Rovibronic bands of the A B 2 2 X B 2 1 transition of C 6 H 5 O and C 6 D 5 O
detected with cavity ringdown absorption near 1.2 m
Chi-Wen Cheng, Henryk Witek, and Yuan-Pern Lee
Citation: The Journal of Chemical Physics 129, 154307 (2008); doi: 10.1063/1.2992077
View online: http://dx.doi.org/10.1063/1.2992077
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/129/15?ver=pdfcov Published by the AIP Publishing
Articles you may be interested in
Photodissociation dynamics of acetylene via the C 1 u electronic state J. Chem. Phys. 133, 014307 (2010); 10.1063/1.3456738
Quantum dynamics of the C ( D 1 ) + HD and C ( D 1 ) + n D 2 reactions on the a A 1 and b A 1 surfaces J. Chem. Phys. 132, 104306 (2010); 10.1063/1.3342061
The spectroscopic characterization of the methoxy radical. II. Rotationally resolved A A 2 1 - X E 2 electronic and X E 2 microwave spectra of the perdeuteromethoxy radical CD 3 O
J. Chem. Phys. 130, 074303 (2009); 10.1063/1.3072105
The spectroscopic characterization of the methoxy radical. I. Rotationally resolved A A 2 1 – X E 2 electronic spectra of CH 3 O
J. Chem. Phys. 130, 074302 (2009); 10.1063/1.3072104
Sub-Doppler spectroscopy of the A 2 + X 2 and B 2 X 2 transitions of NCO J. Chem. Phys. 129, 164301 (2008); 10.1063/1.2987305
Rovibronic bands of the A
˜
2B
2] X˜
2B
1transition of C
6H
5O and C
6D
5O
detected with cavity ringdown absorption near 1.2
m
Chi-Wen Cheng,1Henryk Witek,1and Yuan-Pern Lee1,2,a兲 1
Department of Applied Chemistry and Institute of Molecular Science, National Chiao Tung University, 1001, Ta-Hsueh Rd., Hsinchu 30010, Taiwan
2
Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 10617, Taiwan
共Received 26 June 2008; accepted 8 September 2008; published online 17 October 2008兲
We recorded several rovibronic bands of C6H5O and C6D5O in their A˜ 2B2←X˜2B1transitions in the
range 1.14– 1.31 m with the cavity ringdown technique. While the electronic transition is forbidden, several vibronic bands are observed. By comparison of rovibronic contours of observed and simulated bands to determine their types of transition, and by consideration of vibrational wavenumbers of the upper state based on quantum-chemical calculations, we were able to provide vibronic assignments of observed bands and derive several experimental vibrational wavenumbers 共given as in unit of cm−1 in this paper兲 for the A˜2B2 state, namely, 12= 947, 13= 793, 14
= 417, 15= 964, 16= 866, 17= 723, 18= 680, and 19= 499 for C6H5O, and 12= 772, 13= 626, 14= 365,15= 812,17= 599,18= 532, and19= 436 for C6D5O. Transitions involving vibrationally
excited levels of20 were also observed;20of the A˜ state is greater by 50 cm−1than the X˜ state
of C6H5O. A weak origin at 7681 cm−1for the A˜ ←X˜ transition of C6H5O共7661 cm−1for C6D5O兲
with a c-type contour was observed. Observed isotopic ratios of vibrational wavenumbers for the A˜ state of C6H5O to those of C6D5O are in good agreement with the predictions from
quantum-chemical calculations at the B3LYP/aug-cc-pVTZ level. © 2008 American Institute of
Physics.关DOI:10.1063/1.2992077兴
I. INTRODUCTION
The phenoxy共C6H5O兲 radical is an important
interme-diate in the oxidation of small aromatic compounds that are important constituents of fuels.1–4 Kinetics for reactions of C6H5O with O2, O3, NO, NO2, and CH3have been
investi-gated in detail.5–9Most investigations on kinetics of gaseous C6H5O utilized broad electronic absorption bands of C6H5O
in the UV or visible region. Popular methods for production of C6H5O include laser or flash photolysis of phenol 共C6H5OH兲 or anisol 共C6H5OCH3兲.
C6H5O has a ground electronic state designated as
X ˜ 2B
1. The C⫺O bond was indicated to have a double-bond
character, based on the results of electron spin resonance 共ESR兲 spectroscopy.10–15
Several vibrational modes of C6H5O were identified with resonance-enhanced Raman spectroscopy,16–20 but most vibration fundamentals of C6H5O remained unknown. The infrared analysis of gaseous
C6H5O is unreported, only matrix-isolated C6H5O, produced
upon photolysis of nitrobenzene or nitrosobenzene at 308 nm, has been reported; 26 of the 30 fundamental vibrational modes were identified.21
Electronically excited states of B˜ 2A2, C˜ 2B1, D˜ 2A1, and
E ˜ 2B
1 have been characterized with absorption bands in
re-gions of 510⫺640 nm,8,21–26 370⫺400 nm,9,21,22,25–27 270 ⫺300 nm,5,6,21,23,25,27–29 and 220⫺250 nm,5,6,25,29
respec-tively. Only the B˜ 2A2←X˜2B1 transition shows a resolved
vibrational progression of spacing⬃480 cm−1.24,25
The A˜ 2B2←X˜2B1transition of C6H5O is symmetry
for-bidden but vibronically allowed. The weakness of this tran-sition has precluded detailed investigation of the A˜ 2B2state.
Gunion et al.30 observed a band near 1170 nm共8550 cm−1兲
in the photoelectron spectrum upon photodetachment of gas-eous C6H5O− and assigned it as the A˜ ←X˜ transition of
C6H5O. Radziszewski et al.25 employed ultraviolet-visible
polarization absorption spectroscopy and determined direc-tions of transition dipole for several low-lying electronic states of C6H5O isolated in solid Ar. They reported an
ex-tremely weak band near 1123 nm共8900 cm−1兲 with
undeter-mined direction of the transition dipole and assigned it as the
A
˜ ←X˜ transition. If the observed bands in these experiments
are associated with the same transition, the blueshift of 350 cm−1 for matrix-isolated C6H5O seems to be
uncommon.
Most theoretical calculations on the ground electronic state of C6H5O indicate that C6H5O is planar with a X˜ 2B1
ground state.21,31–34 The predicted bond length of C⫺O is ⬃1.24 Å, closer to the length of a CO double bond of 1.215 Å for acetone than that of a CO single bond of 1.364 Å for phenol,35 consistent with the results from ESR. To our knowledge, quantum-chemical calculations on only vertical excitations but not on the adiabatic excitation energy of electronic excited states of C6H5O are reported. The re-ported vertical excitation energy of the A˜ 2B2 state varied
a兲Author to whom correspondence should be addressed. Electronic mail:
yplee@mail.nctu.edu.tw.
