Open Access
Add to BibSonomyAbstract
Glass ceramic fibers containing Ni2+ doped LiGa5O8 nanocrystals were fabricated by a melt-in-tube method and successive heat treatment. Fiber precursors were prepared by drawing at high temperature where fiber core glass was melted while fiber clad glass was softened. After heat treatment, LiGa5O8 nanocrystals were precipitated in the fiber core. Excited by 980 nm laser, efficient broadband near-infrared emission was observed in the glass ceramic fiber compared to that of precursor fiber. The melt-in-tube method can realize controllable crystallization and is suitable for fabrication of novel glass ceramic fibers. The Ni2+-doped glass ceramic fiber is promising for broadband optical amplification.
© 2015 Optical Society of America
1. Introduction
During the past few decades the telecommunication transmission window has been extended to the range from 1.2 to 1.7μm since great progress has been achieved in the OH elimination of silica fibers [1]. As a result, considerable effort has been devoted to the development of optical fiber amplifiers which can be used to produce optical gains in the whole telecommunication window. Ni2+ ion is an attractive activator for fiber amplifiers because it has a stable valence state and shows broad-band luminescence in near-infrared region when doped in crystal lattices [2–8 ]. However, the difficulty in fabricating single crystal fibers limits their practical application. Glass is an amorphous host and easy to process, but the crystal field is weak, resulting in weak or even no luminescence when Ni2+ ions doped in it. Glass ceramic (GC) is a two-phase composite in which crystals are homogeneously distributed in the glass matrix. It exhibits high transmittance as well as strong crystal field if active ions are doped in crystalline phase. These features make Ni2+ doped GC a suitable matrix for fiber amplifiers to achieve optical gains in the whole telecommunication window. In 2002, Samson et al. reported the investigation of Ni2+ doped GC fiber. The fabrication technique was based on rod-in-tube method, the components of core glass and clad glass were similar, and the fiber was drawn when the glasses were softened [9]. Rod-in-tube is a common method for fabricating multi-component fiber, but it is not suitable for all GC fibers since the drawing temperature is mostly higher than crystallization temperature of core glass, and the uncontrollable crystallization of fiber can be hardly avoided, resulting large optical transmission attenuation.
In order to improve the rod-in-tube process, we developed a novel method for the fabrication of Ni2+ doped GC fiber containing LiGa5O8 nanocrystals. In analog to rod-in-tube process, we call it melt-in-tube method. Electro-probe micro-analyzer (EPMA) images, Raman spectrum and high-resolution transmission electron microscopy (HRTEM) images were measured to characterize the element distribution and microstructures of the fibers. The enhanced broadband emission of GC fibers was verified by the photoluminescence (PL) spectra upon excitation of a 980 nm laser.
2. Experiments
The core glass for precursor fiber was a silicate multi-component glass with the compositions (mol %): 64SiO2-23Ga2O3-13Li2O-0.1NiO. A 200 g reagent grade stoichiometric mixture of SiO2, Ga2O3, Li2CO3 and NiO was mixed thoroughly in an agate mortar and melted in a Pt-Rh crucible at 1650 °C for 2 h. The glass were fabricated by pouring the melt into a preheated mould, the glass was annealed at 570 °C for 3 h to release inner stress. The glass block was cut into two parts, which were cold worked into cylindrical rods with diameter of 15.0 and 3.0 mm in lathe, respectively. The surfaces of the rods were polished and etched by acid in order to remove the contaminated surface layer. The thicker rod (diameter of 15.0 mm) was suspended in graphite furnace to draw fibers without cladding. The thinner rod (diameter of 3.0 mm) was inserted into a high-purity (99.999%) silica glass cylindrical tube with inner diameter of 3.1 mm and external diameter of 25.0 mm. The bottom of tube was sealed to form a preform. Then the preform was suspended in graphite furnace to draw optical fibers. Rising the temperature to about 1830 °C gradually, the core glass was present as a melt when the clad glass became softened. By quickly drawing (15 m/min), the precursor optical fibers were prepared. Finally, the precursor optical fibers were heat treated to obtain GC optical fiber.
