- Email: [email protected]
champagne wine interface
Advances in Colloid and Interface Science 88 Ž2000. 19᎐36
Characterisation by drop tensiometry and by ellipsometry of the adsorption layer formed at the airrchampagne wine interface N. Peron ´ a,b, A. Cagnaa , M. Valade c , R. Marchal d, b A. Maujeand, B. Robillard e, V. Aguie-Beghin , ´ ´ b,U R. Douillard a I.T. Concept, Parc de Chancolan, 69770 Longessaigne, France INRA, Equipe de Biochimie des Macromolecules Vegetales, CRA, 2 Espl. R. Garros, BP 224, ´ ´´ 51686 Reims Cedex 2, France c Comite´ Interprofessionnel du Vin de Champagne, 5 rue Henri-Martin, BP 135, 51204 Epernay Cedex, France d Laboratoire d’Œnologie, URCA, Moulin de la Housse, BP 1039, 51687 Reims Cedex 2, France e Moet ¨ et Chandon, 6 rue Croix de Bussy, 51200 Epernay, France b
Abstract A foam ring composed of small bubbles on the surface of a champagne glass is one of its hallmarks. The equilibrium state of that ring is linked with the rate of formation and of disappearance of bubbles. The stability of bubbles is usually ascribed to the occurrence and to the properties of an adsorption layer formed at the gasrliquid interface. Our goal is to characterise such an adsorption layer at the gasrwine interface in order to understand its role in bubble stability. Alcohol in wine lowers the surface tension to 49 mNrm. The adsorption of other molecules may cause a further decrease of 2 mNrm. Such a situation makes the study of adsorption by surface tension measurement inaccurate. To overcome this problem, we have diluted the wine four times with water before its surface tension
U
Corresponding author. Tel.: q33-3-26-77-35-94; fax: q33-3-26-77-35-99. E-mail address: [email protected] ŽR. Douillard.. 0001-8686r00r$ - see front matter 䊚 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 1 - 8 6 8 6 Ž 0 0 . 0 0 0 3 9 - 7
20
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
measurement by pendant drop shape analysis. In these conditions, ethanol lowers the surface tension to 64 mNrm and the adsorption of other molecules of the wine can be monitored over 6᎐8 mNrm. The usual behaviour of such a diluted wine is a lowering of the surface tension during at least 20 min after drop formation. Since the role of macromolecules on the foaming properties of wine had been previously observed, we have chosen to evaluate the effect of this fraction of the wine molecules on its surface properties. Thus, wines were ultrafiltrated on a membrane with a 10 000 molecular mass cut-off. The ultrafiltrate ŽUF. does not show any decrease of its surface tension over a 20-min period while the ultraconcentrate ŽUC. has a kinetics similar to that of unfiltered wine. Mixtures of UF and UC have behaviours intermediate between those of these products. A technological treatment of the wine with bentonite, believed to lower the content of macromolecules, yields a wine similar to UF. The effect of ultrafiltration was also analysed by spectroscopic ellipsometry. UF has a spectrum similar to that of a waterralcohol mixture with the same ethanol content and its ellipticity is stable during at least 20 min. On the contrary, wine or UC show spectra with the features of an adsorption layer and those characteristics increase during more than 20 min. Two varieties of vine were compared: ‘Chardonnay’ and ‘Pinot noir’. The former is known to have better foaming properties than the latter. Its surface properties measured in this study are also more pronounced than those of Pinot noir. However, the representation of the dilational modulus against the surface pressure Žwhich, in some instances, may be a mathematical transformation of the state equation. puts all the samples Žwines, UF and UC of each. on the same master curve, a fact in favour of a common nature for all the adsorption layers. It can be concluded that surface properties of champagne wines are mostly determined by ethanol and by macromolecules with a molecular mass larger than 10 000. Moreover, the adsorption layers seem to have the same nature, irrespective of the vine variety and of the concentration ratio of the wine. 䊚 2000 Elsevier Science B.V. All rights reserved. Keywords: Foaming behaviour; Airrchampagne interface; Macromolecules; Pendant drop method; Ellipsometry
Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Wine samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Surface tension measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Ellipsometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Protein quantification by the Bradford method . . . . . . . . . . . . . . . . . . 3. Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Effect of wetting on reproducibility . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Water dilution of wine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Effects of macromolecule concentration on the surface properties . . . . . 3.3.1. Kinetics of formation of the adsorption layer . . . . . . . . . . . . . . 3.3.2. Ellipsometric characterisation of the adsorption layer . . . . . . . . . 3.3.3. Dilational modulus of the adsorption layer . . . . . . . . . . . . . . . . 3.4. General discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1. Effect of the adsorption of macromolecules on the surface tension
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. 21 . 21 . 21 . 22 . 23 . 24 . 25 . 26 . 26 . 28 . 29 . 29 . 31 . 33 . 33
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
21
3.4.2. Effect of ultrafiltration on the properties of the products . . . . . . . . . 34 3.4.3. Use of surface tension measurements to monitor the occurrence of surface-active molecules in the wine . . . . . . . . . . . . . . . . . . . . . . . 35 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1. Introduction Small bubbles forming a regular ring around the glass are appreciated by consumers and make the hallmark of champagne. The ability of champagne to form a nice foam is often appreciated using a sparging procedure in a glass tube and measuring the properties of the foam with the ‘Mosalux’ device w1,2x. However, the foam measured in that way is quite different of the subtle collar formed in a glass. Moreover, the ‘Mosalux’ procedure has never been claimed to measure the ability of a ‘base wine’ Žthe wine before the second fermentation. to produce an elegant bubble ring. From a more basic standpoint, the physical phenomena involved in bubble and foam stability are thought to be linked mainly with the properties of the interfaces between wine and gas w3,4x. Thus we have chosen to investigate and characterise the surface properties of base wines in view of studying the correlations between such properties and the foam behaviour of the corresponding sparkling wines. In this study, we take advantage of the fact that treatments such as ultrafiltration w5x or bentonite application w1x are known to decrease drastically the foaming behaviour of the wine: a related decrease of the surface properties is expected. A recent study of protein adsorption in hydro-alcoholic solutions w6x showed that adsorption of macromolecules has not always a strong lowering effect on surface tension. Thus, this study is based on both surface tension and ellipsometry measurements in order to detect a surface layer and to measure its effect on surface tension.
2. Materials and methods 2.1. Wine samples Wines were produced from two vine varieties grown by CIVC ŽComite ´ Interprofessionnel du Vin de Champagne, Epernay.: ‘Chardonnay’ or ‘Pinot noir’ during the 1997 harvest. They were used as still wines before the second fermentation. Ultrafiltration Žultraconcentration . of wine was performed by tangential ultrafiltration on a 1.8-m2 hollow fibre device ŽInceltech, Toulouse, France. with a 10 000 nominal molecular mass cut-off. The ultrafiltration membrane was made of hydrophilic polysulfone. Before ultrafiltration of the volume needed Ž0.8 l., the device was rinsed with 1.6 l wine. The transmembrane pressure was 0.7 atm. The relative
22
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
concentration factor ŽRCFUC . of the ultraconcentrate ŽUC. was calculated with respect to native wine as: RCFUC s Ž VUF q VUC . rVUC
Ž1.
where VUF and VUC are the volumes of ultrafiltrate ŽUF. and of UC of the experiment. RCF of each sample was calculated from the volumes of wine, of UC and of UF mixed to prepare the sample and from their respective RCF: 1 for native wine, RCFUC for UC and zero for UF. Bentonite precipitation has been achieved with 1 grl. Bentonite was first swollen with 20 times its weight of ultrapure water during 24 h at room temperature under mild stirring. Wine and bentonite were mixed and stirred during 20 min before centrifugation Ž15 000 = g, 20 min.. When needed, wine samples were diluted with ultrapure water. All glass vessels were treated with chromosulfuric acid before use. Syringe needles were thoroughly rinsed with ultrapure ethanol and water. 2.2. Surface tension measurements All static and dynamic surface tension measurements were performed with a drop Žbubble. tensiometer from IT Concept, Longessaigne, France w6,7x. A complete description of the experimental set-up has been given by Benjamins et al. w8x. An axisymmetric drop or bubble is formed at the tip of the needle of a syringe whose plunger position is driven by a computer. The image of the bubble Ždrop. is taken from a CCD camera and digitised. The interfacial tension ␥ is calculated by analysing the profile of the bubble according to the Laplace equation: Ž 1rx . d Ž xsin . rd x s 2rb y cz
Ž2.
