Super-Resolution Optical Microscopy

As the electronics and data storage industries build devices on ever decreasing scales, they need advanced imaging techniques that allow them to examine structures less than 10 nanometers in size for manufacturing and quality control. Nanoscale imaging is mostly done today using expensive electron microscopy. There is an alternative approach that uses high resolution optical nanoscale imaging to acquire information nanopixel by nanopixel, but this technique is very time-consuming. NIST researchers are developing a new technique that can create a high resolution optical image by acquiring information from a large portion of a device’s surface at once by using patterned masks instead of a single sharp optical probe in close proximity to the sample surface.

By making device features smaller, the semiconductor and hard disk industries can build faster electronic circuits and higher capacity data storage systems. But to ensure that existing devices perform reliably and to develop still better systems, these industries need to be able to examine their products quickly and at high resolution. Traditional optical microscopes cannot clearly resolve images of structures on scales smaller than the wavelength of visible light -- hundreds of nanometers -- so new techniques are needed. 

One possibility is to use electron or ion microscopy, which can probe scales much smaller than the wavelength of light, but these methods are costly and require the sample to be placed under vacuum. Another option is near-field optical microscopy. In this technique, light passes through a metallized optical fiber and emerges through a 50 nm opening that is brought within nanometers of the surface of interest. By restricting the light to a space much smaller than its wavelength, information on subwavelength scales can be obtained. A complete image of the surface must be built up one tiny pixel at a time, however, making the method slow and expensive.

In this project, NIST researchers are designing a new near-field optical microscopy system that can collect light simultaneously over a larger area. Instead of a single aperture for the light, the system would use a series of patterned masks to break the light into multiple small components covering a surface in a patchwork way. In one scenario, a complete sample surface would be divided into nanoscale regions, or pixels, and each mask would block about half of these pixels, allowing light from the remaining uncovered pixels to be collected by a detector.  A set of masks, each blocking different pixels, will produce a set of distinct detector readings. Another possibility is to move a single mask to different positions above the surface, blocking different pixels for each reading. In either case, mathematical algorithms process the light recorded in each set of exposures to calculate the brightness of all the pixels in the image.

Currently, project researchers are working on a proof-of-concept system covering 100 pixels by 100 pixels, where each pixel is a few tens of nanometers in diameter. Beyond demonstrating that the technique works, this system will allow the researchers to investigate the number and patterns of masks needed to create high resolution composite images. Eventually, the system will be scaled to larger areas.

Because light from distinct larger areas of the sample can be collected simultaneously, the technology could allow several detectors to gather information at either the same or different wavelengths, increasing the method’s speed and versatility.