High Temperature Sapphire Fiber Cladding

Contact: Stephen C. Bates
Thoughtventions Unlimited, LLC
40 Nutmeg Lane
Glastonbury, CT 06033


Sapphire fibers are well suited to the task of sensing harsh environments, tolerating temperatures well above 1500°C and combustion environments. High quality sapphire fibers transmit light without losses from the UV to 6 m m wavelength; ideal for transmission of laser sensing or spectroscopy. A manufacturing process for producing a high-temperature cladding for sapphire fiber sensors is being developed. Previous research at TvU has demonstrated an adherent, thin, and complete cladding that is stable and inert at high temperatures in oxidizing atmospheres. Work will produce clad fibers, measure their transmission, and demonstrate the feasibility of a cladding manufacturing process as well as the performance of the clad fibers. Equipment will be constructed and tested that will allow the practical cladding of sapphire fibers. The optical and mechanical properties of the produced clad fibers will be tested and characterized, and the fibers will be tested for response to combustion environments.


Sapphire fiber cladding that can tolerate high temperatures and harsh chemical environments will lead to a variety of novel commercial sensors for these common industrial environments. Such sensors will provide cheaper and higher quality products through improved process control.

The Problem. Novel technologies, are being sought that will enable development of low-loss, durable infrared fiber optics suitable for use in fire and other combustion environments. Sapphire fibers are well suited to the task of sensing harsh environments, tolerating temperatures well above 1400°C. They are also virtually immune to oxidation or chemical attack. High quality sapphire fibers transmit light without losses from the UV to 6 µ m wavelength. Very high quality sapphire fibers have been produced, but no means has yet been developed to optically isolate these fibers in an industrial high temperature environment. Without a robust protective cladding, the light carried by the fiber can escape out the fiber sides as a result of deposits on the fiber, causing unpredictable light transmission through the fiber. A method and material must be developed to clad sapphire fibers for high temperature operation so that these fibers can be extensively for combustion sensing.

The Innovation. The innovation of this proposal is the development of a stable, high temperature sapphire fiber cladding. Platinum is coated onto the fiber and bonded to the sapphire using a proprietary manufacturing technique. Research at Thoughtventions (TvU) has shown that proper processing of a treated sapphire fiber forms a stable high temperature platinum coating that can be used in oxidizing atmospheres. Furthermore the coating is strongly bound to the sapphire and the binding layer is extremely thin and is appropriate for use as a fiber cladding material at temperatures up to 1450°C. An efficient coating technique has been developed to uniformly coat the fibers with platinum and to conserve the platinum material itself to keep the cladding cost low.

Effort Synopsis. Previous work at TvU has demonstrated that high quality sapphire fiber cladding can be achieved technically. Current work will develop a reliable manufacturing technique to produce high quality clad fibers. The optical and environmental performance of the clad fibers will be measured to contribute to the overall feasibility demonstration. Equipment will be constructed and tested that will allow the continuous, practical coating of sapphire fibers. This equipment will be used to coat fibers for further processing and testing. The cladding heat treatment process used to achieve adhesive bonding will be further developed and used to treat coated fibers. The optical and mechanical properties of the produced clad fibers will then be tested and characterized, and the fibers tested for response to combustion environments. In particular, the upper temperature use limits for the clad fibers will be determined. The data gathered will be used to define the specifications of the clad fiber to provide the applications envelope as well as marketing and sales tools for commercializing clad sapphire fibers.


Sapphire Basics. Sapphire [1] is the hardest material and has the highest melting point (2040°C) of any transparent material that is commonly available. Single crystal sapphire is a widely used material as a result of its transparency, its superior mechanical properties, its chemical and scratch resistance, and the fact that it can be relatively easily manufactured as a grown crystal.

Chemical Properties. Chemically, sapphire is aluminum oxide, Al2O3. Stoichiometrically there is only one oxide of aluminum: alumina, Al2O3, although alumina can form various polymorphs, and hydrated species. In α-Al2O3 (sapphire, vs. γ-Al2O3, the other anhydrous alumina) the oxide ions form a hexagonal close-packed array where the aluminum ions are distributed symmetrically among the octahedral interstices. α-Al2O3 is stable at high temperatures and indefinitely metastable at low temperatures. The most common Al2O3 - that formed on the surface of aluminum metal - has a defect rock-salt structure with an arrangement of Al and O ions in a rock-salt ordering with every third Al ion missing. Sapphire is chemically inert for all practical purposes at room temperature and is extremely resistant to chemical attack at higher temperatures, and has a density of 4 g/cm3. It has a major symmetry axis, the C-axis, and six minor A-axes. Most properties of sapphire are orientation dependent, including both mechanical and thermal properties.

Optical Properties - Optically, sapphire has high internal transmittance from 150 nm to 6000 nm (6 µ m) in wavelength - from the far UV to the middle infrared. It has a high index of refraction, n, at visible wavelengths (1.77 vs. 1.5 for quartz), which causes higher surface reflection losses. It is birefringent; n depends on both the polarization and direction of the incident light.

