New Developments in the Gas-phase Rotational Spectroscopy of High-temperature Species and Unstable Molecules BY MANFRED WINNEWISSER Physikalisch-Chemisches Institut, Justus-Liebig-Universitat Giessen, Heinrich-Buff-Ring 58, D-6300 Giessen, W. Germany Received 10th December, 1 980 The detection of gaseous molecules of unstable and high-temperature species in the interstellar environment has prompted the development in many laboratories of methods of detecting such spe- cies in hostile environments such as flames, plasmas and pyrolysis reaction processes. Our laboratory has participated in the development of a method for producing high-temperature species at ca. 500 K instead of the 1200 to 3000 K to which the solid alkaline earth oxides or sulphides must be heated in order to produce sufficient vapour pressure for millimetre-wave detection.The method is applicable to many molecular systems including KCN and probably LiCN and NaCN whose molecular dynamics is expected to be anomalous and is not yet understood. Included in the discussion is our work on the plasma formation of long-chain cyanoacetylenes and the pyrolysis production of N-cyanoformimine, the millimetre-wave spectrum of which we have studied. In the field of molecular dynamics our results for the v7 manifold of C30S will be presented and the pronounced change of the bending potential functions between C302 and C30S will be shown. 1. INTRODUCTION Flames, plasmas and pyrolysis are well-established chemical processes for oxida- tion reactions, the decomposition of compounds, the generation of heat and/or the formation of new species.Today there are huge industrial operations built around combustion, electrical discharges and pyrolysis or cracking processes. All these pro- cesses, not forgetting stellar chemistry, involve high-temperature phenomena and atomic and/or molecular high-temperature species. I wish to describe in this paper the contribution made primarily in our laboratory towards the development of milli- metre-wave rotational spectroscopy of high-temperature species, unstable species, especially those occurring in plasmas and pyrolysis experiments, and molecular dyna- mics. These heterogeneous-sounding topics are intimately linked, as I hope to show in the course of the discussion, by overlapping aspects of high-temperature gas-phase chemistry, interstellar chemistry and molecular dynamics, and by a growing need for extensive high-precision data concerning the types of molecular species to be discussed here.2. MILLIMETRE WAVES A N D HIGH-TEMPERATURE SPECIES Being associated with the radioastronomical aspects of microwave spectroscopy one learns to appreciate the astronomical fact that at the dawn of our Galaxy all of matter was in the form of hydrogen, some deuterium and possibly helium. Since then, some 10" years ago, 90% of the galaxies' mass (1.1 x 10" M o where 1 M o - 2 x g) has been converted into stars. The remaining 10% of its total mass is in the form of interstellar matter composed of the fossil material of the " big bang " and the exhaust materials returned by the stars in the various stages of their evolution.'32 HIGH-TEMPERATURE SPECIES AND UNSTABLE MOLECULES The distribution of interstellar matter within the galaxy thus follows that of the stars.This also means that most matter finds itself in a state of high temperature, despite the fact that most interstellar molecules so far observed are rather stable species, with the exception of a few radicals, molecular ions and unstable species such as HNC.' Stimulated by the consideration of the calcium-bearing molecules in stars and the interstellar medium, where an apparent depletion of atomic gas-phase calcium was observed [the solar abundance of calcium is n(Ca)/n(H) - 2.14 x 10-6,2 whereas in the interstellar medium it is 1.8 x 10-9-l.l x 10-9,3*4], G. Winnewisser suggested in 1975 the development in our laboratory of millimetre-wave spectroscopy of high- temperature species such as CaO.However, despite all the efforts which will be sketched in this paper, the radioastronomical search for CaO in various stars and molecular clouds and by various research groups was unsuccessful, thus strengthening the suspicion that interstellar Ca occurs in grains.' (1). DEVELOPMENT OF DATA BASE It was realized by that the classical microwave spectroscopy of high-tempera- ture gases as reviewed by Lovas and Lide9 and developed to a high level of sophistica- tion by several research groups runs into difficulties with the refractory oxides and sulphides of the alkaline earth metals Ba, Sr, Ca, Mg and Be. This fact can be appre- ciated by inspecting table 1.As one ascends the alkaline earth group, progressively higher temperatures are required to vaporize the solid oxide or sulphide. The in- creasing tendency towards decomposition proved insurmountable for conventional hot-cell experiments; the vapour above solid BaO at 2200 K consists largely of dia- tomic BaO, but the vapour above SrO at 2500 K is estimated to contain only 5% dia- tomic SrO.'O No data for CaO etc., are available. It was therefore necessary to cir- cumvent some of the high-temperature container problems by synthesizing the high- temperature species directly in the gaseous phase by non-equilibrium gas-phase reac- tions. It was furthermore realized that the high-resolution spectroscopic data base TABLE 1 .-SELECTED DIATOMIC OXIDES AND SULPHIDES PRODUCED BY OXIDATION OF METAL VAPOUR ~~ ~~ m.pt.of m.pt. of m.pt. of metal oxide su Ip hide 1st i.p. cosmic abundance 1°C 1°C 1°C /eV (H = 1, C = 3.3 x Mg 651 2800 d > 2000 7.64 2.6 x 10-5 Ca 843 2580 d - 6.11 2.0 x Sr 769 2430 > 2000 5.69 7.1 x 10-lo Ba 725 1920 1200 5.21 8.9 x A1 660 2045 1100 5.99 2.5 x Fe 1535 1420 1193 7.90 4.0 x 1 0 - 5 Li 181 > 1700 975 5.39 5.0 x (I Li abundance should be much greater in certain stars. for many " exotic " molecular species (closed and open shell molecules, radicals and ions) present in a high-temperature environment should be expanded and that the millimetre-wave region is particularly suitable for such experiments since quasioptical techniques can be employed for the propagation of the electromagnetic radiation.New classes of molecules could be investigated. This situation was recently emphas-M . WINNEWISSER 33 ized by the committee on High Temperature Science and Technology in a report on “ High Temperature Science : Future Needs and Anticipated Developments.” l1 Since the review article by Lovas