Open Access
Add to BibSonomyAbstract
Silicon optical fibers fabricated using the molten core method possess high concentrations of oxygen in the core [Opt. Express 16, 18675 (2008)] due to dissolution of the cladding glass by the core melt. The presence of oxygen in the core can influence scattering, hence propagation losses, as well as limit the performance of the fiber. Accordingly, it is necessary to achieve oxygen-free silicon optical fibers prior to further optimization. In this work, silicon carbide (SiC) is added to the silicon (Si) core to provide an in situ reactive getter of oxygen during the draw process. Scanning electron microscopy (SEM), energy dispersive x-ray spectroscopy (EDX), and powder x-ray diffraction (P-XRD) are used to verify that the glass-clad silicon optical fibers possess very low oxygen concentrations and that the SiC is consumed fully during the reactive molten core fabrication. Optical measurements indicated a reduction in light scattering out of the silicon core as expected. However, the measured attenuation of about 10 dB/cm, which is consistent with existing low-oxygen-content silicon fibers, implies that scattering might not be the dominant source of loss in these molten core-derived silicon fibers. More generally, this work shows that the high temperature processing of optical fibers can be an asset to drive chemical reactions rather than be limited by them.
©2011 Optical Society of America
1. Introduction
A renaissance in optical fibers has begun and is focused on novel combinations of materials. Examples include multi-material fibers comprising polymers, metals, and soft-glasses that provide for a range of new optoelectronic functions [1] as well as novel combinations of highly dissimilar materials that facilitate unique active and passive performance in optical fibers [2–5].
An intriguing subset of this optical materials renaissance are optical fibers that consist of conventional (silica) glass claddings with semiconducting cores. To date, amorphous and crystalline unary (Si and Ge) and crystalline binary (InSb and ZnSe) semiconductor core optical fibers have been realized using, primarily, either a chemical vapor deposition (CVD) approach [6–8] or a molten core method [9–12]; with each having distinct advantages and disadvantages.
In general, CVD techniques utilize gaseous precursors that undergo vapor-phase chemical reactions that yield a different (solid) material. In conventional silica optical fibers, this is exemplified by the reaction: SiCl4(gas) + O2(gas) → SiO2(solid) + 2Cl2(gas). CVD processes rely on precursors that possess a suitable vapor pressure or chemical reactivity at a given temperature necessary to yield the desired material or phase.
While CVD processes employ vapor-phase reactions, the use of other reactive chemistries, such as in the solid state or between a solid and liquid, could have much broader implications for the creation of new and useful optical materials and structures. In differing forms, in situ reactive processing has been employed previously including for crystalline ZnSe [13] and, originally, for all-glass amplifier fibers [14].
This paper further advances this concept and specifically treats solid-liquid (melt) phase chemical reactions during the molten core fabrication of silicon optical fiber in order to achieve low oxygen content cores with reduced scattering. The molten core approach employed in this work requires that the core material melt at a temperature where the cladding glass is sufficiently soft so as to draw into optical fiber. Given the melting points of industrially-relevant unary and binary semiconductors, which can be well in excess of 1000°C (e.g., silicon melts at about 1412°C), the high temperature processing facilitates dissolution of the cladding glass by the core melt such that oxygen and other species are incorporated into the core as it solidifies upon fiber cooling. In the canonical system of a silica glass-clad silicon optical fiber, where only the elements Si and O are present, the issue becomes the removal of oxygen present in the silicon core [9]. The presence of oxygen, as oxide precipitates, is postulated to dominate scattering losses and also limits the minimum core size that can be achieved [12]. Removing or, at least, greatly decreasing the oxygen content would be an important step towards the ultimate realization of single mode silicon optical fibers.
