Monitoring Gaseous CO2 and Ethanol above Champagne Glasses: Flute versus Coupe, and the Role of Temperature

In champagne tasting, gaseous CO2 and volatile organic compounds progressively invade the headspace above glasses, thus progressively modifying the chemical space perceived by the consumer. Simultaneous quantification of gaseous CO2 and ethanol was monitored through micro-gas chromatography (μGC), all along the first 15 minutes following pouring, depending on whether a volume of 100 mL of champagne was served into a flute or into a coupe. The concentration of gaseous CO2 was found to be significantly higher above the flute than above the coupe. Moreover, a recently developed gaseous CO2 visualization technique based on infrared imaging was performed, thus confirming this tendency. The influence of champagne temperature was also tested. As could have been expected, lowering the temperature of champagne was found to decrease ethanol vapor concentrations in the headspace of a glass. Nevertheless, and quite surprisingly, this temperature decrease had no impact on the level of gaseous CO2 found above the glass. Those results were discussed on the basis of a multiparameter model which describes fluxes of gaseous CO2 escaping the liquid phase into the form of bubbles.


Introduction
Since the end of the 17 th century, champagne is a world-wide renowned French sparkling wine. From a strictly chemical point of view, Champagne wines are multicomponent hydroalcoholic systems supersaturated with carbon dioxide (CO 2 ) dissolved gas molecules formed together with ethanol during the second fermentation process, called prise de mousse (promoted by adding yeasts and a certain amount of sugar inside bottles filled with a base wine and sealed with a cap). During the prise de mousse, bottles are sealed, so that CO 2 molecules cannot escape and progressively dissolve into the wine [1][2][3]. Champagne wines therefore hold a concentration of dissolved CO 2 proportional to the level of sugar added to promote this second fermentation. Actually, a standard 75 centiliters champagne bottle typically holds about 9 grams of dissolved CO 2 , which correspond to a volume close to 5 liters of gaseous CO 2 under standard conditions for temperature and pressure [1][2][3].
When champagne is poured into a glass, there are indeed two pathways for progressive CO 2 and volatile organic compounds (VOC) losses. CO 2 and VOCs escape (i) into the form of heterogeneously nucleated bubbles, the so-called effervescence, and (ii) by ''invisible'' diffusion through the free surface of the glass [4][5][6]. Glass-shape, and especially its open aperture, is therefore also suspected to play an important role as concerns the kinetics of CO 2 and flavor release during champagne tasting [5,6]. From the consumer point of view, the role of bubbling is indeed essential in champagne, in sparkling wines, and even in any other carbonated beverage. Without bubbles, champagne would be unrecognizable, beers and sodas would be flat. However, the role of effervescence is suspected to go far beyond the solely aesthetical point of view. Recently, by use of ultrahigh resolution mass spectrometry, it was indeed demonstrated that ascending bubbles radiate a cloud of tiny champagne droplets overconcentrated with compounds known to be aromatic or the precursors of aromas [7]. Moreover, it was also recently demonstrated that the continuous flow of ascending bubbles strongly modifies the mixing and convection conditions of the liquid medium [5,6]. In turn, the CO 2 discharge by diffusion through the free air/champagne interface may be considerably accelerated, as well as the release of the numerous VOCs, which both strongly depend on the mixing flow conditions of the liquid medium [8].
Carbon dioxide is a potent irritant in the nasal cavity, as are many other organic compounds [9][10][11]. Carbonation, or the perception of dissolved CO 2 , involves a truly multimodal stimulus. In addition to the tactile stimulation of mechanoreceptors, CO 2 acts on both trigeminal receptors [12][13][14][15][16] and gustatory receptors [17,18]. Both of these chemically induced sensations involve the carbonic anhydrase enzyme, which can convert CO 2 to carbonic acid. For the sense of taste, the stimulation with CO 2 appears to involve the extracellular anhydrase enzyme and the transient receptor potential (TRP) mechanism of sour receptor cells [17]. This is consistent with the enhancement of sour taste by the presence of CO 2 and the suppression of sweetness [19,20]. Moreover, given the involvement of TRP mechanisms in both nociception and temperature sensing, interactions between carbonation and temperature might be expected. Actually, an enhancement of irritation, tactile sensations, cooling, and cold pain have all been observed with carbonation of solutions served at low temperature (as are indeed Champagne wines, sparkling wines, and carbonated beverages in general) [21][22][23]. For recent and global overviews of how dissolved CO 2 may promote chemically induced sensations in the oral and nasal cavities, see the review by Brand [24], and the most recent edition of the book by Lawless and Heymann [25]. Among all the numerous VOCs found in wines, ethanol is obviously the one which is the most concentrated [26]. Ethanol is an effective gustatory, olfactory and trigeminal stimulus [27]. In recent studies, it has been shown that variation of wine ethanol content significantly contributes to the partitioning of odorants molecules in the wine headspace by modification of their solubility [28][29][30]. Furthermore, from the taster's point of view, the perception of wine flavors was also found to be influenced by glass shape [31,32]. For all the aforementioned reasons, no wonder that a very strong coupling therefore finally exists in champagne and sparkling wines tasting, between dissolved CO 2 , the presence of rising bubbles, glass shape, CO 2 discharge and VOCs release.
