Jan.

5 , 1959 THEXECHANISM
OF THE PRINSREACTION 133

the basis of the extinctioa coefficients, the p- made the infrared absorption spectral measure-
form has a higher enol content than that of the a- ments and of Mrs. Barbara B. Martin who made
form. The fact that the extinction coefficient of the ultraviolet absorption spectral measurements.
the P-form remains fairly constant over a wide The authors are indebted to the Union Carbide
concentration range may be due t o the existence Chemicals Co. for supplying the acetylacetone and
of the P-form in a stable configuration. to Dr. Joseph A. Dixon for supplying the 1,lO-
Acknowledgments.-The authors gratefully ac- dibromodecane.
knowledge the help of 113r. Richard E. Dagle who CXIVERSITY
PARK,PENS.+.

[ COXTRIRUTIOS
FROM THE IXSTITCTE
OF ORGANIC THEUSIVERSITPOF CHICAGO]
CHEMISTRY,

The Mechanism of the Prins Reaction]

B Y N. c. YtlNG, DING-DJUNGH. YANGAND CAMILLA B. ROSS
RECEIVEDJULY 31, 1958

The stannic chloride-catalyzed condensation of some trisubstituted olefins with formaldehyde in chloroform was investi-
gated. The products obtained from 2-methyl-2-butene and from l-ethoxy-3-methyl-2-butene were 2,3-dimethyl-3-buten-
1-01 and l-ethoxy-2-hydroxymethyl-3-methyl-3-butene, respectively, and their corresponding formals. The structures of
these products, established by infrared spectroscopy and chemical degradations, differ from those predicted by a previous
mechanism. X mechanism involving general acid catalysis is suggested for the induced condensation of olefins with alde-
hydes. Catalysts such as anhydrous metal halides and aqueous acids result in different products depending on the solvolytic
action of the reaction media.

