Magnetic-field and Time-resolved Studies of the Electronic Spectrum of HNO BY RICHARD N. DIXON, MARCUS NOBLE AND CAROLINE A. TAYLOR School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 ITS AND MICHEL DELHOUME Paris Observatory, 92190 Meudon, France Receired 29th December, 1980 High-resolution laser-induced fluorescence excitation spectra of HNO have been recorded using both C.W. and pulsed dye lasers. These spectra reveal numerous frequency perturbations, often asso- ciated with marked intensity anomalies. All the branches studied were found to be perturbed to some degree with little regularity to the perturbations. The widths of some of the perturbed lines, and of many apparently unperturbed lines, have been found to be sensitive to a magnetic field. Time- resolved measurements yield a mean zero-pressure excited-state lifetime of 23 ps, and collisional quenching rates varying from (1.1 to 15) x These studies reveal that the intensity anomalies arise through selective higher quenching of perturbed levels, coupled with fast rotational-energy transfer.These aspects are discussed in terms of interactions between the levels of the excited 2, ’A” state and high levels of the ground 2, ‘A’ state, with further perturbations from the d , 3A” state. cm3 molecule-’ s-’. The 2, lA’’-T, ‘A’ band system of HNO in the red and near-infrared regions of the spectrum was first observed by Dalby in absorption using the technique of flash photolysis.’ The bands have the simple appearance of type-c bands of a near prolate asymmetric top, although in his analysis Dalby noted a number of weak perturbations of the rotational structure.Bancroft et a/.’ extended this analysis to a total of 10 bands of HNO and 9 bands of DNO, and noted further perturbations. This band system is also known in eniission from chemiluininescent The emission spectrum exhibits breaking-off due to predissociation to H + NO. In a recent study using laser-induced fluorescence of HNO we have shown’ that at threshold this pre- dissociation occurs through Coriolis coupling of the levels of the A, lA” state to high levels of the ground f, ‘A’ state. Analytical potentials for the 2, 2 and a”, 3A’’ states were derived from spectroscopic, thermochemical and quantum-theoretic data as part of that study. It is now accepted that interstate perturbations play a critical role in the time evolution of the excited states of polyatoinic molecules through the processes of inter- nal conversion, intersystem crossing and predissociation.Much has been learned about these decay routes in small molecules through very many studies of excited- state lifetimes8 and fluorescence quantum yields’ of H,CO (and D,CO), which is the lightest stable polyatomic molecule with a well-characterised sharp spectrum. Even so these processes are not yet capable of quantitative interpretation. HNO is iso- electronic with H,CO, but possesses fewer vibrational degrees of freedom, and there- fore has a much lower density of vibrational states for a given energy. Its dynamic126 ELECTRONIC SPECTRUM OF HNO behaviour should therefore be intermediate between that of formaldehyde and that typical of diatomic molecules. This paper presents improved higher-resolution analyses of those bands of HNO between 641 and 578 nm that can be excited to laser-induced fluorescence. Selected portions of this spectrum have been studied in magnetic fields up t o 10.5 kG.Time- resolved measurements over a range of pressures have then been made for a number of normal levels, perturbed levels and magnetically active perturbed levels. These new observations are discussed in terms of the properties of the perturbing states. EXPERIMENTAL In view of the known quenching behaviour of the x, 'A" state of HN05g6 it is important that a study of excited-state dynamics should include measurements in a collison-free regime if at all possible.Kirby et a/.'' have pioneered the generation of HNO by pyrolysis of its adduct with 9,1O-dimethylanthracene, and have used this in a study of its microwave spec- trum." We have used this method to produce HNO at pressures of 0.5 to ca. 200 mTorr, the products of the 70-80 "C pyrolysis being slowly pumped through the fluorescence cells. A parallel study using photoelectron spectroscopy showed that the gas mixture prepared in this way contained considerable fractions of the disproportionation products N20 and H20. Pressure measurements were made using a capacitance manometer, or at the lowest pressures a Penning gauge calibrated at higher pressure against the capacitance manometer. A Coherent CR 490 C.W. dye laser pumped by an argon-ion laser was used to survey the whole spectrum over the ranges of the dyes Rhodamine 6G and Sulforhodamine B at a resolution of 0.5 cm-'.Selected regions were then recorded at a resolution of 0.1 cm-I by inserting a 0.5 mm fused quartz etalon in the cavity and driving this under microprocessor control in synchronisation with the coarser Lyot birefringent filter. Calibration was achieved using the fluorescence of I2 as a secondary standard,'* and the transmission fringes of a 4 mm fused quartz Fabry-Perot etalon (mean spacing 0.847 cm-') as a means of interpolation between iodine lines. The absolute accuracy of measurement of unblended lines is considered to be $0.01 cm-'. A search was made using this laser for magnetic-field activity in fields up to 10.5 kG, attention being concentrated on those regions where perturbations were apparent.A search was also made for electric-field effects using fields up to 3.5 kV cm- ' : no significant electric-field effects on line profiles or intensities were observed at this resolution, although we have been able to measure the excited-state dipole moment of HNO by? DoppleL-free te~hnique.'