The Neutral Hydrogen Content of Satellite Galaxies in the Local Volume

Ananthan Karunakaran

A thesis submitted to the

Department of Physics, Engineering Physics & Astronomy

in conformity with the requirements for

the degree of Master of Science

Queen’s University

Kingston, Ontario, Canada

November, 2017

The properties of the satellite populations of galaxies are vital for understanding their structure and evolution in a cosmological context. Deep optical imaging surveys have revealed a host of low surface brightness (LSB) features near Local Volume galaxies, that may be satellites of the latter. We conduct atomic gas (HI) observations of 48 such satellite candidate galaxies discovered around 10 different hosts using the Robert C. Byrd Green Bank Telescope. Detections of HI reservoirs in these systems will provide spectroscopic distances that can be used to confirm an association to their putative hosts. Non-detections imply that the objects are gas-poor, irrespective of their distance along the line-of-sight. We first search for HI in six ultra-faint dwarf candidates discovered near NGC 3109. One is detected in HI, placing it in the background as a gas-rich field dwarf, while the remaining five have no detectable HI emission. We also searched for HI along the lines of sight to 27 LSB features that project around M101, and the majority (23/27) were found to be gas-poor. The other four were detected in HI, though only one has a recessional velocity consistent with that of M101. Two of the other galaxies have systemic velocities suggesting an association with a background group and the fourth detection places it in the background. Finally, we searched for HI in 15 satellite candidates around 8 hosts

i ii

finding just one gas-rich satellite. Using the HI and optical properties of all non- detections, we compare them with other samples of dwarf galaxies. We find that in order for our non-detections to be consistent with well-studied scaling relations for gas-rich field galaxies, the most plausible scenario is that the Local Volume LSBs are associated with their hosts, the M101 candidates are associated with it or the background group, and the satellite candidates around NGC 3109 are in fact field galaxies that have distances in the range 17.5-38 Mpc. Therefore, both detections and non-detections place important constraints on the physical properties of candidate satellite galaxies detected in the optical. Acknowledgments

It goes without saying that I would not have had the wonderful opportunity to conduct such interesting research and pursue my passion without the guidance and support from my supervisor, Kristine Spekkens. Thank you. To my friends in the department at Queen’s University, I have no idea where these last two years have gone but I am extremely thankful that I was able to spend it with such amazing people. I would like to thank Matthew Chequers, Zsolt Keszthelyi, Colin Lewis, James Sikora, and Cory Wagner for all of your advice and countless discussions, whether personal or research related. To my friends from FMM and Mac, I will always appreciate your support and for keeping me grounded. This is one step closer to the dream! I don’t know where I would be without the love, encouragement, and support from my family. I would like to speciﬁcally thank my sister, Amirtha, for always being there for me and keeping me focused on what is important.

“Sometimes science is more art than science, Morty. A lot of people don’t get that.”

iii Statement of Co-Authorship

All of the work presented in thesis was completed by the author, including the HI observations, data reduction, and analysis. In Chapter3, the optical imaging campaigns around NGC 3109 and the resulting photometric analysis (from which our HI follow-up target list was drawn) were conducted by collaborator D. Sand at the University of Arizona. The pre-publication optical properties of target galaxies, reduced images (e.g. Figure 3.1), and raw images used in this thesis were provided by him through private communication. In Chapter4, previous and ongoing HST imaging campaigns and preliminary analysis of the targeted M101 satellite candidates were conducted by collaborator P. Bennet at Texas Tech University. Distance information for M101-Dw6 gleaned from a preliminary analysis of those data was provided by him through private communication.

iv Contents

Abstracti

Acknowledgments iii

Statement of Co-Authorship iv

Table of Contentsv

List of Tables viii

List of Figures ix

List of Acronyms xii

List of Symbols xiv

Chapter 1 Introduction1

1.1 Cosmological Structure Formation ...... 6 1.2 Comparing Satellites in Simulations and Observations ...... 9 1.3 The Low Surface Brightness Universe...... 13 1.3.1 The Local Group...... 13 1.3.2 Dwarfs in Nearby Systems...... 15

v CONTENTS vi

1.3.3 Low Surface Brightness Galaxies ...... 18 1.4 HI in Low-Mass Galaxies ...... 22 1.4.1 Neutral Hydrogen...... 22 1.4.2 HI in the Local Group...... 24 1.4.3 Gas-Rich Field Galaxies ...... 27 1.5 This Work...... 31

Chapter 2 Observations and Data Reduction 33

2.1 Observations...... 33 2.1.1 Single-Dish Radio Telescopes ...... 33 2.1.2 The Green Bank Observatory...... 36 2.1.3 Instrument Setup...... 39 2.1.4 Observations with the GBT...... 40 2.2 Data Reduction...... 44 2.3 Conclusions...... 52

Chapter 3 Satellite candidates around NGC 3109 54

3.1 The NGC 3109 System and New Satellite Candidates...... 54 3.2 HI Observations and Analysis...... 59 3.2.1 Observations ...... 59 3.2.2 Analysis...... 60 3.3 Discussion...... 67 3.4 Conclusions...... 74

Chapter 4 The Satellite Population around M101 76

4.1 The M101 Group...... 76 CONTENTS vii

4.2 Observations and Analysis...... 82 4.3 Discussion...... 94 4.3.1 Satellites of M101...... 94 4.3.2 Galaxies in the Background ...... 101 4.4 Conclusions...... 103

Chapter 5 The Satellite Populations of Local Volume Galaxies 105

5.1 Galaxies in the Local Volume ...... 105 5.2 Observations and Analysis...... 108 5.3 Discussion and Conclusions ...... 114

Chapter 6 Discussion 119

6.1 Our Non-Detections as Satellites...... 120 6.2 Our Non-detections as Background Galaxies...... 124 6.3 Constraining Distances to Non-Detections ...... 131

Chapter 7 Conclusions and Future Work 133

7.1 Future Work...... 136

Bibliography 139 List of Tables

3.1 Properties of existing satellites and new satellite candidates around NGC 3109...... 58 3.2 Details of HI observations of the LSB features...... 60 3.3 Upper limits for range of distances of LSB features projected around NGC 3109...... 66 3.4 Properties of Dw1 ...... 66 3.5 Details of potential host galaxies for Dw1...... 68

4.1 M101 satellite candidates and their observations ...... 84 4.2 Upper limits for conﬁrmed satellites, satellite candidates, and background galaxies that are projected around M101...... 92 4.3 Properties of candidate satellites with HI detections ...... 93

5.1 Properties of Local Volume Hosts in our sample...... 107 5.2 Local Volume satellite candidates and their observations ...... 109 5.3 Upper limits for satellite candidates with HI non-detections calculated at the distance of their putative host ...... 112 5.4 Properties of the satellite with HI emission ...... 113

viii List of Figures

1.1 Examples of satellite dwarf galaxies ...... 5 1.2 Milky Way-size dark matter halos from Aquarius simulations .... 8 1.3 Cumulative mass function from ELVIS simulations for Milky Way-like halos...... 11 1.4 Luminosity-Size plane for faint satellites and star clusters of the Milky Way ...... 16 1.5 LSBs around M101 discovered by Merritt et al.(2014) ...... 20 1.6 Ratios of HI mass to V-band luminosity and HI mass to dynamical mass for Milky Way dwarf spheroidals ...... 27 1.7 Gas mass versus stellar mass for isolated and non-isolated systems . . 30

2.1 The Green Bank Telescope (GBT) ...... 37 2.2 Azimuthally averaged beam pattern of GBT...... 38 2.3 Observing script example...... 42 2.4 Software used during observations with the GBT ...... 43 2.5 Examples of radio frequency interference features in our spectra . . . 46 2.6 RFI removal process ...... 49 2.7 RFI removal histogram...... 51 2.8 Examples of ﬁnal spectra for detections and non-detections ...... 52

ix LIST OF FIGURES x

3.1 DECam image of ultra-faint dwarf galaxy candidates around NGC 3109 57 3.2 Spectra for LSB features around NGC 3109 with HI non-detections . 61

3.3 GBT HI spectrum for NGC 3109-Dw1 showing clear emission at vHelio ∼ 1060 kms−1 ...... 62 3.4 Example of our analysis method of HI emission detections ...... 64 3.5 HI-mass and V-band luminosity calculated as a function of distance for LSB features with HI non-detections ...... 72 3.6 Gas richness as a function of V-band luminosity for non-detections and dwarfs from simulations ...... 74

4.1 Distribution of the M101 satellite candidate sample...... 81 4.2 HI non-detections for M101 satellite candidates ...... 86 4.3 Positions of satellite candidates that overlap with the M101 HI disk . 90 4.4 HI detections for satellite candidates around M101...... 91 4.5 Relative velocities of M101 group members and dwM1 ...... 96 4.6 HI mass as a function of stellar mass for the M101 satellite population 100 4.7 Luminosity versus projected distance for M101, Milky Way, and Geha et al.(2017) satellite populations ...... 101

5.1 GBT HI spectra of Local Volume satellite non-detections ...... 110 5.2 GBT spectrum along the line of sight to NGC2683-DGSAT1 . . . . . 113 5.3 HI mass to luminosity ratio as a function of distance for Local Volume satellites...... 116 5.4 HI mass as a function of stellar mass for Local Volume satellites . . . 118 LIST OF FIGURES xi

6.1 Comparison of luminosity versus projected distance for our combined sample, Milky Way, and Geha et al.(2017) satellite populations . . . 121 6.2 Cumulative luminosity functions for M101, the Milky Way, M31, and NGC 2683 ...... 123 6.3 Residuals from the HI mass to stellar mass relation of Bradford et al. (2015) for our combined sample...... 127 6.4 Comparison of the luminosity-size relationship for our combined sample with Local Group dwarfs and Ultra-Diﬀuse galaxies ...... 130 List of Acronyms

ΛCDM Λ- Cold Dark Matter ALFALFA Arecibo Legacy Fast ALFA Survey BCD blue compact dwarf CFHT Canada-France-Hawaii Telescope dE dwarf elliptical galaxy DECam Dark Energy Camera DES Dark Energy Survey DGSAT Dwarf Galaxy Survey with Amateur Telescopes dIrr dwarf irregular galaxy dSph dwarf spheroidal galaxy DSS Dynamic Schedueling System ELVIS Exploring the Local Volume in Simulations FWHM full-width at half maximum GALEX Galaxy Evolution Explorer GBO Green Bank Observatory GBT Green Bank Telescope HIPASS HI Parkes All-Sky Survey HSC Hyper Suprime-Cam HST Hubble Space Telescope IF intermediate frequency JWST James Webb Space Telescope

xii Jy Jansky kpc kiloparsec LAB Leiden/Argentine/Bonn Survey LMC Large Magellanic Cloud LSB low surface brightness Mpc megaparsec NED NASA/Extragalactic Database NGC New Galactic Catalog NSA NASA-Sloan Atlas Pan-STARRS Panoramic Survey Telescope and Rapid Response System pc parsec RFI radio frequency interference RMS root mean square SAGA Satellites around Galactic Analogs SDSS Sloan Digital Sky Survey SMC Small Magellanic Cloud UDG ultra-diﬀuse dalaxy UV ultraviolet VEGAS Versatile GBT Astronomical Spectrometer VLT Very Large Telescope VST VLT Survey Telescope

xiii List of Symbols

c Speed of light D distance DEC Declination fgas gas fraction Hα hydrogen emission, Balmer series: n=3 to n=2 transition HI neutral atomic hydrogen emission

H2 molecular hydrogen

L Solar Luminosity

LV V-band Luminosity Λ Cosmological constant λ wavelength

M Solar mass

Mgas Total gas mass

MHI HI gas mass

Mstellar Stellar mass mV Apparent V-band magnitude

MV Absolute V-band magnitude

xiv MHI LV HI mass to V-band luminosity ratio

Mstellar LV Stellar mass to V-band luminosity ratio ν frequency RA right ascension

Reff Eﬀective radius rvir Virial Radius

S N signal-to-noise ratio

σ∆V RMS noise in spectra at velocity resolution

TA Antenna Temperature

Tsys System Temperature ∆V velocity resolution

Vsys systemic velocity

W50 velocity width at 50% peak ﬂux density

W50,c corrected velocity width at 50% peak ﬂux density z redshift

xv Chapter 1

Introduction

Galaxies are the building blocks of the large-scale structure in our Universe. They are composed of stars of various ages, gas, dust, and dark matter. They come in different shapes (or morphologies) and sizes. Galaxies themselves are thought to form in a hierarchical manner where smaller galaxies merge together to form larger ones. These small galaxies, known as dwarf galaxies, typically have a few thousand to a few billion stars while massive galaxies contain upwards of tens of billions. Dwarf galaxies are also more numerous than massive galaxies and are heavily dark matter dominated. The latter point makes them great arenas to further our understanding of dark matter. However, as they contain few stars, they are less luminous and more difficult to detect, especially at the lowest masses. These galaxies exist in different environments: some are found as satellites of more massive galaxies, while others can be found in the field. We show a compilation of images of dwarf galaxy satellites of the Milky Way from Bullock and Boylan-Kolchin(2017) that span several orders of luminosity and mass in Figure 1.1. Their masses are labeled in the bottom-left corners and a scale bar of 200

1 CHAPTER 1. INTRODUCTION 2

parsecs (1 pc = 3.086 1016m or 650 light-years) is in the bottom-right corners. × ∼ We immediately see that all of these dwarf galaxies have different properties. Some look more red, while others look blue. Some appear to be elongated, while others are round. Some are dense, while others look diffuse. The two brightest satellites of the Milky Way the Large and Small Magellanic Clouds (LMC and SMC) are extremely useful observational tools. Their proximity provides us with a unique opportunity to study a variety of details in an external galaxy. The faintest dwarf galaxies around the Milky Way have only recently been discovered with vastly improved observational technology and data mining techniques. Looking beyond the Milky Way, we find a similar population of dwarf galaxies around our nearest massive neighbour, Andromeda (or M31). Within our Local Group of galaxies defined by the Milky Way, M31, and their satellites, we find a diverse range of morphologies. Dwarf galaxies can be classified as dwarf irregulars (dIrrs), dwarf ellipticals (dEs), or dwarf spheroidals (dSphs). dIrrs are typically the most active of the dwarfs; with on-going star formation and young stellar populations they appear more blue in colour. They are also typically found at large distances from a massive host galaxy or are isolated in the field. Examples of dIrrs are the LMC, WLM, and Pegasus in Figure 1.1. On the other hand, dSphs or dEs consist of old stellar populations with little to no on-going star formation resulting in a redder colour. Most of the nearby satellite galaxies of the Milky Way are classified as dSphs, such as the dwarf galaxies other than the LMC and WLM in Figure 1.1. Beyond our Local Group, the rate of the discovery of satellite galaxies of other systems has increased significantly in the last few years. Detections are made in in projection around massive galaxies in the Local Volume as low surface brightness CHAPTER 1. INTRODUCTION 3

