The New Science of Old Whiskey

Bourbon makers attempt to perfect a 2,000-year-old invention.

Buffalo Trace Distillery

In April 2006, a tornado struck Warehouse C at the Buffalo Trace Distillery in Frankfort, Kentucky. In the aftermath, the building looked like a diorama: part of the roof and one wall had been artfully removed to reveal the 25,000 barrels stacked inside. Miraculously, not a single one of those barrels was damaged—proof, perhaps, of the Major League manager Leo Durocher’s maxim: God watches over drunks and third basemen.

Repairing the warehouse took several months, and during that time the barrels on the upper floors were exposed to rain, heat, and sun. Mark Brown, Buffalo Trace’s president and CEO, joked at the time that the distillery should sell the whiskey as “tornado-surviving bourbon.”

It turned out to be no joke. The barrels were opened about five years later (the liquor inside had then aged for nine to 11 years) and, says Brown, “the darnedest thing is, when we went to taste the whiskey, it was really good. I mean really good.” The company decided to label the bourbon “tornado surviving,” and aficionados—who also found it superior to the usual product—quickly snapped it up. One went so far as to write Buffalo Trace and ask whether it planned to make more. “Not deliberately,” Brown replied.

Yet the tornado bourbon got the distillers wondering: What are the perfect conditions for storing the barrels in which bourbon ages? It’s a question that no one had really asked before, despite the oft-noticed phenomenon that barrels situated near the windows in warehouses have a tendency to become what managers call “honey barrels”—that is, ones that produce above-average whiskey. Moreover, the storage question was a logical follow-up to one that Buffalo Trace had already been pondering: How do you make a perfect barrel?

Both questions are critical for any maker of bourbon. It’s well known that a good portion of a whiskey’s flavor comes from the barrel it ages in, but less well known is just how large a portion we’re talking about. I recently asked a dozen or so people involved in the bourbon industry how much of the flavor comes from the barrel, and how much comes from other elements, such as the grains used or the distillation method. Most said that 60 or 70 percent of the flavor comes from the barrel, and one went as high as 80 percent. No one I spoke with estimated the proportion at less than 50 percent—meaning one of the trendiest liquors on the contemporary cocktail scene owes most of its flavor to a technology that’s thousands of years old.

Liquor barrels are essentially Dickensian nano-factories—dark, sooty, mysterious places from which marvelous things emerge. But they came to that role only after long service as simple containers for shipping and storage.

“The barrel, like the wheel, is one of the outstanding basic inventions of mankind,” wrote the historian William B. Sprague in a 1938 essay. Wooden-stave barrels first appeared millennia ago. In the first century A.D., Pliny the Elder noted their widespread use in the foothills of the Alps.

Used for centuries to transport everything from whale oil to pickles to nails, the barrel is far more ingenious than it appears at first glance. It bows out slightly in the middle (called the bilge), so each individual stave forms a longitudinal arch linking a barrel’s top and bottom. Each stave also abuts two adjoining staves, so they form, collectively, another arch, this one latitudinal around the circumference of the barrel. Barrels are thus remarkably stout and durable: if one tumbled off a ship’s gangway onto a wharf during loading, the impact of the fall was shared by all the staves, reducing the risk of breakage.

Barrel design also evolved such that a single stevedore could move one weighing hundreds of pounds. When standing upright, a barrel can be tilted and rolled on its edge; on its side, it’s even more maneuverable—only a tiny portion of the barrel touches the ground, so it can be spun in any direction, and a light push will start it moving. A skilled worker can bring it upright by rocking it a few times and popping it into position.

It took a long while, but in the past few hundred years, discerning drinkers of spirits noticed that something interesting took place inside containers made of white oak, which were often used for liquids because oak is tightly grained, preventing leakage. Rum shipped from the West Indies to Boston and whiskey shipped from Kentucky to New Orleans were always better when they arrived than when they left.