Differential thermal analysis (DTA) and thermal expansion coefficient were measured at STA449C Jupiter (Netzsch, Bavaria, Germany) in argon atmosphere at a heating rate of 10 K/min in the range from 25 to 1000 °C. The amorphous state and crystalline phase in the glasses were identified by X-ray diffraction (XRD) on a D8 advance X-ray diffractometer (Bruker, Faellanden, Switzerland) with Cu/Kα (λ = 0.1541 nm) radiation. The XRD patterns of the samples were collected in the range of 10° <2θ <80°. For determining the element distribution of the fiber cross section, the precursor fibers were measured by an EPMA system (EPMA-1600, Shimadzu, Kyoto, Japan). Micro-Raman spectra was recorded by a Raman spectrometer (Renishaw inVia, London, UK) excited by a 532 nm laser. The morphology and size distribution of the nanocrystals in GCs were measured by HRTEM (Tecnai G2, FEI, Amsterdam, Netherlands). The PL spectra of fibers were measured by a spectrometer (zolix, Omni k3007, Beijing, China) equipped with AsGaIn photodetector (Hamamatsu R5108, Shizuoka-ken, Japan) and SR830 Stanford Research lock-in amplifier (SRS, SR830, Sunnyvale, CA). All the measurements were performed at room temperature.
3. Experimental results and discussion
Figures 1(a) and 1(b) show images of the glass rod (diameter of 15.0 mm) before and after drawing process near the softening temperature (1360 °C). Before drawing process, the glass rod is transparent in the whole length, no obvious crystal diffraction peak is observed in XRD pattern as shown in Fig. 1(c). With temperature rising, the rod turns to opaque and crystallization appears in the bottom, which located in the high-temperature region of furnace during the drawing process. The crystals can be identified to LiGa5O8 (ICDD PDF No. 012-0535) and LiGaSi2O6 (ICDD PDF No. 014-0505) by XRD pattern as shown in Fig. 1(c). The XRD pattern of GC (heat treated in 800 °C for 5 hours) is also shown in Fig. 1(c); the sharp peaks can be attributed to LiGa5O8 (ICDD PDF No. 012-0535) crystal. The diffraction peaks of the glass rod are narrower than that of GC, which results from the larger size of crystal in glass rod. These results can be explained by the DTA curve of glass presented in Fig. 1(d). The glass starts crystallization at about 720 °C, two main crystallization peaks located at 800 and 890 °C, which are much lower than the drawing temperature of fiber. Thus, the mixed and large crystals precipitated in the glass rod, the crystallization during drawing process is uncontrollable. Similarly, the drawing temperature of rod-in-tube method locates near the softening temperature of glass, which results in the increase of transmission loss caused by uncontrollable crystallization, even results in the quenching of fluorescence. Consequently, rod-in-tube method is not suitable for fabricating LiGa5O8 GC fiber.
Fig. 1 Images of glass rod before (a) and after (b) drawing process. (c) XRD patterns of glass rod before and after drawing process, GC (heat treated in 800 °C for 5 hours) and glass after secondary melting. (d) DTA curve of the precursor glass.
Then the crystallized glass rod was put in furnace and melted again at 1650 °C. By quenching, the glass turned to transparent again, the crystals were dissolved into amorphous state during the secondary melting process, which is confirmed by the XRD pattern as shown in Fig. 1(c). Based on this, we fabricated the precursor optical fiber by a melt-in-tube method, drawing fibers when the core glass was melted and the clad was softened, which is also a secondary melting drawing process.
