supernova remnant N63A
by Viggo van der Roest
Student nr. 11306009
Report Bachelor Project Physics and Astronomy, size 15-18 EC
Conducted between 10-4-2020 and July 23, 2020
Submitted on July 23, 2020
Faculty of Science, University of Amsterdam
snr group, API
Supervisor:
dr. Jacco Vink
Daily Supervisor:
Lei Sun MSc.
Second Examiner:
dr. Oliver Porth
Abstract
The young supernova remnant N63A (∼ 2000−5000 yr) is one of the brightest supernova remnants in the Large Magellanic cloud (LMC). Recently it has come to the attention of the API supernova remnant group that it could be a possible cosmic ray source. To enable cosmic ray research, this project conducted an examination of the Chandra data on N63A to constrain the values for the electron and proton density, measure the temperature and to determine a lower limit on the age of the remnant. In the analysis a model consisting of the Tuebingen-Boulder Interstellar medium absorption model (tbabs) and the non-equilibrium ionization collisional plasma model (nei) was used. The results show that most of the analyzed regions on N63A show abundances of O, Ne and Mg above the LMC levels indicating a core collapse type II origin. The electron density was determined at ∼ 6 − 10 cm−3 for regions that showed abundances above the LMC levels and ∼ 1.6 − 5 cm−3 for regions that did not show elevated abundances. The lower limit on the age has
been determined at 3175−1905 and at 4120+195−92 yr using the Sedov-Taylor model approximation of
uit helium en drie kwart uit waterstof. Alle andere elementen zijn later ontstaan en de enige manier waarop dit kan, is in sterren. In de kern van een ster vindt namelijk kernfusie plaats waarbij de kernen van atomen samensmelten tot nieuwe elementen. In de zon fuseert waterstof voornamelijk tot helium, maar in zwaardere sterren kan dit process veel zwaardere atomen opleveren. Het zwaarste atoom dat in een ster met kernfusie gecre¨eerd kan worden is ijzer, specifiek 56
26Fe. Bij
kernfusie komt er tot en met ijzer energie vrij, daarna kost het energie, vandaar dat het het eindpunt van de fusieketen is. Dit zou betekenen dat elementen als koper, broom, zilver of goud niet kunnen bestaan als sterren aan het einde van hun leven als een nachtkaars uit zouden gaan.
Sterren met een massa ∼ 0.5 − 8 M (=zonsmassa) worden een witte dwerg als alle brandstof
is opgebrand. Dit gebeurt wanneer al het helium is gefuseerd tot zuurstof en koolstof. Als een ster zwaarder is dan 8 M , kan koolstof in de kern verder opbranden tot neon, magnesium en
uiteindelijk ijzer. IJzer kan niet meer verder fuseren. Op een gegeven moment is de massa van het ijzer in de kern meer dan de Chandrasekhar limiet, 1.4 M . De kern kan dan de zwaartekracht van
de omliggende lagen van de ster niet meer tegenhouden en implodeert. De lagen buiten de kern vallen nu opeens naar het binnenste van de ster omdat deze niet meer door de kern naar buiten gedrukt worden. Tijdens dit proces wordt de kern van de ster heel heet, zo rond de 1010K, en komt er 1046 J aan energie in 10 seconden vrij. Op dit moment kan de kern niet meer kleiner worden
omdat de dichtheid van de ingestorte kern de dichtheid van een atoomkern benadert en wordt de implosie omgedraait en samen met alle neerstortende materie in een grote schokgolf naar buiten geblazen. In een korte periode hiervan is het mogelijk om ijzer te laten fuseren tot zwaardere elementen als goud en zilver. De kern van de ster blijft achter als neutronenster of zwart gat met een supernova restant er omheen. Dit wordt ook wel een type II supernova genoemd. Als er geen waterstof in het spectrum van het supernova restant zit, dan is het een type Ib of Ic supernova. Een witte dwerg kan ook als supernova exploderen wanneer deze als dubbelster voorkomt. Dit gebeurt wanneer de witte dwerg massa overneemt van de andere ster. Voor witte dwergen geldt er namelijk een omgekeerd evenredig verband tussen de massa en straal. Hoe zwaarder de witte dwerg dus wordt, hoe kleiner deze wordt. Als de massa de Chandrasekhar limiet bereikt, dan zal de witte dwerg door de zwaartekracht instorten en er een supernova type Ia ontstaan. Deze speciale supernova’s worden in de cosmologie ook gebruikt om de afstanden tot hele verre sterrenstelsels te meten.
Supernova’s zijn dus hele belangrijke en interessante objecten om te bestuderen. Alleen komen ze niet zo heel vaak voor. In een eeuw komen er in een sterrenstelsel als de melkweg maar 2 of 3 supernova’s voor. Dit betekent dat deze moeilijk te bestuderen zijn. Een manier om dit toch te kunnen doen is door te kijken naar het supernova restant wat achterblijft na een supernova. Uit een supernova restant kan heel veel informatie zoals de dichtheid, aanwezige elementen, leeftijd en temperatuur gehaald worden. Dit wordt gedaan door middel van spectroscopie. Een supernova restant zendt in meerdere gebieden van het elektromagnetische spectrum straling uit, maar is het felste in de r¨ontgenband.
In dit onderzoek is het supernova restant N63A onderzocht om de temperatuur te meten en de dichtheid en leeftijd beter te bepalen. N63A ligt in de Grote Maghaelaanse Wolk op een afstand van 160.000 lichtjaar en heeft een doorsnede van 25 lichtjaar. De Grote Maghaelaanse wolk is een klein sterrenstelsel vlakbij de melkweg op ongeveer 160.000 lichtjaar afstand van de Aarde. Men denkt dat N63A ongeveer 2000 tot 5000 jaar oud is en weet nog weinig over de dichtheid. Om hier meer over te weten te komen is een observatie van de Chandra telescoop onderzocht. De Chandra telescoop is een van de drie r¨ontgen telescopen die in een baan om de Aarde zweven. In de analyse zijn er verschillende regio’s onderzocht door naar het spectrum te kijken en hier een model mee te laten rekenen. Het model bestond uit twee delen, ´e´en om de absorptie van de straling tussen N63A en de Aarde te meten en ´e´en om de door N63A uitgezonden straling te kunnen analyseren. Uit deze analyse is gebleken dat N63A een temperatuur van ongeveer 8 miljoen Kelvin en een minimale leeftijd van 3175 jaar heeft. De dichtheid van protonen ligt tussen de 1.60 en 8.29 protonen per cm3 en de dichtheid van de electronen tussen de 1.92 en 9.95 elektronen per cm3.
Acknowledgements
Firstly, I would like to thank my daily supervisor Lei Sun for helping me out with software problems along the way, discussing the results and proofreading the first version of my thesis.
