HERZAND GLICK
2970
solution of ethanolic sodium ethoxide, and 7.5 ml. of absolute ethanol. The resulting solution was refluxed for 20 hr. under nitrogen. At the end of this period 5.6 mg. (0.093 mmole) of glacial acetic acid was added to neutralize the base. Evaporation of the solvent under reduced pressure left a small quantity of liquid which was extracted with two IO-ml. portions of ether. Drying of the ethereal solution over magnesium sulfate and evaporation of the solvent gave 10 mg. (43%) of a colorless oil. Gas-liquid chromatography of the oil on a column containing 20’30 XF-1 150L6on 80-100-mesh Chromosorb P showed two ~
(16) A nitrile-silicon polymer obtained from General Electric.
VOL. 28
peaks in the ratio of 5 to 1. The lesser component was eluted first and had the same retention time as 15A. The larger peak had the same retention time as 15B. Diethyl Bicyclo[2.2.2]octane-l,3-dicarboxylate (16B) from Diethyl Bicyclo [2.2.2]0ct-S-ene-l ,endo-3-dicarboxylate ( 15D).A 12.5-mg. sample of 15D (containing 4.2y0 of 15B) was hydrogenated by platinum oxide in ethanol (1 ml.). Removal of the catalyst and evaporation of the solvent under reduced pressure gave 12.3 mg. (98%) of a colorless liquid giving two peaks on a g.1.p.c. column containing 25Y0 7lOsilicone oil on Chromosorb P. The two peaks were in the ratio of 23 to 1 . The major compo. nent had the same infrared spectrum and retention time as 16B-
The Lactones of cis- and trans-2-HydroxycycloheptaneaceticAcid WERNERHERZAND LAVERNE A. G L I C K ~ Department of Chemistru, The Florida State University, Tallahassee, Florida Received May 6, 1963 The cis- and trans-lactones of 2-hydroxycycloheptaneacetic acid have been prepared and equilibrated a t various temperatures. The cis isomer predominates slightly in the temperature range 384-423 OK. The thermodynamic quantities have been calculated and the conformations are discussed and contrasted with those of the cyrlohexane analogs. Two examples of what appears to be partial cis-electrophilic addition to the cycloheptene double bond have been observed.
The lactone of cis-2-hydroxycyclohexaneacetic a ~ i d ~ (I) is more stable than the trans isomer (II),’the trans compound being convertible to the cis under the influence of sulfuric acid-acetic acid. 3 , 6 An analogous acid-catalyzed rearrangement is well known in the santonin series where the allylic position of the lactone ether oxygen renders the isomerization more facile,* but many other iiistaiices of this stability relationship among lactones of substituted 2-hydroxycyclohexaneacetic acid could be cited.g In the last decade, a group of new sesquiterpene lactones has been discovered in which the y-lactone ring is fused onto the seven-membered ring portion of perhydroazulene skeleton. 10 I n discussions dealing with the stereochemistry of the new lactones, it has been
I11
Qo
a 0
IV
v
QO
VI
~~
(1) Supported in p a r t by a grant from the National Science Foundation ( N S F - G 14396). (2) Abstracted from a thesis subniitted in partial fulfillment of the requirements for the degree Doctor of PhilosoDhy, 1963. (3) M. Y. ru‘ewman a n d C . A. Vander N-erf, J . A m . Chem. Sac., 6 7 , 233 (1945). (4) J. Klein. J . O r g . Chem., IS, 1209 (1958). ( 5 ) J. Klein, J . A m . Chem. Soc., 81, 3611 (1959). (6) E. H. Charlesworth, H. J. Canipbell, and D. L. Stachiw. C a n . J . Chem., 57, 877 (1959). (7) 9. Coffey. R e c . trar. chzm., 42,387 (1923). (8) Y. Abe, T . Miki. hI. Sunii, and T. Toga, Chem. Ind. (London), 953 (1956); H. Ishikawa, J . I’harm. Sor. J a p a n , 7 6 , 504 (1956); XI. Sutni. J . A m . Chem. Soc.. 80, 4869 (19.58); I). H. R . 13arton, J. E. D. Levisalles. a n d J. T . Pinhey. J . Chem. S o c . , 3472 (1962): W. Cocker, 1%. Donnelly. H . Gobinsingli. T . 1%.H.Mc.\furray. and .\I. A . Nisbet, zhzd., 1262 (1963). (9) See, f o r exaiiiple. 31. Hinder and .\I. Stoll. Hrlr,. Chim. A d a . 96, 1495 (1953): FV. Klyne. J . Chem. S o r . 3072 (1953). (10) T h e most recent reviea of tllis railidly moving field is already very much o u t of date, T . Nozoe and S. Ito. F o r t s c h r . Chem. O r e . .Valurstoi7ek, 19, 32 (1961).
