FDK Japanese

Home > Research & Development > Optical Isolators

Optical Isolators

Connectors and other optical components are present in the optical fiber transmission line, and diverse return beams are generated from the end faces of these components. Return beams are known to have a destabilizing effect on oscillation of the laser source and on operation of the optical fiber amplifier, thus resulting in poor transmission performance.
An optical isolator is composed of (1) a magnetic garnet crystal having a Faraday effect, (2) a permanent magnet for applying a designated magnetic field, and (3) polarizing elements whichpermit only forward light to pass while shutting out backward light. For this reason, optical isolators are indispensable devices for eliminating the adverse effect of return beams in high-speed optical fiber transmittance routes and amplifiers. By focusing on optical isolators with their operation wavelengths in the 1.30-1.55 um range used widely in optical fiber communication, we shall examine thier structures, mechanisms, and applications, along with the garnet crystals used in optical isolators.

2-1. Polarization-Dependent Type
Figure 1 shoows the composition of a polarization-dependent optical isolator. The optical isolator consists of two polarization elements (a polarizer and an analyzer having a 45° differential in the direction of their light transmission axes), and of a 45° Faraday rotator interposed between the polarization elements. A forward light passing the optical isolator undergoes the following: (1) when passing through the polarizer, the incident light is transformed into a linearly polarized light; (2) when passing through the Faraday rotator, the polarization plane of the linearly polarized light is rorared 45°; (3) this light passes through the analyzer without loss since its polarization plane is now in the same direction as the light transmission axis of the analyzer, which is tilted 45° from the polarizer in the direction of Faraday rotation.
In contrast, a backward light undergoes a slightly different process: (1) when passing through the analyzer, the vackward light is transformed into a linearly polarized light with a 45° tilt in the transmission axis; (2) when passing thruogh the Faraday rotater, the polarization plane of the backward light is rotated 45°in the same direction as the initial tilt; (3) this light is completely shut out by the polarizer because its polarization plane is now 90° away from the light transmission axis of the polarizer.
Structure of Polarization dependent optical isolator
Polarization-dependent isolator are primarily incorporated in semiconductor laser modules. Accordingly, miniaturization has been an important target. To realize this goal, PBS is being replaced by infrared polarization glass as a polarization element. Bi-substituted iron garnet films are used as Faraday rotators due to their low saturaton magnetization and large rotation capacity. Sm-Co type rare-earth magmets are employed to apply a designated magnetic filed to the Faraday rotator.
The performance of optcal isolators is primarily evaluated by thier insertion losses and isolations, both of which are detetmined by the absorption losses end-face reflectances, and the extinction ratios of optical elements. For this reason, technologies to enhance and actualize the characteristics of optical elements become important. At present, we have adhesive-free and metal-joined type single-tier polarization-dependent isolators measuring 3 x 3 mm in dimension, 0.2 dB in insertion loss, and 40 dB in isolation (see Figure 2).
Photograph of polarization dependent optical isolators
2-2. Pokarization-Independent Type
Figure 3 shows the composition of a polarization-independent optical isolator. A 45° Faraday rotator is interposed between two wedge-shaped birefringent plates, and a lens is placed at both ends of the isolator for junction with optical fibers. While polarization-dependent optical isolators allow only the light polarized in a specific direction, polarization-independent isolators transmit all polarized light. Consequently, these isolators are frequently used in optical fiber amplifiers.
First, a forward incident light is separated into ordinary and extraordinary rays by the No.1 birefringent plate. Second, the polarization planes of these rays are each rotate 45° by the Faraday rotator. Third, ordinary and extraordinary rays pass through the No.2 birefringent plate having such an optic axis that the relationbetween the two types of rays is maintained; consequently, both raysare refracted in an identical parallel direction when exiting from the No.2 birefringent plate. Fourth, these collomate beams are converged into the downstream optical fiber through a lens.
Structure of Polarization dependent optical isolator
On the other hand, a backward light incident on the same optical isolator is separated into ordinary and extraordinary rays whose relation is reversed with that of a forward light due to the non-reciprocality of the Faraday rotation. Consequently, rays passing through the No.1 birefringent plate do not become parallel to each other, so they cannot be converged into the upstream optical fiber.
Thanks to their superior extinction ratios and wide differences between the refraction indexes of ordinary and extraordinary rays, rutile single crystals (TiO2) are mostly used as birefringent plates. Non-spherical lenses, with a focal distance of several millimeters, are employed to link the isolator with oprical fibers.
At present, we have single-tier, single-mode-fiber-coupled and polarization-independent isolators measuring φ6.0 x 33 mm in dimension, 0.5 dB in insertion loss, approx.40 dB in isolotion, 60 dB in return loss, and 0.03 p/sec in polarization-mode dispersin(see Figure 4).
Photograph of polarization dependent optical isolators
2-3. Composite Type
An optical fiber amplifier is comprised of Er-doped (rare earth added) fibers, a wavelength-division multiplexer (WDM), a pumping diode laser, a polarization-independent isolator, and other passive components. Figure 5 illustrates the general composition of an amplifier. Previously, WDM and isolators were separately incorporated in amplifiers, but recently the sections enclosed in the broken lines (1) to (3) have been increasongly modularized, thus eliminating some of the lens systems needed for the been further miniaturized while preventing insertion loss from rising.
The types of optical fiber amplifiers are determined by the directions of signal light and pump light inside Er-doped fibers. The directions are either (1) identical, (2) reverse, or (3) both. They are called, forword pumping, backward pumping and two-way pumping, respectively. Dwpending on which of these three types of amplifiers is necessary, either (1), (2) or both of the modules shown Figure 5 are selected. In addition, there is another module (3) that is combined with a bandpass filter (designed to shut out pump light from the isolator) in order to shut out return beams from the optical circuit.
Schematic diagram of Er-doped fiver amplifier system
Figure 6 presents a composite type optical isolator for forward pumping. This isolator measures 25 x 30 x 8 mm in dimension,0.8 dB in insertion loss for both signal light (wavelength of 1.55 um) and pump light, 40 dG in isolation, and 55 dB or more in reflection attenuation.
Photograph of optical isolator module
3. Magnetic Garnet Crystal
3-1. YIG Single Crystal
A YIG single crystal is grown using the floating zone (FZ) method. Figure 7 shows the deagram of the FZ furnace used. Y2O3 and Fe2O3 are mixed to suit the stoichiometric composition of YIG, and then the mixture is sintered. The resultant sinter is set as a mother stick on one shaft in an FZ furnace, while a YIG seed crystal is set on the remaining shaft. The sintered material of a prescribed formulation is placed in the central area between the mother stick and the seed crystal in order to create the fluid needed to promote the deposition of YIG single crystal. Light from halogen lamps is focused on the central area, shile the two shafts are rotated. The central area, when heated in an oxygenic atmosphere, forms a molten zone. Under this condition, if the mother stick and the seed is moved at a constant speed, if result in the movement of the molten zone along the mother stick, thus gowing single crystals from the YIG sinter.
Since the FZ method grows crystal from a mother stick that is suspended in the air, contamination is precluded and a high-purity crystal is cultivate. The FZ method produces ingots measuring 012 x 120 mm. The specular condition off the coating finish and the trecision of the disk thickness determine the characteristics of optical isolators.
Structure of FZ furnance
3-2. Bi-Substituted Iron Garnet Thick Films
Bi-substitued iron garnet thick films are grown by the liquid phase epitaxy (LPE) method. Figure 8 illustrates the structure of an LPE furnace. Crustal materials and a PbO-B2O3flux are heated and made molten in a platinum crucible. Sigle crystal wafers, such as GdCa)2(GaMgZr)5O12, are soaked on the molten surface while rotated, which causes a Bi-substituted iron garnet thick film to be grown on the wafers. Currently, thick films measuring as much as 3 inches in diameter can be grown.
To obtain 45° Faraday rotators, these films must be ground to a certain thiclness, applied with anti-reflective coating, and then cut into 1-2 mm squares ti fit of the isolators. Having a greater Faraday rotation capacity than YIG single crystals, Bi-substituted iron garnet thicl films must be thinned in the order of 100 um, so higher-precision processing is required.
Structure of LPE Furnance
3-3. Comparison of YIG Single Crystals and Bi-substituted Iron Garnet Thick Films
YIG single crystals fare better than Bi-substituted iron garnet thick films in temperature coefficient and wavelength coefficient for Faraday rotation. If these caracteristics are given higher priority, YIG single crystals are selected as Faraday rotators. On the other hand, the advantages of Bi-substituted iron garnet thick films include: (1) the time required to grow crystal is greatly shortened,(2) the length of rotators can be reduced to one-fifth or even less due to increased rotation capacity, and (3) the saturation magnetization si low.
These features lead to a price reduction and miniaturization of optical isolators; therefore, Bi-substituted iron garnet thick films are mostly used as polarization-dependent isolators incorporated in laser modules. Table 1 below compares the characteristics of YIG single crystals and Bi-substituted iron garnet thick films.
Property Wavelength(nm) YIG single crystal Bi-substituted iron garnetthick film
Material - Y3Fe5O12 (TbBi)3(FeAl)5O12
Saturation magnetization (mT) - 178 60
Insertion (dB) 1310 0.1 0.1
Faraday rotation coefficient (deg./cm) 1310 224 -1570
1550 175 -1060
Faraday rotation temp. coefficient (deg./°C) 1310 0.034 0.054
1550 0.042 0.062
Faraday rotation wavekength coefficient (deg./nm) 1310 0.056 0.089
1550 0.040 0.064
Extinction fario (dB) 1310 > 38 > 41

4. Conclusion
Ultra-speed and large-capacity optical fiber trunk systems are expanding as a result of the development of optical fiber amplifiers. In parallel, the demand for optical isolators is increasing. Demand is also expected to increase, as LAN and other subscriber optical fiber networks expand. It is therefore imperative that isolators and other optical compenents be further improved to achieve higher performance,smaller size, and lower price. Currently, however, a vital part of optical component production depends on human skill and know-how, so this poese limitations on the ability to meet callenge, therefore, is to develop new techonologies for the production of isolators and other optical components.