ACTA UNIVERSITATIS UPSALIENSIS Uppsala Dissertations from the Faculty of Science and Technology 117
Exploring the Universe Using Neutrinos A Search for Point Sources in the Southern Hemisphere Using the IceCube Neutrino Observatory
Rickard Ström
Dissertation presented at Uppsala University to be publicly examined in Ångströmlaboratoriet, Polhemsalen, Lägerhyddsvägen 1, Uppsala, Friday, 18 December 2015 at 13:15 for the degree of Doctor of Philosophy. The examination will be conducted in English. Faculty examiner: Professor Lawrence R. Sulak (Boston University, Boston, USA).
Abstract Ström, R. 2015. Exploring the Universe Using Neutrinos. A Search for Point Sources in the Southern Hemisphere Using the IceCube Neutrino Observatory. Uppsala Dissertations from the Faculty of Science and Technology 117. 254 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-554-9405-6.
Neutrinos are the ideal cosmic messengers, and can be used to explore the most powerful accelerators in the Universe, in particular the mechanisms for producing and accelerating cosmic rays to incredible energies. By studying clustering of neutrino candidate events in the IceCube detector we can discover sites of hadronic acceleration. We present results on searches for point- like sources of astrophysical neutrinos located in the Southern hemisphere, at energies between 100 GeV and a few TeV. The data were collected during the first year of the completed 86- string detector, corresponding to a detector livetime of 329 days. The event selection focuses on identifying events starting inside the instrumented volume, utilizing several advanced veto techniques, successfully reducing the large background of atmospheric muons. An unbinned maximum likelihood method is used to search for clustering of neutrino-like events. We perform a search in the full Southern hemisphere and a dedicated search using a catalog of 96 pre- defined known gamma-ray emitting sources seen in ground-based telescopes. No evidence of neutrino emission from point-like sources is found. The hottest spot is located at R.A. 305.2° and Dec. -8.5°, with a post-trial p-value of 88.1%. The most significant source in the a priori list is QSO 2022-077 with a post-trial p-value of 14.8%. In the absence of evidence for a signal, we calculate upper limits on the flux of muon-neutrinos for a range of spectra. For an unbroken E-2 neutrino spectrum, the observed limits are between 2.8 and 9.4×10-10 TeV cm-2 s-1, while for an E-2 neutrino spectrum with an exponential cut-off at 10 TeV, the observed limits are between 0.6 and 3.6×10-9 TeV cm-2 s-1.
Keywords: astroparticle physics, neutrino sources, neutrino telescopes, IceCube
Rickard Ström, Department of Physics and Astronomy, High Energy Physics, Box 516, Uppsala University, SE-751 20 Uppsala, Sweden.
© Rickard Ström 2015
ISSN 1104-2516 ISBN 978-91-554-9405-6 urn:nbn:se:uu:diva-265522 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-265522) “The probability of success is difficult to estimate but if we never search the chance of success is zero.”
Cocconi and Morrison [1]
Table of Contents
Page
Acknowledgments ...... 9
About this Thesis ...... 13
Abbreviations and Acronyms ...... 17
1 Introduction ...... 23 1.1 Physics Motivation ...... 24 1.2 The Multi-Messenger Approach ...... 26
2 The Standard Model ...... 29 2.1 Matter Particles and Force Carriers ...... 30 2.2 Interaction Strengths ...... 32 2.3 Symmetries ...... 33 2.4 The Weak Interaction ...... 35 2.5 The Higgs Mechanism ...... 37 2.6 Fermion masses ...... 38 2.7 The Parameters of the Standard Model ...... 39 2.8 Beyond the Standard Model ...... 40
3 The Cosmic Ray Puzzle ...... 47 3.1 Energy Spectrum ...... 50 3.2 The Origin of High-Energy Cosmic Rays ...... 53 3.3 Astrophysical Neutrinos ...... 55 3.4 Atmospheric Backgrounds ...... 60 3.5 Acceleration Mechanisms ...... 63 3.6 Potential Acceleration Sites ...... 67
4 Neutrino Detection Principles ...... 75 4.1 Neutrino Cross-Section ...... 76 4.2 Cherenkov Radiation ...... 78 4.3 Energy Loss Mechanisms ...... 80 4.4 Event Topologies ...... 82
5 The IceCube Neutrino Observatory ...... 85 5.1 The IceCube In-Ice Detector ...... 88 5.2 Optical Properties of the Ice at the South Pole ...... 90 5.3 The Digital Optical Module ...... 93 5.4 Data Acquisition System and Triggering ...... 96 5.5 Processing and Filtering ...... 97
6 Event Simulation ...... 99 6.1 The IceCube Simulation Chain ...... 99 6.2 Neutrino Event Weighting ...... 104
7 Reconstruction Techniques ...... 107 7.1 Noise Cleaning ...... 107 7.2 Particle Direction ...... 108 7.3 Angular Uncertainty ...... 118 7.4 Interaction Vertex ...... 121 7.5 Energy ...... 124
8 Opening Up a Neutrino Window to the Southern Hemisphere ...... 129 8.1 The FSS Filter ...... 132
9 Point Source Analysis Methods ...... 139 9.1 Hypothesis Testing ...... 139 9.2 Likelihood and Test Statistic ...... 141 10 A Search for Low-Energy Starting Events from the Southern Hemisphere ...... 147 10.1 Analysis Strategy ...... 149 10.2 Experimental Data ...... 151 10.3 Simulated Data ...... 154 10.4 Event Selection ...... 154 10.5 Event Selection Summary ...... 189 10.6 Likelihood Analysis ...... 192 10.7 Results ...... 201 10.8 Systematic Uncertainties ...... 209
11 Summary and Outlook ...... 213
Summary in Swedish ...... 217
Appendix A: BDT Input Variables ...... 223
Appendix B: Result Tables ...... 227
References ...... 235 Acknowledgements
During my time as a graduate student at Uppsala University I have had the privilege to work with many dedicated, helpful and experienced scientists. We’ve organized meetings and workshops, had lengthy discussions on pecu- liar features of fundamental physics, and had a lot of fun during lunches and fika breaks, discussing everything from foreign and domestic politics to candy wrapping and mushroom picking. I’ve enjoyed the company of each one of you, thank you for this great time. Some of you deserve a special mention. First of all I would like to thank my supervisors Allan Hallgren and Olga Botner. Few graduate students have had the privilege to work close to two such excellent and curious physicists, always available to answer questions no matter the scope. Allan was the main supervisor of my thesis and introduced me to the physics potential of the Southern hemisphere at energies below 100 TeV. Together we developed a new data stream for collecting such low-energy events. After the discovery in 2013 of the first ever high-energy astrophysi- cal neutrinos, the interest in this channel has grown dramatically. Allan has a never-ending stream of ideas and taught me how to be daring and bold in my creativity. He also emphasized the need to stay critical to accepted knowledge. How he can successfully calculate and compare areas of histograms with dif- ferences as small as 1% solely by judging the shape of lines on a plot located 3 meters away, I will never know. That remains a true mystery to me. Olga helped me organize my work and to prioritize. Olga has the rare gift of making people think in new ways by simply listening and asking the right questions. We’ve had countless discussions about everything from V-A cou- plings, the Higgs mechanism, and the nature of astrophysical objects. Thank you for taking time to discuss these things with me over the years, although buried in the work that comes by being the spokesperson of IceCube and a member of the Nobel committee. I’d like to extend my thanks also to my co-supervisor Chad Finley. Thank you for taking an interest in my analysis and for all the discussions we’ve had about OneWeight, likelihood algorithms, and hypothesis testing. I would also like to thank Carlos Perez de los Heros. We’ve discussed many topics, ranging from dark matter to statistical methods and art. Whenever you knocked on my door, I knew I was going to learn something new and unexpected. Henric Taavola, we’ve had as fun as I think two people can have. I want to thank you for five fantastic years, and especially that you put up with me singing, dancing, talking politics and doing white-board science in a more than often weird mix and high pace. I will forever remember those days we
9 pretended it was Christmas although June, the days we walked to the gas sta- tion to buy falafel because we both forgot to bring lunch, the days we took a walk in the beautiful forest nearby, and the ‘glee’ days where we limited all communication to be done through musical numbers only. All of those days have a special place in my heart and so do you! I would like to send a big thanks the people in the IceCube groups in Up- psala and Stockholm, past and present: Olga, Allan, Carlos, Sebastian, Hen- ric, Lisa, Alexander, Christoph, Leif, Olle, David B., Jonathan M., Chad , Christian, Klas, Jon, Samuel, Martin, Marcel, Matthias, Maryon, and the late Per-Olof. Thank you all for being encouraging and helpful. Per-Olof was an inspiration to all scientists and will continue to be so. I remember him as a very determined and intelligent man with a huge pathos and an open heart. He recommended me for the service work I did on the South Pole and Olga and Allan kindly approved. That was an amazing trip and an absolute highlight of my time as a graduate student. Thank you John, Steve, Larissa, Ben R., Felipe, Blaise, Jonathan D., and Hagar for making this trip unforgettable. I also extend my thanks to everyone in the IceCube Collaboration, in partic- ular to those whom I have worked with. A special thanks goes to Antonia, Zig, Carlos, Kim, Megan, and Olivia. Thank you for making me feel like home in Madison. Thanks also: Kai, Volker, Donglian, Mårten, Mike R., David A., Martin B., Matt, Elisa R., Naoko, Stefan, Jake F. , Jake D., John F., Robert, Claudio, Jason, Mike L., Jan B., Jan K., Ryan, Frank, Jim, Laurel, Sandy, Anna P., Anna B., Sarah, Tania, James, Elisa P., Katherine, Joakim, Michael S., Jakob, Klaus, Ben W., Kurt, Moriah, Gwen, Geraldina, Martin C., and Nancy. You are all great people, and I hope I will have the change to work with all of you again someday. Stefan Leupold, you’re an excellent teacher that takes all questions seri- ously and makes the students feel smart. We have had so many interesting discussions, especially about the theoretical aspects of the Standard Model. Thank you Inger and Annica for providing an escape from physics at times needed and for always being helpful. I would like to thank my mother Maria, you are extraordinary in so many ways that I can’t even begin to describe. You are supportive and believe I can do and achieve anything. Thank you also to my sister Linda, to Bazire, my father Roland, and my late grandparents Evert and Edit. I know you are all very proud of me and I can honestly say that I would not have been able to do this without your support. A big thanks to all my friends: Katarina, Laura, Daniel, Sanna, Mikael, Noel, Anna-Karin, Andreas, Jessica, Maja, Lena, Li, Jim, Tove, Filip, and Anette. A special thanks to Katarina with whom I started the science podcast Professor Magenta. You are exceptionally smart and fun to hang around, and you bring out the best of me. I’m looking forward to much more of all of that in future episodes of both the podcast and my life.
10 My biggest thanks goes to Tony. Thank you for being amazing, for cooking, cleaning, and arranging my life outside of physics. Your encouragement and support means a lot. You are smart and brilliant and I have become a better person because of you. Thank you also to the Lennanders foundation for awarding me one of your scholarships enabling me to finish this project, and to all of you who helped me finish this thesis by providing constructive comments and suggestions on how to improve the text: Olga, Allan, Sebastian, David, Stefan, and Henric. The IceCube Collaboration acknowledge the support from the following agencies: U.S. National Science Foundation Office of Polar Programs, U.S. National Science Foundation Physics Division, University of Wisconsin Al- umni Research Foundation, the Grid Laboratory Of Wisconsin (GLOW) grid infrastructure at the University of Wisconsin - Madison, the Open Science Grid (OSG) grid infrastructure; U.S. Department of Energy, and National Energy Research Scientific Computing Center, the Louisiana Optical Net- work Initiative (LONI) grid computing resources; Natural Sciences and Engi- neering Research Council of Canada, WestGrid and Compute/Calcul Canada; Swedish Research Council, Swedish Polar Research Secretariat, Swedish Na- tional Infrastructure for Computing (SNIC), and Knut and Alice Wallenberg Foundation, Sweden; German Ministry for Education and Research (BMBF), Deutsche Forschungsgemeinschaft (DFG), Helmholtz Alliance for Astropar- ticle Physics (HAP), Research Department of Plasmas with Complex Interac- tions (Bochum), Germany; Fund for Scientific Research (FNRS-FWO), FWO Odysseus programme, Flanders Institute to encourage scientific and techno- logical research in industry (IWT), Belgian Federal Science Policy Office (Belspo); University of Oxford, United Kingdom; Marsden Fund, New Ze- aland; Australian Research Council; Japan Society for Promotion of Science (JSPS); the Swiss National Science Foundation (SNSF), Switzerland; National Research Foundation of Korea (NRF); Danish National Research Foundation, Denmark (DNRF).
