Billions of these are inside the chip you're reading this on.
Start here: a sluice gate you never touch the water with
Picture a canal with a sluice gate. Raise the gate and water flows; lower it and the flow stops. Crucially, the lever that raises the gate never touches the water — it is completely separated from what it controls.
A MOSFET works exactly like that, and the separation is the whole invention. Its control terminal — the gate — sits behind a layer of insulating glass. It never touches the current it controls, so it draws essentially no current at all. You control it with voltage alone.
Why that one difference matters so much
⚖️ BJT against MOSFET
BJTControlled by current. The base draws real current continuously, which the previous stage must supply.
MOSFETControlled by voltage. The gate draws essentially nothing, so the previous stage supplies almost no power.
Now scale that up. A chip holds billions of transistors. If each one continuously drew even a tiny control current, the total would be impossible. Because a MOSFET gate draws nothing while it sits still, you can put billions of them on one piece of silicon.
💡 That is why the device you are reading this on is built from MOSFETs rather than BJTs. It is not that MOSFETs are better in every way — BJTs are still preferred for some analogue and high-power work. It is specifically that zero control current is what makes enormous numbers of them practical.
How the gate actually controls anything through insulation
If the gate is insulated, how does it do anything at all? Through electric field, not current.
Put a positive voltage on the gate and it attracts electrons in the silicon underneath the insulator — pulling them up toward the surface to form a thin conducting layer, called the channel. No electron ever crosses the insulator. They are simply attracted into position from the sides.
Raise the gate voltage and the channel gets richer, so more current flows. Lower it and the channel thins and vanishes. The gate is a lever operating at a distance — exactly the sluice gate, where the lever is mechanically separate from the water it moves.
💡 The insulating layer being so thin is also a weakness. Static electricity from your finger can punch straight through it and destroy the device permanently, which is why MOSFETs ship in conductive foam and why factories use earthing straps. The thing that makes them brilliant also makes them fragile.
The BJT is controlled by base current. A MOSFET is controlled by gate voltage — and crucially, the gate is separated from everything else by an insulating oxide layer, so it draws essentially zero current. That single structural difference is why every modern digital chip uses MOSFETs rather than BJTs. No gate current means almost no static power. That is what makes it possible to put billions of them on one die.
Structure and the four terminals
Metal-Oxide-Semiconductor Field Effect Transistor
Gate (G) — metal/polysilicon plate, insulated by SiO₂
Source (S) — where carriers enter
Drain (D) — where carriers leave
Body/Substrate (B) — usually tied to the source
The gate is NOT electrically connected to the channel.
It works purely by the ELECTRIC FIELD it creates through
the oxide — hence "field effect".
Enhancement vs Depletion
🔀 Two types, opposite defaults
Enhancement (E-MOSFET)No channel exists at V_GS = 0 — normally OFF. Applying V_GS > V_th induces (enhances) a channel. This is the type used in all digital logic.
Depletion (D-MOSFET)A channel is physically built in — normally ON. Applying reverse V_GS depletes it to turn the device off. Rarer, used in some RF and analogue circuits.
n-channel vs p-channeln-channel conducts with positive V_GS and uses electrons (faster). p-channel needs negative V_GS and uses holes (slower, ~2–3×). CMOS pairs them.
The operating equations
CUT-OFF: V_GS < V_th → I_D = 0
OHMIC / TRIODE (V_DS < V_GS − V_th) — acts like a resistor:
I_D = k[2(V_GS − V_th)V_DS − V_DS²]
SATURATION (V_DS ≥ V_GS − V_th) — used for amplification:
I_D = k(V_GS − V_th)² ← SQUARE law
where k = ½μ_n C_ox (W/L) depends on geometry
Transconductance:
g_m = dI_D/dV_GS = 2k(V_GS − V_th) = 2√(k·I_D)
A BJT is exponential (I_C ∝ e^V_BE) while a MOSFET is square-law (I_D ∝ (V_GS−V_th)²). That's why a BJT gives more gain per volt of input, but a MOSFET is more linear and far easier to make in huge numbers. Note also the naming trap: for a MOSFET, "saturation" is the useful amplifying region — the opposite of a BJT, where saturation means fully on and useless for amplification.
Worked numerical 1 — saturation current
An n-channel E-MOSFET has V_th = 2 V and k = 0.5 mA/V². Find I_D at V_GS = 5 V (in saturation), and the transconductance there.
I_D = k(V_GS − V_th)²
= 0.5 mA/V² × (5 − 2)²
= 0.5 × 9 = 4.5 mA
Minimum V_DS to stay in saturation:
V_DS ≥ V_GS − V_th = 5 − 2 = 3 V
Transconductance:
g_m = 2k(V_GS − V_th) = 2 × 0.5 mA/V² × 3 V
= 3 mA/V = 3 mS
Check via the other form:
g_m = 2√(k·I_D) = 2√(0.5 × 4.5) = 2√2.25 = 3 mS ✔
Now double the overdrive to V_GS = 8 V:
I_D = 0.5 × (8−2)² = 0.5 × 36 = 18 mA (4× — square law!)
g_m = 2 × 0.5 × 6 = 6 mS (only 2×)
Worked numerical 2 — MOSFET amplifier gain
The MOSFET above (g_m = 3 mS) is used in a common-source amplifier with R_D = 4.7 kΩ. Find the voltage gain and compare with a BJT stage at the same current.
MOSFET common-source gain:
A_v = −g_m × R_D = −3 mS × 4.7 kΩ = −14.1
BJT at the same 4.5 mA collector current:
g_m = I_C/V_T = 4.5/26 = 173 mS
A_v = −g_m R_C = −173 mS × 4.7 kΩ = −813
The BJT gives 58× more gain at the same bias current.
That is the fundamental trade-off: BJTs win on gain
(exponential characteristic), MOSFETs win on input
impedance (~10¹² Ω), power consumption and density.
Worked numerical 3 — MOSFET as a switch
A power MOSFET with R_DS(on) = 25 mΩ switches a 10 A load. Find the conduction loss, and compare with a BJT having V_CE(sat) = 0.8 V.
MOSFET (resistive drop):
V_drop = I × R_DS(on) = 10 × 0.025 = 0.25 V
P_loss = I²R = 10² × 0.025 = 2.5 W
BJT (constant saturation voltage):
V_drop = 0.8 V
P_loss = V × I = 0.8 × 10 = 8 W
MOSFET wastes 3.2× less power — and the advantage grows at
lower currents, because I²R falls faster than V×I.
At 1 A: MOSFET 0.025 W vs BJT 0.8 W (32× better)
This is why every switching power supply, motor driver and
laptop charger uses MOSFETs, not BJTs.
💡 Exam angle: usually 4–6 marks on working principle plus a diagram. Full marks need: the insulated gate (hence near-zero gate current), the induced channel above V_th, the three regions, and the square-law equation. Be careful with the terminology trap — for a MOSFET, saturation is the amplifying region; for a BJT it is the fully-on region. Examiners set that trap deliberately.
Syllabus points
Working principle and applications
Create a free account to tick topics off, take notes as you read, watch the video lessons and get a day-by-day study plan built around your exam date.