(LSB) features in optical images. Due to their larger distance and correspondingly low apparent luminosities, studying them in the same detail as nearby satellites is difficult. Nevertheless, the discovery of satellite populations around nearby systems increases our understanding of their relationship with their hosts and their implications for cosmological galaxy formation. Understanding the relationship between massive galaxies and their low-mass dwarf galaxy satellites is critical in understanding how galaxies form and evolve. Studying the interactions between the host and satellite can be done in different ways. On- going or past interactions can be observed at optical wavelengths by looking for tidal remnants such as coherent streams of stars. Other tracers of interactions can be used as well, such as the gas content and morphology of a dwarf galaxy. Gas is the fuel from which galaxies form stars and evolve. Here, we probe hydrogen in the ground state by way of the 21cm spin-flip transition in the atomic ground state (HI). Due to their low mass and, therefore, shallow potential wells, dwarf galaxies are susceptible to having their gas removed through various processes. As satellites orbit around their massive host galaxy, they can have their gas removed tidally and/or via ram pressure. Satellites can also lose their gas due to internal processes such as feedback from stars or global processes such as universal reionization. Around the Milky Way and M31, the dSphs that are nearby have been shown to be very gas-poor compared to more distant dIrrs. This suggests that there is an environmental dependence on the gas content of satellites. It is, therefore, critical to understand whether this is seen around other Milky Way-like systems. Also, it would prove useful to test whether this trend is seen at smaller scales. That is, would a dwarf galaxy with its own satellites be capable of stripping away their gas? CHAPTER 1. INTRODUCTION 4

In this thesis, we set out to further our understanding of the relationship between satellites and their hosts in the Local Volume (D . 20 Mpc). We aim to accomplish this by searching for HI along the lines of sight to satellite galaxy candidates iden- tiﬁed in deep optical surveys around galaxies in the Local Volume. By comparing our detections and non-detections to existing studies, we constrain the relationship between the satellite candidates and their putative hosts, as well as, that between the gas and the stars in the candidates themselves. The following sections in Chapter 1 present a review of the literature. Chapter 2 introduces our observation and data reduction procedures. Chapters 3-5 present the results for our HI follow-up searches around diﬀerent systems in the Local Volume. Chapter 6 provides a broad discussion of our results, and we conclude and discuss future work in Chapter 7. CHAPTER 1. INTRODUCTION 5

Figure 1.1: Examples of the variety of dwarf galaxies around the Milky Way. Figure from Bullock and Boylan-Kolchin(2017). 1.1. COSMOLOGICAL STRUCTURE FORMATION 6

1.1 Cosmological Structure Formation

The majority of matter in the Universe is in the form of dark matter and only a small fraction is in luminous baryonic matter (Planck Collaboration et al., 2016). It follows that galaxies, which are gravitationally bound systems, are composed of both dark and luminous matter. Galaxies form and live within dark matter halos that are much more massive than their baryonic components. Our current model of cosmology is referred to as ΛCDM, where Λ represents the cosmological constant, a contribution from dark energy, and dark matter is “cold” (velocities much lower than the speed of light) and collisionless. In the ΛCDM formalism, the structure that we see today formed through small perturbations of the hot plasma in the early Universe. This structure is thought to form in a hierarchical process where small halos of dark matter combine to form larger ones (Bullock and Boylan-Kolchin, 2017). Our best method of studying galaxy and structure formation is through simulations. Initial large-scale cosmological simulations, such as the Millennium simulation (Springel et al., 2005), studied the evolution of structure in the Universe only using dark matter particles. Improvements in computing power and eﬃciency allowed the inclusion of baryonic matter and related physics. With this, large cosmological hydrodynamical simulations were conducted. The Evolution and Assembly of GaLaxies and their Environments (EAGLE, Schaye et al., 2015) and the Illustris project (Vogels- berger et al., 2014) are two of the largest cosmological hydrodynamical simulations performed to date. Moving to slightly smaller scales, there are several simulations that look at the local Universe. These include the Constrained Local UniversE Simulations (CLUES, 1.1. COSMOLOGICAL STRUCTURE FORMATION 7

Carlesi et al., 2016), Exploring the Local Volume in Simulations (ELVIS, Garrison- Kimmel et al., 2014), A Project Of Simulating The Local Environment (APOSTLE, Sawala et al., 2016a), and the Aquarius project (Springel et al., 2008). Within these simulations, we can begin to see much more detail both in dark and luminous substructure. Taking a closer look at the Aquarius project, they aimed to simulate galaxies similar to the Milky Way (in dark matter halo mass) and its environment. Their sample consists of six cosmologically simulated Milky Way-like galaxies at multiple resolutions. Figure 1.2 shows the projected dark matter density for these galaxies. It can be seen that while each of these halos has a similar mass to that of the Milky Way, the distribution of subhalos varies. This clearly demonstrates cosmic variance: scatter in observed counts of small-scale structure due to large-scale density ﬂuctua- tions (Moster et al., 2011). These simulations (and those previously mentioned) are essential for testing our galaxy formation and evolution theories, particularly through comparisons with observations. 1.1. COSMOLOGICAL STRUCTURE FORMATION 8 . From Springel et( 2008 ). al. 430 kpc) − 320 ∼ vir r ( Figure 1.2: TheAquarius (dark projected matter dark only)velocity matter simulation dispersion suite. density appear more Brighter of yellow, regionsindicate while 6 indicate the those Milky with regions virial low of Way-size radii velocity higher galaxies, of dispersions density. these are at more halos Regions blue. the of The same high overlaid circles resolution, from the 1.2. COMPARING SATELLITES IN SIMULATIONS AND OBSERVATIONS 9

1.2 Comparing Satellites in Simulations and Obser-

vations

The Local Group of galaxies consists of a pair of massive spiral galaxies, the Milky

10−11 Way and M31. These galaxies have stellar masses on the order of Mstellar = 10 M and reside in even larger and more massive dark matter halos. Each of these galaxies has a system of less massive satellite galaxies which reside in correspondingly less massive dark matter subhalos. We can carry out detailed studies of these systems due to their proximity, more so for the Milky Way than M31. Our most complete sample of satellites is found around the Milky Way, which can be separated into groups depending on when they were discovered; also a tracer of ad- vances in our observational capabilities. The 11 Classical Satellites are in the brightest group ( 19 < MV,Classical < 9), discovered before the advent of the Sloan Digital Sky − − Survey (SDSS, Adelman-McCarthy et al., 2007). The SDSS roughly doubled the number of known satellites; discovering 12 more, much fainter ( 9 < MV,SDSS−DR5 1.5) − ≤ − satellites (Koposov et al., 2008 and references therein). With increased observational capabilities over the last decade ( 1.3 for more details), the number of satellites around § the Milky Way has increased from 23 to 50. The case is similar for the satellites ∼ around M31: more satellites have been discovered in the last two decades due to the increase in observational capabilities. However, it is still diﬃcult to probe to the same luminosities at the distance of M31, resulting in a lower number of observed satellites, 30. ∼ We discussed how cosmological simulations provide us with a new perspective on the formation and evolution of Milky Way-like galaxies. When we compare these 1.2. COMPARING SATELLITES IN SIMULATIONS AND OBSERVATIONS 10

high-resolution simulations with observations of galaxies in the Local Group, we find multiple inconsistencies. The first is the larger number of dark matter subhalos found in simulations compared to the number of satellites found in observations, commonly known as the “missing satellites problem” (Klypin et al., 1999; Moore et al., 1999). This implies that the majority of the Milky Way’s substructure exists as “dark halos” or subhalos of dark matter without any baryonic component (Simpson et al., 2017). The stellar mass functions of Milky Way-like halos from ELVIS (Garrison-Kimmel et al., 2014) are shown in Figure 1.3. For comparison, the authors include the observed stellar mass functions of the satellites around the Milky Way and M31 as cyan solid and dashed lines, respectively. Also included are the stellar mass functions found using the abundance matching relation from Behroozi et al.(2013), orange lines. While modifications to these abundance matching models (black lines) can reduce the differences between simulations and observations, the number of low-mass objects is still overestimated in the simulations. This problem can be addressed in multiple ways: further manipulation of our abundance matching models, including a baryonic component in simulations, pushing our observational capabilities further (see 1.3), § and searching for dark matter halos with little-to-no stellar components. This final point is especially difficult as it requires the ability to detect and distinguish the influence of low-mass halos on existing stellar structures such as streams (e.g. Yoon et al., 2011). 1.2. COMPARING SATELLITES IN SIMULATIONS AND OBSERVATIONS 11

Figure 1.3: A comparison of observed and expected cumulative stellar mass functions from Garrison-Kimmel et al.(2014). The Milky Way and M31 are shown as solid and dashed cyan lines, respectively. Stellar mass functions derived using abundance matching models are shown for the relation from Behroozi et al.(2013) (orange lines) and a modiﬁed relation based on this model (black lines). The thin lines for these models represent each of the halos from the ELVIS dark matter-only simulations and the thick lines show the mean. Hera, a Milky Way-like galaxy, is represented by the solid magenta line.

The “too big to fail” problem was discovered by Boylan-Kolchin et al.(2011). Using the subhalos from the Aquarius (Springel et al., 2008) and Via Lactea II (Diemand 1.2. COMPARING SATELLITES IN SIMULATIONS AND OBSERVATIONS 12

et al., 2007) simulations, they find that the central densities of the largest of these subhalos are too high to host any of the bright satellites of the Milky Way. However, these subhalos are simply too massive to lack the ability to form a galaxy: they are “too big to fail”. Solutions to this problem are not as straightforward as the missing satellites problem. Discovering massive dark subhalos is not a simple task. This problem was also discovered to exist for some of the M31 satellites (Tollerud et al., 2014). Using the ALFALFA survey, Papastergis et al.(2015) show that this problem also exists for field dwarf galaxies. A related expectation from the ΛCDM paradigm, is the steep rise of the density profiles in galaxies. Several dwarf galaxies, which have dark matter as their dominant component, have been shown to have much shallower, essentially constant, central density slopes (McGaugh et al., 2001; de Blok et al., 2008). Both of these problems have suggested solutions that either involve some form of feedback mechanism (e.g.Madau et al., 2014) or external interactions (e.g.Wetzel et al., 2016). “Planes of satellites” are another inconsistency between simulations and observations. The distribution of satellites around the Milky Way has a planar structure which was first suggested by Lynden-Bell(1976). It has since been shown that this is not expected from ΛCDM (Pawlowski et al., 2012). More recently, Ibata et al. (2013) discovered a similar planar distribution of satellites around M31. 1.3. THE LOW SURFACE BRIGHTNESS UNIVERSE 13

1.3 The Low Surface Brightness Universe

1.3.1 The Local Group

As mentioned in the previous section, our advancements in technology have brought forth a new era of observational capability. The SDSS’s systematic survey of large sky areas doubled the number of known Milky Way satellites. More recent large sky surveys have discovered many new faint satellites around the Milky Way. These discoveries are made by detecting stellar over-densities in images and using colour magnitude diagrams made up of individual stars to determine distances. The VLT Survey Telescope (VST, 2.6-meter) ATLAS set out to image approximately 4500 deg2 of the southern sky to a depth similar to that of SDSS (Shanks et al., 2015). The first satellite discovered in the survey, Crater (Belokurov et al., 2014), was separately discovered by Laevens et al.(2014) and was classified as a globular cluster, Laevens I. This was recently confirmed to be a globular cluster (Voggel et al., 2016). Torrealba et al.(2016a,b) recently discovered two more satellites of the Milky Way using the ATLAS survey data. The Dark Energy Survey (DES, 2005; 2016) is a wide-field (5000 deg2) photometric imaging campaign in the southern hemisphere. It uses the Dark Energy Camera (DECam) on the Blanco 4-meter telescope at the Cerro Tololo Inter-American Ob- servatory in Chile. In the first year of the survey, the collaboration presented the discovery of 9 new satellite candidates around the Milky Way (Bechtol et al., 2015; Luque et al., 2016) with 8 fainter satellites discovered in the second year (Drlica- Wagner et al., 2015). These satellites and other Local Group satellites are shown in Figure 1.4 to provide some sense of where they lie in the luminosity and size1. Using 1 MV is the V-band absolute magnitude which is the intrinsic brightness of an object, defined as 1.3. THE LOW SURFACE BRIGHTNESS UNIVERSE 14

the public data release from the first year of the survey, other groups performed an algorithmic search for stellar over-densities that could resemble faint dwarf galaxies (Koposov et al., 2015; Kim and Jerjen, 2015), revealing 10 more satellites around the Milky Way. DECam’s capability to detect faint objects is showcased by the one new satellite discovered in the Magellanic Satellite Survey (Drlica-Wagner et al., 2016) and another in the Survey of the Magellanic Stellar History (Martin et al., 2015). The Panoramic Survey Telescope and Rapid Response System (Pan-STARRS, Kaiser et al., 2002; Chambers et al., 2016) conducted a massive sky survey; imaging the whole sky visible from Hawaii using a 1.8-meter telescope and a camera with a large field-of-view 3.3 deg2. Specifically, the Pan-STARRS1 3π survey saw the discovery of four more faint satellite companions of the Milky Way (Laevens et al., 2015b,a). The Hyper Suprime-Cam (HSC)2 mounted on the 8.2-meter Subaru telescope was one of the most highly anticipated instrument-telescope combinations. The large collecting area of the primary mirror coupled with the wide field-of-view, 1.5 degree, 870 megapixel imager provides the opportunity to observe faint objects with great detail. Two new ultra-faint satellite candidates of the Milky Way have been discovered (Homma et al., 2016, 2017) using data from the first two years of the HSC Subaru Strategic Program (SSP). HSC-SSP is a 300-night imaging survey which began in 2014 with its Wide-field Layer covering 1400 deg2. One of the satellites discovered is

potentially the faintest satellite known to date: Virgo I, MV 0.33. ∼ Similarly the last decade has seen an increase in the number of satellites discovered around our nearest massive neighbour, M31. Initially, smaller scale surveys were the apparent magnitude an object would have if it were placed at a distance of 10 parsecs. 2http://dx.doi.org/10.1117/12. 926844. 1.3. THE LOW SURFACE BRIGHTNESS UNIVERSE 15

conducted around M31 using wide-ﬁeld cameras, Megaprime/MegaCam and Wide Field Camera, mounted respectively on the 3.6-meter Canada-France-Hawaii Tele- scope (CFHT) and the 2.5-meter Isaac Newton Telescope (INT) (Martin et al., 2006; Ibata et al., 2007; Irwin et al., 2008; McConnachie et al., 2008). A dedicated survey, Pan-Andromeda Archaeological Survey (PAndAS, McConnachie et al., 2009a), set out to add to the existing maps of M31 and imaged the entire stellar halo of M31 using the aforementioned MegaCam imager at CFHT. This survey resulted in the discovery of two new dwarfs around M31(Martin et al., 2009) with many more candidate systems discovered (Martin et al., 2013a). Outside of the ﬁeld-of-view of these surveys, the SDSS played a crucial role in the discovery of more satellites around M31 (Zucker et al., 2004, 2007; Slater et al., 2011; Bell et al., 2011). Pan-STARRS1 data had its share of discoveries for the Milky Way and follows suit with three more satellites around M31 (Martin et al., 2013c,b). In total, these surveys and data searches resulted in the discovery of > 15 new satellites and satellite candidates around M31.