The condensation of olefins with aldehydes is corresponding 1,3-glycol and formaldehyde is a re-
catalyzed by acids, heat, light and peroxides, and versible process. Since the 1,3-dioxane and the
those catalyzed by acids are referred to generally as 1,3-glycol are not readily interconvertible under the
the Prins reaction. The reaction may be carried conditions of the condensation, the two products
out in an aqueous or an anhydrous medium, and must be produced i n d e ~ e n d e n t l y . ~A~mechanism
the major products consist of a 1,3-dioxane, 1,3- (mechanism A) which involves the addition of a
glycol or an unsaturated alcohol depending on the protonated formaldehyde molecule to the olefin was
experimental condition^.^,^ Because of this un- p o ~ t u l a t e d . ~The intermediate adduct I1 may re-
usual versatility, the Prins reaction is useful in or- act with formaldehyde or solvent t o give a mixture
ganic syntheses. It offers a new route t o the manu- of the 1,3-dioxane and the 1,3-glycol.
facture of many organic compounds utilizing read- Mechanism A
ily available petroleum product^.^ Despite the
fact that the Prins reaction has been employed fre- HCHO + Ha0e! r’CH?OH
@
+ HsO
quently for synthetic purposes, the condensation in fE
aqueous and anhydrous media was believed to pro- “‘)C=C<; + CH2OH ----f
ceed by different mechanisms,5-’ and different CHa
products were derived accordingly. There is some
controversy as to the actual mechanism involved
and experimental evidence is scarce.K,6b
When an olefin is condensed with formaldehyde Rz H
in an aqueous medium using mineral acids as cata- \ /
R, C
lysts, a mixture of a 1,3-dioxane and a 1,3-glycol is
produced.R If the condensing olefin is trisubsti-
tuted, a 1,3-diene is obtained a d d i t i ~ n a l l y . ~The
acid-catalyzed hydrolysis of a 1,3-dioxane to the
‘C/
(1) Presented in part a t the 133rd Meeting of the American Chemi-
H.,
cal Society, April, 1928, San Francisco, Calif.
R1 IV
( 2 ) For a review on the P r i m reaction, see E. Arundale and L . A
Mikeska, Chem. Revs.,51, 505 (1952).
Ho&CR,HCH20H
(3) A. T. Blomquist, et Q L , THISJ O U R N A L , 79, -1972 (l957), and later
papers. I
CHI I11
(4) E. Arundale and L. A. Mikeska, U. S. Patent 2,312,743, C. A , ,
37, 4749 (1943); E. Arundale and L. A. Mikeska, U. S. Patent 2,360,- When the condensation is carried out under an-
486, C. A , 3 8 , 4 9 5 7 (1944); E. Rosen, U. S. Patent 2,335,029, C. A , ,38, hydrous conditions with stannic chloride as a cata-
3834 (1944).
(5) C . C. Price, “Mechanisms of Reactions a t Carbon-Carbon Dou- lyst, an unsaturated alcohol is produced.1° This
ble Bonds,” Interscience Publishers, Inc., New York, N. Y., 1946, p. 45. alcohol could be dehydrated readily to the corre-
( 6 ) (a) J. W. Baker, J . Chem. Soc., 296 (1944); (b) J. W. Baker, sponding conjugated diene,?b and was claimed to
Saluue, 161, 171 (1948).
(7) (a) Reference 2, p . 516; (b) p. 539.
have the general formula VI,l0J1isomeric with that
(8) H. J. Prins, Pro. A c a d . Sci. Amslerdarn, 22, 51 (1919); H. J . (IO) L. A. Mikeska and E. Arundale, U.S. Patent 2,308,192, C. A , ,
Prins, J . Chem. SOL., 118,4 2 (1920); ref. 2, p. 507. 37, 3450 (1943); L. A. Mikeska and E. Arundale, British Patent
(9) E. Arundale and L. A. Mikeska, U. S. Patent 2,350,485, C. A., 545,191, C. A . , 36, 7030 (1942).
38, 4957 (1944); R. Rosen and E, Arupdale, U. S. Patent 2,368,494, (11) L. A. Mikeska a n d E. Arundale, U.S. Patent 2,253,342, C. A , ,
C. A , , 39, 4529 (1945). 56, 7974 (1941).
predicted by mechanism A. On the basis of this Mechanism C
evidence, a different mechanism (mechanism B) @ E?
which directly involves an “active a-hydrogen” was HCHO + .I I12C--0 -Ai
postulated6
Mechanism B