共2008兲
0021-9606/2008/129共15兲/154307/9/$23.00 129, 154307-1 © 2008 American Institute of Physics
from 5322 to 12685 cm−1.9,25,33,34,24,36–40
The predicted CO bond length of C6H5O increases from⬃1.24 Å in its ground
state to⬃1.35 Å in its A˜2B2 state, indicating a single-bond character for the A˜ state.
We have recently successfully employed the cavity ring-down technique near the 1.2– 1.4 m region to investigate weak rovibronic bands of gaseous CH3OO and CD3OO in their A˜ ←X˜ transitions and reported several vibrational fre-quencies of the A˜ state.41 The high sensitivity of the cavity ringdown technique is suitable to investigate electronically forbidden transitions of free radicals. We report here the first observation of the A˜ 2B2←X˜2B1vibronic system and
iden-tification of several vibrational wavenumbers of the A˜ 2B2
state of C6H5O using the cavity ringdown technique.
II. EXPERIMENTS
The principles and techniques of cavity ringdown have been discussed extensively.42 For absorption spectra based on cavity ringdown one measures the decay of intensity of a laser pulse trapped in an optical cavity formed between two highly reflective mirrors.41,43 When the cavity contains an absorbing medium, which for these experiments is C6H5O or
C6D5O, the ringdown period is decreased at those
wave-lengths at which absorption occurs. The ringdown periodis defined as the temporal interval required for the intensity to decay to 1/e of its initial value. In the presence of an absorb-ing species the rabsorb-ingdown period
⬘
becomes
⬘
= L/关c共1 − R +␣l兲兴, 共1兲in which c is the speed of light, R is the reflectivity of the mirror, l is the path length共in cm兲 of the absorbing medium,
L is the length of the cavity, and ␣=N is the absorption
coefficient, in which is the cross section of absorption in cm2, and N is the number density of absorbing molecules in
cm−3. The absorption per unit length is hence simply related
to the difference in cavity ringdown periods of
⬘
determined when the absorption species was present and when the absorption species was absent,␣=
冉
1 ⬘
− 1 冊
L cl. 共2兲The laser radiation employed for cavity ringdown was generated by shifting the output of a dye laser共Spectra Phys-ics, Model Sirah, with a mixture of exciton dyes LDS 867, LDS 821, and LDS 765兲 with a single-pass Raman shifter employing gaseous H2at⬃15 bars to produce emission near
1.2 m with energy ⬃1 mJ pulse−1. The dye laser was
pumped with a Nd:YAG 共yttrium aluminum garnet兲 laser 共Spectra Physics, PRO-270, 10 Hz兲 to generate output of 20⫺40 mJ in the range 765⫺860 nm. The spectral width of this laser system is ⬃0.06 cm−1 and the wavelength
scan-ning steps were 0.05 nm共⬃0.7 cm−1兲. The ringdown signal
was recorded with an InSb detector共Kolmar, KISDP-1-LJ2, rise time of 520 ns兲 cooled to 77 K. The signal was amplified 共1 MHz bandwidth兲 and recorded with a digital computer oscilloscope共14 bits, Gage Applied Technologies, Compus-cope 14100兲, and subsequently processed with a personal
computer. Typically the waveform was averaged over 30 la-ser pulses at each wavelength. We calibrated the wavelength of the dye laser with a wavemeter 共Burleigh, 4500兲 and the lines of the a˜ 1⌬g←X˜3⌺g
− system of molecular oxygen
ac-cording to theHITRANdatabase.44The accuracy of measure-ments of wavelength is estimated to be⫾0.012 nm.
A flowing system was employed to achieve steady-state conditions of C6H5O. The flow reactor has two rectangular
共2.5⫻15 cm2兲 quartz 共S1UV兲 plates on the side for
photoly-sis. The length of the cavity, formed between two highly reflective mirrors 共Los Gatos Research, R⬵0.999 95 near 1.2 m and R⬵0.9999 near 1.3 m兲, is 63 cm. The cavity mirrors were purged with N2 to avoid possible damage by
the reactants. C6H5O was produced by photolysis of anisol
共C6H5OCH3, Aldrich, 99%兲 or phenetole 共C6H5OC2H5,
Ald-rich, 99%兲 at 193 nm 共Lambda Physik, LPX110i, 10 Hz, 45 mJ兲. The total pressure of the system was typically 220 Torr with C6H5OCH3: N2⬵1:72. C6D5O was produced from
fully deuterated anisol 共Aldrich, 98%兲. III. THEORETICAL CALCULATIONS
Detailed calculations on the geometry, energy, and vibra-tional frequencies of C6H5O and C6D5O in their electronic
ground and first excited states have been reported previously.31The geometry and energy of C6H5O in its A˜ and
X
˜ states were calculated with unrestricted hybrid density
functional method 共B3LYP兲, complete active space self-consistent-field 共CASSCF兲, and complete active space second-order perturbation theory 共CASPT2兲 methods; nine active electrons distributed among eight orbitals, designated as共9,8兲, were used for the latter two methods. Various basis sets were tested; predicted vibrational wavenumbers and electronic excitation energy show little dependence on the basis sets. We employ results using the aug-cc-pVTZ basis set for comparison with experiments.
The ground electronic state of C6H5O has a C2v
symme-try; we define the molecular plane as the yz plane and the direction of the CO bond as the z-axis. The rotational a-, b-, and c-axis coincide with the z-, y-, and x-axis, respectively. Rotational constants of the A˜ and X˜ states based on the pre-dicted geometries are listed in TableI; these values vary less than 1.6% between various basis sets and methods. The re-sults from B3LYP/aug-cc-pVTZ are used in the spectral simulation.
Related molecular orbitals of the A˜ and X˜ states of C6H5O are shown in Fig.1; SOMO indicates the singly
oc-cupied molecular orbital, HOMO indicates the highest dou-bly occupied molecular orbital, and HOMO⫺n indicates the
nth lower doubly occupied molecular orbital. The transition A
˜ 2
B2←X˜2B1 involves excitation of an electron from the
third highest doubly occupied molecular orbital共HOMO⫺2兲 to the SOMO. In the ground electronic state, orbitals of the O atom are extensively involved with bonding of the benzene ring. The HOMO⫺2 involves approximately a nonbonding orbital of O atom, whereas the SOMO has aⴱcharacter for the O atom and the benzene ring. Upon the A˜ ←X˜ transition, the latter orbital becomes doubly occupied and changes its
154307-2 Cheng, Witek, and Lee J. Chem. Phys. 129, 154307共2008兲
nature and energy substantially, as shown in Fig.1. The ap-proximate n→ⴱexcitation implies that the CO bond length of the X˜ state 共1.26 Å兲 becomes greater in the A˜ state 共1.32 Å兲. The bond lengths of C1-C2 and C2-C3 in the X˜ state vary from 1.46 to 1.38 Å, whereas those in the A˜ state are similar共1.41 and 1.39 Å兲; the equilibrium ring structure of the A˜ state of C6H5O is similar to that of benzene in its ground electronic state.
