A heterojunction is an interface between two different semiconductor materials. Because the materials generally have different electronic band structures, their interface can create energy barriers or wells that control the motion and spatial distribution of electrons and holes. Heterojunctions are central to semiconductor lasers, high-frequency transistors, light-emitting devices, photodetectors, and solar cells. A structure containing one or more such interfaces is commonly called a heterostructure.
Unlike a p–n junction, which is defined by a change in conductivity type, a heterojunction is defined by a change in material. A junction can therefore be both a heterojunction and a p–n junction, but heterojunctions can also connect two n-type or two p-type regions.
Physical basis
In a semiconductor, the electronic band structure describes the allowed energies of electrons. The conduction-band minimum, , and valence-band maximum, , are separated by the band gap:
Different semiconductor materials have different band gaps, electron affinities, dielectric properties, and carrier effective masses. When two materials meet, their conduction- and valence-band edges generally do not join continuously. The discontinuities are called band offsets.
These offsets influence electrons in the conduction band and holes in the valence band differently. Electrons preferentially occupy regions with lower conduction-band edges, while holes preferentially occupy regions with higher valence-band edges. Suitable material combinations can therefore confine both carrier types in the same region or separate them into different regions.
Band offsets must be distinguished from band bending. An offset is a discontinuity associated with the change of material; band bending is a spatial variation of band-edge energies caused by an electrostatic potential. A real heterojunction commonly exhibits both.
Equilibrium and charge redistribution
When two semiconductors are brought into electrical contact and allowed to reach thermal equilibrium, charge may transfer between them until their Fermi levels align. The resulting space charge creates an electric field and a built-in potential. Depending on the materials and their doping, the interface may contain depleted regions or an accumulation of mobile carriers.
The equilibrium charge distribution depends on band offsets, doping concentrations, temperature, interface states, and fixed charges. Under an applied voltage or illumination, the junction is out of equilibrium; separate electron and hole quasi-Fermi levels are then used to describe carrier populations.
Band-alignment types
Heterojunctions are commonly classified into three types according to the relative positions of their band edges. This classification describes the underlying band alignment, rather than every detail of an operating device’s band diagram.
| Type | Name | Band arrangement | Typical consequence |
|---|---|---|---|
| I | Straddling gap | The band gap of one material lies within that of the other | Electrons and holes can both be confined in the narrower-gap material |
| II | Staggered gap | The conduction-band minimum and valence-band maximum favored by carriers lie in different materials | Electrons and holes tend to occupy opposite sides of the interface |
| III | Broken gap | The conduction-band minimum of one material lies below the valence-band maximum of the other | Occupied valence-band states and available conduction-band states can overlap in energy |
Type I: straddling alignment
Type-I heterojunctions are particularly useful for light emission because they can confine electrons and holes in the same active region, increasing their spatial overlap and the probability of radiative recombination.
Gallium arsenide surrounded by appropriately chosen aluminum gallium arsenide barriers is a standard example. The alignment depends on alloy composition; material names alone do not always determine the alignment over an entire composition range.
Type II: staggered alignment
In a type-II junction, the lowest-energy location for electrons and the preferred location for holes are in different materials. This spatial separation can assist photogenerated charge separation, although actual extraction efficiency also depends on transport and recombination.
Optical transitions may occur between states on opposite sides of the interface. Such transitions can have energies smaller than either constituent’s bulk band gap and often have reduced electron–hole wavefunction overlap.
Type III: broken-gap alignment
A type-III junction has an energetic overlap between the conduction band of one material and the valence band of the other. The indium arsenide–gallium antimonide material system is a well-known example.
Broken-gap structures are useful for studying and engineering interband tunneling. Their electrical behavior also depends on doping, charge transfer, layer thickness, and applied bias; band overlap alone does not fully specify the current–voltage characteristic.
Band-offset models
A simple estimate of band alignment is provided by the electron-affinity rule, also called the Anderson model. Electron affinity, , is the energy difference between the vacuum level and the conduction-band minimum.
For materials and , define
If their vacuum energy levels are aligned and interface-specific effects are neglected, then
The band gaps and offsets satisfy
The electron-affinity rule is a useful starting point, not a universally accurate prediction. Chemical bonding, interface dipoles, strain, crystal orientation, and intermixing can alter the offsets substantially. Accurate device analysis often requires experimental measurements or detailed electronic-structure calculations.
Materials and fabrication
Heterojunctions can be formed between compound semiconductors, between elemental semiconductors and their alloys, between amorphous and crystalline materials, or between organic semiconductors. Important examples include GaAs/AlGaAs, InGaAs/InP-based structures, AlGaN/GaN, silicon/silicon–germanium, and amorphous-silicon/crystalline-silicon interfaces.
Epitaxial heterojunctions
Many high-performance devices use epitaxy, in which a crystalline layer grows with an orientation determined by the underlying crystal. Major fabrication methods include molecular-beam epitaxy and metalorganic vapor-phase epitaxy.
The difference in the materials’ natural lattice spacings is a critical constraint. A sufficiently thin mismatched layer may remain coherently strained to match the underlying lattice. As thickness increases, strain relaxation can produce misfit dislocations, which may increase leakage or nonradiative recombination.
