- Email: [email protected]
X-ray irradiation effects on a multistep Ge-doped silica fiber produced using different drawing conditions
Journal of Non-Crystalline Solids 357 (2011) 1966–1970
Contents lists available at ScienceDirect
Journal of Non-Crystalline Solids j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j n o n c r y s o l
X-ray irradiation effects on a multistep Ge-doped silica fiber produced using different drawing conditions A. Alessi a,⁎, S. Girard b, C. Marcandella b, S. Agnello c, M. Cannas c, A. Boukenter a, Y. Ouerdane a a b c
Laboratoire H. Curien, UMR CNRS 5516, Université Jean Monnet, 18 rue du Pr.Benoît Lauras 42000, Saint-Etienne, France CEA, DAM, DIF, F-91297 Arpajon, France Dipartimento di Scienze Fisiche ed Astronomiche, Università di Palermo, I-90123 Palermo, Italy
a r t i c l e
i n f o
Article history: Received 29 June 2010 Received in revised form 13 October 2010 Available online 14 December 2010 Keywords: Optical fibers; Drawing effects; Radiation effects; Point defects
a b s t r a c t We report an experimental study based on confocal microscopy luminescence (CML) and electron paramagnetic resonance (EPR) measurements to investigate the effects of the X-ray (from 50 krad to 200 Mrad) on three specific multistep Ge doped fibers obtained from the same preform by changing some of the drawing conditions (tension and speed). CML data show that, both before and after the irradiation, Germanium Lone Pair Center (GLPC) concentrations are similarly distributed along the diameters of the three fibers and they are partially reduced by irradiation. The irradiation induces also the Non Bridging Oxygen Hole Center (NBOHC) investigated by CML and other paramagnetic defects as the Ge(1), Ge(2), E'Ge and E'Si investigated by EPR. We do not observe significant differences in the induced concentrations of these types of defects in the three fibers. All the results suggest that within the range of investigated drawing parameters (usual range for specialty optical fibers), the fiber radiation sensitivity is unaffected by these variations. Consequently, these results show that the drawing parameters (drawing tension and speed) cannot easily be adjusted to improve the radiation hardness of germanosilicate optical fibers. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Silica based optical fibers are used in several technological application fields, telecommunication [1], radiation environments [2], sensors and other optical device production [1]. The fiber design and elaboration processes (dopant types, concentration, drawing conditions…) determine their optical features [3]. It is known that the Germanium doping increases the refractive index of the silica and for this reason it is commonly used to realize the core of the fibers [1]. As regards the production procedures, previous studies have shown that the drawing process can induce the generation of point defects that can act as precursor sites for the generation of optically-active defects under irradiation [4,5]. Since for many fiber-based applications in radiative environments, the radiation response of Ge-doped silica is a crucial aspect, so an important research field is devoted to this domain. Different investigations have shown that the radiation induces new optical absorption bands [6–9], thus increasing the attenuation of the fiber [10], and that it is due to the generation or the conversion of point defects in the glass network [1,6–8]. For this class of materials, the Germanium Lone Pair Center (GLPC) [11] has been widely studied, since a relation between the bleaching of its associated optical absorption (OA) band, peaked at 5.1 eV [1], and
⁎ Corresponding author. E-mail address: [email protected] (A. Alessi). 0022-3093/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jnoncrysol.2010.10.038
the photosensitivity of the Ge-doped silica glass [1,12] has been observed. The GLPC is constituted by a Ge atom bonded with two oxygen atoms by two single bonds, and with an electron lone pair (NGe••, N stands for the two single bonds, whereas • represents an electron) [1,11,13]. The GPLC is also responsible for two photoluminescence (PL) bands which peaked at ~4.3 eV and at ~3.2 eV [1,13]. According to Ref. [13] the 5.1 eV OA band is due to the S0 → S1 transition (S0 being the singlet electronic ground state and S1 the first singlet excited electronic state), the 4.3 eV PL band is due to the inverse transition and the 3.2 eV PL band is related to the T1 → S0 transition (being T1 the first triplet excited electronic state supplied through S1 by an inter-system crossing process [13]). For the data reported in the following it is also important to note that an OA band peaked at ~3.8 eV has been associated with the S0 → T1 transition [7]. In addition to the GLPC and its bleaching, other defects as the Ge(1), Ge(2) and the E'Ge and their generation processes were investigated [1,14–16]. The interest for these defects is due to the proposed relations among them and the induced OA bands [6–8] and to the debate of their relevance regarding transmission attenuation, photosensitivity and non linear effects [1,9,10,17–20]. All these defects are paramagnetic; in particular, the Ge(1) is made up of a fourfold coordinated Ge atom trapping an electron (NGe•b) [6], the E'Ge is formed by a threefold coordinated Ge atom with an unpaired electron ( Ge•) [21]. At variance, for the Ge(2) two structural models have been proposed. In the first one they have a structure similar to the Ge(1), differing for the presence of a second Ge atom as second
A. Alessi et al. / Journal of Non-Crystalline Solids 357 (2011) 1966–1970
1967
neighbor [6]. In the other one the Ge(2) is considered as a ionized GLPC (NGe•) [8]. In this work, we study the influence of X-ray irradiations on a set of specific germanosilicate fibers obtained from the same preform changing drawing conditions. To thoroughly investigate the effects of the drawing conditions on the radiation response of these fibers we also studied the radiation induced amounts of the Non Bridging Oxygen Hole Center (NBOHC Si–O•) [1,22] and of the E'Si ( Si•) [1,23].