- tacitly ~ assumed’ that commonly accepted generalizations about y-lactones fused onto six-membered rings can be extended to ?-lactones fused onto sevenmembered rings. Because of our interest in naturally occurring perhydroazulenic lactones, we decided to examine this supposition by the synthesis and study of the previously unreported title compounds 111 and IV. The energetics of cis- and trans-ring fusion to cycloheptanes have, in the meantime, been considered by Hendrickson,12 with particular reference to the cisand trans-bicyclo [5.3.0]decanes (perhydroazulenes). The conclusion was reached that the energy difference between cis- and trans-perhydroazulene was likely to be virtually negligible, experimental support for this having been provided by Allinger and Zalkow. l a When a lactone is substituted for a five-membered ring, inspection of Dreiding models suggests that in the cis isomer I11 the favored twist-chair conformation12 of the cycloheptane ring may be destabilized somewhat because of angle strain. Of the two trans-forms of IV, the 2e-3e isomer appears to be affected only slightly, the 3e-4e isomer more so. The over-all effect is difficult to assess but would not be expected to alter the stability relationships significantly. Lactone I11 was prepared from VI, or more conveniently from the mixture of ITand VI prepared by cyclization of 2-oxocycloheptaneacetic acid. Lactone I V was synthesized in a maimer similar to that adopted by Xewman and Vander Werf3 for the preparation of 11. However, the first step, the reaction of cycloheptene oxide with maloiiate ion, was exceedingly slow as compared with the analogous reaction of cyclohexeiie oxide which reacts a t least 10 times as rapidly. The reasons for this difference in reactivity are not quite clear. It has been shown that cyclohcxeiir oxide reacts with methoxide ion about l..i times as fast as (11) See, for example. J . IV. €Iufftnan. P ; . r p r i i e i i t i a , 16, 120 (1Y60). T h e conildications have been recowiaeii by .J. 1%. Hendrickson a n d R . Rces, Chem. Ind. (London), 1424 (1962). (12) J. 13. Hendrickson. J . A m . Chem. SOC.,83, 4537 (1Y61). (13) N. L..illinper and V , 1 3 , Zalkow, *had., 68, 1144 (1961).
XOVEMBER, 1963
LACTONES OF
Cis- AND bNt7"r2-HYDROXYCYCLOHEPTANEACETIC
cyclopentene oxide,I4 but no data are available on the reaction of cycloheptene oxide. However, a number of reactions are known in which cyclopentane and cycloheptane derivatives react a t comparable rates which may be greater or less than the rate of the corresponding cyclohexane derivative.I6-l7 The greater susceptibility of cyclohexane oxide to ring opening compared with that of cyclopentene oxide may be attributed to greater relief of eclipsed hydrogen interactions in the cyclohexane oxide system. A further effect is probably operative in cycloheptene oxide where Dreiding models indicate that there may be considerable steric hindrance to nucleophilic displacement of epoxide oxygen by the bulky malonate ion which would result in rate retardation. Nuclear Magnetic Resonance Spectra.-Since the main objective of this work was the determination of the relative thermodynamic stabilities of lactones I11 and IV, a method for the analysis of mixtures of I11 and IV was necessary. In spite of prolonged and tedious efforts, artificial mixtures of I11 and IV could not be separated satisfactorily by gas liquid chromatography. Also, while the infrared spectra of I11 and IV were different, the distinguishing bands overlapped and the analysis by this method would have been difficult indeed. A more than adequate solution to this problem was offered by n.m.r. spectroscopy. Table I lists important n.m.r. peaks of lactones I-IV, whose implications will now be discussed. TABLE I" Lactone
CvHb
HaC
CHid
2.45m 1.2-2.0 e, 1.55 s I 4 . 6 0 q (4) 1.33-2.04e I1 3.97 h (10,4) 2.45n,e 2.66m 1.2-2.0e I11 4.750 2.42m 1.62s IV 4.25n Spectra were determined in CCI, solution on Varian HR-60 or A-60 spectrometers. Values in p,p.m. relative to tetramethylsilane as internal standard. Signals are described as follows: e, envelope; h, sextet; m, multiplet of uncertain multiplicity; n, unresolved inultiplet ; 0 , octet; s, relative sharp signal corresponding to several protons. Numbers in parentheses denote coupling constants in c.p.5. Intensity one proton. Intensity Ring protons other than Hg. two protons.
