Use of PVI-PVP copolymer for wine metal reduction effects on wine characteristics

Summary
We studied the influence of an adsorbent PVI-PVP resin (a copolymer of vinylimidazole and vinylpyrro-lidone), on the removal of heavy metals in wines, main-ly copper (Cu), iron (Fe), lead (Pb), cadmium (Cd) and aluminium (AI). The study also ipvestigated the influ-ence of PVI-PVP on the physical-chemical and sensory characteristics of white and red wines, comparing its effect when applied in the must and in the wine. The removal of metals was more effective when PVI-PVP was applied to the wine than to the must. The removal of Fe and Pb was more effective in white wines than in red wines, while the removal of Cu and Al was higher in red wines. In general, the higher the PVI-PVP dose,the greater the quantity of metallic elements (copper,iron, lead and aluminium) that are removed.PVI-PVP had a minor effect on phenolie composition. The wines showed some decrease in total acidity and an increase in pH with PVI-PVP. The application of PVI-PVP at the dose rates employed here did not affect the wine’ssensory characteristics significantly.

Besides its effect on the metals, PVI-PVP seems to have a relevant action in wine clarification. Here, as well as decreasing the level of hydroxycinnamic acid derivatives,mainly caftaric acid (MATTIVI et al. 1994, 2000; NICOLINI etal. 2004; EDER et al. 2001,2003) it also minimises browning.
The treatment of musts and/or wines with PVI-PVP for lowering levels of metals such as Fe, Cu, Zn and Al, has been under discussion at OIV by experts under both the“Wine Technology”and, also the“Food Security” groups.
The authorisation dossier is currently under evaluation.
In May 2006, the French Agency of Food Security in an advice (AFSSA 2006)commented that utilisation of PVI-PVP at a dosage of 80 g·hl’ did not appear to present any risk to consumer health.
The aim of this work was to study the influence of PVI-PVP on metal reduction (Fe, Cu, Pb, Cd, Al) and, es-pecially to confirm that there were no important changes in the chemical and sensory characteristics of either white or red wines. We also wanted to compare its effects on wine characteristics when it was used either in the must or, later in the wine.
Material and Methods
Fining treatment applied to must and wine: Wines: White wines were prepared using grapes of the variety ‘FernÃo Pires’ (Vitis vinifera L.), and red wines using ‘CastelÃo’ (Vitis vinifera L.). Grapes were harvested in 2002 from the Ribatejo region, Portugal.
PVI-PVP: The PVI-PVP(Divergan®HM,BASF,Germany) was applied at 10 and 20 g·hl’ (must) and 30 and 50 ghl’ (wine). Experiments are summarised inFig. 1. These application doses are those recommended by the supplier.
For the musts, the PVI-PVPwas introduced before fermentation. They were immediately shaken for 20min,

