Perhaps the most important application of compound semiconductor materials to date is in optoelectronic devices. The types of optoelectronic devices that are discussed here can be classified into two main categories, photonic detectors and emitters. In this section, we discuss photonic semiconductor detectors. In Chapter 13 we will discuss emitters.

Many compound semiconductors, such as GaAs, InP, GaInAs, GaN, ZnS, etc., are direct-bandgap materials. As discussed in Chapter 10, optical absorption and emission processes occur to first order in direct-gap systems. Therefore these materials are extremely useful as both detectors and emitters of electromagnetic radiation. Depending on the bandgap energy, different semiconductors can be used to detect radiation from the far infrared to the ultraviolet portion of the spectrum.

In this chapter, we discuss photonic detectors, concentrating on MIS structures (particularly charge-coupled devices, or CCDs), photoconductors, photodiodes, and avalanche photodiodes (APDs). To begin with, some issues in detection are discussed; later it is shown how each of the above-mentioned device types operate.

Basic issues in Photonic Detection

The fundamental purpose of any photonic detector is to convert an input photonic signal into an electrical signal. The application in which the detector is used greatly affects the performance criterion of the detector. For example, a detector can be used for imaging. In these applications, a good detector is defined as one that provides a high degree of spatial resolution, gray-scale resolution (the ability to distinguish different shades on a totally white to totally black contrast scale), etc.

We next consider the dynamics of a collection or ensemble of particles. In the previous sections, we have discussed the dynamics of one particle or systems that can be reformulated as involving only one particle. In general, the macroscopic world consists of many particles simultaneously interacting with one another. For example, the description of the air molecules in a closed room requires tracking at least 1027 particles. Clearly, to attempt to describe the motions of each individual particle would exhaust the computational power of any conceivable machine, not to mention that of the investigator. Therefore it is necessary to construct a means by which the collective behavior of an entire system can be assessed without relying on the complete description of the underlying histories of each individual particle in the ensemble. A statistical approach, called statistical mechanics, has been invented to treat the behavior of a large collection of particles. The basic principles and applications of equilibrium statistical mechanics are the subject of this chapter. In Chapter 6 we will consider treating systems that depart from equilibrium.

Density of States

A large collection of particles is often referred to as an ensemble. To describe the collective behavior of a large ensemble it is necessary to adopt a statistical approach. The best that we can expect to do is to predict, by using statistics, the average values of the macroscopic observables of interest, particularly the energy, the momentum, etc., and the probability that any given particle has a specific value of one of these observables, for example whether a particle has energy E + dE.

The basics of equilibrium statistical mechanics were presented in Chapter 5. In general, most systems of interest in engineering and science are not in equilibrium but are in some nonequilibrium condition. A familiar example of a nonequilibrium state is that of a metal under the application of an applied electric field. An electrical current flows in the metal, resulting in a net transport of charge from one place to another. Such a state is highly unusual; there are relatively few ways in which the system can be arranged so as to provide a specific current flow. Other examples of systems that are in nonequilibrium are systems with a temperature or particle gradient. In these systems, there is a net transport of particles from one part of the system to another in order to establish equilibrium. Hence, in general, nonequilibrium statistical mechanics is concerned with the description of transport phenomena.

How are the states described above different from equilibrium? To answer this question, let us recall the definition of equilibrium from Chapter 5. In the discussion in Chapter 5, it was argued that the most random configuration of a system corresponds to the equilibrium configuration. If we consider a free isolated electron gas, the most random configuration of that gas would be one in which the momentum of the electrons is totally randomized in direction. When the individual values of the momentum are summed over, the net momentum would then be zero, since on average for every electron with a forward-directed momentum, there exists an electron with a compensating negative momentum equal in value.