A bipolar junction transistor (BJT) is a three-terminal semiconductor device used to amplify electrical signals and switch electrical currents. It consists of three differently doped regions—the emitter, base, and collector—arranged as either NPN or PNP. Its operation depends on the injection of charge carriers across one junction and their collection across another, with a thin intervening base allowing the two junctions to interact. The term bipolar indicates that both electrons and holes participate in its operation, although one carrier type normally carries most of the current through the base.
The BJT was central to the replacement of vacuum tubes and the development of electronic computing, telecommunications, and integrated circuits. Although the metal–oxide–semiconductor field-effect transistor dominates modern digital logic, bipolar devices remain important in precision analog circuitry, radio-frequency electronics, power amplification, and specialized switching applications. Their usefulness derives from predictable relationships between voltage, current, and temperature, high transconductance at a given operating current, and the availability of complementary NPN and PNP structures.
Physical structure and terminal roles
A BJT contains two p–n junctions sharing a central semiconductor region. In an NPN transistor, a p-type base lies between n-type emitter and collector regions; in a PNP transistor, the conductivity types are reversed. Doping establishes the concentrations of mobile carriers and fixed ionized impurities in each region. N-type material has electrons as majority carriers, whereas p-type material has holes, which behave as positively charged mobile carriers associated with unoccupied states in the valence band.
The three regions are not normally interchangeable. The emitter is designed to inject carriers efficiently into the base and is commonly more heavily doped than the base. The base is sufficiently thin that most injected carriers cross it without recombining. The collector is designed to collect those carriers and withstand the intended collector–base voltage. It is often more lightly doped and physically larger than the emitter, particularly in devices intended for appreciable voltage or power. Consequently, exchanging emitter and collector usually produces much lower gain and different breakdown limits, even though transistor action can still occur.
At each junction, mobile carriers initially diffuse across the boundary and leave behind charged dopant ions. The resulting depletion region contains relatively few mobile carriers and supports an internal electric field. In thermal equilibrium, that field counterbalances diffusion, producing no net terminal current. Applying an external voltage changes the junction barrier: forward bias lowers it and increases carrier injection, whereas reverse bias generally enlarges the depletion region and suppresses injection until leakage or breakdown mechanisms become significant.
The physically relevant base width is the distance between the edges of the two depletion regions, not simply the distance between the metallurgical junctions. This neutral base width changes with terminal voltages. In a suitably designed transistor it is short compared with the distance minority carriers can diffuse before recombining. Its exact dimensions vary greatly between general-purpose discrete devices, power transistors, and advanced radio-frequency structures; there is no single characteristic BJT base thickness.
In circuit symbols, an arrow on the emitter identifies conventional current direction through the emitter–base junction when forward biased. It points outward for NPN and inward for PNP. Conventional current and electron motion have opposite directions. Package appearance does not reliably identify terminal order: devices with similar outlines can have different lead assignments, and an integrated transistor may have no individually accessible external package terminals.
Carrier transport and operating regions
In the usual forward-active operation of an NPN transistor, the base–emitter junction is forward biased and the base–collector junction is reverse biased. The emitter injects electrons into the p-type base, where they are minority carriers. Their concentration is highest near the emitter and lower near the collector, producing diffusion across the neutral base. Electrons reaching the collector-side depletion region are swept into the collector by its electric field. Thus, a reverse-biased collector junction can carry a substantial current because the emitter supplies carriers that would otherwise be scarce.
Some injected electrons undergo recombination in the base, requiring holes to be supplied through the base terminal. Holes also cross from the base into the emitter, and recombination can occur within the emitter–base depletion region. These processes contribute to base current. High emitter injection efficiency and low recombination in the neutral base allow a relatively small base current to accompany a much larger collector current. PNP operation follows the same principles with carrier types and voltage polarities reversed.
