Filtration Types: Coarse, Fine & Sterile/Membrane

The OIV defines filtration as a physical process of passing wine through filters that retain particles in suspension, with two objectives: clarifying the wine, if necessary by successive stages, and obtaining biological stability by eliminating microorganisms. In practice, per the AWRI, coarse depth filters of earth or cellulose come first, wines are then often filtered through sequentially tighter pad depth filters, and surface filters such as membrane filters are used as a final filter to remove smaller microorganisms. How much to filter is argued over: the AWRI says its commercial-scale trials suggest that common filtration practice has no impact on red wine colour and minimal impact on sensory profiles, while the fashion to bottle 'unfined and unfiltered' has spread across the world, according to Decanter.

Definition and the Filtration Stages

The OIV defines filtration as a 'Physical process consisting of passing the wine through appropriate filters that retain particles in suspension.' Its two stated objectives are clarifying the wine, if necessary by successive stages, and obtaining biological stability of the wine by eliminating microorganisms (sterilising filtration). Coarse depth filtration, the first stage, uses diatomaceous earth (kieselguhr) or cellulose to remove suspended solids. Following that first stage, the AWRI notes, wines are often filtered through sequentially tighter pad depth filters. Sterile membrane filtration, the final stage, uses surface filters made of synthetic polymers such as polyethersulfone (PES) or polyvinylidene fluoride (PVDF) to guarantee microbial removal before bottling.

  • Coarse: depth filters of diatomaceous earth or cellulose, applied to a rotary drum vacuum filter or a plate-and-frame housing; they retain particulate matter but generally do not remove microorganisms (AWRI)
  • Fine/Polish: sequentially tighter pad depth filters, including lenticular filters, which are pad filters reformatted into a filter housing (AWRI)
  • Sterile/Membrane: absolute-rated surface filters such as 0.65 µm and 0.45 µm ('sterile'); the AWRI describes absolute filtration as useful for sterile filtration at the bottling line
Key Facts
  • The OIV International Code of Oenological Practices (sheet II.3.2.2) lists three ways to filter wine: by continuous deposition using additives such as diatomaceous earth, perlite or cellulose; on pads of cellulose or other appropriate materials; and on mineral or organic membranes (microfiltration).
  • Depth filters are not absolute filters, per the AWRI: they trap particles by mechanical retention, adsorption or electric charge, generally within the filtration medium, and include diatomaceous earth, perlite and cellulose pads, with grades of diatomaceous earth ranging in particle size from 2-40 µm.
  • Surface filters are generally absolute filters with precisely defined maximum pore sizes, including 0.65 µm and 0.45 µm or 'sterile' filtration, and they remove particles like a sieve (AWRI).
  • Yeast and bacteria are around 0.5-5 µm in diameter, too small for the human eye, which can detect large particulate matter above 100 µm (AWRI).
  • Bowyer (2018), cited by the AWRI, reports that around 80% of all wine by volume is now sterile filtered, compared with 20% ten years earlier.
  • Most cross-flow membranes are only nominally rated at 0.2 µm and cannot be pressure tested for integrity before use, so wineries still need a membrane filter to achieve sterile filtration (AWRI).
  • In AWRI work on four commercial Cabernet Sauvignon and Shiraz wines (McRae et al. 2017), 0.45 µm filtration lowered polysaccharide concentration while tannin and colour remained unchanged, and the AWRI says the results suggest that commonly applied commercial filtration has no impact on wine colour and minimal impact on the sensory profiles of red wines.

Sensory and Stability Trade-offs

The AWRI calls it 'a long-held belief in the wine industry' that tight surface filtration below 0.45 µm strips aroma and colour from big red wines, and notes that for this reason many winemakers avoid sterile filters, which can cause uncertainty about the microbial stability of the finished wine. Depth filters of diatomaceous earth or cellulose are polar and bind polar colour and phenolic components, so some colour loss can be expected. Most modern surface filters are made of nylon, PES or PVDF, which are non-polar and cause minimal adsorption of colour, per the AWRI citing Bowyer et al. (2013). In AWRI trials on four commercial red wines, published by McRae et al. (2017), cross-flow filtration significantly reduced average particle size and 0.45 µm filtration lowered polysaccharides, while tannin and colour remained unchanged.

