2.1 Flat targets

The most reported type of solid target used in high-power laser experiments – particularly for proton acceleration – is thin, sub 0.5 μm flat foils, made of metals (such as aluminium or gold) or plastic^[ Reference Higginson, Gray, King, Dance, Williamson, Wilson, Capdessus, Armstrong, Green, Hawkes, Martin, Wei, Mirfayzi, Yuan, Kar, Borghesi, Clarke, Neely and McKenna ^{25 ]}, and they require a laser with high temporal contrast. However, thick foils from 0.5 μm and above (commercially available) do not necessarily need a high laser temporal contrast, as pre-pulses can be beneficial for plasma pre-expansion. When irradiated by an ultra-intense laser pulse, a pre-plasma forms at the front surface, and electrons are accelerated through the target. The resulting charge separation generates high electric fields that accelerate ions mostly in the direction normal to the target (thus, giving the name of the TNSA mechanism). These ions come from the contaminant layer on the surfaces of the target (mainly proton, carbon and oxygen ions from water and organic molecules)^[ Reference Roth and Schollmeier ^{2 ]}. Both proton and deuteron beams accelerated by TNSA were obtained from a flat gold target onto which a nm-thick deuterium layer was deposited on the target rear by using a cryogenic system^[ Reference Scott, Carroll, Astbury, Clarke, Hernandez-Gomez, King, Alejo, Arteaga, Dance, Higginson, Hook, Liao, Liu, Mirfayzi, Rusby, Selwood, Spindloe, Tolley, Wagner, Zemaityte, Borghesi, Kar, Li, Roth, McKenna and Neely ^{34 ]}. Aluminium foils, 3 μm thick, have demonstrated proton cut-off energy up to 31 MeV using the 1 PW laser at the Extreme Light Infrastructure–Nuclear Physics (ELI-NP) facility^[ Reference Cernaianu, Ghenuche, Rotaru, Tudor, Chalus, Gheorghiu, Popescu, Gugiu, Balascuta, Magureanu, Tataru, Horný, Corobean, Dancus, Alincutei, Asavei, Dinca, Dreghici and Doria ^{35 ]}. However, improved laser contrast enables the use of thinner targets (tens of nanometres or less), which increases the proton cut-off energy through a combination of cascaded acceleration regimes. Experiments have reported proton energies up to 150 MeV^[ Reference Ziegler, Göthel, Assenbaum, Bernert, Brack, Cowan, Dover, Gaus, Kluge, Kraft, Kroll, Metzkes-Ng, Nishiuchi, Prencipe, Püschel, Rehwald, Reimold, Schlenvoigt, Umlandt and Zeil ^{26 ]}, while simulations predict values exceeding 200 MeV from 10 PW laser pulses at intensities of ${10}^{23}$ W/cm².

Exploratory science experiments that require a succession of single-shot events can make use of individual free-standing targets mounted on wheels or target matrices mounted in raster plates^[ Reference Cernaianu, Ghenuche, Rotaru, Tudor, Chalus, Gheorghiu, Popescu, Gugiu, Balascuta, Magureanu, Tataru, Horný, Corobean, Dancus, Alincutei, Asavei, Dinca, Dreghici and Doria ^{35 ,} Reference Vladisavlevici, Vlachos, Dubois, Haddock, Astbury, Huerta, Agarwal, Ahmed, Apiñaniz, Cernaianu, Gugiu, Krupka, Lera, Morabito, Sangwan, Ursescu, Curcio, Fefeu, Pérez-Hernández and Ehret ^{36 ,} Reference Chagovets, Stanček, Giuffrida, Velyhan, Tryus, Grepl, Istokskaia, Kantarelou, Wiste, Martin, Schillaci and Margarone ^{37 ]}. The fabrication process of thin free-standing metallic targets such as the ones depicted in Figure 2(a) involves thin film coating of the desired material on soluble substrates (e.g., NaCl) and mounting on a frame by a fishing–floating technique. At the ELI-NP Target Laboratory^[ Reference Gheorghiu, Leca, Popa, Cernaianu and Stutman ^{38 ,} Reference Gheorghiu, Ionescu, Zai, Decebal, Burducea, Velisa, Vasile, Ianculescu, Bobeica, Popa and Leca ^{39 ]}, targets with mirror-like flatness, low roughness and high purity have been achieved for aluminium foils of thicknesses between 8 and 400 nm, and for gold foils between 10 and 400 nm, both on target holders with a hole diameter up to 500 μm (see Figure 2(a)). Plastic (CH, formvar) thin films are prepared by dip-coating or spin-coating from a liquid solution followed by a drying and fishing technique. The surface quality of the final target is highly dependent on the substrate quality^[ Reference Cernaianu, Ghenuche, Rotaru, Tudor, Chalus, Gheorghiu, Popescu, Gugiu, Balascuta, Magureanu, Tataru, Horný, Corobean, Dancus, Alincutei, Asavei, Dinca, Dreghici and Doria ^{35 ,} Reference Gheorghiu, Ionescu, Ghenuche, Cernaianu, Doria, Popa and Leca ^{40 ,} Reference Gheorghiu, Cerchez, Aktan, Prasad, Yilmaz, Yilmaz, Popa, Willi and Leca ^{41 ]}.

