Technological Applications of Superconductivity

Superconductivity is the property some materials have of carrying an electric current with zero electrical resistance once they are cooled below a critical temperature. Superconductors also push magnetic fields out of their interior, which is called the Meissner effect. For most of the twentieth century this was a laboratory curiosity that needed liquid helium. Today it is the working core of hospital MRI scanners, the world's largest particle accelerator, the magnets of fusion experiments, a few urban power cables, the most sensitive magnetic sensors ever built, and the leading type of quantum computer.

This article covers where superconductivity is actually used, how each application depends on the physics, which materials are used and why, and where the technology still falls short. It separates working technology from hype, including the brief 2023 excitement over a claimed room-temperature superconductor called LK-99.

A short history of a cold discovery

The discovery and the first decades

On 8 April 1911, the Dutch physicist Heike Kamerlingh Onnes found that the electrical resistance of solid mercury vanished when he cooled it to 4.2 kelvin (K), the boiling point of liquid helium at atmospheric pressure. Onnes had been the first person to liquefy helium a few years earlier, and that is what made the experiment possible. By 1913, tin and lead had been found to superconduct as well, at 3.8 K and 7 K.

In 1933, Walther Meissner and Robert Ochsenfeld showed that a superconductor does more than conduct perfectly. It actively expels an applied magnetic field from its interior. That expulsion is the reason a small magnet can float above a cooled superconductor, the classic demonstration of the effect. In 1935, the brothers Fritz and Heinz London gave a theoretical account of the Meissner effect. In 1950, Ginzburg–Landau theory gave a powerful phenomenological description. In 1957, John Bardeen, Leon Cooper and Robert Schrieffer published what became known as BCS theory, which explained conventional superconductivity in terms of electrons pairing up through their interaction with the crystal lattice. BCS theory earned the 1972 Nobel Prize in Physics.

The materials that made engineering possible

Early superconductors were useless for magnets. Most pure elements are what physicists call Type I superconductors: they have a single critical magnetic field, and above it superconductivity is lost entirely. The breakthrough for engineering came with Type II superconductors, which have two critical fields. Between them, magnetic flux threads through the material in quantized vortices while the rest of the material stays superconducting. Almost all impure and compound superconductors are Type II, and they can tolerate much stronger magnetic fields.

Two results in the early 1960s set the course of the industry. In 1961, J. E. Kunzler and colleagues found that niobium–tin (Nb3Sn) could carry more than 100,000 amperes per square centimetre at 4.2 K in a magnetic field of 8.8 tesla. In 1962, Berlincourt and Hake found that ductile niobium–titanium (NbTi) alloys were usable up to about 10 tesla. Also in 1962, Brian Josephson predicted that a supercurrent could tunnel across a thin insulating barrier between two superconductors. This Josephson effect is the basis of superconducting electronics, sensors and quantum computers, and it earned Josephson a share of the 1973 Nobel Prize.

NbTi became the workhorse material because it is ductile. It can be drawn into fine filaments and embedded in copper to make wire, much like ordinary metal wire. Nb3Sn can carry more current at higher fields, but it is brittle and must usually be reacted with heat after the coil is wound, which makes it much harder to manufacture.

The high-temperature revolution

In 1986, Georg Bednorz and K. Alex Müller at IBM Zürich found superconductivity at about 35 K in a copper-oxide ceramic. This was surprising because ceramics are normally poor conductors. The next year Ching-Wu Chu's group raised the critical temperature to about 93 K with yttrium barium copper oxide (YBCO). Bednorz and Müller received the 1987 Nobel Prize in Physics, one of the fastest awards in the prize's history.

The 93 K figure mattered because it is above 77 K, the boiling point of liquid nitrogen. Superconductors that work above 77 K are called high-temperature superconductors (HTS). The word "high" is relative: 77 K is about −196 °C. But liquid nitrogen is cheap and can even be produced on site, while liquid helium is expensive, scarce and hard to handle. Bismuth strontium calcium copper oxide (BSCCO) followed, with the Bi2Sr2Ca2Cu3O10 variant superconducting at about 110 K. Mercury-based cuprates later passed 130 K at ambient pressure. Wikipedia's summary reports a record of about 133 K for a mercury barium calcium cuprate as of 2021, and cites a March 2026 report of about 151 K in a "pressure-quenched" mercury cuprate, a result that should be treated as new and still being evaluated.

Other families followed. Magnesium diboride (MgB2), found by Jun Akimitsu's group in 2001, superconducts below 39 K. Iron-based superconductors were found from the late 2000s. Hydrogen-rich compounds squeezed to enormous pressures have shown much higher critical temperatures, including 203 K for hydrogen sulfide at roughly 90 gigapascals, about 900,000 times atmospheric pressure. Those pressures exist only inside diamond anvil cells, so hydrides have no engineering use at present.

