2198
COMMUNICATIONS TO THE EDITOR
Vol. 83
such as relative radical stability and ring strain in corresponding 1-ketodicyclopentadienes support dictating direction of cleavage. Rate of cleavage this prediction; we now report that optically ttcof n-alkyl vs. methyl is greater than 10 to 1 (no. tive deuterium-labeled (at the aldehyde hydrogen) 1, 4, 5). Rate of cleavage of sec-alkyl (isopropyl) methacrolein dimer,GI, rearranges to I1 i n a similar exceeds n-alkyl (ethyl) by 20 to 1 (no. 1). This manner, under conditions where there is substantial latter order is reversed when cleavage of "n-alkyl" competitive formation of methacrolein by the reis part of a five-membered ring, I'.c , rate of cleavage verse Diels-Alder reaction. possible coursc of of isopropyl vs. cleavage of one of the two equivalent the rearrangement is ring carbon-carbon bonds in l-isopropylcyclopentoxyl (no. 2 ) is 1 to 14. The bulk of this difference probably is associated with relief of ring strain in the five-membered ring (6 kcal. per mole from heat of combustion data). Hydrogen Abstraction4 vs. Fragmentation.-Examples 4, 5 and 7 indicate that intramolecular ? 111 hydrogen abstraction may compete successfully with fragmentation. The preferred point of abstraction is four carbon atoms removed from the oxygen atom.5 (Intermolecular attack on cyclohexene by the alkoxyl radicals of Table I competes poorly with 1,5-intramolecular hydrogen abstraction or fragmentation'.) Rate of 1,5-intramolecular abstraction of secondary C-H (per hydrogen) DS. fragmentation of n-alkyl (butyl) is 2.7 t o 1 (no. 4); rate of 1,5-abstraction of primary C-H (per H) vs. fragmentation of neopentyl is 1 to 16 (no. 5). If one makes the assumption that the rate constant for abstraction of hydrogen from cyclohexene is the same for the alkoxyl radicals in IV examples 4 and 5, then the data may be extended to indicate a preference for 1,5-abstraction of secondary C-H vs. primary C-H of 10 to l6and a preference for cleavage of neopentyl vs. n-butyl of 4 to 1. Acknowledgment.-We are indebted t o the National Science Foundation and the Sloan Foundation for support of this work. (4) See C. Walling and B. B. Jacknow, J . A m . Chcm. Soc., 82, 6108, 6113 (1960), for a study of intermolecular hydrogen abstraction by t h e Icy!-butoxyl radical (hydrocarbon chlorination by fer!-butyl hyporhlorite). (5) Compare with t h e Hofmann-Lamer-Freytag reaction, E. J. Corey and W. R. Hertler, J. A m . Chcm. SOC.,8 2 , 1657 (1960). (6) Compare with intermolecular abstraction b y I&-butoxyl of tertiary C-H us. secondary C-H vs. primary C-H of 44 t o 8 to 1 (ref. 4).
FREDERICK D. GREENE DEPARTMENT OF CHEMISTRY MAXINEL. SAVITZ INSTITUTE OF TECHNOLOGY HANSH. LAU MASSACHUSETTS CAMBRIDGE 39, MASSACHUSETTS FREDERICK D. OSTERHOLTZ WILLIAMN. SMITH RECEIVED MARCH22, 1961 MECHANISM OF THE DIELS-ALDER REACTION'
Sir:
Species I11 might be the transition state or else The thermal interconversions of a- and p-1- an intermediate, but in either case its geometry is hydroxydicyclopentadiene with syn- and anfi-8- formulated as being highly restricted.' T h e overhydroxydicyclopentadiene, respectively, with re(6) Prepared from inactive unlabeled methacrolein dimer b y a tention of optical activity2 suggest t h a t other Cannizzaro oxidation t o the corresponding acid,"** then resolution Diels-Alder dimers3 might undergo analogous rear- through the brucine salt, conversion of t h e sodium salt of t h e acid t o acid chloride with oxalyl chloride: formation of t h e corresponding rangements. The conversions of S-ketodicyclo- the S-methylanilide and reduction with lithium aluminum deuteride.d pentadiene4 and a chlorinated derivative5 to the R. R. Whetstone, U. S. Patent 2,479,283 (1949); G. G. Stoner (1) Supported in part by t h e Office of Naval Research. ( 2 ) R. B. Woodward and T. J. Katz, Tetrahedron, 5, 70 (1959). (3) Products of Diels-Alder additions in which a single component serves both as diene and dienophile. (4) R. C. Cookson, J. Hudec and R. 0. Williams, Tcrruhcdron Lellys, No. 22, 29 (1960). ( 5 ) P. Yates and P. Eaton, ibid.. No. 11, 5 (1960).
