Chemical Physics 402 (2012) 74–82

Contents lists available at SciVerse ScienceDirect

Chemical Physics

journal homepage: www.elsevier .com/locate /chemphys

Dynamics of CACl bond fission in photodissociation of 2-furoyl chloride at 235 nm

Ankur Saha, Hari P. Upadhyaya, Awadhesh Kumar ⇑, Prakash D. Naik, Parma Nand BajajRadiation and Photochemistry Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400 085, India

a r t i c l e i n f o a b s t r a c t

Article history:Received 10 October 2011In final form 5 April 2012Available online 23 April 2012

Keywords:2-Furoyl chloridePhotodissociation dynamicsREMPIChlorine atom

0301-0104/$ - see front matter � 2012 Elsevier B.V. Ahttp://dx.doi.org/10.1016/j.chemphys.2012.04.009

⇑ Corresponding author.E-mail address: awadesh@barc.gov.in (A. Kumar).

The photodissociation dynamics of 2-furoyl chloride at 235 nm has been investigated, employing reso-nance-enhanced multiphoton ionization technique and time-of-flight mass spectrometry. Both the Clfragments, Cl(2PJ=3/2, relative quantum yield 0.85 ± 0.11) and Cl⁄(2PJ=1/2), have the recoil anisotropyparameter (b) value close to zero, and show bimodal translational energy distributions. The branchingratio of the high kinetic energy CACl bond scission to the low energy CACl scission is 0.78/0.22. The dom-inant high kinetic energy channel arises mainly because of electronic pre-dissociation. But, the lowenergy channel results from the ground electronic state of 2-furoyl chloride, formed subsequent tonon-radiative relaxation of the initially prepared excited state.

� 2012 Elsevier B.V. All rights reserved.

1. Introduction

Furan, a five-membered aromatic heterocyclic compound, andits derivatives have received great attention, because of the pres-ence of these units in a variety of natural and synthetic products.The large number of reports pertaining to their synthesis indicatesthe continuing importance of these compounds in biology, chemis-try, industry and medicine [1]. Furancarboxylic acids have beenused in cosmetic or pharmaceutical formulations. 2-furancarboxy-lic acid also exhibits antitumor properties [2]. The chlorinatedfurans can appear as byproducts in various industrial processes, e.g.,metallurgy, manufacture of paper [3], or in the fly ash from incin-erator plant [4] (due to the burning of PVC). The chlorinated furansare carcinogenic [3]. Furans are emitted into the atmosphere fromburning fossil fuels, waste, and biomass [5]. Methylfurans havebeen identified as biomarkers of recent exposure to cigarettesmoke in exhaled air [6]; the aromatic furan ring is found in a widerange of aroma chemicals [7]. Furan and other oxygen-containingcyclic molecules are expected to be important combustion inter-mediates due to the heavy use of oxygenated fuel additives in gas-oline in recent years [8]. These compounds have a widespreadnatural occurrence and roles in several areas such as air pollution,petroleum refining, and coal liquefaction and gasification pro-cesses. It is, therefore, of interest to study various reactions, partic-ularly dissociation, in which these compounds may be involved.

Thermolysis of furan has been studied in detail [9–11], andmainly two primary dissociation channels have been identified.The formation of propyne + CO is the lower energy pathway, andthat of acetylene + ketene is the higher energy channel. The non-

ll rights reserved.

biradical mechanism, which involves first hydride shift followedby ring opening, has been proposed for these two channels in furan.The dissociation mechanism of furan is different from a biradicalmechanism of tetrahydrofuran dissociation, which gets initiatedwith the CAO bond cleavage [12,13]. Theoretical calculations [14–16] support these experimental results. In addition to these twochannels operating on the ground state potential energy surface(PES), another primary channel involving a radical pathway(HCO + propargyl, H2CAC„CH, radicals) opens up, likely on anexcited PES, on UV excitation of furan [17]. Dissociation of furanhas also been investigated by electron impact excitation, producingelectronically excited atomic and molecular fragments [18].Reasonably good amount of work has been carried out on dissociationof furan, but not on its derivatives. Isomerization and fragmentationof 2-furancarboxylic acid and its 5-substituted derivatives havebeen investigated under dissociative electron capture (DEC) [19].However, thermochemical studies, both experimental and theoret-ical, have been reported on some furan derivatives, such as alkylfu-rans [20], furancarboxylic acids [19,21] and halofurans [22].

In the present work, we have investigated photodissociationdynamics of 2-furoyl chloride (furan-2-carbonyl chloride) on exci-tation at 235 nm, and detected the Cl fragment, in both the groundCl(2P3/2) and the spin–orbit excited Cl⁄(2P1/2) states, employing res-onance-enhanced multiphoton ionization. Each channel showed abimodal velocity distribution of the Cl fragment, with an isotropicangular distribution. We have focused on the dissociation mecha-nism of 2-furoyl chloride to understand whether it involves abiradical (like tetrahydrofuran) [12,13], a non-biradical (like furanfrom the ground electronic state) [9–11] or a different mechanisminvolving the side-chain for Cl generation. The dynamics of Cl for-mation is compared with that in similar a, b-unsaturated carbonylchlorides.

A. Saha et al. / Chemical Physics 402 (2012) 74–82 75

2. Experimental

The photodissociation dynamics of 2-furoyl chloride has beenstudied in a molecular beam (MB), using resonance-enhanced mul-tiphoton ionization (REMPI) with time-of-flight (TOF) mass spec-trometer, and the photoproducts Cl(3P 2P3/2) and Cl⁄(3P 2P1/2)detected state-selectively. The experimental set-up has been dis-cussed in detail in our earlier work [23,24]. Briefly, a pulsed molec-ular beam of 2-furoyl chloride (�5%) in He was generated,employing a solenoid valve (General valve), with 800 lm nozzleand 500 ls opening time, and a conical skimmer (1.9 mm diame-ter). A partially focused laser beam intercepts the molecular beam,and generates ions, which were detected using a detector systemconsisting of a two-stage Wiley–McLaren [25] TOF mass spectrom-eter, with extraction and acceleration regions. The detector systemwas mounted vertically, perpendicular to the horizontal MB. Afterpassing through the acceleration region, the ion packet passedthrough a 1035 mm long field-free flight tube to the detector.Two deflector plates, placed perpendicular to the detector axis(z axis), guided translation of the ion packet in the (x,y) plane andat the center of 18 mm dual microchannel plates (MCP). A single com-pact voltage generator, having multiple output voltage ports, poweredthe TOF ion optics, the deflection plates and the MCP detector.

