Crossflow filtration and the Barboursville Vineyards (Barboursville, VA) experience: A conversation with Luca Paschina, General Manager-Winemaker

Winemakers are largely divided between those who filter their wines and those who do not. Those who eschew filtration are concerned that the practice can strip out aroma and flavor molecules, thus rendering the finished product less-than-optimal. In his argument against traditional filtration, Clark Smith, author of postmodern winemaking, states thusly:

The focus of postmodern philosophy is the creation and preservation of beneficial macromolecular structure. This structure manifests in wine as colloidal particles sometimes nearly as large as a bacterial cell. The benefits of good structure -- profundity, aromatic integration, and graceful longevity -- appear to be lost in sterile filtration, despite the fact that no tannin material may be retained by the filter. While this lack of residue has convinced some of my colleagues that filtration cannot be harmful to wine structure, I do not concur. My hypothesis is that the action of tight filtration somehow disrupts rather than removes structure.

In a January 2003 article, Smith  described a class of filtration systems which he called the Tangential Flow Family of Filtration. This family is shown in the table below, classified based on the molecular weight of particles that pass through the pores.

Tangential Flow Filtration

Filtration System	Application	Molecular Weight Range (Daltons)
Crossflow Clarification		200,000 - 500,000
Ultrafiltration		1000 - 200,000
	Tannin and Browning Removal	10,000 - 200,000
	Protein Removal	10,000 - 40,000
	Decolorization	1,000 - 5,000
Nanofiltration		200 - 1000
Reverse Osmosis		50 - 200

Source: Clark Smith, The Crossflow Manifesto, Wine Business, January 2003.

According to Smith, the idea of tangential flow filters developed in the 1960s. One of the major problems with sterile filtration is the fouling of the membrane which occurs when tight pore sizes are used. This fouling prevents the passage of material through the pores. The effective limit of traditional filtration is 0.1µ. Tangential flow filters use the scrubbing action of the flow across the surface of the membrane to keep it clean thus allowing the utilization of ever-smaller pore sizes.

All of the systems mentioned in the table employ the strategy of pumping the wine across the membrane at high velocity. As the wine flows across the membrane it continually scrubs the surface, removing fouling material. The majority of the feed stream does not pass through the filter but is retained upstream and returned to the tank. This stream, called the retentate, contains all of the high-molecular-weight components. The low-molecular-weight material that passes through the filter is called the permeate. A comparison of traditional (dead-end) versus tangential filtration is shown in the figure below.

Dead-end versus tangential-flow filtration
(Source: cleanflowfilter.nl)

The Holy Grail of tangential filtrations, according to Smith, "is to be able to clarify wine without harming its structure, and crossflow clarification ... continues to gain steam." In his opinion, the technology works well for unstructured whites, "where a little tannin and color stripping is a good thing, but can prove disastrous for structured reds."

I now turn to my discussion of crossflow with Luca Paschina of Barboursville, a follow-up on a wide-ranging discussion of the estate and its wines. As stated above, Smith has some problem with crossflow filtration and I wanted to see whether his views were being validated at Barboursville.

Luca first became aware of crossflow technology around the year 2000 and was immediately impressed. Barboursville first employed the Bucher-produced system in 2004 in an attempt to clarify its wines in a one-step process. Prior to its implementation of crossflow, the winery was going through at least two levels of filtration and wasting a lot of time and product. The process employed for filtration prior to the acquisition of the crossflow technology was diatomaceous earth (DE) and different levels of pad porosity, staples of the dead-end approach. The issues associated with this approach are catalogued below.

In addition to the one-and-done aspect of crossflow, Barboursville also uses it to remove yeast and bacteria in white wines (thus avoiding the potential of malolactic or yeast fermentation of residual sugar in the bottle) and in reds to avoid Brett bloom in the bottle.

The technology is used on white wines after stainless steel aging on lees, allowing for filtering of the fine lees. These wines are filtered the day before bottling. Red wines are transferred from barrels to tank and then filtered the day after (or later if necessary). Crossflow filtration is applied to every wine once per week between mid-January and late July.

Barboursville purchased its crossflow equipment and paid it off over 5 years. The technology is easy to use but requires an operator with an attention to detail and the ability to follow procedures. The winery is very satisfied with the equipment and the associated process. Staff has adapted positively to the implementation of the product.

In terms of additional advantages, Luca can time his bottling schedule with more versatility (such as filtering today and bottling tomorrow), with no surprises arising from filtration difficulties.

In closing, Luca mentioned that he had initially been skeptical of the product but after seeing how the wines were aging gracefully, he is now a firm believer in the technology.

My conclusion: Luca is using crossflow in an effective and traditional manner (that is, replacement for a DE system) and is currently very satisfied. None of the Clark Smith concerns are evident here.


©Wine -- Mise en abyme