Figure 1. GERSTEL MPS with DHS mounted on an Agilent Technologies 7890 GC
Figure 2. Schematic view of the DHS process.
Analysis conditions static headspace
Analysis conditions dynamic headspace DHS
Analysis conditions Cooled Injection System (CIS 4)
Figure 3. Static headspace chromatogram of an aged whiskey, a listing of compounds is shown in table 1.
Figure 4. Dynamic headspace chromatogram of an aged whiskey, a listing of compounds is shown in table 1.
|1||Propanol||23||Heptanoic acid ethyl ester|
|3||Isobutanol||25||ß-Phenyl ethyl alcohol|
|4||3-Methyl butanal||26||Octanoic acid|
|5||2-Methyl butanal||27||Octanoic acid ethyl ester|
|7||1,1-Diethoxy methane||29||ß-Phenyl ethyl acetate|
|8||Propionic acid ethyl ester||30||Nonanoic acid ethyl ester|
|9||n-Propyl acetate||31||Decanoic acid|
|10||3-Methyl-1-butanol||32||Ethyl trans-4 decenoate|
|11||2-Methyl-1-butanol||33||Decanoic acid ethyl ester|
|12||Isobutyric acid ethyl ester||34||Octanoic acid 3-methyl- butyl ester|
|13||Isobutyl acetate||35||1-Ethyl propyl octanoate|
|14||Butyric acid ethyl ester||36||Capric acid isobutyl ester|
|15||Butyric acid 2&3-methyl-ethyl ester||37||Dodecanoic acid|
|16||Isobutyraldehyde diethyl acetate||38||Decanoic acid ethyl ester|
|17||Isoamyl acetate||39||Pentadecanoic acid 3-methyl-butyl ester|
|18||2-Methyl-1-butyl acetate||40||Pentadecanoic acid 2-methyl-butyl ester|
|19||Butyraldehyde diethyl acetal||41||Tetradecanoic acid ethyl ester|
|20||Acetaldehyde ethyl amyl acetal||42||Ethyl-9-hexadecenoate|
|21||Hexanoic acid ethyl ester||43||Hexadecanoic acid ethyl ester|
Table 1. Compound identification
Kevin Mac Namara, Frank McGuigan
Efficient flavor profiling of beverages
that contain involatile matrix
See the big picture - and every little detail
Direct injection for gas chromatographic profiling of alcoholic beverages is usually preferable, but when these contain significant amounts of non-volatile material, pre-treatment is typically required to avoid both inlet and column contamination. This consideration applies in particular to products aged for extended periods in wooden barrels and especially products containing added sugar, as volatile artifacts from sugar decomposition in the hot injection port can also complicate the chromatogram. In this paper a combination of static and dynamic headspace analysis is described for routine profiling of both abundant and trace compounds in alcoholic beverages containing dry extract. Both techniques are performed using one combined analytical instrument and for both techniques the only sample preparation required is dilution of the sample in a headspace vial.
Commercial distilled spirits are complex mixtures of flavor compounds in a dominant ethanol- water matrix [1,2]. These compounds originate from the combined production processes of raw material extraction, fermentation, distillation, and in many cases, ageing in oak barrels. Except for some low volatility compounds originating from wood lignin breakdown during ageing, the majority of flavor compounds in distilled spirits are amenable to gas chromatographic analysis. The matrix composition of distilled spirits is relatively clean and so direct injection without time-consuming sample preparation is possible. Abundant compounds at high mg/L levels can be quantified by simple split injection with flame ionization detection [3,4]. Additional compounds at low mg/L levels (higher esters and acids) can also be assayed by direct injection of 5-10 μL using a PTV injector for both removal of solvent and enrichment of compounds in the liner. This can be extended to 50-100 μL injections for even lower detection limits, but in this case sample introduction must avoid overloading of the injection port liner and subsequent sample loss through the split vent. Speed programmed injection is necessary and recoveries depend on complex interactions between many related sample and instrumental parameters . However, there are many commercial alcoholic beverages which can contain relatively substantial amounts of non-volatile material, and for which direct injection techniques may not be suitable. Fruit spirits and liquors can contain high amounts of added sugar, and very old brandies and whiskies etc. may contain higher than usual amounts of polyphenolic material from wood ageing.
Without frequent liner exchange non-volatile material will accumulate and contaminate both inlet system and column. Added sugar in such products also degrades in the hot inlet to produce artifacts which complicates chromatograms. In these cases there are additional techniques available which can avoid the unwanted effects of non-volatile material. These can be summarized as solid phase micro-extraction (SPME), stir bar sorptive extraction (SBSE) and headspace sorptive extraction (HSSE), static (HSS) and dynamic headspace sampling (DHS). All these techniques have many well documented applications in the literature [6-11].
With SPME a choice of sorbent materials is available but only limited sorbent volumes can be accommodated on the fiber. SBSE and HSSE can use much greater volumes of sorbent material, but this is almost always exclusively apolar polydimethylsiloxane. Headspace application could have the advantage that results may reflect more the actual sensory properties of the product analyzed. Static headspace with intermediate adsorbent trapping was applied to spirit drinks containing dry extract for analysis of the principal abundant secondary alcohols and esters . Automated dynamic headspace using replaceable adsorbent traps was used to profile volatile compounds in beer .
