In Simple Terms
Researchers used an extremely brief flash of light to create tiny pits in silicon, the material used in computer chips. Changing the flash’s energy altered the pits’ size and shape. The test shows the idea can work in a lab, but it does not yet prove it is faster, cheaper, or ready for chip factories.
Why Look for a Different Approach?
Making tiny patterns on silicon chips usually takes several steps and may involve photomasks, which shape light according to a specific design. Making these masks is precise and costly: the process can involve adding light-sensitive layers, exposing them to light, treating them with chemicals, and removing parts of the material. As electronic components get smaller, the precision of the mask itself becomes increasingly important.
Direct laser writing offers another approach. It focuses light on selected spots to remove material without an intermediate mask. But lasers that use visible or near-infrared light face limits on how tightly they can focus their beams. They can also transfer heat into the material, potentially causing unwanted effects such as stress or tiny cracks.
Extreme-ultraviolet light makes it possible to test smaller spots and remove material more precisely. But a short pulse alone does not prove that heat damage has been eliminated. The researchers expected ultrashort flashes to help reduce it, but the study did not directly measure their thermal effects.
A Huge Facility, a Flash Lasting Trillionths of a Second
The experiment took place at Japan’s SACLA facility, using a free-electron laser that produced pulses lasting about 30 femtoseconds. A femtosecond is an extremely small fraction of a second, so each pulse delivered its energy to the sample in a very short time. The photons had an energy of about 92 electron volts and hit a silicon wafer, 525 micrometers thick, straight on.
Under the experimental conditions, silicon reflected almost none of the light, while the light’s absorption depth was about 0.59 micrometers. A two-stage optical system focused the beam onto an area roughly 350 by 700 nanometers across. The setup included a Kirkpatrick–Baez mirror and an elliptical mirror. The samples were mounted on precision stages that let the team aim pulses at different spots.
These details describe an advanced experimental setup, not a compact device that can simply be assumed to exist in an ordinary factory. The results relied on a light source at a large research facility. Making the technique more widely usable will require more practical sources of extreme-ultraviolet light, as well as a way to operate them reliably and efficiently.
The Pulse’s Energy Shapes the Crater
After exposing the samples, the researchers measured their surface features with an atomic force microscope, which detects tiny differences in height. At an absorbed dose of 8 electron volts per atom, equivalent to 6.0 joules per square centimeter, the craters were about 1.5 micrometers deep on average. The deepest points in some craters reached 1.8 micrometers.
The craters were not identical. Most had a broad, bowl-like shape that narrowed toward a central depression. One sample also showed a distinct rim of material that appeared to have been removed and then deposited nearby. The crater dimensions varied by about 13 to 20 percent across repeated pulses, suggesting the results were repeatable to some degree, but not perfectly consistent.
When the dose was lowered to 5 electron volts per atom, or 3.7 joules per square centimeter, the craters averaged about one micrometer deep and 1.1 micrometers wide. Their widths varied by around 10 percent, while their depths varied by 18 to 36 percent. This shows that changing the energy affects not only how much material is removed, but also how consistently the craters form.
At a dose of 2 electron volts per atom, or 1.5 joules per square centimeter, the material removal remained closer to the depth at which the light was first absorbed. The craters were about 0.6 micrometers deep, and some of their dimensions followed the shape of the light spot, which was less than a micrometer wide. Their bottoms, however, were not smooth: they had rough patches and silicon left behind. The lower dose could create a feature narrower than a micrometer in one direction, but it did not produce a perfect surface.
Useful Simulations, but Not a Perfect Match
The team also used a multiscale simulation system called SURFwiX to model what happened. The XTANT program represented how electrons responded and how the material began to change during the first femtoseconds and picoseconds. The researchers then used the LAMMPS molecular-dynamics program to study silicon’s deformation, the movement of pressure waves, and the ejection of material over a longer period, extending into the nanosecond range.
To reduce the computing demands, the researchers made the model smaller while preserving the absorbed dose according to a similarity principle. The simulations captured the overall trend: higher doses were associated with larger craters. But at the highest dose, the simulated craters were smaller than those measured. The model underestimated their depth by about a factor of three and their width by about a factor of two.
The simulations suggested that some silicon at the crater’s edges might be pushed outward, then return and solidify. This could explain the rim of redeposited material. But the gap between the calculations and the measurements shows that the atomic-scale models do not yet capture every detail of the process. They will need improvement before they can accurately predict crater shapes.
From Microscopic Craters to Possible Uses
Direct writing with extreme-ultraviolet light could eventually be an option for making patterns or masks for components with relatively larger features. Possible areas include some applications involving microelectromechanical sensors, motor control, battery management, and medical devices. These are potential areas of interest, not products that the experiment demonstrated how to make.
The approach might reduce reliance on some mask-making steps, including chemical processing, plasma etching, and the removal of light-sensitive layers. But the study did not measure resource use or environmental impact, or compare the process’s cost and speed with existing manufacturing methods. The idea that this could become a dry process that uses fewer resources is therefore a possibility to investigate, not an established result.
One practical challenge is that the experiment depended on a free-electron laser housed at a specialized facility. The study points to high-harmonic generation as one possible route to smaller sources of extreme-ultraviolet light, but industrial use would require further development and testing. The researchers also suggest that even shorter, attosecond-scale pulses might help reduce thermal stress. That remains a possibility, not a benefit measured in this experiment.
The Bottom Line
The experiment showed that a single, ultrashort pulse of extreme-ultraviolet light can remove silicon and create a crater whose depth and shape depend on the dose. Results ranged from craters about 0.6 micrometers deep at the lowest energy to a maximum depth of 1.8 micrometers at the highest, with noticeable variation between measurements. The work’s main achievement is showing that direct, microscale carving is possible—not that the technique is ready for industry. The simulations need improvement, and researchers still need to test speed and repeatability at larger scales, measure thermal and environmental effects, and develop suitable light sources. For now, the flash is a promising laboratory tool, not a proven replacement for chipmaking methods.