~ (The dipole moment changes very little on excitation from the X to the A state.) Lines showing magnetic activity were then examined at the Doppler-limited resolution of ca, 1 GHz (0.035 cm-') using a single-mode Coherent CR 599-21 C.W. dye laser pumped by a krypton-ion laser. The dispersions of the scans with this laser were calibrated by simul- taneously recording the transmission fringes of a 16 cm confocal interferometer, with fringe separation 468 MHz.A particular study has been made of the K' = 4 - K" = 3 sub-band of the 100-000 band, which is the last sub-band with extended J-structure below the predissocia- tion threshold. The time-resolved studies were carried out using a home-built dye laser of Hansch design,I4 with oscillator and amplifier stages, pumped by an Oxford Lasers excimer laser operating on XeCl. With a 5 mm air-gap etalon and pressure tuning this exhibited a line width of ca. 2 GHz (0.07 cm-') and very little background superadiance. This width is only about twice the Doppler width, so that it was possible to make separate studies of the time-evolution of the fluorescence with excitation in various closely resolved components of perturbed lines.The time resolution of these measurements was limited to ca. 0.5 ,us by a combination of the photomultiplier and pre-amplifier responses and by the transient recorder (Datalab DL 920). Signal-averaging was accomplished using a PET microcomputer which also controlled the operation of the laser system. Three tunable laser systems have been used in this work.R . N. DIXON, M. NOBLE, c. A. TAYLOR AND M. DELHOUME 127 RESULTS ASSIGNMENTS Fig. 1 compares a low-resolution LIF excitation spectrum of the 4-3 sub-band of the 100-000 band of HNO with a microdensitometer tracing of the corresponding band photographed in absorption by Bancroft, Hollas and Ramsay (2, hereafter denoted by B.H.R.) Two features are apparent from this figure: (i) the excited-state structure breaks off above J' = 11, and (ii) there is a very clear intensity perturbation in the la) 16 140 16 190 v/cm-' FIG.1.-The 100-000 K = 4-3 sub-band of the 2-2 band system of HNO. (a) Microdensitometer trace of a photograph of the absorption spectrum (after Baxroft et Q L ) . ~ (b) Laser-induced fluores- cence excitation spectrum. The weak RQ4 and RR4 branches in (a) are of the 020-000 K = 5-4 sub- band. excitation spectrum for J' = 7 and 8 that is not apparent in the absorption spectrum. Higher-resolution excitation spectra show that these lines have multiple components arising from rotational perturbations [fig. 2(c) and (41. Even so, the sums of the intensities of these close components are less than expected by interpolation between high and low J values.The extent of this loss of intensity in the excitation spectrum, and the range of J values affected, both increase with increasing pressure. A similar behaviour was noted in a number of other sub-bands. We will return to this feature of the excitation spectra in the discussion of the time-resolved measurements. The spectra recorded included, in order of increasing frequency, the vibronic tran- sitions 011-000, 020-000, 100-000, 101-000 and 030-000. The first three show ex- tended K-structure which could be assigned in part using the tabulations of B.H.R. However, the higher resolving power and sensitivity showed up numerous perturba-128 1 ELECTRONIC SPECTRUM OF HNO 1 I I f I 16 100 5 10 15 v1crn-l n _. I I I I I 16120 25 30 35 LO vlcm-' I I I____ I 1 I 16 150 55 60 16 170 75 83 v/cm-l v/cm-' FIG.2.-Assignments of branches in the HNO fluorescence excitation spectrum close to the breaking- off limit at 16 182 cm-'.R . N . DIXON, M . NOBLE, c. A . TAYLOR AND M. DELHOUME 129 tions of up to 2 cm-l, and extra lines, such that it was necessary to re-analyse much of the spectrum. This analysis was accomplished using ground-state combination dif- ferences, computed using the most recent molecular constants of Johns and McKel- lar15 in assignments of the "R, "Q and "P branches with K' 3 2. The P-type sub- bands and low K' transitions are heavily overlapped, and for these it was not possible to make new unambiguous assignments. A section of the assigned spectrum is pre- sented in fig. 2. The assignments are summarised in the Appendix as a tabulation of upper-state term values, most of which are derived from three branches.Extra lines have only been assigned where they are consistently present in at least two branches. There remains a large number of weak unassigned lines throughout much of the spectrum. These are particularly evident near the breaking off limit where there are fewer main lines. Here they constitute about two-thirds of the total number of lines, but are typically only 1-5x of the intensity of the strong lines. The K-type doubling of the ground-state levels is completely negligible for K" =3 at Doppler-limited resolution. For K" = 2 it reaches the Doppler width by J" = 6, and is very large for K" = 1. Thus for levels with K' == 4 it has not been possible to deduce the upper-state parity even where there