FTIR studies of O(3P) atom reactions with CSe2, SCSe, and OCSe

Article abstract of DOI:10.1021/jp055113p

The overall rate coefficients of the reactions of CSe2, SCSe, and OCSe with O(³P) atom have been determined to be k_CSe2 = (1.4 ± 0.2) × 10^-10 cm³ molecule^-1 s^-1, k_SCSe = (2.8 ± 0.3) × 10^-11 cm³ molecule^-1 s^-1, and k_OCSe = (2-4 ± 0.3) × 10^-11 cm³ molecule ^-1 s^-1 at 301-303 K using Fourier transform infrared (FTIR) absorption spectroscopy. The measurements have been accomplished by calibrating against the literature value of the rate coefficient for O( ³P) with CS₂ (4 × 10^-12 cm³ molecule^-1s^-1). A product channel giving OCSe in 32.0 ± 4.2% yield has been found for the O + CSe₂ reaction. Although CO was also detected, its generation could be attributed to subsequent reactions of OCSe with O atoms. The corresponding reaction for O + SCSe gives OCS and OCSe as observable products, with their yields given as 32.2 ± 4.5 and 30.2 ± 3.3%, respectively. Computational studies using UB3LYP/aug-cc-PVTZ methods have been used particularly to determine the reaction pathways for the channels in which OCS or OCSe is produced.

Full text of DOI:10.1021/jp055113p

J. Phys. Chem. A 2005, 109, 11815-11822
11815
FTIR Studies of O(3P) Atom Reactions with CSe2, SCSe, and OCSe
Shuping Li, Tsz Sian Chwee, and Wai Yip Fan*
Department of Chemistry, National UniVersity of Singapore, 3 Science DriVe 3, Singapore 117543
ReceiVed: September 10, 2005; In Final Form: October 31, 2005
The overall rate coefficients of the reactions of CSe2, SCSe, and OCSe with O(3P) atom have been determined
to be kCSe ) (1.4 ( 0.2) × 10-10 cm3 molecule-1 s-1, kSCSe ) (2.8 ( 0.3) × 10-11 cm3 molecule-1 s-1, and
2
kOCSe ) (2.4 ( 0.3) × 10-11 cm3 molecule-1 s-1 at 301-303 K using Fourier transform infrared (FTIR)
absorption spectroscopy. The measurements have been accomplished by calibrating against the literature value
of the rate coefficient for O(3P) with CS2 (4 × 10-12 cm3 molecule-1s-1). A product channel giving OCSe in
32.0 ( 4.2% yield has been found for the O + CSe2 reaction. Although CO was also detected, its generation
could be attributed to subsequent reactions of OCSe with O atoms. The corresponding reaction for O +
SCSe gives OCS and OCSe as observable products, with their yields given as 32.2 ( 4.5 and 30.2 ( 3.3%,
respectively. Computational studies using UB3LYP/aug-cc-PVTZ methods have been used particularly to
determine the reaction pathways for the channels in which OCS or OCSe is produced.
Introduction
still of interest to compare and contrast the effect of chalcogen
substitution in CX1X2 (X ) S, Se, or O) on their reactions with
The reaction of oxygen (3P) atom with carbon disulfide, CS2,
has received considerable attention for its role in atmospheric
chemistry, especially in the elucidation of global sulfur cycle
leading to acid rain.1 Three main product channels have been
identified for this reaction:2
O(3P) atoms. FTIR spectroscopy is ideal for broadband monitor-
ing of the changes in the chalcogenide precursor and product
concentrations upon O atom reactions. The generation of O(3P)
atoms is accomplished by using UV xenon lamp photolysis of
NO2.2 Apart from examining the primary reactions, some
secondary processes such as the reactions of O atoms with the
product OCSe will be studied as well. Computational studies
of the vibrational frequencies and band strengths of the species
involved and the energies and geometries of reactive intermedi-
ates and transition states in these reactions will be determined.
However, the assumed dominant channel of the O/CSe2 system
which produces CSe and SeO could not be directly investigated
under static cell conditions. Its importance will be inferred by
measuring and then subtracting out the contributions from other
product channels. Such reactions would be best conducted using
flow cell reactors, but the difficulty of obtaining large amount
of CSe2 and SCSe precludes their study here.
O(3P) + CS2 f CS + SO(3Σ-) ∆H ) -80.7 kJ/mol
(1a)
f CO + S2(3Σ- ) ∆H ) -347.8 kJ/mol
g
(1b)
f OCS + S(3P) ∆H ) -227.4 kJ/mol
(1c)
Cheng et al. has studied the dynamics of the dominant channel
1a by examining the vibrational state distributions of CS(1Σ)
and SO(3Σ) using laser-induced fluorescence spectroscopy.3
Cooper et al. used tunable infrared diode laser absorption
spectroscopy (TDLAS) to determine a value of 2.8 for the
branching ratio of (1c)/(1b).2 Other related studies include a
study of the oxidation of CS2 at high temperatures, temperature-
dependent kinetics for the determination of the activation energy
of (1a), reactive scattering using crossed molecular beams, and
chemiluminescence where the S2 product of (1b) was identified.4-7
In addition, the kinetics of the reaction of carbon monosulfide,
CS, with O(3P) atoms were also carried out.8
However we found that oxygen atom reactions with carbon
chalcogenide species such as carbon diselenide, CSe2, and
carbon selenide sulfide, SCSe, have not been explored. Only
one related investigation where the vibrational distribution of
highly excited CO products after UV photolysis of a mixture
containing nitrogen dioxide, NO2, and CSe2 was measured.9 In
this work, Fourier transform infrared (FTIR) absorption spec-
troscopy is used to extend the scope of O(3P) atom reactions to
CSe2 and SCSe under static cell conditions. Although these Se-
containing species are less abundant in the atmosphere, it is
Experimental Section
CS2 was purchased from Ajax Chemicals and distilled before
use. Cylinders of NO, NO2, OCS, and SF6 gases were available
from Linde gas and used without further purification. Because
of the difficulty of obtaining commercial CSe2 and SCSe
samples, their syntheses have to be carried out. A literature
search revealed that CSe2 could be produced by passing
dichloromethane vapor over hot selenium powder at 700 °C
while SCSe was made by passing CS2 over iron(II) selenide at
650 °C.10,11 However attempts using these methods often
resulted in severe contamination by CH2Cl2 or CS2 despite
careful steps taken for sample purification. Therefore an
alternative method was devised in order to obtain gaseous
samples of the chalcogenides from silver selenocyanate salts.