SnC14 I
ti CHa
R1 \ -
VI1 H

\-I
solvent or HCHO
.1 .1
other products RI
Other metal halides including aluminum chloride,
zinc chloride12and boron trifluoride13have been em- EI~C==ACHR~CH~OII+ .I
ployed as catalysts in the Prins reaction. All are YIII
Lewis acids and are electrophilic “with respect to where X = SnC1, or other :~cjds
olefins and carbonyl compound^.^^ The major and careful investigation of the condensation prod-
difference between the two suggested mechanisms ucts. An ordinary kinetic approach was not suc-
may be ascribed to the postulation of an initial as- cessful because of the complicated nature of the
sociation between the proton and the aldehyde or condensation.6a
an association between the metal halide and the Although i t was claimed that the condensation of
olefin which induces the condensation, It is un- formaldehyde with many trisubstituted olefins
likely that acids in different media catalyze the produced alcohols of formula VI, no experimental
condensation by different mechanisms and there is proof of the proposed structures was reported. The
indication that the reaction does proceed by general condensation of formaldehyde with some simple
acid catalysis. Assuming a similar catalytic effect trisubstituted olefins was reinvestigated. In the
for all acids, we wish to propose a general mecha- case of 2-methyl-2-butene (Ia, R1= R2 = CH,), 2,3-
nism (mechanism C) to explain the catalysis by dimethyl-3-buten-1-01 (VIIIa, R1 = Rz = CH3) was
metal halides in anhydrous media. Stannic chloride isolated instead of the reported 3-methyl-3-penten-
reacts initially with formaldehyde to form a polar- 1-01 (VIa, R1 = RP= CH3). n’hile the physical con-
ized addition complex. The complex then adds t o stants of VIIIa are identical to those reported pre-
the olefin in accordance with the Markownikoff viously for TI‘Ia,Io its infrared spectrum shows a
rule. Finally, stannic chloride is regenerated by strong absorption a t 890 ern.-' and no absorption
a simultaneous proton transfer through a six-mem- at 800-840 cni.-’ region indicating the presence of a
bered ring VI1 giving rise to the vinylidene unsaturation. The correct structure of
corresponding 6,y-unsaturated alcohol VIII. This this alcohol was confirmed by degradations. Ozon-
mechanism is applicable to the acid-catalyzed con- olysis of VIIIa followed by hydrogenation yielded
densation of olefins and aldehydes in an aqueous or formaldehyde (46-.%?~Yob) and a C6-methyl ketone.
an aprotic medium. The C6-methyl ketone was shown to be p-hydroxy-
According t o the reaction schemes outlined isopropyl methyl ketone by (1) reverse aldol con-
above, i t is apparent that disubstituted olefins of densation to formaldehyde and 2-butanone, (2) un-
the type CH3CR1=CH2 will give identical alcohols ambiguous synthesis through the aldol condensa-
by mechanisms B and C (VI = VIII if Rz = H), tion of its components and (3) dehydration in acid
whereas trisubstituted olefins of the type C H ~ C R I = solution to isopropenyl methyl ketone.
CHRz will give different alcohols : R*CH=CRI- In addition to the unsaturated alcohol VIIIa
CH2CH20H (VI) by mechanism B and CHF which constituted the major product (26-29%) of
CRICHR2CH20H (VIII) by mechanism C.16 the condensation, there was obtained the formal of
Hence, the two mechanisms can be differentiated VIIIa (19-23%0) and no appreciable quantity of
readily by suitable choice of the reacting olefin the 1,3-dioxane was isolated. The formal X shows
(12) Reference 2, p. 532. a strong infrared absorption a t 885 cm.-’, indica-
(13) D. J. Loder, U. S. Patent 2,158,031, C. A . , 33, 83-13 (1939): tive of a vinylidine double bond, and at 1040 and
D . J. Loder, U. S. Patent 2,289,548, C.A , , 37, 388 (1943).
(14) G. W. Wheland, “Advanced Organic Chemistry,” John Wiley
1110 cm.-l, indicative of an ether function. On hy-
and Sons, Inc., New York, N. Y . , 1949, p. 80. drogenation, X absorbed two equivalents of hydro-
(15) R . T. Arnold, He2u. Chirn. Acto, 31, 134 (1949). gen to give a saturated formal (XI). When X was
(16) The deprotonization step in mechanisms B and C may involve subjected to ozonolysis followed by hydrogenation,
the methylene group in RI, depending on the structure of the alkyl sub-
stituent. In this case, the general statement remains true, and the
formaldehyde and a Cll-methyl ketone were iso-
predicted alcohols will have different formulas accordingly. This lated. The CI1-ketone was identified by its bis-2,4-
possibility is eliminated if Ri = CH:. dinitrophenylhydrazone which could be converted
Jan. 5 , 1939 THEh f E C H A N I S M OF THE P R I N S REACTION 135

