3444
J . Phys. Chem. 1989, 93, 3444-3448
Experimental results are nicely supported by INDO/S CI calculations for I (Figure 2), which reveal the lowest excited singlet state in perpendicular conformation to be indeed an intramolecular charge transfer state. Last but not least, we found an obscure, as yet, nonradiative channel of increasing importance with growing proton donating ability of the solvents and resulting in the lack of fluorescence of the N-protonated and N-methylated cations (the TICT fluorescence maximum of V evaluated from the far extrapolated linear correlation of Fma: vs E1,2r4(A)is about 13 500 cm-l, which
should be detectable with our spectrofluorimeter). Acknowledgment. We express our thanks to Dr. J. Waluk for his kind help in INDOIS calculations and for stimulating discussion. We are also indebted to Mr. J. Karpiuk for the determination of the lifetimes. We are really indebted to one of the referees for his keen and stimulating remarks. The work is done within the Polish research programs CPBP 01.19. and R P 11.13. Registry No. I, 31401-45-3; 11, 56864-93-8; 111, 24260-21-7; IV, 28942-78-1; V*I-, 2228-31-1;VI*[-, 118714-48-0.
The 320-nm Electronic Spectrum of Carbazole in a Jet by Two-Photon Ionization D. M. Lubman, Liang Li, and T. M. Dunn* Department of Chemistry, University of Michigan. Ann Arbor, Michigan 48109-1055 (Received: July 1 , 1988; In Final Form: November 3, 1988)
-
The electronic absorption spectrum of carbazole vapor has been obtained by using one-color resonant two-photon ionization (1 CR2PI) methods in an argon-driven jet. The spectrum has been analyzed from the 0; band (at 30 824 cm-’) to 1350 cm-] higher energy. The analysis confirms the system origin found by Bombach et al. and makes assignments of the various vibronic origins. The lCR2PI spectrum is compared with those obtained by LIF in a jet as well as spectra from low-temperature solid-state absorption and fluorescencestudies where the carbazole molecule is a guest in a fluorene host lattice. The problems of assigning the vibronic transitions obtained from all the above spectra and those obtained by 2CR2PI spectroscopy are discussed.
Introduction The electronic absorption spectrum of carbazole (C,,H,N) has been studied previously in both absorption] and fluorescence2 diluted in a fluorene matrix at -10-15 K, in a jet by 1C (one color) R2PI techniques3 and by LIF of jet-cooled ~ a p o r .These ~~~ last studies4qSwere quite extensive but their spectrum has a different origin than reported here. Thus, the origin was cited4 as being at 30694 cm-I whereas we find no band at that frequency. Bombach et a1.j show the origin of the carbazole molecule at the same frequency as that found in the present study but since their study was concerned with carbazole-hydrate clusters, they did not give a vibronic analysis of the parent system. Their results3 show that the parent spectrum and its polyhydrates have different vibronic structures and the details of the vibronic structure found in the present work are also different from that given by Auty et aL4 Experimental Section The experimental arrangement is similar to that used in previous work.6 It consists of a time-of-flight (TOF) mass spectrometer (R.M. Jordan Co.) mounted vertically in a stainless steel six-port cross and pumped by a 6-in. diffusion pump. The acceleration region of the T O F device is enclosed by a liquid N 2 cooled cryoshield and the flight tube is differentially pumped by a 4-in. diffusion pump. A pulsed supersonic molecular beam expands into the acceleration region of the TOFMS and a laser beam which is perpendicular to both the supersonic jet and the flight tube ionizes the sample. The supersonic jet is produced using a novel hot pulsed valve design which is described elsewhere.’ This valve is capable of operating up to 550 OC with a 330-ps pulse width. In the case of carbazole, an oven temperature of 150 O C was used in order to generate sufficient vapor pressure to obtain a strong RZPI signal. The reservoir pressure was Torr when the pulsed nozzle operated at a 10-Hz repetition rate. The orifice was 800 pm and
-
*To whom all correspondence should be directed.