and Lide9 considerable progress has been made in the field of rotational spectroscopy of high-temperature molecules whose outstand- ing developments were pointed out in 1978 by Torring and Tiemann.12 These authors also emphasized the importance of combining microwave techniques with the available tunable lasers in double-resonance experiments (MODR). (2). LOW-PRESSURE FLAME SPECTROMETERS The first observation of chemiluminescence spectra from the reactions of Ba and Ca with NO, and N20 were reported by Ottinger and Zare in 1970.13 The motivation was to learn about the unrelaxed internal-state distributions of products following a highly exothermic reaction.A plethora of papers has since been published on the spectral analysis of the chemiluminescence and the laser-induced chemiluminescence of metal-vapour plus oxidant reactions 4-24 and optical-optical double-resonance excitation spect r o ~ c o p y . ~ ~ - 27 In our laboratory6-s a modified version of the produc- tion method given by Broida and c o w o r k e r ~ ~ ~ ~ ~ ~ was adapted to the free-space cell type millimetre-wave spe~trometer.~**~~ In these methods diatomic metal oxides and sulphides are produced by gas-phase oxidation of metal vapour entrained in an inert carrier gas.In the most recent paper from this laboratory, the millimetre-wave spec- trum of Bas in a low-pressure flame was reported.8 The formation of gaseous Bas was performed in a flow system within a reaction sphere of 35 cm i.d. by the reaction Ba + COS -+ Bas + CO. A block diagram of the millimetre-wave spectrometer with reaction sphere and free- space absorption cell is reproduced in fig. 1. The low-pressure flame can spread throughout the entire volume of the sphere, allowing the oxidation reaction to take place under conditions minimizing the effects of wall collisions. The reaction zone was characterized by a diffuse but intense greenish-yellow chemiluminescence glow which filled the entire sphere. Therefore, the path of oxidation has been lengthened, producing an enhancement of the metal oxide and sulphide abundances and resulting therefore in a pronounced increase in the signal-to-noise ratio of the absorption lines.This can be appreciated by inspecting fig. 2, where in the upper part the video dis- play is given for the 138Ba160 millimetre-wave spectrum as obtained with our initial system7 which was built with a Pyrex cross-piece of 10 cm i.d. as the low-pressure flame compartment. The lower part of fig. 2 gives an oscilloscope display of a se- quence of rotational lines of 13*Ba3*S, obtained with the third to eighth harmonics of the klystron fundamental frequency, thus covering simultaneously portions of the millimetre-wave spectrum between 100 and 297 GHz. The Bas spectrum was re- corded with the new spectrometer shown in fig.1. From the various measurements carried out on the species BaO, SrO, CaO, BaS and lately SrS it is estimated that our present system is more sensitive by a factor of 50 than the system reported in ref. (7). This was achieved by improving our basic milli- metre-wave system 2 8 9 2 9 in three different respects: (a) Introduction of a reaction sphere with a diameter of 35 cm instead of the usual free-space cell [fig. 1, ref. (7)] for the low-pressure flame. (b) Introduction of a dynamic stabilization scheme for the coherent radiation source while maintaining a fast sweep capability. (c) Use of an InSb photoconducting detector operating at 1.7 K for detection of millinletre-wave power. It should be pointed out that our (1) The entire system is discussed in ref.(8).34 HIGH-TEMPERATURE SPECIES AND UNSTABLE MOLECULES combination of video-detection, dynamic frequency-drift stabilization, He-cooled InSb detector and signal averaging is particularly useful for searching in a dynamic chemical system which requires real-time monitoring of the spectrum while adjusting parameters. It is efficient when only low power levels are available, as is the case if crystal harmonic generators are employed as millimetre radiation sources. CELL FOR IUNING HOLECULT5 NEEDLE VALVES '\ L I n w w w n u 20mr POWER SUPPLY - - - r - y 2 F r < < e l ? s m o u ERROR VOLTAGE KLYSTRON SWEEP STABILIZER 5---7-iL II - 2 UARKER DRIFT u u - I. I SIGNAL - I 1L 3m, START OF MrA ALOUISIlIOII n- STA81112AllON POSITION k 1 - ERROR SIGNAI TRIGGER YbRKER ANG SECOND MARKER 1- TRIGGER PULSE FMI 1 I !5cm-mIGGERI I C W L .] [ C W M L l I L A W R A r O W PERIPUERAL INTERFACE MULTlPLEXER FOR INPUT CHANNELS M E AN0 HOLD AUFiIFIER ANALOG TO OlGlTAL WNVERlER REAL 1lUE CLOCK I[ PDPBII DEDICATED COMPUTER AND PERIPHERALS I1 1) ZSODATA POINTS IN CHANNEL 0 AND I 2 8 ~ 1 SAMPLING RATE SOFlWARE DATA ACOUISlrlONf ~* REOUCTION L !'CENTER FIG. 1 .-Block diagram of the millimetre-wave spectrometer with reaction sphere and free-space absorption cell for the study of low-pressure flames at the Justus Liebig University, Giessen [from ref. (S)]. The logic flow chart of the stabilization is given in fig. 1 together with the data- acquisition system. Fig. 3 presents the sequence of the stabilizer timing signals.The klystron frequency sweep stabilizer produces a correction voltage for each individual sweep so as to keep one point in the klystron frequency sweep fixed. Our results for BaO, SrO, CaO and Bas are collected in ref. (7) and (8). The constants listed there are expansion coefficients of the Dunham series.30 A few significant features of this type of rotational spectroscopy should be men- ti oned.M. WINNEWISSER 35 37396 37403 MHz lx 3 J = 10--9 8-7 6+5 13*Ba 32S I _ - . . ,*- J = 18-17 11-11 36-35 30-29 2!-23 18-17 FIG. 2.-Video display of absorption lines of 138Ba l6O and 138Ba32S. The 13'Ba160 lines were observed with our old system described in ref, (7), while the I3*Baj2S lines were obtained with the spectrometer shown in fig. 1 . (a) The relatively narrow linewidths of the observed millimetre-wave transitlons for the alkaline earth metal oxides and sulphides indicate low concentrations, which are probably offset by the very large dipole moments of these species.(b) From comparison of the relative intensities of the u = 0 and Y = 1 transitions shown in fig. 4 it can be shown that BaO, Bas and also the other oxides' are produced with a rather low vibrational temperature, a fact which was also observed in the chemi- SEOUENCE OF STABILIZER TIMING SIGNALS I I I n I I I I U - I <P I I I I n u ; I U I /I ADJUSTABLE GATE FOR SUPPRESSION OF SECOND MARKER VARIABLE GATE FOR STABILIZATION POSlllON (sp) RESPONSE CONTROL MARK@ SIGNAL EQROR SIGNAL PULSE iNIEGRATtD VOtlAGE (SAMPLE AND HOLD) BUILD-UP OF CORRECTION SIGNAL FINAL ERROR VOllAGE FOR INDlVlOUAl SWEEPS MARKER AFTER 5P MARKLR BEFORE SP FIG.3.