How Rapidly Does the SH Radical React with N2O?

Article abstract of DOI:10.1021/jp9911853

The reaction of SH radical with N₂O was investigated by using two complementary experimental methods: (a) a pulsed photolysis apparatus where SH was detected via laser-induced fluorescence and (b) a discharge flow tube where SH was detected via chemical ionization mass spectrometry. The rate coefficient for this reaction, k₁, was found to be less than 5 ′ 10^-16 cm³ molecule^-1 s^-1 at 298 K, in contrast to the large value reported in the literature [Ravichandran; et al. Chem. Phys. Lett. 1994, 217, 375-379]. Secondary reactions were deduced to be unimportant in our system. Some possible reasons for the previously reported high value are presented. The low value of k₁ makes the oxidation of SH by N₂O and removal of N₂O by SH unimportant in the earth's atmosphere.

Full text of DOI:10.1021/jp9911853

6778
J. Phys. Chem. A 1999, 103, 6778-6785
How Rapidly Does the SH Radical React with N2O?
Scott C. Herndon,† Karl D. Froyd,† Edward R. Lovejoy, and A. R. Ravishankara*,†
National Oceanic and Atmospheric Administration, Aeronomy Laboratory, Boulder, Colorado 80303-3328
ReceiVed: April 8, 1999; In Final Form: June 23, 1999
The reaction of SH radical with N2O was investigated by using two complementary experimental methods:
(a) a pulsed photolysis apparatus where SH was detected via laser-induced fluorescence and (b) a discharge
flow tube where SH was detected via chemical ionization mass spectrometry. The rate coefficient for this
reaction, k1, was found to be less than 5 × 10-16 cm3 molecule-1 s-1 at 298 K, in contrast to the large value
reported in the literature [Ravichandran; et al. Chem. Phys. Lett. 1994, 217, 375-379]. Secondary reactions
were deduced to be unimportant in our system. Some possible reasons for the previously reported high value
are presented. The low value of k1 makes the oxidation of SH by N2O and removal of N2O by SH unimportant
in the earth’s atmosphere.
Introduction
were correct. A short lifetime of N2O (due primarily to seasonal
H2S emissions) is inconsistent with the observed invariability
of its abundance5 with season and location, and other sources
of N2O would be required to balance its atmospheric budget.
The title reaction is also interesting because the analogous
reaction, OH + N2O f HO2 + N2, is also exothermic (∆°Hrxn
≈ -9 kcal mol-1) but very slow,7 ∼3.8 × 10-17 cm3 molecule-1
s-1. (Continuing work in our laboratory places an upper limit
for this rate constant at less than 8 × 10-18 cm3 molecule-1
s-1.) This low value is to be contrasted with the reported high
value of k1.
This paper reports an upper limit of k1 determined using two
experimental methods: pulsed photolysis laser-induced fluo-
rescence (PP-LIF) and discharge flow chemical ionization mass
spectrometry (DF-CIMS). We find k1 is less than 5 × 10-16
cm3 molecule-1 s-1, which sharply contrasts with the report of
Ravichandran et al.4 The possibility that secondary reactions
are responsible for the measured low value of k1 is also discussed
and shown to be very unlikely.
Hydrogen sulfide (H2S), a predominantly terrestrial emission,
is one of the important natural sources of atmospheric sulfur.1
In the troposphere, H2S is removed mostly via its reaction with
the OH radical, which produces the SH radical.2 The reactions
of SH with O3 and NO2 (k(298) ) 3.5 × 10-12 and 6.5 × 10-11
cm3 molecule-1 s-1, respectively3) are the main pathways for
removal of SH and give it an atmospheric lifetime of less than
a second. Oxidation of SH eventually leads to SO2.
It has been reported4 that SH reacts rapidly with N2O, a gas
that is believed to be unreactive in the troposphere,5 to form
HSO.
SH + N2O(0, 0, 0) f HSO + N2;
k1, ∆°Hrxn(298K) ≈ -55 kcal mol-1 (1)
In the above reaction, (0, 0, 0) denotes the vibrational ground
state of N2O. The enthalpy of reaction 1 was calculated from
tabulated data.3 Ravichandran et al.4 report a value of k1 at 298
K of (1.3 ( 0.14) × 10-11 cm3 molecule-1 s-1. They photolyzed
H2S at 193 nm to produce the SH radical, in varying pressures
of N2O, and measured the temporal profiles of the resulting
HSO chemiluminescence to deduce k1. If reaction 1 is as rapid
as they report, the mechanism of H2S oxidation and lifetime of
N2O in the atmosphere will have to be revised significantly. A
rough estimate6 of the average SH radical concentration in the
lower troposphere is 1 × 102 molecules cm-3. Even such a low
concentration of SH would lead to an atmospheric lifetime of
roughly 24 years for N2O with respect to reaction 1 if the value
reported by Ravichandran et al. is used. A lifetime of 24 years
for N2O is 5 times smaller than the currently accepted value of
around 120 years.5 Even if SH is isolated to a few regions (i.e.,
terrestrial boundary layer), it would still significantly reduce
the currently accepted N2O lifetime if the reported value of k1
Experiments and Results
The initial experiments were carried out using pulsed laser
photolysis of H2S as a source of SH. Because we obtained values
of k1 that were significantly lower than that reported in the
literature, we also measured k1 using a different experimental
system that avoids photolysis and the consequent secondary
reactions. For ease of presentation, these two experiments and
their results are discussed below in separate sections.
PP-LIF. The apparatus and procedures used for studying the
reactions of SH are very similar to those used to measure the
rate coefficients for reactions of the OH radical8 and the CH3S
radical.9 Therefore, only the description of the SH detection
and a few details necessary to understand this work are given
here. Figure 1 shows the schematic of the pulsed photolysis
laser-induced fluorescence apparatus. The SH radical was
generated by pulsed laser photolysis of H2S, and its temporal
profile was measured by pulsed laser-induced fluorescence.