The atomic fragments Cl and Cl⁄were probed using a (2 + 1)-RE-MPI scheme in the wavelength region of 234–236 nm. The laserbeam was generated by an Nd:YAG laser (Quantel, YG-981-C-20)pumped dye laser (Quantel, TDL 90), operating on rhodamine101 (LC 6400) dye solution in methanol. The fundamental dye laseroutput was frequency-doubled in a KDP crystal, and then mixedwith 1064 nm of the Nd:YAG laser, to obtain an output in the rangeof 230–236 nm. Since one-color experiment was performed, thesame laser beam was employed as a pump and probe for both pho-todissociation of the parent molecule and ionization of the photo-products Cl and Cl⁄ atoms. The laser beam, with energy of �50 lJ/pulse, was focused by a lens of 280 mm focal length. The TOF spec-tra of Cl and Cl⁄ were measured with three different geometries ofhorizontal (laser polarization perpendicular to the detection axis),vertical (laser polarization parallel to the detection axis) and magicangle (54.7� to the detection axis) to estimate the recoil anisotro-pies, if at all, of these products.

The detector output was fed to a digital storage oscilloscope, fordigitization and averaging, to obtain the time-of-flight profiles of Cland Cl⁄. The REMPI spectra of Cl and Cl⁄ were measured by scan-ning the laser frequencies through their absorption transitions,and feeding the REMPI signal to a gated boxcar integrator. The rel-ative integrated REMPI intensities of Cl and Cl⁄, respectively, at235.34 nm (42492.5 cm�1, 4p 2D32 3p 2P32) and 235.21 nm(42516.1 cm�1, 4p 2P12 3p 2P12), in combination with their corre-sponding two-photon transition probabilities, were used to extracttheir relative quantum yields. In order to minimize cluster forma-tion, photolysis was performed on the rising edge of the molecularbeam pulse. Moreover, the broadened line width (>1.5 cm�1) wasemployed in REMPI experiments. The Nd:YAG laser was operatedwithout an injection seeder, and the dye laser was operating with-out a set of prisms in the laser cavity, used to narrow the linewidth. The laser line width was always greater than the Dopplerprofile of the recoiling fragment.

The 2-furoyl chloride sample (Alfa Aesar, 98% purity) was usedas supplied, after degassing and purification by several freeze–pump–thaw cycles.

3. Theoretical methods

Ab initio molecular electronic structure theories were employedto investigate the Cl formation channel in 2-furoyl chloride on UV

excitation, using Gaussian 03 program [26]. The ground stategeometries of 2-furoyl chloride, along with Cl and its co-fragmentradical, were optimized at the B3LYP/6-311 + G(d,p) level of the-ory, and energies of these geometries were calculated at the Mol-ler–Plesset MP4(sdq) level with the same basis sets. Theharmonic vibrational frequencies and force constants werecalculated, to ensure that the stationary points on the potentialenergy surface are true saddle points, and also for zero point energycorrection. All the reported energy values are zero point energycorrected. The relative energies at MP4(sdq)/6-311 + G(d,p) andQCISD/6-311G(d,p) levels gave the CACl bond dissociation energyof 76.0 and 73.3 kcal/mol, respectively. We have used an averagevalue of 74.6 kcal/mol as the CACl bond dissociation energy in 2-furoyl chloride. The bond dissociation energy can be comparedwith the reported CACl bond dissociation energy of 71.2 kcal/molin allyl chloride [27] and 83.2 and 86.8 kcal/mol in acryloyl chlo-ride [28,29]. The CACl bond dissociation energy of 71.2 kcal/molin allyl chloride is lower than a typical CACl single bond dissocia-tion enthalpy of P84 kcal/mol, due to the resonance stabilizationof the allyl radical produced [27]. The bond dissociation energyin acryloyl chloride suggests almost no effect of the resonancestabilization of the acryloyl radical produced. But, an intermediateeffect of the resonance stabilization on the CACl bond dissociationenergy is observed in 2-furoyl chloride, due to an extended conju-gation of the ring p electrons.

Excited electronic state calculations were performed at the timedependent density functional theory (TD-DFT) level, using cc-pVDZand cc-pVTZ basis sets, and the electronic states of 2-furoyl chlo-ride, accessible at 235 nm, were predicted from the vertical excita-tion energies and the molecular orbital analysis. These calculationspredict the excitation of 2-furoyl chloride at 235 nm is due to thep⁄(ring, C = O) p(ring) transition. Structures of a few lower ex-cited electronic states were optimized at the configuration interac-tion with single excitation (CIS) level, using 6-311 + G basis sets.The S1 state could not be optimized with larger basis sets. Any at-tempt to optimize the S1 state with larger basis sets leads to a tran-sition state (TS) structure with the C atom of the COCl groupmoving out of the molecular plane. Therefore, all excited stategeometry optimization is carried out at the CIS/6-311 + G level oftheory. We searched for the transition state structures for the CAClbond dissociation, and located the same on only the lowest excitedsinglet state, S1 of 2-furoyl chloride.

4. Analysis of experimental results

Photodissociation of 2-furoyl chloride at 235 nm generates Cl,in both the ground (2P32) and the excited (2P12) spin orbit states,which was detected using (2 + 1)-REMPI and TOF mass spectrome-try. A typical time-of-flight mass spectrum measured on photodis-sociation of 2-furoyl chloride at �235 nm, resonant to the Cl REMPIline, is shown in Fig. 1. The observed TOF profiles of Cl and Cl⁄wereanalyzed, using a forward convolution method, which is describedin detail in earlier publications [23,30,31], to extract the speed dis-tribution g(v), translational energy distribution and the recoilanisotropy parameter (b). However, a brief account of the analysisprocedure is provided in this paper. The experimental TOF profiles,I(t,v), at a particular geometry v (the angle between the laserpolarization and the detection axis), are transformed into thevelocity domain, I(vZ,v). Under space focusing conditions, a simplelinear relation [32]:

vz ¼qVexðt � t0Þ

mð1Þ

is used for this transformation, where q and m are the charge andmass of the photofragment Cl, Vex is the electric field in the

Fig. 1. Time-of-flight mass spectra on photodissociation of 2-furoyl chloride on theCl resonant line.