AgNCSe was synthesized from an equimolar reaction of
potassium selenocyanate, KNCSe (97%, Aldrich), with AgNO3
aqueous solution. Typically, CSe2 was generated from the
vacuum decomposition (250-300 °C) of AgNCSe (1-2 g) in
a 15 cm long, 3 cm diameter glass cell equipped with CaF2
* Corresponding
author.
Fax:
65-67791691.
E-mail:
chmfanwy@nus.edu.sg.
10.1021/jp055113p CCC: $30.25 © 2005 American Chemical Society
Published on Web 12/06/2005

11816 J. Phys. Chem. A, Vol. 109, No. 51, 2005
Li et al.
TABLE 1: Experimental and Calculated Vibrational
windows for passage of the FTIR beam. Upon inspection by
FTIR spectroscopy, the main decomposition product was CSe2
Frequencies and Band Intensities (after Scaling) for the
Various Molecules Investigated Herea
(ν ) 1300 cm-1) with a little amount of HCN (ν ) 3314 cm-1
,
c
d
expt Ivib
calc Ivib
<1% CSe2). The latter presumably arose due to reactions with
hydrogen contaminants on the wall of the gas cell. Due to the
differences in boiling points, HCN could be gradually removed
after CSe2 was collected in a dry ice/acetone trap. It was
estimated that, during each synthesis, about 15-20 Torr of CSe2
could be generated in a total volume of 500 cm3; this amount
was sufficient for a few kinetic runs in the static cell. For the
production of SCSe, AgNCSe (0.5-3 g) and silver thiocyanate,
AgNCS (0.5-3 g, Aldrich, 99%), were mixed and decomposed
together at similar temperatures, yielding vapors containing the
three expected products, CS2 (ν ) 1520 cm-1), SCSe (ν ) 1435
cm-1), and CSe2.12 The production of exclusively SCSe was
attempted by varying the ratio of the salt precursors, but it was
unsuccessful because the resulting vapor contained either
copious amounts of CS2 at high ratios of AgNCS to AgNCSe
or very little SCSe at low ratios of AgNCS to AgNCSe.
However, contamination with CSe2 could easily be removed
by UV photodissociation at 254 nm.13 Reddish solids attributed
to selenium and carbon deposits were produced upon photolysis.
Unfortunately, CS2 could not be eliminated using the same
method or by cold traps, so experiments carried out on SCSe
always contained some amount of CS2. Despite this, the simple
method of decomposing the silver salts using their cyanate salts
offered a very convenient way of generating CSe2 or SCSe
vapor.
molecule expt νb(cm-1
)
calc ν(cm-1
)
(km/mol) (km/mol)
CO
CO2
CS2
CSe2
SCSe
OCS
OCSe
2143
2349
1535
1300
1435
2062
2034
2207
2401
1550
1322
1453
2108
2075
62
530
597
80 (65)
675 (547)
684 (554)
546 (442)
626 (507)
763 (618)
769 (623)
634
a All the bands here have been used as spectroscopic probes for
measurement of rates of reaction. The concentration limits of detection
are also given here for each molecule using the respective vibrational
band. b Except for CO, these values represent the asymmetric stretching
vibrational frequencies. c Reference 14. d Scaled intensities in paren-
theses (factor ) 0.81).
spectrum (range, 1000-3500 cm-1; resolution, 1 cm-1; average
16 scans) was collected every 1 min interval. The absorbances
of the reactants were kept low (<20%) by adjusting their vapor
pressures such that Beer-Lambert’s law could be reliably used
to form a linear relationship between absorbance and concentra-
tion. Purging of the spectrometer with N2 was carried out to
reduce interference from water and CO2. This is especially
important in experiments for which the appearance rate of CO2
in the gas cell was to be measured accurately. Whenever
possible, we used initial rate measurements to probe the kinetics
during the early stages of irradiation. Typically this duration
ranged from 0 to 5 min or slightly longer if the irradiation power
was reduced. Initial rates measurements also served to minimize
contributions from secondary reactions caused by product
buildup. Each experiment was repeated 4-6 times, and uncer-
tainty in the rate coefficient or branching ratio was quoted to
3σ deviation. The rate coefficients or product concentration
ratios in their final form, i.e., after data analysis, are listed in
Table 2 together with the previously determined rates from the
O/CS2 and O/OCS systems.
A typical photolysis experiment would begin with the addition
of carbon chalcogenide (3-10 Torr), NO2 (3-20 Torr), and
the buffer gas, SF6 (≈1 bar), into the gas cell placed in the
sample compartment of a Nicolet Nexus 870 FTIR spectrometer
in a dark room. A capacitance gauge (MKS; range, 0.01-1000
Torr; accuracy ≈ 0.005 Torr) was used for gas pressure
determination. O(3P) atoms were generated by a broadband
Xenon lamp (350-380 nm, 100 W) irradiation of NO2. The
lamp was fixed at 2 cm away from the main body of the glass
cell. The temperature recorded by a thermocouple attached onto
the outer wall of the cell showed an increase of 0.1-0.2 K
during irradiation. However the laboratory temperature variation
was larger (≈2 K), and hence the reaction temperature was
quoted as occurring in the range of 301-303 K. A fiber optic
cable placed at a fixed position of 1 cm away from the cell was
used to direct a small portion of the radiation into an Acton
Research monochromator (30 cm focal length) for monitoring
the lamp output. The photolysis of the reaction mixture would
only be initiated once a constant output of radiation from the
lamp could be maintained, as monitored using the fiber optic
cable. Unlike a laser, the broadband radiation is not expected
to induce multiphoton and secondary processes which might
affect the kinetic measurements. Nevertheless, the progress of
the reactions under strong and weak irradiation could be easily
compared by reducing the lamp power five times (≈20 W) but
at the expense of a much longer irradiation time. We found
that different irradiation power gave the same results for the
kinetic runs; hence, multiphoton processes were kept to a
minimum in our experiments. For experiments that required a
mercury (Hg) lamp of 254 nm radiation, a quartz cell of similar
size (15 cm long, 2.5 cm diameter) was used instead, while all
other conditions were maintained as usual.
