This article explores a new technique from the National Institute of Standards and Technology (NIST). The team patterns tantala (tantalum pentoxide) and thin-film lithium niobate onto silicon wafers using heterogeneous integration.
By stacking these materials in a 3D structure, they blend different photonic strengths on one platform. This setup lets the chip work across visible and near-infrared wavelengths and could lead to smaller, more versatile photonic chips.
Overview of the NIST approach
Heterogeneous integration is at the heart of this: placing materials with different optical properties on top of established photonic circuits to create devices that do more than one thing. In this project, tantala gets deposited at room temperature and can be annealed at moderate heat, with low residual stress even in thick films. That makes it pretty friendly to standard silicon-based photonics.
They pair lithium niobate (LN), which brings strong second-order (χ(2)) nonlinearities, with tantala. This combo lets nonlinear optical processes happen on the same wafer without losing the performance of existing circuits.
The researchers patterned both materials in a 3D stacked architecture, basically “sprinkling” tantala right on top of LN on silicon. LN brings robust χ(2) nonlinearities, while tantala adds visible-wavelength operation and its own nonlinear behavior. Together, you get a single photonic platform with a broad spectral range and more features than either material could offer alone.
Materials and architecture
The team used a layering method where upper-layer tantala and lower-layer lithium niobate interact at the device level. By stacking them, the architecture puts nonlinear optics right into today’s and tomorrow’s photonic systems.
Room-temperature processing of tantala and its low intrinsic stress in thick films help keep devices stable, even in complicated circuits. This method tackles a long-standing challenge in integrated photonics: no single material does it all—light generation, frequency conversion, and fast routing across both visible and near-infrared bands.
The 3D stacking approach offers a practical way to use multiple material properties without overhauling existing manufacturing lines.
Demonstration and chip-scale results
For a proof of concept, the researchers combined two nonlinear processes: upper-layer tantala four-wave-mixing optical parametric oscillation and lower-layer lithium niobate second-harmonic generation. This let them achieve arbitrary-wavelength laser conversion, showing how different nonlinear effects can work together on a single wafer.
The team made about 50 chips per wafer, each roughly fingernail-sized. Each wafer held around 10,000 photonic circuits, each set up to output a unique color. That’s a lot of circuits packed into a small space, which really shows this isn’t just a lab curiosity—it’s manufacturable hardware.
Optical functions demonstrated
- They integrated χ(2) nonlinearities from LN with the nonlinear behaviors of tantala in a single stack.
- They pulled off arbitrary-wavelength laser conversion by coordinating four-wave mixing and SHG across different layers.
- This opens the door for visible-to-near-infrared photonic sources, giving access to atomic transitions and wavelength-specific applications.
Implications for photonic systems
This development goes after a stubborn bottleneck in integrated photonics: making a single chip that can generate, convert, and route light across a wide range of wavelengths. The 3D heterogeneous integration method gives a path to scalable, multifunctional photonic systems.
Now, compact, field-ready optical sources can span visible to near-infrared wavelengths. By letting materials with different strengths share the same platform, NIST’s method opens up new design possibilities for on-chip light sources, frequency converters, and nonlinear-optics-based signal processing.
The idea of “sprinkling” compatible materials onto existing circuits could speed up the growth of complex, multi-material photonic ecosystems. That’s important for everything from quantum tech to spectroscopy or atomic physics experiments. It’s an exciting step forward, though of course, there are still challenges ahead.
Future directions and challenges
- Scaling this approach to larger wafers and higher circuit densities. It’s tough to keep the material quality uniform as everything gets bigger.
- Dealing with thermal budgets and those tricky cross-material interactions as circuits keep getting more complex.
- Figuring out how to actually integrate these heterogeneous stacks into packaged systems for the real world. It’s one thing in the lab, but quite another out there.