Silicon

Carbon's heavier cousin

Silicon sits directly below carbon in group 14, with the same four outer electrons, [Ne] 3s² 3p², and the same habit of forming four bonds pointed at the corners of a tetrahedron. Pure silicon crystallises in the structure of diamond, but it is grey, brittle and metallic in lustre — a metalloid, neither metal nor nonmetal. At room temperature it conducts electricity poorly; warmed, lit or doped with traces of other elements, it conducts better, because its band gap of 1.12 eV is narrow enough for electrons to cross.

The resemblance to carbon goes only so far. Silicon–silicon bonds are weak and silicon–oxygen bonds very strong, about 450 kJ per mole, so silicon does not build long chains of itself as carbon does; it builds frameworks with oxygen instead. Carbon dioxide is a gas; silicon dioxide is quartz. That is why life is made of carbon and rocks of silicon. Silicon is one of the few substances that, like water, expand as they freeze, which makes it awkward to cast.

The stuff of rock

Silicon is 28 % of the crust, second only to oxygen, and the two together are three quarters of it. Nearly all of it is in silicates: SiO₄ tetrahedra joined at their corners into chains (the pyroxenes), sheets (mica, clay, talc) or three-dimensional frameworks (feldspar, quartz). Sand, granite, basalt, slate and clay are silicon compounds; so are glass, brick, cement and porcelain, the oldest materials people have made. Silica itself, SiO₂, is quartz, flint, opal, agate and amethyst, and the shells of diatoms, the spicules of sponges and the gritty edges of grasses are silica laid down by living things.

The element is freed by heating quartz with coke in an electric arc furnace at about 2000 °C. Most of the eight million tonnes or so made each year go into aluminium alloys and steel, or become silicones, the flexible polymers of seals, implants and bakeware. A small fraction is purified to better than 99.9999999 % — converted to trichlorosilane, distilled and decomposed — then pulled from the melt as a single crystal by the Czochralski method: an ingot the size of a log, sliced into wafers.

The age of silicon

The first transistor, in 1947, was germanium; the first silicon transistor followed in 1954, and in 1958 and 1959 Jack Kilby and Robert Noyce built the first integrated circuits. Silicon won because its oxide is an excellent insulator that can be grown on its surface at will, so that billions of transistors can now be patterned onto one wafer. The valley south of San Francisco where this was done was named for it in 1971. The same crystal, cut thin and doped, turns sunlight into current: about 95 % of the world's solar panels are silicon.

Silicon has no known role in animal biochemistry, but plants take it up as silicic acid, and the diatoms, which carry out about a fifth of the photosynthesis on Earth, cannot live without it.

Discovery and name

Silica was an earth to the chemists of the eighteenth century. Lavoisier suspected in 1789 that it was the oxide of an unknown element, and Humphry Davy, failing to reduce it in 1808, named that element silicium, from the Latin silex, flint; in 1817 Thomas Thomson proposed silicon, to match boron and carbon rather than the metals. Joseph Gay-Lussac and Louis Thénard had obtained an impure brown powder in 1811 by heating silicon tetrafluoride with potassium, but did not recognise it. Jöns Jacob Berzelius did, in 1824: he reduced potassium fluorosilicate with potassium, purified the product and described the element. English kept Thomson's silicon; most of Europe kept Berzelius's silicium.