Dominated for many years by cork closures, the market for still wine closures has completely changed and is now divided into five major categories of bottle-closure technology. The diagram below (figure 1), based on market data collected by Vinventions’ business intelligence teams, reveals that, despite conventional wisdom thinking, the dominant closure solution is none other than the screw cap with nearly 37% of the market share. Furthermore, another surprising result is that single-piece corks account for only about 10% of this market. Cork-based technical closures account for 20% of this market and are being caught up by micro-agglomerated closures made from small cork particles, clustered in matrices most often composed of polyurethane, with 16% market share. Finally, the category of synthetic closures and PlantcorcTM bio-sourced closures represent 15% of the market. This diversification of the closure solution took place quite quickly because, at the end of the 1990s, cork-based closures accounted for 97% of the market according to our data. The two factors that supported this diversification are undoubtedly the research by all stakeholders in the marketing value chain of wines, for solutions without cork taint and that allow for increased homogeneity between bottles.

1. Current criteria for selecting still wine closures

Long considered a simple way of sealing bottles, the closures for still wines must now fulfil many criteria in order to meet the requirements of the producers, the distributors, but also the consumers. Without assuming their relative importance, the following criteria must be considered by users to ensure fault free wines, kept in the best conditions before being tasted, and with sustainable success in the market.

• The closures must be free of haloanisoles (TCA, TeCA, PCA, TBA, etc.) to avoid cork and moldy taints. As a reminder, haloanisoles above their recognition threshold give rise to defects such as cork taint. But at concentrations between the threshold of perception and the threshold of recognition, these molecules distort the characteristics of the wines by masking the aromas of interest of these wines (Takeuchi et al., 2013). Although many techniques have been developed to eliminate these molecules from cork closures, their effectiveness is not all equivalent, and users should remain vigilant about these aroma masking effects that are responsible for bottle-to-bottle variations. Unfortunately, this subject is not closed.
• Another factor to ask the supplier is the suitability for food contact in liquids containing alcohol. Apart from single-piece closures made of cork that are not subject to serious regulation due to certain empiricism, any other closure solution must be the subject of a food contact study conducted by specialized laboratories. It is important to remember that organoleptic neutrality (having no impact of the closure on the color, aroma, and the taste properties of the wine) is a guarantee of quality.
• The mechanical performance of the closures is essential to optimally preserve the wines. Indeed, the performance during bottling and during the storage of the bottles ensures the absence of gas leakages (oxygen and carbon dioxide) as well as liquid (leaking bottles). For cylindrical closures, it is essential to know the mechanical properties and, in particular, all the physical properties linked to elastic return. For screw caps, it would be unwise to think that the risks of gas or liquid leaks are non-existent, and it is really important to keep in mind that airtightness is only achieved through the screw cap liners. Given the small contact area to ensure this airtightness, we can better understand the crucial influence of fine-tuning when bottling with each change of cap or the shape of the bottle.
• When the mechanical properties of the closures are optimal, then one must consider the aging capacities of the wines for the different technologies considered. When leaks are impossible, the materials used for the different types of closures must regulate gas exchanges (entry of oxygen and loss of carbon dioxide). This theme of gas exchanges will be the main topic of discussion of this article.
• Closure solutions are also at the heart of sustainable development considerations, and it is important to be able to ask the supplier for evidence of their involvement in these topics, both in terms of their production tool and the intrinsic environmental properties of the closures. Legislation and voluntary commitments from retailers and wine producers will create the need for documents that can attest for the carbon footprint, the life cycle, the use of bio-sourced materials, the level of recyclability, biodegradability, etc. of the various closure solutions.
• Finally, the marketing component can also guide the selection of a particular type of closure. Consumer perception remains an element often advanced as a key to decision-making, but there have been few serious studies conducted without cognitive bias or bias in consumer questions that can support these assertions. Consumer perception and the speed with which new closure solutions are adopted vary from country to country. The rapid adoption of the screw cap in Australia or Scandinavia is an example. Aesthetics and originality can also represent the selection criteria, but care must be taken not to compromise the quality of the wines.

