Band Gap

Why band gap matters

• Every electronic device. The on/off switching of a MOSFET depends on whether enough thermal carriers can be excited across the gap. A 1.12 eV silicon gap gives intrinsic n_i around 10^10 cm^-3 at 300 K — six orders of magnitude below dopant concentration, which is why doped silicon behaves like a clean n- or p-type material.
• Solar cells. The Shockley-Queisser limit picks the optimal single-junction band gap at 1.34 eV — close to GaAs (1.42 eV, 33% theoretical efficiency) and silicon (1.12 eV, 32% limit). Wider gaps absorb fewer photons; narrower gaps waste excess energy as heat. Multi-junction tandem cells stack different gaps to extract more of the solar spectrum.
• LEDs and lasers. Photon energy E = h*nu = E_g sets emission color. Red AlGaInP at 1.9 eV, green GaP at 2.3 eV, blue GaN at 2.7 eV — Shuji Nakamura's blue GaN LED (1993, Nobel 2014) finally enabled white-light LEDs and the lighting revolution. Indirect-gap silicon cannot emit efficiently at any color.
• Transparent conductors. Indium tin oxide and zinc oxide have E_g around 3.5 eV — too wide to absorb visible light, but heavily n-type doped to be metallic. Used as the transparent electrode in every smartphone screen, solar panel, and OLED.
• Power electronics. Wide-gap SiC and GaN sustain breakdown fields 10x silicon's, giving thinner drift regions, lower switching losses, and devices that operate at 200 C. The transition to SiC/GaN inverters in EVs and data centers is one of the biggest semiconductor industry shifts of the 2020s.
• Photodetectors. Choice of E_g sets the cutoff wavelength: silicon detects up to 1.1 microns (visible + near IR), InGaAs up to 1.7 microns (telecom), HgCdTe down to 0.1 eV gap for thermal imaging at 10 microns. Tunable narrow-gap alloys cover the entire IR.
• Gemstone color. Pure diamond is colorless because its 5.5 eV gap rejects all visible photons. Defects (nitrogen, boron) introduce intermediate states that absorb specific colors — yellow, blue, pink — turning it from spectroscopic curiosity into mineralogy and gemology.

Common misconceptions

• All insulators are wide-gap. Usually yes, but not always. Sodium chloride (8 eV), diamond (5.5 eV), and quartz (9 eV) all qualify, but small-gap semiconductors like germanium (0.66 eV) become effectively insulating at low T even though their gap is small. The boundary between "semiconductor" and "insulator" is operational, not formal.
• Doping changes E_g. No — doping introduces shallow impurity levels inside the gap (donors near the conduction band, acceptors near the valence band), shifting the Fermi level. The intrinsic gap between the bulk valence and conduction bands stays the same. Only at very high doping (more than 10^19 cm^-3) does band-gap renormalization slightly shrink E_g.
• Metals have no bands. Metals have bands too — they just overlap or are partially filled. Sodium has its 3s band half-full; aluminum has overlapping 3s and 3p; copper has a filled 3d band below an open 4s. Saying metals "have no gap" is correct only because the chemical potential sits inside a continuum of states, not in a forbidden region.
• E_g is fixed. Pressure, temperature, and strain all shift E_g. Diamond's gap drops by about 30 meV from 0 K to 300 K; silicon under 10 GPa pressure crosses to a metallic phase. The Varshni equation captures the temperature dependence empirically.
• Direct gap means easy emission. Direct gap is necessary but not sufficient. Non-radiative defects, surface recombination, and Auger processes can dominate even in direct-gap materials if the crystal quality is poor. GaN's revolution required two decades of materials work to reduce dislocation density to where radiative recombination wins.
• Insulators always look transparent. Only those with E_g above 3.1 eV (the violet end of visible at 400 nm) are transparent across the full visible range. Wider gaps like fused silica (9 eV) reach into the UV; narrower gaps like CdS (2.4 eV) absorb violet and look yellow.

Band-gap engineering

Modern device design treats E_g as a tunable parameter, not a fixed material property. Alloying — Si_x Ge_(1-x), Al_x Ga_(1-x) As, In_x Ga_(1-x) N — shifts E_g continuously with composition x. Heterostructures with abrupt junctions create quantum wells, where confinement adds a particle-in-a-box energy on top of the bulk gap. Strain shifts band edges by tens of meV per percent and is exploited in strained-silicon CMOS and InGaAs detectors. Quantum dots tune E_g via dot diameter — CdSe dots span 1.7 to 2.4 eV (red to green) by changing radius from 6 to 2 nm. The combined toolkit lets engineers hit a target E_g to within 10 meV, enabling everything from three-junction tandem solar cells to monolithic blue-green-red LED panels to mid-IR quantum cascade lasers.