CH, and the corresponding formal. For simple olefins

I SnCla of the type (CH&C=CHR, the results appear to be
CHaC=CHCH3 + HCHO in CHC13 general, and the following conclusions are drawn:
Ia The structure of the unsaturated alcohol supports
CH3 CHI CHI mechanism C. Therefore, the Prins reaction in
I I 1 aqueous and anhydrous media apparently follows
HeC=CCHCHz + HsC=C---CHCHz02CH% a general mechanism which involves an acid-ac-
I tivated formaldehyde intermediate. The observa-
CH20H
VIIIa X tion that only one isomeric unsaturated alcohol is
produced during the condensation favors an intra-
CH3 molecular proton transfer in a cyclic manner as
VIIIa
1 0 3
-& HCHO
I
f O=CCHCH, suggested above.
2, Hz I Acknowledgment.-The authors wish to thank
I X CHzOH Mr. William Saschek for the microanalyses. A
I X e HCHO +
CH3CHzCH3 fellowship from E. I. du Pont de Nemours & co. to
L i TH3
T
CH3C-C=CH2 + HnO
C. B. R. is gratefully acknowledged.
Experimental"
readily to the 2,4-dinitrophenylhydrazoneof isopro- The hins Reaction of 2-Methyl-2-butene (Ia).-To a
stirred mixture of 70.1 g. (1.0 mole) of freshly distilled 2-
penyl methyl ketone in the presence of acid. Direct methyl-2-butene and 28.0 g. (0.31 mole) of trioxymethylene
hydrolysis of X in acid gave formaldehyde and in 60 ml. of anhydrous chloroform,I8 a solution of 7.9 g.
alcohol VIIIa, isolated, respectively, as its dimedone (0.030 mole) of freshly distilled stannic chloride in 15 ml. of
derivative and 3,5-dinitrobenzoate and the structure chloroform was added. The mixture was stirred a t room
temperature for 18 hours. At the end of this period, most of
of formal X was thereby confirmed. the trioxymethylene was dissolved and the unreacted portion
CHI CH3 CHI CH3 (4-7 g.) was removed by filtration. The filtrate was treated
immediately with a mixture of dilute ammonium hydroxide
I I Hz I 1 and ice, and the P H of the mixture was adjusted to 4. T h e
(HzC=C-CHCH20)2CH2 +
i (CHZCH-CHCHz0)2CH2 organic layer, combined with the chloroform extracts of the
SI aqueous layer, was washed thoroughly with 5y0 aqueous
sodium bicarbonate and water, dried, and fractionally dis-
CH3 CH3 CH3 CHa tilled through a tantalum spiral column (5 mm. X 60 cm.).
After the removal of a fore-run which contained mainly the
I I I 1
2HzC=CpCHCH20H '-+
(O=C-CHCH~O)ZCHZ unreacted olefin and some dissolved formaldehyde, two
major fractions were obtained: Fraction I, an alcohol ( V I I I a )
VIIIa +
HCHO XI1 +2HCHO of b.D. 66-66.5' a t 31 mm. or 132-133' a t 749 mm. (b.p. re-
hydrolysis ported'O 135-145' at atmospheric pressure), ~ * O D 1.4391,
bis-2,4-DNP of X I 1 - --f weight 20-23 g. (yield 26-29y0 based on formaldehyde con-
sumed). T h e alcohol shows characteristic infrared absorp-
0 CHs tion a t 3350,3060,1645 and 890 ern-'. Gas chromatography
I1 I of VIIIa through a G. E. Silicone SF96 on firebrick column
2,4-DNP of CHaC-C=CHz 4-HCHO f 2H20 (1/4" X 6 f t . ) and a polypropylene glycol on firebrick
where 2,4-DNP = 2,4-dinitrophenylhydrazone column (1/4" X 6 ft.) shows the presence of only a single
comDonent.
Likewise, the condensation of 1-ethoxy-3-methyl- Anal. Calcd. for C6H120: C, 71.63; H , 12.04. Found:
2-butene (Ib, R1 = CHI, Rz = CHzOC2HJ and C, 71.95; H , 12.08.
formaldehyde gave l-ethoxy-2-hydroxymethyl-3- Fraction 11, a formal (X) of b.p. 77-78" a t 2 mm., nmD
1.4462, weight 11-13 g. (yield 19-23y0 based on formalde-
methyl-3-butene (VIIIb) instead of the reported hyde consumed). The formal shows intense infrared absorp-
5-ethoxy-3-methyl-5-penten-1-01.l1 Structure VIIIb tion a t 3060, 1120,1050 and 890 cm.-'.
was assigned to the unsaturated alcohol on the basis Anal. Calcd. for CL3H2402: C , i3.53; H , 11.39; mol.
of infrared absorption a t 890 cm.-' and degradative wt., 212. Found: C, 73.50, 73.30; H , 11.60, 11.10; mol.
ozonolysis which gave over 40% of formaldehyde wt., The 206.
of 2,3-Dimethyl-3-buten-l-ol
and a methyl hydroxyketone. Because of the in- (VIIIa). a.Characterization &Hydroxyisopropyl Methyl Ketone (IX) .-A
stability of this methyl 0-hydroxy-0-ethoxy ketone slow stream of ozone (0.38 millimole per minute) was passed