Figure 2

Different types of solid targets fabricated within the ELI-NP Target Laboratory: (a) free-standing aluminium ultra-thin film (10 nm) on a 400 μm hole, (b) metallic nanorods, (c) silicon nano-pillars, (d) nanowires, (e) nanotubes, (f), (g) gratings and (h) the multilayer target.

2.2 Structured targets

Adding nano- or micro-structures on top of the flat foil increases the surface area, and thus the laser absorption on the target is enhanced^[ Reference Gheorghiu, Cerchez, Aktan, Prasad, Yilmaz, Yilmaz, Popa, Willi and Leca ^{41 ,} Reference Bargsten, Hollinger, Capeluto, Kaymak, Pukhov, Wang, Rockwood, Wang, Keiss, Tommasini, London, Park, Busquet, Klapisch, Shlyaptsev and Rocca ^{42 ]}. Examples of nano-structures are metallic nanowires and nanotubes (Figures 2(d) and 2(e)), which are being studied for X-ray emission^[ Reference Kulcsár, AlMawlawi, Budnik, Herman, Moskovits, Zhao and Marjoribanks ^{43 ]}, proton acceleration^[ Reference Vallières, Salvadori, Permogorov, Cantono, Svendsen, Chen, Sun, Consoli, D’Humières, Wahlström and Antici ^{44 ]} and terabar pressure formation^[ Reference Bargsten, Hollinger, Capeluto, Kaymak, Pukhov, Wang, Rockwood, Wang, Keiss, Tommasini, London, Park, Busquet, Klapisch, Shlyaptsev and Rocca ^{42 ]}. For their fabrication, elaborated top-down and bottom-up methods are used, such as electrochemical deposition^[ Reference Gheorghiu, Ionescu, Ghenuche, Cernaianu, Doria, Popa and Leca ^{40 ,} Reference Ionescu, Gheorghiu, Lupu, Zai, Magureanu, Dreghici, McCay, Molloy, Ahmed, Borghesi, Popa, Cernaianu, Doria, Tanaka and Leca ^{45 ]}, vapour–liquid–solid growth or optical (ultraviolet (UV) light) and electron beam lithography, followed by the reactive ion etching process^[ Reference Dhindsa, Chia, Boulanger, Khodadad, LaPierre and Saini ^{46 ]}. Another type of nano/micro-structure is the nano-pillar (Figure 2(c)), obtained with the focused ion beam (FIB) technique, which allows direct surface patterning by etching with a xenon ion beam, followed by in situ deposition of low- or high-Z materials (such as carbon or tungsten). Metallic nanorods and free-standing metallic gratings (gold and copper gratings), depicted in Figures 2(b), 2(f) and 2(g), have been manufactured using similar methods as for the top-down approach of the nanowire targets, where electron beam lithography was used to write the pattern on the sample surface, followed by the developing step of the exposed photoresist and dry etching of the unprotected areas by means of the reactive ion etching technique^[ Reference Gheorghiu, Cerchez, Aktan, Prasad, Yilmaz, Yilmaz, Popa, Willi and Leca ^{41 ]}. The self-supported periodic metallic gratings have been used in high-intensity ultrashort relativistic laser pulse experiments for electron acceleration and laser energy coupling investigations^[ Reference Gheorghiu, Cerchez, Aktan, Prasad, Yilmaz, Yilmaz, Popa, Willi and Leca ^{41 ]}. To accelerate carbon ion beams, foil targets made of amorphous carbon (a-C) or diamond-like carbon (DLC), as well as near-critical-density targets such as carbon foams or carbon nanotubes, are commonly employed^[ Reference Chaudhary, Milluzzo, Mcilvenny, Ahmed, Mcmurray, Maiorino-Groeneveld, Polin, Romagnani, Doria, McMahon, Botchway, Rajeev, Prise and Borghesi ^{47 ]}. Various fabrication techniques are utilized for DLC targets, including chemical or physical vapour deposition such as ion beam-assisted deposition processes, pulsed laser deposition, the ion beam-assisted deposition (IBAD) technique, the filtered cathodic vacuum arc (FCVA), ultra-high vacuum (UHV) RF or direct current (DC) sputtering and high-power impulse magnetron sputtering (HiPIMS)^[ Reference Kitagawa, Yamada, Toyoda, Tsubakino, Matsuo, Takaoka and Kirkpatrick ^{48 –} Reference Lin, Sproul, Wei and Chistyakov ^{51 ]}. Regarding the fabrication of carbon nanotubes and low-density, low atomic number polymeric foams, chemical vapour deposition in a specialized high-temperature tubular furnace for carbon nanotubes is used at the ELI-NP facility. Meanwhile, organic (e.g., resorcinol formaldehyde) or inorganic aerogels (e.g., silica aerogels) targets are fabricated via a sol–gel approach followed by the CO₂ supercritical drying method^[ Reference Tamon, Kitamura and Okazaki ^{52 ,} Reference Liang, Sha and Guo ^{53 ]}.