--Sequence of stabilizer timing signals in the dynamic klystron stabilisation system indicated in fig. 1 [from ref. (8)].36 HIGH-TEMPERATURE SPECIES AND UNSTABLE MOLECULES 13' Ba'b J=8-7 4 . . . . . + . . .? ' . * . . . l3*B0 32S 148927.802 MHz V = I 149599.838(301 MHz v = o * . - . . . . . . . 15L096.577 v = 1 154570.323 MHz v = o T I I 191041.354 v = 1 I 191 626.900 MHz v = o FIG. 4.-Computer-averaged absorption lines of 138Ba160 and '''Ba3'S in the states u = 0, I and 2 obtained with the original flame arrangement [first row, ref. (7)] and with the improved spectrometer of fig. 1 (second to fourth rows).M . WINNEWISSER 37 I SUPERCOND. B W 0 - DETECTOR. - - BSORPTION CELL I R - I J 1 300 GHz BIAS PR E AM PL I F.1 I PLL If _I I- FREQUENCY t NORMAL LOCK IN MIXER 5 MHz luminescence Approximate relative intensity measurements on the observed millimetre-wave lines yield an effective vibrational temperature of 500 & 100 K for both BaO and Bas. This is to be contrasted with the temperatures needed for va- porising the oxides or sulphides (see table 1) and even the metals. (c) Rotational spectroscopy of low-pressure flames can be considered a real alter- native to hot-cell spectrometers which operate in the temperature ranges between 1300 and 2300 K. In such systems the high-temperature container problems pose severe technical difficulties. In addition to these problems a great reduction of line strength in the high-temperature regime is suffered as Torring and Tiemann12 have pointed out, since the peak absorption coefficient for a rotational transition of a diatomic molecule, neglecting all hyperfine interactions, is approximately given by = 1.5 x 10-9.~,ic1L21~03p73~v (2) where Av is the linewidth parameter and may be considered independent of tempera- ture T,fv is the fraction of molecules in the vibrational state 21, i, is the fractional iso- --- 30 GHz !I POLYETHYLENE LENGTH 1G3Gmm ~ --+ OVEN +-- FIG.5.-Block diagram of millimetre-wave spectrometer for the study of low-pressure flames at the Free University of Berlin and schematic drawing of the reaction vessel (courtesy of T. Torring). t DISPLAY BWO 1 1 - SIGNAL AVERAGER SATURATION I - EFFECT 11 I38 HIGH-TEMPERATURE SPECIES AND UNSTABLE MOLECULES topic abundance, p is the electric dipole moment in Debye, and vo is the frequency of the absorption line in MHz.In high-temperature systems the absorption path is usually considerably less than in room-temperature systems. Typical cell lengths are 100-20 cm for systems designed for 1000-2000 K. Therefore the peak absorption drops even further. In our flame experiment (fig. 1) the absorption path length is ca. 35-40 cm if we take the full dia- meter of the sphere. This also holds approximately for the millimetre-wave low-pres- sure flame spectrometer built by Torring and Hoeft and their respective coworkers, the block diagram of which is given in fig. 5 . We can conclude from the above dis- cussion that the peak absorption coefficient for molecules in the ground vibrational state is reduced approximately by a factor of 150 if the temperature is raised from 300 to 1300 K.For more complex molecules such as polyatomic linear, symmetric and asymmetric rotors the situation is even worse, owing to the large number of vibrational and rotational states. Therefore, the pivotal requirements, which this type of spectrometer does fulfil for many species, are high spectrometer sensitivity in the milli- metre- and submillimetre-wave region and an efficient production scheme at a mini- mum temperature for the molecules to be investigated. (3). HIGH-TEMPERATURE SPECTROMETER Besides the flame spectrometers discussed above and shown in fig. 1 and 5 Torring and c o - ~ o r k e r s ~ ~ ’ ~ ~ have developed the conventional hot-cell spectrometer to a high degree of sophistication. The temperature range has been expanded to close to 2300 K by a heated absorption cell with a current of Z 21 1000 A.In order to reduce pick- up from stray magnetic fields this current is rectified and the lock-in amplifier is gated in such a way that during the period of no current flow the phase-sensitive detection of the absorption signal takes place. An elegant modulation scheme which was introduced by T o r ~ - i n g ~ ~ makes use of the available high power of 20-100 mW of the backward wave oscillator millimetre- wave radiation sources now commercially available. This is indicated in the block diagram of fig. 6. Torring’s “ saturation-effect ” modulation can be employed when the incident power levels are sufficient to saturate a transition and the absorption becomes non-linear.This non-linearity allows the line signals resulting from the gas absorptions to be separated from the background. The sensitivity of such a spectro- meter employing saturation modulation compares favourably with the sensitivity of a Stark spectrometer. Fig. 7 shows chart-recorder traces for barium and strontium monofluorides obtained with the spectrometer system shown in fig. 6. This type of modulation can be used for any type of absorption cell since no Stark septum is re- quired. Both rectangular and cylindrical cells as shown in the lower part of fig. 6 have been used. The cells were directly heated by a current of a few hundred A. The Berlin group has successfully proved that in the temperature range 1000-2300 K not only the evaporation of highly refractory compounds can be achieved but also many high-temperature reactions can be exploited.In fact all high-temperature, open-shell molecules observed so far in the millimetre-wave region were first observed using this m e t h ~ d . ~ ~ ’ ~ ~ Subsequently, however, Ryzlewicz and Torring have found that both BaF and SrF can be observed more easily in a flame spectrometer. The hot-cell method retains the advantage of populating higher vibrational states sufficiently for their meas~rement.~~M. WINNEWISSER 39 I KLYSTRON OPEN MOUNT -100GHz ABSORPTION CEL Go As ' I 1 DETECTOR , - _ - _ - ~7 MICROWAVE TRANSITION I AND ELECTRICAL GUIOE I s 1000A LOCK IN 50 kHz SATU RAT I0 N I EFFECT , MODULATION FIG.6.