The 3820 A Band System of Propyiial BY J. C . D. BRAND, J. H. CALLOMON AND J. K. G. WATSON The University, Glasgow, W.2, and University College, London, W.C. 1 Received 2nd January, 1962 The first singlet-singlet r * - n transition of gaseous propynal, C2H. CHO, gives rise to a band- system in absorption centred near 3820 A. The rotational and vibrational structures of this spectrum, together with that of the deuterated isotopes C2H. CDO and C2D. CHO, have been extensively analyzed. Of the 12 excited state fundamental frequencies, all but one have been assigned and measured. Rotational analysis reveals a small positive inertial defect in the excited state, as in the (planar) ground state, and it is concluded that the excited state is also planar. All non-totally sym- metric vibrations appear in the spectrum, but only in very short progressions, and this is attributed to " intensity stealing ".Several of the numerous Fermi and Coriolis couplings that are observed have been analyzed. The excited-state vibration21 frequencies show that the effects of electronic excitation are not limited to the carbonyl chromophore, but also transmitted to the furthest atoms in the acetylenic chain. In propynal vapour a weak absorption system extends from about 3820A to- wards shorter waves.19 2 The bands mark a transition analogous to the one-electron rc* --n excitation of formaldehyde, which leads there to a non-planar excited state with an out-of-plane angle estimated to lie between 20" and 2 7 O . 3 ~ 4 It seemed to us of interest to determine whether similar geometrical changes accompany the transition in propynal, where conjugation might be expected to favour a more nearly planar configuration.The structure of propynal is generally favourable for rotational analysis of the ultra-violet bands. As the molecule is very slightly asymmetric ( K = -0.9897 in the ground state 5) only those levels with K = 0 or 1 are displaced appreciably FIG. 1. from the symmetric top values; and the configuration is strongly prolate so that the sub-band origins, except where the degradation produces a head, are well separ- ated from one another. The orientation of inertial axes in the ground state, which is planar, is shown in fig. 1. Since the bands appeared sharp to about 3000 A under low dispersion the prospects for a vibrational analysis appeared good.In the event the detail that could be established exceeded all expectation. 175176 SPECTRUM OF PROPYNAL EXPERIMENTAL The spectra of the normal isotope and of two singly-deuterated derivatives, C2D. CHO and C2H. CDO, were taken in the second order of a 6 m Ebert grating spectrometer described by King.6 At long wavelengths the spectrum is exceedingly sharp, the width of individual rotational lines being less than the instrumental resolution of the grating (ca. 0.09cm-1); but at shorter waves the structure becomes increasingly diffuse, the diffuseness being more pronounced for C2D. CHO than for C2H. CDO or the normal isotope. A number of low-resolution plates were taken with the vapour heated to 200” to determine the effect of temperature on the various bands.The f-value of the system is - 5 x 10-4, roughly twice the intensity of the corresponding transition in formaldehyde ( f = 2 . 4 ~ 10--4).7 RESULTS AND DISCUSSION ROTATIONAL STRUCTURE The principal bands in the spectrum are perpendicular bands of a near-symmetric top. How much detail is visible depends, among other things, on intensity and freedom from overlapping, but in favourable conditions it was immediately obvious that the features to the red of the band centre corresponded to the P-type P- and Q-branches of a perpendicular band. In well-developed sub-bands the PP-branches could be followed for thirty or forty members, so that values of J and K could be assigned unambiguously from the number of missing lines in each branch.2 As the J-structure degrades very slowly to the violet the @branches are prominent, like the Q-branches of an infra-red band.Once the P-type sub-band origins