Low noise resonant amplification by optical injection-locking and residual phase noise cancellation

Antoine Rolland; Brendan M. Heffernan; James Greenberg; Y. Lange Simmons

arxiv: 2604.18772 · v1 · submitted 2026-04-20 · ⚛️ physics.optics

Low noise resonant amplification by optical injection-locking and residual phase noise cancellation

Y. Lange Simmons , James Greenberg , Brendan M. Heffernan , Antoine Rolland This is my paper

Pith reviewed 2026-05-10 03:27 UTC · model grok-4.3

classification ⚛️ physics.optics

keywords optical injection lockingphase noise cancellationfeed-forward correctionresonant optical amplifierdiode laseroptical heterodyne detectionfrequency comb

0 comments

The pith

Combining injection locking with feed-forward phase correction in a diode laser reduces residual phase noise by up to 38 dB.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The authors demonstrate a high-gain resonant optical amplifier that preserves the low phase noise of a weak reference laser. They achieve this by first injection-locking a semiconductor diode laser to enforce phase coherence and then applying feed-forward correction to cancel the remaining phase errors, which are measured using optical heterodyne detection. This approach provides up to 38 dB of additional phase-noise suppression at frequencies above 1 kHz, even when the injection power is very low, and avoids adding amplified spontaneous emission noise.

Core claim

The central discovery is that residual phase noise in an injection-locked laser can be measured via heterodyne detection and canceled with feed-forward correction, yielding up to 38 dB phase noise reduction compared to injection locking alone for injection ratios down to -57 dB. This enables the use of commercial diode lasers as low-noise, high-power amplifiers for weak optical references such as frequency comb lines.

What carries the argument

The key mechanism is the optical heterodyne detection of residual phase error in the injection-locked state, followed by feed-forward phase correction to the amplified output.

If this is right

The method supports wavelength-agnostic amplification using standard diode lasers.
It enables ASE-free amplification of individual lines from optical frequency combs.
High-fidelity phase noise transfer is maintained at large gain and low injection ratios.
Phase noise is reduced by up to 38 dB at Fourier frequencies above 1 kHz.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

This technique could be extended to other types of lasers or amplification schemes where residual phase noise is a limiting factor.
Applications in precision optical metrology may benefit from the clean high-power outputs provided by this amplifier.
Further optimization of the correction bandwidth could potentially extend the noise reduction to lower Fourier frequencies.

Load-bearing premise

The residual phase error can be accurately measured and corrected in feed-forward without adding significant new noise or being restricted by the system's response time at the operating gains and injection levels.

What would settle it

Observing no improvement or an increase in phase noise above 1 kHz when the feed-forward correction is applied, compared to injection locking without correction, would falsify the effectiveness of the cancellation method.

Figures

Figures reproduced from arXiv: 2604.18772 by Antoine Rolland, Brendan M. Heffernan, James Greenberg, Y. Lange Simmons.

**Figure 1.** Figure 1: Schematic of the low-noise optical amplifier and phase-noise measurement setup. Light from a low-noise reference laser is attenuated and split, with one arm injected into a high-power diode laser through an optical circulator to achieve injection locking. The remaining reference light is frequency shifted by AOM 1 and heterodyned with a fraction of the injection-locked output on a balanced photodiode (BPD)… view at source ↗

**Figure 2.** Figure 2: Residual phase noise comparison of (a) the injection-locked oscillator alone, (b) including feed-forward correction and delay line compensation, (c) and feed-forward without delay line compensation. Each trace is labeled in (a) with the injection ratio in dB units. Free running noise of the laser diode is included in each plot. 20 0 20 40 Improvement (dB) (a) 10.0 Hz SMF No SMF (b) 100.0 Hz 20 0 20 40 Impr… view at source ↗

