Electronic Absorption Spectra of HCO and DCO Radicals BY J. W. C. JOHNS,* (Miss) S. H. PRIDDLE t AND D. A. RAMSAY Division of Pure Physics, National Research Council, Ottawa, Canada Received 15th January, 1963 The long wavelength absorption bands of HCO and DCO have been observed with much greater intensity than in the earlier work of Herzberg and Ramsay. Rotational and vibrational analyses of 22 new bands have been carried out and new molecular constants obtained. The numbering of the bands in the principal progression (0, 05, O)--(O, 0,O) has been revised so that the C vibronic levels now correspond to odd values of the quantum number 0;. Analysis of the rotational envelopes of some of the diffuse bands has been carried out. The “ ll-bands ’’ are found to be staggered to lower frequencies relative to the “ 22-bands ’’ and fairly large C-A and T T - 4 splittings are also found.These splittings are roughly quadratic in K, i.e., v$ = vo-GK2, where G-15 cm-1 for HCO and -10 cm-1 for DCO. The magnitudes of these splittings are consistent with those calculated from the theory of Pople and Longuet-Higgins assum- ing that the molecule exhibits a Renner effect. A short discussion of dissociation and predissoci- ation is given. The long wavelength absorption system (7500-4500& of the HCO free radical was first studied by Ramsay,l and Herzberg and Ramsay.2 The electronic transi- tion was found to invdve the ground state in which the molecule is bent (LHCO = 119” 30’) and a low-lying excited state in which the molecule has a linear equilibrium configuration.The spectrum consists principally of a long progression of bands (0, u;, O ) t ( O , 0, 0) involving the bending vibration vi in the excited state. Alter- nate bands in the progression are sharp while the others are difuse. The sharp bands involve the K” = 1 rotational levels of the ground state of the molecule which approximates very closely to a symmetric top, and the K‘ = 0 levels of the excited state. The sharp levels in the excited state are therefore Z vibronic levels. Since the vibrational numbering of the bands was not definitely established by Herzberg and Ramsay, it was not possible to state unambiguously whether the Z vibronic levels are &ectronic X &ibrational, nelectronic X nvibrational Or higher pro- ducts. From the rotational selection rules, however, it was found that the transition moment lies in a direction perpendicular to the molecular plane, hence the bands are vibronicaZZy either X+tA”, or X-cA’.Herzberg and Ramsay gave reasons for assigning the spectrum to a 2Z++-2Af’ electronic transition, but this conclusion was at variance with the predictions of Walsh 3 who from molecular orbital argu- ments concluded that the electronic transition should be 2A” t 2 A ’ . Ramsay 4 later pointed out that if the two combining states are derived from a common 2lI state as is implied by the predictions of Walsh, then the upper state (in which the molecule has a linear equilibrium configuration) is in effect only “ half a rll-state ”, which will be denoted by the symbol 2A”lI. The transition may there- fore be written 2A”n - 2A’.Under these circumstances, the levels associated with the bending vibration of the molecule in the upper state would be expected to show large vibronic splittings (cf. NH2, Dressler and Ramsay 5) due to the effects of elec- tronic-vibrational interaction (Renner effect). * N.R.C. Postdoctorate Fellow 1959-61. -f N.R.C. Summer Research Student 1962. 90J. W. C. JOHNS, S. H. PRIDDLE AND D. A . RAMSAY 91 Unfortunately, for HCO and DCO most of the levels of the excited state are diffuse and no rotational analysis is possible, except for the X vibronic levels. A careful study of the rotational envelopes of the diffuse bands in the present work, however, has permitted the location of several ll, A and iD vibronic levels with an accuracy of a few cm-1.Fairly large vibronic splittings (up to - 100 cm-1) have been ob- served and the origins of the various vibronic bands fit the usual quadratic formula vf = VO- GK2, where G - 15 cm-1 for HCO and G- 10 cm-1 for DCO. The mag- nitudes of these splittings are in moderately good agreement with those calculated from