Comparison of metal oxide-based electronic nose and mass spectrometry-based electronic nose for the prediction of red wine spoilage.

Taints caused by Brettanomyces sp. spoilage are of concern to winemakers and consumers. Typically the taints are described as "barnyard", "sweaty saddle", and "Band-aid" when present in red wine at concentrations of several hundred micrograms per liter or more. The two main components of the taint are 4-ethylphenol (4EP) and 4-ethylguaiacol (4EG), which are metabolites produced by Brettanomyces yeasts. There is a need for a rapid instrumental method to quantify these compounds in wines. In this paper are compared two techniques, the metal oxide sensor-based electronic nose (MOS-Enose) and the mass spectrometry-based electronic nose (MS-Enose). Gas chromatography-mass spectrometry (GC-MS) was used for quantification and prediction purposes. Following ethanol removal, the limits of detection of a MOS-Enose were determined as 44 microg L(-1) for 4EP and 91 microg L(-1) for 4EG, using the SY/gCT sensor. These values are significantly lower than the reported human sensory thresholds. Partial least-squares (PLS) regression of electronic nose signals against known levels of 4EP and 4EG in 46 Australian red wines showed that the MOS-Enose was unable to identify "brett" spoilage reliably because of the response of the gas sensors to intersample variation in volatile compounds other than ethylphenols. Conversely, the MS-Enose was capable of reliably estimating concentrations of 4EP higher than 20 microg L(-1). Correlations (r2) of 0.97 and 0.98 were obtained between estimates of 4EP and 4EG concentrations with the concentrations determined by conventional GC-MS. It is concluded that, following ethanol removal, existing metal oxide sensors are sufficiently sensitive to detect brett taints in wine but lack the selectivity needed to perform this task when the aroma volatile background varies.