A Bragg fiber design with potential for applications in metro networks is proposed for the first time. The average dispersion of the designed fiber is 10 ps/km.nm in the C-band, and in view of its estimated loss being very low, such a Bragg fiber should enable ultra low-loss DWDM transmission over 100 km at 10 Gbits/s. A Bragg fiber based metro network is an attractive proposition because it would not require any amplifier and dispersion compensator for distances ≈100 km. This should significantly reduce installation and operational cost, and complexity of a metro network.
©2005 Optical Society of America
The rapid growth in Internet usage and increasing demand for data services and network efficiency, has led to the development of high-speed metro-networks. A metro network is expected to address features like low installation cost, high degree of scalability, and dynamism that is capable of accommodating unpredicted traffic growth. In addition to voice, video and data, it should also be able to support various protocols such as Ethernet and 10 Gigabit Ethernet, and provide connectivity to a variety of equipments like cell phones, ATM and legacy equipments. With these desirable features, metro-specific fibers have been developed in recent years [1, 2, 3].
It is now well established that in data transport, dense wavelength division multiplexing (DWDM) is the key enabling technology for capacity upgrades of existing as well as new fiber cables. Most DWDM systems use Erbium doped fiber amplifiers (EDFA) for amplification of signals at multiple wavelengths in the C- (1530–1565 nm) and L- (1570–1610 nm) bands. Typically, an EDFA is required after approximately every 80 km in a long-haul network. However, in current metro networks based on SONET/SDH, the signal is regenerated at each node because information is exchanged electronically at the node . Thus, relatively shorter distances are encountered in a metro environment, which seems to make optical amplifiers unnecessary at first sight. However, the large number of add/drop nodes that occur in a metro ring could contribute significantly towards large signal losses, which may necessitate signal amplification. Moreover, futuristic metropolitan networks are likely to be transparent, with flexibility to route/drop off signals at any node in the network, thereby requiring the signal(s) to travel along the entire ring. Thus, the signal(s) should be able to propagate without degradation through distances typically ranging between 120–200 km, before being regenerated. The larger fiber lengths, to be encountered by the signal (at bit rates ≥10 Gbits/s) in such futuristic transparent networks, would tend to increase the required number of network components. Essentially, dispersion and loss would be the primary dictating factors in the choice of the fiber for both current SDH systems or futuristic metro networks that would limit the maximum signal reach. Accordingly, a transmission fiber that would not require insertion of amplifiers and dispersion compensators for distances of ≈100 km or more, would be attractive because it would reduce the overall network cost and complexity. This motivated us to exploit the inherent low-loss feature of a photonic bandgap Bragg fiber and tailor its dispersion characteristics to make it suitable for deployment in transparent metropolitan networks. To the best of our knowledge, this is the first proposal in establishing the potential of Bragg fibers for deployment in transparent metro networks with significant benefits over conventional silica fibers.
Bragg fibers have of late attracted a great deal of interest due to the possibility of guiding light through an air-core by the photonic bandgap effect . The concentric cladding layers of alternate high and low refractive indices that surround the core of a Bragg fiber form a radially periodic structure, which gives rise to the photonic bandgap (Fig. 1). Consequently, light of those frequencies, which correspond to the bandgap, can be confined within the core. Radiation loss and dispersion of the propagating mode depend on the core size, refractive index contrast as well as thickness of the cladding layers. Thus, Bragg fibers offer a wide choice of parametric avenues to tailor their propagation characteristics. The desired characteristics could be achieved through a judicious choice of these parameters.
2. Fiber design
The uncompensated signal span length needs to be maximized in order to minimize the number components in a metro network This has led to the development of metro-specific ITU-compliant G.655/G.655b type non-zero-dispersion shifted fibers (NZDSFs) by well-known fiber manufacturers like Corning  and Alcatel . The reported metro-specific fibers have been designed to operate either as negative dispersion fibers [1, 3] or as positive dispersion fibers [2, 7, 8], across the C-band, with an average |D|≈8 ps/km.nm. These designs can achieve a span length of ≈100 km at 10 Gbits/s, without the need for a dispersion compensating device. The network is, however, loss-limited for all the above-mentioned fiber designs and the signals are required to be amplified after approximately every 80 km. Our proposed design offers the possibility of realizing ultra-low loss transmission, thereby minimizing investment cost on amplifier head, besides providing dispersion figures comparable to those of the available metro-centric NZDSFs.
