Photoionization in KTN deflectors by light in the near-infrared imaging window

Technical Summary: Photoionization in KTN Deflectors by Light in the Near-Infrared Imaging Window

Problem Statement
Electro-optical deflectors (EODs) based on potassium tantalate niobate (KTN) crystals offer high-speed, inertia-free beam scanning with high resolution, making them ideal for laser-scanning microscopy. These devices operate by injecting electrons into the crystal, where they become trapped at defect sites (likely oxygen vacancies), creating a space-charge distribution that modifies the refractive index via the Kerr effect. However, the trapped charges are susceptible to photoionization, where incident photons free the electrons, degrading the internal electric field and reducing the deflection angle.

While visible light is known to rapidly deplete trapped charges, necessitating continuous bias voltage recharging (which introduces angular offsets and limits scan range), the behavior of KTN under near-infrared (NIR) illumination remains poorly characterized. Previous work suggested that NIR light causes minimal photoionization, but quantitative data across the biological imaging windows (NIR-I and NIR-IIa, 700–1300 nm) has been lacking. Without this data, it is difficult to determine if KTN deflectors can operate in discontinuous charging modes (alternating between DC charging and unbiased AC scanning) for deep-tissue imaging without significant charge loss.

Methodology
The authors conducted a quantitative study of photoionization in a KTN deflector across wavelengths from 700 nm to 1300 nm in 50 nm increments.

• Device Setup: A 3.2 mm × 1.2 mm × 4 mm KTN crystal was charged to 450 V to inject and trap electrons. The crystal was maintained at 28 °C to ensure operation in the cubic phase where the Kerr effect is strongest.
• Charge Density Measurement: The authors employed a phase-shifting Mach-Zehnder interferometer using a 1064 nm probe laser. By measuring the birefringent phase retardation induced by the space-charge field, they estimated the trapped charge density. The system utilized a two-beam polarization interferometer to resolve the differential phase between light polarized parallel and perpendicular to the internal field.
• Photoionization Protocol: The charged crystal was exposed to a tunable NIR excitation laser (100 mW transmitted power) for a series of controlled exposures. Between exposures, the charge density was measured. The exposure duration was exponentially increased to sample the rapid initial decay and slower subsequent decay.
• Modeling: The authors developed a drift-diffusion-Poisson model to interpret the data. They initially tested a simplified hypothesis: a single defect species (modeled as a doubly ionized oxygen vacancy capable of holding two electrons) with negligible electron recapture. This model was linearized to allow fitting of photoionization cross-sections ($\sigma_1, \sigma_2$) against cumulative photon fluence.

Key Results

• Wavelength Dependence: The photoionization rate decreased dramatically with increasing wavelength. At 700 nm, the trapped charge was fully depleted by the experimental photon fluence. In contrast, at 1300 nm, the same fluence resulted in less than 5% charge loss.
• Multi-Exponential Decay: The charge density decay curves exhibited multi-exponential behavior. For instance, at 900 nm, the decay was initially sharp but asymptoted around 60% charge density. This behavior could not be explained by a single-trap model without recapture.
• Model Discrepancy: When fitting the experimental data to the single-trap, no-recapture model, the normalized root mean square error (NRMSE) was substantial (median 0.136). The authors concluded that the data is inconsistent with a single defect species, even one with multiple charge states, suggesting the presence of multiple trap species or significant recapture dynamics.
• Two-Photon Absorption: Calculations indicated that two-photon photoionization was unlikely to be the dominant mechanism, as the experimentally observed fluence required for charge loss was orders of magnitude lower than the theoretical threshold for two-photon processes.
• Stability in Darkness: In the absence of illumination, the trapped charge was highly stable, with over 96% of the initial charge remaining after 80 minutes.

Significance and Claims
The paper provides the first quantitative measurements of photoionization in KTN deflectors across the NIR-I and NIR-IIa biological imaging windows. The primary significance of these findings is practical guidance for the operation of KTN-based EODs in advanced imaging applications:

1. Operational Mode: The slow photoionization rates at NIR wavelengths suggest that continuous bias voltage recharging may not be necessary for many applications. Instead, performance can likely be maintained via discontinuous charging (brief recharging pulses between scans), which eliminates angular bias and maximizes the geometric scan angle.
2. Wavelength Selection: The data offers a quantitative basis for selecting operating wavelengths to minimize charge loss, particularly for deep-tissue imaging and nonlinear microscopy where specific wavelengths are required.
3. Defect Characterization: The observation of multi-exponential decay challenges the assumption of a single dominant trap species. The authors posit that the complex dynamics likely arise from multiple trap species or substantial recapture, indicating a need for further characterization using techniques like electron paramagnetic resonance (EPR) or deep-level transient spectroscopy (DLTS).

The authors conclude that these insights are critical for implementing advanced deflection techniques, such as relayed optical loops, where the accumulation of angular bias from continuous charging would otherwise severely degrade system performance.