Results
A. Concentration Analysis. Since the measured infrared
absorbance, Abs(ν) was related to concentration, c, via the
Beer-Lambert law (Abs(ν) ) ꢀ(ν)cl for small absorbances, l
) 15 cm path length), the molar absorptivities, ꢀ(ν), of the
molecular vibrational bands need to be known or estimated,
especially for studying the main products of the reaction.
Although the total vibrational band intensity, Ivib, values for each
of the vibrations of CS2, CO, and OCS have been determined,14
a literature search reveals scarce data for the corresponding
values for other species under investigation here, such as SCSe,
OCSe, and CSe2. The lack of data for these species may be
due to the difficulty of obtaining pure samples; nevertheless,
we have embarked on the calculation of Ivib of all relevant
species using density functional method B3LYP and a fairly
large basis set aug-cc-pvtz in Gaussian 98.15 Molecular geom-
etries were also optimized at this level of theory in order to
obtain a good match between the experimental and calculated
bond lengths and bond angles. Table 1 shows the calculated
vibrational frequencies and band intensities as well as the
previously determined experimental values, if any, for the
molecules under study here. By comparing the experimental
and computed intensities, excellent agreement could be found
for CO, CO2, CS2, and OCS if an arbitrary scaling factor of
0.81 is applied to the calculated values, with a 4.6% discrepancy
[)(1/N)(Σ(Icalc - Iexpt)/Icalc); N ) 4]. A scaling factor of <1 is
FTIR spectroscopy was used to record the decay and
appearance of vibrational bands of reactants and products upon
NO2 photolysis. Table 1 lists the vibrational frequencies of the
molecules monitored during the reaction. Typically, an IR

O(3P) Atom Reactions with CSe2, SCSe, and OCSe
J. Phys. Chem. A, Vol. 109, No. 51, 2005 11817
TABLE 2: Rate Coefficients and Branching Ratios Measured for the Reactions of O (3P) Atoms with CS2, CSe2, SCSe, and
OCSe
no.
reaction
rate coeff (cm3 molecule-1 s-1
)
percentage yield of productsa
1
2
3
O(3P) + CS2 f productsb
O(3P) + OCS f CO + SOc
O(3P) + CSe2 f products
4.0 × 10-12
(1.4 ( 0.2) × 10-10
(2.8 ( 0.3) × 10-11
[OCS]/[CO] ) 2.8
<1.0 × 10-16
[OCSe] ) 23.0 ( 4.0%
[CO] ) 9.0 ( 1.1%
[OCS] ) 32.2 ( 4.5%
[OCSe] ) 20.2 ( 3.0%
[CO] ) 10.0 ( 1.0%
[CO] ≈ 100%
4
O(3P) + SCSe f products
5
O(3P) + OCSe f CO + SeO
(2.4 ( 0.3) × 10-11
a The ratios quoted here are the ones where the effects of secondary reactions have been taken into account. Thus they are not necessarily the
raw experimental values. b Reference 2. c Reference 12.
typical of those used to scale vibrational frequencies as well.
The good match gives some confidence in using the scaling
factor for molecules for which the Ivib have not been determined
yet. The band intensity is related to ꢀ(ν) by Ivib ) ∫ꢀ(ν) δν,
where δν is the vibrational bandwidth. Since a typical bandwidth
of the molecules here is 20-40 cm-1, we have omitted the small
line width contributions from the 1 cm-1 spectral resolution of
the FTIR spectrometer. Thus the concentration can be estimated
using the formula
NO2 and CS2 pressures (5-10 Torr each); this is in good
agreement with the above literature value.
In addition, we have also investigated the reaction of O with
OCS to determine if CO might have been produced from this
reaction. The small value of its rate coefficient (<1 × 10-16
cm3 molecule-1 s-1 12
implies that CO is not expected to be
)
observed over a short irradiation time. A large pressure of OCS
was therefore used with the possibility of generating more CO
molecules for detection. However, the CO transitions were still
unobservable despite a few hours of photolysis of 5 Torr NO2
and 30 Torr of OCS. From these findings, we found that the
data collected were consistent with previous studies of these
reactions, thus lending support to the reliability of our system
for extending the scope of similar reactions.
C. Reactions of O (3P) atoms with CX1X2 (X1 ) Se; X2 )
O, S, or Se). Upon thermal decomposition of AgNCSe and/or
AgNCS, all three carbon chalcogenides could be produced. The
production of CS2 was the most facile, and hence its IR signal
was generally observed to be the strongest of the three species,
while SCSe and CSe2 tend to have similar intensities. Depending
on the amount of silver salts used, it was possible to saturate
the IR signals of all three chalcogenides. The absorbances of
all three infrared bands remained unchanged even after 2-3
days, indicating their stability toward self-decompositions and
wall decompositions.
c ) (Ivibl)-1 Abs(ν) δν
∫
where the integral represents the area under the vibrational band.