I am grateful to dr. Jacco Vink for enabling me to conduct this research with his supernova remnant research group at the API, his solutions offered in the group meetings were very valuable and his proofreading of my first version was very helpful.
Thanks to dr. Oliver Porth for being my examiner on this project.
Finaly, many thanks to my fellow final year bachelor of physics and astronomy students Zeger Ackerman and Niels Ruijter for respectively proofreading my first and final version of this thesis. Their advices were very useful.
2 Background 6
2.1 Supernovae . . . 6
2.2 Supernova remnants . . . 7
2.2.1 Supernova remnant evolution . . . 7
2.2.2 Supernova remnant classification . . . 8
2.3 Shock heating . . . 9
2.4 X-ray emission . . . 10
2.5 X-ray observatories and spectroscopy . . . 10
2.6 N63A . . . 11 3 Analysis 13 3.1 Data preparation . . . 13 3.2 Model . . . 14 4 Results 16 5 Discussion 19 6 Conclusion 21
1 INTRODUCTION 5
1
Introduction
Supernova explosions are among the most spectacular and energetic events in the universe. Their explosions enable nuclear fusions that create the heaviest elements on Earth. [1] A supernova occurs at the end of the life of a massive star. This can either happen when a white dwarf accretes enough matter from a companion star to reach and surpass the upper limit on its mass, the Chandrasekhar limit (∼ 1.4M ), or when the core of a massive star (∼ 8M ) collapses under its
own gravity. A core collapse supernova leaves a supernova remnant with a neutron star or stellar black hole at the center behind after exploding and the white dwarf explodes without leaving a compact remnant behind. Supernovae are used in different fields of physics. Supernovae type Ia are used in cosmology as standard candles to measure the distance to far away galaxies. Due to their brightness, they can be observed for high redshifts up to z ∼ 1.7. [2]
As the explosion itself occurs very quickly, the best way to study nearby supernovae is through supernova remnants as the supernova rate in a typical spiral galaxy as the Milky way in the local group is 2-3 per century. [2] A supernova remnant can be seen as an imprint of the explosion on the interstellar medium. Supernova remnants enable us for example to study the progenitor of the supernova or the nucleosynthesis of the explosion and can help us to understand high Mach number collisionless shocks, which are not yet fully understood.
The best way to study supernova remnants is through X-ray observations. These observations are essential in obtaining the elemental abundances in the remnant through X-ray spectroscopy. This is especially true for elements such as O, Ne, Si, Ar, S, Ca, Mg, Fe and Ni. All of these elements have prominent emission lines in the 0.5 − 10.0 keV range at an emission temperature between 0.2 to 5.0 keV, which includes the typical temperature of a supernova remnant. Apart from the elemental abundances, X-ray spectroscopy can also help to constrain the values of the density and age of the remnant.
This thesis reports on an investigation of the X-ray emission of supernova remnant N63A. N63A is located in the north-eastern part of the Large Magellanic cloud (LMC) and has been estimated to be around 2000-5000 yr old. [3] Recently, it came to the attention of the supernova remnant group at the API that N63A could be a potential cosmic ray source. On the basis of ASCA X-ray observations of LMC supernova remnants taken in the mid 90s, it was concluded that the three brightest X-ray sources -N132D, N63A, and N49- result in unusually high inferred values of initial SN explosion energy. Given that all the three X-ray brightest are at present detected in GeV gamma rays and the X-ray brightest one, N132D, is also detected in TeV gamma rays. N132D is also among the two most luminous supernova remnants in the TeV band. The X-ray analysis of one of these supernova remnants, N63A, with Chanda providing us with finest angular resolution (< 1 arcsec compared with 10 arcmin for TeV observatories) is most relevant for a study of physical processes occurring in gamma-ray emitting supernova remnants. [4] A high gamma-ray luminosity could be due to proton cosmic rays colliding with the supernova remnant plasma, producing pions which decay into gamma rays. Relativistic electron interactions with photons produce gamma rays due to inverse Compton up-scattering and interactions with ions result in brehmsstralung. [2] For the group to investigate the possible cosmic ray source N63A further, it is essential that the density of the remnant is known.
In this thesis the goal is to measure the temperature of the remnant directly from the Chandra observation, to give a constraint on the density and a lower limit on the age of the N63A remnant.
2
Background
2.1
Supernovae
A supernova explosion is the violent end of a massive star (> 8 M ) or white dwarf. Two broad
categories of supernovae are distinguished: type I and II. The absence of hydrogen from the optical spectrum is defining for type I supernovae while type II supernovae do show hydrogen in the visible spectrum. [5] Figure 3 shows what determines the different supernova types. Supernovae can occur when a massive star (> 8 M ) reaches the end of its lifetime. A star with > 8 M will
be able to burn carbon into neon, magnesium, silicon and finally iron. When the core of the star consists entirely of iron,the core will no longer release energy from thermonuclear fusion and start to collapse making the degenerate electrons more relativistic as the pressure increases. [5] If the mass of the core exceeds the Chandrasekhar limit, the degenerate electrons no longer support the core. This causes the core to collapse. When the core collapses, it gains energy by nuclear photodisintegration and electron capture. The photodisintegration is caused by the interaction between iron-56 and gamma rays and results in a complete decay into protons and neutrons. Electron capture allows the energy to be carried away by neutrinos. This process rapidly uses up electrons which causes the the degenerate electron pressure to lower and the core to collapse further. [5] When the density of the core approaches the density comparable to an atomic nucleus, the nuclear forces resist further collapse of the core and rebounds the collapse. This causes a shock wave to go trough the falling material causing much of the stellar envelope to expulse. This is called a supernova explosion. [5] The exact underlying physics are not yet fully understood as most numerical models are not able to explain the shock physics yet. [2]
Figure 1: The determination of supernova types. Type Ia supernovae are associated with CO white dwarfs and type Ib/c and II are core collapse supernovae. The difference between type I and II is in the absence of hydrogen lines in type I supernovae. [6]
Type Ia supernovae progenitors are CO white dwarfs that are acceding mass from a stellar companion in a binary system until the Chandrasekhar limit of ∼ 1.4 M is reached. The
de-generate electron pressure is no longer able to support the mass of the white dwarf and starts to collapse. The density and temperature inside the core of the white dwarf starts to increase prompting the carbon and oxygen to start nuclear fusion up to iron-52. A neutron star will be formed and a small fraction of the initial mass will be ejected by the explosion. [5] The light curve created is characteristic for a type Ia supernova and is similar to all other type Ia supernovae due to the similar progenitors. This makes them excellent as standard candles in cosmology as they can be observed at high redshifts due to their apparent brightness.
