Thyristor (SCR)

1. Theory

• Definition: A thyristor or silicon controlled rectifier (SCR) is a four-layer PNPN switch with anode, cathode and gate terminals that latches on when triggered and stays on until its current is removed.
• Three junctions: $J_1$ (anode P-emitter / $N_1$ base), $J_2$ (central $N_1$/$P_2$ junction that blocks forward voltage) and $J_3$ ($P_2$ base / $N_2$ cathode emitter).
• Two-transistor model: The device acts as a PNP and an NPN transistor whose collectors feed each other’s bases; no new carrier physics beyond the BJT is involved.
• Regenerative feedback: The anode current is leakage amplified by $1/\left[1-({\alpha }_1+{\alpha }_2)\right]$; when ${\alpha }_1+{\alpha }_2$ reaches 1, the current grows until both transistors saturate (latching).
• Current-dependent gains: $\alpha$ rises with emitter current, which is why the device needs a minimum current (latching and holding current) to stay on.
• Wide $N_1$ base: The thick, lightly doped $N_1$ base sets the blocking voltage, and in the on-state it is flooded with plasma like a PIN diode.

2. Working Principle

• Reverse blocking: With the cathode positive, $J_1$ and $J_3$ are reverse biased and only leakage flows.
• Forward blocking: With the anode positive and no gate current, $J_2$ is reverse biased and blocks the voltage.
• Gate triggering: A gate current into $P_2$ raises ${\alpha }_2$, pushing ${\alpha }_1+{\alpha }_2$ to 1 so the device latches on.
• On-state: All junctions are effectively forward biased, and conductivity modulation of the $N_1$ base gives a low voltage drop (about 1–2 V) even in kV-rated devices.
• Turn-off (commutation): The gate cannot turn the device off; the anode current must fall below the holding current, and reverse voltage must be held for the turn-off time $t_q$.
• Unwanted turn-on: Breakover (avalanche of $J_2$) or a fast $dV/dt$ displacement current can fire the device without a gate signal; emitter shorts and snubbers prevent this.

3. Applications

• HVDC transmission: Line-commutated converter valves for the highest-power DC links.
• Phase-controlled rectifiers: MW-scale controlled rectifiers for electrolysis, DC motor drives and industrial heating.
• Static VAR compensators: Thyristor-controlled reactors and thyristor-switched capacitors for grid reactive-power control.
• Soft starters: AC voltage control for starting large induction motors.
• Crowbar protection: Rapidly short-circuits a supply rail to protect loads from overvoltage.

Name	Formula	Related Formulae	Variables/Constants
Two-transistor model currents	$I_{C1}={\alpha }_1{I}_A+I_{CBO1}$ $I_{C2}={\alpha }_2{I}_K+I_{CBO2}$	$I_K=I_A+I_G,I_A=I_{C1}+I_{C2}$	$I_A,I_K,I_G:anode,cathode,andgatecurrents$ $\begin{array}{r}{\alpha }_1,{\alpha }_2:common-base\\ currentgainsofthePNPandNPNhalves\\ (seeBJTpage)\end{array}$ $I_{CBO1},I_{CBO2}:collectorleakagecurrents$
Anode current (regenerative gain)	$I_A=\frac{{\alpha }_2{I}_G+I_{CBO1}+I_{CBO2}}{1-({\alpha }_1+{\alpha }_2)}$		$\begin{array}{r}\text{(forward}\text{blocking:}\text{with}{I}_G=0\text{the}\\ \text{current}\text{is}\text{only}\text{leakage}\text{amplified}\text{by}\\ 1/[1-({\alpha }_1+{\alpha }_2)]\text{)}\end{array}$
Latching (turn-on) condition	${\alpha }_1+{\alpha }_2=1$	${\alpha }_i={\alpha }_i(I_E)\text{increases with emitter current}$	$\begin{array}{r}\text{(at}\text{this}\text{point}\text{the}\text{denominator}\text{of}{I}_A\\ \text{vanishes}\text{and}\text{the}\text{two}\text{transistors}\text{drive}\\ \text{each}\text{other}\text{into}\text{saturation)}\end{array}$
Breakover condition (avalanche-assisted)	$M({\alpha }_1+{\alpha }_2)=1$	$M=\frac{1}{1-(V/V_B{)}^n}\text{(see Avalanche Diode page)}$	$\begin{array}{r}M:avalanchemultiplicationfactorofthe\\ reverse-biasedjunction{J}_2\end{array}$ $\begin{array}{r}{V}_{BO}:forwardbreakovervoltage,the\\ anodevoltageatwhichthisconditionis\\ metwith{I}_G=0\end{array}$
Forward blocking voltage (design)	$V_{BO}\lesssim B{V}_{J2}\approx \frac{{\epsilon }_s{E}_{crit}^2}{2q{N}_D}$	$\text{(see PN Junction Diode / Avalanche Diode pages)}$	$\begin{array}{r}B{V}_{J2}:avalanchebreakdownvoltageof\\ thereverse-biased{J}_{2}junction\end{array}$ $\begin{array}{r}{N}_D:dopingofthelightlydoped{n}_1\\ base,whichsetstheblockingcapability\end{array}$
Holding current	${\alpha }_1(I_H)+{\alpha }_2(I_H)=1$	$I_L>I_H$	$\begin{array}{r}{I}_H:minimumanodecurrentbelowwhich\\ thegainsdropsothat\\ {\alpha }_1+{\alpha }_2<1andthedeviceturns\\ off\end{array}$ $\begin{array}{r}{I}_L:latchingcurrent,thelargercurrent\\ neededatthemomentofturn-on\\ forthedevicetostayonafterthe\\ gatepulseends\end{array}$
Critical rate of voltage rise (dv/dt)	$i_{disp}=C_{J2}\frac{dV}{dt}$	${\left(\frac{dV}{dt}\right)}_{crit}\approx \frac{I_{G,trig}}{C_{J2}}$	$\begin{array}{r}{i}_{disp}:displacementcurrentthroughthe\\ J_{2}depletioncapacitance\end{array}$ $\begin{array}{r}{C}_{J2}:J_{2}depletioncapacitance(seePN\\ JunctionDiodepage)\end{array}$ $\begin{array}{r}{I}_{G,trig}:effectivegatecurrent\\ requiredtotrigger\end{array}$
Turn-on time	$t_{on}=t_d+t_r$		$\begin{array}{r}{t}_d:delaytimeforinjectedcarriersto\\ crossthebasesandraise\\ {\alpha }_1+{\alpha }_{2}tounity\end{array}$ $\begin{array}{r}{t}_r:risetimeastheanodecurrent\\ regenerativelygrowsandtheconducting\\ areaspreads\end{array}$
Turn-off (circuit-commutated) time	$t_q\approx \tau \ln \left(\frac{I_T}{I_H}\right)$	$Q_s=\tau I_T\text{(see PN Junction Diode / PIN Diode pages)}$	$\begin{array}{r}{t}_q:minimumtimetheanodemustbe\\ heldreverseorbelow{I}_{H}beforeforward\\ voltagecanbereappliedwithout\\ retriggering(approximateestimate)\end{array}$ $I_T:on-statecurrentbeforeturn-off$ $\tau :carrierlifetime$

