4. Fermentation Monitoring
Alcoholic fermentation (AF) is a process common to all alcoholic beverages. A straightforward AF helps afford a microbially and chemically stable beverage
[20][26], but AF also affects the aroma profile
[21][27]. AF, especially if conducted with industrial starters, takes place in a reasonably short time span, ranging from a few days to a couple of weeks. Therefore, the process lends itself well to online monitoring using DIMS.
The selection of suitable yeast starter strains is a paramount step for every type of alcoholic beverage. PTR-TOF has been employed to monitor VOC production during growth in a synthetic medium by different
Saccharomyces cerevisiae strains, including two laboratory strains and four wine isolates
[22][28]. The two lab
S. cerevisiae had an identical genetic makeup with the exception of a single frameshift mutation, affecting amino acid transport. This resulted in measurable differences in the production of higher alcohols (3-methyl-1-butanol, 2-methyl-1-butanol), short-chain fatty acids (isovaleric acid, 2-methylbutanoic acid) and mercaptans (methanethiol).
DIMS has been employed to study the alcoholic fermentation of mead (i.e., an alcoholic beverage produced from diluted honey). The work
[23][29] compares PTR-MS with Fourier-Transform Infrared (FTIR) spectroscopy and HPLC. Whereas FTIR allows for online process monitoring, HPLC requires offline sampling and analysis and has limited throughput. PTR-MS represents a valuable compromise between the previously mentioned techniques: in fact, it allows for non-invasive monitoring of fermentation kinetics in real time, at the same time providing information on the production or consumption of key metabolites and aroma compounds, such as ethanol, acetaldehyde, acetic acid and esters.
PTR-TOF has been used as an AF online monitoring tool in experimental brewing to investigate the impact of different yeasts and hops and their combination
[24][30]. The results highlighted complex interaction between factors, with hops affecting aroma profile directly, but also indirectly affected by yeast growth and metabolism.
An approach similar to that applied for brewing has been adopted using PTR-TOF to study fermentation and aroma production kinetics during experimental winemaking
[25][33], with AF being carried out using different microbial strains (
S. cerevisiae,
Metschnikowia pulcherrima and
Torulaspora delbrueckii either alone or in combination) in either fresh must or commercial grape juice. In consideration of the different applications, the interest in DIMS techniques to explore the ‘space’ of interactions between the chemistry of the must/wine and the pro-technological microorganisms is evident, as well as in exploring the different interactions between microorganisms of oenological interest (
Saccharomyces and non-
Saccharomyces yeasts and malolactic bacteria) that can be inoculated with diverse timing.
5. Static and Dynamic Headspace Analysis of Model Ethanolic Systems
During drinking, the first olfactory interaction with the product occurs when aroma compounds reach the nasal epithelium directly through the nose: the process gives rise to the so-called orthonasal perception. The accurate measurement of the composition of the above-the-glass headspace is key to understanding the chemical mechanisms underlying appreciation.
In alcoholic beverages, ethanol may affect the air-to-liquid partitioning of aroma compounds in different ways. APCI-MS was used to characterise the headspace composition of water and ethanol/water model systems, showing that in conditions of static equilibrium
[6][12], ethanol reduces headspace VOC concentration: headspace analysis of water or 12% ethanolic solutions containing different VOCs showed a decrease of up to 42% in headspace concentration for the ethanolic system. The impact of ethanol addition on headspace concentration varies as a function of the compound hydrophobicity, as determined by the log
p value.
A major advantage of the DIMS approach is the possibility of directly performing dynamic measurements, which are more representative of real-life/real-system situations. When APCI-MS was used for dynamic headspace measurements in water and in 12% ethanol solutions
[26][27][35,36], the resulting VOC profiles started with a sharp decrease in measured concentrations, followed by dynamic equilibrium. In ethanolic solutions, VOC concentrations in dynamic equilibrium conditions were generally higher than in water alone; therefore, in such conditions, ethanol helps maintain the compound concentration in the headspace constant and reduces headspace dilution. The magnitude of this effect, which was somewhat affected by the air-to-liquid partitioning coefficient of the target compound, is likely related to the surface-active properties of ethanol. In fact, ethanol evaporation might create local areas of higher evaporation and surface tension, recalling liquid from areas with lower evaporation. The resulting convection phenomenon constantly recalls fresh solution towards the surface, thus helping to keep a constant headspace concentration for most VOCs.
In alcoholic beverages, molecules other than ethanol can also contribute to matrix effects: APCI-MS has been used to study their behaviour. Volatile release has been investigated in model solutions containing odour-active esters and heterocyclic compounds typical of whisky aroma
[28][38]. Ethyl-hexadecanoate, used as representative of long-chain ethyl esters typical of new-make whiskies, reduces aroma release at concentrations above its critical micelle concentration, especially in the case of hydrophobic compounds. This means that whisky dilution may affect perception by altering the balance between hydrophobic and hydrophilic compounds in the headspace, also depending on the content in long-chain ethyl esters. This is particularly relevant in describing real-life conditions, as whisky is often diluted with water to reduce pungency and improve aroma appreciation, both during professional tastings and for everyday consumption.
