We report a dual-wavelength-pumped all-fiber continuous-wave laser operating at the extended wavelength of 3.79 µm that reaches a record output power of 2.0 W. This represents, to the best of our knowledge, the highest output power reported at the longest spectral range for a fiber laser. The laser cavity, made of a heavily erbium-doped fluoride fiber and bounded by two photo-inscribed fiber Bragg gratings, reaches a slope efficiency of 46.5% with respect to the absorbed 1976 nm pump power. The system exhibits an absorption dependency of the 1976 nm pump on the launched 976 nm pump and a quenching behavior dependency on the output coupler reflectivity. The all-fiber design of the cavity allows significant power scaling of the laser and ensures its long-term stability.

Coherent light sources emitting in the mid-infrared region (MIR) of the electromagnetic spectrum located between 2 and 5 µm are in high demand for a wide variety of applications across different fields.1,2 For example, such sources can be of use for polymer processing and detection of gases, such as methane, because of the resonant excitation of vibrational molecular bonds around 3.4 µm.3–7 At longer wavelengths, there is currently an urgent need for the development of multi-watt coherent sources near 3.9 µm in relation to directional countermeasures and free-space communications exploiting the transparent window of the atmosphere centered around this wavelength.8–10 

Promising candidates to meet these needs are rare-earth-doped fluoride glass fiber lasers, given their compact design, high beam quality, and availability of emission over a broad range of wavelengths. Their development is achievable at present, thanks to the commercial availability of high-quality fluoride glass fibers with high concentration of rare-earth dopants and low background losses.11 In continuous-wave operations, rare-earth-doped fluoride fibers allowed the development of multiple record-breaking laser systems, namely 40 W at 2.8 µm and 15 W at 3.55 µm with erbium-doped fluorozirconate (ZrF4) fibers, 10 W at 3.2 µm with dysprosium-doped ZrF4 fibers, and 1.7 W at 3.9 µm with holmium-doped fluoroindate (InF3) fibers.12–15 Even though InF3 fibers are more advantageous at longer wavelengths due to their lower attenuation levels, a laser system was developed in 2016 with erbium-doped ZrF4 fibers using two pump lasers at 976 and 1976 nm capable of producing milliwatt-level emissions up to 3.78 µm. This pumping scheme known as a dual-wavelength-pumped (DWP) scheme replaced the low efficiency pumping scheme of using a single pump laser at 655 nm to stimulate the 4F9/24I9/2 transition of erbium ions in ZrF4 glass.16,17 It proceeds by first populating the long-lived 4I11/2 level through ground-state absorption (GSA) with a 976 nm laser, thus resulting in accumulation of ions in this level. This population then act as a virtual ground-state (VGS), which is next excited up to the 4F9/2 level by the 1976 nm laser through a process identified as virtual ground-state absorption (VGSA). The two pumps also result in various excited-state absorptions (ESA and VESA) between multiple levels, accompanied by interactions between pairs of ions resulting in energy transfer upconversions (ETU) and cross-relaxations (CR) shifting the levels’ population densities. This is represented in the comprehensive energy diagram of the Er3+:ZrF4 DWP system in Fig. 1.

FIG. 1.

Extended energy diagram of the Er3+:ZrF4 DWP system.

FIG. 1.

Extended energy diagram of the Er3+:ZrF4 DWP system.

Close modal

A lot of work was done to study and optimize the DWP system, notably in relation to the quenching phenomenon of the lasing power arising from an improper power ratio between the two pumps, observed in ZrF4 fibers with low levels of erbium-doping,18 and the development of the first monolithic DWP erbium-doped fiber laser with the introduction of robust and highly transmissive single-mode splicing between the silica fiber delivering both pumps and the doped ZrF4 fiber with a laser cavity bounded by two fiber Bragg gratings (FBGs), resulting in the record output power of 15 W at 3.55 µm mentioned above.12 

In this Letter, we report a monolithic heavily erbium-doped DWP fiber laser capable of reaching multi-watt level output powers at the longest operating spectral range of 3.8 µm for a DWP laser system. We also report an absorption dependency of the 1976 nm pump on the launched 976 nm pump, unique to heavily erbium-doped DWP fiber lasers, and a quenching behavior in relation to the output coupler reflectivity.

