We report an experimental study of the doping and drawing effects on the Raman activities of phosphorus (P)-doped silica-based optical fiber and its related preform. Our data reveal a high sensitivity level in the full width at half maximum value of the 1330 cm−1 (O = P) Raman band to the P-doping level. Its increase with the P doping level does not clash with an increase in the disorder of the O = P surrendering matrix. In addition, we observe that in the central core region of the sample (higher doping level), the drawing process decreases the relative band amplitude. We tentatively suggest that this phenomenon is due to the change in the first derivate of the bond polarizability as a function of the normal vibration coordinates.
© 2012 OSA
Optical fibers are more and more frequently used in various technological fields [1–3]. Phosphorous (P)-doping of the silica-based core or cladding is often used for the fabrication of optical fibers. P addition increases the refractive index of the glass  and at the same time allows lowering of the glass melting temperatures . Furthermore, phosphate glasses were studied to meet the increasing demand of UV-transmitting materials for applications in microlithography and laser systems . Moreover, P-doped silica is used as a host matrix for the incorporation of rare-earth elements like Erbium of Ytterbium for designing efficient optical amplifiers and laser sources . Single-mode phosphosilicate fibers are also used to develop stimulated Raman fiber lasers . The interest in studying the fiber and the associated preform is related to the understanding of the drawing effects to adjust the preform features according to the required properties of the final fiber. These types of studies aim to identify the precursor sites of the radiation induced defects responsible for the photodarkening and to control the structural parameter affecting the fiber’s performance. Recently , our group observed that, under irradiation, some of the P-related point defects are induced in lower concentration in the fiber than in its corresponding preform. Currently, the structure of the phosphosilicate glasses is described as a three dimensional random mixture of SiO4 and O = PO3 tetrahedra . In such a structure one of the vertexes of the P tetrahedra is constituted by a non-bridging oxygen atom forming a double bond with the P atom. We note that a Raman band located at about 1330 cm−1 was associated with this bond [9,10].
The present investigation highlights the drawing effects on P-doped silica-based glasses structure to provide evidence for the changes induced by this process. In this study, we recorded confocal micro-Raman spectra on both fiber and its original preform. Thanks to the particular structure of the prototype investigated sample, we also investigated the effects of the P-doping levels on the spectroscopic features of the O = P Raman band.
The investigated fiber, named FP, was manufactured by iXFiber S.A.S starting with a preform (named PP) produced by the modified chemical vapor deposition process (MCVD). The fiber was produced using drawing tension and speed of 60 g and 40m/min respectively.
The doping profiles of the samples are illustrated in Fig. 1 . They consist of five concentric cylindrical layers with different P-doping contents ranging from virtually absent in the pure silica cladding to a maximum of ~7 weight % in the central part of the core. The fiber doping profile reported in Fig. 1 was obtained by electron microprobe analysis, whereas the doping profile of the preform is designed during the layers deposition phase.
We recorded confocal micro-Raman spectra along the sample diameters using an Aramis (Jobin-Yvon) spectrometer. This setup is supplied by a CCD camera, a He–Cd ion laser (energy 3.8 eV and power ~0.15 mW), a step motor and by 40x objectives. The spectra were acquired, at room temperature, with experimental conditions ensuring a spatial resolution of ~5 μm.
3.1 Raman spectra of the preform
In Fig. 2 , we report the Raman spectra recorded in different parts of the PP sample. To compare these measurements, all spectra are normalized to the Raman amplitude detected at about 440 cm−1, as the line shape of this band appears to be less affected by the change of the doping level.We observe the signatures of the silica-related Raman bands [11,12] listed in Table 1 . Furthermore, in zones containing phosphorus, we also observe the signatures of bands associated with P-related linkages [9,11]. The most intense band is detected at ~1330 cm−1 and corresponds to the O = P bond vibration [9,11]. Other clear Raman activities can be seen at ~720 cm−1 and at ~1150 cm−1. Several other bands are known to be related to phosphorus at around 540, 800 and 1020 cm−1 but their presence is more difficult to detect as they have lower relative amplitudes and stronger overlaps with the silica-related bands.
Our measurements reveal that, using the relative amplitude of the 1330 cm−1, it is possible to reconstruct the doping profile along the preform diameter. This is illustrated in Fig. 3a , which compares the normalized P-doping profile and the normalized profile of the 1330 cm−1 amplitude along the preform diameter. Even if it is possible to re-construct the P doping profile via the amplitude of the O = P related Raman band, it appears that its shape evolves in the different doped zones of the preform. This is illustrated in Fig. 3b. By normalizing the spectra with respect to their 1330 cm−1 band, a broadening and a slight frequency shift (about 3 cm−1) of the band is shown.
3.2 Fiber and preform comparison
Figure 4a compares the Raman spectra recorded in the central cores of the fiber and of the preform.
The data clearly indicate a difference in the amplitudes of the 1330 cm−1 Raman band, with a lower amplitude in the fiber. Such a difference appears to be significant only in the highly-doped core region and it is not observed in other zones as proved by the comparison of the spectra recorded in the 4.8 wt% P doped regions (see Fig. 4b). We note that, in the fiber spectra, a slight increase in the D2 Raman band is also visible. However, a quantitative analysis of this effect is limited by the presence of some P related Raman band.
