2528
TfSDROCEN
Crimpounda
W. &JONES !I. TABLE I1 EVOLUTION FROM REACTIONS O F 0-DINITROSOBENZENE (DNB) AND SODIUM BOROHYDRIDE (SBH) G.
Mmole
SBIf,O,b mmole
Hydrogen evolved during reaction c cc. hlmole
1701.
OF
s'7
O-BENZOQUINONE DIOXIME (Ban) \{'ITFI
snii
1 I y d r r ) ~ e ue v o l v e d by acid cc. IImPlr
Excess
SRH,d mmole
consumed i n reactiim, mmolc
n ioo 0 . 5x5 0.761 18.7 DNB 0.0890 0.654 0.745 12.5 0 . 50ge ,191 .564 ,765 DNB ,663 .745 13.5 .550 18.8 ,0902 ,145 5/56 .833 14.2 ,580 .701 20.4 .BO8 BQD ,0839 ,398 373 39.1 1.59 ,254 .701 9.2 .375 BQD .0350 O Samples dissolved in 3 ml. of diglyme purified by distillation from lithium aluminum hydride and redistilled, b.p. 88- 90" (52 mm.). * SBH (0.769 mmole) which assayed 97.OyOused with DNB; SBH (0.769 mmole) which assaved 91.2% 1,sed with BQD. Blank runs with sodium borohydride in purified diglyme gave no more than 0.1 cc. of evolved gas after 2.5 hours a t 35-40". For DNB, reaction time was 3 hours a t 35-40'; for BQD, reaction time was 1.5 hours a t 35-40'; hydroOne-fourth of the mmoles of hydrogen evolved by acid according to the equation S a gen measured a t 27" (i62 mrn.). BHI H~SOI 3H20 4 4Hz HIBOa NaHSOa. eSample calculation: n = (iS2/760)0.0125/O.OS205 (300) = 0.000509 mole = 0.509 mmole.
+
+
+
+
ranged from 0-30%. The initially acidified reaction mixture was made strongly alkaline with sodium hydroxide with the precipitation of 0.1-0.2 g. of a yellow or brown solid which did not melt below 300' and was insoluble in hot methylene chloride or n-hexane but very soluble in hot ethanol. It dissolved readily in cold concentrated hydrochloric acid or glacial acetic acid and gave a wine colored solution in each case; it was insoluble in cold 10% sodium hydroxide. It could be sublimed with difficulty a t 170-180O (0.02 mm.) whereupon a red-orange solid which did not melt below 300" was obtained. Reduction of nitroso derivatives and furoxans is summarized in Table I . I n two experiments, o-dinitrosobenzene was treated with excess sodium borohydride (assayed by hydrogen evolution in a reaction with sulfuric acid) in diglyme following a procedure13 which allowed the collection over mercury of evolved hydrogen, Excess borohydride was removed with dilute sulfuric acid. Based on hydrogen evolution measured be___(13) J. C. Rankin and C. L. Mehltretter, Anal. Ckem., 28, 1012 (1BSfi).
[COSTRIRUTIOY FROM
THE
fore and after treatment with acid, the amount of sodium borohydride used in reduction was calculated. X sample of the gas collected above water during the reduction burned explosively in air and gave negative tests for the presence of boron compounds. Under similar conditions o-benzoquinone dioxime was treated with sodium borohydride. The results obtained with each compound are described in Table 11. Reaction of Bis-(a-hydroxybenzyl)-furoxanwith Sodium Hydroxide.-A suspension of 0.30 g. (1.0 mmole) of the furoxan in 15 ml. (38.0 mmoles) of 10% sodium hydroxide solution was stirred at room temperature for 45 minutes. The solid slowly dissolved as the solution became pink and finally yellow and the odor of benzaldehyde was noted. After acidifying with 5 ml. of glacial acetic acid, a solution of 2,4-dinitrophenylhydrazine was added, the mixture was heated to the boiling point, cooled to 0' and filtered. The crude product recrystallized from methanolic hydrochloric acid as small orange plates of benzaldehyde 2,4-dinitrophenylhydrazone, m.p. and mixture m.p. 243-245", 0.12 g. (21%).
NEW ORLEANS 18, LA.
