Doppler-resolved spectroscopy as an assignment tool in the spectrum of singlet methylene

Article abstract of DOI:10.1021/jp0494133

New spectra of methylene, CH₂, in the near-infrared have been obtained following 308 nm photolysis of ketene, CH₂CO. Nascent photofragment Doppler spectra and thermalization kinetics vary systematically with the energy of the absorbing level, providing additional information to support or refute spectroscopic assignments made on the basis of the frequency measurements and combination differences. New assignments in the 10800 cm ^-1 region extend to higher rotational levels than before and provide new spectroscopic term values for some CH₂ a¹A ₁ state levels. The number and intensity distribution of unassigned lines in the spectrum is consistent with the expected transitions from vibrationally excited and high rotational levels of the a¹A₁ state and transitions due to ¹³CH₂ in natural abundance, and does not require a significant contribution from additional transitions arising from triplet-state perturbations.

Full text of DOI:10.1021/jp0494133

7922
J. Phys. Chem. A 2004, 108, 7922-7927
Doppler-Resolved Spectroscopy as an Assignment Tool in the Spectrum of Singlet
Methylene†
Gregory E. Hall,* Anatoly V. Komissarov, and Trevor J. Sears
Chemistry Department, BrookhaVen National Laboratory, Upton, New York 11973-5000
ReceiVed: February 9, 2004; In Final Form: April 2, 2004
New spectra of methylene, CH2, in the near-infrared have been obtained following 308 nm photolysis of
ketene, CH2CO. Nascent photofragment Doppler spectra and thermalization kinetics vary systematically with
the energy of the absorbing level, providing additional information to support or refute spectroscopic
assignments made on the basis of the frequency measurements and combination differences. New assignments
in the 10800 cm-1 region extend to higher rotational levels than before and provide new spectroscopic term
values for some CH2 a˜1A1 state levels. The number and intensity distribution of unassigned lines in the spectrum
is consistent with the expected transitions from vibrationally excited and high rotational levels of the a˜1A1
state and transitions due to 13CH2 in natural abundance, and does not require a significant contribution from
additional transitions arising from triplet-state perturbations.
1. Introduction
of rotational lines in multiple interleaved vibronic bands,
confounds pattern recognition. Searching for lower state com-
bination differences based on the rotational term values of Petek
et al.8,9 and supplemented by more recent work1,3,10 has been
the most secure path to new spectral assignments, although an
unguided search produces an excessive number of coincidentally
close, but mostly incorrect, possible assignments. Furthermore,
many of the higher rotational energy levels in the a˜ state have
not yet been spectroscopically identified. The large asymmetry
of the a˜ state leads to relatively strong ∆Ka ) 3 subbands11 in
addition to those with ∆Ka ) 1. A few of the a˜ state levels are
mixed with the triplet, X˜ 3B1, leading to doubling of some
lines.9,12 Triplet perturbations of the b˜ state might be expected
to occur even more often due to the higher triplet density of
vibrational states; indeed, about 60% of the assigned lines in
the b˜ r a˜ spectrum between 540 and 640 nm showed magnetic
activity,8 which was originally attributed to pervasive mixing
of b˜ and X˜ state levels.8 However, subsequent analysis by
Duxbury and Jungen13 showed the qualitative patterns of
magnetic activity could be explained by unpaired electronic
angular momentum produced by the Renner-Teller interaction
of the a˜ and b˜ states, without invoking mixing with the triplet
X˜ state. Subsequently, a study of the variations in rovibronically
resolved b˜ state radiative lifetimes was used to characterize the
mixing with dark a˜ and X˜ state levels.14 It was concluded that
the dominant source of lengthened radiative lifetimes was
Renner-Teller and Fermi interactions with the a˜ state, with
additional random variations attributed to spin-orbit mixing
with near-degenerate X˜ state levels. Many of the unassigned
lines in the spectrum are undoubtedly caused by absorption from
vibrationally excited (V1V2V3) ) (010) a˜1A1 levels and higher
rotational levels in the a˜1A1(000) state, whose energies are
presently not well-known. At the high dynamic range of the
present experiments, one can also expect a contribution to the
density of weaker lines due to 13CH2 in natural abundance.
