An empirical model predicting the viscosity of highly concentrated, bimodal dispersions with colloidal interactions
An empirical model predicting the viscosity of highly concentrated, bimodal dispersions with colloidal interactions
复制标题
DOI:
10.1007/s003970100171
复制
发表时间:
2001-09-01
期刊:
影响因子:
2.3
通讯作者:
Willenbacher, N
中科院分区:
文献类型:
--
作者:
Dames, B;Morrison, BR;Willenbacher, N
The relationship between particle size distribution and viscosity of concentrated dispersions is of great industrial importance, since it is the key to get high solids dispersions or suspensions. The problem is treated here experimentally as well as theoretically for the special case of strongly interacting colloidal particles. An empirical model based on a generalized Quemada equation eta = (1 - phi/phi (max))(-epsilon) is used to describe eta as a function of volume fraction for mono- as well as multimodal dispersions. The pre-factor accounts for the shear rate dependence of eta and does not affect the shape of the eta vs phi curves. It is shown here for the first time that colloidal interactions do not show up in the maximum packing parameter and phi (max) can be calculated from the particle size distribution without further knowledge of the interactions among the suspended particles. On the other hand, the exponent epsilon is controlled by the interactions among the particles. Starting from a limiting value of 2 for non-interacting either colloidal or non-colloidal particles, epsilon generally increases strongly with decreasing particle size. For a given particle system it then can be expressed as a function of the number average particle diameter. As a consequence, the viscosity of bimodal dispersions varies not only with the size ratio of large to small particles, but also depends on the absolute particle size going through a minimum as the size ratio increases. Furthermore, the well-known viscosity minimum for bimodal dispersions with volumetric mixing ratios of around 30/70 of small to large particles is shown to vanish if colloidal interactions contribute significantly.