In short
A bipolar junction transistor (BJT) is a three-layer sandwich of doped silicon with two p-n junctions and three external leads: emitter (E), base (B), and collector (C). The two varieties mirror each other:
- npn: heavily doped n emitter, very thin lightly-doped p base, moderately doped n collector.
- pnp: heavily doped p emitter, very thin lightly-doped n base, moderately doped p collector.
The magic of the BJT is the geometric asymmetry: the base is deliberately made thinner than one micrometre and lightly doped, so that most of the carriers injected into it from the emitter do not recombine on the way across — they diffuse through the base and are swept into the collector by the reverse-biased base-collector junction.
In active operation (the only mode useful for amplification) the emitter-base junction is forward biased (about 0.7 V) and the base-collector junction is reverse biased (a few volts). A small base current I_B controls a much larger collector current I_C through the relation
This is current amplification. Push 20 microamperes into the base of a BC547 transistor and 2 milliamperes flow through the collector — one hundred times the signal. Kirchhoff's current law closes the loop: I_E = I_B + I_C \approx I_C. The whole device is a current-controlled current source. Every audio amplifier in every Bajaj home speaker, every switching stage in every Indian set-top box, every high-side driver on every ISRO satellite controller rests on this single three-terminal fact.
Take the p-n junction diode from the previous article and glue a second doped region onto the other side of the p slab. You now have three regions — n, then p, then n again — two junctions, three wires. Done well, this little sandwich does something neither a diode nor a resistor can: it uses a tiny control signal (a few microamperes at the base) to regulate a much bigger main current (a few milliamperes through collector-emitter). That ratio of control-to-controlled — the current gain — is what makes a transistor a transistor. Pull the control away and the main current stops. Push the control a touch harder and the main current grows a hundredfold.
This is the invention that split the history of electronics into before and after. Before: vacuum tubes the size of a rupee coin per logic gate, filament-heated, fragile, expensive. After: a sliver of doped silicon the size of a pinhead, room-temperature, cheap, and replicable a billion times on a single wafer. Every IC on every Tata Electronics chip fabricated at Dholera, every audio amplifier stage in every Bajaj or Ahuja public-address speaker at an Indian wedding, every digital stage inside the Agni missile's inertial guidance computer — built on this one structure.
This article derives the structure, explains the action, and gives you the current-gain equation I_C = \beta I_B from first principles. The next article builds on this foundation to analyse the common-emitter amplifier and the transistor-as-switch. Between them you will have the full mental model for how a transistor earns its keep.
The three-layer structure
Start by picturing a wafer of silicon. Dope one end heavily with phosphorus — make it strongly n-type, with donor density N_d \sim 10^{25}\ \text{m}^{-3} (one phosphorus per 5000 silicons, a degenerate doping). Next to it, dope a very thin layer — typical width W_B \sim 0.5\ \mu\text{m}, roughly one-hundredth the thickness of a human hair — with boron to make it lightly p-type, acceptor density N_a \sim 10^{22}\ \text{m}^{-3}. On the far side, dope a moderately-thick region with phosphorus again, with N_d \sim 10^{23}\ \text{m}^{-3}. Solder thin aluminium contact wires to each of the three regions. You now have a bipolar junction transistor.
The three regions and their external wires have names:
| Region | Doping (typical) | External lead |
|---|---|---|
| Emitter | heavily doped (10^{25} m⁻³) | E |
| Base | very thin, lightly doped (10^{22} m⁻³), width \sim 0.5\ \mu\text{m} | B |
| Collector | moderately doped (10^{23} m⁻³), wide enough to dissipate power | C |
The pnp variant is exactly the mirror image: heavily doped p emitter, thin lightly doped n base, moderately doped p collector. Everything that follows for the npn holds for the pnp with all signs reversed — electrons become holes, forward biases flip, current directions reverse. For clarity the rest of this article works the npn case; the pnp is a relabelling exercise at the end.
Why the base is so thin and so lightly doped
This is the single most important design choice in the BJT, and it is worth pausing on. The thin, lightly-doped base is the mechanical reason the transistor amplifies at all.
Imagine a thick, heavily-doped base. Electrons injected from the emitter into the base would, on average, meet and recombine with a majority hole within a distance of perhaps a micrometre — the diffusion length of electrons in p-type silicon. Make the base 50 μm thick and basically every injected electron recombines before reaching the collector. The current does not make it across. The device is just two diodes back-to-back, with no amplification.