The origin of the coordinates is at the bubble apex; x and z are the Cartesian coordinates at any point of the bubble profile, b is the radius of curvature at the bubble Ždrop. apex, is the angle of the tangent to the bubble Ždrop. profile and c s 2ra2 , where a is the capillary constant Ž a s w2 ␥rŽ g .x1r2 , where is the difference of density between the two phases, and g the acceleration of gravity.. The area of the drop Žbubble. and the surface tension are calculated several times per second. This set-up was made suitable to measure the surface dilational modulus w3x: s d␥rdln A
Ž3.
where A is the surface area of the bubble Ždrop.. This was achieved by fluctuating sinusoidally the area of the drop Žbubble. at a frequency of 0.1 Hz and a relative amplitude of 0.07 which is the lowest value with an acceptable signal-to-noise ratio and where the surface tension amplitude response is in the linear range. No significant change of the modulus was observed by changing the frequency from 0.01 to 0.1 Hz. Moreover, a 10-s period is short enough compared to the duration
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
23
of the adsorption experiment which is of the order of 20 min. The modulus was calculated using a sliding window with a 10-s width Žone period., using the A and ␥ variations. A Fourier transform of the signal is used to extract the main component and avoid the small perturbations linked to the inaccuracies in the movement of the plunger. The surface tension was calculated as the mean value during the same period. All experiments were performed in an air-conditioned room at a fixed temperature of 20 " 1⬚C. Drop formation at the tip of a needle was achieved in a small confining cell Žf 250 ml. fitted with two parallel windows and containing a few millilitres of wine coating its bottom part. This allows for a rapid equilibration of the volume around the drop with wine saturating vapours and avoids evaporation from the drop. The plunger of the syringe is allowed to enter through a toric seal ensuring a gas tight cell. In most experiments, the lowering of the surface tension w ⌬␥Ž t . s ␥Ž0. y ␥Ž t .x was plotted against time instead of the surface pressure which cannot be calculated accurately since the surface tension of the ‘pure’ solvent is not a defined physical quantity in the case of wine. In order to begin the measurements as close as possible to the solvent tension, two drops Žbubbles . were quickly expelled before forming the measuring drop Žbubble.. 2.3. Ellipsometry All measurements have been done using a spectroscopic phase modulated ellipsometer ŽUVISEL, Jobin Yvon, Longjumeau, France.. It is equipped with a xenon arc lamp. Both the polariser and the analyser are set to the 45⬚ configuration angle. The photoelastic modulator, activated at the 50-kHz frequency, is set to the 0⬚ configuration orientation. The spectroscopic measurements are monitored between 250 nm and 700 nm. The incidence angle is set to 53.6⬚. All measurements were done at the airrliquid interface in an air-conditioned room at 20 " 1⬚C. The two ellipsometric angles ⌿ and ⌬ w9x, are linked to the two reflectivity coefficients rp and rs , respectively, in the directions parallel and perpendicular to the incidence plane, by: rprrs s tan⌿exp Ž i⌬ .
Ž4.
The fixed wavelength chosen for the kinetics measurements corresponds to the Brewster conditions of the pure substrate defined by: ⌬ s r2
Ž5.
The coefficient of ellipticity of the adsorption layer measured at the Brewster conditions for the substrate is defined as: s tan⌿sin⌬
Ž6.
The index of refraction n D was measured with an Abbe refractometer ŽRF490, Prolabo, Paris..
24
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
Fig. 1. Effect of the experimental set-up on the reproducibility of the surface tension measurements. Ža. Surface tension was measured using bubbles formed at the tip of a stainless steel needle. Žb. Surface tension was measured using droplets formed at the tip of a Teflon-coated needle.
2.4. Protein quantification by the Bradford method The Bradford method was employed as previously described w10x: 0.2 ml Bio-Rad Protein Assay reagent ŽCoomassie Brilliant Blue, CBB. is added to the 0.4-ml sample. The blue coloration was measured at 595 nm using a spectrophotometer ŽSecoman S250. after 60-min contact when the colour was stable. The protein content was calculated with respect to a calibration curve with bovine serum albumin ŽBSA. Ž0᎐20 mgrl.. Each value corresponded to the average of three measurements. The standard deviation was less than 2%.
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
25
Fig. 2. Effect of the coating of the needle on its wetting by the wine samples. The horizontal line materialises the tip of the needle. Ža. Stainless steel needle; an ascent of the contact line of the drop in an asymmetrical way is visible. Žb. Teflon-coated needle; the contact line of the drop is located at the tip of the needle.
3. Results and discussion Several problems were encountered when trying to analyse the effect of the wine macromolecules on its surface properties. Possible solutions to them are shown here. First is presented the effect of wetting of the needle on the reproducibility of surface tension measurements, then the occurrence of adsorbed material is ob-
26
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
scured by the low surface tension imposed by ethanol in wine and a dilution by water may solve this problem. To confirm the effect of high molecular mass molecules on the surface properties of wine, a comparison of surface tension, surface dilational modulus and ellipsometric parameters was made between wine and samples derived from the base wine particularly by ultrafiltration. 3.1. Effect of wetting on reproducibility At first, surface tension was measured using the bubble procedure which is not sensitive to evaporation at the winerbubble interface. Using a stainless steel needle and wine diluted fourfold by water, the kinetics of evolution of the surface tension of a bubble formed in the liquid was poorly reproducible, as illustrated in Fig. 1a. However, when the kinetics was pursued more than 1 h, the standard deviation of the measured surface pressures was less. Using a drop formed at the tip of a Teflon coated needle ŽFig. 2b., in the gas-tight cell, the reproducibility was much better ŽFig. 1b.. It can be seen using droplets formed with a non-Teflon coated needle that the diluted wine wets the surface of the needle ŽFig. 2a.. That wetting is not completely reproducible and the height of the needle covered by a liquid film is not constant. Thus, it is likely that using a bubble, wetting occurs inside the stainless steel needle and is not visible with the CCD camera equipment of the tensiometer. It is probable that when such a wetting occurs, a film leakage may also happen on the needle, as it has previously been noticed w11,12x, inducing the lack of reproducibility in the surface tension kinetics. In conclusion, care should be given to such phenomena and an evaluation of the wetting is recommended using the drop procedure. 3.2. Water dilution of wine When a new surface is formed at the gasrwine interface, it is mostly free of adsorbed molecules and the surface tension should be that of the solvent. In the case of wine, ethanol adsorbs quickly and, with the procedure used, its adsorption kinetics cannot be resolved. Thus, at the beginning of the kinetics, surface tension is mostly imposed by the concentration of ethanol. Using native wine, it was observed that the surface tension when the drop has just been formed is approximately 2.5 mNrm less than that of the corresponding waterrethanol mixture ŽTable 1.. This seems in favour of molecules other than ethanol adsorbing quickly at the interface or of solvent conditions for ethanol different from those found in pure water. This point stresses the question of the nature of the ‘solvent’ when considering wine. In addition, this unresolved question makes difficult the use of the physical quantity ‘surface pressure’ which refers to the solvent surface tension. Nevertheless, the surface tension decreases less than 0.5 mNrm in the first 20 min of formation ŽFig. 3.. This low decrease does not allow an accurate characterisation of the surface-active molecules of the wine and of the dilational modulus of the interface. Thus, measurements were performed after a dilution of the wine by water to decrease the effect of ethanol on the surface tension ŽFig. 3.. In these
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
27
Table 1 Ethanol concentration and surface tension of wines, of fourfold diluted wines, and of hydroalcoholic solutions with the same ethanol concentration a
Chardonnay Pinot noir Chardonnay, 1r4 dilution Pinot noir, 1r4 dilution a
Ethanol concentration Ž%.