For the last 40 years, the visible spectrum of singlet methylene
has been a proving ground for testing advances in our
understanding of the high-resolution spectroscopy of small
molecules. The b˜1B1-a˜1A1 band system in CH2 is dominated
by an irregular progression in the bending vibrational mode,
extending from the near-infrared through the visible region of
the spectrum. The two states involved form a Renner-Teller
pair: both correlate with a degenerate 1∆g state at linearity. The
ground, X˜ 3B1, state is 3150 cm-1 below the a˜ state, and
perturbations caused by interactions among the three low-lying
states result in an irregular absorption spectrum whose complete
analysis has proved to be very difficult. Garcia-Moreno and
Moore1 summarized the spectroscopic work published before
1993, and our group has recently reported experimental work
on the spectrum at longer wavelengths2-4 where it is weaker,
due to poorer Franck-Condon factors, but somewhat simpler,
since the excited-state level density is reduced. Recent theoretical
work by the Jensen5 and Duxbury6,7 groups has considerably
advanced our understanding of the main features of the spectrum
and the rovibronic interactions that account for its complexity.
However, despite the large volume of experimental and
theoretical work devoted to its understanding, the observed
spectrum at Doppler-limited resolution still contains many more
absorption lines than have been assigned. In the visible region,
Petek et al.8,9 reported the measurement of approximately 104
lines in a 3000 cm-1 spectral region and could assign fewer
than 500 of them. In the near-infrared, Kobayashi et al.3,4 could
assign nearly one-quarter of the observed lines, but the majority
remained unassigned.
There are several sources of difficulty. The Ka-dependent
mixing of a˜ and b˜ vibronic states gives dipole intensity to many
zeroth-order vibrational overtone bands of the a˜ state that
otherwise would have negligible intensity. The mixing also
produces large deviations from asymmetric rigid rotor patterns
in the b˜ state. This irregularity, combined with the wide spacing
In the present work, we have recorded sections of the
methylene spectrum in the b˜ r a˜ (020)-(000), (030)-(000),
and (030)-(010) Ka ) 0-1 subbands following 308 nm
photodissociation of ketene under constant, low-pressure condi-
† Part of the special issue “Richard Bersohn Memorial Issue”.
* To whom correspondence should be addressed. E-mail: gehall@bnl.gov.
10.1021/jp0494133 CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/08/2004

Doppler-Resolved Spectroscopy for Singlet Methylene
J. Phys. Chem. A, Vol. 108, No. 39, 2004 7923
tions. The spectral region around 10800 cm-1 has not been
previously recorded and analyzed, although many of the upper
state level positions have been previously determined from work
at other wavelengths.2-4 The observed nascent line shapes and
their thermalization kinetics are shown to vary systematically
with the internal energy of the absorbing level, and hence
provide considerable help in the identification of observed
spectral features. The results are used to extend and correct some
previous spectral assignments of these bands, while adding new
spectroscopically accurate energies for a few a˜ state rotational
levels. A related method of associating rovibronic transitions
with the energy of the lower state has been reported by Green
et al.15 These authors collected laser-induced fluorescence scans
for CH2 at different ketene photodissociation wavelengths,
chosen to vary the energy available for rotational excitation in
CH2 a˜(000). In principle, a photofragment threshold excitation
scan could provide an accurate energy for the initial state of
any detected rotational line, although the process might be
considered too tedious to be practical.
Figure 1. Selected time slices from the near-infrared transient FM
spectrum of CH2. The decay rates and nascent line widths vary widely
among the observed rotational lines. The four strongest lines shown
are (a) 404-312 in the b˜(020)-a˜(000) band, (b) unassigned line, likely
originating in a low rotational state of a˜(010), (c) unassigned line, likely
originating in a high rotational state of a˜(000), and (d) 404-414 in the
hot band b˜(030) -a˜(010).