Previous calculations provided only vertical excitation energy with values ranging from 5322 to 12 685 cm−1. The
vertical and adiabatic excitation energies of the A˜ ←X˜ tran-sition calculated with B3LYP/aug-cc-pVTZ, CASSCF共9,8兲/ aug-cc-pVTZ, and CASPT2共9,8兲/aug-cc-pVTZ methods are compared with previous results in Table II. The adiabatic excitation energy共zero-point energy corrected兲, with values 7242, 6694, and 8339 cm−1 from B3LYP/aug-cc-pVTZ,
CASSCF共9,8兲/aug-cc-pVTZ, and CASPT2 共9,8兲/aug-cc-pVTZ, respectively, are all smaller than the energy of the observed band at 8550 or 8900 cm−1.25,30
Scaled harmonic vibrational wavenumbers of C6H5O and C6D5O in their A˜ state predicted with the B3LYP/aug-cc-pVTZ method are compared with experimental results in Table III; the ordering of vibrational modes follows that of Herzberg.45 The vibrational wavenumbers scaled by single scaling factors of 0.967 and 0.970 for C6H5O and C6D5O, respectively, are listed in TableIII; the scaling factors were derived by comparison of calculated and experimental values of phenol.31The predicted vibrational wavenumbers of the X˜ state 共not listed兲 are similar to those reported previously for C6H5O and C6D5O isolated in solid Ar;21the deviations are less than 8% and 6%, respectively.
IV. EXPERIMENTAL RESULTS AND DISCUSSION A. Vibronic bands of C6H5O
Upper traces in panels 共A兲–共D兲 of Fig. 2 present ob-served spectra in the region 7660– 8760 cm−1 共1.14–1.31 m兲 upon photolysis of a flowing mixture of
TABLE I. Rotational constants A, B, and C共in cm−1兲 of the X˜ and A˜ states of C
6H5O and C6D5O calculated
with various methods.
Molecule State Rotational constant
B3LYP /aug-cc-pVTZ B3LYP /aug-cc-pVDZ CASSCF /aug-cc-pVDZ C6H5O X˜ A⬙ 0.184 9 0.182 8 0.182 0 B⬙ 0.093 39 0.092 43 0.093 28 C⬙ 0.062 05 0.061 39 0.061 67 A ˜ A⬘ 0.190 4 0.188 1 0.188 6 B⬘ 0.092 46 0.091 55 0.091 55 C⬘ 0.062 24 0.061 57 0.061 63 C6D5O X˜ A⬙ 0.153 4 0.151 7 0.151 2 B⬙ 0.086 10 0.085 18 0.085 97 C⬙ 0.055 00 0.054 55 0.054 80 A ˜ A⬘ 0.157 5 0.155 5 0.156 1 B⬘ 0.085 30 0.084 47 0.084 50 C⬘ 0.055 35 0.054 74 0.054 83 (1) ground state X~ 2B 1
HOMO–3 HOMO–2 HOMO–1 HOMO SOMO
(-0.5340) (-0.4990) (-0.4497) (-0.3511) (-0.1538)
(2) first excited state A~ 2B2
HOMO–3 HOMO–2 HOMO–1 HOMO SOMO
(-0.5451) (-0.4974) (-0.3382) (-0.3175) (-0.2477)
FIG. 1. 共Color online兲 The first few highly occupied molecular orbitals
共HOMOs兲 and singly occupied molecular orbitals 共SOMOs兲 of C6H5O in
their A˜ and X˜ states predicted with CASSCF共9,8兲/cc-pVDZ. The energies in
hartree are listed in parentheses. See details in the Supporting Information
共Table K兲 of Ref.31.
TABLE II. Adiabatic and vertical excitation energies of the A˜ ←X˜ transition
of C6H5O derived with various methods.
Method
Adiabatic
共cm−1兲
Vertical
共cm−1兲 Ref.
Cavity ringdown 7681 This work
Matrix isolation 8900 25 Photoelectron 8550 30 CNDO 5 322 36 TD-B3LYP/aug-cc-pVTZ ⬃8 400 9,25, and40 CASSCF共9,8兲/6-31Gⴱ 10 564 34 MRSDCI 12 685 33 B3LYP/aug-cc-pVTZ 7242 8 548 31 B3LYP/aug-cc-pVDZ 7323 8 605 31 CASSCF共9,8兲/aug-cc-pVTZ 6694 10 927 31 CASPT2共9,8兲/aug-cc-pVTZ 8339 9 952 31
C6H5OCH3/N2
共1/72, 220 Torr兲. These figures represent difference spectra between those recorded before laser irradiation and 2 s after laser irradiation at 193 nm; the scan steps are 0.05 nm. Similar spectra were recorded when a flowing mixture of C6H5OC2H5/N2 共1/145, 220 Torr兲 was used; observed ratio
of signal to noise was smaller with this precursor, so we used data from photolysis of C6H5OCH3 for discussion. Because identical spectra were recorded from both precursors, the carrier of the spectrum should be resulted from C6H5, C6H5O, or their reaction products; the latter is less likely because the intensity of the spectrum was the greatest at near-zero delay after photolysis and decreases slightly at longer delays. According to experimental observations46and theoretical calculations,47there is no electronic state of C6H5
near the 1240 nm共⬃8066 cm−1兲 region; the most likely
car-rier of the observed spectrum is thus C6H5O.
The A˜ 2B2←X˜2B1 transition is electronically forbidden, but some vibronic transitions are allowed due to the Herzberg–Teller effect.48 Vibrationally excited states of the
A ˜ 2
B2 state with a2, b1, and b2 symmetries may be directly
excited from the ground vibronic state X˜ 2B1, showing a-, b-,
and c-type rotational contours, respectively, but modes with
a1symmetry are forbidden. Rotational contours of the a-, b-,
and c-type bands simulated with the PGOPHER program49 show distinct band shapes, as indicated in Fig.2. The a-type band shows P and R branches of approximate equal inten-sity, whereas the b-type band shows a prominent Q branch overlapped with the P branch; the c-type band reveals only narrow Q branches. By comparison of the observed band contours with simulations, the band type may be assigned readily. Observed bands were fitted with calculated rotational parameters predicted with B3LYP/aug-cc-pVTZ 共Table I兲,
Jmax= 200, T = 330 K, and spectral width= 1.0 cm−1 共full
width at half maximum for Gaussian functions兲 to derive the band origins, as listed in Table IV. The spectral width is comparable to the step size共0.05 nm兲 of laser scanning.
