CHAPTER 2: OPTICAL SIGNAL DEGRADATION ‘As an optical signal pulse traveling inside a fiber, there are several factors that can degrade the data transmission. The longer distance an optical pulse goes the less chance the data can get to the receiver end; the faster a pulse is being transmitted the worse the information can be recognized successfillly. These are due to the attenuation and dispersion of a propagating lightwave. The attenuation effect decreases the signal power and the dispersion effect distorts the shape of the pulse as a lightwave propagating down a fiber. The mechanisms causing these effects are discussed in this chapter. 2.1 OPTICAL SIGNAL ATTENUATION Signal attenuation [29] is a very important property in the design of a fiber optical communication system, because it largely determines the maximum transmission distance between a transmitter and a receiver. There are three basic mechanisms causing signal attenuation in a fiber; they are absorption, scattering, and imperfection losses of the optical energy [30-32]. Absorption loss can be classified as two types: intrinsic absorption and extrinsic absorption. The intrinsic absorption is due to the material nature of absorbing specific wavelength regions of light. The intrinsic absorption occurs in both the infrared and ultraviolet ranges. Fortunately, these intrinsic losses are mostly insignificant in the region where fiber systems are operated, but these losses limit the extension of fiber optic ‘communication toward the ultraviolet as well as toward longer wavelength. Extrinsic absorption is caused by the atomic resonance of impurity particles in the fiber. The most important extrinsic absorption is due to water or the hydroxyl ion (OH) bond. Because the bond can absorb incident light at its resonant frequency and harmonics, there are 28absorption peaks at wavelengths of 2.8/(n*+1) jm. For example, the first, second, and third overtones of absorption peaks are at 1.40, 0.93, and 0.70 ym (n=I,2,and 3) [33,34]. ‘There are four kinds of scattering loss in optical fibers: Rayleigh, Mie, Brillouin, and Raman scattering [35]. Rayleigh is the most important scattering loss. During the ‘manufacture process of glass fibers, some localized variations in density may happen due to random motion of molecular. These material density variations may be modeled as small scattering objects embedded in an otherwise homogeneous material. Because these objects are much smaller than the operating wavelengths, when a beam of light passing through these objects, some of its energy is scattered and lost. The scattering-loss dependence is indicated in Fig. The Rayleigh scattering loss can be approximated by the expression: L (0.85, 70854) Qn where 2. is in micrometers and the loss L is in dB/km [36]. As a result, the scattering loss is proportional to 4 *. Therefore, the use of short wavelength in fiber optic communication is severely restricted by Rayleigh scattering, Figure 2-1, Rayleigh scattering, showing attenuation of an incident stream of photons owing to localized variations in refractive index. 29The imperfection loss includes bending, coupling, and splicing losses. Bending loss occurs whenever an optical fiber undergoes a bend of finite radius of curvature. When a fiber is bent, partial energy radiates away through the evanescent field tail in the cladding. Generally, bending loss is not significant and can be neglected, unless the bending curvature is too large or higher than the order of 1 mm’ [37]. Usually, a light signal is also attenuated at a junction of two connected fibers either by a coupler or splicing. The loss is caused by some extrinsic or intrinsic reasons. Extrinsic reasons include misalignment, tlt, end gap, or bad end face quality. Intrinsic reasons are core ellipticity, mismatch in refractive index, or mismatch in mode field diameter. Typically, coupling loss is around 0.2 dB and splicing loss is around 0.05 dB (38). 2.2 MATERIAL DISPERSION As well known, when a light wave travels in a vacuum, it moves at a velocity of ¢ x 10* rms. In any other medium, light waves travel at a slower speed, given by v=c/ 1, where n is the index of refraction of the medium, For the material used to make an optical fiber, the refractive index varies with the wavelength of light traveling inside the fiber. Therefore, a different wavelength (or say color) of light travels at a different speed inside a fiber. The term “dispersion” is used to describe the phenomenon of wavelength dependent velocity of propagation of an electromagnetic wave. When the velocity variation is caused by some property of the material, the effect is called material dispersion. The material dispersion effect can be explained by considering the situation showed in Figure 1-9. A finite linewidth optical source emits a pulse into a dispersive slass fiber. For simplicity, assume the input pulse is composed by three different single 30wavelengths: 1.1, 12, and 43. The tree pure color components