More About Filtration…

Pacific Sun Farms Olive Oil Maker Pablo Voitzuk
Responds to the UC Davis Olive Center Report

The UC Davis Olive Center’s October 2015 report on filtration was a work based on compiling and processing literature on the subject; no field tests were done by the Olive Center or the writers of the paper.

I’d like to present a constructive criticism to that report in the spirit of necessary discussion on the subject, because I consider filtration a key part of quality and shelf life in EVOO.

The main two reasons for filtering are:

  1. To remove water
  2. To remove sediments

Moisture Content As Defined by the International Olive Council

The IOC has identified that in order to guarantee durability, EVOO needs to have a water content of 0.20 % or lower. This can be assessed with a simple lab analysis or at the mill, with the proper equipment. Most olive oils have a water content of 0.40%-0.50% right after they’re made. Thus, something has to be done if a good shelf life from harvest to harvest is expected.

“Suspended solids are usually small olive fragments. They remain in EVOO after the extraction, and they consist of proteins, sugar glycosides, sugar bounded with proteins, phospholipids and phenolic compounds.” [2-3]

Some of the above mentioned compounds have amphiphile behavior showing both lipophilic and hydrophilic properties. They organize themselves in colloidal forms (mainly in reverse micelles and in lamellae, since these forms are energetically preferred) and aggregate at the interphase between water and oil [9]. Colloidal suspended solids have a large reactive surface and, containing both water and oil, allow the presence of both lipophilic and hydrophilic compounds. The size of such colloidal formations have been studied by Papadimitriou et al [5]. who relate the average size of these suspensions with the extraction process. Suspensions give EVOO a cloudy appearance and cause hydrolytic reactions which could result in sensory defects [5].

Furthermore, solids and water in unfiltered EVOO allow the development of a microflora, mainly represented by yeasts [10]. Before olive processing, microflora lives on the olive carpophore, and migrates into the oil together with the olive fragments and water. The substances on the olive fragments and the water allow these micro-organisms to survive. In just a few hours after the olive oil extraction, some yeasts such as Candida and Saccharomyces are able to colonize the environment [11]. The selected microflora could survive during the entire storage period, and could damage the sensory characteristics of EVOO. It has been demonstrated that some lipase-producing yeasts can hydrolyze triglycerides in different manners according to the size of suspended micro-drops of vegetation water [11-12]. The presence of water and solids together with some micro-organisms could lead to the formation of the “muddy-sediment”, “rough”, and “rancid” defects after a short storage time [13].” NOTE: this is a part of an in press chapter on olive oil stabilization in a book entitled “Products from olive tree” by Alessandro Parenti and Lorenzo Guerrini.

Once the latter was understood, different filtration systems were put into practice. Not of all them were successful.

The first attempts at filtering were often faulty. Unfiltered olive oil was kept in tanks and filtered as it was bottled. This was actually counter-productive, since it implied mixing back the sediments that were in the bottom (which exists even after racking) with the cleaner olive oil on the upper levels of the tank. More importantly, that did not prevent all the degrading processes that take place when water content is higher than 0.20% (enzymatic action/hydrolysis, oxidative degradation – of lipids and polyphenols – and fermentative activity from sediments carrying bacteria, sugars, yeast, protein, etc.).

The same can be said about racking and when unfiltered oil is kept in conical tanks that get rid of most sediments through a valve in the bottom that collects them.

So far, the best results are achieved when filtration takes place immediately, and with the proper equipment.

Here is my first objection to the report: it generalizes filtration. While they identify several methods, the conclusions are presented as if any type of filtration would produce those results. This is not valid for either the positive or negative aspects noted in the report.

I’m going to limit myself to what I studied and practiced for several years: filtering with cellulose paper, both with rolls and with a plate filter. This type of filtration aims at shelf-life (decreasing moisture and hygiene of the olive oil) and at quality as well (improving organoleptic characteristics).

Backing my own practice, I based this article on a paper recently produced in Florence (1) by a group of researchers, olive oil tasters, and producers linked to the Laboratory of Chemistry of the Florence Chamber of Commerce, the University of Florence, and published by the European Journal of Lipid Science and Technology.

On My Practical Experience

We’ve been using two filters specifically designed for olive oil filtration.

