bility in hydrocarbon solvents, also appears to support such a formulation. When a phosphonic acid is gradually neutralized by addition of a base, the above formulation would indicate the possibility of initial neutralization of the "free" chain ends and a preservation, in whole or in part, of the essential structure of the chain. In experiments with n-butanephosphonic acid, which is a readily available experimental material, the acid was neutralized in aqueous solution with sodium hydroxide added in small steps. Specimens of solutions so obtained ranging up t o the point of exact half-neutralization in 5yc steps were carefully evaporated to dryness. The products obtained by neutralization up to one-quarter of all available hydrogen ions, Le., up to empirical fonuulatioii of IiPOBHB :RP03HKa, wcre waxy substaiices which melted well below the melting point of the frec acid (1n.p. 100") ; generally the softening bcgan a t OO&'lOo and extended to approximately 1 1 O", with formation of a very viscous transparent liquid. At the point of " 2 S ~ c "neutralization the nature of thc products underwent a decided change. The substance obtained a i this juncture was a frcely rnobilc powder which showed the first sign of partial liquefaction o d y a t 143" and the translucent jelly-like consistency was maintained even a t well above 200'. Extraction of these lower products with hot benzene resulted in slow removal of the free acid from speciniens containing less than W,i1 neutralized material ; the residual matter was identical with the "quarter" salt described above. Since the free acid is rapidly soluble in hot beiuenc on heating, the relative slowness of the extraction of the partial salts (0.3-1.0 hour) may serve as an indication of the time requirement for the separation of the chain fraginents of the residual phosphonic acid chain from the neutralized portioii, possibly through cleavage of monomer units which then associate. The "quarter-neutralized" product was not soluble in hot benzene and its solubility in water a t 25" was appreciably lower than that of the free acid: approximately 17 weight per cent. against 31 weight per cent. This suggests that the essentially chainlinked free acid has been transformed a t this juncture into a ring-linked structure, similar to that found in phosphonous i.e. :.I / o I1 0 , li l'\
110'
' k 0
If
0J1'\\W
Tlic atialogy is obviously incuiiiplctc, siiicc the pliosphonous acids appear to form a trinicric ring, lacking free hydroxyl groups. Products obtained by further neutralization of the acid are freely flowing powders of microcrystalline appearance; they shrink and sinter a t temperatures that rise froiii 1-13' shown by the "qu;trter" salt to 245' showii 1)y the "half" salt, ;.e., product corrcspoiiditig to KI'O$INL(E>/ /CL>/ I I 1 1
3
2
2
IV V VI (l,2&) (2,3-~is) (1,2,3-cis) (Y = H, or H.HBr, or P-COC6H4NO2)
I11 (1,3-cis)
For each of these four derivatives two structures (I or 11) are possible, depending on whether the Trifunctional Aminocyclanols. Synthesis of 3- parent aminediol is 3-aminocyclohexanediol-l,2 Aminocyclohexanediol- 1,2 or 2-aminocyclohexanediol-l,3. Structure I will B Y G. E. M C ~ A S L A NT. D , J . MATCHETTAND hfAKS€iA result if the nucleophilic attack by :NHI on 1HOLLANDER ethoxy-2,3-epoxycyclohexaneoccurs a t position 3, JANUARY 30, 1952 structure I1 if the attack is a t position 2. RECEIVED In order to determine the structure, periodic.acid Through previous studies' on 2-aminocyclanols studies were carried out. The aminediol N-pand 2-aminocyclohexanepentols (inosamines) we nitrobenzoyl derivative consumed nearly the theobecame interested in the synthesis of alicyclic com- retical amount of periodic acid for one -CHOHCHpounds having three neighboring amino and/or hy- OH- grouping, thus supporting structure I.6 droxy functional groups. Of the cyclohexane deFor each structure (I or 11) four diastereomeric rivatives with such substituents, the 1,2,3-triols conjigurations (111, IV, V, VI) are p ~ s s i b l e . ~Since have been thoroughly characterized by Posternak2; oxirane ring opening generally occurs in the trans however, the corresponding triamines and the aldi- manner, the aminediol of structure I might have amines apparently are wholly unknown, and only configuration 111or IV but not V or VI. To decide one investigation3 of the aminediols has been re- between 