4510
COMMUNICATIONS TO THE EDITOR
8.8 (6 H ) . These n.m.r. data rigorously exclude I11 from further consideration as a structure for the major isolable product. Reduction of the product C 2 0 H ~ o N zwith 0 lithium aluminium hydride4gave a colorless oil, CzoHz2Nz,which was purified by V.P.C. (,lnul. Found: C , 82.10; H 7.71) and had no carbonyl band in the infrared. Thus cycloadducts such as IV and V are also eliminated as structures for the product. Cd% I
N-CCBH~
0
V
IV
VI
While the 3-pyrazolidone VI is the structure most consistent with the spectral5and chemical evidence t h a t can be formally derived from I1 and phenyl isocyanate by a cycloaddition, it may not be assigned to the product a t present. The reaction of the cyclopentene-phenyl azide 2-phenyl-2,3,4-triazabicyclo j3.3.0 Ioct-3-ene adduct, (VII), with phenyl isothiocyanate has also been represented as a "1,3-dipolar cycloaddition."3 CsHs
C6H5
CeHs
I
I
VI1
I
I
VI11
CEHS IX
When it became clear t h a t the reaction of the norbornene- phenyl azide adduct with phenyl isocyanate does not follow an analogous pathway, the thermal decomposition of VI1 in the presence of phenyl isothiocyanate was reinvestigated. The major product from the reaction is neither the thiourea IX nor the 3-thiopyrazolidone X, b u t rather the thioanilide X I . 6
a!,! XI'
CsH5 I
SOYES CHEMICAL LABORATORY J O H N E. BALDWIN GARYV. KAISER^^ UNIVERSITY OF ILLINOIS URBANA,ILLINOIS JUDITH A . ROWERSBERGER'~ RECEIVED AUGUST24, 1964
S X
1390, and 1250 em.-' strong;? : :X 386 ( E 24,ooO) and 314 mp ( e 9000). The n.m.r. spectrum of a solution of the product in deuteriochloroform had absorption a t T -3.0 (1 H ) , 2.0-3.4 (11 H ) , a multiplet centered a t 7.25 (4 H), and a multiplet centered a t 8.13 (2 H ) . The infrared, ultraviolet, and n.m.r. spectral data are all incompatible with structure IX.' The spectral data may, however, be rationally correlated with an alternate structural proposal, the thioanilide XI. In particular, the 3390 e m - ' band in the infrared and the T -3.0 proton in the n.m.r. spectra may be assigned to the WH proton of XI, hydrogen bonded to sulfur.8 Conclusive evidence in favor of structure XI was obtained by hydrolysis of the product C18HIRN2S with 707, sulfuric acid to aniline and P-ketothioanilide X I I . The hydrolysis product proved identical with an independently synthesized* authentic sample of XI1 as judged by melting point, mixture melting point, and infrared and n.m.r. spectral criteria. This reaction of triazoline VI1 may involve l-anilinocyclopentene as an i r ~ t e r m e d i a t e or , ~ ,some ~ morecomplex mechanism may obtain. It is clear t h a t , a t least in our hands, the reaction is no cycloaddition, and t h a t seemingly modest variations in structure, as between the cyclopentene- and norbornene- phenyl azide adducts and between phenyl isothiocyanate and phenyl isocyanate, ,can divert one high-yield process to alternative modes of behavior. T h a t the reaction of I with phenyl isocyanate involves the dipole I1 as an intermediate appears extremely doubtful. Preliminary work indicates that the rate of nitrogen evolution from I depends on the phenyl isocyanate concentration. Whether or not our products VI, m.p. 162', and XI, m.p. 128.5-130.5', and those isolated by earlier workers2s3 and assigned structures 111 and IX are identical remains to be seen.Io Kinetic studies of these and similar reactions are in progress. The present results strongly suggest that "1,3dipolar cycloadditions" may be less mechanistically homogeneous than has been previously supposed. (7) For N',N'-diphenylthiourea, 272 mp (log c 4.32); bl. G . Rttlinger a n d J. E. Hodgkins, J . A n . Chem. S O L .7, 7 , 1831 (195.5). (8) W. Reid and W . Kiippeler, A n n . , 673, 132 (1964). (9) G. A. Berchtold, J . Org. C h e m . , 26, 3034 (1861), S . Hunig, K . H u b n e r , and E. Benzing, Chem. B e y . , 96, 926 ( 1 9 6 2 ) , R. Fusco, G. Bianchetti, and S. Rossi, Gazz. chim. i t d , 91, 825 (1961). (10) In ref. 2 a n d 3 , credit for these cycloaddition products is given t o unpublished research of R. Huisgen and R . Grashey, Miinich, 1959-1960, unf o r t u n a t e l y , no melting points or spectral d a t a for t h e adducts "111" and "IX" are given. (11) National Science Foundation Undergraduate Research Participant.