* To whom correspondence should be addressed. NOAA/ERL, R/E/AL2,
325 Broadway, Boulder, CO 80303. E-mail: ravi@al.noaa.gov.
† Also affiliated with Department of Chemistry and Biochemistry,
University of Colorado, Boulder, Colorado, and The Cooperative Institute
for Research in Environmental Sciences, University of Colorado, Boulder,
Colorado.
H2S hν/193 nm8 H + SH
(2)
10.1021/jp9911853 CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/06/1999

SH Radical Reaction with N2O
J. Phys. Chem. A, Vol. 103, No. 34, 1999 6779
Figure 3. Temporal profiles of SH measured in the absence (open
circles) and the presence (triangles) of >1 × 1017 molecules cm-3 N2O
in the PP-LIF system. The upper profiles were measured in the absence
of O2. The increase in SH signal at short times is due to the reaction
of H atoms with H2S. The lower profiles were measured in the presence
of O2 (3.7 Torr), which scavenged H atoms and suppressed the
secondary production of SH. The time constant for the decay of SH
was the same within the precision of the data in all cases. The initial
SH concentrations were (4-7) × 1010 molecules cm-3. The total system
pressure was 93-107 Torr. The signal levels are consistent with the
recorded precursor concentrations and photolysis fluences.
Figure 1. Schematic of the pulsed photolysis apparatus used to study
reactions of SH radicals. SH radicals were generated by pulsed laser
photolysis of H2S and were detected by pulsed laser-induced fluores-
cence. The components of the apparatus are marked in the figure.
defined as a signal-to-noise ratio of 1, where the noise is twice
the standard deviation of the mean of the PMT signal measured
in the absence of SH (i.e., background) and the signal is the
value above this background. The initial concentration of SH
almost never exceeded 8 × 1010 molecules cm-3 (the exception
is noted later).
The concentrations of N2, He, O2, a 2% H2S mixture with
He, and N2O in the reaction cell were calculated from their flow
rates measured using calibrated electronic mass flow meters.
Pressure and temperature were measured using a calibrated
capacitance manometer and a thermocouple probe, respectively.
Both UHP (>99.99995%) He and N2 (>99.9995%) were used
as carrier gases.
Figure 2. Fluorescence excitation spectrum of SH measured in this
study. SH was generated by pulsed photolysis of H2S at 193 nm. This
spectrum was acquired using a 355 ( 12 nm band-pass filter, delay
time of 100 µs, [H2S] ≈ 3 × 1014 molecules cm-3, 80 Torr of N2 total
pressure, and a photolysis energy of 5 mJ pulse-1. The line shapes and
peak positions of this excitation spectrum did not change in the presence
of 12 Torr of N2O. The “bump” beginning near 323.6 nm and tailing
off with longer wavelength is not due to SH but, as will be discussed
in a forthcoming paper, is likely an LIF signal due to SO.
Some typical SH temporal profiles acquired in the PP-LIF
experiments are shown in Figure 3. H2S photolysis at 193 nm
yields SH and H atoms. H atoms react with H2S to generate
more SH on a time scale that depends on the concentration of
H2S.
H + H2S f H2 + SH
(3)
All rate coefficients were measured at 295 ( 3 K in 30-120
Torr of N2 (or 100 Torr of He).
Laser-induced fluorescence of SH (A2Σ r X2Π) was excited
by the output of an excimer-pumped dye laser. A photomultiplier
tube (PMT) was placed at right angles to the photolysis and
the excitation laser beams. A band-pass filter was placed
between the PMT and the reaction zone to allow maximal
transmission at 355 ( 12 nm and thus detection of fluorescence
from the (0 f 1) vibronic manifold. This filter discriminated
against scattered light from the photolysis and probe lasers.
Figure 2 shows an excitation spectrum of SH taken 100 µs after
its production via the 193 nm photolysis of H2S in 80 Torr of
N2 total. This spectrum is in excellent agreement with other
published excitation spectra.10,11 The line shape and peak
locations were unchanged in the presence of 12 Torr of N2O.
The probe laser frequency was scanned to locate the fluores-
The rate coefficient for reaction 3 at 298 K has been meas-
ured12,13 to be ∼7 × 10-13 cm3 molecule-1 s-1. Thus, the
temporal profiles of SH should have a rise due to reaction 3
and a decay due to SH loss. Such biexponential temporal profiles
were indeed observed (Figure 3). Addition of O2 scavenged the
H atoms to make HO2,
H + O2
9
M8 HO2
(4)
and suppressed the secondary production of SH. Therefore, a
single-exponential decay of SH was observed when sufficient
O2 was present to scavenge H atoms (Figure 3). The rate
coefficient for the reaction of HO2 with H2S is less than 3 ×
10-15 cm3 molecule-1 s-1 14
. Even if this reaction produced SH
R
cence peak corresponding to the Q21 (λ ≈ 323 nm) transition
(which is ∼3.3 kcal mol-1 endothermic at 298 K), it could not
regenerate SH on the time scale of our experiments. The first-
order rate constant for the loss of SH, kI, was evaluated both in
the presence and in the absence of O2. This rate constant is
attributed to the sum of the losses of SH due to self-reaction,
reaction with any impurities in the reactor, diffusion out of the
reaction zone (defined as the intersection volume between the
photolysis and the probe laser beams), and reaction with N2O
2
2
in the Σ r Π manifold. Excitation at this peak was used to
monitor SH in kinetics experiments described below.
The sensitivity for SH detection was calculated to be ∼8 ×
108 molecules cm-3 (averaging 100 laser shots) by measuring
the LIF signal from a known concentration of SH. The
concentration of SH was calculated using measured concentra-
tions of H2S and photolysis fluence. The fluence was measured
using a calibrated thermopile detector. The detection limit is

6780 J. Phys. Chem. A, Vol. 103, No. 34, 1999
Herndon et al.
