For linear accelerators like the planned International Linear Collider, the acceleration gradient is the key parameter which determines how long the machine has to be to achieve a given energy. The length is closely related to the overall cost of such a machine, so achieving higher acceleration gradients is a way of getting more bang for the buck. A promising technology for this is plasma wakefield acceleration. In a conventional accelerator, particles are accelerated by the electric field in radio frequence resonators. The maximum field in these resonators is limited by breakdowns on the cavity walls due to the high fields.
This is where the plasma accelerators come in: In a plasma extreme fields on short distances can be achieved, without the risk of breakdowns, by separating electrons from the ions. In so-called beam driven plasma wakefield accelerators this is achieved by shooting a highly compressed, high energetic beam of charged particles into the plasma. The strong electric field of this beam expells the plasma electrons from its path, leaving the much heavier positively charged ions behind, creating a positively charged bubble (see the figure below). The electrons move back towards this bubble behind the driving beam, starting oscillations, creating bubbles of decreasing strenght in the wake of the beam, thus the name of this effect. By placing a bunch of electrons in the rear part of the bubble right before the “wall” of electrons, where the electric field is strongest, they can be accelerated. So these electrons surf the plasma wave, gaining energy like a surfer on a real ocean wave. The noteable difference is that since all particles involved here are relativistic, they are all moving essentially at the speed of light, so the electrons can gain energy without running ahead of the wave.
While this technology is still far from maturity, and applications in accelerators for particle physics are still far away, spectacular successes have already been achieved. Using electrons from the Stanford Linear Accelerator, a doubling of the energy of some of the electrons was achieved over a distance of less than one meter in 2007, an energy increase for which the conventional SLAC machine needs a length of 2 miles. The accelerating gradient was about 50 GV/m, more than a factor of 1000 more than the state-of-the art ILC acceleration structures achieve today.
Apart from the fact that this technology is not ready for use yet, there are other limitations. One issue is that basic principles forbid to give more than twice the energy of the driving bunch to the particles that are accelerated, even in the limit of very few particles. And the highest energy electron accelerator at the moment is at SLAC, with an energy of 50 GeV, limiting the achievable energy increase to 100 GeV in one accelerating stage, far from what is needed to construct a machine at the energy frontier. So either use several plasma acceleration stages behind each other, which is probably quite a challenge, or use particles with much higher energy to excite the plasma wake. Even today, protons at energies of 1 TeV (at the Tevatron) and 7 TeV (at the LHC) are available, which can potentially lead to spectacular electron energies if they can be used as drive beams for a plasma accelerator.

Plasma bubble created by the highly relativistic proton bunch (red), used to accelerate electrons (yellow).
This idea was studied in simulations by a team from the MPI for Physics, the University of Duesseldorf and from the Budker Institute, including myself. Our studies show that it is possible to achieve electron energies of 600 GeV using a 1 TeV proton beam shot into a 400 meter long plasma cell. The figure illustrates how this proton driven plasma wakefield acceleration works, with the electrons riding the plasma wave excited by the protons. Of course, many questions are still unsolved, for example how to get the very short proton bunches needed to make the plasma wake work. Just to set the scale, LHC proton bunches are about 7 cm long, while bunch lengths of 100 micro meters are needed, almost a factor 1000 shorter.
Our results were published last Sunday in Nature Physics, you can get the article here, provided you have a valid subscription. It was also discussed on wired science. Lets see how this technology will develop in the future, it might well be a way to keep the journey to ever higher energies in particle physics going on.