the jet was interrogated 16 cm downstream in order to be in the “free-flow” region. It is important to correctly adjust the time delay between the nozzle pulse and the ionizing laser pulse in order to avoid spectral interference from potential van der Waals complexes. In this work, the time delay has been adjusted so that the ionizing laser beam intersects with the front portion of the gas pulse. Under this circumstance, we have found that the formation of van der Waals complexes is minimized. We saw no Ar-carbazole clusters in the mass spectrum. Another characteristic feature of van der Waals complexes in jet spectroscopy is that they show several peaks at a lower energy than the bare (parent) molecular spectrum. We found no features below the origin frequency (30 824 cm-I) of carbazole. Finally, we found no changes in the spectrum when the carrier gas was changed to carbon dioxide which confirms the absence of peaks due to van der Waals molecules in our spectrum. The laser source consists of the output of a Quanta Ray PDL-1A dye laser pumped by a DCR-3 Nd:YAG laser. Tunable UV radiation is generated by frequency doubling the output of the dye in a phase matched KD*P crystal. This is performed using the Quanta Ray WEX-1 wavelength extension device. The near-UV radiation which has an energy of -0.1-0.3 mJ at 10 Hz is then collimated with a telescope (positive lens, 30 cm focal length; negative lens 10 cm focal length) to a beam -2 mm in diameter. A Quanta Ray CDM-1 control display module is used to control the stepping motor which tunes the grating in the dye cavity. (1) Bree, A.; Zwarich, R. J . Chem. Phys. 1968, 49, 3355. (2) Bree, A.; Zwarich, R. J . Chem. Phys. 1968, 49, 3344. (3) Bombach, R.; Honegger, E.; Leutwyler, S. Chem. Phys. Lett. 1985, 118, 449. (4) Auty, A. R.; Jones, A. C.; Phillips, D. J . Chem. SOC.,Faraday Trans. 2 1986,82, 1219. ( 5 ) Auty, A. R.; Jones, A. C.; Phillips, D. Chem. Phys. 1986, 103, 163. (6) Ternbreull, R.; Lubman, D. M. Anal. Chem. 1987, 59, 1082. (7) Li, L.; Lubman, D. M. Reu. Sci. Instrum., in press.
(8) Kurahashi, M.; Fukuyo, M.; Shirnada, A,; Furusaki, A,; Nitta, I. Bull. Chem. SOC.Jpn. 1966, 39, 2564 (together with a private communication to the authors footnoted in ref 1).
0022-3654/89/2093-3444$0 1.5010 0 1989 American Chemical Society
The Journal of Physical Chemistry, Vol. 93, No. 9, 1989 3445
Electronic Spectrum of Carbazole
TABLE I: lCR2PI Spectrum of Carbazole band 6r, cm-' no. 1 2 3 4 5
6
I
8 9 10
11
12 13
14
15
16 17
P,
cm-l
30824 31015 027 196 234 240 300 32 1 332 342 383 390 421 442 447 465 476 53 1 540 550 590 600 637 668 (sh) 674 677 (sh) 684 720 732 737 740 792 798 (sh) 808 817 837 841 857 862 882 888 929 940 949 962 972 978 32004 016 025 03 3 04 1 049 066 072 097 105 114 147 158 177
assignment
(from 0:)
TABLE 11: Numbering and Classification of the Calculated Fundamentals of Carbazole* species i P,(calcd), cm-l species i e,(calcd), cm-l
OX
191 203 372 410 416 476 497 508 518 559 566 603 618 623 64 1 652 707 716 126 766 776 813 844 850 853 860 896 908 913 916 948 974 984 993 1013 1023 1033 1038 1058 1064 1105 1116 1125 1138 1148 1154 1180 1192 1201 1209 1217 1225 1242 1248 1273 1281 1290 1323 1334 1353
AI
21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
3436 309 1 3068 3046 3024 1648 1612 1535 1491 1419 1388 1302 1240 1213 1133 1041 857 759 629 409 233
A2
30 29 28 27 26 25 24 23 22
989 939 893 77 1 69 1 467 410 23 1 106
Y(A;j*
B2
60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42 41
3091 3068 3046 3024 1635 1623 1539 1511 1431 1424 1296 1252 1226 1152 1051 1001 847 620 559 472
B1
40 39 38 37 36 35 34 33 32 31
1089 935 895 831 737 567 433 342 222 108
