Today I will focus on telecommunications technology for the first decade of the 21st century. Few things incite speculation about the future more than a year ending in the numeral “zero.” With the approaching millennium and its near-mystical three zeroes, everyone everywhere hungers for a glimpse into the future. My vision of the telecommunications future shows a collective of universal capabilities made possible by limitless digital networks with ubiquitous access and functionality. Mobile communications, the Internet, video, telephony and a broad array of end-user services will be the major drivers moving network providers to fuse capability with transparency. Integration of the full spectrum of communication possibilities will be the trend of the 21st century.

This paper presents the current advances in the development of materials in the field of telecommunication wiring together with the background which created the need for the new technologies. The background is somewhat unique to the current socio-economical situation of Japan. The key technical areas, which are addressed in this paper, are; (1) fibers for transmission of over 1 Tbps by utilizing WDM (wavelength division multiplexing), (2) fibers for fiber amplifiers to replace repeaters within trunk lines, (3) high density cables with over 1000 fiber counts, (4) ABF (air blown fiber) for access system, (5) plastic optical fibers wiring in premises or offices operating at the transmission speed of over 200 Mbps.

In this paper we review field data from failures related to materials and devices that are the building blocks of optically amplified dense-wavelength division multiplexed (DWDM) transmission systems. A DWDM transmission system consists of at least three distinct components: an optical multiplexer, one or more optical fiber amplifiers (OFAs), and an optical demultiplexer. A tandem arrangement of these components performs the function currently up to 40 independent high speed optical communications channels (about 2 million phone calls) over a single fiber link. Of these components, OFAs are considered the workhorse of this new technology, and, as with any new technology, their reliability is the focus of considerable attention. Commercially available OFAs are erbium-doped silica or fluoride fibers that operate in the 1550 nm optical communication window. Such OFAs consist of an active Er-doped fiber and pump laser(s) that collectively are referred to as the gain module, passive optical components used for routing, and electronic components for monitoring and operational support. We present the result of the failure analysis for each part as well as their role in the failure of the OFA. Using the data from systems deployed in the field, we summarize the history of failures of OFAs and other components and discuss their impact on the reliability of the DWDM system as a whole. From a material, packaging and system integration point of view, we address the reliability merits of different designs that, by and large, perform the same overall function. Finally, we shall outline how network providers that deploy such DWDM systems would benefit from a manufacturing strategy that incorporates reliability in its initial steps starting with the materials and packaging needs.

In order to guarantee a minimum initial strength, lightguide fibers are proof-tested. However, the strength of these fibers is known to depend on time (due to fatigue and aging) and this time-dependence must be taken into account in the calculation of fiber lifetimes. While hermetic coatings can be applied to eliminate strength degradation with time, they are rarely used. It is thus necessary to understand the time-dependent behavior well enough to ensure a desired lifetime (without an excessive and costly safety factor), or alternatively, it is necessary to validate a fundamental strength limit. This paper discusses the progress that has been made on issues related to such fiber lifetime predictions. In particular, it is pointed out that the possible variations in flaw character and coating chemistry make simple generalizations difficult and require that fundamental understandings be developed.

Stress corrosion in silica may end the lifetime of optical fibers in telecom networks. Since the lifetime demanded usually exceeds 25 years, a careful strategy is necessary in order to preserve cable reliability. The toolbox therefore includes equations for estimating lifetime of fibers, some of them related to the so-called B-parameter. The lifetime is proportional to B, but a value specific for the environment andfiber used has turned out to be difficult to achieve.

Among various experimental methods proposed, the mechanical testing at high stress rates is a physically sound way to confirm the B-value of coated fibers, since testing can be made under similar conditions as prevail during network application. Fatigue tests including high-speed tensile tests have been run on coated silica fibers for a long time. Lately, a new high-stress-rate bending method has been investigated, pushing forward the previous experimental limits.

Values of B from tensile- and bending tests are presented, being 0.004 - 0.10 GPa2 s in tension and considerably less in bending, about 0.0007 GPa2s. A study on some acrylate-coated silica fibers including draw-abraded fibers indicates that the B-value also depends on fiber strength.

Fused silica in service can suffer from strength degradation due to a localized contact load or particle impact that can cause cracking about the indentation or impact site. This cracking generally consists of radial, lateral, and cone cracks and is independent of whether the indenter or particle is sharp or blunt or whether the impact is subsonic or hypervelocity. The impact site is generally characterized by a shallow pit surrounded by an array of microcracks. The pit is formed by the fragmentation of the glass due to the intersecting radial, lateral, and cone cracks. With either static indentation or particle impact, it is the radial crack that controls strength degradation. The applicability of indentation fracture mechanics in predicting this strength degradation is discussed.