travel at different velocities (due to wavelength dependent refractive index) in the fiber. After propagating a distance, they arrived at different time in the receiver end. As a result, the output pulse, the sum of three received single wavelengths, becomes spreading. In a long enough fiber span the dispersion can be sufficiently large so that the adjacent pulses will overlap eventually, this results in inter-symbol-interference (ISI) and producing a high bit error rate in communications. Input Optical Putse Oxiput Optical Pulse Dispersive Optical Fiber —. za — _—_— 0 T 7 eT AL a a a T ¢ eer aw _N A -__ _ o T v ver a a ae ee ee Figure 2-2, Pulse spreading caused by propagation through a dispersive optical waveguide, The complete pulse contains wavelengths A1, A2, and 13; each has different speed due to wavelength ‘dependent indices of refraction. 2.3 WAVEGUIDE DISPERSION Another basic type of dispersion is called waveguide dispersion, which is usually ignored in multimode fiber applications. When a light signal is coupled into a single- 31mode fiber, only about 80 percent of power is confined into the core, and the other 20 Percent of power propagates in the cladding layer [39]. Since the core and cladding have different refractive indices, the two modes of light travel at different speeds. Since the light travels faster in the lower refractive index materials and slower in higher refractive index materials, the light propagating in cladding travels faster than the light confined in the core. Dispersion is then produced. The amount of waveguide dispersion depends on the design of fiber. Waveguide dispersion is a function of the core radius, the refractive index difference between core and cladding, and the shape of the refractive index profile. In fact, the waveguide dispersion can be carefully designed to cancel out the material dispersion at particular wavelength in single-mode fiber design. The resulting optical fibers are known as dispersion-shifted fibers [40-43]. 2.4 MODAL DISPERSION Modal dispersion is a problem only when a multimode fiber is used. Multimode fibers allow different paths or modes for the light to travel in them. The various modes intend to interact with each other in the big fibers. Since different modes propagate at different angles, each of them has a different axial group velocity along the fiber [44]. In other words, they travel at different speeds. This variation in the group velocities of the different modes results in a group delay spread or inter-modal distortion. The modal dispersion limits the speed and distance of an optical communication link. However, this dispersion mechanism can be eliminated when a single-mode fiber is used. That is why some form of single-mode fiber is always used in systems needing the highest speeds and longest spanning. 322.5 POLARIZATION MODE DISPERSION Polarization mode dispersion (PMD) describes a situation in which the electromagnetic wave components that make up an optical signal travel at different speeds within the fiber. This causes a multipath interference at the receiver. PMD is difficult to predict and may possibly vary with temperature and environment, the twisting of the cable as it was pulled, and even between production runs from the same manufacturer. The very high-speed systems that are soon to be deployed (i.e., 10 Gbps OC-192) are more prone to failing in the presence of significant levels of PMD. Some modeling has been performed recently that suggests that PMD will only be a problem for long haul systems (i.e., > 500 km). The analysis took into account that link components such as dispersion shifted fiber (DCF) and EDFAs contribute their own degree of PMD. In addition, Coming Incorporated has performed field tests on various vintages of installed Corning fiber and is confident that its standard single-mode product (SMF-28) has a low enough PMD such that it will support OC-192 on existing routes. Keep in mind that existing routes are limited to distances requiring regenerators rather than the cascaded EDFAs those long haul carriers plan to deploy to maximize bandwidth and economize on signal regeneration. Bell Atlantic has been working with Bellcore to examine the embedded fiber for PMD problems; however, it may not pose any practical difficulty until long haul WDM systems are constructed. 2.6 DISPERSION COMPENSATION As described in previous section, attenuation and dispersion effects can significantly limit the bit rate and the spanning distance of fiber optical communication, 33‘The war against attenuation can be won because the improvement of fiber manufacturing and the invention of EDFA. However, dispersion effects have to be taken into consideration as well. Since PMD is rarely observed, modal dispersion is taken cared by using single-mode fiber, and waveguide dispersion can be controlled by fiber design, it is the material dispersion usually referred as the main factor of limitation of optical ‘networks. In this section, several important fiber technologies used to provide dispersion compensation are described. 