The Chemistry of Photodimers of Maleic and Fumaric Acid Derivatives

Irradiation of dimethyl fumarate in the solid state affords a dimer, cis,fraMs.cisT,2,3,4-tetracarbomethoxycyclobutane, whose stereochemistry can be r...
0 downloads 0 Views 554KB Size
DIMETHYL FUMARATE DIMER

June20, 1961 [CONTRIBUTION

KO. 1652 FROM THE

STERLING CHENISTRY LABORATORY,\‘ALE

2725

~ X I V E R S I T Y ,XEW HAVEN, CONN.]

The Chemistry of Photodimers of Maleic and Fumaric Acid Derivatives.l Fumarate Dimer

1. Dimethyl

BY G. W. GRIFFIN,A. F. VELLTURO~ AND I (.FURUKAWA RECEIVED JASUARY 24, 1961 Irradiation of dimethyl fumarate in the solid state affords a dimer, cis,tmns,cis-1,2,3,-1 tetrac~trl~o~nc:~hos].cyclobutane, whose stereochemistry can be rationalized in terms of direct bond formation between tiearest neighbor molccules in the crystal lattice of the mntiomer. Isomerization of this dimer t o the thermodynamically more stablc trans.trn?~s,tra?fs tetruester is readily achieved thermally. The reduction of t h e trans,trans,lra?rs-isomer a11d the hydrolysis of both isomers as well as their reactions with plienylrnagnesium brotnide were also studied.

The technique used in the dimerization of diI n the cours: of studies directed toward the synthesis of the elusive 1,2,3,4-cyclobutanetetra-methyl funlarate consisted of irradiating a thin carboxylic acid^^-^ to be employed as possible pre- layer of the mononier on the inner surface of a cursors for the unknowii tetramethylenec~clo- glass tube 1iiti-1a tight source which had been inbutane (I),* tetrakis- (diphenylmethylcn e) -cyclo- serted into the tube. ’The dimethyl fumarate was butane (11)9 3s well a s the tricyclic system 111, deposited from solution on the inner surface of we have irradiated several derivatives of maleic the tube by evaporation of solvent under a nitroand fumaric acid in the solid state with formation gen stream. T’he sysiem was thermostated a t 2n” and the irradiation mas continued for a period of of the corresponding cyclcbutane derivatives.’ 1-5 days. Under these cmditions dinieth.1 fumarate undergoes a 59% conversion to a stereoisomer of lJ2,3,4-tetracarbornethoxycyclobutane i m p . 144-143”). I t is noteworthy that only one product could be isolated from the xrddiation although four isomers are posiible. On inspecting thc X-ray crystallographic data for dimethyl fumaratc12 it appeared that the most reasonable structure for the dimer was the czs.tram,cis or “chair” form (Wa). Professor R. Criegte and c o - ~ o r k e r s ~ ~ Although numerous examples of solid state olc- have receiitly synthesized three of the four possible l i i i photodimerizations appear in the literature, lo isomeric 1,2,8,~-cyclobutanetetra~~rboxylic acids no broad study of the scope of this method for thc and thoir iiieth>l esters by ozonolysis of cinnamic preparation of substituted cyclobutanes has been acid dimers of known stereochemistry. The reported. The primary purpose of the work pre- fourth isomer possessing cis,c~s,cissiereocheintstry sented in this report is to determine the structure (1n.p. 203-205^) was obtained in our laboratories ~ i n dstudy the properiies of the dimer obtained from from ces-heptacyciene, an acenaphthylene dimer, dimethyl fumarate. However, it is hoped that by ozonolysis 14,15 S-T‘e are indebted to Professor these data coupled with crystallographic studies Criegee for confirniing by mixture melting point and results obtained with a serics of related olefins and infrared determinations that the dimcthyl will shed some light on the more general problem fumarate dimer does indeed possess the “chair” of the mechanism o€ olefin photodimerizations in structure IVa. the solid state. Of particular interest is the significance of monomer lattice geometry (i.c., iiiterinnleciilar distances and relative orientations) in allowing dimer formation and dictating product stereochemistry. l 1 (1) For a preliminary commitnication on this subject, see G. W , Griffin, J. E. Basinski and A. F.Vellturo, Tcfrahcdron Letters, N o , 3, 13 (1960).

