Article

Cite This: J. Phys. Chem. A 2019, 123, 6130−6143 pubs.acs.org/JPCA

Infrared Emission from Photodissociation of Methyl Formate

[HC(O)OCH3] at 248 and 193 nm: Absence of Roaming Signature
Published as part of The Journal of Physical Chemistry virtual special issue “Hai-Lung Dai Festschrift”.
Lucia Lanfri,† Yen-Lin Wang,‡ Tien V. Pham,‡,¶ Nghia Trong Nguyen,¶,§ Maxi Burgos Paci,†
M. C. Lin,*,‡,§ and Yuan-Pern Lee*,‡,§,∇
†
Facultad de Ciencias Químicas, Universidad Nacional de Córdoba, Ala I - 2do Piso Ciudad Universitaria, Pabellón, Argentina
‡
Department of Applied Chemistry and Institute of Molecular Science, National Chiao Tung University, Hsinchu 30010, Taiwan
¶
School of Chemical Engineering, Hanoi University of Science and Technology, Hanoi, Vietnam
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

§
Center for Emergent Functional Matter Science, National Chiao Tung University, Hsinchu 30010, Taiwan
∇
Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 10617, Taiwan
Downloaded via NATL TAIWAN UNIV on April 29, 2020 at 18:29:35 (UTC).

*
S Supporting Information

ABSTRACT: Following photodissociation at 248 nm of gaseous

methyl formate (HC(O)OCH3, 0.73 Torr) and Ar (0.14 Torr),
temporally resolved vibration−rotational emission spectra of
highly internally excited CO (ν ≤ 11, J ≤ 27) in the 1850−2250
cm−1 region were recorded with a step-scan Fourier-transform
spectrometer. The vibration−rotational distribution of CO is
almost Boltzmann, with a nascent average rotational energy (E0R)
of 3 ± 1 kJ mol−1 and a vibrational energy (E0V) of 76 ± 9 kJ
mol−1. With 3 Torr of Ar added to the system, the average
vibrational energy was decreased to E0V = 61 ± 7 kJ mol−1. We
observed no distinct evidence of a bimodal rotational distribution
for ν = 1 and 2, as reported previously [Lombardi et al., J. Phys.
Chem. A 2016, 129, 5155], as evidence of a roaming mechanism. The vibrational distribution with a temperature of ∼13000 ±
1000 K, however, agrees satisfactorily with trajectory calculations of these authors, who took into account conical intersections
from the S1 state. Highly internally excited CH3OH that is expected to be produced from a roaming mechanism was
unobserved. Following photodissociation at 193 nm of gaseous HC(O)OCH3 (0.42 Torr) and Ar (0.09 Torr), vibration−
rotational emission spectra of CO (ν ≤ 4, J ≤ 38) and CO2 (with two components of varied internal distributions) were
observed, indicating that new channels are open. Quantum-chemical calculations, computed at varied levels of theory, on the
ground electronic potential-energy schemes provide a possible explanation for some of our observations. At 193 nm, the CO2
was produced from secondary dissociation of the products HC(O)O and CH3OCO, and CO was produced primarily from
secondary dissociation of the product HCO produced on the S1 surface or the decomposition to CH3OH + CO on the S0
surface.

I. INTRODUCTION Pereira and Isolani employed infrared multiphoton dissoci-

Methyl formate, HC(O)OCH3, plays important roles in many
aspects, including astrochemistry and organic chemistry. Since
its discovery in the emission spectrum of Sgr B2 in 1975,1
methyl formate was observed in significant amounts in varied
astrophysical environments,2−6 with reported column densities
in the range of (2.6−19) × 1016 cm−2.7 To understand the
observed abundance of methyl formate and its isotopic
fractionation, reaction paths for the formation and destruction
of methyl formate should be considered. Furthermore,
HC(O)OCH3 is the simplest ester and is commonly employed
in organic synthesis. Its catalytic conversion to synthetic
gases8−11 or its pyrolysis to produce CO has been extensively
explored.12,13
ation (IRMPD) of HC(O)OCH3 to detect CO, CO2, CH3OH,
CH4, and C2H6.14 Based on the observed products, they
proposed that the dissociation of HC(O)OCH3 proceeds via
two channels to form CH3OH + CO and CH4 + CO2.14
Francisco employed the CCSD(T)/6-311G(2df,2p)//MP2/6-
311G(2df,2p) method to investigate the decomposition of
HC(O)OCH3 and reported that the experimental products of
decomposition (CH3OH, H2CO, and CO) can be explained
by two competitive parallel reactions forming CH3OH + CO
and H2CO + H2CO, followed by decomposition of H2CO to
Received: May 1, 2019
Revised: July 3, 2019
Published: July 3, 2019

