Smoke Taint Revisited: What Progress Has Been Made?

The following article is a discussion between The Academic Wino and leaders in the wine treatment business in the United States.  Interviews were conducted in November 2011, so any new research since then will not be included in the discussion.

*Edit:  For an update on the most recent research that was published after these interviews were conducted, please read the first comment after this article from Bob Kreisher, who explains in excellent detail the current state of this research within the last month or so.

Since the Mendocino, California wildfires of 2008, researchers have been attempting to find ways to mitigate the negative sensory characteristics of which smoke taint in wine is associated.  Even before these fires, Australian vineyards have been coping with the same problem, as wildfires in that region are even more frequent than they are in California or other places in the United States.

What is Smoke Taint?

Smoke taint is a condition created when grapevines are exposed to large quantities of smoke at a key point in the maturation of the grapes.  Smoke exposure from veraison to harvest results in noticeable smoke taint properties such as “smoky” and “ash” in the final wine.

Smoke volatiles are absorbed by the leaves of the vines and are bound to sugars within the plant.  These glycolated volatiles are moved throughout the plant via the xylem, ultimately accumulating in the berries.  According to Eric Herve from ETS Laboratories (locations in CA, OR, and WA), while in these conjugated forms, smoke taint volatiles are not easily detectable, and could result in false negatives.  In other words, juice samples will test negative for guaiacol and 4-methylguaiacol, currently the only known indicators of smoke taint in wine, when they are bound by sugars.  However, during fermentation, guaiacol and 4-methylguaiacol are released from these conjugated forms and are detectable in the finished wine.

Herve has since established an assay by which the free fraction of guaiacol is measured in whole berry samples instead of the juice, which tends to have much lower levels of guaiacol compared to the whole berry fraction, resulting in a higher likelihood of identifying smoke tainted samples.

What is complicated about smoke taint?

Guaiacol and 4-methylguaiacol are currently the only two compounds in smoke taint that have been identified.  What complicates the issue, however, is that there are actually many compounds that contribute to smoke taint in wines, and no others have been identified.

According to Eric Dalhberg, president of Winesecrets (locations in CA, OR, TX, VA, British Columbia and Ontario), only measuring guaiacol and 4-methylguaiacol to identify smoke tainted wines is problematic.  Some wines can have increased levels of guaiacol with no flavor impact, while other wines that are low in guaiacol can have a negative flavor impact.  Simply measuring the levels of guaiacol and 4-methylguaiacol in wines doesn’t necessarily mean that it is suffering from smoke taint, and that sensory analysis appears to be the only way to determine if smoke taint is present.  Sensory analysis of smoke tainted wines could become problematic if palette fatigue set in and created a situation where a wine is scored negative for smoke taint when in fact it was there.

Bob Kreisher, president of Mavrik North America (MNA, Santa Rosa, CA), stated that there are many compounds involved that produce a wine with the characteristic flavor of smoke taint.

Kreisher believes there is likely a synergistic effect occurring between all of the volatile compounds that contribute to smoke taint.  For example, Kreisher noted that if you take two wines below the sensory threshold for smoke taint and combine them, you might create a wine that is above the sensory threshold.  Though the identity of these compounds is unknown, the team at MNA has successfully identified the chemical characteristics of the volatiles and has applied this knowledge to their amelioration techniques.

The Australian Study

Smoke taint research is not novel, and there have been many papers published on the subject.  Recently, a team of researchers in Australia published an article claiming that they are the first to demonstrate the amelioration of smoke taint from wines using the techniques of reverse phase osmosis and solid phase adsorption to selectively remove guaiacol and 4-methylguaiacol.

The authors used a proprietary spiral wound nanofiltration membrane with a molecular weight cut-off of 150-200amu and a polystyrene based adsorbent resin.  Smoke tainted wine was pumped through under high pressure to attain flow rates of 50mL/min and 600mL/min.  Samples were collected at various times throughout the experiment, and once the process was complete, wine was bottled and stored at 15oC until sensory analysis could be performed.