Rickard Ström, Uppsala, November 2015
11
About this Thesis
In this thesis we present an analysis using data from the IceCube Neutrino Observatory to study the neutrino flux from the Southern hemisphere at en- ergies between 100 GeV and a few TeV. In particular, we search for neutrino point sources as indicators for sites of hadronic acceleration where some of the most energetic particles in the Universe are thought to be produced and/or accelerated. The thesis is divided into the following chapters: Chapter 1 is an intro- duction to high-energy neutrino astrophysics. It also serves as the physics motivation for the analysis. In chapter 2 we discuss the current framework of particle physics, the Standard Model. We focus on the properties of the weak interaction, the only known force (except gravity) that interacts with neutri- nos. In chapter 3 we introduce the so-called cosmic ray puzzle and discuss neutrino production in astrophysical sources as well as in the Earth’s atmo- sphere. Chapter 4 deals with the interaction and energy loss of neutrinos and also introduces the detection principle and the different event topologies seen in IceCube. The IceCube detector, including its optical sensors and data ac- quisition system is discussed in chapter 5. In chapter 6 we describe the event simulations used to produce the signal and background samples that we use in the analysis. In chapter 7 we go through the noise cleaning algorithms applied to data and all reconstruction techniques used to characterize each observed event in terms of e.g. direction and energy. Chapter 8 defines the online filter used at the South Pole to select the neutrino candidate events that constitute the foundation of the analysis presented in chapter 10. The likelihood analy- sis method is introduced in chapter 9 and conclusions and discussions of the results are presented in chapter 11.
The Author’s Contribution Inspired by my work as a Master student in the Fermi collaboration, I started off by searching for dark matter from the Galactic center region, but later switched to a search for point sources in the whole Southern hemisphere. These searches both focus on lower energies, where traditional IceCube meth- ods cannot be applied, in particular the southern sky. I developed a new data stream by constructing the so-called Full Sky Starting (FSS) filter described in detail in chapter 8. A lot of time was spent on tests and verification of this filter.
13 The IceCube crew at the Amundsen-Scott South Pole Station during my stay. From left: Felipe Pedreros, Ben Riedel, Rickard Ström, Jonathan Davies, Steve Barnet, Hagar Landsman, John Kelley, Blaise Kuo Tiong, and Larissa Paul.
I have participated in ten IceCube collaboration meetings, including one in Uppsala where I also was a part of the organizing committee. Further, I have had the privilege to participate in several summer and winter schools discussing neutrino theory, phenomenology and experiments as well as dark matter and high-energy physics in general. I also attended the Astroparticle Physics Conference (ICRC) in 2015, where I presented the results of my anal- ysis. The proceedings are pending publication in Proceedings of Science and are attached in the end of this thesis. Other work, not included in this thesis, where I have contributed during my time as a graduate student include: • IceCube service work. I travelled to the South Pole, Antarctica, in Jan- uary 2013 for three weeks of service work on the in-ice detector. I worked together with John Kelley to recover DOMs disconnected due to various problems, such as communication errors, unstable rates, etc. This was not only an opportunity to work hands-on with the detector sys- tem at the IceCube Lab but also to get an amazing insight in the detector and the IceCube project as a whole.
14 • IceCube monitoring. In 2013 I did two weeks of monitoring of the IceCube detector, and was later asked to develop educational material for coming monitoring shifters as well as to educate them. • Teaching assistant. During my time as a graduate student I have been a teaching assistant on several courses including: particle physics, nuclear physics, and quantum mechanics. Further I have taken a course in ped- agogic training for academic teachers and worked together with several other teaching assistants to improve the laboratory exercises in nuclear physics. • Outreach activities. I participated in several outreach activities and built a permanent exhibition of IceCube at Ångströmlaboratoriet at Uppsala University together with my colleague Henric Taavola. It contains an event display and monitors showing pictures and a documentary of the construction of the detector. • I participated in ‘Gran Sasso Summer Institute 2014 - Hands-On Exper- imental Underground Physics at LNGS’ (Laboratori Nazionali del Gran Sasso), near the town of L’Aquila, Italy, where I performed PMT inves- tigations in the Borexino test facility, together with Osamu Takachio and Oleg Smirnov. The goal was to characterize a large PMT that might be used in several future neutrino experiments. The results were published in Proceedings of Science with identification number PoS(GSSI14)016.
Units and Conventions Throughout this thesis we will use eV (electronvolt) as the standard unit of en- ergy, corresponding to about 1.602 · 10−19 J. It is defined as the energy gained by an electron traversing an electrostatic potential difference of 1 V. Through- out the thesis we also use natural units, i.e., = c = kB = 1, where = h/2π, h is Planck’s constant, c is the speed of light in vacuum, and kB is the Boltzmann constant. This means that we can express particle masses in the same units as energy through the relation E = mc2 = m. Effectively this can be thought of as expressing the masses in units of eV/c2. Throughout this thesis we will use GeV = 109 eV, TeV = 1012 eV, PeV = 1015 eV, and EeV = 1018 eV. When discussing astrophysical objects we sometimes use the erg unit (1 TeV = 1.6 erg).
Cover Illustration The illustration on the cover of this thesis shows the evolution of the Universe from a dense plasma to the structures we see today, such as galaxies, stars, and planets. Credit: European Space Agency (ESA).
15 The author standing on the roof of the IceCube Lab (literally on top of the IceCube detector) at the South Pole, Antarctica. Visible in the background to the left: the 10 m South Pole Telescope (SPT) and the BICEP (Background Imaging of Cosmic Extra- galactic Polarization) telescope. To the right, The Martin A. Pomerantz Observatory (MAPO).
16 Abbreviations and Acronyms
AGILE Astro-rivelatore Gamma a Immagini Leggero AGN Active Galactic Nuclei AHA Additionally Heterogeneous Absorption AMANDA The Antarctic Muon And Neutrino Detector Array AMS Alpha Magnetic Spectrometer ANITA Antarctic Impulse Transient Antenna ANTARES Astronomy with a Neutrino Telescope and Abyss environmental Research project ARA Askaryan Radio Array ARIANNA The Antarctic Ross Ice Shelf Antenna Neutrino Array ASCA The Advanced Satellite for Cosmology and As- trophysics ATLAS A Toroidal LHC ApparatuS ATWD Analog Transient Waveform Digitizer
Baikal-GVD Baikal Gigaton Volume Detector BDT Boosted Decision Tree BDUNT The Baikal Deep Underwater Neutrino Tele- scope
CC Charged Current CERN The European Organization for Nuclear Research (French: Organisation européenne pour la recherche nucléaire) Chandra The Chandra X-ray Observatory CKM The Cabibbo-Kobayashi-Maskawa Matrix C.L. Confidence Level CMB Cosmic Microwave Background CMS The Compact Muon Solenoid COG Center of Gravity CORSIKA COsmic Ray SImulations for KAscade
17 CP Charge Conjugation and Parity Violation CPU Central Processing Unit CR Cosmic Ray CRT Classic RT-Cleaning CTEQ5 The Coordinated Theoretical Experimental Project on QCD
DAQ Data Acquisition System DIS Deep Inelastic Scattering DOM Digital Optical Module DOR DOM Readout DSA Diffusive Shock Acceleration
ESA European Space Agency
fADC fast Analog-to-Digital Converter FSRQ Flat-Spectrum Radio Quasar FSS Full Sky Starting
GC Galactic Center GENIE Generates Events for Neutrino Interaction Ex- periments GPS Global Positioning System GPU Graphical Processing Unit GRB Gamma-Ray Burst GUT Grand Unified Theories GZK Greisen-Zatsepin-Kuzmin
HESE High-Energy Starting Events HESS High Energy Stereoscopic System HLC Hard Local Coincidence HST The Hubble Space Telescope HV High Voltage
18 IC Inverse Compton IceCube The IceCube Neutrino Observatory ICL IceCube Lab IR Infrared ISM The Interstellar Medium ISS International Space Station
Kamiokande Kamioka Nucleon Decay Experiment KM3NeT Cubic Kilometer Neutrino Telescope KS Kolmogorov-Smirnov
LED Light Emitting Diode LEP The Large Electron-Positron Collider LESE Low-Energy Starting Events LHC The Large Hadron Collider LMC Large Magellanic Cloud
MAGIC Major Atmospheric Gamma Imaging Cheren- kov Telescopes MC Monte Carlo MESE Medium-Energy Starting Events MPE Multi PhotoElectron Mu2e The Muon-to-Electron experiment
NC Neutral Current NH Northern Hemisphere NuGen Neutrino-Generator
PAMELA Payload for Antimatter Matter Exploration and Light-nuclei Astrophysics PAO The Pierre Auger Observatory PDF Probability Density Function
19 p.e. photoelectron PINGU The Precision IceCube Next Generation Up- grade PMNS The Pontecorvo-Maki-Nakagawa-Sakata Ma- trix PMT Photo Multiplier Tube PPC Photon Propagation Code PREM Preliminary Reference Earth Model PROPOSAL PRopagator with Optimal Precision and Opti- mized Speed for All Leptons PTP Punch Through Point PWN Pulsar Wind Nebula
QCD Quantum Chromodynamics QED Quantum Electrodynamics QFT Quantum Field Theory
R.A. Right Ascension RMS Root Mean Square
SED Spectral Energy Distribution SH Southern Hemisphere SLC Soft Local Coincidence SM The Standard Model SMT Simple Majority Trigger SNO The Sudbury Neutrino Observatory SNR Supernova Remnant SPATS South Pole Acoustic Test Setup SPE Single PhotoElectron SPICE South Pole ICE SRT Seeded RT-Cleaning SSC Self-Synchrotron Compton STeVE Starting TeV Events
20 SuperDST Super Data Storage and Transmission Super-K Super-Kamiokande SUSY SUperSYmmetry
TWC Time Window Cleaning
UHE Ultra High-Energy UV Ultra-violet
VERITAS Very Energetic Radiation Imaging Telescope Array System VLA The Very Large Array VLT The European Southern Observatory’s Very Large Telescope
WB Waxman-Bahcall WIMP Weakly Interacting Massive Particle
21
1. Introduction
“Somewhere, something incredible is waiting to be known.” Carl Sagan
Intergalactic Magnetic Field
Neutrino N
Photon
W E Proton
S Heavy Nucleus
Figure 1.1. Illustration showing the properties of different astrophysical messenger particles. Starting from the top: Neutrinos (blue) only interact with the weak force and are hence the ideal cosmic messenger, photons (yellow) may scatter or become absorbed on their way to Earth, charged particles (green, red) are affected by magnetic field and do not necessarily point back to their source. Credit: Jamie Yang/WIPAC.
Astroparticle physics denotes the scientific study of elementary particles and their interactions, in the context of our Universe. It is a branch of parti- cle physics dealing with cosmic messengers such as neutrinos, cosmic-rays, and high-energy photons, in many cases produced and accelerated in the most violent and extreme environments of the Universe. Historically astroparticle physics sprung out of optical astronomy and the curiosity to understand and reflect upon the Universe we live in: a gigantic laboratory for astrophysical environments and high-energy physics, that go beyond what we can produce and measure on Earth. Some of the fundamental questions addressed are: Which are the funda- mental building blocks in the Universe and what are their properties? Can the elementary particles and their interactions explain the structures seen in the Universe? What is the origin of cosmic rays? What are the mechanisms behind the acceleration of particles to extremely high energies?
