REDUCTIVE CLEAVAGE WITH METAL IN LIQUID AMMONIA I. T H E SELECTIVE CLEAVAGE OF THE BENZYLTHIO C-S BOND VERSUS THE BENZYLIDENEDIOXY C-0 BOND IN METHYL S-BENZYL-4,6-0-BENZYLIDENE-2-THIO-U-D-ALTROPYRANOSIDES; THE INFLUENCE OF T H E COSOLVENT 1,2-DIMETHOXYETHANE ON THE COURSE OF THE REACTION
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
U. G. T\T;\l':\1<' AND R. I<. BI
1.5 h ) ; under these conditions the benzylidene and benzyl groups are converted to a larger extent into bibenzyl, the rest becoming toluene. The two strong infrared absorption bands (in Nujol) in the region of 766 t o 578 cm-I and 706 to 718 cm-I have been assigned t o the phenyl moiety of the benzylidene group, and the one strong band in the region of 702 cm-I to the phenyl moiety of the S-benzyl group. INTRODUCTION
Birch (1, 2) has pointed out that a carbonyl group can be converted into a non-reducible species by acetal or lietal forination, but only if the resulting alltoxy group involved is neither benzylic nor allylic. Groups such as benzyl, benzyloxycarbonyl, and p-toluenesulfonyl, which are used to protect amino, hydroxy, or mercapto functions, are readily reilloved by the reducing action of sodium and liquid an~monia.Several articles have appeared (3-6) which illustrate the use of this technique in the synthesis of carbohydrate derivatives. A report fro111 this laboratory (3) has described the preparation of sonle monomercaptomonosaccharides by reductive cleavage of S-benzyl-4,6-0-benzylidene-2(or -3-) thioglycopyranosides with sodium and liquid amn~onia.At that time sonle attempts were made, by use of a limited proportion of sodium, to effect a selective cleavage of the S-benzyl group wit11 retention of the protecting 4,6-0-benzylidene moiety (111, Reaction Scheme I ) . However, the only isolable compounds were the unchanged methyl S-benzyl-4,G-Obenzylidene-2- (or -3-) thioglycosidc (I) and the glycoside minus both protecting groups (11, Reaction Scheme 1). T o our Itnowledge few accounts exist wherein attempts a t such selective metalanlmonia hydrogenolysis have been reported. In this coilnection it is note\vorthy that a large excess of Raney nickel removes both the benzylidene and the methylthio groups from illethyl 4,6-O-benzylidene-S-i~~etl~yl-3-thio-a-~-altropyranoside (7) whereas, with only a fourfold excess of the metal, reductive desulfurization talies place unaccoillpanied by extensive removal of the benzylidene group (8). Pinder and Smith (9) found no selective cleavage in the sodium -liquid ammonia reduction of 2-methyl-2-phenyl-1,3-oxathiolane or of 2-i~~ethyl-2-phen~~l-1,3-dioxolane. U~i-dersityof Alberta Postdoctorate Fellow. Canadian Journal of Chemistry. Volume 44 (1966)
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
CANADIAN JOURNAL O F
CI-IEMISTRY.VOL.
44. l9GG
H
m
OH
OCH,
OH
JI
0CH3
I t has been shown, however, that dialkyl thioethers arc readily cleaved by sodi~1111 in liquid an~moniawhereas the a~lalogousoxygen ethers are practically unaffected (10). Thus one would expect, in a competitive reaction, that a benzylthio group should be hydrogenolyzed in preference t o a benzyloxy group. This paper describes the results of a t t e ~ n p t sa t selective cleavage of the benzylthio li~lltagewith sodium and liquid ammonia i11 the presence of the benzylidene dioxy moiety, examples from the field of carbohydrate chemistry being used as substrates. RESULTS AND DISCUSSION
Amoz~ntof Sodizlm Required for Cleavage of the C-S and C-0 Bonds Experilne~ltswere designed to measure the proportion of sodium required to cleave the C-S and C-0 bonds in a molecule such as methyl S-benzyl-4,6-0-benzylidene-2-thio-aD-altropyranoside (I). This infor~natio~l was deemed to be llecessary in approaching the problem of selective bond cleavage. The theoretically required amounts of metal are illustrated in Reaction Scheme 2. I t is seen that C-S cleavage requires 2 g-atorns of sodiu~nif toluene alone is produced but just 1 g-atom of sodiunl if bibenzyl is formed exclusively. \Vhether bibenzyl is made by freeradical coupling as shown in Reaction Scheme 2 or by an anionic process as suggested by Birch (1) is not yet clear, but by either process 1 g-atom of sodiunl is required for C-S cleavage of 1 mole of colnpound I to produce the half ~ n o l eof bibenzyl. For C-0 cleavage, 2 g-atoms of sodiunl are expected t o cleave one leg of the acetal (either