Ioan Dancus

We present a hybrid self-supervised model that integrates a convolutional autoencoder with a Guassian mixture model (GMM) to classify and reconstruct real petawatt laser beam profiles. The model distinguishes between near-field, far-field, and multi-spot beam profiles, demonstrating that self-supervised representations can extract physically meaningful features from unlabeled experimental laser data sets. We achieve 89.4 % classification accuracy on our baseline model, improving the previously unsupervised baseline of 85 %, under comparable experimental conditions. The integration of a probabilistic layer GMM enables more precise clustering of laser beam profiles, achieving 98.9% classification accuracy. Designed for an inference speed up to 1 Hz for live diagnostics on the ELI-NP high power laser system, our architecture enables real-time state detection. Importantly, we show that high classification performance can emerge even when pixel-level reconstructions are imperfect, indicating that the learned latent representations prioritize physically meaningful structure over image fidelity.

High-power laser beamlines typically operate with fixed focusing conditions, limiting the focal spot size and peak intensity. To mitigate these restrictions, prior studies used curved plasma mirrors to adjust the F-number to a specific value. Here, a double-plasma-mirror (DPM) system including spherical optics in a telescope configuration is implemented to adapt the F-number of a multi-petawatt (PW) laser beam resulting in adjustability within a range of intensities. The system is optimized to minimize focal aberrations. A dedicated imaging system is used to evaluate focus quality and the DPM reflectivity at the multi-PW level. Temporal contrast enhancement of the reflected beam is additionally demonstrated, as evidenced by higher particle yield and proton kinetic energy from nanometer-thick foils, compared to results without DPMs. These findings enable multi-PW laser facilities to explore more extreme laser–plasma conditions that require ultra-high temporal contrast and intensity, while expanding their capabilities in intensity adjustment beyond designed specifications.

High-repetition-rate targets present an opportunity for developing diagnostic tools for on-demand calibration at high-power laser facilities for consistent performance and reproducibility during experimental campaigns. The non-linear change in transmission associated with a laser-driven plasma mirror, based on high-repetition rate targets, has been used in a Frequency Resolved Optical Gating (FROG) configuration to analyze the spectral phase for near-infrared pulses far from the Fourier limit. Three types of targets were compared for characterizing pulses in the 1–8 ps range: a glass slide, a polymer tape, and a thin liquid sheet created by two impinging micrometer-scale jets. The thin liquid film had the best mechanical stability and introduced the least spectral distortion, allowing the most robust reconstruction of the temporal intensity profile. The spectral phase was reconstructed using a non-iterative algorithm, which reproduced the second-order phase distortions induced with an acousto-optic programmable dispersive filter with an RMS error of 6.2%, leading to measured pulse durations with an RMS deviation ranging from 1% for pulses of 6.8–7.8 ps up to 7.5% for pulses around 1 ps.

High-power laser systems have opened new frontiers in scientific research and have revolutionized various scientific fields, offering unprecedented capabilities for understanding fundamental physics and allowing unique applications. This paper details the successful commissioning of the 1 PW experimental area at the Extreme Light Infrastructure–Nuclear Physics (ELI-NP) facility in Romania, using both of the available laser arms. The experimental setup featured a short focal parabolic mirror to accelerate protons through the target normal sheath acceleration mechanism. Detailed experiments were conducted using various metallic and diamond-like carbon targets to investigate the dependence of the proton acceleration on different laser parameters. Furthermore, the paper discusses the critical role of the laser temporal profile in optimizing proton acceleration, supported by hydrodynamic simulations that are correlated with experimental outcomes. The findings underscore the potential of the ELI-NP facility to advance research in laser–plasma physics and contribute significantly to high-energy physics applications. The results of this commissioning establish a strong foundation for experiments by future users.

Achieving complex pulses with high-power lasers necessitates rigorous testing of specially designed optical components. The qualification of these components using complementary devices to access both the high-resolution and the large-aperture properties, followed by validation using propagation simulations, is proposed here. In particular, the topology of a large-aperture staircase-like Fresnel phase plate used to generate vortex pulses is qualified using a non-contact optical profiler and a large-aperture wavefront measurement setup based on a Shack–Hartmann sensor. The resulting topography is further used for simulating the focus of laser beams after passing through the phase plate. Step height distribution effects on the doughnut-shaped focus are identified, and avoiding the indicated pitfall in the design of the phase plate provides at least a 10-fold reduction of the irradiance modulation on the circumference of the focus in the super-Gaussian case.

Simultaneous ultra-intense pulses at petawatt laser facilities enable a broad range of experiments in nuclear photonics and strong field quantum electrodynamics. These experiments often require very precise control of the time delays between pulses. We report measurements of the time delay between the two 1 PW outputs of the Extreme Light Infrastructure - Nuclear Physics (ELI-NP) facility in Romania. The short-term standard deviation of the time delay was approximately half of the pulse duration of 23 fs, and the average delay drifted with up to 100 fs/h. The drift and sporadic delay jumps were corrected using a feedback loop, which reduced the long-term standard deviation of the delay close to its short-term value. These results imply that in ELI-NP experiments using two simultaneous pulses, a temporal overlap of better than half of the pulse duration can be achieved for more than two thirds of the shots, which would enable high data rate experiments using simultaneous petawatt pulses.

The high power laser system at Extreme Light Infrastructure—Nuclear Physics has demonstrated 10 PW power shot capability. It can also deliver beams with powers of 1 PW and 100 TW in several different experimental areas that carry out dedicated sets of experiments. An array of diagnostics is deployed to characterize the laser beam spatial profiles and to monitor their evolution during the amplification stages. Some of the essential near-field and far-field profiles acquired with CCD cameras are monitored constantly on a large screen television for visual observation and for decision making concerning the control and tuning of the laser beams. Here, we present results on the beam profile classification obtained from datasets with over 14 600 near-field and far-field images acquired during two days of laser operation at 1 PW and 100 TW. We utilize supervised and unsupervised machine learning models based on trained neural networks and an autoencoder. These results constitute an early demonstration of machine learning being used as a tool in the laser system data classification.

We are showing a significant enhancement in the temporal contrast by reducing the coherent noise of the 10 PW laser system at the Extreme Light Infrastructure - Nuclear Physics facility. The temporal contrast was improved by four orders of magnitude at 10 picoseconds and by more than one order of magnitude at 50 picoseconds before the main peak. This improvement of the picosecond contrast is critical for the experiments using thin solid targets.

We report on the method used to tune the 10 PW High Power Laser System parameters for optimized laser-target interaction at the ELI-NP showing the possibility to reach the 10²³ W/cm² irradiance level.