Splices between materials with dissimilar thermal expansion and melting points are particularly difficult to create. We have developed a method for splicing YAG single crystal fiber to silica fiber. Optical losses associated with the splices were measured for multimode fibers to be 0.33 dB. The splices display greater than 50 kPsi of tensile strength with reaction bonding at the interface. Study of the elemental composition at the splice interface showed formation of a stable intermediate material that provides mechanical strength to the splice. This is a major step toward developing very high power integrated and compact laser systems based on crystals and glass despite their stark dissimilarities in physical and material properties.
© 2016 Optical Society of America
Rare earth doped silica glass fiber lasers possess low thermal conductivities (<1 Wm−1K−1) and low stimulated Brillouin scattering (SBS) thresholds, which limits single frequency glass fiber lasers to about 1.2 kW output power . Scaling to higher power requires materials with high thermal conductivity so that heat can be readily removed from the amplifier, in this case the core of the fiber. Crystals such as yttrium aluminium garnet (YAG) offer several advantages over glass due to their higher thermal conductivity and higher SBS threshold, along with excellent environmental stability and ability to accept higher concentrations of rare earth dopants [2, 3].
One challenge in utilizing YAG crystal fibers for high power applications is integrating the crystal fiber with silica fiber-based components. Currently, this is achieved with mechanical splicing or free-space coupling techniques. Fusion splicing is the preferred technique for coupling fiber components as it offers the potential for robust (increased reliability), higher throughput (reduced Fresnel loss) and better thermal management (no air interface, higher power handling), but it remains challenging to perform with dissimilar fiber materials. For applications beyond the laboratory environment such as fieldable high-power laser systems, demonstrations of robust splices with all these benefits are needed. While we have demonstrated the bonding of dissimilar glass fibers , bonding glass to crystal fiber has its own unique challenges. Splicing YAG crystal to silica glass is particularly difficult because of differences in melting temperature, viscosity-temperature profiles, coefficients of thermal expansion and refractive indices. Table 1 compares these properties for YAG and silica. Unlike glass, which softens gradually when heated, crystalline YAG abruptly melts into a low viscosity liquid at its melting point of 1930°C. Additionally, when molten YAG is cooled from high temperatures, a sharp exothermic maximum around 1600°C has been observed due to spontaneous solidification at a high degree of supercooling . To avoid the complexity of the crystal melting and spontaneous solidification, it is crucial to avoid melting the crystal during splicing. An attempt to join crystal and glass fibers using fusion splicing was reported in 1995 by Barnes et al.  using sapphire and silica fibers with aluminosilicate glass as a joining medium but the losses were high (~10dB).
In this paper, we report a successful direct fusion splicing of the single crystal YAG and silica glass fibers with very low loss without the need of any external coating on the endfaces or performing intermediate steps prior to fusing the fibers together. We also investigated the chemical composition of the spliced interface to understand the spliced interface better.
2. Single crystal fiber fabrication
Single crystal fibers have been grown using the laser heated pedestal growth (LHPG) technique . Single crystal YAG fibers with a typical length of approximately 1 m and diameter ranging from 20 µm to 100 µm were drawn by LHPG at a rate of ~1 cm/min . These fibers show good mechanical strength and flexibility. In general, the fibers have good optical quality with high transparency . The YAG fiber used here is a 100 µm single crystal core-only fiber with no cladding. The fiber is doped with Yb2O3, for lasing and gain at a wavelength of around 1030 nm.
3. Fusion splice between YAG and silica
A filament-based commercial splicer, Vytran GPX-3000, was used for the development of a silica-YAG fiber splice. Preparing the fiber end for splicing is an important step toward minimizing splicing loss. Adverse environmental conditions (such as high humidity, dust) are controlled in order to avoid contamination of the fibers in the vicinity of the splice, potential fiber alignment issues due to interference of dust particles and the associated inherent optical losses. Cleaving is another issue with single crystal fiber as it prefers to cleave along crystalline planes which may or may not be perpendicular to the fiber axis. However, we found that the YAG fiber with 100 µm diameter can be cleaved with standard diamond edge cleaver and the endface quality achieved is comparable to that attained by polishing the YAG fiber (See Fig. 1). A 125 µm diameter multimode silica fiber with a 65 µm diameter Ge-doped core with NA of 0.27 was stripped off its polymer outer jacket and the endfaces were prepared using the commercial cleaver. Once the fiber end faces are prepared, it is very important to complete the splicing process in a short time in order to minimize contamination.
The silica and YAG fibers have more than 10 × difference in their coefficients of thermal expansion. As the two fibers are heated during the fusion splice process, the YAG fiber expands in length much more quickly than the silica fiber, so it is crucial to start with a separation between the two fibers of approximately 10 µm. Because the YAG fiber has a high melting temperature (1930°C) compared to the softening temperature of the silica fiber (~1300°C) the heating element is offset toward the YAG fiber side by 0.5 mm from center to prevent overheating of the silica fiber. During the splicing process, the YAG fiber expands to touch the silica fiber, and due to its high thermal conductivity, it transfers heat directly to the silica fiber endface, softening the silica fiber very quickly. Figure 2 shows an optical micrograph of the splice between the silica and YAG fiber.