The first one, the filter “Più verde” (“Greener” in Italian) consists of two rolls of cellulose paper (one with a smooth surface, the other one with a rough surface) that go together around five plates/serpentines, using a pump, and then gravity. Given that the amount of paper the olive oil runs through is rather small and thin, this is an excellent filter for catching sediments though not enough for removing humidity.

It has been designed to filter “in continuum”, as the olive oils leave the final separator or the decanter, depending on the mill.

A pre-filter made of 3 vertical stainless still cartridges (with holes of 0.50, 0.25 and 0.05 microns) was tested and proved useful to retain sediments and therefore diminish paper use and waste, as well as olive oil loss. Given the amount of olive oil produced at our facility, we discontinued the use of the pre-filter for practical reasons (mostly labor costs and pace of production) even when we had good results in terms of filtration and reducing the use/cost of paper as well as reducing olive oil loss.

The filtration process is completed when the oil is run through approximately 40 square (40 cm x 40 cm) cardboard (cellulose) filter plates. (One side of the plate is smooth and the other side is rough.) This number can be lowered, according to what is needed, by removing some plates.

This second filtration removes the smallest sediments and takes care of humidity, lowering water content drastically. This last passage is ideally done right after or the following day after the first filtration. Since a first filtration took care of the biggest sediments, one setting of paper filter goes a long way, filtering around 750 gallons, depending on cultivar and temperature (colder olive oil slows the process and consumes more paper).

The ideal temperature to work with is between 17C-26C (62-78F).

Once filtered, the oil is ready to be stored or bottled, with no further need for racking (which, by the way, also brings costs and oxygenation).

About the UC Davis Report

Regarding some statements in the report by UC Davis on supposedly negative effects of filtration:

On Oxygenation and Sediments

While is true that there’s some oxygenation and even a degree of emulsion that goes along with filtration, these are a minor impact given the benefits and protection filtration offers.

The report also says that filtration decreases stability due to the removal of suspended solids. Attributing a positive value to those solids is, at least, a debatable if not a questionable idea. Sediments have a clear decomposing action. They’re vehicles for droplets of vegetable water, air, mold, yeast, sugars, proteins, mucosae, microbes, fungi: “The water and suspended particulates also contain microorganisms, such as bacteria, yeasts and mold (2,3), which directly contribute to the transformation of the nutritional component of the oil. These micro-organisms may contain enzymes responsible for the hydrolysis of triglycerides (15), namely lipase, enzymes responsible for fatty acid oxidation (10), peroxidase, and those responsible for the degradation of phenolic compounds, namely b-glycosidase, esterase, and polyphenol oxidase (4). It should be noted that water must be present for enzyme activities (4).”

In fact, after six months of storage, FFA is higher in unfiltered oils whereas it’s stable in filtered ones: “After 6 months of storage, the difference is roughly 0.1% and became significant. Lipases (NA- oxidation of the lipids) act in the interface between water and oil with the following mechanism: the hydrophobic part of the enzyme bounds with the oil, whereas the active site aligns with the substrate and severs the ester bounds of the triglycerides (5). Hence, the free fatty acids increase is probably due to the water content of cloudy oils, which allows lipase enzymes to hydrolyze triglycerides during the storage period. These results are in accordance with the work of Fregapane et al. (6), where they report that filtration reduced the rate of hydrolysis of the triacylglycerol matrix” (1).

Sensory Analysis and Filtration

Sensory analysis confirms the previously mentioned difference regarding FFA and faster decay in unfiltered EVOO, as the evolution of unfiltered olive oils is one of a loss of vitality first (diminished bitterness and pungency due to polyphenol oxidation), then higher viscosity, with oxidation manifesting itself as a greasy mouth feel.

Unfiltered olive oils get the “rough/oxidized” defect, precursor of rancidity, faster than filtered olive oils.

Oxidation can happen for different reasons: low ratio of oleic acid vs linoleic and other acids, lack of phenols, photo-oxidation, etc. In this case it will come from the oxidative processes brought by enzymatic reactions in a too humid environment (hydrolysis and oxidation of lipid and phenols).

This will be more or less perceptible depending on other factors, such as the original health of olives prior to milling, milling temperatures, type of equipment, and cultivars. Nevertheless, unfiltered olive oils, including the ones of cultivars with a high polyphenol count, will be more prone to this greasy mouth-feel, evidencing some degree of oxidation.

