 spectrum shown in Figure 1a. However, the
outbound T16 spectrum, shown in Figure 1b, has siz able
peaks at m/z = 6466 and m/z = 91 92. The first feature
around m/z = 65 belongs to an unaccounted species. The
signal around m/z = 91 is not much higher than the signal at
the surrounding m/z but is statistically significant because of
the much smaller associated error bars. The molecular
weight suggests the chemical formula is C
7
H
8
and the
feature is well fit by an INMS calibration spectrum for
toluene. We therefore consider this to be detection
of toluene in the INMS closed source chamber. Struc-
turally toluene is a benzene ring with a methyl radical
substituted for an H atom, so given the detection of
benzene, the presence of toluene is not surprising. In
addition to the presence of the toluene signature, the
outbound data has a much stronger benzene signature
than the inbound data.
[
11] Results from spectral analysis for the inbound data
from INMS measurements to date are summarized in Table 1
along with geometrical data for each encounter. We do not
include results from T17 or T20 because little neutral data
E05007 VUITTON ET AL.: BENZENE ON TITAN
2of18
E05007
were obtained. Included in Table 1 are the derived mole
fractions for the lowest altitudes sampled (z1) and the
predicted abundance at a constant pressure level of p0 =
1.45 10
5
Pa, which corresponds to an altitude of 1000 km
at a latitude of 60°N. The latter values are calculated by
fitting the altitude variation of the mole fraction for each
pass to an exponential to predict the value at the altitude
(z0) corresponding to p0. The mapping from altitude to
pressure is based on the empirical model described by
Mu¨ller-Wodarg et al. [2008].
[
12] The derived mole fractions are roughly consistent
with the preliminary results presented by Waite et al.
[2007]. For reasons to be explained shortly, we believe that
the derived densities do not represent the abundance of
benzene in Titan’s upper atmosphere, but rather a combina-
tion of the benzene and phenyl abundance that we will refer to
as C
6
H
x
. To derive estimates of the benzene and phenyl mole
fractions, it is necessary to examine the time behavior of the
signals in more detail.
2.2. Time Variation of Signal
[
13] Neutral species in Titan’s upper atmosphere decrease
rapidly with altitude and the temporal response of the INMS
is extremely fast (tens of microseconds), implying that the
measured signal should increase with time up to closest
approach (C/A) and decrease thereafter. For example,
Figure 4a shows that measurements of ethane (C
2
H
6
)at
m/z = 30 during T16 have a maximum at C/A and are fairly
symmetric about this time. The time dependence of the
benzene and toluene signals, shown in Figures 4b 4d, is
quite different. The signals peak 20 s after C/A, rather
than at C/A. Using the data shown in Figures 4b 4d, we
calculate a time shift relative to C/A for m/z = 78, 77, and
92 of 27.0 ± 0.7, 22.5 ± 1.9, and 22.2 ± 6.6 s. The three time
shifts are consistent within the uncertainties and the average
shift is quite large.
[
14] In principle, an asymmetry about C/A could be
caused by large horizontal variations in the atmosphere.
We consider this unlikely for two reasons. First, there is no
evidence that the time shift varies from pass to pass, as
would be expected if it were due to horizontal variations.
This is shown in Figures 5a5c where data from three Titan
passes are plotted versus time from C/A. Despite the gaps in
the measurements near C/A for T18 and T21 (due to other
INMS observations being conducted at these times) the time
behavior of all three encounters appear similar. The geom-
etry of the 3 encounters is quite different however. T21, for
example, has a C/A at 43.3° latitude and the sub-spacecraft
latitude increases monotonically with time during the en-
counter [Mu¨ller-Wodarg et al., 2008] while for T16 the
maximum latitude of 85.5° is reached near C/A, as sum-
marized in Table 1. Second, horizontal variations would
have to be extremely large to cause the observed asymme-
try. The data in Figures 5a5c show that the INMS count
rate in channel 78 is relatively constant from 0 to 50 s after
C/A. The altitude changes by 12 km during that time (for
T16), corresponding to a density decrea se by a factor of 2,
assuming that benzene has a diffusive equilibrium distribu-
tion. A constant signal implies that the horizontal variation
in density must be of the same order over a latitudinal
distance of 0.4°. This large a horizontal variation is unlikely
and is inconsistent with the observed horizontal variations
presented in Table 1. Thus we conclude that the anomalous
time behavior of the signals is due to an internal process in
the instrument rather than a property of the atmosphere.
[
15] The low densities in Titan’s upper atmosphere permit
the development of relatively large densities of radicals. The
photochemical calculations to be discussed in section 4
show that C
6
H
5
is, in fact, more abundant than C
6
H
6
. This
is a consequence of the extremely large C
6
H
6
photolysis
rate. Therefore we argue that the benzene and toluene are
produced partly by radical recombination on the walls of the
Figure 1. Data points represent a closed source neutral spectrum in the m/z = 6099 region obtained
during T16 and averaged over the 960980 km range. The solid line is a model for the benzene and
toluene spectra based on the INMS calibration. (a) Inbound leg. (b) Outbound leg.
E05007 VUITTON ET AL.: BENZENE ON TITAN
3of18
E05007
INMS. This hypothesis is tested with a surface chemistry
model in the next section.
3. Surface Chemistry Model
[16] The time delays in the C
6
H
6
and C
7
H
8
signals
discussed in the previous section and the large abundance
of H, CH
3
, and C
6
H
5
radicals discussed in section 4
suggest that these species are created by recombination of
radicals adsorbed on the chamber walls. This hypothesis
explains the time delay observed for these species be-
cause surface-diffusion limited reactions may require a
significant amount of time. It explains the presence of
both C
6
H
6
and C
7
H
8
because these are the products
produced by reaction of the main radicals in the upper
atmosphere. Methane should also be produced by reaction
of CH
3
and H on the surface, but at a negligible rate near
C/A because the density of atmospheric CH
4
is much
greater than that of CH
3
. The wall reaction hypothesis
helps to reconcile the photochemical models with the
measured C
6
H
6
signal b ecause, as discussed in the
previous section, the phenyl abundance in the upper
atmosphere is predicted to be larger than the benzene
abundance. The INMS is really measuring the sum of
benzene and phenyl and so the required production rate
of benzene is significantly reduced compared with what is
required if it were to be assumed that the signal were due
to benzene alone.
[
17] Recombination of radicals on chamber walls is a
common phenomenon in spaceflight mass spectrometers
and has on occasion been used to good advantage. For
example, Kasprzak et al. [1980] determined the density of
N in the Venusian upper atmosphere through analysis of
the NO signal in the Pioneer Venus mass spectrometer
under the assumption that NO was produced by desorption
of NO, created by recombination of adsorbed N and O on
the walls of the instrument. Other aspects of adsorption/
desorption effects on spacecraft mass spectrometers are
discussed in the work of Hedin et al. [1973]. Here we
construct our own model for this process in order to
interpret the INMS measurements in terms of atmospheric
densities and the time constant for surface chemistry within
the instrument.
[
18] For simplicity of discussion we first consider forma-
tion of benzene. The time constants describing gas flow
inside the instrument, equilibration of the walls and gas
within the instrument, and the ionization and detectio n
systems are all very small and we assume that the only
significant time constant is for the rate of radical recombi-
nation on the walls of the closed-source enhancement
chamber. In this case, the benzene density in the closed-
source chamber is determined by the instantaneous balance
between outflow from the entrance aperture to the chamber
and the combination o f the flow of benzene into the
chamber from Titan atmosphere and desorption of benzene
from the chamber walls. The desorption rate of benzene is
assumed to be proportional to the triple product of phenyl
surface density, H surface density, and the inverse of an
empirical time constant. This time constant is determined by
rate of diffusion over the surface of the chamber and the
intrinsic reaction rate for adsorbed phenyl and H. To solve
for the density of the adsorbed species in the chamber we
Figure 2. C
6
H
6
isomers mass cracking patterns for an
electron energy of 70 eV. (a) INMS calibration data for
benzene. (b) NIST data for benzene. (c) NIST data for
2,4-Hexadiyne. (d) NIST data for 1,5-Hexadiyne. (e) NIST
data for 1,3-Hexadien-5-yne. (f) NIST data for 1,5
Hexadien-3-yne.