is an observed K doubling.However, for K' = 2 and 3 the ground-state combination differences are sufficiently different for pairs of asymmetric top components that the upper state parity could be assigned tak- ing into account the type-c selection rules. Many excited-state levels show asym- metry splittings which are much greater than those predicted from the rotational con- stants and not necessarily of the correct sign: some of these are visible in fig. 2 in the "R2 and "Q2 branches of the 100-000 band. Takagi et al.'6917 have recently observed similar unexpected asymmetry splittings in other sub-bands of the A",'A'' state in micro- wave optical double-resonance spectra. The frequencies of single sharp lines agree with those tabulated by B.H.R.to within h0.05 cm-l, but very many lines were found to be multiple at the higher resolution of our spectra. Upper-state term values of levels not greatly perturbed were found to lie generally within ca. 0.25 cm-I of those calculated with their molecular para- meters. In fig. 3 we present the energy shifts between the term values and those given using the symmetric-top approximation with the B.H.R. constants: AT = T(v, J , K ) - [To + G(u) + ( A , - BL,)K2 + BJ(J + 1) - D,,,K4 - DJK,"K2J(J + 01. (1) Some of the perturbations involve the systematic displacement of a considerable number of consecutive energy levels, some are very localised affecting only one or two levels and yet others consist of multiple perturbations of one J value.The strongest and most easily characterised perturbation occurs at low J in the K r = 4 manifold of the 011 vibronic state [fig. 3(b)]. Two lines are observed for J' = 5,6,7 and 8, with a minimum separation of 2 cm-', but there is no satellite line for J' = 4. The intensity distribution among these pairs, together with the pattern of perturbed energies, clearly indicates that the coupling matrix element increases with increasing J. For a singlet-singlet Coriolis interaction the angular momentum de- pendence of this element is proportional to K for AK = 0, and to [J(J + 1) - K(K & 1)]+ for AK = -& 1. The intensity mixing is closest to that for AK = - 1, and the best fit to the data with this assumption leads to : 01 1, K = 4 ATc,K = $3.26 0.25 cm-I ARff.= -0.074 0.005 cm-': Jeff. = 1.188 cm-I (2) = (0.197 & O.O07)[J(J 4- 1) - 1213 cm-I.130 ELECTRONIC SPECTRUM OF HNO / ox I @I I 0 0 0 0 I ATlcm - m lx E1 o ] 4 -- k 6 - 1" x o 0 I -- m I 0 X @I I I 1, i / I 0 L n 0 1 0 0 3 I L " 0 m 0 0 I 0 0 m 0 0 N n M + 2 's 0 0 0 0 s? 0 0 c-4 n M + 'r: 's 0 0 c 0 IG. 3.-Deviations between the term values for vibronic states of HNO A", 'A" and the model calcu- (b) Parity is not known for K = lations of eqn (1). (a) For K = 2 or 3 : 0, e levels; x , flevels. 4. (--- ) Least-squares fits to the perturbations analysed in eqn (2)-(5).R . N. DIXON, M . NOBLE, C . A . TAYLOR AND M . DELHOUME 131 There is a very similar perturbation of the f levels of the K = 3 manifold. again best fitted as a AK = - 1 perturbation, leading to : 011, K = 3(flevels) This is AT,,K = +4.01 -+_ 0.36 cm-I ABeffe = -0.031 & 0.003 cm-': Befp. = 1.235 cm-I (3) The e levels of this manifold are also perturbed, but less strongly and less systematic- ally, and there is a second region of perturbation for J > 14 which precludes unambi- guous assignments since only the "Q branch is not overlapped.These two values of Jeff* for the perturbing levels are substantially lower than the value of 1.263 cm-' for the upper-state levels, and are significantly different from each other. The overall nature of these two regions of perturbation is inconsistent with interactions with a single upper vibronic level of the g, 'A' state. Alternatively, the perturbing levels could belong to the a", 3A" state.In a molecule of the point group C, the direct spin-orbit interaction between a singlet and a triplet state of the same configuration is not forbidden by symmetry, unlike the case of H,CO of point group Czv. In addition the indirect spin-orbit-orbital-rotation interaction active in H2C0 18*19 may also couple the 2 and a" states, so that the allowed perturbations are given by AK = 0, & l , &2 and AN = 0, & l . Several of these matrix elements have a J-dependence similar to that used in eqn (2) and (3). Thus it is not possible with our present assignments to distinguish between a singlet-singlet and a singlet-triplet interaction. Whereas the perturbations discussed above involve crossings in fig. 3 by lines of negative slope (lower B if AN = 0) two further systematic perturbations in K' = 4 of 020 and K' = 4 of 100 involve crossings by lines of positive slope.Since these are more localised the J-dependence of the interaction matrix element cannot be deduced, and has been assumed constant. In these cases we have again assumed a J(J + 1) energy dependence appropriate to a singlet-singlet interaction, giving : 020, K = 4 H1,2 = (0.068 & O.O04)[J(J + 1) - 6]* cm-'. AT,,K = -2.1 -+ 0.9 cm-' Hl,2 = 0.42 & 0.04 cm-l. ABeff. = +0.055 & 0.020 cm-': Befra = 1.32 cm-' (4) and : 100, K = 4 ATy,& = -1.3 & 0.2 cm-' ABeff. = +0.018 & 0.004 cm-': Jeff. = 1.295 cm-' (5) Hl,2 = 0.088 & 0.016 cm-'. The remaining perturbations are too localised or erratic to permit any deduction con- cerning the properties of the perturbing levels.We note that both Dalby ' and B.H.R.2 encountered the same difficulty with respect to other HNO perturbations. MAGNETIC BEHAVIOUR The majority of the HNO lines show no significant activity at 0.1 cm-' resolution and 10.5 kG, as was to be expected for a non-degenerate singlet-singlet transition. This is also the case for many of the perturbed lines, but a few of these show marked line-broadening or changes in intensity. This underlines that the perturbations are not all of one type. Upon further investigation using the single-mode laser it was found that very many lines showed a small increase in linewidth from the Doppler limit of 1.1 GHz (f.w.h.m.) to 1.5-2.0 GHz, equivalent to a magnetic moment in the field direction of up to ca.0 . 