We also recorded, at room temperature, electron paramagnetic resonance (EPR) spectra using a Bruker EMX-Micro Bay spectrometer working at 9.8 GHz. The defect concentrations were estimated using the EPR signal recorded in a bulk sample with a known concentration of E'Si [24]. The concentrations are estimated with an absolute accuracy of 50% and a relative accuracy of 20% [25].
2. Experimental
In Fig. 2a we report the emission amplitude at 3.2 eV along a fiber diameter recorded in all three investigated samples before irradiation and after a dose of ~100 Mrad. The data indicate that the GLPC contents are similar in the three samples and that their concentration reaches a maximum at the center of the higher doped region. It is important to note that even if the GLPC/Ge ratio in the 9 wt.% doped zone is not constant, it does not depend on the investigated fiber. Such an enhanced generation of defects was previously noticed in the core centers of irradiated samples of pure silica-core step-index fibers [26]. As regards the irradiation effects reported in Fig. 2b, we note that the emission of the GLPC shows a decreasing tendency while the dose increases. In Fig. 2b the point at 0.1 krad represents the initial
We have studied three different types of fibers produced by iXFiber SAS starting from the same preform (made by the modified chemical vapor deposition process) making major changes in the drawing conditions; they will be called FGeD1, FGeD2 and FGeD3 in the text. The Ge doping level along the fiber diameters (~125 μm) was specifically designed to follow a two-step distribution. The first step is doped with ~4.5 wt.%, whereas the second with ~9 wt.% (see Fig. 1). The FGeD1 was obtained using a drawing speed of 70 m/min and a tension of 135 g, the FGeD2 was produced using a drawing speed of 40 m/min and a tension of 70 g, whereas for the FGeD3 they were 22 m/min and 33 g respectively. These drawing conditions have been chosen to cover the range of tension and speed usually applied for the drawing of specialty fibers like radiation-tolerant waveguides, rare-earth doped or polarisationmaintaining optical fibers. However, our experimental procedure does not cover the range of drawing parameters used for the making of Telecom-type fibers. The samples were X-ray (10 keV) irradiated at room temperature with the ARACOR machine at the French atomic energy center (CEA). We recorded confocal microscopy luminescence (CML) measurements to check and follow the evolution of GLPC and NBOHC distribution along the fiber cross section diameter before and after the irradiations. These measurements were acquired using an Aramis (Jobin-Yvon) spectrometer, with a He–Cd ion laser excitation line (energy 3.8 eV, power ~ 0.15 mW) and supplied with a CCD camera, microtranslation stages and a 40x objective. The employed experimental conditions lead to a spatial resolution of ~2 μm. The system adopts a back-reflected geometry. For all the measurements the excitation beam penetrates a few micrometers into the sample. Furthermore we note that using the CML technique it is possible to filter the luminescence from a selected sample region, in this way possible light guiding effects along the fiber diameter are strongly limited.
3. Results
a
b
Fig. 1. Ge doping profile of the three fiber types.
Fig. 2. a) Emission at 3.2 eV as a function of the distance from the fiber center: ( ) FGeD1, (●) FGeD2 and (□) FGeD3 before irradiation, ( ) FGeD1, (▬) FGeD2 and ( ) FGeD3 at ~100 Mrad; b) emission at 3.2 eV recorded at the center of the fibers as a function of the dose ( ) FGeD1, (●) FGeD2 and (□) FGeD3.
1968
A. Alessi et al. / Journal of Non-Crystalline Solids 357 (2011) 1966–1970
amplitude. Even if the error of the emission measurements does not allow us to study such dose dependency in depth, no difference appears in the three investigated fibers. Furthermore considering all the data we deduce that at the dose of 100 Mrad the GLPC concentration is ~30% lower in all the fibers. Finally, comparing the GLPC profiles before and after irradiation we remark that they are very similar. In addition to the 3.2 eV band from the dose of 2 Mrad in the spectra an emission band peaked at ~1.9 eV was also detected. This emission is attributed to the NBOHC [1,22] and in Fig. 3 we report, as a function of the distance from the fibers' core centers, its amplitude at the dose of 100 Mrad. The data show that the NBOHC content is higher in the Ge doped zone, with a maximum at the core centers (similar behavior is observed for lower doses). In the inset of Fig. 3 the 1.9 eV amplitude recorded at the center of the fibers is reported as a function of the dose, we note that no significant differences in the dose dependence are observed in the three fibers and that the NBOHC content increases with a sub-linear dependence on dose. Fig. 4 illustrates the EPR spectra recorded at the dose of 2 Mrad for all the fibers. The spectra are normalized regarding the measurement conditions and the sample weights. We note that the signal amplitude is similar in the three fibers, and we also observe that the line-shape appears independent from the sample. In fact, in all the spectra we can observe the presence of the peaks at ~ 346 mT and at ~ 348.5 mT related Ge(1) and Ge(2), respectively. The identification of the E'Ge is not as easy since their lower concentration and the strong overlapping of the signals [27]. It is important to underline that no detectable concentrations of Ge(1) and Ge(2) have been observed in the fibers before irradiation, whereas an E'Ge concentration of ~ 2 × 1016 defects/ cm3 was detected in all three types of the pristine fibers. The Ge(1) induced concentrations are reported in Fig. 5a as a function of the irradiation dose. We note that, in the three fibers, the Ge(1) defect concentration linearly increases with the dose up to ~ 200 krad and then it tends to reach a limit value of ~ 4 × 1017 defects/ cm3. Similar dose dependence is observed in Fig. 5b for the Ge(2) with a saturation value of ~5 × 1017 defects/cm3. On the contrary, the E'Ge concentration increases with the dose following a sub-linear law without showing a saturation tendency (see Fig. 5c). Even if the dose dependence of the E'Ge is different from those of the Ge(1) and Ge(2) it is important to note that no significant differences are observed in the behavior of the three different fibers. In Fig. 5c we do not report the concentration induced at the first two
Fig. 3. Emission at 1.9 eV as a function of the distance from the fiber center: ( ) FGeD1, (-●-) FGeD2 and (-□-) FGeD3 at the dose of ~100 Mrad; inset emission at 1.9 eV recorded at the center of the fibers as a function of the dose ( ) FGeD1, (●) FGeD2 and (□) FGeD3.
