🔥 0 ⚡ 0 XP

THE MATERIALS ERA

Shrinking alone stopped working

The leak SiO₂ thin enough for good control tunnels too much current (Unit 27)
The depletion Polysilicon gates deplete at the oxide, costing about 0.4 nm of control
The mobility Carriers stopped getting faster as channels got shorter
The answers Strain in 2003 (90 nm), high-k metal gates in 2007 (45 nm), then FinFETs in 2011 (22 nm)
Materials and geometry took over from pure shrinking.

HIGH-K DIELECTRICS

The same control, a thicker film

EOT = tfilm × 3.9 ÷ k

Capacitance depends on the dielectric constant k divided by thickness. Replace SiO₂ (k = 3.9) with hafnium oxide (k near 20) and a film five times thicker has the same capacitance. Tunneling falls exponentially with thickness, so the leakage drops by several orders of magnitude. The equivalent oxide thickness, EOT, says how thick SiO₂ would have to be for the same capacitance.

High-k lets the gate stay strong without letting the electrons through.

METAL GATES

Out with polysilicon

No depletion A metal has no depletion layer, so the 0.4 nm penalty disappears
Work function The metal's work function sets Vth, so threshold no longer needs heavy channel doping
Two metals Different metal stacks serve NMOS and PMOS
Gate-last High-k and metal cannot survive the hottest anneals, so a placeholder gate is replaced after them
High-k only works well together with a metal gate.

STRAINED SILICON

Squeeze or stretch the crystal

Pulling the silicon lattice apart (tensile strain) helps electrons; squeezing it (compressive) helps holes. Chipmakers grow SiGe, whose crystal is larger, in the PMOS source and drain to squeeze the channel, and deposit stressed nitride films over NMOS to stretch it. It raises the drive current without changing the shape of the transistor.

Strain buys speed by changing how carriers move, not how many there are.

CHECK THE UPGRADES

What problem does it solve?

High-k keeps the control, metal fixes the gate, strain speeds the carriers.

UNIT 16 STUDY COMPLETE

🧱

Ready for the Fab Challenge?

You've seen how new materials, not just smaller shapes, kept transistors improving through the 2000s.