C2-H of I gives rise to a quadruplet a t 4.60 p.p.m. characteristic of the A portion of an AX3 system. The multiplicity indicates that the three adjacent protons have approximately the same dihedral angle, a situation which would prevail in a chair conformation in which C1-H is axial and CBH is equatorial. The dihedral angles are then near 60'. This is in reasonable agreement with experimental values observed elsewhere.I8 However, the observed multiplicity of C2-H could arise equally well from the time-averaged values of dihedral angles in a system which changes conformations rapidly compared with the spin frequency. Since the methylene region indicates a fair degree of flexibility
ACID
2971
(vide infra), it is not possible to decide between the two alternatives. The two a-protons give rise to a complex system of bands centered a t 2.45 p.p.m. which appears to contain two triplets a t 2.51 (J = 3.5) and 2.38 p.p.m. (J = 2.2). The methylene region contains an envelope from which there projects a relatively sharp high-intensity band a t 1.55 p.p.m. This suggests a certain amount of methylene proton equivalence and a reasonable degree of ring flexibility as would be expected from a cis isomer. By contrast, the n.m.r. spectrum of I1 exhibits only a broad envelope extending from 1.33 to 2.04 p.p.m., but no strong projecting peak. This is consistent with the assumption of a more rigid ring system. The situation is reminiscent of that prevailing in cis- and transhydrindane,lg with the lactone ring of I and I1 replacing the alicyclic five-membered ring. The aprotons of I1 are unresolved, but C2-H gives rise to a sextet centered a t 3.97 p.p.m. which corresponds to X of an A2BX system ( J A X = 10). This indicates that I1 is a chair in which C1-H and CTH are axial, in accordance with the required trans-diequatorial fusion of five- and six-membered rings. The shielding of CTH in 11, as compared with CTH of I, is as expected; the coupling constants are in agreement with the assumption that the dihedral angles between C r H and the three adjacent protons are approximately 180, 180, and 60'. Comparison of the n.m.r. spectra of I and I1 with those of I11 and IV reveals highly significant differences which point out the danger of drawing analogies between bicyclo [5.3.0] and bicyclo [4.3.0] ring systems. C r H of the cis-fused lactone I11 gives rise to a complex multiplet centered a t 4.75 p.p.m. which can be analyzed as an octet closely approximating X of an ABCX system where J A x = 10, J B X = 7, and JCX = 4 C.P.S. Dreiding models suggest that the most likely conformation is a chair in which C1-H and CTH are quasi-axial and the five-membered ring is quasi-diequatorially oriented. This should give rise to dihedral angles of about 0, 82, and 158°.12 A better fit is obtained by the twist-chair conformation advocated by Hendrickson12; this is shown in the appended formulae where the dihedral angles are markedS20 I11 is quite rigid as shown by the envelope in the 1.2-2-p.p.m. region. The aprotons give rise to a multiplet which could not be analyzed satisfactorily.
H-H la-2a
2a-3a 2a-3@
angle 41 97" 143'
JH-11
7 4 10
8
( 1 4 ) G . Gee, W.C. E . Higginson, P. Levesley, and K. J. Taylor, J . Chem. s o c . . 1338 (1959), (1.5) K.L. Allinrer, J . A m . Chem. S o c . , 81, 5727 (1959). (16) I f . C . 1 1 r o ~ n .K. S. Fletcher, and R. 13. Johannesen, ibid., 79, 212
(1951). (17) I
