Far-infrared Laser Magnetic Resonance Spectra of Vibrationally Excited OH and OD BY P. B. DAVIES,* W. HACK AND H. GG. WAGNER Max-Planck-Institut fur Stromungsforschung, 6-8 Bottingerstrasse, 3400 Gottingen, West Germany Received 14th January, 198 1 Several new far-infrared laser magnetic resonance spectra of OH and OD have been detected in the reactions of H or D atoms with ozone. The strongest spectra have been assigned to rotational tran- sitions in u = 6 OH and u = 4 OD of the X state. The OH radical is one of the most widely studied of all transient gaseous species, and is of fundamental importance in spectroscopy and in reaction kinetics and dynamics. Spectroscopic studies have ranged from the radiofrequency to the ultraviolet yielding a large amount of structural information which has been used to test ab initio calcu- lations.Of particular interest in the present work are the rotational transitions of the X 'II electronic state both in the ground and excited vibrational levels. Rotational term values were first tabulated by Dieke and Crosswhite' for levels up to v = 3 based on the analysis of the A 2C -+ X 211 system. Data on the higher vibrational levels have come from studies of the vibration-rotation Fourier transform spectrum.' Recently, Coxon3 has determined optimum molecular constants for OH based on the wealth of data available from optical, vibration-rotation and microwave spectra. His tabulated term values for OH from u = 0 to 5 are much more accurate than the earlier data and provide an excellent data set for calculating rotational transition wave- numbers.However, these term values exclude data based on the measurement of the rotational transitions themselves. Work by Clyne et al.4" on the A-X system of OD is a useful starting point for predicting approximate frequencies in the lower vibra- tional levels. However, much higher quality data have recently been published 4b from a simultaneous fit to vibration-rotation, electronic and microwave spectra. The transitions within the X 211i manifold of OH and OD that are amenable to study by far-infrared laser magnetic resonance (1.m.r.) can be approximately classified as: (i) pure rotational transitions within the R = 1/2 or 3/2 states and (ii) transitions where AQ = & 1 , which are forbidden electric dipole transitions in the limit of a pure Hund's case (a) molecule.However, although relatively weak the latter were among the first 1.m.r. transitions discovered and as~igned.~ Early work on the ground vibra- tional state at 79.1 ,urn was later extended to u = 1 , 2 and 3 at 78.4 and 79.1 Based on precisely measured or calculated g-factors these studies yielded accurate values of the 2113,2, J = 3 ++ 2111,2, J = + fine-structure spacing with change in vibra- tional quantum state. In contrast the allowed rotational transitions (i) are much stronger and in particular the OH 'rI3,2 u = 0 J = 3 + 4 1.m.r. transition around 118 pm is one of the most common and widely used in kinetics.' However, because the * Permanent address : Department of Physical Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EP.16 FAR-INFRARED L .M . R . OF OH A N D OD rotational interval changes markedly with u, for the pure rotational transitions, it is difficult to obtain 1.m.r. spectra arising from more than one vibrational state with a single laser line. Several examples of both types of transition have recently been des- cribed by Geiger et al. for u = 0 OD(X Fortunately there is a convenient and efficient source of vibrationally excited OH and OD radicals, namely the reaction of H or D atoms with ozone.9 At short reac- tion times it is relatively easy to generate considerable concentrations of vibrationally excited molecules at pressures of 1.3 