For consistency purpose, we have used either the actual
experimental value if possible or the scaled calculated values
for all molecules in the determination of the rate constants in
this work. As discussed later, the determination of many of the
rate coefficients were carried out using pseudo-first-order
approximations, which requires only a change in the concentra-
tion rather than the absolute value themselves. Hence the error
incurred in the band intensity calculations is small except in
cases where the product yields of molecules such as OCSe and
CO are required. The minimum detectable concentration change
of a molecular species could also be obtained by monitoring
its absorbances as the reaction progressed and converting the
values into concentrations using Beer-Lambert’s law. Not
surprisingly, molecules possessing intense vibrational bands such
as OCS and OCSe have the lowest detection limit of concentra-
Having produced a sufficient amount of CSe2, the reaction
of O(3P) atoms with CSe2 can be subsequently studied in the
static cell. A few control experiments were carried out initially
to investigate the effect of interference from unwanted reactions.
First, the xenon lamp radiation was directed into a sample
containing only CSe2 but no photodissociation resulted from
the process. Second, when NO2 was added to CSe2 (5 Torr) in
the cell in the absence of photolysis, the IR spectra of CO and
OCSe could not be detected, implying that the reaction of NO2
with CSe2 was very slow under our experimental conditions.
Since NO is the product of NO2 photolysis, the effect of mixing
an excess of NO (10 Torr) with CSe2 (5 Torr)) was also
examined, but no reaction was observed. It was only upon xenon
lamp photolysis in the presence of both NO2 and CSe2 that IR
signals of CO and OCSe began to appear and increased in
intensity with respect to the irradiation time. This observation
indicated that it was indeed the O atoms produced from NO2
photolysis that were responsible for generating CO and OCSe.
Figure 1 shows a typical IR spectrum taken of the reaction
mixture before and after photolysis, showing the decay of CSe2
and appearance of CO and OCSe upon O(3P) atom reactions.
Since a minor production of OCSe from the thermal decomposi-
tion of AgNCSe was unavoidable, it could be easily subtracted
from the much larger OCSe signals that came from the
tion changes of ≈0.001 Torr or 3 × 1013 molecules cm-3
.
As mentioned, experimental errors were quoted to 3σ
deviation. However for measurements that required the calcu-
lated intensity values such as product yields of OCSe, the error
or discrepancy between the calculated and experimental values
should be included. For example, the relationship between the
errors is given by (∆Z/Z)2 ) (∆A/A)2 + (∆B/B)2, where ∆Z )
overall error in the OCSe yield; Z ) average yield; ∆A/A )
experimental error in OCSe; and ∆B/B ) 0.046, which is the
error/discrepancy in calculated intensity. In all cases here, the
experimental error was larger (>10%) than the error associated
with the intensity calculations (4.6%).
B. Reactions of O (3P) with CS2 and OCS. The reaction of
O(3P) with CS2 was first carried out to test whether our apparatus
could be used to obtain reliable rate measurements. It is known
that UV photolysis of NO2 around 360-380 nm gives a unit
quantum yield of O(3P) atoms.2 Using flash kinetic infrared
spectroscopy, Cooper et al. found the [OCS]/[CO] ratio produced
in this reaction to be 2.7.2 We have used FTIR spectroscopy to
measure the relative rate of appearance of these species by
monitoring the evolution of CO (Vco ) 2143 cm-1) and OCS
(Vocs ) 2062 cm-1) absorptions. Our initial rate measurements
gave an average [OCS]/[CO] ratio of 2.8 ( 0.2 over a range of

11818 J. Phys. Chem. A, Vol. 109, No. 51, 2005
Li et al.
Figure 1. FTIR spectrum of CSe2(10 Torr) + NO2(20 Torr) in SF6
buffer gas (≈1 bar) before photolysis (a) and after photolysis (b). The
CSe2 infrared signal showed depletion, while the appearance of small
signals of OCSe and CO could be observed. The intense bands around
1600-1750 cm-1 belong to NO2.
Figure 2. (a) Plot of OCSe and CO productions vs CSe2 loss during
early stages (5 min) of UV lamp photolysis of O/CSe2; (b) plot of ln
A
CSe /ln A0,CSe versus ln ACS /ln A0,CS during UV lamp photolysis of
2
2
2
photolytic reaction. The product yields of OCSe and CO were
measured over a range of NO2 and CSe2 pressures by comparing
their appearance rates to the decay rate of CSe2,
O/CSe2. A linear relationship was obs2erved until all the CSe2 has been
completely reacted.
Since the amount of O atoms produced was far less than CS2
or CSe2, the reaction proceeds via pseudo-first-order kinetics:
RCO ) kCO/kCSe ) ∆[CO]t/∆[CSe2]t
2
ln([CSe2]0/[CSe2])
) kCSe /kCS
ROCSe ) kOCSe/kCSe ) ∆[OCSe]t/∆[CSe2]t
2
2
ln([CS2]0/[CS2])
2
ln[(A0,CSe /ꢀCSe l)/(ACSe /ꢀCSe l)]
2
2
2
2
where ∆[X]t represents the change in concentration of X at time,
t; kX is the appearance or decay rate of X. The linear plots of
CO and OCSe production versus CSe2 loss are shown in Figure
2. It was found that after repeating the experiments four to six
times, the average yields of [OCSe] and [CO] were determined
to be 23.0 ( 4.0 and 9.0 ( 1.1% during the initial stages of
irradiation.