In case of champagne and other sparkling wine tasting, two quite emblematic types of drinking vessels have coexisted for decades: (i) the classical flute, namely a long-stemmed glass with a deep tapered bowl and a narrow aperture, and (ii) the classical coupe, namely a shallower glass with a much wider aperture. Their very different geometrical properties are believed to confer them completely different sensory profiles. Advantages and disadvantages of both glass shapes have indeed long been debated in popular wine magazines, nevertheless without bringing any analytical data corresponding to each type of drinking vessel. Very recently only, CO 2 fluxes outgassing from a standard Champagne wine were measured, whether it was served in a flute or in a coupe [33].
In the present work, a micro-gas chromatography (mGC) technique coupled with a thermal conductivity detector (TCD) was used, in order to sample the chemical space above a glass poured with champagne, in real tasting conditions. Simultaneous quantification of gaseous CO 2 and ethanol concentrations was monitored, all along the first 15 minutes following pouring, depending on whether champagne was served into a flute or into a coupe. Moreover, a recently developed visualization technique based on infrared imaging has been used, thus revealing the clouds of gaseous CO 2 (completely invisible in the visible light spectrum) desorbing from the liquid phase during champagne tasting, whether champagne is poured into a flute or into a coupe.

Champagne samples
A standard commercial Champagne wine (with 12.5% v/v ethanol), elaborated with a blend of 100% Chardonnay base wines (vintage 2007, cooperative Nogent l'Abbesse, Marne, France), was used for this set of experiments. Since their elaboration, bottles were stored in a cool cellar, at 12uC. Concentration of CO 2 molecules dissolved in Champagne samples (before pouring) was determined using carbonic anhydrase (labeled C2522 Carbonic Anhydrase Isozyme II from bovine erythrocytes, and provided from Sigma-Aldrich, US). This method is thoroughly detailed in two recent papers [33,34]. Before pouring, champagne was found to hold a concentration of dissolved CO 2 of CO 2 ½ ~11:6+0:3 g/ L.
Some classical physicochemical parameters of champagne samples were already determined at 20uC, with samples of champagne first degassed [4]. Its static surface tension, c, was found to be of order of 50 mN m 21 , and its density r was found to be very close to that of water, i.e., 10 3 kg m 23 . In the range of usual champagne tasting temperature (varying from approximately 5uC to 15uC), both surface tension and density of champagne are known to be very slowly temperature-dependent, contrary to its viscosity which is known to be strongly temperature-dependent. The temperature dependence of champagne, measured with a thermostated Ubelhode capillary viscosimeter (with a sample of champagne first degassed), was found to classically obey the following Arrhenius-like equation [4]: where the dynamic viscosity g is expressed in mPa s, and the temperature T is expressed in K.

Glass washing protocol
In order to avoid the randomly located ''bubbling environment'' inevitably provided in glasses showing natural effervescence [35], we decided to use, for this set of experiments, single standards flute and coupe with bubble nucleation sites artificially etched just above the central stem (thus providing a ''standardized'' and artificial effervescence). More details on artificial bubble nucleation provided by laser beam impacts can be found in ref [5]. Between the successive pouring and time series data recordings, the flute and the coupe were systematically thoroughly washed in a dilute aqueous formic acid solution, rinsed using distilled water, and then compressed air dried. This drastic treatment forbids the formation of calcium carbonate crystals on the flute wall as well as the adsorption of any dust particle acting as ''natural'' bubble nucleation sites. Therefore, with such a surface treatment, the CO 2 bubble nucleation is strictly restricted to the bubble nucleation sites of the ring-shaped etching, so that differences in the kinetics of CO 2 release from one type of drinking vessel to another are attributed only to geometrical differences between them.