0 3 ~ ~ . The perturbed lines in the K' = 4 manifolds of132 ELECTRONIC SPECTRUM OF HNO the 01 1 and 020 states, analysed above in eqn (4) and (9, are of this type. Since these show no more magnetic activity than many apparently unperturbed lines this supports the postulate that these perturbations arise from singlet-singlet interactions. A few of the numerous weak lines underlying the main bands show significant shifts, or much greater broadening such that some have disappeared at 10.5 kG. These are consistent with levels of a triplet state. Three lines which show considerable magnetic activity are those with J’ = 7, 8 and 11 in the K’ = 4 manifold of the 100 state, all of which are perturbed. Fig. 4 and 5 0 GHz 10 0 GHz 10 FIG. 4.-Magnetic perturbation of the 100 K = 4 J = 7 level of HNO 2, ‘A!’, recorded in parallel polarisation. (a) Excitation in RQ3(7) at 16 158 cm-’.(b) Excitation in RR3(6) at 16 177 cm-’. show excitation spectra in various magnetic fields for the J’ = 7 and 8 groups of lines run in parallel polarisation in both “Q and RR branches. At zero field the J’ = 7 group consists of one sharp line and one broad line sepa- rated by 5.55 GHz. The splitting between these is one of the perturbations fitted by eqn (5). In “Q excitation, which accentuates the components with high M, the broad line splits into a triplet, the separation between the outer components being 3.25 GHz in 10.5 kG. In “R excitation, which accentuates M = 0, the broad line splits completely into a doublet of spacing 3.20 GHz in 10.5 kG, with a third weaker under- lying component.Thus both the high M and low M components are split by the magnetic field, with splittings which are approximately first order in the field strength. This can only be brought about if both AJ = 0 and AJ = & 1 Zeeman interactions con- tribute to the splitting. This behaviour is qualitatively just that to be expected for interaction of a singlet level with a Hund’s case (b) triplet level in which N and S un- coup!e at high field.” However, the observed splitting is only 0.06 times that pre- dicted for a case (b) triplet level. The weaker sharp line only broadens by 0.6 GHz in 10 kG, so the magnetic activity does not arise from an off-diagonal Zeeman addition to the zero-field perturbation matrix element. We therefore conclude that the main splitting into a doublet, which was analysed in eqn (5), arises from a singlet-singlet perturbation, but that the stronger component is also coincident with a triplet state level which does not conform to Hund’s case (6).At zero field there are clearly five com- ponents spread over 18 GHz (fig. 5), with a sixth more remote line in both the RR and “Q branches. Two of the weak components are strongly active magnetically, one The J’ = 8 group is even more complex.R. N. D I X O N , M. NOBLE, c. A . TAYLOR AND M . DELHOUME 133 broadening, and one shifting by 8 GHz in 10 kG. The remote level is also active but becomes overlapped by stronger lines when the field is applied. The remaining three components, including the most intense are little affected.The number of com- ponents and their magnetic activity requires that, even when K-doubling is taken into -4.99 kG 0 20 GHz FIG. 5.-Magnetic perturbation of the 100 K = 4 J = 8 level of HNO x, ‘A”, recorded in parallel polarisation. Excitation in KR3(7) at 16 178 cm-’. account, there must be two perturbations to this level, probably one singlet-singlet and one singlet-triplet. A further consequence of a magnetic field is to cause a preferential decrease in intensity of a number o f lines in the spectrum. We show below that at the pressures of most of the experiments (a few mTorr) the excited state is partially collisionally quenched, and that this quenching is linked to perturbations. The selection rule for134 ELECTRONIC SPECTRUM OF HNO magnetic field interactions can result in mixing in a field between levels with AJ = &l which cannot perturb one another in zero field.TIME-RESOLVED MEASUREMENTS A study of quenching rate constants and extrapolated fluorescence lifetimes for the A-8 system of HNO has recently been described by Yamada et aL20 The HNO was generated in a discharge flow system through the radical reaction HCO + NO --f HNO + CO in the total pressure range 0.05-0.5 Torr, and excited by a flash-lamp-pumped dye laser. This had a bandwidth which resulted in simultaneous excitation of many J levels within a sub-band. The lifetimes obtained were in the range 6-10 ,us, and the quenching rate constants in the range 4 x 10-Il-2 x lo-'* cm3 molecule-' s-'. Our initial time-resolved measurements, but with single rotational-level excitation, were made over the pressure range 0.5-200 mTorr.Although measurements over the range 10-200 mTorr resulted in similar extrapolated lifetimes and quenching constants to those reported,20 the Stern-Volmer plots showed considerable curvature. The decay curves at higher pressures were also markedly non-exponential. It is now clear from 100 - 4- .3 s G .