76 A. Saha et al. / Chemical Physics 402 (2012) 74–82

extraction region, and t(t0) is the TOF(mean TOF). The observedvelocity domain TOF signal, I(vZ,v), results from the convolutionof the distribution of vZ, f(vZ,v), with an instrument responsefunction. The function f(vZ,v) is given by an equation [30,31,33,34]:

f ðvz;vÞ ¼Z 1

IvzI

gðvÞ2v 1þ bP2ðcos vÞP2

vz

v

� �h idv ; ð2Þ

where vZ is the velocity component along the z axis, v is the recoilspeed of the fragments, b is the anisotropy parameter, P2(cosv) isthe second-order Legendre polynomial, and cosðvÞ ¼ e � z, implyingthe projection of the pump laser electric field e on the detection axisz. The anisotropy parameter b is given by an equation:

b ¼ 2 P2 cos hmð Þh i ¼ 3 cos2 hm � 1; ð3Þ

where hm is the molecular frame angle between the molecular tran-sition dipole moment, l, and the fragment recoil direction, v (thebond dissociation co-ordinate). The instrumental response functionwas determined to be a Gaussian function in the time domain, withFWHM of 27 ns [23], by investigating (1 + 1)-REMPI of aniline beamat 34029 cm�1 (293.77 nm). Thus, under space focusing conditionsit generates a convolution function in the velocity domain, whichdepends linearly on the electric field in the extraction region. In thiswork, we have employed a procedure of non-core sampling data,with a valid assumption that the nature and the shape of the TOFprofiles for the Cl fragment are independent of the probe polariza-tion. But, the presence of atomic v � j correlations can make thisassumption only approximate. However, these correlations are gen-erally weak enough to be neglected [35,36].

To extract the photofragment speed distribution, gi(v), andanisotropy, bi, of each decay channel i contributing to the experi-mental TOF profiles, I(vZ,v), an initial c.m. photofragment speeddistribution gi(v) is assumed for each active decay channel i. TheTOF profile at the magic angle, I(vZ,54.7�), yields an estimate ofthe total c.m. speed distribution g(v), and thus provides an indica-tion of the form of the individual speed distributions gi(v). Theseare usually modeled with the functional form [37,38]

giðvÞ ¼ ðfTÞaii 1� ðfTÞi� �bi ; ð4Þ

where (fT)i is the fraction of the available energy channeled intotranslational modes, (ET/Eavl)i and ai and bi are adjustable parame-ters. Using an adjustable anisotropy parameter bi and weight foreach decay channel, the function f (vZ,v) is simultaneously calcu-lated for the geometries v = 0�, 54.7� and 90�. Convolution withthe instrument response function yields simulated TOF profiles,

which can be compared with the experimental results, I(vZ,v).The parameters are then adjusted to achieve a satisfactory agree-ment with the experimental data. The photofragment speed distri-butions, thus determined, are transformed into centre-of-masstranslational energy distributions P(ET), using the equation:

PðETÞ ¼ gðvÞ dvdET

: ð5Þ

5. Results

2-Furoyl chloride in molecular beams was irradiated at 235 nmto detect the Cl radical, using REMPI-TOF-MS. Both Cl and Cl⁄ couldbe easily detected. But, the (2 + 1)-REMPI signal at m/e = 36(HCl)and MPI signal at m/e = 63(COCl) could not be detected under sim-ilar experimental conditions. The result suggests that HCl molecu-lar elimination is either not taking place, or it is a negligiblechannel. Similarly, the CAC bond cleavage, producing COCl, isprobably not a primary channel.

5.1. REMPI detection of Cl on photodissociation of 2-furoyl chloride at235 nm

Photodissociation of 2-furoyl chloride at 235 nm generates Cl,in both the ground (2P3/2) and the excited (2P1/2) spin orbit states,which was detected using (2 + 1)-REMPI and TOF mass spectrome-try. The laser power dependence of the REMPI intensity has a linearlog–log plot with a slope of �3.0 (one photon for photolysis andtwo for REMPI detection) in the range of laser power used. The re-sult suggests (2 + 1)-REMPI of chlorine atom, produced in a singlephoton process, assuming that the ionization step is saturated.That Cl and Cl⁄ are produced in a single photon process was vali-dated by additional experiments. The shape and the width of TOFprofiles of a halogen atom are systematically measured at variouslaser energies, and observed to be independent of laser energies.However, at high energy a slight increase in the width of the TOFprofile was observed, implying a contribution from additionalsources, such as secondary dissociation or multiphoton process.Hence, all the experiments were performed in the low energy re-gion to ensure that the profiles remain the same, and the singlephoton condition is maintained.

5.1.1. Anisotropy parameters (b) and translational energy distributionsof Cl and Cl⁄

The analysis of the measured time of flight profiles of Cl (shownin Fig. 2) and Cl⁄ (depicted in Fig. 3) with different laser polariza-tions, at the v values of 0�, 54.7� (the magic angle) and 90�, resultsin the b values of �0 for both Cl and Cl⁄ fragments. Thus, the TOFprofiles of Cl and Cl⁄, in photodissociation of 2-furoyl chloride at235 nm, are independent of the laser polarization. The b values,along with other observables, are given in Table 1. The b value ofzero implies an isotropic angular distribution of the fragments,suggesting slow (with respect to the rotational period of 2-furoylchloride) formation of Cl and Cl⁄. Thus, 2-furoyl chloride does notdissociate from an initially excited state to produce Cl and Cl⁄,rather it can dissociate from the ground state or an electronicallyexcited state produced after non-radiative relaxation.

We also obtained photofragment speed distribution from theforward convolution analysis of the TOF profiles of Cl and Cl⁄. Theanalysis reveals two velocity components (fast and slow) in the dis-tribution of both Cl and Cl⁄ produced in photodissociation of 2-fur-oyl chloride at 235 nm. Thus, both Cl and Cl⁄ originate from twodifferent dissociation channels. Using the speed distributions,photofragment translational energy distributions of Cl and Cl⁄ aredetermined, and depicted in Fig. 4. The relative translational energies

Fig. 2. REMPI-TOF profiles of Cl(2P3/2) produced from the 235 nm photodissociationof 2-furoyl chloride. The circles are the experimental data and the solid line is aforward convolution fit. Three panels, namely, upper, middle and lower panels,correspond to v = 90�, 54.7� (magic angle) and 0� experimental geometries,respectively.

Fig. 3. REMPI-TOF profiles of Cl(2P1/2) produced from the 235 nm photodissociationof 2-furoyl chloride. The circles are the experimental data and the solid line is aforward convolution fit. Three panels, namely, upper, middle and lower panels,correspond to v = 90�, 54.7� (magic angle) and 0� experimental geometries,respectively.

A. Saha et al. / Chemical Physics 402 (2012) 74–82 77

for the fast (high energy) and slow (low energy) Cl channels are7.0 ± 2.0 and 1.5 ± 0.5 kcal/mol, respectively. The fractions of thehigh and low translational energy components are determined tobe 75% and 25%, respectively, for the Cl channel. Similarly, the aver-age translational energies for the fast and slow Cl⁄ channels are9.5 ± 2.0 and 0.8 ± 0.5 kcal/mol, respectively. For the Cl⁄ channelas well, the fast component contributes predominantly to the totalyield, accounting for about 95%. This result is similar to that in pho-todissociation of acryloyl chloride at 235 nm, in which the majorityof the translational energy distribution is contributed from fast Clatoms [39]. The fraction of the available energy going to the relativetranslation, fT, of the slow Cl(Cl⁄) channel is �0.03(�0.02). These fT

values are explained well with the statistical models [40,41] of en-ergy partitioning, which predict these values to be less than 6%. Thestatistical dissociation of the CACl bond, leading to slow Cl and Cl⁄

channels, suggests that the slow fragments are produced from theground electronic state of 2-furoyl chloride. However, the fT valuesfor the fast Cl and Cl⁄ channels are higher, 0.15 and 0.21, respec-tively. These fT values are underestimated by the statistical models,but overestimated by the soft impulsive model [42], which predictsthe value to be 0.35. Thus, the fast Cl and Cl⁄ channels do not orig-inate from a repulsive surface of 2-furoyl chloride, rather these areexpected to have exit barriers.