the theory developed by Pople and Longuet-Higgins6 for NH2. Another consequence of the present work is that the principal progressions (0, 05, 0)-(0, 0, 0) for HCO and DCO have been extended to lower quantum numbers so that it is now possible to determine unambiguously the vibrational numbering of the bands. It is found that the I: vibronic levels correspond to odd values of the quantum number vi, and the earlier values of 04 given by Herzberg and Ramsay should be reduced by one.These results leave no doubt that the I: vibronic levels are derived from nelectronic x nvibrational levels, and that the electronic transition may be written 2A”II - 2A’. The recent electron spin resonance measurements of Adrian, Cochran and Bowers 7 provide independent evidence that the ground state is a 2A’ state and not a 2A” state. EXPERIMENTAL The new bands of HCO and DCO were observed during the flash photolysis of CH3CHO and CD3CDO in the manner described by Herzberg and Ramsay. Since weaker bands were sought in the present work, longer absorption paths were used together with flash lamps which were both brighter and shorter in duration. The essential features of the apparatus are as follows : Absorption tube : length 2 m.No. of traversals of the multiple reflection mirror system-up to 32. Photolysis lamps: two 1-m flash lamps in series, fired by two 300pF condenser banks Source lamp : 3 mm int. diam. quartz capillary lamp, fired by a 2 pF condenser charged Time delay - 30 psec. Pressure of acetaldehyde : 50 mm Hg. Spectrograph : 21 ft concave grating ; Eastman Kodak 1-0, I-F and hypersensitized I-N plates. Reference spectrum : provided by iron hollow cathode lamp. charged to f7 kV ; flash duration - 30 psec. to 15 kV ; flash duration - 10 psec. The sharp bands were measured using a photoelectric comparator while measurements of the diffuse bands were taken from microphotometer traces of the bands.92 ABSORPTION SPECTRA OF HCO AND DCO of some of the bands investigated in the earlier work have been appreciably ex- tended and some new band heads observed.The band-head measurements for all the sharp bands of HCO and DCO are summarized in table 1. The vacuum wavenumbers and assignments of the rotational lines are given in a separate publication.8 TABLE BAND HEADS FOR THE SHARP BANDS OF HCO AND DCO HCO DCO &$ir) Y (vac.) int. 8573 3 11660.9 w 8561 -9 11 676.5 8235.5 12139.2 vw 8224.4 . 121 55-6 7560.9 13222-3 m 7551.6 13238.5 7310.7 1367409 vvw 7301.8 13691.4 6774.1 14758.0 s 6766.3 14774.9 6138.0 16287-5 5643.7 17714 17693 vvw 3 5650-3 5624.0 1777601 5574.95 17932.4 m 5570.1 17948'1 5538.0 18052 18036 vvw 3 5542.8 5511.5 18138.7 w 5505.8 18157.5 5201.0 19221.8 s 5195-6 19241.6 5152-5 19402.8 m 5148.1 19419.2 5123.3 19513.2 w 5119.0 19529.6 5105.9 19579.7 vw 5100.3 19601.4 4838.3 20662.8 m 4833.3 20684.0 4795.0 20849.1 w 4791.0 2Q866.6 4769.6 20966-2 vw 4761-2 20997 4757.0 21016 4527.7 22080.2 w 4523.2 22101.9 4488.7 22271.8 vw 4485.0 22290.5 4459.8 22397 22416 vvw 3 4463-6 * overlapped by emission line: I (m-q 3 "R) 3 3 3 $1 3 g} $1 9 3 3 3 3 * * 3 vvw 9 3 3- 6144.7 16269.6 vs 5629-9 17757'3 vs weak, vvw-very very weak.assignment 1 (air) Y (vac.) int. (A) (an-') (0, 390)- 8129'8 12297'0 w (0, 0, 0) 8123'2 12307.1 (0, 1, 0) 7879.5 12687.6 (0, 0, 0) 7386'2 13535'0 (0, 1, 0) 7194'6 13895.5 (0, 0, 0) 67803 14744'1 (0, 0, 0) 6273'3 15936'3 (0, 0, 1) 5842'4 171 11 -5 (0, 0, 0) 5738'3 17421-8 (1, 790)- 5475.9 18256.8 s (0, 0, 0) 5471.9 18270'0 (0, 9, 1)- 5454.6 18328.1 w (0, 0, 0) 5451'4 18338.9 (0, 1, 0) 5385.0 18564-8 (0, 0, 0) 5149'9 19412-6 (0, 0, 0) 5130.1 19487-5 (0, 0, 0) 5076.5 19693.1 (0, 15, 0)- 4870.8 205246 m (0, 1, 0) 4867-2 20540.0 (0, 0, 0) 4848.4 20619.8 (0, 0, 0) 4806-6 20799 (0, 0, 0) 4616-5 21655-4 (0, 1, 0) 4598.7 21739.1 vvw (0, 5,O)- - (0, 5, 0)- 7391.8 13524.7 m (0, 7, 0)- 7199'7 13885'7 vw (0, 7, 0)- 6785-5 14733'2 m (0, 9, 0)- 6277.9 15924.5 s (0, 11, 1)- 5846.7 17098.9 vs (0, 11, 0)- 5742'5 37409.3 vw (0, 13, 0)- 5389.0 18551'2 vw (0, 13, 0)- 5153'6 19398'6 m (1, 9, 0)- 5133'1 19476.1 w (0, 11, 1)- 5080.6 19677.2 vw (0, 15, 0)- 4851.2 20607.9 w (1, 11, 0)- 4809.4 20787 vvw (0, 13, 1)- 4620.8 21635.0 vw (0, 17,O)- 46020 21723.6 vw (0, 17,O)- (0, 0, 0) (1, 13,O)- (0, 030) (0, 15, 1)- ( O , O , O ) rs-very strong, s-strong, m-medium, w-we assignment :ak, vw-very5406.1 A I 5656.9 a I 5432-5 a 1 T Ql-2 T Rl-* 5673.1 A I b i Ql-2 1 RI-2 I QI-0 I 4 - 0 FIG.1.