Thermal Properties - Sapphire has a thermal response similar to some steels. Its thermal conductivity (k = 0.065 cal/cm-sec-°C) is similar to stainless steel, while its thermal diffusivity is closest to plain carbon steel. It has a specific heat of 0.10 cal/g-°C, and a thermal expansion coefficient, CTE, of 8.4 x 10-6 per °C. These values are typically measured at 25°C, 60° to C-axis. Sapphire maintains its structural integrity up to 1600-1700°C, becoming increasingly plastic until it melts at 2050°C. Its compressive strength drops to low values (below its tensile strength) at temperatures around 800°C. Sapphire is susceptible to thermal stress because it has a k and CTE that are both comparatively large.

Mechanical Properties Of all of the properties of sapphire, those related to its strength are the least well defined because the failure of sapphire is statistical. Mean properties cannot be used for design work; design values must be based on the minimum possible strength. With a tensile strength at 25°C of about 410 MPa, sapphire is a relatively strong optical material. Its properties change significantly with crystal orientation; its maximum bending stress increases 50% from its lowest value as the direction of stress is changed. The properties of a sapphire piece are symmetric around the crystal C-axis. Sapphire is the third hardest material known, with a MOH rating of 9.

Sapphire Fibers. Sapphire fibers have been commercially available for over 10 years, but high quality optical fibers have only recently become standard products (Micromaterials, Tampa, FL). Standard fibers have outer diameter variations and material impurities that significantly degrade fiber performance. These fibers are more often used for structural reinforcement of composite materials. Laser-heated pedestal growth fibers have low loss (1 dB/m or less) and the full transmission bandwidth of sapphire. These fibers are expensive relative to standard fiber because sapphire fibers must be made using a much slower crystal growth process, rather than high speed pulling. Short sensor fibers are not expensive, however.

Single crystal sapphire fibers used in this program were prepared at USF using the laser-heated pedestal growth method (LHPG) [2],[3]. The fibers are grown by dipping an oriented single-crystal seed into a molten droplet produced above a feed rod by laser heating. By carefully controlling the ratio of the speeds at which the source rod is pushed into the molten zone and the fiber is pulled out, a reduction ratio of source rod to fiber diameter of 3-4 is typically obtained. The sapphire fibers are grown in air at a speed of about 5 mm/min. The cross section of the c-axis fibers is roughly circular with slight deviations reflecting the trigonal symmetry. Longer fibers are grown by using a two-step reduction. A fiber grown from an approximately 1 mm diameter source rod is used as the source material to grow 100 to 150 µ m diameter fibers as long as 3 m. The fibers grown are unclad (core index = 1.78) and are therefore highly multimode. Despite the large numerical aperture, the measured modal power distribution after propagating a laser beam through a 0.7 m long fiber has a full angle at half intensity of only 11° [4], only weakly sensitive to bends in the fiber and input launching conditions.

Low propagation loss is necessary for optical applications of sapphire fibers. Scattering and absorption are often contributors to the transmission losses in a sapphire fiber. Absorption can arise from point defects in the crystal and from impurities. Extremely pure starting materials, an argon gas jacket during growth, and lack of solid contact with the liquid zone during the growth process eliminates fiber impurities. Inhomogeneities inside the fibers are not detected with optical microscopy, and scattering centers are not apparent under laser illumination.

Diameter variations are the most important cause of the scattering loss in these fibers. Heating power fluctuations and variations in the diameter of the source rod or the feed rate affect the molten zone volume which in turn affects the diameter of the fiber. Diameter variations typically have a maximum at a spatial period of 5 mm for 110 µ m diameter fibers, decrease rapidly for shorter periods. The short period variations efficiently couple power to higher order modes and ultimately lead to scattering loss [5]. The rms diameter variation is 2-3% for spatial periods shorter than 15 mm, primarily resulting from laser power fluctuations. The attenuation loss of these fibers were typically 1 dB/m or less.

Platinum. Pt is known as a "noble" metal as a result of its apparent unreactivity with most materials. Instruments and the chemical industry account for most of the consumption of this expensive metal whose high cost results from its relative scarcity. Pt has a close packed face-centered cubic structure with a density of 21.45 g/cm3. Pure Pt, although soft, has strong interatomic bonds that lead to high thermal parameters; Pt melts at 1773.5°C and boils at 4300°C. It has a specific heat of 0.0314 cal/g-°C, a thermal conductivity of 0.1745 cal/cm2/cm/s/°C, and a coefficient of linear expansion of 9.1 x 10-6 cm/cm/°C in the range 0-100°C. Some optical properties of Pt are given in Table 1 [6], where alloying changes emissivity in rough proportion to conductivity.

Complexity in the properties of Pt arises from the fact that most Pt is not pure, and that the impurities control many of its macroscopic properties. The hardness, strength, and recrystallization temperature of platinum depend strongly on the type and concentration of impurities. High purity Pt recrystallizes at temperatures as low as 200°C, whereas commercial Pt recrystallizes at roughly 600°C [6]. Fine grained (200-400 microns) microstructure occurs if cold deformed Pt is recrystallized at 600°C. Pt is typically annealed for 2 hours at 1400°C to recrystallize and stabilize its microstructure [7]. The behavior of Pt is sensitive to recrystallization because the primary route of chemical attack on the metal is through its grain boundaries. The strength of Pt decreases by a factor of 10 between 20°C and 1250°C.

