The Bipolar Junction Transistor

For the first half of the twentieth century, the machine that made a weak signal strong — the machine that made radio, telephony and the first computers possible — was the vacuum tube: a glass bulb the size of your thumb, with a hot glowing filament that boiled off electrons. Tubes worked, but they were fragile, ran hot, guzzled power, and burned out like light bulbs. An early general-purpose electronic computer, ENIAC, held about eighteen thousand of them and failed somewhere in the machine every few days. Engineers dreamed of a device that could do a tube's job — take a small current and control a large one — out of a cold sliver of solid crystal, with no vacuum and no filament. That device arrived in December 1947, when John Bardeen and Walter Brattain, working under William Shockley at Bell Labs, pressed two gold contacts onto a block of germanium and watched a tiny input current control a much larger output — an amplified signal, from a lump of solid matter. The device — soon christened the transistor (a name coined by Bell Labs engineer John R. Pierce in 1948) — won the three of them the Nobel Prize in 1956. Every microchip since — the billions of switches in the phone in your pocket — is descended from that first crude device. This article builds the workhorse version, the bipolar junction transistor (BJT): two p–n junctions sharing a thin middle layer, wired so that a small base current steers a large collector current. You already own the p–n junction from the previous article; a transistor is just two of them, back to back, with the middle made deliberately thin. From that one geometric trick — a thin, lightly doped base — everything the transistor does follows, using nothing harder than the current-conservation rule $I_E = I_B + I_C$ and school algebra.

Key takeaways

A BJT is two p–n junctions sharing a thin middle layer — three doped regions called the emitter, base and collector, in an n-p-n or p-n-p sandwich. The emitter–base junction is forward biased and the collector–base junction is reverse biased.
The emitter floods the base with carriers; the base is so thin that almost all of them are swept on to the collector. Only a tiny fraction (a percent or two) recombine in the base and leave as base current.
The three currents obey $I_E = I_B + I_C$ — a statement of charge conservation, nothing more. The emitter current splits into a large collector part and a small base part.
Two ratios name the split. $\alpha = I_C/I_E$ is close to $1$ (typically $0.95$ – $0.99$ ); $\beta = I_C/I_B$ is large (typically $20$ – $200$ ). They are linked by $\beta = \alpha/(1-\alpha)$ .
The small base current controls the large collector current — that current-steering is what makes the transistor amplify and switch, taken up in the next article.
Why it works is geometry: a thin, lightly doped base lets carriers cross before they can recombine, so $\alpha \to 1$ and a small base current commands a hundred times more collector current.
Key idea: a transistor is two p–n junctions sharing a base so thin that a small base current controls a large collector current, with the three terminal currents always obeying $I_E = I_B + I_C$ .

The core insight

A bipolar junction transistor is two p–n junctions sharing one deliberately thin, lightly doped base layer, biased so the emitter–base junction is forward and the collector–base junction is reverse; the emitter injects carriers that almost all sweep across the thin base to the collector, so a small base current controls a large collector current, with $I_E = I_B + I_C$ at every instant.

Where this lives today

The bipolar junction transistor you just built is the direct ancestor of essentially all modern electronics. Discrete BJTs still do the heavy lifting wherever you need to drive real current — audio power stages, motor drivers, voltage regulators, radio-frequency amplifiers — because they handle current gracefully and switch fast. But the transistor's world-changing role came from shrinking it: pack thousands, then millions, then billions of transistors onto one chip of silicon and you get the microchip, the microprocessor, and the memory that runs every phone, car and satellite. A modern processor holds tens of billions of transistors, each a descendant of the three-layer sandwich in this article, though most are now the field-effect (MOSFET) cousin optimised for dense digital logic rather than the bipolar type. The next article puts a single transistor to work the two ways that matter most — as an amplifier, where a small base signal is faithfully magnified at the collector, and as a switch, flipping between cut-off (OFF) and saturation (ON) — the switching mode that, chained into logic gates, becomes the language of every computer.

