REVIEW 3 major objections 4 minor 53 references
A Quantum-Enhanced Feedback Oscillator
T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read Squeezed light, injected into an opto-electronic oscillator, suppresses its phase noise by about 1.0 dB—the first demonstration of quantum enhancement in a feedback oscillator.
desk verdict Genuine first demonstration of a quantum-enhanced feedback oscillator, but the headline 1.0 dB suppression is roughly twice what the paper's own eq. (57) predicts from its quoted parameters — the magnitude needs close scrutiny. 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 device is the opto-electronic oscillator loop: a laser field is amplitude-modulated by an electro-optic modulator, delayed by a fiber, detected by a photodetector, filtered, and fed back to the modulator, so the detected amplitude quadrature noise is converted into RF phase noise through square-law detection. The central identity is Eq. (2)/(57): the output phase-noise spectral density factorizes into the closed-loop response times an in-loop noise term proportional to the optical amplitude-quadrature spectrum incident on the photodetector plus detector dark noise. Because the quadrature spectrum is what the photodetector sees, replacing the vacuum on a loss port by squeezed vacuum—with the squeezed quadrature aligned to the amplitude quadrature—directly lowers that term. The paper also defines the standard quantum limit as the phase noise at unit quantum efficiency with vacuum input, Eq. (4), which makes the gap to the SQL a measure of how much room remains for quantum enhancement.
What would settle it
Re-measure the phase-noise spectrum while stepping a calibrated attenuator on the squeezed-light path; the suppression at 6-100 kHz should follow the model's prediction $10\log_{10}\left[(1-\eta_{\rm sqz}^2)(J_0^2+J_1^2)+2\eta_{\rm sqz}^2(J_0^2\bar S^{\rm sqz}_{qq}[\omega_0]+J_1^2\bar S^{\rm sqz}_{qq}[2\omega_0])\right]$ relative to vacuum, where $\eta_{\rm sqz}^2$ and the injected squeezing are measured independently by tomography. A systematic mismatch, or a suppression that does not grow when dark-noise clearance is raised by increasing optical power, would indicate unmodeled technical noise rather than a quantum effect.
Extended reading notes
Core claim
The paper's central claim is that a feedback oscillator's phase stability can be quantum-enhanced, and that the authors have built the first such oscillator. Concretely, the authors operate an opto-electronic oscillator at 9.51 MHz whose phase-noise power spectral density is accounted for by an ab-initio model with no free parameters, with photodetector dark noise and optical amplitude-quadrature quantum noise as the only significant sources. Injecting squeezed light through the OEO's loss port, with its squeezing axis aligned to the measured amplitude quadrature, suppresses the output phase-noise spectrum by 1.0 dB over 6-100 kHz; rotating the squeezing axis to anti-squeezing amplifies it by 2.8 dB, and interspersed vacuum measurements rule out drift. The paper takes this as the first experimental demonstration of quantum-enhanced phase stability in a feedback oscillator, and argues it establishes the principle for evading the standard quantum limit in masers and lasers.
Load-bearing premise
The phase noise is dominated by the modeled optical quantum noise plus photodetector dark noise, with laser intensity noise, fiber length fluctuations, electronics-chain noise, and backscatter from the squeezing path all far below them; quantum noise exceeds electronics noise by only about 0.8 dB, so any significant unmodeled technical noise would weaken the case that the 1.0 dB suppression is quantum in origin.
Editorial extensions
If this is right
- With lower losses along the squeezed-light path and a larger dark-noise clearance, the same OEO should reach and then pass its standard quantum limit.
- Any quantum-noise-limited feedback oscillator should be improvable by injecting squeezed light into its sensing port, giving a general path beyond the Schawlow-Townes limit for lasers and masers.
- The need for more than 60 dB of optical isolation between the squeezing source and the OEO means practical quantum-enhanced oscillators must manage backscatter and parasitic etalons.
- Quantum enhancement is only available after classical and technical noise have been pushed below the quantum floor, so certifying quantum-noise-limited operation is the prerequisite for any such oscillator.
Reading between the lines
- A testable extension follows from the paper's own loss numbers: with the reported ~45% loss and ~5 dB generated squeezing, the observed 1.0 dB suppression and 2.8 dB penalty are mutually consistent, so the same model predicts that cutting the loss in half should roughly double the suppression.
