4TjRR
COMMUNICATIONS TO THE EDITOR
although the absolute magnitudes of these coupling constants are substantially lower than usual. This analysis of the n.m.r. spectrum, together with the fact that full hydrogenation yielded cyclotetradecaneI2 uniquely defines the structure and stereochemistry of the substance as 11. An X-ray crystallographic analysis of 1,8-bisdehydro [14]annulene (11) shows the crystals (from $hloroform) to be monqclinic, a = 8.685 f 0,012 A., b = 7.789 f 0.008 A., c = 8.086 f 0.010 A., /3 = 113’ 48‘; the space group is P21/c (CP,), which, with two molecules of C14H10 per unit cell, corresponds to a density of p calcd. = 1.182g.,/cc. The unit cell dimensions and space group require that the molecule is centrosymmetric in the crystal. The crystal structure has been determined by trial and error, Fourier and least square methods, and fully confirms structure I1 (see Fig. 1). Accurate bond lengths and bond angles cannot, as yet, be reported, since rotational disordering of the molecules in the lattice, as indicated by diffuse scattering and large calculated Debye factors, has not been explicitly considered in the refinement analysis. X three-dimensional analysis of the crystal structure a t 90’K. is being undertaken to obtain more accurate bond distances, and will be reported later (by N.A.B. and R.M.). At this stage, i t is already appacent that all the carbon atoms are coplanar ( f 0.03 A ) . The striking difference in the chemical shifts of the inner and outer protons in the n.m.r. spectrum, together with the over-all geometry of the molecule as determined by the X-ray crystallographic analysis, clearly establishes 1,8-bisdehydro[14]annulene to be aromatic. This is in keeping with expectation, since Hiickel’s rule is obeyed and the carbon skeleton is coplanar. The substance proved to be unusually stable; e.g., i t was completely unchanged after being kept for 1 month in the solid state at room temperature without protection from daylight. Also noteworthy is the position of the highest wave length maximum in the ultraviolet a t 586 mp (in isooctane) . z Various aspects regarding the properties and synthesis of 1,8-bisdehydro[14]annulene are now under investigation. These include magnetic and theoretical studies of the electron distribution in the molecule, as well as the nature of the dehydrogenation involved in its formation.
VOl. 84
been suggested2 that line broadening techniques could also be used with various kinds of optical spectroscopy to measure the rates of even faster reactions. This has now been achieved, using Raman spectra. The range of usefulness seems to be for processes with half times from ~ l 0 - lto~ sec. The latter probably is outside the range of chemical interest but the former is not, and for such rates the present method seems to be, by far, the most convenient when it is applicable. The present communication describes its application to the rate of proton transfer to trifluoroacetate ion from trifluoroacetic acid, probably via the hydronium ion, in aqueous solution. Theory.-The Raman transition is assumed to involve degrees of freedom of the molecule which do not take part in the exchange. The effect of the exchange is to cause a time-dependence of the effective Hamiltonian “seen” by the portion of the molecule of interest: Res = H , when the molecule is in form n (n = 1, 2). A representation is used in which HI is diagonal, and off-diagonal elements of (HI - Hz)are neglected. The problem can now be treated by means of a simple generalization of the methods of Heitler. In second-order time-dependent perturbation theory, the probability amplitude, bf, for a transition in which the molecule changes its state and the angular frequency of a photon is decreased by v obeys the equation 394
in&
=
Kfo(t)exp[i
sd
wro(t’)dt’]
- i&P(t)bi
(1)
