Optical fiber is central to the successful operation of a wide variety of high-speed communications applications, from local area networks (LANs) to long-haul systems to emerging metropolitan area networks. Each of these applications has unique technical requirements, for which a specific optical fiber has been developed to provide optimized performance and high value.
Every optical fiber has the same two basic elements: a core and a cladding. The core is the inner, light-guiding section; the cladding surrounds the core. Because the refractive index of the core is higher than that of the cladding, the index mismatch causes total internal reflection at the interfacelight that enters the core at an angle reflects off the core/cladding boundary and propagates down the length of the fiber. Over distance, the signal experiences some attenuation. Nevertheless, all optical fibers combine low signal loss with very high bandwidth and are lightweight, small, have high tensile strength, and are immune from electromagnetic interference. modes in multiples
There are two main designs for communications-type optical fiber: multimode fiber and singlemode fiber. Each is the size of a human hair, with identical outer diameters of 125 µm. The important difference is in the size of each fiber's core. The core of singlemode fiber is 7 to 10 µm in diameter, while the core diameter of a multimode fiber is much larger, generally 50 µm or 62.5 µm. Core size is critical to how a fiber transmits data.
Light propagates down the fiber core in a stable path known as a mode. Multimode fiber supports hundreds of modes in the core, each of which is a different length (see figure 1). If we launch a single pulse of light into a fiber, the light will excite multiple modes, entering at various angles to bounce off the core/cladding interface. In other words, the light in each mode will travel a different distance depending on the modal path, so the light in some modes will arrive at the far end of the fiber later than others. The pulse will be spread out in time, a phenomenon known as modal dispersion. If multiple pulses are launched into the fiber and they all suffer modal dispersion, adjacent pulses may overlap until the receiver cannot distinguish one pulse from another, which causes bit errors.
Figure 1. Multimode fiber supports light propagation along a large number of paths of varying lengths (top). Grading the index of the fiber can partially correct for modal dispersion (bottom).
Multimode fiber can be optimized to reduce modal dispersion using a graded refractive index profile in which the refractive index of the core glass decreases slowly as a function of the radial distance from the center of the fiber. The lower the index of refraction, the faster light travels; therefore, light travels slower at the center and faster toward the core-clad interface. If this graded index profile were perfect, all modes would arrive at the receiver simultaneously. In practice, modal dispersion can be minimized but not eliminated, and it is the principal bandwidth-limiting factor in multimode fiber.
Nevertheless, with more than sufficient bandwidth for current and future applications, multimode fiber is ideal for the LAN environment. Moreover, its large core enables the use of inexpensive light sources such as light-emitting diodes or vertical-cavity surface-emitting lasers (VCSELs). These sources, combined with new laser-optimized multimode fibers, provide the means to cost effectively transmit data at gigabit speeds over distances required by LANs.
Multimode fiber is available in two core sizes: 50 µm and 62.5 µm. While 62.5 µm has been widely adopted in North America, 50 µm fiber is becoming increasingly important in global premises cabling because it offers up to three times the bandwidth of standard 62.5 µm fiber at 850 nm, the current operating wavelength for VCSELs. The combination of laser-optimized 50 µm multimode fiber and VCSELs represents the lowest-complexity, lowest-cost option for high-speed premises networks as they move steadily into multigigabit data rates. staying single
In singlemode fiber, the core is so narrow that it only supports one mode, eliminating modal dispersion effects. The index profile is a step function that gives good reflection from the core/cladding interface (see figure 2). Singlemode fiber works very well for long-distance communication because it can transmit signals quite far before signal strength diminishes to the point at which amplification is required. In telephony applications, for example, amplifiers typically are placed at intervals of 80 km or more.
Figure 2. Singlemode fiber has a narrow core that only supports one mode (left/top). It can have a step index profile (right/bottom).
The primary bandwidth-limiting effect for singlemode fiber is chromatic dispersion, the spectral spreading of an optical pulse. Light at shorter wavelengths travels down a fiber more rapidly than light at longer wavelengths. Because all laser sources have some finite spectral bandwidth, an optical pulse spreads as it propagates down the fiber, undergoing chromatic dispersion. If pulses spread too much, intersymbol interference and signal-to-noise ratio degradation occur: The receiver cannot distinguish 1's from 0's, and bit errors result.
For years the workhorse of telecommunications networks has been standard singlemode fiber, which is optimized for the minimum dispersion window of 1310 nm. This is a useful wavelength, because high-capacity, low-cost components are readily available for operation around that wavelength. However, signal loss of optical fiber is lowest around 1550 nm, an important wavelength region because it is also in the operating range of erbium-doped fiber amplifiers (EDFAs), which goes from 1520 nm to 1610 nm. EDFAs allow network operators to boost the power of multiple wavelengths simultaneously and extend distance between electronic regenerators. They have been crucial in the development of dense-wavelength-division-multiplexing (DWDM) systems for long-distance transmission. dealing with dispersion
In the mid-1980s, manufacturers introduced dispersion-shifted singlemode fiber (DSF) to take advantage of the 1550 nm transmission window by altering the composition of the fiber core to shift the zero-dispersion wavelength to 1550 nm, where signal loss is lowest. The fiber worked well for single wavelength systems at 1550 nm, but DSF created serious problems for the DWDM systems introduced shortly after.
Local dispersion refers to dispersion of the fiber at a defined wavelength as measured in ps/nm/km. A small, finite amount of dispersion helps to offset optical pulses that would otherwise travel together uniformly. This offset helps to counter the effects of cross-channel nonlinear interactions called four-wave mixing (FWM), in which channels interfere with each other and generate noise. When a network transmits optical signals at different wavelengths, the absence of local dispersion drastically enhances FWM. The amount of power introduced into the fiber also plays a critical role in affecting FWM (and other nonlinear penalties). More power means more nonlinear penalties.
Nonzero dispersion-shifted fiber (NZ-DSF) overcomes FWM by moving the zero-dispersion wavelength outside the 1550 nm operating window. The practical effect of this is to have a small but finite amount of dispersion at 1550 nm, which minimizes four-wave mixing while allowing for 10 Gb/s transmission over distances spanning several hundred kilometers without the need for costly dispersion compensation.
Nonlinear effects are more pronounced at high data rates, however, which challenge even conventional NZ-DSF. Increasing the effective area of a fiber is one method for addressing this problem. Effective area refers to the equivalent area of the fiber in which optical power is transmitted. In the case of singlemode fiber, this is roughly proportional to the core area. Fiber with a large effective area offers reduced optical power density, raising the power threshold for nonlinear penalties. in the metro loop
Metropolitan area networks often consist of rings, or a system of rings, with a combined circumference in excess of 80 km. The principal limiting factors are dispersion, network complexity, and cost.
Although they are an inexpensive alternative, directly modulated (DM) lasers introduce positive transient chirp, which is a slight increase in frequency over time as the pulse is transmitted that causes each pulse to contain a range of wavelengths that travel at slightly different velocities. By adding a negative dispersion profile in the optical amplifier window (1530 nm to 1625 nm), NZ-DSF can be tailored to extend the uncompensated transmission distance of DM lasers without the need for expensive dispersion compensation equipment and regeneration along the network path. The negative dispersion generates pulse compression at the beginning of the signal path, and leads to path lengths exceeding 300 km for DM laser signals.
Optical fiber technology continues to keep pace with speed and bandwidth demands that grow exponentially year to year. A broad selection of multimode and singlemode fibers meets the specific needs of various network applications, providing the means to transport massive amounts of information rapidly and economically. oe
Alan Dowdell is product line manager for high data rate optical fiber at Corning Inc., Corning, NY.