The schematic of the 3.8 µm monolithic Er3+:ZrF4 DWP fiber laser is depicted in Fig. 2. The laser cavity is composed of a 3.75 m long segment of double-clad 7 mol. % Er3+:ZrF4 active fiber (15/240 × 260 µm2, NA = 0.12) manufactured by Le Verre Fluoré.11 Delimiting the cavity is a pair of intra-core photo-inscribed narrowband type-I FBGs of 98% and 66% labeled as high-reflectivity (HR-FBG) and low-reflectivity (LR-FBG), respectively.19 The cavity’s length has been determined to maximize the pumps absorption while keeping it low enough to not cause any parasitic lasing at 2.8 µm. Reabsorption from the lower energy level of the laser transition 4I9/2 did not appear to have an effect on the laser performance, because of its short lifetime. As in Ref. 12, the first pump (P1) is a 60 W source at 976 nm composed of two combined 30 W laser diodes (BWT K980CA3R), while the second one (P2) is a 100 W laser at 1976 nm provided using a homemade Tm3+:SiO2 all-fiber laser (11/260 × 240 µm2, NA = 0.125), which is cladding-pumped using two 100 W laser diodes at 790 nm. P1 and P2 are combined in a passive double-clad silica fiber (11/250 µm) using a custom (2 + 1) × 1 fiber combiner. The two pumps are delivered to the laser cavity after passing through a robust single-mode silica-to-fluoride splice with a transmission of 90% at 1976 nm, an undoped passive ZrF4 relay fiber (14.5/250 µm), and finally, a single-mode ZrF4 splice with a transmission of 92% at 1976 nm joining the relay fiber to the active fiber.20 An UV-cured low index polymer (fluoroacrylate) was used to fix the active fiber and its end tip onto both water-cooled aluminum spool and copper V-groove to prevent any overheating that could affect the long term stability of the laser. At the very end of the setup, any residual pump is removed with wavelength-selective filters. Since no fiber photo-degradation via hydroxide ion (OH) diffusion was observed at this wavelength, no end-cap was required to protect the cleaved fiber tip.21 

FIG. 2.

Experimental setup of the 3.8 µm monolithic Er3+:ZrF4 DWP all-fiber laser system.

FIG. 2.

Experimental setup of the 3.8 µm monolithic Er3+:ZrF4 DWP all-fiber laser system.

Close modal

A ZrF4 fiber with high erbium doping level (7 mol. %) was used as the active fiber for the setup instead of the usual 1 mol. % concentration for DWP lasers to counterbalance the low gain and high attenuation at the longer operating wavelength of 3.8 µm.12,18 However, higher concentrations of Er3+ increase the heat load applied to the fiber. Both reflectors are thus type-I spectrally narrow FBGs chosen to ensure minimum scattering losses that typically occur with type-II femtosecond FBGs.22 In addition, note that 1 mol. % erbium-doped ZrF4 fibers were initially tested to produce an emission at 3.8 µm but with less success than with the 7 mol. % erbium-doped fibers.

Figure 3 shows the output laser spectrum along with the emission cross section of the 4F9/24I9/2 transition and the standard attenuation of a single-mode ZrF4 fiber. As shown, achieving laser emission from a DWP system at longer wavelengths is challenging because of both the decreasing gain from the lowering emission cross section above 3.55 µm and the increasing attenuation in ZrF4 fibers, resulting in a lower efficiency than for DWP lasers near 3.55 µm. Nevertheless, an emission of 2 W centered at 3789 nm with a FWHM of 0.43 nm was achieved from our system, thanks to the high finesse low losses cavity made by highly reflective FBGs. At this wavelength, the emission cross section of the 4F9/24I9/2 transition is 0.193 × 10−25 m2 and the standard attenuation in ZrF4 fibers is about 110 dB/km.

FIG. 3.

Emission cross section of the 4F9/24I9/2 transition of erbium in ZrF4 (green) along with the attenuation of a typical single-mode ZrF4 fiber (yellow) and the laser spectrum at the output power of 2 W (black) measured with an optical spectrum analyzer.11,12

FIG. 3.

