Monitoring Liquefaction Unit Operation in Dry-Grind Ethanol Process: Factors Affecting Hydrolysis and Methods for Analysis

Abstract

Ethanol production in the dry-grind ethanol industry converts starch in corn to ethanol using enzymes and yeast. However, yeast cannot consume starch directly; thus, starch has to first be converted into sugars with the help of enzymes. Liquefaction is the unit operation typically involved in cooking and hydrolysis by enzymes, which facilitates the conversion of starch into sugar. This study was conducted to determine effects of corn hybrids, enzyme types, initial enzyme dose, solids content, and operation time on the liquefaction operation. Chemical assays (Fehling and DNS assays) and HPLC measurements were compared for evaluating the quality of liquefied slurry at the end of the liquefaction operation. In addition, near-infrared spectroscopy (12,000 to 4,000 cm-1) was investigated as a method for monitoring liquefaction. Liquefaction was performed with three different enzymes: Liquozyme SC (Novozymes North America, Inc., Franklinton, N.C.), and Maxaliq ONE and STARGEN 001 (Danisco US, Inc., Rochester, N.Y.). Slurry was prepared by adding water to the corn flour to obtain 25% and 33% dry solids contents. Samples for the dextrose equivalent test were drawn at 40, 80, and 120 min after the start of the liquefaction process. High-pressure liquid chromatography (HPLC), dinitrosalicylic (DNS) assays, and Fehling assays were performed. Samples obtained from the two experiments were scanned on an FT-NIR spectrometer for building the calibrations. It was found that solids content was the most influential factor, followed by operating time. Other factors, including enzyme type, enzyme dose, and interaction between solids content and enzyme type, were also found to be significant. In general, it was found that least squares mean for Fehling assays, DNS assays, and HPLC measurements increased with time of operation, increased with enzyme dose, and increased with solids content. It was found that there was a strong relationship between the assay values and the total soluble sugar as measured from HPLC values. The best calibration model for NIR spectroscopy was found using a PLS1 model with no spectral treatments. Model selection was based on the lowest RMSEP (2.7% DE) and highest RPD value (2.1).