Accelerators use electric fields to accelerate charged particles (electrons, positrons, protons, ions, etc.). For particles to reach high energies, the electric field needs to be in the direction of the particles motion. However, electromagnetic (em) waves have fields mostly perpendicular to their direction of propagation, and in general, particles cannot gain large amounts of energy by interacting, for example, with a high-power laser beam.

In a conventional, radio-frequency (rf) accelerator high power microwave sources feed resonant rf cavities with metallic or superconducting walls. Typical rf frequencies are between 0.5 and 30 GHz). The cavities dimensions are of the order of the microwaves wavelength (60 to 1 cm linear dimensions). The cavities serve two important purposes:

- A resonant cavity can store a large amount of energy, and therefore contain fields many times larger that those produced by the microwave sources themselves.

- In the cavity, the microwave electric field can be made to lie along the cavity axis, which is also chosen to be the direction of propagation of the particles to be accelerated. The cavity eigen modes are either transverse magnetic (TM or E, with their magnetic field component in the directions perpendicular to the cavity axis) or transverse electric (TE or H, with their electric field component in the directions perpendicular to the cavity axis). The cavity dimensions are chosen such that the lowest order TM mode is resonant at the desired microwave frequency.

However, the large electric fields stored in the resonant cavities lead to the first limitation encountered with rf accelerators, known as rf-breakdown. Where there are microscopic imperfections (peaks and pits) in the cavity walls, the electric field is multiplied by the local surface curvature effect, and electrons can be liberated from the cavity walls. These free electrons can be accelerated by the rf field, hit the walls again, and generate more free electrons by secondary emission, as well as free ions. This local plasma destroys the cavity resonance, and the energy of the plasma can be dissipated on the surface of the cavity, creating more and larger imperfections. This arcing phenomena limits the electric field in the cavities and therefore the rate at which particles can be accelerated, and may lead to long term degradation of cavities. Higher gradients can in principle be achieved by using higher frequency microwaves. However, operating at higher resonant frequency cavities leads to the second limitation encountered with rf accelerators: the fabrication tolerances become more and more difficult to maintain since the cavities dimensions are of the order of the rf wavelength. Also, the expected increase in breakdown fields limit expected at higher frequencies has so far not been observed experimentally. The maximum measured accelerating field in rf cavities is currently smaller than 200 MV/m (H. Braun et al., Phys. Rev. Lett. 90, 224801, 2003). The maximum gradient is also limited by long term damage to the cavity metal caused by pulsed heating (D. Pritzkau and R. Siemann, Phys. Rev. ST Accel. Beams 5, 112002, 2002).

Plasma-based accelerators can in principle overcome many of the limitations of conventional accelerators. Plasmas are made of gases or vapors that are ionized. Therefore, the the first (or few first) electron(s) are already free, and the ionization potential for the subsequent electrons is larger that that for the first (few) electron(s). Plasmas can therefore sustains larger field than metallic walls before additional electrons are released. Extremely large longitudinal (accelerating) fields (>10 GV/m)have been excited in plasmas. These fields are those of plasma waves or wakes driven by high-intensity laser pulses or high-current particle bunches. In a plasma accelerator the accelerating structure or the plasma do not have to exist before the drive beam is present. The accelerating structure (or "cavity") is sustained by the plasma electrons and exists only for one (or a few) plasma period(s). Since there is no cavity to fabricate, plasma accelerators can operate at much higher frequencies (100 GHz to many THz!) than rf accelerators (0.5 to 30 GHz). One of the key parameters of a plasma accelerator is the plasma electron density n_e. The frequency of the accelerator is given by the plasma angular frequency ω_pe=(n_ee²/ε₀m_e)^1/2: with n_e=1×10¹⁴ cm^-3, f_pe=ω_pe/2π~90 GHz, while with n_e=1×10¹⁸ cm^-3, f_pe~9 THz, and the accelerating structure size are correspondingly small, of the order of the plasma collisionless skin depth c/ω_pe or ~530 and ~5 μm respectively. The accelerating field of the plasma wake can be estimated from the wave breaking field amplitude E~E_WB=m_ecω_pe/e (J.M. Dawson, Phys. Rev. 113, 383, 1959). These fields are ~1 to ~100 GV/m for the above mentioned plasma densities. Most current particle-beam driven plasma accelerators (also known as Plasma Wakefield Accelerators or PWFAs) use a single bunch to both drive the wake and experience the acceleration. In this case, the largest accelerating gradient is expected to be reached when the plasma wake wavelength is approximately equal to the bunch length. In the linear theory for a longitudinal Gaussian bunch current profile (with an rms width σ_z) this condition can be expressed as k_peσ_z~2^1/2 (k_pe=ω_pe/c). Since the drive beam already has a high energy and is ultra-relativistic, the dephasing of the particles in the accelerating structure is usually much longer that the accelerator length itself.