Fig. 4. EPR spectra at the dose of ~ 2 Mrad for the FGeD1 ( FGeD3 ( ).
), FGeD2 (▬) and
doses since the total measured E'Ge concentrations are too close to those of the pristine samples. We note that the here reported dose dependences agree with the previous investigations [28,29]. Ge(1) and Ge(2) concentrations are determined by a competition between generation and destruction mechanisms involving precursors. By contrast the E'Ge and NBOHC kinetics suggest generation mechanisms involving T–O–T sites (with T = Si/Ge) [30]. Finally, as regards the E'Si, we note that their concentration (under a straightforward model the defects are supposed uniformly distributed in the fibers) starts to be measurable from the dose of 2 Mrad, since the overlapping of the E'Si EPR signal with that of the Ge-related defects. As for the other defects, it is important to underline that no significant differences are observed in the three fibers. The maximum induced concentration is observed at the dose of 200 Mrad and it is ~5 × 1016 defects/cm3 and we bring to mind that the concentration depends on a dose with a law similar to that of the E'Ge. 4. Discussion In a previous study [5], we showed a noticeable difference in the GLPC concentrations between a Ge-doped preform and its corresponding optical fiber. In the case of the fiber, the GLPC generation seems to be enhanced in its core during the drawing process. The data of Fig. 2 show that for the investigated variations of the drawing conditions, the three tested fibers have very similar preirradiation GLPC concentrations (panel a). As a consequence, it appears not possible to reduce the number of pre-existing GLPC defects in the fibers by varying the drawing speed and tension inside the studied range. This is also true for the other investigated precursor or radiation-induced defects: Ge(1), Ge(2), E'Ge, NBOHC and E'Si. All are identically generated in the three fibers. The radiation sensitivity independence from the drawing conditions could seem in contrast to the data reported in Ref. [10]. In that work, we showed that the fiber was more sensitive (in terms of the paramagnetic Ge-related defect generation) at low doses (~200 krad) than its corresponding preform, whereas minor differences were observed at higher doses (~200 Mrad). By changing the drawing condition, a difference in the radiation sensitivity at low doses could be expected. The reported absence of such differences suggests that the variations in the drawing conditions used to obtain the investigated samples, even if quite important for the fiber production, are too small to induce significant changes in the characteristics before or after
A. Alessi et al. / Journal of Non-Crystalline Solids 357 (2011) 1966–1970
a
1969
Furthermore, even if specific studies of the structure and stress of the examined fibers are postponed in future investigations, we can tentatively suppose that the fictive temperature of the three fibers is similar, since in Ref. [35] it was shown that significant variations of this quantity require greater changes in the drawing parameters. From the practical point of view our data indicate that larger changes of drawing tension and speed are required to modify the generation of defects and of the defect precursors to improve the radiation hardness, it is also important to consider the relevance of other drawing parameters. For these reasons the drawing temperature, the atmosphere applied during the drawing, the chemical composition of the fiber, the efficiency of pre-treatments like thermal annealing or hydrogen-loading have to be considered and investigated to design radiation-tolerant germanosilicate glasses. Finally, considering our data and those reported in Ref. [35] we suggest that between the changes in the drawing tension and variations of the drawing speed, these latter appear the most promising for the control improvement of the fiber radiation sensitivity. 5. Conclusions
b
We have reported an experimental investigation regarding three fibers produced from the same preform and differing from the drawing conditions. The study has been devoted to the defects generation–evolution under X-ray irradiation. We have investigated the signals of GLPC, NBOHC, Ge(1), Ge(2), E'Ge and E'Si; from this data the close similarity in the initial values of the various defect concentrations and analogous radiation responses of the fibers have been highlighted. Based on these findings, we suggest that the explored changes in the drawing parameters, usually employed for specialty fibers, do not affect the fiber features. Acknowledgement We acknowledge the members of the LAMP group (http://www. fisica.unipa.it/amorphous/) for support with interesting discussions. References
c
Fig. 5. Concentrations of Ge(1) (panel a), Ge(2) (panel b) and E'Ge (panel c) recorded at the different doses in FGeD1 ( ), FGeD2 (●) and FGeD3 (□).
irradiation. In fact, previous investigations, using larger variations of the drawing parameters, have shown relations between the drawing induced defects and the drawing parameters [7,31–34].