**Figure 3.** Figure 3: The change of residual phase noise versus injection ratio (horizontal axis inverted) of the feed-forward amplifier compared to injection-locking alone. Values are displayed for amplification with and without single mode fiber (SMF) delay compensation. Comparisons are performed at (a) 10 Hz, (b) 100 Hz, (c) 1 kHz, (d) 10 kHz, (e) 100 kHz, and (f) 1 MHz Fourier frequency. noise of the injection-locked diode … view at source ↗

**Figure 4.** Figure 4: Residual phase noise comparison of the low noise amplifier and a dual stage Erbium-dopde fiber amplifier (EDFA), which included spectral filtering to reduce ASE. The total EDFA gain was 55 dB and the low noise amplifier injection ratio shown is -57 dB (corresponding to a gain of 57 dB). from noise contributions in the correction chain, including the finite signal-to-noise ratio of the optical heterodyne … view at source ↗

read the original abstract

We demonstrate a low noise, high-gain, resonant optical amplifier that combines injection locking with feed-forward cancellation of residual phase noise. The wavelength-agnostic architecture uses a commercial semiconductor diode laser as a power amplifier while preserving the spectral purity of a weak reference. Although injection locking enforces phase coherence, finite residual phase noise within the locking regime limits high-fidelity transfer of low phase noise from the reference laser to the injection-locked laser, particularly at large gain. Here, the residual phase error is measured via optical heterodyne detection and canceled using feed-forward phase correction. Compared to injection locking alone, the amplifier achieves up to 38 dB phase-noise reduction at Fourier frequencies above 1 kHz for injection ratios down to -57 dB. This approach enables ASE-free amplification of low-power, low-noise optical references, including individual lines from optical frequency combs.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper demonstrates a working experimental combination of injection locking plus feed-forward phase correction that delivers clear phase-noise improvement in a diode-laser amplifier, though the low-injection-ratio regime still needs closer scrutiny on error-signal quality.

read the letter

The headline result is that adding real-time feed-forward cancellation to an injection-locked diode laser gives up to 38 dB less phase noise above 1 kHz compared with locking alone, and this holds down to injection ratios of -57 dB. The setup uses a commercial slave laser to amplify weak references while keeping the output ASE-free and phase-coherent with the master. That combination is the actual new piece here; both techniques exist separately, but the paper puts them together for resonant, wavelength-agnostic high-gain use on things like individual comb lines. The experimental spectra show the improvement directly, which is the strongest part of the work. They measure the residual phase error with optical heterodyne detection and subtract it via a phase modulator in the feed-forward path. This is straightforward and reproducible in principle, and the quantitative claim is stated clearly enough to be checked. The soft spot is exactly the one the stress-test flags. At -57 dB the injected field is only a few parts per million of the slave power, so the heterodyne beat note is weak. Any technical noise in the detector, electronics, or modulator can dominate the error signal and turn the correction into a noise source rather than a fix. The abstract gives no numbers on heterodyne SNR, loop delay, or actuator bandwidth, so it is not obvious that the cancellation stays effective in the regime where the gain is highest and the injected power is lowest. If the full data show clean error signals and no added noise floor, the result stands; otherwise the 38 dB figure is limited to higher injection ratios. This paper is for experimentalists in precision metrology and frequency-comb applications who need to amplify low-power references without degrading phase noise. A reader who already works with injection locking will see the practical extension and the measured spectra. It is worth sending to peer review because the experimental demonstration is concrete and the architecture is simple enough that others can test it, even if the noise-budget details need tightening.

Referee Report

1 major / 2 minor

Summary. The manuscript demonstrates an optical amplifier architecture that combines injection locking of a commercial semiconductor diode laser with feed-forward cancellation of residual phase noise, measured via optical heterodyne detection. The central experimental claim is up to 38 dB reduction in phase noise (Fourier frequencies >1 kHz) relative to injection locking alone, maintained down to injection ratios of -57 dB, enabling ASE-free high-gain amplification of weak low-noise references such as individual comb lines.

