. l i z d . Calcd. for CIJTlaClnSO C, 42.to; I f , .1.11. (1n.p. 123-124'), gave a dark red oil which could not be crysFound: C, 42.64; H, 4.iii. tallized. Hydrolysis of 0.90 g. (2.2 millimoles) of this crude nitration product with 50 ml. of acetone and 50 ml. of .4 Kutiz hydrolysis6 of this product yielded ii viscous oil 0.1 S aqueous sodium hyroxide5 gave 0.22 g., m . p . 121.5which could not be induced to crystallize. i > ~ - e r y t h o - I -5-Nitro-2( thieny1)-2- dichloroacetamido- 1,3- 1%'. The material was recrystallized successively from 10 ml. of ethylene dichloride, then 10 ml. of water yielding propanediol (VIIIb).-Xitration of the erythro-racemate of \.b (rn.p. 9;3.,S!f4.5") gave the crystalline &nitro compound 0.02 g., m.p. 131.5-132.5". A mixture of this product and (YIb), m.p. 123.3-126O. An analytical sample after rc- the threo-racemate of VIIIb prepared above (m.p. 131.5133") had a m.p. of 132-133'. crystallization from a 2 : 1 mixture of petroleum ether-ethyletle dichloride had a m.p. of 126.5-127.6". Acknowledgments.-We wish to express our ap.Irzul. Calcd. for C I ~ I I j t C l ~ S ~ O ,C, S : 3i.78; H, :3,41; preciation to Drs. Jonathan W. Williams and CI, 17.16. Founcl: C , .37.83; H, 3.60; C1, 1i.46. Li'alter -1.Gregory for many helpful suggestions 01ic gram (2.4 inillimolcsj of 1.1b (1n.p. 137.Z-12X.5c) during the course of this work. We are also inwas hydrolyzed using 50 ml. of acetone and 50 ml. of 0.1 S aqueous sodium hydroxidc.& The product amounted debted to Eugene B. Cooper, Jr., and Robert 0. to 0.51 g., m.p. 125-135'. This was recrystallized twice Bodycot for technical assistance. Microanalyses from 20 nil. of ethylene dichloride yielding 0.10 g., 111.p. were carried out by Dr. J. W. Stillman and his 167--168". It \vas finally recrystallized from 8 ml. of water group in the Chemical Department at the du Pont yielding 0.06 g. of long fine slightly yello\v needles, m.p. Experimental Station, and the ultraviolet absorp'4 mixture of this product and the crythroraceinate of \.IIIb prepared above (n1.p. l W . 3 - - 1 7 ~ ~J , , j otion spectra were determined by l l r . Charles €3. had a 111.1).of 16Y.5-170.5°. LIatthews also in the Chemical Department. DL-tizueo- 1-(5-Nitro- 2- thieny1)- 2- dichloroacetamido- 1,3propanediol (VIIIb\I.---Sitratiori of the tlzrco-raccniate of \'b \ \ r r . ~ ~ s c ; ~ oI s) E. L A ~ - A I W )O.
1 COS~IRIL3L..TIONFROM TIIE 1)EPARI.MES'I.
O F CIIENISTRY, r N I V E R S I T P Olr SOU'THERS CALIFORSIA]
On the Structure of Dehydroacetic Acid €31.
JEROME
-1.BERSON
RECEIVED ~ I A 7, Y 1952
A consideration of the acidic dissociation constants and ultraviolet spectra of dehydroacetic acid, triacetic lactone and certain of their dcrivatives provides confirmatory cvidencc for the presently accepted formulations of these substances.
The structure of dehydroacetic acid has been UII- acetate, formulated as VIa, which on careful questioned since the demonstration1 that the ob- saponification yields the unstable triacetic acid served isomerization of the substance by hot S& (VIb). This acid readily suffers decarboxylation to sulfuric acid to 2,Ci-dimethylpyrone-.l-carboxylic dcetyla~etone.~Tridcetic lactone is regenerated acid-3 (I) was most easily rationalizable in terms when ethyl triacetate is warmed with iOY0 sulfuric of the Feist structure? (TI). The possibility that acid.6 These observdtions leave only two possible itructures for triacetic acid, VIb and VIIb. 0 0 0
CHBCOCHLCOCH~COOK \'Id, R = C-11 YIb, R = I €
/'\
I
CH 4
I
I
I1 \'
0
I
.\
CI1,CO
__
J)
C00Ii C H COCH-COC€I~
~
CH3COCl~2"'0'"U I11
0 '
\ IId, K =
'.COCH.