-Diffuse bands of HCO and DCO photographed using the second order of a 21-ft. concave grating spectrograph; (a) (0, 12, 0)-(0, 0, 0) band of HCO, (b) (0, 14, O)-(O, 0, 0) band of DCO. The reference spectrum is provided by an iron hollow cathode lamp. [To face page 93.J . W. C . JOHNS, S. H. PRIDDLE AND D. A . RAMSAY 93 Improved observations have been made on the diffuse bands. Reproductions of the (0, 12, O)--(O, 0, 0) band of HCO and of the (0, 14, 0)-(0, 0,O) band of DCO are given in fig, 1. Each diffuse band shows four distinct features which may be identified with the intensity maxima of the R- and @branches of the K’ = I t K” = 0 and K‘ = 1i-K‘’ = 2 sub-bands.Microphotometer curves of these bands are given in fig. 2 and 3a. Two further weak maxima were found at the long wave- length end of the DCO “ &band ’’ and were assigned to the intensity maxima of. QI-0 t50 +25 1/, -25 -50 -75 -100 -125 -I50 -175 FIG. 2.-Observed and calculated rotational envelopes for the (0, 12,O)-(O, O, 0) II-band of HCO. The top curve is an original microphotometer curve of the band shown in fig. la. The lower curves are calculated rotational envelopes for this band, assuming a triangular line width function and half-intensity line widths of (a) 10 cm-1, (b) 20 cm-1 and (c) 30 cm-1. the K’ = 3cK” = 2 and K’ = 3 c K ” = 4 sub-bands of the corresponding “ SD- band”. The former intensity maximum can be seen in fig.3a. A microphoto- meter curve of the (0, 15, 0)-(0, 0, 0) band of DCO is given in fig. 3c. In addition to the sharp rotational structure of the “ C-band ”, the intensity maxima of the K’ = 2 t K ” = 1 and K’ = 2 c K ” = 3 sub-bands of the “ A-band ” are clearly visible. Measurements on the intensity maxima of the various diffuse bands of HCO and DCO and their assignments are given in table 2. In addition to the sharp bands and the diffuse bands described above, the experiments gave definite evidence for an absorption continuum in the region from approximately 5000 to 7000A. The intensity of absorption of this continuum was approximately 50 in the strongest region. . .. ..94 ABSORPTION SPECTRA OF HCO AND DCO ANALYSIS SHARP BANDS Rotational constants and vibrational quanta for the upper state X levels of HCO and DCO were obtained by the method of combination differences, viz., R*(J- 1) - R(J- 1) = Q*(J) - Q(J) = P*(J+ 1) - P(J+ 1) = G*(v) - G(v) + (B* - B)J(J+ 1) - (D* - D)J2(J+ 1)2.(1) Quantities marked with an asterisk refer to a new band while the other quantities refer to a previously analyzed band. The left-hand side of eqn. (1) was plotted I I I I 1 17700 Vb50 V600 17550 17500 I 1 1 18250 18200 18150 l8lOO FIG. 3.4bserved and calculated rotational envelopes for some diffuse bands of DCO: (a) original microphotometer curve of the (0,14,0)-(0,0,0) band, showing the 11-band and part of the @-band; (b) calculated rotational envelope for the JI-band, assuming a triangular line width function and a half-intensity line width of 16 cm-1; (c) original microphotometer curve of the (0, 15,0)--(0,0,0) band showing the sharp X-band and the diffuse A-band. against J(J+1) for each new band and the various sets of points were found to lie on straight lines.The values for (D* - 0) may thus be neglected and values for AG and AB were determined from the intercepts and slopes of the straight lines.J . W. C. JOHNS, S . H. PRIDDLE A N D D . A. RAMSAY 95 TABLE 2.