© 2019 American Chemical Society 6130 DOI: 10.1021/acs.jpca.9b04129

J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A
yield CO + H2;15 however, the channel leading to CH4 + CO2
has a large barrier. Lee employed a molecular beam to
investigate the photodissociation of DC(O)OCH3 at 193 nm
and reported four dissociation channels to form CH3O + DCO
(X ∼ 2A′), CH3O + DCO (A ∼ 2A″), CH3OCO + D, and
CH3OD + CO, with estimated branching factors 0.73, 0.06,
0.13, and 0.08, respectively;16 products DCO (A ∼ 2A″) and
C(O)OCH3 further decomposed to D + CO and CH3 + CO2,
respectively.
Chao et al. studied the photodissociation of HC(O)OCH3
at 234 nm using velocity ion imaging coupled with (2 + 1)
resonance-enhanced multiphoton ionization (REMPI) to
detect the CO fragment.17 They selected varied rotational
levels of CO (ν″ = 0) to evaluate the translational and internal
energy disposals in the fragments. Two distinct distributions of
translational energy of CO were reported: for J″ ≥ 24, a broad
distribution resulted from the CH3OH + CO channel with an
average translational energy of 96 kJ mol−1, extending to 293 kJ
mol−1, was reported; for smaller J″, a bimodal distribution
having an additional sharp, low-energy component, with an
energy of 4−8 kJ mol−1, that increased with decreasing J″ was
observed. These authors ascribed this additional low-energy
channel to a roaming mechanism via the radical channel CH3O
+ HCO to form CH3OH + CO. de Wit et al. employed the
phase-space theory (PST) to investigate the photolysis of
methyl formate and indicated that the dynamical signature of
the roaming donor fragment is similar to that of the triple
fragmentation.18 These authors indicated that, although the
small energy distribution of CO might arise from a roaming
reaction, CO from triple fragmentation cannot be discounted;
therefore, finding an unambiguous way to determine the true
mechanism is prevailing.
Tsai et al.19 further investigated the system with photo-
dissociation wavelengths of 225−250 nm to encompass the
theoretical threshold for the triple fragmentation channel lying
near 480.9 kJ mol−1 or 248 nm. With the help of trajectory
dynamic simulations, these authors deconvoluted their
observed kinetic distribution of CO into three components
to correlate a slow component due to the opening of the triple
fragmentation channel, an intermediate one associated with the
roaming mechanism, and a rapid one from the minimal energy
path of the molecular channel. To confirm these results,
Lombardi et al.20 studied the photolysis of HC(O)OCH3 at
248 nm with a time-resolved Fourier-transform infrared
(FTIR) spectrometer to record the emission spectrum of
vibration−rotationally excited CO (ν ≤ 4) and reported two
distinct rotational distributions with rotational temperatures of
∼1000 and 450 K in the ν = 1 and ν = 2 vibrational states;
these two rotational distributions were assigned to the direct
molecular channel and the roaming path, respectively,
according to their dynamical simulations.
Although all reports from these groups of Lin and
Aquilanti17,19,20 seem to point consistently to the existence
of a roaming mechanism for the formation of CO, truly
unambiguous experimental results to support this mechanism,
such as an observation of highly internally excited CH3OH and
experimental data with improved ratios of signal-to-noise to
identify clearly a bimodal distribution of CO, are lacking. The
rotational temperatures estimated from ion-imaging experi-
ments (∼200 K for roaming and ∼420 K for a molecular
channel) are distinct from those derived from the IR emission
experiments (∼450 K for roaming and ∼1000 K for molecular
channel).21 Furthermore, significant vibrational excitation of
6131
Article
CO was predicted from their simulation, but only up to ν = 4
was observed in their experiments.
In this work, we investigated the infrared emission from
products of HC(O)OCH3 irradiated at 248 or 193 nm with a
step-scan Fourier-transform spectrometer. At 248 nm, we
observed emission of CO up to ν = 11, but with no clear
evidence of a small-J component from the roaming channel
and no emission of CH3OH above 1800 cm−1. At 193 nm, an
intense broad feature of CO2 with bimodal distributions and
weaker CO emission with a smaller vibrational excitation (ν ≤
4) were observed. These observations can be rationalized with
quantum-chemical calculations and rate predictions.
II. METHODS
II.A. Experiments. To acquire time-resolved IR emission
spectra, we employed a step-scan FTIR spectrometer coupled
to a vacuum chamber containing a set of Welsh mirrors to
maximize the efficiency of collection of the emission induced
from the ultraviolet (UV)-irradiated flowing gaseous samples.
Since this system has been described with greater details in
previous publications,22−24 only a summary is given here.
The mildly focused photolysis beam, emitted from a KrF
excimer laser (Coherent Model COMPexPro-102F, 248 nm,
17 Hz, 160 mJ pulse−1) or an ArF excimer laser (Coherent
Model COMPexPro-50, 193 nm, 50 Hz, 60 mJ pulse−1),
entered the chamber and propagated in front of the two Welsh
mirrors near the FTIR spectrometer side. The laser beam has
dimensions ∼2.0 × 15.0 mm2 at the photolysis center. An InSb
detector was employed to detect the transient signal, which
was preamplified and further amplified by a factor of 50 with a
voltage amplifier (bandwidth = 1 MHz) before being recorded
with a data-acquisition board (12-bit, 25 ns). Survey spectra
were obtained at a resolution of 6 cm−1 in the spectral region
of 1800−4900 cm−1. To detect the emission of CO at a
resolution of 0.3 cm−1, a filter allowing the passage of light in
the region of 1670−2325 cm−1 was employed; this filtering
enables undersampling to decrease the number of data points
and, hence, the acquisition period. The CO emission was
recorded effectively in the region of 1850−2325 cm−1, because
the lower bound was limited by the detectivity of the InSb
detector. Similarly, to investigate the emission in other spectral
regions, we employed other filters. Temporal profiles at each
mirror step of the FTIR analysis were typically averaged over
30 laser pulses. For analysis with a satisfactory signal-to-noise
(S/N) ratio, 40 consecutive time slices obtained at 25 ns
intervals were typically summed to yield spectra at intervals of
1 μs. In addition, the S/N ratio was improved by averaging 4−
6 sets of spectra recorded under similar experimental
conditions. The spectra were corrected for the transmission
of the filter employed and the instrumental spectral-response
function that was obtained on calibration with blackbody
radiation.
Methyl formate (Alfa Aesar, 97%) was used without
purification, except for degassing through freeze−pump−
thaw cycling. The sample was injected into the vacuum
chamber as a diffusive beam through a slit-shaped inlet.
Typically no buffer gas was added except Ar (AGA Specialty
Gases, 99.999%) to a small extent, which was flowed near the
entrance of the photolysis port to avoid a solid deposit on the
quartz window. The partial pressure of the gases inside the
chamber was maintained typically at 0.42−0.73 and 0.09−0.14
Torr for HC(O)OCH3 and Ar, respectively, so that satisfactory
spectra could be obtained with a minimal pressure. In one set
DOI: 10.1021/acs.jpca.9b04129
J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A
of experiments, 3.0−3.1 Torr of Ar was added to 0.9−1.0 Torr
of HC(O)OCH3 to mimic likely experimental conditions of
the work reported by Lombardi et al.20 Calibrated mass
flowmeters and controllers were employed to determine the
respective flows. Typical flow rates are 0.23−0.45 and 0.04−
1.20 STP cm3 s−1 for HC(O)OCH3 and Ar, respectively. [STP
represents standard temperature (298 K) and pressure (1
atm).]
II.B. Computational Methods. The potential energy
surface (PES) for the dissociation reaction was constructed
using varied methods. The species of interest were initially
optimized with the WB97X/6-311++G(3df,2p) method,25
followed by further refinement with MP2/6-311++G-
(3df,2p),26 B3LYP/6-311++G(3df,2p),27−29 M06-2X/6-311+
+G(3df, 2p),30 and CCSD/6-311++G(d,p).31 To obtain more-
reliable energies, we used the CCSD(T) method32 to calculate
single-point energies of all species involved.
For the loose transition structure (RTS), the energy was
obtained with the CASPT2(10,8)//CASSCF(10,8) /6-
311+G(d,p) method,33−36 similar to those of the variational
transition states (VTS) in bond breaking.37,38 Vibrational
analysis at each stationary point verified the state as either a
local minimum or a transition state. Intrinsic reaction-
coordinate (IRC) calculations39 were used to link each
transition structure with the corresponding intermediates. All
quantum-chemical computations were conducted with the
Gaussian-0940 and Molpro41 program packages.
The energies and the vibration−rotational parameters used
in the calculations of rate coefficients were based on the
highest level of theory, CCSD(T)/6-311++G(3df,2p)//
CCSD/6-311++G(d,p), employed in this work. The energy-
dependent microcanonical rate coefficients for the low-lying
channels were calculated with the RRKM theory. For the
barrierless channels, the Variflex code38 was used to calculate
the conventional rate coefficients, which were used as inputs
for calculations of the microcanonical rate coefficient, as
implemented with the Multiwell code.42 In the kinetic
calculations, vibrational wavenumbers were computed at the
CCSD/6-311++G(d,p) level and treated as a rigid-rotor
harmonic oscillator (RRHO). The effects of hindered internal
rotation associated with the CH3 and CH3O moieties were
included when calculating the partition function for these
small-wavenumber modes. The torsional potential for the
hindered internal rotation of the reactant cis-HC(O)OCH3
and the intermediate state, HOCOCH3, calculated at CCSD/
6-311++G(d,p), are shown in Figures S1 and S2 in the
Supporting Information.
III. RESULTS AND ANALYSIS
The experimental conditions were typically selected to obtain
spectra with a satisfactory S/N ratio for accurate analysis,
which requires a substantial amount of HC(O)OCH3, while
minimizing the effect of collisional quenching on maintaining
the total pressure as small as practical, so that nascent
distributions of internal states of photolysis products were
derivable with only small corrections.
The small absorption cross section of HC(O)OCH3 at 248
nm, with a value of 5.34 × 10−21 cm2 molecule−1,43 requires a
minimal HC(O)OCH3 partial pressure of ∼0.74 Torr for
spectra with a satisfactory S/N ratio. To ensure a single-photon
absorption condition, the power dependence of the emission
signal was evaluated for laser fluence 626−1060 mJ cm−2, as
shown in the log−log plot of Figure S3 in the Supporting
6132
Article
Information; the order of the power dependence was
determined to be 1.02 ± 0.06. The laser fluence employed
in most experiments was <730 mJ cm−2.
At 193 nm, a HC(O)OCH3 partial pressure of ∼0.4 Torr
was required for satisfactory spectra; the absorption cross
section of HC(O)OCH3 is 9.80 × 10−20 cm2 molecule−1 at this
wavelength.43 The fluence employed in most experiments was
<200 mJ cm−2.
III.A. Experiments with HC(O)OCH3/Ar (0.73/0.14
Torr) Irradiated at 248 nm. III.A.a. Survey Emission
Spectra in the Region of 1800−4500 cm−1. Traces (a)−(c)
in Figure 1 present survey emission spectra recorded at a
Figure 1. Emission spectra in the spectral region of 1800−4900 cm−1
recorded 0−1 μs (spectrum (a)), 10−15 μs (spectrum (b)), and 20−
25 μs (spectrum (c)) upon photolysis of a flowing mixture of
HC(O)OCH3 (0.73 Torr) and Ar (0.14 Torr) at 248 nm. Spectral
resolution is 6 cm−1; 30 laser pulses were averaged. Spectrum (d)
corresponds to the region between 4400 and 2400 cm−1 recorded at a
resolution of 1 cm−1. Absorption spectra of H2CO (spectrum (e)),
and CH3OH absorption and HCO emission23 spectra (spectrum (f))
are shown for comparison.
resolution of 6 cm−1 in the spectral region of 1800−4500 cm−1
at intervals of 0−1, 10−15, and 20−25 μs after photolysis of
the flowing mixture of HC(O)OCH3 (0.73 Torr) and Ar (0.14
Torr). The emission in the region of 1850−2200 cm−1 is due
to the unresolved fundamental emission band of CO. The
intensity of this feature remained almost unchanged during the
first 25 μs, consistent with the reported slow vibrational
quenching of CO.44 Two extremely weak features in the
regions of 2600−3000 cm−1 and 3700−4300 cm−1 were also
observed (Figure 1a), but their small intensities and poor
resolution prevented a clear identification of their carriers from
these spectra. The band in region 3700−4300 cm−1 decayed
insignificantly with time, whereas that in the region of 2600−
3000 cm−1 decayed more rapidly.
Spectrum (d) in Figure 1 is a spectrum recorded 0−5 μs
upon irradiation in the region of 2400−4400 cm−1 at a
resolution of 1 cm−1 under similar experimental conditions.
The bands in the region of 3700−4300 cm−1 are resolved and
analysis of these features clearly indicates that they correspond
to the overtone emission Δυ = −2 of CO. A comparison of
observed and predicted spectra is presented in Figure S4 in the
DOI: 10.1021/acs.jpca.9b04129
J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A Article

Figure 2. Emission spectra of CO in the spectral region of 1850−2250 cm−1 recorded 0−1 μs (spectrum (a)) and 3−4 μs (spectrum (b)) upon
photolysis of a flowing mixture of HC(O)OCH3 (0.73 Torr) and Ar (0.14 Torr) at 248 nm. Spectral resolution = 0.3 cm−1; 30 laser pulses were
averaged.