-Block diagram of high-temperature spectrometer with saturation modulation at the Free University of Berlin, and various absorption cell cross-sections and reaction arrangements (a, b and c) that have been tried (courtesy of T. Torring).40 HIGH-TEMPERATURE SPECIES AND UNSTABLE MOLECULES 88Sr "F N =6+5 d L=O R C = 3 s , single sweep i n 50slMHz i I 138g,19~ N = 7-6 I 1 v = o I \ - 1MHz 0 -6 FIG. 7.-Absorption lines of s8Sr'9F and 138Ba19F observed with the spectrometer in fig. 6 (courtesy of T. Torring). (4). COMMENTS ON THE SYNTHESIS OF HIGH-TEMPERATURE SPECIES Torring and TiemannI2 have recently summarized the various methods of produc- ing high-temperature molecules, and the reader is referred to their work. However, a few comments should be made. In standard hot-cell experiments the following types of reaction have been employed for obtaining sufficient vapour pressure (0.01 to 0.1 Torr): (a) direct evaporation : T W s , 1) + AB(g); (3) (b) reduction reaction and evaporation : T AB,(s, 1) + A(s, 1) -+ AB(g) + - - - ; (4) (c) reduction reaction and evaporation when A(s, 1) is the most volatile partner: T AB,(s, 1) + C -+ AB + BC + .. . . Ryzlewicz and TOrring3, have successfully employed the third type of reaction in pro- ducing a sufficient concentration of the 2E-state radicals BaF and SrF by using the react ions T BaF, + A1 -+ BaF + A1F + Ba + . . . , T = 1450 K and T SrF, + A1 --f SrF + AlF + Sr + . . . , T = 1750 K. (7) No lines were found with arrangement (a) in fig. 6. However, with arrangement (b), where the cell walls were coated with A1 as a reducing agent, BaF lines were obtainedM. WINNEWISSER 41 with a signal-to-noise ratio of ca.50 in the ground vibrational state. The BaF, was evaporated directly into the cell from a separate sample boat which is part of the hot cell, so that no temperature gradient between cell and boat should be present. Thus, a non-equilibrium process cannot be excluded. The gas-phase oxidation reactions for producing the alkaline earth oxides and sul- phides have already been discussed in detail. The reaction takes place far from equilibrium conditions in the gaseous phase and is exothermic. Ryzlewicz and Tor- ring32 have employed this type of reaction in their low-pressure flame spectrometer (see fig. 5), Ba + IF6 --f BaF + .. . . (8) They mention that the IF6 concentration is critical since BaF + IF6 -+ BaF, can react further, and under stable conditions a signal-to-noise ratio of 50-100 was obtained for the rotational transitions in the ground state. Our experience agrees with that of Ryzlewicz and Torring that in flame experiments it is much easier to optimize the conditions of the flame reactions, especially since chemiluminescence can be used to monitor the reaction. Our laboratory has used flame reactions to explore a variety of chemical systems, including some which are potentially capable of producing polyatomic high-temperature species. The reaction K(g) -t BrCN 3 KBr + KCN (9) produces two gas-phase constituents, the spectra of both of which could be monitored.The KCN lines of Kuijpers et An attempt to find the polyatomic species CH3Li has failed so far. Only the lines of LiI as seen in fig. 8 and could be seen by video. relative frequency /MHz FIG. 8.-Absorption lines of the J = 5 t 4 transition of 'LiI2'I at 13 1 01 1.85 MHz, LiI observed by video in the reaction CH31 + 2Li (vapour) +LiI + CH3Li. Lower half shows calculated intensity and positions of quadrupole hyperfine components. solid (CH,Li), on the walls could be found. However, it is hoped that with the im- proved system described in this paper the detection of the gas-phase spectrum of CH3Li has become more feasible. As indicated by the above example, the spectroscopy of polyatomic high-tempera- ture species is considerably more difficult. In spite of that, progress has been made in the study of the alkali h y d r o ~ i d e s ~ ~ * ~ ~ with the investigation of the effects of large- amplitude motion.Perhaps the most interesting triatomic species so far studied in detail is KCN. Its rotational spectrum and structure have recently been reported by42 HIGH-TEMPERATURE SPECIES A N D UNSTABLE MOLECULES Torring et u Z . ~ ~ The molecule was found to have a non-linear, T-shaped molecular structure, N with a low-lying bending fundamental w2 157 cm-'. The authors combined mole- cular-beam electric-resonance and microwave data with the earlier millimetre-wave data23 and assigned 63 transitions to the ground vibrational state. The fit was made to a semi-rigid asymmetric rotor model with A = 58 265.62(7) MHz, B = 4 940.064(2) MHz, C = 4 536.208(2) MHz, five quartic centrifugal distortion constants and three sextic distortion coefficients. This result is undoubtedly one of the most interesting ones reported not only in the realm of high-temperature gas-phase spectroscopy but also for triatomic molecules in general.From the comparative ease with which we could obtain KCN lines with our low- pressure flame spectrometer I strongly believe that a wide variety of presently unknown high-temperature species such as LiCN, CH3Li and high-temperature radicals and ions can be investigated. Work on LiCN should be particularly rewarding since ab initio calculation^^^ indicate that the Li+ ion may move freely around the (CN)' anion at room temperature. All of the gas-phase species LiCN, NaCN, KCN, etc., must be characterized by large-amplitude motion of the lowest lying bending mode and should be further investigated by this type of method.3 . MILLIMETRE WAVES, PLASMA CHEMISTRY AND PYROLYSIS (1). INTERSTELLAR PLASMA CHEMISTRY In interstellar space hydrogen cyanide, HCN, and cyanoacetylene, HC3N, are distributed throughout a wide variety of different molecular clouds.' Winnewisser and W a l m ~ l e y ~ ~ have pointed out that the distribution of these molecules and the oc- currence of the cyanopoly-ynes (HC,N, HC7N and HC9N) in sources across the galaxy underline the trend that complex organic molecules are not limited in their occur- rence to the galactic-centre sources only, but appear in sources with very different physical conditions. These molecules are relatively abundant in cold (10 K) dark mole- cular clouds with densities in the order of lo4 ~ m - ~ , in some molecular clouds associ- ated with H I1 regions and in the molecule-rich circumstellar shells of late-type stars.As early as 1973 Buhl and Snyder4' and later Morris et al.41 noticed that the inter- stellar distribution of HNC and HCO+ is particularly highly correlated with that of HCN. Creswell et al.42 in their work on the millimetre-wave spectrum of HNC have summarized the interstellar emission features of HNC, HCN and HCO+. In fig. 9 the interstellar emission signals of OMC-2 as observed by Morris et aZ.41 are shown in the upper part for HCN and HNC. The appropriate laboratory measurements are shown in the lower trace with a summary of the relevant experimental data.It should be noted that in interstellar space the densities of HCN and HNC are roughly the same, while under laboratory conditions measurable HNC concentrations can only be produced by thermal isomerization at 1300 K of HCN43 or by the use of either d . ~ . ~ ~ or radiofrequency discharges with a variety of H, N and C containing molecule^.