TABLE 1: Experimental Conditions and Results in the Pulsed 193 nm Photolysis Laser-Induced Fluorescence System
conditions
label and temp
PL1,a T ) 295 K
photolysis fluence
pressure
(Torr) and gas
[O]0 (atom cm-3
from N2O photolysis
)
[N2O]
(mJ cm-2 pulse-1
0.02
)
(molecule cm-3
)
k′ ( σ (s-1
)
110 (N2)
0
0
36 ( 1
43 ( 1
37 ( 2
38 ( 2
39 ( 2
35 ( 1
36 ( 1
35 ( 1
35 ( 2
36 ( 1
67 ( 13
61 ( 9
54 ( 6
50 ( 8
46 ( 4
51 ( 5
49 ( 4
52 ( 8
61 ( 15
5 × 1011
4 × 1011
2 × 1011
5 × 1011
1 × 1011
7 × 1010
2 × 1010
0
2.8 × 1017
2.3 × 1017
1.0 × 1017
2.8 × 1017
7.7 × 1016
3.9 × 1016
1.3 × 1016
0
5 × 109
7 × 1013
5 × 1013
3 × 1013
9 × 1012
0
3.1 × 1015
1.1 × 1017
8.0 × 1016
4.3 × 1016
1.4 × 1016
0
PL2,b T ) 294 K
7
100 (He)
0
0
9 × 1012
3 × 1013
5 × 1013
0
1.5 × 1016
4.4 × 1016
8.7 × 1016
0
PL3,c T ) 295 K
0.04
0.03
95 (N2)
7300 ( 150
7836 ( 220
8048 ( 103
8156 ( 92
9255 ( 212
8330 ( 151
7773 ( 116
7826 ( 94
8108 ( 100
53 ( 2
1 × 1010
6 × 1011
4 × 1011
0
3.5 × 1015
1.7 × 1017
1.1 × 1017
0
2 × 1011
1 × 1011
4 × 1011
1 × 1012
0
4.3 × 1016
2.8 × 1016
1.2 × 1017
3.7 × 1017
0
PL4,d T ) 295 K
PL5,e T ) 296 K
100 (N2)
82 (He)
4 × 1011
1 × 1011
2 × 1011
0
1.6 × 1017
5.8 × 1016
8.5 × 1016
0
57 ( 2
59 ( 2
53 ( 1
51 ( 1
9 × 109
1 × 1010
1 × 1010
4 × 1010
8 × 1010
2 × 1011
0
3.4 × 1015
2.7 × 1014
3.0 × 1014
9.0 × 1014
1.8 × 1015
4.0 × 1015
0
49 (1
0.5
0.5
0.5
0.5
0.5
0.5
0.06
0.06
24 ( 3
16 ( 1
21 ( 1
27 ( 3
29 ( 5
20 ( 1
0
0
30 ( 1
2 × 1010
3.9 × 1015
31 ( 2
a [H2S] ) 3.4 × 1014 molecules cm-3 and [SH]0 ≈ 5 × 1010 molecules cm-3 with [O2] ) 1.2 × 1017 molecules cm-3
.
b [H2S] ) 7 × 1013
molecules cm-3 and [SH]0 ≈ 3 × 1012 molecules cm-3
.
c [H2S] ) 3.4 × 1014 molecules cm-3 and [SH]0 ≈ 1 × 1010 molecules cm-3 with [NO2]
≈ 1 × 1014 molecules cm-3
.
d The listed k’s are taken from the time constants fit to a biexponential form. [H2S] ) 8.9 × 1013 molecules cm-3 and
[SH]0 ≈ 2 × 1010 molecules cm3. e For the experiments where the photolysis fluence is 0.5 mJ cm-2 pulse-1, the H2S concentration is 1.4 × 1014
molecules cm-3 and where the fluence is 0.06 mJ cm-2 pulse-1, the H2S concentration is 6.3 × 1014 molecules cm-3. This implies that [SH]0 is 4
× 1011 and 2 × 1011 molecules cm-3, respectively.
or other species generated via laser photolysis.
SH + SH f products
constant is given by
kI ) kd + k′ + k′
(I)
(5)
(6)
(7)
(1)
1
X
SH + impurities f products
SH f diffusion from reaction zone
SH + N2O f products
where k′ ) k1[N2O] and k′ represents the sum of the
1
X
first-order rate coefficients for the removal of SH with all
photoproducts. Note that O2 does not react rapidly with SH.3
The concentration of O atoms from the 193 nm photolysis of
O2 was less than 109 atoms cm-3 and did not contribute
significantly to SH loss.
SH + X f products
(8)
The temporal profiles of SH were measured at various
concentrations of N2O and analyzed to extract kI. Experimental
conditions such as the laser fluence (0.02-7 mJ cm-2) and the
SH precursor concentration (1 × 1013 to 4 × 1015 molecules
cm-3) were also varied. Table 1 presents the results of these
experiments. A plot of the measured pseudo-first-order rate
constant for loss of SH vs [N2O] is shown in Figure 4. To place
k1′ from all the experiments on the same plot, the value of kd
measured in the absence of N2O (and associated with each set
X in reaction 8 is a photoproduct (e.g., O(3P) and NO) produced
in the system whose concentration increases proportionately with
that of N2O. Loss of SH via processes 5-7 is represented as
first order in [SH] with an associated first-order rate coefficient
denote by kd. (Reaction 5 is second order in SH; however, over
the time scales of the measurements with the low initial SH
concentrations that were used, it can be roughly represented as
a first-order process.) The overall first-order SH decay rate

SH Radical Reaction with N2O
J. Phys. Chem. A, Vol. 103, No. 34, 1999 6781
Figure 5. Schematic of the discharge flow CIMS apparatus. SH
radicals were generated in a microwave cavity in the sidearm of the
flow reactor, and N2O entered the reactor through a movable injector.
DF-CIMS. A schematic of the discharge flow chemical
ionization mass spectrometer (DF-CIMS) instrument used in
this study is shown in Figure 5. This instrument has been used
previously to measure rate coefficients for various reactions.15,16
Therefore, only the details associated with the present measure-
ments are discussed.
