Theory and Phenomenology of the Neutrino and Gamma-ray Universe
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University
By Bei Zhou, M.S. Graduate Program in Physics
The Ohio State University 2020
Dissertation Committee: Professor John F. Beacom, Advisor
Professor James J. Beatty
Professor Eric A Braaten
Professor Annika H. G. Peter c Copyright by
Bei Zhou
2020 Abstract
Neutrinos, an essential piece of the standard model of particle physics, provide critical infor- mation about the universe. In reverse, astrophysical neutrinos provide indispensable tools to study particle physics. Tremendous progress, including breakthroughs, has been made in the past decades and will likely be made in the next decade. To get there, many aspects of work are needed, including the properties of neutrino sources, neutrino interactions, neu- trino detectors, and more. In general, there are two major directions, one to increase the data from current sources of neutrinos, and the other is to develop new sources.
In this dissertation, I discuss a series of papers pertinent to both directions. First is studying TeV solar gamma-ray emission, as part of a larger program of work to develop the Sun as a new high-energy laboratory, including as a neutrino source. Next is study- ing subdominant interactions for high-energy neutrinos, including W -boson production and trident production, as demanded by the increasing precision of TeV–PeV neutrino astro- physics. The last is understanding atmospheric neutrino foregrounds for searches of the diffuse supernova neutrino background in Super-Kamiokande. This provides a solid the- oretical foundation for further reducing the foregrounds, which will significantly help the
first detection and subsequent precision measurements of the diffuse supernova neutrino background. Together, this series of work advances neutrino astrophysics as a broad and vibrant field.
ii To those who love me.
iii Acknowledgments
I have had an incredible time as a Ph.D. student in CCAPP and the Department of Physics at The Ohio State University, thanks to the help from the following people.
First of all, I would like to thank my advisor, Professor John F. Beacom. He is an incredible advisor, physicist, and leader. He cares about students’ success and is willing to spend plenty of time on students. I can ask for help on research, career, and life even though it was during late nights or weekends. He is very helpful in all these aspects, which is very important for international students like me. He has cultivated excellent students and postdocs who are also extremely helpful to me. His creativity, hard work, responsibility, and rigorousness set a great role model for me as a physicist, as a person, and as an advisor in the future.
I am also thankful to the current and previous graduate students of CCAPP and the physics department. I have got tremendous help and advice from Ranjan Laha, Shirley Li, and Kenny Ng before and after their graduation. I have also learned a lot from frequent chat and discussions with Benjamin Buckman, Chris Cappiello, Yang Cheng, Bin Guo, Liping
He, Paulo Montero, Bowen Shi, Wenjuan Zhang, and Guanying Zhu, and I enjoyed the time with them.
Next, I would like to thank the current and previous postdocs at OSU. I appreciate help from Tim Linden with various things especially my postdoc applications. I have enjoyed discussing particle physics with Juri Smirnov, from whom I have learned a lot. I have had a lot of fun talking about neutrinos, exploring pizza places in Columbus, and other things with Francesco Capozzi. I have also learned a lot from frequent chat and discussions with
Mauricio Bustamante, Junichiro Kawamura, Hong Zhang, Xilin Zhang, and Yinu Zhang.
iv I am also thankful to the professors. I am grateful to Annika Peter for working with us on solar gamma rays, serving on my committee, writing a letter for me, and giving me advice.
I am also grateful to Yuri Kovchegov who has taught me electrodynamics, particle physics, and quantum field theory in his courses, plus discussing questions related to my research.
I thank Eric Braaten who taught me quantum field theory, serving on my committee, and talking about physics with me. I also thank Stuart Raby who talked about physics and other things with me. I appreciated Jim Beatty for serving on my committee and Mary
Hall Reno of the University of Iowa who wrote a letter of reference for me.
The life in graduate school would also be harder without help from administrative people, who have given me all kinds of help in the department. So I would like to thank Jon Pelz,
Kris Dunlap, Phil Davids, Bryan Dunlap, Andrew Canale, and especially CCAPP Program
Coordinator Lisa Colarosa.
Last but not least, I would like to thank my parents for their unconditional care, love, support, patience, and everything. I also thank my friends outside physics. It is a lot of fun to be with them when I am not thinking about physics.
v Vita
2016 – 2020 ...... Graduate Research Associate, The Ohio State University 2015 – 2016 ...... University and Fowler Graduate Fellow, The Ohio State University 2015 ...... M.S. Astrophysics, Purple Mountain Ob- servatory, Chinese Academy of Sciences 2012 ...... B.S. Physics and B.E. Computer science, Guangxi University
Publications
“W -boson and Trident Production in TeV–PeV Neutrino Observatories” Bei Zhou, John F. Beacom Phys. Rev. D 101, 036010 (2020) [arXiv:1910.10720] (Ref. [1] in the bibliography)
“Neutrino-nucleus Cross Sections for W -boson and Trident Production” Bei Zhou, John F. Beacom Phys. Rev. D 101, 036011 (2020) [arXiv:1910.08090] (Ref. [2] in the bibliography)
“Constraints on Spin-Dependent Dark Matter Scattering with Long-Lived Mediators from TeV Observations of the Sun with HAWC” HAWC collaboration, plus J. F. Beacom, R. K. Leane, T. Linden, K. C. Y. Ng, A. H. G. Peter, B. Zhou (alphabetical) Phys. Rev. D 98, 123012 (2018) [arXiv:1808.05624] (Ref. [3] in the bibliography)
“First HAWC Observations of the Sun Constrain Steady TeV Gamma-Ray Emission” HAWC collaboration, plus J. F. Beacom, R. K. Leane, T. Linden, K. C. Y. Ng, A. H. G. Peter, B. Zhou (alphabetical) Phys. Rev. D 98, 123011 (2018) [arXiv:1808.05620] (Ref. [4] in the bibliography)
“Unexpected Dip in the Solar Gamma-Ray Spectrum” Qing-Wen Tang, Kenny C.Y. Ng, Tim Linden, Bei Zhou, John F. Beacom, Annika H.G. Peter Phys. Rev. D 98, 063019 (2018) [arXiv:1804.06846] (Ref. [5] in the bibliography)
vi “Evidence for a New Component of High-Energy Solar Gamma-Ray Production” Tim Linden, Bei Zhou, John F. Beacom, Annika H.G. Peter, Kenny C.Y. Ng, Qing-Wen Tang Phys. Rev. Lett. 121, 131103 (2018) (Editors’ Suggestion; Featured in Physics) [arXiv:1803.05436] (Ref. [6] in the bibliography)
“TeV Solar Gamma Rays From Cosmic-Ray Interactions” Bei Zhou, Kenny C. Y. Ng, John F. Beacom, Annika H. G. Peter Phys. Rev. D 96, 023015 (2017) [arXiv:1612.02420] (Ref. [7] in the bibliography)
“GeV excess in the Milky Way: Depending on Diffuse Galactic Gamma Ray Emission Template?” Bei Zhou, Yunfeng Liang, Xiaoyuan Huang, Xiang Li, Yizhong Fan, Lei Feng, Jin Chang Phys. Rev. D 91, 123010 (2015) [arXiv:1406.6948] (Ref. [8] in the bibliography)
“GRB 131231A: Implications of the GeV Emission” Bin Liu, Wei Chen, Yunfeng Liang, Bei Zhou, Haoning He, Pak-Hin Thomas Tam, Lang Shao, Zhiping Jin, Yizhong Fan, Daming Wei Astrophys. J. Lett. 787, L6 (2014) [arXiv:1401.7283] (Ref. [9] in the bibliography)
“Fast Radio Bursts as a Cosmic Probe?” Bei Zhou, Xiang Li, Tao Wang, Yizhong Fan, Daming Wei Phys, Rev. D, 89, 107303 (2014) [arXiv:1401.2927] (Ref. [10] in the bibliography)
“Model-dependent Estimate in the Connection Between Fast Radio Bursts and Ultra-High Energy Cosmic Rays” Xiang Li, Bei Zhou, Haoning He, Yizhong Fan, Daming Wei Astrophys. J. 797, 33 (2014) [arXiv:1312.5637] (Ref. [11] in the bibliography)
“High Energy Emission of GRB 130821A: Constraining the Density Profile of the Circum- burst Medium as well as the Initial Lorentz Factor of the Outflow” Yunfeng Liang, Bei Zhou, Haoning He, Pak-Hin Thomas Tam, Yizhong Fan, Daming Wei Astrophys. J. 781, 74 (2014) [arXiv:1312.2662] (Ref. [12] in the bibliography)
“High Energy Emission of GRB 130427A: Evidence for Inverse Compton Radiation” Yizhong Fan, Pak-Hin Thomas Tam, Fuwen Zhang, Yunfeng Liang, Haoning He, Bei Zhou, Ruizhi Yang, Zhiping Jin, Daming Wei Astrophys. J. 776, 95 (2013) [arXiv:1305.1261] (Ref. [13] in the bibliography)
Fields of Study
Major Field: Physics Studies in theoretical particle and astroparticle physics: neutrinos, dark matter, as- troparticle physics, gamma rays
vii Table of Contents
Page Abstract...... ii Dedication...... iii Acknowledgments...... iv Vita...... vi List of Figures ...... xi List of Tables ...... xviii
Chapters
1 Introduction1 1.1 What is exciting about neutrinos from the universe...... 1 1.2 Developing the Sun as a new high-energy neutrino source...... 3 1.2.1 Motivations...... 4 1.2.2 Review of prior work...... 5 1.2.3 My work ...... 8 1.3 Diffuse High-Energy Astrophysical Neutrinos ...... 10 1.3.1 Motivations...... 11 1.3.2 Review of prior work...... 12 1.3.3 My work ...... 18 1.4 Diffuse Supernova Neutrino Background...... 19 1.4.1 Motivations...... 20 1.4.2 Review of prior work...... 22 1.4.3 My work ...... 25
2 TeV Solar Gamma Rays From Cosmic-Ray Interactions 27 2.1 Introduction...... 27 2.2 Interplanetary and Solar Magnetic Fields ...... 31 2.3 Hadronic Gamma Rays ...... 33 2.3.1 Calculational framework...... 33 2.3.2 Calculation for the simplified case ...... 35 2.3.3 Calculation for the realistic case ...... 37 2.4 Leptonic Gamma Rays...... 41 2.5 Cosmic-Ray Electrons ...... 44 2.6 Conclusions and Outlook ...... 47
viii 3 Neutrino-nucleus cross sections for W -boson and trident production 50 3.1 Introduction...... 50 3.2 Review of W -boson and trident production ...... 52 3.2.1 Hadronic part...... 53 3.2.2 W -boson production...... 55 3.2.3 Trident production...... 58 3.3 Cross sections between neutrinos and real photons ...... 59 3.3.1 W -boson production...... 59 3.3.2 Trident production...... 62 3.4 Neutrino-nucleus cross sections: coherent and diffractive regimes ...... 64 3.4.1 Framework ...... 64 3.4.2 W -boson production...... 68 3.4.3 Trident production...... 71 3.5 Neutrino-nucleus cross sections: inelastic regime ...... 72 3.5.1 Framework ...... 72 3.5.2 W -boson production...... 73 3.5.3 Trident production...... 77 3.6 Total cross sections of W -boson and trident production, and ratio to CCDIS 79 3.7 Conclusions...... 81 3.8 Supplemental Material...... 84 3.8.1 W -boson production: amplitudes in the Standard Model ...... 84 3.8.2 Trident production: amplitudes in the Standard Model ...... 85 T/L 2 3.8.3 Trident production: kinematics and phase space for σνγ (ˆs, Q ) . . . 86 3.8.4 Trident production: coherent and diffractive cross sections for all channels...... 90
4 W -boson and trident production in TeV–PeV neutrino observatories 92 4.1 Introduction...... 92 4.2 W -boson production cross sections and implications ...... 95 4.2.1 Review of the total cross sections...... 95 4.2.2 New results for the differential cross sections ...... 96 4.2.3 Implication: Cross-section uncertainty...... 98 4.2.4 Implication: Attenuation in Earth ...... 101 4.3 Detectability ...... 102 4.3.1 Larger W -boson yields than Glashow resonance ...... 104 4.3.2 Review of detection in IceCube...... 104 4.3.3 Total shower detection spectrum...... 108 4.3.4 Unique signatures ...... 111 4.4 Conclusions...... 113 4.5 Supplemental Material...... 115 4.5.1 dσνA/dE` and dσνA/dEW ...... 115
5 First Detailed Calculation of Atmospheric-Neutrino Foregrounds for Super-Kamiokande Searches for the Diffuse Supernova Neutrino Back- ground 119 5.1 Introduction...... 120
ix 5.2 Framing the Problem...... 123 5.2.1 Predicted DSNB signals...... 123 5.2.2 Observed atmospheric neutrino backgrounds ...... 124 5.2.3 Atmospheric neutrinos...... 125 5.2.4 Neutrino interactions...... 126 5.2.5 Physics of detection in Super-K...... 128 5.3 Calculation Validation by Matching Super-K High-Energy Atmospheric Neu- trino Data...... 129 5.4 New Results on Super-K Low-Energy Atmospheric data: Invisible-Muon Component...... 134 5.4.1 Theoretical calculations ...... 135 5.4.2 Summary of predictions and uncertainties of the invisible-muon com- ponent...... 140 5.4.3 Comparison with data and implication for nuclear gamma rays . . . 143 5.5 New results on Super-K Low-Energy Atmospheric data: (νe +ν ¯e) CC component...... 143 5.6 Parent Neutrino Distributions...... 146 5.7 Conclusions...... 147
Bibliography 149
x List of Figures
Figure Page
1.1 Spectrum of neutrinos from astrophysical and other sources. The “Back- ground from old supernovae” corresponds to the diffuse supernova neutrino background in this dissertation. Figure taken from [14]...... 2 1.2 Dark matter annihilation at the center of the Sun. Left: standard scenario, in which standard model particles are produced directly and only neutrinos can get out. Right: long-lived dark mediators are produced and decay outside the Sun to standard model particles, including gamma rays and others that produce gamma rays. Figure taken from [15]...... 5 1.3 Integral intensity profiles above 500 MeV for the two components of solar gamma rays. Red points are the observed counts by F ermi-LAT. Dotted magenta line is the disk emission. It extends to larger than 0.26◦ due to the angular resolution of F ermi-LAT. Dashed green line is the halo emis- sion (labeled as IC emission on the figure). Dash-dotted horizontal line is the background. Solid blue line is the sum of the background and the two components of the emission. The shaded areas around the lines show total error estimates. Figure taken from [16]...... 6 1.4 Status of the solar gamma rays after our series of work [3–7]. The dashed green line is the disk flux without magnetic effects [7], which also serves as a theoretical lower bound. The blue and red points show the Fermi- LAT measurements during and outside the solar minimum respectively [5,6]. There is a clear dip feature in each dataset. The nominal prediction from Ref. [17] is shown as the green band. The dashed blue line is a theoretical upper bound of disk flux [6]. The black upper limit is from HAWC 3 years of observation [4]. The dotted cyan line and purple solid line are the sensitivity of future LHAASO [18, 19] and 100000 m2 water Cherenkov detector (WCD) array in the Southern Hemisphere [20] respectively. Figure taken from [21]. 10 1.5 Three possible kinds of messengers to identify the sources of high-energy cosmic rays, including cosmic rays on their own, gamma rays, and neutrinos. Figure taken from Ref. [22]...... 12 1.6 Diagrams for deep-inelastic scattering. Left: CCDIS for neutrino, which is mediated by a W boson. Right: NCDIS, which is mediated by a Z boson. Figure taken from Ref. [23]...... 13
xi 1.7 Cross sections between neutrinos and 16O for CCDIS [24], NCDIS [24], and the Glashow resonance (¯ν e− W −, taking into account eight e → electrons) [25]. Figure is original to the thesis...... 15 1.8 Typical event signatures observed by IceCube. Left: An elongated track, mostly from νµ CCDIS and a small fraction from ντ CCDIS. Middle:A shower event, from NCDIS, νe CCDIS, or ντ CCDIS if the final-state τ lepton does not decay to a muon and no double-bang signature formed. Right: Double-bang, from ντ CCDIS when the τ lepton is above hundreds TeV. Each small colored spheres marks a photomultiplier tube that is triggered by Cherenkov photons, and the size of a sphere indicates the amount of photons registered. The colors represent the relative triggered time of the photomultiplier tubes, and the earlier, the redder, while the later, the bluer. The left and middle are from IceCube’s observation while the third is from simulation. Figure taken from Ref. [26]...... 16 1.9 IceCube’s 6 years of data for astrophysical neutrinos per neutrino flavor (black points with σ errorbar). The atmospheric neutrino components have been subtracted from a combined likelihood fit. The dashed blue line shows the best-fit conventional atmospheric neutrinos flux, and the dashed green line shows the best-fit upper limit on “prompt” neutrinos. The solid blue line with 1σ error band is the best fit using a single power-law spectrum. The pink line with 1σ error band shows the best fit using muon neutrinos only [27]. Figure taken from Ref. [28]...... 17 1.10 DSNBν ¯e spectrum for different neutrino emission spectra, as labeled. For each, two curves are plotted to include the uncertainties mainly due to the cosmic rate of core collapses. The shadings represent backgrounds, and their origins are labeled. Figure taken from Ref. [29]...... 21 1.11 Current status of DSNB searches in Super-K. Solid black lines are the DSNB event rate calculated by Ref. [29] based on an effectiveν ¯e temperature of 6 MeV. Two curves are plotted to include the uncertainties mainly due to the cosmic rate of core collapses. Shaded areas represent the backgrounds as labeled, details below. Colored lines represent atmospheric backgrounds in the current Super-K search window (> 16 MeV). Green lines are for the νµ andν ¯µ components, while red lines are for the νe andν ¯e components. Dashed lines are for current Super-K, while solid lines are for the Super-K-Gd phase. Figure is original to the thesis...... 24
2.1 Prospects for TeV solar gamma-ray observations, illustrated with the disk emission (details in Fig. 2.5). Points: observations with Fermi [16, 30], where the flux difference is due to time variation. Green band: the only theoretical prediction that includes magnetic effects [17]. Dashed lines: the estimated differential point-source sensitivity of HAWC [31] (scaled to one year) and LHAASO [18, 19]...... 29 2.2 Solar mass density as function of height above the photosphere (left axis, blue dashed), as well as the same for optical depth for inelastic proton-proton collisions (right axis, red solid)...... 36
xii 2.3 Solar-limb gamma-ray spectrum produced by hadronic cosmic rays. Red dotted line: semianalytic result for proton-proton interactions with 0 < τ < 0.3. Green dash-dotted line: GEANT4 results for the full range of τ; the gradual cutoff is because it cannot simulate proton interactions above 100 TeV. Blue dashed line: our empirical fit to the GEANT4 results, extrapolated to higher energies. Black solid line: our full prediction, including a correction factor for nuclei. The light grey shading approximately indicates the energies at which magnetic effects, neglected here, should be included...... 38 2.4 Normalized relative contributions of different τ values to the predicted gamma-ray flux, based on our GEANT4 simulation. We show the example of Eγ = 1 TeV; other energies give similar results...... 39 2.5 Gamma-ray spectrum of the Sun. Points: disk observations with Fermi [16, 30], where the flux difference is due to time variation. Green band: the predicted disk flux [17]. Dotted lines: the estimated differential point-source sensitivity of HAWC [31] (scaled to one year) and LHAASO [18, 19]. Our new prediction of the solar-disk signal due to cosmic-ray hadrons (from the limb) is shown by the green solid line. Our new prediction of the solar-halo signal due to inverse-Compton scattering of cosmic-ray electrons is shown by the black solid line for the nominal case and by the dashed lines for enhanced cases from Fig. 2.6...... 43 3 2.6 Diffuse flux (weighted with Ee ) of cosmic-ray electrons. Below about 5 TeV, there are measurements (points, as labeled [32–35]). Above about 70 TeV, there are limits (gray region, which combines many experiments [36, 37]). In between, the spectrum could be as large as the blue solid line, allowing en- hanced contributions (pulsar or dark matter; details are in the text). HAWC should be able to immediately improve sensitivity down to 10−3 (hadronic ∼ rejection) of the proton spectrum (red dashed line)...... 45
3.1 Diagrams for (on-shell) W -boson production via photon exchange. A and A0 are the initial- and final-state nuclei. (See Fig. 3.5 and Sec. 3.3.2 for the connection with trident production.) For antineutrinos, take the CP transformation of the elementary particles...... 52 3.2 Diagrams for trident production via photon exchange in the four-Fermi theory (see Fig. 3.5 for the full Standard Model). For antineutrinos, take the CP transformation of the elementary particles...... 54 3.3 Summary of cross sections for W -boson and trident production from previous work, with the two processes separated as labeled. To simplify the figure, for W -boson production, we show only ν e−W + on 16O (by Seckel [38] and e → Alikhanov [39, 40]), and for trident production, only the coherent regime (the dominant part) of ν ν e−µ+ on 40Ar (by Magill & Plestid [41], Ballett et µ → e al. [42], and Altmannshofer et al. [43]). Also shown, for comparison, is the cross section of charged-current deep inelastic scattering (CCDIS) [24]. . . 56
xiii 3.4 Our cross sections (actually σνγ(sνγ)/sνγ) for W -boson and trident produc- tion, between a neutrino and a real photon as a function of their CM energy. Red, green, and blue lines are νe-, νµ-, and ντ -induced channels, respec- tively. Solid lines are trident CC channels, and dashed lines are trident CC+NC channels (we label only the final states for both). Magenta dotted lines are trident NC channels, which depend on only the final-state charged leptons (we label both the initial and final states). The trident CC, NC, and CC+NC channels correspond to diagrams (1)–(3), (4)–(5), and (1)–(5) of Fig. 3.5. The corresponding antineutrino cross sections (i.e., obtained by CP-transforming the processes shown) are the same. See text for details. . 60 3.5 Diagrams for trident production via photon exchange in the Standard Model, with the order, Tri– Tri, labeled in parentheses, and with the momenta M1 M5 labeled on the fourth diagram. The trident CC, NC, and CC+NC channels correspond to diagrams (1)–(3), (4)–(5), and (1)–(5). For antineutrinos, take the CP transformation of the elementary particles. The first and second diagrams are connected to W -boson production (Fig. 3.1; also see Sec. 3.3.2 for details of the connection)...... 61 3.6 Our coherent and diffractive components of W -boson production cross sec- tions, ν `− + W +, on 16O. Red, green, and blue lines are ν -, ν -, ` → e µ and ντ -induced channels, respectively. Solid: coherent (right bump) and diffractive (left bump) components. Dashed: longitudinal contribution to the coherent regime, which is small, even for the largest case (νe). The ντ line is not shown due to being below the bound of the y axis. Dotted: contribution from neutrons to the diffractive regime, which is small. The corresponding antineutrino cross sections are the same...... 66 3.7 Our coherent and diffractive components of W -boson production cross sec- tions (red solid, from Fig. 3.6 but thicker), for the example of ν `−W + e → on 16O (the flavor with the largest cross section), comparing with our “EPA + no Pauli blocking” results (dashed) and previous calculations (dotted) by Seckel [38] and Alikhanov [39, 40]. Left and right bumps are coherent and diffractive components, respectively. Note our results are substantially smaller, which is important...... 67 3.8 Our coherent (solid lines) and diffractive (dashed lines) components of tri- dent production cross sections on 16O. We show one typical channel for each category, i.e., CC, NC and CC+NC, to make the figure simple. For all the channels, see Appendix 3.8.4. Gray lines are for νe-induced W -boson produc- tion from Fig. 3.6, shown as a comparison. The corresponding antineutrino cross sections (i.e., obtained by CP-transforming the processes shown) are the same...... 70 3.9 Different components of our inelastic neutrino-nucleus cross sections for W - boson production. Only νe is shown to keep the figure simple. For νµ and ντ , the photon-initiated cross sections are smaller (Fig. 3.10), while the quark- initiated cross sections are basically the same. See text for details...... 74
xiv 3.10 Our inelastic neutrino-nucleus cross sections for W -boson production on 16O (solid lines), for all three flavors. Also shown are previous results from Alikhanov [40] and, for comparison, coherent and diffractive cross sections of νe from Fig. 3.6. The corresponding antineutrino cross sections are the same. 75 3.11 Our cross sections for trident production in the inelastic regime. Left: CC channels. Right: CC+NC and NC channels. Solid: the photon-initiated subprocess. Dashed: quark-initiated subprocess. The corresponding an- tineutrino cross sections are the same...... 76 3.12 Our total cross sections (actually σνA/Eν) for W -boson and trident produc- tion on 16O. The colors and line styles are same as in Fig. 3.4( red, green, and blue lines are νe-, νµ-, and ντ -induced channels, respectively; solid lines are CC channels, and dashed lines are CC+NC channels; magenta dotted lines are NC channels, which depend on only the final-state charged leptons). The trident CC, NC, and CC+NC channels correspond to diagrams (1)–(3), (4)–(5), and (1)–(5) of Fig. 3.5. The corresponding antineutrino cross sec- tions (i.e., obtained by CP-transforming the processes shown) are the same. See text for details...... 78 3.13 Ratios of the W -boson production cross sections to those of CCDIS ((ν + ν¯)/2) [24]. Solid lines are for water/ice targets, dotted line for iron targets, and dashed lines are for the Earth’s averaged composition. Color assignment is in the legend. Also shown is the νe (iron) case of Seckel [38], much larger than ours...... 79 3.14 Our elastic cross sections for all trident channels. We add the coherent and diffractive components together to simplify the figure. The colors and line styles are same as in Fig. 3.4( red, green, and blue lines are νe-, νµ-, and ντ -induced channels, respectively; solid lines are CC channels, and dashed lines are CC+NC channels; magenta dotted lines are NC channels, which depend on only the final-state charged leptons). The trident CC, NC, and CC+NC channels correspond to diagrams (1)–(3), (4)–(5), and (1)–(5) of Fig. 3.5. Gray dashed lines are the coherent and diffractive cross sections for W -boson production from Fig. 3.6, shown as a comparison. For an- tineutrinos, which have the same corresponding cross sections, take the CP transformation of the channel labels. See text for details...... 90
4.1 Cross sections between neutrinos and 16O for W -boson production [2], com- pared to those for CCDIS [24], NCDIS [24], and the Glashow resonance (¯ν e− W −, taking into account eight electrons) [25]...... 94 e → 4.2 Left: Differential cross sections for W -boson production in terms of the energy of the charged lepton, shown for each neutrino flavor and two typ- 5 6 ical energies (Eν = 10 GeV and 10 GeV). The y axis is Edσ/dE = −1 (2.3) dσ/d log10 E, matching the log scale on the x axis, so that relative heights of the curves at different energies faithfully show relative contribu- tions to the total cross section. Right: Same, in terms of the energy of the W boson...... 96 4.3 Average energy of the charged lepton (`) and W boson, divided by Eν, for each neutrino flavor...... 99
xv 4.4 Ratios of the W -boson production cross sections [2] to those of CCDIS ((ν + ν¯)/2) [24]. Solid lines are for water/ice targets, dotted for iron targets, and dashed for Earth’s average composition. Color assignments are noted in the legend. For comparison, we also show the νe (iron) result of Seckel [38], which is much larger than ours...... 100 −σC 4.5 Upper: Neutrino attenuation factor, e , for νe in Earth. Dashed lines (ADIS) are for CCDIS and NCDIS without W -boson production. Solid lines (ADIS+WBP) include W -boson production. For attenuation factors below 0.1, the event rate is too low to use, which we denote by using thin lines. Lower: The relative change in the attenuation factor due to W -boson production. 103 4.6 Relative W -boson yields due to W -boson production (νl + A l + W + 0 − − → A ) and the Glashow resonance (¯νe + e W ). We use dΦ/dEν = −2.9 −2 −1 −1 → (Eν/1 GeV) cm s GeV with unit normalization. The yield from W - boson production is 20 times that from the Glashow resonance, which can ' be seen by logarithmic integration of the peaks. The CCDIS (cyan, dashed) and NCDIS (magenta, dashed) cases are shown for comparison, though they do not produce on-shell W bosons...... 105 4.7 Left: Shower spectrum (Upper) and detection significance of W -boson pro- duction (Lower) for the conservative case as regards identifying W -boson production events. Right: Same, but for the optimistic case. The main difference between the two cases is the change with the CCDIS channel. The shaded region below 60 TeV is below the IceCube threshold for cleanly identifying astrophysical neutrinos. We use 0.5 km3 as the approximate fidu- cial volume of IceCube [44], and assume 10 years of IceCube data. For E > 60 TeV, where W -boson production contributes 6 shower events, dep ' the cumulative detection significance should be 1.0σ for the conservative ' case and 3.2σ for the optimistic case. See text for details...... 106
5.1 Super-K’s present DSNB searches, which start at Ee = 16 MeV [45]. Note the green and red lines are in units of per MeV instead of per 4 MeV, to match the DSNB lines. The DSNB signal band is based on a Fermi-Dirac spectrum with a 6-MeV effective (after neutrino mixing) temperature [29]. The backgrounds due to atmospheric neutrinos are based on our calculations (details in Fig. 5.5, and the text). Soon, added Gd will reduce these and other detector backgrounds...... 121 5.2 Our calculated fluxes of atmospheric neutrinos, without and with neutrino mixing, following Refs. [46–48]. Left: Results for νµ andν ¯µ. Right: Results for νe andν ¯e (note that the figure goes to lower energies; the notch is because we use different calculations above and below 0.1 GeV). Forν ¯ in both panels, we multiply by 0.5 for clarity...... 130 5.3 The GENIE cross sections for CC interactions with various interactions and targets as labeled. We only show the most important channels. Left: Results for νµ andν ¯µ. Right: Results for νe andν ¯e (note that the figure goes to lower energies). Below 1 GeV, CCQE dominates, as shown for one specific ∼ channel in the left panel...... 131
xvi 5.4 Our calculated results (lines, as labeled) for charged-lepton spectra induced in Super-K by atmospheric neutrinos, compared their measured data (points with statistical uncertainties only) [49]. Left: Muons, combining FC single- ring, multi-ring and PC data. Right: Electrons, single-ring FC data only, which dominates at lower energies, so at high energies (shaded region), some disagreement is expected. Overall, the agreement is well within the estimated theoretical uncertainties, which are dominated by the systematics (shown as the green arrow in each panel), at the level of a few tens of percent. . . . . 132 5.5 Our final calculation (last column of Table 5.1) of atmospheric neutrino back- grounds compared to Super-K-I data [45] (note the different axis scales com- pared to Fig. 5.1). Even though the data and curves are shown with 4-MeV steps, we have converted them to units of per 1 MeV to match Fig. 5.1. The two components (dashed) and their sum (solid) are shown. Overall, our final calculation — an absolute prediction, not a fit — agrees well with Super-K data...... 141 5.6 Distribution of parent atmospheric neutrinos relevant to the invisible muons in our final calculation (Sec. 5.4.1). Left: (ν +ν ¯ ) CC component ( 80% µ µ ' of total). Right: NC π+ component ( 20% of total)...... 144 ' 5.7 Distribution of parent atmospheric neutrinos relevant to the (νe +ν ¯e) CC component in our final calculation (Sec. 5.5). Left: Results for 16 MeV < Ee < 55 MeV. Right: Results for 55 MeV < Ee < 90 MeV. These comprise about 45% and 55%, respectively, of the events in 16 MeV < Ee < 90 MeV. The boundary of 55 MeV is roughly where the invisible-muon and electron components of the background cross in Fig. 5.5...... 145
xvii List of Tables
Table Page
1.1 List of current and future high-energy neutrino detectors...... 18
3.1 Summary of the features of previous calculations and of this work. “+” and “ ” means “considered” and “not considered” in the calculation respectively. − “Full SM” means using full Standard Model, instead of four-Fermi theory. 57
4.1 Different final state particles, signatures, corresponding fractions, and counts in IceCube. The counts are for greater than 60 TeV deposited energy and 10 years of IceCube observations (or 1 year for IceCube-Gen2). The numbers in the “Channel” column are the maximal ratios to the CCDIS cross section with water/ice. The numbers in “W decay” and “τ decay” columns are the branching ratios. For the “Final state” and “τ decay” columns, we omit the neutrinos; “h” means hadrons. The unique signatures are in boldface. The “/” divides the cases in which the charged lepton from the initial interaction is undetectable or detectable, which, to a good approximation, is half-half. The “Fractions” column shows the fraction of that row relative to the whole channel, which is the multiplication between the branching ratios of W and τ decay...... 110
5.1 Our predicted numbers of decay electrons (16 MeV < Ee < 90 MeV) from invisible muons, using GENIE’s default model set. The “Naive” calculation is defined in Sec. 5.4.1, and its improvement to the “Standard” calculation is defined in Sec. 5.4.1. We build on the latter by including the “Coulomb” corrections (Sec. 5.4.1) and also the “Threshold” corrections (Sec. 5.4.1). We show results for two sets of assumptions about nuclear gamma rays, favoring the second set. All calculations include signal efficiency and other detector effects. Numbers in boldface are bottom-line numbers...... 138 5.2 Same as Table 5.1, but using the GENIE’s EffSFTEM model set...... 138
xviii Chapter 1 Introduction
1.1 What is exciting about neutrinos from the universe
Neutrinos, the most mysterious particle in the standard model of particle physics, provide an indispensable tool to learn about the universe. Tremendous progress, including break- throughs, has been made in the past decades and will likely be made in the next decade.