- The paper does not claim that the squeezing benefit extends below 1 kHz offset, where its data show environmental drift dominating; a natural next experiment would isolate the loop further and test whether quantum enhancement persists in the low-offset band most relevant for timing.
- Because the SQL expression scales as $\hbar\omega_\ell/(P_0\Omega^2\tau_g^2)$, the same injection technique should transfer to oscillators with longer delay lines or higher circulating power, where the un-enhanced floor is lower and the relative quantum gain could be larger.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper reports an opto-electronic oscillator (OEO) whose measured phase noise is described by an ab initio model with independently calibrated parameters. The oscillator is reported to operate within 3.5 dB of a standard quantum limit (SQL) defined by the authors. When squeezed vacuum is injected into the OEO loop, the phase noise is reported to fall by about 1.0 dB between 6 and 100 kHz offset; injecting anti-squeezed light increases it by about 2.8 dB. The authors claim this is the first experimental demonstration of a feedback oscillator whose phase stability is enhanced by quantum engineering.
Significance. If the claims hold, the result is a milestone: it would experimentally establish that engineering the quantum fluctuations of the field inside a feedback oscillator can improve its timing stability, going beyond the usual SQL discussion for lasers and masers. The paper's strengths are its explicit analytic model (Eqs. (2) and (57)), the independent calibration of the loop transfer function, detector responsivity, v_pi, and dark noise, and the use of coherent-control phase locking to set the squeezing angle. The anti-squeezing run is a valuable control because it demonstrates a quadrature-dependent effect on the output phase noise. The main weaknesses are statistical: the headline numbers are point estimates from a selected dataset, and the predicted model curves are not shown, making it difficult to verify the quantitative consistency between the model and the claimed 1.0 dB and 2.8 dB effects.
major comments (3)
- [Quantum-enhanced phase stability; Fig. 4; Fig. 9] The central quantitative claim lacks confidence intervals and rests on a dataset that is explicitly selected. The text states in the Measurement Procedure section that the dataset in Fig. 4 corresponds to the index highlighted in Fig. 9 and was 'verified to have exhibited both a stable average LO power ... and a very stable vacuum phase noise reference.' Because the effect is only 1.0 dB, the reader needs to know the run-to-run scatter of the suppression and the pre-defined selection criteria. Please report the suppression (and anti-squeezing increase) with standard errors for all datasets that used the narrow-bandwidth protocol, or justify the selection as pre-registered.
- [Quantum-enhanced phase stability; Eq. (57); Fig. 4] The model prediction is not shown on Fig. 4. The text states that 'the predicted amount of squeezing and anti-squeezing in the RF phase noise ... agrees well with the measurement data,' but the reader cannot verify this without the model curves and without the exact tomographic values V±, η_sqz^2, and α_SQZ used. Please add the model curves computed from Eq. (57) to Fig. 4, using the tomographic parameters with their uncertainties, and state the residual between model and data at the suppression band.
- [Setup Calibration; Eqs. (64)-(65)] The quoted 'roughly 45%' loss and 'roughly 5 dB' generated squeezing are not sufficient to verify that the same parameter set reproduces both the 1.0 dB suppression and the 2.8 dB anti-squeezing increase. Using Eq. (57) with the paper's vacuum normalization (S_vac = 1/2), η_sqz^2 = 0.55 and 5 dB squeezing (S_sqz = 0.158) predicts roughly 1.0 dB suppression and roughly 2.2 dB anti-squeezing increase; the reported 2.8 dB would require a slightly larger squeezing level or a different η_sqz^2. Please report the measured V± and the extracted η_sqz^2 and α_SQZ with uncertainties, and show that a single parameter set accounts for both observations within the stated errors.
minor comments (4)
- [Introduction; Fig. 2] There are typographical errors in the text, including 'the its open-loop response' and 'transimedance' in the Principle section; please proofread the manuscript.
- [Eq. (2)] The notation ¯S^opt_qq is not explicitly defined in the main text; please define the symmetrized PSD and clarify the factor of 2 in the dark-noise term so that it is consistent with the Supplementary derivation in Eq. (53).
- [Fig. 2(a) caption] The caption says 'blue dashed in a model'; this should read 'blue dashed curve is a model.' The figure would also benefit from error bars or a statement of the measurement uncertainty in the gain and phase data.
- [Setup Calibration] The text states that the laser's free-running intensity noise is 'at the quantum noise level at 9.5 MHz offset,' but no measurement of the laser RIN is shown; a reference to a supplementary figure or a brief calibration trace would help substantiate this assumption.
