Two-color resonant four-wave mixing spectroscopy of highly predissociated levels in the (A)over-tilde(2)A(1) state of CH3S

Two-color resonant four-wave mixing spectroscopy of highly predissociated levels in

the A A 1 2 state of C H 3 S

Ching-Ping Liu, Scott A. Reid, and Yuan-Pern Lee

Citation: The Journal of Chemical Physics 122, 124313 (2005); doi: 10.1063/1.1867333

View online: http://dx.doi.org/10.1063/1.1867333

View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/122/12?ver=pdfcov

Published by the AIP Publishing

Articles you may be interested in

Reaction of C 2 ( a 3 u ) with methanol: Temperature dependence and deuterium isotope effect

J. Chem. Phys. 133, 114306 (2010); 10.1063/1.3480395

Speed dependent rotational angular momentum polarization of the O 2 ( a g 1 ) fragment following ozone photolysis in the wavelength range 248 – 265 nm

J. Chem. Phys. 127, 114308 (2007); 10.1063/1.2775453

Exploring Renner-Teller induced quenching in the reaction H ( S 2 ) + N H ( a 1 ) : A combined experimental and theoretical study

J. Chem. Phys. 126, 034304 (2007); 10.1063/1.2409926

Experimental and theoretical investigations of the reactions NH ( X 3 ) + D ( S 2 ) ND ( X 3 ) + H ( S 2 ) and NH ( X 3 ) + D ( S 2 ) N ( S 4 ) + HD ( X g + 1 )

J. Chem. Phys. 122, 204313 (2005); 10.1063/1.1899563

Experimental and theoretical investigation of the reaction NH ( X 3 ) + H ( S 2 ) N ( S 4 ) + H 2 ( X g + 1 )

J. Chem. Phys. 122, 114301 (2005); 10.1063/1.1862615

Two-color resonant four-wave mixing spectroscopy of highly predissociated

levels in the

A

˜

2

A

₁

state of CH

₃

S

Ching-Ping Liu

Department of Chemistry, National Tsing Hua University, Hsinchu 30013, Taiwan Scott A. Reid

Department of Chemistry, Marquette University, Milwaukee, Wisconsin 53201 Yuan-Pern Leea兲

Department of Applied Chemistry, National Chiao Tung University, Hsinchu 30010, Taiwan and Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 106, Taiwan

共Received 10 November 2004; accepted 13 January 2005; published online 30 March 2005兲

We report results of two-color resonant four-wave mixing experiments on highly predissociated levels of the methylthio共or thiomethoxy兲 radical CH3S in its first excited electronic state A˜

2

A₁. Following photolysis of jet-cooled dimethyl disulfide at 248 nm, the spectra were measured with a hole-burning scheme in which the probe laser excited specific rotational transitions in band 33. The spectral simplification afforded by the two-color method allows accurate determination of line positions and homogeneous linewidths, which are reported for the C–S stretching states 3v共v = 3 – 7兲 and combination states 113v共v=0–2兲, 213v共v=3–6兲, and 11213v共v=0,1兲 involving the symmetric CH3 stretching 共␯1兲 mode and the CH3 umbrella 共␯2兲 mode. The spectra show

pronounced mode specificity, as the homogeneous linewidth of levels with similar energies varies up to two orders of magnitude; ␯3is clearly a promoting mode for dissociation. Derived vibrational

wave numbers␻₁

⬘

,␻₂

⬘

, and␻₃

⬘

of the A˜ state agree satisfactorily with ab initio predictions. © 2005 American Institute of Physics.关DOI: 10.1063/1.1867333兴

I. INTRODUCTION

Oxidation of naturally occurring organic sulfur compounds1,2such as dimethyl sulfide共CH3SCH3兲, dimethyl

disulfide 共CH3SSCH3兲, and methanethiol 共CH3SH兲 in the

atmosphere produces the methylthio 共or thiomethoxy兲 radi-cal, CH3S, as a reactive intermediate.

3,4

As this species is a key intermediate in reactions relevant to modeling the atmo-spheric sulfur cycle,5unambiguous analysis of its spectra and information concerning photochemical processes of its ex-cited states are valuable both for the monitoring of the radi-cal in reactions of environmental importance and for under-standing its chemistry.

From a basic perspective, CH3S is a prototype for

inves-tigation of mode selectivity in predissociation because its first electronically excited state, A˜2A₁, is crossed by quartet and doublet repulsive states correlating with the CH3共X˜2A2

⬙

兲+S共

3_{P兲 asymptote. This situation is analogous}

to that found for the methoxy radical 共CH3O兲 and other

members of this family.6,7 Like methoxy, CH3S has an 2

E ground electronic state and is thus subject to a Jahn–Teller distortion. Theoretical interest in this molecule has been mo-tivated in part by the interactions between Jahn–Teller and spin-orbit effects in the degenerate X˜ 2E ground state.8 Sev-eral ab initio calculations have produced estimates of

ener-gies, geometries, and vibrational wave numbers of the X˜ 2E and/or A˜ 2A₁ states of CH3S.

9–14

There have been many experiments to record spectra of the methylthio radical including electronic absorption15,16 and emission,17 photoelectron and photodetachment,18–20 microwave,21 laser-induced fluorescence 共LIF兲,22–26 photo-fragment yield,27 fluorescence depletion,28 and degenerate four-wave mixing.29These experiments have provided infor-mation on geometries, spin-orbit splitting, and some spectral parameters for CH3S in both X˜ 2E and A˜ 2A1states. Vibronic

analysis of LIF spectra of the A˜ -X˜ system performed in this laboratory identified all six vibrational modes of the X˜ 2E state, but only ␯2 共CH3umbrella兲=1098 cm−1 and ␯3

共C–S stretch兲=401 cm−1 _{were characterized for the A}˜ 2

A₁ state.26 An abrupt decrease of intensity in the fluorescence excitation spectrum was observed above 27 321 cm−1, and attributed to predissociation of the A˜ 2A₁state. No emission was observed above 28 016 cm−1 _{共1490 cm}−1 _{above the A}˜

←X˜ origin兲.