For example, the detection of solar neutrinos (Fig. 1.1) established the picture that nu- clear fusion powers the Sun and other stars. Soon, the solar neutrinos from CNO and hep channels could be measured, and the CNO neutrinos would be the best probe of solar metallicity [23]. The detection of supernova neutrinos from SN1987A (Fig. 1.1), though only about 20 events, verified the basic picture of supernova explosion [50–52]. Now we are much more prepared for the next supernova burst in the Milky Way or its satellite galaxies [53]. For example, for a supernova burst with a distance of 10 kpc, we can detect
104 neutrinos in different flavors. ∼ Recently, the first detection of TeV–PeV astrophysical neutrinos by IceCube has opened a new window for astrophysics [44, 54]. Moreover, detecting those neutrinos is probably the only way to identify the sources of cosmic rays with energies above a few PeV, a hundred- year question of cosmic rays. The number of detected events keeps increasing rapidly so the sources of those neutrinos, and hence the sources of their above-PeV parent cosmic rays, may soon be identified. In fact, there are already candidate source detections in association with a blazar flare [55] and a tidal disruption event [56].
There is also great potential to detect new astrophysical sources of neutrinos. The
1 Figure 1.1: Spectrum of neutrinos from astrophysical and other sources. The “Background from old supernovae” corresponds to the diffuse supernova neutrino background in this dissertation. Figure taken from [14].
diffuse supernova neutrino background (DSNB, Fig. 1.1) will soon be detected by Super
Kamiokande (Super-K) in the new Super-K-Gd phase, thanks to adding Gadolinium into
Super-K water to reduce lots of backgrounds for DSNB searches [57–59]. The solar TeV–
PeV neutrinos could also be detected by IceCube, and theoretical work is still needed to understand the emission [60–64].
In reverse, astrophysical neutrinos provide an indispensable tool to study particle physics.
Same as above, tremendous progress, including breakthroughs, has been made in the past decades and can be made in the next decade. For example, the solar neutrinos and atmo- spheric neutrinos (Fig. 1.1) offered the first hint and evidence for neutrino mixing [65, 66].
Subsequent observations provide the most precise measurements of relevant parameters in the standard model [49, 67–69]. Recently, the cross sections of TeV–PeV neutrinos were
2 first measured using data from IceCube [70, 71], and the precision will increase significantly.
Neutrinos are the only particles in the standard model without a known mass origin, inspiring many physicists to study and to connect this with other big questions of particle physics, e.g., dark matter, matter-anti-matter asymmetry, and more [72–77]. Lots of such new physics have been tested using astrophysics neutrinos, thanks to the fact that they have a wide energy spectrum, directional information, and travel cosmic distances and through extremely high column densities (e.g., through Earth) [78–88].
The work presented in this dissertation focuses on three aspects of those mentioned above. First is calculating the TeV solar gamma-ray fluxes, as a step to understanding how cosmic rays interact with the Sun. This will eventually develop the Sun as a new high-energy laboratory, including as a neutrino source. Second is studying the high-energy neutrino interactions that are subdominant to the deep-inelastic scattering. It is high time to do this, due to our increasing data statistics of the TeV–PeV neutrinos. Third is studying the physics of the backgrounds for DSNB searches. This gives hints for reducing those background, which will significantly help the first detection and subsequent precision measurement of DSNB. The rest of this chapter is organized as follows. In Sec. 1.2, I present the motivation and prior work of developing the Sun as a new high-energy laboratory, including as a neutrino source. This motivates my work in Chapter2. In Sec. 1.3, I present the motivation and prior work of detecting high-energy neutrinos, motivating my work in
Chapters3 and4. In Sec. 1.4, I present the motivation and prior work of detecting the diffuse supernova neutrino background, motivating my work in Chapter5
1.2 Developing the Sun as a new high-energy neutrino source
One way to make progress in neutrino astrophysics is finding new astrophysical sources of neutrinos. There is great potential for the Sun being a source of high-energy neutrinos [60–
64]. These neutrinos are predicted from cosmic rays interacting with the solar atmosphere.
However, the predicted flux has large uncertainties, especially at lower energies, due to our insufficient knowledge of solar magnetic fields and interplanetary magnetic fields, which is
3 hard to study directly.
Solar gamma rays, as also products from cosmic-ray interactions, provide an excellent tool to study how those magnetic fields affect cosmic rays. Moreover, the gamma-ray spec- trum provides a benchmark for the neutrino spectrum because both are from decay of pions from cosmic ray interactions. This is similar to how people predict the neutrino output of other astrophysical objects, like supernovae and active galactic nuclei, using measurements of their electromagnetic output.
Throughout this dissertation, we consider gamma rays only from the quiescent, steady- state Sun. We do not consider gamma rays from solar flares [89–91], as their origin is different in nature than those from the quiet Sun.
1.2.1 Motivations
There are more motivations to study solar gamma rays than developing the Sun as source of high-energy neutrinos.
The most obvious motivation, as mentioned above, is studying the solar magnetic fields and interplanetary magnetic fields on their own and how they affect the propagation of cosmic rays. Energy and location-dependent studies can be done with solar gamma rays, which are hard with directly measuring cosmic rays. There are small cosmic-ray detectors launched into the inner solar system but they have limited ability compared to those on or near the Earth [92, 93]. For example, those detectors have limited reach and only measure the total number of cosmic rays without energy information.
Moreover, studying solar gamma rays is essential to developing the Sun as a laboratory to test new physics [3, 15, 21, 60–64, 94–96]. Searching for new physics usually suffers from backgrounds from astrophysical processes, which should be understood well. For example, it is well-motivated that dark matter particles could be trapped by the Sun’s gravitational potential and accumulate at the center of the Sun. Those dark matter particles could annihilate to long-lived metastable mediators, travel outside the Sun, and eventually decay to gamma rays or standard model particles that produce gamma rays [3, 15]. This popular scenario and, as a comparison, the standard scenario are shown in Fig. 1.2.
4 γ (...... ) γ (extinguished) ν (less attenuated) γ (extinguished) γ ν γ γ › ν ν γ, ν ν (attenuated) (unattenuated)
Short-lived mediators Long-lived mediators
Figure 1.2: Dark matter annihilation at the center of the Sun. Left: standard scenario, in which standard model particles are produced directly and only neutrinos can get out. Right: long-lived dark mediators are produced and decay outside the Sun to standard model particles, including gamma rays and others that produce gamma rays. Figure taken from [15].
The last motivation is more from curiosity. The Sun has been studied in depth in multi- wavelength electromagnetic signals but not in gamma rays. We would like to complete this multi-wavelength in-depth study. Moreover, the Sun is a typical G-type main-sequence star is the universe. Therefore, the above benefits also apply to studying other stars in the universe.
1.2.2 Review of prior work
There are two components of solar gamma rays, with different emission regions, spectra, morphologies, and physical origins. They are solar disk gamma rays and solar halo gamma rays, shown in Fig. 1.3.
The solar disk gamma rays are emitted from the disk of the Sun, with an angular radius of about 0.26◦. It is point-like in terms of the resolution of current gamma-ray telescopes except at above 10 GeV energies by F ermi-LAT. They are likely produced by hadronic cosmic rays interacting with the matter in the Sun’s photosphere and convection zone. Because the physics of gamma-ray production and solar matter density are well
5 Figure 1.3: Integral intensity profiles above 500 MeV for the two components of solar gamma rays. Red points are the observed counts by F ermi-LAT. Dotted magenta line is the disk emission. It extends to larger than 0.26◦ due to the angular resolution of F ermi-LAT. Dashed green line is the halo emission (labeled as IC emission on the figure). Dash-dotted horizontal line is the background. Solid blue line is the sum of the background and the two components of the emission. The shaded areas around the lines show total error estimates. Figure taken from [16].
understood, observing the disk emission provides us with the information of interplanetary
and solar magnetic fields.
Those gamma rays were first mentioned in Ref. [97] and the flux was first estimated
(based on measurements of terrestrial emission) in Ref. [98]. Ref. [99] suggested them to be
detectable by the EGRET experiment on board the Compton Gamma-Ray Observatory. In
1991, Seckel, Stanev, and Gaisser made the first, and so far the only, detailed theoretical
study of this emission [17]. In that work, they modeled the effects of interplanetary magnetic
fields on cosmic rays as a purely spatial diffusion and time-independent problem. The solar
6 magnetic fields are modeled as vertical and homogeneously distributed magnetic flux tubes that root deeply in the convection zone of the Sun and stretch out into corona. Cosmic rays, after being modulated by the interplanetary magnetic fields, which reduces the fluxes, reach the magnetic flux tubes homogeneously and isotropically, get trapped, and propagate along a helical line into the Sun. Due to the mirror effect [100], most of those charged particles are turned around and leave the Sun at some point. When cosmic rays are close to and in the Sun, they interact with the matter of the Sun and develop hadronic showers.
The gamma rays that are produced when the parent particles are leaving the Sun could reach us as the disk emission. Therefore, Ref. [17] predicts a homogenous, isotropic, and steady-state disk emission.
For the observation of disk gamma rays, in 1995, EGRET set an upper limit of 2 × 10−7 cm−2 s−1 on the flux using four years of observation data [101]. The first detection of the disk emission was in 2018, after a reanalysis of the EGRET data [102].
The launch of F ermi-LAT gamma-ray telescope heralded a new era of studying solar gamma rays. In 2011, the F ermi-LAT collaboration analyzed their first 1.5 years of data
(0.1–10 GeV) from the direction of the Sun [16]. Their results show that the disk flux is about 7 times higher, and the spectrum is harder, than the prediction by Ref. [17]. In 2015,
Ref. [30] analyzed 6 years of public Fermi-LAT data and detected the disk emission up to
30 GeV. It also confirmed the high flux and hard spectrum found by Ref. [16]. Surprisingly, the work found an anticorrelation between solar activity and gamma-ray flux of 1–10 GeV, with unknown reasons.
Solar halo gamma rays were first studied by Refs. [102–104]. They are produced by cosmic-ray electrons (electrons means both electrons and positrons in this section) inverse-
Compton scattering the photons ( 1 eV) emitted by the Sun. The solar photons are emit- ∼ ted from the surface of the Sun (in the photosphere) and the density follows 1/r2, where ∼ r is the distance from the Sun’s center. Therefore, the halo emission can extend to tens of degree, and the further from the Sun, the lower the intensity. Because inverse-Compton scattering and solar photon density are well understood, observing the halo emission pro- vides us with the information of interplanetary magnetic fields.
7 On the observation side, the halo emission was also first detected by Ref. [102] using
EGRET data and then by F ermi-LAT [16]. Their results are consistent with theoretical predictions of Refs. [102–104].
1.2.3 My work
In this section, I give an overview of the motivations and key ideas for my work on solar gamma rays.
As mentioned before, the long-term goal is to develop the Sun as a high-energy labora- tory, including to comprehensively understand solar disk and halo gamma-ray emission, to understand how the solar and interplanetary magnetic fields affect the parent cosmic rays, to develop the Sun as a high-energy neutrino source, and to develop the Sun as a lab to test new physics. However, this journey is far from easy, as the solar gamma rays on their own have many problems unsolved, i.e., unexplained observational features including the high
flux and hard spectrum of disk gamma rays mentioned above. (In fact, our further work found more unexplained features, details below.)
A key step is that, because the problems above are highly likely due to insufficient knowledge of solar and interplanetary magnetic fields, we need to know what the emission is like if there were no magnetic effects. At GeV energies where the magnetic effects are strong, comparing the predicted gamma-ray spectrum between without and with magnetic effects and with observation gives us precious information on how magnetic fields play a role. The other key step is understanding the emission at TeV energies. As HAWC [31] and
LHAASO [18, 19] will have data soon, connecting theory and observation will provide us with very valuable information. The TeV flux is not much affected by neglecting magnetic effects, because the parent cosmic rays are too energetic. Therefore, the above two steps can be combined, i.e., calculating solar gamma-ray spectrum at both GeV and TeV ener- gies without magnetic effects, though we need to deal with the disk and halo components separately.
My work in Ref. [7], presented in Chapter2, aims to do the above calculations, which have never been done before. The disk gamma rays without magnetic effects (hence no
8 mirror effects) are only from the solar limb, where cosmic rays graze the Sun and propagate through a column density that is large enough for interactions while small enough for gamma rays to escape. We calculate the flux in two ways, one is semi-analytic and the other is using full simulation by Geant4 [105, 106]. This assures the robustness of the results and helps understand relevant physics. We use the most updated hadronic cosmic-ray data [107–109].
Moreover, our calculation includes the effects from multiple scattering, absorption, cascade processes, and heavier nuclei than proton. The results serve as an important theoretical lower bound on the disk emission at both GeV and TeV energies. Comparing to the case with magnetic effects and observation, it shows that magnetic effects enhance the disk flux by a factor of 10 at GeV energies. HAWC and LHAASO will be able to see if the ∼ enhancement continues at TeV energies and if there are contributions due to new physics.
For the halo gamma rays, we use the StellarICs code [110, 111]. The flux is lower than the sensitivity of HAWC and LHAASO, but they can at least set constraints at these energies, where there are no measurements. Moreover, we show that the halo gamma-ray emission can be used to probe the density of TeV cosmic-ray electrons in the inner solar system, which is not well measured.
The above work [7] led by me is part of a larger program of work to develop the Sun as a new high-energy laboratory. It also sets a theoretical foundation for our further studies on the disk emission which I participated in and are published in Refs. [3–6] (not included in this dissertation). In Refs. [5,6], we analyzed 9 years of F ermi-LAT data and ' 1. Detected significant disk gamma-ray emission from 1 GeV up to 200 GeV (Fig. 1.4);
−2.2 2. Further confirmed the hard spectrum (Eγ , Fig. 1.4) compared to the prediction −2.7 (Eγ ) and the anticorrelation between gamma-ray flux and solar activity. 3. Observed a spectral dip between about 30 and 50 GeV in an otherwise power-law
spectrum (Fig. 1.4), which is surprising and still unexplained;
4. Found that the gamma-ray emission is dominated by the equatorial region of the
disk during solar minimum and by the polar region outside solar minimum, which
strongly suggests that the disk emission is produced by two separate mechanisms.
In Ref. [4], we worked with HAWC collaboration and set strong constraints on the flux
9 10 10− HAWC 95% C.L. (2014–2017) Fermi-LAT (2014–2017) Fermi-LAT (Solar Min.) ] 1 − s 2
− 11 10− CR Upper Bound SSG1991 LHAASO 1yr
12 10− 100,000 m
CR Lower Bound 2 WCD Array 1 yr Energy Flux [ TeV cm
13 10− 2 1 1 10− 10− 1 10 Energy [ TeV ]
Figure 1.4: Status of the solar gamma rays after our series of work [3–7]. The dashed green line is the disk flux without magnetic effects [7], which also serves as a theoretical lower bound. The blue and red points show the Fermi-LAT measurements during and outside the solar minimum respectively [5,6]. There is a clear dip feature in each dataset. The nominal prediction from Ref. [17] is shown as the green band. The dashed blue line is a theoretical upper bound of disk flux [6]. The black upper limit is from HAWC 3 years of observation [4]. The dotted cyan line and purple solid line are the sensitivity of future LHAASO [18, 19] and 100000 m2 water Cherenkov detector (WCD) array in the Southern Hemisphere [20] respectively. Figure taken from [21].
of TeV solar gamma rays (Fig. 1.4). In Ref. [3], also with HAWC collaboration, we set strong constraints on the scenario of a long-lived mediator from dark matter annihilation at the center of the Sun.
1.3 Diffuse High-Energy Astrophysical Neutrinos
The recent detection of extraterrestrial high-energy neutrinos by IceCube is one of the breakthroughs in neutrino physics and astrophysics [44, 54]. It also opened a new era for neutrino astronomy. Since then, we have made tremendous progress on both the theoretical and experimental sides. These include studying high-energy neutrino scatterings, poten- 10 tial production of those neutrinos by various astrophysical objects, detection techniques, source identification, and more. Those studies are still underway and will lead to huge accomplishments.
1.3.1 Motivations
High-energy astrophysical neutrinos provide the best tool to identify the sources of high- energy cosmic rays. Although we have studied cosmic rays for more than 100 years, with tens of experiments, we are still not sure of their origins. This is probably the biggest open question for cosmic rays. For cosmic rays below a few PeV, there is more and more evidence that they are from supernova remnants in the Milky Way. But for higher energy ones, we know very little about their sources except that they should be extragalactic.
Fig. 1.5 illustrates the reasons. We could just use cosmic rays on their own. However, since they are charged, the magnetic fields in the host galaxies, intergalactic medium, and the Milky Way will deflect them. Therefore, they lose their directional information. Another possible messenger is their interaction products like gamma rays and neutrinos which have directional information because they are neutral. For gamma rays, however, since they are high energy and extragalactic, they are absorbed by cosmic microwave background
(CMB) photons or extragalactic background light, via γ +γ e+ +e−. Therefore, CMB/EBL → neutrinos are the best tool to identify the sources of high-energy cosmic rays, as they have
directional information and are not absorbed by matter or radiation during propagation.
Moreover, as a result of above, neutrinos are also the best messenger to study particle
acceleration mechanisms of those sources, which is also a big and long-term question.
Lastly, high-energy neutrinos have a unique power to be used to study neutrino proper-
ties and test new physics [78–88, 112–114]. This is thanks to them having very high energy
and directional information, and traveling cosmic distances and through extremely high
column densities.
11 Figure 1.5: Three possible kinds of messengers to identify the sources of high-energy cosmic rays, including cosmic rays on their own, gamma rays, and neutrinos. Figure taken from Ref. [22].
1.3.2 Review of prior work
To study high-energy neutrinos, we need first detect them, which is a big topic. Neutrinos can not be detected directly, but instead through the secondary particles from neutrino interactions with detector material (for IceCube, it is ice).
The cross section of a neutrino with the nucleus is much larger than that with the elec-
12 Figure 1.6: Diagrams for deep-inelastic scattering. Left: CCDIS for neutrino, which is mediated by a W boson. Right: NCDIS, which is mediated by a Z boson. Figure taken from Ref. [23].
trons, which is suppressed by m /m , where m and m are masses of electron and proton, ∼ e p e p respectively (an exception is the Glashow resonance,ν ¯ + e− W − any final states, e → → details below).
Deep inelastic scattering (DIS) dominates the neutrino-nucleus scattering, and has been a focus of study. It includes charged-current DIS (CCDIS), which can be written as
ν + A `− + X for neutrinos and (1.1) ` → ν¯ + A `+ + X for antineutrinos , (1.2) ` → and neutral current DIS (NCDIS), which can be written as
(−) (−) ν +A ν +X (1.3) ` → ` where A is the nucleus, ` the charged lepton, and X represents any hadronic final states.
Fig. 1.6 shows the diagrams for the CCDIS, mediated by a W boson, and NCDIS, mediated by a Z boson. What the weak boson couples to is actually a quark in the nucleon in a nucleus, because of the large momentum transfer.
Fig. 1.7 shows the cross sections for CCDIS and NCDIS. The cross sections of NCDIS are about 40% of CCDIS. Neutrinos have larger cross sections than antineutrinos. This is
13 because neutrinos couple to d andu ¯ quarks while antineutrinos couple to u and d¯ quarks,
and these quarks have different densities in a nucleon. At higher energies, neutrinos and
antineutrinos have the same cross sections because they dominantly couple to sea quarks.
The cross sections increase following E1 below 103 GeV (not shown) then the increasing ∼ ν ∼ slows down and finally saturates at E0.3 above 106 GeV. Also shown on the figure is the ν ∼ Glashow resonance,ν ¯ + e− W − any final states. It peaks at E = m2 /2m e → → ν W e ' 6.3 106 GeV. The width of the peak is due to the decay width of the W boson. The × Glashow resonance peaks at a very large cross section, but this is only for one flavor and
happens at a very narrow range of neutrino energy.
IceCube is an ice-Cherenkov detector and it detects Cherenkov photons with its pho-
tomultiplier tubes. When a charged particle travels at a speed faster than the light speed
in the medium, it produces Cherenkov radiation. Among the interaction products of DIS,
IceCube and other high-energy neutrino detectors can detect the final-state charged leptons
and hadrons. The volume of IceCube is about 1 km 1 km 1 km. × × Fig. 1.8 shows three possible event signatures that can be formed from the topology of
triggered photomultiplier tubes in the IceCube detector. The first is an elongated track
from a high-energy muon. The second is a shower from a high-energy electron that ini-
tiates electromagnetic cascade or from high-energy hadron(s) that initiate hadronic cas-
cade. Currently, IceCube could not distinguish electromagnetic and hadronic showers, but
this is possible in IceCube-Gen2 [115] (next generation of IceCube) with the echo tech-
nique [116]. For the τ lepton, before decaying to muon, electron, or hadrons, it could
leave a track. However, due to its short lifetime (t 2.9 10−13 s), it only travels ' × γc t 8.7 10−5 (E /1.78 GeV) m. Taking into account the minimal spacing between ' × × τ IceCube’s photomultiplier tubes, 10 m, this means that if the τ lepton is above hundreds ∼ TeV, the ντ CCDIS could form a double-bang signature, the third in the figure, with the first bang from the primary hadrons and the second from the displaced decay of the τ lep-
ton. IceCube has already identified two double-bang candidates [117]. (See Chapter4 for
more about IceCube and its detection technique.)
From those signatures, including their Cherenkov-light yields, spatial and temporal pro-
14 30 10 Glashow resonance ( e) 31 10 ]
2 m c
[
32
10 CCDIS NCDIS 33 10
34 10 4 5 6 7 8 10 10 10 10 10 E [ GeV ]
Figure 1.7: Cross sections between neutrinos and 16O for CCDIS [24], NCDIS [24], and the Glashow resonance (¯ν e− W −, taking into account eight electrons) [25]. Figure is e → original to the thesis.
files, we can reconstruct a neutrino’s energy, arrival direction, and flavor information, be- cause the track and double-bang signatures are mainly from νµ and ντ CCDIS respectively (τ lepton also decays to a muon, with a branching ratio of only 17%). The angular resolution of track events is less than 1 degree while 5–20 degrees for shower events. The difference is due to the topologies of the two signatures described above. The energy resolution of track events is only about a factor of 2, because tracks could leave the detector before depositing all the energies, and for shower events, it is about 15% as most of the showers deposit all of
15 Figure 1.8: Typical event signatures observed by IceCube. Left: An elongated track, mostly from νµ CCDIS and a small fraction from ντ CCDIS. Middle: A shower event, from NCDIS, νe CCDIS, or ντ CCDIS if the final-state τ lepton does not decay to a muon and no double-bang signature formed. Right: Double-bang, from ντ CCDIS when the τ lepton is above hundreds TeV. Each small colored spheres marks a photomultiplier tube that is triggered by Cherenkov photons, and the size of a sphere indicates the amount of photons registered. The colors represent the relative triggered time of the photomultiplier tubes, and the earlier, the redder, while the later, the bluer. The left and middle are from IceCube’s observation while the third is from simulation. Figure taken from Ref. [26].
their energies in the detector. IceCube could not distinguish neutrinos from antineutrinos, as it could not distinguish the charge signs. With the above abilities, IceCube can mea- sure astrophysical neutrinos in terms of fluxes, time variations, directions, spectra, flavor compositions, and more.
