We conduct a broad investigation about stellar mass and population diagnostics in order to formulate novel constraints related to the formation and evolution of galaxies from a nearby cluster environment. Our work is powered by the use of stellar population models which transform galaxy colours and/or absorption line strengths into estimates of its stellar properties. As input to such models, we assemble an extensive compilation of age and chemical abundance information for Galactic globular clusters. This compilation allows a confident expansion of these models into new regions of parameter space that promise to refine our knowledge of galactic chemical evolution. We then draw upon a state-of-the-art spectroscopic and photometric survey of the Virgo galaxy cluster in order to constrain spatial variations of the stellar ages, metallicities, and masses within its member galaxies, and their dynamical masses. We interpret these data in the context of the histories of star formation, chemical enrichment, and stellar mass assembly to formulate a broad picture of the build-up of this cluster’s content over time. In it, the giant early-type galaxies formed through highly dissipational processes at early times that built up most of their stellar mass and drew significant amounts of dark matter within their optical radii. Conversely, dwarf early-types experienced environmental processes that quenched their star formation during either the early stages of cluster assembly or upon infall at later times. Somewhat perplexing is our finding that the internal dynamics of these galaxies are largely explained by their stellar masses. Lastly, Virgo spirals also suffer from their dense environment, through ram pressure stripping and/or tidal harrassment. In addition to quenching, these effects leave an imprint on their internal dynamical evolution too. Late-type spirals exhibit evidence of having ejected significant amounts of baryons from their inner regions, likely via energetic feedback events. Rich as our picture of the history of the Virgo cluster has become, real progress in our understanding of this system will truly benefit from future high-resolution cosmological and hydrodynamic simulations of this environment. Such simulations are still in their infancy, but the data assembled here should soon provide their most direct validation. / Thesis (Ph.D, Physics, Engineering Physics and Astronomy) -- Queen's University, 2013-09-30 23:32:48.575

The recent discovery of a new population of Active Galactic Nuclei (AGN) – the Radio Loud Narrow Line Seyfert 1 (RL NLS1) galaxies – at g-ray energies by Fermi has prompted intense interest among researchers, as evidence mounts that these objects may in fact compose a new class of blazars. If RL NLS1s are indeed a new class of blazars, or at least analogous to them, then the detection of certain blazar-like properties would be expected. These properties include significant variability at all wavelengths across the electromagnetic spectrum and on timescales from minutes to years, significant & variable polarization in the radio and optical regimes, significant & variable high-energy emissions (especially in the g-ray regime), and a double-peaked structure of their spectral energy distributions. This dissertation seeks to characterize several of these properties for RL NLS1s as a class. These include the degree to which these objects are polarized and the variability of this property, the detection and characterization of these sources at g-ray energies with the Large Area Telescope aboard the Fermi spacecraft, and the degree to which these properties are interdependent. A photopolarimetric survey (the first of its kind for this class of objects) and g-ray monitoring program were conducted by the author in order to obtain the data necessary for this project. The measurements obtained via these observations are used to characterize this interesting class of objects with respect to a sample of blazars. In general, it was found that the polarizations of these objects fall between radio-quiet NLS1s and FSRQ-type blazars, and were most similar to HBL-type blazars. The 7 RL NLS1s in this sample that had been detected above a Test Statistic (TS) of 25 by Fermi/LAT were most similar to FSRQs, while 9 objects detected in the interval 9 < TS < 25 shared several properties with HBLs. Two RL NLS1s - J1443+4725 and J1644+2619 - are identified as high-confidence (TS > 25) g-ray sources for the first time, bringing the total number of members of this class firmly detected at g-ray energies to 8. The gamma-ray spectra of RL NLS1s are similar to FSRQs, though some have steeper spectra.

