The Science Behind Vegan Meringues

Working toward a more stable meringue made from the liquid in a can of chickpeas.

Meringue on pink background

Serious Eats / Amanda Suarez

Aquafaba, or the liquid that’s in a can of beans, is one of the most exciting cooking “discoveries” in recent memory. Not only was it found to be a superior vegan substitute for eggs than what had been commercially available; it was found by curious home cooks across the globe, experimenting in their kitchens to find something that could approximate the way egg whites behave in delicate desserts like meringues.

Aside from its wide availability and the fact that it’s very affordable, I like the idea of working with aquafaba because it’s a way to use up something that in many cases would otherwise be poured down the sink. However, in experimenting with aquafaba as an egg substitute, I’ve found that it sometimes works beautifully, as in Kenji’s aquafaba-based pancakes, but it sometimes doesn’t, like the time I tried to substituted it for eggs to make brownies. I decided to take a deeper look at aquafaba and see how it behaved under different conditions, and in the process try to come up with a good aquafaba meringue recipe.

Why Aquafaba Is a Good Egg Substitute

a can of chickpeas draining their liquid (aquafaba) into a glass measuring cup

Nik Sharma

Aquafaba’s utility lies primarily in its ability to make foams. Foams are emulsions, just like a salad dressing or a sauce, but while a sauce like mayonnaise is an oil-in-water emulsion and a vinaigrette is a water-in-oil emulsion, foams are air-in-water emulsions. That is, air molecules are dispersed in and surrounded by water molecules. 

Aquafaba is rich with substances that enhance water’s foaming abilities because legume seeds like chickpeas and soybeans contain albumins and globulins, the same proteins found in eggs that make them so useful in creating culinary foams. They’re also rich in saponins, plant-derived organic chemicals that act like soaps* in that they can dissolve in both fat and water, and they, too, can help to form stable foams. The combined presence of albumins, globulins, and saponins in chickpea cooking liquid means that, when agitated, it will produce a stable foam.

* “Saponin” is derived from “sapo” in Latin, which refers to soap.

Chickpeas happen to have a lot of albumins, globulins, and saponins, but other legume seeds also contain these components and can be used to make aquafaba. However, if you do use other legume aquafabas as egg replacements, you should expect them to produce noticeably different results; legumes will vary in their chemistry, and some aquafabas won’t produce as tight a foam compared to one made from chickpea aquafaba. Also, most studies show that chickpea aquafaba has been found to have the most “neutral” flavor; other aquafabas will look, smell, and taste different. Even different types of chickpeas produce aquafabas that have noticeable differences in color and the ability to produce foams.

A More Stable Aquafaba Meringue

The question about aquafaba I was most interested in pursuing was whether you can manipulate it to make a more stable foam, one that could be used to produce a meringue that didn’t fall apart after a few hours. 

I knew that in an egg-based meringue, proteins like albumin are denatured using a combination of mechanical stress and a lower pH (accomplished by adding cream of tartar, an acid) as the mixture is vigorously aerated to create a foam. It seemed logical that the proteins in aquafaba would behave in the same way as the proteins in egg whites in the presence of cream of tartar.

However, aquafaba doesn’t just contain protein; it also contains those saponins, which I knew created more stable foams at an alkaline pH based on previously published research. Therefore, I reasoned, increasing the pH of aquafaba by using an agent like baking soda should help create a more stable foam. 

I should note that the increased stability would be in addition to the stability provided by the sugar in meringues, which, because of its ability to bind with and hold onto water, helps to prevent proteins on the surface of the foam from drying out too quickly, which increases the overall stability of the foam.

Aquafaba and pH: The Effect on Meringues

Experiment showing aquafaba mixed with cream of tartar, plain, and mixed with baking soda in vials

Nik Sharma

In order to test the relative strengths and weaknesses of meringues made with aquafaba and the addition of acidic and basic ingredients, I used a single meringue recipe but altered the aquafaba used in the recipe in two of three batches, lowering or raising the pH in the aquafaba by adding cream of tartar and baking soda. For the third batch, I used plain aquafaba.

The foams prepared from each experiment were left to sit in order to gauge how long they could last without falling apart. Another set of foams were baked to see how they performed when heated and to gauge how well they could hold their structure in meringue cookies.

In every experiment, I used the liquid from salt-free canned chickpeas, which would not only ensure that the aquafaba was of consistent quality, but would also mean that my results would be easily reproducible for people at home (provided, of course, they also used canned chickpea aquafaba). However, I ran the experiments several times, and I used different brands of chickpeas, which led me to a surprising finding about chickpeas that have been cooked with kombu (more on that below).

The Results: The Untreated Control

a test of plain aquafaba baked into a meringue

Nik Sharma

When aquafaba was whisked with just sugar, the liquid emulsified to form a foam, but it wasn’t very stable. Within an hour, the foam started weeping liquid and began to fall apart. However, re-whipping the mixture helped bring the foam back together, and, when baked, the meringue held its shape. The only issue I had with baking was that the meringue started to take on a brownish hue, since there was nothing to reduce the rate of the caramelization and Maillard reactions. (Also see my results from the Kombu cooked chickpeas).

The Results: Aquafaba With Cream of Tartar

a test of aquafaba mixed with cream of tartar baked into a meringue

Nik Sharma

 Adding cream of tartar to aquafaba yielded the most consistently good and reproducible results. Not only does the low pH from the acid help reduce the brown color developed by the caramelization and the Maillard reactions; it also helps modify the proteins’ structures to create a stable foam that aerates very well and holds its shape both at room temperature and after baking. Even after sitting out on my kitchen counter for 16 hours, the uncooked foam was very stable and didn’t fall apart. The baked meringues made with cream of tartar were often much smoother in appearance than the ones that contained none; the latter looked a bit speckled because the bubbles in the foam were a little unstable. 

The Results: Aquafaba With Baking Soda

a test of aquafaba mixed with baking soda baked into a meringue

Nik Sharma

Saponins form stable foams at an alkaline pH, so I assumed adding baking soda to aquafaba and aerating it would produce a very stable foam. I was wrong: after emulsifying the aquafaba, the foam lost its volume at room temperature within an hour; when it was baked, the foam fell apart. (This batch of meringues, unlike any of the others, also stuck to the parchment paper I’d lined the baking sheet with, and wouldn’t come off.)

A second consequence of using baking soda is that the resulting meringues look and taste a little odd; they both take on more color and have a “toasty” flavor, both of which are undesirable in meringues. The alkaline pH produced by baking soda accelerates the caramelization (a complex set of reactions that involve sugars) and the Maillard reactions (a complex series of reactions that involve the amino acids of proteins and certain sugars called reducing sugars).

Overall, the effect of baking soda on aquafaba meringues seemed to suggest that saponins might play a smaller role in the formation of aquafaba-based foams than the proteins albumin and globulin. 

The Results: Kombu and Aquafaba

While I was working on this recipe, I started to get some conflicting results when I switched my brand of canned chickpeas, and I realized it was because some brands, like Eden, use kombu when they cook their beans. Kombu is often used to cook beans because it contains alpha-galactosidase, an enzyme that breaks down raffinose, the indigestible carbohydrate in beans responsible for flatulence.

It turns out that aquafaba with kombu can make a very stable meringue, even in the absence of cream of tartar. The  reason for this is that kombu contains carrageenan, a carbohydrate that is commonly used as a food additive (you’ll find it in mac and cheese products, or condensed milk) because it acts as an emulsifier and stabilizer. 

However, in the presence of baking soda, it took a very long time to produce stiff peaks when I whipped aquafaba from a can of chickpeas cooked with kombu, the resulting foam wasn’t very stable, and, when baked, the meringues fell apart.

Heating Aquafaba Meringue to Improve Stability

a whisk showing stiff white peaks with aquafaba meringue

Nik Sharma

Inspired by Stella’s Swiss meringue recipe, I also tested whether heating the three different aquafabas could yield a more stable foam. I heated the mixtures to 175°F (79°C) and 183°F (84°C) (the same temperature at which ovalbumin in an egg denatures) in separate trials. The hypothesis here was that the albumin proteins present in the aquafaba might denature around the same temperature as egg albumin and, when whipped, they would create a more stable foam.

I found that I could produce a foam with stiff peaks using the plain aquafaba and the aquafaba with cream of tartar, both with and without kombu (the aquafaba with baking soda performed as poorly as in previous tests), and, when baked, the meringues held their structure. The results were similar to the results of the tests without the heating step, so as a general matter, I don’t think heating is necessary.  However, the heating of the meringue gets rid of most of the chickpea aroma, and the method does yield a silkier marshmallow-type frosting for cakes and pies, one that’s much smoother in texture.

Using Different Kinds of Chickpeas

I tried a variety of brands of chickpeas in these tests, and there was a lot of variance in the results, even if the only ingredients listed on the cans were chickpeas and water. For example, it took me half the time to whip up a meringue with aquafaba from cans of chickpeas sold by the brand Westbrae than any other brand. Eden’s chickpeas, which are cooked with kombu, consistently gave me the best results.

I also tried making meringues with aquafaba from chickpeas I prepared myself from dried, and those experiments didn’t work out very well. I found that the ratio of water to chickpeas that I normally use when preparing dried chickpeas yielded a much less viscous aquafaba. Since canned chickpeas are so widely available, and since I typically season the cooking water for dried chickpeas with spices and aromatics that would be distracting in desserts, I recommend sticking with canned aquafaba for the meringue recipes I’ve developed.

While dehydrated aquafaba powder products are now available, I did not test any of them in this set of experiments.

How to Make Aquafaba Meringues Less 'Beany'

aquafaba-cardamom meringue cookies on a blue plate on a pink background

Nik Sharma

Perhaps the most important question for aquafaba-based meringues is whether or not they taste and smell like beans. As I noted above, cooking the meringue does get rid of a lot of that beany aroma, but it’s still detectable. In the end, I found that the acceptable level of beaniness in a meringue depends primarily on the type of aquafaba you use, how sensitive you are to beany flavors and aromas, and the flavoring agents you use to mask those flavors and aromas.

Aquafaba made from chickpeas is considered to have the most neutral flavor compared to aquafaba made from other legumes. As I tested my recipes, I found 1 teaspoon of high-quality vanilla extract was unable to mask the bean flavor completely, but if I used stronger aromatic spices or flavoring ingredients, like cardamom or rosewater, that bean smell was basically undetectable. Here are some flavorings that I found worked particularly well in each of the recipes:

  • ¼ to ½ teaspoon of ground green cardamom
  • ¼ to ½ teaspoon ground fennel or star anise 
  • ½ teaspoon instant espresso or coffee powder plus 1 teaspoon vanilla extract 
  • ½ teaspoon rosewater