Metal-Coated Fibers

This is a continuation from the previous tutorial - acousto-optic devices and applications.

1. Introduction

An optical fiber has to be defended by some protective coating from mechanical damage during handling and from environmental factors during its use. In many cases, a polymer coating is appropriate for enough protection.

However, there are a number of special applications of optical fibers in which ordinary polymer-coated fibers cannot be used. These applications can be divided into several groups:

1. Increased reliability (hermeticity of the coating is important)
2. High vacuum (when outgassing from the coating is undesirable)
3. Possibility of soldering (embedded fibers, pigtails, inlets to high vacuum)
4. Delivery of high-power laser radiation (polymer can inflame by scattered light)
5. Medical applications (metal-coated fibers can be sterilized using ETO, steam, e-beam, or \(\gamma\)-radiation)
6. Harsh environments
   \(\cdot\) High-temperature environments (>350\(^\circ\)C)
   \(\cdot\) Nuclear radiation (polymer coating decays under radiation)
   \(\cdot\) Chemicals (if they do not cause corrosion of the metal)

Real-life applications (such as sensors, aerospace, chemical industry, deep-well oil-field industry) can belong to several groups simultaneously.

In contrast to carbon (another type of hermetic coating), metal coatings do not need an additional protective polymer coating. Thus, metal coatings have no contender in applications for which the presence of a polymer coating is undesirable.

The known specific applications for metal-coated fibers are as follows:

• Radiation-resistant fiber optic systems intended for use in the nuclear industry (e.g., plasma diagnostic systems in thermonuclear reactors, image guides for visual inspection of nuclear installations). To increase radiation resistance, the fiber can be heated up to approximately 400\(^\circ\)C. Alternatively, its glass can be loaded with molecular hydrogen.
• High-temperature alarm systems remaining functional in accidental conditions (e.g., in case of fire).
• Fiber optic sensors of temperature, vibration, and so on integrated into complicated devices (e.g., jet engines, turbines).
• High-temperature fiber optic systems resistant to hydrogen penetration meant for applications in the chemical and oil-field industries.
• Enhanced-reliability fiber optic devices in which fibers are soldered to connectors (e.g., devices for the space industry).
• Coolable, incombustible fibers for laser-power delivery.

The fiber can be coated by a metal film after drawing in a separate process (off-line) or during drawing (in-line).

Off-line metal-deposition processes, for example, sputtering of trimetal coatings (Ti/Pt/Au), were reported. Another example is an electrolytic plating process for application of Ni/Au coatings.

The aforementioned methods are feasible if only a short length of a polymer-coated fiber (several inches) is to be coated by metal. Such a length is enough if the fiber is to be soldered at the seal location during pig-tailing or packaging.

The off-line slow deposition process can guarantee precise thickness of the metal layer (typical thickness is a few microns), which is important to adjust and to fix the fiber to a high accuracy by soldering.

An attractive possibility is to apply a metal coating in-line (during the drawing process). Attempts to apply a metal (Ni, Mo, Cu, or Ag) coating in-line using sputtering in a vacuum, magnetron sputtering (Cu), ion-plasma deposition (Sn, or In) were not quite successful; the coatings obtained were not hermetic and the fiber strength was low.

In addition, the aforementioned methods required rather expensive equipment and subsequent electrolytic plating was necessary to solder the fiber ends. The continuous plating process of Ni and Cu application on a carbon-coated fiber has also been realized.

A high fiber strength is hardly to be achieved with this method. At present, only the ‘‘freezing’’ method allows application of a metal coating in-line during the drawing process of the fiber.

In this case, the fiber passes through a layer (approximately a few millimeters) of molten metal. If the temperature of the melt is close to the melting point of the metal and the temperature of the fiber is lower, then some layer of the metal can ‘‘freeze’’ on the surface of the fiber.

For this technique, a usual drawing tower can be used with just one modification; a specially constructed metal applicator should replace the polymer die.

