Analysis of Gamma Ray and Neutrino Point Sources with IceCube

IceCube South Pole Neutrino Observatory

Greg Vance

Connecticut College

gvance@conncoll.edu

REU Program — Summer 2013

University of Wisconsin–Madison

Madison, Wisconsin 53706


Background and Introduction | Hadronic Gamma Ray Production | Constraining Gamma Ray Sources | Summary of Results | References and Acknowledgments | Related Links

South Pole IceCube Lab
South Pole IceCube Lab. Photo credit: NSF/F. Pedreros

Background and Introduction

IceCube South Pole Neutrino Observatory

IceCube Diagram
A diagram of IceCube.
IceCube Neutrino Observatory is an array of over 5,000 basketball-sized digital optical modules (DOMs) buried deep in the Antarctic ice at the South Pole. The DOMs are designed to detect weakly interacting particles called neutrinos, which can originate from a variety of different astronomical sources. When a neutrino interacts with the ice at the South Pole, a small burst of light called Cherenkov radiation is emitted. IceCube DOMs contain devices called photomultiplier tubes, which are capable of detecting these tiny flashes of light in the dark environment over a kilometer beneath the ice. The DOMs are spread throughout a cubic kilometer of polar ice, making IceCube the largest neutrino detector in the world.

Data gathered by the IceCube DOMs is sent along wires to the South Pole IceCube Lab (pictured above) where computers filter and store it. Filtered data is then sent north by satellite for analysis at locations like the Wisconsin IceCube Particle Astrophysics Center (WIPAC) in Madison, Wisconsin. A major piece of the analysis process is using IceCube data to reconstruct the actual particle interaction events which created it. This involves the determination of certain physical quantities like the energy of the neutrino or its direction of origin. Once events have been properly reconstructed, they can then be used by scientists for a variety of scientific research projects.

The Search for Neutrino Point Sources

One of the many ongoing areas of research at WIPAC is point source analysis: making use of reconstructed IceCube events to search for specific positions in the sky that are sources of high-energy neutrinos. This task is significantly complicated by the fact that IceCube also detects neutrinos from sources which are not astrophysical in nature. The most significant contribution comes from atmospheric neutrinos, which are created when high-energy cosmic rays interact with particles in Earth's upper atmosphere. Consequently, statistical methods must be used in order to have any chance of differentiating "signal" neutrinos (those originating from astrophysical sources) from the atmospheric background. Four years of IceCube observations have been analyzed so far, but no point sources have yet been found.

It will be at least a few more years before IceCube has collected the amount of data necessary to make definitive point source discoveries, but this does not mean that nothing can be done with the existing data. The four years of observations currently available can already be used to place limits on the types neutrino sources that could possibly exist. The mere fact that IceCube has yet to make any point source discoveries means that there exist no observable neutrino point sources which are producing enough neutrinos so as to be detectable using only four years of data. If such sources did exist, IceCube would have been able to detect them by now. More precisely, it can be asserted that there exist no observable neutrino point sources with a flux above a certain threshold; the exact value of that flux threshold depends on what position in the sky is being considered.

The current IceCube data can be used to calculate neutrino flux upper limits for any position on the sky, but there is little that can be learned by simply calculating limits at arbitrarily chosen points; some system must be used to select points of interest. Where does it make sense to begin looking for high-energy neutrinos, and what can be learned from placing upper limits on the fluxes from these points? One possible answer to this question is to look at known sources of gamma rays, highly energetic photons which are created by some of the most violent events known to modern astronomy. Many sources of very high-energy gamma rays are hypothesized to also be sources of neutrinos, and calculating neutrino flux limits for these sources can reveal something about their nature. By comparing observations of gamma rays and neutrinos for an object, it is possible to begin constraining the mechanisms which might be producing the observed gamma rays.

Hadronic Gamma Ray Production

Gamma Ray Astronomy

Astrophysical gamma rays are the messengers of some of the most energetic events in the known universe. One such type of event is a supernova, the explosive death of a star far more massive than the sun. Supernovae can be accompanied by huge bursts of gamma rays when they occur, and leave behind expanding shells of material called supernova remnants. Supernova remnants can radiate gamma rays for hundreds of years following the explosion that created them. Other gamma ray sources include active galactic nuclei (AGN), supermassive black holes at the centers of galaxies that produce enormous quantities of light as they accrete material from their environments. Many AGN have even been observed to shoot jets of hot, relativistic matter into intergalactic space along their axes of rotation.