E05007 VUITTON ET AL.: BENZENE ON TITAN
4of18
E05007
Figure 3. The variation of density with altitude for C
6
H
6
(m/z = 78) and N
2
(m/z = 28, 14) measured by
INMS. The solid lines indicate isothermal models to the altitude profile. (a) T16. Both C
6
H
6
and N
2
are
well fit by a temperature of 138 K. (b) T23. Both C
6
H
6
and N
2
are well fit by a temperature of 154 K.
Table 1. Geometrical Parameters at C/A (Latitude, Longitude and Solar Zenith Angle) and C
6
H
x
Mole
Fractions for 12 Titan Passes
a
Pass Latitude, deg Longitude, deg SZA, deg z1, km X1 z0, km X0
T5 75.3 98.9 128.9 1038.6 2.2 10
7
987.4 1.4 10
6
T16 87.7 152.3 108.4 953.2 4.0 10
6
968.8 2.2 10
6
T18 76.6 12.6 92.9 972.3 3.8 10
6
988.6 2.7 10
6
T19 66.4 4.8 83.9 1021.4 9.6 10
7
994.8 2.4 10
6
T21 46.4 99.2 127.0 1020.8 8.9 10
7
1004.4 1.5 10
6
T23 33.9 1.0 54.8 1019.8 4.5 10
6
1006.0 6.7 10
6
T25 27.6 17.4 164.2 1004.7 1.0 10
6
1006.3 9.9 10
7
T26 25.7 1.8 155.4 991.9 3.5 10
6
1006.3 2.8 10
6
T27 39.0 0.4 146.3 1025.7 2.6 10
6
1005.9 4.7 10
6
T28 46.8 1.3 141.1 994.8 5.4 10
6
1004.3 4.5 10
6
T29 54.7 2.8 135.1 985.3 5.5 10
6
1001.4 3.9 10
6
T30 62.3 5.7 128.6 974.1 7.0 10
6
997.2 4.2 10
6
a
X1 is the derived mole fraction for the lowest altitude sampled (z1) and X0 is the predicted abundance at a constant pressure
level of p0 = 1.45 10
5
Pa, which corresponds to an altitude z0 of 1000 km at a latitude of 60°N. X1 and X0 represent a
combination of the benzene and phenyl mole fractions (see section 3), they are not the real benzene atmospheric mole fraction.
Uncertainties mostly come from calibration and are estimated to be about 20%.
E05007 VUITTON ET AL.: BENZENE ON TITAN
5of18
E05007
first consider the surface density on the chamber walls:
A
w
_
M
i
¼ N
a
i
tðÞU
sc
A
x
S
1
t
ij
M
i
M
i
M
j
M
j
ð1Þ
where M
i
is the surface density of the adsorbed radical on
the walls of the closed-source chamber, the dot over M
i
indicates a time derivative, t
ij
is the time constant describing
the surface reactions of radical i with radical j, U
sc
is the
spacecraft velocity, A
x
is the area of the aperture of the
closed-source chamber, A
w
is the surface area of the closed
source chamber, and M
i
°
is the surface density for a
monolayer of species i. The quantity N
i
a
(t) is the ambient
atmospheric density of the i
th
radical. The time dependence
is explicit to emphasize that the system is driven by the
passage of the spacecraft through different altitudes on Titan
with varying atmospheric densities. The reaction of species i
and j produces a surface density M
ij
of the product species
(for example, C
6
H
5
+H! C
6
H
6
). We also assume that the
gas phase distribution of the product species in the chamber
is in steady state with the walls; thus the rate at which gas is
desorbed from the walls is equal to the outflow rate from the
closed-source aperture:
N
c
ij
U
w
A
x
4
¼
1
t
ij
M
i
M
i
M
j
M
j
ð2Þ
where U
w
is the thermal velocity at the wall temperature and
N
ij
c
is the gas density in the chamber, i.e., the density
measured by the INMS detection system.
[
19] Solution of this set of equations gives us a model for
the correlation of N
ij
c
with N
i
a
(t). The H and CH
3
densities
used in the wall chemistry model come from our photo-
chemical model, described in the next section. The C
6
H
5
and C
6
H
6
densities are adjusted to fit the data, but the H and
CH
3
are held fixed at the values predicted by photochem-
istry. Our knowledge of surface chemistry for the radicals
considered here is too uncertain to permit determination of
the chemical time constants from first principles, so we treat
them as empirical parameters determined through fits of the
model to the T16 data. The results of such a calculation for
production of C
6
H
6
are shown in Figure 6. A chemical time
constant of 15 s for H + C
6
H
5
provides the best fit to the
data.
[
20] Adequate fits to the data also require a direct
gaseous source of C
6
H
6
, shown as the dashed line in
Figure 6. The required density from this source is 2.5
10
4
cm
3
, corresponding to a mole fraction of 1.3 10
6
at an altitude of 950 km. Recently Vuitton et al. [2007]
analyzed the INMS ion data for T5 and determined a
mixing ratio of a few ppm for C
6
H
6
on Titan’s night side,
in excellent agreement with the value of 1.3 10
6
found
Figure 4. Time dependence of the signal from 200 to
200 s around closest approach for T16. (a) m/z = 30
(ethane). (b) m/z = 78 (benzene). (c) m/z = 77 (benzene).
(d) m/z = 92 (toluene).
Figure 5. Time dependence of the signal from 200 to
200 s around closest approach for m/z = 78. (a) T16.
(b) T18. (c) T21.
E05007 VUITTON ET AL.: BENZENE ON TITAN
6of18
E05007
here. Also, we estimate that the density of benzene in
Titan’s atmosphere is roughly one third that of phenyl,
consistent with the photochemical calculations presented in
the next section.
[
21] The approach described here cannot be used for
passes other than T16 because the spacecraft did not
pene trate deep enough into Titan’s atmosphere and the
signal-to-noise ratio (SNR) of the benzene signal is too
low. As a consequence, we cannot explicitly retrieve the
benzene density for passes other than T16. Additionally, the
signal-to-noise ratio (SNR) of the toluene signal is too low
for us to meaningfully constrain the chemical time constant
for C
6
H
5
+CH
3
. We therefore adopt the time constant that
we determined for H + C
6
H
5
. The fit is adequate but the low
SNR of the measurements prevents any further conclusions
for this species.
4. Chemical Models
[22] The benzene mole fractions derived in the previou s
section are many orders of magnitude larger than that
derived for the stratosphere. Coustenis et al. [2007] report
a value varying with latitude from 1.0 10
10
to 3.5 10
9
at approximately 110 km. This increase of mole fraction with
altitude indicates that benzene is formed in the upper atmo-
sphere. The flux, F, of a minor constituent is related to the
mole fraction, X, through
F ¼ K þ DðÞN
a
dX
dr
DN
a
X
1
H
i
1
H
a

ð3Þ
where, D is the molecular diffusion coefficient, K is the
eddy diffusion coefficient, N
a
is the atmospheric number
density, H
a
is the pressure scale height of the background
atmosphere and H
i
is the diffusive equilibrium scale height
of the minor species. Though we have insufficient data to
accurately compute dX/dr, the separate measurements in the
thermosphere and stratosphere indicate that dX/dr > 0 and
therefore F < 0 and benzene is flowing downward from
above. The fact that the abundance of benzene in the
thermosphere is larger than in the stratosphere indicates that
benzene is synthesized in the thermosphere.
[
23] Lavvas et al. [2008a, 2008b] developed a one-
dimensional radiative-convectiv e/photochemical/micro-
physical model in order to interpret the composition meas-
urements recently obtained by Cassini. As in previous
studies, they find the primary benzene production channel
to be the three-body reaction:
C
3
H
3
þ C
3
H
3
þ M ! C
6
H
6
þ Mk
10
ðÞ:
[24] The model can almost reproduce the benzene strato-
spheric mole fraction but it fails to predict the large mole
fractions in the thermosphere. The predictions are more than
two orders of magnitude less than the observed abundance
in the thermosphere reported here. In fact, three-body
reactions are extremely inefficient at the low densities in
Titan’s thermosphere and it seems clear that a chemical
scheme that utilizes two-body reactions is required to
explain the thermospheric abundance. This raises the ques-
tion of how benzene is produced in the thermosphere.