2. Experimental Details
CH2 radicals were formed by 308 nm excimer laser photolysis
of ketene, synthesized by the pyrolysis of acetic anhydride
flowing past a hot filament (Osram-Sylvania air heater 016501).
A key requirement for these systematic studies of pre-steady-
state kinetics is to maintain a stable and reproducible flow of
ketene and buffer gas at a total pressure low enough that
thermalization occurs in a convenient time window, leading to
distinctive kinetics for different CH2 rotational states. For these
measurements, a He-ketene mixture was prepared continuously
and used immediately after unreacted anhydride and water were
trapped in a dry ice cooled U tube. A 1 sccm flow of helium
used as a carrier gas was bubbled through acetic anhydride at
0 °C via a sintered glass frit. With the filament turned off and
the cold trap passing only He, the pumping speed was adjusted
to maintain a pressure of 100 mTorr in the sample cell. The
heater filament was then turned on and adjusted in temperature
to maintain a constant total pressure of helium plus ketene of
200 mTorr in the absorption cell.
The He-ketene gas mixture was introduced through a series
of pinholes along the length of a 1.5 m long absorption tube
and pumped out through a central exit port using a throttle valve,
liquid nitrogen trap, and mechanical pump. The photolysis and
probe laser beams were aligned nearly coaxially through the
cell, using 45° 308 nm reflecting, near-IR transmitting mirrors
to combine and separate the UV and near-IR beams. The probe
beam was reflected back for a second pass through the sample
cell and then focused onto a fast silicon photodiode receiver.
Details of the frequency-modulated absorption spectrometer used
have been published previously.16,17
Spectra in the range between 10300 and 11500 cm-1 were
measured using a Coherent MBR-110 cw Ti:sapphire probe
laser, pumped by all lines of a Coherent Sabre argon ion laser.
Undesired amplitude modulation of the probe laser at twice the
dither frequency of an intracavity etalon produces a spurious
186 kHz background component in the demodulated transient
signal. The size of the interfering signal was strongly wavelength
dependent, arising from low-finesse etalon effects in the optical
path, although stable in amplitude at a given wavelength. To
minimize the problem, the 10 Hz firing of the photolysis excimer
laser was phase locked to the 93 kHz etalon dither frequency
and a background signal for subtraction was acquired with the
photolysis laser blocked at every wavelength step during a scan.
LabView programs developed in our laboratory controlled the
data acquisition, archiving, and analysis. The data file for each
scan (up to 30 GHz) typically consists of a series of 250-point
transient differential absorption waveforms, spanning a 5 µs time
window, acquired at probe frequencies separated by about 100
MHz, along with wavemeter frequency readout, laser power
normalization channels, a second transient channel measured
in quadrature for phase correction purposes,17 and a background
channel for noise subtraction. Spectra in the 12200 cm-1 region
were recorded using a Coherent 899-29 Ti:sapphire ring laser,
which did not require the background channel.
3. Results and Analysis
3.1. Doppler Widths and Thermalization Kinetics. We
have recorded sections of the spectrum of CH2 between 10538
and 11305 cm-1 under reproducible, low-pressure conditions,
including a continuous region from 10779 to 10855 cm-1. This
region includes strong lines in the b˜(020)0-a˜(000)1 and b˜-
(030)0-a˜(010)1 subbands (the superscript is the Ka quantum
number of the vibronic state). To confirm new assignments,
we also recorded sections of the b˜(030)0-a˜(000)1 subband near
12200 cm-1. This has been recorded previously,18,19 but the
present data are substantially higher quality and were recorded
under the same low-pressure conditions as the longer wavelength
data, to compare line shapes and decays. Figure 1 illustrates
the variation in line widths and decay rates for a cluster of CH2
rotational lines near 10850 cm-1. One can clearly see that line
c at 10850.55 cm-1 decays much faster than the others, and
that the weaker, blended line a at 10850.19 is initially broader
than the rest. Quantitative measurements of such nascent widths
and decay rates for unassigned lines, when calibrated against
the trends in assigned lines, provide additional information to
support or refute tentative assignments of new lines. Lower state
rotational level positions are known from the work of Petek et
al.,8 extended by our previous work,3,4 and spectral assignments
are made using lower state combination differences, comple-
mented by the Doppler and kinetic measurements described
below. New determinations of selected b˜ and a˜ state energy
levels are summarized in Tables 1 and 2. A more extensive list
of observed transition wavenumbers and intensities with partial
assignments is available as Supporting Information.