2. Oxygen management and closures

2.1 Oxygen measurements through the closures

This theme is quite recent, and it is interesting to recall that the study, which is considered to be the first, showing that different permeabilities between types of closures dating from 2001 (Godden et al., 2001), had, in fact, only hypothesized different permeabilities between the closures studied. Due to the lack of precise and easy-to-implement oxygen measurement technology on a large number of simultaneous trials, the authors were unable to demonstrate their hypotheses. The reference method at the time was a method using Mocon technology (Peck et al., 2005) according to a protocol that had been adapted from the ASTM F1307-02 standard (see bibliography) to measure film permeability. This method, never really validated for measuring the permeability of closures, is still in use in material analysis laboratories but has the disadvantages of being expensive, allowing to process only a limited number of tests and requiring the cutting of bottlenecks before sticking them on a metal plate, which does not guarantee the absence of leaks in the system (see photo 1).

2.2 The NomaSense Method of Wine Quality Solutions

Our team began work on the development of a method of measuring the permeability of closures in 2007. A non-invasive technology, based on luminescence using sensors that stick inside transparent bottles, NomaSense (Wine Quality Solutions) technology was chosen. This oxygen measurement technology has been validated according to the recommendations of the OIV (Dieval et al., 2009). We have also published the validation of the measurement of the permeability of closures with the NomaSense method in comparison with the “reference” Mocon method (Dieval et al., 2011). The goal was to develop a fast, affordable, and accurate method. This method has gradually become the reference method, as evidenced by its adoption by large closure-manufacturing groups (Oliveira et al., 2013 ; Chevalier et al., 2019).

2.3 The Protocol of the NomaSense Method

The implementation of the method remains relatively simple but still requires a lot of care given the inerting conditions prior to measurement. We chose to work with a measurement method in the absence of wine because it reacts with oxygen and induces a decrease in the amount of oxygen present in the bottle.

• Equipment: 375 mL clear glass bottles are used. Their internal neck profiles are controlled by an automatic probe (Egitron® PerfiLab®) to make sure the dimensions are within specifications and to calculate the exact volume occupied by the closure. PSt6 sensors from Wine Quality Solutions are glued to the inside of these bottles to allow non-invasive measurement over time. These sensors are calibrated in the factory.
• Closure: Bottle closure remains the critical point of this method. It is very important to close the bottle after a very effective inerting of the bottle, in order to start the measurements with an atmosphere in the bottle at less than 0.5 hPa of oxygen. For that, a GAI bottling line (Model MLE 441) is used, with a compression diameter set at 15.7 mm. The double nitrogen inerting 5.0 Aligal 1 (> 99.99%) of the bottles is made by the vacuum/gas system of the electro-pneumatic filling heads. The closure is made by a vacuum at 0.95 bar, speed at 13 Hz.
• Bottle storage: All bottles are stored in a thermostated cellar at 23 ± 1°C and 70 ± 5% relative humidity (see photo 2). The measurements are carried out for 14 days at a rate of 2 measurements per day.

• Measurement using the NomaSense O₂ P6000
  The NomaSense O₂ P6000 is used to measure the oxygen content in the bottles. The optical fiber is applied to the glass outside the bottle towards the PSt6 sensors glued to the inside, as shown in photo 2. The analyzer emits a blue light that stimulates the luminophore contained in the sensor. The amount of oxygen present on the sensor is proportional to the delay in the emission of red light generated by the return to the basic electronic level of the luminophore. This simple principle of physics is the most specific of the measurement of oxygen among the measurement technologies known to date.
• Data processing and modelling All measurements taken are stored and processed using XLfit software (IDBS, Guildford, Surrey, UK), which is a modelling option of Microsoft® Excel®. The most complicated part of the development of this method was to make it quick. For that, our team developed a predictive model to properly characterize the permeability properties of the closures in 14 days. The model developed is based on the laws of Fick and Henry, as well as Crank’s diffusion equations. The approach and working assumptions are detailed in our 2011 paper (Dieval et al., 2011). This paper also shows the correct correlation between the results predicted after 14 days and those obtained by the Mocon method.

