Sept., 1934 0
4
COMMUNICATIONS TO THE EDITOR 0
2013
form a continuous series of solid solution, the pos(IV) sibility of the separation clearly depends on the relative positions of the liquidus and solidus 0I co I \CH/ curves. Evidently, the negative carbethoxy and lactonic Some graduated fractional freezings were exgroups have so loosened the affinity of the tetra- ecuted, on my suggestion, by Dr. M. Strada, methylene a-and &carbon atoms that the linkage starting from ordinary distilled water. From denbetween them is hydrolyzed by the acid and the sity measurements made by Dr. Strada it was deformed S-hydroxy-carboxylic acid group under- duced that the final fractions contained up to 4 goes lactonization. Results with other a-lactonic per mille of D20. Similar results were found from esters and dilactones, containing mobile a-hydro- water obtained from old electrolytic cells of a gens, indicate that enolate formation is a general hydrogen plant. A short paper was consequently property of such derivatives. published by myself and M. Strada in Rend. Recently, Tschelinzev [Ber., 67, 960 (1934)l Accad. Lincei, 19, 433 (1934). assumed that enolates containing an a-hydrogen Doubts having arisen of the reliability of the react with carbonyl derivatives by intermolecular above experiments, I asked my friend and Assistaldolization, involving the respective hydrocarbon ant Professor A. Quilico to repeat the whole operand carbonyl groups. In accordance mono- ations, during which the greatest care to eliminate alkylated malonic and acetoacetic ester enolates sources of error, or to estimate their influence, was do not condense with ethylene oxide. Tschelinzev taken. Professor Quilico started again from 4000 overlooked that the surmized aldolization process liters of common water, which, by a nine-step was proved experimentally years ago and that his fractional freezing, he reduced to 280 cc. To the generalization is untenable. Enolates with an water 6% of sodium chloride was added each time a-hydrogen unite with phenyl isocyanate with in order to obtain ice as minute crystals and a t intermolecular migration of that hydrogen, form- each step about one-third of the whole was sharply ing well neutralized enolates [Michael, Ber., 38, separated by centrifugation, as ice. The final 22 (1905); Ann., 363, 64 (1908)l: with sodium fraction was distilled eight times in quartz apenol methyl acetoacetic and methyl malonic esters paratus, four times alone and four with the usual deep-seated decompositions occur, giving the well chemical reagents. neutralized salt of a complicated acid [loc. cit., p. On the liquid thus obtained nine independent 241. Sodium enol methyl malonic ester adds to density measurements were made, every time in A-",8-esters by migration of the methyl group to comparison with the original water and four times the carbon atoms [Michael and ROSS,THIS also with the water coming from the fifth freezing. JOURNAL, 52, 4592 (1930); 53, 1633 (1931)l and, The influence of temperature was particularly finally, Bennett [ J . Chem. Soc., 127, 1278 (1925)] considered. The uniform result was that the found that ethylene oxide condenses with (I), former determinations were entirely untrustwhich contains no a-hydrogen to give bis butyro- worthy. The density differences now found were lactone a,a-spiran. The subject is evidently com- in both directions and never exceeded the passible plicated, but the reactions always proceed with error of the measurement. the maximal possible neutralization of the metal. To eliminate the doubt that the chemical treatCONVERSE MEMORIAL LABORATORY ARTHURMICHAEL ment might originate total or partial destruction HARVARD UNIVERSITY NATHAN WEINER of D20, five more density determinations were CAMBRIDGE, MASS. made on samples purified by distillation only. AUGUST22, 1934 RECEIVED The results were the same. The conclusion to be drawn is that if a fractionT H E POSSIBILITY OF SEPARATING HEAVY WATER ation takes place; it amounts certainly t o less than BY FRACTIONAL FREEZING 1/10,000, and probably less than 1/100,000 when Sir : the original mass is reduced to 1/14,300. This behavior is probably due to the fact that On the basis of the fact that D2O has m. p. 3.8", the possibility of separating it from H20 by frac- the crystal lattice of both DtO and HzO is subtional freezing - was taken into consideration. Ad- stantially formed by the O-atoms while the mitting that the two oxides are isomorphous and H-rem. D-atoms do not materiallv affect its
I
I
CHaCHCH2CH (111)
C
