The State-selected Predissociation Dynamics of CCIF NO A ( n ,

The photodissociation of gaseous CClF2NO at 300 K and <20 K in a supersonic jet has been studied in the wavelength range 568 to 705 nm of the S1(n,π*)←S0 electronic transition. Energy disposal into, rates of formation and product-state distributions of the nascent NO(X2Πi) photofragment have been measured in detail, and have been compared with statistical (Prior) calculations. The rotational population distributions of the NO fragment are statistical up to the highest excess energies investigated, while the spin-orbit branching ratios are very non-statistical and non-thermal, favouring the lower component. The lack of evidence for state-specific effects in the energy disposal and appearance times of the NO indicate that excess energy randomises prior to dissociation. Intersystem crossing to T1 dominates the predissociation of all but the lowest S1 levels. Levels close to the O,O origin predissociate either by internal conversion to the ground state or by vibrational predissociation on an S1 potential energy surface that is weakly bound in the dissociation coordinate. A bond dissociation energy of 13 700 ± 350 cm−1 (164 kJ mol−1) has been evaluated for the C—N bond.


INTRODUCTION
The last ten years have seen a dramatic revival of interest in the study of photofragmentation processes, that continues to be fuelled by rapid experimental advances and accompanying progress in theoretical analysis.1'2 This revival was initiated by the now classic work of Busch and Wilson on molecular beam laser photofragment spectroscopy 3 and later by the emergence of pulsed laser dissociation-probe techniques particularly the powerful laser-induced fluorescence (LIF) method.1,4'5 The marriage of molecular beam and laser dissociation- probe techniques now offers an elegant and extremely informative experimental approach to studying photodissociation processes in microscopic detail, provided one of the fragments is amenable to probing by LIF or resonantly enhanced multiphoton ionisation (REMPI).
Among a wide variety of polyatomic species suitable for photo- dissociation-LIF probe studies, molecules producing the NO ground- state fragment are particularly promising candidates.The spectro- scopy of the A ,-X Rydberg transition of NO is known in consider- able detail, and both single-and two-photon LIF excitation spectra on this transition, commonly denoted the y-band system of NO are well characterised.5-7 The ground state of NO (xEII) is split by spin-orbit interaction giving rise to two rotational manifolds separated by roughly 120 cm-.Orbital-rotation interaction causes further fine- structure splitting of each rotational level into A-doublets that can be discriminated readily by means of the optical selection rules.Thus, LIF or (1 + 1) resonantly enhanced two-photon ionisation techniques permit detailed evaluation of electronic branching ratios, rotational, vibrational and fine-structure populations.In addition Doppler- profile studies can reveal energy and momentum disposal into translation.
In consequence considerable attention has focused recently on photodissociation studies of various nitroso-compounds. 8Both predissociation following excitation via structured electronic absorption bands and fast, direct dissociation from repulsive upper electronic states have been studied using LIF of nascent NO. 9-14 The photodisso- ciation of C-nitroso compounds in the visible in particular appears to provide textbook examples of slow predissociation processes.Such studies permit one to test theories of radiationless decay and of the unimolecular decomposition of long-lived complexes.  Thenetrating studies of Dixon and collaborators of the LIF excitation spectra and photophysics of HNO in the visible region have recently provided a detailed model for the spectroscopy and radi- ationless decay of C-nitroso compounds excited on the A XA" --,e'lA (n,rr*) transition in the 500-720 nm region of the visible spectrum, x6,7 Near the dissociation threshold Renner-Teller coupling between the /IA" and A' states of HNO leads to rotationally assisted internal conversion (IC) to degenerate ground-state levels prior to dissoci- ation.At higher energies, as the density of states in the intermediate tiaA state (T1) becomes larger, dissociation from the triplet surface following spin-orbit induced intersystem crossing (ISC) becomes important. 17 seminal work on the photodissociation of CF3NO Bower et al. demonstrated the power of supersonic jet-cooling in combination with pulsed dissociation-delayed LIF to achieve vibrational state-selection and to obtain detailed product-state distributions of the nascent NO (XEHi) fragment.TM By varying the delay between dissociation and probe pulses they were able to show that the rates of appearance of NO equal the fluorescence lifetimes of CFaNO(AIA"), and that they decrease monotonically with increasing excitation energy above the O0 origin of the n,rr* absorption bands.They concluded that both the dissociation and fluorescence decay rates reflect the rates of internal conversion (IC) from S1 levels to highly vibrationally excited levels of S0, which then dissociate in a nearly statistical fashion on the S0 surface.
