3. Special Diodes: A Tour

The pn-junction diode we just analyzed is the basic form. Tweaking the structure or doping gives a family of useful variants.

3.1 The Zener diode: engineered breakdown

A regular diode will die if you reverse-bias it past breakdown. A Zener diode is designed to operate in reverse breakdown, with a stable breakdown voltage. Available from about 2 V up to a few hundred volts.

In reverse breakdown, the voltage across a Zener is essentially constant regardless of the current (as long as the current is above some minimum). This makes them perfect as voltage references:

   V_in (varies) ──[R]──┬──── V_out (= V_zener, stable)
                        │
                        Z (zener)
                        │
                        GND

The series resistor $R$ drops the difference between $V_{in}$ and $V_{zener}$, ensuring some current flows through the Zener to keep it in breakdown. The output sits at $V_{zener}$. This is a 1960s-era voltage regulator; modern designs use more sophisticated circuits, but the Zener still appears as a building block in:

• The reference inside many bandgap circuits.
• Overvoltage protection ("crowbar" circuits — clamp the supply if it spikes too high).
• Audio clipping in guitar pedals.
• USB power-delivery flyback regulators.

A 5.6 V Zener, by happy coincidence, has nearly zero temperature coefficient — the Zener mechanism (negative tempco) and the avalanche mechanism (positive tempco) cancel almost exactly at this voltage. So 5.6 V Zeners are slightly preferred for precision references when nothing fancier is available.

3.2 The light-emitting diode (LED)

In a regular silicon diode, when an electron and hole recombine, the energy gets dumped as heat (silicon is an indirect-bandgap material — the recombination process needs to dump excess momentum into the lattice as a phonon, and a photon almost never comes out).

But in direct-bandgap semiconductors — gallium arsenide (GaAs), gallium nitride (GaN), indium gallium phosphide (InGaP) — recombination does produce a photon, with energy equal to the band gap. This is the LED.

Tune the band gap by alloying:

• Red (about 1.9 eV): GaAlAs, AlInGaP.
• Yellow/orange (2.0–2.1 eV): GaP, AlInGaP.
• Green (2.3–2.5 eV): GaP, InGaN.
• Blue (2.6–2.8 eV): GaN, InGaN. (Nichia's invention of efficient blue GaN LEDs in the 1990s won the 2014 Nobel Prize in Physics — without blue LEDs, no white LEDs, and so no LED lighting revolution.)
• Ultraviolet (>3 eV): AlGaN.
• Infrared (around 1 eV): GaAs — used in TV remotes and optocouplers.

White LEDs are typically blue InGaN LEDs with a phosphor coating that re-emits some of the blue as yellow; the eye perceives the mix as white.

LEDs have their own current/voltage characteristic, with forward voltages that depend on color (about 1.7 V for red, 3.0–3.5 V for blue/white). Always use a current-limiting resistor (or constant-current driver) — driving an LED with a constant voltage source is like driving an avalanche — exponential, ending in a smoke cloud.

3.3 The photodiode

Run an LED in reverse — that is, build the same kind of diode but optimize it for receiving light rather than emitting. When a photon with energy above the band gap is absorbed in the depletion region, it generates an electron-hole pair. The built-in field sweeps them across, yielding a small reverse current proportional to the light intensity.

Photodiodes are used in:

• Every optical-fiber receiver (the converter between light and electrical bits).
• Optical mice (sensing the surface texture).
• Smoke detectors (light-scatter sensing).
• Light meters in cameras.
• Solar cells (a photodiode optimized for area and shorted output).

The PIN photodiode is a specific structure: p-type, intrinsic, n-type. The intrinsic region in the middle is wide, providing more area for photon absorption. Fast (sub-nanosecond response possible) and the standard for all serious optical receiver work.

The avalanche photodiode (APD) runs in deep reverse bias just below avalanche breakdown. A single absorbed photon generates an electron-hole pair that picks up enough energy to ionize more atoms — internal gain of 10× to 100×. Used in long-haul fiber receivers and sensitive scientific instruments.

3.4 The Schottky diode

Built differently from a pn junction: a metal contact directly on a lightly-doped n-type semiconductor. The barrier height is set by the metal-semiconductor work-function difference rather than dopant concentration. Two key advantages:

1. Low forward voltage — typically 0.3 V instead of 0.7 V. So less power lost as heat in rectifier applications.
2. No minority carrier storage — the conduction is by majority carriers only. Reverse recovery is essentially instantaneous (picoseconds).

This makes them perfect for high-speed switching power supplies. The disadvantage: higher reverse leakage and lower breakdown voltage than pn junctions, so they are not always interchangeable.

3.5 The tunnel diode (Esaki diode)

A diode so heavily doped that the depletion region is only nanometers wide — thin enough that quantum tunneling becomes the dominant conduction mechanism. The result is bizarre: at small forward biases, current actually decreases with increasing voltage (a region of negative differential resistance) before resuming the normal exponential rise.

This negative resistance can sustain oscillation in an LC tank circuit, so tunnel diodes were the building block of microwave oscillators in the 1960s. Mostly superseded by Gunn diodes and IMPATT diodes today, but a beautiful demonstration of quantum tunneling at room temperature.

3.6 The SCR (silicon-controlled rectifier)

Not strictly a diode, but built from four alternating layers (PNPN). Behaves like a switch that, once turned on by a small gate pulse, stays on until the current is interrupted. Used in:

• Lamp dimmers (phase-cutting AC waveforms).
• Motor speed controls.
• Crowbar overvoltage protection (when triggered, it shorts the supply to GND, blowing the input fuse to protect the load).

Cousins include the TRIAC (bidirectional SCR for AC switching, in dimmer switches) and the DIAC (used to trigger TRIACs).

3.7 The UJT (unijunction transistor)

A specialty three-terminal device with one pn junction. Behaves like a relaxation-oscillator building block — used in old simple oscillator circuits and pulse generators. Largely obsolete; the 555 timer (Chapter 6) replaced it for almost everything.