Sept., 1950
FACTORS INTERFERING WITH
THE
OPPENAUER OXIDATION OF AMINOALCOHOLS 4085
Starting with Reichstein’s substance L acetate, hydroxy-11-desoxycorticosterone (Reichstein’s now readily available, a convenient partial syn- substance S). thesis of Reichstein’s substance P 2 1-monoacetate Several of these substances are being subjected (XVII) and 3,21-diacetate (XVIII) is described. to a variety of biological tests particularly in The former was carried through the same trans- regard to their potential usefulness as antiformations as performed in the corresponding 21- arthritic agents. desoxy series and afforded two new analogs (XI1 L~~~~~ M~~~~ 413 and XTV) of the important cortical hormone 17a- MEXICO CITY,D. F. RECEIVED JAXUARY 20, 1950
[CONTRIHUTIOh FROM THE C O H B CHEMICAL LABORATORY OF THE UYIVERSITY O F \.IRCIKIA]
Factors Interfering with the Oppenauer Oxidation of Amino Alcohols’ BY ROBERTE.LUTZ,ROBERT H. JORDAN,*^ AND WILLIAM L. TRUETT“,~
In recent attempts3 we found that, like quinine,4aseveral typical aliphatic 1,2-amino alcohols did not undergo the Oppenauer oxidation with aluminum t-butoxide and benzophenone or cyclohexanone. However, aluminum isopropoxide reductions of the corresponding, aiid other, a-amino ketones have proceeded without difficulty in all cases tried except the a-(N-alkyl-ethano1amino)ketones which are cyclic.6 It follows that under the oxidizing conditions the equilibria6 lie well over on the side of the reductants. The explanation offered for this phenomenon, involving acid-base combination between nitrogen and a l ~ m i n u mdid , ~ not seem to us to explain adequately why the alcoholic group, which would still be free, did not undergo the oxidation with reasonable speed when the solubility of the complex was appreciable, and it did not seem to be consistent with the facility of aluminum isopropoxide reductions of a-amino ketones where presumably similar complexes might be formed. The phenomena may be explained in ternis of stable cyclic complexes of the type (I) which if formed would cause displacement of the equilibrium sharply in favor of the amino alcohol, and which would be expected to interfere seriously with (1) The work of this paper was supported in p a r t by 8 grant-in-aid from the National Institutes of Health, recommended b y t h e Natioual Cancer Institute, and it stemmed from t h e program of syntheses of I .2-amino alcohols a s tumor-necrotizing agents. (2) (a) Post-doctorate Fellow‘; (b) Philip Frances d u Pont 1:eIlow; ( c ) as.;i+ted Iry Preston H. Lenke and Rosser L. Wayland, Jr. (3) T h e first unsuccesdul Oppeuauer oxidations in this laboratory were carried out 1)). Dr. K. S. hlurphey usinp aluminum isopropoxide. l t r . C. K. Baucr thru trsed aluruinum I-butoside on I V a which will be described in a M e r publication. (4) (a) Woodward, Wendler ~ n dBrutschy, TMS JOURSAL, 67, 1425 (1945). Cf. (b) Doering and Aschner, ibid., 71, 838 (1949). (c) Doering and Young, ibid., 72, 631 (1950). (5) (a) Lutz, Freek and Murphey, ibid., 70, 2013 (1918); (b) Lutz and Jordan, ibid., 71,906 (1949). (6) Concerning t h e reversihility of this reaction, cj’. (a) Wilde, “Organic Reactions,” Vol. 11, John Wiley and Sons, Inc., New York, N. Y . , 1911, p. 178; (b) “Xewer Methods of Preparative Organic 1948,p. 156; Chemistry,” Intersciencc Publishers, New York, N. Y., (c) Adkins, Elofson, Rosson and Robinson, THISJOURSAL, 71, 3622 (1949); (d) Adkins and Cox, ibid.. 60, 1151 (1938); (e) Baker and Adkins. it&/., 8’2, 3305 (1990); if) Baker and Schafer, i b i , / . , 66, 167.7 ( 1 R 13).
