Laser Stark and Interferometric Studies of Thioformaldehyde and Meth yleneimine BY GEOFFREY DUXBURY * AND HIROSHI KATO Department of Natural Philosophy, University of Strathclyde, 107 Rottenrow, Glasgow G4 ONG AND MICHEL L. LE LERRE School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 ITS Received 22nd December, 1980 Laser Stark and optical-optical double resonance experiments have been carried out on HzCS and on D2CS, and rotational and vibrational constants obtained for the v3 bands. The dipole mo- ments in the v3 and v4 states of HzCS, and in the vj and ground states of D,CS have been measured. Interferometric spectra have been obtained and analysed for the v4, v5 and vg bands of CHZNH, and some laser Stark and interferometric spectra obtained in the perturbed 10 pm region.Very few high-resolution infrared spectra of short-lived molecules have been ob- served using conventional grating spectrometers owing to their inherently low sensi- tivity. Two molecules, H2CS and HBS have been observed in the 3 pm region using these methods, but their observation was time-consuming and difficult. Most recent experiments to observe the infrared spectra of transient molecules have used laser spectrometers, because of the high brightness and narrow linewidths of the sources. The laser spectrometers fall into two categories, those which use fixed-frequency, high- power C02 or CO laser^,^ and those which use tunable low-power lasers such as diode4 and colour-centre lasers. The fixed-frequency spectrometers make use of electric or magnetic fields to tune the molecular transitions into resonance with the lasers, and are known as laser Stark and laser magnetic resonance spectrometers, respectively.The high power of the CO, and CO lasers allows sub-Doppler resolution to be obtained, but only a limited frequency coverage. The diode and F-centre lasers possess good frequency coverage, but the low powers mean that, in general, the resolution is Doppler limited. In the present work we will demonstrate the application of laser Stark spectroscopy, optical-optical double resonance and high-resolution interferometry to the study of short-lived species. High-resolution Michelson interferometers of the Connes type have considerable advantages over grating instruments both in light throughput, the Jacquinot advantage, and in the multiplcx (Fellgett) advantage.Although they do not possess quite the sensitivity of diode laser spectrometers, they allow much greater spectral regions to be covered with comparable resolution, and with very great preci- sion. The use of lhser Stark spectroscopy and of optical-optical double resonance to * Visiting Scientist, Kitt Peak National Observatory, operated by the Association of Universities for Research in Astronomy. Inc., under contract with the National Science Foundation.98 SPECTRA OF HZCS, D2CS AND CH2NH characterise short-lived molecules will be exemplified with the 10 ,urn spectra of H2CS and of D,CS. Methyleneimine, CH,NH, will also be used as an example of the com- bined use of laser Stark spectroscopy and of interferometry, in order to show the com- plementary nature of the methods.EXPERIMENTAL LASER STARK AND OPTICAL-OPTICAL DOUBLE RESONANCE The basic principles of laser Stark spectrometers are described by Freund et al.5 and of this particular spectrometer by Caldow ct a1.6 The use of an acousto-optic modulator to produce a single frequency-shifted sideband is described by Bedwell and D ~ x b u r y . ~ The main differences in the present work are that a more efficient