1.3.2 Dwarfs in Nearby Systems

The Local Group is our best laboratory in which to study galaxies in the magnitude

6 and mass regime MV 5 and Mstellar 10 M . However, we need to expand our & − . search to other Milky Way-like systems in order to properly account for diﬀerences in population due to cosmic variance (see 1.1). This will further our understanding § of whether or not what we observe locally is seen in other systems. As the search for satellites is pushed to fainter magnitudes in the Local Group, we have also initiated the push to search for satellite galaxies in other Local Volume systems by resolving individual stars. 1.3. THE LOW SURFACE BRIGHTNESS UNIVERSE 16

Figure 1.4: Absolute V-band magnitude versus half-light radius for satellite galaxies and globular clusters. Local Group satellite galaxies and globular clusters are represented by blue squares and black crosses, respectively. Candidates discovered by DES are separated by the survey year: Year 1 candidates, red circles, and Year 2 candidates, red triangles. Green diamonds indicate candidates discovered by other groups. Black x’s represent halo star clusters and ambiguously classiﬁed objects. Dashed lines are of constant surface brightness. Figure from Drlica-Wagner et al.(2015). 1.3. THE LOW SURFACE BRIGHTNESS UNIVERSE 17

The Panoramic Imaging Survey of Centaurus and Sculptor (PISCeS) was conducted using the 6.5-meter Magellan telescope along with the Megacam imager. The goal of this survey was to provide a stellar map similar to PAndAS for two of the nearest massive galaxies outside of the Local Group, NGC 253 (D 3.2 Mpc, Radburn- ∼ Smith et al., 2011) and NGC 5128/Centaurus A (D 3.8 Mpc, Harris et al., 2010). ∼ Two satellites have been discovered around NGC 253: Scl-MM-Dw1(MV = 10.1, − Sand et al., 2014) and Scl-MM-Dw2 (MV = 12.1, Toloba et al., 2016). Thirteen − other satellites were found around NGC 5128 ( 13.0 < MV < 7.2, Crnojević et al., − − 2014, 2016). One of these satellites, CenA-MM-Dw3, is particularly interesting as it was found to be tidally disrupting with an extended stellar stream, similar to that of the disrupting Milky way satellite Sagittarius. The M81 group is the nearest analog to the Local Group; a central concentration of galaxies are around M81 with a smaller concentration around NGC 2403. CFHT/MegaCam imaging survey of a 65 deg2 region around the group revealed 22 new faint group member candidates, 14 of which have been conﬁrmed (Chiboucas et al., 2009, 2013). With 22 previously known members, the group has nearly doubled in size. The Magellanic Analog Dwarf Companions and Stellar Halos (MAD- CASH) survey set out to map galaxies in the Local Volume with masses similar to the

LMC (0.33 MLMC < Mstellar < 3 MLMC ). Carlin et al.(2016) present the ﬁrst results from the survey. With observations taken with HSC around NGC 2403, a low-mass spiral galaxy (D 3.0 Mpc, Radburn-Smith et al., 2011), they discovered a faint ∼ (MV = 7.7) dwarf galaxy candidate, MADCASH J072438+652501-dw. − Another system of interest is the NGC 3109 dwarf association (Tully et al., 2006). NGC 3109, the primary galaxy in this system of dwarfs, is an irregular dwarf galaxy 1.3. THE LOW SURFACE BRIGHTNESS UNIVERSE 18

just outside the Local Group with D 1.28 Mpc (McConnachie, 2012). Searching for ∼ satellites of a dwarf can provide crucial information for understanding how a host can inﬂuence its satellites (Dooley et al., 2017). Sand et al.(2015) report the discovery of

Antlia B, a new faint (MV = 9.7) dwarf in the association. Antlia and Antlia B are − suggested to be satellites of NGC 3109 due to their close 3D distances. What makes this system particularly interesting is the gas-richness of these two satellites. This is in contrast to the trends seen in the Milky Way where satellites within the halo viral radius are gas-poor (see 1.4 for details). §

1.3.3 Low Surface Brightness Galaxies

As we expand our search for satellites around hosts at greater distances (D & 4 Mpc), our ground-based imaging becomes less effective at resolving individual stars. These resolved stellar populations are important in determining distances to satellite candidates. Using space-based imaging (i.e. Hubble Space Telescope, HST) is required at these larger distances to resolve individual stars. However, ground-based surveys can discover low-surface brightness galaxies (LSBs) to much larger distances as the integrated surface brightness does not decrease with distance. Several groups have set out to search for LSBs around galaxies in the Local Volume. While the search and study of LSBs is not a new field (Bothun et al., 1987; Impey and Bothun, 1989; Davies et al., 1989; Irwin et al., 1990; Thuan et al., 1991; Bothun et al., 1991; Roukema and Peterson, 1995; Impey et al., 1996; O’Neil et al., 1997a,b, 1999; Matthews et al., 1999b,a), we will cover some of the recent findings here. Small aperture telescopes with a wide field-of-view are well-suited to use in searching for LSBs. Several groups have taken advantage of such telescopes to search around 1.3. THE LOW SURFACE BRIGHTNESS UNIVERSE 19

galaxies in the Local Volume. Tief Belichtete Galaxien (TBG, translates to “very long exposed galaxies”) is a project that was initiated by a German amateur group of astronomers. Later, they began a collaboration with professional astronomers and expanded their goals which saw the inclusion of more amateur astrophotographers. The group makes use of small telescopes (0.1 to 1.1-meter) coupled with high-end CCD cameras. The group recently presented their discovery of 3 LSBs around edge- on spiral galaxy NGC 4631 (Karachentsev et al., 2014), as well as, an LSB search around 13 massive spiral galaxies in the Local Volume (Karachentsev et al., 2015), discovering 27 LSBs around 9 of these hosts. Martínez-Delgado et al.(2008, 2010, 2012) were one of the first to exploit small- aperture telescopes to search for tidal interaction remnants or stellar streams around galaxies in the Local Volume. This project was eventually expanded into the Dwarf Galaxy Survey with Amateur Telescopes (DGSAT, Javanmardi et al., 2016) which used multiple privately-owned observatories around the world, equipped with 0.1- to 0.8-meter telescopes and 16-megapixel CCDs. They searched for dwarf galaxy companions around 6 nearby Milky Way-like galaxies (NGC 2683, NGC 3628, NGC 4594, NGC 4631, M101/NGC 5457, and NGC 7814). In their search of the fields around these galaxies, they discovered 11 previously unknown LSBs. Abraham and van Dokkum(2014) developed the Dragonfly Telephoto Array, a robotic refracting telescope consisting of eight Canon telephoto lenses coupled to science-grade CCDs with a 2.3 1.9 deg field-of-view. This setup is equivalent to × a 0.4-meter optical system with a focal ratio f/1.0. Their science goal is to search for very low surface brightness objects around the Local Volume. Their first target was M101 (the Pinwheel Galaxy), a face-on Milky Way-like spiral. van Dokkum 1.3. THE LOW SURFACE BRIGHTNESS UNIVERSE 20

Figure 1.5: LSBs around M101 discovered by Merritt et al.(2014). Insets show a zoomed cutout of the 7 detected LSBs. Figure from Merritt et al.(2014). et al.(2014) studied the outer regions of M101 and found little evidence for a stellar halo, usually indicative of interactions with satellite galaxies (e.g.McConnachie et al., 2009b). Merritt et al.(2014) discovered seven LSBs found in projection around M101 using the same images taken to study its stellar halo, as shown in Figure 1.5. In follow- up imaging with HST, three of these LSBs were determined to be at the distance of M101 (Danieli et al., 2017) while the remaining four were suggested to be members of a background group (Merritt et al., 2016). 1.3. THE LOW SURFACE BRIGHTNESS UNIVERSE 21

Searches for these features are not limited to new observations taken with small- aperture telescopes. One can use existing survey data to reach the same depths by stacking images together. Bennet et al.(2017) use this method along with a new dwarf detection algorithm in their search of a 9 deg2 region around M101. Using images from the CFHT Legacy Survey they recover the LSBs identified in previous observations (Merritt et al., 2014; Karachentsev et al., 2015; Javanmardi et al., 2016) and discover 38 new LSB dwarf candidates. Müller et al.(2017) follow a similar strategy, using data from SDSS and their own filtering algorithms to search for new LSB dwarfs in the “M101 group complex” (consisting of M101, M51, M63). They discover 15 new dwarf galaxy candidates in a 300 deg2 region. The galaxies they find are much brighter than those previously discovered but are located at larger projected distances from their potential hosts. Section 1.3.1 highlighted the capability of HSC to detect faint objects with the discovery of two new Milky Way companions. Greco et al.(2017) presented the discovery of over 750 low surface brightness objects discovered in 200 deg2 of the HSC-SSP. These galaxies show a variety of morphologies and can be separated into distinct populations according to optical properties such as colour. Some of these objects have been found to be fairly extreme in their properties with

large eﬀective radii (reff & 1.5 kpc) coupled with low surface brightnesses (µg,0 & 24 mag arcsec−2). These objects have luminosities of dwarf galaxies but sizes that exceed those of Milky Way-mass systems. A large number of these objects, dubbed as Ultra-Diﬀuse Galaxies (UDGs), were discovered by van Dokkum et al.(2015a) in the Coma cluster. Due to their elongated morphology, they were initially thought to exist only in high density environments where small galaxies can be tidally disturbed. 1.4. HI IN LOW-MASS GALAXIES 22

UDGs have been found in a variety of environments and the population as a whole has been shown to be quite diverse (Koda et al., 2015; Mihos et al., 2015; Martínez- Delgado et al., 2016; Merritt et al., 2016; Yagi et al., 2016; Trujillo et al., 2017; Leisman et al., 2017; Román and Trujillo, 2017; Shi et al., 2017). It has been shown that the abundance of UDGs scales with the size of the host system; more UDGs are expected in clusters versus groups (van der Burg et al., 2017). The formation mechanisms for UDGs are still uncertain; they may lie at one extreme of the known galaxy distribution (Amorisco and Loeb, 2016) or may form naturally through outflows (Di Cintio et al., 2017). Objects in this surface brightness regime at large distances are too faint to detect resolved stars from ground-based imaging. As a result, obtaining distances to these objects is difficult. Optical follow-up using space-based imaging to obtain distances can be expensive (Bird et al., 2010). However, obtaining low-resolution optical (van Dokkum et al., 2015b) or high-resolution radio (this work) spectroscopic measurements can be more efficient in this regard.

1.4 HI in Low-Mass Galaxies

1.4.1 Neutral Hydrogen

The Universe consists of a small fraction of baryonic matter, most of which is hydrogen. Molecular hydrogen, H2, is difficult to detect as it does not have a permanent dipole moment. This is not the case for neutral hydrogen (HI, pronounced “H-One”) in which the electron-proton pair spins can flip from parallel to anti-parallel. As a result of this spin-flip, a photon is emitted with a frequency of νHI 1.420406 GHz or ≈ 1.4. HI IN LOW-MASS GALAXIES 23

λHI = 21 cm. This frequency/wavelength falls in the radio regime of the electromag- netic spectrum which is unaﬀected by dust in galaxies unlike in the optical/infrared regimes. HI is a spectral line, such that if an object is moving away from or toward us then we can measure the redshift or velocity of an object through the Doppler shift,

λ λ0 z = − , (1.1) λ0

λ λ0 vobject = c − , (1.2) λ0

where z is the redshift, λ is the observed wavelength, λ0 is the rest frame wavelength, vobject is the velocity of the object, and c is the speed of light. The ubiquity of HI throughout the Universe provides us with an excellent tracer of a galaxy’s ability to form stars. It is typically found in spiral galaxies, as well as in star-forming dwarf irregulars and some ellipticals. HI is a major component of the interstellar medium of galaxies, usually found in the form of a disk that is co-planar with the stellar disk. The HI distribution in galaxies extends well beyond the stellar components (Broeils and Rhee, 1997). Consequently, this provides us with the ability to probe the dark matter component in the outer regions of galaxies by measuring the velocity of gas as a function of radius. The HI line is doppler-broadened and the width of the line can be used to estimate the dynamical mass (baryons and dark matter) of the object. Most importantly, we can determine the HI mass of a galaxy using the measured ﬂux (Haynes and Giovanelli, 1984)

5 2 MHI (M ) = 2.356 10 (D) Sdv, (1.3) × ˆ 1.4. HI IN LOW-MASS GALAXIES 24

where D is the distance to the object in Mpc and Sdv is the ﬂux density, the ´ integrated area of the emission line, in (Jy km s−1)3.

