Why It Works
- Using an immersion circulator allows you to precisely heat the ingredients at the optimal temperature for coaxing maximum flavor from the vanilla beans.
- An alcoholic base means your extract will last indefinitely.
- With just minutes of hands-on time, you’ll have a complexly layered, industrial-quality extract—enough for 50 to 150 recipes—at less than a third of the cost.
The enchanting fruit of the vanilla orchid first cast its spell on the Maya, Totonac, and Aztec peoples of its native Mexico at least several hundred years ago. Vanilla complemented these Mesoamericans’ chocolate drinks in pre-Columbian times, and the “beans” eventually made their way to Europe on Spanish ships in the eighteenth century. Vanilla hasn’t outrun its tight association with chocolate, but it does often venture out on its own. In the US, vanilla’s flavor (in its various forms, which we’ll get into shortly) is mainly sought for ice cream―a lot of ice cream―and other sweet treats, with a tiny share ending up in savory dishes.
Most American recipes that include vanilla require it in alcohol-based extract form; we’re used to reaching for those tiny bottles of dark brown spirituous liquid. But as effective as the default store-bought extract may be, I’ve long wondered if it’s the absolute best carrier of concentrated vanilla flavor. In Europe, I’ve seen (or, at least, used to see) a lot more packets of vanilla sugar on grocery shelves than bottles of vanilla extract, so I know that’s one alternative. And as cooks we often hear that fat carries flavor—so why not vanilla butter? I decided to pit store-bought alcohol-based vanilla extract against three homemade flavorings―my own alcohol-based extract, vanilla butter, and vanilla sugar―to determine which delivers the most robust vanilla flavor.
The Science of Vanilla Flavor
Those slender, brown-black, wrinkled vanilla beans filled with miniscule sticky black seeds are actually cured, semi-dried fruits. They started life as green pods resembling overgrown green beans and sprouting along one of three species of orchid vines in the Vanilla genus. The barely ripe pods are shocked with high heat from hot water, steam, or the sun. The goal of this step is to cut short the ripening process, during which sugars and amino acids necessary for curing and flavor development would naturally be exhausted. According to Shawn Gavin, founder of the spice company Slofoodgroup, hot water (over 140°F) is by far the most common method because it is inexpensive; steaming is effective but requires costly equipment; and the sun is, of course, the least controllable and therefore the least relied upon. After this blast of heat that effectively kills the pods and kick-starts the curing process, they’re submitted to intermittent sun exposure and storage over the course of weeks or months. By the end of all this, the volatile molecules that define vanilla’s distinctive aroma and flavor have been freed within the skin, the seeds, and their molasses-like fluid, and cell damage has changed their color from green to dark brown.
The flavor and aroma of vanilla are the concerted effect of some two hundred volatile molecules (a note to my fellow non-chemists out there: molecules called “volatile” aren’t wild and unpredictable; they just readily evaporate under normal conditions). Our brain interprets the molecules to which our nose’s smell receptors and our tongue’s taste buds are exposed. The primary compound that our brain associates with vanilla is vanillin. In addition to vanillin, cured vanilla beans can also emit compounds with floral, anise, smoky, tarry, creamy, and nutty notes. The exact flavor depends partly on the species of vanilla bean, of which there are three that are widely cultivated: common (Vanilla planifolia; labeled Mexican, Bourbon, Madagascar, and Indonesian), Tahitian (Vanilla x tahitensis), and Pompona (Vanilla pompona—a rare Central and South American species). Tahitian vanilla beans have a much more delicate character—more floral and nutty than tarry and smoky—than the other species. (Artificial vanilla flavoring, on the other hand, is usually one-note, made with vanillin synthesized in labs from guaiacol, found in a wood oil called creosote, or from lignin, a by-product of the paper-making industry.)