Phillips acousto-optic modulator has been used, so that up to 60% of the laser power can be obtained in the sideband with a radiofrequency drive power of 10 W, and that the sideband is at higher frequency than the carrier. The Edinburgh Instruments semi-sealed PL3 laser used was run with l2CI6O2, 12C1'02 and 1 3 P 0 2 isotopic species.The small frequency-difference region can be probed if desired by using amplitude modulation of the sideband at 1-5 MHz, via a double balanced mixer before the r.f. power amplifier. A similar method has recently been used in the visible region by Dixon and Noble,' and Orr and Oka have successfully used an electro-optic modu- lator to perform similar experiments. The effective spacing of the Stark electrodes was deter- mined using laser Stark and OODR spectra of I2CH3F and 13CH3F, for which the magnitude of the Stark effect is accurately known.6*'0 The thioformaldehyde, H,CS, and the fully deuterated thioformaldehyde, D2CS, were produced by the pyrolysis of dimethyl disulphide and fully deuterated dimethyl disulphide, respectively.The niethyleneimine, CH2NH, was produced either by the pyrolysis of methyl- amine or of diaminopropane. The pressures used in the Stark cell ranged from 5 to 60 pm, depending upon the Stark voltage used. INTERFEROMETRY The interferometric spectra of CH2NH were obtained using the Solar Telescope Fourier transform spectrometer (FTS), of Kitt Peak National Observatory, Tucson, Arizona. This is a high-resolution rapid scanning machine developed by Dr. J. W. Brault," and which has a maximum unapodised resolution of 0.006 cm-' in the 10 pm region. Spectra were ob- tained between 11 and 2 pm using a KBr beam splitter and various combinations of filters. The ability of the machine to provide rapid scans at medium resolution was crucial in optimis- ing the production of the imine, so that the concentration of methylamine was minimised.Methylamine possesses a very complex spectrum in the region of many of the fundamental bands of CH2NH, so that its concentration must be kept low to prevent the imine spectrum being masked. These molecules possess strong spectra in the 11-2 pm region, but the absorption lines are widely spaced, and the high resolution and precision of the interferometer allowed their lines to be easily assigned. The optimum pressure for the production of CH2NH using an S-bend pyrolysis tube was ca. 500 pm. A total of sixteen traversals of a 60 cm multiple-pass absorption cell were used. The scan time of the interferometer was between 5 and 10 min, depending upon the number of sample points needed. The minimum number of scans used was 8 and the maximum number 16.The Fourier transformation was carried out at KPNO, and the data analysed using the Honeywell Multics system at BristoI University. The line positions were determined by fitting a parabola to the highest three sample points of each line. The accuracy of this procedure was tested on the data run with the InAs filter, with a sample spacing of 0.0027 cm-'. Ground-state combination differences generated from these data had an accuracy of ca. 2 x cm-'. The other impurities produced by the pyrolysis are HCN and NH3.G . DUXBURY, H . KATO AND M . L . LE LERRE 99 RESULTS LASER STARK A N D OPTICAL-OPTICAL DOUBLE RESONANCE SPECTRA OF H2CS A N D OF D2CS The high-resolution infrared spectrum of thioformaldehyde, H2CS, was first ob- served in the 3 pm region by Johns and Olson.' More recently the 10 pm region has been studied by Bedwell et aZ.12 and the lowest frequency fundamental