REVIEW 3 major objections 4 minor 34 references
AOM-based ultra-low noise laser intensity control up to the MHz range
T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash
Pith's one-line read By pairing two feedback loops with a feedforward branch on acousto-optic modulators, this paper achieves -155 dB/Hz relative intensity noise at hundreds of kHz and about 1.5 MHz control bandwidth.
desk verdict Credible AOM-based intensity control pushes bandwidth into the MHz range, but the headline RIN sits only 3 dB above the detector floor—so the absolute number is an upper bound until cross-checked. 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 core object is the two-branch control topology built around two AOMs: a slow DC branch with a digital PI controller and gain-scheduling linearization on a double-pass AOM handles average power and low-frequency drift; a fast AC branch combines an analog PID feedback signal with a feedforward signal on a second AOM, whose zeroth-order transmitted beam is used as the output. The feedforward branch measures noise before a fiber delay line and applies an inverted correction to the same actuator as the feedback, while the beam is focused to a ~50 µm waist near the acoustic transducer to cut the AOM response to 65 ns. The load-bearing relations are the AOM power-transfer response P_out/P_in ∝
What would settle it
Measure the controlled output with a second, independent low-noise detector chain (different photodiode and amplifier) while simultaneously recording beam pointing after the AOM; if the measured floor shifts by more than a few dB or tracks the substitute detector's own noise floor, the -155 dB/Hz level was not the true output RIN.
Extended reading notes
Core claim
The paper's central claim is that an acousto-optic modulator, long considered too slow for high-frequency intensity stabilization, can be made to suppress laser relative intensity noise up to the MHz range. The authors build a two-stage controller: a slow digital feedback loop on a double-pass AOM sets and holds the average power, while a fast analog loop and a feedforward branch act together on a second AOM. The feedforward path picks off the noise before a 27 m fiber delay and applies an inverted correction, giving it a head start; the feedback path cleans what remains. Optimizing the beam through the second AOM—a ~50 µm waist close to the acoustic transducer—cuts the actuator response to
Load-bearing premise
The claimed -155 dB/Hz floor rests on the out-of-loop photodiode PD-OUT faithfully measuring the output intensity noise, with no correlated artifact from beam pointing, polarization drift, amplifier noise, or photodiode nonlinearity hiding or creating that level.
Editorial extensions
If this is right
- AOM-based intensity control can reach about 1.5 MHz closed-loop bandwidth, with total noise reduction at or above 29 dB up to 700 kHz.
- The feedforward branch delivers 23 dB of suppression at 700 kHz, so feedforward, not feedback alone, is the dominant high-frequency noise killer.
- The RIN spectrum stays below -145 dB/Hz up to 1.2 MHz and at -155 dB/Hz in the 10-200 kHz band, meeting the demands of optical lattice experiments at wavelengths where 1064 nm commercial lasers are not usable.
- The system can ramp the average output from zero to maximum in tens of milliseconds, and the shown instability on fast falling edges is a fixable control-signal saturation issue.
- Photodiodes with lower electronic noise and shot noise could lower the achievable RIN below -155 dB/Hz.
Reading between the lines
- Because the architecture uses only free-space AOMs and a fiber delay, the same controller should transfer directly to other wavelengths and multi-watt powers, limited mainly by AOM damage thresholds and fiber choice; the paper demonstrates 170 mW at 841 nm but states the scaling rationale.
- The feedforward branch's 23 dB suppression at 700 kHz suggests that a digitally adjustable delay line, instead of a fixed SMA cable, could tune the feedforward timing in situ and push the usable bandwidth even closer to the AOM's 65 ns response limit.
- If the -155 dB/Hz floor is confirmed by an independent detector, the practical noise bottleneck moves to the photodiode; using lower-gain, higher-power detectors could approach the -158 dB/Hz estimate or lower.
- For quantum-gas experiments, this single platform could replace the common pair of a slow AOM power servo and a separate fast EOM noise eater, simplifying the optical path and increasing transmission, at the cost of the added complexity of the combined controller.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper reports an AOM-based laser intensity stabilization system that combines a slow DC feedback branch with a fast AC branch implementing both feedback and feedforward control on a second AOM. The authors report a relative intensity noise (RIN) suppression down to -155 dB/Hz from 10 kHz to 200 kHz, with the spectrum remaining below -145 dB/Hz up to 1.2 MHz and a claimed closed-loop control bandwidth of approximately 1.5 MHz. They also characterize a gain-scheduling linearization of the DC branch and demonstrate dynamic intensity ramps with 10-50 ms time scales. The measurement relies on an out-of-loop photodiode (PD-OUT) whose estimated noise floor is -158 dB/Hz, only 3 dB below the claimed RIN level.
