Notable improvements on LWFA through precise laser wavefront tuning

概要

Title

Notable improvements on LWFA through precise
laser wavefront tuning

Author(s)

Oumbarek Espinos, Driss; Rondepierre, Alexandre;
Zhidkov, Alexei et al.

Citation

Scientific Reports. 2023, 13, p. 18466

Version Type VoR
URL
rights

https://hdl.handle.net/11094/93325
This article is licensed under a Creative
Commons Attribution 4.0 International License.

Note

Osaka University Knowledge Archive : OUKA
https://ir.library.osaka-u.ac.jp/
Osaka University

www.nature.com/scientificreports

OPEN

Notable improvements on LWFA
through precise laser wavefront
tuning
Driss Oumbarek Espinos 1,2,4*, Alexandre Rondepierre 1,2,4, Alexei Zhidkov 1,2,
Naveen Pathak 1,2, Zhan Jin 1,2, Kai Huang 2,3, Nobuhiko Nakanii 2,3, Izuru Daito 3,
Masaki Kando 2,3 & Tomonao Hosokai 1,2
Laser wakefield acceleration (LWFA) continues to grow and awaken interest worldwide, especially
as in various applications it approaches performance comparable to classical accelerators. However,
numerous challenges still exist until this can be a reality. The complex non-linear nature of the process
of interaction between the laser and the induced plasma remains an obstacle to the widespread
LWFA use as stable and reliable particle sources. It is commonly accepted that the best wavefront is
a perfect Gaussian distribution. However, experimentally, this is not correct and more complicated
ones can potentially give better results. in this work, the effects of tuning the laser wavefront via
the controlled introduction of aberrations are explored for an LWFA accelerator using the shock
injection configuration. Our experiments show the clear unique correlation between the generated
beam transverse characteristics and the different input wavefronts. The electron beams stability,
acceleration and injection are also significantly different. We found that in our case, the best beams
were generated with a specific complex wavefront. A greater understanding of electron generation
as function of the laser input is achieved thanks to this method and hopes towards a higher level of
control on the electrons beams by LWFA is foreseen.
Since the conception of laser wakefield acceleration (LWFA)1 this technique has vastly advanced, to the point
that its use as electron beam source for free electron l­ aser2, notorious for its strict requirements, has been already
­achieved3,4, paving the way to more future uses. In LWFA, a high intensity fs laser propagates inside a gas,
ionizing it and expelling the plasma electrons from its path via the ponderomotive force. An electronless area
(wake) is created behind the laser in which acceleration gradients of up to hundreds of GV/m can be achieved.
The high degree of non-linearity in the components of a LWFA complicates the control, and even the capacity
to understand which conditions to aim, to obtain electron beams capable to equal and surpass the ones of classic accelerators. Simplifying such system to its core components leaves two main non-trivial parts, i.e., the laser
system and the gas target.
Regarding the target, its selection is determined by the kind of used LWFA technique, e.g. ionization i­ njection5
can be done with a flat gas distribution. In addition, mixed ­gases6 substantially improve (main gas + doping) the
technique when in the right proportions. Colliding p
­ ulse7 needs a careful calculation of the space occupied by the
gas and its uniformity to better control both lasers path and interaction time. Shock ­injection8,9 depends critically
on the shock parameters (shock density, position, etc) and the following ramp as well as the laser focusing position to assure a controlled and localized i­ njection10. Thanks to its controlled injection in the wake (low energy
spread), one can separate it from the acceleration part without as much complexity as the colliding scheme.
In all cases, the driver of the LWFA process is the laser pulse. The interaction of a high intensity perfect laser,
e.g., Gaussian, with a mm length plasma is already not a simple process, and therefore, taking into account
a realistic laser beam, which can be quite far from the perfect Gaussian case regarding its phase distribution
­(aberrations11), its beam quality factor (M 2  ) and hence its near and far field intensity pattern can substantially change the laser-plasma interaction. The inclusion of aberrations and real near field patterns (often a

1

Institute of Scientific and Industrial Research (SANKEN), Osaka University, Mihogaoka, Ibaraki, Osaka 565‑0871,
Japan. 2Laser Accelerator R &D, Innovative Light Sources Division, RIKEN SPring-8 Center, 1‑1‑1, Kouto,
Sayo‑cho, Sayo‑gun, Hyogo, Osaka  679‑5148, Japan. 3Kansai Institute for Photon Science (KPSI), National
Institutes for Quantum Science and Technology (QST), 8‑1‑7, Umemidai, Kizugawa, Kyoto  619‑0215,
Japan. 4These authors contributed equally: Driss Oumbarek Espinos and Alexandre Rondepierre. *email:
doumbarek@sanken.osaka-u.ac.jp
Scientific Reports |

