In a new twist on the idea of a flexible microscope, researchers at California Institute of Technology in Pasadena have constructed a functioning microscope out of rubber. Not only does the instrument work, but it also is small enough to be helpful in such applications as handheld diagnostic equipment. Using the rubbery elastomer polydimethylsiloxane (PDMS), the group fabricated a microscope that is only 10 cm long, complete with solid immersion lenses, and used it to image fluorescence in cells. A solid immersion lens is one example of an aberration-free spherical lens. A more familiar version is formed by cutting through the middle of a sphere, creating a hemispherical lens that enhances the numerical aperture of an optical system by the index of refraction of the lens material. Researchers have developed a way to easily produce solid immersion lenses using a rubbery elastomer for both the mold and the lens (left). The lenses can be integrated as imaging elements above a microfluidic channel (right). Images courtesy of Stephen R. Quake and California Institute of Technology. In contrast, a solid immersion lens is formed by slicing a sphere, with the cut made past the midpoint by a distance equal to the sphere’s radius divided by the material’s index of refraction. The result is a sphere missing a chunk at the bottom. Unlike a hemispherical lens, a solid immersion lens enhances an optical system’s numerical aperture by the square of its index of refraction. “The solid immersion lens provides extremely good light-collection ability in a compact optic, and integration with microfluidics allows one to flow objects past the [lens] to perform cytometry,” said Stephen R. Quake, a member of the institute’s research team who is now a professor of bioengineering at Stanford University in California. The adoption of solid immersion lenses has been held back, however, because they are difficult to make — particularly the undercut features found in the overhang at the bottom of each lens — which require labor-intensive hand grinding and polishing to achieve high-performance lenses. As described in the April 24 issue of Applied Physics Letters, in making their rubber lenses, the researchers exploited the physical and optical characteristics of PDMS. Optically, the material has a refractive index of 1.41 and is transparent, and therefore it can function as an optical element. Physically, PDMS is not rigid, and the researchers took advantage of this as well. “Our insight here was that the flexibility of PDMS would allow us to make the highly undercut structures that are needed for solid immersion lenses,” Quake said. The scientists developed a rubber microscope consisting of compound solid immersion lenses manufactured using 395-μm-radius microspheres (A) and other optical elements integrated with a microfluidic chip on one end and a CMOS sensor on the other (B). Only 10 cm long, the microscope imaged two E. coli cells expressing GFP (C). Art and image courtesy of the American Institute of Physics. The ability of elastomers to stretch and return to a previous shape without damage has been exploited to make large solid immersion lenses by embedding stainless steel ball bearings in PDMS, with the top of the ball extending out of the elastomer. Because PDMS stretches, the ball can be removed without damaging the material. Left behind is a form that can be used as a mold for the lenses. Instead of steel bearings, the scientists used ruby microspheres from Edmund Optics of Barrington, N.J., to create the lenses. However, their attempts to extend the molding technique to microscopic dimensions ran into a problem: The surface tension of the PDMS caused it to creep up the side of the spheres. They overcame this by partially curing the PDMS — causing it to shrink back a controlled amount — before extricating the spheres. This process fixed the shape in a highly reproducible manner without trapping the spheres permanently in the elastomer. The researchers then freed each microsphere and finished the curing. In the next step, they poured another layer of PDMS into the mold, cured it and popped free the finished lens. The ruby spheres used as the initial positive mold do not have to be perfect; however, they cannot deviate too much from ideal because they are the basis for the lens quality. The wavelength of the source illumination to be used with the lens plays a part in determining the amount of acceptable imperfection of the sphere. “[The material] needs to be optically smooth, and PDMS can reproduce features much smaller than the wavelength of light,” Quake said. The investigators characterized the resulting lenses by measuring the amount of fluorescence from a latex bead as imaged with and without the solid immersion lenses. They used a microscope that had a numerical aperture of 0.63 without the lens and found that the fluorescence intensity ratio — with and without the lens — was 15.35 to 1, which implied that the numerical aperture with the lens would equal 1.25, close to the theoretical prediction of 1.26. Similar close agreement with predicted performance was found with lenses made with spheres ranging in size from 395 μm to 2 mm. Not content to make a rubber lens, they also constructed a rubber microscope, fabricating six lenses and employing 150-μm ruby microspheres to create two solid immersion lenses. According to Quake, the optics were all fabricated separately and then put into place. One of the lenses was placed atop a microfluidic chip, which also was constructed of PDMS. The other lens formed part of an optical arrangement that began with the first solid immersion lens, continued through the other molded lenses and ended in front of a CMOS sensor. The result was a 10-cm-long structure comprising a microscope and an integrated microfluidic flow cytometer. The only nonrubber parts of the system were the mounting frame, the blue LED light source, a filter to separate excitation from emission light, and the CMOS sensor. Because the device was a monolithic, integrated structure, it did not exhibit the light-robbing problems of more heterogeneous instruments. Using the rubber microscope, the researchers imaged E. coli cells expressing GFP. Without the solid immersion lenses, the microscope could not detect the 1 x 2-μm cells. With the lenses, which boosted the magnification by about a factor of two, the scientists successfully imaged the cells. The elastomer-molding technique could be used in various systems as a way to integrate optics for readout or other uses. PDMS is already used to make a number of microfluidic mechanisms, including valves, pumps, cell sorters and DNA-amplification devices. Adding integrated optics could expand their capabilities.