Today, barrels are central to the making of all manner of alcoholic beverages. Most winemakers use oak barrels that are toasted—that is, lightly browned on the inside through indirect heat. Liquor manufacturers more often use charred oak barrels, which are blackened with direct flame. Every spirit has its own barrel culture. Cognac tends to favor French oak (which is higher in tannins than American oak). A few rum makers, such as Zacapa Rum, from Guatemala, employ used sherry casks for finishing their product after it’s been aged in oak, adding flirtatious complications to the end result. Some American craft distillers, seeking to distinguish themselves from the major producers, have been experimenting with barrels made of alternative woods, including hickory and maple. Even some microbrewers are getting into the game, finishing beer in used bourbon casks.

But bourbon has long been the main driver of the spirits-barrel industry, for one simple reason: to be labeled “bourbon,” a whiskey is required by law to have been aged in new, charred oak barrels, which impart far more flavor than barrels that have been used previously. (After aging bourbon, barrels are typically sold to makers of rum, scotch, or tequila for reuse.) Bourbon producers thus tend to focus more intently on barrels than other spirit makers do. Brown-Forman, which makes Woodford Reserve, Old Forester, and Jack Daniels, has operated its own cooperage since 1945 to ensure a steady stream of quality barrels. “Why would we entrust all our color and 50 to 60 percent of our flavor to an outside vendor who’s also supplying our competitor?” asks Chris Morris, the master distiller of Woodford Reserve.

This barrel flavoring takes place in part because alcohol is a solvent that gradually breaks down elements in the wood over time. White oak in particular has an abundance of appealing flavors, including vanilla, nuts, and coconut, as well as butterscotch notes from sugars in the wood, which are caramelized during charring.

During the summer, heat increases pressure inside a barrel, and some liquor pushes itself through the char in the barrel’s wooden pores, enabling the carbon to filter out impurities. During the winter, the liquor moves in the reverse direction. The process is repeated with less vigor during the heating and cooling cycles of day and night. So merely by sitting in a barrel, whiskey is slowly being filtered through the barrel.

One bourbon maker lashed five barrels on a trawler, where they were subjected to salt air, sun, and constant rocking.

The rise of premium liquors in the 1980s, and a growing consumer willingness to pay more for complex aged spirits, brought new attention to the science of barrel aging. Three well-regarded players in the spirits industry—Brown-Forman, Independent Stave (one of the world’s largest barrel companies), and Buffalo Trace (which makes the super-premium aged bourbon Pappy Van Winkle, as well as Blanton’s, W. L. Weller, and others)—each launched programs to minutely analyze the chemistry of barrels, using new technologies that let them parse ingredients on the parts-per-billion level.

“The thought is to identify every variable, which we may never do,” says Harlen Wheatley, Buffalo Trace’s master distiller. “But we’re on the trail.” He says the company’s chemists have found about 300 different compounds, and have identified some 200 of them. Once all elements are ascertained and understood, some fear it may be a short step to producing synthetic aged bourbon. “You could go to the lab and say, ‘I’m going to take one part per billion of this and one part per billion of this and three parts per billion of that, I’m going to put it all in a test tube, and I’m going to make the perfect bourbon based on our research,’ ” says Brown. He hastens to add, “I don’t think we’ll ever be looking to make a synthetic bourbon.”

Rather, Buffalo Trace clearly revels in tweaking its ancient technology. The company’s most intriguing foray into barrel science has been the Single Oak Project. It was conceived, according to Brown, when a pair of curmudgeonly warehouse managers, Leonard Riddle and Ronnie Eddins, started needling each other during a meeting, with taunts of “my whiskey’s better than your whiskey.” Brown told them, essentially, to put up or shut up by aging a barrel of the same spirit and then comparing the resulting whiskeys.

But as they discussed the rules of engagement, Riddle and Eddins realized that they faced a more vexing problem: finding two identical barrels. The three dozen or so staves in any given barrel likely come from different trees (possibly even from different states), and by dumb luck one of the men might end up with a honey barrel. So Riddle and Eddins started compiling all the variables that might affect a barrel, and soon had listed dozens, ranging from the age and size of the oak used, to the latitude and elevation where it grew.

