Marketing and Delivery of Quality Grains and Bioprocess Coproducts

NON-TECHNICAL SUMMARY: The fermentation process for ethanol production typically runs for about 56 or more hours. During this time yeast uses glucose which is converted to ethanol, but simultaeneously glucoamylase enzymes are making more glucose available as saccarification progresses. Potential exists for lactic acid or acetic acid to rise if bacterial infection occurs. Rapid low cost monitoring of these factors could alert plant operators of lowered yeast growth and ethanol production rates and will greatly enhance production efficiency. <P>PROGRESS: 2007/01 TO 2007/12<BR>
OUTPUTS: The objective of this study was to determine the effect of reaction time, initial enzyme dose and corn hybrid on dextrose equivalent (DE) values at the end of liquefaction; and to investigate the feasibility of developing calibrations for predicting DE values using a Fourier-Transform Near-Infrared (FT-NIR) Perkin Elmer spectrometer. Calibration models using PCA and PLS were investigated. An analysis of variance was performed using a model where enzyme dose (25, 50, 100 and 200 micro-L)/50g slurry, for four hybrids (waxy, hi-amylose, and two yellow dents), and time (1 and 2 hours from the start of liquefaction) were the independent class variables. The interaction terms were Enzyme x Hybrid, Hybrid x Time, and Enzyme x Time, while DE values were treated as the dependent variable. The model had an F- value of 28.08 with 27 degrees of freedom, and R2 values of 0.78 indicating that the independent variables had a statistically significant influence on the dependent variable. The individual variables (enzyme dose, hybrid, time, and interaction) were statistically significant (0.05 level) based on F-values and p-values obtained for each term. Least Significant Difference tests (t-tests) were performed to determine significant differences between levels for enzyme dose, hybrid, and time. Least significance difference for DE among Hybrids was 0.79. Of the four hybrids, DE values for yellow dent corn were highest with means varying from 12.8 to 13.9%. Waxy corn mean was 11.2% DE and high amylose corn was lowest with 9.7% DE. For enzyme dose, least significance difference was 0.71. All the enzyme dose levels were significantly different from each other. Mean levels of DE values for enzyme doses of 25, 50, 100 and 200 micro-L/50 g were 8.5, 11.1, 12.9, and 16.4 %, respectively. Time for 1 hour and 2 hours after the start of liquefaction were both statistically different with mean DE levels of 11.8% and 12.6% for 1 and 2 hours, respectively. All of the interaction terms were found to be statistically significant. About 230 samples were scanned in duplicate in the range of 12,000 to 4,000 cm-1 (833 to 2500 nm) with 2 cm-1 intervals. Most of the variations in the spectra were in the region 7,500 to 5,500 cm-1 (1,333 to 1,818 nm) and 5,200 to 4,000 cm-1 (1,923 to 2,500 nm). The Unscrambler (CAMO Software AS) software was used for regression analysis, and PLS1 and PCR methods were used for building the calibration. The Jackknife method with full cross-validation was used for elimination of the non-significant variables. In the Jackknife method, standard errors for the coefficients were calculated, and wavelengths with standard errors that crossed the zero lines were removed from the calibration. The model with lowest RMSEP (1.5% DE) was found by using the raw spectra with PLS1. Its offset was smallest with 1.4, slope (0.89) was closest to 1.0 over the DE range from 5.4 to 22.4 %, and the PLS1 model had a RPD value of 2.7. <BR>PARTICIPANTS: S.S. Rathore is a Ph.D. graduate student at the University of Illinois at Urbana Champaign. He was advised by Dr. M.R. Paulsen and Dr. V. Singh. This project provided an opportunity for research training for Mr. Rathore. <BR>TARGET AUDIENCES: The target audience is the dry-grind ethanol industry.
<P>
IMPACT: 2007/01 TO 2007/12<BR>
U.S. ethanol production capacity was 4.4 billion gallons per year in February of 2006 and as of October 2007 the Renewable Fuels Association (RFA) estimates that 131 plants were producing 7.02 billion gallons of ethanol annually. An additional 6.49 billion gallons of ethanol production is expected by the end of 2009 based on 72 more new plants and 10 plant expansions already underway (RFA http://www.ethanolrfa.org/industry/locations/ ). The ability to rapidly measure and monitor liquefaction and fermentation processes and variability in DDGS quality is needed.

