Laser-Driven Capacitor Coils

• Laser-driven capacitor coils are targets where laser irradiation releases hot electrons to charge a capacitor-like structure, inducing a transient current in a sub-millimeter coil.
• They employ diverse geometrical designs and equivalent-circuit models (RL, underdamped or overdamped RLC) to tailor magnetic field strength, rise time, and spatial profile.
• Advanced diagnostics such as proton radiography and Faraday rotation are used to measure the generated magnetic fields, though separating electric and magnetic contributions remains challenging.

Searching arXiv for recent and foundational work on laser-driven capacitor-coil targets. Laser-driven capacitor coils are laser-irradiated targets in which laser-produced hot-electron escape charges a metallic structure and drives a transient discharge current through a sub-millimeter coil, thereby generating strong magnetostatic or quasi-static magnetic fields in an open geometry. In the canonical realization, two metallic plates or disks are connected by a single-turn or double-U coil; in related realizations, a foil-coil or simple welded loop provides the current path. Reported fields include $40\,\mathrm{T}$ with $>100\,\mathrm{ps}$ lifetime in a femtosecond-driven Ni capacity-coil, $600$–$800\,\mathrm{T}$ in nanosecond-driven Cu/Ni capacitor-coils, and $250\pm 30\,\mathrm{T}$ in a short-pulse Cu double-coil geometry, while modeling and simulation studies extend the concept toward kilotesla-class and ultrafast multi-kT regimes (Wang et al., 2014, Santos et al., 2015, Gao et al., 5 May 2025, Brantov et al., 2019).

1. Canonical target geometries and architectural variants

The most widely used capacitor-coil architecture consists of two parallel metallic plates connected by a shaped conductor. In the nanosecond platform of Santos et al., each target comprises two parallel metallic disks of diameter $3.5\,\mathrm{mm}$ and thickness $50\,\mu\mathrm{m}$, separated by $d_0\approx 900\pm 200\,\mu\mathrm{m}$; the front disk has a laser-entry hole of $1.0$–$1.75\,\mathrm{mm}$ diameter, and a single-turn square-section wire of cross-section $>100\,\mathrm{ps}$0 is bent into a loop of radius $>100\,\mathrm{ps}$1 to connect the disks. Targets were made of Cu, Ni, or Al (Santos et al., 2015).

Later short-pulse and reconnection platforms converged on a related Cu geometry using parallel square foils and paired U-coils. In the Osaka interferometry study, two parallel Cu foils of thickness $>100\,\mathrm{ps}$2 and area $>100\,\mathrm{ps}$3 are separated by $>100\,\mathrm{ps}$4 and joined by two identical U-shaped Cu coils with wire cross-section $>100\,\mathrm{ps}$5, straight legs $>100\,\mathrm{ps}$6 long, half-circle radius $>100\,\mathrm{ps}$7, and coil-to-coil separation $>100\,\mathrm{ps}$8; a $>100\,\mathrm{ps}$9-radius hole in the front foil admits the drive laser (Zhang et al., 30 Jan 2026). The record-field short-pulse experiment used two $600$0-thick Cu foils, each $600$1, separated by $600$2 and bridged by two identical one-turn U-shaped Cu wires of cross-section $600$3, each with two $600$4 legs and a semicircular turn of radius $600$5 (Gao et al., 5 May 2025). Micro-MRX employed a closely related double-U geometry with two $600$6 copper plates, each $600$7 thick, connected by two identical U-shaped Cu coils separated by $600$8 center-to-center (Ji et al., 2024).

A distinct but related family uses a simpler foil-plus-loop topology rather than two explicit plates. In the simple Cu-coil study, a single-turn loop machined from $600$9-diameter Cu wire with loop diameter $800\,\mathrm{T}$0 is welded directly onto a planar Cu foil of thickness $800\,\mathrm{T}$1; the foil acts as the electron-emission surface and the loop as the return-current path (Zhu et al., 2015). Another related class comprises single-piece, laser-cut Cu targets connected to ground through a stalk, with a one-turn loop of radius $800\,\mathrm{T}$2 formed in the target rod; these were used to study guided electromagnetic discharge pulses rather than the plate-to-plate capacitor discharge alone (Ehret et al., 2022). At much smaller scale, 3D PIC simulations have also treated a single-turn loop made from a thin conducting foil with an emitter and collector separated by a $800\,\mathrm{T}$3 vacuum gap, all contained in a $800\,\mathrm{T}$4 box (Brantov et al., 2019).