Four idealized operating regions are distinguished by the bias states of the two junctions. In cutoff, neither junction is appreciably forward biased, and terminal currents are dominated by leakage. In forward active, the emitter–base junction is forward biased and the collector–base junction reverse biased. In saturation, both junctions are forward biased, so the collector current is no longer described adequately by the forward-active current-gain relation. In reverse active, the collector–base junction injects carriers and the emitter–base junction collects them. Because practical structures are asymmetric, reverse-active gain is usually much smaller than forward-active gain.
For an NPN transistor in forward-active operation, the conventional emitter-current magnitude satisfies
The common-base current gain and common-emitter current gain are defined, respectively, as
A transport efficiency close to unity therefore permits a much larger common-emitter gain: , for example, corresponds to . Actual gain depends on collector current, temperature, collector voltage, manufacturing variation, and frequency.
The familiar description of a BJT as a “current-controlled device” is useful in circuit calculations but does not fully describe its physics. Collector current is strongly related to base–emitter voltage through carrier injection, while base current reflects injection efficiency and recombination losses. A complete description must therefore include both voltage-dependent transport and the currents required to maintain that transport. At sufficiently high current, high-level injection and internal voltage drops further modify these relationships.
Current–voltage equations and device models
For a silicon BJT in forward-active operation, an important first approximation is
where is a scale current determined by device geometry, material properties, and doping; is Boltzmann’s constant; is absolute temperature; and is the elementary charge magnitude. At , the thermal voltage is approximately . When collector current is much larger than , increasing by about raises the ideal collector current by a factor of ten.
This exponential law explains why a silicon transistor does not have a fixed “turn-on voltage.” A quoted base–emitter voltage of approximately – describes common operating conditions, not a universal threshold. At lower currents the voltage can be substantially smaller, and at high current it can be larger. Temperature, emitter area, series resistance, and the detailed transistor structure also affect it. At constant collector current, the base–emitter voltage of many silicon BJTs decreases by roughly near room temperature, but the coefficient is operating-point dependent.
Collector current is not perfectly independent of collector voltage. Increasing reverse bias across the collector–base junction widens its depletion region into the base, reducing the neutral base width and increasing carrier transport. This base-width modulation is called the Early effect. A common approximate expression is
where is the Early-voltage parameter. The model gives an incremental output resistance of approximately when the other relevant conditions are held fixed and the simplifying assumptions apply.
The Ebers–Moll model, introduced by John Ebers and John Moll in 1954, represents both junctions and their coupled carrier transport. Unlike a single forward-active equation, it describes cutoff, saturation, and reverse operation within a unified large-signal framework. More elaborate models account for effects such as high-level injection, current-dependent gain, base resistance, charge storage, and the variation of junction capacitances with bias. The Gummel–Poon model became especially influential in integrated-circuit simulation.
For small signals, nonlinear equations are linearized around a direct-current operating point. For larger excursions, switching, or distortion analysis, time-dependent nonlinear models are necessary. Circuit simulators such as SPICE use model parameters rather than a single gain value to represent this behavior. A model fitted for low-frequency analog work may not capture microwave operation, power-device quasi-saturation, or self-heating adequately; model validity depends on both the represented physics and the range over which parameters were characterized.
Small-signal amplification and circuit configurations
A BJT becomes an amplifier when a bias network establishes an operating point and a small input signal produces a controlled change in collector current. Its key incremental parameter is transconductance, the change in output current per change in controlling voltage. Under the ideal forward-active approximation,
At and , this gives approximately . The value depends primarily on operating current and temperature rather than on current gain, an important advantage in analog design.
The hybrid- small-signal model represents the base–emitter input by an incremental resistance , the collector response by a controlled current source , and the Early effect by an output resistance . In the basic model,
Here is the incremental current gain at the operating point, which need not equal a large-signal gain measured under different conditions. At higher frequencies, base–emitter and base–collector capacitances, base resistance, and additional parasitic elements must be included.
The common-emitter configuration normally provides both voltage gain and current gain. A rise in base voltage increases collector current, increasing the voltage drop across a collector load and thereby lowering collector voltage. Its output is therefore inverted relative to its input. With an effectively grounded emitter and a resistive collector load, the approximate voltage gain is , before allowing for external loading. An emitter resistor introduces local negative feedback, reducing gain while improving linearity and making the operating point less sensitive to gain variation.