The physical picture

Take the p–n junction you met last article and imagine sandwiching a third region onto it, so you have three doped layers in a row — say n, then a thin p, then n again. That is an n-p-n transistor. Swap the doping everywhere and you get its mirror twin, the p-n-p transistor. Either way the three regions have names that describe their jobs: the emitter (it emits, i.e. injects, the carriers), the base (the thin controlling middle), and the collector (it collects the carriers that make it across). The three regions are deliberately unequal. The emitter is heavily doped — it is the source, so it must be rich in carriers. The base is made very thin and lightly doped — this is the whole secret of the device, and we will see why in a moment. The collector is moderately doped and physically the largest region, because it has to soak up the arriving carriers and get rid of the heat. The two internal boundaries are ordinary p–n junctions: the emitter–base junction and the collector–base junction. A transistor is, quite literally, two diodes sharing a middle terminal.

Figure

An n-p-n transistor with a heavily doped emitter, a thin lightly doped base and a moderate collector, the emitter-base junction forward biased and the collector-base junction reverse biased, so electrons injected by the emitter sweep across the thin base to the collector while only a few recombine as base current

Now put it to work. Bias the two junctions the way the transistor demands: forward-bias the emitter–base junction (the small battery, about $0.7\ \text{V}$ for silicon) and reverse-bias the collector–base junction (a larger battery). The forward-biased emitter junction behaves exactly like a forward diode: its barrier drops, and the heavily doped emitter pours carriers into the base — in an n-p-n, that means electrons flooding into the thin p-type base. Here is where the thin base earns its keep. In an ordinary diode those injected electrons would wander into the p-region and quickly recombine with holes. But this base is only about a micrometre thick — far thinner than the distance an electron typically travels before recombining. So the flood of electrons barely gets started across the base before it runs into the second junction, the collector–base junction, which is reverse biased. A reverse-biased junction is a brick wall to the base's own carriers, but it is a downhill slide for these visiting electrons: its field grabs them and sweeps them straight into the collector. The upshot is that roughly $98\%$ of the electrons the emitter injects shoot clean across the base and out the collector, and only about $2\%$ find a hole to recombine with on the way. Those few recombinations must be topped up by the external base wire, and that trickle is the base current. The lesson in one line: a small base current is the price of admission that lets a hundred-times-larger collector current flow.

The three currents and the rule that binds them

Trace the wires. Current $I_E$ flows in at the emitter terminal; it leaves by two exits — a large part $I_C$ out of the collector, and a small part $I_B$ out of the base. Nothing accumulates inside the crystal in steady operation, so whatever flows in must flow out. That is the transistor's master equation:

I_E = I_B + I_C. \qquad (1)

where: $I_E$ is the emitter current, $I_B$ the base current and $I_C$ the collector current. This is simply Kirchhoff's current law applied to the transistor as a single node — a statement of charge conservation, not a special property of semiconductors. Its physical reading is the whole device in miniature: the emitter's output stream splits into a dominant collector stream and a meagre base stream, and because $I_B$ is tiny, $I_C$ is only a hair less than $I_E$ . To make the split quantitative we name two ratios. The first compares the collector to the emitter — the fraction of the injected flood that survives the trip:

\alpha = \frac{I_C}{I_E}. \qquad (2)

where: $\alpha$ (the common-base current gain) is the fraction of emitter current that reaches the collector. Because almost every carrier makes it across the thin base, $\alpha$ sits just below $1$ — typically between $0.95$ and $0.99$ . An $\alpha$ of $0.98$ says $98\%$ of the emitter's carriers reach the collector and $2\%$ are lost to recombination in the base. Notice $\alpha$ can never quite reach $1$ : some recombination is unavoidable, so there is always a base current. The second ratio is the one that makes the transistor a useful amplifier — it compares the large collector current to the small base current that controls it:

\beta = \frac{I_C}{I_B}. \qquad (3)

where: $\beta$ (the common-emitter current gain, often written $h_{FE}$ ) tells you how many times larger the collector current is than the base current. If $2\%$ of the carriers recombine, then for every $2$ units of base current there are $98$ units of collector current, so $\beta = 98/2 = 49$ . Real transistors have $\beta$ anywhere from about $20$ to $200$ . The punchline: change $I_B$ a little and $I_C$ changes $\beta$ times as much — the small current commands the large one. The two ratios are not independent; they are two views of the same split, related by pure algebra (derived below):

\beta = \frac{\alpha}{1-\alpha}, \qquad \alpha = \frac{\beta}{1+\beta}. \qquad (4)

where: these convert between the common-base gain $\alpha$ and the common-emitter gain $\beta$ . The physical reading: as $\alpha$ creeps toward $1$ (a thinner, cleaner base), the denominator $1-\alpha$ shrinks and $\beta$ shoots up. A base that passes $99\%$ of carriers ( $\alpha = 0.99$ ) gives $\beta = 99$ ; passing $98\%$ gives $\beta = 49$ . A tiny improvement in $\alpha$ near $1$ is a big change in $\beta$ .

Derivation AThe current-splitting rule and the α–β relation

The goal is to show, from charge conservation alone, that the three terminal currents obey $I_E = I_B + I_C$ , and then to derive the exact algebraic bridge $\beta = \alpha/(1-\alpha)$ between the two current gains — using nothing but the definitions of $\alpha$ and $\beta$ and school algebra. No calculus, no diode equation, no band theory enters; the transport physics (how carriers cross the base) is captured entirely in the single measured number $\alpha$ .

Step 1. Treat the transistor as one circuit node in steady state. Charge cannot pile up anywhere inside the crystal indefinitely, so the total current entering equals the total leaving. Current enters at the emitter ( $I_E$ ) and leaves at the base ( $I_B$ ) and collector ( $I_C$ ). Hence $I_E = I_B + I_C$ , which is equation (1). This holds instant by instant, for n-p-n or p-n-p, biased or not — it is Kirchhoff's current law, and it is exact.

Step 2. Bring in the definitions $\alpha = I_C/I_E$ and $\beta = I_C/I_B$ . From the first, $I_C = \alpha\,I_E$ ; from equation (1), $I_B = I_E - I_C = I_E - \alpha I_E = (1-\alpha)\,I_E$ . So the base current is the lost fraction $(1-\alpha)$ of the emitter current, exactly as the recombination picture demands.

Step 3. Form $\beta$ from those two results: $\beta = \dfrac{I_C}{I_B} = \dfrac{\alpha I_E}{(1-\alpha) I_E} = \dfrac{\alpha}{1-\alpha}$ . The emitter current cancels, leaving equation (4). Read it as a warning: as $\alpha \to 1$ , $\beta \to \infty$ , which is why small differences in base quality produce wildly different $\beta$ values between transistors of the same type.

Step 4. Invert it for the return trip. Starting from $\beta(1-\alpha) = \alpha$ , expand to $\beta - \beta\alpha = \alpha$ , gather the $\alpha$ terms as $\beta = \alpha(1+\beta)$ , and divide to get $\alpha = \dfrac{\beta}{1+\beta}$ . As a check, $\alpha = 0.98$ gives $\beta = 0.98/0.02 = 49$ , and running it back, $\beta = 49$ gives $\alpha = 49/50 = 0.98$ — the two formulas are consistent inverses. Verified: $I_B = (1-\alpha)I_E$ , $\beta = \alpha/(1-\alpha)$ and $\alpha = \beta/(1+\beta)$ recover each other for $\alpha = 0.98 \leftrightarrow \beta = 49$ and $\alpha = 0.99 \leftrightarrow \beta = 99$ — recomputed in verify_hs2-transistor-bjt.py.

