Scienceintermediate

Transistors: BJTs and MOSFETs

The two transistor families that built modern electronics. BJTs as current-controlled amplifiers, MOSFETs as voltage-controlled switches with a capacitor for a gate, the three operating regions of each, and the four reasons MOSFETs ended up running every digital chip.

semiconductors transistors bjt mosfet circuits cmos amplifiers

Part of:⚡Semiconductors: from band theory to modern chips

Not signed in — your progress and quiz score won't be saved.

Lesson progress1 / 8

What a transistor does

A transistor is a three-terminal semiconductor device whose behavior at one pair of terminals is controlled by a signal at the third terminal. That gives you two essential functions:

Amplification: a small input controls a large output — the same signal shape, bigger.
Switching: a digital input either fully cuts off or fully turns on the output. The basis of every gate, register, and CPU on Earth.

Two dominant families exist. The Bipolar Junction Transistor (BJT), invented at Bell Labs in 1947, is current-controlled: a small base current modulates a large collector current. The Metal-Oxide-Semiconductor Field-Effect Transistor (MOSFET), mainstream by the late 1960s, is voltage-controlled: a gate voltage modulates a channel that carries current between source and drain.

BJTs ruled analog and early digital. MOSFETs run essentially every digital chip today. The rest of this lesson is why.

Full lesson text

All 8 steps on one page — for reading, reference, and search.

Show

1. What a transistor does

A transistor is a three-terminal semiconductor device whose behavior at one pair of terminals is controlled by a signal at the third terminal. That gives you two essential functions:

Amplification: a small input controls a large output — the same signal shape, bigger.
Switching: a digital input either fully cuts off or fully turns on the output. The basis of every gate, register, and CPU on Earth.

BJTs ruled analog and early digital. MOSFETs run essentially every digital chip today. The rest of this lesson is why.

2. BJT structure

A BJT is three doped layers in a row, forming two PN junctions. Two flavors:

NPN: n-emitter, p-base, n-collector. Conducts when the base is more positive than the emitter (by ~0.7 V).
PNP: p-emitter, n-base, p-collector. Mirror image.

The key trick is that the base is thin and lightly doped. When the base-emitter junction is forward-biased, electrons flood from emitter into base. Most of them don't recombine (the base is too thin) — they get swept into the collector by the reverse-biased base-collector junction.

The ratio of collector current to base current is the current gain $\beta$ (often 50–500). Pump 1 mA into the base of an NPN with $\beta = 100$ , and 100 mA flows through the collector. The base current controls the collector current. That's a BJT.

3. BJT operating regions

A BJT lives in one of three operating regions, defined by which junctions are forward-biased:

Region	BE junction	BC junction	Behavior
Cutoff	Reverse	Reverse	No current. "Off."
Active	Forward	Reverse	$I_C \approx \beta I_B$ . Amplifier.
Saturation	Forward	Forward	$I_C$ capped by external circuit. "On."

In active region the BJT amplifies — the IV characteristic is roughly a flat horizontal line whose height is set by $I_B$ . In saturation, $V_{CE}$ collapses to ~0.2 V and the transistor acts like a closed switch.

Digital logic uses cutoff and saturation. Analog circuits camp in active. Reverse-active (BE reverse, BC forward) is the fourth quadrant — exists, but rarely useful.

4. BJT and MOSFET, side by side

Three-terminal devices, different control mechanisms: BJT current-in-base, MOSFET voltage-on-gate.

flowchart LR
  A["BJT NPN: emitter, base, collector"] --> B["Base current controls collector current"]
  C["MOSFET NMOS: source, gate, drain"] --> D["Gate voltage modulates channel"]
  B --> E["Used in: amplifiers, analog, power"]
  D --> F["Used in: digital, CMOS logic"]

5. MOSFET structure

A MOSFET is a layered sandwich on a silicon substrate:

Source and drain: two heavily doped regions (n+ for NMOS, p+ for PMOS) implanted in a lightly doped substrate of the opposite type.
Gate: a conductive layer (originally aluminum, today polysilicon or metal) sitting on top.
Gate oxide: a thin insulating layer of SiO₂ (or modern high-K dielectric) between gate and substrate.

The magic is in the oxide. The gate doesn't touch the silicon — it forms a capacitor with the substrate. Apply a voltage to the gate and you electrostatically induce a channel of carriers in the substrate just below the oxide. The carriers flow from source to drain. No gate current is drawn in steady state — only a tiny leakage through the (extremely thin) oxide.

That's why MOSFETs are voltage-controlled: the gate sees only displacement current. Pure capacitor action.