<P>
PROGRESS: 2006/01/01 TO 2006/12/31<BR>
The objective of this study was to investigate the feasibility of developing calibrations for ethanol yield based on ground corn samples using a Fourier-Transform Near Infrared (FT-NIR) Perkin Elmer Model 6750 Spectrum One NTS spectrometer. Calibration models using PCA, PLS, and discriminant PLS were investigated. Fourier-Transform NIR calibrations were made based on HPLC tests of fermented corn slurries. The wavenumbers (wavelengths) of 5170 cm-1 (1934 nm), 4764 cm-1 (2100 nm), 4518 cm-1 (2210 nm), 5782 cm-1 (1729 nm), and 5938 cm-1 (1684 nm) were found to be important for the classification of corn hybrids for ethanol values. The Discriminant PLS model with five factors had a correlation coefficient of 0.82 and RMSEP of 2.04. One principal component was able to explain 60% of the variation in ethanol among samples. Discriminate PLS analysis for each group had a population overlap with adjacent groups; but the highest ethanol group could be separated from the lowest ethanol group. The PLS model with five factors had a correlation coefficient of 0.67 and RMSEP of 0.6 % v/v. The calibration model using the PCR method had a higher RMSEP as compared to that of PLS1; thus, it provided no advantage in terms of prediction. The model using the PLS1 method usually gave a lower RMSEP than those built using PCR. However, with more samples at the upper and lower range of ethanol values, there is a possibility of obtaining improved results. Fermentability of corn not only depends on the starch constituents, but also on the process parameters such as pH and temperature. Factors such as variation in yeast nutrients, and activity of saccharification enzymes can also greatly affect yield of ethanol.
<P>
IMPACT: 2006/01/01 TO 2006/12/31<BR>
U.S. ethanol production capacity was 4.4 billion gallons per year in February of 2006 and is expected to reach 7 billion gallons by the end of 2007, creating a huge surplus of DDGS. Value of DDGS can be improved by modified processes that reduce fiber and increase protein, but rapid measurement and monitoring of fermentation processes and variability in DDGS quality is lacking. Near-infrared spectroscopy provides a rapid technology for on-line quality monitoring, but robust calibrations for constituents in fermentation broth and DDGS are first needed.
<P>
PROGRESS: 2005/01/01 TO 2005/12/31<BR>
The ability to monitor and modify the fermentation process used for dry-grind ethanol production is important. The objective of this research is to determine suitable Near-Infrared (NIR) wavelength regions to identify ethanol, total soluble sugars, glycerol, and organic acids in the fermentation broth for a corn dry grind process. Effects of temperature on spectral absorption were identified and related to loadings of principal component scores. The wavenumber region between 12,000 to 6,400 cm-1 had low correlation coefficients for a 0.5-mm short pathlength transflectance accessory. Wavenumbers having high correlation coefficients for ethanol were found in the region between 4,712 to 4,200 cm-1, while that for total sugars were from 5,000 to 4,286 cm-1. High RPD (ratio of laboratory standard deviation to the RMSEP) and low RMSEP (root mean standard error of prediction) values for both ethanol and total soluble sugars validated the suitability of the wavelength regions found. Lactic acid and acetic acid were present at very low levels and did not appear to separate from the other fermentation constituents for the given pathlength and concentrations observed. Distillers dried grains with solubles (DDGS) is the primary co-product of dry grind fermentation of corn. DDGS has been mainly used as a ruminant animal feed, but its use for non-ruminants is important. However, significant variation in DDGS composition (protein, fat, starch, and fiber) occurs. Certain modified dry-grind corn processes, such as quick germ, quick fiber, and enzymatic milling reduce fiber content of DDGS from 11% down to 2% and increase protein from 28% up to 58%. Rapid techniques such as Near Infrared Reflectance (NIR) for measurement of nutrient levels in the DDGS could be useful for analytical and quality control. The objective of this work was to evaluate a sample preparation method based on enriching protein in DDGS by removal of fiber using elutriation, a separation process using the upward flow of air. Regression methods: partial least squares (PLS1 and PLS2), and principal components (PCR), were evaluated. Best results for most constituents were obtained with the PLS1 method. Cross validation results predicted crude protein with a RMSEP (root mean square error for prediction) of 1.2 for a range of 10.4 to 41.5%; Neutral Detergent Fiber (NDF) with a RMSEP of 2.7 for a range of 21 to 69%; crude fiber with a RMSEP of 1.7 for a range of 4.1 to 16.8%; crude fat with a RMSEP of 1.1 for a range of 4.4 to 27.7%; acid detergent fiber (ADF) with a RMSEP 1.9 for a range of 10.4 to 23.2%; ash content with a RMSEP of 0.31 for a range of 2.6 to 5.5%; and total digestible nutrients with a RMSEP of 1.7 for a range of 71.1 to 93.5%.
<P>
IMPACT: 2005/01/01 TO 2005/12/31<BR>
Current U.S. ethanol production of 3.9 billion gallons per year is expected to soon double, creating a huge surplus of DDGS (distillers dried grains with solubles). Value of DDGS can be improved by modified processes that reduce fiber and increase protein, but rapid measurement and monitoring of fermentation processes and variability in DDGS quality is lacking. Near-infrared spectroscopy provides a rapid technology for on-line quality monitoring, but robust calibrations for constituents in fermentation broth and DDGS are first needed.