In a more recent re-investigation of the spectroscopy, photophysics and photodissociation dynamics of CF3NO we have demonstrated the influence of SI-T1 coupling of all but the lowest lying levels of the,4 A" state.Our work has shown that the absence of significant deviations from statistical (Prior) product-state distributions cannot be taken as evidence for well-behaved, adiabatic unimolecular decay on the ground-state surface. 9A clear example of the influence of ISC on the predissociation dynamics of C-nitroso compounds has been unravelled with jet-cooled (CHa)aCNO by Noble et al. 2 Here excitation of parent features below a 650 cmbarrier above the dissociation threshold leads to slow unimolecular dissociation on the ground-state surface and statistical NO(X2H) product-state distributions.Above this 650 cmbarrier prompt dissociation of (CHa)aCNO produces non- statistical product-state distributions and is proposed to occur from the a3A" surface.This predissociation therefore proceeds in close analogy to the mechanism suggested by Dixon et al. for HNO and DNO, 16,17 but could be demonstrated beautifully with product-state distributions and bi-exponential appearance rates of the nascent NO(X2Hi) photo- fragment. 2s part of a systematic investigation of the spectroscopy, photophy- sics and predissociation dynamics of the AA states of C-nitroso compounds we present here studies of the dissociation dynamics of chlorodifluoronitroso-methane (CC1F2NO).Despite the lower sym- metry, this molecule closely resembles CF3NO revealing a very similar vibrationally structured A --" n,r* electronic transition in the 500-700 nm wavelength region with similar oscillator strength.It predissociates like CF3NO and (CH3)3CNO as summarised in Eq. (1).
CC1FENO (,g) CC1FENO (fi n,r*) CCIF2 + NO(X2rI,) (1) CCIF2NO was prepared in a three-step synthesis starting with the vacuum pyrolysis of PTFE powder to produce tetra-fluoroethene.Addition of NO2 to the olefin in the gas phase and subsequent hydrolysis of the addition products provides difluoronitroacetic acid.The decarboxylation of this acid with hot hydrochloric acid generates crude gaseous CCIF2NO.Various modifications of this synthesis and the purification of CC1F2NO have been described in detail elsewhere.21-24 The compound was stored at -20C in the absence of light and air and repurified prior to use by several trap-to-trap distill- ations in vacuo.Infrared and visible absorption spectra of the blue gas were used to monitor its purity.
Room temperature photodissociation experiments CCIF2NO vapour at a pressure <150 mTorr was pumped slowly from a sample container kept near -100C by means of a slush bath, through a small room temperature LIF flow cell.This cell was equipped with baffle arms, Suprasil I windows (Heraeus) for admission of the laser beams and for observing at right angle the UV-fluorescence of NO(A2y+) excited by the probe laser.A Nd-YAG pumped dye laser (JK 2000) loaned from the laser support facility of the SERC provided the red dissociation beam with fixed wavelengths of 568 nm (Rhoda-mine 6G in methanol) and 645 nm (DCM in methanol) at pulse energies of <25 mJ and <10 mJ, respectively.A planoconvex lens of 500 mm focal length focused the dissociation beam into the centre of the cell to a spot size of ca. 1.5 mm as judged from burn patterns.The probe laser, a Lambda-Physik EMG 50 excimer pumped dye laser (FL 2002) delivered between 0.5 to 1 mJ pulse energy for interrogation of nascent NO on the A 2y.+ --X21-Ii transition using the dyes coumarin 47 and 107 in methanol.A Photon Control model 25 power meter and a Molectron pyroelectric detector (J3-05) continuously monitored the average pulse energies of dissociation and probe beams.A temporal delay of 100 ns between the dissociation and probe pulses was used unless specified otherwise resulting in a combined jitter of + 20 ns.The pulse durations were 15 ns for the dissociation and 5 ns for the probe laser system.An in-house constructed multiple delay generator con- trolled the synchronisation of the lasers as described in the following section.

Photodissociation of jet-cooled CCIFzNO
The experimental set-up is shown in Figure 1.The expansion chamber consists of a small in-house constructed and anodised aluminium cube of 145 mm edge equipped with five 'O' ring sealed vacuum flanges and is mounted via a cylindrical A1 pipe on a 100 mm liquid nitrogen trapped oil diffusion pump (Edwards EO4K and NTC 4).A quarter swing valve (Edwards QSB 100) between the top of the trap and the neck of the chamber allows the apparatus to be isolated from the diffusion and the mechanical rotary backing pump (Edwards E2M18).