(2) Predoctoral N.I.H. Fellow, 1960-1961. (3) 1’.Shibata, Ber., 43, ’2619 (1910). (4) J. Owen and J. I.. Sinionsen, J . Chem. SOL,1424 (1932); 1225 (1933). ( 5 ) S. R. Ranganathan, J . I n d i a n Chrm. S O L .19, , 419 (1930). ( 6 ) E. B. Reid and M. Sack, J . Am. Chem. Soc.. 73, 1985 (1951). (7) E. B. Reid, Chcmisfry 6’Industry, 846 (1953). (8) J. D. Roberts, A. Streitwieser, Jr., and C. M. Regan, J. Ani. Chcm. Soc., 74, 4579 (1952). (9) H. Shechter and R . 0. Uhler have recently published data which indicate that the photodimer of tetraphenylbutatriene possesses this structure; see Disserfalion Absfr., 21, 765 (1960). (IO) A. Muatafa, Chem. Reus., 61, 1 (1952). (11) For a preliminary report on a study of the effect of monomer crystal structure in determining the course of solid state reactions, see G. hl. J. Schmidt, Abstracts, 4th International Congress, International Union of Crystallography, Montreal, July, 1957, p. 65. I t is reported that the dimerizations of the alpha and beta modificatinns of franscinnamic acid to a-truxillic and 8-trurinic acids, respectively, may be

COAR IVa,R=CH3 b, R - H

The cis,tmns,cis-isorner is thermally stable up to 200” and for short periods a t 2 5 O C , but undergoes thermal isomerization at 300° in 50% yield rationalized in terms of the packing arrmgement of nearest neighbor molecules in the crystal structure of the monomer. (12) I . E. Knaggs and K. Lonsdale, J. Chem. Soc., 417 (1942). (13) R. Criegee, Private Communication; R . Criegee and H . Hdver, Bcr., 93, 2821 (1960). (14) G. W. Gri5n and D. Veber, J. Am. Chern. SOC.. 82, 6417 (1960). (15) Professor Criegre has recently been successful in obtaining cis,cis,cis-tetracarbomethoxycyclohutanc by ozonolysis of peritruxillic acid anhydride and subsequent esterification of the oxidation product with diazomethane. This compound has an infrared spectrum identical in all respects to that of 0111product.

2726

Vol. 83

G. UT. GRIFFIN,A. F. VELLTURO AND K. FURUKAWA

to an isomeric ester, m.p. 126-127". The thermally rearranged product was assigned structure Ira, having the tmns,trans,trans stereochemistry-, on the basis of melting point and infrared data for autheiitic T'a supplied by Professor Criegee.I3 Further confirmation of this assignment was obtained by n.m.r. analysis. The renrrangemeiit of IVa to the inore stable isomer can also be effected with sodium methoxide, albeit in much lower yields (18%). Essentially quantitative hydrolysis of the tetramethyl esters to IVb and Vb can be achieved with concentrated hydrochloric acid. Esterification of IYb with diazomethane to IVa established that the less stable isomeric ester IVa does not rearrange under these hydrolysis conditions. \Then heated a t its melting point in a sublimator the tetrabasic acid IVb loses water and affords a dianhydride which has been assigned the tram structure VI. Conversion of IITbt o T'I can be accomplished more effectively with thioiq-1 chloride. T h e assignment

Conversion of the tetraester IVa to the corresponding tetraacetate VIIIa was accoinplisheti in a one-step process designed to avoid the problenis associated with isolating t h e wcitcr-soluble alcohol YIIIb from an aqueous system In this procedure the alcoholate obtained from the reduction of IVa with lithium aluminum hydride in t h e inverse inanner was treated directly with acetyl chloride The resulting tetraacetate VIIIa after purific a t'i n i , by distillation, was coiir-ertccl to the benzoate ('1-I 2 N

("2X Y I I l a . x-=O A I ~ b. S - O H c, S ~ = O R / .