2 BACKGROUND 7
Type Ib/c and II supernovae are core collapse supernova explosions. Type Ib and Ic do not show hydrogen in their visible spectrum as the hydrogen envelope has been stripped from the progenitor star. Absence of silicon lines are characterizing these supernovae. The difference between type Ib and Ic supernovae is the absence of strong helium lines in the Ic supernova spectra. [6]
Type II supernovae show clear hydrogen lines in their spectra. Most type II supernovae are of the type IIp or IIl. The defining feature is the shape of the optical light curves. Type IIp supernova luminosity stops declining after the maximum of the light curve and forms a plateau on the luminosity curve for 2-3 months, while type IIl supernovae show a linear decline in the luminosity curve. [6] This is probably due to the mass of the hydrogen envelope of the star. If the hydrogen envelope is larger, a plateau is formed in the luminosity curve as the recombination wave moves trough the envelope. [6] Some type II supernovae show narrow emission lines and are therefore called type IIn. The spectra evolve slowly and are dominated by strong Balmer emission lines. [6]
A mixing between type II and Ib/c supernovae is called a type IIb supernova. These resemble a type II supernova in the early phases of the light curve evolution, but have late spectra that resemble type Ib supernovae. [6]
2.2
Supernova remnants
The structure of a supernova remnant can be simplified as two shocked shells. The inner shell consists of material ejected during the supernova explosion, the ejecta, and the outer shell consists of the shocked interstellar medium. The swept up interstellar medium is accelerated, compressed and heated by the shock preceding the ejecta. It must be noted that the shock front is not a well defined line, but instead more an area at the edge of the remnant where the interstellar medium is affected by the shock. The shocked interstellar medium pushes back onto the ejecta which decelerates, compresses and heats the ejecta. This is called the reverse shock. [7] The basic structure can be observed in figure 2.
Figure 2: The structure of a supernova remnant. From the outside in: the un-shocked ambient medium, the forward shock(FS), the shocked ambient medium, the contact discontinuity(CD), the shocked ejecta, the reverse shock(RS) and the freely expanding un-shocked ejecta (This drawing is not to scale). [7]
2.2.1 Supernova remnant evolution
The evolution of a supernova remnant can be divided into four different phases: the ejecta domi-nated phase, the Sedov-Taylor self-similar phase, the snowplough phase and the momentum con-serving shell phase. [8] These distinct phases are simplifying the evolution as a whole. It could also be that the remnant is in different phases in different areas. [2] The first phase, the ejecta dominated phase is characterized by an almost free expansion of the remnant into the surrounding interstellar medium as in this phase the remnant still experiences a loss of energy. The phase is
best described by the Chevalier model. [9] The Chevalier model is roughly described by a power-law density distribution and describes the early evolution of the supernova remnant when the reverse shock is still sweeping through the ejecta. [2] This phase continues until the mass of the ejecta, Mej, is no longer larger than the mass of the swept up interstellar medium, Msw.
When Mej< Mswthe second phase, the Sedov-Taylor phase, takes over. This phase is described
by the Sedov-Taylor self-similar analytical model. [10, 11]This model does not take into account the structure of the ejecta itself, but does present a decent approximation as soon as Msw> Mej.
In the Sedov-Taylor phase, the reverse shock has passed the inner most ejecta and the interior is heated to such extend that the radiative energy losses are very small as all atoms are ionized and recombination is non existent. As there is no radiative loss, the Sedov-Taylor phase can be regarded as adiabatic. This implies that the cooling of the gas is only due to the remnant expanding. Both the Sedov-Taylor model and the Chevalier model can be combined in the Truelove and Mckee analytical model. This model takes the early as well as the late phase of the remnant in account. [12]
As the remnant grows, the forward shock velocity drops as more interstellar medium becomes swept up by the remnant. Eventually the remnant evolution will proceed to the third phase:: the snowplough phase. At this point radiative losses have become important and the evolution shifts from being governed by energy conservation to momentum conservation. This happens when the shock velocity Vs ≈ 200 km/s. Around that boundary the post-shock temperature falls below
∼ 5 × 105K at which point oxygen line emission starts acting as an important cooling factor in. [2]
The remnant enters the last phase when the shock velocity reaches a value below the local speed of sound and starts to merge with the interstellar medium. In this stage a remnant is no longer distinguishable from the random motions in the interstellar medium.
2.2.2 Supernova remnant classification
Supernova explosions are organized in two different types: thermonuclear, type Ia, and core col-lapse supernovae with several subdivisions. Supernova remnants however are not equally classi-fiable. Especially for older remnants the origin is not always determinable. The presence of a neutron star in the center of a remnant was first used as an indicator of a core collapse origin. However, this could also happen by chance in the case of a later enclosed neutron star during the remnant expansion. The position of a remnant could also be a clue. Core collapse supernovae are expected in or near a star forming region for example, but this is not definite proof of their origin. Type Ia supernovae can be expected to be found above the galactic plain. [2] Recently, X-ray spectroscopy has made it possible to distinct type Ia supernovae from core collapse remnants by looking at the ejected oxygen and iron mass. If the ejected oxygen mass in young supernovae is over 0.2 M , it is an indicator for a core collapse origin. [7] Similarly when the ejected iron mass
is more than 0.4 M , it is likely that the remnant is the result of a type Ia explosion. [7]
In most cases the criteria above are insufficient for determining whether a remnant has a core collapse or type Ia origin. Therefore a different classification is used for supernovae remnants based on observations. This classification is mostly based on morphology and consists of four classes: the shell type remnant, pulsar wind nebulae, composite and mixed morphology remnants. [8] The shell type remnant is characterized by a shell consisting of shocked material emitting thermal X-ray emission from a circular rim in the sky, an example can be seen in figure 3. For a core collapse supernova however, a fast rotating neutron star is expected in the middle of the remnant. The rotating neutron star would produce a wind consisting of relativistic electrons and positrons which terminates in a shock and this process accelerates the particles to highly relativistic energies. [7] This results in emission of synchrotron radiation and inverse Compton scattering. [2] Well known examples of pulsar wind nebulae are the crab nebula and 3C58 shown in figure 3. The third type of supernova remnants is a composite remnant. This consists of a pulsar wind nebula surrounded by a thermal X-ray emission shell like the Kes 75 remnant in figure 3. The fourth type is the mixed morphology or thermal-composite remnant, which are usually located in high density interstellar medium surroundings or can be associated with GeV gamma ray sources like remnant W44. [7] [13]
2 BACKGROUND 9
Figure 3: Examples of the four types of supernova remnants. The top left remnant is a shell type remnant called the Cygnus loop with a diameter of 3◦ observed by ROSAT PSPC. Red ∼ 0.1 − 0.4 keV, green ∼ 0.5 − 1.2 keV and blue ∼ 1.2 − 2.2 keV. The top right shows a pulsar wind nebula called 3C58 with a long axis of 70 observed by Chandra. The bottom left is a composite remnant called Kes 75. The inside consists of a pulsar wind nebula with a hard X-ray spectrum and is surrounded by a partial shell with a radius of 1.40. Red (Ne and Mg emission) ∼ 1 − 1.7 keV, green (Si and S emission) ∼ 1.7 − 2.5keV and blue (mostly pulsar emission) ∼ 2.5 − 5 keV. The bottom right is the remnant W28 observed in radio (orange) by the VLA and in X-ray (blue) observed by ROSAT. This is an example of a thermal-composite or mixed morphology supernova remnant. [2]
2.3
Shock heating
As a supernova starts expanding, it is preceded by a shock wave as the ejecta travel way faster ∼ 104km/s than the local speed of sound. This shock heats the interstellar medium surrounding
the progenitor star by changing kinetic energy into thermal energy. The local interstellar medium has a very low density (∼ 1 cm−3). A particle-particle interaction like what would happen in a pressure shock on earth is not possible in the interstellar medium because the mean free path is long compared to the size of the remnant itself. Therefore the heating process can not be through particle-particle interaction. [2] Instead, a collisionless shocking mechanism heats the interstellar medium due to fluctuating electric and magnetic fields. The average temperature of the remnant can be used to determine the shock velocity of the remnant
kT = 3 16µmpv 2 s ≈ 1.2( Vs 1000 km/s) 2 keV (1)
with kT the average plasma temperature in keV, µ the average particle mass, mpthe proton mass
and Vs the shock velocity in km/s. This equation assumes a high Mach number (M > 5) for the
shock velocity.