Bragg fibers are highly polarization sensitive, and a vector approach is required to solve for the eigen modes supported by them. However, recently a semi-asymptotic matrix approach involving much less complexity and computational time has been applied with success . In this theory, the field in a certain number of inner layers is assumed to be exact Bessel solutions while the field in the rest of the outer layers is assumed to have the asymptotic form of Bessel functions. The Bloch theorem is then applied to set up the eigen value equation, solutions of which give the effective indices of the guided modes. In the present paper, we apply the same formalism to design a Bragg fiber, which should be useful in metro networks. Since the proposed fiber has a large air-core and cladding layer refractive index values that are sufficiently higher than that of air, the semi-asymptotic matrix method could be employed to determine the mode effective indices accurately .
The refractive index contrast and corresponding thickness of the cladding layers of a Bragg fiber are critical in deciding its dispersion and propagation loss characteristics. For a given number of cladding layers, the quarter-wave stack condition (Eq. 1) minimizes the radiation loss of the TE01 mode of the Bragg fiber .
where , j=1,2; λ0 is the central wavelength of the bandgap corresponding to the cladding layers, and (l 1, l 2) and (n 1, n 2) are the thickness and refractive index of the cladding layers.
Typically, a Bragg fiber based on the above-mentioned quarter-wave stack condition exhibits very large positive dispersion at the wavelength, λ0, which is undesirable for metropoliton applications. However, we observed that the dispersion of the fiber reduces towards zero, as the operating wavelength is reduced. Below the zero-dispersion wavelength, the fiber becomes highly dispersive and lossy. Thus, to achieve small dispersion that is imperative for metro networks, we envisaged operating the fiber at wavelengths smaller than the central wavelength corresponding to Eq. 1. Accordingly, we chose the cladding thickness of the fiber such that it satisfied the quarter-wave stack condition at a wavelength λ′0, which is slightly larger than the desired operating wavelength. To fulfill the simultaneous requirement of small dispersion slope, we chose a large core radius of 10 µm. In this context, we should mention that this choice of the core radius is not restrictive, though it is more difficult to achieve small dispersion slope with smaller core radii. Large core radius reduces the dispersion slope of the fiber structure, while increasing the zero-dispersion wavelength. Thus, for operation in the C-band, we chose the thickness of the cladding layers in accordance with the quarter-wave stack condition, satisfied at 1800 nm (=λ′0). Subsequently, we fine-tuned the dispersion and dispersion slope by increasing the thickness of the high-index layer so as to achieve positive dispersion with reduced dispersion slope.
As already mentioned, negative dispersion fibers (e.g. Corning’s MetroCor™ fiber) are the other option available for achieving increased uncompensated span length in a metropolitan network. The negative dispersion of these fibers serves to cancel out the positive chirp inherent in most of the directly-modulated DFB lasers, which are commonly used as the light source in optical networks. Thus, the light pulse carrying the signal can propagate through a longer distance in the fiber without broadening, as compared to the distance that it would have travelled in a standard single-mode fiber (SMF). Consequently, span length ≈100 km, without a dispersion compensating device, can be achieved with such a fiber design. For longer lengths, standard SMFs could be used for dispersion compensation. However, due to the small positive dispersion coefficient of SMFs, the required SMF lengths would be relatively long, which would result in increased loss penalty. Nevertheless, negative dispersion metro fibers are an attractive option. It was recently reported that highly flattened dispersion could be achieved in Bragg fibers by introducing a defect layer in its otherwise periodic cladding . For comparison purpose, by exploiting a similar route, we report here a negative non-zero dispersion shifted Bragg fiber, whose dispersion spectrum is comparable to that of the MetroCor ™ fiber.
3. Results and discussion
As mentioned in the previous section, the radius of the air-core of the proposed fiber is 10.0 µm, and the cladding consists of 20 bi-layers, having refractive indices of 2.8 and 1.5. The choice of these indices was driven by the fact that they correspond to materials, which have recently been used in the fabrication of Bragg fibers . Figure 2 shows the dispersion spectrum of the lowest order TE mode of the proposed fiber.