Pushkarsky et al.28 employed fluorescence depletion to extend observed features of A˜2A₁ to 2979 cm−1 above the transition origin; the two features of greatest wave numbers, 37 _{and 2}1₃5_{, have lifetimes ⬃0.5 ps. Bise et al.}27

recorded photofragment yield spectra of CH3S produced on laser

pho-todetachment of a rapid beam of mass-selected CH3S−, and

observed extended vibronic bands of the A←X system up to 31 763 cm−1 _{共5237 cm}−1 _{above the origin兲 which were} as-a兲_{Author to whom correspondence should be addressed; Electronic mail:}

yplee@mail.nctu.edu.tw

0021-9606/2005/122共12兲/124313/8/$22.50 122, 124313-1 © 2005 American Institute of Physics

signed to vibrational progressions of 3v _{and 2}1₃v _共0艋v 艋15兲. We employed degenerate four-wave mixing to

inves-tigate highly predissociative levels of CH3S in the A˜ 2

A₁ state29 and observed bands from 1180 to 5020 cm−1 above the origin; the results showed that the vibrational structure at higher energies was inconsistent with a continuation of the 3v and 213v progressions, and we tentatively identified new progressions 113v, 413v, and 213v41 involving symmetric CH3 stretching 共␯1兲 and asymmetric CH3 stretching 共␯4兲

modes.

As the spectroscopy of CH3S has progressed, so has

re-search on its photochemistry. Measurements of radiative life-time show a significant decrease for vibrational levels of the A

˜ state 艌800 cm−1 _{above the origin.}23,24,26,30

Bise et al.27 used an indirect method to measure the lifetimes of highly predissociative levels of the A˜ 2A₁ state; their reported life-times are significantly longer than those reported by Push-karsky et al.28 who applied fluorescence depletion to mea-sure lifetime broadening for the 3v共v艋7兲 and 213v共v艋5兲 progressions. The latter authors demonstrated that the predis-sociation was mode selective, with ␯3 a promoting mode,

28

and also proposed that, in the region in which excitation involves less than three quanta of ␯3, a second nonradiative

decay channel for the A˜ 2A₁ state, possibly leading to CH2S + H, might exist. In our previous work of DFWM,

29

homogeneous broadening of rovibronic lines leading to over-lapped band structures prevented detailed information on lifetimes of these predissociative states from being obtained. Two-color resonant four-wave mixing共TC-RFWM兲 has been demonstrated to be an excellent tool to investigate highly predissociative states; its double-resonance nature has advantages over DFWM in selecting a specific rovibronic state, hence providing unambiguous spectral assignments and linewidth measurements. When used in a hole-burning scheme, in which the pump and probe beams share a com-mon lower level and the signal is thus derived from the ground-state grating, the signal intensity is unaffected by the lifetime of the upper level.31We have applied TC-RFWM to detect highly predissociative electronic states B2⌺+, C2⌺+, and D2⌸ of CH radical in a flame;32–35 the observed rovi-bronic transitions numbered two to three times those re-ported previously with laser-induced fluorescence or conven-tional absorption methods. We demonstrated also that, although its signal has a quadratic dependence on concentra-tion of species of interest, TC-RFWM can be applied to both stable36,37 and unstable38 species in supersonic jets. In this work we applied TC-RFWM in the hole-burning scheme to investigate highly predissociative levels of CH3S in a

super-sonic jet. In addition to confirming our assignment of pro-gressions involving the symmetric CH3stretching共␯1兲 mode,

lifetimes of these states are measured for the first time, con-firming the mode-selective nature of the predissociation.

II. EXPERIMENTS

We employed TC-RFWM with a ground-state grating共or hole-burning兲 scheme in which the pump 共grating forming兲 and probe transitions shared a common lower level. For the probe transition, we selected specific rotational lines in the 33

band of the A˜ ←X˜ transition. Although this band is predisso-ciative, as evidenced by a minute fluorescence quantum yield,26 it shows resolved rotational structure in DFWM spectra.29 Moreover, the intense 33 _{band allows the use of a}

weaker probe beam; scattered light due to the probe beam is hence diminished. For all recorded TC-RFWM spectra, we set the probe wavelength to achieve resonance with a se-lected rotational transition in the 33 _{band and scanned the}

wavelength of the grating laser.

Details of the TC-RFWM experiment have been reported previously.32,33The pump beams were generated with a dye laser 共Lambda Physik, Scanmate 2E, tunable in a spectral region 338– 361 nm兲 pumped with a XeCl excimer laser at 308 nm共Lambda Physik, LPX 105兲. The frequency-doubled output of a dye laser 共Lambda Physik, Scanmate 2E-OG,

⬃361 nm兲 pumped with the second harmonic 共532 nm兲 of a

Nd:YAG 共YAG—yttrium aluminum garnet兲 laser 共Spectra Physics, GCR-5兲 was employed as a probe. The dye lasers have temporal pulse widths 6 – 7 ns and fundamental spectral linewidths ⬃0.1 cm−1_{. In some critical experiments when}

small jitter between pump and probe beams was required, the 532 nm output of the Nd:YAG laser was split into two beams with an energy ratio⬇2:1 to pump two dye lasers 共Spectra Physics PDL-3, and Scanmate 2E-OG兲; their outputs were frequency-doubled with separate BBO crystals to provide re-quired wavelengths for grating and probe beams. The laser wavelengths were calibrated with a wavemeter 共Burleigh WA-5500; accuracy ±0.2 cm−1兲.

