SENSITIVITY OF THE CHERENKOV TELESCOPE ARRAY TO THE DETECTION OF AXION-LIKE PARTICLES MANUEL MEYER MARCH 13, 2015 BLAZAR WORKSHOP SLAC MANUEL.MEYER@FYSIK.SU.SE
EXTRAGALACTIC BACKGROUND LIGHT (EBL): STARLIGHT AND DUST Emitted spectrum ɣ Observed spectrum e - e + [Nikishov 1962; Jelley 1966; Gould & Schréder 1966,1967] 2
ɣ B AXION-LIKE PARTICLES ALPs: pseudo-nambu Goldstone bosons, arise in extensions of Standard Model Similar to axions that solve strong CP problem in QCD Couple to photons in magnetic fields mixing with ɣ rays: ALP mass < 1 μev Dark matter candidate a L a = 1 4 g a F µ F µ a = g a EBa [Peccei & Quinn, 1977; Weinberg, 1978; Wilczek, 1978; Raffelt & Stodolsky 1988; Csaki et al. 2003; Hooper & Serpico, 2007] 3 B ɣ
jet lobes: B ~ 1μG, λ ~ kpc MAGNETIC FIELDS ALONG LINE AGN jet: B ~ 1G ɣ host galaxy: B ~ 1μG, λ ~ 100 pc OF SIGHT galaxy cluster: B ~ 1μG, λ ~ kpc B a Intergalactic: B 1nG Milky Way: B ~ 1μG [De Angelis et al., 2007,2011; Mirizzi et al., 2007; Hochmuth & Sigl 2007; Simet et al. 2008; Mirizzi & Montatnino 2009; Sanchez-Conde et al. 2009; Horns et al. 2012; Tavechhio et al. 2012, 2014; Wouters & Brun 2013; Abramowski et al. 2013; MM et al. 2014, MM & Conrad 2014] 4 B ɣ
EXAMPLE: MIXING IN CLUSTER AND MILKY WAY ɣ Photon-ALP conversion probability B a CRITICAL ENERGY E crit / m2 a! plasma 2 g a B ɣ g 11 = g a ɣ / 10-11 GeV -1 = 5 m nev = m a / nev = 10 5 B
EXAMPLE: MIXING IN CLUSTER AND MILKY WAY MANY RANDOM REALISATIONS ɣ Photon-ALP conversion probability B a g 11 = 5 m nev = 10 6 B ɣ
EXAMPLE: MIXING IN CLUSTER AND MILKY WAY PHOTON SURVIVAL PROBABILITY ɣ B a z = 0.4 g 11 = 5 m nev = 10 7 B ɣ
EXAMPLE: MIXING IN CLUSTER AND MILKY WAY PHOTON SURVIVAL PROBABILITY ɣ Expected spectral features: B 1. Irregularities + reduced photon flux at E crit 2. a Reduced opacity at high optical depth z = 0.4 g 11 = 5 m nev = 10 8 B ɣ
CHERENKOV TELESCOPE ARRAY Factor ~10 improvement of point source sensitivity over currently operating Cherenkov telescopes Northern and southern array for full-sky coverage Southern array: multiple telescope designs, e.g.: 4 large (23m) telescopes, energy threshold ~30 GeV 23 mid-sized (12m) telescopes + U.S. Schwarzschild-Couder telescopes Small (4m) telescopes, covering > 3km 2, large collection area for E > 10 TeV 9
ANALYSIS FOR SPECTRAL IRREGULARITIES IN NGC 1275 IN PERSEUS CLUSTER Simulate 50 hrs CTA observation w/ and w/o ALPs Simulated CTA Spectrum with ALP mixing Preliminary work with M. Wood Fit smooth spectrum use spectral residuals to search for irregularities (chisquared) Repeat for many random B-field realizations m a = 22.9 x 10-9 ev g aγ = 0.43 x 10-11 GeV -1 Smooth Spectral Model ALP Mixing Spectrum folded with CTA Energy Resolution No sys. uncertainties included 10 [Approach similar to Wouters & Brun, 2013; H.E.S.S. Collaboration 2013]
SENSITIVITY FOR SPECTRAL IRREGULARITIES CAST SN ɣ-ray burst H.E.S.S. Globular clusters TeV transparency ALPS II CTA Irreg. ADMX IAXO ALP DM 95% exclusion (preliminary) QCD Axion [Preliminary work with Matthew Wood] 11 LIMITS SENSITIVITIES THEORY PREDICTIONS
ANALYSIS FOR REDUCED OPACITY Extrapolate measured blazar spectrum to high optical depths g 11 = 4.3, m nev = 12.2 Simulate CTA observation w/ and w/o ALP contribution Compare the simulations with log-likelihood ratio test Repeat for many random B- field realizations; for several sources with different magnetic field scenarios 12 [MM & Conrad, 2014]
SOURCES USED FOR OPACITY SENSITIVITY ANALYSIS Source Redshift Assumed T obs [hours] Comment 1ES 0229+200 0.139 41 Within ~700 kpc of cluster of Wen et al. (2012) catalog PG 1553+113 > 0.4 20 PKS 1424+240 > 0.6035 67 Flare 2012, reaching 100% Crab [ATel 4069], within 1.5 Mpc of cluster of Hao et al. (2010) catalog (GMBC catalog) Observations beyond τ 5, [VERITAS Collab. 2014], use mixing in BL Lac jet PKS 0426-380 1.11 70 Long flaring period, two photons with E > 100 GeV [Tanaka+ 2013], use mixing in lobes 13 [MM & Conrad, 2014]
SENSITIVITY FOR REDUCED OPACITY SN ɣ-ray burst 3σ detection CTA Opacity H.E.S.S. CTA Irreg. Globular clusters TeV transparency ADMX CAST ALPS II IAXO ALP DM preliminary QCD Axion [Preliminary work with Matthew Wood; MM & Conrad, 2014] 14 LIMITS SENSITIVITIES THEORY PREDICTIONS