Fig. 1.9 shows IceCube’s data for astrophysical neutrinos per flavor. The data have already had the atmospheric neutrino components subtracted from a combined likelihood
fit, which is consistent with the theoretical prediction. The astrophysical neutrino spectrum can be fitted using a single power law,
+0.2 −(2.89−0.19) dΦν +1.46 Ev −18 −1 −2 −1 −1 = 6.45−0.46 10 GeV cm s sr , (1.4) dEν 100 TeV ·
16 Figure 1.9: IceCube’s 6 years of data for astrophysical neutrinos per neutrino flavor (black points with σ errorbar). The atmospheric neutrino components have been subtracted from a combined likelihood fit. The dashed blue line shows the best-fit conventional atmospheric neutrinos flux, and the dashed green line shows the best-fit upper limit on “prompt” neutri- nos. The solid blue line with 1σ error band is the best fit using a single power-law spectrum. The pink line with 1σ error band shows the best fit using muon neutrinos only [27]. Figure taken from Ref. [28].
for all three flavors.
As shown in the figure, the spectrum is not well measured, due to a lack of data. The
flavor compositions are also poorly constrained. Moreover, the origin of those astrophysical neutrinos are not known yet, except that they should mostly be extragalactic. There are candidate source detections in association with a blazar flare [55], but the significance is less than 4σ and we are not sure if they could contribute all those astrophysical neutrinos.
Therefore, we need more data, hence more and larger detectors, for high-energy neutri- nos.
17 1.3.3 My work
In this section, I give an overview of the motivations and key ideas for my work on high- energy neutrinos.
Table 1.1: List of current and future high-energy neutrino detectors.
Detector Size Detector material Lcoation Status IceCube 1 km3 Ice South Antarctica Running for 8 years ' KM3NeT 1 km3 Water Mediterranean Sea Upcoming Baikai-GVD 1 km3 Water Lake Baikal Upcoming IceCube-Gen2 10 km3 Ice South Antarctica Proposed
Table 1.1 shows current and future high-energy neutrino detectors, which shows that the future of high-energy neutrinos is promising. We are about to have more and much larger detectors, hence much more data. The rapidly increasing amount of data demand a deeper understanding of neutrino interactions, which plays a key role in neutrino detections, in order to have more precise measurements. The deep inelastic scattering has been a focus for theorists. After tremendous effort, the claimed theoretical precision (from the parton- distribution functions) has reached 2% for TeV–PeV neutrinos. ' However, there are subdominant interactions that are larger than the claimed uncer- tainties of DIS, and even large enough that they should be included when analyzing data.
The largest subdominant processes are actually through photon coupling to the nucleus, as a neutrino can split into a charged lepton and a charged boson that can couple to a pho- ton. Specifically, they are W -boson production and trident production. Since CCDIS is the dominant process for detecting high-energy neutrinos, we take it for comparison. Previous calculations of trident production show that the cross sections are only 10−5 of CCDIS ∼ at GeV energies, but they increase with energy much faster than CCDIS [41–43, 118, 119].
So it is very important to know how large they are at TeV–PeV energies. On the other hand, previous calculations of W -boson production shows that the cross sections are as large as 10% of CCDIS [38–40]. However, those calculations have shortcomings so a new
18 and complete calculation is eagerly needed. Then, if our calculations show the cross sections are large, we should study how they affect the detection of high-energy neutrinos.
In my first work on this topic [2], presented in Chapter3, I did the first calculation of trident production at TeV–PeV neutrino energies and the first complete calculation of
W -boson production. We divided every process into three regimes, in which the whole nucleus, a nucleon, and a quark is involved, respectively. The cross section of a process is the sum of the cross sections in all three regimes. For the W -boson production, we found that the previously used equivalent photon approximation is not valid, so we use a complete formalism. More improvements in the calculations compared to earlier work are in
Chapter3. As a result, our cross sections of W -boson production are about half of before, which is very important in terms of the future statistics of high-energy neutrinos.
For the trident production, previous calculations only used four-Fermi theory, which significantly simplifies the calculation, because they were interested in lower energies. How- ever, for TeV–PeV energies, the four-Fermi theory does not work so we need to use the full standard model. Our calculation shows the cross sections are as large as 0.5% of CCDIS.
More interestingly, we found that trident and W -boson production are related processes for
Eν & 7 TeV. In my second work on this topic [1], presented in Chapter4, we studied how W -boson production affects the detection of TeV–PeV neutrinos, because in Ref. [2] we found their cross sections are large enough for detecting. We found that W -boson production should be taken into account by IceCube, KM3NeT, and Baikal-GVD, and must be taken into account by IceCube-Gen2.
1.4 Diffuse Supernova Neutrino Background
Another new astrophysical source of neutrinos that we can detect is the diffuse supernova neutrino background (DSNB) [120].
A supernova burst can produce an enormous amount of neutrinos. For a galactic su- pernova, we expect to detect thousands of neutrinos, but the burst rate is only a few per
19 century. We definitely do not want to wait that long. Instead, we integrate all the super- nova in the whole universe. Though the detectable neutrinos per burst is tiny, the burst rate is huge. This gives us a guaranteed steady neutrino flux, the DSNB. Detecting DSNB is vital to neutrino physics and astrophysics. In this section, I first give a more detailed introduction of DSNB and, as a result, the importance of detecting it (Sec. 1.4.1). Then,
I review the prior work on searching for DSNB by Super-Kamiokande (Sec. 1.4.2). Last, I briefly introduce my work (Sec. 1.4.3).
1.4.1 Motivations
DSNB is a predicted isotropic flux of neutrinos and antineutrinos in all flavors produced from all the massive-star core collapses throughout the universe. Here the core collapses include both supernovae and those that directly produce back holes without detectable electromagnetic emission, i.e., dark collapses [121]. Such dark collapses can not be studied by electromagnetic signals at all.
The DSNB flux can be calculated by a cosmological line-of-sight integral [29, 120],
Z ∞ dφ c dt (Eν) = [(1 + z) ϕ (Ev(1 + z))] R(z) dz , (1.5) dEν 0 dz where z is redshift and c is the light speed. dt/dz is the differential distance and dt/dz −1 = | | | | 31/2 H0(1+z) ΩΛ + Ωm(1 + z) , where H0,Ωm, and ΩΛ are cosmological parameters, which are relatively very well known.
There are two major ingredients in Eq. (1.5). First is ϕ (Ev(1 + z)), the neutrino emis- sion per core collapse, which is the primary observable, because it can only be determined by neutrino experiments. It is also the most uncertain part, as we only observed about 20 neutrinos from SN1987A [50–52] and the explosion mechanism and neutrino mixings are poorly understood theoretically.
The other ingredient is R(z), the cosmic core-collapse rate, which is relatively well known, as it can be measured by electromagnetic observations. The dominant contribution to DSNB is from z 1, as the corresponding core-collapse rate is 10 times the present ∼ ∼ rate.
20 10
8 MeV 6 MeV ] 4 MeV -1 SN1987A 1 MeV
-1 Invisible µ s -2 [cm ν 0.1 / dE φ
d Invisible µ + Reactor νe Spallation 0.01 0 10 20 30 40 Eν [MeV]
Figure 1.10: DSNBν ¯e spectrum for different neutrino emission spectra, as labeled. For each, two curves are plotted to include the uncertainties mainly due to the cosmic rate of core collapses. The shadings represent backgrounds, and their origins are labeled. Figure taken from Ref. [29].
Figure 1.10 shows the spectrum of DSNBν ¯e, which is the most promising flavor to be detected, calculated by Ref. [29]. The calculation was based on a widely-used Fermi-Dirac
spectrum with zero chemical potential [122, 123], which characterizes the time-integrated
ν¯e spectrum per supernova very well. Different styles of lines are for different effective
ν¯e temperature, which is a key parameter of Fermi-Dirac spectrum. Also shown is the calculation based on the measured SN1987A spectrum. Overall, the DSNB spans from
E 0 to 40 MeV, and the energy-integrated flux per flavor is 10 cm−2 s−1. Moreover, the ν ' ∼ uncertainty in the DSNB spectrum is large. The cosmic core-collapse rate mainly affects the
21 flux normalization, which has less than a factor of 2 uncertainty, while the neutrino emission per core collapse mainly affects the spectral shape, which has a much larger uncertainty.
Therefore, measuring DSNB will provide us with precious information about the neu- trino emission spectrum per collapse, which can not be learned from electromagnetic signals.
The data will be extremely valuable to learn about explosion mechanisms of supernovae and dark collapses, neutrino mixings (in extremely dense environment), and more. On the other hand, we could also learn this from detecting neutrinos from a galactic supernova burst.
However, the burst rate is only a few per century and we can only measure one burst.
On the contrary, DSNB is a steady and guaranteed flux, so we can measure it any time and it probes the average neutrino emission per core collapse, especially including the dark collapses.
Moreover, DSNB also provides precious information that a galactic supernova burst can- not, for example, the cosmic rate of core collapses, dark collapses, and star formation [124].
We could also study particle physics using DSNB, such as electric dipole, magnetic moment, new light dark-sector particles, and many kinds of BSM physics in the neutrino sector [125–127]. Moreover, if the guaranteed DSNB flux is not found, there must be surprising new physics or astrophysics [120].
1.4.2 Review of prior work
To detect DSNB, the most promising detector is Super-Kamiokande (Super-K) [128]. Super-
K is a water-Cherenkov detector. It contains 50 ktons of ultrapure water in a cylinder.
The inner detector has 32 ktons water and the fiducial volume has 22.5 kton. Super-K has 104 photomultiplier tubes which can detect Cherenkov radiation from relativistic ' charged particles. Super-K cannot distinguish charge signs. (See Chapter5 for more about
Super-K.)
Super-K can detect DSNB mainly through the inverse-beta decay process,
ν¯ + p n + e+ , (1.6) e → due to its large cross section [129, 130]. Interactions with nuclei are suppressed by binding
22 effects and interactions with electrons are suppressed by their small mass. The total cross section is σ(E ) 10−43 cm2 (E 1.3 MeV)2, with the outgoing electron (we use electron ν ' ν¯e − to mean both electrons and positrons in this section, unless we specify otherwise) carrying
E E 1.3 MeV and emitted near-isotropically. Super-K has a trigger threshold of e ' ν¯e − 5 MeV, so the relativistic positron is easy to trigger the detector. The neutron is hard to ∼ detect [131], because the neutrons are neutral and mostly captured by hydrogen, producing an only 2.2 MeV gamma ray [132]. As predicted, Super-K has collected 100 DSNB events ∼ after more than 20 years of running. However, these events are hidden in huge backgrounds, so that Super-K has not identified them yet. (In principle, anything that produces electrons or positrons in DSNB energy range could be a background for DSNB.)
Fig. 1.11 summarizes the current status of searching for DSNB in Super-K [45, 131,
133]. The search window is for above 16 MeV electron energy, Ee, due to the backgrounds from reactor neutrinos, spallation events, and NC elastic events (from atmospheric neutrino interactions) which are hard to reduce. Above 16 MeV, the situation is better but there are still large backgrounds caused by atmospheric neutrinos. The dominant component
(dashed green line) is due to atmospheric νµ andν ¯µ, which interact with Super-K water mainly through,
ν (¯ν ) + H/O X + µ− µ+ , (1.7) µ µ → where X represents any final state nuclei or hadrons. If the muon is below 55 MeV kinetic energy, it is invisible in Super-K because it will not produce Cherenkov radiation. Then the muon decays to a visible relativistic electron that mimics a DSNB event. (This corresponds to the “invisible µ” background in Fig. 1.10.) Another component comes from atmospheric
νe andν ¯e (dashed red line) interacting through
ν (¯ν ) + H/O X + e− e+ , (1.8) e e → and those electrons are backgrounds for DSNB events. Therefore, in the figure, we can see that those backgrounds leave only a small window for DSNB searches.
The Super-K collaboration has been working hard on searching for DSNB, by separat-
23 NC elastic
] 1.2
1 & V
e Spallation 1.0 M
1 Reactor e r y
1 0.8 ) n o t
k 0.6 ( ) 5 . 2 2 (
[ 0.4 E d / DSNB flux N 0.2 d e( e)
0.0 0 10 16 20 30 40 Ee [ MeV ]
Figure 1.11: Current status of DSNB searches in Super-K. Solid black lines are the DSNB event rate calculated by Ref. [29] based on an effectiveν ¯e temperature of 6 MeV. Two curves are plotted to include the uncertainties mainly due to the cosmic rate of core collapses. Shaded areas represent the backgrounds as labeled, details below. Colored lines represent atmospheric backgrounds in the current Super-K search window (> 16 MeV). Green lines are for the νµ andν ¯µ components, while red lines are for the νe andν ¯e components. Dashed lines are for current Super-K, while solid lines are for the Super-K-Gd phase. Figure is original to the thesis.
ing them from backgrounds [45, 131, 133]. Though DSNB events are not identified yet,
tremendous progress has been made. The first search was published in 2003 [133], in which
they used 1496 days of Super-K-I data with a search window of Ee > 18 MeV. Their upper
−2 −1 limit on DSNBν ¯e flux, 1.2 cm s for Eν > 19.3 MeV, was already in tension with DSNB model with an 8 MeV effectiveν ¯e spectrum [29]. The second and much more sophisticated search was published in 2012 [45], in which 2853 days of data from Super-K-I to Super-K-III
24 were used. Moreover, the searching window was pushed down to 16 MeV, signal efficiency improved especially at lower energies, and much more sophisticated cuts to the backgrounds
−2 −1 were performed. Those gave an upper limit on DSNBν ¯e flux to be 2.8–3.1 cm s for
Eν > 17.3 MeV. In 2008, Super-K was upgraded from Super-K-III to Super-K-IV, which could detect the coincident neutrons from the inverse-beta decay (Eq. (1.6)). In 2015, a search using only Super-K-IV data (960 days) was published [131]. In the search window of 12 MeV < Ee < 31.3 MeV, the neutron detection efficiency is only about 17.7%. The upper limit was higher than that of Ref. [45], due to less data, high backgrounds at lower energies, low neutron detection efficiency, etc.
Right now, Super-K is being upgraded to Super-K-Gd with the addition of dissolved gadolinium as 0.2% (by mass) gadolinium sulfate, Gd2(SO4)3 [57–59]. Adding gadolin- ium significantly improves Super-K’s ability to detect neutrons. Due to the huge neutron- capture cross section of gadolinium, in the Super-K-Gd phase, more than 90% of the neu- trons are captured on gadolinium, and the energy release is about 8 MeV in a few gamma rays [132], which are much easier to detect. This neutron-tagging technique significantly improves DSNB detectability [57]. This is because each DSNB event has exactly one neu- tron (Eq. (1.6)), but for atmospheric neutrino backgrounds, more than half of them have zero or multiple neutrons, and those can easily be reduced. This is shown as the solid green and dashed lines in Fig. 1.11. This enlarges the DSNB detection window, though still small.
1.4.3 My work
In this section, I give an overview of the motivations and key ideas for my work on DSNB detection.
According to above, the key next step for DSNB is to further enlarge the detection window of DSNB, which is helpful to both the first detection and the subsequent precision measurement of DSNB. Therefore, we need to further reduce the backgrounds. One way forward is to further reduce the spallation backgrounds between about 10 and 16 MeV [134–
136]. Another way forward is to further reduce the atmospheric neutrino backgrounds at higher energies (Fig. 1.11), which is my work. On the other hand, the atmospheric neutrino
25 events are interesting as a signal, which provides important tests of neutrino-interaction modeling and neutrino mixings [137].
Chapter5 presents my thorough study of the underlying physics of atmospheric neu- trino backgrounds for DSNB detection. This is crucial to both seeking ideas to reduce those backgrounds and using them as a signal (e.g., Ref. [137]). In this study, I devel- oped a detailed framework to calculate those backgrounds (both the νµ/ν¯µ component and the νe/ν¯e component) in Super-K for the first time. The framework combined all the un- derlying physics pieces, including atmospheric neutrino fluxes, neutrino mixings, neutrino interactions (mainly using GENIE [138–140], plus our own corrections for Coulomb dis- tortion), propagation of secondaries in water (using FLUKA [141, 142]), interpretations of
Super-K analysis cuts, and detector physics. Using the framework, I successfully repro- duced those atmospheric neutrino backgrounds. From this process, I have found several ideas that will significantly help reduce the backgrounds further. The most promising one stems from the neutrons from atmospheric neutrinos backgrounds (tens MeV) being usu- ally much more energetic than those from DSNB interactions (. 2 MeV). Neutrons with higher energies propagate further in water than those with lower energies, and the distances can be measured in the Super-K-Gd phase, so that the atmospheric neutrino backgrounds can be reduced further. Our study in Chapter5 also sets a solid theoretical foundation for reducing atmospheric neutrino backgrounds for other searches in Super-K like proton decay [143–145], neutron-antineutron oscillation [146], dark matter [147–149], and more.
26 Chapter 2 TeV Solar Gamma Rays From Cosmic-Ray Interactions
The Sun is a bright source of GeV gamma rays, due to cosmic rays interacting with solar matter and photons. Key aspects of the underlying processes remain mysterious. The emission in the TeV range, for which there are neither observa- tional nor theoretical studies, could provide crucial clues. The new experiments HAWC (running) and LHAASO (planned) can look at the Sun with unprece- dented sensitivity. In this paper, we predict the very high-energy (up to 1000 TeV) gamma-ray flux from the solar disk and halo, due to cosmic-ray hadrons and electrons (e+ + e−), respectively. We neglect solar magnetic effects, which is valid at TeV energies; at lower energies, this gives a theoretical lower bound on the disk flux and a theoretical upper bound on the halo flux. We show that the solar-halo gamma-ray flux allows the first test of the 5–70 TeV cosmic- ∼ ray electron spectrum. Further, we show that HAWC can immediately make an even stronger test with nondirectional observations of cosmic-ray electrons. Together, these gamma-ray and electron studies will provide new insights about the local density of cosmic rays and their interactions with the Sun and its mag- netic environment. These studies will also be an important input to tests of new physics, including dark matter.
The contents of this chapter were published in [7].
2.1 Introduction
The Sun is a passive detector for cosmic rays in the inner Solar System, where direct measurements are limited. It shines in gamma rays from its disk and from a diffuse halo [16, 17, 30, 102–104]. Disk emission is expected due to cosmic-ray hadrons inter- acting with solar matter, which produces pions and other secondaries of which the decays
27 and interactions lead to gamma rays. Halo emission is expected due to cosmic-ray electrons
(e+ + e−) interacting with solar photons via inverse-Compton scattering. There are no other important astrophysical mechanisms for steady solar gamma-ray production; solar-
flare gamma rays are episodic, and are observed up to only a few GeV [89, 150–152].
Gamma-ray observations thus open the possibility of detailed cosmic-ray measurements near the Sun. The hadronic and leptonic components can be distinguished because the disk and halo emission can be separated by direction. Further, the energy spectra of the cosmic rays can be inferred from the gamma-ray spectra, which can be measured over a wide energy range. This would give a significant advance compared to typical satellite detectors in the inner Solar System, which only measure the energy-integrated all-particle flux (e.g.,
Refs [153, 154]), and are thus dominated by low-energy particles. Further, gamma-ray data can trace the full solar cycle, testing how solar modulation of cosmic rays depends on energy and position [155, 156].
Figure 2.1 shows that the prospects for measuring TeV solar gamma rays are promising.
The solar-disk fluxes measured in the GeV range with Fermi data [16, 30] are high, and naive extrapolation suggests that HAWC and LHAASO may detect gamma rays in the
TeV range. Further, the GeV observations are significantly higher than the theoretical prediction of Seckel et al. [17], who proposed a compelling mechanism by which the solar- disk gamma-ray flux could be enhanced by magnetic effects. Evidently, even this expected enhancement is not enough, which increases the need for new observations to reveal the underlying physical processes. Even if HAWC and LHAASO only set limits on the TeV gamma-ray flux, that would be important.
Our goal here is to provide a theoretical foundation to quantitatively assess the TeV detection prospects. At low energies, cosmic rays are affected by magnetic modulation in the inner Solar System, as well as by magnetic fields in the solar atmosphere, all of which are complicated [17]. At high energies, where magnetic effects can be neglected, the calculations are relatively straightforward but have not been done before this paper.
The energy separating the two regimes is not known. We estimate that neglecting magnetic effects is appropriate for TeV–PeV gamma rays and show that it leads to useful benchmarks
28 10-7 Fermi2011 ]
1 Ng2016 − s
2 -8 − 10 m c V e G
[ SSG1991 prediction γ HAWC 1 yr
E -9 d 10 / F d 2 γ E LHAASO 1 yr
10-10 -1 1 4 5 10 100 10 102 103 10 10
Eγ [ GeV ]
Figure 2.1: Prospects for TeV solar gamma-ray observations, illustrated with the disk emission (details in Fig. 2.5). Points: observations with Fermi [16, 30], where the flux difference is due to time variation. Green band: the only theoretical prediction that includes magnetic effects [17]. Dashed lines: the estimated differential point-source sensitivity of HAWC [31] (scaled to one year) and LHAASO [18, 19].
for GeV–TeV gamma rays. In future work, we will treat magnetic effects in detail. For a broader context on our program of work on understanding the gamma-ray emission of the
Sun — aimed toward eventual new measurements of cosmic rays, among other goals — see
Ref. [30].
We now provide more information about gamma-ray observations and prospects. Over
29 the 0.1 GeV–TeV range, the Sun has been well observed. Following the upper limits given by EGRET [157] and the first detection using EGRET archival data [102], more detailed measurements were reported in Ref. [16] by the Fermi Collaboration, based on 1.5 years of data. Over the range 0.1–10 GeV, they separately measured the disk and halo fluxes,
finding spectra E−2, plus a hint of time variation in the disk flux. In Ref. [30], where ∼ we used six years of Fermi data and a newer version of the data processing (Pass 7 vs Pass
6), we detected the disk flux up to 100 GeV, finding that its spectrum falls more steeply than E−2. We also made the first robust detection of time variation, showing that the disk flux decreased by a factor of 2.5 from solar minimum to maximum. While the solar- halo gamma-ray flux is reasonably well understood, our results deepen the mysteries of the solar-disk gamma-ray flux. New observations are needed, especially at higher energies, which will critically test emissions models. However, this is difficult with Fermi due to the low gamma-ray flux.
In the TeV–PeV range, the only ground-based gamma-ray experiments that can observe the Sun are those that directly detect shower particles. (For air-Cherenkov detectors, based on detecting optical photons, the Sun is too bright.) The HAWC experiment began full operations in 2015, and is now reporting first results. The LHAASO experiment, under construction, is expected to begin operations in 2020. These experiments will greatly im- prove upon the energy range and flux sensitivity of their predecessors, e.g., Milagro [158],
ARGO-YBJ [159], and Tibet AS-gamma [160]. Those and other experiments have observed the “Sun shadow,” a deficit of shower particles caused by the solar disk blocking cosmic-ray hadrons [161, 162], but none have detected an excess gamma-ray emission from the Sun.
The shadow is displaced by 1◦ from the Sun’s position due to magnetic deflections of ∼ cosmic rays en route to Earth, but the gamma-ray excess will be centered on the Sun.
HAWC and LHAASO observations in the TeV range, combined with Fermi observations in the GeV range, will provide a long lever arm to test models of solar gamma-ray emission.
This paper makes steps toward a comprehensive understanding of solar gamma rays. In
Sec. 2.2, we discuss the effects of magnetic fields and justify why we can neglect them here.
The next three sections are ordered by the directionality of the signals. In Sec. 2.3, we detail
30 our calculation of the hadronic gamma-ray emission from the limb of the solar disk. This calculation has not been done before. We also estimate the flux of other secondary products
(electrons, positrons, and neutrons), discussing if they are significant background for the gamma rays. In Sec. 2.4, we detail our calculation of the leptonic gamma-ray emission from the solar halo. We extend earlier calculations to higher energies and are the first to include allowed new contributions to the electron spectrum. In Sec. 2.5, we discuss the all-sky signal of directly detected cosmic-ray electrons. Our points about these prospects are new and exciting. In Sec. 2.6, we present our conclusions and the outlook for further work.
2.2 Interplanetary and Solar Magnetic Fields
The flux of cosmic rays near the Sun is altered by magnetic effects. Throughout the So- lar System, there are magnetic disturbances sourced by the Sun and carried by the solar wind [163, 164]. These form an interplanetary magnetic field (IMF) that repels galactic cosmic rays (“solar modulation”) [155, 156]; the effects and their uncertainties increase at low energies and at small distances from the Sun. In addition, near the Sun, within approx- imately 0.1 AU, there are solar magnetic fields (SMF) that are quite strong, especially ∼ in the photosphere and corona [165]. Because the SMF are complex and not completely measured, their effects may be varied and are quite uncertain.
In this paper, we focus on gamma-ray signals in the energy regime where magnetic effects can be neglected. When this is appropriate for the solar-disk signal, it will be even more so for the solar-halo signal, for which cosmic rays interact farther from the Sun. We begin by discussing SMF effects on the solar-disk signal, as these turn out to be dominant over IMF effects.
SMF effects enhance gamma-ray production from the solar disk. A likely physical mech- anism was proposed in Ref. [17], although the authors’ predictions still fall far below obser- vations [16, 30]. The enhancement is due to the mirror effect of solar magnetic flux tubes on charged hadronic cosmic rays, which can reverse the directions of cosmic rays before they interact, thus producing outgoing gamma rays that are not absorbed by the Sun. At
31 high enough energies, this mirroring becomes ineffective, and the enhancement ends. To estimate the critical energy Ec for this transition, where magnetic-field effects on cosmic rays can be neglected, we compute the Larmor radius, L, using the typical SMF strength
10 near the Sun, B 1 G, and the solar radius, R 7 10 cm [163, 164], finding ∼ ' × 4 L B Ec 10 GeV . (2.1) ∼ R 1 G
A similar value is obtained for a single flux tube, for which the magnetic field strength can be
103 times larger and the distance scale 103 times smaller [163, 164]. (Ref. [17] estimated ∼ ∼ E to be between 3 102 GeV and 2 104 GeV, so our choice is conservative.) Because c ' × ' × E 0.1E for typical hadronic interactions, SMF effects should therefore be negligible γ ∼ p for gamma-ray energies above about 1 TeV. However, SMF models are uncertain, and it is important to test them with new data.
IMF effects reduce gamma-ray fluxes. Near Earth, IMF effects on the cosmic-ray spec- trum are well described by the widely used force-field approximation [16, 166, 167] and detailed simulations [168, 169], which are informed by extensive measurements. For cosmic rays in the inner Solar System, both modeling and data are sparse. A key clue is that the
MESSENGER probe to Mercury found only . 10% modulation of the cosmic-ray spectrum above 0.125 GeV near solar distances around 0.4 AU [153, 154]. Using a force-field model with appropriate parameters to be consistent with these data (potentials . 400 MV), we find that IMF effects can be neglected for cosmic rays with energies above 100 GeV (and thus gamma rays above 10 GeV), even near the solar surface. However, IMF models are also uncertain, heightening the need for new data.
