5 Part (iii) MICROWAVE SPECTROSCOPY By J. E. Parkin (Department of Chemistry, University College, London) SEVERAL recent developments seem to point to microwave spectroscopy (or molecular rotational resonance as the subject may well become known) being on the threshold of a minor boom. The number of papers considered represents an increase of almost 50 % over those for the comparable period a year ag0.l Not only have more investigations of a traditional nature been reported (determinations of complete or partial molecular structures, barriers to internal rotation, dipole moments, and nuclear quadrupole coupling constants) but the technique has been applied to new kinds of problem. The introduction of commercial instruments with much greater ease of operation than those which have evolved in research laboratories will certainly be reflected in the kind of investigations reported in the next few years.The study of collision processes in gases by microwave spectroscopy has been reported by several authors, and the potentialities as an analytical tool have been pointed to and preliminary work carried out. Microwave Spectrometers.-The most important development of recent years which has become most marked in 1968 is the introduction of the first complete commercial spectrometers2 which incorporate some features absent from all but the most sophisticated research instruments. The backward- wave oscillator source has largely superseded the original klystron source over which it has the great advantages of a much wider frequency-range and tuning at a wide range of speeds.It is an interesting irony that microwave spectroscopy began as a very high-resolution technique and only after about 20 years has a low-resolution scanning instrument become available. The possibility of distinguishing the wood from the trees is of great benefit especially in the spectrum of relatively large molecules where the low- resolution spectrum is an extremely simple linear function of the single molecular parameter proportional to the mean of the reciprocals of the principal moments of inertia Ib and Ic. Since the three principal moments of inertia are sensitive functions of the geometry and atomic masses of the molecule, such a low-resolution spectrum can often distinguish between isotopic or conformational modifications.The spectra then correspond to the band spectra of other regions, but being a property of the molecule as a whole do not have the inherent structural information associated with func- J. E. Parkin, Ann. Reports, 1967, 64, 181. Hewlett-Packard Company, Palo Alto, California, U.S.A. ; Tracerlab Ltd., Wey- bridge, Surrey, England. 111112 J. I? Parkin tional groups in these spectra. Since the relative intensities of a given micro- wave transition, i.e. with the same upper- and lower-state quantum numbers, for different isotopic, vibrational or conformational species in equilibrium are proportional to population (along with certain other normally estimable quantities), this measurement enables relative concentrations or, via a Boltzmann factor, vibrational and conformational energy differences to be obtained.The intensity problem which has plagued microwave spectroscopy has largely been overcome by Harrington3* and is discussed later in more detail. The commercial instruments have been designed with this intensity problem in mind and incorporate many features beyond the scope of this article. These, together with the facility for direct-frequency measurement, rapid scanning and location at a particular frequency, although present individually in earlier instruments, have, as combined here, provided a tool of great versatility and ease of operation comparable to the larger instru- ments designed for other spectral regions. Microwave Spectroscopy in Routine Chemistry.-Since their application to molecular spectroscopy, microwave techniques have been recognised to have many advantages as a routine analytical tool.Dailey5 for instance reviewed such possibilities in 1949 when the inherent drawbacks were also recognised. It is significant that as late as 1966 and 1968 Lide6 and Millen7 make many of the same points. Actual practical application has come only in the last year or so and then only in a preliminary way. The microwave spectrum of a gaseous molecule at high resolution is a very sensitive and unique function of the inertial parameters and therefore of its detailed structure. As such, microwave spectroscopy can be used in a straightforward manner for the empirical characterisation