Co-doped Dy3+ and Pr3+ Ga Ge Sb S fibers 5 20 10 65 for ...1 Co-doped Dy3+ and Pr3+ Ga 5 Ge 20 Sb 10 S 65 fibers for mid-infrared broad emission J. Ari, 1,2 F. Starecki,1,3 C. Boussard-Plédel,1

1

Co-doped Dy3+ and Pr3+ Ga5Ge20Sb10S65 fibersfor mid-infrared broad emission

J. Ari,1,2 F. Starecki,1,3 C. Boussard-Plédel,1 Y. Ledemi,2 Y. Messaddeq,2 J.-L. Doualan,3 A. Braud,3 B. Bureau,1 And V. Nazabal1

1 Institut des Sciences Chimiques de Rennes, (ISCR), UMR-CNRS 6226, Université de Rennes 1, France2 Centre d’Optique, Photonique et Laser (COPL), 2375 rue de la Terrasse, Université Laval, Québec (Qc), Canada3 Laboratoire CIMAP UMR6252 CEA CNRS Ensicaen, 6 boulevard Maréchal Juin, 14050 Caen Cedex, France

Accepted for publication in Optics Letters, May 2018

Rare earth ion doped materials are means to obtain cost-effective infrared light sources, with enough brilliance for applications such as gas sensing. Within a sulfide matrix, the simultaneous luminescence of both Pr3+ and Dy3+ in the Ga5Ge20Sb10S65 glass is reported.The use of these two rare earths is giving rise to a broad continuous luminescence in the 2.2–5.5 μm wavelength range, which could be used as a mid-infrared light source for gas-sensing applications. The demonstration of CO2 and CH4 detection using a fiber drawn from these materials is reported.

Since chalcogenide glasses have intrinsic low phonon energies, these materials are largely investigated for infrared (IR) applications [1–3]. Rare earth (RE) ions can be embedded in such glass matrices to generate mid-IR emissions, as these low-energies emission lines would be quenched in oxide-based materials. However, the incorporation of an appropriate amount of Ga in the glass composition is required to allow a suitable introduction and dispersion of the RE ions in the glass network [4,5]. Such sulfide-, selenide-, and telluride-based glasses are appropriate to generate the mid-IR emission, leading to a broadband light source required for IR applications such as gas sensing, biomolecules fingerprint detection, and medical diagnostics such as tissue identification and early detection of cancer [6–10]. RE ion doped chalcogenide glass can potentially be used as a laser source operating in the mid-IR. Thus, population inversion outcomes are still investigated, and models are showing potential mid-IR laser operation [6,11–13]. The development of RE ion doped chalcogenide glass lasers is actually still relevant [14]. The broad emissions of RE ions in the mid-IR can also directly be used as incoherent sources because the interest in broad compact luminescence sources in the mid-IR is very high for commercial purposes. Spectrophotometers are mostly relying on blackbody sources, even if more brilliant sources could be proposed in this wavelength domain, such as mid-IR quantum cascade lasers and optical parametric oscillator [15]. Supercontinuum lasers using chalcogenide materials are still under development, and they would provide an intense broad IR signal [16,17]. Gas sensing in the mid-IR has already been achieved for multiple targets, for example, CO2,CO,and CH4 using the combination of optical filters and pyroelectric detectors [18]. Other works were focused on the detection part, as the association of a micro-bolometer and a micro-electro-mechanical systems source, leading to interesting detection ranges and sensitivities [19]. However, some other sensors are using passive chalcogenide to propagate the IR signal through fibers, based on the IR transmission, diffuse reflectance, or evanescent spectroscopy [3,20,21].