were established the A-type origins could be identified from combination differences of the form, A2F”(K) = RQK-l-PQK+I = 4(Ag-B;;)K-8DiK(K2+1) using microwave values for the constants A: and z;.5 Since the K-structure de- grades rather sharply towards the red the R-type sub-bands form a head of Q-branches at quite low values of K. Fig 2 shows the central region of the 3821 A band (the electronic origin) of C2H . CHO. From the constants A0 and Bo and the assumption of a small positive inertial defect it is possible to calculate the position of lines in the Q-branches of low K. It then emerges that the accumulation of intensity near the band origin is accounted for if, and only if, the transition moment responsible for this band coincides with the axis of greatest inertia (type C band).This result might have been obtained intuitively for the very slight degradation in the J-structure implies that the band centre will have a contour similar to an infra-red band, so that the high intensity about the origin has essentially the same explanation as the central maximum of a type C vibration-rotation band. Although all the strong bands in the spectrum (including the electronic origin) and most of the weaker ones are type C there are a few weak bands of different con- struction. Of them, some are perpendicular bands with a central minimum and can be explained along the lines of fig.2 as due to hybrid transitions polarized almost parallel to the inertial b-axis, while others are parallel bands, type A (AK = 0), the transition- moment lying nearly along the a-axis. A very few bands are A-B hybrids with the intensity divided roughly equally between the perpendicular and parallel components. Rotational constants relating to the vibrationless levels of the combining states in the light isotope are set out in table 1. The lower state K-structure constant (A; - 5;) agrees well with the more accurate values calculated from the microwave spectrum,s although the centrifugal constant D&, which could not be separately determined by microwaves, was there included in the effective principal constant A,”.€30. 2.-The electronic origin band of C2H.CHO, with Fortrat curves for Q-branches near the band-centre. Length of cell, 1 m ; pressure [To face page 176. of gas, 5 mm.J . C. D. BRAND, J . H. CALLOMON A N D J . K . G. WATSON 177 TABLE 1 .-SYSTEM ORIGIN AND ZERO-POINT ROTATIONAL CONSTANTS (C2H. CHO) Too 261 62.89 GROUND STATE 2.1 1387 0.1 55538 2 . 9 0 ~ 10-4 - 0.9897 +0.1718 a.m.u. .@ EXCITED STATE 1.7340 0.1 5660 f0.00001 1.8906 0.1 6336 f0.00002 0.14987 f0.00002 2.73 x 10-4 - 0.9844 (4 +Om40 f0.04 a.m.u. & (f l (a) AU units cm-1, except those labelled (e) and 0; (c) = +(I?+ C) ; ( d ) after correcting the effective microwave (= Ao- B,) ; (e) asymmetry-parameter (dimensionless) ; (b) From the microwave spectrum; (f) inertial defect (&--la-&), a.m.u. A2. ELECTRONIC A N D VIBRATIONAL ASSIGNMENTS The isotope shifts and the assignment of hot bands in the spectrum show that the strong type C band near 3821 A (see fig.2) is definitely at the electronic origin. As the rotational analysis contains no evidence of electronic angular momentum in the excited state the results collectively indicate that the transition is 1A’’-lA’ in nomenclature appropriate to planar configurations, consistent with its interpretation as a one-electron 71”-n transition. A number of aids were used in order to extract the vibrational frequencies from the spectrum. Apart from isotope shifts and the effect of temperature on intensity these aids were mostly concerned with the presence or otherwise of regular perturba- tions. In all the stronger bands the K-structure provided at least several values of the initial and final state combination differences, AzF”(K) and A2F’(K), which were evaluated as fully as the band development allowed.When there is an appreciable component of Coriolis coupling along the a-axis in either state the vibrational angular momentum is reflected in the A2F(K), whose values may differ sufficiently from expectation to characterize the state in question. As an example, we give in table 2 a few results for two bands close to the electronic origin. Both bands are less intense at 200” than at 20” and so must emanate from the vibration- less ground state or from a vibrational level close to the ground state. The first possibility is ruled out by the fact that the A2F”(K) are different from those observed for the ground state, thus both bands must be hot bands: in fact, the A2F”(K) are symmetrically disposed above and below the ground state values and hence the initial states are mutually coupled by Coriolis forces.The only two ground state vibrational levels which meet this specification are vF and v1;’ so that the initial state of each transition can be identified. Moreover, analysis of the coupling gives the constant I [G,i2 I from which one may calculate approximate values of the A2F”(K) for transitions emanating from the overtone and combination levels 2v&178 SPECTRUM OF PROPYNAL vF+v;; and 2vy2, which do give rise to observable transitions. Effectively, the vg, vy2 coupling serves as a fingerprint. The form of analysis is equally successful for certain pairs of vibrations which are coupled in the excited electronic state.TABLE 2.-GROUND STATE K-STRUCTURE COMBINATION DIFFERENCES (C2H. CHO) A2F”(K) (cm-1) = “ Q K - ~ - ~ Q K + ~ for : K vy2 = 1 v; = 1 zero-point mean of col. 2 and 3 3 4 5 6 7 8 9 10 27-3 23.6 36.1 31.5 44.9 39.5 53-5 47.4 62.1 55.3 63.2 71.2 78-9 I 55.72 1 = 0.639 25-38 25.4 33.61 33.8 41-98 42.2 50.27 50-4 58-43 58.7 66-41 74.39 82-14 Another criterion is connected with the anharmonic resonances which are wide- spread in the A” state. For instance, Fermi resonance between v i and v;+t$ is seen in the spectra of C2H. CHO and C2D . CHO and its effects extend to all over- tones and combinations containing vi. Since vi is the most active excited state vibration the resonance allows a large number of levels to be recognized easily as combinations including v i : while, conversely, any level higher than v; but not TABLE 3.-EXAMPLE OF FERMI RESONANCE : Vi-Vi-Vi IN C2H.CDO Assumed : unperturbed GO(V2,04,t’6) = C W f U j + C X i V i V i ; V (Fermi) = hC.kz46qzq4q6; the q’ are dimensionless normal co-ordinates, and k246 is a cubic force-constant, in cm-1. i=2,4,6 i,j=2,4,6 level calc. obs. unperturbed perturbed v4 1267.4 - 1267-4 2v4 25 19.6 - 25 19.4 3v4 3756-6 - 3756.8 v6 941 -9 - 941-9 2v6 1878.5 - 1878.5 2204.0 2 190.4 2 1 89-8 2208.0 2221 -6 2221.3 3471.0 3487-6 3488-6 3458.7 34421 3442.9 4722-6 4738.9 4738-0 469443 4678.5 4678.5 3 130.8 31 14.3 3115.1 3143-3 3 159.8 3 159-0 4330.0 4322.2 4322-5 43963 4428.2 4427-2 4392.7 4368-8 4369.5 w$ = 1275-0, wg = 94455cm-1, ~ 4 4 = -7.6, x66 = -265, x26 = -15.1, x46 = -1.3, I k246 1 = 43.7.J .C . D . BRAND, J . H . CALLOMON AND J . K . G. WATSON 179 affected by resonance is either a fundamental or a combination not involving v;. Thus a band at 0 + 1945.5 cm-1 (0 + 1850.0 and 0 + 1945-8 cm-1 for C2D . CHO and C2H. CDO, respectively) must be associated with an upper-state fundamental, and its isotope shift then identifies it with the C r C stretching mode v;. Table 3 gives a representative analysis of the v;, vi-tv; Fermi resonance in the spectrum of C2H. CDO. Table 4 lists the assigned fundamentals in the ground and excited states. TABLE 4.-FUNDAMENTAL FREQUENCIES OF PROPYNAL approximate description CHZ stretch CHI stretch CzC stretch C=O stretch CHI rock C-C stretch L CCH bend L CCO bend L C-C E C bend IA’ (ground) state 1A” (excited) state C2H .CHO C2D. CHO C2H. CDO C2H . CI-IO C2D . CHO C2H . CDO 3326 2858.2 2106 1696.9 1389 943.7 650.0 613.7 205.3 2605 2858.6 1977 1697.0 1387.6 933.6 507.9 609.0 195.6 3326 2118 2101 1679 6 1080.0 876.5 649.7 61 1.9 201.5 C2H. CHO C2H. CHO C2D. CHO C2H. CDO product rule for a‘ fundamentals : ~~ obs. 1.861 1.936 - harmonic value 1 -892 1-937 - v1o(a”) CHI wag 981.2 980.9 841.0 4621 459.0 q 1 ( d ’ ) LCzCH bend 692.7 548.6 6913 389.7 291 vlz(a”) L