How semiconductor doping work

01 Pure silicon is not quite enough

Silicon is useful because its conductivity sits between that of a metal and an insulator, and because its crystal can be manufactured with extraordinary uniformity. But an undoped crystal is a poor switch. Its electrons and holes are created mainly by temperature, so the number of available carriers changes with conditions rather than with a circuit designer's intent.

Doping solves that problem by inserting a very small concentration of chosen impurity atoms into the lattice. The impurity does not replace silicon's crystal structure wholesale. It changes the local electron count and adds energy states that make one type of mobile carrier much easier to produce.

02 Donors make electrons available

Silicon forms four covalent bonds. Substitute a pentavalent atom such as phosphorus, arsenic or antimony and four of its outer electrons can participate in those bonds while a fifth is only weakly held. At ordinary operating temperatures, that extra electron can be promoted into the conduction band. The impurity is called a donor, and the resulting material is n-type.

The important idea is not that every dopant atom becomes a free electron. Activation, compensation and temperature determine the fraction that contributes. In a simple, lightly doped region, though, donor concentration is a practical control knob for electron concentration and therefore conductivity.

03 Acceptors leave holes behind

A trivalent atom such as boron has only three valence electrons available for silicon's four-bond pattern. The missing bond electron behaves as a mobile hole: an absence that can be filled by a neighboring electron, making the apparent positive charge move through the lattice. This creates p-type silicon.

Holes are not literal particles drifting through empty space. They are a useful and experimentally predictive description of how the valence-band electron population rearranges. A designer can therefore choose whether a region conducts mainly through electrons or holes, and can combine the two in one crystal.

04 The junction creates a barrier

Put p-type and n-type regions next to one another and carriers diffuse across the boundary: electrons move toward the p side and holes toward the n side. Near the interface they recombine, exposing fixed ionized dopants. Those charges create an electric field that opposes further diffusion.

The result is a depletion region with few mobile carriers and a built-in potential. Forward and reverse bias change the width and height of this barrier. That asymmetry is why a p-n junction can rectify current, emit light, detect photons or become part of a transistor.

05 From junctions to transistor control

A transistor arranges multiple doped regions so a small electrical input changes a much larger current. In a bipolar transistor, the emitter and collector are separated by a thin base; in a MOSFET, a gate's electric field creates or removes a conducting channel near the surface. In both cases, doping sets the background carrier population, junction depths and fields that make the control geometry work.

This is why “how a transistor works” cannot be separated from materials engineering. A schematic symbol hides concentration gradients measured in atoms per cubic centimeter, nanometer-scale interfaces and thermal histories that determine whether the intended profile survives fabrication.

06 Concentration is a spectrum, not a switch

Real chips use many doses. Light doping can create a wide, low-field region; heavier doping lowers resistance for contacts; carefully graded profiles manage breakdown voltage and capacitance. At very high concentration, the semiconductor can become degenerate: its behavior approaches a metal-like carrier population and simple textbook approximations need correction.

07 What the factory must preserve

Dopants are introduced by diffusion or ion implantation, then activated with heat. Implantation offers precise dose and depth control but damages the lattice; annealing repairs much of that damage and moves atoms onto electrically useful lattice sites. Cleaning, masks and metrology keep the profile from wandering into neighboring features.

Electrical tests such as sheet-resistance and carrier-profile measurements turn the invisible chemistry into process data. If a dose is correct but the junction diffuses too far, a transistor can leak; if activation is poor, its channel or contact becomes too resistive. Doping works because fabrication treats it as a measured process, not a one-time recipe.

08 The small change that scales to computation

At the atomic level, doping is a sparse substitution. At the system level, it lets engineers place barriers, channels and contacts with repeatable electrical behavior. Billions of such regions can then cooperate as logic gates, memory cells, sensors and light sources.

That scale is the central lesson: semiconductor doping is a way of converting chemistry into geometry and geometry into computation. The transistor video above shows the device-level consequence; the dopant profile is the quiet precondition that makes the consequence manufacturable.