Circularity Check
No significant circularity: the phase-noise model is built from independent calibrations and the squeezing enhancement is predicted from independent tomography; self-citations are contextual, not load-bearing.
full rationale
The paper's central derivation is an ab-initio phase-noise model (eqs. 2 and 54-58) whose inputs—open-loop transfer function, EOAM half-wave voltage, photodetector quantum efficiency, dark-noise spectrum, and optical power—are all independently calibrated, with no free parameters. The vacuum phase-noise comparison in Fig. 3 is therefore a genuine test of the model, not a fit. The squeezed-light enhancement in Fig. 4 is likewise predicted from independent state tomography: eqs. (64)-(65) infer the squeezing-path efficiency and generated squeezing from measured variances V+ and V- at the in-loop photodetector, and those values, together with a frequency-dependent squeezing model from ref. [51], are used in eq. (57) to compute the expected phase noise. The phase-noise spectra used for the suppression/amplification claim are measured separately from the tomography, so the comparison is not circular. The self-citations [38,39] appear only as theoretical background for the existence of an SQL and for proposals to evade it; the OEO-specific SQL and noise model are derived in the present paper and cross-checked against the independent ref. [40]. No step in the derivation reduces by construction to its own input. Any quantitative discrepancy between the reported 1.0 dB suppression and the quoted dark-noise clearance, loss, and squeezing level would be a correctness or calibration concern, not a circularity, and no circular step can be identified from the manuscript text.
Assumptions & free parameters
free parameters (4)
- Photodetector responsivity R =
0.72 A/W
- EOAM half-wave voltage v_pi =
1.5 V
- Squeezed-light path transmission eta_sqz^2 =
about 0.55 (45% loss)
- Generated squeezing level alpha_SQZ =
about 5 dB at 9.5 MHz
assumptions (6)
- standard math Quadrature operators obey the canonical commutation relation [q(t), p(t')] = i delta(t - t').
- domain assumption The EOAM behaves as a variable beam splitter with amplitude transmissivity cos(pi v/(2 v_pi) + phi_bias).
- domain assumption The photodetector current is proportional to field intensity, with broadband transimpedance gain and no significant nonlinearity in the operating regime.
- domain assumption The RF filter rejects harmonics n*omega_0 for n != 1, and the oscillator operates in a single-mode steady state.
- domain assumption The squeezed state is phase-locked to the local oscillator with negligible residual phase noise, and the tomography model neglects LO phase noise.
- domain assumption All non-quantum, non-dark noise sources are sufficiently suppressed to be negligible.
Cite this review
Pith. "Pith review of A Quantum-Enhanced Feedback Oscillator." pith.science (2026). https://pith.science/paper/UDCBLPUW
@misc{pith2026260807753,
author = {Pith},
title = {Pith review of: A Quantum-Enhanced Feedback Oscillator},
year = {2026},
howpublished = {\url{https://pith.science/paper/UDCBLPUW}},
note = {Machine review of arXiv:2608.07753}
}
read the original abstract
Feedback oscillators, such as lasers and masers, serve as time references in modern computing, communication, and measurement. Quantum fluctuations ultimately limit their phase stability and ability to keep time precisely; in the absence of quantum engineering, their phase stability is bounded by a standard quantum limit (SQL). Techniques to improve the frequency stability of feedback oscillators beyond the SQL have been theorized, but have not yet been demonstrated. We demonstrate an opto-electronic oscillator (OEO), a type of feedback oscillator, with phase stability approaching the SQL. We then engineer the OEO's quantum state to improve its phase stability, thereby demonstrating the essential principle quantum-enhancement of feedback oscillators. Similar techniques may be employed to evade the SQL in other feedback oscillators such as masers and lasers.