7924 J. Phys. Chem. A, Vol. 108, No. 39, 2004
TABLE 1: Selected b˜1B1 State Level Positions (cm-1 a
Hall et al.
)
(020)0
energy
(030)0
energy
(010)4
energy
b
JK K
JK K
a
c
a
c
000
101
202
303
404
10826.989
10842.403
10873.233
10919.495
10980.844
10982.179c
11058.001
11150.307
11257.959
12219.637
12235.147
12266.142
12312.561
12374.526
541
542
11460.905c
11460.921c
505
606
707
12452.104
12546.415c
12651.167c
12651.371c
12775.004c
12920.433c
808
909
11380.833c
11519.343c
a Relative to the a˜1A1(000) 000 level. Estimated relative errors are
less than 0.01 cm-1, absolute error (0.02 cm-1 b 404 in (020) is
observed to be split into two levels separated by 1.335 cm-1, while 707
in (030) is observed to be split by 0.204 cm-1 c New or corrected value.
.
Figure 2. Average translational energy of singlet CH2 states based on
nonthermal fits to prompt FM Doppler spectra. The internal energy of
the detected states includes the 1352 cm-1 vibrational energy for
rotational levels in the (010) state.
.
TABLE 2: Newly Determined a˜1A1 State Rotational Level
Positions (cm-1
)
JK K
energy
JK K
a c
energy
a
c
8
18 (010)a
817(000)b
562.741
671.287
919(000)b
918(000)b
699.287
819.841
a Relative to the a˜1A1(010) 000 level, which is 1352.45(2) cm-1 above
the a˜(000) 000 level.8 b Relative to the a˜1A1(000) 000 level. Estimated
absolute errors (0.01 cm-1
.
The nascent state distribution of singlet CH2 following
photodissociation of ketene at a number of wavelengths has been
characterized by the LIF experiments of Garcia-Moreno et al.20,21
The rotational distributions of each vibrational state could be
empirically described by a rotational temperature that varied
smoothly with the excess energy above the dissociation thresh-
old. Interpolating their results to a dissociation wavelength of
308 nm, the singlet CH2 formed in the vibrational ground state
can be characterized by a rotational temperature of about 800
K. About 25% of the singlet CH2 is formed in the (010)
vibrationally excited state with a rotational temperature of about
350 K. The yield of triplet CH2 has been determined to be about
6% at this wavelength.22
Figure 3. Normalized log decay plots of selected singlet CH2 rotational
state populations. The JK Kc quantum numbers for the illustrated states
a
(in order of increasing rotational energy) are (a) 110, (b) 414, (c) 515,
(d) 717, and (e) 919.
Prior to thermalizing collisions, the velocity distribution of
selected CH2 states reflects the internal energy distribution of
the coincident CO fragments. In a supersonic jet environment,
the speed distribution can be inverted to estimate these
coincident state distributions.22,23 In a room temperature “bulb”
experiment, the thermal parent velocity distribution limits the
information that can be determined reliably by direct inversion,
yet the mean translational energy can be extracted accurately
from the nascent Doppler line shapes, providing a distinctive
marker for the CH2 state energy. Figure 2 illustrates the average
translational energy derived from the nascent FM Doppler lines
of assigned CH2 states measured in the 10800 cm-1 region.