It can be seen from the figure that the average dispersion of the fiber across the C band is 10.4 ps/km.nm with a dispersion slope of 0.17 ps/nm2.km at 1550 nm. This should enable a dispersion-limited fiber length of ≈96 km in the C band at 10 Gbits/s , assuming dispersion power penalty of 1 dB (for 2 dB power penalty, it could be ≈175 km). Its average dispersion across the L-band is 15.9 ps/km.nm with a dispersion slope of 0.09 ps/nm2.km at 1590 nm. The accumulated positive dispersion of the NZDSF after installation in a network system can be easily compensated by conventional dispersion compensating fibers or through another Bragg fiber based dispersion compensator . Differential loss suffered by higher order modes of the fiber were found to be orders of magnitude larger than the TE01 mode, thus making the fiber effectively single moded. Moreover, the TE01 mode of the Bragg fiber is non-degenerate, which makes the design inherently free from polarization mode dispersion. The strong confinement of the TE01 mode within the air-core allows the material-related losses in the cladding to be ignored as compared to the radiation loss . Wavelength dependence of radiation loss (of TE01 mode) across the C and L bands, are tabulated in Table 1. As can be seen from the table,
the average radiation loss of the TE01 mode across the C band is only 0.03 dB/km while it is nearly an order of magnitude lower across the L band. To achieve efficient optical coupling, the Bragg fiber can be tapered to attain significant modal overlap between its TE01 mode and the LP 01 mode of the fiber pigtail of a laser diode. Additionally, the air-core of the Bragg fiber enhances the threshold for optical nonlinearity while simultaneously allowing a large mode effective area as compared to conventional fibers - both these features are attractive for large power throughput with low sensitivity to detrimental nonlinear effects. Large power throughput serves to offset the signal distribution losses at the nodes of a metro and/regional network rings, which is advantageous from the point of view of power budgeting options.
The negative dispersion Bragg metro fiber has been designed by introducing an aperiodic layer in the cladding. It has a core radius of 40 µm with cladding thicknesses of 0.33 µm and 0.19 µm, respectively (for the same material system used above). The thickness of the second layer of this Bragg fiber is chosen to be twice the thickness of the corresponding layer in the periodic cladding, i.e. 0.38 µm. This Bragg fiber exhibits dispersion of - 9.4≤D≤- 0.2 ps/km.nm within the wavelength range 1520–1600 nm (Fig. 3), thereby yielding average dispersion of -3.3 ps/km.nm and dispersion slope of 0.13 ps/nm 2.km, in the C-band. In comparison, Corning’s MetroCor™ fiber aimed at metropolitan networks, exhibits a total dispersion of -10≤D≤- 1 ps/km.nm, over the wavelength range of 1530 - 1605 nm. Thus, the dispersion characteristics of our proposed negative-dispersion Bragg fiber are similar to the MetroCor ™ fiber and should yield a dispersion-limited span length of ≈100 km. Average loss of MetroCor ™ fiber in the C-band is ≈0.5 dB/km, implying a loss-limited fiber length of ≈75 km, after which an amplifier is necessitated. The proposed Bragg fiber exhibits negative dispersion due to interaction between the core mode and the defect mode. Consequently, the field must penetrate into the defect layer. Fallout of this is that it may result in larger radiation and material-related loss for the structure. This can be investigated through actual fiber fabrication. Nevertheless, the distinct advantage of increased uncompensated fiber span length (≈100 km devoid of any loss or dispersion compensator) offered by the positive dispersion-flattened Bragg fiber proposed above, may be difficult to accomplish in the case of the negative dispersion-flattened Bragg fiber.
We have designed a positive dispersion-flattened Bragg fiber suitable for metropolitan networks having an average dispersion of ≈10 ps/km.nm in the C-band and ≈16.0 ps/km.nm in the L-band of the EDFA. The estimated average radiation loss of the fiber across this entire wavelength range is ≈0.02 dB/km. This should allow a fiber span length of ≈100 km without requiring any dispersion compensating device and amplifier. These features should significantly reduce the installation and operational costs, and complexity of the metropolitan network. Moreover, detrimental nonlinearity issues like four wave mixing in DWDM systems, and PMD could be ignored.
S. Dasgupta thanks the Council of Scientific and Industrial Research, India, for the award of Shyama Prasad Mukherjee fellowship. B. P. Pal thanks B. Culshaw, University of Strathclyde, for his keen interest in the work.