1.4.2 HI in the Local Group

The Local Group is our best opportunity to study the HI content of low-mass galaxies and how they are aﬀected by a massive host. As a reminder, dwarf galaxies can be separated into two primary groups: dSph and dIrrs. These two groups have diﬀerent physical properties where the former are typically composed of older stellar populations with little to no HI, while the latter are actively star-forming and have sizeable HI reservoirs (e.g. McConnachie, 2012). Grcevich and Putman(2009) studied the HI content of dwarf galaxies in the Lo- cal Group using survey data from the HI Parkes All-Sky Survey (HIPASS) and the Leiden/Argentine/Bonn Survey (LAB). They found that most of the satellites with distances beyond 270 kpc of the Milky Way or M31 were detected in HI while the majority of those within this distance were not. This suggests a strong environmental (i.e. distance) dependence on the HI content of satellites around massive hosts. Spekkens et al.(2014) expanded on this study, taking a detailed look at Milky Way dSphs. They obtained more stringent limits on the HI masses of these galaxies by using higher sensitivity data from the Arecibo Legacy Fast ALFA Survey (ALFALFA), the Galactic All-Sky Survey, and obtained deeper data for some galaxies using the Robert C. Byrd Green Bank Telescope (GBT). Figure 1.6 shows their main result where the ratio of HI mass to V-band luminosity (panel a) and HI mass to dynamical mass (panel b) are plotted as a function of galactocentric radius. In both cases, we

3A Jansky or Jy is a measure of ﬂux density; 1 Jy = 10−26 W m−2 Hz−1 1.4. HI IN LOW-MASS GALAXIES 25

can see that upper limits on the gas content of the dSphs which are within the virial radius of the Milky Way (dotted line, 300 kpc) implying a negligible gas reservoir in these satellites, which strengthens the case for an environmental dependence. These studies (and many others) point towards ram-pressure stripping as a key mechanism for the removal of gas from the satellites around the Milky Way and M31. In this mechanism, as a satellite moves through its orbit and begins its passage through the hot gas corona of its massive host, its gas reservoir is stripped away by ram pressure. Gatto et al.(2013) provide a criterion for stripping to occur, based on the work by Gunn and Gott(1972),

2 2 ρcor vorbital & ρgasσ , (1.4)

where ρcor is the density of the massive galaxy’s corona, vorbital is the orbital velocity of the dwarf galaxy, ρgas is the density of the gas in the dwarf, and σ is the velocity dispersion of the dwarf. This final term reflects the mass of the object. Therefore, low-mass objects would be more susceptible to this mechanism of gas removal. As a satellite moves closer towards its host, it will also begin to feel a stronger gravitational influence. Tidal mechanisms have also been proposed as a process for gas removal in satellites. However, tidal forces and ram pressure have been shown to be more effective at stripping when working together than individually (Mayer et al., 2006). Fillingham et al.(2016) construct an analytical framework to test how effective stripping is in a Milky Way-like environment. Using HI density profiles of nearby galaxies along with orbital conditions from the ELVIS simulations, their results show that ram pressure and turbulent viscous stripping are effective

9 at lower stellar masses (. 10 M ). Simpson et al.(2017) used the Auriga suite of 1.4. HI IN LOW-MASS GALAXIES 26

zoom-in cosmological simulations to study the diﬀerent mechanisms that could quench satellites around 30 Milky Way-like hosts. They, too, ﬁnd that ram-pressure stripping appears to be the most impactful quenching mechanism. Other mechanisms that can remove gas from a system include stellar feedback (i.e. stellar winds, supernovae) or reionization (i.e. photoevaporation of gas). The latter is critical in low-mass halos (Sawala et al., 2016b). All of these mechanisms play some role in the removal of gas throughout the lifetime of low-mass galaxies. To properly understand the trends seen in the Local Group, we need to study the gas content in low-mass galaxies in other systems. 1.4. HI IN LOW-MASS GALAXIES 27

Figure 1.6: Panel (a) shows the ratio of HI mass to V-band luminosity for dwarf spheroidals (red circles with downward arrows) and Local Volume dwarfs (blue triangles, circles with downward arrows) as a function of galactocentric distance. Panel (b) shows the HI mass to dynamical mass ratio as a function of galactocentric distance for dwarf spheroidals (red circles with downward arrows) and Local Volume dwarfs. Blue circles and green squares represent dynamical masses estimated using stars and gas, respectively. The dotted vertical line in both panels represents the virial radius of the Milky Way, Rvir = 300 kpc. From Spekkens et al.(2014), ﬁgure 2.

1.4.3 Gas-Rich Field Galaxies

7 The HI properties of ﬁeld galaxies with stellar masses Mstellar > 10 M are well- known. Using the ALFALFA 40% catalog (Haynes et al., 2011), Huang et al.(2012) 1.4. HI IN LOW-MASS GALAXIES 28

select galaxies with low HI masses (log [MHI ] < 7.7 M ) and combine them with photometric and spectroscopic data from the SDSS as well as UV imaging from GALEX (the Galaxy Evolution EXplorer). All of these galaxies will be gas-rich by definition as they would not have been detected otherwise. The SDSS and GALEX data provide corresponding details about the stellar components of the HI-selected galaxies and information about their star formation. The stellar mass estimates of these galaxies are in the range of 5.31 M < log (Mstellar) < 9.54 M . Their sample of low HI- mass (and correspondingly low stellar mass) galaxies have comparable gas fractions MHI fgas = Mstellar to those with higher stellar masses. They also find that the star formation rates scale more drastically (i.e. steeper slope) with HI mass than stellar mass. A detailed analysis of whether these galaxies are in an environment that can affect their HI content is not performed. Although, they do separate galaxies into Virgo cluster members and non-members, where the former would be expected to be devoid of HI as the intracluster medium is conducive to stripping. Geha et al.(2012) take a different approach to understanding environmental trends in low-mass galaxies. The main goal of their study was slightly different: to look for trends in the quenching of these galaxies as opposed to their gas content. These two properties are correlated, since galaxies that are star-forming typically have detectable gas reservoirs. The authors use the NASA-Sloan Atlas (NSA, Blanton et al., 2011) which is a reprocessed version of data from the SDSS catalog combined with GALEX UV images, optimized to search for faint, extended objects nearby. They selected

7 9 galaxies with stellar masses in the range 10 M < Mstellar < 10 M . In order to estimate the environment of each galaxy in the sample, they searched several catalogs

10 for luminous ( Mstellar 2.5 10 M ) neighbours. They deﬁne galaxies as isolated if ∼ × 1.4. HI IN LOW-MASS GALAXIES 29

they are 1.5 Mpc from their nearest massive neighbour and determine whether or not they are quenched (i.e. galaxies with little to no star formation) using Hα emission and a criterion based on the 4000A˚ break. In the analysis of their sample they ﬁnd that quenched galaxies are preferentially located near a massive host. They also show that quenched galaxies in the ﬁeld are virtually non-existent below a stellar mass of

9 Mstellar < 10 M . Bradford et al.(2015) study the baryon content of isolated low-mass galaxies and compare their ﬁndings with galaxies both of similar and greater mass in isolated and denser environments. To do this, they follow a similar route to Geha et al.(2012) using the NSA catalog to deﬁne low-mass galaxy samples. They combine this with HI data from the ALFALFA 40% catalog and new pointed HI observations with the Arecibo Telescope and the GBT. Their samples of isolated and non-isolated galaxies have relationships between stellar mass and atomic gas mass (Mgas = 1.4MHI , accounting for helium abundance). Both samples show a break in this relationship at a stellar

8.6 mass of Mstellar 10 M . We show Figure 5 from their work in Figure 1.7. We can ∼ see that, by and large, both isolated and non-isolated galaxies have similar stellar and gas masses. Galaxies with HI reservoirs more than 1 dex below the fits were classified as gas-depleted. Interestingly, this small fraction of galaxies are only found in the non-isolated sample. A recent study by Geha et al.(2017) has the potential to provide us with a vast sample of satellites around Milky Way-like hosts. They presented their first results from the Satellites Around Galactic Analogs (SAGA) survey which aims to study the distribution of satellites around 100 Milky Way-like systems down to luminosities of

Leo I (Mr < 12.7 or MV < 12, McConnachie, 2012). They conduct a search in the − − 1.4. HI IN LOW-MASS GALAXIES 30

Figure 1.7: Logarithm of atomic gas mass as a function of logarithm of stellar mass for the isolated (red squares, upper limits are thick, peach downward arrows) and non- isolated (black circles, upper limits are thin, grey downward arrows) galaxies from Bradford et al.(2015). The median error bar is shown as an open point in the top- 8.6 left. A broken-power law is fit to the two samples with the break at M∗ = 10 M . The parameters of the fit are shown at the bottom of the figure. The blue circles are galaxies classified as gas-depleted because they lie more than 1 dex below the fits (dashed blue line). From Bradford et al.(2015), Figure 5. 1.5. THIS WORK 31

Hyper-LEDA2 database (Makarov et al., 2014), supplemented with data from other catalogs as necessary, for Milky Way analogs restricted to distances of 20 to 40 Mpc. They find a total of 71 hosts that satisfy all of their criteria and plan to use other imaging surveys to reach their target of 100 hosts. To search for satellites around these hosts, SDSS Data Release 12 photometry is used with the NSA catalog. Hα spectroscopic observations are used to confirm an association between their satellite candidates and their putative hosts. Around 8 hosts, they find 14 new satellites that fall within their completeness limits and 11 that do not. Their complete sample has 27 satellites, 14 new and 13 previously known, around these 8 hosts. The most intriguing result from this survey is that the vast majority (26/27) of these satellites are star-forming. This is in contrast to satellite populations of the Milky Way and M31 (Grcevich and Putman, 2009; Spekkens et al., 2014), although, this may stem from the different characteristic masses of the sample galaxies.

1.5 This Work

As discussed in the previous section, much work has been done on understanding the relationship between low-mass galaxies and massive hosts in terms of their gas content and star formation. It should be made clear that these studies do not suﬃciently probe

7 below stellar masses of Mstellar . 10 M . It is in this mass regime that the majority of satellites reside in the Local Group (e.g. Figure 1.3). While there is beneﬁt in studying more massive satellites, we should take advantage of the recent discoveries of low-surface brightness (presumably low-mass) galaxies in the Local Volume. In this thesis, we aim to study the HI content of 48 ultra-faint or LSB galaxies discovered around galaxies in the Local Volume: NGC 3109, M101/NGC 5457, NGC 1.5. THIS WORK 32

672, NGC 891, NGC 1156, NGC 2683, NGC 3344, NGC 4594, NGC 4625, and NGC 7814. To accomplish this, we obtain HI spectra using the GBT of these optically identiﬁed galaxies. The majority fall within the projected virial radius of their putative hosts. With a detection of HI emission, we obtain information about the HI mass of target, systemic velocity, and its dynamical mass. For HI non-detections, we place stringent upper limits on the HI mass along the line of sight to the target. We can then conduct multiple comparisons to the Local Group, particularly the distance dependence on the HI content of satellites (Figure 1.6), as well as to other studies of the HI content in low-mass galaxies in isolation and around massive hosts (Figure 1.7). This thesis is organized into 7 chapters. Chapter 2 introduces our HI observations using the GBT and the data reduction process. Chapter 3 presents the results from our observations of 6 ultra-faint dwarf candidates around NGC 3109. In this chapter, we also describe the data analysis method that we use throughout the thesis. Chapter 4 presents the results from our observations of 27 satellite candidates discovered around M101. Chapter 5 presents the results from our observations of 15 satellite candidates around 8 Local Volume galaxies. Within Chapters 3-5 we provide brief discussions of the implications of the results of each study. In Chapter 6, we broadly discuss our results when the sample as a whole is considered, and their implications for the gas richness of ﬁeld and satellite galaxies. We provide our conclusions and future work in Chapter 7. Chapter 2

Observations and Data Reduction

The goal of this chapter is to explain our observation and data reduction process. We cover these topics in a chapter of their own as the observation method and data reduction techniques are common between the projects discussed in the later chapters. In Section 2.1, we describe our observations which include explanations of single-dish radio telescopes, the Green Bank Observatory, our instrumental setup, and the tools we use during the observations. In Section 2.2, we provide an overview of the data reduction process which consists of two major components: calibration and radio frequency interference (RFI) removal.

2.1 Observations

2.1.1 Single-Dish Radio Telescopes

Radio telescopes can be separated into two classes, single-dish telescopes and interferometers. The latter consist of an array of antennas whose signals are correlated. Each

33 2.1. OBSERVATIONS 34

has its advantages and disadvantages but for the purposes of the research presented in this thesis we have made use of a single dish because of its higher point-source sensitivity compared to interferometers. The basic components of a simple single-dish radio telescope are an antenna, a receiver, and a backend/detector. The antennas of most radio telescopes are paraboloids that reflect incoming radiation to a single focal point. Depending on the configuration of the telescope, this radiation is either reflected again into the receiver by a secondary reflecting surface (subreflector) at the focal point or directly focused into a receiver at that location. After being focused, the reflected radiation travels through a feed horn/waveguide where it is measured as an oscillating current by a dipole at the end of the feed horn. The incoming signal is extremely weak and needs to be amplified. The first amplification is through a low-noise amplifier whose purpose, as its name suggests, is to amplify the signal while adding as little noise as possible. The signal is then sent through a mixer. Here, the signal is combined with a sine wave that is generated by a local oscillator. The output of the mixer is the signal at an intermediate frequency (IF) that is more easily amplified and reduces transmission loss over large distances. Another advantage of using an IF is that, by converting signals at different frequencies to a common IF, the same transmission hardware can be used. The IF signal is amplified and then sent through to a backend. The choice of backend depends on the science goals. We use a spectrometer that divides the signal into many spectral channels over a given frequency width. The power in each channel is found by taking square of the input signal through a square- law detector. The final step is sending the squared signal through an integrator 2.1. OBSERVATIONS 35

to ﬁnd the time-averaged power in each channel. In modern spectrometers, this power spectrum is generated by taking the Fourier transform of the signals and then accumulating them over time. The power, P , received per unit frequency is typically presented as an antenna

temperature, TA:

P TA = , (2.1) k

where k is Boltzmann’s constant. Measuring a signal from a source of interest,

Tsource, is dependent on the noise ﬂuctuations within the system itself. The system

temperature, Tsys, has a number of contributions such as the Cosmic Microwave Background, other background astronomical sources, Earth’s atmosphere, radiation from the ground, and the receiver.