Figure

Circuit symbols for the n-p-n and p-n-p transistors with the emitter arrow marking conventional current direction, and the current node showing the emitter current splitting into a large collector current and a small base current so that I_E equals I_B plus I_C

Why the base must be thin and lightly doped

Everything the transistor does hangs on one design choice, so it is worth isolating. Suppose the base were thick, as in an ordinary pair of diodes wired together. Then the electrons injected by the emitter would have a long way to travel through p-type material full of holes, and almost all of them would recombine before reaching the collector junction. The collector would see almost nothing; $\alpha$ would collapse toward zero; and the "transistor" would just be two back-to-back diodes with no useful coupling between them — no current-steering, no gain. Make the base thin — thinner than the typical distance a carrier diffuses before recombining — and the electrons cross it in a flash, long before they can find a hole. Make it lightly doped — few holes to recombine with — and you cut the losses further. Between them, these two choices push $\alpha$ up toward $1$ and $\beta$ up into the tens or hundreds. This is why a transistor cannot be built by simply soldering two diodes together: the two junctions must share one continuous, deliberately thin base, so that carriers injected by one junction are still "alive" when they reach the other. That single crystal-growing subtlety is the entire reason the transistor took from the p–n junction's discovery in 1940 to its own invention in 1947 to get right. One more consequence worth stating. Because $I_C = \alpha I_E$ with $\alpha$ nearly constant, the collector current barely depends on the collector voltage once the collector junction is reverse biased enough to sweep up every arriving carrier. The collector current is set by the emitter side — that is, by the base current that controls the emitter injection — not by the collector battery. A transistor is therefore a current-controlled current source: the base current is the tap, the collector current is the flow, and the collector voltage is almost a bystander. That property is what the next article turns into amplification and switching.

Limiting cases

case	result	meaning
$\alpha \to 1$	$\beta \to \infty$	ideal thin, loss-free base — every emitter carrier reaches the collector
$\alpha = 0.98$	$\beta = 49$	typical silicon BJT — $2\%$ recombines in the base
$\alpha \to 0$	$\beta \to 0$	thick base — carriers recombine before crossing; no useful transistor
$I_B = 0$ (base open)	$I_C \approx 0$	cut-off — no base current, so essentially no collector current (transistor OFF)
$I_B$ large	$I_C = \beta I_B$ until saturation	active region — collector current tracks base current, then clamps
emitter junction reverse biased	device does not inject	wrong bias — no emitter injection, transistor idle
collector junction forward biased	saturation	both junctions forward — transistor fully ON, like a closed switch

Common trapA transistor is not two loose diodes soldered together, and the base current does not "become" the collector current

Two traps live here. First, the seductive picture of a transistor as "just two diodes back to back." Electrically the junctions are diodes, yes — but if you actually wire two separate diodes anode-to-anode you get no transistor action at all, because each diode has its own thick base and the carriers injected by one never survive to reach the other. The magic is the shared, ultra-thin base: the two junctions must be a single crystal so that carriers injected across the first are swept across the second before they recombine. Second, students often say the base current "turns into" or "adds up to" the collector current — as if $I_B$ were the source of $I_C$ . It is not. The collector current comes from the emitter's injected flood; the base current is merely the small recombination leftover, and it controls $I_C$ the way a tap controls a torrent without being the water. Keep the causal arrow straight: emitter injects, collector collects, base steers.

Worked examples

Board level — finding the collector current and both gains (CBSE and ICSE pattern)

In an n-p-n transistor the emitter current is $I_E = 2\ \text{mA}$ and the base current is $I_B = 20\ \mu\text{A}$ . Find the collector current $I_C$ , the common-base gain $\alpha$ , and the common-emitter gain $\beta$ .

Step 1. Use the master equation $I_E = I_B + I_C$ to get the collector current: $I_C = I_E - I_B = 2\ \text{mA} - 0.02\ \text{mA} = \boxed{1.98\ \text{mA}}$ . (Convert the base current first: $20\ \mu\text{A} = 0.02\ \text{mA}$ .)