mbar or less by this reaction. In the present work we describe an extension of earlier studies6 on vibrationally excited OH at 78.4 and 79.1 pm to pure rotational transitions in vibrationally excited OH and OD using other water-vapour laser lines.The 1.m.r. studies complement earlier electron paramagnetic resonance measure- m e n t ~ ~ - ' ~ of vibrationally excited OH and OD which provide useful tests of the theo- retical accuracy of the g-factors necessary to calculate the Zeeman parameters. EXPERIMENTAL The far-infrared laser magnetic resonance spectrometer has been described in detail else- where.13 The fast flow system used to generate OHt was similar to that employed in the earlier studies6 but modified to maximise the effective sample volume by displacing the en- trance and exit ports of the flow system in the magnetic-field region. The nominal pumping speed was 4000 dm3 min-'. Optimum signals of OH? were produced by discharging He (99.996 %), which had been passed through Teflon tubing slightly porous to air, in a 2450 MHz discharge and mixing the discharge products with ozone as close to the laser axis as possible.Ozone was prepared by passing pure O2 through a commercial ozoniser (Argentox mbH), storing on silica gel (-78 "C) and displacing it from the trap as required with a flow of pure He. The concentration of vibrationally excited OH was found to be critically dependent on the amount of H2 present in accord with the known rapid relaxation14 of OH? by Hz. The addition of H2 to the discharge was only of marginal value in increasing [OH?] and often yielded considerably smaller concentrations. Vibrationally excited OD was produced by discharging D2 + He mixtures and reacting the products with 03.For both OH and OD, concentrations of vibrationally excited species were relatively inert to pressure between 0.4 and 0.9 mbar. Laser lines used in this study were those from H20 and D20 gas discharge lasers, namely: HzO D20 78.4, 79.1 and 118.6 ,uni 84, 108 and 171 ,urn. The frequencies of these lines are accurately known ( f l MHz) with the exception of the D20 line at 171 pm. This line does not appear to have been used before in published 1.m.r. work and Benedict et a1.16 give its wavenumber as 58.25 cm-'. Interchange between laser lines and suppression of higher order modes was greatly facilitated by incorporating an intracavity iris. The magnetic field was calibrated by measuring the positions of other well- known spectra.Examples of the strongest spectra obtained, with a He cooled bolometer detector, field modulation at 90 Hz and phase-sensitive detection are shown in fig. 1, 2 and 3 and discussed below. RESULTS ASSIGNMENT AND ZERO-FIELD SPACINGS The assignment of the new spectra reported here to " pure " rotational transitions within the 2113,t manifold of OH and OD is relatively straightforward. Allowed rotational transitions for levels up to u = 5 were first derived from Coxon's3 termP . B . DAVIES, W . HACK A N D H . G g . WAGNER rs < + t t I 'L -lr a t a -. I t , IN I 1 1 N I nlN f- t c: 3 4 n 5 .t -? N 3 c 0 1718 FAR-INFRARED L.M.R. OF OH AND OD I U 1 1 .o 1.1 magnetic field intensity/T FIG, 2.-High-field portion of the 108 pm c spectrum showing two components (0) of thef-tf series assigned in fig.4 and the M,. = + 1 /2 component of the e + e series shown in fig. 1 . values. For a particular transition the rotational spacing varies linearly with vibra- tional quantum number. The v = 0 to 5 data were fitted with a linear regression for- mula (r2 = 0.999 98) and then used to predict transition wavenumbers for levels from u = 6 to 9, which can also be populated in the H + O3 reaction9 Results are given in table 1. Transitions in the lower vibrational levels of OD taken from the term values of Amiot et are also given in table 1. These data are sufficient to assign the strongest spectra and examples for OH and OD are given below. i AM, = * I 0 0.5 1 .o magnetic field intensity/T FIG.3.