^^*^^*^^ In interstellar space the Taurus Molecular Cloud 1 (TMC 1) is the only source where HC3N,47 HC5N,48 HC7N49 and HC,N5' have been discovered to be present simul- taneously with HCN and CN.39 As can be seen from the comparision of the high-re- solution spectra (see fig. 10) taken with the Effelsberg radiotelescopeS1 the molecular concentrations drop only slowly as one goes up the series, which indicates that theM .WINNEWISSER 43 chemistry in this source must be special with regard to the production of long-chain molecules. Churchwell et aZ.47 suggested that one single reaction involving C2Hz+ or C2H3+ may transform the lower cyanoacetylenes into HC5N, HC7N and HC,N. Mitchell et aZ.52 showed that the observed abundances of cyanopoly-ynes can be ob- HCN HNC 14 a ysR/km s-' A LABORATORY 1 - 1 - 0 0-1 188633.954 MHz 2-1 188 631.871 MHz 1 1-1 I 88 630.431 MHz Sampling rate: 0 Sweep duration: 20 ms Sweep rate : 15.5 Hz Sweeps accumulated : 1 Pressure : (static) Ton 14 a v , / km 5-1 .n. . . . . . . '4 SMOOTHING I I I I I I I -1.0 *l.OMHz 90 663.602(50)MHz 28 .us 15 ms 15.5 Hz 2000 (flow) 3.5 x lo-' Torr (total) N4 + CHjI-tHNC + . . . F~G. 9.-Interstellar emission (upper) and laboratory absorption (lower) lines observed for HCN and HNC, showing relative intensity which reflects relative abundance.Data-acquisition parameters are given for laboratory lines [interstellar lines from ref. (41)]. tained by the gas-phase reaction pathways of H2CN+ with C2H2. Another possibi- lity, discussed and pursued by us in the l a b ~ r a t o r y , ~ ~ is that polyacetylene chains54 are built up first and then react with HCN or H,CN+ to give the cyanopoly-ynes. (2). PLASMA DIAGNOSTICS In order to shed more light upon these problems our laboratory started to investi- gate various gas-phase plasmas composed of hydrocarbons and nitrogen-containing molecules. In the production of HNC it was noticed that besides HNC and HCN large amounts of HCC-CN and other, not yet identified species were produced.The formation of cyanoacetylene by the reaction r.f. discharge HCCH + HCN - HCC-CN + 2H' (10)44 HIGH-TEMPERATURE SPECIES AND UNSTABLE MOLECIJLES I was then used to produce enriched HCC-l3CNS5 in quantities adequate for rotational spectroscopy, as can be seen in fig. 1 1 . HC5N can be synthesized by a continuous or flash discharge in a 2: 1 mixture of HCCH + HCN as Winnewisser et have shown. Our results are summarized in fig. 12. Under continuous-flow conditions at a pressure of 10-2-10-1 Torr the 2: 1 mixture of HCCH + HCN is passed through a 25 cm long r.f. discharge tube powered by a 300-1000 W, 4.5 MHz, r.f. generator. HC3N is produced in preparative quantities and HC5N in The gas-phase production of higher cyanopoly-ynes is also possible.I I RESOLUTION +I+- 0.027 krn 5”’ TA T HC3 N I = l-O,F=2-1 / \ HCs N /=9---0 +k- 0.041 krn s-’ Ti +I+- HC7 N 0.042 krn 5’ I 21-20 I I I I I I I 4.0 5.0 6.0 7.0 U ‘ S R h s-’ FIG. 10.-Interstellar emission lines of HC3N, HCsN and HC7N from TMC 1, u = 04 h 38 m 38 s, 6 = 25” 36’ 00”, indicating relative intensity and thus abundance, since transition moments are similar [from ref. (51)].M . WINNEWISSER 45 J=20-19 SAMPLE HC 13CCN HCC 13CN CHEM I CAI LY NATURAL ABUNDANCE PREPARE 0 1.1% l3C DISCHARGE GENERATE0 H CC H* H13C N I , I , I 1 181 170 181 190 frequency/MHz FIG. 11 .-Laboratory absorption lines of cyanoacetylene prepared chemically (upper trace) and pre- pared in a discharge (lower trace) of acetylene with 13C enriched HCN [from ref.(55)l. DIRECT OBSERVATION (FREE-SPACE CELL) HCC-CCH} . . . . HC, N 1 . . . . HC5 N . . . . ? H ~ N 0-1 I N DI R E C T 0 BS E RVAT I 0 N (PRODUCTS TRAPPED AND FRACTIONALLY DlSTl l l E O 1 24-23 16-15 i 1x6 1x5 1 x 4 1x3 HARMONIC I I 36.387 36.391 GHz 42-41 I 1x3 HARMONIC I 37.274 G H t ROTATIONAL -TRANSITIONS J.1- J 65-64 39-38 I I 78-77 1 55154 1 26j25 I c . . . . ? 1 x 6 1.5 1 x 4 1x3 1 x 2 HARMONIC I I 34.605 34.613 GHz FJG. 12.-Detection of HCSN in the products of a discharge of HCN with acetylene (lower trace) and with diacetylene (upper pair of traces). In the former case HCSN was concentrated by fractional dis- tillation to obtain an observable amount, while in the latter case it could be seen video jn the dis- charge products directly.The HCsN line has roughly the same intensity as the HC3N line; owing to the large number of states which are populated in HCsN this means a larger concentration of HCsN, supporting the theory relying on p~lyacetylenes.~~46 HIGH-TEMPERATURE SPECIES AND UNSTABLE MOLECULES TABLE 2.-ROTATIONAL FREQUENCIES OBSERVED IN THE LABORATORY TOGETHER WITH CAL- CULATED FREQUENCIES AND CONSTANTS OBTAINED FOR HCSN IN THE GROUND VIBRATIONAL STATE ~~ ~ ~~ ~ J" v(o bs)/MHz v(calc)/MHz v(0bs) - v(calc)/kHz 0" 1" 3" 6 7" 8" 9 10 11 12 13 14 18 25 30" 32 " 36" 38 39 43 51 54 64 65 77 10 650.657 18 638.615 21 301.262' 23 963.8968 26 426.543' 29 289.152 " 31 951.772' 34 614.387 " 37 276.994" 39 939.591 50 589.81 5 " 69 227.1824 103 836.8169 " 106 498.9104" 117 147.0163 " 138 441.6788 146 426.5550" 173 040.1865 " 175 701.2968 " 207 630.7692 2 662.6653 5 325.3299 10 650.6540 18 638.6167 21 301.2618 23 963.9011 26 626.5339 29 289.1595 31 951.7771 34 614.3860 37 276.9856 39 939.5751 50 589.8173 69 227.1 849 82 539.0412 87 863.6321 98 512.5219 103 836.8093 106 498.91 11 117 147.0221 138 441.6722 146 426.5659 173 040.1862 175 701.3019 207 630.7643 3 .O - 1.7 0.2 -4.3 9.1 - 7.5 -5.1 1 .o 8.4 15.9 - 2.3 - 2.5 7.6 - 0.7 - 5.8 6.6 - 10.9 0.3 -5.1 4.9 Bo = 1 331.332 714(46) MHz Do = 30.101 6(58) HZ " Lines observed in interstellar sources;' ref. (53); ref.