The words used in Italian and Spanish are quite interesting in terms of describing oils which have suffered oxidation: grossolano and basto, both meaning “coarse”. These oils often leave a greasy film on the lips of the taster, a major sign of oxidation.

Dr. Lorenzo Guerrini, a researcher from Italy, explains:

“Based in our tests, the only differences that could be detected immediately after filtration are related to turbidity and total solids. In the subsequent analyses, filtered oil retains superior nutritional and sensory characteristics than unfiltered oil. More importantly, filtered olive oil retains the standards required for categorization as “extra virgin olive oil” for longer.”

“In one of the tests two cultivars have been analyzed: Arbequina and Coratina. Samples were taken after one month and after six months from production. In Arbequina, after one month the filtered EVOOs have lower K232 (lower concentration of conjugated dynes) and lower tocopherols than unfiltered EVOOs. After six months passed tocopherol count was the same. On the other hand, filtered EVOOs showed lower K232 and lower peroxides number (10.3 vs 7.9 meqO2/kg). In Coratina, no difference was found after one month. After six months passed we found lower content of peroxides in filtered oils. A higher number of tyrosol and hydroxytyrosol was found in the unfiltered olive oil. This also indicates a higher oxidative activity because these secondary phenols appear as a consequence of the degradation of the secoiridoidic fraction [9] (oleoeuropein and ligstrosides and their derivatives)”. (NA-: This is explained in length in the next subject).

On Polyphenol Loss Due to Filtration

The report correctly points out that there’s a reduction of water-soluble phenols (tyrosol and hydroxytyrosol) when filtration takes place. Then, it assumes that this will reduce shelf-life.

While it’s true that filtration implies the loss of some water-soluble phenols (7-8%), this is altogether an insignificant loss, offset by the gains filtration brings. It’s worth noting that not all phenols are equal in their anti-oxidant action. Water soluble ones, still important anti-oxidants, are more prone to deterioration than fat-soluble phenols, being these last ones of higher power as anti-oxidants.

Filtration provides oil-soluble polyphenols (oleoeuropein and ligstrosides) with serious protection against oxidation triggered by enzymatic reactions enhanced by water. These phenols are the ones really in charge of anti-oxidative action: “Other works pointed out that a reduction in the total phenolic content [7] or in certain phenolic alcohols [8] could occur immediately as a result of water removal. In the present work, the total phenolic content was 337 mg/kg in cloudy oil, and 313 mg/kg in ?ltered. This difference is not signi?cant at the paired t-test. Over time, ?ltration affects both total phenolic content and each phenolic compound. The effect was particularly pronounced for phenolic alcohols, namely tyrosol and hydroxytyrosol, which content in unfiltered samples reaches values up to an order of magnitude greater than that in ?ltered oils. The increase in phenolic alcohols in unfiltered oils over time is a consequence of the degradation of the secoiridoidic fraction [9] (oleoeuropein and its derivatives, and derivatives of ligstrosides), which decreases in unfiltered oils (…) Regarding oleuropein derivatives, there was a signi?cant increase in hydroxytyrosol and a concomitant decrease of dialdehydic form of decarboxymethyloleuropein aglycones in cloudy oil. On the other hand, in ?ltered oil, hydroxytyrosol did not change signi?cantly, whereas there was a signi?cant increase in oleuropein aglycones similarly to unfiltered oils. The different changes in ?ltered and unfiltered oils of oleuropein derivatives could be explained by the deglycosilation of oleuropein, due to the action of b-glucosidase, mediated by water [4]. The formation of tyrosol and hydroxytyrosol is due to the rupture of the ester bond in the dialdehydic form of decarboxyme- thylligstroside aglycon and the dialdehydic form of decar- boxymethyloleuropein aglycon, respectively. This is a hydrolytic reaction, favored by the presence of water [4]. For this reason, the degradation of phenolic compounds is more pronounced in unfiltered olive oil, in which there was a signi?cant increase in tyrosol and hydroxytyrosol, and a concomitant decrease in the dialdehydic form of decarboxymethylligstroside aglycon and the dialdehydic form of decarboxymethyloleuropein aglycon. As a consequence of these phenomena, the nutritional value of unfiltered oils rapidly decreases, whereas the ?ltration allows to maintain a high nutritional value for the filtered oils over time (1).