The splice loss was determined by measuring the attenuation of a 1.6 µm laser source through the splice. The single mode CW laser light exiting a standard single mode silica fiber (NA = 0.12) was injected into the multimode silica fiber (NA = 0.27) and the power output from the YAG fiber (NA = 1.5 for air clad) was actively measured and monitored during the fusion splicing. A splice loss of 0.33 dB ( ± 0.02 dB) was measured. The measurement technique does not correct for the Fresnel loss (around 0.06 dB) at the silica-YAG spliced interface. Once the splice technique was optimized, we performed 4 splices between YAG and glass fiber, including, phosphate doped silica and germanate doped silica, all with the losses about 0.3-0.4 dB.The mechanical strength of the splice was measured using a Vytran tensile strength tester (GPX-3000) in order to confirm the robustness and usefulness of the splice for practical applications. The silica-YAG splice was tested to the maximum applied static load of the instrument, more than 400 g of tension in the direction of the fibers axes before fracturing which is equivalent to ~50 kpsi of tensile strength for the 100 μm diameter fiber. Even though the measured strength of the splice between YAG and silica fibers is smaller than the splice between silica fibers, it is sufficient for use in the fabrication of practical devices and industrial prototypes based on silica and YAG fibers and alleviates the need for mechanical splicing or free-space coupling. While it was not measured in this work, we do anticipate that the thermal expansion mismatch between these dissimilar fiber materials will play an important role in the long term durability of the fiber splice when used in a device, especially with regard to thermal cycling. This is an area of interest for ongoing study.
4. Elemental analysis of spliced region
To better understand the bonding mechanism behind the strength of the splice, electron probe microanalysis (EPMA, JEOL Superproble 733) was performed to measure the elemental profile across the spliced-region. A robust silica-YAG splice was mounted in epoxy and material was ground away and polished exposing the fiber midplane as shown in the SEM image in Fig. 3. Wavelength dispersive spectroscopic (WDS) line scanning was performed to measure the concentration of known elements: Si, Ge, Y, Al, Yb and O in the vicinity of the splice. Scans were performed in the direction of the fiber axis at the fiber center (within the silica fiber core region) and near the fiber edge (within the silica fiber cladding region) over a distance of 100 µm in 5 µm increments in order to accurately locate the interface region. The scan interval was reduced to 1 µm (the resolution limit of the instrument) for the 30 µm region near the interface. The SEM micrograph in Fig. 3 clearly shows the formation of an intermediate material at the splice interface (called out by the arrow).
The elemental profile for the scan performed in the “core region” is shown in Fig. 4 on a percent atomic basis. The elemental profiles for Si, Y, Al, and Yb show concentration gradients over distances of about 9 µm in the axial direction. The profile for germanium however varies over a distance of only 2 µm, which is essentially a discontinuity considering the 1 µm spatial resolution of the instrument. This discontinuity in the germanium profile was used as a reference to define the location of the splice interface for all profiles.
The oxide concentration profile, on a weight percent basis, is shown in Fig. 5. To improve clarity of the plot, the dopants (Yb2O3 and GeO2) have been combined with their hosts (Y2O3 and SiO2, respectively). The intermediate material at the splice is delimited by the markers A and B. The yttrium-alumina-silica ternary is a well-studied system, the phase diagram is shown in Fig. 6 and was adapted from the 1963 paper by Bondar and Gapakhov . This phase diagram shows the liquidus surface and has eight invariant points indicated by the intersections of the thick lines, including a low-melting eutectic composition with 32% of Y2O3, 22% of Al2O3 and 46% of SiO2 (in wt. %) at 1345°C. The oxide profile measurement points between markers A and B (within the splice region in Fig. 5) have been overlaid on the ternary phase diagram in Fig. 6 as circumscribed x’s. It should be acknowledged that the reference phase diagram in Fig. 6 does not account for the dopants (Yb, Ge) that are present in the fibers here. The presence of these dopants could affect the compositions and temperatures of the eutectic and reaction points and add additional phases to the phase diagram region of interest. For example, the eutectic for Yb2O3–Al2O3–SiO2 is about 100°C higher than YAS and has similar SiO2 solubility . However the dopant levels in our YAG fiber are low and their profiles within the splice region mimic those of their hosts in Fig. 5, so for the purpose of this discussion, the dopants have been grouped with their hosts and treated as the host oxide.