The optical micrograph of the precursor fiber cross section is shown in Fig. 2(a) . There exist no vacuum bubbles and microcracks due to the thermal expansion coefficient mismatching of core (5.4 × 10−6/K) and clad glass (6.3 × 10−7/K in 30-700°C). The fiber core shows faint yellow color, and the boundary between fiber core and fiber clad is clear. The diameter of fiber core and clad is about 12.3 and 125.1 μm, respectively. In order to determine element distribution of the precursor fiber core and clad after high temperature drawing, EPMA measurement was performed on the cross section of the precursor fiber. It can be observed from Figs. 2(b) and 2(c) that the element of Ga distributes in the core and the relative concentration of Si shows an abrupt change due to the different concentration of Si in the core and clad. The distribution boundary of each element forms a circle and exhibits similar size as that of the precursor fiber core as shown in Fig. 2(a). These results indicate that the core-clad structure of fiber is preserved completely. In addition, Li and Ni elements are not detected by EPMA because of the light weight and low concentration, respectively.
Fig. 2 (a) Optical image of the precursor fiber in cross section, (b) and (c) the EPMA images of the precursor fiber in local cross section.
According to the DTA curve of glass, the fiber was heat treated in 800 °C for 10 hours to obtain GC fibers. Raman spectra of fiber and glass are shown in Fig. 3 . The focus diameter and the power of the probe laser were 1 μm and 25 mW, respectively. The Raman spectrum of precursor fiber core consists of several bands without any sharp crystal peak, which is a Raman spectrum typical of oxide glasses and consistent with that of precursor glass. The broad band located around 500 cm−1 is mainly due to O-Si-O rocking modes of SiO4 tetrahedra, and the Si-O stretching bands located around 800 cm−1 [10, 11 ]. This result indicates that the fiber core maintains amorphous state due to high cooling rate of the thin core glass during the drawing process. For the core of GC fiber, sharp crystal peaks are observed in the Raman spectrum, which is in good agreement with GC Raman spectrum. The observed Raman peak at 260 cm−1 is mainly due to the Li-O framework and Li-O stretching and O-Li-O bending modes [12, 13 ]. The Raman peak at 415 cm−1 belongs to deformation of GaI(OI)2 octahedra; the Raman peaks observed at 651 and 770 cm−1 are related to symmetric stretching bands of GaO4 tetrahedra [14, 15 ]. The sharp peaks are in good agreement with the Raman spectrum of LiGa5O8 crystals. These results indicate that LiGa5O8 crystals have been precipitated in the fiber by heat treatment at 800 °C for 10 hours. As a reference, the Raman spectrum of the GC fiber clad is also shown in Fig. 3, which presents as several bands corresponding to the characteristic Raman peaks of silica glass [16]. Therefore, the crystallization process is controllable and the crystallization of LiGa5O8 crystals can be well confined within the GC fiber core.
Fig. 3 (a) Raman spectra of (1) precursor fiber core, (2) GC fiber clad, (3) GC fiber core, (4) precursor glass and (5) GC heat treated in 800 °C for 5 hours. (b) Raman spectra at different positions of GC fiber. The inset of (b) is the image of GC fiber cross section.
In addition, the Raman spectra at different positions of GC fiber are shown in Fig. 3(b). The crystal Raman peaks can be observed in all the positions of fiber core (1, 2, 3 and 4). This result indicates that LiGa5O8 crystals distribute both in the center and surface of fiber core and the crystallization in fiber is not a surface crystallization.
As for GC fiber, high transmission is important as well as controllable crystallization. For reducing the transmission loss caused by light scattering of particles with different refractive index from glass matrix, the size of crystals in GC fiber must be much smaller than the emission wavelength. To observe the morphology and size distribution of crystals in the GC fiber, we ground the GC fiber samples into powders and measured the HRTEM images. As presented in Fig. 4(a) , the particles are dispersed in the glassy matrix with diameter ranging from 0.5 to 7.6 nm. The inset of Fig. 4(a) is the corresponding selected area electron diffraction (SAED) pattern. Owing to the different orientation of LiGa5O8 nanocrystals in the glass matrix, the GC fiber composites show a polycrystalline character. The HRTEM image is shown in Fig. 4(b). The crystal lattice fringes are clear, which are different from that of the amorphous glass matrix. The interval of the crystal lattice fringes d can be measured directly, and its value is about 0.207 nm, which corresponds to the (400) crystal facet of LiGa5O8. Tick et al. proposed that the scattering losses achieved to a minimum when the size of nanocrystal kept small (usually of the order of 10 nm) [17]. Therefore, the transmission loss in the GC fibers caused by light scattering of the nanoparticles is almost negligible.
