GB2151369A - Optical fibres - Google Patents

Abstract

An optical fibre comprising a core and a cladding, which substantially prevents the passage of hydrogen through the cladding to the core. Such fibres can be produced by coating with silicon nitride, silicon, gold, silver, copper or SnO2, TiO2- doped glass.

Description

SPECIFICATION Optical fibres This invention relates to optical fibres. In particular, it relates to fibres intended for long distance communication, having low loss, e.g.

in the form of optical fibre cables for submarine use.

Stone et al., Optics Letts, 7 (1982) 297299, observed, during a study of Raman gain mechanisms in optical fibres, that hydrogen diffusion gave rise to a set of absorption peaks, between 1.08 and 1.24 ym, which were attributed to the first overtone spectrum of molecular hydrogen. Subsequently, Mochizuki et al., Electron. Letts. 19 (1983) 743745 quantified the same set of peaks, in a first experiment where an optical fibre cable was soaked in water and in a second experiment where a single-mode fibre was placed in a hydrogen atmosphere. Mochizuki et al. conclude by warning that, when laying cables, hydrogen should not be introduced, and that water propagation in the cable should be prevented. Loss increases in optical fibre cables filled with water are again reported by Uesugi et al., Electron. Letts. 19 (1983) 782763.This article concludes that the loss increase is caused by molecular hydrogen, itself generated by electrolytic corrosion of metals in the optical cable, which diffuses into the optical fibre and is held interstitially.

US-A-4118211 and US-A-4319803 disclose that static fatigue and abrasion/corrosion, respectively, can be reduced in an optical fibre by coating it with, for example, silicon nitride. The exact nature of the fibres for which the desired properties are shown is not clearly disclosed. US-4118211 suggests that a function of the improved strength consequent on the provision of an external coating may be reduced attack by a number of materials, including water, acids, alkalis and gases. US-A-4319803 discloses that a silicon nitride or silicon oxynitride coating, applied by chemical vapour deposition at 200-1400 C, has the advantage of being amorphous rather than crystalline or polycrystalline, and thereby less subject to fracture, corrosion and electrical isolation breakdown.

GB-A-1592234 discloses that an optical fibre comprising a core and a cladding respectively having compositions of differently doped silica may be coated with silicon nitride. It is stated that improved fatigue resistance can be achieved, although no specific data, or specific fibres, are given.

Accordingly, it has been demonstrated that a ceramic coating, e.g. of silicon nitride, can be provided on an optical fibre and thereby increase its strength characteristics. It appears that known ceramic-coated optical fibres are not of the type, i.e. with low loss, suitable for long distance communication.

None of the documents discussed above provides a solution to the problem of attenuation in long distance optical fibre communication as the result of hydrogen diffusion. It has now been found, surprisingly, that this problem can be satisfactorily solved for low loss optical fibres of considerable length.

An optical fibre according to the present invention comprises a core and a cladding, and has a loss of less than 1 dS/km at an attenuation minimum, and is substantially impervious to the passage of hydrogen through the cladding to the core. The fibre should be capable of mono-mode transmission, possibly also with a wavelength of zero dispersion, in a transmission window corresponding to the attenuation minimum.

As is usual in optical fibres, the core and cladding compositions will generally be of substantially similar composition, one or other having been modified such that the former has a higher refractive index. The difference in refractive index between the core and the cladding, in fibres of the invention, may be determined as necessary.

The glass compositions may be of the, say, borosilicate, aluminosilicate, germanosilicate, silica or fluoride type. The general nature of such compositions will be well known to those skilled in the art. The transmission window at which fibres of the invention can operate will depend on the nature of the glasses. This may be above 2.5 ym for fluoride glasses, or about 1.2 to 1.6 ym for silica glasses. For the purpose of illustration, the invention will be discussed below in terms of silica glasses optimised for transmission in the given range, e.g. a 1.3 or 1.55 Lm window. Transmission need not be exactly at an attenuation minimum.