The existence of low melting ternary eutectic and reaction points enables the co-melting of YAG and silica at a temperature (~1350°C) much lower than the melting points of the individual fibers and allows for robust fusion splicing of these dissimilar fibers. When the fibers are heated during splicing, the interface reaches this eutectic or reaction temperature forming a molten zone at the interface. The slope of the profile curves between points A and B in Fig. 5 can be attributed to co-diffusion within this molten zone and indicates the formation of an intermediate material with a graded composition. The intermediate material in the splice region is the product of cooling from this molten zone and creates a strong bond between the two fibers. While the elemental composition of the interface material was measured using EPMA, as shown in Fig. 5, the specific phases present, i.e. the crystalline or amorphous nature of the material was not determined and is an area of interest for future study. The elemental composition of the YAS intermediate material is consistent with amorphous and nano-crystalline material produced by others pursuing crystal-glass “hybrid” fibers recently [11, 12]. Ballato et al demonstrated amorphous YAS by co-drawing YAG inside a silica preform on a fiber optic draw tower at 2025°C . Lo et al have reported an intermediate material containing nano-crystals in an amorphous silica matrix when co-drawing single-crystal YAG fibers inside silica capillary using LHPG .
Rare-earth doped single crystal YAG fibers are an attractive alternative to bulk crystals and doped glass fibers for very high power lasers on the order of tens of kW, but their integration with conventional and ubiquitous silica fiber components and devices has thus far been limited to mechanical splicing and free-space coupling techniques. The robust fusion splice of single crystal Yb:YAG and silica fibers was demonstrated for the first time using a commercial filament splicer. The splice was sufficiently robust for handling and incorporation in integrated systems with a splice strength in excess of 50 kpsi and a splice loss of 0.33 dB. The composition profile in the splice region was measured using EPMA and found to contain a graded YAS composition consistent with low temperature eutectic or reaction melting at the fiber interface during the fusion splicing process. Development of this robust splicing method between single crystal YAG fiber and silica glass fibers provides a major step toward the production of very high power integrated and compact laser systems based on integration of crystal and glass fiber components, in the order of tens of kW, despite the stark dissimiliraties in physical and material properties between the two fibers.
References and links
1. J. W. Dawson, M. J. Messerly, R. J. Beach, M. Y. Shverdin, E. A. Stappaerts, A. K. Sridharan, P. H. Pax, J. E. Heebner, C. W. Siders, and C. P. J. Barty, “Analysis of the scalability of diffraction-limited fiber lasers and amplifiers to high average power,” Opt. Express 16(17), 13240–13266 (2008). [CrossRef] [PubMed]
4. R. Thapa, R. R. Gattass, V. Nguyen, G. Chin, D. Gibson, W. Kim, L. B. Shaw, and J. S. Sanghera, “Low-loss, robust fusion splicing of silica to chalcogenide fiber for integrated mid-infrared laser technology development,” Opt. Lett. 40(21), 5074–5077 (2015). [CrossRef] [PubMed]
5. J. L. Caslavsky, D. J. Viechnicki, M. Army, and M. A. Mechanics Research Center Watertown, “Study of the Melting Behavior of YAG Single Crystal by Optical Differential Thermal Analysis,” (1979).
7. M. M. Fejer, J. L. Nightingale, G. A. Magel, and R. L. Byer, “Laser‐heated miniature pedestal growth apparatus for single‐crystal optical fibers,” Rev. Sci. Instrum. 55(11), 1791–1796 (1984). [CrossRef]
8. W. Kim, C. Baker, G. Villalobos, C. Florea, D. Gibson, L. B. Shaw, S. Bowman, S. Bayya, B. Sadowski, M. Hunt, C. Askins, J. Peele, I. D. Aggarwal, and J. S. Sanghera, “Recent advancements in transparent ceramics and crystal fibers for high power lasers,” Proc. SPIE 8733, 87330V (2013). [CrossRef]
9. F. Y. G. I. A. Bonder, “Phase Equilibria in the System Y2O3-Al2O3-SiO2,” Izv. Akad. Nauk SSSR, Ser. Khim. 7, 1325–1326 (pp. 1231–1322 in English translation) (1964).
10. U. Kolitsch, H. Seifert, and F. Aldinger, “Phase relationships in the systems RE2O3-Al2O3-SiO2 (RE = rare earth element, Y, and Sc),” J. Phase Equilibria 19(5), 426–433 (1998). [CrossRef]
11. C. C. Lai, K.-Y. Huang, H.-J. Tsai, K.-Y. Hsu, S.-K. Liu, C.-T. Cheng, K.-D. Ji, C.-P. Ke, S.-R. Lin, and S.-L. Huang, “Yb3+:YAG silica fiber laser,” Opt. Lett. 34(15), 2357–2359 (2009). [CrossRef] [PubMed]
12. L. Chien-Chih, L. Yen-Sheng, H. Kuang-Yao, and H. Sheng-Lung, “Study on the core/cladding interface in Cr:YAG double-clad crystal fibers grown by the codrawing laser-heated pedestal growth method,” J. Appl. Phys. 108, 054308 (2010).
13. J. Ballato, T. Hawkins, P. Foy, B. Kokuoz, R. Stolen, C. McMillen, M. Daw, Z. Su, T. M. 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]
14. C. Y. Lo, K. Y. Huang, J. C. Chen, C. Y. Chuang, C. C. Lai, S. L. Huang, Y. S. Lin, and P. S. Yeh, “Double-clad Cr4+:YAG crystal fiber amplifier,” Opt. Lett. 30(2), 129–131 (2005). [CrossRef] [PubMed]