How much baking powder to use

I posted previously about a dramatic biscuit failure I experienced when I forgot to include baking powder, due to a careless reading of the recipe. Instead of regular flour, it called for “self-rising flour”, which already has baking powder mixed in. The strange thing is, no one seems to agree on exactly how much baking powder should go into self-rising flour. Casual googling turned up recommendations for 0.5 tsp, 1 tsp, 1.25 tsp, and 1.5 tsp (same as 0.5 Tbsp) as the amount of baking powder to add for each 1 cup of flour.

Now the difference between 0.5 and 1.5 tsp may not sound like a lot, but consider that it represents a 50% increase or decrease from a middle value of 1.0 tsp. For something as sensitive to stoichiometry as baking is, I’d expect that to make a difference. Then again, it seems reasonable that the desired amount would vary depending on the item being baked and how much loft you hope to get — which sort of defeats the purpose of a pre-mixed flour-baking powder product.

But even specialized biscuit recipes disagree on this, but seem to choose either 1 tsp or 0.5 Tbsp (1.5 tsp, in agreement with my mom’s recipe). (As a side note, they also vary widely on how much shortening or butter to use, as well as how much milk or buttermilk to use and whether or not to chill the dry ingredients + butter. The number of permutations sent me into a brief paralysis (gosh darn it, shouldn’t we have converged on a solution by now?!) until I decided to give up on the web and just use my mom’s recipe.)

I decided that a scientific test was called for. I split up the flour involved in a batch of biscuits (2 cups) into three bowls, for baking powder:flour ratios of 1 tsp:1 cup, 1.5 tsp:1 cup, and 2 tsp:1 cup. There was enough material to make two full biscuits of each type, plus some extra left over for a partial-biscuit. I measured the biscuit height before and after baking. The following chart shows the average (across two biscuits) difference in height I observed (data points in blue, average in red).


A clear difference emerged! It’s even almost linear, which is a bit surprising given the small sample size. Now it would be interesting to try even smaller and larger amounts of baking powder… the curve is likely to have an interesting shape at both ends. But for food, one of the most important measures of success is not size, but taste. I sampled all of the results and found that I couldn’t really tell a difference between the 1.5- and 2-tsp results, but the 1-tsp biscuits were noticeably less fluffy. I conclude that the wise biscuit baker should avoid self-rising flour with less than 1.5 tsp of baking powder per cup of flour (or avoid it altogether and just add your own ingredients).

Some other notes:

  • One of the annoyingly tedious parts of making biscuits or scones is the “cutting in” step that gets the fat (butter or shortening or whatnot) into the flour. I used a tip from my friend Evan: freeze the butter, then use a cheese grater to shred it into the flour. Mix with fingers. Works like a charm!
  • Some biscuit recipes call for milk, some for milk with lemon juice added, and some for buttermilk. Ever wondered why? Well, adding lemon juice or using buttermilk lowers the pH of the liquid (makes it more acidic). And chemical leaveners such as baking powder and baking soda are basic, therefore in theory should react more strongly in an acidic environment (giving your baked good more “rise”). But baking powder is baking soda pre-combined with its own acid (cream of tartar). So you shouldn’t actually need an acidic liquid. I tested this by dropping some baking powder in water, then in buttermilk. If anything, the baking powder reacted more to the water than the buttermilk. (I should do the same test with baking soda.) This also explains (maybe) why some recipes use baking powder and others use baking soda + cream of tartar — the latter want control over the ratio, just like the self-rising-flour issue!
  • Biscuit aficionados recommend the use of flour with a lower protein content (to get even more loft) such as cake flour; see How to make the best Buttermilk Biscuits from Scratch. I haven’t tried this one yet, either.
  • I actually did a parallel experiment, with the same three types of mixtures, but first chilling the dry ingredients + butter. However, a distraction at a critical moment caused me to forget to measure the biscuits before they went into the oven! So I only have their post-baking heights. If anything, the relationship seemed weaker, with less rising action. While a firm conclusion should await more reliable data, for now I’m going on the assumption that the chilling step is unnecessary. (Taste is unaffected, too ;) )

Clearly, the field of interesting experiments with ingredient combinations is a rich and open one, even just for making biscuits!

7 Comments
7 of 7 people learned something from this entry.

  1. jim said,

    April 25, 2009 at 1:29 pm

    (Learned something new!)

    Nice work and thank you for the analysis!

    Thanks, also, for relaying the tip of freezing the butter then shredding it with a cheese grater. I’d just used a pastry cutter to mash, mash, mash it down.

    (By the way, freezing ginger root makes it much easier to grate.)

  2. Jimmy Huynh said,

    April 25, 2009 at 8:34 pm

    (Learned something new!)

    Thanks for this! I am actually wondering the same thing but for chinese steamed buns.

    Would be curious to see when rise decreases and when taste is affected.

  3. Susan said,

    April 26, 2009 at 7:36 am

    (Learned something new!)

    I have nothing of substance to say, just that this was an awe-inspiring exercise in food geekery that I will actually use. I’d never even thought to do a structured experiment. Bravo!

  4. Terran said,

    April 26, 2009 at 2:44 pm

    (Learned something new!)

    This is incredible! I am in awe of your kitchen science fu. I have never been so principled about my kitchen experimentation. (Result: much more futzing around to arrive at much more heuristic recipes.) I will be sure to apply your baking powder guidelines.

    Did you check the p and r^2 values for your data fit? ;-)

    A couple of auxiliary notes:

    – I don’t fully understand the chemistry of baking powder (although Wikipedia appears to have a very nice article. But most commercial baking powders are more sophisticated than baking soda+cream of tartar/tartaric acid. They contain an additional acid that activates at high temperature, making them “double acting”. The point, as I understand, is to cause a secondary rise cycle, yielding higher rise in the oven. Plus it desensitizes the batter from the timing a bit.
    – The relationship would be predicted to be linear because the amount of C02 produced by the leavening should be proportional to the mass of leavening used, up to structural constraints. I conjecture that the linearity will fail when the gas volume exceeds the structural capacity of the dough or batter (i.e., the item falls). That’s an empirical limit, however.
    – Note that the curve should be parametrized by the ambient air pressure. In Albuquerque, for example, I predict a sharper slope and an earlier structural failure point. (Part of why baking at altitude is a challenge — very, very few recipes are properly calibrated.)

    Finally, on the construction of biscuits, I have observed a significant dependence on the form of the “rolling out” step. I have found (through non-rigorous, non-quantitative) experimentation that a process of repeated folding is important. This has some of the effect that steel folding in Katana construction or Damascus steel does — it creates a many-layered structure, with “softer” (butter) material separating layers of “harder” (flour dough) material. While it doesn’t improve the hardness or temper of biscuits, per se, this technique does yield nicely flaky layers. ;-) I have found empirically that 4-6 folds is about optimal; fewer folds yields a noticeably less layered and flaky pastry, while more folds toughens the dough as the gluten begins to develop.

    If the whole “senior research scientist at JPL” gig doesn’t pan out for you, you clearly have a place in America’s Test Kitchen. ;-)

  5. Emma said,

    March 12, 2019 at 5:15 am

    (Learned something new!)

    I have learnt something new today.

  6. iimran ali said,

    April 5, 2019 at 10:47 am

    (Learned something new!)

    dear sir i have topic on data graph of baking plzzz send some thing for this topic sir

  7. Ty said,

    June 10, 2021 at 9:36 pm

    (Learned something new!)

    Going to try this in a vegan whole wheat drop biscuit recipe to see if I can lower the density. Thanks a ton

Post a Comment

I knew this already. I learned something new!