ITU-T profile of optical wavelength bands in telecommunications.
This chapter will provide background information in the development of BDFs and their applications in optical communication systems. Herein, the main focus is briefly described previous studies on BDFs that have attracted much interest over the last two decades. This necessary information and concepts are very much relevant to understanding this book, mainly due to the doping of Bi in the studied bismuth and erbium-doped silicate fibers (BEDFs). The remaining chapter is consisting of the following sections: Sec.2: General introduction about optical fibers. Sec. 3 discusses the general spectral characteristics of BDFs. Sec.4: Including the active centers (namely the bismuth (Bi) active centers (BACs)) responsible for the spectral properties in Bi-doped fibers. Sec.4 Discusses the Bismuth Doped Fiber Amplifier (BDFA).
- bismuth-doped fibers (BDFs)
- optical communication
- Bi active centers (BACs)
- bismuth doped fiber amplifier (BDFA)
A previous work by C.Kato et al. stated that the loss of 20 dB/km in a dielectric waveguide could be minimized if it uses pure dielectric material . This encourages many researchers to discover methods to minimize failure in optical silica fibers. Since its proper mechanical and optical properties, silica was used as the material for the improvement of optical fibers. A low-loss was established in 1970 by merging with a 20 dB/km loss .
Besides, with ultra-pure precursors with impurities of ppb-order (parts per billion), the growth of low-loss (0.2 dB/km) single-mode fibers about 1550 μm is demonstrated to be possible . Figure 1 displays the attenuating spectrum of conventional single-mode fiber. The wavelength range of 1260–1625 nm is split into several subwavelength bands. A specific ITU-T wavelength range is allocated to each band, as listed in Table 1.
After the introduction of low-loss fiber, erbium (Er)-doped (EDFA) technology was introduced in 1987 as an advance to revolutionize optical fiber communication. It allowed transatlantic fiber communication . The progression of fiber development to advance low-loss fiber in conjunction with EDFA technology led to the use of the low-loss range for optical fiber communications from 1530 to 1625 nm (C L band). Throughout the years, several techniques including dense wavelength division multiplexing (DWDM), wavelength multiplexing division (WDM), and coarse wavelength multiplexing division (CWDM) with new modulation formats as quadrature phase-shift keying (QPSK) as 16-QPSK, Quadrature Phase-shift Keying (QPSK) and 64-QPSK used to increase capacities for existing silica optical fibers, in the C + L band. Nonetheless, the stated transmission capacity of Standard Single-Mode Fiber (SMF) achieves the non-linear Shannon limit . Figure 2 shows growth in the global population’s use of the internet every ten years. More than 60% of the world’s population is projected to be connected by 2020. The enhanced internet connectivity and comprehensive Internet-based applications such as cloud computing, social media, e-commerce, and e-learning are the drivers of significant demand for fiber transmission. New approaches to improving the potential for existing optical fiber networks must also be considered economically rather than merely adding more traditional single-mode fibers to meet end-user needs. Many researchers have proposed many approaches. One of these approaches is using a low-loss window (1260–1625 nm) of silica optical fibers by improving useful fiber lasers and amplifiers. The other is the development of new fibers such as multi-mode fibers (MMF), multi-core fibers (MCF), and multi-element fibers (MEF) for space-division multiplexing (SDM) in a C band from 1530 to 1565 nm. A further challenge is looking for a different transmission band in a 2000 nm wavelength region with new fibers . The early stages of fiber optic communication in C-bands of about 1550 nm employed silica fibers with a loss of about 0.2 dB / km and a peak of about 1380 nm as shown in Figure 1. Recently, Lucent technologies and Optical Fiber Solutions (OFS) have delivered low-loss optical fiber in the entire 1260–1625 nm wavelength range. Such ultra-low-loss fibers are known as dry fibers and give a bandwidth of about 53THz for optical fiber. The mitigation for these ultra-low loss fibers in the 1260–1625 nm wavelength range is less than 0.4 dB/km, as shown in Figure 1 [4, 8]. Besides, new optical networks have been developed. For nearly every feature, they provide dramatically improved performance over traditional single-mode fiber, including increased usable bandwidth, superior macro-bend performance, and ultra-low dispersion polarization mode (PMD).