[1] G. Pacchioni, L. Skuja, D.L. Griscom, Defects in SiO2 and Related Dielectrics: Science and Technology, Kluwer Academic, Dordrecht, 2000. [2] E.J. Friebele, C.G. Askins, M.E. Gingerich, Appl. Opt. 23 (1984) 4202. [3] S. Girard, J.-P. Meunier, Y. Ouerdane, A. Boukenter, B. Vincent, A. Boudrioua, Appl. Phys. Lett. 84 (2004) 4215. [4] E.J. Friebele, G.H. Sigel Jr., D.L. Griscom, Appl. Phys. Lett. 78 (1976) 516. [5] G. Origlio, M. Cannas, S. Girard, R. Boscaino, A. Boukenter, Y. Ouerdane, Opt. Lett. 15 (2009) 2282. [6] E.J. Friebele, D.L. Griscom, Mater. Res. Soc. Symp. Proc. 61 (1986) 319. [7] NeustruevV.B. , J. Phys. Condens. Matter 6 (1994) 6901. [8] M. Fujimaki, T. Kasahara, S. Shimoto, N. Miyazaki, S. Tokuhiro, K.S. Seol, Y. Ohki, Phys. Rev. B 60 (1999) 4682. [9] L. Dong, J.L. Archambault, L. Reekie, P.St. Russell, D.N. Payne, Appl. Opt. 34 (1995) 3436. [10] S. Girard, Y. Ouerdane, G. Origlio, C. Marcandella, A. Boukenter, N. Richard, J. Baggio, P. Paillet, M. Cannas, J. Bisutti, J.-P. Meunier, R. Boscaino, IEEE Trans. Nucl. Sci. 55 (2008) 3473. [11] K. Awazu, H. Kawazoe, M. Yamane, J. Appl. Phys. 68 (1990) 2713. [12] M. Essid, J. Albert, J.L. Brebner, K. Awazu, J. Non-Cryst. Solids 246 (1999) 39. [13] L.N. Skuja, J. Non-Cryst. Solids 149 (1992) 77. [14] M. Fujimaki, T. Katoh, T. Kasahara, N. Miyazaki, Y. Ohki, J. Phys. Condens. Matter 11 (1999) 2589. [15] H. Shigemura, Y. Kawamoto, J. Nishii, M. Takahashj, J. Appl. Phys. 85 (1999) 3413. [16] J. Nishii, K. Kintaka, H. Hosono, H. Kawazoe, M. Kato, K. Muta, Phys. Rev. B 60 (1999) 7166. [17] K.D. Simmons, S. LaRochelle, V. Mizrahi, G.I. Stegeman, D.L. Griscom, Opt. Lett. 16 (1991) 141. [18] D.P. Hand, P.St. Russell, Opt. Lett. 15 (1990) 102. [19] T.E. Tsai, M.A. Saifi, E.J. Friebele, D.L. Griscom, U. Osterberg, Opt. Lett. 14 (1989) 1023. [20] M. Gallagher, U. Osterberg, J. Appl. Phys. 74 (1993) 2771. [21] Y. Watanabe, H. Kawazoe, K. Shibuya, K. Muta, Jpn. J. Appl. Phys. 25 (1986) 425. [22] L. Skuja, J. Non-Cryst. Solids 179 (1994) 51. [23] R.A. Weeks, J. Appl. Phys. 27 (1956) 1376. [24] S. Agnello, R. Boscaino, M. Cannas, F.M. Gelardi, Phys. Rev. B 64 (2001) 174423.
1970
A. Alessi et al. / Journal of Non-Crystalline Solids 357 (2011) 1966–1970
[25] For all the fibers the defects have been considered to be located in the higher doped region and in half of the 4.5% doped zone. The difference in the positioning the bulk reference sample and the fibers inside the microwave cavity is one of the most important error sources affecting the absolute concentrations in the fiber. [26] S. Girard, Y. Ouerdane, B. Vincent, J. Baggio, K. Medjahdi, J. Bisutti, B. Brichard, A. Boukenter, A. Boudrioua, J.-P. Meunier, IEEE Trans. Nucl. Sci. 54 (2007) 1136. [27] The identification of the E'Ge defects requires the decomposition of the experimental EPR spectra. These decompositions have been done using the Ge (1), Ge(2) and E'Ge reference signals isolated in a sample doped with 12% of Ge. Variations in the reference signals due to the different doping levels can affect the concentrations within the 20%, however the comparison between the defect concentrations induced in the three fibers is not affected since we used the same reference signals for the decomposition of all the spectra.
[28] S. Agnello, A. Alessi, F.M. Gelardi, R. Boscaino, A. Parlato, S. Grandi, A. Magistris, Eur. Phys. J. B 61 (2008) 25. [29] L. Vaccaro, M. Cannas, B. Boizot, ParlatoA. , J. Non-Cryst. Solids 353 (2007) 586. [30] F.L. Galeener, D.B. Kerwin, A.J. Miller, J.C. Mikkelsen Jr., Phys. Rev. B 47 (1993) 7760. [31] H. Hanafusa, Y. Hibino, F. Yamamoto, J. Appl. Phys. 58 (1985) 1356. [32] Y. Hibino, H. Hanafusa, J. Appl. Phys. 60 (1986) 1797. [33] H. Hanafusa, Y. Hibino, YamamotoF. , Phys. Rev. B 35 (1987) 7646. [34] Y. Hibino, H. Hanafusa, J. Non-Cryst. Solids 95 (1987) 343. [35] J. Hong, FTIR investigation of amorphous silica fibers and nano-size particles, ProQuest Dissertations and Theses; Thesis (Ph.D.) (Rensselaer Polytechnic Institute, Troy, 2003).
