Laser-driven ion acceleration is a highly promising novel particle acceleration method, owing to its compactness and efficiency.1–3 It can produce particle bunches with more than 10¹³ particles4 and high ion energies (up to 160 MeV5) with low emittance6 and short pulse durations.7 Applications of laser-accelerated ions are diverse, ranging from isochoric heating8,9 and isotope production10 to neutron generation,11 plasma radiography,12,13 fast-ignition nuclear fusion,14,15 materials science,16 and the analysis of cultural heritage artifacts.17,18

One of the simplest and most widely used laser ion acceleration mechanisms is target normal sheath acceleration (TNSA).19,20 In this process, a high-intensity laser beam interacts with a solid target, accelerating electrons from the front side of the target. Some electrons escape the target, leaving behind positively charged ions, while others are retained by the positive charge, creating a strong sheath electric field on the rear surface of the target. This field, of the order of several TV/m, accelerates ions to energies of tens of MeV in the target-normal direction, typically within a semi-angular distribution of approximately ±20°. However, this wide angular divergence of TNSA-accelerated ions and its broad energy spread can pose significant challenges for medical and industrial applications where dose and/or energy needs to be concentrated onto specific spots.

To address these limitations, a variety of advanced schemes have been proposed with the aim of improving ion beam collimation,8,21,22 energy selection,21,23 and post-acceleration.24 All of these acceleration schemes have many drawbacks, including complex fabrication, high cost, and the need for expensive additional hardware. One promising alternative approach involves the use of helical coil (HC) targets.25–31 This concept draws on the physics of wave–particle interactions similar to those observed in traveling-wave tube amplifiers,32,33 where an electron beam interacts with a Radio frequency (RF) signal within a waveguide, leading to signal amplification. In the HC target setup, the interaction between the laser-induced plasma and a helical coil connected to the target generates an electromagnetic pulse (EMP) that post-accelerates and focuses the proton beam.25 This method offers the possibility of simultaneously focusing, post-accelerating, and bunching the proton beam to overcome the limitations of traditional TNSA beams. As described in previous publications,26,27 the implementation of a tube around the helical coil with tube (HCT) changes the impedance of the system to mitigate the current dispersion and increase the application length of the accelerator field, resulting in the production of huge proton energy bunching.

In this study, we focus on an experimental investigation of HC targets using the Advanced Laser Light Source (ALLS) at the Institut National de la Recherche Scientifique (INRS) in Varennes, Canada. Our experimental results are complemented by simulations to provide a comprehensive understanding of the underlying phenomena. For the first time, we examine the effect of HC targets on carbon ions, expanding the scope of HC-target research beyond merely proton beams. This experimental–numerical comparison confirms the potential of HC targets for optimizing ion acceleration for different ion species (protons and carbon ions).

In the experiments, we used solid tantalum foils of 2 μm thickness as interaction targets for producing the ion acceleration. A schematic of the experimental setup for the ion acceleration using an HC is depicted in Fig. 1. The HC target setup was realized by inserting a circular piece of foil with the same radius as the HC, which had been cut out of a 2 μm tantalum foil by ultrafast laser micromachining. Figure 2 shows the helical coil (a) without and (b) with the target foil. The fabrication method for the cutting of the target foil is presented in Sec. II A and the ion acceleration experiment with HC using a Ti:sapphire laser is presented in Sec. II B.

