Fiber-based mid-infrared frequency-swept laser at 50 MScans/s via frequency down-conversion of time-stretched pulses

Kazuki Hashimoto; Makoto Shoshin; Takahiro Kageyama; Takuma Nakamura; Takuro Ideguchi

REVIEW 3 major objections 5 minor 1 cited by

Fiber-based mid-infrared frequency-swept laser at 50 MScans/s via frequency down-conversion of time-stretched pulses

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A fiber-based time-stretch source uses difference-frequency generation to produce mid-infrared frequency sweeps at 50 million scans per second, demonstrated by methane spectroscopy with 220 spectral elements at 3.4 µm.

desk verdict A genuine engineering advance in MIR time-stretch spectroscopy—worth a serious referee, but the 220-element claim needs an explicit edge-calibration check. read the letter →

arxiv 2501.01641 v1 pith:XCMKNJSY submitted 2025-01-03 physics.optics

Makoto Shoshin , Takahiro Kageyama , Takuma Nakamura , Kazuki Hashimoto , Takuro Ideguchi This is my paper · ORCID

classification physics.optics PACS 42.65.Ky42.55.Wd

keywords mid-infraredspectroscopyfrequency-sweptlasertime-stretchdifferencefrequencygenerationfibermethanesensingdispersioncompensationultrafastoptics

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that the speed and practicality of mid-infrared frequency-swept spectroscopy can be pushed far beyond what active scanning allows by stretching near-infrared pulses in ordinary telecom fiber and then down-converting them to the mid-infrared. The authors build a 50-MHz pulse train of 20-ns chirped pulses at 3.4 µm, sweep 19.0 cm⁻¹ per pulse, and resolve 220 spectral elements at 0.086 cm⁻¹ resolution. They demonstrate the system by recording methane absorption spectra at 50 million spectra per second and verifying the chirp against fiber dispersion parameters. If the claim holds, bulky femtosecond OPO systems can be replaced with compact, off-the-shelf fiber components, which matters for field applications such as combustion diagnostics and mid-infrared OCT.

What carries the argument

The central mechanism is time-stretch frequency down-conversion. A 50-MHz erbium-doped mode-locked fiber laser emits 1.5-µm pulses that are stretched to 20 ns by 20 km of dispersion-compensating fiber, imprinting a chirp whose instantaneous wavenumber is given by $\nu(t) = \nu_0 - \frac{1}{2\pi c}\left(\frac{t}{\phi_2} - \frac{\phi_3 t^2}{2\phi_2^3}\right)$ (Eq. 1), with $\phi_2$ the group-delay dispersion and $\phi_3$ the third-order dispersion. The stretched pulses are amplified and mixed in a periodically poled lithium niobate (PPLN) waveguide with a 10-MHz-linewidth continuous-wave 1.064-µm pump, generating 3.4-µm pulses that inherit the chirp through difference-frequency generation. The chirp is verified with a Michelson interferometer, giving $\phi_2 = 5113.0$ ps² and $\phi_3 = -31.7$ ps³, and the transmittance spectra are recovered with an iterative gradient-descent algorithm.

What would settle it

Record methane absorption lines that span the full 19.0 cm⁻¹ window and compare the retrieved line positions against a reference spectrum; any systematic deviation larger than 0.086 cm⁻¹ at the band edges would show the chirp mapping is not one-to-one.

Watch

Extended reading notes

Core claim

The central discovery is that difference-frequency generation of a time-stretched near-infrared pulse with a narrow-linewidth continuous-wave pump preserves the time-frequency chirp while shifting it to the mid-infrared. Using a 50-MHz erbium-doped mode-locked fiber laser, 20 km of dispersion-compensating fiber, and a 20-mm PPLN waveguide pumped at 1.064 µm, the authors obtain 3.4-µm pulses with an instantaneous wavenumber that sweeps nearly linearly through 19.0 cm⁻¹ in 20 ns. Michelson-interferometer measurements yield a group-delay dispersion of 5113 ps² and third-order dispersion of −31.7 ps³, matching the fiber's specifications. The resulting source runs at 50 MScans/s, and methane spectra retrieved with a gradient-descent algorithm agree with HITRAN simulations at 0.086 cm⁻¹ resolution.

