Ultrasound enhanced micromixing: the effect of frequency and probe tip surface area

1. Introduction Ultrasound is often used to significantly improve mixing on a molecular scale (micromixing) [1-3]. This micromixing impacts the reaction rate and selectivity of several chemical reactions like precipitation, neutralization, combustion or polymerization reactions. The Villermaux-Dushman reaction is a widely used technique to study the micromixing efficiency. This research shows, however, that ultrasound can interfere with this reaction by sonolysis of potassium iodide, which is excessively available in the Villermaux-Dushman solution, into triiodide ions. Hence, the mixing performance will be underestimated when no corrective actions are used. Furthermore, the ultrasonic parameters which optimize micromixing are not clear. Some authors claim best micromixing at high frequencies while others observed it at lower ones. Also the effect of the probe tip surface area on micromixing was not studied before [1–3]. This research contributes to a better understanding of ultrasound enhanced micromixing by investigating the effect of both the frequency and probe tip surface area on micromixing in a single reactor geometry, at similar power dissipation, stirring speed and cavitation bubble type. Finally, these results are used to evaluate and improve the existing ultrasound micromixing model of Rahimi et al. [2]. 2. Materials and methods A single batch reactor with an ultrasonic horn (Hielscher UP50H) at 30 kHz and ultrasonic transducers at 41, 94, 165, 577, 860 and 1135 kHz were used. The probe tip surface area was varied by using probes with different diameters on the ultrasonic horn. Calorimetric power calibrations were performed to assure that the power inside the reactor was constant at 47 W/kg for all different sources. First, the sonolysis effect of ultrasound was studied by sonication of solely the Villermaux-Dushman buffered solution. The I3- concentration was defined by UV/VIS spectrophotometry (Shimadzu UV-1601) at 353 nm. Next, the Villermaux-Dushman reaction was performed, the I3- concentration measured and the segregation indices calculated. Finally, the dominant bubble type was defined with sonoluminescence experiments to define which cavitation bubble type was dominant for each experiment. 3. Results and discussion The sonolysis experiments show clearly I3- production in the buffer solution by sonication. This effect is more prominent for longer sonication times and frequencies of 577 kHz or above. Frequencies below 165 kHz do not show any significant effect while it is maximized at 577 kHz. Possible explanations and corrective measures will be discussed at the conference. Secondly, segregation indices were studied, with the corrective actions in place, in function of the applied frequency as shown in Figure 1. Higher frequencies resulted in lower segregation indices and hence better micromixing for stable bubbles. A similar level of micromixing was observed for stable bubbles at 1135 kHz and transient ones at 30 kHz. Furthermore, comparable segregation indices were observed during the tests with different probe tip surface areas. This indicates that the probe tip diameter has an insignificant effect on micromixing. Finally, these results were used to investigate the ultrasound micromixing model of Rahimi et al. [2]. This model was not able to predict micromixing accurately and it will be shown at the conference that the addition of the frequency, as an additional parameter, is needed to improve the simulations. 4. Conclusion It was shown that ultrasound can interfere with the segregation indices obtained by the Villermaux-Dushman reaction through sonolysis. Therefore corrective actions are needed, especially at frequencies around 570 kHz. Next, best micromixing was obtained at the highest frequency for stable bubbles. Also, it was shown that the diameter of the probe had no effect on the degree of micromixing within the tested conditions. Finally, the power law model of Rahimi et al. was tested and improvement of the model, by inclusion of the frequency as an additional parameter besides the energy dissipation, is needed.status: publishe