A photomask is a flat plate of glass with a precise pattern etched onto it, used as a stencil to transfer circuit designs onto silicon wafers during chip manufacturing. Think of it like a photographic negative: light shines through the mask, and the pattern of light and shadow imprints a design onto the wafer below. Every microchip in every phone, computer, and car starts with a photomask.
What a Photomask Is Made Of
A standard photomask starts with a substrate of high-purity quartz glass, chosen because it lets ultraviolet light pass through with minimal distortion. On top of the quartz sits a thin layer of chromium, typically about 100 nanometers thick, which blocks UV light completely. That chromium layer is selectively removed in certain areas to create the pattern. Where chromium remains, light is blocked. Where it’s been etched away, light passes through. The result is an extremely precise stencil of the circuit layout.
On top of the chromium, there’s usually a thinner layer of chromium oxide (about 26 nanometers) that reduces unwanted reflections. Research has shown that as little as 40 nanometers of chromium is enough to fully block UV light, but manufacturers use thicker layers to ensure durability and scratch resistance during repeated use.
How Photomasks Transfer Patterns to Chips
The process of using a photomask is called photolithography, and it works in a few key steps. First, a silicon wafer is coated with a light-sensitive material called photoresist. Then the photomask is positioned above the wafer and UV light is directed through the mask. Wherever the mask is transparent, UV light hits the photoresist and changes its chemical structure. Wherever chromium blocks the light, the photoresist stays unchanged.
After exposure, the wafer is washed with a chemical developer. Depending on the type of photoresist used, either the exposed or unexposed portions dissolve away, leaving behind a precise pattern on the wafer’s surface. That pattern then serves as a guide for the next processing step, whether that’s etching trenches into the silicon, depositing metal wires, or adding insulating layers. A single modern chip requires dozens of different photomasks, one for each layer of the circuit.
There are two main approaches to how the mask and wafer interact. In contact lithography, the mask is pressed directly against the wafer to minimize light-scattering effects. In projection lithography, a system of lenses sits between the mask and the wafer, projecting and shrinking the mask’s pattern down to a fraction of its original size. Projection systems are standard in advanced chip manufacturing because the mask pattern can be four or five times larger than the final feature on the wafer, making it easier to fabricate accurately.
How Photomask Patterns Are Created
The patterns on a photomask are written using tightly focused beams of electrons, a process called electron-beam (e-beam) writing. This is necessary because the features are far too small for conventional methods. Modern multi-beam mask writers fire over 262,000 individually programmable electron beams simultaneously, achieving sub-nanometer positioning precision and resolving features as small as the low tens of nanometers. This combination of extreme accuracy and reasonable speed makes it possible to write the billions of features required for advanced chip designs.
The pattern on the mask doesn’t look exactly like the final circuit, though. At the tiny scales involved, light bends and distorts as it passes through narrow openings, a phenomenon called diffraction. To compensate, engineers add tiny corrective shapes to the mask pattern, a technique known as optical proximity correction. These sub-resolution assist features, small serifs, notches, and helper bars, pre-warp the design so that after light diffraction does its work, the pattern that lands on the wafer matches the intended circuit layout.
Phase-Shift Masks for Sharper Patterns
As chip features have shrunk below the wavelength of the light used to print them, standard binary masks (fully opaque or fully transparent) struggle to produce sharp images. Phase-shift masks solve this by manipulating not just where light passes through, but how the light waves align when they arrive at the wafer.
In an attenuated phase-shift mask, the normally opaque areas are replaced with a semi-transparent material that lets a small amount of light through, but shifts that light’s wave by exactly half a cycle. When this phase-shifted light meets the unshifted light from a neighboring clear opening, the two waves cancel each other out at the boundary. This destructive interference creates a sharper, darker edge between bright and dark regions on the wafer. The result is higher contrast, crisper features, and a wider margin for focus errors during printing.
EUV Masks Work Differently
The newest generation of chip manufacturing uses extreme ultraviolet (EUV) light, which has a wavelength of just 13.5 nanometers, roughly 14 times shorter than the deep ultraviolet light used in older systems. At this wavelength, virtually all materials absorb light rather than transmitting it, so EUV masks can’t work as stencils that light passes through. Instead, they work as mirrors.
An EUV mask is built on a substrate coated with 40 alternating pairs of molybdenum and silicon layers, each pair only about 7 nanometers thick. These ultrathin layers work together to reflect EUV light, similar to how a stack of partially reflective coatings in high-end optics can bounce back nearly all incoming light at a specific wavelength. The individual molybdenum layers are about 2.8 nanometers thick, and the silicon layers about 4.1 nanometers. On top of this reflective stack, an absorber material is patterned to selectively block reflection in certain areas, creating the circuit pattern.
Protecting Masks From Contamination
Because photomasks are reused thousands of times and even a speck of dust can ruin a chip’s circuit, they’re protected by a thin transparent membrane called a pellicle. The pellicle is mounted a few millimeters above the mask surface. Any particle that lands on the pellicle sits far enough from the pattern plane that it stays out of focus and doesn’t print onto the wafer.
For EUV lithography, pellicle engineering is especially challenging. The membrane must be thin enough to transmit EUV light efficiently while surviving intense heat from the high-powered light source. Current research focuses on membranes made from molybdenum-silicon compounds. Adjusting the ratio of molybdenum to silicon lets engineers trade off between mechanical strength, thermal durability, and optical transparency. Compositions richer in molybdenum are stronger and more heat-resistant, while silicon-rich versions transmit more light.
Finding and Fixing Defects
Quality control for photomasks is extraordinarily demanding. A single defect smaller than a virus particle can cause a flaw that repeats across every chip printed from that mask. Inspection tools scan the entire mask surface and compare it against the intended design, flagging any deviation.
When defects are found, they need to be repaired without damaging the surrounding pattern. Traditional repair methods use nanoscale tips or focused electron beams to add or remove material at specific locations. A newer approach uses femtosecond laser pulses, bursts of light lasting just quadrillionths of a second, to vaporize unwanted material from the mask surface. This laser method is significantly faster and cheaper than older techniques, and it can remove surface contaminants from EUV masks without damaging the delicate multilayer structure underneath.
The Scale of the Photomask Industry
The global photomask market was valued at $4 billion in 2018 and was projected to reach nearly $5 billion by 2026. Growth is driven by rising demand for semiconductors across consumer electronics, automotive systems, and connected devices. Display manufacturing, for screens in phones, TVs, and monitors, represents the largest and fastest-growing segment of photomask demand, since display panels also rely on lithographic patterning to create their pixel circuits.
Each new generation of smaller, more complex chips requires more mask layers and tighter tolerances, pushing up the cost per mask set. A complete set of EUV photomasks for a cutting-edge processor design can cost millions of dollars. That cost is justified because a single mask set can print patterns on millions of wafers, making it one of the most leveraged investments in all of manufacturing.