Pressure-independent Brillouin Fiber Optic Sensors for temperature measurements

NOC-16853; No of Pages 4 Journal of Non-Crystalline Solids xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol

Pressure-independent Brillouin Fiber Optic Sensors for temperature measurements C. Sonneville, S. Degioanni, C. Martinet, D. de Ligny, V. Martinez, A.-M. Jurdyc, A. Braunn, L. Raffaelly, B. Champagnon ⁎, D. Vouagner Institut Lumière Matière, UMR5306 Université Lyon 1-CNRS, Université de Lyon, 69622 Villeurbanne Cedex, France

a r t i c l e

i n f o

Article history: Received 30 September 2013 Received in revised form 23 December 2013 Available online xxxx Keywords: Brillouin; Sensors; High pressure; Optical fibers; Silica anomaly

a b s t r a c t Fiber Optic Sensors (FOSs) based on Brillouin scattering are widely used in large infrastructures to detect modifications over large distances. In doped silica fibers the Brillouin Frequency Shift (BFS) is proportional both to temperature and strains. In this work we establish that the sensitivity of FOSs to hydrostatic pressure can be forecast from the behavior of the glass under hydrostatic compressions in a diamond anvil cell. It is shown that the BFS under a hydrostatic pressure is a manifestation of the elastic anomaly observed in silica glass. This anomaly vanishes in GeO2 glass and accounts for the decrease of the sensor sensitivity when the GeO2 doping concentration increases in a silica fiber. The progressive vanishing of the anomaly in sodium aluminosilicate glasses which contain the same amount of silicon dioxide (75%) but differ in the Na2O and Al2O3 ratio allows to determine the composition of a glass with a BFS independent of the pressure. Such a glass composition will provide a pressureindependent temperature FOSs. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The fiber optic domain is at present in considerable expansion and reaches the market place fast. Fiber Optic Sensors (FOSs) have undergone considerable improvement during the last 25 years with developments allowing them to monitor varied environmental parameters (temperature, pressure, strain, humidity, chemicals…) and to be applied in several fields of technology (aerospace, medicine, chemistry, telecommunications…) [1–3]. Among fiber sensor systems sensitive to temperature and strains variations, one of the most extensively developed is based on Brillouin scattering because of its strong dependence on these two environmental factors [4]. Temperature and strain Brillouin distributed fiber sensors are used in civil infrastructure (bridges, railways, land monitoring…), geotechnical structure and pipeline monitoring, and also in larger variety of structures (competition yachts, experimental vehicles, aircrafts…) [5] to detect modifications over large distances (20–30 km). These measurements are performed in the infrared, mainly at 1.550 μm, which corresponds to the minimum attenuation of the silica optical fibers. Brillouin scattering is an inelastic light scattering corresponding to the interaction of light with thermally excited acoustic phonons [5,6] and give information on the elastic behavior of the material.

⁎ Corresponding author. Tel.: +33 472448334. E-mail address: [email protected] (B. Champagnon).

In the case of optical fibers, backscattering geometry is used. In this geometry the Brillouin Frequency Shift (BFS) ν180 is linked to the longitudinal sound velocity (VL) [7]

ν 180 ¼

2n V L λ0

ð1Þ

where λ0 and n are respectively the laser excitation wavelength and the optical index. VL is a function of the density ρ, the bulk modulus K and the Young modulus E according to the relation: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3Kð3K þ EÞ : Vι ¼ ρð9K−EÞ

ð2Þ

In a first approximation the BFS variation (Δν) with temperature (T) and longitudinal strain (ε) can be expressed linearly: Δν ¼ Ct  T þ Cε  ε:

ð3Þ

With C t , the frequency–temperature coefficient and C ε , the frequency–longitudinal strain coefficient. Typical values for silica fibers at λ 0 = 1550 nm are C t = 1 MHz/°C and C ε = 0.05 MHz/με [8].