Such near-critical-density targets (e.g., foams/aerogels) contribute to a better coupling with the laser pulse and can be tailored to produce electrons, by the direct laser acceleration (DLA) mechanism^[ Reference Rosmej, Gyrdymov, Andreev, Tavana, Popov, Borisenko, Gromov, Gus’kov, Yakhin, Vegunova, Bukharskii, Korneev, Cikhardt, Zähter, Busch, Jacoby, Pimenov, Spielmann and Pukhov ^{6 ]}, with twice the proton energy and an increased bremsstrahlung yield compared with flat foils^[ Reference Prencipe, Metzkes-Ng, Pazzaglia, Bernert, Dellasega, Fedeli, Formenti, Garten, Kluge, Kraft, Garcia, Maffini, Obst-Huebl, Rehwald, Sobiella, Zeil, Schramm, Cowan and Passoni ^{54 ]}, high-energy photons^[ Reference Shou, Wang, Lee, Rhee, Lee, Yoon, Sung, Lee, Pan, Kong, Mei, Liu, Xu, Deng, Zhou, Tajima, Choi, Yan, Nam and Ma ^{27 ]} or neutron sources for nuclear applications^[ Reference Günther, Rosmej, Tavana, Gyrdymov, Skobliakov, Kantsyrev, Zähter, Borisenko, Pukhov and Andreev ^{55 ]}. Carbon ions with energies up to 48 MeV per nucleon have been accelerated using double-layer targets made of carbon foam on DLC thin film and cascaded acceleration mechanism of RPA and TNSA, which initially ionizes the carbon ions from the near-critical-density foam by RPA followed by their acceleration through the DLC thin layer by TNSA. Tailoring the thickness and density of the foam and the thickness of the DLC film leads to finding an optimum for the acceleration process with a high-contrast fs laser pulse^[ Reference Ma, Kim, Yu, Choi, Singh, Lee, Sung, Lee, Lin, Liao, Zhu, Lu, Liu, Wang, Xu, He, Chen, Zepf, Schreiber, Yan and Nam ^{56 ]}.

Multilayer design, used mostly for anti-reflective (AR) coatings, can be fabricated by alternative sputter deposition processes (RF, DC or HiPIMS) of different material layers, such as oxides, nitrides and metals (Figure 2(h)). In this case, an interconnected multi-chamber UHV system is the ideal solution to reducing interface contamination between layers and obtaining high-purity multilayered targets or optical coatings. For proton–boron fusion experiments, boron targets can be prepared either as thick films, boron nanoparticles or nano-structures^[ Reference Turcu, Margarone, Giuffrida, Picciotto, Spindloe, Robinson and Batani ^{57 ]}, and are used directly as targets or as the proton catcher in the pitcher–catcher scheme, which involves another flat foil target for proton production. Using the alpha particles produced in the fusion reaction, radioisotopes are generated^[ Reference Batani, Rodrigues, Bonasera, Cipriani, Consoli, Filippi, Scisciò, Giuffrida, Kantarelou, Stancek, Lera, Pérez-Hernández, Volpe, Turcu, Passoni, Vavassori, Dellasega, Maffini, Huault and Batani ^{58 ]}.