VERITAS Cas A
Gamma ray spectrum of Cassiopeia A as measured in 2007. Source: VERITAS [1]
There are a number of currently active gamma ray telescopes, including such projects as MAGIC, VERITAS, H.E.S.S., and the Fermi LAT. These instruments make various observations of astronomical sources of gamma rays, including measurements of their spectra. For gamma ray sources, the spectrum is a plot of how the flux changes with photon energy. Spectral data can be used together with neutrino observations to learn about the kinds of processes occurring in the source. The image at the right shows the gamma ray spectrum of supernova remnant Cassiopeia A, as measured by VERITAS in 2007. The horizontal axis is the photon energy, measured in tera-electronvolts (TeV). The vertical axis measures dN/dE, the photon flux from the source for a given energy.

pp Interaction

One possible method of gamma ray production is a process called proton-proton interaction (or pp interaction). In supernova remnants or the jets of particularly powerful AGN, there is commonly a shockwave of very energetic material present. Because such environments often involve strong magnetic fields, charged particles like protons can become trapped at the shock front by magnetic forces. The shockwave has the effect of accelerating the protons, and is able to do so repeatedly since they are trapped by the magnetic fields. Eventually the protons gain enough energy to escape from the shockwave, and travel away at relativistic speeds. If there happens to be a nearby gas cloud, relativistic protons from the shockwave could interact with protons in the cloud. Because the accelerated protons have such high energy, such collisions can create additional particles called pions.

Hadronic Production
Hadronic production of gamma rays and neutrinos via shock acceleration and pion decay.

Pions come in two varieties: neutral pions (π0) and charged pions (π+ or π) When protons interact, there is a chance that either type will be created. Neither variety is very long-lived; each quickly decays into other particles. Neutral pions decay into gamma ray photons, but charged pions decay into neutrinos and muons, which decay into additional neutrinos and electrons or positrons. The important point is that this process is capable of creating both gamma rays and neutrinos, and that the relationship between them can be determined given the proportions in which neutral and charged pions are made. For a typical set of assumptions and a model from Kappes et al. (2007) [2], there are about 2.67 gamma ray photons for each neutrino produced, and the gamma rays each tend to have twice the energy of the neutrino.

Constraining Gamma Ray Sources

Galactic Sources

The predicted relationship between neutrinos and gamma rays under the assumptions of pp interaction allows for the comparison of gamma ray and neutrino data. The relationship makes it possible to convert from one to the other, so IceCube flux upper limits can be calculated for positions of known gamma ray sources and then converted from neutrinos to gamma rays. These converted limits can then be use to constrain the nature of the gamma ray source at those positions. The two types of data can now be properly compared to one another and conclusions can begin to drawn.

Supernova remnant Cassiopeia A is a known source of gamma rays located in the northern sky. Cassiopeia A is a galactic source, meaning that it is within the Milky Way galaxy; on a cosmic scale, is it relatively nearby. Below is a gamma ray spectrum of this source, including data from multiple different observations. The colored points on the plot indicate spectral gamma ray data for this source as measured by various gamma ray telescopes at various point in time. As in the previous graph from VERITAS, the horizontal axis measures photon energy and the vertical axis measures flux. For this plot however, the flux has been multiplied by the energy squared.

The reasoning behind this decision is somewhat complicated. Examining this plot will reveal that both axes are on a log scale; log scales are convenient for displaying numbers like these, which vary greatly in magnitude, but that is not the only reason for this choice of scale. The processes of gamma ray/neutrino sources are often suspected to yield a power law spectrum, one where the flux is proportional to the energy to some power. Such power laws appear as straight lines on a log-log plot like this. In particular, a baseline assumption is that the flux is proportional to one divided by the energy squared (a power law with exponent -2). Multiplying the flux by the energy squared causes such a spectrum to appear as a horizontal line on this plot and exaggerates any deviation from the baseline assumption.

