Beam-driven plasma wakefield accelerators (PWFA) offer a promising path towards significantly increasing the acceleration gradient of particle beams compared to conventional radiofrequency (RF) accelerators. In a PWFA, a relativistic particle beam propagating through a plasma excites a wakefield, which can then accelerate a trailing witness bunch. The blowout regime, where plasma electrons are completely expelled from the driver's vicinity, is particularly attractive due to its potential for generating high accelerating gradients and uniform focusing fields. However, current PWFA implementations rely on kilometer-scale RF accelerators to produce the high-current drive beams needed to achieve the blowout regime. This limitation restricts the widespread adoption and research of PWFA technology. Compact laser-wakefield accelerators (LWFA), on the other hand, are becoming increasingly accessible, capable of generating GeV-class electron beams with peak currents exceeding 10 kA. These high-current, short-duration beams are ideal drivers for PWFA. Combining LWFA and PWFA in a staged approach, referred to as LWFA-driven PWFA (LPWFA), is a compelling strategy to overcome the limitations of both individual technologies. An LPWFA system would potentially yield a compact high-brightness electron source for applications such as free-electron lasers. Moreover, this hybrid system provides a valuable platform for investigating fundamental PWFA physics in a more accessible setting. This paper reports on experimental demonstrations of an LPWFA system. In contrast to previous attempts that relied on uncontrolled laser pump depletion to transition from laser- to beam-driven acceleration, this work utilizes two separate gas jets for the LWFA and PWFA stages, enabling independent control and optimization of the two processes. This separation is crucial for unambiguous demonstration of beam-driven wakefield acceleration in the PWFA stage and is a key step toward realizing the full potential of LPWFAs.

Extensive research has explored both laser- and beam-driven plasma wakefield acceleration. Chen et al. (1985) demonstrated the basic principle of electron acceleration by the interaction of a bunched electron beam with a plasma. Subsequent work, such as Rosenzweig et al. (1991), focused on understanding the nonlinear dynamics of plasma wakefield generation and its impact on beam acceleration and focusing. Lotov (2004) characterized the blowout regime, a crucial operating condition for efficient energy transfer to the witness bunch. Tzoufras et al. (2008) explored beam loading effects in the nonlinear regime of plasma-based acceleration. While RF-based PWFA have shown promising results, their large-scale nature limits their accessibility. The advent of compact LWFA (Tajima and Dawson, 1979) offering high peak-current beams, such as those demonstrated by Gonsalves et al. (2019), and Li et al. (2017), Couperus et al. (2017), and Lundh et al. (2011), opened possibilities for a staged hybrid approach. Earlier works exploring the transition from laser- to beam-driven modes (Corde et al., 2011; Masson-Laborde et al., 2014; Wu et al., 2019) often relied on uncontrolled laser pump depletion. This approach limited the independent control and optimization necessary for detailed investigation and application development. Previous works also investigated the use of hybrid LWFA-PWFA systems for beam energy and brightness transformation (Hidding et al., 2010; Martinez de la Ossa et al., 2015, 2019). Studies on ionization injection techniques (Hidding et al., 2012; Martinez de la Ossa et al., 2013; Li et al., 2013; Wittig et al., 2015; Martinez de la Ossa et al., 2017; Zhang et al., 2019) and driver beam instabilities (Mehrling et al., 2017; Martinez de la Ossa et al., 2018) further enriched the understanding of PWFA physics. Direct observation of beam-driven plasma wakefields and induced ion motion have also been reported (Gilljohann et al., 2019). High-efficiency acceleration in RF-based PWFA has also been demonstrated (Blumenfeld et al., 2007; Litos et al., 2014).