(~H,?i I X a . ST o.\c b. S- 013

VIIIc by basic hydrol;,-sis arid subsequent ti-i.atnient with benzoyl chloride. T h e results of L: study on the pyrolysis of the tetraacetate and thc. preparation of 1S:r. and 1x7)will be reported at :i later date. In connection with studies on the tetraacetate VIIIa i t is noteworthy that an isomer (1n.p. 10-1 105') tentatively assigned a cyclobutane structure of structure is supported by the infrared spectrum was obtairied by Raphael?O as a by-product from of VI which exhibits the characteristic unconjugated the partial catalytic hydrogenation of 1,4-diacetox~-anhydride peaks a t 3.38 and 5.60p. l6 Furthermore, I~utyiie-2. On the basis of physical properties it this anhydride is identical in all respects t o the is clear t h a t if this product is indeed a cyclobutailc. dimer obtained Yrom the solid state irradiation of derivative, i t does not possess the stereocheinistr:: maleic anhydride.' It is interesting, however, t h a t of V l I I a . -1preliminary examination, however, the "all-trans" isoineric tetrabasic acid Vb under- indicates that the n.m.r. spectrum of the tetragoes thermal isomerization and dehydration a t benzoate obtained by Raphael froni the tetraacc265" under reduced pressure to form VI. -4s yet tate is too complex to arise froiii a cyclobutane conditions have not been found which allow isola- structure such as V I I I C . ~ ~ tion of the pure trans,fmns,trans-dianhydride VIT, Further studies 011 the tetraesters IVa and 1':~i i i although its formation from the "all-trans'' acid eluded their reactions with phenyl Grignard reagent and acetyl chloride is indicated by infrared analysis. aiid with sodium ineta!. The inverse addition o l -1pplying the lead dioxide bis-decarboxylativc phenylinagnesiuin bronlidc to the 1,2,3,-Ltetr;ttc'chnique of Doering and co-workers" to the d - carbomethoxycyclobutaiie p~isscssing the "chair" mhydridc VI could formall;: lead to cyclobutadieiie ; stereochemistry afforded a compouiid wliich cis however ~ i to p the present tiuic cxperinieiits aloiig hibited both c:irboiiyl :tiid hvdrosyl l ~ a t ~ dins it5 these lines have proved unrewarding. Allthough iiiirared spectrum. 111coiitrast, the correspoiidi1ig iiiore thaii one mole of carbon dioxide was liber- Grigtiarcl reaction \rith the " a l l - f i n ~ s "isomer Va ated when the anhydride 1-1 was heated with gave :L compountl possessing iio c:trboiiyl BbSOrjlactive lead dioxide, no products could be collected tion in its spectrum. The latter product has in the low temperature traps aiid no precipitation been assigned the "all-trans" structure X on the occurred when the effluent nitrogen carrier gas was basis nf its iiifriircd atid (*ombustion a11al~sis.?' passed through :t concentrated silver nitratc 'I'hc dift'erciice in beha\-ior show1 by the two esters solution.'* The use of lead tetraacetate in pyridine as a iiiilder tlecarboxylative agent for the tetra11 (,'OII . l < . ( ' O I i basic acid IVb also provrd iiieffectual, probably I If('(1H; because of the low solubility of the polybasic ' IiiC'(.) acid in acetonitrile and benzene which were used rt,('o~ I t i(.xoH as solvents. l 9 ()

A

0

0

\

*_

(16) The doublet for the anhydride from ris,cis-3-meth~lcyclobutane-1,2-dicarboxylic acid is at 5.41 and B . 6 3 ~ . See H ..-'i Cripps, J. K. Williams, V. Tullio and W. H. Sharkey, J . A m . Chem. S O L . , 81, 4904 (1959). (17) 1%'. v. E. Doering. hl. Farher and A . Sayigh, ibid.. 74, 4370 (1952). (18) M . .%\-ram, E . lIarica a n d C. D. Nenitzescu, Bpi., 92, 1088 (1959). (19) C . A . Groli, & I . Olitd nnrl A . \i'ciss, .47igr.a. Cifer?r., 7 0 , 313

:l!,aa).

S:F I l ,

.

XI. I? -:('bkl,,

.. - .

(20) R.A. Raphael, .I. Cl!s>n.S O L ,401 (195.7). (21) W e are indebted t o R. A. Raphael for a sample of this material a n d t o Kugent Chamberlain of t h e Humble Oil and Refining C o . . Baytown, T e x . , for n.m.r. analvsis. ( 2 2 ) T h e tetraalcohol X h n s also been 1ire1txd from lr~l1!s,lralf.v, ii.a?z~-1,2,3,I~tetrabenzii?.lc?.clohuta,,eh y trenlinent rvith n h c n r l m?gnesium bromide. l'he cxperimental details nil1 be described i n d later paver.

