REVIEW 2 major objections 6 minor 38 references
Ultra-narrow linewidth light generation based on an optoelectronic oscillator
T0 review · 2 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read An optoelectronic oscillator built around a high-finesse cavity and a short RF link directly generates light with a 0.23-Hz integrated linewidth and −100 dBc/Hz phase noise at 1 kHz.
desk verdict A well-modeled OEO-locking scheme for narrow-linewidth light, but the headline linewidth and phase-noise numbers are measured on an unvalidated frequency-shifted pump surrogate, not on the cavity-stored light. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing element is the phase-noise redistribution identity $\varphi_{\mathrm{RF}}(t) = \frac{\tau_O}{\tau_{\mathrm{RF}}+\tau_O}\varphi_{\mathrm{PL}}(t)$ and $\varphi_O(t) = \frac{\tau_{\mathrm{RF}}}{\tau_{\mathrm{RF}}+\tau_O}\varphi_{\mathrm{PL}}(t)$, where $\tau_O$ is the group delay of the Fabry–Pérot cavity at resonance, $\tau_{\mathrm{RF}}$ is the delay of the optoelectronic link, and $\varphi_{\mathrm{PL}}$ is the pump laser's phase fluctuation. It follows from the Barkhausen phase condition for the composite optoelectronic cavity and the relation $\omega_{\mathrm{RF}} = |\omega_{\mathrm{PL}} - \omega_O|$. When the cavity delay dominates the link delay, the RF carries almost all of the pump's phase noise and the optical oscillation is nearly immune to it; the RF is then used both to frequency-shift the pump for a low-noise output and to drive a feedback loop that locks the pump. The equivalent RF bandpass filter created by PM-IM conversion (same window as the FP cavity) provides self-alignment of the optical oscillation to the cavity resonance and single-frequency operation in both domains.
What would settle it
Improve the cavity coupling enough to collect the transmitted optical oscillation directly with usable power, and measure its phase noise and beat linewidth against an independent reference; if the directly extracted light shows higher phase noise than the frequency-shifted pump replica at any offset below 10 kHz, or a beat linewidth clearly above 0.23 Hz, the central claim that the OEO generates the reported low-noise optical oscillation would be falsified.
Extended reading notes
Core claim
The central claim is that a PM-IM OEO with a high-quality optical resonator and a minimized optoelectronic-link delay acts as an optoelectronic frequency converter: the pump laser's phase fluctuations are almost entirely transferred to the RF oscillation, while the optical oscillation inside the cavity is left with only a tiny residual ($\tau_{\mathrm{RF}}/\tau_{\mathrm{RT}}$) of the pump noise. The RF oscillation therefore provides a faithful, fast copy of the pump's phase noise, which the authors use to phase-lock the pump to a reference. With two such systems, the free-running frequency-shifted pump already reaches a 5-Hz linewidth, and with feedback the estimated individual linewidth is 0.23 Hz, with phase noise reaching the cavity thermal-noise floor below 1 kHz. A direct comparison on the same cavities shows the OEO-locked configuration beats PDH by 26 dB at 1 kHz and 39 dB at 10 kHz offsets.
Load-bearing premise
The headline linewidth and phase-noise values are measured on a frequency-shifted copy of the pump laser, not on the light actually stored in the cavity, and the paper's conclusions assume that this copy faithfully reproduces the optical oscillation's noise performance.
Editorial extensions
If this is right
- A few-kilohertz-linewidth pump laser can be converted into sub-hertz-integrated-linewidth light without a gain medium in the reference cavity, reaching the passive cavity's thermal noise floor at low offsets.
- The OEO-locking scheme suppresses pump phase noise by roughly the cavity-to-link delay ratio (about 74 dB here) plus feedback, and outperforms PDH on the same cavities by 26 dB at 1 kHz and 39 dB at 10 kHz offsets.
- Shortening the optoelectronic link further increases both the suppression (up to $20\log_{10}(\text{finesse})$, exceeding 120 dB for a finesse of $10^6$) and the pump-frequency drift tolerance, easing the demands on the pump.
- Since the OEO is a voltage-controlled RF oscillator, chip-scale phase-locked loops can be added to remove residual pump noise, pointing toward compact integrated low-noise light sources.
- The measured $1.73 \times 10^{-15}$ frequency instability at 1 s approaches the thermal noise limit of the FP cavity, making the scheme applicable to optical clocks and precision metrology.
Reading between the lines
- Inference: the reported 0.23-Hz linewidth and −100 dBc/Hz are measured on the frequency-shifted pump replica, which the authors state carries extra high-offset noise; light extracted directly from the cavity transmission should be even quieter at high offsets once coupling efficiency is improved.