2.3 High-repetition-rate targets

For experiments that require large statistics or aim at generating a large throughput of secondary radiation, solid target delivery has to approach the HRR alongside quasi-continuous operation of PW-class laser systems. An example is the tape-drive target^[ Reference McKenna, Ledingham, Spencer, McCany, Singhal, Ziener, Foster, Divall, Hooker, Neely, Langley, Clarke, Norreys, Krushelnick and Clark ^{59 ,} Reference Nayuki, Oishi, Fujii, Nemoto, Kayoiji, Okano, Hironaka, Nakamura, Kondo and Ueda ^{60 ]}. Mechanical stability is crucial for precise operation; the tape oscillations have to be below the Rayleigh length of the focused laser pulse to keep the tape surface in focus^[ Reference Ehret, de Luis, Apiñaniz, Henares, Lera, Pérez-Hernández, Puyuelo-Valdes, Volpe and Gatti ^{61 ]} and rupture has to be avoided by choosing wide enough tapes that exceed the diameter of the through-holes remaining after the interaction, resulting from the energy deposition of recirculating electrons^[ Reference Ehret, Apiñaniz, Henares, Lera, de Luis, Pérez-Hernández, Volpe and Gatti ^{62 ]} and shocks^[ Reference Caroll ^{63 ]}. Also essential are robust electronic systems that withstand the extreme environment around a relativistic laser–plasma interaction. Strong auto-generated kA-scale return currents have to be mitigated^[ Reference Ehret, Cikhardt, Bradford, Vladisavlevici, Burian, de Luis, Henares, Martín, Apiñaniz, Lera, Pérez-Hernández, Santos and Gatti ^{64 ]}. The generation of TNSA ion beams at 1 Hz was demonstrated with 10 μm aluminium and 7 μm copper tapes^[ Reference Ehret, Vladisavlevici, Bradford, Cikhardt, Filippov, Henares, Martín, de Luis, Pérez-Hernández, Vicente, Burian, García-García, Hernández, Mendez, Ruíz, Varela, Frías, Santos and Gatti ^{28 ]} by employing a short-pulse PW-class laser. The repetition rate is currently limited by circuit interrupter systems, which are part of the electromagnetic pulse (EMP) protection of motors and controllers for the tape drive. Lower-energy short-pulse lasers with fewer EMPs allow for up to 100 Hz repetition rates^[ Reference Xu, Streeter, Ettlinger, Ahmed, Astbury, Borghesi, Bourgeois, Curry, Dann, Dover, Dzelzainis, Istokskaia, Gauthier, Giuffrida, Glenn, Glenzer, Gray, Green, Hicks and Najmudin ^{65 ]} and have shown good shot-to-shot stability. Laser-driven secondary sources have been demonstrated at various laser systems with different repetition rates while using a wide range of materials, that is, VHS tape^[ Reference Noaman-ul-Haq, Ahmed, Sokollik, Yu, Liu, Yuan, Yuan, Mirzaie, Ge, Chen and Zhang ^{66 ]}, Kapton tape^[ Reference Ehret, de Luis, Apiñaniz, Henares, Lera, Pérez-Hernández, Puyuelo-Valdes, Volpe and Gatti ^{61 ,} Reference Ehret, Cikhardt, Bradford, Vladisavlevici, Burian, de Luis, Henares, Martín, Apiñaniz, Lera, Pérez-Hernández, Santos and Gatti ^{64 ]}, plastic films^[ Reference McKenna, Ledingham, Spencer, McCany, Singhal, Ziener, Foster, Divall, Hooker, Neely, Langley, Clarke, Norreys, Krushelnick and Clark ^{59 ]} and metal foils^[ Reference Ehret, Vladisavlevici, Bradford, Cikhardt, Filippov, Henares, Martín, de Luis, Pérez-Hernández, Vicente, Burian, García-García, Hernández, Mendez, Ruíz, Varela, Frías, Santos and Gatti ^{28 ,} Reference Nayuki, Oishi, Fujii, Nemoto, Kayoiji, Okano, Hironaka, Nakamura, Kondo and Ueda ^{60 ,} Reference Ehret, de Luis, Apiñaniz, Henares, Lera, Pérez-Hernández, Puyuelo-Valdes, Volpe and Gatti ^{61 ,} Reference Ehret, Cikhardt, Bradford, Vladisavlevici, Burian, de Luis, Henares, Martín, Apiñaniz, Lera, Pérez-Hernández, Santos and Gatti ^{64 ,} Reference Raschke, Spickermann and Toncian ^{67 –} Reference Condamine, Jourdain, Hernandez, Taylor, Bohlin, Fajstavr, Jeong, Kumar, Laštovička, Renner and Weber ^{69 ]}. Tape target systems allow not only for the supply of flat foils but also for structured targets and embedding^[ Reference Forsman, Do, Haid, Keaty, Manuel, Lovelace, Williams, Carlson, Alexander, Condamine, Fauvel, Hernandez, Laštovička, Rubovič, Singh and Weber ^{70 ]}. Complex, three-dimensional (3D) printed foam targets were produced by two photon polymerization technique on a tape, allowing for one shot every 3 minutes, limited by the alignment procedure of individually embedded foams.