10 THz belongs to the Er amplification band between 1530 and 1625 nm with 53THz bandwidth provided by these dry fibers. A bandwidth can, therefore, be increased four times by using dry fibers. All of these dry fiber technologies have provided the ultimate flexibility in network construction and provided economic approaches to optimize data transition capability. However, the industry requires useful fiber amplifiers and lasers to use this utterly low-loss belt for optical communications, which are central to communication between an optical fiber. Regrettably, there are no active rare-earth (RE) doped fiber amplifiers and lasers that support the band of 1260–1530 nm. Given numerous attempts to implement various RE dopers in silica fibers to improve lasers and amplifiers in this wavelength range, potential devices available for use are still lacking. The amplifiers development and lasers in the 1260–1625 nm wavelength band, which uses bismuth (Bi) and Er-doped fibers to use the ultra-low loss window of ultra-low loss optical fiber.
2. Bismuth-doped fibers (BDFs)
Doping of the core zone with RE material is necessary for optical fiber fabricated to improve lasers and amplifiers. Various RE elements and their favorite emission bands in silica host are shown in Figure 3. Conventional RE elements, including ytterbium (Yb), Er, and holmium (Ho) or thulium (Tm), cover the wavelength bands about 1000, 1500, and 2000 nm . However, any RE-doped silica does not cover the band between Er and Yb. The Energy level diagram of Yb, Er, Tm is shown in Figure 4. neodymium (Nd) and Praseodymium (Pr)-doped fibers explored extensively in silica host to improve lasers and amplifiers about 1300 nm. Nevertheless, highly phonon energy in a silica host made the dopants ineffective. They changed to a low phonon energy host, as fluoride glass enabled them reasonably active. Fluoride glasses, though, are not appropriate for splicing with traditional silica fibers as required in many applications for all-fiber compact devices. Despite this, the fabrication of these fibers is complex and not mature sufficient, in contrast to conventional changed chemical vapor deposition (MCVD)-solution doping technique. Furthermore, as shown in Figure 4, in the case of Pr, ground-state absorption (3H4-3F4), excited-state absorption (ESA) (1G4-1D2) at the signal wavelength, and pump ESA (1G4-3P0) are prejudicial to improve effective lasers and Pr-doped fiber amplifiers at 1300 nm.
For neodymium (Nd), the ESA at a signal wavelength (4F3/2-4G7/2) with a get competition between 1005 nm (4F3/2-4F11/2) and 1.34 μm (4F3/2-4F13/2) are detrimental to improve 1.3 μm amplifiers and lasers . The last research subjects on BDFs covered the main areas of (i) design of BDFs with new materials and compositions, (ii) studying the fabrication techniques and processes to make BDFs with better properties, (ii) spectral characterization method to assess BDFs for desired applications, (iii) post-fabrication techniques and their impacts on spectral properties, (iv) applications of BDFs into gain media, or sensing device or any other, (v) evaluation of optical devices made with BDFs using various configurations and operational conditions, (vi) studying the nature of BACs responsible for producing the emission in BDFs, (vii) unraveling the energy levels of the active centers, and many other similar topics. Among these subjects, the nature of BACs is essential for understanding the spectral characteristics and their effects on applications; however, the issue has not been resolved yet and required more addressing.
3. Characteristics and applications of BDFs
3.1 Spectral emission properties and Bi active centers (BACs)
Visible emissions from Bi-doped glasses and crystals were deeply clarified [12, 13, 14, 15, 16, 17]. Recently, broadband emissions in the near-infrared (NIR) region from 1200 nm to 1500 nm have been demonstrated in Bi-doped glasses (BDGs) by Fujimoto et al. . Peng also reported broad NIR emission from Bi glasses. et al. as well . With around 1300 nm of optical amplification shown produced by the BDG.s . The first Bi-doped fiber (BDF) and a laser-based on that BDF were reported by Dianov et al. in 2005 . Different BDFs and BDGs have been improved for potential applications as lasers and amplifiers in the second telecommunication window range at around 1.3 μm. It was found that the bandwidth and intensity of emission produced from the BDFs are influenced mainly by the host type and other co-doping materials. Bi can produce emissions in various wavelengths in the NIR wavelength region over (1200–1500) nm, as reported in BDFs and BDGs. For instance, emissions in several BDFs and BDGs (such as silicate, germanate, alumino-silicate, alumino-phosphate, barium-aluminoborate types) have been reported in earlier work [19, 22, 23, 24, 25, 26, 27]. Typical emissions of Bi in some recent work are presented in Figure 5.
