Haitao Yu Columbia Astrophysics Laboratory

1. Indirect Dark Matter Search with Antideuterons: Progress and Future Prospects for General Antiparticle Spectrometer (GAPS) Haitao Yu Columbia Astrophysics Laboratory The General Antiparticle Spectrometer (GAPS) is a novel approach for indirect dark matter (DM) searches that exploits the antideuterons produced in neutralino-neutralino annihilations.  GAPS complements existing and planned direct DM searches as well as other indirect techniques, probing a different, and often unique, region of parameter space in a variety of proposed DM models.  The GAPS method involves capturing antiparticles into a target material with the subsequent formation of exotic atoms.  These exotic atoms decay with the emission of X-rays of precisely defined energy and a correlated pion signature from nuclear annihilation.  This signature uniquely characterizes the captured antiparticle.  I will report on a GAPS prototype tested in an antiproton beam at the KEK accelerator in Japan that confirms the multi-X-ray/pion topology.  I will also outline the steps that are being taken to develop GAPS for a long duration balloon experiment.

2. Accelerator testing of the general antiparticle spectrometer; a novel approach to indirect dark matter detection T Aramaki, C J Hailey, J Jou, J E Koglin, H T YuColumbia Astrophysics Laboratory W W Craig, L Fabris, N Madden, K P ZiockLawrence Livermore National Laboratory F GahbauerUniversity of Latvia/Columbia U. K MoriCITA/Toronto • Recently Joined: • T Yoshida and H Fukeof BESS collaboration • S Boggsof UC Berkeley • D Protic of Institut fur Kernphysik, Germany • Rene Ong of UCLA Online at stacks.iop.org/JCAP/2006/i=01/a=007

3. Outline • Cosmology Overview • Why AntiDeuteron? Why lowest energy is preferred? • GAPS Detection Concept • Atomic Physics of exotic atoms; what can the X-rays tell us? • GAPS Prototype Experiment • Testing the GAPS concept on a beamline with antiprotons • Data Analysis and Results • Picking out a needle from the bottom of sea • Future Works

4. Cosmology Overview NASA, WMAP

5. The composition of the universe is known in detail… • Except we don’t know what “Dark Matter” is • However, we know • It is neutral (“Dark”) • It is Weakly Interacting • It is composed of Massive Particles (WIMPs)* • The particles are non-relativistic (CDM, or Cold Dark Matter)

6. X _p c _D _n q,h,W… c Dark MatterModel HadronizationMonte Carlo CoalescenceModel The Most promising DM candidates includes: • Neutralino, the Lightest SuperSymmetric Partner in most SUSY theories; • Kaluza-Klein States predicted in 5-dimensional theory, which are motivated by string theory • And many other particles, e.g. axion • The neutralino is a Majorana particle  It will annihilate with itself • Pair annihilating WIMPS produce:g, n, e+… • Donato et al. (2000) suggest antideuteron signal _ p… _ D…

7. Why an antideuteron search for dark matter? • Growing number of papers in past six years suggesting antideuterons can probe very significant volume of parameter space in many SUSY and other DM models. For example: • Edsjo, Schelke and Ullio (2004) – antideuterons are the best means to find the neutralino and constrain mSUGRA • Baer and Profumo (2005) – antideuterons best means to search for UED Kaluza-Klein particles • Baer et.al. (2006) – balloon-borne antideuteron search sensitive to DM from reduced SU(3) gaugino mass models • Barrau & Ponthieu (2004) – antideuterons from primordial black holes provide limit on gravitino mass • Antideuteron coverage of parameter space is complementary with other approaches such as direct searches, gamma-rays (including positron annihilation), neutrinos & antiprotons • Avoids well-known limitations of antiproton searches • There are lots of direct searches and no antideuteron searches! • Maybe you don’t find Dark Matter …but discover something unexpected!!!