DIMETHYL FUMARATE DIMER

June20, 1961

2727

is consistent with the observations made by iilber- absorption in the region T = 3-4 attests t o t h e absence of vinyl protons. T h e major peak in the mass spectrum of the man and Kipping on the reactions of the dimethyl dimef is at M / e 257 corresponding to loss of a n -0Me fragesters of cis- and trans-l,2-cyclobutanedicarboxylic ment from the molecule.29 acids with phenylmagnesium bromide.23 While The tetrabeazyl ester was prepared by transesterification the trans-diester reacted with four moles of Grig- and was purified by two recrystallizations from benzenehexane; m.p. 107.5-108.5". T h e n.m.r. spectrum is connard reagent to give the diol XI1 which was sub- sistent with the proposed structure (three peaks in the ratio sequently converted to the diolefin, the cis isomer, of 1: 2 : 5 ) . presumably because of hindrance, reacted with Anal. Calcd. for C36H3208: C, 72.96; H, 5.44. Found: only three moles affording the ketoalcohol X I . C, i3.22; H , 5.28. Hydrogenolysis of X to 1,2,3,4-tetrabenzhydryl- Isomerization of cis.trans,cis-1,2,3,4-Tetracarbomethoxycyclobutane (XIII) was achieved with copper cyclobutane (IVa) .-.I Pyrex tube containing 1 .O g. (0.004 chromite catalyst in an autoclave a t a pressure of mole) of the dimer IVa was sealed a t a pressure of 0.1 mm. 2700 p.s.i. and a temperature of 250". The sim- The tube was then heated in a muffle furnace a t 300" for 20 After cooling, the dark solid was recrystallized twice plicity of the n.m.r. spectrum of XI11 as well as the hours. from benzene ( S o r i t ) t o give V a i n 50y0yield, m . p . 123-125" splitting and the relative peak areas all attest to (reportedI3 12i'). the fact that the cyclobutane ring has been preStereochemical equilibration was also effected by heating served in the Giignard reaction and subsequent the dimer 0.28 E[. (0.001mole) with 0.3 E[ (0.006 mole) of sodium methoxiae in methanol for 2 hours. T h e resulting reduction. Experiments designed to convert X reaction mixture, after treatment with 10 ml. of a 10% amto the tetrakis-(diphenylmethylene)-cyclobutane monium chloride solution, was evaporated t o dryness. T h e (11)are now in progress. organic material was separated from the inorganic solids by '

(1'350).

sublimation at 80" (0.01mm.) and subsequently purified by recrystallization from water; m.p. 127' ( ~ 7 ~ ) . The infrared spectrum of this substance determined in potassium bromide shows absorption a t 3.36(w),5.77(s) and 8.25(s) p . The n.m.r. spectrum has bands a t T = 6.67 and T = 6.28in the proper ratio of 1 :3,respectively. Hydrolysis of cis , i ~ a n,cis-l,2,3,4-Tetracarbomethoxys cyclobutane (IVa).-Five grams (0.017mole) of the dimer IVa was heated on a steam-bath with concentrated hydrochloric acid until complete solution mas attained. T h e mixture was then evaporated t o dryness under reduced pressure. T h e resulting solid after recrystallization from acetone-hexane lost water a t 220-225'; reportedla m.p. 226" dec. Treatment of the acid with diazomethane in ether provided the original dimer demonstrating t h a t no isomerization had occurred during the hydrolysis stage. Hydrolysis of ti.ans,trans,tvans-l,2,3,4-Tetracarbomethoxycyclobutane (Va).--A mixture of 0.28 g. (0.001mole) of t h e dimer and 5 nil. of concentrated hydrochloric acid was heated on n steam-bath until a homogeneous solution was obtained. This solution was then evaporated t o dryness under reduced pressure. The residual solid was crystallized from either acetic acid-hexane or tetrahydrofuran-hexane solvent pairs; m.p. 261-261" dec., reported'3 261-264" dec. Sublimation of a sample of this tetrabasic acid Vb a t 265" (0.05 mm.) afforded a mixture of starting material and the "chair" dianhydride. Dehydration of cis,trans,ris-1,2,3,4-Tetracarboxycyclobutane (IVb).--.2 0.10-g. (0.0005 inole) sample of the tetracarboxycyclobutane was heated !\.it11 5.0 nil. of thionyl chloride for 3 hours. The excess tliioii!.l chloride was removed by distillation under reduced pressure and the residue after being washed wirh hexane was collected on a filter and finally purified by sublimation. The tetrabasic acid may also be dehydrated thermally by heating at 225-230" in a sublimator a t 0.05 mtn. X second sublimation is needed in order t o free the anhydride from acid which also sublimes. The infrared spectrum ex!ii!>its tlic ciiaracteristic anhydride doublet a t 5.38 and 5.GO u and is identical t o the spectrum of the product obtained from the pliotodimerization of maleic anhydritle in t h e solid state. Attempted Bis-decarboxylation of the Dianhydride VI ;ln intimate mixture of 0.30g. of the dianhydride, 2.50 g. of active lead oxide and 10 g. of powlcred pyreu glass was heated a t 250" for 0 . 2 5 hour while being stirred in a stream of nitrogen. The efRuent gases were passed in turn through silver nitrate and barium hydroxide traps. Although carbon dioxide was generated, there was no indication of a n y reaction occuring in the silver nitrate t r a p and repeated extraction of the residual material with solvents yielded no soluble organic products.

(26) Unless otherwise indicated all melting points are uncorrected. Analyses were performed by Schwarzkopf Microanalyytical Laboratories, Woodside 77, N. Y . We are indebted t o R. B. Hager and R. IYillcott for fi0mc. n.m r. spectra. (27) Maximum ultraviolet radiation (9570) a t 253.7 m p . (28) G. V. D . Tiers, J . P h y s . Chcm., 62, 1151 (1958); L. hl. Jack-

man, "Applications of h-uclear Magnetic Resonance Spectroscopy in Organic Chemistry," Pergamon Press, London, 1959, p. 50. (29) We are indebted t o A. H. Struck and Miss R Herberich, S t a m ford Research Laboratories, American Cyanamid Co., S t a m f o r d , Conn , for determination of the mass spectrum.

R2(IOH N I I . I1 = C ' J ,

R,CH XIII, R = C,;H,

IVhereas the "all-trans" ester Va reacted siiioothly with phenyl Grignard reagent, no isolable products were obtained when it w a s treated with s ~ d i u n i - o n - a l u m i n aor ~ ~ sodium dispersed in tolueneZ5under acyloin conditions.

Experimental26 Dimerization of Dimethyl Fumarate .-The solvent from a solution of 10 g. (0.07 mole) of dimethyl fumarate in acetone was evaporated under a nitrogen stream while the container, a 500-ml. graduate cylinder, was rotated in a nearly horizontal position. -4 Westinghouse 15TS Germicidal Sterilampz? v a s then inserted and irradiation was continued for a period of 1-5 days. During the course of the irradiation the temperature was maintained a t 25-30" by immersing the tube in a large water-bath. After the irradiation was completed the dimer was separated from unreacted dimethyl fumarate aiid polymer b y extraction with benzene, and finally recrgstallized from the same solvent (Norit); m.p. 144-145' (60 yo). A mixture melting point determination showed this product t o be identical t o a n authentic sample of the "chair" form of 1,2,3,4-tetracarbo1nethoxycyclobutane( IVa) obtained by Criegee and co-workers from a cinnamic acid dimer of this stereochemistry. .lnal. Calcd. for C12H1608: C, 50.00; 13, 5.60; mol. wt., 388. Found: C, 50.28; H, 5.55; mol. wt., 263 ( R a s t ) . The infrared spectrum of the compound, determined in potassium bromide, shows two bands characteristic of a carbonyl function a t 5.74(s) and 5.80(m) p . These bands merge when the spectrum is taken in chloroform. Other significant bauds in the spectrum are found a t 7.72(s),8.33 ( m ) , 9 . 7 9 ( a ) , 10.55(w), 11.85(w), and 12.21(w) p . The i1.m.r. spectrum determined in deuteriochloroforni shows only two types of protons in the espected ratio of 1: 3 LI-ith the less intense peak a t T = 6.15 and the other at 7 = 6.20. The position of the latter peak is consistent with t h a t of t h e -0CH3 protons of methyl acetate (7 = 6.35).**The lack of ( 2 3 ) K. B. Alberman and F. R. Kipping, J . Chem. Soc., 779 (1951). (24) W. v. E. Doering and P. LaFlamme, Tefvahedron, 2 , 75 (1958). (25) A . C. Cope and E. C. Herrick, J . A m . Chem. SOL, 72, 983

