2.05-µm all-fiber laser source designed for CO 2 and wind coherent lidar measurement

TL;DR: An all-fiber pulsed laser source for simultaneous remote sensing of CO2 concentration and wind velocity in the 2.05 µm region based on a polarization-maintaining master oscillator power amplifier (MOPA) architecture is reported on.

Abstract: This work reports on an all-fiber pulsed laser source for simultaneous remote sensing of CO2 concentration and wind velocity in the 2.05 µm region. The source is based on a polarization-maintaining master oscillator power amplifier (MOPA) architecture. Two narrow-linewidth master oscillators for ON-line/OFF-line CO2 differential absorption lidar operation alternately seed a four-stage amplifier chain at a fast switching rate up to 20 kHz. The MOPA architecture delivers laser pulses of 120 µJ energy, 200 ns duration (600 W peak power) at 20 kHz pulse repetition rate (2.4 W average power). The output linewidth is lower than 5 MHz, close to the pulse Fourier transform limit, and the beam quality factor is M2=1.12. The source also provides a pre-amplified 20 mW local oscillator with a relative intensity noise of −160dB/Hz that ensures optimal performance for future coherent detection.

Summary

1. Introduction

• Carbon dioxide (CO2) is widely acknowledged as the most important anthropogenic greenhouse gas in the atmosphere.
• To improve their knowledge of the CO2 life cycle, and open ways to control anthropogenic emissions, it is necessary to quantify CO2 fluxes around sources and sinks, at a local scale with ground-based instruments, and if possible, at the global scale with space-borne instruments.
• In the perspective of ground-based lidar systems, a fiber pulsed laser source at 2.05 µm could therefore offer a robust solution to measure simultaneously range-resolved profiles of the CO2 Volume Mixing Ratio (VMR) and the wind speed, which is ideal for autonomous CO2 flux rate measurements.
• These systems could also find industrial interests.

2. Laser objectives

• The CO2 absorption line centered on 2050.97 nm (R30) has been identified in previous studies as one of the most promising for a space-borne lidar instrument [17].
• The ON and OFF wavelengths must also be close enough (typically < 1nm) to guarantee similar backscattering amplitude by the atmospheric aerosols or by hard-target surfaces.
• All these features play an important role on the Carrier-to-Noise Ratio (CNR) (see section 4).
• To minimize the bias made on the CO2 VMR measurement, the spectral drift within the measurement time should be limited (or at least be monitored) to keep the bias as low as possible.

3. Experimental setup

• The source architecture is based on a Polarization Maintaining (PM) Master Oscillator Power Fiber Amplifier made of four Thulium Doped Fiber Amplifiers (TDFA) pumped at 793 nm.
• The pre-amplifier (TDFA1, detailed in section 3.1) delivers a continuous signal.
• The optical power at its output is split in two parts, one delivering the LO power and one seeding the second amplifier.
• An AOM (AcoustoOptic Modulator) shapes the signal into pulses and adds an optical frequency offset for heterodyne detection.
• Since the last two amplifiers, (including ASE filtering, Stimulated Brillouin Scattering (SBS) monitoring, and fiber strain gradient for SBS gain reduction) are identical to those reported in [22], only general features of TDFA3 and TDFA4 are reminded in section 3.3.

3.1 Continuous wave pre-amplification for LO derivation

• In many coherent fiber lidar designs, a fraction of the CW (Continuous Wave) laser seed is directly derived to be used as LO, and the rest is shaped before amplification [23].
• Moreover, DIAL measurement may also require a wavelength calibration or monitoring channel that would require another fraction of the DFB-LD CW power.
• As the laser PRF requirement must be under 50 kHz, it allows for pulse-to-pulse switching if necessary.
• This issue and its expected impact on the lidar performance is discussed later in section 4.5.
• Compared to the pre-amplifier previously reported in [22], this new pre-amplifier has lower power (80 mW instead of 1 W) but it exhibits much lower RIN and thermal fluctuations.

