Gallium arsenide (GaAs) is a III-V compound semiconductor made from gallium and arsenic. It matters because it does two things silicon does poorly: its electrons move about six times faster, which suits radio-frequency amplifiers and switches, and its direct bandgap lets it emit light efficiently, which suits LEDs, lasers and high-efficiency solar cells. Silicon still wins on cost, wafer size and integration density, so GaAs lives in the parts of a system where speed, light or efficiency matter more than price: the RF front end of a phone, the laser in an optical link, the solar panel on a satellite.
Key takeaways
- GaAs is a III-V semiconductor with a zincblende crystal structure, a direct bandgap of about 1.42 eV at room temperature and electron mobility of roughly 8,500 cm²/V·s.
- High mobility gives fast transistors (HEMTs, pHEMTs, HBTs) for RF and microwave work; the direct bandgap gives efficient infrared LEDs and lasers.
- Its weak points are cost, small and brittle wafers, poor hole mobility, lower thermal conductivity than silicon and the toxicity of arsenic.
- For high-power switching, GaN and SiC have taken the lead; GaAs remains strong in phone power amplifiers, RF switches, VCSELs and space solar cells.
- Gallium supply is concentrated: most primary gallium is refined in China, which has restricted exports since 2023.
What is gallium arsenide?
Gallium arsenide is a III-V compound semiconductor, meaning it is formed by combining elements from the third and fifth groups of the periodic table: gallium (Ga) and arsenic (As). Unlike elemental semiconductors such as silicon or germanium, GaAs is a compound with a zincblende crystal structure. Each gallium atom sits bonded to four arsenic atoms and the reverse, which gives it distinct electrical and optical properties.
Because the lattice is built from two elements, related compounds can be grown on top of it with almost the same atomic spacing. Aluminium gallium arsenide (AlGaAs) is the classic example. Stacking GaAs and AlGaAs layers lets engineers build heterojunctions, the layered structures behind high electron mobility transistors, quantum well lasers and multi-junction solar cells.
Key properties of GaAs
- High electron mobility: GaAs has an electron mobility of approximately 8,500 cm²/V·s, significantly higher than silicon's (about 1,400 to 1,450 cm²/V·s). Electrons move faster through the material, enabling high-speed devices.
- Direct bandgap: GaAs has a direct bandgap of about 1.42 to 1.43 eV at room temperature. Electrons can drop across the gap and emit a photon without needing help from a lattice vibration, which makes it ideal for LEDs and lasers. The matching wavelength is in the near infrared, close to 870 nm.
- High frequency performance: GaAs transistors can operate at frequencies exceeding 100 GHz, making the material suitable for RF and microwave applications.
- Semi-insulating substrate: GaAs can be grown with very high resistivity. Passive parts such as inductors and transmission lines can then sit on the same die with low loss, which is why monolithic microwave integrated circuits (MMICs) are often built on GaAs.
- Thermal behaviour: GaAs has lower thermal conductivity than silicon, but its wider bandgap means devices keep working at higher junction temperatures before intrinsic leakage takes over.
- Breakdown field: GaAs has a somewhat higher breakdown electric field than silicon (roughly 0.4 MV/cm against 0.3 MV/cm), so it can handle a little more voltage across the same thickness before failing.
- Hole mobility: the weak point. Hole mobility is only about 400 cm²/V·s, so GaAs has no practical equivalent of silicon CMOS. That one fact explains why GaAs never replaced silicon for digital logic.
How GaAs compares with other semiconductors
The table gives approximate room-temperature textbook values. Exact numbers vary with doping, crystal quality and source, so treat them as orders of magnitude.
| Property | Silicon | GaAs | GaN | SiC (4H) |
|---|---|---|---|---|
| Bandgap (eV) | 1.12 (indirect) | 1.42 (direct) | 3.4 (direct) | 3.26 (indirect) |
| Electron mobility (cm²/V·s) | about 1,400 | about 8,500 | about 1,000 bulk, near 2,000 in a HEMT channel | about 900 |
| Breakdown field (MV/cm) | about 0.3 | about 0.4 | about 3.3 | about 3 |
| Thermal conductivity (W/cm·K) | about 1.5 | about 0.5 | about 1.3 | about 3.7 to 4.9 |
| Main uses | Logic, memory, most analog | RF front ends, LEDs, lasers, space solar | Power conversion, RF power, blue and white LEDs | High-voltage power devices |
Read the table as a set of trade-offs. GaAs beats silicon on speed and light emission, but GaN and SiC, with bandgaps more than twice as wide, beat GaAs on voltage and power density. That is why the industry uses all four.