Significance. If the reported performance holds, the result is significant for precision optics and metrology: it provides a practical, wavelength-agnostic route to high-gain amplification while preserving the spectral purity of a low-power reference, addressing a known limitation of pure injection locking at large gain. The use of commercial components and direct experimental comparison to injection locking alone strengthens the practical value.

major comments (1)

[Abstract and §4] Abstract and §4 (results): The 38 dB improvement at injection ratios down to -57 dB rests on the feed-forward path successfully canceling residual phase error without adding noise. At -57 dB the injected field is only ~2e-6 of slave power, so heterodyne SNR, photodetector noise, electronics noise, and actuator bandwidth/delay become critical; the manuscript provides no quantitative bounds on these quantities, leaving the central claim vulnerable to the possibility that the correction path adds rather than subtracts noise in the reported regime.

minor comments (2)

[Figure captions and §3] Figure captions and §3 (setup): The heterodyne detection schematic and phase-noise spectra would benefit from explicit labels for the injection ratio values corresponding to each trace and from inclusion of the measured heterodyne beat-note power or SNR.
[§2] Notation: Injection ratio is reported in dB but the definition (power ratio or field ratio) should be stated once in §2 for clarity.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful reading of our manuscript and for highlighting the need for quantitative noise analysis in the feed-forward path. We address the major comment below and have revised the manuscript to strengthen the central claim.

read point-by-point responses

Referee: [Abstract and §4] Abstract and §4 (results): The 38 dB improvement at injection ratios down to -57 dB rests on the feed-forward path successfully canceling residual phase error without adding noise. At -57 dB the injected field is only ~2e-6 of slave power, so heterodyne SNR, photodetector noise, electronics noise, and actuator bandwidth/delay become critical; the manuscript provides no quantitative bounds on these quantities, leaving the central claim vulnerable to the possibility that the correction path adds rather than subtracts noise in the reported regime.

Authors: We agree that explicit quantitative bounds on the noise sources in the heterodyne detection and feed-forward electronics are necessary to fully substantiate the 38 dB reduction at injection ratios as low as -57 dB. The reported improvement is obtained from direct comparison of measured phase-noise spectra (injection locking alone versus with feed-forward active), and the feed-forward spectra lie below the injection-locking floor across the band of interest, which is only possible if the correction subtracts rather than adds noise. Nevertheless, to address the referee’s concern directly, the revised manuscript includes a new noise-budget subsection in §4. This analysis provides: (i) the heterodyne beat-note power and resulting SNR at -57 dB injection ratio, (ii) the measured and calculated contributions of photodetector shot noise and transimpedance-amplifier voltage noise referred to phase, (iii) the open-loop gain and delay of the feed-forward actuator (electro-optic modulator), and (iv) confirmation that the correction bandwidth exceeds 1 kHz while the added noise remains at least 10 dB below the residual phase error being canceled. These bounds demonstrate that the feed-forward path operates well above its own noise floor in the reported regime. revision: yes

Circularity Check

0 steps flagged

No circularity: experimental demonstration with direct measurements

full rationale

The paper reports an experimental result: measured phase-noise reduction of up to 38 dB via combined injection locking and feed-forward cancellation, quantified at specific injection ratios down to -57 dB. No derivation chain, first-principles prediction, or theoretical model is presented that reduces to fitted inputs, self-definitions, or self-citations. The central claim rests on heterodyne measurements and direct comparison to injection-locking alone, which are independent of any internal fitting or renaming. This is the expected non-finding for a measurement-focused optics paper.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim is an experimental demonstration that relies on standard laser physics and interferometry; no new free parameters, ad-hoc axioms, or invented entities are introduced in the abstract.

axioms (2)

domain assumption Injection locking enforces phase coherence between master and slave lasers
Standard result in laser physics invoked to explain the base locking behavior.
domain assumption Optical heterodyne detection can accurately measure residual phase error
Established interferometric technique used for the feed-forward measurement.

pith-pipeline@v0.9.0 · 5454 in / 1250 out tokens · 34475 ms · 2026-05-10T03:27:13.791885+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

[1]

X. Xie, R. Bouchand, D. Nicolodi,et al., Nat. Photon11, 44 (2017)

work page 2017
[2]