Supporting Information. Spectrum (e) in Figure 1 is an

absorption spectrum of H2CO recorded in this laboratory. The
band position is near the weak feature that we observed in the
region of 2600−3000 cm−1 and also in the emission band of
H2CO from the reaction of C2H3 with O2.45 However, definite
assignment of this feature to H2CO or other species is
impracticable, because of its weakness. Spectrum (f) in Figure
1 is the absorption spectra of two other possible products:
CH3OH (blue trace) and HCO (black trace).23 The rapidly
decaying weak feature near 2300 cm−1 might be associated
with internally excited HCO, but a definite assignment is
unachievable. Upon comparison with the absorption spectrum
of CH3OH, we were unable to find emission of CH3OH in this
spectral region.
III.A.b. Emission of CO. Figure 2 shows representative
emission spectra in the region of 1850−2250 cm−1 obtained
0−1 μs and 3−4 μs after photolysis of HC(O)OCH3 in argon,
recorded at a resolution of 0.3 cm−1. These spectra exhibit an
excellent S/N ratio. Upon comparison of these two spectra,
rotational quenching is only slightly perceptible. Figure S5 in
the Supporting Information shows the emission spectra,
together with a stick diagram indicating each transition. Each
observed vibration−rotational line was assigned to a transition
according to spectral parameters reported by Ogilvie et al.46
We observed emission from levels of CO up to J ≤ 27 and ν ≤
11.
Each line was integrated and divided by the Einstein
coefficient for the corresponding transition to obtain the
relative population Pv(J) for each vibration−rotational level, in
which ν and J denote the vibrational and rotational quantum
numbers of the emitting (upper) state, respectively. For
partially overlapped lines, we obtained their individual
intensities through curve fitting. For severely overlapped
lines, no datum was analyzed, because curve fitting might
yield large errors. Figure 3 presents semilogarithmic plots of
Pv(J)/(2J + 1), as a function of rotational energy ER for the
spectrum recorded 0−1 μs after photolysis for each vibrational
level (ν = 1−10) of CO; error bars of each Pv(J) were
estimated by taking into account the error induced by the
baseline due to noise. A reliable linear fit could not be achieved
for vibrational level ν = 11, because of small signal strengths;
therefore, we excluded ν = 11 from our analysis. From Figure
3, the rotational distribution of CO appeared to be Boltzmann-
like. The rotational temperature (TR) for each vibrational state
was fitted and is listed in Figure 3 and Table 1. The population
values that were missing because of severely overlapped lines
6133
Figure 3. Relative rotational populations of CO (ν = 1−10) recorded
0−1 μs after photolysis at 248 nm of a flowing mixture of
HC(O)OCH3 (0.73 Torr) and Ar (0.14 Torr). Solid lines represent
least-squares fits. Each dataset was shifted downward by 10−(ν−1) (for
ν = 2−5) and 10−(ν−6) (for ν = 7−10), respectively, for clarity.
were estimated by performing an interpolation within the fitted
linear curve. The average rotational energy for each vibrational
level listed in Table 1 was obtained by summing the energy of
each rotational level multiplied by its normalized population.
Similar procedures were undertaken for spectra averaged
over 1−2, 2−3, and 3−4 μs (Figures S6−S8 in the Supporting
Information). The derived rotational temperatures TR and
averaged rotational energies ER for CO (ν = 1−10) in these
four periods are listed in Table S1 in the Supporting
Information. Even though we employed the smallest
practicable pressure, we must consider the effect of rotational
quenching to calculate the nascent rotational energy (E0R). For
this purpose, we plotted the variations in rotational temper-
ature TR for each vibrational level, as a function of time (Figure
S9 in the Supporting Information). Assuming an exponential
decay of TR to 298 K, we fitted the decay (red lines) and
extrapolated the rotational temperature to time zero to obtain
the T0R value; the estimated T0R values are 620 ± 120, 390 ± 20,
400 ± 10, 450 ± 20, 410 ± 20, 390 ± 20, 360 ± 10, 370 ± 10,
470 ± 10, and 470 ± 60 K for ν = 1−10, respectively; the
errors reflect only the standard deviations of fitting. The
average ratio of the nascent rotational temperature to those
DOI: 10.1021/acs.jpca.9b04129
J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A Article

Table 1. Summary of Observed Rotational and Vibrational Temperatures and Average Energies, and Vibrational Populations
Pv of CO Produced upon Photolysis of HC(O)OCH3 (0.73 Torr) and Ar (0.14 Torr) at 248 nm
Rotation Vibration
ν TRa [K] TRb [K] ER [kJ mol−1] Pva [%] Pv × EV [kJ mol−1] TV [K] TVd [K]
13160 ± 850 10000
0 [20.5] 0.0
1 480 ± 140 470 ± 50 (1080 ± 70) 3.2 16.6 4.2
2 390 ± 40 430 ± 50 (920 ± 50) 3.1 13.8 7.1
3 390 ± 40 (800 ± 50) 3.2 11.6 8.8
4 440 ± 30 (730 ± 40) 3.3 9.2 9.3
5 390 ± 30 2.8 8.7 10.8
6 370 ± 20 2.9 6.6 9.8
7 350 ± 50 2.9 5.0 8.6
8 370 ± 30 2.8 3.6 7.2
9 430 ± 110 2.8 2.5 5.5
10 440 ± 110 2.8 1.9 4.4
average E 3.1 ± 0.3 76 ± 9
nascent E0c 3.3 ± 0.6 76 ± 9
a
Observed 0−1 μs after irradiation. The Pv value of ν = 0 was estimated by fitting the Boltzmann distribution. bData taken from ref 20 for the
small-J component; TR values for the large-J components are listed in parentheses. cExtrapolated to t = 0; see text. dData taken from ref 20;
predicted value.

determined for the 0−1 μs period is 1.06 ± 0.11, indicating
that the correction for quenching is only ∼6%.
We calculated the relative vibrational population for each ν
as Pv = ∑J Pv(J) by counting all levels up to the observed
maximal J values. A semilogarithmic plot of relative population
Pv, as a function of vibrational energy EV, indicated an almost-
linear relationship. Assuming a Boltzmann vibrational dis-
tribution, we estimated the relative vibrational population of
CO(ν = 0) to be 1.23 ± 0.09 times greater than that of CO(ν
= 1). By taking into account the relative population estimated
for CO (ν = 0), we normalized the values of Pv to yield the
normalized vibrational population distribution, as listed in
Table 1 and depicted in Figure 4 (red open square); the
extrapolated population of CO(ν = 0) is presented with a red
solid square. The fitted vibrational temperature of CO is 13160
Figure 4. Vibrational distributions of CO observed after photolysis at
248 nm. [Legend: red square, total pressure = 0.87 Torr; black circle,
total pressure = 4.0 Torr. Population of ν = 0 indicated with the filled
symbol was derived by extrapolation according to the fitted
Boltzmann distribution, shown as solid lines. Blue triangle:
experimental data from Lombardi et al.;20 solid blue line represents
their trajectory calculations.]
6134
± 850 K, and the average vibrational energy is EV = 76 ± 9 kJ
mol−1 (see Table 1); the error limit takes into account of
possible contributions of unobserved higher vibrational levels.
As vibrational quenching of CO during period 0−1 μs is
negligible, the nascent vibrational energy is E0V = 76 ± 9 kJ
mol−1.
The average rotational energy was then derived by summing
the rotational energy and the associated normalized vibrational
population over all vibrational levels. An average rotational
energy 3.1 ± 0.3 kJ mol−1 for 0−1 μs was thus derived; the
level ν = 0 was excluded because of the nature of the emission
experiment. Hence, the estimated nascent rotational energy for
the CO fragment is E0R = 3.3 ± 0.6 kJ mol−1 (Table 1), using a
correction factor of 1.06 ± 0.15, derived from the rotational
quenching.
III.B. Experiments with HC(O)OCH3/Ar (1.00/3.00
Torr) Irradiated at 248 nm. We tried to perform
experiments under conditions similar to Lombardi et al.,20
but could not find the detailed experimental conditions from
the reference. The experimental condition for the reported IR
emission spectrum of HCO was recorded with an unknown
amount of HC(O)OCH3 and 3 Torr of Ar.19 Hence, we also
performed experiments with HC(O)OCH3/Ar (1.00/3.00
Torr) at 248 nm to mimic their experimental conditions and
tried to understand the effect of the added quencher at higher
pressure.
Detailed discussions are presented in Section A of the
Supporting Information, including Figures S10−S12. The
results are similar to those observed at low pressure (section
III.A) except for some quenching effects, as listed in Table 2.
We obtained TR near 300 K and the nascent values of E0R = 2.4
± 0.3 kJ mol−1. The observed vibrational population
distribution is presented as black open circles in Figure 4;
the extrapolated population of CO (ν = 0) is represented by a
black solid circle. The vibrational temperature of CO is 12210
± 530 K and the average vibrational energy is Evib = 61 ± 7 kJ
mol−1.
III.C. Experiments with HC(O)OCH3/Ar (0.41/0.09
Torr) Irradiated at 193 nm. Traces (a)−(c) in Figure S13
in the Supporting Information present survey emission spectra
DOI: 10.1021/acs.jpca.9b04129
J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A Article

Table 2. Summary of Observed Rotational and Vibrational Temperatures and Average Energies, and Vibrational Populations
Pv of CO Produced upon Photolysis of HC(O)OCH3 (1.00 Torr) and Ar (3.00 Torr) at 248 nm
Rotation Vibration
ν TRa [K] TRb [K] ER [kJ mol−1] Pva [%] Pv × EV [kJ mol−1] TV [K] TVd [K]
12210 ± 530 10000
0 [24.7] 0.0
1 300 ± 10 470 ± 50 (1000 ± 70) 2.5 18.3 4.6
2 330 ± 20 430 ± 50 (920 ± 50) 2.6 16.5 8.2
3 290 ± 10 (800 ± 0) 2.4 11.2 8.3
4 310 ± 10 (730 ± 40) 2.2 8.5 8.3
5 310 ± 20 2.5 7.2 8.8
6 300 ± 10 2.4 5.7 8.4
7 320 ± 30 2.4 4.4 7.5
8 360 ± 20 2.5 3.3 6.5
average E 2.4 ± 0.3 61 ± 7
nascent E0c 2.4 ± 0.3 61 ± 7
a
Observed 0−1 μs after irradiation. The Pv value of ν = 0 was estimated from fitting the Boltzmann distribution. bData taken from ref 20 for the
small-J component; TR for the large-J components are listed in parentheses. cExtrapolated to t = 0; see text. dData taken from ref 20; predicted
value.

