January 28, 2005

Does filtration make vodka better?

A massive extracurricular science project in Hyde Park and around the country has intrepid college students asking the question, "Can my roommate's water filter improve cheap vodka?" Some feel that Brita plus Skol equal Grey Goose, while others think Dmitri plus Pur equal placebo effect. Though results vary and theoretical analysis is not quite conclusive, the phenomenon shows no signs of slowing.

Of the experimenters on campus, two undergraduates shared their experience with the Maroon. Russ Hittinger, a second-year in the College, involved his roommates in his pursuit of better vodka. After lots of testing, Hittinger and company reached positive conclusions. They have since made Brita filtering a regular part of their drinking.

A similar but separate team led by third-year Nick Bailey came to different conclusions: At-home vodka filtration is "an experiment I'll only ever repeat as a parlor trick, should I sink that low." According to a beverage science expert—Professor Karl Siebert, Cornell biochemistry researcher—more rigorous scientific consideration tacitly supports filter effectiveness.

Like so many other campus memes, vodka filtration's beginnings are mysterious. Hittinger heard about it in his hometown of Tulsa, from "20-something no-ones who like to drink."

Bailey first heard about it "through the ether of the Internet"—more specifically, the blog-like website (OMGIB). Filled almost entirely with vodka filtration content, the site provides a comprehensive "lab report" from an informal evening of "boozahology" that include non-blind taste tests and Saltine "science" crackers for palate cleansing.

Procedures vary slightly, but all sources recommend multiple passes through a pitcher-style water filter. Warm or cold, Hittinger and OMGIB agree with Bailey, who states that vodka filtration takes "a very long time, much longer than it takes to filter water"—about 15 to 20 minutes for the standard 1.75 liters. While performing a demonstration, Hittinger provided his own explanation of the slow drip: "That means it's working."

Once filtered, the difficult task of testing begins. Actual procedures varied significantly. Both Bailey's team and OMGIB sampled progressively, simultaneously comparing unfiltered, filtered, and control (expensive) vodkas. Bailey's team went a step farther, using a single-blind testing methodology to avoid bias.

Given that the federal government defines vodka as "neutral spirits…without distinctive character, aroma, taste or color," experimenters struggled to accurately convey their experiences.

After five passes, Bailey's testers "could very reliably tell the difference between the different samples." The OMGIB team noted reduction of "offensive taste and odor" starting on the first pass and continuing through the fifth. One OMGIB tester felt that the inexpensive but multiply filtered "Vladimir Vodka" surpassed the control vodka—a pricey bottle of Ketel One.

Hittinger reported that after four passes, his $13 Popov tasted like vodka for which he'd expect to pay twice as much. When asked to describe results in more detail, Hittinger and other team members, who chose to remain anonymous, opted for "smoother," "less throat burning," and "more drinkable."

Were these results valid? One Hittinger team member suspected the placebo effect, and all testers pointed out the uncontrolled variable of intoxication.

When polled by e-mail for a more rigorous scientific perspective, professor Laurie Butler—the only respondent amongst seventeen University Chemistry faculty members—stated that "students can filter their vodka however they'd like."

A less subjective perspective was provided by Siebert, who currents the Cornell undergraduate course, "Understanding Wine and Beer." In addition to working at the Stroh Brewery Company for 18 years, Siebert has in his travels toured a Stolichnaya factory outside of Moscow.

From Siebert's scientific perspective, vodka is "as close to pure ethanol as you'll get," albeit diluted with water. Extensive distillation and filtration creates this purity, and variations in these two processes account for nearly all differences between vodkas. From an analytic chemistry perspective, even a single rough distillation creates an alcoholic spirit that is "99 percent pure." Unfortunately, these trace impurities "can be strongly flavored," in a way that sensation scientists don't completely understand.

To "clean up" these remaining impurities, producers pass vodka through filters filled with active carbon—the same active carbon which fills consumer water filters. Could the Brita filter pick out impurities missed by industrial carbon filtering processes? "Good vodkas," according to Siebert, "are normally passed through ‘mountains of charcoal,'" thereby reducing the possibility of consumer filter effectiveness. For cheap vodkas, though, Siebert speculates that corners are cut on both distillation and filtration, creating room for a Brita intervention.

Does this analysis agree with on-campus investigations?

Bailey's team, though noting significant differences, decided against further trials or permanent implementation. Bailey, an economics concentrator, summarized his personal dissatisfaction with an alchemical allusion: "Rather than lead into gold, we spun lead into slightly less filthy and offensive lead."

After the testing phase, Hittinger's team was so impressed they moved to permanent implementation. Surplus filters allowed for a two-filter switching system that quickly solved what one roommate termed the "milky water or smooth vodka" tradeoff.

Much as these experiments have shown that the vodka world can be finicky, The New York Times Dining and Wine section published a taste test of high-end domestic and imported vodkas in a late breaking development. Smirnoff beat out other brands for the top spot. Grey Goose did not even make the top 10.

All experimenters agree that filtration makes a perceptible difference, yet some found the exercise essentially pointless. Without a rigorous diagnostic analysis, one recourse seems to remain—more research.