Specifically, in this work, silicon carbide (SiC) is blended into silicon (Si) and used to reduce silica (SiO2) that enters the silicon core melt from the cladding via the reaction: SiO2 + SiC → Si + SiO + CO. The latter two reaction products (SiO and CO) are vapors and evolve out of the fluent melt such that the resultant optical fiber core is oxygen-free. This work further advances molten core fiber fabrication to include chemical reactivity whereby the high temperature processing, often considered a disadvantage, is employed to drive the thermodynamics rather than be limited by it.
2. Experimental procedures
2.1 Optical fiber fabrication
Silicon rods, equivalent to those used in [9], were pulverized, mixed with SiC powder (ABO Switzerland Co., Ltd., China), and the resultant powder blend placed into a high purity silica cladding tube (VitroCom, Inc., Mountain Lakes, NJ) of 30 mm outer diameter and about 3.2 mm inner diameter. The amount of SiC that was added to the Si powder was determined based on the SiO2 – SiC phase diagram, which is discussed later, given the oxygen content found in the original molten core silicon optical fibers of [9]. Approximately 5 meters of fiber with core size of about 200 μm (to ease in subsequent x-ray analysis) was drawn, largely limited in length by the initial amount of Si + SiC powder, at a speed of about 2 meters/minute at a temperature of 1925°C using a carbon resistance furnace purged with argon (Clemson University, Clemson, SC). Generally speaking, this molten core approach yields a fluent melt that flows along with the cladding tube as it draws into fiber [15]. Hereafter, said fibers will be referred to as “Si + SiC”-derived silicon optical fibers.
2.2 Electron microscopy and elemental analysis
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, in order to evaluate the elemental profile across the core/clad interface and to determine the effect of the SiC on the oxygen content in the core. Prior to microscopic evaluation, samples were polished to a surface finish of about 0.5 micrometer.
2.3 X-ray diffraction
Powder x-ray diffraction (PXRD) was performed on pulverized samples of both the precursor powder blend and the as-drawn fibers using a Rigaku Ultima IV powder diffractometer with Cu Kα radiation (λ = 1.5418 Å). Diffraction patterns were collected from 5 to 65° in 2-theta at a rate of 0.75 deg/min for the precursor blend and 0.1 deg/min for the powdered fiber. The powdered sample of the fiber contained material from both the crystalline core and glass cladding, and a background correction was applied to remove the amorphous roll contributed by the cladding glass. A slower scan (0.025 deg/min) also was performed on the powdered fiber sample from 33 to 39° in 2-theta, targeting the region that would contain the strongest diagnostic peaks of SiC to examine explicitly the sample for consumption of SiC.
2.4 Optical Transmission
Transmission measurements were made at 1.3 μm in a manner equivalent to that described in [9]. The output beam was imaged under 20 × magnification using an optical microscope and viewed with a Find-R-Scope 84499(A)-5 IR viewer. Given uncertainties in the exact transmission values due to light observed to propagate in the cladding, qualitative scattering measurements also were performed.
The experimental setup for the scattering measurements is shown in Fig. 1 . Briefly, a single-mode tapered fiber, mounted on a three dimensional translation stage, was used to couple light from a tunable laser (1.55 μm wavelength) into the sample. Since the tapered fiber can focus the input laser light into a spot of about two microns in diameter, it offers an effective way to couple light into either the core or cladding of the “Si + SiC”-derived silicon optical fiber. The output end of the sample is placed in front of a near-infrared (NIR) objective (50 × ) in order to form a magnified near field image in the infrared camera. The light intensity distribution at the sample output facet can be obtained using this setup. A second optical path includes a white light source, a beam-splitter, a beam sampler, and a visible CCD camera and is used to image the output facet of this sample.
Fig. 1 Schematic of the experimental set-up used for qualitatively evaluating transmission and scattering by the silicon optical fiber.
