Abstract
The off-flavor compounds geosmin and 2-methylisoborneol (2-MIB) are well-known to impact the quality of farmed freshwater fish species, but little is known about off-flavors in marine aquaculture. To begin addressing this knowledge gap, a method for determining geosmin and 2-MIB using LC with atmospheric pressure chemical ionization (APCI) MS detection was developed. While 2-MIB was readily detected using LC-APCI/MS, geosmin exhibited on-column degradation that was independent of column chemistry and could not be eliminated. Optimized conditions were identified that balanced the separation and ionization efficiency of 2-MIB and geosmin while minimizing geosmin degradation, but the overall method sensitivity for geosmin was reduced by the on-column losses. The method was used with direct aqueous injections to determine the volatilization rates of geosmin and 2-MIB at ppb levels during aeration under laboratory conditions in both salt water and pure water to simulate marine and fresh water aquaculture, respectively. The volatilization rates of both compounds were 30% faster in salt water than in fresh water with or without aeration, but aeration was found to enhance the rate by a factor of 2.5 in both water types. The LC-APCI/MS method was combined with stir bar sorptive extraction (SBSE) to achieve greater sensitivity for determining off-flavors in recirculating aquaculture system (RAS) water. Using SBSE-LC-APCI/MS, the LODs for geosmin and 2-MIB were 70 ng/kg (part per trillion) and 6 ng/kg, respectively. The on-column losses resulted in a relatively high LOD for geosmin that renders this method unsuitable for determining geosmin at the low ng/kg levels expected in RAS. SBSE using both grab water samples and an in-situ diving unit were used to evaluate 2-MIB levels in the culture water of two separate marine RAS that were supporting the growth of European sea bass but had differing levels of water treatment. 2-MIB was readily detected using both SBSE approaches in the RAS with less sophisticated treatment when the animal stocking density was at its highest (50 kg/m3) but was not detected in the more sophisticated RAS regardless of stocking density. Geosmin was not detected in either system, but the results were inconclusive given its higher LOD. These limited results suggest that the anaerobic water treatment components, present only in the more sophisticated RAS, maintained the level of 2-MIB below the LOD.