-INTENSITY MAXIMA FOR THE DIFFUSE BANDS OF HCO AND DCO HCO DCO assignment ir (air) Y (vac.) (A> ( a - 1 ) A (air) v (vac.) assignment (A) (cm-1) 6482.2 15422.6 Ql-2 (0, 8, 0)- 63222 15813 Q2-1 (0, 11, 0)- 6181.1 16174 ( 0 7 9, 0)- 6043.7 16505 6057.2* 16542 16948 6168.1 16208 5906.4 5874.1 5863.2 17051 5645 17709 (0, 11, 0)- 5755 5638 17732 56941- 5432-5 18402.4 5673.1 17622.1 Q1-21 54244 18430.1 (0, 12, 0)- 5665.0 17647.5 R1 2 (0, 14, 0)- 5406.1 18492.3 Ql-0 (0, 0, 0) 5656.9 17672-6 Qi-o 1 (0, b, 0) 5398-3* 18519.2 Rl-0 5648-1 17700.1 R1-o J 5213.7 19175 Q2-1 (0, 13, 0)- 5525 18094 (0, 0, 0) 5491 18207 5395 18530 5353 18677 5325.0 18774.1 531 7.4 5312.5 6449.7 15500.2 Qi-o } (0, 0, 0) ( 0 7 0, 0) 53043* 18847.4 R1-o * shoulder.These numbers are less certain. Since the present method of determination yields more accurate values for AB and AD than the method used by Herzberg and Ramsay, the earlier measurements were re-evaluated by the present procedure. Bands of the principal progression (0, u;, O)--(O, 0, 0) were referred to the (0, 9, 0)-(0, 0, 0) band of HCO or to the (0,13,0)-(0, 0, 0) band of DCO as standards and the effective band origins, voefi., and B’ values are listed in table 3.Other bands were referred to appropriate standards as indicated in table 4 in which the values for AG and AB are also given. TABLE 3 .-EFFECTIVE BAND ORIGINS AND UPPER-STATE ROTATIONAL CONSTANTS FOR THE HCO AND DCO BANDS HCO DCO band voeff* ( a - 9 B’ (cm-1) yoeff. ( c m 3 B’ (cm-1) 11661.06f0.05 13222.37 f0.04 14758.02 f0-04 16269.59 17757031 f0.03 19221.77 &Om05 20662.80 f0-06 22080.30 f0.06 1.3436 f0-0005 1-3476 f0.0003 1.3517 f0.0003 1-3565 1.3619 310-0003 1-3673 f0.0006 1.3729 f0*0006 1-3782 f0.0010 12297.17 f0.05 13524.77 f0.06 14733.27 f0.05 15924.62 50.05 17098.87 18256.86 f0.05 19398.75 f0.05 20524.73 50.06 21635.12fO-10 1.1 11 1 f0.0005 1*115Ort0~0005 1.1212 f0.0005 1 * 1270 f O.OOO5 1.1348 1.1408 f0-0005 1.1481 f0.0005 1 * 1 572 k O.OOO8 1.1657 f0.0010 All errors refer to errors obtained from combination difference plots, using the (0, 9, O)-(O, 0, 0) band as the standard band for HCO and the (0, 13, O)--(O, 0, 0) band as the standard band for DCO.The values quoted for the standard bands are taken from Herzberg and Ramsay (1955).bimds TABLE (1.-vIBRATIONAL MTERVALS HCO AG (cm-1) u (cm-1) 3174.55 f0.06 3 133.39 f0-05 3091.93 h0.08 3050.15 f0.08 1756.00 50.05 1745.57 f0.10 1083.00 f0.08 a; = Om74 &O*OOO6 a; = 0.0068 f0-0004 a; = 0.0064 &040008 a; = 0-0059 f0.0008 a; = 0.01 15 f010005 a; = 0,0123 &0*0010 afRW = 04033 f0.0006 C I ~ " = - O*Oo24 fO.OOO6 AND @-VALUES FOR HCO AND DCO DCO bands AG (cm-1) a (cm-1) (1, 11, 0)-(0, 0, 0) 2403.62jz0.05 a; = 0.0061 f0-0005 (1, 13, 0)-(0, 0, 0) 2377-33f0.05 a; = 0.0062f0*0005 (0, 11, w-0, 0, 0) 1 (0, 1 3 , 0 ) ~ 0 , 0 , 0 ) 1 (0, 15, w-0, 0,O) 1 (1, 15,O)-(O, O, 0) 2351*16f0.08 ai = 0*0042f0*0008 (1, 17,0)-(0, 0, 0) 2325.03 f0.05 ai = 0.0050 50-0005 (0, 17,0)40, OYO)J.W. C. JOHNS, S. H . PRIDDLE AND D. A. RAMSAY 97 The rotational constants were fitted to the equation : using the method of least squares 9 to determine the coefficients of (uj+ 1). The values for the constants are given in table 5. To obtain the value of Bi for DCO it was assumed that the value for a; is the same as for HCO. This assumption is TABLE 5.-MOLECULAR CONSTANTS FOR HCO AND DCO state constant HCO DCO units xzo3 Too0 Y52 Y222 cm-1 cm-1 cm-1 cm-1 cm-1 cm-1 A A cm-1 cm-1 cm-1 cm-1 cm-1 cm-1 cm-1 cm-1 cm-1 assuming the same value for a3 as for HCO; b assumed value; C from Teller-RedIich product rule.questionable, since for HCN and DCN the corresponding a-values differ by -35 % (Douglas and Sharmalo). From the rotational constants Bz for HCO and DCO, the equilibrium bond lengths r:(CH) and rk(C0) were calculated. The following values were obtained : r:(CH) = 1.044&0*03 A, r:(CO) = 1.1866 +0*008 A. The rather low value for r:(C€€) is probably due mainly to inaccuracies in the deter- mination of Bi for DCO caused by the moderately long extrapolation of the &, "i.0 values and the assumption of the value for a;. A closer estimate of the CO bond length can probably be obtained by using the HCO data alone, assuming an ap- propriate value for r;(C€€).Since the excited state of HCO is very similar to the ground state of HCN both in vibration frequencies and electronic structure,4 we shall assume that r;(CH) = 1.065+_0-01 A (cf. Douglas and Sharma lo), whence r:(CO) = I -1 82 & 0.002 A. D98 ABSORPTION SPECTRA OF HCO AND DCO The effective band origins listed in table 3 and the vibrational intervals given in table 4 were fitted to the equation : It should be noted that Tieff = T& constants and for Tho are given in table 5. The values for the vibrational The effective rotational constants a;R*’ and ~$2‘‘ and the vibrational frequency for the bending vibrational level of the ground state were obtained from the combination relations : R(J) - R’(J) = P(J) - P*(J) = v2 ” -aZpR’’ J ( J + l ) , (4) ( 5 ) Q(J) - Q*(J) = V; - a y J ( J + 1), where quantities marked with an asterisk refer to a “ hot ” band originating in the level u;‘ = 1 and the corresponding unmarked quantities refer to a band originating in the ground state.The left-hand-sides of these equations were plotted against J(J+ 1) and values for a:R.