CeH5
I
Ir01. S(i
XI
HNCeH5 XI1
The triazoline VI1 with an excess of phenyl isothiocyanate in chlorobenzene a t 110' gave the product ClsH18X2S [.-lnul. Found: C, 73.36; H , 6.12; N, 9.39; mol. wt., 296 (osmometric in benzene)] as bright yellow needles, m . p . 128.5-130.5' (78y0 yield, from cyclohexane) ; v:ty3 3390, 1612, 1595, 1580, 1900,
The Influence of Unsaturation and of Fluorine Substitution on Ketone-Alcohol Equilibrium Constants. A Measure of a$-Unsaturated Ketone Resonance Energy and of Halogen Destabilization'
Sir : T h e destabilization of a carbonyl group by a halogen atom located a to a saturated ketone or a , 6, or y to
(4) Compare R. L. Hinmaii. J . A m , Chem. S O L .7, 8 , 1645 (1956).
( 5 ) B . J. R . Nicolaiis. e / a l . . H i i v . Cliim. A c t a , 44, 2055 (1961). (6) Possible tautomerization hetween XI a n d X I ' is here neglected
(1) Supported in p a r t by G r a n t F-185, American Cancer Society, and A4044, National Institiites of Health.
COMMCNICATIONS TO THE EDITOR
Oct. 20, 1964 TABLE I Alcohol
cm,
A F s O . kcal
(eq. 1)','
(es. 3)
0.01
+8 5
1734
+5.8
1704
+5.5
1688; 169i5
Ketone
F ' J y y 0
CHCli
K,, X 10-8
Ymax
ketone
,
-05
B. H OH
''
OH
*'
1.8
F-&
&H F
d5 1
H
no
0
N4
+5.0
1685
&
f 4 0
1680
+3 2
1675
+1.7
1660
0
F
0
OH H
d5
G' H
22
H
OH a Equilibrium constants were reproducible t o about 1 2 0 y 0 on repeated runs. * Doublet. Infrared spectra were determined by M r . George Scrimshaw on a Beckman IR-7 with Bausch and Lomb replica grating. e Equilibrium constants were determined spectrophotometrically in 0.03 M phosphate buffer by following the appearance or disappearance of D P K H a t 340 mp essentially according t o the procedure of Talalay and Marcus (ref. 4). In the case of entry G equilibrium was approached from both sides while the others were determined by reduction of the ketone.
an a,PTunsaturated ketone is well-recognized, although the magnitude of this effect has not been directly open to experimental determination. Recently we have found t h a t the 3a-hydroxysteroid dehydrogenase of Pseudomonas testosteroni4 catalyzes the reversible, D P N H dependent conversion of A4-3-ketosteroids to the corresponding A4-3a-hydroxy compound (eq. 1) and t h a t a halogen substituent on or adjacent to the unsaturated ketone system facilitates the r e d ~ c t i o n . ~ We have now measured the equilibrium constants6 for the conversion of testosterone and a number of its fluorinated derivatives to the A4-3a-alcohols. This has enabled determination of the magnitude of the halogen effect and, further, comparison of the respective equilibrium constants for the saturated and unsaturated ketone-alcohol systems provides an estimate of the extra resonance energy of an a,@-unsaturated 3-ketosteroid.