Automatic alignment technology is being developed to increase the repetition rate to one shot every minute^[ Reference Collins IV, Jaris, Dahlke, Do, Kotian, Sabouri, Haid, McGuffey, Forsman and Manuel ^{71 ]}. All solid metal targets are sources of debris, which is detrimental to beam transport elements and diagnostics^[ Reference Vladisavlevici, Vlachos, Dubois, Haddock, Astbury, Huerta, Agarwal, Ahmed, Apiñaniz, Cernaianu, Gugiu, Krupka, Lera, Morabito, Sangwan, Ursescu, Curcio, Fefeu, Pérez-Hernández and Ehret ^{36 ]}. Mitigation of debris is especially important for tape targets, which can emit hundreds of micrograms per interaction, and therefore several grams per day of operation. A high-speed spinning multi-disc debris reducing system was proposed by Chen et al. ^[ Reference Chen, Zhao, Xu, Peng, Ma, Gao, Chen, Liu, Xu, Pan, Liang, Song, Xu, Yan and Ma ^{72 ]} and will be implemented at the 2 PW laser from Peking University, aiming to reduce the damage caused by the coating of the optical components. All HRR target approaches discussed in this manuscript are summarized in Table 1.

Table 1

Comparison of high-repetition-rate (HRR) target approaches.

2.4 Summary and remarks

Solid target fabrication has a strong base in material science research and brings together different types of methods that combined with precise engineering lead to high reproducibility, constant geometrical parameters and a wide variety of possibilities (material, thickness, density, structuring). While flat targets pave the way for the particle acceleration processes, and rely on straightforward fabrication techniques, structured and near-critical-density materials enhance the laser–matter interaction by increasing the conversion efficiency, allowing the production of tailored radiation sources, while requiring extensive fabrication process development and careful experimental planning for clean temporal contrast of the laser pulse. For structured targets, 3D PIC and hydrodynamic studies would model the interaction in a realistic manner, although they require huge amounts of computational time, which is usually a limited resource, making the experimental approach an attractive and faster option. Therefore, detailed parameter studies are needed in order to discover the optimal target geometrical characteristics for each type of laser system, which require a wide range of readily available targets. Developing systems such as the tape target to reach a higher repetition rate begins to increase number of the data sets and, with that, the understanding of the process, and opens the path to application-oriented efforts. Such tape systems currently work at a repetition rate of more than 1 Hz with some limitations related to the thickness of the target, which is in the range of a few micrometres, and deal with the resulting debris. For more sophisticated targets, holders are yet to be developed, while for the debris, solutions as pellicles in front of the optics and rotating discs are explored. The solid target fabrication methods, along with advantages and disadvantages, as well as their applications, are summarized in Table 2.

Table 2

Target fabrication methods and applications for single-shot experiments ^a .

^a HHG, high-order harmonic generation; RIE, reactive ion etching; VLS, vapor liquid solid; PLD, pulsed laser deposition; CVD, chemical vapor deposition; FIB, focused ion beam; CNT, carbon nanotube; DLC, diamond-like carbon; AR, anti-reflective.

3.1 Gas jet

A gas jet system operates via a pulsed solenoid valve that controls gas flow through the convergent–divergent (de Laval) nozzle^[ Reference Semushin and Malka ^{77 ]}. Gas jets are popular among the LPA community due to their relatively simple implementation and tunable gas density profile. These profiles can be tailored by coupling the gas flow to different nozzle geometries, such as circular^[ Reference Hosokai, Kinoshita, Watanabe, Yoshii, Ueda, Zhidokov, Uesaka, Nakajima, Kando and Kotaki ^{78 ]}, slit^[ Reference Couperus, Köhler, Wolterink, Jochmann, Zarini, Bastiaens, Boller, Irman and Schramm ^{79 ,} Reference Miao, Shrock, Feder, Hollinger, Morrison, Nedbailo, Picksley, Song, Wang, Rocca and Milchberg ^{80 ]}, shock-front^[ Reference Buck, Wenz, Xu, Khrennikov, Schmid, Heigoldt, Mikhailova, Geissler, Shen, Krausz, Karsch and Veisz ^{81 ]}, reaching the typical electron density, n _e  $\sim$ 10¹⁷–10²⁰ cm^–3. The diagnostics of inner plasma processes are straightforward due to their open geometry. However, a high gas flow burdens the vacuum systems and limits operation to a low repetition rate ( $<$ 1 Hz) to prevent the gas overloading. The density fluctuation in gas jets is typically higher compared to the gas cell and capillary discharge^[ Reference Osterhoff, Popp, Major, Marx, Rowlands-Rees, Fuchs, Geissler, Hörlein, Hidding, Becker, Peralta, Schramm, Grüner, Habs, Krausz, Hooker and Karsch ^{76 ]}.