3. Results and discussion
3.1 Electron microscopy and energy dispersive spectroscopy
Scanning electron microscopy was employed to image the cross-section of the as-drawn and polished fiber. As can be seen in Fig. 2(a) , there is good core circularity and core/clad concentricity and interfacial integrity. Figures 2(b) and 2(c) provide spatial maps of the elemental distribution of silicon and oxygen, respectively, obtained using energy dispersive x-ray spectroscopy (EDX).
Fig. 2 (a) Scanning electron microscope (SEM) image of a silica glass-clad silicon core optical fiber drawn using a reactive molten core of Si + SiC. Energy dispersive spectroscopic (EDX) spatial maps qualitatively show the concentration of silicon (b) and oxygen (c) where a brighter region indicates higher elemental content.
Most striking is the lack of oxygen in the core [Fig. 2(c)]. Albeit qualitative, these results provide an initial indication as to the efficacy by which SiC can getter oxygen in this reactive molten core approach. Quantitative determination of composition also was performed using energy dispersive x-ray spectroscopy (EDX) by conducting elemental analysis at numerous points in a straight line across the fiber. As is shown in Fig. 3 , the cladding is stoichiometric SiO2 and the core has negligible oxygen, certainly compared to the about 18 atom percent of previous molten core fibers [9]. There are indications of oxygen present near the core/clad interface and it is uncertain if this influences optical scattering. Interestingly, oxide precipitates in silicon do have the beneficial effect of helping to pin dislocations thereby enhancing strength [16].
Fig. 3 Elemental profile across the silica glass-clad silicon core optical fiber drawn using a reactive molten core of Si + SiC.
3.2 X-ray diffraction
Figure 4 provides powder x-ray diffraction (PXRD) results on the initial blend of Si and SiC powders that subsequently were loaded into the pure silica cladding tube and drawn into fiber. As is expected, there is a match between the Si and SiC (cubic, moissanite-3C phase) contributions with their respective standard reflection indices.
Fig. 4 Powder x-ray diffraction scans of the core precursor powder blend of Si and SiC with crystallographic indices given. Figure 4(a) is the full intensity range whereas Fig. 4(b) is a magnified intensity range to accentuate the reflections and fit to the Si and SiC standards.
Figure 5 provides the PXRD spectra from the drawn optical fiber. Clearly identified in Fig. 5 are the peaks of silicon; the narrowness in linewidth further indicating a high degree of crystallinity. Figure 5(b) shows, to the precision of XRD, that the SiC has been consumed during the fiber draw process since no reflections from SiC are observed whereas there most certainly was SiC present in the initial preform (Fig. 4).
Fig. 5 (a) Powder x-ray diffraction (P-XRD) scan of the “Si + SiC”-derived silicon optical fiber with Si standard overlay and crystallographic indices for comparison; spectrum corrected to remove amorphous signature from cladding. (b) P-XRD scan of “Si + SiC”-derived silicon optical fiber over the selected two-theta region where SiC reflections should exist including an overlay of the SiC reflections. No reflections from SiC are observed in the drawn optical fiber.
3.3 Optical transmission
For 1.55 μm light incident on the center of the core of a silicon fiber processed without the SiC, a scattering pattern was observed in the cladding; i.e., light scattered out of the core and propagating in the cladding. However, for the “Si + SiC”-derived silicon optical fiber, there was no such scattering pattern in the cladding. This indicates, albeit qualitatively, that the core of Si + SiC fiber scattered less light. However, the total attenuation, measured at 1.3 μm, was 9.7 dB/cm after subtracting Fresnel reflections. This attenuation value is consistent with published reports on low-oxygen content silicon optical fibers, which range from about 15 db/cm to about 45 dB/cm, depending on fiber processing [17], at a similar wavelength.
These measurements imply that the removal of oxygen from the core significantly reduces the scattered light in the cladding. However, that there is no significant improvement in overall transmission suggests that the primary loss mechanism in these molten-core-derived silicon fibers might not be scattering from oxide precipitates as originally postulated [9,12]. Further, these findings indicate that the removal of oxygen from the core of the silicon fiber is just one piece in the larger puzzle associated with the transmission loss of (poly)crystalline semiconductor core fibers. Recent work has indicated that scattering is a major contributor to loss in CVD-derived silicon fibers [18] highlighting a difference in fiber behavior associated with the processing method.