IY the structurc (111) proposed by Collie' or some other alternative (e.g., 11') might also be convertible to I under the drastic conditions of the isomerization, as well as the observation ( o d e znfru) of certain apparent anomalies in the ultraviolet absorption behavior of the dehydroacetate ion have led us to re-examine the structural question. With 90yo sulfuric acid, dehydroacetic acid is converted to triacetic lactone, formulated as V.4 Alcoholysis of triacetic lactone leads to ethyl tri11, C
r
Ra,cariler
,rid
R
4dnrii
lrrr>
11424)
2758
69
fii
I\'l
?AI
17'1 l \ ' J l
/o
'
(I-COCIL \ 111
p'-L,ictone structures for triacetic ldctone cdn be rejected in view of the relatively high stability of the lactone to acids and bases.' There remain, thereforc, only two possible structures for tridcetic lactone derivable from VIa or VIIa by the loss of alcohol, namely, V and VIII. Triacetic lactone readily undergoes 0-alkylation,' is a inonobasic acid, gives a blood-red color with ferric chloride and is therefore to be considered Ill1
I 8'10 I 1 I Y L i > l l l < / tho,! \ > ( I I \ (C,lIil , / I , / 59 l,l'l
I r r ~ \ a i 46
CLH,
YlIb, R = H
__
R I \Titter a n d F H St0t7 T Afof Chrm 176 485 11948) I' R. J o h n \ c ~ n II h f U a r i i k i , i r d \ \1 \ f c L I \ d n i 1 HI, J u r Y 62 ' I O i 1'110 31, I (/titi \ 89 I I Y 8 I I ~ I I I I ~ I ) i I l r t i 7 6 I I l h I'III,
Oct. 20, 1952
O N THE STRUCTURE O F DEHYDROACETIC ACID
highly enolic. Of the three possible enols of VI11 (VIIIa, VIIIb, VIIIc), both VIIIa and VIIIb contain the enolized B-diketone svstem characteristic of the acidic dimer of methyl'ketenes (IX). The
ofl'-JCOCK3 VIIIa
VIIIc
51'73
L
3.5
-
3.0
-
2.5
-
VIIIb
IX
ultraviolet absorption spectrum of triacetic lactone in neutral or acid solution (Fig. 1, Xmax 283 mp) indicates a considerably higher degree of conjugation in the absorbing species than is normally encountered in simple enolized P-diketone~.~This result is consistent with the formulation of triacetic lactone as an enol of VI11 only if all three carbonyl groups are chromophorically active and thus part of the conjugated system. This argument logically implies that VIIIa and VIIIc will form the same anion and consequently that any of the enols (VIIIa, VIIIb or VIIIc) will be acids of a t least the same order of strength as IX.l0 Potentiometric titration of triacetic lactone shows that is a weaker acid (pK 5.0) than IX (pK 2.8) by a factor of about 160. In terms of the cyclobutenolone structures (VIIIa-c) for triacetic lactone, there seems to be only one reasonable explanation for this marked discrepancy in acidity, namely, that there is very effective intramolecular hydrogen-bonding between the enolic hydrogen and one of the carbonyl groups. This type of effect is exemplified by the sharp diminution in acidity of acetylacetone (pK 8.938) as compared to dimethyldihydroresorcinol (pK 3.253) . I 1 The existence of strong hydrogen-bonding in the present cases (VIIIa-c) seems unlikely because of the unusual steric requirements resulting from partial inclusion of the enolized P-diketone system in the four-membered ring. I n addition, one would expect a marked change in the ultraviolet spectrum of the enol ether