“SHARP” microscope will lead to next-gen microchips

Ashley N. Paddock, ashley.paddock@photonics.com

A new microscope called SHARP is poised to advance photolithography technology through higher resolution, speed, illumination uniformity and coherence control than existing instruments provide. The extreme-ultraviolet (EUV) photomask-imaging microscope will push research toward semiconductors with features as small as 13.5 nm.

The name stands for Semiconductor High-NA Actinic Reticle Review Project, and it refers to a $4.1 million, 1 1/2-year-long collaboration between scientists at Lawrence Berkeley National Laboratory and semiconductor manufacturers. The partnership is led by Kenneth Goldberg, deputy director of the Center for X-Ray Optics in Berkeley Lab’s Materials Science Div.

Kenneth Goldberg is seen in the reflective coating of a photolithography mask that he is about to measure at the Advanced Light Source’s 11.3.2 beamline. Inset at lower right shows a mask’s extreme-ultraviolet (EUV) absorbing layer, printed on a 6-in.-sq glass coated with multiple layers of molybdenum and silicon to reflect unwanted EUV. The patterned layer represents one level of a working microprocessor or memory chip, which may have 20 or more such levels. Courtesy of Lawrence Berkeley National Laboratory.

Initially, the instrument will be used in parallel with operations at an existing microscope on beamline 11.3.2 of Berkeley Lab’s Advanced Light Source (ALS). But by the last quarter of 2012, SHARP will replace the beamline’s aging facilities.

Within a few years, semiconductor devices will be measured in dimensions of 16, 11 or 8 nm. To mass-produce them, industry is pushing a photolithography process that uses EUV light with a wavelength of just 13.5 nm. At such a short wavelength, it will be possible to print and image circuit patterns at nanometer-length scales, Goldberg said.

The 8-year-old microscope at beamline 11.3.2, called the AIT (Actinic Inspection Tool), has unique imaging capabilities, but the fast-moving nature of semiconductor technology means that its future is limited. With its higher-performance specifications, SHARP should exceed the AIT’s performance and enable forward-looking research years before commercial tools become available.

At top left are programmed pattern defects, including the large protrusion in the center, developed to evaluate printing sensitivity. An image of the defects created with the Actinic Inspection Tool (AIT) is shown at center left. The final stage of the AIT’s beam path is shown bottom left: The synchrotron’s EUV beam enters from the top, reflects from the mask, is focused by an array of zone plates and then reflects from a turning mirror to an EUV-sensitive camera (not shown). The more efficient SHARP beam path at right removes the turning mirror and a window from the array of zone plates to increase brightness.

Like AIT, SHARP is also an actinic microscope because it relies on the same EUV wavelengths used in production. Thus, the new EUV microscope will enable semiconductor company researchers to better evaluate defects and repair strategies, mask materials and architectures, and advanced pattern features.

A special feature of the new microscope will be illumination coherence control. The ALS produces an EUV beam with laserlike coherence, ideal for many experiments. For microscopy, however, the image resolution can be improved by a factor of two by carefully re-engineering the illumination into a state called partial coherence. Microscopists have recognized the importance of partial coherence for years, and the synchrotron community is now catching up.

Working with EUV light is difficult because every material absorbs it strongly, Goldberg explained. Instead of glass lenses, EUV optical systems must rely on specialized mirrors with atomic-scale smoothness, topped by multilayer coatings for high reflectivity. To maintain efficiency, the entire optical system must be placed in a high-vacuum environment.

Minute imperfections or tiny particles of dust, if not found and cleaned or fixed, ultimately cause semiconductor chips to fail. Goldberg and his team have shown that defects and patterns can appear very different when viewed with non-EUV inspection tools such as electron microscopes, making EUV microscopy essential for the development of EUV photolithography masks. Only in this way can damaging particles and other defects be identified reliably.