The discovery of the Spite plateau in the abundances of 7Li for metal-poor stars led to the determination of an observationally deduced primordial lithium abundance. However, with the determination of the baryon density, Omega_B_h^2, from the Wilkinson Microwave Anisotropy Probe (WMAP) data, a discrepancy arose between observationally determined and theoretically determined abundances of 7Li. This is what has become known as the “lithium problem”. Of all the uncertain factors in determining a stellar Li abundance, the effective temperature is the most important. This thesis is concerned with determining an accurate effective temperature scale for metal-poor halo dwarfs, paying specific attention to eliminating any possible systematic errors. This is done by utilising the exponential term, Chi/T, of the Boltzmann equation. Two assumptions are adopted; firstly the simplifying assumptions of local thermodynamic equilibrium (LTE), and secondly the more sophisticated techniques of non-local thermodynamic equilibrium (NLTE). The temperature scales are compared to others derived using different techniques; a photometric scale, where I find comparable Teff in LTE and hotter temperatures by an average of ~ 150 K in NLTE; a scale derived using Balmer lines, for which I have comparable values in LTE and hotter Teff values, by typically 110 K – 160 K, in NLTE; and finally a scale derived using an infrared flux method (IRFM). Here I find their Teff values are hotter by ~ 250 K for LTE and ~ 190 K in NLTE. Lithium abundances are then calculated for the program stars and a mean Li abundance is derived. I find values ranging from A(Li) = 2.10 dex – 2.16 dex with the LTE scales and A(Li) = 2.19 dex – 2.21 dex for the NLTE scales. These mean Li abundances are compared to other observationally deduced abundances, for which I find comparable results in LTE and higher values in NLTE, and to the WMAP + big bang nucleosynthesis calculated Li abundance. I find that my new values are still considerably lower than the WMAP value and are therefore unable to reconcile the lithium problem. Second to this primary investigation, I use Ti as an independent test of the derived Teff values and log g’s. I find that Ti is not a useful constraint on the temperatures or, therefore, on the lithium problem. I also assess the impact of the new Teff scales on the different models of Galactic chemical evolution (GCE), comparing newly calculated abundances with GCE determined abundances. It was found that trends exist in several of the elements; however, these were not statistically relevant. Also a larger degree of scatter was found in the abundances compared to the Arnone et al. (2005). This scatter was not to the degree found in the Argast et al. (2000). Reasons for the differences have been discussed.

In this thesis I use X-ray observations to study the cores and extended structures of radio-loud AGN, to determine their structure, accretion properties and the impact they have on their surroundings. I use new Chandra data and archival XMM-Newton observations ofMarkarian 6 to look for evidence of emission from shocked gas around the external radio bubbles, using spatially resolved regions in Chandra and spectral analysis of the XMM data. The results show that the bubbles in Mrk 6 are indeed driving a shock into the halo of the host galaxy, with a Mach number of 3.9. I also find that the spectrum of the AGN has a variable absorbing column, which changes from 8 × 1021 atoms cm−2 to 3 × 1023 atoms cm−2 on short timescales (2-6 years). This is probably caused by a clump of gas close to the central AGN, passing in front of us at the moment of the observation. Using new and archival Chandra observations of the Circinus galaxy, I match them to pre-existing radio, infrared and optical data to study the kpc-scale emission. As for Mrk 6, I find that the radio bubbles in Circinus are driving a shock into the interstellar medium of the host galaxy, with Mach numbers M 2.7–3.6 and M 2.8–5.3 for the W and E shells respectively. Comparing the results with those we previously obtained for Centaurus A, NGC 3801 and Mrk 6, I show that the total energy in the lobes (thermal+kinetic) scales approximately with the radio power of the parent AGN. The spatial coincidence between the X-ray and edge-brightened radio emission in Circinus resembles the morphology of some SNR shocks, a parallel that has been expected for AGN, but has never been observed before. I investigate what underlying mechanisms both types of systems may have in common, arguing that, in Circinus, the edge-brightening in the shells may be accounted for by a B field enhancement caused by shock compression, but do not preclude some local particle acceleration. I also carry out a systematic study of the X-ray emission from the cores in the 0.02 < z < 0.7 2Jy sample, using Chandra and XMM-Newton observations. I combine the results with the mid-IR, optical emission line and radio luminosities, and compare them with those of the 3CRR sources, to show that the low-excitation objects in our sample show all the signs of radiatively inefficient accretion. I study the effect of the jet-related emission on the various luminosities, confirming that it is the main source of soft X-ray emission for our sources. I also find strong correlations between the accretion-related luminosities, and identify several sources whose optical classification is incompatible with their accretion properties. I derive the bolometric and jet kinetic luminosities for the sample and find a difference in the total Eddington rate between the low and high-excitation populations, with the former peaking at 1 per cent and the latter at 20 per cent Eddington. There is, however, an overlap between the two, indicating that a simple Eddington switch may not be possible. The apparent independence of jet kinetic power and radiative luminosity in the highexcitation population in our plots allows us to test the hypothesis in which jet production and radiatively efficient accretion are in fact independent processes that can coexist in high-excitation objects.