These experiments utilized a forward-box geometry, in which a beam splitter and a few total reflectors were em-ployed to obtain two nearly parallel grating beams共␻1 and

␻2兲 that cross at a small angle 共⬃1°兲 near the nozzle of a

supersonic jet. A temporally coincident probe beam 共␻3兲

propagating in the same direction crossed the grating beams at an angle satisfying the phase-matching condition. The three input beams should overlap spatially and temporally in the medium of interest. The resultant signal beam共␻4兲 was

allowed to travel 2 – 3 m before being spatially filtered with an iris, convex lens 共f=10 cm兲 and a pinhole 共diameter 0.15 mm兲 in combination, and was subsequently filtered with a suitable bandpass interference filter 关passband ⬃365 nm, full width at half maximum 共FWHM兲 10 nm兴 or a mono-chromator 共Jobin–Yvon, HR 320, 0.32 m focal length, 600 grooves mm−1_{兲 before detection with a photomultiplier}

tube 共Hamamatsu, R955P兲. The photomultiplier signal was amplified and averaged with a gated boxcar averager 共Stan-ford Research Systems, SR250兲 and the data were subse-quently transferred to a computer for further processing. The relative timing among the lasers, pulsed nozzle, and data acquisition systems was controlled with a digital delay gen-erator 共Stanford Research Systems, DG535兲. To diminish scattered laser light, we employed a crossed polarization scheme 共YXXY兲, using Fresnel rhombs to rotate the polar-ization and Glan-laser polarizers to select the polarpolar-ization of each beam. The polarization notation used here employs the conventional labeling scheme 共␻4,␻1,␻3,␻2兲.

39

A mixture of dimethyl disulfide共DMDS兲 in helium was generated by bubbling helium through a liquid DMDS sample at ⬃296 K. Additional helium was added

down-124313-2 Liu, Reid, and Lee J. Chem. Phys. 122, 124313共2005兲

stream to dilute further the mixture before expansion through a pulsed nozzle共General Valve, orifice diameter 1 mm兲. The concentration of DMDS was estimated to be less than 1%. The opening duration of the pulse valve was 500␮s and the stagnation pressure was typically 2 atm, which led to a typi-cal background pressure ⬃1⫻10−4 _{Torr. CH}

3S radicals

were produced on photolysis of DMDS at 248 nm using a KrF excimer laser 共Lambda Physik, LPX 150T兲. The pho-tolysis laser beam⬃8 mJ pulse−1_{was loosely focused with a}

cylindrical lens to a point several nozzle diameters down-stream from the orifice. From its laser-induced fluorescence spectrum, the temperature of CH3S was estimated to be

10– 15 K, depending on experimental conditions. DMDS

共Aldrich, 99%兲 and He 共Scientific Gas Products, 99.9995%兲

were used without further purification.

III. RESULTS AND DISCUSSION

We reported previously DFWM spectra of jet-cooled CH3S in its A˜ 2A1 state.

29

The first prominent feature ob-served in the DFWM spectrum was the 33 _{band, of which}

Fig. 1共a兲 displays a high-resolution scan showing the rota-tional structure. TC-RFWM scans were performed on tuning the wavelength of the grating beams with the probe beam tuned to excite rotational lines at 27 707.2 and 27 703.8 cm−1, in the 33band observed in DFWM关marked a and b, respectively, in Fig. 1共a兲兴. These two lines correspond mainly to transitions K_a

⬘

= 0←K_a

⬙

= 1 from J

⬙

= 9 / 2 and 3 / 2

共parity ⫹兲 and J

⬙

= 9 / 2 共parity ⫺兲, respectively, according to spectral simulation with SpecView.40 Figure 1共b兲 illustrates TC-RFWM spectra recorded by scanning the wavelength of grating beams over the 34and 35bands; one dominant tran-sition and some additional weaker features were observed. Observed rovibronic lines for all vibrational bands measured in this work were thus assumed to correspond to rotational transitions identical to the selected probe transition; specifi-cally, when a line corresponding to v₃

⬘

= 3 and K_a

⬘

= 0 is probed, observed TC-RFWM lines correspond to upper lev-els with v3

⬘

艌3 and Ka

⬘

= 0. For extremely predissociated

lines, acceptable ratios of signal to noise were achieved only when line b was probed; hence we report vibrational wave numbers of all vibrational levels of the A˜ 2A₁state based on experiments with line b probed. Because several states were marked by the grating beams, occasionally additional weak lines were observed共e.g., the 35 _{line in Fig. 1 and the 2}1₃3

line in Fig. 2兲. According to spectrum simulation, they are likely due to transitions from J

⬙

= 7 / 2, K_a

⬙

= 2 and J

⬙

= 5 / 2, K_a

⬙

= 2 for lines at low-energy and high-energy sides, respec-tively.

A portion of the TC-RFWM spectra of CH3S is

illus-trated in Fig. 2 for the combination bands 21₃v_{共v=3–5兲 and}

11₃v _{共v=0–2兲. The state-selected TC-RFWM spectra are}

much simpler than those from LIF or DFWM, allowing more accurate determination of line positions and homogeneous line widths. These spectra clearly demonstrate a pronounced increase in line width with increasing quanta of v3,

consis-tent with previous reports.27,28 In the following discussion, we focus first on the vibrational assignments and second on the lifetimes determined from measured linewidths. The

fit-ting procedure used to extract wave numbers and homoge-neous widths of lines is discussed in Sec. III C.