The dew point temperature (DPT) is defined as the temperature at which condensation will form from a gas containing moisture. The DPT, which is a measure of the water vapor reactivity, depends on the relative humidity (RH) and the temperature of the gas. The strength of optical fiber has been shown to have a dependence on the humidity and the temperature of the test environment. The results of one study imply that the strength over a range of temperature and humidity can be expressed as a function of the DPT only. While attractive for its simplicity, this idea suggests that the strength depends only on the thermodynamics of the equilibrium between the vapor and condensed phases of water and does not depend on the kinetics of the reaction between water and silica. In the work described here, the strength has been measured at various humidities and temperatures in order to test the dependence on dew point temperature. It was found that the strength degradation is not solely described by the DPT. The strength and dynamic fatigue data were also interpreted by a stress assisted chemical kinetics model. The degradation reaction order with respect to humidity at different air temperatures was found to be approximately two.

In order to achieve the optimum optical performance of fiber optic connectors, it is critical that there be good physical contact at the fiber-to-fiber interface. Loss of physical contact results in increased insertion loss and reflectance. Today's connectors are designed to have mating surfaces with tightly controlled endface geometries to ensure physical contact during temperature variations and repeated matings. However, material degradation of the fiber, ferrule, or the adhesive which holds the fiber into the ferrule, can change the endface geometry and cause eventual failure of the connector. Tests have been developed or are under development to quantify mechanical changes in these materials, including adhesive creep, fiber pistoning, ferrule deformation and roughening, and fiber deterioration. This paper will describe the latest test results acquired from applying these procedures.

Lucent Technologies manufactures and markets a family of SC and ST singlemode optical attenuator products which are low cost and easy to use. The attenuator elements are available in the attenuation range of 3 to 20 dB and are housed in buildout barrels which snap into companion buildout blocks to form complete optical couplings. The attenuator element is a 0.086-inch diameter thermoplastic material which, depending upon the attenuation level, has a thickness from 0.006 to 0.065 inches. In order to meet the full range of customer performance requirements, the attenuator element should not creep at temperatures up to 80°C and exhibit a reflection of less than -55 dB at optical telecommunication wavelengths. Commercially available thermoplastic materials cannot meet these requirements.

Very fine oxide dispersoids have been incorporated into solder alloys to create a new, improved solder structure with an ultrafine grain size of ˜200 -500 nm and significantly enhanced mechanical properties. The microstructure is seen to be quite stable upon high temperature exposure (e.g.. 120°C), which is attributed to the presence of very fine dispersoid particles which impede grain boundary sliding and dislocation movement. As much as three orders of magnitude reduction in the steady state creep rate has been achieved. The new solders also exhibit improved (4–5 times higher) tensile strength at low strain rates and improved ductility under high strain rate deformation. It is demonstrated that with a dispersion of TiO2 particles, the Pb-Sn eutectic solder with a low melting point of 183°C can be made more creep-resistant than the 80Au-2OSn eutectic solder with a much higher melting point of 278°C. The new creep-resistant solders can be useful for optical or optoelectronic packaging in which the dimensional stability of the assembled structure is essential.

Solder is often used as an adhesive to attach optical fibers to a circuit board. In this proceeding we will discuss efforts to model the motion of an optical fiber during the wetting and solidification of the adhesive solder droplet. The extent of motion is determined by several competing forces, during three “stages” of solder joint formation. First, capillary forces of the liquid phase control the fiber position. Second, during solidification, the presence of the liquid-solid-vapor triple line as well as a reduced liquid solder volume leads to a change in the net capillary force on the optical fiber. Finally, the solidification front itself impinges on the fiber. Publicly-available finite element models are used to calculate the time-dependent position of the solidification front and shape of the free surface.

Two types of plastic modules for laser diodes and photodiodes are developed and tested under various environmental conditions. The stable operations for both types of modules are confirmed under various long-term aging tests, high humidity high temperature tests (85°C, 85% RH), and temperature cycling tests (-40 / +85°C). These plastic moduless are potentially suitable for economic access networks.

Newly developed low loss, graded-index polymer optical fibers (GI-POF) are being investigated as high bandwidth media for data transmission in local area networks and for interconnection of large telecommunications and computer systems. These fibers are fabricated by distributing refractive index-raising dopants to form a core profile in a transparent glassy polymer fiber. The long term reliability of the GI-POF depends upon the degree to which the dopant remains immobilized in the polymer glass in service environments. This paper presents a method for assessing the diffusion kinetics of dopants in polymer optical fibers and for relating these measurements to GI-POF profile stability through numerical simulation. Simulated and measured index profile changes are compared to validate the approach.

The reliability assurance program for the BeamBoxTM thermo-optic solid state optical switches is presented. The different phases of the program are discussed, and some of the results are described.