<P>
PROGRESS: 2004/01/01 TO 2004/12/31<BR>
A calibration for extractable starch in maize was developed for the Foss Infratec 1229 near infrared transmission (NIT) grain analyzer. The calibration was based on over 2,267 samples collected over five crop years. Extractable starch in maize can be predicted using the Infratec 1229 NIT spectrophotometer with a standard error of prediction (SEP) of 1.24, coefficient of determination (R2) of 0.79, and ratio of laboratory standard deviation to SEP (RPD) of 2.15 for a validation set using one elimination pass. With no elimination passes, 8 of 389 samples were not predicted well, giving an SEP of 1.34, R2 of 0.80 and a RPD of 2.2. Regression coefficient peaks corresponded to some of the wavelengths known for starch-cellulose, water, and negative protein and oil absorption bands. A calibration for a Fourier-Transform Near-infrared Reflectance (FT-NIR) spectrometer was developed for predicting soybean oil properties for biodiesel fuel. Biodiesel fuel depends on fatty acid composition of the soybean oil. Vegetable oils (soybean oil) are converted into methyl esters by trans-esterification prior to use as biodiesel. Four types of biodiesel, including two types of soybean methyl ester, yellow grease methyl ester, and genetically modified soy methyl ester differing in fatty acid compositions were blended with biodiesel levels from 0-100%. These samples were scanned at room temperature (20-22C) on a FT-NIR machine using a transmission accessory cell with a 1-cm pathlength. Three regression methods: partial least square regression (PLS1), principal component regression (PCR), and multiple linear regression (MLR) were evaluated as calibration models for predicting the percentage blend, specific gravity and viscosity of biodiesel fuel. Principal Component Analysis (PCA) of the NIR absorption data revealed two major variations were captured in the first two principal components, PC1 variation was due to the varying amount of blend level in diesel and PC2 was due to the presence of differences in the biodiesel source. Two principal components accounted for (PC1 89%, PC2 8%) 97% of the total variation in the NIR data.
<P>
IMPACT: 2004/01/01 TO 2004/12/31<BR>
The near-infrared transmission calibration for extractable starch is the primary means used by wet millers, seed companies, and the grain trade for determining the percentage of extractable starch in corn hybrids that are used in the high extractable starch market. This potentially affects about 19% of the corn grown in the U.S. today.
<P>

