Experimental Investigation on the Dynamics of On-Demand Ferrofluid Drop Formation under a Pulse-Width-Modulated Nonuniform Magnetic Field.
Mohamad Ali BijarchiMohammad Behshad ShafiiPublished in: Langmuir : the ACS journal of surfaces and colloids (2020)
Drop formation has been the focus of many studies because of its vast application in biomedicine and engineering, as well as its rich underlying physics. Applying a magnetic force on ferrofluids can provide more control over the formation process of the droplet. In this study, a time-dependent, nonuniform magnetic field was used for the formation of ferrofluid droplets using a nozzle. A pulse-width-modulation signal (PWM) was utilized to induce the time-dependent magnetic field, and a drop-on-demand system was designed using the capability of the PWM magnetic field. Three kinds of drop formation regimes under the PWM magnetic field were seen. Also, a new droplet generation regime was observed in which the drop is formed while it bounces back to the nozzle during the off-time period of the magnetic excitation. As compared to other techniques, the main advantage of droplet formation in this regime is that there will be no satellite droplet during the pinch-off. The regime map of drop formation based on the magnetic Bond number and the dimensionless induced frequency was obtained. Also, the effect of the duty cycle, the induced frequency, the magnetic induction, and the vertical interval between the coil's top surface and the nozzle on the drop formation evolution, the equivalent diameter of the droplets, the frequency of droplet formation, and the pulses that are necessary to form a drop was studied. Additionally, it was illustrated that by prolonging the duty cycle, the magnetic induction, or by decreasing the induced frequency, the equivalent diameter of the drop and the pulses that are necessary to form a drop reduce, while the frequency of drop formation increases. Eventually, a correlation for predicting the nondimensionalized diameter of the droplet, based on dimensionless variables, was presented with a maximum relative error of 8.1% and an average relative error equal to 2.2%.