The Semiconductor Revolution
No material has transformed modern civilization more profoundly than silicon. Every smartphone, laptop, and server in the world depends on silicon's peculiar electronic properties — properties that arise from quantum mechanics and can be precisely tailored through chemistry.
A semiconductor is a material with electrical conductivity between that of a metal (~10⁷ S/m) and an insulator (~10⁻¹⁵ S/m). Silicon's conductivity (~10⁻³ S/m in pure form) is unremarkable. What makes semiconductors extraordinary is that their conductivity can be tuned over many orders of magnitude by adding tiny concentrations of impurities — a process called doping.
Band Theory of Solids
To understand semiconductors, we need band theory — a quantum mechanical description of how electrons behave in solids.
In an isolated atom, electrons occupy discrete energy levels. When billions of atoms form a solid, these levels broaden into continuous energy bands. Two bands are critical:
- Valence band: The highest energy band fully occupied by electrons at absolute zero
- Conduction band: The next higher band, empty at absolute zero
- Band gap (Eₘ): The forbidden energy range between valence and conduction bands
The band gap determines a material's electrical character: - Metals: No band gap (or overlapping bands); electrons flow freely → conductor - Semiconductors: Small band gap (0.1–3 eV); thermal energy can excite electrons across the gap - Insulators: Large band gap (>5 eV); electrons cannot jump the gap under normal conditions
Silicon has a band gap of 1.12 eV at room temperature. Diamond (carbon, also group 14) has a band gap of 5.5 eV — an insulator. Germanium (also group 14) has 0.67 eV — a narrower-gap semiconductor.
Silicon's Crystal Structure
Silicon adopts the diamond cubic structure: each silicon atom is covalently bonded to four neighbors in a tetrahedral arrangement, with each neighbor also bonded to four others. This creates an interlocking 3D network of Si–Si single bonds.
Pure silicon is an intrinsic semiconductor. At absolute zero it is a perfect insulator. At room temperature, thermal energy (~0.026 eV, or kT) excites a tiny fraction of electrons from the valence band into the conduction band, leaving behind positively charged holes in the valence band. Both electrons and holes carry current.
The intrinsic carrier concentration in silicon at 25°C is approximately 1.5 × 10¹⁰ cm⁻³ — tiny compared to metals (~10²² carriers/cm³).
Doping: Engineering Conductivity
Doping means deliberately introducing impurity atoms into the silicon lattice to dramatically increase carrier concentration.
n-Type Doping
Phosphorus (P) and arsenic (As) are group 15 elements — they have five valence electrons, one more than silicon's four. When a phosphorus atom substitutes for a silicon atom, it bonds to four silicon neighbors and has one "extra" electron loosely bound. This electron requires only ~0.045 eV to ionize into the conduction band — much less than silicon's 1.12 eV band gap.
At room temperature, essentially all dopant electrons are ionized. With a phosphorus doping level of 10¹⁶ cm⁻³, the electron concentration jumps from 10¹⁰ (intrinsic) to 10¹⁶ cm⁻³ — a millionfold increase. Silicon doped with group 15 elements is called n-type (n for negative majority carriers).
p-Type Doping
Boron (B) is a group 13 element with three valence electrons. A boron atom substituting for silicon is one electron short of completing four covalent bonds. It readily accepts an electron from a neighboring Si–Si bond, creating a hole that moves through the lattice as positive charge. Boron doping creates p-type silicon.
The p–n Junction
When p-type and n-type silicon are joined, electrons from the n-side diffuse across the junction and fill holes on the p-side, creating a depletion region depleted of mobile carriers. This establishes a built-in electric field that opposes further diffusion — equilibrium.
Applying a forward bias (positive voltage on p-side) narrows the depletion region and allows current to flow easily. Reverse bias widens it and blocks current. This rectifying behavior is the diode — the simplest semiconductor device.
The bipolar junction transistor (BJT) and MOSFET (metal-oxide-semiconductor field-effect transistor) build on this principle to amplify signals and switch states. A modern processor contains tens of billions of MOSFETs on a chip smaller than a fingernail.
Moore's Law and the Limits of Silicon
Moore's Law — the observation that transistor count doubles roughly every two years — has driven semiconductor progress for six decades. Current leading-edge chips (e.g., TSMC 3nm process node) pack ~100 million transistors per mm², with gate lengths around 3–5 nm (roughly 10–15 silicon atoms wide).
At these scales, quantum mechanical effects like electron tunneling become significant, and heat dissipation becomes critical. Silicon's physical limits are being approached, driving research into alternative materials.
Beyond Silicon
Compound Semiconductors
Gallium arsenide (GaAs) has higher electron mobility than silicon and a direct band gap, making it efficient for light emission. It dominates in laser diodes, solar cells for satellites, and radio-frequency transistors in cellular phones.
Indium phosphide (InP) is used in ultrafast electronics and fiber-optic communications.
Gallium nitride (GaN) has a wide band gap (3.4 eV) enabling high-power, high-frequency transistors. GaN transistors are replacing silicon in EV chargers, power supplies, and 5G base stations.
Wide-Bandgap Semiconductors
Silicon carbide (SiC) (band gap: 3.26 eV) can operate at higher voltages, temperatures, and frequencies than silicon. Tesla and other EV manufacturers use SiC MOSFETs in inverters, achieving greater range and efficiency.
Two-Dimensional Materials
Graphene — a single atomic layer of carbon in a hexagonal lattice — has extraordinary electron mobility (~200,000 cm²/V·s, versus ~1,400 for silicon). Molybdenum disulfide (MoS₂) and related 2D materials can be only one atom thick yet still exhibit a band gap, making them candidates for ultimately miniaturized transistors.
Silicon Manufacturing: From Sand to Chip
The journey from raw material to microchip is remarkable:
- Metallurgical-grade silicon: SiO₂ (quartz sand) + carbon → Si + CO₂, giving ~98% pure silicon
- Electronic-grade silicon: Si + HCl → SiHCl₃, then purified by fractional distillation and reduced back to ultrapure Si (99.9999999% — "nine nines")
- Czochralski process: A seed crystal is dipped into molten silicon and slowly pulled upward while rotating, growing a single-crystal boule 300 mm in diameter
- Wafer slicing: The boule is sliced into 300 mm diameter, ~775 μm thick wafers and polished to atomic flatness
- Lithography and doping: Patterns are printed using extreme ultraviolet (EUV) light at 13.5 nm wavelength; dopants are implanted by ion bombardment
The semiconductor industry is arguably the most precise manufacturing operation humanity has ever undertaken.