2.6.1 DISPERSION SHIFTED FIBER ‘The standard single-mode fiber deployed today is manufactured to optimize transmission at 1310 nm by effectively eliminating dispersion at that wavelength. The dispersion in the 1550 nm window far exceeds that for 1310 nm on standard fiber and hence is a limiting factor in single channel or DWDM systems operating in that window. Dispersion shifted fiber (DSF) differs from standard fiber in that the zero dispersion point is shifted from 1310 nm to 1550 nm by constructing a single-mode fiber with a triangular-shaped refractive index variation (instead of a step-index or graded-index variation) [45]. It is best suited for applications involving single channel transmission at 1550 nm providing the benefits of zero dispersion as well as taking advantage of the lower attenuation occurring at that wavelength. WDM systems do not perform as well as single channel systems on DSF due to a phenomenon known as four-wave mit Because a fiber's refractive index is nonlinear, two or more optical carriers can combine and produce several mixing products. This has a cascading effect and can result in unwanted products occurring at the operating carrier wavelengths. This process is more 34intense when there is zero dispersion at the operating wavelengths because the unwanted products will be moving at the same speed (i.e., in phase) with the desired signals causing significant interference and thereby hampering system performance. There have been some successes with techniques that involve wavelength allocations that avoid mixing products occurring at the signal wavelengths; however, four-wave mixing remains a concer for WDM using DSF. 2.6.2 NON-ZERO DISPERSION SHIFTED FIBER In order to support WDM systems, a fiber was developed that lowered the chromatic dispersion at 1550 nm but not to the extent that would encourage four-wave mixing as DSF does. This fiber is called Non-Zero Dispersion Shifted Fiber (NZDSF) because of the small non-zero amount of dispersion that occurs in the 1550 nm window. By now it may be apparent that there exists a sort of "Catch-22" situation with TDM and WDM. systems. A large telecommunications provider desires to maximize the TDM rates placed on as many WDM optical carriers as possible to maximize the bandwidth of fiber optic facilities. Very high speed TDM requires zero dispersion; however, WDM or DWDM requires some small amount of dispersion to avoid mixing effects. One philosophy is that by a careful combination of NZDSF and standard single-mode fiber in the same link, the benefits of local non-zero dispersion for WDM are realized while also achieving zero dispersion for the total link so that very high TDM rates are possible. 2.6.3 DISPERSION COMPENSATING FIBER. Another philosophy is to use standard single-mode fibers in combination with a new type of fiber. DCF (dispersion compensating fiber) is a new specialty fiber that has a very 35high negative value of dispersion [46]. It can actually reverse the effects of chromatic dispersion suffered by 1550 nm signals that traverse standard single-mode fiber. It is used a5 a sort of inline pre- or post-equalization in the form of a fiber spool of a particular length placed at one end of a link. However, DCF is not an optimal dispersion compensating solution for DWDM systems, because itis efficient for only single wavelength. For a DCF integrated DWDM system, channels located farther from the central wavelength of the DCF get less compensating amount than channels located near the central wavelength. Another disadvantage of DCF is that it cannot handle high optical power due to its relative small effective area. In DCF designs, an increased negative dispersion is achieved by shrinking the core diameter of the fiber. A fiber with a small core diameter experiences high light intensity, or high power per unit area, and, consequently, a system utilizes the fiber will suffer a high level of non-linear effects. Because DCF’s high non-linear penalties, its optical power tolerance is limited. 2.7 SUMMARY To meet growing demand for bandwidth at the lowest possible cost, long haul carriers are migrating to next generation optical networks characterized by high channel counts of 200 or more, high bit rates of 10 Gbps or 40 Gbps, and link distances of thousands of kilometers. Chromatic dispersion has emerged as a key obstacle for system designers in the path to the next generation optical network. The ideal solution should be able to correct dispersion at different wavelength (different channel in DWDM system) and provide continuous flexibility to accommodate network changes and upgrades (for example, migrating from 8 to 16 channels). There are two main technologies providing 36channelized dispersion management: fiber Bragg gratings (FBG) and virtually imaged phase array (VIPA), and the FBG based solution is what this research focus on. 37