The sensitivity of a radio telescope is dependent on Tsys and is described by the radiometer equation

Tsys σT , (2.2) ≈ √∆νt

where ∆ν is the frequency bandwidth of the observation and t is the integration time. However, it is more useful to use the system equivalent ﬂux density, SEFD,

which is the ﬂux density of a source that would produce Tsys. SEFD is deﬁned as

2ηkTsys SEFD = , (2.3) Ae

where η is the eﬃciency (typically 1 for dishes with unblocked apertures), k is ∼ Boltzmann’s constant, and Ae is the eﬀective area of the telescope. With this one 2.1. OBSERVATIONS 36

could rewrite Eq.2.2 as SEFD σT , (2.4) ≈ √∆νt

which can be used to compare between telescopes. While the sensitivity of a telescope is one of multiple factors to consider, the power pattern is another. The power pattern is defined by the Fourier transform of the illumination of the antenna. For telescopes with feed/subreflector structures blocking the dish, the Fourier transform and, therefore, the power pattern will be affected. Telescopes with an unblocked aperture will have cleaner power patterns with lower sidelobes (Spekkens et al., 2013). This is where the GBT (Figure 2.1) can outperform many other single-dish radio telescopes.

2.1.2 The Green Bank Observatory

We perform our observations with the GBT at the Green Bank Observatory (GBO) located in Green Bank, West Virginia, USA. The GBT is the largest fully steerable telescope in the world with a 100 meter by 110 meter parabolic dish. Its parabolic shape results in an effective collecting area equivalent to a 100 meter circular dish. Its large aperture and offset receiver cabin results in the beam response pattern shown in Figure 2.2. The solid black line shows the beam model for the GBT and the red dashed line shows a Gaussian with a full-width half max (FWHM) of 9.12 arcmin. The main beam is where the antenna is most sensitive and its FWHM is typically simply called the “beam”. For the GBT, the beam extends down to -45 dB and ∼ out to a radius of r 0.25 deg from the pointing center. The first sidelobe of the ≈ GBT is has a peak response of -30 dB at 0.35 deg from the center of the beam, ∼ ∼ 2.1. OBSERVATIONS 37

Figure 2.1: The Robert C. Byrd Green Bank Telescope (GBT). A 100m by 110m parabolic dish with the oﬀset subreﬂector and receiver cabin in the top-middle of the image. Image courtesy of NRAO/AUI. 2.1. OBSERVATIONS 38

Figure 2.2: Azimuthally averaged beam model for the GBT at 1.4GHz (black line) from Spekkens et al.(2013). The GBT beam is well-modeled by a Gaussian (red dashed line) with a FWHM of 9.12 arcmin at I > 7dB. The ﬁrst sidelobe of the − beam is located at -30 dB which is 0.1% of the central intensity. ∼ 2.1. OBSERVATIONS 39

or 0.1 of the central intensity. This implies that emission beyond the main beam of the GBT needs to be at least 1000 times brighter than that at the point center to contribute equally to the measure signal. The GBT, therefore, excels at searching for faint sources in the vicinity of bright ones. Another important feature of the GBT is its location in the National Radio Quiet Zone, a 34,000 km2 area between Virginia and West Virginia that has strict regulations on radio frequency usage.

2.1.3 Instrument Setup

For our observations, we use the L-Band receiver in combination with the Versatile GBT Astronomical Spectrometer (VEGAS). The L-Band receiver is a cryogenically cooled, single-beam receiver that operates at frequencies ranging from 1.15GHz - 1.73GHz. The signal is then sent to one of eight spectrometers that comprise VEGAS: a highly flexible instrument with a variety of modes consisting of various combinations of spectrometer bandwidth and resolution. We use VEGAS in two modes: Mode 7 and Mode 15. Mode 7 has a bandwidth of 100 MHz ( 21, 000 km s−1) with 32768 ∼ spectrometer channels resulting in a frequency resolution of 3.1 kHz ( 0.6 km s−1). ∼ Mode 15 has the same number of channels as Mode 7, however, its bandwidth is 11.72 MHz ( 2500 km s−1) which results in a finer spectral resolution of 0.4 kHz ∼ ( 0.08 km s−1). We use Mode 15 in cases where the redshift (or distance) of the ∼ target is known, allowing us to appropriately center our bandpass. An advantage of the narrower bandwidth in Mode 15 is the reduced likelihood of radio frequency interference (RFI) contaminating the spectra (see 2.2). Mode 7 is typically used when § the redshift of the source is unknown and we need to cover as much redshift/velocity- space as possible. The difference in the raw VEGAS resolutions does not affect our 2.1. OBSERVATIONS 40

science results since the spectra are smoothed to much lower resolutions to search for detections (see 2.2). §

2.1.4 Observations with the GBT

We conduct our observations via a remote connection to computers at the GBO site. On-site training to observe remotely with the GBT was completed in May 2015. Our GBT observations are dynamically scheduled through the NRAO Dynamic Scheduling System (DSS). The DSS uses weather forecasts to maximize the efficiency of the observations that are scheduled. A new schedule is made each day for the following 24-48 hour period. Observing sessions during which data were collected for this thesis were 1 hour to 13 hours in duration. A total of 175 hours of observations ∼ were conducted by the author for the projects presented in this work. The author managed the scheduling, observing setup, and remote observations for all of them. We perform pointed, position-switched (ON-OFF) observations of our targets with the GBT. ON-OFF observations involve pointing the telescope on the target of interest (ON position) for a set amount of time. We then point the telescope to an emission-free, reference position the sky (OFF position) in the vicinity of the target for an equal amount of time. Since the system temperature and bandpass shape are stable throughout the ON and OFF observations, the difference between the two scans is a measure of the HI emission along the line of sight. During ON and OFF pointings, a noise diode with a known temperature contribution, TND, is used as a calibrator to determine Tsys during each measurement. This process is similar for both ON and OFF pointings and consists of two measurements, one with the noise diode on, NDON , and one with the noise diode off, NDOFF . Tsys can be found by 2.1. OBSERVATIONS 41

comparing the known noise diode temperature with the temperature measured as a

fraction of Tsys, NDOFF Tsys = TND, (2.5) NDON NDOFF − (e.g. O’Neil, 2002). The data collected at these two positions are then calibrated as described in the next section. Before each observing block, observing scripts, target catalogs, and configuration files need to be prepared. Our target catalogs contain information about our targets such as the target name, their sky position, and the frequency/velocity at which the spectrum will be centered. We use two types of configuration files. The first performs the standard AutoPeakFocus procedure and the second configures VEGAS. Both files contain similar information such as our choice of receiver, backend, the rest frequency of our observations, the length of a single integration, and other parameters. The AutoPeak procedure involves mapping a known radio source to calibrate the pointing of the antenna and to adjust the position of the receiver. After this, we execute our observing scripts, Figure 2.3. First, we load the catalog of targets (line 1). Next, we configure the telescope to the specified backend using the VEGAS configuration files (lines 3 and 4). Once the telescope and backend are set up, we center the telescope on the target and balance the IF system (lines 6 and 7). The Balance() command ensures that the devices in the IF system and the backend are operating within the linear response regime. Finally, the commands for the OnOff scans for observing the target are run (lines 9 to 14). When observing we use the Astronomer’s Integrated Desktop (Astrid; Figure 2.4, middle). In Astrid, we manage and execute our observing scripts in the Observation Management tab and monitor the telescope in the GbtStatus tab. Data that are 2.1. OBSERVATIONS 42

Figure 2.3: Example observing script for position-switched (ON-OFF) observations with VEGAS. being taken can also be displayed in Astrid for some instruments/receivers. Another set of useful tools during observations are found in the Control Library for Engineers and Operators, CLEO. Our primary tools in CLEO are T alk and Draw (Figure 2.4, top left), the VEGAS Status and Power monitors (Figure 2.4, bottom left and right, respectively). The former is a text-based communication interface between observers and telescope operators. With the status monitor we obtain information on whether VEGAS is in the correct mode, the correct banks are being used, etc. The power monitor provides an active power time series graph and histogram for both polarizations. This allows us to actively monitor the power that is being received to make sure there aren’t any anomalies. 2.1. OBSERVATIONS 43 Figure 2.4: Screenshot of theleft:VEGAS tools status used monitor. during a Middle: remote Astrid. observing session. Right: Top VEGAS left: Power Talk Monitor. & Draw window. Bottom 2.2. DATA REDUCTION 44

2.2 Data Reduction

We make use of the data reduction machines (standard desktops with a UNIX system) that are made available by the GBO. The data output is in the format of a Single Dish FITS (Flexible Image Transport System), SDFITS1, ﬁle. We inspect, reduce, and perform preliminary analysis on our data using GBTIDL2, an interactive package written in IDL primarily for spectral line observations at the GBT. The following will describe our data reduction process. We calibrate our data using the standard getps function in GBTIDL that is used for position-switched, total power observations. This function works at the integration level. It averages a single integration from the ON position with the calibrator on and one with the calibrator oﬀ to obtain the total power. This is repeated for the OFF position. The resulting total power data from the ON and OFF positions are then combined:

DataON DataOFF DataONOFF = − Tsys,OF F , (2.6) DataOFF

where DataON and DataOFF are the total power data from the ON and OFF positions and Tsys,OF F is the system temperature in the OFF position. The calibrated data are in units of antenna temperature. The user can specify more commonly used units, such as Jy, by using the dcsetunits procedure which applies the following scale factor h i τ0 exp sin(el) , (2.7) 2.85ηAηl

1https://ﬁts.gsfc.nasa.gov/registry/sdﬁts.html 2http://gbtidl.nrao.edu/ 2.2. DATA REDUCTION 45

where τ0 is the zenith opacity, el is the telescope elevation, ηA is the aperture efficiency, and ηl is the correction for spillover, ohmic loss, and blockage efficiency. τ0 and ηA are calculated using the observing frequency and are set to 0.008 and 0.710, respectively. ηl is set to 0.99 in GBTIDL. The factor of 2.85 in the denominator comes from the Boltzmann constant and the physical collecting area of the telescope. This process is done for all integrations in a given scan and polarization. All of the integrations in a given scan are then averaged and we are left with a calibrated scan in one polarization. From here, we can move on to address any contamination in our data. Any sort of RFI (car stereo, camera, cell phone, etc.) during an integration can show up in the data as anomalous features that span a few to several spectra channels. Broadband RFI can have both large and small amplitudes, spanning many spectral channels. Narrowband RFI spans one to a few channels and have small amplitudes at the integration level which allows them to be hidden by the noise of the spectrum. When the integrations are combined, the narrowband RFI features emerge from the noise and have larger amplitudes at the scan level. Examples of both types of RFI are shown in Figure 2.5. To remove broadband sources of RFI we flag the entire contaminated integration or scan to ensure it is not included in any further analysis. The most common flag used in our case is for the GPS L3 signal that comes from a communication satellite emitting at 1.381GHz (see Figure 2.5, panel a and b). This is a rather strong signal and spans many channels, and can affect a large portion of the spectrum. This signal is monitored by NRAO and, therefore, is well understood. In our experience, the signal can contaminate anywhere from 5 seconds to several minutes of data in 2.2. DATA REDUCTION 46

3.5 (a)

1.5

0.5 − (b) 0.75

0.75 Flux (Jy) −

1.36 1.38 1.40 1.42 1.44 (c) 0.1

0.1 −

1.36 1.38 1.40 1.42 1.44 Frequency (GHz)

Figure 2.5: Examples of RFI in our scans. Panel (a) shows the GPS L3 feature at 1.38 GHz clearly extending above the noise. Panel (b) shows examples of ∼ narrowband RFI signals inside rectangles. Panel (c) shows the spectrum from the middle panel smoothed with a Gaussian to 71kHz. When smoothed, we can see low ∼ amplitude, broad RFI sources inside the rectangle around 1.39 to 1.43GHz. In all ∼ panels, the Milky Way emission is seen at 1.42GHz. ∼ 2.2. DATA REDUCTION 47

a typical observing run. This feature only appears in observations conducted with VEGAS in Mode 7. To remove narrowband RFI features we replace the affected channels. Our RFI removal script was written by the author in IDL/GBTIDL and is shown visually in Figure 2.6. As mentioned above, larger narrowband signals are built up by smaller ones at the integration level. To replace these signals, we search through a given integration for values that exceed five times the median absolute deviation (MAD) and replace them with the local median value. Panel (a) in Figure 2.6 shows values above this threshold as blue squares. The MAD is defined as

MAD = median ( ai median(a) ) , (2.8) | − |

where a is our data and ai are individual entries. We use the MAD as opposed to other measures (e.g. standard deviation) as it is not as heavily weighted towards outliers in a data set. We have tested other multiples of the MAD in this process and we ﬁnd, on average, that multiples lower than 5 can remove too many genuine noise spikes in the spectrum, whereas multiples above 5 remove too few. We perform this process on each integration and then take the average of all integrations within a given scan (panel c in Figure 2.6). Next, we perform this same replacement process on the scan itself. Panel (c) in Figure 2.6 shows values above the threshold in the scan, similar to Panel (a). This shows that even after our removal of outliers in individual integrations, potential RFI signals are able to build up. The repetition of the removal process at the scan level is therefore necessary. Due to the data container structure of GBTIDL, we average 2 to 6 scans (1 to 3 in each polarization) at a time. These intermediately averaged scans are convolved with a Gaussian (i.e. smoothed) 2.2. DATA REDUCTION 48

to a lower spectral resolution (305kHz or 155km/s) to reveal large scale variations in the spectrum to which we ﬁt a baseline (10th-order polynomial). This baseline is then subtracted from the full resolution spectrum to ensure that any variations are removed. The intermediate scans are then averaged and we are left with the total raw spectrum, Panel (e). A raw spectrum is typically smoothed to a desired spectral resolution depending on the science requirements. Panel (f) shows the spectrum in panel (e) smoothed to 71.3 kHz (15 km s−1). As we go through the reduction process, we can see a decrease in the noise level of the spectrum as more data are co-added. The feature at 1.45 GHz is a cellular communication signal and is located at a ∼ frequency that does not aﬀect our science goals. 2.2. DATA REDUCTION 49

(a) 0.5

0.5 − 0.4 (b)

