Sniffing has come a long way. In the odour industry that is. From the humble beginnings when panels of expert sniffers had only their noses to rely on, their job has become easier by the introduction of instrumental accessories, notably the gas chromatograph. Now, rather than trying to describe the odour characteristics of one particular source, such as cooked food, in its entirety, the odour contributors can be split up and characterised in their own right. This type of job was never easy, but GC certainly helped.
The output from the gas chromatograph is split to a detector and a sniffing port, so that the odour characteristics identified by the sniffer can be correlated with the GC retention time. Later developments placed several sniffing ports on the same instrument, so that a panel of experts could sniff and describe the same odour at the same time, saving much post-sniffing correlation time between individuals.
The next logical step was to insert another split from the gas chromatograph to a mass spectrometer to allow the simultaneous identification of the sniffed compounds from mass spectral library searching. Although this was a notable improvement, the use of a single GC column was often insufficient to separate all of the odorants in a complex mixture.
That led to two-dimensional GC systems (GCxGC), involving the use of two columns of different properties to effect far better separation of closely eluting components. For best performance, a sniffing port is added but this can require the use of a special interface that slows down peak elution from the second column to allow the sniffers plenty of time to sample each aroma as it emerges.
These techniques have been backed up by GCxGC/MS using the mass spectrometer rather than the sniffer to identify the compound but it remains desirable to link the odour to the compound. And herein lies a problem. How to correlate the data from the different systems so that all data for a given compound line up.
That dilemma was faced by scientists from the Swiss company Firmenich SA in Geneva, an organisation specialising in flavour and perfumery chemicals. Alain Chaintreau, Sabine Rochat and J. Egger set out to analyse the aroma of cooked shrimps, one of the more important seafoods. They too had to resort to the use of more than one instrument to dissect the complex odour.
They analysed the odour from shrimp heads cooked in water, a popular ingredient of shrimp sauces, and shrimp shell powder, which is a raw material for the food processing industry and a source of chitinase. The odour compounds were collected by the purge-and-trap headspace technique.
In the first instance, simple GC-O/MS resulted in the tentative identification of many compounds from their linear retention indices (LRIs) and olfactory descriptors held in an in-house database. However, many compounds with high odour thresholds (that could be sensed from especially small amounts) could not be identified because they were below the detection limit of the mass spectrometer.
The next step was to use a GCxGC/MS system with a time of flight (TOF) detector, with sample introduction via a SPME fibre. This set up provided excellent resolution of the aroma volatiles and many more compounds were detected. However, the LRIs were different for the two mass spectrometry systems so that it was difficult to match a peak in the GC-O/MS system with the corresponding peak in the GCxGC/MS system. To complicate matters further, the LRIs within the in-house database had been collected on a third type of instrument by unidimensional GC/MS.
The solution involved plotting the LRIs of known compounds from the GC-O/MS system against those from the GCxGC/MS system. An excellent straight-line correlation was obtained, allowing the aroma compound LRIs to be found on one system when measured on the other. In the same way, the experimental LRIs from the GCxGC/MS system were correlated to those in the database, with excellent linearity. With small error limits, the identities of many compounds were confirmed.
Three peaks that strayed from the straight line in the GC-O/MS vs/ GCxGC/MS correlation had been initially mis-assigned from their mass spectra. The LRIs from GC-O/MS were used to determine those in the GCxGC/MS system and, finally, proposals were made from the mass spectral library at the calculated indices.
Some very potent odorants, notably some heterocyclic compounds, were even below the detection limits of the more sensitive GCxGC/MS system with the TOF. Other compounds like low-molecular-mass aldehydes, were not detected so they were confirmed by preparing substituted hydrazone derivatives.
For one particular compound in the shrimp powder, identification was hampered by a poor mass spectrum due to co-elution with other compounds. It was finally solved using a third system, a GCxGC/MS with a single quadrupole mass spectrometer, a field ionisation detector and a sniffer. A special interface was in place to slow down transfer between the two columns and allow time for sniffing to take place. The compound was eventually declared to be 2-ethyl-3,5-dimethylpyrazine, a well-known and potent aroma compound.
The new retention index correlation strategy has allowed the data from three related techniques to become fully complementary. It led to the identification of many components of the complex shrimp aroma that remained unknown or unsure by the common GC-O/MS technique and should be of great value in the aroma industry.
Related Links:
Journal of Chromatography A 2009 (Article in Press): "Strategy for the identification of key odorants: Application to shrimp aroma"
The views represented in this article are solely those of the author and do not necessarily represent those of John Wiley and Sons, Ltd.