- 3 10 W rn Q, +- .3 8 u, Q, a: 1 0 25 50 75 1 t ime/ps FIG. 6.--Serni-logarithmic decay curves for excitation in the RR3(4) line of the 100-000 band of HNO, (a) 0.5 mTorr, (6) 5.4 mTorr, (c) 10.3 mTorr total pressure. our new measurements that energy transfer plays an important role in the decay of HNO fluorescence at pressures of tens of mTorr or higher. We have therefore con- centrated on the pressure range 0.5-1 5 mTorr, and have studied the decay from ca.30 different pumped levels. These have been chosen to include apparently unperturbed levels, levels within systematically perturbed series and randomly perturbed levels. Decay curves were accumulated in the store of the PET microprocessor for 50 to 100 laser pulses, and then output on a recorder in both linear and logarithmic form.R . N . DIXON, M. NOBLE, c . A . TAYLOR AND M . DELHOUME 135 At the lowest pressures of 0.5 to 1 mTorr the logarithmic decay curves were found to be close to linear for at least one decade in all cases where excitation was to a single level. At ca. 10 mTorr the linear portion of many logarithmic decay curves was limited to about half a decade (see fig.6). The initial logarithmic decay constants were determined graphically from the linear portions of the curves if these were suffi- ciently long, or by fitting the decays to a double exponential when necessary and com- puting the initial slopes. The initial slopes were then used in constructing Stern- Volmer plots. The results are summarised in table 1. TABLE 1 .-FLUORESCENCE DECAY RATES AND QUENCHING RATE CONSTANTS FOR SINGLE ROVI- BRONIC EXCITATION OF HNO U' K branch v/cm-' J' kf/104 s-1 k,/10-10 crn3 S-1 01 1 020 1 00 100 15 750.17 751.66 752.60 754.75 755.97 756.98 16 116.06 115.18 113.51 113.75 114.36 110.93 109.15 109.33 16 113.20 112.55 110.77 103.59 097.58 16 172.01 173.87 175.62 177.06 177.25 " 178.28 178.56 178.86" 179.82 180.91" 181.79 181.92 4 5 s 6 m 6 m 7 s 8 m 4 5 6 w 6 m 6 w 8 9 m 9 m 3 4 6 11 14 4 5 6 7 m 7 s 8 w 8 s 8 w 9 s 10 s 11 w 11 m 4.9 i 1.9 4.4 f 0.6 3.8 f 0.4 4.2 i 0.7 6.1 i 1.6 2.7 i 1.8 4.7 i 0.3 5.3 i 0.4 4.1 f 0.5 3.9 i 0.2 3.6 i 0.2 4.2 f 0.2 5.3 f 0.3 4.1 i 0.7 4.4 f 0.6 4.5 i 0.2 4.4 f 0.2 4.1 f 0.5 3.8 f 0.5 4.56 f 0.13 4.48 f 0.33 4.17 f 0.55 4.48 0.42 4.14 f 0.95 4.31 i 0.48 4.54 f 0.42 3.50 f 0.40 4.87 0.48 4.66 f 0.72 w 295 183 rt 16 5.5 & 1.2b 2.6 f 0.7 2.94 f 0.22 2.86 f 0.13 2.92 f 0.36 2.86 i- 0.6 1.64 f 0.16 1.81 f 0.17 1.88 f 0.32 2.83 f 0.16 2.04 f 0.09 2.46 f 0.10 2.28 f 0.17 2.95 f 0.45 1.79 f 0.26 1.46 f 0.09 1.73 f 0.10 2.52 f 0.31 3.65 f 0.34 1.10 f 0.11 1.32 f 0.20 1.76 f 0.30 3.03 & 0.26 2.44 & 0.72 3.56 f 0.28 3.10 & 0.22 2.24 f 0.22 2.59 i 0.24 2.58 f 0.44 5 * 3 1.1 f 0.6b 2.67 f 0.09 1 - " Magnetically sensitive transition.s = Our extrapolated zero-pressure lifetimes are all longer than 20 ,us and are thus several times greater than those reported by Yamada et a1." The quenching rate con- stants range from I x lo-'' to 5 x lo-'' cm3 molecule-' s-', and correspond to half- quenching pressures of 10 to 2 mTorr. The dominant quenching partner in the measurements of ref. (20) was N,, whereas we have a mixture of HNO, N20 and Slow component of strongly double exponential. strong, m = medium, w = weak component of perturbed level.136 ELECTRONIC SPECTRUM OF HNO H20. Since our high-pressure measurements were similar to those of ref. (20), the higher quenching rates measured at low pressure indicate a difference of decay regime, rather than the difference in collision partner.The results in table 1 show that in a number of cases there is a marked discrimina- tion in measured quenching rates with excitation in close-lying levels. In each of these cases one or more of the levels are perturbed. Thus energy transfer cannot be so fast that the decaying excited-state population distribution lacks a memory of the initially populated level. In order to assess the importance of energy transfer we have sought to make a separate measurement of its rate. One excited level makes this possible. We have shown that the J' = 0 level of the 101 vibronic state lies 520 cm-I above the dissocia- tion limit to H + NO, but is stable and emits fluorescence. The excitation spectrum of the 101-000 band at low pressure consists of the one line pP,(l) that leads to J' = 0.All other rotational levels in this state are sufficiently strongly coupled to high levels of the ground state through orbital-rotational coupling that they are predissociated. This one line is too weak to study at the lowest possible pressure, but we have been able to measure a decay rate of (31 j: 3) x lo4 s-I at 5.2 mTorr by summing 350 pulses. If we assume that the fluorescence rate for this level has the typical value of ca. 5 x lo4 s-' then the sum of the quenching and energy-transfer rates from this level is ca. 1.5 x cm3 molecule-' s-'. This rate constant is about three times higher than the highest measured quenching rate constant for a level below the dissociation limit.Thus the fluorescence decay at pressures of tens of mTorr or higher must be dominated by the more persistent levels in a relaxed population. In a short paper on the predissociation of the 101-000 absorption