OPTICAL FIBER DESIGNS FOR AMPLIFIERS AND SENSORS

Abstract

Over the last few decades information superhighway is paved with glass. The frustrating bottlenecks in the information superhighway are becoming less, day by day, allowing the services such as telemedicine, internet telephony, distance education, e-commerce, high-speed data and video conference. This scenario enhances the bandwidth demand in a geometric progression. To meet this over-whelming demand for bandwidth, the way is to introduce concurrency among multiple user transmissions into the optical communication network. This con-currency may be provided by the deployment of Dense Wavelength Division Mul-tiplexing (DWDM) technology, a new and crucial milestone in network evolution. DWDM referring essentially to the closer spacing of channels is the current fa-vorite multiplexing technology for long-haul communications, where the data rate within a single channel is terabytes per second. In long haul telecommunication link, use of erbium doped fiber amplifiers (EDFA) to compensate the fiber propa-gation losses, is limited by the chromatic dispersion which increases the bit-error rate. The signal distortion due to dispersion and optical non linearity can be avoided by an erbium-doped dual-concentric-core fiber and erbium-doped ultra large-core segmented cladding fiber. The key element that can potentially fur-ther revolutionize the effectiveness of DWDM network is rare earth doped fiber amplifier, specifically (EDFA). EDFA technology allows for hundreds of optical channels to be amplified simultaneously without electrical regeneration. As a result, the number of channels offered in commercially available DWDM systems Abstract has climbed from 8 to 160 within years. Unlike optoelectronic regenerators, an optical amplifier does not need high-speed electronic circuitry and has the ability to work at high data rates, which significantly reduces cost. In addition, ED-FAs also provide high gain, high saturation power, low noise figure, immunity to crosstalk and high operating bandwidth. Despite from these excellent characteristics, one problem with EDFA is the gain non-uniformity which causes divergence of DWDM channels in a long chain. Thus, a flatter gain spectrum across the whole usable bandwidth in EDFAs is desirable for DWDM optical network, which has been already manifested by the large current research exertion. Early wavelength division multiplexing (WDM) networks utilized the wavelengths between 1540 nm and 1560 nm, where the EDFA gain is somewhat flat. Next, to fully utilize the EDFA gain band 1530-1560 nm, tremendous efforts have been made in literature. In 1996, an attempt was made using alumino-germanosilicate EDF with a low Al content (1% by weight) to obtain an EDFA. In this fiber the gain spectrum has been flattened and the gain excursion was within + 0.3 dB for the wavelength band 1542-1552 nm. In addition, gain flattening filters (GFF), such as long period fiber grating filters (LPFG), fiber Bragg gratings (FBG), acousto-optic filter, side-polished fiber based filter, the fiber-loop mirror (FLM) filter, Mach-Zehnder (MZ) optical filter and split-beam fourier filters have evolved to flatten the amplifier gain spec-trum. Majority of these gain-flattening filters usually require complex procedures to tune their filter characteristics. Further these techniques would whittle down the original gain values in the effective operating range. In long haul networks, the wavelength dependent gain may increase the amplified spontaneous emission (ASE) powers that can self-saturate the amplifier and reduce the pumping effi-ciency. To tackle this problem LPGs written in EDF itself have been proposed . Such LPG significantly suppresses both forward and backward ASE and flattens the forward ASE spectrum. ii Abstract However, the explosion of information traffic and the need for flexible net-work requires the amplifier bandwidth to be wider to meet this vast demand on high capacity communication networks for DWDM long-haul applications. This extension involves additional transmission windows which are namely as short-wavelength band (S- band) that lies from 1460 to 1530 nm and long-wavelength band (L- band) that lies from 1560 to 1620 nm. Recently, ultrawide band gain flattened EDFAs combining both C- (1530-1560 nrn) and L- band have been with a bandwidth of 85 nrn in Wavelength range 1528-1613 nm demonstrated. Still, the non-uniformity of EDFAs gain spectrum in the long wavelength region (1460-1625 nm) becomes a real issue especially for long chains of amplifiers where even small variations in gain will grow into huge differences among individual channels. It has also been proposed that the use of an inherently gain flattened EDFA in the design of certain special networks like a metro network would signifi-cantly reduce installation cost. This motivated some investigators to investi-gate the designs for inherently gain flattened EDFAs. Some of these designs include highly asymmetric concentric dual-core fiber for C- band; co-axial fiber, depressed cladding fiber (DCF) and segmented cladding fiber(SCF) designs for S-band. Since the gain in S-band is highly non-flat ±10 dB gain ripple), sep- arate gain- flattening mechanisms need to be adopted for DCF design. Whereas in case of co-axial fiber and SCF designs, the noise figure is high and the design parameters are also very sensitive. In addition to wider bandwidth, low noise optical amplification is also the ultimate goal of optical amplifier research. In the present thesis we have proposed a novel leaky fiber design for gain flattening of EDFAs, that provide high gain, low noise figure (NF)covering 1490-1620 nm wavelength region. A low NF and wide-dynamic-range EDFA with an inherent flat-gain spectrum allows cost-effective mechanisms to cater to future DWDM networks. iii Abstract The main focus of the present thesis is to exploit the specialty leaky fibers such as dual-core resonant fiber as amplifiers for optical communication networks and for refractive index (RI) sensing applications. We present a novel co-axial dual-core resonant leaky fiber (DCRLF) fiber design for gain flattening of erbium doped fiber amplifier (EDFA) that covers an ultra-wideband including S-, C- and L-bands and then we extended its application towards RI sensing. In DCRLF, at resonance condition power couples from inner core to outer core, so that there is a sharp increase in the wavelength dependent leakage loss near the resonant wavelength. Such a spectral loss variation of the structure has been utilized to suppress gain peak and, thus, flatten overall gain profile in the C-band. For this we have carried out the comprehensive study of the tunable behavior of leakage loss of the DCRLF. We have analyzed the DCRLF using the transfer matrix method to calculate the leakage loss. The gain of the EDFA has been modeled by using a standard three level rate equation model. We achieved a 18 dB flat gain with ± 0.7 dB ripple utilizing two cascaded DCRLF filters. After designing the DCRLF based gain flattening filter for C- band, we have implemented this type of design for inherent gain flattening in S- band. Reso-nance tail of leakage loss of the fiber into the S-band region is utilized to suppress the amplified spontaneous emission (ASE) in the C-band and to lift the gain in the S-band of an EDFA. The structure provides high gain and low noise figure which is essential for long haul transmission systems. This fiber has shown 23.5 dB inherent flat gain with + 0.9 dB ripple over 30 nm bandwidth (1490 -1520 nm) and noise figure less than 3.5 dB. Behind this, we have also presented the use of DCRLF for gain equalization in L- band. With this design we achieved 18.5+2.0 dB inherently flat gain in 50 nm wavelength band (1570-1620 nm). Whereas in 25 nm wavelength band, the same fiber gives 19+0.9 dB flat gain. The use of DCRLF as a filter gives 15+1.0 dB flat gain in the region 1560-1590 iv