Significance. If the headline RIN value is robust, this is a significant advance for AOM-based intensity control, extending the usable bandwidth beyond the typical few hundred kHz and combining two feedback loops with feedforward on a single actuator. The paper is careful in several respects: it provides a noise-floor model for the photodiodes, measures step responses of the AOM and summing amplifier, releases data on Zenodo, and compares spectra with and without feedforward. However, the central quantitative claim of -155 dB/Hz rests on a single out-of-loop photodiode whose estimated floor is only 3 dB below the reported value, and no independent verification of the detector floor is presented. This gap must be addressed before the absolute RIN claim can be accepted.
major comments (3)
- [Section 4 / Methods 8.3] The claimed RIN floor of -155 dB/Hz is only 3 dB above the estimated PD-OUT floor of -158 dB/Hz (S_min = 2(S_el + S_sn), Methods 8.3). The paper itself attributes the 3 dB discrepancy to electronic noise from the summing amplifier or other components, which is exactly the mechanism that would raise the actual detector floor to -155 dB/Hz. No independent out-of-loop verification is shown: there is no calibrated intensity-modulation injection, no second photodiode in the same beam, and no cross-spectral measurement. As the headline result is the absolute RIN level, the measurement must distinguish optical RIN from detector and electronics noise. I recommend adding an in-situ calibration of the measurement floor, e.g., by injecting a known intensity modulation and measuring the photodiode response, or by performing a two-detector cross-spectral measurement, and reporting the resulting confi
- [Section 2] PD-OUT and PD-FB2 are placed in the two output ports of the final HWP-PBS (Figure 1a). If the control loop reduces common-mode intensity noise but differential polarization or beam-pointing fluctuations affect the transmitted and reflected ports unequally, the out-of-loop port may not represent the actual stabilized output. This is a load-bearing assumption for the absolute RIN measurement. The manuscript does not quantify the polarization or pointing stability of the output beam or compare the two ports simultaneously. A simple test would be to swap the roles of the two photodiodes or to record the RIN from both ports concurrently and show that they agree within the claimed accuracy.
- [Section 4 / Figure 3] Figure 3 shows RIN spectra without error bars or repeated measurements. Given that the spectrum flattens at -155 dB/Hz, only 3 dB above the estimated detector floor, the lack of statistical uncertainty is particularly limiting. The flattening could be a detector-floor artifact rather than an optical noise floor. Please provide multiple traces or a statistical confidence interval, and preferably an in-situ floor measurement as described above, so that the reader can assess whether the -155 dB/Hz level is real optical RIN.
minor comments (4)
- [Section 4] The statement 'the closed-loop control bandwidth is approximately 1.5 MHz, with the servo bump located at around 4.8 MHz' needs a clear definition. Is this the unity-gain frequency, the -3 dB suppression point, or inferred from the step response? The step response in Methods 8.1 measures the AOM-plus-photodiode response, not the closed-loop transfer function. Please clarify how the bandwidth is determined.
- [Figure 3] Adding a legend to the figure would improve readability. The colors are described in the caption, but a direct label on the figure would help.
- [Methods 8.3] Equation (2) defines S_sn(f) = 2e/I_p, which is white noise and does not depend on frequency. The notation S_sn(f) with a frequency argument is slightly misleading; consider writing S_sn = 2e/I_p and noting that it is frequency-independent.
- [Throughout] There are minor typographical issues such as '0 th' instead of '0th' and 'P opt' inconsistently formatted. Also, 'F ALC' appears without a clear abbreviation expansion; consider defining it at first use.