J. Liu, E. Lucas, A. S. Raja,et al., Nat. Photonics14, 486 (2020)

work page 2020
[3]

Hoghooghi, M

N. Hoghooghi, M. Mazur, N. Fontaine,et al., Optica12, 894 (2025)

work page 2025
[4]

M. L. Day, P . J. Low, B. White,et al., npj Quantum Inf8, 72 (2022)

work page 2022
[5]

Krinner, K

L. Krinner, K. Dietze, L. Pelzer,et al., Opt. Express32, 15912 (2024)

work page 2024
[6]

X. Fang, Y . Zhu, X. Cai,et al., Nat Commun15, 6339 (2024)

work page 2024
[7]

Chang, Y

Z. Chang, Y . Xu, J. Chen,et al., IEEE Commun. Surv. Tutorials27, 912 (2025)

work page 2025
[8]

Noise and gain of an erbium- doped fiber amplifier for delay-line optoelectronic oscillator,

K. Mikitchuk, A. Chizh, and S. Malyshev, “Noise and gain of an erbium- doped fiber amplifier for delay-line optoelectronic oscillator,” in2017 International Conference on Noise and Fluctuations (ICNF),(IEEE, Vilnius, Lithuania, 2017), pp. 1–4

work page 2017
[9]

Desurvire,Erbium-doped fiber amplifiers: principles and applications (Wiley-Interscience, Hoboken, N.J, 2002)

E. Desurvire,Erbium-doped fiber amplifiers: principles and applications (Wiley-Interscience, Hoboken, N.J, 2002)

work page 2002
[10]

Semiconductor Optical Amplifiers (SOAs),

M. J. Connelly, “Semiconductor Optical Amplifiers (SOAs),” inHand- book of Optoelectronic Device Modeling and Simulation,(ImprintCRC Press, 2017), 1st ed

work page 2017
[11]

Pedersen, Opt Quant Electron26, S273 (1994)

B. Pedersen, Opt Quant Electron26, S273 (1994)

work page 1994
[12]

F . M. Gardner,Phaselock techniques(Wiley-Interscience, Hoboken, N.J, 2005), 3rd ed

work page 2005
[13]

W. W. Chow, Opt. Lett.7, 417 (1982)

work page 1982
[14]

Vaughan,et al., IEEE Trans

Heng-Chia Chang, Xudong Cao, M. Vaughan,et al., IEEE Trans. Mi- crow. Theory Techn.45, 2035 (1997)

work page 2035
[15]

Razavi, IEEE J

B. Razavi, IEEE J. Solid-State Circuits39, 1415 (2004)

work page 2004
[16]

Chao, Z.-X

Y .-X. Chao, Z.-X. Hua, X.-H. Liang,et al., Optica11, 945 (2024)

work page 2024
[17]

L. Li, W. Huie, N. Chen,et al., Phys. Rev. Appl.18, 064005 (2022)

work page 2022
[18]

Liu and R

Z. Liu and R. Slavik, J. Light. Technol.38, 43 (2020)

work page 2020
[19]

Buczek, R

C. Buczek, R. Freiberg, and M. Skolnick, Proc. IEEE61, 1411 (1973)

work page 1973
[20]

B. M. Heffernan, J. Greenberg, T. Hori,et al., Nat. Photon.18, 1263 (2024)

work page 2024
[21]

Zhang, Y

Z. Zhang, Y . Dai, P . Ou,et al., IEEE Photonics J.8, 1 (2016)

work page 2016
[22]

Teyssieux, M

D. Teyssieux, M. Callejo, J. Millo,et al., J. Light. Technol.44, 222 (2026)

work page 2026
[23]

Schiemangk, S

M. Schiemangk, S. Spießberger, A. Wicht,et al., Appl. Opt.53, 7138 (2014)

work page 2014

[1] [1]

X. Xie, R. Bouchand, D. Nicolodi,et al., Nat. Photon11, 44 (2017)

work page 2017

[2] [2]