It was demonstrated that metal-coated fibers fabricated by the ‘‘freezing’’ technique were indeed hermetically sealed. It means that because of the absence of water vapor under the metal, the fiber strength can be twice as high as that of polymer-coated fibers (5.5 GPa).

This value for metal-coated fibers approaches the glass strength in liquid nitrogen (~ 14 GPa). Fatigue parameter \(n\) in this case will be also high (>100), as compared to \(n\) ~ 20 for polymer-coated fibers.

Carbon-coated fibers also demonstrate such a high fatigue parameter but cannot reach such a high strength because of cracking of the brittle carbon coating during the tests at high elongation (approximately >5%).

Unfortunately, the freezing method of the metal-coating application has some restrictions:

1. Only metals with a comparative low (\(\le\) 1400\(^\circ\)C) melting point (In, Sn, Pb, Zn, Al, Ag, Cu, Au, Ni) can be applied without technological problems resulting in strength reduction. It appears that it is impossible to achieve satisfactory application of such metals as Ti, Co, and Pd by the ‘‘freezing’’ method.
2. There are only a very limited number of alloys that have been optimized for application by this method. Compositions of such optimized alloys are usually far from the conventional alloys, such as corrosion-resistant alloys.
3. Stable metal application is possible in a limited range of the coating thickness (e.g., 15–25 \(\mu\)m for a fiber diameter of 125 \(\mu\)m, or approximately 50 \(\mu\)m for the fiber diameter of 250 \(\mu\)m).

Our 20 years of experience shows that for many special applications Al-coated fibers are quite suitable. In some cases, related to soldering or extremely high temperatures, Cu-coated or Au-coated fibers are better candidates. Ni-coated fibers could be used as sensors of magnetic field.

Use of long-length metal-coated fibers is associated with the problem of microbending optical losses because of a high expansion modulus of the metal and a very high difference in the thermal expansion coefficients of silica and metals.

Preliminary thermo-cycling is usually used to stabilize optical losses in the reduced temperature range. Fibers of a thick diameter and/or a high aperture are less sensitive to the microbending effect.

In the following section, we provide more detailed information on the ‘‘freezing’’ technology and the properties of metal-coated fibers obtained by this technique.

2. Freezing Technique

In the 1960s, Arridge et al. and Arridge and Heywood demonstrated, for the first time, aluminum application on silica fibers using the freezing process. These nonoptical fibers were meant for use in fiber-reinforced aluminum constructions. Later, this method was used for aluminum coating deposition on optical fibers.

At present, only the freezing method allows one to apply a metal coating in-line during drawing of a fiber of any length. For this technique, usual drawing towers can be used with just one modification: A specially constructed metal applicator should replace the polymer die (Fig. 15.1).

In this technique, the fiber passes through a layer (approximately few millimeters) of molten metal. If the temperature of the melt is close to the melting point of the metal and the temperature of the fiber is somewhat lower, a layer of the metal can freeze on the surface of the fiber.

To obtain a stable uniform metal film, the duration of the contact of the fiber with the molten metal in the metallizer should be shorter than the time of fiber heating to the metal melting point. Otherwise, the frozen layer will melt again and the fiber will pass through the metallizer without any coating.

A detailed schematic of the process is given in Fig. 15.2. At the moment of the first contact of the fiber with the melt, the layer of the frozen metal arises on the fiber surface. Its thickness rapidly grows to the value depending on the energy the fiber can take from the melt.

In other words, the coating thickness depends on the fiber thickness and its temperature, as well as on the temperature of the melt. After the appearance of a coating, during the further passage through the melt, the thickness of the metal layer gradually decreases as a result of melting of the frozen metal.

Because all metals are well wetted with their melts, some amount of molten metal is carried out from the metallizer as a liquid film on the solid surface of the frozen metal.

A mathematical description of the process is quite complicated. Nevertheless, there are a few publications on this issue, in which the calculated results are close to the experimental ones.