3.2 Pulse shaping and pre-amplification

• An AOM is placed before TDFA2 to perform simultaneously the pulse shaping and the 80 MHz frequency shift.
• The electronic command for the AOM driver is a trapezoidal pulse with a slower leading edge to pre-compensate the gain alteration during the pulse amplification [26].
• While testing their lidar coherent detection, the authors found that the CNR was degraded by a residual bounce of the AOM-transmitted light, about 4 µs after trigger time.
• Its amplitude was about 48 dB lower than the main pulse (typical AOM characteristic).
• At this position the EOM also filters about 9 dB of the ASE power produced by the TDFA2 between pulses.

3.3 Power amplification

• The power amplification gain is distributed over two stages (TDFA3 and TDFA4) in order to limit the fiber warming.
• Compared to [22], the TDFA3 is seeded with a higher peak power.
• Therefore, a lower pump power is required to reach the same output power, thus limiting the thermal heating and signal power fluctuations.
• The limitation of the output power by Brillouin scattering is alleviated by introducing a strain gradient along the fiber.
• The SBS threshold peak power has been measured to 800 W at the TDFA4 output, but in order to prevent laser pump warming over long time measurement, a working point of 120 µJ/pulse, 200 ns, 20 kHz was finally chosen (600 W peak power).

4.1 Laser-induced noise sources

• The quality of coherent lidar signals is often expressed through the Carrier-to-Noise Ratio (CNR), defined by: 2 2 (z) het noise i CNR i (1).
• The angle brackets denote the mean value, i²het is the heterodyne current power and i²noise the noise current power.
• The CNR expression can thus be developed to highlight the most important features as followed:.
• The optimal sensitivity for heterodyne detection is achieved when the noise current is dominated by the shot noise (quantum limit).
• The RIN of the DFBLD is lower than -170 dB/Hz around 80 MHz, and therefore it represents a negligible contribution to the noise power.

4.2 Spectral linewidth

• Its linewidth is determined by the LO linewidth, the pulse profile, and the pulseto-pulse beat note frequency jitter in case of time averaging.
• The beat note between the output laser signal and the LO is obtained by mixing a small fraction of the emitted pulses with the LO through a fiber coupler.
• The average spectral broadening caused by pulse shaping and amplification is 3 MHz at FWHM.
• For a lidar configuration, where the LO and signal path differences cannot be balanced, the total spectral linewidth is the convolution of the DFB-LD PSD (induced by phase noise with Lorentzian profile) and the spectral broadening mentioned before.
• The Allan standard deviation of the beat note frequency is shown on Fig 4.

4.3 Beam quality and polarization

• At the final fiber output, the laser beam is collimated by an aspheric lens that does not alter the M2 factor.
• The measurement is made by fitting the beam diameter (D4σ definition) along the propagation axis (Fig 5).
• The Polarization Extinction Ratio (PER) has also been measured and is higher than 16 dB.
• As the interference between the LO and the lidar signal requires the same polarization state, the higher is the PER, the higher is the heterodyne efficiency.
• The relative degradation of the CNR induced by the PER is expressed as: (PER) 1 1 ( ) CNR CNR PER (3) In their case, the PER degrades the CNR by less than 0.11 dB compared to a perfectly linearly polarized beam.

4.4 Frequency tuning and stability

• The laser spectral tuning range is shown on Fig 6, together with the absorption lines of CO2 and H2O.
• The authors DFB-LD is tunable over this range using the temperature setting.
• Such a multispecies DIAL-Doppler lidar would then be able to monitor simultaneously two major greenhouse gases [28].
• The authors find a peak-to-peak drift lower than 50 MHz over ten minutes (standard deviation of 20 MHz).
• This error would convert into a 0.2% bias on a CO2 VMR measurement, in the case of an ON-line wavelength set on the center of the R30 absorption line.

4.5 Spectral purity impact on DIAL measurement

• Fig 7 represents the optical spectrum at the OS output for the two switch positions.
• The cross-talk is -23 dB for both positions.
• The authors found however that the cross-talk value is sensitive to temperature fluctuations.
• Due to the presence of two distinct wavelengths in the amplifiers, Four-Waves Mixing (FWM) occurs and produces additional sidebands.