Surface tension ŽmNrm. Wine
Ethanol᎐water solution
10.8 11.6 2.7 2.9
46.9 46.1 60.9 61.2
49.4 48.6 64.6 64.1
Surface tension was measured immediately after forming the drop.
conditions it is clear that surface-active molecules adsorb and lower the surface tension to a value significantly smaller than that measured at the beginning of the kinetics Ž2᎐4 mNrm.. Several dilutions were tried. It is likely that when the wine is diluted 10 times, the concentration of surface-active molecules is so low that the kinetics is slowed down. In the case of a fourfold dilution, the lowering of surface tension is larger than for a twofold dilution. For these reasons, it seems finally convenient to use a fourfold dilution for all samples. An important conclusion of these water dilution experiments is that, apart from ethanol, surface-active molecules occur in champagne base wine and that it is possible to characterise them by surface tension measurements.
Fig. 3. Effect of wine dilution on its surface tension kinetics. Chardonnay wine was eventually diluted with ultrapure water and the kinetics of its surface tension recorded during 20 min. The dilution is quoted on each curve.
28
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
Fig. 4. Kinetics of surface tension of various wine samples. All samples were diluted four times with ultrapure water. `, native wine; q, ultraconcentrate; ⽧, ultrafiltrate; I, bentonite-treated wine; =, reconstituted wine from UF and UC; ', native wine diluted twice by UF. Ža. Chardonnay wine; Žb. Pinot noir wine.
3.3. Effects of macromolecule concentration on the surface properties To characterise the effect of macromolecules on the surface properties, a study has been done comparing the properties of native base wine, ultrafiltrate ŽUF., ultraconcentrate ŽUC., reconstituted wine from UF and UC with a macromolecular concentration identical to that of the native wine, native wine diluted twice by UF, and bentonite-treated wine.
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
29
Fig. 5. Effect of the relative macromolecular concentration on the surface tension reached after 5 min. The zero time of the kinetics is the formation of the drop at the needle tip. The macromolecular concentration is relative to the macromolecular concentration of the native still wine. All samples were prepared by mixing UF and UC. From `, Chardonnay; =, Pinot noir.
3.3.1. Kinetics of formation of the adsorption layer The kinetics of evolution of the surface tension of the various samples show that UF has a very small decrease of surface tension Ž ⌬␥ ., while native wine or UC have a significant and rather similar decrease. Wine diluted twice by UF has an intermediate behaviour and bentonite-treated wine is close to UF ŽFig. 4.. Moreover, all the results obtained with Pinot noir wine show a smaller decrease of the surface tension than those obtained with Chardonnay. All those results indicate that when macromolecules are extracted from the wine, its surface properties are reduced and that the bentonite treatment has an effect similar to macromolecule concentration lowering on the surface properties. Moreover, the effect of the vine variety is coherent with the usual observation by oenologists that the former gives less foam stability than the latter. In conclusion, these results show that there is a positive correlation between the macromolecular concentration and the kinetics of surface tension lowering ŽFig. 5.. They also seem to confirm the effect of bentonite on the wine and to confirm all the same the macromolecular difference between the two vine varieties. The exact nature of the macromolecules retained in the ultraconcentrate is not presently known in detail. Some evidences show that glycoproteins are involved w10x, but some colloids with large enough dimensions may also be candidates. 3.3.2. Ellipsometric characterisation of the adsorption layer Some typical samples studied by tensiometry have also been analysed by ellipsometry. They are wine, UF and UC. The ellipsometric spectrum of UF measured
30
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
Fig. 6. Ellipsometric spectra of ethanolrwater mixture, of wine and of wine ultrafiltrate. Spectra were determined between 250 and 700 nm on a 10.8% ethanol solution in water Ž=.; on Chardonnay Ž`. and on Chardonnay UF Žq.. The spectra of Pinot noir and of Pinot noir UF are very similar to those of Chardonnay. The calculated spectrum of a dioptre between air and a medium with the same index as the ethanolrwater solution is also included Žcontinuous line..
30 min after pouring the sample in the cuvette is very similar to that of a waterrethanol mixture with the alcohol concentration of wine ŽFig. 6.. Contrarily, the spectra of the wine or of UC are quite different. A simulation shows that the spectra of UF or of the waterrethanol mixture are very close to the spectrum of a plain dioptre. Thus the adsorption layer of UF, if any, is of the same order of magnitude as that of ethanol in model systems w13x. This conclusion should be valid only if the adsorption layer is not formed by molecules with the same refractive index as that of the solvent, a situation which is not very likely to occur. Another feature of the spectra of UF and waterrethanol mixture is that the former is red shifted with respect to the latter, this comes from the refractive index of UF ŽChardonnay, n D s 1.3416; Pinot noir, n D s 1.3415. which is slightly larger than that of the model mixture Ž10.8% ethanol in water, n D s 1.3382.. By contrast, the spectrum of the native wine is very different from the spectra of UF or of the ethanolrwater mixture. In particular, the minimum of ⌿ is much larger than in the previous case. This shows unambiguously that an adsorption layer is formed at the winerair interface. The kinetics of evolution of the ellipsometric parameters of the wines, of UF and of UC have been determined at the wavelength of the Brewster conditions ŽFig. 7.. The ellipticity of the UF is positive, indicating that rugosity is the main contribution to that physical quantity w14x. The consequence is that practically no adsorption layer is formed on UF in the time course of the measurement. For the wines and UC, the ellipticity is negative, showing that an adsorption layer has a major
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
31
Fig. 7. Kinetics of evolution of the ellipticity of wine, UF and UC. I, Pinot noir wine; `, Chardonnay wine; q, UF; and ⽧, UC. Measurements were started a few seconds after pouring the sample in the ellipsometer cuvette.
contribution to the ellipsometric measurement. The formation of the adsorption layer seems very rapid since practically half of the signal variation Žwith respect to the UF kinetics. is achieved before the first measurement Ža few seconds after pouring the wine in the cuvette.. The kinetics seem to be completed after 20 min, a time where two spectra recorded consecutively are practically identical. It can also be noted that the ellipticity of Pinot noir increases slower than that of Chardonnay, an observation which is coherent with the difference observed by champagne makers on foam stability between the two vine varieties w15x. In addition, it is surprising to realise that the properties of UC are quite similar to those of wine. The surface tension measurements on wine and UC diluted four times by water are also in favour of quite similar properties between these two kinds of products ŽFig. 4.. Comparing Figs. 4 and 7, it is also visible that the surface tension kinetics Žon wines diluted four times. is much slower than that of the ellipticity variation measured on undiluted wines. Thus, ellipticity has also been monitored on diluted wines and the evolution of the surface properties Žas compared to those of diluted UF which are also constant. is even faster than in the case of undiluted wine Ždata not shown.. Thus the difference of kinetics between surface tension and ellipticity cannot be ascribed to the dilution of wine. The reason may be that the surface in the ellipsometric device is created by pouring wine in the trough while a new surface of wine is created by forming the drop used for surface tension measurement. 3.3.3. Dilational modulus of the adsorption layer During the kinetics of lowering of the surface tension, a sinewave deformation of
32
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
the area of the drop was applied ŽFig. 8.. This procedure allows the calculation of the dilational modulus wEq. Ž3.x at the imposed frequency. The phase angle between strain and stress varies between 5⬚ and 10⬚. Thus the viscous component is low and has not been studied in detail. The modulus measured with UF is very small during the whole time-course of the kinetics. Contrarily, with wine or UC, the modulus increases steadily while the surface tension decreases. Thus, the plot of vs. ⌬␥ Žthe decrease of surface tension. yields a line which is close to a straight line and which looks like a single straight line whatever the sample used ŽFig. 9.. This last point is not completely obvious from our results, but it is clear that both and ⌬␥ should be zero at the same time and that all the lines of Fig. 9 should pass through the origin of the axis, a fact which is probably slightly obscured by the accuracy of determination of and ⌬␥. Nevertheless, as long as all these results fit a single master curve, they seem to indicate that there is no qualitative difference between all the adsorption layers of the samples obtained from different vine varieties or from UC. If this conclusion is true, it would mean that the difference of properties of the adsorption layers is mainly of a quantitative nature, or, in other words, mostly linked with the kinetics of surface properties evolution. The occurrence of a master curve, when plotting vs. ⌬␥ Žor in model systems., has previously been observed with proteins w8x and interpreted as characteristic of the nature of the adsorbed molecule. Moreover, it has been shown that such a master curve is a simple mathematical transform of the state equation and that its slope, assuming a polymer behaviour of the adsorbed molecules is linked to the Flory exponent of the isolated polymer w16x which, itself, is the reciprocal of the fractal dimension of the polymer in the adsorption layer. In the present study, the
Fig. 8. Kinetics of evolution of the surface tension during a sinewave variation of the area of the drop. The drop is made of ultraconcentrated Chardonnay wine diluted four times with water. The kinetics starts immediately after the formation of the drop. 䉫, area of the drop; `, surface tension.