I have used blue near-infrared colours to select a group of UKIDSS T dwarfs with spectral types later than T4. From amongst these I identify two kinematic halo T-dwarf candi- dates. Blue near-infrared colours have been attributed to collisionally-induced hydrogen absorption, which is enhanced by either high surface gravity or low metallicity. Proper motions are measured and distances estimated, allowing the determination of tangential velocities. U and V components are estimated for our objects by assuming Vrad = 0. From this, ULAS J0926+0835 is found to have U = 62 kms−1 and V = −140 kms−1 and ULAS J1319+1209 is found to have U = 192 kms−1 and V = −92 kms−1. These values are consistent with potential halo membership. However, surprisingly, these are not the bluest objects in the selection. The bluest is ULAS J1233+1219, with J −K = −1.16±0.07, and surprisingly this object is found to have thin disc-like U and V . Our sample also contains Hip 73786B, which I find to be a companion to the star Hip 73786. Hip 73786 is a metal- poor star, with [Fe/H]= −0.3 ± 0.1 and is located at a distance of 19±0.7 pc. U, V,W space velocity components are calculated for Hip 73786A and B, finding that U = −48±7 kms−1, V = −75 ± 4 kms−1 and W = −44 ± 8 kms−1. From the properties of the pri- mary, Hip 73786B is found to be at least 1.6Gyr old. As a metal poor object, Hip 73786B represents an important addition to the sample of known T dwarf benchmarks. Using mid-infrared data from WISE, I also identify T dwarfs with abnormally-red H − W2 and consider possible causes for their extreme colours. In particular I exam- ine three prominent examples of this phenomenon, ULAS J1416+1348B, 2MASS J0939- 2448 and BD+01o 2920B. A plot of spectral type against MW2-magnitude suggests that ULAS J1416+1348B is potentially an unresolved binary, similar to 2MASS J0939-2448. However, the plot also indicates that BD+01o 2920B is not an unresolved binary. I also present new FIRE spectroscopy for ULAS J1416+1348B and 2MASS J0939-2448. These data show that ULAS J1416+1348B has a similar shape to the Y -band spectrum to that of BD+01o 2920B, thus suggesting that the two objects have a similar metallicity, whereas 2MASS J0939-2448 appears to be a more metal-rich object. Using a new parallactic dis- tance, I derive a luminosity of (6.9±0.7)×1020W for ULAS J1416+1348B. I also find a radial velocity of −39 ± 1 kms−1 for this object. The agreement between this and that of the L dwarf SDSS J1416+1348A confirms that these two objects are physically-associated. I also present a set of simulated unresolved binaries; the colours of these systems do not appear to redden significantly with the addition of cooler companions. From this, I suggest that the colours of ULAS J1416+1348B and BD+01o 2920B cannot be solely attributed to any possible unresolved companions; for these two objects, composition and/or surface gravity must be playing a substantial role. Consideration of model predictions provides extra evidence for this argument, showing as it does that high log g and low metallicity can redden H − W2 colours by as much as »0.5mag as compared to a high-metallicity and low log g object of the same effective temperature. I also present kinematics and photometry for several more new candidate low-metallicity T dwarfs. Spectra are also presented, where available. In addition I provide new follow-up JHK spectroscopy for ULAS J0926+0835, ULAS J1233+1219 and ULAS J1319+1209. These new spectra allow full JHK-based spectral typing for these objects.