0.4 − 0.2 (c)

0.2 −

Flux (Jy) (d) 0.1

0.1 − 0.03 (e)

0.03 − 1.36 1.38 1.40 1.42 1.44 (f) 0.002

0.002 − 1.36 1.38 1.40 1.42 1.44 Frequency (GHz)

Figure 2.6: RFI removal process. Each panel shows a different stage in the process. (a): Spectrum generated for a single integration in one polarization. Blue squares indicate channels that exceed the 5-MAD threshold. (b): Same as (a) but with the affected channels replaced. (c): Spectrum generated for a single scan (this one specifically has 60 integrations) in one polarization. Blue squares indicate channels that exceed the 5-MAD threshold. (d): Same as (c) but with affected channels replaced. (e) Total raw spectrum of the source composed of all scans. (f): Spectrum from (e) smoothed to 71.3 kHz (15 km/s). The large spike at 1.42 GHz is HI ∼ emission from the Milky Way. The feature at 1.45 GHz is a cellular communication ∼ signal that is seen is most of our data. It is located at a frequency that does not affect our work. 2.2. DATA REDUCTION 50

We keep track of the channels that we have replaced during the RFI removal process. With this information we can then generate a histogram to see how many times we have had to replace a value in a given channel. An example of this is shown in Figure 2.7 at both the integration and scan levels. This allows us to check for overlap between channels that contain a potential signal from a target of interest and channels with signiﬁcant RFI removal. In the following chapters, we indicate the frequencies at which we have removed > 15% of integrations. After completing the data reduction process for a source, we are left with ﬁnal spectra that we use in our analysis. Figure 2.8 shows examples of a spectrum with a detection of HI from NGC 3109-Dw1 and a non-detection in the spectrum of NGC 3109-Dw3. A consequence of observations using the ON-OFF technique is that the emission from the Milky Way (which covers the whole sky) is captured in both scans. This results in the atypical feature seen in Figure 2.8 (b) at V 0 km s−1. We ∼ will describe our analysis procedure for spectra with detections of HI emission and non-detections in the next chapter. 2.2. DATA REDUCTION 51

Figure 2.7: Histogram of frequencies at which RFI was removed in our data reduction process for M101-Dw33 normalized to the total number of integrations/scans. Top: Fraction of integrations with RFI removed at a given spectral channel. Bottom: Same as top but for the fraction of scans. 2.3. CONCLUSIONS 52

Dw1 0.005 (a)

0.003

0.001 Flux (Jy)

0.001 − 500 1000 1500 2000

Dw3 (b) 0.005

Flux (Jy) 0.005 −

1000 0 1000 2000 3000 4000 5000 − VHelio(km/s)

Figure 2.8: Final spectrum of a detection, (a), and a non-detection, (b), that we use in our analysis of the targets around NGC 3109. In all panels, the Milky Way emission is seen at around 0 km s−1.

2.3 Conclusions

In this section, we have described the basic components of single dish radio telescopes and their functions, and how the GBT is the best choice for our science goals due to its well-understood, clean beam pattern and its radio-quiet environment. We 2.3. CONCLUSIONS 53

provided an overview of our observational process and the tools that we use to conduct our observations. Finally, we presented our data reduction process which includes temperature calibration and RFI removal. We use the data collection and analysis procedures described in this chapter in the three that follow. Chapter 3

Satellite candidates around NGC 3109

In this chapter, we describe the system centered around the dwarf irregular galaxy NGC 3109. In Section 3.1, we will review the properties of the system’s host, its existing satellite population, and the discovery of new satellite candidates. In Section 3.2, we will describe our observations of their gas content and our analysis. We move on to interpreting our results in Section 3.3 and we conclude in Section 3.4.

3.1 The NGC 3109 System and New Satellite Can-

didates

As discussed in 1.1, simulations provides us with a unique opportunity to understand § our relationship between massive galaxies and their low-mass satellite companions (e.g. Simpson et al., 2017). They show that there are numerous subhalos of dark

54 3.1. THE NGC 3109 SYSTEM AND NEW SATELLITE CANDIDATES 55

matter that live within their host halo. For the most part, our observations of satellites in the Local Group agree with this (Kim and Jerjen, 2015; Koposov et al., 2015; Bechtol et al., 2015; Drlica-Wagner et al., 2015; Homma et al., 2016). However, a prediction from simulations that has not been thoroughly studied through observations is satellites of dwarf galaxies (or satellites of satellites, Dooley et al., 2017). These subhalos are thought to merge and form more massive halos in a hierarchical process. Studying these objects is crucial in properly understanding cosmological galaxy formation. The low-mass galaxies residing in these subhalos are much fainter than normal satellites, resulting in their lack of detection. Recent improvements in our observational instrumentation has allowed for some of these objects to be revealed in the Local Group around NGC 3109. NGC 3109 is a dwarf irregular galaxy (dIrr) located at the edge of the Local

Group, D = 1.28 Mpc, with MV = 14.9 (McConnachie, 2012). Its distance from the − center of the Local Group places it in a lower density environment relative to other dwarf galaxies. NGC 3109 has been studied extensively at both the optical and radio

8 wavelengths. It has an estimated stellar mass of 0.76 2.1 10 M (McConnachie, ∼ − × 8 2012; Jobin and Carignan, 1990) and an HI mass of 3.8 4.9 10 M (Barnes and ∼ − × de Blok, 2001; Koribalski et al., 2004a; Jobin and Carignan, 1990). NGC 3109’s higher gas mass relative to its stellar mass is consistent with other dwarf galaxies that are located at large distances from more massive galaxies (Grcevich and Putman, 2009; Spekkens et al., 2014; McConnachie, 2012). The NGC 3109 system is classiﬁed as a dwarf association (Tully et al., 2006): it is a collection of galaxies that lie at similar distances and in similar regions of the sky. The other dwarf galaxies that belong to the association are Sextans A and B 3.1. THE NGC 3109 SYSTEM AND NEW SATELLITE CANDIDATES 56

(Zwicky, 1942; Holmberg, 1958), Antlia (Whiting et al., 1997; Tully et al., 2006), Leo P (Giovanelli et al., 2013; McQuinn et al., 2015), and Antlia B (Sand et al., 2015). The latter two are more recently discovered dwarfs found in HI and optical observing campaigns. Leo P was detected by ALFALFA (Giovanelli et al., 2005) and confirmed to be a member of the dwarf association in optical imaging (Rhode et al., 2013; McQuinn et al., 2013). Antlia B was found in a broadband optical imaging campaign around NGC 3109 with DECam (Sand et al., 2015). Antlia and Antlia B project within the virial radius of NGC 3109 and would be expected to be gas-poor, assuming the trend seen in Milky Way satellites extends to NGC 3109 (Grcevich and Putman, 2009; Spekkens et al., 2014). However, this is not the case as both have been found to be gas-rich satellites (Barnes and de Blok, 2001; Sand et al., 2015). In the same optical imaging campaign in which Antlia B was discovered, six more dwarf galaxies candidates were also discovered. These candidates, denoted as Dw1- Dw6, all have similar apparent V-band magnitudes and have projected distances from NGC 3109 between 40 80 kpc (Table 3.1). The optical images of these candidates ∼ − are shown in Figure 3.1. These dwarf galaxies candidates fall in the regime of ultra- faint dwarf galaxies (Geha et al., 2009). These galaxies are likely the most abundant type of galaxy and are the most dark matter dominated (Simon and Geha, 2007). If confirmed, these would be the first ultra-faint dwarfs discovered outside of the Milky Way environment (e.g. McConnachie, 2012) and would, therefore, provide potential insight into their evolution. After the HI observations presented in this chapter were obtained, follow-up observations with the Hubble Space Telescope (HST) were obtained (PID: 14078 PI: J. Hargis) for five of the six candidates, Dw2-Dw6. The preliminary analysis of the HST 3.1. THE NGC 3109 SYSTEM AND NEW SATELLITE CANDIDATES 57

Figure 3.1: 3 arcmin by 3 arcmin DECam images of the satellite candidates found around NGC 3109. Their names are labeled in the top-left of each image. The red circles denote a half-light radius of rh 0.3 arcmin,( 100 pc at their expected ≈ ∼ distance) typical of similar Milky Way satellites (McConnachie, 2012). 3.1. THE NGC 3109 SYSTEM AND NEW SATELLITE CANDIDATES 58

follow-up (D. Sand, private communication) shows no obvious sign of resolved stars which implies that these objects must lie beyond 17.5Mpc (Merritt et al., 2016). ∼ This lower distance limit implies that the candidates are not in fact satellites of NGC 3109; however, unless they lie at much larger distances than this limit, the objects remain in the dwarf regime. Our HI observations provide us with an opportunity to constrain the properties of these LSB features. If they are gas-rich, then their HI reservoirs provide estimates of their distances and insight to the dark matter halos in which they reside. Conversely, HI non-detections need to be understood in the context of gas scaling relations for nearby dwarf (Huang et al., 2012; Bradford et al., 2015), which show ﬁeld galaxies to be near-universally gas-rich.

Table 3.1: Properties of existing satellites and new satellite candidates around NGC 3109 Name RA Dec mV Dproj (hh:mm:ss) (dd:mm:ss) (mag) (kpc) Antlia 10:04:04 -27:19:51 15.1 40 Antlia B 09:48:56 −25:59:24 15.8 73 Dw1 10:03:42 -27:57:11 20.0 40.1 Dw2 09:57:59 -22:53:27 20.0 77.5 Dw3 09:54:51 -24:22:47 20.5 57.6 Dw4 09:51:40 -24:47:05 20.5 65.3 Dw5 09:51:58 -27:47:43 20.5 66.4 Dw6 10:15:12 -27:45:09 20.5 69.8 3.2. HI OBSERVATIONS AND ANALYSIS 59

3.2 HI Observations and Analysis

3.2.1 Observations

We conducted a total of 20 hours of position-switched HI observations of these LSB ∼ features with the GBT (PI: Spekkens, program AGBT16A-186) on multiple nights from May 2016 to July 2016. We used the instrumental setup described in Section 2.1.3 with VEGAS in Mode 7 (100 MHz bandwidth, 3.1 kHz channels). This allows us to search for HI emission at recessional velocities ranging from 7000 km s−1 to ∼ − ∼ 14000 km s−1. The integration times were determined by inputting the desired ﬂux sensitivities into the GBT Sensitivity Calculator1 available online. We calculate these

MHI = 1 sensitivities using the V-band luminosities and assuming LV (Grcevich and Putman, 2009; Spekkens et al., 2014) which gives us an expected HI mass. We can approximate the last term in Equation 1.3 to estimate the HI mass of a 5-sigma detection for a given velocity width:

5 2 MHI (M ) = 2.356 10 (D) 5σ (∆V ) , (3.1) ×

where ∆V is the velocity width of the putative HI emission spectrum and σ is our requisite sensitivity. We rearrange Equation 3.1 for σ and use this as our input to the GBT Sensitivity Calculator. The observation time for each target is listed in Table 3.2. These total times are split equally between the ON and OFF positions. We also include the smoothed velocity resolution of the spectra, ∆V , and the RMS noise at that velocity resolution,

σ∆V . 1https://dss.gb.nrao.edu/docs/Calculator_ug.pdf 3.2. HI OBSERVATIONS AND ANALYSIS 60

Table 3.2: Details of HI observations of the LSB features Name ON+OFF Obs. Time ∆V σ∆V (Hours) (km s−1) (mJy) Dw1 2.8 10 0.62 Dw2 4.0 15 0.49 Dw3 3.2 15 0.76 Dw4 3.3 15 0.45 Dw5 3.2 15 0.86 Dw6 3.3 15 0.57

3.2.2 Analysis

Our primary technical goal is to search for HI emission along the line of sight targeted by each observation. As explained in Section 2.2, we smooth our full-resolution spectra to a desired spectral resolution. We expect these satellite candidates to have Gaussian- like HI emission proﬁles (e.g. Sand et al., 2015), which is why we convolve our spectra with Gaussians. This process is more formally known as a matched ﬁltering technique (Andrews et al., 1970). A key component of this technique is that it reduces the noise

S of the spectra and, therefore, increases the signal-to-noise ratio ( N ) if there is a signal. After we smooth our spectra, we search for any emission features. For Dw2-Dw6 (Figure 3.2), we examine the spectra in 1000 km/s intervals, applying a baseline subtraction to each one. This subtraction is done in order to remove any large-scale (i.e. relative to the expected signal) variability that could reduce the amplitude of a signal. If there is still no sign of an emission feature, then we conﬁrm a non-detection

S and calculate upper limits. For Dw1, the search is rather short due to the high N ratio of the detected emission, shown in Figure 3.3. 3.2. HI OBSERVATIONS AND ANALYSIS 61

(a) Dw2 0.005

0.005 −

(b) Dw3 0.005

0.005 −

0.005 − Flux (Jy) (d) Dw5 0.005

0.005 −

(e) Dw6 0.005

0.005 − 1000 0 1000 2000 3000 4000 5000 − VHelio(km/s)

Figure 3.2: Spectra along the line of sight to Dw2 - Dw6. Milky Way emission is seen at velocities between -100 km/s and 300 km/s. Vertical dashed lines indicate channels that have had > 15% of scans in which RFI was removed. The dashed lines −1 that repeat in all panels around VHelio = 4300 kms are for a known RFI source (e.g. −1 cellular network signal). The two emission features in (e), at VHelio = 2500 kms and 4100 kms−1, are massive background galaxies that fall within the beam of the GBT. We ﬁnd no emission features along these lines of sight that can be attributed to Dw2 - Dw6: they are non-detections. 3.2. HI OBSERVATIONS AND ANALYSIS 62

Dw1 0.005

0.003

Flux (Jy) 0.001

0.001 − 500 1000 1500 2000 VHelio(km/s)

Figure 3.3: GBT HI spectrum for NGC 3109-Dw1 showing clear emission at vHelio ∼ 1060 kms−1.