Key superconducting materials at a glance

Material	Type	Approximate critical temperature	Main uses
Niobium–titanium (NbTi)	Low-temperature, ductile alloy	about 9.3 K (under 10 K)	MRI magnets, LHC dipoles, ITER poloidal field coils, most commercial magnets
Niobium–tin (Nb3Sn)	Low-temperature, brittle compound	18.3 K	ITER toroidal field coils and central solenoid, HL-LHC quadrupoles, high-field NMR
Magnesium diboride (MgB2)	Conventional, cheap to make	39 K	Experimental cryogen-free magnets, wire
YBCO and other REBCO	Cuprate high-temperature	about 92–93 K	Coated-conductor tape, fusion magnets, NMR inserts, cables
BSCCO (Bi-2223)	Cuprate high-temperature	about 110 K	Early HTS wire, current leads, cables
Mercury cuprates	Cuprate high-temperature	over 130 K	Research only

How superconductivity becomes useful

Every application exploits one or more of four properties.

Zero DC resistance. A superconducting coil carrying a steady current dissipates no heat in the winding itself. A copper magnet producing a strong field over a large volume would need megawatts of power and heavy water cooling, and most of that power would turn into heat. A superconducting magnet, once charged, can run in "persistent mode" with the current circulating indefinitely. The power bill shifts to the refrigerator that keeps it cold.

High current density. Superconducting wire can carry far more current per unit cross-section than copper. That allows compact windings that produce fields of many tesla. For comparison, a fridge magnet is around a hundredth of a tesla.

Flux quantization and the Josephson effect. Magnetic flux through a superconducting loop is quantized in units of the flux quantum. Combined with Josephson junctions, this produces devices whose electrical behaviour changes with extremely small changes in magnetic flux. That is the basis of SQUID sensors, superconducting digital logic, and superconducting qubits.

Sharp transition to resistance. When a superconductor exceeds its critical current, field or temperature, it abruptly becomes resistive. Engineers usually fear this (it is called a quench), but fault current limiters use it on purpose.

Each property comes with three limits that shape every design: critical temperature, critical magnetic field, and critical current density. Push any one of them too far and superconductivity disappears. In practice, the higher the field and the current, the colder a given conductor must be. This is why the LHC runs its niobium–titanium magnets at 1.9 K instead of 4.2 K: the extra cooling lets the same well-proven conductor reach a higher field.

Medical imaging: MRI and NMR

MRI is the largest market

Magnetic resonance imaging is the biggest commercial application of superconductivity, in a market worth multiple billions of US dollars. The Wikipedia article on the subject names Oxford Instruments and Siemens among the major companies in the magnet business. A typical clinical scanner uses a large solenoid of NbTi wire bathed in liquid helium at about 4 K, producing a uniform field of 1.5 or 3 tesla across the bore where the patient lies.

MRI needs a field that is strong, extremely uniform over a large volume, and very stable over time. A persistent superconducting magnet fits that well. Once ramped up, the current circulates for years with negligible decay, and the field is always on. That is why MRI suites have strict rules about metal objects, and why the field does not switch off when the scanner is idle.

MRI magnets still overwhelmingly use low-temperature superconductors. High-temperature superconductors are not yet cheap enough to wind at the field strength and volume a whole-body scanner needs. NbTi is mature, reliable and comparatively inexpensive, and its main drawback is the need for liquid helium.

The helium problem and the move toward "dry" magnets

A conventional 1.5 T or 3 T NbTi magnet is typically quoted as holding about 1,200 to 2,000 litres of liquid helium. Helium is a finite resource extracted from natural gas, and supply disruptions have pushed up prices at various times. Manufacturers have responded with sealed and low-helium designs. Philips' BlueSeal 1.5 T magnet has been reported at about 7 litres of helium, and Siemens' DryCool 0.55 T system at about 0.7 litres. Some cryogen-free magnets use a Gifford–McMahon cryocooler with no liquid bath at all. Wikipedia notes one 1.5 T whole-body example. The trade-off is that a cryogen-free magnet cannot keep running as long during a power outage as one sitting in a large helium reservoir, because there is no stored cold to fall back on.

Quenches

If part of an MRI winding loses superconductivity, the stored magnetic energy turns into heat in that region. The heat spreads, more of the coil goes normal, and the helium boils off rapidly. Scanners have a quench pipe that vents the helium gas safely outside the building. A deliberate emergency quench is also how the field is shut down if a person is pinned against the magnet by a metal object. Quenches are rare in normal operation, but they are expensive: the helium has to be replaced and the magnet has to be re-energized.