On Aromas

The UC Davis study also says that filtration “eliminates desirable compounds that affect the aroma of the oil”.

Here is what this recent study, done in Florence, says about the chemical side of this issue: “On the other hand, the volatile fraction was influenced by filtration. Table 2 shows the mean values of some volatile molecules, selected on the basis of the study by Kalua et al. [10]. It has been established [11,12] that the most abundant volatile compound related to the LOX pathway is E-2-hexenal. The content of this molecule, and that of Z-3-=hexenal, is higher in just filtered oils than in unfiltered oils. Over time, these aldehydes significantly decreased in unfiltered oil, whereas they remained constant in filtered oil. As the fruity attribute is closely linked to these molecules [11], we conclude that filtration enhances this sensory attribute. On the other hand, the alcohols E-2-exenol and hexenal, and the ester E-2-hexenyl acetate were more abundant in cloudy oils. This phenomenon could be explained by the inhibition of enzyme activity of alcohol dehydrogenase and alcohol acetyl transferase due to the removal of water during filtration, which, over time, preserves C6 aldehydes known to be responsible for the “green” aroma [13]. The concentration of ethyl acetate, considered to be a winey marker [38, 42] was consistently higher in unfiltered oils than filtered oils; moreover, it only increased over time in the case of non-filtered oils (1)”.

More on Sensory Analysis

Another negative point, according UC Davis, is that filtration “decreases positive attributes (fruitiness, bitterness, pungency).”

This statement provides for an interesting discussion that will address the previous point as well. Aside from the chemical data offered by the study in Florence, sensory analysis can be of help in this case, since we’re discussing exactly that. And here is where a some of these studies cited by UC Davis are outdated: the only way to determine if filtered oils fare better or worse than unfiltered ones in terms of organoleptic virtues, is by comparison in panel tasting sessions. Proper filtration and oils of true excellence are a very recent phenomena and literature falls short on this type of study, aside from the study done in Florence (and this one clearly favored filtered oils as storage-time passed by).

Since filtered olive oils of high quality are a rather new thing, it’s no surprise that the European panels in contact with them (the majority in Italy and some in Spain) are the ones abler to more rapidly identify the virtues of these in contrast to the old fashioned characteristics of unfiltered ones. This trend fell onto panels trained primarily for detecting quality (or the absence of it), not just for EVOO certification. These panels would appreciate specific, central key elements in the evaluation of EVOO such as freshness, cleanliness, neatness, structure, durability, and greenness. Since filtration is functional and synergetic to these characteristics, tasters familiarized with these concepts and values can acknowledge the benefits of filtration more easily than tasters who have not been exposed to this approach.

In California, 98%-99% of the EVOO produced is unfiltered and therefore, local tasters have not had many opportunities to learn how to discriminate one lot from the other. Interestingly enough, the practice of filtration gained serious traction in Italy when blind tasting sessions consistently favored filtered olive oils.

Filtered oils have quite a distinct texture: leaner, cleaner, and crisper than unfiltered ones, because filtration removes non-olive oil matter and thus allows for a neater experience of positive attributes.

Tasters trained to distinguish filtered vs unfiltered become sensitive to the negative sensory aspects of the non-removed vegetable water, while tasters unaware of it tend to compute its presence positively, as a “harmonizing” factor. In reality, what the remnants of vegetable water do is numb flavors while bringing a viscosity that Italians describe as pastosità (doughy-ness).

It’s worth noting that for many years, some tasting schools (associated with areas where mass-marketed olive oil was produced and not much quality olive oil) denied that texture was an element of olive oil tasting. In fact, the “rough” defect is a recent addendum by the IOC to their tasting sheet. This also shows how sensory analysis of EVOO is an evolving discipline which has been changing in recent years, along with newer milling technologies becoming available.

Bitterness gets rounded up with filtration, contributing to what some experts in Italy call “olive bitterness” (though achieving this requires optimal ripeness and proper milling as well). The latter is quite different from bitterness from vegetable water, wood, debris or more commonly, from secondary oxidations of oleoeuropein (which renders the tannic, harsh bitterness).

The removal of waste water and sediments allow for flavors to be sharper and overall the filtered olive oil gains in elegance.