Another field for the application of DIMS is the direct measurement of carbonated beverages, such as beer or sparkling wine. In such systems, the opening of the bottle creates rapidly evolving dynamic conditions. Therefore, static measurements may provide results which are not representative of the real conditions. When model beers with different compositions were measured using APCI-MS under static equilibrium conditions
[29][40], no statistically significant effect of alcohol, hop acids or carbonation on aroma release was observed.
6. Nosespace Analysis
After wine is swallowed, the opening of the epiglottis and the velum–pharynx barrier allows VOCs to travel from the mouth to the olfactory epithelium; in a later phase, VOCs released from liquid remainders adhering to the mouth and throat mucosal surfaces sustain perception over time. This mode of perception is labelled ‘retronasal’ and generally follows orthonasal perception. The possibility of employing DIMS to analyse ’in vivo’ aroma release was first explored by developing a dedicated APCI-MS source
[30][42]. The technique, generally referred to as ‘nosespace’ analysis, is performed by connecting the instrument inlet to the assessor’s nose by means of a custom nosepiece. The results typically obtained consist of time-resolved release profiles (
Figure 23); from these, some key metrics of particular significance can be extracted (e.g., ‘area’, ‘maximum’ or ‘t.max’) and used to provide an analytical determination capable of supporting the description of the evolution of perception over time.
Figure 23. Example of a nosespace profile. Different metrics can be extracted from each profile, such as the area under the curve (area), the maximum intensity (max) and the time after which maximum intensity is reached (t.max). The vertical dotted line marks the moment of sample introduction.
DIMS has been used to perform nosespace measurements on ethanolic model systems
[29][31][40,43], as well as real alcoholic beverages, including wine
[32][44], brandy
[33][45], vodka
[34][46] and palm wine
[35][47].
When APCI-MS nosespace measurements were performed on model systems mimicking VOC release in beer
[29][40], the maximum and total area of release profiles were increased by both ethanol and CO
2 addition. In fact, both compounds are surface-active and might increase release directly or by reducing VOC adhesion to mucosal surfaces. The effect of carbonation has also been observed in non-alcoholic beverages during PTR-MS measurements carried out in vivo
[36][48] and using artificial mouths
[37][49].
7. Headspace Analysis of Wine and Other Alcoholic Beverages
Whenever a new DIMS-based approach is developed for the analysis of ethanolic systems, wine discrimination has often been employed as a case study. Proof of concept for the PTR-MS ethanol saturation method was first obtained, discriminating commercial red and white wines from different regions and grape varieties
[8][14]. PTR-MS, with the dilution method, was used in the separation of experimental wines obtained from Cabernet Sauvignon and Pinot Noir grapes harvested on different days and at different altitudes. The wine samples from the higher altitude sites and at later harvesting dates displayed higher signal intensities for a number of VOCs as compared to earlier harvested, lowland samples.
The PTR-MS high E/N method was tested in the discrimination of eight French brandies from different producers with 2–6 years of ageing
[38][58]. The authors compared results obtained at E/N = 145 and 454 Td. Samples were well separated in both instances, but discrimination took place according to different ions and when the same ions were present in both sets of conditions, they were not likely to come from the same compound. This is not unexpected, as with high E/N, PTR-MS reactivity is definitely shifted more towards dissociation; furthermore, under such conditions, mass spectral interpretation is difficult because multiple analytes may easily give rise to the same fragment.
In the case of APCI-MS, application to real alcoholic beverages has been less frequent, with applications published on the targeted analysis of wort, commercial beer and cider samples
[39][59] and palm wine
[40][60]. PTR-TOF has instead been used more often, possibly due to its improved characteristics in terms of mass and time resolution, guaranteeing improved performance in untargeted analysis. PTR-TOF with argon as buffer gas has been employed in the separation of experimental wines resulting from malolactic fermentation (MLF) performed using three different
Oenococcus oeni strains
[10][16]. The main drawback when using argon was the reduction in sensitivity (sometimes as high as ten-fold) but the method allowed to preserve standard PTR-TOF chemistry even in the presence of high amounts of headspace ethanol.
8. Quality Control
DIMS techniques have been used in quality control of ingredients, materials or finished products. For instance, SIFT-MS has been employed to detect methylamine contamination in agricultural ethanol
[7][13] or to determine methanol content in beer, wine, gin and whisky
[41][62]. Interestingly, the reactivity conditions observed when measuring ethanolic samples with SIFT-MS are close to those occurring in APCI-MS or PTR-MS under ethanol saturation conditions. VOCUS-CI-MS has been used in combination with a dedicated autosampler with the aim to quantify 2,4,6-Thrichloroanisole (TCA), the most widespread cork stopper contaminant
[42][63]. This custom setup allowed the high-throughput analysis of TCA on intact cork stoppers with outstanding results: 5000 stoppers were analysed in a little over 4 h, whereas for a subset of these corks, a good correlation was observed with TCA concentrations obtained according to the ‘releasable TCA’ method
[43][64].