PROGRESS: 2003/01/01 TO 2003/12/31<BR>
In 2003, research continued on using near-infrared spectroscopy to measure extractable starch in corn. During the 1997 to 2002 crop years, over 2,600 samples of corn were scanned on the Foss Infratec 1229 near-infrared transmittance (NIT) unit. Extractable starch was predicted using the Infratec 1229 NIT spectrophotometer with a standard error of prediction (SEP) of 1.34, R2 of 0.80 and a RPD of 2.2. The extractable starch calibration has been licensed to Foss North America. A study was conducted to compare moisture meter readings to the 103 degree C 72-hr air oven reference method for naturally wet corn during the fall of 2002. A total of 178 samples were collected from 12 counties in Illinois. The oven moistures of the samples ranged from 12.6 to 39.6% wet basis. Each sample was tested by the Illinois Department of Agriculture in triplicate in air ovens and in the Dickey-john GAC 2100, GAC 2000, and GAC II; Motomco 919, 919E, and 919ES; Steinlite SL 95, SB 900, SS 250, RC, RCT; and Burrows 700 moisture meters. Compared to the oven, most of the meters performed within the acceptable limits up to approximately 25% moisture content on combine and hand-shelled corn. Above 25% moisture, most of the meters read lower than the air oven on hand-shelled corn. Based on only one crop year with predominately hand-shelled samples no recommendations for changes in moisture meter calibrations could be made.
<P>
IMPACT: 2003/01/01 TO 2003/12/31<BR>
The near-infrared transmission calibration for extractable starch is the primary means used by wet millers, seed companies, and the grain trade for determining the percentage of extractable starch in corn hybrids that are used in the high extractable starch market. This potentially affects about 19% of the corn grown in the U.S. today.

<P>
PROGRESS: 2002/01/01 TO 2002/12/31<BR>
A calibration for extractable starch in corn was developed for the Foss Infratec 1229 Near Infrared Transmission (NIT) grain analyzer. The calibration was based on 2,267 samples collected over five crop years (1997 through 2001). Extractable starch in corn was predicted using the Infratec 1229 NIT spectrophotometer with a standard error of prediction (SEP) of 1.24, R2 of 0.79, and RPD of 2.15 for a validation set using one elimination pass. With no elimination passes, 8 of 389 samples were not predicted well, giving an SEP of 1.34, R2 of 0.80 and a RPD of 2.2. Beta coefficient peaks corresponded to some of the wavelengths known for starch-cellulose, water, and negative protein and oil absorption bands. Work continues to commercially test the calibration and to add more diverse genotypes, and more drying treatments. Calibrations for other instruments are also being developed. Global positioning systems provide precise field locations for fertilizer and herbicide applications as well as moisture and yield monitoring. To determine variability in corn protein, oil, starch, and extractable starch, near-infrared transmission (NIT) was used on samples from 640 sub plots that were planted with two varieties at four population levels, five nitrogen levels, and two nitrogen application methods. A GLM analyses of variance indicated NIT protein content increased significantly with increases in nitrogen rate and application method. Protein ranged from 5.7 to 11.0% over all plots and treatments. The highest protein content occurred on plots receiving 202 kg N/ha; while the lowest occurred on plots receiving no nitrogen. Oil percentages did not vary significantly with nitrogen rate and they ranged from 2.2 to 4.3%. Starch content and extractable starch decreased significantly as nitrogen rate increased. Extractable starch ranged from 63.4 to 72.1%. The highest extractable starch level of 72.1% occurred on the plots with no nitrogen application; while the lowest extractable starch occurred on plots receiving 202 kg N/ha. Starch content ranged from 72.0 to 76.1%. Yield increased from a mean of 10.5 tonnes/ha at the 0 nitrogen level to 11.5 tonnes/ha at the 202 kg N/ha level. Extractable starch had a negative correlation (R2 = -0.76) with protein and a positive correlation (R2 = 0.65) with starch content.
<P>
IMPACT: 2002/01/01 TO 2002/12/31<BR>
The near-infrared transmission calibration for extractable starch is the primary means used by wet millers, seed companies, and the grain trade for determining the percentage of extractable starch in corn hybrids that are used in the high extractable starch market. This potentially affects about 18% of the corn grown in the U.S. today.
<P>

PROGRESS: 2001/01/01 TO 2001/12/31<BR>
A population of 1,042 samples were scanned using NIRSystems 6500 and 5000 spectrophotometers to develop a calibration for amylose content in corn. The NIRSystems 6500 calibration using the 3,5,5,1 math treatment with unground samples predicted its validation set with an R2 of 0.83, SEP of 7.6 and an RPD of 2.2. The NIRSystems 5000 calibration using the same math treatment with ground samples predicted its validation set with an R2 of 0.73, SEP of 3.0 and an RPD of 1.9. About 1,705 samples from 1997 through 2000 crop years were scanned on an Infratec 1229 NIT instrument. Predictions for extractable starch were made with a SEP of about 1.3 percentage points, and a R2 of 0.86.
<P>
IMPACT: 2001/01/01 TO 2001/12/31<BR>
The NIR calibration for amylose content in corn reduces the time needed to check amylose content. The NIT calibration for extractable starch helps wet millers and ethanol processors select corn ideally suited for use and it helps indicate if starch recovery has been reduced due to excessively high temperatures during drying.