Load-bearing premise

The scheme assumes the nonlinear crystal transfers the stretched near-infrared pulse's frequency sweep to the mid-infrared pulse without distortion; if the crystal's phase-matching bandpass reshapes the chirp, the stated 0.086 cm⁻¹ resolution and 220 spectral elements would not hold.

Editorial extensions

If this is right

Scan rates above 50 MHz are reachable by using a mode-locked laser with a higher repetition rate, provided the detection bandwidth is increased to preserve the 0.086 cm⁻¹ resolution.
The spectral window can be shifted or widened by temperature tuning the PPLN or choosing different poling periods, and sparse MIR absorption bands can be covered by adding multiple CW pump wavelengths.
The source can become fully fiber-connected using a fiber-coupled PPLN waveguide, making the entire system compact and robust enough for field use.
The demonstrated rate and resolution are suited to real-time MIR-OCT and on-site combustion diagnosis, the applications named in the paper.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The one-to-one spectral-transfer assumption could be tested edge-to-edge by comparing retrieved methane line positions near both ends of the 19.0 cm⁻¹ window; residual shifts would quantify PPLN phase-matching distortion.
The same architecture should extend to other MIR regions by only changing the CW pump wavelength and poling period, since the stretching fiber remains the same.
With a single-shot SNR of 14, quantitative single-shot sensing is plausible only for strong absorbers; field instruments may need averaging or up-conversion detection to reach trace-gas sensitivity.
Because the chirp nonlinearity from third-order dispersion is already below the instrument resolution, a faster detector would immediately translate into finer spectral resolution rather than requiring a new source.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this paper deserves a serious referee. The architecture is genuinely new: stretch 1.5 µm pulses in 20 km of telecom fiber, amplify, then difference-frequency-generate to 3.4 µm. That yields a compact MIR swept source at 50 MSpectra/s with 220 spectral elements over 19 cm⁻¹, replacing the group's previous fs-OPO/FACED system with its ~30 elements. This is a real step toward field-deployable high-speed MIR spectroscopy, not a rehash of a known result.

What the paper does well: the chirp characterization is the strongest part. The measured φ₂ and φ₃ from the Michelson interference agree with manufacturer specs to better than 1%, and the retrieved methane spectrum matches HITRAN after iterative retrieval. The use of telecom fiber plus a CW pump for DFG is sensible, and the discussion of up-conversion vs. down-conversion is honest about trade-offs. Citations look appropriate, with the retrieval and near-field effect properly traced to earlier work.

Soft spots, in order of weight. First, the absolute wavenumber axis is calibrated from a methane absorption line and then validated against the same methane spectrum. That is partially circular, and the retrieved spectrum has no error bars, so the agreement is qualitative. Second—the one I would press on—the 220-element/0.086 cm⁻¹ claim assumes the time-to-wavenumber map in Eq. (1) holds across the full 19 cm⁻¹ window, but the Michelson fit is described over ±5 ns around pulse center, and the PPLN phase-matching phase at the band edges is not characterized. The 10 MHz CW pump linewidth addresses the pump, not the PPLN transfer function that sets the -20 dB edges. If phase-matching adds curvature or sidelobe structure at the edges, the edge mapping could deviate by more than half a resolution element. I want to be fair: the data are consistent with a nearly linear chirp, and the HITRAN comparison would probably catch gross edge errors, so this is fixable with an explicit edge-calibration measurement rather than a fundamental flaw. Third, no data or code are shipped; 'available on request' is thin but acceptable for a proof-of-concept.

Bottom line: the time-stretch spectroscopy community should read and cite this. It needs a revision that makes the calibration chain non-circular and either measures or bounds the phase-matching contribution at the band edges. That is normal referee work, so I would send it out.

Referee Report

3 major / 5 minor

Summary. The paper reports a fiber-based mid-infrared frequency-swept source that reaches a scan rate of 50 MScans/s by stretching 1.5-μm pulses in a 20-km dispersion-compensating fiber, amplifying them, and difference-frequency mixing with a 1.064-μm CW laser in a PPLN waveguide to generate chirped pulses near 3.4 μm. The authors characterize the instantaneous wavenumber with a Michelson interferometer, fit the GDD and TOD, and demonstrate methane spectroscopy at 50 MSpectra/s with a claimed bandwidth of 19.0 cm⁻¹, a resolution of 0.086 cm⁻¹, and 220 spectral elements. The central claim is that this compact, passively scanned system is a practical alternative to bulkier OPO-based time-stretch MIR sources.