NMR spectroscopy and ultra-high fields

Nuclear magnetic resonance spectroscopy uses the same physics as MRI to study molecules rather than bodies. Chemists and structural biologists want ever-higher fields because resolution and sensitivity improve with field strength. NMR magnets are labelled by the proton resonance frequency they produce. Bruker installed the first 1.1 GHz (25.8 T) NMR spectrometer at St. Jude Children's Research Hospital in 2019 and the first 1.2 GHz (28.2 T) system at the University of Florence, Italy, in 2020.

Those fields are beyond what NbTi and Nb3Sn can reach alone. The 1.2 GHz magnets are hybrids: low-temperature superconducting outer coils with an insert of REBCO high-temperature superconductor tape at the centre, still run in persistent mode. This "HTS insert" approach is one of the main ways high-temperature superconductors entered commercial instruments.

Research magnets go higher. The US National High Magnetic Field Laboratory's all-superconducting 32 T user magnet, which combines LTS and HTS coils, reached full field in 2017. Its report confirmed a peak of 32.103 T. A different all-superconducting magnet holds a reported record of 32.35 T.

Particle accelerators

The Large Hadron Collider

CERN's Large Hadron Collider (LHC) is the best-known superconducting machine. It uses 1,232 main dipole magnets to bend proton beams around its ring. Each dipole is designed for a field of about 8.3 tesla (CERN material gives 8.33 T) at a nominal current of 11,850 amperes. The coils are made of niobium–titanium cable and sit in a bath of superfluid helium at 1.9 K at atmospheric pressure. Operating at 1.9 K rather than 4.2 K is what allows the established NbTi technology to reach that field.

The dipoles use a "two-in-one" design: both beam pipes, carrying protons in opposite directions, share a single iron yoke and cryostat. This saves space and cost compared with two separate magnet rings.

The scale was enormous for the wire industry. Building the LHC magnets used more than 28% of the world's niobium–titanium wire production over five years. Accelerators have always been among the most demanding customers for superconducting wire, and their procurement shaped how the industry grew.

The High-Luminosity LHC and niobium–tin

The High-Luminosity LHC upgrade will replace the focusing magnets on either side of the main experiments with new quadrupoles made of Nb3Sn. CERN Courier reports that Nb3Sn allows fields of about 11.5 to 12 tesla, compared with roughly 8.5 T for the NbTi bending magnets elsewhere in the ring. That allows the magnet aperture to be doubled while the inner-triplet assembly grows only from 23 m to 32 m. The work is shared between CERN and a US collaboration (the HL-LHC Accelerator Upgrade Project), which supplies the shorter Q1 and Q3 cryo-assemblies, each containing two 4.2 m quadrupoles. CERN builds the longer Q2 versions, each with one 7.2 m quadrupole. CERN reports that installation in the tunnel is planned for 2028–2029, during Long Shutdown 3.

This is a milestone for the field. Nb3Sn has been known since the early 1960s, but its brittleness kept it out of accelerators until now.

Superconducting radio-frequency cavities

Accelerators also use superconductors to give particles energy, not just to steer them. Superconducting radio-frequency (SRF) cavities, usually made of pure niobium, build up strong oscillating electric fields with very low power loss. They are used in many modern linear accelerators and light sources. The physics differs slightly from DC magnets: at radio frequencies a superconductor has a small but nonzero surface resistance, which is still far lower than copper's.

Fusion energy

Why fusion needs superconducting magnets

A tokamak confines a plasma hotter than the core of the Sun inside a doughnut-shaped magnetic "bottle." The magnets must run continuously for long periods and produce strong fields over large volumes. Copper magnets can do this only for short pulses and at great power cost. Wikipedia notes that conventional copper tokamaks such as JET, MAST and NSTX-U are limited to about 1 to 3 tesla. For a power plant, superconducting magnets are the only realistic option.

ITER

ITER, the international fusion experiment being built in southern France, has the largest superconducting magnet system ever built. ITER gives these figures:

ITER magnet system	Number	Superconductor	Maximum field	Stored energy
Toroidal field coils	18 D-shaped coils, 330 tonnes each	Nb3Sn	11.8 T	41 GJ
Central solenoid	6 independent modules, 1,000 tonnes in total	Nb3Sn	13 T	6.4 GJ
Poloidal field coils	6 ring coils, largest 24 m across	NbTi	6 T	4 GJ
Correction coils	18	NbTi	—	—

The whole magnet system weighs about 10,000 tonnes and stores about 51 gigajoules of magnetic energy. It is cooled with supercritical helium to about 4 K. The conductor is "cable-in-conduit": superconducting strands mixed with copper, packed inside a steel jacket through which the helium flows. ITER bought about 500 tonnes of Nb3Sn strand, more than 100,000 km of it, from nine suppliers between 2008 and 2015. ITER reports that this pushed world production capacity from about 15 to 100 tonnes per year.