2.4
X-ray emission
The X-ray emission from a supernova remnant comes from thermal and non thermal emission. Thermal emission can come from bremsstrahlung, recombination continuum (or free bound emis-sion) and two-photon emission caused by the electron quantum level transition. [2] Most of the continuum emission comes from bremsstrahlung. Bremsstrahlung is dominated by helium ions and protons colliding with electrons for sub-solar abundances expected in the Large Magellanic cloud. The collision between shocked ejecta electrons and ions is also a large contributor to the continuum emission. [2] Free bound emission is emitted when an electron is caught in the atomic shells and can result in narrow emission peaks near series limits of lines. [2] Two-photon emission comes from the de-excitation of electrons in metastable states as the 2s state. [2] Sometimes line emission can be caused by radioactive decay when gamma rays or X-rays emitted during the decay. Nickel-56 could be taken as an example. The amount of Ni-56 is for instance directly related to the luminosity of type Ia supernovae. [6] Non-thermal emission can either be synchrotron radiation or non-thermal bremsstrahlung. X-ray synchrotron is emitted by relativistic electrons in a magnetic field. Synchrotron sources can emit frequencies between low radio and ray bands. Usually X-ray synchrotron radiation is associated with a pulsar wind nebula. Radio synchrotron radiation is mostly found in the shell of the remnant. [2] Non-thermal radiation is not thought to emit enough radiation to be identified in the hard X-ray spectrum, but could still be contributing. [2]
2.5
X-ray observatories and spectroscopy
The brightest emission from supernova remnants is in the X-ray band, therefore X-ray analysis of the remnants is a valuable source of information. Currently, there are three major observa-tories in orbit. These are the American Chandra (lauched: 23 July 1999), European XMM-Newton (launched December 1999) and Japanese Suzaku observatory (launched 10 July 2005). All three observatories carry CCD-detectors as their main instruments. The Chandra Advanced CCD Imaging Spectrometer (ACIS) provides the highest spatial resolution of the three with a pixel size of 0.4900. [14] [2] The XMM-Newton has the European Photon Imaging Camera (EPIC) behind its telescopes. EPIC-MOS is a CCD-detector with a p-doped metal-oxide-semiconductor (MOS) and EPIC-pn is a CCD-detector based on a pn-junction. XMM-Newton contains two EPIC-MOS detectors and one EPIC-pn detector. The use of the detectors depends on the obser-vation made. [15] The Suzaku observatory consists of five soft X-ray and one hard X-ray detector. The main CCD-detector is called the X-ray Imaging Spectrometer (XIS). [16] Both the Chandra and XMM-Newton also carry grating spectrometers. Chandra carries the High Energy Transmis-sion Grating Spectrometer (HETGS) and the Low Energy TransmisTransmis-sion Grating Spectrometer (LETGS) and XMM-Newton caries the Reflection Grating Spectrometer (RGS). [2]
Both the CCD-detectors and grating spectrometers are used to detect the emission lines in the observed spectra. For supernova remnants the main lines that can be expected in the spectrum are the emission lines of elements like oxygen ∼ 0.55 − 0.60 keV, neon ∼ 0.9 − 1.1 keV, iron ∼ 0.9 − 1.1 and ∼ 6.6 − 7.0 keV, magnesium ∼ 1.3 − 1.4 keV, silicon∼ 1.8 − 1.9 keV, sulfur ∼ 2.4 − 2.5 keV, argon ∼ 3.1 − 3.2 keV and calcium ∼ 3.8 − 4.1 keV. [17, 18] Most of these lines can be observed in the spectrum of Cassiopeia A as seen in figure 4
2 BACKGROUND 11
Figure 4: The spectrum of supernova remnant Cassiopeia A with the emission lines of O, Ne, Mg, Si, S, Ar, Ca, Fe shown as detected by Chandra. Image retrieved from: https://chandra.harvard.edu/edu/ formal/snr/bg5.html
2.6
N63A
The young core collapse supernova remnant N63A, also known as MCSNR 0535-6602, LHA 120-N 63A or S120-NR B0535-66.0) is located in the north-eastern part of the Large Magellanic Cloud (05:35:44.10 -66:02:14.00), see figure 5a. It is estimated to be around 2000-5000 years old. [3] The diameter of N63A is ∼ 8.2 pc. In 1998, Hughes et al. used ASCA data to estimate the shock age of the remnant at 4200 − 4500 years with respect to the Sedov-Taylor model. [19] There is no young energetic pulsar in the middle of N63A as found by Warren et al. in 2003, which is ruled out by hard limits on the hard X-ray emission. A pulsar wind nebula with a luminosity less than ∼ 4 × 1034 erg/s is allowed in the center of the remnant. [3]
Figure 5b shows that in the middle right part of the remnant there is a lack of x-ray emission. This can be explained by a cloud of dense cold molecular gas between Earth and N63A that absorbs most of the X-ray photons emitted in this part of the remnant. The structure of N63A also reveals ”crescent”-shaped structures that lie outside of the main shell of the remnant, see figure 5b. These structures have only been seen at one other remnant: the Vela supernova remnant. These have been identified as high speed ejecta clumps that interact with the surrounding interstellar medium. [3]
(a) N63A in the Large Magellanic cloud (b) RGB-image of N63A
Figure 5: Left: The Large Magellanic cloud. The white arrow points at the approximate location of N63A. Image retrieved from: https://www.eso.org/public/images/eso1021d/ Right:RGB image of N63A with shown in red the radio band (ACTA), green the optical band (Hubble space telescope) and blue the X-ray band (Chandra). Image retrieved from:https://chandra.harvard.edu/photo/2003/n63a/
Table 1 shows the elemental abundances found by Warren et al. and Hughes et al. with respect to the solar abundances. Since Warren et al. used the same Chandra observation (ObsID: 777) as was used in this thesis, it is expected that approximately the same abundances are expected to be found. Warren et al. however used the abundances found by Russel and Dopita in 1992, while this study has used the abundances found by Wilms et al.. [20, 21] Both results also showed an absorption of NH∼ 0.8 − 1.7 × 1021 cm−2 where the mean values of 1.4 × 1021cm−2 found by
Hughes et al. and 1.68 × 1021cm−2 found by Warren et al. are consistent with each other. Abundance Hughes et al. Warren et al. soft Warren et al. mean
O 0.28 ± 0.05 0.21 ± 0.02 0.18 ± 0.06
Ne 0.45 ± 0.08 0.39 ± 0.04 0.26 ± 0.04
Mg 0.46 ± 0.08 0.28 ± 0.05 0.16 ± 0.02
Si 0.27 ± 0.09 0.19 ± 0.04 0.15 ± 0.02
Fe 0.30 ± 0.04 0.17 ± 0.03 0.13 ± 0.01
Table 1: The elemental abundances with respect to the solar abundance for N63A found by Hughes et al.. [19] Warren et al. [3] found abundances for the soft X-ray band and mean spectrum. Both are with respect to the solar abundances found by Russel and Dopita in 1992. [20] All errors quoted are for a 90% confidence range.
