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

Electro-optical deflectors (EODs) offer unparalleled scanning speed for laser-scanning microscopy and other applications, but suffer from limited deflection range. EODs based on potassium tantalate niobate (KTN) crystals feature some of the highest number of resolvable spots. These deflectors rely on internal electric fields generated by trapped electrons to enable beam scanning. However, visible light induces rapid photoionization of trapped charges, thus KTN-based deflectors are typically continuously recharged with a bias voltage that effectively limits the range of the deflector. Recent work has proposed the use of KTN-based EODs for biological imaging with infrared excitation light, but quantitative data on near-infrared photoionization is lacking. Here, we present quantitative measurements of photoionization in KTN deflectors across the NIR-I and NIR-IIa biological imaging windows (700 - 1300 nm), a range that is particularly important for deep tissue imaging and nonlinear microscopy. Using a two-beam polarization interferometer, we measured trapped charge density as a function of photon fluence. We observed that the photoionization rate decreases dramatically with increasing wavelength. The charge density decay curves exhibit multi-exponential behavior that cannot be explained by a single-trap model without recapture, indicating the presence of multiple trap species or substantial recapture. These measurements provide critical quantitative guidance for selecting operating wavelengths and charge-scan duty cycles for KTN-based EODs in near-infrared imaging applications.

💡 Research Summary

This paper presents a comprehensive quantitative study of photo‑ionization in potassium tantalate niobate (KTN) electro‑optical deflectors (EODs) when illuminated with near‑infrared (NIR‑I and NIR‑IIa) light spanning 700 nm to 1300 nm, a spectral region of great relevance for deep‑tissue and two‑photon microscopy. KTN‑based EODs rely on space‑charge fields generated by electrons trapped in crystal defects; these fields produce a Kerr‑type refractive‑index change that enables rapid beam scanning. Visible illumination rapidly empties the traps, so conventional operation uses a continuous bias voltage to replenish charge, which limits the usable scan angle. The authors therefore investigated whether NIR excitation could mitigate this limitation.

A 3.2 × 1.2 × 4 mm KTN crystal with Ti/Pt/Au electrodes was temperature‑stabilized at 28 °C (above the cubic‑phase transition). The crystal was first charged by applying 450 V DC, filling the traps with electrons. After removing the bias, the device was exposed to a tunable pulsed laser (Spectra‑Physics InSight X3) delivering 100 mW average power at an 80 MHz repetition rate. Pulse duration was nominally 100 fs; group‑delay dispersion (GDD) was deliberately added using four 5 mm ZnSe windows at Brewster’s angle to suppress two‑photon absorption and ensure that measured ionization originates from single‑photon processes. For each wavelength (in 50 nm steps) 470 exposure events were performed. Exposure times followed an exponential schedule, starting at 50 ms and increasing to provide dense sampling when the charge density changes most rapidly. The total measurement sequence for a single wavelength lasted about 9 hours.

To monitor the trapped‑charge density, the authors built a phase‑shifting Mach‑Zehnder interferometer using a low‑power 1064 nm probe beam (≈1 µW). The probe was linearly polarized at 45°, split into orthogonal polarizations, and recombined after traversing the crystal. A piezo‑mounted mirror introduced five phase steps (π/4 increments) and a CMOS camera recorded the resulting interferograms. An error‑compensating algorithm extracted the wrapped phase, which was then unwrapped and fitted to the quadratic phase profile expected from the Kerr effect: Δϕ(y)=−κL_z²n₀³(g₁₁−g₁₂)ρ²y². From the fitted curvature the absolute charge density ρ was obtained.

The measured charge‑decay curves showed a strong wavelength dependence. The photon fluence required to reduce the trapped charge by a factor of e (Fγ,1/e) decreased from ~5 × 10²⁶ photons cm⁻² at 700 nm to ~1 × 10²⁶ photons cm⁻² at 1300 nm, indicating that the single‑photon ionization cross‑section falls roughly as λ⁻⁴. Importantly, the decay was not well described by a single exponential; instead, a sum of two (or more) exponentials provided a significantly better fit. This multi‑exponential behavior suggests the presence of several trap species with distinct optical cross‑sections, and/or substantial recapture of freed electrons back into traps during the measurement interval.

To interpret these observations, the authors developed a one‑dimensional drift‑diffusion‑Poisson model that includes M distinct trap families, each capable of holding up to Ki electrons. The model accounts for photo‑ionization (cross‑section σ_i,k), electron capture (C_i,k), free‑electron drift under the self‑consistent electric field, and diffusion. Initial conditions assume uniform trap occupancy and negligible free carriers after the bias is removed. Numerical integration of the coupled equations reproduces the experimentally observed multi‑exponential decay and the wavelength‑dependent ionization rates, allowing extraction of plausible σ and C values for the dominant trap families.

The practical implications are clear. Because photo‑ionization rates drop dramatically beyond ~900 nm, KTN EODs can be operated in a “discontinuous charging” mode—alternating short DC charging pulses with unbiased AC scanning—without significant loss of trapped charge during the scanning window. This enables larger geometric scan angles, higher scan speeds, and eliminates the bias‑induced angular offset that limits aperture‑clipped systems. For deep‑tissue two‑photon imaging, where excitation wavelengths of 900–1100 nm are common, the measured ionization rates imply that a few‑millisecond scanning pulses will retain >95 % of the initial charge, preserving scan fidelity over many frames. Moreover, the identification of multiple trap species informs crystal growth and post‑processing strategies aimed at reducing the most photo‑sensitive traps, further extending the usable NIR range.

In summary, the paper delivers the first systematic, wavelength‑resolved quantification of NIR photo‑ionization in KTN electro‑optical deflectors, demonstrates that the process is strongly suppressed at longer NIR wavelengths, and provides a physics‑based model that captures multi‑trap dynamics and recapture effects. These results furnish designers of high‑speed, high‑resolution microscopy and photonic‑steering systems with concrete guidelines for selecting operating wavelengths, duty cycles, and charging protocols to maximize deflection range while minimizing charge loss. The work thus bridges a critical knowledge gap and paves the way for practical deployment of KTN‑based EODs in advanced biomedical imaging applications.