of triacetic lactone as compared to that of the free enol.12 The absence of any such marked effect in the spectrum of (8) R. B. Woodward a n d G. Small, THISJ O U R N A L , 72, 1297 (1950). (9) Compare dimethyldihydroresorcinol (Amx 255 mp) [E.R. Blout, V. A. Eager a n d D. C. Silverman, ibid., 68,566 (1946)l. (10) Woodward a n d Small (ref. 8) have suggested a n interesting explanation for t h e inordinate enhance m e n t of acidity resulting from t h e incorporation of t h e P-diketone system in t h e four-membered ring of IX. (11) G. Schwarzenhach and K. Lutz, Helu. C k i m . Acla, 23, 1147 (1940). (12) Derivatives of enolized P-diketones a n d o-hydroxyacetophenones in which intramolecular hydrogen-bonding is prevented ( e , # , . t h e enol ethers or acetates) show a pronounced shift of t h e principal absorption maximum toward t h e violet. T h u s , (a) acetylacetone absorbs a t 270 m p while t h e acetate absorbs a t 230 m p [R.S. Rasmussen, D. D. Tunnicliff a n d R. R . B r a t t a i n , THISJ O U R N A L , 71, 1068 (1949)I ; (b) dibenzoylmethane absorbs a t 345 mp a n d cis-P-methoxystyryl phenyl ketone a t 290 m p [B. Eistert, F. Weygand and E. Csendes, Chem. B e y . , 84, 745 (1951)l; (c) R . A. Morton, A. Hassan a n d T. C. Calloway. J . C h r m . .Tor.. 883 (1934),report ,A, 345 mp for dihenzoylmethane and Anlax 300 m p for 8-ethoxystyryl phenyl ketone; (d) 2hydroxy-4-methoxyacetophenone absorbs a t 274 mp and 2,4-dimethirxyacetophcrione at %ti7 inp [ U . J. C r a m and .'1 W. Cranz, THIS . r U l J K N A l . , 72, i!,j ( l ' , I > o ) 1.
ui M
2.0
-
I
360
320 A
Fig. 1.-A,
280
240
(w).
triacetic lactone methyl ether; B, desoxodehydroacetic acid: C, triacetic lactone.
triacetic lactone methyl ether (Fig. 1, Xmax 2SO mp) we interpret as strong evidence against the hydrogen-bonding hypothesis. Having thus eliminated all reasonable interpretations of the properties of this substance in terms of enols derived from VIII, we conclude that triacetic lactone must be an enol of V. The direct acetylation of triacetic lactone to dehydroacetic acid13 implies a close structural relationship between these substances. The relationship is clarified by a consideration of the properties of the compound CgHloOs, m.p. 188" (hereinafter designated as desoxodehydroacetic acid) formed on catalytic reduction of dehydroacetic acid14 and assigned structure X. Although triacetic lactone potassium salt could not be alkylated to desoxodehydroacetic acid, the 0-ethyl ether being obtained instead,14 the acidity of desoxodehydroacetic acid (pK 5.3) and, more particularly, the ultraviolet spectrum (Fig. 1) leave little doubt that the substance is indeed a C-ethyl derivative of V. Of the three possible structures (X, X I and XII), only X and XI are consistent with the observed14hydrolysis to 2,4-heptanedione. Since the reduction of dehydroacetic acid involves the change C-COCH3 +- C-CH.LCH~, there remain only two possible structures for dehydroacetic acid, namely, I1 and 111. With this delimitation of the structural alterna(13) W.Dieckmann a n d F. Breest, Bev., 37,3387 11904). (14) R. Malachowski a n d T. WdnCZUrd, Bull. in1criL. acud. p d o i l i l i s r , lSSSA, 547 IC. A . , 28, 4421 (1934)l.