a t Cd-0 or C6-0) to form the 0-benzpl group, after proton abstraction from the solvent or any alcoholic group present. This is follo~vedby consu~nptionof either 1 or 2 g-atoms of sodium to produce bibenzyl or toluene respectively. Thus, 3 or 4 g-atoms of sodium are required per ~noleof I depending upon the fate of the benzylidene group (as bibellzyl or toluene). These predictions were substantiated by the results of experinlents which measured the proporti011 of sodium metal required for the reductive cleavage of lnethyl 4,6-0-benzylidene-a-D-glu~op~ranoside (IV), methyl S-benzyl-4,6-O-ethylidene-3-thio-cr-~-altropyra1loside (Ira), and methyl S-benzyl-4,6-O-ethylidene-2-0-metl~yl-3-thio-a-~-altropyranoside (Vb) (Table I , experiments 1-5). The ethylidene group in cornpounds V a and
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
N A Y A K A N D BROWN: REDUCTIVE CLEAVAGE. I
C - 0 cleavage 2Na
v
7
C-S
cleavage 1Na
OH
H @ ~ N Hor~ ROH
( CH ,, 0CH3
CH2 I2
C6H5CH2 :
OCH,
'
'bH5 CH3
0 [email protected] OH
OCH,
'bH5
CH2:
NH30r C6H5CH3 ROH
Vb is unaffected by the metal, and hence permitted estimation of the sodium consumption in the cleavage of the benzyl group from sulfur. In liquid arnmonia, the 4,G-0benzylidene group required 4 equivalents of sodium2 for cleavage, yielding essentially toluene along with a minor amount (3%) of bibenzyl .(expt. I). But in the nixed solvent NH3-1,2-dimethoxyethane somewhat less than 4 equivalents of sodium were consumed and a larger amount of bibenzyl was for~ned(16%, expt. 2). Accordingly, as expected, methyl S-benzyl-4,G-O-benzylidene-2-thio-a-~-altropyranoside (I) in liquid anlrnonia required G equivalents of sodium for cotnplete reduction to methyl 2-thio-a-D-altropyranoside (11) (Table I , expt. 6). Selective Cleavage of the C-S Rather Than the C-0 Bond When 2 equivalents of sodium was added to 1 equivalent of methyl S-benzyl-4,G-0benzylidene-2-thio-a-n-altropyranoside (I) in liquid ammonia, only methyl 2-thio-a-Daltropyranoside (11), bibenzyl, and much unchanged I were obtained. This lack of selective reduction was due to the sparing solubility of I in liquid a l n n ~ o n i aWhen .~ 1,2-di~nethoxy2The hydroxyl groups i n I V and V a cool~sztmedn o inetal under oz~rconditions. SA separate test sl~owedthat the bulk of a 100 mg quantity of I failed to dissolve in 500 vz1 of liquid alnmonin
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
CANADIAN JOURNAL O F CIIEMISTRY. VOL. 14, l 9 G G
a n.rn n.n. ir, n.
w
m
m
A A
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
NAYAK A N D BROWK: REDUCTIVE CLEAVAGE. I
595
ethane was added as a cosolvent, to improve the solubility of I in liquid ammonia (4, 11, 13), selective reductive cleavage of the C-S bond did occur (Table I , expt. 7). Optimum results \\.ere obtained \\hen the ratio of 1,2-dimethoxyethane to NE13 was 4 or 5 to 6 (by volume). Several interesting features became evident when 1,2-dimethoxj~ethanewas used as cosolvent. (i) For the quantities of reagents taken (see Table 11), the time required for consumption of sodium increased from 15 to 20 min when only ammonia was used, to lh-3% h in the mixed solvent. A larger ratio of 1,2-diinethoxyethane to NI-I3 increased the time to several hours. (ii) The proportion of bibenzyl increased 1narl;edly (>50yo of theory) over that (-9%) obtained in just liquid alnmonia. The yield of toluene dropped sharpll. as a consequence. (iii) The proportion of sodium metal required for C-S cleavage decreased also. This is in keeping with the requireinent of 1 g-atom of metal when bibenzyl is the product of C-S cleavage, and 2 g-atoms when only toluene is obtained (see Reaction Scheme 2). The influence of solvent con~positionand ainount of metal on the proportion of products obtained is shown in Table 11. I t is note\vorthy that the use of lithium instead of sodium, under otherwise identical conditions, gave a larger yield (76%) of the selectively cleaved product 111. Table I11 shows the effects, on the reduction of Va, of solvent composition and proportion of metal to substrate, with particular attention to bibenzyl (and therefore toluene) production. Assigjznzent of Absorption Bands for the Benzylthio and Benzylidene Dioxy G r o ~ ~ p s The selective re~novalof the benzyl group froin I has permitted the assignment of infrared absorption bands caused by the two non no substituted phenyl