The quality of vanilla flavor in our foods is, as with any flavor, highly affected by the medium within which the flavor molecules reside. Harold McGee points out in On Food and Cooking:
"Aroma chemicals are generally more soluble in oils, fats, and alcohol than in water….Oils and fats dissolve more aroma molecules than water during cooking, but also hang on to them during eating, so that their flavor appears more gradually and persists longer. Alcohol also extracts aromas more efficiently [than water], but because it too is volatile, it releases them relatively quickly."
Stella Parks has written that high-quality store-bought vanilla extracts deliver far more intense and complex vanilla flavor than whatever most homemade vanilla extract recipes can. This is true. Most homemade vanilla extract recipes have you steep the pods in alcohol for a long time at room temperature, and for all the investment of time and ingredient expenses, what comes out the other end is a mere shadow of what good bottled products offer. Commercial producers know what they’re doing and use experience, research, and equipment we don’t have at home to extract maximal flavor molecules from vanilla beans. But let's take a closer look at what exactly makes some homemade alcohol-based vanilla extracts—and other vanilla flavorings, for that matter—weaker than store-bought extract.
- Concentration: The FDA requires a minimum 1 part vanilla beans for every 10 parts ethyl mixture (ethyl alcohol, at least 35%, mixed with water) by mass. That’s a lot of vanilla beans―25 beans for each liter of liquid―and judging from many recipes available online, most homemade “extracts” include a lower proportion of beans.
- Temperature: As Stella points out, most homemade “extracts” are in reality imprecisely flavored vodka. Extraction requires precise heating: too low of a temperature and the flavor volatiles won’t be pulled out from the vanilla beans; too high of a temperature and some of the volatiles will be undesirably altered or destroyed.
- Extraction Medium: The volatile molecules that define flavors have different solubility characteristics. Most are more soluble in alcohol than in water, while some are more soluble still in fats or oils. And when it comes to infusing granulated sugar with vanilla’s flavor, the marriage seems to be a more superficial one. “Over time,” writes Kye Ameden at King Arthur Baking Company, “the oils and small particles from the ingredient impart aroma and flavor into each spoonful of sugar.”
- Vanilla Bean Quality: You want to make sure that your vanilla beans come from a trustworthy source—one that has cured and stored the fruits under optimal conditions—and haven’t been sitting around long enough to lose their potency. As soon as you open a package of vanilla beans, their signature heady aroma should waft its way to your nose. Grade A beans should be plump and supple with a somewhat glossy finish. Grade B beans are perfect for extract; they are a tad drier than their Grade A counterparts and may have superficial imperfections, but they are less expensive and contain all of the flavor you need. As for the species and origin of vanilla you choose, that’s a matter of taste. Vanilla planifolia is the most vanillin-forward, while V. tahitensis is subtler and more floral in character. If you’re lucky enough to come across the rare V. pompona (I’ve only read of it), you may find truth in descriptions of it as complex, spicy, sweet, and fruity.
Taking into account all of the above, I hypothesized that fat or oil would be a superior flavor extraction medium for vanilla, with more flavor stability and intensity than in an alcohol-based extract. If you’ve ever cooked an Indian dish, you’ve likely made a tadka. That’s the technique (and product) of blooming spices in hot oil or fat; it wakes up the spices’ flavors and diffuses them through the fatty medium in which they’ve been heated. It’s taken as a given in Indian cooking that heat and fat are essential to get the most out of spices.
With some tailoring in terms of temperature and time to accommodate vanilla’s delicate volatile molecules, I can’t see why the concept behind tadka wouldn’t apply to vanilla extraction. Extracting vanilla flavor using fat or oil might produce a more intense flavor than what we find in our usual alcohol-based extract.
But while I suspected fat was going to be the best medium for vanilla flavor extraction, I also wanted to test other techniques, including precision-heated extraction (using an immersion circulator a.k.a. a sous-vide device) in alcohol and, separately, granulated sugar, to see if there was a worthy alternative to store-bought vanilla extract.