(56); ref. (47); ref. (75). small amounts (ca. 10 mmol). CH,CN and H2C2HCN were not detected. HCSN was isolated by trapping the discharge products and then using a low-temperature distillation for separation.HC7N and higher poly-ynes have not yet been detected in these mixtures. In the lower part of fig. 12 the results of the flash discharge reaction is shown using diacetylene, HCC-CCH, and HCN as starting materials. HC5N is produced imme- diately with approximately the same yield as HC3N. The duration of the single dis- charge flash is 0.5-1 s. In table 2 our measurements of HCSN in the millimetre-wave region are combined with earlier published results making precision line positions available throughout the millimetre-wavelength range for HC5N.s6 HCC-CN, in particular, seems to be a chemical sink. We have observed it in the reactions just discussed, in all reactions producing HNC and in the following discharge reactions: We are no longer surprised at finding the longer cyanoacetylenes. r.f.discharge A HCC-CN + 2H' \ /" (4 // c = c H \ H' CN (b) HCCH + HCNO -H- HCC-CNO + 2H' (12) r.f. discharge HCC-CN + H20.M . WINNEWISSER 47 Apparently, the discharge of any combination of hydrocarbons and a source of nitro- gen will generate HCN and HCC-CN. The latter experiment (b) brought another aspect of the interstellar chemistry to light; in the reducing atmosphere of a hydrogen-containing cell (or cloud), hydrogen atoms will scavenge the oxygen atoms from such molecules as HCNO. The milli- metre-wave H,O line at 183 GHz appears immediately when the discharge is turned on in the mixture of acetylene and fulminic acid, HCNO. Therefore HCC-CNO was not found. It has long been noted that NNO is absent from the list of interstellar molecules.The NO bond in N,O, like that in HCNO, is not stable in such an atmo- sphere. A search was also initiated to obtain the millimetre-wave spectrum of isocyano- acetylene, HC=C--NC, since equal interstellar concentrations of HCN and HNC (see fig. 9) strongly indicate that isocyanopoly-ynes might be constituents of the inter- stellar gas, particularly in the molecular envelope of IRC-10216 or similar carbon stars.57 Therefore, experiments are under way in our laboratory to synthesize and detect this species for example uia the reactim H H \ / r.f. discharge + HC EE C - NC + 2H- (13) NC H F="\ in analogy to the formation of HC3N discussed above. (3). PYROLYSIS EXPERIMENTS The stability of HCN, HCC-CN and vinyl cyanide, H,CCHCN, in the inter- stellar atmosphere and in r.f.discharges leads us to believe that N-cyanoformimine H,CNCN, as an addition product of HCN, might be a species responsible for some TABLE 3.-ROTATIONAL AND CENTRIFUGAL DISTORTION CONSTANTS IN THE S-REDUCED ROTA- TIONAL HAMILTONIAN CHLNCN" CHzCHCN 63 373.235(58) 5 449.339 62(110) 5 009.563 26(98) 0.166 872 7 2.386 O(22) - 123.970(19) 5870( 17) -0.480 69(88) -0.032 14(30) -0.003 4(22) - 0.441 (25) -27.79(22) 7470( 1240) 0.001 45(70) 0.000 43( 19) - 0.000 54(46) -0.010 74(71) 119 0.054 49 850.6982(87) 4 971.164 06(70) 4 513.877 20(76) 0.161 219 6 2.182 l(17) - 85.076( 18) 2 717.1(11) -0.457 14(58) -0.030 53(20) - 0.343(26) - 7.09( 12) 0.004 4( 13) 433( 37) 0.002 91(71) 0.000 15(48) -0.000 12(38) - 129 0.044 ~~ " Ref.(61): * Ref. (64).P 00 I I I I I I I 9 7 1 8 ' , ' I 1 I A A FIG. 13.-Video mosaic of the J = 10 t- 9 a-type R-branch transitions of N-cyanoformimine, H,CNCN, and stick diagram of ground-state lines. Fortrat diagram follows the band head, where K represents K,. Note quadrupole hyperfine splitting of K, = 8 and 9 lines.61 r m mM. WINNEWISSER 49 of the still unassigned lines observed in the discharge experiments yielding HNC, HC3N, etc. The addition product N-cyanoformimine was first produced by Wentrup'' in 1971 and identified by mass spectrometry of the products of the pyrolysis of trimethylene- tetra-azole at 800 K and a pressure of 0.01 Torr (1.3 Pa). The first microwave detec- tion was carried out by Bak et aLS9 in 1978 using a different pyrolysis precursor. Bak and Svanholt6' realized that the interstellar emission line at 10 458.634 MHz, identi- fied as the J = 18 t 17 transition of HC9N," coincides closely with the lo, t Ooo transition of H,C=NCN at 10 458.49 MHz.We have extended the rotational tran- sition measurements into the millimetre-wave region where lines have been assigned in both the a-type and b-type spectrum, thus providing the data base for identifying the molecule in plasmas and in interstellar clouds.61 Fig. 13 shows the J = 10 t 9 rotational transitions with the characteristic K, band-head structure of the a-type R- branch spectrum.62 In table 3 the preliminary rotational and centrifugal distortion constants in the S-reduced rotational H a m i l t ~ n i a n ~ ~ are presented together with the corresponding constants obtained from the available data on vinyl cyanide.64 A re- fined rotational analysis must await completion of the analysis of the quadrupole N - CYANOFORMIM INE VINYLCYANIDE - + / d- / I I 1 I ' / i - + I I / I L, - + 2639 M H z 1320 MHz LLO MHz L572 MHz 27LL MHz 1372 MHz L57 MHz L l / / L - - - * FIG.14.-Rotational-energy-level diagram for H,CNCN and HzCCHCN showing lowest-lying states and probable interstellar transitions. The 211 -+ 212 and 10o,lo+- 90,9 (not shown) transitions of HzCCHCN have been observed in Sgr B2.'HIGH-TEMPERATURE SPECIES AND UNSTABLE MOLECULES HNCS LINEAR I 0 / I / / 3 l - 7 3 3 = r I I I , I + ' i ; I I I I I I I I I , I I I 1 I I I. / NEARLY BENT I I I I I I f I t I I I I I I I I I I I 4-?: I 1 f / X = O j ; 3 4 s I I I I I \ I \ I \ I I I I I ' < , ' I I 1 1 2 3 4 q / {'' hyperfine structure in the spectrum, in which we are now engaged.Fig. 14 shows a part of the rotational energy level scheme of N-cyanoformimine and vinyl cyanide showing the possible transitions opportune for a radioastronomical search. 4. MOLECULAR DYNAMICS The polytopic molecules KCN, NaCN and LiCN mentioned in Section 2 are examples of a special group of molecules from the point of view of molecular dynamics and structure. Because of a very low-lying, large-amplitude bending mode, theM . WINNEWISSER 51 analytical form of the vibrational potential surface is not self-evident, but is necessary for the understanding of their spectra and structure. Some aspects of these problems are touched on in the following brief discussion of the effects of quasilinearity on mole- cular energy levels.