A. Progressions 3v_{and 2}1₃v

Two intense progressions, assigned as 3vand 213v, have been characterized previously by LIF,26 photofragment yield spectra,27fluorescence depletion spectra共FDS兲,28and our re-cent work with DFWM.29Wave numbers reported previously for the progression 3v _{are nearly identical to those}

deter-mined using TC-RFWM for v = 4 – 7 and within 4 cm−1 _for

36 _{and 3}7 _{bands. The slight discrepancy for line positions}

between our TC-RFWM and DFWM experiments reflects the fact that in our DFWM work we could only estimate line positions from the contour of highly predissociative bands showing no rotational structure. For both measurements un-certainty in determining peak positions is greater for higher vibrational levels because of the severe lifetime broadening. Vibrational wave numbers, derived by subtracting 26 523.2 cm−1_{共the corresponding line b in the origin band of}

LIF兲 from an observed wave number of each line, and their

FIG. 1. 共a兲 DFWM spectrum of the 33 _{band of the A˜ ←X˜ transition of}

jet-cooled CH3S. Labels “a” corresponds mainly 兩J⬘, parity, Ka⬘,⌺典 −兩J⬙, parity, Ka⬙,⌺⬘典=兩9/2,−,0,1/2典−兩9/2, + ,1,1/2典, 兩1/2,−,0,1/2典 −兩3/2, + ,1,1/2典, 兩3/2,−,0,−1/2典−兩3/2, + ,1,1/2典, and 兩11/2,−,0, −1 / 2典−兩9/2, + ,1,1/2典 and “b” corresponds to 兩9/2, + ,0,−1/2典−兩9/2, − , 1 , 1 / 2典, 兩7/2, + ,0,1/2典−兩9/2,−,1,1/2典, 兩3/2, + ,1,−1/2典−兩5/2, − , 2 , 1 / 2典, and 兩3/2,−,1,−1/2典−兩5/2, + ,2,1/2典. 共b兲 TC-RFWM spectra re-corded with a YXXY polarization scheme; lines a and b in共a兲 were selected separately as the probe transition while scanning the wavelength of the grating beams.

corresponding uncertainties in measurements are listed in Table I. The observed uncertainty in peak position was de-termined from the fit共see below兲; for narrow lines the error was taken as the uncertainty of wave number measurements, 0.2 cm−1_{. The origin of this transition 共␯}

00兲 lies at

26 526.7 cm−1_,29

corresponding to the a line 共transition K_a

⬘

= 0←K_a

⬙

= 1, parity⫹兲 in the 00_band.

We combined the results of LIF, DFWM, and TC-RFWM to fit the 3v _{progression by least squares to an}

equa-tion,

␯=␯00+共␻3

⬘

共0兲− x33

⬘

兲v3

⬘

− x33

⬘

共v3

⬘

兲2, 共1兲

in which v3

⬘

is the vibrational quantum number of the C–S

stretching共␯3兲 mode and␻3

⬘

共0兲indicates the harmonic

vibra-tional wave number when only␯₃is considered. Fitted spec-tral parameters are listed in Table I. Unless otherwise noted, all listed uncertainties represent one standard deviation in fitting. We found that observed high-energy levels共v=7 and 8兲 deviate from Eq. 共1兲 beyond uncertainties in wave number measurements; these lines were therefore excluded in the fit-ting. Deviations of experimental from calculated values are less than 0.7 cm−1_{forv艋6. Wave numbers predicted for 3}7

and 38_{according to Eq.共1兲 with derived spectral parameters}

are listed in brackets in Table I for comparison.

The 21₃v_{progression in the TC-RFWM spectrum begins}

at v = 3 and extends to v = 6. Our previous DFWM experiments29 showed 21₃v _{bands up to v = 7, with wave}

numbers that deviate within 3 cm−1_{for 2}1₃5 _{and 2}1₃6 _bands

from this work. The discrepancy is more significant for un-resolved DFWM bands associated with the higher vibrational levels, as also found for the 3v_{progression discussed above.}

Vibrational energies of upper states in the 3v共v艋6兲 and 21₃v_{共v艋4兲 progressions are described with an equation} 共with i=2, x22

⬘

= 0, andv2

⬘

= 1 in this case兲:

⌬␯= A +␻₃

⬘

共v₃

⬘

+ 0.5兲 − x₃₃

⬘

共v₃

⬘

+ 0.5兲2+␻_i

⬘

共v_i

⬘

+ 0.5兲 − x_ii

⬘

共v_i

⬘

+ 0.5兲2_{− x}

i3

⬘

共vi

⬘

+ 0.5兲共v3

⬘

+ 0.5兲, 共2兲

in which ⌬␯ is a vibrational wave number relative to the origin, A is a constant to account for zero-point energy, and

␻3

⬘

=␻3

⬘

_共0兲+ 0.5xi3

⬘

. We exclude x22

⬘

from the fitting because

there is no information on x₂₂

⬘

from this progression. Ob-served vibrational wave numbers and fitted spectral param-eters are listed in Table I. Derived spectral paramparam-eters in progressions 3v and 213v are essentially the same as those from our previously reported DFWM work,29but deviations between observed and calculated wave numbers, less than 1 cm−1_{, are diminished through accurate determination of}

band positions. Similar to the progression 3v_{, lines}

corre-sponding to 21₃v_{共v=5–7兲 were excluded from the fitting}

be-cause of large deviations from Eq. 共2兲; wave numbers pre-dicted for these lines are listed in brackets in Table I for comparison. Predicted wave numbers of these lines in the 21₃v _{progression are greater than observed values, whereas}

predicted values of lines 37_{and 3}8_{are smaller than observed}

values. A Fermi resonance between members of 3v+3 _and

21₃v _{共v艌4 or 5兲 might take place.} B. New progressions 11₃v_{and 1}1₂1₃v

In our DFWM experiments,29 from observed spacings the intense lines denoted A

⬘

共29 492.2 cm−1_{兲 and B}

_⬘

共29 826.6 cm−1_{兲 appeared to belong to one progression,}

whereas lines C共30 180.3 cm−1_{兲, D 共30 492.9 cm}−1_{兲, and E}

共30 797.5 cm−1_{兲 appeared to belong to another progression;}

they were tentatively assigned to 11₃v_{and 4}1₃v_{, respectively.}

These bands 共except band E兲 are compared in Fig. 3 with TC-RFWM spectra recorded in this work. Based on this as-signment, three features 38 _{共29 514 cm}−1_{兲, 2}1₃5

共29 503 cm−1_{兲, and 1}1_{共29 492 cm}−1_{兲 might appear and}

over-lap each other in the region near 29 500 cm−1_{. We observed}

only two features in TC-RFWM spectra when line b was probed: one at 29 462.2 cm−1 _{has a laser-limited linewidth}

and the other at 29 500.4 cm−1_{is significantly broader 共Fig.}

2兲. The broad feature is readily assigned to the 21₃5 _band.