Cassiopeia A
Constraint plot for Cassiopeia A. Gamma ray flux data sources: MAGIC [3], VERITAS [4], Fermi [5]

The blue and purple lines in the upper right of the plot are neutrino limits calculated by IceCube for the position of Cassiopeia A and fully converted to fluxes and energies in gamma rays using pp interaction assumptions. The horizontal blue line is an integrated limit, which uses all four years of IceCube observations to limit the possible flux from this point in the sky. Calculation of the limit assumes a power law with exponent -2, so it appears as a horizontal line. The line is drawn beginning at 1 TeV because IceCube is not very sensitive at energies below 1 TeV. The purple line is a differential upper limit, which uses the same data as the blue line, but bins the data by energy range. By binning the data in this way, some strength in the limit is lost (it moves upwards from the level of the integrated limit), but the result is a better picture of IceCube's sensitivity, which is greatest between about 10 and 100 TeV, and falls off at lower or higher energies. The differential limit is also more model-independent; its overall curved shape does not rely on the assumption of any particular power law spectral index.

Knowing all of this, it can now be seen that this particular plot is not really very interesting; it properly demonstrates the concepts discussed so far, but it is not a very exciting result. At closest approach, the flux upper limits are roughly an order of magnitude away from the gamma ray data points. The limits constrain the area above themselves; if a neutrino source with a flux above these lines were present at this position in the sky, IceCube would have detected it by now. IceCube will need more years of data to be able to lower the limits further on this source. Until the limits closely approach the gamma ray data, no constraints can be put on this source.

EBL Absorption

Before moving on to more interesting sources, a certain concept relevant to extragalactic gamma ray sources must be discussed. For galactic gamma ray sources, the fluxes observed on Earth are not noticeably different from the actual flux "at the source." For more distant sources, this is not necessarily the case. When studying extragalactic sources like AGN, it becomes necessary to account for absorption by the extragalactic background light (EBL). The EBL is an excess of background photons, mostly in the infrared and microwave parts of the electromagnetic spectrum, that fills intergalactic space. When extremely high energy gamma ray photons (like those being considered) travel through intergalactic space, they can be absorbed by the EBL. If a gamma ray interacts with an EBL photon, the gamma ray possesses such a huge amount of energy that, regardless of the energy of the other photon, pair production can occur, turning the photons into an electron-positron pair. When this happens, the gamma ray will not be observed on Earth and is said to have been absorbed by the EBL.

EBL Absorption
Graph demonstrating the effect of EBL absorption for photons of various energies. Created by interpolating optical depth values from Franceschini et al. (2008) [6]
The effect of EBL absorption increases quickly as a function of the photon energy and the distance to the source, measured using redshift. The graph at the right shows the effect of the absorption as a function of redshift for photons with energies 0.1, 1, 10, and 100 TeV. The absorption effect is here measured by a quantity called the optical depth. The optical depth is such that the gamma ray flux seen at Earth is equal to the flux at the source divided by the exponential of the optical depth. For gamma rays at very high energies and moderate redshift, the absorption can make a dramatic difference of one or more orders of magnitude in flux.

High-energy gamma ray photons are affected quite strongly by EBL absorption. Neutrinos, by comparison, interact so extraordinarily weakly with other particles and themselves that they do not suffer from any type of absorption effect. Therefore, it is not possible to directly compare neutrino and gamma ray fluxes from extragalactic sources. Fortunately, there are models (like that used to create the graph at the right) which predict the effects of EBL absorption. Using these models, we can correct for the effects that the EBL has on gamma rays, a process called deabsorption. Once extragalactic gamma ray data has been deabsorbed, it can then be properly compared to neutrino data from extragalactic sources.

Extragalactic Sources

With the ability to correct for gamma ray EBL absorption, analysis can now be performed for extragalactic sources. One such source is an object named 1ES 1959+650. 1ES 1959+650 is a blazar, a particularly bright type of AGN which happens to have a powerful jet pointed in our direction. This makes the object unusually bright (across the electromagnetic spectrum) for its massive distance from earth and makes it a particularly interesting target for observations. With a redshift of z = 0.048, it is certainly distant enough that deabsorption is necessary for gamma ray data from it. The animation below shows the effect of deabsorbing gamma ray data obtained for this source. The flux value of several points changes drastically, especially at the high-energy end of the spectrum.

Deabsorption of 1ES 1959+650
Gamma ray data from blazar 1ES 1959+650 before and after deabsorption. Gamma ray flux data sources: HEGRA [7], Whipple [8], MAGIC [9]

The next plot is a static image of the deabsorbed spectrum of 1ES 1959+650. This plot is much more interesting than the one for Cassiopeia A, because, in this case, the limits are approaching and even going below some of the data points. The odd-looking spectrum of 1ES 1959+650 requires some degree of explanation. The normal flux observed from this source is represented by the circular data points, which are currently below the IceCube limits. The other points, with up to 10 times more flux at certain energies, are data gathered during a flare that occurred in 2002. For a short time that year, the source's flux was dramatically higher. It is the data from this event that is above the IceCube limits.