Step 2. The common-base gain is the collector-to-emitter ratio: $\alpha = I_C/I_E = 1.98/2.00 = \boxed{0.99}$ . So $99\%$ of the emitter's carriers reach the collector.

Step 3. The common-emitter gain is the collector-to-base ratio: $\beta = I_C/I_B = 1.98/0.02 = \boxed{99}$ . Cross-check with equation (4): $\beta = \alpha/(1-\alpha) = 0.99/0.01 = 99$ — consistent. A base current of just $20\ \mu\text{A}$ commands a collector current $99$ times larger.

Board level — going the other way from emitter current and α (CBSE and ICSE pattern)

A transistor has $\alpha = 0.98$ and is run with an emitter current $I_E = 5\ \text{mA}$ . Find the collector current $I_C$ , the base current $I_B$ , and the common-emitter gain $\beta$ .

Step 1. From $\alpha = I_C/I_E$ , the collector current is $I_C = \alpha\,I_E = 0.98 \times 5\ \text{mA} = \boxed{4.9\ \text{mA}}$ .

Step 2. The base current is what is left over: $I_B = I_E - I_C = 5 - 4.9 = 0.1\ \text{mA} = \boxed{100\ \mu\text{A}}$ . Equivalently $I_B = (1-\alpha)I_E = 0.02 \times 5 = 0.1\ \text{mA}$ .

Step 3. The current gain is $\beta = I_C/I_B = 4.9/0.1 = \boxed{49}$ , matching $\beta = \alpha/(1-\alpha) = 0.98/0.02 = 49$ . Notice how a modest drop in $\alpha$ from $0.99$ to $0.98$ halves $\beta$ from $99$ to $49$ — the steepness near $\alpha = 1$ in action.

Competitive level — building the currents up from α alone (JEE and NEET pattern)

A transistor with $\alpha = 0.95$ is driven with a base current $I_B = 25\ \mu\text{A}$ . Find $\beta$ , the collector current $I_C$ , and the emitter current $I_E$ . Confirm your $\alpha$ from the resulting currents.

Step 1. Convert the gain: $\beta = \alpha/(1-\alpha) = 0.95/0.05 = \boxed{19}$ .

Step 2. The collector current follows from $\beta = I_C/I_B$ : $I_C = \beta\,I_B = 19 \times 25\ \mu\text{A} = 475\ \mu\text{A} = \boxed{0.475\ \text{mA}}$ .

Step 3. The emitter current closes the node: $I_E = I_B + I_C = 25 + 475 = 500\ \mu\text{A} = \boxed{0.5\ \text{mA}}$ .

Step 4. Confirm: $\alpha = I_C/I_E = 475/500 = 0.95$ — back to where we started, so the numbers are self-consistent. Timing note: about a minute. The fast route is $\beta = \alpha/(1-\alpha)$ first, then $I_C = \beta I_B$ and $I_E = I_B + I_C$ — three one-line steps. Do not detour through $I_E$ before you have $\beta$ ; going $\alpha \to \beta \to I_C \to I_E$ is the shortest path, and the final $\alpha = I_C/I_E$ check costs five seconds and catches any slip.

Competitive level — from β and collector current (JEE Advanced pattern)

A silicon n-p-n transistor has a common-emitter gain $\beta = 100$ and carries a collector current $I_C = 5\ \text{mA}$ (take leakage as negligible). Find the base current $I_B$ , the emitter current $I_E$ , and the common-base gain $\alpha$ .

Step 1. From $\beta = I_C/I_B$ , the base current is $I_B = I_C/\beta = 5\ \text{mA}/100 = 0.05\ \text{mA} = \boxed{50\ \mu\text{A}}$ .

Step 2. The emitter current is the sum: $I_E = I_B + I_C = 0.05 + 5 = \boxed{5.05\ \text{mA}}$ .

