Described herein are glass-clad optical fibers, fabricated using a molten core fiber draw process, comprising oxide cores in the Bi2O3 – GeO2 system. More specifically, the fibers utilized a borosilicate glass cladding with core compositions in the initial preform ranging from un-reacted crystalline Bi2O3-rich (Bi2O3 + GeO2) powders to stoichiometric crystalline Bi12GeO20. Fibers drawn from the as-purchased crystalline Bi2O3-rich powders were amorphous with a transmission of about 80% at 1.3 μm. Fibers drawn from the crystalline Bi12GeO20 core contained a mixture of crystalline bismuth germanate (Bi2GeO5) and bismuth oxide (δ-Bi2O3/BiO2-x). While representing an initial proof-of-concept, this work shows that commercially-relevant draw processing can be employed to yield fibers with core composition that are very difficult to fabricate using conventional methods and that the molten core method further enables in situ reactive chemistry to take place during fiberization resulting in amorphous or crystalline oxide core fibers depending on initial core composition. Perhaps more importantly is that optical fibers possessing acentric, hence optically nonlinear, oxide crystals can be realized in a scalable manufacturing manner though further optimization is required both of the core chemistry and process conditions in order to achieve a single phase and single crystalline fiber.
©2012 Optical Society of America
Recent efforts in glass-clad crystalline core optical fibers have focused on semiconductors [1; and references therein]. While applications do abound for semiconductor core optical fibers, such as those enabled by their potential mid-wave infrared transparency and high Raman gain coefficients, an equally rich opportunity exists for crystalline oxide core optical fibers. Crystalline oxide fibers have been grown previously, principally using laser-heated pedestal growth or micro-pull down techniques . However, by conventional optical fiber fabrication standards, those techniques are slow (~0.1 mm/min) and can be problematic for cladding the crystalline core.
Of the approaches employed to make semiconductor core optical fibers, the molten core method  seems most practical for achieving long lengths where the real value of fiber over planar analogs is realized. In this process, a core composition is chosen that melts at a temperature below the draw temperature of the surrounding glass cladding. As the cladding is drawn into fiber, the core is a fluent liquid which solidifies as the fiber cools. More recently, reactive chemistry has been performed in the melt during the fiber draw process  opening the door to a wide variety of other potential core materials that cannot easily be directly melted or drawn into fiber; i.e., incongruent melting or high vapor pressure compound compounds.
The purpose of this work is to examine further this concept of the reactive molten core fabrication of optical fiber. However, contrary to past efforts, the focus here is on oxide core compositions, specifically two Bi2O3-rich compounds in the Bi2O3 – GeO2 family. This family is known to form glasses  as well as numerous stoichiometric crystals  that possess a range of useful optoelectronic properties. Accordingly, it provides a means to gauge the broader utility of the molten core method to the fabrication of unusual glass and, potentially, crystalline oxide core optical fibers. Compositions of about 6:1 Bi2O3:GeO2 were selected for this proof-of-concept effort since the equivalent crystal, Bi12GeO20, possesses a relatively low melting point and has useful scintillation behavior . Further, given the volatility of Bi2O3, which is known from micro-pull down growth of Bi4Ge3O12 (BGO) fibers [8–10], the 6:1 Bi2O3:GeO2 precursor might ultimately yield the BGO phase (equivalent to 2:3 Bi2O3:GeO2). BGO is a cubic crystal (space group I-43d) and exhibits a Kerr electro-optic coefficient of r41 ~3.8 pm/V . If the resultant fiber core is amorphous, rather than crystalline, high Bi2O3 content glasses are of interest for their high nonlinear refractive index and chemical stability [12,13]. As such, this method might result in a scalable method for fabricating χ(2) or χ(3) nonlinear oxide core optical fibers. More generally, though, the present effort seeks to determine if the molten core approach can yield glass-clad crystalline oxide optical fibers. Previous efforts on oxide crystals, such as YAG (Y3Al5O12), were unsuccessful in achieving a crystalline core likely due to several factors including dissolution of silica into the core melt from the cladding , the crystallographic complexity of YAG and its ability to be considerably under-cooled without crystallizing .
