Viviana Scherini: The Neutrino Search of the Pierre Auger Observatory
The most extreme window for astronomy can be opened based on multi-messenger observation: charged cosmic rays, gamma, and neutrinos. UHE neutrinos can be produced at the sources, or during propagation. The latter we call cosmogenic neutrinos. Neutrinos could reveal the acceleration of nuclear primaries in astrophysical objects. They can test alternative models to the origin of ultra-high-energy cosmic rays.
The Pierre Auger Observatory is made up by 500 members in 18 institutions. The detector is made up by an array of 1660 cherenkov stations on a 1.5-km triangular grid, covering 3000 square kilometers. The detectors sample the lateral distribution of particles. There are four sites equipped with fluorescent detectors, which only work during clean moonless nights. These are important to calibrate the energy measurement of the full data sample, although they only have a 15% duty cycle.
There are two low-energy extensions of this array: Amiga and HEAT. The first is a denser array instrumented with muon detectors. The second are three further detectors at high elevation.
Each surface detector is a plastic tank filled with purified watered, 12 cubic meters, and equipped with three PMTs which detect the Cherenkov light. They have independent power supply and communication is provided by radio.
When a ultra-high energy neutrino comes in and hits the detector one sees a signal which is spread in time over several microseconds. Typical of a signal with a large EM component. Or, after traversing the atmosphere, the shower gets older and the EM component is absorbed. What you then see is a peak contained in a couple hundred nanoseconds. This can be used to distinguish hadronic-induced showers from neutrino-induced showers. So the signal for neutrinos is a young shower with significant EM component with a broad signal.
The detection channels are two. Earth-skimming neutrinos are almost upward-going or horizontal. This gives high interaction probability and sensitivity to tau decays occuring just above the detector. This has no background but little signal and exposure-limited. The other channel is the normal downward-going signal, which is still targeted to inclined events to be sensitive to reduce the hadronic component, zenith angles of 75 to 90 degrees.
The identification of neutrino candidates involves the selection of inclined events with significant EM component. Many stations must record a broad signal. At minimum 60% of the stations satisfy a specific trigger of time bins over threshold. Quality cuts for neutrino identification are on the elongated footprint, the speed along the footprint consistent with the speed of light, and small RMS of the speed. A further discriminating observable which is sensitive to the shape of the recorded signal is the signal integral over the peak value, again targeting the width of the signal.
The analysis strategy for downgoing candidates makes use of all the above observables, and these are combined with a Fisher linear discriminant. You want no neutrino events in the selected data, and the cut can be tuned to allow for none of it. The cut is set at <1 event of background expected in 20 years of running.
The result is that no candidates are found in the 5.5 years of data taking since 2004. With this result they can set limits on the flux as a function of neutrino energy. This is one order of magnitude higher than the flux predicted by cosmogenic models. Max sensitivity is for energies of 10^18 electronvolts. There they are competitive with IceCube.