Laser-heated Pedestal Growth
The technique that we have used to grow SC fibers is the LHPG method. This crucibleless technique closely resembles the float-zone method of crystal growth. In the float-zone method, the molten zone is freely supported between the two ends of the crystal rod. LHPG is inherently the best technique for growing high optical-quality SC fibers because the molten zone is held in place by surface tension, eliminating the need for a crucible, which could be a possible source of contamination. Furthermore, a CO2 laser beam, which provides a uniform, ultra-clean heat source, is used to melt the starting rod. Unlike the EFG technique, this growth method allows only one fiber to be grown at a time, so commercialization is difficult. LHPG, however, is the method used to produce the lowest loss SC fibers.
In LHPG SC fiber growth, a CO2 laser beam is focused onto the tip of a source rod creating a small molten bead of the oxide crystal source rod. A seed fiber is dipped into the molten region, shown schematically in Fig. 1, and slowly pulled upward forming the single-crystal fiber. The source rod, which may be single crystal, polycrystalline, sintered, or a pressed powder, is simultaneously fed upward to replenish the supply of molten material. The shape of the molten zone is a function of the laser power, the diameter reduction, and the material being grown. In general, the length of the maximum stable zone is approximately 3 times the fiber diameter. It is much more difficult to produce SC fibers with smooth surfaces than it is for a glass fiber. This is because the viscosity of a glass is very high during drawing whereas the viscosity of crystalline material at the molten zone is very low and, thus, sensitive to any minor perturbations in the system. Since the molten region is held in place simply by surface tension, any air currents, vibrations, laser-power fluctuations, etc. will have enormous effects on the stability of growth. For this reason it is necessary to use a very stable laser source; have a small molten zone; and grow with a source-to-fiber reduction ratio of 3:1.
Figure 1 – Schematic of source rod, molten zone, and fiber in LHPG.
Our LHPG apparatus, shown in Fig. 2 (a), was built to grow fibers 1 m or greater in length. The mechanical drive mechanisms involved the use of dc-motordriven belt drives to smoothly translate the source rod and seed fibers. A precision V groove in a hard anodized metal block provided the low-friction guiding surface for the SC source and fiber. The groove angle was 90°, and the depth depended on the diameter of the fiber or source rod. We found that smaller groove depths—so that slightly more than half the diameter of the fiber or source protruded beyond the surface of the block—provided smoother translation. We also found that using a stainless steel guide tube, with a bore slightly larger than the fiber diameter and mounted on the end of the fiber translator, helped to eliminate side-to-side motion. The diameter of the fiber was measured in line by a LaserMike, Inc. laser micrometer, with a resolution of <0.1 µm, which was fitted with a notch filter so that the detector was not saturated by the intense white light from the molten SC material. The measurements from the laser micrometer were fed to the computer and used in a feedback loop to precisely control the fiber diameter during growth. The laser source used was a Coherent Model 42 flowing-gas CO2 laser. This laser could deliver as much as 50 W of near TEM00 power. The 10-mm-diameter output beam was expanded to ~16 mm and directed into the reflaxicon optical arrangement shown schematically in Fig. 2 (b). Normally the power stability of this laser is ±5% with temperature-controlled cooling water. Because this is too large an amplitude fluctuation for fiber growth, we added a feedback system to control electronically the laser power. Using this feedback system, we were able to stabilize the laser power to ±0.5%. Typically, 10 W was needed to melt our 1- mm-diameter source rods, and the power fluctuation was less than ±0.05 W. Finally, the entire system was sealed within a Plexiglas enclosure to block out any air currents that could disrupt the growth process and to allow for growth in different atmospheres.
Figure 2 – Schematics of (a) our continuous-feed LHPG apparatus and (b) our reflaxicon
optical arrangement. The source rod (as much as 2 m long) is fed through a hole drilled
in the table, and the entire system is sealed within an airtight Plexiglas enclosure.
The optical system used to focus a ring of laser light onto the source rod employed the following elements: a reflaxicon, a turning flat, and a parabolic mirror. The second element of the reflaxicon was made out of a ZnSe window with a Ag-coated, diamond-turned cone in the center. By using a window and reflecting cone, we eliminated the need for supporting spokes for the cone and thus our laser beam was not obscured. The entire optical system could produce a calculated spot size as small as 22 mm. In practice, however, we adjusted the spacing between the reflaxicon elements so that we could produce a larger focused spot that led to more stable growth from the large 1-mm-diameter source rods.
The control of fiber diameter was crucial to the growth of uniform, low-loss fiber. Our LHPG apparatus was controlled by a Labview program that simultaneously monitored and controlled the fiber diameter and laser power, each to more than ±0.5% stability. All control and process variables were recorded automatically during growth, and the program could also automatically stop fiber growth if necessary. At the typical sapphire fiber growth rate of 2 mm/min, a 1-m long sapphire fiber took nearly 8.5 h to grow. The longest sapphire fiber was 5 m, which took nearly 40 h to grow. For YAG SC fibers, the typical growth rate is about 1 mm/min. The seed fiber was either a sapphire or YAG fiber or a Pt wire. For this work the final fiber diameters were about 400 µm when a 1-mm source was used. Smaller diameters can be obtained by regrowing these 400 µm fibers. The longest length of YAG fiber grown was 60 cm and all fibers were core only, i.e. unclad. Fig. 3 shows the dramatic improvement to diameter stability for a sapphire fiber as a result of the computer-controlled feedback. Without the feedback, diameter fluctuations were ~±5% (i.e., ±15 µm for a 300-µm diameter fiber); a factor of 10 higher than achieved with the feedback. We were able to grow 300-µm diameter fibers with peak-to-peak diameter fluctuations less than 2 µm.
Figure 3 – Diameter fluctuations in LHPG sapphire fibers with and without active feedback control.