These geometries share two defining features: a laser-facing emitter region that produces hot electrons, and a low-inductance closed path that converts target charging into a coil current. A plausible implication is that later geometrical diversification was driven less by the need to create the basic magnetic pulse than by the need to tailor field topology, diagnostic access, and coupling to a secondary experiment.

2. Charging physics and equivalent-circuit descriptions

In the standard capacitor-coil picture, laser irradiation of one plate generates supra-thermal or hot electrons, some fraction of which escape and charge the opposing conductor, establishing a potential difference $800\,\mathrm{T}$5 or $800\,\mathrm{T}$6. The target then behaves as a lumped capacitor in series with an inductance and resistance, and the coil current rises during the laser pulse as the capacitor discharges through the connecting wire. For the Santos platform, the current obeys

$800\,\mathrm{T}$7

with step-like charging giving the usual RL rise,

$800\,\mathrm{T}$8

Typical orders of magnitude were reported as $800\,\mathrm{T}$9, $250\pm 30\,\mathrm{T}$0, and $250\pm 30\,\mathrm{T}$1, so that $250\pm 30\,\mathrm{T}$2 matches the observed rise time (Santos et al., 2015).

Other experiments and models use the full RLC form. In the femtosecond Ni capacity-coil study, the structure was treated as an under-damped LRC circuit with

$250\pm 30\,\mathrm{T}$3

using $250\pm 30\,\mathrm{T}$4, $250\pm 30\,\mathrm{T}$5, and a resistance rising from $250\pm 30\,\mathrm{T}$6 to $250\pm 30\,\mathrm{T}$7 as the coil heats (Wang et al., 2014). The 2025 diagnostic study similarly wrote the discharge current for a U-coil target as

$250\pm 30\,\mathrm{T}$8

with an underdamped solution

$250\pm 30\,\mathrm{T}$9

and an approximate peak current

$3.5\,\mathrm{mm}$0

for $3.5\,\mathrm{mm}$1 (Zhang et al., 5 May 2025). By contrast, micro-MRX reported an overdamped regime, with

$3.5\,\mathrm{mm}$2

and a monotonic rise during the $3.5\,\mathrm{mm}$3 laser pulse followed by single-exponential decay, using order-of-magnitude parameters $3.5\,\mathrm{mm}$4, $3.5\,\mathrm{mm}$5, and $3.5\,\mathrm{mm}$6 (Ji et al., 2024).

The hot-electron source term is also parameterized differently across regimes. In the Osaka study, a $3.5\,\mathrm{mm}$7, $3.5\,\mathrm{mm}$8 drive at $3.5\,\mathrm{mm}$9 produced hot electrons with $50\,\mu\mathrm{m}$0 hundreds of keV, charging the front foil negatively and establishing a capacitor-like voltage $50\,\mu\mathrm{m}$1 of order $50\,\mu\mathrm{m}$2 (Zhang et al., 30 Jan 2026). In the record-field short-pulse study, the hot-electron temperature followed the ponderomotive scaling

$50\,\mu\mathrm{m}$3

yielding $50\,\mu\mathrm{m}$4 and a capacitor potential $50\,\mu\mathrm{m}$5 for a $50\,\mu\mathrm{m}$6, $50\,\mu\mathrm{m}$7, $50\,\mu\mathrm{m}$8 drive at $50\,\mu\mathrm{m}$9 (Gao et al., 5 May 2025). The simple Cu-coil work emphasized a different language: hot-electron escape creates a positive potential on the irradiated foil, which then pulls cold conduction electrons through the attached loop, so the transient return current is carried by cold electrons rather than by thermionic or plasma-sheath currents alone (Zhu et al., 2015).