Emissions in BDFs and BDGs are accompanied by active centers or BACs. BACs largely determine the spectroscopic properties, performance, and operations of BDFs. Hence, to design a fiber for a specific application, it is essential to know the BAC details and their relationships to the spectral properties. As an approach to studying the nature of BACs can be broadly categorized into two main groups: (a) spectral analyses to determine the type of BAC and (b) instrumental analyses to identify the chemical bonding and electronic states of Bi, which is responsible for forming a particular type of BAC. The spectral analysis gives information about a specific BAC’s spectral characteristics, determined mainly by the composition types in the fiber or glass. Absorption, emission (luminescence or fluorescence), and emission lifetime are some of the most common and basic spectral properties analyzed to identify the BAC types. Besides, an emission-excitation spectroscopic graph described recently was found to be very important to recognize the particular emission . At the initial phase of Bi fiber research, it was thought that Bi emission could be produced only in the presence of Al . However, later, Bi emission was detected in pure silica-doped fibers without any Al co-doping, and lasing was realized in such BDFs . The concept that BACs are linked with other ions in addition to Bi was subsequently considered through the investigation into several Bi fibers and glasses, which were fabricated with different compositions and their relations. Following this advance, Bi fibers with the most straightforward glass compositions were made and BACs in them were characterized, enabling the identification of certain BAC types, e.g., BAC-Si, BAC-Al, BAC-Ge and BAC-P associated respectively with Si, Al, Ge, and P, along with their respective characteristic spectral properties [29, 30, 31, 32, 33]. Instrumental analyses of the BAC study concentrate on how to reveal possible valence and the electronic configuration of specific BAC. This approach included analyses of the nuclear magnetic resonance (NMR), X-ray photoelectron spectroscopy (XPS), electron spin resonance (ESR), X-ray diffraction, and extended X-ray absorption fine structure (EXAFS),
Although BAC-Si and BAC-Ge were described with much information about their kind and spectral characteristics, in-depth knowledge of BAC-Al and BAC-P lags far behind. BAC-Al and BAC-P showed notably different spectral behavior in BDFs than those of BAC-Si and BAC-Ge. One of the essential distinctive features is the Stokes shifts in the emission and excitation produced from these two centers, while for BAC-Si and BAC-Ge, no such Stokes shifts were pronounced . Electron-photon interaction giving increase to the detrimental emission processes joined with these two centers was expected as the reason for the Stokes shifts. Furthermore, BAC-Al was quite different from BAC-Si and BAC-Ge based on their susceptibility to electron irradiation . All of these specifications made it hard to discover the complete energy diagram of these active centers.
3.2 Luminescence properties of BDFs
The properties of near-infrared (NIR) broadband luminescence of Bi-doped silica fibers are sensitive to the composition of glass in addition to the conditions of fabrication. The partial shielding of the unfilled subshell 4f by filled 5 s and 5p in RE ions prevents significant interaction from the host environment. In comparison to RE elements, Bi has filled internal subshells, and external 6 s and 6p electrons interact significantly with the host, showing host-dependent absorption and emission properties. Co-dopants can also dramatically alter the spectrum of luminescence, as shown in Figure 7.
Herein, Bi-doped fibers with aluminosilicate hosts are described as Bi-doped aluminosilicate fibers (BASEs). They have shown a luminescence peak at 1150 nm. In contrast, Bi-doped fibers with phosphorus silicate and Germane silicate core composition hereafter described as Bi-doped phosphorus silicate fiber and Bi-doped German silicate fibers (BGSFs) are known to transfer the emission band to a longer wavelength side of about 1300 and 1450. Recently, by using high GeO2 concentration (50 mol%) within the fiber core, the Bi emission window is extending to a covered wavelength band of 1600–1800 nm. Luminescence in Bi-doped phosphor germane silicate fibers (BPGSFs) is broad compared to phosphorous (P) and other Bi-doped fibers co-doped with germanium (Ge). This makes it possible by changing the fiber core composition using Bi-doped optical fibers to show lasers and amplifiers for cover the full spectrum region about 1150–1800 nm. In Figure 6, the pump wavelength bands are referred to by the arrow. Wavelength bands of the pump about 1050, 1230, and 1310 nm for BASFs, BPSFs, and BGSFs, respectively, usually are utilized. In the case of BGSFs with a high concentration of GeO2, the pump wavelength band is about 1550 nm [39, 40].
3.3 Applications of BDFs
BDFs have many applications in different optical systems because of their wavelength based discriminative transmission losses . A 1651 nm single-frequency laser diode is providing with a Bi-doped fiber power amplifier, and its performance is evaluated experimentally . This amplifier is using to increase the performance of methane detection systems, such as remote stand-off systems or photoacoustic spectroscopy (methane has a molecular absorption line of almost 1651 nm). Two amplifier configurations are shown and output power of larger than 80 mW in both cases. The results obtained provide valuable perspective for compact and straightforward fiber amplifiers in the spectral region of ~1630–1750 nm with a power output of over 100 mW. The experimental set-up schematic diagrams are displayed in Figure 8(a) and (b). Two configurations are investigating, both with signal and pump radiation continues to using a WDM coupler to launch pump light at 1550 nm into the BDF and offer the amplified radiation output at 1651 nm. Standard, commercially available WDM coupler that is proposed to operating at 1625 nm and 1550 nm (from the Opto-Link Corporation). Its transmission is calculated to be ~95% at 1550 nm, but only ~65% at 1651 nm. A polarization-independent optical circulator is using in a second configuration (Figure 8(b)) broadband as an alternative a WDM coupler. The circulator (including from OptoLink Corporation) has been proposed for a wavelength of 1525–1610 nm. Nevertheless, its 1550 nm (port 1 - port 2) and 1651 nm (port 2 - port 3) transmissions were quite the same, almost 80 percent.
The Er-Doped Fiber Amplifier (EDFA) is currently one of the most crucial elements for various fiber optic systems . Nevertheless, driven by a continually growing demand, network traffic, which has grown exponentially for decades, would result in the overloading of these”capacity crunch” systems, since EDFA is limited to 1530–1610 nm in the spectrum. New technology will need to be pursued, and the development of optical amplifiers for new spectral regions can be a promising solution. Some of the fiber-optic amplifiers are made with materials doped with rare-earth. As a result, full bands for the gain band of Er-doped fibers are still available in shorter (1150–1530 nm) and longer wavelength (1600–1750 nm) regions. A novel fiber amplifier is operated at a spectrum region of 1640–1770 nm pumping via commercial laser diodes at 1550 nm. This amplifier is achieved by use Bi-doped high-Germania silicate fibers fabricating by an MCVD technique. The experimental set-up of a typical Bi-doped fiber amplifier is illustrated schematically in Figure 9. The BDFA is built using a scheme with Bi-directional pumping (backward and forward pumping). As a pumping source, commercial laser diodes are using with an ultimate output power of 150 mW each. The active fiber was core-pumped through commercial WDMs based on SMF-28. WDMs Transmission spectra are shown in Figure 8. Optical isolators are spliced to the amplifier of the input and output. The first is using to reduce the effect of amplified spontaneous emission (ASE) of a Bi fiber on the signal source. The second prevented possible lasing. Bold points act as splices, where an optical loss is ~1 dB because of the difference between the active fiber mode field diameter and that of conventional fibers (SMF-28).
The proposed dual pumping scheme of 830 and 980 nm aims at broadening and flattening Bi/Er multicomponent fiber (BEDF) spectral performance . The distinct BACs of germanium (BAC-Ge), aluminum (BAC-Al), phosphorus (BAC-P), and silique (BAC-Si) have spectral properties characterized by single pumping, respectively, of 980 nm and 830 nm. Depending on the BAC-Al (∼1100 nm) and BAC-Si (∼1430 nm) emission slope efficiencies under the single pumping of 830 and 980 nm, the dual pumping scheme with an optimizing pump power ratio of 25 (980 nm VS 830 nm) is determined to realize flat, ultra-broadband luminescence spectra covered the wavelength range 950-1600 nm. The dual pumping scheme is more illustrated on the on-off gain BEDF performance. The gain spectrum is flattening and broadening over 300 nm (1300-1600 nm) with an average gain coefficient of ∼1.5 dBm−1, due to the pump power ratio of ∼8 (980 VS 830 nm). The spectral covering is about 1.5 and 3 times more expansive compared to single pumping of 830 and 980 nm pumping, respectively. Due to the optical characteristic of the energy level diagrams of 830 and 980 nm, the advantage of dual pumping is clarified. The proposed dual 830 and 980 nm pump scheme with the BEDF multicomponent shows great promise potential in a range of broadband optic applications, such as NIR-band tunable laser, a standard ASE source, and broadband amplifier. The luminescence measurement set-up is shown in Figure 10. The 980 and 830 nm pigtailed laser diodes (LD) were launched into the input end of 3 dB 808 couplers, and the 810/1310 WDM, the BEDF, was spliced with the output end of 1310 beam with ∼1 dB splice loss. The backward optical spectral analyzer (OSA) was used to record the emission signal to remove the effect of residual pump power. To monitor the unabsorbed pump power, a digital power meter was placed at the end of BEDF, and A short length (∼40 cm) of BEDF was tested.
