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

Geun Ho Ahn 1,2,† , Ziyu Wang 1,† , Devin J. Dean 1 , Hubert S. Stokowski 1 , Taewon Park 1,2 , Martin M. Fejer 1 , Jonathan Simon 1,3 , Amir H. Safavi-Naeini 1,∗ 1 Department of Applied Physics and Ginzton laboratory, Stanford University, Stanford, California 94305, USA 2 Department of Electrical Engineering, Stanford University, Stanford, California 94305, USA 2 Department of Physics, Stanford University, Stanford, California 94305, USA ∗ [email protected] · 2026

Read Ultrabroadband Passive Laser Noise Suppression to Quantum Noise Limit through on-chip Second Harmonic Generation by Geun Ho Ahn 1,2,† , Ziyu Wang 1,† , Devin J. Dean 1 , Hubert S. Stokowski 1 , Taewon Park 1,2 , Martin M. Fejer 1 , Jonathan Simon 1,3 , Amir H. Safavi-Naeini 1,∗ 1 Department of Applied Physics and Ginzton laboratory, Stanford University, Stanford, California 94305, USA 2 Department of Electrical Engineering, Stanford University, Stanford, California 94305, USA 2 Department of Physics, Stanford University, Stanford, California 94305, USA ∗ [email protected] (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-25801.txt.

Introduction

Ahn and co-authors demonstrate a chip device that calms laser intensity noise without any feedback loop. Laser noise limits quantum sensing, metrology, and computing. Feedback loops are slow, and optical resonators are narrow and need locking. Their device uses second-harmonic generation in nanophotonic lithium niobate waveguides. It runs at a pump-depletion point where input swings barely move the output. The authors report noise cuts of 25 to 60 decibels from DC past 10 gigahertz. A noisy fiber amplifier output was steadied to the shot-noise limit.

Catalog record

CREATORGeun Ho Ahn 1,2,† , Ziyu Wang 1,† , Devin J. Dean 1 , Hubert S. Stokowski 1 , Taewon Park 1,2 , Martin M. Fejer 1 , Jonathan Simon 1,3 , Amir H. Safavi-Naeini 1,∗ 1 Department of Applied Physics and Ginzton laboratory, Stanford University, Stanford, California 94305, USA 2 Department of Electrical Engineering, Stanford University, Stanford, California 94305, USA 2 Department of Physics, Stanford University, Stanford, California 94305, USA ∗ [email protected]
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE61150
LICENSEhttp://creativecommons.org/publicdomain/zero/1.0/
RIGHTSCC0
SOURCEarXiv