The chamber and its jet-cooling performance have been described in detail elsewhere. 25Briefly, two identical light-collection telescopes are mounted by flanges on opposite faces of the cube.Suprasil I windows (Heraeus) transmit the fluorescence light collected by condensor lenses, that are positioned inside the chamber.One telescope images fluorescence in the visible and near-IR collected by an f/0.8 BK7 crown glass aspheric condensor lens (Spindler and Hoyer, G6ttingen) onto the photocathode of a GaAs photomultiplier tube (RCA 4832).This serves to detect laser-excited fluorescence in the visible, while fluore- scence from laser-excited NO(AEye) is collected by Suprasil-I optics (Heraeus) consisting of a doublet of planoconvex spherical lenses and a field lens, and is viewed through the opposite telescope by a solar Figure 1 Artist's view of experimental set-up for dissociation-probe experiments of jet-cooled CCIF2NO, laser 1: dye laser beam for pulsed dissociation, laser 2: dye laser beam for pulsed delayed probing of NO by two-photon LIF on the ,-bands, P: prism, M: mirror, L1 and L2: lenses, F: glass blocking filter, PMT: photomultiplier tubes.
blind photomultiplier (Hamamatsu R166 UH).Filters (Schott RG 715 and UVR-250) are mounted in the telescopes, and baffles equipped with a series of conically shaped irises as well as Brewster-angled quartz windows serve to reduce scattered laser light.
A pulsed solenoid valve (NRC model BV-100) mounted in the centre of the top flange delivers gas pulses of 150/s duration.During routine operation of the valve at 10 Hz pulse repetition rate and at 500 Torr stagnation pressure the chamber vacuum remains near 2 x 10 -4 Torr.Under the same expansion conditions NO or NO2 in mixtures of less than 1% by volume in Ar are cooled to a rotational temperature of less than 10 K.The dissociation laser system (laser 1 in Figure 1) consisted of the same JK 2000 Nd-YAG pumped dye laser used for the room- temperature experiments and provided pulse energies <8 mJ in the 670-720 nm region with pyridine 1 in methanol or <12 mJ in the 600--680 nm region with DCM in methanol.The Lambda Physik EMG50/FL2002 coumarin dye laser (laser 2 in Figure 1) provided ca. 1 mJ for probing nascent NO on the two-photon ),-bands.The colinearly aligned and counter-propagating beams were focused by the lenses L1 (f 500 mm) and L2 (f 200 mm) such that their foci overlapped in the centre of the chamber and observation region ca 5 mm below the nozzle orifice.Suitable glass blocking filters prevent the dissociation beam from entering the probe dye laser.Mixtures of CCIF2NO (5-10% by volume) with Ar were prepared on a glass vacuum line equipped with greaseless Teflon taps and an MKS baratron capaci- tance gauge.The mixtures were stored at 296 K in 5 dm 3 blackened glass bulks connected to the pulsed valve.
Signal acquisition and synchronisation of lasers Our signal processing system consists of a 100 MHz 8 bit digital transient recorder (Transiac TR 2001) and crate controller housed in and powered by a Grenson CAMAC crate.Stray 308 nm light in the probe dye laser triggers signal acquisition under the control of an IBM-PC microcomputer using CAMAC commands to simulate gated detection and boxcar integration of transient fluorescence signals.A mains phase-locked and crystal-controlled oscillator built in-house acts as the master clock and provides TTL level pulses for the opening of the molecular beam valve, the firing of the Nd-YAG flash lamps, the triggering of the Q-switch and the firing of the excimer laser via an in-house constructed analogue multiple delay generator.The pulse repetition rate used was 10 Hz.Scanning of the dye laser grating may be controlled externally by computer or internally via a home-built stepper motor controller.Spectra can be corrected for variation of the dye laser output during extensive scans over a wide wavelength interval by the microcomputer.Such wide spectral scans are recorded using a gated integrator and boxcar averager (Stanford Research Systems) rather than the transient recorder.An A to D converter and the microcomputer in conjunction with a Hewlett-Packard digital plotter (HP7470) are employed for displaying spectra.
Electronic absorption spectra of CCIF2NO The visible absorption spectrum of gaseous CC1F2NO was recorded at a resolution of 2 , with a dual-beam microprocessor controlled spectrophotometer (Shimadzu UV-240) using a 10 cm fused silica cylindrical cell.Molar extinction co-efficients were calculated using pressure readings from an MKS baratron gauge.

