Some time after I published a technical paper on the measurement of copper in wine in 1983, I was called by a winemaker who wished to discuss the chemistry of copper additions to remove the aroma of hydrogen sulfide. The winemaker’s concern was that after the addition of copper(II) sulfate, the off-aroma of H2S disappeared, but it returned after about one month in tank. The winemaker said, ‘I am trained in chemistry and spent many years working as an industrial chemist. The whole thing doesn’t make chemical sense, as I know the solubility product of copper sulfide and it should simply precipitate and be removed’. I agreed and said ‘leave it with me and I will get back with an answer’.
It is now about 35 years after that phone call; I have retired, as has the winemaker. Fortunately, Dr Andrew Clark of the National Wine and Grape Industry Centre (NWGIC) at Charles Sturt University has continued working on the copper–sulfide problem and is now unravelling the intriguing chemistry of this wine chemistry issue. Andrew, together with Nikos Kontoudakis and Anque Guo, has examined the complexities of copper sulfide filtration using model wines and white wines. The results take the understanding of the copper sulfide saga into the area of copper sulfide nanochemistry and the association of these nanoparticles with various wine components (see Aust. J. Grape Wine Res. 2019, vol. 25, pp. 53–61).
To model industry practice of adding copper(II) (see April 2015, p. 39), the NWGIC group used a tartaric acid/ethanol model wine to which CuSO4 and Na2S were added to give a final Cu:S ratio of 1:2. A 10-millilitre syringe filter was used in association with 0.45 and 0.20 μm pore size regenerated-cellulose filters. The copper concentration in the filtrate was measured to reflect the effectiveness of the filtration.
With the model wine, about 4% of the total copper in the system was found in the filtrate, seemingly all good news. However, when as little as 1% white wine was included with the model wine, about 70% of copper made its way into the filtrate, increasing to more than 90% with 10% white wine in the model. In separate experiments, it was found that the copper in the filtrate was present as copper sulfide, which I will write about in my next column. So, at least we have an explanation as to why the H2S aroma can return to the wine over time: the sulfide is not actually being removed as copper sulfide – rather, it is being sequestered. An explanation in part perhaps, but not overly helpful to the winemaker. None-the-less, the results opened up several areas for research, including the influence of wine components on the filtration step, the size of the copper sulfide particles themselves and what the actual mechanism of filtration is: size discrimination or adsorption.
A search for wine components that influence the filtration step showed that white wine proteins and polysaccharides exerted a major influence on the filtration process. When added to the model wine, these macromolecules allowed close to 100% of the sulfide-bound copper to pass through the filter compared to only 4% of sulfide-bound copper in the absence of the macromolecule. Phenolic compounds allowed 30–35% of the sulfide-bound copper to pass through the membrane, a much lower effect than with the wine macromolecules. So, while copper–phenolic chemistry is important for wine oxidation, it is less critical for copper–sulfide chemistry.
Nanoparticle tracking measurements were carried out through a NWGIC collaboration with Agnieszka Mierczynska-Vasilev, Paul Smith and Eric Wilkes at the Australian Wine Research Institute. The results showed that when H2S (from Na2S) and CuII were added to the model wine, the particles were much lower than 0.2 μm, markedly smaller than for white wine, where the majority of the particles were just below 0.2 μm with about 20% in the 0.2–0.3 μm range. The existence of smaller particles in the model wine was sufficient for the study authors to conclude that size exclusion was not the mechanism involved in the different filtration results for copper sulfide in model and white wines.
Using different filter materials, it became clear that the removal of copper sulfide is actually by an adsorption process, a process that is hindered by white wine proteins and polysaccharides. The impact of the macromolecules is most significant with the regenerated cellulose filters, while more effective, but not total, copper sulfide removal was observed with nylon and polyethersulfone filters. Differences in the hydrophilicity and/or polar activity of the filters is a question still be explored.
Recognition of the chemistry of copper sulfide nanoparticles in white wine and the interference by wine macromolecules on the filtration process has significantly advanced our understanding of the copper(II)/sulfide problem. The study authors do note that their experiments need to be repeated on a larger scale with selected filter materials, recognising that the type of filter most suitable for copper sulfide removal may not be effective for other wine filtration needs.