Figures
Figures from the paper (6 more)
Reference graph
Works this paper leans on
-
[1]
A. D. Ludlow, M. M. Boyd, J. Ye, E. Peik, and P. O. Schmidt, Reviews of Modern Physics87, 637 (2015)
work page 2015
- [2]
-
[3]
I. M. Reid, Progress in Earth and Planetary Science2, 33 (2015)
work page 2015
-
[4]
Abichet al., Physical Review Letters123, 031101 (2019)
K. Abichet al., Physical Review Letters123, 031101 (2019)
work page 2019
- [5]
-
[6]
C. Clivati, R. Ambrosini, T. Artz, A. Bertarini, C. Bor- tolotti, M. Frittelli, F. Levi, A. Mura, G. Maccaferri, M. Nanni, M. Negusini, F. Perini, M. Roma, M. Stagni, M. Zucco, and D. Calonico, Scientific Reports7, 40992 (2017)
work page 2017
-
[7]
C. Cahillane, G. L. Mansell, and D. Sigg, Optics Express 29, 42144 (2021)
work page 2021
-
[8]
E. D. Caldwell, L. C. Sinclair, N. R. Newbury, and J.-D. Deschenes, Nature610, 667 (2022)
work page 2022
Show all 53 references
-
[9]
Br¨ uning, H
O. Br¨ uning, H. Gray, K. Klein, M. Lamont, M. Narain, R. Polifka, and L. Rossi, Reports on Progress in Physics 85(2022)
2022
-
[10]
W. C. Lindsey,Synchronization systems in communica- tion and control(Prentice-Hall, 1972)
1972
-
[11]
Xanthopoulos, ed.,Clocking in Modern VLSI Systems (Springer, 2009)
T. Xanthopoulos, ed.,Clocking in Modern VLSI Systems (Springer, 2009)
2009
-
[12]
Murmann, IEEE Solid-State Circuits Magazine7, 58 (2015)
B. Murmann, IEEE Solid-State Circuits Magazine7, 58 (2015)
2015
-
[13]
D. J. Wineland, C. Monroe, W. M. Itano, D. Leibfried, B. E. King, and D. M. Meekhof, Journal of Research of the National Institute of Standards and Technology103, 259 (1998)
1998
-
[14]
H. Ball, W. D. Oliver, and M. J. Biercuk, npj Quantum Information2, 16033 (2016)
2016
-
[15]
A. J. Viterbi,Principles of Coherent Communication (McGraw-Hill, 1966)
1966
-
[16]
Meyr and G
H. Meyr and G. Ascheid,Synchronization in digital com- munication(John Wiley, 1990)
1990
-
[17]
Kroupa,Phase Lock Loops and Frequency Synthesis (Wiley, 2003)
V. Kroupa,Phase Lock Loops and Frequency Synthesis (Wiley, 2003)
2003
-
[18]
Huygens,Horologium Oscillatorium: Sive de Motu Pendulorum ad Horologia Aptato Demonstrationes Ge- ometricae(F
C. Huygens,Horologium Oscillatorium: Sive de Motu Pendulorum ad Horologia Aptato Demonstrationes Ge- ometricae(F. Muguet, Parisiis, 1673)
-
[19]
W. G. Cady, Proceedings of the Institute of Radio Engi- neers10, 83 (1922)
1922
-
[20]
R. H. Varian and S. F. Varian, Journal of Applied Physics 10, 321 (1939)
1939
-
[21]
Boot and J
H. Boot and J. Randall, Journal of the Institution of Elec- trical Engineers - Part IIIA: Radiolocation93(1946)
1946
-
[22]
Essen and J
L. Essen and J. V. L. Parry, Nature176, 280 (1955)
1955
-
[23]
H. M. Goldenberg, D. Kleppner, and N. F. Ramsey, Phys. Rev. Lett.5, 361 (1960)
1960
-
[24]
E. H. Armstrong, Proceedings of the Institute of Radio Engineers24, 689 (1936)
1936
-
[25]
T. Udem, R. Holzwarth, and T. W. Hansch, Nature416, 233 (2002)
2002
-
[26]
J. M. Robinson, E. Oelker, W. R. Milner, W. Zhang, T. Legero, D. G. Matei, F. Riehle, U. Sterr, and J. Ye, Optica6, 240 (2019)
2019
-
[27]
A. E. Siegman and R. Arrathoon, Physical Review Let- ters20, 901 (1968)
1968
-
[28]
Fritschel, A
P. Fritschel, A. Jeffries, and T. J. Kane, Opt. Lett.14, 993 (1989)
1989
-
[29]
Benmessai, D
K. Benmessai, D. L. Creedon, M. E. Tobar, P.-Y. Bour- geois, Y. Kersal´ e, and V. Giordano, Phys. Rev. Lett.100, 233901 (2008)
2008
-
[30]
