The Truth about Filtration

To filter or not to filter, that is a question that winemakers must grapple with continuously. It is often said that when wine is filtered it is never as good as it was before filtration. US critic Robert Parker went on a crusade against filtration some years ago, and this probably encouraged some winemakers to stop filtering, with results both good and bad. Yes, filtration does take something out of the wine apart from the particles it’s intended to remove – colour, aroma, flavour, tannin – but one of the chief uses of filtration is to make a wine microbially stable. It could be argued that the ‘unfiltered’ fashion contributed to an increase in brettanomyces and other undesirable microbial growths in wine, which can end up seriously spoiling the taste.

Chapel Hill winemaker Michael Fragos recently demonstrated a series of winemaking trials to a group of Sydney retailers. Among them were unfiltered and filtered versions of the same wine, a 2008 McLaren Vale cabernet sauvignon (tastings). It was obvious that the filtered wine, which had been run through a diatomaceous earth filter, had been significantly diminished by the process. The unfiltered version had a slightly denser color, was richer and more succulent on the tongue with more extract, length and harmony, as well as a more detailed bouquet. The filtered wine was less opulent and its alcohol showed more. To many tasters, the differences might not be great, but it’s the summary of all the little one-percent differences that distinguishes a great wine from a wine that’s merely very good.

The unfiltered wine had been clarified by gravity settling and racking (pumping or draining the wine off its lees), a gentle and natural process which may be done up to four times during the 20 months barrel aging, but which can result in a less-polished wine with a slightly less limpid appearance. “When you strip away the filter-cake, you find it has a lot of colour and smells terrific,” said Fragos. “So you know the wine has lost a lot of its important constituents to the filter medium.”

Fragos also showed us four alcohol trials on a 2009 McLaren Vale shiraz (tastings). Samples of the same wine had been passed through a reverse osmosis (RO) machine – a modern invention that can remove alcohol. The samples were at 8%, 10%, 12% and 14% alcohol by volume. The 14 percenter was the natural, untreated wine. The differences were astonishing. The 8% wine was so out of balance that it tasted unpalatably sour. The wines all contained the same acidity but this one tasted higher because of the loss of balance. Of course, no-one would market a wine like that: it was simply for our instruction, but it was interesting that the 10% wine tasted quite good and the 12% sample was my preferred wine – the 14% version being a teensy fraction too alcohol-warm and clumsy for my personal taste. Fragos said he’d used RO in the past but not any longer, because “it does knock the wine around”. He said achieving lower alcoholic strengths was not as simple as picking the grapes earlier. “Flavour and tannin take longer to ripen than sugar takes to accumulate. If you just pick earlier you’ll get greener (astringent) tannin and less-ripe flavours. The challenge is to get flavour and tannin ripeness without pushing the alcohol up too much.”

RO units force the alcohol out through a membrane, but not the rest of the wine. The winemaker can back-blend the reduced alcohol fraction to the body of the wine to achieve the desired strength. Fragos did the trials because he wanted to see what impact alcohol had on tactile sensations.

He also showed the results of his trials on fermenter types and extraction techniques for red wine. “In winemaking you only get one shot each year; one vintage. I do trials to find better ways to do things or to audition new equipment. My aim is to bring finesse and complexity into our wines. In McLaren Vale, we tend to get fruit weight and richness every year, but we could improve texture, style and balance. Ideally, the end-result would be more food-friendly wines and more enjoyable wines.”

Perhaps the most fascinating of the five mini-tastings he staged for us was to trial three extraction techniques – three ways to extract tannin, color and flavour from the grape during red fermentations. The same 2010 McLaren Vale shiraz (tastings) was extracted in three ways. A roto-fermenter works like a cement-mixer, is totally mechanical and is the cheapest and most efficient method. But the wine had the lightest color, the simplest and most reductive nose, the tannins were disjointed and light, and the flavour was lightest and simplest. It’s a good way to make cheap red wine.

‘Pumping-over’ in a static tank is where the juice is recirculated by pumping it out the bottom and over the top of the ‘cap’ of skins – which tends to rise because of CO2 pressure. This resulted in deeper color, a more serious bouquet and deeper flavour with good, albeit slightly tough, tannins.

Finally, a static fermenter in which the skins are ‘hand-plunged’: the most costly, slow, and laborious method, but also the best. The result was a wine so superior to the roto-fermented portion that you’d never guess that grapes from the same batch were used. Deep colour, rich chocolaty bouquet; deep, full-bodied yet elegant palate with ample, smooth tannins which saturated the entire mouth, providing terrific texture, structure and contributing to persistence. It had the most tannin, but also the finest quality tannin.

It goes without saying the best Chapel Hill reds are made using this method!


First published in Sydney Morning Herald, Good Living – 27 July 2010.

Leave a Reply

Your email address will not be published. Required fields are marked *