Hi Rolnor, Kriggy and the others,
I should have told it more clearly. The paper's single step achieves a strong replacement of H by D in the target molecule because D2O is in large excess indeed.
To my understanding, less D2O would limit the achieved D/H proportion at the target as water loses some D exchanged for H. There is no selectivity here, only some levelling among the molecules.
For some uses, for instance to produce deuterated molecules for sale, the consumption of D2O would be a waste. To bring the same amount of D to the target, exploiting D2O from <100% purity to >98% means <2% of the bought D is used, letting consume 50* too much D2O.
My multi-step (or continuous) proposal, not quite convenient at lab scale for sure, lets save on the D2O consumption by depleting it further and still achieving high D proportion in the target.