3 ANALYSIS 13
3
Analysis
3.1
Data preparation
The data on N63A was obtained from the Chandra observatory with observation ID 777. The observation started at 16 October 2000 at 20:25:59 and lasted for 4.335 × 104 seconds. The
spectrometer used to observe N63A was the ACIS-S instrument and no grating instrument was used. The data was publicly released on 3 November 2001. In the analysis of the data the SAOImage DS9, Ciao-4.12 by Chandra and Heasoft-6.27.2 by NASA were used.
Firstly the downloaded observation files were reproduced with the chandra repro function in Ciao-4.12 to create new level 2 event, pha and bad pixel files with the appropriate response files. With the use of DS9, the remnant was divided in 110 regions scattered across the whole remnant, see figure 6. Each region must at least contain 1000 photon counts to be used as it enables the spectrum to be divided into at least 40 groups of a minimum of 25 counts. This obtains a group with S/N ∼ 5σ to be able to do a good fit. For each region the spectrum is extracted by hand using the dmextract function to extract the spectrum from the region. Afterwards the auxiliary response file (ARF) is obtained by using the mkwarf function and the redistribution matrix file (RMF) is obtained by the mkacisrmf function in ciao-4.12. The same result can also be obtained by using the specextract function, but, since this gives less freedom to extract the files in a preferred way, it has not been used. Each region would then be grouped by the grppha function in heasoft-6.27.2 in groups of 25 photon counts. XSPEC 12.11.0 is used to perform the spectral fitting.
Figure 6: Image of N63A on with the photon count on a square root scale, blue equals a low and yellow a high photon count. The green circles and rectangles are the analyzed regions.
3.2
Model
The spectral fits have been made with a two component model consisting of an absorption com-ponent and an emission comcom-ponent. The absorption comcom-ponent used was the Tuebingen-Boulder Interstellar medium absorption model (tbabs) which estimates the X-ray absorption from the grain and gas-phase interstellar medium and the molecules in the interstellar medium from which only molecular hydrogen is considered. [22] The emission component is the non-equilibrium ionization collisional plasma model(nei). The nei model allows for the calculation of the plasma tempera-ture in keV, the abundances for hydrogen, helium, carbon, nitrogen, oxygen, neon, magnesium, silicon, sulfur, argon, calcium, iron and nickel with respect to the solar abundance, the redshift, the ionization timescale in s · cm−3 and the normalization of the function. [23]
From the plasma temperature (keV) the temperature (K) can be calculated by dividing the value by the Boltzmann constant. The abundances used in this thesis are fitted with the values found by Wilms et al. in 2000. [21] Since the Large Magellanic cloud is close to Earth, 49.97 ± 1.30 kpc [24], redshift is not relevant for the fit and can therefore be ignored (z = 0). The ionization timescale can be used to calculate the age of the remnant. It is defined as τ = net and may be used
as an indication of the equilibration of the ions and electrons in the shocked interstellar medium. By dividing τ by ne, the age of the remnant can be obtained. For non-cosmological sources the
normalization is given by normalization = 10 −14 4πd2 Z nenHdV, (2)
in which d is the proper distance in cm, nethe electron density in cm−3, nHthe proton density in
cm−3 and V the volume of the analyzed region in cm3. [23] For the proper distance to the Large
Magellanic cloud 49.97 ± 1.30 kpc is taken. [24] When the volume is known, the value of nenH
can be obtained using formula 2. The volume of a region is assumed to be a cylinder taken from the remnant which is assumed to be spherical. The height of the cylinder is then assumed to be the chord length of a central angle in a circle as seen in figure 7. The chord length can then be calculated by
c =p4h(2R − h), (3)
where c is the chord length, h the height of the circular section and R the radius of the circle. By assuming that ne ≈ 1.2nH, ne can be calculated together with the age of the remnant. It must
be noted that the highest found age is taken as a lower limit for the age of the remnant, as it is the region that has been shocked first because the ionization age is the time past since the gas was shocked. There should also be an upper limit for the age of the remnant, but that can not be constrained directly by the results that are obtained.
3 ANALYSIS 15
Figure 7: An illustration of the chord length with c the chord length, d the height of the triangular section, h the height of the circular section, θ the central angle and S the arc length. Image retrieved from: https://en.wikipedia.org/wiki/Circular_segment
Another way to determine the age of the remnant from the results, is by using equation 1. The average plasma temperature can be obtained from the results. From the Sedov-Taylor model follows that the age can be calculated by
t = 2 5
Rs
Vs
, (4)
with t the age of the remnant in seconds, Rs the shock radius in kilometers and Vs the shock
velocity in km/s. This value must also be treated as a lower value for the age of the remnant as the real temperature can be higher than the electron temperature due to temperature non-equilibrium. A higher temperature results in a higher shock velocity and therefore a lower age of the remnant as it is dependant on the electron temperature rather than the real temperature. The real temperature is however a decent estimate of the electron temperature.
