The Physics of Semiconductor Devices: Bandgaps, Doping, and Real-World Failures

The physics of semiconductor devices describes how manipulating the atomic crystal lattice and electron energy bands of materials like silicon controls electrical conductivity to switch, amplify, or rectify current. It is the fundamental reason your multimeter reads a 0.6V drop across a diode and the reason a MOSFET will violently self-destruct if you misuse it in its linear region. Understanding this physics changes how you design real circuits by dictating forward voltage drops, leakage currents, thermal limits, and switching speeds.

The Core Mechanism: Bandgaps and Charge Carriers

In a pure silicon crystal, electrons are locked into covalent bonds. To conduct electricity, an electron must break free from this bond and jump into the conduction band. The energy required to make this jump is the bandgap. For silicon at room temperature (300K), this bandgap is exactly 1.12 eV (electron volts).

Think of the bandgap as a toll booth on a highway. Electrons in the valence band are parked cars; they need exactly 1.12 eV of energy (the toll) to jump into the conduction band (the express lane) and carry current. Thermal energy (heat) or photon energy (light) pays this toll. This is why semiconductors conduct better when hot—the ambient thermal energy pays the toll for more electrons—whereas metallic conductors (like copper wire) conduct worse when hot due to increased atomic lattice scattering.

Doping and the PN Junction: A Numeric Breakdown

Pure silicon is a poor conductor. We alter its physics by 'doping' it with impurities. Adding phosphorus (5 valence electrons) creates N-type silicon with excess free electrons. Adding boron (3 valence electrons) creates P-type silicon with excess 'holes' (missing electrons). When you press N-type and P-type together, you get a PN junction—the heart of every diode and bipolar transistor.

Let's look at the math governing a standard 1N4148 signal diode using the Shockley diode equation:

I = I_s × (e^{V_D / (n × V_T)} - 1)

Worked Numeric Example:
What is the current through a 1N4148 if we apply exactly 0.6V across it at room temperature?

1. Calculate the thermal voltage multiplier: n × V_T = 1.75 × 0.02585V = 0.04523V.
2. Divide the applied voltage by this multiplier: 0.6V / 0.04523V = 13.26.
3. Calculate the exponential term: e^13.26 ≈ 577,336.
4. Subtract 1 (which is negligible here) and multiply by I_s: 2.52 nA × 577,336 = 1,454,886 nA.
5. Convert to milliamps: 1.45 mA.

If you bump the diode up to 0.7V, the current jumps to roughly 10.5 mA. This exponential relationship is why a tiny voltage increase causes a massive current spike, necessitating a current-limiting resistor in your circuit.

Where You Meet This in Practice

The quantum physics of the bandgap directly dictates macroscopic component selection on your bench. Here is what the physics changes in a real installation:

• LED Color Selection: The bandgap determines the photon energy released when an electron recombines with a hole. A standard red LED uses GaAsP (bandgap ~1.8 eV), while a blue LED requires GaN (bandgap ~3.4 eV). You cannot make a blue LED out of silicon; the physics forbids it.
• Power Dissipation: The 1.12 eV bandgap of silicon results in a practical forward voltage (V_f) of about 0.6V to 0.7V. In a 10A rectifier circuit, that V_f means 6W to 7W of heat per diode that you must sink away.
• Leakage Current Scaling: Because thermal energy excites carriers across the bandgap, reverse leakage current in a silicon PN junction roughly doubles for every 10°C rise in temperature. A circuit that draws 1 µA of leakage at 25°C will draw 16 µA at 65°C.

Material	Bandgap (eV)	Typical Vf / Application	Max Junction Temp
Germanium (Ge)	0.67	~0.3V / RF detection	85°C
Silicon (Si)	1.12	~0.7V / General purpose	150°C
Silicon Carbide (SiC)	3.26	~1.5V / High-voltage EV inverters	175°C - 200°C
Gallium Nitride (GaN)	3.40	N/A (FETs) / Blue LEDs	150°C - 200°C

For deeper reading on atomic lattice structures, the Georgia State University HyperPhysics database provides excellent foundational diagrams. For practical circuit applications, the All About Circuits semiconductor textbook bridges the gap between theory and bench work.

Real-World Scenario Walkthrough: The MOSFET Thermal Runaway

Ignoring the internal physics of a semiconductor device is the fastest way to let the magic smoke out. Here is a classic bench failure involving the physics of the parasitic BJT hidden inside every power MOSFET.

The Numbers: The gate-source voltage (V_GS) is set to 4.5V. At 25°C, the IRF520 passes about 2A. The voltage drop across the MOSFET is 6V. Power dissipation is P = V × I = 6V × 2A = 12W. The hobbyist attaches a small heatsink rated for 5°C/W. The junction temperature rises by 60°C, settling at 85°C.

The Outcome: Suddenly, the current spikes to 15A, the MOSFET cracks audibly, and the 12V rail sags. The silicon die has suffered second breakdown.

What Went Wrong (The Physics): This is known as the Spirito effect. A power MOSFET is not a single monolithic switch; it is thousands of tiny parallel MOSFET cells on the die. The threshold voltage (V_th) of a silicon MOSFET has a negative temperature coefficient of roughly -10 mV/°C. As the die heats up, the local hotspots require less gate voltage to turn on. Because V_GS is fixed at 4.5V, the hotter cells draw disproportionately more current. This causes them to heat up faster, drawing even more current, in a localized thermal runaway loop. Eventually, the local temperature triggers the parasitic NPN bipolar transistor inherent in the MOSFET's physical structure, shorting the drain to the source and destroying the part. Modern SiC MOSFETs (like the C3M0060120D) are highly susceptible to this if not driven with proper negative turn-off gate bias.

Common Confusions and Bench Mistakes

When makers misunderstand semiconductor physics, they make predictable mistakes:

• Confusing Macroscopic Resistance with Channel Resistance: A resistor obems Ohm's law linearly. A semiconductor channel does not. You cannot calculate a MOSFET's power dissipation in the linear region using simple I²R without accounting for the severe temperature coefficients of the silicon lattice.
• The '0.7V Diode' Myth: People assume all diodes drop 0.7V. This is only true for standard silicon PN junctions at moderate currents. A Schottky diode (metal-semiconductor junction) drops ~0.3V. A SiC diode drops ~1.5V. Always check the datasheet's V_f vs I_f curve.
• Avalanche vs. Zener Breakdown: People use 'Zener diode' as a catch-all term. Physically, true Zener breakdown is a quantum tunneling effect that only happens in diodes rated below 5V. Diodes rated above 5V (like a 12V 'Zener') actually rely on avalanche breakdown (impact ionization), which has a positive temperature coefficient, whereas true Zener tunneling has a negative temperature coefficient.

Frequently Asked Questions