Figure 5. Emission spectra of CO and CO2 in the spectral region of 1850−2325 cm−1 recorded 0−1 μs (spectrum (a)) and 3−4 μs (spectrum (b))
upon photolysis of a flowing mixture of HC(O)OCH3 (0.42 Torr) and Ar (0.09 Torr) at 193 nm. Spectral resolution = 0.3 cm−1; 30 laser pulses
were averaged.

recorded at a resolution of 6 cm−1 in the spectral region of

1800−4500 cm−1 at 5-μs intervals after photolysis at 193 nm of
the flowing mixture of HC(O)OCH3 (0.42 Torr) and Ar (0.09
Torr). From a comparison with Figure 1 for photolysis at 248
nm, an obvious difference is that an intense broad emission
band of CO2 near 2250 cm−1 appeared and the emission of
CO is relatively weak. Weak emission in the regions of 2600−
3100 and 3200−3800 cm−1 were also observed, but the small
intensity of these features prevents a definitive assignment.
III.C.a. Emission of CO Fundamental. Figure 5 shows
representative emission spectra at a resolution of 0.3 cm−1 in
the region of 1850−2325 cm−1 recorded 0−1 μs and 3−4 μs
after photolysis of HC(O)OCH3 (0.42 Torr) and Ar (0.09
Torr). In comparison with Figure 2 for photolysis at 248 nm,
in addition to the prominent continuous CO2 emission, the R-
branch of CO appears to extend to larger wavenumbers (J ≤
38 for ν = 1) and was quenched more significantly during the Figure 6. Relative rotational populations of CO (ν = 1−4) recorded
period of 0−4 μs, indicating greater rotational excitation. In 0−1 μs after photolysis at 193 nm of a flowing mixture of
contrast, the vibrational excitation (up to ν = 4) was less than HC(O)OCH3 (0.42 Torr) and Ar (0.09 Torr). Solid lines represent
that at 248 nm. Because the emission of CO2 appears least-squares fits. Each dataset was shifted downward by 10−2(ν−1) (for
continuous whereas lines of CO are sharp, we were able to ν = 2−4) for clarity.
separate emission of CO2 from that of CO using numerical
filtering to facilitate the analysis. were undertaken for spectra averaged over 1−2, 2−3, and 3−4
Figure 6 presents semilogarithmic plots of Pv(J)/(2J + 1), as μs, as shown in Figures S14−S16 in the Supporting
a function of rotational energy ER for the spectrum recorded Information. The derived rotational temperatures and averaged
0−1 μs after photolysis for each vibrational level (ν = 1−4) of rotational energies for CO (ν = 1−4) in these four periods are
CO; the rotational temperature for each vibrational state was listed in Table S3 in the Supporting Information. The variation
derived and is listed in Table 3. The rotational distribution of in rotational temperature for each vibrational level, as a
CO was found to be almost Boltzmann. Similar procedures function of time, is presented in Figure S17 in the Supporting
6135 DOI: 10.1021/acs.jpca.9b04129
J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A
Table 3. Summary of Observed Rotational and Vibrational
Temperatures and Average Energies, and Vibrational
Populations Pv of CO Produced upon Photolysis of
HC(O)OCH3 (0.42 Torr) and Ar (0.09 Torr) at 193 nm
Rotation Vibration
−1
ν a
TR [K] ER [kJ mol ] Pva [%] Pv × EV [kJ mol−1]
0 [57.8]
1 1200 ± 20 8.6 24.6
2 1250 ± 20 8.6 10.3
3 1400 ± 30 9.9 3.5
4 1400 ± 30 9.7 3.4
average E 8.8 ± 0.8
nascent E0b 13 ± 2
a
Observed 0−1 μs after irradiation. The PV value of ν = 0 was
estimated from fitting the Boltzmann distribution. bExtrapolated to t
= 0; see text.
Information. Assuming an exponential decay of the temper-
ature to 298 K, we fitted the data (red lines) and performed a
short extrapolation to t = 0; the estimated nascent rotational
temperatures for ν = 1−4 are listed in Table S3. The average
ratio of the nascent rotational temperature TR0 to the
temperatures determined for the 0−1 μs period is 1.44 ±
0.20, indicating that the correction is significant (∼44%).
Assuming a Boltzmann vibrational distribution, we estimated
the relative vibrational population of CO (ν = 0) to be 1.58 ±
0.23 times that of CO (ν = 1). We normalized the values of Pv
(including ν = 0) to yield the normalized vibrational
population distribution, as listed in Table 3 and depicted in
Figure 7. The vibrational temperature of CO is 3530 ± 380 K;
Figure 7. Vibrational distributions of CO observed 0−1 μs after
photolysis at 193 nm of HC(O)OCH3 (0.42 Torr) and Ar (0.09
Torr). Population of ν = 0 (indicated by the filled symbol) was
derived by extrapolation, according to fitted Boltzmann distribution,
shown as solid lines.
the average vibrational energy is Evib = 18 ± 5 kJ mol−1; the
error limit takes into account of possible contributions of
unobserved higher vibrational levels. Following a method
similar to that in section III.A, we derived an average rotational
energy of 8.8 ± 0.8 kJ mol−1 for 0−1 μs and a nascent
rotational energy E0R = 13 ± 2 kJ mol−1 for the CO fragment.
III.C.b. Emission of CO2. The black trace in Figure 8 shows
emission spectra of CO2 in the region of 1920−2320 cm−1 for
Article
0.0
6.3
5.3
2.6
3.4
18 ± 5
18 ± 5
Figure 8. Observed and simulated CO2 emission spectra. Black line:
spectrum observed 0−1 μs after photolysis at 193 nm of HC(O)-
OCH3 (0.42 Torr) and Ar (0.09 Torr). Blue line represents the
spectrum of the small-energy component (E = 13800 cm−1). Red line
represents the spectrum of the large-energy component (E = 22000
cm−1). Orange dashed line represents the simulated spectrum of CO2
with these two components.
the period of 0−1 μs after removing the sharp emission lines of
CO with numerical filtering; only one broad emission band
with a maximum near 2230 cm−1 and full width at half-
maximum (fwhm) of ∼150 cm−1 was observed. This emission
is assigned to transitions involving a single quantum of CO2 in
the ν3 mode, (ν1, ν2S , ν3) → (ν1, ν2S , ν3 − 1). Although only
emission involving Δν3 = −1 was observed, these transitions
contain the information on the internal excitation of CO2 in all
vibrational modes, including the symmetric stretching (ν1) and
bending (ν2) vibrations. We simulated the observed feature
with an approximate method employed previously for analysis
of the ν3 emission of highly vibrationally excited CO2 produced
from several systems such as O (1D) + CO2,47 O (1D and 3P)
+ OCS,48 and O (1D) + HCOOH.49
We introduced a polyad quantum number νb = 2ν1 + ν2 with
degeneracy gb to describe the vibrational levels of CO2 as (νb,
ν3); ν1 and ν2 are quantum numbers of strongly Fermi-coupled
vibrational modes ν1 and ν2, respectively. In our approx-
imation, the transition is treated as transition (νb, ν3) → (νb, ν3
− 1) of which the wavenumber of the transition is formulated
according to a Dunham expansion, as described previously.
The emission intensity of transition (νb, ν3) → (νb, ν3 − 1), I
[(νb, ν3) → (νb, ν3 − 1)], is proportional to ωN(νb, ν3) A[(νb,
ν3) → (νb, ν3 − 1)], in which ω is the transition wavenumber
for (νb, ν3) → (νb, ν3 − 1), N(νb, ν3) is the population of the
(νb, ν3) state, and A is the Einstein coefficient that is assumed
to be proportional to ν3 but independent of νb. To estimate the
population distribution N(νb, ν3) in the vibrational state (νb,
ν3), we assumed that a maximal available energy E* is
partitioned statistically into the rotation and vibrational
degrees of freedom of CO2 and the relative translational
degree of freedom of CO2 with its counter products, as
described previously.47
The observed feature was decomposed into two components
with the statistically partitioned energy E*1 = 165 kJ mol−1
(equivalent to 13800 cm−1) and E*2 = 263 kJ mol−1 (22000
cm−1), as shown in blue and red traces of Figure 8,
respectively, with a population ratio of ∼74:26. The combined
6136 DOI: 10.1021/acs.jpca.9b04129
J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A Article

Figure 9. Representative potential-energy schemes for the reaction computed at level CCSD(T)/6-311++G(3df,2p)//CCSD/6-311++G(d,p) of
theory. All values are given in kJ mol−1.

spectrum of these two simulated components is presented with electronic surface, including the path via the roaming transition
a red dashed line; it agrees satisfactorily with the experimental structure (RTS), are shown in Figure 9. The full PES related to
observation. The average energies of these two channels are molecular and radical channels are presented in Figures S18
∼17 and 32 kJ mol−1, respectively. Rotational temperatures and S19, respectively, in the Supporting Information. The
500 ± 100 and 800 ± 100 K for the small-E and large-E intrinsic reaction coordinate (IRC) for the RTS is presented in
components, respectively, produces the best fit of the Figure S20 in the Supporting Information. The energies
calculated envelope with the observed band profile. computed with varied methods are compared in Tables S4 and
S5 in the Supporting Information. The geometries of the key
IV. THEORETICAL CALCULATIONS species, including cis-HC(O)OCH3, transition states TS1 and
RTS optimized with varied methods agree with experimental
Methyl formate has two isomers, cis- and trans-HC(O)OCH3.
data,52 as compared in Figure S21 in the Supporting
John and Radom50 and Cui et al.51 have shown that the cis-
Information. Structures TS1−TS11, TS15−20, HOCOCH3,
isomer is more stable than the trans-isomer by ∼14 kJ mol−1;
cyc-H2COC(H)OH, and HOC(H)OCH2 are presented in
one expects that cis-HC(O)OCH3 be dominant at 300 K. We
Figure 10.
considered only the cis-isomer, with the CO and CH3
As shown in Figure 9, as well as Figures S18 and S19, cis-
moieties on the same side, throughout our calculations.
HC(O)OCH3 can decompose via several radical channels
Cui et al. predicted the electronically excited states S1 and T1
through direct cleavage of a bond or through several molecular
to lie ∼430 and 410 kJ mol−1, respectively, above the S0
channels via well-defined transition states; the barriers of the
minimum with the CASSCF-MRCI-SD-ZPE method.51 The
latter are generally smaller than those of the former. The
transition states for the fission of bond C−O are located at
following molecular channels are possible:
∼474 and 441 kJ mol−1, respectively, for the S1 and T1
surfaces; those for the fission of bond C−H are higher in TS1
energy and, hence, unimportant. These authors also located a HC(O)OCH3 ⎯⎯⎯→ CO + CH3OH
minimum-energy conical intersection near 433 kJ mol−1 and ΔH = 38.7 kJ mol−1 (1a)
predicted that, for the C−O bond fission, part of the parent
molecules in the S1 state can decay to the ground state via this
RTS
conical intersection before dissociation. The lack of minimum- HC(O)OCH3 ⎯⎯⎯→ CO + CH3OH
energy S1/T1 crossing point in the Franck−Condon region
causes the dissociation channels on the T1 surface to be ΔH = 38.7 kJ mol−1 (1b)
unimportant. At 248 nm (482 kJ mol−1), because the energy is
only ∼8 kJ mol−1 above that predicted for the transition state TS2
of the C−O fission on the S1 surface, one expects a significant HC(O)OCH3 ⎯⎯⎯→ CO + H 2 + H 2CO
fraction of HC(O)OCH3 to proceed through the conical ΔH = 113.9 kJ mol−1 (2)
intersection and dissociate on the S0 surface. However, at 193
nm (619 kJ mol−1), most HC(O)OCH3 decomposes via S1
and T1 surfaces, as supported by experiments of Lee.16 The TS3
HC(O)OCH3 ⎯⎯⎯→ HOCOCH3 ΔH = 174.0 kJ mol−1
following theoretical calculations was meant to help under-
(3)
standing the decomposition of HC(O)OCH3 on the S0 surface
at 248 nm. TS3 → TS4
IV.A. Potential-Energy Schemes and Decomposition HC(O)OCH3 ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ CO + CH3OH
Channels. The representative potential-energy schemes
(PES) for the decomposition of methyl formate on the ground ΔH = 38.7 kJ mol−1 (4)