hydrogen transfer if the reaction were an intramolecular one involving a quasi-ring intermediate or transition state such as II.*l7 Five-membered
I R’ = H or CsHS
I1
ring complexes (I) might well involve a significant degree of added stabilization through second order or ‘‘nobond” resonance. Analogous six-membered ring, complexes (111) based on 1,3-amino alcohols would be conceivable although resonance stabilization would be excluded or a t least greatly diminished because of the break in conjugation involved in the extra methylene group. Larger rings or linear polymeric complexes presumably would not be stable. Thus the amino alcohols might fall 111
,C”\ K’CH
I
i-NK”,: ~
C6H:CH
-Al(ORI,
‘0,
into three categories; the 1,%types which should not undergo the standard Oppenauer oxidatioii ; 1,4 and l , 5 and longer amino alcohols which should generally be oxidizable without difficulty; and the intermediate 1,3-amino alcohols where a less predictable reaction might depend on structural and steric effects. Preliminary studies, successful as far as they have gone, have been made to test these predictions. Nine typical lJ2-aminoalcohols (IVa-c, 17, and XIII-XVII of Table I) including one diasteroisomeric pair (IVb and c), were recovered unchanged employing aluminum t-butoxide and benzophenone or cyclohexanone; three of these were used in the form of the free bases, four as the hydrochlo(7) (a) Baker and Linn, ibid., 71, 1399 (1949); (b) Lutz a n d Gillespie, ibid., 72, 314 (19.50); (c) Jackman and Mills. Valure, 164. 7R0 (1949).
lt. E;. LUTZ,K . H.
4080
JOXDAN AND
rides, and two both as hydrochlorides and as bases. A s a check on these findings a reduction in a typical case (XIV.HC1) was carried out, successfully, under conditions similar to those employed in the Oppenauer oxidations using toluene, with aluminum fbutoxide arid ben7hytlrol in place of aluniinuiii isopropoxide and is01)ropyt dcuhol Thc equilibria in the presence of ~'xcessaluniirium f-butoxide are thus shown to be consistent1.i far o\ cr on the side of the amino alcoliol~ 1 !'i
cir
~'iic
OII\K.
XRz = ( a ) S ( C L H ~ ) C H ~ C H ~ O C H ~ ' , Ib) SC,H,o(iiorner-A), ( c i isorner-R, id) SHC6H5, SrCH3)C6N,
C,H ,CWCW~\C,,I-l, j I
\1
)H
UC,Hlie = piperidyl)
It is noteworthy that no stereochemical equilibration of the epimeric amino alcohols (11% and c) occurred in the attempted Oppenauer oxidations using aluminum t-butoxide, a fact which in itself indicates that the oxidation rate is exceedinglv low..'b The I,%-anilinoalcohols are of interest because oi their comparatively weakly basic nitrogen. It1 two cases (IVd and e ) the Oppeiiauer oxidatioii under the usual conditions gave mixtures seemingly the result of partial reaction; in the case of IVd a small amount of the pure a-anilino ketone (XXI) was isolated and identified; and in the case of IVe some pure unchanged material was finally isolated from the mixture and identified. Thus it appears that N-aryl a-amino nitrogen is not as effective as aliphatic alkylarnino nitrogen in interfering with the oxidation. In three cases chosen arbitrarily (IVb, I\'c aiid V), under comparable conditions, successful oxidations of 1,2-aniino alcohols to the amino ketones were accomplished using an excess of potassiuni t-butoxide. In one of these cases (IVb, a catalytic amount of potassium t-butoxide was used, with similar results a t the higher temperature tieces sary to achieve reaction withiri a reasonable time, this experiment showed that the oxidation had not proceeded by virtue o f tabilkation o f the. amino ketone in the form o an addition coni pound or an enolate. I t therefore follows that the true 1,2-amino ketone-amino alcohol equilibrium is not as far different froin Ihc equilibria o f simple ketone-alcohol systems as might have been supposed6cbut is sharply tlisplaccd in the d i m tion of the amino alcohol by an excesh of .l!uini num t-butoxide. A typical 1,4-aniiiio alcohol (VI) was prepared As was anticipated