band of HNO Freedman2' noted that the linewidths in this band were pressure dependent, showing an increased broadening at 80 Torr total pressure compared with 10 Torr. At 80 Torr a rate constant of 1.5 x low9 cm3 molecule-1 s-l would lead to a lifetime broadening increase of the linewidth (f.w.h.m.) of 0.02 cm-l, which is just at the limit of Freedman's measurements. Note, however, that collisionally induced line- widths depend on the rate of phase-changing collisions, which is usually higher than the rate of population decay. With energy transfer taken into account the evolution of the population n(i) of an excited level will be given by: where kf is the first-order fluorescence rate constant and k , and k, are second-order quenching and energy-transfer rate constants, respectively, N being the total molecular number density.At this microscopic level we define quenching as irreversible loss from the excited state over the timescale of the measurements. The effective first- order rate constant for the initial fluorescence decay with excitation in level i is then: The slope of the Stern-Volmer plot is therefore given not just by kq(i), but includes contributions from energy transfer. For transfer between levels of equal fluorescence rate this second contribution vanishes, but it can be important where perturbed levels are involved, for which the k,(i) will not all be equal.Thus excitation to a regular level, followed by transfer to a perturbed level with low kf(j), increases the effective quenching constant to above k,(i). Conversely, excitation to a perturbed level withR . N . DIXON, M . NOBLE, c . A . TAYLOR AND M . DELHOUME 137 low k,(i), followed by transfer to a regular level, tends to reduce the effective quenching constant to below the appropriate kq(i). The true microscopic quenching constants of the perturbed levels must therefore be even higher than those in table 1 derived from Stern-Volmer plots of the initial decay rates. We can now assess the measured decay parameters. 011, K = 4 The J-dependent perturbation noted above for this manifold of levels affects all the levels studied with the possible exception of J = 4.The quenching constants for this series are consistently higher than those of unperturbed levels in the 020 and 100 vi- bronic states, 020, K = 4 The two lowest levels ( J = 4,5) are the only clearly unperturbed levels in this series, and have the lowest quenching constants. The higher members are mainly split into doublets with somewhat erratic splittings. Since the lines in each pair have approxi- mately equal intensity we presume that this is an anomalous K-doubling, but this can- not be proved from combination differences. Such perturbations must arise from interactions with more remote levels with large asymmetry splittings. For J = 6 we have one unperturbed level and one almost symmetrically split doublet (see fig.2) but, surprisingly, the highest quenching is for the central level. In general there is a corre- lation between the extent of perturbation and increased quenching. 100, K = 3 The three low J lines studied from this manifold are all clearly unperturbed, and have fairly low quenching constants. The J = 11 level has an anomalously large K- doubling, as also does the level with J = 14. This series of levels breaks off above J = 16. For J = 14 energy transfer upwards in energy can therefore lead to collision- ally induced predissociation. This process may contribute to the high quenching rate. 100, K = 4 The levels in this manifold show a systematic increase in quenching rate up to the predissociation limit at J = 11. Superimposed on this trend there is a maximum quenching for J = 7 and 8, which marks the centre of the local perturbation analysed above.We may comment here that for both these J values the quenching of the mag- netically active levels (highest energy components in both cases) is less than that of the perturbed levels with little magnetic activity, although the differences are not large. The decay with excitation in J’ = 11 is strikingly different from all other cases (fig. 7). The logarithmic decay curve has two linear portions with one very fast decay constant and a slower “ normal ” decay. This fast decay persists at 0.7 mTorr, and arises from intramolecular predissociation. The smaller proportional increase in this decay with increasing pressure gives a similar quenching rate to that for J = 10.We have also found that the intensity in the C.W. excitation spectra of the weak lines lead- ing to J ’ = 11 increases relative to other J values with increasing pressure, as does the intensity of the slow component of the decay curve. This indicates the importance of collisional stabilisation of unstable levels to the intensity distribution in excitation spectra. One general feature of the results in table 1 is the small range of zero-pressure fluorescence rates k,. We had expected that the rates for the perturbed levels would have been reduced rather more strongly below the mean value of 4.35 x lo4 s-’138 ELECTRONIC SPECTRUM OF HNO too 10 1 0 5 10 15 20 25 timelps FIG. 7.