DEPARTMENT O F CHEMISTRY, THEUNIVERSITS O F
FLORIDA]
Amine Borane Reductions. The Stereochemistry of the Reduction of 4-t-Butylcyclohexanone with Trimethylamine Borane in the Presence and Absence of Boron Fluoride BY W. M. JONES RECEIVED OCTOBER 9, 1959 The reduction to the alcohol of 4-t-butylcyclohexanone with trimethylamine borane and diborane in neutral, non-aqueous solvents was found to give approximately 16% of the cis isomer in each case. These results suggest that the two reactions have a common intermediate On the other hand, the reduction of 4-t-butglcvclohexanone with trimethvlamine borane and diborane in the presence of boron fluoride gave, under the same conditions, 46-52% and 16% cis-4-t-hvt~~lc~clohexan~1, respectively. These resvlts are rationalized in terms of the trimethylamine borane reduction proceeding by initial complexing of the boron fluoride with the carbonyl oxygen followed by intermolecular hydride transfer from the anline borane. Evidence is also presented which indicates that the resulting ionic intermediate I decompose to 4-t-butylcyclohexyldifluoroboronite I1 and 111.
Of the many substituted boranes and borohydrides that have been found in the past few years to be effective reducing agents' one which we feel has potential for becoming especially useful due to its stability1bs2and convenience to deal with is trimethylamine borane.s The possibility that (1) (a) N. G . Gaylord, "Reduction with Complex Metal Hydrides," Interscience Publishers, Inc., New York, N. Y..1956; (b) Callery Chemical Co., Technical Bulletin C-200,Callery, Penna., April 1, 1958. (2) A. B. Burg a n d H. I. Schlesinger, TEISJOURNAL, 59, 780 (1939). (3) Commercially available from Callery Chemical Co., Callery, Penna. T h e trimethylamine borane employed in these investigations was most generously supplied by this company.
carbonyl reductions involving this reagent might well proceed by either of the general paths R R a0 I I RC=O f Me3NBHs + R-C-0-BHs II
@
e
--+ RCOBI-I2 I
R
R
R
II
RCH R
May 20, 1960
AMINE BORANE REDUCTIONS
2529
intrigued us to undertake a rather thorough examiTABLE I nation of this reaction. The first approach which --Molar ratio-we elected to pursue was the stereochemical conTrimethylcistranssequences of the reduction of 4-t-butylcyclohex- Reamine Boron Reduction Reduc- Alco- Alcotone borane fluoride time tion, % hol, % hol, % anone4 in the presence and absence of Lewis 46 54 2 min. 100 1.0 1.0 1.0 acids.5 46 54 30 min. 100 1.0 2.0 1.0 Reduction of 4-t-Butylcyclohexanone with Tri46 54 30 min. 1.0 4.0 1.0 100 methylamine Borane and Diborane in the Absence 30 min. 100 46 54 1.0 2.0 3.0 of Lewis Acids.-The reduction of 4-t-butylcyclo1 min. 1.0 0.2 1.0 39 hexanone with trimethylamine borane in neutral, 1.0 .2 1.0 12 min. 40 non-aqueous solvents (diglyme, benzene) was found 1.0 2 min. 68 49 51 1 .o .33 t o be quite slow even at steam-bath temperature. 2 min. .33 0.67 1 .o 67 49 51 The stereochemical consequences of this reaction 100 52 48 1 0 .33 1.0 16 hr. were not appreciably different in the two solvents, 1.0 1.0 0.3 1 rnin. 31 the reductions in benzene and diglyme being found 1.0 1.0 .33 10 min. 35 to give mixtures of the two isomeric alcohols con.33 31 0.33 10 sec. sisting of 15% and 17y0cis-4-t-butylcyclohexanol, 1.0 20 min. .33 .33 35 1.1) respectively. Reduction of the ketone was also 42 .33 .44 1.0 1 min. effected in diglyme by employing diborane as the .33 .44 45 12 min. 1.0 reducing agent.6 Whereas a t low temperature (ice-bath) diborane caused quantitative reduction reduction gave a mixture of alcohols consisting to a mixture of 8.5% cis-4-t-butylcyclohexanol and of 25% cis-4-t-butylcyclohexanol and 75% of 91.5% of the trans isomer, reduction effected by passing diborane into a hot solution (steam-bath) the trans-alcohol. Discussion of the ketone in diglyme gave a mixture of 16% cis-5-t-butylcyclohexanoland 84% of the trans Diborane and Uncatalyzed Trimethylamine alcohol. Borane Reductions.