Detecting Supernovae with Ice Cube Kael Hanson University

Detecting Supernovae with Ice. Cube Kael Hanson University of Wisconsin HEP Seminar March 26, 2007

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Overview 2 • The Ice. Cube neutrino observatory is currently under construction at the South Pole. To date approximately 30% of the detector is installed and over 50% of the instrumentation has been produced. • Science operation begins 4/1/2007 • As a UHE telescope, Ice. Cube is gigaton detector with effective area in excess of 1 km 2 to > Te. V muons. • First proposed by Jacobsen, Halzen, Zas PRD 49 (1994), possibility to use background counting of optical detectors for detection of Me. Vscale neutrinos from galactic supernovae. • This technique utilized by AMANDA detector since 1998. • With est. 3 Mton effective volume for low-energy ν, Ice. Cube potentially provides detailed information for modeling supernovae: – Bigger, better, lower-noise optical detectors – Data acquisition at fine timescales - 1. 6 ms / bin • Supernova detection would provide high resolution, high statistics data for supernova burst models. • Sensitivity may be sufficient for particle physics investigations

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 The Ice. Cube UHE Neutrino Observatory Both representations not to scale - nevertheless nicely illustrate salient detector characteristics. 3

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Ice. Cube UHE Neutrino Observatory (2) 4 • • • 1 st string deployed Jan-2005 Last string deployed Jan-2011 4500 deep-ice optical modules along 75 strings Depth ranges 1450 m - 2450 m with 17 m spacing. 320 surface modules in 80 stations - Ice. Top Detector is hybridizing already - this year radio and acoustic test modules deployed and are working well. • Ice. Cube optimized for detection of Te. V and Pe. V-scale neutrinos of cosmic origin: point sources, diffuse HE neutrinos, GRBs, &c.

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Ice. Cube Integrated Volume (Projected) 5 § Graph shows cumulative km 3·yr of exposure × volume § # of strings per year is based on latest “best guess” deployment rate of 12 strings 2007 (13) and 14 strings per season thereafter. § 1 km 3·yr reached 2 years before detector is completed § Close to 4 km 3·yr at the beginning of 2 nd year of full array operation.

AMANDA • 677 analog OMs deployed along 19 strings Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 – – – 6 • • • 10 strings 1997 (AMANDA B 10) 3 strings 1998 (AMANDA B 13) 6 strings 2000 (AMANDA II) AMANDA supernova analyses typically employ between 400 -500 of these channels due to instabilities in some. Analog PMT signals using electrical and optical transmission lines. 200 m diameter, 500 meters height; AMANDA II encompasses 20 Mton instrumented ice volume: 6 times more dense than Ice. Cube. AMANDA will remain operational and form Ice. Cube Inner Core Detector for low E physics (~ 100 Ge. V WIMPs, &c) Ice. Cube surrounding strings provide effective veto – lower background and can push AMANDA energy threshold down. Conventional TDC / ADC technology for AMANDA has been entirely replaced by TWR system. Beginning 2007 season, AMANDA / Ice. Cube data streams are conjoined; detector subsystems will share trigger information.

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 The AMANDA/Ice. Cube Authors 7

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Optical Properties of Glacial Ice Why deploy in ice? Deep glacial ice is optically transparent. Two mechanisms: scattering length ~ 20 m, absorption ~ O(100) m. Ice has several layers of dust from prehistoric events. Monte Carlo detector simulation must account for this. Reconstruction methods involving maximum likelihood tests against hypotheses have been developed to overcome difficulties posed by photon scattering. Plots above from in situ measurements using artificial light sources in AMANDA. “Hole ice” around deployed modules must also be taken into account. 8

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 The Enhanced Hot Water Drill (EHWD) EHWD designed to drill a 2450 m × 60 cm hole in ~30 hr. Fuel budget is 7200 gal per hole. Shown above is drill camp and tower site (inset), both mobile field arrays. Everything must fit into LC-130 for transport to Pole. 9 Supply: 200 GPM @ 1000 psi, 190 °F Return: 192 GPM @ 33 °F Make-Up: 8 GPM @ 33 °F Thermal Power: 4. 5 Megawatt

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Drilling Ted Schultz Top layer of packed snow is called firn. Hot water drill designed for ice drilling – it gets starter hole from firn drill (lower right). (Top left and top right) EHWD drill head entering hole. 10

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Deployment 11

2005, 2006, 2007 Deployments AMANDA 80 79 78 Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 77 72 71 74 73 2005 - 1 string 67 66 65 64 56 55 59 58 57 56 50 49 48 47 2006 - 8 strings 46 40 39 38 30 29 2007 - 13 strings! Ice. Top only 2007 21 1424 DOMs deployed to date 1320 DOMs on 22 deep ice strings >99% DOM survival rate Following deployment years will target ≥ 14 strings / year until a ~75 strings are installed. Final detector will include ~4500 in-ice optical modules and 500 surface modules. 12

The Ice. Cube Digital Optical Module (DOM) DOM Highlights - Optical Large Area Photocathode 10” (500 cm 2) Hamamatsu R 7081 -02 bialkali PMT (peak QE 24% @ 420 nm) Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Low noise < 300 Hz background counting rate in-ice (with deadtime - see later) Glass / Gel Improvements Better transmission in 330 - 400 nm relative to AMANDA OM Optical calibration Each DOM is calibrated ε(λ) in the lab to about 7%; in-situ flasher board additionally permits in -ice measurements DOM Highlights - Electronics The DOM has been in production at UW, DESY-Zeuthen, and Stockholm since mid-2004 with very little change. It has met or exceeded design requirements, and, despite harsh re-freeze conditions 2500 m deep, we’ve lost < 0. 5 % of the units during this critical phase. To date, 3000 of approx. 5000 units have been produced. 13 “Smart” sensor digital technology Versatile FPGA design with option to expand / change programming at any point in lifecycle. Core of supernova DAQ resides inside DOM itself. Array Timing Handled in DOM logic - DOM-to-DOM timing good to 2 -3 ns using RAPCal method. Low power - 3. 75 W / DOM

Reducing Background Counting - Deadtime Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Correlated noise in PMTs 14 Measurements counting all PMT pulses in deployed DOMs yield average rate of approx. 700 Hz / module. Study of the time structure of this noise clearly indicates correlated noise patterns: PMT Afterpulses Well-known that ionized residual gases in PMTs cause afterpulses on timescale (for large PMTs) of 6 -10 µs. Glass Scintillation Small contamination of rare-earth oxides produces further tail out to 100’s of µs due to scintillation processes. To reduce noise rate and restore Poissonian behavior to fluctuations , it is necessary to apply afterpulse inhibit (deadtime) when pulse counting. Analysis of real data demonstrates that 200 µs afterpulse suppression window reduces background by factor of 4 while sacrificing only few percent of signal.

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 DOM Temperatures in the Ice 15 One DOM didn’t freeze-in until May!

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Singles Counting Rates vs. Depth 16 Importance of noise rates: 1. ) noise rate w/o dead time: 700 Hz, important for DAQ bandwidth 2. ) noise rate w/suppression of 50µs: 300 Hz, important for event reconstruction and in particular for supernova sensitivity. Two Icecube strings equivalent or more sensitive than all of AMANDA to SN.

The Supernova DAQ In the ice - counting Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 FPGA Integrate into 1. 6384 ms (216 / 40 MHz) bins SPE/MPE discriminator crossings with application of 200 µs deadtime. This counting operation executes in parallel with digitization and readout activity of main DAQ core and does not depend on its state. CPU Interrupted every 6. 5 ms (4 bins). Bins are timestamped with ~ 10 ns precision and copied into SDRAM where they await ~1 Hz commands from surface to readout accumulated bins. At the surface - assembly Surface DAQ for supernova is 2 -stage assembly of disparate data packets from individual DOM channels: String. Hub Issue periodic readouts to string of 60 DOMs. Use RAPCal information to translate DOM timestamps to UTC then perform merge and sort into single stream of data which is sent downstream. Supernova. Builder Further merge-and-sort of String. Hub streams into final stream written to tape. 17 Began supernova-mode data-taking last September with 9 -string 540 -module detector. This year’s run is set to begin 5/1. Taking full supernova data (1 MB/sec) since 3/18.