Recent experiments using the Stanford Linear Accelerator Center (SLAC) beam with an energy of 28.5 or 42 GeV and 1.2-1.8×10¹⁰ particles per bunch have produced very important results:

-The acceleration of electrons by more than 42 GeV over a plasma length of ~90 cm ...

-... and therefore the sustained excitation of an accelerating gradient of more than 45 GV/m over ~90 cm of plasma with n_e=2.6×10¹⁷ cm^-3.

-The first demonstration of the acceleration of positrons by a plasma (B.E. Blue et al., Phys. Rev. Lett. 90, 214801, 2003).

A characteristic of these proof-of-principle, single-bunch experiments is that the bunch particles cover all the phases of the accelerating structure, leading to a very wide energy spread. One of the next milestones in PWFA experiments will therefore be the demonstration of the acceleration of a particle bunch with a finite energy spread. This can naturally be done by using a drive bunch to excite the plasma wake, followed by a witness bunch that experiences the acceleration. Another milestone will be the acceleration of a positron bunch in a meter-long plasma with a multi-GV/m accelerating gradient. These kinds of experiments will be performed in the near future at the SLAC SABER facility.

An extensive list of publications related to these experiments can be found in the Publications page. Find more details about recent results here.

In the plasma afterburner scheme recently proposed (S. Lee et al., Phys. Rev. ST Accel. Beams 5, 121301 (2002)), short (when compared to the original accelerator) sections of plasma are placed after the conventional accelerator to double the energy of a witness beam. This scheme could be used to either double the energy reach of a collider without doubling its length, or to double the effective accelerating gradient of a collider thereby significantly reducing its size. the afterburner uses high-gradient plasma wakefield acceleration (PWFA). The plasma wake is excited by a train of low quality drive bunches, while a high-quality train of witness bunches, that will be collided, are accelerated. The emittance and energy spread of the drive bunch train can be modest because this train is disposed of before the collision point. This allow for a choice of parameters suitable to maximize the quality and stability of the plasma wake. The witness bunch train follows the drive bunch train and has the parameters required for the high-luminosity required for particle physics studies. While the afterburner concept is in its early research state, its potentials are very attractive, and significant experimental progress has been made recently. In particular, high-gradient acceleration (>50GV/m) of electrons leading to very high energy gain (>42GeV) has been demonstrated (I. Blumenfeld et al., Nature 445, 741, 15 February 2007). Acceleration of positron in plasmas has also been demonstrated, although at low gradient (70MV/m) has also been demonstrated (B.E. Blue et al., Phys. Rev. Lett. 90, 214801, 2003). The low gradient is the result of the non-availability of short positron bunches. Short positron bunches appropriate for high gradient acceleration experiment will be available soon at the SLAC FACET facility that was recently proposed. It seems more economical to build a beam-driven linear collider (or PWFA-LC) based on a multi-stage scheme rater than on an energy doubler scheme. In that new scheme a train of 25GeV bunches drive successsive meter-long plasma stages where the witness bunch gains 25GeV per stage. Preliminary studies reveal that a high quality witness bunch suitable for high energy physics can be produced with high efficiency. The next significant experimental steps toward the realization of a plasma-based linear collider include:

- High gradient acceleration of short positron bunches

- Accelleration of electron and/or positron bunches with a narrow energy spread

- Demonstration of the preservation of the incoming beam emiitance in conjunction with large energy gain

- Optimization of the acceleration process including loading of the plasma wake by the witness bunch