In this paper we describe the sequential application of static and dynamic headspace to profiling both abundant and trace compounds in an aged whiskey. Maximum sensitivity for each mode is achieved by using a PTV injector in solvent vent mode where the liner can also act as a cold trap. Use of a short 0.15 mm I.D. apolar capillary column with a phase ratio of 19 allows fast analysis with excellent separation of both abundant and trace compounds. All operations for both modes of analysis are amenable to total automation for unattended sequence operation.
Analyses were performed using a 7890 GC equipped with a 5975 Mass Selective Detector (Agilent Technologies), Thermal Desorption Unit (TDU, GERSTEL), PTV inlet (CIS 4, GERSTEL) and MPS 2 with headspace and DHS option (GERSTEL) as shown in figure 1.
Aqueous and high water content samples can often be problematic for headspace analysis. The presence of water vapor in the headspace above the sample can lead to poor precision. Operating the PTV inlet in solvent vent mode using a Tenax-filled liner significantly reduces the amount of water transferred to the analytical column.
The GERSTEL Dynamic Headspace System (DHS) is an option for the MultiPurpose Sampler (MPS) which enables dynamic purging of the headspace above a sample. Analytes in the purged headspace are trapped onto a 2 cm sorbent bed in a compact glass tube, an optional dry purge step allows reduction of water content. The thermal desorption tube is then placed into the Thermal Desorption Unit (TDU) and thermally desorbed into the pre-cooled CIS 4 inlet, where the analytes are cryofocused to improve peak shape before introduction into the column. Applying the solvent vent mode in the TDU before transfer of analytes to the CIS provides an additional venting step of e.g. fusel alcohols. Figure 2 shows a schematic of the trapping and desorption process.
Sample Preparation. No sample preparation other than transferring the samples into empty 10 mL screw cap headspace vials is necessary.
Results and Discussion
Figure 3 shows a typical trace obtained using the headspace approach. The fusel or higher alcohols together with ethyl acetate and the principal straight chain fatty acid esters up to dodecanoic acid ethyl ester dominate the chromatogram. Interesting also in the initial elution space are clear peaks for important trace aldehydes, ethyl esters and acetals. Of particular importance are the ethyl esters of short chain fatty acids, called fruit esters due to their pleasant aromas. Pungent aldehydes and their sweet acetals with various alcohols can also affect perceived aroma.
Figure 4 in turn shows the chromatogram obtained when the same sample is injected after dynamic headspace stripping. Analytes up to the C5 alcohols have been partially vented in the TDU since their elution in the chromatogram would give only limited information due to chromatographic crowding, but now much more compound detail is apparent in the remaining elution space. Many interesting esters of both straight and branched chain higher esters are visible and it is even possible to profile some acids. Nonanal and Decanal have been reported previously in beer, wine and cognac, and both are used in the flavor and fragrance business. Both injection modes are very reproducible and do not normally require use of internal standards.
A combination of static and dynamic headspace techniques offers a useful complimentary approach for profiling major and minor components in alcoholic beverages, especially those with substantial levels of dissolved solids. All constituent parts of each analysis are automated using the described instrumentation and no off-line sample preparation is required.
 Aroma of Beer, Wine and Distilled Beverages, L. Nykänen, H. Suomalainen, Eds. Akademie-Verlag. Berlin (1983).
 R. de Rijke, R. ter Heide, Flavour of Distilled Beverages; J. Piggott Ed.; Ellis Horwood: Chichester (1983) 192.
 K. Mac Namara, J. High Res. Chrom. 7 (1984) 641.
 R. Madera, B. Suárez Valles, J. Chrom. Sci. 45 (2007) 428.
 J. Staniewski, J. Rijks, J. Chrom. A 623 (1992) 105-113.
 A. Zalacain, J. Marín, G.L. Alonso, M.R. Salinas, Talanta 71 (2007).
 K. Schulz, J. Dressler, E-M. Sohnius, D.W. Lachenmeier, J. Chrom. A 1145 (2007) 204-209.
 J.C.R. Demyttenaere, J.L. Sánchez Martínez, M.J. Téllez Valdés, R. Verhé, P. Sandra, Proceedings of the 25th ISCC, Riva del Garda, Italy, (2002).
 P. Salvadeo, R. Boggia, F. Evangelisti, P. Zunin, Food Chem., 105 (2007) 1228.
 B. Tienpont, F. David, C. Bicchi, P. Sandra, J. Microcol. Sep. 12(11) (2002) 577-584.
 C. Bicchi, C. Cordero, E. Liberto, P. Rubiolo, B. Sgorbini, P. Sandra, J. Chrom. A 1071 (2005) 111- 118.
 J.R. Stuff, J.A. Whitecavage, A. Hoffmann, Gerstel Application Note (2008) 04/2008.