bands, v,, v4 and v6 located.Bedwell and Duxbury also observed optical-optical double reson- ance, OODR, signals between ground-state levels of H2CS. We have extended the study of the v3 band using 12C'80, laser lines, so that improved molecular constants have been derived, and we have also observed several new OODR signals. In symmetric tops or near-symmetric tops, when the asymmetry splitting is negli- gible, the OODR signals form a characteristic doublet pattern of the type first des- cribed by Brewer.13 When a single-sideband modulator is used, the double-resonance condition for signals involving co-propagating beams, and the overall selection rule AMJ = &2 is given by where E is the electric field, SZRF the modulator drive frequency and ,u is the molecular dipole moment.6 Certain simple relationships between the pair of signals hold for the following cases, where E2 > El.(a) Parallel infrared transitions with AKa = 0: El E2 PI (i) QQ lines with no asymmetry splitting, - = (ii) QR or QP lines, 5 !z J (for low J ) . E2 J + 2 (b) Perpendicular infrared transitions with AKa = & l , and AJ = 0: El - Ka E2 Ka + 1 * -- In transitions where the asymmetry splitting is comparable with the r.f. frequency, signals showing resolved M structure can be seen. An example of this is the pQ( 11, 3) transition in the v g band of H2CS reported by Bedwell and D ~ x b u r y .~ Two pairs of OODR signals observed in H,CS are very interesting, since one pair allows the dipole moment difference between the ground state and v3 to be checked, and the second pair allows the dipole moment of the v4 state to be determined accur- ately. The signals involving v4 are shown in fig. 1, and illustrate the sub-Doppler resolution achievable in this type of experiment. By a serendipity both infrared tran- sitions involve the same ground-state rotational level, so that a test of the internal con- sistency of the data can be made. The infrared spectrum of D,CS is much less well-characterised than that of H2CS, and only approximate values of the ground-state vibration frequencies have been determined by Jacox and Milligan l4 using matrix-isolation spectroscopy, We have located and analysed the v3 band using lines from a 13C1602 laser.One of these lines, the 10R32 line at 936.1359 cm-', lies very close to the v3 band origin, and allows the characteristic QQ Stark patterns to be observed, as shown in fig. 2. The intensities of the K, even and odd transitions are reversed from those of H2CS given by Bedwell and The results are summarised in table 1.100 SPECTRA OF H2CS, D2CS AND CH2NH 500 1000 EjkV cm-’ FIG. l.-OODR spectrum of the RR(12, 7) transition of HzCS using the 9P(28) line of a 1zC1s02 laser. The path length was 3 m, the total gas pressure was 40 mTorr, the modulation field 1 V cm-l and the time constant of the detection system 100 ms. The radio frequencies used were (a) 52.06, (b) 60.03 and (c) 68.05 MHz.Duxbury,12 as expected from the nuclear spin statistics when H2 is replaced by D2. The band was analysed with the aid of the computer program DEFIN, which is suitable for C2, molecules, and which was used for H2CS.l2 The Hamiltonian is given in eqn (1)-(5) of ref. (1 5 ) , and uses the Watson centrifugal distortion constants, including the HK term. In the analysis of the v3 band of D2CS, 23 vibration-rotation transitions were used with 61 AM components. The data are suniinarised in table 2 and the derived vibra- tion-rotation frequencies in table 3. The ground-state rotation constants were fixed at the values given by Judge et a1.,16 and the excited-state centrifugal distortion con- stants were fixed at the ground-state values. The initial value of the ground-state TABLE 1 .