Significance. If the performance claims are correct, this is a significant advance for high-speed mid-infrared spectroscopy: it replaces a bulky fs-OPO and free-space stretcher with an off-the-shelf fiber laser and telecom fiber, while increasing the number of spectral elements from roughly 30 in the previous OPO-based demonstration to 220. The work is clearly presented, the measured dispersion values agree with manufacturer specifications, and the methane spectra show visible agreement with HITRAN. The main strengths are the passive scanning principle, the use of low-loss NIR fiber for stretching, and the demonstration of real spectroscopic measurements at 50 MSpectra/s. However, several load-bearing validation steps are missing: the full-window chirp map is not verified, the absolute wavenumber calibration is not described and is partly circular, and the retrieved spectrum is not accompanied by uncertainty or quantitative residuals.

major comments (3)

[Results, chirp characterization (Fig. 2, Eqs. (1)-(2))] The chirp-rate fit is performed only over ±5 ns around pulse center, which is roughly half of the 20-ns stretched-pulse window, whereas the 220-spectral-element and 19.0 cm⁻¹ claims apply to the full window. The statement that the 10 MHz CW pump ensures a 'precise one-to-one spectral transfer' addresses the pump linewidth, not the PPLN phase-matching transfer function; the amplitude and phase of the DFG process across the -20 dB band edges are not measured or simulated. The edge mapping therefore remains unverified, which directly affects the 0.086 cm⁻¹ resolution and 220-element count. Please provide a full-window chirp characterization or an independent multi-line reference validation, and quantify the DFG phase-matching contribution to the instantaneous wavenumber.
[Results, absolute wavenumber calibration (Fig. 2 caption and methane spectroscopy section)] The text says, 'The absolute wavenumber was calibrated by measuring an absorption line of methane gas, as detailed later,' but no later section contains the calibration procedure. Because the same methane dataset is then compared with the HITRAN spectrum in Fig. 4, the absolute wavenumber axis is not independently validated. Please specify which methane line was used, how the offset was determined, and validate the axis using lines not involved in the calibration or with a different gas sample.
[Results, Fig. 4 and retrieval] The retrieved transmittance spectrum is shown without error bars or residuals. Given the reported single-shot SNR of 14 and the nonlinear gradient-descent retrieval, the agreement with HITRAN should be quantified, for example by reporting residual RMS and line-position deviations across the full 19.0 cm⁻¹ window. Without this, the spectral resolution of 0.086 cm⁻¹ and the claim of 220 usable spectral elements cannot be fully assessed.

minor comments (5)

[Results, DFG setup] The conversion efficiency is quoted as '1.3%/W, including the coupling loss'; please specify whether the efficiency is referenced to the incident or coupled pump and signal powers.
[Fig. 2(c)] The caption says 'The black line represents the fitted curve of the data,' but the black line is difficult to distinguish from the data points in the printed figure; please use a different line style or color.
[References] The heading 'Reference' should be 'References'.
[Throughout] The term 'MScans/s' is used interchangeably with 'MSpectra/s'; please define both and use a consistent unit throughout.
[Data availability] The data availability statement is acceptable, but depositing the averaged waveforms and the retrieval code would strengthen reproducibility.

Circularity Check

1 steps flagged · score 2.0 of 10

Partial circularity only in the methane-based absolute wavenumber calibration; the central source-performance claims are direct measurements.

fitted input called prediction [Results, 'Evaluation of the instantaneous wavenumber' (Fig. 2d) and 'MIR absorption spectroscopy' (Fig. 4)]
""The absolute wavenumber was calibrated by measuring an absorption line of methane gas, as detailed later." ... "For comparison, a reference spectrum calculated based on the HITRAN database with the same resolution is also displayed in Fig. 4, demonstrating good agreement with the measured spectrum.""