<P>

PROGRESS: 2000/01/01 TO 2000/12/31<BR>
Four types of corn - high oil, high amylose, waxy, and dent were scanned using the Foss Infratec 1229 Grain Analyzer. Extractable starch was determined using the 100-g laboratory wet milling method. Regression analyses were conducted using The Unscrambler. A model was developed to predict extractable starch in corn and compared to InfraSoft International\'s WINISI software. Root mean square error of calibration (RMSEC) and root mean square error of prediction (RMSEP) were calculated to evaluate the model. This study enables a better understanding of the calibration process, and provides a means for testing of extractable corn starch using NIT analyzers.
<P>
IMPACT: 2000/01/01 TO 2000/12/31<BR>
Extractable starch is a measure of how much starch can be obtained from corn for wet milling or ethanol production purposes. The Unscrambler is a software package which can be used to predict extractable starch using near infrared spectroscopy.

<P>
PROGRESS: 1999/01/01 TO 1999/12/31<BR>
Laboratory-dried corn had a 0.9 to 16.0 percentage point range in extractable starch reduction due to drying at 90 to 100 C rather than ambient temperatures. Corn obtained from commercial crossflow dryers had a maximum reduction of 3.1 percentage points for corn dried at 110 C. The crossflow configuration did not reduce extractable starch as much as the laboratory dryers. Each percentage point gain in extractable starch is estimated to be worth a minimum of 4 to 6 cents per bushel. The NIR and NIT calibrations may be used as a rapid method for checking dried corn for extractable starch.
<P>
IMPACT: 1999/01/01 TO 1999/12/31<BR>
A NIR or NIT calibration for extractable starch can be used to channel corn with high extractable starch to enhanced-value markets.
<P>

PROGRESS: 1997/10/01 TO 1998/09/30<BR>
Hard, medium, and soft endosperm corn hybrids were grown on the Agricultural Engineering Research Farm in 1998. They were combine harvested at five moisture levels, and tested for 100-g extractable starch and scanned from 850 to 1050 nm on an Infratec 1229 NIT unit. Over 500 additional samples were scanned, covering a wide range of drying conditions, varieties and growing locations. Accompanying 100-g extractable starch tests have enabled development of a NIT calibration for extractable starch on the Infratec 1229.
<P>
PROGRESS: 1995/10 TO 1996/09<BR>
The degree of whiteness of corn kernels was measured using machine vision. RGB images of kernels were converted to YCrCb (luminance, chrominance red, and chrominance blue). Kernel germ regions were deleted from images to obtain comparable results regardless of whether the germ was up or down. Whiteness index values averaged between 18 to 80 as corn samples varied from deep yellow/orange to bright white, with above 54 indicating white corn. After training, human-inspected samples were ranked in color from yellow to white. Machine vision inspections of the samples maintained the same order of ranking as the human-inspected samples, and were more consistent than human inspectors.

<P>
PROGRESS: 1994/10 TO 1995/09<BR>
Corn quality characteristics desired by Japanese millers and feed processors were investigated. Wet millers prefer corn that has not been dried at temperatures sufficiently high to cause protein denaturation or starch gelatinization. Stress crack percentages less than 20% are an indirect test for acceptable corn meeting the heat restrictions. Dry millers prefer hard endosperm corn in the true density range of 1.25 to 1.28 g/cm3, in order to obtain a high yield of large flaking grits. High oil corn may be a growth area for feed processors if agronomic yields can be increased significantly beyond the normal 85 to 95% of normal dent corn yields.
<P>

PROGRESS: 1993/10 TO 1994/09<BR>
Variation in moisture content exits between single kernels of corn. Even on the same ear, moistures can vary by as much as 6 percentage points. When corn lots of different moistures are handled, moisture blending naturally occurs. In this study corn at 29% and 10% wet base moisture were blended and stored at 25deg.C and 74% relative humidity in sealed containers. At the end of 1, 3 and 7 days after blending, individual kernel moistures ranged from 13 - 24, 15 - 21, and 16.5 - 19%, respectively. After 7 days, the equilibration moisture range no longer narrowed.