-Brown, Schlesingerab and Reduction in the Presence of Lewis Acids.-In Burg have suggested that diborane reductions of the presence of added BFa (as the etherate), tri- ketones most likely proceed by an initial complexing methylamine borane was found to be a much more of the BH3 moiety with the carbonyl oxygen foleffective reducing agent. Not only was the rate lowed by intramolecular hydride shift. of the reduction drastically increased, but the 0 O-BHs O-BHs stereochemistry of the reduction was found to be I1 le I RCR BHa +RCR +RCR quite different from that observed in the absence e 1 of added Lewis acid. Thus, for example, when H ketone, trimethylamine borane and boron fluoride etherate were mixed in a 1:1:1 molar ratio, the This mechanism gives a reasonable explanation ketone was quantitatively reduced to 4-t-butylcy- for the stereochemical consequences of the diclohexanol in a matter of seconds. The product borane reduction of 4-t-butylcyclohexanone in was found to contain 46% of the cis-alcohol and diglyme. Since the product-determining tran5470 of the trans isomer. It was also found that sition state is reminiscent of the product, it would the amount of reduction (in two minutes) was not be expected that I1 would be favored over I and affected by changing the molar ratios of starting the trans isomer would predominate. The fact materials as long as the acid was kept equal to or in excess of the ketone and as long as the trimethylamine borane :ketone ratio was one-half or greater. The stereochemical consequences of the reduction were not affected by changes in the concentration of trimethylamine borane as long as the trimethylamine borane:ketone ratio was I I1 equal to or greater than one. Reduction of this ratio below one caused small increases in the that the diborane and the uncatalyzed trimethylamount of the cis isomer produced (see Discussion). amine borane reductions gave essentially identical Exploratory reductions were also run with other ratios of isomeric alcohols (17% cis and 16% Lewis acids ; e.g., aluminum chloride. Although cis, respectively) suggests that the two reactions the majority of these results will be reported in might well have a common intermediate. Thus, a future paper, it is of interest and perhaps sig- it is suggested that the uncatalyzed reduction of nificant that the aluminum chloride-catalyzed 4-t-butylcyclohexanone with trimethylamine borane proceeds by initial slow formation of the ke(4) This ketone was selected in order to ensure the formation of contone-borane complex by either initial dissociation formationally pure cis- and trans-cyclohexanol products; c j . s. Winof the Me3NBH3or, perhaps, by participation of the stein and N. J . Holness, THISJOURNAL, 7 7 , 5502 (1955). carbonyl oxygen, followed by rapid intramolecu(5) E. L. Eliel and M. N. Rerick (ibid., 82, 1362 (1960)) have recently reported a similar study of the effect of aluminum chloride on lar hydride shift. the stereochemistry of cyclohexanone reductions employing lithium Kinetic investigations are in progress to gain aluminum hydride as the reducing agent. further evidence for this suggested reaction path (6) (a) H. C. Brown and B. C. Subba Rao, J . Org. Chem., 22, 1135 (1957); (b) H. C. Brown, H. I. Schlesinger and A. B. Burg, THIS as well as to determine the role of the ketone, if JOURNAL, 61, 673 (1939). any, in the breaking of the boron-nitrogen bond.
