Fresenius Zeitschrift fiJr Fresenius Z Anal Chem (1982) 311:578-580

�9 Springer-Verlag 1982

Use of 13C/12C Ratios for Studying the Origin of CO 2 in Sparkling Wines

John Dunbar 1

Chemistry Department, University of Waikato, Hamilton, New Zealand

Anwendung von 13C/12C-Isotopenverh~Utnismessungen znr Bestimmung der CO2-Herkunft in Handelssekt

Zusammenfassung. Die Messung von ~3C/~2C-Isotopen- verhfiltnissen wurde dazu benutzt, um in neuseelfindischem Handelssekt den Ursprung des CO2 zu bestimmen. Mit dieser Methode konnte zwischen CO2 aus C-3- und aus C-4- Pflanzenprodukten und CO2 technischen Ursprungs un- terschieden werden. Bei zusiitzlichen ~3C/~2C-Verhiiltnis- messungen an Ethanol von Sektproben war es m6glich, Riickschlfisse auf die Art der Sektherstellung zu ziehen.

Summary. 13C/aZC ratios were measured in an attempt to determine the origin of the CO2 in New Zealand commercial sparkling wines. This technique enabled a distinction to be made between CO2 originating from C-3 plants, C-4 plants and that originating from an industrial source. With the additional measurement of the alcohol 13C/~2C ratios, it was possible to draw conclusions as to the wine making techniques used.

Introduction

The methods used for the production of a sparkling wine can be divided into three general categories. These are a) the Champagne process, b) the tank fermentation (or Charmat) process or c) direct carbonation. The cost of producing wines by each of these methods varies greatly because of the differing labour content involved in each case. It could therefore be a temptation for a wine producer to make a sparkling wine by one of the less labour intensive methods and then claim it to be made by one of the other more complicated processes, hence earning more for his product. It would therefore be desirable to have an analytical method that could determine the method by which a sparkling wine had been made.

A brief description of the methods of sparkling wine production follows: a) Champagne. This is a two-stage fermentation process with step one being the fermentation of the must (grape juice) to dryness as is done in still wine production. All but a small portion of the yeast cells are removed and the wine is then bottled along with a small quantity of sugar solution. The fermentation then restarts in the bottle and continues until the sugar has been used up. During fermentation the bottles are placed approximately

1 Present address: Chemisches Untersuchungsamt, Maximineracht 11 A, 5500 Trier, Federal Republic of Germany

30~ above horizontal and turned daily so that the yeast growth is slowly transferred to the neck of the bottle. When all of the yeast is resting against the cork, the wine is chilled and the cork quickly removed. A small quantity of liquer (wine with approximately 20 % alcohol and 50 % sugar or a brandy and sugar solution) is added to replace the lost wine. In addition, up to 150 ppm sulphur dioxide is added to prevent further fermentation. The bottle is then recorked.

b) Tank Fermentation (Charmat Process). There are many variations possible with this process but basically it involves fermenting the wine in bulk in large tanks under pressure. The wine is then cleared of yeast and residual sweetening is added. Sulphur dioxide is added to prevent further fermentation and the already sparkling wine is bottled under pressure.

The advantages of a system such as this is that it can be done on a large scale without the high labour content that is required with the bottle fermentation method to rack bottles and remove corks. However, the quality of sparkling wine produced by this method is not considered to be of the same standard as that produced by the traditional champagne method.

c) Carbonation. Carbonated wines are produced by directly injecting "food grade" cylinder CO2 into a still wine prior to bottling. This system is the cheapest of all and produces a wine that is generally considered to be of inferior quality [1].

It is known that two basic pathways exist by which a plant is able to fix CO2, and that the isotope discrimination against 13CO2 by each of these pathways is different. The C-3 (or Calvin) pathway uses ribulose-l,5-diphosphate as the CO2 acceptor molecule and the carbon of plants fixed via this pathway have 613C values (definition of 613C given in "Methods" section) in the range of-22~ to - 33O/oo (parts per thousand) [2]. The alternative pathway called the C-4 or Hatch-Slack pathway uses oxaloacetate as the CO2 acceptor molecule, and plants utilising this method of photosynthesis have ~13C values in the range of -10~ to -200/00 [2]. The C-4 pathway is however present in only a minority of plants such as the tropical grasses e.g. corn and sugar cane. As grapes use the C-3 method of CO2 fixation it is thus possible to distinguish CO2 produced from the fermentation of grape sugar from that produced from the fermentation of cane sugar.

As stated previously it is also possible that the CO2 in a sparkling wine can originate from a cylinder, i. e. carbonated wines, so it is necessary to determine the gas's origin. If the CO2 has come from the air then it should have a 613C value in the range -6.7O/oo to -7.4O/oo as was found by Keeling [3] for rural air. However if this CO2 came from city air which

0016-1152/82/0311/0578/$01.00

579

has been contaminated by the burning of fossil fuels (C-3), then this range could be changed considerably. The most probable source for cylinder "food grade" CO2 is however from the combustion of fossil fuels. The resultant 613 C value of the CO2 produced then depends upon the type of fossil fuel used. For coal, oil and natural gas the following range of 613C values have been found: Oil: --25O/o0 to -31~ Coal: Distribution as for modern plants i.e. - 22%o to -33%0 Natural gas : --250/00 to -75~ (depending upon the origin).