W. L. TRUETT
Vol. 7 2
oxidized to the y-amino ketone (XX) under the Oppenauer conditions using cyclohexanone and atuminuni t-butoxide. Other pertinent examples are the successful aluminum phenoxide-cyclohexanone oxidations of the alkaloids yohimbine and corynanthine which are 1,Ganiino alcohols.8~6hI t IS coiiclucled therefore that in those cases where the amine nitrogen ant1 alcoholic hydroxyl are sepmntetl froin each other by a chain of a t least Iour carhotis Oppmaiier oxidations generally will I,? suc*cc~isful In thc case O i the intermediate 1,3-type amino alcohols thr situation is coinplicated by the tendericy of the $-amino ketones to lose nitrogeng p-Chloro and p-bromo-/3-aminopropiophenones10c are extensively deaminated during aluminum isopropoxidr reductions. Other amino ketones of this type are rleaminated during catalytic reductions, 10.1 !J The reaction between aluminum isopropoxide and l,%-diphenpl-3-piperidyl-l-propanone (VIIL proceeded exceedingly slowly with deamination as the chief result, but catalytic reduction was reasonably successful arid gave the amino alcohol i1'1 I1 j . 'I'his ainino alcohol either as the free base or '1s Ihc sdlt wal; rccovered unchanged ( 3 . 5 - - K 3 ~ o ~ (:,,€I CC)CHCHjUC,III,, --+ Pt t H-
ChHjCH-CHCH?SCjHI, INC',Hlo = piperidyl)
OH
I
CsHj 1'111
after treatment with aluminum f-butoxide and cyclohexanone or benzophenone, but it was destroyed using potassium t-butoxide as the catalyst. I t should be noted that because this atnino alcohol is relatively stable under the usual Oppenauer oxidation conditions, it follows that the amino ketone (VII), which is largely albeit slowly deaminated under alutiiiiiuni alkoxide reduction conditions, has not in this case first undergone reduction ; the deamination must have involved the ketone, directly, in a reaction (perhaps initially enolization) which competed successfully with a very slow reduction. It is clear from these experiments that neither aluminum alkoxide oxidation nor reduction goes readily here. In the case of 3-piperidylpropiophenonehydrochloride ( I X ) ,in contrast with the #-chloro and p bromo analogsI0' referred to above, aluminum is0 I)ropoxiclc reduction gave a 657{ yield of the deiired muno alcohol (XII), and a 20yo yield of a imi-basic product which was identified a s phenyl vinyl carbinol (XI). The latter compound was ( 8 ) id, Janot and Goutar?l. Bull PUC c h t m , 509 (1949), (b) a'itkup. A m , 564, 83 (1943) (9) Cf la) hlannifh and Hetlner, Bri ., 66, 363 (1922), (b) Blicke and B u r c k h d t e r , THISJ o u n ~ r 64, ~ , 451 (1942), (e) Snyder and
Ore Chew , 11, 10.; (1946). L) I , U t / , P I ai , r h d 12, 6R0
Sept., 1950
FACTORS INTERFERING WITH
THE
OPPENAUER OXIDATION O F AMINO ALCOHOLS
4087
gation of the C-0 or C=O groups with the C-N linkage (inductive effect). The 2-nitrogen should increase the activity of the carbonyl group toward aluminum alkoxide reduction and a t the same time diminish the oxidizability of the carbinol a-hyAI(OR)3 drogen. These influences would operate in the * [CeH6COCH=CH2] C~H~COCH~CH~KC~HIO 1,3- and longer amino alcohol and ketone systems IX with progressively diminishing intensities, and perhaps effectively only when coupled with apC ~ H ~ C H C H ~ C H & C ~fJ H~O CsHbCHCH=CHz preciable steric effects (e. g., VIII). This alterI I nate hypothesis however does not explain the XI1 OH XI OH shift of the equilibrium toward the amino ketone that, as would have been expected, the para sub- in a typical case in the 1,2-types (IVb) when a stitution of halogen in IX favored the deamination catalytic amount of potassium t-butoxide was reaction.l°C The amino alcohol (XII) either in the used. form of the base or the hydrochloride was conI n a recent paper by Adkins, et a1.,6con “oxiverted into non-crystalline non-basic products in dation potentials” of a series of ketones, calculated an Oppenauer oxidation using aluminum