SUMMARY AND CONCLUSION Oscillation between ɣ rays and ALPs leave imprint on spectra: 1. Irregularities around critical energy CTA will be sensitive to ALPs with g a ɣ 10-12 GeV -1 at a mass of m a ~ 10 nev probe parameter space where ALPs are DM candidates 2. Reduced opacity at high optical depths CTA will be sensitive to ALPs with g a ɣ 2 x 10-11 GeV -1 and masses m a 100 nev probe parameter space where ALPs can explain reduced opacity Complementary to future laboratory searches with, e.g., ALPS II and IAXO 15
DISCUSSION: HINT FOR A REDUCED OPACITY? Situation at the moment: different statistical analyses with different samples lead to different results Observed spectra corrected for absorption are too hard [De Angelis et al. 2009,2011,2013; Dominguez & Sanchez-Conde 2011, Rubtsov & Troitsky 2014] Statistical analysis yields ~4σ indication [Horns & MM 2012] Recent analysis of so far largest blazar sample does not yield a hint [Biteau & Williams, 2015] How to resolve this? Agree on a hypothesis test to search for an opacity anomaly at high optical depths The dream : full likelihood analysis of all spectra from all IACTs that have data points at high opacities 16
BACK UP SLIDES 17
MODEL ASSUMPTIONS FOR OPACITY STUDY 18 [MM & Conrad, 2014]
RESIDUALS OF FITS TO ALL SPECTRA 19
OPACITY ANALYSIS USING KS TEST INCLUDING SPECTRA PUBLISHED IN 2013 20
OPACITY ANALYSIS USING T TEST INCLUDING SPECTRA PUBLISHED IN 2013 21
OPACITY ANALYSIS USING T TEST INCLUDING SPECTRA PUBLISHED IN 2013 NOT A NORMAL DISTRIBUTION LONG TERM VERITAS SPECTRUM OF 1ES1218+304 SPECTRUM HAS LARGEST EFFECT WITH 5 DATA POINTS WITH τ > 2 [MADHAVAN ET AL. 2013] 22
SIMULATED SPECTRA 23 [MM & Conrad, 2014]
LIKELIHOOD RATIO TEST Likelihood for observing N ON (source region) and N OFF (background region) in i-th energy bin: Expected counts from source: μ, expected counts from background: b Likelihood ratio test with Asimov data set [Cowan et al. 2011], i.e. N ON = μ + b and N OFF b / α: sum over all sources and all bins for which τ > 2 and significance > 2σ expected counts w/o ALPs expected counts w/alps that maximise L Test statistic follows χ 2 distribution with ~ 7 d.o.f. 24 Background counts that maximise L for fixed μ background counts that maximise L [MM, Montanino, Conrad 2014; MM & Conrad, 2014]
TEST STATISTIC DISTRIBUTION FOR RANDOM MAGNETIC FIELDS 25 [MM, Montanino, Conrad 2014]
CONTRIBUTION OF EACH SOURCE TO SENSITIVITY 26 [MM & Conrad, 2014]
ASSESSMENT OF MODEL ASSUMPTIONS RANDOM MAGNETIC FIELD 27 [MM & Conrad, 2014]
ASSESSMENT OF MODEL ASSUMPTIONS CHANGING THE MODEL ASSUMPTIONS 28 [MM & Conrad, 2014]
MAGNETIC FIELDS IN GALAXY CLUSTERS Observational evidence: Non-thermal (synchrotron) emission Rotation measure Field strengths: 0.1 and 10 µg Extent: up to few Mpc Magnetic field follows electron distribution n e (r) Rotation measure map with 5 GHz contours of galaxy NGC 4869 in the Coma cluster Simulated B field (blue) and analytical profile (magenta) of the Coma cluster Rotation measure [Figure from Bonafede et al., 2010; see, e.g., Feretti et al., 2012, for a review] 29 Change of polarisation angle
TURBULENT B -FIELDS IN GALAXY GROUPS Fanaroff-Riley Type I (misaligned BL Lacs) often found in groups of galaxies Turbulent B fields, O(μG) Turbulence spectrum: power law on kpc scales [Guidetti et al. 2010] 30
B FIELD IN BL LAC LOBES Residuals of rotation measure Inferred from rotation measures of radio sources behind lobes Turbulent Centaurus A: B ~ 1 μg λ ~ 20 kpc total path length ~ 180 kpc 31 [Feain et al. 2009]
MAGNETIC [Jansson & Farrar 2012] FIELD OF THE MILKY WAY Turbulent and coherent component Photon-ALP mixing: only coherent component relevant g 11 = 5, E > E crit [Pshirkov et al. 2011] Electron density given by NE2001 code [Cordes & Lazio 2003] [Horns, Maccione, MM, Mirizzi, Montatino, Roncadelli 2012] 32
PHOTON-ALP MIXING IN BL LAC JETS Modelling follows Tavecchio et al. (2012,2014) and Mena & Razzaque (2013) Environment modelled between VHE emission zone r VHE to some outer radius R max Magnetic field (coherent), profile: B(r) =B 0 r r VHE p p = 1 for toroidal field, p = 2 for poloidal field [Blandford et al., 1984; Rees, 1987; Lobanov 1998; O Sullivan & Gabuzda, 2009] Ambient electron density: n(r) =n 0 r r VHE s s = 2 assuming equipartition [Lobanov 1998; O Sullivan & Gabuzda, 2009, see also the review by Pudritz et al. 2012] Values for B and r VHE from SSC modelling 33