Emission cross section of the 4F9/24I9/2 transition of erbium in ZrF4 (green) along with the attenuation of a typical single-mode ZrF4 fiber (yellow) and the laser spectrum at the output power of 2 W (black) measured with an optical spectrum analyzer.11,12

Close modal

Figure 4 shows the experimental output power of the 3.8 µm laser emission as a function of both the launched P1 and the launched P2. The 2D map actually showcases experimental data produced from the unique combination of both launched pumps with the launched P1 ranging from 3.8 to 21.5 W and the launched P2 ranging from 5.8 to 32.8 W. It allows us to clearly illustrate the unique behavior of the DWP laser in ZrF4 fibers with a high erbium doping level as a function of both launched pumps. The scaling of P2 for P1 < 11.9 W (bottom half of the heat map) exhibits the same behavior previously reported in Ref. 22, which shows an increase in the slope efficiency and thus in the output power and a decrease in the lasing threshold as the launched P1 is increased. For the scaling of the launched P2 for P1 > 11.9 W (upper half of the heat map), the slope efficiency slowly decreases and the threshold slowly increases back as the launched P1 is increased. The quenching of the output power specific to the DWP system, which was to this day only reported with 1 mol. % Er3+:ZrF4 active fibers, is also present for the scaling of P2 for P1 > 11.9 W, but in this case, the quenching appears earlier during the scaling of the launched P2 as the launched P1 is increased, instead of later in ZrF4 fibers with a lower erbium doping level.

FIG. 4.

Experimental heat map of the 3.8 µm output power as a function of both launched pump powers.

FIG. 4.

Experimental heat map of the 3.8 µm output power as a function of both launched pump powers.

Close modal

Figure 5 shows the effect of the increase in the launched P1 on the slope efficiency and threshold of the 3.8 µm laser emission. Figure 5(a) is the same data presented in Fig. 4 for P1 values that exhibit no quenching of the 3.8 µm output power. Figure 5(b), on the other hand, is the same data but as a function of the absorbed P2 power, which was determined by subtracting the residual P2 power to the launched P2. The residual P2 was calculated by measuring the laser output without any wavelength-selective filter and subtracting the residual P1 power measured for P2 = 0 W and the 3.8 µm output power measured with the wavelength-selective filters. The 3.8 µm output power as a function of the absorbed P2 exhibits a lasing behavior with a constant slope efficiency and power threshold of 32.4% and 3.92 W, respectively. Such a behavior arises from the varying absorption of P2 depending on the launched P1. The absorption of P2 increases with the increase in the launched P1. This phenomenon can easily be explained by the fact that larger amounts of the launched P1 result in higher population densities on the VGS caused by more ground state absorption, and higher population densities on this energy level result in more interactions between the launched P2 and ions, so more of the launched P2 will be absorbed. This behavior seems to be unique to DWP lasing with heavily erbium-doped ZrF4 fibers since it has never been reported for DWP lasing with lightly erbium-doped ZrF4 fibers. This phenomenon is also observed in the numerical modeling of the laser system. In fact, the output power of the 3.8 µm emission for different values of the launched P1 as a function of the absorbed P2 of the same laser system with different output couplers (R2 = 56% and R3 = 45%) were also measured and fairly well reproduced by our numerical simulations, as shown in Fig. 6(a). Figures 5(a) and 6(b) also show clearly the quenching of the output power hindering the power scaling ability of the DWP laser system. This phenomenon has been known to limit the power output of such lasers around the emission peak of 3.55 µm, and it can be seen now affecting results at the presented magnitude of output power for the extended wavelength of 3.8 µm.

FIG. 5.

(a) Output power of the 3.8 µm laser for different values of the launched P1 ranging from 5.5 to 11.3 W as a function of the launched P2. (b) Output power of the 3.8 µm laser for different values of the launched P1 ranging from 5.5 to 11.3 W now as a function of the absorbed P2.

FIG. 5.

(a) Output power of the 3.8 µm laser for different values of the launched P1 ranging from 5.5 to 11.3 W as a function of the launched P2. (b) Output power of the 3.8 µm laser for different values of the launched P1 ranging from 5.5 to 11.3 W now as a function of the absorbed P2.

Close modal
FIG. 6.