Thus it is possible that the 613C value of CO2 produced from fossil fuels is going to be in the range of -22%0 to -750/00 . Therefore unless the ~513C value is < -33~ it will be impossible to distinguish it from the CO2 produced naturally from the fermentation of grape sugars.

As an additional comparison the 613C value of the ethanol in the wine was also measured. By using both the ethanol and CO2 values it should then be possible to gain further information as to the wine making process used.

Methods

Gas Sampling. The CO2 samples (2 ml) were obtained directly with a gas syringe from the headspace of an unopened bottle of sparkling wine. As all of the bottles sampled had plastic stoppers this method presented no great problems. The gas sample was then injected via a rubber septum into a U-tube in a vacuum line, cooled to liquid nitrogen temperature and the air present pumped away. The CO: was dried by three distillations using an ethanol slush bath at - 110 ~ C, then the 613C values were measured using a Micromass 602C isotope ratio mass spectrometer.

In addition to sampling the headspace, samples of the CO2 dissolved in the wine were also taken. This was done by quickly removing 1 ml of wine from a freshly opened bottle and transferring it to a 20 ml flask which was then attached to a vacuum line. Using a process of successive cycles of freezing with an ethanol slush bath and then thawing, the CO2 was removed from the wine. It was then transferred to another section of the line and dried as described previously.

Ethanol 13C/12C Analysis. The ethanol present in 15 ~tl of completely degassed wine was distilled into a standard combustion line where it was combusted over a platinum catalyst at 500 ~ C. The gases produced were passed through a CuO and silver wool furnace to ensure complete oxidation to CO2 which was subsequently dried by distillation using an ethanol slush bath at -110~ The 613C value was then measured.

13C/12C Measurememts. The results of the isotope de- terminations are measured in the 6-notation where:

613C F13C/12C sample 1 : L - i_ •

13C/12CpDB is the 13C/12C ratio of the international standard PDB (Pee Dee Belemnite) the absolute isotope ratio of which has been determined as 13C/12C = 0.0112372 [5].

Results

Comparison Between Sampling Methods

As stated previously, CO: from the wine was sampled either in the gas phase (in the headspace) or from the solution

(dissolved). Mook et al. [6] found that a fractionation existed between dissolved CO2 and gaseous CO2 and at 25~ this was - 1.06%o with the gas being more enriched. In contrast Deuser and Degens [7] found this fractionation to be negli- gible. In the case of wine however the situation is more complicated because a) it contains many dissolved solutes e.g. sugars, and b) it is acidic, hence no appreciable amount of bicarbonate exists.

Both of these sampling methods were carried out on 3 wines, the results of which are shown in Table 1.

Also shown in Table 1 is a column headed A (gas-solution) which expresses the difference in 613C ratio between the gaseous CO2 and that in solution. Errors are also listed. For wine I there is no difference in 6laC values between the gaseous and dissolved CO2 whereas for wines 2 and 3 there are differences of +0.5 and +0.4%o, respectively. In each case the gas is heavier. These differences are lower than those found by Mook et al. [6], however because of the relatively crude sampling technique, the combined errors are large so it is difficult to draw any significant conclusions. The error is such that the fractionation could range between 0 and +0.8%0.

Because both the gaseous and dissolved CO2 have similar 61sc values it was decided to use the headspace sampling method because of its relative simplicity. Ten commercial sparkling wines were surveyed in addition to a bottle of fermented wine which had been produced by the Wine Research Unit of the Ministry of Agriculture and Fisheries at Ruakura (New Zealand). This control wine had been pro- duced by the traditional champagne process i.e. the CO2 in the wine had been produced naturally be secondary fermenta- tion within the bottle. The initial fermentation was done without the addition of sugar however cane sugar was added to allow the secondary fermentation to take place. This sugar served then only as a source of carbonation CO2.

The results of these determinations are shown in Table 2. As can be seen from Table 2, a wide range of values were

obtained for the carbonation CO2 samples measured. The Te Kauwhata Research Centre Champagnes, Nos. 1, 2 and to a large extent No. 3, have had their CO2 derived from a C-4 source which was most probably cane sugar. Nos. 3, 4, 5, and 6 all appear to have had their CO2 derived from a C-3 source e. g. pure grape sugar. This would then imply fermentation by the Charmat process. However, this CO2 was not derived from the wine, it was used to carbonate as is obvious when the 613 C values of the ethanol of these wines are considered. In all 4 cases the ethanol 613C ratios show that the wine was made with large additions of cane sugar (>60%). This sugar addition would then be reflected in the CO2 produced during the fermentation, so if this COz was used to produce the sparkle in the wines then both the alcohol and CO2 should have similar 613C values.