Femtosecond laser micromachining is a universal tool allowing very precise processing of any material thanks to the extremely short interaction time of the laser beam with the material.34 Thermal diffusion is suppressed in this regime and thus reduces the formation of a heat-affected zone even in high-thermal-conductivity materials such as metals.35 Moreover, transparent materials such as dielectrics can be modified very locally in the bulk by nonlinear absorption within the laser focus.36

For the target fabrication, microcutting of small disks of 1–3 mm diameter out of a 2 μm-thickness tantalum foil was performed [see Fig. 2(b)] using a commercially available Yb-doped femtosecond laser source (Tangor 100, Amplitude) operating at 1030 nm wavelength37 coupled to a home-built micromachining station.38 The cutting process was performed using a repetition rate of 20 kHz and an average power of 0.67 W, corresponding to a fluence of 8.5 J/cm². The trajectories for cutting the circles out of the metal foil were generated using a galvo scanner (IntelliScan III-14, Scanlab) equipped with a 100 mm focusing lens. The scanning velocity was 10 mm/s, and the cutting was done in a single pass. The foil was self-supported by two spring pliers with a distance of 12 mm between the pliers. Once the foil had been cut, the disks fell down into a plastic cup. All disks were then sorted with a microsuction device and stored in a membrane box. Finally, the dimensions of the disks were measured with a measuring microscope equipped with a 20× objective (MF-B1010D, Mitutoyo).

This target fabrication method, by ultrafast laser processing, allows for extremely precise cutting without causing damage, thereby preserving the quality of the TNSA during the experiments. TNSA relies on perfectly flat targets with minimal imperfections, which optimize the acceleration process. In other cutting processes, targets may become folded, creased, or torn, which can disrupt the laser–plasma interaction, reducing the ion energy cutoff and the number of accelerated ions.

The experiment was performed on the laser-driven ion acceleration beamline of the ALLS. We used the ALLS 150 TW Ti:sapphire laser (λ₀ = 800 nm) system delivering 3.2 J on target with a pulse duration of τ_FWHM = 22 fs at full-width half-maximum (FWHM). The experimental setup (Fig. 1) used an f/3 off-axis parabola (OAP) to focus the 95 × 95 mm² beam (at e⁻²) down to a spot size of w_FWHM ≈ 5 μm. Parabola alignment and wavefront optimization, using a feedback loop between a wavefront sensor and a deformable mirror, were both performed at full laser power, which allowed compensation of aberrations arising from thermalization in the laser system. This resulted in a peak intensity I₀ around 1.3 × 10²⁰ W/cm².2,39 Before entry to the second CPA amplification stage, a cross-wave polarizer (XPW) and a booster stage relying on a saturable absorber were employed to clean the incoming laser beam, achieving an amplified spontaneous emission (ASE) prepulse contrast <10−10 at −100 ps before the main pulse, along with a steep power rise with contrast <10−6 at −3 ps.

The p-polarized laser pulses were incident at an angle of 20° with respect to the target normal on the circular 2 μm-thick tantalum foils at the tip of the HC to produce ion beams using TNSA. The energy distribution of the accelerated ions was measured using a calibrated Thomson parabola spectrometer positioned at 0° with respect to the target-normal axis, employing a double microchannel plate (MCP)2 detector in chevron configuration with a P43 phosphor screen, allowing acquisition of images of the ion spectra for each laser shot. Additionally, time-of-flight (ToF) spectrometers were employed at 9° and 180° from the main axis to monitor the number of protons and their energies in correlation with the Thomson parabola spectrometer. The primary purpose of these ToF spectrometers was to obtain online verification of the laser–plasma interaction during high-repetition-rate experiments.40 Finally, radiochromic films (RCFs) of type EBT3 were used to measure the angular distribution of the ion beam. The films were positioned at a distance of 50 cm centered on the proton beam axis. After irradiation, the RCFs were scanned with an Epson Perfection 2450 scanner.

Figure 3(a) presents the proton spectrum obtained using an HC of length 5 mm length (red curve), compared with the classical TNSA spectrum obtained at INRS (black curve) and with the background noise (blue dashed curve). The HC geometry was chosen on the basis of previous simulations26,27 to synchronize the electric field propagation velocity with the proton velocity for energy EHC=12mpVHC2=4 MeV (the stable energy within the proton spectrum for the ion beamline at ALLS), where VHC=hc/h2+4π2a2 is the phase velocity and m_p is the proton mass.