Bismuth oxide (Bi2O3) and germanium oxide (GeO2) powders in a molar proportion of about 6:1 were placed into a glass vial and vortex-mixed for 2 min. The powder mixture then was calcined in an alumina crucible at a temperature of 650°C for 1 hour in order to reduce hydroxyl content. For comparison, a single crystal of Bi12GeO20 (equivalent to 6:1 molar ratio of Bi2O3:GeO2) also was employed, which was grown by one of the authors (O.S.) more than 30 years ago by the Czochralski method using individual oxides melted together in an atmosphere of air.
Optical fibers were fabricated at Clemson University wherein a sodium borosilicate (containing a low concentration of alumina) glass tube was drawn at a temperature of about 1020°C, which exceeds the 925°C melting point of the precursor 6:1 Bi2O3:GeO2 core crystalline compounds, in a nitrogen atmosphere at ambient pressure . As such, the core is a liquid during the drawing of the glass cladding into fiber and solidifies as the fiber cools. In the case of the fiber drawn from the Bi12GeO20, the single crystal was core-drilled into a solid bulk rod that then was sleeved into the borosilicate cladding tube. In both cases, i.e., the fiber drawn from the Bi2O3 + GeO2 powder or from the solid Bi12GeO20 single crystalline rod, the draw speed was 0.5 m/min. The resultant fibers had core and total diameters, respectively, of about 330 μm and about 2.0 mm for the powder-derived samples and about 734 μm and about 1.94 mm for the single crystal-derived sample. The different ratio of core/clad diameters results from differing borosilicate cladding tubes being used as one size was no longer commercially available. Based on prior results in other crystalline core fiber systems , this different in core/clad ratio is not expected to plan an important role in core chemical reactions. No bubbles, potentially a result of residual porosity in the powder-derived fiber, were observed in the drawn fiber.
Powder x-ray diffraction (PXRD) was performed on both the precursor materials and on the as-drawn fiber using a Scintag XDS 2000-2 powder diffractometer with Cu Kα radiation (λ = 1.5418 Å). Diffraction patterns were collected from 5° - 65° in 2-theta. A background correction was applied to the data from the optical fiber in order to remove the amorphous roll contributed by the borosilicate cladding glass. Figure 1 provides the PXRD scans for the two precursor materials used in this work along with reference standards [16–19].
Characterization using electron microscopy was performed using a Hitachi S-3400 variable pressure scanning electron microscope (SEM) operating under variable pressure at 20kV and a working distance of about 10 mm. Elemental analysis was conducted under high vacuum, using energy dispersive x-ray (EDX) spectroscopy in secondary electron (SE) mode.
Transmission measurements were conducted at wavelengths of 1306 and 1534 nm using an equivalent method to that described in . As is described below, only the amorphous core fiber yielded transmission results.
Results and discussion
Bismuth-rich glass core optical fibers
Figure 2(a) shows the electron micrograph image of the borosilicate-glass clad optical fiber derived from the powder bismuth germanate mixture. As is observed, there is good circularity and uniformity as well as being no obvious discontinuities at the core/clad interface. PXRD analysis of the fiber yielded no crystalline peaks; suggesting that the bismuth-rich melt was quenched to an amorphous form.
Figure 3 provides the EDX compositional profile, both in elemental and compound atom percentages, across the entire fiber. As can be seen, there is some volatilization of the GeO2 and dissolution of the SiO2 into the molten core from the glass cladding. As a result, the core composition is better represented as a bismuth silicate glass rather than a bismuth germanate based on the initial core composition. It should be noted for completeness that boron is too light an element to be reliably detected using EDX and so the B2O3 is not shown in this elemental analysis. It is expected that there also is B2O3 present in the core which likely substitutes structurally for the Bi2O3.
Of particular interest is the fact that the core possesses a composition of 70Bi2O3-30SiO2 (mole percent), which is nearly 95Bi2O3-5SiO2 in weight percent (discounting the ~1 percent by weight contributions of the GeO2 and Al2O3. This is considerably higher than the 78 weight % bismuth silicate glass previous studied as a highly nonlinear optical fiber for all-optical signal processing . These high Bi2O3-content glasses are to be differentiated from the low-doping level silicates studied as potential near-infrared fiber amplifiers and lasers [22,23].