The narrow line at 29 462.2 cm−1_{, showing a vibrational}

wave number of 2939.0 cm−1_{, corresponds to a weak}

shoul-der in the DFWM spectrum. Possible assignments for this energy level are␯1,␯4, and 2␯5. We reject the assignment to

␯4and 2␯5because small activity is expected for transitions

involving ␯4, ␯5, and ␯6 due to the small Jahn–Teller

inter-action; we thus assign this line together with lines at 29 828.7 and 30 182.8 cm−1 _{observed in TC-RFWM}

关corre-lated with bands B

⬘

and C in previous DFWM spectrum

FIG. 2. Representative TC-RFWM spectra of the A˜ ←X˜ system of jet-cooled CH3S recorded with a YXXY polarization scheme. The probe beam

was fixed at 27 703.8 cm−1_{共line b in Fig. 1兲 and the scan steps are 0.001 nm}

for narrow bands and 0.01 nm for broad bands. Spectra in the left and right panels show 21₃v_{and 1}1₃v_{progressions, respectively.}

124313-4 Liu, Reid, and Lee J. Chem. Phys. 122, 124313共2005兲

共Ref. 29兲兴 to the 11₃v_{progression withv = 0 – 2. Fitting these}

peak positions by least squares to Eq. 共2兲 with i=1, x₁₁

⬘

= 0, andv1

⬘

= 1 yields␻3

⬘

共0兲= 410± 1 cm −1_{, x} 33

⬘

= 4.1± 0.1 cm−1_,␻ 1

⬘

= 2940± 1 cm−1, and x₁₃

⬘

= 37.5± 0.7 cm−1, as listed in Table I. The derived frequency is consistent with ab initio predic-tions, as discussed in detail below, thus confirming our origi-nal assignment.

Two additional lines at 30 489.4 and 30 799.5 cm−1 above the 1132 line at 30 182.8 cm−1 were observed in the TC-RFWM spectra, which correlate with bands D and E in previous DFWM spectrum.29Although these seemed at first to be higher members of the 113vprogression, we ultimately

rejected this assignment because: first, the intervals do not follow the pattern observed for the low-lying members of this progression, and second, we found that the first of these bands is narrower than the 1132band, which is unexpected if this band is 11₃3_{. Thus, it is clear that these two bands belong}

to a new progression. As the line at 30 489.4 cm−1 corre-sponds to a vibrational wave number of 3966.2 cm−1, pos-sible assignments for this vibrational level are ␯1+␯2, ␯4

+ 3␯₃,␯₂+ 2␯₅, and␯₄+␯₆. Except the assignment to␯₁+␯₂, all others involve at least one nontotally symmetric mode and are expected to be weak. We thus assign these two fea-tures to the 1121 and 112131 bands.

TABLE I. Assignments and vibrational wave numbers 共in cm−1_{兲 of observed TC-RFWM transitions A˜}2 A1 ←X˜2_E

3/2of CH3S. Relative to the rotational line at 26 523.2 cm−1in the origin band; line b in Fig. 1 was

probed. Uncertainties in measurements are listed in parentheses and predicted wave numbers are listed in brackets. v 3v ₂1₃v ₁1₃v ₁1₂1₃v ₂1₃1₄v 0 0a 1095.9a 2939.0 共0.2兲 3966.2共0.2兲 1 401.1a 1489.9a 3305.5 共0.2兲 4276.3b共0.4兲 4358.5c 关4322.1兴 2 794.7a 1874.5c 3659.6 共0.3兲 4667.5c 3 1180.6 共0.2兲 2252.5 共0.2兲 4970.8c 4 1558.0 共0.2兲 2619.3 共0.2兲 5 1925.6 共0.2兲 2973.7b 共0.3兲 关2980.3兴 6 2287.2 共0.4兲 3324.7b 共1.3兲 关3332.3兴 7 2643.9b共0.4兲 关2639.5兴 3669.0b,c 关3676.0兴 8 2987.3b,c 关2983.9兴 ␻3⬘共0兲 409.9± 0.4 410.2± 0.6 410± 1 410± 0 x33⬘ 4.10± 0.06 4.14± 0.07 4.1± 0.1 4.15± 0.09 ␻1⬘共0兲 2940± 1 2940± 2 ␻2⬘共0兲 1096.7± 0.8 1097± 2 x23⬘ 8.5± 0.3 8.6± 0.4 x13⬘ 37.5± 0.7 37.5± 0.7 x12⬘ 70± 1 a

Observed in LIF共Ref. 29兲.

b

These lines were excluded from the fitting; see text.

c

Observed only in DFWM共Ref. 29兲.

FIG. 3. Comparison of DFWM共upper trace, Ref. 29兲 and TC-RFWM spectra of the A˜ -X˜ system of CH3S with scan steps 0.01 nm. TC-RFWM bands 11_vs

21₃5_{and 1}1₃1_{vs 2}1₃6_{demonstrate the}

pronounced mode specificity observed in this system, with ␯3 a promoting

mode for dissociation.