5. Conclusion

• The authors demonstrated a 2.05µm all-fiber laser source that meets the requirements of a DIAL-Doppler emitter for remote sensing of atmospheric CO2 and wind.
• The addition of all the noise contributions listed above reduces the CNR by less than 2.6 dB.
• A single-pass solid-state amplification is under assessment, which would minimize alignment issues.

Figures

Fig 2. Schematic of the DFB-LD with the OS seeding the pre-amplifier (TDFA1). The two boxes with right-directed arrows are optical isolators

Fig 6. CO2 (red) and H2O (blue) horizontal transmission for 2 km propagation through standard atmosphere. The white background represents the total laser spectral tuning.

Fig 7. Optical spectrum of the DFB-LD at the OS output.

Table 2: Laser-induced CNR loss/VMR bias budget

Fig 5. M2 estimation of laser beam at 120 µJ in two perpendicular directions. Laser beam is

Table 3. Parameters used for theoretical calculation of CNR

Table 1: Main laser features

Fig 4. Allan standard deviation (blue) of the beat note frequency calculated from a dataset of 5 minutes (plotted over 10 s to minimize the statistical error) and -1/2 slope (black, dashed) representing the stationary behavior. LO and signal are delayed by 1 km. σ: Allan standard deviation, ν0: beat note frequency

Fig 3. (a) Pulse shape at the TDFA4 output (blue) and Gaussian fit (red, dashed). (b) Beat note between the LO and the emitted pulse and (c) the PSD averaged over 200 shots. Δt: pulse duration at FWHM, Δν: spectral linewidth at FWHM, ν0: beat note center frequency

Fig 1: Schematic of the MOPFA. Energy values are given for 200 ns pulses at 20 kHz repetition rate. DFB-LD: Distributed Feedback Laser Diode, OS: Optical Switch, LO: Local Oscillator, TDFA: Thulium Doped Fiber Amplifier, AOM: Acousto-Optic Modulator, EOM: Electro-Optic Modulator, ASE: Amplified Spontaneous Emission, HR-FBG: High Reflectivity Fiber Bragg Grating, SBS: Stimulated Brillouin Scattering, LMA: Large Mode Area.

Full text

HAL Id: hal-03202735
https://hal.archives-ouvertes.fr/hal-03202735
Submitted on 2 Jun 2021
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
205-µm all-ber laser source designed for CO 2 and wind
coherent lidar measurement
Julien Lahyani, Julien Le Gouët, Fabien Gibert, Nicolas Cézard
To cite this version:
Julien Lahyani, Julien Le Gouët, Fabien Gibert, Nicolas Cézard. 205-µm all-ber laser source designed
for CO 2 and wind coherent lidar measurement. Applied optics, Optical Society of America, 2021, 60
(15), pp.C12-C19. �10.1364/AO.416821�. �hal-03202735�

1
2.05-µm all-fiber laser source designed for
CO
2
and wind coherent lidar measurement
JULIEN LAHYANI
1
, JULIEN LE GOUËT
1,*
, FABIEN GIBERT
2
AND NICOLAS
CEZARD
3
1
DOTA, ONERA, Université Paris-Saclay, F-91123 Palaiseau, France
2
Laboratoire de Météorologie Dynamique (LMD), Centre National de Recherche Scientifique (CNRS),
Ecole Polytechnique, FR-91128 Palaiseau cedex, France
3
ONERA/DOTA, Université de Toulouse, F-31055 Toulouse, France
*julien.le_gouet@onera.fr
Abstract: This work reports on an all-fiber pulsed laser source for simultaneous remote
sensing of CO
2
concentration and wind velocity in the 2.05 µm region. The source is based on
a polarization maintaining Master Oscillator Power Amplifier (MOPA) architecture. Two
narrow-linewidth master oscillators for ON-line/OFF-line CO
2
DIAL operation alternately
seed a four-stage amplifier chain at a fast switching rate up to 20 kHz. The MOPA
architecture delivers laser pulses of 120 μJ energy, 200 ns duration (600 W peak power) at
20 kHz pulse repetition rate (2.4 W average power). The output linewidth is lower than 5
MHz, close to the pulse Fourier-Transform limit, and the beam quality factor is M² = 1.12.
The source also provides a pre-amplified 20 mW local oscillator with a relative intensity
noise of -160 dB/Hz that ensures optimal performance for future coherent detection.
1. Introduction
Carbon dioxide (CO
2
) is widely acknowledged as the most important anthropogenic
greenhouse gas in the atmosphere. Yet our understanding of its impact on future climate
evolution still suffers some uncertainties [1]. To improve our knowledge of the CO
2
life
cycle, and open ways to control anthropogenic emissions, it is necessary to quantify CO
2
fluxes around sources and sinks, at a local scale with ground-based instruments, and if
possible, at the global scale with space-borne instruments.
Lidar sensors are especially attractive for such tasks, and several lidar systems based on high
energy solid-state lasers already demonstrated good capabilities for CO
2
monitoring at
2 µm [2–5] and 1.5 µm [6–9], using DIfferential Absorption Lidar (DIAL) or Integrated Path
Differential Absorption (IPDA). However, solid-state architectures generally involve large
numbers of free-space optics that can raise substantial thermal and mechanical alignment
issues when designing the observation system. All-fiber laser architectures alleviate those
problems. In an assessment for a space-borne lidar (ASCENDS program), NASA Goddard
Space Flight Center [8], based on a Fibertek heritage [10,11], and the Information
Technology R&D Center of Mitsubishi Electric [12] both demonstrated for instance the
interest of an all-fiber architecture in the 1.57 µm region. As a benchmark regarding pulse
energy levels and repetition rates performed in previous works, solid-state cavities at 2.05 µm
can reach tens of mJ at moderate repetition rate (usually lower than 1 kHz) [4], while all-
fiber systems at 1.57 µm and 2.05 µm (including our system) reach hundreds of µJ at higher
repetition rate (usually tens of kHz) [10]. Both technologies thus offer different trade-offs.
For space-borne monitoring with the IPDA technique, the 2.05 µm band is attractive, for it
allows relaxing the requirement on the random error compared to1.57 µm. Indeed, the
pressure dependence of the CO
2
R30 absorption line at 2051 nm offers a more favorable

2
Weighting Function (WF) in the low troposphere [13], where sources and sinks are
localized [14]. Therefore, the development of a 2.05 µm all-fiber pulsed laser source, suitable
for CO
2
monitoring, appears highly desirable in the perspective of future space-borne lidar
system.
Fiber lidar systems also have the advantage to facilitate the use of coherent detection to
perform range-resolved measurements of the wind speed. In the perspective of ground-based
lidar systems, a fiber pulsed laser source at 2.05 µm could therefore offer a robust solution to
measure simultaneously range-resolved profiles of the CO
2
Volume Mixing Ratio (VMR) and
the wind speed, which is ideal for autonomous CO
2
flux rate measurements. Such dual
function DIAL-Doppler lidars have already been reported for CO
2
[2,3] using solid-state
lasers.
These systems could also find industrial interests. For example, the Physics Department of
Montana state University [15] reported surface monitoring of CO
2
sequestration sites with a
1.57 µm fiber-based DIAL lidar, but they used direct detection and could not perform wind
measurement simultaneously. Recently, our research group reported a fiber-based DIAL-
Doppler lidar at 1.64 µm for industrial methane leaks monitoring, and demonstrated
simultaneous CH
4
/wind range-resolved profiles [16].
This study follows a bottom-up approach, with multiple potential applications that will not be
discussed in details here. Of course, space requirements are very demanding and could not be
fulfilled using the presented system. Ground-based industrial applications would generally
require relaxed requirements, but with large variations depending of the scenario. Whatever
the final application, building a powerful all-fiber laser source is a step forward, simpler and
easier-to-deploy, for future lidar systems.
In this paper we report on the design and performance of a high peak-power, narrow-
linewidth, all-fiber pulsed laser source at 2.05 µm, designed to be suitable for standalone km-
range ground-based CO
2
/wind measurement using coherent detection. To the best of our
knowledge, this is the first all-fiber laser system at 2.05 µm that allows for such possibilities.
In the first part, we describe and justify the objectives we had for the 2 µm fiber laser source,
and remind relevant results of the literature in that field. In the second part, we describe the
laser design. The last part is dedicated to experimental characteristics obtained with the laser
source, in the perspective of upcoming DIAL-Doppler measurements.
2. Laser objectives
The CO
2
absorption line centered on 2050.97 nm (R30) has been identified in previous
studies as one of the most promising for a space-borne lidar instrument [17]. According to the
HITRAN database, the CO
2
R30 transition is about 4.3 GHz wide (at Full Width Half
Maximum - FWHM) at 1 bar/293 °K in standard atmosphere, and is separated by 40 GHz
from the nearest CO
2
absorption line. Therefore, the laser must offer narrow-linewidth in
comparison to the absorption line width, and a tuning range of 20 GHz is desirable to allow
full coverage of the R30 absorption line sideband. The ON and OFF wavelengths must also
be close enough (typically < 1nm) to guarantee similar backscattering amplitude by the
atmospheric aerosols or by hard-target surfaces.
To maximize the measurement accuracy and the lidar range, high laser pulse energy is
necessary. Typically hundreds of µJ are required for range-resolved DIAL measurement with
kilometer range in the boundary layer [2]. The pulse length should be between 100 ns and
1 µs, for the range-resolution to be between 15 and 150 m. The Pulse Repetition Rate (PRF)
should be below 50 kHz to raise the ambiguity range up to 3 km. Fast wavelength switching
is also required to ensure high atmospheric correlation between ON and OFF signals. For the