) kCSe /kCS
2
2
ln[(A0,CS /ꢀCS l)/(ACS /ꢀCS l)]
2
2
2
2
ln(A0,CSe /ACSe ])
2
2
) kCSe /kCS
2
2
ln(A0,CS /ACS
)
2
2
where the subscript 0 denotes the concentration [X] or absor-
bance A of species X at time zero (before photolysis). A linear
relationship was indeed observed between ln(A0,CSe /ACSe ) and
The overall reaction rate constants of CSe2 with O atom could
also be measured relative to the reaction between CS2 and O
atom. This was accomplished by deliberately adding CS2 into
the cell containing NO2 and CSe2 and allowing both CSe2 and
CS2 to react competitively with O atoms. Since the rate
coefficient of the latter reaction is known previously, the rate
coefficient of the former reaction could then be determined. The
relative rate kCSe /kCS was obtained by observing the disap-
2
2
ln(A0,CS /ACS ) and the ratio kCSe2/kCS was estimated from the
2
2
2
initial linear slope of the plots shown in Figure 2b. A value for
the overall reaction rate of O with CSe2, was then determined
to be kCSe ) (1.4 ( 0.2) × 10-10 cm3 molecule-1 s-1 after
2
substituting the value for kCS ) 4.0 × 10-12 cm3 molecule-1
2
s-1.6 When the rates of change of the three species (CSe2, OCSe,
2
2
pearance of CSe2 and CS2 during the initial stages of irradiation.
and CO) were compared, a ratio of (100 ( 2):(23.0 ( 4.0):(9.0

O(3P) Atom Reactions with CSe2, SCSe, and OCSe
J. Phys. Chem. A, Vol. 109, No. 51, 2005 11819
determined that CS2 decays 7 times slower than SCSe and if
the branching ratio for the production of OCS from CS2 or SCSe
is similar, then the OCS contributions from the O/CS2 system
are ≈14% of the total amount. However, it was possible in our
experiments to synthesize about 50% more SCSe than CS2; this
reaction mixture was usually used for a better differentiation
of the OCS branching ratio. We further estimate the contribu-
tions to OCS by CS2 or SCSe by increasing the ratio of CS2 to
a fixed amount of SCSe in the reaction mixture and remeasuring
the OCSe/OCS ratio again. If the O/CS2 reaction contributes
significantly, the signals of OCS will also increase correspond-
ingly compared to OCSe, which comes only from SCSe. From
the results of varying the reactant ratios ranging from [SCSe]/
[CS2] ) 1.5:1 to 1:4, we could revise the overall [SCSe]:[OCS]:
[OCSe]:[CO] ratio to be (100 ( 2.2):(32.2 ( 4.5):(20.2 ( 3.0):
(10 ( 1.0).
The importance of secondary reactions involving OCSe was
also considered since they could contribute to the generation
of CO and perhaps CO2 upon subsequent reactions with O
atoms. The preparation of OCSe free from CSe2 or SCSe
contamination could be carried out in this manner. Starting from
a reaction mixture containing CSe2 and excess NO2, 254 nm
UV irradiation of the sample over time would result in OCSe
being formed and CSe2 being completed reacted. The major
components of the gas mixture at the end of the 254 nm
irradiation would be NO2, NO, CO, and OCSe. Further
photolysis of NO2 would then allow the reaction of O with OCSe
to proceed and its products to be monitored. A known pressure
of CS2 could also be added into the reaction mixture in a
separate experiment run so that the rate coefficient of the
reaction could be calibrated against kCS . It was found that the
2
OCSe IR signal decreased further upon reactions with O atoms
but at a rate of about 5.8 times slower compared to O and CSe2
reactions. This resulted in a rate coefficient for the O and OCSe
reactions to be kOCSe ) (2.4 ( 0.3) × 10-11 cm3 molecule-1
s-1 (Figure 4a). In addition, only the increase of CO infrared
signals (subtracted from CO already present in the mixture) was
registered without any detectable formation of CO2. Interestingly
the magnitude of the decay rate of OCSe appeared to be very
close to the appearance rate of CO, judging from a near-unity
value (0.94) for the gradient of Figure 4b.
Figure 3. (a) Plot of ln ASCSe/ln A0,SCSe versus ln ACS /ln A0,CS2 during
2
UV lamp photolysis of O/SCSe. A linear relationship was observed
until all the SCSe has been completely reacted. (b) Plots of OCSe,
OCS, and CO formation vs SCSe decay for the reaction of O + SCSe
in the presence of CS2 during photolysis.
( 1.1) was obtained for CSe2:OCSe:CO, respectively (with the
negative sign indicating a signal decay).
Discussion
The control experiments described earlier for the O/CSe2
system were also performed for the O/SCSe/CS2 system. Similar
results were obtained, whereby it was only upon photolysis of
NO2 in a mixture of CS2 and SCSe that CO, OCS, and OCSe
first appeared. Although the production of SCSe was always
accompanied by the presence of CS2, this is advantageous
because the competitive reaction of O(3P) with CS2 may again
be used as the internal standard for determining the overall rate
coefficient of O(3P) with SCSe.
The decays of equal pressures of SCSe and CS2 (5 Torr)
signals were compared, and it was found that SCSe reacted 7
times faster with O atoms, hence with an overall rate constant
of kSCSe ) (2.8 ( 0.3) × 10-11 cm3 molecule-1 s-1 (Figure
3a). The unsymmetrical SCSe molecule may react with O(3P)
to form either OCS or OCSe. The rates for the production of
OCS, OCSe, and CO were measured to be about 3, 5, and 10
times slower, respectively, if compared against the decay rate
of SCSe during the initial stages of irradiation (Figure 3b). In
other words, the [SCSe]:[OCS]:[OCSe]:[CO] ratio could also
be expressed as (100 ( 2.2):(34.5 ( 5.0):(20.2 ( 3.0):(10 (
1.0). However, OCS could also be generated from the reaction
of O with CS2 albeit with a different rate. Since it has been
The reaction rate coefficients kCSe ((1.4 ( 0.2) × 10-10 cm3
2
molecule-1 s-1) and kSCSe ((2.8 ( 0.3) × 10-11 cm3 molecule-1
s-1) have been measured to be 34 and 7 times larger than that
of kCS (4.0 × 10-12 cm3 molecule-1 s-1); this may be a result
2
of lower activation barriers for these reactions. Qualitatively, it
is not surprising to observe an increase in rates since the CdSe
bond of CSe2 or SCSe is weaker than the CdS bond of CS2
and hence yielding relatively lower energy transition states as
well. Because of the weaker CdSe bond, the O and OCSe (kOCSe
) (2.4 ( 0.3) × 10-11 cm3 molecule-1 s-1) reaction is also
observed to proceed much faster than the corresponding O and
OCS reaction (k < 1.0 × 10-16 cm3 molecule-1 s-1). Evidence
gathered by the excellent match between the decay rate of OCSe
with the appearance rate of CO upon O atom reactions suggests
that only a single channel is open and the measured rate
coefficient may be used to describe the single reaction, O +
OCSe f CO + SeO.