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
33
Fig. 9. Relation between the dilational modulus and the surface tension during the formation of the adsorption layer. Both parameters were measured on the same droplet. ⌬␥ is the difference between the surface tension at t s 0 and at the measuring time. =, Chardonnay; q, Chardonnay UC; ⽧, Pinot noir; B, Pinot noir UC.
slope is 5.5, a value indicative, in model studies of surface properties dominated by a purely two-dimensional behaviour of the polymer and of a two-dimensional conformation intermediate between ‘good solvent’ and ‘ solvent’ conditions w16x. Another point of interest is the absolute value of , which reaches 25᎐30 mNrm, while the surface tension of the wine is of the order of 47 mNrm ŽTable 1.. Although the modulus is measured on diluted wine, it can be noted that it is larger than half the surface tension, a value which is critical for bubbles whose stability is driven by the Laplace equation w3x. Such a high value of the modulus with respect to the surface tension should avoid disproportionation ŽOstwald ripening. of bubbles. 3.4. General discussion 3.4.1. Effect of the adsorption of macromolecules on the surface tension As pointed out in the introduction, the surface tension of wines is so low that adsorption of macromolecules induces only a minute decrease of that physical quantity w6x. However, when macromolecules adsorb at interfaces, their concentration in the bulk has often only a logarithmic effect on the surface tension lowering w17x. Thus it can be expected that the dilution of a hydroalcoholic solution by water increases the surface tension because of ethanol concentration lowering but does not change much the surface properties of the macromolecules, resulting in a larger effect of the macromolecules on the surface tension Žwith respect to the solution without macromolecule adsorption layer.. The occurrence of surface-
34
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
active macromolecules in wine has been confirmed by ellipsometry but is actually detected by surface tension measurement after dilution with water. 3.4.2. Effect of ultrafiltration on the properties of the products As expected, after ultrafiltration, the protein concentration measured in the UF is lowered ŽTable 2.. The measured value could reflect molecules with a molecular mass lower than 10 000 but reacting as proteins in the assay used. Thus the amount of macromolecules with a molecular mass larger than 10 000 was evaluated by subtracting the effect of the low molecular mass compounds. Nevertheless, the UC concentration calculated from the ultrafiltrated volumes or from the protein concentrations are, respectively, 2.5 and 1.6 ŽTable 2.. This points probably to the fact that the ultrafiltration process using large membrane areas Ž1.8 m2 in the device used. may modify the protein pattern of the solutions either by adsorption or by conformation change. This puzzling effect of ultrafiltration on the protein concentration of UC is also probably reflected by the surface tension kinetics of UC which is often lower than that of the native wine and by that of reconstituted wine which is much lower than that of native wine ŽFig. 4. while results larger than those of native wine in the former case and equal to it in the latter case would have been expected if the ultrafiltration procedure had only a sieving effect. In conclusion, it should be remembered that ultrafiltration has some side effects on proteins but these effects do not modify the conclusions drawn from the samples obtained by that procedure. These results, pointing to the low macromolecular content of the UF and to the related effect on the surface concentration of the adsorption layer, are consistent with the lowering of the foaming properties previously observed on wines whose macromolecular content had been lowered by ultrafiltration w5x. Thus, UF seems to Table 2 Protein concentration and ratios of concentration of the samples estimated from protein concentrations or ultrafiltration volumesa Vine variety
Wine sample
Protein concentration Žmgrl.
Macromolecular Žprotein. concentration ratio From protein measurements
From the VUF and VUC values
Chardonnay
Native UC UF
18 Ž10.4. 24 Ž16.4. 7.6 Ž0.
1 Ž1. 1.3 Ž1.6. 0.42 Ž0.
1 2.31 0
Pinot noir
Native UC UF
22 Ž12. 30 Ž20. 10 Ž0.
1 Ž1. 1.4 Ž1.7. 0.45 Ž0.
1 2.56 0
a Protein concentration was determined by the Bradford method. The concentration ratio of macromolecules Žproteins. was calculated with respect to native wine. The measured protein concentration in the UF corresponds to low molecular weight compounds. The values in brackets have been calculated assuming that the protein content of the UF is zero.
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
35
be a product devoid of macromolecular adsorption layer and of foaming properties and could be used as a matrix to test the ability of the various macromolecules of wine to adsorb and to stabilise bubbles. 3.4.3. Use of surface tension measurements to monitor the occurrence of surface-acti¨ e molecules in the wine Since ellipsometric measurements on native wines completely confirm the surface tension measurements performed on diluted wines, these drop tensiometer measurements are good quantitative evidences for the occurrence of surface-active Žmacro.molecules in the wine samples. Another point is that the occurrence of a master curve r⌬␥ ŽFig. 9. implies that the modulus varies as ⌬␥ and consequently as the adsorption of molecules at the interface. Thus, the measurement of ⌬␥, whatever the sample used, may give a measure of . This may have important practical applications. Moreover, this tensiometer procedure needs a less sophisticated equipment than the ellipsometric one. In addition, it should be pointed out that the surface activity measured by this procedure is higher for ‘Chardonnay’, than for ‘Pinot noir’, just as oenologists recognise that the former has usually more foam than the latter w15x.
Acknowledgements Our warm thanks to B. Monties for his support to this program, to N. Puff for his help in the experimental work and to European Community Contract no. IN 10381 D for financial support.
References w1x A. Maujean, P. Poinsaut, H. Dantan, F. Brissonnet, E. Cossiez, Bull. OIV 61 Ž701r712. Ž1990. 405. w2x R. Marchal, G. Bocquillon Liger-Belair, L. Berthier, F. Brissonnet, P. Jeandet, A. Maujean, Meeting ‘Bubbles in Food’, UMIST, Manchester UK, June 1998. w3x J. Lucassen, in: E.H. Lucassen-Reynders ŽEd.., Anionic Surfactants, Physical Chemistry of Surfactant Action, Ch. 6, Marcel Dekker, Inc, New York, 1981. w4x A. Dussaud, M. Vignes-Adler, J. Colloid Interface Sci. 167 Ž1994. 266. w5x J. Malvy, B. Robillard, B. Duteurtre, Sci. Aliments 14 Ž1994. 87. w6x N. Puff, A. Cagna, V. Aguie-Beghin, R. Douillard, J. Colloid Interface Sci. 208 Ž1998. 405. ´ ´ w7x S. Labourdenne, N. Gaudry-Rolland, S. Letellier et al., Chem. Phys. Lipids 71 Ž1994. 163. w8x J. Benjamins, A. Cagna, E.H. Lucassen-Reynders, Colloids Surf. A: Physicochem. Eng. Aspects 114 Ž1996. 245. w9x R.M.A. Azzam, N.M. Bashara, Ellipsometry and Polarized Light, North Holland, Amsterdam, 1977. w10x R. Marchal, V. Seguin, A. Maujean, Am. J. Enol. Vitic. 48 Ž1997. 303. w11x G. Putz, J. Goerke, H.W. Taeusch, J.A. Clements, J. Appl. Physiol. 76 Ž1994. 1425. w12x R.M. Prokop, A. Jyoti, M. Eslamian et al., Colloids Surf. A: Physicochem. Eng. Aspects 131 Ž1998. 231.
