To provide info regarding some of this discussion:
Here is a very basic handout on the science of soap - ScienceofSoap.pdf. CMC is the surfactant concentration when the surface tension of the water decreases no further and the surfactant begins to produce micelles - geometric structures and it's the micelles that provide the detergency. Each type of surfactant, such as anionic and nonionic has different functions. Nonionic is known for its ability to solubilize oils which is why you will see these in cosmetics, they are used to blend the essential oils. Anionics are the backbone of the dish detergents such a Dawn but their CMC is much higher than nonionic and use requires much high concentrations. The rule of thumb for using surfactants for detergency is to use at least 2-3X the CMC.
While UT cleaning can easily exceed that of vacuum-RCM using standard record cleaning products, consider that vacuum-RCM is nothing more than machine assisted (drying) manual cleaning. Pre-clean with a good brush, good technique with a neutral-pH anionic detergent such as Alconox Liquinox used at 1.0%, rinsed, final clean with 0.5% nonionic surfactant only, rinse, dry will achieve very close to UT.
As far a decrease in CMC with temperature - this article shows it for the Polysorbate family Effect of Temperature on the Critical Micelle Concentration and Micellization Thermodynamic of Nonionic Surfactants: Polyoxyethylene Sorbitan Fatty Acid Esters and between room temperature 25C and 35C is not that much. The benefit for UT cleaning is that at the operating temp 35C a CMC concentration at 25C heated to 35C may add some detergency depending on the specific micelle development, but its marginal at best.
High powered tabletop UT tanks like the Elmasonic P-series will heat the water in use. A common small tank is the Elmasonic P60H P-series-P60H-brochure.pdf. I have worked with people using the P60H (and larger P120H) and for those doing serial type cleaning we install a recirculated pump filter-radiator to cool the tank. The reason you do not see it with the inexpensive Chinese units is that they are not very powerful. Although they may be advertised at 160-180W (three 60W transducers) that is very optimistic. People have measured the power draw (UT only) with a watt meter and measure about 110W. The Elmasonic P-series can damage a record - I have seen visible evidence; but it was at >40C, very slow (<0.4-rpm) and duration about 20-min.
As far as the "theoretical" temperatures developed during the cavitation-implosion event - that is shown in this computer simulation video - Inertial collapse of a single bubble near a solid surface - Bing video. However, the implosion energy is dependent on the UT frequency and the tank power. These are some of the basic design rules for UT tanks.
-The power to produce cavitation is proportional to the kHz, so a 120kHz UT needs more power than a 40kHz.
-For ultrasonic tanks, the bubble diameter is inversely proportional to the kHz, so a 40 kHz UT produces a large bubble than a 120kHz UT.
-The cavitation intensity is proportional to the bubble diameter and the tank power (watts/L) but there is a maximum power above which no addition cavitation intensity is obtained.
-The number of cavitation bubbles produced is proportional to kHz, so a 120kHz produces more bubbles than a 40kHz, but smaller bubbles.
-The smaller the tank volume, the more power that is required. It has to do with the ratio of the tank volume to its interior surface area.
-For lower kHz units (<60kHz), if the tank bath flow rate (from filtering or spinning) >50% of the tank volume per minute, cavitation intensity decreases.
Hope this adds to the conversation.