At energies where magnetic effects can be neglected, the solar-disk signal should thus be wholly due to the limb contribution. This emission is caused by cosmic rays that graze the Sun, encountering a column density that is large enough for them to interact but small enough for their gamma rays to escape. Because this signal can be calculated with minimal uncertainty, a gamma-ray measurement consistent with its flux prediction would confirm that magnetic effects are negligible. In principle, this could also be checked by the angular distribution of the signal, where the Sun would appear as a bright ring with
32 a dark center, although planned TeV–PeV experiments may not have adequate angular resolution [18, 19, 31, 170]. Finally, tests could also be made by the time variation, as there should be none.
At lower energies, where magnetic effects are important, several distinctive signatures of the solar-disk signal should emerge. The flux should be larger, as SMF effects that enhance the gamma-ray flux dominate over IMF effects that decrease it [17]. That is, our solar-disk prediction neglecting magnetic effects provides a theoretical lower bound on
the disk flux, which is especially interesting at GeV–TeV energies. The angular variation
of the signal should tend toward illumination of the full disk. And there should be time
variations that reveal the nature of the dominant magnetic effects. IMF effects decrease
gamma-ray production near solar maximum, due to cosmic-ray modulation [16, 30]. Perhaps
surprisingly, SMF effects must act in the same sense, as the IMF effects are too small to
explain the observed time variation [30].
For the solar-halo signal, IMF effects dominate over SMF effects [16], except perhaps
very near the Sun. The comparison of disk and halo signals will thus help disentangle IMF
and SMF effects. It also means that neglecting magnetic effects provides a theoretical upper
bound on the halo flux.
2.3 Hadronic Gamma Rays
2.3.1 Calculational framework
In the direction of the solar disk, the dominant source of gamma rays is the interactions of
hadronic cosmic rays with matter in the solar atmosphere [16, 17, 30]. Of these interactions,
the most important are inelastic proton-proton collisions that produce neutral pions, which
promptly decay to gamma rays. (In Sec. 2.4, we calculate gamma-ray production by leptonic
cosmic rays, including near the direction of the solar disk, although the interactions occur
well away from the solar surface.)
Here we calculate the gamma-ray emission from the solar limb — the small fraction of
the Sun encountered by cosmic rays that just graze its surface on trajectories toward Earth.
33 We use the straight-line approximation, where gamma rays maintain the direction of their parent hadrons, appropriate because the particle energies are so high. We ignore emission from the disk because we neglect magnetic effects that can reverse the directions of cosmic rays before they interact [17] and because the contributions of back-scattered pions are tiny [171]. As the ingredients of the calculation are reasonably well known, the predicted limb emission is robust and, as noted, sets a theoretical lower bound on the solar-disk flux.
We calculate the total flux from the limb, integrating over its solid angle. Here we assume that it cannot be resolved, as single-shower angular resolution of HAWC and LHAASO near
1 TeV is comparable to the 0.5◦-diameter of the Sun [18, 19, 31, 170]. The solid angle of the limb is tiny, 10−3 of that of the solar disk. If the limb could be resolved, it would appear ∼ as a thin, bright ring, with the intensity (flux per solid angle) enhanced by 103 over the ∼ intensity averaged over the solar disk. The angular resolutions of HAWC and LHAASO
improve at higher energies, which may allow partial resolution of the limb, especially with
stricter cuts to select events with the best angular resolution. In the long term, hardware
upgrades to improve this should be considered.
We begin in Sec. 2.3.2 by discussing gamma-ray production in a simplified case —
proton-proton production of neutral pions in the thin-target limit — which can be handled
semianalytically, following Ref. [172]. Then, in Sec. 2.3.3, we include the effects of multiple
scattering and absorption, cascade processes, and nuclear composition through a simulation
using GEANT4 [105, 106]. In ths simplified case, the flux is
dF Z Z Z dI dN (Eγ) = dΩ ds np(~s) dEp (Ep) σinel(Ep) (Ep,Eγ), (2.2) dEγ dEp dEγ where np is the number density of target protons at the line-of-sight coordinate ~s, dI/dEp is
the cosmic-ray proton intensity, σinel is the inelastic proton-proton scattering cross section, and dN/dEγ is the spectrum of gamma rays per interaction. The length of the chord
1/2 4 through the solar atmosphere is ∆s (8R h ) 2.6 10 km, where the 8 comes from ∼ 0 ∼ × 5 geometry, R 7 10 km is the radius of the Sun, and h 120 km is the scale height ' × 0 ' of the solar matter density in the photosphere. In the realistic case, the most important interactions occur at proton optical depths τ 2, so this simplified case is not adequate for ∼ 34 our full results, although it does introduce the framework well.
2.3.2 Calculation for the simplified case
Figure 2.2 shows the solar mass density ρ from Refs. [173, 174]. Above the photosphere, the density declines exponentially, following the Boltzmann distribution of gravitational potential energy in the nearly isothermal atmosphere. Figure 2.2 also shows the proton optical depth τ as a function of height above the photosphere. The cross section for inelastic proton-proton collisions changes only modestly with energy and is 30 70 mb for proton ' − energies 1 107 GeV [108]. In the optically thin limit, gamma-ray production is dominated − by the decay of neutral pions, which, at the low densities considered here, always decay in
flight before interacting. Kelner, Aharonian, and Bugayov [172] have extensively studied the yields of secondaries in proton-proton collisions, in which their results are based on a
fit to data and to particle-interaction simulations. The yield of gamma rays has a broad energy spectrum, but the most important gamma ray typically has E 0.1E . The shape γ ∼ p of τ(h) closely follows that of ρ(h), due to the exponential dependence, with the conversion
factor σ s/m 5 107 cm3 g−1. ' inel p ' × The cosmic-ray flux can be taken to be that at Earth, as we neglect magnetic effects.
(Technically, the flux at Earth includes some modulation effects, but these are negligible
at such high energies.) Up to 1 TeV, we use the precisely measured proton spectrum from
the Alpha Magnetic Spectrometer (AMS-02) [107]. At higher energies, it is sufficient to
extrapolate this using dI/dE 1 (E/GeV)−2.7 cm−2 s−1 sr−1 GeV−1 [108]. ' Figure 2.3 shows the resulting gamma-ray spectrum for the case where we integrated over 0 < τ < 0.3, up to roughly the largest value for which an optically thin calculation is appropriate (the probability for a proton to interact twice is then . 10%, so the gamma-ray spectrum scales linearly with τ). We checked the results of our semianalytic calculation by a Monte Carlo simulation with the particle-interaction code GEANT4 [105, 106], for which the results matched to within . 10%. This shows that effects beyond those in Ref. [172], e.g., particle cascades in the medium, are unimportant.
Lastly, compared to the Sun, the gamma-ray flux from the limb of the Earth’s atmo-
35 Figure 2.2: Solar mass density as function of height above the photosphere (left axis, blue dashed), as well as the same for optical depth for inelastic proton-proton collisions (right axis, red solid).
sphere has been measured by Fermi up to nearly 1 TeV and compared to simulations,
finding good agreement with predictions, which demonstrates the robustness of theoretical calculations [175, 176]. In principle, in the thin-target limit, the limb flux from the Sun could simply be expressed in terms of the limb flux from Earth, nullifying several potential uncertainties, such as the energy spectrum, composition, and cross section.
36 2.3.3 Calculation for the realistic case
To include proton-proton interactions in the optically thick case, we use GEANT4 [105, 106].
This allows protons to interact several times, and takes into account their particle and energy
losses from all processes. It also includes gamma-ray production by cascade processes, such
as bremsstrahlung by electrons. The number density of target photons is 104 times ∼ smaller than that of solar matter so energy losses and gamma-ray production by inverse-
Compton processes can be neglected [177]. The density is low enough that charged pions
below 1 PeV will typically decay in flight before interacting. Neutrons and muons may
escape, and the neutrons may survive to Earth without decay.
Figure 2.4 shows the range of τ values that contribute most to gamma-ray production,
based on our GEANT4 simulation. The y axis is weighted to properly compare different
logarithmic ranges of τ. The peak is near τ 2, where 90% of cosmic-ray protons will ∼ ∼ interact at least once. To the left of the peak, the linear decline is due to reduced optical
depth. To the right, the exponential decline is due to proton cooling and especially gamma-
ray absorption, which happen to have similar interaction lengths (for pion production and
electron-positron pair production, respectively). In combination, about 90% of the total
flux arises in the range τ 0.1 10. ∼ − Using this range of τ values, we use Fig. 2.2 to determine the corresponding range of
heights above the photosphere and corresponding mass densities, finding h 60 600 km ∼ − and ρ 10−7 10−9 g cm−3. This leads to important insights about the physical conditions ∼ − in which interactions occur. In this range, the solar properties are reasonably well known
and are stable in time. The conditions for the production of solar atmospheric gamma rays
are quite different from those for Earth atmospheric neutrinos [178]; for the latter, ρ 10−4 ∼ g cm−3 at an altitude of 10 km, and the distance scales are short but the proton optical
depth is high (τ 20). Lastly, this information will be useful for assessing interactions in ∼ the presence of magnetic effects, which we will consider in future work.
With GEANT4, we can simulate proton interactions only up to a laboratory energy of
100 TeV, which leads to a gradual cutoff of the gamma-ray spectrum near 10 TeV. To
37 Figure 2.3: Solar-limb gamma-ray spectrum produced by hadronic cosmic rays. Red dotted line: semianalytic result for proton-proton interactions with 0 < τ < 0.3. Green dash-dotted line: GEANT4 results for the full range of τ; the gradual cutoff is because it cannot simulate proton interactions above 100 TeV. Blue dashed line: our empirical fit to the GEANT4 results, extrapolated to higher energies. Black solid line: our full prediction, including a correction factor for nuclei. The light grey shading approximately indicates the energies at which magnetic effects, neglected here, should be included.
extend our results to higher energies, we develop an empirical fit to the GEANT4 results at
lower energies. We modify our semianalytic approach, Eq. (2.2), by including a correction
factor, e−ατ(Ep,~s), that only becomes important in the optically thick regime. For the free
parameter α, we find that 0.65 gives a good match to the GEANT4 results. This is shown in
38 Fig. 2.3.
Figure 2.4: Normalized relative contributions of different τ values to the predicted gamma- ray flux, based on our GEANT4 simulation. We show the example of Eγ = 1 TeV; other energies give similar results.
Finally, we consider the effect of nuclei in the cosmic rays and in the solar atmosphere.
Besides protons, the only important constituent is helium, which has a 10% relative ' number abundance in both the beam and target [108, 173, 174]. We use the cosmic-ray helium data from Ref. [109] up to 1 TeV. Above that, we use a power law and extrapolate 39 up to 10 PeV with spectral index 2.7, which roughly describes the data compilation in ∼ Ref. [108]. Following Ref. [179], we calculate the gamma-ray flux enhancement factor due to cosmic-ray helium. We find that the gamma-ray flux is increased by an overall factor
1.8, with a small energy dependence due to a slightly different spectral shape between the ' proton and helium. We also consider the case in which the helium spectrum may be harder than 2.7 at high energies [180–182]. If we use a spectral index of 2.58 [181] for extrapolation, our result changes by less than 20% near 10 TeV. Thus, we can safely ignore the spectral hardening.
Figure 2.3 shows our full prediction for the gamma-ray spectrum from the solar limb.
The gamma-ray spectrum closely follows the cosmic-ray proton spectrum [172]. This is because the pions and gamma rays typically carry fixed fractions of the parent proton energy, the cross sections and multiplicities for (high-energy) pion production and gamma- ray absorption have only mild energy dependence, and the pions decay before interacting.
(For the same reasons, Earth atmospheric neutrinos at sub-TeV energies also follow the cosmic-ray spectrum [178].) In Fig. 2.3, the gamma-ray flux has normalization 2 10−9 ' × sr times the proton intensity (flux per solid angle). This factor can be roughly reproduced
using ∆Ω τ (0.1)1.7, where ∆Ω 10−7 sr is the relevant solid angle of the limb, τ 1 × × ∼ ∼ is a typical value, and the last factor comes from assuming that each proton produces 1 ∼ gamma-ray at E 0.1E . γ ∼ p The hadronic-interaction processes discussed here also produce neutrinos, electrons (in-
cluding positrons), and neutrons [17]. The neutrino flux [183–186] is an important back-
ground for dark matter searches with neutrino telescopes [187–189], and constitutes a sen-
sitivity floor [60–62]. The other species could be useful messengers to study cosmic-ray
interactions with the Sun, using detectors such as Fermi [190], AMS-02 [191], CALET [192],
and DAMPE [193]. A dedicated study of their detectability, is beyond the scope of this
paper, and will be considered elsewhere. Here we briefly comment on their relevance to
gamma-ray observations.
Electrons can be effectively separated from gamma rays in space-borne detectors. How-
ever, this separation is difficult for ground-based experiments, as both particles induce
40 electromagnetic showers in the atmosphere. In principle, the inclusion of electrons en- hances the detectability of the Sun for ground-based experiments. The flux of the electrons can be estimated similarly to that of gamma rays, described above, also by first ignoring magnetic-field effects. The electron flux is found to 2 times lower than that of the gamma ∼ rays, due to receiving a smaller fraction of the pion energy. Further, the detection of these secondary electrons with ground-based experiments is more complicated than gamma rays, as the effects of solar, interplanetary, and Earth magnetic fields all need to be taken into ac- count, demonstrated by cosmic-ray shadow studies [161, 162]. The deflections and diffusion they cause will reduce the electron flux per solid angle. Therefore, for the current study, we neglect the addition of the electron flux to the total electromagnetic signal observable by ground-based experiments.
Neutrons, the most important secondary hadrons, travel without being affected by the magnetic fields. The Sun is therefore a point source of neutrons, and could in principle be detectable by ground-based experiments. Compared to gamma-ray production in pionic processes, secondary neutrons carry a smaller fraction of the primary energy. However, spal- lation of helium is efficient at producing secondary neutrons. Combining these two factors, the limb neutron flux is comparable to that of the gamma rays (also the disk flux [17]). In practice, it is difficult for these neutrons to be confused with gamma rays by ground-based experiments, due to the excellent hadron rejection factor, 10−3. The detection in the ∼ hadron channel is also likely to be difficult due to the much higher background, compared
to that of gamma rays and electrons. A more careful treatment of hadrons, in particular at
lower energies, is the subject of a separate paper (Zhou et al., in prep.).
2.4 Leptonic Gamma Rays
In directions away from the solar disk, there is a solar halo of gamma-ray emission, of
which the dominant source is the interactions of cosmic-ray electrons (e+ + e−) with solar
photons [16, 102–104]. Of these interactions, the most important is inverse-Compton scat-
tering. There is also a contribution in the direction of the solar disk. We estimate that
41 other interactions with solar photons are irrelevant; these include Bethe-Heitler [194, 195] and photo-pion interactions of protons [196] and deexcitation interactions of nuclei following photodisintegration [197–200].
Here we calculate this leptonic gamma-ray emission, mostly following prior work [102–
104, 201]. For the first time, we calculate results up to 1 PeV and show that uncertainties in the electron spectrum at very high energies allow larger signals than in the nominal case
(a broken power-law spectrum for cosmic-ray electrons). As above, we neglect magnetic effects and assume straight-line propagation for the parent-daughter kinematics. Although the solar halo flux is present in all directions, its intensity (flux per solid angle) is greatest near the Sun, falling approximately as θ−1 [102–104] , where θ is the angle away from the center of the Sun. The flux within a given angle thus grows as θ, but the backgrounds — especially significant for ground-based detectors — grow as θ2. Therefore, the solar-halo signal is most interesting at relatively small angles. We calculate the leptonic signal within
1.5◦ degrees of the solar center; this value matches what we used for our Fermi analysis [30] and will allow HAWC and LHAASO to treat it as a near-point source.
In the optically thin regime, the gamma-ray flux from the inverse-Compton interactions of cosmic-ray electrons is
Z Z Z Z dF dnph dσ dI = dΩ ds dEph (Eph, ~s) dEe (Ee,Eph,Eγ) (Ee, ~s) (2.3) dEγ dEph dEγ dEe
where dnph/dEph(~s) is the number-density spectrum of target photons at the line-of-sight
coordinate ~s, dI/dEe is the cosmic-ray intensity, and dσ/dEγ is the electron-photon differ- ential cross section including Klein-Nishina effects.
The column density of the solar photon field is n R2 /Dθ for small angles θ [102], ∼ ph where nph is the number density of photons at the solar surface and D = 1 AU. For electron energies below about 0.25 TeV, the inverse-Compton cross section is in the Thompson
regime, where the total cross section is constant with energy. At higher energies, it is
in the Klein-Nishina regime, where the total cross section falls with increasing energy.
An electron passing close to the Sun has an optical depth of 10−2 (in the Thompson ∼ regime; less at higher energies), so the optically thin assumption of Eq. (2.3) is appropriate.
42 10-7 Fermi2011 Ng2016 S ] -8 SG1 991 IC Max
1 10 I − − nv s er se 2 C − -9 h om 10 a p C 1 yr m d HAW r to c o n ni 1 c, .5 V lim ◦ e b LHAASO 1 yr -10 o G n 10 ly [ γ E
d -11
/ 10 F d 2 γ -12 E 10 IC Pulsar − 10-13 1 4 5 100 10 102 103 10 10 106
Eγ [ GeV ]
Figure 2.5: Gamma-ray spectrum of the Sun. Points: disk observations with Fermi [16, 30], where the flux difference is due to time variation. Green band: the predicted disk flux [17]. Dotted lines: the estimated differential point-source sensitivity of HAWC [31] (scaled to one year) and LHAASO [18, 19]. Our new prediction of the solar-disk signal due to cosmic-ray hadrons (from the limb) is shown by the green solid line. Our new prediction of the solar- halo signal due to inverse-Compton scattering of cosmic-ray electrons is shown by the black solid line for the nominal case and by the dashed lines for enhanced cases from Fig. 2.6.
To calculate the gamma-ray spectrum, we use the StellarICs code [110, 111], slightly modified to include a parametrization of the electron spectrum at the highest energies.
The solar photons are taken to have a blackbody spectrum with temperature 5780 K and corresponding typical energy of 1 eV. The photon density falls as distance squared far ∼ 43 from the Sun but less quickly near its surface, where it varies as with radial distance r as
[1 (1 R2 /r2)] [102, 103]. The cosmic-ray electron flux has been precisely measured by − − AMS-02 up to almost 1 TeV [32], and measured moderately well by H.E.S.S. [33, 34] and
VERITAS [35] up to 5 TeV. We use a broken power-law fit to these data. As discussed in detail in Sec. 2.5, the electron spectrum at very high energies might be much larger than expected from this nominal case, in which the flux above 5 TeV is assumed to fall off quickly. Our calculation is the first to show how allowed contributions to the electron spectrum above 5 TeV would enhance the solar-halo gamma-ray signal.
2.5 Cosmic-Ray Electrons
Figure 2.5 shows our results for the leptonic gamma-ray emission in the nominal case plus some enhanced cases. (Below 10 GeV, where there are measurements from Fermi [16], not shown here, our prediction is consistent.) In the Thomson regime, the gamma-ray spectrum is less steep than the electron spectrum due to the nature of the differential cross section. In the Klein-Nishina regime, the gamma-ray spectrum steepens sharply due to the suppression of the total cross section (in addition to the steepening electron spectrum). The nominal predictions are not detectable with HAWC and LHAASO. In fact, only the most extreme enhanced scenarios — with the cosmic-ray electron flux as large as the proton flux — are
(lines labeled “Max” in Figs. 2.5 and 2.6). If no solar-halo signals are detected, as is likely, that will make it easier to isolate hadronic gamma-ray flux in the direction of the solar disk.
Section 2.5 introduces a better way to probe cosmic-ray electrons.
Figure 2.5 also recaps our result for the hadronic gamma-ray emission from the solar limb. This is well below the leptonic gamma-ray emission from the solar halo near the disk
(below about 1 TeV), as well as the sensitivity of HAWC and LHAASO. However, this prediction leads to several important points. The gamma rays observed from the solar disk must be hadronic, with their flux enhanced by magnetic effects, and the ratio of the data to our limb prediction provides a first direct measure of the strength of that enhancement.
The hadronic gamma-ray spectrum must eventually bend toward and join with our limb
44 3 Figure 2.6: Diffuse flux (weighted with Ee ) of cosmic-ray electrons. Below about 5 TeV, there are measurements (points, as labeled [32–35]). Above about 70 TeV, there are limits (gray region, which combines many experiments [36, 37]). In between, the spectrum could be as large as the blue solid line, allowing enhanced contributions (pulsar or dark matter; details are in the text). HAWC should be able to immediately improve sensitivity down to 10−3 (hadronic rejection) of the proton spectrum (red dashed line). ∼ prediction. Until the energy at which that occurs, there is positive evidence for interesting processes (magnetic effects) beyond the limb emission. It may be that the leptonic gamma- ray emission is never dominant in the data, despite its apparent dominance in Fig. 2.5.
Here we show that HAWC and LHAASO can directly measure the cosmic-ray electron
45 (e+ + e−) spectrum, which is of great interest [33–37, 190–193, 202–206]. Compared to the method of Sec. 2.4, this is simpler and more powerful. The flux is expected to be isotropic. If a nearby pulsar or dark matter halo contributes significantly, the resulting anisotropy would enhance the detection prospects, but we neglect this possibility. Because cosmic-ray electrons lose energy quickly, by synchrotron and inverse-Compton processes, the highest-energy electrons must come from quite nearby, e.g., a few hundred pc at 10
TeV.
Figure 2.6 summarizes present knowledge of the cosmic-ray electron spectrum. Below 5
TeV, there are measurements from various detectors, including AMS-02 [32], H.E.S.S. [33,
34], and VERITAS [35]. Above 70 TeV, there are strong limits from ground-based arrays
(summarized in Refs. [36, 37]). Importantly, at 5–70 TeV, there have been no experimental probes, as emphasized in Ref. [36]. At those energies, the only limit, which is quite weak, comes from requiring that the electron flux not exceed the all-particle flux. New sensitivity is needed to probe the electron spectrum in this energy range, where new components could appear. Intriguingly, there are hints of a new component starting to emerge at 5 TeV, seen by both the southern-sky H.E.S.S. [34] and the northern-sky VERITAS [35].
HAWC and LHAASO detect electrons and gamma rays with comparable efficiency [207].
However, the flux sensitivity for electrons is worse because, like the background protons, they are isotropic. The sensitivity depends on just the hadronic rejection factor. (Gamma rays are not a background, except in the direction of point sources; the diffuse flux of TeV electrons, even in the nominal case, exceeds that of gamma rays, even in the direction of the Milky Way plane [208].) We assume a hadronic rejection factor of 103, which should ∼ be reachable (Segev BenZvi, private communication). Performance close to this has been demonstrated by some analyses with a partially complete HAWC detector [209, 210]. More importantly, HAWC has already shown preliminary limits that approach our estimated sensitivity [211].
Figure 2.6 shows the estimated HAWC sensitivity to the electron flux (LHAASO’s will likely be similar), along with possible enhancements to the 5–70 TeV electron spectrum.
HAWC and LHAASO can reach higher energies than air-Cherenkov detectors because of
46 their huge advantages in field of view and uptime.
Probing the 5–70 TeV cosmic-ray electron spectrum for the first time will allow inter- esting tests of pulsars, dark matter, and possible surprises. For pulsars, we use predictions from Refs. [204, 205], which may explain the positron excess [191, 203]. (Even larger fluxes can be found in Ref. [212].) For dark matter, we use the PPPC4DMID code [213, 214] to calculate the electron spectra from dark matter decay, in this case with a mass of 100 TeV and a lifetime of 2 1026 s, which is comparable to current constraints [215–217]. × While simple, our results are important. Although the gap in coverage of the cosmic- ray electron spectrum was known [36], as was the possibility of using HAWC to detect electrons [207], this paper is the first to combine those points and quantify the prospects.
In the near future, there will be good sensitivity to high-energy cosmic-ray electrons from the CALET [192], DAMPE [193] and CTA [218] experiments. Even so, they may only reach . 20 TeV. With more than a year of data already collected, HAWC has a unique opportunity now, and we encourage swift action to complete an analysis.
2.6 Conclusions and Outlook
The Sun’s high-energy gamma-ray emission — seemingly due to irradiation by cosmic rays
— is not well understood. Above 10 GeV, the Sun is one of the 20 brightest sources ∼ detected by Fermi, and its disk emission is nearly an order of magnitude brighter [16, 30] than predicted [17]. In the TeV range, there have been no theoretical or observational studies.
Now there is a convergence of two opportunities: the recognition that the high-energy
Sun can reveal important physics and the unprecedented sensitivity of the already running
HAWC experiment. These opportunities will be enhanced by ongoing theoretical work and the sensitivity gain due to the coming LHAASO experiment.
This paper has three main results. The first calculation of the gamma-ray emission due to hadronic cosmic rays interacting with the solar limb. At high enough energies (& 1 TeV), magnetic effects can be neglected,
47 and the complete emission from the solar disk should be from only the thin ring of the limb. This flux can be robustly calculated. Further, it serves as an important theoretical lower bound on the solar-disk emission at all energies. The enhancement of the disk flux by magnetic fields can be deduced by the ratio of the observed flux to this prediction. In the
GeV range, this is a factor 10. As illustrated in Fig. 2.1, HAWC and LHAASO will provide ∼ new sensitivity to solar gamma rays in the TeV range, and can test if this enhancement continues, plus if there are new contributions, e.g., due to dark matter. (Limits from ARGO-
YBJ [159] are already in preparation [219], and is about one order of magnitude weaker than HAWC sensitivity at TeV energies.) Finally, the limb flux would be significantly more detectable if the solar disk could be resolved, due to lower backgrounds per solid angle.
Although we have conservatively neglected this possibility, it seems attainable. New results on the gamma-ray emission due to cosmic-ray electrons interacting with solar photons. This emission forms a gamma-ray halo around the Sun, and the intensity peaks near the disk. For the first time, we calculate the TeV–PeV gamma-ray flux, including the possibility of new components in the 5–70 TeV electron spectrum. HAWC and LHAASO can at least set constraints at these energies, where there are no measurements. A new perspective on allowed enhancements to the cosmic-ray electron spectrum and direct tests of such. Lastly, we show that direct observations of electromagnetic showers by
HAWC and LHAASO can provide unprecedented sensitivity to the 5–70 TeV cosmic-ray electron spectrum. This search, based on nondirectional signals, will be a powerful probe of the high-energy electron spectrum, testing some realistic models.
This paper is part of a larger program of work to develop the Sun as a new high-energy laboratory (see Ref. [30] for further discussion). With a good theoretical understanding of magnetic effects, the Sun could be used as a passive detector for cosmic rays in the inner
Solar System, allowing measurements that are differential in particle type and energy, a capability unmatched by any existing or planned detector. Currently, the major roadblock to this goal is taking into account the complicated magnetic field effects, but this problem is tractable in principle, and progress is being made (Zhou et al., in preparation). The Sun is already a calibration source for direction, and could become one for flux. Interestingly, un-
48 like any other astrophysical source, the Sun’s hadronic and leptonic emission can be clearly separated using angular information alone. Finally, a thorough understanding of cosmic-ray interactions with the Sun is crucial for testing dark matter and neutrino physics [15, 95].
Acknowledgments We thank Andrea Albert, Mauricio Bustamante, Rebecca Leane, Shirley Li, Shoko
Miyake, Carsten Rott, Qingwen Tang, and especially Segev BenZvi, Igor Moskalenko, Elena
Orlando, and Andrew Strong for helpful discussions. BZ was supported by Ohio State
University’s Fowler and University Fellowships. KCYN was supported by NASA Grant No.
NNX13AP49G, Ohio State’s Presidential Fellowship, and NSF Grant No. PHY-1404311.
JFB was supported by NSF Grant No. PHY-1404311. AHGP was supported by NASA
Grant No. NNX13AP49G.