Optical fibres of silica glasses may conveniently be prepared by chemical vapour deposition within a silica tube, e.g. MCVD. In this, layers of cladding and then of core material are deposited from an appropriate vapour mixture onto the inside of a silica tube which is then collapsed to yield the preform, which may be sleeved with another silica tube before drawing, so as to achieve a particular desired aspect ratio. The vapour mixtures that may be used as appropriate are mixtures of pure oxygen with one or more of SiCI4, GeCI4, POCI3 and CCI2F2 (these latter compounds providing Si, Ge, P and E respectively). Advantageously, chlorine is present as a drying agent during collapsing of the tube.

Among other methods which may be used for producing the preform are the rod-andtube method, outside vapour phase oxidation, vapour axial deposition, and plasma-modified chemical vapour deposition.

The cladding may consist of pure silica or it may include dopants. Dopants may comprise a refractive index-depressing dopant, e.g.

fluorine, and a refractive index-raising dopant such as phosphorus. It is desirable to reduce the P and E concentrations immediately adjacent to the core, so as substantially to avoid absorption due to phosphorus.

On the one hand, the amounts of such dopants in the cladding can be such that their respective effects on refractive index substantially cancel; this expedient permits the use of lower vapour deposition temperatures. In this case, a silica core must be doped so that it has a refractive index higher than pure silica; germanium dioxide is a suitable dopant, e.g.

in an amount of 2 to 12.5 mole %.

On the other hand, the cladding may have a refractive index lower than that of silica by a substantial amount (by 0.001 or more). This can be achieved by the use of a refractive index-depressing dopant, e.g. fluorine, not fully compensated for by refractive index-raising dopants, e.g. phosphorus. The core may be pure or doped silica.

The difference in refractive index between core and cladding (A71) may be, for example, from 0.003 to 0.02. The overall fibre diameter may be of the order of 100 pom. A vapour-deposited cladding thickness may be about 20 to 50 sum. Fibres of the invention are preferably intended for mono-mode operation in the transmission window; the core diameter is therefore usually 5 to 15, e.g. 8 to 10, m, and 10 lim is the preferred maximum.

A primary use for hydrogen-impervious optical fibres may be in long-distance submarine communication. Communication lengths, between repeaters, may be at least 20, 50 or even 100 km. Individual fibres may be jointed together to give such lengths.

A cable, adapted for submarine use, according to the present invention, includes an optical fibre which has a loss of less than 20, preferably less than 5, dB/km on transmission at an attenuation minimum, which comprises a core-and a cladding, and in which there can be substantially no hydrogen passage to the core. The fibre may have a loss less than 3, preferably less than 2, and most preferably less than 1, dB/km at an attenuation minimum. The fibre may also have any of the characteristics described above.

Submarine cables of the invention may be prepared in conventional form, e.g. comprising a steel wire and an aluminium tube in addition to one or more optical fibres. More particularly, a submarine cable can comprise a plurality of optical fibres supported by a high strength, e.g. steel, fibre, and which are supported within a metal, e.g. aluminium, sheath which is often laminated with a plastics, e.g.

polyethylene film.

Another use for hydrogen-impervious optical fibres may be in applications and loci where high temperatures, e.g. above 50 C, are met.

Such high temperatures may be the environment of, say, aerial fibres or fibre cables in tropical regions or of fibres within power transmision cables. At elevated temperatures, not only is the evolution of hydrogen from cable components likely to be increased, but also the susceptibility of the glass constituents of the fibre to hydrogen degradation by accelerated chemical reaction is increased; this affects attenuation. Of course, fibres in such environments may be long, with optical communication lengths as described above.

A power transmission cable according to the invention includes an optical fibre which has a loss of less than 20, preferably less than 5, dB/km on transmission at an attenuation minimum, which comprises a core and a cladding, and in which there can be substantially no hydrogen passage to the core. The fibre may also have any of the characteristics defined above.

The transmission cable itself may be of a conventional type. It may comprise, for example, one or more optical fibres aligned along the axis of the cable, which are surrounded by one or more power transmission lines within an insulating surround.

It is a feature of the present invention that a fibre of the invention is substantially impervious to the passage of hydrogen through the cladding to the core. As will be shown below, a fibre of the invention can be prepared which, after exposure to a hydrogen atmosphere under the given conditions, exhibited no observable loss at the fundamental hydrogen absorption wavelength of 2.42,um. As a result, fibres of the invention can be produced which can easily be designed to meet tolerance margins of, say, 0.05 dB/km over 25 years.