-L.m.r. spectrum assigned to v = 4 OD X 2r13,2 in perpendicular polarisation with the 171 pm D20 laser.P. B. DAVIES, W. HACK AND H. Gg. WAGNER 19 TABLE 1 .-ROTATIONAL TRANSITION ENERGIES IN THE LOWER VIBRATIONAL LEVELS OF OH AND OD (X 2n3,2) OH" parity e - e f-f e - e f - f e + + e f -f v = o 1 2 3 4 5 6 7 8 9 83.723 80.733 77.772 74.828 71.888 68.931 65.97 63.01 60.06 57.102 83.869 80.868 77.897 74.944 7 1.994 69.027 66.06 63.09 60.12 57.16 1 18.208 113.904 109.649 105.427 101.21 6 96.990 92.73 88.49 84.25 80.01 1 18.453 114.136 109.865 105.627 101.401 97.160 92.88 88.63 84.3 7 80.12 153.189 147.524 141.930 136.385 130.859 125.319 119.71 114.14 108.57 103.01 153.535 147.850 142.236 136.670 13 1.123 125.562 119.94 114.35 108.76 103.17 OD v = o 46.384 46.414 65.063 65.117 83.833 83.916 1 45.159 45.187 63.333 63.384 81.587 8 1.666 2 43.944 43.970 61 -61 8 61.666 79.363 78.437 3 42.737 42.761 59.91 5 59.960 77.154 77.224 4 41.534 41.557 58.219 58.262 74.958 75.024 5 40.333 40.354 56.526 56.566 72.765 72.827 a Data for u = 0-5 taken from ref.(3) estimated uncertainty ca. 0.002 cm-'. Data for D = 6-9 from a linear extrapolation of the u = 0-5 data with an estimated error < 0.05 cm-'. Data taken from ref. (46) with stated accuracy 0.005 cm-'. THE 108 ,urn SPECTRUM OF OH The spectrum recorded with this laser line in perpendicular polarisation is shown in fig. 1 and 2. The frequency of the 108 pm D20 laser l5 is 2 783 066.6 Ifi 1 MHz = 92.833 109 cm-', close to the J = 3 -+ transition in u = 6 OH (table 1).The g- factors required to calculate the Zeeman effect in these rotational levels have not been measured but can be calculated using Radford's formulae.'' Lee and Tam9 have shown that the theoretical g-factors calculated in this way are in excellent agreement TABLE 2.-MOLECULAR PARAMETERS FOR OH V = 6 AND J = 512 ++ 7/2 ROTATIONAL INTERVALS vL = 2 783 066.6 -i 1 MHz = 92.833 lo9 cm-' E = 27 627.0 cm-' Lee and Tam9 V (ITIALy + 2 BLylZ>o = -56.16 cm-' <nlBLyl z>o R = ($) - - 9.812 J = 512 J = 712 g; calc 0.466 30 0.310 05 g7 calc 0.467 85 0.311 91 f i e - fk f i r - f 1 r 92.71 5" 92.874" " Estimated error <0.002 cm-'.20 FAR-INFRARED L.M.R. OF OH A N D OD with experimental e.p.r. measurements of g $ for OH X2113,,, J = 4 for levels ZJ = 5 to 9.We have adopted the same procedures to calculate the g-factors for J = 3 and z, u = 6, and these are given in table 2. The parameters required in the calculation were taken from Lee and Tam9 and are also given in table 2, together with the zero- field spacings which are derived from the 1.m.r. spectra. The results are in excellent agreement with predictions from the extrapolation of Coxon's work. By calculating the A doubling frequencies for both rotational levels it was confirmed that the laser frequency falls between the rotational transition frequencies. An alternative assign- ment in which both transitions are higher or lower in frequency than the laser gives an unacceptable value for vL(+) - vA(3). THE 171 pm SPECTRUM OF OD The much smaller A doubling in the corresponding rotational levels of OD results in a more compact 1.m.r.spectrum. This can be seen in the 171 pm D20 laser spec- trum shown in fig. 3, where the two series arising from thef-fand e t) e transitions are correspondingly closer together than in OH. The wavenumber given by Benedict et aZ.,16 58.25 crn-l, is close to the J = 9 -+ $ transition in u = 4 OD (table 1) and this strong spectrum can be positively assigned to this transition. The g-factors re- quired to calculate the Zeeman pattern were derived in a similar manner to OH and based on the data of Rashid et a1.12 (table 3). TABLE 3,-cALCULATED &'-FACTORS FOR J = 5/2 AND 712 OD, x ' ~ J / z IN THE U = 4 LEVEL J = 512 J = 712 g : 0.4329 0.2807 gJ 0 . 