111 and IV one must know not only the ported. position of nucleophilic attack, but also the conA suitable starting material for preparation of figuration of the ethoxy-epoxide used. The conthe aminediols appeared to be l-ethoxy-2,3-epoxy- figuration of the epoxide must be either d,l-V or cyclohexane, whose preparation via the i ~ d o h y d r i n , ~d,l-VI, since the cyclohexane-oxirane ring-fusion is and also directly from 1-ethoxycyclohexene-2 by necessarily ~ i s . ~ , 8 a peracid t r e a t n ~ e n thad , ~ been described. We found Possible evidence on the configuration of the that a procedure using the chlorohydrin is somewhat ethoxy-epoxide may be found in the work of Postermore convenient, but the peracid method does give nak,2 who showed that i t is hydrolyzed by water to much better yields. give (after ether-cleavage by hydrobromic acid, The reaction of the ethoxy-epoxide with ethanolic which presumably causes no inversion) predomiammonia gave us an aminediol monoethyl ether6 nantly the meso-triol of configuration Unforof m.p. 132-134'. Its N-p-nitrobenzoyl derivative tunately, i t is not known whether the presumed numelted a t 201-203'. Cleavage of the aminediol ether with hydrobromic acid gave the aminediol as cleophilic attack by :OH2 on the oxide ring is a t its hydrobromide, m.p. 167-168O. The aminediol position 2 , or a t position 3. Hence the configurawas further characterized by conversion of the hy- tions of the ethoxy-epoxide and of our aminediol drobromide to the N-p-nitrobenzoyl derivative, derivatives remain uncertain. An earlier attempt to prepare an aminediol of which melted a t 182-183'. structure I1 from the corresponding nitrodiol was
.
(1) For related publications see G. E. McCasland a n d Donald A. Smith, THISJOURNAL, 78, 5164 (1951), and references there cited. (2) T. Posternak, Hclu. Chim. Acta, 3 0 , 441 (1947). (3) M. Mousseron, F. Winternitz and G. Combes, Bull. SOC. chim.. [5] 14, 79 (1947); M. Mousseron and R. Granger, ibid., [5] 14, 850 (1047); M. Mousseron, G. Manon a n d G. Combes, ibid., [5] 16, 396 (1940). (4) L. Brunel Compl. r e n d . , 160, 986 (1910); A n n . chim. Phys., [SI 6 , 269 (1805). ( 5 ) hlousseron, et d . , a reportedly obtained by the amination of this same ethoxy-epoxide a n aminediol monoethyl ether of m.p. l l O n , b.p. 148-150' (15 mm.), whose hydrochloride melted a t 160' and contained 9.7% N (calcd. 7.2%). The free base reportedly reacted with nitrous acid i o give 1-ethoxy-2-formylcyclopentane (semicarbazone m.p 204-205"). Yields were not reported, and the compounds were not lurther chdracterirrd. The product of m.p. 110' could be a n isomer. but is perhaps rnr~elya less pure sample, of the compound of m.p. 134' iiow r r p x t e d . 111 early 1947 hlousseron, el al., assigned their product structure I , but later wavered between I a n d 11. I t should be noted that, according to current interpretations of the pinacolic rearrangement ( e g., see G. E. McCasland, THISJOURNAL, 73, 2293 (1051)). both structures I and I 1 might be expected to yield the above aldehyde.
(6) It is conceivable t h a t a periodic acid solution of low $H might cause N -.f 0 acyl migration in a compound of structure -CHOHCH(NHC0R)CHOH-. This would produce a structure -CH(OCOR)CHNHiCHOH- scissionable by periodate. So far a s we know, no one has found, or looked for, such an effect. However, since in any case the periodate oxidation of a 2-aminocyclanol is very slow a t low PHI, we regard it as improbable t h a t the observed periodate uptake was due to acyl migration. (7) For structure 11, configurations IV and V become identical when both hydroxyl groups are free; therefore only three diastereomers would be predicted for such derivatives. (8) The possible presence of both diastereomers in the ethoxy-epoxide as ordinarily prepared has not been entirely excluded. ( s a ) FUOTNOTB ADDBU I N PRuOF.--Dr. J . A. McHae and Dr. R. Y. hloir of Queens University (Kingston), advise in a private communication t h a t they have actually found two diastereomers in the ethoxyepoxide prepared from the chlorohydrin, aud w i l l describe the separation and identification of these diastereomers in a forthcoming publication. (9) A small proportion of the d,l-triol of configuration I V ( = V ) is simultaneously formed. The meso-triol of the all-cis configuration VI is obtained by hydrogeuating pyrogallol.