-OBSERVED RESONANCE ELECTRIC FIELDS FOR OPTICAL-OPTICAL DOUBLE RESONANCE SIGNALS IN H,CS, WITH OVERALL SELECTION RULE \bhf.~I = 2 laser line transition r.f./MHz E‘/V cm-’ p’/D E”/V cm-l ,a”/D 1161“9R10 QR(12, 7 ) ~ 3 52.06 810.6 60.03 937.2 68.05 1060.3 1.654 699.3 1.658 806.2 1.657 915.6 1.648 1.648 1.645 average 11 81 “9P28 RR(12, 7)v4 52.53 732.1 60.41 841.6 68.07 948.1 1.656(3) 1.621 705.9 1.622 811.6 1.622 915.6 1.647( 3) 1.647 1.647 1.646 average 1.622( 3) 1.646(3) a 1161 l2Cl6O 2, (181 12c’80,.G .D U X B U R Y , H . KATO AND M . L . LE LERRE 101 TABLE 2.-oBSERVED STARK RESONANCES IN THE V 3 BAND OF D2CS USING A 13c02 LASER laser line laser frequency resonant field assignment 1cm-l /kV cm-l J ' K t K c t t Jg attK,tr M'-M" lOR(38) 939.9500 1.293 * 1.929 * 3.874 * 0.889 1.249 2.477 * 7.517 * 2,576 * 7.190 * 6.646 10.032 4.394 5.720 8.068 13.178 * 13.926 lOR(32) 936.3705 34.993 * 14.370 * 14.998 * 17.743 * 18.776 * 20.984 * 21.436 * 21.901 * 22.386 * 21.484 * 21.718 * 21.949 * 22.178 * 22.386 * 22.619 * 15.420 * 15.493 * 15.569 * 15.640 * 15.712 * 15.782 * 15.864 * 15.936 * b 0.700 * 24.056 * 24.74 1 SO9 * b 2.905 * 32.73 5.01 3 * 8.063 * 12.216 * - 110 +- 110 111 +- 111 220 f- 220 221 +- 220 33, +- 331 43 +- 43 55 t 55 5 4 +- 5 4 65 +- 65 64 +- 64 75 t 75 76 +- 76 8 5 +- 85 95 +- 95 lo5 t lo5 66 66 3 t 3 24-2 1t1 2 t 3 1-5-2 O t l 4 t 3 3 t 2 1t0 2t1 3473 2 t 2 2 t 3 1t2 O t l 4 t 3 1t0 1t0 1t0 O t l 1t0 O t l 2 - 1 1 - 0 0 - 1 1 - 2 3 t 2 2t1 1t0 04-1 1t2 2 t 3 4 t 3 3 t 2 2 + 1 1t0 04-1 1t2 2 t 3 3 + 4102 SPECTRA OF HZCS, DZCS AND CHzNH TABLE 2.-Continued laser line laser frequency resonant field assignment /cm- /kV cm-I J I K a t K c t -+ JKattKcN M‘-M’’ 1 OR(30) 935.1359 10.926 1 OR(24) 93 1.3093 26.88 16.783 * 34.312 * 29.634 * 38.757 * 29.848 * 39.562 * 20.42 24.445 30.741 43.614 * b 432f- 533 38.730 * 20.1 1 23.730 10R20 39.131 32.060 41.319 * 21.49 25.280 3 1.020 39.900 928.6568 19.520 24.140 * 32.058 * 27.717 * 38.551 * 37.807 12.775 14.880 17.848 * 22.641 * 30.031 * b 441 +- 542 c 734 + 835 C 735 836 744 f- 84s 16.341 20.035 25.774 35.719 * 26.164 * 743 +- 844 36.601 * O t O O t l 4 - 4 3 - 3 4 t 5 3 t - 4 4 - 4 3 - 3 4 - 3 3 - 2 2 - 1 1 - 0 4 - 4 3 - 3 4 - 3 3 - 2 2 - 1 1 - 0 4 - 4 3 - 3 4 - 3 3 - 2 2 - 1 1 - 0 7 - 7 6 - 6 5 - 5 4 - 4 3 - 3 7 t 8 6 6 7 04-1 l t 2 7 - 7 6 - 6 5 - 5 4 - 4 3 t 3 7 t 6 6 + 5 5 t 4 4 + 3 3 t 2 24-1 1 c o 7 t 8 6 t 7 O t l 1 t - 2 * Lamb dip: a components unresolved: overlapped; very weak.G .D U X B U R Y , H . KATO AND M . L . LE LERRE 103 TABLE 3.-THE V 3 BAND OF CDZS transition Vobs/Cm-' vcatclcm-' (v, - v,)/cm-' laser line 940.0049 a 940.0047 939.9586 939.9548 936.3673 936.3673 936.283 1 936.2833 936.281 1 936.28 12 936.21 69 936.21 72 936.1698 936.1700 936.1889 936.141 8 936.141 3 936.141 3 936.13 10 935.1864 931.5480 93 1.4788 93 1.431 1 93 1.2869 940.0042 940.0042 9 3 9.9 5 8 8 939.9549 936.3682 936.3682 936.2832 936.2832 936.2806 936.2806 936.21 70 936.2170 936.1699 936.1699 936.1890 936.1419 936.141 5 936.1415 936.1320 935.1851 931.5477 931.4783 931.431 1 93 1.2867 0.0007 0.0005 - 0.0002 -0.0001 -0.0009 - 0.0009 -0.0001 0.0001 0.0005 0.0006 - 0.0001 0.0002 -0.0001 0.0001 -0.0001 - 0.0001 - 0.0002 - 0.0002 -0.0010 0.0013 0.0003 0.0005 0.0000 0.0002 11 3 I lOR(3 8) I 1 3 I 1 OR( 30) 11311OR(24) 9 9 7 9 7 9 a Obtained by adding the calculated Stark shift to the laser-line frequency.dipole moment was taken from recent work of Cox and Hubbard,17 but was finally allowed to be determined in the least-squares fit. The standard deviation of the fit- ting was 0.0005 cm-'. The vibrational and rotational constants are given in table 4, where they are compared with those of H,CS. The dipole moments of H2CS and of DzCS are compared in table 5. The ground-state dipole moment is taken from the molecular-beam experiments of Fabricant et aZ.19 TABLE 4.