The transmission of the amorphous high Bi2O3-content core fiber was 81% and 75% at wavelengths of 1.3 and 1.5 μm, respectively, not adjusting for Fresnel reflections. Not knowing the refractive index at these wavelengths precluded a more precise calculation of a loss value. In these initial fibers, losses likely are dominated by extrinsic absorptions since high purity powders were not used and enhancements to attenuation, although reasonable to-date, represent an opportunity for future advancement.
By way of an initial summary for this section, borosilicate glass-clad optical fibers drawn with a bismuth germanate powder mixture in the core yielded a very high Bi2O3-content silicate core. Such a core should possess higher nonlinearity than the high bismuth-content glasses reported previously and will be a topic of future study.
Bismuth germanate crystalline core optical fiber
Figure 4 provides the powder x-ray diffraction results on the optical fiber drawn from the core of single crystalline Bi12GeO20 (Fig. 1(b) shows the PXRD scan for this crystal, which is well-matched to its established crystallographic reflections ). The as-drawn fiber does indeed exhibit crystalline x-ray reflections though they are found to correspond to a mixture of two different phases: Bi2GeO5 and bismuth oxide.
Interestingly, the precursor Bi12GeO20 (6:1 Bi2O3:GeO2) did not dissociate to the desirous BGO phase (2:3 Bi2O3:GeO2) but, rather, formed the slightly more Bi2O3-rich phase of Bi2GeO5 (1:1 Bi2O3:GeO2) and bismuth oxide. Thus, the general reaction occurring during the fiber draw process is:
Upon careful investigation of the observed peak positions, we note that the peaks corresponding to the bismuth oxide product in the as-drawn fiber are slightly shifted to fall between the indexed peak positions for δ-Bi2O3  (a cubic phase of Bi2O3 stable above 730°C, a = 5.6595Å, likely metastable under the experimental conditions of this study) and BiO2 (a = 5.539 Å) . This suggests that the above reaction is enriched in oxygen to cause partial oxidation of Bi3+ to Bi5+, yet not to the extent where a BiO2 composition is obtained. So the actual reaction occurring as the fiber is drawn is more accurately approximated by:
The presence of enriched oxygen in the core would not be surprising given the propensity for oxygen diffusion from the cladding glass observed in our previous studies using the same cladding glass and draw temperature (for example, ~5 at.% in Ge optical fiber ).
Such a reaction is supported by the elemental profiles of the fiber cross-section obtained using energy dispersive x-ray (EDX) spectroscopy and shown in Fig. 5 . A Ge:Bi:O ratio of ~1:12:31 is observed confirming oxygen enrichment in the core compared to what would be observed (1:12:20) if the reaction proceeded according to Eq. (1). Furthermore we also observe Si diffusion into the core, ~9 atom percent, which is considerably less than the ~30 atom percent found above in the powder-derived amorphous core optical fiber. As there was no additional silicate phase identified by PXRD, the borosilicate cladding glass likely provides the majority of this contribution. Higher concentrations of Si and O near the core-cladding interface are also characteristic of a diffusion process.
The greater concentration of silica in the powder-derived amorphous core fiber, in comparison to the Bi12GeO20-derived crystalline core fiber, has two potential origins. Bi12GeO20 melts at about 925°C whereas Bi2O3, being the main component in the alternative fiber draw, melts at 817°C (GeO2 does not melt by itself until 1115°C but one can assume that the molten Bi2O3 will dissolve the GeO2). It is likely that this reduction in melting point, which offers greater time for the melt to interact with the cladding glass thereby dissolving more of it into the core leads to the greater SiO2 content in the amorphous core optical fiber. Alternatively, this longer time that the Bi2O3-rich melt has to interact with the cladding glass might allow for more Bi2O3 volatilization and shift the Bi:Si ratio leading to the effectively higher SiO2 concentration. Either way, such reactive chemistry in the melt, and between the core and clad offer addition degrees of flexibility in the materials chemistry and phase equilibria (or non-equilibria in the case of the amorphous core) for the design of novel optical fibers.