Figure 4. Plot of the first-order rate coefficient for the loss of SH
(after subtracting the contribution due to removal in the absence of
N2O) measured in the PP-LIF system as a function of N2O concentra-
tion. The fitted slope plus the 95% confidence interval (dashed line) is
2 × 10-16 cm3 molecule-1 s-1. The solid black line is the upper limit
this work (PP-LIF and DF-CIMS) places on k1 (5 × 10-16 cm3
molecule-1 s-1). The various symbols represent different sections of
data found in Table 1. The open triangles are the PL2 data, the squares
are the PL4 data, and the diamonds are the PL1 data. The PL3 data are
omitted from this figure because the imprecision in that data would
make the data presented here difficult to see. The PL5 is omitted as
well; it spans a limited concentration range and is clearly affected by
secondary reactions (see Discussion).
A 1% mixture of H2 in UHP He was passed through an active
metal trap to remove O2 and H2O impurities before flowing
through a microwave cavity. The H atoms generated in the
microwave discharge reacted with an excess of ethylene sulfide
((5-20) × 1013 molecules cm-3) in a sidearm reactor to produce
the SH radical:17,18
of measured k′ values) has been subtracted. The data in Figure
4 are fit to eq II,
1
H + (CH2)2S f SH + C2H4
(9)
Reaction 9 (k9 ) 1.2 × 10-12 cm3 molecule-1 s-1) was allowed
to go to completion (i.e., typically greater than three lifetimes
in [H]) before SH came into contact with N2O. N2O was added
through a 9.5 mm movable Teflon injector downstream of the
SH generation region. Thus, SH generation by reaction 9 was
insignificant after the mixture encountered N2O. A 2.22 cm i.d.
Teflon sleeve lined the inside of the flow reactor to minimize
radical loss on the reactor walls. Flow rates of N2O (0.001-
1.1 STP cm3 s-1; STP ) 1 atm, 273 K) were calculated from
the time rate of change of pressure in a calibrated volume.
Concentrations of N2O in the flow reactor were 1.5 × 1013 to
2.2 × 1016 molecules cm-3. Up to 10 STP cm3 s-1 of UHP
(>99.9995%) He was passed through a liquid nitrogen trap and
added to the flow reactor to serve as a carrier gas. The pressure
in the middle of the 50 cm reaction zone, 1.2-1.9 Torr, was
measured using a calibrated capacitance manometer.
kI - kd ) k1[N2O] + k′
(II)
X
using an unweighted linear least-squares method. Note that k′
X
may vary linearly with [N2O] and may contribute to the
measured slope. The slope of this plot, plus the 95% confidence
interval (derived from the precision of the fit), is 2 × 10-16
cm3 molecule-1 s-1. This is the upper limit for k1 based only
on the PP-LIF measurements, excluding the possible systematic
influences considered in Discussion.
The temporal profiles of SH measured in He (see Table 1,
PL2 subset) using a photolysis fluence of 7 mJ pulse-1 cm-2
did not obey pseudo first-order kinetics. The pseudo first-order
rate constants were determined by fitting the temporal profile
after excluding the sharper initial decay and are shown in Table
1 and Figure 4 for the PL2 experiments. In all cases, some N2O
was unavoidably photolyzed to produce O(1D); however, in the
PL2 experiments, the fluence was so large that the concentration
of initially produced O(1D) was ∼1013 molecules cm-3. The
implications of the generation of the high concentration of O(1D)
will be discussed later.
In the ion flow tube, SH radicals exiting the radical flow
-
reactor were converted to SH- via charge transfer with the SF6
reagent ion:
SF6- + SH f SF6 + SH-
(10)
Another set of experiments was conducted with a roughly
constant amount of NO2 present in the reaction mixture (PL3).
They are not shown because there is a large offset in k′ vs [N2O]
for these data (7300 ( 150 s-1), due to the reaction of SH with
NO2. The values of kI for these data show no clear positive
The SF6- ion was generated at the upstream end of the ion flow
tube by passing a small flow of SF6 over an electron-emitting
filament. Up to 120 STP cm3 s-1 of He was added to the 7.30
cm i.d. stainless steel ion flow tube, where the total pressure
was approximately 0.6 Torr. SH- was mass-selected with a
quadrupole mass filter and detected by an electron multiplier.
SH concentrations of less than 1 × 1010 molecules cm-3 in the
flow reactor were estimated for all the kinetics experiments on
the basis of the measured SH- signal levels and an assumed
influence of N2O (3.5 × 1015 to 3.7 × 1017 molecules cm-3
)
on the loss rate constants of SH. These data considered alone
yield a rate constant for the title reaction of 8 × 10-16 cm3
molecule-1 s-1; however, these experiments were performed
strictly as a crude measurement of k1 in a system where SH is
lost rapidly. Additional implications of these experiments are
considered in Discussion.
rate coefficient for the reaction of SH with SF6 of 2 × 10-9
-

6782 J. Phys. Chem. A, Vol. 103, No. 34, 1999
Herndon et al.
TABLE 2: Experimental Conditions and Results in the
Discharge Flow Chemical Ionization Mass Spectrometry
System
[N2O]
(molecule cm-3
k′ ( σ
(s-1
)
)
DF1
1.28-1.77 Torr
pressure
temperature
0
0
0
-4.5 ( 0.6
-4.6 ( 0.4
-17.3 ( 0.5
-15.3 ( 0.8
-4.3 ( 0.9
-7.7 ( 0.6
-21.0 ( 0.5
295 K
main flow velocity 1169-1465 cm s-1
3.27 × 1015
9.81 × 1015
1.71 × 1016
2.19 × 1016
DF2
pressure
temperature
1.30-1.57 Torr
0
0
-14.7 ( 1.0
-13.9 ( 1.0
-20.9 ( 1.1
-16.4 ( 1.1
-20.0 ( 1.2
295 K
main flow velocity 1219-1357 cm s-1
6.02 × 1015
8.93 × 1015
1.65 × 1016
Figure 6. Plot of k′ - kloss, measured using the discharge flow CIMS
apparatus, as a function of [N2O] in the reactor. The solid line is a
weighted linear fit to the data yielding (-2.5 ( 1.5) × 10-16 cm3
molecule-1 s-1. Error bars represent the combined 95% confidence
intervals for the fits of k′ and kloss. The reported limit of k1 < 5 ×
10-16 cm3 molecule-1 s-1 is plotted as a dashed line.