Now make the base 0.5 μm thick — much shorter than the electron diffusion length in the base. Injected electrons have almost no time to find a hole. They diffuse straight through, reach the edge of the base, and are then swept into the collector by the strong reverse-bias field of the base-collector junction. A fraction \alpha \approx 0.99 of the carriers injected from the emitter make it to the collector; only about 1% recombine in the base and are collected at the base lead. That 1% is the base current; the 99% is the collector current; their ratio is \beta = 99/1 = 99. A thin base gives a large \beta.
The light doping in the base serves a related purpose: it suppresses reverse hole-injection from the base into the emitter. With N_d^E \gg N_a^B, carriers crossing the E-B junction are dominantly electrons going p-to-n-ward from emitter to base, not holes going the other way. This keeps the "useful" current fraction high.
Circuit symbols and the arrow convention
Before going further it helps to know how BJTs appear in a circuit diagram. The symbol is a circle (sometimes omitted) containing the three leads, with an arrow on the emitter lead indicating the direction of conventional current flow during normal (active) operation.
A mnemonic Indian students often use: "npn — not pointing in." If the emitter arrow points out of the device, it is an npn; if it points into the base, it is a pnp. Every BC547 (npn, small-signal), every 2N2222 (npn, general purpose), every BC557 (pnp complement to BC547) you will encounter on a school breadboard follows this convention.
Biasing: putting the junctions into "active" mode
A BJT is only useful when each of its two junctions is biased correctly. There are four combinations — the two junctions can each be forward or reverse biased — and they have names:
| Emitter-base | Base-collector | Region | Use |
|---|---|---|---|
| reverse | reverse | cutoff | both junctions off; I_C \approx 0 |
| forward | reverse | active | amplification |
| forward | forward | saturation | both junctions on; V_{CE} small |
| reverse | forward | inverse active | rarely used; low \beta |
The only mode useful for linear amplification is the active region: emitter-base forward biased (about V_{BE} = 0.7\ \text{V} for silicon, same as a diode's turn-on voltage), base-collector reverse biased (a few volts). In a switching circuit (see the next article) the BJT alternates between cutoff (fully off, no current) and saturation (fully on, small voltage drop) — binary digital behaviour.
For an npn in the active region, the voltage conventions are: V_{CE} > V_{BE} > 0, so the collector sits at the highest potential, the base next, the emitter at ground. All three terminal currents flow into the device in the conventional sense shown by the symbol arrow:
Why: whatever flows in at the base and collector terminals must come out at the emitter terminal (or vice-versa for a pnp). The transistor is just a node as far as Kirchhoff is concerned — no charge accumulates inside.
This equation is the first of two we will need. The second — the defining equation — is the current gain relation between I_C and I_B.
How the carriers move: the microscopic story
Let's walk through what happens at the atomic level once the biases are applied. Track the electrons (majority carriers in the n-type emitter) as they move through the device.
Stage 1: emitter injects electrons into the base. The forward bias V_{BE} \approx 0.7\ \text{V} lowers the emitter-base barrier. Electrons from the heavily-doped n emitter diffuse across the junction into the p base. Because the emitter is much more heavily doped than the base (N_d^E \gg N_a^B), this current is dominated by electrons n→p, not holes p→n. Call this injected electron current density J_n^E.
Stage 2: electrons diffuse across the thin base. Once in the base, these electrons are minority carriers (the base is p-type, so electrons are outnumbered). They diffuse from the high-concentration edge (emitter side) to the low-concentration edge (collector side) of the base. Because the base width W_B is much less than the electron diffusion length L_n in the base, most electrons — fraction \alpha_T called the base transport factor, typically \alpha_T \approx 0.99 — make it across without recombining.
Stage 3: the reverse-biased base-collector junction sweeps them into the collector. At the base-collector junction, any electron arriving from the base side sees a steep downward potential slope (the reverse bias). It is swept across into the collector the way a skater drops into a half-pipe. These electrons leave the device through the collector lead as I_C.
Stage 4: the small fraction that recombined. A few percent of the electrons injected into the base found a hole in transit and recombined. That recombination used up a hole, so a hole must be delivered from the external base lead to maintain charge neutrality in the base. This replenishing hole current is the external base current I_B.
So of 100 electrons injected from the emitter, about 99 reach the collector and 1 recombines in the base. That is the entire microscopic story. What you call amplification at the terminal level is simply the geometric fact that the base is thin.