t-butox- from equilibria, it is reported that a-(N-piperidyl)ide. The stability of the amino alcohol and the acetophenone has an oxidation potential higher instability of the amino ketone under the reaction by 85 mv. than that of acetophenone itself, but conditions were demonstrated in independent ex- that the equilibration rate is considerably lower. periments in which the oxidant was omitted. It Unfortunately equimolar or slightly larger is therefore inferred that the Oppenauer oxidation amounts rather than catalytic amounts of alumitook place with appreciable speed in this series, but num t-butoxide were used in the equilibrations.6c no inference can be drawn as to the relative posi- The observed and very low equilibration (i. e., tion of the oxidation-reduction equilibrium. oxidation) rate would be an expected consequence Steric hindrance to reaction may be an impor- of the formation of a stable amino alcohol comtant or dominant factor in the 1,3-amino alcohol plex such as I. It is doubtful that catalytic series where there is little or no activation of the amounts of aluminum t-butoxide would have been carbonyl groups. This might explain the fact adequate. However, catalytic amounts of potasthat the amino alcohol VI11 is not oxidized read- sium t-butoxide probably could have been used ily nor is the amino ketone VI1 readily reduced successfully (as here in the case of IVb) because using aluminum alkoxides under the usual condi- of the absence of impeding complex formation tions. On the other hand if a six-membered ring and because of the high hydride-a-hydrogen complex (111) were possible here, then the differ- activity in, and greater concentration of, the ing behaviors of the two compounds VI11 and XI1 alkoxide ions produced by exchange with t-bumight be attributed to an understandable though toxide ions; and this might have led to results perhaps not predictable difference in ease of cycli- more nearly representative of the true amino kezation or stability of the complex (cf. the strik- tone-amino alcohol oxidation-reduction system. ing effect of N-substitution in facilitating cycliThe ideas discussed above suggested other zation of a-(ethano1amino)-ketones to hydroxy- problems of interest in this field, e. g., the effect of morpholines) .6 an adjacent hydroxyl, alkoxy1 or carbonyl on the It is possible that equilibria in all of these reac- Oppenauer oxidation. Allopregnane-3,17,20-triol, tions may be affected significantly by differences which is a typical example, is only partially oxiin conditions, solvent and reagent, under which dized by aluminum phenoxide-acetone-benzene to the two opposite reactions are studied. The sta- the 3-keto compound, l 2 possibly because of probilities and solubilities of the nitrogen-aluminum tection of the 17 and 20-hydroxyls in the form of a complexes might well be affected by the basicity five-membered phenoxyaluminum glycolate ring. of nitrogen which should be greater in the 1,2- The sharp raising of “oxidation potentials” of amino alcohols than in the a-amino ketones. ketones by alpha6c or orthoBf-methoxyl groups However, since the oxidations of the 1,4- and might be explained in part on the basis of chelate longer amino alcohols proceed normally in spite complexes analogous t o I. of their presumably still higher basicities1’ and Experiments are in progress to put the idea of their presumably readier tendency to form simple cyclic complexes to other tests. If the idea is coracid-base complexes, the basicity effects alone rect then the trans-cyclopropane, cyclobutane, would appear to be negligible in this connection. and other related 1,2-amino alcohols, should An alternative to the ring-complex explanation undergo the Oppenauer oxidation, whereas the cis of the interference with the Oppenauer oxidation of 1,2-amino alcohols may be suggested in terms isomers should not (and incidentally this reaction of the effect of resonance involved in the conju- should offer a dependable means of determining configurations in such types). On the other (11) The basicities of two typical 1,Z-amino alcohols are signifihand, should the cis and trans isomers both fail cantly lower than those of the corresponding phenethylamines to undergo the Oppenauer oxidation, i t would (Spencer,Leffler and Burger, unpublished results). presumed to have resulted from deamination of the 6-amino ketone to phenyl vinyl ketone (X) and subsequent reduction of the carbonyl group. Incidentally i t is interesting to note a t this point,
