Chip-scale Microscopy for Biological Imaging

Seung Ah Lee (2009-2013)

Optofluidic microscope and color imaging of malaria parasites

Optofluidic microscope (OFM) is a chip-scale imaging platform based on the combination of a low-cost CMOS image sensor and microfluidics. The image sensor captures shadow images of the sample as it flows through the microfluidic channel mounted on top of an image sensor. The shadow images have limited resolution (determined by the image sensor's pixel size). However, we can reconstruct a high-resolution image from a series of raw shadow images using the pixel super-resolution algorithm.  

Color imaging can be done with RGB illumination. We used switching LEDs to obtain three channel monochrome images which are combined into a single color image. 

Low res to high res image reconstruction using pixel super-resolution reconstruct

Multiframe shadow images are taken at high framerate as the cell flows in the microfluidic channel, so that each low res frame has sub-pixel shifts between each frame. Here is the direct shadow images of the red blood cells flowing through an optofluidic microscope(40x slowed).  

Color OFM scheme can be useful for imaging malaria infected red blood cells. Blood sample can be treated with Toluidine blue stain, which selectively stains Plasmodium parasites within the red blood cells into purple color. Pretreated sample can be injected into an OFM device and the resulting high-res images can be used for identifying malaria infection. 

Monitoring motile microorganisms on a chip 

ePetri is a digital petri-dish platform with imaging capabilities. We use a CMOS image sensor as a substrate for cell growth, which is monitored real-time via our lensless imaging techniques. In this work, we cultured motile microorganisms (euglena) on our chip-scale microscope and used the inherent motion of the cells for high-resolution imaging.  The system is very simple - a CMOS image sensor with a microchamber and a control PC. We can simply grow microorganisms in this platform and continuously observe the cells without invasion. Large field-of-view (5mm x 5mm) of the image sensor lets us monitor a large number of cells. 

Video 2. Live cell imaging of euglena culture

With simple image analysis based on particle tracking, we can extract some useful information about the viability of the cells - such as cell number, cell shape and motility. Such information provides quantified cell-viability under different culture environment and external stimuli. 

Fluorescence chip-scale microscope using silo-filter CMOS imager

Fluorescence imaging on a chip-scale microscope can be tricky because 1) fluorescence emission is incoherent and 2) we need a good filter to block out the excitation light. A thick absorptive emission filter between the sample and the sensor causes fluorescence signal to diverge and worsens the image resolution. To prevent resolution loss through the filter layer, we constructed a silo-filter structure with a high-aspect-ratio metal grid which confines the fluorescence emission to a single pixel of the CMOS image sensor. We have fabricated a silo-filter system on a commercial CMOS image sensor via SU8 photolithography and electroplating. As a result, the silo-filter sensor can perform large-area (~4mm x 4mm) fluorescence imaging with the resolution determined by the image sensor's pixel size (5.2um in this experiment). The fluorescence sample (Texas red and mCherry RFP) resting on top of the silo-filter sensor is illuminated with high-power LED (565nm) for excitation. High-resolution brightfield imaging is also possible with LED array illumination and pixel-super resolution algorithm. Altogether, we can construct a compact lensless fluorescence imaging system (10cm x 10cm x 10cm) for long-term monitoring of biological samples. Here are some time-lapse videos of mCherry-HEK cells cultured on our silo-filter sensor. (bright-field and fluorescence overlay) 

Comparison between uniform-filter and silo-filter structure (Sample : 2.5um fluorescent and 2um non-fluorescent beads)

Live cell imaging of HEK-mCherry cell division

We also built a smartphone-based lensless imaging device using sunlight illumination. It works by loading the sample on the surface of the image sensor on a smartphone's camera module and taking shadow images under sunlight (or any single light source, like a flashlight) while manually tilting the phone around the sun. We then perform pixel-super resolution image reconstruction of these low-resolution (resolution = 2*pixel size) images to obtain sub-micron resolution microscopy images over an ultra-wide field of view (4mm x 5mm). We use the back camera module of a smartphone by removing the lens module to reveal the image sensor pixels. It does not require any complex optical components (maybe just an IR filter for the sun), making it a truly portable and low-cost imaging device for field applications.

Image acquisition and reconstruction of our smartphone-microscope are all performed in the custom-built application (powered by OpenCV). The application measures the illumination angle in real-time as the user takes images by moving the device around the light source. Once the photos are taken, it performs image reconstruction based on pixel-super resolution algorithm and provides resolution-improved images of the sample. The figure above shows red blood cell images taken with our prototype device. 

With its portability and simple imaging principle, we believe that our device addresses microscopic imaging needs in resource-limited settings, such as field diagnostics of blood-borne and water-borne infections and environmental monitoring. See the microscope in action -

References