Microfluidic on CMOS with Laser Cut Adhesive Tape

Wednesday, May 14, 2014: 10:20
Gilchrist, Ground Level (Hilton Orlando Bonnet Creek)
A. Shanmugam and C. D. Salthouse (University of Massachusetts, Amherst)
CMOS integrated circuits are an ideal platform for microfluidic-based sensing because they offer small feature size at low cost and can incorporate control and sensing circuitry1.  Fabricating microfluidic devices on top of CMOS presents two challenges: bonding the channel to the CMOS surface and working with the small area that is usually available. This paper presents and compares two techniques for building microfluidic networks on CMOS.


The goal of this project is develop an efficient way of prototyping microfluidic devices on CMOS that can be translated into production.

New Results

Poly(dimethylsiloxane) (PDMS) has been widely used for making microfluidic devices on glass substrates1. Conventional microfluidic devices are formed by pouring uncured PDMS over a master. The cured slab is the peeled off from the master to punch ports and laid on a glass slide. The chamber is bonded by exposing the slide to oxygen plasma2. This technique works when the sample is analyzed under a microscope, but cannot be used on a CMOS chip. 

In this work, this technique was modified by using uncured PDMS as a glue to bond the cured PDMS chambers to the CMOS. To make the device, the cured chamber with punched ports was placed on the sensor and held down to hold it in place for a few seconds. Uncured PDMS was deposited on the outer edges of the channel so that the bottom of the channel is a clean CMOS sensor. The PDMS was then flash cured with a hot air gun. Tubing was then inserted at the ports before imaging. This set up was capable of imaging fluorescence sample but fabrication could not be consistently repeated. Images of the master, chamber, and the imager is shown in Fig 1.  

For the second method, Pressure Sensitive Adhesive (PSA) tape was used to make the microfluidic device3. For testing, we used a CO2 laser to cut the microfluidic channels out of the tape. The tape defines a channel which is 46.26 µm. In production, a die cutting process could be used. For conventional microfluidic device, the tape was stuck on a glass slide and the top layer was formed by placing a piece of clear acrylic plastic or a slab of PDMS. The thickness of the top layer serves to hold the inlet and outlet tubing. We have observed that PDMS forms a tighter seal around the tubing than acrylic plastic.

When working on CMOS chips, the tape was stuck to the sensor and then the punched PDMS slab was placed on top to seal the chamber. Apart from occasional leaks at the input and output ports, acrylic plastic also filtered most of the UV light during fluorescence imaging. An image of the PDMS, PSA tape, and the imager is shown in Fig 2.

Figure1: Microfluidic device made from PDMS (A) Master made from copper plates (B) Cured PDMS microfluidic chamber (C) Sealed chamber on CMOS imager

Figure2: Microfluidic device made from PSA tape (A) Top layer made of PDMS (B) Patterned PSA tape (C) Sealed chamber on CMOS imager


                This microfluidic fabrication technique is repeatable and can be used to make microfluidic chambers with any dimensions and patterns. These on-sensor lensless imaging devices can be used for sample detection during fluorescence imaging and white light imaging, studying sample interaction, fluid dynamics, and for implementation of a low cost, portable, flow cytometer.


1   D Psaltis, S Quake, C Yang, “Development optofluidic technology through the fusion of microfluidics and optics”, Nature, Vol 442, 2006

2   J Friend, A Yeo, “Fabrication of microfluidic device using polydimethylsiloxane”, Biomicrofluidics, 4, 2010

3   T Merian, F He, H Yan, D Chu, J Talbert, J Goddard, S Nugen, “Development and surface characterization of an electrowetting valve for capillary-driven microfluidics”, Colloids and Surfaces A: Physiochem. Eng Aspects 414, pp. 251-258, 2012