Understanding and predicting alcohol yield from wheat (PhD)

Abstract

Bioethanol is a promising renewable biofuel and wheat is currently the main candidate as the feedstock for its production in the UK context. In general, the end-use quality determination of wheat in terms of alcohol yield has not been thoroughly investigated; breeders and growers are unclear about which characteristics of wheat are relevant to alcohol production and how to breed and cultivate for these characteristics.

This work focused on understanding and predicting the alcohol yield from wheat according to its physical, physicochemical and chemical characteristics. The research used the wheat samples of the GREEN Grain project, which consisted of a range of wheat varieties, agronomic regimes and growing sites from the four harvests years 2005-2008.

High alcohol-yielding wheats typically have high starch, mealiness and albumin+globulin fraction, and also low protein, gliadin fraction and hardness. They also have larger and more spherical kernels. The application of N fertiliser increases the protein components and yields smaller and more elongated kernels. High alcohol-yielding varieties tend to be softer with lower protein and larger and more spherical kernels. Alcohol yield could be predicted based on SKCS-reported values of hardness and diameter plus protein.

It is frequently hypothesised that larger and more rounded kernels produce more alcohol because they have a smaller relative amount of the unfermentable outer layers. To test this hypothesis, the pericarp thicknesses and crease characteristics of the wheat samples were measured. It appears that larger kernels tend to have thicker pericarp which largely eliminates the hypothesised benefit.

The Perten Single Kernel Characterisation System (SKCS) fundamental crushing force data (rather than the reported hardness values) were further analysed in an effort to improve alcohol yield predictions. It was found that the initial peak in the averaged Crush Response Profiles (aCRP) does not occur because of breakage of the "shell" (i.e. the bran layers), but is principally influenced by the crease; a deep crease gives kernels that are easier to break and that therefore exhibit a lower peak force. The aCRP parameters could improve the alcohol yield predictability of the GREEN Grain wheats to an R2 of about 82% with a standard error of the regression of 6.3 l/dry ton.