23 In this introductory chapter we will present the physics motivation for this thesis, introduce its main actors - the neutrinos - and discuss their unique role in the exploration of the high-energy Universe. Further we will briefly mention the efforts to connect different kinds of observations of the same astrophysical objects into one unified description, known as the multi-messenger approach.
1.1 Physics Motivation The first humans studied the Universe with their own eyes looking for an- swers in the optical light emitted by stars and galaxies. In fact almost every- thing we know today about the Universe is the result of an exploration of the electromagnetic spectra at different wavelengths. As the techniques and meth- ods grew more sophisticated we were able to study the Universe as seen in radio, infrared, X-ray and eventually γ-ray radiation, the latter consisting of extremely energetic particles. Since photons are massless and electrically neutral they are unaffected by the magnetic fields in the Universe and travel in straight lines, in space-time, from the point where they escape their production site to Earth where they can be observed. Indeed they are excellent messenger particles, carrying in- formation about the physics and constituents of other parts of the Universe. However, they may be absorbed in dense environments on the way or at the production site itself, scattered in dust clouds, and captured by ambient radia- tion and matter in the interstellar medium (ISM). Astronomy is also possible with charged particles, so-called cosmic rays, but their directions are randomized by interstellar and intergalactic magnetic fields for all but the highest energies. Figure 1.2 shows the energy of particles as a function of the observable dis- tance. The blue shaded region illustrates the parameter space where photons are absorbed through interactions with the Cosmic Microwave Background (CMB) and Infrared (IR) background radiation. The red shaded area shows the region where high-energy protons are absorbed by interactions with the CMB. Low-energy protons do not point back to their source, due to magnetic deflection. Further, the upper (lower) horizontal dashed black line indicate the energy of the highest proton (photon) ever observed. The mean free path of photons decreases quickly with increasing energy and above 1 PeV (1015 eV) it is shorter than the typical distances of our own galaxy. The colored bars in the bottom part indicate the position of the Galactic plane, the local galaxy group, some of the closest Active Galactic Nucleis (AGNs), etc. The idea of using neutrinos as cosmic messengers has been around since the late 1930s. They are electrically neutral and about one million times lighter than the electron [2]. Further they only interact with the weak nuclear force and gravitationally. Indeed neutrinos are the perfect astrophysical messenger with the only drawback that the low cross-section also makes them very diffi-
24 Figure 1.2. Particle energy as a function of the observable distance. The blue (red) shaded region illustrates the parameter space where photons (protons) are absorbed by interaction with the CMB and IR background radiation. The upper (lower) hori- zontal dashed black line indicate the energy of the highest photon (proton) ever ob- served. Credit: P. Gorham, 1st International Workshop on the Saltdome Shower Array (SalSA), SLAC, (2005). cult to detect. In fact, it turns out that one needs detectors of cubic-kilometer size to see even a handful of astrophysical neutrinos per year. Further, the presence of a large atmospheric background requires neutrino detectors to be built underground. The first discovery of astrophysical neutrinos was made in the early 1970s, when Ray Davis and collaborators observed neutrinos coming from the Sun using a detector at the Homestake Gold Mine in South Dakota. Davis was awarded the Nobel Prize in 2002 together with Masatoshi Koshiba who led the design and construction of the Kamiokande (Kamioka Nucleon Decay Ex- periment), a water imaging Cherenkov detector in Japan which later showed that the neutrinos actually pointed to the Sun and was further one of the de- tectors that observed neutrinos from supernova SN1987A in the Large Magel- lanic Cloud (LMC) in early 1987. These were the first neutrinos ever observed from outside the Solar system. Since then, many new experiments have been built and observations show an ever increasing number of peculiar facts about the neutrinos. Most notably they seem to have tiny masses giving rise to so-called neutrino oscillations,
25 responsible for their chameleonic nature: they transform from one type into another as they propagate through space. These observations successfully ex- plained one of the longest standing problems of neutrino astrophysics, namely the presence of a large deficit in the number of observed electron-neutrinos from the Sun, see section 2.8.2. One of the next big breakthroughs in neutrino astrophysics was made in 2013 with the discovery of a diffuse flux of high-energy neutrinos of astro- physical origin by The IceCube Neutrino Observatory (hereafter IceCube) [3]. Localized sources of high-energy neutrinos have not yet been seen outside of our solar system and remain one of the most wished-for discoveries. In par- ticular since neutrinos are indicators for sites of hadronic acceleration where some of the most energetic particles in the Universe are thought to be produced and accelerated.
1.2 The Multi-Messenger Approach With more pieces of the great cosmic puzzle we can build a greater picture. This is the essence of what is called the multi-messenger approach in astron- omy: combining observations with different messengers and at various ener- gies to reach a deeper understanding of an event or process in the Universe. In practice this means that traditional observations of electromagnetic ra- diation are combined with observations of high-energy γ-rays and neutrinos to provide complementary information. In the future gravitational waves may play an equally important role. A stunning example of what can be learned is shown in figure 1.3 where we present multi-wavelength images of the nearby galaxy Centaurus A, reveal- ing enormous jets of relativistic particles perpendicular to the accretion disc. Localized objects of neutrino emission may reveal other mind-boggling facts about the most energetic sources in the Universe. Further, neutrino flux predictions build upon the close association between the production and acceleration of cosmic rays and the non-thermal photon emission from astrophysical sources. This close connection is covered in more depth in chapter 3.
26 Figure 1.3. Multi-wavelength images of the nearby galaxy Centaurus A. Top right: X- ray data from Chandra. Mid right: Radio data from VLA. Bottom right: Optical data from the ESO’s Wide-Field Imager (WFI) camera at the ESO/MPG 2.2-m telescope on La Silla, Chile. Left: Combined X-ray, radio, and optical data. Credit: X-ray: NASA/CXC/CfA/R.Kraft et al; Radio: NSF/VLA/Univ.Hertfordshire/M.Hardcastle; Optical: ESO/WFI/M.Rejkuba et al.
27
2. The Standard Model
“Young man, if I could remember the names of these particles, I would have been a botanist.” Enrico Fermi
The visible Universe consists of mainly three elementary particles; the elec- tron, the up-quark and the down-quark. The two latter are the building blocks of the atomic nuclei forming the more familiar proton and neutron, while the electron orbits the nucleus and gives rise to electricity and complex atomic and molecular bonds. These particles are the fundamental constituents of proteins, cells, and humans, but also astrophysical objects such as stars, planets and the vast ISM. When combined with the electromagnetic force we end up with a set of tools that can explain most of the physics of everyday life. But it was realized already in the early 1930s, after careful studies of cosmic-rays (see chapter 3) that more particles and forces were needed to explain the large variety of radiation and phenomena observed [2]. The first major step towards creating a unified standard model in particle physics was taken in the 1960s by Glashow, Weinberg, and Salam [4]. They managed to combine the electromagnetic interaction, Quantum Electrodynam- ics (QED), with the weak interaction, creating a unified electroweak theory in the same spirit as Maxwell in the 1860s, who formulated the theory of the electromagnetic field encompassing the previously separate field theories: electricity and magnetism. The modern Standard Model (SM) of particle physics is a collection of quantum field theories that gives us a unique insight into the structure of matter in terms of a plethora of 301 elementary particles and the interactions among them. Elementary particles are particles thought to be definitive constituents of matter: the smallest building blocks of nature. The SM describes the elec- troweak force and the strong nuclear force (a.k.a. the strong force), the latter in the theory of Quantum Chromodynamics (QCD). Further the SM provides tools to calculate and predict interactions of com- binations of elementary particles, e.g., observed in the incredibly rich zoo of bound quark states, so-called hadrons.
1This number is derived by counting the particles shown in figure 2.1: 12 fermions, 12 anti- fermions, and 6 bosons (the hypothetical graviton not included).
29 Figure 2.1. The Standard Model of elementary particles. Credit: CERN.
The SM is by far the most accurate theory ever constructed [4] and its crowning achievement was made only years ago in 2012 when the ATLAS (A Toroidal LHC ApparatuS) and CMS (The Compact Muon Solenoid) Col- laborations together announced the discovery of the Higgs particle in data from proton-proton collisions at the Large Hadron Collider (LHC) complex at CERN [5, 6]. In the following sections we will present the constituent particles and fields of the SM. We will discuss the importance of symmetries in the underlying theories and further explore the weak interaction that is essential in the de- scription of the interactions and phenomena of neutrinos. We conclude the chapter by discussing theories that go beyond the SM.
2.1 Matter Particles and Force Carriers The SM is presented schematically in figure 2.1 where each square represent the observed particle. Each particle is associated with a set of quantum num- bers such as e.g. electric charge, mass2, and spin. These determine the way the
2 Strictly speaking, neutrinos (νe, νμ, and ντ) are not mass eigenstates [4].
30 particles interact and propagate through space. For a thorough introduction to the particles and interactions in the SM see e.g. [4]. The 3 columns on the left hand side of the figure consist of families (also known as generations) of half-integer spin particles called fermions. These are often thought of as the particles that constitute matter. The particles in the first column (u, d, e, and νe) are stable on large time scales, while the 2nd and 3rd column contain particles seen almost exclusively at accelerators or in particle showers resulting from cosmic ray interactions in the Earth’s atmo- sphere. The three generations are literally copies of the first generation, with the only observable difference being higher mass for the rare particles. The two top rows show the six quarks that all have fractional charges. These are not observed as free particles in nature but rather in bound states of either three quarks, so-called baryons, or a pair consisting of a quark and an anti-quark, so- called meson3. This is a manifestation of a phenomenon called confinement which is a part of QCD. The two bottom rows show the six leptons: three charged leptons (e−, μ−, and τ−) each with a corresponding neutral particle, the neutrino (νe, νμ, and ντ). As indicated in the figure, the fermions also have corresponding antiparticles4. The right hand side of the figure shows integer spin particles called bosons. In the SM a fundamental force is the result of continuous exchange of such force mediating particles. The large rectangles indicate to what extent the forces act upon the fermions. Starting from bottom we have the weak force ± 0 mediators W and Z . These couple to particles with weak iso-spin IW and weak hypercharge Y. In particular, they couple to all fermions and constitute the only SM coupling to neutrinos resulting in their unique role in astroparticle physics. The charged part of the weak force mediated by W± is responsible for radioactive decays of subatomic particles. The γ (photon) is the mediator of the electromagnetic force and couples to all charged particles (incl. W±), while the eight (8) massless color-state gluons (g) only interact with the quarks (and themselves) to give rise to the strong force responsible for the attractive bonds in hadrons and nuclei. As with the quarks, gluons have never been observed free in nature, but hypothetical gluon-only states called glue-balls could exist and might have been observed [8]. The distinction between fermions as the constituents of matter on one side and bosons as force mediating particles on the other, is somewhat simplified. E.g. the valence quarks alone cannot account for the mass or spin in composite particles such as protons. To determine the properties of such bound hadronic states we need to consider the impact of both the quark- and gluon-sea. John Wheeler, a famous theoretical physicist known for his large contribu- tions to the field of general relativity and also for coining the term ’wormhole’
3Indications for the existence of more exotic configurations have been suggested lately, see [7] for a recent review on the topic. 4Neutrinos might actually be their own antiparticles, but this remains to be seen in experiment.
31 60 U(1) 50
40 SU(2) -1 α 30
20
10 SU(3)
0 2 4 6 8 10 12 14 16 18 Log10(Q/GeV)
Figure 2.2. The inverse of the gauge couplings as function of energy for the three interactions in the SM (dashed lines) and including SUSY (solid lines), extrapolated to the GUT scale. The two SUSY lines represent different values for the energy scale of the SUSY particles (500 GeV and 1.5 TeV). Figure taken from [9].
(theoretical shortcuts in four-dimensional space-time), once said “Mass tells space-time how to curve, and space-time tells mass how to move”. The in- terplay between fermions and bosons is very much the same: just like yin and yang, two opposite and contrary forces that work together to complete a wholeness.