The 6~3C values of the carbonation CO2 for sample Nos. 7, 8, 9 and 10 show that the COz has originated from a non-grape, non-cane sugar source, i.e. the sparkle has probably been introduced into the wine by direct injection of industrial "food grade" CO:.

The ethanol 613C values of these wines also show that cane sugar has been used during the wine making process. In the case of sample No. 7 the 613C value is almost identical to that of pure C-4 sugar indicating that cane sugar has been used almost exclusively as the source of sugar in this wine.

The control wine (Te Kauwhata Research Centre) shows a ~13C value expected from a "legitimate wine". The ethanol

580

Table 1. Comparison of ~13C values for both gaseous and dissolved carbon dioxide

Wine pH Headspace (gas) No. of Solution No. of A (gas-solution) No. analyses analyses

~ c %0 ~ %o ~ c %0 ~ %0 ~ c %0 ~ %0

1 3.0 -26.3 0.4 4 -26.3 0.07 4 0 0.5 2 3.0 - 41.7 0.2 4 - 42.2 0.05 4 + 0.5 0.25 3 3.3 -27.1 0.2 3 -27.5 0.1 4 +0.4 0.3

m e a n A c ] 1 3 C = +0.30/00

Table 2. 13C/12C ratios of C O 2 and ethanol in some New Zealand sparkling wines

Sample number pH CO2 pressure CO2 Number of CO2 Ethanol (atmospheres) ~13Cpo B ~ o determinations source 613CpDB ~ o

1 3.0 - -12.6_+0.1 4 C4 -20.0 2 3.0 - -16.4_+0.3 4 C4 -15.0 3 3.3 1.3 -21.3 _+ 1.3 4 C 3/C4 - 14.8 4 2.8 - -24.1 _+0.2 2 C3 -15.9 5 3.3 - -27.1 _+0.2 3 C 3 - 15.5 6 3.3 1.2 -27.7_+0.7 3 C3 -15.4 7 3.0 1.0 - 41.7 _+ 0.1 4 industrial - 12.9 8 - - 41.7 _+ 0.5 4 industrial - 13.3 9 3.2 1.5 - 43.6_+ 0.3 3 industrial - 21.0

10 3.0 1.8 - 43.9 + 0.4 3 industrial - 20.3

Te kauwhata 3.3 1.5 -10.4-+0.3 4 C4 -25.9 wine research centre champagne

Errors shown as _+ 1 a

has been derived exclusively f rom grape sugar whereas the ca rbona t ion CO2 originates f rom cane sugar that was added to al low the secondary bot t le fe rmenta t ion to take place.

Conclusions

By measur ing bo th the ca rbona t ion CO2 and the e thanol ~13C values of a sparkl ing wine, it is possible to gain in fo rmat ion as to the methods used in its product ion . O f the ten commerc ia l sparkl ing wines s tudied two had their CO2 derived f rom a C-4 source, three f rom a C-3 source, one f rom a C-3/C-4 mixture and four f rom an industr ial source. By addi t ional ly measur ing the e thanol c513 C values o f these wines it is possible to draw conclusions as to the wine m a k i n g techniques used. F o r the wines conta in ing CO2 with ~13C values o f < - 3 0 % o it is obvious that a direct ca rbona t ion me thod has been used. The wines conta in ing C-3 CO2 also p robab ly had COz directly injected f rom a cylinder because the e thanol 613C values show large addi t ions o f cane sugar i.e. the CO2 and e thanol do not originate f rom the same source.

The two wines conta in ing C-4 CO2, but showing their a lcohol to or iginate mainly f rom a C-3 source, have p robab ly been made by a t ank fe rmenta t ion (i. e. Charmat ) process. In bo th cases cane sugar has been added to a wine in which the

fe rmenta t ion has been restar ted so as to produce the CO2 sparkle.

As a fur ther deve lopment of this w o r k it wou ld be necessary to completely survey the sources of commerc ia l ly available " f o o d g rade" CO2 so as to determine the range o f c~ 13C values possible. The measurement o f CO2 613C values could then be used as a m e t h o d to determine the source of CO2 present in a bot t le o f sparkl ing wine.

Acknowledgements. The author would like to thank the University Grants Committee of New Zealand for financial support made available during this work.

References

1. Amerine MA, Berg HW, Cruess WV (1972) The technology of wine making, The Av Publishing Co.

2. Bender M (1971) Phytochem 10:1239 3. Keeling CD (1961) Geochim Cosmochim Acta 124:277 4. Hoers J (1980) Stable Isotope Geochemistry, 2nd Ed. Springer,

Berlin Heidelberg New York, pp 127-138 5. Craig H (1957) Geochim Cosmochim Acta 12:133 6. Mook WG, Bommerson JC, Staverman WH (1974) Earth Planet Sci

Lett 22:169 7. Deuser WG, Degens ET (1967) Nature 215:1033

Received October 19, 1981