However, after revisiting the ameliorated wines 6, 12, and 30 months after the treatment, the authors noted that the smoke taint characteristics gradually returned over time.

What has been done in the US?

Representatives at both Winesecrets and MNA claimed that none of the wines they treated from the 2008 wildfires suffered recurrence of smoke taint, unlike the results of the Australian study.  At Winesecrets, activated carbon is used in the solid phase adsorption step, instead of the resin that used in the Australian study.  According to Dahlberg, using activated carbon is much more effective than using resin for the removal of smoke taint from wine.

At MNA, instead of only focusing on guaiacol and 4-methylguaiacol, they targeted a range of volatile compounds by creating membranes that selectively remove compounds based on several variables.  One problem with the Australian study, according to Kreisher, is that they used membranes that only selected compounds based on molecular weight.

The problem is that there are many ways in which a compound can pass through a membrane, which are not solely dependent upon molecular weights.  By removing a range of volatile compounds, and by manipulating other variables such as electron attraction and flow rate in the solid phase adsorption step, MNA is successfully able to remove the smoke taint from the wines, without dramatically reducing the desired volatiles, and none have returned with a recurrence of smoke taint.

What caused the recurrence of smoke taint in the Australian study?

The Australian researchers claimed there was a recurrence of smoke taint as a result of 4-ethylphenol and 4-ethylguaiacol being released into the wine over time.  The problem with this claim, Kreisher says, relates to how glycobound phenols are released.  Simply put, guaiacol, 4-ethyguaiacol, and other phenols, are released from their glycolated forms during the fermentation process.  Once fermentation is complete, there is not enough sugar to release guaiacol and 4-ethylguaiacol at the levels found in the Australian study.

Kreisher believes the Australians found a recurrence of smoke taint in their wines due to the devolution of other volatile phenols into guaiacol and 4-methylguaiacol.  If they had targeted the entire range of volatile compounds to begin with, he says, they would not have found a recurrence of smoke taint in their wines.

Another mechanism for this recurrence, Kreisher elaborated, is the possible synergistic effect of the smoke taint volatile phenols that were not removed and a Brettanomyces infection.  There is a possibility that a treated wine with smoke taint volatile levels below threshold that becomes infected with Brettanomyces, an infection by which 4-ethylphenol and 4-ethylguaiacol are increased; could have a synergistic effect that results in the wine showing smoke taint above sensory threshold levels.

Concluding Remarks

Even though the Australian study is the first study demonstrating the amelioration of smoke taint from wines using the aforementioned methods, work has been occurring simultaneously in the US.  Companies in the United States appear to have established smoke taint amelioration methods that are similar to those used in the Australian study, but with key differences that eliminate the problem of smoke taint recurrence in wine.

An independent review of all known methods may be the next step toward finding the most effective smoke taint amelioration technique.

The Academic Wino would like to thank Eric Herve, Eric Dahlberg, and Bob Kreisher for taking the time to speak with us and provide us with important information for this post.

2 comments for “Smoke Taint Revisited: What Progress Has Been Made?

  1. January 10, 2012 at 3:28 pm

    I want to elaborate on the "volatile compounds". I would be more specific that it is "volatile phenols". This group of compounds is well known when it comes from microbial sources like brettanomyces and dekkera yeast. It was true when Rebecca and I spoke that, beyond guaiacol and 4-methylguaiacol these had not yet been identified. I have since heard that the State of Victoria (Aus) Department of Primary Industries has published a study identifying at least some of these volatile phenols, but I have not yet seen that article.