and J. S. McNulty, J . A m Chem. Sos., 72, 1531 (1980); A . L. Wilds F. Weygand, G. Eberhardt, and C. H. Shunk. i b i d . , TO, 2427 (1948); 13. Linden, F. Schafer and I. Eigen, Angew. Chem., 66, 525 (1953). (7) T h e bicyclooctane furmalism (I-IV) is used only t o make clear t h e general character of t h e transformation which is taking place and should not be taken a s a description of t h e precise geometry of t h e species involved. Actually, as has been pointed out t o us by Professor
May 5, 1961
COhlMUNICATIONS TO THE
2199
EDITOR
Fig. 1.-?rotan n.nl.r. spectra a t 60 Mc. of rearranged dimers: chemical shifts are in cps. from tetramethylsilane ( S O cps.) as internal standard: (A) distXed dimer mixture after 9 hours a t 171' in a n open system: (B) crude product mixture after 7 hours a t 180" in a sealed tube. The inset enlargements of the aldehyde and vinyl regions mere taken a t a reduced sweep rate. One of the rinyl hydrogen resonances of methacrolein is buried under t h a t of the dimer. For comment on the unusual splitting of the dimer aldehyde proton resonance, see D. R. Davis, R. P. Lutz and J. D. Roberts, J . A m . Chew. Snc., 83, 246 (1961).
all process converts the aldehyde deuterium to a vinyl substituent and would occur with retelztiorh of optical configuration. Rearrangement by way of a diradical intermediate such as IV would be expected t o lead to more or less racemization. The experimental data for rearrangement of I a t 171' are suinniarized in t h e diagram (I -+ VI). Methacrolein, formed by dissociation of the dimer, distilled from the reaction mixture and was condensed a t -78". No other by-products were detected. T h e optical activity of the rearranged dimer, which represents 99.5y0 retention of configuration, was determined for a sample twice vacuum-distilled to reiiiove traces of methacrolein which remained persistently dissolved. The relative amounts of I and I1 were determined by electronic integration of the n.m.r. absorption signals of the aldehyde and vinyl hydrogens in the distilled dimer mixture (see Fig. 1-A). Another rearrangement mas carried out in a sealed n.m.r. tube; the experimental results (given in mole per cent.) are summarized in (I -+ [V -i- VI1). The relative amounts of rearranged and unrearranged dimers and of methacrolein formed were R. B. Wondmard, t h e substituents attached t o the two carbon-carbon partial d w b l e bonds of 111 probably would be very nearly planar. (8) Thwmometer bulb immersed in t h e boiling liquid; rearrangem e n t proceeded a t a n inconveniently slow rate a t 1130~. Hvdroquinone (0 8 % ) was used t o inhibit polymerization; nonetheless, 4.lc7, of polymeric residue was formed. I n a sealed tube in t h e absence of inhibitor, 30% of the starting material was converted t o polymer in 7 hours a t 180". (9) Root-mean square error.
I
'ihours a t 180"1n P
-71.44 =t0 0,j" (neat, I = 1 din.)
a%
i-
I
1
11
\'I 1I 2 h'
,
J
CPD -60.83 & 0.05nii (neat, 1 = 1 tlrn.)
determined directly in the reactioii mixture by n.m.r. (see Fig. 1-B). The amounts of VI1 and VIII, expected through random recombination of V and VI by a forward Diels-Alder reaction (with a corresponding loss of optical activity), were determined with a mass spectrometer. Assuming the reactions are first order, we calcumin.-' late approximately t h a t kr = 1.8 X for non-dissociative rearrangement and that k d (10) Hydroquinone (0.9%) added:
