• Winkler, Georg
• Prinz, Maximilian
• Truong, Gar-Wing
• Nguyen, Catherine
• Follman, David
• Bailey, D. Michelle
• Heckl, Oliver H.
• Fleisher, Adam J.
• Zhao, Gang
• Cole, Garrett D.
• Perner, Lukas W.

Abstract

Precision spectroscopy, time-frequency metrology, stable optical reference signal generation, interferometry and other uses of optical cavities generally benefit from mirror coatings with vanishingly low excess loss (absorption A and scatter S). In the near-infrared and visible wavelengths, A+S&lt;1 ppm and cavity finesses F&gt;1 000 000 [1] have already been achieved. However, using traditional physical vapor deposition techniques, the lowest reported mid-infrared (MIR) A+S is ~30 ppm (with transmittance T of ~120 ppm) [2], resulting in F≈20900 and on-resonance cavity transmission of ~64% near 4.5 μm.<br/>We use a novel crystalline coating technique [3] to produce distributed Bragg reflectors (DBR) at 4.54 μm with A+S&lt;10 ppm [4]. A refined fabrication procedure allows us to minimize A+S by growing the DBR as two “half mirrors”, which are later bonded to form a 34.5 period GaAs/AlGaAs mirror. This buries growth defects in the middle of the DBR, thereby reducing excess losses. These wafers are then processed to 12 mm discs and transfer-bonded to super-polished, concave silicon (Si) optical substrates.<br/>As summarized in Fig. 1(a), we measured a total single-mirror loss L=151±3 ppm and T=144±2 ppm (where L=T+A+S=1-R, with reflectance R) via two independent cavity-ringdown experiments and two independent techniques for isolating T [4]. As S is negligible for crystalline MIR DBRs, we infer A=7±4 ppm, which we confirmed by photothermal common-path interferometry. This results in a finesse F=20 805±413 and a cavity transmittance &gt;92%. This is key for future optical resonators with F&gt;100 000 in the MIR, which can now be achieved by an increase in DBR period count (reducing T).<br/>Fabrication of such DBRs with more periods requires precise knowledge and control of many parameters, among them the MIR refractive index n of GaAs/AlGaAs. We used an as-grown half mirror to probe n of both materials simultaneously by means of a transfer-matrix model (TMM), based on precision data obtained from Fourier-transform spectrometry (FTS) and cross-sectional scanning electron microscopy (SEM). Modelling both indexes as single effective oscillators [5], uncertainty propagation via a Monte Carlo method suggests ~10-3 uncertainties over a wide wavelength range, see Fig. 1(b).<br/>Figure 1:(a) Total loss 1-R as measured in two CRD setups at NIST and CDL, fitted with polynomials; design transmittance T (dashed line); T via modelling pinned to measured coating parameters (red line); direct T measurement (triangles) (b) GaAs/AlGaAs refractive indices, via TMM fit based on precision FTS and SEM.<br/>References:<br/>[1] G. Rempe et al., Measurement of ultralow losses in an optical interferometer. Opt. Lett. 17, 363 (1992).<br/>[2] M. G. Delli Santi et al., Biogenic fraction determination in fuel blends by laser-based 14CO2 detection. Adv. Phot. Res. 2, 2000069 (2021).<br/>[3] G. D. Cole et al., High-performance near- and mid-infrared crystalline coatings. Optica 3, 647–656 (2016).<br/>[4] G. Winkler et al., Mid-infrared interference coatings with excess optical loss below 10 ppm. Optica 8, 686–696 (2021).<br/>[5] M. A. Afromowitz, Refractive Index of Ga1-xAlxAs. Solid State Comm. 15, 59–63 (1974)

Topics

• impedance spectroscopy
• scanning electron microscopy
• experiment
• physical vapor deposition
• Silicon
• defect
• Monte Carlo method
• spectrometry
• interferometry