6137 DOI: 10.1021/acs.jpca.9b04129

J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A Article

Among all transition structures, TS1 has the smallest energy

(Figure 9); in this reaction, the H atom in the HCO group
migrates toward the O atom of the CH3O group to form
CH3OH + CO. The energies of TS1 (286.4 kJ mol−1) and the
products CH3OH + CO (38.7 kJ mol−1) predicted at the
CCSD(T)/6-311++G(3df,2p)//CCSD/6-311++G(d,p) level
are near values 285.8 and 37.7 kJ mol−1, respectively, obtained
at the CCSD(T)//cc-pVTZ//B3LYP/cc-pVTZ level by Chao
et al.17 The products CH3OH + CO can also be formed via a
roaming-like transition structure (RTS), lying 397.5 kJ mol−1
above the reactant. The geometries of the RTS predicted with
various methods, including CCSD, MP2, M06-2X, and
WB97X, are compared in Figure S21 in the Supporting
Information. However, the contribution of the RTS channel is
negligible, because of its large energy; the CH3OH + CO
products are hence considered to be formed primarily via TS1
(reaction 1a).
In addition, three decomposition or rearrangement molec-
ular channels might also be involved via TS2 (302.4 kJ mol−1),
TS3 (317.2 kJ mol−1), or TS7 (349.3 kJ mol−1) to form CO +
H2 + H2CO (113.9 kJ mol−1), HOCOCH3 (174 kJ mol−1), or
CO2 + CH4 (−117.7 kJ mol−1), respectively. The intermediate
state HOCOCH3 (structure shown in Figure 10) can
decompose to products CO + CH3OH, CO2 + CH4, or CO
+ H2 + H2CO via TS4 (313.5 kJ mol−1) or TS6 (303.3 kJ
mol−1) or TS10 (379.3 kJ mol−1, shown in Figure S18); it can
also break the H3C−O bond without barrier to yield trans-
HOCO + CH3 (309.3 kJ mol−1). A second decomposition
channel of HOCOCH3 to form CO + CH3OH via TS5 (334.7
kJ mol−1) is negligible, because of the large barrier.
Intermediate HOCOCH3 plays a significant role in the
predicted rate coefficients, because of multiple reflections
above the well of the intermediate,53,54 to be discussed later.
Three additional channels to produce H2CO + H2CO, H2CO
+ HCOH, and cyc-H2COC(H)OH (Figure S18) via TS8
(324.0 kJ mol−1), TS9 (435.5 kJ mol−1), and TS11 (387.6 kJ
mol−1), respectively, are negligible, because of the associated
large barriers.
The predicted energies of TS7 (349.3 kJ mol−1) and its
associated products CO2 + CH4 (−117.7 kJ mol−1) are in
satisfactory agreement with corresponding values (348.5−
361.9 kJ mol−1) and −(115.5−128.9) kJ mol−1, respectively,
Figure 10. Structures of transition states and some reaction products
calculated with the CCSD/6-311++G(d,p) method. reported by Francisco19 using varied methods. This channel is
unimportant under our experimental conditions, because of a
TS3 → TS5 large barrier (349.3 kJ mol−1). In contrast, the products CO2 +
HC(O)OCH3 ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ CO + CH3OH CH4 might be produced via TS3 (317.2 kJ mol−1) to form
ΔH = 38.7 kJ mol−1 (5)
excited HOCOCH3, followed by decomposition via TS6
(303.3 kJ mol−1), because of their smaller barriers.
TS3 → TS6 The parent molecule cis-HC(O)OCH3 can also sever its C−
HC(O)OCH3 ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ CO2 + CH4 O or C−H bond directly to form radical products CH3 +
ΔH = −117.7 kJ mol−1 (6) HC(O)O, H + CH3OCO, H + HC(O)OCH2, and HCO +
CH3O, similar to those for direct bond fission of HC(O)-
TS7 OCH2CH3 to form HC(O)O + C2H5, H + C2H5OCO, H +
HC(O)OCH3 ⎯⎯⎯→ CO2 + CH4 HC(O)OCHCH3, and HCO + C2H5O, as reported by Ning et
ΔH = −117.7 kJ mol−1 (7) al.55 These radical channels are summarized below:

TS8 HC(O)OCH3 → CH3 + HC(O)O ΔH = 370.0 kJ mol−1

HC(O)OCH3 ⎯⎯⎯→ H 2CO + H 2CO (10)
ΔH = 129.7 kJ mol−1 (8) HC(O)OCH3 → H + CH3OCO ΔH = 401.8 kJ mol−1
TS9 (11)
HC(O)OCH3 ⎯⎯⎯→ H 2CO + HCOH
HC(O)OCH3 → H + HC(O)OCH 2 ΔH = 405.7 kJ mol−1
ΔH = 348.8 kJ mol−1 (9) (12)