Considering sulfide glasses, the Pr3+, Er3+,and Dy3+ mid-IR luminescence was already reported from sulfide GaLaS fibers allowing a high RE concentration by Schweizer et al. [7], while the sulfide Ga5Ge20Sb10S65 matrix has been selected for drawing fibers taking into account its satisfactory thermo-mechanical properties and its higher stability against crystallization [9,22–26]. In Fig. 1, some representative examples of RE lumi-nescence in Ga5Ge20Sb10S65 fibers in the mid-IR wavelength domain are reported, matching some pollutants absorption bands in this region. The Ga5Ge20Sb10S65 glass preform has been doped with 500 ppm of various RE. Under the proper pumping scheme, the Tm3+, Ho3+, and Nd3+ luminescence was recorded at 3.8 μm, 4.0 μm, and 5.1 μm, respectively. Er3+, Dy3+,andPr3+ luminescence in these glasses was already published [8,25–27]. In regard to the gases presented in Fig. 1, these RE could be used for gas detection [9].

Our goal was to activate two RE within the same fiber to produce a broad mid-IR emission band for gas detection. Energy transfers could occur, depending on both RE concentrations and species. Some interactions with Dy3+ ions were effectively described in case of the introduction of a Pr3+, Tm3+, Ho3+, or Tb3+ sensitizer for an enhancement of the 2.95 μm emission from a Dy3+-doped selenide glass matrix [28]. Therefore, a proper combination of two RE and a suitable concentration must be found and could provide a broader spectrum in the mid-IR.

In this Letter, we report on the broadband luminescence between 2.2 and 5.5 μm emitted from a co-doped Pr3+ and Dy3+ chalcogenide fiber. The aim is to use appropriate Pr3+ and Dy3+ doping rates to optimize the mid-IR emission efficiency and to adapt pumping powers to obtain a broad and almost continuous emission between 2.2 and 5.5 μm, which could be suitable for multiple gas sensing. The Ga5Ge20Sb10S65 chalcogenide matrix was selected as the individual emission from Pr3+ and Dy3+ in these sulfide fibers has already been demonstrated successfully [8,9,25]. Depending on the

2

pollutant absorption cross-section and the targeted detection thresholds, this co-doped material is designed for a specific targeted gas, such as a mix of CH4, CO, and CO2 among other gases absorbing in the 2.2–5.5 μm wavelength range.

The Pr3+ and Dy3+ ions were inserted in the Ga5Ge20Sb10S65 sulfide matrix. All glass samples were doped with a fixed 1000 ppmw Pr3+ concentration, while the Dy3+ doping concentration has been varied from 200 to 1000 ppmw. Sulfide glasses were prepared by the conventional melt-quenching method, and this method is described in more detail elsewhere in the literature [8,9]. High-purity Ga (99,99999% Alfa), Ge (99.999% Umicore), Sb (99.999% Alfa), S (99.999% Strem), and for the RE Dy2S3 (99.9%–99.99%) and Pr2S3 (99.9%–99.99%), were used as starting elements, stored and weighted in a glove box under Ar. A 350-μm-unclad single-index fiber is obtained by drawing the Pr3+∕Dy3+-doped sulfide glass preform, and 4-mm-thick bulk samples were also prepared for spectroscopic characterizations.

The absorption coefficient of the 1000 ppm Pr3+- and 1000 ppm Dy3+-doped sulfide glass as a function of the wavelength is reported in Fig. 2(a). The absorption coefficient spectrum of the Dy3+∕Pr3+ co-doped sample results from the perfect overlap of the individual absorption coefficient of the Dy3+ and Pr3+ ions considering the same doping concentration in the Ga5Ge20Sb10S65 bulk glasses [8,9]. At the wavelength of 1700 nm, the 6H11∕2 level from Dy3+ and the 3F3 level from Pr3+ absorption bands are overlapped.

For a simultaneous pumping of the Dy3+ and Pr3+ ions embedded in the sulfide glass fiber, two laser sources have been synchronized to achieve the dual wavelength optical pumping. Two silica fibers (105/125 μm) were bundled into a single FC/PC connector and then connected to a FC/PC-terminated 200 μm single-index silica fiber. The pumping beams from the laser diodes are then mixed in the 1-m-long 200 μm silica fiber, and they are injected in the 350 μm single-index chalcogenide fiber, as illustrated in Fig. 2(b).