(a) Output power of the 3.8 µm emission of multiple launched P1 powers for three different output couplers (R1 = 66%, R2 = 56%, and R3 = 45%, respectively) as a function of the absorbed P2 pump with slope efficiency labeled. Results of numerical modeling are represented by solid curves. (b) Output power of the 3.8 µm emission for the launched P1 pump achieving the highest output (P1 = 11.9, 10.8, and 10.8 W, respectively) for the same three output couplers as a function of the launched P2.

FIG. 6.

(a) Output power of the 3.8 µm emission of multiple launched P1 powers for three different output couplers (R1 = 66%, R2 = 56%, and R3 = 45%, respectively) as a function of the absorbed P2 pump with slope efficiency labeled. Results of numerical modeling are represented by solid curves. (b) Output power of the 3.8 µm emission for the launched P1 pump achieving the highest output (P1 = 11.9, 10.8, and 10.8 W, respectively) for the same three output couplers as a function of the launched P2.

Close modal

As shown in Fig. 6(a), lowering the output coupler’s reflectivity from R = 66% to R = 56% and then to R = 45% increases both the efficiency and the lasing threshold with respect to the absorbed P2, which makes sense since an output coupler with a lower reflectivity causes less feedback in the cavity and lets out more signal. Unfortunately, it also causes quenching of the output power to appear at lower launched P2 pump power values, presented in Fig. 6(b), which is detrimental to any power scaling means of reaching the highest output power at 3.8 µm. The solid curves shown in Fig. 6(a) represent the results of the numerical modeling of the system replicating the experimental data of the 3.8 µm laser emission as a function of the absorbed P2 for the three different output couplers. Some parameters of our model had to be adjusted in order to produce a good fit of the experimental data, but these adjustments are deemed reasonable because they are all within uncertainty range of each parameter, like the one predicted by the Füchtbauer–Ladenburg analysis for the cross section parameters.18 As seen in Fig. 6(b), there is no numerical replication of the experimental data for the 3.8 µm laser emission as a function of the launched P2 for the different output couplers because neither the ETU3 nor the VESA parameters (Fig. 1), which were introduced to explain the quenching of the output power in DWP systems made from lightly erbium-doped ZrF4 fibers, are able to replicate numerically the quenching observed in the experimental results presented in this paper without major adjustments that are not within the acceptable range.

The 3.8 µm laser emission was kept operating at a high power for 2 h to measure its stability. P1 and P2 were set at 11.3 and 19.9 W, respectively. The stability test was made at this power level to steer clear of any quenching effect, which affects greatly the output’s stability.18 Over this 2 h period, the output power maintained an average power of 1.07 W and had a root-mean-square (rms) deviation of 0.58%. The emission’s stability was affected mainly by two factors: the rising temperature of the two 790 nm diodes that pump the 1976 nm laser and the periodic air conditioning cycle over around half an hour of the laboratory. The prolonged operation of the 790 nm diodes caused them to heat up and, as a result, to red-shift their operating wavelength. This causes a decrease in the pumping efficiency of the Tm3+:SiO2 fiber laser, which translates into a drop over time of the 1976 nm laser and, then, a similar drop of the 3.8 µm output power because of its reduced pumping. To avoid this problem, the temperature of the 790 nm diodes was kept low with conductive heat transfer methods to stabilize their wavelength. On the other hand, the temperature effects of the periodic air conditioning cycle could not be eliminated and are shown in Fig. 7. Previous analysis of a DWP laser with a heavily erbium-doped fluoride fiber concluded that for power-scaling ends, a lightly erbium-doped fiber is preferable because of the lower heat load applied on the system that puts the splices and FBGs at a lower risk of critical failure.22 The results presented here show that with proper heat load management, a heavily erbium-doped fiber can also deliver stable power-scaling results, since a prolonged high power operation was achieved without any harmful effects on these key components of the setup, such as splices and FBGs, and its stability is comparable to one of a DWP laser made from a lightly erbium-doped fiber.12 

FIG. 7.

Stability test of the 3.8 µm output power over 2 h at about 1 W output power.

FIG. 7.

Stability test of the 3.8 µm output power over 2 h at about 1 W output power.