6. MOSFET operating regions

A MOSFET also has three regions, controlled by gate-source voltage $V_{GS}$ and drain-source voltage $V_{DS}$ . For an NMOS with threshold voltage $V_T$ (typically 0.3–0.7 V):

Region	Condition	Behavior
Cutoff	$V_{GS} < V_T$	No channel. No current. "Off."
Triode	$V_{GS} > V_T$ and $V_{DS} < V_{GS} - V_T$	Acts like a resistor set by $V_{GS}$ .
Saturation	$V_{GS} > V_T$ and $V_{DS} > V_{GS} - V_T$	Current roughly constant; set by $V_{GS}$ . Amplifier.

In saturation: $I_D \propto (V_{GS} - V_T)^2$ — quadratic in the gate overdrive. Compare to the BJT's exponential in $V_{BE}$ .

Digital logic uses cutoff (off) and triode (on, low resistance). Analog camps in saturation. PMOS works identically with all polarities flipped.

7. BJT vs MOSFET tradeoffs

Side by side, the differences come down to four:

Control: BJT needs a steady base current (always wastes some power). MOSFET draws gate current only during switching.
Speed: MOSFET switches faster at the same process node — no minority-carrier storage to flush out.
On-resistance: BJT has a ~0.2 V $V_{CE,sat}$ floor in saturation. MOSFET on-resistance can drop below 1 mΩ; power loss is $I^2 R$ instead of $I \cdot V$ .
Density: MOSFET is one device on one piece of silicon with a thin oxide on top. BJT needs three vertical doped layers with careful isolation. MOSFETs pack denser by roughly 10×.

BJTs still win in a few niches: low-noise RF amplifiers (the exponential IV is great for analog gain), high-current linear regulators, and discrete power circuits where the BJT's gentler turn-on smooths transients.

8. Why MOSFETs won digital

In the 1960s, BJTs were the obvious choice. They had higher transconductance, faster transistors, and lower noise. MOSFETs were slow and noisy.

Three things flipped the race over the next two decades:

Static power matters more than speed at scale. A million BJTs each leaking a small base current adds up. A million MOSFETs in steady state draw essentially zero gate current.
CMOS exists. Combining NMOS and PMOS makes an inverter that draws power only when switching. There's no BJT equivalent — bipolar logic always pulls steady current.
MOSFETs scale. The gate oxide can be made thinner (more transconductance) and the channel shorter (more speed) with the same process knobs. BJTs hit physical limits faster.

By the 1980s every microprocessor and DRAM was CMOS. The next lesson covers exactly how that inverter works.

Check your understanding

The lesson ends with a 5-question quiz. Take it in the player above to see your score.

An NPN BJT has $\beta = 200$. You inject 50 μA into the base. Approximately what flows through the collector?
- 100 μA.
- 10 mA.
- 200 mA.
- 50 μA.
An NMOS has $V_T = 0.5$ V. You apply $V_{GS} = 1.5$ V and $V_{DS} = 0.3$ V. Which region is it in?
- Cutoff.
- Triode (linear).
- Saturation.
- Reverse-active.
Why do MOSFETs dominate digital logic but BJTs survive in low-noise RF amplifiers?
- BJTs are cheaper.
- BJTs switch faster than MOSFETs at the same process node.
- BJTs' exponential IV gives higher transconductance per unit current — better for low-noise analog gain — while MOSFETs' near-zero gate current and dense scaling win for digital.
- BJTs handle higher voltages.
What physical structure prevents a MOSFET gate from drawing current in steady state?
- A reverse-biased PN junction between gate and channel.
- The thin SiO₂ insulating layer between the gate and the channel.
- The doped polysilicon used as the gate material.
- The drain-substrate junction.
Why is a million-transistor digital chip built with MOSFETs in CMOS rather than BJTs?
- MOSFETs are individually cheaper to manufacture than BJTs.
- MOSFETs draw essentially no steady-state gate current and combine into CMOS gates that consume power only when switching.
- BJTs can't be integrated on silicon at all.
- BJTs have a smaller forward voltage drop than MOSFETs.

Related lessons

Science

advanced

Entanglement, gates, and the circuit model

What entanglement is mathematically, how Bell states are built from Hadamard and CNOT, how quantum circuits compose, and why measurement on entangled subsystems looks correlated regardless of separation.

8 steps·~12 min

History

advanced

Concentration and risk: single points of failure

The geographic and corporate concentration of the chip supply chain expressed as engineering risk — single points of failure, the cost of redundancy, hedging strategies, and the customer-side concentration that mirrors the supply side. Structural analysis, not prediction.

8 steps·~12 min

History

advanced

Catchup: what closing a node gap actually requires

The structural ingredients required to move a national or corporate semiconductor capability up the node ladder — capital, equipment access, process know-how, and talent — and the historical examples (Korea, Taiwan) of how earlier catchup campaigns played out.

8 steps·~12 min

History

advanced

Export controls as policy: instruments and mechanics

How the US, Dutch, and Japanese export-control regimes for advanced semiconductors actually function — the legal instruments, the parameter thresholds in published regulations, the extraterritorial Foreign Direct Product Rule, and the licensing-and-enforcement mechanics.

8 steps·~12 min