- Inference: the mechanism is not limited to Fabry–Pérot cavities; any high-Q resonator with large group delay relative to the link delay (whispering-gallery, integrated microresonators) should exhibit the same noise redistribution, so the approach could transfer to chip-scale platforms.
- Inference: the RF output is itself a fast, wideband discriminator of the pump frequency against the cavity resonance; this could be exploited as a self-calibrating frequency reference or a simple optical-frequency readout without separate PDH electronics.
- Inference: a decisive test of the replica assumption would be to split the OEO-locked light and heterodyne it against a third independent ultra-stable laser whose noise is far below the claimed levels; if the beat linewidth is not $\le 0.23$ Hz, the estimate would need revision.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes that a phase-modulation-to-intensity-modulation optoelectronic oscillator (PM-IM OEO) with a high-finesse Fabry–Pérot (FP) cavity and a deliberately shortened optoelectronic link can act as an optoelectronic frequency converter, directly generating low-noise, narrow-linewidth light while also producing an RF signal that replicates the pump laser's phase fluctuations. This RF signal is then used in a feedback loop (called OEO-locking) to stabilize the pump, further suppressing residual noise. The central theoretical result is Eq. (1), which predicts that the pump's phase noise is split between the optical and RF oscillations with coefficients tau_O/(tau_RF+tau_O) and tau_RF/(tau_RF+tau_O), respectively; with tau_O >> tau_RF, the optical oscillation inherits almost none of the pump noise. The authors report a phase-noise suppression of -73.5 dB compared to the free-running pump, close to the predicted -74.1 dB, and claim an outstanding phase noise of -100 dBc/Hz at 1 kHz offset and an integrated linewidth of 0.23 Hz after activating OEO-locking. They also compare OEO-locking with conventional Pound–Drever–Hall (PDH) locking, reporting superior phase noise and frequency stability. However, all headline noise and linewidth values are measured on a frequency-shifted version of the pump laser (FS-PL), not on the optical oscillation stored in the FP cavity, and the paper explicitly notes that this is a workaround for low cavity transmission coupling.
Significance. If the claimed performance is correct, this is a conceptually new and practically interesting method for generating ultra-narrow-linewidth light. The theoretical framework is plausible and well-developed in the supplementary information, and the measured suppression factor matches the prediction, giving confidence in the underlying mechanism. The OEO-locking scheme, which leverages the RF replica for feedback, is an innovative twist that may offer broadband suppression beyond what PDH can achieve. The paper is also candid about the FS-PL workaround, which is a strength. Nevertheless, the experimental support for the central 'direct generation' claim rests on the unverified assumption that the FS-PL faithfully represents the optical oscillation inside the cavity; the paper would be significantly strengthened by direct validation or a clear, quantified statement of the surrogate's limitations.
major comments (2)
- [Implementation and Measurement; Fig. 5] The headline phase noise of -100 dBc/Hz at 1 kHz and the integrated linewidth of 0.23 Hz are measured on the frequency-shifted pump (FS-PL), not on the optical oscillation inside the FP cavity. The paper states that 'collecting the low-noise optical oscillation with sufficient power from the cavity transmission poses a challenge due to the low coupling efficiency... As a workaround, we use the generated RFs to frequency-shift the corresponding PLs.' This surrogate is never directly verified against the cavity-stored light. The equivalence requires (i) the AOM shift to subtract exactly the RF phase fluctuations from the PL phase and (ii) the AOM/driver chain to add negligible phase noise at the measured offsets; neither condition is experimentally demonstrated. The measured -73.5 dB suppression (Fig. 5A) validates the predicted redistribution for the replica, but it does not establish that the replica's absolute noise equals that of the optical oscillation. To support the claim of 'direct generation of low-noise light,' the authors should either measure the cavity transmission (even at low power) or characterize the AOM chain's added phase noise and show explicitly that the FS-PL noise spectrum matches the predicted optical-oscillation spectrum across the offsets that contribute to the 0.23 Hz linewidth.