ITER also uses high-temperature superconductors in a quieter role. Its 31 magnet feeders use HTS current leads, the links that carry current from room-temperature power supplies into the cold magnets. HTS conducts heat poorly, so it lets current in while letting little heat leak along the lead. This is one of the most widespread practical uses of HTS.

HTS and compact tokamaks

The strength of a tokamak's magnetic field strongly affects how large the machine must be for a given performance. If the field doubles, the device can be much smaller. That is the case for high-temperature superconductors in fusion. REBCO tape keeps a high critical current in very strong fields, especially when cooled well below 77 K, to around 20 K.

On 5 September 2021, Commonwealth Fusion Systems and MIT's Plasma Science and Fusion Center tested a large D-shaped HTS magnet that held a field of more than 20 tesla in steady state for about five hours. The results were announced on 8 September. The magnet weighed about 10 tons and was built as a prototype for the toroidal field coils of SPARC, a compact tokamak the company was building in Devens, Massachusetts. A ring of 17 such magnets is intended to confine SPARC's plasma. Tokamak Energy, a British company, had reported 24 T from an HTS magnet in 2019, but at a smaller scale and not in a tokamak D-shape.

These magnet results are real engineering achievements. They do not prove that fusion power plants will be economical. Plasma performance, materials that survive neutron bombardment, tritium breeding and cost all remain unsolved at commercial scale. Timelines announced by fusion companies have often slipped, so readers should check current status rather than relying on earlier target dates.

Transportation: magnetic levitation

The Japanese SCMaglev

Japan's SCMaglev system uses superconducting magnets mounted on the train's bogies, which interact with coils in the guideway walls. The train runs on rubber wheels at low speed and lifts off to levitate about 100 millimetres above the guideway once it reaches about 150 km/h. The superconducting magnets produce much stronger fields than conventional electromagnets, which allows that large air gap. A large gap is more tolerant of guideway imperfections and earthquake movement.

On 21 April 2015, a seven-car L0 series train operated by Central Japan Railway Company reached 603 km/h on the Yamanashi test line. Guinness World Records lists this as the fastest maglev speed. It beat the same train's 590 km/h run a week earlier and the previous rail record of 581 km/h.

SCMaglev is being built for the Chūō Shinkansen between Tokyo and Nagoya, with an extension to Osaka planned. Planned service speed is reported as about 500 to 505 km/h. The line was originally meant to open in 2027 but has been delayed, with Wikipedia currently giving 2034 for Tokyo–Nagoya and 2037 for Osaka. The delays came from construction and permitting problems, not from the superconducting technology.

Other maglev approaches

Not all maglev trains use superconductors. The Shanghai Transrapid line, for example, uses conventional electromagnets that attract the train toward the guideway, held at a small, actively controlled gap. Superconducting maglev is a different design choice, not a requirement of magnetic levitation. Research groups have also built small demonstration vehicles that float on bulk HTS blocks above permanent-magnet tracks using flux pinning, but these remain demonstrations.

Ultra-sensitive sensors: SQUIDs

How a SQUID works

A superconducting quantum interference device (SQUID) is a superconducting loop interrupted by one or two Josephson junctions. Because magnetic flux through a superconducting loop is quantized, the current and voltage across the junctions vary periodically as the flux through the loop changes. With suitable electronics, a SQUID can detect a tiny fraction of one flux quantum. That makes it the most sensitive magnetic field detector in routine use.

The DC SQUID, with two junctions, was invented in 1964 at Ford Research Laboratories by Robert Jaklevic, John Lambe, James Mercereau and Arnold Silver. The RF SQUID, with a single junction, followed in 1967, with James Zimmerman joining the Ford team. DC SQUIDs are more sensitive. RF SQUIDs are cheaper and easier to make in small numbers.

Wikipedia reports that SQUIDs can detect fields as small as 5 × 10⁻¹⁸ tesla with several days of averaging, and that noise levels as low as 3 femtotesla per root hertz have been reported.

Magnetoencephalography

The firing of neurons in the brain produces magnetic fields measured in femtotesla. Published estimates for signals outside the head range from about 10 femtotesla to a few hundred femtotesla, depending on the source and where it is measured. Magnetoencephalography (MEG) uses arrays of SQUIDs in a helmet-shaped cryostat to map this activity with millisecond timing. It is used in epilepsy surgery planning and in neuroscience research. Because the signals are so small, MEG systems sit in magnetically shielded rooms.