4
Results
Figure 8a shows the spectrum of N63A with (orange) and without background (blue). The calcu-lated emission lines are shown in figure 8b together with the soft and medium X-ray band. When compared to the spectrum of Cassiopeia A, see figure 4, two very different spectra are observed. The spectrum of N63A shows a lot of photon counts in the soft and medium X-ray band, but the hard X-ray emission is ∼ 102− 103 times lower than the soft and medium bands, which can also
be seen in figure 9a. The photon counts in the hard X-ray band in individual regions were often below 40 counts and not sufficient for the fit. Therefore, this research focused mainly on the soft and medium X-ray bands ranging from 0.5 − 2.0 keV. This range nicely contains most of the main emission lines for oxygen, neon, iron, magnesium, and silicon as seen in sec. 2.1 and figure 9b. Figure 8a shows an emission line at around 7 keV for the observed spectrum, but this vanishes with the background subtraction as no clear signal but noise can be distinguished around 7 keV.
(a) The full spectrum of N63A (b) The soft and medium X-ray band of N63A with
emis-sion lines
Figure 8: The spectrum of N63A for the whole energy band and for the soft and medium X-ray bands. Blue is the observed spectrum and orange is the spectrum with the background subtracted.
(a) Inverted RGB image of the soft, medium and hard X-ray bands.
(b) Inverted RGB image of the main emission lines of O, Ne, Fe and Si.
Figure 9: Left: RGB image of N63A with shown in blue the soft ∼ 0.5-1.2 keV, green the medium ∼ 1.2-2.0 keV and red the hard X-ray band ∼ 2.0-7.0 keV. Right: RGB image of N63A with blue the main oxygen lines ∼ 0.55-0.65 keV, green the main neon and iron lines ∼ 0.9-1.1 keV and red the main silicon lines ∼ 1.7-1.9 keV.
4 RESULTS 17
From the 110 regions analyzed, the 12 best fits were selected for further calculations, see figure 10. Other regions showed to0 high values of reduced χ2 to be considered good fits. Table 2 and 3 show the fitting results. The ejecta are found in the regions where the abundance is higher than the LMC abundance. Most regions observed show higher values for oxygen, neon and magnesium, while silicon and iron ejecta are scarce.
O Ne Mg Si Fe χ˜2 Reg 2 0.23+0.13−0.07 0.31+0.14−0.07 0.27+0.24−0.16 0.24+0.27−0.19 0.20+0.14−0.08 0.888 Reg 3 0.20+0.11−0.03 0.27+0.15−0.06 0.19+0.16−0.08 0.19+0.14−0.13 0.12+0.05−0.03 1.214 Reg 10 1.52+1.20−0.69 0.79+1.11−0.55 0.81+0.96−0.43 0.56+0.49−0.43 0.28+0.17−0.10 0.847 Reg 20 0.34+0.83−0.18 0.25+0.26−0.16 0.38+0.33−0.20 0.87+0.59−0.35 0.25+0.09−0.06 1.096 Reg 50 0.40+0.14−0.10 0.23+0.14−0.08 0.20+0.29−0.20 0.23+0.84−0.23 0.07+0.23−0.03 0.949 Reg 70 1.11+0.24−0.24 0.81+0.19−0.19 0.45+0.17−0.17 0.45+0.28−0.28 0.23+0.10−0.10 0.906 Reg 77 1.28+0.35−0.30 0.86+0.39−0.62 0.36+0.26−0.21 0.25+0.13−0.11 0.25+0.04−0.05 1.19 Reg 97 0.19+0.11−0.07 0.28+0.20−0.16 0.36+0.23−0.16 0.49+0.35−0.24 0.17+0.07−0.05 1.182 Reg 98 0.50+1.08−0.34 0.46+0.78−0.22 0.49+0.54−0.25 0.51+0.49−0.27 0.22+0.12−0.07 0.895 Reg 104 0.38+0.92−0.27 0.17+0.47−0.13 0.23+0.27−0.16 0.55+0.41−0.28 0.22+0.09−0.06 1.089 Reg 108 0.22+0.63−0.10 0.44+0.23−0.15 0.43+0.35−0.20 0.94+0.58−0.39 0.23+0.11−0.07 1.309 Reg 109 0.15+0.70−0.08 0.29+0.22−0.11 0.23+0.19−0.12 0.47+0.26−0.22 0.19+0.08−0.07 0.902 LMC 0.21 0.28 0.33 0.69 0.35
Table 2: The elemental abundances with respect to the solar abundance found by Wilms et al. [21] for different regions in N63A. The LMC values are below to allow for comparison. All errors are for a 90% confidence range.
Figure 10: The 12 regions with the best fitting results projected on N63A. The photon count scale is linear for this image.
As shown in table 3 the values for the absorption are mostly around the expected value of ∼ 1.5 × 1021. The temperature of all regions is between 0.51 and 1.01 keV with an average of
0.735+0.087−0.041keV. All regions have an the ionization age ranging from 0.62 − 99.7 × 1011s cm−3. In
some cases only a lower bound was found by XSPEC because XSPEC calculates the error for the ionization age on a logarithmic scale. Region 77 shows the highest ionization age of all regions. Region 3 is not in table 4 as it lies outside of the remnant in one of the in sec. 2.4 mentioned ”crescent”-shaped structures. The theory was that, similar to the Vela supernova remnant, these structures are high speed ejecta clumps interacting with the interstellar medium. [3] However, table 2 shows us that the abundances for this area lie around or below the Large Magellanic cloud abundances. [25]
nH (1021cm−2) kT (keV) tau (1011s cm−3) normalization (10−4) χ˜2 Reg 2 0.69+0.06−0.04 1.01+0.42−0.24 0.62+0.38−0.27 0.69+0.40−0.28 0.888 Reg 3 1.33+0.40−0.34 0.61+0.12−0.11 0.99+0.62−0.34 3.20+1.44−1.04 1.214 Reg 10 1.65+0.72−0.55 0.66+0.06−0.09 7.76−4.95 1.57+1.18−0.62 0.847 Reg 20 1.07+0.63−0.49 0.70+0.07−0.08 1.63+1.67−0.69 1.23+0.55−0.41 1.096 Reg 50 2.47+1.79−2.39 0.51+0.56−0.19 1.74−1.13 2.90+6.52−2.89 0.949 Reg 70 0.69+0.18−0.12 0.73+0.03−0.03 4.72+2.11−1.17 2.05+0.10−0.11 0.906 Reg 77 0.94+0.27−0.14 0.75+0.04−0.03 99.7−56.4 4.15+1.27−0.65 1.19 Reg 97 (3.86 · 10-6)+0.07 −1.40·10-6 0.75 +0.73 −0.12 1.08 +0.76 −0.38 1.10 +0.49 −0.27 1.182 Reg 98 0.83+0.98−0.71 0.74+0.07−0.11 1.98+3.16−0.88 1.40+0.86−0.57 0.895 Reg 104 0.40+1.16−0.40 0.73+0.07−0.10 1.94+9.56−1.04 2.30+1.06−0.65 1.089 Reg 108 0.75+1.57−0.74 0.67+0.11−0.16 1.22+1.00−0.47 1.95+1.92−0.69 0.902 Reg 109 1.08+1.95−0.94 0.69+0.10−0.23 1.44+2.32−0.46 4.05+7.01−1.29 0.902
Table 3: The fitted values for the absorption, temperature, ionization age and normalization for the best-fit regions. All errors are for a 90% confidence range.