6138 DOI: 10.1021/acs.jpca.9b04129

J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A Article

HC(O)OCH3 → HCO + CH3O ΔH = 408.1 kJ mol−1

(13)
The predicted energies for these radical channels are in
satisfactory agreement with the literature values, as compared
in Table S4 in the Supporting Information. For example, the
energy of HCO + CH3O, 408.1 kJ mol−1 predicted in this
work, is near the value reported by Chao et al. (406.3 kJ
mol−1).17
Moreover, some internally excited radical products can
further decompose, as shown in Figure S19 in the Supporting
Information. For example, the HCO radical can decompose to
H + CO via TS15a or TS15b with a barrier of 72.8 kJ mol−1
and has an endothermicity of 53.8 kJ mol−1 (reaction 14).
Similarly, the HC(O)O radical can also fragment to H + CO2
via TS16 with a small barrier of 0.8 kJ mol−1 and has an
exothermicity of 64.3 kJ mol−1 (reaction 15). In addition, the
CH3OCO radical can break its C−O bond to form CH3 + CO2
via TS17 with a barrier of 142.7 kJ mol−1 and an exothermicity
of 96.1 kJ mol−1 (reaction 16) or form CH3O + CO via TS18 Figure 11. Microcanonical rate coefficients k(E) for the major
channels in the decomposition of HC(O)OCH3, as a function of
with a barrier of 86.2 kJ mol−1 and an endothermicity of 60.1 energy E. The energies of two photolysis wavelengths482 kJ mol−1
kJ mol−1 (reaction 17). Lastly, the HC(O)OCH2 radical can for 248 nm and 619 kJ mol−1 for 193 nmare marked by dotted
decompose to form products (H + CO + CH3O) and (2H + lines.
CO + H2CO) via (TS19, TS18) and (TS20, TS15b),
respectively (see reactions 18 and 19, respectively). 6.4 × 1010 s−1 agree with the values reported by Francisco (2.4
TS15 × 109 and 8.9 × 1010 s−1, respectively). At 482 kJ mol−1 (248
HCO ⎯⎯⎯⎯→ H + CO ΔH = 53.8 kJ mol−1 (14) nm), the next important molecular channel, reaction 2 via TS2,
TS16
has k(E) = 1.5 × 109 s−1, which is ∼53% of that observed for
HC(O)O ⎯⎯⎯⎯→ H + CO2 ΔH = −64.3 kJ mol−1 reaction 1a. Rate coefficients of reactions 6, 4, and 7 are 4.4 ×
(15) 108, 4.1 × 108, and 0.27 × 108 s−1, respectively, which are
<16% of the value observed for reaction 1a.
TS17
CH3OCO ⎯⎯⎯⎯→ CH3 + CO2 ΔH = − 96.1 kJ mol−1 For the radical channels, formation of CH3 + HC(O)O
(16) (reaction 10) is the most important (see Figure 11, as well as
Table S6 in the Supporting Information). At 482 kJ mol−1, the
TS18
CH3OCO ⎯⎯⎯⎯→ CH3O + CO ΔH = 60.1 kJ mol−1 value of k(E) for this channel is 1.1 × 109 s−1, which is ∼40%
(17) of the value observed for reaction 1a. The rate coefficients of
other radical channels are <17% that of reaction 1a. According
TS19/TS18 to the rate predictions on the ground electronic surface, the
HC(O)OCH 2 ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ CH3O + CO
branching ratios of various product channels at 248 nm are
ΔH = 56.2 kJ mol−1 (18) listed in Table 4. Upon excitation of HC(O)OCH3 at 248 nm,
TS20/TS15b
HC(O)OCH 2 ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ H 2CO + CO + H Table 4. Branching Ratios of Various Product Channels of
Photolysis of Methyl Formate at 248 and 193 nm on the
ΔH = 133.7 kJ mol−1 (19) Ground (S0) Surface
IV.B. Calculations of Rate Coefficients. The micro- Branching Ratio
canonical rate coefficients k(E) for the decomposition of cis- reaction product channel 248 nm 193 nm
HC(O)OCH3 on the ground electronic surface were calculated
1a CH3OH + CO (via TS1) 0.390 0.060
for both molecular channels (reactions 1−9) and radical
2 CO + H2 + H2CO (via TS2) 0.205 0.034
channels (reactions 10−13). The predicted rate coefficients for
3/4 CH3OH + CO (via TS4) 0.058 0.011
the channels passing through TS4 and TS6 via TS3 were
3/6 CO2 + CH4 (via TS6) 0.062 0.012
computed by including the multireflection effect.53,54 The
7 CO2 + CH4 (via TS7) 0.004 0.002
computed rate coefficients of the important channels are
10 CH3 + HC(O)O 0.156 0.510
plotted in Figure 11 and are listed in Table S6 in the
11 H + CH3OCO 0.043 0.092
Supporting Information.
12 H + HC(O)CH2 0.019 0.051
For all channels, reaction 1a is the most important in the
13 HCO + CH3O 0.065 0.230
energy range of 290−500 kJ mol−1, because TS1 has the
smallest energy among all transition states. The k(E) value
calculated in this work for the molecular channels via TS1,
TS7, TS8, and TS9 are compared with that calculated by the branching ratios of the main channels are ∼39% for
Francisco15 in Figure S22 in the Supporting Information; the reaction 1a (producing CO + CH3OH), ∼21% for reaction 2
agreement is satisfactory, especially for E ≥ 470 kJ mol−1. For (producing CO + H2 + H2CO), ∼16% for reaction 10
example, at 482 and 619 kJ mol−1 (corresponding to excitation (producing CH3 + HC(O)O), ∼6.6% for reaction 13
at 248 and 193 nm, respectively), our values of 2.8 × 109 and (producing HCO + CH3O), ∼6.2% for reactions 3 and 6
6139 DOI: 10.1021/acs.jpca.9b04129
J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A
(producing CO2 + CH4), and ∼5.8% for reactions 3 and 4
(also producing CO + CH3OH).
V. DISCUSSION
V.A. Photolysis of HC(O)OCH3/Ar at 248 nm. In
photolysis of HC(O)OCH3 at 248 nm, highly vibrationally
excited CO (ν ≤ 11) was observed. Because the rotational
excitation of CO was insignificant, distinguishing the various
channels from rotational distributions is challenging. According
to the listed mechanism on S0, reactions 1a, 1b, 2, 4, and 5 all
produce CO. According to the calculations of the micro-
canonical rate coefficients, at 248 nm (482 kJ mol−1), the three
most important channels for the decomposition of HC(O)-
OCH3 are reactions 1a, 2, and 10, with branching ratios ∼0.39,
0.21, and 0.16 (Table 4). Hence, the CO emission is expected
to be mainly from reactions 1a and 2.
An examination of the structure of TS1 (Figure 10) reveals
that the CO bond length (1.146 Å) is similar to the
equilibrium bond length of CO (1.123 Å), implying that
vibrational excitation of CO would be limited if the reaction
proceeds via TS1. Similarly, reaction via TS2 (Figure 10),
which has a CO bond length of 1.136 Å, is expected to
produce CO with insignificant vibrational excitation. For
reaction 4 (following reaction 3) via TS4, the CO bond
length (1.241 Å) is much greater than the equilibrium bond
length of CO; highly vibrationally excited CO is expected to be
produced via this transition structure. However, according to
the rate calculations on the ground electronic surface, the
branching of this channel is only ∼6%. If the structure of the
transition state governs the vibrational excitation of the
products, our calculations then provide only weak support
for the production of highly vibrationally excited CO.
Lombardi et al.20 employed classical trajectory calculations
and predicted that the vibrational distribution of CO is
Boltzmann with a vibrational temperature of ∼10000 K, shown
as a blue solid line in Figure 4. Our experimental data at 4 Torr
with a vibrational temperature of ∼12210 ± 530 K and at 0.87
Torr with a vibrational temperature of ∼13160 ± 850 K agree
satisfactorily with this prediction. The fact that our observed
vibrational distribution is similar to that predicted with
trajectory calculations that take into account conical
intersection from electronic excited state S1 to the ground-
state surface S0 indicates that this internal conversion likely
plays an important role so that the parts of the PES other than
TS1 might be sampled; decomposition via those regions in
PES might subsequently yield highly vibrationally excited CO.
Finally, we observed no emission of CO2, in agreement with
the small branching ratios of reactions 6 and 7, which were
predicted to be 6.2% and 0.4%, respectively.
V.B. Photolysis of HC(O)OCH3/Ar at 193 nm. Our
observation of CO up to only ν = 4 indicates that the reaction
path differs from that upon photolysis at 248 nm, of which
highly vibrationally excited CO up to ν = 11 was observed. The
observed emission of CO upon photolysis at 193 nm is likely
produced mainly from the secondary decomposition of HCO,
produced via the C−O fission on the S1 surface or via reaction
1a on the S0 surface. As discussed previously, the CO produced
via reaction 1a is expected have have limited vibrational
excitation. The secondary decomposition of HCO proceeds via
TS15a (reaction 14). The C−O bond length of TS15a is 1.127
Å, near a length of 1.123 Å for CO; hence, small vibrational
excitation of CO is expected.
6140
Article
Lee investigated the photodissociation dynamics of DC(O)-
OCH3 in a molecular beam at 193 nm with photo-
fragmentation translational spectroscopy.16 They observed
four channels, corresponding to the production of CH3O +
DCO (X) on the S1 surface, CH3OCO + D on the T1 surface,
CH3O + DCO (A) on other excited surfaces, and CH3OD +
CO on the S0 surface, with branching ratios of 0.73, 0.13, 0.06,
and 0.08, respectively. This author reported that ∼8% of the
DCO further decomposes to D + CO and the ratio of the two
components of CO, from DCO secondary decomposition and
from decomposition on the S0 surface, is 45:55. However, we
identified only one component of CO, likely because the
limited rotational and vibrational excitations of these two
components are indistinguishable.
Lee also reported that all CH3OCO, produced on the T1
surface, decomposed to CH3 + CO2; they probed CH3 instead
of CO2, so it is unclear if two components of CO2 exist in their
system. If we assume that some CO was produced from the S0
surface (reaction 1a), as reported by Lee,16 then it is possible
that the predicted main channel to form CH3 + HC(O)O
(reaction 10) will also be present, according to calculations for
excitation at 193 nm on the S0 surface (Table 4); HC(O)O
can further decompose to H + CO2, yielding the second
component of CO2. Our observed CO2 might hence be
produced from secondary dissociation of CH3OCO (reaction
16 via TS17) and HC(O)O (reaction 15 via TS16). In the
transition structures, the OCO angle is larger for TS16
(158.0°) than for TS17 (147.8°). One would expect more
bending vibrational excitation for CO2 produced from
decomposition of CH3OCO than that produced from the
decomposition of HC(O)O when the transition structures are
considered. Because we do not know the fraction of secondary
decomposition of HC(O)O and CH3OCO, we are unable to
deduce if more HC(O)O (yielding low-internal-energy CO2)
than CH3OCO (yielding high-internal-energy CO2) is
produced. However, the observed major channel CH3 +
HC(O)O (probed with CO2 via secondary decomposition)
was unreported in the molecular-beam experiments, presum-
ably because these authors probed only CH3O, D, and CO to
derive the branching ratios.
V.C. Presence of the Roaming Channel? Figure 12
shows the rotational distributions for ν = 1 and 2 (blue open
triangles) of CO observed upon photolysis of HC(O)OCH3 at
248 nm reported by Lombardi et al.,20 in comparison with our
results at 0.87 Torr (red open circles) and 4.0 Torr (black
open squares). Lombardi et al. observed a bimodal rotational
distribution of CO for ν = 1 and 2, and they proposed that the
small-J component is associated with a roaming mechanism.20
In contrast, we observed an almost-Boltzmann rotational
distribution, even with added Ar to attain a total pressure 4.0
Torr to mimic the experimental conditions of Lombardi et al.20
Clearly identifying a distinct low-J component from our data is
difficult; a slightly enhanced proportion of the low J-states for ν
= 1 and a slightly smaller TR observed at 4.0 Torr might be due
to rotational quenching, because we did not observe such a
component at 0.87 Torr. As shown in Figure 12, for the low-J
levels, the three sets of data are generally similar to each other
within experimental uncertainties. For large-J levels, the
differences in slopes (rotational temperature) are significant,
especially for ν = 2. The details of the analysis method that was
reported by Lombardi et al. were unavailable; hence, the
reason why these authors observed the large-J component is
unclear. However, our results of a rotational temperature near
DOI: 10.1021/acs.jpca.9b04129
J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A
Figure 12. Observed rotational distributions of CO upon photolysis
of HC(O)OCH3 at 248 nm. [Symbols: red circle, total pressure of