8. Antiprotons are produced in neutralino-neutralino annihilation However, they are difficult to distinguish from cosmic-ray produced antiprotons • Primaries: _ χ+χ→ p+p • Secondaries: _ p+p → p+p+p+p _ p+He → p+He+p+p • Tertiaries: _ p with diffusive energy loss to low energies Picozza & Morselli 2002

9. Primary component:  neutralino annihilation _ X+X → D+etc Secondary component: spallation _ p+H → p+H+D+etc _ p+He → p+He+D+etc Clean signature @ low E, but see Duperry et al. ‘05 However, sensitivity demand is daunting Low energy, neutralino-neutralino producedantideuterons are near background free Antideuteron flux at the earth (with propagation and solar modulation) Primary Secondary

10. _ b b BESS limit 10 -4 GAPS W-W+ balloon AMS-02 KK DM 10 -6 LZP Antideuteron Flux [m-2 s-1 sr-1 GeV-1] GAPS 10 -8 10 -10 10 -12 AMS BESS Sensitivity needs of antideuteron searches for DM will require next generation experiments From Baer & Profumo ‘06 • Current premier techniques utilize magnetic spectrometers from balloons (BESS), satellite (PAMELA) & space station (AMS) with grasp ~ 0.2 m2-sr  have reached practical limits of mass/grasp • Optimal antideuteron search requires ~2-20 m2-sr in package of ~ 1-2 metric tons • First upper limit for antideuteron search recently reported by BESS (Fuke et.al. 2003)! • AMS will begin to obtain interesting sensitivity at E >~ 1 GeV/n (where background is higher)  Need for GAPS at low energy 0.1 1 10 100 Kinetic Energy per Nucleon [GeV/n]

11. GAPS Detection Concept GAPS concept, as shown in original GAPS proposal

12. Atomic Transitions no,lo n=nK~15 n=6 g n=5 g n=4 g Degrader n=3 n=2 p+ n=1 p+ p- p- po GAPS is based on radiative emission ofantiparticles captured into exotic atoms A time of flight (TOF) system tags candidate events and records velocity Auger e- Refilling e- The antiparticle stops in the target & forms an excited exotic atom with near unity probability Deexcitation X-rays provide signature Pions from annihilation provide added background suppression LadderDeexcitations Dn=1, Dl=1 Segmented Detector p* 19 keV TOF Target _ D p* Nuclear Annihilation 75 keV p* 35 keV

13. Exotic Atoms has been studied extensively Hartmann et.al, 1976 Wiegand et.al, 1976 And many models were developed to model the cascade process Ishiwatari et.al, 2004

14. Although the atomic physics is well known… • Many experiments didn’t measure yield • From the known data (of Kaonic atom), yield depends on atomic number • Antiproton yields were NOT measured for most elements! • Yield also known to depend on chemical composition, gas pressure, etc • No Antiproton yield known for liquid/solid targets • All experiments done with a single / a few detector • GAPS concept (spontaneous detection of x-rays) not tested before

15. There is still much to be learned from accelerator testing of a GAPS prototype • What is the detection efficiency for GAPS like detector? • measure yield of X-rays per capture • Investigate flight representative data acquisition electronics and detector approaches • Investigate signatures generated by particle backgrounds • protons, pions, electrons etc. • Investigate specific target materials • Investigate processes not well characterized by previous experiment • higher energy X-ray transitions, nuclear g-rays • Investigate solid targets • original concept used gas…

16. GAPS Prototype Experiment • Designed, built, and first tested within 18mo of proposal acceptance • Two beam tests performed in KEK pi2 beamline in 2004 and 2005 • On a shoe-string budget

17. GAPS Prototype Detector • 16 NaI(Tl) detector modules covering 40 cm long x 12 cm diameter target cell • Each modular 4x2 arrays of 25mm x 25mm x 5 mm thick crystals (128 total) • Solid angle coverage ~30% (average) Target

18. GAPS Detector C2F6 (gas) S C (Aerogel) Gas, Liquid, and Solid Target can be Tested Al (Wool)

19. A Custom low-power DAQ suitable for a balloon experiment was developed in LLNL • Directly handles signals from the 128 NaI detector phototubes. • Pulse processing system is built around a gated integrator with a parallel fast channel. • Fast channel discriminator recognizes an event and initiates the signal processing cycle. • Charge from PMT is integrated & digitized with a low power, 16 bit ADC. • External trigger mode: trigger generated from TOF. • Self-triggered mode: each channel is read out every time a signal is recognized • 32 extra ADC channels digitized for the timing & energy deposit of the trigger counters • 16 logic signals for each event trigger. • Fast PCI interface for high-speed data transfer between the electronics & PC – can sustain a rate of over 100 kHz per channel. • Total power consumption for the data acquisition system is 150 W