From the observation a radius of 34.035±0.200was obtained with a scale of 100: 7.4720±0.1944×1017 cm. The error is due to N63A not being perfectly spherical. The distance d, see sec. 3.2, was determined by measuring the distance from the center of the remnant to the center of the region with an error of 0.0100 in the measurement. The radius of the region is obtained from the region
information in DS9 and has a precision of 0.00100 so the error is estimated at 0.000500. ne (cm−3) nH(cm−3) age (yr) temperature (106K)
Reg 2 1.92+1.14−0.81 1.60+0.95−0.68 1030+833−623 11.68+4.85−2.77 Reg 10 4.66+3.54−1.89 3.98+2.95−1.58 5276−3988 7.69+0.74−1.00 Reg 20 6.77+3.04−2.30 5.64+2.53−1.91 765+855−415 8.12+0.88−0.91 Reg 50 9.53+21.4−9.56 7.94+17.9−7.94 577−690 5.92+6.45−2.24 Reg 70 5.16+0.43−0.44 4.30+0.35−0.36 2902+1315−757 8.51+0.34−0.35 Reg 77 9.95+3.30−1.98 8.29+2.75−1.65 3175−1905 8.75+0.50−0.37 Reg 97 3.78+1.69−0.98 3.15+1.41−0.81 906+753−395 8.65+1.48−1.44 Reg 98 4.26+2.62−1.75 3.55+2.19−1.46 1804+2398−732 8.58+0.85−1.29 Reg 104 7.47+3.49−2.21 6.22+2.91−1.85 823+4075−503 8.48+0.85−1.14 Reg 108 5.22+5.13−1.88 4.35+4.28−1.56 745+953−391 7.83+1.25−1.82 Reg 109 8.60+14.9−0.74 7.17+12.4−0.61 532+1257−175 7.83+1.21−2.64
Table 4: The calculated values for the electron and proton density, age and temperature of N63A. Region 3 is not included as it lies outside of the shock wave. All errors are for a 90% confidence range.
All densities found are higher than the density of the interstellar medium ∼ 1 cm−3. The lowest electron density is found for region 2 at 1.92+1.14−0.81cm−3 and the highest value is 9.95+3.30−1.98cm−3 for region 77. The same applies for the proton density since it was assumed in sec. 3.2 that ne ∼ 1.2nH. The highest value found for the age of N63A is 5276−3988 year which is above the
expected age of 2000-5000 years, but still inside considering the error on the age. The second highest age is 3175−1905 year for region 77. By using the Sedov-Taylor model, see equation 4, the
5 DISCUSSION 19
5
Discussion
In determining the volume of the selected regions, it has been assumed that these have a uniform density for simplicity. However, as a strong shock compresses the volume with a factor 4, this is not a completely accurate assumption. The thickness of the filled shell is ∼ 9.1% of the radius of N63A, implying that close to the edge of the remnant the assumption of uniform density is acceptable, but near the middle of N63A this could potentially result in lower values for the density than the actual density. Therefore, a second approach to the volume could change the results. A smaller sphere with radius r = 0.909Rs can be used to calculate the part of the chord
length that is not in the shocked shell c1. The volume of the region is now defined by
V = 2πr2reg∗ (c − c1) for d < r and V = 2πr2reg∗ c for d > r (5)
with V the volume of the region in cm3, r
reg the radius of the region in cm, c the chord length in
cm (see sec.3.2) and c1 the chord length in the smaller sphere in cm. This estimate is acceptable
for regions near the center of the remnant. Nearer to the edge however, this leads to a problem with the filling factor as the cylindrical surface is not perpendicular to the surface of the sphere. A second option is to divide the volume of the region by 1 + 3√r2− d2 for r > d with r the radius
of the same sphere to account for the factor four in compression, but this lacks accuracy and is therefore not a better estimate.
The values found for the lower limit on the age of N63A are not consistent as seen in table 4. This could be explained by taking a look at the definition of the ionization age. The ionization age is defined as the time elapsed since the area was shocked multiplied by the electron density. If the region has not been shocked right after the supernova explosion, the time elapsed is less than the age of the remnant and thus the ionization time should be lower for a region with the same electron density. The age found with the Sedov-Taylor model might be a better lower limit. The region with the highest age lies in the middle of the area with higher expected absorption values due to a dense cold cloud between N63A and Earth and shows a large error on the ionization age which can indicate a possible fitting difficulty. [3] Behind region 10, region 70 and 77 had the highest respective ages of 2902+1315−757 and 3175−1905yr, the upper limit is not constraint as XSPEC
calculates the error on a logarithmic scale which reached a hard limit in the error calculation. As these values are close and better constrained, they can be assumed to be better estimates for a lower limit on the age of N63A. Because the value of the age found with the Sedov-Taylor model is model dependent and because the Sedov-Taylor assumption of the shock velocity only works well for fully equilibrated plasma, it can only be taken as an indicator and not as a lower limit. [26] The results in table 3 have large errors. This can be due to the fact that a degeneracy between some of the fit parameters exists. Especially a degeneracy between the absorption and the tem-perature, see figure 11a, the absorption and the normalization, see figure 11b and the temperature and ionization time, see figure 11c. If the parameters were virtually independent, the contours are expected to be circle shaped. For all three contour plots it is clear that the shape is not spherical. Figure 11a and 11b are both shaped elliptical indicating a dependence of the param-eters. The dependence of the absorption and temperature is because a higher absorption could indicate less emission and therefore a lower temperature. The dependence between the absorption and normalization can be explained because the model assumes more emission coming from the emission lines for different elements due to higher absorption and thus increases the normalization of the fit. For the contour plot of the temperature and ionization, figure 11c the shape differs from the other two contour plots, more leg shaped than elliptical as shown in figure 11. As the ionization age approaches the boundary of ∼ 1012s cm−3, a higher temperature is beneficial for the ionization as the available energy to equilibrate the ions and electrons increases. Therefore, lower temperatures do not allow for high ionization ages because not enough time has elapsed yet to achieve full equilibration of ions and electrons. Another explanation for the high error values is that XSPEC calculates most errors on a logarithmic scale, quickly increasing in step size. This can lead to unnecessary large errors and possibly explain some of the large errors found for the
ionization age and normalization. An example of this is that XSPEC sometimes was not able to find upper limits on the ionization age because this limit could quickly surpass the boundary at ∼ 1012s · cm−3 which caused the fit to get stuck at a high limit set by the nei model.