[
25] We can gain some insight into benzene production
by examining the required time constants. The benzene
density distribution should only reflect the scale height for a
mass of 78 amu if benzene is close to diffusive equilibrium.
The time constant characterizing diffusion can be estimated
through t
D
= H
2
/D, where H is the benzene scale height and
D is the molecular diffusion coefficient. At 1000 km the
benzene diffusion coefficient is D =3.1 10
9
cm
2
s
1
and
the measured sc al e height i s H = 21 km, i mpl ying a
Figure 6. Time dependence model fit for benzene (m/z = 78). Data points represent the density in the
closed source obtained during T16 from 200 to 400 s around closest approach. The dotted line
corresponds to the recombination of radicals on the walls of the closed source. The dashed line represents
the direct gaseous source. The solid line is the sum of the dotted and dashed lines and represents the best
fit of the databased on the surface chemistry model.
E05007 VUITTON ET AL.: BENZENE ON TITAN
7of18
E05007
diffusion time constant of t
D
= H
2
/D 1400 s. If we assume
that the benzene ring is built up by sequential reaction with
CH
4
and the intermediary molecules have roughly the same
abundance as benzene, then 5 reactions are needed and the
required reaction rate is roughly k 5/N(CH
4
)t
D
5
10
10
cm
3
s
1
. If we assume that benzene is synthesized in
three reactions involving molecules that contain 2 and 3
carbon atoms present at the ppm level in the atmosphere
then the required reaction rate is k 2 10
9
cm
3
s
1
.
These reaction rates are consistent with ion-neutral reaction
rates but not neutral-neutral reaction rates, which are much
slower. This indicates that the benzene in the ionosphere is
likely synthesized by ion chemistry. The likely precursor is
protonated benzene, i.e., C
6
H
7
+
, a significant ion in Titan’s
ionosphere [Vuitton et al., 2007]. Chemical models are
discussed in the next subsection, where we argue that
protonated and neutral benzene are closely related.
4.1. Ion Chemistry
4.1.1. Description
[
26] We use an ion chemistry model based on the 0D model
presented by Vuitton et al. [2006a, 2007]. The model is
extended to 1D in order to allow retrieval of integrated
vertical production rates. We use daytime cond itions because
the solar EUV energy flux is about 10 times the energy flux
deposited by magnetospheric electrons [Cravens et al., 2005;
Galand et al., 2006] and as a consequence complex species
are more efficiently produced on the dayside. The model
represents daytime conditions with a local solar zenith
angle of 45°, as appropriate for T17. We concentrate on
this pass because it is the on ly one that occurred in
sunlight. Unfortun ately, the neutral data for T17 were
compromised and we cannot directly compare benzene
and C
6
H
7
+
densities.
[
27] The background atmosphere is derived from an
empirical model based on the INMS data for the thermo-
sphere [Mu¨ller-Wodarg et al., 2008], CIRS results for the
stratosphere [Vinatier et al., 2007], and Huygens results for
the troposphere [Fulchignoni et al., 2005; Niemann et al.,
2005], as described in the work of Yelle et al. [2008].
Profiles of N
2
and CH
4
densities and the atmospheric
temperature profile are presented on Figure 7. The densities
and temperature are representative of the spacecraft location
at C/A for T17. To calculate the relative production rates of
N
x
+
(x = 1,2) and CH
y
+
(y = 0 4), we use the photoionization
cross sections of Samson et al. [1987] and Stolte et al.
[1998] for N
2
, and of Lee and Chiang [1983] and Samson et
al. [1989] for CH
4
. The solar flux is based on the solar EUV
flux model for Aeronomic Calculations (EUVAC), scaled to
the location of Titan [Richards et al., 1994a, 1994b]. We
can be fairly confident that ionization by energetic magne-
tospheric electrons and ions is small compared with solar
ionization [Cravens et al., 2005; Galand et al., 2006] and is
therefore neglected. We assume chemical equilibrium for
the ion species. This is a good assumption on the dayside
and at the altitudes that primarily concern us here.
[
28] The chemical network includes 150 ions contain-
ing up to 9 heavy atoms (carbon and nitrogen), 40 neutrals
and 1250 ion-molecule reactions. The densities of heavier
neutrals are fixed in the model at values either measured by
INMS in the neutral mode or determined from the observed
ion densities [Vuitton et al., 2007]. Reaction rate constants
and products are mostly retrieved from existing compila-
tions of experimental data [Anicich, 1993; Anicich and
McEwan, 1997; McEwan and Anicich, 2007]. Titan’s iono-
spheric chemistry is driven by proton exchange reactions
where ionization flows from neutrals having small proton
affinities (PA) to neutrals having larger PA [Fox and Yelle,
1997; Vuitton et al., 2006a]. We assume that CH
5
+
,C
2
H
5
+
and
HCNH
+
readily exchange a proton with neutrals having a
higher proton affinity. When the reaction has not been
studied, we assume t he rate constant to be equal to the
collision rate and calculate Langevin rates at 300 K. Use of
Langevin rates for a temperature of 150 K lead only to minor
changes in the modeled ion densities. Ion-electron recombi-
nation rates a re retrieved from the literature. When the
electron temperature dependence has not been determined,
we assume the rate constant to be proportional to T
e
0.7
and
when no data are available, we estimate the recombination
rate to be equal to 7 10
7
(300/T
e
)
0.7
cm
3
s
1
. Thermal
electron te mperature, as measured by the Cassini Radio
Plasma Wave Spectrometer (RPWS) during the T17 encoun-
ter, are shown on Figure 7 [J.-E. Wahlund, personal com-
munication, 2007].
[
29] According to our chemical model C
6
H
6
is produced
primarily from the sequence of reactions presented on
Figure 8 and listed in Table 2. For most reactions, the
structure of the ions produced is not known because exper-
imental data are usually based on mass spectrometry. The
form of C
6
H
5
+
and C
6
H
7
+
, whether cyclic or not, likely
determines if electron recombination produces benzene or
another C
6
H
6
isomer. The reaction C
4
H
3
+
+C
2
H
2
! C
6
H
5
+
+
hn (k
0
11
) produces 2 isomers, one cyclic and one linear. The
ground state isomer c-C
6
H
5
+
is reactive but l-C
6
H
5
+
is not
[Scottetal., 1997]. Generally, when there is radiative
stabilization, rin g closure occurs and the cyclic form domi-
nates [McEwan et al., 1999]. Accordingly, we assume that
C
6
H
5
+
and C
6
H
7
+
are cyclic and we use rate constants
corresponding to c-C
6
H
5
+
and c-C
6
H
7
+
, if available.
Figure 7. N
2
(solid line) and CH
4
(dashed) density, and
neutral T
n
(dotted line) and electron temperature T
e
(dot-
dashed line) used in the ion chemistry model.
E05007 VUITTON ET AL.: BENZENE ON TITAN
8of18
E05007
[30] The electron recombination rate of c-C
6
H
7
+
has been
measured for electron temperatures of 300 to 800 K but
products are not known [McLain et al., 2004]. We assume
that benzene and H are the only products. This is a key
assumption in our model and laboratory measurements are
desperately needed. It is also possible that products other than
benzene would be quickly converted by subsequent reactions
to benzene. Other recombination of heavy ions may also lead
to benzene production, or to molecules that get converted to
benzene, but these possibilities are not considered here.
