A Two-Mirror Faceted Projection System for EUV Lithography

The modern world runs on tiny silicon chips, the invisible engines inside everything from smartphones to the systems guiding global internet traffic. To make these chips faster and more powerful, engineers must shrink the transistors that power them, packing billions onto a single grain of sand. This process of miniaturization relies on a technique called lithography, which acts like a high-precision camera, projecting a pattern from a master template onto a silicon wafer to carve out the circuitry. The smaller the features the camera can print, the smaller the transistors can be. For decades, the industry has pushed the limits of light, moving from visible light to ultraviolet, and finally to extreme ultraviolet light, a wavelength so short it is measured in nanometers. However, as the demand for even smaller features grows, the traditional tools used to print these patterns are hitting a wall. The mirrors required to focus this light are so complex and numerous that they absorb most of the energy, making the process inefficient and incredibly expensive.

A team of researchers from the University of Nizhny Novgorod in Russia has proposed a radically different way to solve this problem. Instead of using a long chain of curved mirrors to focus the light, they designed a system that uses just two flat surfaces, each covered in thousands of tiny, flat tiles. In their simulations, this new arrangement successfully projected a pattern onto a silicon wafer with a level of detail that current machines struggle to achieve, all while keeping the light much brighter. The researchers did not build a physical machine for this study; rather, they used powerful computer models to prove that the idea works in theory. Their work suggests a path toward printing features smaller than six nanometers, a size that would be impossible with the current generation of equipment.

The core of the problem lies in how light behaves when it is extremely short. To print tiny patterns, the light must be focused with extreme precision, which usually requires a series of six to ten curved mirrors. Each time the light bounces off a mirror, a significant portion of its energy is absorbed or lost. By the time the light reaches the silicon wafer, less than fifteen percent of the original power remains. This loss forces the machines to use massive amounts of electricity to generate enough light, driving up costs and limiting how small the features can be. The researchers asked a simple question: could they achieve the same result with far fewer reflections? Their answer was to abandon the idea of a single, continuous mirror surface in favor of a "faceted" approach.

Imagine a mirror not as a single sheet of glass, but as a mosaic made of thousands of tiny, flat tiles. In the system proposed by the researchers, the light scattered by the mask is broken down into many distinct beams, each carrying a specific piece of the pattern. Instead of trying to bend all these beams with a single curved surface, the system assigns each beam to its own dedicated pair of flat tiles. The first set of tiles catches the beam and sends it straight down, while the second set catches it again and aims it precisely at the wafer. Because the light only bounces twice, the system retains about half of the light's power, a massive improvement over the current technology. This efficiency is crucial because it means the machine could potentially print much smaller features without needing a light source that is impossibly powerful.

The geometry of this system is intricate. The researchers had to calculate the exact angle and position for every single tile to ensure that all the different beams arrived at the wafer at the same time and with the correct focus. If the timing were off, the pattern would blur. They designed the system so that the distance each beam travels is identical, regardless of which tile it hits. This ensures that the waves of light reinforce each other to create a sharp image. They also had to account for the fact that the light hits the tiles at different angles. To handle this, they designed a special coating for each tile, made of alternating layers of different metals. These layers act like a filter, reflecting the specific color of light needed for that particular tile while absorbing the rest. By customizing the coating for every single tile, they ensured that the system works efficiently across the entire range of angles required.

The team tested their design using two different types of light wavelengths, one at 13.5 nanometers and another at 11.2 nanometers. They simulated the printing of patterns that included lines as thin as six nanometers and isolated peaks with a width of about 5.4 nanometers. In the computer models, these tiny features remained sharp and distinct even when the focus was slightly off, a sign that the system is robust. The simulations showed that the system could achieve a numerical aperture, a measure of how much detail the lens can capture, of nearly 1.0. This is a value that is theoretically possible but has never been reached in a practical machine, as current designs are limited to about 0.55.

While the results are promising, the researchers are careful to note that this is a theoretical breakthrough. The system relies on a two-dimensional pattern, and extending it to the complex, three-dimensional shapes found on real computer chips presents new challenges. Furthermore, the sheer number of tiles—over two thousand in the 3D version—means that manufacturing and aligning them would be a monumental engineering task. The researchers acknowledge that building such a machine would require solving problems related to the precise placement of each tile and the stability of the coatings. However, their work proves that the fundamental physics allows for a system with only two reflections, a concept that could eventually lead to a new generation of lithography tools.

The implications of this work extend beyond just making smaller chips. By demonstrating that a system with fewer mirrors can achieve higher resolution, the researchers have opened a door to a different way of thinking about optical design. The traditional approach has been to add more mirrors to correct for errors, but this new method suggests that the solution might be to simplify the path of the light itself. The ability to retain more light energy means that future machines could be more efficient and potentially less expensive to operate. While the road from a computer simulation to a working factory machine is long and filled with obstacles, this study provides a clear map of a path that was previously thought to be blocked. It shows that with enough ingenuity, the laws of physics can be bent to allow us to see and build things at a scale that was once considered impossible.