Taken together, these studies show that “laser-driven capacitor coil” is not a single circuit regime. Depending on geometry, drive duration, and evolving target conductivity, the effective behavior can be RL-like, underdamped RLC, or overdamped RLC.

3. Magnetic-field generation, scaling, and reported performance

For a single-turn loop, the on-axis field at the center is commonly estimated as

$d_0\approx 900\pm 200\,\mu\mathrm{m}$0

or, more generally along the axis,

$d_0\approx 900\pm 200\,\mu\mathrm{m}$1

This scaling underlies essentially all capacitor-coil analyses (Santos et al., 2015, Zhang et al., 5 May 2025).

In the nanosecond experiments of Santos et al., $d_0\approx 900\pm 200\,\mu\mathrm{m}$2, $d_0\approx 900\pm 200\,\mu\mathrm{m}$3 laser pulses at $d_0\approx 900\pm 200\,\mu\mathrm{m}$4 generated peak on-axis fields of $d_0\approx 900\pm 200\,\mu\mathrm{m}$5 for Cu, $d_0\approx 900\pm 200\,\mu\mathrm{m}$6 for Ni, and $d_0\approx 900\pm 200\,\mu\mathrm{m}$7 for Al, with a rise time of $d_0\approx 900\pm 200\,\mu\mathrm{m}$8 consistent with the laser duration. The field distribution was dipole-like, homogeneous within the coil interior, and extended over a characteristic volume of $d_0\approx 900\pm 200\,\mu\mathrm{m}$9. Reported laser-to-magnetic conversion efficiencies were $1.0$0 for Cu, $1.0$1 for Ni, and $1.0$2 for Al (Santos et al., 2015).

A broader 2017 review of ns-driven capacitor-coils reported discharge currents of a few $1.0$3 and magnetostatic fields in excess of $1.0$4 over ns time scales, with Ni targets yielding reproducible peak fields $1.0$5 up to $1.0$6 for peak current $1.0$7. In that account, the total magnetic energy in the coil region corresponded to $1.0$8 of the driver-laser energy, and the major empirical control parameter was the laser irradiance $1.0$9, leading to the statement $1.75\,\mathrm{mm}$0 (Santos et al., 2017).

Shorter-pulse experiments operate at lower absolute field than the best ns-driven Cu targets but with faster formation. The femtosecond Ni capacity-coil experiment measured a peak loop current $1.75\,\mathrm{mm}$1 and a peak field $1.75\,\mathrm{mm}$2, with rise to maximum at $1.75\,\mathrm{mm}$3–$1.75\,\mathrm{mm}$4 and decay to half amplitude in $1.75\,\mathrm{mm}$5. The magnetic-field energy inside a $1.75\,\mathrm{mm}$6 volume was estimated as $1.75\,\mathrm{mm}$7, giving $1.75\,\mathrm{mm}$8 for one shot and typical $1.75\,\mathrm{mm}$9 across the reported shots (Wang et al., 2014). The record short-pulse Cu double-coil experiment, driven by a $>100\,\mathrm{ps}$00, $>100\,\mathrm{ps}$01 laser, inferred a peak current of $>100\,\mathrm{ps}$02 and a magnetic field of $>100\,\mathrm{ps}$03 at the coil center, with magnetic energy $>100\,\mathrm{ps}$04 and a laser-to-magnetic energy conversion efficiency of $>100\,\mathrm{ps}$05–$>100\,\mathrm{ps}$06 (Gao et al., 5 May 2025).

Still other studies focus on current-field topology rather than absolute record field. Micro-MRX inferred $>100\,\mathrm{ps}$07, $>100\,\mathrm{ps}$08 for one coil, and an anti-parallel upstream field $>100\,\mathrm{ps}$09 between the two coils, with $>100\,\mathrm{ps}$10 for magnetic energy relative to the $>100\,\mathrm{ps}$11 laser drive (Ji et al., 2024). The short-pulse Osaka interferometry study reported proton-radiography measurements at $>100\,\mathrm{ps}$12 giving $>100\,\mathrm{ps}$13 and noted that tens of kA through a $>100\,\mathrm{ps}$14-radius loop produce quasi-static magnetic fields of tens to hundreds of tesla in the coil interior (Zhang et al., 30 Jan 2026).