C-Cr C bend 260-6 248-5 249-9 345.9 346.9 (1.5) (2.1) (4) (0) (3.0) (2.5) C2H. CHO C2H. CHO C2H. CHO C2H. CHO product rule for a” fundamentals : - _ _ _ ~ C2D . CHO C2H. CDO C2D . CHO C2H. CDO obs. 1.324 1.219 1.344 1 -255 harmonic value 1.336 1-21 8 1.331 1 -242 0 relative intensity of first quanta, origin band = 10.6 corrected for Fermi resonance ; c precise value uncertain. GROUND-STATE FUNDAMENTALS All ground state fundamentals below 2000cm-1 occur at least once in the electronic spectrum (as 0-1 or 1-1 transitions) though vI; is seen only for C2D. CHO and v;l only for C2H. CDO. The fundamental frequencies in table 4 differ from earlier infra-red measurements 8 in two respects. First, a re-examination of the 2100 cm-1 region with an infra-red spectromster of higher resolving power has prompted some revision of the v;’ and v; fundamentals of C2H. CDO, but the changes are small and comparatively unimportant. Secondly, however, the180 SPECTRUM OF PROPYNAL electronic spectrum requires the original assignments of the L C E CH in- and out- of-plane bending modes, v;(a') and vyl(a"), to be interchanged. These two funda- mentals are strongly coupled by Coriolis forces, and are seen in the infra-red as two overlapping perpendicular bands, the former predominantly type B, the latter wholly type C.An absorption-minimum near 688 cm-1 was previously ascribed to the band-centre of v7, type B, and a maximum at 661 cm-1 to v11, type C.8 On this assumption, the partially-resolved and weak K-structure on both wings of the ab- sorption could be numbered and analyzed to yield a value for the Coriolis coupling- constant. This simple interpretation neglected, however, the effect of the Coriolis coupling on the intensities of the bands, which can result in a strong asymmetry in intensity between AK = + 1 and AK = -1 sub-bands: in the limit in which the two coupled vibrations become merely two components of a degenerate mode of a prolate symmetric top or linear molecule, each component of the degeneracy provides only either the X-type or the P-type sub-bands, i.e., two half-bands having the appearance of a single whole perpendicular band.9 In the present case a strong Coriolis intensity-effect is undoubtedly present in the infra-red v7, v11 bands, making their appearance misleading.The correct assignment of vYl is deduced from four ultra-violet transitions having states in common, a fact which could be established from K-type combination- differences. Four vibrational levels are involved (fig. 3), two of which could be identified with the known via and vY2, also by combination-differences (table 2).FIG. 3.-Cross-sequences in the spectrum of C2H . CHO. The relative strengLs are obtainer visual estimates of the intensities corrected by Boltzmann factors. from Together with the electronic origin band, sufficient information is available to cal- culate the frequencies of sub-bands of v11 in the infra-red, and hence to renumber those resolved there. Similar arguments can be applied to the infra-red v7, and the coupled bands reanalyzed. The revised band-origins are at 692.2 cm-1 (vY1 : C2H . CHO) and 649.5 cm-1 (v; : C2H. CHO). The v;&") origin lies close to an intensity-maximum as expected, but the minimum to be expected for the centre of v"(a') lies on a steeply-falling part of the intensity-curve, and although recog- nizable, is not prominent.The magnitudes of the revised Coriolis constants are [;,yl (C2H . CHO) = 0.922 ; (C2D . CHO) = 0.892 ; and (C2H . CDO) = 0.915, close to the value of unity for degenerate vibrations of a linear molecule.J . C. D . BRAND, J . H. CALLOMON A N D J . K . G. WATSON 181 EXCITED-STATE FUNDAMENTALS Besides the excited state fundamental frequencies, table 4 lists the relative in- tensities (on a scale which accords the system origin an intensity of 10) with which the 1-0 transitions appear in the spectrum. The outstanding feature is the high intensity of the CO stretching fundamental v;, implying a large change in the CO bond distance on excitation. The