4.1.2. Results
[
31] The calculated altitude profiles for some key ions in
benzene production are compared to the measured densities
from T17 inbound in Figures 9a9b. For the lighter species
(Figure 9a), the model results are in good agreement with
the observations, suggesting that solar UV deposition and
primary ion chemistry are well described by the model. For
Table 2. Principal Ion Reactions for Benzene Formation
R# Reaction Rate Constant
a
Reference
k
0
1
CH
2
+
+CH
4
! C
2
H
4
+
+H
2
9.1 10
10
[Anicich, 1993]
k
0
2
CH
3
+
+CH
4
! C
2
H
5
+
+H
2
1.1 10
9
[Anicich et al., 2003]
k
0
3
CH
4
+
+CH
4
! CH
5
+
+CH
3
1.1 10
9
[Anicich, 1993]
k
0
4
CH
5
+
+ HCN ! HCNH
+
+CH
4
4.1 10
9
estimate
k
0
5
C
2
H
4
+
+C
2
H
2
! C
4
H
5
+
+H 1.9 10
10
[Anicich, 1993]
k
0
6
C
2
H
5
+
+C
2
H
2
! C
4
H
5
+
+H
2
1.2 10
10
[Kim et al., 1977]
k
0
7
C
2
H
5
+
+C
4
H
2
! C
4
H
3
+
+C
2
H
4
1.0 10
9
estimate
k
0
8
C
2
H
5
+
+C
6
H
6
! C
6
H
7
+
+C
2
H
4
1.1 10
9
estimate
k
0
9
C
2
H
5
+
+ HCN ! HCNH
+
+C
2
H
4
2.7 10
9
[Mackay et al., 1980]
k
0
10
C
4
H
2
+
+C
2
H
4
! C
6
H
5
+
+H 7.2 10
10
[Anicich et al., 2006]
k
0
11
C
4
H
3
+
+C
2
H
2
! C
6
H
5
+
+hv 2.2 10
10
[Anicich, 1993]
k
0
12
C
4
H
3
+
+C
2
H
4
! C
6
H
5
+
+H
2
1.2 10
10
[Anicich et al., 2006]
k
0
13
C
4
H
3
+
+C
6
H
6
! C
6
H
7
+
+C
4
H
2
1.3 10
9
[Deakyne et al., 1987; Lifshitz and Weiss, 1980]
k
0
14
C
4
H
5
+
+C
2
H
4
! C
6
H
7
+
+H
2
7.3 10
11
[Anicich et al., 2006]
k
0
15
C
6
H
5
+
+H
2
! C
6
H
7
+
+hv 6.0 10
11
[McEwan et al., 1999]
k
0
16
C
6
H
5
+
+C
2
H
4
! C
6
H
7
+
+C
2
H
2
1.0 10
10
[Anicich et al., 2006]
k
0
17a
N
+
+CH
4
! CH
3
+
+NH 5.8 10
10
[Anicich, 1993]
k
0
17b
N
+
+CH
4
! CH
4
+
+N 5.8 10
11
[Anicich, 1993]
k
0
17c
N
+
+CH
4
! HCNH
+
+H
2
4.1 10
10
[Anicich, 1993]
k
0
18
N
2
+
+H
2
! N
2
H
+
+H 2.0 10
9
[Anicich, 1993]
k
0
19a
N
2
+
+CH
4
! CH
2
+
+N
2
+H
2
1.0 10
10
[Anicich et al., 2006]
k
0
19b
N
2
+
+CH
4
! CH
3
+
+N
2
+H 1.0 10
9
[Anicich et al., 2006]
k
0
20
N
2
H
+
+CH
4
! CH
5
+
+N
2
8.9 10
10
[Burt et al., 1970]
k
0
21
HCNH
+
+C
4
H
2
! C
4
H
3
+
+ HCN 1.8 10
9
[Petrie et al., 1991]
k
0
22
HCNH
+
+C
6
H
6
! C
6
H
7
+
+ HCN 1.1 10
9
estimate
a
1
C
6
H
7
+
+e
! C
6
H
6
+H 2.4 10
6
(300/T
e
)
1.3
[McLain et al., 2004]
a
Bimolecular rate constant (cm
3
s
1
).
Figure 8. Flowchart of the principal ion chemistry reactions for benzene formation.
E05007 VUITTON ET AL.: BENZENE ON TITAN
9of18
E05007
the heavier species (Figure 9b), the model reproduces the
data down to 1100 km but at lower altitude, the model
predictions systematically overestimates the observed ion
densities. The modeled ion density peak decreases in
altitude with the ion mass because of the smaller scale
height of the heavier neutrals. However, the observed ion
density peak is located around 1100 km for all the species,
independently of their mass. Observational effects might
cause this unexpected behavior. Because the width of the
velocity distribution of heavy ions is narrower than the
width of the velocity distribution of lighter ions, any
mistuning of the lenses of the mass spectrometer caused
by strong winds, etc. would be reflected as an underesti-
mation of the higher masses density.
[
32] Some chemistry is possibly missing in our model but
it is not clear if this could explain the discrepancy observed
at low altitude. The Electron Spectrometer (ELS), one of
Cassini Plasma Spectrometer (CAPS) instruments, detected
negative ions below 1150 km in all Titan flybys [Coates et
al., 2007]. Even if the negative ion density were a substan-
tial fraction of the electron density, positive-negative ion
recombination would not substantially reduce the positive
ion density, because recombination with negative ions is
about 10 times slower than recombination with electrons.
Also, if the chemistry is complex enough to lead to the
formation of nanometer size aerosols, as suggested by the
detection of high molecular mass (>10,000 amu) negative
ions during T16 [Coates et al., 2007; Waite et al., 2007], ion
reactions on the surface of aerosols could act as a sink for
ions. Generally, addition reactions of heavier ions are faster
than addition reaction s of lighter ions because of their
higher number of vibrational modes required to redistribute
the energy liberated by the reaction. Better observational
constrain on both negative ions and heavier molecular mass
positive ions are required before these species can be taken
into account in the model.
[
33] Our model cannot reproduce the observed ion den-
sity at m/z = 50. As explained in the work of Vuitton et al.
[2007], this mass is attributed to C
4
H
2
+
but ion-molecule
reactions do not efficiently produce C
4
H
2
+
. Here we improve
the previous model by including direct photoionization of
C
4
H
2
.WeusetheC
4
H
2
mole fraction inferred from the
C
4
H
3
+
density, the photoionization cross sections of Cool et
al. [2005] and the solar flux described previously. This
process forms a significant amount of C
4
H
2
+
but the modeled
density profile still falls short by about one order of
magnitude. As represented on Figure 8 and discussed in
the next paragraph, C
4
H
2
+
is an important intermediate in
C
6
H
7
+
formation and in order to estimate the benzene
production rate it is crucial that the model accurately
predicts the C
4
H
2
+
density. Fragmentation of heavier ions
could possibly produce C
4
H
2
+
ions but without furth er
experimental constraints, it is not possible to include such
reactions in the model. As a consequence, we scale the
C
4
H
2
+
production rate by a factor of 10 in order to get a
forced agreement between observed and calculated C
4
H
2
+
densities, as showed on Figure 9b.
[
34] On the night-side, Vuitton et al. [2007] showed that
because the ionization rate by magnetospheric electrons is
small, C
6
H
7
+
is mostly produced by proton exchange reac-
tions between the major ions and C
6
H
6
, and lost by electron
recombination to form back C
6
H
6
. As a consequence, C
6
H
7
+
is constantly recycled and ion chemistry does not contribute
to a net production of C
6
H
6
. However, on the dayside, N
2
and CH
4
are rapidly photoionized and C
6
H
7
+
is efficiently
produced by ion chemistry reactions. Figure 9b shows the
altitude profile for C
6
H
7
+
ions on T17 inbound. The lower
peak in the C
6
H
7
+
density is from
C
6
H
þ
5
þ C
2
H
4
! C
6
H
þ
7
þ C
2
H
2
k
0
16

;
and
C
6
H
þ
5
þ H
2
! C
6
H
þ
7
þ hn k
0
15

;
Figure 9. Observed vertical profiles of some key ions
during T17 inbound. The solid lines represent the predic-
tions of the ion chemistry model. (a) m/z = 14, 15, 16, 17,
28, 29. (b) m/z = 50, 51, 53, 77, 79. The C
4
H
2
+
curve has
been scaled by a factor of 10.
E05007 VUITTON ET AL.: BENZENE ON TITAN
10 of 18
E05007
with a total column production rate of 3.2 10
7
cm
2
s
1
.
C
6
H
5
+
comes primarily from
C
4
H
þ
2
þ C
2
H
4
! C
6
H
þ
5
þ Hk
0
10

;
so it is necessary to accurately match the C
4
H
2
+
density, as
discussed above. The upper peak in the C
6
H
7
+
density is
from
C
4
H
þ
5
þ C
2
H
4
! C
6
H
þ
7
þ H
2
k
0
14

:
with a total column production rate of 1.6 10
6
cm
2
s
1
.
