Abstract
Magnetophotonic crystals (MPCs) often utilize a microcavity structure where a thin magnetic garnet film—bismuth-substituted yttrium iron garnet (Bi:YIG)—is sandwiched between two Bragg mirrors (BMs) made of $Ta_2O_5$ and $SiO_2$. For MPCs fabricated by sputtering, annealing of as-sputtered Bi:YIG films is essential for garnet crystallization. However, $Ta_2O_5$ is also crystallized at garnet crystallization temperatures when utilizing conventional annealing in an electric furnace. $Ta_2O_5$ crystallization worsens optical parameters of BMs and consequently properties of an MPC. To overcome this problem, we selectively annealed as-sputtered Bi:YIG/BM samples by a high-power laser. Experiments showed that Bi:YIG can be crystallized while keeping $Ta_2O_5$ in the amorphous state.
Introduction: The Challenge of Global Heating
Magnetophotonic crystals are critical components in advanced optical devices like spatial light modulators. Their performance relies on the high magneto-optical activity of the central Bi:YIG layer and the high reflectivity of the surrounding Bragg mirrors. However, the standard fabrication process involves sputtering, which leaves the Bi:YIG in a non-magnetic amorphous state.
The traditional solution—heating the entire sample in an electric furnace at 600–800°C—presents a significant drawback. At these temperatures, the $Ta_2O_5$ layers in the mirrors also undergo crystallization. This creates grain boundaries that scatter light, introducing optical noise and reducing the efficiency of the MPC. The researchers needed a way to heat the garnet without heating the mirror.
Methods: Selective Thermal Processing
The researchers proposed selective laser annealing to heat the garnet layer locally without affecting the dielectric mirrors. This selectivity is achieved by choosing a laser wavelength (532 nm) that exploits the differences in optical absorption between the materials.
As shown in the absorption spectra, Bi:YIG exhibits a high absorption coefficient at 532 nm, whereas $Ta_2O_5$ and $SiO_2$ are virtually transparent. Consequently, when the multilayer stack is illuminated, the laser energy is primarily deposited into the Bi:YIG layer. The experimental setup utilized a continuous-wave (CW) laser mounted on a 3-axis stage, allowing for precise control over power density (20–140 $W/cm^2$) and scan speed (10–500 µm/s).
Key Results: Finding the "Goldilocks" Zone
The effectiveness of the laser annealing was evaluated by measuring the Faraday rotation of the processed films. The study identified a specific range of parameters required to achieve optimal crystallization without damaging the material.
The data indicates that Faraday rotation—a proxy for successful crystallization and magnetic performance—peaks at a power density of approximately 53 $W/cm^2$. At this point, the rotation reaches roughly -3.4 deg/µm, outperforming typical furnace-annealed samples (-2.8 deg/µm). However, increasing the power density beyond 100 $W/cm^2$ leads to a sharp decline in performance, likely due to thermal degradation or ablation of the garnet layer.
Structural analysis via X-ray diffraction (XRD) and Transmission Electron Microscopy (TEM) confirmed these findings. While the Bi:YIG layer showed clear crystalline peaks, the $Ta_2O_5$ layers remained amorphous, effectively eliminating the light-scattering issues associated with furnace annealing.
Conclusion and Future Outlook
Selective laser annealing provides a viable pathway for fabricating high-quality MPCs by decoupling the thermal requirements of the magnetic layer from the stability of the dielectric mirrors. While the current use of a Gaussian beam profile introduces some spatial non-uniformity in crystallization, future work involving beam shaping and Finite Element Method (FEM) thermal modeling aims to refine the process. This technique holds promise for the development of high-performance Magneto-Optical Spatial Light Modulators (MOSLMs) for next-generation display and sensing technologies.
Summary updated