fact that all the a’ fundamentals except vl are active suggests that geometrical changes are rather widespread; but with the ex- ception of vi the 2-0 transitions tend to be weak, so that most of the changes must be relatively small.There is evidence that some of the intensity associated with the 1 4 transitions in v; and vi is explained by the presence in the normal co-ordinate of the internal co-ordinate for CO bond stretching. Of interest is the occurrence, as mentioned above, of bands of all possible polar- ization types in the spectrum, and analysis identifies progressions in all three of the upper state fundamentals which correspond to the non-totally symmetric (out-of- plane, a”) modes in the planar ground-state. (The 1-0 transition in vil is exceed- ingly weak, but as vil occurs also in sequences its numerical value is not in question.) Two explanations are possible : (i) that the excited state is non-planar, i.e., retains no elements of symmetry, all fundamentals being strictly speaking totally sym- metric and hence potentially active in progressions according to the Franck-Condon principle ; (ii) that the non-totally symmetric vibrations a” perturb the planar excited-state electronic wave-function A” and mix it to some degree with another higher state A‘, thereby acquiring non-zero intensity in the spectrum by virtue of “ intensity-borrowing ” from transitions to this higher state.The shortness of the progressions and the small positive inertial defect of the excited state rule out serious departures from planarity ; potential maxima separating possible double potential minima about the molecular plane cannot, therefore, be high, and if present at all should be reflected in appreciable anharmonicities in the out-of-plane vibrations.Judged from these viewpoints the skeletal vibration vi2 is certainly active through intensity-borrowing, for it is seen in sequences to be harmonic up to three quanta, and the intensity of the 2-0 transition is of the order of magnitude to be expected from the ratio of frequencies vi2/vY2. The fundamental v;l is so weak as to argue against any change of out-of-plane angle at the terminal carbon atom in the CCH chain. The large ratio vi,/vyl implies that the 2 4 transition in v11 should have appreciable intensity, but no transition turns up at 0+2vil so that there could be some anharmonicity in the potential.vi0 is of importance in that any non-planarity in the CHO group, such as occurs in excited formaldehyde, would be detected from the behaviour of this fundamental. The 2 4 transition in v10 does appear in the spectrum though most of its intensity (I = 2.0 on the scale used in table 4) is explained by Fermi resonance with v;. After correction for the resonance the location of the 2-0 transition, at 0+949.0 cm-1, is about 25 cm-1 higher than the two-fold fundamental frequency (2 x 462.1 = 924.2) ; thus there is anharmonicity in the same sense, though not nearly so large, as observed with formaldehyde.3.4 The an- harmonicity in via, and the possible anharmonicity in vil, may explain why the pro- duct rule ratio for the excited state a’’ frequencies is in excess of the harmonic value.To see what the vi0 anharmonicity might mean in terms of non-planarity in the CHO group the two observed quanta have been fitted to an equation of the form, v = hcv[+q2 +a exp ( -pq2)] in which a Gaussian function is added to the ordinary harmonic potential.10 Here, v is the limiting harmonic wavenumber for large v, 4 is the (dimensionless) normal182 SPECTRUM OF PROPYNAL co-ordinate for the mode via, and a and p are parameters. The potential has a double minimum unless 2ap < 1. With only two quanta observed, several arbitrary values were assigned to p and the equations solved for the parameters v and a. It emerges that the potential does not necessarily have a double minimum in order to fit the observed results, though a double minimum is by no means ruled out.Assuming an effective mass of 1 a.m.u. the allowed range of values