References and links
1. I. Tomkos, B. Hallock, I. Roudas, R. Hesse, A. Boskovic, J. Nakano, and R. Vodhanel, “10-Gb/s transmission of 1.55-µm directly modulated signal over 100 km of negative dispersion fiber,” IEEE Photon. Tech. Lett. 13, 735–737 (2001). [CrossRef]
2. J. Ryan, “Fiber considerations for metropolitan networks,” Alcatel Telecom. Rev. 1st quarter, 52–56 (2002).
3. T. Okuno, H. Hatayama, K. Soma, T. Sasaki, M. Onishi, and M. Shigematsu, “Negative dispersion-flattened fibre suitable for Gbit/s directly modulated signal transmission in whole telecommunication band,” Elec. Lett. 40, 723–724 (2004). [CrossRef]
4. P. Yeh, A. Yariv, and E. Marom, “Theory of Bragg fibers,” J. Opt. Soc. Am. 68, 1196–1201 (1978). [CrossRef]
5. D. Culverhouse, A. Kruse, C. Wang, K. Ennser, and R. Vodhanel, “Corning MetroCor fiber and its application in metropolitan networks,” (2002). White paper at http://www.corning.com/docs/opticalfiber/wp5078_7_00.pdf.
6. “Alcatel 6911, TeraLightTM Metro Fiber,” (2002). http://www.alcatelcable.com/Products/Fiber/data-sheets/6911_ds_rev0.pdf.
7. M. Tanaka, T. Okuno, H. Omori, T. Kato, Y. Yokoyama, S. Takaoka, K. Kunitake, K. Uchiyama, S. Hanazuka, and M. Nishimura, “Water-peak-suppressed non-zero dispersion shifted fiber for full spectrum coarse WDM transmission in metro network,” in OFC 2002 Technical Digest, (Optical Society of America, Washington, D.C., 2002), pp. 171–173.
8. I. Sogawa, “Study on full spectrum directly modulated CWDM transmission of 10 Gb/s per channel over water-peak-suppressed nonzero dispersion shifted fiber,” in Proc. of ECOC 2002.8.2.1 (Copenhagen, Denmark, 2002).
9. Y. Xu, G. X. Ouyang, R. K. Lee, and A. Yariv, “Asymptotic matrix theory of Bragg fibers,” J. Lightwave Tech. 20, 428–439 (2002). [CrossRef]
10. S. Guo, S. Albin, and R. S. Rogowski, “Comparative analysis of Bragg fibers,” Opt. Express 12, 198–207 (2004). http://www.opticsexpress.org/abstract.cfm?URI=OPEX-12-1-198. [CrossRef] [PubMed]
11. T. D. Engeness, M. Ibanescu, S. G. Johnson, O. Weisberg, M. Skorobogatiy, S. Jacobs, and Y. Fink, “Dispersion tailoring and compensation by modal interactions in omniguide fibers,” Opt. Express 11, 1175 (2003). http://www.opticsexpress.org/abstract.cfm?URI=OPEX-11-10-1175. [CrossRef] [PubMed]
12. S. D. Hart, G. R. Maskaly, B. Temelkuran, P. H. Prideaux, J. D. Joannopoulos, and Y. Fink, “External reflection from omnidirectional dielectric mirror fibers,” Science 296, 510–513 (2002). [CrossRef] [PubMed]
13. Y. Danziger and D. Askegard, “High-order-mode fiber - an innovative approach to chromatic dispersion management that enables optical networking in long-haul high-speed transmission systems,” Opt. Networks Mag. 2, 40–50 (2001).
14. S. Dasgupta, B. P. Pal, and M. R. Shenoy, “Design of a Low Loss Bragg Fiber with High Negative Dispersion for the TE01 Mode,” in Frontiers in Optics 2004 Technical Digest, FWH49 (Rochester, USA, 2004).
15. Y. Xu, A. Yariv, J. G. Fleming, and S. Y. Lin, “Asymptotic analysis of silicon based Bragg fibers,” Opt. Express 11, 1039–1049 (2003). http://www.opticsexpress.org/abstract.cfm?URI=OPEX-11-9-1039. [CrossRef] [PubMed]