Filtration Explored and Explained

In our recent piece on wine fining, we touched on filtration and promised a more thorough discussion of that topic. Here it is, at last. There are two principal reasons to filter wine, but some consumers are so afraid of filtration that many producers choose not to filter. We will do our best to present the facts and let you draw your own conclusions.

One reason to filter a wine is to improve its appearance. Allowing the wine to settle in barrel or tank is a great way to clarify it, but sometimes the spent yeast and other particles simply won’t settle to the bottom in the alloted time. Filtration can remove those particles and leave the wine limpid and bright. Another reason to filter is for microbial stability. No one wants to open a bottle and find it still fermenting, or nearly explosive with foul-smelling gases. A third type of filtration is performed by larger wineries only, to recover wine from the lees that settle to the bottom of the maturation vessel. This requires special equipment and enough lees to justify the investment, which most smaller producers just don’t have. We will ignore that here, but if you have questions about it, ask away.

Filtration tools and techniques—pad filtration

To understand how and why we might filter for appearance or for stability, it is helpful to understand how filtration works. There are two technologies in widespread use today. Pad filtration is the older technology. A series of cellulose (wood pulp, essentially paper) sheets are stacked in a frame and wine is forced through them. The tighter the weave of the cellulose, the more solids are left behind. One obvious disadvantage is that cellulose has flavor, and it is not a flavor that we would enjoy in our wine. This is easily overcome, however, by running (a lot of) water through the pads prior to introducing the wine. The water emerging from the pads is tasted and, when there is no more “papery” taste, it’s time for the wine. Another disadvantage is that the size of the gaps in the cellulose is imprecise. The filter sheets are sold with a nominal pore size, but there will always be a range of sizes. This can be of special concern when filtering for stability.

Typically a wine will go through several sets of pads, starting with bigger (looser) pores, and ending with tighter or smaller pores.

Crossflow filtration

A newer filtration technology is called crossflow because, rather than pushing the wine through the filter medium, it gradually passes through as it circulates. The filter medium is a membrane and it can have very precise pore sizes, making this a better choice for filtration for stability. Typically only one pass is needed, as opposed to one pass for each pore size as with pad filtration. Membrane filtration is also a necessary first step to treating a wine with reverse osmosis, which we will discuss below as a special application of crossflow filtration.

Filtration for appearance

To clean up or “polish” a wine, either technology will work. For pad filtration, the wine might first pass through pads with a nominal pore size of 5-7 microns (0.005-0.007mm) and then, if not “bright enough,” through another set of pads with 2-3 micron pores. Some winemakers who do this will still confidently claim that there wine has not been filtered, for reasons that elude us (but may be elucidated below).

Filtration for stability

Naturally, if the concern is to remove microbes, the pores have to be tighter. Wine is a wonderful beverage in that virtually no human pathogens can survive in it. Nevertheless, some microbes can survive in it and, while they pose no threat to human health, they can pose a threat to the taste of the wine. Lactobacillus spp., for instance, can metabolize any sugars left in the wine and produce acetic acid (vinegar smell) and carbon dioxide (gas, foaming). The acetic acid may form an ester with the ethanol in the wine to create ethyl acetate, which smells like nail polish remover. Brettanomyces yeast can use ethanol as fuel and transform polyphenol precursuors into nasty vinyl phenols that smell of anything from stale hay to horse manure. Not what we want to drink!

A 0.45 micron pore size is generally considered tight enough to remove bacteria (much larger yeast will have been removed at larger pore sizes). With pad filtration it usually takes 4 passes to filter to this level, but with crossflow filtration it can be achieved in one pass. Some winemakers prefer to go tighter still, to ensure that no bacterial spores can persist in the wine. 0.2 microns is considered safe for this. Again, the pore size for filter pads is nominal, and even at the tightest level there may be larger pores that could allow bad agents to pass, but at a minimum the filtration will dramatically reduce their population and thus the risk of contamination.

The downsides

Filtration sounds great, doesn’t it? So why are some consumers afraid of it, and why do some producers proudly claim they forego it? Much of the hostility towards filtration seems to be based in romantic ideas of a wine’s integrity, or even soul. Wine certainly inspires mystical thinking and that is part of its charm. But do odd bits of grape pulp and particles of dead yeast really comprise an integral part of the beverage? Dead yeast actually can contribute pleasant flavors and textures to a wine. That is properly the subject of another piece, but in brief, as the yeast cells die and autolyze, they release compounds into the wine that can enhance mouthfeel, making the wine feel more silky and rich on the palate. Still, that can and should happen before the wine goes into bottle. Leaving the spent yeast in the bottled wine just makes for a cloudy drink.

Another downside of filtration is the risk of oxidation. Anytime a wine is handled there is the chance for oxygen pickup. At best oxygen can make a wine seem prematurely “tired” or flat. Alcohol is oxidized to acetaldehyde, which smells like stale apples. The risk of oxidation is greater with pad than with crossflow filtration, not least because of the multiple passes required for pad filtration. With a skilled operator, however, the oxygen pickup is minimal.

Perhaps the biggest concern with filtration has to do with its affect on a wine’s structure. But this, we will see, is based on a misunderstanding. Many elements contribute to a wine’s structure, which is how it appears on the palate, or its “three-dimensionality.” Tannins give the wine its grip, and ethanol can make it seem somewhat thick, as can glycerol and any residual sugar.

Colloids also contribute to wine’s structure and mouthfeel. Colloids are associations of large molecules, and they cannot pass through a filter intact. Tasting a wine that has just passed through a filter can be a shock. The mouthfeel is completely changed, usually for the worse. We suspect it is this mistake—tasting just after filtration—that contributes most to filtration’s bad rap. However, the components of the colloids do pass through the filter, and with time they re-form the colloids. All better now.

Why we filter

We started filtration with the 2013 vintage. We have had no reports or complaints regarding the appearance or stability of our earlier, non-filtered wines, or for that matter, of our later, filtered wines. We decided to filter because we wanted to be sure that the wines you enjoy are as we intended them, with nothing nefarious happening before you open that bottle. We hope that this piece has helped you to understand our rational for filtering, and what the process entails. Please chime in with any comments or questions you have on the subject.

Post-script—reverse osmosis

Crossflow filtration can be so tight that it can discriminate among molecules by size. The principal components of wine: water, ethanol, and organic acids such as tartaric and malic acids, are all quite small by molecular weight. So are most volatile compounds, which contribute aromas. When bad things happen to good wines, unpleasant volatile compounds are often produced. Some examples are acetic acid and ethyl acetate, the volatile phenols mentioned above, and compounds related to “smoke taint,” such as 4-methylguaiacol. Wines can pass through a crossflow filter and then over resins that selectively bind these compounds. With the bad actors gone the treated path is reunited with what did not pass through the filter to re-create the now restored wine. This technique can also be used to adjust the alcohol content in the wine. The permeate (the portion that passed through the filter) is distilled to remove the alcohol before being rejoined with the retentate (the portion that did not pass through the filter). Pretty crazy stuff!