Eq. (1) is to be compared with Heitler’s eqs. (12) and (13).3 The last term represents the broadening due to the natural line width. Otherwise, the only difference is in the time-dependence of Kro and W F due ~ to the exchange. When the molecule is in form n, K f o ( t ) K,, wro(t) fW n - v, p(t) Pn. An explicit expression for Ki, is given by Heitler’s eq. (7).4 The Raman intensity a t frequency v is given by eq. (2). I ( v ) d v = 2Adv lim @ ( t ) lbi ( t ) i2) t
(2)
dm
Here ( ) indicates an average and A is a constant. I n the present case, the molecule jumps randomly back and forth between forms 1 and 2. If T~ is the mean time of a single sojourn in form n, 1 , ’ ~is’ kl, the (pseudo) first-order rate constant for the DANIELSIEFFRESEARCH INSTITUTE reaction (form 1) +-(form 2). WEIZMANN IXSTITUTE OF SCIENCE F. SOXDHEIMER Equations (1) and ( 2 ) have been solved for the REHOVOTH, ISRAEL Y. GAONI Tn >> 1, case of relatively slow exchanye: IWI DEPARTMENT OF CHEMISTRY and all Prn.m >> 1. Since ,a1 - w2i may be as L. &I. JACKMAN IXPERIALCOLLEGE OF AT. A. BAILEY large as several hundred cm.-’, this approximation SCIENCE AND TECHKOLOGY R . MASON should hold fairly well until T’ and 7 2 approach LONDON S. W. 7 , ENGLASD RECEIVED OCTOBER 6, 1962 lo-’* sec. The result is eq. (3). Simplification MEASUREMENT OF VERY FAST REACTION RATES BY RAMAN LINE BROADENING
Sir: Line broadening measurements in magnetic resonance spectroscopy have been widely used to measure rates of fast chemical reactions.’ I t has (1) H. Strehlow, in “Investigation of Rates and Reaction Mechanisms, P a r t 11.” S. L . Friess, E S . Lewis and A. Weissberger, Editors, Interscicnce Publishers, Inc., New York, N . l’.,1962, Chapter 7.
+
resultsif iK2!%>Pn I/T”. In these cases i t is shown readily that eq. (-4) gives (2) M. Eigen, private conversations. (3) W. Heitler, “ T h e Quantum Theory of Rad~ati,,n.”3rd editicin, Clarendon Press, Oxford, 19.54, p p . I3F-1.45. (-1) Illid., pp. 189-196.
Dec. 5, 1962
COMMUNICATIONS TO THE EDITOR
4507
dence supports the interpretation that the broadT~ in terms of p1and 0, the half width a t half height of the broadened line. ening is due to exchange. 1/71 = 8 - 81 (4) (8) Alfred P. Sloan Foundation Fellow, 1960-1964. Application.-Raman spectra of sodium trifluoro- DEPARTMENT OF CHEMISTRY acetate ion in the presence of excess sodium hy- UNIVERSITY OF MINKESOTA MAURICE M. KREEVOY~ c. ALDEN MEAD 14, MINNESOTA droxide show an intense peak a t 1433 cm.-l (prob- MINNEAPOLIS RECEIVED SEPTEMBER 4, 1962 ably the symmetrical C=O stretching f r e q ~ e n c y ) ~ with a width a t half height of 15.0 an.-'. In the presence of excess base the width is independent GLYCEROL AS A PRECURSOR OF RICININE' of trifluoroacetate concentration. Trifluoroacetate Sir: also has a fairly intense peak a t 843 cm.-', with By means of tracer experiments with young a width of 16.2 cm.-'. Trifluoroacetic acid in the presence of excess HC1 gives a strong peak a t 816 Ricinus communis L. we have recently established2s3 cm.-' with a width of 28.8 cm.-l. The latter two that succinic acid, or a closely related four-carbon are presumably the C-C stretching frequencies.j dicarboxylic acid found in the Krebs tricarboxylic I n solutions containing trifluoroacetic acid, its acid cycle, is a specific precursor of carbon atoms 2, conjugate base, and hydronium ion, the trifluoro- 3, and 7 in the biosynthesis of ricinine I. Since an origin outside of the Krebs cycle was acetate peak a t 1433 ern.-' is broadened by as much as 7.2 cm.-', and the two peaks around 830 indicated for carbon atoms 4,5, and 6, g l y ~ e r o l - l - ~ ~ C cm.-l both broaden and tend to merge. Both and glycer01-2-'~C were fed in separate experithe 1433 cm.-l peak and the 816 cm.