Step 3. The common-base gain is $\alpha = I_C/I_E = 5/5.05 = \boxed{0.990}$ , which matches $\alpha = \beta/(1+\beta) = 100/101 = 0.990$ . The lesson: a $\beta$ of $100$ means the base sips only $50\ \mu\text{A}$ to steer a collector current a hundred times larger — the essence of current gain. Timing note: under a minute. Memorise $I_B = I_C/\beta$ and $\alpha = \beta/(1+\beta)$ so you never have to solve equation (4) from scratch under exam pressure. The $\alpha = \beta/(1+\beta)$ form is the one examiners reach for when they give you $\beta$ and want $\alpha$ .

Exam traps

Reversing the two biases. The examiner's favourite: the emitter–base junction is forward biased and the collector–base junction is reverse biased — always, in normal (active) operation. Swap them and the device does nothing useful. A one-line memory hook: the emitter emits (so its junction must be forward, to inject) and the collector collects (so its junction must be reverse, to sweep carriers in). Writing "both junctions forward biased" describes the saturation switch-ON state, not normal amplifying operation.

Confusing $\alpha$ and $\beta$ , or which is bigger. $\alpha = I_C/I_E$ is just under $1$ ; $\beta = I_C/I_B$ is large (tens to hundreds). If your $\alpha$ comes out above $1$ or your $\beta$ below $1$ , you have swapped the currents. Anchor: $\alpha$ compares collector to the bigger emitter (so $<1$ ); $\beta$ compares collector to the tiny base (so $\gg 1$ ).

Dropping the microamp/milliamp conversion. Base current is usually in microamps ( $\mu\text{A}$ ) while collector and emitter currents are in milliamps ( $\text{mA}$ ). Forgetting that $20\ \mu\text{A} = 0.02\ \text{mA}$ throws every ratio out by a factor of a thousand. Convert everything to one unit before you divide.

Thinking the base current "becomes" the collector current. The collector current is supplied by the emitter's injected carriers, not by the base wire. The base current is the small recombination leftover that controls $I_C$ . Say "the base current controls the collector current," never "the base current turns into the collector current."

Treating a transistor as two independent diodes. The junctions are diodes, but the device works only because they share one ultra-thin base. Two separately packaged diodes wired back to back give no transistor action.

Board scope. The BJT — its structure, the two biasing rules, $I_E = I_B + I_C$ , and the $\alpha$ / $\beta$ relations — is core on CBSE, ICSE (ISC), JEE Main, JEE Advanced, and NEET. A-Level treatment varies by board: CIE 9702 scopes its electronics to the diode/rectifier, the operational amplifier and logic gates, and does not carry the BJT or its $\alpha$ / $\beta$ gains — so do not assume this material for a 9702 student. It is not on AP Physics 2, which carries no semiconductor electronics at all, so do not quote AP-2 conventions here. IGCSE 0625 stops at qualitative diodes, LEDs and logic gates and does not carry the transistor or its current gains — an IGCSE student needs none of the $\alpha$ / $\beta$ algebra. Current IB Physics has moved electronics out of the core, so treat the BJT as India- and A-Level-facing; do not tell an IB student it is examinable for them.

Origins · 1 minJohn Bardeen and how this was discovered

John Bardeen

1908–1991 · Madison, Wisconsin and Murray Hill, New Jersey and Urbana, Illinois

John Bardeen is the only person ever to win the Nobel Prize in Physics twice. Trained as an electrical engineer and then a physicist, he joined Bell Labs after the war and, with Walter Brattain, tackled why a solid crystal stubbornly refused to amplify. Bardeen's crucial insight was the role of surface states — trapped charges on a semiconductor's surface that screened out any applied field — and getting around them led the pair, in December 1947, to press two closely spaced gold points onto germanium and demonstrate the first transistor, a point-contact amplifier. He shared the 1956 Nobel Prize with Brattain and Shockley for it. Bardeen then moved to the University of Illinois and won a second Nobel in 1972, with Cooper and Schrieffer, for the BCS theory of superconductivity — a rare double that ties the transistor's inventor to two of modern physics' deepest results.