Spectroscopy of CCIF2NO
The electronic absorption spectrum of 9.95 Torr of CC1F2NO vapour at 295 K is shown in Figure 2. We calculate a maximum molar decadic absorption coefficient of 22 dmamol cm-1 corresponding to a cross- section of 8.4 x 10 -20 cm-2.From the integrated absorption we deduce an oscillator strengthf 2.0 +_ 0.1 x 10 -4 indicating that the electronic transition responsible for the blue colour of this gas is of the same type (n,rr*) as in CFaNO and CH3NO. 19 '26'27 The structure clearly discerni- ble on the long wavelength side of the spectrum is vibrational and mirrors transitions between torsional levels in the ground and elec- O-500 600 700 Xlnm Figure 2 Electronic absorption spectrum of CCIF2NO vapour at 295 K.The two wavelengths chosen for the room temperature photodissociation studies (645 and 568 nm) and the region of the LIF excitation spectrum of the jet-cooled molecule (Figure 3) tronically excited state.Due to the bent CNO grouping two distinct conformers with the O atom eclipsing either an F or a CI atom are expected to be present at 300 K. 22 They execute large amplitude torsional vibrations within potential wells separated by barriers in the range 200-1000 cm-.'22'28 Near 300 K a large fraction of the thermal population occupies "hot" torsional levels.In conjunction with the eclipsed-staggered conformational change on electronic excitation that is characteristic for this class of molecules, 9'26-29 progressions in the torsional mode are observed, that lead to marked vibrational congestion at 300 K.This "hot" band congestion is more pronounced than the sequence congestion com- monly encountered in aromatic systems since the Franck-Condon factors for out-of-phase or minimum to maximum electronic transi- tions v', v" increase rapidly with increasing torsional quantum number v". 27 On cooling to 200 K the spectrum sharpens up considerably indicating clearly that the apparent broadening or diffuseness is due to thermal congestion.
Figure 3 shows the LIF excitation spectrum of CC1F2NO jet-cooled with Ar.][10][11] The fluorescence excitation spectrum of jet-cooled CC1F2NO has been recorded and discussed previously by Ernsting. 28 We have re-examined it and have studied the fluorescence decay of a range of single vibronic levels as part of a systematic investigation of the spectroscopy and photophysics of these compounds. 29While we agree with the previous assignment of the electronic origin and of the major features of the spectrum, we have deduced substantially differ- ent potential curves for the ground and excited state.The conform- ational analysis of the A and ."states of CCIF2NO is beyond the scope of the present work and will be reported in detail elsewhere. 29Briefly, we conclude, that the excitation spectrum of the jet-cooled molecule is carried almost entirely by a single energetically favoured but two-fold degenerate "gauche" conformer with the O atom eclipsing either of two equivalent F atoms in the ground state.The 300 K spectrum is also dominated by the gauche ground state rotamer but the cis-rotamer contributes roughly 15 % of the integrated absorption.We have found no evidence for dimerisation or polymerisation of the parent in the jet under our expansion conditions as mentioned previously. 2 Emission can only be excited in the small portion of the absorption spectrum close to the electronic origin of the transition as indicated in Figure 2, and rapid non-radiative processes clearly dominate the photophysics of the molecule at wavelengths shorter than 650 nm.
Table I summarises the assignment of the major features in the fluorescence excitation spectrum of CC1F2NO.The detailed vib- rational analysis will be given elsewhere. 29Normal mode 12 is the lowest a" (in C symmetry) vibration and corresponds to the torsion.Modes 8 and 7 are the lowest frequency in-plane (a' in C symmetry) skeletal bending/rocking modes.
Photodissociation at 300 K The two wavelengths 645 and 568 nm chosen for the photodissociation experiments with the 300 K vapour are indicated in the absorption spectrum of Figure 2 by bold arrows.The spectral congestion clearly precludes any state selection.At both wavelengths nascent NO(X2II) was identified readily by the characteristic ),(O,O) two-photon LIF band near 450 nm.By changing the delay between pump (dissociation) and probe pulses we established that the formation of NO at both dissociation wavelengths occurs within the dissociation pulse (5 ns).
An initial sharp rise due to the appearance of the fragment is followed by a slow decrease associated with diffusion out of the observation region and reaction.At these wavelengths the rates of appearance of NO can therefore not be measured with our equipment.By scanning the probe-laser through the A2 / (v' 0) X2IIi(v 0) two-photon transition of NO with a fixed delay of 50 ns we recorded rotationally resolved two-photon excitation spectra of the nascent fragment.Col- lisional relaxation effects were shown to be negligible by examining the effect of increasing temporal delay between pump and probe pulses on the relative intensity of the well isolated O2-branch near 454.5 nm.We were unable to detect NO(v" 1) in the 645 nm photolysis.
