Taken together, the results of the PTR-MS and SPME GC-MS techniques provide a more definitive description of the VOCs produced by A. sarcoides for product (e.g. biofuel) development. However, the challenges associated with the quantitative and qualitative analysis of complex VOC mixtures in an aqueous solution are non-trivial, and neither method had the ability to both identify and quantify the breadth of compound diversity synthesized by this complex fungal system.
The A. sarcoides culture consistently produced numerous VOCs. There was little quantitative agreement in VOC concentrations between the SPME technique and the PTR-MS for the suite of compounds produced by A. sarcoides. Both techniques have advantages for exploring potential fuel products from fungal or other cultures. SPME effectively identifies a range of compounds produced by the culture, and at modest concentrations, has adequate fiber capacity for many relevant compounds. The use of PTR-MS gives quantitative data on the evolution of volatiles permitting temporal resolution of microbial volatile production. These data can be used to help elucidate real-time metabolism shifts of an organism for product applications (excluding alkanes) and metabolic modeling. The PTR-MS results may also be valuable in guiding genetic engineering efforts to improve production rates of specific compounds..
To effectively use PTR-MS and SPME GC-MS, the inherent differences in signal acquistion must be considered. Foremost, PTR-MS is a continuous measurement of VOCs in the gaseous stream stripped from an aqueous culture; SPME is a measurement of the headspace compounds that develop above a liquid aliquot after a 45 minute exposure and equilibration period. However, SPME is not truly an equilibrium measurement, because the concentration of the headspace is potentially being depleted by compounds adsorbing to the fiber, lowering their vapor pressure and thereby constantly shifting the headspace composition (Eq. 1) [Zeng and Noblet 2002
]. SPME adsorbs VOCs from the headspace then estimates a liquid concentration based on one internal standard. Therefore, we anticipate differences in distribution of VOCs detected by the two methods (Figures and ). The measurement variability that exists between the two techniques should decrease if the gaseous effluent measured by the PTR-MS is at or near equilibrium with the compounds in the liquid. Based on the abiotic experimental results with a known mixture of VOC standards, equilibrium was assumed for the experimental reactor system.
A limitation of PTR-MS is that alkanes are not ionized by H3
and therefore are not detectable making it unsuitable for some biofuel applications. Additional quantitative complications can arise due to the presence of fragment ions (decomposition products of the protonated molecule). While ionization via proton transfer with H3
is considered to be a relatively soft ionization method, many of the components produced by A. sarcoides
fragment extensively. Most notable are the alcohols that, except for methanol, fragment by the elimination of H2
O from the protonated molecule [Buhr et al. 2002
]. Thus, the alcohols are detected as an ion corresponding to (M-OH)+
. Despite these considerations, the results clearly demonstrate the applicability of the PTR-MS for time series resolution of volatiles from biological cultures.
SPME analysis should reflect the chemical composition and capacity on the SPME fiber if possible. However, not all compounds adsorb to the fiber material with equal affinities or at equal rates, such that the use of one internal standard is not sufficient to determine all constituent concentrations [Pawliszyn 1997
]. The preferential adsorption by some compounds appears to be further exacerbated if the capacity of the fiber is exceeded. The least volatile, highest affinity compounds will be extracted from the headspace-liquid matrix last, therefore adequate extraction times must be observed [Pawliszyn 1999
]. Therefore, drawing conclusions about the concentration of constituents in complex mixtures is very difficult due to fiber selectivity and saturation characteristics. In general, the SPME fiber appeared to preferentially adsorb the higher molecular weight compounds and therefore over predicts their abundance (Table ). While these ions may have been detected by the PTR-MS, signals were low in concentration either because their headspace concentrations were low or because of possible adsorptive losses in the sample lines and therefore were not included in Table . An outcome of this preferential adsorption was seen in the 1-butanol, 3-methyl- data. In the fungal culture, the SPME data correlated well with the PTR-MS signal (Figure ), but not in the abiotic data (Figure ) because the culture headspace environment was dominated by lower molecular weight compounds. The negative trend seen in the abiotic results was most likely because the higher molecular weight compounds were preferentially adsorbed by the SPME fiber, leaving few adsorption sites for the other compounds (Figure ). The extent of fiber bias for the lower volatility, higher molecular weight compounds is not easily predicted or measured for complex mixtures. However, SPME appears to be more applicable in the fungal headspace environment than in conditions where higher concentrations of higher molecular weight compounds were present, such as in the abiotic system.
The PTR-MS and SPME results demonstrated the quantitative limits of the SPME technique even when an internal standard was used. However, use of SPME to estimate relative concentrations of higher molecular weight compounds may be reasonable if application is restricted to qualitative temporal trends and not individual quantitative time points. A possibility for improving the SPME technique is to use replicate samples with additional fiber types to better capture and characterize the breadth of compounds present.
This study shows that despite their shortcomings, the combined use of PTR-MS with a qualitative technique such as GC-MS by headspace SPME is a valuable method for obtaining more detailed analysis of the VOCs produced by microbial systems and can be applied to biofuel product development. However, when quantifying compounds as produced by metabolic processes, especially when those compounds are in a complex mixture, the combination of multiple techniques is warranted until a single more robust qualitative and quantitative analytical method is developed.