([alcohol][ D P N f ] ) ] and the standard free energy change, AF3', corresponding t o eq. 3 (alcohol 2H). T h e latter value is readily obtained4 ketone from AF,' by subtracting AF?' ( + 5 . 2 kcal.), the standard free energy change for the conversion of D P N + to D P N H (eq. 2 ) which has been determined by Burton and Wilson.7 Testosterone (Table I , G) exhibited K,, = 900 X 10F. Introduction of a fluorine atom a t the Ba- (equatorial), 6@- (axial), 4-, and 2-positions, respectively, led to progressive shifts of the equilibrium constants toward the alcohol. In the case of 2a-fluorotestosterone (C) there was a 500fold shift relative to testosterone, K = 1.8 X 1W8, which was of the order of magnitude of the saturated 5a-3-ketone B, K = 1 x 1 F 8(900-fold shift). Introduction of a 2a-fluor0 substituent into the saturated ketone A resulted in a further 100-fold displacement of the equilibrium constant toward the alcohol. One may derive certain interesting data by making the reasonable assumption that the effects noted are due primarily to electronic stabilization or destabilization of the ketone rather than of the alcohol. Table I1 lists the effects (AAF3') of the various groups on the equilibrium constants. Subtraction of the free-energy change for the conversion of the 5a-3a-01 to the saturated 5a-3-one from t h a t of the conversion of the A4-3a-ol to the A4-3-one gives a AAF3' of -4.1 kcal., which may be taken as a first approximation of the additional resonance energy of the a,@-unsaturated ketone. An additional 0.3 kcal. should be added to this figure for a net a$-unsaturated ketone resonance energy of about 4.4 kcal.8 since the saturated 3a-01 has one additional 1,3-diaxial interaction ( 3 a - O H , 5 a - H ) , which is not present in the A4-3a-~l.
+
TABLE I1 VARIOUS GROUPSO S EQUILIBRIUM: ALCOHOLs KETONE 2H
*t3
4- D P S + *ketone + D P S H + H + + 2H S D P N H + H * alcohol * ketone + 2H AFao = A K " alcohol
AK"
(1)
DPN+
AFz0
(2)
AF," ( 3 )
Table I lists the equilibrium constants at 27' corresponding to eq. 1 [ K = ( [ k e t o n e ] [ D P N H ] [ H + ] ) / Cf, E. J . Corey, J. A m . Chem. Soc., 76, 2301 (1953); J . Allinger and Allinger. Tetrahedron, 2, 64 (1958); N. L . Allinger, M A. DaRooge, Miller, and B. Waegell, J . Org C h e m , 28, 780 (1963). H. J . Ringold a n d A. Bowers, E x p e r i r n l i n , 17, 6.5 (1961). (1) P. Talalay and P. I . Marcus, J . Bioi. C h r m . , 218, 675 (1956) ( 5 ) H . J . Ringold, J. Graves, hl. Hayano, and If. Lawrence, J r . , Biochem. B i o p h y s . Res. Commun., 13, 162 (1963). (6) See ref. 4 for the enzymatic determination of the equilibrium constants of saturated alcohol-ketone systems. (2) N. L . M. A. (3)
451 1
INFLUENCE OF
Group
+
E n t r y numbers from Table I
4(5)-Double bond G-B 2a-Fluoro (saturated ketone) A-B 2a-Fluoro ( A4-3-ketone) C-G 4-Fluor0 ( A4-3-ketone) D-G 68-Fluoro ( A4-3-ketone) E-G 6a-Fluoro ( A4-3-ketone) F-G 5 Estimated accuracy 1 0 . 1 5 kcal
AAFa', kcal."
-4.1 +2.7 +3.8 +3 3 +2 3 +1 5
T h e equatorial a-halo ketone effect, commonly attributed to halogen-oxygen dipole-dipole repulsion, is readily recognized by the bathochromic shift of the carbonyl stretching frequency in the infrared.g This effect is seen in the 2a-fluoro-saturated ketone which is destabilized by 2.7 kcal. relative to the unsubstituted saturated ketone, a figure identical with the electrostatic interaction calculated by Corey? for an equatorial a-bromocyclohexanone. The 3.8-kcal. destabilization effect of the 2a-fluor0 group and the 3.3-kcal. effect of the 4-fluoro atom in the a,@-unsaturated ketones would appear to be due to a combination of the aforementioned electrostatic interaction and an inhibition of double-bond polariza(7) K . Burton and T H . Wilson, Biochem. J , 6 4 , 86 il95:3). (8) hf. M . Kreevoy and I