‘, ctf” and vg were readily deduced (see table 4). A very weak band was found underlying the P-branch of the (0, 11, o)-(o, 0, 0) band of HCO and might be due to HC13O in natural abundance; but a calculation of the expected isotope shift gives a value of 75 5 cm-1 whereas the observed shift is -63 cm-1. A more reasonable assignment for this band is that it is a ‘‘ hot ” band (0, 11, 1)-(0, 0, l), originating in the ui’ = 1 level of the ground state.From the difference between the frequencies of the bands (0,ll , l)---(O, 0, 0) and (0, 11 , 1)- (0, 0, 1) we find v;I = 1820-2 cm-1. This value is smaller than the value (1860 cm-1) quoted by Ewing, Thompson and Pimentel 11 on the basis of matrix studies in the infra-red. DIFFUSE BANDS The band origins for the diffuse bands were determined by fitting calculated rotational envelopes to the observed band contours. Since the rotational selection rules are AK = f 1, AJ = 0, & 1, the “ II-bands ” consist of two sub-bands, viz., K‘ = l+K” = 0 and K‘ = 1 t K ” = 2, each sub-band consisting of a P, Q and R-branch. The sub-band origins are separated by 4(A” -I?’), i.e., by N 84 cm-1 for HCO and -50 cm-1 for DCO.The calculated rotational structures for the (0, 12, O)--(O, 0, 0) band of HCO and the (0, 14, O)--(O, 0, 0) band of DCO are shown in fig. 2 and 3b respectively. The following approximations were made: (a) the B values for the II-levels were obtained by interpolating between those for the C.-levels, (b) the K-splittings of the IT-levels were neglected, (c) the Honl-London formulae (Herzberg 12) were used for the line intensities, and (d) the rotational temperature was taken as 50°C, the temperature of the absorption tube during the experiment. The ground-state energy levels were calculated by substituting the A” , B” and C” values given by Herzberg and Ramsay into the expressions derived by P0l0.13 The band envelopes were calculated by assuming a triangular line shape function and various values for the half-intensity line width, Avt.Three curves are shown in fig. 2 for different values of Ava, viz., 10,20 and 30 cm-1. The agreement between the experimental band envelope and the theoretical contour calculated with Av) = 20 cm-f leaves no doubt as to the correct assignment of this band. Furthermore,J . W. C . JOHNS, S. H. PRIDDLE AND D. A. RAMSAY 99 since the shape of the calculated envelope varies appreciably as Avt changes, we may obtain a fairly reliable estimate for the half intensity line width, viz., Avt = 20 + 3 cm-1. In a similar way the calculated rotational envelope for the (0, 14, 0)- (0, 0, 0) band of DCO fits well with the experimental curve if we assume that Avt = 16+ 3 cm-1 (see fig.3a and 3b). The band origins for the various II-bands of HCO and DCO were determined by finding the best fit of the calculated envelopes to the experimental curves. The values are given in the second column of table 6. TABLE 6.-sUB-BAND ORIGINS AND G-VALUES FOR HCO AND DCO sub-band origin (cm-1) 15537.7 15511 f 5 16269.6 16181 f8 17037.4 17023 f 3 17757.3 17716 f5 18513.5 18498 f 3 19221.8 19182 f8 15924.5 (C) 15820f8 (A) 16526.2 (L‘ C ”) 16514f5 (n) 17098.9 (C) 1705815 (A) 17692.3 (“ E ”) 17682f2 (n) 17692.3 (“ I: ”) 17579f10 (@) 18256.8 (C) 18216f5 (A) 18842.2 (“ C ”) 1883064 (n) 18842.2 (“ C ”) 18737i-10 (a) G(o bs.) G(calc.) (cm-1) (cm-1) 26.1 1 2 12.2 f5 10.2 f l - 2 10s3 k 2 1 2-6 i- 1.1 10.2 f 1.2 12.2 f 4 11.7 f 1.1 17-2 15.7 14.4 13-2 13.2 12.2 11.3 11.3 The origins of the hypothetical “ C ” sub-bands are given relative to the J = 0, K = 0 level of the ground state.The origins of the other Z sub-bands are taken from table 3 and have not been corrected by the term (A”--S”). The A-bands