3.2 Gas cell

A gas cell is an enclosed interaction chamber filled with gas in which the laser propagates through two apertures at the cell’s entrance and exit. These small apertures mitigate the leaks, allowing lower operational density compared to gas jets, in the range of approximately 10¹⁵–10¹⁹ cm^–3, under the steady-state condition^[ Reference Osterhoff, Popp, Major, Marx, Rowlands-Rees, Fuchs, Geissler, Hörlein, Hidding, Becker, Peralta, Schramm, Grüner, Habs, Krausz, Hooker and Karsch ^{76 ]}. Simple gas cell design employs cylindrical capillaries embedded in a solid block, with gas inlets located on the side^[ Reference Genoud, Cassou, Wojda, Ferrari, Kamperidis, Burza, Persson, Uhlig, Kneip, Mangles, Lifschitz, Cros and Wahlström ^{82 ,} Reference Audet, Hansson, Lee, Desforges, Maynard, Dufrénoy, Lehe, Vay, Aurand, Persson, González, Maitrallain, Monot, Wahlström, Lundh and Cros ^{83 ]}, while the square gas cells are assembled from polished, micro-machined sapphire plates that are held within a housing. The polished surfaces are transparent to laser transmission for probing the gas or plasma density diagnostics^[ Reference Kirchen, Jalas, Messner, Winkler, Eichner, Hübner, Hülsenbusch, Jeppe, Parikh, Schnepp and Maier ^{84 –} Reference Aniculaesei, Kim, Yoo, Oh and Nam ^{86 ]}. However, these apertures and inner walls are susceptible to erosion and damage from laser misalignment and plasma, consequently requiring periodic replacement. Recent advancements in 3D-printing technology have significantly reduced the cost and time for nozzle fabrication, enabling the production of tailored and complex target geometries – such as hybrid jet–cell designs – that are otherwise infeasible to manufacture using conventional machining, thereby enhancing flexibility in LWFA experimental setups^[ Reference Döpp, Guillaume, Thaury, Gautier, Phuoc and Malka ^{87 –} Reference Andrianaki, Grigoriadis, Skoulakis, Tazes, Mancelli, Fitilis, Dimitriou, Benis, Papadogiannis, Tatarakis and Nikolos ^{91 ]}.

3.3 Capillary discharge

The capillary discharge shares similarities with gas cells, where gas is confined to a cavity. This allows uniform gas distribution and minimal turbulence. Typical operating densities range from approximately 10¹⁷ to 10¹⁹ cm^–3. When the high-voltage pulse is applied across the electrodes inserted at both ends of the capillary, this triggers an electrical discharge through the gas, producing a fully ionized plasma with a parabolic electron density profile ${n}_{\rm e}={n}_{\mathrm{o}}+\Delta n{\left(r/{r}_{\mathrm{o}}\right)}^2$ that varies radially, peaking on-axis ( $r=0$ ) and gradually decreasing outward^[ Reference Lee, Kwon, Nam, Cho, Jang, Suk and Kim ^{75 ,} Reference Kim, Jang, Uhm, Hwang, Lee and Suk ^{92 ,} Reference Kruchinin, Mondal, Sasorov, Zimmermann, Niekrasz and Molodozhentsev ^{93 ]}. Due to this density gradient, the refractive index $\eta (r)$ is highest on-axis and diminishes radially, creating a focusing force along the capillary, maintaining laser focus over extended distances to overcome both diffraction and dephasing limitations. Hence, the first measurement of laser-produced electron beams in the GeV range through a hydrogen-filled capillary waveguide by Leemans et al. ^[ Reference Leemans, Nagler, Gonsalves, Tóth, Nakamura, Geddes, Esarey, Schroeder and Hooker ^{21 ]} was demonstrated. However, this system demands sub-nanosecond synchronization between the gas injection and high-voltage discharge, significantly increasing the system’s complexity.

3.4 Modular targets

As single-stage targets approach their limits in acceleration length, the LPA community has shifted its attention to the development of the stackable, modular unit-cell design target system to increase the effective acceleration length^[ Reference Kim, Pae, Cha, Kim, Yu, Sung, Lee, Jeong and Lee ^{94 –} Reference Kroupp, Queller, Levine and Malka ^{99 ]}. The approach demonstrated by Miao et al. ^[ Reference Miao, Shrock, Rockafellow, Sloss and Milchberg ^{98 ]} uses a single-plenum modular unit-cell, specifically a 30-cm-long single-module supersonic slit nozzle fed by nine solenoid valves. This design enables the extensive stacking of acceleration stages to achieve metre-scale interaction lengths. On the other hand, Kroupp et al. ^[ Reference Kroupp, Queller, Levine and Malka ^{99 ]} demonstrated an alternative approach: a single integrated system featuring a multi-slit nozzle valve with a five-independent plena design. Both designs are highly versatile, allowing tailored density profiles over a wide range of modulations through the selection of gas species and backing pressures, while supporting both synchronous sequencing and simultaneous single-puff operation. Despite the different architecture approaches, the smoothing effect could be a significant shift towards resolving gas density stability in jet-like targets, potentially making such designs more favourable in the community.