It is preferred that a fibre of the invention includes a hydrogen-impervious layer. This may be deposited on the outside of the fibre, during or following drawing of, say, a preform. Alternatively, a hydrogen-impervious layer may be deposited on the inside of a silica tube or sleeve (see the description of deposition processes, above). It is desirable that the layer should not be immediately adjacent to the core, and indeed outside the optical transmission zone.

A suitable material for the hydrogen-impervious layer is a ceramic such as that provided by the thermal dissociation of silicon and nitrogen compounds. Depending on the conditions, this reaction may give a ceramic coating, on an optical fibre, comprising silicon nitride or silicon oxynitride, or a mixture of the two. Such a ceramic will be described herein as a silicon oxynitride. It is known that such a material can be produced by the thermal deposition of ammonia and silane. Other materials might be used to deposit this or a similar ceramic layer in situ.

A further aspect of the present invention lies in the use of a silicon oxynitride coating for the prevention of hydrogen absorption into an optical fibre.

Alternative hydrogen-impervious layers may be of metallic composition. For example, a metal having a diffusion coefficient for hydrogen substantially less (e.g. a factor of less than 0.1 %) than silica may be used to coat a suitable low loss optical fibre, for use in the invention. For example, an optical fibre may be coated with Si, Au, Cu or Ag. Metal-coated fibres of the invention may be of particular utility within submarine repeaters.

Effective hydrogen-impervious layers may be formed from materials in addition to those described above. For example, a diamond-like form of carbon may be used; the glass composition may be a silica glass doped with, for example, TiO2 or SnO2.

Hydrogen diffusion effects may also be minimised by the nature of the core or cladding composition; this may be chemically determined, by adjusting the level of, say, Ce, Sb, As, Ge and/or P in order to influence the concentration of defect sites in the glass. This concentration may also be influenced by the temperature, speed, tension or other drawing conditions. A fibre whose core and/or cladding has been modified in this way may be a fibre as defined elsehwere herein; however, it may also be a fibre in which no measure has been taken substantially to prevent the passage of hydrogen through the cladding to the core but, rather, a fibre in which any absorbed H2 is rendered substantially non-absorbing.

The means described above for the prevention of hydrogen diffusion are, effectively, physical. Hydrogen diffusion may also be effectively prevented chemically, again by the provision of a suitable layer in or on the cladding, but which reacts with rather than being simply impervious to, hydrogen. To this end, a cladding composition may be prepared having an increased P205 or B203 content at a certain point through the thickness of the cladding.

The degree to which hydrogen diffusion into an optical fibre can be reduced will now be illustrated, in three series of experiments, as follows: In the first experiments, tests were conducted on an ultra-violet-cured epoxy acrylate-coated optical fibre comprising a germania-doped silica core and a P- and F-doped silica cladding prepared by MCVD. The fibre had a core diameter of about 8 time The fibre was of the type generally suitable for monomode transmission in the 1.3 or 1.55 lim window, with low loss.

In a first test, a silicon nitride coating was deposited by thermal decomposition of ammonia and silane as N and Si sources, and the coated fibre was then held under hydrogen, at a 0.74 atm. (74 kpa) partial pressure of H2, at room temperature. Transmission measurements were conducted periodically at the hydrogen fundamental absorption wavelength of 2.42,us. No change was detected after 700 hours, indicating that there was substantially no uptake of molecular hydrogen by, and substantially no concentration of H2 in, the fibre.

For the purpose of comparison, the same procedure was conducted in a second test, except that no ceramic coating was applied to the fibre. The expected incremental loss at 2.42 cm, caused by hydrogen absorption, was observed. This indicated that the fibre had taken up molecular hydrogen, to the expected degree of concentration.

In the second experiments, results of which are given in the three graphs of the accompanying drawings, both mono-mode and multi-mode silica-based fibres were drawn using a ZrO2 induction furnace. Fibres were coated with amorphous films of a silicon nitride by the CVD reaction of silane, carbon dioxide and ammonia, at about 1000 C.

