Headspace SPME will only extract the most volatile, non-polar compounds from the food matrix you want to analyze for volatile compounds, and thus give a partial picture of the volatile aroma profile.
I might suggest using Stir Bar Sorptive Extraction (SBSE), a method similar to SPME, but far more sensitive and reproducible than SPME, as well as more durable than the fragile SPME fibres. SBSE utilizes a magnetic stir bar coated with polydimethylsiloxane coating (or other more polar films), where you would place a stir bar in an aqueous solution with your sample and stir for a set period of time, then thermally desorb the extracted volatile organic compounds onto the GC column for analysis.
Because of the greater amount of PDMS film and surface area than SPME, a greater concentration and selection of compounds may be absorbed onto the stir bar. Since PDMS is non polar it has low affinity with water, and will only extract the more hydrophobic volatiles. This means the partition coefficient (Log P for octanol:water) of the volatiles will determine % recovery for each compound. Thus, samples with high fats or alcohol must be diluted to under 2-3% fat or 10% ethanol. Because of the greater amount of PDMS film, they say an increase of up to 1000 times sensitivity can be achieved as well as being able to extract far more polar compounds. The graph below shows how compounds with much lower partition coefficients can be extracted. In my experience, I have seen small peaks for ethanol and acetic acid, both very polar compounds in wine samples.
As for quantitation, you will have to use internal standards in all sample solutions and compare these internal standards to internal standards in a standard solution with known amounts of the compounds you wish to be able to quantitate. Since GC uses very small amounts of sample (unlike HPLC where a fixed volume loop lends well to external standard curves), internal standards must be used to compensate for instrument and extraction (sample prep) variation. It must be empphasized that since the absorption of each compound to the PDMS coating is an equilibrium effect, different recoveries for each compound will be noticed (same as SPME), and therefore all compounds you want to quantitate should be in the standard solution at known quantities and this solution can be run with several dilutions to obtain a concentration curve, once normalized using the IS. Although it is theoretically possible (but not recommended) to estimate the concentration of compounds' recovery (if compounds not commercially available) based on their Log P and similarity to other quantitated molecules used in the standard solution.
I would suggest gently comminuting your sample and adding deionized RO water and 20% NaCl (this aids in extracting the more polar volatiles), and stirring with stir bar for 1 hour (1000 rpm). Then use a thermal desorption unit to release volatiles from stir bar onto GC and run analysis. If you do not have access to a thermal desorption unit, some researchers have actually inserted PDMS stir bar into GC insert liner before run, and then thermally desorbing onto column at high injection port temp concentrating onto column using low oven temp before GC analysis. Another option, although I am not very familiar with this aspect, is to use an organic solvent to desorb extracted volatile compounds from SBSE stir bar and then do a straight injection onto GC column.
These links will provide you with more information:
http://chromatographyonline.findanalyti ... ail/599227
http://www.gerstel.com/en/twister-stir- ... action.htm