It also is noted that the PXRD peaks corresponding to Bi2GeO5 are shifted to slightly higher angles (by ~0.1 degrees in 2Θ). This could indicate a small amount of substitution by B or Si for Ge also made possible by cladding-core diffusion. In particular, boron substitution to produce Bi2Ge1-xBxO5-.5x would make additional oxygen available for oxidation of Bi2O3. Some degree of non-stoichiometry is suggested by the pale yellow color exhibited by the core in the as-drawn fiber. Larger degrees of boron or silicon (as Bi2SiO5 is iso-structural with Bi2GeO5) substitution may also be indicated by the shoulders on the higher angle side of the observed Bi2GeO5 peaks in Fig. 4. Though not anticipated, these results speak to the chemical flexibility of B, Si and Ge as well as the complexity of the Bi2O3 – GeO2 phase diagram  given the poly-valency of bismuth and the volatility of bismuth oxides.
The crystalline Bi2GeO5 of the core is polar acentric (orthorhombic symmetry; space group Cmc21) and, as a result, is optically-active, pyroelectric, piezoelectric, optically-nonlinear, and electro-optic . Unfortunately, given the biphasic mixture in the core, their respectively high refractive indices and optical anisotropies , and cracks (Fig. 2(b)) likely due to thermal expansion or elastic constant differences, precluded any waveguide or transmission measurements from these crystalline oxide core fibers.
For completeness, it is worth noting that Bi4Ge3O12 and Bi2GeO5 have been realized by the controlled phase-separation of a Bi2O3-GeO2 glass . Such a post-fabrication process could be performed on a drawn fiber, such as has been done for fiber amplifiers , to realize an electro-optic fiber. However, in this work, the as-drawn fiber exhibited the crystalline oxide phases without the need for secondary processing. Additionally, this work provides a further example of reactive chemistry that can occur during the draw to yield optical fibers containing compounds in their core that would be very difficult to realize using other methods .
Clearly, further optimization is required in order to achieve a phase-pure acentric single crystalline optical fiber. It is known from recent work that annealing and core geometry can significantly influence single crystallinity in these crystalline core optical fibers [29,30]. That said, such a crystalline core fiber could offer a unique palette of optical and optoelectronic properties heretofore unavailable including χ(2) nonlinearities, piezo-optic, second-harmonic generation, parametric amplification, optical modulation, electric field sensing and scintillation, to name just a few. Additionally, the molten core fabrication approach is especially useful for manufacturing long continuous lengths of optical fiber, which would be required for practical implementation of the resultant fiber-based devices. Further, BGO (Bi12GeO20) continues to be of great technological interest in that it possesses an extremely wide transparency range (600 nm to over 9 microns) and has a strong photo refractive effect . Fibers with BGO cores and glass cladding could be used to control infrared lasers via index gratings written into the core through the side using visible light in undoped BGO or by near infrared radiation in doped material .
In conclusion, a reactive molten core process was employed to make optical fibers with cores in the bismuth germanate family. Depending upon the nature of the precursors, for similar compositions, either amorphous or crystalline core compositions were realized under equivalent processing conditions. Perhaps more importantly, this proof-of-concept work establishes, for the first time to the best of our knowledge, the direct fiber draw fabrication of optical fibers comprising a crystalline oxide core. The core, derived from Bi12GeO20, underwent reactive chemistry during the fiber draw process and ultimately contained a biphasic crystalline mixture of acentric Bi2GeO5 and cubic bismuth oxide (δ-Bi2O3/BiO2-x). Future work will focus on tailoring the initial core composition, based on the Bi2O3-GeO2 phase diagram, to yield a reaction pathway to optical fibers ranging from highly nonlinear high-Bi2O3-content glasses to phase-pure Bi2GeO5 or, potentially, the equally interesting compounds of Bi12GeO20 or Bi4Ge3O12.
The authors wish to acknowledge the efforts of Laura Burka and Courtney Kucera (Clemson University) and financial support from the Defense Advanced Research Projects Agency, under contract HR0011-08-C-0137, and the Northrop Grumman Corporation.