The same goes for aromas. In order to appreciate them, tasters need to be trained in discriminating aromas coming solely from olives from aromas and odors (even pleasant ones) coming from other sources. This applies to decaying processes such as oxidation, that produces – momentarily – over-ripe fruit aromas that can cause some intensity for a while.

The effect filtration has on some cultivars, including Mission and Manzanillo (particularly when working with older trees), is remarkable, helping to reduce some rustic sides of the olive oils coming from these varieties. The same can be said for robust cultivars such as Coratina, Picual or Moraiolo.

Finally, and even this being not proof of anything, the fact that the vast majority of winners of European competitions, as well as winners of the international section of US competitions, are filtered oils is something to consider (14).

On Color and Chlorophyll Content

The next two points made by UC Davis are the more curious ones. They said that filtration diminishes the green color of the olive oil.

I believe that the confusion originates when filtration was done before bottling and sediments were remixed with the olive oil. I imagine that some papers were produced after those experiences. Another possible assumption is the belief that by removing some chlorophyll present in the waste water greenness will be diminished.

Suspended particles and the waste water removed have some chlorophyll, though this rapidly turns yellow (meaning it has oxidized) due to hydrolysis. Thus, the removal of sediments and water only removes some greenness in the moment of filtration, though guarantees a greener oil over time.

The only thing to consider about this issue is that removing the floating particles and sediments leaves the filtered oil more vulnerable to photo-oxidation, as the particles screen a degree of the light. As we well know, this is best to avoid with both filtered and un-filtered olive oils. Proper storage will take care of this. 

It’s also true that after filtration, olive oil is more transparent and glowing, presenting an altogether different shade of green. Nevertheless, it is not less green, but rather the opposite, since sediments tend to brown the color of the olive oil. One could wonder why the Italian filter we use is called “Greener” if using it would produce the opposite.

(NA—It’s also pertinent to say that color in EVOO depends also of other factors: cultivars, ripeness, type of crusher, milling practices, etc.)

Final Considerations

It’s only fair to mention that the report acknowledges some studies that see filtration, or aspects of it, in a positive manner. It lists some important benefits such as diminishing moisture, increasing stability, diminishing FFA, decreasing hydrolysis and therefore extending shelf life, and that filtration eliminates undesired compounds that affect aromas. While all these are true, each positive effect is presented with an equal or greater negative effect, giving an overall negative view of the practice.

Filtration would be a relative inexpensive step forward for the segment of the Californian industry aiming at having premium quality EVOO. More importantly, it would be critical for the vast majority of Californian olive oils in terms of improving shelf life and stabilizing positive attributes, something that has not been discussed or even addressed much in our industry.

Filtration has its downsides – labor costs, loss of oil trapped in the filters, cost of equipment, time consumption – though it has proved so far to be the best method of reducing the moisture/water content below the IOC standard of 0.20% (it goes from 0.44%-0.22% to 0.07%-0.10%) while improving the organoleptic characteristics of EVOO.

During a recent visit to Italy, I met with a group of producers, researchers, and tasters, some of them authors of the study where most of the quotations in this article were sourced. When I mentioned the UC Davis report to them, their response was that nobody making premium EVOO in Italy questions filtration any longer; it’s an outdated discussion. The current discussion is about how to filter, improvements in filtration equipment, and so on.

Summary

  • Filtration is the more effective method known so far to reduce water content in olive oil below 0.20%, as recommended by IOC.
  • Not all filtration methods produce the same result.
  • Filtration must happen immediately after olive oil is made.
  • Oxygenation is a minor negative aspect of filtration, offset by the many positives it effects.
  • Suspended solids are to be removed, there are no significant positives associated with their presence in olive oil.
  • While is true that filtration causes a loss of water-soluble phenols, this is a minor loss, given that filtration is able to protect the remaining antioxidant fraction for longer.
  • Filtration does not remove any desirable aromas nor does it diminishes positive attributes. On the contrary, by removing non-olive oil matter (of negative progression), filtration contributes to positive sensory attributes and preserves them longer.
  • There’s no significant decrease of chlorophyll content, aside from the chlorophyll present in sediments and water, which is prone to rapid oxidation. Unfiltered olive oils are not greener by any means.
  • Filtration reduces moisture, extends shelf-life, impedes hydrolysis, and delays oxidation of lipids and phenols.