The chemistry of Pt is complex. Pt is inert to most chemicals, but it is widely used to catalyze chemical reactions. At elevated temperature Pt alloys with many metals and must be used with care. For instance, there are numerous alloys of Pt and Al exist at 1000°C [8], and phosphorus, lead, silicon, and boron form low melting point eutectics with Pt [7]. Pt5P2 forms at 6 weight % P and melts at 590°C [7]. Iron is known to cause intercrystalline oxidation in platinum, reducing its hot strength [7].

For this program the important facet of Pt chemistry is its reaction with oxygen. In practice Pt only begins to react significantly with O2 at temperatures above 1400°C, which is the basic reason Pt is so widely used as a material and a thermocouple material. This apparent inertness to oxididative attack is misleading, because the surface of Pt is protected like aluminum with an oxide layer. There are three known oxides of Pt: PtO, a violet black compound that dissociates at 550°C; PtO2, the most common oxide of Pt, a black compound with a melting point of 430°C, and PtO3, a red brown oxide less thermodynamically favored. The most stable oxide of platinum is PtO2 [9]. Platinum oxide (PtO2) has several crystallographic forms [10]. It is a volatile oxide at high temperatures, and is responsible for most of the mass loss of Pt at high temperatures in air at temperatures above 1400°C, where Pt becomes permeable to O2. It should also be noted that at temperatures above 700°C Pt is highly permeable to H2.

Fiber Optics. At this time fiber optics is a massive industry as a result of the replacement of copper telephone lines with silica fiber optic light guides. The advantages of fiber optics in terms of signals/cable and high transmission rate at competitive cost are so large that optical communication is rapidly replacing conventional systems. There is thus a large literature and commercial base in silica fibers. Fiber optic sensing has become an important field, [11], [12] because light-based sensors are immune from the electrical interferences that is usually the bane of conventional high-sensitivity sensors.

Optical fibers contain light within a thin, very long cylinder, functioning as an optical wave guide.

Most fibers (NOT sapphire) are made by pulling a cylinder through a die to successively smaller diameters. Silica fibers have long been used for image bundles and laser power delivery, but the use of nonsilica fibers is more recent. The scientific, medical, and engineering aspects of fibers made from crystalline, UV and IR glass, plastic, and hollow materials has become important. This includes IR fibers transmitting from 2 to 20 µ m, fibers that are useful in transmitting UV wavelengths shorter than 350 nm, and plastic fibers. Fabrication, characterization and applications of these specialty fibers are areas of expanding work.




















Temperature(°C)........Total Emission






Table 1. Optical properties of platinum.

Fiber Cladding. Almost all fibers in a fiber optic system must be clad, or coated, in some way. Not only must the light be kept in, but the environment must be kept out to maintain fiber integrity. Cladding is also used to increase fiber strength. Common optical fibers are clad in a lower index of refraction transparent material. Most cladding is plastic and used at low temperatures. For a glass fiber this is relatively easy process, since other glasses are available with the same coefficient of thermal expansion and the proper index of refraction.

Fiber Coating Process

This discussion begins with a description of past research on sapphire fiber cladding, concentrating on work previously performed at TvU [13]. This discussion is meant to form a background for TvU's work rather than a comprehensive review of all of the work that has been done in the field. A cladding method for sapphire fiber optics at 1000°C or hotter, does not at present exist. Such a cladding is necessary to prevent light loss if the fiber comes in contact with some light absorbing medium of sufficiently large index of refraction.

The essence of the sapphire fiber cladding problem is that an appropriate transparent cladding material does not exist at first inspection. Any practical cladding material must have an index of refraction less than that of sapphire (allowing total internal reflection at the fiber/clad boundary), while simultaneously possessing a compatible coefficient of thermal expansion (CTE), adequate adhesion to the fiber, and a tolerance of temperatures over 1000°C. Glass coatings fail because they have a much lower CTE than sapphire and are subject to chemical attack at high temperatures. Achieving adhesion to the sapphire is also a major problem; since sapphire is such an inert material it resists chemical bonding. Very reactive materials do bond with sapphire at high temperature, but they also damage the fiber. A lack of adhesion during the coating process causes coatings to not cover the fiber completely; gaps appear during processing or use at high temperature.

Cladding Research. The development of a high temperature sapphire fiber cladding has been pursued by a number of groups. Transparent sapphire fiber claddings have been attempted in previous research. USF (University of South Florida) and researchers in China [14] have tested a polycrystalline alumina deposition-coated onto sapphire fibers. A smooth, continuous coating was achieved, but was too thin to provide internal reflection at visible wavelengths of light. USF attempted to use a sol-gel process to coat the fibers. If a sufficiently thick alumina coating could be obtained, it would have a sufficiently different index of refraction from the single crystal sapphire to form an effective cladding. The USF experiments found that the alumina coating was only 0.5 µ m thick; 10 successive coatings were required to build up a 5 µ m layer [15], and this is commercially impractical. The Chinese research had similar problems. Procedures that give a much thicker single coating of alumina also appear impractical.