- [Data measurement; Fig. 4C] The individual linewidth of 0.23 Hz for each OEO-locked FS-PL is inferred from a 0.33 Hz beat between two OEO-locked FS-PLs by assuming equal noise contributions from the two systems. This equal-partition assumption is not validated; if one system is noisier than the other, the individual linewidths will differ. The linewidth is extracted from a single 8-second FFT record at 0.125 Hz RBW with no repeated measurements or uncertainty analysis. Since 'integrated linewidth as narrow as 0.23 Hz' is a headline claim in the abstract, the authors should either measure against an independent reference (e.g., a third cavity-stabilized laser) or explicitly present 0.23 Hz as an estimate under the equal-partition assumption, with a bound on the error.
minor comments (6)
- [Abstract and Discussion] The phrase 'direct generation of low-noise light' overstates what is measured; the output characterized in the experiments is the frequency-shifted pump (FS-PL), not the cavity-stored field. Please qualify this wording, e.g., 'directly generates a low-noise optical oscillation, which we characterize via a frequency-shifted replica.'
- [Implementation and Measurement] The claim that 'the 5-Hz linewidth is the narrowest value reported to date in free-running optical oscillators' needs clarification: if 'free-running' is meant to exclude passive self-injection locking, this should be stated explicitly, since self-injection-locked lasers can have narrower linewidths without active electronic feedback.
- [Fig. 4B] The traces in Fig. 4B are difficult to distinguish; please label each trace more clearly and include a legend or explicit annotation for the PL, RF, and FS-PL frequency fluctuations.
- [Fig. 5B] The text refers to 'modified Allen deviation' but the correct term is 'modified Allan deviation' (two l's); please correct this in the main text and supplementary.
- [Supplementary, Eq. S6] The equations in Section II of the supplementary are typeset in a garbled fashion in the submitted version; please ensure they are rendered cleanly so that the transfer functions are legible.
- [Data measurement] In the measurement of the free-running PL linewidth (top of Fig. 4C), the beat is taken against an OEO-locked FS-PL assuming the OEO-locked FS-PL's contribution is negligible; this assumption should be stated in the main text as well as in the methods.
Circularity Check
No significant circularity: the suppression formula follows from the Barkhausen phase condition and measured cavity delay, not from the output; the main caveat is that headline measurements are on a frequency-shifted replica, which is a validation gap rather than a circular derivation.
full rationale
The central derivation, Eq. (1), is obtained from the Barkhausen roundtrip phase condition and the frequency relation ω_RF = |ω_PL − ω_O|, with τ_RF (optoelectronic link delay) and τ_O (FP group delay) as independent physical parameters. The predicted suppression factor 20log10(τ_RF/(τ_RF+τ_O)) = −74.1 dB is not fitted to the measured −73.5 dB or to the 0.23 Hz linewidth; it depends only on the measured 6.26-kHz cavity linewidth and the roughly 10-ns link delay. The measured frequency-suppression factor of about 1/5000 likewise matches τ_RF/τ_RT with no free parameter. The intrinsic-noise floor is taken from the standard OEO treatment (Ref. 28), and the PDH comparison is a transfer-function simulation with stated parameters, not a fit; the comparison does not reduce to the claimed result. The self-citations (Refs. 26–30, 38) support components such as PM-IM conversion, OEO noise modeling, and cavity mounting, but they are not used to define the new suppression result. The significant caveat is flagged in the paper itself: 'collecting the low-noise optical oscillation with sufficient power from the cavity transmission poses a challenge due to the low coupling efficiency... As a workaround, we use the generated RFs to frequency-shift the corresponding PLs, thereby obtaining the low-noise, narrow-linewidth lights that share similar noise performance with the corresponding optical oscillations...' Because the FS-PL is synthesized by subtracting the RF, which itself carries τ_O/(τ_RT) of the PL phase noise, from the PL, its measured suppression is a consistency check of the same phase-noise partition rather than an independent measurement of the intra-cavity optical oscillation. This limits the strength of the 'direct generation' claim and the attribution of the 0.23-Hz linewidth, but it is not a circular derivation: the suppression formula is not defined in terms of the measured linewidth, and no fitted parameter is renamed as a prediction. Hence the circularity score is low.
Assumptions & free parameters
free parameters (1)
- optoelectronic-link delay tau_RF =
about 10 ns
assumptions (3)
- domain assumption The FP cavity can be modeled as lossless and symmetric, and the leakage-field transfer function shares the cavity bandwidth and group delay (SI I, Eqs. S1-S2).
- domain assumption The Barkhausen phase condition applies, with roundtrip phase equal to 2N pi, leading to Eq. (1) for phase-noise redistribution (SI II and SI IV).
- ad hoc to paper The frequency-shifted pump (FS-PL) replicates the noise of the optical oscillation except at high offsets.
Cite this review
Pith. "Pith review of Ultra-narrow linewidth light generation based on an optoelectronic oscillator." pith.science (2026). https://pith.science/paper/3I6FW3DN
@misc{pith2026241202490,
author = {Pith},
title = {Pith review of: Ultra-narrow linewidth light generation based on an optoelectronic oscillator},
year = {2026},
howpublished = {\url{https://pith.science/paper/3I6FW3DN}},
note = {Machine review of arXiv:2412.02490}
}
read the original abstract
Narrow-linewidth light sources are essential for both fundamental research and various technological applications, yet they are challenging to generate directly due to instabilities in active laser cavities and the misalignment between closely spaced resonator modes and broad gain bandwidth. In this study, we demonstrate the direct generation of low-noise light from an optoelectronic oscillator (OEO). By minimizing the delay in the optoelectronic link and employing a high-quality optical resonator, an OEO can simultaneously generate both a low-phase-noise optical oscillation that is resilient to phase fluctuations of its pump laser, and a radio frequency (RF) oscillation that captures these fluctuations. We leverage these RF-domain fluctuations to implement a high-performance feedback loop, which stabilizes the pump laser and significantly enhances the performance of optical oscillation. This approach achieves an outstanding phase noise level of -100 dBc/Hz at a 1 kHz offset and an integrated linewidth as narrow as 0.23 Hz. The integration of optoelectronic oscillation and feedback loop provides significant broadband noise suppression compared to conventional schemes for generating narrow-linewidth light, paving the way for developments in fields such as coherent optical communications, atomic spectroscopy, metrology, and quantum optics.
Figures
Reference graph
Works this paper leans on
-
[1]
Kikuchi, Fundamentals of Coherent Optical Fiber Communications
K. Kikuchi, Fundamentals of Coherent Optical Fiber Communications. Journal of Lightwave Technology 34, 157-179 (2016)
work page 2016
-
[2]
Juddbrian, Operator Techniques in Atomic Spectroscopy
R. Juddbrian, Operator Techniques in Atomic Spectroscopy . (Operator Techniques in Atomic Spectroscopy, 1998)
work page 1998
-
[3]
J. Aasi et al., Advanced ligo. Classical and quantum gravity 32, 074001 (2015)
work page 2015
-
[4]
K. J. Gå svik, Optical Metrology, Third Edition . (Optical Metrology, Third Edition, 2003)
work page 2003
-
[5]
A. D. Ludlow, M. M. Boyd, J. Ye, E. Peik, P. O. Schmidt, Optical atomic clocks. Reviews of Modern Physics 87, 637-701 (2015)
work page 2015
-
[6]
Hinkley et al., An atomic clock with 10 –18 instability
N. Hinkley et al., An atomic clock with 10 –18 instability. Science 341, 1215- 1218 (2013)
work page 2013
-
[7]
C. C. Gerry, P. L. Knight, Introductory quantum optics. (Cambridge university press, 2023)
work page 2023
-
[8]
Svelto, Orazio, Principles of Lasers. (Principles of Lasers, 2010)
work page 2010
Show all 38 references
-
[9]
Kessler et al., A sub-40-mHz-linewidth laser based on a silicon single-crystal optical cavity
T. Kessler et al., A sub-40-mHz-linewidth laser based on a silicon single-crystal optical cavity. Nature Photonics 6, 687-692 (2012). 9 / 18
2012
-
[10]
Jin et al
N. Jin et al. , Micro -fabricated mirrors with finesse exceeding one million. Optica 9, 965-970 (2022)
2022
-
[11]
Liang et al
W. Liang et al. , Whispering -gallery-mode-resonator-based ultranarrow linewidth external -cavity semiconductor laser. Optics letters 35, 2822 -2824 (2010)
2010
-
[12]
Kondratiev et al., Self-injection locking of a laser diode to a high -Q WGM microresonator
N. Kondratiev et al., Self-injection locking of a laser diode to a high -Q WGM microresonator. Optics Express 25, 28167-28178 (2017)
2017
-
[13]
Wu et al
L. Wu et al. , Greater than one billion Q factor for on -chip microresonators. Optics Letters 45, 5129-5131 (2020)
2020
-
[14]
M. W. Puckett et al., 422 Million intrinsic quality factor planar integrated all - waveguide resonator with sub-MHz linewidth. Nature communications 12, 934 (2021)
2021
-
[15]
Kessler, T
T. Kessler, T. Legero, U. Sterr, Thermal noise in optical cavities revisited. Journal of the Optical Society of America B 29, 178-184 (2011)