(I). POTENTIAL FUNCTION AND CORRELATION PARAMETER yo Quasilinearity is just one of the forms of large-amplitude motion which complicate molecular structure determination, and even structure definition. Yamada and Win- n e w i s ~ e r ~ ~ studied the correlation of the rotation-bending energy levels of linear and bent triatomic or longer chain-type and introduced a correlation para- meter yo to give a quantitative, empirical measure of quasilinearity, (14) E(1owest excited state with K or I = 0) E (lowest state with K or I = 1) p - y - 4 - ( where E represents the energy of a given state measured from the ground state. This parameter has the value of - 1 for an ideal linear molecule and a value of + 1 for an ideal bent molecule.Tt traces in fig. 15 from left to right the transition of a vibrational degree of freedom into a rotational degree of freedom. Although the molecules entered in fig. 15 are a real job-lot of chemical composition and properties they have the common physical property of one low-lying, large-amplitude bending mode. As 300 - 200 - d I E Y 4 100 - v7 1, Elcrn-' 5 1 191.1100 5 3 181.10" 5 5 164.55" 4 0 144.2985 4 2 136.9487 4 4 120.52' 3 1 97.1401 3 3 79.9312 0 60.7022 2 45.7977 1 18.1795 0 - P had FIG. 16.-Energy-level manifold of the v7 bending mode of GO2. The 1, 1 level was determined in ref. (67), the levels denoted by ' are from ref. (69) and the remaining levels are from ref. (68). can be seen from fig.15 carbon suboxide, C302, is the most quasilinear molecule with a low-lying bending mode of 18.1795 cm-' which has recently been observed in the sub-millimetre-wave region by Krupnov and The vibrational manifold of the quasilinear bending mode v, is reproduced in fig. 16 together with the energy levels as determined from sub-rnillimetre-wa~e,~~ infrared6* and Raman spectro~copy.~~52 HIGH-TEMPERATURE SPECIES AND UNSTABLE MOLECULES ---r- (2). V7 VIBRATIONAL MANIFOLD OF C3OS In our early work7* on the pure rotational spectrum of tricarbon oxide sulphide, C,OS, we thought we would be extending the study of molecules exhibiting quasilinea- rity. However, it was determined at a very early stage that C,OS showed no quasi- linear behaviour. Tn order to explore the vibrational manifold of v7, the lowest-lying bending mode, which is well-separated from the other two bending 71 Winnewisser et aZ.72 have carried out precision relative intensity measurements of the vibrational satellite lines in the microwave region from 8 to 40 GHz using a Hewlett- Packard bridge-type spectrometer.The results of our studies are summarized in fig. - -- -200 I "7 - 100 e 561,9 496,4 4 89 , 5;: 485,l 483,7 416 ,4 410,l 406,9 338,O 332,9 331,4 251,2 248 ,O 167,4 165,8 82,9 0 L - 600 - r- 400 Ho I V ( d / v; I ;f- 7ef i- J-"f I I 1 I I 3 1 I I I -3kf I I : ' i--l;f FIG. 17.-Energy-level manifold of the v7 bending mode of C30S, determined from relative intensity measurements in the pure rotation spectrum. The energy level 6$ indicated by * was derived via the /-resonance effect.12M .WINNEWISSER 53 17, showing a vibrational manifold belonging to a two-dimensional isotropic oscillator with only slight anharmonicity of the potential function, but without any flattening or barrier in the bottom of the potential. This is in sharp contrast to the C302 potential function and v7 manifold reproduced in fig. 16. It must be concluded that the substitution of one oxygen by sulphur in C,O, changes the delicate balance of the electron density on the central carbon atom notice- ably in such a way that a stiffening of the bonds occurs, despite the fact that the bend- ing force-constant corresponding to the central bend is still surprising small,71 indicat- ing a large-amplitude motion but one for which the Born-Oppenheimer separation of rotation and bending is still valid.(3). EXTENSION OF THE CORRELATION PARAMETER y o TO THE FREE INTERNAL ROTOR In a recent contribution to the investigation of quasilinearity Bunker and Howe 73 have extended the correlation parameter of eqn (14) to (15) le, 1) - E(OO, 0) + E(OO, 1) - E(0 ) E(2O, 0) - E(OO, 0) O , O ) , (E(l y o = l - 4 in order to include the free internal-rotor limit. The notation is that for the energy levels of a linear molecule, E(d, J ) . Free internal rotation in a triatomic molecule means one atom being free to move 360” around the line joining the other two more rigidly bound atoms from one linear configuration to the other. van der Waals mole- cules, for example, are in this category. The ideal limiting values of - 1 and + 1 for linear and bent molecules are unaffected but a new range of values is covered, namely between either +1 or -1 and -3, the value assumed by a free internal rotor.The general definition put forward by Bunker and Howe is Y u = 1 - 4 b ” + W C ” l (16) where a, is equal to the energy difference between the (3v + 1)th rotation-bending level having J = 1 and the uth rotation-bending level having J = 0, bu corresponds to the energy difference between the (3u + 2)th rotation-bending level having J = 1 and the 0th rotation-bending level having J = 0, and c, is equal to the difference between the ( u + 1)th and uth rotation-bending energy levels having J = 0. From the above definition it follows that a series of parameters y, with v = 1, 2, 3, .. . can be obtained as a function of the degree of bending excitation. Furthermore the y, parameter will vary with the degree of excitation of the stretching vibrations. The polytopic molecules LiCN and KCN were used as model systems by Bunker and Howe7j and it will be interesting to see, when their spectra are obtained and analysed, which y, values and thus which potential functions will describe the dyna- mics of these intriguing high-temperature species. I express my gratitude to Prof. T. Torring for making available preprints and re- prints of the recent work carried out in the Berlin High Temperature Microwave Laboratory. I also would like to thank him very much for providing some figures for this paper, I am grateful to Dr. Brenda P. Winnewisser for numerous discussions and comments on this paper and to Dr.Gisbert Winnewisser for permission to use some results of forthcoming papers. The experimental work which was carried out in our laboratory was in part sup- ported by funds from the Deutsche Forschungsgemeinschaft, the Max-Planck-Institut fur Radioastronomie in Bonn and the Fonds der Chemischen Industrie.54 HIGH - TEMPE RAT U R E S P E C I ES A N D UNSTABLE M.OLECUEES G. Winnewisser, E. Churchwell and C. M. Walmsley, Modern Aspects of Microwave Spectro- scopy, ed. G. W. Chantry (Academic Press, London, 1979), chap. 6, p. 314. G. L. Withbroe, The Menzel Symposium, NBS Spec. Pub. No. 353, ed. K. G. Gebbie (U.S. Government Printing