Benzene is subsequently produced by electron recombina-
tion of C
6
H
7
+
:
C
6
H
þ
7
þ e
! C
6
H
6
þ H a
1
ðÞ:
[35] The measured electron density and temperature, the
density of C
6
H
7
+
, and the recombination coefficient imply a
C
6
H
6
production rate of 1.0 10
1
cm
3
s
1
at 1000 km.
The column-integrated production rate at a solar zenith
angle of 45° derived from the ionospheric model is 3.7
10
7
cm
2
s
1
, which corresponds to a globally averaged rate
of 1.3 10
7
cm
2
s
1
. These rates and all the subsequent
column integrated rates are referred to the surface. Approx-
imately 95% of the benzene production occurs below the
altitude for which we have measurements and thus represents
an extrapolation based on our calculation of solar ionization
and our chemical reaction list. Nevertheless, our calculated
ionospheric production rate is sufficient to produce large
benzene densities in Titan’s upper atmosphere, of the same
order as those measured by the INMS.
4.2. Neutral Chemistry
4.2.1. Description
[
36] To investigate this further we have constructed a
neutral photochemical/diffusion model for the distribution
of benzene. The model includes eddy and molecular diffu-
sion, photolysis, and chemistry. Our photolysis rates are
based on a solar spectrum obtained with the EUV flux
model for Aeronomic Calculations (EUVAC), scaled to the
location of Titan [Richards et al., 1994a, 1994b]. We adopt
a solar zenith angle of 60° and divide the rates by a factor of
2, to represent a global average. The temperature and total
density are representative of the equatorial region, as
determined in the empirical model of Mu¨ller-Wodarg et
al. [2008]. The N
2
and CH
4
densities are presented on
Figure 10 and the temperature profile is shown in the work
of Yelle et al. [2008]. Molecular diffusion coefficients are
taken from measurements presented in the literature if
possible [Mason and Marrero, 1970] or calculated from
molecular force constants using the standard approach
[Hirschfelder et al., 1964] if no measurements are available.
We found no measurements for the diffusion coefficient of
benzene in N
2
. We calculate that the diffusion coefficient
has a value of DN =2.3 10
19
cm
1
s
1
at a temperature
of 150 K. The eddy mixing profile is constrained by INMS
measurements of the
40
Ar distribution in the upper atmo-
sphere and by CIRS measurements of the C
2
H
6
distribution
in the stratosphere [Yelle et al., 2008] and is presented on
Figure 10. The 1D chemical-diffusion mode l does not
include ions or ion chemistry. We specify an ion-molecule
production rate for benzene consistent with that inferred
from our ion-molecule reaction model. The production rate
is specified as a Chapman production profile with a peak at
830 km of 1.3 cm
3
s
1
and a scale height of 75 km. The
column-integrated production rate is 7 10
6
cm
2
s
1
,
which is on the same order as the rate of 1.3 10
7
cm
2
s
1
predicted by the ion chemistry model.
[
37] The neutral chemical network is based on the reac-
tion list presented in the work of Vuitton et al. [2006b]. It
consists of 40 hydrocarbon species, 130 neutral-neutral
reactions and 40 photoreactions. Chemical reactions and
photolysis of C
6
H
6
(benzene), C
7
H
8
(toluene), C
8
H
10
(eth-
ylbenzene) and their associated radicals is described explic-
itly. Oxygen and nitrogen bearing species are not abundant
enough to significantly affect the hydrocarbon densities and
in order to avoid a significant increase of the chemical
network complexity, we do not consider the nitrogen or
oxygen chemistry in this study. Molecules with more than
6 carbon atoms are assumed to ultimately form soot and are
no longer involved in the chemistry. Because we expect
benzene rings to be conserved, complex species produced
by reaction of C
6
H
5
or other aromatic radicals are included
in a sub-category of soot called ‘rings’’. ‘aromatics’ refers
to C
6
H
6
,C
7
H
8
,C
8
H
10
and ‘rings’’, ‘solid aromatics’ refers
to condensed C
6
H
6
,C
7
H
8
,C
8
H
10
, and ‘rings’’. Reactions
relevant to benzene production and destruction are listed in
Table 3.
[
38] Besides the ionospheric source, we include produc-
tion through neutral chemistry and extend the chemical
network considered in previous models. There have been
three rece nt studies of the kinetics of the propargyl
(C
3
H
3
) self-reaction at 295 K, reporting rate coefficients of
4 10
11
cm
3
s
1
[Atkinson and Hudgens, 1999; De Sain
and Taatjes, 2003; Fahr and Nayak, 2000]. These new
results seem to supersede the threefold higher v alue
obtained by Morter et al. [1994]. This rate coefficient is
independent of total pressure and buffer choice (Ar, He, N
2
)
over the pressure range 3 133 mbar, indicating that the
Figure 10. N
2
(solid line), CH
4
(dashed line) and eddy
diffusion K (dotted line) pro files used in the neutral
chemistry model.
E05007 VUITTON ET AL.: BENZENE ON TITAN
11 of 18
E05007
reaction is at the high-pressure limit in the experimental
conditions. The rate coefficient exhibits a slight temperature
dependence from 373 to 1000 K [Giri et al., 2003; Shafir et
al., 2003] but no measurement is available below room
temperature. As a consequence, we used a temperature
independent rate coefficient in our model. The reaction rate
has never been measured in the low-pressure range. We
assume that it is equal to 10 times the recombination rate of
methyl (CH
3
) radicals because of the higher number of
channels available to redistribute the excess energy. We base
our estimate on the CH
3
recombination rate of Wang et al.
[2003], which was obtained from a combination of theoret-
ical and experimental investigations down to 200 K.
[
39]TheC
6
H
6
produced in the reaction of propargyl
radicals is not necessarily in the form of benzene. Fahr
and Nayak [2000] analyzed products for propargyl recom-
bination at T = 298 K and P = 67 mbar. They detected two
linear C
6
H
6
(1,5hexadiyne and 1,2hexadiene5yne)
with yields of 60 and 25%, respectively and a third
unidentified isomer. Howe and Fahr [2003] performed
product studies at a pressure range of 27 mbar to 933 mbar
and at a temperature range of 295 623 K. The same two
linear and three new cyclic C
6
H
6
isomers (dimethylenecy-
clobutene, fulvene and benzene) were identified. Benzene
was a major product particularly at low pressure (25%).
The authors suggest that extrapolation of the product yield
to low pressures (<15 mbar) gives a yield of about 40%,
irrespective of the temperature. This is in qualitative agree-
ment with the theoretical results of Miller and Klippenstein
[2003]. In the absence of any data concerning the isomer-
ization or reactivity of the different C
6
H
6
isomers, we
assume that cyclization quickly occurs and that the only
C
6
H
6
product is benzene.
[
40] Reactions of the C
3
H
2
,C
3
H
3
,C
4
H
3
,C
4
H
5
and
C
6
H
5
radicals with C
2
H
2
have been considered as a
potential source of C
6
H
6
[Lebonnois, 2005; Wilson et al.,
2003]. However, a large energy barrier makes them
unimportant at low temperatures [Moses et al., 2005;
Vuitton et al., 2006b] and we do not consider these
reactions in our model. Instead, we include in the chemical
network the reaction between C
2
H and 1,3-butadiene
(C
4
H
6
). The reaction is fast (k =3 10
10
cm
3
s
1
)
and 1,3-butadiene is efficiently produced in the recombi-
nation of C
2
H
3
radicals [Fahr et al., 1991; Nizamov and
Leone, 2004]. The products have not been determined but
it is usually believed that the reaction of C
2
H with
unsaturated molecules proceeds via attack of a p-orbital
followed by rapid loss of an H atom by the energized
complex initially formed. It follows that the product of
reaction k
4
is most probably C
6
H
6
.