JEKOAIU 0
0
I t is now pertinent to examine the consistency of the Feist structure (11) with certain further properties of dehydroacetic acid. The striking similarity of the ultraviolet absorption spectra of dehydroacetic acid and its ethyl ether (Fig. 2) provides cornpelliug evidence that dehydroacetic acid exists as a virtually completely enolized species in solution. The pronounced bathochroinic shifts of the maxima in the spectra of dehydroacetic acid (Ama, 310 mp) and the dehydroacetate ion (Ama, 294 nip) relative to the maxima for triacetic lactone (A, 283 mp) and the 278 mp) are consistent triacetic lactonate ion (A,, with the chroriiophorically active situation of the tivcs, the available chemical evidencc stroiigly in- acetyl group in each of the enol variants of the dicates the correctness of the Feist structure (11). Feist structure (IIa, IIb, IIc: R = H). I n addition to the isomerization of dehydroacetic ()I< 0 x i t i by sulfuric acid (aide supra), the demonstration that the dehydration product of dehydroacetic acid phenylhydrazone has the structure XI11 is, barring an intervening rearrangement, clearly consistent only with I I. 0
I
I
I
I
,
0K
I
I t is noteworthy that maximal absorption for the dehyclroacetate and triacetic lactonate ions occurs a t sizorter wave lengths than for the corresponding un-ionized acids. This is in contrast to the observedg bathochromic effect of ionization on simple 8-dicarbonyl systems. On the basis of the absorption data (Fig. 2), it seems certain that the conversion of dehydroacetic acid to its ethyl ether occurs with minimum structural change, ;.e., with simple replacement of hydrogen by ethyl. *llthough the ethyl ether IIa (I< = C2H5)might conceivably be expected to show acidic properties since the methyl group is conjugated with both the lactone and side-chain carbonyl groups, both IIb and IIc (K = C2Hs) require the absence of acidic character in dehydroacetic acid ethyl ether. The observations that the ether tloes not consume alkali upon titration, gives a negative ferric chloride reaction and exhibits virtually unchanged spectral characteristics in alkaline solution are thus consistent with at least two of the expressions derived from the Feist structure. l7 (16) T h e circumstance t h a t notie of the enol variants of t h e Cullic structure (111) accommodates t h e combined requirements (i) Lhat tlic acetyl group of dehydroacetic acid he chromophorically active and iii) that dehydroacetic acid ethyl ether he non-acidic, provides further direct evidence of the inadequacy of 111. (17) T h e ftirmiilation of t h e ethyl ether as a C-ethyl derivative of I1 ( X l V l also accounts for t h e lark of acidic character. I t is, however, incoriristcni with t h e faciic hydroly3is t o the parent enol.” F u r t h e r , XI\’ fails t u account for t h e ultraviolet absorption behavior since i t contains only t h e simple @-acylory-n,0-unsaturated ketone chromophore (compare acetylacetone enol acetale, hmal 230 mplza).
0
Fig. 2.
~
dehydroacetic acid ethyl ether; €3, dehydro-
acetic acid.
- _...___ i1.5) li. Beuary, L k . ,43, 1070 ~ l ! ) l O ~ .
SELF-CONDENSATION OF PHENACYL BROMIDE
Oct. 20, 1952
5175
droxide (one or more equivalents of alkali), the maximum occurred a t 278 mp, log e 3.84. Values for neutral equivalent (129, calcd. 126) and pK (5.00)were obtained by potentiometric titration of the lactone in water solution with standard sodium hydroxide. Triacetic lactone methyl ether,'* white needles from ether, ,,A 280 mp, m.p. 78-80' (reported" m.p. 81'), showed Experimental18 log e 3.76. Dehydroacetic acid ethyl ether,Z2 was obtained as fine, Dehydroacetic acidls (11) was obtained as white needles white needles from ether, m.p. 90-91.2" (reportedz' m.?. from ethanol, m.p. 108-109' (reportedI9 m.p. 108"). The 93-94'). The compound failed to develop color with ferric ultraviolet spectrum showed maxima a t 225 mp, log e 3.99 chloride a t room temperature. On boiling the solution a and a t 310 mp, log e 4.05.z0 In the presence of one or more short time, the color deepened from the pale yellow of the equivalents of sodium hydroxide, the spectrum showed Laxferric chloride solution t o a n orange-brown which was visu294 mp, log e 3.91. ally indistinguishable from the color generated by dehydroSince the acid is sparingly soluble in water, the potentio- acetic acid in ferric chloride. metric titration (Beckman P H meter) was performed by Potentiometric titration of the ether with standard sodissolving 0.0358 g. of acid in 15.32 cc. of 0.0225 N sodium dium hydroxide failed to reveal acidic character. The addihydroxide and back titrating with 0.0395 N hydrochloric tion of 0.3 equivalent of alkali caused the p H to rise t o 11. acid. From the titration curve were obtained a neutral After 0.5 equivalent of alkali had been added, the solution equivalent 170 (calcd. 