To test all their variables, they calculated, they would need to make about 3,000 barrels. So they narrowed down the list to what they believed was a more manageable number of salient qualities, including growth rings per inch, soil quality, and whether the oak grew on a north- or south-facing hillside. Then they set off into the Ozarks to select trees. They returned with 96 white-oak logs, which were then cut in half to test another persistent myth: that a barrel produced from the upper half of an oak tree makes better bourbon than one produced from the bottom half. The logs were cut and dried and assembled as 192 barrels, each representing a different set of variables. The barrels were filled with freshly distilled whiskey, and left to age for about eight years.

The bourbon was bottled (which stops the aging process), and for the past two and a half years Buffalo Trace has rolled out an experimental batch about every three months. Consumers are tipped off about releases via the company’s Web site and are asked to give feedback; at least one informal tasting club has sprung up for this purpose. (Buffalo Trace is also analyzing the bourbon in-house.) The 11th release is due this month, and the remaining single-oak bourbons will be out within the next two years. “We’ve already seen reviews where people have said wheat [bourbon] with the top half of the tree is really good, and rye [bourbon] with the bottom half of the tree is really good,” Brown says. “We’re going to have to think about that.”

Once all the bourbon has been released and the feedback is in, the company will crunch the data and figure out which oak—and which part of the oak—makes the best bourbon. Then the real challenge begins. Buffalo Trace plans to launch a Single Oak brand, and has said it will produce the bourbon using barrels made from the same type of wood that prevailed in the test project. “If you drive through a forest in Missouri and the top half of a bunch of trees are missing,” Brown told me, “you’ll know exactly what happened.”

Apart from the question of how to make the best barrel, there’s a lot of experimentation taking place that involves ways of managing a barrel’s interaction with its contents. If you visit a craft distillery, for instance, one of the first things you may notice is that the barrels are very small—in some cases as tiny as five gallons. (The standard spirits barrel is 53 gallons.) The thinking is that increasing the ratio of wood surface to liquid will accelerate barrel aging.

Cutting down aging time is particularly important for start-ups trying to get their spirits to market quickly. After all, if you were to hand a potential investor a business plan that called for spending hundreds of thousands of dollars to produce something that would sit in a dark room for several years before it earned its first dime, you’d probably find yourself faced with a long and unsettling silence. (This is also why many start-up distillers launch with gin, vodka, or white whiskey, which can move from still to liquor store without delay.)

But while the increased wood-to-liquor ratio speeds the infusion of the wood flavors, that’s only part of the aging process. The other part occurs during oxidation, when air gets in through the barrel’s semi-porous wood and interacts with the spirit. Corn (bourbon is by law at least 51 percent corn) has long, complex, oily molecules, which give young bourbon its round and vegetal funk. During a chemical reaction called esterification, these chains break down and reconfigure, then react anew with the wood, adding depth and additional flavors. This oxidation process can’t be rushed with smaller barrels.

Many craft distillers have been simultaneously aging liquor in barrels both small and large, and as time passes and stock from standard-size barrels makes it to market, smaller barrels may naturally fade away. Yet some of the better craft distillers are likely to stick with small barrels—they’ve become a tool that allows the distillers to fine-tune distinctive taste profiles.

“What I see happening is people saying, ‘Well, let’s not give up on small barrels,’ ” says Clay Risen, who writes about spirits for The New York Times and is the author of the forthcoming guide American Whiskey, Bourbon & Rye. “If you think of a clear, unaged distillate as a rough product, a first draft, barrels can be a sort of editing process—they can take the edge off, remove certain flavors, impart certain flavors over a relatively short period of time.”

The federal regulations that define bourbon don’t leave much room for monkeying with the construction of barrels themselves. But they do leave some latitude for experimentation. Black Swan, a four-year-old artisanal cooperage in Minnesota, is best known for its Honey Comb Barrels, which are, as the name suggests, honeycombed with shallow perforations on the inside. As with smaller barrels, the idea is to increase the surface area in contact with the whiskey, to speed up the aging process.