Co-pumping operation was performed by synchronizing two Thorlabs ITC4020 laser diode controllers. The Lumics LU0915T090 and Q-Photonics QSM-1550-3 laser diodes are the pumping sources, emitting at the wavelengths of 920 and 1550 nm, respectively. Some pump beam coupling losses arise when mating the two bundled fibers to the 200 μm fiber, and the two laser diodes pumping powers were measured and adjusted at the output of the 200 μm silica fiber. This co-injection device was used to pump either a chalcogenide bulk or a 350-μm-diameter fiber.

Emission properties of this Pr3+∕Dy3+: Ga5Ge20Sb10S65 glass material were investigated, and fluorescence lifetimes were recorded on bulk samples. For the Dy3+ lifetimes measurement, 4.3 μm–200 nm (central wavelength: FWHM) and 3.05 μm–100 nm filters were used for the 6H11∕2 and 6H13∕2 emissions, respectively. For the Pr3+ mid-IR emitting manifold lifetimes, 2.36 μm–80 nm and 4.64 μm–260 nm filters were used to isolate the 3H6 and 3H5 emissions, respectively. The recorded lifetimes are reported in Table 1.

The Dy3+ and Pr3+ lifetimes in these co-doped Ga5Ge20Sb10S65 glasses are equivalent to singly doped glasses containing 1200, 1500, and 2000 ppm of either praseodymium or dysprosium, for which the lifetime is decreasing with concentration [22,27]. This statement is based on the decrease of the Pr3+ and Dy3+ fluorescence lifetime as the Dy3+ doping rate is increasing. Also, considering the relatively small doping amount of Dy3+

ions and the relatively significant lifetime decrease, some energy transfers between the Dy3+ and Pr3+ ions must take place, thereby influencing the fluorescence lifetimes of each ion.

Emission spectra of the drawn fibers were recorded using a Horiba MicroHR monochromator equipped with gold-coated optics. The infrared detector is a nitrogen-cooled Hamamatsu InAsSb chip, connected to a pre-amplifier. The electrical signal is then analyzed using an EG&G DSP7265 lock-in amplifier, and the modulation frequency is 21 Hz. The same grating and detector were used for all these measurements. The proper measurement on such a broad wavelength range requires to record spectra free of harmonic contributions. Three long-pass filters were used for this purpose: a germanium wafer used as 2 μm long-pass filter, and two long-pass Spectrogon LP-2800 nm and LP-3300 nm filters. The acquisition chain spectral response was calibrated using a referenced Arcoptix ArcLight MIR black-body for each filter used. The almost continuous emission spectrum of the 1000 ppmw Pr3+/200 ppmw Dy3+-doped Ga5Ge20Sb10S65 glass fiber obtained from a 2 to 6 μm spectral range is presented in Fig. 3 plotted with a logarithmic scale for presentation clarity.

As can be seen in Fig. 3, the Pr3+∕Dy3+ sulfide fiber emission is more intense in the 2.2–3.3 μm region than in the 3.3–5.5 μm region. This was already noticed for both singly doped glass bulks and fibers. In the 2.2–3.3 μm wavelength range, the 2.4 μm (labeled a) and 2.9 μm (labeled g) emission bands are attributed to radiative transitions of the Pr3+ ( 3H6 → 3H4) and Dy3+ (6H13∕2 → 6H15∕2), respectively. A shoul der at ≈3.2 μm (labeled d, e) can be ascribed to a minor contribution from the Dy3+ (6H5∕2;

6F7∕2 → 6H9∕2; 6F11∕2) and ( 6H7∕2;