Because the band origin of the A˜ ←X˜ transition of C6H5O is expected to be missing due to its symmetry-forbidden nature, and because the electronic transition en-ergy calculated quantum chemically is not accurate enough to predict the exact position of vibronic bands, it is difficult to assign these bands by direct comparison of band positions between experiments and calculations. However, we can
TABLE III. Comparison of scaled harmonic vibrational wavenumbers共in cm−1兲 of C
6H5O and C6D5O in their
A
˜ states calculated with various methods.
Mode Sym.
C6H5O C6D5O
Calc.a Calc.b This work Calc.a Calc.b This work
1 a1 3100 3107 2306 2312 2 a1 3091 3098 2295 2301 3 a1 3062 3070 2266 2272 4 a1 1556 1556 1521 1525 5 a1 1410 1393 1274 1269 6 a1 1202 1201 1183 1179 7 a1 1152 1143 930 922 8 a1 1011 1008 862 853 9 a1 961 953 824 816 10 a1 802 797 744 738 11 a1 505 501 495 491 12 a2 941 928 947⫾2 768 758 772⫾2 13 a2 799 787 793⫾6 624 615 626⫾3 14 a2 416 414 417⫾2 364 361 365⫾2 15 b1 955 943 964⫾5 804 799 812⫾7 16 b1 861 847 866⫾3 726 716 17 b1 710 697 723⫾2 585 576 599⫾2 18 b1 673 674 680⫾2 544 539 532⫾2 19 b1 496 490 499⫾5 429 424 436⫾5 20 b1 219 218 208 206 21 b2 3099 3106 2301 2307 22 b2 3067 3075 2273 2279 23 b2 1529 1529 1493 1498 24 b2 1408 1395 1312 1314 25 b2 1304 1291 1218 1228 26 b2 1222 1232 1014 996 27 b2 1139 1130 829 820 28 b2 1058 1052 809 801 29 b2 604 599 581 576 30 b2 358 354 346 343
aScaled vibrational wavenumbers calculated with B3LYP/aug-cc-pVTZ. See text and Ref.31.
bScaled vibrational wavenumbers calculated with B3LYP/aug-cc-pVDZ. See text and Ref.31.
154307-4 Cheng, Witek, and Lee J. Chem. Phys. 129, 154307共2008兲
identify the band type by its rotational contour to provide additional information. There are only three and six vibra-tional modes of C6H5O with a2 and b1 symmetries,
respec-tively. By positive identification of some a-type bands char-acteristic of the a2 vibrational modes and b-type bands
characteristic of the b1vibrational modes of C6H5O, vibronic
assignments are possible.
For the three a-type bands in the region 8080– 8220 cm−1 关Fig.2共B兲兴, the spacing of 50 cm−1 does
not correspond to differences among predicted vibrational wavenumbers, 941, 799, and 416 cm−1共after scaling, Table
III兲, for the three vibrational modes of the A˜ state with a2
symmetry. The separation of⬃50 cm−1agrees satisfactorily
with the energy difference of 38 cm−1 between the
vibra-tional modes of the A˜ state 共20= 219 cm−1兲 and the X˜ state
共20= 181 cm−1兲 with the least energy. The observed relative
intensities of 1.00:0.54:0.14 for these three bands also agree satisfactorily with a Boltzmann distribution at 298 K for lev-els with energy separation⬃200⫾70 cm−1. Hence, we
ten-tatively assigned bands near 8148⫾1 and 8199⫾2 cm−1as
hot bands involving one and two vibrational quanta of 20, whereas the band near 8098⫾1 cm−1is a pure vibronic
tran-sition from the ground level of the X˜ state. Similar hot bands
were also observed for other vibrational modes with a similar spacing and relative intensity, as discussed in following paragraphs.
The next intense a-type band was observed at 8628⫾1 cm−1; hot bands involving transitions 20
1 1 and 20
2 2
were observed at 8678⫾3 and 8728⫾3 cm−1, respectively
关Fig.2共D兲兴. Observed a-type bands at 8098 and 8628 cm−1
may be tentatively assigned to 1401and 1201, respectively, be-cause the separation between these two a-type bands, 530 cm−1, is close to the difference in quantum chemically calculated vibrational wavenumbers of 14= 416 cm−1 and 12= 941 cm−1. The bands at 8678 and 8728 cm−1 are
as-signed as hot bands 12012011 and 12012022, respectively. The remaining a-type band 1301, expected to lie⬃8098 + 799− 416= 8481 cm−1, is partially overlapped with other
b-type bands, but observed contour suggests that this a-type
band might lie near 8474⫾5 cm−1 关Fig.2共C兲兴.
The four prominent b-type bands lie at 8361⫾1, 8404⫾1, 8451⫾2, and 8500⫾3 cm−1 关Fig. 2共C兲兴. The
separation of ⬃48 cm−1 for the last three bands also
indi-cates that bands at 8451 and 8500 cm−1 are hot bands
in-volving transitions 2011 and 2022, respectively, whereas the bands at 8361⫾1 and 8404⫾1 cm−1 are associated with
two separate b1vibrational modes. The intensity of the band
at 8404 cm−1, which is greater than that of the band at
8361 cm−1, also supports that the former is not a hot band of
the latter transition. The six b1 vibrational wavenumbers, 15–20, were predicted to be 955, 861, 710, 673, 496, and
219 cm−1, respectively 共Table III兲. The separation of these
two b-type bands, 43 cm−1, is close to the difference in
vi-brational wavenumbers of 18= 673 cm−1 and 17
= 710 cm−1. The separations of these bands from the tenta-tively assigned 1401band at 8098, 306 and 263 cm−1, respec-8520 8540 8560 8580 8600 8620 8640 8660 8680 8700 8720 8740 8760 0 2 a-type b-type b-type a. u. Wavenumber / cm-1 0 4 8 121 020 2 2 121 020 1 1 121 0 Ab s. /ppm 8280 8300 8320 8340 8360 8380 8400 8420 8440 8460 8480 8500 8520 0 2 a. u. 0 4 8 17102022 Ab s. /ppm 8020 8040 8060 8080 8100 8120 8140 8160 8180 8200 8220 8240 0 2 a. u. 0 4 8 1910 Ab s. /ppm 7660 7680 7700 7720 7740 7760 7780 7800 7820 7840 7860 7880 79000 5 2011 161 020 1 1 15 1 020 2 2 151 020 1 1 131 0 000 1610 15 1 0 171 020 1 1 171 0 181 0 14102022 14102011 1410 a. u. 0 4 8 c-type c-type b-type a-type a-type a-type a-type b-type b-type b-type b-type a-type a-type b-type b-type b-type Simulation Simulation Simulation Simulation D Experiment C Experiment B Experiment Ab s. /p pm (A) Experiment
FIG. 2. Comparison of experimental spectra and simulated spectra of
C6H5O in the region 7660– 8760 cm−1. The notation of the assignments is
described in the text.
TABLE IV. Assignments for observed vibronic bands of the A˜ ←X˜
transi-tion of C6H5O in the 7650– 8750 cm−1region. The experimental shifts from
the origin band 共000兲 are compared with those calculated
quantum-chemically.
