🔬 optics

Dielectric-loaded guided-wave electro-optic modulator using bulk rubidium titanyl phosphate

This paper proposes a dielectric-loaded guided-wave electro-optic modulator that overcomes the lack of a mature thin-film RTP platform by utilizing a silicon nitride strip on a bulk rubidium titanyl phosphate crystal to achieve high-performance modulation at 1064 nm.

Original authors: David Trop, Boris Desiatov

Published 2026-07-22

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: David Trop, Boris Desiatov

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the world of light as a super-fast highway where information travels in beams of laser light. To make this highway useful for things like the internet or powerful lasers, we need to build "traffic lights" that can turn the light on and off or change its speed in a split second. These traffic lights are called electro-optic modulators. They work by using electricity to nudge the light, a trick known as the Pockels effect. For a long time, the best materials for building these tiny traffic lights have been crystals like lithium niobate. However, these materials have a downside: they act like a sponge for electricity, which creates heat and noise, especially when you try to switch them on and off very quickly or with high power.

Enter Rubidium Titanyl Phosphate, or RTP for short. Think of RTP as the "quiet, tough athlete" of the crystal world. It's already famous for being used in big, bulky devices that handle massive amounts of laser power without breaking a sweat. It doesn't get hot, it doesn't make annoying electrical noise, and it can handle high voltage like a champ. The problem is, nobody knew how to shrink RTP down to the tiny, microscopic size needed for modern computer chips. Previous attempts to carve RTP into tiny waveguides (the roads for light) were messy and complicated. This paper asks a simple but bold question: Can we keep RTP in its big, tough, bulk form and just build a tiny road on top of it, instead of carving the road into it?

The authors of this paper propose a clever solution they call a "dielectric-loaded guided-wave modulator." Instead of digging a trench into the RTP crystal, they simply place a thin strip of silicon nitride (a type of glass-like material) right on the polished surface of a big chunk of RTP. Imagine laying a narrow, slightly sticky tape on a smooth table; the light gets "stuck" to this tape and travels along it, skimming the surface of the crystal underneath. This setup allows the light to interact with the crystal's special properties without needing to cut the crystal itself.

To make this work, the researchers had to play a game of 3D Tetris with the crystal's orientation. The RTP crystal has a specific internal structure, like a grid, and the electricity needs to push on the light in just the right direction to work. They tested different ways to slice the crystal (called X-cut, Y-cut, and Z-cut) and found that the "X-cut" was the winner. In this orientation, the light, the electricity, and the crystal's internal grid all line up perfectly to create a strong effect. They simulated placing gold electrodes (the metal wires that carry the electricity) on the sides of the silicon strip to push the light.

Through detailed computer simulations, the team found that this design works surprisingly well. They discovered that by choosing the right width for the silicon strip (1.2 micrometers) and the right gap between the metal wires (5.25 micrometers), they could create a device that is much more efficient than the old, bulky versions. Specifically, they calculated that a 10-millimeter-long version of this new device would need only about 4.85 volts to switch the light, whereas a traditional bulk RTP device of the same length would need thousands of volts. That's a massive difference!

The paper also looked at the potential downsides. They found that if the metal wires are placed too close to the light, the light gets absorbed and lost, kind of like a car hitting a wall. So, they had to find a "sweet spot" where the wires are close enough to be effective but far enough away to keep the light safe. Their simulations suggest that with this balance, the device would lose only about 1.3 decibels of light per centimeter, which is a very acceptable amount for a high-performance device.

Importantly, the authors are careful to note that these results come from computer simulations, not from a physical device they built in a lab yet. They haven't measured the actual light loss or the speed in a real-world test. However, the simulations are rigorous, checking everything from how the light bends to how the electricity flows. They also ruled out other crystal orientations, showing that if you don't align the crystal correctly, the light won't stay trapped on the surface, and the device won't work.

In conclusion, this paper suggests a promising new way to bring the high-power, low-noise benefits of RTP crystals into the world of tiny, integrated chips. By simply loading a strip of material onto a polished crystal instead of carving it up, they might be able to create the next generation of ultra-fast, high-power light switches. While it's not a finished product yet, the math looks solid, and it opens the door for a future where we can control powerful lasers with the same ease as we control the lights in our smartphones.

Technical Summary: Dielectric-Loaded Guided-Wave Electro-Optic Modulator Using Bulk Rubidium Titanyl Phosphate

Problem Statement
Rubidium titanyl phosphate (RTP, RbTiOPO₄) is a well-established material for bulk Pockels cells, prized for its low dielectric permittivity, high electrical resistivity, negligible piezoelectric ringing, and high optical-damage tolerance. These properties make it ideal for high-power, high-repetition-rate, and intracavity laser systems where acoustic artifacts and RF-driven losses are detrimental. However, the application of RTP in integrated or guided-wave electro-optic (EO) modulators has been hindered by the lack of a mature thin-film RTP platform. Previous attempts using ion exchange or reactive-ion etching differ significantly from the desired architecture, often compromising the bulk material advantages or adding fabrication complexity. Consequently, RTP modulators have remained largely as free-space devices, limiting mode control, device footprint, and integration with planar optical systems.