Spectra were selected from a time window from 20 to 60 ns
after the photolysis laser and fit with a two-parameter function
to represent the center-of-mass velocity distribution: xa(1 -
x)b, with x ) V/Vmax. The empirical parameters a and b were
used to generate a trial distribution for one of three identical
Cartesian components of the velocity in the center-of-mass
frame. For assigned transitions, Vmax is fixed by energy and
momentum conservation, and for unassigned lines, Vmax is treated
as a third adjustable parameter. The trial velocity distribution
was numerically converted to a center-of-mass Doppler profile,
and convolved with the room temperature velocity distribution
of ketene to make a trial laboratory-frame Doppler line shape.
Finally the FM line shape was computed from the finite
differences at the sideband frequencies to compare to the phase-
corrected experimental spectrum. Parameters a and b were
adjusted for the best least-squares fit, and the average transla-
tional energy was derived from the second moment of the best-
fit laboratory-frame absorption line shape. A simpler, Gaussian
form provides accurate line shape and population fits for times
greater that 1 µs, but systematically overestimates both the
average translational energy and the populations at earlier times,
when the line shapes are nonthermal.
The first few collisions of nascent CH2 with buffer gas and
undissociated ketene smooth the CH2 velocity distribution (and
Doppler line shape) toward a Gaussian form, which then narrows
as it approaches the bath temperature. The rotational distribution
of CH2 also relaxes from its initially hot distribution toward
the bath temperature on a similar time scale. Meanwhile, the
total population of singlet CH2 is depleted by reaction with
ketene and collision-induced intersystem crossing to the ground
triplet state. The disappearance of the vibrationally excited a˜-
(010) state is not appreciably faster than that of a˜(000).24 The
Doppler-integrated intensity has been determined as a function
of time from the same line shape fits described above for
evaluating the initial average translational energy. Figure 3
shows the time-dependent populations for a series of rotational

Doppler-Resolved Spectroscopy for Singlet Methylene
J. Phys. Chem. A, Vol. 108, No. 39, 2004 7925
rotational states appear with faster decays and lower translational
energies (up and to the left). Most of the unassigned lines (open
symbols) can be confidently attributed to either cold- or hot-
band transitions, and the rotational energy of the state respon-
sible for the absorption can be estimated within a few hundred
cm-1. This additional information is often enough to refute a
tentative assignment, or to suggest a preferable nearby candidate
transition, subject to confirmation by combination differences.
Making use of this behavior, we have been able to extend the
spectral assignments in the regions studied. Table 1 lists some
new or corrected rovibronic energies in the b˜1B1 state. A linked
set of combination differences has also allowed us to determine
several new spectroscopic term values for previously unknown
rotational levels of a˜1A1 CH2. These are summarized in Table
2.
3.2. Density of Spectral Lines and the Role of Triplet
Perturbations in the b˜ State. Some considerable help in the
assignment of the spectra also comes from theoretical calcula-
tions. For example, Jensen and co-workers5 treat the Renner-
Teller and anharmonic coupling between a˜ and b˜ state levels,
and compute vibronic band origins that agree with experimental
values within a few cm-1 for low Ka levels and within tens of
cm-1 for higher Ka levels. A search through calculated line lists
can provide a restricted, but still large, set of plausible
assignments for the stronger lines. Even without complete
assignments, however, some statistical properties of the unas-
signed lines can be profitably compared to calculated spectra
to assess the extent of triplet mixing in the upper level.
Figure 4. Decay rates during the 1-2 µs time window following ketene
photolysis for selected CH2 rotational levels in vibrational states (000)
(triangles) and (010) (squares).
The large dynamic range in the present data permits a
comparison of the density and intensity distribution of lines in
the observed spectrum with the calculations of Jensen and co-
workers.5,25 Since these calculations ignore the perturbations
induced by spin-orbit mixing with background triplet levels, a
comparison of the density and intensity distribution of lines with
experimental spectra allows an independent test of the proposi-
tion13 that strong perturbation of b˜ state levels by the dense X˜
state levels is infrequent, and that the pervasive magnetic activity
of the b˜ state8 is predominantly due to unquenched orbital
angular momentum of the mixed a˜ and b˜ singlet states, unrelated
to triplet perturbations.