DF3
pressure
temperature
1.31-1.44 Torr
0
0
0
-10.5 ( 0.2
-12.3 ( 0.5
-11.9 ( 0.3
-12.0 ( 0.4
-12.5 ( 0.3
-11.5 ( 0.4
-11.5 ( 0.6
-12.0 ( 0.3
295 K
main flow velocity 1175-1301 cm s-1
1.57 × 1013
1.57 × 1013
6.69 × 1014
1.15 × 1015
5.76 × 1015
where kloss is the first-order rate constant for removal of SH in
the absence of N2O, d is the relative reaction distance, V is the
average linear gas flow velocity in the flow reactor downstream
of the injector, and C represents an experimental background
signal at this mass. The DF-CIMS results are presented in Table
2. Loss of SH radicals on the outer wall of the injector led to
a rise in SH- signal as the injector was retracted so that the
measured k′ values were negative. This is generally observed
in flow tube studies where the monitored radical is in the main
body of the flow tube and the excess reactant is introduced
through a movable injector, as in the present work. Decay
constants obtained in the absence of N2O (i.e., kloss) were
subtracted from those measured in the presence of N2O (k′).
The rate constant difference (k′ - kloss) ranged from -7.4 to
1.7 s-1 and typically encompassed zero within 2σ of the
DF4
1.32-1.56 Torr
295 K
pressure
temperature
0
-20.4 ( 0.6
-17.0 ( 0.6
-18.4 ( 0.7
-17.9 ( 0.7
-19.2 ( 1.1
-19.0 ( 0.9
1.85 × 1013
1.01 × 1014
4.22 × 1014
1.47 × 1015
5.52 × 1015
main flow velocity 1369-1439 cm s-1
[N2]
k′ ( σ
molecule cm-3
(s-1
)
DF5
1.30-1.49 Torr
295 K
pressure
temperature
0
-17.7 ( 1.0
-14.9 ( 0.7
-15.9 ( 0.7
-12.5 ( 0.5
-12.8 ( 0.7
-20.0 ( 0.8
1.93 × 1013
9.80 × 1013
1.23 × 1014
4.08 × 1014
4.98 × 1015
main flow velocity 1377-1480 cm s-1
measured value of k′. Figure 6 shows a plot of (k′ - kloss
)
measured at various concentrations of N2O. It is clear that there
is no discernible change in (k′ - kloss) upon addition of N2O. A
weighted linear fit of these data yields a slope of (-2.5 ( 1.5)
× 10-16 cm3 molecule-1 s-1 at room temperature. Error bars
are determined by combining the 95% confidence intervals of
[NO2]
k′ ( σ
molecule cm-3
(s-1
)
DF6
1.27-1.36 Torr
343 K
pressure
temperature
0
-8.25 ( 0.3
-5.68 ( 0.5
19.65 ( 0.9
16.68 ( 0.9
59.69 ( 1.1
69.24 ( 2.2
105.78 ( 3.5
9.63 × 1010
2.64 × 1011
3.72 × 1011
8.38 × 1011
1.20 × 1012
1.61 × 1012
the fits to eq III for k′ and kloss
.
main flow velocity 1552-1650 cm s-1
Obviously, a rate coefficient cannot be negative. The negative
slope may be due to complex SH diffusion behavior in the flow
reactor. With only the He carrier gas present, SH radicals that
are lost on the flow reactor walls are replaced via fast diffusion.
This first-order wall loss of radicals is independent of injector
position and is not represented in the observed decays. However,
when a significant amount of N2O is added through the injector,
the effective diffusion constant for SH in the flow reactor
decreases and radical loss at the reactor wall is inhibited.
Retracting the injector exposes more of the SH stream to N2O,
resulting in a greater increase of SH signal than without N2O
present. Thus, high concentrations of N2O may serve to make
k′ more negative, as seen in Figure 6. In the DF-CIMS
experiments, N2O was as much as 40% of the flow reactor
content. A simulation of the conditions in the flow reactor
suggests that up to 30% of the observed (k′ - kloss) can be
accounted for by depressed radical wall loss.
cm3 molecule-1 s-1. Typically, exothermic electron-transfer
reactions such as reaction 10 proceed at the Langevin collisional
rate.
Reaction 1 was studied under pseudo-first-order conditions
in SH by monitoring the SH- signal as a function of injector
position. The SH- signal level is proportional to the concentra-
tion of SH at the end of the ion flow tube. The variation of the
SH- signal with reaction distance was fitted to expression III
using a weighted least-squares method,
SSH- ) A exp(-k′d/V) + C
In this expression,
(III)
To test that the negative k1 value was due to SH diffusion
effects and not to secondary chemistry in the flow reactor, a
k′ ) k1[N2O] + kloss
(IV)

SH Radical Reaction with N2O
J. Phys. Chem. A, Vol. 103, No. 34, 1999 6783
series of experiments was performed with N2 (an inert species)
in place of N2O. The variation of k′ with [N2] is similar to that
seen with N2O. Since N2 does not react with SH, this experiment
lends support to the low measured value of k1. The line
representing a k1 value of 5 × 10-16 cm3 molecule-1 s-1 is
also shown in the figure. Clearly, the measured values are con-
in the presence of N2O. In addition, when O2 was added in
sufficient quantities to scavenge H atoms, the decay rate constant
for SH did not change within the precision of the measurements;
however, the initial rise in SH due to reaction 3 was essentially
eliminated, as one anticipates (see Figure 3). The lifetime of H
in 3.5 Torr of O2 and 1 mTorr of H2S with a total pressure of
100 Torr of N2 is 44 µs, on the basis of the known rate
coefficients3 for reactions 3 and 4. This time constant is much
shorter than the lifetime of SH observed in our experiments.
