Laser modulation offers choices for optical communications networks
Laser modulation offers choices for optical communications networks
Direct, external and electroabsorptive modulation techniques for laser light sources offer users a myriad of application choices
Wei Shin Tsay
To compete viably in the fiber-optic communications market, system manufacturers are differentiating themselves more and more through incremental features and migration strategies to next-
generation products. They can no longer distinguish their optical transmission products based on proprietary transmission systems in the globally standardized environment of synchronous optical network/synchronous digital hierarchy, or Sonet/SDH.
One of the key differentiating features is the need for longer unrepeatered span length. There is an emerging need for span length that is longer than the current interface standards (greater than 100 kilometers), especially for emerging markets such as Australia, China and Saudi Arabia.
System customers also mandate a clear migration path to the next-generation, higher-capacity system. To enable the next generation of fiber-optic systems to transmit at the OC-192 rate of 10 gigabits per second, time-division multiplexing, or TDM, is expected to be used to directly modulate a single optical carrier. However, for now, wavelength-division multiplexing, or WDM, systems that transmit multiple optical carriers down a single fiber (for example, four wavelengths, each running at 2.5 Gbits/sec) appear more practical.
Depending on an SDH system manufacturer`s product differentiation and migration strategies, and the life cycle of current products, three laser-modulation technologies offer different benefits and pose distinctive challenges. These technologies--directly modulated distributed feedback, or DFB, lasers; externally modulated continuous-wave lasers; and electroabsorptive-modulated lasers, or EMLs--offer system manufacturers tradeoff choices in terms of performance, cost, availability, ease of use and upgradability.
All three modulating technologies can be used in the pursuit of longer transmission-span length and higher-capacity transmission. Several component manufacturers are perfecting laser device designs and fabrication processes to ramp up manufacturing volume and move along the learning curve of these devices. To that end, component suppliers, system manufacturers and system customers can benefit from working together to realize the optimal laser devices.
Component manufacturers that can introduce--in a timely manner--products with these capabilities at equitable prices are figured to attain a competitive market edge.
Traditionally, two primary techniques are implemented for modulating optical signals into the fiber: direct modulation and external modulation. The direct modulation of a laser source by electrical signals is mainly used for shorter-distance links (to 100 kilometers) because of simpler transmitter design and component technology. External modulation using a lithium-niobate modulator allows transmission over distances exceeding 100 km.
Recently, researchers have begun combining the best features of both modulation techniques to create EMLs, which have a monolithically integrated laser and modulator. These laser devices have been available in the laboratory for several years. Until recently, though, their performance has been inconsistent because of immature design and production technologies. However, improvements in manufacturing techniques are making EMLs a viable option for field deployment.
For short-distance applications operating at 2.5 Gbits/sec or less, the directly modulated laser has become the technology of choice. However, it presents a transmission limitation at 1550-nanometer communications because of dispersion effects in the optical fiber.
Dispersion causes optical signals of different wavelengths to travel at different speeds down the fiber. This effect degrades the signal and limits the system bit rate and the fiber link length. It becomes a problem, however, only for 1550-nm communications because most of the installed fiber possesses zero dispersion at 1310 nm.
Directly modulated lasers tend to have wider spectral width due to chirping; the optical energy is spread out over a range of wavelengths. As these different wavelengths travel down the fiber, dispersion spreads them out in time, making the signal more difficult to distinguish at the receiver.
Most directly modulated lasers are used for links that are under 100 km. However, laser manufacturers have succeeded in improving the technology and are now achieving fiber distances longer than 200 km. These low-dispersion-penalty lasers can be manufactured using existing, field-qualified fabrication processes.
A prime advantage of these high-performance directly modulated lasers deals with their shorter time-to-market because usually no transmitter redesign is needed. These laser types also cost less and are available in production quantities because they are based on a mature technology.
For very-long-distance links, externally modulated lasers have proven successful due to their small spectral chirping. These devices commonly consist of a continuous-wave laser module interconnected to a lithium-niobate modulator via a polarization-maintaining fiber pigtail. A critical parameter for the continuous-wave laser modules is a stable polarization-extinction ratio over the ambient temperature range.
Field applications demonstrate that externally modulated systems can transmit optical signals over distances greater than 500 km for 2.5-Gbit/sec systems. In addition, they can pack more wavelengths into a WDM network, thanks to reduced chirp. Cost-effective 10-Gbit/sec WDM systems are currently being deployed for commercial applications.
Lithium-niobate devices themselves are generally larger--4.7ٲ.5ٲ.4 inches--than the common butterfly packages used for directly modulated lasers. More space is taken up by interconnecting the device to the laser via the fiber pigtail.
Currently, the cost of manufacturing lithium-niobate modulators typically exceeds that for making distributed feedback lasers because it is a less-mature technology. However, the higher costs are expected to decrease as market competition and product volume increase.
Lithium-niobate devices tend to attenuate optical signals more than do other modulation technologies. The lithium-niobate modulator itself presents a loss of 4 decibels. In operation, the device produces another 3-dB loss. On the positive side, these losses are not limitations because higher-powered lasers can be used and the signals are often optically amplified before they travel down the fiber.
An electroabsorptive modulator monolithically integrates a DFB laser with a voltage-controlled optical modulator. The laser section is operated with constant current, thereby reducing frequency chirp, while the modulator provides high-speed light modulation. In principle, the electroabsorptive modulator combines the low-chirp characteristics of the lithium-niobate device and the small size and low cost of the directly modulated DFB laser. This combination produces an efficient light source for high-bit-rate, long-fiber-link communications systems.
Because the modulator and the laser reside on the same chip, the complexity and materials needed to package and interconnect the two as required by lithium-niobate devices are eliminated, decreasing size and costs for an optical transmitter.
An electroabsorptive modulator laser can later be extended to 10-Gbit/sec applications due to its low chirp and ease of integration. It suits both TDM and WDM systems.
Electroabsorptive technology, however, faces several challenges. Because the laser and modulator sections are fabricated at the same time, manufacturing constraints can affect simultaneous performance optimization of both sections. These constraints also burden manufacturing process control. In addition, the physics of the indium-gallium-arsenide-phosphide material generates a lower output power level for the electroabsorptive modulator device than that of a directly modulated device.
These devices also require a different transmitter design. EMLs are voltage-driven, whereas directly modulated lasers are current-driven.
Over time, the cost of EMLs is expected to decrease, thereby increasing their use in short-distance links. For these short-haul links, less testing should be needed for EMLs because of their intrinsically lower chirp. In the long run, as their performance improves and cost decreases, the EMLs might take the place of both directly modulated and externally modulated devices for both long and short links.
For WDM applications, tunable EMLs across a range of wavelengths will be very desirable. They can greatly simplify the manufacture, installation and maintenance of the communications system.
Directly modulated lasers are presently the lowest-cost lasers available. They are commonly used for short-range links to 100 km at 2.5 Gbits/sec. Moreover, refinements in directly modulated DFB laser production have pushed their maximum range beyond 200 km.
For 2.5-Gbit/sec links, externally modulated laser are being deployed successfully in longer links, especially in the WDM configuration. The electroabsorptive modulator lasers are also becoming a practical choice for communications networks from 200 to 600 km, even though these devices are still in the early stage of their manufacturing learning curve.
For 10-Gbit/sec TDM links, EML and externally modulated laser devices will be capable of links to 100 km. Advanced electroabsorptive modulator laser and lithium-niobate devices are being developed in the laboratory for system deployment this decade. u
Wei Shin Tsay is product manager of digital laser modules for the optoelectronics strategic business unit of AT&T Microelectronics, Breinigsville, PA.
Electroabsorptive Modulated Laser Is Grown in a Single Process
All of the components of the electroabsorptive modulated laser, including the modulator, the laser and the isolation region in which the trench that separates them is formed, are grown continuously in a single selective growth process. Consequently, there is no optical gap between them.
The trench is etched into the surface of the chip to provide electrical isolation between the distributed feedback laser and the electroabsorptive modulator section. During operation, a separate electrical connection is made to the top of the laser and the modulator. A common electrical ground is attached to the bottom of both sections.
The light is generated in the multiquantum well layers of the distributed feedback section and passed to the modulator section. The period of the grating provides wavelength control for the laser. When a positive current is applied to the laser, it emanates a continuous signal down the optical cavity.
The electroabsorptive modulator is a voltage-controlled device. With no electrical signal, it is transparent to the light. When a negative voltage is applied to the modulator section, it turns opaque and absorbs light from the laser section, much like a shutter.
The entire device is mounted on a good thermal conductor to remove excess heat. Although the laser is only powered with a few tenths of a watt of power, it is so small that even this amount of power could pose a problem without a heat sink. The entire device is 750 microns long and approximately 100 microns thick.