The common-collector configuration, usually called an emitter follower, takes its output from the emitter. Its voltage gain is somewhat less than unity, but it can transform a relatively high input impedance into a lower output impedance and provide substantial current gain. The common-base configuration holds the base approximately fixed for the signal and applies the input at the emitter. Its low input resistance and limited interaction between input and collector make it useful in some wideband circuits. A cascode combines common-emitter and common-base stages to improve output resistance and reduce feedback through collector–base capacitance.
Practical amplifiers often use a differential pair rather than an isolated transistor. Two matched BJTs share a tail current, and their current division responds to the difference between their base voltages. Current mirrors establish or reproduce bias currents using matched transistor characteristics. These building blocks are central to operational amplifiers, comparators, and analog signal-processing circuits, where matching and feedback are generally more reliable design resources than an assumed exact transistor gain.
Switching behavior and high-frequency response
As a switch, a BJT is commonly operated between cutoff and saturation. In a simple NPN low-side switch, the emitter is connected to the low-potential supply rail, the load lies between the positive supply and the collector, and base drive controls conduction. With adequate drive, the collector–emitter voltage can fall to a small fraction of a volt at modest current. The actual saturation voltage depends on collector current, base current, device construction, and temperature; it is not a fixed material constant.
A saturated transistor cannot be analyzed by indefinitely extending . Once the base–collector junction becomes forward biased, additional base drive increasingly stores charge rather than producing a proportional increase in collector current. Switching specifications therefore often give a forced beta, the ratio of collector current to applied base current under the stated test conditions. This ratio describes the drive condition and should not be confused with the transistor’s forward-active current gain.
Stored charge makes turn-off more complicated than simply removing the input voltage. In saturation, excess carriers occupy the base and, in some structures, parts of the collector region. These carriers must be removed or recombine before the collector voltage can rise fully. The resulting storage interval can dominate switching delay. Reverse base current can accelerate charge removal, while clamp circuits can prevent deep saturation. A Schottky-clamped transistor uses a metal–semiconductor diode to divert base drive before substantial forward bias develops across the base–collector junction.
Small-signal high-frequency behavior is characterized in part by the transition frequency , at which the magnitude of short-circuit common-emitter current gain falls to unity. In a simplified charge-control description,
where represents the combined effects of carrier transit and charging delays. The transition frequency depends strongly on bias current and collector voltage. It is neither the highest frequency at which a transistor can have power gain nor a direct specification of a complete circuit’s switching speed. The maximum oscillation frequency, , is a separate high-frequency figure related to available power gain.
Collector–base capacitance is especially significant in an inverting voltage amplifier because changes in collector voltage feed charge back toward the input. The resulting Miller effect can make its effective input contribution much larger than its physical value. At high current, collector resistance and the Kirk effect—an effective displacement of the base boundary into the collector—can increase transit time. Package inductance, contact resistance, interconnect capacitance, and the surrounding circuit then become as consequential as the intrinsic transistor structure.
Materials, fabrication, and specialized structures
Early junction transistors used germanium, but silicon became the principal material for general-purpose BJTs because of its lower leakage at ordinary operating temperatures, greater temperature capability, and compatibility with effective oxide passivation. Modern fabrication controls doping profiles and dimensions using processes such as diffusion, ion implantation, epitaxial growth, and photolithography. The required sequence depends on whether the goal is a low-cost discrete transistor, an integrated analog device, a microwave transistor, or a high-voltage power component.
A common integrated vertical NPN structure places an n-type emitter inside a p-type base formed in an n-type collector region. An epitaxial collector layer can provide voltage support, while a heavily doped buried layer reduces collector resistance. Isolation structures separate neighboring components electrically. In many traditional bipolar and mixed-signal processes, lateral PNP transistors are easier to incorporate than high-performance vertical PNP devices, but they often have lower gain or poorer high-frequency performance than the process’s optimized NPN devices. Complementary bipolar processes are designed to reduce this disparity.
