7 years of IceCube data Christian Haack , RWTH Aachen University For - - PowerPoint PPT Presentation
Constraints on diffuse neutrino emission from the Galactic Plane with 7 years of IceCube data Christian Haack , RWTH Aachen University For the IceCube Collaboration ICRC 2017, 07/15/2017 Galactic Cosmic Rays During propagation
Christian Haack, RWTH Aachen University ꟷ For the IceCube Collaboration ꟷ ICRC 2017, 07/15/2017
▪ During propagation protons interact with material near the source or interstellar gas ▪ Interactions produce pions which decay into γ and ν →Diffuse γ / ν emission
07/15/2017
𝜌+/− 𝜌0 p 𝜈 𝜉𝜈 𝛿 𝛿
Christian Haack 2
▪ During propagation protons interact with material near the source or interstellar gas ▪ Interactions produce pions which decay into γ and ν →Diffuse γ / ν emission
07/15/2017
𝜌+/− 𝜌0 p 𝜈 𝜉𝜈 𝛿 𝛿
Christian Haack 3
γ / ν are tracers of acceleration, propagation and interaction mechanisms
Simple model:
Spatial: 𝜌0-component of Fermi diffuse γ background model Energy: 𝐹−𝛿 powerlaw → No prediction for flux normalization Two free parameters: Normalization & Spectral Index
07/15/2017
Christian Haack 4
▪ Diffuse Galactic γ and ν are created by π-decays ▪ Fermi provides model of diffuse Galactic γ emission
▪ Model by Gaggero et. al. provides consistent picture of ν and γ diffuse emission ▪ Based on KRAγ CR-diffusion model
Assumes diffusion coefficient depending on galiocentric radius
▪ Developed to solve problems of conventional propagation models (e.g. “Milagro excess“) ▪ 5 PeV or 50 PeV CR cutoff
07/15/2017
Christian Haack 5 Daniele Gaggero et al 2015 ApJL 815 L25
▪ Model by Gaggero et. al. provides consistent picture of ν and γ diffuse emission ▪ Based on KRAγ CR-diffusion model
Assumes diffusion coefficient depending on galiocentric radius
▪ Developed to solve problems of conventional propagation models (e.g. “Milagro excess“) ▪ 5 PeV or 50 PeV CR cutoff
07/15/2017
Christian Haack 6 Daniele Gaggero et al 2015 ApJL 815 L25
→ν measurement can help constrain diffusion models
07/15/2017
▪ Cherenkov detector at the geographic South Pole ▪ 5160 Digital Optical Modules (PMT with onboard digitization) ▪ 86 Strings in a depth of 1450m to 2450m ▪ 125m string spacing ▪ Detection Principle: Cherenkov emission of secondary particles produced by ν-interaction in or near the detector ▪ Energy threshold ~10GeV (with DeepCore)
𝜉𝜈 𝜈 𝜈 𝜉𝜈
Christian Haack 7
▪ Allows precise characterization of the isotropic astrophysical neutrino flux
Astrophys.J. 833 (2016) no. 1, 3
▪ No associated neutrino point sources found ▪ Measured flux parameters differ from other channels ▪ Might be an indication for second component (Galactic?) in neutrino spectrum ▪ How robust is this result against a subdominant Galactic component?
07/15/2017
Christian Haack 8
Update to 8 years
New eight year result!