after which they remained at rest during fermentation. Sep-aration occurred at first racking. The wines were shaken for 15 min immediately after introduction of the PVI-PVP,
after which they too remained at rest for 48 h.
Fining treatment applied in wine with metal additions: With the aim of con-firming the influence of PVI-PVP on Fe and Cu we added these elements to matured white and red wines: ‘Addi-
tion A’ (5 mg·l’ of Fe and 0.5 mgl’ of Cu). ‘Addition B'(15 mgl’ of Fe and 1 mgl’ of Cu). We also added two levels of PVI-PVP 25 and 50 g-hl’ (the amounts ‘1’ and ‘2’respectively).
Mineral composition: The concentration of potassium (K), calcium (Ca), sodium (Na), magnesium (Mg), iron (Fe) and copper (Cu) was assayed by Flame
Atomic Absorption Spectrometry (FAAS) using a Varian Spectra 10/20(Victoria,Australia), according to the offi-cial method of OIV (OIV 1990).
The Pb, Cd and Al content were determined by Elec-trothermal Atomic Absorption Spectrometry (ETAAS), ac-cording to the method described by CAIARINO (2000) and
CATARINO et al. (2002).
Colour and phenolic composition:
The total content of phenolic compounds was measured by the absorbance at 280nm (RIBÉREAU-GAYON1970); colour intensity by the sum (A2o+As2o+^g0), where tonality is de-
fined by the ratio (Ae/ ^s2g). For the white wine, colour in-tensity was measured using absorbance at420 nm. Colour
measurements of the wines were also performed according to the CIELAB 76 method (McLAREN 1980). Spectral read-ings, transmittance every 10 nm over the visible spectrum
380-770 nm, were performed with a UV4 Unican Visible Spectrometer (Cambridge, UK), using quartz cells of path length 1 mm (red wine), and 10 mm (white wine). The soft-
ware Chroma 2.0 colour measurement was used to calcu-late the CIELAB coordinates directly. The L*, a* and b*values describe a three-dimensional colour space.
For the red wine, the total pigments were estimated us-ing the method of SomeRs and EvaNs (1977); the polymeric pigments index was determined by a method proposed by
GLORIES (1978) and total anthocyanins were evaluated ac-cording to the method of RIBÉREAU-GAYON and STONESTREET(1965).
Organic acid composition:Organic acids (tartaric, malic, citric, lactic and shikimic acid) were analysed by High Performance Liquid Chromatography
(HPLC) using a method described by TussEAU and BENOIT(1986 a, b). Two columns with revcrsc phasc Lichrospher 100RP 8 (Merck, Darmstadt,Germany) (particle size 5 μm, 250 x 4 mm) were used. Detection was made with a UVIS 206 PHD (KONIK Instruments, Barcelona, Spain) set at 210 om, and the peak areas were determined with Konikchrom 5.2 software.
Current analysis: pH, titratable and vola-tile acidity were determined by Fourier Transform Infrared
Spectrometry-FTIR (WineScan FT120, Foss, Slangerup-gade, Denmark) (MoREIRA et al. 2002 a,b).
Sensory analysis: Sensory analyses were performed by six expert panellists who were members of“ComissÃo Vitivinicola Regional of Ribatejo (CVRR)”.
These persons were all properly trained and had good pre-vious experience. We asked panellists to look for differ-ences between the control wine and each sample of treated wine. The wine attributes included visual, nose and taste senses, as well as overall (global) appreciation.
Statistic analysis: The data were analyzed by Analysis of Variance using SPSS 12.0 for Windows. The treatment means were separated by the Scheffée test at the 5 % significance level.
Results and Discussion
Fining treatment applied in must and wine:Must and white wine: The results of applying PVI-PVP in must and wines were analysed sepa-rately. The summary of our variance analysis results with white wine is shown in Tab.1.’
For phenolic composition we confirm that PVI-PVP,either in the must or in wine caused a slight, but signif-cant decrease in the level of phenolic compounds in the wine. This was expressed in a reduction in the total phenols index (Ipt). However, no significant differences appeared between the two PVI-PVP dose rates used. This is consist-ent with the observations of MATTIVI et al. (2000). Accord-ing to these authors, PVI-PVP removes also some phenolic compounds from the wine and, in particular the low mo-lecular weight ones. It is these that are mainly involved in the oxidative browning of white wines. Also NicolINI et al.
(2001,2004) and Eder et al. (2001, 2003) refer to the influ-ence of PVI-PVP on the reduction of hydroxycinnamates,thus improving colour stability.
In fact, with regard to the wines’ chromatic character-istics, we confir a slight, but not significant decrease in A420 and in the co-ordinates L*, a* and b*.
As for mineral composition, we found that PVI-PVP had no effect on Ca, Mg and Na contents, regardless of whether the product was used in the must or, later in the wine. This result is in line with that of other authors (EDER et al. 2001, 2003; NicOLNI et al. 2004). However, the ap-plication of PVI-PVP in the must did have a slight, but sig-nificant effect on the potassium content when compared with the control. This seems to indicate that the application of PVI-PVP in the must causes a premature elimination of this cation, and this probably contributes to tartaric stabili-sation in the subsequent wine.
In the particular case of Cu, the application of PVI-PVP in the must or in thc winc did not havc a significant effect on the content of this metal. This result may be due to the very low copper content of our wine, or to our exper-imental conditions. Also, the discontinuous shaking may have resulted in insufficient contact between PVI-PVP and the wine. Our result could be explained by that of NIcOLINI et al. (2004), who found that removal was faster and more complete when mixing was more thorough. Also, EdER et al. (2003), suggest that metal removal can occur quite rapidly allowing PVI-PVP to be separated from the wine after only 16 h of contact.
Under our conditions, no effect on Fe content was detected with PVI-PVP either in white must or in wine;
For more data, please contact our customer service。