Close modal

While these results are a major improvement over any other report for the DWP system to this day and are promising for the future of development of a laser system emitting in the transparency window of the atmosphere located near 3.9 µm, there is still room for improvement to achieve higher output powers. The aforementioned issue of high attenuation in ZrF4 fibers in this spectral range could be addressed by using an erbium-doped InF3 fiber instead of the doped ZrF4, which has a much lower attenuation of about 10 dB/km around 3.8 µm.11 Such a design emitting at 3.91 µm has been simulated already, but using the far less efficient single pump scheme with a 635 nm dye laser instead of the DWP scheme.23 

In conclusion, we report an all-fiber Er3+:ZrF4 DWP system with a heavily erbium-doped fluoride fiber achieving a record stable CW emission of 2.0 W in the longest operating wavelength range of 3.8 µm for this system. The all-fiber design of the system provides stability at a high output power, with a rms deviation of only 0.58% over 2 h of operation at around 1 W. We also report a dependency of the 1976 nm pump’s absorption on the 976 nm pump and the quenching of output power with a dependency on the output coupler reflectivity. Both limit the output of the system, and the first one was well reproduced by numerical modeling under various pumping conditions.

This work was supported by Fonds de recherche du Québec - Nature et Technologies (Grand No. FT114976) and Natural Sciences and Engineering Research Council of Canada (Grant Nos. CRDPJ-543631-19 and RGPIN-2016-05877).

The authors have no conflicts to disclose.

Maxime Lemieux-Tanguay: Conceptualization (equal); Data curation (main); Formal analysis (equal); Investigation (equal); Methodology (equal); Software (equal); Validation (equal); Visualization (main); Writing – original draft (main); Writing – review & editing (main). Tommy Boilard: Investigation (equal); Writing – review & editing (equal). Pascal Paradis: Software (equal); Writing – review & editing (equal). Réal Vallée: Conceptualization (equal); Methodology (equal); Supervision (equal); Validation (equal); Writing – review & editing (main). Martin Bernier: Conceptualization (equal); Funding acquisition (main); Methodology (equal); Project administration (main); Resource (equal); Supervision (main); Validation (equal); Writing – review & editing (main).