of molecules without the need for detailed spectroscopic analysis. The chief limitations of the technique are firstly that the molecule must be polar, as absorption intensity is proportional to the square of the dipole moment, and secondly that it should have sufficient vapour pressure at not too elevated tempera- tures.In practice, this means dipole moments >0-1 Debye and vapour pressures of a few microns at temperatures up to 3 0 0 " ~ . The advantages as an analytical technique are that very small samples of the order of micro- moles can be investigated, it is non-destructive, and it is rapid and repro- ducible. High resolution is a feature of all microwave spectrometers and it is probable that of the order of half-a-dozen lines and their relative intensities would be sufficient for characterisation. In reality, most molecules have many thousands of such transitions so that the problem is one of documentation rather than of inherent impracticality.Unfortunately no collection of charac- teristic lines has been published since 1950.8 The use of commercial instru- H. W. Harrington, J. Chem. Phys., 1967, 46, 3698. H. W. Harrington, J. Chem. Phys., 1968, 49, 3023. B. P. Dailey, Analyt. Chem., 1949, 21, 541. D. R. Lide, jun., Ado. Analyt. Chem. Instrumen., 1966, 5 , 235. D. J. Millen, Chem. in Britain, 1968, 4, 202. P. Kisliuk and C. H. Townes, J. Res. Nut. Bur. Stand., 1950,44,611.Micro wave Spectroscopy 113 ments with their advantage of speed may render a similar but up-to-date compilation extremely valuable. Potentially the most valuable feature of the newer instruments is the possibility of their use in quantitative analysis of polar gas mixtures. If practical, the method has the advantage of being fast, precise, and unambi- guous.In order to have quantitative significance; intensity data must bear some simple reproducible relation to the concentration of the absorbing species. This condition is met in conventional spectroscopy by Beer’s law in which the absorption intensity is directly proportional to the path-length. Deviations from Beer’s law are well known but comparatively unimportant except in special cases. In microwave spectrometers however, with essentially monochromatic sources, such deviations occur as soon as the microwave power is increased beyond a relatively low level, when saturation of a transi- tion at resonance takes place. At still higher power levels, the absorption intensity decreases and there is a characteristic maximum in the intensity- power level curve. Power saturation has been a troublesome problem since the early days of the subject and spectrometers are operated normally at a level sufficiently low to reduce its effects.Under such conditions they lose much of their sensitivity, especially when the gas is being monitored at low concentrations. A second problem arises when pressures are increased beyond a certain point, typically 20 to 50 microns. The absorption intensity then no longer increases with pressure and the lines merely broaden. Harrington3p has overcome these problems by abandoning the Beer’s law intensity-coefficient, y, defined by AP,/Po = yL, where PO is the incident power level and APg is the loss in power for a waveguide cell of length L.He recognises that a new intensity coefficient I?, defined by AP,/Pd = rL, has the property that its maximum attainable value for a given line is directly proportional to the concentration of the absorbing species and, most importantly, is independent of the broadening collision time T. He further shows that this intensity coefficient I’ is directly proportional to the signal amplitude S for a con- ventional Stark modulated absorption cell.9 In many cases insufficient power is available to fully saturate a transition. Harrington overcomes this difficulty by recognising that the shape of the curve of log S against log PO, which he calls the intensity law, is a function only of the instrument except for a change of origin due to concentration of the absorbing species (proportional to a shift in the log S direction) and to the broadening relaxation time T (proportional to a shift in the log PO direc- tion).Experimental points, obtained well before saturation is maximised, can be translated to this theoretical curve by an amount proportional to the relative concentration of the absorbing species. Calibration can then be effected in a number of ways. This mode of operation applies directly to the case where a variation in molecular concentration is being monitored by a R. H. Hughes and