Film thicknesses were measured by Auger spectroscopy combined with argon ion sputter profiling and were found to be typically about 20 nm. Strength and dynamic fatigue measurements were conducted using the bend test technique reported by Duncan et awl., supra, and was similar to those reported in that paper and by Hanson et al., "Properties of silicon oxynitride-coated fatigue-resistant fibres", 3rd IOOC, Washington (1981) paper MG3.

Samples 5 m long were exposed to hydrogen at pressures of 0.74 and 65 atm. (74 and 6500 kpa) at room temperature. The molecular hydrogen fundamental absorption peak at 2.42 ym was used to monitor the diffusion of hydrogen through the oxynitride coating into the fibres. Samples without the oxynitride coating were used as reference standard, using the first overtone of the hydrogen absorption at 1.24 zm as the monitor in the high pressure experiments. The attenuation measurements were performed using conventional apparatus, with a pbs detector being used for wavelengths above 1.6,us. For 5 m samples, the limit of sensitivity of the equipment at 2.42 lim is about 1 5 dB/km.

Figure 1 shows the time-dependence of the loss increases at two hydrogen, related peak wavelengths (1.24 and 1.69 yam, respectively plotted as dots and squares) for an uncoated fibre in 0.74 atm. (74 kpa) hydrogen, at room temperature. The line through the data corresponds to the loss increases expected for a diffusion coefficient D(H2) = 1.7 X 1011 cm2/sec. The fractional loss was calculated in a way similar to that reported by Fox et al., Electron. Lett. 19(1983)916-917. The more complex case of a coated fibre is treated as a composite cylinder of different materials. The relative hydrogen concentrations, as a function of radius and time, were computed numerically using the standard diffusion equation with a Taylor series expansion for the concentration.The boundary conditions were that, at the interface, the flux was continuous and Kcl = c2, where c, is the concentration in the silicon oxynitride, c2 is the concentration in the silica and K is defined by the ratio of the equilibrium concentrations. The value of K is not known but is assumed to be at least 1 since silica has a more open structure than a silicon nitride.

The uncoated fibre has a 2.42 pm peak of about 1 200 dB/km when in equilibrium at room temperature at 0.74 atm. (74 kpa) hydrogen (after 500 hours). An oxynitride-coated fibre, tested under the same conditions, showed no measurable loss in increase after 40 days. Since the limit of sensitivity is about 1 5 dB/km, this null result indicates that the fractional loss increase, i.e. the loss measured normalised by the loss of an uncoated fibre in equilibrium at the same pressure of hydrogen, is about or below 102. Analysis shows that D(H2) for the silicon oxynitride coating is at least 5 orders of magnitude lower than for silica.However, even for this large difference in diffusion coefficients, calculations predict that the increase in hydrogen concentration after 25 years, a typical system lifetime, would be more than one-half of the equilibrium value of an uncoated fibre, and therefore of limited practical value.

In order to increase the sensitivity of these experiments, and thus obtain an improved estimate of D(H2) in the oxynitride, fibres were tested in 65 atm. (6.5 Mpa) hydrogen at room temperature. Uncoated fibres showed a loss of 600 dB/km at 1.24,us, corresponding to approximately 1 x 105 dB/km at 2.42 m. Coated fibres showed no measurable increase in loss at 2.42 ym after 1 73 days. The fractional loss increment now detectable has been reduced to about 104.

Figure 2 shows the theoretical loss increase (for K = 1) for a range of diffusion coefficients, as a function of time. In Figure 2, the diffusion coefficients for the four curves are (a) 1 X 1018, (b) 5 X 1019, (c) 2.5 X 1019 and (d) 1 X 1019, cm2/sec; the vertical line bridging curve (d) only is the experimental data point for p(H2) = 65 atm. (6.5 MPa) and an oxynitride coating 20 nm thick. This indicates that the upper limit for D(H2) in the oxynitride coating is about 1 x 1019 cm2/sec, compared with 1.7 X 10 cm2/sec for silica.

Figure 8 shows the fractional concentration increase (which is equal to the fractional interstitial loss increase) expected for longer periods, up to 25 years, for a range of values of D(H2), K = 1 and a film thickness of 20 nm.

