Matching Tin Plasma Emission to Molybdenum-Silicon Bragg Mirrors: Why Extreme Ultraviolet Lithography Operates at 13.5 Nanometers

Every high-end smartphone and artificial intelligence accelerator manufactured today exists because of a single, highly specific physical alignment. The transistors inside them are carved by light measuring exactly 13.5 nanometers in wavelength, a dimension that allows foundries to print features just a few dozen atoms wide. This is the foundation of extreme ultraviolet lithography, the technology that rescued the semiconductor industry from physical stagnation.[4]

The shift to this wavelength was not a gradual step down from the 193-nanometer deep ultraviolet lasers that dominated the 2000s and 2010s. It was a sheer cliff. At 13.5 nanometers, light behaves entirely differently, absorbing into almost all known matter rather than passing cleanly through it.[4]

Foundries could not simply swap out a lightbulb to achieve this resolution. They had to reinvent the entire optical pipeline, replacing transmissive glass lenses with reflective mirrors operating inside a hard vacuum. The resulting machines, built exclusively by the Dutch firm ASML, cost upwards of $200 million each and require megawatts of power to operate.[5]

Yet, despite the massive engineering effort, the 13.5-nanometer figure itself was never a variable that designers could freely tune. It is a hard physical constraint. The entire extreme ultraviolet ecosystem is trapped at this specific wavelength because it is the only point where two independent physical phenomena—the emission spectrum of vaporized tin and the reflectivity of molybdenum and silicon—perfectly overlap.[5]

The Physics of the Tin Plasma Source

Generating light at such an extreme frequency requires conditions rarely found outside of a star. Traditional lasers cannot directly emit 13.5-nanometer photons with the power required for commercial chipmaking. Instead, the industry relies on a laser-produced plasma source, which uses a secondary material to generate the required glow.[3]

That material is pure tin. Inside the machine's source chamber, a generator fires microscopic droplets of molten tin—each about 30 micrometers across—at a rate of 50,000 droplets per second. A high-power carbon dioxide laser then tracks and strikes these droplets in mid-air.

The strike happens in two distinct phases. A low-energy prepulse hits the spherical droplet first, flattening it into a disk to maximize its surface area. Microseconds later, a 30-kilowatt main pulse vaporizes the tin, stripping electrons from the atoms and creating a highly ionized plasma.

This plasma reaches temperatures of approximately 250,000 degrees Celsius, or 20 to 30 electron volts. At this extreme thermal state, the tin ions—specifically those stripped of eight to fourteen electrons—undergo atomic transitions that emit a brilliant burst of extreme ultraviolet radiation.[3]

Crucially, this emission is not a uniform white light. The multiply charged tin plasma ions act as the atomic sources of extreme ultraviolet radiation, peaking sharply near the 13.5-nanometer mark. Researchers have found that this specific plasma state offers a high conversion efficiency, turning the infrared laser's energy into the exact spectral band needed for lithography.[3]

Why Glass Fails at Extreme Wavelengths

Generating the light is only the first half of the problem; delivering it to the silicon wafer is the second. In older lithography systems, light passed through a series of highly polished quartz or fused silica lenses that focused the circuit pattern onto the chip.[5]

At 13.5 nanometers, that approach immediately fails. Photons carrying 92.4 electron volts of energy are highly ionizing, meaning they interact violently with the electrons in almost any material they touch. If you shine extreme ultraviolet light at a glass lens, the glass simply absorbs the energy and turns it into heat.

Even the air we breathe is opaque at this wavelength. A beam of 13.5-nanometer light traveling through a standard room will be entirely absorbed by atmospheric gases within millimeters. Consequently, the entire optical path inside an extreme ultraviolet scanner must be maintained in a near-perfect vacuum.

