Related Experiment Video
Updated: Aug 12, 2026

Separating Beads and Cells in Multi-channel Microfluidic Devices Using Dielectrophoresis and Laminar Flow
Published on: February 4, 2011
Splitting dynamics of ferrofluid droplets inside a microfluidic T-junction using a pulse-width modulated magnetic
Mohamad Ali Bijarchi1, Mahdi Dizani1, Mohammadmahdi Honarmand1
1Department of Mechanical Engineering, Sharif University of Technology, Tehran, Iran. behshad@sharif.edu.
Abstract:
Micro-magnetofluidics offers a promising tool for better control over the ferrofluid droplet manipulation which has been vastly utilized in biomedical applications in recent years. In this study, the ferrofluid droplet splitting under an asymmetric Pulse-Width-Modulated (PWM) magnetic field in a T-junction is numerically investigated using a finite volume method and VOF two-phase model. By utilizing the PWM magnetic field, two novel regimes of ferrofluid droplet splitting named as Flowing through the Same Branch (FSB) and Double Splitting (DS) have been observed for the first time. In the FSB regime, the daughter droplets move out of the same microchannel outlet, and in the DS regime, the droplet splitting occurs two times which results in generating three daughter droplets. The main problem related to the asymmetric droplet splitting under a steady magnetic field is daughter droplet trapping. By using a PWM magnetic field, this issue is resolved and the trapped/escaped regions are obtained in terms of the duty cycle and dimensionless magnetic field frequency. The effects of six important dimensionless parameters on the splitting ratio, including magnetic Bond number, duty cycle, dimensionless magnetic field frequency, capillary number, dimensionless mother droplet length, and dimensionless dipole position are investigated. The results showed that the splitting ratio increases with increasing magnetic Bond number or duty cycle, or decreasing the dimensionless magnetic field frequency. Eventually, a correlation is offered for the splitting ratio based on the dimensionless variables with an average relative error of 2.67%.