The main competition is optically pumped atomic magnetometers, which do not need cryogenic cooling and can be placed closer to the scalp. They are an active research area, but SQUID-based MEG remains the established clinical technology.

Other SQUID uses

Heart imaging. SQUIDs can map the heart's magnetic field without contact.
Geophysics. SQUIDs are used in mineral exploration, oil prospecting and geothermal surveys.
Fundamental physics. NASA's Gravity Probe B, a test of general relativity, used a SQUID to read out each of its four gyroscopes.
Materials science. SQUID magnetometers are standard laboratory instruments for measuring the magnetic properties of samples.
Astronomy and particle detection. SQUIDs read out arrays of superconducting detectors, such as transition-edge sensors, used in telescopes and particle physics.

High-temperature SQUIDs made from YBCO and cooled with liquid nitrogen were developed from the late 1980s. They are less sensitive than niobium devices at 4 K but are cheaper and easier to run.

Quantum computing

Superconducting qubits

The leading type of quantum computer by number of qubits built uses superconducting circuits. A superconducting qubit is a tiny electrical circuit patterned on a chip that contains one or more Josephson junctions. The junction acts as a nonlinear inductor with no energy loss. That turns the circuit into an artificial atom with discrete energy levels, two of which are used as the qubit's 0 and 1. The most common design is called the transmon. The chips run at millikelvin temperatures inside dilution refrigerators, much colder than an MRI magnet, because heat and stray radiation easily disturb the fragile quantum states.

Superconductivity matters here for two reasons. Zero resistance means the circuit does not dissipate energy and destroy its quantum state. The Josephson junction provides the nonlinearity that makes a controllable two-level system possible.

The state of the art

In December 2024, Google Quantum AI announced its Willow chip, with 105 superconducting qubits. In a Nature paper titled "Quantum error correction below the surface code threshold," the team stored a single logical qubit in square grids of physical qubits of increasing size: 3×3, 5×5 and 7×7. Each step up roughly halved the logical error rate. Earlier experiments had often seen errors grow as more qubits were added. Showing that errors fall as the code grows is a necessary step toward fault-tolerant quantum computing. Google also reported that Willow's qubits hold their quantum states about five times longer than those of its earlier Sycamore processor.

Context is important. A logical qubit built from dozens of physical qubits is still a long way from the thousands of high-quality logical qubits needed for many proposed useful algorithms. Press coverage has cited estimates of around 1,000 physical qubits per logical qubit at useful error rates. Superconducting qubits also compete with trapped ions, neutral atoms and photonic approaches, each with different strengths.

Electric power systems

Superconductors could, in principle, transform how electricity is moved and managed. In practice, grid applications have been the hardest to make pay for themselves. Copper and aluminium conductors are cheap and familiar, and a superconducting system needs a cryogenic plant that must keep running reliably for decades.

Superconducting power cables

A superconducting cable carries current in HTS tapes cooled by liquid nitrogen flowing through the cable. Its main advantage is power density. A superconducting cable can carry the same power as a much larger conventional cable, or carry it at a lower voltage. That matters in crowded city centres where digging new tunnels is very expensive.

Notable projects include:

Long Island, New York (2008). The Long Island Power Authority commissioned a roughly 600-metre underground HTS cable at its Holbrook substation in late June 2008, using wire from American Superconductor and liquid-nitrogen cooling. It operated for about two years. Wikipedia flags parts of the project data as disputed or out of date.
AmpaCity, Essen, Germany (2014). A 1-kilometre, 10-kilovolt, three-phase concentric HTS cable rated at 40 MVA links two substations in Essen's city centre. It was built by Nexans with the utility RWE and the Karlsruhe Institute of Technology, with funding from Germany's federal economics ministry. Nexans describes it as the world's longest superconducting cable link at the time. It operated for seven years. The system also included a resistive superconducting fault current limiter. A key point was that a 10 kV superconducting cable could replace a 110 kV conventional line, which removes the need for some high-voltage transformers and substations in the city.
Industrial direct current. Wikipedia reports that in 2020 an aluminium plant in Voerde, Germany, announced plans to use superconducting cables carrying 200 kiloamperes. Aluminium smelting uses huge direct currents, which makes it a natural early market.
Tres Amigas (proposed, not built). Proposed in 2009 as a superconducting DC hub joining the three main US power grids with 5 gigawatts of capacity, the project lapsed in 2015 without construction starting.

The economics come down to up-front cost versus lifetime savings. HTS tape is expensive, and the cryogenic system adds cost and maintenance. The savings come from lower losses, smaller rights-of-way and avoided substations. Supporters expect costs to converge over time. City utilities have so far found the start-up costs hard to justify except in specific cases.

