OVERVIEW

From its inception in 1962, the semiconductor laser based on GaAs and related III–V compounds has evolved into an extraordinary and versatile optoelectronic device. Today we find this ubiquitous, relatively low-cost, high-efficiency coherent light source used in a wide spectrum of applications ranging from long-haul optical communications systems to high-density optical data storage. The evolution of these devices to their present state of sophistication – with wall-plug efficiencies of 50% and greater in the conversion of DC electric current to coherent photons (as one illustrative characteristic) – did not happen overnight, however. Nearly two decades of research and development were spent before the edge-emitting GaAs-based diode laser emerged as a mature optoelectronic technology suited for applications and integration. Once the basic guiding scientific principles were established in the design of these devices, making specific use of optical and electronic confinement in heterostructures, a major challenge remained at the basic materials science level. This challenge concerned the role of crystal defects, both point and extended-state in nature, which were found to lead to degradation and early burnout of the lasers. However, advances in epitaxial techniques, in the III–V arsenide and phosphide systems, together with improvements in the device fabrication techniques, led to the development of low threshold current density quantum well (QW) heterostructure lasers, which are now contributing to the revolution in optical network technology.

The history of semiconductor light emitters at the short visible wavelengths stands in sharp contrast.

Introduction

Today's communicatios industries are facing tremendous challenges and new opportunities brought about by deregulation, competition, and emerging technologies. The opportunities in broadband subscriber access networks have been stimulating large scale business efforts and technology evolution. Network operators and services providers must continuously upgrade their “embedded infrastructure” in order to protect their current revenue while searching for new markets. Depending on the economic situation and projected service/revenue potentials, different network operators and service providers may chose different network upgrade paths and business strategies utilizing different technologies.

With the cost being the primary consideration, the embedded metallic last-mile drop may exist longer than people would hope in wired networks. This therefore makes it necessary to embrace RF transmission technology in those access networks. Even in all-fiber networks (e.g., PON: passive optical networks), the capability of broadcasting multichannel TV signals is critical, therefore making it desirable to carry certain types of RF signals. All these also become very attractive thanks to the innovations in the wireless industry that continually improve the bits/Hz and bits/dollar ratios of RF technologies.

In this chapter, we discuss technology and system issues associated with RF links in the access networks. Instead of discussing all the physical details with complicated equations, we try to provide an intuitive view of those issues, their impacts on system performance, and on-going technology innovations that continually drive architecture evolution and revolution. Section 2.2 describes the different access networks and the roles of RF subcarrier links in those networks.

Introduction

Review of oscillators

Oscillators are devices that convert energy from a continuous source to a periodically varying signal. They represent the physical realization of a fundamental basis of all physics, the harmonic oscillator, and they are perhaps the most widely used devices in modern day society. Today a variety of mechanical (such as the pendulum), electromagnetic (such as LC and cavity based), and atomic (such as maser and laser) oscillators provide a diverse range in the approximation to the realization of the ideal harmonic oscillator. The degree of spectral purity and stability of the output signal of the oscillator is the measure of the accuracy of this approximation, and is fundamentally dependent on the energy storage ability of the oscillator, determined by the resistive loss (generally frequency dependent) of the various elements in the oscillator.

An important type of oscillator widely used today is the electronic oscillator. The first such oscillator was invented by L. De Forest in 1912, shortly after the development of the vacuum tube. In this triode based device known as the “van der Pol oscillator” the flux of electrons emitted by the cathode and flowing to the anode is modulated by the potential on the intervening grid. This potential is derived from the feedback of the current in the anode circuit containing an energy storage element (i.e., the frequency selecting LC filter) to the grid, as shown in Fig. 9.1a.

Introduction

The optical detector plays an important role in an analog fiber link as its performance determines the baseline characteristics of the link. In the past, the development of the detector has evolved around on-off operation, high speed and high responsivity aspects due to its extensive uses in digital fiber links. However, for analog applications, as discussed in Chapter 1, the detector has to incorporate additional designs to ensure high power and high linear dynamic range operation, both of which are essential for meeting the high link gain and low noise figure requirements.

The organization of this chapter goes as follows. In the remainder of this section an overall view of optical detection is presented including key terminology and figures of merit. Section 8.2 compares the baseline properties of different photodiodes. Section 8.3 briefly discusses the noise sources for photodiodes, with special attention to shot noise and noise attributed to laser relative intensity noise. These noise sources are dependent on the optical power. Section 8.4 highlights the research status of photodiode nonlinearity. Section 8.5 presents some recent advances in photodiodes related to analog fiber link applications.

Definitions

Semiconductor materials are commonly employed for optical detectors mainly because the resultant detectors are usually small, low noise, operate with a low voltage, and can be easily assembled with the rest of the receiver circuits. Photons interact with electrons in semiconductors and are absorbed in processes whereby electrons make a transition from a lower to a higher energy state.

Introduction

Microwave links serve important communication, signal processing and radar functions in many commerical and military applications. However, the attenuation of microwave RF signals in cables and waveguides increases rapidly as the frequency of the signal increases, and it is specially high in the millimeter wave range. Optical fibers offer the potential for avoiding these limitations for the transmission of RF signals.

Photonic microwave links employ optical carriers that are intensity modulated by the microwave signals and transmitted or distributed to optical receivers via optical fibers. Since the optical loss for fibers is very low, the distance for photonic transmission and distribution of microwaves can be very long. When the modulation of an optical carrier is detected at a receiver, the RF signal is regenerated. Figure 1.1 illustrates the basic components of a simple photonic microwave link.

Since the objective of a photonic microwave link is to reproduce the RF signal at the receiver, the link can convey a wide variety of signal formats. In some applications the RF signal is an unmodulated carrier – as for example in the distribution of local oscillator signals in a radar or communication system. In other applications the RF signal consists of a carrier modulated with an analog or digital signal.

Introduction

Performance of multiple quantum well (MQW) electroabsorption (EA) modulators can best be evaluated in terms of a basic RF photonic link. A basic RF photonic link consists of a laser transmitter with its optical intensity modulated by an RF signal, an optical fiber transmission line and a receiver. In the externally modulated link shown in Fig. 6.1, the RF source supplies the signal to the EA modulator. The CW laser radiation is coupled directly or through a pigtailed fiber to the modulator input. The fiber transmission line couples the modulated output to the receiver. The receiver detects the optical radiation and converts the intensity modulation back into RF power. Because of the low transmission loss of fibers, a major advantage of an RF photonic link is its low RF transmission loss, especially for long distances and at high RF frequencies. Many RF channels at widely different frequencies can also share the same optical carrier. More sophisticated links may employ optical or electronic amplification, distribute the modulated optical carrier in a fiber network, down or up convert the RF frequencies using opto-electronic components. However, the fundamental effect of a component such as a modulator can most clearly be understood through the basic link.

In Fig. 6.1, the optical carrier is obtained from CW laser and modulated by an external modulator, based on semiconductors or ferroelectrics such as LiNbO3 or polymers.

Introduction

The laser is one of the most important elements in fiber optic links since it generates the coherent optical wave that carries the signal. Typical laser wavelengths are 1.3 μm and 1.55 μm corresponding to the dispersion and absorption minimum, respectively, of silica fibers. The laser frequency is about 200 THz and the RF (10 kHz-300 MHz) or microwave (300 MHz-300 GHz) signal can be modulated onto the laser beam either directly or externally. This chapter focuses on direct modulation. It is simpler to implement than external modulation but the usable bandwidth is limited to a few GHz. Applications of direct analog laser modulation include cable TV, base station links for mobile communication, and antenna remoting. Laser performance requirements include high slope efficiency to obtain high link gain, low laser noise to keep the link noise figure low, and low distortion to achieve a large spurious free dynamic range (SFDR).

Section 3.2 outlines basic physical mechanisms of semiconductor lasers. We emphasize the quantum nature of electrons and photons which helps to understand efficiency and noise issues. The slope efficiency is discussed in detail. Section 3.3 presents the laser rate equations which are the common basis for the analysis of analog performance. Numerical solutions to the rate equations allow for an exploration of a wide spectrum of lasing effects. However, analytical formulas based on the small signal approximation are valid in many cases and they are given throughout this chapter.

Introduction

Microwave frequency conversion techniques using analog photonic link technology are reviewed in this chapter. Opto-electronic or photonic radio-frequency (RF) signal mixing refers to converting intensity modulation of an optical carrier at one modulation frequency to intensity modulation or an electrical output at a different frequency. Frequency conversion optical links integrate the functions of electrical frequency mixing, traditionally provided by electronics, together with the transport of the RF carrier by the optical link. Photonic RF signal mixing using fiber optic link technology has recently become a topic of interest for reducing front-end hardware complexity of antenna systems and efficiently extending link frequency coverage into the millimeter-wave (MMW, 30–300 GHz) range. As commercial and military systems push to higher operating frequencies, microwave optical transmission and signal conversion techniques offer attractive benefits to designers of RF systems for communications, radar and electronic warfare applications

The frequency converting link diagram displayed in Fig. 10.1 can be used to introduce the concept of photonic link signal mixing as well as to introduce some nomenclature that will be used throughout the chapter. Here, an example antenna remoting configuration is shown that incorporates both optical RF up-conversion for transmit operation and optical RF down-conversion for receive operation. Referring to the transmit-mode up-conversion path of Fig. 10.1, the photonic frequency conversion occurs by multiplying the MMW optical local oscillator (LO) signal at fLO1, with the lower frequency RF input or information bearing signal at fIF1 in the integrated optical modulator (IOM).

Introduction

As the modulation frequency is increased in a traveling wave electro-optic modulator, good performance becomes more and more difficult to realize. There are generally three reasons for this: (1) velocity mismatch, (2) electrode loss, and (3) parasitic inductance and capacitance in the connections to the electrodes. Several schemes have been invented to deal with these limitations, some of which have already been discussed in Chapter 5. In this chapter we introduce a new scheme that offers improvement in all three limiting aspects, but is practical only at very high microwave frequencies. The basic scheme is simple. The transmission line electrodes of the modulator are broken into N sections, and each section is connected to an on-substrate antenna. The array of antennas thus formed is excited by a plane wave incident from the substrate side, as illustrated in Fig. 11.1. The illumination angle is chosen so that the phase velocity from antenna to antenna is the same as the phase velocity of the light in the optical waveguide. This assures velocity matching from segment to segment, even though the phase velocities of RF and optical waves are not matched within a segment. This overcomes limitation (1), so that the modulator may be made as long as desired. While the optical power is now divided by N (at best), the attenuation along each segment is now α L/N compared to α L for a simple modulator of length L with transmission line loss of α nepers/unit length.

Introduction

This chapter will cover the design of modulators for high performance externally modulated analog links. This includes development of a model of the externally modulated analog link so that the connection between modulator design parameters and link performance is clear. The only type of link considered here is one using amplitude modulation of the light and direct detection.

Section 4.2 will go through the basic modulator designs in detail. Section 4.3 will relate the modulator performance to link performance, including an explanation of modulator linearization and an overview of several types of linearized modulator.

The link analysis in Section 4.3 is applicable to any externally modulated link using a modulator that can be characterized by a transfer function that depends on voltage (i.e., the optical transmission depends on voltage). It applies to modulators of any type in any material that meet this criterion. The modulator designs given in Section 4.2, and the linearized modulators described in Section 4.3.2, can be built on a variety of materials.

Lithium niobate (LiNbO3) presently is the material in widest use as a modulator substrate for modulators used in analog links. It is an insulating crystal whose refractive index changes with voltage. It has a high electro-optic coefficient, it is stable at normal electronic operating temperatures, low-loss (0.1 dB/cm) optical waveguides can be made in it, there are manufacturable ways to attach optical fibers with low coupling loss, and its dielectric constant is low enough that high speed modulators can be made without too much trouble.

Introduction

The application of RF photonics to RF antenna systems is a thread of development that runs parallel to the development of high performance analog links and components. This chapter deals with the promise of high performance analog photonic links applied to the field of phased arrays or RF manifolds.

In the early 1980s RF design engineers began to view photonics as a promising system option because of the possibility of modulating RF signals onto an optical carrier and the advantage of fiber optic transmission for low loss over large distances. As a cable replacement, fiber optics offers extremely wide bandwidth with no dispersion, significant weight reduction, loss reduction over a distance greater than 100 m, and immunity from electromagnetic interference or cross coupling. The challenges for achieving cable replacement have been lowering the overall conversion loss, and achieving high spur-free dynamic range, and high dynamic range.

In the late 1980s the focus shifted to applying photonics to antenna systems, as the development of optical networking techniques suggested the possibilities of routing optical modulated carriers to perform different RF functions. In particular the systems designers were interested in the possibility of performing several RF functions with photonics, namely beam steering, null steering, channelization, and local oscillator distribution over an optical manifold. To implement a photonic architecture, we assume the existence of high fidelity links and networks and device schemes for signal gathering, beam control, phase steering and true time delay steering, multichannel remoting, and pre-processing in the optical domain.

Benefits of polymer modulators

Polymer electro-optic modulators offer several important advantages over more mature technologies such as lithium niobate interferometric modulators or semiconductor electroabsorption modulators. Velocity matching between RF and optical waves is much simpler because there is a close match between the dielectric constants at optical and radio frequencies, enabling flat response to higher frequencies. The relatively low dielectric constant enables 50-Ωdrive electrodes to be easily achieved with simultaneous velocity matching. Sensitivity as much as an order of magnitude greater than in lithium niobate may be achievable, primarily through material engineering to achieve very large electro-optic coefficients, but also through the drive field concentration enabled by parallel-plate or microstrip drive electrodes. Greater sensitivity can result in greater RF gain and smaller noise figure of an RF link incorporating an external modulator. While the electro-optic effect employed is resonant, the wavelength of operation is far from resonance, so that there is little wavelength sensitivity or thermal sensitivity, in contrast to electroabsorption modulators which are operated close to resonance. The refractive index of polymer materials is nearly matched to that of glass optical fibers, enabling small Fresnel losses at interfaces. The material cost of polymer modulators may be low because they can be made of thin films of high-value polymer material on top of inexpensive substrates. Because polymer waveguide layers can be fabricated on a variety of substrates, including flexible substrates, by spin-coating, integration of polymer waveguides with other components is possible, and conformable devices could be made.

Introduction

Traveling wave electro-optic devices were first proposed in 1963 and demonstrated, in LiNbO3 waveguide modulators in the late seventies. The basic requirements for broadband operation were known at the outset: velocity matching between the microwave and optical signals, low electrical loss in the microwave guide, and impedance matching between the microwave waveguide and external electrical connectors. However, these conditions were not immediately realized because the dielectric constants and refractive index of LiNbO3 made velocity mismatch inevitable, within the constraints of existing electrode geometries. Eventually, accurate modeling of the multilayer structure using numerical methods, combined with processing improvements, led to devices which were nearly velocity matched. Electrode thickness increased as the structures evolved, which naturally decreased the microwave loss, further extending device bandwidth.

Impedance matching has been considered a lower priority than velocity matching and low losses and was not achieved in a velocity matched device until etched ridge LiNbO3 devices were introduced. These devices concurrently satisfy the conditions for broadband operation, and operation beyond 100 GHz has recently been attained.

The development of broadband, low Vπ modulators is a final consideration. While the drive voltage can be decreased by lengthening the electrodes, this also increases electrode loss, resulting in a trade-off between drive voltage and bandwidth. Clearly, requirements for a particular application must be considered in modulator design.

This chapter covers the impressive progress that has been made in broadband LiNbO3 modulators over several decades, and addresses current challenges which face researchers.

RF technology is at the heart of our information and electronic technology. Traditionally, RF signals are transmitted and distributed electronically, via electrical cables and waveguides. Optical fiber systems have now replaced electrical systems in telecommunications. In telecommunication, RF signals are digitalized, the on/off digitally modulated optical carriers are then transmitted and distributed via optical fibers. However, RF signals often need to be transmitted, distributed and processed, directly, without going through the digital encoding process. RF photonic technology provides such an alternative. It will transmit and distribute RF signals (including microwave and millimeter wave signals) at low cost, over long distance and at low attenuation.

RF photonic links contain, typically, optical carriers modulated, in an analog manner, by RF subcarriers. After transmission and distribution, these modulated optical carriers are detected and demodulated at a receiver in order to recover the RF signals. The transmission characteristics of RF photonic links must compete directly with traditional electrical transmission and distribution systems. Therefore the performance of an RF photonic transmission or distribution system should be evaluated in terms of its efficiency, dynamic range and its signal-to-noise ratio.

RF photonic links are attractive in three types of applications. (1) In commercial communication applications, hybrid fiber coax (HFC) systems, including both the broadcast and switched networks, provide the low cost network for distribution of RF signals to and from users. RF photonic technology has already replaced cables in commercial applications such as CATV.