Circularity Check
No circularity: the headline RIN is a measured spectrum, not a derived or fitted prediction; stated detector-floor comparison is a limitation, not an input.
full rationale
The paper reports an empirical laser-intensity-noise measurement. The central claim (-155 dB/Hz RIN) is a measured spectrum from PD-OUT (Fig. 3 and Sec. 4), not derived from a fitted model. The feedforward gain K and 7 m cable delay are manually tuned for best suppression, but the resulting noise spectrum is measured, not computed from those settings. The only quoted 'estimate' is S_min = 2(S_el + S_sn) from ref [29], used as a post-hoc comparison: the paper explicitly states that the measured floor is 3 dB above this estimate and attributes it to summing-amplifier electronic noise (Sec. 4). This is a candid measurement-fidelity limitation (could PD-OUT be detector-limited at -155 dB/Hz?), which is a correctness risk, not circularity: the measured spectrum is not equal by construction to S_min, and the claim is not inferred from S_min. Same-author citations [8] and [23] are contextual (tune-out wavelength and dataset availability), not load-bearing. No equation defines the result in terms of its own inputs, no fitted parameter is renamed as a prediction, and no uniqueness theorem or ansatz is imported from prior work. Thus no circular step exists.
Assumptions & free parameters
free parameters (5)
- Feedforward gain K =
0.01 to 1 (continuous)
- Bias voltage V_b =
160 mV
- Beam waist in AC-AOM =
50 µm
- Feedforward cable length =
7 m
- PI gains (DC branch) =
G_P=0.02, G_I=60000 (optimized); defaults G_P=0, G_I=3000
assumptions (4)
- standard math AOM diffraction efficiency follows P_out/P_in proportional to sin^2(sqrt(P_RF/P_sat)) (Eq. 1).
- standard math Shot noise formula S_sn = 2e/I_p (Eq. 2).
- domain assumption The intensity noise measured by PD-FF before the fiber delay is representative of the noise to be corrected at the output.
- domain assumption The out-of-loop PD-OUT measurement is not corrupted by electronic or optical artifacts at the claimed RIN level.
Cite this review
Pith. "Pith review of AOM-based ultra-low noise laser intensity control up to the MHz range." pith.science (2026). https://pith.science/paper/5XHIYHUP
@misc{pith2026260803547,
author = {Pith},
title = {Pith review of: AOM-based ultra-low noise laser intensity control up to the MHz range},
year = {2026},
howpublished = {\url{https://pith.science/paper/5XHIYHUP}},
note = {Machine review of arXiv:2608.03547}
}
abstract
High-power laser sources that exhibit low relative intensity noise and allow simultaneous dynamic control of their light level are required for a broad range of applications in various fields of physics. Acousto-optic modulators (AOMs) are widely used for active power stabilization and regulation due to their simple drive electronics requirements and high optical power handling capability in free-space. However, the rather slow propagation speed of the sound wave within the AOM crystal typically limits their control bandwidth to a few hundred kHz. In this work, we present a novel AOM-based control system that is capable of significantly suppressing intensity noise of high-power lasers up to the MHz range. By combining two standard feedback loops with one feedforward control branch and optimizing the beam path in the AOM crystal, ultra-low relative intensity noise levels down to $-155\, \text{dB}\,\text{Hz}^{-1}$ even at several hundred kHz are achieved. Our results are relevant for applications that require ultra-low intensity noise at Fourier frequencies up to the MHz range, such as optical lattice experiments with light ultracold atoms.
