Kinetics of the reaction of SH and SD with NO2

Article abstract of DOI:10.1021/jp053918r

The rate constants for the reactions of NO₂ with SH and SD were measured between 250 and 360 K to be 2.8 × 10^-11 exp{(270 ± 40)/T(K)} and 2.6 × 10^-11 exp{(285 ± 20)/T(K)} cm³ molecule^-1 s^-1, respectively. SH(SD) radicals were generated by pulsed laser photolysis of H₂S(D ₂S) or CH₃SH and detected via pulsed laser-induced fluorescence. The laser-induced fluorescence excitation spectrum of SH was found to be contaminated by the presence of the SO radical. This contamination is suggested as a possible reason for differences among some of the reported values of k₁ in the literature. The title reaction influences the atmospheric lifetime of the SH radical when NO₂ is greater than 100 pptv, but the revised value of k₁ does not significantly alter our current understanding of SH oxidation in the atmosphere.

Full text of DOI:10.1021/jp053918r

106
J. Phys. Chem. A 2006, 110, 106-113
Kinetics of the Reaction of SH and SD with NO2
Scott C. Herndon† and A. R. Ravishankara*
National Oceanic and Atmospheric Administration, Earth System Research Laboratory, Boulder, CO 80305,
Department of Chemistry and Biochemistry, UniVersity of Colorado, Boulder, CO 80309, and The CooperatiVe
Institute for Research in EnVironmental Sciences, UniVersity of Colorado, Boulder, CO 80309
ReceiVed: July 16, 2005; In Final Form: October 10, 2005
The rate constants for the reactions of NO2 with SH and SD were measured between 250 and 360 K to be
2.8 × 10-11 exp{(270 ( 40)/T(K)} and 2.6 × 10-11 exp{(285 ( 20)/T(K)} cm3 molecule-1 s-1, respectively.
SH(SD) radicals were generated by pulsed laser photolysis of H2S(D2S) or CH3SH and detected via pulsed
laser-induced fluorescence. The laser-induced fluorescence excitation spectrum of SH was found to be
contaminated by the presence of the SO radical. This contamination is suggested as a possible reason for
differences among some of the reported values of k1 in the literature. The title reaction influences the
atmospheric lifetime of the SH radical when NO2 is greater than 100 pptv, but the revised value of k1 does
not significantly alter our current understanding of SH oxidation in the atmosphere.
Introduction
by Wang et al.5 using a flow system. Stachnik and Molina took
great care to minimize secondary reactions in their system. The
lower value of Stachnik and Molina could be due to the
suppression of secondary reactions or due to the complexing
of SH with O2 to form HSOO. If HSOO has a lower reactivity
with NO2 than SH, the loss of SH via reaction 1 would be offset
by the dissociation of HSOO. The formation of HSOO has been
suggested6 on the basis of analogy with the reactivity of CH3-
SOO formed in the addition of CH3S with O27 and by ab initio
calculations.6
Hydrogen sulfide (H2S), emitted mostly by terrestrial sources,
is an important natural reduced sulfur species in the troposphere.
The tropospheric lifetime of H2S is determined by its reaction
with the OH radical (∼2.5 days), which leads to the formation
of the SH radical.1-3 Thermochemical considerations rule out
H atom abstraction from aliphatic hydrocarbons by SH. The
tropospheric loss of SH is governed by its reactions with O3
and NO2 and to a lesser extent NO. Accurate values of the rate
coefficients are needed to assess the lifetime of the SH radical
and to fully elucidate the mechanism for atmospheric oxidation
of H2S.
In this work, the rate coefficient for reaction 1 was measured
as a function of temperature. The rate coefficient for NO2
reaction with SD, the deuterated analogue of SH, was also
determined.
There are some differences between previously reported
values of the rate coefficients for the reactions of SH with NO2:
SD + NO2 f DSO + NO
M8 DSNO2
(2a)
(2b)
SH + NO2 f HSO + NO; k1a
M8 HSNO2; k1b
(1a)
(1b)
9
9
Reported values of k1 (k1 ) k1a + k1b) range from 2.5 × 10-11
to 12 × 10-11 cm3 molecule-1 s-1. The range of literature values
for k1 could be due to the experimental difficulties involved in
measuring this reaction rate coefficient. First, there is a dark
reaction (likely on the walls) between NO2 and H2S, the
commonly used precursor of SH. This potentially leads to
uncertainties in the concentration of NO2 in the reaction cell as
well as production of species that enhance the loss or secondary
production of SH. Second, SH reacts very rapidly with itself
and other free radicals, such as OH and HO2, requiring the use
of low concentrations of SH in kinetics studies. Third, in
photolysis systems, H atoms are unavoidably produced due to
H2S photodissociation. This could regenerate SH via the reaction
H + H2S f SH + H2 and lead to an apparent value of k1 that
is lower than the true value (see Results and Discussion for
more details). Stachnik and Molina4 used molecular oxygen to
scavenge H atoms and obtained a lower value than that reported
The differences between k1 and k2 are expected to be very small
because only a secondary kinetic isotope effect is anticipated.
The rate coefficient for reaction 2, k2, was measured as a further
check on k1.
During the course of our investigation of the SH detection
via laser-induced fluorescence, we found the observed laser-
R
induced fluorescence (LIF) spectrum near 323.7 nm, the Q21
and R1 band heads of SH, to be contaminated by a contribution
by another species, which we believe is SO. These were the
wavelengths used to monitor SH in some of the previous
determinations of k1. The possible impact of this spectral
contamination on the measured value of k1 was also investigated.
Experimental Description
The SH radicals were produced by pulsed 193 nm laser
photolysis of H2S or CH3SH and were detected by laser-induced
fluorescence. The temporal profiles of SH were measured in
the presence of a large excess of NO2, whose concentration in
the reactor was determined using flow and/or absorption
methods.
* Corresponding author. Present address: NOAA/ESRL R/CSD 2, 325
Broadway, Boulder, CO 80305. E-mail: a.r.ravishankara@noaa.gov.