2.2 Interaction Strengths The coupling strength between fermions and bosons is determined by dimen- sionless gauge couplings α, so that the probability of an interaction includes a factor α for each interaction vertex. Effectively these couplings are energy dependent, see figure 2.2 where the inverse gauge couplings (‘strengths’) are shown as function of energy for the three interactions in the SM (dashed lines) and including supersymmetry (SUSY) (solid lines), see section 2.8. The energy dependence can be understood in terms of vacuum polariza- tions where virtual particle-antiparticle pairs with the charge(s) relevant for each force partially screen the fundamental coupling in the field surrounding a central charge. For QED the effect gets smaller the closer we get to the central charge, i.e., the coupling grows with energy. QCD is fundamentally
32 different in that the corresponding gauge bosons, the gluons, themselves carry (color) charge. The polarization of virtual gluons instead augment the field, an effect that in the end wins over the color charge screening caused by virtual quark-antiquark pairs. QCD is therefore referred to as being asymptotically free in that the bonds become asymptotically weaker with increasing energy. A similar balancing act is observed for the weak interaction with the self- interacting W± and Z0, but is less pronounced due to the finite masses of the gauge bosons. Further the number of gauge bosons is smaller leading to an additional weakening of the effect. At energies close to the electroweak scale O(100 GeV), we obtain the fol- lowing relations between the fundamental forces: α−1 α−1 α−1 ≈ , EM : W : S 128:30:9 (2.1) where all gauge couplings are sufficiently small for perturbation theory to ap- ply. Somewhat lower in energy, at O(1 GeV), perturbation theory breaks down for QCD and we enter the non-perturbative region with bound hadronic states and the final levels of hadronization processes in which hadrons are formed out of jets of quarks and gluons as a consequence of color confinement.
2.3 Symmetries Symmetries play an essential role in modern physics and in particular in the SM. The first to understand their significance was Emmy Noether who in 1915 proved that continuous symmetries give rise to conserved quantities [10]. The SM is constrained by a set of local and global symmetries. These con- straints are motivated by experimental evidence and observations of particles interactions. In section 2.3.2 we describe how local symmetries dynamically give rise to mass-less force carriers and the conservation of so-called good quantum numbers. But we will begin by considering a number of global symmetries in section 2.3.1. In the framework of the SM these are often thought of as accidental in contrast to the local gauge symmetries.
2.3.1 Global Symmetries One of the most important laws of conservation is that of the baryon number B. In particular because it ensures the stability of the proton by prohibiting processes like5: + p e + νe + ν¯e. (2.2) Baryons (anti-baryons) are assigned baryon number B = 1(B = −1) or equiv- alently quarks (anti-quarks) can be assigned B = 1/3(B = −1/3). Mesons,
5Also forbidden by lepton number conservation.
33 consisting of a quark and anti-quark pair, therefore have baryon number 0 and − − can therefore decay into the lepton sector (e.g. π → μ + ν¯μ)6. Separate quantum numbers can also be defined for the different quark fla- vors. These are all seen to be conserved separately for both the electromag- netic and strong interaction. We will see below that the weak interaction cou- ples to isospin doublets with mixing between the three families of quarks and can therefore violate the individual quark flavor quantum numbers. To distinguish neutrinos from anti-neutrinos we introduce a total lepton = − number L nl nl¯, where nl(lˆ) is the number of leptons (anti-leptons). This has been observed to be conserved in all known interactions. Further we observe that separate lepton numbers for each family Le, Lμ, and Lτ are also conserved with the exception of considering the oscillations of neutrino eigenstates in the weak sector. Electromagnetic interactions that vio- late lepton family numbers, e.g., μ± → e± + γ, are searched for but have never been observed. The Mu2e (Muon-to-Electron) experiment at Fermilab (US) began construction in early 2015 and expects preliminary results around 2020 [11]. It will have unprecedented sensitivity for such decays and is particularly interesting since such conversions are expected at low rates in many models of physics beyond the SM, e.g., SUSY discussed briefly in section 2.8.
2.3.2 Local Gauge Symmetries It can be shown that a system described by the Lagrangian L is invariant under local phase transformations if we add terms to L that cancels the ones from the derivatives of the local phase itself. This is called the principle of local gauge invariance [12]. The additional terms introduce matter-gauge field interactions, realized by the exchange of massless spin-1 bosons as force mediators. In particular they introduce conserved currents with associated conserved quantum numbers. E.g. the electromagnetic interaction described by the U(1)7 local gauge sym- metry is mediated by the massless γ and further leads to the conservation of electric charge8, a quantum number that is non-zero for particles participating in the electromagnetic interaction. Further, the strong force mediated by eight massless gluons acts on particles with quantum numbers called color charge, a consequence of the SU(3)C[4] local gauge symmetry of the terms describing the strong interaction. The weak interaction is discussed in detail below in section 2.4. For a more complete discussion about local gauge symmetries see textbooks in particle physics and quantum field theory, e.g., [12] and [4].
6Leptons have baryon number B = 0. 7The Unitary group U(n) with n = 1 consist of all complex numbers with absolute value 1 under multiplication (the so-called circle group) [4]. 8The conservation of charge is in fact a consequence already of global U(1) symmetry.
34 2.4 The Weak Interaction The weak force is unique in the following ways: Two of the three force carri- ≈ . ± ≈ . ers have electric charge, they are all massive (mZ0 91 2 GeV [2], mW 80 4 GeV [2]) unlike the force carriers of QED and QCD, and the interactions do not conserve parity nor charge conjugation. Further, weak interactions can be categorized into two separate processes: Charged Current (CC) interac- tions mediated by W± and Neutral Current (NC) interactions mediated by Z0. The latter were first identified in the Gargamelle bubble chamber exper- iment at CERN in 1974 [13] using a beam of muon-neutrinos produced in pion decay. The Gargamelle Collaboration observed two types of NC events: muon-neutrino scattering from a hadron without turning into a muon, as well as events characterized by a single electron track, see figure 2.3. The weak bosons themselves were discovered at the Large Electron-Positron Collider (LEP) at CERN in 1983 [14, 15, 16]. The charge current interactions violate parity (mirror symmetry) and charge conjugation maximally and only interact with left-handed particles and right- handed antiparticles9. This is condensed into a so-called “vector minus axial- vector” (V-A) theory. This chiral theory can be described by an SU(2)L [4] local gauge symmetry where right-handed particles and left-handed antipar- ticles have been placed in weak-isospin singlet states, i.e., with weak isospin 0:
(ψe)R,(ψμ)R,(ψτ)R and (ψu)R,(ψc)R,(ψt)R,(ψd)R,(ψs)R,(ψb)R. (2.3) Note that there are no right-handed neutrinos in the SM [4]. Left-handed par- ticles (and right-handed antiparticles) are arranged in weak isospin doublets differing by one unit of charge: ψν ψν ψν ψ ψ ψ e , μ , τ and u , c , t . (2.4) ψ ψ ψ ψ ψ ψ e L μ L τ L d L s L b L The flavor eigenstates in the quark sector differ from the corresponding mass states giving rise to so-called quark mixing, parametrized by the Cabibbo- Kobayashi-Maskawa (CKM) matrix. Mixing is also seen in the lepton sector and is described by the Pontecorvo-Maki-Nakagawa-Sakata (PMNS) matrix. While charged leptons are eigenstates of both flavor and mass, neutrinos are not, resulting in neutrino oscillations10. The mass of the neutrinos is discussed is section 2.8.2.
9The handedness of the particles can either refer to the helicity, the projection of the spin onto the direction of momentum or as here chirality, an abstract concept related to helicity (for massless particles they are in fact the same) that describes how particle states transform un- der Lorentz transformations. 10We consider the weak mixing to take place between the neutrinos and not the charged leptons but this is merely a matter of definition. However, mixing among the neutrinos is the natural choice since mixing between charge leptons would be unobservable technically due to the large mass splitting.
35
Figure 2.3. The first neutral current event ever detected using the Gargamelle experi- ment at CERN in 1973. Credit: Picture printed with permission from CERN. Particle trace by Christine Sutton (Editor, CERN Courier).
Complex phases in these matrices give rise to Charge-Parity (CP) viola- tion effects. This has been observed in the CKM matrix and leads to weak processes that occur at different rates for particles and mirrored antiparticles. CP-violations are particularly interesting since they might be connected to the difference between the content of matter and anti-matter in the Universe. CP- violation has yet to be observed in the lepton sector and is the focus of the next generation of long-baseline neutrino oscillation experiments. 1 2 3 SU(2)L has three massless gauge bosons W , W , and W [12]. Linear combinations of the first two can be identified as the charged W-bosons (es- sentially corresponding to the raising and lowering operators of weak isospin) and it would be tempting to associate the third boson with the observed Z0. The problem is that W3 breaks parity maximally unlike Z0 that does couple to right-handed particles and left-handed antiparticles although with a different strength from left-handed particles and right-handed antiparticles. Hence, fol- lowing the same procedure as for QED and QCD leads to a theory in tension with experimental observations. The solution came through the unification of the SU(2)L gauge theory and a modified version of the electromagnetic gauge theory, U(1)Y with a new gauge
36 field B0 and an associated weak hypercharge Y. With linear combinations of the fields W3 and B0 we can describe the observed γ and Z0 bosons, with the caveat that the Z0 should still be massless. Note that although the theory is unified it does not follow that the gauge couplings for individual components unite11, see figure 2.2. However, the theory gives relations between the masses of the massive bosons as well as for their gauge couplings. These relations are determined by a parameter, the weak mixing angle θw, experimentally mea- sured to be 28.74◦[2].
2.5 The Higgs Mechanism Both the force carriers and the fermions subject to the weak interaction are massive which is in tension with local gauge symmetry arguments. This means either that the underlying symmetries must be broken below some very high energy not yet accessible or that the theory is wrong. Considering the for- mer, we can introduce a scalar field, the so-called Higgs field, to the SM in a way that spontaneously breaks the underlying SU(2)L ⊗ U(1)Y symmetry of the electroweak interaction. This triggers the so-called Higgs mechanism: a beautiful way of giving mass to the boson fields W± and Z0 through their interaction with the Higgs field. In the SM spontaneous symmetry breaking can happen when a field ac- quires a non-zero vacuum expectation value. Considering a simplified case (following [12]) of a global U(1) symmetry of the Lagrangian, for a complex √1 scalar field Φ= (Φ1 + iΦ2) with the appropriate choice of potential, a so- 2 called ‘Mexican-hat’, it can be shown that we obtain an infinite set of minima along a constant radius. This is depicted in figure 2.4 where the potential is illustrated as a function of the real scalar components Φ1 and Φ2. The vacuum state is the lowest energy state of the field Φ and can be chosen in the real direction (Φ1, Φ2) = (v, 0) without loss in generality. The Lagrangian for the complex scalar field can be rewritten in terms of two new scalar fields (η and ξ) following the expansion around the minimum12. From this we can identify several interaction terms between η and ξ, but more importantly both a massive and a massless scalar field. The latter corresponds to a so-called Goldstone boson predicted by Goldstone’s theorem [17], and can be identified with excitations in the direction along the constant circle of minima. The former instead describes excitations in the radial direction corresponding to a massive gauge boson. These arguments can be promoted to hold for the local gauge symmetry U(1) by replacing the standard derivatives with appropriate covariant deriva- tives at the cost of introducing a new massless gauge field Bμ. The result is
11 The theory is described by the product group SU(2)L ⊗ U(1). 12In Quantum Field Theory (QFT) we define particles as excitations of the participating fields when expanded around the vacuum state.
