Baking soda is sodium bicarbonate, which already includes one proton and so has a limited ability to take up more. But if you heat baking soda, its molecules react with one another to give off water and carbon dioxide and form solid sodium carbonate, which is proton-free.
Actually, the reaction he's talking about is relatively simple. It's the lime kiln effect, done all the time to make lime from calcium carbonate:
CaCO3 --> CaO + CO2Similarly, using baking soda,
NaHCO3 --> NaOH + CO2
I'm afraid Harold McGee doesn't understand what he's talking about - claiming that it goes to sodium carbonate (aka washing soda) rather than to sodium hydroxide means that he gives a bunch of cockamamie advice from his bad information.
Actually, it IS lye, just probably a mixture of lye and baking soda. Some of the weight might come off of it in the form of water due to hydrated crystals, but most of it will be carbon dioxide. I'd wager if you put your baking soda through an oven self-cleaning cycle, on the other hand, you'd end up with a much higher proportion of NaOH versus baking soda, possibly over 90%. So don't handle it as if it's slightly stronger than lye, handle it for what it is: lye.Just spread a layer of soda on a foil-covered baking sheet and bake it at 250 to 300 degrees for an hour. You’ll lose about a third of the soda’s weight in water and carbon dioxide, but you gain a stronger alkali. Keep baked soda in a tightly sealed jar to prevent it from absorbing moisture from the air. And avoid touching or spilling it. It’s not lye, but it’s strong enough to irritate.
Oh, also you can make food grade lye by baking baking soda. Who knew?