† Present address: Aerodyne Research, Inc., Billerica, MA 01821.
10.1021/jp053918r CCC: $33.50 © 2006 American Chemical Society
Published on Web 12/09/2005

Kinetics of the Reaction of SH and SD with NO2
J. Phys. Chem. A, Vol. 110, No. 1, 2006 107
Figure 1. Schematic of the experimental apparatus used to measure k1 and k2. See text for details regarding the points marked A and B where NO2
was added.
Figure 1 shows a schematic of the apparatus employed to
measure k1 and k2. Dilute mixtures of the SH/SD precursor
(RSH/RSD, where R is CH3/CD3 or H/D) in N2 and/or He were
flowed through the cell. Varying amounts of NO2 were added
to this mixture just upstream of the reaction cell. The flow rates
of the carrier gases were measured using calibrated electronic
mass flow meters. The gas mixture passed through the reaction
cell with a linear flow velocity between 13 and 50 cm s-1. For
several experiments, O2 was also added to this gas mixture. The
total pressure in the reactor was between 30 and 120 Torr.
Pressure was measured using a calibrated capacitance manom-
eter. A heated or cooled liquid from a temperature-controlled
circulator was flowed through the jacket around the reaction
cell to maintain the reactor at any desired temperature. The
temperature of the gas mixture in the middle of the reactor was
measured using a retractable calibrated chromel-alumel ther-
mocouple. The temperature was varied from 240 to 360 K.
Ultrahigh purity grade N2, He, and O2 (purities of >99.9995%,
>99.9%, and >99.99%, respectively) as well as the photolytic
precursors H2S, D2S, and CH3SH (manufacturer’s stated purities
of >99% each) from Scott Specialty Gases were used without
further purification. NO2 was prepared by reacting purified NO
with excess O2, which was introduced through a molecular sieve
trap maintained at dry ice temperatures. NO2 was collected in
a trap cooled to dry ice temperature. It was purified by trap-
to-trap distillation in the presence of O2 until a pure white solid
remained. The major impurities in our sample of NO2 were NO
and O2, each at a level of <0.1%. The rate coefficients4,8 for
the reactions of SH/SD with NO and O2 are respectively <2 ×
10-12 and 4 × 10-19 cm3 molecule-1 s-1. At the concentrations
of NO2 used in our experiments, the NO and O2 impurities in
NO2 were too small to make significant contributions to the
measured SH/SD loss rate coefficients. The ACS spectropho-
tometric grade dimethyl sulfoxide (>99.9% pure), used to
generate the excitation spectrum of SO, was purchased from
Aldrich Chemicals and used without further purification.
Over the course of the experiments, we became suspicious
that H2S and NO2 were reacting on the surfaces of the gas
manifold because we saw a feature in the LIF spectrum that
did not belong to SH. Therefore, we took a great deal of care
to minimize the reaction of H2S with NO2 and also used
CH3SH as a source of SH in some experiments. These factors
are discussed later. NO2 concentrations were determined by
absorption at 365 nm. Two different schemes were employed
for the reactant and photolytic precursor; they are described later.
Photolysis of H2S and CH3SH at 193 nm was used to generate
SH radicals. The fluence of the photolysis laser was measured
using a calibrated thermopile detector. For all experiments, the
initial radical concentration was estimated by using the known
absorption cross section of H2S/CH3SH and the quantum yield
for SH production, along with the measured photolysis fluence
and the concentration of H2S/CH3SH. In all cases, the initial
concentration of SH, [SH]0, was less than 3 × 1012 molecules
cm-3, and typically 1 × 1011 cm-3. The higher concentrations
(1012 cm-3) were used when SH was detected using weaker
spectral features. The absorption cross sections at 193 nm of
H2S, D2S, and CH3SH, used to estimate [SH]0 and [SD]0, were
6.4 × 10-18, 7.4 × 10-18, and 1.86 × 10-18 cm2, respectively.9-13

108 J. Phys. Chem. A, Vol. 110, No. 1, 2006
Herndon and Ravishankara
limited its detection sensitivity. A benefit of the short lifetime
was that the effect of fluorescence quenching by the bath gases
(such as N2) was essentially negligible and the detection
sensitivity was nearly independent of pressure.
In kinetics measurements, the probe laser was tuned to a
specific feature in the LIF spectrum and left on that feature while
the delay time between the photolysis and probe lasers was
varied to obtain a temporal profile. All temporal profiles of SH
and SD were measured under pseudo-first-order conditions. The
temporal profiles of the LIF signal followed equation I (when
SH detection was not contaminated by an interfering species,
as discussed later):
I
St ) S0e-k t
(I)
where St and S0 are LIF signals at time t and zero and kI is the
first-order rate coefficient for the loss of SH or SD. Some typical
temporal profiles of both SH that is not influenced by interfer-
ences noted later and SD are shown in Figure 2. The bimolecular
rate constants, k1 and k2, were determined using the following
analysis of the observed temporal profiles. To avoid redundancy,
the determination of k1 is detailed below, but identical proce-
dures were used to obtain k2. LIF signals at various reaction
times were fit to eq II using a linear least-squares routine.
ln(St) ) ln(S0) - kIt
(II)
Figure 2. Temporal profiles of SH (a) and SD (b) at room tempera-
tures. These figures show the fluorescence of SH and SD as a function
of delay time.
This first-order rate coefficient is attributed to the sum of the
following reactions that contribute to the loss of SH: reaction
with NO2, self-reaction, reaction with any impurities in the
reactor, diffusion out of the reaction zone, and reaction with
other species generated via laser photolysis.