3
atmosphere to be considered as ‘frozen’, a switching rate of 1 kHz or more is typically
required for ground-based systems [18]. In Table 1, we summarize the main laser features.
Table 1: Main laser features
Feature
Results in this study
Laser wavelength
2.05 µm
Spectral tunability
70 GHz (1 nm)
Pulse energy
120 µJ
Peak – Average power
600 W – 2.4 W
Pulse duration
200 ns
PRF
20 kHz
ON-OFF switch rate
Up to 20 kHz
Heterodyne measurement requires a Local Oscillator (LO) with low RIN (Relative Intensity
Noise) in the analysis bandwidth. The fiber laser should exhibit a nearly Fourier-transform
limited linewidth, a beam quality close to the diffraction limit (M
2
=1) and a linear
polarization. All these features play an important role on the Carrier-to-Noise Ratio (CNR)
(see section 4). To minimize the bias made on the CO
2
VMR measurement, the spectral drift
within the measurement time should be limited (or at least be monitored) to keep the bias as
low as possible. High cross-talk isolation between ON-line and OFF-line beams and high
Side-Mode Suppression Ratio (SMSR) are also required. In Table 2, we show the measured
characteristics for all these parameters, with the associated error budgets in terms of CNR loss
(random error) and CO
2
VMR bias budget (systematic error). These numbers are discussed in
section 4.
Table 2: Laser-induced CNR loss/VMR bias budget
Features
Results in this study
CNR loss budget
RIN (around AOM
frequency shift)
-160 dB/Hz
< 1 dB
Spectral Linewidth
<5 MHz
< 1 dB
Signal -LO beat frequency
stability @ 10 ms
100 kHz
Negligible
Beam quality M
2
1,12
0.5 dB
Polarization, Polarization
Extinction Ratio (PER)
Linear, PER > 16 dB
< 0.11 dB
VMR bias budget
Frequency drift over 10 min
< 50 MHz (peak-to-peak)
<0.2%
Cross-Talk
-23 dB
<0.1% up to 3 km
SMSR
> 45dB
Negligible
Usually, power scaling of fiber-laser is limited by the extractible power or Stimulated
Brillouin Scattering (SBS). As illustrated in [19], Tm-doped fiber amplifiers can deliver very
high powers (up to 1 kW), and SBS can be partly circumvented by numerous methods [20].
In [21], a monolithic all-fiber amplifier delivering a high peak power of 10 kW for 100 ns
pulses at 2.05 µm has also been demonstrated, but using a non-single-frequency seeder. We
previously developed a 2.05 µm, single-frequency, 110 µJ, 110 ns, 20 kHz fiber laser source
(2.2 W average power) [22]. However, this source was developed for optical parametric
oscillator pumping, was mono-wavelength, and did not provide any LO output. Moreover,