6F9∕2 → 6H11∕2) manifolds. These transitions are mostly phonon assisted, and the radiative transitions with low intensity contribute to form a quasi-continuous emission spectrum. The 3.3–5.5 μm wavelength range shown in Fig. 3 is composed of many overlapping emission bands: (i) the Pr3+:3H6 → 3H5 (3.6 μm, labeled b), (ii) the Dy3+:6H11∕2 → 6H13∕2 (4.3 μm, labeled f),(iii) the Pr3+:3H5 → 3H4 (4.8 μm, labeled c) with maybe a minor contribution from the Dy3+:6H9∕2 → 6H11∕2 (shoulder at 5.1 μm). The first minimum of emission intensity is observed at the wavelength of 3.4 μm, corresponding to the overlap of the luminescence tail from the Dy3+ emission at 2.9 μm and the rise of the Pr3+ emission at 3.6 μm[8,9]. With a single pumping beam, the weak intensity at 3.4 μm persists in both the cases and is not affected by the co-doping. On the other hand, when the two pumps are switched on, the summation of the Dy3+ and Pr3+ fluorescence intensities give rise to a non-negligible fluorescence signal. The minimum remains at the wavelength of 3.42 μm. A spectrum with a resolution of 10 nm was also recorded, thus demonstrating the continuity of the fluorescence even in this low-intensity domain. Moreover, the lowest fluorescence intensity recorded at the wavelength of 4.0 μm corresponds to reabsorbed fluorescence by the S–H bond impurities. Thus, this Dy3+∕Pr3+ co-doped sulfide glass shows a quasi-continuous emission in the 2.2–5.5 μm wavelength spectral range.

From this broad mid-IR emission, several gases could be de-tected using such a co-doped chalcogenide fiber. A sensor has been developed to achieve a suitable measurement of CO2 and CH4 at their 4.3-μm- and 3.4-μm-centered absorption bands, respec-tively. The CO2 monitoring from the use of the Dy3+ fluorescence has already been demonstrated [9], but the multiple gases detection using a RE- ion-co-doped single-emitting chalcogenide fiber has not been proposed yet. The Dy3+∕Pr3+ co-doped sulfide luminescent mid-IR sensor scheme is described in Fig. 4.

The 45 cm long Z -folded gas cell with a regulated gas flow formed by two gold-coated mirrors is equipped with CaF2 1” wafers dispatched on its sides. The mid-IR emitted beam from the Pr33+∕Dy3+:Ga5Ge20Sb10S65 fiber is collimated and passed through the gas cell. Exiting the gas cell, the mid-IR signal is then alternatively passed through two band-pass filters matching the CO2 and CH4 absorptions. The central wavelength and the

3

FWHM of these filters are 4.3 μm (200 nm) and 3.4 μm (217 nm), respectively. Then, the signal is focused on the IR detector. To acquire a pair of the CH4∕CO2 transmitted intensities, the whole process monitored by a Labview program lasts 8 s, as long delays were used here due to the lock-in amplifier time constant. The gases outlet is connected to an ADC LFG20 monitoring device. The CH4 rate could be adjusted by controlling the Ar and CH4 flowrates. Two gas bottles of 200 ppmw and 1500 ppmw of CO2 were used. The operation of the Pr3+∕Dy3+:Ga5Ge20Sb10S65 fiber as the mid-IR source for CH4 and CO2 simultaneous detection is reported in Fig. 5. Labels (0) to (7) correspond to opening or closing of gas valves. At t = 0 min, the cell is first purged using Ar (0). Then (Segment 1), the CH4 valve is opened, and the Ar − CH4 mix is adjusted to about 4.5 vol. % of CH4 using the two flow-meters. Then, the gas cell is flushed with Ar (Segment 2), and CO2 and CH4 signal intensities are back to the reference values. Afterward, the analyte is changed to the 1500 ppmw CO2 (Segment 3). The gas inlet line is the same as that used for CH4, and the remaining volume of the line (containing pure CH4) is pushed into the cell when the valve is opened, which is remarkably noticable in the CH4 measurement line (3). The 1500 ppmw CO2 mix is progressively filling the gas cell and then reaching the equilibrium. The 200 ppmw CO2 is then injected (Segment 4). During these CO2 injections, except when the remaining CH4 intensity had been pushed into the gas cell, the CH4 line remains stable at its Ar-purged cell level, proving the ability of this detector at measuring separately the concentrations of two gases. When the 200 ppmw CO2 signal becomes stable, the gas cell is purged using Ar (Segment 5). Afterward, a CH4 injection has been performed at the rate fixed at the beginning of the experiment. The CH4 is stabilized at 4.7 vol.% and slowly decreased to 4.5 vol.%. Both the LFG20 commercial sensor and our mid-IR luminescent sensor were measuring this slight change in the CH4 concentration, and then a last Ar purge ended up this experiment.