We calculate HI mass upper limits from the non-detections: Dw2, Dw3, Dw4,

Dw5, and Dw6. To do this, we use Equation 3.1 and substitute σ∆V in Table 3.2 of the spectrum for the sensitivity, σ. As mentioned in 3.1, these objects were not § resolved into stars by HST, which sets a lower limit on their distances, D > 17.5 Mpc (Merritt et al., 2016). Similarly, our observations are limited by the bandwidth of

our spectrum which has a corresponding distance limit of D . 200 Mpc. Due to the uncertainty of the exact distances of these objects, we calculate 5σ upper limits in HI mass for the range of distances in between these two values. We list our calculated values in Table 3.3. Luminosity and HI mass estimates are determined in the distance range, as well as the upper limit on the HI mass to luminosity ratio. Recall that the latter property is distance-independent, and as a result does not have a range of values. 3.2. HI OBSERVATIONS AND ANALYSIS 63

For the Dw1 detection, we must first ensure that this emission feature is likely to stem from our target of interest. There is a potential for an HI gas-rich background source to be detected in our spectra if it is close enough in projection. We check the NASA/IPAC Extragalactic Database (NED)2 for other galaxies with HI emission in a 20 arcmin radius around our sources. We also check the deep DECam images directly for interlopers that may not be listed in NED. We find no plausible sources for the HI emission other than Dw1, and we move forward with our analysis under the assumption that we have detected its HI counterpart. We follow the method of Springob et al.(2005) to estimate the systemic velocity, velocity width, and the integrated flux of the detection. This is shown visually in Figure 3.4. We fit a polynomial (n=1) to the rising and falling sides of feature

between 15%-85% of the peak ﬂux minus the rms noise of the spectrum, fp rms, − as shown by the blue and green solid lines in Figure 3.4, respectively. The width of the feature is found using

W50 = VH,50 VL,50, (3.2) −

where VH,50 and VL,50 are the velocities corresponding to 50% of fp rms value for − the high and low velocity edges of the feature, respectively. We correct the velocity width for instrumental broadening, spectral smoothing, and redshift using Eq. (8) from Springob et al.(2005),

W50 2∆V Γ W50,c = − , (3.3) 1 + z

where ∆V is the spectral channel width in km s−1, Γ is a velocity width-dependent 2https://ned.ipac.caltech.edu/ 3.2. HI OBSERVATIONS AND ANALYSIS 64

10.0km/s 0.007

0.005

0.003 Flux (Jy) 0.001

0.001 − 800 1000 1200 1400 1600 VHelio(km/s)

Figure 3.4: Example of analysis method from Springob et al.(2005). We show the smoothed spectrum of NGC3109-Dw1 clearly revealing the emission feature. The blue and green solid lines represent the polynomial ﬁt (n=1) to the high- and low-velocity edges of the feature between 15%-85% of the fp rms value. − correction parameter (from Table 2 of Springob et al., 2005), and z is the redshift of the source. The second term in the numerator of Equation 3.3 applies the instrumental and smoothing corrections. We assume the error on W50,c to be 30% or 50% of this term if ∆V = 10 km s−1 or if ∆V > 10 km s−1, respectively. The systemic velocity of the feature is found using VH,50 + VL,50 Vsys = . (3.4) 2

We assume the error on Vsys to be the error on W50,c divided by √2. 3.2. HI OBSERVATIONS AND ANALYSIS 65

Distances are calculated using Hubble’s Law:

Vsys D = , (3.5) H0

−1 −1 adopting a Hubble’s constant of H0 = 70 1 km s Mpc (e.g. Jarosik et al., ± 2011). The integrated ﬂux, Sdv, is found by numerically integrating the data ´ corresponding to the emission in the spectrum. We substitute this value along with the distance into Equation 1.3 to determine the HI mass of the sources. The values derived from our analysis for Dw1 are shown in Table 3.4. 3.2. HI OBSERVATIONS AND ANALYSIS 66 (g) 3

. L V M HI L M )

(f) L 8 20 6 . V,D 10 L × ) c ( 02 48 48 01 86 . . . . .

2 0 1 2 1 2 1 V L HI . M L 0 M < < < < < (e) ± ) D 2 b 54 34 34 34 34 . ( . . . . . ) ) 8 8 8 8 8 v

L − − − − − L ) (Mpc) ( 41 21 21 21 21 02 15

...... log ( 0 (d) M 8 ± HI 10 ) 55) 6 73) 6 51) 6 64) 6 61) 6 M a 24 . . . . . × ( . 8 8 8 8 8 ) )( HI

− − − − − 9 1 M M (c) . ( 42 61 38 52 48 . . . . . S N (6 (6 (6 (6 (6 log ( )( Table 3.4: Properties of Dw1 < < < < < 2 8 1 (b) − ,c ± 50 W Dw2 Dw3 Dw4 Dw5 Dw6 Name (a): Log of HI(b): mass Log upper-limit of range V-band(c): luminosity HI-mass range to V-band luminosity ratio upper-limit 2 32 ) (km s 1 − (a) ± sys V (km s 1066 Dw1 Name (a): Heliocentric systemic velocity (b): Velocity width at(c): 50% Peak of signal-to-noise proﬁle ratio peak,(d): corrected HI for mass redshift(e): and instrumental Distance broadening calculated using(f): Hubble’s V-band Law luminosity calculated(g): at HI the mass distance to in V-band (e) luminosity ratio Table 3.3: Upper limits for range of distances of LSB features projected around NGC 3109 3.3. DISCUSSION 67

3.3 Discussion

We present the results from our HI observations of six dwarf galaxy candidates discovered in projection around NGC 3109 in the previous section. We find no HI emission in the spectra of five of the six dwarfs, Dw2-Dw6, and HI emission in the spectrum of Dw1. Following the analysis methods of Springob et al.(2005) we estimate the systemic velocity, Vsys, corrected velocity width, W50,c, and HI line flux, Sdv which ´ we use to estimate the distance and HI mass of this system. These values are listed in Table 3.4. Here, we consider the implications of our five non-detections and our one detection.

−1 With a systemic velocity of Vsys = 1066 km s , Dw1 is not physically associated

−1 with NGC 3109 (Vsys = 403 km s , McConnachie, 2012). While isolated gas-rich dwarfs are ubiquitous (e.g. Bradford et al., 2015), we search for any nearby galaxies using NED that may be associated with Dw1. Our criteria are objects with systemic velocities that are within 100 kms−1 and have angular separations of < 1.89 degrees (or 500 kpc at the distance of Dw1). For the purpose of this work, we conduct ∼ this search with fairly conservative constraints. We ﬁnd ﬁve galaxies that satisfy our criteria and list them in Table 3.5. 3.3. DISCUSSION 68

Table 3.5: Details of potential host galaxies for Dw1.

Name Vsys DHubble θsep Dsep log (MHI ) Ref. −1 (kms ) (Mpc) (deg) (kpc) (M ) Dw1 1066 2 15.2 0.2 — — 8.09 — ± ± NGC 3113 1083 15.5 0.52 137 9.35 (1) ESO 435- G 039 1045 14.9 0.65 173 8.99 (2) ESO 435- IG 020 976 13.9 0.97 258 — — ESO 435- G 016 974 13.9 1.27 338 — — NGC 3137 1104 15.8 1.63 432 — — Ref.: (1) Courtois and Tully(2015) (2) Koribalski et al.(2004b)

Interestingly, the two galaxies that project nearest to Dw1 also have systemic velocities that are within 20 kms−1 suggesting that they may be associated. NGC ∼ 3113 and ESO 435- G 039 are both gas-rich galaxies and we list their HI masses from the literature in Table 3.5. With the addition of Dw1, it is possible that these objects may form an association of gas-rich objects. However, it is difficult to be certain of their association with one another without the direct distances measurements (i.e. HST) and, therefore, separations of these objects. Our non-detections combined with the preliminary HST results rule out the possibility of an association between these satellites, Dw2-Dw6, with NGC 3109. This leaves two main possibilities for where the locations of these objects along the line of sight. The first is that they are associated with a background galaxy and the second is that they are located in the field. From our detailed studies of the Local Group, we know that gas-poor dwarf galaxies are typically satellites of a more massive galaxy (see 1.4). With this information, § our first step is to search for nearby background galaxies that could potentially host 3.3. DISCUSSION 69

these dwarfs. We search NED for galaxies within 1 degree (60 arcmin) of each dwarf,

corresponding to a projected separation of Dsep 300 kpc at the distance limit set by ∼ HST (D > 17.5 Mpc; Merritt et al., 2016) and Dsep 3.5 Mpc at the distance limit ∼ set by our observations (D 200 Mpc). For the purpose of this work, we consider ∼ galaxies with projected separations of less than 300 kpc (the virial radius of the Milky Way) at all distances. The following discussion presents the galaxies that may play host to our satellites one at a time. We ﬁnd three potential hosts within 300 kpc of Dw2. TOLOLO 0957-225 (Bohuski et al., 1978) and NGC 3081 (Dreyer, 1888) are located at similar distances of 33.4 Mpc and 32.5 Mpc, respectively. Both have similar projected separations from Dw2 of Dsep 200 kpc. The third potential host, TSK2008 0499 (Tully et al., 2008), is ∼ a galaxy group located at a distance of 37.7 Mpc and has a separation of 236 kpc. Unfortunately, none of these objects have easily accessible mass estimates in order to predict the satellite populations around them and the likelihood of association with our satellite candidate (e.g. Dooley et al., 2017). Dw4 has a single potential host within 300 kpc, ESO 499-G 008 (Bottinelli et al., 1993) located at a distance of 38 Mpc and a projected separation of 293 kpc. If

Dw4 were a satellite of this galaxy it would have an HI mass upper limit of MHI <

7 6 Mstellar M 1.1 10 M and a stellar mass of Mstellar = 7.8 10 M (assuming = 1 ; × × LV L e.g. Martin et al., 2008). ESO 499-G 008 has HI measurements from the HIPASS

catalog (Doyle et al., 2005); using their ﬂux measurements, we estimate MHI =

9 3.1 10 M . We can take this a step further by using the gas mass to stellar mass × 9 relations of Bradford et al.(2015) to give us a stellar mass of Mstellar = 2.2 10 M , × 3.3. DISCUSSION 70

similar to the LMC. Predictions of the satellite populations around LMC and SMC- like galaxies were made by Dooley et al.(2017), who make use of abundance matching and reionization models on the Caterpillar simulations (Griﬀen et al., 2016). Their results show that for an LMC sized host galaxy there should be < 2 satellites with

6 Mstellar 7 10 M . This prediction is not consistent with our observational results, ≈ × however, concrete conclusions about their association cannot be made. There are two potential hosts nearby Dw6. ESO 436-G 006 is at a distance of 38.2 Mpc and has a projected separation of 135 kpc. NGC 3173 is spiral galaxy located at a distance of 39.1 Mpc and has a separation of 103 kpc. While the former has no mass information, we estimate the HI mass of the latter using measurements from the

9 HIPASS catalog (Doyle et al., 2005). We find MHI = 5.3 10 M and estimate its × 9 stellar mass using the relations in Bradford et al.(2015) to be Mstellar = 7.1 10 M . × 7 We estimate that Dw6 would have an HI mass upper limit of MHI < 1.5 10 M and × 6 a stellar mass of Mstellar = 8.2 10 if it were a satellite of NGC 3173. Dooley et al. × 5 (2017) also provide an estimate for the number of satellites with Mstellar > 10 M as a function of host mass. For NGC 3173, we would expect to find 0 5 satellites with ∼ − 5 Mstellar > 10 M . These estimates are in agreement with our observational results but concrete associations cannot be made. We show the upper limits that these dwarfs would have at the distance of their potential hosts in the upper panel of Figure 3.5. The two potential hosts for Dw5, NGC 3037 and NGC 3056, have distances that fall short of the HST distance limit. We find no hosts within 300 kpc of Dw3 in our search. Another possibility is that these objects have no massive companions but are instead in the field. Geha et al.(2012) studied the star formation properties of dwarf 3.3. DISCUSSION 71

galaxies in the field using data from the NASA-Sloan Atlas, a reprocessed version of the Sloan Digital Sky Survey imaging catalog with the addition of ultraviolet data from GALEX (Galaxy Evolution Explorer). They find that dwarf galaxies in the field

7 9 with 10 M < Mstellar < 10 M are almost entirely actively star forming; < 0.06% have no active star formation and there are 0 quenched sample galaxies (galaxies with no recent star formation). The HI gas that we search for is the primary fuel source in the complex star formation process. If our dwarfs are in fact located in the field with distances greater than 30 to 40 Mpc, then our non-detections are in tension with the findings of Geha et al.(2012). However, placing these dwarfs at smaller distances is not inconsistent with their findings since the stellar mass ranges being studied are different. 3.3. DISCUSSION 72

9.0

8.5 ) ⊙ 8.0 M )( 7.5 HI Dw2 M 7.0 Dw3

log( Dw4 6.5 Dw5 Dw6 ) ⊙ 9.0 M

)( 8.5 M 8.0 log( 7.5 ∼ )

⊙ 7.0 L

)( 6.5 Dw2 V

L 6.0 Dw3-Dw6 Geha et al. 2012

log( 5.5 50 100 150 200 D (Mpc)

Figure 3.5: Distance-dependent properties of our dwarf galaxies in projection around NGC 3109. The upper panel shows the upper limits on the HI mass as a function of distance. The lower panels shows the luminosity of the dwarfs as a function of Mstellar = 1 M distance. Assuming LV L , the lower panel can be used as a proxy for stellar mass as a function of distance. Stars indicate the distance limit set by HST (17.5 Mpc) and circles indicate the distance limit set by our observations with the GBT (202 Mpc). Squares with downward arrows indicate the upper limits that our dwarfs would have if they were associated with the hosts that project within 300 kpc along the line-of-sight. The dash-dotted lines in the lower panel show the range of stellar masses in the Geha et al.(2012) sample.