Relative spectral intensities in the two-photon LIF excitation spec- tra were converted to rotationally resolved population distributions by means of the relationship /rel C q'v" Sy,y,, i, nflg Figure 4 Rotational distributions of NO(v" 0) from the photolysis of 300 K CCIF2NO at 645 nm.The solid lines show calculated statistical distributions except for the spin-orbit ratio, which has been scaled to permit better comparison with experiment.where C is a constant, q,,, stands for the Franck-Condon factor of the particular ),-band of NO, S,,, for the two-photon H6nl-London factor and i, for the wavenumber of the transition, g 2J + 1 denotes the rotational degeneracy of the ground-state levels.Details of this method and of the two-photon LIF spectrum of NO have been given previously.6'7'25 Since the predissociation studied here is slow on the timescale of the parent rotation, the photofragments cannot be alig- ned, and the spatial distribution of NO is isotropic.Observed line intensities were corrected for variations in the dissociation and probe laser power as well as for variations in the pressure of the parent molecules during the extensive spectral scans, that were necessary to record complete bands.
Figure 4 shows the NO(v" 0) population distribution obtained in this way from the photolysis of CC1F2NO at 645 nm and 300 K.While NO(v" 1) is absent at 645 nm, a significant fraction of NO(X) formed in the 568 nm photolysis appears as NO(v" 1).In both cases, the product-state distributions are dominated by low-J populations, and the spin-orbit population distributions are cold and not equilibrated.
Photolysis at 645 nm is well behaved and shows a linear dependence of the NO signal on dissociation pulse energy.Log-log plots of NO signal measured at the band head of the PI -b O21 two-photon branch versus pulse energy are linear and provide a slope of 1.0 + 0.1 at 645 nm and slope of 2.0 + 0.1 at 568 nm as shown in Figure 5. Clearly, two or possibly more photons must be involved in the absorption step prior to dissociation at 568 nm.
To simulate statistical product-state (Prior) distributions we used a procedure outlined in a previous publication 9 and adapted from Bogan and Setser. 3The probability for formation of NO in a given quantum state is calculated using energy conservation but neglecting angular momentum conservation, and by evaluating the density of product states for different available excess energies E+.The vib- rational frequencies of the CC1F2 radical are available from the matrix isolation studies of Milligan et al. 31 and Prochaska and Andrews. 32The average parent internal energy was estimated using the known vib- rational frequencies 22 and classical statistical mechanics.
Calculations were performed with the excess available energy E / as the only variable until the average rotational energy coincided with the experimentally observed value.Errors introduced by neglecting the strict angular momentum conservation of more sophisticated B45 nm 5BB nm Relotive Dissociotion Loser Power Figure 5 Log-log plot of the NO two-photon LIF signal excited on the band head of the Pll + O21 branch near 452.3 nm versus 645 nm photolysis pulse energy.approaches like phase space (PST) or statistical adiabatic channel theory (SACT) are likely to be minor in comparison to the errors associated with the uncertainty of the vibrational properties and the structure of the CC1F2 radical.
Experimental and calculated average rotational energies were evaluated by summation (Erot) njErot/ pj J,f J, PREDISSOCIATION DYNAMICS OF CCIF2NO .(n,a'*) Table II Photodissocation of 300K CCIF2NO: Observed and (in brackets) calculated average rotational ((Erot)) and vibrational (Evib) energies and spin-orbit ratios of the nascent NO photoproduct formed by pulsed photolysis at the wavelengths/diss E+ are the excess energies above the dissociation threshold used in the statistical calculations (see text). 391(392) 580 (585) (Evib)/cm -1 <19 ( 4) 24 (57) F/F2(v" 0) 2.01 (1.35) 1. 30 (1.24) E + 2600 4300 using the constants and term values for the intermediate coupling case to evaluate the rotational level energies Erot.The spin-orbit ratios given are similarly obtained by summation F1/F2 X nfl X nj fl--1/2 Table II summarises observed and calculated E, V and R energy disposal into nascent NO in the photodissociation of CCIF2NO at the two wavelengths.
Photodissociation in the free jet The use of two separate electronically synchronised dissociation and probe laser systems permits one to carry out three distinct types of measurement with predissociating jet-cooled nitroso compounds.By scanning the dissociation laser through the structured absorption features of the cold parent while monitoring the undispersed fluore- scence of the NO photoproduct excited at a fixed resonant wavelength of the probe laser, a photofragment yield spectrum can be obtained.