5.1.2. Relative quantum yields of the Cl and Cl⁄ productsThe spectral profiles of REMPI lines of Cl and Cl⁄ were measured

at 235.34 and 235.21 nm, respectively (shown in Fig. 5), and the

relative quantum yields of Cl, U(Cl), and Cl⁄, U(Cl⁄), were evalu-ated. The relative quantum yields were extracted from the inte-grated line intensities, I(Cl)/I(Cl⁄), which are directly proportionalto the number density of Cl/Cl⁄, N(Cl)/N(Cl⁄), and can be expressedas in Eq. (6):

NðCl�ÞNðClÞ ¼ k

IðCl�ÞIðClÞ ; ð6Þ

where the constant k is the relative ionization probability, i.e., therelative REMPI line strength of Cl as compared to that of Cl⁄. The va-lue of k is taken to be 0.85 ± 0.10 from literature [43]. The value ofU(Cl⁄) was determined from the number density ratio to be0.15 ± 0.02, using the expression:

U Cl�ð Þ ¼ NðCl�ÞNðClÞ þ N Cl�ð Þ

and UðClÞ ¼ 1�U Cl�ð Þ: ð7Þ

These values are reported in Table 1. Similar relative yield of Cl⁄

(0.19) has been reported in photodissociation of acryloyl chlorideat 235 nm [39]. In addition to the spin–orbit branching ratio, theCl product population distribution is also sub-divided based onthe relative amounts of the measured low and high kinetic energycomponents. The branching ratio of the high kinetic energy CAClbond scission to the low energy CACl bond scission is 0.78/0.22.

Table 1The excitation energy (hm), dissociation energy of the CACl bond (D0), available energy (Eavl), the average translational energy, hETi value, the relative quantum yield (U) and therecoil anisotropy parameter (b) for photodissociation of 2-furoyl chloride at 235 nm, leading to Cl and its co-fragment radical. All energies are in kcal/mol.

Reaction channel hm D0 (CACl) Eavl hETi U b

Fast Slow

c-C4H3O CO + Cl 122 74.6 47.4 7.0 ± 2.0 1.5 ± 0.5 0.85 ± 0.11 �0c-C4H3O CO + Cl⁄ 122 74.6 45.0 9.5 ± 2.0 0.8 ± 0.5 0.15 ± 0.02 �0

Fig. 4. Centre-of-mass recoil translational energy distributions for Cl(2P3/2) andCl⁄(2P1/2) derived from Figs. 2 and 3, respectively, in the photodissociation of 2-furoyl chloride at 235 nm. The dashed lines indicate the speed distributions for thefast and slow components of the chlorine atom formation channel; the solid lineshows the sum.

Fig. 5. REMPI spectra of Cl and Cl⁄ atoms produced in the 235 nm laser photolysis of2-furoyl chloride used for the determination of their ratio.

78 A. Saha et al. / Chemical Physics 402 (2012) 74–82

5.2. Theoretical results

The optimized structures of the ground state of 2-furoyl chlo-ride suggest two conformers, s-trans and s-syn with respect tothe CAC single bond, of almost equal stability (within 0.4 kcal/mol, with a greater stability for trans). Both syn and trans struc-tures are planar (dihedral angle ClCCO of 0� and 180�, respectively)with the barrier of rotation to be 6.9 kcal/mol. These trans, syn andTS structures are depicted in Fig. 6 as S0-trans, S0-syn and S0–TS,respectively.

Molecular orbital calculations predict that the S1 and the T2

states originate from the p⁄(C@O, ring) n(Cl,O) excitation,involving the non-bonding electrons on mainly Cl or O atom ofthe COCl group to the antibonding p orbitals of C@O and the ring.The optimized S1 geometry (structure marked as S1 in Fig. 6) re-veals an appreciable increase in the C@O bond length from 1.18in S0 to 1.27 Å. The angle ClCO is decreased from 120.3 to 113.4.These significant structural changes in the S1 state suggest thatthe electronic excitation remains mostly localized to the C@Ogroup of the COCl side-chain, and the furan ring remains almostunaffected. The nature of excitation of the T2 state of 2-furoyl chlo-ride is similar to that of the S1 state.

MO calculations predict that both the S2 and T1 states originatefrom the p⁄(C@O, ring) p(ring) excitation, with the excitation

mostly localized in the ring. The S2 structure (depicted as S2 inFig. 6) suggests that the C@C bond lengths of the ring are immen-sely affected, with the C@C bond length (1.37 Å) of the ring, at-tached to the COCl group, acquiring a single bond character(1.43 Å) in the S2 state. The structures of the S2 and the T1 statesare almost similar.

Since the translational energy release predicts presence of exitbarriers for fast Cl and Cl⁄ channels, we attempted to locate corre-sponding TS structures on different PESs of 2-furoyl chloride, andlocated the same on only the S1 state. The TS structure, optimizedat CIS/6-311 + G level of theory, has greatly extended CACl bond of2.35 Å (shown as S1–TS in Fig. 6) and reduced angle ClCO to 89.0�from 113.4� in the S1 state. Although the TS structure is non-planarwith the dihedral angle ClC1C2O of 100�, the planarity of the furanring, including the C atom of the COCl group, is still retained. TheTS for the CACl bond cleavage is predicted to have relative energiesof �30 and �120 kcal/mol with respect to the S1 and S0 minima,respectively. These energies of the TS are approximate (shown inFig. 7), since theoretical calculations are at the CIS level with smal-ler basis sets. However, characterization of the TS for the CAClbond scission establishes presence of an exit barrier (�44 kcal/mol). The TS for the CACl bond cleavage is also expected on theT2 surface, since the nature of the T2 state is similar to that of theS1 state. However, we failed to locate the TS on the T2 surface.Higher level CASSCF calculations are required for prediction ofaccurate energies of TS, and to understand whether the CACl bondscission takes place only on the S1 state of 2-furoyl chloride, orsome other states are also involved.