36
w13x w14x w15x w16x w17x
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36 Z.X. Li, J.R. Lu, D.A. Styrkas, R.K. Thomas, A.R. Rennie, J. Penfold, Mol. Phys. 80 Ž1993. 925. J. Meunier, J. Phys. 48 Ž1987. 1819. B. Robillard, L. Viaux, B. Duteurtre, Vigneron champenois, No. 2 fevrier 1995, 17. ´ V. Aguie-Beghin, E. Leclerc, M. Daoud, R. Douillard, J. Colloid Interface Sci. 214 Ž1999. 143. ´ ´ B. Harzallah, V. Aguie-Beghin, R. Douillard, L. Bosio, Int. J. Biol. Macromol. 23 Ž1998. 73. ´ ´
Characterisation by drop tensiometry and by ellipsometry of the adsorption layer formed at the airrchampagne wine interface N. Peron ´ a,b, A. Cagnaa , M. Valade c , R. Marchal d, b A. Maujeand, B. Robillard e, V. Aguie-Beghin , ´ ´ b,U R. Douillard a I.T. Concept, Parc de Chancolan, 69770 Longessaigne, France INRA, Equipe de Biochimie des Macromolecules Vegetales, CRA, 2 Espl. R. Garros, BP 224, ´ ´´ 51686 Reims Cedex 2, France c Comite´ Interprofessionnel du Vin de Champagne, 5 rue Henri-Martin, BP 135, 51204 Epernay Cedex, France d Laboratoire d’Œnologie, URCA, Moulin de la Housse, BP 1039, 51687 Reims Cedex 2, France e Moet ¨ et Chandon, 6 rue Croix de Bussy, 51200 Epernay, France b
Abstract A foam ring composed of small bubbles on the surface of a champagne glass is one of its hallmarks. The equilibrium state of that ring is linked with the rate of formation and of disappearance of bubbles. The stability of bubbles is usually ascribed to the occurrence and to the properties of an adsorption layer formed at the gasrliquid interface. Our goal is to characterise such an adsorption layer at the gasrwine interface in order to understand its role in bubble stability. Alcohol in wine lowers the surface tension to 49 mNrm. The adsorption of other molecules may cause a further decrease of 2 mNrm. Such a situation makes the study of adsorption by surface tension measurement inaccurate. To overcome this problem, we have diluted the wine four times with water before its surface tension
U
Corresponding author. Tel.: q33-3-26-77-35-94; fax: q33-3-26-77-35-99. E-mail address: [email protected] ŽR. Douillard.. 0001-8686r00r$ - see front matter 䊚 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 1 - 8 6 8 6 Ž 0 0 . 0 0 0 3 9 - 7
20
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
measurement by pendant drop shape analysis. In these conditions, ethanol lowers the surface tension to 64 mNrm and the adsorption of other molecules of the wine can be monitored over 6᎐8 mNrm. The usual behaviour of such a diluted wine is a lowering of the surface tension during at least 20 min after drop formation. Since the role of macromolecules on the foaming properties of wine had been previously observed, we have chosen to evaluate the effect of this fraction of the wine molecules on its surface properties. Thus, wines were ultrafiltrated on a membrane with a 10 000 molecular mass cut-off. The ultrafiltrate ŽUF. does not show any decrease of its surface tension over a 20-min period while the ultraconcentrate ŽUC. has a kinetics similar to that of unfiltered wine. Mixtures of UF and UC have behaviours intermediate between those of these products. A technological treatment of the wine with bentonite, believed to lower the content of macromolecules, yields a wine similar to UF. The effect of ultrafiltration was also analysed by spectroscopic ellipsometry. UF has a spectrum similar to that of a waterralcohol mixture with the same ethanol content and its ellipticity is stable during at least 20 min. On the contrary, wine or UC show spectra with the features of an adsorption layer and those characteristics increase during more than 20 min. Two varieties of vine were compared: ‘Chardonnay’ and ‘Pinot noir’. The former is known to have better foaming properties than the latter. Its surface properties measured in this study are also more pronounced than those of Pinot noir. However, the representation of the dilational modulus against the surface pressure Žwhich, in some instances, may be a mathematical transformation of the state equation. puts all the samples Žwines, UF and UC of each. on the same master curve, a fact in favour of a common nature for all the adsorption layers. It can be concluded that surface properties of champagne wines are mostly determined by ethanol and by macromolecules with a molecular mass larger than 10 000. Moreover, the adsorption layers seem to have the same nature, irrespective of the vine variety and of the concentration ratio of the wine. 䊚 2000 Elsevier Science B.V. All rights reserved. Keywords: Foaming behaviour; Airrchampagne interface; Macromolecules; Pendant drop method; Ellipsometry
Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Wine samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Surface tension measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Ellipsometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Protein quantification by the Bradford method . . . . . . . . . . . . . . . . . . 3. Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Effect of wetting on reproducibility . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Water dilution of wine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Effects of macromolecule concentration on the surface properties . . . . . 3.3.1. Kinetics of formation of the adsorption layer . . . . . . . . . . . . . . 3.3.2. Ellipsometric characterisation of the adsorption layer . . . . . . . . . 3.3.3. Dilational modulus of the adsorption layer . . . . . . . . . . . . . . . . 3.4. General discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1. Effect of the adsorption of macromolecules on the surface tension
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. 21 . 21 . 21 . 22 . 23 . 24 . 25 . 26 . 26 . 28 . 29 . 29 . 31 . 33 . 33
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
21
3.4.2. Effect of ultrafiltration on the properties of the products . . . . . . . . . 34 3.4.3. Use of surface tension measurements to monitor the occurrence of surface-active molecules in the wine . . . . . . . . . . . . . . . . . . . . . . . 35 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1. Introduction Small bubbles forming a regular ring around the glass are appreciated by consumers and make the hallmark of champagne. The ability of champagne to form a nice foam is often appreciated using a sparging procedure in a glass tube and measuring the properties of the foam with the ‘Mosalux’ device w1,2x. However, the foam measured in that way is quite different of the subtle collar formed in a glass. Moreover, the ‘Mosalux’ procedure has never been claimed to measure the ability of a ‘base wine’ Žthe wine before the second fermentation. to produce an elegant bubble ring. From a more basic standpoint, the physical phenomena involved in bubble and foam stability are thought to be linked mainly with the properties of the interfaces between wine and gas w3,4x. Thus we have chosen to investigate and characterise the surface properties of base wines in view of studying the correlations between such properties and the foam behaviour of the corresponding sparkling wines. In this study, we take advantage of the fact that treatments such as ultrafiltration w5x or bentonite application w1x are known to decrease drastically the foaming behaviour of the wine: a related decrease of the surface properties is expected. A recent study of protein adsorption in hydro-alcoholic solutions w6x showed that adsorption of macromolecules has not always a strong lowering effect on surface tension. Thus, this study is based on both surface tension and ellipsometry measurements in order to detect a surface layer and to measure its effect on surface tension.
2. Materials and methods 2.1. Wine samples Wines were produced from two vine varieties grown by CIVC ŽComite ´ Interprofessionnel du Vin de Champagne, Epernay.: ‘Chardonnay’ or ‘Pinot noir’ during the 1997 harvest. They were used as still wines before the second fermentation. Ultrafiltration Žultraconcentration . of wine was performed by tangential ultrafiltration on a 1.8-m2 hollow fibre device ŽInceltech, Toulouse, France. with a 10 000 nominal molecular mass cut-off. The ultrafiltration membrane was made of hydrophilic polysulfone. Before ultrafiltration of the volume needed Ž0.8 l., the device was rinsed with 1.6 l wine. The transmembrane pressure was 0.7 atm. The relative
22
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
concentration factor ŽRCFUC . of the ultraconcentrate ŽUC. was calculated with respect to native wine as: RCFUC s Ž VUF q VUC . rVUC
Ž1.
where VUF and VUC are the volumes of ultrafiltrate ŽUF. and of UC of the experiment. RCF of each sample was calculated from the volumes of wine, of UC and of UF mixed to prepare the sample and from their respective RCF: 1 for native wine, RCFUC for UC and zero for UF. Bentonite precipitation has been achieved with 1 grl. Bentonite was first swollen with 20 times its weight of ultrapure water during 24 h at room temperature under mild stirring. Wine and bentonite were mixed and stirred during 20 min before centrifugation Ž15 000 = g, 20 min.. When needed, wine samples were diluted with ultrapure water. All glass vessels were treated with chromosulfuric acid before use. Syringe needles were thoroughly rinsed with ultrapure ethanol and water. 2.2. Surface tension measurements All static and dynamic surface tension measurements were performed with a drop Žbubble. tensiometer from IT Concept, Longessaigne, France w6,7x. A complete description of the experimental set-up has been given by Benjamins et al. w8x. An axisymmetric drop or bubble is formed at the tip of the needle of a syringe whose plunger position is driven by a computer. The image of the bubble Ždrop. is taken from a CCD camera and digitised. The interfacial tension ␥ is calculated by analysing the profile of the bubble according to the Laplace equation: Ž 1rx . d Ž xsin . rd x s 2rb y cz
Ž2.