3.4 Thermodynamics in the SiO2 + SiC system relevant to silicon optical fiber
Underlying these results is the technologically important thermodynamics of the Si-O-C system. The oxidation of SiC has been studied for over half a century [19,20; and references therein] and can be considerably more complicated than the simple stoichiometric reactions suggest. It was noted in [12] that the phase diagrams in the Si-O and Ge-O system are insightful into both their similarities and differences that are relevant to the fabrication of Si and Ge core optical fibers. This is equally true for SiC.
It is illustrative to discuss the Si-O-C phase diagram and, for this, an under-appreciated gem is [21]. Although the present work is focused only on the Si-O-C system [21], treats the more complete Si-O-C-N system (i.e., Si, SiO2, SiC, Si3N4, etc.). Consideration of nitrogen could be useful since nitrogen is a known additive into Czochralski-grown silicon and Si3N4 precipitates have been found in silicon optical fibers; see [12] and citations 44 and 45 therein.
As is shown schematically in Fig. 6(a) , a ternary phase diagram having Si, O, and C at its apexes has a smaller ternary phase field comprising Si, SiO2 (along the Si – O binary), and SiC (along the Si – C binary). The tie-line between the SiO2 and SiC, and its phase dependence with temperature, is critical to this discussion.
Fig. 6 Schematic representations of (a) the generalized Si-O-C ternary phase diagram, with smaller Si-SiO2-SiC ternary phase field highlighted, recast from [21] and (b) the SiO2 – SiC phase diagram recast and using the terminology from [19].
Figure 6(b) recasts the SiO2 – SiC phase diagram based on [21] and, for consistency, uses the terminology from said reference as follows: G = gas (vapor), CR = crystalline SiO2 (either cristobalite or tridymite), SiC = silicon carbide (SiC), Si = solid silicon (Si), C = carbon (graphite; though presumed not present in the case treated in this work), LS = liquid salt, and LM = metallic liquid.
As suggested by the phase diagram, the SiO2 + SiC = Si + SiO(g) + CO(g) stoichiometric reaction is more complex in practice. Table 1 provides a summary of the reactions and reaction temperatures that underlie the SiO2 + SiC thermodynamics. As the mixture of Si + SiC in the core of the silicon optical fiber preform is heated towards the draw temperature (1925°C or, about, 2200K), the first reaction is the melting of the silicon at 1685K (1412°C). This liquid Si phase, further facilitated by the progressively higher temperatures, begins to dissolve the softened SiO2 cladding above its glass transition temperature (about 1600°C or, about, 1873K). Amorphous silica, such as the cladding glass, does react more rapidly than crystalline SiO2 with SiC [22]. If crystalline SiO2 is present, or nucleates at the core/clad interface, then it melts at about 1996K (1723°C). Subsequently, any remnant SiC, which still is solid, is surrounded by a reactive melt and begins to liberate vapors at 2086K (1813°C). Bubbles of these vapors should fine (i.e., be removed by buoyancy) through the fluent melt at the draw temperature since the viscosity of silicon above its melt is less than that for water [23]. The reaction continues until either the SiC is consumed by the SiO2 or the fiber is drawn and cools thereby hindering the thermal driving forces; ultimately the core crystallizes upon solidification.