-' peak also ments to young Ricinus plants. The percentages tend to shift toward higher frequency by 2-5 of the incorporated radioactivity located a t the 0cm. in mixed solutions. Spectra were measured methyl, N-methyl, and nitrile carbon atoms of the on a Cary model 81 photoelectric Raman spectro- isolated ricinine have been reported already.2 We now wish to describe degradative procedures photometer, equipped with two Toronto-type mercury arc lamps. Solution temperatures were to isolate the carbon atoms a t positions 4, 5, and around 40'. The lines are roughly Lorentzian in 6 of the ricinine molecule, and applications of these procedures to the ricinine obtained from both shape. C glycerol-2-14C, as well as, in The broadening of the 1433 cm.-l peak was g l y ~ e r o l - l - * ~and used to obtain k1 for trifluoroacetate ion by means part, to ricinine from the previously reported sucof eq. (4). These results are shown in Table I. cinic acid-2,3-I4C and sodium acetate-2-I4Cfeeding Solutions were made up simply by mixing tri- experiments. Ricinine I was converted to 4-methoxy- l-methylfluoroacetic acid and water. Concentrations were obtained by interpolation, using the degree of dis- 2-pyridone 11, m.p. 114-116°,4 which was reduced lithium aluminum hydride to l-methyl-4sociation results of Redlich and c o - ~ o r k e r s . ~ ~with ' Concentrations could also be inferred from the in- piperidone 111, b.p. 55-60' (12 mm.),5 identified tensity of the 1433 cm.-' line in the Raman spectra. by comparison with an authentic sample. ReacThe average difference between the two estimates tion of I11 with phenyllithium provided the known IV, m.p. of the trifluoroacetate concentration was 10%. 4-hydroxy- 1-methyl-4-phen~lpiperidine~ This strongly suggests that the intensity of the 115-116', which was oxidized with potassium per1433 cm.-l band is proportional to the trifluoro- manganate to benzoic acid V, m.p. 121-122', with acetate concentration and otherwise independent the carboxyl carbon atom originating from carbon of the composition of the solution. Table I also 4 of ricinine. Nitration of ricinine I afforded the 5-nitro derivgives the rate constant k-l for the conversion of trifluoroacetic acid, as determined from the re- ative VI, m.p. 163-164' (anal. Calcd. for CsH704N3: C, 45.94; H, 3.37; N, 20.09. Found: C, quirement that equilibrium be maintained. 45.94; H, 3.39; N, 20.19) which on treatment TABLE I with calcium hypobromite yielded tribromonitroAVERAGE LIFETIME OF TRIFLUOROACETATE IONIN VARIOUS methane (bromopicrin) VII, with the carbon atom SOLUTIONS originating from position 5 of ricinine. The bromoMolar 10-11 ki 10-11 k-1 picrin was identified by comparison with an CFaCOzCFsC02H Hf sec.-1 sec. -1 authentic sample prepared from picric acid.7 1.8 3.5 1.8 6.9 3.5 Catalytic reduction of the pyridone I1 yielded 1.5 1.5 1.5 4.7 4.7 1-methyl-2-piperidone VIII, b.p. 110-105° (12 mm.) 0.81 0.24 0.81 1.6 5.4 (mercuric chloride derivative, m.p. 117-120°8) which was hydrolyzed with hydrochloric acid, The rate constants cited above were used to (1) N.R.C. No. 7107. calculate the shape of the doublet centering around (2) P. F. Juby and Leo Marion, Biochem. Bioghys. Res. Comm.. 6, 830 cm.-l. The general shape of the band is 461 (1961). moderately well reproduced. The band shape ( 3 ) P . F. Juby and Leo Marion, Can. J . Chem., in the press. (4) E . Winterstein. J. Keller and A. B. Weinhagen, Arch. Pharm., calculated by adding the unbroadened lines is 513 (1917). worse, but the difference is not great. This evi- ass, (5) R . E. Lyle, R . E. Adel and G. G. Lyle, J . Org. Chcm., 14, 342 ~
~~~~
7
(5) R . E . Robinson and R . C. Taylor, Sgectrochimico A d a , 18, 1093 (1962). (6) 0. Redlich and G . C. Hood, Discussions Faraday Soc., No. 2 4 , 8 7 (1957). (7) G. C. Hood, 0. Redlich and C. A. Reilly, J . Chem. P h y s . , as, 2229 (19.5.5).
(1959). (6) Aktieselskabet "Ferrosan," Danish Patent 60,592 (1943) : Chem. A b s . , 40, 408@ (1946). (7) J. Baddilep, G. Ehrensvlrd, E. Klein, L. Reio and E . Salriste, J . Biol Chcm., 183, 777 (1950). (8) C . R l t h , A n n . , 489, 107 (1931).
