Sifting segregation of ideal blends in a two-hopper tester: Segregation profiles and segregation magnitudes

摘要

The main aims of this work were: 1) to increase the understanding of the segregation process that occurs during a segregation test that follows the ASTM sifting segregation procedure, and 2) to understand the effect of powder load, fines mass fraction and particle diameter ratio on the segregation of ideal blends. Binary blends of microcrystalline cellulose particles (mean diameter ranging from 161 to 661 mu m) were used. The segregation tendency of the blends was characterized with a tester built according to the ASTM D 6940 standard. Additional experiments were performed using only the upper or the lower hopper of the tester, to better understand the contribution of the set-up to segregation. The fines fraction in the samples collected during the segregation experiments was determined with a QicPic particle size analyser, and segregation profiles were obtained. Most of the segregation occurred when the upper hopper was discharged into the lower hopper. For most of the profiles, an initial phase presenting a significant oscillation in the fines mass fraction was followed by a second phase where the fines mass fraction was rather constant and, finally, by a third phase where the fines mass fraction decreased. Segregation was quantified calculating complementary coefficients: the standard deviation of the normalized fines mass fraction, the fines mass fraction ratio of the last to the first sample, the fines mass fraction ratio of the sample with the lowest fines fraction to the sample with the highest fines fraction. The segregation amount was rather constant reducing the powder load from 838 to 665 g, while segregation largely increased when the powder load was further reduced to 435 g. The segregation tendency largely increased when the fines mass fraction in the initial blend decreased from 75 to 25% w/w. Finally, the segregation tendency increased when the particle diameter ratio increased from 1.9 to 4.1. The generated data are useful also for comparison when blends of increased complexity are used. (C) 2018 Elsevier B.V. All rights reserved.