1: \section{Tau events simulation}
2: \subsection{Interactions in earth}
3: A Monte-Carlo technique has been used to simulate the tau neutrino or charged lepton
4: interactions and propagation inside the Earth.
5: The lepton may interact several times through deep inelastic scattering, changing charge in most cases,
6: or eventually decay, but, in all cases, a tau neutrino or charged lepton is
7: present in the final state. Some energy is lost at each interaction,
8: as well as continuously along the paths. However, in our energy range, the
9: initial direction of the incoming neutrino is always
10: conserved (Figure~\ref{simul}).
11: \begin{figure}[!t]
12: \begin{center}
13: \includegraphics[bbllx=1,bburx=548,bblly=210,bbury=490,width=8cm,clip=]{simul.eps}
14: \end{center}
15: \vspace*{-3.0cm}
16: \caption{Chain of interactions producing an observable shower.}
17: \vspace*{0.5cm}
18: \label{simul}
19: \end{figure}
20: \par In our model, we assumed an isotropic incident flux of
21: neutrinos, an homogeneous Earth with density 2.15 g/cm$^3$~\cite{PDG}, and
22: a very high energy parameterization of the charged current (CC) cross section
23: accurate to within 10\% as given by the results of the CTEQ4-DIS parton distributions~:
24: \[
25: \sigma_{\,cc}^{\,\nu N}=
26: 1.0\,\left(\frac{E_\nu}{1\,EeV}\right)^{\!0.363}\!\!\!\!\!10^{-32}\mbox{\text cm}^2
27: %5.53\,\left(\frac{E_\nu}{1\,GeV}\right)^{\!0.363}\!\!\!\!\!10^{-36}\mbox{\text cm}^2
28: \]
29: %\subsection{Interactions in earth}
30: A step-by-step method is used; at each step, the probabilities of
31: different interactions and of the decay are evaluated as functions of the
32: energy. Both CC and NC interactions are taken into account, with
33: the cross section given above (and $\sigma_{\,nc}=0.4\,\sigma_{\,cc}$)~\cite{Gandhi1998} ;
34: the energy of the outgoing lepton is computed using a parametrization of the inelasticity at 1~EeV.
35: %the energy of the outgoing lepton is set to 80\% of the incoming one.
36:
37: \par
38: Energy losses have been calculated including Bremsstrahlung (BS) and Pair
39: Production (PP) as well as Deep Inelastic Scattering (DIS).
40: The energy loss model is of the form~:
41: \[
42: -\frac{dE}{dx} = a + b(E)E
43: \]
44: where the second term is dominant above a few 100 GeV.
45: \par
46: Contributions from BS and PP have been rescaled from the muon values given
47: in reference~\cite{PDG} and~\cite{Ginneken}, leading to $b = 0.08\times10^{-7}$ and
48: $1.4\times10^{-7}$g$^{-1}$cm$^2$ respectively. We then obtained an
49: attenuation length $L=(\rho\sum b)^{-1}$ of 31~km.
50: DIS contributions rely on parameterization of the photo-nucleon cross sections
51: as well as on the proper modelisation of the nucleon structure functions at very low $x$
52: and/or very large $Q^2$. This subject being still tentative, we decided to use
53: two different estimates, an energy independant contribution
54: (DIS-low, $b = 10^{-7}$g$^{-1}$cm$^2$) rescaled from the muon behavior given in~\cite{Ginneken}
55: and an energy dependant one (DIS-high, $b = 6\,E_{18}^{0.2}\times 10^{-7}$g$^{-1}$cm$^2$
56: which dominates energy losses above $10^{15}$eV) as a parameterisation of the
57: recent calculation from~\cite{Dutta}. This later case gives an attenuation length as low as
58: 6~km at \E{18} strongly reducing the penetrating power of the tau.
59:
60: %The cross section used for continuous radiative electromagnetic
61: %interactions are based on QED calculations for bremsstrahlung, pair
62: %production and photo-nuclear interactions~\cite{PDG}.
63:
64: % \subsection{Decay of the tau}
65: The $\tau$ is assumed to decay according to the relative
66: probabilities into one of the most frequent modes~: $e$,
67: $\mu$, $\pi$, $\pi\pi^0$, $\pi\pi^+\pi^-$ and $\pi\pi^0\pi^0$, which cover
68: 90\% of the total decays.
69: We simplified
70: the kinematical distribution of the decay products, reproducing only the essential feature, namely the
71: fraction of
72: initial energy going into the electromagnetic shower and into the
73: hadronic one\footnote{Note that at 1~EeV the decay length of a $\pi^0$ is
74: 200 m, therefore comparable to the interaction length in air
75: (750m)}). A more accurate description would not modify our results as
76: can be seen in Fig.~\ref{ground_spot} where the central ground region of
77: a pure electromagnetic shower, although different, still compares with the
78: hadronic one.
79: The muons are considered to be unobservable. A possible effect of the
80: longitudinal polarization of the tau was ignored.