The four curves show the fractional increase in concentration (at the fibre axis) for oxynitride-coated fibres for the same four values of D(H2) as for Figure 2. D(H2) is assumed to be constant over the 25 years. For D(H2) = 1 X 1 019 cm2/sec, the fractional concentration gradually builds up to about 0.01 after 25 years, whereas the corresponding time for the uncoated fibre is only of the order of 10 hrs.

This discussion has assumed K = 1; if K is greater than 1, there are different combinations of K and D(H2) which fit the experimental observation. However, the long-term extrapolation for all these combinations with K at a value greater than 1 will result in lower fractional concentrations than for the case of curve (b) in Figure 3 (where K = 1).

The implications of the use of oxynitride coatings for optical systems operating in a hydrogen-containing atmosphere are important. For the silicon oxynitride-coated fibre, the fractional concentration after 25 years is about or less than 0.01 and, for most of the system lifetime, it is significantly less than the 25 year value. There would thus be an increase by a factor of at least 100 in the safe working pressure of hydrogen, compared to that for uncoated fibres. On the assumption that, for future mono-mode fibres, the interstitial effects will outweigh the permanent changes, and that a loss increase of 0.03 dB/km at 1.3 ym can be tolerated within initial systems budgets, then a system using uncoated fibre will have a safe working pressure of 0.1 atm. (10 kpa) hydrogen. By comparison, a coated fibre would survive with 10 atm. (1 MPa) hydrogen in the cable.

A third series of experiments concerns the coated fibres at elevated temperatures. Specifically, by monitoring the 2.42 jum H2 fundamental peak, no H2 ingress into the coated fibre was observed after 70 hours at 230 C in 45 atm. (4.5 Mpa) H2. The equilibrium time for uncoated fibres at this temperature is less than 10 hours.

In conclusion, it appears that fibres of the invention may effectively prevent hydrogen diffusion, thereby reducing all hydrogen-related optical aging mechanisms.

Claims (16)

1. An optical fibre comprising a core and a cladding, which is capable of mono-mode transmission in a transmission window corresponding to an attenuation minimum, which has a loss of less than 1 dB/km at the attenuation minimum, and which substantially prevents the passage of hydrogen through the cladding to the core.

2. An optical fibre comprising a core and a cladding, which is capable of mono-mode transmission in, with a wavelength of zero dispersion in, a transmission window corresponding to an attenuation minimum, which has a loss of less than 1 dB/km at the attenuation minimum, and which substantially prevents the passage of hydrogen through the cladding to the core.

3. A fibre according to claim 1 or claim 2, which comprises a hydrogen-impervious coating on the outside of the cladding.

4. A fibre according to any preceding claim, which comprises a hydrogen-impervious layer within the cladding.

5. A fibre according to claim 3 or claim 4, in which the coating or layer comprises a ceramic material.

6. A fibre according to claim 5, in which the ceramic material is a silicon nitride or oxynitride.

7. A fibre according to any preceding claim, in which the diameter of the core is no more than 10,l4m.

8. A fibre according to any preceding claim, in which the core and cladding are silica glasses and the transmission window is at 1.3 or 1.55 item.

9. A fibre according to any preceding claim, in submarine use.

10. A fibre according to any of claims 1 to 8, in use at a temperature above 50 C.

11. A submarine or power transmission cable including an optical fibre according to any of claims 1 to 8.

1 2. A cable, adapted for submarine use, including an optical fibre which has a loss of less than 20 dB/km on transmission at an attenuation minimum, which comprises a core and a cladding, and in which there can be substantially no passage of hydrogen through the cladding to the core.

1 3. A power transmission cable including an optical fibre which has a loss of less than 20 dB/km on transmission at an attenuation minimum , which comprises a core and a cladding, and in which there can be substantially no passage through the cladding to the core.

1 4. A cable according to claim 1 2 or claim 13, in which the fibre has a loss of less than 5 dB/km and/or any or all of the features defined in claims 1 to 8.

1 5. The use of a silicon nitride or oxynitride coating for the prevention of hydrogen absorption into an optical fibre.

16. The use according to claim 15, in which the fibre has any or all of the characteristics defined in claims 1 to 8.