1648 J.Org. Chem., Vol. 41, No.

9,1976 Notes

Stability of 2,4,6-Tri-tert-butylphenol in Acetic Acid. Sodi- dium bicarbonate (to neutralize the acetic acid formed
um bismuthate (33.0 g, 0.118 mol) was added to a solution of I11 which otherwise gave side reactions with 2): we could,
(8.73g, 0.033 mol) in 50 ml of benzene. The mixture was stirred for however, increase the yield of 2 to 80% (VPC analysis of the
6 h at room temperature under nitrogen. The mixture was filtered
rapidly. The dark blue filtrate was mixed with 100 ml of glacial reaction mixture). Furthermore, the resulting reaction mix-
acetic acid and kept under nitrogen for 2 weeks. The color persist- ture, containing also some residual 1 (87% conversion7),
ed until the nitrogen supply was exhausted. could be used, after filtration from the salts, directly for
Stability of Hydroxycyclohexadienone IX in Acetic Acid. the reaction with cupric chloride.
Compound IX was prepared from 2,4,6-tri-tert-butyl-4-nitrocyclo- In preliminary attempts to convert 2 to useful precursors
hexadien-2,5-onegby the method of Muller and Ley.lo In 5 ml of of 4 we speculated that 2, by free-radical reaction with tert-
glacial acetic acid was dissolved 0.330 g of IX. The solution was
kept at room temperature for 3 days. The solution was poured into butyl hypochlorite, could give 4-chloro-3,4-epoxy-3-
75 ml of water. The precipitate was filtered, dried, and weighed at methyl-1-butene ( 5 ) 8 and that the latter, by rearrangement
0.294 g (89%)and found to be identical with the starting material (possibly in situ),g could afford either 3 or its 2 stereoiso-
by an infrared spectrum and mixture melting point determination. mer (6) or the constitutional isomer 2-chloro-2-methyl-3-
Registry No.-I, 7218-21-5; 11, 4906-22-3; 111, 732-26-3; IV, butenal(7).
2525-39-5; V, 20778-61-4; VI, 20778-58-9; VII, 1975-14-0; VIII,
719-22-2; IX, 4971-61-3; sodium bismuthate, 12125-43-8;2,6-xyle-
nol, 576-26-1; 2,4,6-tri-tert-butyl-4-nitrocyclohexadien-2,5-one,
1665-87-8.
References and Notes 5
(1) E. Kon and E. McNelis, J. Ora. Chem., 40, 1515 (1975).
E. Adler, K. Holmberg, andI.-O. Ryrfors, Acta'Chem. Scand., Ser. B.
28, 888 (1974).
C. J. R. Adderley and F. R. Hewgill, J. Chem. SOC.C,2710 (1968).
W. A. Waters, J. Chem. SOC.B, 2026 (1971).
J. Derkosch and W. Kaltenegger, Monatshefte, 90, 874 (1959).
E. McNelis, J. Org. Chsm., 31, 1255 (1966). 6 7
T. Matsuro and K. Ogura, Tetrahedron, 24, 6157 (1968).
E. Muller and K. Lev. Chem. Ber.. 87. 922 119541. In fact, after gas-phase reactionlo over alumina of an
C. D. Cook and R. C.Woodworth: J. Am. Chem.'Soc., 75, 6242 (1953). equimolar mixture of 211 and tert -butyl hypochlorite
E. Muller and K. Ley, Chem. Ber., 88, 602 (1955).
(added dropwise, from the top, to a heated (125 "C)column
charged with alumina pellets, using nitrogen as carrier gas
and collecting the product a t the bottom of the column in a
liquid air cooled trap), VPC analysis of the reaction mix-
ture revealed formation of a product (37% yield) with the
A Two-step Synthesis of same retention time as 3. The same analysis revealed also
( E )-4-Chloro-2-methylcrotonaldehyde formation of a product (37% yield) with retention time as
from Isoprene. An Unprecedented Oxidative for (E)-2-methylcrotonaldehyde (8). Structures 3 and 8 for
Chlorination of a 1,3-Diene Monoepoxide
by Cupric Chloride
Giancarlo Eletti-Bianchi,Felice Centini, and Lucian0 Re*
OHC- G
Snamprogetti S.p.A., L.P.M.,00015 Monterotondo (Rome), Italy 8
Received J u n e 30,1975 these two products were confirmed by the NMR and ir
spectra and the boiling points of the compounds isolated,
A key intermediate in the Pommer industrial synthesis in low yields, by fractionation. Derivative 8, the rearrange-
of vitamin A acetate from p-ionone is (E)-4-acetoxy-2- ment product of 2, was obtained in quantitative yields
methylcrotonaldehyde (4).l However, the known syntheses when the gas-phase reaction of 2 over alumina was carried
of the latter derivative are multistep and/or low-yield pro- out in the absence of the hypochlorite. On the other hand,
cedure~.~-~ when 8 was treated with the hypochlorite under the condi-
The scope of the present paper is to report a more effi- tions used for 2, no 3 was produced, supporting the inter-
cient route to 4 involving epoxidation of isoprene (1) with mediate formation of 5, rather than 8, in the reaction of 2
peracetic acid t o 3,4-epoxy-3-methyl-l-butene (2) followed with the hypochlorite to give 3.
by oxidative chlorination of the latter with cupric chloride Although it is known that by oxidative chlorination with
to afford (E)-4-chloro-2-methylcrotonaldehyde(3), a cupric chloride 8 can be converted to 3 in moderate y j e l d ~ , ~