A methane absorption line is used to set the absolute wavenumber (frequency offset) of the measured axis, and the same methane dataset is then compared with a HITRAN reference spectrum as validation. For the calibration line itself, the position match with HITRAN is enforced by construction, so the line-position agreement in Fig. 4 does not independently confirm the absolute frequency axis. The relative spacings of the other lines, their depths and widths, the measured chirp-rate shape, and the 50 MScans/s and 19.0 cm-1 performance claims remain independent, so the circularity is partial and not load-bearing for the main source-performance claims.

full rationale

The main derivation chain is not circular: the NIR pulses are stretched in fiber, down-converted by DFG, and the instantaneous wavenumber is measured by Michelson interferometry. The chirp-rate data are fitted to Eq. (2) to obtain phi2 and phi3, and those values are compared with manufacturer-specified GVD/TOD coefficients, which is an external consistency check. The 220 spectral elements are simply the ratio of the measured 19.0 cm-1 bandwidth to the measured 0.086 cm-1 resolution, not a fitted prediction. The only identifiable circular step is the absolute wavenumber calibration: one methane line anchors the frequency axis, and the same methane data are later compared with HITRAN, making the absolute position of that line agree by construction. This does not compromise the source-rate, bandwidth, or resolution claims. The concern about uncharacterized PPLN phase-matching at the band edges is a correctness/validation gap rather than circularity, because the paper does not use the phase-matching model to derive the claimed 220-element window; it directly measures the chirp and spectra.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claims rest on two fitted dispersion coefficients, a scalar wavenumber calibration, and standard time-stretch/DFG assumptions. No new physical entities are introduced. The fitted phi2/phi3 are benchmarked against manufacturer values, which is good, but the methane-based absolute calibration creates a mild circularity in the validation.

free parameters (3)

phi2 (group delay dispersion) = 5113.0 +/- 0.4 ps2
Fitted to the measured chirp rate using Eq. (2) in Fig. 2(c); used to establish the instantaneous wavenumber curve and to compare with manufacturer GVD.
phi3 (third-order dispersion) = -31.7 +/- 0.7 ps3
Fitted together with phi2 to the chirp rate; the comparison with the manufacturer TOD is a consistency check, not an independent prediction.
Absolute wavenumber offset = not stated
The instantaneous wavenumber curve is calibrated by matching a methane absorption line, introducing a scalar frequency offset that is not independently measured.

assumptions (5)

domain assumption The stretched-pulse dispersion relation Eq. (1), including GDD and TOD only, describes the instantaneous wavenumber.
Invoked in the chirp evaluation and curve fitting; higher-order dispersion is asserted to be negligible.
domain assumption The CW 1.064 um pump has negligible linewidth and the PPLN waveguide preserves the NIR chirp one-to-one in the DFG process.
Stated in the Results section as the basis for spectral transfer from NIR to MIR.
domain assumption The methane sample's response follows the near-field propagation effect model from ref [21], used by the gradient-descent retrieval.
The retrieval of the transmittance spectrum relies on this model and the HITRAN simulation.
domain assumption Averaging 100 waveforms and normalizing by a baseline without sample removes systematic errors without bias.
Used to produce the displayed spectrum; no uncertainty propagation is provided.
standard math Hilbert transform and Fourier analysis give the instantaneous frequency from the interference spectrum.
Used in the chirp-rate extraction in Fig. 2(c).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Fiber-based mid-infrared frequency-swept laser at 50 MScans/s via frequency down-conversion of time-stretched pulses." pith.science (2026). https://pith.science/paper/XCMKNJSY

@misc{pith2026250101641,
  author       = {Pith},
  title        = {Pith review of: Fiber-based mid-infrared frequency-swept laser at 50 MScans/s via frequency down-conversion of time-stretched pulses},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XCMKNJSY}},
  note         = {Machine review of arXiv:2501.01641}
}

read the original abstract

Increasing the sweep rate of mid-infrared (MIR) frequency-swept sources offers significant potential for various high-speed spectroscopy-based applications. While continuous-wave frequency-swept lasers have achieved sweep rates up to 1 MHz, a recently demonstrated time-stretched ultrashort pulsed laser has reached a significantly higher sweep rate, up to tens of MHz. However, the previous system relied on a bulky femtosecond optical parametric oscillator and produced only ~30 discrete spectral elements due to the use of a free-space time stretcher. In this work, we present a fiber-based frequency-swept MIR source that utilizes the frequency down-conversion of time-stretched near-infrared pulses, employing a compact mode-locked fiber laser and telecommunication fiber. As a proof-of-concept demonstration, we performed MIR spectroscopy of methane gas around 3.4 um at a rate of 50 MSpectra/s, capturing 220 spectral elements over a range of 19.0 cm-1. This compact and robust high-speed MIR frequency-swept laser system holds the potential for deployment in field applications.