(a) Confidence contour plot of the absorption and temperature.
(b) Confidence contour plot of the absorption and normalization
(c) Confidence contour plot of the temperature and ion-ization age.
Figure 11: Confidence contour plots showing the red 67% (1σ), green 95% (2σ) and blue 100% (3σ) confidence ranges. The cross marks the best fit position.
6 CONCLUSION 21
6
Conclusion
The study performed in this thesis is to measure the temperature, constrain the electron and proton density and determine the age of N63A. This is done to enable the supernova remnant research group at API to study the possible cosmic ray source in N63A. Important to perform this study is the density of N63A. An analysis of 110 regions around N63A was conducted with a two component model consisting of an absorption component, the Tuebingen-Boulder Interstellar medium absorption model (tbabs), and an emission component, the non-equilibrium ionization collisional plasma model(nei). The average temperature of the electron density was found to be ∼ 1.92+1.14
−0.81 − 9.95 +3.30
−1.98cm−3 and the proton density ∼ 1.60 +0.95
−0.68 − 8.29 +2.75
−1.65 cm−3. The age
of N63A is estimated between 2000-5000 yr. [3] Using the Sedov-Taylor model an assumption for the lower limit of the age is found at 4120+195−92 yr. The Chandra data gave a lower limit of 3175−1905yr. A higher value for the age was found, 5276−3988yr, but since this lies outside of the
estimated age of 2000-5000 yr and above the Sedov-Taylor solution which assumes full electron-ion equilibratation that has not been reached by N63A as found in this study, this value has been omitted as lower limit. [26]
References
[1] J. A. Johnson, “Populating the periodic table: Nucleosynthesis of the elements,” Science, vol. 363, no. 6426, pp. 474–478, 2019.
[2] J. Vink, “Supernova remnants: the X-ray perspective,” , vol. 20, p. 49, Dec. 2012.
[3] J. S. Warren, J. P. Hughes, and P. O. Slane, “Revealing New Physical Structures in the Supernova Remnant N63A through Chandra Imaging Spectroscopy,” , vol. 583, pp. 260–266, Jan. 2003.
[4] J. Hinton and W. Hofmann, “Teraelectronvolt astronomy,” Annual Review of Astronomy and Astrophysics, vol. 47, p. 523–565, Sep 2009.
[5] S. G. Ryan and A. J. Norton, Stellar evolution and nucleosynthesis. Cambridge University Press, 2010.
[6] E. Cappellaro and M. Turatto, Supernova Types and Rates, pp. 199–214. Dordrecht: Springer Netherlands, 2001.
[7] S. Broersen, X-ray spectral analysis of non-equilibrium plasmas in supernova remnants. PhD thesis, 2014.
[8] D. Leahy, “Supernova remnant evolution: from explosion to dissipation,” 2016.
[9] R. A. Chevalier, “The radio and X-ray emission from type II supernovae.,” , vol. 259, pp. 302– 310, Aug. 1982.
[10] L. I. Sedov, Similarity and Dimensional Methods in Mechanics. 1959.
[11] G. Taylor, “The Formation of a Blast Wave by a Very Intense Explosion. I. Theoretical Discussion,” Proceedings of the Royal Society of London Series A, vol. 201, pp. 159–174, Mar. 1950.
[12] J. K. Truelove and C. F. McKee, “Evolution of Nonradiative Supernova Remnants,” , vol. 120, pp. 299–326, Feb. 1999.
[13] Y. Uchiyama, S. Funk, J. Katsuta, H. Katagiri, M. Lemoine-Goumard, H. Tajima, T. Tanaka, and D. Torres, “Fermi large area telescope discovery of gev gamma-ray emission from the vicinity of snr w44,” Astrophysical Journal Letters, vol. 749, 04 2012.
[14] “About chandra: Chandra mission.” https://chandra.harvard.edu/about/axaf_ mission.html, July 2014, accessed at July 23, 2020.
[15] “The european photon imaging camera (epic) onboard xmm-newton.” https://www.cosmos. esa.int/web/xmm-newton/technical-details-epic, accessed at July 23, 2020.
[16] “Suzaku: Spacecraft.” http://www.isas.jaxa.jp/en/missions/spacecraft/others/ suzaku.html, accessed at July 23, 2020.
[17] K. A. Arnaud, R. K. Smith, and A. Siemiginowska, Handbook of X-ray astronomy. Cambridge University Press, 2011.
[18] “Emission lines.” http://www.atomdb.org/Webguide/webguide.php, accessed at July 23, 2020.
[19] J. P. Hughes, I. Hayashi, and K. Koyama, “Ascax-ray spectroscopy of large magellanic cloud supernova remnants and the metal abundances of the large magellanic cloud,” The Astro-physical Journal, vol. 505, no. 2, p. 732–748, 1998.
REFERENCES 23
[20] S. C. Russell and M. A. Dopita, “Abundances of the Heavy Elements in the Magellanic Clouds. III. Interpretation of Results,” , vol. 384, p. 508, Jan. 1992.
[21] J. Wilms, A. Allen, and R. McCray, “On the Absorption of X-Rays in the Interstellar Medium,” , vol. 542, pp. 914–924, Oct. 2000.
[22] “tbabs, ztbabs, tbfeo, tbgas, tbgrain, tbpcf, tbvarabs, tbrel: Ism grain absorption.” https:// heasarc.gsfc.nasa.gov/docs/xanadu/xspec/manual/node265.html, accessed on July 23, 2020.
[23] “nei, vnei, vvnei: collisional plasma, non-equilibrium, constant temperature.” =https://heasarc.gsfc.nasa.gov/docs/xanadu/xspec/manual/node195.html, accessed on July 23, 2020.
[24] G. Pietrzy´nski, D. Graczyk, W. Gieren, I. B. Thompson, B. Pilecki, A. Udalski, I. Soszy´nski, S. Koz lowski, P. Konorski, K. Suchomska, and et al., “An eclipsing-binary distance to the large magellanic cloud accurate to two per cent,” Nature, vol. 495, p. 76–79, Mar 2013. [25] P. Maggi, F. Haberl, P. J. Kavanagh, M. Sasaki, L. M. Bozzetto, M. D. Filipovi´c, G.
Vasilopou-los, W. Pietsch, S. D. Points, Y. H. Chu, J. Dickel, M. Ehle, R. Williams, and J. Greiner, “The population of X-ray supernova remnants in the Large Magellanic Cloud,” , vol. 585, p. A162, Jan. 2016.
[26] J. Vink and L. Kuiper, “Supernova remnant energetics and magnetars: no evidence in favour of millisecond proto-neutron stars,” , vol. 370, pp. L14–L18, July 2006.