-Semi-logarithmic decay curve for excitation in the RR3(10) line (high-frequency component of the 100-000 band of HNO.Pressure 2.6 mTorr. (z = 23 ps). at 0.5 mTorr. Energy transfer may still be causing some rotational scrambling even FLUORESCENCE YIELD The explanation for the intensity anomaly noted in fig. 1 is now quite clear. At the centre of the perturbation ( J = 7 and 8) the increased quenching compared with low and high J levels leads to a lower quantum yield for fluorescence. As the pressure is increased excitation in neighbouring lines is also subject to this loss through energy transfer into J = 7 and 8. This loss is particularly evident in C.W. laser excitation since the fluorescence is then integrated over the complete excited-state decay. We have been able to simulate this behaviour by numerically integrating the Master equa- tions [eqn (6)] for a simple model approximating to the full set of levels in the 100 K = 4 manifold.This model also reproduces the observed departure from single exponential decay at high pressures. A similar striking intensity anomaly is evident in the fluorescence excitation spec- trum of the ~,1A”(050)-~,1A’(000) band of HCCl reported by Kakimoto et aZ.22 Lines from four J values centred on K’ = 0, J’ = 9 and 10 are weak. In the presence of a magnetic field of 10 kG the perturbed lines decrease further in intensity and some in- crease in width by 1-2 GHz. Lifetime studies on this system would confirm whether this is another example of preferential collisional quenching. DISCUSSION The levels of the 2, lA” state of HNO considered in this paper lie ca.16000 cm” above the lowest level of the 2, ‘A’ state, ca. 9000 cm-’ above the presumed origin5 of the a“, 3A” state, and within 700 cm-l of the ground-state dissociation limit at 16 450 cm-l. We have computed the average density of vibrational levels at this energy from our analytical potentials using the Thomas-Fermi statistical approxima- tion, giving one level per 20 cm-’ for the 8 state and one level per 50 cm-’ for the a” state. An important feature of the potentials is that, whereas the 8 state is the groundR. N. DIXON, M. NOBLE, C. A . TAYLOR AND M. DELHOUME 139 state near the equilibrium geometries of the 8 and 2 states, the 8 state potential rises above that of the a” state for modest displacements of the H atom. Consequently the 8 and a” potential surfaces intersect over a wide range of energies.The weak magnetic character of many apparently unperturbed levels of the 2 state is far too large to be consistent with a pure singlet state, but no strong perturbed line has a magnetic moment consistent with a pure triplet state interaction. We therefore propose that most, if not all, of the perturbations arise from the following mechanism. The levels of the 2, ’A” state are mixed through orbital-rotational interaction with nearby levels of the 8, ‘A’ state, and these are in turn extensively mixed through spin- orbit coupling to nearby levels of the a“, 3A“ state. The perturbations then arise when there are AJ = 0 degeneracies between levels of the 2 and 2 states, but with some triplet state contamination of the wavefunctions.In addition there may be isolated direct 3-a” perturbations, but we have found that the only levels which are strongly magnetic are members of multiply perturbed groups. We have already concluded that predissociation of the 2 state proceeds through this coupling to high levels of the 2 state, with no selection rule on K, because of the very large vibrational amplitudes of these 8 state levels. Further evidence is that the greatest predissociation line- widths are for the 101 and the strongest perturbations occur in the 001 and 01 1 levels.2 In all three cases one quantum of the bending vibration v, is excited. Since the x a n d 8 states become degenerate in linear HNO, at energies not much higher than these states, excitation of v3 will enhance the Franck-Condon factors for the 2-8 per- turbations.Gelbart and Freed23 have put forward a model for fluorescence quenching which specifically attributes the rates of collision-induced electronic quenching to the mixed character of the wavefunctions of perturbed levels. In a first-order treatment this rate is given by (p’) times the rate of rotational relaxation, where (p’j is the sum of the squares of the coefficients of pure final-state wavefunctions in the mixed-state wave- function. The quantitative validity of this model has been questioned for intersystem crossing in CO and an alternative formulation For apparently unperturbed levels our quenching rates are 8-15% of the rate of rotational relaxation, rising to ca. 20-300/, for perturbed levels.The rates are in quali- tative agreement with the Gelbart and Freed model but do not correlate with (p2) where this can be estimated. Perhaps this difficulty arises because no level of the 2 state with J > 0 is truly unperturbed. Furthermore, the random nature of the pertur- bations is probably a reflection of the random pattern of levels of the A? state close to dissociation. The pattern of perturbing levels may be more regular at the energy of the 000 level of the x state, permitting a more quantitative interpretation of time- resolved measurements. In H2C0 the lifetimes of many of the levels of the 2,”’’ excited state are greatly shortened by collision-free internal conversion to high levels of the 8, ‘ A , state, which form a “ lumpy continuum ” at this energy.Weisshaar and Moore have found that these lifetimes can be modified by Stark tuning of the interacting levels into or out of resonance.