corresponds to an out-of-plane angle between the CH bond and the CCO plane of 0-4" : thus the potential function is probably rather flat for about 4" on either side of the planar configuration, possibly with a central maximum. An interesting feature of the spectrum is the occurrence of transitions which seemingly exchange quanta ofvll and v12. Both vibrations are active in n-n sequences of the usual type, but in addition we observe " false " sequences in which transitions from an initial state with, for example, vy2 = 1 lead to a final state with vil = 1. The assignments are established beyond doubt by combination differences and are shown schematically in fig.3. The explanation is that, whereas vYl and vY2 are essentially pure C=CH and CCC out-of-plane bending modes, the excited state vibrations are roughly equal mixtures of the two co-ordinates : therefore, transitions from the level vY2 = 1 may lead to the CCC bending component in either vi2 or v ; , with relative intensities determined by the coefficients for CCC bending in the normal co-ordinates for ui2 and vil. This mechanism does not seem to have been noticed previously but is probably to be found quite commonly in polyatomic spectra. ELECTRONIC STRUCTURE OF THE EXCITED STATE The dominant frequency changes on excitation are those affecting v4, v10, v11 and v12. The effect on v4 and v10 might have been expected by analogy with formal- dehyde ; but the change in v11, almost as great as that in v10, shows conclusively that the excitation is not localized in the CHO group.Using conventional valence formulae the results suggest that the excited state should be represented as a hybrid of /" H / \- \- H-C=C-C* and H-C-C-C 0- 0. with roughly equal weights for the two structures. The implied delocalization possible in propynal partly relieves the strain which forces formaldehyde out of plane, leaving excited propynal planar or very nearly so. Estimates of the change in bond distance can be made for those normal vibrations which are essentially group vibrations using Badger's rule or Clark's rule.11 Results for the CO, C r C and formyl CH bonds (table 5) are consistent for the three isotopes, suggesting that this simple analysis has some validity. From the relative intensity TABLE 5.-BOND DISTANCES (A) IN EXCITED PROPYNAL r"a r' (Clark) r' (Badger) r' (mean) bond 1 I1 111 I 11 111 C=O 1.215 1.327 1.328 1.334 1.318 1.319 1-325 1.325 G C 1.209 1.241 1.236 1.240 1.238 1.233 1.237 1.238 CHI 1.106 1.094 1-094 1.092 1.090 1.090 1.087 1.091 (fOrnY1) I = C2H.CHO I1 = C2D. CHO I11 = C2H. CDO = ref. (5).J . C . D. BRAND, J . H. CALLOMON AND J . K . G . WATSON 183 of the principal bands in solution (vi progression) the calculated increment in the CO bond distance is -0.10A,1 and a similar result is obtained from the vapour intensities : thus it appears very probable that the excited state CO bond distance is virtually the same as in excited formaldehyde (1-32 A).3,4 It is difficult at present to say anything about the changes in bond angle but a complete set of rotational constants will provide further structural information and this work is in progress. We thank D.S.I.R., the Carnegie Trust, and Esso Research Ltd., for financial support, and Dr. C . C. Costain for a sample of deuterated propynal. 1 Howe and Goldstein, J. Amer. Chem. Soc., 1958, 80,4846. 2 Brand, Callomon and Watson, Can. J. Physics, 1961, 39, 1508. 3 Brand, J. Chem. SOC., 1956, 858. 4 Robinson, Can. J. Physics, 1956, 34, 699. Robinson and Di Giorgio, Can. J. Chem., 1958, 5 Costain and Morton, J. Chem. Physics, 1959, 31, 389. 6 King, J. Sci. Instr., 1958, 35, 11. 7 Duncan and House, quoted by Pople and Sidman, J . Chem. Physics, 1957, 27, 1270. 8 Brand and Watson, Trans. Faruday SOC., 1960, 56, 1582. King and Moule, Spectrochim. 9 Teller and Tisza, 2. Physik, 1932, 73, 791. 36, 31. Acta, 1961, 17, 286. 10 Chan, Zinn and Gwinn, J. Chern. Physics, 1961, 34, 1319. 11 Herzberg, Spectra of Diatomic Molecules (Van Nostrand, New York, 2nd edn., 1950), p. 457.