groups. Strong absorption for monosubstituted phenpl in coinpound I in Nujol is found a t 770 cm-l, 710 cin-I, and 702 cm-l. For inethyl 4,6-0-benzylidene-2-thio-a-D-altropyrmoside (I1 I), absorption (in Kujol) occurred a t 766 cm-I and 706 c~n-I caused by the benzylideile moietl-. The completely reduced substance, methyl 2-thio-a-D-altropyranoside (11), showed none of these three absorption bands. Accordingly, for these altrosides, the 4,Ci-0-benzj lidene group absorbs in the region of 766-770 cm-I and 706-710 cm-I (Acm-I = -60) \\hereas the 2-benzylthio group absorbs a t 702 cnl-I. In agreement with this, methyl 4,6-O-benzylidene-2-0-111ethyl-a-~-altropyranoside (14) sho~vsabsorption bands a t 778 cm-I and 718 cm-I (Acm-I = GO), methyl 4,6-0-benz~~lidene-2,3-di-O-i11ethyl-cr-~altropyranoside (14) exhibits strong absorption a t both 766 cm-I and 706 cm-I (Acm-I = GO), and methyl 4,6-O-benzylidene-2-deox~~-a-~-allopyraioside absorbs a t 778 cm-' and 718 cm-I (Acin-I = 60) (Fig. 1). EXPERIMENTAL hlelting p o i ~ ~are t s i~ncorrected.Infrared spectra were taken with a Perkin-Elmer nod el 4'21 instrument. Nuclear magnetic resonance spectra were obtained with a Varian -Associates A-60 instrument. Gas-liquid chromatography \vas acco~nplishedwith a Burrell I<-2 ICromo-Tog by using a 2.5 ni colunln packed with 257; silicone rubber on Gas Chrom P (60-80 mesh) and a helium flow rate of 100 1n1/1ni11. Sy?ztheses of Corrrpoz~nds dfethyl ~,6-O-Etl~ylidene-2-O-(p-toluenesz~lfonyl)-oi-~-glz~copymnoside The published procedure (12, 15) was e~nployed,b u t with the follo\ving modifications which i~nprovedthe isolation and more than tripled the yield. (a) The methyl 4,6-0-ethylidene-a-D-glucopyranoside(12) was ,treated with a 1595 rather than 487, excess of p-toluenesulfonyl chloride. ( B ) The chloroform extract of the reaction mixture was first washed with ice-cold 6 IV I-ICI to remove pyridine, and then washed with bicarbonate solution and water. (c) The oil left upon removal of the chloroform from the decolorized solution was dissolved in ether. U;hen this was left in a refrigerator overnight a solid was obtained. \r hot absolute ethanol solution of this solid, when cooled, gave 1l.'iYOof pure methyl 4,6-0-eth ylidene-2,3-di-O-(p-toluenesulfo11yl)a-D-glucop)r:unoside, m.p. 156' (lit. 1n.p. 156-157' (12), 154-155' (16)). (d) ?'he brownish solid left after re~novalof solvent from theabove nlcoholic filtrate, when dissolved in chloroform and diluted with petrole~un
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
CA4S.\I1I.1S J O L K S A L 01; CIIE?JISTKY. VOL. 44. 10GG
--2O
1
i I
:.2 - I *z
wtn rw lmwcid *rnL?rb & cod mm
.-a-
>yj
-n
O g
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
NAYAIC A X D BROWN: REDUCTIVE CLEXVrZGE. I
CANADIAN JOUKNAL O F CHEMISTRY. VOL. 44. 19liO
OH
OCH,
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
u
111
c6~5-\~&
11
1 '1
H
li
H Orl
OCH,
m
OH
OCH,
Infrared absorption in Nujol of a number of altropyranosides in the region of 700 to '780 an-'. ether (GO-SOo) to faint turbidity, gave 42.37, of niethyl 4,G-O-ethylidene-2-0-(p-toluenesulfon~-l)-~-~glucopyranoside, m.p. 151' (lit. m.p. 150-151' (12)). (e! T h e mother liquors from the precipitated mono- and di-p-toluenesulfonyl co~npounds,freed from solvent and the residue acetylated (12), gave 11% of methyl 3-0-acetyl-4,6-0-ethylidene-2-0-(p-toluenesulfonyl)-a-~-g~ucopyralioside a s long needles, m.p. 171" (lit. 1n.p. 169-171" (13)). illethyl 2,3-Anhydro-/t,6-O-ethylidetze-a-~-n2atztzopyranoside T h e following modifications of the published procedure (12) greatly simplified the preparation of pure product. Methyl 4,6-0-ethylide1~e-~-(p-to~uenesulfony~)-a-~-glucopyranoside (or its acetate) \\.as treated \vith a 26% rather than 10070 excess of sodium methoside in relluxing methanol for 4 h. T h e cooled mixture (needles separated) was poured into ice water. T h e water mixture was extracted with chloroform and the extract dried (KazSOn) and freed fro111 s o l v e ~ ~'The t . residual solid was crystallized from methanol to give needles (84%) melting a t 100' (lit. m.p. 100-100.5" (12)). il[ctl~ylS-Benzyl-~,6-O-ethylidene-S-tl~io-~-~-atroyranosdc (Va) Published directions (12) \vere followed for most of this preparation. However, the isolation and crystallization procedures were modified advantageously. T h e precipitate obtained \\,hen the reaction mixture \\.as poured into ice water was washed with water and petroleum ether a s described ( l a ) , dissolved in diethyl ether, and dried (NazS0.1).T h e filtered ethereal solution mas diluted with a n equal volume of petroleum ether (60-80") and the resulting solution gently refluxed on a water bath. When much of the ether had been