4. Bismuth doped fiber amplifier (BDFA)
Doped fiber amplifiers (DFAs) are optical amplifiers that use a doped optical fiber as a gain medium to amplify an optical signal. They are related to fiber lasers. The pump laser and the signal to be amplified are multiplexed into the doped fiber, and the signal is amplified through interaction with the doping ions. As an example is the (BDFA), where the core of a silica fiber is doped with Bi ions and can be efficiently pumped with a laser. 1120 nm diode-pumped Bi-doped fiber amplifier is fabricated by N. K. Thipparapu et al. . Bi-doped aluminosilicate fiber is fabricated using an MCVD solution doping and is distinguished by its gain and unsaturated loss. The amplifier performance was compared with the traditional pumping wavelength region of 1047 nm for a novel pumping wavelength of 1120 nm. Unsaturable losses in 1047 and 1120 nm pump wavelengths were 65% and 35%, respectively. At 1180 nm, the maximum gain of about 8 dB was observed for 100 m fibers with pumping at 1120 nm. The 1120 nm pump produced an enhancement of gain of 70% compared with the 1047 nm pump. An additional 3.5 dB gain was achieved at 1047 and 1120 nm simultaneous pumping. Figure 11 shows the schematic experimental set-up used for the measurement of gain in Bi-doped fiber. The set-up consists of an 1180 nm LD as input signal source and fiber pigtailed 1120 nm LD and/or 1047 nm Nd-YLF laser as the pump source. To abstain from the wavelength division multiplexers (WDMs) and back reflections to combining pump and signal sources, Isolators (ISO) was used. The spectrum of the output and input signal is taking through an OSA. The input signal was obtained little before the fiber under trial. In contrast, the output signal is determined by consideration of the WDM loss using to separating the pump from the signal.
O-band has recently been widely used for low-cost data transmission. The feature of O-band is that the transmitter wavelength(s) are laying close to the zero dispersive fiber wavelength (λ0), and no compensation is therefore needed for either optical or electronic chromatic dispersion. O-band transponder total bit rate is enhanced to 425 Gb/s, for instance, using 8 LAN WDM 26.6 Gbaud/s PAM-4 modulated channels . Using WDM and complex modulation formats minimize the power of both the receiver sensitivity and receiver per channel such that optical amplification is appropriate. O-band signal boosting can be achieved by employing Semiconductor Optical Amplifiers (SOA). However, they create considerable distortions due to the modulation of self- and cross- gain . Although most of dispersion broadened channels are amplified , SOAs are not approbated for intensity modulation formats transmission like PAM-4 operating near λ0 with relatively small channel count. Praseodymium doped fiber amplifiers (PDFA) with a bandwidth of 1280–1320 nm  were shown, but non-silica host glass was required, which makes PDFA complicated and costly. O-band amplification has been extensively studied in Bismuth-doped silica fibers . A 150 m long BDFA with bandwidths of 1320-1360 nm has been reported using a complex dual-wavelength pumping system with only 6 × 10 Gb/s OOK channels . Simple silica-based BDFA with 80 nm 6-dB gain-bandwidth flexibly centered within 1305-1325 nm, and parameters comparable to EDFAs is developed by V. Mikhailov et al. . The amplifier can extend 400GBASE-LR8 transmission (8 × 26.6 Gbaud/s PAM-4 channels) beyond 50 km of G.652 fiber. The active fiber core consists of phosphosilicate glass doped with Bi (>0.01 mol%) produced using the MCVD process. The preform was cladding by a Heraeus F300 tube while all the core components, including P, Si, and Bi, were deposited from the gas phase. The index difference between the fiber core and the cladding was approx. 6 × 10–3, and the cut-off wavelength was near 1.1 μm. As the core diameter of 7 μm of fiber offered reasonable slice capability with silica-based fibers, the regular automatic splicer was used to splice G.652and Bi- fibers. In order to investigate emission properties, 80 meters of Bi-fiber were subsequently pumped with a 3 dB wideband coupler with lasers of 1155, 1175, 1195, 1215, and 1235 nm (Figure 12).