5.3. Discussion

5.3.1. Excitation of 2-furoyl chloride at 235 nm2-Furoyl chloride is expected to have similar absorption spectra

as that of benzoyl chloride or probably benzoic acid, since in both

Fig. 6. The optimized structures of the ground (first, showing two conformers with corresponding TS, and third, depicting TS for CO elimination, rows) and excited electronic(second row, showing S1, TS for Cl formation from S1, and S2 states) states of 2-furoyl chloride, along with transition states. Details are given in the text. A few important bondlengths (in Å) and dihedral angles are marked on structures. Dissociating and forming bonds are depicted as dotted lines.

Fig. 7. Relative potential energy diagrams for formation of Cl from the excitedelectronic state (S1), and that of CO from the ground state of 2-furoyl chloride.Details are given in the text. All the energies are in kcal/mol.

Fig. 8. Gas phase UV absorption spectra of 2-furoyl chloride (0.6 Torr) recorded atroom temperature.

A. Saha et al. / Chemical Physics 402 (2012) 74–82 79

the cases p electrons of the ring are in conjugation with that of theC@O group of carboxylic acid or acid chloride. In carboxylic acid,the S1 state arises from an electronic excitation located in theC@O group, and it is of np⁄ character [44]. The S2 state is statedto have pp⁄ character. Since the UV absorption spectra of 2-furoylchloride in the vapor phase are not available in literature, the samehave been measured (shown in Fig. 8). The room temperatureabsorption spectra are somewhat structured and show two bands.

One band from 275 to �230 nm with the maximum absorption at256 nm, and the other weak band from �230 to 200 nm with abroad maximum at 221 nm. In addition, very weak absorption isobserved near 280 nm. In comparison to absorption spectra of ben-zoic acid, the absorption bands around 280 and 221 nm are verymuch reduced and the band at 256 nm is enhanced in 2-furoylchloride absorption. These absorption features have been assignedbased on molecular orbital (MO) calculations and compared withreported UV spectra of benzoic acid and acryloyl chloride.

Vertical excitation energies were calculated, employing time-dependent (TD) density functional theory (DFT), using the cc-pVDZand cc-pVTZ basis sets for various electronic transitions of 2-furoylchloride to understand the nature of excitation at 235 nm. Molec-ular orbitals participating in different electronic transitions were

80 A. Saha et al. / Chemical Physics 402 (2012) 74–82

analyzed, and found that the S1 state involves excitation of non-bonding electrons on Cl and O atoms of the COCl group to the anti-bonding p⁄ orbitals, mainly of the side-chain C@O group. However,the S2 state involves the excitation of the p electrons of the ring tothe antibonding p⁄ orbitals, mainly C@C of the ring. The charactersof triplet T1 (3pp⁄) and T2 (3np⁄) states are similar to that of S2

(1pp⁄) and S1 (1np⁄) states, respectively. The calculated verticalexcitation energies from the ground state to the S1, S2 and S3 stateswith the cc-pVDZ basis sets are 106.4 (268.6 nm), 116.9 (244.6 nm)and 130.6 (218.9 nm) kcal/mol, respectively, along with theirrespective oscillator strengths of 0.0000, 0.37 and 0.0001. How-ever, with the cc-pVTZ basis sets the calculated vertical excitationenergies to the S1, S2 and S3 states are 105.8 (270.1 nm), 114.5(249.8 nm) and 129.6 (220.6 nm) kcal/mol, respectively, along withsimilar respective oscillator strengths. These calculated peak posi-tions of 270.1, 249.8 and 220.6 nm, corresponding to S1, S2 and S3

states respectively, with higher basis sets are in a relatively betteragreement with the experimental values of 280, 256 and 221 nm.Although the S1(1np⁄) state (with oscillator strength of zero) in 2-furoyl chloride is not expected to be observed experimentally inthe energy region below the S2(1pp⁄) state, the observed weakshoulder at �280 nm can be assigned to it. A non-zero transitionprobability for the S1 state can be due to vibrational coupling. Thus,the excitation of 2-furoyl chloride at 235 nm (122 kcal/mol) isexpected to lead the molecules mainly to the S2 state (1pp⁄). Thenature and vertical excitation energies of the lower excited statesof 2-furoyl chloride are similar to that reported for benzoic acid[44,45]. Electronic transitions in 2-furoyl chloride can also becompared to that in acryloyl chloride (H2C@CHCOCl), since theseare a, b-unsaturated carbonyl chlorides. Theoretical calculations[29] on acryloyl chloride predict that both S1 and T2 states originatefrom the C@O p⁄ n excitation, whereas the S2 and T1 statesoriginate from the C@C p⁄ p excitation. Thus, the assignmentof these four transitions is similar to that in 2-furoyl chloride.

5.3.2. Mechanism of Cl formationUV excitation of saturated alkyl chlorides generally leads to the

CACl bond fission from a repulsive state r⁄(CACl) with an impul-sive mechanism, producing Cl and Cl⁄ with an anisotropic angulardistribution. However, the dynamics of the CACl bond fission inunsaturated alkyl chlorides involving a p⁄ p transition, such asallyl chloride [27], propargyl chloride [46] and 2-chloropropene[47] are different. In these cases, bimodal velocity distributionsof Cl and Cl⁄ are observed. In these two CACl bond scission chan-nels, the high translational energy channel is assigned to originatefrom an electronic pre-dissociation, whereas the low energy chan-nel operates from the ground electronic state, following internalconversion (IC), from an initially excited state. Even in unsaturatedalkyl acid chlorides, such as fumaryl chloride [48] and acryloylchloride [28], two similar CACl bond fission channels are observedon UV excitation.

2-Furoyl chloride (c-C4H3O COCl, with c-standing for cyclic), anunsaturated acid chloride, also has two CACl bond fission channelson excitation at 235 nm. But, what is the mechanism of Cl atomformation from 2-furoyl chloride? It can dissociate to produce Clin broadly three possible pathways. The Cl fragment can be pro-duced as a primary (reaction (8)), or a secondary (reaction (9))product from the primary product COCl.

c-C4H3O COCl ! c-C4H3O COðR1Þ þ Cl 2P3=2;2P1=2

� �; ð8Þ

c-C4H3O COCl ! c-C4H3OðR2Þ þ COCl; ð9aÞCOCl ! COþ Cl 2P3=2;

2P1=2� �

: ð9bÞ

These two dissociation channels involve only the side-chain with-out ring-opening. The third possible mechanism can involve firstthe ring-opening by a CAO bond cleavage, and the primary birad-

ical can subsequently undergo dissociation or isomerization, fol-lowed by dissociation to produce Cl in one or more pathways.Our calculations at MP4(sdq)/6-311 + G(d,p)//B3LYP)/6-311 + G(d,p) level predict that reactions 8 and 9a are endothermicby 76.0 and 107.7 kcal/mol, respectively. Similarly, the ring-open-ing of 2-furoyl chloride involving the CAO bond cleavage is endo-thermic by 81.1 kcal/mol. Thus, present calculations suggest thatCl formation is more feasible energetically as a primary product(reaction (8)). Another source of Cl (secondary) involving theCACOCl fission (reaction (9)) is a higher energy channel and canproceed with much difficulty, because of the conjugation interac-tion [29]. Thus, although Cl formation involves a side-chain reac-tion, the furan ring cleavage in 2-furoyl chloride is expected tobe a high energy channel.