Contents lists available at ScienceDirect
Journal of Non-Crystalline Solids j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j n o n c r y s o l
X-ray irradiation effects on a multistep Ge-doped silica fiber produced using different drawing conditions A. Alessi a,⁎, S. Girard b, C. Marcandella b, S. Agnello c, M. Cannas c, A. Boukenter a, Y. Ouerdane a a b c
Laboratoire H. Curien, UMR CNRS 5516, Université Jean Monnet, 18 rue du Pr.Benoît Lauras 42000, Saint-Etienne, France CEA, DAM, DIF, F-91297 Arpajon, France Dipartimento di Scienze Fisiche ed Astronomiche, Università di Palermo, I-90123 Palermo, Italy
a r t i c l e
i n f o
Article history: Received 29 June 2010 Received in revised form 13 October 2010 Available online 14 December 2010 Keywords: Optical fibers; Drawing effects; Radiation effects; Point defects
a b s t r a c t We report an experimental study based on confocal microscopy luminescence (CML) and electron paramagnetic resonance (EPR) measurements to investigate the effects of the X-ray (from 50 krad to 200 Mrad) on three specific multistep Ge doped fibers obtained from the same preform by changing some of the drawing conditions (tension and speed). CML data show that, both before and after the irradiation, Germanium Lone Pair Center (GLPC) concentrations are similarly distributed along the diameters of the three fibers and they are partially reduced by irradiation. The irradiation induces also the Non Bridging Oxygen Hole Center (NBOHC) investigated by CML and other paramagnetic defects as the Ge(1), Ge(2), E'Ge and E'Si investigated by EPR. We do not observe significant differences in the induced concentrations of these types of defects in the three fibers. All the results suggest that within the range of investigated drawing parameters (usual range for specialty optical fibers), the fiber radiation sensitivity is unaffected by these variations. Consequently, these results show that the drawing parameters (drawing tension and speed) cannot easily be adjusted to improve the radiation hardness of germanosilicate optical fibers. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Silica based optical fibers are used in several technological application fields, telecommunication [1], radiation environments [2], sensors and other optical device production [1]. The fiber design and elaboration processes (dopant types, concentration, drawing conditions…) determine their optical features [3]. It is known that the Germanium doping increases the refractive index of the silica and for this reason it is commonly used to realize the core of the fibers [1]. As regards the production procedures, previous studies have shown that the drawing process can induce the generation of point defects that can act as precursor sites for the generation of optically-active defects under irradiation [4,5]. Since for many fiber-based applications in radiative environments, the radiation response of Ge-doped silica is a crucial aspect, so an important research field is devoted to this domain. Different investigations have shown that the radiation induces new optical absorption bands [6–9], thus increasing the attenuation of the fiber [10], and that it is due to the generation or the conversion of point defects in the glass network [1,6–8]. For this class of materials, the Germanium Lone Pair Center (GLPC) [11] has been widely studied, since a relation between the bleaching of its associated optical absorption (OA) band, peaked at 5.1 eV [1], and
⁎ Corresponding author. E-mail address: [email protected] (A. Alessi). 0022-3093/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jnoncrysol.2010.10.038
the photosensitivity of the Ge-doped silica glass [1,12] has been observed. The GLPC is constituted by a Ge atom bonded with two oxygen atoms by two single bonds, and with an electron lone pair (NGe••, N stands for the two single bonds, whereas • represents an electron) [1,11,13]. The GPLC is also responsible for two photoluminescence (PL) bands which peaked at ~4.3 eV and at ~3.2 eV [1,13]. According to Ref. [13] the 5.1 eV OA band is due to the S0 → S1 transition (S0 being the singlet electronic ground state and S1 the first singlet excited electronic state), the 4.3 eV PL band is due to the inverse transition and the 3.2 eV PL band is related to the T1 → S0 transition (being T1 the first triplet excited electronic state supplied through S1 by an inter-system crossing process [13]). For the data reported in the following it is also important to note that an OA band peaked at ~3.8 eV has been associated with the S0 → T1 transition [7]. In addition to the GLPC and its bleaching, other defects as the Ge(1), Ge(2) and the E'Ge and their generation processes were investigated [1,14–16]. The interest for these defects is due to the proposed relations among them and the induced OA bands [6–8] and to the debate of their relevance regarding transmission attenuation, photosensitivity and non linear effects [1,9,10,17–20]. All these defects are paramagnetic; in particular, the Ge(1) is made up of a fourfold coordinated Ge atom trapping an electron (NGe•b) [6], the E'Ge is formed by a threefold coordinated Ge atom with an unpaired electron ( Ge•) [21]. At variance, for the Ge(2) two structural models have been proposed. In the first one they have a structure similar to the Ge(1), differing for the presence of a second Ge atom as second
A. Alessi et al. / Journal of Non-Crystalline Solids 357 (2011) 1966–1970
1967
neighbor [6]. In the other one the Ge(2) is considered as a ionized GLPC (NGe•) [8]. In this work, we study the influence of X-ray irradiations on a set of specific germanosilicate fibers obtained from the same preform changing drawing conditions. To thoroughly investigate the effects of the drawing conditions on the radiation response of these fibers we also studied the radiation induced amounts of the Non Bridging Oxygen Hole Center (NBOHC Si–O•) [1,22] and of the E'Si ( Si•) [1,23].