2011
-
[16]
Y. I. Khanin, Principles of laser dynamics. (Newnes, 2012)
2012
-
[17]
Li et al
B. Li et al. , Reaching fiber -laser coherence in integrated photonics. Optics Letters 46, 5201-5204 (2021)
2021
-
[18]
Dahmani, L
B. Dahmani, L. Hollberg, R. Drullinger, Frequency stabilization of semiconductor lasers by resonant optical feedback. Optics letters 12, 876-878 (1987)
1987
-
[19]
Cheng et al., Harnessing micro -Fabry-Perot reference cavities in photonic integrated circuits
H. Cheng et al., Harnessing micro -Fabry-Perot reference cavities in photonic integrated circuits. arXiv preprint arXiv:2410.01095, (2024)
2024 arXiv
-
[20]
Liang, Y
W. Liang, Y. Liu, Compact sub -hertz linewidth laser enabled by self -injection lock to a sub-milliliter FP cavity. Optics Letters 48, 1323-1326 (2023)
2023
-
[21]
Henry, Theory of the linewidth of semiconductor lasers
C. Henry, Theory of the linewidth of semiconductor lasers. IEEE Journal of Quantum Electronics 18, 259-264 (1982)
1982
-
[22]
Fleming, A
M. Fleming, A. Mooradian, Spectral characteristics of external -cavity controlled semiconductor lasers. IEEE Journal of Quantum Electronics 17, 44- 59 (1981)
1981
-
[23]
R. W. Drever et al., Laser phase and frequency stabilization using an optical resonator. Applied Physics B 31, 97-105 (1983)
1983
-
[24]
Matei et al
D. Matei et al. , 1.5 μ m lasers with sub -10 mHz linewidth. Physical review letters 118, 263202 (2017). 10 / 18
2017
-
[25]
E. D. Black, An introduction to Pound –Drever–Hall laser frequency stabilization. American journal of physics 69, 79-87 (2001)
2001
-
[26]
A. Wolf, B. Bodermann, H. Telle, Diode laser frequency-noise suppression by> 50 dB by use of electro -optic parametric master oscillators. Optics Letters 25, 1098-1100 (2000)
2000
-
[27]
A. Wolf, H. Telle, Generation of coherent optical radiation by electronic means:? the electro-optical parametric oscillator. Optics letters 23, 1775-1777 (1998)
1998
-
[28]
X. S. Yao, L. Maleki, Optoelectronic microwave oscillator. JOSA B 13, 1725- 1735 (1996)
1996
-
[29]
W. Li, J. Yao, A wideband frequency tunable optoelectronic oscillator incorporating a tunable microwave photonic filter based on phase -modulation to intensity-modulation conversion using a phase -shifted fiber Bragg grating. IEEE Transactions on Microwave Theory and Technique...
2012
-
[30]
Hao et al., Breaking the limitation of mode building time in an optoelectronic oscillator
T. Hao et al., Breaking the limitation of mode building time in an optoelectronic oscillator. Nature communications 9, 1839 (2018)
2018
-
[31]
W. Li, M. Li, J. Yao, A narrow -passband and frequency -tunable microwave photonic filter based on phase -modulation to intensity -modulation conversion using a phase -shifted fiber Bragg grating. IEEE Transactions on Microwave Theory and Techniques 60, 1287-1296 (2012)
2012
-
[32]
Rubiola, Phase noise and frequency stability in oscillators
E. Rubiola, Phase noise and frequency stability in oscillators . (Cambridge University Press, 2008)
2008
-
[33]
M. J. Madou, Fundamentals of microfabrication: the science of miniaturization. (CRC press, 2018)
2018
-
[34]
D. J. Blumenthal, Photonic integration for UV to IR applications. APL Photonics 5, (2020)
2020
-
[35]
Lu et al., Emerging integrated laser technologies in the visible and short near- infrared regimes
X. Lu et al., Emerging integrated laser technologies in the visible and short near- infrared regimes. Nature Photonics, 1-14 (2024)
2024
-
[36]
Zhou et al., Prospects and applications of on-chip lasers
Z. Zhou et al., Prospects and applications of on-chip lasers. Elight 3, 1 (2023)
2023
-
[37]
Collins, Integrated PLLs and VCOs for Wireless Applications
I. Collins, Integrated PLLs and VCOs for Wireless Applications. Radio Electronics, (2010)
2010
-
[38]
Jiao et al
D. Jiao et al. , Highly vibration -resistant sub -Hertz ultra -stable laser passing over 1700 km transport test. Infrared Physics & Technology 130, 104608 (2023). 11 / 18 Methods FP cavity fabrication: Two perpendicular FP cavities are positioned horizontally in a cubic optica...
2023
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.