Figure 5. Correlation of nascent translational energy and thermaliza-
tion/decay rates for assigned and unassigned CH2 spectral lines.
Assigned lines from rotational levels in the (010) state are solid squares
in the left region. Assigned lines from rotational levels of the vibrational
ground state (000) are solid triangles in the right region. Unassigned
lines are shown as open squares.
For a 40 cm-1 section of spectrum starting at 10780 cm-1
,
levels in the vibrational ground a˜(000) state. The logarithmic
plots have been normalized at early times, and illustrate the
systematic variation in decay kinetics with the rotational energy
of the detected state. An arbitrary, but distinctive, attribute
derived from such kinetic plots is the local slope of the
logarithmic decay during the time interval from 1 to 2 µs after
photolysis, when the rotational populations have not yet
completely thermalized. Figure 4 shows the variation of this
decay parameter with the internal energy of CH2 states with
assigned transitions in the present data set. Long-term stability
and day-to-day reproducibility of the pyrolysis and flow
conditions dominate the errors in these measurements, which
occasionally differed by as much as 0.2 µs-1 in replicate runs
on different days.
For lines in the experimental spectrum with relative intensities
down to 0.005 of the most intense absorptions in the region,
we are able to extract both translational energy and decay
parameters. A correlation plot of these two parameters, as shown
in Figure 5, provides a useful tool for identifying the vibrational
and approximate rotational energy of the lower state involved
in an unknown transition. The assigned lines originating in V2
) 0 and V2 ) 1 (solid symbols) fall into two clusters, each
dispersed according to rotational excitation. Within each cluster,
denoted by the tilted ellipses, transitions from higher energy
we sorted the measured absorption lines by nascent intensity
and computed the cumulative density of lines as a function of
threshold intensity. The results are plotted as the symbols in
Figure 6, which illustrate, for example, that in this spectral
region we find 0.3 line/cm-1 with an intensity at least 10% of
that of the strongest line, and 1.6 lines/cm-1 with an intensity
at least 1% of that of the strongest line. The upward curvature
of this plot reflects an increasing number of additional weak
lines detected for each constant factor of increased sensitivity.
The line densities of the weaker lines should be considered lower
limits, due to some blended weak lines that may not have been
recognized. From asymmetries in the time-resolved Doppler-
broadened FM spectra, pairs of lines separated by as little as
0.025 cm-1 (about half the room temperature Doppler width)
are quite easy to recognize if the intensities differ by less than
a factor of 10. We are confident in identifying nearly all the
lines with intensities above 0.15% of that of the strongest line.
Baseline noise levels with 20 shot averaging makes the counting
of weaker lines less reliable.
The experimental distribution can be compared to calcula-
tions5,25 of the frequencies and intensities for all transitions in
this spectral region originating in levels up to J ) 10 of 12CH2
in the b˜-a˜ system. The calculated intensities were rescaled to

7926 J. Phys. Chem. A, Vol. 108, No. 39, 2004
Hall et al.
Positive identification of instances of X˜ state mixing with
the b˜ state is similarly lacking in the b˜ state radiative lifetime
measurements by Garcia-Moreno et al.14 When both members
of a perturbed doublet were identified, neither line ever showed
lifetimes as long as twice the unperturbed value. We take this
as evidence that the perturbing state cannot be the triplet.
Complete mixing with a dark state would double the lifetime
of both components, and incomplete mixing would lead to a
still longer lifetime for the doublet component with the larger
dark state character. There are many instances of isolated
perturbations within the singlet system that lead to a doubling
of zeroth-order b˜ state rotational levels, where the perturbing
level has enough b˜ state character to result in only modest
lengthening of the radiative lifetime. Indeed, the rovibronically
resolved transition moments calculated by Gu et al.5,25 show
J-dependent variations within different Ka states of selected
vibronic levels. These small, apparently random fluctuations,
as well as strong isolated doublings, arise in a detailed
calculation of the Renner-Teller coupled singlet a˜-b˜ system.