2.4 Definitions and components of permeability

Let us take a moment to recall some points and establish some important definitions so as to avoid confusion that may have arisen due to a lack of uniformity of communication in the area of the permeability of closures.

The NomaSense method that we developed, as well as the approach we followed, enabled us to be the first to consider the aspects linked to the supply of oxygen through the closures as a whole. Under the experimental conditions described above, the classic curve obtained for the oxygen supply of closure is shown in figure 2 below. This curve, which follows the model developed based on the Crank equation, makes it possible to distinguish two distinct but yet interconnected phases by the Crank equation:

• the desorption phase (oxygen contained in the cork and released into the bottle after the closure)
• the phase reached at equilibrium, commonly called OTR (for Oxygen Transmission Rate).

The two phases observed can be explained (cf. figure 3) by recalling that the force favoring the entry of oxygen from one compartment to another is the difference in partial oxygen pressure between these two compartments. A cylindrical closure is a porous cylinder which therefore contains air and, when inserting it into a bottle neck (18.5 mm internal diameter as reference dimension), the oxygen pressure in the cylinder increases from 21% to more than 30%. Despite oxygen release to the outside of the bottle, oxygen will tend to enter the bottle quickly after closure because the distance to travel and the pressure gradient is more favorable than when the balance is reached (when the removal in the compressed closure has disappeared).

Even if these phenomena had already been observed, we were the first to describe them mathematically and to correlate the role of desorption with the early stages of wine evolution. This work, carried out in collaboration with AWRI, was published in 2011 (Ugliano et al., 2011) and was behind the creation of the first cylindrical closure treated against desorption, now called the Nomacorc Select Green 100, and was marketed in 2011. Since then, the Nomacorc Reserva closure has benefited from the same patented treatment. To our knowledge, these two closures remain the only ones with these advanced oxygen management characteristics. The advantage of eliminating desorption is that the oxygen-sensitive aromas are better preserved. In addition, unlike closure solutions with very low OTRs, the higher oxygen intakes thereafter can be used to counter the reduction phenomena that may appear during the aging of the wines. (Ugliano et al., 2011 & 2012).

It is worth specifying 2 important points regarding desorption. Firstly, oxygen is not brutally flushed into the bottle (in more trivial terms, there is no “pscchttt” of oxygen into the bottle). The materials used according to the type of closure have characteristics of diffusion intrinsic to oxygen that will define the velocity of oxygen progression. Secondly, depending on the thickness of the materials and the cellular or molecular organization of these materials, the apparent diffusion velocity can vary greatly. In other, more colorful terms, oxygen has to work its way through successive layers of material before it can be released into the bottle. As we have stated previously, desorption and OTR are bound by the Crank equation. Thus, the lower the OTR of a closure, the longer the desorption will last. For example, we evaluated the time it takes for the desorption phase to be finished for different closures in the Nomacorc Plantcorc range. (cf. table 1). Although a recent publication proposes to set the desorption after 6 months (Chevalier et al., 2019), it does not seem appropriate to set a duration for desorption a priori based on the results of the table below.

Furthermore, any leakage to the glass/cork interface impacts the resulting curves, and the model can no longer be used. Photo 3 below illustrates the different sources of oxygen that can enter a wine bottle at the cylindrical closure. In the absence of leakage, the oxygen supply from the closure is, therefore, the sum of the desorption and the OTR.

2.5 The units used

If there is one point that may have created confusion in the market, it is that of the units used by different authors and suppliers.