A different technique was tried at NASA, where a single-crystal alumina fiber was coated with the refractory oxide zirconia. For this process the fiber was first passed through a slurry of zirconium dioxide particles suspended in the molten vehicle, guanidine 2-ethyl hexanoate. After the fiber was removed from the melt, it was fired at a temperature of 1,500°C to burn organic material and densify the zirconia coat. Incomplete overall coverage was obtained.

Another approach to fiber cladding is to add a metal coating to the fiber. The metal cladding on the fiber keeps the light in the fiber by mirror reflection, rather than total internal reflection. This technique has the weakness that the best metal mirrors have a few percent absorption loss after each reflection, so there will be losses in the fiber as a result of the many bounces of the light passing down the fiber channel. The target application is relatively short fiber sensors, where some light loss in the fiber is acceptable. Actual losses in a sensor fiber have been shown at TvU to be tolerable as is discussed below. The phenomenon of the propagation of light down a fiber is a complex interaction of the electromagnetic radiation in the fiber material with a boundary formed by the cladding. Glancing incidence reflection losses can be much lower than normal incidence reflection losses.

A summary of platinum metallizing techniques is given by Baumgartner and Raub [16]: "Thin layers of metallic platinum can be produced in a variety of ways: by electrolytic deposition from aqueous- electrolytes or salt melts, by thermal evaporation or cathodic sputtering (the physical vapor deposition or PVD processes) or by firing printed pastes. Depending on the deposition process and the parameters involved, the properties of the coating can vary considerably. Those deposited from aqueous electrolytes at thicknesses of a few microns or less are highly stressed and porous, although bright, hard and wear resistant. Deposits from molten salts are generally more than 1 µ m thick; they are low stressed and have low porosity, being semibright in appearance, soft and highly ductile. Layers produced by the PVD processes are generally in the thickness range of 0.1 µ m or less, and therefore rather porous. Fired ceramic pastes often form thick layers with internal porosity." Work at TvU confirms these observations.

A recent NASA program created one type of high temperature Pt sapphire fiber coating process [17]. Soaps made from the strong organic base guanidine [(NH2)2C:NH] and organic fatty acids were used as vehicles and binders for coating ceramic fibers with thin precious-metal or metal-oxide films. This coating was intended to act as a barrier to diffusion in fiber matrix ceramic composite materials. A layer of Pt 3 µ m thick was applied to a sapphire fiber 0.13 mm in diameter by dipping the fiber and firing it at a temperature of 1200°C to burn off the organic materials, leaving pure Pt without residues on the fiber. The Pt coating had numerous gaps in the coating over the entire fiber as a result of poor adhesion to the fiber, making this coating process inappropriate for cladding of sapphire fibers.

Platinum Cladding: Work at TvU has focused on developing techniques for cladding a sapphire fiber with an adherent, complete layer of Pt. Work has focussed on a Pt cladding as a result of its excellent high temperature behavior, and the existence of a strong, high temperature, thin, interface bond between Pt and sapphire. Further research at TvU has demonstrated evidence of higher reflectivity of the Pt/alumina bond layer, which makes this cladding even more attractive.

Pt/Sapphire Bonding. Dr. Bates' research into bonding a variety of materials to sapphire has been in progress since 1985, primarily attempting to create a durable high temperature sapphire/metal seal. Initially the application was to seal a sapphire internal combustion engine liner [18]. More recently the goal has been a transparent furnace seal and high temperature, partially reflective coatings [19]. Sealing temperatures of at least 1200°C have been sought, with mixed success. A search for possible joining materials found a complicated and expensive niobium seal used for fabricating fuel cells, a glass frit used for high temperature lamps, and noble metal foil seals. Only the noble metal foils are able to absorb the thermal stresses in an alumina/metal joint. Dr. Bates has performed extensive research into the conditions and mechanism of this noble metal bonding to sapphire, centering on Pt as a result of its excellent high temperature performance under oxidizing conditions.

Although the details are proprietary to TvU, the salient feature of the research at TvU is that there is a strong bond formed between Pt and alumina formed at temperatures around 1500°C. Furthermore, this bond is stable under oxidizing conditions. These characteristics make Pt the material of choice for cladding sapphire fibers in this program.

Cladding Research at TvU. A goal of work at TvU has been to create a platinum coating on sapphire optical fibers. The target coating thickness is in the range of 1-5° m thick. A highly uniform coating is desirable, but not necessary; only a complete, opaque coating is necessary. A variety of techniques have been investigated and are discussed below.

Cladding Facilities. Two processing facilities were used in previous small-scale cladding research and are available for this program. A small, insulated SiC-element furnace with central access allows continuous or batch fiber processing. This furnace was operated reliably and repeatably in multiple experiments at over 1500°C. Fibers can be inserted and removed from the hot furnace for inspection during operation.