-MOLECULAR CONSTANTS OF H2CS AND D2CS (Cm-') HzCS DZCS E" A B c 0 9.726 991 (216) 0.590 391 9 (33) 0.555 451 4 (33) 0.7622 (207) 0.1784 (64) 0.659 (13) 0.1109 (89) 0.406 (83) a v3 1059.2051 (11) 9.715 75 (22) 0.587 34 (36) 0.552 34 (37) 0.7558 (76) b b b b a 0 4.882 06 0.497 186 0.450 168 0.168 2 0.119 3 0.435 6 0.086 68 0.662 4 C v3 936.132 65 (3) 4.891 54 (3) 0.497 12 (9) 0.449 72 (9) b x 10-3 b x 10-4 b x 10-4 b x 10-7 x Ref.(12); fixed at ground-state values; ref. (16).104 SPECTRA OF HZCS, D2CS AND CH2NH I- --- I C (10,5) 1 Q (3,3) I I 1 t 20 25 30 E/kV cm-' 3 5 FIG. 2.-Observed AM = The laser line used is the 1131 lOR(32) line at 936.3705 cm-'. The time constant for detection was 100 m s, the It can be seen that the dipole moment difference, aP = pV3 - po, is much smaller in D2CS than in H,CS.This contrasts with the result for the analogous bands of H2C0 and D2C0,19 but is similar to the results obtained for CH3F and CD3F.20 A similar explanation to that given in the methyl fluorides is proposed, namely that since the v2 and v3 states are separated by ca. 400 cm-' in the hydride, but by ca. 200 cm-' in the deuteride, the increased mixing between Y , and v3 results in a fall in the value of 1 transitions for the Q-branch series of the v3 band of D,CS. modulation amplitude 5 V cm-', and the sample pressure ca. 10 mTorr. TABLE 5.-DIPOLE MOMENTS (D) FOR THIOFORMALDEHYDE AND DEUTERATED THIOFORMALDE- HYDE IN VARIOUS VIBRATIONAL STATES vibrational state H2CS ref.D2CS ref. ground state 1.6491 (4) 18 1.658 (3) this work 1.647 (3) this work v 3 = 1 1.6576 (12) 12 1.661 (3) this work 1.656 (3) this work v4 = 1 1.622 (3) this work v6 = 1 1.642 ( 5 ) 12 8p/8q3. This fall does not appear to occur where the C-O(S) stretching vibration is " isolated ", as it appears to be in H,CO.l9 INTERFEROMETRIC AND LASER STARK SPECTRA OF METHYLENEIMINE Although the infrared spectrum of methyleneimine, CH,NH, has been studied by matrix-isolation spectroscopy by Jacox and Milligan,,' the only high-resolution gas-G . DUXBURY, H . KATO AND M . L . LE LERRE 105 phase spectrum reported has been the laser Stark study of the v, band at 1638.3 cm-' by Allegrini et aLZ2 In the latter paper they also reported an attempt to observe the infrared spectrum using a grating spectrometer, with which they were just able to identify the QQ branches of the v, band.In the present study we have observed the high-resolution gas-phase spectrum of CH2NH using both interferometry and laser Stark spectroscopy. The interferometric spectra spanned the region from 11 to 2 pm, and the laser Stark spectra the 10 ,urn region. Unfortunately the spectra observed in the 10 ,urn region are very complex owing to a strong Coriolis interaction between the v7 and vg vibrations, and a weaker interaction with the v8 vibrational state. The Stark spectra are therefore much less easy to interpret than those of thioformaldehyde described in the previous section, and only fragments of these spectra are so far assigned. All the band assigned here are of A' symmetry, and so in principle possess both type A and type B electric dipole transitions.The v4 band has been obtained using the interferometer with a resolution of 0.006 cm-I. Whereas the transitions seen by Allegrini et a1.22 had J' and K,' values < 3, the interferometric spectra extend up to much higher values, and hence allow a better evaluation of the rotational and centrifugal distortion constants in the v4 state. The 11 l:* a 7 6 ; 5 ' 4 3 t I ' K = 0 3 4 5 6 7 1655.4 1656.2 1657.0 1657.8 wavenumber/cm- FIG. 3.