The quantum yield of SH from the 193 nm photolysis of H2S
is assumed to be 1.13 We measured the quantum yield for SH
in CH3SH photolysis (at 193 nm) to be 0.5 ( 0.1. This value is
consistent with the quantum yield for CH3S from CH3SH
photolysis (0.48 ( 0.05).11
SH + NO2 f products
SH + SH f products
(1)
(3)
(4)
The temporal profiles of SH and SD were monitored by
pulsed laser-induced fluorescence (LIF). The output from an
excimer pumped dye laser was frequency doubled to obtain
wavelengths of 320 to 330 nm. The dye laser was scanned to
pump the individual rotational lines of the (0,0) band of the
(A2Σ, V′ ) 0) r (X2Π, V′′ ) 0) transition of the SH or SD
radical. The probe beam had a pulse width of ∼15 ns (fwhm)
and a line width of ∼0.005 nm (0.1 cm-1). The probe beam
was passed through the reaction cell to intersect the pho-
tolysis beam; the volume created by the intersection of the two
beams is referred to here as the reaction zone. The fluorescence
of the thiol radical corresponding to the (A2Σ, V′ ) 0) f (X2Π,
V′′ ) 1) transitions was collected along the third orthogonal
axis using a set of lenses. The fluorescence was passed through
a band-pass filter (355 ( 12 and 345 ( 10 nm for SH and SD,
respectively) before it was imaged onto a photomultiplier tube
(PMT). The output of the PMT was fed into a gated charge
integrator. The integrated signal was sent to a personal computer
via an A/D converter for signal averaging. Typically, signals
from 100 laser pulses were averaged at each reaction time.
The detection sensitivities of both SH and SD were estimated
using the initial signal level (temporal profile was extrapolated
back to t ) 0), the measured photolysis fluence, and the
measured precursor concentration. In 100 Torr of N2, the
detection limit for SH was <1 × 109 molecules cm-3, while
that for SD was <5 × 107 molecules cm-3; the better sensitivity
for SD is due to the much longer fluorescence lifetime of SD
(τ ) 260 ns14), which predissociates much less than SH. The
A2Σ+ state of SH is highly predissociative with a radiative
lifetime of only 3 ns14 for V′ ) 0. This short lifetime ultimately
SH + impurities f products
SH f diffusion from reaction zone
SH + X f products
(5)
(6)
X in reaction 6 is a photoproduct (e.g., O(3P) and NO) produced
in the system whose concentration increases proportionately with
that of NO2. Loss of SH via processes 3-5 is represented as
the first-order rate coefficient kd. Note that reaction 3 is second
order in SH; however, over the time scales of the measurements
and the initial SH concentrations used in our experiments, we
approximate this as a first-order process. The overall first-order
SH decay rate constant is described by
kI ) kd + k′1 + k′X ) kd + (k′X + k1) × [NO2] (III)
The first-order loss rate coefficient kI was measured as a function
of NO2 concentration. Typically, NO2 was varied between 0.5
× 1013 and 20 × 1013 molecules cm-3. kd was assumed to be
constant, while NO2 was varied. The rate coefficient (k1 + kX)
was derived from a linear least-squares fit of the measured
values of kI at various concentrations of NO2 to eq III. In Figure
3, kI is plotted against NO2 under several experimental condi-
tions, for both SH and SD at room temperature. The measured
value of k1 + kX was independent of photolysis fluence, and

Kinetics of the Reaction of SH and SD with NO2
J. Phys. Chem. A, Vol. 110, No. 1, 2006 109
Figure 3. Plots of measured values of kI for the removal of SH (lower
line) and SD (upper line) versus NO2 concentration at room temperature.
The diamonds representing k2 (upper line) have been offset by 104 s-1
for visual clarity. The points in the lower curve were measured under
different concentrations of O2 in the reaction cell and different photolysis
fluences: (a) In 80 Torr of O2; circles ∼1 mJ pulse-1 cm-2; squares
∼0.1 mJ pulse-1 cm-2; diamonds ∼2 mJ pulse-1 cm-2. (b) In the
absence of O2; triangles. Each of these measurements was carried out
by either properly accounting for the SO interference beneath the SH
LIF transition or in the absence of such an interference, as described
in the text. The figure shows that k1 is approximately equal to k2.
Figure 4. Excitation spectra of SH and SD. The two traces shown are
the thermal excitation spectra of SH (upper trace) and SD (lower trace).
The data in these spectra were taken at a delay time of 100 µs in a
total pressure of 100 Torr of N2. The photolysis of H2S and D2S at 193
nm generated SH and SD, respectively. The concentrations of precursor
and photolysis fluence were such that [SH] ∼ 3 × 1012 molecules cm-3
and [SD] ∼ 5 × 109 molecules cm-3
.
therefore, [X] is assumed to be negligible. The measured value
of the slope of the plots, as depicted in Figure 3, is therefore
only k1.
Results and Discussion
Different types of experiments were carried out during this
study. The results of each type of study are described and
discussed together. First, the excitation spectra of SH and SD,
along with the interference in the detection of SH, is described.
Second, the experimental methods used to minimize the
influence of the interference in deriving kinetics data are
presented. Finally, the influences of secondary reactions on the
measured values of k1 and k2 are considered and our values of
k1 and k2 are compared with those from previous reports.
Excitation Spectra of SH and SD. The excitation spectra
of SH and SD are shown in Figure 4. Each of the six rotational
Figure 5. Spectra obtained from the photolysis of DMSO and mixtures
of NO2 and H2S. The spectrum marked “a” shows the result of the
photolysis of dimethyl sulfoxide at 193 nm without any change in the
instrument settings optimized for detection of the SH radical. Trace a
is the excitation spectrum of the SO radical. The spectrum “b” shows
the excitation spectra when NO2 and H2S are photolyzed together. The
spectrum “c” shows the excitation spectrum taken after cleaning the
glassware involved in the instrument but prior to the introduction of
NO2 to the system.