The current-gain relations
Define two gain parameters:
Common-base current gain \alpha: the fraction of emitter current that reaches the collector.
Why: \alpha is the single number that captures the base-transit efficiency and the emitter-injection efficiency together. It is slightly less than 1 because some of I_E is "stolen" by recombination in the base, and a tiny part is hole-injection back into the emitter.
Common-emitter current gain \beta: the ratio of collector current to base current.
Why: \beta is what matters for an amplifier. If you drive the base with a small signal current, the collector responds with a current \beta times larger. A small wiggle at the base produces a big wiggle at the collector.
Relating \alpha and \beta
These two numbers are not independent — they are two views of the same physics. Here is the short derivation.
Step 1. Start with Kirchhoff (equation 1): I_E = I_B + I_C.
Step 2. Divide both sides by I_C.
Why: we want to express things in terms of I_B/I_C because that ratio is exactly 1/\beta.
Step 3. Recognise I_E/I_C = 1/\alpha (from equation 2) and I_B/I_C = 1/\beta (from equation 3).
Step 4. Rearrange to isolate \beta.
Why: this is the key algebraic link between the two gains. \alpha is close to 1 and \beta is large because a denominator 1 - \alpha that is close to zero gives a large quotient. For \alpha = 0.99: \beta = 0.99 / 0.01 = 99. For \alpha = 0.995: \beta = 0.995 / 0.005 = 199. Every extra decimal of \alpha doubles \beta.
Why this ratio is so sensitive
Equation (4) tells you something important about transistor design. To make \beta = 100 you need \alpha = 0.99 — one recombination loss in a hundred. To make \beta = 300 you need \alpha = 0.9967 — only three losses in a thousand. A tiny change in manufacturing quality (base thickness uniformity, dopant purity, surface cleanliness) produces a big change in \beta. That is why the \beta values of "identical" BC547 transistors straight off the reel in a 2025 Bajaj Electronics procurement lot can vary from 110 to 800 — a spread of nearly an order of magnitude. Good transistor design and good engineering practice both account for this: a well-designed amplifier circuit has a gain that depends on external resistors, not on \beta, so the circuit works correctly for any transistor in the reel (see the next article for the common-emitter derivation).
Explore the I_C = \beta I_B relation
The interactive figure below lets you slide the base current and watch the collector current rise in proportion. The slope of the line is \beta; here it is fixed at \beta = 150 (a typical BC547 value at room temperature in the active region). This is the transfer characteristic of the common-emitter configuration.
Worked examples
Example 1: A BC547 driving a red LED
A BC547 transistor (current gain \beta = 200) is wired as a switch to drive a red LED. The base is pushed with I_B = 25\ \mu\text{A} by a resistor from a 5 V logic line. Find (a) the collector current I_C and (b) the emitter current I_E, assuming the transistor stays in the active region.
Step 1. Write the defining relation.
Why: equation (3). In active mode the collector current is the base current amplified by \beta. Assuming we stay in active mode (the collector has not saturated because the LED and R_C drop does not pull V_{CE} close to zero), this is all the physics we need.
Step 2. Substitute.
Why: a microampere times a (dimensionless) number is still microamperes. 5000 μA is a convenient 5 mA — a typical LED drive current that gives bright red light on a standard 20 mA-rated indicator LED.
Step 3. Apply Kirchhoff's current law for I_E.
Why: equation (1). Since I_B is a tiny fraction of I_C, the emitter current is essentially equal to the collector current. This is why circuit designers often say "I_E \approx I_C" without losing meaningful accuracy.
Result: I_C = 5\ \text{mA}, I_E = 5.03\ \text{mA}. The LED is driven brightly by a base signal that a 100 kΩ resistor can deliver from any 5 V microcontroller output pin.
What this shows: The transistor has amplified a 25 μA signal into a 5 mA drive — a factor of 200. That factor is \beta, and it is determined entirely by the transistor's internal geometry (base width, doping), not by any external resistor. This is the fundamental asymmetry that makes the BJT useful.
Example 2: Running a transistor backward — why pnp signs flip
A pnp transistor (the BC557, \beta = 180) is connected with its emitter at +5 V, collector at 0 V through a resistor, and base at about +4.3 V (0.7 V below the emitter). A base current |I_B| = 40\ \mu\text{A} flows out of the base into the external circuit. Find the collector current and its direction.