37 again a massive scalar field η, a massless Goldstone boson ξ and additionally a mass term for Bμ. Hence, spontaneous symmetry breaking can give mass to the gauge boson of the local gauge theory. η can be eliminated from the La- grangian by making an appropriate gauge transformation and it can be shown that the η is “eaten” by the massive gauge boson to provide the longitudinal polarization, see e.g. [12]. For the full electroweak symmetry we introduce a weak isospin doublet of complex scalar fields φ = (φ+,φ0), called the Higgs field. The field φ has four real-valued components φi (i = 1,2,3,4). The Lagrangian is given by:
† μ 2 † † 2 LHiggs = (∂μφ) (∂ φ) − (μ φ φ + λ(φ φ) ) (2.5) where the second term is the Higgs potential. For μ2 < 0 it has an infinite set of degenerate minima. By construction, the Lagrangian is invariant under global transformations of both U(1)Y and SU(2)L. The two components in the dou- blet differ with one unit of charge just like the left-handed fermion doublets in the weak charge current interaction. Since the photon should remain mass- less, the minimum of the Higgs potential should be chosen to give a non-zero value in the neutral component φ0 only. The Lagrangian can be made symmet- ric under the full local gauge symmetry by introducing covariant derivatives i involving both Bμ and Wμ (i = 1,2,3) fields. The spontaneous symmetry breaking of the full electroweak Lagrangian re- 3 sults in the observed mixing of the Bμ and Wμ fields when considering the masses of the physical gauge bosons. Technically the breaking gives a mas- sive scalar and three massless Goldstone bosons. The latter correspond to the longitudinal polarizations of the massive electroweak bosons W± and Z0. The massive scalar is called the Higgs boson. When such a particle was discovered at CERN in 2012, it provided an important missing piece of the electroweak puzzle13. The combination of the unified electroweak theory with the Higgs mechanism is known as the Glashow-Weinberg-Salam model.
2.6 Fermion masses Fermion masses are particularly tricky in the SM since such terms are not invariant under the electroweak gauge symmetry. The Higgs mechanism gen- erates the masses of the massive weak bosons but can also give mass to the fermions. Couplings between the Higgs field and the fermion fields, con- structed through combinations of left- and right-handed particle states, can be shown to generate gauge invariant mass terms at the strength of flavor depen- dent so-called Yukawa couplings proportional to the vacuum expectation value of the Higgs potential (v ∼ 246 GeV) and the fermion masses.
13Note that the actual mass of the Higgs boson was not predicted by the theory. The window of possible masses was rather narrowed down by over three decades of experimental observations.
38 V (φ)
A
B
Re(φ) Im(φ)
Figure 2.4. An illustration of the Higgs potential in a simplified model: the classic shape of a Mexican-hat.
The coupling to a fermion with mass m f can be expressed as [12]:
m f g ∝ . (2.6) f v The couplings to the massive fermions turns out to be quite different in size. E.g. the coupling to taus are in the order of O(1) while the coupling to electrons are O(10−6). Further, without right-handed neutrinos we cannot generate mass to the neutrinos through the Higgs mechanism at all and even if such particles ex- ist, the corresponding Yukawa couplings would be 10−12. This might hint to that the neutrino masses are provided by a different mechanism. The so-called see-saw mechanism is a well-loved possibility for such, see section 2.8.2.
2.7 The Parameters of the Standard Model The expanded SM including also neutrino oscillations has 26 free parame- ters [12]. These are all measured by experiments and include the twelve (12) Yukawa couplings mentioned in section 2.6 or equivalently the fermion masses
(mν1 , mν2 , mν3 , me, mμ, mτ, mu, ...), the three (3) gauge couplings of the electromagnetic, weak, and strong interactions, the Higgs vacuum expectation value and Higgs mass, four (4) parameters (mixing angles, phases) for each of the CKM and PMNS matrices and finally a strong CP phase related to a fine-tuning problem in the theory of the strong interaction [12]. The latter is not covered in this thesis, see e.g. [12].
39 We note that the fermion masses (except for the masses of the neutrinos) are of the same order within each family and we have already seen in figure 2.2 that the coupling strengths of the gauge fields are similar in size at the electroweak scale. Indeed there seems to be patterns among these parameters and it is tantalizing to think of them as hints of physics and symmetries beyond the SM.
2.8 Beyond the Standard Model Although incredibly successful the SM is not the end of particle physics. A growing number of phenomena cannot be explained within the current frame- work but requires extensions of the SM: most notably the observation of neu- trino masses, the only firm evidence for physics beyond the SM. These topics are discussed below, in particular we discuss the notion of neutrino oscillations and means to give the neutrinos mass. The flavor sector of the SM is a source of several unexplained phenomena. E.g. there is no deeper explanation of why there are exactly three families of fermions and why the PMNS matrix is relatively flat compared to the close to diagonal CKM matrix. Further, the CP violation terms in these matrices are likely too small to explain the observed matter-antimatter asymmetry in the Universe.
Figure 2.5. Tip of the iceberg. Illustration of the energy budget of the Universe. Only about 5% consists of ordinary visible matter such as electrons, protons, and photons. About 25% consist of so-called dark matter and the remaining part of mysterious so- called dark energy. Credit: [18].
40 GUT attempt to unify the electroweak interaction with the strong interac- tion, i.e., with focus on restoring the apparently broken symmetry between colorless leptons and colored quarks [2]. The simplest GUT would be SU(5) which is the smallest Lie group that can contain all of the symmetries of the SM (U(1)Y ⊗ SU(2)L ⊗ SU(3)C) [2]. This gives rise to twelve (12) colored X-bosons with GUT scale masses, which generally cause problems with the stability of the proton, and predicts a weak angle θW in tension with obser- vations. Note that GUTs lead to a common gauge coupling for all three (3) interactions above some high energy scale O(1016 GeV) where the symmetry is broken. This unification of forces becomes even more accurate when the framework of SUSY is included. SUSY is an extension of the SM, building on the success of consider- ing symmetries as a means to understand interactions and derive conserva- tion laws. By incorporating a fundamental symmetry between fermions and bosons, it doubles the number of particles in the SM: each SM particle gets a super-particle companion differing by half a unit in spin. In particular it provides several viable dark matter particle candidates, see section 2.8.1 and references therein. Since we have not yet observed such super-particles at the same mass scale as the SM particles, this symmetry is thought to be broken at some unknown energy scale. The governing force in the Universe is gravity, but at lengths typical for particle physics and relevant at the subatomic scale, it is negligible, with a strength that is 1040 times weaker than the electromagnetic force. There is nothing a priori that says that the strengths should be different, hence this con- stitutes a so-called hierarchy problem. To solve this, some theories include extra-dimensions into which parts of the initially strong gravitational interac- tion leak. Attempts to include gravity in the SM include so-called supergravity theories where the quantum world and the general theory of relativity are com- bined into a single framework. The corresponding gauge particle is called the “graviton” but has never been observed. Another hierarchy problem arises in the Higgs sector where a large fine- tuning is required to obtain the observed bare mass O(100 GeV). The gauge and chiral symmetries of the SM prevent the spin-1 and spin-1/2 particles from getting mass (except for the dynamical creation described in section 2.5) but there is nothing that protects the spin-0 Higgs particle. Since SUSY connects a scalar to a fermion, it can provide such protection by introducing comple- mentary quantum correction terms [4].
2.8.1 The Dark Sector Physics beyond the SM is also found by studying the large-scale structures of the Universe. Measurements of velocity dispersions in clusters of galaxies [19], galaxy rotational curves [20, 21, 22], weak gravitational lensing [23], and
41 Figure 2.6. “Pandora’s cluster” - A collision of at least four galaxy clusters re- vealing a plethora of complex structures and astrophysical effects.. Credit: X- ray: NASA/CXC/ITA/INAF/J.Merten et al, Lensing: NASA/STScI; NAOJ/Subaru; ESO/VLT, Optical: NASA/STScI/R.Dupke.
the CMB [24] provide experimental evidence for a dark sector that dominates the energy budget of the Universe. The CMB reveals tiny fluctuations in the temperature (energy) of photons from different arrival directions. These photons were produced at the time of recombination about 380,000 years after the Big Bang. These fluctua- tions contain an imprint of the large-scale structures in the early Universe that started as primordial quantum fluctuations and were then amplified during an epoch of rapid expansion known as inflation and further by gravitational forces and the expansion of the Universe. The magnitude and spatial structure of the fluctuations reflect dynamics requiring much more gravitational mass than what can be provided for by the visible Universe. In particular, they show that baryonic non-luminous matter like brown dwarfs cannot make up but a fraction of the dark matter abundance. In fact, ordinary baryonic matter only accounts for about 5% [24] of the total energy density of the Universe. 26% consists of so-called dark matter and the remaining 69% seems to correspond to some kind of dark energy [24]. The two dark components are distinct in that the former originates from the observed mismatch between luminous and gravitational matter while dark en- ergy is introduced to explain the accelerating expansion of the Universe, i.e., dark matter clusters while dark energy does not. In the standard model of cos-
42 mology ΛCDM, dark energy is attributed to a non-zero so-called cosmological constant that emerges in the equations of general relativity. The concept of dark energy is beyond the scope of this thesis, but we’ll say a few words about dark matter since it is potentially closely related to the physics described by the SM. If dark matter consists of some kind of particles these could be a relic den- sity of stable neutral particles that decoupled in the early hot Universe as it expanded and cooled down (for a review of these concepts see e.g. [25]). Particles that were relativistic (non-relativistic) at the time of freeze-out con- stitute so-called hot (cold) dark matter. Hot dark matter, e.g. SM neutrinos, cannot make up a large fraction of the dark matter content since they would have ‘free-streamed’ out of small density enhancements and suppressed them. Cold dark matter is very interesting and compatible with N-body simulations in which one simulates the large scale formation of the Universe. Particular interest has been directed towards so-called Weakly Interacting Massive Par- ticles (WIMPs). These arise in a large number of theories, such as SUSY and theories with extra curled-up dimensions. Figure 2.6 shows the so-called ‘Pandora’s Cluster’14 located about 3.5 bil- lion light years from Earth. It shows the collision of at least four galaxy clus- ters in a composite image consisting of X-ray data from Chandra in red tracing the gas content (gas with temperatures of millions of degrees) and gravitational lensing data from the Hubble Space Telescope (HST), the European Southern Observatory’s Very Large Telescope, and the Subaru telescope tracing the to- tal mass concentration in blue on top of an optical image from HST and VLT. The center region shows a separation between gas and matter, where the for- mer slows down as a result of friction in the collisions of the hot gas clouds, induced by the electromagnetic force. The main fraction of the matter only seems to interact gravitationally giving empirical proof of WIMPs.
2.8.2 Neutrino masses and oscillations The neutrinos included in the SM are massless and left-handed. Historically this agreed well with observations and the latter even holds today: only left- handed neutrinos have been observed in nature. But experiments studying neutrino oscillations have confirmed that neutrinos are indeed massive albeit with a tiny mass in the O(1 eV) [2]. Neutrino oscillations have been seen in measurements of neutrinos from the Sun, from nuclear reactors and from neutrinos produced in the atmosphere. The first indication came from a deficit in the flux of solar neutrinos compared to the Standard Solar Model and was measured in Ray Davis’ Homestake Ex- periment in the late 1960s [26]. The concept of neutrino oscillations had been
14Officially known as Abell 2744.
43 put forward by Bruno Pontecorvo already in 1957 and was later revised by Pontecorvo, Maki, Nakagawa, and Sakata. The observations of the Super-Kamiokande (Super-K) detector in Japan and the Sudbury Neutrino Observatory (SNO) in the U.S., around the turn of the millennium, provided the first evidence for neutrino oscillations. For this dis- covery, Takaaki Kajita and Arthur B. McDonald, were rewarded the Nobel prize in physics 2015. Considering vacuum mixing between two flavors: the probability to mea- sure a certain flavor νβ given να is a functions of the total particle energy E and time t. Since neutrinos are close to relativistic we can parametrize t in terms of distance travelled L. The probability is given by [27]: Δ 2 2 2 2 m [eV ]L[km] P(να → νβ) = sin (2θ)sin 1.27 , (2.7) E[GeV] Δ 2 = 2 − 2 θ where m m2 m1 and is the mixing angle. This result can be generalized to three flavors by considering the full PMNS matrix [27]. The results of oscillations with three flavors in vacuum is shown in figure 2.7 given oscillation parameters consistent with current measurements, see figure caption. The plots show the probability for a muon-neutrino to ap- pear in any of the three neutrino flavors as a function of the ratio L/E. The upper plot show a zoom in on small values of L/E relevant for atmospheric oscillations while the lower plot show a larger range relevant for solar oscil- lations. The effect of neutrino oscillations is only relevant for macroscopic distances and can safely be ignored at the small distances typical for e.g. col- lider experiments, see equation 2.7. The oscillation probability is dramatically increased if we consider neutrino propagation in matter through the so-called MSW (Mikheyev-Smirnov-Wolfenstein) effect [27]. In particular this effect was crucial for understanding and solving the solar neutrino problem, a deficit in the number of electron-neutrinos arriving at Earth. In order to give mass to the particles through Yukawa couplings to the Higgs field we need terms with mixed Dirac fermion fields (right- and left-handed). ψ = ,μ,τ But even if we were to include right-handed neutrino fields ( νl )R (l e ) (or the corresponding left-handed antineutrino fields) in the SM the Higgs cou- plings would have to be extremely small to explain the small neutrino masses. These neutrinos would only interact gravitationally and are therefore referred to as ‘sterile’ neutrinos15. Sterile neutrinos can be searched for in neutrino oscillation experiments, where they would mix with the left-handed particles. That effect is particularly strong for propagation in a dense medium such as the Earth [27]. Right-handed neutrinos and left-handed antineutrinos are invariant under the SM gauge symmetries, why we can add terms to the Lagrangian formed
15Technically the term ‘sterile’ neutrino is used to distinguish any non-interacting neutrino from the ‘active’ neutrinos that participate in the SM.