3.5 Summary and remarks

The selection of a gas target involves a fundamental trade-off among diagnostic accessibility, fluid control, plasma stability and the desired acceleration length. As the LPA community moves towards multi-GeV energy gains, the transition from monolithic, single-stage targets to complex, multi-stage systems has become a central research priority. The primary limitation across all traditional single-stage targets is the coupling of the injection and acceleration phases. Recent advancements suggest that the ‘game-changer’ resolving this lies in advanced fluid engineering and modularity. This approach incorporates design features specifically engineered to improve gas density stability in jet-like geometries. Hybrid complex internal geometries, achieved through 3D-printed targets, could be seen as potential hardware design improvements, which, combined with active feedback loops and machine learning for real-time density control, represent the next frontier in target development. Ultimately, the choice between target architectures remains dependent on the experimental goal: a single-unit integrated system might offer high-precision stability for injection studies, while modular assemblies provide the scalability necessary for the beyond-GeV frontier. This integration must minimize technical constraints, such as alignment procedures, vacuum pressure recovery speeds that affect repetition rates, the physical footprint and the mechanical complexity of the system within the vacuum chamber.

4.1 Cryogenic liquid targets

Liquid targets are well-suited for laser experiments at HRRs due to their unique combination of fast renewability and shot-to-shot reproducibility^[ Reference Gamaiunova, Tryus, Grepl, Velyhan, Stancek, Kantarelou, Cirrone, Margarone, Giuffrida and Chagovets ^{100 ]}. Among them, cryogenic liquid targets offer a continuous, debris- and contaminant-free source of accelerated ions, which is crucial for secondary beam quality and compatibility with sensitive optical components^[ Reference Chagovets, Viswanathan, Tryus, Grepl, Velyhan, Stancek, Giuffrida, Schillaci, Cupal, Koubikova, Garcia, Manzagol, Bonnay, Souris, Chatain, Girard and Margarone ^{101 ]}. In addition, such targets are favourable for simulations due to the very high purity of the target material. The cryogenic target system developed at the ELI Beamlines (BL) Cryogenic Laboratory is designed to support kHz laser operation. It is based on a closed-cycle helium-free cryocooler coupled with a custom cell capable of maintaining temperatures down to 4 K. A cold head placed inside a helium exchange gas volume ensures low vibration and efficient thermal coupling to a cold plate, which in turn supports a brass or copper condensation cell. The cell features an exchangeable nozzle with 5–25 μm apertures, determining the liquid jet diameter. The setup enables long-term, stable, high-purity, reproducible cryogenic liquid flow extrusion. Ongoing work focuses on expanding the range of liquefied gases^[ Reference Chagovets ^{102 ]} and minimizing spatial fluctuations of the jet. Cryojets of N₂, Ar and Kr have already been obtained. Current efforts concentrate on optimizing the performance of the target. For instance, a stable 10 μm N₂ column jet with adjustable velocity (5–20 m/s) was achieved at 70 K and 0.5–7 bar, while a chamber pressure level of 10^–4–10^–3 mbar remains suitable for laser-driven ion acceleration at HRRs (see Figure 4). With a 5 μm hydrogen jet and a petawatt-class laser, protons with energy up to 80 MeV were produced, using induced pre-pulses for plasma pre-expansion^[ Reference Rehwald, Assenbaum, Bernert, Brack, Bussmann, Cowan, Curry, Fiuza, Garten, Gaus, Gauthier, Göde, Göthel, Glenzer, Huang, Huebl, Kim, Kluge, Kraft and Zeil ^{29 ]}.

Figure 4

(a) Assembled cryocell inserted into vacuum chamber. (b) Cylindrical cryogenic N₂ jet with the thickness of 10 μm. (c) Overlay of 200 images of N₂ 10 μm jet flow giving lateral fluctuations ±7 μm 2 mm away from the nozzle.

4.2 Cryogenic cluster targets

From the point of view of practical application, high-purity proton beams with HRR operation are quite advantageous. At the National Institutes for Quantum and Radiological Science and Technology, Kansai Photon Science Institute (QST-KPSI), we have developed micrometre-sized hydrogen cluster targets, which can produce impurity-free, highly reproducible and robust proton beams^[ Reference Jinno, Kanasaki, Asai, Matsui, Pirozhkov, Ogura, Sagisaka, Miyasaka, Nakanii, Kando, Kitagawa, Morishima, Kodaira, Kishimoto, Yamauchi, Uesaka, Kiriyama and Fukuda ^{30 –} Reference Jinno, Tanaka, Matsui, Kanasaki, Sakaki, Kando, Kondo, Sugiyama, Uesaka, Kishimoto and Fukuda ^{103 ]}.

