Impact of Metal Screw Cap Closures on Trace Element Profiles in White Wines After One Year in Bottle

Abstract

In addition to the minerals naturally present in grapes, wine can acquire additional minerals during its production and storage from materials that come into contact with it, including bottling materials. This study aimed to evaluate the concentration of a wide range of elements in white wine samples packaged in 0.75 L green glass bottles sealed with two different closure systems: natural cork stoppers and metal screw caps with a plastic liner. No statistically significant differences were observed between the two closure types for most elements (Li, Be, Fe, Co, Ni, Zn, Se, Rb, Sr, Mo, Sb, Cs, Ba, and Tl). For V, Cr, Mn, Cu, As, Cd, and Pb, some differences were observed, but without a clear pattern. However, the concentration of Sn was significantly higher in wines packaged in bottles sealed with metal screw caps plus plastic liner. Elemental analysis of the original, unused liners showed negligible content of Sn and other studied elements, suggesting that the Sn in the wine comes from the Sn-plated steel screw cap, despite the presence of the plastic liner. Although the changes in the natural elemental composition under these bottling conditions are not very high and unlikely to pose a health risk to consumers, they may still influence wine stability and sensory attributes. Understanding these effects is important for both wine producers and consumers to ensure optimal wine quality and preservation.

1. Introduction

Wine is one of the oldest and most widely consumed alcoholic beverages worldwide [1]. According to estimates from the International Organisation of Vine and Wine, global wine consumption reached 221 million hectolitres in 2023 [2]. European countries accounted for 48% of this total, making them the largest consumers.

Wine is a complex mixture of water, ethanol, glycerol, polysaccharides, organic acids, polyphenols, minerals and other minor compounds [3,4]. The concentration of these components is influenced not only by the composition of the grapes but also by the techniques, equipment and materials used in wine production and storage, as well as by the yeasts involved in the fermentation process and other factors [3,5,6,7,8]. Minerals, in particular, have been used to assess authenticity and determine provenance [6]. However, due to their ubiquitous presence in materials that come into contact with grapes (e.g., stainless steel, wood, glass) and the addition of other substances (e.g., bentonite) during the winemaking process, completely eliminating all sources of contamination is virtually impossible [6,9,10,11].

Throughout history, various bottle designs and closure methods have been used for wine bottling. In ancient times, the Greeks and Romans employed amphoras to effectively prevent wine oxidation [12]. Natural cork closures, made from the bark of Quercus suber L., have been used since at least the 17th century, when glass bottles were first manufactured, mainly because other materials could not provide an airtight seal [12,13]. Cork stoppers allow some oxygen to enter the wine, primarily from the air trapped within the cork cells. Metal screw caps are also used to seal wine bottles, though they require specially designed bottles to accommodate the screw cap setting [14]. The inner structure of a screw cap liner typically includes a polyvinylidene chloride (PVDC) film in direct contact with the wine, a polyethylene (PE) wad and, in some cases, a thin Sn sheet between the PVDC and PE layers [15].

The type of packaging used for wine storage also appears to influence its mineral composition. Hopfer et al. (2013) observed higher concentrations of Cr, Cu and Sn in wine samples bottled in 0.75 L green glass bottles compared to those stored in bag-in-box containers [16]. In the same study, the authors reported a higher Sn concentration in wine bottles sealed with metal screw caps (aluminium STELVIN^® cap, 30 mm × 60 mm, with 28.6 mm × 2 mm Sn-PVDC liner) compared to those sealed with natural cork stoppers or stored as bag-in-box, possibly due to leaching from the Sn layer in the screw cap liner [16].

This study aimed to evaluate the concentration of various elements in wine samples bottled in 0.75 L green glass bottles, comparing two different closure systems: natural cork stoppers and metal screw caps. In their study, Hopfer et al. (2013) hypothesised that wine was able to leach Sn from the Sn layer of the metal screw cap liner [16]. For this reason, in the present study, the metal screw caps contained a Sn-free liner composed of a PE layer covered by PVDC on both sides.