Step 1. Write the current-gain relation. For a pnp, it is the same magnitude:
Why: the physics is symmetric under swapping electrons and holes. Only the directions reverse. We will track magnitudes first, then put directions in at the end.
Step 2. Compute the collector current magnitude.
Step 3. Fix the direction. For a pnp in active mode, conventional current flows out of the emitter (holes, which are positive, exit the emitter internally). So at the external terminals, conventional current:
- flows into the emitter lead (from the +5 V supply),
- flows out of the base lead (the 40 μA),
- flows out of the collector lead (the 7.2 mA, returning to ground through R_C).
Why: holes are the majority carriers in a pnp. Holes are positive. Hole flow and conventional current point the same way. So every arrow points opposite to its npn cousin.
Step 4. Check Kirchhoff.
Why: Kirchhoff's current law works just as well for a pnp, only with the reversed sign convention. Whether the charge carriers are electrons or holes, charge into a node equals charge out.
Result: Collector current magnitude 7.2 mA, flowing out of the collector lead (and into ground through R_C). Emitter current magnitude 7.24 mA, flowing into the emitter lead from the +5 V supply.
What this shows: The pnp is the complementary mirror of the npn. Every sign flips, every equation keeps its form, every physical interpretation survives. Circuit designers exploit this: a complementary pair (one npn + one pnp with matched \beta) is the building block of the push-pull output stage in every audio amplifier, from the humblest Bajaj speaker in a Pune college auditorium to the studio monitors at All India Radio.
Common confusions
-
"The base current is a leakage current." No. The base current is essential — it is the replenishment of holes needed to keep the base electrically neutral as carriers transit. Starve the base of current and the transistor shuts off completely (this is the cutoff mode, which is exactly how a transistor is used as an "off" switch).
-
"\beta is a fundamental constant like \pi." No. \beta is a manufacturing parameter that depends on the base width, dopant profiles, surface recombination, and temperature. A BC547 can have \beta anywhere from 110 to 800 depending on the specific "bin" (grade) you bought. Well-designed circuits make their gain depend on external resistors, not on \beta.
-
"An npn transistor is just two back-to-back diodes." In a curve-tracer test it looks like it (E-B junction is a diode, B-C junction is a diode), but the interaction between the two junctions — which relies on the thin base and is not present in two discrete diodes wired together — is what makes it a transistor. Two back-to-back diodes wired in a breadboard will not amplify. The continuous crystal is the whole point.
-
"A forward-biased B-C junction means active mode." Backwards. Active mode is: emitter-base forward biased, base-collector reverse biased. If both junctions are forward biased you are in saturation — the collector voltage has dropped so low that the B-C junction forward biases, and I_C can no longer grow with I_B. (This is the "on" state when the transistor is used as a switch.)
-
"The collector is the input, the emitter is the output." Backwards. The emitter injects carriers (it is the carrier source in active mode). The collector collects them. The base is the control. In a common-emitter amplifier, the signal is applied between base and emitter, and the amplified output is taken between collector and emitter — see the next article for this configuration.
-
"Silicon transistors work at any temperature." Not quite. Below about -40\ ^\circ\text{C}, dopant freeze-out reduces the carrier density. Above about +125\ ^\circ\text{C} (standard industrial grade), intrinsic carrier density n_i starts to compete with the doping, and the transistor loses its well-defined doping identity. ISRO's space-qualified transistors for the Chandrayaan missions are specially rated to +200\ ^\circ\text{C} using wider-bandgap silicon carbide — but at a significant cost premium.
If you came here to understand what a transistor is and how it amplifies, you have what you need. What follows is the derivation of \beta from the base-transit physics, the Ebers-Moll model that unifies all four operating regions, and the thermal-runaway failure mode that keeps power-electronics engineers awake at night.
The base transport factor from diffusion
The claim "99% of electrons injected into the base reach the collector" can be made quantitative. Inside the base, the excess minority-electron concentration n(x) (where x runs from 0 at the emitter-base edge to W_B at the base-collector edge) satisfies the steady-state diffusion equation with recombination:
where D_n \sim 25\ \text{cm}^2/\text{s} is the electron diffusion coefficient in p-silicon and \tau_n \sim 10\ \mu\text{s} is the minority-carrier lifetime. Define the diffusion length L_n = \sqrt{D_n \tau_n} \sim 160\ \mu\text{m}.