49 Chapter 3 Neutrino-nucleus cross sections for W -boson and trident production
The physics of neutrino-nucleus cross sections is a critical probe of the Standard Model and beyond. A precise understanding is also needed to accurately deduce astrophysical neutrino spectra. At energies above 5 GeV, the cross section ∼ is dominated by deep inelastic scattering, mediated by weak bosons. In addi- tion, there are subdominant processes where the hadronic coupling is through virtual photons, γ∗: (on-shell) W -boson production (e.g., where the underly- ing interaction is ν + γ∗ `− + W +) and trident production (e.g., where it is ` → ν+γ∗ ν+`− +`+). These processes become increasingly relevant at TeV–PeV → 1 2 energies. We undertake the first systematic approach to these processes (and those with hadronic couplings through virtual W and Z bosons), treating them together, avoiding common approximations, considering all neutrino flavors and final states, and covering the energy range 10 –108 GeV. In particular, we present the first complete calculation of W -boson production and the first calculation of trident production at TeV–PeV energies. When we use the same assumptions as in prior work, we recover all of their major results. In a companion paper [1], we show that these processes should be taken into account for IceCube-Gen2.
The contents of this chapter were published in [2].
3.1 Introduction
The interactions of neutrinos with quarks, nucleons, and nuclei are a cornerstone of the
Standard Model. These test neutrino couplings to hadrons and probe the internal structure of hadronic states [138, 220–225]. Increasingly precise measurements of cross sections allow increasingly precise tests of neutrino mixing and beyond the Standard Model physics [78, 88, 50 112, 226–229]. Understanding the cross section is also crucial to neutrino astrophysics [24,
230–239]. In the laboratory, neutrino scattering has been well measured up to E 102 ν ∼ GeV [240–242]. Above 5 GeV, the dominant interaction is deep inelastic scattering (DIS), ∼ where neutrinos couple via weak bosons to the quark degrees of freedom, with the nucleon
and nuclear structure being less important but still relevant.
New scientific opportunities have arisen with IceCube, as atmospheric and astrophysical
neutrinos have been detected up to E 107 GeV [54, 243]. Even though the spectra ν ∼ are not known a priori, and the statistics are low, important progress can be made. For example, the neutrino cross section can be determined by comparing the event spectra due to neutrinos that have propagated through substantial Earth matter or not [70, 71, 237, 244–
246]. And to the extent that the cross section is understood — e.g., the claimed theoretical precision (from the parton-distribution functions) at 107 GeV is 2% [24] or 1.5% [237] ' ' — the measured event spectra can be used to accurately deduce neutrino spectra and
flavor ratios, allowing tests of both astrophysical emission models and neutrino properties
(e.g., Refs. [85–87, 247–257]). As IceCube accumulates statistics, and larger detectors are under consideration [115, 258], the opportunities — and the need for a better theoretical understanding of neutrino-nucleus scattering — increase.
There are neutrino-nucleus interactions in which the hadronic coupling is via a virtual photon, γ∗, and the diagrams are more complex than in ordinary DIS. Although these photon interactions are subdominant, their importance grows rapidly with energy, becoming relevant in the TeV–PeV range. In (on-shell) W -boson production (Fig. 3.1), the neutrino interacts with a virtual photon from the nucleus to produce a W boson and a charged lepton. The cross section for this process has been claimed to reach 10% of the DIS ∼ cross section at 105–107 GeV [38–40]. (To set a scale, in the past 7.5 years, IceCube ∼ has detected 60 starting events with reconstructed energies above 60 TeV [259, 260].) More careful calculations are needed. In trident production (Fig. 3.2), the neutrino interacts with a virtual photon from the nucleus to produce a neutrino, a charged lepton, and a charged lepton of opposite sign [41–43, 118, 119, 226]. The cross section for this process has never been calculated at TeV–PeV energies. A first calculation is needed.
51 p − + l k1 1 l l W + − W p + l − 2 W l q PP’ A A’ AA’
Figure 3.1: Diagrams for (on-shell) W -boson production via photon exchange. A and A0 are the initial- and final-state nuclei. (See Fig. 3.5 and Sec. 3.3.2 for the connection with trident production.) For antineutrinos, take the CP transformation of the elementary particles.
In this paper, we provide the first full calculations of both processes. We treat them
in a unified way, avoiding common approximations, considering all neutrino flavors and
final states, and covering the energy range 10–108 GeV. We recover all previous major
results when we adopt their inputs. In our companion paper [1], we detail the implications
for IceCube-Gen2 measurements of neutrino spectra and flavor ratios, tests of neutrino
properties, and tests of new physics.
In Sec. 3.2, we review the W -boson and trident production processes, identifying the
shortcomings of previous work. In Sec. 3.3, we calculate the neutrino-real photon cross
sections for both processes. The more complicated neutrino-nucleus cross sections in differ-
ent regimes are calculated in Secs. 3.4 and 3.5, then added up (Fig. 3.12) and discussed in
Sec. 3.6. We conclude in Sec. 4.4.
3.2 Review of W -boson and trident production
The (on-shell) W -boson and trident production processes are, respectively,
ν + A `− + W + + A0 , (3.1) ` →
52 ν + A ν + `− + `+ + A0 , (3.2) → 1 2 where A and A0 are the initial and final-state nuclei and ` is a charged lepton. For trident production, for now we simplify the flavor information (for details, see Eq. (3.5)). For antineutrinos, take the CP transformation of the elementary particles.
Figures 3.1 and 3.2 show the diagrams for W -boson and trident production processes, respectively. We also calculate diagrams, not shown, with W and Z boson couplings to the hadronic side; this is discussed in Sec. 3.5. For trident production, (for Fig. 3.2 only) we use the four-Fermi theory for simplicity, the diagrams of which nicely show the “trident” feature though hiding the connection to W -boson production (see Fig. 3.5 for the full Standard-
Model diagrams, on which our calculation is based). In both processes, a neutrino splits into charged particles (leptonic part) that couple to the photon from the nucleus (hadronic part). The leptonic part is straightforward but depends on the process, while the hadronic part is complicated but independent of the process.
In the rest of this section, we review the hadronic part (Sec. 3.2.1), which also sets the foundation, then discuss the two processes respectively (Secs. 3.2.2 and 3.2.3).
3.2.1 Hadronic part
At most energies, the hadronic part is connected by a virtual photon from the nucleus.
8 Above & 10 GeV, the contributions of virtual weak bosons from the nucleus and mixing with the photon are not negligible (see Sec. 3.5).
The hadronic coupling can be in different regimes, including coherent (σ Z2), diffrac- ∝ tive (σ Z), and inelastic (σ Z), in which the virtual photon couples to the whole ∝ ∝ nucleus, nucleon, and a single quark, respectively. (These three regimes are analogous to
the coherent elastic neutrino-nucleus scattering, quasi-elastic scattering, and deep-inelastic
scattering, respectively, for the usual neutrino-nucleus interaction, in which the hadronic
coupling is through a W/Z boson.) Adding the cross sections in different regimes gives the
total cross section.
53 l l
+ − l l l l − l l+
A A’ A A’
Figure 3.2: Diagrams for trident production via photon exchange in the four-Fermi theory (see Fig. 3.5 for the full Standard Model). For antineutrinos, take the CP transformation of the elementary particles.
The coherent (A0 = A) and diffractive (A0 = A) regimes are both elastic, on the nucleus 6 and nucleon, respectively. The former is usually described by a nuclear form factor and the latter by a nucleon form factor.
Two different calculational frameworks have been used in previous work, i.e., using or not using the equivalent photon approximation (EPA, or Weizs¨acker–Williams Approxima- tion) [261–263]. The EPA assumes the photon mediator (to the nucleus) to be on shell, i.e., q2 = 0. This is motivated by the fact that the photon is usually very soft when the beam particle is very energetic (e.g., high-energy electron scattering on nuclei). Using real photons significantly simplifies the calculation, because then one does not need to take into consideration the photon virtuality or longitudinal polarization. However, Refs. [42, 264–
266] pointed out that EPA is not valid for most cases, especially for electron final states, leading to an overestimation of the cross section that, in some cases, is by more than 200%. The simplest reason is that, though the beam neutrino is very energetic, the charged particle that directly couples to the photon may not be.
Inelastic scattering (A0 = A) could also happen, with nucleon breakup. The hadronic 6 part is usually described by parton-distribution functions (PDFs) for photon, quarks, etc.
The inelastic regime has two subprocesses, photon-initiated (related to the photon PDF) and quark-initiated (related to the quark PDFs) [267]. See Sec. 3.5 for details.
54 3.2.2 W -boson production
The W -boson production process (Fig. 3.1) initially raised interest in the 1960s and 1970s.
The hypothetical (at that time) W boson could be directly produced by a beam of νµ scattering off the Coulomb field of a nucleus (Fig. 3.1; e.g., Refs. [268–273]). If W bosons
were not detected, a lower bound on their mass could be set. Later, the discovery of the
W boson at a proton-antiproton collider [274], and especially its large mass, significantly
reduced the motivation to search for this process at fixed-target neutrino experiments.
The interest in this process came back due to high-energy astrophysical neutrino de-
tectors [54, 258, 275], and was studied by Seckel [38] and Alikhanov [39, 40]. In Ref. [38],
for the neutrino-nucleus cross sections, only the ratio of ν e−W + to charged-current e → (CC) DIS on 16O and 56Fe were shown, and only the coherent regime was considered (see
Table 3.1). In Refs. [39, 40], all three flavors were considered and shown, and all three
scattering regimes were considered. However, all three regimes used EPA, and nuclear ef-
fects (mainly Pauli blocking) were not included. Moreover, for the inelastic regime, only
the photon-initiated subprocess was calculated (see Table 3.1).
Figure 3.3 shows their results. All three scattering regimes are important. The high
threshold is set by the W -boson mass and the hadronic structure functions. The diffractive
regime has a lower threshold than the coherent regime because larger Q2 ( q2; virtuality ≡ − of the photon) can be probed by the nucleon form factor than the nuclear form factor.
Above threshold, the coherent cross section ( Z2) is larger than the diffractive cross ∝ section ( Z). ∝ The coherent cross section of Seckel [38] is about two times that of Alikhanov [39, 40],
possibly due to their treating the nuclear form factor differently (as pointed out by Ref. [40]).
(The origin of the factor of two between them could not be traced, as the details of the
calculations are not given in Ref. [38].)
Importantly, this cross section is claimed to be 10% of the charged-current deep ∼ inelastic scattering (CCDIS) cross section [24], indicating this process is detectable by high-
energy neutrino detectors like IceCube [243], KM3NeT [258], and especially the forthcoming
55 37 10 Coherent Diffractive 38 CCDIS 10 Inelastic Alikhanov 2016
] 39
1 10 V
e 40
G 10
2
m 41 c
10 Magill & Plestid 2017
[ Altmanshofer+ 2019
42 Seckel 1998 E
/ 10 (coherent only) 43 10 Ballett+ 2018 + + 44 Ar e e Ar e O e W X 10 Tridents, coherent only W boson production 0.1 1 10 102 103 104 105 106 107 108 E [ GeV ]
Figure 3.3: Summary of cross sections for W -boson and trident production from previous work, with the two processes separated as labeled. To simplify the figure, for W -boson production, we show only ν e−W + on 16O (by Seckel [38] and Alikhanov [39, 40]), and e → for trident production, only the coherent regime (the dominant part) of ν ν e−µ+ on 40Ar µ → e (by Magill & Plestid [41], Ballett et al. [42], and Altmannshofer et al. [43]). Also shown, for comparison, is the cross section of charged-current deep inelastic scattering (CCDIS) [24].
IceCube-Gen2 [115]. With 60 starting events with energies above 60 TeV [259, 260], Ice-
Cube already has a nominal precision scale of 13%, and IceCube-Gen2 would be 10 times larger. However, on the theory side, due to the limitations above, more complete and careful calculations are needed (see Table 3.1).
56 Table 3.1: Summary of the features of previous calculations and of this work. “+” and “ ” means “considered” and “not considered” − in the calculation respectively. “Full SM” means using full Standard Model, instead of four-Fermi theory.
Full SM Coherent Diffractive Beyond EPA Pauli blocking Inel., photon Inel., quark W -boson Seckel [38] + + − − − − −
57 production Alikhanov [39, 40] + + + + − − − Altmannshofer et al. [226] + − − − − − − Magill & Plestid [41] + + + Trident − − − − Ge et al. [118] + production − − − − − − Ballett et al. [42] + + + + − − − Altmannshofer et al. [43] + + + + − − − Both, unified This work + + + + + + + 3.2.3 Trident production
The trident processes (Fig. 3.2) raised interest at a similar time to W -boson production, also as a process to probe the then-hypothetical W/Z propagators in the weak interactions.
Even if the weak bosons were not produced directly due to, e.g., their large masses, their existence could make the trident production rate different from that of the pure V-A theory
(e.g., Refs. [264–266, 276–281])
So far, only the ν ν µ−µ+ process has been observed, by the Charm-II [282] and µ → µ CCFR [283] experiments; NuTeV [284] set an upper limit. These results are consistent with
SM predictions.
The trident processes have been popular again in recent years, due to currently running and upcoming accelerator neutrino experiments (e.g., Refs. [285–288]) as well as Ref. [226] showing first trident constraints on new physics such as Z0 models. Table 3.1 summarizes the calculations of trident cross sections by Refs. [41, 118, 226] using EPA and by Refs. [42, 43] using an improved calculation. Usually, only the electron and muon flavors are considered, as the tau flavor is rare for accelerator neutrinos. The inelastic regime is very small, so not considered (except in Ref. [41]).
All previous work used the four-Fermi theory, instead of the full Standard Model (see
Table 3.1). One reason is that these papers focused on accelerator neutrinos below ∼ 100 GeV. Another reason is that the hadronic part (Sec. 3.2.1) complicates the calculation a lot, so using the four-Fermi theory for leptonic part is significantly simpler.
Figure 3.3 summarizes previous calculations. The threshold is set by the final-state lep- ton masses and hadronic structure functions. The difference between Magill & Plestid [41] and Ballett et al. [42], Altmannshofer et al. [43] is due to the former using EPA, while the latter two not.
Though at GeV energies the cross sections are 10−5 of CCDIS [24], they increase ∼ quickly. Therefore, it is interesting and important to know the cross sections at TeV–PeV energies. To this end, the full Standard Model is needed instead of the four-Fermi theory.
In addition, our calculations fix several other shortcomings (see Table 3.1).
58 3.3 Cross sections between neutrinos and real photons
In this section, we calculate the cross sections of W -boson and trident production between a neutrino and a real photon. This shows the underlying physics and the basic behavior of the cross sections. The connection between the two processes is also clearly revealed.
For the cross sections between elementary particles, we calculate the matrix elements and phase space integrals ourselves, and check the results using the public tools MadGraph
(v2.6.4) [289] and CalcHEP (v3.7.1) [290]. The calculational procedures set the basis for the off-shell cross sections in Sec. 3.4.
3.3.1 W -boson production
The leptonic part of W -boson production is (Fig. 3.1),
ν + γ `− + W + . (3.3) ` →
The cross section can be calculated using
Z 1 1 X WBP 2 σνγ(sνγ) = dPS2 , (3.4) 2sνγ 2 |M | spins where 1/2s is the Lorentz-invariant flux factor, s (k + q)2, 1 P WBP 2 the νγ νγ ≡ 1 2 spins |M | photon-spin averaged matrix element (Appendix 3.8.1), and dPS (= √pCM d cos θ ) is the two- 2 sνγ 8π body phase space, of which pCM is the momentum of the outgoing particle in the CM frame, with angle θ respect to the incoming particle. This process has been calculated by
Refs. [38, 39]. Our calculation gives the same results.
The diagrams for Eq. (3.3) are similar to those in Fig. 3.1, but replacing the photon from the nucleus with a free (real) photon. Both diagrams, with a relative minus sign, need to be included to assure gauge invariance. Numerically, the first diagram dominates at small sνγ, while the second dominates at large sνγ. Neutrinos and antineutrinos have the same total and differential cross sections, as the matrix element is invariant under CP transformation.
Figure 3.4 shows σνγ(sνγ)/sνγ for W -boson production. We divide out sνγ, the dominant trend, to highlight the deviations over the wide range of the x axis. The threshold is set
59 37 10 e +
38 + W 10 -b Tridents (CCo and CC+NC) so + n p
] r
39 o d 2 10 u c ti V o
e n G
+
2 40 10 ee m
c +
[ e + ee e Tridents (NC)
s 41 / 10 e + + , e + + , + +
+ e e, e, 42 e 10
, e
+ + 43 10 , e
e 10 6 10 5 10 4 10 3 10 2 0.1 1 10 102 103 104 105 106 s [ GeV2 ]
Figure 3.4: Our cross sections (actually σνγ(sνγ)/sνγ) for W -boson and trident production, between a neutrino and a real photon as a function of their CM energy. Red, green, and blue lines are νe-, νµ-, and ντ -induced channels, respectively. Solid lines are trident CC channels, and dashed lines are trident CC+NC channels (we label only the final states for both). Magenta dotted lines are trident NC channels, which depend on only the final- state charged leptons (we label both the initial and final states). The trident CC, NC, and CC+NC channels correspond to diagrams (1)–(3), (4)–(5), and (1)–(5) of Fig. 3.5. The corresponding antineutrino cross sections (i.e., obtained by CP-transforming the processes shown) are the same. See text for details.
2 by sνγ = (mW + m`) . Just above threshold, the lepton propagator in the first diagram (Fig. 3.1) gives a logarithmic term, log[(...)/m2] , which leads to σ > σ > σ [38]. ∼ ` νeγ νµγ ντ γ 6 2 For sνγ > 10 GeV , the cross sections become constant and different flavors converge, with σ 2√2αG 10−34 cm−2. νγ ' F '
60 − l1 l1 + l2
l1 W + + W l2 l2 W+ − + l1 (1) (2) l2
A A’ A A’ − l1 l1 W+ l2
(3) + l2
A A’
l1 k1 k2 l1 l1 l1 Z + Z − p1 l2 l2
(4) p − (5) + 2 l2 l2 q P P’ A A’ A A’ Figure 3.5: Diagrams for trident production via photon exchange in the Standard Model, with the order, Tri– Tri, labeled in parentheses, and with the momenta labeled on the M1 M5 fourth diagram. The trident CC, NC, and CC+NC channels correspond to diagrams (1)– (3), (4)–(5), and (1)–(5). For antineutrinos, take the CP transformation of the elementary particles. The first and second diagrams are connected to W -boson production (Fig. 3.1; also see Sec. 3.3.2 for details of the connection).
61 3.3.2 Trident production
The leptonic part of trident production, for each incoming neutrino flavor, is (Fig. 3.5)
ν + γ `− + ν + `+ (CC), (3.5a) `1 → 1 `2 2 ν + γ ν + `− + `+ (NC), (3.5b) `1 → `1 2 2 ν + γ ν + `− + `+ (CC+NC), (3.5c) ` → `
where ` , ` = e, µ or τ and ` = ` . So there are two, two and one CC, NC and CC+NC 1 2 1 6 2 channels, respectively; details below. For antineutrinos, take the CP transformation of the
elementary particles; details below.
The cross section can be calculated using [41, 226, 291]
Z 1 1 X Tri 2 σνγ(sνγ) = dPS3 , (3.6) 2sνγ 2 |M | spins where 1/2s is the Lorentz-invariant flux factor, 1 P Tri 2 the photon-spin averaged νγ 2 spins |M | matrix element, and dPS3 is the three-body phase space (see below). Different from previous calculations, here we need to use the full Standard Model, instead of the four-Fermi theory, as we are also interested in TeV–PeV energies.
Figure 3.5 shows the five possible diagrams of trident production in the Standard Model.
Note that the diagram involving a W W γ vertex is not included by the four-Fermi theory, though it is suppressed at low energies. When ` = ` , the top three diagrams (exclusively 1 6 2 mediated by W ) lead to CC channels, and the bottom two (exclusively mediated by Z) lead to NC channels. When `1 = `2 = `, all five diagrams give the same final states, which lead to the CC+NC channels.
We work in the unitarity gauge, which is simpler for the tree level. The amplitudes for each diagram, Tri— Tri, can be found in Appendix 3.8.2, and relative signs between M1 M5 these diagrams are
Tri = ( Tri Tri + Tri) ( Tri + Tri) . (3.7) M M1 − M2 M3 − M4 M5
The matrix element is calculated using FeynCalc [292, 293].
62 For antineutrinos, the total cross sections are the same as neutrinos, due to CP in- variance [118, 266]. For the differential cross sections, they are only the same for the NC channels due to interchange symmetry of two charged leptons, which is not the case for CC and CC+NC channels [118]. Therefore, for the following discussion, we take neutrinos only.
The three-body phase space in the case of real photon is [41, 226, 291]
1 1 dt dl dΩ00 dPS3 = 2 β(l) , (3.8) 2 (4π) 2sνγ 2π 4π where t 2q (k k ), l (p + p )2,Ω00 the solid angle with respect to q in the rest frame ≡ · 1 − 2 ≡ 1 2 of p + p , and 1 2 r 2(m2 + m2) (m2 m2)2 β(l) = 1 1 2 + 1 − 2 , (3.9) − l l2
2 where m1, 2 is the mass of p1, 2. The integration over l is done from (m1 + m2) to sνγ,
and t from l to sνγ. Using these variables, we find that the numerical integration converges reasonably fast for both the four-Fermi theory case and the Standard-Model case.
Figure 3.4 shows σνγ(sνγ)/sνγ for all 15 trident channels. The thresholds are set by the
2 masses of final states, i.e., sνγ = (m1 + m2) . The cross sections increase from threshold until 106 GeV2. For s > 106 GeV2, same as for W -boson production, the cross sections ∼ νγ become constant.
For the CC and CC+NC channels, very interestingly, just above s = m2 6.5 νγ W ' × 103 GeV2, there is a sharp increase. This is due to the s-channel like part of the first and
2 second diagrams in Fig. 3.5, which are mediated by W bosons. For sνγ > mW , W -boson production is turned on. In the view of trident production, this is a W -boson resonance followed by decay to a neutrino and a charged lepton. Therefore the CC and CC+NC trident cross sections are enhanced by the W -boson production cross section (of same incoming neutrino flavor) times the corresponding decay branching ratio, Γ − /Γ ( 11%, with W →ν`` W ' slight deviation for specific flavors [242]). The W -resonance contribution keeps dominating
2 for sνγ > mW . This is different from usual resonance features, like the Glashow resonance ν¯ + e− W − or e−e+ Z. The reason is that the charged lepton, `−, could take e → → 2 2 away additional 4-momentum, keeping the s-channel W propagator on shell (qW = mW ).
63 Contributions from the non-resonant part and from the other three non-resonant diagrams
2 are negligible. So for sνγ > mW , the six CC and three CC+NC channels basically form 6 2 three groups due to the three neutrino flavors. For sνγ > 10 GeV , they all converge and
−35 −2 become constant, i.e., σ 2√2αG Γ − /Γ , which 10 cm . ' F × W →ν`` W ' For the NC channels, the cross sections are much smaller, as there is no resonance. The channels that have same charged lepton final states have same cross sections, independent of incoming neutrino flavor, due to the same couplings and lepton propagators.
3.4 Neutrino-nucleus cross sections: coherent and diffractive regimes
In this section, we calculate the photon-mediated neutrino-nucleus cross sections, σνA, for W -boson and trident production in the coherent and diffractive regimes, which are elastic on the nucleus and nucleon, respectively. We focus on the hadronic coupling through virtual photons, as the contribution through weak bosons is highly suppressed due to their large masses. We first describe the framework (Sec. 3.4.1), which is independent of the leptonic part. Then we calculate the W -boson (Sec. 3.4.2) and trident (Sec. 3.4.3) production processes.
The framework we use, which is from Ballett et al. [42], is a complete treatment of the hadronic part instead of using EPA, which is known to be not a good approximation for trident production (see Fig. 3.3 and Sec. 3.2). Moreover, the major nuclear effect, Pauli blocking, is included (see Refs. [42, 266] for details). In this work, for the first time, we show that the EPA also does not work well for W -boson production. Moreover, for trident production, we calculate all 15 possible channels, including for the τ flavor, and go to
TeV–PeV energies, using the full Standard Model instead of the four-Fermi theory.
3.4.1 Framework
Both the coherent and diffractive cross sections can be calculated using [42]
d2σ 1 1 νX = [hT Q2, sˆ σT Q2, sˆ + hL Q2, sˆ σL Q2, sˆ] , (3.10) dQ2dsˆ 32π2 sQˆ 2 X νγ X νγ 64 where X is to distinguish coherent (X = c) and diffractive (X = d) regimes, Q2 q2 the ≡ − photon virtuality, ands ˆ 2(p q) = s + Q2. Note that Eq. (3.10) decomposes the σ ≡ 1 · νγ νX T/L into 2 2 parts: transverse (“T”) and longitudinal (“L”), leptonic (σνγ ) and hadronic × T/L (hX ). T/L 2 The leptonic parts, σνγ (Q , sˆ), may be viewed as the cross sections between a neutrino and an off-shell photon, and it can be calculated as
1 Z 1 X 4Q2 σ = gµν + kµkν L L∗ dPS , (3.11a) T 2ˆs 2 − sˆ2 1 1 µ ν n spins 1 Z X 4Q2 σ = kµkνL L∗ dPS , (3.11b) L sˆ sˆ2 1 1 µ ν n spins
where Lµ is the leptonic amplitudes, details in Appendices 3.8.1 and 3.8.2, and dPSn is the phase space of the leptonic part, with n = 2, 3 for the W -boson and trident production processes, respectively .
A factor of 1/2 appears in the first equation because a virtual photon has two transverse polarizations. The Q2 dependence should also be included in both the leptonic matrix element and phase space, which are process dependent. In the limit Q2 = 0, the transverse cross section is the same as the real-photon case (Eqs. (3.4) and (3.6)), and the longitudinal cross section vanishes.
T/L 2 The hadronic parts, hX (Q , sˆ) are dimensionless factors that involve the nu- clear/nucleon form factors. For the coherent regime, we use the Woods-Saxon (nuclear)
form factor. For the diffractive regime, we use the nucleon form factors that have a dipole
parametrization. More details are given in Ref. [42].
For the diffractive regime, in addition, the Pauli-blocking effects are included by mul-
tiplying Eq. (3.10) by a factor derived from modeling the nucleus as ideal (global) Fermi
gas of protons and neutrons with equal density, which is (derived by Ref. [270] and used by
Refs. [42, 43, 276]),
3 3 ~q 1 ~q | | | | , if ~q < 2 kF , f( ~q ) = 2 2 kF − 2 2 kF | | (3.12) | | 1, if ~q 2 k , | | ≥ F
65 38 Coherent and diffractive only 10
39 10 C CoherentCD
] e
IS
1 × 0. Diffractive 1 V 40 e 10 G
2 m c
[
41 10 Diff. (neutrons only) E
/ Coh. (longitudinal only) A
42 10
43 10 4 5 6 7 8 9 10 10 10 10 10 10 E [ GeV ]
Figure 3.6: Our coherent and diffractive components of W -boson production cross sec- tions, ν `− + W +, on 16O. Red, green, and blue lines are ν -, ν -, and ν -induced ` → e µ τ channels, respectively. Solid: coherent (right bump) and diffractive (left bump) compo- nents. Dashed: longitudinal contribution to the coherent regime, which is small, even for the largest case (νe). The ντ line is not shown due to being below the bound of the y axis. Dotted: contribution from neutrons to the diffractive regime, which is small. The corresponding antineutrino cross sections are the same.
where kF = 235 MeV is the Fermi momentum of the gas, which sets the kinetic boundary for the final states, and ~q is the magnitude of the transferred 3-momentum in the lab | | p 2 2 2 frame, which can be derived to be, for the virtual photon case, (Q /2MN ) + Q , where
66 Coherent and diffractive only
Seckel, coh. only Alikhanov (EPA) Our EPA 39 Our full calculation 10 Coherent ]
1 C C
V D
e IS
G × 0 2 .1
m Diffractive c
[
E
/ 40
A 10
4 5 6 7 8 10 10 10 10 10 E [ GeV ]
Figure 3.7: Our coherent and diffractive components of W -boson production cross sections (red solid, from Fig. 3.6 but thicker), for the example of ν `−W + on 16O (the flavor e → with the largest cross section), comparing with our “EPA + no Pauli blocking” results (dashed) and previous calculations (dotted) by Seckel [38] and Alikhanov [39, 40]. Left and right bumps are coherent and diffractive components, respectively. Note our results are substantially smaller, which is important.