[
41] The most important chemical loss process for ben-
zene is photolysis, which extends as far as 270 nm [Pantos
et al., 1978]. To estimate the photolysis rate, we use the
cross section published by Koch and Otto [1972] and
Pantos et al. [1978]. However, the photolysis of benzene
is complicated, and the products are not completely char-
acterized. Deactivation of the excited electronic states
formed by absorption of ultraviolet photons lead to a
transient highly excited state of the ground electronic state
(C
6
H
6
*). This state has a lifetime of up to 2 ms and
subsequently fragments into the photodissociation products
[Tsai et al.,2000].C
6
H
6
* can also be deactivated by
collisions with N
2
[Nakashima and Yoshihara, 1982]. As
a consequence, C
6
H
6
* can be collisionally stabilized if the
background atmospheric pressure is high enough, whereas
dissociation occurs if collisions are rare. Our model
Table 3. Principal Neutral Reactions for Benzene Formation
R# Reaction Rate Constant
a
T Range
b
Reference
k
1
H+C
6
H
5
! C
6
H
6
1.7 10
23
T
1.8
300
10
00 10 k
0
(H + CH
3
)[Baulch et al., 1994]
6.9 10
11
T
+0.15
200 2000 [Harding et al., 2005]
k
2
CH
3
+C
6
H
5
! C
7
H
8
2.2 10
15
T
3.75
exp(494/T) 200 1350 10 k
0
(CH
3
+CH
3
)[Wang et al., 2003]
2.3 10
11
exp(23/T) 300
9
80 [Tokmakov et al., 1999]
k
3
C
2
+C
6
H
6
! C
8
H
5
+H 5.2 10
10
300 [Reisler et al., 1980]
k
4
C
2
H+C
4
H
6
! C
6
H
6
+H 3.0 10
10
104 296 [Nizamov and Leone, 2004]
k
5
C
2
H+C
6
H
6
! C
8
H
6
+H 9.1 10
10
T
0.18
105 298 [Goulay and Leone, 2006]
k
6a
C
2
H
3
+C
4
H
3
! C
6
H
6
2.2 10
15
T
3.75
exp(494/T) 200 1350 10 k
0
(CH
3
+CH
3
)[Wang et al., 2003]
4.8 10
10
exp(411/T) 700 1300 [Duran, 1988]
k
6b
C
2
H
3
+C
4
H
3
! C
2
H
4
+C
4
H
2
1.2 10
11
298 k(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
6c
C
2
H
3
+C
4
H
3
! C
2
H
2
+C
4
H
4
1.2 10
11
298 k(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
7a
C
2
H
3
+C
4
H
5
! C
6
H
8
2.2 10
15
T
3.75
exp(494/T) 200 1350 10 k
0
(CH
3
+CH
3
)[Wang et al., 2003]
1.2 10
10
298 k
1
(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
7b
C
2
H
3
+C
4
H
5
! C
2
H
4
+C
4
H
4
1.2 10
11
298 k(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
7c
C
2
H
3
+C
4
H
5
! C
2
H
2
+C
4
H
6
1.2 10
11
298 k(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
8a
C
2
H
3
+C
6
H
5
! C
8
H
8
2.2 10
15
T
3.75
exp(494/T) 200 1350 10 k
0
(CH
3
+CH
3
)[Wang et al., 2003]
1.2 10
10
298 k
1
(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
8b
C
2
H
3
+C
6
H
5
! C
2
H
4
+C
6
H
4
1.2 10
11
298 k(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
8c
C
2
H
3
+C
6
H
5
! C
2
H
2
+C
6
H
6
1.2 10
11
298 k(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
9a
C
2
H
5
+C
6
H
5
! C
8
H
10
2.2 10
15
T
3.75
exp(494/T) 200 1350 10 k
0
(CH
3
+CH
3
)[Wang et al., 2003]
1.2 10
10
298 k
1
(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
9b
C
2
H
5
+C
6
H
5
! C
2
H
6
+C
6
H
4
1.2 10
11
298 k(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
9c
C
2
H
5
+C
6
H
5
! C
2
H
4
+C
6
H
6
1.2 10
11
298 k(C
2
H
3
+C
2
H
3
)[Fahr et al., 1991]
k
10
C
3
H
3
+C
3
H
3
! C
6
H
6
2.2 10
15
T
3.75
exp(494/T) 200 1350 10 k
0
(CH
3
+CH
3
)[Wang et al., 2003]
4.0 10
11
295 [Fahr and Nayak, 2000]
k
11
C
6
H
5
+C
6
H
5
! C
12
H
10
2.2 10
15
T
3.75
exp(494/T) 200 1350 10 k
0
(CH
3
+CH
3
)[Wang et al., 2003]
2.3 10
11
exp(56/T) 295 [Park and Lin, 1997]
a
Plain: bimolecular rate constant (cm
3
s
1
) /Bold: trimolecular rate constant (cm
6
s
2
).
b
Temperature range for which the reaction has been studied (K).
E05007 VUITTON ET AL.: BENZENE ON TITAN
12 of 18
E05007
explicitly calculates the C
6
H
6
* density. The lifetime of
2 ms and the de-excitation coefficient (k
2
*) imply that
dissociation occurs above 350 km and stabilization below.
C
6
H
6
* could also react with other molecules (k
3
*) or
absorb a photon (J*), leading to different products from
single-photon absorption. We assume a reaction rate con-
stant k
3
*of5.5 10
10
cm
3
s
1
and a dissociation
coefficient J* equal to that of benzene. Estimations of the
efficiency of these processes at 900 and 200 km are
summarized in Table 4.
[
42] The main contribution to the photolysis rate comes
from the peak in the absorption cross section near 220 nm,
which is also the region used in the laboratory studies of
dissociation pathways. In the near ultraviolet, we assume
photodissociation products to be exclusively C
6
H
5
and H
[Kislov et al. , 2004]. At shorter wavelength (l < 135 nm),
ionization represents more than 50% of the total absorp-
tion [Person, 1965; Rennie et al., 1998] but both ion and
neutral products are ill-defined. The fact that the benzene
photoabsorption cross section extends to long wavelengths
coupled with the rapid increase in solar flux with increas-
ing waveleng th implies that 80 90% of the absorption
occurs at wavelengths above 180 nm. As a consequence,
we make the assumption that photodissociation products
are identical over the whole spectral range.
[
43] The fate of the C
6
H
5
produced by benzene photolysis
is to reform benzene through
C
6
H
5
þ H þ M ! C
6
H
6
þ Mk
1
ðÞ;
or to produce toluene through
C
6
H
5
þ CH
3
þ M ! C
7
H
8
þ Mk
2
ðÞ
:
[44] There is little kinetic data on the reaction of C
6
H
5
with other species. Since C
6
H
5
is a rather low energy radical
(DH
f
= 339 kJ/mol), we consider it unlikely that it could
react with stable species in the atmosphere, and hypothesize
that the other main losses are through radical-radical reac-
tions such as
C
6
H
5
þ C
2
H
3
þ M ! C
8
H
8
þ Mk
8a
ðÞ;
or
C
6
H
5
þ C
2
H
5
þ M ! C
8
H
10
þ Mk
9a
ðÞ:
[45] We assume that the rate for these reactions is 1.2
10
10
cm
3
s
1
, based on analogy with
C
2
H
3
þ C
2
H
3
þ M ! C
4
H
6
þ M:
4.2.2. Results
[
46] In Figure 11a, the vertical profiles of the principal
neutrals are displayed for a simulation using the solar flux,
chemical network and eddy diffusion coefficient stated
previously. The
40
Ar profile is compared to the INMS [Yelle
et al., 2008] and GCMS [Niemann et al., 2005] result s in
the upper and lower atmosphere, respectively. The C
2
H
2
,
C
2
H
6
and C
4
H
2
profiles are compared to the CIRS measure-
ments at 15°S[Vinatier et al., 2007]. The model matches
closely both the measured
40
Ar and C
2
H
6
mole fractions,
thus validating the choice of eddy diffusion coefficient. The
C
2
H
2
profile is in extremely good agreement with the
observations as well. The C
4
H
2
profile is a factor of 2 3
too high, but that is still a fairly good agreement considering
the numerous uncertainties in the C
4
H
2
formation and loss
processes [Vuitton et al., 2006b].