The principal theory concerning the origin of the elements heavier than the Fe-peak, such as Ba, strongly suggest that for old, metal-poor environments, the rapid (r-) process is the most likely path taken in their synthesis, while the slow (s-) process becomes more substantial in younger, more metal-rich stellar populations. In this work I test this theory by evaluating the isotope ratios of Ba. It is understood that Ba consists of seven stable isotopes, five of which are synthesised by the two neutron-capture processes. The two odd isotopes, 135,137Ba, as well as 138Ba are synthesised via both the r- and s-processes while two of the even isotopes, 134,136Ba are synthesised via the s-process only. The relative contribution of the r- and s-process to these isotopes can be understood via nucleosynthesis calculations and is described using the parameter fodd, where fodd = [N (135Ba) + N (137Ba)] /N (Ba). Low values of fodd (~0.11) indicate an s-process regime, while high values of fodd (~0.46) indicate an r-process regime. In the Ba II 4554 A line the even isotopes lie close to the line centre, while the odd isotopes, which are hyperfine split because of their non-zero nuclear spin, lie in the wings of the line. From an analysis of the line profile shape, one can determine whether Ba has been synthesised primarily through the r-process or s-process; a broad, asymmetric line would indicate a high r-process contribution, while a line with a deeper core and shallower wings would indicate a high s-process contribution. Using the radiative transfer code ATLAS, which assumes local thermodynamic equilibrium (LTE) and employs 1-dimensional (1D) KURUCZ06 model atmospheres, I synthesised line profiles for six metal-poor stars: HD140283, HD122563, HD88609, HD84937, BD-04 3208 and BD+26 3578 - for a range of isotope ratios. All six are of sufficiently low metallicity that Ba was expected to have an r-process origin. These were fit to high resolution (R\equiv \lamda/\Delta\lamda = 90 000 - 95 000), high signal-to-noise to the Ba II 4554 A line which has multiple components. In the first test, synthetic spectra were computed using the non local thermodynamic equilibrium (NLTE) radiative transfer code MULTI. The synthetic line profiles were fit to a number of lines in HD140283. Although this technique might have improved the fit in the line core, it was found that such a treatment did not improve upon fitting errors associated with the best fit 1D LTE synthetic profiles. The second test used a 3-dimensional (3D) radiative transfer code (LINFOR3D) that employed 3D, time-dependent atmospheres produced with CO5BOLD. The 3D synthetic pro les were fit to a selection of Fe lines and improvements over the poor fits produced by the 1D LTE synthesis were seen. It was found that the 3D synthesis could almost completely reproduce the line asymmetries seen in the observed stellar spectrum. This result suggests that further work to refine the 3D calculations and synthesis code would be valuable.

