Patent EP1238715
The above patent goes into exotic details of enhancing the effectiveness of sonochemistry, but it doesn't have to be so complicated.
A 555 timer IC chip could produce the 20 or 20+KHz signal (see Forrest Mims' mini-books at Radio Shack, "555 Timer Circuits", etc.), then a suitable amplifier (plans and datasheets everywhere, op amps, jfet, mosfet) could provide the power levels. Transucers are available everywhere also.
As with other chemical equipment, the price of ready made is sky high and inflated to the nth degree.
I mean, yes, it has space-age contours and specially-formed stainless steel, however a transducer attached to the outside of a flask could easily suffice, and in fact I have seen pics of that very type of setup.
It's too bad there isn't more on espacenet and pubmed.
I frequently find, as time has progressed (see literature and patents), chemists have adopted more complicated, expensive and dangerous methods to "keep up with the Joneses", much as in other branches of science, abandoning simple cheap methods. Sonochemistry must work too well and be too simple. As I heard in a good movie recently, some people "are just too hip to be happy".
I may try it sometime in some months, as one of the precursors I'm interested is a bit pricey for my taste, but tricky to make, my way, due to steric hindrance and the general inactivity of the moiety. I've read sonochemistry can increase some chemical reaction rates a million fold, and shorten reaction times spectacularly.
I was also thinking, on the subject of ready-made vs homemade equipment, that a reaction kettle is perfect (but, sticker shock awaits those who look for them) for most things, easier to clean out than a flask, more balanced to help prevent embarrassing and dangerous breakage accidents, plenty of room for a large stirrer, multiple inlets, flat-bottomed to obviate the need for a mantle, etc. I thought of making a lid/stopper for a beaker out of rubber or a chem- and heat-resistant plastic.