Principles and applications of compact laser–plasma accelerators

Abstract

Rapid progress in the development of high-intensity laser systems has extended our ability to study light–matter interactions far into the relativistic domain, in which electrons are driven to velocities close to the speed of light. As well as being of fundamental interest in their own right, these interactions enable the generation of high-energy particle beams that are short, bright and have good spatial quality. Along with steady improvements in the size, cost and repetition rate of high-intensity lasers, the unique characteristics of laser-driven particle beams are expected to be useful for a wide range of contexts, including proton therapy for the treatment of cancers, materials characterization, radiation-driven chemistry, border security through the detection of explosives, narcotics and other dangerous substances, and of course high-energy particle physics. Here, we review progress that has been made towards realizing such possibilities and the principles that underlie them.

Introduction

The development of laser–plasma accelerators began in the early 1980s, inspired by the pioneering work of Tajima and Dawson1. Key to their operation is the fact that unlike the superconducting radiofrequency cavities on which conventional accelerators are based, a plasma can support immense electric fields of 100 GV m- 1 and greater, which can be generated by separating the ion and electron charges with a high-intensity laser. Static fields of this order generated in a solid target can be used to accelerate protons and ions, whereas ‘travelling’ electric fields supported by the creation of electron plasma waves by a related process can be used to accelerate lighter particles such as electrons or positrons. And even more exotic processes arising from the generation of relativistic electrons within a target can be exploited to produce not just particle beams, but novel sources of X-ray radiation.

The purpose of this review article is to explain the physical processes involved in such laser–plasma accelerators, to underline the uniqueness of the resulting particle and radiation beams and to stress their relevance for fundamental and societal applications. We will restrict the scope of our article to the development of laser-driven plasma accelerators, leaving aside plasma accelerators driven by electron or positron beams2, 3. In the remainder of this article, we first describe the physical processes involved in the generation of high-quality electron, proton and X-ray beams. We then discuss the most promising applications that have recently been demonstrated or identified and finally conclude with perspectives on beam developments. Previous review papers providing a broad coverage of the different theoretical processes involved in laser–plasma accelerators4 and of the development of experimental techniques and results in this domain5, 6, 7 can be used to gain a more thorough and indepth view of this field and its recent developments.

Electron beams

In laser–plasma electron accelerators, a longitudinal accelerating electric field is generated by the ponderomotive force of an ultrashort and ultraintense laser. This force, proportional to the gradient of the laser intensity, pushes the plasma electrons out of the laser beam path, separating them from the ions. This creates a travelling longitudinal electric field, in the wake of the laser beam, with a phase velocity close to the speed of light, most suitable for accelerating particles to relativistic energies. This electric field can reach amplitudes of several hundred gigavolts per metre. In addition, the characteristic scale length of the wakefield is the plasma wavelength, 10–30 m for electron densities ne=1018–1019 cm- 3. Consequently, if we manage to inject and accelerate electrons into a single period of the wakefield, it will lead to ultrashort electron bunches, with length shorter than the plasma wavelength. Electrons need to be injected into the wakefield with a sufficient initial energy so that they can be trapped and accelerated. Experimentally, two injection mechanisms have recently demonstrated the generation of high-quality quasi-monoenergetic electron beams. In the first mechanism, a single laser pulse is used to drive the wakefield to large enough amplitude such that electrons are injected into the rear of the first wake oscillation through transverse breaking of the plasma wave. The electrons then surf the wake and after outrunning the wave they form a monoenergetic electron bunch. This is referred to as the ‘bubble’ regime8. So far, laser parameters used in published experimental results have been unable to directly access this regime. Instead, the conditions for transverse wave breaking are eventually met as a result of laser pulse evolution as it propagates in the plasma. With current laser technology, electron beams in the 100 MeV range have been produced over millimetre distances9, 10, 11, with relative energy spreads of the order of 5–10% and a charge of hundreds of picocoulombs. A 1 GeV electron beam has been reported in a recent experiment, where the laser pulse was guided and evolved over a few centimetres in a capillary plasma discharge12. The second mechanism is based on the use of several laser pulses13. In its simplest form, the scheme uses two counterpropagating ultrashort pulses with the same wavelength and polarization. The first laser pulse, the ‘pump’ pulse, creates a wakefield, whereas the second laser, the ‘injection’ pulse, is only used for injecting electrons into this wakefield. The laser pulses collide in the plasma and their interference creates an electromagnetic beatwave pattern that preaccelerates some electrons. A fraction of these have enough energy to be trapped in the wakefield driven by the pump pulse and further accelerated to relativistic energies. Although this scheme is more complicated experimentally, it also offers more flexibility: experiments have shown that the electron beam energy can be tuned continuously from 10 to 250 MeV. The electron beam has a quasi-monoenergetic distribution with energy spread in the 5–10% range, charges in the 10–100 pC range and its parameters are stable within 5–10% . This approach is promising for the control of the electron beam parameters, and might enable tuning of both the charge and the energy spread. For instance, increasing the beam energy to the gigaelectronvolt range should decrease the relative energy spread to the 1% level. The electron bunch duration has never been measured experimentally with sufficient resolution, but simulations show that it might be shorter than 10 fs.