UHE photons and neutrinos at the Pierre Auger

UHE photons and neutrinos at the Pierre Auger Observatory Enrique Zas Departamento de Física de Partículas & Instituto Galego de Física de Altas Enerxías, Universidad de Santiago de Compostela Universidade de Santiago de Compostela, SPAIN for the Pierre Auger Collaboration

A Hybrid detector Two techniques: Fluorescence (FD) Particle detector array (SD) ht g Redundant ~ 10% of events are observed with both: wealth of information about shower development & exploit SD FD e sc e li ce n Fl or u SD E. Zas

Communications antenna SD Units GPS Low consumption electronics Solar panels Battery box Calibrated online regularly using signals induced by atmospheric muons 3 photomultiplier tubes Rotomolded plastic tank 12 tons of purified water Digitised signals: FADC 25 ns time bins Time [ns]

Rich SD data: Useful observables which can be correlated with hybrid data • At detector level – – – Signal: Number of particles Start time: timing Rise time: Sperad of particle arrival Area over Peak: low for single muons Structure: jumps -> muon counting. . • At shower level – – Shower size Direction Xmax (FD) Curvature of particle front

Signal (VEM) “Slow & broad signal” produced by EM component “Fast & narrow signal” produced by muonic component Signal (VEM) Time (ns) 25 ns time resolution allows distinction between broad and narrow signals Time (ns)

Xmax and curvature are related Xmax Larger Xmax => larger curvature (smaller radius) L C, RAV, AAW, EZ (Ap Phys 2004)

Apparent for ms in hadronic showers (ns) 200 400 Time delay of 300 first muon 200 (curvature) 100 & average 600 150 700 100 50 0 0 Distance to core 80 120 80 800 870 40 40 0 60 20 1000 m

Risetime also related to Xmax Two main reasons: 1. Z range (production) 2. m less delayed than e & g Deep showers have more em component muons travel in straight lines em component straggles Risetime

Photon Search

Basis: Xmax discrimination P. Homola for the Auger Collab. , ICRC 2009 g-induced showers reach maximum deeper in the atmosphere than nucleonic ones

Use Surface Detector data Astroparticle Physics 29 (2008) 243 -256 Two discriminating observables • Radius of curvature of shower front • Time structure of shower front (Risetime) (both correlated to Xmax) Rise time is the time it takes to go from 10% to 50% of the total signal Signal (VEM) 50% of integrated signal Time (ns)

Deviation of Curvature w. r. t. to mean [s units] Surface Detector Cut: Median of distribution Principal component analysis MC photons 5% Data MC photons Deviation of Risetime w. r. t. to mean [s units] Data Cut

Direct Xmax search: Hybrid Astroparticle Physics 27 (2007) 155 & ar. Xiv 0903. 1127 v 2 Quality cuts • More than 6 PMTs • Shower axis distance to highest signal SD station <1. 5 km • Reduced c 2 (profile fit) <6 and ratio to c 2 (line fit) <0. 9 • Xmax within field of view Fiducial volume cuts avoid biasses: • Zenith> 350 +g 1(E) [35+7 @ 1019. 7 • Distance to telescope < 24 km +g 2(E) • Viewing to shower axis angle >150 (Cherenkov rejection) • E>2, 3, 5 and 10 Ee. V

Full simulations made (Corsika, QGSJET 01, FLUKA): • Fotons • Protons • Iron Hybrid Quality cuts Fiducial volume cuts

Hybrid search: g candidates E cand p(Fe) 2 Ee. V 8 30 (0. 3) 3 1 12 (0. 2) 5 0 4 (0. 1) 10 0 1 (0) Uncertainties: s(Xmax) ~ 16 g cm-2 s(E)/E ~ 22 % Cut: Median of the simulated photon Xmax distribution 5% of protons simulated with QGSJET 01 above this line

Deepest event observed

Limits on g fractions: SD & Hybrid P. Homola for the Auger Collab. , ICRC 2009 31 % 3. 8% 5. 1% 2. 4% 3. 5% 2. 0% A 1, A 2 = AGASA HP = Haverah Park Y = Yakutsk Strong constraints on: Super-Heavy DM & Topological Defect models g fraction constrained in Energy - range 2 Ee. V → 40 Ee. V

Neutrino Search

Cosmogenic ns Cosmic rays at ultra high energy (neutrino? ) V. S. Berezinsky, G. T. Zatsepin Academy of Sciences of the USSR, Physical Institute, Moscow, Russia Physics Letters B Vol. 28, Issue 6, pp. 423 -424 (1969) Received: 8 November 1968 Published: 6 January 1969 Abstract: The neutrino spectrum produced by protons on microwave photons is calculated. A spectrum of extensive air shower primaries can have no cut-off at an energy E>3 1019 e. V, if the neutrino-nucleon total cross-section rises up to the geometrical one of a nucleon.