2. Materials and Methods

2.1. Sample Preparation

Five different wineries (A, B, C, D, E) provided 10 bottles each of their own white wine. Five of these bottles were sealed with natural cork stoppers, and five were sealed with metal screw caps (Sn-plated steel) with a Sn-free Saranex™ liner. Bottling took place between February and August of 2023. The wine bottles were stored horizontally in the wineries’ cellars (to maintain contact between the wine and the closure system), in naturally quite stable conditions, typical of traditional wine cellars, with consistent temperature and humidity, and protected from light and vibrations. In May 2024, all bottles (n = 50; 5 × 10 bottles; 25 with natural cork stoppers and 25 with metal screw caps) were carefully opened, and two 10 mL aliquots were transferred to 15 mL HDPE tubes and placed in an ultrasonic bath for degassing. Additionally, ten 15 mL tubes containing 10 mL of ultrapure water were also placed in the ultrasound bath to control contamination during the degassing process and were later analysed as procedure blanks. After degassing, the samples were stored at 4 °C until analysis.

2.2. Elemental Analysis

The concentration of 22 elements was determined by inductively coupled plasma with mass spectrometry (ICP-MS). The instrument used was an iCAP™ Q (Thermo Fisher Scientific, Bremen, Germany), equipped with a Meinhard^® TQ+ quartz concentric nebuliser (Golden, CO, USA), a Peltier-cooled, high-purity quartz, baffled cyclonic spray chamber and a demountable quartz torch with a 2.5 mm i.d. quartz injector. The interface consisted of two Ni cones (sampler and skimmer). High-purity argon (99.9997%) supplied by Gasin (Matosinhos, Portugal) was used as nebuliser, auxiliary and plasma gas. Prior to analytical runs, the instrument was tuned for maximum sensitivity and signal stability, as well as for minimal formation of oxides and doubly charged ions. The main operating parameters of the ICP-MS instrument were as follows: nebuliser gas flow, 1.15 L/min; auxiliary gas flow, 0.79 L/min; plasma gas flow, 13.9 L/min; radio frequency generator power, 1550 W; and dwell time, 10 ms.

All solutions were prepared with ultrapure water (>18.2 MΩ·cm at 25 °C) obtained from a Sartorius Arium^® pro water purification system (Göttingen, Germany). Nitric acid (HNO₃, 67–69% w/w, TraceMetal^® Grade) was obtained from Fisher Scientific (Leicestershire, UK). Absolute anhydrous ethanol was acquired from Carlo Erba Reagents (Val de Reuil, France), and Triton X-100 was obtained from Sigma-Aldrich (St. Louis, MO, USA). Calibration standards were prepared by diluting single-element stock solutions of Cs and Ba (1000 mg/L, PlasmaCAL^®, SCP Science, Baie-D’Urfé, QC, Canada) and two multielement stock solutions: ICP multi-element Standard Solution XVI (100 mg/L, Certipur^®, Supelco, Steinhein, Germany) and Metalloid and Non-Metal Mix for ICP (100 mg/L, Certipur^®, Supelco) in 2% HNO₃. Internal Standard Mix 1—SCP-IS7 (10 mg/L, PlasmaCAL^®, SCP Science) and single-element stock solutions of Ga, Rh and Ir (1000 mg/L, PlasmaCAL^®, SCP Science) were used as internal standards (IS). All laboratory ware (bottles, tubes, volumetric flasks) was made of polypropylene or high-density polyethylene (HDPE) and was properly decontaminated by immersion in a 10% v/v HNO₃ solution for at least 24 h, followed by extensive rinsing with ultrapure water and drying under dust-free conditions at room temperature.