References and links
1. J. Ballato, T. Hawkins, P. Foy, B. Yazgan-Kokuoz, C. McMillen, L. Burka, S. Morris, R. Stolen, and R. Rice, “Advancements in semiconductor core optical fiber,” Opt. Fiber Technol. 16(6), 399–408 (2010). [CrossRef]
2. R. Feigelson, “Pulling optical fibers,” J. Cryst. Growth 79(1-3), 669–680 (1986). [CrossRef]
4. S. Morris, T. Hawkins, P. Foy, C. McMillen, J. Fan, L. Zhu, R. Stolen, R. Rice, and J. Ballato, “Reactive Molten Core Fabrication of Silicon Optical Fiber,” Opt. Mater. Express 1(6), 1141–1149 (2011). [CrossRef]
5. B. Kusz, K. Trzebiatowski, and R. Barczynski, “Ionic conductivity of bismuth silicate and bismuth germanate glasses,” Solid State Ion. 159(3-4), 293–299 (2003). [CrossRef]
6. A. Kaplun and A. Meshalkin, “Stable and metastable phase equilibrium in the system Bi2O3 – GeO2,” J. Cryst. Growth 167(1-2), 171–175 (1996). [CrossRef]
7. M. Weber and R. Monchamp, “Luminescence of Bi4Ge3O12: Spectral and Decay Properties,” J. Appl. Phys. 44(12), 5495–5499 (1973). [CrossRef]
8. J. Shim, J. Leea, A. Yoshikawaa, M. Niklb, D. Yoonc, and T. Fukudaa, “Growth of Bi4Ge3O12 single crystal by the micro-pulling-down method from bismuth rich composition,” J. Cryst. Growth 243(1), 157–163 (2002). [CrossRef]
9. M. Zhuravleva, V. Chani, T. Yanagida, and A. Yoshikawa, “The micro-pulling-down growth of Bi4Si3O12 (BSO) and Bi4Ge3O12 (BGO) fiber crystals and their scintillation efficiency,” J. Cryst. Growth 310(7-9), 2152–2156 (2008). [CrossRef]
10. H. Farhi, S. Belkahla, K. Lebbou, and C. Dujardin, “BGO fibers growth by μ-pulling down technique and study of light propagation,” Phys. Procedia 2(3), 819–825 (2009). [CrossRef]
11. D. Bortfeld and H. Meier, “Refractive indices and electro-optic coefficients of the eulitities Bi4Ge3O12 and Bi4Si3O12,” J. Appl. Phys. 43(12), 5110–5111 (1972). [CrossRef]
12. N. Sugimoto, H. Kanbara, S. Fujiwara, K. Tanaka, Y. Shimizugawa, and K. Hirao, “Third-order optical nonlinearities and their ultrafast response in Bi2O3-B2O3-SiO2 glasses,” J. Opt. Soc. Am. B 16(11), 1904–1908 (1999). [CrossRef]
13. Y. Kuroiwa, N. Sugimoto, K. Ochiai, S. Ohara, Y. Fukasawa, S. Ito, S. Tanabe, and T. Hanada, “Fusion Spliceable and High Efficient Bi2O3-based EDF for Short-length and Broadband Application Pumped at 1480 nm,” in Optical Fiber Communication Conference, 2001 OSA Technical Digest Series (Optical Society of America, 2001), paper TuI5.
14. J. Ballato, T. Hawkins, P. Foy, B. Kokuoz, R. Stolen, C. McMillen, M. Daw, Z. Su, T. Tritt, M. Dubinskii, J. Zhang, T. Sanamyan, and M. J. Matthewson, “On the Fabrication of All-Glass Optical Fibers from Crystals,” J. Appl. Phys. 105(5), 053110 (2009). [CrossRef]
15. B. Cockayne, “The uses and enigmas of the Al2O3-Y2O3 phase system,” J. Less Common Met. 114(1), 199–206 (1985). [CrossRef]
16. Powder Diffraction Standard Number: 00–027–0053 (International Centre for Diffraction Data (ICDD), Newtown Square, PA, USA).