Before commenting the difference of the amplitude of the 1330 cm−1, we highlight the variation of the full width at half maximum (FHWM) of this band as a function of the P content. The FWHM increases with the P-doping level in similar ways both for the fiber and the preform. This is illustrated in Fig. 5a . We also note that, as reported in Fig. 3b, such an increase of the FWHM takes place both on right and left side of the peak. This consideration is relevant since in  the authors have shown the existence of two types of O = P3 structure named single and double phosphorous centers. These structures have two Raman bands peaking in this frequency range but with different peak positions. So, the broadening at both sides of the peak indicates that the changes of the FWHM are not only related to a relative change of the center populations. We then suggest that by increasing the P content the distribution of the structural parameter affecting this band becomes bigger. For example, the FWHM increase could be related to a larger degree of disorder in the matrix because of an increasing number of atoms different from O and Si atoms. The lack of evident effects on the FWHM of other bands related to Si-O-Si or Si-O-P can be explained by the fact that their distributions are already large, as shown by the Raman signals, whereas the O = P band is narrow as expected since the O atom is a non-bridging one. We also note that in ref  where the authors have studied phosphosilicate glass films differing by their P content, an increase in the FWHM with the doping was also noted.The main effect of the drawing is the change in the relative amplitude of the 1330 cm−1 band. As a first consequence of this drawing related change, in the fiber core region, the P doping profile is not exactly re-constructed by the distribution of the O = P band amplitude. This change cannot be attributed to significant variation of P content in the central core since the electron microprobe analysis clearly indicates that the P content remains unchanged.
The two other possibilities that could explain this change are a decrease in the O = P bond concentration and the modification of its scattering properties. To understand if O = P bonds are destroyed, we have to consider the investigation presented in  where it was suggested that the transformation O = P(O-Si)3→P(O-Si)5 has an energy barrier of ~0.3 eV, whereas the inverse one has an energy barrier of ~0.7 eV. This transformation should be accompanied by the decrease in the Raman signal at about 1150, 680 and 530 cm−1 and by the increase of the Raman signal at about 1060 and 890 cm−1. Since our measurement provides no evidence of any negative component in the difference between the spectra of the fiber and of the preform (see Fig. 5b) we can exclude this possibility. Furthermore, it is also assumed that O = P are not destroyed and transformed in the paramagnetic ●O-P (r-POHC  or l-POCH ). This assumption is supported by the following considerations. In the core, about 3 1021 P atoms/cm3 are present and most of them are involved in O = P linkages , so that a 20% of decrease in the 1330 cm−1 amplitude should induce the formation of several 1020 defects/cm3 whereas in  with electron paramagnetic resonance (EPR) measurements we did not observe the presence of detectable signal before any irradiation. Since the stable POHC [13,15] (r-POHC) have the [(O-)2P( = O)2] as more probable precursors [13,15], whereas the room temperature not stable POHC [13,15] (l-POHC) have the [(O-)3P = O] as more probable precursors [13,15], one could speculate that the decrease in the Raman band is due to the generation of the second types of POHC. To exclude this hypothesis, we also have to consider the following data previously reported in ref . The l-POHC defects were observed at room temperature in the EPR spectra of the preform. In ref , based on their calculation, the authors do not rule out that a l-POHC (metastable) can become a r-POHC (stable) after an internal conversion involving the braking of a near Si-O bond.
As a consequence, we suggest that the main reason for the lower amplitude value of the 1330 cm−1 observed in the fiber is a modification of the O = P bond and/or of its surrendering environment, which modifies the dα/dQ, being α and Q the polarizability tensor and the vibration normal coordinate, respectively. Similar considerations can be applied to the other differences in the Raman activity of central core. In fact, even if it is less evident, by subtracting the Raman spectrum of the fiber central core to that of the preform, it is possible to note other differences at about 700, 800 and 1150 cm−1, see Fig. 5b. Then, it is important to remember that a band peaked at ~700 cm−1 was attributed to O = P-O [10,16], one at ~800 cm−1 P-O-P [10,16], and one at about ~1150 cm−1 was related to P-O-P in  and to P-O-Si in , so that the variations reported in Fig. 5b suggest variations not only of the O = P bond but also in its surrounding matrix. Moreover, since the P-O-P linkages appear mainly involved in the attribution of these bands, the changes may be related to slight differences in the structure of the double-bonded phosphorous center. Anyway, further experimental and computational investigations as well as a quantitative spectroscopic analysis of the Raman spectra are necessary to identify these modifications and to characterize them.
Another interesting result is that the decrease the 1330 cm−1 band amplitude is observed only in the higher doped region. This may suggest that a higher sensitivity of the central core to the drawing process. This may be related to greater variations of the mechanical properties of the network  and/or to the lower average distance among the P atoms due to the higher P content. More detail should arise by changing the value of doping in the central core.