All samples, calibration standards and blanks were diluted 1:10 with a diluent solution containing 2% v/v HNO₃, 1.5% v/v ethanol and 10 µg/L IS. To control matrix effects and isobaric interferences, a “synthetic wine” solution containing 7.0 g/L citric acid monohydrate, 3 g/L sucrose, 10 g/L glycerol, 100 mg/L CaCl₂, 100 mg/L MgCl₂, and 10% (v/v) ethanol was added to the calibration standards. The elemental isotopes ⁷Li, ⁹Be, ⁵¹V, ⁵³Cr, ⁵⁵Mn, ⁵⁹Co, ⁶⁰Ni, ⁶⁵Cu, ⁶⁶Zn, ⁷⁵As, ⁸²Se, ⁸⁵Rb, ⁸⁸Sr, ⁹⁸Mo, ¹¹¹Cd, ¹¹⁸Sn, ¹²¹Sb, ¹³³Cs, ¹³⁷Ba, ²⁰⁵Tl, ²⁰⁶Pb, ²⁰⁷Pb and ²⁰⁸Pb were measured for analytical determination, and the elemental isotopes ⁶Li, ⁷¹Ga, ⁸⁹Y, ¹⁰³Rh and ¹⁹³Ir were monitored as IS. The elemental isotopes ¹³C, ³⁵Cl, ⁷⁷Se, ⁷⁹Br and ⁸³Kr were measured to correct for interferences on the elemental isotopes ⁵¹V, ⁵³Cr, ⁷⁵As and ⁸²Se. The concentration of Pb was estimated by summing the signal intensity of ²⁰⁶Pb, ²⁰⁷Pb and ²⁰⁸Pb. A rinse solution consisting of 2% v/v HNO³, 1.5% v/v ethanol and 0.002% v/v Triton X-100 was pumped through the sample introduction system between each sample to prevent carryover. A certified reference material (EnviroMAT-Drinking Water High Matrix from SCP Science, Baie-D’Urfé, QC, Canada) was matrix-spiked as for the calibration standards and analysed together with the wine samples for analytical quality control. The limits of detection (LD) were calculated using the equation: LD = 3.3 × σ_blanks, where σ_blanks is the standard deviation (SD) of the blanks (matrix spiked as for the calibration standards).

Diluted samples and calibration standards were introduced into the ICP-MS instrument using a CETAC ASX-520 auto sampler (Teledyne CETAC Technologies, Omaha, NE, USA), after thorough homogenization in a vortex mixer.

In addition to the wine sample analysis, three liners were carefully removed from unused metal screw caps and mineralised by closed vessel microwave-assisted acid digestion in an ETHOS™ EASY microwave oven (Milestone, Sorisole, Italy) equipped with an SK-15 EasyTEMP high-pressure rotor. The liners were carefully cut into small pieces with clean stainless-steel scissors, and approximately 0.15 g were accurately weighed into three microwave TFM (modified polytetrafluorethylene) vessels. For the mineralisation procedure, 9 mL of HNO₃ (67–69% w/w) and 1 mL of hydrogen peroxide (H₂O₂, 30% w/w, Suprapur^®, Supelco, Darmstadt, Germany) were added to the vessels. After closing with a calibrated tension wrench, the vessels were placed inside the microwave oven, and samples were mineralised under the following programme: a gradual temperature increase for 15 min until 210 °C, followed by 20 min at 210 °C. After cooling to room temperature, the vessels were carefully opened, the contents transferred into 50 mL calibrated tubes (+/− 0.25 mL), and 40 mL of ultrapure water was added for a final solution volume of 50 mL. Three procedure blanks (only the reagents) were included in the mineralisation batch and were later analysed to estimate the LD as described previously. Elemental analysis was performed by ICP-MS as described for wine samples, but without the addition of ethanol to the diluent and synthetic wine matrix to the calibration standards.

2.3. Statistical Analysis

Statistical analysis was performed using SPSS Statistics v.27.0 (IBM Corporation, Armonk, NY, USA). The elements’ concentrations are presented as mean (SD). Levene’s test was used to assess the homogeneity of variance. The significance of differences between wine samples bottled with natural cork stoppers and metal screw caps was determined using either Student’s t-test or Welch’s t-test, depending on variance equality. Multivariate linear discriminant analysis was performed to identify the elements that differentiate between the two wine closure types, using XLSTAT software version 2020.2.3 (Addinsoft, Paris, France). Statistical significance was set at p < 0.05.

3. Results

The concentration of a total of 22 elements was determined. The LD ranged from 0.003 µg/L (for Tl) to 22.1 µg/L (for Fe) (

Table 1. Detection limits (LD) of the analytical procedure used for wine sample analysis (µg/L).