E. B. Wilson, jun., Phys. Rev., 1947, 71, 562.114 J . E. Parkin single transition. In measuring the relative intensities of lines separated in frequency, account must be taken of the instrumental variation in power loss with frequency due to the waveguide itself.This is a common difficulty in microwave investigations where the relative intensities of vibrational satellites are of interest in providing their vibrational energy separation. Harrington shows how large errors can result from neglect of this function, especially when the Stark absorption-cell has not been carefully designed to reduce or eliminate reflections and other frequency-dependent losses. In the first paperlo to appear where the technique has been applied to an analytical problem are demonstrated many of the possibilities of the method. Ways of tackling the intensity measurement problem are investigated by tracing the linearity of three different intensity measures with the pressure of the absorbing gas.These measures were firstly the maximum signal obtained for the transition being monitored (the method suggested by Harrington4), secondly the signal at a relatively low but constant power level insufficient to fully saturate the transition, and thirdly the total peak area under these constant-power-level conditions. They found that all the signals were indeed linear over pressure ranges from 0-ca. 25 microns but at higher total pres- sures the curves become appreciably non-linear. The third method gave linearity over an appreciably longer range of pressure. For instance a linear signal was obtained for a 5 % mixture of acetone in nitrogen up to 50 microns total pressure. Despite these deviations at higher pressures, the curves are quite reproducible and are ideally suited to the analysis of certain kinds of gas mixtures.The example given indicates the analysis to be quite as precise as that obtained mass spectrometrically. Vibrational and Conformational Energy Differences.-The measurement of relative intensities is hindered by several factors, some of which have already been alluded to. Frequency-dependent power losses in the waveguide cell render it imperative that a calibration curve be obtained for each Stark absorption-cell, i.e. a cell background-intensity function. Even when these experimental corrections are made, relative intensities are still not directly comparable with the populations and therefore, from the Boltzmann factors, with the energy differences.Rotational line intensities, y, can be shown (see e.g. ref. 11, p. 343) to be given by : 8x2N f 3ckTAv Y = ___ 1pijp vo2 where the constants c, k, and T have their usual meaning, N is the number of molecules per ~ m . ~ in the absorption cell, f is the population factor, vo is the resonant frequency of absorption, and Av the halfwidth. l p j 1 2 is the square of the dipole-moment matrix element for the transition summed over three lo J. T. Funkhouser, S. Armstrong, and H. W. Harrington, Analyt. Chem., 1968,40,22. l1 C. H. Townes and A. W. Schawlow, 'Microwave Spectroscopy,' McGraw-Hill, New York, 1955.Micro wave Spectroscopy 115 perpendicular directions in space, and is a function of the rotational quantum numbers, the molecular asymmetry parameter, and the square of the total dipole moment.In comparing the intensities of lines of two species in equi- librium, in order to determine the relative populations, these factors must be known. Assuming the same rotational quantum numbers for both lines, the additional factors will include the ratios of the dipole moments and transition frequencies squared, the inverse ratio of the linewidths, and a very compli- cated function of the molecular asymmetry parameters. In the case of dif- ferent vibrational species these ratios are normally sufficiently close to unity to be neglected, and relative populations accurate to a few % are easily obtained. In practice, vibration rotation interaction is sufficient to change the inertial parameters of a molecule, in excited vibrational states, to render the satellite lines easily resolvable in the microwave region, as opposed to the i.r.where sequence structure is often quite unanalysable. This comparison is especially favourable for the microwave method when the molecule has many low-lying vibrational levels arising from hindered torsions or ring-puckering vibrations. Accurate intensity-data for a number of these vibrational satellites gives values for the energy levels with sufficient precision to give quite good molecular potentials. Harrington4 considers this problem in some detail with special reference to the low-lying excited states of the out-of-plane bending vibration of trimethylene sulphide. In an earlier paper12 he obtained the vibrational energies of the first four excited-levels as 0, 141, 156, and 241 cm.