  • Depth filters (diatomaceous earth, cellulose): polar materials that bind colour and phenolic components, so some colour loss can be expected (AWRI)
  • Membrane filters (nylon, PES, PVDF): non-polar, causing minimal adsorption of colour (AWRI)
  • Microbial control: the AWRI states that microorganisms are typically controlled by a combination of pH/SO₂ and filtration
Ad hidden for this session.

Filtration Methods and Technical Execution

Traditional depth filters use an earth or cellulose matrix in a rotary drum vacuum unit or a plate-and-frame housing. Today many winemakers bypass depth filtration with centrifugation, flotation or cross-flow filtration, and the AWRI notes that depth filters have gradually been replaced by centrifuge and cross-flow alternatives, mainly to avoid the use of diatomaceous earth (kieselguhr). In filtration by continuous deposition, the OIV describes a filter bed that is fed constantly by a continuous addition of filter material such as diatomaceous earth, perlite or cellulose, with the dose set by the turbidity of the wine and the clarification sought. Depth filter sheets, per the OIV Codex, are made from cellulose fibres with filter aids such as diatomaceous earth and perlite, and are applied to particles ranging from 0.1 to 40 µm. Cross-flow (tangential-flow) filtration is a surface filter that allows high throughput because its tangential flow continually cleans the filter surface. The AWRI calls it important to test the integrity of a filter membrane, to confirm there are no holes or leaks in or around the filter, generally with non-destructive techniques such as the bubble point test, which can be run at the start of a bottling run, or the pressure-hold test.

  • Sheets and lenticular modules: the OIV Codex notes that depth filter sheets can be used as flat sheets or in lenticular filter modules
  • Turbidity guide: supplier recommendations collated by the AWRI run from coarse filtration at 35-100 NTU and polishing at 12-35 NTU to final filtration of dry wines at 1.0-3.0 NTU, with sweet or higher-risk wines filtered below 1.0 NTU
  • Temperature: the AWRI notes that filtering at 8-12°C doubles the difficulty compared with 20°C and recommends aiming for 15-21°C, or around 18°C
  • Cross-flow: most cross-flow membranes are nominally rated at 0.2 µm and cannot be pressure tested before use, so a membrane filter is still needed for sterile filtration (AWRI)
  • Blockages: wines below 1.0 NTU can still block membranes, most often because of polysaccharide complexes, more commonly associated with Botrytis infection; glucanases or other enzymes can break these down (AWRI)

Identifying Filtration Impact in the Glass

The AWRI treats a wine as 'cellar bright', or visually clear to the eye, below approximately 30 NTU, and a commonly used threshold for bottling is below 1.0 NTU. Filtered or not, Steven Spurrier wrote in Decanter, many fine wines, especially reds, throw a natural deposit of tannins and colouring pigments, and some white wines shed a crystalline deposit due to a precipitation of tartrates. Aroma and texture effects are harder to pin down: in the AWRI's red wine trials, sensory analysis showed no consistent filtration-related trends in textural attributes, though some aroma or flavour attributes differed between filtration grades within each wine. Bordeaux consultant Denis Dubourdieu, speaking at a tasting reported by Decanter, cited the 'horse-like' aroma of brettanomyces as a risk of unfiltered wine.

  • Clarity: the AWRI's table, after Ribéreau-Gayon et al. (2006), rates white wine brilliant below 1.1 NTU, rosé below 1.4 NTU and red below 2.0 NTU
  • Deposit: many fine wines, especially reds, throw a natural deposit whether filtered or not (Decanter)
  • Texture: in AWRI trials, 0.45 µm filtration lowered polysaccharide concentration, but sensory analysis found no consistent filtration-related trends in textural attributes
  • Bottle ageing: after 18 months in bottle, filtered and unfiltered 2013 wines in the AWRI trial had similar average particle sizes, while the filtered 2014 wines contained smaller particles than the unfiltered wines

Philosophy and the Producer Divide

According to Steven Spurrier in Decanter, the fashion to bottle 'unfined and unfiltered' began among careful growers in Burgundy and the Rhône Valley and has spread across the world. Decanter also notes that many natural wine producers do not fine or filter their wines. On the other side, Bordeaux consultant Denis Dubourdieu said, as reported by Decanter, that leaving wine unfiltered is the equivalent of landing a plane without help from the control tower, and argued: 'Great wines have nothing to gain from the absence of fining and filtration.' He made the case at a tasting of Chateau d'Yquem, Chateau Cheval Blanc and other properties he consulted for, all of which were either filtered or fined. The AWRI puts the industry trend in numbers: Bowyer (2018) reports that around 80% of all wine by volume is now sterile filtered, compared with 20% ten years earlier.

