Lensless Fluorescence Imaging System to Measure Surface Sample Flow

Microfluidic systems offer portable sensing with ever increasing sensitivity, complexity, and control over the system¹.  Sensors with single molecule limits of detection have been demonstrated where the movement of particles within the microfluidic channel is the limiting factor for system performance.  This work presents a technique for accurately measuring the flow of microparticles at the surface of a CMOS sensor fed by a microfluidic network.  This technique can be used to develop microfluidic networks for improved sample flow.

Objectives

            Conventionally, flow of particles in microfluidic networks is imaged using a fluorescence microscope.  In streak imaging, a long exposure is taken and the paths of particles are viewed as long streaks in the image.  Using higher speed cameras, Particle Image Velocimetry (PIV) captures sequential images of particles in the channel so that flow patterns including the speed of particles can be determined by measuring the distance covered between successive frames ².  Both streak imaging and PIV do a good job of imaging particles in the middle of the channel, but are poor at discriminating between particles at the bottom of the channel that may interact with a sensor and those in the middle of the channel.

Total Internal Reflection Fluorescence Microscopy³(TIRFM) solves this problem by only imaging particles that are very close to the bottom surface.  Like streak imaging and PIV, it is limited to imaging in a microscope so microfluidic networks must be prepared on a glass slide.  This is fine for investigating generic properties of microfluidic networks, but is poorly suited in optimizing a system, which must contend with the challenges of building microfluidics on a CMOS integrated circuit. Our solution to this problem was to develop a system with microfluidics fabricated on CMOS that can track samples and determine the flow rate.

New Results

Lensless fluorescence imaging is highly sensitive to the separation between the sample and the sensor⁴.  In the past this has been seen as a limitation, but we have recently shown that this effect can be used to measure the separation of flowing particles⁵.  In this work, we are demonstrating that this technique can be used specifically to optimize the design of microfluidic networks on CMOS chips.

The microfluidic chamber is laid directly on a bare CMOS imager to make the device. For this system, a PSA double sided tape was patterned using a laser and placed on the sensor. Input and output ports were punched on a slab of PDMS and placed on the tape to seal the device, as shown in Fig 1a. The sample was injected with a syringe pump and the chamber was illuminated with a hand held UV source. A video of the sample was captured by the sensor and analyzed in Matlab. One frame of the video is shown in Fig 1b. The code identifies the sample in each frame and tracks them as they flow on the sensor. The rate of each sample is calculated from its trace. The design of the chamber is independent of the device’s performance and any type of microfluidic chamber can be integrated with the device. The device does not consists of any optics or filters making it robust and low cost. An image of the complete system is shown in Fig 2.

Fig1: Lensless fluorescence imaging system. (A) On sensor microfluidic device (B) Frame captured by the system showing fluorescence beads inside the microfluidic device

Fig2: Complete lensless fluorescence imaging system with source, syringe pump, on-sensor microfluidic device, and source holder. The source holder shown above contains a low cost UV filter as the hand held source emitted low levels of visible light.

Conclusions

            Existing tools are well suited for studying the flow of liquids and particles in microfluidic networks built on glass slides.  The results from these experiments likely agree well with the behavior at the center of microfluidic channels where surface effects are small.  For CMOS sensing applications, however, it is the movement at the surface of the chip that is most important.  Lensless fluorescence image tracking is an ideal technique for measuring the flow across a true CMOS surface.

References

1   W Tian, E Finehout, “Microfluidics for Biological Applications”, Springer, New York, ch.1, pp. 1-5.

2   J. G. Santiago, S. T. Wereley, C. D. Meinhart, D. J. Beebe, R. J. Adrian, “A particle image velocimetry system for microfluidics”, Experiments in Fluids, 25, pp. 316-319, 1998

3   T Wazawa, M Ueda, “Total Internal Reflection Fluorescence Microscopy in Single Molecule Nanobioscience”, Adv Biochem Engin/Biotechnol, 95, 77-106, 2005

4 J Wu, G Zheng, L Lee, “Optical imaging techniques in microfluidics and their applications”, Lab on a Chip, 12, 3566-3575, 2012

5 A Shanmugam, C Salthouse, “Lensless Fluorescence Imaging with height Calculations”, Journal of Biomedical Optics, unpublished