17. J. Haines, O. Cambon, E. Philippot, L. Chapon, and S. Hull, “A neutron diffraction study of the thermal stability of the alpha-quartz-type structure in germanium dioxide,” J. Solid State Chem. 166(2), 434–441 (2002). [CrossRef]
18. Powder Diffraction Standard Number: 01–073–9108 (International Centre for Diffraction Data (ICDD), Newtown Square, PA, USA).
19. M. Morris, H. McMurdie, E. Evans, B. Paretzkin, H. Parker, N. Pyrros, and C. Hubbard, “Standard x-ray diffraction patterns - data for 71 substances,” NBS Monograph 25(Sec. 20), 19 (1984).
20. J. Ballato, T. Hawkins, P. Foy, R. Stolen, B. Kokuoz, M. Ellison, C. McMillen, J. Reppert, A. M. Rao, M. Daw, S. R. Sharma, R. Shori, O. Stafsudd, R. R. Rice, and D. R. Powers, “Silicon optical fiber,” Opt. Express 16(23), 18675–18683 (2008). [CrossRef] [PubMed]
21. K. Kikuchi, K. Taira, and N. Sugimoto, “Highly-nonlinear Bismuth Oxide-based glass fibers for all-optical signal processing,” in Optical Fiber Communications Conference, A. Sawchuk, ed., Vol. 70 of OSA Trends in Optics and Photonics (Optical Society of America, 2002), paper ThY6.
22. M. A. Melkumov, I. A. Bufetov, A. V. Shubin, S. V. Firstov, V. F. Khopin, A. N. Guryanov, and E. M. Dianov, “Laser diode pumped bismuth-doped optical fiber amplifier for 1430 nm band,” Opt. Lett. 36(13), 2408–2410 (2011). [CrossRef] [PubMed]
23. V. V. Dvoyrin, V. M. Mashinsky, L. I. Bulatov, I. A. Bufetov, A. V. Shubin, M. A. Melkumov, E. F. Kustov, E. M. Dianov, A. A. Umnikov, V. F. Khopin, M. V. Yashkov, and A. N. Guryanov, “Bismuth-doped-glass optical fibers--a new active medium for lasers and amplifiers,” Opt. Lett. 31(20), 2966–2968 (2006). [CrossRef] [PubMed]
24. H. Hartwig, “On the Structure of Bismuthsesquioxide: The α, β, γ, and δ-phase,” Z. Anorg. Allg. Chem. 444(1), 151–166 (1978). [CrossRef]
25. B. Begeman and M. Jansen, “Bi407, the First Defined Binary. Bismuth(lll,V)-Oxide,” J. Less-Common Met. 156, 123–135 (1989).
26. P. Halasyamani and K. Poeppelmeier, “Noncentrosymmetric Oxides,” Chem. Mater. 10(10), 2753–2769 (1998). [CrossRef]
27. L. Dimesso, G. Gnappi, A. Montenero, P. Fabeni, and G. P. Pazzi, “The crystallization behaviour of bismuth germanate glasses,” J. Mater. Sci. 26(15), 4215–4219 (1991). [CrossRef]
29. N. Gupta, C. McMillen, R. Singh, R. Podila, A. M. Rao, T. Hawkins, P. Foy, S. Morris, R. Rice, K. F. Poole, L. Zhu, and J. Ballato, “Annealing of Silicon Optical Fibers,” J. Appl. Phys. 110(9), 093107 (2011). [CrossRef]
30. S. Morris, C. McMillen, T. Hawkins, P. Foy, R. Stolen, R. Rice, and J. Ballato, “The influence of core geometry on the crystallography of silicon optical fiber,” J. Cryst. Growth (to be published), doi:. [CrossRef]
31. J. White and A. Yariv, “Real-time image processing via four-wave mixing in a photorefractive medium,” Appl. Phys. Lett. 37(1), 5–7 (1980). [CrossRef]
32. J. Martin, I. Foldvari, and C. Hunt, “The low-temperature photochromic response of bismuth germanium oxide,” J. Appl. Phys. 70(12), 7554–7559 (1991). [CrossRef]