Figures

Figures reproduced from arXiv: 2501.01641 by the authors.

**Figure 1.** Schematic of the MIR frequency-swept source via frequency down-conversion of time-stretched NIR pulses. The inset displays the down-converted MIR spectrum on a linear scale, measured using a Fourier-transform infrared spectrometer. DCF, dispersion compensating fiber; EDFA, erbium-doped fiber amplifier; DBR-LD, distributed Bragg reflector laser diode; YDFA, ytterbium-doped fiber amplifier; CW, continuous wave; PPLN, … view at source ↗

**Figure 2.** Evaluation of the instantaneous wavenumber of [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** MIR absorption spectroscopy of methane gas (a) Continuously measured time-stretched MIR pulses at a rate of 50 MSpectra/s. (b) Temporal waveform averaged over 100 measurements. (c) Comparison of dispersive absorption lines in the measured spectrum and calculated spectrum based on the HITRAN database. The transmittance spectrum was retrieved from the measured spectrum using an iterative gradient-descent (GD) algorith… view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Transmittance spectrum of methane gas retrieved from the 100-times averaged temporal waveform shown in [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Ultra-rapid broadband mid-infrared spectral tuning and sensing
physics.optics 2025-01 conditional novelty 6.0 of 10

A spectral-focusing difference-frequency source with a frequency-modulated electro-optic comb achieves pulse-to-pulse mid-infrared tuning at up to 60 THz/us and microsecond-scale broadband absorption measurements.

Reference graph

Works this paper leans on

24 extracted references · 23 canonical work pages · cited by 1 Pith paper

[1]

Gmachl, A

C. Gmachl, A. Straub, R. Colombelli, F. Capasso, D. L. Sivco, A. M. Sergent, and A. Y . Cho, IEEE J. Quantum Electron. 38, 569 (2002)

work page 2002
[2]

Mohan, A

A. Mohan, A. Wittmann, A. Hugi, S. Blaser, M. Giovannini, and J. Faist, Opt. Lett. 32, 2792 (2007)

work page 2007
[3]

Silva and I

A. Silva and I. D. Lindsay, Appl. Phys. B 107, 685 (2012)

work page 2012
[4]

Yumoto, N

M. Yumoto, N. Saito, U. Takagi, and S. Wada, Opt. Express 23, 25009 (2015)

work page 2015
[5]

Martyshkin, T

D. Martyshkin, T. Kesterson, V . V . Fedorov, and S. Mirov, in Solid State Lasers XXVI: Technology and Devices (SPIE, 2017), 10082, pp. 73–78

work page 2017
[6]

R. I. Woodward, M. R. Majewski, D. D. Hudson, and S. D. Jackson, APL Photonics 4, 020801 (2019)

work page 2019
[7]

Z. E. Loparo, A. V . Muraviev, P. Figueiredo, A. Lyakh, R. E. Peale, K. Ahmed, and S. S. Vasu, J. Energy Res. Technol. 140, 112202 (2018)

work page 2018
[8]

C. K. N. Patel, in Micro- and Nanotechnology Sensors, Systems, and Applications VIII (SPIE, 2016), 9836, pp. 411–419

work page 2016

Show all 24 references

[9]

D. T. D. Childs, R. A. Hogg, D. G. Revin, I. U. Rehman, J. W. Cockburn, and S. J. Matcher, Appl. Spectrosc. Rev. 50, 822 (2015)

2015
[10]

M. Abe, Y . Nishida, O. Tadanaga, A. Tokura, and H. Takenouchi, Opt. Lett. 41, 1380 (2016)

2016
[11]

Butschek, S

L. Butschek, S. Hugger, J. Jarvis, and M. Haertelt, Optical (2018)

2018
[12]

