Semiconductor

• 1. Introduction

  ›

  A semiconductor is a material whose ability to carry electricity sits somewhere between a conductor and an insulator. That middle ground sounds unremarkable. It is the reason microchips, computer processors, and most modern electronics exist at all. The trick is that a semiconductor does not stay in the middle. By adding tiny amounts of impurities to its crystal structure, a process called doping, its conductivity can be tuned. Add roughly one foreign atom for every hundred million, and silicon starts to behave in a new way. How does adding dirt to a pure crystal make it more useful, not less? Why does heating a semiconductor make it conduct better, the opposite of what a metal does? And how did scattered observations from the early 19th century turn into the diode, the transistor, and the integrated circuit?

• 2. The Logic Of Doping

  ›

  One in 10 to the 8th power. That is the order of magnitude of impurity that turns silicon from a poor conductor into a controllable one. The added atoms are either pentavalent, like antimony, phosphorus, or arsenic, or trivalent, like boron, gallium, or indium. The resulting materials are called doped, or extrinsic, semiconductors. Group V elements, with five valence electrons, behave like donors. They create free electrons, producing what is called n-type doping. Group III elements, with three valence electrons, behave like acceptors. They create free holes, producing p-type doping. Silicon itself has four valence electrons bonding each atom to its neighbors. A silicon crystal doped with boron becomes a p-type semiconductor. One doped with phosphorus becomes n-type. The n and p labels mark which carrier is the majority carrier; the opposite carrier, called the minority carrier, exists at much lower concentration due to thermal excitation. The numbers are stark. A 1 cubic centimeter sample of pure germanium at 20 degrees Celsius holds only a handful of free electrons and holes. Adding 0.001 percent arsenic donates an extra 10 to the 17th free electrons in that same volume and raises conductivity by a factor of 10,000. Using a hot-point probe, a technician can quickly tell whether a given sample is p-type or n-type.

• 3. Electrons, Holes, And Junctions

  ›

  Charge carriers are the heart of the story, and they come in three kinds: electrons, ions, and electron holes. A hole is the absence of an electron in the valence band, and it behaves like a positively charged particle with its own positive effective mass. It responds to electric and magnetic fields just as a real positive charge would in a vacuum. Electrons in the conduction band, by contrast, are dilute enough to be treated like a classical ideal gas, flying around freely with a different effective mass, described by simple ideas such as the Drude model and electron mobility. When two differently doped regions meet inside the same crystal, they form a junction, and this is where the useful behavior begins. A homojunction can be built from p-doped and n-doped germanium joined together. Electrons and holes are exchanged across the boundary until equilibrium is reached through recombination, leaving a narrow strip of immobile ions that sets up an electric field across the junction. The p-n junctions between regions are responsible for the useful electronic behavior, and a single device crystal can contain many p-type and n-type regions at once. The behavior of carriers at these junctions is the basis of diodes, transistors, and most modern electronics.

• 4. Light, Heat, And Energy

  ›

  In certain semiconductors, an excited electron relaxes not by giving off heat but by emitting light. Controlling the material's composition and the electrical current allows the properties of that emitted light to be shaped. These materials are used to build light-emitting diodes and fluorescent quantum dots. Heat is the other half of the thermal story. Semiconductors with high thermal conductivity move heat away from circuits, improving thermal management in electric vehicles, high-brightness LEDs, and power modules. Thermal energy can also be converted directly. Semiconductors carry large thermoelectric power factors, which makes them useful in thermoelectric generators, and high thermoelectric figures of merit, which makes them useful in thermoelectric coolers. The same family of materials therefore spans emission, dissipation, and conversion, all governed by how electrons and holes are generated and recombine.