Power transistors place particular emphasis on collector thickness, current distribution, heat flow, and contact resistance. Their active area may contain many parallel emitter cells or interdigitated emitter and base contacts rather than one simple junction. This geometry limits the distance that base current must travel through resistive material and helps distribute collector current. Device area alone does not determine useful power capability: the package, die attachment, thermal path, and susceptibility to localized current concentration are also important.
A heterojunction bipolar transistor uses different semiconductor materials on the two sides of at least one junction, usually the emitter–base junction. Band-gap engineering suppresses unwanted carrier injection from base to emitter, allowing a heavily doped base with low resistance while retaining useful gain. Silicon–germanium HBTs and III–V devices, including gallium arsenide-based structures, are important in high-frequency electronics. SiGe processes can combine fast bipolar transistors with CMOS circuitry in a BiCMOS technology.
Not every component marketed as a bipolar transistor contains only one transistor. A Darlington pair connects two BJTs so that the first supplies base current to the second, producing a very high overall current gain but generally increasing the required input voltage and complicating turn-off. Phototransistors use absorbed light to generate carriers that control transistor current. These devices are related to a photodiode coupled to transistor action, although their detailed structures and response characteristics differ from those of ordinary three-terminal BJTs.
Breakdown, temperature, noise, and reliability
A reverse-biased junction can enter avalanche breakdown when carriers accelerated by its electric field generate additional electron–hole pairs through impact ionization. BJT voltage ratings distinguish among different terminal conditions. Collector–emitter breakdown with the base open, commonly denoted , is often lower than collector–base breakdown with the emitter open, . Avalanche-generated carriers can supply base drive and undergo transistor amplification, explaining why the complete three-terminal structure cannot be treated as two independent diode breakdown ratings.
Reverse base–emitter voltage is another limitation. In many silicon BJTs, its permissible value is much lower than the collector-related voltage ratings. Repeated or substantial reverse emitter–base breakdown can degrade gain and increase noise, even when immediate catastrophic failure does not occur. The precise behavior is device dependent, which is why terminal voltage, current, and power ratings must be considered separately rather than reduced to one maximum operating number.
Temperature influences both electrical behavior and allowable power dissipation. At a fixed base–emitter voltage, collector current generally rises strongly as temperature increases. If the associated dissipation causes further heating, positive feedback can produce thermal runaway. At fixed collector current, by contrast, base–emitter voltage normally decreases with temperature. Emitter resistances, thermally coupled bias structures, and controlled current sources are common ways circuits counteract these dependencies. Parallel power BJTs are particularly sensitive to unequal current sharing unless their electrical and thermal design limits local imbalance.
The safe operating area specifies permissible combinations of current, voltage, and duration. In addition to maximum current and average heating, power BJTs may be limited by secondary breakdown, in which localized current concentration creates a hot region that can destroy the device even though total power appears acceptable. A simplified steady-state thermal relation is
where is junction temperature, ambient temperature, dissipation, and an applicable junction-to-ambient thermal resistance. Its usefulness depends on matching the assumed mounting and cooling conditions; pulses require transient thermal analysis.
Noise is important when amplifying small signals. Shot noise arises from discrete charge transport, with the ideal collector-current noise spectral density commonly written in . Base current contributes another shot-noise source, base resistance contributes thermal noise, and low-frequency behavior can include flicker and generation–recombination noise. Consequently, a transistor with high gain is not automatically the lowest-noise choice for every source impedance. Input current noise, input voltage noise, bias current, device area, and frequency all influence the result.
Applications and relationships to other transistor technologies
Bipolar transistors are widely used where controlled analog currents and predictable exponential behavior are valuable. Discrete examples include microphone preamplifiers, audio output stages, oscillators, current drivers, and interface circuits. Integrated examples include operational amplifiers, voltage references, comparators, data-converter support circuits, and analog control systems. In many such applications the BJT is not operating as an isolated gain element but as part of a matched network whose behavior depends on device ratios, thermal coupling, and feedback.
