High-power handling and bias stability of thin-film Lithium Tantalate microring and coupling resonators

1. Problem Statement

The primary challenge addressed in this work is the photorefractive (PR) effect inherent to thin-film lithium niobate (TFLN) platforms. While TFLN is a leading candidate for next-generation optical communication and quantum devices due to its strong electro-optic (EO) coefficients, the PR effect causes:

• Severe power handling limitations: High optical intensities induce refractive index changes, limiting on-chip power to low levels.
• DC bias instability: The PR effect causes drift in bias points, making stable low-power bias circuits difficult to implement.
• Thermal management issues: Mitigating PR drift often requires the thermo-optic (TO) effect, which is weak in LN, leading to high static power consumption.

These instabilities hinder the transition of TFLN devices from proof-of-principle demonstrations to practical, high-power, and stable commercial applications (e.g., quantum transducers and high-speed modulators).

2. Methodology

The authors investigated Thin-Film Lithium Tantalate (TFLT) as a superior alternative to TFLN, leveraging its similar EO properties but significantly higher optical damage threshold.

• Device Fabrication:
  • Material: 4-inch TFLT wafers on silicon substrates with a 4.7 µm buried oxide (BOX).
  • Structure: Microring resonators and coupling modulators were defined via photolithography and etched using Ar+ plasma.
  • Cladding: A 1.5 µm oxide cladding was deposited via PECVD to ensure velocity matching for high-speed modulation and to protect the waveguide.
  • Annealing: Devices underwent thermal annealing at 500°C for 2 hours to suppress the PR effect.
• Experimental Setup:
  • High-Power Characterization: A tunable laser (Santec-570) was amplified by an EDFA to generate high circulating powers (up to 4 W). Transmission spectra were measured in both forward and backward wavelength scans to distinguish between PR (blue-shift) and TO (red-shift) effects.
  • Modulator Testing: A 2 mm long coupling modulator was driven by an arbitrary waveform generator (AWG) to test bias stability and phase control in the C-band.

3. Key Contributions

1. Demonstration of Ultra-High Power Handling: The paper establishes that TFLT microring resonators can handle circulating powers of ~4 Watts with minimal frequency shift (< 1 GHz) and no observable PR effect after annealing.
2. Suppression of PR Effect via Annealing: The study proves that thermal annealing effectively suppresses the PR effect in TFLT to a level where it becomes unobservable, even in the presence of the TO effect.
3. First TFLT Coupling Modulator: The authors report the first resonant coupling modulator on the TFLT platform, achieving a low half-wave voltage ($V_\pi$) of 3 V over a 2 mm interaction length.
4. Bias and Phase Stability: The work demonstrates stable DC bias operation and phase control, overcoming the drift issues that plague TFLN devices.

4. Key Results

A. Power Handling and Stability

• Pre-Annealing: Unannealed devices showed significant resonance distortion and a maximum frequency shift of ~3.5 GHz at 1 W intracavity power. Forward and backward scans showed different dynamics, indicating a mix of PR and TO effects.
• Post-Annealing:
  • Frequency Shift: At 1 W, the shift dropped to ~160 MHz. At 4 W, the shift remained only ~1 GHz (well within the resonance linewidth).
  • Symmetry: Forward and backward scans became nearly identical, confirming the PR effect was negligible compared to the TO effect.
  • Quality Factor: The device $Q$-factor increased by a factor of ~2 after annealing.
• Comparative Analysis: Table I in the paper compares the "Stability Factor" (normalized TO-induced frequency shift vs. energy density).
  • TFLT (After Anneal): $1.8 \times 10^{-7} \text{ mJ}^{-1}\text{cm}^3$.
  • TFLN (Oxide Coated): $4.2 \times 10^{2} \text{ mJ}^{-1}\text{cm}^3$.
  • Result: TFLT performs approximately 9 orders of magnitude better than TFLN in terms of power stability. It outperforms AlGaAs and Ta2O5, though SiN remains slightly better.

B. Coupling Modulator Performance

• Efficiency: The 2 mm device achieved a $V_\pi$ of 3 V, resulting in a $V_\pi L$ product of 0.6 V·cm. This represents a ~4.7x improvement over a straight waveguide modulator (14 V).
• Extinction Ratio: By tuning the bias from 0 V to 3 V, the device switched from critical coupling to an "uncoupled" state with an extinction ratio > 15 dB.
• Stability:
  • Bias Stability: Transmission remained stable over 40 minutes with stepped DC bias voltages (0 V to 3 V).
  • Phase Stability: When applying high DC voltages (±50 V) to the resonator arm, the resonance position shifted by less than 0.1 pm over 25 minutes, demonstrating exceptional long-term stability.

5. Significance

This work represents a critical step toward the commercialization of thin-film photonic platforms:

• Practical High-Power Applications: By enabling watt-level power handling, TFLT opens the door for high-power nonlinear optics, efficient frequency combs, and high-power optical amplification on-chip, which were previously impossible on TFLN due to PR damage.
• Quantum Computing: The stability of TFLT is crucial for quantum applications (e.g., microwave-to-optical transduction) that require cryogenic environments and stable bias circuits without the high power consumption associated with thermal compensation in LN.
• Telecommunications: The demonstrated coupling modulator offers a path to ultra-high-speed, low-energy optical communication with stable resonant EO modulators, solving the drift issues that have limited TFLN modulators in real-world deployments.

In summary, the authors have successfully transitioned TFLT from a theoretical alternative to a practical, high-performance platform that combines the high efficiency of lithium niobate with the robust power handling and stability required for next-generation photonic integrated circuits.