We are considering three plasma sources in the context of the AWAKE project.

The first one is a metal vapor source ionized by a short and intense laser pulse. It will be used for the initial AWAKE experiments because it offers the density uniformity required for the resonant driving of plasma wakefields by a self-modulated bunch much longer than the plasma wave perions (σ_z>>λ_pe).

The above figure shows the principle of the lithium heat-pipe oven [Vidal et al., J. Appl. Phys. 40(8), 3370 (1969)] that was developed for the PWFA [Muggli et al., IEEE Trans. on Plasma Science 27(3), 791 (1999)] at SLAC.

The second is an electrical discharge source, for example in argon (Ar). This source is developed by N. Lopes and his group at IST in Portugal in close colaboration with MPP. It is much simpler and affordable, and can potentially be scaled to tens of meters in length. However, it has not yet demonstrated the requires density uniformity.

The third one is a helicon source developed by O. Grulke and his group at the IPP in Greifswald, Germany. It is attractive in the long term because it can in principle be replicated and staged many times (cookie-cutter approach) to reach plasma lengths in the 100 to many 100s of meter. However, current helicon sources operate a much lower densities (~10¹³cm^-3) than required (~10¹⁴-10¹⁵cm^-3). They also require large amounts of radio-frequency (rf) power to initiate and sustain the plasma.

The metal vapor source (or a variation of it, using rubibium instead of lithium) will be used in early experiments. It provides the density uniformity required and also allows for seeding of the self-modulation instability through the ionization front located within the p⁺ bunch. However it is relatively costly due to the expensive laser needed for ionization. Also, it does not scale well to long lengths. However, it may always be used for the sort (~3-5m) length of the "self-modulator".
The next (in time and space) source may be a discharge source. This relatively simple technology source has been demontrated to work at the densities and lengths contemplated for these experiments.
The helicon source has not yet demontrated operation at the required density, however because it has no real end it can be stacked or staged many times and can therefore in principle be deployed over very long lengths.

Since the plasma frequency, the wakefield periods and the bunch self modulation frequency are all equal and in the 100-300GHz frequency range, similar techniques can be used as diagnostics. For example broadband radiation can be generated using photo-mixing of two laser beams at different frequencies. When the frequency of one or two lasers is varied, the transmission of the radiation through the plasma can be measured as a function of frequency. One can in principle determine the plasma fgrequency and temperature by measuring the cut-off at or about the plasma frequency. This diagnostic method is developped for application to the discharge and helicon sources [R. Tarkeshian et al., AIP Conf. Proc. 1507, 639 (2013)]