The exponential current–voltage relationship has direct computational uses. For two matched transistors at different collector currents,
under the assumptions of the ideal model. A fixed current-density ratio therefore generates a voltage proportional to absolute temperature. Bandgap voltage references combine a scaled version of this voltage with a base–emitter voltage having a negative temperature coefficient. The resulting first-order temperature compensation gives a reference near the familiar silicon bandgap-reference value of roughly . Related principles support temperature sensors, logarithmic amplifiers, and translinear circuits.
BJT-based digital families include transistor–transistor logic and emitter-coupled logic. In conventional TTL, bipolar transistors perform logic and drive outputs, with circuit variants differing in speed, power consumption, and saturation control. Emitter-coupled logic steers a nearly constant current through a differential stage and avoids transistor saturation, enabling rapid transitions at the cost of continuous power consumption. These families were important in computing and communications before CMOS became dominant in large-scale digital integration.
Compared with an insulated-gate transistor, a BJT requires sustained base current in normal operation and has a lower incremental input resistance. Its transconductance per unit bias current is particularly attractive for many analog tasks, while its exponential characteristic supports accurate current ratios. MOSFETs offer very high direct-current input impedance and particularly favorable scaling and static-power characteristics for digital logic. Comparisons depend on operating region, process, voltage, noise requirements, and circuit topology; neither device type is uniformly superior across all uses.
In power switching, MOSFETs and the insulated-gate bipolar transistor have replaced conventional power BJTs in many roles because their gate drive is easier to implement and their overall characteristics suit particular voltage and frequency ranges. The IGBT combines an insulated-gate input with bipolar carrier injection in its power-conduction structure; it is not simply a BJT with a different terminal label. In radio-frequency systems, meanwhile, silicon–germanium and compound-semiconductor HBTs remain important alongside CMOS, compound-semiconductor field-effect devices, and other specialized technologies.
Invention and historical development
The BJT emerged from research aimed at understanding and exploiting semiconductor behavior for amplification. At Bell Laboratories in December 1947, John Bardeen and Walter Brattain demonstrated the point-contact transistor. That device established practical transistor amplification but was not a junction transistor in the later structural sense. William Shockley subsequently developed the junction-transistor concept and its theoretical basis, recognizing how a thin base between two semiconductor junctions could control carrier transport.
Shockley’s junction-transistor work in 1948 was followed by the production of grown-junction devices at Bell Labs, with an important demonstration in 1950 and public announcement in 1951. In a grown-junction transistor, changes in dopant composition during crystal growth produced the successive conductivity regions. Alloy-junction methods offered another manufacturing route, particularly for germanium devices. These early techniques made transistor production possible but offered less control over geometry and high-frequency performance than later diffusion and planar processes.
Silicon transistors became commercially significant during the 1950s. Gordon Teal’s team at Texas Instruments announced a commercial silicon transistor in 1954. Silicon devices were especially attractive for operation at higher temperatures than germanium devices could readily tolerate. During the same period, improved diffusion techniques enabled more precise junction formation. Jean Hoerni’s planar process, developed at Fairchild Semiconductor in 1959, used silicon dioxide to protect the semiconductor surface and support patterned processing, becoming foundational to reliable silicon devices and integrated circuits.
Theoretical and circuit developments accompanied manufacturing advances. The Ebers–Moll model supplied a general large-signal description in 1954, while later charge-control and Gummel–Poon approaches improved representation of dynamic and nonideal behavior. Bipolar integrated circuits became central to analog electronics and to early digital logic families. Bardeen, Brattain, and Shockley received the 1956 Nobel Prize in Physics for their semiconductor research and discovery of the transistor effect.
During the later twentieth century, MOS technology and especially CMOS displaced bipolar logic from most high-density digital computing because of their integration density and power characteristics. Bipolar development nevertheless continued in precision analog, power, and microwave technologies. Silicon–germanium BiCMOS brought high-performance heterojunction bipolar devices into processes that also supported dense MOS circuitry. In these platforms, digital control, memory, analog biasing, and radio-frequency signal paths can share a chip while using different transistor structures for their distinct electrical requirements.