Binned Method
▪ Binned poissonian template fit in reconstructed neutrino energy and direction ▪ Signal & background PDF calculated from MC ▪ Systematic uncertainties included as continuous nuisance parameters ▪ Fits neutrino flux parameters of conventional + prompt atmospheric, isotropic + Galactic astrophysical
Unbinned Method
▪ Unbinned spatial LH fit with energy weighting ▪ Data-driven background model („scrambled data“) ▪ Fits anisotropic Galactic plane contribution in isotropic background
07/15/2017
Q: Does the data include a contribution of Galactic neutrinos? ~10% better sensitivity Consistent picture of all neutrino fluxes
Christian Haack 9
07/15/2017
Baseline Model
Fermi 𝜌0 spatial template 𝐹−𝛿 energy spectrum (baseline: 𝛿 = 2.5)
KRAγ(50 PeV cutoff)
Spatial template from tuned diffusion model 𝐹−𝛿 energy spectrum (𝛿~2.45) Prediction: ~ 30 𝜉𝜈/yr
Christian Haack 10
07/15/2017
Baseline Model
Fermi 𝜌0 spatial template 𝐹−𝛿 energy spectrum (baseline: 𝛿 = 2.5) Normalization: 10−5 𝐻𝑓𝑊−1𝑑𝑛−2𝑡−1@1GeV 75 ν/a
KRAγ
Fermi 𝜌0 spatial template 𝐹−𝛿 energy spectrum (baseline: 𝛿 = 2.5)
Downgoing Upgoing KRAγ is stronger ↔ Fermi is stronger
Christian Haack 11
07/15/2017
▪ Using the unbinned method, we set an upper limit of: 1.2 x 𝐋𝐒𝐁𝛅 (50PeV cutoff) ▪ For the Fermi 𝜌0 model:
Φ90 = 2.97 ⋅ 10−18
𝐹 100 𝑈𝑓𝑊 −2.5
GeV−1cm−2s−1
▪ Not more than 14% of the diffuse flux @ 𝛿 = 2.5 from the Galactic plane ▪ New ANTARES (arXiv: 1705.00497) limit: 1.3 x 𝐋𝐒𝐁𝛅 (50PeV cutoff)
http://arxiv.org/abs/1707.03416
Christian Haack 12
07/15/2017
Using the binned method and the Fermi 𝜌0 model, we obtain an insignificant non-zero best fit: ▪ P-Value of no Galactic flux: 7% ▪ Consistent with unbinned method
Christian Haack 13
Φ = Φ0 ⋅ 10−18
𝐹 100𝑈𝑓𝑊 −𝛿
GeV−1cm−2s−1
Φ0 = 3.13 ±1.85
2.11, 𝛿 = 2.07 ±0.25 0.22
http://arxiv.org/abs/1707.03416
07/15/2017
▪ The binned method delivers consistent picture of isotropic & galactic astrophysical fluxes ▪ Check the influence of inclusion of galactic component on isotropic flux measurement ▪ Scan isotropic flux parameters with galactic plane parameters (norm. + spectral index) free to float ▪ Significance of isotropic astrophysical component still > 3σ
affect ability to constrain isotropic astrophysical flux.
Christian Haack 14
http://arxiv.org/abs/1707.03416
▪ A measurement of a diffuse galactic neutrino emission can provide valuable insight into CR propagation mechanisms ▪ IceCube is already able to probe models for diffuse 𝜉𝜈 emission: 90% UL: 1.2 x KRAγ (50PeV) For Fermi 𝜌0 model: Φ90 = 2.97 ⋅ 10−18
𝐹 100 𝑈𝑓𝑊 −2.5
GeV−1cm−2s−1 ▪ Binned analysis results in an overfluctuation (~7% p-value) ▪ The best-fit galactic spectral index for Fermi 𝜌0 is suprisingly hard (2.07 ±0.25
0.22)
→ But no conclusions can be drawn yet ▪ Measurement of the isotropic astrophysical flux is robust against a subdominent galactic component
Outlook: Global analysis combining multiple detection channels & combined analysis with Antares
07/15/2017
Christian Haack 15
Paper submitted to ApJ http://arxiv.org/abs/1707.03416
▪ Sensitivity is defined as median upper limit, when no signal is present ▪ Neyman construction:
▪ Generate BG & Signal pseudoexperiments ▪ 90% UL is found when 𝑅50 𝑈𝑇𝐶𝐻 ෝ = 𝑅10 𝑈𝑇𝑇𝑗𝑜𝑏𝑚
07/15/2017
Signal distribution 𝑜𝑗𝑜𝑘 = 𝑜𝑉𝑀 Signal distribution 𝑜𝑗𝑜𝑘 = 𝑜𝐸𝑗𝑡𝑑𝑄𝑝𝑢
