First Published: ASMS (Atlanta) 2nd June 2019. This publication is also available in Spanish and Japanese.
Introduction
Advances in Mass Spectrometry and MS data processing are enabling the analysis of micro samples and unknown components that were not previously observed. As the volume of information acquired from MS instruments increases, researchers are calling for more simple techniques to analyse the numerous components observed. As a result, there is a rise in demand for comprehensive analytical techniques and data processing including multiple classification analysis. Here we present a novel technique of non-targeted analysis using high resolution GC-MS with “hard” and “soft” ionization as well as component analysis using AnalyzerPro, an end to end MS software solution.
Experimental
Four different types of commercial coffee beans were selected as samples (A: Indonesian; B: Ethiopian; C: Guatemalan; D: Brazilian). Each sample was measured 5 times. The samples were prepared as follows:
1. One gram of coffee beans was added to a 22 mL headspace (HS) vial, 15 mL of 100°C water was added, and then the vial was sealed.
2. After ambient cooling to room temperature, 10 mL of the supernatant was loaded into another HS vial, and 2 μL of internal standard (p-Bromofluorobenzene) was added to the sample.
3. 2 mL of the above solution was sealed in a vial and used for headspace analysis.
Table 1 shows the experimental condions for the Headspace and GC-MS analysis.
Results
Figure 1 shows the TIC chromatograms acquired from the Field Ionisaon (FI) headspace GC-MS runs for each of the 4 coffee beans. There are visual differences between the components detected at high intensity, however, it would be an extremely lengthy process to manually examine all of the detected components
This process is highly subjective so the use of software capable of objective chromatographic deconvolution and peak detection is more effective in comparing the components between samples. For comprehensive analysis of volatile components in coffee, SpectralWorks’ AnalyzerPro was used.
AnalyzerPro inially extracts the components in question from the chromatogram through chromatographic deconvolution. The program automatically searches the NIST libraries for all mass spectra of the components. The results are tabulated, and the resulting data subjected to Principal Component Analysis (PCA). Figure 2 shows the results of PCA for the headspace EI GC-MS samples normalized to the p-Bromofluorobenzene internal standard peak area in each sample.
The PCA score plot shows good grouping of the sample data according to where the coffee beans originated. The corresponding Loading Plot is shown in Figure 3. It is a simple step to select components from the Loadings Plot that contribute the most to the separation of the categories.
For example, selecting the item furthest to the right of the Loadings Plot will show the component which accounts for the greatest discrimination between the Indonesian coffee and the other coffees. This can be easily confirmed by looking at the average areas plot for this component as shown in Figure 4, indicating that this is a characteristic component on the Indonesian coffee compared to the other three.
Summary
This is just one example of the many characteristic differences that can be identified in these data. The powerful combination of chromatographic deconvolution and PCA available in AnalyzerPro can quickly and easily extract characteristic components that distinguish multiple sample groups. Hard and soft ionisation techniques as well as high or low resolution MS data can be processed.
Acknowledgement
JEOL Ltd for the provision of the data from their MS-62070 STRAP Headspace autosampler and their JMS-T200GC AccuTOF GCx.