The proposed mechanism of Cl formation from 2-furoyl chloridecannot be compared with literature, since no information is avail-able on photodissociation of 2-furoyl chloride or even furancarb-oxylic acids. However, the dissociation mechanism can becompared to dissociative electron capture (DEC) of 2-furancarb-oxylic acid, in which only two primary processes are centered atthe carboxylic acid group, and the ring does not open up. Theseprocesses involve COOAH and COAOH bond cleavage, producingH and OH, respectively, along with their co-fragment ions [19].Even the carboxylic acid group is not detached from the furan ring.It is not surprising since the bond between the ring carbon to thecarboxylic carbon is strengthened (acquires a double bond charac-ter), because of resonance between the furan O atom and the COOHgroup through the ring. Similar resonance is possible between thefuran O atom and the COCl group of 2-furoyl chloride, and thusrendering the COCl elimination, as a primary product (reaction(9a)), difficult energetically on excitation of 2-furoyl chloride at235 nm. Even the bonds of the ring-carbon with hydrogen, carbon,fluorine and chlorine atoms in substituted furans are unusuallystrong, and the ring-carbonAH bonds in alkylfurans are amongthe strongest ever CAH bonds recorded [20]. It appears that thephotodissociation of 2-furoyl chloride is grossly similar to the dis-sociation of 2-furancarboxylic acid in DEC. However, the dissocia-tion dynamics of 2-furancarboxylic acid induced by UV light candiffer from that by electrons.

The photodissociation of 2-furoyl chloride can also be comparedto that of benzoyl chloride/benzoic acid, since the furan ring in theformer is also aromatic, like the benzene ring in the latter. Sincethe photodissociation of benzoyl chloride is not reported, the com-parison can be made with that of benzoic acid [44,45,49]. The pho-todissociation of benzoic acid at 248 and 193 nm also involves thecarboxylic group only, and the ring remains intact. Two channelsare observed at 248 nm: the CAOH bond cleavage, generating OHalong with its co-fragment, and the decarboxylation with molecu-lar CO2 elimination [45]. At higher energy, on excitation of benzoicacid at 193 nm, an additional channel, leading to formation ofCOOH and its co-fragment, opens up. It is suggested that manyelectronic states of benzoic acid are involved in the photodissocia-tion processes. Thus, similar to the photodissociation of benzoicacid that of 2-furoyl chloride also involves reactions mainly atthe side-chain. It appears that the photodissociation of benzoylchloride also should lead to the side-chain reaction alone, leavingthe ring uncleaved. We are in the process of investigating the pho-todissociation dynamics of benzoyl chloride at 235 nm [50].

Thus, our earlier discussion suggests that reaction 8 (primaryreaction involving CACl bond dissociation) is responsible for Cl for-mation on excitation of 2-furoyl chloride at 235 nm, with corre-sponding TS in the S1 state. Since the CACl dissociation channelin the S1 state has an exit barrier, this can be responsible for thefast Cl and Cl⁄ formation. A fraction of 2-furoyl chloride in the S1

state can escape dissociation, and undergo non-radiative relaxationto the ground electronic state, from which slow Cl and Cl⁄ can be

S ( )2 1 *ππ S ( n )1 1 *π S0IC IC

Cl / Cl (fast)*

Cl / Cl (slow)*S0hν

Scheme 1. A scheme for chlorine atom formation from 2-furoyl chloride.

A. Saha et al. / Chemical Physics 402 (2012) 74–82 81

produced. The scheme for Cl and Cl⁄ formation (see Scheme 1) from2-furoyl chloride on excitation at 235 nm can be given as follows:

The branching ratio of the high kinetic energy (fast) to the lowenergy (slow) CACl bond scission is measured to be 0.78/0.22,implying, at least, 78% yield of the S1 state of 2-furoyl chloride dis-sociates to produce mainly Cl and Cl⁄. Thus, relaxation of the S1

state to the S0 state by internal conversion is a minor process in2-furoyl chloride. Unlike in 2-furoyl chloride, dissociation in cyclicethers is initiated by the ring cleavage involving the CAO bondcleavage, because the CAO bond cleavage energy in cyclic ethers,such as tetrahydrofuran (74.6/74.8 kcal/mol) [13,51], is lower.Therefore, the thermal and photochemistry of cyclic ethers giverise to unsaturated hydrocarbons and carbonyl compounds. Theultraviolet photolysis of cyclic ethers has been interpreted toinvolve the cleavage of the carbonAoxygen bond as the majorprimary process [12].

In acryloyl chloride also, the CACl bond cleavage is predicted totake place in the S1 and T2 states on UV excitation [29], and re-ported to be a primary dissociation channel [28,39]. The CAOHbond (analogous to the CACl bond of an acid chloride) dissociationis also the major channel from benzoic acid on excitation at248 nm [45], and predicted to occur in the T2 state after IC fromS2 to S1 followed by ISC [44]. The CACOOH bond (analogous tothe CACOCl bond of an acid chloride) dissociation could be ob-served in benzoic acid at higher excitation energy of 193 nm irra-diation, predictably from the T1 state. The role of the S0 state ofbenzoic acid in photodissociation of the CAOH and CACOOH bondsis stated to be negligible.

5.3.3. Comparison of Cl formation in 2-furoyl chloride with other a, b-unsaturated chlorides and acetyl chloride

The dynamics of Cl formation in photodissociation of 2-furoylchloride (which can be considered as an a, b-unsaturated chloride)has several similar aspects as that of a, b-unsaturated chlorides,such as allyl, propargyl and particularly acryloyl chloride with sim-ilar functional group as 2-furoyl chloride. In these chlorides, theCACl bond scission is a primary channel, and both high and low ki-netic energy Cl and Cl⁄ are produced, with predominance of theformer. In 2-furoyl chloride, the low energy component contributesto 25% and 5% only to the Cl and Cl⁄ channels, respectively. But, inallyl [52], propargyl [46] and acryloyl [39] chlorides, the slow com-ponent of the Cl⁄ channel is too small to be detected. The relativequantum yields of Cl and Cl⁄ in 2-furoyl chloride (0.85,0.15) arecomparable to that in acryloyl chloride (0.81,0.19) [39], but are dif-ferent from that in allyl chloride (0.67,0.33) [52,53]. An isotropicangular distributions of Cl and Cl⁄ are observed in these chlorides,with the anisotropy parameter b value close to zero. In the photo-dissociation of acryloyl chloride, the b values of Cl and Cl⁄ are notreported, but these are expected to be nearly zero.

