- Dough is easy to roll out
- Crust bakes up tender and flaky
- Consistent results every time
Pie crust in a nutshell: Mix flour, salt, and sugar together, cut in some fat, add water just until the dough sticks together, roll it out, and bake it. A study in simplicity. Yet it can all go wrong so easily. The dough is almost always too dry and crumbly to roll out successfully. The crust is either flaky but leathery or tender with no flakes. And the results are seemingly random: The recipe that gave you a perfect crust last month resulted in a tough-as-nails crust when you followed it this week.
I wanted to figure out exactly where a crust goes south, so I set out to sort through all the dubious science, purported secret ingredients, and perennial pie crust theories to separate fact from fiction and create a recipe that not only bakes up tender and flaky every single time, but also rolls out easily.
The first question was what type of fat to use. The test kitchen likes the rich flavor of an all-butter crust. Problem: Butter starts to soften at around 50 degrees and fully melts at around 100 degrees, which means the crust has to be worked very quickly. Also, butter’s high water content (about 20 percent; the rest is fat) can lead to leathery crusts, as too much water will stimulate the formation of gluten, the protein matrix that provides structure in baked goods. Enter hydrogenated vegetable shortening, a soft fat that doesn’t melt until a relatively high temperature and contains no water, just fat. But although crusts made with shortening are very tender, they have virtually no flavor. I ultimately found that a combination of butter and shortening provided the best balance of flavor and tenderness.
I moved on to the next step: cutting the fat into the flour. Of all the methods I tried (food processor, standing mixer, pastry blender, and by hand), the food processor was the fastest and most consistent. Even so, I ran into my first major hurdle—some recipes call for cutting the butter into walnut-sized pieces, and others say to incorporate the fat until it resembles wet sand. Which approach is better? And once you determine which method to use, is it possible to produce same-sized pieces of butter time after time?
What if I ran the food processor until the flour and fat were completely combined? This is simple to repeat every time, and there’s no way to over-process it. But dough is supposed to have pockets of fat in it, which melt upon baking to leave behind the gaps that create flaky layers. By fully incorporating the fat, I was left with no pockets, and sure enough, my dough baked up with no flakes. My next attempt was to process only a portion of the fat completely into the flour, freeze the rest of the fat, and grate it into the mixed dough to create those fat pockets. Consistent? Yes. But despite the fact that there were plenty of pockets of unmixed fat, my crust still came out flake-free.
While I was testing methods for incorporating the fat into the flour, I had been dealing with the frustrating issue of how much water to add to the dough. Some recipes call for a range of water that can vary by as much as 100 percent, claiming that a hot or humid day can throw measurements off. This excuse seemed a little suspicious, and I was eventually able to dismiss the theory by measuring the effects of humidity on flour. It was time to step back and examine the structure of a pie crust.
When fat is being cut into flour, the flour is separated into two groups; some of the flour is coated with a layer of fat, which protects it from absorbing any water, while the uncoated flour will absorb water and form gluten. When the dough is rolled out, this gluten stretches into sheets separated by pockets of unmixed fat that melt upon baking, leaving behind crisp, separated sheets. The problem is that depending on who’s making the crust, the exact temperature of the fat, and even the type of food processor being used, the ratio of fat-coated flour to uncoated flour can change drastically from batch to batch. This means a pie crust recipe that barely absorbed 1⁄4 cup of water one time might readily absorb 1⁄2 cup the next. It also explains why the same recipe is flaky one day but not the next: For consistent flakiness, you need the same ratio of fat-coated flour to uncoated flour.
It’s not just the chunks of fat that create flakiness. It’s also the uncoated flour that mixes with water and forms gluten that guarantees a flaky crust. This explained the failure of the test in which I combined all the flour with some of the butter, then added grated butter to the dough. You need at least some flour that hasn’t been coated with butter in the dough in order to create the gluten layers that form flakes. When processing the fat in a traditional crust, leaving some chunks of butter in the dough is a good sign that the dough hasn’t been overprocessed (that is, chunks of butter in the dough are an indication that there is enough uncoated flour left to combine with water and create a flaky crust).