4
this previous source unfortunately exhibited high RIN and also suffered power instabilities.
The following sections explain how the source has been re-designed, upgraded, and fully
characterized to comply with our objectives for combined lidar measurement of CO
2
and
wind.
3. Experimental setup
The amplification chain is shown on Fig 1 and is detailed in following subsections. The
source architecture is based on a Polarization Maintaining (PM) Master Oscillator Power
Fiber Amplifier (MOPFA) made of four Thulium Doped Fiber Amplifiers (TDFA) pumped at
793 nm. The MOPFA is seeded alternatively by two narrow-linewidth Distributed Feed-Back
Laser Diodes (DFB-LD) using an Optical Switch (OS). The pre-amplifier (TDFA1, detailed
in section 3.1) delivers a continuous signal. The optical power at its output is split in two
parts, one delivering the LO power and one seeding the second amplifier. An AOM (Acousto-
Optic Modulator) shapes the signal into pulses and adds an optical frequency offset for
heterodyne detection. The TDFA2 and the pulse shaping system are described in section 3.2.
At its output, a Mach-Zehnder Electro Optic Modulator (EOM) is used as a time-gated
attenuator to filter amplified AOM parasitic spikes. Since the last two amplifiers, (including
ASE filtering, Stimulated Brillouin Scattering (SBS) monitoring, and fiber strain gradient for
SBS gain reduction) are identical to those reported in [22], only general features of TDFA3
and TDFA4 are reminded in section 3.3. Finally, the output of TDFA4 is spliced to a simple-
clad LMA fiber that is crimped into a FC-APC connector, allowing for an easy coupling to
the lidar emission optics.
Fig 1: Schematic of the MOPFA. Energy values are given for 200 ns pulses at 20 kHz
repetition rate. DFB-LD: Distributed Feedback Laser Diode, OS: Optical Switch, LO: Local
Oscillator, TDFA: Thulium Doped Fiber Amplifier, AOM: Acousto-Optic Modulator, EOM:
Electro-Optic Modulator, ASE: Amplified Spontaneous Emission, HR-FBG: High Reflectivity
Fiber Bragg Grating, SBS: Stimulated Brillouin Scattering, LMA: Large Mode Area.
3.1 Continuous wave pre-amplification for LO derivation
In many coherent fiber lidar designs, a fraction of the CW (Continuous Wave) laser seed is
directly derived to be used as LO, and the rest is shaped before amplification [23]. In our
case, a LO power of 20 mW (able to provide 1 mW power at the detector after a 95:5 coupler)
is necessary for the heterodyne measurement. However, the power delivered by the DFB-LD
is limited to 10 mW, and the optical switch introduces losses. Moreover, DIAL measurement
may also require a wavelength calibration or monitoring channel that would require another
fraction of the DFB-LD CW power. Therefore, it was necessary to pre-amplify the signal,
without degrading the intensity noise. The seeders, OS and CW pre-amplifier are illustrated in
Fig 2.

Frequently asked questions

Q1. What have the authors contributed in "205-μm all-fiber laser source designed for co 2 and wind coherent lidar measurement" ?

This work reports on an all-fiber pulsed laser source for simultaneous remote sensing of CO2 concentration and wind velocity in the 2. 05 μm region. The source also provides a pre-amplified 20 mW local oscillator with a relative intensity noise of -160 dB/Hz that ensures optimal performance for future coherent detection. 

Q2. What are the future works mentioned in the paper "205-μm all-fiber laser source designed for co 2 and wind coherent lidar measurement" ?

A future step of this work is therefore to inject the fiber output beam into a solid-state amplifier for peak-power upscaling.