Important Questions

Ques: What is actually different about a thyristor compared with a single BJT?

Nothing in the junction physics: each half is an ordinary bipolar transistor described by the BJT page. What is new is the connection. In a BJT the base current is supplied externally, so output current is proportional to input current ($I_C=\beta I_B$) and there is no latching. In a thyristor the collector current of each transistor is the base current of the other, which closes a positive-feedback loop with loop gain ${\alpha }_1+{\alpha }_2$. As long as this sum is below 1, the anode current is only leakage amplified by $1/[1-({\alpha }_1+{\alpha }_2)]$; once it reaches 1, the denominator vanishes and the current grows regeneratively until both transistors saturate and the whole structure is flooded with carriers. At that point the gate has no further control, which is why an SCR stays on after the gate pulse is removed. Reference: S.M. Sze and Kwok K. Ng, Physics of Semiconductor Devices, 3rd ed., Wiley, 2007.

Ques: Why does a holding current exist, and why do ${\alpha }_1$ and ${\alpha }_2$ depend on current?

At low emitter currents, a large fraction of injected carriers recombine inside the emitter-base depletion region (the same recombination-dominated, ideality-factor-near-2 regime described on the PN Junction Diode page), so the emitter injection efficiency, and hence $\alpha$, is small. As current rises, diffusion current dominates over depletion-region recombination and $\alpha$ climbs toward its high-level value. A thyristor in the on-state therefore depends on its anode current to keep ${\alpha }_1+{\alpha }_2\ge 1$; if the external circuit lets $I_A$ fall below $I_H$, the gains drop below the latching threshold and the loop collapses. This is the physical reason an SCR stays latched on until the anode current is reduced below $I_H$ (commutation). The failure of gate turn-off is a separate matter, caused by the low turn-off gain and the lateral base resistance described on the GTO page. Reference: S.M. Sze and Kwok K. Ng, Physics of Semiconductor Devices, 3rd ed., Wiley, 2007.

Ques: Why can a thyristor turn on without any gate current, and how is this prevented?

Two mechanisms push ${\alpha }_1+{\alpha }_2$ toward 1 with $I_G=0$. First, if the forward voltage across $J_2$ approaches its avalanche breakdown, multiplication $M$ amplifies the leakage current and satisfies $M({\alpha }_1+{\alpha }_2)=1$ (breakover), which is an unintended turn-on if it happens below the rated blocking voltage. Second, a fast rising anode voltage drives a displacement current $C_{J2}dV/dt$ through the depleted $J_2$ junction, which acts like gate current into $p_2$; if it exceeds the trigger level the device fires at far below $V_{BO}$. Practical devices raise $(dV/dt{)}_{crit}$ by adding emitter shorts (shorting the $p_2$ base to the cathode metal so that displacement current bypasses the emitter junction) and by using external RC snubbers across the device. Reference: S.M. Sze and Kwok K. Ng, Physics of Semiconductor Devices, 3rd ed., Wiley, 2007.

Ques: Why is the on-state voltage drop of a thyristor low even though it blocks thousands of volts?

In the on-state all three junctions are effectively forward biased and the lightly doped, wide $n_1$ base is flooded with an electron-hole plasma, so the device behaves like a forward-biased PIN diode with conductivity modulation of the wide intrinsic-like region (see the PIN Diode page). This lets the thick, lightly doped base that provides high blocking voltage conduct with very low resistance once latched. The price is the same as in a PIN diode: large stored charge in that plasma, which is why turn-off ($t_q$) is slow compared with unipolar devices. Reference: S.M. Sze and Kwok K. Ng, Physics of Semiconductor Devices, 3rd ed., Wiley, 2007.

Related pages in this vault: Gate Turn Off (GTO), Silicon Controlled Switch (SCS), Integrated Gate-Commutated Thyristor (IGCT), Bipolar Junction Transistor (BJT), Diode for Alternating Current (DIAC), Insulated Gate Bipolar Transistor (IGBT), Silicon Power MOSFET, IGBT and Thyristors, On-Chip ESD Protection Devices, Power Device Compact Models.