Cladding Techniques. Initial cladding efforts at TvU attempted simply to form a complete platinum coating on a sapphire fiber, but this work was unsuccessful as a result of the difficulty of achieving adhesion of the Pt to the sapphire. The generic possible routes to clad sapphire fibers include chemical, physical, and mechanical deposition. Chemical deposition efforts involved coating the fiber with a chemical then reacting it with various atmospheres to create a metallic coating. Physical deposition efforts included evaporation and sputtering. Mechanical deposition efforts consisted of wrapping the fiber in foil and brazing the foil at high temperature for a monolithic encapsulation. High quality sapphire fibers, 15 cm in length and 180 µ m in diameter were supplied by USF for this work. The fibers were usually cleaned in an ultrasonic bath followed by dipping in a 10% HF solution. Clad fibers were not usually cleaned, due to the weak cladding adherence. The fiber cladding experiments were ultimately successful a practical and proprietary cladding technique has been developed.

Coating Adhesion to Sapphire. Adhesion to sapphire surfaces is the key to the achieving a successful fiber cladding. Research at TvU has shown that straightforward techniques used to apply a coating to the sapphire test fibers did not result in coatings that were adherent, and furthermore that simple heat treatment or high temperature coating processes are also not effective. Sputter coating is known for its adhesion, but the sputter coating on test fibers simply flaked off on light contact. Foil wrapped around a sapphire fiber did not become adherent when heated to the temperatures that would normally cause the ceramic-metal bonding as discussed above. Chemical and liquid coatings at low temperature did not coat uniformly; numerous gaps occurred in the coating.

Adhesion to sapphire is normally achieved by adding an interface material that reacts with the sapphire at elevated temperature. Titanium is the standard example. The materials used are extremely reactive, and the reaction depth is typically substantial - on the order of the radius of the fibers used here. For this reason, standard frits and braze preparations used for alumina are inappropriate for creating fiber claddings. It does seem that the lack of adhesion to sapphire has been a generic problem that has prevented successful sapphire cladding efforts in the past. Adhesion of the pure Pt fiber coatings was ultimately achieved by heating them to high temperatures at atmospheric pressure, but only if the Pt was in intimate contact with the sapphire fiber as a result of the coating process.

Adhesion Testing. Cladding mechanical testing centered on testing the adhesion of the cladding to the sapphire fiber. There is no accepted quantitative test of fiber cladding [23], although the tape test is usually used. A successful test occurs when scotch tape is pressed against the coating and peeled off without carrying away any coating. All of the attempted claddings passed the tape test, except for the sputter coatings before they were heat-treated. The next level of adhesion testing included (in order from most mild to severe) washing, skin and paper rubbing, finger nail rubbing, abrasion with a dull copper blade, a sharp copper blade, and, finally, a sharp steel blade. The cladding of the heat treated sputter and ink coated fibers showed resistance to distortion by the sharp copper blade, and were unaffected by the less severe tests. They also showed resistance to scraping by the steel blade; the platinum itself was soft enough to scrape off, as was the case with the foil, but the Pt did not separate from the fiber. The final Pt coating that was placed in the flame could not be removed with a steel blade.

Clad Fiber Testing. The performance of the clad fibers must be experimentally mapped in terms of their chemical/environmental, physical, mechanical, thermal, and optical properties. Testing the high temperature behavior/performance of the clad fibers will be emphasized. The clad fibers must demonstrate adhesion, completeness and opacity of the coating, and optical transmission at high temperature. Extensive testing of bare and clad sapphire fibers has already been performed at TvU.

Chemical/Environmental Testing. Both sapphire and Pt have been thoroughly tested for chemical resistance, both are inert (except Pt catalytically) to almost all chemicals except at high temperatures. In the context of a Pt-clad fiber the major performance questions concern high temperature exposure to gaseous species, and degradation of the Pt/sapphire interface. A concern is the exposure to metal vapors at high temperatures in reducing environments. Basic testing of the clad fiber response to oxidizing, reducing atmospheres at 1000°C will be done.

The clad fibers will be tested for their response to various cleaning methods. Cleaning by standard solvent liquids such as soap and water, alcohol, acetone, toluene, HCl, and HF will be tested to determine whether these procedures degrade the cladding. Ultrasonic cleaning and standard abrasive methods such as brushes will be tested for any attack or encroachment on the fiber/cladding interface.

Physical Testing. Two primary techniques have been and will be used to examine the bare and clad fibers. A variable magnification stereo microscope (up to 60X) will be used to study the color, transparency and gross morphology of the coating. SEM imaging will be used to examine the microscopic morphology of the surface. Previous cladding attempts have not had obvious results. Often extremely thin coatings were obtained that were almost transparent. The curvature and high index of refraction of the fiber make it difficult to distinguish between metal coatings and uncoated surfaces.

The physical specifications of the clad fiber that will be measured include fiber diameter, cladding thickness, and variation of these parameters along the fiber length. Also important are the properties of the fiber ends, including flatness, optical polish, perpendicularity of the fiber faces with respect to the fiber axis, and edge imperfections in the end faces. The fibers will be inspected in TvU's SEM, which will also measure the uniformity, thickness, and grain size. Deformation of the coating as a result of testing will also be recorded by SEM imaging if appropriate.