Additional couplings with the triplet likely occur, but will be
difficult to identify unambiguously.
Figure 6. Spectral line density (lines/cm-1) as a function of threshold
intensity, relative to the strongest line. Symbols represent the observed
line density in the 10800 cm-1 region. The dashed line is based on
calculations of Jensen and co-workers.5,25 The solid line includes a
correction for triplet perturbations of the lower levels.
4. Conclusion
a rotational temperature of 800 K for the a˜(000) level. We have
included a separate contribution of calculated 13CH2 transitions
in this region, weighted by the 0.01 natural abundance. A set
of lines tentatively assigned to the Q branch of the b˜(020)0-
a˜(000)1 subband in 13CH2 is observed 12-15 cm-1 to the red
of the analogous 12CH2 transitions. The intensities, kinetics, and
line widths are consistent with this assignment, and the isotopic
shift is consistent with unpublished calculations by Jensen.25
We have not, however, systematically searched for other lines
necessary to confirm the 13CH2 assignments. A contribution to
account for hot band transitions in the nascent spectrum was
approximated using the experimental21 vibrational branching
ratio and rotational distribution to scale a frequency-offset copy
of the calculated a˜(000) transitions. Uncertainty in this hot band
contribution probably dominates the error in the calculated
intensity distribution. The resulting calculated intensity distribu-
tion is given by the dashed line in Figure 6, which agrees fairly
well with the observations. The effects of spin-orbit mixing
in the lower state have been estimated using a distribution of
mixing coefficients similar to that computed by Bley and
Temps.26 A fraction of the original, unperturbed lines are
replaced by pairs of lines, which share the original intensity
according to the ratio of the corresponding mixing coefficients.
The frequency and strength of triplet mixing in the a˜(000) state
should increase the density of weaker spectral lines as shown
by the small change from the dashed to the solid line in Figure
6. Most of the extra lines with dominant triplet character are
calculated to contribute to the line density only at still lower
intensities. Triplet perturbations in the b˜ state would generate a
further upward shift in the calculated line density, with a
magnitude reflecting the relative frequency and strength of
perturbations. The observed density of weak lines in this spectral
region is no higher than that of the singlet calculations of
Jensen,5,25 implying that there is no need to invoke frequent
strong triplet perturbations in the b˜ state to account for the
observed density of spectral lines, at least in this spectral region
and with a dynamic range of 1000:1. We must be clear that
these observations cannot be taken as firm evidence against any
triplet perturbations in the b˜ state, only that the perturbations
within the singlet system are already complex enough to account
for the observed density of weak lines.
In a reversal of the usual roles of producer and consumer,
we have applied some tools of photodissociation dynamics and
kinetics to the solution of problems in spectroscopy. Following
ultraviolet photodissociation of ketene, nascent photofragment
Doppler profiles of unassigned rotational lines can be compared
with those of known transitions to provide additional evidence
to support or refute a trial assignment. When time-dependent
spectra are collected with a reproducible flowing sample of
ketene and inert gas, the kinetics of thermalization provide
another characteristic marker of the rotational energy of the
carrier state. Using these tools, we have extended and corrected
some spectral assignments for CH2 in the b˜-a˜ system.
Acknowledgment. This work was performed at Brookhaven
National Laboratory under Contract No. DE_AC02_98CH10886
with the U.S. Department of Energy and supported by its
Division of Chemical Sciences. We thank Prof. Per Jensen
(Wuppertal) for providing us with a detailed list of calculated
transition frequencies and intensities for 12CH2 and 13CH2 based
on ref 5.
Supporting Information Available: Table containing ob-
served CH2 transitions and intensities (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
References and Notes
(1) Garcia-Moreno, I.; Moore, C. B. J. Chem. Phys. 1993, 99, 6429.