44 Figure 2.7. Probability for a muon-neutrino to appear in any of the three neutrino flavors as a function of the ratio L/E. The upper plot show a zoom in on small values of L/E relevant for atmospheric oscillations while the lower plot show a larger range relevant for solar oscillations. The black line illustrates the probability to observe the particle as an electron-neutrino, blue a muon-neutrino, and red a tau-neutrino. The 2 oscillation parameters used are consistent with current measurements (sin θ13 = 0.10, 2 θ = . 2 θ = . δ = Δ 2 = . · −5 2 Δ 2 ≈ Δ 2 = sin 23 0 97, sin 12 0 861, 0, m12 7 59 10 eV , and m32 m13 2.32 · 10−3 eV2). Normal hierarchy is assumed.
45 by their fields without breaking the local gauge invariance. In fact, the full neutrino mass Lagrangian can be written as a combination of Dirac (mass terms generated by the spontaneous symmetry breaking) and so-called Majo- rana mass terms. The Majorana term can in principle lead to processes with total lepton number violation that could be searched for in experiments, but since the helicity and chirality is almost identical for neutrinos such processes are suppressed [12]. Instead experiments focus on observing neutrinoless dou- ble β-decays, a process that is only possible for so-called Majorana neutrinos, neutrinos that are their own antiparticles. The Lagrangian is defined as: ⎛ ⎞⎛ ⎞ 1 ⎜ 0 m ⎟⎜ νc ⎟ L = − ν νc ⎜ D ⎟⎜ L ⎟ + . . DM L R ⎝ ⎠⎝ ⎠ h c (2.8) 2 mD M νR
= 1 ± + 2 / 2 ν where the eigenvalues are m± 2 (M M 1 4mD M ) and L,R are the left- νc and right-handed neutrino states respectively and R/L is the CP conjugate of νR/L [12]. If mD would be of the same order as the lepton masses and M mD, then m− ≈ mD · (MD/M) and m+ ≈ M, i.e., m− gets suppressed by the scale of M. This way of giving the observed neutrinos light masses is called the see-saw mechanism and typically predicts extremely high masses (in GUT M ∼ 1015 GeV) for the other eigenstate, far above the reach of current experi- ments [12]. If Majorana mass terms exist, the mechanism predicts one massive neutrino O(M) for each light SM neutrino O(1 eV).
Summary After this review of the SM and in particular of some of the particle physics aspects of the neutrinos, we are ready to introduce neutrinos in a wider context; namely to discuss the important role of neutrinos in the determination of the origin and production of cosmic rays.
46 3. The Cosmic Ray Puzzle
“Thus it is possible to say that each one of us and all of us are truly and literally a little bit of stardust.” William A. Fowler The Quest for the Origin of the Elements, 1984
The Earth’s atmosphere is constantly bombarded by charged particles and nuclei, so-called Cosmic Rays (CRs), at the astonishing rate of 1,000 particles per square meter per second [28]. These are primarily protons and helium nuclei with a small contribution from electrons. In general these particles are relativistic but they arrive with a huge variation of kinetic energies, the most energetic ones with an energy as large as that of a brick falling from a roof. The highest energies observed so far belongs to so-called Ultra High-Energy (UHE) CR that have an energy of O(1020)eV [29]. When CRs where discovered in the early 1910s they provided the first evi- dence of a more exciting Universe consisting of something more than the stars and gas observed in optical telescopes. The discovery was made through a series of high-altitude balloon flights by Victor Hess (1911-12, < 5.3 km) [30] and subsequently by Werner Kolhörster (1913-14, < 9 km) [30], both measur- ing an increasing flux of ionizing particles with increasing altitude, thus ruling out the Earth itself as a source. Their discovery became one of the biggest breakthroughs in particle physics and started an era of CR experiments that would lead to the discovery of many new particles, most notably the positron (anti-electron) in 1932 [31], the muon in 1936 [32], the pion in 1947 [30], and later kaons and lambdas that were the first particles discovered that contained the strange quark [30]. The majority of the CRs, were confirmed to have positive charge which was illustrated by studies of the so-called east-west effect, the result of deflection in the geomagnetic field [33, 34, 35]. It was also realized in the late 1930s (Rossi, Auger, Bethe, Heitler) that the particles we observe at the ground are in fact so-called secondaries, produced by primary CR interactions with air nuclei causing electromagnetic and hadronic cascades, that propagate like showers in the atmosphere, see figure 3.1.
47 Figure 3.1. Artistic impression of a CR entering Earth’s atmosphere where it interacts with air nuclei to form electromagnetic and hadronic showers. Credit: CERN/Asimmetrie/Infn.
Countless experiments have been built since then and they have provided us with many new clues about the nature of CRs, in particular about their com- position and propagation through the ISM. But the most important questions remain partly unanswered: the origin and the production mechanisms are still to a large extent unknown. Two important features of the CR data that can be used to learn more about their origin is their composition, i.e., the relative abundance of different nu- clei, and their energy spectra. While the latter may be a characteristic of a particular acceleration mechanism, the former can be used to infer what kind of sources are involved in the acceleration process. This is done by compar- ing the composition models and data from different astrophysical objects [28]. Further, the identification of a localized source of CRs, either direct of indirect, may provide the most important clue, in particular if the CRs are observed in connection with known emitters of electromagnetic radiation such as X-rays, γ-rays, etc. Galactic CRs are usually modeled by the so-called leaky box model. It assumes that CRs are confined by magnetic fields within the galactic disk (∼ 10−7 G) but gradually leak out. Since the resulting gyro-radii of CRs be- low 1015 GeV (‘the knee’) are much smaller than the size of the Galaxy, they become trapped for times on the order of 106 years [30]. Scattering on inho- mogeneities in the magnetic fields randomizes their directions and leads to a high degree of isotropy, hence they will have no memory of their origin when reaching Earth. Since CRs above the knee cannot be contained by the galactic magnetic fields they are not trapped for long enough time to accelerate to high energies.
48
Figure 3.2. CR anisotropy observed by IceCube using 6 years of data. The plots are preliminary and show the relative intensity in equatorial coordinates. The median energy of the primary CR is shown in the upper left corner of each plot. Each map was smoothed using a 20◦ radius. Interestingly the deficit and excess switch positions as we go above ∼ 100 TeV. The reason for this is unknown. Figure is taken from [36].
They are assumed to be extragalactic of origin or alternatively associated with strong local sources. Anisotropies in the arrival directions of CRs can be introduced either through large magnetic fields with significant deviation from an isotropic diffusion (through modified propagation models), by local sources, or by signifiant es- cape from the Galactic disk1. The predicted anisotropy from these effects features large regions of relative excess and deficit of number of events, with amplitudes on the order of 10−2 to 10−5 [38]. Further there are effects due to the motion of the observer relative to the CRs, e.g. due to the motion of
1At TeV energies propagation effects could give rise to large-scale anomalies [37], but it is not clear if these can also explain observations made at higher energies.
49 the Earth relative to the Sun, and potentially a so-called Compton-Getting ef- fect [39], a dipole anisotropy due to the relative motion of the solar system with respect to the rest frame of the CRs2. However, the predicted ampli- tude and phase of the latter is not consistent with data from current air shower experiments [40], i.e., if the Compton-Getting effect is present in data it is overshadowed by stronger effects and may be one of several contributions to the CR anisotropy. Indeed, observations of CR arrival directions show a nearly isotropic dis- tribution at most energies [2]. However, energy dependent anisotropies at the level of ∼ 10−3 have been observed by several experiments amongst others: IceCube [40] and Milagro [41] (for median primary CR energies of 6 TeV). Figure 3.2 show the CR anisotropy observed by IceCube using 6 years of data. The plots are preliminary and show the relative intensity in equatorial coordi- nates for three different median energies. Interestingly the deficit and excess switch positions as we go above ∼ 100 TeV. The reason for this is still un- known. In the following sections we will focus on the observed energy spectra and the possible acceleration mechanisms and sources involved in the production of high-energy CRs. In particular we will investigate the close link between the acceleration of high-energy CRs (> 100 MeV) and the production of high- energy astrophysical neutrinos in section 3.2. For a more complete review of CR, see e.g. [28, 30, 38].
3.1 Energy Spectrum One of the most striking facts about CRs is their enormous span in energy and flux. The spectrum covers several decades in energy, from 108 to 1020 eV, and flux, from 104 to 10−27 GeV−1 m−2 s−1 sr −1. The spectrum approximately3 follows an unbroken power-law E−γ with a slope γ ≈ 2.7 above 1010 eV. Fur- ther it has several interesting features; a knee around 1015 − 1016 eV, a second knee around 1017 eV, an ankle around 1018 − 1019 eV and finally a sharp cut- off above 1020 eV. Figure 3.3 shows the flux of CRs multiplied with E2 as a function of kinetic energy. Further, on the horizontal axis, it illustrates the typical collision energy of man-made accelerators such as the LHC at CERN and Tevatron at Fermilab (Fermi National Accelerator Laboratory). Above the knee we observe roughly 1 particle per m2 per year, while above the ankle we only observe 1 particle per km2 per year. The spectral features can be seen more clearly in figure 3.4 that shows the CR flux multiplied by E2.6. This represents the differential energy spectrum as a function of reconstructed energy-per-nucleus, from direct measurements
2Note that the rest frame of the Galactic CRs is not known. 3The spectral index depends on the compilation of data, in particular the different composition models assumed [29].
50 Figure 3.3. CR particle flux as a function of energy. Credit: Bakhtiyar Ruzy- bayev/University of Delaware.
below 1014 eV and from air-shower events above 1014 eV. The lowest energy CRs can be detected directly, while the flux is so low at high-energies that a direct measurement is no longer feasible and we instead study air-showers produced when CRs interact in the atmosphere. Air-showers are detected using mainly three different techniques: number density and lateral distances of charged particles on the ground are measured in air-shower arrays, Cherenkov radiation from the charged particles as they propagate in the atmosphere, is measured in air Cherenkov detectors, and fluo- rescence detectors study Ultra-violet (UV) light created as CRs interacts with primarily nitrogen in the atmosphere. In modern facilities these techniques are usually combined, see e.g. the Pierre Auger Observatory (PAO) [42] utilizing 1,600 surface Cherenkov detectors combined with 24 fluorescence detectors, together covering a detection area of 3,000 km2. At low energies, E < 10 GeV, the CR spectrum is largely influenced by the solar wind consisting of electrons and protons with energies in the range
51 1 < E < 10 keV [2, 29]. At even lower energies the spectrum is modulated by Earth’s geomagnetic field. Further, in the event of a solar flare, the spectrum is dominated by particles from the Sun in the range of a few tens of keV to few GeV. The low-energy CR abundance is hence largely dependent on both location and time. While CRs with energies between a few keV to a few GeV are known to be created in large amounts in the Sun, the birthplace of the high-energy CRs (a few GeV to 1020 eV) is still under debate. Indeed, the origin of CRs is a rather complex question mainly because many different processes are likely at play simultaneously throughout the full energy spectrum. One of the key questions is to figure out whether the spectral features are caused by processes happening during creation (acceleration) or propagation through the Universe. Typically we associate events to the left of the knee to the Galaxy. We will see in section 3.5 that shock acceleration in Supernova Remnants (SNRs) can explain particle energies up to PeV (∼ 1015 eV) energies. Further, events to the right of the knee are thought to show the onset of an extragalactic component. Objects such as AGNs and Gamma-Ray Bursts (GRBs) have the necessary power to provide the energy in this region.