2728

G. W. GRIFFIN,A F. I

r

~AND K. ~

FURI:K.\Wt'rl ~ d

~

~

~

Vol. ~ s3

aqueoris ph;i,;e eu:ractcrl repeate.ii; 11 i ~ i A i - m i . pori.ioii.; < ) f Preparation of cis,trans,cis-1,2,3,4-Tetraacetoxy~~~eti~~~ibenzene. After the combined organic laycrs liad beea driecl cyclobutane (VIIIa) .-A 5.8-g. (0.02 mole) sample of the over Drierite, the volatile solverits were rcinoveti. Tile tetraester IVa dissolved in a mixture of 150 ml. of benzene residual solid was heated with enough hmzeiie t o effect soluand 150 i d . of dry ether was reduced in the inverse manner tion. \&'hen the tieiimie enchai-igeit \vi111 iiietii!.l hy addipg 3.0 g. (0.08 mole) of lithium aluminum hydride. ted, ivhicii when rccrys?a X superior procedure for accomplishing this consists of ex- hexane 0.5 g. nf crystals se from benzene melted a t 285-387' dec. tracting the lithium alurnitliiin hydride into the stirred ester trum of this compound 113s bands assigti solution from a Soxhlet thimb!e with the refluxing solvent. After the addition was completed 23 ml. of acetyl chloride in function (3.03 p ) and a:i aromatic cnrbor Condensation of t~~ns,trans,trai?s-l.'2,3,4-'~etracarbo60 mil. of ether was added dropwise and heating was contiiiucd for ail additional 13 hours. T h e reduction mixture was methoxycyclobutane (Va) with Phenylmagnesiuin Bromide. -The Grigriard reagent prepared in tetrahydrofuran from 1t i firially decomposed by adding 300 ml. of water followed by g. (0.70 g. atom) of magnesium turnings and 72 ~n!.of brolna100 ml. of 10yosulfuric acid. The organiclayer was removed benzene (0.70 inole) in 75 rul. of dry tetraliylrofurnri was :rnd the aqueous phase extracted three times with 50-m!. portioris of benzene. 'The combined organic layers were added dropvise t o a well stirred solutioil of thc tetrawter Y:i then tlried over Drierite and finally concentrated t o a dark in 250 ml. of tetrahydrofuran. After the addition -n:i i cntripleted, the solution was heated under reflux for 2' hot1is, arid oil (4.5 g.). T'acuum di5tilla:ion of this oil (178-180' then stirred a ? additional 12 hours :tt r n v n ieinperaturc. (0.02rnm.)) provided 2.0 g. of t h e polyacetate VIIIa which s l o ~ ~ vsolidified !y on standing. The infrared spectrum of this T h e tetraliydrolur,in !vas then dicplxed xvitli beiizew :Ind, C ( I I ! I ~ O ~ I Iineasured I~ i n chloroform has bands a t 5.80(s) and after being cooled, the reaction rniut solution of ainmoniu~nchloride ( T O g . X.O9(s) mp. T h e n.m.r. spectrum has peaks at T = 5.88, After stirring 2 - 4 hours the resulting T = 7.42 and 7 = 8.00 in the ratio of 2 : 1 :3 which is consistfuged and the so!iJ which appeared elit with t h e assigned structure. A n d . C;ilcd. for. CIeii2,Qa: C , 55.80; W, 7.03. Found: collected by decantiiig the solvents. T:iis so?id ivr!b drie,-I, transferred t o a Soxhlct thimble and estr;icted in tilrn w i t h c, 55.26; 11, 7.21.. pentane t o reinove hip1lcn::l; benzene, to wtracL hvtlrt)uyPreparation of ci.s.tuem,cis-1,2,3,4-Tetrabenzoyloxy- ketones; and fi:ia!ly with tetrnliydrofura~i. .Liter 3 to 5 niethyicyclobutane (VIIlc).--\ solution of 0.90 g. (0.0026 hours of estrsctioii, t!ie id 2.0 g. of potassium hyin ole^ nf the tetraacetate VI from the hoiliiig tetrah drciside i i i 60 nil. ot riiethnii heated under reflux for 2 extraction was co'iipietc liours. Tlic methano! WE removed and the residue when the eutract \vas d \vas hezicil for 3 iiours witii 50 nil. of benzoyl chloride. After above 300". Ti:e i: liydrcll! si5 of the excess beiizoyl chloride the reaction mix- 3.0!)(S), 3.1N