Appearance rates of NO can be measured by varying the delay between dissociation and probe pulses, and finally, excitation spectra of the NO(X) photoproduct and hence product-state distributions can be recorded by scanning the probe laser, while keeping the dissoci- ation wavelength fixed on the parent transition selected using a fixed electronic delay.Figure 6 presents the photofragment yield spectrum of jet-cooled CCIF2NO obtained by monitoring NO near 452.3 nm on the P + O2 obtained with a delay of 100 ns between the dissociation and probe pulses, and consequently features at wavelengths longer than 680 nm may be progressively attenuated and may appear weaker as the NO appearance times become longer with decreasing excitation energy.
Any parent lost by flying out of the focal region of the probe beam before dissociation will not contribute to the NO signal.A detailed comparison of the LIF excitation spectrum (Figure 3) and the photo- fragment yield spectrum leads us to conclude that there are no drastic differences that might indicate state specific effects.
In spite of the complications associated with fly-out effects and the use of a focused probe beam, we selected a number of vibronic features for estimating crude NO appearance rates from plots of the NO yield versus the delay time.The results of these measurements are collected in Table III.They are only meant to indicate general trends,,and no firm conclusions can be drawn from this data.Under the free-jet conditions used parent molecules will cross the 0.1 mm focal region of the probe laser beam in roughly 100 ns.Thus, molecules dissociating on a time scale longer than this, move out of the observation region, and NO subsequently produced does not contribute to the signal.Houston and, subsequently, Reisler, Wittig and their respective colla- borators were able to overcome this experimental problem by using single-photon LIF and wide unfocused probe beams 18'2 enabling them to measure NO appearance times well in excess of 200 ns.In our case Table III Kinetics of the appearance of nascent NO following the state-selected photodissociation of jet-cooled CCIF2NO at the wavelengths/dis/nm (transition wave- number ). :ise -rise time of the NO  the values collected in Table III give a reasonable estimate of the fast NO appearance rates (rd < 100 ns) but any slow component of a bi-exponential rise curve would be missed under our conditions.
Risetimes longer than 100 ns quoted in Table III should be considered as lower limits only.Nevertheless, we may conclude, that the rate of appearance of NO increases monotonically with excitation energy, and for most vibronic transitions studied, the mean appearance times exceed the duration of the dissociation pulses.
Having obtained rough NO appearance times, energy disposal into rotation, the product-state distributions and electronic branching ratios of nascent NO(v" 0) were measured.Figure 7 presents a section of a typical two-photon LIF spectrum of nascent NO(v" 0) from the photolysis of CC1FENO on the strong feature at 646.6 nm.
Low J transitions clearly dominate the spectra.In the example shown in Figure 7, band heads associated with the Oll and Pll + O21 branches dominate the spectrum, whereas the O12 and 022 -I-PiE band heads at longer wavelengths are much weaker.This indicates qualitatively, that NO is formed preferentially in the F1 (2ri1/2) state.We recorded complete y(O,O) two-photon LIF spectra of nascent NO(v" 0) from ,/nm Figure 7 Section of the A2X (v' 0) <--X2Ili (v" 0) two-photon LIF excitation spectrum of nascent NO from the photolysis of the jet-cooled parent molecule on the unassigned 646.6 nm vibronic transition.the photodissociation of CC1F2NO from 14 single vibronic levels of the A(n, rr*) state and evaluated average rotational energies (Erot) and spin-orbit ratios F/F2 from the population distributions deduced as described above.Table IV summarises the results of this considerable experimental effort and also includes calculated statistical F/F2 ratios and the excess energies E + required to fit statistical (Prior) distribu- tions to the observed rotational product-state distributions.Compari- sons of typical observed and simulated distributions are shown in Figures 8 and 9. Note, that the observed F/F2 distributions are cold showing a clear non-statistical preference for the lower F1 (21-11/2) component.The calculated statistical distributions in Figures 8 and 9 have been scaled to the experimental n versus J distributions to permit better comparison.The Prior F1/F2 ratios are much closer to unity as evident from Table IV.These results clearly show that more energy appears in NO rotation as excess energy increases, but that the spin-orbit ratio remains non-statistical and very cold showing no sign NO(v" 0) photoproduct from the dissociation of jet-cooled CCIFENO on the 663.5 nm vibronic feature.For additional explanations see legend of Figure 4.
photolysis wavelength E + may be identified by locating the highest rotational state of the nascent NO fragment.Due to the poor sensiti- vity of the two-photon LIF probe technique these thresholds were buried in noise and could not be located with confidence.The less accurate alternative approach demonstrated first by Bower et al. 18 in the case of CFaNO had to be used.Here the excess energies E + available for a statistical disposal into the photofragments are esti- mated by fitting Prior distributions to rotational product-state distribu- tions observed at several photolysis wavelengths.A D0(CN) value of 13 700 + 350 cm -1 or 164 kJ mo1-1 was obtained from the intercept of E + versus the excitation energy.Figure 9 Observed and calculated statistical rotational distributions of nascent NO(v" 0) from the photolysis of the jet-cooled parent at 641.8 nm.For additional explanations see legend of Figure 4.