-l, with a precision of some 3 cm.-l.There is every reason to suppose that this precision might be improved with more advanced equipment. When the different species are conformational isomers intensity compari- sons are not quite so favourable. The nature of a conformational change is often sufficient to modify greatly such factors as the molecular symmetry and the dipole moment, especially the projection of the latter along a given symmetry axis. This is well illustrated by a preliminary communication13 which reports studies of the two conformational isomers of piperidine, CSHXN, having axial and equatorial imino-hydrogen atoms. The relative intensity data, using admittedly less sophisticated equipment than was employed in the previous example, pointed to a population ratio for the axial and equatorial conformers of 2 : 3 yet to a dipole-moment ratio (actu- ally its pa component) in the ratio 3.2 : 1.Since the latter appears squared in the intensity expression, this gives an observed intensity ratio of some 6.5 : 1 in favour of the less-abundant species. The energy difference between the two isomers then corresponds to a separation of 245 It 150 crn.-l, the equatorial conformer being the more abundant. Collision Processes in Gases.-Microwave transitions are particularly sensitive to the phenomenon of pressure broadening because, as opposed to l2 H. W. Harrington, J . Cliem. Phys., 1966, 44, 3481. l3 P. J. Buckley, C . C. Costain, and J. E. Parkin, Clzem. Cumtn., 1968, 668.116 J. E.Parkirt other spectroscopic techniques, the effective slitwidth is very much smaller than the linewidth. The broadening problem was treated in classical terms14 and later modified15 to give the intensity profile for a line in terms of the variables of equation (l), and the mean collision time T. This standard expres- sion (see ref. 11, p. 342) is the starting point for most treatments of collision- broadening data. A consequence of the theory, borne out in practice, is that for pressures between ca. 0.5 mm. and 10 cm. the peak intensity of a line remains constant but the halfwidth increases linearly with pressure. The broadening parameter can be interpeted in terms of T and hence effective collision-diameters may be determined. These values are considerably larger than those obtained from kinetic theory and it is their interpretation in terms of long- and short-range molecular-interaction potentials which forms the basis of the interest in this kind of study.Several groups have reported work recently on these lines, and it will certainly gain impetus when commercial spectrometers become more readily available. The most fruitful approach so far to the interpretation of pressure-broaden- ing data is based on the work of Anderson16 as modified later by Tsao and Curnutte.17 Interactions considered in their theory include dipole-dipole, dipole-quadrupole, quadrupole-quadrupole, and dispersion forces, where the emitting species is self-broadened or broadened by a foreign species. It is thus one of the few methods of obtaining molecular quadrupole moments, although in practice their evaluation is critically dependent on the validity of the theory.The different methods of determining quadrupole moments, including the microwave method have been reviewed.18 In a series of paper@ a further refinement of the theory is presented in which some of Anderson’s approximations are removed, although this results in very unwieldy expressions. The usefulness of the treatment was demon- strated by application to earlier data on the self-broadening of CH3C1, CH3F, CHF3, and PF3. For these systems the interaction potentials are entirely dominated by the dipole-dipole terms and since these quantities are known experimentally very accurately, the theoretical broadening parameter can be calculated and compared directly with the experimental value.The compari- son was extremely good being within the limits of experimental error, and the new theory appears to represent a great improvement on Anderson’s theory. The treatment is applied to data for the more general cases of line broadening of a number of linear and symmetric rotors by a collision with a range of linear molecules and inert gas atoms. In such cases shorter-range forces such as dipole-qudrupole and quadrupole-quadrupole interactions become important and quadrupole moments may be estimated from the l4 H. A. Lorentz, Proc. Amst. Akad. Sci., 1906, 8, 591. l5 J. H. Van Vleck and V. F. Weisskopf, Rev. Mod. Phys., 1945, 17, 227. l6 P. W. Anderson, Phys. Rec., 1949, 76, 647. 