+=-
known direct precursor of 4.2 we thought that a more direct and possibly more efficient
synthesis of 3 (or 6) from 2 might be achieved by reaction
=I-= CH COOOH
L* cuc!, of 2 itself with cupric chloride. To our knowledge, oxidative
halogenation of an epoxide with a cupric halide has no
1 precedent in the literature.12 We speculated, however, that
2 reaction of 2 with cupric chloride could afford 3 (or 6), ei-
ther via rearrangement in situ (cupric chloride acting as a

OHC OHC- u""" Lewis acid catalyst) of 2 to 8 (or to its 2 stereoisomer) or

via cupric alcoholate 9.
In fact, when the chloroform solution of 2 obtained from
the peracetic acid epoxidation of 1 was treated, after addi-
3 4
tion of an equal volume of ethyl acetate,l3 with cupric chlo-
Although the peracetic acid epoxidation of 1 to 2 has ride in the presence of lithium chloride,14 VPC analysis of
been already described, the yield claimed is only 42%.5 By the organic extracts revealed formation of a major product
using a modified procedure, i.e., carrying out the reaction (80% yield) with the same retention time as 3. The reten-
a t about 5 OC in chloroform solution in the presence of so- tion time of the only by-product (20% yield) was as for 8.
Notes J.Org. Chern., Vol. 41, No. 9,1976 1649

'?'L-
,.c1 3,4-Epoxy-3-methyl-l-butene (2). To a solution of 20.4 g (0.30

-
Cl.,

+,
mol) of 1 in 240 ml of chloroform was added 30.2 g (0.36 mol) of
2 + CUClZ
NaHC03 and a trace of radical inhibitor 2,6-di-tert -butyl-4-meth-
ylphenol. To the stirred suspension was added dropwise over a 6-h
period under ice-bath cooling and Nz atmosphere 27.8 ml of 82%
peracetic acid (corresponding to 0.30 mol of peracid). After an ad-
ditional 24 h of stirring under the same conditions, the undissolved
salts were removed. by filtration. Gas chromatography-mass spec-
trometry analysis of the filtrate confirmed structure 2 for the reac-
tion products on comparing its mass spectrum, m/e 84 (9), 83 (14),
69 (16), 55 (98), 53 [42), 43 (96), 41 (22), 39 (loo), 29 (61), and 27
(41), to the one of authentic 2 (prepared ad cited under ref 11).
Quantitative VPC analysis (column A at 70 + 180 "C (30 OC/min)

-
9 and with 60 ml/min of He, using n-octane a2 intefnal standard) of
the filtrate revealed that 2 was formed in 80% yield and that 13%
cuo + cuc1, 2CuCl of the charged 1 was still present in the solution. This solution was
used directly for the preparation of 3 by reactjon with cupric chlo-
ride.
Structures 3 and 8 for these two products were confirmed (E)-4-Chloro-2-methylcrotonaldehyde(3). To the chloro-
by the NMR and ir spectra and the boiling points of the form solution obtained from the epoxidation of 1 and containing
isolated pure-compounds. However, for the a c e t ~ x y l a t i o n ~ ~20.2 g (0.24 mol) of 2 was added 240 ml of Qthylacetate followed by
step the isolated crude 3 (78%pure, yield corrected for pure 81.6 g (0.48 mol) of CuC1~2H20and 10.0 g (0.24 mol) of LiC1. The
mixture was refluxed (90 "C bath temperature) for about 15 min
3: 80%) was not distilled since the impurities consisted and then poured onto 240 g of ice. After filtration from the CuC1,
mainly of residual solvents and the distillation resulted in the organic phase was separated and the aqueous layer extracted
considerable decomposition of the product (57% yield). The with 240 ml of hexane. The combined organic extracts were
acetoxylation afforded, after removal of low-boiling im-
purities from the isolated crude product, a 90% yield of 4,
which did not require distillation, being already quite pure
VPC analysis (column B a t 60 OC for 1 min then 60
OC/min) and with 15 ml/min of He, using
-
washed to neutrality with water and dried (NazS04). Quantitative