Two products from the O/CSe2 system were detected, i.e.,
OCSe and CO with the former being more abundant with
product yields of 23.0 ( 4.0% compared to 9.0 ( 1.1%. The
combined appearance rates of OCSe and CO are thus only a

11820 J. Phys. Chem. A, Vol. 109, No. 51, 2005
Li et al.
Figure 5. Plots of CSe2, OCSe, and CO concentrations as a function
of photolysis time in a reaction mixture containing 5 Torr CSe2, 10
Torr NO2. Arrows pointing upward indicate the tangents drawn for the
estimation of O atom concentration. Arrows pointing downward indicate
the sampling of three different times for the estimation of the
contribution of O/OCSe reaction to CO production.
determine the relative importance of these two channels, the O
atom concentrations need to be estimated first. This can be
accomplished by examining the rate of decay of CSe2 during
the early stages of irradiation (see Figure 5), where
d[CSe2]/dt ) -kCSe [O][CSe2]
2
The concentration change of CSe2 could be estimated by
drawing a tangent to its decay curve at certain time intervals,
and since both kCSe and [CSe2] are known, the O atom
2
Figure 4. (a) Plot showing the decay of OCSe against the decay of
CS2 signals upon O atom reactions. (b) Plot of CO production vs OCSe
loss during UV lamp photolysis (360-380 nm) of an O/OCSe mixture.
No CO2 was observed during the reaction.
concentration could then be determined. Calculations carried
out at three different early times as indicated in Figure 5 gave
an average value of [O] ) (5.0 ( 0.5) × 107 cm-3. As expected,
the concentration is rather low because of the low power
irradiation used for the experiments. The CO contribution from
the O/OCSe reaction could be approximated now that the O
atom concentration is known. We first assume that all CO came
from this reaction, thereby ignoring reaction 2b;
fraction, about 32%, of the CSe2 decay rate. This certainly
implies that at least another product channel is open for the
O/CSe2 system. By analogy to reaction 1a of the O/CS2 system,
the channel could be attributed to O (3P) + CSe2 f CSe +
SeO. The large discrepancy in the above measured rates would
suggest that the production of CSe and SeO is the main reaction
channel. The ratio of products generated by the O/CSe2 system
could then be expressed as CSe:OCSe:CO ) (68 ( 5):(23.0 (
4.0):(9.0 ( 1.1).
It is tempting at this stage to ascribe three reaction channels
(represented by CSe, OCSe, and CO products) for the O/CSe2
system, similar to those observed for the O/CS2 system.
However OCSe continues to react further with O atoms to yield
exclusively CO with a rather large rate coefficient. This is
certainly unlike the O/CS2 system, where the slow reaction
between O and OCS does not contribute significantly to the
production of CO. For clarity sake, the possible reactions in
the O/CSe2 system may be written as
d[CO]/dt ) k3[O][OCSe]
Since k3 has been measured and both [OCSe] and d[CO]/dt can
be evaluated directly from Figure 5, the values for the left-
hand and right-hand sides of the equation can be compared at
different times of the reaction. When comparisons were made
for three different times as indicated in the graph, it turns out
that the left- and right-hand-side values were equal and thus
the reaction of O with OCSe could perhaps fully account for
the production of CO! If this were the case, reaction 2b is at
most a minor channel on the basis of this analysis. Since some
OCSe have reacted further to generate all the CO detected in
the system, the branching ratio of reaction 2c would have been
underestimated, and hence, it should be revised accordingly to
include the CO contributions to 23% + 9% ) 32%.
Similar analysis could also be carried out on the O/SCSe
system. The product ratio of [OCS]:[OCSe]:[CO] has been found
to be (32.2 ( 4.5):(20.2 ( 3.0):(10 ( 1.0). The probability of
direct CO production from O and SCSe reaction is again in
doubt. As usual, the [O] concentration could be determined to
be 4.5 × 107 cm-3 from the early stage of irradiation for a NO2-
(15 Torr)/SCSe(7 Torr)/CS2(9 Torr) mixture. Indeed, for the
given concentration of [O], it appears that the O + OCSe
O(3P) + CSe2 f CSe + SeO(3Σ-)
(2a)
(2b)
f CO + Se2(3Σ- )
g
f OCSe + Se(3P)
O (3P) + OCSe f CO + SeO
(2c)
(3)
The CO production could come from reactions 2b and 3. To

O(3P) Atom Reactions with CSe2, SCSe, and OCSe
J. Phys. Chem. A, Vol. 109, No. 51, 2005 11821
TABLE 3: Optimized Structures and Vibrational
Frequencies of the Intermediates and Transition States in
the O/CS2, O/SCSe, and O/CSe2 Systems Determined at the
UB3LYP/aug-cc-PVTZ Level of Theorya
θ1 (deg),
θ2 (deg)
structure P, Q
a, b, c (Å)
vibrational freq (cm-1
)
A
B
C
D
E
F
G
H
I
S, S 1.95, 1.61,1.56 79, 162 431i, 247, 328, 402, 635, 1429
S, S 1.19, 1.78, 1.78 133, 94 289, 381, 508, 547, 606, 1798
S, S 1.16, 2.28, 1.64 119, 87 307i, 222, 394, 503, 750, 1992
Se, Se 2.07, 1.69, 1.75 76, 165 354i, 193, 242, 326, 378, 1240
Se, Se 1.18, 1.94,1.94 133, 94 178, 279, 450, 459, 495, 1812
Se, Se 1.16, 2.49, 1.79 118, 87 228i, 137, 321, 450, 561, 1993
S, Se 1.98, 1.56, 1.72 97, 170 473i, 177, 262, 334, 493, 1349
S, Se 1.18,1.72, 2.03 139, 93 209, 261, 457, 489, 666, 1830
S, Se 1.17, 2.30, 1.65 121, 90 211i, 205, 374, 492, 727, 1945
Se, S 1.15, 2.43, 1.78 117, 84 265i, 164, 348, 457, 576, 2020
Se, O 2.03, 1.75, 1.16 57, 179 395i, 361, 469, 637, 690, 2106
Figure 7. Energy level diagram for the O(3P) + CSe2 f OCSe +
Se(3P) potential energy surface showing the intermediate and transition
states computed at the UB3LYP/aug-cc-PVTZ level of theory.
