Dispersion of nanoparticles is precisely why we got the bath (CNT's to be exact). It seems to be the least destructive method for CNT dispersion.
I haven't had the time I would like to do much work with spontaneous gold particle formation via ultrasonics. I read a paper, and I continuously kick myself for losing it, that detailed a method for formation in pentanol I believe.
I use the US bath to aid inter-micellular interaction. This allows me to produce two separate solutions of a surfactant at CMC (or Heptane and Propanol w/CTAB) one of which has an aqueous solution of HAuCl4, and the other containing the reducing agent/s. I've used citrates, and maltodextrin. My hope, and I've yet to get any samples under the AFM to verrify this, is that the ultrasonication will insure a more uniform micelle-micelle mixing, resulting in a more monodisperse colloid. This process can be done without the US bath though. Simply shaking the container containing both parts is enough to produce.
I've yet to work out the bugs in this method. My ultimate goal, is to find method that would allow a continuous production process, preferably in a variety of solvents. This is the reason i'm looking into ultrasonics.