81:
82: \par Once a $\tau$ emerges from earth, and if a decay occurs within an
83: altitude of 3 km above the ground, an atmospheric shower is simulated as
84: described in Sec.~\ref{shower_sim}.
85: The detector response is then evaluated through a
86: simulation (outlined below) of the interactions of incident particles in water.
87:
88: \par Interactions in the mountains surrounding the detector were also simulated, using a detailed description of the relief.
89: Their contribution was found to be much less than the material below sea level, whatever the energy.
90: On the other hand we did not account for the lower density in Pacific Ocean, 250 km West from the southern site.
91: The overall correction is less than 10 \%.
92:
93: \subsection{Detector response}
94: The set of weighted ground particles in a ``sampling region'' around each
95: station is used to regenerate a set of particles entering the tank,
96: statistically reproducing all significant characteristics of the incident
97: flux~: global normalization of the different particles, distribution in
98: energy and direction.
99: \par
100: Then a simplified simulation is performed for interactions (cascade of
101: Compton scattering and pair production for photons, energy loss for
102: charged particles) and \v{C}erenkov emission in the water.
103: The production of \v{C}erenkov photons and their propagation in the tank
104: is performed until they hit a PMT or are absorbed in the water or in the tank walls.
105: The PMT response is assumed to be proportional to the amount of light
106: emitted. This is a good approximation in most cases, in particular for the sum
107: over the three PMTs collecting the light from the tank.
108: \par The level of the local trigger (one tank) is set to 4 {\em vem}
109: (vertical equivalent muons), and a global trigger is built if at least 4
110: stations are locally triggered within 20 $\mu$s with a relatively compact
111: topology. For exemple at least two stations must be within 3 km from a ``central'' one,
112: and an additional one within 6 km.
113: If needed, some long-shaped configurations with nearly aligned stations
114: and the right time spacing could be added to the global trigger
115: processor. These allow to a gain of up to 50~\% in the acceptance at energies between 0.1
116: and 0.3 EeV. However, we did not include these additions in this study.
117: \begin{figure}[!t]
118: \begin{center}
119: \vspace*{-1cm}
120: \hspace*{-1cm} \includegraphics[width=8cm]{evt1.eps}
121: \end{center}
122: \vspace*{-2cm}
123: \caption{Simulation of the ground trace left by a $\tau$ decay shower as
124: produced by a 5$\times$\E{17} tau neutrino.
125: Each thick circle represents a triggered station with a surface
126: proportional to the Cerenkov signal.
127: The $\tau$ shower had an energy of 3.63$\times$\E{17} and decayed 390
128: meters above the ground. In this particular exemple energy deposition in triggered tanks
129: ranges from 4 to almost 100 {\em vem}.}\label{tau-events}
130: \end{figure}
131:
132: \par
133: Fig.~\ref{tau-events} shows a simulation
134: of the ground trace of a tau, produced by a $5\times$\E{17}
135: neutrino, as sampled by the Auger stations.
136: The signal is clearly visible and 10 stations pass the 4 {\em vem}
137: trigger requirement (thick circles).
138:
139: \par
140: The probability to detect a shower with a given visible energy
141: depends essentially on the altitude of the core at the maximal lateral
142: development. It is not very sensitive to the exact definition of the local trigger
143: threshold nor to the global configuration.
144: For example, detecting all events with 3 locally triggered stations
145: would not increase sensibly the rate, except on the edges of the
146: acceptance below a few times 0.1 EeV. This is illustrated on Fig.\ref{detect},
147: where the equivalent area for detecting a shower has been plotted for various trigger
148: conditions~: triangles 3 stations, squares 4 stations and circles 4
149: stations plus the above condition on the global configuration. %Note that
150: %the efficiency is normalized to the area of the array (3000 km$^2$) but
151: %the simulated shower were scattered over a larger region (4200 km$^2$) to
152: %allow for edge effect. Therefore the efficiency may exceed one, as
153: %under good condition a shower that does not land above the nominal Auger
154: %surface may still be visible.
155:
156: \begin{figure}[!ht]
157: \hspace*{-1.1cm}
158: \includegraphics[width=9.0cm]{eff_new.eps}
159: \vspace*{-1.0cm}\caption{Equivalent detection area of tau showers averaged over all decay
160: channels versus the shower center (defined 10 km after the decay point)
161: altitude. Triangles : 3 stations global trigger, squares 4 stations, circles 4 stations
162: and the compact topology restriction.}
163: \label{detect}
164: \end{figure}
165:
166:
167: \subsection{Reconstruction}
168: The direction of origin may be estimated from the times of arrival of the
169: shower front on the stations, which is, as a first approximation, a plane
170: moving at speed $c$.
171: The precision on the azimuthal angle $\varphi$ is of the order of 1 deg,
172: and could be improved by taking into account the front curvature and by weighting
173: each station contribution according to its integrated amplitude.
