The major problem with ultrasound as a mechanism of purification is that it doesn't break molecules. Heat at least denatures proteins and breaks hydrogen bonds, but ultrasound is of a just smaller order of magnitude of energy at the atomic scale, which can be of the order of the adhesive forces holding the liquid together, but not of stronger molecular bonds.
But I think you can do it a different way: use a sound waves intensity gradient to move the biological impurities in the water to a part of the container, by having them walk down an effective potential gradient, like optical tweezers move molecules. This requires only that the density/stiffness of the molecules be different from water, so that the sound energy at a given mode is different inside the molecules than in the water. If it is greater, the molecule will move to the regions of greater intensity. If it is less, it will move towards the regions of smaller intensity. By arranging the sound wave to have an intensity gradient, you can make all the molecules segregate towards/away from the microphone, leaving water in the middle with only ionic or small molecule impurities, which are not affected by the sound.
You can flush the sides away, and repeat to make a purer water. This might work for getting rid of prions, which aren't disinfected by boiling.
This article is the only thing I found that might be relevant, but it is paywalled:
I think this might be a very useful idea.