0022-3093/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jnoncrysol.2014.01.029

Please cite this article as: C. Sonneville, et al., Pressure-independent Brillouin Fiber Optic Sensors for temperature measurements, J. Non-Cryst. Solids (2014), http://dx.doi.org/10.1016/j.jnoncrysol.2014.01.029

2

C. Sonneville et al. / Journal of Non-Crystalline Solids xxx (2014) xxx–xxx

34

Δν p ¼ Ct  T þ Cp  P

ð4Þ

with Cp, the frequency–pressure coefficient. In the case of optical fiber sensors, the stimulated Brillouin scattering (SBS) is often used. SBS is driven through an electrostriction process (material compression under the influence of the electric field) generating an acoustic wave when the fiber core is excited by the incident pump signal [9–11]. Stimulation corresponds to the enhancement of the Brillouin signal by the presence of another intense electromagnetic wave (called the pump wave) that reinforces the spontaneous scattering. With a 10 ns pulse pump laser the time distance conversion allows a resolution of 1 m and measurements on several tens of kilometers [8]. Most of the sensors are based on optical telecommunication fibers and composed of a silica core doped by Germanium dioxide (GeO2). Brillouin optical fiber sensors are sensitive both to strains and temperature effects and it is a challenge to separate these effects as in practical use both occur [12]. Prediction of glass compositions that give rise to strain and temperature independent Brillouin frequency shifts was recently proposed in BaO-silicate fibers [13] and a-thermal Brillouin frequency shift was obtained for sapphire all glass optical fibers containing 38 mol% alumina [14]. In this paper we will discuss the potentiality of silicate based fibers taking into account the role of the fiber composition to monitor the response sensitivity of the Brillouin sensor to hydrostatic pressure. These sensors can be of interest in various high pressure environments like in submarine measurements. This discussion is based on recent works which studied the in situ behavior of the silica BFS by hydrostatic compression experiments performed on silicate glasses and germania glass [15,16] and molecular dynamics simulations [17,18]. The possibility to obtain silicate fibers with hydrostatic pressure independent BFS is shown. 2. Experimental set-up A diamond anvil cell (DAC) high pressure device, Chervin type from the Laboratoire de Physique des Milieux Condensés — University Pierre et Marie Curie (France), equipped with ultra-low fluorescence diamonds allows in situ Brillouin compression experiments. Using liquid argon as a transmitting medium, a quasi-hydrostatic pressure is applied on the samples. The use of a DAC cell, where the external pressure is applied by a membrane loaded with a high pressure gas, allows a very progressive increase or decrease of the pressure applied on the sample with steps down to 0.2 GPa. Brillouin experiments were performed with a JRS Sandercock Fabry–Perot tandem interferometer coupled with a microscope in back scattering geometry using the 532 nm frequency-doubled excitation of a YAG:Nd3+ laser.

Brillouin Fréquency shift (GHz)

For a BFS submitted to a high hydrostatic pressure P a similar formula can be written:

33 32 31 30 29 28 0

1

2

3

4

5

6

Pressure (GPa) Fig. 1. Evolution of the Brillouin frequency shift for pure silica glass during a compression up to 6 GPa. The black arrow corresponds to the initial slope at Patm and is equal to −3.3 GHz/GPa.

(Fig. 2) show that the bulk modulus K of silica decreases when the pressure increases and has a minimum at 7 GPa. This minimum, described as the silica anomaly, is only in qualitative agreement with experimental observations as often observed in molecular dynamics simulations (experimental minimum at 2.5 GPa). The important result obtained from Huang and Kieffer simulations [17] (Fig. 2) is that the bulk modulus decreases for an increasing pressure. The bulk modulus K is related to the BFS through Eqs. (1) and (2). A positive BFS variation Δν (Eq. (3)) as function of the tensile stress increase is observed whereas a negative variation results from a hydrostatic compression (Eq. (4)). This is a manifestation of the elastic anomaly corresponding to an increase of the bulk modulus for negative pressure (Fig. 2). Indeed, for a “normal glass” the elastic moduli and also the longitudinal sound velocity increase when the pressure increases. In a normal glass, the sound velocity decreases as function of the tensile stress intensity. Fig. 2 also shows that the linear approximation (cf Eq. (4)) is only valid close to the atmospheric pressure. Our experimental results can be used to predict the BFS for a compression of an optical fiber at 1.550 μm from the Brillouin shift at 0.532 μm which is equal to − 3.3 GHz/GPa. Taking into account the ratio of both wavelengths the pressure coefficient Cp at 1.550 μm is: Cp ¼ −3:3 ð0:532=1:550Þ ¼ −1:1 GHz=GPa:

3. Results and discussion In the following, we will use the in situ high pressure compressive Brillouin experiments realized on silica and sodium aluminosilicate glasses to predict what would be the pressure sensitivity of Brillouin sensors made with such materials. 3.1. Silica The evolution of pure silica glass BFS at 532 nm during an hydrostatic compression in a DAC up to 6 GPa is shown on Fig. 1. The minimum of the BFS at 2.2 GPa corresponds to the well-known elastic anomaly [19] and the initial slope at Patm is equal to −3.3GHz/GPa. Molecular dynamics simulations by Huang and Kieffer [17] and Mantisi et al. [18] describing silica glass under hydrostatic compression

Fig. 2. Silica anomaly: bulk modulus and density evolution versus pressure determined by Molecular Dynamic Simulations (from [17]).

Please cite this article as: C. Sonneville, et al., Pressure-independent Brillouin Fiber Optic Sensors for temperature measurements, J. Non-Cryst. Solids (2014), http://dx.doi.org/10.1016/j.jnoncrysol.2014.01.029

C. Sonneville et al. / Journal of Non-Crystalline Solids xxx (2014) xxx–xxx

35

7512 Brillouin frequency shift (GHz)

These results are in reasonable agreement with the previous determinations on standard telecommunication single mode fibers −0.91 MHZ/MPa [20] and −0.742 MHz/MPa [21] taking into account that the core doping, the pressure and the wavelength dependences of the refractive index have not been taken into account. The drawing process and the high fictive temperature of the optical fibers could also be considered as a possible reason for this discrepancy but recent experimental Brillouin measurements on silica with different fictive temperatures have demonstrated that these effects are much smaller than the one observed in the present experiments [22]. We can conclude that the behavior of a doped fiber under hydrostatic pressure can be predicted from the study of the glass under hydrostatic pressure. In the following we have applied this statement for GeO2 and sodium aluminosilicate glasses and studied the glass composition influence.

3

34

33

32

31

30

-1

0

1

2

3

4

5

6

Pressure (GPa) Fig. 4. BFS evolution for the 7512 glass during a compression up to 6 GPa.

3.2. Influence of the glass composition 3.2.1. GeO2 glass Silica anomaly is observed not only for silica but also for different glasses. GeO2 glass has a structure similar to silica but it is more compact allowing less free volume. In situ high pressure compression experiments on pure GeO2 glass show that the elastic anomaly almost vanishes (Fig. 3). The BFS of a GeO2 glass is quasi-independent (−0.1 GHz/GPa) of the pressure. This results demonstrate that the doping of the silica core fiber by GeO2 decreases the pressure sensitivity of the BFS. Further more in [12] it is shown that for a tensile stress the sensitivity decreases when the GeO2 content increases (Cε = 0.035 MHz/με at 1.550 μm in a 28% GeO2 doped silica fiber). Zou et al. [23] demonstrated that this decrease is different for stress and temperature coefficients.

3.2.2. Sodium aluminosilicate glasses In the following, we consider three silicate glasses (7500, 7506 and 7512) which contain the same amount of silicon dioxide (75 mol%) but differ in the Na2O and Al2O3 ratio (see [24] for more details on the glass structure). The depolymerization (corresponding to an increase of Non-Bridging Oxygen) increases when the Na2O/Al2O3 molar ratio increases. In a recent paper [16] we show that these three glasses with different degrees of polymerization behave differently during hydrostatic compression. For instance, the silica anomaly has been observed to progressively vanish with the depolymerization corresponding to an increase of Non-Bridging Oxygen due to an increase of the Na2O/Al2O3 ratio.