2
branches anticipated for the (2Σ r Π) transitions are present
in both spectra. The line positions are consistent with other
published spectra.10,15 The relative intensities of the rovibronic
lines in this spectrum show that SH/SD is rotationally thermal-
ized. This indicates that a delay time of 100 µs in ∼100 Torr
of N2 was sufficient for rotational thermalization. The blue shift
of the SD band heads is expected on the basis of the potential
energy surface of the 2Σ state. Also, as expected, the centrifugal
distortion effect on the rotational line spacing is more apparent
in SD than in SH.
under some conditions used in this work as well as in the study
of Black.16 Black noted unusual behavior of the “SH signal”
level when NO2 and H2S were mixed but did not investigate it
further. We found that qualitatively the magnitude of this feature
was greatly diminished when the glassware in the system was
cleaned. Furthermore, this feature was visibly enhanced by the
presence of NO2. In Figure 5, the scale of the excitation spectra
is expanded in wavelength to clearly show this feature. In the
lowest panel, part c, a “clean” excitation spectrum obtained in
the absence of NO2 is shown. In the middle panel, part b,
excitation spectra obtained in varying amounts of NO2 are
shown. The feature is clearly visible, and its magnitude increases
with the NO2 concentration ((0.6-3.8) × 1014 cm-3). Note that,
in one of the traces, the photolysis laser was temporarily blocked
while the excitation wavelength continued to scan. The signal
clearly returns to zero. When the photolysis laser was unblocked,
Typically, when measuring rate constants by laser-induced
fluorescence, the excitation wavelength is set to match the
strongest feature in the spectrum and is not changed during the
kinetics experiments. In SH, the transition that led to high
R
Q
detection sensitivity was the beginning of the Q21 and R21
branches at 323.7 nm. When H2S and NO2 were flowed together
and photolyzed, a fluorescence feature emerged near 323.4 nm
that is not due to SH. The broad feature is inconsistent with the
anticipated excitation spectrum of SH, a diatomic molecule with
a rotational constant greater than 8 cm-1. Also, it is not present
in the LIF spectrum of SH reported by Weiner et al.,10 who
generated SH in the absence of NO2. This feature was present

110 J. Phys. Chem. A, Vol. 110, No. 1, 2006
TABLE 1: Experimental Results (k1 and k2)
Herndon and Ravishankara
concn
(molecules (mJ cm-2 (molecules
cm-3 pulse-1 cm-3
photolysis
[SH]
LIF
total
[NO2]
ki ( σ
LIF
line
wavelengtha [O2] pressure NO2
range (1013
# of
k′
temp
(10-11 cm3
source
H2S
)
)
)
(nm)
(Torr) (Torr) methodb molecules cm-3
)
(K) molecule-1 s-1
)
9 × 1013 0.4
CH3SH 5 × 1014 0.3
2 × 1011 R1
7 × 1011 P3/2
3 × 1011 R1
323.7
323.0
323.7
323.0
323.7
323.7
323.7
323.7
323.7
323.7
323.7
323.7
323.0
323.0
323.0
322.8
322.8
322.8
322.8
322.8
322.8
322.8
83
91
85
92
98
99
85
70
50
91
100
99
85
87
88
56
57
57
87
60
70
68
flow
abs
0.3-21
3-22
0.5-15
4-26
15
9
5
c
248
252
253
254
263
272
295
296
296
297
297
297
317
335
358
246
256
277
297
298
329
359
8.1 ( 0.4
8.1 ( 0.5
8.6 ( 0.7
8.6 ( 0.6
7.6 ( 0.9
7.8 ( 0.7
7.0 ( 0.3
7.5 ( 0.6
7.1 ( 0.5
7.4 ( 0.8
4.9 ( 0.5
7.2 ( 0.3
6.7 ( 0.8
6.3 ( 0.7
5.9 ( 0.7
8.4 ( 0.3
7.8 ( 0.4
7.1 ( 0.3
6.3 ( 0.3
6.7 ( 0.3
6.7 ( 0.4
5.5 ( 0.4
CH3SH 3 × 1014 1.3
80
flow
abs
CH3SH 3 × 1014
2
1 × 1012 P3/2
8 × 1010 R1
>1
5
CH3SH 1 × 1014 0.05-1.1
flow
flow
flow
flow
abs
0.3-28
0.25-26
2.5-25
0.7-11
0.2-12
0.7-19
0.9-6
0.5-32
0.4-27
3-17
12
10
27
7
9
9
CH3SH 2 × 1014 0.1
4 × 1010 R1
H2S
H2S
H2S
H2S
H2Sd
7 × 1013
*
<1 × 1012 R1
c
c
c
c
d
1 × 1014 0.3
4 × 1013 1.5
1 × 1014 0.3
8 × 1013 0.15
2 × 1011 R1
52
4 × 1011 R1
2 × 1011 R1
flow
flow
flow
abs
abs
abs
8 × 1010 R1
4
CH3SH 2 × 1014 0.1
CH3SH 1 × 1015 1.2
3 × 1010 R1
12
13
8
5
9
7
10
9
7
5
2 × 1012 P3/2
1 × 1012 P3/2
3 × 1012 P3/2
2 × 1010 R1(D2S)
3 × 1010 R1(D2S)
2 × 1010 R1(D2S)
1 × 1010 R1(D2S)
3 × 1010 R1(D2S)
6 × 1010 R1(D2S)
4 × 1010 R1(D2S)
>0.013
>0.14
0.5
CH3SH 7 × 1014
CH3SH 9 × 1014
1
2
1-23
D2S
D2S
D2S
D2S
D2S
D2S
D2S
3 × 1013 0.4
3 × 1013 0.5
3 × 1013 0.3
8 × 1013 0.1
2 × 1014 0.08
8 × 1013 0.4
7 × 1013 0.3
10
10
abs
abs
5-35
5-40
10
>0.05
10
abs
abs
abs
6-32
3-19
4-22
10
10
abs
abs
4-28
3-26
6
a The approximate wavelength of the LIF transition is given in this column. b This column specifies the method used to determine [NO2]. “Flow”
implies the concentration was determined through a measurement of the standard mass flow rate; “abs” implies the concentration was determined
through an absorption measurement. See text for additional details. c This measurement has accounted for the SO interference in the detection of
SH by measuring the SO signal at long reaction time. See text for additional details. d This measurement did not account for the SO interference
in detecting SH. See text for details. This measurement is not included in the calculation of our recommended value of k1.