are given in Table I together with assignments. The free carbazole molecule (C12H9N)belongs, as far as is known, to the point group C,. It has 60 vibrational degrees of freedom which are distributed among the four irreducible representations as 21A1 + 9A2 + 20B1 10B2. In this description we choose the molecular axes such that x(BJ is normal to the plane, y(BJ is the least axis of inertia, and z(Al) is the 2-fold in-plane axis passing through the nitrogen atom. The crystal structure shows the molecule to be slightly nonplanar (1.6' out of plane) but it is not known whether this is a crystal structure effect or a property of the isolated gas-phase molecule. In our analysis, we restrict the model to the planar (C,) form since there is no compelling evidence otherwise (see below). As noted by Bree and Zwarich,2 the spectrum is somewhat unusual in terms of the absence of long progressions and the presence of many origins which can only be analyzed as vibronic in character. They also performed a normal-coordinate calculation using a valence force field transferred from phenanthrene and, even though there are bound to be some differences between this and the real force field, many of their conclusions are undoubtedly correct, at least with respect to the number of low-frequency (25%) change in frequency. There is a sufficient prima facie case for this assumption from the work of Bree and Zwarich1s2 since neither their absorption nor emission spectra
+
3'X(B2) 1'48'(B2) 2'46'(B2) 1261(A1)
+ 12461(B2) + 122'43'(B2)
The flight time of the molecular beam which is -500 1.1s is properly synchronized to the laser pulse by using a delay generator to set the correct delay between the pulsed valve and laser. The laser then fires and triggers the oscilloscope and gated integrator. A Stanford Research System SRS-250 gated integrator unit was used to monitor the molecular ion peak in the TOFMS as a function of wavelength and the signal was displayed on a strip chart recorder. N o correction was made for dye laser intensity in the spectra so obtained.
Results and Discussion We have observed 5 0 well-defined and 23 less well-defined features of which 61 are listed here. Their frequencies (in cm-')
+
-
-
+
+
3446
Lubman et al.
The Journal of Physical Chemistry, Vol. 93, No. 9, 1989
Carbazole
Figure 1. The 1CR2PI spectrum of carbazole (C,*H9N).
showed the presence of long progressions (and, indeed, only a single progression of the lowest a l fundamental, to two quanta). As discussed below, however, we are much more circumspect in terms of gross reductions in force constant (and, consequently, frequency) in the out-of-plane a2 and bl modes. The spectrum is reproduced in Figure 1 where some of the bands have been arbitrarily numbered in order to facilitate comparison with the tabulated frequencies. It is clear that there are no extended sets of progressions in this 1350-cm-I interval apart from that of which the band at 3 1 027-cm-I (band 2) is the first member. This progression can be followed to at least three members and is also found in combination with at least one other vibronic origin (that at 31 332 cm-I). Since there are 12 bands (two of them very weak) in the 0-600-crn-’ interval and two of them are clearly identifiable as 1’ and l 2 with an upper-state fundamental frequency of -203 cm-l, the other seven must be totally symmetric vibronic origins or else vibronic origins of active b2 species. Reference to Table I1 shows that there are only two other possible a i fundamentals ( u 2 and u3) and three b2 ( ~ 4 1 ; and ~ 4 1 )which have any likelihood (ut supra) of being identified with the bands found in this range. In addition to fundamentals, there is also the possibility of vibronic origins arising either from two quanta of low-frequency excited-state a2 or bl modes (which have overall A, symmetry) or a combination of low-frequency excited-state a2 and bl modes with an overall B2 symmetry. While jet spectra have the advantage of greatly reducing the number and extent (rotational) of the vibronic structure in the spectra of polyatomic molecules, they suffer from the disadvantage that little or no evidence is obtained regarding sequence structure from which, knowing the groundstate vibration frequencies, one can deduce the excited-state frequencies of nontotally symmetric modes which are not Herzberg-Teller active in single quanta. In the case of planar aromatic hydrocarbons, there are usually some modes which suffer a catastrophic reduction in frequency in the excited stateparticularly those of the out-of-plane species-and which are, therefore, often active as overtones of their greatly reduced excited-state frequencies. The archetypes of this group are the very well-knowngJ0e2u( v ” ~=~ 298 cm-], dI6= 173 cm-I) and elg (v’’]~