The origin of the coordinates is at the bubble apex; x and z are the Cartesian coordinates at any point of the bubble profile, b is the radius of curvature at the bubble Ždrop. apex, is the angle of the tangent to the bubble Ždrop. profile and c s 2ra2 , where a is the capillary constant Ž a s w2 ␥rŽ g .x1r2 , where is the difference of density between the two phases, and g the acceleration of gravity.. The area of the drop Žbubble. and the surface tension are calculated several times per second. This set-up was made suitable to measure the surface dilational modulus w3x: s d␥rdln A
Ž3.
where A is the surface area of the bubble Ždrop.. This was achieved by fluctuating sinusoidally the area of the drop Žbubble. at a frequency of 0.1 Hz and a relative amplitude of 0.07 which is the lowest value with an acceptable signal-to-noise ratio and where the surface tension amplitude response is in the linear range. No significant change of the modulus was observed by changing the frequency from 0.01 to 0.1 Hz. Moreover, a 10-s period is short enough compared to the duration
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
23
of the adsorption experiment which is of the order of 20 min. The modulus was calculated using a sliding window with a 10-s width Žone period., using the A and ␥ variations. A Fourier transform of the signal is used to extract the main component and avoid the small perturbations linked to the inaccuracies in the movement of the plunger. The surface tension was calculated as the mean value during the same period. All experiments were performed in an air-conditioned room at a fixed temperature of 20 " 1⬚C. Drop formation at the tip of a needle was achieved in a small confining cell Žf 250 ml. fitted with two parallel windows and containing a few millilitres of wine coating its bottom part. This allows for a rapid equilibration of the volume around the drop with wine saturating vapours and avoids evaporation from the drop. The plunger of the syringe is allowed to enter through a toric seal ensuring a gas tight cell. In most experiments, the lowering of the surface tension w ⌬␥Ž t . s ␥Ž0. y ␥Ž t .x was plotted against time instead of the surface pressure which cannot be calculated accurately since the surface tension of the ‘pure’ solvent is not a defined physical quantity in the case of wine. In order to begin the measurements as close as possible to the solvent tension, two drops Žbubbles . were quickly expelled before forming the measuring drop Žbubble.. 2.3. Ellipsometry All measurements have been done using a spectroscopic phase modulated ellipsometer ŽUVISEL, Jobin Yvon, Longjumeau, France.. It is equipped with a xenon arc lamp. Both the polariser and the analyser are set to the 45⬚ configuration angle. The photoelastic modulator, activated at the 50-kHz frequency, is set to the 0⬚ configuration orientation. The spectroscopic measurements are monitored between 250 nm and 700 nm. The incidence angle is set to 53.6⬚. All measurements were done at the airrliquid interface in an air-conditioned room at 20 " 1⬚C. The two ellipsometric angles ⌿ and ⌬ w9x, are linked to the two reflectivity coefficients rp and rs , respectively, in the directions parallel and perpendicular to the incidence plane, by: rprrs s tan⌿exp Ž i⌬ .
Ž4.
The fixed wavelength chosen for the kinetics measurements corresponds to the Brewster conditions of the pure substrate defined by: ⌬ s r2
Ž5.
The coefficient of ellipticity of the adsorption layer measured at the Brewster conditions for the substrate is defined as: s tan⌿sin⌬
Ž6.
The index of refraction n D was measured with an Abbe refractometer ŽRF490, Prolabo, Paris..
24
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
Fig. 1. Effect of the experimental set-up on the reproducibility of the surface tension measurements. Ža. Surface tension was measured using bubbles formed at the tip of a stainless steel needle. Žb. Surface tension was measured using droplets formed at the tip of a Teflon-coated needle.
2.4. Protein quantification by the Bradford method The Bradford method was employed as previously described w10x: 0.2 ml Bio-Rad Protein Assay reagent ŽCoomassie Brilliant Blue, CBB. is added to the 0.4-ml sample. The blue coloration was measured at 595 nm using a spectrophotometer ŽSecoman S250. after 60-min contact when the colour was stable. The protein content was calculated with respect to a calibration curve with bovine serum albumin ŽBSA. Ž0᎐20 mgrl.. Each value corresponded to the average of three measurements. The standard deviation was less than 2%.
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
25
Fig. 2. Effect of the coating of the needle on its wetting by the wine samples. The horizontal line materialises the tip of the needle. Ža. Stainless steel needle; an ascent of the contact line of the drop in an asymmetrical way is visible. Žb. Teflon-coated needle; the contact line of the drop is located at the tip of the needle.
3. Results and discussion Several problems were encountered when trying to analyse the effect of the wine macromolecules on its surface properties. Possible solutions to them are shown here. First is presented the effect of wetting of the needle on the reproducibility of surface tension measurements, then the occurrence of adsorbed material is ob-
26
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
scured by the low surface tension imposed by ethanol in wine and a dilution by water may solve this problem. To confirm the effect of high molecular mass molecules on the surface properties of wine, a comparison of surface tension, surface dilational modulus and ellipsometric parameters was made between wine and samples derived from the base wine particularly by ultrafiltration. 3.1. Effect of wetting on reproducibility At first, surface tension was measured using the bubble procedure which is not sensitive to evaporation at the winerbubble interface. Using a stainless steel needle and wine diluted fourfold by water, the kinetics of evolution of the surface tension of a bubble formed in the liquid was poorly reproducible, as illustrated in Fig. 1a. However, when the kinetics was pursued more than 1 h, the standard deviation of the measured surface pressures was less. Using a drop formed at the tip of a Teflon coated needle ŽFig. 2b., in the gas-tight cell, the reproducibility was much better ŽFig. 1b.. It can be seen using droplets formed with a non-Teflon coated needle that the diluted wine wets the surface of the needle ŽFig. 2a.. That wetting is not completely reproducible and the height of the needle covered by a liquid film is not constant. Thus, it is likely that using a bubble, wetting occurs inside the stainless steel needle and is not visible with the CCD camera equipment of the tensiometer. It is probable that when such a wetting occurs, a film leakage may also happen on the needle, as it has previously been noticed w11,12x, inducing the lack of reproducibility in the surface tension kinetics. In conclusion, care should be given to such phenomena and an evaluation of the wetting is recommended using the drop procedure. 3.2. Water dilution of wine When a new surface is formed at the gasrwine interface, it is mostly free of adsorbed molecules and the surface tension should be that of the solvent. In the case of wine, ethanol adsorbs quickly and, with the procedure used, its adsorption kinetics cannot be resolved. Thus, at the beginning of the kinetics, surface tension is mostly imposed by the concentration of ethanol. Using native wine, it was observed that the surface tension when the drop has just been formed is approximately 2.5 mNrm less than that of the corresponding waterrethanol mixture ŽTable 1.. This seems in favour of molecules other than ethanol adsorbing quickly at the interface or of solvent conditions for ethanol different from those found in pure water. This point stresses the question of the nature of the ‘solvent’ when considering wine. In addition, this unresolved question makes difficult the use of the physical quantity ‘surface pressure’ which refers to the solvent surface tension. Nevertheless, the surface tension decreases less than 0.5 mNrm in the first 20 min of formation ŽFig. 3.. This low decrease does not allow an accurate characterisation of the surface-active molecules of the wine and of the dilational modulus of the interface. Thus, measurements were performed after a dilution of the wine by water to decrease the effect of ethanol on the surface tension ŽFig. 3.. In these
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
27
Table 1 Ethanol concentration and surface tension of wines, of fourfold diluted wines, and of hydroalcoholic solutions with the same ethanol concentration a
Chardonnay Pinot noir Chardonnay, 1r4 dilution Pinot noir, 1r4 dilution a
Ethanol concentration Ž%.
Surface tension ŽmNrm. Wine
Ethanol᎐water solution
10.8 11.6 2.7 2.9
46.9 46.1 60.9 61.2
49.4 48.6 64.6 64.1
Surface tension was measured immediately after forming the drop.
conditions it is clear that surface-active molecules adsorb and lower the surface tension to a value significantly smaller than that measured at the beginning of the kinetics Ž2᎐4 mNrm.. Several dilutions were tried. It is likely that when the wine is diluted 10 times, the concentration of surface-active molecules is so low that the kinetics is slowed down. In the case of a fourfold dilution, the lowering of surface tension is larger than for a twofold dilution. For these reasons, it seems finally convenient to use a fourfold dilution for all samples. An important conclusion of these water dilution experiments is that, apart from ethanol, surface-active molecules occur in champagne base wine and that it is possible to characterise them by surface tension measurements.