MN is the mass of the nucleon. This reduces the diffractive cross section by about 50% for protons and 20% for neutrons.
2 T/L 2 The EPA formalism can be obtained by setting Q = 0 in σνγ (Q , sˆ) of Eq. (3.10). This is basically the same as that initially derived by Ref. [281] and later used by Refs. [41, 226].
67 Below, we also show the “EPA + no Pauli blocking” results for comparison.
As a validation of our understanding of the formalism, we reproduced the cross sec- tion results of Ballett et al. [42] using the four-Fermi theory and other same input. Our calculations agree with theirs to within a few percent, with the remaining differences due to numerical precision. (Note that we decompose the phase-space (Appendix 3.8.3) in a different, but equivalent, way from them [42, 266].)
3.4.2 W -boson production
T/L 2 Off-shell cross sections, σνγ (ˆs, Q )
The process is the same as Eq. (3.3), but replacing the real photon, γ, by a virtual photon
∗ T/L 2 γ . We calculate the off-shell cross sections, σνγ (ˆs, Q ), in the CM frame (consistent with Sec. 3.3.1), using Eq. (3.11). For the leptonic matrix element, the photon virtuality can be included by writing
2 2 sνγ + Q sνγ + Q k1 = , 0, 0, , (3.13a) 2√sνγ 2√sνγ s Q2 s + Q2 q = νγ − , 0, 0, νγ . (3.13b) 2√sνγ − 2√sνγ
The 4-momenta of the outgoing particles do not have Q2 dependence, and the phase space is the same as that in Eq. (3.4). When Q2 = 0, all results return to the real-photon case
(Sec. 3.3.1).
T/L 2 T 2 2 The major features of σνγ (ˆs, Q ) are the following. First, σνγ(ˆs, Q ) decreases with Q , 2 2 especially when Q & m` , because Q enters the denominator of the lepton propagators which L 2 2 suppresses the cross section [42]. Second, σνγ(ˆs, Q ) increases with Q , due to the factor 4Q2/sˆ2, then becomes flat when Q2 m2. Third, when Q2 is nearings ˆ (m + m )2, & ` − W ` T L an exponential cutoff happens in both σνγ and σνγ, due to running out of phase space (s sˆ Q2 < (m + m )2). νγ ≡ − W `
68 σνA and discussion
The coherent (σνc) and diffractive (σνd) cross sections are then calculated with Eq. (3.10), by convolving the leptonic parts with the hadronic parts.
Figure 3.6 shows the cross sections. The features discussed in Sec. 3.3.1 mostly appear here (e.g., σ > σ > σ ). The threshold here is effectively set by E m2 /2Qeff , νeA νµA ντ A ν ∼ W max where Qeff is the effectively maximum Q of the form factors, which is 0.1 GeV for max ∼ 16O (coherent) and 1 GeV for nucleons (diffractive). Similarly, the typical momentum ∼ transfer, Q, for each E , is between m2 /2E and Qeff for both regimes. The sharp peak ν ∼ W ν max in σνγ (Fig. 3.4) is here smeared due to convolving with the form factors. The coherent cross sections ( Z2) are larger than diffractive ones ( Z), which is similar to Fig. 3.3, which ∝ ∝ uses EPA. For both regimes, the transverse part dominates, while the longitudinal part, as shown on the figure, is suppressed by Q2/sˆ. For the diffractive regime, the contribution ∼ from protons dominates, due to its electric form factor.
We tested the sensitivity of our cross section results to the choices of form factors. For
2 2 the coherent component, we also tried using a Gaussian form factor, e−Q /2a , which is
sometimes used for lighter nuclei. For 16O, we find that the cross section is changed by
15% for E 105 GeV (where the inelastic component dominates anyway) and by 5% . ν ∼ . for E 106 GeV. For the diffractive component, we explored changing the vector mass in ν ∼ the form factor. For any reasonable change, the effect on our calculated cross sections is negligible.
Figure 3.7 compares our results with previous ones from Ref. [38] (only the coherent regime was considered) and Refs. [39, 40] (EPA was used, and Pauli-blocking effect was not included for the diffractive component), and our “EPA + no Pauli blocking” results discussed above. Comparing to Refs. [39, 40], which uses a different EPA formalism, our
EPA result is close. Surprisingly, the result from the full calculation is only about half as large, for both coherent and diffractive regimes. This means that the EPA is still not valid at even such high-energy scales. The reason is that, as discussed in Sec. 3.4.2, the
2 T 2 T 2 nonzero Q suppresses the transverse cross section, σνγ(ˆs, Q ), compared to σνγ(ˆs, Q = 0)
69 Coherent and diffractive only 39 10 e C e C W 40 DI + 10 S ×
] 0 .01 1 41 V
e 10 G
+ 2 e
m 42 e c
10 [
+
E e, e,
/ 43 +
A 10
44 10
45 10 1 2 3 4 5 6 7 8 10 10 10 10 10 10 10 10 E [ GeV ]
Figure 3.8: Our coherent (solid lines) and diffractive (dashed lines) components of trident production cross sections on 16O. We show one typical channel for each category, i.e., CC, NC and CC+NC, to make the figure simple. For all the channels, see Appendix 3.8.4. Gray lines are for νe-induced W -boson production from Fig. 3.6, shown as a comparison. The corresponding antineutrino cross sections (i.e., obtained by CP-transforming the processes shown) are the same.
used in the EPA. The larger Q2, the larger the suppression. Physically, this is because,
although the incoming neutrino is very energetic, the charged particle that directly couples
to the photon may not be. The difference between the full calculation and the EPA in
the diffractive regime is larger than that in coherent regime is because the nucleon form
70 factor probes larger Q2 than the nuclear form factor, also because the Pauli-blocking effect suppresses the diffractive cross section. Another feature is that, for a specific Q2, the larger
2 2 the charged-lepton mass, the smaller the suppression, which is small when Q . m` . So the difference between full calculation and EPA is smaller for the muon and tau flavors.
3.4.3 Trident production
T/L 2 Off-shell cross sections, σνγ (ˆs, Q )
The processes are the same as in Eq. (3.5), but again replacing the real photon, γ, by a
virtual photon γ∗. Same as above, we work in the CM frame, and both the phase space
term dPS3 and the leptonic matrix element are modified due to nonzero photon virtuality. The leptonic matrix element is modified due to the modification of the 4-momenta, the
details of which can be found in Appendix 3.8.3.
The phase space integration can be done by decomposing the three-body phase space
into two two-body phase spaces [41, 294]. The result is the same as Eq. (3.8), but replacing
s bys ˆ. The integration range is now ((m + m )2, sˆ Q2) for l, and νγ 1 2 − l l + Q2, sˆ Q2 + 2 Q2 (3.14) − − sˆ Q2 − for t. See Appendix 3.8.3 for details.
The major features due to the nonzero Q2 is the same as those of W -boson production
(Sec. 3.4.2).
σνA and discussion
Figure 3.8 shows the cross sections for the typical channels (for all channels, see Ap-
pendix 3.8.4). We start from Eν = 10 GeV, as below this energy, the cross sections have been shown in Refs. [42, 43]. Our results agree with theirs. Same as before, the threshold
here is effectively set by E (m + m )2/2Qeff , which is 0.1 GeV for 16O and 1 GeV ν ∼ 1 2 max ∼ ∼ for nucleons. The sharp peak in σνγ (Fig. 3.4) is smeared here due to convolving with the form factors. Other features and the physics are the same as those discussed in Sec. 3.3.2
71 for Fig. 3.4 and in Sec. 3.4.2 for Fig. 3.6.
3.5 Neutrino-nucleus cross sections: inelastic regime
In this section, we calculate the neutrino-nucleus cross sections, σνA, for W -boson and trident production in the inelastic regime, in which the partons of nucleons are probed.
3.5.1 Framework
The inelastic regime has two contributions, photon-initiated subprocess and quark-initiated subprocess [267].
The photon-initiated subprocess is that the hadronic coupling is through a virtual pho- ton, which is similar to Sec. 3.4, but with larger photon virtuality Q2. Calculation of this subprocess involves the photon PDF, which describes the photon content of the nucleon
(e.g., Refs. [267, 295–302]). The photon PDF consists of elastic and inelastic components.
The elastic component corresponds to the diffractive regime of Sec. 3.4, and can be calcu- lated from the nucleon electromagnetic form factors. The inelastic photon PDF consists of nonperturbative and perturbative parts, and the resonance region is included in the former [267, 300, 301]. For the W -boson production, this component was calculated by
Alikhanov [40]. For trident, this has never been considered.
The quark-initiated subprocess is that a quark of a nucleon is explicitly involved as an initial state of the scattering process. The propagator that couples to a quark can be photon,
W or Z boson. For W -boson production, this was mentioned in Alikhanov [40] (diagrams were also shown in its Fig. 8, plus another one from replacing the Z by a photon in the upper middle diagram) but not calculated. For trident production, this was calculated in
Ref. [41].
Those two subprocesses are at the same order though they may not seem to be, as the photon propagator to quark has an additional αEM. The reason is that the photon PDF has a factor of αEM implicitly [267]. A double-counting problem occurs if summing up the two subprocesses for the total
72 inelastic cross section. This is because the contribution from the photon propagator to the quark of the quark-initiated subprocess is already included in the inelastic photon PDF
(perturbative part) of the photon-initiated subprocess. We deal with this problem below.
For the PDF set, we use CT14qed [267, 303], which provides the inelastic photon, quark, and gluon PDFs self-consistently. The inelastic photon PDF of CT14qed is modeled as emission from the quarks using quark PDFs and further constrained by comparing with
ZEUS data on the DIS process ep eγ + X [304]. The quark PDFs are obtained by the → usual method and constrained by DIS and other data.
The reasons that we choose CT14qed are the following. First, it is the only PDF set that provides the inelastic component of photon PDF only. For the elastic part, we do not use the elastic photon PDF, which is obtained using EPA and does not include the neutron magnetic component form factor, as the treatment in Sec. 3.4 (diffractive regime) is better.
Second, it is also the only PDF set that provides the inelastic photon PDF for both proton and neutron. Finally, though the uncertainty of photon PDF is larger than the later ones by LUXqeD [300, 301] and NNPDF31luxQED [302], the central value is very close.
We use MadGraph (v2.6.4) [289] to do the calculation, which handles the PDFs and hard processes systematically. We remove kinematic cuts to get the total cross section of both processes. The model we choose in MadGraph is “sm-lepton masses”, which includes the masses of charged leptons, while the default “sm” does not. Moreover, this model uses diag(1, 1, 1) for the CKM matrix and ignores the masses of u, d, s, c quarks, which are good approximations for us. Note that for the initial-state neutrinos, which have only a left-handed chirality, we need “set polbeam1 = -100” (+100 for antineutrinos) to fully
polarize the beam, otherwise the cross section will be mistakenly halved. As a check of above
configuration, we calculate the neutrino CCDIS cross sections and the result is consistent
with Refs. [24, 233, 237, 239, 305–307] within uncertainties.
3.5.2 W -boson production
For photon-initiated process, the diagrams are shown in Fig. 3.1. The factorization and
renormalization scales are chosen to be √sνγ. Our choice is consistent and has no ambiguity 73 Inelastic only
photon-initiated subprocess photon-initiated, protons only 39 photon-initiated, neutrons only 10 quark-initiated subprocess quark-initiated, photon propagators only quark-initiated, W & Z propagators only ]
1 C C D V IS e ×
G 0 .
2 1
m 40 c
[ 10
E /
41 10 4 5 6 7 8 10 10 10 10 10 E [ GeV ]
Figure 3.9: Different components of our inelastic neutrino-nucleus cross sections for W - boson production. Only νe is shown to keep the figure simple. For νµ and ντ , the photon- initiated cross sections are smaller (Fig. 3.10), while the quark-initiated cross sections are basically the same. See text for details.
p for both diagrams of Fig. 3.1 compared to (k p )2 (motivated by the first diagram) − 1 − 1 p or (k p )2 (motivated by the second diagram). The result is only 10% larger than − 1 − 2 ' that using the default factorization and renormalization scales of MadGraph. Changing both scales to 2√sνγ or √sνγ/2 would increase or decrease the cross section by 15%. For quark- ∼ initiated subprocess, the diagrams can be found in Fig. 8 of Ref. [40], plus another one from
74 Inelastic only
Alikhanov, e 39 10 e C C D IS ]
×
1 0 .1 V e G
2
m 40 c
[ 10
e E D / i f , f t ra n c e t r iv e e h , o e C
41 10 4 5 6 7 8 10 10 10 10 10 E [ GeV ]
Figure 3.10: Our inelastic neutrino-nucleus cross sections for W -boson production on 16O (solid lines), for all three flavors. Also shown are previous results from Alikhanov [40] and, for comparison, coherent and diffractive cross sections of νe from Fig. 3.6. The corresponding antineutrino cross sections are the same.
replacing the Z by a photon in its upper middle diagram. We use the default factorization and renormalization scales of MadGraph, as using √sν,quark causes calculational problems.
Figure 3.9 shows the results for νe. (For νµ and ντ , the discussion below also ap- plies.) The photon-initiated subprocess is much larger than the quark-initiated process, because softer photons are favored (photons in the low-Q2 region). The contribution
75 Inelastic only Inelastic only
40 40 10 10
41 ] ] 41 10 1 10 1 V V e e 42 G G 10 2 2
m 42 m c c
10 [ [
43 10 E E / / + + A A e, e, + + 43 e, e, 10 + 44 + ee , , e e + 10 e e + + ee + + + e e 44 45 e ee e 10 0 1 2 3 4 5 6 7 8 10 0 1 2 3 4 5 6 7 8 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 E [ GeV ] E [ GeV ]
Figure 3.11: Our cross sections for trident production in the inelastic regime. Left: CC channels. Right: CC+NC and NC channels. Solid: the photon-initiated subprocess. Dashed: quark-initiated subprocess. The corresponding antineutrino cross sections are the same.
from protons is larger than that from neutrons, and their ratio is similar for both subprocesses. For the quark-initiated subprocess, also shown are the contributions from
photon-propagator diagrams 2 and from the W/Z-propagator diagrams 2, with the former | | | | being much larger than the latter. This indicates the relative importance, though the calculation of each component separately would break gauge invariance, especially above
E 108 GeV where it is not numerically stable and the mixings between photon and ν ∼ weak bosons are large. However, because the photon-propagator diagrams dominate the quark-initiated subprocess and, as mentioned in the last subsection, are already included in the photon-initiated process (inelastic photon PDF), we can ignore the quark-initiated subprocess in our calculation. (In other words, as long as the solid lines in Fig. 3.10 are much larger than the dot-dashed line in Fig. 3.9, the quark-initiated process can be ignored.) This also avoids the double-counting problem mentioned above. A more complete treatment that includes both subprocesses while avoiding double counting is beyond the scope of this work. One way is to use the W and Z PDFs [308–310], which may appear in future PDF sets.
76 Figure 3.10 shows the inelastic cross sections for all three flavors. As before, σνeA >
σνµA > σντ A. Also shown are νe coherent and diffractive cross sections, for comparison. The inelastic cross section is the largest for most energies. Our result is smaller than the result of Alikhanov [40]. The difference is due to multiple reasons. First, we use much more up-to-date photon PDFs. Second, we use the dynamic scale √sνγ, which is more appropriate, while they used the fixed scale, mW . Third, we use MadGraph which does a full systematic calculation while they used the EPA.
3.5.3 Trident production
Figure 3.11 shows the cross sections for trident production processes in the inelastic regime.
The calculational choice is the same as that for W -boson production in the last subsection.
Different from before, here we separate CC channels (left) and CC+NC, NC channels (right) into two different panels. Our calculation is the first to include the nonperturbative part of the photon PDF for the trident production in which the resonance region is included. So here we start from Eν = 1 GeV. For CC channels, similar to those of W -boson production, the photon-initiated subprocess is much larger than the quark-initiated subprocess, and the latter is dominated by photon propagators to the quarks (not shown, for simplicity), which is already included by the former. Therefore, for the total inelastic cross section, we can also ignore the quark-initiated subprocess.
For CC+NC and NC channels, interestingly, the quark-initiated subprocess could be
(much) larger than the photon-initiated subprocess. The reason is that a pair of charged leptons can be split by a virtual photon emitted from the initial- or final-state quark on top of the NC DIS process, which does not happen in the CC channels as the two charged leptons have different flavors. (Radiative corrections through virtual W /Z bosons also exist but are suppressed by their large masses.) Therefore, the quark-initiated subprocesses of CC+NC and NC channels are enhanced compared to those of CC channels, and the lighter the final-state charged leptons, the larger the enhancement. For the total inelastic cross section of CC+NC and NC channels, we can sum the two subprocesses up, as the contribution from the double-counting region is negligible.
77 38 10 e
39 + C 10 CD IS × + 0.1 ] e e
1 W- 40 bos on V 10 pro e du ctio
G n Tridents (CC and CC+NC) 2 +
m e 41 ee , c e 10 , e e + [
+ ee e, Tridents (NC)
E + e, + /
A 42 e 10 + + + + ee e, e, + 43 10
44 10 1 2 3 4 5 6 7 8 10 10 10 10 10 10 10 10 E [ GeV ]
Figure 3.12: Our total cross sections (actually σνA/Eν) for W -boson and trident production on 16O. The colors and line styles are same as in Fig. 3.4( red, green, and blue lines are νe-, νµ-, and ντ -induced channels, respectively; solid lines are CC channels, and dashed lines are CC+NC channels; magenta dotted lines are NC channels, which depend on only the final-state charged leptons). The trident CC, NC, and CC+NC channels correspond to diagrams (1)–(3), (4)–(5), and (1)–(5) of Fig. 3.5. The corresponding antineutrino cross sections (i.e., obtained by CP-transforming the processes shown) are the same. See text for details.
Our result is different from the DIS cross sections calculated in Ref. [41]. With limited details provided by Ref. [41], it is hard to trace the exact reason. In addition to the quark PDFs considered by Ref. [41], we also consider the photon PDFs, which has the nonperturbative part. For the quark-initiated process, we may include more diagrams, including the radiative-correction diagrams mentioned above.
78 0.25 + e + A e W A + A W + A + A W + A 0.20 Seckel e (iron)
0.15 Our e (iron)
0.10
0.05 Ratio to CCDIS cross section
0.00 4 5 6 7 8 10 10 10 10 10 E [ GeV ]
Figure 3.13: Ratios of the W -boson production cross sections to those of CCDIS ((ν + ν¯)/2) [24]. Solid lines are for water/ice targets, dotted line for iron targets, and dashed lines are for the Earth’s averaged composition. Color assignment is in the legend. Also shown is the νe (iron) case of Seckel [38], much larger than ours.
3.6 Total cross sections of W -boson and trident production, and ratio to CCDIS
16 Figure 3.12 shows the total cross sections (σνA/Eν) with O for W -boson and trident production, which summarize our calculations in previous sections. We divide out Eν, the dominant trend, to highlight the deviations over the wide range of the x axis. Specifically, 79 the total cross sections are obtained by summing up the coherent, diffractive and inelas- tic components, for each interaction channel. For the W -boson production and trident
CC channels, the inelastic cross section is the photon-initiated subprocess, while for tri- dent CC+NC and NC channels, it is from summing up both photon- and quark-initiated subprocesses. Reasons are discussed in Sec. 3.5.3.
We estimate the uncertainties of the W -boson and trident cross sections. For the coher- ent regime, it is about 6%, coming from higher order electroweak corrections (dominant) and nuclear form factors (see Ref. [43] for more details). For the diffractive regime, the nuclear effects lead to larger uncertainties. Our calculations include the Pauli-blocking fac- tor derived from ideal Fermi gas model for nucleus and ignore other subdominant effects.
Therefore, to be conservative, we assume 30% uncertainty, following Ref. [43]. (Further de- tails on the sensitivity to the choices of nuclear and nucleon form factors are given above.)
For the inelastic regime, the uncertainty mainly comes from choosing the factorization and renormalization scale, which is 15%, and mixing between photon and weak boson which ' matters more at higher energy, for which we give 25% to be conservative (Sec. 3.5). There ' is no study of the nuclear uncertainties of photon PDF, but they should be subdominant especially for light nucleus, like 16O, considering that of the quark PDFs are subdomi- nant [311–314]. Their combination gives 30% for the inelastic regime. Combining the ' three regimes in quadrature, the uncertainty is estimated to be 15%. ' Figure 3.13 shows the ratios of the important channels to the neutrino CCDIS cross sec-
CCDIS CCDIS tions ((σν + σν¯ )/2) [24], for different targets (right panel), including water/ice (for neutrino detectors), iron and the Earth’s averaged composition (for neutrino propagation).
The coherent cross sections ( Z2) for different isotopes are calculated in the same way ∝ as for 16O. For diffractive and inelastic cross sections, one can just rescale by atomic and mass numbers from 16O, as the cross section on single nucleon is independent of nucleus in above formalism (the Pauli-blocking factor in the diffractive cross section derived from ideal
Fermi-gas model [42, 43, 276] is nucleus independent). For hydrogen (1H) only, there is no coherent component, and the Pauli-blocking factor should not be included. The CCDIS cross sections for different isotopes are calculated by multiplying their mass number by the
80 CCDIS cross section on isoscalar nucleon target [24]. The bias caused by the fact that some isotopes have a slightly different number of protons and neutrons and by nuclear effects on
PDFs is negligible.
The maximum ratios of W -boson production to CCDIS are 7.5% (ν ), 5% (ν ), ' e ' µ and 3.5% (ν ) on water/ice target, respectively. For trident production (not shown ' τ on the figure), the CC+NC and NC channels are large enough to matter, i.e., 0.75% ' (ν -induced), 0.5% (ν -induced), and 0.35% (ν -induced). This is a factor 0.1 of corre- e µ τ ' sponding W -boson production channels, as the dominant contribution at these energies are
from the W -boson production followed by W ν + ` (Sec. 3.3.2). These trident channels → ` and also hadronic decay of those W bosons could produce distinct signatures in the high-
energy neutrino detectors like IceCube [243], KM3NeT [258], and especially the forthcoming
IceCube-Gen2 [115]. In IceCube, there could already be a few W -boson production events.
IceCube-Gen2 will have much larger yields. The detectability and implications are detailed
in our companion paper [1]. (For clarification, the W -resonance enhancement part of the trident, which dominates above 104 GeV, is the W -boson production followed by leptonic ∼ decay of the W boson, so they would not be studied separately for detection.)
The maximum ratios of W -boson production to CCDIS are 14% (ν ), 10% (ν ), ' e ' µ and 7% (ν ) on iron, and 11% (ν ), 7.5% (ν ), and 5% (ν ) on the Earth’s ' τ ' e ' µ ' τ averaged composition. (The larger the charge number of a nucleus, the larger the ratio is, due to the coherent component Z2.) Trident CC and CC+NC channels (not shown) ∝ are 0.1 of these numbers, same as above. (As a comparison, the ratio for ν -iron case ' e by Seckel [38] is 25%, much larger than ours (14%).) This affects the absorption rate of high-energy neutrinos when propagating through the Earth, which affects the measurement of neutrino cross sections by IceCube [70, 71, 244–246].
3.7 Conclusions
The interactions of neutrinos with elementary particles, nucleons, and nuclei are a corner- stone of the Standard Model, and a crucial input for studying neutrino mixing, neutrino
81 astrophysics, and new physics. Above E 5 GeV, the neutrino-nucleus cross section ν ∼ is dominated by deep inelastic scattering, in which neutrinos couple via weak bosons to the quarks. However, additionally, there are two processes where the hadronic coupling is through a virtual photon: W -boson and trident production, the cross sections of which
increase rapidly and become relevant at TeV–PeV energies.
In this paper, we do a complete calculation of the W -boson and trident production
processes. We significantly improve the completeness and precision of prior calculations.
We start by giving a systematic review of both processes, pointing out the improvements
that can/should be made compared to previous calculations (Sec. 3.2 and Table 3.1).
Our results can be put into three major categories. The neutrino-real photon cross sections for these two processes over a wide energy • range (Sec. 3.3). This sets the foundation for our neutrino-nucleus cross section
calculation and for discussing the underlying physics (Sec. 3.3). For trident pro-
duction, there are three different categories of interaction channels (Eq. (3.5)):
CC, CC+NC, and NC, arising from different groups of diagrams (Fig. 3.5). Inter-
2 estingly, for the CC and CC+NC channels, above sνγ = mW , the cross sections are enhanced by two orders of magnitude (Fig. 3.4). The reason is that the s-channel
like W -boson propagators in the trident diagrams (Fig. 3.5) can be produced on
shell (i.e., W -boson production in Fig. 3.1) and then decay leptonically. This indicates the unification of these two processes.
A first complete calculation of neutrino-nucleus W -boson production cross sec- • tions. For the neutrino-nucleus cross sections, we handle them in different regimes:
elastic (Sec. 3.4; including coherent and diffractive) and inelastic (Sec. 3.5; includ-
ing both photon- and quark-initiated subprocesses). For W -boson production, we show that the previously used equivalent photon approximation [39, 40] is not good
for its cross section calculation. For the νe-induced channel, our results are about half of those in Refs. [39, 40] for both coherent and diffractive regimes, and 1/4 ∼ of Ref. [38] for coherent regime. The reasons for this are largely understood. A
82 significant factor is the nonzero photon virtuality, Q2, which suppresses the cross
section (EPA assumes Q2 = 0 as used by previous works). Also, the Pauli-blocking
effect suppresses the diffractive cross section. For the inelastic regime, we use an
up-to-date photon PDF and more reasonably dynamic factorization scale, √sνγ. Moreover, we do a first calculation of the quark-initiated subprocess and find that
they can be neglected below 108 GeV. '
A first calculation of neutrino-nucleus trident production cross sections at TeV– • PeV energies. The full Standard Model is used in order to also study the TeV–
PeV behavior, compared to the four-Fermi theory used by previous work. The
equivalent photon approximation is also avoided. Moreover, we do a more careful
treatment of the inelastic regime, and, importantly, we for the first time use the
inelastic photon PDF [267, 303] for trident calculation, which includes the reso-
nance region as a component of the nonperturbative part of the inelastic photon
PDF. More improvements are detailed in the previous sections.
These cross section are large enough to matter (Sec. 3.6). For a water/ice target, the W - boson production cross sections are 7.5% (ν ), 5% (ν ), and 3.5% (ν ) of CCDIS [24]. ' e ' µ ' τ (For the corresponding CC and CC+NC trident channels, they are 0.1 times the num- ' bers above.) This means these processes are detectable, or will be detectable, by high- energy neutrino detectors like IceCube [243], KM3NeT [258], and especially the forthcoming
IceCube-Gen2 [115], with distinct signatures [1]. For the iron target or the Earth’s averaged composition, the W -boson production cross sections are 14%/11% (νe), 10%/7.5% (νµ), and
7%/5% (ντ ) of CCDIS [24]. This affects the absorption rate of high-energy neutrinos dur- ing propagation. Moreover, the DIS cross sections extracted from in-Earth absorption as seen by IceCube [70, 71, 244–246] contain a contribution from W-boson production. Note the fact that the cross section affects the absorption rate exponentially may make these processes even more important than that shown by the numbers above.
This paper sets the theoretical framework and calculates the cross sections of these processes. In our companion paper [1], we discuss the phenomenological consequences, in-
83 cluding the effects mentioned above and other aspects of high-energy neutrino astrophysics that these processes make a difference, such as neutrino cross-section and spectrum mea- surement, flavor ratio determination, neutrino mixing, and new physics.