the “interference” returned. This sequence of experiments
provides further proof that the feature is due to laser-induced
fluorescence of some species generated upon photolysis either
due to photodissociation of a species other than H2S or a rapid
reaction of a photolytically produced molecule. Furthermore,
the prompt production of the signal strongly suggests that it is
formed by the photolysis of a species produced via heteroge-
neous reaction between H2S and NO2. Note that when H2S (SH
precursor) was turned off, the feature was greatly diminished,
even in the presence of NO2. The top panel, spectrum a, shows
the laser-induced fluorescence spectrum of SO (∼2 × 1012
cm-3) generated through the photolysis of dimethyl sulfoxide
((CH3)2SO, DMSO) at 193 nm.17 The concentration of SO was
estimated using the known UV cross section of DMSO, the mea-
sured laser fluence, and an assumed quantum yield of unity for
SO in DMSO photolysis. The similarities between the feature
of SO and the interference suggests that SO is being generated
and detected when H2S and NO2 are mixed and photolyzed.
Accuracy of the Measured Values of k1 and k2. When
measuring temporal profiles of a reacting species, the probe laser
detecting the species is typically tuned onto the signal peak and
parked there when the temporal profiles are measured. In our
experiments here, such a procedure would measure not only
SH but also the interference “underneath” the SH signal.
Therefore, to minimize the contribution of the interference to
the measured value of k1, we employed two approaches. In the
first, we tuned the probe laser to one of the rotational transitions
where the interference is not present. In the second, we
subtracted out the contribution of the interference by detecting
it at a long reaction time (five lifetimes of SH with respect to
reaction 1) and assuming that the abundance of interference did
not change during the course of SH reaction. The signal-to-
noise ratio was lower in the first method, but no assumptions
are made about the temporal behavior of the interference. The
signal-to-noise ratio was larger in the second method, but it
relied on the concentration of the interference species being
constant. Table 1 summarizes the results of these experiments
and notes the transition used. Clearly, both methods yielded
essentially the same results, confirming that the contribution of
the interference could be minimized. Further, the table clearly
shows that the measured value of k1 is ∼20% lower when we
did not account for the contribution due to the interference which
we have identified to be SO and fitted the observed temporal
profile to eq I.
The rate coefficients k1 and k2 were measured at various
temperatures, and the obtained values are also listed in Table
1. Figures 6 and 7 shows our measured values of k1 and k2,
respectively as a function of temperature, along with those from
previous publications. Figure 6 also shows the lower value of
k1 at 298 K that we determined when we did not account for
the SO interference. Clearly, when the interference is not
accounted for, the measurement of k1 is lower by 20%.
The SD line positions are blue shifted relative to those of
SH by ∼1 nm; this shift is sufficient to completely avoid
detection of the SO interference. The detection sensitivity for
SD was also about 2 orders of magnitude better than that for
SH, and therefore, any possible contribution by the interfering
species was further reduced. Thus, it was relatively straight-
forward to measure k2. The measured values of k2 are also given
in Table 1. As seen in the table, within the precision of these
experiments, the measured values of k1 and k2 are the same.
In some of the experiments, NO2 was added to the reaction
mixture prior to the main flows entering the dilution volume
(added at “A” in Figure 1). In other experiments, the flow from
the NO2 bulb was diluted with N2 and flowed through an
absorption cell where the concentration of NO2 was measured.
(The absorption cross section used for NO2 at 365 nm (in reality,
two Hg lines at 365.02 and 365.48 nm) was 5.75 × 10-19 cm2.18)
The flow out of the absorption cell was mixed into the main
flow very close to the reaction cell (added at “B” in Figure 1).
This arrangement minimized the contact time between NO2 and
H2S and, visibly, reduced the SO interference. Tests carried out
using Fourier transform infrared (FTIR) spectroscopy indi-
cated that NO2 did react with H2S, presumably on surfaces. A

Kinetics of the Reaction of SH and SD with NO2
J. Phys. Chem. A, Vol. 110, No. 1, 2006 111
Figure 7. Arrhenius plot showing the measurements of k2 vs 1000/T.
The filled circles are the results of this work with D2S photolysis as
the SD source. See Table 1 for additional details regarding the filled
circles in this figure. Other measurements also depicted are the
following: solid triangle, Wang et al.; open diamonds, Fenter and
Anderson.
Figure 6. Arrhenius plot showing the measured values of k1 vs
1000/T. The solid points are from this work; the squares were measured
with CH3SH photolysis as the SH source. The solid triangles used H2S
as the SH source. Studies of previous data from Wang et al. on k1 are
shown as open squares. The 298 K data from previous studies are the
following: open triangle, Schoenle et al. (corrected by Schindler and
Benter); open diamond, Stachnik and Molina; ×, Black; open circle,
Friedl et al.; bow tie, Bulatov et al.
at the typical concentrations of H2S, the first-order rate constant
for SH production was 500 s-1. Under these conditions, with a
constant [H2S], a plot of the measured values of kI versus NO2
would have a negative intercept due to the regeneration of SH.
Such negative intercepts were observed in the experiments where
only H2S is used as the SH source. Large concentrations of O2
(>3 × 1017 cm-3) suppressed regeneration because H atoms
were preferentially converted to HO2 and minimized OH pro-
duction via reaction 9. Negative intercepts were not observed
in the presence of large concentrations of O2. When SH was
generated via photolysis of CH3SH, regeneration of SH was
avoided. The value of k1 measured using CH3SH photolysis
agreed with that determined using H2S photolysis in the pres-
ence of O2. The values of k1 reported here were obtained by
either using H2S photolysis in the presence of O2, accounting
for the interference by subtracting the residual signal at long
reaction times, or using CH3SH photolysis as the source of SH
(Table 1).