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
S A Y t \ I i ASD BROWN: REDUCTIVE CLICAVAGE. I
5'39
removed and the boiling point of the solvent had reached SOo, the flask and contents were allowed to cool t o room temperature, whereupon Va separated a s long colorless needles, yield 93%, m.p. 133" (lit. m.p. 133-133" (12)). Metlzyl S-Benzyl-Q,6-0-etlzylide?ze-d-O-nretlzyl-3-thio-a-~-aLtropyranoside (Vb) T o a solution of 16.3 g (0.05 mole) of Va in 90 ml of methyl iodide was added 24 g of silver oxide. T h e mixture, first stirred for 1.5 h a t room temperature and then for 1.5 h a t reflux, \I1as filtered (cold) and the silver salts were washed with 50 rnl of dry ether. T h e combined methyl iodide and ether solutions were freed from solvent under vacuum. T h e viscous oily residue showed no absorption for OH in the infrared spectrum. Chromatography on neutral alumina, activity grade I (Hlupharm Chemicals, Xew Orleans, Louisiana), gave, by benzene elution, 11.1 g (65.3Y0) of a n oil which solidified when cooled. Crystallization from cold hexane ! (c, 0.5 in chloroform). T h e infrared spectrum in Kujol gave a compound melting a t 84-85O; [ a ] ~ ?-91" showed bands a t 710 cm-I (phenyl) but none for OH. Anal. Calcd. for CITH?IOSS: C, 59.97; H , 7.11; S , 9.42. Found: C , 59.84; H , 6.88; S, 9.68. Ether elution of the colunln gave 2.5 g (2.5y0) of a compound melting a t 98-99' (from methanol). The infrared spectrum in Nujol was identical with t h a t of authentic methyl 2,3-anhydro-4,6-0-ethylidene-a-Drnannopyranoside (ref. 12; see also above). No absorption for OH was evident. A mixed melting point was undepressed. iV!etlzyl S-Benzyl-~,6-O-benzylidene-d-tlz~o-a-~-atroyranoside (I) This compound was prepared by a n improved variation of the literature directions ( 5 ) .Commercial sodillin ~nethoxide(Fisher Scientific Co.) (21.9 g, 0.115 mole) was dissolved in 200 ml of pure, dry methanol. 2. continuous current of nitrogen was passed through the apparatus, and benzyl mercaptan (67.9 g, 0.58 mole) was then added. When the solution had been stirred for 15 min, 26.4 g (0.1 mole) of methyl 2,s-anhydro-4,60-benzylidene-a-D-allop)~ranoside (14, 17) was added and the solution refluxed for 24 h (N?).When cooled and poured with stirrir~ginto 1 1 of ice water, the solution gave a precipitate of I. T h e solid w ~ separated ~ s by filtration and dissolved in chloroform; this was combined with chloroform extracts (2 X 250 ml) of the filtrate. After the organic solution had been washed with water, it was dried (Na?SOI), separated from the drying agent, and freed from solvent under vacuum in a rotary evaporator. The sen~isolidresidue was triturated with 500 ml of petroleun~ether (60-SO"), whereupon solidification occurred, yield (crude) 29 g (-10070). Recrystallization fro111250 ml of absolute ethanol gave needles (35 g, 9070) melting a t 136-137'; [ffjDZS $94.7' (c, 1 in chloroform) (lit. m.p. 136-137"; [ a ] ~+90.G0 ~' (c, 1.668 in chloroform) (3)). T h e infrared spectrum in Nujol showed absorption a t 3 505 cm-I (OH), 770 cni-I and 710 cm-I (phenyl of the benzylidene group), and 702 cm-I (phenyl of the benzyl group). Sodiz~?n -Liquid A~rcnloniaReduction Methyl /t,6-~-Etlzy~ide?ze-d-0-?netIzy~-S-tIzio-a-~-a~tropyrnnoside (VIIb) frovz illretlzyl S-Benzyl-4,fi-0-ethylide~ze-2-0-nzetlzyl-3-t~~io-a-~-altropy,zosidc (Vb) (a Typical I'rocedzlre for Esti??cationof tlze Consz~nzptior~ of Sodium by Reaction in Liquid A?iznzo?