I
4
H . E. L I ' T ~ 31. , H
408s
JORDAN A N D
Vol. 72
iy. L. TRI'ETT
then be clear that ring complexes arc. i i * > t i n - compounds were specifically sought among the products ~ / m ~ r \ W l t nc1c not founti. volved. A similar study of some t i ,inti cyclic l,%glycols and glycol monoall,\ I e t h r s 1% S I 1 I = ~ - R r C ~ t l ~ C I I O H C H ~ N C ~ H ~ O ( i n o r p h o l i n y l i YI \ = -I-IC~F€~CHOHCH~N(n-butyl)~ also projected in this coniiec'ticiii \ i '< C1 I--CII,C~I-IICHOHCHJS(CH3)C€€~~~~-I Acknowledgment.--Exte T'c' colltrlbutlvas i (1 ('I iOHCH2Y (GHr,)2 the experimental work were inade ti?. I'restcm FI \\! = CI Leake and Rosstar I,. \Y:LyIanci, Jr
Experimental''Aluminum and potassiu
utoxide oxidations, m i i riiarized in Table I , werr e t i ottr ,iccording t o t i i t following general procetlur he rextion mixtur. c o i l sisted of 0.01 inole of the conipounct i n thc form of t h i h,iw o r the hytirochlortrlr l t 3 i nil of . . i ) l \ r a i i t . drv tolui!i( in I ZULL I OPPEXAUER OSIUAi i o \
A) dry bcii~eiie],oxidant, (B) 0.4 inole of cycloheu~noxie, or iC) 0.05 mole or ID) 0.1 mole of benzophenone, [or t!) no oxidant a t all but with 0.4 mole of added l-but,inol
for the purpoie of testing the effect of reaction condi0.0341.04 mole of aluniiiium lor iii ioine c~sey P) potdssiuni] t-butoxide. This mixture W V ~ Srcfluxed for c C;) 17-30, (H) 24 or (I) 40 hours, poured into ice w'tter, i i d e alkaline with l o g sodium hydroxide to dissolvc aluminum hydroxide, and separated, with appropriate extraction by added ether where needed. The combined organic ioirent extract was washed and the basic material rcrnovecl by extractions with portions of 2 .V hydrochloric icid until test showed the extraction to be completed. 1 i i i s o m ~tabes ( J ) the combined organic 5olvent extract i z 15 instcad dried over sodium sulfate And treated with cthere,tl hydrogen chloride t o precipit'ite the hydrochloride.] The combined aqueouq extract was cooled by addition of ice and the bas? was precipitated by addition of 10-3051 sodium hydroxide, or in d few c a s s by sodium iarI)otiate If the product solidified (e)it w'ts filtered m t l I ecrystillized from ethanol ; otherwise the resulting oil m't'* extracted with ether, and the 5olution WAS dried o\ cr -odium sulfate and either (IJ) evdpordted under reducctl pre>sure followed by crystallization of the base from cth,iiiol, or (M) the hydrochloride was precipitated by xldition of ethereal hydrogen chloride aiid recrystallized. were iclentified by mixture melting points field seem generally to be reliable even for d l the experiments listed, evcept two a? iiidii,itecl, only the one crystdine product indicated %-a> t,iiiic.d with no indication that the correipondirig .imino t o r i t tt'cts present as J significant impurity; the yields 'The rrst of t h t tcil rc-frr t o ielativrly pure trrateri.il LiollsJ j , . i d
'l'hc bteiizeiie et ,~por,~teri during the e\periiileiit , c i i t l ini\turt >toot1 at ciz. 150' for over ten hour,. 'tilother experiment tising xylene a? the medium xid iv.ictioii
L!IL
i ~ i tt-