It can be seen that the use of the HC increases the cutoff energy from 4 to 5 MeV, enhances the fluence for protons above 2.4 MeV, and reduces the number of low-energy protons. This reduction of low-energy protons is attributed to space-charge effects, as discussed in previous work.26,41

Figure 3(b) depicts a RCF, placed at 50 cm from the TNSA target, which clearly shows a reduced angular dispersion, about 2.3° with our HC compared with the 20° obtained with classical TNSA. The opening angle of the HC is 6.8°, which confirms that the angular dispersion is reduced thanks to the focusing effect of the coil and is not just a filtering effect.

During TNSA, multiple ion species are accelerated, predominantly protons and carbon ions. Up until now, no laser experiments have demonstrated an effect of HC targets on carbon ions. This is probably because, in most cases, previous experiments were conducted with a stack of RCFs as the sole diagnostic tool, without the use of a Thomson parabola spectrometer. Carbon ion spectra are presented in Fig. 4(a), where the dashed lines represent spectra obtained in the presence of an HC, while the solid lines represent those obtained in the absence of an HC. The black solid line represents the background noise for carbon ions. It can be seen that there is a decrease in the number of carbon ions, from 10¹⁰ with TNSA alone to 10⁹ when an HC is present. The cutoff energy of the carbon ions from TNSA alone is 2.5 MeV for C¹⁺ and 4 MeV for C²⁺, C³⁺, C⁴⁺. By contrast, with the HC, the signals above 2.5 MeV are indistinguishable from background noise.

Figure 4(b) compares carbon ion spectra obtained using TNSA together with a short-circuited HC (dotted lines) and those obtained using TNSA without an HC (solid lines). A short-circuited HC means that the HC is connected to the tube, i.e., Δr = 0 in Fig. 1(b). Therefore, the major part of the current pulse does not propagate through the helix. The short-circuited HC serves as a reference to isolate the effect of the coil itself. Indeed, when the number of carbon ions observed with a short-circuited HC is similar to the background noise, this indicates that the degradation of laser–plasma interaction quality caused by the HC is responsible for the reduced number of carbon ions. Conversely, if the carbon ion yield with the short-circuited HC is comparable to that from TNSA alone, the reduced carbon ion production can be attributed to the discharge current propagating through the coil.

Figure 4(b) demonstrates that the cutoff energy of carbon ions remains similar between the HC and TNSA cases, while the total number of carbon ions is of the same order of magnitude but reduced by a factor of two owing to shot-to-shot fluctuation.

These comparisons reveal a clear effect of the HC on the carbon ions. In the case of a short-circuited HC, carbon ion spectra similar to those obtained with standard TNSA are observed. However, when an HC that has not been short-circuited is employed, the number of carbon ions becomes comparable to the background noise level, indicating a notable suppression of carbon ion production. This suppression results from the increased space-charge induced by the HC.41 The HC focuses ions and defocuses electrons, which induces a beam deneutralization. This creates a significant space-charge effect, causing the dispersion of those ions that are not synchronized with the discharge current, which induces an explosion of the ion beam. For particles synchronized with the discharge current, the radial electric field generated by the current propagation through the helix27 partially compensates for the space-charge. However, for nonsynchronized particles, no such compensation occurs, leading to beam dispersion and a reduction in the number of carbon ions by approximately a factor of 100. Finally, the particle beam produced at the helix exit is composed of only one ion species, namely, protons.

To shed light on current propagation within the helix, we used the code SOPHIE.42 SOPHIE is a 3D FDTD-PIC simulation code developed to solve Maxwell’s equations in both vacuum and material environments. The code supports boundary conditions for perfect electric conductors, dielectrics, and magnetic materials. Relativistic particle motion in vacuum is governed by Newton’s second law, solved using the Boris algorithm.43 To ensure self-consistency, Buneman current collectors are implemented. The Buneman instability occurs when electrons move significantly faster than ions, generating electrostatic waves that enhance energy transfer and current dissipation. In a PIC-FDTD code like SOPHIE, implementation of Buneman current collectors refers to a technique ensuring consistency between the current carried by charged particles and the current induced in the electromagnetic grid of the simulation. This helps accurately model plasma–conductor interactions and prevent numerical artifacts related to improper current representation.