The main parameters defining the thickness of the metal coating are the diameter and the temperature of the fiber, the temperature of the melt (i.e., how far it is from the melting point), the longitudinal length of the bath of the melt, and the speed of the fiber drawing. Typical results are shown in Fig. 15.3.

In contrast to the polymer coating process, the diameters of the inlet and outlet do not significantly influence the diameter of the coating. These diameters may be much greater than that of the resultant metal-coated fiber. Molten metal does not stream up or down from the metallizer through the inlet or outlet owing to surface tension.

Surface tension keeps metal from streaming, only if the melt does not wet the material of the metallizer and does not react with it. An oxidizing film can also be a problem for surface tension, so an oxygen-free atmosphere is very desirable, at least, at the outlet.

The maximum thickness of the coating strongly depends on the fiber diameter and its temperature, that is, on the energy the fiber can take from the melt when heated to the melting point. For a 125-\(\mu\)m fiber, we obtained a maximum thickness of approximately 25 \(\mu\)m for most of the studied metals; for a 250-\(\mu\)m fiber, it was about 60 \(\mu\)m.

Changing the process parameters (e.g., the temperature of the melt), we could change the thickness of the coating in the range between the maximum value and half of that. It is possible to maintain the chosen coating thickness along the fiber length with a typical accuracy of \(\pm 2\) \(\mu\)m.

In accordance with the theory, we could reduce the coating thickness down to a few microns. Nevertheless, if we tried to obtain the thickness less than half of the maximum value, the application process became rather unstable and sensitive to small perturbations of the drawing parameters.

Under abnormal coating application conditions, uncoated fiber spans (several millimeters in length) became possible. In addition, with a coating thickness smaller than a certain critical value, there arose holes in the coating (Fig. 15.4).

Figure 15.5 presents a view of a metal-coated fiber after stripping the coating from the front side. More or less regular cavities in the metal are seen through the transparent fiber glass.

Figure 15.6 is an SEM picture of a copper coating taken off of a fiber. Cavities can be seen as well forming orthagonal bands along the inner surface of the copper metal coating.

Cavities in the glass–metal interface are inherent in the freezing technique. Usually they appear regularly with a period comparable to the fiber diameter. This phenomenon is due to hydrodynamic instability of the melt’s flow near the inlet meniscus. Oscillations of the meniscus can be described as standing capillary waves (Fig. 15.7).

These standing waves are excited by fluctuation of the parameters of the process, such as the fiber-drawing speed, diameter, position, and so on. The depth of the cavities can be about 5 \(\mu\)m. Thus, this effect limits the minimum possible thickness of a continuously applied metal layer onto a fiber.

The presence of cavities makes the strength and fatigue of a metal-coated fiber sensitive to the atmosphere over the metallizer, because the cavities are filled with the corresponding gases (see the next section).

Moreover, microbending optical losses in metal-coated fibers can be caused by the cavities or, at least, increased by their presence.

Only metals with a comparatively low melting point (In, Sn, Pb, Zn, Al, Ag, Cu, Au) can be applied by the freezing method without a noticeable problem with the fiber strength due to the reaction of the melt with silica.

Very high stresses can arise in the metal film and in the silica fiber during cooling after metal application because of a big difference in thermal expansion coefficients of silica and the metal.

Fortunately, it is not a problem for pure metals, thanks to quick stress relaxation at a high temperature due to mobile dislocations. However, for nonoptimal alloys, the difference in thermal expansion coefficients may be a problem, because mobility of dislocations may be significantly lower. As a result, the fiber or the metal film can crack during cooling (Fig. 15.8).

A metal coating usually has a glossy smooth surface, but its structure can be various. If this surface is etched in a special way, the structure of the metal can be brought out.

We found that in some cases (at some regimens), the metal film consisted of small polycrystalline grains (Fig. 15.9), whereas at other regimens, it looked like a surface of a monocrystal (Fig. 15.10).

