Infrared Fiber Technology: Its Promise and Status

Optical fibers with attenuations much lower than those attainable in present-day silicate waveguides are possible in theory if mid-infrared operating wavelengths are used. The loss mechanisms, materials, and problems inherent in obtaining such fibers are discussed. Among several candidates, heavy metal fluoride glasses are suggested as being must suited to low and moderate loss fiber applications in the 2-5 micrometer spectral region.


INTRODUCTION
The successful development of optical fibers based on silicate glasses has recently prompted interest in the extension of this technology to longer wavelengths in the infrared. A variety of communications and non-communications applications have been envisioned for such fibers, many based on the ultra-low optical absorption (perhaps 0.01 to 0.001 dB/km) which may theoretically be acheived in certain non-silicate materials. Very low attenuations offer the prospect of repeaterless transoceanic or transcontinental fiber links, as well as waveguides for infrared power transmission, IR imaging, and sensing (e.g., remote pyrometry). Although efforts in this area are still very much in the research and development stage, a number of appropriate fiber materials have been identified and are under active examination at laboratories worldwide. This paper endeavors to briefly assess the state of the art in IR fiber technology, and examines the mechanisms, materials, and problems inherent in achi eving long wavelength lightguides.

LOSS MECHANISMS
Fibers which operate at wavelengths in excess of about 2 micrometers could in principle exhibit very low optical loss Although the atoms in a glass are considered to be randomly arranged, localized microscopic fluctuations in the refractive index are frozen into the material when it is cooled from a liquid state.
This small-scale "granularity" causes Rayleigh scattering of light. The amount of Rayleigh scattering decreases with the reciprocal fourth power of the wavelength and can reach very low values in the mid-infrared ( Fig.1 Schematic illustration of the three intrinsic loss-inducing mechanisms which govern the total attenuation observed in a transparent solid or optical fiber.
Absorption due to electronic transitions, which is strongest in the ultraviolet for many materials, decreases rapidly with increasing wavelength. Small localized fluctuations in the refractive index of the material lead to Rayleigh scattering of light. The intensity of this effect decreases with the reciprocal fourth power of the wavelength. At very long wavelengths, vibrations of the atoms in the material's structural framework or "lattice" lead to high absorption. Combinations and overtones of these fundamental vibrations give rise to the infrared vibrational or "multiphonon" edge, whose intensity decreases with decreasing wavelength. The conjunction of these three factors leads to an overall attenuation curve which exhibits a "vee" shape.
The wavelength of minimum attenuation is governed by the slopes and separation of the three loss-inducing mechanisms which in turn are a function of materials properties.

3.6.2
The wavelength of minimum attenuation is governed by the slopes and separation of the parts of the curve due to the three loss-inducing mechanisms, which in turn depend on the composition of the material. To obtain very low loss fibers, it is necessary to chose materials with infrared edges at long wavelengths and thus take advantage of the The heavy metal fluoride glasses exhibit a broad range of high transparency which spans the ultraviolet to the midinfrared.
In the vicinity of 2-4 micrometers, they show a minimum in absorption coefficient; fiber losses of less than 10.0 dB/km have been demonstrated and are expected to decrease further.
The current status of infrared optical fiber technology can be examined by comparing the loss levels found in experimental fibers and bulk glasses with the attenuations predicted by theory on the basis of material properties. This has been done in Fig. 2, which summarizes data taken from a variety of sources (3)(4)(5)(6)(7)(8)(9)(10).
The theoretical "vee" curves for heavy metal fluoride glasses, chalcogenide glasses such as arsenic trisulphide, and thallium bromoiodide (KRS-5) crystals (all shown as dashed lines in the bottom of Fig. 2) suggest that losses of 0.01 to 0.001 dB/km could in principle be obtained in the 3 to 7 micrometer region (4,5,9). Only with silicate glasses, however, have fibers with experimental attenuations (solid line) near the theoretical limit of about 0.2 dB/km been prepared (9). This is due in part to the availibility of chemical vapor deposition techniques for silicate fibers, which allow preparation of very high purity glass by precipitation and sintering of silicate "soots" from gas phase reactions. Data obtained from bulk silica glass show the fundamental vibrational absorption in the 8-15 micrometer region decreasing rapidly along the infrared edge to meet with fiber data at shorter wavelengths (10).
The best heavy metal fluoride fibers, prepared by workers at the U.S. Naval Research Laboratory, have losses of 4 dB/km near 2.5 micrometers.
In the example shown in Fig.  2 (dashed line), a minimum attenuation of 8.5 dB/km was observed (7). The disparity between these results and the theoretical prediction is due to extrinsic absorptions from transition metal, rare earth and hydroxyl impurities in the starting materials, and to non-Rayleigh scattering from crystallites and other defects. As was the Case with silicates, the infrared edge data (8) suggests that improvements could be made through materials purification. Despite the presence of hydrogen impurities, arsenic trisulphide chalcogenide glass fibers with losses of about 35 dB/km near 2.4 micrometers have been fabricated (dotted line); their long wavelength behavior shows good agreement with theory. It has recently been suggested, however, that a weak intrinsic electronic absorption tail which extends into the infrared may limit the minimum attainable attenuation in chalcogenide fibers to about 10 dB/km (6).  I  I  I  I  I  I  I  I  I  I  I  I   2  4  6  8  10  12  14  16 18 20 22 WAVELENGTH (micrometers) FIGURE 2: Attenuation versus wavelength for a number of prospective infrared transmitting fiber materials, based on data contained in Refs. 3-10. The theoretical intrinsic attenuation attainable in each material is indicated by the dashed "vee" curves at the bottom of the figure.
Above them, experimental data for fibers and bulk glasses are shown. These data represent the "best" values reported in the literature for a given material as of June, 1985. With the exception of silica-based fibers, the loss levels in all the materials are far above the intrinsic limit. In large part, this is due to the presence of extrinsic impurities and non-Rayleigh scattering effects. 3.6.4 transmission; their range of transparency, however, exceeds that of the other candidates.
It is clear from Fig. 2 that heavy metal fluoride glasses have exhibited the lowest experimental attenuations at wavelengths beyond 2 micrometers. Despite their promise, these materials are still in a fundamental stage of development and a considerable amount of basic and applied research remains to be done. Problems include glass stability, the preparation of large fiber preforms and long fiber lengths, light scattering behavior, chemical durability and the reduction of impurities such as transition metals, rare earths, and hydroxyl and oxide species. The wavelength position and absorptivities of various metallic species, for example, have been determined and it is now recognized that such contaminants must be reduced to the parts-per-billion level. Successful efforts to reduce hydroxyl and oxide impurities have focused on high temperature glass processing or the use of reactive atmospheres such as carbon tetrachl ori de.
Fibers and bulk glasses have been prepared whose Rayleigh scattering behavior closely follows the predictions shown in Fig.   2. While current mid-IR loss levels in the fluoride glasses are adequate for many short length applications, the progression towards ultra-low optical absorption is presenting a considerable challenge to the applied sciences.