History

1874

— Karl Ferdinand Braun discovers that metal–crystal contacts rectify current, the first solid-state semiconductor device.

1940

— Russell Ohl at Bell Labs identifies the silicon p–n junction, the building block every transistor is made from.

1947

— John Bardeen and Walter Brattain demonstrate the first working transistor, a point-contact device, at Bell Labs in December.

1948

— William Shockley conceives the bipolar junction transistor (first working junction device 1951), the layered n-p-n / p-n-p design used ever since.

1954

— Gordon Teal at Texas Instruments makes the first commercial silicon transistor, opening the way to mass production.

1956

— Bardeen, Brattain and Shockley share the Nobel Prize in Physics for the invention of the transistor.

Going further

Texts. NCERT Physics Class 12, Chapter 14 ("Semiconductor Electronics"), the junction-transistor sections · H. C. Verma, Concepts of Physics Vol. 2, "Semiconductors and Semiconductor Devices" · Sedra and Smith, Microelectronic Circuits, for the full BJT model beyond this article · Boylestad and Nashelsky, Electronic Devices and Circuit Theory.

Provenance. LANE_SPEC anchors. Recompute in verify_hs2-transistor-bjt.py: master relation $I_E = I_B + I_C$ ; $\alpha = I_C/I_E$ (range $0.95$ – $0.99$ ); $\beta = I_C/I_B$ (range $20$ – $200$ ); $\beta = \alpha/(1-\alpha)$ and $\alpha = \beta/(1+\beta)$ with $\alpha = 0.98 \leftrightarrow \beta = 49$ , $\alpha = 0.99 \leftrightarrow \beta = 99$ , $\beta = 100 \leftrightarrow \alpha = 0.990$ ; the $98\%/2\%$ carrier split giving $\beta = 98/2 = 49$ ; worked example 1 ( $I_E = 2\ \text{mA}$ , $I_B = 20\ \mu\text{A} \Rightarrow I_C = 1.98\ \text{mA}$ , $\alpha = 0.99$ , $\beta = 99$ ); example 2 ( $\alpha = 0.98$ , $I_E = 5\ \text{mA} \Rightarrow I_C = 4.9\ \text{mA}$ , $I_B = 100\ \mu\text{A}$ , $\beta = 49$ ); example 3 ( $\alpha = 0.95 \Rightarrow \beta = 19$ , $I_B = 25\ \mu\text{A} \Rightarrow I_C = 475\ \mu\text{A}$ , $I_E = 500\ \mu\text{A}$ ); example 4 ( $\beta = 100$ , $I_C = 5\ \text{mA} \Rightarrow I_B = 50\ \mu\text{A}$ , $I_E = 5.05\ \text{mA}$ , $\alpha = 0.990$ ); Si forward knee $0.7\ \text{V}$ ; base thickness about $1\ \mu\text{m}$ .

Prerequisites. The p–n Junction — the single junction, its depletion region and barrier potential, and the forward/reverse bias behaviour that a transistor uses twice over. Without the reverse-biased junction's carrier-sweeping action, the collector could not collect. Leads to. Transistor as Amplifier and Switch, which takes the current gain $\beta$ built here and turns it into voltage gain and digital switching. For the full quantitative device physics — the Ebers–Moll model, the diffusion of minority carriers across the base, and the temperature dependence of $\beta$ — the university seam on semiconductor devices treats the transport properly (go-deeper only, never an exam dependency here).

Why is it called a bipolar transistor?