1ES 1959+650 Deabsorbed
Deabsorbed spectrum of blazar 1ES 1959+650.

Unfortunately, this flare happened several years before IceCube was built. However, something can still be learned from this plot; if 1ES 1959+650 undergoes a flare of that magnitude again, IceCube should be able to see it. Were such an event to happen, IceCube scientists could use special time-dependent analysis techniques to check for an accompanying neutrino flare. Over such a short period of time, the number of background detections would be limited, significantly increasing IceCube's sensitivity, and thus increasing the chance of seeing something. That is, assuming that there is something to see. It is possible that 1ES 1959+650 is not creating gamma rays via pp interaction or other hadronic methods. If this is the case, IceCube can begin to constrain the source, concluding that some or all of its gamma rays are not hadronic in origin.

Another famous blazar named Markarian 421 has also been examined using IceCube limits. Markarian 421 is an extremely bright blazar and is historically quite variable in flux, as is apparent on its spectral plot, shown below. The gamma ray data shown spans a 15-year period from 1995 to 2010, and the flux of Markarian 421 seems to change quite sporadically over that time period. This plot is quite promising because several of the data points are exceeding the limits imposed by IceCube. In April 2013, Markarian 421 even underwent a flare, but the IceCube data from this time period has yet to be analyzed for evidence of a corresponding neutrino flare. Under any circumstances, it should soon be possible for IceCube to begin constraining Markarian 421, determining the proportion of its emission that has hadronic origins and perhaps detecting a statistically significant neutrino point source at this location.

Markarian 421
Deabsorbed spectrum of blazar Markarian 421. Gamma ray flux data sources: Whipple [10], HEGRA [11][12], TACTIC [13][14]

Summary of Results

In conclusion, IceCube does not yet have enough data to be able to conclusively detect point sources of neutrinos, but the data can still constrain what sources could exist. Under the assumptions of pp interaction, high-energy astrophysical gamma rays are related to high-energy neutrinos. With this knowledge, we can look at known gamma ray sources and constrain the gamma ray production methods. For a few bright sources like Markarian 421, IceCube is finally reaching the point where real constraints can be made and conclusions can be drawn. With more years of data, the limits will only get better, and IceCube will be able to not only limit more and more high-energy gamma ray sources, but begin independently searching for neutrino point sources.

References and Acknowledgments

[1] Acciari et al. (2010) arXiv:1002.2974v1
[2] Kappes et al. (2007) arXiv:astro-ph/0607286v3
[3] Albert et al. (2007) arXiv:0706.4065v2
[4] Acciari et al. (2010) arXiv:1002.2974v1
[5] Abdo et al. (2010) arXiv:1001.1419v2
[6] Franceschini et al. (2008) arXiv:0805.1841v2
[7] Aharonian et al. (2003) arXiv:astro-ph/0305275v2
[8] Daniel et al. (2005) arXiv:astro-ph/0503085v1
[9] Albert et al. (2005) arXiv:astro-ph/0508543v1
[10] Krennrich et al. (1998) arXiv:astro-ph/9808333v1
[11] Aharonian et al. (1999) arXiv:astro-ph/9905032v1
[12] Aharonian et al. (2002) arXiv:astro-ph/0205499v1
[13] Yadav et al. (2007) arXiv:astro-ph/0701890v1
[14] Chandra et al. (2012) arXiv:1202.2984v1

There are several people I would like to extend thanks to for their contributions to this summer's REU program. I would like to thank my WIPAC mentors Jake Feintzeig and Albrecht Karle for all their help, and for a great summer of astrophysical research in Madison. I'd like to thank Eric Hooper, who organized the UW–Madison astrophysics REU program and kept everything running smoothly. Thank you also to the IceCube Collaboration, the University of Wisconsin–Madison, and the National Science Foundation, who funded the program. Finally, thank you to all of my fellow REU students who helped to make my summer in Wisconsin the fantastically enjoyable experience that it was.

Related Links

Wisconsin IceCube Particle Astrophysics Center

IceCube South Pole Neutrino Observatory

TeVCat Online TeV Gamma Ray Astronomy Catalog

arXiv.org Research Article Archive

Research Projects of Other REU Students


Greg Vance — University of Wisconsin–Madison — Wisconsin IceCube Particle Astrophysics Center — 2013