[
47] Th e vertical profil es of the aromatic speci es are
displayed in Figure 11b. Averages of the INMS C
6
H
x
mole
fractions over the 12 passes presented in Table 1 as well as the
C
6
H
6
mole fraction retrieved by CIRS at 15°S[Coustenis
et al., 2007] are plotted for comparison. Figure 11c shows
the vertical profiles of the principal radicals and Figure 12
the rates of the reactions that have a significant impact on
the benzene mole fraction. The calculated profiles repro-
duce the main characteristics of the measurements. Near
900 km, electron recombination of C
6
H
7
+
forms C
6
H
6
(a
1
),
which is rapidly depleted by photolysis (J
1
) to produce
C
6
H
5
. The C
6
H
5
density in the model at 1000 km is 2.0
10
4
cm
3
, corresponding to a mole fraction of 1.4 10
6
,
as displayed in Figure 11c. This is 2.5 times the benzene
density at the same altitude. The INMS observations reveal
a mole fraction of C
6
H
x
of 2 10
6
at 1000 km in good
agreement with the model results. Other abundant radicals
at 1000 km include hydrogen (H) and met hyl (CH
3
)
radicals. They are produced mainly through methane pho-
tolysis above 700 km.
[
48] As shown in Figure 12, photolysis is the main loss
for benzene around 900 km. C
6
H
5
, the main photodissoci-
ation product, diffuses downward and is converted to
heavier aromatics in the 350800 km region (k
2
,k
8a
,
k
9a
). This process is responsible for the rapid decrease in
benzene mole fraction below 900 km shown in Figure 11b.
As the pressure increases, stabilization of t he transient
benzene becomes more important and it becomes compa-
rable to its dissociation lifetime (k
1
*/k
2
* 1) at a density of
Table 4. Formation and Loss Processes of Hot Benzene C
6
H
6
*
R# Reaction Rate Constant
a
Lifetime (s)
Reference900 km 300 km
J
1
C
6
H
6
+hn ! C
6
H
6
* 2 10
5
5 10
5
[Koch and Otto, 1972; Pantos et al., 1978]
k
1
*C
6
H
6
* ! C
6
H
5
+H >5.0 10
2
<2 10
3
<2 10
3
[Tsai et al., 2000]
J* C
6
H
6
*+hn ! products 2 10
5
5 10
5
estimated equal to J
1
k
2
*C
6
H
6
*+N
2
! C
6
H
6
5.5 10
13
3 10
1
6 10
5
[Nakashima and Yoshihara, 1982]
k
3
*C
6
H
6
*+C
x
H
y
! products <5.5 10
10
>1 10
0
>3 10
6
estimate
b
a
Plain: bimolecular rate constant (cm
3
s
1
) /Bold: unimolecular rate constant (s
1
).
b
C
x
H
y
taken as CH
4
(2% N
2
). The lifetime obtained is a lower limit since C
6
H
6
* may only react with heavier, less abundant species.
E05007 VUITTON ET AL.: BENZENE ON TITAN
13 of 18
E05007
10
15
cm
3
(350 km). Thus in the lower atmosphere,
stabilization of the transient state is the dominan t process.
Moreover, at 350 km, the recombination of H and C
6
H
5
(k
1
)
becomes equal to the photolysis rate and destruction of
C
6
H
6
by photolysis stops. Below 300 km, C
6
H
6
is created
by recombination of C
2
H
3
and C
4
H
3
(k
6a
), and C
3
H
3
(k
10
),
producing the low altitude peak in C
6
H
6
density observed
by CIRS. The mole fraction in the lower stratosphere is
30 ppb, as shown in Figure 11b, which is consistent with
that inferred from the mid-IR spectroscopy.
[
49] Reactions k
6a
and k
10
contribute to most of the
benzene formation via neutral chemistry, with a total inte-
grated production rate of 4 10
6
cm
2
s
1
. Vuitton et al.
[2007] suggested that the reaction between C
2
H and C
4
H
6
(k
5
) could produce a significant amount of benzene.
Because the C
4
H
6
density is quite low, it appears that the
column-integrated value is only 1 10
4
cm
2
s
1
, 200 times
less than the other channels. The integrated benzene pro-
duction rate from the electron recombination of C
6
H
7
+
is
3 times larger than the total production rate from neutral
chemistry. Because of the general agreement with the
observations, we believe that the model results strongly
Figure 11. Modeled neutral vertical profiles and compar-
ison with observations. (a) Principal neutrals.
40
Ar: the
stratospheric measurement was obtained by GCMS
[Niemann et al., 2005] and thermospheric measure ments
by INMS [Yelle et al., 2008]. C
2
H
2
,C
2
H
6
and C
4
H
2
: CIRS
measurements at 15°S[Vinatier et al., 2007]. (b) Aromatic
species. The observed C
6
H
6
mole fraction in the strato-
sphere was retrieved by CIRS at 15°S[Coustenis et al.,
2007]. Data points represent averages over 12 Titan passes
of the C
6
H
x
density measured by INMS and presented in
Table 1. Error bars include statistic uncertainties but not
calibration uncertainties. ‘Aromatics refers to heavy
aromatic species that are not tracked in the model.
(c) Principal radicals.
Figure 12. Modeled reaction rates for some key reactions.
Labels refer to the reactions listed in Tables 2, 3 and 4.
E05007 VUITTON ET AL.: BENZENE ON TITAN
14 of 18
E05007
support th e hypoth esis that benze ne is c reated by ion
chemistry.
5. Discussion
[50] The results from the ion and neutral chemistry
models presented in the previous sections show that the
ionospheric production rate of C
6
H
6
is of the right order to
explain the INMS observat ions. These ar e 1D glob ally
averaged models and do not address possible variations of
C
6
H
6
with latitude or local time. The strong dependence of
both production and loss mechanisms on solar insulation
suggests that horizontal variations could be large, whereas
the results sh own i n Table 1 reveal that the measured
variations are mild. Three-dimensional simulations that
couple dynamics and chemistry are required to fully inves-
tigate horizontal variations but we can gain some insight
from the timescales for chemical and transport processes.
[
51] Three timescales are of interest, those for chemical
processes, horizontal transport, and vertical diffusion. The
chemical and diffusion time constants from our model are
plotted in Figure 13. The calculated photolysis rate indicates
that the chemical time constant in the upper thermosphere is
10
5
s. The diffusion time constant at 1300 km is two orders
of magnitude shorter than the time constant for photolysis; it
increases exponentially with decreasing altitude and
becomes equal to the photolysis time constant at an altitude
of 950 km. Thus chemistry is too slow to affect the altitude
distribution of benzene above 950 km and it should be close
to diffusive equilibrium, as calculated in the numerical
models and observed by INMS.
[
52] As shown in Figure 13, the diffusion time constant
and chemical time constant are both within a factor of
several of the diurnal time constant (equal to the length of a
day divided by 2p) in the 9001100 km region. This
implies that diurnal variations are possible but should not
be too extreme. The data shown in Table 1 are consistent
with this. The observed variation is a factor of several. C
6
H
6
is not absent on the night side, though the one measurement
on the dayside (T23) implies a larger than average mole
fraction. Since benzene is mostly formed on the dayside, our
night side results imply that it is transported to the night side
with no large loss.
[
53] The situa tion is more complicated than simple diur-
nal variability however. The dynamics of Titan’s upper
atmosphere appear to be quite vigorous [Mu¨ller-Wodarg et
al., 2008] with strong meridional as well as vertical winds.
The meridional winds will affect the latitudinal distribution
of C
6
H
6
and the vertical winds may alter both latitudinal
and local time distributions, depending on the precise
circulation pattern. Strong zonal winds will smooth diurnal
variations. Longitude variations are also possible if some
component of the circulation pattern is longitude dependent.
Further observations and interpretation with 3-D simulation
of dynamics and chemistry is needed to investigate these
questions. At the present time, we can only say that the
similarity in chemical, diffusive, and diurnal time constants
is grossly consistent with the observed mild variations in
abundance.
[
54] Wilson et al. [2003] performed a study dedicated to
the mechanisms for the formation of benzene in the atmo-
sphere of Titan. They considered ion-neutral reactions and
found that electron recombination of C
6
H
7
+
controls the
benzene production in the upper atmosphere. However,
the production rate at 1000 km is only 10
2
cm
3
s
1
,
one order of magnitude less than the production rate
calculated here. As a consequence, the model fails to
reproduce the observed benzene abundance in the thermo-
sphere [Wilson and Atreya, 2003]. It is difficult to track
down the discrepancy between their model and ours but the
fact that the major production reaction in the ionosphere
(k
0
10
) is missing in their model is probably one of the
reasons. Only 13 out of the 22 reactions identified here as
important for benzene formation (Table 2, Figure 8) are
present in their reaction list.