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
S. D.
Jackson
, “
Towards high-power mid-infrared emission from a fibre laser
,”
Nat. Photonics
6
,
423
(
2012
).
2.
S. D.
Jackson
and
R. K.
Jain
, “
Fiber-based sources of coherent MIR radiation: Key advances and future prospects (invited)
,”
Opt. Express
28
,
30964
31019
(
2020
).
3.
L.-C.
Michaud
,
T.
Boilard
,
S.
Magnan-Saucier
,
P.
Paradis
,
L.
Talbot
,
A.
Thiboult
,
D. F.
Nadeau
,
R.
Vallée
, and
M.
Bernier
, “
Towards real-time active imaging of greenhouse gases using tunable mid-infrared all-fiber lasers
,”
Appl. Opt.
62
,
G69
(
2023
).
4.
A. E.
Klingbeil
,
J. B.
Jeffries
, and
R. K.
Hanson
, “
Temperature- and pressure-dependent absorption cross sections of gaseous hydrocarbons at 3.39 µm
,”
Meas. Sci. Technol.
17
,
1950
(
2006
).
5.
M. R.
Majewski
,
G.
Bharathan
,
A.
Fuerbach
, and
S. D.
Jackson
, “
Long wavelength operation of a dysprosium fiber laser for polymer processing
,”
Opt. Lett.
46
,
600
603
(
2021
).
6.
C.
Frayssinous
,
V.
Fortin
,
J.-P.
Bérubé
,
A.
Fraser
, and
R.
Vallée
, “
Resonant polymer ablation using a compact 3.44 μm fiber laser
,”
J. Mater. Process. Technol.
252
,
813
820
(
2018
).
7.
H.
Ma
,
A.-Y.
Jen
, and
L.
Dalton
, “
Polymer-based optical waveguides: Materials, processing, and devices
,”
Adv. Mater.
14
,
1339
1365
(
2002
).
8.
B.
Molocher
, “
Countermeasure laser development
,”
Proc. SPIE
5989
,
598902
(
2005
).
9.
H.
Bekman
,
J.
van den Heuvel
,
F.
van Putten
, and
H.
Schleijpen
, “
Development of a mid-infrared laser for study of infrared countermeasures techniques
,”
Proc. SPIE
5615
,
27
38
(
2004
).
10.
J. J.
Liu
,
B. L.
Stann
,
K. K.
Klett
,
P. S.
Cho
, and
P. M.
Pellegrino
, “
Mid and long-wave infrared free-space optical communication
,” in
Laser Communication and Propagation through the Atmosphere and Oceans VIII
,
11133
(
SPIE
,
2019
), pp.
1
18
.
11.
Le Verre Fluoré, Le Verre Fluoré Catalog 2023, https://leverrefluore.com/wp-content/uploads/2023/04/LVF-Catalog-2023.pdf (
2023
).
12.
M.
Lemieux-Tanguay
,
V.
Fortin
,
T.
Boilard
,
P.
Paradis
,
F.
Maes
,
L.
Talbot
,
R.
Vallée
, and
M.
Bernier
, “
15 w monolithic fiber laser at 3.55 μm
,”
Opt. Lett.
42
,
289
292
(
2022
).
13.
Y. O.
Aydın
,
V.
Fortin
,
F.
Maes
,
F.
Jobin
,
S. D.
Jackson
,
R.
Vallée
, and
M.
Bernier
, “
Diode-pumped mid-infrared fiber laser with 50% slope efficiency
,”
Optica
4
,
235
238
(
2017
).
14.
V.
Fortin
,
F.
Jobin
,
M.
Larose
,
M.
Bernier
, and
R.
Vallée
, “
10-W-level monolithic dysprosium-doped fiber laser at 324 μm
,”
Opt. Lett.
44
,
491
494
(
2019
).
15.
T.
Boilard
,
V.
Fortin
,
M.
Lemieux-Tanguay
,
P.
Paradis
,
P.
Du Teilleul
,
J.-Y.
Carrée
,
R.
Vallée
, and
M.
Bernier
, “
1.7 W holmium-doped fluoroindate fiber laser at 3920 nm
,”
Opt. Lett.
49
,
2677
2680
(
2024
).
16.
H.
Többen
, “
Room temperature cw fibre laser at 3.5 μm in Er3+-doped ZBLAN glass
,”
Electron. Lett.
28
,
1361
1362
(
1992
).
17.
O.
Henderson-Sapir
,
S. D.
Jackson
, and
D. J.
Ottaway
, “
Versatile and widely tunable mid-infrared erbium doped ZBLAN fiber laser
,”
Opt. Lett.
41
,
1676
1679
(
2016
).
18.
F.
Maes
,
V.
Fortin
,
M.
Bernier
, and
R.
Vallée
, “
Quenching of 3.4 µm dual-wavelength pumped erbium doped fiber lasers
,”
IEEE J. Quantum Electron.
53
,
1
(
2017
).
19.
M.
Bernier
,
D.
Faucher
,
R.
Vallée
,
A.
Saliminia
,
G.
Androz
,
Y.
Sheng
, and
S. L.
Chin
, “
Bragg gratings photoinduced in ZBLAN fibers by femtosecond pulses at 800 nm
,”
Opt. Lett.
32
,
454
456
(
2007
).
20.
S.
Cozic
,
S.
Boivinet
,
C.
Pierre
, and
J.
Boulet
, “
Splicing fluoride glass and silica optical fibers
,”
EPJ Web Conf.
215
,
04003
(
2019
).
21.
N.
Caron
,
M.
Bernier
,
D.
Faucher
, and
R.
Vallée
, “
Understanding the fiber tip thermal runaway present in 3µm fluoride glass fiber lasers
,”
Opt. Express
20
,
22188
22194
(
2012
).
22.
F.
Maes
,
C.
Stihler
,
L.-P.
Pleau
,
V.
Fortin
,
J.
Limpert
,
M.
Bernier
, and
R.
Vallée
, “
3.42 μm lasing in heavily-erbium-doped fluoride fibers
,”
Optics Express
27
,
2170
(
2019
).
23.
M. C.
Falconi
,
A. M.
Loconsole
,
A.
Annunziato
,
S.
Cozic
,
S.
Poulain
, and
F.
Prudenzano
, “
Design of a broadband erbium-doped fluoroindate fiber laser emitting up to 3.91 μm
,”
J. Lightwave Technol.
41
,
6065
6072
(
2023
).