Source: https://arxiv.org/abs/2603.25801

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Ultrabroadband Passive Laser Noise Suppression to Quantum Noise Limit through on-chip Second Harmonic Generation Geun Ho Ahn 1,2,† , Ziyu Wang 1,† , Devin J. Dean 1 , Hubert S. Stokowski 1 , Taewon Park 1,2 , Martin M. Fejer 1 , Jonathan Simon 1,3 , Amir H. Safavi-Naeini 1,∗ 1 Department of Applied Physics and Ginzton laboratory, Stanford University, Stanford, California 94305, USA 2 Department of Electrical Engineering, Stanford University, Stanford, California 94305, USA 2 Department of Physics, Stanford University, Stanford, California 94305, USA ∗ [email protected] 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. Laser amplitude fluctuations are a pervasive limitation in precision sciences. Intensity noise reduces the signal-to-noise ratio in measurements, masking weak signals in optical sensing 8 , degrades atom-trapping lifetimes by parametric heating 10 ; 34 , and thereby limits performance in quantum computing 6 , metrology 23 ; 28 , communications and optical-links 1 ; 31 . In most precision applications, laser-intensity stabilization is employed to reduce intensity noise, thereby improving precision and performance. The key figures of merit for intensity-stabilization methods include the noise-suppression ratio, suppression bandwidth, insertion loss, and overall method complexity 12 . Active noise stabilizers based on electronic feedback loops driving electro-optic or acousto-optic modulators are widely used remedies, sensing instantaneous power and providing a correction to reduce excursions of the output power from the desired level 5 ; 24 . However, such active controls require high loop bandwidth, which poses challenges for broadband noise suppression 5 . Moreover, it is even more challenging to integrate the required discrete components of active stabilizers into a chip-scale platform. Engineering internal dynamics of a laser via injection locking to narrow bandwidth feedback or filtering with a high finesse resonator has been shown to reduce phase noise and moderately reduce intensity noise 32 ; 20 . However, the performance depends sensitively on the coupling conditions, back reflections, and resonator drift, and the high finesse of the resonator limits the achievable bandwidth 18 . Furthermore, the requirement to access and control the laser’s internal dynamics limits this approach to a relatively narrow class of devices. Optical nonlinearity offers a compelling route to intensity-noise reduction by exchanging noise between modes or dissipating it into an effective harmonic bath 27 ; 35 ; 2 ; 13 . While intracavity second harmonic generation (SHG) has demonstrated noise buffering 15 ; 4 , reliance on high-finesse cavities fundamentally restricts the suppression bandwidth and necessitates complex locking. Traveling-wave interactions in single-pass devices can overcome these limitations, offering passive, broadband suppression 17 . However, this approach requires exceptional nonlinear efficiency to operate at practical power levels. Integrated thin-film lithium niobate (TFLN) waveguides address this challenge, combining tight mode confinement with extended interaction lengths to enable highly efficient, scalable nonlinear devices 30 ; 7 ; 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 . Here, we demonstrate a passive and broadband noise-suppression method using a non-resonant SHG waveguide that is generalizable across wavelengths, power levels, and applications: At high conversion efficiency, with increasing input pump (FH) power, the depletion due to conversion to the SH increases, reaching a turning point where the output FH power begins to decrease 17 . Operating at this stationary point where the input fluctuations are eliminated to first order, the photonic intensity noise eater (PINE) achieves > 25 dB reduction in relative intensity noise (RIN), exhibits an ultra-wide suppression band ( > 10 GHz), and stabilizes the output of a noisy fiber amplifier to the standard quantum limit (SQL). To quantify the dependence on RF bandwidth, wavelength, and pump power, we develop new analytical and numerical sideband noise models for PINE that show excellent agreement with experiment. Compared with state-of-the-art active servo noise eaters, PINE eliminates feedback-loop bandwidth ceilings and actuator dynamic-range limits, remains compatible with a wide optical-power range and diverse wavelengths, and can be integrated upstream or downstream of other stabilization approaches. This stabilization technology, based on \chi^{(2)} , thus provides a new building block for intensity-stable laser systems suitable for atom- and ion- control, squeezed-light generation over audio-to-RF frequencies, and low-RIN sources for precision science and optical communications. I Theory of Operation In guided wave systems, the SHG process is described by : \displaystyle\frac{\text{d}A}{\text{d}z} \displaystyle=i\kappa A^{*}Be^{i\Delta kz} (1) \displaystyle\frac{\text{d}B}{\text{d}z} \displaystyle=i\kappa A^{2}e^{-i\Delta kz} where A and B are the amplitudes of the pump and second harmonic fields normalized such that |A|^{2} , and |B|^{2} are the power carried by the mode, \kappa with units of [\mathrm{W}^{-1/2}\mathrm{cm}^{-1}] is the second order nonlinear coefficient, and \Delta k=k_{B}-2k_{A}-2\pi/\Lambda is the quasi-phase mismatch 14 , where k_{A} and k_{B} are the propagation constants of the modes, and \Lambda is the period of the quasi-phase-matched (QPM) grating. For a phase matched interaction ( \Delta k=0 ), we analytically solve for A and B and find the corresponding relation between input and output powers: \displaystyle P_{p,\text{out}} \displaystyle=|A(L)|^{2}=P_{p,\text{in}}\sech^{2}(\kappa\sqrt{P_{p,\text{in}}}L) (2) \displaystyle P_{s,\text{out}} \displaystyle=|B(L)|^{2}=P_{p,\text{in}}\tanh^{2}(\kappa\sqrt{P_{p,\text{in}}}L) Using these solutions, we can relate the change in output pump power to a change in input pump power 17 : \displaystyle\frac{\partial P_{p,\text{out}}}{\partial P_{p,\text{in}}}=\sech^{2}\left(\zeta\right)\left[1-\zeta\tanh{\left(\zeta\right)}\right] (3) where \zeta=\kappa\sqrt{P_{p,\text{in}}}L . Equation ( 3 ) shows that input power fluctuations are suppressed to first order when \zeta\tanh{\zeta}=1 is satisfied 17 . This condition defines a critical pump power, P_{\text{in, c}} , at a critical value \zeta_{c}\approx 1.2 . At this operating point, the system reaches a conversion efficiency (CE) of \approx 70\% , derived from the general depletion formula P_{p,\text{out}}/P_{p,\text{in}}=1-\text{CE}=\sech^{2}(\zeta) . Consequently, the passive stabilization comes with a fixed intrinsic insertion loss of -5.2 dB, a universal value

Read the complete text