Material selection therefore involves trade-offs among band alignment, lattice compatibility, strain, thermal expansion, chemical stability, and growth conditions. An alloy’s composition can be varied to tune several of these properties.
Abrupt and graded interfaces
An abrupt heterojunction has a composition change over a short distance, ideally approaching the atomic scale. A graded heterojunction changes composition progressively through a finite region.
Composition grading can spread an energy barrier over a larger distance, assist carrier transport, or manage lattice mismatch. Real interfaces are never perfectly ideal: roughness, diffusion, alloy fluctuations, and chemical reactions can modify their behavior.
Van der Waals heterostructures
Atomically thin materials can be stacked to create van der Waals heterostructures. Because adjacent layers can interact without forming a conventional network of interfacial covalent bonds, strict lattice matching is less restrictive than in conventional epitaxy.
Their properties depend on layer composition, stacking order, relative rotation, contamination, and the dielectric environment. Such structures extend heterojunction engineering to combinations that are difficult to realize through conventional crystal growth.
Carrier confinement and transport
A thin layer bounded by suitable heterojunction barriers can form a quantum well. When the confinement dimension becomes sufficiently small, carrier motion perpendicular to the interfaces is quantized into discrete subbands.
Repeated wells and barriers form multiple-quantum-well structures or, when interwell coupling is important, a superlattice. Layer thickness and composition provide additional control over electronic and optical properties beyond those available in bulk materials.
Carrier transport across heterojunctions can involve:
- Diffusion and drift, driven by carrier-density gradients and electric fields.
- Thermionic emission, in which carriers have enough thermal energy to cross a barrier.
- Tunneling, through sufficiently thin barriers or between suitable energy bands.
- Recombination, including radiative transitions and defect-assisted processes.
The dominant mechanism depends on the band profile, temperature, doping, applied voltage, and interface quality.
Device applications
Semiconductor lasers and light-emitting diodes
A semiconductor laser commonly uses a double heterostructure: a lower-band-gap active layer lies between wider-band-gap cladding layers. These layers can confine injected carriers and, through a suitable refractive-index profile, guide the optical field. The arrangement reduces carrier escape and improves the conditions for stimulated emission.
Quantum-well structures provide further control over the available electronic states and emission wavelength. Many light-emitting diodes likewise use heterostructures to confine carriers in an emitting region.
Heterojunction bipolar transistors
A heterojunction bipolar transistor uses a heterojunction, commonly at the emitter–base interface, to improve carrier injection. A wider-band-gap emitter can suppress unwanted injection from the base into the emitter.
This allows different doping and geometry choices from those available in a conventional bipolar junction transistor, supporting high-frequency operation. Silicon–germanium-base devices are used in radio-frequency and mixed-signal circuits, while compound-semiconductor variants serve demanding high-speed applications.
High-electron-mobility transistors
A high-electron-mobility transistor exploits a heterostructure to create a conducting channel near an interface. In modulation-doped structures, the dopant atoms supplying electrons can be spatially separated from the channel, reducing scattering by ionized impurities.
GaAs/AlGaAs is a classic material system. In AlGaN/GaN structures, polarization-related charges can generate a high-density electron channel. Such transistors are important in microwave electronics and power conversion.
Photodetectors and solar cells
Heterojunctions can combine an absorbing material with layers designed for carrier collection, confinement, or surface passivation. They are used in several forms of photodiode, including devices that separate light absorption from carrier multiplication.
In solar cells, heterojunctions can provide selective contacts that favor collection of one carrier type while limiting recombination. Silicon heterojunction cells commonly combine crystalline silicon with thin hydrogenated amorphous-silicon layers, including intrinsic passivation layers. Multijunction solar cells use semiconductor subcells with different band gaps to absorb different parts of the solar spectrum.
Historical development
Heterojunction device concepts developed during the early decades of semiconductor electronics. In 1957, Herbert Kroemer proposed using heterojunctions to improve bipolar-transistor performance.
In 1963, Kroemer and, independently, Zhores Alferov and Rudolf Kazarinov proposed double-heterostructure semiconductor laser concepts. Improvements in crystal growth subsequently enabled practical structures with controlled composition and high-quality interfaces. Room-temperature continuous-wave operation of double-heterostructure semiconductor lasers was achieved in 1970.
The 2000 Nobel Prize in Physics recognized Alferov and Kroemer for developing semiconductor heterostructures used in high-speed electronics and optoelectronics; the other half of the prize recognized Jack Kilby’s contribution to the integrated circuit.
Limitations and characterization
The advantages of a heterojunction depend strongly on its interface quality. Defects can trap carriers, introduce nonradiative recombination pathways, scatter carriers, or produce leakage currents. An energy barrier that improves confinement in one part of a device can impede useful carrier injection or extraction elsewhere.
Practical limitations include lattice mismatch, thermal instability, interdiffusion, contact resistance, processing complexity, and variability across a semiconductor wafer. Polarization fields and interface charge can be useful design tools but may also distort the intended band profile.
No single measurement fully characterizes a heterojunction. Structural methods examine composition, strain, and atomic arrangement; optical measurements probe transitions and recombination; electrical measurements reveal transport and charge response. Photoelectron spectroscopy can help determine band offsets, while microscopy and diffraction assess interface structure. Interpreting these measurements requires accounting for band bending, defects, and the complete device architecture.