Mechanical Testing. Goals of mechanical testing include cladding adhesion, cladding hardness, clad fiber strength, and resistance to abrasion and erosion. Mechanical testing will consist primarily of adhesion testing. Tape pull-off tests will be done. Scrape tests with progressively harder materials (plastic, copper, steel) will be done. The scrape tests will measure hardness of the cladding in terms of whether the coating can be pierced, and they will measure adhesion by how much force is required to pull the edge along the sapphire surface of the fiber. There are no standard tests for adhesion of fiber cladding; current practice will be investigated further. The approximate strength of the clad fibers will be measured to compare with bare fiber strength. Tensile strength will be measured by suspending weights from a clamped, vertical fiber. Bending strength will be measured by resting (not clamping) a horizontal fiber between two flat supports and supporting a weight from the center. Minimum bending radius of the clad fiber will be measured.

Thermal Testing. Thermal testing will consist of testing the tolerance of the clad fiber to isothermal environments, temperature cycling, temperature gradients, thermal shock, and insertion into a flame. All of these measurements will initially be done as part of testing to determine the cladding adhesion heat treatment process.

Previous experiments and research has shown that Pt metal undergoes an environmental weight loss as a result of evaporation of Pt oxide at temperatures above 1500°C, so this will be the testing limit for long term isothermal testing. Isothermal testing at 100°C intervals up to 1500°C will be done and the clad fibers inspected for degradation.

Temperature gradients and temperature cycling tests will be done by operating the fiber furnace at a constant temperature while the clad fibers are inserted and removed multiple times. The fibers will be inspected for degradation as the tests proceed.

Thermal shock testing will be done using progressively more harsh quenching measures. Insertion into a furnace followed by fast removal to room air is the mildest test, and has already been done many times. The next and most severe test will be water quenching. Detecting cracking and spalling-off of the coating will be emphasized, together with clues to the cause of this behavior if it does occur.

Hydrocarbon flame testing will be an important part of clad fiber thermal testing because this is such an important application area. By changing the type of hydrocarbon, the dilution, and the oxygen content of the oxidizer, flame temperature can be changed from 1000°C (subatmospheric pressure) to 3000°C. Pt evaporates progressively more rapidly above 1500°C and melts at 1770°C, limiting the performance of unalloyed Pt. The actual temperature of the fiber in the flame is actually significantly less than the flame temperature as a result of radiation cooling. Tests for cladding degradation in progressively hotter flames will be done, together with calculations and measurements of the actual temperature of the fiber. Pt alloy thermocouples can probably be used to simulate the radiation cooling of the fiber, but adjustments will have to be made for a larger diameter and a somewhat different emissivity.

Previous Bare-Fiber Thermal Testing. Extensive testing has already been done on USF sapphire fibers at TvU, showing them to be unaffected by static temperatures below 1500°C in air, by axial temperature gradients on the order of 1000°C/cm, and by thermal cycling. A fiber was inserted into the 1000°C air furnace to test its physical integrity. It was inserted multiple times for periods between 1 and 10 sec at various depths. No change was observed in the fiber. The fibers were thus shown to survive a 1000°C environment, the large temperature gradients associated with penetrating the walls protecting this type of environment, and the thermal cycling of furnace operation or fiber insertion.

Temperature gradients in the fiber optics are strictly axial as a result of the small diameter of the fibers. This implies that thermal stresses are negligible, since the scale on which they occur is that of the diameter of the fiber; the high thermal conductivity of sapphire will equalize the temperature across the fiber. The strength of a fiber is also higher than the bulk material because the diameter of the fiber is of the same order of size as that of the flaws that are usually responsible for the failure of large pieces. Similar temperature gradients are involved in the fabrication of the fibers by the LHPG method.

Previous Thermal Cycling Tests. A large number of thermal cycling tests have been done on the fibers. The standard form of testing for the effects of temperature was to heat the fiber furnace to progressively higher temperatures, insert the fiber from room temperature, then withdraw it for inspection. A fiber tested between 700°C and 1300°C, would be inserted and removed every 100°C, so that at least 7 cycles would have been performed during this exploratory test. For thickly coated fibers, no change in the Pt coating was observed as a result of cycling.

Optical Testing. The goal of optical testing will be to determine transmission vs. wavelength, transmission vs. bending radius, transmission vs. input convergence angle, and the extent of light leakage from the sides of the clad fiber. Because previous testing indicates that the reflectivity of the heat-treated Pt/sapphire boundary has a significantly higher reflectivity than that of Pt itself, measurements will be made to quantify this improvement. There may also be some gradual performance increase with the degree of surface bonding achieved. If this were the case, transmission measurements might be a good technique for verifying that the Pt had been properly bonded to the sapphire.

Measurements can be made with a HeNe laser operating at 633 nm, and with a pulsed Er:YAG laser operating at 3 µ m, varying beam convergences into the fiber. A laser is preferred, so that the effect of varying the angle of convergence of the incident beam can be examined. Transmission is generally reduced for large incidence angles, because the beam makes more bounces off the wall before exiting. Since many high temperature applications require collecting light from a large viewing area, this is an important consideration. For measuring the attenuation spectrum of thin fibers, a quartz-tungsten-halogen lamp spectrometer is set up to cover the 250-4000 nm spectral range. Measurements are accurate to a few 1%.