Micro gas chromatography procedure mGC is generally employed to monitor gas of environmental interest such as CO 2 , N 2 O, CH 4 [36,37]. GC coupled with a thermal conductivity detector has already been applied to the analysis of CO 2 , N 2 and O 2 in beverage headspace, the respective concentration of each gas present being determined with a headspace sampler developed to puncture the beverage package (carbonated beverages or still wines) [38]. In the present study, analyses were conducted on a dual channel (A and B) micro gas chromatograph equipped with thermal conductivity detectors (TCD) (MicroGC 200, Agilent, SRA Instruments, France). On channel A, a PoraPlot U (PPU) column was set at 140uC for determination of CO 2 while analysis of ethanol was performed on channel B with a OV-1 column at 100uC. Helium was used as a carrier gas in the two columns. The injection time on both columns was 50 ms. Chromatograms were obtained every 60 s. Peaks areas were quantified using the SOPRANE software (version 2.6.5). The chromatographic conditions for the analysis of CO 2 and ethanol and the peak integration parameters used were the same as those previously described (Table S1) [39]. The quantity of CO 2 was determined by means of a calibration curve using two bottles containing respectively 10% and 1% of standard CO 2 (supplied by Linde gas, France) and air (<0.038% of CO 2 ) was used as a control. Ethanol was quantified with a bottle of gas holding 0.25% of standard ethanol (supplied by Linde gas, France). Calibration with the standard bottles were made with direct connection of the bottle to the mGC sample injection valve using stainless steel tubing, avoiding any gas loss or any disturbing airstream, thus keeping constant the concentration of standard gas. The gas delivery pressure was also kept constant to 1 bar. Then, analyses and quantification of standard gases for calibration were made with the same parameters as those used for the samples ( Table S1). Calibration curves plotted the relative area of CO 2 or ethanol versus the concentration of the standard. The procedure finally developed has shown a high reproducibility: 0.18% relative standard deviation (RSD) for CO 2 10%, 0.54% RSD for CO 2 1% and 0.28% RSD for ethanol.
Experimental set-up and procedure used to measure the concentration of CO 2 and ethanol above a glass poured with champagne A volume of 10062 mL of champagne was carefully poured into the glass previously level-marked. During the standard champagne-like way of serving, champagne vertically falls and hits the bottom of the glass. Characteristic geometrical dimensions and liquid levels of the flute and coupe filled with 100 mL of champagne are displayed in Figure 1. Immediately after pouring, the glass was manually placed at a well-defined position under the injection valve of the chromatograph (see the scheme displayed in Figure 1). Then, the chromatographic analysis was started, and the sampling of champagne headspace above the glass was performed during 10 s and was repeated every 60 s, during 15 min following the pouring process. Figure 2 presents a photographic detail of our experimental set-up. It is worth nothing that, in usual tasting conditions, the consumer rather ''sniffs'' at the edge of the glass. Therefore, we have decided to analyze CO 2 and ethanol concentrations above the champagne surface, close to the edge of the glass for the following experiments.
Experiments were conducted at room temperature (2362uC). Champagne wines were stored at 12 or 2061uC for one day before the experiment. Between the successive pourings, bottles were hermetically closed and stored at the appropriate temperature. To enable a statistical treatment, six successive pouring and successive data recordings on three distinct bottles were done for both glass types.
Statistical analysis was done by Student's t test (two-tailed, two sample unequal variance) to determine whether concentrations of CO 2 or ethanol were significantly different for each time after pouring. Differences at P,0.05 were considered as significant.
Infrared imaging technique used to visualize the flow of gaseous CO 2 desorbing from champagne A visualization technique based on the InfraRed (IR) thermography principle has been used to film the gaseous CO 2 fluxes outgassing from champagne (invisible in the visible light spectrum) [40]. The CO 2 absorptions observable by the IR camera are quite weak because this gas molecule has only a strong absorption peak in the detector bandwidth at 4.245 mm. Consequently, the best way to visualize the flow of gaseous CO 2 desorbing from champagne is to fit the IR video camera with a band-pass filter (centered on the CO 2 emission peak). The experimental device consists of a CEDIP middlewaves Titanium HD560M IR video camera, coupled with a CO 2 filter (Ø 50.8 mm61 mm thick -Laser Components SAS). In complement, the technique involves an extended high-emissivity (0.97) blackbody (CI systems provided by POLYTEC PI), used at a controlled uniform temperature of 80uC, and placed approximately 30 cm behind the glass. The IR video camera was used at a 10 frames per second (fps) filming rate.