J. Liu, E. Lucas, A. S. Raja,et al., Nat. Photonics14, 486 (2020)

work page 2020

[3] [3]

Hoghooghi, M

N. Hoghooghi, M. Mazur, N. Fontaine,et al., Optica12, 894 (2025)

work page 2025

[4] [4]

M. L. Day, P . J. Low, B. White,et al., npj Quantum Inf8, 72 (2022)

work page 2022

[5] [5]

Krinner, K

L. Krinner, K. Dietze, L. Pelzer,et al., Opt. Express32, 15912 (2024)

work page 2024

[6] [6]

X. Fang, Y . Zhu, X. Cai,et al., Nat Commun15, 6339 (2024)

work page 2024

[7] [7]

Chang, Y

Z. Chang, Y . Xu, J. Chen,et al., IEEE Commun. Surv. Tutorials27, 912 (2025)

work page 2025

[8] [8]

Noise and gain of an erbium- doped fiber amplifier for delay-line optoelectronic oscillator,

K. Mikitchuk, A. Chizh, and S. Malyshev, “Noise and gain of an erbium- doped fiber amplifier for delay-line optoelectronic oscillator,” in2017 International Conference on Noise and Fluctuations (ICNF),(IEEE, Vilnius, Lithuania, 2017), pp. 1–4

work page 2017

[9] [9]

Desurvire,Erbium-doped fiber amplifiers: principles and applications (Wiley-Interscience, Hoboken, N.J, 2002)

E. Desurvire,Erbium-doped fiber amplifiers: principles and applications (Wiley-Interscience, Hoboken, N.J, 2002)

work page 2002

[10] [10]

Semiconductor Optical Amplifiers (SOAs),

M. J. Connelly, “Semiconductor Optical Amplifiers (SOAs),” inHand- book of Optoelectronic Device Modeling and Simulation,(ImprintCRC Press, 2017), 1st ed

work page 2017

[11] [11]

Pedersen, Opt Quant Electron26, S273 (1994)

B. Pedersen, Opt Quant Electron26, S273 (1994)

work page 1994

[12] [12]

F . M. Gardner,Phaselock techniques(Wiley-Interscience, Hoboken, N.J, 2005), 3rd ed

work page 2005

[13] [13]

W. W. Chow, Opt. Lett.7, 417 (1982)

work page 1982

[14] [14]

Vaughan,et al., IEEE Trans

Heng-Chia Chang, Xudong Cao, M. Vaughan,et al., IEEE Trans. Mi- crow. Theory Techn.45, 2035 (1997)

work page 2035

[15] [15]

Razavi, IEEE J

B. Razavi, IEEE J. Solid-State Circuits39, 1415 (2004)

work page 2004

[16] [16]

Chao, Z.-X

Y .-X. Chao, Z.-X. Hua, X.-H. Liang,et al., Optica11, 945 (2024)

work page 2024

[17] [17]

L. Li, W. Huie, N. Chen,et al., Phys. Rev. Appl.18, 064005 (2022)

work page 2022

[18] [18]

Liu and R

Z. Liu and R. Slavik, J. Light. Technol.38, 43 (2020)

work page 2020

[19] [19]

Buczek, R

C. Buczek, R. Freiberg, and M. Skolnick, Proc. IEEE61, 1411 (1973)

work page 1973

[20] [20]

B. M. Heffernan, J. Greenberg, T. Hori,et al., Nat. Photon.18, 1263 (2024)

work page 2024

[21] [21]

Zhang, Y

Z. Zhang, Y . Dai, P . Ou,et al., IEEE Photonics J.8, 1 (2016)

work page 2016

[22] [22]

Teyssieux, M

D. Teyssieux, M. Callejo, J. Millo,et al., J. Light. Technol.44, 222 (2026)

work page 2026

[23] [23]

Schiemangk, S

M. Schiemangk, S. Spießberger, A. Wicht,et al., Appl. Opt.53, 7138 (2014)

work page 2014