X-ray analysis confirmed this visual observation (Fig. 15.11). Our experience shows that both mechanical and optical properties of metal-coated fibers are significantly better in the case of a ‘‘monocrystalline’’ structure of the coating.

3. Strength and Reliability

The overall strength and reliability of silica-based optical fibers depend on the fatigue effect. It means that flaws in the glass subjected to a tensile stress in the presence of moisture grow subcritically before failure.

Because of this, the strength of standard polymer-coated fiber (a typical value is ~ 5.5 GPa at 50%RH and room temperature) is significantly less than that at the liquid nitrogen temperature (~ 14 GPa), when the influence of water vapor is minimized.

Fatigue is the reason of limited reliability of polymer-coated optical fiber under static stress. A power law with fatigue parameter \(n\) ~ 20 is usually used to describe the fatigue effects and to predict the time to failure of the fiber in service.

In the case of strength testing at different loading rates (dynamic fatigue), the power law gives

\[\tag{15.1}\frac{\sigma_1'}{\sigma_2'}=\left(\frac{\sigma_{d1}}{\sigma_{d2}}\right)^{n+1}\]

where \(\sigma_{d1}\) and \(\sigma_{d2}\) are the tensile strength of similar samples at loading rates \(\sigma_1'\) and \(\sigma_2'\), respectively. That is, for \(n\) = 20, the tensile strength increases by a factor of approximately 1.12 if the loading rate increases by a factor of 10.

In the case of tests at a constant stress (static fatigue), time to failure \(t\) will increase by a factor of 10 if applied stress \(\sigma_s\) is decreased by a factor of about 1.12 (for \(n\) = 20), according to the following relation:

\[\tag{15.2}\frac{t_2}{t_1}=\left(\frac{\sigma_{s1}}{\sigma_{s2}}\right)^n\]

It was predicted that in the absence of moisture on the fiber surface (e.g., under a hermetic coating), slow crack growth still could take place under stress, because of thermofluctuations.

Time to failure \(t\) under static stress \(\sigma_s\) in that case can be evaluated by the following expression:

\[\tag{15.3}t=t_0\exp\left[\frac{U_0}{kT}\left(1-\frac{\sigma_s}{S_i}\right)\right]\]

where \(t_0\) is the value close to the period of atomic thermal fluctuations (~ 10\(^{-13}\)s); \(U_0\) is the energy of Si-O bonds in silica glass (~ 110 kcal/mol); \(S_i\) is the fiber initial strength (in the absence of thermofluctuations at \(T=0^\circ\)K).

The thermofluctuation model predicts that the strength of a hermetically coated fiber must be about 0.84 \(S_i\) at room temperature and about 0.95 \(S_i\) in liquid nitrogen. Because these estimations are based on the exponential description, parameter \(n\) depends on time and applied stress.

We can estimate the \(n\) value using (15.3)

\[\tag{15.4}n=-\frac{d(\ln[t_s])}{d(\ln[\sigma_s])}=\frac{U_0}{kT}\frac{\sigma_s}{S_i}\]

It can be calculated from Eqs. (15.3) and (15.4) that the \(n\) value decreases under laboratory conditions from 155 to 135, when time to failure changes from 1 second to 30 years.

Thus, the main features of an ideal hermetically coated fiber are a high strength (approaching that in liquid nitrogen) and a high \(n\) value.

In real cases, there exists a problem of correctly measuring the strength of metal-coated fibers. It is quite difficult to measure such a high strength by a usual tensile testing machine because of the difficulty of fixing the fiber ends during the test without damaging the fiber surface.

Because all the fiber length between the holders usually molders away after the failure, it is impossible to detect the position of the initial place of break, whether it is in the holder or between the holders.

The two-point bending technique is free of this problem, but in this case, only a very short fiber length (~ 1 mm) is subjected to stress.

In addition, correct calculation of the bending radius and the respective strength in the case of the two-point bending technique is hampered by a nonlinear behavior of Young’s modulus of silica at elongations up to 8% and by the absence of experimental data at higher elongations (failure in liquid nitrogen occurs at ~ 15% elongation).