Dissociation reaction in 2-furoyl chloride takes place only in theside chain, and the excited molecule relaxes to the S1 (1np⁄) state toproduce the high kinetic energy Cl channel. Since acetyl chloridealso dissociates from the 1np⁄ state to produce the Cl channel[54,55], it is worth discussing dynamics of Cl formation in 2-furoylchloride relative to that in acetyl chloride on excitation at 235 nm.In acetyl chloride also [54,56], a bimodal distribution of Cl(2P3/2) isobserved, with a minor contribution from the low kinetic energycomponent. However, in contrast to 2-furoyl chloride, acetyl chlo-ride produces anisotropic Cl and Cl⁄ distributions, indicating a

prompt dissociation relative to the molecular rotation. An isotropicdistribution of Cl in 2-furoyl chloride suggests its slow dissociation.This difference probably arises because of different excitation of2-furoyl chloride and acetyl chloride at 235 nm. The excitationpopulates the 1(pp⁄) state of furoyl chloride, which subsequentlyrelaxes to the 1(np⁄) state before dissociation to produce chlorine.But, the excitation of acetyl chloride directly populates the 1(np⁄)state.

5.3.4. Mechanism of CO formationIn addition to the nascent Cl product, detected by REMPI, stable

CO and C2H2 products were detected by FT-IR absorption. The min-or product C2H2 is expected to be produced, as a secondary prod-uct, from the furan ring cleavage. We have not investigated themechanism of its formation. The other stable product CO can beproduced as a secondary product (reaction (10)) from the primaryproduct c-C4H3O CO (R1 produced in reaction (8)), or COCl (reac-tion (9a)) through reaction (9b).

c-C4H3O COðR1Þ ! c-C4H3OðR2Þ þ CO: ð10Þ

Since in the previous section, it is pointed out that reaction (9a) can-not compete energetically with reaction (8), the formation of COinvolving reaction (9a) cannot be considered. Thus, the radical c-C4H3O CO (R1) can undergo the CAO bond cleavage via a TS(marked as TSACO in Fig. 6) to produce CO. The TS has a planargeometry with the CACO bond stretched to 2.56 Å. The activationbarrier and the endoergicity of the reaction are predicted to be31.9 and 29.1 kcal/mol, respectively. The predicted energy of TSbeing very close to the endoergicity of the reaction, suggests a looseTS with the saddle point lying in the exit channel. This barrier for COelimination in 2-furoyl chloride (31.9 kcal/mol) is higher than thatin acryloyl chloride (23.0 kcal/mol) [39] with a similar dissociationmechanism.

The CO product can also be generated as a primary productthrough reaction (11), involving a TS, shown as TSACOACl inFig. 6, with concerted Cl migration to the ring and CO elimination:

c-C4H3O COCl ! c-C4H3O ClðCFÞ þ CO: ð11Þ

In the TS the COCl group is not in the plane of the furan ring. Theactivation barrier and endoergicity of the reaction (shown inFig. 7) are calculated to be 75.6 and 5.4 kcal/mol, respectively. Thus,energetics suggests this mechanism of CO production to be moreplausible. REMPI detection of the nascent CO fragment is expectedto provide in depth dynamical information, and this work isplanned for future.

6. Conclusions

The photodissociation dynamics of 2-furoyl chloride has beeninvestigated, on excitation at 235 nm to its S2 state with 1(pp⁄)character, using resonance-enhanced multiphoton ionization(REMPI) and time-of-flight mass spectrometry. The atomic Clfragments, detected by (2 + 1)-REMPI, are produced in both thespin–orbit ground Cl(2P3/2) and excited Cl⁄(2P1/2) states, with relativequantum yields of 0.85 and 0.15, respectively. These fragments areprimary products, produced from the CACl bond scission of theCOCl side-chain, and show bimodal translational energy distribu-tions with the average high (low) kinetic energy of 7.0 (1.5) and9.5 (0.8) kcal/mol, respectively. The high kinetic energy channelis the dominant one (78%) and arises mainly because of electronicpre-dissociation. Excited state MO computations suggest the CAClbond scission has a barrier, and the corresponding transition statewas located in the S1 state, with the 1(np⁄) character, of 2-furoylchloride. The relative translational energies in the fast Cl and Cl⁄

channels are greater than the statistical limit, but smaller than

82 A. Saha et al. / Chemical Physics 402 (2012) 74–82

the impulsive limit of the energy partitioning models, as expectedfor the dissociation with an exit barrier. The minor CACl bondfission channel results from the ground electronic state of 2-furoylchloride, subsequent to non-radiative relaxation of the initiallyexcited state, and is explained well with the statistical mechanism.The angular distributions of both Cl(2P3/2) and Cl⁄(2P1/2)are nearlyisotropic with the recoil anisotropy parameter (b) value close tozero. Thus, the CACl bond fragmentation of even the fast Cl andCl⁄ is not fast enough to avoid molecular rotation for the parentmolecule, and thus leading to washing out of initial anisotropy.Dynamics of the CACl bond fission in 2-furoyl chloride is grosslysimilar to that in acryloyl chloride, since both are a, b-unsaturatedcarbonyl chlorides. However, unlike in acryloyl chloride, the HClproduct is not observed in 2-furoyl chloride.

Acknowledgments

We thank Dr. S.K. Sarkar and Dr. T. Mukherjee for their constantsupport and encouragement throughout this work. Assistance fromMr. Yogesh Indulkar, Pune University, in theoretical calculations isduly acknowledged.

References

[1] A.R. Katrizky, C.W. Rees, E.F.V. Scriven (Eds.), Comprehensive HeterocyclicChemistry II, vol. 2, Pergamon Press, Oxford, UK, 1996.

[2] Z. Sajadi, M.M. Abrishami, Paricher-Mohseni, J.M. Chapman Jr., I.H. Hall, J.Pharm. Sci. 73 (1984) 266.

[3] L.B. Sonnenberg, K.M. Nichols, Chemosphere 31 (1995) 4207.[4] R. Bechtler, L. Stieglitz, G. Zwick, R. Will, W. Roth, K. Hedwig, Chemosphere 37

(1998) 2261.[5] E.D. Pellizzari, J.E. Bunch, R.E. Berkley, J. McRae, Anal. Chem. 48 (1976) 803.[6] J.J.B.N. Van Berkel, J.W. Dallinga, G.M. Moller, R.W.L. Godschalk, E. Moonen,

E.F.M. Wouters, F.J. Van Schooten, J. Chromatogr. B 101 (2008) 861.[7] D. Rowe, Chem. Biodiversity 1 (2004) 2034.[8] K. Owen, Gasoline and Diesel Fuel Additives, Wiley, New York, 1989.[9] A. Lifshitz, M. Bidani, S. Bidani, J. Phys. Chem. 90 (1986) 5373.