J
K
potential energy surface. The calculated enthalpy of formation,
∆Hr, of the overall reaction is -235 kJ/mol, which puts it in
very good agreement with previously determined experimental
values.2 Several intermediates or transition states were found
along this pathway as well. The first optimized transition state
A depicts the oxygen atom approaching CS2 at an OCS angle
of about 79°. Despite several attempts, we have been unable to
find an optimized structure in which the O atom is directed
perpendicularly ( OCS ) 90°) toward the C atom. Incidentally,
a bent transition state rather than a collinear one was also found
previously for the primary channel 1a.3 An earlier study showed
that the average kinetic energy of the O(3P) atom following 355
nm photolysis corresponding to NO(X2Π,V)0) state was 1833
a The imaginary frequencies of the transition states are denoted by
i.
cm-1 15
. As the calculated activation energy turns out to be about
23 kJ/mol (1922 cm-1), it is just low enough for the reaction to
take place at rates measurable over a short time under our
experimental conditions. Not surprisingly, this value is higher
than the activation energy of the primary channel 1a, which
was estimated to be about 420 cm-1 16
The transition-state
.
Figure 6. Energy level diagram for the O(3P) + CS2 f OCS + S(3P)
potential energy surface showing the intermediate and transition states
computed at the UB3LYP/aug-cc-PVTZ level of theory.
structure also shows that the two sulfur atoms of the CS2 moiety
have bent away from the incoming O atom and will eventually
lead to the intermediate COS2 denoted as B. This intermediate
B, which sits in a relatively deep potential energy well (250
kJ/mol) with respect to structure A, is also believed to be the
precursor to channel 2b, where CO and S2 are produced.2 The
CdO bond distance of 1.19 Å for B is only slightly longer than
a typical carbonyl bond of organic molecules (1.15-1.17 Å).
Further along the pathway, a second transition state C was found
in which one of the CdS bonds lengthens as the sulfur atom of
that bond begins to pull away, leaving behind an OCS moiety
which has a quasilinear structure with an obtuse angle of 154°.
The dissociative C-S bond distance at 2.28 Å is considered
rather long but consistent with C being a late transition state
and hence productlike in structure.
As expected, similar transition states and intermediates of
the reaction pathway were also computed for the O/CSe2 system
with ∆Hr ) -323 kJ/mol (Figure 7). The first activation energy
leading to structure D for this system is only 2.1 kJ/mol. This
almost barrierless reaction explains the experimental data in
which the production of OCSe from CSe2 is >37 times faster
than the production of OCS from CS2. The transition states, D
and F, also closely resemble their sulfur counterparts in structure.
The intermediate COSe2 denoted by E occupies a deep potential
well relative to D and forms the lowest energy state of the
reaction pathway for the O/CSe2 system.
secondary reactions could account for the entire CO produced
during the initial stages of photolysis. Thus, the ratio of [OCS]:
[OCSe]:[CO] is revised to 1.6:1.5:≈0. However the total
appearance rate of these species could only account for 62% of
the decay rate of SCSe. Similar to the O/CS2 system, two other
product channels could be assigned to the production of SeO
and CS, and SO and CSe. Unfortunately the relative importance
of the two channels could not be separately determined here;
hence only one branching ratio of 38 ( 5% will be used to
describe the contribution of these channels.
D. Computational Studies. Ab initio computational studies
have been performed using Gaussian 98 to further understand
and substantiate the experimental findings. Since OCSe has been
detected as the main observable product in this work, we place
emphasis on the computation of the (1c), (2c), and related
reaction pathways for the O/SCSe system. The density functional
theory (DFT) method of UB3LYP with aug-cc-PVTZ basis sets
has been used for the optimization of the structures and energies
of reactive intermediates and transition states in all the reactions
of O(3P) with CS2, CSe2, and SCSe (Table 3).16
Comparisons with the experimental values of the enthalpies
of formation of the three channels of the O/CS2 system were
first made to check the reliability of the calculations. As shown
in Figure 6, the reaction pathway leading to OCS and S (channel
1c) has been calculated for the O/CS2 system on a triplet
For the O/SCSe system shown in Figure 8, the enthalpies of
reaction for the production of OCSe and S and for the production

11822 J. Phys. Chem. A, Vol. 109, No. 51, 2005
Li et al.
Summary
We have determined the rate coefficients of the reactions of
the carbon chalcogenides, CSe2, SCSe, and OCSe with O(3P)
atom to be kCSe ) (1.4 ( 0.2) × 10-10 cm3 molecule-1 s-1
,
)
2
kSCSe ) (2.8 ( 0.3) × 10-11 cm3 molecule-1 s-1, and kOCSe
(2.4 ( 0.3) × 10-11 cm3 molecule-1 s-1 at 301-303 K using
the technique of Fourier transform infrared (FTIR) absorption
spectroscopy. These measurements have been carried out by
using the known value of the rate coefficient for O(3P) with
CS2 (4 × 10-12 cm3 molecule-1 s-1) as the internal calibrant.
