Degree Type


Date of Award


Degree Name

Doctor of Philosophy


Mechanical Engineering

First Advisor

Sriram Sundararajan


The rapid growth of applications of lab-on-chip and bio-MEMES (Micro-electro-mechanical systems), arranging from cell culture and drug delivery to the microscale energy harvesting, microelectronics cooling, requires an understanding of microfluidic flow. Surface roughness, even at micro and nano scale, both deterministic and random, is considered an important parameter to affect fluid flow at the microscale. Over the past several decades, a variety of studies have been carried out on the impact of surface roughness on microfluidic flow behavior, including pressure drop, friction factor, heat transfer, laminar-turbulent transition, etc. However, most of the current studies focus on the large, deterministic roughness with patterned shapes formed by micromachining or microfabrication techniques. The effect of micro/nano scale random roughness on micro fluid flow remains relatively unexplored. As the size of microfluidic chip shrinks, the effect of realistic surface roughness on the microchannel walls cannot be simply ignored. In this work, the effect of realistic random surface roughness on fluid flow at the microscale was investigated through experiment and numerical simulation.

Methods for roughness generation were explored. Chemical wet etching was used to vary the surface roughness on glass substrate. The evolution of roughness with etching parameters and the surface distribution was studied to provide a better understanding of random roughness characterization. In addition to this study, a hybrid surface tailoring process combining particle masking and reactive ion etching was developed to generate and control roughness on quartz substrate. A mathematical model was built to predict the roughness parameters, which was validated by the Atomic force microscopy (AFM) measurement. It was proved mathematically that the process is able to produce a random surface with controllable amplitude and spatial parameters.

Based on the surface tailoring techniques, microfluidic devices were fabricated with desired surface roughness inside. This was made possible through incorporating the chemical wet etching treatment to the microfluidic device fabrication. The device is made from PDMS and glass substrate using soft lithography. Different levels of roughness were generated on the glass substrates for the study.

Micro particle image velocimetry (microPIV) was utilized as an analysis tool for the microfluidic flow characterization inside rough channels. Measurement was carried out on both Newtonian fluid (deionized water) and non-Newtonian fluid (dilute xanthan gum solution) at low flow rate condition. The roughness data obtained from optical profilometry was integrated with Direct Numerical Simulation (DNS) to resolve the effect of the realistic nanoscale roughness. This work demonstrated that realistic random surface roughness, even at the nanoscale, has a measurable effect on microfluidic flow. This effect can be characterized through microPIV measurement and full-scale coupled experimental-computational analysis. Particularly, this study provides promising potential of surface roughness harness for respective lab-on-chip applications, such as chemical mixing, cell manipulation and drug delivery, etc.

Copyright Owner

Jing Ren



File Format


File Size

173 pages

Included in

Engineering Commons