J. D. Breeze, E. Salvadori, J. Sathian, N. M. Alford, and C. W. M. Kay, Nature555, 493 (2018)
2018
-
[31]
H. C. Rawlings,The Science of Clocks and Watches (British Horological Institute, 1948)
1948
-
[32]
Hajimiri and T
A. Hajimiri and T. H. Lee,The Design of Low Noise Oscillators(Kluwer, 2003)
2003
-
[33]
A. L. Schawlow and C. H. Townes, Physical Review112, 1940 (1958)
1958
-
[34]
T. H. Maiman, Nature187, 493 (1960)
1960
-
[35]
R. E. Slusher, Rev. Mod. Phys.71, S471 (1999)
1999
-
[36]
X. S. Yao and L. Maleki, IEEE Journal of Quantum Elec- tronics32, 1141 (1996)
1996
-
[37]
Maleki, Nature Photonics5, 728 (2011)
L. Maleki, Nature Photonics5, 728 (2011)
2011
-
[38]
H. A. Loughlin and V. Sudhir, Nature Communications 14, 7083 (2023)
2023
-
[39]
Loughlin and V
H. Loughlin and V. Sudhir, Phys. Rev. A113, 043717 (2026)
2026
-
[40]
X. S. Yao and L. Maleki, JOSA B13, 1725 (1996). 6
1996
-
[41]
Romisch and A
S. Romisch and A. de Marchi, inProceedings of the 1999 Joint Meeting of the European Frequency and Time Fo- rum and the IEEE International Frequency Control Sym- posium (Cat. No.99CH36313), Vol. 2 (1999) pp. 1100– 1104 vol.2
1999
-
[42]
Romisch, J
S. Romisch, J. Kitching, E. Ferre-Pikal, L. Hollberg, and F. Walls, IEEE Transactions on Ultrasonics, Fer- roelectrics, and Frequency Control47, 1159 (2000)
2000
-
[43]
Vahlbruch, S
H. Vahlbruch, S. Chelkowski, B. Hage, A. Franzen, K. Danzmann, and R. Schnabel, Phys. Rev. Lett.97, 011101 (2006)
2006
-
[44]
Ganapathy, W
D. Ganapathy, W. Jia, M. Nakano, V. Xu, N. Aritomi, T. Cullen, N. Kijbunchoo, S. E. Dwyer, A. Mullavey, L. McCuller,et al.(LIGO O4 Detector Collaboration), Phys. Rev. X13, 041021 (2023)
2023
-
[45]
K. J. Blow, R. Loudon, S. J. D. Phoenix, and T. J. Shep- herd, Phys. Rev. A42, 4102 (1990)
1990
-
[46]
S. L. Danilishin and F. Y. Khalili, Living Reviews in Relativity15, 5 (2012)
2012
-
[47]
A. A. Clerk, M. H. Devoret, S. M. Girvin, F. Marquardt, and R. J. Schoelkopf, arXiv:0810.4729 (2010)
2010 arXiv
-
[48]
Sudhir,Quantum limits on measurement and control of a mechanical oscillator(Springer, 2017)
V. Sudhir,Quantum limits on measurement and control of a mechanical oscillator(Springer, 2017)
2017
-
[49]
A. F. Galiana,Development of Precision, Field- Deployable, Opto-Mechanical Instrumentation: Accessi- bility as a Functional Requirement, Ph.D. thesis, Mas- sachusetts Institute of Technology (2022)
2022
-
[50]
L. A. Bui, inFrontiers in Optics / Laser Science(OSA, Washington, DC, 2018) p. JW4A.83
2018
-
[51]
S. E. Dwyer,Quantum noise reduction using squeezed states in LIGO, Ph.D. thesis, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139 (2013)
2013
-
[52]
Ganapathy, V
D. Ganapathy, V. Xu, W. Jia, C. Whittle, M. Tse, L. Barsotti, M. Evans, and L. McCuller, Physical Review D105, 122005 (2022). Amplitude and phase modulation quadratures The magnitude of a quantized electromagnetic field with angular frequencyω ℓ, normalized to its mode profile...
2022
-
[53]
If saturation occurs primarily due to the modulator’s nonlinearity, we can model the OEO’s saturating behavior analytically
Generally, we will have|H OL[ω0]|>1 initially, and saturation effects will reduce the magnitude ofH OL until |HOL[ω0]|= 1. If saturation occurs primarily due to the modulator’s nonlinearity, we can model the OEO’s saturating behavior analytically. For oscillation amplitudes sa...
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.