Evidence is presented showing that it should be possible to depolymerize a
purified cellulose to produce anhydro-oligosaccharides containing small num
bers of glucosidic units if suitable reaction conditions are chosen. Experi
ments were carried out in a stream of nitrogen in a downflow cocurrent tubu
lar reactor with a high temperature wall, Two sizes of particles of Avicel
pH102 were used and two sizes of reactor tube with wall temperatures from 8
50 to 1200 degrees C and residence times from 35 to 75 ms. From the solid r
esidue of the reaction, a water soluble fraction was recovered which contai
ned anhydrosugars. Yields of soluble solids of up to 44% of the cellulose f
ed were obtained in a small diameter tube, and up to 34% in a larger diamet
er reactor, only a narrow window of reaction conditions exists which is opt
imal for maximum yields. Carbohydrate analyses showed significant yields or
anhydro-oligosaccharides from G(2) to G(7), the limit of resolution for th
e HPLC column used, Normally, these compounds made up similar to 50% of the
soluble solids fraction. Substantial amounts of material > G(7) were also
produced but could not be identified, Tests were also done with pretreated
Avicel, but only one method gave improved results (slow preheating) Given t
he probable high unit value of anhydro-ologosaccharides, it would appear th
at this method of flash pyrolysis may be worth further exploitation. (C) 20
00 Elsevier Science B.V. All rights reserved.