Higher magnetic fields depend on materials, mechanical restraint and safe energy release. From CERN’s accelerator magnets to fusion coils, recent tests expose the engineering behind a usable superconducting system.
A field record needs a location
A superconducting magnet can carry an enormous electrical current while its conductor remains in the superconducting state. That capability supports particle accelerators, magnetic resonance instruments and fusion experiments. Yet the headline field strength does not tell a prospective user how much space is available, how uniform the field is or what happens when the current can no longer flow without resistance.
Comparisons also need a location. Peak field on the conductor, field at the centre of a coil and the field experienced by a plasma or sample are different quantities. A compact insert operating inside another magnet cannot be compared directly with a complete large-aperture system using only the largest number of teslas reported. Aperture, temperature, geometry and duty cycle belong beside that number.
For Europe’s magnet industry, these distinctions determine what must actually be manufactured and guaranteed. Improving a superconducting strand is valuable, but converting it into an instrument introduces cable joints, supports, cooling, power supplies and protection. Recent programmes show how each can become the limiting factor.
Stronger conductors demand different assembly
CERN’s move from niobium–titanium to niobium–tin for its new final-focusing quadrupoles illustrates the mechanical challenge. Niobium–tin offers access to higher fields but is brittle after reaction. In December 2023, CERN reported that its third full-length 7.2-metre prototype reached the target current at both 1.9 and 4.5 kelvin. Earlier prototypes had fallen short at the higher test temperature.
The improvement followed a production pause, changes to the outer shell, reduced peak stress during assembly and changes to coil manufacturing parameters. This was not simply a better piece of superconducting wire. The result depended on how the coil and its surrounding structure behaved together.
That history argues for preserving manufacturing detail alongside electromagnetic test results. Winding, heat treatment, dimensional inspection and assembly loading are parts of the performance specification. A successful material substitution can require a different production process, even when the finished magnet is intended to perform the same broad function.
Integrated tests expose the stored energy
In July 2026, CERN reported that the new quadrupoles at its Inner Triplet String facility reached their nominal 16,230-ampere current without further quenches. The magnets had already been trained individually; the result demonstrated retention of that trained performance in the integrated test arrangement. Their aperture is 150 millimetres, compared with 70 millimetres for the earlier generation, and CERN reports an operating field of 11.3 teslas.
A quench is a transition out of the superconducting state. Local heating can then become dangerous unless the stored magnetic energy is managed. In the same commissioning account, CERN reported that protection systems safely handled up to 38 megajoules. Analysis extended across powering, cryogenics, vacuum, controls and alignment as well as the magnets themselves.
The significance is broader than reaching a target current once. An accelerator needs reproducible behaviour through repeated cycles and interventions. Commissioning evidence should therefore distinguish an individual magnet’s test, the integrated circuit’s response and eventual accelerator operation. The July milestone establishes the second of those; it does not represent years of operating experience in the upgraded collider.
High temperature does not mean room temperature
Rare-earth barium copper oxide, usually shortened to REBCO, is central to many high-field magnet programmes. Calling it a high-temperature superconductor describes its superconducting properties relative to conventional materials. High-field applications can still require substantial cryogenic infrastructure, as the SPARC toroidal field model coil programme demonstrates.
The MIT and Commonwealth Fusion Systems programme reported 20.1 teslas as the peak field on the conductor, at 40.5 kiloamperes. Its structural case accommodated mechanical stress approaching one gigapascal. The associated test system provided 600 watts of cooling at 20 kelvin. Those results belong to a representative-scale model coil and its test facility, rather than a complete operating fusion power station.
The programme also assessed passive quench protection in a coil without conventional turn-to-turn insulation. The paper reports localised thermal damage during an intentionally induced event, while explaining that the model coil was not optimised for quench resilience. This qualification matters: a successful high-field demonstration and a protection challenge can coexist in the same programme.
For a buyer or research partner, the sensible request is a documented operating envelope. How much margin remains at the specified temperature and field? Which fault conditions were tested? How was damage assessed afterwards? Answering those questions is more informative than describing a magnet technology as inherently safe or ready for every application.
Small coils reveal large mechanical problems
A 2024 study of the Little Big Coil programme makes the distinction between field generation and durability particularly clear. Researchers tested a REBCO insert in a high background field. Their fourth coil reached a combined field of 44 teslas without a quench, although dissipation was present; a later ramp produced a quench at 43.5 teslas associated with excessive heating in a joint.
The researchers examined conductor damage and the way critical current varied across the tape’s width. They found that the orientation of slit tape edges and the resulting current distribution mattered to screening-current stresses. Their fourth coil showed substantially less characteristic wavy damage than its predecessor. The finding concerns this conductor and coil arrangement; it is not a universal prescription for every REBCO winding.
Such studies give manufacturers a reason to characterise spatial variation, rather than rely solely on an average current rating. The engineering inference is that incoming material inspection, winding orientation and mechanical modelling need compatible assumptions. A model can only use the properties that its designers know to measure.
Volume and operating purpose change the comparison
ITER demonstrates a different industrial scale. Its official magnet description records procurement of 500 tonnes of niobium–tin strand from nine suppliers between 2008 and 2015. The 18 toroidal field coils are designed for a maximum field of 11.8 teslas and a combined stored magnetic energy of 41 gigajoules. These are design characteristics of a large confinement system, not evidence that a higher-field compact coil is automatically a substitute.
At the other end of the specification, CERN’s SHiP magnet development targets a moderate field across a large volume. A one-metre magnesium diboride demonstrator coil passed operating tests at 20–30 kelvin, CERN reported in 2024. The planned application needs approximately 0.5 tesla. CERN presented large potential power savings compared with a resistive magnet, while explicitly noting that further development remained.
The commercial lesson is application selection. Maximum field, accessible volume, operating hours and refrigeration demand can favour different conductors and architectures. Any energy comparison should include the cooling system and realistic duty cycle; eliminating resistive heating in the winding does not make the entire installation consume no electricity.
Procurement can reflect those differences by separating material qualification from acceptance of the delivered magnet. A strand manufacturer can certify properties under specified test conditions; the system supplier must account for joints, assembly and thermal performance in the intended configuration. Research customers should also establish who owns the test records, how a replacement part would be qualified and which operating limits apply after a repair. A magnet may remain scientifically useful long after the team that built it has changed. Maintenance then depends on accessible records that connect the manufactured hardware to its verified operating conditions.
A usable field is an instrument capability
CEA’s Iseult programme offers a final reminder that the magnet is part of a larger instrument. CEA released its first in vivo human brain images at 11.7 teslas in April 2024, following years of magnet and imaging development. The images establish an imaging capability under the study conditions; they do not by themselves demonstrate improved clinical outcomes or routine hospital deployment.
For magnet developers, the endpoint should be expressed in the user’s terms: a stable field over an agreed volume, reproducible ramping, acceptable recovery after faults and manageable operating requirements. Field strength remains fundamental, but the magnet becomes useful when that strength is delivered reliably within a complete engineering system.