An important feature of SOPHIE is its ability to simulate current propagation through a helix using boundary conditions, such as perfect electric conductors, without requiring macroparticles. This design choice reduces computational constraints and facilitates large-scale simulations. Unlike traditional codes that resolve the laser wavelength, SOPHIE is optimized for simulations on grids spanning several centimeters and time intervals on the nanosecond scale. These capabilities make SOPHIE an optimal tool for studying the physics of HC targets.

In our SOPHIE simulations, an HC was positioned at the rear side of a foil target. The target, coil, and holder were considered to be perfect conductors and were meshed at full scale with a cell size of Δx = Δy = Δz = 20 μm, yielding a total of 2 × 10⁹ cells across a volume of 72 cm³. Electrons, protons, and carbon ions were represented using 10⁷ macroparticles each.

Inputs for ion energy and angular distributions were taken from experimental data obtained with TNSA without an HC. The input proton energy distribution was that represented by the black curve in Fig. 3(a), and the ion distributions were those represented by the blue and red curves for C³⁺ and C⁴⁺, respectively, in Fig. 4(a). The angular distribution was modeled using a super-Gaussian function given by dNp/dθ∝exp[−(θ/θp)10/2], where θ_p = 19°. The temporal emission profile dN/dt had a Gaussian shape with an FWHM duration τ of 7 ps. The emission zones for protons and carbon ions were both set to a transverse size of 200 μm.

The proton and carbon ion dynamics within the HC are determined solely by the electromagnetic fields generated by the discharge current. To simulate this current, electrons were emitted isotropically with an energy distribution similar to that of the protons. The transverse size of the electron emission zone matched that of the protons and carbon ions (200 μm).

In the simulations, the electron charge emission was set to 40 nC, while the proton charge emission was 5.6 nC, both emitted over a duration of 7 ps. Carbon ions were emitted with total charges of 6.9 and 3.8 nC for C³⁺ and C⁴⁺, respectively, over the same emission duration as that for protons.

The electron emission induces a discharge current in the helix of amplitude 1.8 kA and a pulse duration of 15 ps (FWHM), which matches perfectly with the values reported in the experiment of Kar et al.,25 who measured 60 nC of electron charge and 15 ps of pulse duration for the discharge current. This current profile is shown in Fig. 5, where a rise time of 5 ps and a decay time of 10 ps are highlighted, which again correspond to the experimental results of Kar et al.25

From the comparison of experimental and simulated results for protons presented in Fig. 6(a), it can be seen that the simulation overestimates the proton cutoff energy compared with the experimental data. During the experiments, the laser–plasma interaction created a charge separation generating an electric field that accelerated the protons, with the most energetic electrons escaping. However, in the simulations, we had to take into account a particle emission similar to that obtained by TNSA. Therefore, we created a fictitious TNSA-like electric field that slightly accelerated protons over a distance of a few hundreds of micrometers with a field of the order of a hundred GV/m. This added field is an artifact, but it cannot be removed from the simulation. In conclusion, Fig. 6(a) shows an excellent agreement between experiment and simulation, with both showing the same number of protons in the plateau level and a cutoff energy of 5 MeV. This strong correlation validates the accuracy of our model, enabling us to robustly and reliably design HCs specifically for carbon acceleration.

In Fig. 6(b), the SOPHIE results for carbon ions show a decreased number of carbon ions, lower than the background noise level. Experimentally, the number of carbon ions obtained with the HC designed for protons is of the same order as the background noise. This comparison confirms the validity of the results presented in Fig. 6(a) and the input data used in the SOPHIE simulations.