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 The Online Trigger and the SNEWS Connection Real-time detection of SNe SNEWS and AMANDA Don’t stop there: real-time detection and reporting of burst candidates gives “heads-up” to optical observers. AMANDA official participant since 6/2005. In addition, combining many, distributed observatories makes more sensitive, robust alert. Super. Nova Early Warning System [NJP 6 (2004)114] … Strict requirement that individual detector produce no more than 1 false alarm / week. SNEWS and Ice. Cube Mainz supernova group adapting AMANDA online trigger to handle data feed from Ice. Cube DAQ. Necessary to thoroughly evaluate system to demonstrate that we meet the strict requirement of < 1 false alarm / week. 22 -string Ice. Cube will join end of year 2007. 18 Near real-time delivery of alerts possible through Iridium link to pole (24/7; low-bandwidth) Noise rates in AMANDA can be highly variable. Real-time detection of noisy channels necessary; additionally, detector chi-square reports uniformity of signal in detector (good SN totally uniform). Statistics based on Gaussian approx. ; OK for large bins.

Supernovae • Categorization of SNe historically based on presence of hydrogen in spectroscopic lines: Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 – Type I SNe don’t have it 19 • Type Ia probably from dwarf stars accreting mass to Chandrasekhar limit, then exploding. • Type Ib/c core collapse SNe but have lost hydrogen envelope because of stellar winds (Wolf-Rayet stars) or perhaps mass transfer to companion star. Incidentally, recent investigations associate these types with GRBs (e. g. Woosley and Bloom astro-ph/0609142); variations with slower jets may further be γ dark but still produce Te. V ν detectable in Ice. Cube and may be much more common. – Type II SNe have the hydrogen lines - these are likely massive stars with hydrogen envelopes intact that undergo core collapse • Intense neutrino luminosities only with core collapse SNe Type Ib/c and Type II. • Galactic rate of core collapse SNe given by INTEGRAL measurements - 1 -3 per century (Nature 439 (2006) 45). Optimistically, given Ice. Cube lifetime of 15 years - 40% chance of observing galactic event.

Core Collapse - Basic Features • Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 – Massive star (> 8 M⊙) has developed 1. 5 M⊙ Fe core which is beginning to neutronize. – “Homologous” core becomes unstable as Fe nuclei leech out electrons and photodisintegration processes occur and collapses to ~30 km where nuclear repulsion causes ‘core bounce. ’ Lots of neutrinos of all flavors inside simmering protoneutron star. – Shock wave from rebound and collapsing outer portion of core. When shock penetrates “neutrinosphere, ” initial neutronization burst escapes. – Core accretes infalling material and begins to radiate ~1053 erg in neutrinos over the next ~ 0. 5 s until explosion. – The remainder of the energy is emitted over timescales of 10’s of seconds as the newly-formed neutron star cools • • 20 Details of physics of core collapse not completely understood due to lack of observational data. 19 neutrino events detected in Kamiokande-II and IMB from SN 1987 A bolster support for general model but were too few to provide specific insights (see Yuksel and Beacom astro-ph/0702613 v 2 for recent discussion of SN 1987 A data). Optical emission some ~10 hr later. Models still fail, unmanipulated, to provoke explosion - this is continuing mystery which might be unraveled given adequate input data provided by Ice. Cube or future SN neutrino detector.

