Fig. 1. Fabrication process and microscope characterization of the toroid resonators. (a) Fabrication process flow of large diameter toroid resonators. (b) 3D atomic force microscopy (AFM) scan on the surface of toroid. The RMS roughness and the correlation length are determined to be 0.3 nm and 30 nm, respectively. (c) Typical scanning electron microscopy (SEM) image of a toroid resonator with 2.8 mm diameter.

Fig. 2. Characterization of intrinsic quality factors (Q0). (a) Left column: linewidth measurement of a 23-GHz free spectral range (FSR) toroid resonator at 1560 nm. The upper (lower) panel presents a resonance transmission with its Lorentzian line shape fitting for the singlet (doublet) mode. The sinusoidal trace in each panel is a frequency calibration from a Mach–Zehnder interferometer (MZI) with a 201.21-kHz FSR. Right column: corresponding ringing measurement of the modes in left column. (b) Corresponding linewidth and ringing measurements of the same device in (a) at 1064 nm.

Fig. 3. The material absorption loss measurement. (a) Typical normalized transmission spectra of thermal triangle under different input powers. The spectra are obtained under a sufficiently slow laser frequency scanning speed of 20 MHz/s to prevent the distortion of the line shape. (b) Measured resonance frequency shift versus microresonator temperature offset at 1560 nm and 1064 nm with linear fittings. (c) Measured resonance frequency shift versus intracavity energy density for four modes at 1560 nm and 1064 nm. The identical slope of the linear fittings is expressed as δω0∼(α+g)ρ. (d) Material absorption-limited Q-factors (Qabs) of the four modes in (c) at 1560 nm (red) and 1064 nm (royal).

Fig. 4. Q factor and corresponding optical loss versus wavelength. (a) Scatter plot of Q-factors measured in five samples from 1480 nm to 1570 nm and from 1020 nm to 1070 nm. (b) Corresponding intrinsic loss (dB/km) in (a) and loss curves originating from different sources. The red and blue dashed lines represent the measured material absorption and simulated scattering loss, respectively, while the green dashed line represents the simulated water absorption loss, and the purple solid line represents a combination of these three losses.

Fig. 5. Measurement of parametric oscillation threshold and soliton generation. (a) Plot of parametric oscillation power versus input power (1560 nm). The oscillation threshold is seen at 31.9 μW. The inset presents a parametric oscillation spectrum at the input power of 74.3 μW. (b) Optical spectrum of a single soliton state with sech2 fitting (green solid line) at different pump power levels. The inset shows the transmission of the pump laser with a power of 220 μW.

Fig. 6. Resonance splitting. (a) Scatter plot of the splitting rate versus the measured intrinsic linewidth for two wavelength bands. Red dot and gray dot represent the wavelength bands of 1560 nm and 1064 nm, respectively. (b) Distribution of resonance splitting at two wavelength bands.

Fig. 7. The simulation of surface water absorption. (a) Mode profiles of the toroid resonator with a water layer. The diameters of the resonator and the ring core are 2.8 mm and 55 μm, respectively. (b) The simulated quality factor (Qwater) of toroid resonators with different water layer thicknesses versus wavelength. (c) Qwater of different mode families versus water layer thickness at 1560 nm and 1064 nm.