Interesting article but left me wondering why they don't use ¹H₂¹⁶O as the "standard water"? Surely you can make that (relatively) easily in a centrifuge by spinning out the lightest fraction of water.
At the very least that has the advantage that it's possible to make more of it, so you aren't dependent on a limited sample of standard water.
18 O enriched water is available to buy but the pricing suggests it's quite expensive to separate 16 O from 18 O. And you get much less variability in the ratio of isotopes from mixing compared to separating a single isotope as much as possible.