fluxing for 20 hour:, a 45y0 yield of crude XVIII-HCl w,is obtained. purified 116%) and identified. * A quarter of .III equivalent of potassium t-butoxide (0.0025 mole) m ~ l 5 used. ' Crude but nearly pure. This run, in toluene as the solvent, using 0.3 mole of benzhydrol rather than benzophenone, and refluxing for three hours, resulted in recovery of unchanged material, but under refluxing for twenty-one hours gave the amino alcohol XIV.HC1 in 1'3'' yield (purified). It should be noted that therc wa5 iderable resiniiicdtion in this experiment. 'Thest
__
- .__
1 7
1 2 ) Microaaaljsc~ 1,) ,I. t w a l 1 I h O ! xt
rcf 5:,
Sept., 1%0
FACTORS INTERFERING
WITH THE O P P E N A U E R OXIDATION OF AAMINO -ALCOHOLS
which was recrystallized from ethanol; 12 g. (84ojO); m. p. 122-124" ( V ~ i g t , 119"). '~~ N-Desyl-N-methylaniline (XXII) was made by condensation of 23 E. of desvl chloride and 35 ml. of methylaniline (100' for fair hours), and crystallized from ethaflol ; pale yellow; yield 12 g. (40Y0); m. p. 99.5-101.5 Anal. Calcd. for C*lHlsT\;O: C, 83.69; H , 6.35. Found: C, 83.74; H , 6.51. 1,2-Diphenyl-2-(N-methylanilino) -ethanol (IVe) .-Reduction of 12 g. of X X I I by aluminum isopropoxide (as above) and crystallization of the groduct from ethanol gave 4.2 g. (34%) ; m. p. 110.5-112 Anal. Calcd. for C21H21KO: C, 83.13; H , 6.98. Found: C, 83.18; H, 6.58. a-(Piperidylmethyl)-benzyl Alcohol (V) .-Reduction of 5 g. of XI>;" (see Table 1) with 45 ml. of 1.8 M aluminum isopropoxide under slow distillation for three hours, and recovery of material by hydrolysis in dilute sodium hydroxide, extraction with ether, precipitation of the salt by ethereal hydrogen chloride, liberation of the base by dilute sodium carbonate and extraction by ether, gave 3 g. of crystals of m. p. 67-68" (60%) (B. and B.,15 69-70"). Preparation from styrene bromohydrin in refluxing benzene (48 hours); yield 3670. The hydrochloride was crystallized from alcohol by addition of ether; m. p. 192194" (B. and B.,14 193-194'). Anal. Calcd. for C 1 ~ H l ~ S O ~ H CC,I : 65.13; H, 7.57. Found: C, 65.21; H , 7.45. a-Piperidylacetophenone hydrochloride16 ( X I X , Table I ) , when recovered in the several experiments, was identified by mixture melting point with material prepared from phenacyl bromide and piperidine. We found the melting point to be somewhat higher than reportedl6; it began to soften a t 210 O and melted over 228-230' (initial bath temperature 200' and heating rate 1O per minute). 1,2-Dipheny1-3-piperidylpropanol-l(VIII).-The amino k e t ~ n e ' ~ (VLI) , ' ~ (12 9.) in ethanol, acidified with 5 ml. of concd. hydrochloric acid beyond that necessary to give the congo reaction, was reduced16 with platinum and hydrogen a t 3 atmospheres (1.5 hours) ; the theoretical amount of hydrogen was absorbed. The crystalline precipitate which had appeared was dissolved by heating, and the solution was ofiltered and cooled; yield 6.4 g. (53%); m. p. 254-256 Anal. Calcd. for C2oH2aNO.HCl: C, 72.38; H , 7.90. Found: C, 72.36; H, 8.21. The base (VIII) was liberated by sodiup carbonate and crystallized from isooctane; m. p. 91-93 Anal. Calcd. for CZOHZBSO: C, 81.31; H, 8.53. Found: C, 81.47; H , 8.22. Reaction of the amino ketone (VII) with aluminum isopropoxide proceeded only very slowly. Precipitation of a complex from which VI1 could be recovered on hydrolysis, occurred quickly ; this changed slowly upon long heating (eight hours) into a complex which dissolved completely in aqueous sodium hydroxide. No basic product was precipitated, and it appeared that deamination was the chief result. p-Piperidylpropiophenone (IX) hydrochloride was reducedl6.l6 in ethanol with platinum and hydrogen a t 3 atmospheres with absorption of the calculated amount in 35 minutes. The product (XI1 hydrochloride) crystallized upon addition of ether t o the filtered solution; yield 67%. Aluminum isopropoxide reductionlo of 20 g. of IX with 50 g. of reagent in 300 ml. of isopropyl alcohol under reflux for six hours, followed by hydrolysis in ice and hydrochloric acid, and extraction with ether, gave an oil [ 2 g. (2OY0), collected a t 70-90" ( 5 mm.); n z 2 h P.5201 which was identified as (XI)" by conversion into tri-
.