17 C. J. Tsao and B. Curnutte, J . Quant.Spectroscopy Radicitiue Transfer, 1962, 2, 41. l8 Krishnaji and V. Prakash, Reo. Mod. Phys., 1966, 38, 690. l 9 J. S . Murphy and J. E. Boggs, J . Chem. Plzys., 1967, 47, 691, 4142; 1968,49, 3333.Microwave Spectroscopy 117 data. In general, the values obtained are an improvement on previous deter- rninations from microwave data and compare favourably with values ob- tained by other experimental methods, e.g. the value for the N2O quadrupole moment is 6.07 & 0.31 Debye A, compared with values of 6.0 obtained from induced birefringence measurements, 8.50 from nuclear spin relaxation, 5.56 from the second virial coefficient, and theoretical calculations in the region 7-40-7-50. Although this is the most favourable comparison, the agreement for other molecules is quite good and is in general inside the spread of other values.New pressure broadening data have been obtained20 for OCS, CHF3, and N2O in collision with a number of foreign gases by a novel experimental technique. The data are interpreted on the basis of Anderson’s theory and it is concluded, not surprisingly, that it is inadequate and in particular tends to overestimate the dipole-dipole contributions. Further data have been obtained21 for OCS broadened by 0 2 and Nz. A more thorough treatment of this improved data, using the theory of ref. 19, is awaited with interest. The conclusion to be drawn from this work is that although microwave pressure-broadening studies will be a valuable source of molecular quadru- pole moments, much critical work, both theoretical and experimental, needs to be done.Concerned with a different aspect of the collision problem, Oka22 has continued the work reported last year1 on the direct study of collision- induced transitions. He uses a high-power double-resonance technique to pump molecules into a non-Boltzmann distribution and observes changes in the absorption intensity of lines involving quite specific levels other than those being pumped. This change in intensity can only result from preferred collision-induced transitions which appear to obey quite definite selection rules of the dipole or quadrupole type. For instance collisions of ammonia molecules with rare-gas atoms and other ammonia molecules appear to give quite different types of selection rule. Some of the transitions induced by the rare-gas collisions seem to be caused by the octopole moment of ammonia.A large amount of data accumulated in this work points, as did much of the pressure-broadening data, to the inadequacy of current theories of molecular collision phenomena, Double Resonance.-In a 1967 paper received after last year’s report was written, Ronn and Lide23 describe the first successful i.r.-microwave double- resonance experiment. Radiation from a carbon dioxide laser delivering 50 watts in the 10.6 micron region was passed through a conventional Stark- modulated absorption cell containing methyl bromide. The P(20) laser line at 944.1 8 cm.-l was strongly absorbed, most probably by the methyl bromide transition from J,K = 9,l in the ground-state to J,K = 8,0 in the vibration- ally excited-state. They observed simultaneously that microwave lines involv- 2o B.Th. Berendts and A. Dymanus, J . Chem. Phys., 1968, 48, 1361; 1968, 49, 2632. 21 K . Srivastava and S. L. Srivastava, J . Chem. Phys., 1967,47, 1885. 22 T. Oka, J . Chem. Phys., 1967, 47, 4852; 1968,48,4919; 1968,49, 3135. 23 A. M. Ronn and D. R. Lide, jun., J . Chem. Phys.; 1967, 47, 3669.118 J. E. Parkin ing the J = 1 and 2, K = 1 levels in the ground-state decreased in intensity whereas the corresponding K = 0 levels were unaffected. This indicated that the disturbance to the equilibrium Boltzmann distribution was transmitted from J,K = 9,l over a wide range of J values but not to levels of different K. This behaviour is in accord with Oka’s findings on similar systems (see refs.in ref. 22). They were not able to identify increased intensity in transitions of the vibrationally excited state due to the overlapping of stronger lines. If this technique can be extended, possibility of the detailed study of simultaneous vibrational and rotational energy transfer will be opened up. The limitation appears to be that the coincidence of the