n-octane
200 "C (30
as internal
standard) of the organic solution revealed formation of 3 and 8 in
(95%). 80 and 20% yield, respectively. The greater part of the solvents and
When 2 was treated with cupric chloride in chloroform- the aldehyde 8 were then distilled from the organic solution by
ethyl acetate in the absence of lithium chloride the reaction concentration, first at 10-20 mm and room temperature and final-
was quite slower (2.5 h instead of 10-15 min), the amount ly a t 120 mm and 40 OC bath temperature (in both cases the distil-
of 8 formed was practically the same (23%), and the purity lation was monitored by VPC to prevent distillation of 3), leaving a
as well as the yield of the isolated crude 3 were somewhat residue, 29 g, of crude 3 (78% pure by VPC, the balance to 100%
being mainly residual solvents; corresponding to an 80% corrected
lower (70% purity, 72% corrected yield). The yield of the yield). Distillation of the crude product gave 16.2 g (57%) of 95%
distilled 3 was 50%. pure (VPC) 3: bp 39-42 "C (0.5 mm) [lit.ls 41-43 "C (0.5 mm)];
By treatment of 8 with cupric chloride in the presence NMR (CDC13) T 0.57 (s, CHO), 3.43 (tq, J = 6.5 and 1.5 Hz,
(or absence) of lithium chloride and in chloroform-ethyl C=CH), 5.67 (dq, J = 6.5 and 1 Hz, CHZCl), 8.23 (dt, J = 1.5 and
acetate solution, using the same reaction conditions as for 1 Hz, CH3); ir (film) 2710, 1685, 1645 cm-'.
When the distillates obtained in the isolation of the crude 3 (see
2, the starting material was recovered practically un-
above) from a few runs were combined and concentrated first a t
changed.16 Therefore, in the oxidative chlorination of 2 to 3 atmospheric pressure and 85 "C bath temperature (until incipient
under these conditions the intermediacy of 8 can be exclud- distillation of 8, VPC monitoring) and then a t 40 mm and 0 "C (to
ed and the formation of the latter attributed only to a side a 1:l mixture of 8 and ethyl acetate, VPC monitoring), and the res-
reaction. idue distilled at atmospheric pressure, some pure 8 could be ob-
Finally, when the cuprous and lithium chlorides recov- tained: bp 116-117 OC (lit.19 116.2 "C); NMR and ir spectra identi-
ered together from the oxidative chlorination of 2 were sub- cal with the ones reported in literature for 8.z0,z1
When the oxidative chlorination 'with CuC12-2HzO of the chloro-
mitted to air oxidation in aqueous hydrogen chloride (to re- form solution of 2 was repeated in the absence of LiCl', 3 and 8
generate the cupric chloride) and recycled several times, were formed, after 2.5 h reflux, in 72 and 23% yields, respectively
the yields of 3 (and 8) remained unchanged. (quantitative VPC analysis of the organic extracts). The yields of
In conclusion, peracetic acid epoxidation of 1 followed by the isolated, 70% pure, 3 amounted to 72% (corrected for pure 3)
cupric chloride oxidative halogenation of the resulting solu- and the one of distilled, 95% pure, 3 to 50%.
tion and final acetoxylation of the q u d e 3 represents, on Acetoxylation of 11.85 g of the 78% pure 3 (corresponding to
0.078 mol of 3) dissolved in 100 ml of anhydrous ethanol was car-
comparison with the previously known ~ y n t h e s e s , l -the ~ ried out with 8.42 g (0.086 mol) of potassium acetate keeping the
most efficient route to 4. mixture at reflux for 8 h. From the cooled (0 "C) mixture the pre-
cipitated KC1 was removed by filkration, the filtrate concentrated
Experimental Section under vacuum, and the residue taken up in ether in order to dis-
Materials. Isoprene obtained from Fluka (purum a) was puri- solve the product from residual KCl. Evaporation of the solvent
fied by distillation over sodium; 82% peracetic acid was prepared from the filtered solution gave, after removal of low-boiling im-
according to Swern.17 purities a t 2-3 mm and room temperature, 10.5 g (90%, based on
General. Gas chromatography-mass spectrometry analysis was
performed with a Varian MAT 111 instrument under the following ready quite pure (95% by'\iPC using column B at 80 -.
0.078 mol of 3) of 4, which did not require distillation, being al-
180 PC (10