We also recorded, at room temperature, electron paramagnetic resonance (EPR) spectra using a Bruker EMX-Micro Bay spectrometer working at 9.8 GHz. The defect concentrations were estimated using the EPR signal recorded in a bulk sample with a known concentration of E'Si [24]. The concentrations are estimated with an absolute accuracy of 50% and a relative accuracy of 20% [25].
2. Experimental
In Fig. 2a we report the emission amplitude at 3.2 eV along a fiber diameter recorded in all three investigated samples before irradiation and after a dose of ~100 Mrad. The data indicate that the GLPC contents are similar in the three samples and that their concentration reaches a maximum at the center of the higher doped region. It is important to note that even if the GLPC/Ge ratio in the 9 wt.% doped zone is not constant, it does not depend on the investigated fiber. Such an enhanced generation of defects was previously noticed in the core centers of irradiated samples of pure silica-core step-index fibers [26]. As regards the irradiation effects reported in Fig. 2b, we note that the emission of the GLPC shows a decreasing tendency while the dose increases. In Fig. 2b the point at 0.1 krad represents the initial
We have studied three different types of fibers produced by iXFiber SAS starting from the same preform (made by the modified chemical vapor deposition process) making major changes in the drawing conditions; they will be called FGeD1, FGeD2 and FGeD3 in the text. The Ge doping level along the fiber diameters (~125 μm) was specifically designed to follow a two-step distribution. The first step is doped with ~4.5 wt.%, whereas the second with ~9 wt.% (see Fig. 1). The FGeD1 was obtained using a drawing speed of 70 m/min and a tension of 135 g, the FGeD2 was produced using a drawing speed of 40 m/min and a tension of 70 g, whereas for the FGeD3 they were 22 m/min and 33 g respectively. These drawing conditions have been chosen to cover the range of tension and speed usually applied for the drawing of specialty fibers like radiation-tolerant waveguides, rare-earth doped or polarisationmaintaining optical fibers. However, our experimental procedure does not cover the range of drawing parameters used for the making of Telecom-type fibers. The samples were X-ray (10 keV) irradiated at room temperature with the ARACOR machine at the French atomic energy center (CEA). We recorded confocal microscopy luminescence (CML) measurements to check and follow the evolution of GLPC and NBOHC distribution along the fiber cross section diameter before and after the irradiations. These measurements were acquired using an Aramis (Jobin-Yvon) spectrometer, with a He–Cd ion laser excitation line (energy 3.8 eV, power ~ 0.15 mW) and supplied with a CCD camera, microtranslation stages and a 40x objective. The employed experimental conditions lead to a spatial resolution of ~2 μm. The system adopts a back-reflected geometry. For all the measurements the excitation beam penetrates a few micrometers into the sample. Furthermore we note that using the CML technique it is possible to filter the luminescence from a selected sample region, in this way possible light guiding effects along the fiber diameter are strongly limited.
3. Results
a
b
Fig. 1. Ge doping profile of the three fiber types.
Fig. 2. a) Emission at 3.2 eV as a function of the distance from the fiber center: ( ) FGeD1, (●) FGeD2 and (□) FGeD3 before irradiation, ( ) FGeD1, (▬) FGeD2 and ( ) FGeD3 at ~100 Mrad; b) emission at 3.2 eV recorded at the center of the fibers as a function of the dose ( ) FGeD1, (●) FGeD2 and (□) FGeD3.
1968
A. Alessi et al. / Journal of Non-Crystalline Solids 357 (2011) 1966–1970
amplitude. Even if the error of the emission measurements does not allow us to study such dose dependency in depth, no difference appears in the three investigated fibers. Furthermore considering all the data we deduce that at the dose of 100 Mrad the GLPC concentration is ~30% lower in all the fibers. Finally, comparing the GLPC profiles before and after irradiation we remark that they are very similar. In addition to the 3.2 eV band from the dose of 2 Mrad in the spectra an emission band peaked at ~1.9 eV was also detected. This emission is attributed to the NBOHC [1,22] and in Fig. 3 we report, as a function of the distance from the fibers' core centers, its amplitude at the dose of 100 Mrad. The data show that the NBOHC content is higher in the Ge doped zone, with a maximum at the core centers (similar behavior is observed for lower doses). In the inset of Fig. 3 the 1.9 eV amplitude recorded at the center of the fibers is reported as a function of the dose, we note that no significant differences in the dose dependence are observed in the three fibers and that the NBOHC content increases with a sub-linear dependence on dose. Fig. 4 illustrates the EPR spectra recorded at the dose of 2 Mrad for all the fibers. The spectra are normalized regarding the measurement conditions and the sample weights. We note that the signal amplitude is similar in the three fibers, and we also observe that the line-shape appears independent from the sample. In fact, in all the spectra we can observe the presence of the peaks at ~ 346 mT and at ~ 348.5 mT related Ge(1) and Ge(2), respectively. The identification of the E'Ge is not as easy since their lower concentration and the strong overlapping of the signals [27]. It is important to underline that no detectable concentrations of Ge(1) and Ge(2) have been observed in the fibers before irradiation, whereas an E'Ge concentration of ~ 2 × 1016 defects/ cm3 was detected in all three types of the pristine fibers. The Ge(1) induced concentrations are reported in Fig. 5a as a function of the irradiation dose. We note that, in the three fibers, the Ge(1) defect concentration linearly increases with the dose up to ~ 200 krad and then it tends to reach a limit value of ~ 4 × 1017 defects/ cm3. Similar dose dependence is observed in Fig. 5b for the Ge(2) with a saturation value of ~5 × 1017 defects/cm3. On the contrary, the E'Ge concentration increases with the dose following a sub-linear law without showing a saturation tendency (see Fig. 5c). Even if the dose dependence of the E'Ge is different from those of the Ge(1) and Ge(2) it is important to note that no significant differences are observed in the behavior of the three different fibers. In Fig. 5c we do not report the concentration induced at the first two
Fig. 3. Emission at 1.9 eV as a function of the distance from the fiber center: ( ) FGeD1, (-●-) FGeD2 and (-□-) FGeD3 at the dose of ~100 Mrad; inset emission at 1.9 eV recorded at the center of the fibers as a function of the dose ( ) FGeD1, (●) FGeD2 and (□) FGeD3.