D. G. Revin and S. J. Matcher, Opt. Express 30, 21843 (2022)

2022
[13]

Lyakh, R

A. Lyakh, R. Barron-Jimenez, I. Dunayevskiy, R. Go, E. Tsvid, and C. K. N. Patel, Photonics 3, 19 (2016)

2016
[14]

Welzel, F

S. Welzel, F. Hempel, M. Hübner, N. Lang, P. B. Davies, and J. Röpcke, Sensors 10, 6861 (2010)

2010
[15]

R. S. M. Chrystie, E. F. Nasir, and A. Farooq, Opt. Lett. 39, 6620 (2014)

2014
[16]

Wu, Y .-Q

J.-L. Wu, Y .-Q. Xu, J.-J. Xu, X.-M. Wei, A. C. Chan, A. H. Tang, A. K. Lau, B. M. Chung, H. Cheung Shum, E. Y . Lam, K. K. Wong, and K. K. Tsia, Light Sci Appl 6, e16196 (2017)

2017
[17]

Kawai, K

A. Kawai, K. Hashimoto, T. Dougakiuchi, V . R. Badarla, T. Imamura, T. Edamura, and T. Ideguchi, Communications Physics 3, 1 (2020)

2020
[18]

Dougakiuchi, A

T. Dougakiuchi, A. Ito, M. Hitaka, K. Fujita, and M. Yamanishi, Appl. Phys. Lett. 118, 041101 (2021)

2021
[19]

S. W. Jolly, N. H. Matlis, F. Ahr, V . Leroux, T. Eichner, A.-L. Calendron, H. Ishizuki, T. Taira, F. X. Kärtner, and A. R. Maier, Nat. Commun. 10, 2591 (2019)

2019
[20]

Savitzky and M

A. Savitzky and M. J. E. Golay, Anal. Chem. 36, 1627 (1964)

1964
[21]

Hashimoto, T

K. Hashimoto, T. Nakamura, T. Kageyama, V . R. Badarla, H. Shimada, R. Horisaki, and T. Ideguchi, Light Sci. Appl. 12, 48 (2023)

2023
[22]

D. P. Kingma and J. Ba, arXiv:1412.6980v9 [cs.LG] (2014)

2014 arXiv
[23]

F. Y u, P. Song, D. Wu, T. Birks, D. Bird, and J. Knight, APL Photonics 4, 080803 (2019)

2019
[24]

S. Yagi, T. Nakamura, K. Hashimoto, S. Kawano, and T. Ideguchi, APL Photonics 9, 051301 (2024)

2024

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

Gmachl, A

C. Gmachl, A. Straub, R. Colombelli, F. Capasso, D. L. Sivco, A. M. Sergent, and A. Y . Cho, IEEE J. Quantum Electron. 38, 569 (2002)

work page 2002

[2] [2]

Mohan, A

A. Mohan, A. Wittmann, A. Hugi, S. Blaser, M. Giovannini, and J. Faist, Opt. Lett. 32, 2792 (2007)

work page 2007

[3] [3]

Silva and I

A. Silva and I. D. Lindsay, Appl. Phys. B 107, 685 (2012)

work page 2012

[4] [4]

Yumoto, N

M. Yumoto, N. Saito, U. Takagi, and S. Wada, Opt. Express 23, 25009 (2015)

work page 2015

[5] [5]

Martyshkin, T

D. Martyshkin, T. Kesterson, V . V . Fedorov, and S. Mirov, in Solid State Lasers XXVI: Technology and Devices (SPIE, 2017), 10082, pp. 73–78

work page 2017

[6] [6]

R. I. Woodward, M. R. Majewski, D. D. Hudson, and S. D. Jackson, APL Photonics 4, 020801 (2019)

work page 2019

[7] [7]

Z. E. Loparo, A. V . Muraviev, P. Figueiredo, A. Lyakh, R. E. Peale, K. Ahmed, and S. S. Vasu, J. Energy Res. Technol. 140, 112202 (2018)

work page 2018

[8] [8]

C. K. N. Patel, in Micro- and Nanotechnology Sensors, Systems, and Applications VIII (SPIE, 2016), 9836, pp. 411–419

work page 2016