4 3 3 3 0.2812 Assuming that the laser frequency is sufficiently accurate, i.e., that the laser fre- quency is greater than the e t) e spacing and less than thef-fspacing, this leads to the assignment in fig.3. This was confirmed by frequency-pulling experiments, in which the laser frequency was changed slightly by cavity length adjustment and the field shifts examined. Supporting evidence for this ordering of the transition fre- quencies and laser frequency comes from a qualitative calculation of the A doubling intervals.12 DISCUSSION Assignment of the more intense spectra whether of the rotational or fine-structure type is relatively straightforward. However, we have detected several weaker spectra which can be attributed to 160H but which are accompanied by spectra of similar intensity probably due to HO, (DO,), other isotopic forms of OH and trace impurities.This makes assignment of the weaker 160H and 160D spectra difficult with the rela- tively few lines available from H20 and D20 lasers. For example, one of the more prominent of the weaker spectra (5' : N z 10 : 1) appears with the 118.6 pm H20 laser and could be due to the J = 3 -+ 3 transition in u = 8 OH. It is clear from table 1 that by using the more versatile optically pumped spectrometers there are many more accessible 1.m.r. spectra in the higher vibrational levels. This should permit rotational, fine-structure and A-doubling parameters to be measured for levels higher than. u =5 in OH for which there are already high-qualityP. B. DAVIES, W . HACK AND H . G g . WAGNER 21 0.7 0.6 0.5 0.4 0.3 0.2 4 0.1 E J; =I 2 0 - 7 - 0.1 cd -0.2 cd -0.3 -0.4 2 I .X J; - 5 2 2 - 0 .3 * U - 0 . 5 -0.6 I 1 0.5 10 magnetic field intensity/T FIG. 4.-Zeeman components of the J = 5/2 --f 7/2 ( f + f ) transition in OH X 'ILi2, u = 6. The DzO laser frequency has been subtracted from the zero-field frequency: 0, perpendicular; A, parallel 1.m.r. transitions. G. H. Dieke and H. M. Crosswhite, J . Quatit. Spectrosc. Radiat. Tratisfer, 1962, 2, 97. J. P. Maillard, J. Chauville and A. W. Mantz, J . Mol. Spectrosc., 1976, 63, 120. J. A. Coxon, Can. f. Phys., 1980, 58, 933. ( a ) M. A. A. Clyne, J. A. Coxon and A. R . Woon Fat, f. Mol. Spectrosc., 1973, 46, 146. (6) C. Amiot, J-P. Maillard and J. Chauville, J . Mol. Spectrosc., 1981, 87, 196. K. M. Evenson, J. S . Wells and H. E. Radford, Phys. Rev. Lett., 1970, 25, 199. P. B. Davies, W. Hack, A. W. Preuss and F. Temps, Chetn. Phys. Lett., 1979, 64, 94. C. J. Howard and K. M. Evenson, J . Cherii. Phys., 1974, 61, 1943. K. P. Lee and W. G. Tam, Cheni. Phys., 1974,4,434. * J. S. Geiger, D. R. Smith and J. D. Bonnett, Chein. Phys. Lett., 1980, 70, 600. lo K. P. Lee, W. G. Tam, R. Larouche and G. A. Woonton, Cait. J . Phys., 1971, 49,2207. l1 P. N. Clough, A. H. Curran and B. A. Thrush, Proc. R. Suc. Lorzdun, Ser. A, 1971, 323, 541. l2 M. H. Rashid, K. P. Lee and K. V. L. N. Sastry, J . Mol. Spectrosc., 1977, 68, 299. l3 W. Hack, A. W. Preuss and H. Gg. Wagner, Ber. Butisetrges. Phys. Cheni., 1978, 82, 1167. l4 J. E. Spencer, H. Ender and G. P. Glass, Sixteenth Znt. Synip. Cotiibiistiorz (The Combustion l5 J. T. Hougen, H. E. Radford, K. M. Evenson and C. J. Howard, J . Mul. Spectrosc., 1975, 56, l6 W. S. Benedict, M. A. Pollack and W. J. Tomlinson, ZEEE J . Qirantiim Electron., 1969, QE-5, l7 H. E. Radford, Phys. Rev., 1961, 122, 114. Institute, Pittsburg, 1976), p. 829. 210. 108. J. A. Coxon, K. V. L. N. Sastry, J. A. Austin and D. H. Levy, Can. J . Phys., 1979, 57, 619.