Eric Mayer and Dirk Basting, Lambda Physik USA Inc.
Fiber Bragg gratings (FBGs), which incorporate sophisticated optical functionality in compact, rugged, fiber-based packages, offer significant advantages to designers of both telecommunications and sensing systems. The explosive growth in the use of FBGs has led to a tremendous research effort aimed at characterizing all aspects of the FBG production process, as well as at extending the capabilities of the various production techniques. Laser-system manufacturers are playing a major role in this effort by offering systems specifically optimized for FBG production.
A fiber Bragg grating consists of a longitudinal, periodic variation in the refractive index in the core of an optical fiber. When light propagates through a FBG, Bragg diffraction causes one wavelength to be selectively reflected. The wavelength at which high reflectivity occurs is determined by the period of the grating.
Because FBGs are an integral part of the fiber itself, they deliver a high level of mechanical convenience, simplicity, and economy to the user. These characteristics have made them very attractive to designers of tele communications systems. Their ability to selectively separate (by reflection) closely spaced wavelengths, for example, makes them a useful component of add/drop and other wavelength-division multiplexing devices. Also, FBGs can perform dispersion compensation in fibers and gain flattening in amplifiers
Another emerging application for FBGs is sensing. Fiber sensors exploit the high sensitivity of the FBG reflected wavelength to changes in the effective grating spacing. Mechanical strain and thermal expansion in structures, such as airplane frames, oil tanks, and bridges, can be measured precisely from changes in the reflected spectrum caused by physical deformations of the embedded fibers. New research is attempting to extend the utility of FBG sensors by characterizing their response to bending from shear forces.
Fiber Bragg gratings are produced by exposing a step-indexed, germanosilica fiber to intense ultraviolet (UV) light, typically from a krypton fluoride (KrF) excimer laser (at 248 nm) or frequency-doubled argon-ion laser (at 244 nm). Absorption of this light in the germanium-doped core causes a permanent change in the refractive index of the fiber. There are three basic techniques currently in use that can produce the required high-frequency index modulation with the necessary accuracy (see Figure 1).
In the interferometric method, a single laser beam is split into two components, which are subsequently recombined at the fiber to produce an interference pattern. The primary disadvantage of this technique is that the interference-fringe spacing and placement is highly sensitive to the optical alignment of the system. Furthermore, maintaining adequate fringe contrast requires high mechanical stability and isolation from ambient vibration. Finally, this ap proach cannot be used to create variably spaced (chirped) gratings. An advantage, though, is that the technique is flexible, allowing grating parameters to be changed quickly.
The phase-mask approach utilizes a diffraction grating to split a single laser beam into several diffractive orders. Interference between the various orders creates the required pattern in the fiber. The phase-mask technique does not have the flexibility of the interferometric method but is far less sensitive to vibration and alignment, making it generally more suitable for production environments. Additionally, the use of phase masks enables the production of chirped gratings.
The third FBG fabrication approach utilizes a mask-projection technique. In this method, a laser beam is homogenized and passed through a mask. This illuminated, striped pattern is then projected at a high reduction ratio onto the fiber. The advantage of mask projection is that it can be used to produce virtually any type of complex periodic or even nonperiodic structure. But, unlike with the other methods, it is difficult to achieve resolutions that produce submicron features.
Recent advances in FBG production methods are allowing fabrication of longer FBGs. The methods described here typically can produce FBGs that range up to several centimeters in length. Some telecommunications applications, however, would benefit from longer gratings. Wavelength dispersion in long-haul fibers,
for example, limits performance capabilities of high-bit-rate systems. This dispersion can be corrected with a long, chirped FBG in which wavelength reflectivity is designed to vary along the grating. The design allows "faster" wavelengths to travel farther into the FBG before being reflected back and recombined, thus compensating for dispersion. For each nanometer of signal bandwidth in a 10-Gbit/sec system, roughly about 100 mm of FBG length are required to compensate for the dispersion experienced over 80 km of fiber transmission. This leads to a requirement for FBGs of several meters in length for typical telecommunications systems.
One way to produce longer gratings is with a step-and-repeat approach, which involves making an exposure via the phase-mask technique, then using precise positioning equipment to index the fiber for a subsequent exposure. A fiber Bragg grating of virtually any length can, in theory, be "stitched" together using this method. In reality, the total travel of the high-resolution positioners currently available limits the maximum achievable length to a few meters. Most positioners are also unable to exactly match the beginning of each grating to the end of the previous one within the required submicron-positioning tolerance.
Now, though, researchers at Nortel Networks (Harlow, England) have been able to greatly increase the accuracy with which separate exposures can be stitched together.1 The group developed a technique based on the fact that fibers fluoresce when excited by UV light and that the unexposed region of fiber, between grating fringes, will produce greater fluorescence than the exposed area. A low-intensity probe beam is passed through a phase mask identical to that used for writing, guaranteeing that the probe fringe pattern exactly matches the refractive index profile of the FBG. When the mask is moved relative to the fiber, the resulting sinusoidal variation in fluorescence reveals the phase of the grating and position with high accuracy.
In this particular case, the phase mask consists of a series of smooth chirp profiles. Each patch on the phase mask begins with a 5-mm-long segment that is identical to the last 5 mm in the previous patch. After writing one segment, the fiber or phase mask is moved exactly one patch length, and the probe beam is then directed through this 5-mm section for positional registration via the fluorescence signal. This alignment process takes about 2 sec and yields stitch errors of less than 5 nm for gratings of up to 2.5 m in length.
A different approach taken by investigators at 3M Telecom Systems Div. (Austin, TX) eliminates the need for high-precision positioners or any position-feedback systems.2 The basic optical setup for this approach is interferometric. In this case, however, the fiber is run over a spool mounted on a rotary stage and translated through the stationary interferogram at a constant velocity. The writing laser is then sinusoidally modulated to continuously write the FBG. The advantage of this method is that a precise fiber travel velocity can be maintained with a large flywheel on the spool. Furthermore, a chirped grating can be produced this way by simply changing the laser modulation frequency so that it is not exactly in synchronization with the spacing of the interferogram. This technique has already yielded FBGs of more than 10 m long and should be scalable to even longer structures.
The creation of a new generation of FBGs has also required development of novel laser sources. The ideal laser for phase-mask FBG writing methods exhibits high UV output power, excellent energy stability, high spatial coherence, and high beam-pointing stability. Traditional excimer lasers supply plenty of UV power, but have limited coherence.
One approach to reducing divergence and increasing spatial coherence in lasers with large volume gain media, such as excimers, is to use an unstable resonator. However, the beam expansion that occurs in an unstable resonator can lower beam intensity relative to the amplified spontaneous emission (ASE). Typically, spatial filtering is used to eliminate the effects of ASE, but the high power density at the focus of an excimer laser makes this problematic because it can damage the aperture.
Our L1-FBG excimer laser addresses these problems by specifically optimizing the laser for FBG writing. High coherence is required along one axis, so the oscillator is based on a cylindrical unstable-resonator (CUR) design to expand the beam in one axis only (see Figure 2). This provides a high degree of spatial coherence along the critical long axis, while minimizing gain losses. In fact, the CUR oscillator reduces losses by an order of magnitude when compared with a spherical unstable resonator of the same magnification. This robust system combines an integrated single-dimensional spatial filter before power stabilization to provide unsurpassed pulse-to-pulse energy stability, a superior beam profile, and a minimum of ASE. The resulting spatial coherence of greater than 1 mm meets the needs of typical FBG writing schemes (see Figure 3).
- H. Rourke et al., Optical Society of America 1999 Technical Digest, 321 (1999).
- J. Brennan and D. LaBrake, Optical Society of America 1999 Technical Digest, 35 (1999).
Eric Mayer is product manager and Dirk Basting is president and CEO of Lambda Physik USA Inc. (Ft. Lauderdale, FL).
This article appeared in the April 2000 issue of Laser Focus World, Lightwave's sister publication.