+
2530
1’01. S%
$+
than was observed. However, in the present case we are reckoning with a rather large “effective oxygen” size (oxygen complexed with BF3) which M~,NBH, - - ,\le,N-+p~oducts apparently offsets the steric control enough to make the two effects of about equal importance. Consistent with these suggestions is the fact t-Bu t Bu that reduction in the presence of aluminum chloLewis Acid-catalyzed Reductions.-The catalytic ride, which should create even a larger effective effect of boron fluoride in the trimethylamine oxygen size, was found to give a mixture of 257* borane reductions apparently does not arise from ris- and 75y0 tram-alcohols. Thus, applying the initial reaction of the Lewis acid with the amine above ideas, in this case the product development borane to give diboraneIb followed by reduction control becomes appreciably more important than since this reaction should lead to the same stereo- the steric approach control. chemical results as the reduction with diborane. T o this point, the discussion on acid-catalyzed Instead, as indicated in Table I, this reaction gave reductions has dealt with only one of the three ‘1 mixture consisting of 46yo cis-4-t-butylcyclopotential reducing hydrogens oi trimethylamine hexanol and 54yc of the trans isomer. The known borane. It, was therefore of interest to investigate facile reaction of ketones with boron fluoride6b this reaction further in an attempt to determine suggests that the reaction probably proceeds by the number of amine borane hydrogens actually initial coniplexing of the carbonyl oxygen with the involved in the reduction with an excess of the acid followed by an intermolecular hydride transfer amine borane as well as to determine the condiiron1 the amine borane. tions necessary for employing all three of these available hydrogens. It was mentioned earlier that the amount oi reduction (in two minutes) was not affected by changing the molar ratios of st,arting materials as long as fluoride was kept in excess and the aniine boraneketone ratio was 0.5 or greater. If, howt-Bu I-Bu I Bu I ever, this ratio was made less than one-half, reThe stereochemistry of the reaction is then duction was not complete in u few minutes and the readily explained by the sequence summarized amount of reduction was found to be equal to by Dauben, Fonken and Noyce.8 These investi- twice the amount oi amine borane present. For gators pointed out that the stereochemical con- esainple, in the presence of a full mole of €3F:; sequerices of the reduction of a cyclohexanoiie and ketone, 0.2 mole of amine borane led to 4070 proceeding by an intermolecular hydride transfer reduction in one minute or twelve minutes; oneshould be governed by twu competing factors third of a mole of the amine borane led to 68% which they chose to call “sterac upproaclz control reduction, etc. Thus, i t is clear that in a short and “product deaelopmerct control.” It was sug- time and in the presence of a limited amount of gested t h a t steric approach control should favor amine borane, only two of the available reducing formation of the u.wzuZ-alcohol due to interference hydrogens are employed. Further, it is obvious arising from interaction between the incoming that in the presence of excess aniiiie borane, at reducing species and ~ , ~ - U Y Z hydrogens. C L ~ On the most two oi the hydrogens are used and possibly other hand, product development control should only one. To gain information on the relative rates of the favor formation o f the thermodynamically more first two steps, the products from the reductions stable alcohol (in this case, the ~ Y C L cI Z product) .9 Thus, applying this idea to hydride transfer to involving a limited amount of trimethylamine the coinplexed ketone, two possible transition borane were ana!yzed. It was found that in those cases where the amine borane was forced to use states need to be coilsidered. two of its reducing hydrogens, the reduction gave a mixture of the stereoisomeric alcohols consisting of 497c of the ris isomer and 51% of the truns material. Although this is just outside the region of experiniental error ( *1%), these results were consistently obtained and we feel real confidence in their validity. This change in stereochemistry can be explained by assuming that, in the presence In view of the bulky nature of the reducing of excess trimethylamine borane, reduction inagelit it might have been expected that steric ap- volves only one o l the amine borane reducing proach control would be even more important hydrogens to gi1-c l ( j % cis-alcohol’”; however, O-BH3
(7) V. Gasselin, A i i i i . chiin. p k y s , , [;I 3, 58 (1894). (8)W.G. Dauben, G. J. Fonken and D. S. Noyce, THISJOURNAL, 7 8 , 2 5 i S (1956). (9) K. D. H a r d y and R . J. Wicker ( i h i d . , 80, 640 (1958)), have recently pointed o u t t h a t this argument does not lead t o a totally consistent explanation of t h e observed stereochemistry of cyclohexanone reductions with lithium aluminum hydride, sodium borohydride and various other reducing agents. However, we still feel t h a t it offers a reasonahle approach t o t h e system under question.
2 triinethylainine horanca
t 3 B F j --+ BzHs
+ 2Me3?rTBF3
~
(10) T h e possibility th.rt reduction in t h e presence of excess trimerhylaminr borane actually involves competing reactions employing both t h r first and second a\-xilahIe hydrogens was eliminated by varyi n i t h e amine borane:krione ratio ( t o a s high as 4 : l ) . I t was found t h a t thestereochemistry of t h e reaction wasinsensitive t o these changes. These observations ;rlso eliminate the possibility t h a t t h e decrease i n t h r a n i o i n l t of ,-isis