-Interferometric spectrum of the QR(8, K ) transitions of the v4 band of methyleneimine, CH2NH. This spectrum was obtained using an InAs filter, with the fore-optics purged, with sixteen scans, and with a sample spacing of 0.0027 cm-'.The total gas pressure of the pyrolysis products in the 60 cm multiple pass absorption cell was ca. 500 mTorr, and sixteen traversals were used. Peaks marked with an asterisk are those identified by a peak-finding computer program. The pattern seen resembles that of a high-resolution microwave spectrum, with obvious K doubling for K, = 1 and 2 transit ions.106 SPECTRA OF H2CS, D2CS AND CH2NH QQ branches were easily identified, but as the changes in rotation constants are very small, the central Q feature is very compact and hence only partially resolved. The low J and K, QR and QP lines in the spectrum were easily identified using the para- meters of Allegrini et al., but the high J lines were difficult to assign. A stepwise approach was therefore used when the improved constants were used to extend the predictions, until finally 446 QR and QP transitions were assigned.The line positions and assignments of the v4 band are given in Supplementary Publication no. SUP 23024 (11 pp).* A portion of the spectrum is shown in fig. 3, where the asymmetry splitting of the transitions with low values of K, can be seen easily. No type B transi- tions with AK, = &1 were observed, so that the infrared transition moment of this band, which is primarily associated with the C-N stretching vibration, must lie al- most completely along the a molecular axis. In the final least-squares fitting of the data the ground-state parameters were frozen at the values determined from the fit to the microwave spectral data.23,24 The vibrational and rotational constants are given in table 5.The v5 and v g bands at ca. 1453 and 1347 cm-' have not previously been observed in the gas phase. Portions of the spectra in the QQ branch regions are shown in fig. 4 and 5. It can be seen that these branches are considerably degraded, and are similar K = l 11 1 3 2 1 1 2 3 101 I 5j q 33 ! t 2 1 I ~ = 2 8 7 6 5 4 4 , 5 6 7 8 K = 3 1347.2 1348 0 1348.8 FIG. 4.-QQ transitions of the v6 band of CH2NH. The experimental conditions are as those for fig. 3. The K, = 1 and 2 patterns are dominated by asymmetry splitting, and those with large values of J are considerably degraded. in the two bands. The asymmetry splitting of the sub-bands with K, = 1 and 2 can be easily seen in the v6 band. In both bands the transition involving high values of IC,' are perturbed, so that the overall fit is not as good as for the v4 band.The assignments were made with the aid of the ground-state combination differences, which have a very high accuracy as stated in the experimental section. No type B component could be found in the vs band, but one could be identified in the v6 band, where the type B lines were a factor of five to ten times weaker than the corresponding * See Notice to Authors, J. Chem. SOC., Fnrnday Trans., 1980, 76, index issue.G . D U X B U R Y , H . KATO A N D M. L . LE LERRE 107 TABLE 6.