Assignment Band type
Obs. 共cm−1兲 Expt. shift 共cm−1兲 Calc. shift 共cm−1兲 000 c 7681⫾1 0 0 2011 c 7732⫾1 51 38 1401 a 8098⫾1 14= 417 416 140120 1 1 a 8148⫾1 467 454 1901 b 8180⫾4 19= 499 496 140120 2 2 a 8199⫾2 518 492 1801 b 8361⫾1 18= 680 673 1701 b 8404⫾1 17= 723 710 17012011 b 8451⫾2 770 748 1301 a 8474⫾5 13= 793 799 17012022 b 8500⫾3 819 786 1601 b 8547⫾2 16= 866 861 160120 1 1 b 8595⫾10 914 899 1201 a 8628⫾1 12= 947 941 1501 b 8645⫾4 15= 964 955 12012011 a 8678⫾3 997 979 150120 1 1 b 8694⫾4 977 993 12012022 a 8728⫾3 1047 1017 150120 2 2 b 8744⫾4 1028 1031
tively, are also close to the quantum-chemically predicted differences of 294 and 257 cm−1 in harmonic vibrational
wavenumbers. Hence the observed b-type bands at 8361 and 8404 cm−1 may be tentatively assigned to 18
0 1 and 17
0 1,
re-spectively. The bands at 8451 and 8500 cm−1are assigned as
17012011 and 17012022, respectively.
Additional b-type bands were observed at 8547⫾2 and 8595⫾10 cm−1关Fig.2共D兲兴; the latter is a hot band
involv-ing 2011. Following a similar argument, the bands at 8547 and 8595 cm−1 are tentatively assigned as 16
0 1 and 16 0 120 1 1,
re-spectively, because the vibrational wavenumber of 866 cm−1 for the former is close to the predicted value of 16
= 861 cm−1.
These five tentatively assigned bands of 1401 at 8098 cm−1, 1201 at 8628 cm−1, 1801 at 8361 cm−1, 1701 at 8404 cm−1, and 16
0
1 at 8547 cm−1 imply a band origin at
7682, 7687, 7688, 7694, and 7686 cm−1, respectively; the
average value of 7688⫾4 cm−1 shows a standard deviation
within expected error of predicted vibrational wavenumbers. Based on the predicted origin of electronic transition at 7688⫾4 cm−1, we performed a careful search near this
re-gion and observed a weak narrow line at 7681⫾1 cm−1 showing a c-type contour关Fig.2共A兲兴. Consideration of pos-sible hot bands with c-type contours, such as 19102001, does not yield satisfactory assignment for this band; hence we tentatively assign this band at 7681 cm−1to the origin of the
A
˜ ←X˜ transition. It is unclear how this band derives its
in-tensity. The possibility of this band being induced by mag-netic dipole transition or Coriolis-induced transition may be ruled out because an extremely weak a-type band is expected from such a mechanism. Based on the assignment of 0
= 7681 cm−1 for C
6H5O, we derived 12= 947⫾2 cm−1, 13= 793⫾6 cm−1, 14= 417⫾2 cm−1, 16= 866⫾3 cm−1, 17= 723⫾2 cm−1, and18= 680⫾2 cm−1.
The P branches of the 150120n
n 共n=0–2兲 series b-type
bands overlapped with the R branches of the 120120nn 共n = 0 – 2兲 series; hence the overlapped regions appear to have more intensity than expected 关Fig. 2共D兲兴. We were able to derive bands at 8645⫾4, 8694⫾4, and 8744⫾4 cm−1 by
curve fitting and assign them to 1501, 15012011, and 15012022, respectively. The observed vibrational wavenumber of 964⫾4 cm−1 is consistent with the predicted value of
15
= 955 cm−1.
Another b-type band, 1901, expected to lie near 7681 + 496= 8177 cm−1, might be attributed to the weak band near 8180⫾4 cm−1 that overlapped with the 14
0 1
2022 band near 8199 cm−1. The last b-type band, 2001, is expected to lie near 7681+ 219= 7900 cm−1. We did not observe this band,
pre-sumably due to poor Frank–Condon overlap and poor signal-to-noise ratio in this region.
We present in Fig. 3 an overview spectrum of the A˜
←X˜ transition of C6H5O 关panel 共A兲兴, and stick diagrams
showing tentatively assigned vibronic bands excluding hot bands 关top of panel 共B兲兴 and predicted vibrational energy levels of the A˜ state and their band types 关bottom of panel 共B兲兴 for comparison; the band types are represented with different line types. Ignoring the weak origin band, we varied the calculated wavenumbers relative to experimental values
to obtain the best overall match and concluded that the only satisfactory match is to assign the aforementioned a-type bands at 8098 and 8629 cm−1to 1401and 1201, respectively, as indicated in Fig. 3共B兲. All a-type and b-type bands except 2001 were observed.
The intensity borrowing of a-type and b-type bands re-quires interaction of the A˜ state with nearby B1and A2
elec-tronic states, respectively. The X˜ 2B1, B˜ 2A2, and C˜ 2B1states
might just serve the purpose. The absence of c-type bands, for transitions to vibrational modes 21–30 of the A˜ state
with b2 symmetry, may be explained by the absence of a
nearby A1 electronic state. The 2A1 state was predicted to
have a vertical excitation ⬃4.88 eV 共39 400 cm−1兲 above
the ground state;31 it might correspond to the absorption band in the region 270⫺300 nm, but whether the symmetry of the upper state is A1 or A2 remains to be
determined.5,6,21,23,25,27–29
The wavenumbers of the observed band origins, their assignments, and the associated vibrational wavenumbers are summarized in Table IV; the latter are also compared with scaled harmonic vibrational wavenumbers calculated quantum-chemically in Table III. All observed vibrational wavenumbers agree with the calculations to less than 4.4% 共without scaling兲 and 1.8% 共with scaling兲.
One would expect to observe vibrational progressions associated with the CO stretching mode of C6H5O because
the A˜ ←X˜ electronic transition involves a substantial change in the length of a CO bond. However, the CO stretching共6兲
mode at 1202 cm−1 has an a
1symmetry and is vibronically
forbidden. Combination bands associated with 6 are likely
to be observed, but they are outside of our range of detection. B. Vibronic bands of C6D5O
Upper traces in panels 共A兲–共D兲 of Fig. 4 present the observed difference spectra in the region 7600– 8620 cm−1
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 0 2 4 6 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 0 2 4 6 8 107600 7700 7800 7900 8000 8100 8200 8300 8400 8500 8600 8700 8800 8900 Vibrational wavenumber / cm-1 251 0 281 0 151 0 161 0 171 0 181 0 191 0 201 0 12 1 0 131 0 141 0 291 0 301 0 T00 261 0 271 0 A bs. /ppm (D) (C) (B) (A) Ab s. /p pm 261 0 271 0 281 0 151 0 161 0 171 0 181 0 191 0 201 0 12 1 0 131 0 141 0 291 0 301 0 T00 Wavenumber / cm-1
FIG. 3.共Color online兲 Comparison of experimental spectrum and calculated
stick spectrum of C6H5O关panels 共A兲 and 共B兲兴 and C6D5O关panels 共C兲 and
共D兲兴 in the region 7660–8900 cm−1. The stick spectra on top are
experi-mental data and those at the bottom are harmonic vibrational wavenumbers, predicted with the B3LYP/aug-cc-pVTZ method, built upon the
experimen-tal transition origin at 7681 cm−1of C
6H5O共7661 cm−1for C6D5O兲; a-, b-,
and c-type bands are indicated with solid, dashed, and short solid lines, respectively.