3.8 Supplemental Material
3.8.1 W -boson production: amplitudes in the Standard Model
The leptonic amplitudes of the diagrams of Fig. 3.1, in the unitarity gauge, are µ egW µ k/ p/2 + m` ν 5 L = u¯ (p1) γ − γ (1 γ )u (k) ν(W ) , (3.15a) 1 2√2 (k p )2 m2 − × − 2 − ` (k−p1)µ(k−p1)ν gµν 2 µ egW ν 5 − mW L = u¯ (p1) γ 1 γ u (k) 2 − 2√2 − × (k p )2 m2 − 1 − W h i gρλ(q + p )µ + gλµ(p p k)ρ + gµρ(k p q)λ (W ) , (3.15b) × 2 1 − 2 − − 1 − λ
where the m` and mW are masses of the charged lepton (p1) and the W boson (p2), respec- tively, and (W ) the polarization vector of the W boson, for which we have
4 X pµpν ν(W )ν(W ) = gµν + 2 2 . (3.16) i i − m2 i=1 W
The amplitudes above are used for both Sec. 3.3.1( WBP = (Lµ Lµ), where µ is the M µ 1 − 2 photon polarization vector) and Sec. 3.4.2(L µ = Lµ Lµ). 1 − 2
84 3.8.2 Trident production: amplitudes in the Standard Model
The leptonic amplitudes of the diagrams of Fig. 3.5, in the unitarity gauge, are
(p1+k2)α(p1+k2)β 2 eg gαβ m2 Lµ = W u¯ (k ) γβ 1 γ5 v (p ) − W 1 8 2 1 2 2 − − × (p1 + k2) mW + imW ΓW − µ p/2 /q + m2 α 5 u¯ (p2) γ − γ (1 γ )u (k1) , (3.17a) × (p q)2 m2 − 2 − − 2
(k1−p2)α(k1−p2)β 2 gαβ 2 µ egW α 5 − mW L = u¯ (p2) γ 1 γ u (k1) 2 8 − × (k p )2 m2 1 − 2 − W h i gγβ( p k k + p )µ + gβµ(k p q)γ + gµγ(q + p + k )β × − 1 − 2 − 1 2 1 − 2 − 1 2 (p1+k2)γ (p1+k2)δ gγδ 2 − mW δ 5 u¯ (k2) γ (1 γ )v (p1) , (3.17b) × (p + k )2 m2 + im Γ × − 1 2 − W W W
(k1−p2)α(k1−p2)β 2 eg gαβ m2 Lµ = W u¯ (p ) γα 1 γ5 u (k ) − W 3 4 2 1 2 2 − − × (k1 p2) mW − − β 5 /q p/1 + m1 µ u¯ (k2) γ (1 γ ) − γ v (p1) , (3.17c) × − (q p )2 m2 − 1 − 1
(k1−k2)α(k1−k2)β 2 5 eg 1 γ gαβ m2 Lµ = Z u¯ (k ) γα u (k ) − Z 4 4 2 2 2 1 2 2 − − × (k1 k2) mZ − − p/ /q + m2 µ 2 − β 1 2 1 5 u¯ (p2) γ γ + 2 sin θW + γ v (p1) , (3.17d) × (p q)2 m2 −2 2 2 − − 2
(k1−k2)α(k1−k2)β 2 5 gαβ 2 µ egZ α 1 γ − mZ L = u¯ (k2) γ u (k1) 5 − 4 2 − 2 × (k k )2 m2 1 − 2 − Z β 1 2 1 5 /q p/1 + m1 µ u¯ (p2) γ + 2 sin θW + γ − γ v (p1) . (3.17e) × −2 2 (q p )2 m2 − 1 − 1 µ µ Note that for L1 and L2 , the width term of the W propagator, imW ΓW , should be included. These are used for both Sec. 3.3.2( Tri = ((Lµ Lµ +Lµ) (Lµ +Lµ)), same as Eq. (3.7)) Mi µ 1 − 2 3 − 4 5 and Sec. 3.4.3(L µ = (Lµ Lµ + Lµ) (Lµ + Lµ)). i 1 − 2 3 − 4 5
85 T/L 2 3.8.3 Trident production: kinematics and phase space for σνγ (ˆs, Q )
Following Refs. [41, 291], we give more details of the kinematics and the three-body phase
T/L 2 space of σνγ (ˆs, Q ) for trident production, and derive the case for the virtual photon. The momenta are labeled in Fig. 3.5.
In the CM frame of the ν-A interaction and treating the two charged leptons together
(p p + p , i.e., the total momentum of `+ and `−), we can write the 4-momenta by ≡ 1 2 s + Q2 k = (1, sin θ, 0, cos θ) , (3.18a) 1 2√s − s + Q2 s Q2 q = − , sin θ, 0, cos θ , (3.18b) 2√s s + Q2 − s l k = − (1, 0, 0, 1) , (3.18c) 2 2√s − s + l s l p = , 0, 0, − , (3.18d) 2√s 2√s
where Q2 q2 is the photon virtuality, l p2, s s (k + q)2 (for simplicity, we use ≡ − ≡ ≡ νγ ≡ 1 s s in this section only), and θ is the angle of the incoming particles with respect to ≡ νγ the direction of p, which is chosen to be the z axis.
Define another variable,
l Q2(cosθ 1) + s + scosθ + s Q2(3 cosθ) + s scosθ t 2q (k k ) = − − − . (3.19) ≡ · 1 − 2 2s
This relation allows us to rewrite sin θ and cos θ in terms of t, which is Lorentz invariant,
86 then putting back into Eq.( 3.18), we obtain p ! Q2 + s (l + Q2 t)(lQ2 + s( 2Q2 s + t)) l(s Q2) + s(3Q2 + s 2t) k = , − − − , 0, − − , 1 2√s s l 2√s(l s) − − (3.20a) p ! s Q2 (l + Q2 t)(lQ2 + s( 2Q2 s + t)) l(s Q2) + s(3(Q2 + s 2t)) q = − , − − − , 0, − − , 2√s l s 2√s(s l) − − (3.20b) s l k = − (1, 0, 0, 1) , (3.20c) 2 2√s − l + s s l p = , 0, 0, − . (3.20d) 2√s 2√s
To find the expression of p1 and p2, it is easier to go to the rest frame of p. We
do a Lorentz transformation to boost to this frame, using β = s+l , γ = s√+l for the s−l 2 sl transformation matrix,
2 p 2 2 2 2 2 ! 0 l + 2Q + s t (l + Q t)(lQ + s( 2Q s + t)) l lt + s(2Q + s t) k1 = − , − − − , 0, − − , 2√l s l 2√l(l s) − − (3.21a) p ! t 2Q2 (l + Q2 t)(lQ2 + s( 2Q2 s + t)) (2ls lt + 2Q2s st) q0 = − , − − − , 0, − − , 2√l l s − 2√l(l s) − − (3.21b) s l k0 = − (1, 0, 0, 1) , (3.21c) 2 2√s − p0 =(√l, 0, 0, 0) . (3.21d)
The situation can be further simplified if we work in the frame where q0 is along the z axis. So we need to do a rotation of above. The rotation angle, ηq, can be determined by
q0[2] q0[4] sin ηq = p , and cos ηq = p . (3.22) q0[2]2 + q0[4]2 q0[2]2 + q0[4]2
87 So, finally, we have s 2 2 2 2 2 2 2 00 l + 2Q + s − t (l + Q − t)(lQ + s(−2Q − s + t)) l(4Q + 2s − t) + (2Q − t)(2Q + s − t) k1 = √ , − , 0, , 2 2 2 r 2 2 l 4lQ + (t − 2Q ) t−2Q2 ( ) 2 −2l l + 4Q (3.23a) r ! t − 2Q2 1 (t − 2Q2)2 q00 = √ , 0, 0, + 4Q2 , (3.23b) 2 l 2 l
p 2 2 2 2 ! 00 s − l (l + Q − t)(lQ + s(−2Q − s + t)) (2ls − lt + 2Q s − st) k2 = √ , − , 0, − , (3.23c) 2 l p(t − 2Q2)2 + 4lQ2 2p(t − 2Q2)2 + 4lQ2 √ p00 = ( l, 0, 0, 0) . (3.23d)
In this frame, the p1,2 can be written as
00 00 00 00 00 00 p1 = E1, +ρ sin θ cos φ , +ρ sin θ sin φ , +ρ cos θ , (3.24a)
p00 = E , ρ sin θ00 cos φ00, ρ sin θ00 sin φ00, ρ cos θ00 , (3.24b) 2 2 − − −
where θ00 and φ00 are the angles with respect to the photon, q00, in the current frame, q 2 2 E1,2 = ρ + m1,2, and
l2 2l m2 + m2 + m2 m22 ρ2 = − 1 2 1 − 2 . (3.25) 4l
The three-body phase space can be done by decomposing it into two two-body phase spaces [41, 294] (each is independently Lorentz invariant),
dl dPS (k , p , p ) = dPS (k , p) dPS (p , p ) (3.26) 3 2 1 2 2π 2 2 2 1 2
with dΩ dPS (x , x ) = β(x , x ) (3.27) 2 1 2 1 2 32π2 which is frame independent, and s 2 2 2 2 2 2(x1 + x2) (x1 x2) β(x1, x2) = 1 2 + − 4 . (3.28) − (x1 + x2) (x1 + x2)
88 The dPS2(k2, p), in the CM frame, can be written as
dΩ s l dΩ dPS (k , p) = β(k , p) = − , (3.29) 2 2 2 32π2 2s 16π2 and dΩ d cos θdφ 1 s 2 = = dt , (3.30) 16π2 16π2 8π s + Q2 s l − which is Lorentz invariant, where the first step uses the azimuthal symmetry of the system in the CM frame, and the second step uses Eq. (3.19).
The dPS2(p1, p2), in the rest frame of p, can be written as
dΩ00 dPS (p , p ) = β(p , p ) (3.31) 2 1 2 1 2 32π2 where β(p1, p2) can be derived using Eq. (3.28), i.e., r 2(m2 + m2) (m2 m2)2 β(p , p ) β(l) = 1 1 2 + 1 − 2 , (3.32) 1 2 ≡ − l l2
i.e., Eq. (3.9), and dΩ00 = d cos θ00dφ00 is the solid angle with respect to the photon in the
rest frame of p (c.f. Eq. (3.24)).
T/L 2 Putting above together, we get the phase space of the off-shell cross section, σνγ (ˆs, Q ), for the trident production,
1 1 dl dt dΩ00 1 1 dl dt dΩ00 dPS3 = 2 β(l) 2 , or dPS3 = 2 β(l) . (3.33) 2 (4π) 2π 2(sνγ + Q ) 4π 2 (4π) 2π 2ˆs 4π
This is the same as the phase space for the real photon case, Eq. (3.8), but replac- ing the s bys ˆ s + Q2. The integration range of l is now (m + m )2, s or νγ ≡ νγ 1 2 νγ (m + m )2, sˆ Q2. And the integration range of t can be obtain from Eq. (3.19), which 1 2 − gives
l l l + Q2, s + 2 Q2 , or l + Q2, sˆ Q2 + 2 Q2 . (3.34) νγ − s − − sˆ Q2 νγ − When Q2 = 0, all the above return to the on-shell photon case.
89 3.8.4 Trident production: coherent and diffractive cross sections for all channels
Figure 3.14 shows our elastic (coherent + diffractive components) cross sections for all trident channels.
39 Coherent + diffractive only 10 + e e W W + W + + 40 , , e e 10 + e, e, + e, e, +
] e e
1 + 41 e e
V +
e 10 ee
G +
2 +
m ee c
+ [
42 + e ee e E 10 / + A +
43 10
44 10 1 2 3 4 5 6 7 8 10 10 10 10 10 10 10 10 E [ GeV ]
Figure 3.14: Our elastic cross sections for all trident channels. We add the coherent and diffractive components together to simplify the figure. The colors and line styles are same as in Fig. 3.4( red, green, and blue lines are νe-, νµ-, and ντ -induced channels, respectively; solid lines are CC channels, and dashed lines are CC+NC channels; magenta dotted lines are NC channels, which depend on only the final-state charged leptons). The trident CC, NC, and CC+NC channels correspond to diagrams (1)–(3), (4)–(5), and (1)–(5) of Fig. 3.5. Gray dashed lines are the coherent and diffractive cross sections for W -boson production from Fig. 3.6, shown as a comparison. For antineutrinos, which have the same corresponding cross sections, take the CP transformation of the channel labels. See text for details.
90 Acknowledgments We are grateful for helpful discussions with Brian Batell, Eric Braaten, Mauricio Bus- tamante, Richard Furnstahl, Bin Guo, Liping He, Matheus Hostert, Junichiro Kawamura,
Matthew Kistler, Yuri Kovchegov, Gordan Krnjaic, Shirley Li, Pedro Machado, Aneesh
Manohar, Kenny Ng, Alexander Pukhov, Stuart Raby, Subir Sarkar, Juri Smirnov, Carl
Schmidt, Xilin Zhang, and especially Spencer Klein, Olivier Mattelaer, Sergio Palomares-
Ruiz, Yuber Perez-Gonzalez, Ryan Plestid, Mary Hall Reno, David Seckel, and Keping
Xie.
We used FeynCalc [292, 293], MadGraph [289] and CalcHEP [290] for some calculations.
This work was supported by NSF grant PHY-1714479 to JFB. BZ was also supported in part by a University Fellowship from The Ohio State University.
91 Chapter 4 W -boson and trident production in TeV–PeV neutrino observatories
Detecting TeV–PeV cosmic neutrinos provides crucial tests of neutrino physics and astrophysics. The statistics of IceCube and the larger proposed IceCube- Gen2 demand calculations of neutrino-nucleus interactions subdominant to deep-inelastic scattering, which is mediated by weak-boson couplings to nuclei. The largest such interactions are W -boson and trident production, which are mediated instead through photon couplings to nuclei. In a companion paper [2], we make the most comprehensive and precise calculations of those interactions at high energies. In this paper, we study their phenomenological consequences. We find that: (1) These interactions are dominated by the production of on-shell W -bosons, which carry most of the neutrino energy, (2) The cross section on water/iron can be as large as 7.5%/14% that of charged-current deep-inelastic scattering, much larger than the quoted uncertainty on the latter, (3) Attenuation in Earth is increased by as much as 15%, (4) W -boson production on nuclei exceeds that through the Glashow resonance on electrons by a factor of 20 for the best-fit IceCube spectrum, (5) The primary signals ' are showers that will significantly affect the detection rate in IceCube-Gen2; a small fraction of events give unique signatures that may be detected sooner.
The contents of this chapter were published in [1].
4.1 Introduction
The recent detections of TeV–PeV neutrinos by IceCube [28, 44, 117, 243, 259, 260] are a breakthrough in neutrino astrophysics. Though the sources of the diffuse flux have not been identified, important constraints on their properties have been determined [248, 249,
92 251, 256, 315–325]. In addition, there is a candidate source detection in association with a blazar flare [55, 326]. The IceCube detections are also a breakthrough in neutrino physics.
By comparing the observed spectra of events that have traveled through Earth or not, the cross section can be measured at energies far above the reach of laboratory experiments [70,
71, 237, 244–246]. And many models of new physics have been powerfully limited by the
IceCube data [78–88, 112–114].
With new detectors — KM3NeT [258], Baikal-GVD [327], and especially the proposed
IceCube-Gen2 (about 10 times bigger than IceCube) [115] — the discovery prospects will be greatly increased, due to improvements in statistics, energy range, and flavor information.
At high energies, neutrino-nucleus interactions are dominated by deep inelastic scattering
(DIS) mediated by weak-boson couplings to nuclei [231, 233]. For charged-current (CC) interactions, νe leads to a shower, νµ leads to a shower and a long muon track, and ντ leads to two showers that begin to separate spatially at 100 TeV [117]. For neutral-current ∼ (NC) interactions of all flavors, showers are produced. Cherenkov light is produced by muon tracks and through the production of numerous low-energy electrons and positrons in showers.
With these coming improved detection prospects, new questions can be asked, including the role of subdominant interactions. We focus on those in which the coupling to the nucleus and its constituents is through a virtual photon, γ∗, instead of a weak boson [38–
43, 118, 119, 226, 266, 269, 272, 273, 276, 277, 280, 281]. The most important processes are
on-shell W -boson production, in which the underlying interaction is ν + γ∗ ` + W , and ` → trident production, in which it is ν + γ∗ ν + `− + `+. → 1 2 In a companion paper [2], we make the most comprehensive and precise calculations
of these cross sections at high energies. The cross section of W -boson production can
be as large as 7.5% of the DIS cross section for water/ice targets (and as large as 14%
for iron targets, relevant for neutrino propagation through Earth’s core) [2]. For trident
production, the most important channels are a subset of W -boson production followed by
leptonic decays [2]. To set a scale, IceCube has identified 60 events above 60 TeV in 7.5
years of operation [259, 260], so taking these subdominant processes into account will be
93 30 10 Glashow resonance ( e) 31 10 ]
2 32
m 10
c CCDIS
[
NCDIS 33 10 rod. on p ) -bos , / W , / 34 ( e/ e 10
35 10 4 5 6 7 8 10 10 10 10 10 E [ GeV ]
Figure 4.1: Cross sections between neutrinos and 16O for W -boson production [2], compared to those for CCDIS [24], NCDIS [24], and the Glashow resonance (¯ν e− W −, taking into e → account eight electrons) [25].
essential for IceCube-Gen2. Moreover, the W -boson and trident events have complex final states, which may allow their detection even sooner, in IceCube.
In this paper, we detail the phenomenological consequences of these processes. In
Sec. 4.2, we focus on their cross sections. In Sec. 4.3, we focus on their detectability.
We conclude in Sec. 4.4.
94 4.2 W -boson production cross sections and implications
In this section, we briefly review the total cross section for W -boson production (Sec. 4.2.1; details are given in our companion paper [2]), and present new calculations of the differential cross sections (Sec. 4.2.2). Then we talk about the implications, including the cross section uncertainties (Sec. 4.2.3) and the effects on neutrino attenuation in Earth (Sec. 4.2.4).
4.2.1 Review of the total cross sections
The nuclear production processes for on-shell W bosons are
ν + A `− + W + + A0 , (4.1a) ` → ν¯ + A `+ + W − + A0 , (4.1b) ` → where A and A0 are the initial and final-state nuclei and ` is a charged lepton. The neutrino- and antineutrino-induced processes have the same total and differential cross sections, but there is flavor dependence. The coupling to the nucleus and its constituents is through a virtual photon, γ∗ (contributions from virtual W and Z bosons are only important for
E > 108 GeV [2]). The process has a high threshold, E 5 103 GeV, due to the large ν ν ' × mass of the W boson, though much lower than the threshold for the Glashow resonance, which peaks at 6.3 PeV. Above threshold, the leptonic decays of the W boson (branching ' ratio 11% to each flavor) lead to the dominant contributions to trident production. ' The interactions happen in three different scattering regimes — coherent, diffractive, and inelastic — in which the virtual photon couples to the whole nucleus, a nucleon, and a quark, respectively. The corresponding cross sections are calculated separately and added to give the total cross section. For the coherent and diffractive regimes, we deal with the hadronic part in a complete way, which takes into account the photon virtuality, instead of using the equivalent photon approximation (as in, e.g., Refs. [39, 40]). Moreover, in the diffractive regime, we include the Pauli-blocking effects that reduce the cross section [42, 43, 270, 276].
For the inelastic regime, we point out that there are two subprocesses: photon-initiated and quark-initiated. For the former, we use the up-to-date inelastic photon PDF of proton and
95 32 10
34 6 E = 106 GeV 10 E = 10 GeV 33 10 e ] ]
5 GeV
2 E = 10 2 m m
c e c
e [ [
34
5 GeV 10
E = 10 W 35 E E
d 10 d / / d d e W 35 E E 10
36 10 36 10 3 2 1 0 1 2 3 4 5 6 4 5 6 10 10 10 10 10 10 10 10 10 10 10 10 10 E [ GeV ] EW [ GeV ]
Figure 4.2: Left: Differential cross sections for W -boson production in terms of the energy 5 of the charged lepton, shown for each neutrino flavor and two typical energies (Eν = 10 GeV 6 −1 and 10 GeV). The y axis is Edσ/dE = (2.3) dσ/d log10 E, matching the log scale on the x axis, so that relative heights of the curves at different energies faithfully show relative contributions to the total cross section. Right: Same, in terms of the energy of the W boson.
neutron [267, 303] and dynamical factorization and renormalization scales. For the latter, we do the first calculation and find that this sub-process can be neglected below 108 GeV. ' A key result is that our W -boson production cross section is smaller than that of previous work [38–40].
Figure 4.1 shows our W -boson production cross sections on 16O for different neutrino
flavors, along with other relevant processes. The width of the Glashow resonance is due to the intrinsic decay width of the W boson.
4.2.2 New results for the differential cross sections
For the differential cross sections, the most relevant results to detection are the energy distributions of the charged lepton (E`) and the W boson (EW ). The energy that goes to the hadronic part is negligible (see next paragraph). As above, the differential cross sections are calculated separately for the three regimes and summed. For the coherent and diffractive regimes, the phase-space variables we chose to calculate the total cross section in
96 Ref. [2] are not directly related to the energies of the final states, so some transformations are needed; see Appendix 4.5.1. For the inelastic regime, following Ref. [2], we use MadGraph
(v2.6.4) [289] and analyze the event distributions in terms of the relevant quantities.
2 The energy that goes to the hadronic part, ∆Eh = Q /2mh (Appendix 4.5.1), is neg- ligible compared to the detection threshold, which is 100 GeV for showers in IceCube. ∼ Here Q2 q2 is the photon virtuality; the hadronic mass, m , is the nuclear mass in ≡ − h the coherent regime and the nucleon mass in the diffractive and inelastic regimes. For the coherent and diffractive regimes, the corresponding nuclear and nucleon form factors are highly suppressed above Q 0.1 GeV and Q 1 GeV, respectively, which leads to ∼ ∼ ∆E (0.1)2/2/16 0.0003 GeV and ∆E (1)2/2/1 0.5 GeV. For the inelastic h . ' h . ' regime, although Q2 could be much larger, the cross section is still dominated by the low-
Q2 region (Q2 10 GeV2, i.e., ∆E 10/2/1 5 GeV) because the nonperturbative part . h . ' of the inelastic photon PDF [267, 300] dominates the cross section (see Sec. V.B of Ref. [2]).
Above is very different the DIS, in which the energy transferred to the nucleus is 25%Eν on average [24, 231, 233, 237].
Therefore, E E + E is an excellent approximation for the coherent and diffractive ν ' ` W regimes and a good approximation for the inelastic regime. We checked this through the distribution of the sum of EW and E` , finding that this is nearly a delta function at Eν. Figure 4.2 shows the differential cross sections for the charged lepton (left) and W
5 6 boson (right), for each neutrino flavor and two typical energies, Eν = 10 and 10 GeV, summed over the contributions from all three scattering regimes. For the charged lepton, the differential cross section is relatively flat when plotted as Edσ/dE, which means that no specific energy range is particularly favored; the narrow bump near Eν does not contribute significantly to the total cross section. For the W boson, the differential cross section favors the highest possible energy, E E m . The differences between different flavors are W ' ν − ` due only to the charged lepton mass, m`, which sets the lower limit of the distribution. Therefore, when E m , the results for different flavors converge. This induces the ` ` opposite feature for the distribution of W boson, where the results for different flavors converge at lower energies.
97 Figure 4.3 shows the average energy for the charged lepton ` and the W boson for each
flavor of initial neutrino. This is calculated as
R dσ E 1 dE E (E,Eν) h i = dE , (4.2) Eν Eν σ(Eν) where E = E` or EW . As can be expected from Fig. 4.2, the W boson is typically much more energetic than the charged lepton, except for when Eν is very large. In a crude approximation, at the main neutrino energies relevant to detection in IceCube or IceCube-
Gen2, all the neutrino energy goes to the W boson, with some dependence on neutrino
flavor. Below, we provide more careful calculations.
4.2.3 Implication: Cross-section uncertainty
Figure 4.4 shows the ratios of the W -boson production to the neutrino CCDIS cross sections
CCDIS CCDIS ((σν + σν¯ )/2) [24]. We neglect the NCDIS cross section because it is smaller (see Fig. 4.1) and because the energy deposition is only 0.25E , which suppresses its ' ν importance [328]. What we show is most relevant for detection. For a water/ice target, the maximum ratios of W -boson production to CCDIS [24] are 7.5% (ν ), 5% (ν ), ' e ' µ and 3.5% (ν ). For an iron target or Earth’s average composition, the maximum ratios ' τ are 14%/11% (νe), 10%/7.5% (νµ), and 7%/5% (ντ ). This is more relevant to propagation (affected by both CCDIS and NCDIS) than detection (dominated by CCDIS); see Sec. 4.2.4 for details. The larger the charge number of a nucleus, the larger the ratio is. The coherent component is Z2, while the diffractive and inelastic components are Z, the same as for ∝ ∝ DIS. As noted, our results are significantly smaller than those of Seckel [38].
For the CCDIS cross section, the claimed uncertainties (from the parton-distribution functions) in 104–108 GeV are 1.5–4.5% in Ref. [24] and 1–6% in Ref. [237] (see also
Refs. [239, 307]). The impact of W -boson production is thus significant and thus should be included in future calculations of neutrino-nucleus cross sections. Further, as IceCube has detected 60 events above 60 TeV (deposited energy) in the past 7.5 years [259, 260], this means that taking W -boson production into account is relevant for IceCube and essential for IceCube-Gen2, which would be about 10 times larger. This is detailed in Sec. 4.3.
98 1.0 0.9 ( ) 0.8 W e W ( ) 0.7 ( ) 0.6 W E
/ 0.5 E 0.4 ( ) 0.3 ( ) 0.2 ( e) 0.1
0.0 4 5 6 7 8 10 10 10 10 10 E [ GeV ]
Figure 4.3: Average energy of the charged lepton (`) and W boson, divided by Eν, for each neutrino flavor.
For future calculations of the cross-section uncertainties, aiming to reach the few-percent scale, we note some other corrections that should be taken into account. The DIS calcu- lations are done at next-to-leading order in QCD, using corresponding parton distribu- tion functions. However, as far as we are aware, next-to-leading order electroweak cor- rections [329–333] are not included. At the highest energies, nonperturbative electroweak cascades [334] may be a significant effect. Finally, there are other processes, such as tri-
99 charged lepton production [335–338], that become increasingly important at high energies.
In addition, in Ref. [339], Klein notes that going beyond assuming isoscalar nucleon targets and nuclear effects on the parton distribution function should also be considered.
0.25 + e + A e W A + A W + A + A W + A 0.20 Seckel e (iron)
0.15 Our e (iron)
0.10
0.05 Ratio to CCDIS cross section
0.00 4 5 6 7 8 10 10 10 10 10 E [ GeV ]
Figure 4.4: Ratios of the W -boson production cross sections [2] to those of CCDIS ((ν + ν¯)/2) [24]. Solid lines are for water/ice targets, dotted for iron targets, and dashed for Earth’s average composition. Color assignments are noted in the legend. For comparison, we also show the νe (iron) result of Seckel [38], which is much larger than ours.
100 4.2.4 Implication: Attenuation in Earth
Starting in the TeV range, neutrinos may be significantly attenuated while passing through
Earth. (For a path along an Earth diameter, τ = 1 at E 40 TeV.) Attenuation depends ν ' on the total CCDIS + NCDIS cross section, σ(Eν). Taking into account NCDIS increases the cross section by a factor 1.4 compared to CCDIS only [24, 231, 233, 237]. We ignore ' neutrino regeneration because of the steeply falling neutrino spectra.