-MOLECULAR CONSTANTS OF METHYLENEIMINE (Cm- ') ' ground state V4 v5 v6 ).'O 1638.298 8 (56) A 6.544 896 6.549 45 (11) b6.549 56 (19) B 1.155 557 1.151 3847 (92) d1.151 3766 C 0.979 028 0.972 767 0 (76) A K 213.6 215.3 (41) A j K 19.4 17.54 (12) A j 2.0 1.860 (9) d K 0.3 0.247 (8) 65 16.6 14.1 (7) HK 0.556 0.101 (66) 1638.297 6 1452.053 (12) 1453 6.589 4 (18) 1.593 (18) 0.978 4 (18) 444.(120) -64. (30) d d d 3.6 (2.1) 1344.261 2 (23) 1347 6.778 48 (81) 1.158 335 (86) 0.972 788 (70) 53.7 (12) x 10- 400. (36) x 2.01 (16) x 0.28 (12) x 30. (14) x -1.1 (8) x ~~ Uncertainties in parentheses are three standard deviations in units of the last digit of the para- meter; ref. (22); ref. (21); fixed at the ground-state value. K- 3 4 5 I" ! HCN 2v2R(13) 1.452.2 1.4 53.0 1453.8 1454.6 wavenumber/cm- FIG. 5.-QQ transitions of the vs band of CH2NH, with the same experimental conditions as in fig. 3. Note the strong resemblance to the structure seen in the v6 band.108 SPECTRA OF HzCS, D2CS AND CHZNH ,M .J G 1 - - --___1__- 1 25 30 35 40 EjkV cm-I FIG.6.-AM = fl transitions of a QQ(3, 3) line of CHzNH observed using the 9P(22) line of a COz laser. The time constant was 100 ms, the gas pressure was 5 mTorr, the modulation field 5 V cm-' and both first (If) and second (2f) derivative spectra are shown. This pattern is very similar to those observed in CH3F ( 5 ) and in HzCO.I9 L -__i_ ~ V J 4.0 4 5 5 ,O E= 4.405 V cm-' 824MHzl- 1 FIG. 7.-Saturation spectrum of a QQ(5, 4) transition of CHzNH observed using the 9P18 line of a 12C1B02 laser. The frequency offset is cn. 381 MHz, and the radio frequency 60.09 MHz. (a) Spec- trum produced by carrier and sideband, the central group comprises the counter-propagating double- resonance signals.(h) Saturation spectrum (Lamb dips) with the sideband only. (c) Frequency scan of the laser through the Stark pattern at fixed electric field. The longtiudinal mode spacing of the laser is indicated. The electric field used was 4405 V cm-'.G . DUXBURY, H. KATO AND M. L . LE LERRE 109 type A transitions. The preliminary constants derived from the analyses are given in table 5. The region of the v, and vg bands contain complicated patterns of lines, but some transitions with recognisable structure have been obtained in the laser Stark spectra. In particular a QQ(3, 3) transition has been observed with both the 9P(22) line of the 12C'60, laser, as shown in fig. 6, and with the 9P(48) line of the 12C180, laser. It has also been identified on the interferometric spectrum.Similar transitions, OQ(6, 4) and QQ(5, 4), have been observed using the 9RlO line of the 12C1602 laser and the 9P18 line of the 12C1s0, laser. The frequency offset has been measured using the acousto-optically generated sideband, as shown in fig. 7. The outer and inner patterns of Lamb dips correspond to the transitions induced by the sideband and by the carrier, whsreas the central pattern is due to counter propagat- ing double-resonance signals of the type observed by Mattick et aZ.24 The sub-Doppler Lamb dip pattern can also be observed by scanning the frequency of the laser using a ramp voltage on the end piezoceramic as shown in the lower part of fig. 7. Rackley and Butcher 25 have recently made use of radiofrequency Stark modulation of the sample gas to produce sidebands with a known r.f.separation. However, in the present example only a single sideband is generated, and hence the three sets of signals described above are easily identified. It would be much more difficult