20. Beamline Elements Atomic Transitions • Objective: • Pick out antiprotons stopped in the target Absolute Yield = (# of x-ray)/(# of antiproton stopped) • Challenges: • Low concentration of antiproton in the beam • High beam energy == thick degrader • LOTS of annihilation in degrader • Serious scattering • Range Straggling Exotic Atom TOF Trigger Counters ShowerCounters Degrader Nuclear Annihilation

21. TOF Counters • Each Beam spill contains 20-30 antiprotons, along with ~105 other particles (pions, kaons, electrons) • Three TOF counters allows clean separation of antiprotons and pion accidental background • Electronic triggers on antiproton events and/or scaled pion events TOF Trigger Counters

22. Degrader and Shower Counters • Beam lose intensity below 1GeV/c, need >15cm lead degrader to stop antiproton at that energy • Scattering is significant so TOF not possible after degrader • >99% antiprotons annihilated in flight inside degrader, often cause hadronic showers • Shower counters (0.5cm plastic scintillator counters sandwiched with ¼ inch lead sheets) installed to monitor possible hadronic shower development • Veto and exit counters give additional check on antiproton stops Atomic Transitions Exotic Atom ShowerCounters Degrader Nuclear Annihilation

23. Data Analysis • All Channels (crystal, TOF, pulse height) has been calibrated with known sources / delays • Performance consistent with typical values • NaI(Tl) has energy resolution of 15KeV (FWHM) at 60KeV • Timing Channels resolution is 1.0ns (FWHM) • Non-linearity effects removed according to literature

24. Range Curve Calibration • Range curve measured with positive beam (pion + proton) • All Channels needs to be calibrated for dead time to obtain correct event rate • The curve is verified with GEANT simulation • More careful simulation with true experiment geometry to determine antiproton capture rate

25. proton 25 20 15 c2 10 5 0 120 160 200 240 S1 Energy [keV] Advanced selection of Pbar events with ‘Likelihood’ function based on energy deposition pion • Stopping antiprotons: energy deposit Di in beam counters is approximately gaussian • Probability is product of individual counters • Variance si is also approximately constant • Log likelihood becomes c2 problem • Minimize c2 to determine energy at the S1 counter where Di are a function of ES1 1500 2500 500 3500 ADC Channel c2

26. Simulations shows “likelihood function” is efficient in identifying stopped events Simulation Measurement

27. Using advanced selection of likely Pbar stops X-ray transitions are evident with no detector cuts Simulated Cut Efficiencyxcut = 25% Detector Solid AngleWdet = 31% X-ray Line Widths ~ (6 keV)*sqrt[E/60] edet: X-ray Escape & Detection Efficiency Nx: Number of fit X-rays 53±24% 66±14% 66±14% 63±29 127±27 148±31 Yield of X-ray Transitions Yx = Nx/ xcut edet Wdet Nstop Measured X-ray Yields consistent with Theory!

28. Similar Results for the Sulfur Target • Different Target Configuration • Less target = less stopping • Higher Z = higher absorption • Discrete geometry 106±40 153±39 240±46 0.50±0.20 0.35±0.09 0.42±0.08 Measured X-ray Yields consistent with ~50%

29. NaI Detector [keV] p* 96 17 52 20 32 0.3 0.2 8.9 10.3 Sample Al event This event has all three higher energy x-rays as predicted in atomic cascade model, and one charged pion from annihilation of antiproton. The TOF and pulse height readouts confirms that the event is an antiproton survived from degrader, with correct energy to stop in target. TOF & ShowerBeam Counters TOF [ns] dE/dx [MeV/g cm2] 0.5 0.1 5.1 5.8 6.4 7.7 P4 P1 P2 S1 S2 S3 S4 P3 P5