154307-6 Cheng, Witek, and Lee J. Chem. Phys. 129, 154307共2008兲
共1.16–1.32 m兲 upon photolysis of a flowing mixture of C6D5OCD3/N2共1/72, 220 Torr兲. Observed bands were fitted
with calculated rotational parameters of C6D5O listed in
Table I, Jmax= 200, T = 330 K, and width =1.0 cm−1 using
the PGOPHERprogram49 to derive the band origins, as listed
in TableV. Simulated spectra thus obtained are also shown in Fig.4 for comparison.
The weak band origin of the A˜ ←X˜ transition is shifted from 7681⫾1 cm−1 for C6H5O to 7661⫾1 cm−1 for
C6D5O关Fig.4共A兲兴. The shift of −20 cm−1is consistent with
the quantum-chemical predictions of −15 cm−1based on the scaled harmonic vibrational wavenumbers calculated with B3LYP/aug-cc-pVTZ.31 The a-type 1401 band of C6D5O
shifted from that of C6H5O toward smaller wavenumbers by
⬃72 cm−1; bands at 8026⫾1, 8074⫾2, and 8114⫾5 cm−1
are assigned to 1401 and its associated hot bands 140120n n 共n
= 1 , 2兲 of C6D5O关Fig.4共B兲兴. The spacing of ⬃50 cm−1 for
these a-type bands of C6H5O decreases slightly to
44⫾4 cm−1for C
6D5O; the latter separation agrees
satisfac-torily with the calculated energy difference between 20
= 208 cm−1of the A˜ state and20= 171 cm−1 of the X˜ state.
We thus derived 14= 365 cm−1 for C6D5O, consistent with
the harmonic vibrational wavenumber of14= 364 cm−1
pre-dicted quantum chemically. The experimental isotopic ratio of 365/417=0.875 for C6D5O/C6H5O is nearly identical to
the ratio of 364/416=0.875 derived from the scaled har-monic vibrational wavenumbers. These bands are not as well resolved as those of C6H5O, indicating that an additional
band near 8097⫾4 cm−1, 436 cm−1 above the origin, and its associated hot bands might overlap with these bands. Ac-cording to quantum-chemical calculations, the only b1
vibra-tional mode that has vibravibra-tional wavenumbers⬃436 cm−1is 19 共429 cm−1兲. We thus tentatively assign this band at
8097⫾4 cm−1 to 1901 of C6D5O. The isotopic ratio of
436/499=0.874 for 1901 is close to the ratio of 429/496 = 0.865 from the scaled harmonic vibrational wavenumbers.
The second a-type band at 8630⫾1 cm−1, 12 0 1 of
C6H5O, shifted to 8433⫾1 cm−1 for C
6D5O; hot bands of
12012011, 12012022, and 12012033 were observed at 8480⫾2, 8527⫾5, and 8573⫾8 cm−1, respectively关Fig. 4共D兲兴. The
value of 12= 772 cm−1 implies an isotopic ratio of
772/947=0.815, which agrees with the ratio of 768/941 = 0.816 calculated from the scaled harmonic vibrational wavenumbers for12 of C6D5O/C6H5O. The b-type 150120n n
共n=0–2兲 series of C6D5O overlapped with the 120120n n
series, with the 1501 band separated from the 1201 by ⬃36 cm−1,
according to calculated harmonic vibrational wavenumbers; this separation is close to the separation from the hot band involving 2011. With spectral simulation, we were able to lo-cate the 1501 band at 8473⫾6 cm−1, yielding
15
= 812 cm−1, similar to the predicted value of 804 cm−1. The
isotopic ratio of 812/964=0.842 for 1501is nearly identical to
the ratio of 804/955=0.842 from the scaled harmonic vibra-tional wavenumbers. The 1601 band was predicted to have wavenumbers ⬃42 cm−1 smaller than that of the 1201 band, but we were unable to provide a positive identification of this band.
8400 8420 8440 8460 8480 8500 8520 8540 8560 8580 8600 86200 2
4
b-type b-type a-type a-type b-type a-type a-type a. u. wavenumber / cm-1 0 2 4 6 12102022 121 020 1 1 1210 Ab s. /ppm 8180 8200 8220 8240 8260 8280 8300 8320 8340 8360 8380 84000 2 4 b-type b-type b-type b-type a. u. 0 2 4 6 A bs ./p pm 7960 7980 8000 8020 8040 8060 8080 8100 8120 8140 8160 81800 2 4 a-type a-type b-type a-type a. u. 0 2 4 6 191 0 Ab s. /p pm 7600 7620 7640 7660 7680 7700 7720 7740 7760 7780 7800 78200 2 4 6 17102022 c-type 201 1 c-type 121 020 3 3 15102022 151 020 1 1 131 0 00 0 1510 171 020 1 1 171 0 181 0 14102022 14102011 141 0 a. u. 0 2 4 6 A Experiment a-type Simulation Simulation Simulation Simulation D Experiment C Experiment B Experiment Ab s. /pp m
FIG. 4. Comparison of experimental spectra and simulated spectra of
C6D5O in the region 7600– 8620 cm−1. The notation of the assignments is
described in the text.
TABLE V. Assignments for the observed vibronic bands in the
7650– 8600 cm−1region for the A˜ ←X˜ transition of C
6D5O. The
experimen-tal shifts from the origin band 共000兲 are compared with those calculated
quantum chemically.
Assignment Band type
Obs. 共cm−1兲 Expt. shift 共cm−1兲 Calc. shift 共cm−1兲 000 c 7661⫾1 0 0 2011 c 7710⫾1 49 37 1401 a 8026⫾1 14= 365 364 140120 1 1 a 8074⫾2 413 401 1901 b 8097⫾4 19= 436 429 140120 2 2 a 8114⫾5 453 438 1801 b 8193⫾1 18= 532 544 1701 b 8260⫾1 17= 599 585 17012011 b 8305⫾1 644 622 1301 a 8287⫾2 13= 626 624 17012022 8351⫾5 700 659 1501 b 8473⫾6 15= 812 804 1201 a 8433⫾1 12= 772 768 150120 1 1 b 8518⫾3 857 841 120120 1 1 a 8480⫾2 819 805 15012022 b 8560⫾8 899 878 120120 2 2 a 8527⫾5 866 842 12012033 a 8573⫾8 912 879
According to calculations, the third a-type band is ex-pected to lie near 7661+ 624= 8285 cm−1. In this region a
band near 8287⫾2 cm−1 was observed to lie between two
b-type bands 共assigned as 1701 and 17012011 in the next para-graph兲. We thus tentatively assign this band to 1301of C6D5O.