Fig. 4 (a) TEM image of GC fiber, the inset is the corresponding SAED pattern. (b) HRTEM of GC fiber.
Figure 5 shows the PL spectra of the precursor fiber and GC fiber excited by a 980 nm laser. The length of the precursor fiber and the GC fiber are both about 20 cm. It is noticed that, an intense broad emission band centering at 1260 nm, which covers from 1100 to 1600 nm (the full width at half maximum (FWHM) is about 230 nm), is observed in the GC fiber. Such an emission can be attributed to the transition from the 3T2 (F) excited state to the 3A2 (F) ground state of octahedral Ni2+ [18]. Indicating Ni2+ ions have entered LiGa5O8 crystal lattices and substituted for Ga3+ ions. In contrast, the precursor fiber does not show any emission in near-infrared region. The enhanced emission is ascribed to strong crystal field and high transmittance of the GC fiber, it verifies that the GC fiber containing Ni2+ doped LiGa5O8 nanocrystals is a promising material for broadband optical amplifier, and the melt-in-tube method is highly suitable for fabricating novel functional GC fibers.
4. Conclusion
In this work, the GC fibers containing Ni2+ doped LiGa5O8 nanocrystals were prepared by the melt-in-tube method and successive heat treatment. The fiber core glass was melted while the clad was softened at drawing temperature. The precursor fibers were prepared without crystallization. After heat treatment, the GC fibers were obtained with Ni2+ doped LiGa5O8 nanocrystals precipitation in the fiber core, the size of nanocrystals is between 0.5 and 7.6 nm. Excited by 980 nm laser, enhanced broadband emission with FWHM of about 230 nm was observed in the GC fiber comparing to that of precursor fiber, which makes this GC fiber a promising material for broadband fiber amplifier. Comparing to rod-in-tube method, the melt-in-tube method can realize controllable crystallization, and may open a new gate towards the fabrication of novel GC fibers.
Acknowledgements
This work was financially supported by the National Natural Science Foundation of China (Grant Nos. 51102096, 51302087), Guangdong Natural Science Foundation (Grant No. S2011030001349), the China Postdoctoral Science Foundation (Grant No. 2015T80903), Fundamental Research Funds for the Central Universities (Grant No. 2013ZM0001). Shupei Zheng and Zaijin Fang contributed equally to this work.
References and links
1. S. Zhou, H. Dong, H. Zeng, B. Wu, B. Zhu, H. Yang, S. Xu, Z. Wan, and J. Qiu, “Broadband near-infrared emission from transparent Ni2+-doped silicate glass ceramics,” J. Appl. Phys. 102(6), 063106 (2007). [CrossRef]
2. T. Suzuki, G. S. Murugan, and Y. Ohishi, “Optical properties of transparent Li2O-Ga2O3-SiO2 glass- ceramics embedding Ni-doped nanocrystals,” Appl. Phys. Lett. 86(13), 131903 (2005). [CrossRef]
3. T. Suzuki, K. Horibuchi, and Y. Ohishi, “Structural and optical properties of ZnO-Al2O3-SiO2 system glass-ceramics containing Ni2+-doped nanocrystals,” J. Non-Cryst. Solids 351(27–29), 2304–2309 (2005). [CrossRef]