Fig. 3. Effect of wine dilution on its surface tension kinetics. Chardonnay wine was eventually diluted with ultrapure water and the kinetics of its surface tension recorded during 20 min. The dilution is quoted on each curve.
28
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
Fig. 4. Kinetics of surface tension of various wine samples. All samples were diluted four times with ultrapure water. `, native wine; q, ultraconcentrate; ⽧, ultrafiltrate; I, bentonite-treated wine; =, reconstituted wine from UF and UC; ', native wine diluted twice by UF. Ža. Chardonnay wine; Žb. Pinot noir wine.
3.3. Effects of macromolecule concentration on the surface properties To characterise the effect of macromolecules on the surface properties, a study has been done comparing the properties of native base wine, ultrafiltrate ŽUF., ultraconcentrate ŽUC., reconstituted wine from UF and UC with a macromolecular concentration identical to that of the native wine, native wine diluted twice by UF, and bentonite-treated wine.
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
29
Fig. 5. Effect of the relative macromolecular concentration on the surface tension reached after 5 min. The zero time of the kinetics is the formation of the drop at the needle tip. The macromolecular concentration is relative to the macromolecular concentration of the native still wine. All samples were prepared by mixing UF and UC. From `, Chardonnay; =, Pinot noir.
3.3.1. Kinetics of formation of the adsorption layer The kinetics of evolution of the surface tension of the various samples show that UF has a very small decrease of surface tension Ž ⌬␥ ., while native wine or UC have a significant and rather similar decrease. Wine diluted twice by UF has an intermediate behaviour and bentonite-treated wine is close to UF ŽFig. 4.. Moreover, all the results obtained with Pinot noir wine show a smaller decrease of the surface tension than those obtained with Chardonnay. All those results indicate that when macromolecules are extracted from the wine, its surface properties are reduced and that the bentonite treatment has an effect similar to macromolecule concentration lowering on the surface properties. Moreover, the effect of the vine variety is coherent with the usual observation by oenologists that the former gives less foam stability than the latter. In conclusion, these results show that there is a positive correlation between the macromolecular concentration and the kinetics of surface tension lowering ŽFig. 5.. They also seem to confirm the effect of bentonite on the wine and to confirm all the same the macromolecular difference between the two vine varieties. The exact nature of the macromolecules retained in the ultraconcentrate is not presently known in detail. Some evidences show that glycoproteins are involved w10x, but some colloids with large enough dimensions may also be candidates. 3.3.2. Ellipsometric characterisation of the adsorption layer Some typical samples studied by tensiometry have also been analysed by ellipsometry. They are wine, UF and UC. The ellipsometric spectrum of UF measured
30
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
Fig. 6. Ellipsometric spectra of ethanolrwater mixture, of wine and of wine ultrafiltrate. Spectra were determined between 250 and 700 nm on a 10.8% ethanol solution in water Ž=.; on Chardonnay Ž`. and on Chardonnay UF Žq.. The spectra of Pinot noir and of Pinot noir UF are very similar to those of Chardonnay. The calculated spectrum of a dioptre between air and a medium with the same index as the ethanolrwater solution is also included Žcontinuous line..
30 min after pouring the sample in the cuvette is very similar to that of a waterrethanol mixture with the alcohol concentration of wine ŽFig. 6.. Contrarily, the spectra of the wine or of UC are quite different. A simulation shows that the spectra of UF or of the waterrethanol mixture are very close to the spectrum of a plain dioptre. Thus the adsorption layer of UF, if any, is of the same order of magnitude as that of ethanol in model systems w13x. This conclusion should be valid only if the adsorption layer is not formed by molecules with the same refractive index as that of the solvent, a situation which is not very likely to occur. Another feature of the spectra of UF and waterrethanol mixture is that the former is red shifted with respect to the latter, this comes from the refractive index of UF ŽChardonnay, n D s 1.3416; Pinot noir, n D s 1.3415. which is slightly larger than that of the model mixture Ž10.8% ethanol in water, n D s 1.3382.. By contrast, the spectrum of the native wine is very different from the spectra of UF or of the ethanolrwater mixture. In particular, the minimum of ⌿ is much larger than in the previous case. This shows unambiguously that an adsorption layer is formed at the winerair interface. The kinetics of evolution of the ellipsometric parameters of the wines, of UF and of UC have been determined at the wavelength of the Brewster conditions ŽFig. 7.. The ellipticity of the UF is positive, indicating that rugosity is the main contribution to that physical quantity w14x. The consequence is that practically no adsorption layer is formed on UF in the time course of the measurement. For the wines and UC, the ellipticity is negative, showing that an adsorption layer has a major
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
31
Fig. 7. Kinetics of evolution of the ellipticity of wine, UF and UC. I, Pinot noir wine; `, Chardonnay wine; q, UF; and ⽧, UC. Measurements were started a few seconds after pouring the sample in the ellipsometer cuvette.
contribution to the ellipsometric measurement. The formation of the adsorption layer seems very rapid since practically half of the signal variation Žwith respect to the UF kinetics. is achieved before the first measurement Ža few seconds after pouring the wine in the cuvette.. The kinetics seem to be completed after 20 min, a time where two spectra recorded consecutively are practically identical. It can also be noted that the ellipticity of Pinot noir increases slower than that of Chardonnay, an observation which is coherent with the difference observed by champagne makers on foam stability between the two vine varieties w15x. In addition, it is surprising to realise that the properties of UC are quite similar to those of wine. The surface tension measurements on wine and UC diluted four times by water are also in favour of quite similar properties between these two kinds of products ŽFig. 4.. Comparing Figs. 4 and 7, it is also visible that the surface tension kinetics Žon wines diluted four times. is much slower than that of the ellipticity variation measured on undiluted wines. Thus, ellipticity has also been monitored on diluted wines and the evolution of the surface properties Žas compared to those of diluted UF which are also constant. is even faster than in the case of undiluted wine Ždata not shown.. Thus the difference of kinetics between surface tension and ellipticity cannot be ascribed to the dilution of wine. The reason may be that the surface in the ellipsometric device is created by pouring wine in the trough while a new surface of wine is created by forming the drop used for surface tension measurement. 3.3.3. Dilational modulus of the adsorption layer During the kinetics of lowering of the surface tension, a sinewave deformation of
32
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
the area of the drop was applied ŽFig. 8.. This procedure allows the calculation of the dilational modulus wEq. Ž3.x at the imposed frequency. The phase angle between strain and stress varies between 5⬚ and 10⬚. Thus the viscous component is low and has not been studied in detail. The modulus measured with UF is very small during the whole time-course of the kinetics. Contrarily, with wine or UC, the modulus increases steadily while the surface tension decreases. Thus, the plot of vs. ⌬␥ Žthe decrease of surface tension. yields a line which is close to a straight line and which looks like a single straight line whatever the sample used ŽFig. 9.. This last point is not completely obvious from our results, but it is clear that both and ⌬␥ should be zero at the same time and that all the lines of Fig. 9 should pass through the origin of the axis, a fact which is probably slightly obscured by the accuracy of determination of and ⌬␥. Nevertheless, as long as all these results fit a single master curve, they seem to indicate that there is no qualitative difference between all the adsorption layers of the samples obtained from different vine varieties or from UC. If this conclusion is true, it would mean that the difference of properties of the adsorption layers is mainly of a quantitative nature, or, in other words, mostly linked with the kinetics of surface properties evolution. The occurrence of a master curve, when plotting vs. ⌬␥ Žor in model systems., has previously been observed with proteins w8x and interpreted as characteristic of the nature of the adsorbed molecule. Moreover, it has been shown that such a master curve is a simple mathematical transform of the state equation and that its slope, assuming a polymer behaviour of the adsorbed molecules is linked to the Flory exponent of the isolated polymer w16x which, itself, is the reciprocal of the fractal dimension of the polymer in the adsorption layer. In the present study, the
Fig. 8. Kinetics of evolution of the surface tension during a sinewave variation of the area of the drop. The drop is made of ultraconcentrated Chardonnay wine diluted four times with water. The kinetics starts immediately after the formation of the drop. 䉫, area of the drop; `, surface tension.