-
= 846 cm-’, di0= 585 cm-’) of benzene and their analogues in p-difluorobenzene,IiJ2 a,, ( v ” ~= 420 cm-l, v ’ ~= 173 cm-I) and bl, ( v ” ~= 800 ern-', dg = 528 cm-I). All of these out-of-plane upper-state fundamentals form active overtone (totally symmetric) origins in their respective spectra with the characteristic that their “second” quantum (Le., 4 v ’ ~ )has a greatly reduced intensity vis-a-vis the “fundamental” (Le., 2 ~ ’ ~ ) . In traditional optical absorption and emission spectroscopy, the appearance of a vibronic origin of a totally symmetrical vibration having an intensity comparable to the origin signals a progression of several members, for reasons which are well understood. It is, therefore, easy to distinguish between the totally symmetric overtone and fundamental origins in terms of their relative Franck-Condon progression intensities, but this is no longer either appropriate or necessarily obvious in spectra of the R2PI type. Further, it is clear that there are significant differences in the relative intensities of bands which are quite close together in frequency when spectra of 1CR2PI and 2CR2PI types are compared.I3 This has to do with the ionization efficiency as a function of frequency and it can be expected-even if not always found in practice-to vary significantly over the several thousands of cm-’ characterizing the usual electronic excitations of large classes of organic and inorganic molecules alike. Hays et a1.,I5 in a multiphoton ionization study of 3-methylindole by both 1C and 2C methods, found that the ionization threshold of the 0-0 transition of the ‘Lb state is remarkably narrow (- 1.5 meV wide) and this kind of behavior must be considered as the more likely one in large molecules where there is little change in geometry or size between the ground and the excited states. The results found by Hager and Wallace13in their 2C study of the methylindoles confirm this expectation. Accordingly one cannot necessarily rely upon the traditionally expected regular Franck-Condon intensity variations in any analysis of an R2PI spectrum. Additionally, in most R2PI spectra there is a significant variation of the lasing efficiency of the dye ~~~
~~~~
( 1 1) Childs, A. F.; Dunn, T.M.; Francis, A. H. J . Mol. Spectrosc 1983, 102, 56.
(12) Zimmerman, R L.; Dunn, T M., unpublished data on p-C6H4F2
vapor. (9) Garforth, F. M.; Ingold, C. K. J . Chem. Soc. 1948, 417. (IO) Wunsch, L.; Metz, F.; Neusser, H. J.; Schlag, E. W. J . Chem. Phys. 1977, 66(2), 386.
(13) Hager, J. W.; Wallace, S. C. Anal. Chem. 1988, 60, 5 . (14) Lubman, D. M.; Li, L.; Dunn, T. M., to be published. ( 1 5 ) Hays, T.R.; Henke, W. E.; Selzle, H. L.; Schlag, E. W. Chem. Phys. Lett. 1983, 97(4,5), 347.
Electronic Spectrum of Carbazole over the range of spectral activity and this may further complicate the criteria for distinguishing between the different kinds of origins. Finally, in jet spectroscopy (R2PI or LIF), the characteristic rotational contour which would serve to indicate the direction of the transition moment is also lacking and this further constrains the amount of evidence available for distinguishing between the electronically and vibronically allowed components. Fortunately, in aromatic hydrocarbons and their simple derivatives, one may at least disregard the likelihood of transitions with out-of-plane transition moments. In this study we have been fortunate to have the polarized doped crystal fluorescence and absorption spectra’,* of carbazole and this has greatly assisted the current analysis since the B2(y) polarized transitions are strongly polarized along the c crystal axis and are essentially lacking in the b axis spectra while the opposite is true for the A,(z) polarization. Thus, the assignments given in Table I have been made in the light of the polarization data, the presence, rather than the regularity, of extended progressions and the approximate numerical data obtained by Bree and Zwarich in their force field analysis.