For the SD experiments, the secondary chemistry concerns
are analogous to those listed above. In this work, all of the
experiments were conducted in the presence of ∼10 Torr of
O2, except for one measurement at room temperature. The rate
constant measured in the absence of O2 was slightly less than
that measured with O2; however, it was essentially the same
within our experimental precision. Since SO was not detected
at the wavelength used for SD detection, we did not have to
sulfoxide, however, was not conclusively identified as a prod-
uct of such a reaction. We did not pursue further the pos-
sible reaction of H2S with NO2. All of the evidence sug-
gested that the photolysis of the product of a reaction of H2S
with NO2 in our system was the source of the interference
and that the interference was most likely SO. We could avoid
the effect of the interference as discussed above when determin-
ing k1.
Possible Influences of Secondary Reactions on the Mea-
sured Values of k1 and k2. When H2S/D2S is used as the
SH/SD source, secondary reactions can potentially regenerate
SH/SD on time scales of its loss due to reactions 1 and 2. The
following reactions, depicted for the case of SH, are involved
in determining the extent SH regeneration.
H2S + hν(193 nm) f H + SH
H2S + H f H2 + SH
(7)
(8)
H + NO2 f NO + OH
OH + H2S f H2O + SH
H + O2 + M f HO2 + M
OH + NO2 f HNO3 + M
(9)
(10)
(11)
(12)
account for the formation of SO.
The absorption cross section of NO2
20,21
In the absence of O2, reactions 7, 9, and 10 lead to SH
regeneration. Although H atoms can react with H2S directly to
form SH and H2, this rate coefficient is rather small (8.2 ×
10-13 cm3 molecule-1 s-1 at 298 K) and, hence, H atom removal
by this reaction was negligible compared to that by NO2 and
O2 because the concentration of H2S used in our studies was
very small. In these experiments, as explained earlier, the NO2
concentration was varied such that the lifetimes of H atoms with
respect to loss due to reaction 9 varied from 33 to 2000 µs.19
At room temperature, when 10-0.5 Torr of O2 was present,
the lifetimes of H atoms with respect to loss via O2 association
with O2 were 10-150 µs.19 The rate constant for reaction 10 at
298 K is ∼5 × 10-12 cm3 molecule-1 s-1 19 which means that,
at 193 nm is less
than 5 × 10-19 cm2. At the photolysis fluences and the
concentrations of NO2 employed in our experiments, at most,
1 × 1011 molecules cm-3 of O(1D) and/or O(3P) could be
formed. At the total pressures of N2 used in these experiments,
all O(1D) is expected to be quenched to O(3P) within 1 µs.19
Because the reaction of O(3P) with H2S is rather slow, the
possible regeneration of SH would be at most 2 s-1. The reaction
of SH with O atoms could contribute at most 10 s-1 toward the
SH loss rate constant. These effects are suppressed in the
presence of NO2, since O atoms would react with NO2 with a
pseudo-first-order loss rate constant of 200-2400 s-1. The
presence of ∼10 Torr of O2 further enhances the removal of O

112 J. Phys. Chem. A, Vol. 110, No. 1, 2006
TABLE 2: Summary of k1 and k2 Determinations
Herndon and Ravishankara
reference
reactor pressure
(Torr)
species
monitored
k1 or k2 at 298 Ka
Ea/R
method
(10-11 cm3 molecule-1 s-1
)
(K-1
)
k1
FP-LIF
30-300 He
100 Ar
SH
HSO
SH
SH
SH
SH
SH
3.5 ( 0.4
2.4 ( 0.2
4.8 ( 1.0
3.0 ( 0.8
8.6 ( 0.9
6.7 ( 1.0
7.0 ( 0.8
k2
Black16
FP-ICLA
FP-LPA
DF-LIF
DF-MS
DF-LMR
FP-LIF
Bulatov et al.25
Stachnik and Molina4
Friedl et al.24
Schoenle et al.b 23
Wang et al.5
30-730 O2/N2
2-8 He
2-5 He
1 He
-240 ( 50
-270 ( 40
60-130 O2/N2
This Work
DF-LIF
DF-LMR
FP-LIF
2 He
1 He
70 N2
SD
SD
SD
6.8 ( 0.6
7.3 ( 1.1
6.6 ( 0.5
-210 ( 70
-279 ( 20
Fenter and Anderson22
Wang et al.5
This Work
a The reported errors are those from the reports and may be 1σ or 2σ. Our quoted errors are at the 2σ level and included estimated systematic
errors. b The number quoted here is the value reported by Schindler and Benter in an erratum, correcting their previously reported values in Schoenle
et al. A weighted (according to their reported uncertainty in the measured first-order rate coefficient) linear least-squares analysis of the data
reported by Schoenle et al. and correcting for the pressure measurement error noted in the erratum (Schindler and Benter) yields k1 ) (7.1 ( 0.6)
× 10-11 cm3 molecule-1 s-1. FP, flash photolysis; DF, discharge flow; LIF, laser-induced fluorescence; ICLA, intracavity laser absorption; LPA,
long path absorption; MS, mass spectrometry; LMR, laser magnetic resonance.
atoms. Thus, the reactions of O atoms should not lead to errone-
ous values of k1 or k2. Variation of the photolysis laser fluence
by a factor greater than 2 showed no change in the measured
values of k1 and k2 (Table 1), confirming the negligible
contribution of the O(3P) + H2S reaction to SH regeneration.
Photolysis of H2S at 193 nm can yield vibrationally excited
SH radicals (up to V′′ ) 12). We estimated the quantum yield
of SH(V′′ ) 0) by following the temporal profile of SH(V′′ )
0) at short reaction times to be ∼70% of the total SH produced
in the 193 nm photolysis of H2S; this estimate agrees with the
detailed vibrational distribution of SH reported by Continetti
et al.13 We also observed that the vibrationally excited SH was
quenched to the ground state by 100 Torr of N2 with a time
constant of ∼15 µs. This observation leads to a phenomenologi-
cal rate coefficient of ∼2 × 10-14 cm3 molecule-1 s-1 for the
quenching of higher vibrational levels to the ground state.
Therefore, we always employed at least 30 Torr of N2 as the
bath gas and measured the temporal profiles of SH and SD after
∼60 µs while determining k1 and k2. The rotational distribution
of photolytically produced SH was very rapidly thermalized,
as shown by the invariance of the peak intensities in the exci-
tation spectra taken at different delay times between the pho-
tolysis and probe laser. The measured values of k1 and, by anal-
ogy, k2 were not affected by vibrational relaxation of SH/SD
during the course of SH/SD removal via reactions 1 and 2.
