Ultrabroadband Passive Laser Noise Suppression to Quantum Noise Limit through on-chip Second Harmonic Generation

Abstract

Laser intensity noise limits performance in quantum sensing, metrology, and computing. Existing stabilization methods face a trade-off between bandwidth and complexity: electronic feedback loops are speed-limited, while optical resonators are constrained by narrow linewidths and locking requirements. Here, we demonstrate an all-optical "noise eater" that passively suppresses intensity fluctuations from DC to >10 gigahertz. By leveraging high-efficiency second-harmonic generation in nanophotonic lithium niobate waveguides, we operate at a pump-depletion stationary point where input fluctuations are decoupled from the output to first order. This passive and nonresonant nanophotonic device suppresses relative intensity noise by 25 to 60 dB over the full measurement bandwidth and stabilizes a noisy fiber amplifier output to the shot-noise limit. Our results establish a scalable, wide-bandwidth paradigm for laser stabilization essential for high-throughput quantum technologies and deployable photonic sensing systems.

Figures (22)

• Figure 1: Theory of operation of PINE.a. Conceptual schematic showing the operation of the nonlinear optical noise eater, where a noisy input beam is coupled into PPTFLN and the noise on the output pump power is strongly suppressed at the critical point. The alternating colors of the PPTFLN conceptually represent inverted domains. b. Simulated normalized first harmonic (FH) output power, which corresponds to output pump, and conversion efficiency (CE) into second harmonic (SH) as a function of normalized pump input power. At the CE $\approx 70\%$, a stationary point in normalized output power is observed. c. An image of experimental setup with PINE chip, where input and output are coupled through lensed fibers. This $10$ mm $\times$$10$ mm chip has 20 PINE devices, and each PINE section has a footprint around $0.05$ mm $\times$$10$ mm. The inset image shows an SHG microscope image of periodically poled domains. The scale bar in the inset represents 30 $\mu m$.
• Figure 2: Highly efficient integrated second harmonic generators.a. Normalized second harmonic generation intensity, Norm. $P_{s,\text{out}}$, (top panel) and normalized pump output, Norm. $P_{p,\text{out}}$, (bottom panel) as a function of pump wavelength detuning. The pump laser wavelength is swept around 1555.6 nm and the output pump (green) and second harmonic (blue) intensities measured, with a moving-average applied to the solid line, and the raw data plotted in fainter colors. For pump output scans, data from three different pump powers ($71.1$ mW, $137$ mW, $232$ mW) are plotted. At higher pump power, we observe lower normalized transmission at QPM wavelength, indicating greater pump depletion. The transfer function scan for SH was measured with $\approx 20$ mW of pump power on chip. b. Experimentally measured pump output power and CE into second harmonic as a function of input pump power at the QPM wavelength, indicated by the gray box in (a). The CE is estimated using pump depletion at QPM wavelength, following CE = $(P_{p, \text{in}} - P_{p,\text{out}})/P_{p, \text{in}}$. The dotted lines show analytical estimation of CE and pump depletion using the experimentally derived $\eta$ and $L$ = 10 mm
• Figure 3: Ultra-Broadband Intensity Noise Reduction.a. Experimental NRR over the broadband RF frequency of the VNA indicated by the scatter points. The measurement was performed with injected RIN peak values of -38.7 dBc/Hz (navey blue), -48.7 dBc/Hz (light blue), and -58.7 dBc/Hz (light green).
• Figure 4: Operation windows of PINE.a. A heatmap of simulated noise reduction ratio with respect to input pump power and pump wavelength detuning using experimentally derived $\eta$ and $L$ values. Three white lines indicate line slices at different input power levels. b. Simulated (solid line) and measured (scatter points) NRR dependence on wavelength detuning from the QPM wavelength at three different pump powers (104 mW, 112 mW, 116 mW). The measured NRR are taken at 1 GHz RF frequency. c. $P_{\text{p, in}}$ vs. $P_{\text{p, out}}$ dependence on different poling length on a single PINE chip. Faded scatter points indicate experimentally measured $P_{\text{p, out}}$ data. Inset showing the chip with different lengths of poling electrodes. The black dashed line indicates the critical power line that can be achieved by different poling lengths, and intersecting points with analytical $P_{\text{p, out}}$ solutions of the three different lengths are indicated by solid scatter points. d. Temperature tuning of pump depletion at critical power, where the normalized FH intensity is $\approx 0.3$ at the QPM wavelength. The wavelength window shifts approximately linearly with the temperature of PINE chip.
• Figure 5: Application of PINE approaching the quantum limit.a. A simplified schematic of the measurement set up for detecting shot noise from PINE. The setup is kept identical for measurements of EDFA and EDFA+PINE with an exception of bypassing PINE Chip (setup details shown in Fig. \ref{['Sfig:Shot_Noise_Setup']}). A fiber wavelength de-multiplexer (WDM) is used to separate FH and SH light. The FH output of WDM still has noticeable SH unfiltered. Therefore, a fiber bandpass filter (BPF) with 10 nm bandwidth is employed to filter the residual SH. Importantly, both measurements of EDFA only and EDFA with PINE see this BPF. Filtered FH power is connected to an output VOA. b. Optical power dependence of PSD at the offset frequency of 100 MHz for EDFA and EDFA+PINE. The scatter points represent measured electrical PSD calibrated with averaged value of RF amplifier response over 1000 measurements, and the error bars represent the measured standard deviation. The navy scatter and fit line represent electrical PSD as a function of detected optical power for EDFA output alone, and the light green scatter and fit line represent corresponding data for EDFA output after PINE. The shaded region represents shot noise estimated at the given optical power.