Nevertheless, it was demonstrated that metal-coated fibers fabricated by the ‘‘freezing’’ technique could be really hermetically sealed. It means that the strength at room temperature is close to that in liquid nitrogen and n value is higher than 100.

Bending tests show that to obtain the best result, the metal film has to be of ‘‘monocrystalline’’ structure (Fig. 15.12) and water vapor has to be excluded from the atmosphere over the metallizer, where the fiber is still uncoated (Fig. 15.13).

Tensile tests of fibers coated by various metals do not give so high strength levels (Fig. 15.14). Nevertheless, all those fibers were stronger than the conventional polymer-coated fibers.

Carbon-coated fibers also demonstrate fatigue parameter \(n\) >100, but their strength is less than that of standard fibers. The reason is brittleness of the carbon film. It cannot survive a high elongation (approximately >5%). Carbon-coating cracks during strength tests at a higher elongation were followed by a failure of the fiber itself.

Metal-coated fibers made by other methods (sputtering, plating, etc.) are usually nonhermetic or may contain water under the coating. Thus, only metal-coated fibers fabricated by the ‘‘freezing’’ technique demonstrated a strength higher than that of standard fibers.

Unfortunately, it is difficult to take advantage of this unique strength of metal-coated fibers in the case of a long fiber length. The problem to obtain a long-length metal-coated fiber of a uniform high strength is significantly more complicated than that solved for polymer-coated fibers.

In addition to usual sources of possible defects, a number of new types of defects, such as inclusions of hard particles in the metal, pores or holes in the metal film, and so on, should be identified, systematized, and eliminated.

In addition, to ascertain a high strength of the whole length, the fiber is to be proof-tested (rewound from one reel to another) at some load (strain). However, the process of proof-testing itself at loads over approximately 5–10 N (for a 125-\(\mu\)m fiber) can damage the metal coating and the fiber.

Moreover, tin- or indium-coated fibers survive proof-testing only in the case of an additional polymer coating applied over the metal. This problem is less acute, if only bending at low tension on a certain set of wheels is used as the proof-test procedure instead of tensile load.

4. Degradation at High Temperature

Ideally, in the absence of corrosion or oxidation effects, a metal coating remains at work at temperatures close to the melting temperature of the metal. However, in reality, reaction of metal with silica can drastically reduce the fiber strength at an elevated temperature.

It should be noted that we did not observe this effect for metals with a low melting point (In, Sn, Pb, Zn). Some reduction of strength of tin-coated fibers at high temperatures was explained by thermofluctuation effects.

However, for an Al coating (one of the best candidates for most applications), a reaction between aluminum and silica significantly reduces the working temperature.

Initially an interaction between silica and aluminum at high temperatures was observed in bulk aluminum reinforced by silica fibers. Molten aluminum easily reacts with silica in the following way:

\[\tag{15.5}4\text{Al}+3\text{SiO}_2\rightarrow2\text{Al}_2\text{O}_3+3\text{Si}\]

This reaction also takes place at the interface at temperatures lower than the melting point (~ 660\(^\circ\)C) with a rate decreasing with decreasing the temperature.

Because of this, the strength of Al-coated fibers degrades rapidly at temperatures of about 500\(^\circ\)C. Even a carbon film at the interface cannot significantly slow this process (Fig. 15.15).

Activation energy of 250–290 kJ/mol was obtained from experiments on strength degradation at different temperatures. A conclusion can be drawn from extrapolation to lower temperatures that Al-coated fibers can be used in long-term applications only at temperatures less than 400\(^\circ\)C.

Unfortunately, the process at the metal–silica interface was not so well studied for metals with a higher melting point (Au, Ag, Cu, Ni). These metals do not react chemically with silica at a high temperature, in contrast to aluminum.

