Lensless microscopy technology developed at the University of Connecticut, USA. The screen panels depict the setup; wide-field-of-view, hi-res lensless imaging of a cell culture; and zoomed in views of the same culture. [Image: Sean Flynn, UConn Photo]
Most microscopy systems present users with a choice: you can have a wide field of view, or you can see the most minuscule details possible. Researchers at a U.S. university have developed a lensless on-chip microscope that delivers both a broad and detailed glimpse of tissue samples and other biological specimens (Lab Chip, doi: 10.1039/C9LC01027K).
Instead of an objective lens, the ptychographic microscope developed at the University of Connecticut (UConn), USA, uses a randomly moving diffuser positioned only about 1 mm above the sample being imaged. The resulting ultra-high Fresnel number of 50,000 allows the system to record the position of the diffuser directly from the raw images it captures.
Simplifying ptychography
For the past 20 years, but especially in the last decade, scientists have explored multiple types of lensless processes to produce high-resolution images for lab-on-a-chip applications. Many of these approaches involve ptychography, a computational method for deriving images from interference patterns generated from a function—a screen or aperture—moving across the field of view. Ptychography has been used to boost the resolution limit of electron microscopes, but the technique also works with visible and shorter wavelengths of light.
The UConn biomedical-engineering researchers created a diffuser by sticking polystyrene beads, roughly 1 μm in diameter, onto a standard cover slip for a microscope slide and placed it about 1 mm away from the imaging target. They illuminated the target with a 532-nm-wavelength laser diode. Simple, off-the-shelf microcontrollers—the kind marketed to electronics hobbyists—moved the diffuser randomly in the x-y plane. The raw images were 1024×1024 pixels.
Cutting down processing time
According to the researchers, the computational phase-retrieval process that generated the images from the raw data converged within two or three iterations of the algorithm. Other ptychographic experiments have required hundreds of iterations to generate images. The high Fresnel number—the area of the aperture divided by the product of the illumination wavelength and the distance from the target to the diffuser—may be responsible for the computational speedup.
The UConn engineers took pictures of several types of materials, from standard test patterns to a thin slice of mouse kidney to a comparatively thick slab of potato. They also used an automated cell-segmentation routine to count the number of yeast cells in the imaging field of view, 6.4 mm by 4.6 mm.