laser line with the i.r. transition must be almost exact. Tunable lasers are becoming available and there seems to be some hope for new developments in the near future. Microwave-microwave double resonance has now become a standard technique as predicted in last year’s report. F l ~ n n ~ ~ has given a useful review of several quantitative aspects of double resonance spectroscopy including the line shape to be expected.There is no doubt that in analysing congested spectra of large molecules or mixtures of molecules double-resonance tech- niques have considerable advantages, although in general their sensitivity is less than in the conventional spectrometer. Vibration-Rotation Interactions.-Several rather thorough investigations of the ground-states of molecules have been reported recently in which i.r. and microwave data are combined to provide the best molecular force-field. As is well known (ref. 11, p. 105) rotational energy can be expressed as the sum of the rigid-rotor energy and that due to centrifugal distortion. The latter is to first order a function of five inde~endent~~ distortion constants, (four in the case of a planar molecule) which are in turn linearly related to the elements of the inverse force-constant matrix.They can thus be combined with i.r. frequencies to provide a consistent set of molecular force-constants. As obtained from microwave work, distortion constants often have precision sufficient to rank with weights comparable with the frequency data and can render determinable a previously underdetermined problem. A typical recent example is a study of the NSF molecule26 in which earlier data27 are extended to higher quantum numbers to give better determination of the distortion constants. The general force-field contains six harmonic. potential constants f1 andfi associated with S-N and S-F bond stretching, fa the bending constant, and three interaction constants f12, fia, and fia Using the i.r.or microwave data alone, only the three diagonal force-constants could be determined. Combination of the data allowedfia also to be deter- mined, the two remaining constants being insignificantly small. A more interesting way of considering the data is to examine how far the distortion constants could be used to fix the vibrational assignments. Of six possibilities only one is consistent with the microwave data and must be preferred even 24 G. W. Flynn, J. Mol. Spectroscopy, 1968, 28, 1. 25 J. K. G. Watson, J. Chem. Phys., 1966, 45, 1360; 1967, 46, 1935. 26 R. L. Cook and W. H. Kirchhoff, J. Chem. Phys., 1967, 47, 4521. 27 W. H. Kirchhoff and E. B. Wilson, jun., J. Amer. Chem. SOC., 1963, 85, 1726.Microwave Spectroscopy 119 though this results in an abnormally low value for fz of 2.85 mdyne/& nor- mally between 4 and 5 in other molecules containing the S-F bond.Other Investigations.-As mentioned in the introduction, considerably more structure determinations have been reported this year than last. Rather than attempt to document these, a few interesting examples chosen more or less at random will be summarized. Some years ago Costain and SrivastavaZ8 investigated the microwave spectra of gas-phase hydrogen-bonded dimers of trifluoroacetic acid with acetic, monofluoracetic, and formic acids. The spectra consisted of broad evenly spaced bands assigned to successive transitions of the type J + 1 t J . No K structure was observed but some structural information could be inferred. The spectrum of the formic acid dimer has been rein~estigated~~ by measuring the susceptibility dispersion curve for the J = 8 +- 7 transition, and splitting attributable to K structure has been found.The reason why this structure should be apparent in the one technique and not the other is not clear, but a value of 20 D for the dipole moment is obtained from the data and the inertial parameters are not unreasonable. It is significant that these two are the only microwave investigations on hydrogen-bonded dimers reported in the literature although other systems are well known from i.r. studies. It is possible that the broad-band spectra expected for these species have been missed by use of high-resolution instruments ; low-resolution scanning instmments may well alter this situation.The threefold barrier to internal rotation for methyl alcohol changes from 375.6, to 371.8 and 370.3 for the isotopic species CH30H, CD30H, and CH30D. Lees and Baker30 have reported a thorough reinvestigation of the spectrum of this molecule in the millimeter wave region. They were unable to separate the sixfold from the threefold term in the barrier potential, the two moments of inertia about the near-symmetry axis, and two interaction para- meters; they cast some doubt on determinations of this term in other mole- cules. They also obtain a complete geometry for the molecule from the inertial parameters, with significant increase in precision over earlier deter- minations. The bond lengths are YCH 1.094 & 0-003, YOH 0.945 & 0.03, and TCO 1.425 rf 0.002 A, and the methyl tilt is 3’16’ & 11’.The spectrum of nitrosomethane has been investigated31 and a value of 1137 cm.