Glucose with dibenzylamine

Although there is no positive evidence in support of it, a plausible alternative to the above reaction sequence is enolization of the aldose before condensation with the amine. In such a sequence, the Y-substituted glycosylamine would not be a precursor of the aminodeoxyketose. This postulate merits investigation because, in spite of attempts to isolate Y, Y-dibenzyl-D-glucosylamine, this compound could not be found in the reaction product of D-glucose with dibenzylamine only D-glucose or the Amadori rearrangement product was isolated. ... 

To a suspension of D-glucose (59.8 g, 332.0 mmol) and dibenzylamine (64.0 mL, 332.0 mmol) in absolute ethanol (350 mL) was added acetic acid (57.0 mL, 995.7 mmol). The mixture was heated to reflux with an oil bath (temperamre not exceeding 90 °C) and stirred continuously for 3 h. After cooling to 0°C, the mixture was filtered, washed to white with ethanol, and dried under vacuum to give the rearranged product as a white solid (98.4 g, 83 %). 

The reaction of dibenzylamine with glucose 200 in ethanol, in the presence of acetic acid, gave a high yield of Amadori ketose 201, which was subsequently treated with calcium hydroxide in water and afforded, after acidic work-up, the branched ribono-1,4-lac tone 202a. No epimeric arabinonolactone 202b was formed under these conditions [74] (O Scheme 55). 

---