Knee 104
Grigorov 2nd Knee ]
-1 JACEE MGU sr 3 -1 10 Tien-Shan s
-2 Tibet07 Ankle
m Akeno 1.6 CASA-MIA 102 HEGRA
[GeV Fly’s Eye Kascade
F(E) Kascade Grande 2.6 IceTop-73 E 10 HiRes 1 HiRes 2 Telescope Array Auger 1 1013 1014 1015 1016 1017 1018 1019 1020 E [eV]
Figure 3.4. CR particle flux as a function of energy-per-nucleus. The data points show the all-particle flux from a number of air-shower experiment. Figure from [2].
52 3.2 The Origin of High-Energy Cosmic Rays Technically CRs are divided into different categories based on their production site, but this division is not sharp and often not decisive4. We will use the following definition: primary CRs denote particles that are accelerated in astrophysical sources: protons, electrons, and nuclei such as helium, carbon, oxygen, and iron. Secondary CRs denote particles that are created through interactions of primary CRs with the gas and dust in the ISM, in the Earth’s atmosphere or in regions close to the accelerators themselves, so-called astrophysical beam dumps. When comparing the composition of CRs with the abundance of elements in the Solar system, we see two main differences. In general there is an en- richment, in the CRs, in the heaviest elements relative to H and He. This is not entirely understood but could be due to a different composition at the sources [28]. Further, we observe two groups of secondaries - Li, Be, B and Sc, Ti, V, Cr, Mn - having a much larger relative abundance among the CRs compared to the relative abundance in the solar system. These two groups of elements are rare end-products of stellar nucleosynthesis, but are present in the CR flux as a result of so-called spallation of primary CRs in the ISM5. Further, the ratio of secondary to primary nuclei is observed to decrease with energy [2]. This is interpreted in terms of high-energy particles escaping the Galaxy to a higher degree. These observations give us a handle on the parameters in propagation models and the lifetime of the CRs in our Galaxy. Another source of secondary CRs is the production of electron-positron pairs, that becomes relevant for protons above 1 EeV, and further dominates the energy loss up to around 70 EeV where pion-production takes over. CRs with energies above 50 EeV may interact with Cosmic Microwave + Background (CMB) photons to produce pions through p + γCMB → Δ → n(p)+π+(π0). This provide the mechanism for the so-called Greisen-Zatsepin- Kuzmin (GZK) cut off [47, 48], potentially the reason for the sharp cut-off in the energy spectra of CRs around 50 EeV [49, 50]. The neutrinos produced through the decay of these charged pions are very highly energetic and are referred to as GZK-neutrinos. The GZK flux prediction is sensitive to the fraction of heavy nuclei present in the UHE CRs. These lose energy through photo-disintegration and hence lower the neutrino production efficiency [51]. The search for GZK-neutrinos has been conducted with a large variety of ex- periments: large radio telescopes such as e.g. LUNASKA (The Lunar UHE
4E.g. the excess of positrons measured by PAMELA (Payload for Antimatter Matter Explo- ration and Light-nuclei Astrophysics) and later confirmed by Fermi and the AMS-01 (Alpha Magnetic Spectrometer)-01 and AMS-02 experiments at the International Space Station (ISS) has been interpreted as evidence for dark matter [43], in terms of modified propagation models [44], and as evidence for a nearby astrophysical source [45]. See [29] and references therein for a brief review of the topic. 5In this context spallation refers to the process in which nuclei emit nucleons after being hit by incoming high-energy particles.
53 Neutrino Astrophysics using the Square Kilometer Array) searching for radio pulses from neutrino-induced particle cascades in the Moon [52], balloon ex- periments such as the ANITA (Antarctic Impulse Transient Antenna), satellite experiments such as FORTE (Fast On-orbit Recording of Transient Events) [53], etc. Future detectors include the ARA (Askaryan Radio Array) [54, 55] and the ARIANNA (Antarctic Ross Ice Shelf Antenna Neutrino Array) [56]. These all search for radio signals created by the so-called Askaryan effect [57]. Further, the SPATS (South Pole Acoustic Test Setup) studies acoustic signals formed in dense neutrino-induced cascades in the Antarctic ice, see e.g. [58] for a review on this topic. In section 3.3 and 3.4, we will show how secondary CRs are produced in re- gions close to astrophysical sites and the relatively dense environments of the Earth’s atmosphere. These sites also produce high-energy neutrinos, the main topic of this thesis. Figure 3.5 illustrate the sources of neutrinos through-out the Universe. These include both natural sources such as the hypothetical relic neutrinos from the Big Bang (’Cosmological ν’), solar neutrinos, atmospheric neutrinos and neutrinos produced in violent processes related to supernova bursts and AGNs, as well as man-made sources, such as reactor neutrinos produced in radioactive decays of heavy isotopes used for fission energy gen- eration.
Figure 3.5. Measured and expected neutrino fluxes from several sources, both natural and man-made. The incredible energy range covers both time-dependent and time- independent sources from μeV all the way up to EeV. The figure is taken from [46] with kind permission from Springer Science and Business Media.
54 Figure 3.6. Cross-section for photo-production of neutral and charged pions. The single-pion production (black solid and red dashed line) is dominated by the resonance + of Δ at mΔ ≈ 1232 MeV [60]. Figure from [60].
3.3 Astrophysical Neutrinos Astrophysical neutrinos can be produced as a result of interactions between accelerated primary CRs and the matter and radiation fields that surround the acceleration site. The important processes are the hadronic reactions6: p + γ → Δ+ → n(p) + π+(π0) (3.1a) p + p → p + n(p) + π+(π0) (3.1b) where equation 3.1a is commonly referred to as a photo-production process where the protons interact with ambient photons, typically produced by ac- celerated electrons. Photo-production is the foundation of the so-called Δ- approximation, see figure 3.6, where we assume that the proton-γ cross-section can be described by an analytic Breit-Wigner resonance at mΔ ≈ 1232 MeV [60]. Equation 3.1b is referred to as an astrophysical beam dump process due to the similarity with accelerator-based experiments [29]. The balance between these reactions is related to the cross-section and number density of each particle present [2]. + + Neutrinos are produced in the decay of charged pions, e.g., π → μ + νμ + + (99.99%) followed by μ → e +νe +ν¯μ (≈100%). The typical energy for each of the three neutrinos is: Eν ∼ 1/4Eπ ∼ 1/20 Ep, i.e., protons at the CR knee creates ∼ 100 TeV neutrinos [61]. γ-rays are produced in the decay of neutral pions π0 → γ +γ (98.82%). Further, escaping neutrons may decay through the
6This set of reactions can also proceed with neutrons and/or higher mass mesons and baryons in the final states. Further, see [59] for a review of channels with resonances.
55 − weak interaction n → p + e + ν¯e (100%). All branching ratios are taken from [2]. The energy emission of astrophysical sources consists of escaping high- energy protons, γ-rays, and neutrinos, all with highly correlated continuous non-thermal energy spectra. The pion spectrum from photo-production, and hence the neutrino spectrum, follows closely the primary proton spectrum in case the photons are thermal, but is somewhat steeper if the process proceeds with synchrotron radiation, produced by electrons loosing energy in strong magnetic fields [62]. Note that these arguments require that the sources are transparent enough for protons to interact at least once before decaying and that mesons decay before interacting. For optically thin sources photons are emitted in coincidence with the neutrinos, but for thick sources they avalanche to lower energies until they escape, giving rise to a distribution of sub-TeV photons. In models one typically normalize the expected neutrino flux to the observed flux of γ-rays or CRs given assumptions of the production efficiency and absorption in the sources [63]. The neutrino flux for a generic transparent source of extragalactic CRs is called the Waxman-Bahcall (WB) limit or WB bound [64, 65]. By integrat- ing the energy spectrum from figure 3.4 above the ankle assuming an E−2 spectrum and a GZK cut-off at 50 EeV we arrive at a CR energy density ρ ∼ 3 × 10−19 erg cm−3 [63] or ∼ 3 × 1037 erg s−1Mpc−3 over the Hubble time 1010 years. Converted to power for different populations of sources this gives [66]: • ∼ 3 × 1039 erg s−1 per galaxy, • ∼ 2 × 1042 erg s−1 per cluster of galaxies, • ∼ 2 × 1044 erg s−1 per AGN, or • ∼ 2 × 1052 erg per cosmological GRB. These numbers are of the same order of magnitude as the observed electromag- netic emission from the listed sources [63]. Further, they are consistent with a transparent model where the most important process is the photo-production through the Δ-resonance and the energy emission is approximately equally dis- tributed among the CRs, γ-rays, and neutrinos. These sources have emerged as leading candidates for the extragalactic component of the CRs and the derived neutrino flux is loosely confined in the range [63]:
2 −8 −2 −1 −1 EνdΦWB/dEν = 1 ∼ 5 × 10 GeV cm s sr , (3.2) where ΦWB is the WB bound. The precise value depends on a number of assumptions: the minimum energy considered in the integration of the CR spectrum, details of the production process (in particular the energy distribu- tions), cosmological evolution of the CRs sources, etc. For details see [63], [59] (Δ-approximation), [67] (simulations of pγ processes), and [68] (simula- tions of pp processes). The Δ-approximation is crucial in order to construct
56 approximate pion-production cross-sections and to determine gross features like the number ratio of γ to neutrinos produced [59]. A similar exercise can be done for the Galactic part of the CR energy spec- trum, leading to an energy density ρ ∼ 10−12 erg cm−3 (∼ 1eVcm−3) [28, 63]. Assuming a volume, corresponding roughly to the size of the Galactic disk, V = πR2d ∼ π(15 kpc)2(200 pc) ∼ 4 · 1066 cm3 we can derive the power re- quired to accelerate these galactic CRs [28]: ρ = V ∼ · 40 −1, W τ 5 10 erg s (3.3) where τ is the active time in the source region, typically of the order of 106 years [28]. Interestingly this power estimate is similar to the power provided by typical supernova scenarios, if we assume a supernova rate corresponding to one every 30 years and that roughly 10% of the power produced by super- novae, each releasing 1051 erg, can be used to accelerate new particles through the expanding shock wave [63]. The predicted neutrino rate give a handful of detectable neutrinos per decade of energy per km2 and year [63], assuming that the sources extend to 100 TeV with an E−2 spectrum. The relative ratio of different neutrino flavors produced according to equa- tions 3.1 is 1:2:0 for νe:νμ:ντ. Due to flavor oscillations during propagation this implies that the astrophysical neutrinos arriving at Earth should have the ratio 1:1:1 given the so-called long baseline approximation [69, 70]. The observation of localized neutrinos from known γ-ray emitters would undoubtedly confirm the existence of hadronic accelerators. However, re- cently the Fermi collaboration published results identifying hadronic accel- erators solely based on the presence of γ-rays produced from the decay of π0 in the Spectral Energy Distribution (SED) [71]. Assuming an E−2 spectrum the pion-decay bump is characterized by a sharp rise at 70-200 MeV, tracing the parent population above ∼ 1 GeV [72]. The height depends on the maximal proton energy and the fraction of energy transferred to the γ-rays. The standard model for production of high-energy photons is the so-called Self-Synchrotron Compton (SSC) mechanism, see e.g. [73]. This constitutes the basis for leptonic production where high-energy photons are produced through so-called synchrotron emission, caused by the movement of electrons in the strong magnetic fields of jets. The generated photon spectrum is peaked in the IR to X-ray region, and constitute the so-called synchrotron peak. The synchrotron photons subsequently get accelerated by the same parent electron distribution through Inverse Compton (IC) scattering. Further, high-energy photons are also generated through Bremsstrahlung as charged particles are decelerated. These latter two contributions result in photon energies in the range GeV-TeV. The non-thermal emission spectra is hence characterized by a double peak distribution, see e.g. figure 3.7. Note that the SSC model has an intrinsic energy limit set by the decrease of the IC cross-section in the Klein-Nishina regime [74]. This is in contract to the hadronic scenarios where
57 IC 443 10-10 ) -1 s -2
10-11
dN/dE (erg cm Best-fit broken power law 2 Fermi-LAT E VERITAS (30) -12 MAGIC (29) 10 AGILE (31) π0-decay Bremsstrahlung Bremsstrahlung with Break
108 109 1010 1011 1012 Energy (eV)
) 0 decay
-1 π
s Mixed model -2 10-10 Inverse Compton Bremsstrahlung Bremsstrahlung with Break
10-11 dN/dE (erg cm 2
10-12
Gamma-ray flux E 10-13 10-7 10-5 10-3 10-1 10 103 105 107 109 1011 1012 Energy (eV)
Figure 3.7. Spectral Energy Distribution (SED) of SNR ‘IC 443’. The figures show best-fit values to Fermi-LAT data (circles) of both leptonic and hadronic components (the radio band is modeled using synchrotron emission), the upper figure showing a zoom in on the γ-ray band. The solid lines represent the best-fit pion-decay spectrum, and the dashed lines show the best-fit Bremsstrahlung spectrum. The inverse Compton component is shown as a long-dashed line in the lower plot, together with a mixed model (Bremsstrahlung and pion-decay), shown as a dotted line. See details in legend. Figure taken from [71]. Reprinted with permission from AAAS.