168) and a pK 5.30. T h e excellent was a t p H 11.2. We do not attach any quantitative signifiagreement with the value pK 5.28 foundz1by the conducto- cance t o this observation since the p H was beyond the remetric technique is presumably fortuitous, since our figure liable range of the cohventional Beckman glass electrode is not corrected for salt effect. instrument. Triacetic Lactone (V).-As obtained by the procedure of The ultraviolet spectrum showed AXrnx 22Gmp, log 3.90 and Collie,4 the lactone, glistening buff needles from water, had 312 mplog e 3.89. The position and intensity of the maxima m.p. 186-186.5' (reported4 m.p. 188-189'). The spectrum were unchanged by the addition of 1.25 equivalents of alkali. in ethanol alone or in the presence of a twenty-molar excess Desoxodehydroacetic Acid.-This substance was obtained of sulfuric acid showed AXmax 283 mp, log e 3.78 and 345 mp, by catalytic reduction of dehydroacetic acid as rugged, log E 2.45. (Witter and Stotz6 report Amax 282 mp, log e chalk-white prisms, m.p. 188-188.1" (reported14 m.p. 3.89 for an aqueous solution.) In ethanolic sodium hy185'). Potentiometric titration was carried out as for dehydroacetic acid. The curve was not quite steep enough a t (18) All melting points are corrected. The ultraviolet spectra were the equivalence point for accurate determination of a neudetermined in 95% ethanol with the Beckman spectrophotometer, tral equivalent. The pK calculated from the curve was model DU. 237 mp, 5.30. The ultraviolet spectrum showed , ,XA (19) F . Amdt, Org. Syntheses, 20, 26 (1940). log e 2.95 and 290 mp, log E 3.92. (20) The spectrum in 2,2,4-trimethylpentane solution above 250 mp
Although presently available information does not allow a definite choice among the three enols, we consider that the expression of the properties of dehydroacetic acid as an enol of the Feist structure (11)is confirmed.
has been reported to show Amax at 311 mr, log c 4.3 [M. Calvin, T. T. Magel and C. D. Hurd, THISJOURNAL, 63, 2174 (1941)l. (21) W. Ostwald, Z . ghysik. Chem., 8, 401 (1889).
[CONTRIBUTION FROM THE
(22) J. N. Collie and H. R. Le Sueur. J . Chem. S O L . 66, , 254 (1894).
Los ANGELES7. CALIFORNIA
DEPARTMENT OF CHEMISTRY,
UNIVERSITY O F SOUTHERN CALIFORNIA]
The Self-Condensation of Phenacyl Bromide. The Structure of the Halogen Diphenacyls BY JEROME A. BERSON RECEIVED MAY13, 1952 T h e 3,4-epoxytetrahydrofuranstructures proposed by Widman for thc a- arid 13-bromodiphenacyls are shown to be incorrect. These substances are formulated as epimeric a,P-epoxyketones resulting from normal Darzens-type self-condensation of phenacyl bromide.
The sodium ethoxide catalyzed self-condensation of phenacyl bromide, which occurs according to the stoichiometry 2C6H5C0CH2Br + C16H1302Br HBr, leads to two isomeric products: a-bromodiphenacyl (m.p. 129") and 0-bromodiphenacyl ( m p . 159').ls2 The investigation of the structure of these product^^-^ culminated in the proposal by Widrnan7 that they were stereoisomers of the structure I.
+
A
(1) (2) (3) (4) (5) (6) (7)
V. Fritz, Ber., 28, 3028 (1895). C. Paal and C. Demeler, ibid., 29, 2092 (1896). C. Paal and H. Stern, i b i d . , 84, 1611 (1901). C. Paal and H. Schulze, ibid., 86, 2405, 2415. 2416 (1903). W. L. Evans, A m . Chcm. J . , 36, 115 (1906). 0. Widman, Bcr., 44, 3261 (1909). 0. Widman, Ann., 400, 86 (1913).
The formation of I admittedly7 requires a singular condensation mechanism. I n addition, the reported chemistry of the bromodiphenacyls conflicts seriously in several respects with that predicted by I. These substances are obtained in strongly basic solution, survive prolonged boiling in neutral alcoh01,~and react sluggishly with water, potassium iodide, potassium cyanide, silver acetate4 and alcoholic silver nitrate. These properties are in direct contrast to the well-known8-11behavior of a-haloethers, which are inordinately responsive to reagents of this type. Further, no mechanism is readily apparent for the facile ba~e-catalyzed~ conversion of a-bromodiphenacyl to the p-isomer. Although the possibility of the presence of a ( 8 ) A. Geuther and H . Laatsch, ibid., 218, 36 (1883). (9) F . M. Litterscheid, ibid., 380, 118, 123 (1904). (10) H. R. Henze and J. T. Murchison, THISJOURNAL, 63, 4077 (193 1). (11) J. B. Conant, W. R. Kirner and R. E. Hussey, ibid., 47, 488 ( 1925).