Boundary conditions. At x = 0 (emitter edge), the forward bias injects excess electrons: n(0) = n_{p0}\bigl(e^{eV_{BE}/k_BT} - 1\bigr) where n_{p0} is the equilibrium electron density in the base. At x = W_B (collector edge), the reverse bias sweeps electrons out the instant they arrive, so n(W_B) \approx 0.
Solution. The general solution is n(x) = A \sinh((W_B - x)/L_n) + B \sinh(x/L_n). Applying the boundary conditions:
Why: at x = 0 the hyperbolic-sine numerator equals \sinh(W_B/L_n), matching the denominator to give n(0). At x = W_B the numerator is \sinh(0) = 0, matching the zero at that boundary.
Electron current at each edge. Diffusion current is J_n = eD_n \, dn/dx. Evaluating at x = 0 and x = W_B:
The base transport factor is the ratio:
Why: the coth over 1/sinh simplifies using \cosh = \cosh. For W_B / L_n \ll 1 (thin base), Taylor-expand: \cosh(x) \approx 1 + x^2/2, so \alpha_T \approx 1 - (W_B/L_n)^2/2. For W_B = 0.5\ \mu\text{m} and L_n = 160\ \mu\text{m}: \alpha_T \approx 1 - (0.0031)^2/2 \approx 1 - 5 \times 10^{-6} — transit losses are microscopic.
The emitter injection efficiency
Not all of the emitter current is electrons injected into the base; a small hole current goes the wrong way from base into emitter. Define the emitter injection efficiency \gamma = J_n / (J_n + J_p), the fraction of the E-B junction current carried by electrons. For N_d^E \gg N_a^B (heavy emitter doping):
The common-base gain is then \alpha = \gamma \cdot \alpha_T. In practice both factors are individually near 1, so \alpha is very close to 1, and \beta = \alpha/(1-\alpha) is large.
The Ebers-Moll model in one glance
The four operating regions (cutoff, active, saturation, inverse-active) can all be captured in one pair of equations that generalises the Shockley diode equation to the two-junction BJT:
Here V_T = k_B T/e \approx 25.85 mV at 300 K, I_{ES} and I_{CS} are the saturation currents of the emitter-base and base-collector junctions in isolation, and \alpha_F, \alpha_R are the forward and reverse common-base gains. The reciprocity relation \alpha_F I_{ES} = \alpha_R I_{CS} ties them together by time-reversal symmetry of the device.
Pick any bias combination — say V_{BE} = +0.7 V and V_{BC} = -3 V (active mode) — substitute, and you get the same I_C = \beta I_B we derived from carrier-counting. The Ebers-Moll model is just that derivation written in compact form for all four quadrants simultaneously.
Thermal runaway — why power BJTs fail
Collector current has a hidden temperature dependence: I_C at fixed V_{BE} grows exponentially with T, because the saturation current I_S \propto n_i^2 \propto e^{-E_g/k_BT} grows roughly as \exp(+0.1T/^\circ\text{C}) per degree. A power transistor dissipating 10 W heats up; the junction temperature rises; I_C rises at fixed V_{BE}; dissipation P = V_{CE} I_C rises further; temperature rises further. If the heat sink cannot extract heat fast enough, the junction temperature runs away to destruction — the semiconductor cracks, the junction short-circuits, the transistor explodes (audibly, on a lab bench). This is thermal runaway. Every audio amplifier design in every Bajaj or Ahuja public-address system includes an emitter resistor R_E (0.1–1 Ω in the output stage) that provides negative feedback — as I_C rises, the voltage across R_E rises, V_{BE} falls, I_C drops back. This is why professional amplifiers survive an Indian summer with its 45 °C ambient temperatures while cheap knock-offs burn out.
Where this leads next
- Transistor as Amplifier and Switch — the common-emitter configuration, voltage gain A_v = -\beta R_C / R_B, the input-output characteristic curves, and the cutoff-saturation switching mode.
- Logic Gates and Digital Electronics — how combinations of transistors in saturation and cutoff realise the Boolean operations OR, AND, NOT, and the universal gates NAND and NOR.
- The p-n Junction Diode — the one-junction device that this article's two-junction device is built from. The Shockley equation, depletion regions, and built-in potentials all carry over directly.
- Semiconductors — Intrinsic and Extrinsic — the doping story that explains why the emitter, base, and collector differ so dramatically in their majority-carrier populations.
- Diode as Rectifier — the one-transistor circuit's simpler cousin, where the single junction converts AC to DC in every Indian power adapter.