.
.
.
(14) Blicke and Blake, THIS JOURNAL, 61, 235 (19301. (15) Mannich and Lammerling, Ber., 66, 3524 (1922). (16) Cf. reduction of 6-piperidylhenzylacetophenone: J. D. Smith, Doctorate Dissertation, University of Virginia (1946). (17) (a) Klages and Klenk, Bet.., 89, 2554 (1906); (b) Meinenheimrr and Link, Ann., 4T9, 246 (lQaO),
4089
bromphenylpropanel" and the p-nitrobenzoateIsb (identified by m. p. 123-124" and 46-48', respectively). The acid solution of the amino alcohol (XII) was treated with excess dilute sodium hydroxide and chilled, and the precipitated solid base was fi1tei;ed and recrystallized from petroleum ether; m. p. 63-64 ;. yield 11.3 g. (65%). In one experiment the base was distilled; b. p. 170" (50 mm.) . I t gave no mixture melting point depression with the sample prepared by catalytic reduction (above). 4-Piperidyl-1 -phenylbutanol-1 (VI) made by the method of Marxer,lSawas obtained under difficulties similar to those experienced by HauserlTain analogous syntheses. The hydrochloride crystallized from ethanol by addition of ether, and melted a t 127.5-128.5" (MarxerlBareported 109-111 ") . I t contained persistent water of crystallization which was eliminated upon drying in a desiccator over ascarite for several days. Anal. Calcd. for C15H28N0.HC1: C, 66.67; H , 8.99. Found: C, 66.77; H , 9.01. The base was an oil,Isa b. p. 152-153" (0.5 mm.); nZ6D 1.5297. Anal. Calcd. for Cl5H23NO*HCl: C, 77.20; H , 9.93; N, 6.00. Found: C, 77.36; H , 10.08; N, 6.28. It crystallized as the monohydrate when exposed to the atmosphere; recrystallization from n-hexane gave photoIt lost its solvent of sensitive crystals; m. p. 53-54'. crystallization and reverted t o an oil on standing over calcium chloride in a desiccator. -4nal. Calcd. for C14H23NO.H?O: C, 71.67; H , 10.02. Found: C, 71.94; H , 10.38. 7-Piperidylbutyrophenone (XX of Table I) was extracted by 4 N hydrochloric acid in the usual way (above) hut it partially crystallized from this solution; the material was filtered a t this point and combined with the rest of the product which was subsequently isolated in the usual way (see M above). I t was recrystallized froom ethyl aceDrying a t tate by addition of ether; m. p. 185-186 60" in zucuo was necessary to remove water of crystallization which was partially regained with great rapidity upon exposure to the atmosphere. Anal. Calcd. for ClhH21NO.HCl: C, 67.20; H, 8.28. Found: C, 66.90; H , 8.06.
.
Summary Nine typical 1,2-amino alcohols failed to undergo the Oppenauer oxidation under the usual conditions using aluminum t-butoxide ; but three were successfully oxidized using potassium t-butoxide. Two 1,2-anilino alcohols seemed to be partially oxidized under the usual Oppenauer conditions. One 1,3-amino alcohol failed to undergo the Oppenauer reaction, but another apparently did react. A typical 1,4-amino alcohol was oxidized under the usual Oppenauer conditions without difficulty. The deamination product obtained in the attempted aluminum isopropoxide reduction of ppiperidylpropiophenone was identified as phenylvinylcarbinol. Explanation of the interference with the Oppenauer oxidation of aliphatic 1,2-amino alcohols is offered in terms of five-membered chelate complexes involving aluminum, the alcoholic hydroxyl and amine nitrogen. CHARLOTTSVILLE, VIRGINIA RECEIVEDJANUARY 26, 1950 (18) (a) Marxer, Rclo. C h i n . Acto, 94, 2223 (1941); (b) cf. also Brealow, Walker, Yoit and Rawer, Taxi J O ~ N A L ,W, 1471 (1B46).