More detailed measurements will be done in for small convergence angles to determine the dependence of transmission in the case of only a few bounces off the fiber boundaries. This will be important in determining clad fiber transmission as a function of fiber length for applications using laser light sensing. Bending transmission tests will also be done. Cladding opacity testing will be done at TvU using index matching fluid. This testing will determine whether there are light leaks in the cladding.

High temperature emission and transmission testing will be performed in future work. The purpose of emission tests is to determine the amount of the thermal emission from the both the fiber and the cladding, together with its effect on fiber transmission. Emission into the environment will determine the fiber operating temperature. Thermal emission into the fiber and out the end of the fiber will determine the optical noise for high temperature fiber diagnostics. The sapphire itself will have emission at wavelengths greater than about 5 µ m, decreasing with temperature. At some temperature above 1500°C sapphire becomes opaque. Transmission testing vs. temperature is important to determine the temperature dependence of the absorption of the cladding.

The physical boundary for the light at the edge of the sapphire itself is very complex and varies with the cladding technique. It is not known what effect the surface roughness of the fiber has on transmission. Standard silica fibers have very small roughness because they are solidified from a melt. Sapphire fibers have small-scale non-uniformities in the crystallization at the outside of the fiber. SEM images indicate that this roughness is at most the wavelength of the HeNe laser and may be less. This roughness is unimportant only if it is much less than this wavelength. For the clad fibers the boundary is a Pt oxide a few lattice dimensions (nm) thick.

Previous Transmission Testing. A series of high quality sapphire fibers about 14 cm long and 180 µ m in diameter produced by USF were tested before and after cladding experiments [13]. Transmission of the bare fibers varied from the mid 70% s to the high 60% s. A high quality 15 cm fiber with low losses (face and internal) would have a transmission of about 80% from the UV to the mid-IR, where losses are a result of polishing and surface (Fresnel) reflection. Initial transmission measurements were done at 633 nm wavelength (HeNe laser) for 40, 67, and 133 mrad input laser beam convergence into the fiber, and were taken before and after the fiber was clad. The reflectivity of Pt is only 66% at 633 nm, increasing at longer wavelengths to 88% at 3 µ m. For this reason fiber transmission was also measured at 3 µ m with a pulsed Er:YAG laser at a full convergence angle of 60 mrad. Bending transmission measurements were done when possible.

Two sputter-coated fibers, one heat treated, one not, were transmission tested. The fibers were 140 mm long and, as a result of processing problems, were only about half covered with Pt, such that detailed results are not given here. Relative to bare fibers, transmission loss after cladding was found to increase much more rapidly with entrance angle, evidence that most of the transmission is axial radiation, rather than bouncing radiation, which is absorbed in the cladding. Transmission at 633 nm through the unheat-treated clad fiber was 23% at 40 mrad, 19% at 67 mrad, 13% at 133 mrad, and 20% at 3 µ m and 60 mrad. The comparable transmission at 633 nm and 3 µ m also indicates that the Pt reflectivity is not the primary controlling factor for transmission, since otherwise transmission would have increased significantly. Transmission through the heat-treated, adherent clad fiber at 633 nm varied from 47% at 40 mrad, 32% at 67 mrad, and 19% at 133 mrad had more than twice the transmission that the fiber that had not been heat-treated. The heat treatment that gives the Pt cladding adherence and provides the Pt oxide bond at the surface, also greatly increase the transmission through the fiber. The details of this effect are not known. The heat treatment temperature was probably too low to affect the fiber surface roughness.

Transmission through a bent clad fiber was greatly decreased relative to a straight fiber, supporting the conclusion that the cladding absorption is significant, and that transmission through the clad fibers occurs primarily as a result of limited numbers of bounces off of the cladding/sapphire boundary. For a straight fiber this also implies that the transmission will not drop rapidly with length. Unclad sapphire fibers in air maintain transmission when bent; a 190 mm long, 180 µ m diameter fiber, measured at 633 nm for a 10 mrad full convergence angle had a transmission drop from 84% to 81% associated with a 1/4 turn of 2.5 cm radius. For the heat treated clad fiber at 3 µ m and 10 mrad full angle convergence, transmission dropped from 19% to 8% for a 1/4 turn bend of 2.5 cm radius. This confirms axial ray transmission, demonstrating a strong effect of the lower reflectivity of Pt. However, the Pt oxide layer seems to greatly improve the transmission of the fiber in all cases.

Aluminum Sputter-coated Fibers. The only other data point that is available for metal-clad fibers is an aluminum sputter-clad a fiber about 20 cm long that had transmission between 50 and 60% as measured at USF during the fiber imaging program [24]. The reflectivity of Al is much higher than Pt, which probably accounts for the difference in transmission of non-adherent coatings.


1. L. M. Belyaev (Editor), "Ruby and Sapphire," Nauka Publishers, Moscow, Available From National Technical Info. Service, Springfield, VA 22161, (1974).

2. M. M. Fejer, G. A. Magel, and R. L. Byer, "Rev. Sci. Instrum., 55, 1791 (1984).

3. G. N. Merberg and J. A. Harrington, "Optical and mechanical properties of single-crystal sapphire optical fibers," Appl. Opt., 32, 18, 3201 (1993).

