Vol. 85, No. 2, 2008 207 Laboratory Wet Milling of Corn: Milling Fraction Correlations and Variations Among Crop Years Amit Arora, 1 Yuxian Niu, 1 M. E. Tumbleson, 1 and Kent D. Rausch 1,2 ABSTRACT Cereal Chem. 85(2):207–210 Several coproducts result from fractionating corn in the wet-milling process. Because small changes in product composition and milling char- acteristics can have a major impact on coproduct yields and values, test- ing is done to anticipate final product yields. Using small sample size and controlled conditions, a laboratory wet-milling method proved to be a useful tool for wet milling and genetics industries. A wet-milling process (100-g batches) was used for data collection. Data collected during 11 years (1994–2004) were observed for samples used as benchmarks to verify process precision and accuracy and determine correlations among wet-milling yields. More than 400 milling tests were performed on benchmark samples. Data from benchmark samples also were pooled. Co- efficients of variation were low (<6%) for mean yields; year-to-year stan- dard deviations of benchmark sample yield means were homogenous and implied precision of the procedure. Some differences were detected in mean yields among years (P ≤ 0.05) for benchmark data due to combined effects of hybrid and environment. A negative correlation (r = –0.58) was observed between starch and gluten yield for pooled benchmark data. Four years (2002–2005) of milling data from commercially available hybrids were analyzed using the milling procedure. For pooled commer- cial data, the correlation between starch and fiber yield was (r = –0.80); correlation between starch and gluten was (r = –0.76). The U.S. wet-milling industry is a large processor of corn, us- ing 1.6 billion bushels (40.6 million tonnes) of corn annually (CRA 2006). Corn wet milling converts corn into a variety of coproducts including starch and starch-derived products (e.g., high-fructose corn syrup and ethanol), corn oil, corn gluten meal, and corn gluten feed. Genetics companies play a major role in producing hybrids with improved wet-milling characteristics. Milling evaluation of hybrids assists genetics companies and corn producers in choosing quality hybrids. The commercial wet- milling process is too large for determining milling quality of hybrids; laboratory-scale milling procedures have been used to evaluate milling characteristics (Watson et al 1951; Pelshenke and Lindemann 1954; Anderson 1963; Weller et al 1988; Steinke and Johnson 1991; Steinke et al 1991; Eckhoff et al 1993; Rausch et al 1993; Singh and Eckhoff 1995; Eckhoff et al 1996; Vignaux et al 2006). Industrial testing or pilot-plant-scale studies are more expensive, require larger quantities of sample, and are suitable only when relatively large amounts of coproducts are required for subsequent testing (Singh and Eckhoff 1996). Due to small sample size and controlled conditions, a labora- tory method proved to be a useful tool for the wet-milling and genetics industries. There are several laboratory methods that require samples of 300–1,500 g (Zipf et al 1950; Watson et al 1951; Anderson 1963; White et al 1990; Steinke and Johnson 1991; Eckhoff et al 1993). To fulfill the need of a process that used smaller amounts of corn, a 100-g laboratory wet-milling procedure was developed with reproducible results for product yields and a high degree of precision. Resulting data were repre- sentative of commercial-scale wet-milling operations and could be compared with large-scale milling methods (Eckhoff et al 1996). Recently, a 10-g laboratory method was developed to measure wet-milling yields of corn hybrids (Vignaux et al 2006). Small-scale milling procedures have been used to measure wet milling changes due to hybrid (Pelshenke and Lindemann 1954; White et al 1990; Zehr et al 1995; Rausch et al 1999; Singh et al 2001). Efforts have been made to correlate starch recovery with various quality factors to quantitatively relate corn quality to wet- milling characteristics (Freeman 1973; Vojnovich et al 1975). Previous research has measured correlations of physical proper- ties and wet-milling yields (Vojnovich et al 1975; Brown et al 1979; Wight 1981; Weller et al 1988). These studies revealed that physical properties were not highly correlated to milling yields. Correlations were established among starch recovery and quanti- tative measures of assorted quality factors of corn hybrids, but the best model had relatively low correlation (R 2 = 0.60) (Weller et al 1988). The most accurate method of predicting industrial wet- milling results was to process the corn in a milling procedure. There are few published data comparing wet-milling yields for a broad range of hybrids and for extended time periods. We used a 100-g wet-milling laboratory procedure to determine yield varia- tions due to hybrid, growing location, and crop year. MATERIALS AND METHODS Wet-Milling Procedure A wet-milling procedure described by Eckhoff et al (1996) was used to obtain yields of steepwater, germ, fiber, starch, and gluten. Each corn sample (100 g) was placed in 180 mL of steep solution; 2,000 ppm of sulfur dioxide and 0.55% w/w lactic acid were added. Steeping was done at 52°C for 24 hr. Fiber yields from pericarp and endosperm sources were combined and reported as a single quantity. Yields were calculated as a percentage (db) of original corn solids using standard methods to determine solids contents (AACC International 2000). Benchmark Data Benchmark samples were from hybrids grown at the Agricul- tural and Biological Engineering Research Farm at the University of Illinois (1994–2004) and were used to verify precision of the laboratory process. A hybrid of commercially available yellow dent corn was used each year as a reference (benchmark) sample for the laboratory wet-milling process (Pioneer Hi-Bred Interna- tional, Johnston, IA). Hybrids (FR1064 × LH59, 3394, 33A13, and 33A14) were selected based on their use in commercial corn production; new benchmarks were chosen as new hybrids came into mainstream production. Therefore, variations year-to-year were due to hybrid and environmental effects. Samples were milled approximately weekly throughout each year (Table I). Client Data Client data were those data generated by samples submitted from genetics companies and represented a relatively broad range of wet-milling characteristics compared with benchmark hybrids. 1 Agricultural and Biological Engineering, University of Illinois at Urbana-Cham- paign, Urbana, IL 61801. 2 Corresponding author. Phone: 217-265-0697. Fax: 217-244-0323. E-mail address: krausch@uiuc.edu doi:10.1094/ CCHEM-85-2-0207 © 2008 AACC International, Inc.