Simulation studies extend the accessible parameter space. In 3D PIC calculations of a foil-coil target driven by a $>100\,\mathrm{ps}$15, $>100\,\mathrm{ps}$16 pulse at $>100\,\mathrm{ps}$17, the azimuthal field component inside the loop reached $>100\,\mathrm{ps}$18–$>100\,\mathrm{ps}$19 for $>100\,\mathrm{ps}$20, $>100\,\mathrm{ps}$21 for $>100\,\mathrm{ps}$22, and $>100\,\mathrm{ps}$23 for $>100\,\mathrm{ps}$24, with a collisionless decay time of $>100\,\mathrm{ps}$25 and an expected $>100\,\mathrm{ps}$26–$>100\,\mathrm{ps}$27 lifetime in a real metal coil (Brantov et al., 2019). In the bifilar helical undulator concept, a $>100\,\mathrm{ps}$28, $>100\,\mathrm{ps}$29 charging laser produced a modeled peak current $>100\,\mathrm{ps}$30, corresponding to an on-axis transverse amplitude $>100\,\mathrm{ps}$31 for a period $>100\,\mathrm{ps}$32 and internal diameter $>100\,\mathrm{ps}$33 (Tan et al., 2019).

Reported energy-conversion efficiencies therefore range from $>100\,\mathrm{ps}$34 in micro-MRX to $>100\,\mathrm{ps}$35 in femtosecond proton-radiography shots (Ji et al., 2024, Wang et al., 2014). This suggests that quoted efficiency depends strongly on topology, pulse duration, coupling physics, and on how the magnetic-energy volume is defined.

4. Diagnostics and the problem of separating magnetic and electric fields

The diagnostic canon for laser-driven capacitor coils is built around inductive probes, optical polarimetry, and proton radiography. In the Santos experiments, inductive pickup coils with $>100\,\mathrm{ps}$36 bandwidth were placed $>100\,\mathrm{ps}$37–$>100\,\mathrm{ps}$38 from the coil center, with $>100\,\mathrm{ps}$39–$>100\,\mathrm{ps}$40 time resolution. The raw signal, proportional to $>100\,\mathrm{ps}$41, was band-pass filtered from $>100\,\mathrm{ps}$42 to $>100\,\mathrm{ps}$43–$>100\,\mathrm{ps}$44 to remove DC drifts and EMP spikes, then numerically integrated to obtain $>100\,\mathrm{ps}$45; absolute calibration used FFT-noise analysis, and the spatial uncertainty was reported as $>100\,\mathrm{ps}$46 from magnetostatic Radia modeling (Santos et al., 2015). The same work used Faraday rotation of a $>100\,\mathrm{ps}$47, $>100\,\mathrm{ps}$48 probe through two $>100\,\mathrm{ps}$49-thick TGG crystals at $>100\,\mathrm{ps}$50 from the coil plane, with rotation inferred from the ratio $>100\,\mathrm{ps}$51 and $>100\,\mathrm{ps}$52, using $>100\,\mathrm{ps}$53 (Santos et al., 2015).

Proton deflectometry and proton radiography remain the most flexible field-mapping methods. In the 2015 ns platform, a $>100\,\mathrm{ps}$54, $>100\,\mathrm{ps}$55 laser on Au foil produced $>100\,\mathrm{ps}$56 TNSA protons, a 42-pitch mesh cast a shadow onto a 15-layer RCF stack $>100\,\mathrm{ps}$57 downstream, and Monte Carlo ray tracing through 3D Radia-computed field maps was used to infer $>100\,\mathrm{ps}$58 (Santos et al., 2015). The femtosecond Ni capacity-coil study used time-gated proton radiography with a mesh at $>100\,\mathrm{ps}$59 from the proton source, the coil at $>100\,\mathrm{ps}$60, and RCF at $>100\,\mathrm{ps}$61, reconstructing the field by matching 3D particle-tracking simulations to the measured proton traces (Wang et al., 2014). The 2025 record-field paper used axial probing with a TNSA proton beam of peak energies $>100\,\mathrm{ps}$62–$>100\,\mathrm{ps}$63, a $>100\,\mathrm{ps}$64-pitch Au mesh, and stacked RCF; clockwise rotation of the mesh inside the coils was treated as an unambiguous magnetic signature, and synthetic radiographs were fit to infer $>100\,\mathrm{ps}$65 (Gao et al., 5 May 2025).