174:
175: \par As a horizontal array is only sensitive to $\sin\theta$ the zenith angle
176: $\theta$ is quite difficult to obtain precisely.
177: However, taus are all produced with $|\theta-90|<5$~deg. Therefore one can isolate them
178: from the standard horizontal neutrino shower as can be seen on Fig.~\ref{sin_theta}.
179: %The normalization of the two curves on the figure is arbitrary and was
180: %chosen in order to emphasize the differences of the two distributions. In
181: %the full mixing oscillation hypothesis the standard neutrino
182: %distribution should be about a factor 40 smaller than the tau neutrino one.
183: \par
184: The reconstruction of the energy $E_i$ of the incident neutrino is
185: much more delicate~:
186: \begin{itemize}
187: \item The energy $E_{\tau}$ of the emerging tau may be much less
188: than $E_i$, in particular for $E_i \gg 1$ EeV, where many intermediate
189: interactions may have occurred reducing $E_{\tau}$ to a few 0.1 EeV.
190: As $\theta$ is not well known, it is difficult to evaluate even an order
191: of magnitude of the energy loss.
192: \item An arbitrary fraction of $E_{\tau}$ goes into neutrinos
193: and will not be visible while the decay type will influence
194: the hadronic to electromagnetic ratio of the decay products.
195: This may be corrected for only if the tail of the shower is visible on
196: the ground.
197: \item The estimation of the shower energy depends strongly on the
198: altitude of the shower core which is {\em a priori} unknown.
199: If many stations are hit, there is a hope to evaluate it from
200: the transverse distribution.
201: \end{itemize}
202: Given these difficulties, we can predict the rate of events knowing
203: the energy spectrum of the original neutrinos.
204: The inverse will be difficult.
205: \par
206: A careful statistical analysis of all observable characteristics such as
207: tank multiplicity, longitudinal and transverse profile of the ground spot and
208: time structure will certainly give additional information on the original spectrum.
209: We also beleive that for events where a arge number of tanks are struck we can obtain an estimate
210: of the neutrino energy but those studies need to be done.
211: Of course, the hybrid reconstruction (involving both the ground array
212: and the fluorescence detector of Auger) will be extremely valuable to
213: remove some ambiguities (zenith angle, visible energy),
214: but such ``golden'' events are expected to be less than 10\% of the total
215: event rate.
216:
217: \subsection{Evaluation of the acceptance}
218: The rate of observable events on a given surface $A$ (surface covered by the Ground Array)
219: is simply the rate over the whole earth, multiplied by $A/(4\pi R_T^2)$ , where $R_T$ is
220: the radius of the earth. This rate may be evaluated from a parallel flux
221: crossing the earth section ($\pi R_T^2$) as the integration over the
222: solid angle just gives an additional factor of $4\pi$.
223: \begin{figure}[!ht]
224: \vspace*{-1.3cm}
225: \hspace*{-0.8cm}
226: \includegraphics[width=9cm]{distang.eps}
227: \vspace*{-1.8cm}
228: \caption{$Sin\theta$ distribution of accepted events. Solid line~:
229: neutrino interaction in the atmosphere, filled area~: tau neutrino
230: interaction in the ground. The normalization between the two curves
231: (1:20) roughly corresponds to $\nu_\mu\leftrightarrow\nu_\tau$ oscillations with full
232: mixing and an $E^{-2}$ flux. The RMS on the reconstructed $sin\theta$ is at most 0.005.}
233: \label{sin_theta}
234: \end{figure}
235: \par
236: A tau emerging with an angle $\alpha$ over the horizon greater than
237: $\alpha_m=0.3$~rad has no chance of producing an observable shower at
238: ground level. Therefore we only simulated incident neutrinos close to the earth
239: surface for which $\alpha \leq 0.3$~rad (hatched area on
240: Fig.\ref{simul}).
241: For various incident energies, $N_{sim} = 10^6$ neutrinos were simulated
242: and the complete history up to the trigger was followed, giving the total
243: number $N_{acc}$ of accepted events.
244: \par The apperture at a given energy may then be defined
245: as:\label{effective_area}
246: \begin{eqnarray*}
247: A_{\text{\small eff}} & = & 4\pi~\pi R_T^2 \sin^2\alpha_m~\frac{A}{4\pi
248: R_T^2}~\frac{N_{\text{\small acc}}}{N_{\text{\small sim}}}\\
249: & = & \pi A \sin^2\alpha_m~\frac{N_{\text{\small acc}}}{N_{\text{\small sim}}}
250: \end{eqnarray*}
251: and the rate of events (integrated over the solid angle) coming from neutrinos of energy between $E_1$ and
252: $E_2$ as~:
253: \[
254: \frac{dN_{\text{\small acc}}}{dt} = \int_{E_1}^{E_2} f(E)\,A_{\text{\small eff}}(E)\,dE
255: \]
256: where $f(E)$ is the incident flux.
257: