Semiconductors: Atomic-level Behavior

Type of physical science: Condensed matter physics

Field of study: Solids

Semiconductors are materials whose electrical conductivity lies between that of metals and insulators. They are the basis for contemporary electronic devices such as diodes, transistors, and integrated circuits.

Overview

Semiconductors are materials whose electrical conductivity lies between that of conductors and insulators. Common semiconductors include elements such as silicon and germanium and compounds such as zinc sulfide and gallium arsenide. Because of their moderate ability to conduct electricity--about 10^-5 times that of metals--semiconductors have found many uses in solid-state devices such as transistors, diodes, and integrated circuits.

Most semiconductors are produced from crystalline materials. To exhibit the proper electrical properties, it is necessary to produce these semiconductors as single crystals, not polycrystalline materials. The production of very pure and large single crystals requires precise manufacturing conditions. There are two types of semiconductors: intrinsic and extrinsic. The electrical properties of both types, and the distinction between conductors, insulators, and semiconductors, can be explained by considering the band theory of solids.

Isolated, noninteracting atoms consist of nuclei surrounded by electrons. In accordance with modern atomic theory, the electrons can be found only at discrete, distinct energy levels.

Each atom has its own particular set of energy levels, and all like atoms have identical energy levels. The electrons fill these levels starting from the lowest energy level and proceeding upward in energy, until all electrons have been placed into one of the available levels. The electrons which are found in the highest level are called the valence electrons.

When large numbers of atoms join together to form a solid, there is a certain amount of interaction between the individual atoms. This interaction causes the discrete energy levels found in isolated atoms to broaden into bands. With semiconductors, the band corresponding to the highest energy level, which contains electrons (the valence band) is relatively close to the lowest energy level, which is empty (the conduction band). The energy gap between the two bands is sufficient to prevent the valence electrons from moving freely into the conduction band, as happens with conductors, but not so great as to eliminate totally valence electrons from the conduction band, as happens with insulators. Only those electrons located in the conduction band are capable of conducting electricity.

Pure semiconductors, which conduct electricity solely because of the size of the energy gap between the valence band and the conduction band, are called intrinsic semiconductors. The conduction of intrinsic semiconductors is low and strongly temperature-dependent. The intrinsic semiconductors of major commercial importance--silicon and germanium--are both found in group IVA of the periodic table. Gray tin, which is also in group IVA, is a semiconductor as well, but its transformation into white tin at near-ambient temperatures prevents its commercial application. Other elements, such as boron from group IIIA and tellurium from group VIA, are intrinsic semiconductors. Extrinsic semiconductors are formed from intrinsic semiconductors when the inherent energy gap is modified by adding very small amounts (typically 1 part per million) of impurity atoms within the crystal. This process of adding impurities is called "doping." The conductivity of extrinsic semiconductors can be varied widely by varying the type of impurity and its concentration.

In a semiconductor such as silicon, each silicon atom has four available electrons, which can form chemical bonds with adjacent atoms. In crystalline silicon, each silicon atom forms bonds with four adjacent silicon atoms. If one of the silicon atoms is replaced by an impurity atom with a different bonding capability, this simple view of four identical bonds has to be modified.

Assume that one of the silicon atoms is replaced by an atom such as arsenic. Arsenic can form bonds with five adjacent atoms; however, only four neighboring silicon atoms are available for bonding. The fifth bonding electron from the arsenic would be very weakly bound and essentially free to move about the crystal. Considering the band theory, the excess bonding electrons (one of which would be available from each impurity atom) would be at an energy state immediately below the conduction band. As a result, it would be easy to promote these electrons into the conduction band. Extrinsic semiconductors with impurities capable of providing an excess, nonbonding electron are called "n-type" semiconductors. The nonbonding electrons move throughout the crystal transferring the negative electrical charge. Other common n-type dopants in addition to arsenic are antimony, bismuth, and phosphorus.

Assume that one of the silicon atoms is replaced by an atom such as boron. Since boron can bond with a maximum of three other atoms, only three of the neighboring silicon atoms could bond with the boron impurity. The fourth neighboring silicon would be able to form only an incomplete bond. One incomplete bond such as this is associated with each impurity atom. In terms of the band theory, this electron vacancy (or hole) would generate an energy level immediately above the valence band. This empty energy band can serve as a stepping stone for electrons, aiding their promotion to the conduction band. Impurities such as boron, which lack an adequate number of bonding electrons, generate "p-type" semiconductors. The vacancy, which has an effective positive charge, can itself move throughout the crystal, transferring positive electrical charge. Indium, gallium, thallium, and aluminum are other p-type dopants.

Semiconductors can be doped in three ways: First, the semiconductor can be alloyed, or mixed, directly with the impurity material prior to crystallization. Second, the impurity can be diffused into the semiconductor by putting the solid semiconductor in contact with a gaseous impurity. Finally, the semiconductor surfaces can be bombarded with a high-energy beam of impurity atoms, leading to ion implantation. The amount of impurity added need not be large.

For example, only 0.01 percent boron is necessary to increase the conductivity of silicon 1 million times. The nature of the doping--whether the material is an n-type or a p-type--can be determined by a classic experiment known as the Hall effect. A magnetic field applied at right angles will deflect the charge in a direction determined by whether the current is transported by electrons or holes, causing a voltage buildup across the semiconductor.

Not all semiconductors are based on elemental crystals. Semiconductors can also be composed of compounds from elements near group IVA. Examples include boron phosphide, gallium phosphide, gallium arsenide, zinc sulfide, zinc selenide, and cadmium sulfide. These compounds average electronically to the group IVA structure, in that they are compounds formed from IIIA and VA elements or from IIA and VIA elements. Both of these classes of compounds are intrinsic semiconductors which can be doped to form extrinsic semiconductors.

Though most semiconductors are found in the crystalline state, advances have produced so-called amorphous semiconductors. Amorphous, or noncrystalline, semiconductors have one major advantage commercially: They are less expensive to manufacture. They can be produced from materials similar to the crystalline semiconductors, such as gallium arsenide, silicon, or a silicon/hydrogen alloy. Semiconductive behavior is also exhibited by some organic solids. These semiconductors are quite different structurally from the conventional inorganic semiconductors.

No generally accepted theory has been developed concerning their conduction mechanism.

Applications

The major application for semiconductors is in solid-state electronics. Solid-state devices are all based on semiconductors. The simplest of these is the diode or rectifier, which permits electrical current to flow in only one direction. A diode consists of a single "p-n junction," where a p-type semiconductor is placed in intimate contact with an n-type semiconductor. The boundary region between the two is called the junction.

When voltage is applied across a diode so that the negative electrode is attached to the p-type material and the positive electrode is attached to the n-type material, no electricity flows.

The holes are attracted to the negative electrode and the electrons to the positive electrode, but no current travels through the junction. This condition is called "reverse biased." On the other hand, if the electrodes are reversed, a net current passes through the junction. The positive electrode would repel the holes toward the center and attract the excess electrons from the n-type layer across the junction. The negative electrode from the other side would repel the electrons toward the center but attract the holes from the p-type layer. The holes and electrons would meet at the junction, where they would recombine continuously and lead to a flow of both positive and negative charge carriers in the overall circuit. This condition is referred to as "forward biased."

If an alternating current is applied to a diode, the current is rectified, or converted into a direct current. The current will flow through the diode only in one direction, corresponding to the forward bias, the direction which is determined by the orientation of the diode.

Adding a third semiconductor layer forms a triode. The most common triode device is the transistor. A transistor consists of two junctions, either p-n-p or its complementary form, n-p-n. A transistor is formed by the junction of two p-n diodes. Assume that the transistor is of p-n-p structure. The n-type material is a very narrow region, with relatively low impurity content, sandwiched between two p-type regions of higher impurity content. Each region has a terminal, so that it can be connected with the external circuit. These three regions are termed, respectively, the emitter, base, and collector. The first junction, between the emitter and base, is forward biased; the second junction, between the base and collector, is reverse biased. When positive current is applied to the (p-type) emitter, the holes within the emitter move freely toward the base. When they reach the base, rather than immediately recombining with the electrons, many of the charge carriers overshoot and pass through into the collector, where they can again move freely. The extent of overshoot through the base is determined by the voltage to the base. Since small changes in voltage to the base can alter the collector current significantly, the transistor acts as an amplifier. A comparable situation exists for n-p-n structures, with electrons rather than holes being the overall charge carrier.

In addition to ordinary diodes and transistors, many other semiconductor devices have been developed. For example, the light emitting diode (LED) is commonly used in calculators, digital display panels, and watches. In an LED, the energy released by the recombination of an electron and hole at a p-n junction occurs as a photon of light. Thyristors, which are four-layer devices (p-n-p-n), can serve as switches and are used in control devices (motor, temperature, light, pressure, and level control), in timing circuits (oven timers, oscilloscopes, and industrial process control), and in many similar applications. Variable resistors (varistors) are based on the nonlinear resistance obtained when a suitable contact is made between a semiconductor and a metal. Thermally dependent resistors (thermistors) utilize the large temperature coefficient of resistivity of semiconductors. Common applications have been found in meteorology, medical thermometry, and microwave and high-frequency power meters. Photoconductive devices are based on the liberation of electrons from semiconductors by the absorption of light energy and the corresponding change in resistance; photovoltaic devices operate as a result of the voltage generated when p-n junctions are illuminated.

Increasingly, semiconductor devices are fabricated in arrays on tiny chips of substrate material to produce integrated circuits. An integrated circuit combines a number of interconnected circuit elements, such as diodes, resistors, capacitors, and transistors. There are three basic types of integrated circuits: Monolithic integrated circuits are formed of components that are prepared within and upon a semiconductor substrate. At least one of the components must be within the substrate itself. Monolithic integrated circuits are produced by the successive depositing of layers of semiconductor materials and the subsequent etching of patterns that define current paths. In this way, the function of many electronic components is replicated on a single three-dimensional microscopic square of semiconductor material.

Multichip integrated circuits are formed by the assembly of two or more tiny pieces, or chips, of semiconductor material on a substrate. Each of these chips may contain a single component, or they themselves can be monolithic integrated circuits. The substrate serves to separate and interconnect chips, but does not actually contain any circuit components.

In film integrated circuits, the various components are produced from films deposited on an insulating substrate such as glass or ceramic. Film integrated circuits can be either thin film or thick film. Thin-film technology utilizes films of a wide variety of substances, including nickel-chromium alloys, tantalum, tantalum nitride, metal silicides, and mixtures of metals and insulators. The components produced can be very tiny, typically less than 10 microns in width.

Thick-film circuits utilize special inks printed in patterns through masks, as in silk-screen printing. When the inks are fired at high temperature, they form electronic components, conductors, and insulating layers several microns thick and at least 100 microns wide. The film integrated circuits are restricted to passive components such as arrays of resistors and capacitors.

The inability to produce a generally useful active component such as a transistor has limited their commercial use.

Often the various types of integrated circuits are used in conjunction with one another; for example, several monolithic integrated circuits may be combined with thin-film components on the same substrate. Though integrated circuits have many functions, the majority are used in digital logic for data processing, such as in computers and calculators.

Context

The properties of semiconductors, particularly their photoconduction and strongly temperature-dependent conductivity, have been recognized since the mid-1800's. The first practical application, however, was probably in early radios, which utilized a crystal rectifier--usually a lead sulfide crystal--as the receiver. Semiconductor rectifiers were introduced commercially in the 1930's, many years before the p-n junction mechanism was understood.

These early diodes were made from copper oxide, and employed a copper/cuprous oxide junction. This composition has been seen only rarely since the later introduction of the less-expensive selenium rectifier with a selenium/tin-cadmium alloy junction.

World War II highlighted the requirement for miniature electronic systems. During this period, the first integrated circuit concepts were developed. Arrays of interconnected resistors were fabricated by modifying screen printing techniques using resistive inks, silver paste, and ceramic substrates. During the late 1940's, significant advances were made regarding other techniques for producing integrated circuits. Of particular future importance was the development of etched circuits to produce various conduction pathways.

The next major advance in semiconductor electronics was the transistor. The transistor was invented by three Bell Laboratory scientists, John Bardeen, Walter H. Brattain, and William Shockley, in 1948. The device was initially named for the phrase "transfer-resistor" because of its behavior in circuits. It was the first practical solid-state device capable of signal amplification.

Early transistors, rather than employing p-n junctions, used two closely spaced fine wire springs, in contact with a semiconductor surface, as the rectifying junctions. Because of this structure, they were sometimes known as point-contact devices. Shortly after their introduction, transistors began replacing vacuum tubes as the basic electronic component in many different types of circuits.

By the late 1950's, all the basics for integrated circuits had been developed. The first monolithic integrated circuit was introduced in 1958. It consisted of a transistor, two resistors, and a resistor-capacitor network. By 1959, thin-film conductors, over an oxide insulator, had been developed to serve as an interconnection structure. At about the same time, vacuum evaporation techniques (previously used to coat lenses and mirrors) were extended for the deposition of semiconductor thin films. Since that time, semiconductor electronics have proliferated. By the 1980's, vacuum tubes had essentially vanished, and semiconductor devices became an integral part of modern life-style and technology.

In general, solid-state devices have many advantages over conventional electronic components and/or systems: They are reliable, rugged, and compact; they have low power dissipation and thus operate at relatively low temperatures and conserve energy; they are versatile, because their applications cover the entire electromagnetic spectrum, from ultraviolet frequencies to direct current, and they can be used at powers ranging from very low to extremely high levels; they are sensitive, and thus can operate from low-level signal sources; and they are economical to purchase and operate.

Principal terms

BAND THEORY OF SOLIDS: the atomic-level explanation, from a physical perspective, of the conductivity exhibited by semiconductors

CONDUCTION ELECTRON: an electron that has been promoted from the valence band in a solid to a higher energy state, which frees it to migrate through a solid, transferring electrical current

DOPING: the intentional addition of small amounts (typically 1 part per million) of impurity to an intrinsic semiconductor to enhance and control its conductivity

EXTRINSIC SEMICONDUCTOR: a semiconductor that is doped with small amounts of impurity atoms to enhance and control its conductivity

INTRINSIC SEMICONDUCTOR: a semiconductor that conducts electricity as a result of the energy gap between the valence band and the conduction band

N-TYPE SEMICONDUCTOR: an extrinsic semiconductor that is doped with impurity atoms with more valence electrons than the matrix material, causing it to conduct electrical charge by the migration of loosely bound electrons

P-N JUNCTION: the boundary region created when a p-type semiconductor is placed into intimate contact with an n-type semiconductor

P-TYPE SEMICONDUCTOR: an extrinsic semiconductor that is doped with impurity atoms with fewer valence electrons than the matrix material, causing it to conduct electrical charge by the migration of holes

VALENCE ELECTRON: an electron that is found in the outermost electronic energy level in an atom; as a result, such electrons fill the valence band of solids

Bibliography

Brodsky, Marc H. "Progress in Gallium Arsenide Semiconductors." SCIENTIFIC AMERICAN 262 (February, 1990): 68-75. Well-written description of one particular type of semiconductor, gallium arsenide. Gallium arsenide has been considered the "technology of the future." The speed at which electrons move through the material renders it suitable for applications in computing, television reception, and fiber-optic networks. Specifically describes gallium arsenide's properties, manufacturing technology, and applications.

Grove, A. S. PHYSICS AND TECHNOLOGY OF SEMICONDUCTOR DEVICES. New York: John Wiley & Sons, 1967. This book reviews the principles of solid-state technology, including vapor-phase growth and diffusion and then discusses basic semiconductor physics and the properties of p-n junctions. Focuses on some of the more sophisticated devices such as field-effect transistors, with an emphasis on surface-controlled devices. Recommended for the more scientifically trained reader.

Hamakawa, Yoshihiro. "Photovoltaic Power." SCIENTIFIC AMERICAN 256 (April, 1987): 86-92. Semiconductors of silicon, copper indium diselenide, cadmium telluride, and gallium arsenide can be used to convert solar energy into electricity. Focuses on both the theoretical aspects of their operation as well as many of the practical considerations, including cost factors. For the general reader.

Kane, Philip F., and Graydon B. Larrabee. CHARACTERIZATION OF SEMICONDUCTOR MATERIALS. New York: McGraw-Hill, 1970. Discusses the properties of semiconductor materials, including the history and early development, principles of functioning (bands, intrinsic and extrinsic conduction, p-n junctions), properties of specific semiconductors (silicon, germanium, gallium, indium, and arsenic), growth of single crystals, analysis for imperfections (both chemical and physical), and characterization of surfaces and films. In contrast to most books in the field, the material is presented at a level comprehensible to the general reader.

Katon, J. E., ed. ORGANIC SEMICONDUCTING POLYMERS. New York: Marcel Dekker, 1968. A series of monographs focusing on the potential development of organic macromolecular semiconductors. After a general review of semiconductors, the monographs concentrate on theoretical and experimental aspects of their general electrical behavior, metal-containing polymers, and biological polymers.

Meindl, James D. "Chips for Advanced Computing." SCIENTIFIC AMERICAN 257 (October, 1987): 78-88. Many of the advances in semiconductors are now related to the miniaturization of increasingly complex integrated circuits for handling information. Considers the theoretical and practical aspects of gigascale integration, including its implications for the future.

Sze, S. M. SEMICONDUCTOR DEVICES, PHYSICS, AND TECHNOLOGY. New York: John Wiley & Sons, 1985. Thorough, extensive look at the structure and properties of semiconductors and semiconductor devices. Includes discussions of more advanced concepts such as tunnel diodes, junction transistors, thin-film devices, and optoelectronic devices. The mathematics level is fairly complex, but the book starts with the basics, such as crystal structure, and then develops the topics. Many diagrams, graphs, and figures add to the explanations.

Tauc, J., ed. AMORPHOUS AND LIQUID SEMICONDUCTORS. New York: Plenum Press, 1974. Treats amorphous and liquid semiconductors, which are much more difficult to discuss theoretically than crystalline semiconductors. A fundamental discussion of the nature and structure of amorphous materials is followed by consideration of their physical and electronic structure and their electronic and optical properties. Also discusses the structure and electronic properties of liquid semiconductors. Mathematically sophisticated.

Charges and Currents

Electrical Properties of Solids