The 1301 band of C6H5O was estimated to lie
⬃8474⫾5 cm−1, 793⫾6 cm−1 from the origin. The
corre-sponding experimental isotopic ratio of 626/793=0.789 agrees with the ratio of 624/799=0.781 calculated from the scaled harmonic vibrational wavenumbers.
The b-type bands 1801, 1701, and 17102011 of C6D5O were
readily identified at 8193⫾1, 8260⫾1, and 8305⫾1 cm−1,
respectively 关Fig. 4共C兲兴. Hence 18= 532 cm−1 and 17
= 599 cm−1 for C6D5O, consistent with the values of 18
= 544 cm−1 and 17= 585 cm−1 from the calculations. The
corresponding experimental isotopic ratios of 532/680 = 0.782 and 599/723=0.828 agree with the ratios of 544/673=0.808 and 585/710=0.824 calculated from the harmonic vibrational wavenumbers 共TableIII兲.
We also presented in Fig.3an overview spectrum of the
A
˜ ←X˜ transition of C6D5O 关panel 共C兲兴, and two stick
dia-grams showing assigned vibronic bands with hot bands ex-cluded and predicted vibrational energy levels of the A˜ state and their band types 关panel 共D兲兴 for comparison. The satis-factory agreement of the band type and positions between experiments and calculations provides further support for the assignments of observed bands to the A˜ ←X˜ transition of C6H5O and C6D5O. The wavenumbers of the observed band
origins, their assignments, and the associated vibrational wavenumbers of C6D5O are listed in Table V to compare
with harmonic vibrational wavenumbers calculated quantum chemically. All observed vibrational wavenumbers agree with calculations to within 2.4%.
C. Electronic transition energy and the transition origin
High-level calculations predicted the origin to be 7242 and 8339 cm−1 at the B3LYP/aug-cc-pVTZ and
CASPT2共9,8兲/aug-cc-pVTZ level of theory. Experimentally observed origin of the A˜ ←X˜ transition of C6H5O at
7681 cm−1 deviates from these values by 439 and
−658 cm−1, respectively, but is close to their average value
共Table II兲. The 0 value of 7323 cm−1 predicted with
B3LYP/aug-cc-pVTZ is the closest to our experimental value.
The value of 8550 cm−1 reported from photoelectron
experiments30 corresponds approximately to the average value of the two most prominent bands, 1701 at 8404 cm−1
and 1201at 8629 cm−1; an unresolved broadband spectrum in this region would show a peak near 8530 cm−1. This value is also consistent with the vertical excitation energy of 8548 cm−1predicted with the B3LYP/aug-cc-pVTZ method, but much smaller than a value of 9952 cm−1predicted with the CASPT2共9,8兲/aug-cc-pVTZ method. It is unclear why the matrix experiment25yielded a band at 8900 cm−1as it is outside our detection region.
Observation of the origin band of the symmetry-forbidden A˜ 2B2←X˜2B1 transition at 7681 and 7661 cm−1
for C6H5O and C6D5O, respectively, is unexpected.
Al-though we could not positively rule out the possibility that these bands are due to impurity bands, we favor our present tentative assignments because these bands were observed for both C6H5O and C6D5O, with a deuterium isotopic shift of
−20 cm−1 close to the value of −15 cm−1 predicted from
quantum-chemical calculations. However, it should be noted that if the observed c-type band for the origin is true, it implies that the electronic transition 共direct product of
A ˜ 2B
2←X˜2B1兲 might have a relatively strong b1 character.
Consequently the original a-type band 共excitation to the a2
vibrational mode of the A˜ state兲 would have some b-type character, the b-type band 共excitation to the b1 vibrational
mode of the A˜ state兲 would have some a-type character, the originally forbidden transitions to the a1vibrational modes of the A˜ state would have some c-type character, and excitation to the b2vibrational modes of the A˜ state might diminish in
intensity. Our observation of no band associated with excita-tion to b2 vibrational modes of the A˜ state is consistent with
such a scheme, but the mixing between a-type and b-type bands are not obvious, and no c-type bands associated with 901, 1001, and 1101共expected near 8641, 8482, and 8185 cm−1兲 can be positively identified, even though they might be hid-den among other bands. Further theoretical and experimental investigations, such as high resolution spectroscopy, are de-sired to decipher this problem.
V. CONCLUSION
We have recorded in the region 1.14– 1.31 m several rovibronic bands of normal and perdeuterated phenoxy radi-cals in their A˜ ←X˜ transitions with the cavity ringdown tech-nique. For the A˜ 2B2 state, vibrational wavenumbers 共in cm−1兲 of
12= 947, 13= 793, 14= 417, 15= 964, 16= 866, 17= 723, 18= 680, and19= 499 for C6H5O, and 12= 772, 13= 626, 14= 365, 15= 812, 17= 599, 18= 532, and 19
= 436 for C6D5O were determined. Transitions from vibra-tionally excited levels of 20 in the X˜ state were also
ob-served; 20 of the A˜ state is greater by 50 cm−1 than the X˜
state. These experimental results are all in agreement with the quantum-chemical predictions using B3LYP/aug-cc-pVTZ. A weak origin at 7681 cm−1 for the A˜ ←X˜ transition of C6H5O共7661 cm−1for C6D5O兲 with a c-type contour was
observed. Quantum-chemically calculated A˜ ←X˜ electronic transition energies using various methods deviate from ex-periments; the closest value, 7323 cm−1 with a deviation of
⬃5%, was derived with the B3LYP/aug-cc-pVDZ method.
Note added in proof. Our recent MRCI/aug-cc-pVTZ
calculations of the 0-0 adiabatic A←X excitation energy of C6H5O yielded a value of 7243 cm−1, similar to the B3LYP
value reported in Ref.31
ACKNOWLEDGMENTS
We thank C. Western for providing thePGOPHERspectral simulation program, the National Center for High-Performance Computing of Taiwan for computer facilities, and the National Science Council of Taiwan 共Grant Nos.
154307-8 Cheng, Witek, and Lee J. Chem. Phys. 129, 154307共2008兲
NSC96-2113-M-009-025 and NSC96-2113-M-009-022兲 and the ATU project of the Ministry of Education, Taiwan for support.
1D. S. Haynes, in Fossil Fuel Combustion, edited by W. Bartok and A. F.