58 Figure 3.8. Best-fit neutrino spectra from IceCube global fit paper [77], assuming a single power-law and including all flavors. The red (blue) shaded are correspond to 68% C.L. allowed regions for the astrophysical (atmospheric conventional) flux component, while the green line represent the 90% C.L. upper limit to the prompt atmospheric flux fitted to zero. Figure taken from [77].
the maximal energy is set by the maximum proton acceleration energy of the source. The leptonically produced photons can outnumber the photons from the hadronic production by large amounts, in particular for sources with a very low matter density around the acceleration region. Sources known to be inter- acting with clouds of dense material therefore provide good targets for where to detect evidence of CR acceleration. SNRs interacting with molecular clouds belong to this class of interesting sources, producing γ-ray via the decay of neutral pions produced in the dense gas surrounding the sources [75]. One of the best examples of such SNR-cloud interactions in the Galaxy is SNR IC-443, with an estimated age of about 10,000 years, located at a distance of 1.5 kpc [76]. The plots in figure 3.7 show the measured energy spectrum for SNR ‘IC 443’. The circles show the best-fit values to Fermi-LAT data of both leptonic and hadronic components. The solid lines denote the best-fit pion- decay γ-ray spectra. See further details in figure caption. The conclusion from this search was that the γ-rays are most likely produced through the hadronic channels, while leptonic production can still account for the synchrotron peak. These models predict a neutrino luminosity that can be searched for by neu- trino detectors like IceCube.
59 The IceCube Collaboration reported the detection of a diffuse flux of astro- physical neutrinos in 2013 [3, 78]. Evidence for astrophysical neutrinos were also seen in a separate analysis looking for a diffuse astrophysical neutrino flux using tracks from the Northern hemisphere [79]. The best-fit all-flavor as- trophysical neutrino flux calculated using a global fit to six different IceCube searches is given as [77]:
φ = . +1.1 · −18 −1 −2 −1 −1 . (6 7−1.2) 10 GeV cm s sr (at 100 TeV) (3.4) where γ = 2.50 ± 0.09 [77] (valid for neutrino energies between 25 TeV and 2.8 PeV [77]). The atmospheric-only hypothesis was rejected with 7.7 σ as- suming a χ2-distribution. Apart from a statistically weak cluster close to the Galactic center, the arrival directions of these neutrinos are consistent with an isotropic distribution, but the actual origin is unknown. Since the events ex- tend to large Galactic latitudes many suggestions include extragalactic sources such as AGNs embedded in molecular clouds, intense star-formation regions, etc. See [80] for a review and references therein. Note that since high-energy photons have a short absorption length such emission is not to be expected in association with the IceCube flux, unless there is a significant Galactic contri- bution [61]. Instead, we can search for association with the possible sub-TeV extension of the signal, as mentioned in the beginning of this section.
3.4 Atmospheric Backgrounds Atmospheric muons constitute the by far most dominant part of the event yield in large-volume underground particle detectors such as IceCube. Analyses searching for astrophysical neutrino-induced muons starting inside of the de- tector are particularly plagued by atmospheric muons mimicking truly start- ing events. Further, most analyses also struggle with a close to irreducible background of atmospheric neutrinos. The mimicking muons are particularly troublesome for searches for low-energy neutrinos, relevant for this work, or the diffuse astrophysical flux, but less so when searching for clusters of local- ized high-energy neutrinos. In this section we discuss these backgrounds and the important interplay between decay and interaction that takes place in the Earth’s atmosphere. The interactions of high-energy nuclei and protons in the Earth’s atmo- sphere give rise to electromagnetic and hadronic air-showers, see illustration in figure 3.9. The CR flux through the atmosphere can be approximated with a set of coupled cascade equations [2], but Monte Carlo simulations are needed to accurately account for the decay and interactions of the secondary particles as well as the spectral index of the primary CRs.
60 The main reactions generating muons and neutrinos are: p + N → π±(K±) + X ± → μ + νμ(¯νμ) (3.5) ± → e + νe(¯νe) + ν¯μ(νμ), where p is a proton, N is the target nucleus, and X represents the final state hadron(s). The decays of hadrons induce both electromagnetic cascades of high-energy photons, electrons, and positrons as well as highly penetrating muons and neutrinos. The balance between interaction and decay in the atmosphere is energy de- pendent. The critical energy is defined as the energy where the interaction probability equals the decay probability, under the assumption of an isother- mal atmosphere. The so-called conventional flux of muons and neutrinos is primarily produced in decays of π± or K±7, with critical energies of 115 GeV and 855 GeV [81], respectively. Above the critical energy these particles tend to interact in the atmosphere before they decay, why a steepening (softening) of the muon spectra with about one power relative to the primary spectrum of CRs, is observed above these energies, i.e., γconv. ∼ 3.7 [28]. Note that this effect is less pronounced for zenith angles close to 90◦ since the mesons then travel longer in the low density of the upper atmosphere where they do not interact [2]. The flux from semi-leptonic decays of short-lived charmed hadrons (generally heavy quark particles with lifetimes smaller than 10−12 s) is referred to as prompt and is characterized by a flatter muon spectrum, i.e., relatively more muons at higher energies. This is because the charmed hadrons, in contrast to the conventional component, in general decay before losing energy in interactions. E.g., D± has a critical energy of 3.8 · 107 GeV [29]. Current estimates of the cross-over, from the conventional to the prompt component, predict that it occurs ∼ 1 PeV, see e.g. [82]. Consequently the muon and neutrino spectra below O(100 GeV) where decay dominate follow the CR spectrum and are well described by a power-law spectrum: dΦν ∝ E−γ, (3.6) dEν where γ = 2.7. Above O(100 GeV) the spectra become steeper eventually reaching γ = 3.7. While the conventional production, described in equation 3.5, produces neutrinos and anti-neutrinos with flavor ratio νe:νμ:ντ = 1:2:0, the prompt com- ponent yields approximately equal amounts of electron and muon flavored 8 9 neutrinos . Further the ratio of νe/νμ decreases with energy above O(GeV) ,
7The fraction of pion- relative to kaon-induced muons is energy dependent. About 8% (20%) of vertical muons come from decay of kaons at 100 GeV (1 TeV) (27% asymptotically) [28]. 8The large mass of τ suppresses the production of ντ andν ¯τ from atmospheric processes. 9This is in contrast to the lower density environments considered in section 3.3 where the ratio stays constant to much higher energies [28].
61 Figure 3.9. Production of muons and neutrinos in the Earth’s atmosphere.
62 due to the increasing muon decay length [28]. Further, there is an asymmetry in the yield of neutrinos compared to anti-neutrinos. This is due to the excess of positively charged pions and kaons in the forward fragmentation region of proton-initiated interactions, and further due to the larger amount of protons over neutrons in the primary CR spectrum.
Particle Flux at Sea Level Since muons are relatively long-lived (livetime 2.2 μs [2]) some penetrate to sea level and further through kilometers of Earth material such as ice, rock, and water. The muon flux on the surface is around 100 particles m−2s−1sr−1. The abundance of particles in the air-shower as it moves through the atmosphere is shown on the left side of figure 3.10. Muons and neutrinos dominate the picture at sea level and any remaining electrons, hadrons, or photons will be absorbed in the Earth material. At the depths of IceCube the muons outnumber the atmospheric neutrinos by a factor of ∼ 106. The plot on the right side of figure 3.10 shows the vertical intensity of muons as a function of depth expressed in kilometers water equivalent (km w.e.). IceCube is located at a depth of about 1.8 km w.e. and is subjected to a large number of atmospheric muons from the Southern hemisphere. An efficient way to reduce this flux is to only study particles from the Northern hemisphere, where the Earth itself can be used as a filter, blocking most muons. At about 20 km w.e., the muon flux becomes constant, constituting an isotropic neutrino-induced background.
3.5 Acceleration Mechanisms The power-law nature of the CR energy spectrum discussed in section 3.1 is indicative of non-thermal10 processes where a relatively small number of par- ticles is accelerated by a focused energy out-flow from a powerful source [30]. In this section we will discuss mechanisms needed to transfer this macroscopic energy to individual particles, in particular the so-called standard model of CR acceleration: Diffusive Shock Acceleration (DSA). The acceleration of high-energy particles in the Solar system occurs both on large scales, e.g. in interplanetary shock waves associated with the solar wind, and in the vicinity of point sources such as the particle acceleration to GeV en- ergies observed in connection with solar flares [28]. Similarly, both extended and point sources are likely to play a role in the acceleration of Galactic and extragalactic CRs. It is also possible that acceleration (initial or further) takes place during propagation, through interactions with large gas clouds containing magnetic irregularities. In general, acceleration to high-energies (TeV energies and above) requires a massive bulk flow of relativistic charged particles [63]. The
10Emission not described by a black body spectrum with a given temperature.
63 Altitude (km) 15 10 5 3 2 1 0 10000
1000
] _ –1 νμ + νμ sr 100 μ+ + μ− –1 s
–2
10 p + n 11250
1 e+ + e− Vertical flux [m π+ + π− 0.1
0.01 0 200 400 600 800 1000 1 10 100 Atmospheric depth [g cm–2]
Figure 3.10. Left: Vertical CR fluxes in the atmosphere above 1 GeV (displays the integrated flux above 1 GeV for all particles except electrons, for which the threshold Ee > 81MeV applies instead) as a function of atmospheric altitude/depth. Data points show experimental measurements of μ−. Figure taken from [2]. Right: Vertical muon intensity as a function of kilometer water equivalent depth. For data points see caption of figure 28.7 in [2]. The shaded area at large depths shows neutrino-induced muons with an energy above 2 GeV. Figure taken from [2]. presence of relativistic gas in the ISM has been established through mainly two key observations: synchrotron radiation induced by relativistic electrons (and protons) and the presence of γ-rays from the decay of neutral pions. Earth-bound accelerators use electric fields to give energy to particles typ- ically confined in circular orbits by strong magnetic fields. But in most as- trophysics environment, such electric fields would immediately short-circuit since the particles are in a plasma state with very high electric conductiv- ity [29]. Instead, we consider so-called shock acceleration where low-energy particles are given higher energies through repeated stochastic encounters with magnetic irregularities. We start by deriving the basic concept of shock acceleration, closely fol- lowing the discussion in [28]: A test particle undergoing stochastic collisions gains an energy ΔE = ξE in each encounter. After n such collisions the en- n 0 ergy is En = E0(1+ξ) , where E is the initial (injection) energy. Assuming an escape probability, Pesc, from the finite acceleration region per encounter, the proportion of particles with an energy greater than E is [28]: