So after experimenting with homemade nut milk ice creams for a while, I wanted to go back and do a few test runs of improvements I’ve had in my head for some of my oat milk-based ones. The first thing I wanted to see was how well maltodextrin (frequently used in sorbets just to add to the raw amount of solids) helps compensate for some of the difference in protein levels between oat milk and dairy milk. While I was trying that out, I ran out of pea protein powder and decided to just see if the maltodextrin could also compensate for that drop in protein. The batch I made was essentially just ‘fior de oat milk’, if you will – it also included a changed fat ratio and a pre-whipping step before the churn. The result was one of the smoothest, creamiest batches of ice cream I have made – period. And furthermore, the flavour was just really nice! Despite being just oat milk and… hang on a second, why would the flavour be different?

And then I realised: the pea protein I was using in my ice creams was adding a slight aftertaste to all my batches. I was used to covering it up, and it usually wasn’t as obvious when the ice cream was frozen, but it was there. I just assumed it was me just not being used to oat milk, but it turns out that I actually quite like oat milk – when it doesn’t have pea protein in it.

Whoops.

I felt immediately stupid for not working it out sooner – all the evidence was in front of me! During earlier vegan dulce de leche experiments, I tried adding pea protein to one batch, hoping increased protein would promote the Maillard Reaction. The result was quite bitter; I thought I’d just burnt it, but now I clearly recognise the flavour as being the result of the pea protein. It was also contributing to the bitterness in my batches I was doing directly over the stove top. I guess it turns out that pea protein just scorches easily.

I had to verify it wasn’t just a matter of the brand of oat milk, though, so I tried some pea protein in cold oat milk, and then neat oat milk. What do you know – there was that aftertaste again. I went on to make a couple of batches of my new vanilla recipe with maltrodextrin instead of pea protein and the other changes I mentioned above produced batches that I’m quite happy with. A chocolate batch made using the refined recipe was described as ‘having no right to be that good’, which I quite liked hearing.

So, with that all in mind, I’ve gone through and implemented all those changes in all my recipes. The fat balance change means they are now much softer (perhaps even a bit too soft if your freezer isn’t cold enough) and the maltodextrin has helped make them far creamier. I’ve also done some tests with giving the custard a quick whipping before churning but found it to be largely unproductive – for more air incorporation, letting the base thicken in the fridge beforehand seems to be more effective.

I’ve also started testing a black sesame ice-cream, my second request, that shouldn’t take too long to refine – my first two batches came out pretty amazing!