The Supernova - GRB Connection Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 GRBs Mounting evidence that Type Ib/c SNe produce GRB. Rate of GRB is 1%. May be many more with mildly relativistic jets (Γ ~ few) which don’t produce significant EM component. Te. V neutrinos detectable from nearby objects (D < 20 Mpc) at the rate of ≈ 1 SN / year. [ Razzaque, Mezaros, Waxman PRL (2004) ] Ando & Beacom PRL 2005 SNe ? MK, A. Mohr, astro-ph/0701618 21 Te. V Neutrinos from Nearby SNe Sensitivity can be doubled by optical follow-up!

Me. V Supernova Neutrinos in AMANDA Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 2000 -2003 preliminary SN neutrino signal simulation center of galaxy, normalized to SN 1987 A 22

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Me. V Neutrinos in Ice. Cube 23 • Jacobsen, Halzen, Zas PRD 49 (1994) 1758 first proposed. Followup calculation for Ice. Cube presented in JCAP 6 (2003) - Dighe, Keil, Raffelt. Unfortunately both groups have some incorrect assumptions. • Ice. Cube supernova analysis group has done preliminary update to Ice. Cube of more detailed work from AMANDA Ph. D. thesis by T. Feser. • Some points from all works: – Neutrino effective volume ∝ E 3; 2 powers from σ 1 power from electron/positron tracklength. Thus, detection is sensitive to neutrino energy spectra - or stated another way, effective volumes are all dependent on SN models / oscillations, &c. – Effective volume ∝ Λabs the optical pathlength in the ice – For SN models this yields approx. 700 m 3 Veff - or a sphere around each module of 5 m radius. As such, each module may be treated independently. – Pri� ncipal detection channel is inverse beta decay (ref cross-sections slide) - this makes detection of neutronization peak difficult. – Currently, with ~ 1300 deep ice modules detector mass is 800 kton. – Full Ice. Cube ~ 4500 modules detector mass 3 Mton.

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Me. V Neutrinos in Ice. Cube (continued) 24 • The detection comes from increase in background counts across the entire array. • Disadvantages - you have no pointing or energy reconstruction as in Super-K • Advantage is that enormous volume provides high-statistics measurement. Time binning can be made fine. • Signal from 1987 A SN at galactic center would produce 475 k excess counts in ~10 sec window on a background of 12 × 106 counts from noise - S/N ~ 150: 1 in full Ice. Cube; • In current 22 -string detector signal is still very significant: 140 k excess counts giving S/N ~ 75: 1. • Gain comes when decimating signal in time to study time evolution

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Me. V Neutrino Cross Sections on Water 25

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Signal Predictions 26 From Dighe, et. al. JCAP 06 (2003)005. Note that predicted rate should be scaled dow by 3 x - cf. previous slide. Also statistical errors for 50 ms bins are +/- 250 counts. From Dighe, et. al. JCAP 06 (2003)005. Their prediction of earth modulation effect for particular choice of neutrino mixing angle θ 13 could render Ice. Cube sensitive to neutrino mass hierarchy. Ice. Cube detector sensitive to modulation effect; however, modeldependence of the neutrino luminosity would almost certainly require contemporaneous detection of SN at another detector such as Super-K or Hyper-K.

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Signal Predictions - Ice. Cube SN group calculation 27

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Detection of neutronization peak Neutronization fluence is largely independent of SN model and progenitor mass - useful as a neutrino standard candle. Ice. Cube detection is marginal. 28

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Neutrino Oscillation in Star From Kachelreiß & Tomàs PRD 71 (2005) 29

Detecing Supernovae with Ice. Cube K. Hanson - UW HEP Seminar 2007 -03 -26 Conclusions 30 • Ice. Cube gigaton high energy neutrino observatory doubles nicely as megaton low energy detector • High-statistics measurements are possible which can at the very least provide detailed measurements of neutrino luminosity vs time. • Ice. Cube longevity gives good chance of observing significant galactic core collapse event. • Ice. Cube is operating and is sensitive to GC events NOW! With 1300 modules deployed in ice this year sensitive to 100% of Milky Way. • SN analysis group working to improve outdated simulations and bring up online supernova trigger • Participation in SNEWS later this year.