Now that we have demonstrated that HCs act not only on protons, but also on carbon ions, we can design new HC target geometries adapted to carbon ions. To optimize the HC geometries for accelerating protons, the phase speed V_HC, which corresponds to the longitudinal speed of a current pulse propagating at the speed of light along the helical coil, is computed. It is given by VHC=hc/h2+4π2a2, where h is the pitch and a the radius of the helix. This phase speed determines the group of particles that is synchronized with the electromagnetic pulse, having an initial energy EHC=12miVHC2, where m_i is the ion mass. To predict the final energy of the ion bunch, we apply the scaling law presented in our previous work.26 Here, we propose to use the same methodology for the design of new HC target geometries optimized for carbon ion acceleration, focusing, and bunching.

We have performed further SOPHIE simulations using the measured C⁴⁺ and C³⁺ ion spectra [see Fig. 4(b)] as input data. Figure 6(c) presents the potential effects of an HCT on carbon ions, showing an increase in energy up to 9 MeV, energy bunching up to 7 MeV, and improved focusing, as indicated by the increased number of carbon ions per solid angle. This simulation result, based on experimental input data, demonstrates that an appropriate HC target can indeed also accelerate, focus, and bunch carbon ions.

This study has demonstrated the impact of HC targets on the shaping of proton and carbon ion beams in laser-driven ion acceleration experiments. On the ALLS laser facility, we have observed improvements in proton acceleration, increased cutoff energy, and enhanced beam focusing. These results align with our previous theoretical and simulation-based studies, confirming the potential of HC targets for optimizing proton acceleration schemes.

This experiment is also the first to demonstrate that HC targets can act on heavy ions. During the HC acceleration scheme, electrons are defocused, producing a deneutralization of the particle beam and creating a strong space-charge. If an HC is designed for protons, the propagation of the radial electric field that it produces is not synchronized with that of carbon ions. Therefore, these ions are only subjected to their space-charge, which defocuses and suppresses them from the main particle beam. As observed during this experiment, the particle beam at the helix exit is a pure monospecies, beam composed only of protons.

Our new PIC simulations, based on experimental input data measured during this experiment, suggest that appropriate tuning of the HC geometry can selectively synchronize with ion species, enabling energy bunching, focusing, and post-acceleration of carbon ions.

The use of HCs has shown great potential in producing pure monospecies ion beams, which is of great interest for medical applications. The ability of HCs to double the energy of carbon ions is especially beneficial for oncological applications. Cancer treatments involving ion therapy require both precise energy control and high carbon ion energies to effectively target and treat deep-seated tumors. Achieving doubled carbon ion energy on petawatt laser facilities could represent a significant step forward in meeting these stringent requirements. This breakthrough would enable enhanced energy modulation capabilities, making laser-driven carbon ion acceleration a promising approach for advanced oncological treatments.

Acknowledgment. The authors would like to acknowledge the staff of the ALLS facility—Joël Maltais, William Lévesque, and Mandy Doly—for their technical assistance during the experiment, as well as members of the CELIA Workshop—Laurent Merzeau and Franck Blais—for their target manufacture contribution. The authors are grateful to V. Tikhonchuk and O. Cessenat for their help and for useful discussions. This work is supported by the CEA/DAM Laser Plasma Experiments Validation Project and the CEA/DAM Basic Technical and Scientific Studies Project. This work is also supported by the National Sciences and Engineering Research Council of Canada (NSERC) (Grant Nos. RGPIN-2023-05459 and ALLRP 556340–20), the Digital Research Alliance of Canada (Job pve-323-ac), the Canada Foundation for Innovation (CFI), and the Ministère de l’Économie, de l’Innovation et de l’Énergie (MEIE) from Québec. This study was granted access to the HPC resources of IRENE under allocation Grant No. A0170512899 made by GENCI. We acknowledge the financial support of the IdEx University of Bordeaux/Grand Research Program “GPR LIGHT” and of the Graduate Program on Light Sciences and Technologies of the University of Bordeaux.