likewise consist of two sub-bands, viz., K’ = 2-K” = 1 and K’ = 2 t K ” = 3 separated by - 167 cm-1 for HCO and -99 cm-1 for DCO. Both sub-bands have been observed for DCO but only the stronger sub-band, viz., K’ = 2 c K ” = 1 has been identified for HCO. The rotational contours were calculated in the same way as for the II-bands and rotational line widths of -20 cm-1 were assumed. The experimental band contours are less well-defined than those for the n-bands, especially as the K‘ = 2 t K ‘ = 1 sub-bands are overlapped by100 ABSORPTION SPECTRA OF HCO A N D DCO parts of the Z-bands (see fig.3c). Nevertheless, the observed and calculated con- tours could be fitted with an accuracy of - 10 cm-1 and values for the origins of the A-bands obtained (see table 6). No HCO bands with K‘>2 could be positively identified but assignments for two @-bands of DCO are given in table 2. DISCUSSION VIBRATIONAL NUMBERING All three upper state frequencies for HCO are now known, together with the ground-state frequencies v y and v;. The only unknown frequency is thus the ground- state C-H stretching frequency vi‘. For DCO, only two upper-state frequencies, v; and vi. are known but the third frequency may be calculated using the Teller-Redlich product rule,l4 viz., Since this equation is valid for zero-order frequencies, and the anharmonic corrections for the v1 vibrations are large, the experimental values for v;(CH) and v; (CD) were corrected using the anharmonic constants determined for the HCN and DCN molecules (Douglas and Sharma 10).From the above equation we then obtain v;(DCO) = 1713 cm-1. For the ground state of DCO, only the bending frequency v;‘ is known. The Teller-Redlich product rule gives the relation If we assume that (u;’(DCO) -o;’(HCO), then a value for mi’( DCO) may be calculated from eqn. (7) if we assume a value for o;’(HCO). Provided that the difference coj’(HC0) -(uy(DCO) is small (<lo0 cm-I), the sum +[m;‘(DCO) +co;’(DCO)] is approximately constant (t 10 cm-1) and is dependent only on the value assumed for coi’(HC0) in eqn.(7). Thus, in calculating the zero-point energies for HCO and DCO in their ground and excited states, there is in effect only one unknown frequency, viz., coi’(HC0). We can now discuss the vibrational numbering for the principal progressions (0, u;, O)--(O, 0, 0) of HCO and DCO. If each progression is extrapolated to its zero band, then the difference in the frequencies of the two (000)-(000) bands must be consistent with the difference in the zero-point energies for the two molecules. It is necessary to consider only two alternatives, viz., Tioo = 8489 cm-1 for HCO and 8523 cm-1 for DCO as given by Herzberg and Ramsay, or Tioo = 9294 cm-1 for HCO and 9161 cm-1 for DCO as preferred in the present paper. According to the former assignment T~oo(HCO)-T&o(DCO) = -34 cm-1 while for the latter assignment T600(HCO) - Tim(DCO) = + 133 cm-1.Since all the relevant fre- quencies are known or are related to w;’(CH), we can use the two possible assignments to calculate the corresponding values for (ui’(C€€). On the basis of the Herzberg and Ramsay assignment we find co;’(CH)-4000 cm-1 which is clearly too large. For the revised assignment we find co;’(CH) -2700 cm-1 which is quite acceptable. To estimate the error in this determination we need to consider the various errors involved in the extrapolation of the principal progressions and in the neglect of the contributions from some of the anharmonic terms, which together are probably -10 cm-1. We then find that (u;’(CH),= 2700+100 cm-1. A value for 7’; may now be calculated to be 8690+ 50 cxn-1.J . W.C . JOHNS, S. H. PRIDDLE AND D . A. RAMSAY 101 The vibrational numbering thus shows that the upper state is derived from a n and not a X- or A-electronic state. Further confirmation of this conclusion is afforded by the vibronic structure discussed in the following section. : I I I I I I I VIBRONIC STRUCTURE A schematic diagram of part of the DCO spectrum is given in fig. 4. Two points are worthy of note. First, the II-bands are staggered to lower frequencies with respect to the E-bands. Second, in a group of levels with the same value of u;, L I I I I I I I I I the levels with higher K are found at lower frequencies and the separations between the levels are roughly quadratic in K, i.e., Thus the Z-A and II-4 splittings are equal to 4G and 8G respectively, while the staggering of the Il-bands with respect to the X-bands is equal to G.The experi- mental values of G for HCO and DCO are given in column 4 of table 6 . V: = V O - GK2. (8) FIG. 5.-Potential curves for the upper and lower states of HCO and DCO. r represents the normal co-ordinate for the bending vibration. We now calculate the vibronic structure expected if the observed splittings are due to electronic-vibrational interaction, is., to the Renner effect. We assume that the potential curves are of the form shown in fig. 5, and use the theory developed102 ABSORPTION SPECTRA OF HCO AND DCO by Pople and Longuet-Higgins.6 Following these authors we denote the upper potential curve by U+ = 3r2+hr4 and the lower curve by U- = (+-f)rZ+gr4, where f>$, g>O, and r is the normal co-ordinate associated with the bending vibra- tion.According to their theory, the energy levels of the upper state for small values of K (i.e., K<u2) are given by The quantity h is determined from the anharmonicity of the upper state while f and g are determined from the height of the barrier (8690 cm-1) valency angle (119" 30') in the ground state. For HCO and DCO we find Comparing eqn. (8) and (9), f = 1.515, 9 = 0.025, h = -0.0026. Evaluating this equation we find that G(theor.) = (:20:: -- 3.8) for HCO, and G(theor.) = ( tz4:y - - 3.0) for DCO. The theoretical values for G are given in column 5 of table 6. The experimental values are, on the whole, slightly smaller than the theoretical values, and show a trend towards lower values at higher values of 0;.The variation of G (obs.) with u;, however, is by no means regular and the irregularities are probably caused by resonances with the higher vibrational levels of the ground state. Such resonances can affect all of the levels of the excited state except the I: levels. These results are very similar to those which have been found for NH2 and ND2 (Dressler and Ramsay,s Eaton, Johns and Ramsay 15). The general agreement between theory and experiment, both in magnitude and sign, leaves little doubt that the observed vibronic splittings are due to the Renner effect. To be more rigorous, it is necessary to show that the K-dependence of the energy levels is greater than that given in the equation (Pople and Longuet-Higgins), which would be obtained in the absence of a Renner effect.Indeed, the observed K-dependence is roughly ten times greater in magnitude and opposite in sign to that given by eqn. (13). Hence we may conclude that HCO furnishes another example of the Renner effect, similar to that which was first established for NH2. An interesting point which arises is that there should be no (0, 0, 0) level in the excited state (see Dressler and Ramsay) ; the first level should be the (0, 1, 0) &level. It would be very difficult, however, to verify this point in absorption studies, because of the adverse Franck-Condon factor and the low frequency of the transition (-9000 cm-1). E + ( u ~ , K) = ((v, + l)++h[3(V2 + 1)2 - K 2 ] ) ~ 2 , (13) PREDISSOCIATION The observation of the (0, 3, O)--(O, 0,O) and (0, 5,O)-(O, O, 0) E-bands at high dispersion, but not of the (0,4,0)-(0,0,0) n-band, indicates that the (0,4,0)J .W. C . JOHNS, S. H . PRIDDLE AND D. A. RAMSAY 103 II-level is diffuse. Since this level is the lowest level which has been shown to be diffuse,” the dissociation energy of the molecule must be less than the energy of the (0, 4, 0) II-level, i.e., Do(HC0) < 12400 cm-1 (< 35.4 kcal/mole, < 1-54 ev). This value, which is an upper limit, is slightly lower than the limit given by Herzberg and Ramsay, but is still higher than the current chemical values of - 13 kcal/mole or -28 kcal/mole (see discussion by Cottrell 16). A correlation diagram showing the observed states of HCO and their probable dissociation productsis given in fig.6. cm-’ I There are no states of the products-between k ca I /mole H+CO *s +311 \ 40 000 20 000 200 I 5 0 I00 5 0 0 FIG. 6.