Reliability and durability are essential for optical fibres in both telecommunication networks and sensor applications. We report on stress-free accelerated ageing of modem telecommunication fibres in a temperature range from 60 °C to 150 °;C in dry air, relative humidity (rh) up to 85 %, and deionised water. Fibre fracture stress was measured by dynamic tensile testing according to IEC 60793-1-3 standard requirements. Ambient humidity effects on fibre fracture strength are shown. High temperature and relative humidity ageing behaviour of fibres and acrylate coatings is discussed. Fibre fracture strengths were fitted as functions of ageing time and ageing temperature using single process Arrhenius-based models. Acceleration factors can differ by several orders of magnitude, depending on the activation energy of the process. High temperatures damage acrylate coatings, which might deteriorate its adhesion to the fibre, thus introducing new ageing mechanisms not experienced under service conditions. Nevertheless, short time ageing at high temperature is useful for estimating relative lifetime of different products and for optimizing production process.

The fatigue behavior of binary oxide glasses in the germania-silica system is interesting since it influences the long-term mechanical reliability of products made of these glasses. Dynamic fatigue tests were carried out in four-point bending in 100% RH using tumble abraded rods. The fatigue data were obtained as a function of GeO2 content which ranged from 0 to 24 mole %. In addition, other physical properties of these glasses, such as the thermal expansion coefficient, were measured as a function of GeO2 content. The results of these tests and analyses are highlighted in this paper.

There have been numerous studies of the fatigue and strength behavior of fused silica optical fibers. However, no coherent model has emerged that self-consistently describes the simultaneous effects of stress, temperature and activity of the corroding species (e.g. water). A power law degradation kinetics model (relating the crack growth rate to the applied stress intensity factor, KI) is widely used although various exponential forms based on chemical rate theory have also been proposed. The dependence of fatigue on parameters such as humidity, pH and temperature, has usually been treated in an empirical manner. Sometimes it is even ignored -for example, the service environment is often assumed to be the same as the proof test environment when making lifetime predictions, thus avoiding the need for understanding the humidity dependence; this assumption is often unjustified. This paper reviews the dependence of fatigue on environmental factors and highlights some of the inconsistencies in published data. It is then attempted to present a coherent kinetics model that simultaneously accounts for stress temperature, humidity, etc. Several possible forms of the model are compared to a range of experimental data of several different types. The comparison is made using fitting techniques that account for correlation between fit parameters. It is found that a simple exponential form of the degradation kinetics model gives the best overall description of the temperature, humidity and pH effects on static and dynamic fatigue. It should be noted that the exponential form predicts shorter lifetimes than the ubiquitous power law model. Therefore, under some circumstances, the predictions of “worst case” models based on power law kinetics are unduly optimistic.

Today, during their use, optical fibers are subject to harsher installation and environmental conditions. So to evaluate optical fiber lifetime with higher precision, it is necessary to study the mechanical behaviour of optical fibers in extreme conditions : high stress and harsh environment.

This paper presents the results of static bending tests in distilled water at temperatures between 20°C and 70°C and in buffer solutions with a pH between 4 and 9.The stresses applied on optical fiber were between 3 and 3.5 GPa.

We observed a dependence of the failure time on temperature. This dependence can be described by the Arrhenius model, where the activation energy is one of the main parameters. The n-value seems also to change regularly with temperature. The results obtained in buffer solutions show that the pH influence on optical fiber strength may depend on coating characteristics.

Median strength 4 GPa and Weibull slope 5 were obtained after stripping the coating from fibers with a mechanical stripping technique. However, the median strength was found to decrease to 1.6 GPa following fusion splicing. Most of the fractures in a tensile test occurred at a distance of 1 mm from the splice point. The internal axial tensile stress in the fiber cladding was found to be removed over a 2 mm section over the splice point. The surface fictive temperature was found to be decreased only at the zones where the fibers fracture. The reason for the fiber weakening during fusion is therefore thought to be due to the intrinsic structural changes in the silica indicated by the fictive temperature decrease. The magnitude of the decrease can be reduced by using a quicker fusion process. The corrosion velocity around a fusion splice point during aging in hot tap water was found to be significantly higher than along the fiber elsewhere. Silica layers with a thickness of hundreds nanometers were removed uniformly. However, the weakest points (roughest surface) after water aging were located in the ends of the recoated section.

Dynamic and static fatigue results are presented for the first time for a new type of optical fiber, whose modified sol-gel cladding is a fenestrated (micro porous) form of silica. Unjacketed fibers have mean Weibull strengths in bending of 6.5 to 7.6 GPa with Weibull slopes in the 40 to 60 range. The strength decrease with decreasing strain rate is similar for both jacketed and unjacketed fibers. Even the unjacketed fibers tested in ambient water or in boiling water retain a majority of their mean strength, 7.6 GPa (ambient air) versus 6.5 GPa (ambient water) versus 5.5 GPa (boiling water). Exposure to boiling water for 1 hour has no measurable affect on the strength of these fibers. Exposure to boiling water for 8 hours, however, does significantly broaden the low strength end for the unjacketed fiber. The dynamic fatigue and static fatigue parameters in ambient water are substantially the same, ND = 22±2 and Ns 21±4, for the jacketed and unjacketed fibers. Possible mechanisms are discussed to explain the strength and fatigue behavior of these fibers in light of their unique structure.