The identification of the relic particles which presumably constitute cold dark matter is a key challenge for astroparticle physics. Indirect methods for their detection using high energy astro- physical probes such as cosmic rays have been much discussed. In particular, recent ‘excesses’ in cosmic ray electron and positron fluxes, as well as in microwave sky maps, have been claimed to be due to the annihilation or decay of dark matter. In this thesis, we argue however that these signals are plagued by irreducible astrophysical backgrounds and show how plausible con- ventional physics can mimic the alleged dark matter signals. In chapter 1, we review evidence of, and possible particle candidates for, cold dark matter, as well as our current understanding of galactic cosmic rays and the state-of-the-art in indirect detection. All other chapters contain original work, mainly based on the author’s journal publications. In particular, in chapter 2, we consider the possibility that the rise in the positron fraction observed by the PAMELA satellite is due to the production through (hadronic) cosmic ray spallation and subsequent acceleration of positrons, in the same sources as the primary cosmic rays. We present a new (unpublished) analytical estimate of the range of possible fluctuations in the high energy electron flux due to the discreteness of plausible cosmic ray sources such as supernova remnants. Fitting our result for the total electron-positron flux measured by the Fermi satellite allows us to fix the only free parameter of the model and make an independent prediction for the positron fraction. Our explanation relies on a large number of supernova remnants nearby which are accelerating hadronic cosmic rays. Turning the argument around, we find encouraging prospects for the observation of neutrinos from such sources in km^3-scale detectors such as IceCube. Chapter 3 presents a test of this model by considering similar effects expected for nuclear secondary-to-primary ratios such as B/C. A rise predicted above O(100)GeV/n would be an unique confirmation of our explanation for a rising positron fraction and rule out the dark matter explanation. In chapter 4, we review the assumptions made in the extraction of the `WMAP haze' which has also been claimed to be due to electrons and positrons from dark matter annihilation in the Galactic centre region. We argue that the energy-dependence of their diffusion means that the extraction of the haze through fitting to templates of low frequency diffuse galactic radio emission is unreliable. The systematic effects introduced by this can, under specific circumstances, reproduce the residual, suggesting that the ‘haze’ may be just an artefact of the template subtraction. We present a summary and thoughts about further work in the epilogue.

The two SCUBA HAlf-Degree Extragalactic Survey (SHADES) fields are amongst the richest places in the sky in terms of multi-wavelength coverage. They comprise an eastern section of the Lockman Hole (LH) and the central portion of the Subaru- XMM/Newton Deep Field (SXDF). In this thesis, I have obtained extremely deep, multi-frequency radio imaging of the SHADES fields using the GiantMetre-wave Radio Telescope (GMRT) and the Very Large Array (VLA), at 610MHz and 1.4GHz, respectively. These data are used to analyse the nature of the sub-milliJansky (sub-mJy) radio population, which has been hotly debated in the last few years: are they powered by star-forming or nuclear activity? To tackle the problem, I employ different approaches making use of the large variety of multi-wavelength data in the SHADES fields. I begin by analysing the spectral index, α610MHz 1.4GHz , of radio sources detected in the LH, to explore the dominant emission mechanism. Based on a robust 10 σ detection criterion, I find a constantmedian spectral index of α610MHz 1.4GHz ≈ −0.6 to −0.7 for sources between S1.4GHz ≈ 200 μJy and 10mJy. This result suggests that the galaxy population in the sub-mJy regime is powered by optically-thin synchrotron emission – starforming galaxies or lobe-dominated active galactic nuclei (AGN). Making use of X-ray observations in the LH, I show that the fraction of radio sources detected in the hard X-ray band (between 2 and 10 keV) decreases from 50 to 15 per cent between S1.4GHz ≈ 1mJy and . 100 μJy, which strongly suggests a transition from AGN to star-forming galaxies. Based on the deep, multi-wavelength coverage of the SXDF, I explore the behaviour of the far-infrared (FIR)/radio correlation as a function of redshift. I combine the q24 factor – the logarithmic flux density ratio between Spitzer 24-μm and VLA 1.4- GHz flux densities – with available photometric redshifts and find strong evidence that the correlation holds out to z ≈ 3.5. Based on M82-like k-corrections and using a high-significance (S1.4GHz > 300 μJy) radio sub-sample, I find a mean and scatter of q24 = 0.71 ± 0.47. Monte-Carlo simulations based on these findings show that fewer sources deviate from the correlation at fainter flux densities (i.e. fewer radioloud AGN). I predict that the radio-loud fraction drops from 50 per cent at ∼ 1mJy to zero at . 100 μJy. The validity of the FIR/radio correlation out to very high redshifts adds credibility to identifications of sub-millimetre (submm) galaxies (SMGs) made at radio wavelengths. Based on a sample of 45 radio-identified SMGs in the LH, I find a median radio spectral index of α610MHz 1.4GHz = −0.72 ± 0.07, which suggests that optically-thin synchrotron is the dominant radio emission mechanism. Finally, as anAppendix I include a theoretical treatment that constrains the average geometry of the dusty, torus-like structures believed to obscure a large fraction of the AGN population. I use the distribution of column densities (NH) obtained from deep ∼ 1Msec X-ray observations in the Chandra Deep Field South. I find that to reproduce the wide observed range of NH, the best torus model is given by a classical “donut”- shaped distribution with an exponential angular dependency of the density profile.