<P>
PROGRESS: 1992/10 TO 1993/09<BR>
A machine vision system capable of performing identification of corn kernel profiles for shape and breakage was developed for automated grain quality inspection. The profile of a corn kernel was resampled into a sequence of one-dimensional digital signals from its binary image. Shape parameters were selected by analyzing the kernel profile and were sent to a machine learning algorithm to train for a shape membership function of broken versus whole kernels. The inspection algorithm provided a successful classification of 93% and 91% for whole and broken kernels, respectively.

<P>
PROGRESS: 1991/10 TO 1992/09<BR>
A machine vision system capable of performing identification of corn kernel profiles for shape and breakage was developed for automated grain quality inspection. The profile of a corn kernel was resampled into a sequence of one-dimensional digital signals from its binary image. Shape parameters were selected by analyzing the kernel profile and were sent to a machine learning algorithm to train for a shape membership function of broken versus whole kernels. The inspection algorithm provided a successful classification of 934 and 91% for whole and broken kernels, respectively.

<P>

PROGRESS: 1990/10 TO 1991/09<BR>
Machine vision was used to detect breakage and pericarp damage of corn kernels. Corn kernels were dyed using 0.1% Brilliant Blue R-250 dye. Exposed starch and damaged areas readily absorb the blue dye. HSI, (hue, saturation, and intensity) images of individual corn kernels were taken. The hue component was processed and thresholded to isolate the blue-dyed regions from the whole kernel using an on-board real-time feature extraction algorithm. An AI algorithm (ID3 machine learning algorithm) was used to find the damaged areas (class membership functions). A decision tree classifier was built to classify the corn kernels into three categories of pericarp damage: severe, minor, and negligible damage. The algorithm required about one second to run.

<P>
PROGRESS: 1989/10 TO 1990/09<BR>
Present US Corn Grading Standards do not discriminate between broken corn (BC) and foreign material (FM) in corn. The Grain Quality Improvement Act of 1986 provides separate definitions of BC and FM. The fraction of a corn sample passing through a 2.38 mm (6/64 in). sieve was defined as being FM. This study involves collecting and analyzing corn samples from each place in the market channel beginning with country elevators through export elevators to determine percentages of material which will be classified as FM and as BC under this new definition. Data for predicting actual cleaning efficiency relative to laboratory cleaning was collected. Early results indicate that about 20 % of a sample containing BC and FM is actually FM under the above definition.
<P>
PROGRESS: 1988/10 TO 1989/09<BR>
The effect of corn harvest moisture content (34, 30, 24, 20%) and different drying temperatures (ambient 40, 50, 60, 80 and 120C) were determined on ethanol soluble protein, albumin & globulin proteins, degree of starch gelatinization, starch damage observation with the scanning electronic microscope, lab-scale wet millability, kernel density, and other physical quality measurements for FR1141 x Va22 (hard-endosperm), FR1130 x FR1193 (medium hard), and FR27 x FRMo17 (soft) corn genotypes. Starch damage increased as kernel temperatures, harvest moisture, and heating time increased. Starch granules in corneous endosperm and in hard genotypes appeared to be more susceptible to damage than those in floury endosperm. High termperature drying was highly detrimental to starch recovery.