0.87 Torr (this work); black square, 4.0 Torr (this work); blue
triangle, Lombardi et al.20. Red and black lines are fitted Boltzmann
distributions of our data. The blue lines are fitted distributions of the
large-J component reported in ref 20.]
400 K is similar to that reported in the ion imaging method for
the molecular channel of CO (∼420 K), but smaller than that
from their FTIR experiments (∼1000 K).20 The superiority of
our data, in terms of spectral resolution, temporal resolution,
and S/N ratio, allows us to have definitive assignments of
observed lines and integrate each individual line to obtain the
relative population of each internal state; this superiority is
demonstrated by the fact that we were able to observe emission
from vibrational levels up to ν = 11, whereas Lombardi et al.20
observed up to only ν = 4. However, we observed no
significant large-J component that could be clearly separated
from the main, small-J component, even though we could not
exclude the possibility that a roaming mechanism might
produce CO with small vibration−rotational energy that is
indistinguishable from that of the observed major channel.
Furthermore, the predicted energy of the roaming transition
state (RTS), 397.5 kJ mol−1, is much higher than that of TS1
(286.4 kJ mol−1); as a result, the roaming mechanism is
expected to be unimportant, according to microcanonical rate
predictions.
A unique signature for a roaming mechanism of HCO +
CH3O to form CO + CH3OH would be the formation of
internally excited CH3OH; the OH-stretching mode is
expected to be highly excited, similar to what was observed
for CH4 in the roaming of CH3 + HCO to form CH4 + CO
from the photolysis of acetaldehyde.56 However, we have not
observed emission in either the OH-stretching or the CH-
stretching region of CH3OH, as indicated in Figure 1. We
conclude that internally excited CH3OH was not produced in
our experiments, unless it had intense emission only below
1800 cm−1, in the region that which we had poor detectivity.
Our observation of neither highly internally excited CH3OH
nor a bimodal rotational distribution of CO provides no
positive support for a roaming mechanism, even though we
could not exclude this possibility.
VI. CONCLUSION
We recorded time-resolved FTIR emission spectra of CO in
the region of 1850−2250 cm−1 upon the photolysis of flowing
6141
Article
mixtures of HC(O)OCH3 (0.73 Torr) and Ar (0.14 Torr) at
248 nm. With an excellent signal-to-noise ratio, we observed
CO emission up to v = 11, with an average vibrational energy
of 76 ± 9 kJ mol−1, in contrast with the previous results of
emission up to ν = 4, as reported by Lombardi et al.20
However, our vibrational distribution agrees with their
trajectory calculations, taking into account the conical
intersection from the S1 to the S0 electronic state. We
observed a near-Boltzmann rotational distribution for all
vibrational states with a small rotational energy, ∼3 kJ mol−1,
in contrast to the bimodal distribution reported by Lombardi
et al. The absence of significant emission in the OH- and CH-
stretching regions of CH3OH and the absence of a definitive
bimodal rotational distribution of CO provide no direct
support for the proposed roaming mechanism, even though we
could not exclude the presence of a roaming mechanism.
Adding 3 Torr of Ar to the system to mimic previously
reported experiments altered the observed results insignif-
icantly, except for enhanced quenching.
Photolysis of a flowing mixture of HC(O)OCH3 and Ar at
193 nm produced intense emission of CO2 and weaker
emission of CO (ν ≤ 4). The rotational distribution of CO is
near-Boltzmann with a rotational energy of 13 ± 2 kJ mol−1
and an average vibrational energy of 18 ± 5 kJ mol−1,
consistent with an expectation from the dissociation of
HC(O)OCH3 to HCO + CH3O, followed by secondary
decomposition of HCO to form CO; this radical channel to
form HCO + CH3O was observed in molecular-beam
experiments to be one of the major channels. Another
possibility is the decomposition to CH3OH + CO on the S0
surface, which was also observed by Lee.16 The internal
distribution of CO2 was deconvoluted to two distributions with
available energies of ∼165 and 263 kJ mol−1 (corresponding to
13800 and 22000 cm−1, respectively), with a population ratio
of ∼74:26. These two channels likely correspond to
dissociation of HC(O)OCH3 to form CH3 + HC(O)O,
followed by secondary decomposition of HC(O)O to H +
CO2, and dissociation to form H + CH3OCO, which further
decomposes to CH3 + CO2, respectively.
■
*
ASSOCIATED CONTENT
S Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jpca.9b04129.
Description of experimental results with HC(O)OCH3/
Ar (1.00/3.00 Torr) irradiated at 248 nm; rotational
temperatures at 0−5 μs determined at 1-μs intervals
upon photolysis at 248 and 193 nm of a flowing mixture
of HC(O)OCH3 and Ar under varied conditions;
relative energies of representative channels and all
species calculated at varied levels of theory; micro-
canonical rate coefficients for decomposition paths of
HC(O)OCH3 as a function of internal energy; torsional
potential for HC(O)OCH3 calculated with the CCSD/
6-311++G(d,p) method; power dependence of observed
emission signal; comparison of observed and simulated
overtone (Δν = 2) emission of CO; assignments of
observed IR emission spectra of CO produced upon
photolysis; relative rotational populations of CO(ν) at 1-
μs intervals 1−5 μs after photolysis at 248 and 193 nm
of flowing mixtures of HC(O)OCH3 and Ar; rotational
temperature TR of CO (v) from photolysis at 248 and
DOI: 10.1021/acs.jpca.9b04129
J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A
193 nm as a function of time; emission spectra in
spectral region 1800−4500 cm−1 recorded upon
photolysis of a flowing mixture of HC(O)OCH3 (1.00
Torr) and Ar (3.00 Torr) at 248 nm at resolution 6 and
0.3 cm−1; potential-energy scheme for molecular
decomposition channels and radical channels of HC-
(O)OCH3 computed at the CCSD(T)/6-311++G-
(3df,2p)//CCSD/6-311++G(d,p) level of theory; in-
trinsic reaction coordinate profile for a roaming
transition state (RTS) calculated at the WB97X/6-
311++G(d,p) level of theory; comparison of calculated
geometries of the reactant, TS1 and RTS at varied levels
of theory with literature values; comparison of predicted
k(E) for several reaction channels with those reported by
Francisco; 6 supplemental tables and 22 supplemental
figures (PDF)
■
AUTHOR INFORMATION
Corresponding Authors
*Tel.: +886-3-5131696. E-mail: chemmcl@emory.edu (M. C.
Lin).
*Tel.: +886-3-5131459. E-mail: yplee@nctu.edu.tw (Y.-P.
Lee).
ORCID
Tien V. Pham: 0000-0002-2067-9028
Maxi Burgos Paci: 0000-0003-2002-7481
M. C. Lin: 0000-0003-3963-6017
Yuan-Pern Lee: 0000-0001-6418-7378
Notes
The authors declare no competing financial interest.
■ ACKNOWLEDGMENTS
This work was supported by Ministry of Science and
Technology, Taiwan (Grant Nos. MOST106-2745-M009-
001-ASP and MOST107-3017-F009-003) and the Center for
Emergent Functional Matter Science of National Chiao Tung
University from The Featured Areas Research Center Program
within the framework of the Higher Education Sprout Project
by the Ministry of Education (MOE) in Taiwan. The National
Center for High-Performance Computation provided com-
puter time. L.L. and M.A.B.P. acknowledge support from
CONICET and FONCYT, Argentina.
■ REFERENCES
(1) Brown, R. D.; Crofts, J. G.; Gardner, F. F.; Godfrey, P. D.;
Robinson, B. J.; Whiteoak, J. B. Discovery of Interstellar Methyl
Formate. Astrophys. J. 1975, 197, L29−L31.
(2) Macdonald, G. H.; Gibb, A. G.; Habing, R. J.; Millar, T. J. A
330−360 GHz Spectral Survey of G 34.3 + 0.15. I. Data and Physical
Analysis. Astron. Astrophys., Suppl. Ser. 1996, 119, 333−367.
(3) Favre, C.; Despois, D.; Brouillet, N.; Baudry, A.; Combes, F.;
Guelin, M.; Wootten, A.; Wlodarczak, G. HCOOCH3 as a Probe of
Temperature and Structure in Orion-KL. Astron. Astrophys. 2011, 532,
A32.
(4) Cazaux, S.; Tielens, A. G. G. M.; Ceccarelli, C.; Castets, A.;
Wakelam, V.; Caux, E.; Parise, B.; Teyssier, D. The Hot Core Around
the Low-Mass Protostar IRAS 16293−2422: Scoundrels Rule!
Astrophys. J. 2003, 593, L51−L55.
(5) Remijan, A. J.; Wyrowski, F.; Friedel, D. N.; Meier, D. S.;
Snyder, L. E. A Survey of Large Molecules toward the Proto-Planetary
Nebula CRL 618. Astrophys. J. 2005, 626, 233−244.
(6) Crovisier, J.; Bockelee-Morvan, D.; Colom, P.; Biver, N.;
Despois, D.; Lis, D. C. The Composition of Ices in Comet C/1995
6142
Article
O1 (Hale-Bopp) from Radio Spectroscopy. Further Results and
Upper Limits on Undetected Species. Astron. Astrophys. 2004, 418,
1141−1157.
(7) Liu, S.-Y.; Mehringer, D. M.; Snyder, L. E. Observations of
Formic Acid in Hot Molecular Cores. Astrophys. J. 2001, 552, 654−
663.
(8) Worley, S. D.; Yates, J. T., Jr A Thermal Desorption Study of
Methyl Formate from W(100) and its Relevance to the Catalytic
Production of Methane. J. Catal. 1977, 48, 395−403.
(9) Barteau, M. A.; Madix, R. J. Decomposition of Methyl Formate
on W(100), W(100)-(5 × 1)C, and W(100)-CO(β) Surfaces. J.
Catal. 1980, 62, 329−340.
(10) Ushikubo, T.; Hattori, H.; Tanabe, K. High Catalytic Activities
of Specially Treated MgO and Na/MgO for Decomposition of
Methyl Formate. Chem. Lett. 1984, 13 (4), 649−652.
(11) Ma, F.-Q.; Lu, D.-S.; Guo, Z.-Y. Base-Catalyzed Decomposition
of Methyl Formate to Carbon Monoxide and Methanol over Zeolite
Catalysts. J. Mol. Catal. 1993, 78, 309−325.
(12) Steacie, E. W. R. The kinetics of the Heterogeneous Thermal
Decomposition of Methyl Formate. Proc. R. Soc. London, Ser. A 1930,
127, 314−330.
(13) Jain, D. V. S.; Murwaha, B. S. Kinetics of Thermal
Decomposition of Methyl Formate in Gas Phase. Indian J. Chem.
1969, 7, 901.
(14) Pereira, R. C. L.; Isolani, P. C. Multiphoton Gas Phase
Dissociation of Methyl and Ethyl Formates. J. Photochem. Photobiol., A
1988, 42, 51−61.
(15) Francisco, J. S. Mechanistic Study of the Gas-phase
Decomposition of Methyl Formate. J. Am. Chem. Soc. 2003, 125,
10475−10480.
(16) Lee, S. H. Photodissociation Dynamics of Methyl Formate at
193.3 nm: Branching Ratios, Kinetic-Energy Distributions, and
Angular Anisotropies of Products. J. Chem. Phys. 2008, 129,
194304−8.
(17) Chao, M.-H.; Tsai, P.-Y.; Lin, K.-C. Molecular Elimination of
Methyl Formate in Photolysis at 234 nm: Roaming vs. Transition
State-Type Mechanism. Phys. Chem. Chem. Phys. 2011, 13, 7154−
7161.
(18) de Wit, G.; Heazlewood, B. R.; Quinn, M. S.; Maccarone, A. T.;
Nauta, K.; Reid, S. A.; Jordan, M. J. T.; Kable, S. H. Product State and
Speed Distributions in Photochemical Triple Fragmentations. Faraday
Discuss. 2012, 157, 227−241.
(19) Tsai, P.-Y.; Chao, M.-H.; Kasai, T.; Lin, K.-C.; Lombardi, A.;
Palazzetti, F.; Aquilanti, V. Roads Leading to Roam. Role of Triple
Fragmentation and of Conical Intersections in Photochemical
Reactions: Experiments and Theory on Methyl Formate. Phys.
Chem. Chem. Phys. 2014, 16, 2854−2865.
(20) Lombardi, A.; Palazzetti, F.; Aquilanti, V.; Li, H.-K.; Tsai, P.-Y.;
Kasai, T.; Lin, K.-C. J. Phys. Chem. A 2016, 120, 5155−5162.
(21) Nakamura, M.; Tsai, P.-Y.; Kasai, T.; Lin, K.-C.; Palazzetti, F.;
Lombardi, A.; Aquilanti, V. Dynamical, Spectroscopic and Computa-
tional Imaging of Bond Breaking in Photodissociation: Roaming and
Role of Conical Intersections. Faraday Discuss. 2015, 177, 77−98.
(22) Yeh, P.-S.; Leu, G.-H.; Lee, Y.-P.; Chen, I.-C. Photodissociation
of HNO3 at 193 nm: Near-Infrared Emission of NO Detected by
Time-resolved Fourier Transform Spectroscopy. J. Chem. Phys. 1995,
103, 4879−4886.
(23) Bagchi, A.; Huang, Y.-H.; Xu, Z. F.; Raghunath, P.; Lee, Y. T.;
Ni, C.-K.; Lin, M. C.; Lee, Y.-P. Photodissociation Dynamics of
Benzaldehyde (C6H5CHO) at 266, 248, and 193 nm. Chem. - Asian J.
2011, 6, 2961−2976.
(24) Huang, Y.-H.; Chen, J.-D.; Hsu, K.-H.; Chu, L.-K.; Lee, Y.-P.
Transient Infrared Absorption Spectra of Reaction Intermediates
Detected with a Step-scan Fourier-Transform Infrared Spectrometer.
J. Chin. Chem. Soc. 2014, 61, 47−58.
(25) Chai, J. D.; Head-Gordon, M. Systematic Optimization of
Long-Range Corrected Hybrid Density Functionals. J. Chem. Phys.
2008, 128, 084106−15.
DOI: 10.1021/acs.jpca.9b04129
J. Phys. Chem. A 2019, 123, 6130−6143
The Journal of Physical Chemistry A Article