Also, in one series of experiments, NO2 was added to scavenge
H atoms and the obtained value of k1 was essentially the same.
On the basis of these arguments, we rule out the possibility of
SH regeneration via reaction 3 as a significant source of error
in measuring k1.
sistent with an upper limit of 5 × 10-16 cm3 molecule-1 s-1
.
Discussion
As shown above, we did not observe a reaction between SH
and N2O in the pulsed photolysis system. An upper limit for k1
is 2 × 10-16 cm3 molecule-1 s-1, based on the pulsed photolysis
experiments. This result is consistent with the upper limit of 5
× 10-16 cm3 molecule-1 s-1 derived from the flow tube studies.
Similar results from two distinctly different experimental
methods that do not have the same sources of systematic errors
lend confidence to our conclusion that reaction 1 is slow. It is
conceivable, though, that reaction 1 occurs rapidly, but in these
experiments, SH does not appear to react with N2O. Therefore,
we examined possible mechanisms that would mask the
occurrence of reaction 1 in our systems. The quantification of
these considerations leads to our final result, k1 < 5 × 10-16
cm3 molecule-1 s-1. We also explore below a possible reason
for the large value of k1 reported by Ravichandran et al.4
Secondary Chemistry in the PP-LIF System. The SH
radical may be regenerated by secondary reactions whose rates
in our systems scale with N2O concentration and thus obscure
reaction 1. This possibility is explored below for the pulsed
photolysis system. The reactions that can occur following
photolysis of a mixture of H2S and N2O at 193 nm are
Influence of the Reaction of O + H2S. The concentration of
O(1D) produced through the unavoidable photolysis of N2O at
193 nm scales with N2O concentration. In the PP-LIF reactor,
O(1D) is collisionally quenched to O(3P) by N2 within 0.2 µs
(calculated five lifetimes for 30 Torr, the lowest pressure that
was employed). A fraction of O(1D) also reacts with N2O to
make N2, O2, and NO (reaction 13). The presence of O atoms
can alter the temporal behavior of SH in two ways: (a) it can
generate SH via reaction 14 and (b) it can deplete SH via
reaction 15 and, to a lesser extent, via addition to NO. At a
fixed concentration of N2O, the laser fluence was varied by a
factor of 25 to proportionately change the initial O atom
concentration ([O]max < 1 × 1012 atom cm-3) and, hence, its
contribution to SH loss. The contribution of O atoms was
identified to be negligible because the measured value of k1,
within the precision of the measurements, did not change (see
PL1 and PL5 in Table 1). The regeneration of SH via reaction
14, even though it results in the production of two SH radicals
(because the OH product in this reaction will rapidly react with
H2S to give a second SH), is relatively slow (k14(298K) ≈ 2 ×
10-14 cm3 molecule-1 s-1) and should be negligible; i.e., the
first-order rate constant for regeneration of SH is less than 2
s-1. Changing the concentration of H2S did not change the
measured value of k1 and again confirms the expected negligible
contribution of reaction 14 (see Table 1).
H2S hν/193 nm8 H + SH
HS + N2O f HSO + N2
H + H2S f HS + H2
(2)
(1)
(3)
N2O hν/193 nm8 N2 + O(1D)
O(1D) + N2 f O(3P) + N2
O(1D) + N2O f O2 + N2
O(1D) + N2O f 2NO
(11)
(12)
With the exception of the data taken at high fluence and/or
He bath gas, photolysis of N2O generated less than 1 × 1012
molecules cm-3 of O(3P). The rate constant for the reaction of
O(3P) with SH is approximately gas kinetic19 and can lead to
an initial rapid loss of SH with a first-order rate constant of up
to 100 s-1. We scavenged O(3P) by adding O2 and reduced the
SH loss via reaction with O atom by a factor of at least 3; the
measured first-order rate coefficient for SH loss did not change.
Therefore, we conclude that the O atom reactions did not
contribute to the measured small value of k1. If HO2 reacts with
the SH radical as quickly as it does with the OH radical3 (1.1
× 10-11 cm3 molecule-1 s-1), the rate constant for the initial
(13a)
(13b)
(14)
O(3P) + H2S f OH + SH
O(3P) + HS f OH + S
(15a)
O(3P) + HS f SO + H
(15b)
(16)
loss of SH radical could be enhanced by as much as 10 s-1
.
HSO + N2O f N2 + O2 + SH
This enhancement was estimated from the calculated concentra-
tion of HO2 generated in the experiments when O2 was added
to scavenge H atoms. The increase is a negligible change to
the decay rate constants.
Influence of the Reaction of H + H2S. If SH is regenerated
by reaction 3 on the time scale of reaction 1, we may not observe
a loss of SH in our systems. Analyses of the biexponential
temporal profiles (see Figure 3) yielded time constants for the
rise that agreed with the rate coefficient for reaction 3.
Furthermore, as expected, the rise time for SH increased with
H2S concentration. The rise time of SH should scale only with
[H2S] because H atoms do not react rapidly with N2O or O
atoms. The rate coefficient for the loss of SH derived from the
biexponential fits in the absence of N2O was the same as those
It should be noted that none of the secondary reactions
mentioned above (reactions 3, 14, and 15) are important in the
DF-CIMS experiment, which yielded the same results as the
pulsed photolysis measurements.
SH Radical Detection in the DF-CIMS System. To validate
the SH detection methodology in the flow tube experiments,
the rate coefficient for the reaction of SH with NO2,

6784 J. Phys. Chem. A, Vol. 103, No. 34, 1999
SH + NO2 f HSO + NO
Herndon et al.