to inter- pret a complicated Lamb dip pattern of this type had the multiple-sideband method been used. CONCLUSION In this paper we have compared some of the methods used for studying short-lived molecular species, in order to show the relative merits of fixed-frequency laser spectro- scopy and of interferometry. The laser based spectrometers are capable of much higher resolution and of higher sensitivity, but possess a very limited tuning range. The interferometers can give an overall picture of the spectrum with good sensitivity, precision of measurement and with reasonable intensity information.It can also be seen that high-resolution Michelson interferometers are very competitive with diode laser spectrometers for studies in this region, particularly since the resolution achieved is comparable, and the region covered in a single experiment is an order of magnitude greater . A complete study of any one system would probably utilise all three spectroscopic approaches. We are indebted to the S.R.C. for the provision of much of the equipment, and for the support of two of us (M. L. L. and H. K.). We also wish to thank Dr. J. W. Brault and Mr. R. Hubbard of K.P.N.O. for their help in obtaining the Fourier transform spectra, and Mr. A. J. Fox of Phillips Research Laboratories, Redhill, for assistance with the A.O.M.We are also grateful to the Royal Society and the U.S. Army for travel support. We would also like to thank Mr. M. Taylor for his help with the 12C's02 Stark spectrum of H,CS, and Dr. A. R. W. McKellar for his unpublished data on the v4 band of CH,NH. J. W. C . Johns and W. B. Olson, J. Mol. Spectrosc., 1971, 39, 479. R. L. Sams and A. G. Maki, J. Mol. Struct., 1975, 26, 107. J. W. C . Johns, A. R. W. McKellar and M. Riggin, J. Chem. Phys., 1978, 68, 3957. C . Yamada and E. Hirota, J. Mol. Spectrosc., 1979, 74, 203. S. M. Freund, G. Duxbury, M. Romheld, J. T. Tiedje and T. Okay J. Mol. Spectrosc., 1974,52, 38.110 SPECTRA OF HZCS, D2CS AND CH2NH G . L. Caldow, G. Duxbury and L. A. Evans, J . Mol. Spectrosc., 1978,69,239. R. N. Dixon and M. Noble, Chern. Phys., 1980, SO, 331. B. J. Orr and T. Oka, Appl. Phys., 1980, 21, 293. ’ D. J. Bedwell and G . Duxbury, Chem. Phys., 1979, 37, 445. lo R. L, Shoemaker, S. Stenholm and R. G . Brewer, Phys. Reu. A , 1974, 10, 2037. l 1 J. W. Brault, Proceedings of the Workshop on Future Solar Optical Observations-Needs and l2 D. J. Bedwell and G . Duxbury, J . Mol. Spectrosc., 1980, 84, 531 ; M. Taylor, B.Sc. Thesis(Uni- l3 R. G. Brewer, Phys. Rev. Lett., 1970, 25, 1639. l4 M. E, Jacox and D. E. Milligan, J. Mol. Spectrosc., 1975, 58, 142. l5 J. W. C. Johns and A. R. W. McKellar, J . Chem. Phys., 1977, 66, 1217. l6 R. H. Judge, D. C. Moule and G. W. King, J . Mol. Spectrosc., 1980, 81, 37. l7 A. P. Cox and S. Hubbard, personal communication. B. Fabricant, D. Krieger and J. S . Muenter, J. Chem. Phys., 1977, 67, 1576. l9 J. W. C. Johns and A. R. W. McKellar, J . Mof. Spectrosc., 1973, 48, 354. 2o G. Duxbury, S. M. Freund and J. W. C. Johns, J . Mol. Spectrosc., 1976, 62, 99. 21 M. E. Jacox and D. E. Milligan, J. Mof. Spectrosc., 1975, 56, 333. 22 M. Allegrini, J. W. C. Johns and A. R. W. McKellar, J . Chem. Phys., 1979, 70, 2829. 23 W. H . Kirchoff, D. R. Johnson and F. J. Lovas, J . Phys. Chem. Ref. Data, 1973, 2, 1. 24 A. T. Mattick, A. Sanchez, N. A. Kurnit and A. Javan, Appl. Phys. Lett., 1973, 23, 675. 25 S. A. Rackley and R. J. Butcher, Mol. Phys., 1980, 39, 1265. Constraints, Florence, Italy, November 7-10, 1978. versity of Bristol, 1980).