Figs. 4, 5 and 6 represent the evolution of the BFS during compression up to 6 GPa for the three samples (7512, 7506 and 7500). Intensity of the Brillouin lines at atmospheric pressure is of the same order of magnitude [15]. The slopes at atmospheric pressure (black arrow) have been determined and summarized in Table 1. For the 7512 glass, the elastic anomaly exists with a minimum at 1.5 GPa. From this curve, we can make a linear extrapolation for negative pressure and predict that a silicate fiber with this composition will have a sensitivity for a linear tensile stress much smaller (−0.6 GHz/ GPa) than a pure silica fiber. Figs. 5 and 6 show that the sample containing 19% of Na2O has a smaller elastic anomaly than the sample containing 12% of Na20 whereas the sample containing 25% of Na2O (7500 glass) does not show the elastic anomaly anymore and behaves as a normal glass: i.e. the BFS increases when the pressure increases. Fig. 7 represents the variation of the slopes at atmospheric pressure of the BFS as function of the Na2O content. From this figure, we can predict for example that a glass with approximately 75% of SiO2, 21% of Na2O and 4% of Al2O3 (this composition is represented by a red square in Fig. 7), will correspond to a glass where the BFS is independent of the hydrostatic pressure. It has been shown [23] that in GeO2-doped optical fibers the strain and temperature coefficients increase with different rates with GeO2 concentration. If we make the hypothesis that the Na2O concentration has also a different influence on pressure and temperature coefficients we can predict that a fiber sensor with this particular composition will be a temperature sensor pressure-independent. This composition is likely not unique but defines a domain in the ternary system Al2O3–SiO2–Na2O where this property can be adjusted.

7506 Brillouin frequency shift (GHz)

Brillouin Frequency shift (GHz)

35

GeO2 24

23

34

33

32

31

22 -1

0

1

2

Pressure (GPa) Fig. 3. BFS evolution for pure GeO2 glass during a compression up to 3 GPa.

3

30 -1

0

1

2

3

4

5

6

Pressure (GPa) Fig. 5. BFS evolution for the 7506 glass during a compression up to 6 GPa.

Please cite this article as: C. Sonneville, et al., Pressure-independent Brillouin Fiber Optic Sensors for temperature measurements, J. Non-Cryst. Solids (2014), http://dx.doi.org/10.1016/j.jnoncrysol.2014.01.029

4

C. Sonneville et al. / Journal of Non-Crystalline Solids xxx (2014) xxx–xxx

0,8

35 34

0,6

33

0,4

Slope (GHz/GPa)

Brillouin frequency shift (GHz)

7500

32 31 30

0,2 0,0

Na2O=21%

-0,2 -0,4

29 -0,6 28 -1

0

1

2

3

4

5

12

6

Pressure (GPa)

14

16

18

20

22

24

26

% Na2O

Fig. 6. BFS evolution for the 7500 glass during a compression up to 6 GPa. Table 1 Molar composition in % of the three samples and the initial slope at atmospheric pressure of the Brillouin frequency as function of an hydrostatic pressure. Name

%SiO2

%Al2O3

%Na2O

Slope (GHz/GPa)

7500 7506 7512

75 75 75

0 6 12

25 19 13

0.4 −0.3 −0.6

The important point on which we want to focus is that these Brillouin experiments in a diamond anvil cell are much easier and less time consuming than that of the experiments on fibers to predict range of compositions with monitored strain properties. This procedure can be used to define glass composition of new Brillouin temperature sensors unaffected by compression. 4. Conclusion We have shown in this paper, that the Brillouin fiber-optic sensor sensitivity to compression can be determined from DAC hydrostatic compression experiments. The negative value of the BFS with compression for silica sensors is a manifestation of the silica anomaly; this anomaly is reduced in GeO2 glass. Silicate fibers with different sensitivities to compression can be made by varying the glass composition. A pressureindependent glass composition (75%–SiO2–21%–Na2O–4%–Al2O3) BFS is proposed. These results allow us to anticipate the use of temperature sensors based on Brillouin effect and independent of pressure for example in submarine environment. More generally the approach illustrated above can be extended to various glass compositions which could be studied easily in DAC under hydrostatic pressure. Acknowledgment We would like to acknowledge the Direction Generale de l'Armement (DGA) and Centre National de la Recherche Scientifique (CNRS) for fundings (S.D. doctoral grant) of this work. The authors also thank the Centre COmmun de Microscopie Optique (CECOMO), vibrational spectroscopy platform of Institut de Chimie de Lyon (ICL). These experiments