Considering the signal-to-noise ratio for both gases measurements, the sensitivity could be estimated under some approxima-tions. For this assessment, we assume a linear response of the mid-IR sensor to the gas concentration changes at low concentrations. The sensitivity (S) is defined as S = 2RΔB∕ΔS,where Sis the sensitivity, R the calibrated gas concentration, ΔB the highest noise measured, and ΔS the signal difference between the intensities with and without gas presence. Accordingly, the CO2 and CH4 sensitivities are of 15 ppm and 300 ppm, respectively.

To conclude, these Pr3+∕Dy3+:Ga5Ge20Sb10S65 fibers are suitable materials for the development of mid-IR sensor devoted to gas detection. The quasi-continuous emission generated has been demonstrated in the 2.2–5.5 μm wavelength range arising from the luminescence addition of the Pr3+ and Dy3+ ions doping sulfide fibers with some minor energy transfer effects. The efficient use of bundled silica fibers enables the simultaneous injection of two wavelengths for a dual pumping at 920 and 1550 nm of the Pr3+∕Dy3+ co-doped chalcogenide fibers. Gas detection is demonstrated with the use of 200 ppmw of CO2 and 4.5 vol.% of CH4, leading to detection sensitivities of 15 ppm and 300 ppm, respectively. The detection is not limited to CO2 or CH4, but potentially any of the absorption bands presented in Fig. 1 could be detected simultaneous using the Pr3+ and Dy3+ GaGeSbS fiber luminescence with some modifications, including the monochromator use. The Letter illustrates theses potentialities through the example of CO2 or CH4.

Funding

Agence de l’Environnement et de la Maîtrise de l’Energie (ADEME) (COPTIK); Agence Nationale de la Recherche (ANR) ((ANR-15-CE39-0007).

Acknowledgment

ADEME COPTIK and Optigas pro-jects (ANR-15-CE39-0007), Brittany region and CERC in Photonic Innovations for PhD fundings. We thank the IDIL Fibres Optiques company for the silica fiber optical bundle on demand development.

References

1. J. L. Adam, L. Calvez, J. Troles, and V. Nazabal, Int. J. Appl. Glass Sci. 6, 287 (2015).

2. B. J. Eggleton, B. Luther-Davies, and K. Richardson, Nat. Photonics 5, 141 (2011).

3. J. S. Sanghera and I. D. Aggarwal, J. Non-Cryst. Solids 257, 6 (1999).

4. J. K. Kim, B. Kyou Jin, W. J. Chung, B. J. Park, J. Heo, and Y. G. Choi, J. Phys. Chem. Solids 72, 1386 (2011).

5. B. G. Aitken, C. W. Ponader, and R. S. Quimby, C. R. Chim. 5, 865 (2002).

6. L. B. Shaw, B. Cole, P. A. Thielen, J. S. Sanghera, and I. D. Aggarwal, IEEE J. Quantum Electron. 37, 1127 (2001).

7. T. Schweizer, D. W. Hewak, B. N. Samson, and D. N. Payne, J. Lumin.72–74, 419 (1997).

8. R. Chahal, F. Starecki, C. Boussard-Pledel, J.-L. Doualan, K. Michel, L. Brilland, A. Braud, P. Camy, B. Bureau, and V. Nazabal, Sens. Actuators B 229, 209 (2016).

9. F. Starecki, F. Charpentier, J. L. Doualan, L. Quetel, K. Michel, R. Chahal, J. Troles, B. Bureau, A. Braud, P. Camy, V. Moizan, and V. Nazabal, Sens. Actuators B 207, 518 (2015).

10. S. Cui, R. Chahal, Y. Shpotyuk, C. Boussard, J. Lucas, F. Charpentier, H. Tariel, O. Loreal, V. Nazabal, O. Sire, V. Monbet, Z. Y. Yang, P. Lucas, and B. Bureau, Optical Fibers and Sensors for Medical Diagnostics and Treatment Applications XIV (2014), Vol. 8938.