(2023) 13:18466

| https://doi.org/10.1038/s41598-023-45737-5

1
Vol.:(0123456789)

www.nature.com/scientificreports/
non-homogeneous flat top distribution) greatly affects the laser intensity distribution along its ­propagation11,
crucial in the wakefield creation and dynamics. On top of this, the multiple processes occurring inside the plasma
­(etching12, self-focusing13, ­filamentation14, etc) originates a highly complex evolution of the laser propagation
compared to vacuum conditions. Therefore, in experiment, where laser and gas are not perfect, the LWFA process
shows some shot-to-shot instabilities and also puts into question the superiority of aiming for a perfect Gaussian
laser pulse when interacting with an imperfect gas target as already hinted in some ­works11,15–17. Beaurepaire
et al.16 presents a qualitative comparison between two simple laser transverse distributions, reconstructed from
experiment, with real and flat wavefront. However, the conclusion is that the wavefront is important for LPA
but lacks any deeper study. Similarly, Ferri et al.18 measure their non-Gaussian transverse laser wavefront and
phase. Through simulations they show that using an initial Gaussian or a non-Gaussian wavefront give up to a
23% difference in the number of photons emitted by the plasma. Nevertheless, the simulations do not take into
account a full characterization of the laser pulse, thus making its longitudinal evolution incorrect, and from this
work the authors conclude that “improving the laser spot quality would also lead to an important benefit”, thus
coming back to the belief that every situation could be improved with a Gaussian ideal laser beam. Lin et al.19
through a semi-random optimization (genetic algorithm) applied to a deformable mirror and a change in mid-IR
laser focus position finds mainly the possibility of increasing the charge of the low energy electrons (1–4 MeV)
however, the same inaccuracy on the simulations is committed and the LPA configuration is quite different (low
energy, huge divergence, high charge) from the other works, including ours. Another thing in common in these
works is the use of the ionization injection scheme without gas density tailoring.
In this work, we show how by using a complex understood wavefront intensity distribution, defined by the
Zernike polynomial terms, a higher quality electron beam has been achieved with respect to the “no aberrations”
case (which is the usual target) and a possibility of simultaneous multiple electron beam generation. First, we present the effects on the LWFA electron beam generation of tuning the laser pulse transverse intensity distribution
in the vicinity of the waist position by adding aberrations in a controlled fashion. We demonstrate the capabilities
to change and even improve the beam characteristics without altering anything on the target side. Furthermore,
we explore the distinctive electron beam patterns observed related to different aberration configurations.

Results

Configuration

For this experiment, the second beam line of the LAPLACIAN (Laser Acceleration Platform as a Coordinated
Innovative Anchor) facility, located at RIKEN SPring-8 Center, was used. For the LWFA, a 800 nm, 23 fs full
width half maximum (FWHM) laser with 0.7 J on target energy interacts with a pure He gas target under vacuum.
The laser aberrations are modified by a deformable mirror that uses the measurements of a wavefront sensor as
feedback. A F/20 parabola focuses the laser onto the gas target with a waist diameter of 20 µm . The gas target is
prepared in the shock injection configuration as seen in Fig. 1, with a 4 mm long conical supersonic gas jet 4.5
mm under the laser axis and a simple blade 3.5 mm above the jet and with ≈ 20% coverage (percentage of the
gas distribution covered by the blade in the longitudinal ­direction20). The gas density is set to around 2 × 1018
cm−3 . The generated electron beams transverse distribution and relative charge are observed in a beam monitor
955 mm after the gas jet and their energy and charge on an electron spectrometer positioned 2 m after the gas jet.

Concept

When a laser pulse is focused, the pattern evolution from its origin up to the focal plane depends on the amplitude
and phase patterns. For a simple Gaussian beam free of aberrations, the beam is focused without any disturbance
and will remain Gaussian at every position. However, when either a phase is added or the initial profile is differing from the Gaussian one, the beam pattern will become more c­ omplex11. Figure 2 shows the beam intensity
pattern 2 mm around the focal plane during propagation in-vacuum when using the experimental laser: intensity
pattern at LAPLACIAN with a flat phase (obtained after correction with a Strehl Ratio of 0.92), and considering
an additional phase error. The added phase were /20 of both trefoil 0 and 45 (Fig. 2a), /20 of first order astigmatism 0 (Fig. 2b) and /14 of second order astigmatism 0 (Fig. 2c). ...

この論文で使われている画像