Christian Haack 18
Requirements ▪ Pure sample of 𝜉𝜈 (tracks) (>99%) ▪ High neutrino statistics ▪ Well reconstructed events (good pointing) Diffuse Sample ▪ Six (Seven) years of data combining multiple detector configurations ▪ Purity: >99.7% ▪ High neutrino statistics: 350.000 events (2009 – 2015) ▪ High signal statistics: ~500 astrophysical 𝜉𝜈 (2009-2015)
07/15/2017
Christian Haack 19
Requirements ▪ Pure sample of 𝜉𝜈 (tracks) (>99%) ▪ High neutrino statistics ▪ Well reconstructed events (good pointing) Aachen Diffuse Sample (developed by Sebastian S. & Leif R.) ▪ Six (Seven) years of data combining multiple detector configurations ▪ Purity: >99.7% ▪ High neutrino statistics: 350.000 events (2009 – 2015) ▪ High signal statistics: ~500 astrophysical 𝜉𝜈 (2009-2015)
07/15/2017
Neutral Currents
Christian Haack 20
07/15/2017
Atmospheric Neutrinos ▪ Prompt (heavy meson decay) ▪ Conventional (pion / kaon decay) Astrophysical Neutrinos ▪ Diffuse Galactic (powerlaw) ▪ Diffuse isotropic (powerlaw)
Christian Haack 21
07/15/2017
Atmospheric Neutrinos ▪ Prompt (heavy meson decay) ▪ Conventional (pion / kaon decay) Astrophysical Neutrinos ▪ Diffuse Galactic (powerlaw) ▪ Diffuse isotropic (powerlaw)
Christian Haack 22
Detector Effects:
Ice properties, optical sensor efficiency
Flux Uncertainties:
Rate, shape and composition of the CR flux, rate of pion-to-kaon decay in air showers, neutrino cross sections Influence of every sys. effect on analysis variables is parametrized continously and implemented as nuisance parameters
07/15/2017
position change shape change
Detector is symmetric in azimuth (and located at South Pole) → RA not influenced by sys. Effects
Christian Haack 23
07/15/2017
Atmospheric Neutrinos ▪ Conventional (pion / kaon decay) ▪ Prompt (heavy meson decay) Astrophysical Neutrinos ▪ Diffuse Galactic (powerlaw) ▪ Diffuse isotropic (powerlaw)
Christian Haack 24
Q: Does the data include a contribution of galactic neutrinos? ▪ 3D histogram of observables: zenith angle, right ascension, energy ▪ Forward-folding template fit using poissonian likelihood: ▪ Likelihood ratio test: TS =
ℒ ො 𝑜𝑏𝑚𝑏𝑑𝑢𝑗𝑑, ො 𝑜𝑏𝑡𝑢𝑠𝑝 𝜄𝑜𝑣𝑗𝑡𝑏𝑜𝑑𝑓) ℒ 𝑜𝑏𝑚𝑏𝑑𝑢𝑗𝑑=0, ො 𝑜𝑏𝑡𝑢𝑠𝑝 𝜄𝑜𝑣𝑗𝑡𝑏𝑜𝑑𝑓)
𝑀 𝝂 𝜾, 𝝄 𝒐 = ෑ
𝑗=1 𝑂 (𝝂𝒋 𝜾, 𝝄 )𝑜𝑗
𝑜𝑗! exp(−𝝂𝒋 𝜾, 𝝄 )
07/15/2017
Bin number MC-expectation # data events Signal & Nuisance Parameters
Christian Haack 25
High Purity Sample („Diffuse“) ▪ Upgoing track-like events ▪ Six years of data combining multiple detector configurations ▪ Purity: >99.7% ▪ ~350,000 events ▪ Allows measurement of spectral properties of astrophysical 𝜉𝜈-flux
07/15/2017
High Statistics Sample („Point Source“) ▪ Upgoing, downgoing and starting track- like events ▪ Seven years of data combining multiple detector configurations ▪ Contamination from atmos. Muons and neutral current (+𝜉𝑓) interactions ▪ ~700,000 events Good pointing (<0.5° @ PeV)
Christian Haack 26
▪ Fermi LAT provides a full model for Galactic γ emission ▪ Modelling of CR injection, transport and interaction with interstellar matter
07/15/2017
𝜌0 IC Brems Diffuse Sum
PS Total Sum
→ Model has difficulties explaining the total diffuse γ emission accurately
Christian Haack 27
07/15/2017
Two Strategies
Upgoing Neutrinos (𝜉𝜈)
Starting Events (all-flavor)
misreconstructed µ
Downgoing Upgoing
Christian Haack 28