Element	LD	Element	LD	Element	LD
Li	0.347	Cu	1.33	Sn	0.061
Be	0.007	Zn	0.435	Sb	0.051
V	0.883	As	0.135	Cs	0.021
Cr	3.04	Se	0.502	Ba	0.258
Mn	0.294	Rb	0.040	Tl	0.003
Fe	22.1	Sr	0.513	Pb	0.016
Co	0.043	Mo	0.020
Ni	0.137	Cd	0.018

). The analytical recovery of all elements varied between 82% and 107%, and the relative standard deviation (RSD, %) was below 5% for all elements except V (6.6%) (Table S1).

The mean (SD) concentrations of the various elements in the wine samples bottled under the two different conditions—0.75 L green glass bottles with natural cork stoppers or metal screw caps—are shown in Figure 1 and Table S2 separately for each wine (winery). No statistically significant differences were observed between wines (A–E) bottled under the two conditions for Li, Be, Fe, Co, Ni, Zn, Se, Rb, Sr, Mo, Sb, Cs, Ba and Tl (Table S3). The Sn concentration was significantly higher in wines bottled with a metal screw cap (p < 0.001, except for Wine D, where p = 0.014; Figure 1 and Table S2). Some wines showed statistically significant differences (natural cork stoppers versus metal screw caps) for V, Cr, Mn, Cu, As, Cd, and Pb, but without a clear pattern (Figure 1 and Table S2).

Multivariate linear discriminant analysis revealed only one discriminant function, which explains 100% of the variance (eigenvalue = 6.96, Wilk’s Λ = 0.126, χ² = 23.1, p < 0.001) observed between wines bottled with natural cork stoppers or metal screw caps (Figure 2). The element Sn was strongly correlated with the discriminant function (structure coefficient = 0.882), while the remaining elements had negligible correlations, with structure coefficients ranging from −0.029 (for Cd) to 0.053 (for Mo).

To better interpret these results, the elemental content of the liners from the metal screw caps was also assessed. The LD ranged from 0.001 µg/g for Tl to 13.8 µg/g for Fe (Table S4). The mean (SD) Sn content was 0.015 (0.007) µg/g. The content of Li, Be, V, Cr, Mn, Fe, Co, Ni, As, Se, Rb, Mo, Cd, Sb, Cs, Tl, and Pb were below the LD. The mean (SD) content of Cu, Zn, Sr, and Ba was 0.057 (0.031), 3.54 (0.35), 1.29 (0.21), and 0.924 (0.113) µg/g, respectively (

Table 2. Mean (SD) elemental content (µg/g) of the liners.

Element	Content	Element	Content	Element	Content
Li	<0.027	Cu	0.057 (0.031)	Sn	0.015 (0.007)
Be	<0.001	Zn	3.54 (0.35)	Sb	<0.002
V	<0.242	As	<0.082	Cs	<0.008
Cr	<0.343	Se	<0.088	Ba	0.924 (0.113)
Mn	<0.242	Rb	<0.029	Tl	<0.001
Fe	<13.8	Sr	1.29 (0.21)	Pb	<0.058
Co	<0.008	Mo	<0.033
Ni	<0.015	Cd	<0.001

).

4. Discussion

The concentration of the various components in wine varies greatly, not only due to the different composition of the grapes but also because of the use of different yeasts, techniques, equipment, and materials during the production and storage of wine [3,5,6,7,8]. Minerals, in particular, beyond their natural presence in grapes, can be introduced into wine during the production and storage phases due to their ubiquitous presence in the environment, from materials that come into contact with the grapes and wine (e.g., stainless steel, wood, glass), as well as through the addition of some ingredients (e.g., bentonites) [6,9,10,11].

In this study, we aimed to investigate the effect of two different bottle closures—natural cork stoppers and metal screw caps with Sn-free liners (i.e., liners composed of PE coated with PVDC on both sides)—on the mineral composition of wine. The concentration of 22 elements was determined. Of these, 14 elements showed no statistically significant differences (Li, Be, Fe, Co, Ni, Zn, Se, Rb, Sr, Mo, Sb, Cs, Ba, and Tl; Table S3). For V, Cr, Mn, Cu, As, Cd, and Pb, statistically significant differences were observed in some samples, but no clear pattern emerged (Figure 1). For example, the concentration of Cu and Pb was significantly higher in Wine B bottled with a natural cork stopper compared to the same wine bottled with a metal screw cap, whereas the opposite was observed for Wine D (Figure 1). Although statistically significant, the differences were very small and likely the result of small amounts of contamination during the bottling process.