30. So… We are Happy

31. Auger e- Atomic Transitions Refilling e- no,lo n=nK~15 n=6 LadderDeexcitations Dn=1, Dl=1 g n=5 g n=4 g n=3 n=2 p+ n=1 Nuclear Annihilation p+ p- p- po A little bit more to do on data analysis • Fit yield data with atomic cascade models to determine physical parameters such as initial angular momentum distribution, electron refilling rate • Extrapolate yields in known elements into other elements of interest

32. The Collaboration is Actively Working on Design of Flight Module Original Design Plastic TOF Trigger 3x3x3 segmented NaI Detectors Gas Target New Design Option Plastic TOF Trigger 10 Alternating Layers Si (Li) Detectors Teflon Target Detector Target TOF (cut away)

33. Current focus is on Si(Li) wafers formed into p-i-n detectors with coarse (~2-4 cm) strips on both sides and thickness of ~5mm Very good energy and timing information Complexity is fairly low because of very coarse pixels Simple to produce – challenge is large number of wafers required (~thousands) Two Si(Li) detector samples has been tested, with extrapolated performance of 1.2-1.5KeV resolution Several (cost effective?) detector schemes are under study for use in an actual flight experiment

34. Flight Module will have much better particle ID power than prototype experiment • Much better energy resolution = much less accidental events • Rejection power ~ (dE)3 • In addition, pion tracks can be reconstructed • Number of pions helps identify antideuterons from antiprotons • Antideuteron generates much more pions in annihilation compared to antiprotons • Depth information also give significant rejection power • Antideuteron of same velocity travels further than antiproton • Flight module can easily achieve proposed particle ID power

35. Explorer Satellite LDB Antartica GAPS optimization principles have been discussed and experiments designed for both long duration balloon & Explorer missions • Mori et. al. 2002, Astrophys. J. 566, 604 • Hailey et. al. 2004, Nucl. Instr. Meth B, 214, 122. • GAPS collaborators are actively optimizing new designs

36. GAPS development plan culminates in a long-duration balloon (LDB) experiment • Flight proposal to NASA in Spring 2007 • Flight of GAPS prototype to test flight representative prototype detector and characterize space background from Sanriku facility in Japan (late 2008-early 2009) • First GAPS flight from Antarctica in 2010; second & third flights 2011-2012; each flight ~3 weeks • Experiment design will be ultra long duration (ULDB) capable to exploit such a launch if it becomes available; flight duration >100 days NASA/ULDB

37. GAPS is complementary with both direct & other indirect methods as well as collider experiments A variety of methods will be required to: • Probe unique regions of parameter space • Narrow down the available phase space & eliminate models • Define the type and sensitivity of future experiments

38. Antideuteron search experiments for other models such as extra-dimensional DM models • A balloon mission will provide discovery potential for a number of theoretical models such as 5D Warped GUT (RH Neutrino) • Other models such as U.E.D Kaluza-Klein require a satellite GAPS mission Baer & Profumo 2005

39. Evaporation of Primordial Black Holes p e+ e- _ p GAPS is also sensitive to Primordial Black Holes Antideuterons from PBH evaporation have similar spectrum as generated in neutralino annihilations

40. Summary • We had a fruitful GAPS prototype experiment • Measured Yield on target candidates • Demonstrated GAPS detection concept (multiple x-ray events) • Tested flight-compactable electronics • We are working hard on GAPS flight module • New design, better energy resolution, more information for ID • Optimization and detector test under way • Much exciting physics to be discovered!

41. New Calculation of antideuteron flux R. Duperry et al., 2005

42. Thermal Design of balloon payload • Mechanical Engineer working on the design; a few options available for evaluation

44. 25 20 15 c2 10 5 0 120 160 200 240 S1 Energy [keV] Statistical distribution demonstrates that assumptions are fairly good c2 Probability Distribution: • Initial Energy is only fit parameter k = Npaddles – 1 • Good agreement with expected distribution as more paddles are added • More advanced, but more time consuming analysis methods possible c2

45. As an additional check,number of gold/silver events agrees with prediction