-
with respect t o 0-elimination, it was not further into a solution of 4.3 g. (0.043 mole) of VIIIa in 50 ml. of
characterized. ethyl acetate a t -45'. When the uptake of ozone had sub-
sided, the solution was diluted with a n equal volume of 9570
SnC14 ethanol and hydrogenated in the presence of 2 % palladium
-t H C H O oxide on C a C 0 3(0.45 g.). The reaction mixture was filtered
in CHC13 and the filtrate was concentrated by distillation a t atmos-
Ib CHI CH20C2Hs pheric pressure. T h e distillate was collected under the sur-
\ / face of 10 ml. of aqueous methanol (1: 1). To the residue,
C-C-H 20 ml. of methanol was added, and the distillation was con-
/ \ tinued in the same manner. A good yield (53%) of the di-
HzC CHIOH medone derivative of formaldehyde (m.p. 109-191O) could be
VIIIb isolated from aliquots of the distillate mixture.
T h e residual oil was fractionally distilled through a semi-
T o avoid effects of small traces of moisture and micro Vigreux column and 2.3 g. (30y0) of 0-hydroxyisn-
protolytic solvents on the reaction mechanisms,6b propyl methyl ketone (IX) was obtained, b.p. 64-66' a t 7
mm., n Z o 1.4320.
the condensation was carried out in absolute chloro- test. ~ T h e ketone gave a positive iodoform
I t s infrared absorption spectrum was identical to
form utilizing trioxymethylene and freshly distilled
stannic chloride. Under these conditions, the con- (17) T h e melting points reported were not corrected, and the
molecular weights were determined by cryoscopic measurements in
densation products invariably consisted of the same benzene.
alcohol with unsaturation a t the terminal position, (18) Analytical reagent, Fisher Scientific Co.
136 N. C. I'ANG,
DING-DJL-SG
H. I-.ING
;\KD C.\MILLA
E . ROSS i'ol. SI

that of a n authentic sample (b.p. 70" a t 9 nim., nZoD1.4318) Xu 8O-nig. saiiiple uf the bis-8,4-dinitrophen~.Ihydrazone
prepared by the aldol condensation of formaldehyde and 2- ( X I I ) dissolved in a mixture of 0.20 ml. of concentrated
butanone.1g hydrochloric acid, 5 ml. of dioxane and two drops o f water
.A 2,4-dinitrophenylhydrazi~neof the ketone was prepared as warmed gently on a steam-bath for 25 minutes. After
f r m i a methanolic hydrochloric acid solution, and ail 8.55; removal of solvents a t reduced pressure, the residue was re-
yield of red needles was isolated, m.p. 185-186' dec. The crystallized from ethyl acetate-ethanol mixture and 2<5 mg.
derivative was shown to be the 2,4-dinitrophenylhydrazone of fine red needles was obtained. A mixed melting point de-
( i f the dehydrated isopropenyl methyl ketone by analyses termination with a n authentic sample of the 2,4-dinitro-
and infrared spectrum. phenylhydrazone of isopropenyl meth>-l ketone gave no de-
Anal. Calcd. for CllHI2iS404: C, 49.99; H , 4.58; S . pression (m.p. 184-lS5"j. I t s infrared absorption spectrum
21.21. Found: C, 49.97; H . 4.62; h7,21.26. in KBr was identical with that of an authentic sample.
b. The Reverse Aldol Condensation of IX.-Two grama c. Hydrolysis of X. --.\ solution of 0.34 g . (0.001; miile)
of IX isolated from the ozonolyses of VIIIa was treated for of S dissolr-ed in 5 nil. r i f aquccius dioxane was treated with
20 minutes over a steam-bath with 30 tnl. of 3('b aqueous three drops [ i f concentrated hydri)chIciric acid and the mis-
ii(tassium bicarbonate. The mixture was steam distilled lure was \varmcti on a steam-bath for 1 L5 minutes. .lfter
until 150 ml. of the distillate had been collected. After re- neutralizaticni with piitassium carbonate, the mixture was
moval of a small quantity of a n immiscible oil, the distillate divided into two equal portions from which the dimedone
was treated with a lOy, excess of Tollens reagent. The re- derivative of formaldehyde (m.p. 189-191") and a 3,5-dini-
sulting clear solution, separated from the reduced metal b y trobenozate were prepared. Two recrystallizations of the
decantation, was neutralized with 6 N hydrochloric acid and benzoate from aqueous methanol yielded 15 mg. of glittering
the precipitated silver chloride was removed by centrifuga- flakes (m.p. 40-41') identified as the 3,5-dinitrobenzoate of
tion. Fractional distillation of the aqueous layer a t atmos- %,3-dimethyl-3-buten-l-o1 ( V I I I a ) by mixed melting point
pheric pressure gave a ketone fraction boiling between 80- determination.
90". -12,4-dinitrophenylhydrazone of the ketone was pre- Anal. Calcd. for Ci3Hl4S2O6:C, 53.06; H , 4.80; N,
pared; weight 0.39 g., corresponding t o a n over-all yield of 9.52. Found: C , 53.14; H , 4.99; S ,9.70.
EGG. The melting point was not depressed on admixture The Prins Reaction of 1-Ethoxy-3-methyl-2-butene ( I b ) .-
with a n authentic sample of the 2,3-dinitrophen!-lhI\-drazone .I stirred suspension of 17.1.g.(0.19 mole) of trioxymethylene
of 2-butanone (m.p. 11;-118"). in 51.5.g. (0.45 mole) of IbM and 167 ml. of anhydrous chloro-
The Characterization of X, the Formal of 2,3-Dimethyl-3- form was treated at room temperature with a solution of 7
buten-1-01. a. Hydrogenation of X.--A solution of 1.09 g. ml. of chloroform containing 6.6 g. (0.0245 mole) of freshly
(0.0050 mole) of the unsaturated formal X in 15 ml. of gla- distilled stannic chloride. .It the end of 20 hours, the un-
cial acetic acid was hydrogenated in the presence of platinum reacted trioxymethylene (3.5 g.) was removed and the reac-
oxide and the hydrogenation was complete after two equiva- tion mixture was worked up by the regular procedure. The
lents of hydrogen was absorbed. The reaction mixture was crude product was fractionally distilled through a