(2) Marr, A. J.; Sears, T. J.; Chang, B.-C. J. Chem. Phys. 1998, 109,
3431.
(3) Kobayashi, K.; Pride, L. D.; Sears, T. J. J. Phys. Chem. A 2000,
104, 10119.
(4) Kobayashi, K.; Sears, T. J. Can. J. Phys. 2001, 79, 347.
(5) Gu, J.-P.; Hirsch, G.; Buenker, R. J.; Brumm, M.; Osmann, G.;
Bunker, P. R.; Jensen, P. J. Mol. Struct. 2000, 517-8, 247.
(6) Duxbury, G.; McDonald, B. D.; van Gough, M.; Alijah, A.; Jungen,
C.; Paliva, H. J. Chem. Phys. 1998, 108, 2336.
(7) Duxbury, G.; Alijah, A.; McDonald, B. D.; Jungen, C. J. Chem.
Phys. 1998, 108, 2351.
(8) Petek, H.; Nesbitt, D. J.; Darwin, D. C.; Moore, C. B. J. Chem.
Phys. 1987, 86, 1172.
(9) Petek, H.; Nesbitt, D. J.; Moore, C. B.; Birss, F. W.; Ramsay, D.
A. J. Chem. Phys. 1987, 86, 1189.

Doppler-Resolved Spectroscopy for Singlet Methylene
J. Phys. Chem. A, Vol. 108, No. 39, 2004 7927
(10) Hartland, G. V.; Qin, D.; Dai, H.-L. J. Chem. Phys. 1995, 102,
6641.
(11) Hartland, G. V.; Xie, W.; Qin, D.; Dai, H.-L. J. Chem. Phys. 1992,
97, 7010.
(19) Chang, B. C.; Wu, M.; Hall, G. E.; Sears, T. J. J. Chem. Phys.
1994, 101, 9236.
(20) Garcia-Moreno, I.; Lovejoy, E. R.; Moore, C. B. J. Chem. Phys.
1994, 100, 8890.
(12) McKellar, A. R. W.; Bunker, P. R.; Sears, T. J.; Evenson, K. M.;
Saykally, R. J.; Langhoff, S. R. J. Chem. Phys. 1983, 79, 5251.
(13) Duxbury, G.; Jungen, C. Mol. Phys. 1988, 63, 981.
(14) Garcia-Moreno, I.; Lovejoy, E. R.; Moore, C. B.; Duxbury, G. J.
Chem. Phys. 1993, 98, 873.
(15) Green, W. H., Jr.; Chen, I.-C.; Bitto, H.; Guyer, D. R.; Moore, C.
B. J. Mol. Spectrosc. 1989, 138, 614.
(16) Chang, B.-C.; Costen, M. L.; Marr, A. J.; Ritchie, G.; Hall, G. E.;
Sears, T. J. J. Mol. Spectrosc. 2000, 202, 131.
(17) Hall, G. E.; North, S. W. Annu. ReV. Phys. Chem. 2000, 51, 243.
(18) Herzberg, G.; Johns, J. W. C. Proc. R. Soc. London, Ser. A 1966,
295, 107.
(21) Garcia-Moreno, I.; Lovejoy, E. R.; Moore, C. B. J. Chem. Phys.
1994, 100, 8902.
(22) Morgan, G. C.; Drabbels, M.; Wodtke, A. M. J. Chem. Phys. 1996,
104, 7460.
(23) Costen, M. L.; Katayanagi, H.; Hall, G. E. J. Phys. Chem. A 2000,
104, 10 247.
(24) Langford, A. O.; Petek, H.; Moore, C. B. J. Chem. Phys. 1983, 78,
6650.
(25) Jensen, P. Private communication, 2003.
(26) Bley, U.; Temps, F. J. Chem. Phys. 1993, 98, 1058.
(27) Bunker, P. R.; Jensen, P.; Kraemer, W. P.; Beardsworth, R. J. Chem.
Phys. 1986, 85, 3724.