Reference graph
Works this paper leans on
-
[32]
Optics Express 28(21), 31209 (2020) https://doi.org/10.1364/OE.405002 16
Wang, Y., Wang, K., Fenton, E.F., Lin, Y.-W., Ni, K.-K., Hood, J.D.: Reduction of laser intensity noise over 1 MHz band for single atom trapping. Optics Express 28(21), 31209 (2020) https://doi.org/10.1364/OE.405002 16
-
[1]
Physical Review A 94(6), 063632 (2016) https://doi.org/10.1103/PhysRevA.94.063632
Anisimovas, E., Raˇ ci¯ unas, M., Str¨ ater, C., Eckardt, A., Spielman, I.B., Juzeli¯ unas, G.: Semisynthetic zigzag optical lattice for ultracold bosons. Physical Review A 94(6), 063632 (2016) https://doi.org/10.1103/PhysRevA.94.063632
-
[2]
Physical Review A92(2), 021402 (2015) https://doi.org/10.1103/PhysRevA.92.021402
Blatt, S., Mazurenko, A., Parsons, M.F., Chiu, C.S., Huber, F., Greiner, M.: Low- noise optical lattices for ultracold Li 6. Physical Review A92(2), 021402 (2015) https://doi.org/10.1103/PhysRevA.92.021402
-
[3]
Classical and Quantum Gravity 22(20), 4279–4283 (2005) https://doi.org/10.1088/0264-9381/22/20/009
Barr, B.W., Strain, K.A., Killow, C.J.: Laser amplitude stabilization for advanced interferometric gravitational wave detectors. Classical and Quantum Gravity 22(20), 4279–4283 (2005) https://doi.org/10.1088/0264-9381/22/20/009
-
[4]
Physical Review Applied23(1), 011005 (2025) https: //doi.org/10.1103/PhysRevApplied.23.L011005
Chao, Y.-X., Hua, Z.-X., Liang, X.-H., Yue, Z.-P., Jia, C., You, L., Tey, M.K.: Robust suppression of high-frequency laser phase noise by adaptive Pound- Drever-Hall feedforward. Physical Review Applied23(1), 011005 (2025) https: //doi.org/10.1103/PhysRevApplied.23.L011005
-
[5]
Optics Express23(25), 32481 (2015) https: //doi.org/10.1364/OE.23.032481
Duan, L., Fang, J., Li, R., Jiang, L., Ding, M., Wang, W.: Light intensity stabilization based on the second harmonic of the photoelastic modulator detec- tion in the atomic magnetometer. Optics Express23(25), 32481 (2015) https: //doi.org/10.1364/OE.23.032481
-
[6]
Review of Scientific Instruments76(6), 063112 (2005) https://doi.org/10.1063/1.1930095
Donley, E.A., Heavner, T.P., Levi, F., Tataw, M.O., Jefferts, S.R.: Double-pass acousto-optic modulator system. Review of Scientific Instruments76(6), 063112 (2005) https://doi.org/10.1063/1.1930095
-
[7]
npj Quantum Information8(1), 72 (2022) https://doi
Day, M.L., Low, P.J., White, B., Islam, R., Senko, C.: Limits on atomic qubit control from laser noise. npj Quantum Information8(1), 72 (2022) https://doi. org/10.1038/s41534-022-00586-4
Show all 34 references
-
[8]
Physical Review A112(5), 051304 (2025) https://doi.org/10.1103/rykg-qn6s
De Martino, A., Kiesel, F., Auch, J., Karpov, K., Gross, C.: Dissipationless tune- out trapping for a lanthanide–alkali-metal quantum gas mixture. Physical Review A112(5), 051304 (2025) https://doi.org/10.1103/rykg-qn6s
2025 doi
-
[9]
Physical Review Research8(2), 023327 (2026) https://doi.org/10.1103/48zr-2t1p
Dawel, F., Pelzer, L., Dietze, K., Kramer, J., Hild, M., King, S.A., Spethmann, N.C.H., Klose, J., Stahl, K., D¨ orscher, S., Benkler, E., Lisdat, C., Porsev, S.G., Safronova, M.S., Schmidt, P.O.: High-stability optical clock based on a continu- ously ground-state cooled Al + ...
2026 doi
-
[10]
Optics Express26(5), 5275 (2018) https://doi.org/10.1364/OE.26.005275 14
Dong, J., Zhang, L., Jiang, H., Yang, X., Pan, W., Cui, S., Gu, X., Feng, Y.: High order cascaded Raman random fiber laser with high spectral purity. Optics Express26(5), 5275 (2018) https://doi.org/10.1364/OE.26.005275 14
2018 doi
-
[11]
Physical Review A58(5), 3914–3921 (1998) https: //doi.org/10.1103/PhysRevA.58.3914
Gehm, M.E., O’Hara, K.M., Savard, T.A., Thomas, J.E.: Dynamics of noise- induced heating in atom traps. Physical Review A58(5), 3914–3921 (1998) https: //doi.org/10.1103/PhysRevA.58.3914
1998 doi
-
[12]
Version Number: 1 (2026)
Holman, A., Sun, X., Seo, B., Corn, J., Zhu, Z., Xu, Y., Wu, J., Yu, N., Filin, D., Safronova, M., Will, S.: A Mid-Infrared Platform Based on Strontium Tweezer Arrays. Version Number: 1 (2026). https://doi.org/arXiv:2606.02560
2026 arXiv
-
[13]
Physical Review A107(4), 042611 (2023) https://doi.org/10.1103/PhysRevA.107.042611
Jiang, X., Scott, J., Friesen, M., Saffman, M.: Sensitivity of quantum gate fidelity to laser phase and intensity noise. Physical Review A107(4), 042611 (2023) https://doi.org/10.1103/PhysRevA.107.042611
2023 doi
-
[14]
Review of Scientific Instruments96(3), 033002 (2025) https://doi.org/10
Karami, M., Hatefi, M., Ahangaran, Z.G., Heydarinasab, Z., Samadi, D., Tameh, H.S., Tasnim, A., Sarreshtedari, F.: A novel laser intensity stabilization and mod- ulation method based on utilization of Zeeman–Faraday effect in atomic vapor cells. Review of Scientific Instrument...