History of gallium arsenide
The story of gallium arsenide begins with the discovery of its constituent elements. Gallium was discovered in 1875 by French chemist Paul-Émile Lecoq de Boisbaudran, who named it after Gallia, the Latin name for France. Arsenic compounds have been known since ancient times, and the element itself is usually credited as being isolated in the Middle Ages. The potential of combining the two to form a semiconductor was not realised until the 20th century.
The first version of this article said that German physicist Walter Schottky theorised about the electronic properties of compound semiconductors in 1926. We have not been able to confirm that date. Schottky's best known contribution here is his theory of the metal-semiconductor barrier from the late 1930s, the basis of the Schottky gate in GaAs MESFETs. GaAs itself is reported to have been first synthesised by Victor Goldschmidt in 1929.
The key step came in the early 1950s. Heinrich Welker at Siemens showed in 1952 that III-V compounds such as GaAs and indium antimonide behave as semiconductors, and research groups, including Bell Laboratories, began growing GaAs crystals in that period (the earlier version of this article dated Bell Labs crystal growth to 1952; that exact date is not re-verified). The material's direct bandgap and high electron mobility were soon recognised as promising.
The 1960s saw the first GaAs devices. Texas Instruments engineers James Biard and Gary Pittman developed a GaAs infrared LED, and TI is credited with the first commercial GaAs infrared LED in 1962. The same year, several laboratories (General Electric, IBM and MIT Lincoln Laboratory) demonstrated GaAs semiconductor lasers. In 1963 J. B. Gunn at IBM observed microwave oscillations in GaAs, the Gunn effect. The first version of this article also credited an IBM researcher, Gunther Fenner, with the first GaAs field-effect transistor in 1963; we could not find a record of that. The usual reference points are Carver Mead's proposal of the MESFET in 1966 and a working GaAs MESFET reported by Hooper and Lehrer at Texas Instruments in 1967.
The 1970s and 1980s were pivotal decades for GaAs. Molecular beam epitaxy (MBE), developed by Alfred Y. Cho and John Arthur at Bell Labs in the late 1960s, gave atomic-layer control over crystal growth. That made possible quantum well lasers and, in 1979 to 1980, the high electron mobility transistor (HEMT), first shown by Takashi Mimura at Fujitsu using AlGaAs on GaAs. By the 1980s, GaAs was being used in microwave and RF applications, particularly in defence and satellite communications. Metal-organic chemical vapour deposition (MOCVD) further improved GaAs production, making it viable for commercial volumes.
In the 1990s, GaAs found its way into consumer electronics, especially mobile phones, where its high-frequency performance enabled compact, efficient RF power amplifiers. In the 2000s, GaAs-based multi-junction solar cells became the standard for space applications because of their high efficiency and radiation tolerance. Today, research continues on integrating GaAs with silicon and on its use in quantum devices and future wireless systems.
How GaAs wafers and devices are made
Producing high-quality GaAs wafers is a complex and costly process, which is one reason silicon remains more common. The main steps are as follows.
1. Raw material purification
Gallium is mostly recovered as a by-product of alumina refining from bauxite, with smaller amounts from zinc processing. Arsenic is mostly recovered as a by-product of copper and gold smelting, from minerals such as arsenopyrite. Both must be purified to very high levels (99.9999 per cent, "6N", or better) to produce low-defect crystals.