To manipulate the light without absorbing it, engineers had to abandon lenses entirely and switch to mirrors. But standard household mirrors, which rely on a single layer of silver or aluminum, also absorb extreme ultraviolet radiation. A new type of reflective surface was required, one that relied on wave interference rather than simple metallic reflection.[5]

This led the industry to the Bragg mirror, a complex optical structure built from alternating layers of different materials. By stacking materials with different refractive indices, a Bragg mirror forces the light waves reflecting off each boundary to constructively interfere, bouncing the beam back rather than absorbing it.[1]

The Molybdenum-Silicon Bragg Mirror

Finding the right materials for these alternating layers was a monumental materials science challenge. The layers had to be nearly transparent to extreme ultraviolet light, yet possess enough of a refractive difference to trigger the Bragg interference effect.[5]

The solution was a pairing of molybdenum and silicon. When deposited in alternating nanometer-thick layers, this combination creates a highly efficient reflector for soft X-rays and extreme ultraviolet light. The operational wavelength of 13.5 nanometers was selected precisely because periodic molybdenum-silicon Bragg multilayer coatings exhibit record near-normal reflectivities.[1]

A standard extreme ultraviolet mirror consists of 40 to 50 of these bilayers. Each layer must be deposited with sub-nanometer precision, as any deviation in thickness destroys the constructive interference. When properly aligned, the stack achieves a theoretical peak reflectivity of approximately 70 percent per mirror surface.[1]

However, this peak is incredibly narrow. The full width at half maximum bandwidth for these mirrors is only about 0.65 to 0.78 nanometers. If the incoming light deviates even slightly from the 13.5-nanometer target, the reflectivity plummets, and the mirror absorbs the energy instead of reflecting it.[2]

The physical stability of these mirrors is also a constant battle. At high operating temperatures, the molybdenum and silicon atoms tend to intermix at the boundaries, forming silicide compounds that degrade the mirror's optical properties. Manufacturers now engineer the interfaces with thin barrier layers, such as boron carbide, to prevent this diffusion and maintain the 70 percent reflectivity target.[2]

The 13.5-Nanometer Intersection

The entire architecture of modern semiconductor manufacturing rests on the overlap of these two distinct physical curves. The tin plasma emission spectrum provides a massive spike of photons at 13.5 nanometers, and the molybdenum-silicon Bragg mirror provides a narrow window of 70 percent reflectivity at that exact same wavelength.[5]

If tin emitted light at 11 nanometers, the molybdenum-silicon mirrors would absorb it. If the mirrors peaked at 15 nanometers, the tin plasma would not provide enough photons to expose the wafer. The 13.5-nanometer standard is the only commercially viable intersection of a high-efficiency plasma source and a high-reflectivity mirror.[5]

Other material pairings exist, but they carry fatal flaws. For example, a Bragg mirror made of alternating molybdenum and beryllium layers can achieve over 70 percent reflectivity at 11.4 nanometers. However, beryllium is highly toxic, making it entirely unsuitable for the rigorous, high-volume manufacturing environments of commercial semiconductor fabs.[5]

Similarly, researchers have investigated other plasma sources, such as xenon or lithium, which emit at different wavelengths. Yet none offer the same combination of conversion efficiency, safety, and manageable debris as tin. The industry is effectively locked into the tin-molybdenum-silicon triad.[5]

The Engineering Reality of Photon Loss

Even at this optimal intersection, the physics of extreme ultraviolet lithography are brutally inefficient. A 70 percent reflectivity per mirror sounds impressive, but a modern scanner requires a complex optical train to shape and focus the beam.[1]

The light must bounce off a collector mirror, pass through a series of illumination optics, reflect off the patterned circuit mask, and then navigate a projection optic system before finally hitting the silicon wafer. In total, the beam may reflect off a dozen different molybdenum-silicon surfaces.[5]

Because each mirror absorbs 30 percent of the light it receives, the losses compound exponentially. By the time the beam reaches the photoresist on the wafer, only a tiny fraction—often less than two percent—of the original extreme ultraviolet photons remain.[5]

This staggering loss rate explains why the initial carbon dioxide laser must be so powerful, and why the tin droplets must be vaporized 50,000 times a second. To deliver enough 13.5-nanometer photons to print a chip quickly, the source must generate an overwhelming surplus of light, brute-forcing its way past the fundamental limits of Bragg mirror reflectivity.[5]