Results and Discussion
Losses of dissolved CO 2 during the service of champagne in each type of drinking vessel As recently shown in a previous article, the pouring process is far from being inconsequential with regard to the concentration of CO 2 dissolved into the wine [41]. During the several seconds of the pouring process, champagne undergoes highly turbulent and swirling flows. During this phase, champagne loses a very significant part of its initial content in dissolved CO 2 . Gray scale infrared thermography time-sequences displayed in Figure 3 illustrate the progressive losses of dissolved CO 2 desorbing from the liquid phase into the form of a cloud of gaseous CO 2 , whether champagne is poured in a flute or in a coupe. Clouds of gaseous CO 2 escaping from the liquid phase clearly appear. Consequently, at the beginning of the time series (i.e., at t = 0, after the glass was poured with champagne and manually placed below the sampling valve of the chromatograph), champagne holds a level of dissolved CO 2 well below 11:6+0:3 g L 21 (as chemically measured inside a bottle, after uncorking, but before pouring). In the present work, the initial bulk concentration of dissolved CO 2 after pouring, denoted c i , was also chemically accessed by using carbonic anhydrase. To enable a statistical treatment, six successive CO 2 -dissolved measurements were systematically done for each type of drinking vessel, after six successive pouring (from six distinct bottles). When served at 20uC, champagne was found to initially hold (at t = 0, after pouring) a concentration of CO 2 -dissolved molecules of c flute i~7 :4+ 0:4 g L 21 in the flute, and c coupe i~7 :4+0:5 g L 21 in the coupe (i.e., approximately 4 g L 21 less in both types of drinking vessel after pouring than inside the bottle, before pouring).
Gaseous CO 2 and ethanol content found in the headspace above each type of drinking vessel Concentrations of gaseous CO 2 found above the wine surface were monitored during the first 15 minutes following pouring, as displayed in Figure 4, whether champagne was served into the flute or into the coupe. All along the first 15 minutes following pouring, concentrations of gaseous CO 2 found close to the edge of the flute are approximately between two and three times higher than those reached above the coupe. This observation is self-consistent with some recent data about volume fluxes of gaseous CO 2 measurements above glasses poured with champagne, including a flute and a coupe (as seen in the graph displayed in Figure 5) [41].
Fluxes of gaseous CO 2 per unit surface area offered to gas discharging are indeed significantly higher above the surface of the flute than above the surface of the coupe because the same total amount of dissolved CO 2 (<0.7 gram for both glass types after pouring) has to be released by bubbles from a narrower surface, thus concentrating in turn more gaseous CO 2 in the headspace above the flute. The observation reported in Figure 4 is the analytical proof of a situation well-known by champagne and sparkling wines tasters. Actually, due to higher concentrations of gaseous CO 2 above the flute than above the coupe, the smell of champagne, and especially its first nose, is always more irritating when champagne is served into a flute. It is indeed well-known that a sudden and abundant quantity of CO 2 (a strong trigeminal stimulus) may irritate the nose during the evaluation of aromas [3]. Moreover, it clearly appears from Figure 4 that the concentration of gaseous CO 2 in the headspace above glasses progressively and quickly decreases as time proceeds. This observation betrays the fact that the dissolved CO 2 content in the liquid phase also quickly decreases as time proceeds -from 7.4 g L 21 after pouring to about 3 g L 21 15 minutes later in the case of the flute, as shown in a previous article [41] -thus decreasing in turn the rate at which gaseous CO 2 escapes from the champagne surface.
In a previous article, a model was proposed, which links the total flux of gaseous CO 2 molecules desorbing from the liquid phase into the form of bubbles (denoted dV =dt) with several parameters of the liquid medium [34]: with T being the liquid temperature, g being the champagne viscosity, r being the champagne density, g being the acceleration due to gravity, c L being the bulk concentration of dissolved CO 2 , k H being the solubility of CO 2 molecules in champagne, and P being the ambient pressure. It is worth noting from eq (2) that the lower the dissolved CO 2 concentration is in champagne, the lower the flux of gaseous CO 2 desorbing from champagne into the form of collapsing bubbles. This is totally self-consistent with our mGC data displayed in Figure 4 showing a progressive decrease of gaseous CO 2 concentrations in the headspace above glasses (because the concentration of dissolved CO 2 progressively decreases in champagne all along the first 15 minutes following pouring).