2. Crystal growth
Arsenic evaporates readily from molten GaAs, so plain Czochralski pulling as used for silicon does not work. The liquid encapsulated Czochralski (LEC) method covers the melt with a layer of boron oxide to hold the arsenic in while a single crystal is pulled. The vertical gradient freeze (VGF) technique, and the related vertical Bridgman method, solidify the crystal slowly in a crucible with a controlled temperature gradient, which produces fewer dislocations. VGF is now common for the substrates used in RF and laser production.
3. Wafer slicing and polishing
The ingot is sliced into thin wafers with diamond saws and polished to a mirror finish. GaAs wafers are typically 100 mm or 150 mm in diameter, against 300 mm for leading silicon fabs, and they are much more brittle, which adds to the cost per device.
4. Epitaxial growth
To create working devices, thin layers of GaAs or related compounds (such as AlGaAs or InGaP) are deposited on the wafer by MBE or MOCVD. These methods control layer thickness and composition down to a few atomic layers, which is what makes heterojunction devices possible.
5. Device fabrication
The wafers then go through photolithography, etching, implantation and metallisation to form transistors, diodes and passive parts. The two main GaAs transistor types are the high electron mobility transistor (HEMT, and its pseudomorphic form pHEMT), used for low-noise amplifiers and RF switches, and the heterojunction bipolar transistor (HBT), used for phone power amplifiers. Both rely on the heterojunctions grown in step 4.
Despite advances in manufacturing, GaAs production remains expensive because of the cost of raw materials, the smaller wafers, more complex processing and lower yields than silicon.
Applications of gallium arsenide
GaAs's properties make it hard to replace in several areas that need high performance.
1. Telecommunications and RF
GaAs is widely used in RF and microwave devices, such as power amplifiers and low-noise amplifiers in smartphones, satellite communications and 4G and 5G networks. A typical phone contains several GaAs HBT power amplifier dies inside its front-end modules. The ability to operate efficiently at high frequency gives better signal quality and longer battery life.
2. Optoelectronics
The direct bandgap of GaAs makes it well suited to light-emitting and light-sensing devices:
- LEDs and lasers: GaAs-based infrared LEDs and laser diodes are used in fibre-optic links, remote controls and optical storage (for example CD players). Vertical-cavity surface-emitting lasers (VCSELs) on GaAs are used in short-reach data centre links and in the face-recognition and proximity sensors of phones.
- Photodetectors: GaAs photodiodes are used in optical communication systems and sensors because of their sensitivity and speed.
3. Solar cells
GaAs solar cells are among the most efficient made. Single-junction GaAs cells have reached a reported record of about 29 per cent, close to the theoretical limit for one junction. Multi-junction cells, which stack layers such as InGaP, GaAs and germanium with different bandgaps, capture more of the spectrum: they routinely exceed 30 per cent, and laboratory records under concentrated light are above 47 per cent, as listed on the NREL best research-cell efficiency chart. Their light weight, efficiency and radiation tolerance make them the standard power source for satellites and space probes.
A worked example shows why this matters in space. Sunlight above the atmosphere delivers about 1,361 W/m². A panel of 1 m² at 30 per cent efficiency gives about 408 W; at the 20 per cent typical of a good silicon space cell it would give about 272 W. For the same power, the GaAs-based panel is about a third smaller, and launch mass is expensive. (Illustrative figures, ignoring temperature, packing and ageing losses.)
4. Defence and aerospace
GaAs is used in radar, missile guidance and electronic warfare because of its high-frequency performance and radiation tolerance. Its ability to work in harsh environments makes it a staple of military and aerospace electronics, although GaN has taken over many new high-power radar designs.
5. Consumer electronics
GaAs is found in RF front-end modules in smartphones and Wi-Fi routers, where it provides compact, efficient power amplifiers and low-loss switches.
Advantages of gallium arsenide over silicon
While silicon dominates the semiconductor industry thanks to its abundance and mature manufacturing, GaAs offers several advantages:
- Speed: higher electron mobility gives faster transistors, which suits high-frequency circuits.
- Efficient light emission: the direct bandgap enables efficient light emission; silicon's indirect bandgap does not.
- Efficiency: GaAs solar cells and RF amplifiers outperform silicon equivalents in efficiency, which is critical for power-sensitive applications.