Though they mention that LIF may be used to detect HSO
and SH, Ravichandran et al. did not present such data in their
paper. Therefore, we will discuss only their chemiluminescence
data. The high value of the rate coefficient reported by
Ravichandran et al. may be due to an incorrect interpretation
of their observations. These authors measured the chemilumi-
nescence from HSO(A2A′) upon photolysis of mixtures of H2S
and N2O at total pressures of 0.5-4 Torr. Ravichandran et al.
appear to have relied on the temporal profile of the HSO*
chemiluminescence signal to extract a value of k1. The authors
ruled out the possibility that reactions of SH with O(1D) or O(3P)
are responsible for the HSO signal. It appears from their
published paper that they did not consider the possibility that
the reaction of O(1D) with H2S could yield HSO(A2A′):
(17)
was also measured in a procedure analogous to that used to
measure k1. At 343 K and 1.3 Torr, the rate coefficient for
reaction 17 was measured to be (7.0 ( 0.9) ×10-11 cm3
molecule-1 s-1, in reasonable agreement with the accepted
value17 of (5.9 ( 1.0) × 10-11 cm3 molecule-1 s-1. This
experiment confirms that the SH- signal seen in our system is
due primarily to the SH radical inside the flow reactor.
Possibility of SH Regeneration from Reaction of HSO(A2A′)
with N2O. It is possible that the reaction of SH with N2O
produces HSO(A2A′).
SH + N2O f HSO(A2A′) + N2
∆rH°(298K) ) -7.8 kcal mol-1 (18)
O(1D) + H2S f HSO(A2A′) + H
∆rH°(298) ) -2 kcal mol-1 (20)
The heat of formation of HSO(A2A′) was calculated from the
known energy separation, ∼47 kcal mol-1, between HSO(X)
and HSO(A) states. Further, the reaction of HSO(A2A′) with
N2O to give SH (reaction 19) is exothermic:
This reaction may occur because O(1D) has a propensity to insert
into stable molecules and the excited HSOH may eliminate an
H atom. Note that only a very small fraction of the O(1D)-
H2S encounter needs to produce chemiluminescence to obtain
a strong signal.
HSO(A2A′) + N2O f SH + N2 + O2
∆rH°(298K) ) -31 kcal mol-1 (19)
The apparent rate constants for the rise and decay of the HSO
chemiluminescence signals in Figure 3 of the Ravichandran et
al. paper are on the order of 10 and 1 µs-1, respectively. In
their Figure 4, they identified the rise as due to production of
HSO* (their panel a) and decay as to loss of HSO* (their panel
b). Figure 4a (labeled as production of HSO) has a slope that
would correspond to a bimolecular rate constant of 1.3 × 10-11
cm3 molecule-1 s-1, identified by them as the rate coefficient
for the reaction of SH with N2O. Their Figure 4b (labeled as
loss of HSO) has a slope corresponding to a bimolecular rate
constant of 1.5 × 10-10 cm3 molecule-1 s-1, which the authors
associated with the rate coefficient for the removal of HSO*.
Their assignment may be in contradiction with their Figure 3.
Without external calibration of the signal and/or a priori
knowledge of the rates of processes, it is not possible to
determine which part of the biexponential profile corresponds
to the formation of HSO* and which corresponds to its
quenching.21 Their reported value of k1 ) 1.3 × 10-11 cm3
molecule-1 s-1 may be the rate coefficient for the quenching
of HSO(A2A′) by N2O (reaction 21), while the faster rate
coefficient may be for the reaction of O(1D) with N2O in the
following sequence of processes in their system:
The reaction of HSO(X) with N2O is endothermic by ∼16 kcal
mol-1 and is not expected to occur (see below). If reactions 18
and 19 both occur, they could mask reaction 1. The quenching
rate coefficients20 of HSO(A2A′) by He and N2 are ∼2.5 × 10-12
and ∼5 × 10-12 cm3 molecule-1 s-1, respectively. The fraction
of HSO(A2A′) that could react with N2O is ∼0.6, in 100 Torr
of N2 and 10 Torr of N2O, if we assume k19 to be 1 × 10-10
cm3 molecule-1 s-1. In our experiments, the measured rate
constant did not change with N2O concentration. Even with 10
Torr of N2O, the fraction of HSO(A2A′) that is quenched is
∼0.4, and therefore, this scenario could mask the measurement
of k1 by a factor of at most 2.5, not orders of magnitude from
the real value. At lower concentrations of N2O, this quenching
fraction would be smaller and regeneration, if it occurs at all,
would be lower. Therefore, we conclude that the possible
reaction of HSO(A2A′) with N2O is not responsible for the low
measured value of k1 compared to that previously reported. This
conclusion is supported by the upper limit determined using
the flow tube where N2O concentrations were in the milliTorr
range.
In one series of experiments (PL3), ∼1 × 1014 molecules
cm-3 of NO2 was added. In this case, HSO is produced by
reaction 17. Reaction 17 is exothermic by ∼21 kcal mol-1. If
HSO(X), the product of reaction 17, contained all the excess
energy of the reaction, it could react with N2O to produce SH.
In such a case, the measured SH decays would have been
reduced upon addition of N2O. Such a decrease was not seen
(see PL3). This lack of regeneration is not proof that energetic
HSO does not react with N2O but is consistent with the absence
of such a reaction.
Possible Reason for the Discrepancy between Our Values
and Those of the Previous Report. We quote an upper limit
for k1 of less than 5 × 10-16 cm3 molecule-1 s-1 at 298 K.
This upper limit is 4 orders of magnitude smaller than the value
reported by Ravichandran et al.4 We offer the following
mechanism/scenario as a possible explanation for the observa-
tions of Ravichandran et al. We emphasize that we do not have
sufficient information or familiarity with their apparatus and
experiments to properly interpret their observations.