Simultaneously, the concentration of ethanol was monitored with the same successive samplings of the champagne gaseous headspace, analyzed with the second module of the mGC. The successive levels of ethanol found in the headspace above both glass types, all along the first 15 min after pouring, are displayed in Figure 6. Quite surprisingly at first sight, and whatever the glass type, it is worth noting from Figures 4 and 6 that the concentration of ethanol vapors remains roughly constant all along the 15 min following pouring, whereas the concentration of gaseous CO 2 progressively and quickly decreases. The case of ethanol is indeed basically different. Actually, from a strictly chemical point of view, champagne is a highly complex water/ ethanol mixture (at 12.5u v/v). Ethanol being more volatile than water, it evaporates more rapidly than water does. Nevertheless, during the first 15 minutes following pouring, the concentration of ethanol in the liquid phase remains roughly constant, because the ''reservoir'' of ethanol is huge (<10 grams per glass) compared to the small reservoir of dissolved CO 2 (<0.7 gram per glass after pouring) which quickly decreases as time proceeds. Therefore, the rate at which ethanol vapors escape from the champagne bulk  By using time-sequences provided through infrared imaging, the gaseous CO 2 desorbing from champagne and progressively invading the headspace above glasses was made visible in a false color scale (see Figure 7). Such an image processing analysis provides a better visualization of the relative differences in the CO 2 concentration field between both glass types, as shown in the thermography images displayed in Figure 8. Zones highly concentrated in gaseous CO 2 appear in black and dark blue, whereas zones slightly concentrated in gaseous CO 2 appear in red. The concentration of CO 2 found above the flute (close to the edge) is indeed always significantly higher than that found above the coupe, thus confirming the tendency underscored through the mGC measurements. It can be noted for example, through infrared imaging, that the headspace (above the champagne surface, but below the glass edge) remains black during the first 3 min following pouring in case of the flute, whereas it progressively turns blue in case of the coupe. Moreover, it is also worth noting from infrared imaging time-sequences that the cloud of gaseous CO 2 escaping from champagne tends to stagnate above the glass, or even tends to flow down from the edge of glasses by ''licking'' the glass walls (rather than diffuse isotropically around them). These observations conducted through infrared imaging betray the fact that gaseous CO 2 is approximately 1.5 times denser (r CO2 &1:8 g L {1 at 20uC) than dry air is (r air &1:2 g L {1 at 20uC), and therefore tends to naturally flow down.

The impact of champagne temperature
The impact of champagne temperature on the progressive release of gaseous CO 2 and ethanol desorbing from a flute was also investigated, through micro gas-chromatography. Champagne was served at 12uC, and the concentrations of gaseous CO 2 and ethanol found in the headspace above champagne surface was compared with mGC data provided by the same champagne served at 20uC ( Figure 9). As could have been expected, the headspace above the champagne surface is significantly less concentrated in vapors of ethanol when champagne is served at 12uC. Actually, the saturated vapor pressure of a liquid phase is indeed strongly temperature-dependent.
Generally speaking, the lower the temperature is, the lower the saturated vapor pressure of a liquid phase, and therefore the lower the rate of evaporation of the liquid phase. It is clearly selfconsistent with the lower concentration of gaseous ethanol found at 12uC above the champagne surface rather than at 20uC, as seen in Figure 9a.