DISCUSSION
Our results of the photolysis of CC1F2NO at ambient temperature indicate a substantial dependence of the dissociation mechanism on the wavelength.At 645 nm the expected single-photon process domin- ates as judged from the dependence of the NO yield on pulse energy.The statistical model reproduces energy disposal into the rotation and vibration of NO reasonably as shown in Table II, and the rotational population distribution of NO agrees well with a statistical distribution calculated with an excess energy E + of 2600 cm-1.The 568 nm Re I at Ive O eeocl at i on La6er Power Figure 10 Log-log plot of the NO two-photon LIF signal excited near 452.3 nm versus dissociation pulse energy at 694.8 nm (12 vibronic transition).
photodissociation, in contrast, leads to a rotational population dis- tribution, that can only be simulated poorly, and much less energy appears in the vibration of NO than predicted by the model.This may imply, that energy disposal at this wavelength may have some dyna- mical basis.The spin-orbit distributions are too cold at either wave- length to mirror a statistical disposal of excess energy into the electronic degree of freedom of the NO product.The higher order power dependence observed at 568 nm indicates that dissociation via S1 and via a higher state reached by sequential absorption of another photon leads to competition between quite unrelated dissociation channels.
The comparison of our results for the photodissociation of CC1F2NO at 300 K and in the jet clearly shows the advantages of jet-cooling for achieving state-selective excitation and for studying the unimolecular decomposition of monoenergetic excited molecules.The data on energy disposal in the photodissociation of CCIFENO following excita- tion of well defined single vibronic levels of the (n, rr*) state (Table IV) show that the average energy appearing as rotation of NO increases with increasing available energy.There are no abrupt changes that might indicate state-specific effects, as the vibrational character of the initially excited level changes, and the observed product rotational distributions conform very well to statistical dis- tributions as evident by the comparison between calculated and observed distributions in Figures 8 and 9.Even at the highest excess energy, ca 1900 cm -1 above the dissociation threshold, we cannot see any clear indication of non-statistical or dynamical effects in the NO rotational distribution.The spin-orbit distributions, however, are far too cold to reflect a statistical partitioning of excess energy and remain cold even at the highest excess energies involved.Similar product-state distributions for nascent NO are found in the case of CFaNO, where only slight deviations from statistical nj-distributions appear at high excess energy, andwhere the F1/F2 (I'I1/2/I'I3/2) ratios remain high.ts't9 In the photodissociation of (CHa)3CNO, on the other hand, Noble et al.
were able to distinguish two excitation regions leading to distinctly different NO product-state distributions. 2 Here, dissociation near threshold apparently proceeds on So on a time scale >3.5 #s and yields NO with a statistical product-state distribution, while fast predissoci- ation via the triplet state T1 predominates from $1 levels located more than 650 cm -1 above threshold.This rapid predissociation by inter- system crossing (ISC) gives rise to NO with decidedly cold and non-statistical rotational and spin-orbit distributions.Wittig and collaborators have summarised possible reasons for cold-spin-orbit populations of the nascent NO from the predissoci- ation of 2-methyl-2-nitrosopropane. 2 In our case the statistical nature of the NO rotational distributions indicates that there is no significant barrier in the exit channel leading to the fragments.The colder than statistical spin-orbit ratios, on the other hand, may be the consequence of the absence of strong coupling between the II1/2 and FI3/2 surfaces correlating parent and separated fragments.Alternatively, the branching is controlled by the dynamics of the nonadiabatic transfer associated with the crossing of potential energy surfaces.