The concentration of Sn was significantly higher in all wines bottled with metal screw caps (Figure 1). Multivariate linear discriminant analysis revealed that Sn concentration in wine accounted for nearly all the variance between wines bottled with either a natural cork stopper or a metal screw cap (Figure 2). In a similar study, Hopfer et al. (2013) also observed higher Sn concentrations in wine bottled with metal screw caps compared to natural cork stoppers [16], likely due to leaching from the thin Sn layer of the liner of the metal screw caps. In the present study, wines were bottled with metal screw caps containing a Saranex™ liner, which does not have an Sn layer. However, the concentration of Sn was increased in wines bottled with a metal screw cap. The liner is composed of a layer of PE covered on both sides with PVDC [15]. To clarify this, the content of the same elements analysed in the wine was also determined in the liners (taken from unused metal screw caps). The content of most elements, including Sn, was negligible compared to the concentrations observed in wine (Table 2). The liners were carefully removed from the metal screw caps, but the trace amount of Sn found [mean (SD) = 0.015 (0.007) µg/g] may simply be the result of contamination due to their close contact with the Sn-plated steel of the screw cap. These liners weigh roughly 1–2 g and contribute 15–30 ng of Sn to the wine. The concentration of Sn in wine (0.75 cL) bottled with metal screw caps ranged from 3.34 to 12.1 µg/L, which amounts to 2.51–9.08 µg of Sn per bottle. Therefore, the contribution of the liners used in this study to the total amount of Sn in wine bottled with metal screw caps is negligible. Taken together, the results of this study suggest that wine can leach Sn from Sn-plated steel metal screw caps even with the presence of a plastic liner, and this would explain the increased Sn levels found in wines bottled with such closure systems.

Tin has been used in food packaging, such as canned foods, since the 19th century [17]. Oral bioavailability studies in humans have shown that Sn, especially in its inorganic form, is poorly absorbed in the gastrointestinal tract (<10%) [18]. Dietary exposure to Sn in European countries has been estimated at 0.65 mg/day to 6.3 mg/day, well below levels known to cause deleterious effects [19]. The ATSDR has derived an intermediate-duration oral exposure minimal risk level (MRL) of 0.3 mg/kg/day for inorganic Sn [18], equivalent to 21 mg/day for a 70 kg adult. Assuming all Sn in the wines was inorganic, a 70 kg individual would need to consume about 1700 L/day of the wine with the highest Sn concentration found (12.1 µg/L) to reach that MRL. Moreover, the recent Commission Regulation (EU) 2023/915 on maximum levels for certain contaminants in food sets the maximum inorganic Sn level in canned beverages at 100 mg/kg [20], which is four orders of magnitude higher than the concentrations found in our samples. Thus, the Sn levels found in the wines studied do not appear to pose a risk to human health. However, Sn may interact with other components of wine (e.g., organic acids, polyphenols), potentially leading to unwanted changes in its organoleptic properties during storage. The influence of other metals in wine chemistry has been studied in the past [21], but, to our knowledge, there are no published articles studying the impact of Sn on wine chemical and sensorial properties.

The present study provides comparative evidence on the influence of two types of glass bottle closure systems on the elemental composition of wine. Future studies should investigate other types of wine (white and red; different compositions), other types of screw caps and liners, as well as the influence of variables such as temperature and storage duration on the content of Sn and other elements in wine. While the concentrations observed with either closure system do not pose a significant risk to consumer health, they do modify the natural elemental composition of the wine. The impact of Sn on the chemical composition and organoleptic properties of wine is unknown. Further studies are needed to understand how Sn interacts with wine components and whether these changes may affect wine quality or stability.

5. Conclusions

The concentration of a wide range of elements did not significantly differ between white wines bottled in glass bottles sealed with natural cork stoppers or metal screw caps with plastic liners. However, the Sn concentration was consistently higher under the second condition. As expected, elemental analysis of original, unused liners revealed negligible content of Sn and other elements, suggesting that the Sn in the wine comes from the Sn-plated steel screw cap, despite the presence of the plastic liner.

Although changes in the elemental composition of wine have been shown to be minimal and unlikely to pose a health risk to consumers, they may still affect its stability and sensory attributes.