Because its operation relies on both kinds of charge carrier — electrons and holes — moving through the device. In an n-p-n transistor, electrons injected from the emitter are the carriers that cross the p-type base, but the base is a sea of holes, and the small base current is a flow of holes replacing those lost to recombination. Both polarities of carrier take part, hence "bipolar." This distinguishes the BJT from the field-effect transistor (FET), which is unipolar: only one carrier type carries the current, and the device is controlled by a voltage on an insulated gate rather than by a base current. Both are transistors; they just steer current by different means.

What is the difference between n-p-n and p-n-p, and which should I use?

They are mirror images. An n-p-n has n-type emitter and collector around a p-type base; a p-n-p has p-type emitter and collector around an n-type base. Everything reverses: in a p-n-p you flip both battery polarities, and the majority carriers doing the crossing are holes, not electrons. The current directions in the symbol flip too — the emitter arrow points into the p-n-p and out of the n-p-n (a handy memory aid: the arrow points the way conventional current flows in the emitter lead). In practice n-p-n types are more common because electrons move faster than holes, giving slightly better high-frequency performance, but the two are used together (as "complementary" pairs) in many amplifier and switching circuits.

Why does a small base current control such a large collector current?

Because the base is thin, almost all the carriers the emitter injects sail straight across it to the collector; only a small fraction — a percent or two — get caught by recombination on the way. Those caught carriers are what the base wire has to replenish, so the base current is small. But that recombination fraction is exactly what limits how much the emitter can inject: reduce the base current and you starve the recombination, which throttles the emitter injection, which cuts the collector current. So the small base current acts as a valve on the large collector current. The ratio between them is $\beta$ , and because only $2\%$ or so recombines, $\beta$ naturally comes out around $50$ – $100$ .

Is a transistor really just two diodes back to back?

Electrically, the two junctions are diodes — you can even measure them with a multimeter's diode test. But you cannot build a working transistor by soldering two separate diodes together, and this is the key subtlety. Each stand-alone diode has a thick base region, so carriers injected across one junction recombine long before they could reach the other. Transistor action needs the two junctions to share one continuous, deliberately thin base, grown as a single crystal, so that carriers injected by the emitter are still mobile when they arrive at the collector junction. The "two diodes" picture is fine for predicting which way each junction conducts, but it completely misses the current-steering that makes the device useful.

What happens if I reverse the collector and emitter by mistake?

The transistor still works, but badly. Because the emitter is heavily doped and the collector only moderately doped, the device is not symmetric — swapping them (running it "inverted") gives a much lower current gain, often a $\beta$ of only a few instead of tens or hundreds. This is a direct consequence of the design: the heavy emitter doping is what makes injection efficient, so using the lightly doped collector as an emitter cripples the injection. The lesson is that emitter and collector are not interchangeable, even though both are n-type in an n-p-n. Always wire the terminals as the datasheet specifies.

Why did the transistor take so long to invent after the p-n junction?

The p–n junction was identified in 1940, but the transistor did not follow until 1947 — seven years — and the reason is the thin base. Getting two junctions to share a base thin and clean enough for carriers to cross without recombining was a serious materials-science problem. Bardeen and Brattain first had to understand why applied fields seemed to do nothing at a semiconductor surface — the answer was surface states, trapped charges that screened the field — before they could coax amplification out of germanium at all. Their first device was a fiddly point-contact affair; Shockley's more robust layered junction design was conceived a year later (with a working junction transistor by 1951), and reliable silicon versions took until 1954. The physics was clear; the crystal-growing craft to realise it was not.

Does the collector voltage set the collector current?

Almost not at all, and this surprises people. Once the collector–base junction is reverse biased enough to sweep up every carrier arriving from the base, raising the collector voltage further does very little — the collector current is set by how many carriers the emitter injects, which is controlled by the base current, not by the collector battery. This is why a transistor behaves as a current-controlled current source: the base current is the control knob and the collector current follows it, staying almost flat as the collector voltage changes (the horizontal part of the output characteristic). That near-independence is exactly the property the next article exploits to make a clean amplifier.