A summary of all of the variations in experimental parameters
used in this work is shown in Table 1. They included variations
in photolysis fluence and photolyte concentration, use of a
different photolytic precursor, addition of O2, and use of
different spectral features for SH detection. The measured values
of k1 and k2 did not change with the experimental precision.
Therefore, we believe that our measured values of k1 and k2
are accurate. The average value of k1 at 298 K measured here
is (7.24 ( 0.62) × 10-14 cm3 molecule-1 s-1, where the quoted
error bars are 2σ and include an estimated uncertainty of 8% in
the determination of NO2 concentration.
In the above expression, the quoted uncertainties are 2σ. The
uncertainty in A was obtained by the expression σA ) AσlnA
and by adding in quadrature an estimated systematic error of
8% due to the possible errors in the measurement of the NO2
concentration in the reactor. A similar analysis of the data for
k2 yields
k2 ) (2.60 ( 0.50) ×
10-11 exp[(285 ( 40)/T] cm3 molecule-1 s-1
The measured values of k1 and k2 are essentially the same
because the reaction involves only a secondary kinetic isotope
effect. The H or D atom is essentially a spectator if the reaction
proceeds via abstraction and should not influence the reaction
if it is an addition process. Therefore, the measured value of k2
is a good indicator of the value of k1.
Discussion of This Work in Relation to Literature Values.
When measuring rate constants, it is typically convenient to
work under pseudo-first-order conditions. The strength of this
method lies in the relative ease of measuring a purely expo-
nential decay (or rise) of the reactant (or product) and the ability
to ascribe any deviation from exponential behavior to systematic
causes. The weakness is that any systematic error in the deter-
mination of the concentration of NO2 (in this work) translates
into a systematic bias in the determined k1 value. Regarding
the measurement of NO2, there are some issues of note: the
photolability of NO2, the dimerization of NO2 (where mano-
metric preparation of dilute NO2 or temperature dependent
studies are concerned), and the “dark” reactivity of NO2 itself.
Due to these considerations, measuring [NO2] itself can be
difficult and may lend to the range of measured values shown
in Table 2.
Depending on which type of experiment was being performed,
that is, SH was generated photolytically or via discharge flow
chemistry, reactions 7 and 8 act as SH sources. Reactions 8
and 10 regenerate SH, potentially hampering the determination
of k1. To suppress these regeneration sources, the addition of
sufficient O2 was employed4 with the constraint k11[O2][M] .
k9[NO2]. It has been proposed that O2 could “add” to SH (similar
to the addition of O2 to CH3S6,7). The effect of this adduct would
be to depress the measurement of k1. The Stachnik and Molina4
result is ∼20% less than the result of Wang et al.,5 which is in
accord with the idea that the adduct could be reducing the
measured value of k1. Work in this lab2 has determined the
Figure 6 shows our measured values of k1 in an Arrhenius
form, that is, as a plot of k1 on a logarithmic scale versus 1/T.
These results indicate a negative temperature dependence. A
fit of our data to the expression ln(k) ) ln(A) - Ea/(RT) yields
the following expression:
k1 ) (2.87 ( 0.50) ×
10-11 exp[(270 ( 50)/T] cm3 molecule-1 s-1

Kinetics of the Reaction of SH and SD with NO2
J. Phys. Chem. A, Vol. 110, No. 1, 2006 113
HSOO adduct to be very weak and unlikely to affect the
room temperature measurement of k1. (The upper limit for
∆0H298(association) for the formation of HSOO from HS and
O2, determined in this lab, is ∼6.5 kcal mol-1 and will be
published in a future paper.)
product of the atmospheric oxidation of SH by both O3 and
NO2. A SH radical in the remote clean troposphere has a lifetime
of ∼0.5 s (assuming 20 ppb O3 and 0.1 ppb NO2). This lifetime
assumes that HSO, formed in the reaction of SH with O3, will
not react further with ozone to regenerate SH. In regions with
higher NOx, the lifetime of SH decreases dramatically because
k1 ∼ 20 × k(O3 + SH).
Only one previous study, Fenter and Anderson,22 reported a
temperature dependence of k2. They employed a discharge flow
tube to generate SD using two different sources. One was the
abstraction of a deuterium atom from D2S by a H atom, and
the other was to react a deuterium atom with ethylene sulfide,
which yielded SD and ethene. NO2 was added through a
movable injector at the end of the flow tube. SD was detected
via laser-induced fluorescence. The results of our work are in
excellent agreement with the Fenter and Anderson22 study. The
results of Wang et al.5 on k2 at room temperature also agree
within the combined uncertainties with our value.
The history of k1 measurements is richer than that of SD with
several published studies (see Table 2). The range of measure-
ments is great, from the product study of Bulatov et al., which
reported 2.4 × 10-11 cm3 molecule-1 s-1 to the value of 12 ×
10-11 cm3 molecule-1 s-1 found by Schoenle et al.23 The
Stachnik and Molina study suggests the measurements of Friedl
et al.24 and Black16 suffered from the SH regeneration, as
outlined earlier, and added O2 to the reaction mixture to suppress
those regeneration schemes. The Wang et al.5 study measured
k1 as a function of temperature, using two sources of SH; the H
+ ethylene sulfide, already discussed, and the reaction of F with
H2S. Due to the rapidity of F + H2S relative to the H + H2S
reaction, the regeneration of SH via reactions 8 and 10 can be
minimized. It should be noted that the F atoms generated in
the Wang et al. study were created through the microwave
discharge of a CF4/He mixture. Using F2 creates difficulties, as
it participates in SH regeneration via the F2 + SH reaction. Our
results agree with those of Wang et al. The source chemistry
and regeneration of SH are discussed by Wang et al. and
Stachnik and Molina. Therefore, the following discussion
addresses only the apparent influence of SO on the measured
rate coefficient.