-Correlation diagram between HCO and its dissociation products H+CO. Since the value for the dissociation energy of HCO is still in doubt, both sides of the diagram have been plotted relative to the same arbitrary zero level. the ground-state combination H(2S) + CO(lX+) and the combination H(2S) + CO(3II) at 48687 cm-1 ( G 139.2 kcal/mole, = 6.03 ev). The ground-state combination gives rise to a single state with species 2X+ or 2A’, depending on whether the molecule is linear or non-linear. This state is probably repulsive and accounts for the ob- served predissociation. The two known stable states of HCO presumably “want to dissociate ” into H(2S)+CO(W), though the problem of the crossing or the non- crossing of these potential surfaces with the repulsive potential surface needs to be considered.? While a complete discussion cannot be given at this time, two important points emerge.First, all the observed upper state levels except the X-levels are strongly predissociated. The probability of predissociation does not appear to depend strongly on the K-value (contrast the predissociation in HCN, Herzberg * The (0, 3, 0)-(0, 0, 0) A-band lies just beyond the limit of the present observations. t Teller 17 has shown that for a polyatomic molecule two potential surfaces of the same symmetry may cross in a special type of “ conical intersection ”. This behaviour is in marked contrast to that found in diatomic molecules where an “ avoided crossing ” results.104 ABSORPTION SPECTRA OF HCO AND DCO and Innes 18) since the rotational line widths in both the II- and A-levels for HCO and DCO are -20 cm-1.From the uncertainty relation Av (cm-1) = 1/2zcz, we deduce that the lifetime of the molecule in the excited state is -2.7 x 10-13 sec, which corresponds to the period of -27 C-H stretching vibrations. The second point concerns the E-levels which are sharp at low J-values and are presumably not predissociated. This sharpness can be understood if the C-levels are antisym- metric with respect to the molecular plane (Z- or A”) since the state formed from H(%S)+CO(lZ+) is symmetric (Z+ or A’), and symmetric states do not predissociate antisymmetric states.A breakdown of this selection rule, however, may be caused by the coupling of rotation to the vibronic motions of the molecule. Indeed, in the E-bands there is evidence that the rotational lines are slightly diffuse at the highest J-values observed (J- 15-20). Some slight broadening or doubling might be ex- pected due to the effects of spin uncoupling, but the widths of some of the lines at high J-values appear to be too great to be explained on this basis. A final point concerns the continuum which has been observed in the same region as the main absorption bands. This continuum might be due to the overlapping of numerous of the higher sub-bands. An estimation of intensities, however, suggests that the continuum is probably too strong to be explained on this basis. An alter- native mechanism which is more plausible, is that the continuum is due to a direct transition from the ground state to the repulsive 2Z+ (or 2A’) state. We wish to acknowledge the valuable assistance of Mr. W. Goetz with some of the experimental work. 1 Ramsay, J. Chem. Physics, 1953,21,960. 2 Herzberg and Ramsay, Proc. Roy. SOC. A, 1955,233,34. 3 Walsh, J. Chem. SOC., 1953,2292. 4 Ramsay, Adv. Spectroscopy (Interscience Publishers, New York and London, vol. 1, 1959), 5 Dressler and Ramsay, Phil. Trans., A, 1959, 251, 553. 6 Pople and Longuet-Higgins, MoZ. Physics, 1958, 1, 372. 7 Adrian, Cochran and Bowers, J. Chem. Physics, 1962, 36, 1661. 8 Johns, Priddle and Ramsay, to be published. 9 Birge, Rev. Mod. Physics, 1947, 19, 298. p. 29 ff. 10 Douglas and Sharma, J. Chem. Physics, 1953, 21,448. 11 Ewing, Thompson and Pimentel, J. Chem. Physics, 1960,32,927. 12 Herzberg, Infra-red and Ramun Spectra of PoIyatomic Molecules (Van Nostrand Co., Inc. 13 Polo, Can. J. Physics, 1957, 35, 880. 14 ref. (12), p. 231 ff. 15 Eaton, Johns and Ramsay, to be published. 16 Cottrell, The Strengths of Chemical Bonds (Butterworths Scientific Publications, London, 17 Teller, J. Physic. Chem., 1937, 41, 109. 18 Herzberg and fnnes, Can. J. Physics, 1957, 35, 842. New York, 1945), p. 426. 2nd ed. 1958), p. 185.