~ia) A dry, 500 nil, three-necked, round-bottomed flasli (A) was htted with a gas irllet tube (C), mercury seal stirrer, and dry ice condenser (B) whose outlet accommodated a tube containing sodium hydroxide pellets a s protection against moisture from the atmosphere. The apparatus (:I and B) was gently heated with a free flame for 5 min while a current of dry nitrogen was passed through via C to drive out surface moisture. When the system had been cooled (~lndera constant flow of dry nitrogen) gaseous alninorlia fro111a cylinder of commercial anhydrous liquid ammonia was collected (150 ml) in a dry ice - acetone cooled receiver (D) fitted with a dry ice - acetone condenser (E) whose outlet was protected by a sodium hydroxide pellet tube. Metallic sodium was added t o the liquid in D t o destroy moisture, and then the ammonia was distilled, b y using a cold water bath gradually xvarmed t o 50°, into the dried vessel (A) previously cooled in a d r y ice - acetone bath. A few milligrams of metallic sodium were added a t intervals t o A until the blue color persisted (for about 1 h). T o this flask was then added 3.4 g (0.01 mole) of Vb. A clear solution was obtained. Freshly cut sodium (0.46 g) was weighed in dry toluene, and from this four sn7all pieces of about 100 mg each were added to the reaction mixture in -4, one a t a time only when the blue color caused by each preceding piece of metal had disappeared. Then the size of each s u b s e q ~ l e piece ~ ~ t of metal was reduced to -10 mg and these were added until the blue color persisted for a t least 15 mill. In all, 440 mg was necessary, over a total reactiori time of about 20 min. T h e reactiori mixture was then decomposed with excess ammonium chloride (5 g) and the ammonia allolved t o evaporate in a current of nitrogen. The residue was dissolved in chloroform, and filtered free of , solid; then the solvent was ren~ovedunder vacuum. 1 he remaining solid, when crystaliized fro111cold ether, gave 2.35 g (98%) of VIIb, n1.p. 103-104", [ a ] ~ ~ + l 1 3 . 7 (c, ~ 1 in chloroform). Absorption in the infrared (Nujol) occurred a t 2 600 cn1-I (SH) but none between 700 and SO0 cm-I for phenyl. Anal. Calcd. for CloI-Il8O5S:C, 47.98; I-I, 7.35; S , 12.51. F O L I I IC~ ,: 45.13; I-I, 6.96; S, 12.72. The residual ethereal filtrate was evaporated and the residue triturated with 20 in1 of petroleum ether (60-80") to dissolve bibcnzyl. 'The petroleu~nether extract was freed from solvent t o give SO mg (8.3%) of bibenzyl, m.p. 51" (mixed melting point undepressed).
.
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
600
CAS.4DIAN JOURNAL O F CHEMISTRY. VOL. 4-1. 1966
iUlethyl ~,6-0-Ethylide?ze-3-t1~io-a-~-a~tropyranoside ( VIIa,) froin ilCethyl S-Be~~zyl-~,6-0-etlzylide?~e-3-tl~ioa-D-altropyranoside ( Va) illethod 11: with sodi~r?n and liqztid antn~onia.-The reduction of this compound was carried out b y using the apparatus and typical procedure described above. T h e 3.26 g (0.01 mole) of Va required 440 mg (0.0'2 mole) of sodiu~nover a 15 to 20 min period. The reaction mixture was worked up accordingly t o the preceding description to give a solid, after removal of the chloroform solvent, which, when dissolved in ether and diluted with petroleunl ether (60-80") t o slight turbidity, gave 2.26 g (95.5%) of VIIa a s needles melting a t 106-107"; [ a ] ~ ?-123" .' (c, 1 in chloroform). Anal. Calcd. for C9HIGOIS:C, 45.75; H , 6.83; S , 13.57. Found: C , 45.83; M, 6.91; S, 13.72. Infrared absorption occurred a t 3 465 cm-I (01-1) and 2 600 cm-I (SH). There was no absorption for phenyl. The residual ethereal filtrate, worked u p a s before, gave 80 mg of bibenzyl. ~l~etlrod B:with sodium aitd liquid a~n?izoniaand ?~zodz$cutionto isolate the prodz~cttoluene.-The preparation of the system and the reduction of the thioether Va (3.26 g, 0.01 mole) were carried out exactly a s in t h e typical procedure above, except t h a t the sodiu~nwas weighed and cut up in Sliellysolve B. T h e work-up differed as follows. After the decomposition of the reaction mixture with ammonium chloride, the condenser (B) was filled with a n ice-salt freezing mixture to retain the toluene b u t permit vaporization of the ammonia. T h e stream of nitrogen that was passed through the solution was discontinued, since it might promote loss of toluene by entrainment. T h e ammonia was allo\i~edto volatilize and escape through B, without t h e use of a ~ v a r ~ n ewater d bath, by allowing the apparatus t o remain overnight exposed to room temperature conditions. The residue in A was stirred with dry ether (200 ml) for half an hour and the solution then filtered. Careful fractionation renloved the ether (which showed 110 peak for toluene by gas-liquid chromatography) until the residue was concentrated to about 15 ml. T o this residue was added 0.499 g of ethylbenzene a s internal standard; the resulting solution was analyzed b y gas-liquid chromatography. An authentic mixture of lcnown amounts of toluene and ethylbenzene was used for comparison. H total of 540 mg (58%) of toluene was thus determined. Removal of the ethylbenzene, toluene, and ether by distillation i n vaczro gave a residue which was washed with water to eliminate the water-soluble VIIa. T h e solid remaining (100 mg, l l y o ) was bibenzyl, m.p. 51". ilJethod C: with sodi?rn~and liquid anzmonia diluted witlr 1,2-diiizetl~oxyetlta?te.