2025
-
[15]
Review of Scientific Instruments78(10), 103110 (2007) https://doi.org/10.1063/1.2801012
Kim, D.I., Rhee, H.-G., Song, J.-B., Lee, Y.-W.: Laser output power stabilization for direct laser writing system by using an acousto-optic modulator. Review of Scientific Instruments78(10), 103110 (2007) https://doi.org/10.1063/1.2801012
2007 doi
-
[16]
Applied Physics B102(3), 515–522 (2011) https://doi.org/10.1007/ s00340-011-4399-1
Kwee, P., Willke, B., Danzmann, K.: New concepts and results in laser power stabilization. Applied Physics B102(3), 515–522 (2011) https://doi.org/10.1007/ s00340-011-4399-1
2011
-
[17]
Physical Review Applied 18(6), 064005 (2022) https://doi.org/10.1103/PhysRevApplied.18.064005
Li, L., Huie, W., Chen, N., DeMarco, B., Covey, J.P.: Active Cancellation of Servo- Induced Noise on Stabilized Lasers via Feedforward. Physical Review Applied 18(6), 064005 (2022) https://doi.org/10.1103/PhysRevApplied.18.064005
2022 doi
-
[18]
Optics & Laser Technology 45, 775–781 (2013) https://doi.org/10.1016/j.optlastec.2012.04.037
Liu, F., Wang, C., Li, L., Chen, L.: Long-term and wideband laser intensity sta- bilization with an electro-optic amplitude modulator. Optics & Laser Technology 45, 775–781 (2013) https://doi.org/10.1016/j.optlastec.2012.04.037
2013 doi
-
[19]
Review of Scientific Instruments90(3), 033101 (2019) https:// doi.org/10.1063/1.5066623
Mazurenko, A., Blatt, S., Huber, F., Parsons, M.F., Chiu, C.S., Ji, G., Greif, D., Greiner, M.: Implementation of a stable, high-power optical lattice for quantum gas microscopy. Review of Scientific Instruments90(3), 033101 (2019) https:// doi.org/10.1063/1.5066623
2019 doi
-
[20]
Optics Letters40(7), 1334 (2015) https://doi.org/10.1364/OL.40.001334
Michael, E.A., Pallanca, L.: Broadband near-to-shot-noise suppression of arbi- trary cw-laser excess intensity noise in the gigahertz range. Optics Letters40(7), 1334 (2015) https://doi.org/10.1364/OL.40.001334
2015 doi
-
[21]
IEEE Photonics Technology Letters20(18), 1542–1544 (2008) https://doi.org/10.1109/LPT.2008.928838 15
Nelson, C.W., Hati, A., Howe, D.A.: Relative Intensity Noise Suppression for RF Photonic Links. IEEE Photonics Technology Letters20(18), 1542–1544 (2008) https://doi.org/10.1109/LPT.2008.928838 15
2008
-
[22]
Applied Physics B71(2), 163–168 (2000) https://doi.org/10.1007/s003400000340
Ottaway, D.J., Veitch, P.J., Hollitt, C., Mudge, D., Hamilton, M.W., Munch, J.: Frequency and intensity noise of an injection-locked Nd:YAG ring laser. Applied Physics B71(2), 163–168 (2000) https://doi.org/10.1007/s003400000340
2000 doi
-
[23]
https: //doi.org/10.5281/zenodo.21290198
Partes, C., Auch, J., Heidt, E., Karpov, K., Kiesel, F., Groß, C.: AOM-based ultra-low noise laser intensity control up to the MHz range [dataset] (2026). https: //doi.org/10.5281/zenodo.21290198
2026 doi
-
[24]
Journal of Physics: Conference Series1380(1), 012034 (2019) https: //doi.org/10.1088/1742-6596/1380/1/012034
Phrompao, J., Pongvuthithum, R., Sucharitakul, T., Srakaew, K., Anukool, W.: Real-time and versatile laser-power stabilization with arbitrary amplitude mod- ulation. Journal of Physics: Conference Series1380(1), 012034 (2019) https: //doi.org/10.1088/1742-6596/1380/1/012034