Office, Washington, D.C., 1971), p. 127. L. M. Hobbs, Astrophys. J . , 1974, 191, 381.E. B. Jenkins and B. D. Savage, Astrophys. J . , 1974, 187, 243. W. H. Hocking, G. Winnewisser, E. Churchwell and J. Percival, Astron. Astrophys., 1979, 75, 268. R. A. Creswell, W. H. Hocking and E. F. Pearson, Chem. Phys. Lett., 1977, 48, 369. W. H. Hocking, E. F. Pearson, R. A. Creswell and G. Winnewisser, J. Chem. Phys., 1978, 68, 1128. D. A. Helms, M. Winnewisser, and G. Winnewisser, J. Phys. Chem., 1980, 84, 1758. demic Press, New York, 1971), vol. 3, p. 177. High Temperature Science: Future Needs and Anticipated Deuelopments, Report by the Com- mittee on High Temperature Science and Technology, Assembly of Mathematical and Physical Sciences, National Research Council, National Academy of Sciences, Washington D.C., 1979. T. Torring and E. Tiemann, Rotational Spectroscopy of High Temperature Molecules, in NBS Spec.Pub. No. 561. Proceedings of the 10th Materials Research Symposium on Characteriza- tion of High Temperature Vapors and Gases held at NBS, Gaithersburg, Maryland, Sept. 18- 22, 1978, issued Oct. 1979. ‘ F. L. Lovas and D. R. Lide Jr, Aduances in High Temperature Chemistry, ed. L. Eyring (Aca- lo M. Kaufman, L. Wharton and W. Klemperer, J . Chem. Phys., 1965, 43, 943. l3 Ch. Ottinger and R. N. Zare, Chem. Phys. Lett., 1970, 5, 243. l4 C. D. Jonah, R. N. Zare and Ch. Ottinger, J . Chem. Phys., 1972, 56, 263. l5 F. Engelke, R. K. Sander and R. N. Zare, J. Chem. Phys., 1976, 65, 1146. l6 R. N. Zare and P. J. Dagdigian, Science, 1974, 185, 739. l7 D. J. Wren and M. Menzinger, J . Chem. Phys., 1975, 63, 4557.l9 C. R. Jones and H. P. Broida, J . Chem. Phys., 1974, 60, 4369. ’O J. B. West and H. P. Broida, J . Chem. Phys., 1975, 62, 2566. *l G. A. Capelle, C. T. Jones, J. Zorskie and H. P. Broida, J. Chem. Phys., 1974, 61, 4777. 22 G. A. Capelle, H. P. Broida and R. W. Field, J. Chem. Phys., 1975, 62, 3131. 23 R. W. Field, in Molecular Spectroscopy, Modern Research (ed. K. N. Rao) (Academic Press, 24 J, B. West, R. S. Bradford, J. D. Eversole and C . R. Jones, Reu. Sci. Instr., 1975, 46, 164. 25 R. A. Gottscho, P. S. Wein and R. W. Field, J. Mol. Spectrosc., 1980, 82, 283. 26 R. A. Gottscho, J. B. Koffend and R. W. Field, J . Mol. Spectrosc., 1980, 82, 310. ” P. F. Bernath and R. W. Field, J. Mol. Spectrosc., 1980, 82, 339. 28 M. Winnewisser, 2. Angew. Phys., 1971, 30, 359.29 M. Winnewisser and B. P. Winnewisser, 2. Naturforsch., Teil A , 1974, 29, 633. 30 J. L. Dunham, Phys. Rev., 1932, 41, 721. 31 E. Tiemann, M. Bojaschweski, Ch. Sauter-Servaes and T. Torring, 2. Naturforsch., Teil A, 1974, 32 Ch. Ryzlewicz and T. Torring, Chem. Phys., in press. 33 T. Torring, J. Mol. Spectrusc., 1973, 48, 148. 34 P. Kuijpers, T. Torring and A. Dymanus, Chem. Phys. Lett., 1976, 42, 423. 35 P. Kuijpers, T. Torring and A. Dymanus, 2. Naturforsch., Teif A , 1975,30, 1256; 1977,32,930. E. F. Pearson, B. P. Winnewisser and M. B. Trueblood, 2. Naturforsch., Teil A, 1976,31, 1259. 37 T. Torring, J. P. Bekooy, W. L. Meerts, J. Hoeft, E. Tiemann and A. Dymanus, J. Chem. Phys., 1980,73,4875. 38 E. Clementi, H. Kistenmacher and H. Popkie, J .Chem. Phys., 1973, 58, 2460. 39 G. Winnewisser and C. M. Walmsley, Astrophys. Space Sci., 1979, 65, 83. 40 D. Buhl and L. E. Snyder, Astrophys. J., 1973, 180, 791. 41 M. Morris, B. Zuckermann, B. E. Turner and P. Palmer, Astrophys. J., 1974, 192, L 27. 42 R. A. Creswell, E. F. Pearson, M. Winnewisser and G. Winnewisser, 2. Naturforsch., Teil A , 43 G. L. Blackmann, R. D. Brown, P. D. Godfrey and H. I. Gunn, Nature (London), 1976, 261, 44 R. J. Saykally, P. G. Szanto, T. G. Anderson and R. C. Woods, Astrophys. J., 1976, 204, 143. 45 E. F. Pearson, R. A. Creswell, M. Winnewisser and G. Winnewisser, 2. Naturforsch., Teil A , 46 C . A. Arrington and E. A. Ogryzlo, J. Chem. Phys., 1975, 63, 3670. 47 E. Churchwell, G. Winnewisser and C. M. Walmsley, Astron. Astrophys., 1978, 67, 139. D. J. Benard, W. D. Slofer and J. Hecht, J . Chem. Phys., 1977, 66, 1012. New York, 1976), vol. 11, chap. 5.3. 29, 1692. 1976, 31, 221. 395. 1976, 31, 1394.M. WINNEWISSER 55 48 J. M. MacLeod, L. W. Avery and N. W I Broten, Astrophys. J., 1979, 233, 584. 49 H. W. Kroto, C. Kirby, D. R. M. Walton, L. W. Avery, N. W. Broten, J. M. MacLeod and T. Oka, Astrophys. J . , 1978, 219, L133. N. W. Broten, T. Oka, L. W. Avery, J. M. MacLeod and H. W. Kroto, Astrophys. J., 1978,223, L105. F. Tolle, H. Ungerechts, C . M. Walmsley, G. Winnewisser and E. Churchwell, Astron. Astro- phys., 1981, in press. 52 G. F. Mitchell, W. T. Huntress Jr and S . S . Prasad, Astrophys. J., 1979, 233, 102. 53 G. Winnewisser, R. A. Creswell and M. Winnewisser, 2. Naturjursch., Ted A , 1978, 33, 1169. 54 M. J. Vasile and G. Smolinksy, Int. J . Muss. Spectrom. IOFZ Phys., 1977, 24, 1 1 . 55 R. A. Creswell, G. Winnewisser and M. C. L. Gerry, J. Mol. Spectrosc., 1977, 65, 420. 56 G. Winnewisser, M. Winnewisser, J. J. Christiansen and J. Vogt, personal communication. 57 S . Wilson, Astrophys. J, 1978, 220, 363. ’* C. Wentrup, Tetrahedron, 1971, 27, 1281. 59 B. Bak, 0. J. Nielsen and H. Svanholt, Chem. Phys. Lett., 1978, 59, 330. 6o B. Bak and H. Svanholt, Chem. Phys. Lett., 1979, 66, 387. 61 M. Winnewisser, B. P. Winnewisser and C. Wentrup, personal communication. 62 K. Yamada and M. Winnewisser, Z . Naturforsch., Teil A, 1975, 30, 672. 63 J. K. G. Watson, in Vibrational Spectra and Structures, ed. J. R. Durig (Elsevier, Amsterdam, 64 M. C . L. Gerry, K. Yamada and G. Winnewisser, J. Phys. Chem. Ref. Data, 1979, 8, 107. 65 K. Yamada and M. Winnewisser, 2. Naturforsch., Teil A , 1976, 31, 139. 66 K. Yamada, B. P. Winnewisser and M. Winnewisser, J. Mol. Spectrosc., 1975, 56, 449. 67 A. V. Burenin, E. N. Karyakin, A. F. Krupnov and S. M. Shapin, J. Mol. Spectrosc., 1979,78, 1977), vol. 6, chap. 1. 181. W. H. Weber, J. Mol. Spectrosc., 1980, 79, 396. 69 J. E. Lolck and S. Brodersen, J. Mol. Spectrosc., 1978, 72, 445. ’O M. Winnewisser and J. Johs. Christiansen, Chem. Phys. Lett., 1976, 37, 270. 71 F. M. Nicolaisen and J. Johs. Christiansen, J. Mol. Struct., 1979, 52, 157. 72 M. Winnewisser, W. Peau and J. Johs. Christiansen, to be published; W. Peau, Diplom. Thesis (Justus-Liebig-Universitat, Giessen, 1979). P. R. Bunker and D. J. Howe, J . Mol. Spectrosc., 1980, 83, 288. 74 B. P. Winnewisser and M. Winnewisser, Topics Current Chem., in press. ” A. J. Alexander, H. W. Kroto and D. R. M. Walton, J. Mol. Spectrosc., 1976, 62, 175.