We have investigated the Raman activity of a 4- steps P doped fiber and its original preform. Our data indicate that the FWHM of the Raman band of the O = P bond increases with the P content. This result may suggest larger distributions of the structural parameters of the O = P bond or of its surrounding matrix which affect its Raman activity. Furthermore, we observe significant change in the Raman spectra after the drawing only in the highly P-doped central zone. Such variation concerns the relative amplitude of the O = P component. After excluding various possible reasons, we tentatively suggest that the first derivate of the polarizability as a function of the normal vibration coordinates is changed because of structure modifications. Further investigations will be performed to confirm this hypothesis as well as to clarify the fact that this effect is observed only in the central part of the core. Anyway, our data provide evidence for the drawing effects on the P doped silica structure, which have to be taken into account for fiber production.
We acknowledge the members of the LAMP group (http://www.fisica.unipa.it/amorphous/) for support with interesting discussions.
References and links
1. G. Pacchioni, L. Skuja, and D. L. Griscom, eds., Defects in SiO2 and Related Dielectrics: Science and Technology (Kluwer Academic, Dordrecht, 2000).
2. R. Kitamura, L. Pilon, and M. Jonasz, “Optical constants of silica glass from extreme ultraviolet to far infrared at near room temperature,” Appl. Opt. 46(33), 8118–8133 (2007). [CrossRef] [PubMed]
3. A. J. Ikushima, T. Fujiwara, and K. Saito, “Silica glass: A material for photonics,” J. Appl. Phys. 88(3), 1201–1213 (2000). [CrossRef]
4. R. A. B. Devine, J. P. Duraud, and E. Dooryhée, eds., Structure and Imperfections in Amorphous and Crystalline Silicon Dioxide (John Wiley & Sons, LTD, New York, 2000).
5. D. Ehrt, P. Ebeling, and U. Natura, “UV Transmission and radiation-induced defects in phosphate and fluoride-phosphate glasses,” J. Non-Cryst. Solids 263-264, 240–250 (2000). [CrossRef]
6. P. L. Kelly, I. Kaminov, and G. Agrawal, eds., Erbium-Doped Fiber Amplifiers, Fundamental and Technology (Academic, London, 1999).
7. E. M. Dianov, M. V. Grekov, I. A. Bufetov, S. A. Vasiliev, O. I. Medvedkov, V. G. Plotnichenko, V. V. Koltashev, A. V. Belov, M. M. Bubnov, S. L. Semjonov, and A. M. Prokhorov, “CW high power 1.24 μm and 1.48 μm Raman lasers based on low loss phosphosilicate fibre,” Electron. Lett. 33(18), 1542–1544 (1997). [CrossRef]
8. G. Origlio, F. Messina, S. Girard, M. Cannas, A. Boukenter, and Y. Ouerdane, “Spectroscopic studies of the origin of radiation-induced degradation in phosphorus-doped optical fibers and preforms,” J. Appl. Phys. 108(12), 123103 (2010). [CrossRef]
9. V. G. Plotnichenko, V. O. Sokolov, V. V. Koltashev, and E. M. Dianov, “On the structure of phosphosilicate glasses,” J. Non-Cryst. Solids 306(3), 209–226 (2002). [CrossRef]
10. N. Shibata, M. Horigudhi, and T. Edahiro, “Raman spectra of binary high-silica glasses and fibers containing GeO2, P2O5 and B2O3,” J. Non-Cryst. Solids 45(1), 115–126 (1981). [CrossRef]
11. F. L. Galeener and A. E. Geissberger, “Vibrational dynamics in 30Si-substituted vitreous SiO2,” Phys. Rev. B 27(10), 6199–6204 (1983). [CrossRef]
12. F. L. Galeener and G. Lucovsky, “Longitudinal optical vibrations in glasses: GeO2 and SiO2,” Phys. Rev. Lett. 37(22), 1474–1478 (1976). [CrossRef]
13. M. Fanciulli, E. Bonera, S. Nokhrin, and G. Pacchioni, “Phosphorous–oxygen hole centers in phosphosilicate glass films,” Phys. Rev. B 74(13), 134102 (2006). [CrossRef]
14. V. G. Plotnichenko, V. O. Sokolov, V. V. Koltashev, V. B. Sulimov, and E. M. Dianov, “UV-irradiation-induced structural transformation in phosphosilicate glass fiber,” Opt. Lett. 23(18), 1447–1449 (1998). [CrossRef] [PubMed]
15. D. L. Griscom, E. J. Friebele, K. J. Long, and J. W. Fleming, “Fundamental defect centers in glass: Electron spin resonance and optical absorption studies of irradiated phosphorus-doped silica glass and optical fibers,” J. Appl. Phys. 54(7), 3743–3762 (1983). [CrossRef]
16. E. M. Dianov, V. V. Koltashev, P. G. Plotnichenko, V. O. Sokolov, and V. B. Sulimov, “UV irradiation-induced structural transformation in phosphosilicate glass,” J. Non-Cryst. Solids 249(1), 29–40 (1999). [CrossRef]