Error fields

Overview

In magnetic confinement fusion, an error field (EF) is any deviation from the intended, perfectly symmetric magnetic field geometry. These small, spatially varying magnetic perturbations are an unavoidable consequence of engineering tolerances in the construction and assembly of complex devices like tokamaks and stellarators. Sources include minute variations in the shape and position of magnetic coils, electrical leads and busbars, and the presence of nearby ferromagnetic materials.

Although the magnitude of these fields is typically small, often on the order of 10⁻⁴ of the main toroidal field (δB/B₀ ≈ 10⁻⁴), their impact on plasma performance can be severe. Error fields can apply electromagnetic torques to the plasma, slowing its rotation. They can also directly drive the growth of magnetohydrodynamic (MHD) instabilities by resonating with rational magnetic flux surfaces. This process can lead to the formation of magnetic islands, which degrade thermal confinement, and can ultimately trigger a locked mode that grows to cause a major plasma disruption. Consequently, the characterization, control, and correction of error fields are critical for achieving stable, high-performance operation in current and future fusion devices, including ITER.

Physics / Mechanism

The detrimental effects of error fields are primarily linked to their interaction with rational magnetic surfaces, where the magnetic field lines close on themselves after a rational number of turns. The safety factor, q, defines the pitch of the field lines, and these rational surfaces exist at locations where q = m/n, with m and n being the poloidal and toroidal mode numbers, respectively.

An external, non-axisymmetric error field can be decomposed into a spectrum of Fourier components, each with a specific (m, n) helicity. When a component of the error field has the same helicity as a rational surface within the plasma, it is termed a resonant magnetic perturbation. This resonant field can induce magnetic reconnection, leading to the formation of a static magnetic island at that surface. These islands break the nested magnetic flux surfaces, creating a shortcut for heat and particles to escape from the core to the edge, thereby degrading confinement.

Furthermore, error fields exert an electromagnetic torque on the plasma. In a typical tokamak, the plasma rotates toroidally due to external momentum injection (e.g., from neutral beams) or intrinsic effects. Naturally occurring, rotating MHD instabilities like tearing modes can also exist. The external, static error field applies a braking torque on these rotating structures. If the error field is large enough, or if the plasma rotation is slow enough, this torque can overcome the plasma's inertia and lock the mode to the wall frame. A locked tearing mode often grows uncontrollably, leading to a thermal quench and a disruptive termination of the plasma discharge. The threshold for this locking is particularly low at low plasma densities, establishing an operational limit known as the locked mode density limit.

In high-performance plasmas, error fields can also amplify otherwise stable modes, such as the resistive wall mode (RWM), by providing a non-ideal boundary condition that the mode can couple to. This effect, known as resonant field amplification (RFA), is a key diagnostic tool for understanding plasma stability and the proximity to no-wall beta limits.

Historical development

The significance of error fields became increasingly apparent in the early 1990s. Experiments on the COMPASS-C and DIII-D tokamaks systematically demonstrated a strong correlation between the magnitude of specific error field components and the onset of locked modes, particularly at low plasma densities. A key 1992 study by J.T. Scoville et al. on DIII-D established a scaling law for the critical error field strength required to cause a disruption, showing its dependence on plasma density and toroidal field strength. These findings transformed error fields from a suspected nuisance into a well-defined engineering and physics challenge for future devices.

This understanding spurred the development of error field correction coils (EFCCs). Early experiments on devices like DIII-D and JET involved placing large, simple coils around the vacuum vessel to produce a magnetic field that would cancel the dominant, low-n error field components (e.g., n=1 and n=2). These efforts were highly successful, demonstrating a significant expansion of the operational space to lower densities and higher performance. The development of more sophisticated feedback control systems and a deeper understanding of the plasma's response, including the RFA effect, further refined these techniques throughout the 2000s. The work of physicists like /scientists/rob-la-haye was instrumental in developing the physics basis for mode locking and its mitigation.

For the design of ITER, error field control was identified as a first-order requirement from the outset. Extensive modeling and analysis, benchmarked against experiments on existing machines, were used to set stringent tolerances on coil manufacturing and alignment (typically <1 mm). This analysis also drove the design of a dedicated set of 27 correction coils (6 upper, 9 mid-plane, 12 lower) to provide robust control over the expected intrinsic error fields and enable advanced operational scenarios.

Current status

As of 2026, error field correction is a standard and essential technique on all major tokamaks, including DIII-D, JET, ASDEX Upgrade, and KSTAR. The state of the art has moved beyond simple cancellation of intrinsic, or 'vacuum', error fields. Modern approaches account for the plasma's own magnetic response, which can amplify the applied field. The RFA phenomenon is now routinely used as a diagnostic to measure the plasma's stability to external perturbations.