Nevertheless, and quite surprisingly, the case of gaseous CO 2 seems rather different than that of ethanol, as can be seen in Figure 9b. Concentrations of gaseous CO 2 found above the flute are of the same order of magnitude, and show the same general trend along the first 15 minutes following pouring, whether champagne is served at 12uC or at 20uC. The desorption of CO 2 molecules from the liquid phase is fundamentally different than the desorption of ethanol. Ethanol progressively and continuously evaporates from champagne (as water does also) simply because the opened atmosphere above the flute never reaches the saturated vapor pressure of the binary water/ethanol mixture, whereas CO 2 progressively diffuse from the liquid phase because champagne is supersaturated with dissolved CO 2 molecules. By replacing in eq (2) the viscosity by its Arrhenius-like expression given in eq (1), eq (2) transforms as: The higher the champagne temperature is, the higher the flux of gaseous CO 2 desorbing from the liquid phase into the form of bubbles. Theoretically, and following eq (3), the flux of gaseous CO 2 desorbing from champagne into the form of bubbles should be about 40% higher at 20uC than at 12uC, every other parameter being equal under the same operating conditions. Nevertheless, and despite this strong dependence of CO 2 fluxes on the champagne temperature, the concentration of gaseous CO 2 above the champagne surface experimentally determined through the mGC procedure shows the same general trend and order of magnitude whether champagne is served at 12uC or at 20uC. We are surprised and logically tempted to wonder why. An attempt to explain this experimental observation invokes both the role of the dissolved CO 2 concentration c L found in champagne, and the temperature dependence of the gaseous CO 2 density. Actually, as shown in eq (3), the flux of gaseous CO 2 desorbing from champagne depends on several parameters, including c L in the last term. Now, it is worth noting that losses of dissolved CO 2 during the pouring of champagne into the flute strongly depend on the champagne temperature itself [41,42]. Immediately after pouring and during the first 10 min following, champagne holds almost 1 g L 21 more dissolved CO 2 when it is served at 12uC than when it is served at 20uC [41,42]. Thus, at lower champagne temperatures, the last term of eq (3) should increase, thus counterbalancing the 40% decrease of the two first terms. Actually, by reducing the champagne temperature from 20uC to 12uC, the last term of eq (3) (which includes the ''dissolved CO 2 '' effect) experiences a rough 20% increase. The other 20% of gaseous CO 2 needed to theoretically counterbalance the 40% decrease in the flux of gaseous CO 2 desorbing from the champagne surface at 12uC could originate from a ''density'' Figure 8. Close-up on gaseous CO 2 desorbing above both glass types. False color IR time-sequences showing close-up snapshots of CO 2 clouds desorbing above the flute and the coupe, respectively, immediately after pouring (a), 1 min after pouring (b), and 3 minutes after pouring (c); By using the color scale which provides a correspondence between the relative abundance of gaseous CO 2 and the temperature detected by the IR sensor of the camera after absorption by the gaseous headspace above glasses, it clearly appears that gaseous CO 2 is always more concentrated above the flute than above the coupe. doi:10.1371/journal.pone.0030628.g008 Figure 9. Monitoring ethanol and CO 2 above a flute poured with champagne served at 20 or 126C. Ethanol (A) and CO 2 (B) concentrations found in the headspace above the champagne surface, all along the first 15 min after pouring champagne, depending on whether champagne is served at 20uC or at 12uC; each datum of each time series is the arithmetic average of six successive values recorded from six successive pouring; standard deviations correspond to the root-mean-square deviations of the values provided by the six successive data recordings. doi:10.1371/journal.pone.0030628.g009 effect. Actually, because the density of gaseous CO 2 increases at lower temperature, the flux of gaseous CO 2 desorbing from champagne at 12uC tends to naturally stagnate even more easily above the champagne surface than at 20uC. It could therefore lead to higher gaseous CO 2 concentrations above the champagne surface than theoretically expected from fluxes of gaseous CO 2 as described by eq (3).
In conclusions, simultaneous monitoring of gaseous CO 2 and ethanol was conducted, through micro-gas chromatography (mGC), all along the first 15 minutes following pouring, depending on whether a volume of 100 mL of champagne was served into a flute or into a coupe. The concentration of gaseous CO 2 was found to be significantly higher above the flute than above the coupe. Moreover, a recently developed gaseous CO 2 visualization technique based on infrared imaging was performed, thus confirming this tendency. Those analytical results are selfconsistent with sensory analysis of champagne and sparkling wines, since it is generally accepted that the smell of champagne, and especially its first nose, is always more irritating (because more concentrated in gaseous CO 2 which is a strong trigeminal stimulus) when champagne is served into a flute than when it is served into a coupe. The influence of champagne temperature was also tested. As could have been expected, lowering the temperature of champagne was found to decrease ethanol vapor concentrations in the headspace of a glass. Nevertheless, and quite surprisingly, this temperature decrease had no impact on the level of gaseous CO 2 found above the glass. Those results were discussed on the basis of a multiparameter model which describes fluxes of gaseous CO 2 escaping the liquid phase into the form of bubbles.

Supporting Information
Table S1 Listing of the chromatographic parameters used for the simultaneous analysis of CO 2 and ethanol above the free surface of champagne glasses (redrawn from [39]).