A prototype sensor system for the early detection of microbially linked spoilage instored wheat grain

Abstract

Sensors based on composites of metal oxides were fabricated and testedextensively under high-humidity and high-flow conditions with exposure tovapours reported to increase in the headspace of wheat grain (Triticum aestivumcv Hereward) colonized by fungi. The sensors that exhibited high sensitivity totarget vapours combined with high stability were selected for inclusion into afour-sensor array prototype system. A sampling protocol aligned to parallel gaschromatography–mass spectrometry and human olfactory assessment studies wasestablished for use with the sensor system. The sensor system was utilized toassess irradiated wheat samples that had been conditioned to 25% moisturecontent and inoculated with pathogens known to cause spoilage of grain instorage. These included the fungi Penicillium aurantiogriseum, Penicilliumvulpinum, Penicillium verrucosum, Fusarium culmorum, Aspergillus niger, andAspergillus flavus and the actinomycete, Streptomyces griseus. The sensor systemsuccessfully tracked the progress of the infections from a very early stage andthe results were compared with human olfactory assessment panels runconcurrently. A series of dilution studies were undertaken using previouslyinfected grain mixed with sound grain, to improve the sensitivity andmaximize the differentiation of the sensor system. An optimum set ofconditions including incubation temperature, incubation time, sampling time,and flow rate were ascertained utilizing this method. The sensor systemdifferentiated samples of sound grain from samples of sound grain with 1% (w/w)fungus infected grain added. Following laboratory trials, the prototypesensor system was evaluated in a commercial wheat grain intake facility.Thresholds calculated from laboratory tests were used to differentiate betweensound and infected samples (classified by intake laboratory technicians)collected routinely from trucks delivering grain for use in food manufacture.All samples identified as having an odour-related problem by the intakelaboratory gave a total system output above the set threshold and weretherefore rejected by the prototype system. A number of samples passed bythe intake laboratory were rejected by the prototype system, resultingin what appeared to be false positive results. However, the thresholdswere selected on the basis of a limited number of samples and may needto be adjusted to minimize false positives. The output from the sensorsystem was also compared with moisture content values for the wheat(where available) to demonstrate that the system was not simply measuringdifferences in moisture. A separate study (carried out at the intake facility)assessed 37 newly harvested wheat samples of different varieties and fromdifferent geographic locations within the UK. These samples were analysedby the sensor system, using the same thresholds as before. Six samplesrejected by the system were then assessed by the wheat intake laboratory,where only one sample was rejected. This rejected sample had given thehighest output when exposed to the sensor system. The commercial trialhighlighted the promise of this prototype for the detection of spoilage in wheatgrain and a larger trial should ascertain the reliability and long-termstability of the device and therefore confirm its usefulness to the industry.