The centre of the Milky Way, commonly referred to as the Galactic Centre, is roughly that region within 500 pc of the central black hole, Sagittarius A*. Within the innermost parsec around the supermassive black hole Sagittarius A* are more than a hundred massive young stars whose orbits align to form one or possibly two discs. At about 100 pc is a ring containing more than ten million solar masses of molecular gas which could be the origin of some of the most massive star clusters in the Galaxy. I have performed a number of numerical simulations to help us understand how it is that these structures may have been formed. I firstly describe and test an improvement to the smoothed particle hydrodynamics code I used. This improves conservation of energy and momentum in certain situations such as in strong shocks from supernovae, which were to be included in a later chapter. The discs of massive stars around Sagittarius A* are believed to have been born there within fragmenting gaseous discs. This is problematic, as the formation of two stellar discs would require two gaseous counterparts. A method is described of forming multiple discs around a black hole from a single cloud's infall and subsequent tidal destruction. This is due to its prolate shape providing a naturally large distribution in the direction of the angular momentum vectors within the cloud. The resulting discs may then go on to form stars. Energetically, it would appear that a sequence of supernovae could potentially cause a giant molecular cloud to fall inwards towards the central black hole from an originally large orbit around the Galactic Centre. I simulate the impact on a giant molecular cloud of supernovae originating from a massive stellar cluster located a parsec away. Ultimately, the supernovae are found to have little effect. Finally, I simulate the formation of the dense ring of clouds observed in the Central Molec- ular Zone at a distance of about 100 pc from Sgr A*. Infalling gas is shown to be subject to such extreme tidal forces that a single cloud of gas is extended to form a long stream. The ribbon grows to the point that it self-intersects and forms a ring-like structure. Its complexity depends on the orbit of the original cloud. The position-velocity data is compared with observations, and similarities are noted.

All massive nearby galaxies, including our own, host supermassive black holes. Active galactic nuclei (AGN) are seen when such black holes accrete, and when they produce powerful jets of synchrotron-emitting plasma, they are termed radio-loud AGN. The close correlation between black hole mass and galaxy bulge mass in elliptical galaxies indicates that AGN feedback may be the key to the regulation of galaxy formation. It is thus necessary to fully understand the structure of AGN, the way that they are fuelled, and their duty cycle, in order to study the feedback processes and get a clear picture of galaxy formation. In this thesis, independent methods are developed to constrain the accretion disk and radio jet angles to the line of sight. H IX emission from a sub-sample of high-redshift quasars is measured from near-infrared spectroscopy and modelled as sums of different components, including the characteristic double-peaked profile which results from a thin, rotating accretion disk. Comparing the models using Bayesian evidence, almost all quasars were found to have infrared spectra consistent with the presence of a disk. The jet inclination angles of the same set of quasars were constrained by fitting a model, including the effect of Doppler boosting and the receding torus model for dust obscuration, to the radio \ spectral energy distribution. The fitted disk and jet angles correlate strongly, and are consistent with a model in which the radio jets are launched orthogonally to the plane of the accretion disk, as expected if the jet is powered by energy drawn from the spin of the black hole. Both disk and jet angles correlate with the observed linear source size, which is a projection effect; when deprojected using the fitted angles, the distribution of source sizes agrees with a scenario in which the sources expand into the surrounding medium at a constant rate up to ~ 1 Mpc and then shut off, probably as the nuclei become quiescent. The accretion disk angle was found to correlate weakly with the low-frequency radio luminosity, which provides direct, albeit tenuous, evidence for the receding torus model.