-l found for the barrier to internal rotation. The dipole moment is 2.300 D but the struc- tural parameters are not well determined. Structure determinations have been reported32* 33 of the related compounds SeF4 and SeOFz. The former has an unsymmetrical structure, as implied by the observation of a microwave spectrum, which can be thought of as a 28 C. C. Costain and G. P. Srivastava, J. Chem. Phys., 1964, 41, 1620. 29 G. P. Srivastava and M. L. Goyal, Phys. Rev., 1968, 168, 104. 30 R. M. Lees and J. G. Baker, J. Chem. Phys., 1968.48, 5299. 31 D. Coffey, jun., C. 0. Britt, and J. E. Boggs, J . Chem. Phys., 1968, 49, 591. 32 I. C. Bowater, R. D.Brown, and F. R . Burden, J. Mol. Spectroscopy, 1968,28,454. 33 I. C. Bowater, R . D. Brown, and F. R . Burden, J. Mol. Spectroscopy, 1967, 23,272; 1968, 28, 461.120 J. E. Parkin distorted octahedral structure with selenium lone-pair orbitals occupying two of the positions. The axial and equatorial Se-F bonds have lengths 1.771 and 1.638 A, and are at angles of 169.20 and 100.55" respectively. The dipole moment is 1.779 D. The structure is very similar to that reported34 for SF4. %OF2 has the structural parameters rseo 1.576, rseF 1.730 A, F-Se-F 92*22", and 0-Se-F 104.82". The dipole moment is 2-84 D. The molecules formally have sp3d hybridisation and this data will give insight into the electronic properties of selenium in this ~tate.~5 A spectrum attributed to cyclopropanone has been reported.36 The ob- served moments of inertia are in agreement with a structure with the para- meters, rc(l)o 1-18, rc(qc(2) 1-49, rc(z)c(3) 1.58, YCH 1.085 A, C(3)-C(l)-C(2) 64 and H-C-H 117'35'.The molecule has CZ, symmetry and the only non-zero dipole moment component pa is 2.67 & 0.1 D. These data rule out alternative structures such as allene oxide and the oxyallyl radical, in spite of the indication of Huckel calculations that cyclopropanone should be unstable with respect to the latter. Some of the structural parameters of cyclopropyl bromide have been determined37 as well as the quadrupole coupling constants, which indicate 22% ionic character for the C-Br bond, although this is probably an overestimate. The similarity of the coupling constants with those of vinyl bromide, perhaps indicating a similar C-Br bond, is noted. The ground and low-lying excited vibrational states of the ring-puckering vibration of cyclobutanone and methylene cyclobutane have been investi- gated.3*y39 Barrier heights obtained for the motion are 7.6 and 160 cm.-l respectively indicating considerably more preference for the non-planar structure in the latter compound. The partial structures of cyclohexene have been determined40 and the half-chair conformation and CZ symmetry con- firmed. The dipole moment is 0.331 D. The spectra have been investigated41 of some 1 -halogenoadamantanes, C~OHISX, probably the largest compounds yet to be studied by microwave spectroscopy. The assumption that all the angles are tetrahedral and all the C-H bond-lengths 1.09A leads to the reasonable C-C bond-length of 1.541 & 0.001 A for the chloro-compound. In view of the highly symmetrical cage-structure of adamantane, these assumptions are quite reasonable. The carbon-halogen bond-lengths are 1.370, 1.790, and 1-947 8, for C-F, C-CI, and C-Br, and are in accord with the analogous methyl, ethyl and t-butyl bond-lengths. 34 W. M. Tolles and W. D. Gwinn, J . Chem. Phys., 1962, 36, 1 1 19. 35 R. D. Brown and J. B. Peel, Ausfral. J . Chern., 1968, 21, pp. 2589, 2605, 2617. 3% J. M. Pochan, 3. E. Baldwin, and W. H. Flygare, J . Amer. Chem. SOC., 1968,90, 1072. 37 F. M. K . Lam and B. P. Dailey, J . Chem. Phys., 1968, 49, 1588. 38 L. H. Scharpen and V. W. Laurie, J . Chem. Phys., 1968,49, 221. 39 L. H. Scharpen and V. W. Laurie, J. Chem. Phys., 1968,49, 3041. 40 L. H. Scharpen, J. E. Wallrab, and D. P. Ames, J . Chem. Phys., 1968,49, 2368. 41 D. Chadwick, A. C. Legon, and D. J. Millen, J . Chem. Sot. ( A ) , 1968,1116.