Besides its downsides (costs, time and labor expenses, loss of oil trapped in filters, and expense of equipment), filtering is such a positive practice and a key part in the production of quality olive oil, that it has been assumed by most premium brands in Europe and by other segments of the olive oil industry there.

I hope to bring about a friendly discussion on this subject. My experience with filtration has been so overwhelmingly positive that I naturally want to share it, with my best wishes for our growing industry.

~ Pablo Voitzuk

(1) Martina Fortini, Marzia Migliorini, Chiara Cherubini, Lorenzo Cecchi, Lorenzo Guerrini, Piernicola Masella and Alessandro Parenti. Shelf life and quality of olive oil filtered without vertical centrifugation. –
(2) Ciafardini, G., Zullo, B., Microbiological activity in stored olive oil.
Int. J Food Microb. 2002,75, 111–118.
(3) Ciafardini, G., Zullo, B., Iride, A., Lipase production by yeasts from extra virgin olive oil. Food Microb. 2006,23,
(4) Cecchi, L., Migliorini, M., Cherubini, C., Innocenti, M., Mulinacci, N., Whole lyophilized olives as sources of unexpectedly high amounts of secoiridoids: The case of three Tuscan cultivars. J. Agric. Food Chem. 2015, 63, 1175–1185.
(5) Brockerhoff, H., in: R. G. Jensen (Ed.), Lypolytic Enzymes,
Academic Press, New York 1976, pp. 25–125.
(6) Fregapane, G., Lavelli, V., Leon, S., Kapuralin, J., Desamparados Salvador, M., Effect of filtration on virgin olive oil stability during storage.
Eur. J. Lipid Sci. Technol. 2006, 108, 134–142.
(7) Koidis, A., Boskou, D., The contents of proteins and phospholipids in cloudy (veiled) virgin olive oils. Eur. J. Lipid. Sci. Technol. 2006, 108, 323–328
(8) Lozano Sanchez, J., Segura-Carretero, A., Fernandez -Gutierrez, A., Characterization of the phenolic compounds retained in different organic and inorganic ?lter aids used for ?ltration of extra virgin olive oil. Food Chem. 2011, 124, 1146–1150.
(9) Brenes, M., Garc?a, A., Garc?a, P., Garrido, A., Acid hydrolysis of secoiridoid aglycons during storage of virgin olive oil. J. Agric. Food Chem. 2001, 49, 5609–5614.
(10) Kalua, C. M., Allen, M. S., Bedgood, D. R., Bishop, A. G., et al., Olive oil volatile compounds, flavor development and quality: A critical review. Food Chem. 2007, 100, 273–286.
(11) Di Giacinto, L., Di Loreto, G., Di Natale, C., Gianni, G., et al., Caratterizzazione analitica degli attributi sensoriali degli oli vergini di oliva. Project report. Camera di Commercio di Firenze. 2010.
(12) Tamborrino, A., Clodoveo, M. L., Leone, A., Amirante, P. Paice, A. G., in: Preedy, V. R., Watson, R. R. (Ed.) Olives and Olive Oil in Health and Disease Prevention, Academic Press, San Diego 2010, pp. 77–83, ISBN 9780123744203.
(13) Angerosa, F., Servili, M., Selvaggini, R., Taticchi, A., et al., Volatile compounds in virgin olive oil: Occurrence and their relationship with quality. A 2004, 1054, 17–31.
(14) These are some premium EVOO from Europe, consistent award winners: Dievole, Altomena, Balduccio, De Carlo, Intini, Caputo, Torre Bianca, Olio della Madonina, Fattoria Ramerino, Laudemio, Franci, Fonte di Foiano, Castillo de Canena, Melgarejo, Marqués de Grignón, Crudo.
From California: Apollo, Katz, Isern & Sons, Saltonstall, Moonshadow, Pacific Sun.
(15) Capella, P., Fedeli, E., Bonaga, G., Lercker, G., Il Manuale degli Oli e dei Grassi, Tecniche Nuove, Milan (Italy) 1997.

Guerrini, L., Masella, P., Migliorini, M., Cherubini, C., & Parenti, A. (2015). Addition of a steel pre-filter to improve plate filter-press performance in olive oil filtration. Journal of Food Engineering, 157, 84-87.