We fit observed lines to an equation involving three vi-brational modes 共with x₁₁

⬘

= x₂₂

⬘

= 0, and v₁

⬘

=v₂

⬘

= 1 in this case兲, ⌬␯= A +

兺

i=1 3 ␻i

⬘

共vi

⬘

+ 0.5兲 −

兺

i=1 3 x_i3

⬘

共v_i

⬘

+ 0.5兲共v₃

⬘

+ 0.5兲 − x12

⬘

共v1

⬘

+ 0.5兲共v2

⬘

+ 0.5兲, 共3兲

in which⌬␯and A have the same meaning as in Eq.共2兲, and

␻3

⬘

=␻3

⬘

共0兲+ 0.5x

⬘

13+ 0.5x23

⬘

. We exclude x11

⬘

and x22

⬘

from the

fitting because information on x₁₁

⬘

and x₂₂

⬘

from these pro-gressions is lacking. The fitted value x₁₂

⬘

= 70± 1 cm−1 indi-cates a strong interaction between ␯1 and␯2. Other

param-eters are similar to those derived from fitting two progressions, as listed in Table I.

Equation共3兲 was sufficient to fit the wave numbers of all observed lines involving no more than two different vibra-tional modes. It fails to predict as accurately wave numbers of the 11₂1₃1 _{line that involves all three totally symmetric}

vibrational modes. This line was excluded from the final fit, but is still listed in Table I, along with the corresponding wave number predicted with Eq.共3兲. The fitting of this line requires the introduction of additional interaction parameters, but such a fitting is unfeasible with only one additional line available for the progression 11₂1₃v_{. A similar situation was}

also observed in the A˜ -X˜ system of CF₃S.41

Lines at 30 881.7, 31 190.7, and 31 494.0 cm−1_,

ob-served in our previous DFWM spectrum, were tentatively assigned as either the 21₃v₄1_{or 1}1₂1₃v_{共v=1–3兲.}29

Now that the latter has been assigned, the only possible assignment for these lines is 21₃v₄1_{. However, wave numbers of these lines}

cannot be fitted satisfactorily with Eq.共3兲, as in the case of 11₂1₃v _{discussed above.}

It is reasonable to ask why the 11₃v_{progression seems to}

terminate abruptly atv = 3, when for the 21₃v_{and 3}v

progres-sions the Franck–Condon maximum is near v = 3. As is ex-plained below, the expected line width of 11₃3 _{based on the}

trend observed for the 11₃v _{progression exceeds 100 cm}−1_.

Thus, a truncation of the 11₃v _{progression at v = 2 is likely}

due to a broadening of this feature beyond our detection limits.

Table II summarizes observed vibrational wave numbers

␻i

⬘

共0兲of states A˜

2

A₁and X˜2E of CH3S and CH3O; 6,42

the raw and scaled results of theoretical calculations13 on the A˜ 2A₁ state are listed for comparison. Experimental vibrational wave numbers of A˜ 2A₁states of CH3S and CH3O agree well

with theoretical calculations when predicted values are scaled by a factor of 0.935; this scaling factor was derived by comparison of predicted values with experiments of CH3O.

Our observation of ␻₁

⬘

_共0兲= 2940 cm−1 _共␻ 1

⬘

= 2994 cm−1_{兲 for}

CH3S is slightly greater than the scaled theoretical value of

2978 cm−1 and the observation of ␻₂

⬘

_共0兲= 1097 cm−1 共␻₂

⬘

= 1136 cm−1_{兲 is slightly smaller than the scaled theoretical}

value 1160 cm−1_{both are within expected error limits.}

C. Predissociation of theA˜ 2A₁

In general, the total line-shape function of the TC-RFWM signal involves integration over the velocity distri-bution and is difficult to express in a simple analytic form, being dependent on relaxation processes and lifetimes of the states involved.43 However, in the limiting case in which homogeneous broadening dominates, the line-shape function can be approximated with a simple Lorentzian squared func-tion, as shown previously.39,44To account for the laser line width, we fit the observed line profile to the square of a Voigt function using a nonlinear least-squares routine. The Gauss-ian component was fixed to reproduce the laser-limited line width 共⬃0.2 cm−1兲 observed in the four lines 共33, 34, 2133, and 11兲 for which no lifetime broadening was observed.

The predissociation lifetime␶is related to the Lorentzian line width by

␶= 1/2␲c⌫ 共4兲

in which c is the speed of light in cm s−1_{and⌫ is the FWHM}

in cm−1_{. Lifetimes of predissociative levels thus estimated}

from observed line widths are listed in Table III; each listed value is the mean of at least five independent measurements and the errors represent one standard deviation. The data are plotted in Fig. 4 for comparison. We did not make a con-certed effort to explore the power dependence of the

TC-TABLE II. Comparison of vibrational wave numbers共in cm−1_{兲 of X˜}2

E3/2and A˜ 2 A1states of CH3S and CH3O. Vibrational mode CH3O CH3S X ˜2 E3/2 A˜ 2 A1 X˜ 2 E3/2 A˜ 2 A1

Expt. Expt. Calc.a Expt. Expt. Calc.b

␯1共a1兲 2840 2947.8 3120共2916兲 c 2940 3186共2978兲 ␯2共a1兲 1412 1289.3 1392共1300兲 1313 1097 1240共1159兲 ␯3共a1兲 1047 662.4 759 共709兲 727 410 439共410兲 ␯4共e兲 2885 3077.8 3281共3066兲 c 3367共3147兲 ␯5共e兲 1465 1403.0 1494共1396兲 1496 1488共1391兲 ␯6共e兲 1210 929.5 1034 共966兲 586 d 746共697兲

Ref. 6 6, 42 13 26 29, this work 13

a

EOM-IP/TZ2P共Ref. 13兲; values scaled by 0.935 are listed in parentheses.

b

EOM-IP/ 6-31G共d,p兲 共Ref. 13兲; values scaled by 0.935 are listed in parentheses.