Homemade Vanilla Flavorings: Common Techniques
I made each of my three homemade vanilla flavorings—vanilla extract, vanilla butter, and vanilla sugar—under conditions that took into account my research into commercial vanilla extraction methods and studies of optimal vanillin extraction. I wanted all of my flavorings to be made using the same single-origin species of vanilla bean to ensure a level playing field in terms of flavor profile. I went with Tahitian vanilla simply because Nielsen-Massey makes a Tahitian vanilla extract recommended by Stella Parks, and my trusted vanilla bean source, Native Vanilla, sells extract-grade Tahitian beans.
For the vanilla butter, I went with a European butter, as it has a higher fat content than conventional American butter (about 85% versus 79%). For the homemade alcohol-based vanilla extract, I chose Tito’s vodka. And for the vanilla sugar, I used organic cane crystals, to keep it in line with the quality and flavorfulness of the other mediums.
To evaluate how the flavorings performed, I chose two recipes to make that called for substantial amounts of vanilla extract and would allow the vanilla flavor to shine through: pound cake and sugar cookies. I tailored the vanilla concentration of my homemade elements for the cake and cookie recipes for which these elements were destined. Both recipes call for 1 tablespoon (15 grams) of vanilla extract, which my research suggests would have been extracted from about 1.5-1.65 grams of vanilla bean. So, the extracted flavor molecules of 1.65 grams of vanilla bean (I went with the high end of the estimate) are what I wanted to end up in 145 grams of my vanilla butter, in 250 grams of my vanilla sugar, and in 15 grams of my homemade alcohol-based extract.
To make the same quantity of each element and achieve the desired concentrations, I combined the following, in three separate jars:
- Alcohol-based vanilla extract: 750 grams vodka + 82.5 grams vanilla bean
- Vanilla butter: 750 grams butter + 8.53 grams vanilla bean
- Vanilla sugar: 750 grams sugar + 4.95 grams vanilla bean
I split the vanilla beans in half and placed them in the jar with their respective extraction medium. For the sugar, I also scraped the seeds from the pods in order to give it a fighting chance for flavor distribution against its more fluid competitors. (I included granulated sugar as a medium because, as I wrote above, vanilla sugar is a common ingredient in Europe. But if I were to run this experiment again, I would add a simple syrup to the lineup to find out whether being in liquid form would make sugar a better carrier.)
I placed the three jars in a water bath whose surface level reached just below the jar caps, and I used an immersion circulator to bring the water to 146 degrees Fahrenheit. I covered the whole apparatus with aluminum foil to trap the heat, and I left the vodka, butter, and sugar to do their thing (that is, attempt to extract as much flavor as possible from the beans) for three hours, after which I removed them from the hot water and let them slowly come to room temperature on the kitchen counter. (The butter, which had separated during heating, required occasional agitation in order to re-emulsify it). They rested overnight, and in the morning I put the vanilla butter in the refrigerator.
Testing the Vanilla Flavorings
For each recipe (the pound cake and sugar cookies), I made five versions: one using Nielsen-Massey’s Tahitian Pure Vanilla Extract (a Stella Parks favorite), one with the homemade alcohol-based extract, one with the homemade vanilla butter, one with the homemade vanilla sugar, and one without any vanilla flavoring as a control. For the recipes flavored with vanilla butter and vanilla sugar, I added a tablespoon of vodka to make up for the omitted alcohol-based vanilla extract. For the cakes, I replaced the coconut oil with butter or vanilla butter, and I made up for the fat loss by swapping in crème fraîche for a portion of the sour cream.
I baked all of the cakes and cookies in a single day, and over the course of three days a partner and I performed two blind taste tests and two blind smell tests, each time ranking the samples from most to least intense vanilla flavor and/or aroma. The first taste and smell tests were performed the day after baking, when everything had cooled to room temperature and settled overnight. The observation continued over the two days that followed. For the first taste test, the cake was served at room temperature, and for the second, it was warmed. For the smell tests, the cake and cookie samples were kept in airtight containers which their aromas could fill before being briefly opened for sniffing.
Our sniffing and tasting revealed the vanilla butter to have the strongest vanilla flavor and aroma, followed in descending order by the homemade alcohol-based extract, the store-bought extract, and finally the vanilla sugar. The vanilla-free control cake and cookies offered a good reference for the flavors and aromas that were not coming from vanilla.