Selected developments in neutrino search with EAS: • • • 1969 Inclined showers for neutrino detection Berezinsky, Zatsepin 1987 n bound with Fly’s Eye 1991 -97 n bounds with Tokyo data Halzen, EZ, . . . 1996 Auger UHE n possibilities shown Capelle, Cronin, Parente, EZ 1999 Earth skimming nt effect Fargion / Lettessier-Selvon & Bertou, Billoir 2007 First earth skimming experimental bound Auger / Hi. Res

Inclined showers Protons, nuclei, g: Shower g’s e+’s and e-’s do not reach ground level Only muons

Inclined hadron Air Showers vertical atmospheric depth g e +e m 0 2000 4000 6000 8000 Depth (g/cm 2) 10000 12000

Case 1: down-going n Air shower n Detection (deep)=>inclined Case 2: Earth-skimming nt nt Earth Air shower t Upgoing: detection=>inclined) Complex three stage process • Attenuation through Earth and regeneration: NC CC & t CC CC & t decay • CC interaction, t energy loss and no decay • Exit and t decay in the atmosphere

“Earth skimming” nt Auger results: PRL 100 (2008) 211101 Jan 04 - Aug 07 PRD 79 (2009) 102001 Jan 04 - Apr 08 Low t loss => large target volume Large density: Earth’s crust Only sensitivity to nt CC channel Small zenith angle range (50) (solid angle)

“Down-going” n Low density target Zenith angle range 750(600? )-900 All channels and flavors. Relative contributions: Channel • CC ne interactions • NC n interactions • CC nm nt interactions • CC nt the t decay s x flv 3 x 2 1 x 6 3 x 4 3 x 2 Shower Energy Transfer • EM + Hadronic 100% • Hadronic 25% • t decay 40% • Resonant ne interact • qq • ene • mnm • tnt 6 x 1 1 x 1 • Hadronic • EM • NO shower • t decay 100% 25% 50%

Searching for n in data: general criteria D. Gora for the Auger Collab. , ICRC 2009 (1) Search for Inclined Showers Footprint of the shower on ground compatible with that of an inclined shower: Elongated pattern (large Length over Width). “Speed of propagation of signal” along Length, close to speed of light. Angular reconstruction.

(2) Search for showers with large electromg component Inclined proton/nuclei showers induced high in the atmosphere: “Fast & narrow signal” (mainly) of muons at ground. produced by muonic component Neutrinos: inclined showers with broad traces “Slow & broad signal” produced by EM component

Selection for earth skimming neutrinos • Trace cleaning (remove random muons) • Inclined • signal pattern length/width>5 (elongated) • 0. 31 m/ns > ground speed > 0. 29 m/ns (horizontal) • r. m. s. (ground speed) < 0. 08 m/ns (compatible) • Electromagnetic • >60% of stations satisfy “Offline To. T” (Time over threshod: 13 bins above 0. 2 VEM) • Signal over peak>1. 4 • Central trigger condition only to Off Tot stations • Quality trigger (T 5)

Selection for down going neutrinos • Only events of 4 or more stations • Trace cleaning (remove random muons) • Inclined • • signal pattern length/width>3 (elongated) 0. 313 m/ns > ground speed > 0. 29 m/ns (horizontal) r. m. s. (ground speed) < 0. 08 m/ns (compatible) Zenith reconstructed < 750 • Electromagnetic • Fisher discriminant analysis on ten variables (related)

Acceptance (Monte Carlo) Earth skimming • Earth conversion of nt to t • t decay in the atmosphere • extensive air shower • Trigger and identification efficiency (Et, h 10 km) • detector exposure (integration over running array) Down-going • Atmospheric interactions

Down-going neutrino channels

Fisher discriminant analysis Maximise discrimination power using multivariate analysis (Fisher discriminant). Very simple idea: – Find “projection line” for maximal hadrons & n separation Simple example in 2 D var 2 Neutrinos F is a linear combination F = a 1·var 1 +a 2·var 2 HAS var 1

Fisher discriminant analysis F is the linear combination of discriminating variables used maximising the ratio: “mean” of F for HAS and neutrinos maximally SEPARATED relative to Variance of F for HAS Variance of F for neutrinos

Variables for Fisher method Exploit that neutrino showers have: (1) Broad signals in the early part (2) Asymmetry in time spread of signals between early and late parts. Signal (VEM) Useful variable: AOP = integrated signal over peak signal Broad signal Large AOP Training data 01 Jan 04 -31 Oct 07 Area Over Peak of the first T 2 tank Narrow signal Small AOP (black) and Nu showers (red) AOP Product of the first four T 2 tanks Time (ns) Ten discriminating variables: First 4 AOPs First 4 (AOPs)2. Product of the first 4 AOPs. An asymmetry parameter: “Mean[early AOP] - Mean[late AOP]”. Time (ns)

Asymmetry in time spread: Neutrinos interacting deep in the atmosphere “Early” region “Late” region

Spread in time of the FADC trace of each station in an event Simulated down-going neutrino Spread in time of the signal [ns] Real inclined event Each dot represents a station in the event early late early (broad signals) late (narrow signals) (μs) Attenuation of the EM component of the shower from the earliest to the latest station (μs)

Example distributions: Inclined real events (black) Simulated nu showers (red) AOP of the 1 st tank in the event early – late asymmetry parameter of the event

Blind search for neutrinos: Data from 01 Jan 04 to 31 Oct 07 used to “train” Fisher method: • Select the best discriminating observables. • Set cuts in Fisher variable above which an event is a n candidate. Data from 01 Nov 07 to 28 Feb 09 to do a blind search for neutrinos No neutrino candidates in the search period

Flux limits for a E-2 neutrino spectrum COSMOGENIC ns AUGER limits Down 01 Nov 07 - 28 Feb 09 Up 01 Jan 04 -28 Feb 09 K [Ge. V cm-2 s-1 sr-1] 3. 2 x 10 -7 4. 7 x 10 -8 J. Tiffenberg for the Auger Collab. , ICRC 2009