Why "forward" and "reverse" actually change the physics of the junction — not just which way current flows.
It's tempting to think of forward and reverse bias the way you'd think of flipping a battery on a resistor — same circuit, current just runs the other way. A PN junction doesn't work like that. The direction of the applied voltage doesn't just redirect current — it physically reshapes the depletion region at the junction, either tearing down the barrier to current or reinforcing it. That reshaping is the entire reason a diode conducts one way and (almost) not the other.
The moment p-type and n-type silicon are joined, nothing external has to happen for a barrier to form. The p-side has an abundance of free holes; the n-side has an abundance of free electrons. Right at the interface, electrons diffuse into the p-side and holes diffuse into the n-side, and they recombine there. That recombination strips the region right around the junction of its free carriers, leaving behind only the fixed, immobile dopant ions — negative acceptor ions on the p-side, positive donor ions on the n-side. Those exposed ions set up a built-in electric field pointing from n toward p, and a built-in potential barrier (roughly 0.6–0.7V for silicon at room temperature) that opposes any further diffusion. Left alone, the junction reaches equilibrium: diffusion pushing carriers across is exactly balanced by the built-in field pushing them back. That equilibrium depletion region — carrier-free, with a fixed width — is what an applied bias voltage then either shrinks or grows.
The built-in field at an unbiased junction exists to hold diffusion in check — it's doing real work even with no external voltage applied. Forward bias (+ to p, − to n) pushes an external field in the opposite direction of the built-in field, partially cancelling it. As the net barrier shrinks, diffusion current wins out over drift, the depletion region narrows, and once the applied voltage climbs past roughly 0.7V(silicon's built-in potential), majority carriers flood across essentially unopposed and current rises exponentially with voltage. Reverse bias does the opposite: it adds to the built-in field instead of opposing it, which pulls majority carriers even further from the junction, widens the depletion region well beyond its equilibrium width, and raises the barrier higher. With a taller, wider barrier and no majority carriers left nearby to cross it, only the trickle of minority carriers that thermal energy happens to generate near the junction can get through — a nearly bias-independent leakage current, typically nanoamps to microamps, until the field gets strong enough to trigger avalanche or Zener breakdown.
This is the single most common way people over-generalize Ohm's law onto a device it doesn't apply to. Flip a battery across a resistor and nothing about the resistor changes — the current simply reverses direction with the exact same magnitude, because a resistor is symmetric: I = V/R holds identically in either direction. A PN junction is fundamentally asymmetric, and that asymmetry is the whole point. Reversing the applied voltage doesn't just flip the sign of the current — it changes the physical structure of the device: the depletion region narrows in one direction and widens substantially in the other, so the two directions present wildly different resistance to current flow (roughly mA–A of forward current versus nA–µA of reverse leakage, easily six or more orders of magnitude apart). That's not a detail — it's the entire reason a diode can rectify AC into DC instead of just passing an alternating current through unchanged in both directions. A resistor could never do that job, no matter which way you wire the battery.
Explains why applying a forward or reverse voltage to a PN junction physically changes the width of its depletion region — not just the direction of current — using a side-by-side comparison of the junction under forward and reverse bias, and why that asymmetry is what makes a diode rectify AC in the first place.
Most people meet Ohm's law first, where a device's behavior is symmetric — reverse the voltage and the current simply reverses too, at the same magnitude. It's natural to assume a diode behaves the same way with the labels "forward" and "reverse" just describing which way current happens to go. In reality, a PN junction has no fixed resistance at all; its effective resistance depends enormously on bias direction because the applied voltage restructures the depletion region itself, not merely the current's direction.
At equilibrium (no applied bias), diffusion of majority carriers across the junction and recombination near it leave behind a carrier-depleted region bounded by exposed dopant ions, producing a built-in electric field and potential barrier (~0.6–0.7V for silicon). Forward bias (+ to p, − to n) opposes this built-in field, shrinking the depletion region; once the applied voltage exceeds the built-in potential, majority carriers cross freely and current rises exponentially with voltage. Reverse bias (+ to n, − to p) reinforces the built-in field instead, pulling majority carriers further from the junction and widening the depletion region well beyond equilibrium, which raises the barrier and permits only a small, roughly voltage-independent minority-carrier leakage current — until the electric field becomes strong enough to trigger avalanche or Zener breakdown.
This forward/reverse asymmetry is the working principle behind every rectifier diode, LED, photodiode, varactor (which exploits the bias-dependent depletion width as a voltage-controlled capacitance), and every BJT and MOSFET junction used for switching or amplification. Understanding depletion-region behavior — not just "diodes pass current one way" — is what lets you reason correctly about diode capacitance, reverse recovery time, breakdown voltage ratings, and why real diodes need a minimum forward voltage before they conduct meaningfully at all.
That figure is approximately the built-in potential of a silicon PN junction at room temperature, set by the doping concentrations and silicon's bandgap. Below it, the depletion region hasn't been narrowed enough for majority carriers to cross in significant numbers; once applied voltage approaches and exceeds it, current rises exponentially, which is why the forward curve looks like a sharp "knee" near 0.7V rather than a gradual ramp. Other semiconductors have different built-in potentials — germanium is closer to 0.3V, and typical LEDs (wider bandgap materials) range from about 1.8V to 3.3V or more depending on color.
It's physically real and measurable — typically tens of nanometers to around a micron wide depending on doping levels, and it's exactly why PN junctions are used as voltage-controlled capacitors (varactor diodes): the depletion region acts as a dielectric gap whose width (and therefore capacitance) changes with reverse bias voltage.
Once reverse voltage is large enough, one of two mechanisms takes over: avalanche breakdown, where the electric field accelerates minority carriers enough that they knock loose additional carriers through impact ionization, multiplying current rapidly; or Zener breakdown, dominant in heavily-doped junctions with very thin depletion regions, where the field becomes strong enough to directly pull electrons out of covalent bonds (quantum tunneling). Both cause reverse current to rise sharply at a fairly well-defined voltage, which is exactly the effect exploited deliberately in Zener diodes for voltage regulation.
Thermal energy continuously generates electron-hole pairs throughout the semiconductor, including within and near the depletion region. Under reverse bias, the electric field readily sweeps any minority carriers generated near the junction across it, producing a small reverse saturation current that depends mainly on temperature and material properties, not on the exact reverse voltage (until breakdown is approached). It's small, but it's never exactly zero.
Yes, the same PN junction and depletion region are central to both, just used in different directions. An LED is a forward-biased junction where recombination across the narrowed depletion region releases energy as photons instead of just letting carriers pass. A photovoltaic solar cell is the reverse concept: light striking the junction (often operated near zero or reverse-ish bias) generates electron-hole pairs that the built-in field separates and drives out as current, without needing an external voltage source to start the process.
Try our Semiconductor & Electronics Studio
More calculators, simulators, and guides for this discipline.