c

Dispersed fluorescence lists␯1⬙= 2774 cm −1_and␯

4

⬙= 2706 cm−1_{共Ref. 26兲.} d

Proposed to be 635 cm−1_{in LIF共Ref. 26兲.}

124313-6 Liu, Reid, and Lee J. Chem. Phys. 122, 124313共2005兲

RFWM signal; however, based on our qualitative observa-tions of the signal to laser power, it seems clear that we were not in the power saturation regime. The linewidths observed for the strongest transition 共33兲 are consistent with expected laser linewidth; it would be more difficult to saturate highly dissociative transitions. Our results for the 3v and 213v

pro-gressions are consistent with measurements of Pushkarsky and Miller using fluorescence depletion,28which are listed in Table III for comparison. This is important confirmation of the validity of our data reduction procedures. We also in-clude in Table III the results of Bise et al.,27 who used an indirect method to estimate the lifetimes. Their reported val-ues are overestimated in comparison with those obtained in our measurements and those of Pushkarsky and Miller.

The lifetimes of the 113v and 11213v progressions were determined here for the first time. Previous work has shown that levels above 27 321.4 cm−1_共32_{兲 are highly}

predissocia-tive, but the 11line at 29 462.2 cm−1 showed no evidence of lifetime broadening. However, the addition of one quanta of

␯3 leads to a tenfold decrease in lifetime, illustrating the

mode specificity in the dissociation previously revealed for the 3v_{and 2}1₃v_{progressions by Pushkarsky et al.}28

Figure 3 shows other examples of this mode specificity, and the trend observed in each progression in Fig. 4 clearly demonstrates that␯3is a promoting mode for dissociation. Comparing the

levels 21 _{共␶⬃480 ns兲, 1}1 _{共10 ps⬍␶⬍10 ns兲, and 1}1₃1 _共␶

⬃0.46 ps兲, we discern that the rate of predissociation is also

a function of energy. We look forward to theoretical studies on the predissociation rates in this system.

IV. CONCLUSION

We demonstrate the advantages of two-color resonant four-wave mixing in investigating highly predissociated

lev-TABLE III. Vibrational wave numbers, assignments, and lifetimes for the A˜2A1state of the CH3S radical.

Wave numbers

共cm−1_{兲 Assignment}

Lifetimes Chiang and Leea

共ns兲 Pushkarsky et al.b 共ns兲 This work 共ps兲 Bise et al.c 共ps兲 26 526.7d 00 _{1130共70兲 1090 共55兲} 26 927.6d 31 _{860共30兲 870 共40兲} 27 321.4d 32 _{250共20兲 300 共30兲} 27 622.6d 21 _{480共30兲 460 共30兲} 27 707.2e 33 _{72共30兲 艋35} f 28 015.6d 21₃1 _{85共15兲 60 共20兲} 28 084.7e 34 _{0.010⬍␶⬍10} f 25 28 398.9d 21₃2 _{0.010⬍␶⬍10} 28 452.6e 35 _{0.0015 共3兲 1.55共14兲 10} 28 779.1e 21₃3 _{0.010⬍␶⬍10} f 25 28 812.6e 36 _{0.0004 共1兲 0.48共10兲 4} 29 145.9e 21₃4 _{0.0010 共2兲 1.09共14兲 8} 29 170.0e 37 _{0.0005 共1兲 0.53共13兲 2} 29 462.2 11 f 29 500.4 21₃5 _{0.0004 共1兲 0.35 共6兲} 29 828.7 11₃1 _{1.5 共2兲} 29 851.4 21₃6 _{0.09 共5兲} 30 182.8 11₃2 _{0.28 共3兲} 30 489.4 11₂1 _{0.46 共3兲} 30 799.5 11₂1₃1 _0.35共10兲 a Reference 26. b Reference 28. c Reference 27. d

Observed only in LIF or DFWM共Ref. 29兲.

e

Listed wave numbers are from TC-RFWM spectra with line a in Fig. 1 probed, whereas their widths were averages of those measured from TC-RFWM spectra with both lines a and b in Fig. 1 probed. Lines above 29 170 cm−1_{are from spectra with line b probed.}

f

Laser-limited linewidth.

FIG. 4. Dependence of lifetimes in the A˜ ←X˜ system on vibrational mode and energy in TC-RFWM experiments. Progressions 3v_{共䊐, v=3–7兲, 2}1₃v

共쎲, v=3–6兲, 11₃v_{共䉭, v=0–2兲, and 1}1₂1₃v_{共⽧, v=0 and 1兲 are marked}

with different symbols. The arrows indicate bands for which no lifetime broadening was observed, indicating that the lifetime exceeds⬃10 ps.

els of CH3S in its A˜ 2

A₁. We confirmed vibrational assign-ments of new progressions 113v and 11213v involving the symmetric CH3stretching mode共␯1兲 and measured

quantita-tively their lifetimes. The vibrational wave number is ␯1

= 2940 cm−1, consistent with theoretical calculations. The spectra show pronounced mode specificity with ␯3 a clear

promoting mode for dissociation.

ACKNOWLEDGMENTS

The authors thank the National Science Council of Tai-wan 共Grant No. NSC93-2119-M-009-002兲 and MOE Pro-gram for Promoting Academic Excellence of Universities

共Grant No. 89-FA04-AA兲 for support. S.A.R. thanks the

Na-tional Science Council of Taiwan for a visiting professorship at the National Tsing Hua University and acknowledges the Petroleum Research Fund, administered by the American Chemical Society.

1

R. J. Charlson, J. E. Lovelock, M. O. Andreae, and S. G. Warren, Nature

共London兲 326, 655 共1987兲.

2

T. S. Bates, B. K. Lamb, A. Guenther, J. Dignon, and R. E. Stoiber, J. Atmos. Chem. 14, 315共1992兲.

3

A. R. Ravishankara, Y. Rudich, R. Talukdar, and S. B. Barone, Philos. Trans. R. Soc. London, Ser. B 352, 171共1997兲.

4

G. S. Tyndall and A. R. Ravishankara, Int. J. Chem. Kinet. 23, 483

共1991兲.