Fig. 4. EPR spectra at the dose of ~ 2 Mrad for the FGeD1 ( FGeD3 ( ).
), FGeD2 (▬) and
doses since the total measured E'Ge concentrations are too close to those of the pristine samples. We note that the here reported dose dependences agree with the previous investigations [28,29]. Ge(1) and Ge(2) concentrations are determined by a competition between generation and destruction mechanisms involving precursors. By contrast the E'Ge and NBOHC kinetics suggest generation mechanisms involving T–O–T sites (with T = Si/Ge) [30]. Finally, as regards the E'Si, we note that their concentration (under a straightforward model the defects are supposed uniformly distributed in the fibers) starts to be measurable from the dose of 2 Mrad, since the overlapping of the E'Si EPR signal with that of the Ge-related defects. As for the other defects, it is important to underline that no significant differences are observed in the three fibers. The maximum induced concentration is observed at the dose of 200 Mrad and it is ~5 × 1016 defects/cm3 and we bring to mind that the concentration depends on a dose with a law similar to that of the E'Ge. 4. Discussion In a previous study [5], we showed a noticeable difference in the GLPC concentrations between a Ge-doped preform and its corresponding optical fiber. In the case of the fiber, the GLPC generation seems to be enhanced in its core during the drawing process. The data of Fig. 2 show that for the investigated variations of the drawing conditions, the three tested fibers have very similar preirradiation GLPC concentrations (panel a). As a consequence, it appears not possible to reduce the number of pre-existing GLPC defects in the fibers by varying the drawing speed and tension inside the studied range. This is also true for the other investigated precursor or radiation-induced defects: Ge(1), Ge(2), E'Ge, NBOHC and E'Si. All are identically generated in the three fibers. The radiation sensitivity independence from the drawing conditions could seem in contrast to the data reported in Ref. [10]. In that work, we showed that the fiber was more sensitive (in terms of the paramagnetic Ge-related defect generation) at low doses (~200 krad) than its corresponding preform, whereas minor differences were observed at higher doses (~200 Mrad). By changing the drawing condition, a difference in the radiation sensitivity at low doses could be expected. The reported absence of such differences suggests that the variations in the drawing conditions used to obtain the investigated samples, even if quite important for the fiber production, are too small to induce significant changes in the characteristics before or after
A. Alessi et al. / Journal of Non-Crystalline Solids 357 (2011) 1966–1970
a
1969
Furthermore, even if specific studies of the structure and stress of the examined fibers are postponed in future investigations, we can tentatively suppose that the fictive temperature of the three fibers is similar, since in Ref. [35] it was shown that significant variations of this quantity require greater changes in the drawing parameters. From the practical point of view our data indicate that larger changes of drawing tension and speed are required to modify the generation of defects and of the defect precursors to improve the radiation hardness, it is also important to consider the relevance of other drawing parameters. For these reasons the drawing temperature, the atmosphere applied during the drawing, the chemical composition of the fiber, the efficiency of pre-treatments like thermal annealing or hydrogen-loading have to be considered and investigated to design radiation-tolerant germanosilicate glasses. Finally, considering our data and those reported in Ref. [35] we suggest that between the changes in the drawing tension and variations of the drawing speed, these latter appear the most promising for the control improvement of the fiber radiation sensitivity. 5. Conclusions
b
We have reported an experimental investigation regarding three fibers produced from the same preform and differing from the drawing conditions. The study has been devoted to the defects generation–evolution under X-ray irradiation. We have investigated the signals of GLPC, NBOHC, Ge(1), Ge(2), E'Ge and E'Si; from this data the close similarity in the initial values of the various defect concentrations and analogous radiation responses of the fibers have been highlighted. Based on these findings, we suggest that the explored changes in the drawing parameters, usually employed for specialty fibers, do not affect the fiber features. Acknowledgement We acknowledge the members of the LAMP group (http://www. fisica.unipa.it/amorphous/) for support with interesting discussions. References
c
Fig. 5. Concentrations of Ge(1) (panel a), Ge(2) (panel b) and E'Ge (panel c) recorded at the different doses in FGeD1 ( ), FGeD2 (●) and FGeD3 (□).
irradiation. In fact, previous investigations, using larger variations of the drawing parameters, have shown relations between the drawing induced defects and the drawing parameters [7,31–34].