Our discovery of an interference by SO in the detection of
SH suggests that the previous studies of Stachnik and Molina
and of Black may have inadvertently influenced their measured
values. The results of Black et al. were influenced by secondary
SH generation, as noted by others before. When the influence
of the SO interference is not accounted for (see Table 1), we
measured a k1 value which is close to the result of Stachnik
and Molina. It should be stated that a direct comparison of the
influence of SO on the measurement of k1 is not possible because
this technique, like the work of Black and Friedl et al., is based
on LIF of SH, while that of Stachnik and Molina is based on
the absorption of SH. In the absorption measurement, the
influence would depend on the absorption cross section of SO
and SH at 323.7 nm, whereas in the LIF measurements, the
relative detection sensitivity is a convolution of the absorption,
fluorescence strength, and collection efficiency (mostly filler
transmission). These comments regarding the influence SO has
on the measurement of k1 should not be considered as anything
more than another possible difficulty associated with the
measurement of k1, by LIF at the R1 bandhead.
If HSO were removed mostly via its reaction with O3 in the
atmosphere, the conversion of SH to HSO followed by conver-
sion of HSO with O3 to give SH will form a catalytic ozone
destruction cycle. However, given that only a fraction of the
HSO reaction with O3 leads to SH and that the abundances (and
emissions) of H2S to the atmosphere are rather small, we do
not expect SH reaction to be a significant catalytic removal
pathway for ozone. If SH is oxidized in NOx rich regions, the
formed HSO will likely be converted to HSO2 and eventually
to SO2. Thus, in the case of either O3 reaction or NO2 reaction,
we expect SH to be rapidly oxidized to SO2.
Acknowledgment. This work was funded in part by
NOAA’s Air Quality Program.
References and Notes
(1) Berresheim, H.; Wine, P. H.; Davis, D. D. In Composition,
Chemistry and Climate of the Atmosphere; Singh, H. B., Ed.; Van Nostrand
Reinhold: New York, 1995; pp 251-307.
(2) Finlayson-Pitts, B. J.; Pitts, J. N. J. Atmospheric Chemistry:
Fundamentals and Experimental Techniques; Wiley: New York, 1986.
(3) Seinfeld, J. H. Atmospheric Chemistry and Physics of Air Pollution;
Wiley: New York, 1986.
(4) Stachnik, R. A.; Molina, M. J. J. Phys. Chem. 1987, 91, 4603-
4606.
(5) Wang, N. S.; Lovejoy, E. R.; Howard, C. J. J. Phys. Chem. 1987,
91, 5743-5749.
(6) Laasko, D.; Smith, C. E.; Goumri, A.; Rocha, J.-D. R.; Marshall,
P. Chem. Phys. Lett. 1994, 227, 377-383.
(7) Turnipseed, A. A.; Barone, S. B.; Ravishankara, A. R. J. Phys.
Chem. 1992, 96, 7502-7505.
(8) Black, G.; Patrick, R.; Jusinski, L. E.; Slanger, T. G. J. Chem. Phys.
1984, 80, 4065-4070.
(9) Robert Wu, C. Y.; Chen, F. Z. J. Quant. Spectrosc. Radiat. Transfer
1998, 60, 17-23.
(10) Weiner, B. R.; Lenene, H. B.; Valentini, J. J.; Baronavski, A. P. J.
Chem. Phys. 1989, 90, 1403-1413.
(11) Barone, S. B.; Turnipseed, A. A.; Gierczak, T.; Ravishankara, A.
R. J. Phys. Chem. 1994, 98, 11969-11977.
(12) Butler, L. J. Chem. Phys. Lett. 1991, 182, 393-399.
(13) Continetti, R. E.; Balko, B. A.; Lee, Y. T. Chem. Phys. Lett. 1991,
182, 400-405.
(14) Friedl, R., R.; Brune Wm, H.; Anderson, J. G. ′J. Chem. Phys.
1983, 79, 4227-4236.
(15) Ramsay, D. A. J. Chem. Phys. 1952, 20, 1920.
(16) Black, G. J. Chem. Phys. 1984, 80, 1103-1107.
(17) Chen, X.; Wang, H.; Weiner, B. R.; Hawley, M.; Nelson, H. H. J.
Phys. Chem. 1993, 97, 12269-12274.
(18) Wine, P. H.; Kreutter, N. M.; Ravishankara, A. R. J. Phys. Chem.
1979, 83, 3191-3195.
(19) Sander, S. P.; Friedl, R. R.; Golden, D. M.; Kurylo, M. J.; Huie,
R. E.; Orkin, V. L.; Moortgat, G. K.; Ravishankara, A. R.; Kolb, C. E.;
Molina, M. J.; Finlayson-Pitts, B. J. Chemical Kinetics and Photochemical
Data for Use in Atmospheric Studies, Evaluation No. 14 of NASA Panel
for Data Evaluation, JPL Publication 02-25, February 2003, Pasadena CA.
Available at http://jpldataeval.jpl.nasa.gov.
(20) Johnston, H. S.; Graham, R. Can. J. Chem. 1974, 52, 1415-1423.
(21) Bass, A. M.; Ledford, A. E.; Laufer, A. H. J. Res. Natl. Bur. Stand.
(U.S.) 1976, 80A, 145-166.
(22) Fenter, F. F.; Anderson, J. G. Int. J. Chem. Kinet. 1994, 26, 801-
812.
(23) Schoenle, G.; Rahman, M. M.; Schindler, R. N. Ber. Bunsen-Ges.
Phys. Chem. 1987, 91, 66.
(24) Friedl, R. R.; Brune, W. H.; Anderson, J. G. J. Phys. Chem. 1985,
89, 5505-5510.
(25) Bulatov, V. P.; Kozliner, M. Z.; Sarkisov, O. M. Khim. Fiz. 1984,
3, 1300-1305.
Atmospheric Implications
The atmospheric lifetime of SH is determined primarily by
its reaction with O3 and NO2. HSO is believed to be the primary