However, for fibers coated with copper with protection from oxidizing, we observed strength degradation after aging at temperatures of 600–1000\(^\circ\)C during a few hours.

The reason for this degradation was crystallization of the surface of silica glass. The effect of degradation was higher if oxygen penetrated into the cavities at the silica–copper interface to form a copper oxide film at the silica surface. Strength degradation of Ni-coated fibers at 600\(^\circ\)C was even stronger than that of copper-coated fibers.

This phenomenon is in qualitative agreement with the data on crystallization of bulk silica, which is significantly accelerated by contaminations on the silica surface.

In the case of metal coatings, all the surface of the fiber is ‘‘contaminated’’ by a metal. Thus, crystallization of the silica surface at elevated temperatures (>600\(^\circ\)C) is a fundamental factor resulting in strength degradation.

Significant reduction of strength at 600\(^\circ\)C usually occurs during approximately 1 week or faster depending on the specific metal.

5. Optical Properties of Metal-Coated Fibers

Because the drawing technology of metal-coated fibers differs from that of polymer-coated fibers only in the fact that the fiber is heated for a short time to the melting temperature of the metal, metal-coated fibers feature approximately the same optical loss as polymer-coated fibers.

In some cases, the loss in metal-coated fibers is even lower. For example, in Al-coated low-OH pure silica core fibers, the 0.63-\(\mu\)m band inherent in polymer-coated fiber disappears because of thermal annealing in the metallizer. However, a relatively thick metal coating causes additional microbending loss.

The physical mechanisms of this effect include

1. A high Young modulus of metals, close to that of silica
2. A large difference in the thermal expansion coefficients of silica and metals
3. A low threshold of plastic deformation of pure metals (<10 MPa)

Losses in as-drawn normal numerical aperture (NA) Al-coated fibers can be as high as 20–100 dB/km at room temperature. Certain treatment or temperature cycling can change the level of microbending loss.

For example, rewinding the fiber to another reel can either increase or decrease the excess loss depending on a number of parameters (rewinding tension, diameter of the guiding rollers, the reel diameter, etc.).

Although the excess loss can be minimized, it poses a problem for some applications of metal-coated fibers.

It is known that high-NA metal-coated fibers demonstrate virtually no excess loss due to microbending.

A properly selected regimen of temperature cycling in the range from \(-20\) to \(+60^\circ\)C led to a reduction of microbending loss to approximately 0.1 dB/km in a multimode Al-coated fiber (cladding diameter 125 \(\mu\)m, NA ~ 0.2).

Further thermal treatment of this fiber in the range 5–40\(^\circ\)C gave an excess loss of no more than 0.2 dB/km. The increase of loss in such a fiber due to rewinding can be suppressed by a subsequent temperature cycling.

A typical example of loss variation during temperature cycling in the range from \(-120\) to \(+300^\circ\)C is given in Fig. 15.16.

It is seen that the added loss strongly depends on the thermal prehistory. The temperature range for each type of metal-coated fiber can be roughly divided into three regions (Figs. 15.16 and 15.17).

• Region 1. Loss in this region does not depend on the thermal prehistory and does not exceed several decibels per kilometer.
• Region 2. Loss depends on temperature and thermal prehistory in a complicated way. Loss relaxes to an acceptable level of several decibels per kilometer during a period of 1 hour to 1 month.
• Region 3. The excess loss level is high (>20–40 dB/km) and does not change in the course of thermal annealing.

Region 2 corresponds to temperatures of \(-70\) to \(-20^\circ\)C for Sn coating, to \(-50\) to \(-10^\circ\)C for Pb coating, to \(+20\) to \(+200^\circ\)C for Al coating, to \(+20\) to \(+250^\circ\)C for Cu coating, and to \(+20\) to \(+150^\circ\)C for Au coating.

At higher temperatures (Region 1), the microbending loss problem does not exist. At lower temperatures (Region 3), loss relaxation takes too long a time. Although the boundaries of the regions are not determined exactly, they do not strongly vary for different fiber types and are governed by the type of coating.