(26) Head-Gordon, M.; Head-Gordon, T. Analytic MP2 Frequen- (48) Chiang, H.-C.; Wang, N. S.; Tsuchiya, S.; Chen, H.-T.; Lee, Y.-
cies Without Fifth-order Storage: Theory and Application to P.; Lin, M. C. Reaction Dynamics of O(1D,3P) + OCS Studied with
Bifurcated Hydrogen Bonds in the Water Hexamer. Chem. Phys. Time-Resolved Fourier Transform Infrared Spectroscopy and
Lett. 1994, 220, 122−128. Quantum Chemical Calculations. J. Phys. Chem. A 2009, 113,
(27) Becke, A. D. Density-Functional Thermochemistry. I. The 13260−13272.
Effect of the Exchange-only Gradient Correction. J. Chem. Phys. 1992, (49) Huang, S.-C.; Nghia, N. T.; Putikam, R.; Nguyen, H. M. T.;
96, 2155−2160. Lin, M. C.; Tsuchiya, S.; Lee, Y.-P. Reaction Dynamics of O(1D) +
(28) Becke, A. D. Density-Functional Thermochemistry. II. The HCOOD/DCOOH Investigated with Time-Resolved Fourier-Trans-
Effect of the Perdew−Wang Generalized-Gradient Correlation form Infrared Emission Spectroscopy. J. Chem. Phys. 2014, 141,
Correction. J. Chem. Phys. 1992, 97, 9173−9177. 154313−15.
(29) Becke, A. D. Density-Functional Thermochemistry. III. The (50) John, I. G.; Radom, L. Molecular Conformations of Methyl
Role of Exact Exchange. J. Chem. Phys. 1993, 98, 5648−5652. Formate and Methyl Vinyl Ether from Ab Initio Molecular Orbital
(30) Zhao, Y.; Truhlar, D. G. Density Functionals with Broad Calculations. J. Mol. Struct. 1977, 36, 133−147.
Applicability in Chemistry. Acc. Chem. Res. 2008, 41, 157−167. (51) Cui, G.; Zhang, F.; Fang, W. Insights into the Mechanistic
(31) Scuseria, G. E.; Schaefer, H. F. Is Coupled Cluster Singles and Photodissociation of Methyl Formate. J. Chem. Phys. 2010, 132,
Doubles (CCSD) More Computationally Intensive than Quadratic 034306−5.
Configuration-interaction (QCISD)? J. Chem. Phys. 1989, 90, 3700− (52) Hellwege, K. H.; Hellwege, A. M. (ed.). Landolt-Bornstein:
3703. Group II: Atomic and Molecular Physics Vol. 7: Structural Data of Free
(32) Pople, J. A.; Head-Gordon, M.; Raghavachari, K. Quadratic Polyatomic Molecules. Springer-Verlag. Berlin, Germany. 1976.
(53) Hirschfelder, J. O.; Wigner, E. Some Quantum-Mechanical
Configuration Interaction - A General Technique for Determining
Considerations in the Theory of Reactions Involving an Activation
Electron Correlation Energies. J. Chem. Phys. 1987, 87, 5968−5975.
Energy. J. Chem. Phys. 1939, 7, 616−628.
(33) Andersson, K.; Malmqvist, P.; Roos, B.; Sadlej, A. J.; Wolinski,
(54) Miller, W. H. Unified Statistical Model for “Complex” and
K. Second-order Perturbation Theory with a CASSCF Reference
“Direct” Reaction Mechanisms. J. Chem. Phys. 1976, 65, 2216−2223.
Function. J. Phys. Chem. 1990, 94, 5483−5488. (55) Ning, H.; Wu, J.; Ma, L.; Ren, W.; Davidson, D. F.; Hanson, R.
(34) Andersson, K.; Malmqvist, P. A.; Roos, B. O. Second-Order K. Combined Ab Initio, Kinetic Modeling and Shock Tube Study of
Perturbation Theory with a Complete Active Space Self-consistent the Thermal Decomposition of Ethyl Formate. J. Phys. Chem. A 2017,
Field Reference Function. J. Chem. Phys. 1992, 96, 1218−1226. 121, 6568−6579.
(35) Fernandez-Ramos, A.; Miller, J. A.; Klippenstein, S. J.; Truhlar, (56) Heazlewood, B. R.; Jordan, M. J. T.; Kable, S. H.; Selby, T. M.;
D. G. Modeling the Kinetics of Bimolecular Reactions. Chem. Rev. Osborn, D. L.; Shepler, B. C.; Braams, B. J.; Bowman, J. M. Roaming
2006, 106, 4518−4584. is the Dominant Mechanism for Molecular Products in Acetaldehyde
(36) Klene, M.; Robb, M. A.; Frisch, M. J.; Celani, P. Parallel Photodissociation. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 12719−
Implementation of the CI-vector Evaluation in Full CI/CAS-SCF. J. 12724.
Chem. Phys. 2000, 113, 5653−5665.
(37) Baer, T.; Hase, W. L. Unimolecular Reactions Dynamics: Theory
and Experiment; Oxford University Press: New York, USA, 1996.
(38) Klippenstein, S. J., Wagner, A. F., Dunbar, R. C., Wardlaw, D.
M.; Robertson, S. H. VARIFLEX, Version 1.00; Argonne National
Laboratory, Argonne, IL, USA, 1999.
(39) Gonzalez, C.; Schlegel, H. B. An Improved Algorithm for
Reaction Path Following. J. Chem. Phys. 1989, 90, 2154−2161.
(40) Frisch, M. J.; Trucks, G. W.; Schlgel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B. et al. Gaussian 09, revision B.01; Gaussian, Inc.,
Wallingford, CT, USA, 2010.
(41) Werner, H.; Knowles, P.; Knizia, G.; Manby, F.; Schutz, M.;
Celani, P.; Korona, T.; Lindh, R.; Mitrushenkov, A.; Rauhut, G. et al.
Molpro, version 2012.1, A Package of ab initio Programs. http://www.
molpro.net, 2012.
(42) Barker, J. R. Multiple-Well, Multiple-Path Unimolecular
Reaction Systems. I. MultiWell Computer Program Suite. Int. J.
Chem. Kinet. 2001, 33, 232−245.
(43) Keller-Rudek, H.; Moortgat, G. K.; Sander, R.; Sörensen, R.
The MPI-Mainz UV/VIS spectral atlas of gaseous molecules of
atmospheric interest. Earth Syst. Sci. Data 2013, 5, 365−373.
(44) Hancock, G.; Smith, I. W. M. Quenching of Infrared
Chemiluminescence. 1: The Rates of De-Excitation of CO (4 ≤ ν
≤ 13) by He, CO, NO, N2, O2, OCS, N2O, and CO2. Appl. Opt. 1971,
10, 1827−1842.
(45) Wang, H.; Wang, B.; He, Y.; Kong, F. The Gaseous Reaction of
Vinyl Radical with Oxygen. J. Chem. Phys. 2001, 115, 1742−1746.
(46) Ogilvie, J. F.; Cheah, S.-L.; Lee, Y.-P.; Sauer, S. P. A. Infrared
Spectra of CO in Absorption and Evaluation of Radial Functions for
Potential Energy and Electric Dipolar Moment. Theor. Chem. Acc.
2002, 108, 85−97.
(47) Chen, H.-F.; Chiang, H.-C.; Matsui, H.; Tsuchiya, S.; Lee, Y.-P.
Distribution of Vibrational States of CO2 in the Reaction O(1D) +
CO2 from Time-Resolved Fourier Transform Infrared Emission
Spectra. J. Phys. Chem. A 2009, 113, 3431−3437.

6143 DOI: 10.1021/acs.jpca.9b04129

J. Phys. Chem. A 2019, 123, 6130−6143