N2O hν/193 nm8 N2 + O(1D)
O(1D) + N2O f O2 + N2
(11)
(13a)
(13b)
(20)
O(1D) + N2O f 2NO
O(1D) + H2S f HSO(A2A′) + H
HSO(A2A′) + N2O f HSO(X) + N2O
(21)
In the above scheme, the rise time would increase with N2O
because both reactions 13 and 20 control that time constant. Of
course, in their experiments the signal level (due to HSO(A2A′))
may not change with N2O concentration because N2O is both
the source of O(1D) and the reactant for O(1D). Ravichandran
et al. also reported that they detected vibrationally excited
HSO(A), i.e., HSO(A2A′, V′ e 5). If the thermochemistry of

SH Radical Reaction with N2O
J. Phys. Chem. A, Vol. 103, No. 34, 1999 6785
References and Notes
HSO(X) is correct and O(1D) is thermalized in their system,
then reaction 20 cannot be the source of their observed HSO-
(A2A′, V′ e 5). However, it is worth noting that thermalizing
the translational energy of O(1D) in a system where it is reacting
almost on every collision with a reactant is difficult if only a
small amount of a thermalizing agent, such as an unreactive
bath gas, is added.
The same authors also report that vibrational excitation of
N2O to the (1, 0, 0) vibrational state raises their measured rate
constant by a factor of 2.2. It is possible that the vibrational
excitation of N2O increases the absorption of 193 nm and yields
more O(1D). However, the time constant for the decay of
HSO(A2A′) signal should not change, unless vibrationally
excited N2O quenched HSO(A2A′) more efficiently. If N2O(1,
0, 0) was produced after HS was produced and the same
enhancement was observed, our explanation of what could have
caused the faster appearance of HSO(A) would be wrong. The
main reason we proposed a plausible scheme is to show that
there are reasons to suspect the results of Ravichandran et al.
and not to unequivocally identify where their measurements or
interpretations were in error. An incorrect explanation on our
part for the large value of the rate constant reported by
Ravichandran et al. does not affect our conclusion that SH does
not react rapidly with N2O.
(1) Berresheim, H.; Wine, P. H.; Davis, D. D. Sulfur in the atmosphere.
In Composition, Chemistry and Climate of the Atmosphere; Singh, H. B.,
Ed.; Van Nostrand Reinhold: New York, 1995; pp 251-307.
(2) Tyndall, G. S.; Ravishankara, A. R. Int. J. Chem. Kinet. 1991, 23,
483-527.
(3) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. Chemical Kinetics and Photochemical Data for Use in Stratospheric
Modeling; Evaluation 12, JPL Publication 97-4; Jet Propulsion Labora-
tory: Pasadena, CA, 1997.
(4) Ravichandran, K.; Williams, R.; Fletcher, T. R. Chem. Phys. Lett.
1994, 217, 375-379.
(5) World Meteorological Organization, Report 44, Scientific Assess-
ment of Ozone Depletion: 1998; Geneva, Switzerland, 1999, p 1.3.
(6) Assuming a globally averaged H2S emission rate of 1 Tg of S per
year, a lifetime of SH of 0.3 s, and steady state in H2S and SH
concentrations, we calculate the [SH]av to be 115 molecules cm-3 at ground
level. The concentration of SH will approximately fall off linearly with
number density as the altitude increases.
(7) Biermann, H. W.; Zetzsch, C.; Stuhl, F. Ber. Bunsen-Ges. Phys.
Chem. 1976, 80, 909.
(8) Vaghjiani, G. L.; Ravishankara, A. R. J. Phys. Chem. 1989, 93,
1948-1959.
(9) Turnipseed, A. A.; Barone, S. B.; Ravishankara, A. R. J. Phys.
Chem. 1994, 97, 5926-5934.
(10) Weiner, B. R.; Levene, H. B.; Valentini, J. J.; Baronavski, A. P. J.
Chem. Phys. 1989, 90, 1403.
(11) Kakimoto, M.; Saito, S.; Hirota, E. J. Mol. Spectrosc. 1980, 80,
334-350.
Atmospheric Implications. Our upper limit of k1 at 298 K
is less than 5 × 10-16 cm3 molecule-1 s-1. This value should
be also applicable under most atmospheric temperatures. Our
small rate coefficient suggests that reaction 1 does not impact
(<4 × 10-4%) the loss rate of SH in the troposphere. Further,
this reaction should not significantly alter the calculated lifetime
of N2O in the atmosphere. Even if we take a very high globally
(12) Kurylo, M. J.; Peterson, N. C.; Brown, W. J. Chem. Phys. 1971,
54, 943.
(13) Bradley, J. N.; Trueman, S. P.; Whytock, D. A.; Zaleski, T. A. J.
Chem. Soc., Faraday Trans. I 1973, 69, 416.
(14) Mellouki, A.; Ravishankara, A. R. Int. J. Chem. Kinet. 1994, 26,
355-365.
(15) Lovejoy, E. R.; Hanson, D. R.; Huey, L. G. J. Phys. Chem. 1996,
100, 19911.
(16) Froyd, K. D.; Lovejoy, E. R. Int. J. Chem. Kinet. 1999, 31, 221.
(17) Wang, N. S.; Lovejoy, E. R.; Howard, C. J. J. Phys. Chem. 1987,
91, 5743-5749.
(18) Lee, J. H.; Stief, L. J.; Timmons, R. B. J. Phys. Chem. 1977, 67,
1705-1709.
averaged tropospheric concentration of SH of 1 × 103 cm-3
,
the lifetime of N2O with respect to this reaction would be
∼60 000 years, based on our measured upper limit for k1. Such
a slow chemical sink in the troposphere would reduce the
atmospheric lifetime of N2O by at most 0.2%. Therefore, we
conclude that the reaction of SH with N2O is not significant to
the chemistry of earth’s atmosphere.
(19) Cupitt, L. T.; Glass, G. P. Int. J. Chem. Kinet. 1975, 1, 39-50.
(20) Hung, W.-C.; Lee, Y.-P. J. Chin. Chem. Soc. 1993, 40, 407-412.
(21) Haugen, H. K.; Pence, W. H.; Leone, S. R. J. Chem. Phys. 1984,
80, 1839-1852.