Overall, the vanilla flavor was more pronounced in the pound cake than in the sugar cookies. And the reheated cake samples had a more pronounced vanilla flavor than when they were served at room temperature. To explain this, Serious Eats Senior Culinary Director Daniel Gritzer offers that volatile molecule activity increases with temperature, and he theorizes that pound cake's spongy texture (versus that of the denser cookies) offers more surface area for the release of aroma molecules.
The differences of flavor and aroma among the cakes and cookies with vanilla were subtle. But by the end of days spent sniffing and tasting, it had become clear that the vanilla sugar didn’t really hold up to baking; that the homemade alcohol-based vanilla extract was slightly more intense than the store-bought; and that the vanilla butter delivered on all fronts―intensified vanilla and butter flavors and aromas.
Considering just how intoxicatingly fragrant the vanilla sugar was in its raw state—before being baked into a recipe—I was surprised by the weak results it delivered in the finished cake and cookies. By way of explanation, there’s the fact that sugar is the most neutral-smelling of the substances I infused, so in a raw smell test, the vanilla gets to shine without any buttery or alcoholic distractions. And as Stella Parks points out, vanilla sugar’s delicious aroma “is like the top note of perfume: fleeting at best, and too volatile to survive the rigors of boiling or baking.” For this reason, I would use vanilla sugar in applications that require no cooking: to coat doughnuts or to sweeten whipped cream or to top oatmeal. (Vanilla confectioner’s sugar might be more useful, as it could dust all kinds of dessert dishes for an aromatic finishing touch.)
So, you’re thinking, I should switch from alcohol-based vanilla extract to vanilla butter, right? Well, that’s where things get a little complicated. The main reasons to stick with alcohol-based extract over vanilla butter are about food safety and shelf life.
I consulted a number of food safety experts and several of them confirmed the validity of my worries. "The bottom line," wrote Dr. Don Schaffner, co-host of the podcast Risky Or Not?, "is that the process you’re using has not been validated for control of pathogens like Clostridium botulinum [the bacteria that can cause botulism], which might well be present on the vanilla beans." And low-oxygen environments like oil and fat, the experts explained, are favored breeding grounds for pathogens that wish to proliferate. Vanilla bean being a dried fruit, the existence of harmful pathogens is a possibility, however minimal. That doesn’t mean that vanilla beans are dangerous; it just means that any tiny risk can multiply into a much bigger one in the right environment, which is why canning has such an aura of danger associated with it.
The heat used in my extraction process (146°F) is not high enough to kill all pathogenic microorganisms—but raising the temperature enough to do so would destroy the very flavor molecules we seek to coax from the beans. Refrigerating the vanilla butter after production would reduce the risk, but even then one expert warned against keeping it beyond four days. With all the effort and expense that goes into extracting vanilla flavor at home, I think it's pretty clear we want it to have a long and happy shelf life—well beyond four days.
This is where the alcohol-based extractions come out ahead. Microorganisms detest alcohol about as much as they love a room-temperature bath of fat. As a hospitable environment for flavor volatiles and an inhospitable one for pathogens, an alcoholic vanilla extract’s shelf life is effectively indefinite.
All this testing and research just to find out that I should stick to alcohol-based vanilla extract? Well, yes and no. The homemade vanilla extract performed better than the store-bought, and that’s a big deal. What it means is that while most recipes for homemade extract are a waste of time and money (as Stella has argued before), a more controlled approach can be worth doing.
My tests showed that there is a much more effective method for making vanilla extract at home than the internet-ubiquitous guidance to simply long-infuse vodka with vanilla beans. I’ve found that a really good homemade alcohol-based vanilla extract—more intensely flavored than even a top-quality commercial extract—is achievable; it just requires a bit of temperature control and the right ratio of vodka to vanilla beans. Plus, it cost me less than one-third the amount of money per ounce to make this vanilla extract than to purchase the one I used in my tests: about $1.80 per ounce for the homemade versus $5.50 or more per ounce for the store-bought, based on my purchases of vanilla beans from Native Vanilla, vanilla extract from Amazon, and vodka from the local liquor store.