A recurrent difficulty is that proton radiography generally measures the combined Lorentz force,

$>100\,\mathrm{ps}$66

so electric and magnetic deflections overlap. The 2025 field-separation study systematized this issue for U-coil targets by comparing face-on and side-on radiographs under magnetic-only, electric-only, and combined-field conditions. It identified distinct signatures: magnetic-only face-on radiographs show a central void and tip or bar features at the legs, whereas electric-only face-on images show uniform deflection of grid lines; magnetic-only side-on radiographs show asymmetric leg widths and rotation of the central mesh pattern, whereas electric-only side-on radiographs show symmetric widening or narrowing without grid rotation. On that basis, the side-on grid-rotation method was proposed, with

$>100\,\mathrm{ps}$67

and

$>100\,\mathrm{ps}$68

using $>100\,\mathrm{ps}$69, followed by residual fitting to extract line-charge density $>100\,\mathrm{ps}$70 (Zhang et al., 5 May 2025).

The misconception that proton radiography directly yields magnetic field without ambiguity is therefore not supported by the later literature. The diagnostic record instead shows that reliable inference requires geometry-aware forward modeling and, in many cases, explicit electric-field separation.

5. Plasma loading, non-ideal effects, and temporal limitations

Capacitor-coil targets do not evolve in vacuum once the discharge begins. Time-resolved interferometry has now shown that the coil region is loaded by at least two distinct plasma sources. In the Osaka measurements, two-dimensional electron-density maps at $>100\,\mathrm{ps}$71 were reconstructed using

$>100\,\mathrm{ps}$72

with $>100\,\mathrm{ps}$73 and a typical phase-retrieval error $>100\,\mathrm{ps}$74 corresponding to an uncertainty $>100\,\mathrm{ps}$75. At $>100\,\mathrm{ps}$76 in the coil target, dense lobes adjacent to the coils reached $>100\,\mathrm{ps}$77–$>100\,\mathrm{ps}$78, while a plate-ablation component outside the lobes reached $>100\,\mathrm{ps}$79; by $>100\,\mathrm{ps}$80 the entire inter-coil volume was loaded to $>100\,\mathrm{ps}$81 (Zhang et al., 30 Jan 2026).

The same study identified two sources of plasma loading: coil-wire plasma produced by inductive electric fields, ohmic heating, and energetic-particle or X-ray irradiation of the wire, and a plate-ablation plume expanding into the coil interior. The coil-wire plasma alters effective resistance and skin depth, modifying current distribution and reducing peak field and rise time relative to an ideal vacuum circuit, while the plate plume can impede magnetic-field penetration, alter coupling to secondary plasmas, and introduce additional inductive loading (Zhang et al., 30 Jan 2026). This directly challenges the simplified view that the coil interior remains an empty, passively magnetized volume during the useful part of the pulse.

Other measurements delimit the valid diagnostic window. In the Santos experiments, B-dot, Faraday, and proton measurements agreed for $>100\,\mathrm{ps}$82 when extrapolated by a full 3D magnetostatic model of the exact target geometry; beyond $>100\,\mathrm{ps}$83, Faraday rotation broke down because of crystal ionization by X-rays and fast particles, and proton traces were perturbed by accumulating magnetized electrons near the coil that built a negative electrostatic potential partially neutralizing proton deflections. The same work also reported that the circuit eventually “shorts” in $>100\,\mathrm{ps}$84–$>100\,\mathrm{ps}$85 as plasma plumes or thermally expanded wires bridge the disk gap (Santos et al., 2015).