[1] G. Pacchioni, L. Skuja, D.L. Griscom, Defects in SiO2 and Related Dielectrics: Science and Technology, Kluwer Academic, Dordrecht, 2000. [2] E.J. Friebele, C.G. Askins, M.E. Gingerich, Appl. Opt. 23 (1984) 4202. [3] S. Girard, J.-P. Meunier, Y. Ouerdane, A. Boukenter, B. Vincent, A. Boudrioua, Appl. Phys. Lett. 84 (2004) 4215. [4] E.J. Friebele, G.H. Sigel Jr., D.L. Griscom, Appl. Phys. Lett. 78 (1976) 516. [5] G. Origlio, M. Cannas, S. Girard, R. Boscaino, A. Boukenter, Y. Ouerdane, Opt. Lett. 15 (2009) 2282. [6] E.J. Friebele, D.L. Griscom, Mater. Res. Soc. Symp. Proc. 61 (1986) 319. [7] NeustruevV.B. , J. Phys. Condens. Matter 6 (1994) 6901. [8] M. Fujimaki, T. Kasahara, S. Shimoto, N. Miyazaki, S. Tokuhiro, K.S. Seol, Y. Ohki, Phys. Rev. B 60 (1999) 4682. [9] L. Dong, J.L. Archambault, L. Reekie, P.St. Russell, D.N. Payne, Appl. Opt. 34 (1995) 3436. [10] S. Girard, Y. Ouerdane, G. Origlio, C. Marcandella, A. Boukenter, N. Richard, J. Baggio, P. Paillet, M. Cannas, J. Bisutti, J.-P. Meunier, R. Boscaino, IEEE Trans. Nucl. Sci. 55 (2008) 3473. [11] K. Awazu, H. Kawazoe, M. Yamane, J. Appl. Phys. 68 (1990) 2713. [12] M. Essid, J. Albert, J.L. Brebner, K. Awazu, J. Non-Cryst. Solids 246 (1999) 39. [13] L.N. Skuja, J. Non-Cryst. Solids 149 (1992) 77. [14] M. Fujimaki, T. Katoh, T. Kasahara, N. Miyazaki, Y. Ohki, J. Phys. Condens. Matter 11 (1999) 2589. [15] H. Shigemura, Y. Kawamoto, J. Nishii, M. Takahashj, J. Appl. Phys. 85 (1999) 3413. [16] J. Nishii, K. Kintaka, H. Hosono, H. Kawazoe, M. Kato, K. Muta, Phys. Rev. B 60 (1999) 7166. [17] K.D. Simmons, S. LaRochelle, V. Mizrahi, G.I. Stegeman, D.L. Griscom, Opt. Lett. 16 (1991) 141. [18] D.P. Hand, P.St. Russell, Opt. Lett. 15 (1990) 102. [19] T.E. Tsai, M.A. Saifi, E.J. Friebele, D.L. Griscom, U. Osterberg, Opt. Lett. 14 (1989) 1023. [20] M. Gallagher, U. Osterberg, J. Appl. Phys. 74 (1993) 2771. [21] Y. Watanabe, H. Kawazoe, K. Shibuya, K. Muta, Jpn. J. Appl. Phys. 25 (1986) 425. [22] L. Skuja, J. Non-Cryst. Solids 179 (1994) 51. [23] R.A. Weeks, J. Appl. Phys. 27 (1956) 1376. [24] S. Agnello, R. Boscaino, M. Cannas, F.M. Gelardi, Phys. Rev. B 64 (2001) 174423.
1970
A. Alessi et al. / Journal of Non-Crystalline Solids 357 (2011) 1966–1970
[25] For all the fibers the defects have been considered to be located in the higher doped region and in half of the 4.5% doped zone. The difference in the positioning the bulk reference sample and the fibers inside the microwave cavity is one of the most important error sources affecting the absolute concentrations in the fiber. [26] S. Girard, Y. Ouerdane, B. Vincent, J. Baggio, K. Medjahdi, J. Bisutti, B. Brichard, A. Boukenter, A. Boudrioua, J.-P. Meunier, IEEE Trans. Nucl. Sci. 54 (2007) 1136. [27] The identification of the E'Ge defects requires the decomposition of the experimental EPR spectra. These decompositions have been done using the Ge (1), Ge(2) and E'Ge reference signals isolated in a sample doped with 12% of Ge. Variations in the reference signals due to the different doping levels can affect the concentrations within the 20%, however the comparison between the defect concentrations induced in the three fibers is not affected since we used the same reference signals for the decomposition of all the spectra.
[28] S. Agnello, A. Alessi, F.M. Gelardi, R. Boscaino, A. Parlato, S. Grandi, A. Magistris, Eur. Phys. J. B 61 (2008) 25. [29] L. Vaccaro, M. Cannas, B. Boizot, ParlatoA. , J. Non-Cryst. Solids 353 (2007) 586. [30] F.L. Galeener, D.B. Kerwin, A.J. Miller, J.C. Mikkelsen Jr., Phys. Rev. B 47 (1993) 7760. [31] H. Hanafusa, Y. Hibino, F. Yamamoto, J. Appl. Phys. 58 (1985) 1356. [32] Y. Hibino, H. Hanafusa, J. Appl. Phys. 60 (1986) 1797. [33] H. Hanafusa, Y. Hibino, YamamotoF. , Phys. Rev. B 35 (1987) 7646. [34] Y. Hibino, H. Hanafusa, J. Non-Cryst. Solids 95 (1987) 343. [35] J. Hong, FTIR investigation of amorphous silica fibers and nano-size particles, ProQuest Dissertations and Theses; Thesis (Ph.D.) (Rensselaer Polytechnic Institute, Troy, 2003).