What if I measured out the two types of flour— the portion I wanted coated with fat and the portion I wanted to remain uncoated—separately? Rather than starting with all the flour in the processor, I put aside 1 cup of flour, then placed the remaining 1 1⁄2 cups of flour in the food processor with all of the fat and processed it until it formed a unified paste. I then added the cup of reserved flour back to the bowl and pulsed it just until it was evenly distributed around the bowl. This would guarantee the dough had a constant amount of uncoated flour to mix with the water. After mixing in the water and rolling out the dough, I now theoretically had a dough with two distinct parts: long sheets of gluten separated by a flour-fat paste.
The dough baked up as flaky as could be. And since the stage in which the fat gets processed into the flour was no longer ambiguous, my new crusts came out identically, time and again.
I had guaranteed flakiness, but tenderness was still a crapshoot. Most recipes with 2 1⁄2 cups of flour call for 6 to 8 tablespoons of ice water. If I kept the water at the lower end of this range, the dough baked up very tender but was dry and hard to roll out. When I used the full 8 tablespoons, the dough was smooth and easy to roll out but baked up tough—too much gluten was forming. I had to figure out a way to tenderize the finished crust without reducing the amount of water I used.
Scanning through recipes turned up a common “miracle ingredient”—acid. Many recipes say that a teaspoon of vinegar or lemon juice can tenderize dough, claiming that gluten formation is inhibited at lower pH values. But after consulting our science editor, I learned that gluten formation is actually increased in slightly acidic environments (a pH of between 5 and 6) and doesn’t begin to decrease until the pH drops below 5. This required replacing nearly half the water with lemon juice, by which point the crust was inedibly sour.
What about using lower-protein cake or pastry flour? No good. The crusts baked up sandy and too short. What about adding cream cheese or sour cream? Although this made the crust more tender, it had a strange, soft chewiness.
Let’s review: In order to roll easily, dough needs more water, but more water makes crusts tough. Therefore, I needed something that’s not water but is still wet. As the aromas from a nearby pan of reducing wine reached my nose, the answer hit me like a bottle to the head: alcohol.
Water Alone: In pie dough made with water alone, the water interacts with the flour to form a good deal of gluten, which can cause the dough to be tough.
Vodka and Water: Gluten does not form in alcohol. Therefore pie dough made with a mixture of vodka and water has less gluten and is much more tender.
Eighty-proof vodka is essentially 40 percent ethanol and 60 percent water. As it happens, gluten cannot form in alcohol, which means that for every tablespoon of vodka I added, only 60 percent of it contributed to gluten development.
I made a batch of pie dough with 4 tablespoons each of cold vodka and water. The resulting dough was as smooth as Play-Doh, and I couldn’t have made it crack even if I’d wanted to. I was tempted to toss it, thinking it would bake up tough as leather, but giving good science the benefit of the doubt, I baked it anyway. It was an unparalleled success. The dough baked up every bit as tender and flaky as any crust I’d ever had, without a hint of booziness to give away its secret. One hundred forty-eight pie crusts later, I’d finally come up with a recipe that is 100 percent reliable.
Dough is easy to roll out
Replacing some of the ice water with chilled vodka lets you add more total liquid so the dough is smooth and easy to handle without baking up tough. Working with chilled, slightly softened dough ensures a successful roll-out.
Crust bakes up tender and flaky
We cut the butter with some shortening, a pure fat that doesn’t contain water (which encourages gluten development). Alcohol doesn’t trigger gluten development, so the vodka-spiked dough still bakes up tender.
Consistent results every time
A food processor is essential for making this dough—it’s not only fast but it also ensures consistent results. By adding the flour in two distinct steps, the ratio of fat-coated flour to uncoated flour stays constant from batch to batch.