5

J. M. C. Plane, in Biogenic Sulfur and in the Environment, edited by E. S. Saltzman and W. J. Cooper 共American Chemical Society, Washington, D.C., 1989兲.

6

D. E. Powers, M. B. Pushkarsky, and T. A. Miller, J. Chem. Phys. 106, 6878共1997兲.

7

B. E. Applegate, M. B. Pushkarksky, and T. A. Miller, J. Phys. Chem. A

103, 1538共1999兲. 8

G. D. Bent, J. Chem. Phys. 92, 1547共1990兲.

9

G. D. Bent, J. Chem. Phys. 89, 7298共1988兲.

10

R. Fournier and A. E. Depristo, J. Chem. Phys. 96, 1183共1992兲.

11

C.-W. Hsu, C.-L. Liao, Z.-X. Ma, P. J. H. Tjossem, and C. Y. Ng, J. Chem. Phys. 97, 6283共1992兲.

12

A. C. Curtiss, R. H. Nobes, J. A. Pople, and L. Radom, J. Chem. Phys. 97, 6766共1992兲.

13

Q. Cui and K. Morokuma, Chem. Phys. Lett. 263, 54共1996兲.

14

R. D. El Bouzaidi, A. El Hammadi, A. Boutalib, and M. El Mouhtadi, J. Mol. Struct.: THEOCHEM 497, 197共2000兲.

15

A. B. Callear, J. Connor, and D. R. Dickson, Nature共London兲 221, 1238

共1969兲.

16

A. B. Callear and D. R. Dickson, Trans. Faraday Soc. 66, 1987共1970兲.

17

K. Ohbayashi, H. Akimoto, and I. Tanaka, Chem. Phys. Lett. 52, 47

共1977兲.

18

P. C. Engelking, G. B. Ellison, and W. C. Lineberger, J. Chem. Phys. 69, 1826共1978兲.

19

B. K. Janousek and J. I. Brauman, J. Chem. Phys. 72, 694共1980兲.

20

S. Moran and G. B. Ellison, J. Phys. Chem. 92, 1794共1988兲.

21

Y. Endo, S. Saito, and E. Hirota, J. Chem. Phys. 85, 1770共1986兲.

22

M. Suzuki, G. Inoue, and H. Akimoto, J. Chem. Phys. 81, 5405共1984兲.

23

G. Black and L. E. Jusinski, J. Chem. Soc., Faraday Trans. 2 82, 2143

共1986兲.

24

G. Black and L. E. Jusinski, J. Chem. Phys. 85, 5379共1986兲.

25

Y.-C. Hsu, X. Liu, and T. A. Miller, J. Chem. Phys. 90, 6852共1989兲.

26

S.-Y. Chiang and Y.-P. Lee, J. Chem. Phys. 95, 66共1991兲.

27

R. T. Bise, H. Choi, H. B. Pedersen, D. H. Mordaunt, and D. M. Neumark, J. Chem. Phys. 110, 805共1999兲.

28

M. B. Pushkarsky, B. E. Applegate, and T. A. Miller, J. Chem. Phys. 113, 9649共2000兲.

29

C.-P. Liu, Y. Matsuda, and Y.-P. Lee, J. Chem. Phys. 119, 12335共2003兲.

30

Y.-Y. Lee, S.-Y. Chiang, and Y.-P. Lee, J. Chem. Phys. 93, 4487共1990兲.

31

R. L. Farrow, D. J. Rakestraw, and T. Dreier, J. Opt. Soc. Am. B 9, 1770

共1992兲.

32

W.-C. Hung, M.-L. Huang, Y.-C. Lee, and Y.-P. Lee, J. Chem. Phys. 103, 9941共1995兲.

33

A. Kumar, C.-C. Hsiao, W.-C. Hung, and Y.-P. Lee, J. Chem. Phys. 109, 3824共1998兲.

34

X. Li, A. Kumar, C.-C. Hsiao, and Y.-P. Lee, J. Phys. Chem. A 103, 6162

共1999兲.

35

X. Li and Y.-P. Lee, J. Chem. Phys. 111, 4942共1999兲.

36

A. Kumar, C.-C. Hsiao, W.-C. Hung, and Y.-P. Lee, Chem. Phys. Lett.

269, 22共1997兲. 37

Y. Tang, J. P. Schmidt, and S. A. Reid, J. Chem. Phys. 110, 5734共1999兲.

38

A. Kumar, C.-C Hsiao, Y.-Y. Lee, and Y.-P. Lee, Chem. Phys. Lett. 297, 300共1998兲.

39

S. Williams, J. D. Tobiason, J. R. Dunlop, and E. A. Rohlfing, J. Chem. Phys. 103, 9941共1995兲.

40

SpecView: Simulation and Fitting of Rotational Structure of Electronic and Vibronic Bands. V. Stakhursky and T. A. Miller, 56th Molecular Spectroscopy Symposium. http://molspect.mps.ohio-state.edu/goes/ specview.html

41

D. E. Powers, M. B. Pushkarsky, M.-C. Yang, and T. A. Miller, J. Phys. Chem. A 101, 9846共1997兲.

42

S. C. Foster, P. Misra, T.-Y. D. Lin, C. P. Damo, C. C. Carter, and T. A. Miller, J. Phys. Chem. 92, 5914共1988兲.

43

S. Williams, E. A. Rohlfing, L. A. Rahn, and R. N. Zare, J. Chem. Phys.

106, 3090共1997兲. 44

S. Williams, R. N. Zare, and L. A. Rahn, J. Chem. Phys. 101, 1072

共1994兲.

124313-8 Liu, Reid, and Lee J. Chem. Phys. 122, 124313共2005兲