[10] D. Fulle, A. Dib, J.H. Kiefer, Q. Zhang, J. Yao, R.D. Kern, J. Phys. Chem. A 102(1998) 7480.

[11] P.P. Organ, J.C.J. Mackie, J. Chem. Soc. Faraday Trans. 87 (1991) 815.[12] A.A. Scala, E.W.-G. Diau, Z.H. Kim, A.H. Zewail, J. Chem. Phys. 108 (1998) 7933.[13] S. SenGupta, H.P. Upadhyaya, A. Kumar, P.D. Naik, P.N. Bajaj, J. Chem. Phys. 122

(2005) 124309.[14] R. Liu, X. Zhou, L. Zhai, J. Comput. Chem. 19 (1998) 240.[15] R. Liu, X. Zhou, T. Zuo, Chem. Phys. Lett. 325 (2000) 457.[16] K. Sendt, G.B. Bacskay, J.C. Mackie, J. Phys. Chem. A 104 (2000) 1861.[17] O. Sorkhabi, F. Qi, A.H. Rizvi, A.G. Suits, J. Chem. Phys. 111 (1999) 100.

[18] M. Dampc, M. Zubek, Int. J. Mass Spectrom. 277 (2008) 52.[19] V.V. Takhistov, M.V. Muftakhov, A.A. Krivoruchko, V.A. Mazunov, Izv. Akad.

Nauk SSSR. Ser. Khim. 9 (1991) 2049.[20] J.M. Simmie, H.J. Curran, J. Phys. Chem. A 113 (2009) 5128.[21] M.V. Roux, M. Temprado, P. Jiménez, C. Foces-Foces, M.V. Garcıa, M.I. Redondo,

Thermochim. Acta 420 (2004) 59.[22] I. Novak, J. Org. Chem. 66 (2001) 9041.[23] H.P. Upadhyaya, A. Saha, A. Kumar, T. Bandyopadhyay, P.D. Naik, P.N. Bajaj,

J. Phys. Chem. A 114 (2010) 5271.[24] Y.N. Indulkar, A. Saha, H.P. Upadhyaya, A. Kumar, S.B. Waghmode, P.D. Naik,

P.N. Bajaj, J. Chem. Phys. 134 (2011) 194313.[25] W.C. Wiley, I.H. McLaren, Rev. Sci. Instrum. 26 (1955) 1150.[26] M.J. Frisch, Gaussian 03, Revision C.02, Gaussian Inc. in: R.C.G.I. Gaussian 03

(Ed.), Wallingford CT, 2004.[27] M.L. Morton, L.J. Butler, T.A. Stephenson, F. Qi, J. Chem. Phys. 116 (2002) 2763.[28] D.E. Szpunar, J.L. Miller, L.J. Butler, F. Qi, J. Chem. Phys. 120 (2004) 4223.[29] G.-L. Cui, Q.-S. Li, F. Zhang, W.-H. Fang, J.-G. Yu, J. Phys. Chem. A 110 (2006)

11839.[30] K. Bergmann, R.T. Carter, G.E. Hall, J.R. Huber, J. Chem. Phys. 109 (1998) 474.[31] W.S. McGivern, R. Li, P. Zou, S.W. North, J. Chem. Phys. 111 (1999) 5771.[32] J.A. Syage, J. Chem. Phys. 105 (1996) 1007.[33] R. Vasudev, R.N. Zare, R.N. Dixon, J. Chem. Phys. 80 (1984) 4863.[34] M. Dubs, U. Bruhlmann, J.R. Huber, J. Chem. Phys. 84 (1986) 3106.[35] R.J. Wilson, J.A. Mueller, P.L. Houston, J. Phys. Chem. A 101 (1997) 7593.[36] A.S. Bracker, E.R. Wouters, A.G. Suits, Y.T. Lee, O.S. Vasyutinskii, Phys. Rev. Lett.

80 (1998) 1626.[37] A.S. Bracker, S.W. North, A.G. Suits, Y.T. Lee, J. Chem. Phys. 109 (1998) 7238.[38] S.W. North, A.J. Marr, A. Furlan, G.E. Hall, J. Phys. Chem. A 101 (1997) 9224.[39] K.-C. Lau, Y. Liu, L.J. Butler, J. Chem. Phys. 123 (2005) 054322.[40] J.T. Muckermann, J. Phys. Chem. 93 (1989) 179.[41] C.E. Klots, J. Chem. Phys. 58 (1973) 5364.[42] A.F. Tuck, J. Chem. Soc. Faraday Trans. 2 (73) (1977) 689.[43] R. Liyanage, Y.A. Yang, S. Hashimoto, R. Gordon, R.W. Field, J. Chem. Phys. 103

(1995) 6811.[44] J. Li, F. Zhang, W.H. Fang, J. Phys. Chem. A 109 (2005) 7718.[45] Y.A. Dyakov, A. Bagchi, Y.T. Lee, C.-K. Ni, J. Chem. Phys. 132 (2010) 14305.[46] L.R. McCunn, D.I.G. Bennett, L.J. Butler, H. Fan, F. Aguirre, S.T. Pratt, J. Phys.

Chem. A 110 (2006) 843.[47] J.A. Mueller, B.F. Parsons, L.J. Butler, F. Qi, O. Sorkhabi, A.G. Suits, J. Chem. Phys.

114 (2001) 4505.[48] M. Kawade, A. Saha, H.P. Upadhyaya, A. Kumar, P.D. Naik, P.N. Bajaj, J. Phys.

Chem. A 115 (2011) 1538.[49] Q. Wei, J.L. Sun, X.F. Yue, S.B. Cheng, C.H. Zhou, H.M. Yin, K.L. Han, J. Phys.

Chem. A 112 (2008) 4727.[50] A. Saha, M. Kawade, H.P. Upadhyaya, A. Kumar, P.D. Naik, P.N. Bajaj,

unpublished work.[51] J. Kramer, J. Phys. Chem. 86 (1982) 26.[52] Y. Liu, L.J. Butler, J. Chem. Phys. 121 (2004) 11016.[53] M.S. Park, K.W. Lee, K.-H. Jung, J. Chem. Phys. 114 (2001) 10368.[54] S. Deshmukh, J.D. Myers, S.S. Xantheas, W.P. Hess, J. Phys. Chem. 98 (1994)

12535.[55] M.D. Person, P.W. Kash, L.J. Butler, J. Chem. Phys. 97 (1992) 355.[56] X. Tang, B.J. Ratliff, B.L. FitzPatrick, L.J. Butler, J. Phys. Chem. B 112 (2008)

16050.