* Such a process cannot occur for HNO since, even if J and parity are the only good quantum numbers for the high ground-state levels, the level spacing is still far coarser than the level widths. This behaviour of HNO can be contrasted with that of formaldehyde. We are indebted to the S.R.C. for funds to purchase the three laser systems, and Keith Rosser’s technical We also for financial support for two of us (M. N. and C. A. T.). assistance was invaluable in the construction of much of the equipment. thank Dr. D. A. Ramsay for the original of fig. l(a).140 ELECTRONIC SPECTRUM OF HNO APPENDIX Table 2 presents the mean upper-state term values. Where K-doublings have been re- solved, and the parities assigned, these are denoted by the e, f notation of Brown et For perturbed levels only the relative intensities of the various lines in the excitation spectra are indicated by s-strong, m-medium, w-weak and vw-very weak.TABLE 2.-EXCITED-STATE TERM VALUES J parity T/cm - J parity T/cm - (a) 011, K = 3 3 4 5 5 5 6 6 7 7 8 8 9 9 10 10 10 11 4 5 5 6 6 7 7 8 8 8 9 10 11 3 4 5 6 7 7 8 8 9 9 10 10 11 11 (b) 011, K = 4 (c) 020, K = 3 e f e f e f e f e f e f e f f f e f e f e e f f e 15 758.24 768.35 780.69 w 780.94 m 781.13 m 796.14 796.27 81 3.81 81 3.99 834.02 834.29 856.68 m 857.03 s 881.62 m 882.31 s 883.33 w 909.20 w 15 920.17 932.53 s 934.51 vw 947.04 m 949.19 m 963.93 mw 966.70 s 983.10 vw 986.75 m 986.99 m 16 009.39 034.65 062.20 16 150.05 160.13 172.73 187.85 205.43 205.50 225.62 225.67 248.32 248.39 273.50 273.51 301.24 301.25 11 11 12 12 12 13 13 13 14 14 14 15 15 15 16 17 18 12 12 13 14 14 15 15 16 16 17 17 17 12 12 13 13 14 14 15 15 16 16 17 17 18 18 e f f e f f e f e f e e e f e e e f e e f e f e f e f f e f e 910.11 s 910.76 w 939.05 m 940.42 s 940.88 w 971.23 w 973.24 s 973.67 m 16 008.55 w 009.12 m 009.28 m 045.52 w 046.78 w 16047.11 m 087.31 m 130.38 m 175.93 m 092.58 m 092.84 m 125.43 160.73 m 161.77 ms 198.44 ms 199.50 m 236.86 ms 239.89 ms 281.77 m 282.29 w 282.89 m 33 1.44 331.46 364.17 364.18 399.43 399.45 436.98 437.17 477.35 477.40 520.15 520.22 565.43 565.55R .N. DIXON, M. NOBLE, c. A . TAYLOR AND M. DELHOUME 141 TABLE 2.-continued J parity T/cm-' J parity T1cm-l 2(d) 020, K = 4 4 5 6 6 6 7 7 8 2 e 2 f 3 e 3 f 4 e 4 f 5 e 5 f 5 e 6 f 6 e 6 f 7 e 7 f 8 e 8 f 9 e 9 f 9 e 10 e 2(e) 100, K = 2 2 ( f ) 100, K = 3 3 4 5 6 e 6 f 7 e 7 f 8 e 8 f 9 e 9 f 10 e 10 f 11 e 2(g) 100, K = 4 4 5 6 7 7 8 8 8 16 296.92 309.62 324.25 w 324.49 m 325.10 w 342.11 m 342.62 w 362.41 16 093.78 093.81 101.52 101.53 1 1 1.77 1 1 1.95 124.67 w 124.68 s 124.90 m 139.81 m 140.02 s 140.03 m 157.93 157.96 178.37 178.42 201.42 m 201.51 s 201.74 m 226.87 m 16 197.89 208.12 220.92 236.22 236.26 254.09 254.13 274.56 274.64 297.43 297.65 323.18 323.18 351.25 16 342.01 352.74 370.07 387.81 mw 388.00 m 408.02 w 408.22 w 408.31 m 9 9 10 10 11 11 12 13 10 10 11 1 1 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 11 12 12 13 13 14 14 15 15 16 16 17 17 8 8 9 9 10 10 11 11 f f f f e f e f e e f e f f f e e e e e f f f f f e e f e f e e e 385.08 m 385.26 m 410.31 m 410.53 m 438.11 m 438.35 m 468.42 501.25 227.20 s 227.27 w 255.29 255.42 286.23 286.88 319.55 320.28 355.58 356.27 394.08 394.38 435.26 435.21 478.66 478.88 524.62 525.1 1 573.22 573.82 351.54 382.10 382.12 415.26 41 5.37 450.96 451.10 489.52 489.67 530.45 530.53 573.31 573.79 408.42 w 408.61 w 430.71 vw 431.31 s 455.98 vw 456.84 s 484.89 w 485.02 mw142 ELECTRONIC SPECTRUM OF HNO F.W. Dalby, Can. J. Phys., 1958, 36, 1336. J. L. Bancroft, J. M. Hollas and D. A. Ramsay, Can. J. Phys., 1962, 40, 322. J. K. Cashion and J. C. Polanyi, J. Chem. Phys., 1959, 30, 317. M. J. Y. Clement and D. A. Ramsay, Can. J. Phys., 1961, 39, 205. T. Ishiwata, I. Tanaka and H. Akimoto, J. Phys. Chem., 1978, 82, 1336. M. A. A. Clyne and B. A. Thrush, Discuss. Faraday Soc., 1962, 33, 139. J. C. Weisshaar and C. B. Moore, J. Chem. Phys., 1980, 72, 5415. K. Shibuya and E. K. C. Lee, J . Chem. Phys., 1978, 69, 5558. ' R. N. Dixon, K. B. Jones, M. Noble and S. Carter, Mol. Phys., 1981, 42, 455. lo G. W. Kirby and J. G. Sweeney, J. Chem. Soc., Chem. Commun., 1973, 704. l1 J. E. T. Corrie, G. W. Kirby, A. E. Laird, L. W. MacKinnon and J. K. Tyler, J. Chem. SOC., l2 S. Gerstenkorn and P. Lucy Atlas du Spectre d'absorption de la molecule d'iode (C.N.R.S., 1978). l3 R. N. Dixon and M. Noble, Chem. Phys., 1980,50, 331. l4 T. W. Hansch, Appl. Optics, 1975, 11, 895. l5 J. W. C. Johns and A. R. W. McKellar, J . Chem. Phys., 1977, 66, 1217. l6 K. Takagi, S. Saito, M. Kakimoto and E. Hirota, Proc. 14th Int. Symposium on Free Radicals Chem. Commun., 1978, 275. (Sanda, Japan, 1979), p. 254. K. Takagi, S. Saito, M. Kakimoto and E. Hirota, J . Chem. Phys., 1980,73, 2570. la C. G . Stevens and J. C. D. Brand, J. Chem. Phys., 1973,58, 3324. l9 J. C. D. Brand and C. G. Stevens, J . Chem. Phys., 1973, 58, 3331. 2o F. Yamada, T. Ishiwata, M. Kawasaki, K. Tsukiyama, K. Obi and I. Tanaka, Proc. Proc 14th Int. Symposium on Free Radicals (Sanda, Japan, 1979), p. 275. 21 P. A. Freedman, Chem. Phys. Lett., 1976, 44, 605. 22 M. Kakimoto, S. Saito and E. Hirota, Annual Review of the Institute for Molecular Science 23 W. M. Gelbart and K. F. Freed, Chem. Phys. Lett., 1973, 18, 470. 24 D. Grimbert, M. LavollCe, A. Nitzan and A. Tramer, Chem. Phys. Lett., 1978, 57, 45. 25 J. M. Brown, J. T. Hougen, K.-P. Huber, J. W. C . Johns, I. Kopp, H. Lefebvre-Brion, A. J. Merer, D. A. Ramsay, J. Rostas and R. N. Zare, J . Mol. Spectrosc., 1975, 55, 500. (Okazaki, Japan, 1979), p. 54.