A difference of opinion exists as to the molecular weight and constitution of this substance. Iwanoff [[1909, 1]] regards it as a triosephosphoric acid, C3H5O2(PO4H2), basing this view on the preparation of an osazone which melted at 142°, but when recrystallised from benzene gave a product melting at 127°–8°, which had the same appearance, melting-point, and nitrogen content as the triosazone formed by the action of phenylhydrazine on the oxidation products of glycerol. Neither Lebedeff [[1909]] nor Young could obtain Iwanoff's osazone, and all attempts to reduce the acid with formation of glycerol either by sodium amalgam or hydriodic acid were unsuccessful (Young). There is therefore practically no serious experimental evidence in favour of Iwanoff's view.
On the other hand, Harden and Young regard the acid as a diphosphoric ester of a hexose. This view is based on the fact that when the acid is boiled with water, or an acid, free phosphoric acid is produced along with a levo-rotatory solution containing fructose and possibly a small proportion of some other sugar or sugars. (Euler and Fodor however did not obtain a hexose in this way [[1911]].) The acid itself only reduces Fehling's solution after some hours in the cold, rapidly when boiled, whereas when its solution is first boiled, and then treated with Fehling's solution in the cold, the products of decomposition bring about reduction in a few minutes. The reduction brought about when the acid is boiled with Fehling's solution is considerably less (33 per cent.) than that produced by an equivalent amount of glucose. The behaviour of the compound towards phenylhydrazine is also in complete agreement [p050] with this view. Lebedeff found [[1909], [1910]] that the acid or its salts heated with phenylhydrazine in presence of acetic acid gave an insoluble compound which was ultimately found to be the phenylhydrazine salt of hexosemonophosphoric acid osazone
C6H5NH·NH2·H2PO4·C4H5(OH)3·C(N2HC6H5)CH(N2HC6H5)
[Lebedeff, [1910]; [1911, 6]; Young, [1911]]. After recrystallisation from alcohol this compound forms yellow needles, melting at 151°–152°. It is decomposed by caustic soda yielding a sodium salt
Na2PO4·C4H5(OH)3·(CN2HC6H5)·CH(N2HC6H5)
and on boiling with caustic soda decomposes giving a hexosazone (free from phosphorus) which is probably glucosazone, and in addition glyoxalosazone, probably as the result of a secondary decomposition. Towards acids it is remarkably stable yielding with hydrochloric acid a hexosonephosphoric ester from which the original osazone can be regenerated (Lebedeff). Lebedeff at first [[1910]] argued from the formation of this osazone that the original hexosephosphate contained only one phosphoric acid group per molecule of hexose. It was however shown by Young [[1911]] and subsequently confirmed by Lebedeff [[1911, 6]] that one molecule of phosphoric acid is split off during the formation of the osazone, even in neutral solution. Moreover it has been found that in the cold hexosediphosphoric acid reacts with 3 molecules of phenylhydrazine forming the diphenylhydrazine salt of hexosediphosphoric acid phenylhydrazone
(C6H5NH·NH2·H2PO4)2·C6H7(OH)3·N2HC6H5.
This compound crystallises out when 1 volume of alcohol is added to a solution of 3 molecules of phenylhydrazine in one of the acid and forms colourless needles melting at 115°–117°. p-Bromophenylhydrazine yields an analogous compound melting at 127°–128°.
Precisely the same products are given with phenylhydrazine by the hexosephosphoric acid prepared from glucose, mannose, and fructose, proving that all these sugars yield the same hexosediphosphoric acid, a point of fundamental importance.
Direct measurements of the molecular weight of the acid by the freezing-point method, combined with the determination of the degree of dissociation by the rate of cane-sugar inversion, are indecisive, but indicate that the acid has a molecular weight considerably higher than that required for a triosephosphoric acid.