Initially, the units to define the OTRs came from the ASTM standard and were expressed in cc/day/closure at 100% oxygen (meaning that the closures were exposed to an atmosphere of 100% oxygen instead of the 21% concentration in the air). The first level of confusion arose when the units in cc/day/closure at 21% oxygen were brought to light. Some took advantage of this fivefold difference to advance lower OTR values. Then, new, more complicated units emerged.

We proposed to simplify all this a decade ago and to use a simple unit understood by all winemakers, namely the mg/bottle. In fact, technicians in the wine industry are used to measuring oxygen intakes during winemaking in the order of mg per month and are accustomed to concentrations of important molecules of the wine in mg/L, as is the case for Free SO₂. We invested a lot of time to better understand the link between oxygen supply through corks and the evolution of bottled wines, so it was clear to us that we needed to use units that were understood by all and to establish correlations more easily. Other closure manufacturers are now following suit, as evidenced by recent publications (Oliveira et al., 2013 ; Chevalier et al., 2019).

3. Oxygen intakes and homogeneity

As already mentioned in the introduction to this review, the problems of bottle-to-bottle variations were one of the main causes of the diversification of the closure technologies for still wines. These variations may be due to very variable oxygen supply levels from one closure to another (M. Ugliano et al., 2015). Once again, to highlight these variations, a large number of modalities had to be measured, which was possible using the NomaSense method.

For more than a decade, we have been relaying the message that determining the values of oxygen supplies is not enough to give ourselves the means to properly characterize its choices of closure but that it was really necessary to take into account the coefficient of variation around the average value. We have been able to analyze all the closure solutions available on the market and can relay the main observations. Figure 4 below gives an overview of the results obtained for different closure technologies. We will comment only on the results of corks, micro-agglomerates, and co-extruded closures (synthetic and Plantcorc).

The category of corks has the greatest variability between closures among all the categories we measured. Moreover, the study published by Amorim’s teams in collaboration with the University of Bordeaux (Oliveira et al., 2013) confirmed that the variability that may exist within a population of 600 tubed corks in only 2 cork planks could be very significant and could vary from 0.5 to more than 4 mg of oxygen over 1 year, as illustrated by the results of Table 2.

We measured the same levels of variability within our laboratory, even in cases where the closures were pre-selected on the basis of their respective masses, not allowing the link between the permeability of the corks and their density.

For the micro-agglomerated class, a large number of measures show that the median for the category is 0.5 mg per year, which corresponds to the low value of OTR measured on natural closures. The variability can be observed between different producers of micro-agglomerates but also between batches of the same producer, which should encourage a good understanding of this variability even for this class of industrial closures.

For the other large category of cylindrical closures, the co-extrusion process, used in the manufacture of synthetic closures and Plantcorc, is a continuous means of production that achieves the highest level of homogeneity from one closure to the other as illustrated by the results of the previous graph. This mode of production, therefore, gives the best homogeneity performance to the class of co-extruded closures.

Conclusion

Many criteria are now taken into account in the selection of a still wine closure solution and allow responding to both the requirements of the producers as well as those of the different markets. Among these criteria, the management of oxygen through the seal is increasingly put forward by different manufacturers of closures. After developing mathematical models that record the oxygen supply of cylindrical closures, we want to draw attention to the following:

• the OTR is not enough to describe the oxygen management performance of cylindrical closures
• the phenomenon of desorption must also be considered to account for the oxygen supply by the closures
• desorption lasts longer when the OTR is low
• finally, the absence of intrinsic variability of selected closures is a key element in ensuring bottle-to-bottle homogeneity in line with market requirements.

Research carried out in collaboration with academic institutes in the wine industry worldwide has allowed us to better understand the influence of oxygen on the evolution of bottled wine and to develop a range of closures (Green Line and Reserva) with oxygen management characteristics still unique in the closure market in terms of oxygen supply and homogeneity.