Fig. 7. Evolution of the slope of the variation BFS as function of pressure and as a function of the Na2O content. The line corresponds to a linear guide for reader's eye and the red square to a composition with a zero slope. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

would not be possible without the samples provided by Professor D. R. Neuville and C. Le Losq. References [1] K. Fidanboylu, H.S. Efendioglu, “Fiber Optic Sensors and Their Applications”, 5th International Advanced Technologies Symposium (IATS'09), May 13–15, 2009, Karabuk, Turkey, May 13–15, 2009. [2] D. Mukherjee, A. De, T. Patra, Int. J. Latest Trends Eng. Technol. 2 (2013) 221. [3] N. Sabri, S.A. Aljunid, M.S. Salim, R.B. Ahmad, R. Kamaruddin, J. Phys. Conf. Ser. 423 (2013) 012064. [4] A. Rogers, in: J.M. Lopez-Higuera (Ed.), Optical Fibre Sensing Technology, Chapter 14, JohnWiley & Sons, New York, USA, 2002. [5] C.A. Galindez-Jamioy, J.M. Lopez-Higuera, J. Sens. (2012) 204121. [6] G. Carlotti, D. Fioretto, L. Palmieri, G. Socino, L. Verdini, E. Verona, IEEE Trans. Ultrason. Ferroelectr. Freq. Control 38 (1991) 56. [7] N. Hayashi, Y. Mizuno, K. Nakamura, Opt. Express 20 (2012) 21101. [8] S. Delepine-Lesoille, S. Mayer, S. Buschaert, J.-P. Dubois, J. Bertrand, E. Merliot, F. Taillade, J.-R. Courivaud, J. Salin, S. Blairon, J.-M. Henault, G. Moreau, Adv. Civ. Eng. (2010) 930796. [9] D.A. Fishman, J.A. Nagel, J. Lightwave Technol. 11 (1993) 1721. [10] J.E. McElhenny, R.K. Pattnaik, J. Toulouse, K. Saitoh, M. Koshiba, J. Opt. Soc. Am. B 25 (2008) 582. [11] A. Boh-Ruffin, NIST-SOFM, 2004. 1. [12] Y.S. Mamdem, Telecom Paris Tech, Ph.D. Thesis 2012. [13] P. Drajic, C. Kucera, J. Furtick, J. Guerrier, T. Hawkins, J. Ballato, Opt. Express 21 (2013) 010924. [14] P. Drajic, T. Hawkins, P. Foy, S. Morris, J. Ballato, Nat. Photonics 6 (2012) 627. [15] C. Sonneville, Université Lyon1, Ph.D. Thesis 2013. [16] C. Sonneville, D. de Ligny, A. Mermet, B. Champagnon, C. Martinet, G.H. Henderson, T. Deschamps, J. Margueritat, E. Barthel, J. Chem. Phys. 139 (2013) 074501. [17] L. Huang, J. Kieffer, Phys. Rev. B 69 (2004) 224203. [18] B. Mantisi, A. Tanguy, G. Kermouche, E. Barthel Eur, Phys. J. B 85 (2012) 304. [19] M. Grimsditch, Phys. Rev. Lett. 52 (1984) 2379. [20] S. Le Floch, Université Bretagne Occidentale, PhD Thesis 2001. [21] H. Gu, H. Dong, G. Zhang, J. He, N. Xu, D.J. Brown, Chin. Opt. Lett. 10 (2012) 100604. [22] V. Martinez, C. Martinet to be published. [23] W. Zou, Z. He, J. Light Wave Technol. 26 (2008) 1854. [24] C. Le Losq, University Paris Diderot, Ph.D. thesis 2012.

Please cite this article as: C. Sonneville, et al., Pressure-independent Brillouin Fiber Optic Sensors for temperature measurements, J. Non-Cryst. Solids (2014), http://dx.doi.org/10.1016/j.jnoncrysol.2014.01.029