Vibrations, movement, and pressure can also accelerate the interaction between liquor and wood. Cleveland Whiskey, a new brand, has a proprietary system involving “stainless-steel, pressure-capable tanks” that its inventors claim can reduce the aging process from years to as little as six months. (The verdict’s still out.) Meanwhile, Tuthilltown Spirits, in the Hudson Valley, takes a more minimalist approach, regularly playing loud music in its warehouses in an effort to shake its whiskey up.

And then there’s Jefferson’s Ocean. A few years ago Trey Zoeller, the founder and master blender of Jefferson’s Bourbon, took five barrels of whiskey and lashed them above deck on a Russian trawler that had been converted into a shark-research vessel. They were subjected to salt air, heat, sun, and constant rocking for three and a half years. Such an approach may actually have had historic precedent: 19th-century rum from Medford, Massachusetts, was famed for its quality, with one writer explaining, “It never left the bonded warehouse on Riverside Avenue until it had passed a severe test and was shipped across the Atlantic and back again, in wood, to age it.”

Zoeller’s bourbon roamed widely, from the Panama Canal (six transits) through the Pacific and up to Puget Sound. The steel hoops on two barrels corroded and the staves fell apart. (“At least that’s what I was told,” Zoeller says, noting that the crew also subscribed to a vague concept it called “Juan’s share.”) When the barrels were finally retrieved, what came out was nearly black from constant sloshing in charred wood. But once the char was filtered out, a remarkably smooth and supple bourbon emerged, with a slight briny tang reminiscent of scotch.

Zoeller is currently repeating the ocean-aging experiment: last spring he loaded 72 barrels on a freighter in a defensible container. (“It’s locked up, and we’ve got a Webcam,” Zoeller says.) Most of the bourbon had already spent between four and eight years in the barrels, so this is more of a finishing run. Zoeller plans to meet up with the ship in various ports and sample the bourbon throughout the fall. Bottles could start appearing on the market soon after.

Most distillers I spoke with estimated that 60 or 70 percent of a bourbon’s flavor comes from the barrel.

Which brings us back to the question of how best to store barrels. It’s long been known, for instance, that climate affects aging time. Rum and tequila are typically aged in the steamier West Indies or Mexico, which results in shorter aging times. Dave Pickerell, a former master distiller at Maker’s Mark, told me about an experiment that took place some years ago when a barrel of bourbon was shipped to Scotland to age, and a barrel of scotch was shipped to Kentucky. The verdict? One year of aging in warmer Kentucky was the equivalent of about three or four years of aging in cooler Scotland.

Last summer, inspired in part by the experience of its unexpectedly tasty tornado-surviving bourbon, Buffalo Trace began a new experimental project to better understand the aging environment. It’s building a barrel-aging laboratory called Warehouse X, which will be clad in stone and feature four south-facing interior chambers, each with a sunroof regulated by louvers. In addition to light, three other elements will be independently regulated in each chamber: airflow, temperature, and humidity. There’s also a fifth storage area, a breezeway, that serves as a sort of control: it’s covered but otherwise exposed to the elements, and thus more or less simulates the conditions that produced the tornado whiskey. “No one has built the perfect aging warehouse,” Mark Brown muses. “Is it a greenhouse? Or do you build something that has no windows and is completely controlled? Or do you build something in between?”

Warehouse X will hold about 100 barrels—a tiny fraction of the 25,000 in Warehouse C when it was hit by the tornado—and this limited-edition bourbon will be tasted and tested, then bottled and sold after an aging period of six years or more. Brown adds, “But only if it’s any good.”

He doesn’t hazard any guesses about which of the barrels in the Warehouse X experiment will turn out best, but I get the sense that he’s rooting for the breezeway, which is essentially at the mercy of the elements. “It would be funny,” he says, “if all any idiot ever needed to do is have a field with an armed guard, and just plant the barrels out there. Maybe that’s what Buffalo Trace will do—buy 5,000 acres and have a barrel farm.”