The pH of a batter decides its colour, its set and its crumb

A narrow dial with disproportionate effects

Most cake batters sit somewhere between mildly acidic and mildly alkaline, in a band perhaps two or three points wide on the pH scale. Because that scale is logarithmic, a shift that looks small on paper represents a tenfold change in the concentration of hydrogen ions at each step. Proteins, starches and the browning reactions are all sensitive to that concentration, so a modest change in the acid content of a recipe reaches several different parts of the outcome at once.

This is why swapping buttermilk for plain milk, or natural cocoa for an alkalised one, so often produces a cake that is wrong in ways nobody predicted. The substitution was made on grounds of flavour or availability, and it silently moved a variable that the rest of the formula had been balanced against. Recipes rarely mention pH at all, which leaves the baker adjusting it constantly without a name for the thing being adjusted.

Colour is the most visible consequence

The browning reactions between sugars and amino acids run faster in alkaline conditions, which is why a batter carrying excess bicarbonate produces a darker, more deeply coloured crust than the same batter with the acid balanced. Pretzels take this to its extreme with a strongly alkaline dip before baking, and the resulting mahogany surface has no equivalent in ordinary bread. In a cake the effect is gentler but the direction is identical, and an unexpectedly dark sponge is often a leavening imbalance announcing itself.

Cocoa is where this becomes a genuine choice. Natural cocoa is acidic, and its pigments read reddish in an acidic batter — the traditional reason a devil’s food cake had a red cast long before any colouring was added to one. Cocoa treated with an alkali is darker, less bitter, milder in flavour and close to neutral, which is why it dissolves more agreeably into a drink and why it looks almost black in a biscuit. Neither is better. They are different ingredients wearing the same name.

Acidity moves the moment a cake sets

Egg proteins coagulate most readily near the pH at which they carry no net charge, and moving a batter towards acidity generally brings that setting point forward. An alkaline batter sets later, giving the bubbles more time to expand before anything rigid contains them. That extra time sounds like an advantage and occasionally is, but past a certain point it produces the classic over-alkaline crumb: coarse, open, tunnelled, and inclined to sink in the middle because the structure arrived too late.

Acid also alters starch behaviour, since starch granules swelling in an acidic environment break down more readily and thicken slightly less. In a cake this shows up as tenderness rather than as a failure, which is one reason cultured dairy has such a good reputation in baking. Buttermilk, yoghurt and sour cream are contributing acid, fat and protein simultaneously, and the tenderising credit usually given to fat alone is shared with the acid alongside it.

Acid weakens gluten, which is often the aim

A moderately acidic environment interferes with gluten development, so a batter or dough made with an acidic liquid tends to be softer and less elastic than the same one made with water. In cakes and muffins that is precisely what is wanted, since gluten there is a liability rather than an asset. The effect is not linear, though, and strongly acidic conditions eventually break down protein far enough to weaken the structure outright, which is a real possibility in heavily fruited batters.

Sourdough sits at the interesting end of this. The lactic and acetic acids produced during a long fermentation soften the dough noticeably towards the end of the process, which is one reason a very extended proof can tip from developed into slack with little warning. Bakers manage it by shortening the bulk stage or lowering the temperature rather than by adding anything, which is a good illustration of pH as a variable that accumulates over time rather than being set at mixing.

Where the trade stops being worth it

Everything above has a limit, and the limits are mostly flavour. Excess acid tastes sour and can curdle dairy in the batter before it reaches the oven; excess alkali gives a soapy, faintly metallic note along with the darkened crumb, and it destroys some of the vitamins present in the flour. The window in which the structural benefits arrive without the flavour penalty is not especially wide, and most traditional recipes have already been tuned to sit inside it.

The practical upshot is to treat any acid-bearing ingredient as part of the leavening system rather than as flavouring. If you replace buttermilk with milk, the bicarbonate in the recipe no longer has a partner and something else has to supply one. If you switch cocoas, expect the colour and the rise to move together. None of this requires a meter on the worktop. It requires noticing that the ingredient you changed was doing chemistry as well as tasting of something.