A main product channel yielding OCSe has been found for the
O + CSe2 reaction, similar to the O/CS2 system. CO was also
observed in the O/CSe2 system, although its generation could
be attributed to subsequent reactions of OCSe with O atoms.
The corresponding reaction for O + SCSe gives roughly equal
amounts of OCS and OCSe as products, although the dominant
channel is still inferred to be the one that produces CSe and
SO, and CS and SeO. Ab initio studies using density functional
methods such as UB3LYP/aug-cc-PVTZ have been used
particularly to determine the reaction pathways for the channels
in which OCS or OCSe is produced.
Figure 8. Energy level diagram for the O(3P) + SCSe f OCS +
Se(3P) and OCSe + S(3P) potential energy surfaces showing the
intermediate and transition states computed at the UB3LYP/aug-cc-
PVTZ level of theory.
of OCS and Se have been determined to be -238 and -318
kJ/mol, respectively. Interestingly, these values are very close
to the corresponding values for the O/CS2 and O/CSe2 systems,
indicating that the strength of a particular CdS or CdSe bond
is almost the same in any of the carbon chalcogenides. The
calculated activation barrier leading to its first transition state
G is 9.3 kJ/mol, which is between the values for the O/CS2 and
O/CSe2 systems. This again agrees with the experimental data
in which decay of SCSe to OCS and OCSe is intermediate in
rate compared to CS2 and CSe2. The transition state shows the
O atom being located at almost right angle to the SCSe moiety.
The intermediate COSeS denoted by H also possesses a
relatively deep potential well, although this time it is no longer
the lowest energy state of the reaction pathway. Further along
the pathway, we managed to locate two more transition states,
I and J which correspond to either the cleavage of the C-S
bond to produce OCSe or the cleavage of the C-Se bond to
produce OCS. As expected, the latter transition state, J is
energetically lower by 53 kJ/mol since cleavage of the weaker
CdSe bond is easier to achieve. However both transition states
I and J are still lower in energy compared to G and may not
play a critical role in determining the OCS/OCSe branching
ratios. Experimental observations showed that the appearance
rate of OCS is only slightly higher than the corresponding rate
for OCS.
Acknowledgment. S.L. acknowledges a research studentship
from the National University of Singapore. We thank P. Li for
valuable discussions. This work is supported under Grant Nos.
143-000-231/210-112 from the Faculty of Science, National
University of Singapore.
References and Notes
(1) Tyndall, G. S.; Ravishankara, A. R. Int. J. Chem. Kinet. 1991, 23,
483.
(2) Cooper, W. F.; Hershberger, J. F. J. Phys. Chem. 1992, 96, 5405.
(3) Cheng, Y.; Han, J.; Chen, X.; Ishikawa, Y.; Weiner, B. R. J. Phys.
Chem. A 2001, 105, 3693.
(4) Murakami, Y.; Kosugi, M.; Susa, K.; Kobayashi, T.; Fujii, N. Bull.
Chem. Soc. Jpn. 2001, 74, 1233.
(5) Wei, C. N.; Timmons, R. B. J. Chem. Phys. 1975, 62, 3240.
(6) Rochford, J. J.; Powell, L. J.; Grice, R. J. Phys. Chem. 1995, 99,
15369.
(7) Naik, P. D.; Pavanaja, U. B.; Sapre, A. V.; Ramarao, K. V. S.;
Mittal, J. P. Chem. Phys. Lett. 1991, 186, 565.
(8) Smith, I. W. M. Discuss. Faraday Soc. 1967, 43-44, 194.
(9) Morley, C.; Ridley, B. A.; Smith, I. W. M. J. Chem. Soc., Faraday
Trans. 2 1972, 68, 2127.
Rochford et al. has indicated from the transition-state structure
that an angular attack of the O atom on OCS has taken place
with an activation barrier of 24 kJ/mol for the production of
CO and SO.6 We have also assumed similar reaction pathways
for the reaction of O with OCSe. Unfortunately an optimized
transition-state structure could not be found despite a few
attempts. Nevertheless, it is not surprising to expect the barrier
to be lower than the O/OCS case by virtue of the weaker Cd
Se bond. Hence the reaction of O with OCSe should proceed
much faster and in fact account for CO production in the system.
However, CO2 and Se may also be generated from the reaction
of O with OCSe with a large release of energy (∆Hr ) 375
kJ/mol, also computed using UB3LYP/aug-cc-PVTZ). By virtue
of its exothermicity, this channel could be deemed important
although CO2 could not be observed in the experiment. Hence,
the same UB3LYP/aug-cc-PVTZ computations were carried out
to trace its reaction pathway, and indeed, the transition state K
was found to be located at about 85 kJ/mol higher than the
reactants O and OCSe. The large value for the activation barrier
appears to have hindered the formation of CO2 under our
experimental conditions.
(10) Pan, W. H.; Fackler, J. P., Jr. Inorg. Synth. 1982, 21, 6.
(11) Henriksen, L. Synthesis 1985, 2, 204.
(12) NIST Standard Reference Database No. 69, March, 2003. Release
at: http://webbook.nist.gov/chemistry.
(13) Callear, A. B.; Tyerman, W. J. R. Trans. Faraday Soc. 1965, 61,
2395.
(14) Molecular Spectroscopy: Modern Research; Rao, K. N., Mathews,
C. W., Eds.; Academic Press: New York, 1972-1985; Vol. 2, p 179.
(15) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-
Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe,
M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.;
Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. GAUSSIAN
98; Gaussian Inc.: Pittsburgh, PA, 1998.
(16) Hsu, D. S. Y.; Shaub, W. M.; Burks, T. L.; Lin, M. C. Chem. Phys.
1974, 44, 143.