Methodology
The authors propose and numerically evaluate a dielectric-loaded guided-wave architecture that avoids etching the active RTP crystal. The device consists of a silicon nitride (SiN) strip deposited on a polished, unpatterned bulk RTP substrate, flanked by gold ground-signal-ground (GSG) side electrodes.

Simulation Framework: All simulations were performed using Tidy3D (Flexcompute) at a wavelength of 1064 nm. The framework integrates full-vector optical eigenmode calculations, electrostatic and RF field extraction, and tensor-aware electro-optic perturbation analysis.
Material Modeling: The RTP substrate is modeled as an anisotropic crystal using Sellmeier equations for refractive indices and specific EO tensor coefficients (primarily $r_{33}$ ). The SiN strip is treated as an isotropic dielectric ( $n \approx 2.00$ ).
Design Co-Optimization: The study systematically co-designs the crystal orientation (X-cut, Y-cut, Z-cut), the optical mode polarization, and the electrode field direction. The goal is to maximize the overlap between the guided optical mode and the specific RF field component that drives the largest EO tensor element ( $r_{33}$ ) while maintaining low capacitance.
Performance Metrics: Key figures of merit include the half-wave voltage-length product ( $V_\pi L$ ), capacitance per unit length, microwave effective index, RF loss, and electrode-induced optical loss. The analysis specifically tracks the "tensor-weighted overlap" rather than simple scalar field overlaps.

Key Contributions

Novel Architecture: The paper introduces a dielectric-loaded waveguide on bulk RTP, creating a bound near-surface optical mode without requiring thin-film transfer or crystal etching. This preserves the bulk crystal's material advantages.
Orientation Co-Design: The authors demonstrate that for this specific side-electrode geometry, X-cut RTP is the optimal orientation. It aligns the in-plane guided mode polarization and the lateral electrode field ( $E_Z$ ) with the dominant $r_{33}$ tensor coefficient, whereas Y-cut and Z-cut orientations suffer from poor confinement or weaker tensor projections.
Comprehensive Trade-off Analysis: The study maps the design space to identify an operating point that balances modulation efficiency ( $V_\pi L$ ), optical confinement in the bulk RTP, and electrode-induced optical loss. It explicitly avoids the "minimum $V_\pi L$ " trap, which would require electrode gaps so small that optical loss becomes prohibitive.

Results
The optimized design features an X-cut RTP substrate with a 1.2 µm wide, 400 nm thick SiN strip and a 5.25 µm electrode gap.

Modulation Efficiency: The device achieves a single-arm $V_\pi L$ of approximately 4.85 V·cm. This represents a significant reduction in drive voltage compared to bulk devices (which typically require kilovolts), while maintaining a straight phase-shifter geometry.
Optical Confinement: The optical mode is well-confined, with ~48.6% of the optical energy residing in the bulk RTP and ~49.3% in the SiN strip. The active-axis field fraction in RTP is ~99.9%.
RF Performance: The GSG electrode structure is designed to be impedance-matched near 50 Ω ( $Z_0 \approx 49 \Omega$ ) with a microwave effective index ( $n_{RF} \approx 2.43$ ) close to the optical group index, suggesting potential for velocity matching in millimeter-scale devices. The estimated RC bandwidth for a 1 mm device is ~19 GHz.
Loss: The electrode-induced optical loss is calculated at ~0.286 dB/cm. Including a representative budget for SiN film absorption and scattering, the estimated total propagation loss is ~1.3 dB/cm.
Fabrication Tolerance: The design shows robustness against realistic fabrication variations. While SiN thickness and electrode gap are the most sensitive parameters, the device remains stable against lateral electrode offsets of ±0.25 µm.

Significance and Claims
The paper claims that this dielectric-loaded approach successfully bridges the gap between the high-power capabilities of bulk RTP and the integration benefits of guided-wave photonics. By avoiding direct etching of the RTP crystal, the platform retains the low capacitance, high optical-damage tolerance, and acoustic cleanliness that make RTP attractive for high-power and pulsed laser applications.

The authors position this work as a viable alternative for straight or weakly curved phase modulators in high-power, pulsed, and intracavity photonics. They explicitly note that the weak lateral confinement inherent to dielectric loading limits the structure's suitability for dense photonic integration (e.g., tight bends), but it is well-suited for applications where clean switching of large optical powers is the primary requirement. The study validates that competitive EO performance can be achieved on bulk crystals using established SiN processing, offering a practical route to guided-wave RTP modulators without the complexity of thin-film fabrication.

Technical Summary: Dielectric-Loaded Guided-Wave Electro-Optic Modulator Using Bulk Rubidium Titanyl Phosphate

More like this