  • Unfined and unfiltered: a fashion that began among careful growers in Burgundy and the Rhône Valley (Decanter)
  • Natural wine: many natural wine producers do not fine or filter their wines (Decanter)
  • Compromise approach: Bordeaux consultant Denis Dubourdieu said, as reported by Decanter, that 'Filtration promotes microbiological stability by keeping out yeasts and bacteria, while being just sufficient to leave smaller elements in the wine useful to the flavor.'
From the WineShopAffiliate

Open one bottle, taste it for weeks.

The Coravin Pivot+ is the simpler Coravin, for wines you will finish inside a few weeks, which is how I taste a producer across vintages without finishing every bottle. It takes five seconds to use.$199.

See the Pivot+

Practical Implications for Collectors and Consumers

Steven Spurrier wrote in Decanter that with many fine Burgundies bottled unfiltered, decanting is necessary for clarity, and he advised standing the bottle up a few hours beforehand so the sediment falls into the punt. For winemakers, residual sugar raises the stakes: the AWRI recommends that sweet wines be sterile filtered (0.45 µm membrane filtered) at bottling, and notes that sweet wines may need an additional treatment because of the higher risk of refermentation if an infection occurs after sterile filtration. The AWRI also advises that wines being sterile filtered be bottled immediately, since filtering to a packaging tank and bottling several days later allows time for microbiological growth.

  • Unfiltered bottles: stand the bottle up a few hours before serving, and stop pouring when the sediment flows towards the neck (Decanter)
  • Sweet wines: sterile filtration (0.45 µm) at bottling is the AWRI recommendation; apart from SO₂, the main chemical treatments available are dimethyl dicarbonate (DMDC) and sorbic acid
  • Sweetened wines: the AWRI advises bottling within a week of sweetening, filtered in line to the bottling line

Sources and last verified

Primary sources (the OIV and the Australian Wine Research Institute): https://www.oiv.int/standards/international-code-of-oenological-practices/part-ii-oenological-treatments-and-practices/wines/filtration, https://www.oiv.int/standards/international-code-of-oenological-practices/part-ii-oenological-treatments-and-practices/wines/filtration-by-continuous-deposition, https://www.oiv.int/standards/international-oenological-codex/part-i-monographs/monographs/filter-sheet, https://www.awri.com.au/industry_support/winemaking_resources/storage-and-packaging/pre-packaging-preparation/filtration-physical-removal-of-microorganisms/, https://www.awri.com.au/industry_support/winemaking_resources/storage-and-packaging/pre-packaging-preparation/microbiological-stabilisation/. Secondary sources (Decanter): https://www.decanter.com/wine-news/non-filtration-like-flying-unaided-dubourdieu-94761/, https://www.decanter.com/learn/winemaking-the-facts-2-54064/8/, https://www.decanter.com/learn/advice/what-is-fining-51651/, https://www.decanter.com/learn/how-to-decant-53904/. Last verified 2026-09-28

Ad hidden for this session.
Exam Study NotesWSET / CMS
  • OIV sheet II.3.2.2 defines filtration as a physical process that passes wine through filters retaining particles in suspension; its objectives are clarifying filtration and sterilising filtration (eliminating microorganisms for biological stability).
  • Stages (AWRI): coarse depth filters of earth or cellulose first, then sequentially tighter pad depth filters, then surface filters such as membrane filters as the final filter to remove smaller microorganisms.
  • Depth filters are not absolute filters and trap particles generally within the medium; surface filters are generally absolute, with precisely defined maximum pore sizes such as 0.65 µm and 0.45 µm ('sterile').
  • Most cross-flow membranes are only nominally rated at 0.2 µm and cannot be pressure tested before use, so sterile filtration still needs a membrane filter; membrane integrity testing generally uses non-destructive techniques such as the bubble point test or the pressure-hold test.
  • AWRI trials on commercial red wines (McRae et al. 2017) suggest common commercial filtration has no impact on colour and minimal impact on sensory profiles; the AWRI recommends sterile (0.45 µm) filtration at bottling for sweet wines.

Put this into practice

Blind tastings, flashcards, and trivia in the Wine with Seth app.

Try the app free