Totally Symmetric Fundamental Origins We have identified seven of these ranging in frequency from -200 cm-l (v”) to 993 cm-’ (v’,). The v’’ vibration forms a progression of at least three members, with a possible fourth member, blended with a stronger band. This mode is also observed in combination with all of the other a , fundamentals and, in fact, with most of the origins of b2 or overtone in type. Comparison with the solid-state spectrum reveals two significant differences with the lCR2PI spectra involving J1. First of all, the electronic origin in the solid-state spectrum is by far the strongest band in the spectrum while the R2PI spectrum shows 1’ (band 2) to be somewhat stronger than the 0;. In fact, l 3 was not observed in the matrix spectrum, although it would have been obscured by the first quantum of their 648 cm-’ (a,) fundamental (our J3with a frequency of 641 cm-I). Second, there is a considerable difference in frequency of d l in the solid-state and R2PI spectra with the former 218 cm-l and the latter 203 cm-’ for the first quantum. Despite these differences there does not seem to be any doubt that these are the same modes. It is also of interest that, in the spectrum analyzed by Auty et al.,435J1is 21 1 an-’( v f l is not given by Bombach et aL3) despite the fact that the van der Waals molecule is a “mass loaded” carbazole. The second quantum of J 1is almost coincident with vf2and there may well be some Fermi interaction, although that this is minimal is indicated by the fact that the second quantum of v f l has a higher frequency than the first (207 cm-I) while the third quantum has an even higher value of 213 cm-I. The presence of a “negative” anharmonicity is completely consistent with the vibration form of this lowest a, fundamental, as illustrated by Bree and Zwarich.’ The band at 31 240 cm-I is identified as 2’ (band 3) in agreement with its displacement of 416 cm-’ from the origin compared with the (very much weaker) equivalent band in the solid-state spectrum at 0; + 418 cm-’. The combination band 1121is displaced 202 cm-l from 2,. No trace of 22 was found and the principal reason for assigning the 416 cm-’ interval as d2 is based upon its appearance with the appropriate polarization in the solid-state spectrum, coupled with the fact that other bands in the spectrum involving both a , and b2 fundamentals form combination bands with it (31 737, 31 798, 31 808, 31 882 cm-I, etc.). There is, however, no compelling reason not to assign it as the overtone of a nontotally symmetric mode except for its frequent occurrence in combination with other modes. Many of the same comments apply to the identification of 3, (band 6) as involving v ’ ~ .The band 32 is not seen despite the very strong appearance of its analogue in the solid-state spectrum. There is, indeed, an extraordinarily weak band displaced by 1281 cm-’ from 0; but it is difficult, if the ordinary Franck-Condon considerations apply, (ut supra) to so designate it, at least in this study. On the other hand, the 641-cm-’ interval appears frequently in combinations and it has, therefore, been assigned as an excited-state fundamental.
The Journal of Physical Chemistry, Vol. 93, No. 9, 1989 3447 The same kinds of considerations apply to the assignments of 4’, 5l, 6I, and 7’.
Overtone Origins In their solid-state analysis, Bree and Zwarichl*2remarked upon the fact that the carbazole spectrum has no long progressions but, rather, a very large variety of a , and b2 fundamentals. In the solid state Bree and Zwarich assigned 10 a l and seven bl origins in fluorescence and seven a l and six b2 origins in absorption in a 1350-cm-I range from the electronic origin. Almost the same situation holds for the R2PI spectrum. The multiplicity of origins and the lack of long progressions point to the importance of Duschinsky in this molecule; Le., despite the fact that the upper and lower electronic states are very similar geometrically, the significant force field difference between the two requires a renormalization of any set of normal coordinates calculated for either the upper or lower state separately. All of the b2 vibronic and most of the a l origins are weaker than the electronic origin, which supports this general interpretation, rather than any major, or even relatively minor, change in geometry (shape or size) between the ground and excited state. There are, however, in our spectrum, more bands with a displacement from the origin e 6 0 0 cm-’ than can be accounted for as b, fundamentals or as low-frequency a l modes. In accordance with assignments made in similar we assign a number of these vibronic origins as the overtones of nontotally symmetric modes of the out-of-plane species a, and b,. Very low frequency shifted bands of this type would be obscured in the solid-state spectra by Iattice/phonon structure and are, therefore, more easily seen in R2P1 or LIF jet spectra. Perusal of Table I1 indicates that there are four a2 modes and five bl modes with ground-state frequencies