4

11. R. S. Quimby, L. B. Shaw, J. S. Sanghera, and I. D. Aggarwal, IEEE Photon. Technol. Lett. 20, 123 (2008).

12. S. S. Kim, C. Young, and B. Mizaikoff, Anal. Bioanal. Chem. 390, 231 (2008).

13. L. Sojka, Z. Tang, H. Zhu, E. Beres-Pawlik, D. Furniss, A. B. Seddon, T. M. Benson, and S. Sujecki, Opt. Mater. Express 2, 1632 (2012).

14. Z. Q. Tang, D. Furniss, M. Fay, H. Sakr, L. Sojka, N. Neate, N. Weston, S. Sujecki, T. M. Benson, and A. B. Seddon, Opt. Mater. Express 5, 870 (2015).

15. C. Gmachl, H. Y. Hwang, R. Paiella, D. L. Sivco, J. N. Baillargeon, F. Capasso, and A. Y. Cho, IEEE Photon. Technol. Lett. 13, 182 (2001).

16. R. A. Martinez, G. Plant, K. W. Guo, B. Janiszewski, M. J. Freeman, R. L. Maynard, M. N. Islam, F. L. Terry, O. Alvarez, F. Chenard, R. Bedford, R. Gibson, and A. I. Ifarraguerri, Opt. Lett. 43, 296 (2018).

17. U. Moller, Y. Yu, I. Kubat, C. R. Petersen, X. Gai, L. Brilland, D. Mechin, C. Caillaud, J. Troles, B. Luther-Davies, and O. Bang, Opt. Express 23, 3282 (2015).

18. Q. Tan, L. Tang, M. Yang, C. Xue, W. Zhang, J. Liu, and J. Xiong, Opt. Lasers Eng. 74, 103 (2015).

19. P. Barritault, M. Brun, O. Lartigue, J. Willemin, J.-L. Ouvrier-Buffet, S. Pocas, and S. Nicoletti, Sens. Actuators B 182, 565 (2013).

20. X. Dai, X. Liu, L. Liu, B. Zhu, and Z. Fang, Sens. Actuators B 220, 414 (2015).

21. J. D. Albert, V. Monbet, A. Jolivet-Gougeon, N. Fatih, M. Le Corvec, M. Seck, F. Charpentier, G. Coiffier, C. Boussard-Pledel, B. Bureau, P. Guggenbuhl, and O. Loreal, Joint Bone Spine 83, 318 (2016).

22. R. Chahal, F. Starecki, C. Boussard-Plédel, J.-L. Doualan, K. Michel, L. Brilland, A. Braud, P. Camy, B. Bureau, and V. Nazabal, Sens Actuators B 229, 209 (2016).

23. A. L. Pele, A. Braud, J. L. Doualan, F. Starecki, V. Nazabal, R. Chahal, C. Boussard-Pledel, B. Bureau, R. Moncorge, and P. Camy, Opt. Mater. 61, 37 (2016).

24. A. L. Pelé, A. Braud, J. L. Doualan, R. Chahal, V. Nazabal, C. Boussard-Plédel, B. Bureau, R. Moncorgé, and P. Camy, Opt. Express 23, 4163 (2015).

25. F. Charpentier, F. Starecki, J. L. Doualan, P. Jovari, P. Camy, J. Troles, S. Belin, B. Bureau, and V. Nazabal, Mater. Lett. 101, 21 (2013).

26. V. Moizan, V. Nazabal, J. Troles, P. Houizot, J.-L. Adam, J.-L. Doualan, R. Moncorgé, F. Smektala, G. Gadret, S. Pitois, and G. Canat, Opt. Mater. 31, 39 (2008).

27. F. Starecki, S. Morais, R. Chahal, C. Boussard-Pledel, B. Bureau, F. Palencia, C. Lecoutre, Y. Garrabos, S. Marre, and V. Nazabal, Int. J. Greenhouse Gas Control 55, 36 (2016).

28. B. J. Park, H. S. Seo, J. T. Ahn, Y. G. Choi, J. Heo, and W. J. Chung, J. Ceram. Soc. Jpn. 116, 1087 (2008).

5

6

7

8

9

10