Superconducting fault current limiters

When a short circuit happens on a power grid, current can spike to many times its normal value within milliseconds. Circuit breakers and other equipment must be rated to survive that surge. As grids add more generation and interconnections, potential fault currents rise, which can force expensive equipment upgrades.

A superconducting fault current limiter (SFCL) is a device that is invisible during normal operation and becomes a large resistance during a fault. In the resistive design, normal current flows through a superconducting element with essentially no loss. When a fault pushes the current past the element's critical current, it quenches and its resistance jumps within milliseconds, limiting the surge until a breaker can open. The device then has to cool back down before it is ready again. Other designs use superconductors to control the saturation of an iron core, changing the device's inductance during a fault.

SFCLs have been demonstrated in field installations, including the AmpaCity project. Wikipedia describes the technology as still under development rather than in routine use, and recovery time after a fault, cost and cryogenic reliability remain the main hurdles.

Superconducting magnetic energy storage

Superconducting magnetic energy storage (SMES) stores energy in the magnetic field of a large superconducting coil. Because the coil has no resistance, the stored energy stays put until it is needed, and it can be released very fast. That makes SMES attractive for power quality and very short-term grid stabilisation, not bulk storage. SMES coils typically use low-temperature superconductors, usually NbTi, because LTS can handle higher fields than HTS at affordable cost. The amount of energy SMES can economically store is small compared with batteries or pumped hydro, so it has remained a niche technology.

Motors and generators

Superconducting windings let a motor or generator produce a strong magnetic field in a smaller, lighter machine. That is attractive where size and weight matter, such as ship propulsion and offshore wind turbines, where a lighter generator at the top of the tower reduces the cost of the whole structure. NbTi coils cooled by liquid helium have been used in some generator designs, including offshore wind turbines, and Wikipedia cites a 2025 paper describing a 20 MW low-temperature superconducting wind generator concept. These machines remain demonstration projects or concepts. The main barriers are reliable cryogenics on a rotating machine and cost compared with permanent-magnet generators.

Other industrial and scientific uses

Magnetic separation. Superconducting magnets can pull weakly magnetic particles out of a flow of material. Wikipedia notes their use in coal purification, where a superconducting electromagnet iron separator using liquid-helium-cooled LTS coils removes iron and other ferromagnetic material. Similar high-gradient magnetic separators have been used for clay purification and water treatment.

Trapped-field magnets. A bulk superconductor, such as a disc of YBCO, can "trap" a magnetic field after brief exposure to it and then act like a permanent magnet while it stays cold. Magnetized bulk HTS can hold fields stronger than the best rare-earth permanent magnets. Uses being explored include compact NMR, motors and magnetic bearings.

Superconducting electronics. Josephson junctions can switch extremely fast with very low energy. Rapid single flux quantum (RSFQ) logic represents bits as single quanta of magnetic flux and has been demonstrated at very high clock rates. Voltage standards based on arrays of Josephson junctions define the volt in national metrology laboratories, because the Josephson effect links voltage directly to frequency and fundamental constants. Superconducting digital electronics have not displaced silicon for general computing, mainly because of cooling requirements and the scale of the silicon industry.

RF and microwave filters. HTS thin films have very low losses at microwave frequencies. That allows filters with sharp cut-offs, which have been used in some mobile phone base stations to reject interference.

Particle detectors. Large particle physics experiments use superconducting magnets to bend charged particles so their momentum can be measured. The CMS experiment at the LHC is named for its large superconducting solenoid ("Compact Muon Solenoid").

Low-temperature versus high-temperature superconductors in practice

It might seem that high-temperature superconductors should have replaced the old helium-cooled materials by now. They have not, for several reasons.

Consideration	Low-temperature (NbTi, Nb3Sn)	High-temperature (REBCO, BSCCO)
Cooling	Liquid helium or cryocoolers near 4 K	Liquid nitrogen at 77 K, or colder (around 20 K) for high fields
Conductor form	Round multifilament wire; NbTi is ductile	Mostly flat tape (REBCO coated conductor) or powder-in-tube (BSCCO); ceramics are brittle
Cost per amp-metre	Low and mature for NbTi	Much higher, though falling
Performance in very high field	Limited (NbTi around 10 T, Nb3Sn somewhat higher)	Excellent, especially at low temperature
Maturity	Decades of industrial use	Younger, fewer suppliers, less long-term data
Typical uses	MRI, LHC, ITER, most lab magnets	HTS current leads, NMR inserts, compact fusion magnets, cables, fault current limiters

High-temperature superconductors win where they offer something low-temperature conductors cannot. Examples are very high fields (NMR inserts and compact fusion magnets), current leads that must not leak heat, and devices that must run on liquid nitrogen (cables, some SQUIDs and filters). Elsewhere, cheap and reliable NbTi still dominates.