The optical depth τ = Cσ, where C(cos θz) is the target number column density in- tegrated along the line of sight, which depends on the zenith angle, θz. We use Earth’s average composition. The flux is attenuated by a factor
A = e−τ(Eν , cos θz). (4.3)
The column density in the direction of the zenith angle θz is reasonably well known [340]. Though the change in the cross section due to W -boson production is not large, it affects the argument of the exponential. For W -boson production, we calculate τ as the sum of results for the three regimes, taking into account that the targets in the coherent regime are nuclei, while in the diffractive and inelastic regimes they are nucleons.
Figure 4.5 (upper panel) shows the neutrino attenuation factor without (ADIS) and with
DIS+WBP (A ) W -boson production. For simplicity, we consider only νe, which has the largest such cross section (see Fig. 4.1).
Figure 4.5 (lower panel) shows the relative change to the attenuation factor,
1 ADIS+WBP/ADIS. For high energies and long paths through Earth, this can be − quite large, even though the W -boson production cross section is small compared to the CCDIS cross section. However, for A that is too small, the event rate would be too low to matter; accordingly, we use thin lines for where A 0.1. Even ≤ avoiding these regions, the change in A can be as large as 15%. This follows from
1 ADIS+WBP/ADIS = 1 exp( Cσ ) Cσ σ /σ τ σ /σ , which − − − WBP ' DIS × WBP DIS ' × WBP DIS is the multiplication of the optical depth and the cross section ratio, which 10%/1.4 7% ' ' for νe (from Fig. 4.4). The factor 1.4 roughly accounts for including NCDIS in addition to
101 CCDIS. For example, when A = 0.1, i.e., τ 2.3, 1 ADIS+WBP/ADIS 2.3 7% 15%, ' − ' × ' as above.
Interestingly, in IceCube’s through-going muon analysis, there was an unknown 2% deficit of straight up-going events, compared with their Monte Carlo simulation [341]. Tak- ing W -boson production into account may explain this deficit.
The Earth attenuation effect allows a measurement of the neutrino-nucleus cross section at TeV–PeV energies in IceCube [70, 71, 237, 244–246]. The energy scales probed are far above those of laboratory experiments, for which the highest beam energies are 350 ' GeV [240–242]. In essence, the downgoing neutrino event rate depends on φσ, while the upgoing event rate depends on φσe−τ , and taking a ratio cancels the flux and the detection cross section.
In present IceCube measurements of the cross section, the uncertainty is 35% when ' only the cross section normalization is checked (one wide energy bin for all data, plus assuming the shape of the standard model cross section) [70] or a factor of 4 when these ' assumptions are relaxed (several energy bins, no prior on the cross section shape) [71].
The energy ranges of both are comparable to where W -boson production is important.
However, the measurement uncertainties will decrease. In addition, in Ref. [70] the ratio of the measured cross section to DIS prediction is 1.3 0.45. The central value would be ± about 0.1 smaller if the contribution from W -boson production were included.
Last, attenuation effects also lead to slightly altered flavor ratios because the W -boson production cross sections are flavor dependent.
4.3 Detectability
In this section, we calculate the detection prospects. We focus on W -boson production.
The most important channels of trident production are a subset of W -boson production followed by leptonic decays [2]. We first calculate the W -boson yields compared to those through the Glashow resonance (Sec. 4.3.1). Then, after a brief review of IceCube detection
(Sec. 4.3.2), we calculate the detectability of W -boson production from the shower spectrum
102 0 10
E = 105 GeV
1 6 7 10 10 GeV 10 GeV
Neutrino attenuation factor 2
S 10 I
D 0.2 A / P B W
+ 0.1 S I D A 0.0
1 1.0 0.8 0.6 0.4 0.2 0.0 Neutrino zenith angle, cos z
−σC DIS Figure 4.5: Upper: Neutrino attenuation factor, e , for νe in Earth. Dashed lines (A ) are for CCDIS and NCDIS without W -boson production. Solid lines (ADIS+WBP) include W -boson production. For attenuation factors below 0.1, the event rate is too low to use, which we denote by using thin lines. Lower: The relative change in the attenuation factor due to W -boson production.
(Sec. 4.3.3) and from unique signatures (Sec. 4.3.4).
103 4.3.1 Larger W -boson yields than Glashow resonance
The Glashow resonance (¯ν +e− W −) is well known for producing on-shell W bosons with e → a narrow feature in the cross section around Eν = 6.3 PeV. The maximum cross section is 10−30 cm2, a factor of about 100 larger than that of DIS. Once the intrinsic and detector ∼ energy resolution are taken into account, the effects on the total event spectrum are less
dramatic but still important.
Surprisingly, we find that on-shell W -boson production is actually dominated by neutrino-nucleus interactions where the coupling to the nucleus is through a photon. The cross section is much smaller, but it involves all six neutrino flavors, and acts over a much wider energy range, in particular at lower energies, where the neutrino fluxes are much larger.
Figure 4.6 illustrates this. We multiply the cross sections by a power-law flux,
1−α −2 −1 −1 Eν dΦ/dEν = (Eν/1 GeV) cm s GeV , with unit normalization, and plot results versus neutrino energy. We use α = 2.9, which matches the astrophysical neutrino spectrum from fitting IceCube data [28, 260]. (Below, we calculate more realistic expectations for detection.) We multiply the flux by a factor Eν so that the relative heights on the y axis faithfully display the relative numbers of events per logarithmic energy bin.
The yield from neutrino-nucleus W -boson production is a factor of 20 times that for ' the Glashow resonance. For α = 2.5 and 2.0, the factor is 3.5 and 0.5 respectively. ' ' Therefore, for TeV–PeV neutrino observatories, neutrino-nucleus W -boson production is the dominant source of on-shell W bosons unless the spectrum is very hard. This is a new and interesting physics point. When it comes to detection, the W bosons are not detected directly, due to their short lifetimes, and are instead detected by their decay products.
4.3.2 Review of detection in IceCube
We briefly summarize the neutrino detection techniques used in IceCube and similar detec- tors [44, 116, 243, 342]. Neutrinos interact with nuclei and electrons, producing relativistic particles that emit Cherenkov light that is detected by photomultiplier tubes.
104 42 10
] Glashow resonance ( )
e 1 s
[
E d / 43
d 10 E × ) W-boson production E (
A (sum of 6 flavors) 44 10
3 4 5 6 7 8 10 10 10 10 10 10 E [ GeV ]
Figure 4.6: Relative W -boson yields due to W -boson production (νl + A l + 0 − − → W + A ) and the Glashow resonance (¯νe + e W ). We use dΦ/dEν = −2.9 −2 −1 −1 → (Eν/1 GeV) cm s GeV with unit normalization. The yield from W -boson pro- duction is 20 times that from the Glashow resonance, which can be seen by logarithmic ' integration of the peaks. The CCDIS (cyan, dashed) and NCDIS (magenta, dashed) cases are shown for comparison, though they do not produce on-shell W bosons.
A νe CCDIS event produces an electron that carries most of the neutrino energy and hadrons that carry the remainder. The electron initiates an electromagnetic shower of electrons, positrons, and gamma rays, with most of the Cherenkov emission coming from the most numerous low-energy but still relativistic charged particles. The hadrons initiate a hadronic shower that consists primarily of pions. The charged pions continue the hadronic
105 3 Conservative case 3 Optimistic case 10 10 C C 2 D 2 10 IS 10 , p p
e NCDIS, sume , of 6 flavors e NCDIS, sum of 6 flavors d d 1 e a 1 E nd E d 10 d 10 / / , N N
d 0 Glashow d 0 Glashow p p
e 10 e 10
d resonance d resonance
E W-boson production E W-boson production 1 1 10 (sum of 6 flavors) 10 (sum of 6 flavors)
2 2 10 10 0 0 10 10 B B 1 1
/ 10 / 10 S S
2 2 10 3 4 5 6 7 8 10 3 4 5 6 7 8 10 10 10 10 10 10 10 10 10 10 10 10 Edep [ GeV ] Edep [ GeV ]
Figure 4.7: Left: Shower spectrum (Upper) and detection significance of W -boson produc- tion (Lower) for the conservative case as regards identifying W -boson production events. Right: Same, but for the optimistic case. The main difference between the two cases is the change with the CCDIS channel. The shaded region below 60 TeV is below the IceCube threshold for cleanly identifying astrophysical neutrinos. We use 0.5 km3 as the approximate fiducial volume of IceCube [44], and assume 10 years of IceCube data. For E > 60 TeV, where W -boson production contributes 6 shower events, the cumulative dep ' detection significance should be 1.0σ for the conservative case and 3.2σ for the optimistic ' case. See text for details.
shower, but the neutral pions decay promptly, feeding the electromagnetic shower. The shower components induced by a νe event have high and comparable light yields, so that the total amount of Cherenkov light is proportional to Eν. Because of the light scattering in ice, a shower looks like a large ( 100 m), round blob, even though the shower is a narrow ∼ cigar-shaped blob of length 10 m. Forν ¯ , the total and differential cross sections are ∼ e slightly different, but the detection principles are the same.
For ντ (andν ¯τ ) CCDIS events, the results can be very similar to those for νe (andν ¯e). The τ decay produces a hadronic shower that is displaced in time and position, though
these displacements start to become identifiable in IceCube only above 100 TeV [117]. ∼ At lower energies, ντ events are nominally indistinguishable from νe events. A way forward
could be possible using muon and neutron echoes [116]. For ντ (andν ¯τ ) events, the average
106 deposited energy is 20% less than E due to losses of neutrinos from τ leptonic decays. ' ν In addition, 17% of τ decays produce muon tracks, producing separable events.
All six flavors of neutrinos cause NCDIS events that also produce showers. These appear identical to the other shower events above, though the energy deposition is typically only
0.25E . Because of the falling neutrino spectra, these events matter much less in the ' ν detection spectra [328].
For νµ (andν ¯µ) CCDIS events, the topologies are quite different because the muon range is so long, already > 1 km at E 200 GeV. For events where the neutrino interaction µ ' is inside the detector (known as a contained-vertex or starting event), there is a hadronic
shower and a long muon track, which itself produces small showers along its length. Though
the muon is not contained, its energy can be estimated from its energy-loss fluctuations, so
that the neutrino energy can be estimated. There can also be events where the neutrino
interacts far outside the detector, and only the throughgoing muon is detected. This enlarges
the effective volume of the detector, but then only a lower limit on the neutrino energy can
be set.
Shower events are especially important because of the ability to faithfully reconstruct
the neutrino spectrum. The shower spectrum can be estimated as [87, 238, 315]
Z 1 dN ρiceVfidNA dΦ −τ(Eν , cos θz) Edep = 2π T d cos θz Eν (Eν) σ(Eν) e , (4.4) dEdep 18 −1 × dEν where E is the energy deposited in the detector from a shower, ρ 0.92 g cm−3 the dep ice ' 3 density of ice, Vfid = 0.5 km is the approximate fiducial volume of IceCube [44], NA the Avo- gadro number (ρiceVfidNA/18 gives the number of water targets), and T is the exposure time. For the neutrino flux, dΦ , we use that of Ref. [28], which includes both the atmospheric dEν
(dominated by νµ andν ¯µ) and astrophysical neutrino fluxes. The best-fit astrophysical flux, assuming 1:1:1 flavor ratios, is (2.46 0.8) 10−18(E/100 TeV)−2.92 GeV−1 cm−2 s−1 sr−1 ± × for each flavor (ν +ν ¯), which is consistent with a more recent result [260]. The σ is the
cross section between neutrino and water for different interaction channels. For CCDIS
and NCDIS, we multiply the cross section on isoscalar nucleon targets [24] by 18, the mass
number for water. For the Glashow resonance, we multiply the cross section on electrons by
107 10, the charge number for water. For the attenuation factor, e−τ(Eν , cos θz), we use Ref. [343] with modification to include the cross section for W -boson production. Once we obtain the shower spectra, we convolve them with a detector energy resolution of 15% [344].
4.3.3 Total shower detection spectrum
For the detection of W -boson production events, we consider two general scenarios. In this subsection, we consider final states that contribute to the overall shower spectrum.
(The prospects for detection via track events are not favorable.) In the next subsection, we consider unique final states that can be individually identified. We focus on IceCube [243], the largest detector for TeV–PeV neutrinos. Our results can be scaled to the proposed
IceCube-Gen2, whose instrumented volume is expected to be 10 times that of IceCube [115], and fiducial volume may be more, though with a higher energy threshold.
Table 4.1 shows the complex possibilities for W -boson production events, including pure shower, track, and other unique signatures, depending on the decay modes of the W boson and τ leptons. For the charged leptons from the initial interactions, their detectability de- pends on IceCube’s trigger threshold ( 100 GeV; note the analysis threshold of IceCube, ' 1 TeV, is less relevant because one would be searching for a lower-energy event in asso- ' ciation with a higher-energy event) [116]. Fig. 4.2 (left) is, up to an overall factor 2.3−1, the probability distribution in log E , and it is roughly flat, with a median 100 GeV. 10 ` ∼ Therefore, we assume that half of the primary leptons are detectable and half not. In the
“Signatures” column of Table 4.1, we use “/” to distinguish the two cases. (The decaying
W ’s are always detectable.)
First, we calculate the change to the overall shower spectrum in the conservative case where ντ events appear as showers and where all showers are indistinguishable. We ignore events with an energetic muon track: ν CCDIS, ν CCDIS with τ µ, half of ν -induced µ τ → µ W -boson production, ν - and ν -induced W -boson production with W µ, τ µ, or e τ → → W τ µ, and Glashow resonance events with W µ or W τ µ. → → → → → Second, we calculate for an optimistic case where ντ CCDIS events are identifiable through a double-bang or double-pulse signature (above 100 TeV, this is becoming real- ∼ 108 istic with current technology [117]), and where electromagnetic and hadronic showers can be separated using echo techniques (this is not yet possible with IceCube, but it may be with IceCube-Gen2). Therefore, for CCDIS, we remove the remaining νe- and ντ -induced channels. For the W -boson production, we remove the channels that give tracks, pure EM showers (EM means electromagnetic), and other unique signatures (see Table 4.1).
For ν CC, E E , as both the final state electron and hadrons produce showers. e dep ' ν For ν CC, E [ y + 0.7(1 y )] E 0.8E , where y 0.25, for both CCDIS and τ dep ' h i − h i ν ' ν h i ' NCDIS, is the average inelasticity, which is the fraction of neutrino energy transferred to the hadrons [231]. The factor 0.7 above is due to about 30% of the energy is taken away by neutrinos from τ decay. For all-flavor NCDIS, E y E 0.25E . The ratios for dep ' h i ν ' ν CCDIS and NCDIS above are the similar to those used in Refs. [87, 116, 238].
For W -boson production, we use E E if the W decays hadronically and E dep ' ν dep ' 0.5Eν if the W decays leptonically. At the relevant energies, the W boson takes nearly all of the neutrino energy; even when it does not, the approximations here are good for the total energy deposition, because the charged lepton (e or τ) from the initial interaction deposits most of its energy. We make the same assumptions for W bosons produced via the Glashow resonance.
Figure 4.7 shows the total shower spectrum for the conservative (left) and optimistic
(right) cases for IceCube observations with T = 10 years (or 1 year of IceCube-Gen2 [115]).
W -boson production is subdominant, especially in the conservative case. The CCDIS events dominate, due to the large cross section and energy deposition. The NCDIS events are reduced in importance due the small energy deposition. The two-peaks feature of the
Glashow resonance is due to the leptonic and hadronic decays of W bosons.
The lower panels show the detection significance. For each bin, this is calculated by the number of W -boson production events divided by the square root of CCDIS+NCDIS events. The cumulative significance for detecting W -boson production, combining all the bins above E = 60 TeV, is 1.0σ for the conservative case and 3.2σ for the optimistic dep ' case, for 10 years of IceCube observations. The Glashow resonance events are not included because doing so would not appreciably affect the results.
109 Table 4.1: Different final state particles, signatures, corresponding fractions, and counts in IceCube. The counts are for greater than 60 TeV deposited energy and 10 years of IceCube observations (or 1 year for IceCube-Gen2). The numbers in the “Channel” column are the maximal ratios to the CCDIS cross section with water/ice. The numbers in “W decay” and “τ decay” columns are the branching ratios. For the “Final state” and “τ decay” columns, we omit the neutrinos; “h” means hadrons. The unique signatures are in boldface. The “/” divides the cases in which the charged lepton from the initial interaction is undetectable or detectable, which, to a good approximation, is half-half. The “Fractions” column shows the fraction of that row relative to the whole channel, which is the multiplication between the branching ratios of W and τ decay.
Final Channel W decay τ decay Signature Fraction Counts state
eνe, 11% e e Pure EM shower 11% 0.34 µνµ, 11% e µ Track without/with shower 11% 0.34 νe → eW e, 18% Pure EM shower 2.0% 0.06 (7.5% rel. τντ , 11% e τ µ, 17% Track without/with (displaced) shower 1.9% 0.06 to CCDIS) h, 65% Shower 7.2% 0.22 qq¯, 67% e h Shower 67% 2.08
eνe, 11% µ e Pure EM shower/Track with shower 11% 0.56 µνµ, 11% µ µ Single/Double tracks without shower 11% 0.56 e, 18% Pure EM shower/Track with (displaced) shower 2.0% 0.10 νµ → µW τντ , 11% µ τ µ, 17% Single/Double tracks without shower 1.9% 0.10 (5.0% rel. h, 65% Shower/Shower with (displaced) track 7.2% 0.36 to CCDIS) qq¯, 67% µ h Shower/Shower with track 67% 3.41 e, 18% Pure EM shower 2.0% 0.02 eνe, 11% τ e µ, 17% Pure EM shower/Track with (displaced) shower 1.9% 0.02 h, 65% Pure EM shower/Shower 7.2% 0.09 µ, 17% Single/Double tracks without shower 1.9% 0.02 µνµ, 11% τ µ e or h, 83% Track without shower/with (displaced) shower 9.1% 0.11 ντ → τW e e, 3% Pure EM shower 0.4% 0.004 (3.5% rel. µ µ, 3% Single/Double tracks without shower 0.3% 0.004 to CCDIS) τντ , 11% τ τ µ e/h, 29% Track without shower/with (displaced) shower 3.1% 0.04 h h/e, 65% Shower/Double bang 7.2% 0.09 e or h, 83% Shower 56% 0.69 qq¯, 67% τ h µ, 17% Shower/Shower with track 11% 0.14 Total counts 9.44
In summary, for Edep > 60 TeV and for 10 years of IceCube observations, W -boson production contributes 6 shower events, and the cumulative detection significance should ' be 1.0σ for the conservative case and 3.2σ for the optimistic case. With 10 years of ' IceCube-Gen2, the counts would improve by a factor of 10 and the significances by a ' factor √10 3.2. '
110 4.3.4 Unique signatures
Table 4.1 also shows the unique signatures (in boldface) that W -boson production could give in IceCube, including fractions and counts. The events that give unique signatures are from leptonic decays following W -boson production, therefore they are also trident events. The counts are calculated using Eq. (4.4). Some of them are background free compared to other standard model processes (DIS and the Glashow resonance). We focus on IceCube’s high- energy analysis, so the counts are for greater than 60 TeV deposited energy and also for 10 years of IceCube observations. For lower-energy analysis of IceCube, such as the medium- energy starting events [345] and the Enhanced Starting Track Event Selection [346], these unique signatures could also exist.
There could be “Double tracks” signatures, with one muon track from the initial inter- action (µ or τ µ) and the other from the decay of W boson. See Table 4.1 for all the → contributing channels. The calculated counts are 0.34. Double tracks could also come ' from outside the detector, which would increase the counts. Some of the “Double tracks” events may have separation angles (θ) too small for them to be distinguished from a single track [273]. According to Refs. [347, 348], IceCube’s resolution for double tracks is about
150 m. Therefore, for the “Double tracks” traveling 1 km, as long as the cos θ . 0.99, they can be separated. Moreover, the “Double tracks” would be a background for dimuon-type new physics searches [348–350].
These “Double tracks” may also be identified because they are a subset of “track without shower” signatures. The no-shower feature is because the energy transferred to the hadronic part in W -boson production is mostly negligible (Sec. 4.2.2). The “track without shower” signatures are mostly background free, because in CCDIS the energy transferred to hadronic
5 part, for Eν > 10 GeV, is mostly above the IceCube threshold (100 GeV), according to the dσCCDIS/dy of Ref. [231]. Moreover, these signatures could also come from other channels of W -boson production (see Table 4.1 for details). The calculated total counts are 0.96. ' Interestingly, there is a “track without shower” candidate (Event 5) in the IceCube event list in Ref. [44] (arXiv version, page 15). It has no obvious shower activity, while all
111 seven other track events have prominent showers at their starting points. It is important to quantify the probability of this event coming from CCDIS. Event 5 deposited 71.4 TeV energy in IceCube, so the neutrino energy is 105 GeV. We can require that the hadronic ∼ energy be smaller than the full energy lost by muon in the initial 100 m of its path length.
This is conservative, as it would double the average energy deposited in the first 100 m of the muon track compared to the second 100 m, which would be visible, unlike for Event 5.
According to Ref. [351], the corresponding energy loss of a 105 GeV muon would be about
5 4 TeV. From dσCCDIS/dy of Ref. [231] for Eν = 10 GeV, we can estimate the probability for having a hadronic energy smaller than a given value. For 100 GeV, it is 0.3%. For 4 ' TeV, it is 10%. Therefore, the probability for Event 5 being induced by CCDIS is small, ' even in conservative cases.
Moreover, “track with shower” events from W -boson production would also look differ- ent from CCDIS in terms of the inelasticity distribution. For CCDIS, the dominant energy,
(1 y)E , goes to the track, with the smaller remainder, yE , going to the hadrons. For − ν ν those W -boson production events that are analyzed as CCDIS events with tracks, the non-
track energy comes primarily from the W -boson decay, and this is typically much larger
than the energy going to the track, in contrast to CCDIS. Therefore, W -boson production
events should be included in theoretical expectations of attempts to better measure the
νµ toν ¯µ flux ratio and neutrino charged-current charm production [352]. This may also provide a way to detect W -boson production.
There could also be “Pure EM shower” signatures, where EM means electromagnetic.
The no-hadronic-shower feature is because the energy transferred to the hadronic part in
W -boson production is mostly negligible (Sec. 4.2.2). The major contributing channels are
W e following W -boson production induced by ν or by other flavors with the initial → e charged lepton below the trigger threshold (see Table 4.1 for details). The calculated counts
are 0.82. This signature could be background free with the echo technique [116], same ' reason as for the “Track without shower” signature.
At last, there could also be the “Track + displaced shower” signatures (Table 4.1). The
calculated counts are 0.35. However, due to the short lifetime of the τ, it may be hard ' 112 to identify them. The ντ CC events with τ decay to muon will be a background. These unique signatures would help flavor identification. For example, the “Double track” and “Track without shower” signatures are dominated by νµ-induced W -boson pro- duction; the “Pure EM shower” are dominated by νe-induced W -boson production, and the “Track with displaced shower” is dominated by ντ -induced W -boson production. The unique signatures are also backgrounds for exotic signals.
4.4 Conclusions
It is time to study the role of subdominant neutrino-nucleus interactions for TeV–PeV neutrinos. These cosmic neutrinos provide essential probes for neutrino astrophysics and physics, and the statistics of IceCube and especially IceCube-Gen2 demand greater precision in the theoretical predictions used to interpret the data.
The most important subdominant processes not yet taken into account are those where the interaction of a neutrino with a nucleus and its constituents is through a virtual photon,
γ∗. These processes are W -boson production (ν + A l + W + A0) and trident production l → (ν + A ν + `− + `+ + A0). In a companion paper, we present the more comprehensive → 1 2 and precise calculations of these cross sections at high energies [2].
In this paper, we study the phenomenological consequences of these processes at TeV–
PeV energies for IceCube and related experiments. We have five major results:
1. These interactions are dominated by the production of on-shell W -bosons, which carry
most of the neutrino energy. The most important trident channels are a subset of W -
boson production followed by leptonic decays. The energy partition follows from the
calculation of the differential cross sections (Fig. 4.2) and the average energies (Fig. 4.3)
of the final states. The lepton takes a modest amount of the neutrino energy and, in
stark contrast to DIS, the hadronic final state takes almost none.
2. The cross section on water/iron can be as large as 7.5%/14% that of charged-current deep-
inelastic scattering, much larger than the quoted uncertainty on the latter. From Fig. 4.4,
the maximum ratios of W -boson production to CCDIS cross sections for water/ice targets
113 are 7.5% (ν ), 5% (ν ), and 3.5% (ν ). For iron targets, these are 14%, 10%, ' e ' µ ' τ ' ' and 7%. These are significantly smaller than the early predictions of Seckel [38]. On ' the other hand, these ratios are much larger than the quoted uncertainties on the deep-
4 8 inelastic scattering cross section for Eν = 10 –10 GeV, which are 1.5–4.5% in Ref. [24] and 1–6% in Ref. [237]. We also point out other corrections to DIS that should be taken
into account for future calculations.
3. Attenuation in Earth is increased by as much as 15% due to these cross sections (Fig. 4.5).
They are also an inseparable part of the measured neutrino cross section. Though the
uncertainty of measured cross sections by IceCube is larger than the change in the cross
section due to W -boson production [70, 71], the measured uncertainties will decrease. In
addition, in Ref. [70] the ratio of the measured cross section to DIS prediction is 1.3 0.45. ± The central value would be about 0.1 smaller if the contribution from W -boson production
were included.
4. W -boson production on nuclei exceeds that through the Glashow resonance on electrons
by a factor of 20. From Fig. 4.6, the former produces on-shell W bosons 20 times ' ' more efficiently than the Glashow resonance if the neutrino spectrum index is 2.9, the
nominal value from fitting IceCube data [28, 260]. This point was not previously known.
5. The primary signals are showers that will significantly affect the detection rate in IceCube-
Gen2; a small fraction of events give unique signatures that may be detected sooner. The
overall shower spectrum is changed by W -boson production. Based on the calculations
in Fig. 4.7, we show that this could be detected with 10 years of IceCube data above 60
TeV with significance 1.0σ and 3.2σ for conservative and optimistic cases. In 10 years
of IceCube-Gen2, these would improve by a factor √10 3.2. We also note unique ' signatures that may be identified sooner, including with IceCube. (Though not explored
here, it would be interesting to consider their impact on detectors for ultra-high-energy
neutrinos.)
Since 2013, IceCube has opened the field of high-energy neutrino astronomy, probing neutrino-nucleus interactions well above 1 PeV, far beyond the reach of laboratory experi-
114 ments. Now, only six years later, it is becoming important to take into account subdominant neutrino-nucleus interactions. This rapid progress hints at the discovery prospects of larger detectors, both for increased precision in probing astrophysics and the cross section as well as in searches for new physics.
4.5 Supplemental Material
4.5.1 dσνA/dE` and dσνA/dEW
In this section, we detail the calculations of the differential cross section for W -boson pro- duction in the coherent and diffractive regimes. For these two regimes, we use the formalism in Refs. [42, 266], and need to deal with the phase space by ourselves. For the inelastic regime, the phase space is handled with MadGraph [289].
In our companion paper [2], we use the center-of-momentum (CM) frame between the neutrino and the virtual photon, which is the most convenient for calculating the total cross section. In this frame, the 4-momentum can be easily written as
s + Q2 s + Q2 k = , 0, 0, , (4.5a) 1 2√s 2√s s Q2 s + Q2 q = − , 0, 0, , (4.5b) 2√s − 2√s
p = (E , 0, p sin θ, p cos θ) , (4.5c) 1 1 − − p = (E , 0, p sin θ, p cos θ) , (4.5d) 2 2 − −
where k1, q, p1, and p2 are the 4-momenta of the neutrino, virtual photon, charged lepton, and W boson, respectively, s = (k + q)2, Q2 = q2, and p = 1 − s (m + m )2 s (m m )2 1/2 /2√s. − l W − l − W However, to get the differential cross sections, dσνA/dE` and dσνA/dEW , we should
115 transform to the lab frame (nucleus-rest frame), in which
k1 = (Eν, 0, 0,Eν) , (4.6a)