-A quantity of dry ammonia (250 ml) was put into a 1 I, three-necked, round-bottomed flaslc according to the typical procedure above. T o this was added, via a dried dropping funnel, 100 ml of d r y 1,2-dimethoxyethane (distilled from potassium metal) followed by a solution of 3.26 g (0.01 mole) of Va in 100 ml of dry 1,2-dimethoxyethane. An additional 50 ml of 1,2-dirnethosyethane was used to rinse the funnel, thusgiving a 1:1 mixture (by volume) of ammonia and 1,2-dimethoxyethane in the reaction flasli. T o this solution was added freshly cut sodiu~n(230 ~ n g0.01 , mole) in four pieces, one a t a time, according t o the typical procedure. The time required for this sodium t o react was 45 mill. The mixture was then decomposed with arnrnoniu~nchloride a s usual and the a ~ n m o n i a allowed to escape overnight, without the use of a warmed water bath, b y using a constant current of d r y nitrogen gas. About 150 ml of chloroform was then added and the mixture stirred for 1 h. This process eliminated the rest of the anlnlonia (PI-1 paper test). T h e material was filtered, and chloroform (100 nil) was used to rinse the flasli and to wash the collected solid. Removal of the chloroforn~and 1,2-dirnethoxyethane on a rotary evaporator under vacuum left a semisolid. Trituration with water dissolved the mercaptan, and unreacted Va and bibenzyl remained as a solid. The mixture was filtered, and the aqueous filtrate concentrated i n vacuo uutil the residue solidified. T h e crude yield was 1.59 g (67.3y0). Crystallization fro111a mixture of ether and petroleum ether (60-80") gave 1.55 g (66%) of material melting a t 106-107°. T h e infrared spectruni and mixed melting point showed this to be identical with VIIa. Anal. Calcd. for C O H I G O ~CS, :45.75; M, 6.83; S, 13.57. Found: C, 46.00; M, 6.87; S, 13.66. The mixture of uilreacted material and bibenzyl left after removal of t h e aqueous solution was heated in 50 rnl of refluxing petroleunl ether (60-SOo). The cooled mixture was filtered, leaving 1.07 g (32.Gy0) of solid melting a t 132-133Oand identical (infrared, mixed melting point) with starting material (Va). The petrole~un ether filtrate was freed from solvent, leaving 300 mg (327,) of solid bibenzyl melting a t 51'. When the solvent mas co~nposedof 250 ml of ammonia and 200 rnl of 1,2-dimethoxyethane, and 3.26 (0.01 mole) of Ira and 0.31 g (0.0134 mole) of sodium were used, an 86% yield of VIIa, 0.451 g (14y0) of unchanged \k,and 400 mg (42y0) of b i b e ~ ~ z ywere l obtained. Table 111 illustrates the influence of solvent composition and amount of metal 011 the proportion of products in the ammonia-l,2-dimethoxyethane-sodium reduction of Va to VIIa.
AJethyl4,6-0-Benzylidene-2-thio-a-~-o~ltro~ra~~oside (111) from iUletltyl S-Bepzzyl-4,6-0-benzylide?te-2-tlzio-a:D-altropyranoside (I) Metlzod A : with sodirrnz and liquid ali~?izonia.-Details of the procedure followed those in method A for the reduction of the 3-benzylthioaltroside Va described above. The 19.4 g (0.05 mole) of I in 600 ml of ammonia consumed a total of 6.7 g (0.29 mole) of sodium in 200 mg pieces over a period of 1 5 min. T h e solution, after decomposition with excess ammonium chloride, was extracted with hot chloroform (4 X 250 ml). T h e solid left after removal of the chloroform from these extracts was triturated with 100 ml of petroleurn ether (60-80") t o remove a n y bibenzyl, and the residue was dissolved in 50 ml of hot ethanol. When cooled overnight in a refrigerator, the solution yielded 11 a s long colorless needles (8.4 g). A second crop (I .37 g) was
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
NAYAK AND BROWN: REDUCTIVE CLEAVAGE. I
601