The cuprate superconductors are brittle ceramics that are costly to make and hard to form into wire. Modern REBCO tape solves this by depositing a thin, highly aligned superconducting layer on a flexible metal substrate. That is an elaborate, multi-layer manufacturing process. The crystal grains must be closely aligned because current flows poorly across misaligned grain boundaries in cuprates. That is a major reason HTS tape is expensive.

Magnesium diboride sits between the two groups. With a critical temperature of 39 K it can run at 10 to 20 K in cryogen-free magnets cooled by cryocoolers. Its raw materials are cheap, and Wikipedia reports that its cost per unit of current-carrying capacity is in the LTS range and far below BSCCO and YBCO. Its tolerance of strong magnetic fields is limited at those temperatures, so it suits lower-field magnets such as some MRI designs.

Limits and practical challenges

Cryogenics

Every superconducting system is really two systems: the superconductor and the refrigerator that keeps it cold. Refrigeration near 4 K is inherently inefficient: thermodynamics dictates that removing each watt of heat at liquid-helium temperature costs many watts of electrical power at room temperature, and the colder the target, the worse the ratio. Cryostats must be carefully designed to limit heat leaking in through supports, wiring and radiation. Cryogenic failures, not the superconductor itself, are a frequent cause of downtime.

Helium supply

Liquid helium is a by-product of certain natural gas fields and is concentrated in a few countries. Shortages and price spikes have hit hospitals and laboratories several times. This has pushed MRI makers toward sealed, low-helium magnets and is one reason HTS and MgB2 are attractive.

Quench protection

A large superconducting magnet stores a lot of energy: 51 gigajoules for ITER's full system, roughly the chemical energy in about 12 tonnes of TNT (using the standard figure of 4.184 GJ per tonne of TNT). If superconductivity is lost in one spot, that energy must be safely removed before it overheats the conductor. Magnet designers use quench detection, protection heaters that spread the quench evenly, and external dump resistors. HTS magnets bring a new challenge: quenches spread slowly in HTS, so a hot spot can form and burn out the conductor before the problem is detected.

Mechanical forces

Strong magnetic fields push on the very currents that create them. In large magnets the forces reach millions of newtons. ITER's central solenoid support structure is designed for about 60 meganewtons. Brittle conductors such as Nb3Sn and REBCO can lose performance under strain, so structural design is as important as the superconductor.

Alternating current losses

Zero resistance applies strictly to direct current. With alternating current, superconductors still have small losses from the movement of magnetic flux vortices. For power cables and AC machines these losses are much smaller than in copper but must still be removed by the cryogenic system, which is expensive at low temperatures.

Misconceptions

"Superconductors are perfect conductors and nothing more." Superconductivity is a separate state of matter, not just very low resistance. The Meissner effect, the expulsion of magnetic fields, does not follow from perfect conductivity alone.

"High-temperature superconductors work at room temperature." No. "High temperature" means above 77 K, about −196 °C. All practical superconductors still need serious cooling.

"Superconducting cables would end grid losses." Transmission and distribution losses are a modest share of the electricity generated, and a superconducting cable must power its own cryogenic plant. Superconducting cables make sense where space, power density or avoided infrastructure is the main issue, not as a blanket replacement for overhead lines.

"Maglev requires superconductors." Some maglev systems, including Shanghai's, use conventional electromagnets. Superconducting maglev is one option among several.

"Superconductors float magnets because they repel them." In Type II materials, levitation demonstrations often rely on flux pinning, which locks the magnet in position, rather than pure repulsion. That is why a pinned magnet can hang below a superconductor as well as float above it.

"A room-temperature superconductor has been found." As of this writing, no claim of a room-temperature superconductor at ambient pressure has survived independent replication. The most widely publicised recent case is described below.

The LK-99 episode

In July 2023, a claim that a material called LK-99 was a superconductor at room temperature and ambient pressure spread rapidly online. The material had first been described in April 2023 in the Journal of the Korean Crystal Growth and Crystal Technology. On 22 July 2023, two preprints were posted to the arXiv server by Sukbae Lee, Ji-Hoon Kim and colleagues in Korea and the US. Hyun-Tak Kim of the College of William and Mary was the corresponding author of the second. Neither preprint had been peer reviewed. The material is a copper-doped lead apatite with the formula Pb10−xCux(PO4)6O, where x is between 0.9 and 1.1. The name comes from the discoverers' initials.