<P>
PROGRESS: 1987/01 TO 1988/09<BR>
Effects of temperature, airflow, and initial moisture content on drying time andon germination percentages for corn genotypes were investigated for laboratory-scale dryers and for a full-size seedcorn drying plant. Software was developed to monitor temperatures, relative humidities, and static pressures of input and exhaust air for prediction of moisture removal rates. A laboratory wet-milling process was developed for use in determining starch recovery of yellow dent corn. The effects of harvest moisture and drying air temperature on starch recovery were investigated for four hybrids ranging from hard to soft endosperm. A split-plot experimental design was used to evaluate the effects. Starch recovery was not significantly different among the hybrids; however, it did decrease as both harvest moisture and drying air temperature increased. Starch recovery was correlated with quantitative measures of several quality factors. Regression analysis indicated starch recovery was a function starch content, test weight, and ethanolsoluble protein.
<P>
PROGRESS: 1986/10 TO 1987/09<BR>
Two ocean vessel shipments of 15.0% moisture corn from New Orleans to Japan had starting levels of broken corn and foreign material (BCFM) below 4.0%. After handling and dropping into the ship, BCFM ranged from 4.9 to 7.1%. After unloading with pneumatic unloaders into barges at Japan, BCFM averaged 7.7 to 9.3%. Thus, approximately 2.2 to 2.8 percentage points of additional BCFM may be contributed to vessel unloading and conveying equipment at the destination port. The effects of corn variety, harvest moisture (18, 24, and 30%), and drying air temperature (22 to 93C) on twelve quality factors related to wet millability of corn were investigated using a split-split plot experimental design. Changes in protein conformation due to increased drying temperature were initiated between 49 and 71C. Increased drying temperature generally decreased kernel density, germination, equilibrium moisture content, and ethanol soluble protein, and increased near infrared absorbance. Thus, increased drying temperature above 49C was detrimental to recovery of starch for wet milling.
<P>
PROGRESS: 1986/01 TO 1986/09<BR>
A vessel of 57,000 metric tons of corn was shipped from a New Orleans port elevator to Japan during the Spring of 1986. Thermcouple temperature sensors were placed at 38 locations in two holds of the vessel to monitor initial and final grain temperatures. Temperatures were measured every two hours automatically by a battery powered recorder throughout the 32 day shipment. In the locations measured, initial corn temperature averaged about 13 degres C, with some temperatures as low as 8 degrees C. Vessel wall temperatures fluctuated with ocean water temperatures for the portion below the draft line. Corn located beyond 1 m from the vessel walls, bottom, or upper surface changed negligibly in temperature during the voyage. The only exception to this finding was in a few isolated locations where microorganisms found sufficient moisture to grow rapidly and produce heat. In one location, microorganisms heated the grain as high as 31 degrees C. Air temperatures on the grain surface but below the hatch covers reached 30 degrees C during passage through the Panama Canal, causing corn at the 0.3 m depth to warm to a high of 22 degrees C. As the vessel moved into cooler water, air temperature below the hatch cover dropped at a low of 13 degrees C. Grain at the 0.3 m depth arrived at 16 degrees C, while grain at the 3 m depth arrived unchanged in temperature at 13 degrees C. To assure safe shipment, maximum corn moisture must be specified below 15.
<P>
PROGRESS: 1985/01 TO 1985/12<BR>
Corn transported by ocean vessel from a Gulf port elevator to Japan increased inbroken corn and foreign material from an average of 3.4 percent to 4.9 percent. Barges and trucks loaded from the vessel in Japan averaged 7.7 and 6.9 percent, respectively. Temperature sensors installed in one half of the vessel indicated only small fluctuations in temperature along the sides and bottom of the hold, but significant increases from 23C to as high as 47C were measured in the upper one meter of corn. The corn was shipped at about 15 percent moisture content, wet basis.
<P>PROGRESS: 1984/01 TO 1984/12<BR>
Argentine corn was found to be high in density. Kernel densities averaged 1.29 g/cm·. Bulk density (test weight) was also high (765-785 kg/m·). Floaters tests in a 1.275 sp. gr. solution indicated Argentine corn is very thich in vitreous endosperm. Stress cracks ranged from 43 to 55% on artificially dried corn and 4 to 24% on corn dried with little or no heat added. One hundred-kernel dry weights average 23.4 to 24.8 g. Whole kernel percentages were high (93-95%). In a commerical dry mill, yields of large flaking grits were significantly increased by selecting corn with low breakage susceptibility and high bulk densities. The breakage susceptibility test also correlated significantly with stress crack percentages, indicating that highly stress-cracked corn is not desirable for dry milling. Based on 1982 crop year corn.

<P>
PROGRESS: 1983/01 TO 1983/12<BR>
Initial selection for breakage susceptibility and kernel density of corn delivered to a dry milling plant can improve prime product yield of large flaking grits. Four tests were performed using corn segregated into categories of test weight above and below 732 kg/m3 and either high or low breakage susceptibility using a Wisconsin-type 1800 rpm impactor. Tests on over 4000 truckloads provided a breakage susceptibility prediction equation of the form y = a ebx, where y = breakage susceptibility using a 6.35-mm sieve, %; x = wet base moisture content, %; and a, b = regression coefficients. Composite samples collected every 30 minutes during the four 24-hour milling runs were analyzed for stress cracks, 100-kernel weights, true density, test weight, floaters in a 1.275 sp. gr. solution, breakage susceptibility, fine material and moisture content. True density and test weight were positively correlated; true density and test weight were negatively correlated with floaters; and stress cracks were positively correlated with breakage susceptibility.

