Torque measurements reveal large process differences between materials during high solid enzymatic hydrolysis of pretreated lignocellulose

BACKGROUND: A common trend in the research on 2(nd) generation bioethanol is the focus on intensifying the process and increasing the concentration of water insoluble solids (WIS) throughout the process. However, increasing the WIS content is not without problems. For example, the viscosity of pretreated lignocellulosic materials is known to increase drastically with increasing WIS content. Further, at elevated viscosities, problems arise related to poor mixing of the material, such as poor distribution of the enzymes and/or difficulties with temperature and pH control, which results in possible yield reduction. Achieving good mixing is unfortunately not without cost, since the power requirements needed to operate the impeller at high viscosities can be substantial. This highly important scale-up problem can easily be overlooked. RESULTS: In this work, we monitor the impeller torque (and hence power input) in a stirred tank reactor throughout high solid enzymatic hydrolysis (< 20% WIS) of steam-pretreated Arundo donax and spruce. Two different process modes were evaluated, where either the impeller speed or the impeller power input was kept constant. Results from hydrolysis experiments at a fixed impeller speed of 10 rpm show that a very rapid decrease in impeller torque is experienced during hydrolysis of pretreated arundo (i.e. it loses its fiber network strength), whereas the fiber strength is retained for a longer time within the spruce material. This translates into a relatively low, rather WIS independent, energy input for arundo whereas the stirring power demand for spruce is substantially larger and quite WIS dependent. By operating the impeller at a constant power input (instead of a constant impeller speed) it is shown that power input greatly affects the glucose yield of pretreated spruce whereas the hydrolysis of arundo seems unaffected. CONCLUSIONS: The results clearly highlight the large differences between the arundo and spruce materials, both in terms of needed energy input, and glucose yields. The impact of power input on glucose yield is furthermore shown to vary significantly between the materials, with spruce being very affected while arundo is not. These findings emphasize the need for substrate specific process solutions, where a short pre-hydrolysis (or viscosity reduction) might be favorable for arundo whereas fed-batch might be a better solution for spruce.