2019 doi
-
[25]
Review of Scientific Instruments91(8), 083001 (2020) https://doi.org/10.1063/5.0009524
Preuschoff, T., Schlosser, M., Birkl, G.: Digital laser frequency and intensity sta- bilization based on the STEMlab platform (originally Red Pitaya). Review of Scientific Instruments91(8), 083001 (2020) https://doi.org/10.1063/5.0009524
2020 doi
-
[26]
Applied Physics B Photophysics and Laser Chemistry39(3), 149–153 (1986) https://doi.org/10.1007/BF00697412
Robertson, N.A., Hoggan, S., Mangan, J.B., Hough, J.: Intensity stabilisation of an argon laser using an electro-optic modulator - Performance and limitations. Applied Physics B Photophysics and Laser Chemistry39(3), 149–153 (1986) https://doi.org/10.1007/BF00697412
1986 doi
-
[27]
Applied Sciences10(2), 659 (2020) https://doi.org/10.3390/ app10020659
Sun, R., Wang, X., Zhang, K., He, J., Wang, J.: Influence of Laser Intensity Fluctuation on Single-Cesium Atom Trapping Lifetime in a 1064-nm Microscopic Optical Tweezer. Applied Sciences10(2), 659 (2020) https://doi.org/10.3390/ app10020659
2020
-
[28]
Journal of Physics: Conference Series122, 012016 (2008) https://doi.org/10.1088/1742-6596/122/1/012016
Takahashi, K., Ando, M., Tsubono, K.: Stabilization of laser intensity and fre- quency using optical fiber. Journal of Physics: Conference Series122, 012016 (2008) https://doi.org/10.1088/1742-6596/122/1/012016
2008 doi
-
[29]
Review of Scientific Instruments89(11), 113112 (2018) https://doi.org/10.1063/1.5046852
Tricot, F., Phung, D.H., Lours, M., Gu´ erandel, S., De Clercq, E.: Power stabi- lization of a diode laser with an acousto-optic modulator. Review of Scientific Instruments89(11), 113112 (2018) https://doi.org/10.1063/1.5046852
2018 doi
-
[30]
Review of Scientific Instru- ments68(1), 1–7 (1997) https://doi.org/10.1063/1.1147810
Talvitie, H., Pietil¨ ainen, A., Ludvigsen, H., Ikonen, E.: Passive frequency and intensity stabilization of extended-cavity diode lasers. Review of Scientific Instru- ments68(1), 1–7 (1997) https://doi.org/10.1063/1.1147810
1997 doi
-
[31]
Version Number: 2 (2026)
Villenas, F., G¨ ultekin, Y.C., Alvarado, A.: Optical Communications with Relative Intensity Noise: Channel Modeling and Information Rates. Version Number: 2 (2026). https://doi.org/arXiv:2603.08411
2026 arXiv
-
[33]
Optics Communications569, 130850 (2024) https://doi.org/10.1016/j.optcom.2024.130850
Wang, P., Wang, C., Li, J.: Relative intensity noise reduction for noise-like pulses based on synchronous intensity modulation. Optics Communications569, 130850 (2024) https://doi.org/10.1016/j.optcom.2024.130850
2024
-
[34]
Doctoral dissertation, Harvard University Graduate School of Arts and Sciences (2024)
Xu, M.: Quantum phases in Fermi Hubbard systems with tunable frustration. Doctoral dissertation, Harvard University Graduate School of Arts and Sciences (2024). https://nrs.harvard.edu/URN-3:HUL.INSTREPOS:37378930 17
2024
Reviewed August 5, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.