wiirked up 1))- the usual procedure, and the crude residue, \.igreux column and there was obtained 14.9 g. (23y0based
dissolved in a small amount of petroleum ether, was chro- on formaldehyde consumed) of T'IIIb (b.p. 66-67'" at 2.8
matographed through aluminum oxide. The major frac- nim.). The product showed infrared absorption maxima a t
tion eluted with petroleum ether was concentrated and the 3120,3070, 1646, 1115, 1035 and 890 crn.-'. -Isample was
residue WAS distilled a t 20 mm. pressure and a n oil-bath tem- refractionated through a 60-cm. tantalum spiral column for
perature of 90-9.5". The product showed intense absorption analyses; b.p. j3-53.5" at 1.75 mm. (b.p. reported" 89-90'
a t 1120 and 1050 cm. - 1 and no absorption in the double bond :it 1.6nini.), X ~ " D1.4433.
region; n z O ~1.3401. Alncz/. Calcd. for CsHIGO?:C, 66.62; H , 11.19. Found:
d n d . Calcd. for CI3H&: C , 72.16; H, 13.05; i ~ i o l . C,66.54; H , 11.06.
wt., 216. Found: C, 71.79; H, 12.82; mol. wt., 209. -1second fraction (5.4 g., b.p. 72-85" at 0.05 mm.) which
b. The Formal of p-Hydroxyisopropyl Methyl Ketone.- showed characteristic absorption a t 3060, 1646, 1120, 1035
.i solution of 1.09 g. (0.0050 mole) of X in 12 ml. of ethyl and 890 ern.?, and the high boiling residue (11.0 g.) were
acetate was subjected t o ozonolysis followed by catalytic not investigated.
hydrogenation in the presence of 2yo palladium oxide on The Characterization of l-Ethoxy-Z-hydroxymethyl-3-
CaC03 (0.20 9.) and ethanol (15 m1.j. The reaction mix- methyl-3-butene (VIIIbj.-*1 solution of 5.0 g. (0.035 mole)
ture was concentrated by the method described previously of V I I I b dissolved in 45 ml. of ethyl acetate was treated with
for the isolation of P-hydroxyisopropyl methyl ketone. A ozone at -45O, and hydrogenated in the presence of 0.50 g.
good yield of the dimedone derivative of formaldehyde of platinum oxide and 45 ml. of 957, ethanol. The mixture
(weight 1.8 g., yield, SOY0, m.p. 190-19lo) was isolated from was concentrated by distillation a t atmospheric pressure
the distillate and a diketone (0.37 g., b.p. 7rj-80° at 0.25 according to the procedure described previously for the isola-
inm.) mas isolated from the residue. The diketone gave a tion of j3-hydroxyisopropyl methyl ketone, and 4.1 g. (40%)
positive iodoform test. -1 bis-2,4-dinitrophenylhydrazone of the dimedone derivative of formaldehyde (m.p. 191-192')
( X I ) was prepared at roum temperature by treatment with was isolated from the distillate. Fractional distillation of
a slight excess of a methanolic 2,4-dinitrophenylhydrazine the residue through a semimicro Vigreux column gave 2.4 g.
hydrochloride reagent. Two recrystallizations of the deriv- (51y0) of a hydroxy ketone (b.p. 78-86" at 14 mm.; y,,, a t
ative from chloroform-ethanol afforded a n analytical sam- 3440, 1707, 1115 and 1040 cm.-l). T h e ketone gave a posi-
ple (m.p. 178-179'), the analyses of which agreed with those tive iodoform test and decomposed readily to a resin on
required for the bis-2,4-dinitrophenylhydrazone derived standing. When the ketone was treated with a 2,4-dini-
from the formal of &hydroxyisopropyl methyl ketone. trophenylhydrazine reagent at room temperature, a mixture
Anal. Calcd. for C2aHdVaOlo: C , 47.91; H , 4.90; N, of 2,4-dinitrophenylhydrazones (m.p. 230-250') was ob-
19.44. Found: C , 47.89, 47.92; H , 4.87, 4.95; N, tained which could not be purified by recrystallization from
19.32. common organic solvents or chromatography.
(19) C. T. Morgan, N. J. L. Megson and K. L. Peper, Chemistry b CHICAGO 37, ILL.
~~.
..
I n d u s t r y , 6 1 , 885 (1938); C. T. Morgan and E. I,, Home?, J . Chem.
Sor., 2667 (1932). (20) 1,. Claisen, J . p r i i k t . Chcm., [ Z ] 106,80 (1922).