[
55] Lebonnois [2005] performed a comparative study of
benzene production in Titan and Jupiter atmosphere. His
model does not consider ion chemistry and the primary
mechanism respons ible for benzene formation is the recom-
bination of C
3
H
3
radicals (reaction k
10
in Table 3). In the
nominal simulation, the mole fraction of benzene reaches
10
6
at 1000 km, a factor of 200 higher than obtained
here with neutral chemistry alone. For reaction k
10
,
Lebonnois [2005] adopts a value of 1.2 10
10
cm
3
s
1
for k
1
[Morter et al., 1994] and 4.4 10
23
cm
6
s
2
for k
0
at 150 K (10 times the rate for CH
3
+CH
3
retrieved by
McPherson et al. [1983]). These rate constants are 3 and
80 times higher than the rate constants used here. As in our
model, benzene is efficiently photodissociated into C
6
H
5
radicals but in the model of Lebonnois [2005], the most
probable fate of C
6
H
5
is to form back benzene by reaction with
H atoms (reaction AR7 in his Table 1). As a consequence,
there is no efficient chemical loss for C
6
H
5
and benzene
accumulates in the atmosphere. As discussed in section 4.2,
even if these reactions have yet to be studied, it is very likely
that C
6
H
5
does react with radicals other than H and that
benzene is ultimately lost to heavier species. It follows that
the high benzene abundance found by Lebonnois [2005]
probably results from the inadequacy of the reaction list.
[
56] On Earth, benzene is a major component in flames,
where it appears to be associated with the formation of
PAHs and soot [Frenklach, 20 02]. Various mechanisms
Figure 13. Modeled chemical (solid line) and diffusion
(dashed line) time constants for C
6
H
6
. The dotted line
corresponds to Titan diurnal time constant.
E05007 VUITTON ET AL.: BENZENE ON TITAN
15 of 18
E05007
involving either neutral [Wang and Frenklach, 1994] or ion
chemistry [Calcote and Keil, 1990] have been proposed to
explain the formation of benzene and PAHs in combustion
processes. Building on these studies, Lebonnois [2005] and
Wilson and Atreya [2003 ] suggested that PAHs were a
major component of Titan’s haze. However, they extrapo-
late kinetic data obtained at high temperature and neglect
the fact that the neutral reactions involved have an activa-
tion barrier that cannot be overcome at cold temperature. As
a consequence, the neutral reactions leading to PAHs in
combustion processes will be inefficient on Titan. Instead,
we suggest that reactions of aromatic radicals with simpler
radicals (CH
3
,C
2
H
3
,C
2
H
5
, etc.) along with self-reactions of
aromatic radicals (C
6
H
5
+C
6
H
5
) produce heavier aromatics
that eventually condense and form haze that settles to the
surface. We calculate a total production rate of solid
aromatics of 10
15
gcm
2
s
1
. 2/3 of this mass flux
is traceable to benzene produced by ion chemistry. When
this process is not taken into consideration, the aromatic
production rate decreases by a factor of 3, as shown in
Figure 11b. Experimental studies on the reactivity of C
6
H
5
,
C
7
H
7
(benzyl) and other aromatic radicals are required to
better constrain formation processes. Also, as shown in
Table 4, reaction of hot benzene (k
3
*) is possible at lower
altitudes and some reactions could lead to the formation of
more complex aromatics. Finally, considering the high
density and variety of ions detected in the upper atmo-
sphere, it is also likely that there may be other pathways in
ion-neutral chemistry that could lead to aromatics.
[
57] Solid benzene has been tentatively identified on
Titan’s surface [Niemann et al., 2005; R.N. Clark, manuscript
in preparation, 2008]. We calculate a benzene condensation
rate of 1 10
6
cm
2
s
1
, about 1/10 of the benzene
production rate. Condensation at such a rate over the age of
the solar system corresponds to an average layer of only
20 cm of solid benzene on the surface of Titan at present.
This condensate will be mixed with other more abundant
deposits (for example, the acetylen e layer is predicted to be
125 times thicker), rendering its detection by remote sensing
more challenging. However, solid benzene could also be
formed on the surface by heterogeneous chemistry if
enough energy is available.
[
58] Measurements by the CAPS Electron Spectrometer
imply the existence in Titan’s upper atmosphere of various
negative ions ranging from 15 to 10,000 amu [Coates et
al., 2007]. No formation processes or attribution for these
ions have been proposed to date. Dissociative attachment
following low energy (9 eV) electron impact on benzene
produces C
6
H
5
[Fenzlaff and Illenberger, 1984], which has
a fairly high electron affinity (1.1 eV) [Gunion et al., 1992].
More generally, dissociative electron attachment of the type:
PAH þ e
< 15eVðÞ!PAH
þ H
is an efficient process with cross sections ranging from
10
17
to 10
14
cm
2
[Tobita et al., 1992]. As a consequence,
C
6
H
5
as well as PAH
.
are potential candidates for the
negative ions observed by ELS. Complex negative ion
chemistry models are necessary to further investigate this
possibility.
[
59] Benzene has been observed in other astronomical
objects as well. It is present in the atmosphere of Jupiter
[Kim et al., 1985], Saturn [Be´zard et al., 2001] and in the
proto-planetary nebula CRL 618 [Cernicharo et al., 2001].
Several photochemical models have tried to reproduce the
quantities of benzene observed on Jupiter and Saturn. At
low-latitude, recombination of propargyl (C
3
H
3
) radicals is
found to be responsible for most of the C
6
H
6
production. In
both planets, this reaction leads to abundance profiles that
are 50% weaker than needed to reproduce the benzene
observations [Moses et al., 2000, 2005]. Under Jupiter high-
latitude conditions where precipitation of energetic elec-
trons is important, benzene is produced, a s on Titan,
through successive ion-neutral reactions followed by elec-
tron recombination of C
6
H
7
+
[Wong et al., 2003]. Since the
benzene abundance in Jupiters auroral regions is much
higher than at low latitudes, this might indicate a possible
role of meridional transport [Lebonnois, 2005]. Also, the
models should use the two following reactions that have not
been included so far:
C
2
H þ C
4
H
6
! C
6
H
6
þ H; k
4
ðÞ
and
C
2
H
3
þ C
4
H
3
! C
6
H
6
; k
6a
ðÞ:
[60] Benzene has also been formed in laboratory simu-
lations of Titan’s chemistry. Imanaka and Smith [2007]
investigated the formation of complex species from a
N
2
/CH
4
gas mixture as a function of irradiation wavelengths
from 50 nm to 150 nm. The formation of benzene is
observed at wavelengths less than 80 nm, which corre-
sponds to the photoionization threshold of N
2
. This result is
qualitatively consistent with the benzene formation path-
ways presented on Figure 8. Above 80 nm, only methane
can be ionized and the formation of benzene requires the
presence of species such as hydrogen cyanide (HCN) and
diacetylene (C
4
H
2
). These are quite complex molecules that
are only present at low concentration in the laboratory gas
mixture, if present at all (see Figure 4 in the work of
Imanaka and Smith [2007]). However, if N
2
+
is present,
the formation of benzene only requires the presence of more
simple species such as acetylene (C
2
H
2
) and ethylene
(C
2
H
4
), which are readily available in the gas mixture.
Modeling of the laboratory experiment would be useful in
order to test this explanation in a more quantitative way.
6. Conclusions
[61] The results from analysis of the C
6
H
6
time profiles
combined with photochemical models of the atmosphere
are:
[
62] (1) The time dependence of the m/z = 77, 78 and 92
signals implies that benzene and toluene are mostly formed
in the INMS chamber through H + C
6
H
5
radicals and CH
3
+
C
6
H
5
recombination on the walls.
[
63] (2) The amplitude of the m/z = 77 and 78 signals
during T16 can be matched by a benzene mole fraction at
950 km of 1.3 10
6
.
[
64] (3) The globally integrated production rate of ben-
zene in the ionosphere, is 10
7
cm