- Radiation tolerance: GaAs is more resistant to radiation damage, which suits space and military use.
- Low-loss substrate: semi-insulating GaAs lets passives and transistors share one die in MMICs.
Challenges and limitations
Despite its advantages, GaAs faces several challenges:
- High cost: GaAs wafers are much more expensive than silicon because of costly raw materials, smaller wafers and complex processing.
- Brittleness: GaAs is mechanically fragile, which increases the risk of wafer breakage in production.
- Thermal conductivity: at roughly a third of silicon's, it demands careful thermal design in high-power parts, for example thinned wafers and through-wafer vias to the heat sink.
- No CMOS: poor hole mobility rules out dense complementary logic, so control and digital functions stay in silicon.
- Scalability: silicon benefits from decades of process optimisation and enormous volume; GaAs remains a specialist material with smaller economies of scale.
- Environmental and safety concerns: arsenic and its compounds, especially arsine gas used in MOCVD, are toxic. Fabs need gas monitoring, abatement and controlled disposal of waste.
- Supply concentration: according to the US Geological Survey Mineral Commodity Summaries, China accounts for the large majority (reported at about 98 per cent) of primary gallium production. China introduced export licensing for gallium and germanium from August 2023 and, in December 2024, announced a ban on their export to the United States.
Common misconceptions
- "GaAs will replace silicon." It was called the material of the future in the 1980s, and it did not happen for digital logic, mainly because of hole mobility, cost and the absence of a good native oxide. It complements silicon instead.
- "GaAs is a power semiconductor." For power conversion in electric vehicles and chargers, the wide-bandgap materials GaN and SiC are the main choices. GaAs is a high-frequency and optoelectronic material.
- "GaN has made GaAs obsolete." GaN leads in high-power RF and power conversion, but GaAs remains cost-effective and well understood for handset power amplifiers, switches and VCSELs.
The future of gallium arsenide
As technology advances, GaAs is expected to keep a role in several fields:
- 6G and beyond: demand for higher-frequency components in future networks may drive further GaAs use, although GaN, SiGe and indium phosphide will compete at the highest frequencies.
- Quantum computing: GaAs-based quantum dots are being explored for spin qubits, and GaAs quantum dots are also studied as single-photon sources.
- Electric vehicles: the original version of this article suggested GaAs power electronics could improve EV charging and power management. In practice GaN and SiC dominate EV power conversion; GaAs's likely role in vehicles is in sensors, lidar lasers and RF links.
- Space exploration: as missions grow more ambitious, GaAs-based solar cells and radiation-tolerant electronics will remain important.
- Integration with silicon: growing or bonding GaAs layers on silicon could combine the strengths of both materials, reducing cost while keeping GaAs performance where it is needed.
Research also focuses on lowering manufacturing cost and defect density. Improvements in VGF crystal growth and MOCVD are making GaAs more competitive, and recovering gallium from production scrap and electronic waste could ease supply concerns.
FAQ
Why is GaAs faster than silicon?
Electrons in GaAs have a smaller effective mass, so they reach higher velocity for a given electric field. Electron mobility is about six times that of silicon, which gives transistors with higher gain at high frequency.
Why can GaAs emit light when silicon cannot?
GaAs has a direct bandgap: the lowest point of the conduction band sits directly above the highest point of the valence band, so an electron can recombine with a hole and release a photon. In silicon the two points are offset in momentum, so recombination needs a lattice vibration and mostly produces heat.
What is the difference between GaAs and GaN?
GaN has a bandgap of about 3.4 eV against 1.42 eV for GaAs, so it withstands far higher electric fields and runs at higher power density and temperature. GaAs has higher electron mobility and lower-cost mature processes. GaN suits power conversion and high-power RF; GaAs suits handset RF, infrared lasers and solar cells.
Is gallium arsenide toxic?
Solid GaAs in a finished device poses little risk in normal use, but GaAs dust and the arsenic compounds used in manufacturing, such as arsine gas, are hazardous. Fabs handle them under strict controls.
Your comments will be moderated before it appears here.