OPTICAL SORTING OF WHOLE KERNELS CONTAMINATED WITH AFLATOXIN OR FUMONISN

Authors: Pearson, Thomas; Wicklow, Donald

Submitted to: Multicrop Aflatoxin and Fumonisin Elimination and Fungal Genomics Workshop-The Peanut Foundation
Publication Type: Abstract Only
Publication Acceptance Date: 9/11/2003
Publication Date: 10/13/2003
Citation: Pearson, T.C., D.T. Wicklow. Optical sorting of whole corn kernels contaminated with aflatoxin or fumonisin. Abstract only.

Interpretive Summary:

Technical Abstract: Near infrared and reflectance spectra (550-1700 nm) were analyzed to determine if they could be used to identify single whole corn kernels contaminated with aflatoxin and fumonisin. Kernels used for the study were collected from ears that were wound-inoculated with Aspergillus flavus or Fusarium verticillioides in the late milk to early dough stage of kernel maturity (1998 and 1999 harvest). Shortly after harvest, the ears were shelled by hand, the wounded kernels discarded, and the whole non-wounded corn kernels examined for visible symptoms of Fusarium or A. flavus infection. Friable kernels and fragments were not included in this study as they are usually removed by existing cleaning equipment at grain elevators. Spectra were acquired on individual kernels showing different levels of BGYF and or discolorations indicative of Fusarium infestation. After spectra acquisition, aflatoxin and fumonisin was measured in each individual kernel by standard chemical methods. For high speed sorting operations, whole spectra cannot be acquired at throughput rates that are economically feasible. Most commercial sorting machines are able to only measure one spectral band of light while some machines can measure two bands. Discriminate analysis was used to select the optimal pair of wavelengths to identify kernels containing aflatoxin. It was found that using the wavelength pair of 750nm and 1200nm, more than 97% of the aflatoxin contaminated kernels were correctly classified as containing either high (>100 ppb) or low (<10 ppb) levels of aflatoxin. Using these pair of wavelengths, nearly 100% of the fumonisin contaminated kernels were correctly classified as containing high (>100 ppm) or low (<10 ppm) levels of fumonisin. Classification accuracy for kernels with intermediate levels of aflatoxin (10 ppb to 100 ppb) and fumonisin (10 ppm to 100 ppm) was poor. A high speed sorter was tested using samples of inoculated and un-innoculated corn from a field in Central Illinois that was harvested by hand, by a Hege field plot harvester, and a commercial harvester (2002 harvest). Additionally, commercially grown and harvested drought stressed corn from Eastern Kansas (2002 harvest) was tested. The sorter was set up to measure reflectance of single kernels at 750 and 1200nm and accept or reject kernels based on these two readings. The sorter was tested at two reject threshold settings. One setting rejected approximately 5% of heavily mold infested corn, the other setting rejected about 10%. For the Kansas corn, the sorter was able to reduce the aflatoxin levels from an average of ~50 ppb to an average of ~10 ppb in the accept stream. There was little difference in results between the 10% and 5% reject settings. Fumonisin levels were reduced from an average of 17 ppm to 2.7 ppm. However, the 10% reject setting removed significantly more fumonisin. For un-innocualted Illinois corn, the sorter reduced mycotoxins from an initial level of 3.1 ppb to 1.4 ppb aflatoxin and 10.8 ppm to 1.4 ppm fumonisin. For corn inoculated with A. flavus, the sorter reduced aflatoxin from an initial level of 398 ppb to 30 ppb for samples harvested by the Hege combine.