Matter Assembly Computation Lab
CU Boulder, College of Engineering
An OpenFOAM Framework to Model Thermal Bubble-Driven Micro-Pumps
Summary
Inertial pumps hold great promise to democratize microfluidics by integrating scalable, mass producible micro-pumps with no moving parts directly in microfluidic channels. However, it is difficult to simulate these micro-pumps due to the multiphysics coupling of Joule heating, thermal bubble nucleation, phase change, and multiphase flow. As such, most simulation approaches simplify the physics by neglecting Joule heating, nucleation, and phase change effects as done in this study. To date, there are no readily available, reduced physics open-source modeling tools that can resolve both pre-collapse (defined as when the bubble is expanding and collapsing) and post-collapse (defined as when the bubble has re-dissolved back into the subcooled fluid) bubble and flow dynamics. In this work, an OpenFOAM framework for modeling thermal bubble-driven micro-pumps is presented, validated, and applied. We envision the developed OpenFOAM framework as an open-source CFD toolkit for microfluidic designers to simulate devices with thermal bubble-driven micro-pumps.
Main Learnings
-
OpenFOAM compressibleInterFoam solver is modified with computation heuristics to successfully model both pre-collapse and post-collapse bubble and flow physics of thermal bubble-driven micro-pumps
-
The shape of the transient velocity profile during a pump cycle varies substantially from theoretical Poiseuille flow during pre-collapse but is within 25% of the theoretical flow profile during post-collapse. We find that this deviation is due to flow never becoming fully developed during each pump cycle
-
OpenFOAM can be used as an open-source alternative to commercial CFD codes for simulation of thermal bubble-driven micro-pumps
Model Heuristics
OpenFOAM's compressibleInterFoam solver cannot account for mass transfer which is a crucial compnent of vapor bubble dissolution upon vapor bubble collapse. As such, physics-based heuristics are used to simulate vapor bubble collapse. Primarily, the following model assumptions are made and implemented using model heuristics:
-
Thermal bubble nucleation is represented by an initial vapor layer of set thickness, pressure, and temperature with area Wr x Lr, where Wr and Lr are the resistor width and length, respectively
-
At the maximum expansion extent, all vapor cells are assigned a sub-atmospheric collapse pressure, p = 0.03po, where po is atmospheric pressure, to allow the vapor bubble to uniformly collapse
-
When the vapor bubble short leg collides with that of the long leg, all vapor cells are turned back into liquid to simulate vapor “dissolution” upon bubble collapse.​
.jpg)
Model Validation
The OpenFOAM model was validated against FLOW-3D and experimental data from literature and found to be in agreement. The net volume displaced as a function of time was used to compare OpenFOAM simulation data to that from literature and FLOW-3D. A mesh analysis study verified that 1 um mesh elements were sufficient for accuracy and convergence.
Transient Flow Profile Analysis
Past work has assumed that the flow profile from thermal bubble-driven micro-pumps can be modeled using Poiseuille flow in a rectangular channel. However, there has been a lack of a detailed analysis validating this assumption. In this work, we utilize the OpenFOAM model and FLOW-3D model for thermal bubble-driven micro-pumps to understand the transient velocity profile. It was found that the velocity profile deviates substantially from fully developed Poiseuille flow during the pre-collapse stage of the pump cycle but closely resembles fully developed Poiseuille flow during the post-collapse stage. Furthermore, it was found that flow does not become fully developed over the pump cycle; as such, deviations between theoretical Poiseuille flow and the transient velocity profile should be expected when using thermal bubble-driven micro-pumps.