Micrometre-scale hydrogen clusters were produced by expanding supercooled (25 K), high-pressure (6 MPa) molecular hydrogen gas into vacuum through a conical nozzle connected to an Even–Lavie pulsed valve, which can be operated at 600 Hz with a strong pumping system ( $\gg$ 1000 L/min).^[ Reference Even ^{105 ]}. To cool the gas temperature, the pulsed valve was attached to a compact closed-cycle helium refrigerator (see Figure 5). The temperature of the pulsed valve and the nozzle was monitored using a silicon diode temperature sensor and stabilized using two high-power thick-film resistors connected to a temperature controller. The size distribution of the micrometre-scale hydrogen clusters was evaluated by Mie scattering measurements^[ Reference Jinno, Tanaka, Matsui, Kanasaki, Sakaki, Kando, Kondo, Sugiyama, Uesaka, Kishimoto and Fukuda ^{103 ,} Reference Jinno, Kanasaki, Uno, Matsui, Uesaka, Kishimoto and Fukuda ^{104 ]}. Since the distance between each cluster is several micrometres, if a ‘skimmer’ is used, it is possible that laser pulses can hit one cluster with a short-focus off-axis parabola (OAP) (F/1.5–F/3) having a focus spot diameter of a few micrometres, although the probability of hitting is low. With a long-focus OAP (>F/10) having a focus spot diameter of tens of micrometres, the laser pulse can hit tens of clusters in one shot.

Figure 5

(a) Cryogenic micrometre-scale hydrogen cluster target generation system. (b) Conical nozzle with orifice diameter of 250 μm having a conical angle of 5 degrees connected to an Even–Lavie pulsed valve. (c) Schlieren image of sprayed hydrogen cluster target in vacuum.

Multi-MeV high-purity proton acceleration by using a hydrogen cluster target irradiated with relativistic intensity laser pulses has been demonstrated with a long-focus OAP (F/10) with the J-KAREN-P laser system at QST-KPSI^[ Reference Jinno, Kanasaki, Asai, Matsui, Pirozhkov, Ogura, Sagisaka, Miyasaka, Nakanii, Kando, Kitagawa, Morishima, Kodaira, Kishimoto, Yamauchi, Uesaka, Kiriyama and Fukuda ^{30 ]}.

As a future prospect, the 3D PIC simulation predicts the generation of directed proton beams over 100 MeV via the anisotropic Coulomb explosion^[ Reference Jinno, Kanasaki, Uno, Matsui, Uesaka, Kishimoto and Fukuda ^{104 ]} or collisionless shock dynamics inside the cluster coupled with the relativistically induced transparency effect^[ Reference Matsui, Fukuda and Kishimoto ^{106 ]}. Since the micrometre-scale clusters are less sensitive to the laser pre-pulses, it can be an attractive target for laser-driven ion acceleration aiming at over 100 MeV and other applications.

4.3 Summary and remarks

Cryogenic targets are a key enabling technology for HRR laser–matter interaction experiments because they offer debris-free operation and high material purity compared with conventional solid targets. They bridge the gap between conventional solid targets and gaseous media, offering solid-density interaction conditions with improved reproducibility and long-term stability. Within this category, two complementary approaches, cryogenic liquid targets and cryogenic cluster targets, address different experimental requirements. Cryogenic liquid targets provide a continuous, macroscopic, solid-density medium with well-defined geometry and excellent shot-to-shot stability, with access to light elements such as hydrogen as well as noble gases, resulting in high-purity proton beams. Compared to solid targets, they eliminate debris and surface contamination, while offering significantly higher effective densities than gas targets. These characteristics make them especially suitable for systematic ion-acceleration studies and for application-oriented experiments that demand long-term, HRR operation. Their main limitations stem from the complexity of cryogenic operation, residual jet instabilities, constraints on achievable target thickness and geometry, sensitivity to laser-induced hydrodynamic perturbations and vacuum compatibility, particularly at very HRRs. Cryogenic cluster targets constitute a complementary approach, enabling interactions with discrete, micrometre-scale overdense structures that are intrinsically robust against laser pre-pulses. They grant access to distinct acceleration mechanisms, including anisotropic Coulomb explosion and collisionless shock dynamics, and show strong potential for generating high-purity, high-energy protons. However, their stochastic spatial distribution reduces interaction reproducibility and limits precise control over the effective interaction volume. In summary, cryogenic liquid targets prioritize stability, reproducibility and scalability to HRRs, whereas cryogenic cluster targets emphasize robustness and access to novel laser–plasma interaction regimes. Future work should focus on enhancing flow stability and material versatility in liquid targets, and on improving control over cluster size, density and spatial distribution in cluster-based systems.