Table 1. Thermal Progression of Reactions in the SiO2 + SiC System*
This “road-map” behind the SiO2 + SiC = Si + SiO(g) + CO(g) reaction, taken with the phase diagram of Fig. 6(b), suggest a specific non-stoichiometric route to gettering the oxygen out of the silicon optical fiber. The reaction noted above specifies equimolar quantities of SiO2 and SiC. However, according to the phase diagram, such an equimolar ratio (50 mole % SiC in the Fig. 6(b)) will still have solid SiC as a thermodynamically stable phase at the draw temperature (denoted in Fig. 6(b) as the dotted horizontal line). Ideally, the only phases present at the draw temperature are the metallic liquid and vapor, that latter evolving out of the molten core. Accordingly, as was done in this work, slightly less than an equimolar amount of SiC, with respect to the amount of oxygen (SiO2) measured in the core of the [9] silicon optical fiber, was used.
4. Conclusion
Molten core silicon optical fibers suffer from dissolution of the cladding glass by the core melt. This dissolution brings oxygen into the core, which leads presumably to oxide precipitates which scatter light and generally degrade performance. Since dissolution is a thermally-driven process, the high draw temperatures of silica glass-clad silicon optical fibers further aggravates this issue. Employed here, for the first time to the best of our knowledge, is a reactive molten core approach whereby silicon carbide (SiC) is incorporated into the silicon core such that the reduction of SiO2 getters oxygen at the draw temperature. Such “Si + SiC”-derived silicon optical fibers were fabricated and were found by elemental analysis to exhibit no measurable oxygen in the core of the resultant fiber. Further, x-ray diffraction showed the SiC to be consumed during the draw and that the fiber core was comprised only of highly crystalline silicon. Scattering was reduced though the total attenuation, about 10 dB/cm at 1.3 μm, did not improve suggesting that oxygen in the core is not necessarily the major source of loss in the molten-core-derived silicon fibers. This work shows that the high temperature processing of optical fibers can be judiciously employed to take advantage of various chemical reactions to yield counter-intuitive results. Exemplifying this is the realization here of an oxygen-free silicon core optical fiber drawn in an oxide glass cladding at 2000°C; an important advance in the continued evolution of these crystalline semiconductor optical fibers.
Acknowledgments
The authors acknowledge financial support from Clemson University, the Northrop Grumman Corporation, and the Raytheon Company.
References and links
1. A. F. Abouraddy, M. Bayindir, G. Benoit, S. D. Hart, K. Kuriki, N. Orf, O. Shapira, F. Sorin, B. Temelkuran, and Y. Fink, “Towards multimaterial multifunctional fibres that see, hear, sense and communicate,” Nat. Mater. 6(5), 336–347 (2007). [CrossRef] [PubMed]
2. N. K. Goel, R. H. Stolen, S. Morgan, J.-K. Kim, D. Kominsky, and G. Pickrell, “Core-suction technique for the fabrication of optical fiber preforms,” Opt. Lett. 31(4), 438–440 (2006). [CrossRef] [PubMed]
3. N. Da, L. Wondraczek, M. Schmidt, N. Granzow, and P. Russell, “High index-contrast all-solid photonic crystal fibers by pressure-assisted melt infiltration of silica matrices,” J. Non-Cryst. Solids 356(35-36), 1829–1836 (2010). [CrossRef]
4. J. Ballato, T. Hawkins, P. Foy, S. Morris, N. K. Hon, B. Jalali, and R. Rice, “Silica-clad crystalline germanium core optical fibers,” Opt. Lett. 36(5), 687–688 (2011). [CrossRef] [PubMed]