The near-threshold dissociation of C-nitroso alkanes offers an inter- esting opportunity to study the unimolecular decomposition of long- lived, weakly bound complexes.By combining state-selective excita- tion with detailed mapping of product-state distributions, dissociation rate and fluorescence decay measurements, it is possible to obtain a fairly complete picture of the predissociation dynamics.Unfortunately our knowledge of the potential energy surfaces connecting the three relevant bound states of bent RuN=O f( So, A(n,r*) $1 and (n, r*) T1 with the fragments is inadequate and rests entirely on the analogy with HNO, where the surfaces and correlations are known relatively well. 33There are essentially three mechanisms that need to be considered in principle in the predissociation of RNO A(n,r*) where R is an alkyl group: 1) predissociation by internal conversion (IC) and unimolecular decomposition on the ground state surface (So).
3) vibrational predissociation on an A(n, zr*) S surface, that is quasi-bound or has only a shallow minimum of a few hundred cm -1 along the dissociation coordinate.Mechanisms 1 and 2 have been invoked previously, 8'2'18-2 but the third possibility has not hitherto been considered.
Houston and collaborators concluded from the statistical behaviour of the NO product rotation and the monotonic increase of the NO appearance and fluorescence day rate of A state CF3NO with excita- tion energy, that mechanism 1 predominates, but did not exclude a contribution from mechanism 2 at higher energy.TM We have recently reported evidence for significant involvement of ISC (mechanism 2) and the importance of T in the predissociation of CF3NO.S levels higher than 700 cm-1 above threshold appear to decay mainly by ISC.In the large molecule (CH3)3CNO Noble et al. identified both mechan- ism 1 and 2 with the latter providing the dominant route for E + > 650 cm-1.The predissociation of HNO A 1A" from levels within 700 cm -1 of the dissociation threshold by mechanisms 1 and 2 is understood in considerable detail based on the work of Dixon et al. 16,17,33 Mechanism 3 remains unimportant in HNO since a sizeable barrier separates the, state minimum from the dissociation asymptote.The assumption that a similar barrier is present in the (n,r*) states of C-nitroso com- pounds is probably incorrect.There is evidence that the ,4 states of C-nitrosoalkanes exhibit only very shallow potential minima along the minimum energy path leading to the separate fragments. 34We con- clude therefore, that mechanism 3 also needs to be considered for the near-threshold dissociation of CC1F2NO.
We have observed pronounced biexponential fluorescence decay behaviour for all ,4(n, rr*) levels of CC1FENO situated higher than 200 cm -1 above the v' 0 vibrational level 29 indicating strong mixing of these levels with T1.ISC is therefore an important and for the higher levels probably the dominating pathway for the fi state predissoci- ation, which must show the same temporal characteristics as the fluorescence decay.We anticipate therefore that the NO appearance curves will reveal a fast and slow component on careful measurement with a better technique than was available to us.In CFaNO/] state levels higher than 700 cm -1 predissociate predominantly by ISC.In CC1F2NO dissociation by mechanism 2, i.e. via ISC under the influ- ence of spin-orbit coupling and in the absence of a significant barrier produces statistical rotational NO distributions.Levels close to the O,O electronic origin may decay either by IC (mechanism 1) or by vibrational predissociation (mechanism 3) on an $1 surface with a very shallow minimum.The observed product-state distributions do not permit us to discriminate between these mechanisms.

CONCLUSION
The predissociation dynamics of CC1F2NO (n, rr*) strongly resembles that of the parent molecule CFaNO, and the dynamics of all but the lowest/] state levels is dominated by ISC.Levels close to the O,O origin of thesystem predissociate either by internal conversion or by direct vibrational predissociation on the weakly bound $1 surface.Fluorescence decay measurements provide more information about the involvement of T1 than the rotational and spin-orbit populations of the nascent NO photoproduct.Little evidence for state-specific beha- viour is revealed by the monotonic dependence of the rotational product-state distribution and energy disposal into rotation on the quantum state excited in the A(n,r*) state of the parent.
C-nitrosoalkanes offer the possibility of studying in detail the disso- ciation dynamics of weakly bound complexes.Fluorescence excitation and laser photofragment yield spectra provide information on the level structure and some information on the potential surface of the parent, while detailed dynamical information can be gleaned from time- resolved fluorescence and appearance rate measurement and NO product-state distributions.

Figure 6
Figure6NO(X, v" 0) photofragment yield spectrum of jet-cooled CCIF2NO.The probe laser excites the Pll + O21 band head of the ,(O,O) A2y --21-Ii two-photon transition near 452.3 nm while the dissociation laser is tuned through the absorption bands of the jet-cooled parent.

Figure $
Figure $ Observed and calculated statistical rotational distributions of the nascent

Table I
Uncorrected wavelengths, transition wavenumbers 9, intervals and assign- ments of the n, rr* LIF excitation spectrum of jet-cooled CCIF2NO in the visible