obtained upon concentration of the mother liquor. The total yield was 9.87 g (94%) of material melting a t 146-147" (lit. m.p. 145-146" (3), 145-148" (4)). ?'he infrared spectrum in Nujol showed absorption a t 3 415 cm-1 (OH) and 2 570 cm-1 (SH) b u t no phenyl absorption. Method B: with sodiziv~and liquid a ~ i ~ ? r ~ oand n i a?izodi$fications to isolate tolz~e?zeand IriIrcnsy1.-'l'hc procedure used in method B for the reduction of the :3-benz>~lthioaItrosideVa was followed here. total of 1.3-1g (0.058 mole) of sodium was required by 3.88 g (0.01 mole) of compound I in 150 ml of amn~onia..After decomposition of the reaction mixture with ammonium chloride, 250 ml of dry ether was added to the flas1.r and the ammonia allowed to evaporate through the condenser B cooled by a salt-ice freezing mixture. T h e residual ethereal mixture was filtered and the solid residue extracted with 200 ml of hot chloroform. Ren~oval of the chlorofor~nfrom this extract and crystallization of the resulting material from ethanol gave I1 as colorless needles (1.822 g, 87%), m.p. 146-147". The ethereal extract was reduced in volume and ethylbenzene added as internal standard according to the details of method B for compound Va; 1.02 g (-56%) of toluene and 0.163 g (-9%) of solid bibenzyl, n1.p. 51-52", were obtained. iMethod C: with sodium and liquid a?it?rtonia dilz~tedwith 1,2-di?itetho~yetltane.-~I'I1e apparatus and the procedure used for the preparation of the ammoi1ia-1,2-dimethoxyethane solution and subsequent reductioi~ were the same as those in method C described for the reduction of Va. T o a mixture of 300 ml of ammonia and 250 ml of 1,2-dimethoxyetha~lecontaining 7.76 g (0.02 mole) of I was added 0.69 g (0.03 mole) of sodium cut into seven pieces, each piece being added in accordance with previous directions. The time required for total addition was 2.5 h. The decomposition and solvent elimination also followed cited directions. The solid t h a t was obtained was dissolved in 50 n ~ofl hot ethanol (rather than being triturated with water), and the solution was cooled to room temperature to give 2.3 g of the selectively reduced compound 111 as needles, m.p. 167-169". Theethanol was removed from the filtrate with a rotary evaporator under vacuum, and the residue was dissolved in benzene and subjected to chromatography on 100 g of neutral alumina, activity grade 11 (Alupharm Chemicals, New Orleans, Louisiana). Initial benzene elution gave bibenzyl and unreacted I. With continued elution by benzene the mercaptan 111 began to appear, as shown by infrared absorption a t 2 500 cm-I, whereupon ether replaced benzene as the eluant to hasten the exit of the rest of component 111 still on the column (1.9 g). The total yield of 111 was 4.2 g (70.5%). Crystallization of the combined material from ethanol raised the melting point slightly ~ . ~ (c, 1 in chloroform). to 168-169'; [ a ] ~+S0.5" Anal. Calcd. for C14H180SS:C, 56.37; El, 6.08; S , 10.74. Found: C , 56.23; M,6.04; S, 10.76. The infrared spectrum in Nujol showed absorption a t 3 483 cin-I (OH), 2 540 cm-I (SH), and 766 cm-1 and 706 cm-1 (monosubstituted phenyl for the benzylidene group), but none a t 708 cni-I (monosubstituted phenyl for benzyl group). The integrated nuclear magnetic resonance spectrum in CDCIJ, with tetramethylsilane as reference, agreed with the structure assigned. Signals appeared a t T = 2.35 to 2.75 ( m ~ ~ l t i p l for et five phenyl protons), 7 = 6.58 (si~lgletfor three methyl protons), and T = 7.95 and 8.15 (J = 10 c.p.5.) (doublet for SI-I proton). The combined benzene eluates, freed from solvent, gave a solid which was heated with 50 n ~ of l refluxing petroleum ether (60-SOo). Separation by filtration left 1.61 g (21y0) of unreacted I (undepressed melting point and identical infrared spectrum). From the petroleum ether filtrate was obtained, upon solvent removal, 0.87 g (48%) of bibenzyl, m.p. 51". Final methanol eIution of the column above gave 0.47 g of unidentified oil which showed absorption for SH (2 540 cm-I). Table I1 illustrates the effect on product yield of different proportions of mixed solvent and starting materials with compound I as the reducible substaiice. Selective Reduction of I with Litltizlnt in Liquid Antrrronia - 1,2-Di?