<P>
PROGRESS: 1982/01 TO 1982/12<BR>
Tests for kernel density and breakage susceptibility were performed on commonly grown corn genotypes to relate the measurement of initial quality factors to end-use value. As kernel density increases there is an apparent increase in kernel hardness or the amount of horny endosperm relative to floury endosperm. Kernel density increases as test weight increases; and it decreases as percentages of floaters in a 1.275 specific gravity solution increases. Kernel density varies significantly among corn genotypes. The increased desirability of dense-kernel genotypes for dry milling is being investigated. Breakage susceptibility, which predicts breakage likely to occur during a specific handling operation, varies significantly among corn genotypes also. Tests are underway for predicting dry milling performance using breakage susceptibility tests. Comparisons of the new Wisconsin breakage susceptibility tester to the Stein CK-2M tester and to the Illinois breakage susceptibility tester are being made as a function of corn moisture content.
<P>
PROGRESS: 1981/01 TO 1981/12<BR>
The objective of this research was to determine the accuracy of corn moisture meters compared to the 103~ C 72-hour USDA method for determining moisture. On corn samples tested with Illinois calibrations during the Fall of 1980, moisture meter error ranged from +0.6 percentage points for low moisture corn to -3.0 percentage points for high moisture corn. Revised calibrations that went into effect by moisture meter manufacturers on August 1, 1981 are being compared for 1981 crop corn with the air-oven moisture determination method. Corn samples from 14 states other than Illinois are also being tested.

<P>
PROGRESS: 1981/01 TO 1981/12<BR>
Corn genotype of (FR4AXFR4C)XMo17 and those with Mo17 as a female parent had lower breakage susceptibility than the popular B73XM017 genotype. Breakage susceptibility was determined by centrifugally impacting kernels at an impactor peripheral velocity of 23.4 m/sec. Breakage susceptibility was determined by sieving the sample before and after impact with a 4.76mm round-hole sieve. Breakage susceptibility for corn dried at air temperatures of 60 degrees was 2 to 4 times greater than that for the same genotype dried at 24 degrees C. Kernel density varied significantly among genotypes. Correlations among density (which is often referred to as kernel hardness or softness), breakage susceptibility, and dry milling yield of large grits are being investigated. The desired objective is to develop quick mechanical tests that will enable corn varietal selection for reduced breakage during handling and to identify high quality corn lots that will have exceptionally low breakage when handled.
<P>
PROGRESS: 1980/01 TO 1980/12<BR>
Significant differences in the breakage susceptibility of corn samples, among varieties commonly grown in Illinois, were found when individual kernels were propelled against a rigid steel cylinder wall at about 26 m/sec. The corn samples were tested at controlled moisture levels and were sieved before and after impact treatment to determine breakage. Repeated breakage susceptibility tests are being performed for corn equilibrated to 14% moisture content. Correlations between breakage susceptibility and dry milling quality are being determined. Tests for dry milling quality include kernel density and specific gravity measurements, and yield of large flaking grits on laboratory-scale dry milling equipment. In addition, corn samples were collected from eleven vessels by mechanical diverter samplers at U.S. export terminal elevators and at an import elevator in England. Breakage susceptibility, measured with a centrifugal impacter and with 2 and 4-minute Stein breakage tests, was compared to the actual change in broken corn and foreign material (BCFM) between origin and destination. Average BCFM percentages increased by 1.7 to 6.9 percentage points. Good correlations were obtained between breakage susceptibility for vessels loaded from the same origin terminal elevator and actual BCFM increases, after adjustments were made for moisture content.
<P>
PROGRESS: 1979/01 TO 1979/12<BR>
To obtain a measure of corn kernel susceptibility to breakage in handling, corn inbreds and hybrids commonly grown in Illinois were impacted at controlled velocities in a centrifugal impacter. Breakage is proportional to kernel velocity at impact and each sample is evaluated by sieving before and after impact. Significant differences in breakage were found among the inbreds and hybrids and are being explored through further testing. Fast green dye tests are used to identify kernel pericarp damage and the effects of kernel size and shape, and high temperature drying on breakage are being evaluated.