mesh) column, at 40 OC for 3 min then 40 -

conditions: 6 ft X 0.125 in. 3% OV-1 on Chromosorb W (80-100
180 "C (20 "Clmin)
and with 24 ml/min of He; ionizing energy, 70 eV. Quantitative
OC/min) with 15 ml/min of He and n-decane as internal standard).
The NMR and ir spectra and the boiling point of the product ob-
tained were identical with the ones reported in the literature for
VPC analyses ' were performed with a Hewlett-Packard Model 4.22'
7620-A gas chromatograph equipped with a thermal conductivity
detector. The following columns Gere employed: (A) 10 f t X 0.125 Acknowledgments. The authors wish to express their
in. 10% Carboaax 20M on silanized Chromosorb G (60-80 mesh); thanks to Mr. B. Biancini for his technical assistance in
(B) 6 ft X 0.125 in. 4% SE-30 on silanized Clpomosorb G (60-80 this investigation. Thanks are due to Dr. A. Robertiello for
mesh). The 60-MHz NMR spectra were recorded on a Varian T-60 the gas chromatography-mass spectrometry analysis and to
spectrometer with tetramethylsilane as internal standard. The fol- Dr. L. Settembri for the NMR spectra.
lowing designations were used: s, singlet; dt, doublet of triplets; dq,
doublet of quartets; tq, triplet of quartets; qq, quartet of quartets. Registry No.-1, 78-79-5; 2, 1838-94-4; 3, 26394-25-2; 8, 497-
The ir spectra were taken with a Perkin-Elmer 457 spectrometer. 03-0; CUC12,7447-39-4.
1650 J.Org. Chem., Vol. 41, No. 9, 1976 Notes

References and Notes hydroxyl group in oxirane ring opening. Indeed, this was
(1) W. Reif and H. Grassner, Chem.-hg.-Tech., 45, 646 (1973). found to be the case since epoxidation of l a with m-chloro-
(2) H. Mayer and 0. lsler in "Carotenoids", 0. Mer, Ed., Birkhauser-Verlag, perbenzoic acid in methylene chloride followed by aqueous
Basel, Switzerland, 1971, pp 394-397. work-up resulted in a mixture of the desired epoxide 2a
(3) Badlsche Anilin- & Soda-Fabrik A.-G., French Patent 1 587 943 (1970);
Chem. Abstr., 74, 31556t (1971). and the glycol 4d. It is interesting to note that previous
(4) H. K. Dietl. J. R. Normark, D. A. Pavne, J. G. Thweatt, and D. A. Youna. workers have reported similar problems in the synthesis of
Tetrahedron Lett., 1719 (1973).
(5) A. N. Pudovik and B. E. Ivanov, Zh. Obshch. Khlm., 26, 2771 (1956). this epoxide.2 In contrast, no difficulty was experienced in
(6)For epoxidations with peracetic acid in the presence of bases see M. the epoxidation of picrotoxinin 6-acetate (lb) to give 2b by
Korach, D. R. Nielsen, and W. H. Rideout, J. Am. Chem. SOC.,82, 4328 the same procedure. Hence picrotoxinin epoxide (2a) ap-
(1960).
(7) Use of excess peracetic acid in order to complete the conversion of 1 pears to be much more reactive than 2b. Because of this re-
resulted In lower yields of 2. activity the epoxide 2a was not isolated in pure form but
(8) Ethylene and propylene oxides give thq corresponding chioro epoxldes
by liquid-phase photochiorination with tert-butyl hypochlorite; see C.
was used, after a modified work-up procedure, directly in
Walling and P. S.Fredricks, J. Am. Chem. SOC.,84, 3326 (1962). subsequent reactions with amines. On treatment of 2a with
(9) For related rearrangements of chloro epoxides see H. 0. House, "Mod- pyrrolidine or diethylamine at room temperature, the cor-
ern Synthetic Reactions", 2d ed, W. A. Benjamin, Menio Park, Calif.,
1972, pp 313-314. responding amine derivatives 4a and 4b were obtained
(IO) Liquid-phase photochlorination of 2 with tert-butyl hypochlorite under whereas chloroform reflux conditions were required to ob-
various conditions was unsuccessful. tain 4c from 2b with diethylamine.
(11) For these experlments, because of difficulties in isolating the product
from the solvent in the peracetic acid epoxidation of 1, pure 2 was best
prepared according to E. J. Reist, I. G. Junga, and B. R. Baker, J. Org.
Chem., 25, 1673 (1960); the only modification being that the dehydro-
bromination was performed by treatment of the distilled Isoprene brom-
ohydrin (bp 60 OC, 10 mm) with solid sodium hydroxide at 90 OC and at-
mospheric pressure in a distillation flask equiped with a Vigreux frac-
tionating column. Sodium sulfate drying of the distillate (bp 72-73 "C)
afforded pure 2.
(12) For a review on oxidative halogenations with cupric halides see W. G.
Nigh in "Oxldation in Organic Chemistry", Vol. 5,Part B, W. S. Trahan-
ovsky, Ed., Academic Press, New York, N.Y., 1973, pp 67-84.
(13) For the successful use of the chloroform-ethyl acetate solvent system
in oxldative halogenations of carbonyl compounds with a cupric halide
see literature cited under ref 12. The use of other solvents in the cupric
chloride halogenation of 2l gave less satisfactory results. la, R = H (picrotoxinin) 2a,R=H
(14) Lithium chloride is a catalyst of oxldative halogenations of carbonyl
compounds with cupric chloride: see literature cited under ref 12. b,R = COCH, b, R = COCHj
(15) For the acetoxylation of 3 to 4, a procedure (see Experimental Section)
different from the one cited under ref 2 gave more satisfactory results.
(16) The reaction conditions, especially wlth regard to solvents, used by the
authors cited in ref 4 for the conversion of 8 to 3 with cupric chloride
were quite different.
(17) D. Swern, "Organic Peroxides", Vol. 1, Wiley-Interscience, New York,
N.Y. 1970, p 481.
(18) Badische Anilin- & Soda-Fabrik A,-G., German Patent 1 188 577 (1965);
Chem. Abstr., 82, 13049c (1965).
(19) M. T. Rogers, J. Am. Chem. SOC.,69, 1243 (1947).
(20) A. W. Douglas and J. H. Goldstein, J. Mol. Spectrosc.. 16, 1 (1965).
(21) S. Satsumabayashi, K. Nakajo, R. Soneda, and S.Motoki, Bull. Chem.
SOC.Jpn., 43,-1586 (1970). A
(22) G. Pattenden, J. E. Way, and B. C. L. Weedon, J. Chem. SOC.C, 235
(1970). 3 4a,R = H; R' = N
3
b, R = H; R' = N( CZH,),
C, R = COCH,; R' = N(CZH6)t

d, R = H R' = OH
e, R = H; R' = H (picrotin)

The ease of opening of the epoxide in 2a with amines to

Hydroxyl Assisted Epoxide Opening in Picrotoxins yield compounds of type 4 can be ascribed to the neigh-
Haldean C. Dalzell, Raj K. Razdan,* and Reuben Sawdaye boring 6-hydroxyl group participation in an "anionic" pro-
cess. Only a few examples of such participation are known
Sheehan Institute for Research, Inc., such as the neighboring hydroxy group participation in the
Cambridge, Massachusetts 02138 alkaline hydrolysis of and more recently the open-
Received December 9,1975 ing of epoxides by nucleophiles in steroids.6 Neighboring
group participation in ('cationic'' reactions, on the other
It is well documented that picrotoxinin (la) is the potent hand, has been known for many years and is well under-
analeptic component of picr0toxin.l In order to modify the ~tood.~,~,~
analeptic properties of la, we were interested in the intro- The structural assignment of these compounds (4a-c)
duction of a basic nitrogen moiety in the picrotoxinin mole- seems secure on the basis of spectral data. The NMR spec-
cule. Picrotoxinin possesses an epoxide ring which should tra of 4a and 4b are compared in Table I with those of pi-
theoretically be opened with amines or other nucleophiles, crotoxinin ( l a ) and picrotin (4e).The NMR spectra of the
but is known to be very resistant to intermolecular nucleo- latter two important compounds have apparently not been
philic attack. This unusual feature of the picrotoxinin reported previously. Infrared spectra of the amines show
structure is explained on the basis of shielding of the epox- that the lactone groupings are maintained in the picrotoxi-
ide ring from rearward attack by the lactone groupings. nin epoxide molecule during reaction with amines. The
However, the close proximity of the axial C-6 a-hydroxyl mass spectrum of 4a showed principal ions a t mle 380 (M
group to the C-4 a-isopropenyl group in 1 suggested to us + l)+,379 (M-+),280, 128, and 84 (base). The base peak a t
that the 8,g-epoxide of picrotoxinin (Le., 2a) would be mle 84 and the peak at mle 128 correspond to the ions i
prone to nucleophilic attack due to participation by the C-6 and ii, respectively, both of which confirm the presence and