Abstract: An anisotropic model for the fabrication of annealed and reverse proton exchange (RPE) waveguides in lithium niobate (LN) is presented. We characterized the anisotropic diffusion properties of proton exchange, annealing, and RPE in Z-cut and X-cut substrates using planar waveguides. Using this model, we fabricated high-quality channel waveguides with propagation losses as low as 0.086 dB/cm and a coupling efficiency with optical fiber of 90% at 1550 nm. The splitting ratio of a set of directional couplers is predicted with an accuracy of ±0.06.
Introduction
Lithium Niobate is often referred to as the "silicon" of photonics due to its exceptional electro-optic and nonlinear properties. To create integrated optical circuits, engineers frequently use Proton Exchange (PE)—a process where lithium ions are swapped for hydrogen ions to modify the refractive index. However, designing these waveguides has historically been challenging. Most existing models assume isotropic diffusion, meaning protons move at the same rate in every direction. Since LN is a uniaxial crystal, its physical properties vary by orientation, making the isotropic assumption inaccurate for designing complex components like directional couplers where lateral diffusion is critical.
Methods and Characterization
The researchers developed a comprehensive anisotropic model by characterizing the diffusion properties of Annealed Proton Exchange (APE) and RPE across two different crystal orientations: Z-cut and X-cut.
- Planar Waveguide Characterization: Planar waveguides were fabricated to isolate diffusion along specific axes. The effective refractive indices of the guided modes were measured using a prism coupling technique.
- Modeling Diffusion: A one-dimensional nonlinear diffusion equation was used to describe the evolution of proton concentration ($C$) over time. Diffusion coefficients were treated as variables dependent on both concentration and crystal orientation.
- 2D Simulation: These parameters were then applied to a two-dimensional diffusion equation to predict the refractive index profiles of channel waveguides. This allowed for the simulation of optical modes in buried waveguides created via RPE.
- Experimental Validation: To verify the model, the team fabricated straight waveguides and 15 directional couplers on a Z-cut substrate, comparing experimental performance against their simulations.
Key Insights and Anisotropy
A central finding of the study is that anisotropy is non-negotiable for accurate waveguide design. The diffusion coefficient for X-cut substrates ($D_{PE,X} = 0.098 \, \mu\text{m}^2/\text{h}$) was found to be nearly double that of Z-cut substrates ($D_{PE,Z} = 0.056 \, \mu\text{m}^2/\text{h}$).
Ignoring these differences leads to significant errors in predicting the lateral "bulge" of a waveguide, which is vital for determining the coupling strength between adjacent waveguides in a circuit. Furthermore, the use of Reverse Proton Exchange (RPE) was shown to "bury" the waveguide, pulling the optical mode away from the surface. This reduces surface scattering losses and creates a more circular mode shape, optimizing coupling with standard optical fibers.
Key Results
The application of this anisotropic model yielded several record-breaking results:
- Ultra-Low Propagation Loss: Fabricated channel waveguides achieved losses as low as 0.086 dB/cm at 1550 nm.
- High Coupling Efficiency: A 90% coupling efficiency with standard optical fibers was achieved due to the symmetric mode profiles enabled by RPE.
- Predictive Accuracy: The overlap between measured and calculated optical modes reached 96%. Most impressively, the model predicted the splitting ratios of directional couplers with an RMS error of only 0.06, proving the accuracy of the anisotropic parameters.
Conclusion and Future Work
While this model is highly effective for the α-phase of proton exchange—the most desirable phase for low-loss telecommunications—future adjustments may be needed for other crystallographic phases used in high-index-contrast devices. This work provides a robust foundation for the predictive design of complex integrated photonic structures, such as Mach-Zehnder interferometers and ring resonators, where sub-micron precision is essential for performance.
Summary updated