More broadly, guided-discharge experiments along grounded wires indicate that surface plasma can determine propagation characteristics of the pulse itself. In the PHELIX study, proton deflectometry measured a guided pulse group velocity $>100\,\mathrm{ps}$86, charge per unit length peaking at $>100\,\mathrm{ps}$87, total charge $>100\,\mathrm{ps}$88, and peak current $>100\,\mathrm{ps}$89, with radial electric fields of $>100\,\mathrm{ps}$90–$>100\,\mathrm{ps}$91 and azimuthal magnetic field at the coil center of $>100\,\mathrm{ps}$92. Analytical and PIC modeling attributed the dispersion to a TM-like Sommerfeld wave guided by a hot surface plasma of density $>100\,\mathrm{ps}$93 and temperature $>100\,\mathrm{ps}$94 (Ehret et al., 2022). A plausible implication is that even in nominally metallic targets, evolving surface plasma can no longer be treated as a perturbation once the pulse duration reaches tens of picoseconds or longer.

6. Applications, derivative concepts, and research directions

The principal scientific value of laser-driven capacitor coils is that they generate strong magnetic fields in an open geometry that remains accessible to secondary drivers and diagnostics. This has already enabled direct applications to beam transport. In the 2017 magnetostatic-field study, a $>100\,\mathrm{ps}$95 CH slab with a $>100\,\mathrm{ps}$96 Cu rear coating was placed in the coil plane and magnetized by a field of approximately $>100\,\mathrm{ps}$97 before irradiation by a $>100\,\mathrm{ps}$98, $>100\,\mathrm{ps}$99 laser. The imposed field radially pinched the relativistic electron beam, increased the coherent transition radiation signal by a factor of 8, and yielded an approximately $600$00 enhancement of time-integrated relativistic-electron-beam energy-density flux at $600$01 depth; the background-electron temperature at the rear rose to $600$02 versus $600$03 without magnetization (Santos et al., 2017).

Magnetic reconnection has become another major application domain. Micro-MRX uses parallel U-coils driven by OMEGA-EP long-pulse beams to produce anti-parallel upstream fields of about $600$04 on millimeter scales and nanosecond durations, in a quasi-axisymmetric geometry analogous to MRX. That platform has reported direct measurement of accelerated electrons and observation of ion acoustic waves during anti-parallel reconnection, supported by FLASH and VPIC simulations (Ji et al., 2024). A 2026 extension proposed injecting MeV electron-positron pairs into a capacitor-coil reconnection layer; in VPIC simulations with prescribed current rising linearly to $600$05 over $600$06, pair loading increased the reconnection rate by a factor of approximately 8, broadened the diffusion region to the positron Larmor-radius scale, and produced particle energy gains up to $600$07 for injected positrons (Russell et al., 17 Mar 2026).

Capacitor-coil concepts have also been adapted as insertion devices. The bifilar capacitor-coil undulator concept used a helical double-coil with period $600$08, internal diameter $600$09, and $600$10 turns, driven by a modeled peak current of $600$11 to produce an on-axis field amplitude $600$12 and undulator parameter $600$13. Coupled to state-of-the-art laser wakefield accelerators, WAVE simulations predicted quasi-monochromatic X-rays tunable from approximately $600$14 to $600$15 with peak brightness up to $600$16, and a 1D FEL gain length $600$17 for representative beam parameters (Tan et al., 2019).

Related self-discharging helical-coil targets have been proposed for proton post-acceleration rather than direct magnetic-field production. In that scheme, a picosecond-scale current pulse propagates as a surface wave along an aluminum helical coil, producing a transient longitudinal electric field. A two-stage helical-coil design with an inserted drift section increased the proton energy gain from $600$18 to $600$19, a four-fold enhancement, and projected proton energies of $600$20 with a $600$21 laser and $600$22 with a petawatt laser (Liu et al., 2022).

Across the literature, the recurring outlook is toward reproducible, well-characterized magnetic pulses suitable for magnetized high-energy-density physics, inertial-fusion studies, laboratory astrophysics, and compact radiation sources (Santos et al., 2015, Santos et al., 2017). The more recent plasma-loading and diagnostic-separation studies suggest that future progress will depend at least as much on controlling the evolving plasma-field environment as on increasing the nominal peak current.