    In any event, it is a finite class of compounds with well-known characteristics. As Rebecca points out, these compounds work synergistically. Guaiacol and 4-methylguaiacol themselves don't seem to be particularly active in the sensory expression of smoke taint. In fact, they evolve from barrel aging too. And it is important to note, if you have not had the displeasure of tasting smoke-tainted wine, that smoke taint does not taste anything remotely like "really oaky wine". Some occasionally find very low levels "interesting", but anywhere over threshold levels is just plain offensive. There is also a tactile component on the back of the palate that is not unlike walking into a room which had, several hours prior, been filled with smokers and then taking a deep breath. You feel it on your tongue, you don't just taste it.

    However, guaiacol and 4-methylguaiacol ARE very useful markers which we use to accurately predict how much processing is required to remove the smoke taint characters. When testing grapes or juice, where mostly glycolated precursors can't always be perceived sensorily, a positive hit on g/4mg is considered a positive for smoke taint. Likewise, in young wine–if it has not touched oak of any kind, g/4mg presence is a positive test. If it has touched oak, we can calculate, based on volumes, time, type/form of oak, etc, where the g/4mg should be due to oak and can make useful inferences based on this.

    It is important to understand that, for whatever reason, sensory characteristics don't fully express themselves until some weeks post-malolactic fermentation. Any wine that may have been exposed should be assessed at this time, even if it was not previously expressing smoke taint characters.

    As I mentioned above, g/4mg, and other desirable volatile phenols (furfural, eugenol, etc.) are associated with oak aging. So we took the challenge of removing the undesirable volatile phenols while leaving behind the desirable ones, mostly untouched. We noticed that there were differences between the desirable volatile phenols and the rest of all known volatile phenols. We exploited that difference to develop a proprietary adsorbent which left behind most of the desirables. Commercial activated carbon was too indiscriminate–we found it to remove too much desirables. On the other hand, we found polystyrene resin to potentially remove too many other unrelated desirable elements, while actually removing too little of the offending volatile phenols.

    As for the Australian study Rebecca cited, in general the best research on the subject has come out of Australia's fine academic and independent researchers. Although I seriously doubt the researchers speculation about glucose-bound volatile phenols slowly being released (there just isn't enough residual glucose to account for this), this is just an end note speculation, not the actual product of the scientific research. Otherwise, the study itself is flawless as long as it is taken in context. It is a study of one Australian service's way of removing smoke taint. As Rebecca perceptively points out, at least some techniques in this country vary greatly from the technique studied in the article.

    The "return of smoke taint" phenomenon was already well-known in Australia, at least anecdotally. This is why we sought out a way to remove the higher order volatile phenols that actually caused the smoke taint character while leaving unmolested the desirable oak-related volatile phenols which winemakers paid good money to introduce to the wine in the first place.

    Finally, I want to address the claim that Rebecca attributed to me: that none of our treated wines have experienced a return of smoke taint. This is not entirely accurate. I know of two cases where we treated wine before malolactic fermentation and those wines expressed some lesser degree of smoke character after ml. I know of two cases where blending of treated wines resulted in a subtle new expression of smoke taint. And I know of one case where the test result numbers were reported incorrectly. In this case the wine had higher g/4mg results than were actually reported. So they were underprocessed initially (which we later corrected). Finally, one winemaker asked us to stop at only 60% of our forecasted treatment (our model is highly accurate and correlates well with sensory evaluation by customers). He said he could no longer detect the smoke taint. We could still detect it and furthermore we believe in our model. We advised him to continue with the planned treatment, but he declined. Later, he reported that he could once again detect the smoke taint.

  2. January 10, 2012 at 3:47 pm

    Thank you, Bob, for these great comments and elaborations on statements in this post, and also what has changed within the past month or so regarding the state of smoke taint research. I am going to keep my eyes peeled for that new article from the State of Victoria Department of Primary Industries, and hopefully write a blog post presenting those data and results.

    Also, thank you for clearing up the mistake I made regarding the return of smoke taint in some of the wines your company has treated.

    I'm hoping your insightful comments will spur some discussion here among other readers of this blog. Thank you again!

Comments are closed.