~retlzosyetlta~ze A solution of compound I in ammonia-] ,2-dimethoxyethane, u ~ i t hthe same quantities of materials as in the preceding experiment for method C , was treated with 0.208 g (0.03 mole) of lithium according to the detailed directions in method C. The solid obtained upon decomposition of the reaction mixture and removal of the 1,2-dimethoxyethane and chloroform used for extraction was heated with benzene to perinit separation of the lithium chloride by filtration. The benzene was removed from the filtrate and the solid residue was dissolved in 100 ml of hot ethanol. When cooled, the solution deposited 2.7 g of I11 as needles, 1n.p. 168-169". Chromatography of the residue as described above (method C) gave a further 1.81 g of 111. The total yield of the lnercaptan was 4.51 g (75.7%). Unreduced starting material amounted to 1.248 g (16%), and bibenzyl to 1.09 g (59%). Here again the same unidentified oil (420 mg) was obtained via the final methanol elution. Redzlction of Methyl 4,6-0-Benzylide?te-or-D-glucopy,.anoside (IV) Method A: with sodiunt i?z liqz~ida?,z?lzo?zia.-Details of the procedure were the same as those in method B for the reduction of I to obtain the products 11, toluene, and bibenzyl. Methyl 4,6-0-benzylidene-ol-nglucopyranoside (11) (5.64 g, 0.02 mole) in 300 ml of ammonia required 1.785 g (0.0776 mole) of sodium over a period of about 15 min. The ethereal extract obtained after removal of the solid by filtration gave, b y gas-liquid chromatographic analysis, 1.14 g (62%) of toluene and 0.053 g (-3%) of bibe~lzyl. Methanol extraction of the solid obtained after separation of the ether filtrate gave 3.08 g (79%) of methyl or-D-glucopyranoside melting a t 164-165° (lit. m.p. 165" (11)). Method B: with sodium and liquid a?nnzonia containiltg 1,2-dimethoxyetkane.-Details for this experiment
follo\ved those in method C for the reduction of I to obtain 111. A solution of 6.G4 g (0.02 mole) of IV i r ~ 300 ml of ammonia diluted with 200 ml of l,9-dimethosyethane consumed 1.770 g (0.0756 mole) of sodium over a period of 4 11. After evaporation of the ammonia, however, the residue \\.as extracted with ether t o remove bibenzyl:l 'l'his amounted to 0.30 g (16%). ?'he glycoside that \\.as produced \\.as not isolated in this case and was believed to be only methyl a-D-glucopyra~losidc.\Ye were collcerned in this esperilncnt only \\:ith the a r n o ~ ~ noft bibellzyl formecl ill a n ammoni;1-1,2-dimctl~osyetha11emisturc compnrecl with the amouilt produced in ammonia a l o ~ ~ under e otherwise identical conditior~s.
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 46.3.205.140 on 12/31/17 For personal use only.
ACICSOWLEDGMENTS
T h e financial assistance given by the Natioilal Research C o ~ ~ n cof i l Canada for part of this work is greatly appreciated. Infrared and nuclear magnetic resonance spectra were inade bl- JIrs. G. Conu-ay of the Departineilt of Chemistry, University of Alberta, Edmonton, Alberta. Elemental analyses were performed by Dr. C. Daessle, Organic lIicroanalyses, 5757 Decelles Avenue, Alontreal, Quebec, and by Dr. F. Pascher, Mikroanalytisches Laboratoriu~n,Buschstrasse, Bonn, Germany. REFEliENCES 1. 2. 3. 4.
5, 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 1G. 17.
A. J. BIRCH.Quart. Rev. London, 4, 69 (1950). A. J. BIRCH and H. SIIITII. Quart. Rev. London, 12, 17 (1958). Y. C. JAMIESON and I<. I<. UIIOXVK. Can. J. Chem. 39, 1765 (1961). J. E. CHRISTEXSET and L. GOODA~AX. J. Am. Chern. Soc. 82, 3827 (1961). . 86, 1427 (1964). G. CASINI and L. Goo~ar~\x.J. Am. C h e n ~ Soc. E. J. R ~ r s rV. , J. BARTUSKA, and L. Goonlr.4~. J. Org. Chem. 29, 3725 (1964). H. I<. BOLLIGER and 13. rZ. PRIES. Helv. Chim. Acta, 29, 1061 (1946). Helv. Chi~n.i-\cta, 30, 496 (1947). A. C. MAEI.ILY and ?'. REICI-ISTBIN. A. R. PIXDER and M. SAIITII.J . Chem. Soc. 113 (1954). J. Am. Chem. Soc. 53, 352 (1031). F. E. ~ V I L L I A ~and I S E. GEBAUER-F~~LNEGG. N. IC. RICHTIIEYBR. Methods Carbohydrate Chem. 1, 107 (19G2). L. G O O D ~and ~ A J. N E. CHRIS TENSE^'. 1. Org. Chem. 28, 158 (1963). N. D. SCOTT, T. F. WALKER, and V. I,. I~ANSLEY.J. Am. Chem. Soc. 58, 2442 (1936). . 1199 (1935). G. T. ROBEI~TSOE and C. I.: GRIFFITII.J. C h e ~ n Soc. H. R. R o ~ ~ r c eand n D. A. PRINS.I-Ielv. Chim. ilcta, 28, 465 (1945). E. G. ANSELL and J. HONEYAIAE. 1. Chem. Soc. 9778 (1952). N. I<. ~1c11~ArEnr~ and C. S. M G D ~ ~J.X Am. . Chem. Soc. 63, 1727 (7941).
" N o atter)~pt7uas made to isolate or to estimate tolirene o n the assrr?r~ptionthat i t mzrst l~aoebeen present as Part of the prodrrct. I t s anzolint zuas assl~lrledto be eqrial to t l ~ a tportion of tile bejzzylidetze groz~p7uliirlz did not f o r l i ~ bibenzyl. Tlze latter corjzpoz~?zdwas readily obtained al))rost qlrant~tatively.