4. D. H. Jundt, M. M. Fejer, and R. L. Byer, "Characterization of single-crystal sapphire fibers for optical power delivery systems," Appl. Phys. Lett., 55, 21, 2170 (1989).

5. D. Gloge, "Bell Syst. Tech. J., 51, 1767 (1972).

6. E. Savitsky, et. al., Physical Metallurgy of Platinum Metals, Pergamon Press, NY, NY, (1978).

7. B. Fischer, "Reduction of Platinum Corrosion in Molten Glass," Plat. Met. Rev., 36, 1, 14 (1992).

8. M. Charcosset, Rev. Inst. Fr. Pet., 34, 969 (1979).

9. J.C. Chaston, Platinum Met. Rev., 8, 50 (1964).

10. R.W.G. Wychoff, Crystal Structures, Interscience, NY, (1965).

11. E. Udd (Ed.), Fiber Optic Sensors, J. Wiley & Sons, New York, (1991).

12. J. Daikin and B. Culshaw, Optical Fiber Sensors: Principles and Components Vol. 1, Artech House, Boston, MA, (1988).

13. S.C. Bates, "High Temperature Sapphire Fiber Cladding," STTR Final Report, March, AFOSR Contract # F49620-98-C-0063, (1999).

14. Y. Shen, L. Tong, and S. Chen, "Performance stability of the sapphire fiber and cladding under high temperature ," Proc. SPIE, 3852, 134 (1999).

15. R. Nubling, R. Kozodov, and J. Harrington, "Optical properties of clad and unclad sapphire fiber," Proceedings SPIE Vol. 2131, 56, (1994).

16. M. Baumgartner and Ch.J. Raub, "The Corrosion Behavior of Objects Electroplated with Platinum," Plat. Met. Rev., 29, 4, 155 (1985).

17. Platinum Interfacial Coatings for Sapphire/Al2O3 Composites," NASA Tech. Memorandum 106091, (1991).

18. S. C. Bates, "A Transparent Engine for Flow and Combustion Visualization Studies," SAE Paper 880520, (1988).

19. S.C. Bates, "High Temperature Transparent Furnace Development," SBIR Phase II Final Report, NASA Contract NAS3-27664, (1997).

20. H.J. De Bruin, et. al., "Ceramic-metal reaction welding," J. Matl. Sci., 7, 909 (1972).

21. H.J. de Bruin et. al., "Solid-state reaction welding," Patent #4050956, (1977).

22. J.M. Herrmann, "Electronic Effects in Strong Metal-Support Interactions on Titania Deposited Metal Catalysts," J. Catal., 89, 404 (1984).

23. S. Suzuki and et.al., "Measurement of the adhesion strength of metal oxide and metal nitride thin films sputtered onto glass by the direct pull-off method.," J. Adh. Sci. & Techn., 11, 8, 1137 (1998).

24. S.C. Bates and R.F. Chang, "High Temperature Fiber Optic Imaging," Fiber and Integrated Optics, 6, 6, 387 (1997).

25. M. A. Serio, H. Teng, K. S. Knight, and S. C. Bates, "Insitu Fiber optic FT-IR spectroscopy for coal liquefaction processes," SPIE Paper No. 2069, (1993).

26. S. C. Bates, "High Temperature Fiber Optic Imaging," SBIR Phase I Final Report, NASA Contract NAS3-26566, (1995).

27. J. L. Dunlap, B. A. Carreras, V. K. Pare, J. A. Holmes, S. C. Bates, J. D. Bell, H. R. Hicks, V. E. Lynch, A. P. Navarro, "Magnetohydro-dynamic instability with neutral-beam heating in the ISX-B Tokamak," Phys. Rev. Lett., 48, 8, 538 (1982).

28. P. R. Solomon, S. C. Bates, and Y. P. Zhang, "Mild Gasification Transport Reactors, " SBIR Phase I Final Report, USDOE Contract DE-FG05-09ER80877, (1991).

29. S. C. Bates, "Controlled Crystal Growth Using Auxiliary Optical Heating and Optical Diagnostics," SBIR Phase I Final Report, NASA Contract NAS8-40546, (1995).

30. D. W. Yoel and S. C. Bates, "Visual Monitoring of MIM Debinding and Sintering ," International Conference on Powder Metallurgy and Particulate Materials, Seattle, WA, (1995).

31. S.C. Bates, "A Novel Binder for Reactive Metal Injection Molding," SBIR Final Report, July, NSF Grant #9760106, (1998).

32. S. C. Bates, "Insights into Spark-Ignition Four-Stroke Combustion Using Direct Flame Imaging," Combustion and Flame, 85, 3 & 4, 331 (1991).

33. S. C. Bates and L. Liou, "High Performance Sapphire Windows," Technology 2002, NASA Conference, Baltimore, MD, (1992).

34. S.C. Bates, "Low Loss Sapphire Windows for High Power Microwave Transmission," DOE STTR Final Report, Grant # DE-FG02-95ER-86038, (1999).

35. S.C. Bates and P. R. Solomon, "Elevated Temperature Oxygen Index Measurements and Apparatus," Journal of Fire Sciences, 11, 271 (1993).