The authors reported a drop in resistance, partial levitation and other signatures they attributed to superconductivity. Laboratories around the world tried to reproduce the results within weeks. Physics World's August 2023 summary described the outcome:

A group at the Beijing National Laboratory for Condensed Matter Physics argued that copper sulfide (Cu2S) impurities explain the sharp resistivity drop near 400 K. Cu2S undergoes a structural phase transition at about 385 K. The Korean team acknowledged Cu2S was present in its samples.
A Peking University group led by Shuang Jia showed that ferromagnetism could explain the partial levitation without superconductivity.
Groups at Beihang University, India's CSIR–National Physical Laboratory, Princeton University and Charles University in Prague reported samples that did not superconduct. One Beihang sample behaved like a semiconductor with high resistance at room temperature.

In December 2023, a verification committee of the Korean Society of Superconductivity and Cryogenics reported that eight research teams, including one at Seoul National University, had found no superconductivity at room temperature in LK-99. The committee also noted that the original group had not provided samples for cross-checking.

The LK-99 episode was a fast, public demonstration of how replication works. It also showed why a single preprint, however exciting, is not evidence that a technology exists. None of the applications in this article depend on room-temperature superconductivity. They all work with materials that have been studied, replicated and manufactured for decades.

Frequently asked questions

What is the most common everyday use of superconductivity? MRI scanners. Most hospital MRI machines contain a superconducting NbTi magnet cooled by liquid helium. It is the largest commercial market for superconducting wire.

Why do MRI machines need helium? The NbTi wire in most MRI magnets becomes superconducting only below about 9 K. Liquid helium boils at 4.2 K, so a bath of it keeps the magnet cold enough with a safety margin. Newer sealed and cryogen-free designs use far less helium or none.

What temperature does the LHC run at? The main LHC dipole magnets run at 1.9 K in superfluid helium.

Are high-temperature superconductors actually used anywhere? Yes, in specific roles: current leads that feed current into cold magnets (including at ITER), inserts in the highest-field NMR magnets, some power cable demonstrations, fault current limiters, some SQUIDs and RF filters, and prototype fusion magnets such as the 20 T magnet tested in 2021.

Why not make all power lines superconducting? Cost and complexity. A superconducting line needs continuous cryogenic cooling, which costs money and energy and must be extremely reliable. For most routes, conventional overhead lines are far cheaper. Superconducting cables are being used where space is tight or where they let a utility avoid building high-voltage substations.

How do superconducting quantum computers relate to MRI magnets? They use the same phenomenon in very different ways. MRI uses large amounts of superconducting wire to carry big currents. Quantum computers use tiny superconducting circuits with Josephson junctions to make controllable quantum states. Quantum chips also run much colder, at millikelvin temperatures.

Is fusion power close because of superconducting magnets? HTS magnets have made compact, high-field tokamaks plausible, and large magnets like the 20 T CFS/MIT prototype have been built and tested. But magnets are only one of several hard problems. Net electricity from fusion on a commercial basis has not yet been demonstrated.

What happens if a superconducting magnet quenches? Part of the coil becomes resistive, the stored magnetic energy turns into heat, and any liquid helium boils off quickly. Well-designed magnets detect this and dump the energy safely. MRI rooms have vent pipes to carry helium gas outside. A quench usually means downtime and the cost of replacing helium rather than lasting damage, though serious faults can damage magnets.

Was LK-99 a superconductor? The evidence says no. Replication attempts found that its resistance drop came from copper sulfide impurities and its partial levitation from ferromagnetism. A Korean expert committee reported in December 2023 that eight teams had found no room-temperature superconductivity.

What would a true room-temperature superconductor change? Removing the need for cryogenics would make many applications far cheaper, from MRI to power cables. But a useful conductor also needs a high critical current, tolerance of magnetic fields and a material that can be made into wire at reasonable cost. Critical temperature alone is not enough.

Summary

Superconductivity has moved from a strange result in a Leiden laboratory in 1911 to a working technology in hospitals, accelerators and fusion experiments. Its main commercial value today is in magnets: MRI and NMR, particle accelerators like the LHC, and fusion devices like ITER, nearly all built from niobium–titanium and niobium–tin cooled by helium. Josephson junctions add a second family of uses in SQUID sensors, voltage standards and superconducting quantum computers. High-temperature superconductors, discovered in 1986, have found places where their high-field performance or liquid-nitrogen operation is worth their cost, and REBCO tape has made compact high-field fusion magnets possible. Grid applications such as cables, fault current limiters and energy storage work technically but still struggle on cost. Through all of this, the main constraint has not changed: superconductors must be kept cold, and the cost and reliability of that cooling decide which applications succeed.