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
33
Fig. 9. Relation between the dilational modulus and the surface tension during the formation of the adsorption layer. Both parameters were measured on the same droplet. ⌬␥ is the difference between the surface tension at t s 0 and at the measuring time. =, Chardonnay; q, Chardonnay UC; ⽧, Pinot noir; B, Pinot noir UC.
slope is 5.5, a value indicative, in model studies of surface properties dominated by a purely two-dimensional behaviour of the polymer and of a two-dimensional conformation intermediate between ‘good solvent’ and ‘ solvent’ conditions w16x. Another point of interest is the absolute value of , which reaches 25᎐30 mNrm, while the surface tension of the wine is of the order of 47 mNrm ŽTable 1.. Although the modulus is measured on diluted wine, it can be noted that it is larger than half the surface tension, a value which is critical for bubbles whose stability is driven by the Laplace equation w3x. Such a high value of the modulus with respect to the surface tension should avoid disproportionation ŽOstwald ripening. of bubbles. 3.4. General discussion 3.4.1. Effect of the adsorption of macromolecules on the surface tension As pointed out in the introduction, the surface tension of wines is so low that adsorption of macromolecules induces only a minute decrease of that physical quantity w6x. However, when macromolecules adsorb at interfaces, their concentration in the bulk has often only a logarithmic effect on the surface tension lowering w17x. Thus it can be expected that the dilution of a hydroalcoholic solution by water increases the surface tension because of ethanol concentration lowering but does not change much the surface properties of the macromolecules, resulting in a larger effect of the macromolecules on the surface tension Žwith respect to the solution without macromolecule adsorption layer.. The occurrence of surface-
34
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
active macromolecules in wine has been confirmed by ellipsometry but is actually detected by surface tension measurement after dilution with water. 3.4.2. Effect of ultrafiltration on the properties of the products As expected, after ultrafiltration, the protein concentration measured in the UF is lowered ŽTable 2.. The measured value could reflect molecules with a molecular mass lower than 10 000 but reacting as proteins in the assay used. Thus the amount of macromolecules with a molecular mass larger than 10 000 was evaluated by subtracting the effect of the low molecular mass compounds. Nevertheless, the UC concentration calculated from the ultrafiltrated volumes or from the protein concentrations are, respectively, 2.5 and 1.6 ŽTable 2.. This points probably to the fact that the ultrafiltration process using large membrane areas Ž1.8 m2 in the device used. may modify the protein pattern of the solutions either by adsorption or by conformation change. This puzzling effect of ultrafiltration on the protein concentration of UC is also probably reflected by the surface tension kinetics of UC which is often lower than that of the native wine and by that of reconstituted wine which is much lower than that of native wine ŽFig. 4. while results larger than those of native wine in the former case and equal to it in the latter case would have been expected if the ultrafiltration procedure had only a sieving effect. In conclusion, it should be remembered that ultrafiltration has some side effects on proteins but these effects do not modify the conclusions drawn from the samples obtained by that procedure. These results, pointing to the low macromolecular content of the UF and to the related effect on the surface concentration of the adsorption layer, are consistent with the lowering of the foaming properties previously observed on wines whose macromolecular content had been lowered by ultrafiltration w5x. Thus, UF seems to Table 2 Protein concentration and ratios of concentration of the samples estimated from protein concentrations or ultrafiltration volumesa Vine variety
Wine sample
Protein concentration Žmgrl.
Macromolecular Žprotein. concentration ratio From protein measurements
From the VUF and VUC values
Chardonnay
Native UC UF
18 Ž10.4. 24 Ž16.4. 7.6 Ž0.
1 Ž1. 1.3 Ž1.6. 0.42 Ž0.
1 2.31 0
Pinot noir
Native UC UF
22 Ž12. 30 Ž20. 10 Ž0.
1 Ž1. 1.4 Ž1.7. 0.45 Ž0.
1 2.56 0
a Protein concentration was determined by the Bradford method. The concentration ratio of macromolecules Žproteins. was calculated with respect to native wine. The measured protein concentration in the UF corresponds to low molecular weight compounds. The values in brackets have been calculated assuming that the protein content of the UF is zero.
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36
35
be a product devoid of macromolecular adsorption layer and of foaming properties and could be used as a matrix to test the ability of the various macromolecules of wine to adsorb and to stabilise bubbles. 3.4.3. Use of surface tension measurements to monitor the occurrence of surface-acti¨ e molecules in the wine Since ellipsometric measurements on native wines completely confirm the surface tension measurements performed on diluted wines, these drop tensiometer measurements are good quantitative evidences for the occurrence of surface-active Žmacro.molecules in the wine samples. Another point is that the occurrence of a master curve r⌬␥ ŽFig. 9. implies that the modulus varies as ⌬␥ and consequently as the adsorption of molecules at the interface. Thus, the measurement of ⌬␥, whatever the sample used, may give a measure of . This may have important practical applications. Moreover, this tensiometer procedure needs a less sophisticated equipment than the ellipsometric one. In addition, it should be pointed out that the surface activity measured by this procedure is higher for ‘Chardonnay’, than for ‘Pinot noir’, just as oenologists recognise that the former has usually more foam than the latter w15x.
Acknowledgements Our warm thanks to B. Monties for his support to this program, to N. Puff for his help in the experimental work and to European Community Contract no. IN 10381 D for financial support.
References w1x A. Maujean, P. Poinsaut, H. Dantan, F. Brissonnet, E. Cossiez, Bull. OIV 61 Ž701r712. Ž1990. 405. w2x R. Marchal, G. Bocquillon Liger-Belair, L. Berthier, F. Brissonnet, P. Jeandet, A. Maujean, Meeting ‘Bubbles in Food’, UMIST, Manchester UK, June 1998. w3x J. Lucassen, in: E.H. Lucassen-Reynders ŽEd.., Anionic Surfactants, Physical Chemistry of Surfactant Action, Ch. 6, Marcel Dekker, Inc, New York, 1981. w4x A. Dussaud, M. Vignes-Adler, J. Colloid Interface Sci. 167 Ž1994. 266. w5x J. Malvy, B. Robillard, B. Duteurtre, Sci. Aliments 14 Ž1994. 87. w6x N. Puff, A. Cagna, V. Aguie-Beghin, R. Douillard, J. Colloid Interface Sci. 208 Ž1998. 405. ´ ´ w7x S. Labourdenne, N. Gaudry-Rolland, S. Letellier et al., Chem. Phys. Lipids 71 Ž1994. 163. w8x J. Benjamins, A. Cagna, E.H. Lucassen-Reynders, Colloids Surf. A: Physicochem. Eng. Aspects 114 Ž1996. 245. w9x R.M.A. Azzam, N.M. Bashara, Ellipsometry and Polarized Light, North Holland, Amsterdam, 1977. w10x R. Marchal, V. Seguin, A. Maujean, Am. J. Enol. Vitic. 48 Ž1997. 303. w11x G. Putz, J. Goerke, H.W. Taeusch, J.A. Clements, J. Appl. Physiol. 76 Ž1994. 1425. w12x R.M. Prokop, A. Jyoti, M. Eslamian et al., Colloids Surf. A: Physicochem. Eng. Aspects 131 Ž1998. 231.
36
w13x w14x w15x w16x w17x
N. Peron ´ et al. r Ad¨ ances in Colloid and Interface Science 88 (2000) 19᎐36 Z.X. Li, J.R. Lu, D.A. Styrkas, R.K. Thomas, A.R. Rennie, J. Penfold, Mol. Phys. 80 Ž1993. 925. J. Meunier, J. Phys. 48 Ž1987. 1819. B. Robillard, L. Viaux, B. Duteurtre, Vigneron champenois, No. 2 fevrier 1995, 17. ´ V. Aguie-Beghin, E. Leclerc, M. Daoud, R. Douillard, J. Colloid Interface Sci. 214 Ž1999. 143. ´ ´ B. Harzallah, V. Aguie-Beghin, R. Douillard, L. Bosio, Int. J. Biol. Macromol. 23 Ž1998. 73. ´ ´