4. S. Zhou, G. Feng, B. Wu, S. Xu, and J. Qiu, “Transparent Ni2+-doped lithium-alumino-silicate glass- ceramics for broadband near-infrared light source,” J. Phys. D 40(8), 2472–2475 (2007). [CrossRef]
5. C. Jiang, “Ultra-broadband amplification properties of Ni2+-doped glass-ceramics amplifiers,” Opt. Express 17(8), 6759–6769 (2009). [CrossRef] [PubMed]
6. S. Zhou, H. Dong, G. Feng, B. Wu, H. Zeng, and J. Qiu, “Broadband optical amplification in silicate glass-ceramic containing β-Ga2O3:Ni2+ nanocrystals,” Opt. Express 15(9), 5477–5481 (2007). [CrossRef] [PubMed]
7. S. Zhou, N. Jiang, H. Dong, H. Zeng, J. Hao, and J. Qiu, “Size-induced crystal field parameter change and tunable infrared luminescence in Ni2+-doped high-gallium nanocrystals embedded glass ceramics,” Nanotechnology 19(1), 015702 (2008). [CrossRef] [PubMed]
8. B. Wu, S. Zhou, J. Ruan, Y. Qiao, D. Chen, C. Zhu, and J. Qiu, “Enhanced broadband near-infrared luminescence from transparent Yb3+/Ni2+ codoped silicate glass ceramics,” Opt. Express 16(3), 1879–1884 (2008). [CrossRef] [PubMed]
9. B. N. Samson, L. R. Pinckney, J. Wang, G. H. Beall, and N. F. Borrelli, “Nickel-doped nanocrystalline glass-ceramic fiber,” Opt. Lett. 27(15), 1309–1311 (2002). [CrossRef] [PubMed]
10. F. L. Galeener, “Band limits and the vibrational spectra of tetrahedral glasses,” Phys. Rev. B 19(8), 4292–4297 (1979). [CrossRef]
11. R. Ceccato, R. Dal Maschio, S. Gialanella, G. Mariotto, M. Montagna, F. Rossi, M. Ferrari, K. E. Lipinska-Kalita, and Y. Ohki, “Nucleation of Ga2O3 nanocrystals in the K2O-Ga2O3-SiO2 glass system,” J. Appl. Phys. 90(5), 2522 (2001). [CrossRef]
12. Y. Repelin, E. Husson, F. Bennani, and C. Proust, “Raman spectroscopy of lithium niobate and lithium tantalate. Force field calculations,” J. Phys. Chem. Solids 60(6), 819–825 (1999). [CrossRef]
13. P. W. Jaschin and K. B. R. Varma, “Pyroelectric and second harmonic responses from LiTaO3 nanocrystallites evolved in a Li2O-B2O3-Ta2O5 glass system,” CrystEngComm 17(11), 2327–2335 (2015). [CrossRef]
14. I. López, A. D. Utrilla, E. Nogales, B. Mendez, J. Piqueras, A. Peche, J. Ramirez-Castellanos, and J. M. Gonzalez-Calbet, “In-doped gallium oxide micro- and nanostructures: morphology, structure, and luminescence properties,” J. Phys. Chem. C 116(6), 3935–3943 (2012). [CrossRef]
15. Y. Wang, N. Li, P. Duan, X. Sun, B. Chu, and Q. He, “Properties and photocatalytic activity of β-Ga2O3 nanorods under simulated solar irradiation,” J. Nanomater. 10(5), 191793 (2015).
16. P. González, J. Serra, S. Liste, S. Chiussi, B. Leon, and M. Perez-Amor, “Raman spectroscopic study of bioactive silica based glasses,” J. Non-Cryst. Solids 320(1–3), 92–99 (2003). [CrossRef]
17. P. A. Tick, N. F. Borrelli, and I. M. Reaney, “The relationship between structure and transparency in glass-ceramic materials,” Opt. Mater. 15(1), 81–91 (2000). [CrossRef]
18. S. Zhou, G. Feng, B. Wu, N. Jiang, S. Xu, and J. Qiu, “Intense infrared luminescence in transparent glass-ceramic containing β-Ga2O3: Ni2+ nanocrystals,” J. Phys. Chem. C 111(20), 7335–7338 (2007). [CrossRef]