Control schemes have become highly sophisticated. Feed-forward systems use pre-programmed waveforms in the EFCCs based on extensive experimental databases and models of known error sources. These are often combined with real-time feedback systems that use magnetic sensors to detect the growth of a locked mode and dynamically adjust the correction currents to suppress it. The focus has shifted from merely avoiding disruptions to optimizing confinement and stability in high-beta, high-confinement (H-mode) scenarios, where even small residual error fields can degrade performance.

Research also focuses on understanding the full 3D field effects, including the impact of ferritic steel components like the ITER Test Blanket Modules (TBMs), which can introduce significant localized error fields. Experiments and modeling are underway to develop control strategies for these complex, dynamic error sources.

Notable implementations

• ITER: The International Thermonuclear Experimental Reactor has the most advanced and comprehensive error field control system planned. Its design specifications for coil tolerances are among the tightest ever attempted for a large-scale magnet system. The in-vessel EFCCs are designed to correct the n=1, 2, 3, and 4 error field harmonics with high fidelity, a necessity for achieving its goal of Q_plasma ≥ 10. The design is based on decades of research from the global fusion community.
• DIII-D National Fusion Facility: Operated by General Atomics, DIII-D has been a world leader in error field research. Its flexible set of 12 external and 6 internal control coils (the 'I-coils') has been instrumental in pioneering many of the foundational concepts of EF correction, RFA diagnostics, and the use of 3D fields for controlling edge localized modes (ELMs).
• JET (Joint European Torus): As the largest operating tokamak for many years, JET's work on error fields has been crucial for validating models and scaling laws to ITER-relevant sizes and parameters. Its Error Field Correction Coils were vital for accessing and sustaining high-performance scenarios, particularly with its ITER-like wall of beryllium and tungsten.
• ASDEX Upgrade: This facility at the Max Planck Institute for Plasma Physics in Germany has made significant contributions to understanding the physics of mode penetration and the plasma response to 3D fields, providing key data for benchmarking models used to predict ITER's error field correction requirements.

Open challenges

Despite significant progress, several challenges remain in the understanding and control of error fields.

1. Dynamic Error Fields: While static, intrinsic errors are well-understood, dynamic error fields created by events like minor disruptions, ELMs, or the movement of mechanical structures are more difficult to characterize and correct in real time.
2. Plasma Response Modeling: Accurately predicting the plasma's response to an applied 3D field remains a complex computational problem. While linear models work well in many regimes, fully non-linear, multi-fluid MHD models are needed for a complete predictive capability, especially in scenarios with strong plasma flows and complex magnetic topology.
3. Error Field Metrology: Precisely measuring the error field sources in a fully assembled, cryogenically cooled tokamak is extremely difficult. Current methods rely on a combination of Hall probe measurements during construction and 'compass scans' (mapping the field with a low-strength toroidal field before operation), but uncertainties remain.
4. Stellarator Optimization: While this article focuses on tokamaks, stellarators are intrinsically 3D devices where the concept of an 'error' is more complex. Ensuring the constructed 3D field matches the design with sufficient accuracy to maintain good confinement is a major engineering challenge, as demonstrated by the meticulous assembly of the Wendelstein 7-X stellarator.

Outlook

Over the next 5-15 years, the field will focus on integrating advanced error field control into the operational scenarios of next-generation devices. For ITER, the commissioning and initial operation phases will provide the first full-scale test of the extensive modeling and design efforts. The successful correction of its intrinsic error fields will be a critical early milestone.

Research will increasingly leverage machine learning and artificial intelligence techniques to develop more sophisticated, real-time feedback control systems. These systems will be capable of identifying and suppressing instabilities driven by complex, dynamic error fields, moving beyond pre-programmed corrections. Furthermore, the deliberate use of 3D magnetic fields, which began with error field correction, will continue to evolve as a primary tool for controlling other plasma phenomena, most notably for the suppression of ELMs.

Ultimately, the lessons learned from controlling error fields on ITER and other devices will be essential for the design of a demonstration power plant (DEMO). A successful fusion power plant will require extremely high reliability and availability, making robust, automated control of all non-axisymmetric field perturbations a non-negotiable engineering requirement.

References

1. Error field mode-locking in tokamaks — Physics of Plasmas (1999)
2. Requirements on error fields for ITER — Nuclear Fusion (2005)
3. Error field experiments in DIII-D — Nuclear Fusion (1992)
4. Plasma response to non-axisymmetric magnetic fields in tokamaks — Nuclear Fusion (2011)
5. ITER Physics Basis Editors, Chapter 3: MHD stability, operational limits and disruptions — Nuclear Fusion (1999)
6. Correction of error fields in existing tokamaks and ITER — Fusion Engineering and Design (2001)
7. ITER error field control — ITER Organization (2011)
8. Resistive wall mode physics in tokamaks — Physics of Plasmas (2004)