5. N. Healy, J. R. Sparks, M. N. Petrovich, P. J. Sazio, J. V. Badding, and A. C. Peacock, “Large mode area silicon microstructured fiber with robust dual mode guidance,” Opt. Express 17(20), 18076–18082 (2009). [CrossRef] [PubMed]
6. D. Won, M. Ramirez, H. Kang, V. Gopalan, N. Baril, J. Calkins, J. Badding, and P. Sazio, “All-optical modulation of laser light in amorphous silicon-filled microstructured optical fibers,” Appl. Phys. Lett. 91(16), 161112 (2007). [CrossRef]
7. C. Finlayson, A. Amezcua-Correa, P. Sazio, N. Baril, and J. Badding, “Electrical and Raman characterization of silicon and germanium-filled microstructured optical fibers,” Appl. Phys. Lett. 90(13), 132110 (2007). [CrossRef]
8. J. R. Sparks, R. He, N. Healy, M. Krishnamurthi, A. C. Peacock, P. J. Sazio, V. Gopalan, and J. V. Badding, “Zinc selenide optical fibers,” Adv. Mater. (Deerfield Beach Fla.) 23(14), 1647–1651 (2011). [CrossRef] [PubMed]
9. 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]
10. J. Ballato, T. Hawkins, P. Foy, B. Yazgan-Kokuoz, R. Stolen, C. McMillen, N. K. Hon, B. Jalali, and R. Rice, “Glass-clad single-crystal germanium optical fiber,” Opt. Express 17(10), 8029–8035 (2009). [CrossRef] [PubMed]
11. J. Ballato, T. Hawkins, P. Foy, C. McMillen, L. Burka, J. Reppert, R. Podila, A. M. Rao, and R. R. Rice, “Binary III-V semiconductor core optical fiber,” Opt. Express 18(5), 4972–4979 (2010). [CrossRef] [PubMed]
12. 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]
13. N. Orf, O. Shapira, F. Sorin, S. Danto, M. Baldo, J. Joannopoulos, and Y. Fink, “Fiber draw synthesis,” Proc. Natl. Acad. Sci. U.S.A. 108(12), 4743–4747 (2011). [CrossRef]
14. E. Snitzer and R. Tumminelli, “SiO2-clad fibers with selectively volatilized soft-glass cores,” Opt. Lett. 14(14), 757–759 (1989). [CrossRef] [PubMed]
15. J. Ballato and E. Snitzer, “Fabrication of fibers with high rare-earth concentrations for Faraday isolator applications,” Appl. Opt. 34(30), 6848–6854 (1995). [CrossRef] [PubMed]
16. S. Hu, “Dislocation pinning effect of oxygen atoms in silicon,” Appl. Phys. Lett. 31(2), 53–55 (1977). [CrossRef]
17. L. Lagonigro, N. Healy, J. Sparks, N. Baril, P. Sazio, J. Badding, and A. Peacock, “Low loss silicon fibers for photonics applications,” Appl. Phys. Lett. 96(4), 041105 (2010). [CrossRef]
18. N. Healy, L. Lagonigro, J. R. Sparks, S. Boden, P. J. Sazio, J. V. Badding, and A. C. Peacock, “Polycrystalline silicon optical fibers with atomically smooth surfaces,” Opt. Lett. 36(13), 2480–2482 (2011). [CrossRef] [PubMed]
19. G. Ervin, “Oxidation behavior of silicon carbide,” J. Am. Ceram. Soc. 41(9), 347–352 (1958). [CrossRef]
20. W. Pultz and W. Hertl, “SiO2 + SiC reaction at elevated temperatures. Part 1— Kinetics and mechanism,” Trans. Faraday Soc. 62, 2499–2504 (1966). [CrossRef]
21. J. Weiss, H. Lukas, J. Lorenz, G. Petzow, and H. Krieg, “Calculations of heterogeneous phase equilibria in oxide-nitride systems. I. The Quaternary System C–Si–N–O,” Calphad 5(2), 125–140 (1981). [CrossRef]
22. M. Hakamada, Y. Fukunaka, T. Oishi, T. Nishiyama, and H. Kusuda, “Carbothermic reduction of amorphous silica refined from diatomaceous earth,” Metall. Mater. Trans., B, Process Metall. Mater. Proc. Sci. 41(2), 350–358 (2010). [CrossRef]
23. S. Nakamura and T. Hibiya, “Thermophysical properties data on molten semiconductors,” Int. J. Thermophys. 13(6), 1061–1084 (1992). [CrossRef]