Plus, alcohol-based vanilla extract is highly versatile. Unlike vanilla butter, which relies on a recipe that requires (or at least works with) butter, alcohol-based extract can be added to a much wider variety of foods. I’ve also found the extract to be the perfect storage environment for the beans used to make it; the alcohol keeps the beans plump for a future second use while the extract only gets smoother and more intense. I highly recommend giving my recipe for precision-heated vanilla extract a shot. Use the resulting liquid as a 1:1 replacement for store-bought extract. It will change the way you think about the DIY stuff.
- 83g grade B vanilla beans (2.9 ounces; about 31 beans), halved lengthwise (see note)
- One 750ml bottle of 80-proof vodka (25.4 fluid ounces; 3 cups plus 2 tablespoons), such as Tito’s
Set up an immersion circulator and preheat water bath to 146°F (63.5°C).
Place the halved vanilla beans in a 1-liter (33.8-ounce) glass jar. Add the vodka to the jar and use a fork or slotted spoon to push the vanilla beans down so that they are completely submerged in the vodka. Seal jar and gently lower into water bath; the water should reach just below the lid of the jar. Set a heavy mug, small bowl (filled with water, if necessary), or other weight on top of the jar to keep it from floating, and cover the top of the water bath with aluminum foil followed by a kitchen towel. Cook for 3 hours.
Using tongs, carefully remove jar from water bath (if you don’t want to deal with the hot water, you can also turn off the immersion circulator and let the water cool before removing the jar—this just takes a while). Allow vanilla extract to cool to room temperature (overnight is best). At this point, you can remove the beans from the extract and reserve them for other uses, or you can leave them in the liquid. (I recommend leaving them in the liquid; it’s the perfect place to park them until you need to fish them out for another use, and the flavor of the extract will only get better over time.) Either way, store the extract in a cool, dark place.
immersion circulator (sous vide machine), 1-liter (33.8-oz) glass jar with a tight lid
Grade B vanilla beans, also sold as “extract grade,” are a tad drier than their Grade A counterparts and may have superficial imperfections, but they are less expensive and contain all the flavor you need. They are available from Native Vanilla and Slofoodgroup.
Make-Ahead and Storage
Vanilla extract can be stored in an airtight jar or bottle in a cool, dark area. The vanilla extract won’t necessarily go “bad,” but for optimal flavor, the beans should be removed within 2 years and the liquid extract should be used within 5 years.
Lubinsky, Pesach, Séverine Bory, Juan Hernández Hernández, Seung-Chul Kim, and Arturo Gómez-Pompa. “Origins and Dispersal of Cultivated Vanilla (Vanilla Planifolia Jacks. [Orchidaceae]).” Economic Botany 62, no. 2 (2008): 127–38. http://www.jstor.org/stable/40390612.
McGee, Harold. On Food and Cooking: The Science and Lore of the Kitchen. United States: Scribner, 2007.
Editors of the American Heritage Dictionaries. The American Heritage Dictionary of the English Language, Fifth Edition. United States: Houghton Mifflin Harcourt, 2011.
McGee, Harold. Nose Dive: A Field Guide to the World’s Smells. United States: Penguin Press, 2020.
Ehrenberg, Rachel. “Yeast Bred to Bear Artificial Vanilla: Scientists Co-Opt Fungi to Produce Flavor More Efficiently.” Science News 175, no. 11 (2009): 9–9. http://www.jstor.org/stable/20494770.
Cicchetti, Esmeralda, and Alain Chaintreau. “Comparison of extraction techniques and modeling of accelerated solvent extraction for the authentication of natural vanilla flavors.” Journal of separation science vol. 32,11 (2009): 1957-64. doi:10.1002/jssc.200800650.
Experts from Cornell, Rutgers, and Purdue Universities, and U. of Illinois, provided opinions by e-mail.