Zeroth excess loss occurs in a fiber when no mechanical stress is applied to the fiber by the coating—in other words, when no microbending takes place.

In this case, the internal stresses in the metal must be close to zero.

Metals are subject to plastic deformation in a wide temperature range, from zero to the melting point. The free energy of a crystalline material rises during deformation because of the presence of dislocations and interfaces.

A material containing such defects is thermodynamically unstable. If the material is subsequently heated to a high temperature (annealed), thermally activated processes, such as solid-state diffusion, promote elimination of defect.

In the course of recrystallization, because of the motion of dislocations, their annihilation, and migration of the grains’ boundaries, the structure of grains is restored or a new structure arises with a very low density of dislocations.

This means that the residual stresses in the coating are minimal after recrystallization. The relationship between the recrystallization rate and the temperature is given by the Arrhenius equation: The temperature of recrystallization decreases with increasing the annealing time.

Data on the recrystallization temperature for some metals is given:

Al, 130---200\(^\circ\)C; Cu, 180---240\(^\circ\)C; Au, 160---200\(^\circ\)C; Sn, \(-70^\circ\)C; Pb, \(-50^\circ\)C

The recrystallization temperature ranges coincide well with boundary of Region 2, in which a noticeable reduction of losses occurs in metal-coated fibers. However, the recrystallization temperature and rate may strongly vary depending on the specific crystalline structure of the metal, composition of impurities, degree of deformation, and the sample shape. It was found that fibers with a polycrystalline structure of the metal coating have a high excess loss and relaxation is not so efficient.

For this reason, it is desirable to apply a metal coating with a monocrystalline structure. Another way to overcome the microbending loss problem is to increase the fiber diameter. In this case, the coating cannot bend the fiber so strongly, despite the increased thickness of the coating.

For Al-coated fibers with NA ~ 0.2 and cladding diameter more than 250 \(\mu\)m, losses larger than several decibels per kilometer are observed only at \(T\lt20^\circ\)C. For fibers with a cladding diameter of 300–1000 \(\mu\)m, the microbending loss problem does not virtually exist.

Apart from the reversible microbending loss, irreversible loss growth occurs in metal-coated fibers in the temperature range 300–1000\(^\circ\)C. Such effects have been detected but have not been investigated thoroughly.

It is known that the irreversible loss strongly depends on the core and cladding chemical composition and increase as a result of hydrogen penetration from outside.

6. Summary

During the last 20 years of research on metal-coated fiber technology, many fundamental works have been performed aimed at obtaining coatings with genuine hermeticity. With such coatings, the static fatigue phenomena can be minimized and the fiber strength can be close to the theoretical limit for silica glass.

Although some problems remain unsolved, hermetically metal-coated fibers have gained recognition as an important fiber type for applications in harsh environments, where fibers with ordinary coatings cannot be used.

A variety of metals with the melting points of 1400\(^\circ\)C or less have been mastered as coating materials to endow the fibers with unique properties.

A variety of applications have been determined for which hermetically metal-coated fibers is the best-suited fiber type. Such applications include

• Radiation-resistant fiber optic systems intended for use in the nuclear industry (e.g., plasma diagnostic systems in thermonuclear reactors, image guides for visual inspection of nuclear installations). To increase radiation resistance, the fiber can be heated up to approximately 400\(^\circ\)C. Alternatively, its glass can be loaded with molecular hydrogen.
• High-temperature alarm systems remaining functional in accidental conditions (e.g., in case of fire).
• Fiber optic sensors of temperature, vibration, and so on integrated into complicated devices (jet engines, turbines).
• High-temperature fiber optic systems resistant to hydrogen penetration meant for applications in the chemical and oil-field industries.
• Enhanced-reliability fiber optic devices in which fibers are soldered to connectors (e.g., devices for the space industry).
• Coolable, incombustible fibers for laser-power delivery.

The next tutorial introduces optical beams and resonators.