Monolithic high density integrated photonics on bulk lithium niobate

Light is the engine of modern information, carrying data across oceans and through the air at the speed of light. To harness this power for computers and communication, scientists have spent decades shrinking massive optical components down to the size of a fingernail, creating tiny circuits where light travels instead of electricity. For these circuits to be truly useful, they need to be cheap to make, small enough to pack millions onto a single chip, and capable of changing their behavior quickly. While silicon has been the workhorse of this industry, it struggles to change light on command without losing energy. A crystal called lithium niobate has long been the ideal material for this task because it can manipulate light with incredible speed and efficiency, but it has remained stubbornly difficult to use in mass production. The traditional way to build with it involves slicing the crystal into paper-thin layers and gluing them to other materials, a process that is expensive, fragile, and difficult to scale up.

A team of researchers at Delft University of Technology has found a way to bypass these difficulties entirely. Instead of slicing and gluing, they built a fully integrated circuit directly on a solid block of the crystal. They grew a thin, amorphous silicon carbide layer on top of a standard, thick piece of lithium niobate and etched tiny channels into it to guide the light. This approach keeps the manufacturing simple and compatible with the factories that already make computer chips, while still allowing the light to interact with the powerful properties of the crystal underneath. The result is a platform that is not only cheaper and easier to make but also performs better than previous attempts, achieving a level of control over light that was previously thought to require much more complex setups.

The core of this achievement lies in how the researchers handled the materials. Lithium niobate is a robust crystal, but it is notoriously hard to cut and shape with the precision needed for modern electronics. Previous methods tried to solve this by creating a thin film of the crystal and bonding it to a silicon wafer, a technique that requires high temperatures and delicate handling. The new method skips the thin film entirely. The researchers started with a standard, thick slice of lithium niobate and deposited a layer of amorphous silicon carbide on top. This material is transparent to light and can be patterned with extreme precision using standard tools. By keeping the entire process below a temperature of 300 degrees Celsius, they ensured that the underlying crystal remained untouched and unharmed. This means the light travels through a hybrid path: mostly in the silicon carbide layer, but with a significant portion of its energy dipping down into the lithium niobate, where it can be controlled.

To prove this system works, the team built tiny ring-shaped circuits where light circulates continuously. These rings act as sensitive detectors for how well the light is being guided and how easily it can be tuned. They measured the loss of light as it traveled through the silicon carbide channels and found it to be remarkably low, indicating that the light moves smoothly without scattering or fading. They also tested how well the circuit could be adjusted using heat and electricity. When they applied a small amount of heat, the light shifted its color, a property known as thermo-optic tuning. More impressively, when they applied an electric voltage, the crystal's internal structure shifted just enough to change the light's path. They measured the efficiency of this electrical control and found that it required a voltage-length product of 2.87 volt-centimeters to switch the light, a record high for circuits built directly on bulk lithium niobate. This means the device can switch light on and off very quickly with very little power.

The researchers did not stop at simple rings; they wanted to see if they could control the flow of light even more precisely. To do this, they carved a series of tiny, elliptical holes into the silicon carbide layer, creating a structure known as a photonic crystal. Imagine a row of pillars that light must weave through; by spacing these pillars just right, they can create a barrier that stops certain colors of light from passing while letting others through. The team used this technique to build filters and mirrors that could trap light in a small space. They created a cavity, or a trap, for light between two of these mirror structures. Inside this trap, the light bounced back and forth, and the researchers measured how long it stayed there. They found that the light was confined with a quality factor of 76,437, a number that indicates the light circulates many times before escaping. This level of confinement is usually only possible with much more complex, suspended structures, proving that their flat, solid approach works just as well.

The most striking result came when they combined these photonic crystals with the electrical tuning. By slowing down the light as it passed through the crystal structure, they made the light spend more time interacting with the electric field. This is similar to how a car moving slowly through a muddy field gets stuck more than one speeding by, but in this case, the "stuck" light interacts more strongly with the material. This slowing effect boosted the efficiency of the electrical control by a factor of 8.5. At specific points near the edge of the crystal's operating range, the device became incredibly sensitive, requiring even less voltage to switch the light. This demonstrated that the platform could not only guide light but also engineer its behavior to maximize performance.

The significance of this work extends beyond just one type of crystal. The method used to create these circuits is entirely compatible with the standard manufacturing processes used for computer chips, which means it can be scaled up to produce millions of devices at a low cost. It avoids the need for the expensive and complex steps of slicing thin films or bonding different materials together. The researchers showed that this approach could potentially be applied to other functional crystals used in optics, opening the door to a new generation of photonic devices. By proving that high-performance, tunable circuits can be built directly on solid blocks of material, they have removed a major barrier that has kept advanced optical technology from becoming a standard part of everyday electronics. The path forward is now clearer, with a route to high-density, low-cost photonics that relies on simple, robust fabrication rather than delicate, expensive tricks.