Sarofim共Wiley, New York, 1991兲, p. 261.
2H. Bockhorn, Soot Formation in Combustion 共Springer-Verlag, New
York, 1995兲.
3C. Y. Lin and M. C. Lin,J. Phys. Chem. 90, 425共1986兲.
4R. G. W. Norrish and G. W. Taylor, Proc. R. Soc. London, Ser. A 234,
160共1956兲.
5F. Berho and R. Lesclaux,Chem. Phys. Lett. 279, 289共1997兲.
6J. Platz, O. J. Nielsen, T. J. Wallington, J. C. Ball, M. D. Hurley, A. M.
Straccia, W. F. Schneider, and J. Sehested,J. Phys. Chem. A 102, 7964
共1998兲.
7Z. Tao and Z. Li,Int. J. Chem. Kinet. 31, 65共1999兲.
8T. Yu, A. M. Mebel, and M. C. Lin,J. Phys. Org. Chem. 8, 47共1995兲.
9K. Tonokura, T. Ogura, and M. Koshi, J. Phys. Chem. A 108, 7801
共2004兲.
10T. J. Stone and W. A. Waters, Proc. Chem. Soc., London 1962, 253.
11W. T. Dixon and R. O. C. Norman, Proc. Chem. Soc., London 1963, 97.
12T. J. Stone and W. A. Waters,J. Chem. Soc. 1964, 213.
13W. T. Dixon and R. O. C. Norman, J. Chem. Soc. 1964, 4857.
14P. Neta and R. W. Fessenden,J. Phys. Chem. 78, 523共1974兲.
15W. T. Dixon and D. Murphy,J. Chem. Soc., Faraday Trans. 2 72, 1221
共1976兲.
16A. Mukherjee, M. L. McGlashen, and T. G. Spiro,J. Phys. Chem. 99,
4912共1995兲.
17G. R. Johnson, M. Ludwig, and S. A. Asher,J. Am. Chem. Soc.108, 905
共1986兲.
18S. M. Beck and L. E. Brus,J. Chem. Phys. 76, 4700共1982兲.
19G. N. R. Tripathi and R. H. Schuler,J. Chem. Phys. 81, 113共1984兲.
20G. N. R. Tripathi and R. H. Schuler,J. Phys. Chem. 92, 5129共1988兲.
21J. Spanget-Larsen, M. Gil, A. Gorski, D. M. Blake, J. Waluk, and J. G.
Radziszewski,J. Am. Chem. Soc. 123, 11253共2001兲.
22D. Pullin and L. Andrews,J. Mol. Struct. 95, 181共1982兲.
23B. Ward,Spectrochim. Acta, Part A 24, 813共1968兲.
24L. J. Johnston, N. Mathivanan, F. Negri, W. Siebrand, and F. Zerbetto,
Can. J. Chem. 71, 1655共1993兲.
25J. G. Radziszewski, M. Gile, A. Gorski, J. Spanget-Larsen, J. Waluk, and
B. J. Mroz,J. Chem. Phys. 115, 9733共2001兲.
26K. Kesper, F. Diehl, J. G. G. Simon, H. Specht, and A. Schweig,Chem.
Phys. 153, 511共1991兲.
27G. Porter and F. J. Wright,Trans. Faraday Soc. 51, 1469共1955兲.
28F. Bayrakceken, S. Aktas, M. Toptan, and A. Unlugedik, Spectrochim.
Acta, Part A 59, 135共2003兲.
29Y. Kajii, K. Obi, N. Nakashima, and K. Yoshima,J. Chem. Phys. 87,
5059共1987兲.
30R. F. Gunion, M. K. Gilles, M. L. Polak, and W. C. Lineberger,Int. J.
Mass Spectrom. Ion Process. 117, 601共1992兲.
31C.-W. Cheng, Y.-P. Lee, and H. A. Witek, J. Phys. Chem. A 112, 2648
共2008兲.
32D. M. Chipman, R. Liu, X. Zhou, and P. Pulay,J. Chem. Phys.100, 5023
共1994兲.
33J. Takahashi, T. Momose, and T. Shida,Bull. Chem. Soc. Jpn. 67, 964
共1994兲.
34R. Liu, K. Morokuma, A. M. Mebel, and M. C. Lin,J. Phys. Chem.100,
9314共1996兲.
35NIST Computational Chemistry Comparison and Benchmark Database,
NIST Standard Reference Database No. 101, edited by R. D. Johnson III 共National Institute of Standards and Technology, Gaithersburg, 2005兲, http://srdata.nist.gov/cccbdb.
36S. Hirata and M. Head-Gordon,Chem. Phys. Lett. 302, 375共1999兲.
37C. Adamo and V. Barone,Chem. Phys. Lett. 314, 152共1999兲.
38Z. Rinkevicius, I. Tunell, P. Sauek, O. Vahtras, and H. Ågren,J. Chem.
Phys. 119, 34共2003兲.
39H. M. Chang, H. H. Jaffe, and C. A. Masmanidis,J. Phys. Chem. 79,
1118共1975兲.
40M. Dierksen and S. Grimme,J. Chem. Phys. 120, 3544共2004兲.
41C.-Y. Chung, C.-W. Cheng, Y.-P. Lee, H.-Y. Liao, E. N. Sharp, P. Rupper,
and T. A. Miller,J. Chem. Phys. 127, 044311共2007兲.
42K. W. Busch and M. A. Busch, Cavity-Ringdown Spectroscopy: An
Ultratrace-Absorption Measurement Technique共American Chemical
So-ciety, Washington, DC, 1999兲.
43C.-Y. Chung, J. F. Ogilvie, and Y.-P. Lee,J. Phys. Chem. A 109, 7854
共2005兲.
44L. S. Rothman, C. P. Rinsland, A. Goldman, S. T. Massie, D. P. Edwards,
J.-M. Flaud, A. Perrin, C. Camy-Peyret, V. Dana, J.-Y. Mandin, J. Schroeder, A. McCann, R. R. Gamache, R. B. Wattson, K. Yoshino, K. V.
Chance, K. W. Jucks, L. R. Brown, V. Nemtchinov, and P. Varanasi,J.
Quant. Spectrosc. Radiat. Transf. 60, 665共1998兲.
45G. Herzberg, Molecular Spectra and Molecular Structure, Infrared and
Raman Spectra of Polyatomic Molecules, Vol. II 共Krieger, Malabar,
1991兲.
46J. G. Radziszewski,Chem. Phys. Lett. 301, 565共1999兲.
47S. M. Mattar, J. Phys. Chem. A 111, 251共2007兲.
48G. Herzberg and E. Teller, Z. Phys. Chem. Abt. B 21, 410共1933兲.
49C. M. Western,
PGOPHER, a program for simulating rotational structure, University of Bristol, http://pgopher.chm.bris.ac.uk.