Two complementary experiments were conducted using two different 100 TW-class short-pulse laser systems: DRACO at Helmholtz-Zentrum Dresden-Rossendorf (HZDR) and ATLAS at Ludwig-Maximilians-Universität (LMU) München. Both experiments used a staged approach with distinct LWFA and PWFA stages. **Experiment 1 (High-Gradient LPWFA):** * **LWFA Stage:** A 3 mm long helium gas jet (doped with 3% nitrogen) was used. The self-truncated ionization-induced injection scheme was employed to generate a high peak-current drive electron beam. The DRACO laser delivered 1.7 J pulses (after accounting for pre-ionization and probing pulses) focused to a 19.5 µm FWHM spot size. * **PWFA Stage:** A 3 mm long hydrogen gas jet (doped with 10% helium) was placed directly behind the LWFA stage, with a 12.5 µm thick steel foil positioned at the entrance to reflect the spent laser pulse and allow the electron beam to pass through and drive the wakefield. The PWFA stage was either self-ionized by the driver beam or pre-ionized by a counter-propagating laser pulse. * **Diagnostics:** Electron beam spectra were measured using a permanent magnet spectrometer. Plasma wakefields were visualized using shadowgraphy with a synchronized few-cycle laser pulse. **Experiment 2 (Drive-Witness Bunch Pair):** * **LWFA Stage:** A 5 mm long hydrogen gas jet was used. The shock-front injection technique was optimized to produce a controlled drive-witness electron bunch pair with a separation of approximately one plasma wavelength. * **PWFA Stage:** A 1 mm long hydrogen gas jet was placed downstream, with a 6 mm vacuum gap between the stages to reduce the laser intensity on the PWFA stage. The PWFA stage was pre-ionized by the remnant laser pulse. The plasma density in the PWFA stage was adjusted to control the witness beam's acceleration phase. * **Diagnostics:** Electron beam spectra were measured using a permanent magnet spectrometer. The plasma wavelength and density profiles were independently confirmed. **Simulations:** Particle-in-cell (PIC) simulations using PIConGPU and FBPIC codes were used to model the LPWFA processes in both experiments. The simulations included realistic parameters from experiments (gas density, laser profiles, etc.).

**Experiment 1:** * In the self-ionized PWFA stage, witness beams were observed in 28% of shots with an average energy of 62 ± 4 MeV. * In the pre-ionized PWFA stage, witness beams were observed in 37% of shots, with a significantly higher average energy of 100 ± 5 MeV, and a maximum energy of 128 MeV. * An effective accelerating gradient of ~50 GV m⁻¹ (a conservative lower limit) was estimated in the pre-ionized case. * A driver-to-witness energy transfer efficiency of 2.9% was achieved, exceeding previous results in RF-based PWFAs by a factor of 5. * PIC simulations using PIConGPU code reproduced the experimental observations, showing peak accelerating fields exceeding 100 GV m⁻¹. **Experiment 2:** * A controlled drive-witness electron bunch pair was generated in the LWFA stage using the shock-front injection technique. * The witness bunch was accelerated to 133 ± 1 MeV in the PWFA stage (14 MeV energy gain). This demonstrates a controlled energy transfer from the drive to the witness bunch. * A charge capture efficiency of close to 70% was estimated which is significantly higher than in previous RF-based PWFAs. * Simulations using FBPIC code supported the experimental findings, showing that the laser remnant primarily acted as a pre-ionizer for the PWFA stage.

The results clearly demonstrate the successful acceleration of witness electron beams in two independent LPWFA experiments. The independent control of the LWFA and PWFA stages through the use of separate gas jets and a laser blocker (or increased stage separation) was crucial for unambiguously attributing the witness beam acceleration to the beam-driven plasma wakefields. The significantly higher witness beam energies observed in the pre-ionized scenario compared to the self-ionized scenario, along with consistent driver beam degradation, strongly supports this conclusion. The achievement of a high energy transfer efficiency and high charge capture efficiency in the dual-bunch experiment highlights the potential of LPWFAs for generating high-quality electron beams. The compact nature of the LPWFA system greatly enhances the accessibility of PWFA research and opens up new avenues for advancing the field. The ability to independently control and optimize both the LWFA and PWFA stages allows future experiments to explore various advanced injection schemes tailored for high-quality beam generation. The high wakefield amplitudes generated in LPWFA enable the implementation of advanced injection methods, potentially leading to ultra-high-brightness electron beams for applications like compact free-electron lasers.

This work successfully demonstrated a compact, millimeter-scale plasma accelerator driven by laser-accelerated electron beams. The experiments achieved significant electron beam acceleration with high energy transfer efficiency and charge capture efficiency. The ability to independently control the LWFA and PWFA stages through separate gas jets allows for precise tuning of the acceleration process and opens up the potential for advanced injection techniques. This development is highly significant for advancing PWFA research and its applications, particularly in creating compact sources of high-brightness electron beams for diverse applications including free-electron lasers. Future research will focus on improving several aspects of the staging concept, including mitigating emittance growth and optimizing the length and density of the PWFA stage.

While the experiments demonstrated the feasibility of the LPWFA concept, several limitations should be noted. The shot-to-shot fluctuations in beam parameters were relatively large, especially in the self-ionized PWFA cases, potentially influencing the accuracy of the results. The steel foil used in Experiment 1 introduced some beam divergence and energy spread, which could be mitigated with alternative laser blocking methods. Further optimization of the gas jets and beam focusing would enhance the overall efficiency and beam quality. While the simulations accurately reproduce the experimental observations qualitatively, quantitative discrepancies exist which warrant further investigation and code refinements to fully capture the nuances of the experimental conditions.