• 5. The Material Palette

  ›

  Group 14 of the periodic table holds the two most commercially important pure semiconducting elements: silicon and germanium. Both carry four valence electrons in their outermost shell, letting them gain or lose electrons equally. Beyond these elements lies a wide palette of compounds. Binary compounds form between groups 13 and 15, such as gallium arsenide, and between groups 12 and 16, groups 14 and 16, and different group-14 elements, as in silicon carbide. Gallium arsenide is the second-most common semiconductor after silicon. It appears in laser diodes, solar cells, microwave-frequency integrated circuits, and more. The list extends to ternary compounds, oxides, alloys, organic semiconductors built from organic compounds, and semiconducting metal-organic frameworks. Not every semiconductor is a tidy crystal. Amorphous and liquid semiconductors exist too, including hydrogenated amorphous silicon and mixtures of arsenic, selenium, and tellurium in varying proportions. These disordered materials lack a rigid crystalline structure but share intermediate conductivity, rapid variation of conductivity with temperature, and occasional negative resistance. They are generally used in thin film structures, which tolerate lower electronic quality and are relatively insensitive to impurities and radiation damage. The word semiconductor even reaches into power cabling, describing plastic XLPE, cross-linked polyethylene, mixed with carbon black in the insulation of medium- to high-voltage cables.

• 6. From Sand To Chip

  ›

  Chemical purity is paramount. At the scale these materials operate, any small imperfection can drastically change how a semiconductor behaves, and crystalline faults like dislocations, twins, and stacking faults are a major cause of defective devices. The larger the crystal, the harder that perfection is to reach. Mass production uses crystal ingots between 100 and 300 millimeters in diameter, grown as cylinders and sliced into wafers. The round shape of a wafer comes from single-crystal ingots usually produced by the Czochralski method. Silicon wafers were first introduced in the 1940s. Turning a wafer into a circuit takes a sequence of steps. Thermal oxidation forms silicon dioxide on the surface, serving as a gate insulator and field oxide. Photomasks and photolithography come next, where ultraviolet light and a photoresist layer trigger a chemical change that writes the circuit pattern. Etching follows. Plasma etching pumps an etch gas into a low-pressure chamber, often chlorofluorocarbon, commonly known as Freon, and a high radio-frequency voltage between cathode and anode creates the plasma. The wafer sits on the cathode and is struck by positively charged ions, leaving silicon etched anisotropically. The final step is diffusion, another name for doping. The wafer enters a 1,100 degree Celsius chamber where impure atoms are injected and diffuse into the silicon, forming the p-n junction. Once the wafer cools to room temperature, the semiconducting wafer is almost prepared.

• 7. The Long Road To The Transistor

  ›

  In 1821, Thomas Johann Seebeck noticed that semiconductors produced a much stronger result in experiments on what became the Seebeck effect. The discoveries piled up across the 19th century. In 1833, Michael Faraday reported that the resistance of silver sulfide falls when heated, the reverse of copper. In 1839, Alexandre Edmond Becquerel observed a voltage between a solid and a liquid electrolyte struck by light, the photovoltaic effect. In 1873, Willoughby Smith found selenium resistors lose resistance under light, and in 1874 Karl Ferdinand Braun observed conduction and rectification in metallic sulfides, an effect Peter Munck af Rosenschold had reported in 1835 only to be ignored. Simon Sze called Braun's work the earliest systematic study of semiconductor devices. Theory lagged behind the experiments. Josef Weiss introduced the term Halbleiter in his Ph.D. thesis in 1910, and Johan Koenigsberger classified solids as metals, insulators, and variable conductors in 1914. By 1931, Alan Herries Wilson had established the band theory of conduction and developed the concept of band gaps. John Bardeen later explained why early predictions matched experiments so poorly: semiconductors are extremely structure sensitive, their properties shifting with tiny amounts of impurity. Devices ran ahead of theory. The 1904 cat's-whisker detector, a primitive semiconductor diode, served early radio receivers, and Jagadish Chandra Bose used a lead-sulfide point-contact rectifier in 1904. In 1906, H.J. Round saw light emission when current passed through silicon carbide crystals, the principle behind the light-emitting diode. The breakthrough came in 1947, when John Bardeen, Walter Houser Brattain, and William Shockley invented the first working transistor, a point-contact device, at Bell Labs. In 1954, physical chemist Morris Tanenbaum fabricated the first silicon junction transistor there. Those early junction transistors were bulky and hard to mass-produce, which is why the integrated circuit of 1958 mattered so much.