For further insight, we can also turn to the simulation work conducted by Shen et al.(2014). The authors perform cosmological simulations of a group of seven ﬁeld

8 10 dwarf galaxies with virial masses 4.38 10 M < Mvir < 3.59 10 M . Only × × four of the dwarf galaxies were able to form stars with stellar masses in the range of 3.3. DISCUSSION 73

4 8 5 7 9.6 10 1.15 10 M and HI masses of 5.4 10 2.34 10 M . We compare × − × × − × the upper limits on MHI MHI of our dwarfs with the simulated sample in Figure LV ∼ Mstellar 3.6. The simulated sample have decreasing MHI MHI as they increase in stellar LV ∼ Mstellar mass. If our dwarfs have log (LV ) log (Mstellar) < 7.5, then their upper limits would ∼ follow this trend. Recall that our range of luminosities are distance dependent. If we expected our dwarfs to follow this trend, then they must have distances lower than D 75 Mpc. However, these are upper limits and the actual values of MHI may be ∼ LV much lower than this trend would predict. 3.4. CONCLUSIONS 74

1.5 Dw2 Dw3 Dw4 Dw5 Dw6

) 1.0 Shen 2014 stellar /M

HI 0.5 M log( ∼ )

V 0.0 /L HI M

log( 0.5 −

4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 log(LV )(L ) log(Mstellar)(M ) ⊙ ∼ ⊙

Figure 3.6: HI mass to V-band luminosity ratio as a function of V-band luminosity. Mstellar = 1 M We assume a LV L in order to use the luminosity of our dwarfs as a proxy for their stellar masses. Our dwarfs have upper limits that span a range of luminosities and are shown as horizontal lines with downward arrows at their ends. The simulated dwarfs from Shen et al.(2014) are shown as stars.

3.4 Conclusions

In this chapter, we presented our HI observations and results of LSB features projected around NGC 3109. We ﬁnd that ﬁve of the six galaxies observed are gas-poor (with HI non-detections) and that one is gas-rich (HI detection). None of these candidates are 3.4. CONCLUSIONS 75

associated with the Local Group dwarf galaxy NGC 3109 since HST doesn’t resolve individual stars in these systems, implying D > 17.5 Mpc (Merritt et al., 2016). We suspect that Dw1 may be part of a loose association with two background galaxies, NGC 3113 and ESO 435-G 039. For 3 of the dwarfs with HI non-detections, we find potential background host galaxies. Without proper distance measurements there is still the possibility that these dwarf galaxies are located in the field. In this case, we show that to avoid tension with the properties of gas-rich field dwarfs these gas-poor dwarfs must be located at a distance between 17.5 Mpc and 40 Mpc. Chapter 4

The Satellite Population around M101

In this chapter, we describe the M101 satellite population. In Section 4.1, we provide an overview of the M101 group, including its existing satellites as well as new satellite candidates. We describe our observations and analysis in Section 4.2. We discuss our results in Section 4.3 and provide our conclusions in Section 4.4.

4.1 The M101 Group

M101 is a nearly face-on spiral galaxy located in the Local Volume (DM101 = 7 Mpc; Tikhonov et al., 2015) and is one of the most well-studied nearby massive galaxies. M101 is considered to be a Milky Way analog due its similar size and stellar mass

10 Mstellar = 5.3 10 M (van Dokkum et al., 2014). M101 is an interesting galaxy due × to some of its unique physical properties, particularly its massive asymmetric spiral structure.

76 4.1. THE M101 GROUP 77

In their HI observations of M101, Mihos et al.(2012) found an extended plume of HI approximately 100 kpc in length, showing signs of diffuse gas between M101 and its nearby companion NGC 5474. This suggests that these galaxies may have interacted in the past and that the gas from M101’s extended HI disk is tidally disturbed. Mihos et al.(2013) follow up their HI study with an optical imaging campaign to study the stellar populations in its outer disk. They find no extended optical features that would further suggest that M101 has had any recent interactions. However, they state that asymmetry in M101 and NGC 5474 along with the extended HI emission in M101 suggest some interaction history. The outskirts of M101 were studied by van Dokkum et al.(2014) using the Drag- onfly Telephoto Array. They do not find photometric evidence for a stellar halo, implying a stellar halo mass fraction that is an order of magnitude lower than those in the Milky Way and M31. Since the stellar halos of galaxies are thought to form through close interactions with accreting satellites (e.g. McConnachie et al., 2009b), the faint stellar halo in M101 suggests that galaxies with similar stellar masses can have different satellite accretion histories. This is consistent with the cosmic scatter observed from numerical simulations of halo assembly (Springel et al., 2008; Garrison- Kimmel et al., 2014). Aside from NGC 5474, the M101 group consists of three more galaxies: NGC 5477, Holmberg IV (UGC 8837), and UGC 9405. They are somewhat similar in their properties to the brightest Milky Way satellites (Tikhonov et al., 2015). However, to compare M101’s satellite population to the Milky Way we need to perform deeper searches for objects representative of the fainter end of the Milky Way’s satellite mass function (e.g. Figure 1.3). 4.1. THE M101 GROUP 78

Many recent searches have been conducted around M101 using small-aperture telescopes to perform wide-field imaging searches for low surface brightness features. Several of these surveys have searched the region around M101, finding many satellite candidates (Merritt et al., 2014, hereafter M14; Karachentsev et al.(2015), hereafter K15; Javanmardi et al.(2016), hereafter J16). M14 use the Dragonfly Telephoto array to search a 3.3 2.8 deg2 field finding 7 low-surface brightness features (DF1-DF7; see × Figure 1.5). K15 use a series of amateur telescopes (10 to 110 cm in diameter) to conduct surveys around more than 50 galaxies out to 10 Mpc. They recovered the objects first reported by M14 along with 4 new satellite candidates around M101 (dwA-dwD). J16 present the Dwarf Galaxy Survey with Amateur Telescopes (DGSAT) which uses telescopes with sizes ranging from 0.36 meters to 0.5 meters. In their 2.0 1.3 deg2 × field around M101, they recover 5 of the 7 M14 satellite candidates as well as 1 of the K15 satellite candidates (dwA/DGSAT1). One can also mine existing imaging survey data to find stellar over-densities that correspond to satellite candidates. This is what was done by Müller et al.(2017); hereafter M17, and Bennet et al.(2017); hereafter B17. M17 use Sloan Digital Sky Survey (SDSS) to algorithmically survey 330 deg2 around M101, M51, and M63. These three massive spiral galaxies form a “planar structure” in the sky. They find 6 new satellite candidates (we have labeled them dwM1-dwM6) located at much larger projected separations than previous studies. B17 perform a similar search using Canada-France-Hawaii-Telescope Legacy Survey data. A total of nine 1 deg2 fields were searched, resulting in the discovery of 38 new satellite candidates and the recovery of 11 of those previously reported by in M14, K15, and J16. An interesting outcome of all of these searches is the asymmetric distribution of 4.1. THE M101 GROUP 79

satellite candidates around M101. A large fraction of these galaxies are located on the East side, with only 1 located in the South-West region (see Figure 4.1). This type of subhalo distribution is not expected in the ΛCDM paradigm and is not seen in simulations (Ibata et al., 2013; Pawlowski et al., 2012). While it is tempting to make a connection between this over-density of satellite candidates and M101, there is also a background group in this region. NGC 5485 and NGC 5473 are two massive elliptical galaxies that are part of a background group located at distance of 28 Mpc. ∼ We can attempt to ﬁnd associations between these newly discovered satellite candidates and either the M101 or NGC 5485/5473 group by searching for HI emission. If these galaxies are gas-rich, then the resulting spectra will provide redshift information that can be used to infer distances and therefore group membership. On the other hand, the majority of these galaxies are located within the projected virial radius of M101 ( 260 kpc; Merritt et al., 2014). If the the satellite population of M101 has ∼ the same gas properties as those of the Milky Way, then these satellites should be gas-poor (see Figure 1.6). Our sample consists of satellite candidates with apparent magnitudes mV . 20 mag with a distribution around M101 shown in Figure 4.1. Some of these galaxies have follow-up observations that have or will provide a strong case for their association. All of the M14 galaxies have HST observations with only 3 being conﬁrmed as satellites of M101 (Danieli et al., 2017). The authors of B17 have existing and forthcoming HST observations of 15 of their satellite candidates as well as the 4 found by K15. Preliminary analysis of 2 of the satellites (K15: dwB and B17: Dw6) suggest that they are in the background due to the lack of resolved stars in their HST images (P. Bennet, private communications). In this chapter, we treat all LSB features around M101 for which distance information is not yet available as 4.1. THE M101 GROUP 80

satellite candidates. 4.1. THE M101 GROUP 81 200 205 210 Right Ascension (deg) 215 220 M14/K15/J16 B17 M17 NGC 5485 M101

225

59 58 57 56 55 54 53 52 elnto (deg) Declination Figure 4.1: DistributionM14, of K15, M101 or satellite J16.M17. candidates Red The in circles yellow show our star candidatesthe sample. indicates discovered background. the by Blue position B17. circles of Magenta show M101 circles and candidates show the candidates discovered black discovered by by triangle indicates the position of NGC 5485 in 4.2. OBSERVATIONS AND ANALYSIS 82

4.2 Observations and Analysis

We conducted a total of 67 hours of observations for 27 of these satellite candi- ∼ dates with the GBT (PI: Spekkens, programs AGBT16B-046, AGBT17A-188, PI: Karunakaran, program AGBT17B-235) on multiple nights from August 2016 to Au- gust 2017. We use the instrumental setup described in 2.1.3 with VEGAS in Mode § 7 for all but one target. Recall that Mode 7 allows us to search for gas-rich objects with recessional velocities ranging from 7000 kms−1 to 14000 kms−1. For DF3 ∼ − ∼ +0.25 (distance measured using HST imaging, D = 6.25−0.27 Mpc; Danieli et al., 2017), we use VEGAS in Mode 15 with a bandwidth of 11.72 MHz corresponding to recessional velocities ranging from 950 kms−1 to 1400 kms−1. ∼ − ∼ We determined the integration times in a similar manner to the method described in 3.2. However, there are two distinct diﬀerences. For the satellites in our sample § MHI = 0.7 M from M14, K15, and J16, we assume LV L which is 2 standard deviations below the mean of gas-rich dwarf galaxies (Bradford et al., 2015; Huang et al., 2012). This HI mass-to-light ratio was chosen for a broader group of satellites around multiple

MHI = 2 M hosts. For the satellites in our sample from B17 and M17, we assume LV L , the same as the gas-rich dwarf Leo P which has a similar luminosity to the satellites in this subset of our sample. The former method results in deeper searches while the latter still sufficiently probes the potential HI reservoirs but drastically reduces the integration times. We do not expect our results to be significantly affected by these two different methods of determining integration times. We show the properties and details of the observations of our satellite candidate sample in Table 4.1. Our detections and non-detections are treated in the same way as those in 3.2. Briefly, for non-detections we compute 5σ upper limits on the HI § 4.2. OBSERVATIONS AND ANALYSIS 83

mass of a satellite along that line-of-sight. For detections we follow the methods of Springob et al.(2005) to determine systemic velocities, velocity widths, and flux integrals. In Figure 4.2 we show the spectra for our non-detections. A total of 23 target galaxies have no HI emission along their lines-of-sight that we associate to them. We find 4 satellite candidates (DF1, dwC, Dw8, and Dw38) that are close enough to the HI disk of M101 to have its HI emission detected during the observation. Three other satellite candidates (Dw13, Dw31, and Dw32) show slight signs of emission which we also attribute to the HI disk of M101. Figure 4.3 displays the positions of these objects relative to the extended HI disk of M101 observed by Mihos et al. (2012). Background galaxies are also detected along the lines-of-sight to our targets of interest: these are labeled in Figure 4.2, panels (k) and (u), and identified by searching NED as described in 3.3. We summarize our upper limits in Table 4.2. § Note that for galaxies that have clipped the M101 HI disk our upper limits are only valid outside this velocity range. We provide ranges for the two galaxies (DF5 and Dw6) that have been confirmed as background galaxies through HST observations. The spectra for our satellites with HI emission is shown Figure 4.4 and their derived properties are in Table 4.3. Two of these galaxies (dwB and dwM3) have velocities that are consistent with the NGC 5485 group in the background. Another galaxy, Dw26, is at a greater distance in the background. The final detection is of dwM1 which is the only satellite candidate that we find to have a systemic velocity consistent with that of the M101 group. 4.2. OBSERVATIONS AND ANALYSIS 84 (f) 39 42 23 52 43 22 39 39 39 25 20 38 23 34 26 V ...... ∆ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 σ ) (mJy) 1 − (e) V ∆ (km s (d) Time . 4.7 25 1.8 25 4.2 25 ON + OFF Obs (c) V m (b) . Ref (a) Dec (a) Table 4.1: M101 satellite candidates and their observations RA (hh:mm:ss) (dd:mm:ss) (mag) (Hours) Dw7 14:07:21 55:03:51 B17 20.1 Dw8 14:04:25 55:06:13 B17 19.5 DF1 14:03:45 53:56:40 M14 18.6 1.4 25 DF2 14:08:38 54:19:31 M14 19.1 4.8 25 DF3 14:03:06 53:36:56 M14 17.5 1.3 25 DF5 14:04:28 55:37:00 M14 17.7 0.8 25 dwB 14:08:43 55:09:57 K15 20.3 3.2 15 Dw3 14:08:46 55:17:14 B17 19.5 1.7 25 Dw4 14:13:02 55:11:16 B17 20.0 3.6 25 Dw5 14:04:13 55:43:34 B17 20.2 6.0 25 Dw6 14:02:20 55:39:17 B17 19.6 1.6 25 dwC 14:05:18 54:53:56 K15 20.1 5.2 25 dwD 14:04:24 53:16:19 K15 19.3 1.0 25 Dw13 14:08:01 54:22:30 B17 20.1 Name DGSAT1 (dwA) 14:06:50 53:44:29 J15 (K15) 18.7 2.6 25 4.2. OBSERVATIONS AND ANALYSIS 85 V (f) ∆ . 34 33 75 54 68 35 89 36 52 65 47 V . . . . 0 ...... ∆ 0 0 0 0 0 2 0 1 0 0 1 σ ) (mJy) 1 − (e) V ∆ (km s (d) Time . ON + OFF Obs (c) V m (b) . Table 4.1 Continued Ref (a) Dec (a) RA (hh:mm:ss) (dd:mm:ss) (mag) (Hours) * - indicates objects(a): with Satellite incomplete candidate observations at coordinates,Karachentsev the (b): et time al. ( 2015 ); Discovery of J16, paper, writing MüllerJavanmardi M14, et etMerritt al. ( 2016 ); al. ( 2017 ), et B17, (c): al. ( 2014 );dueBennet K15, V-band to et apparent RFI,); al. ( 2017 magnitude, (e): M17, (d): Velocity Total resolution observing of time final including spectrum, time (f): lost RMS noise in spectrum at given Dw19 14:10:20Dw26 14:08:50Dw31 54:45:50 14:07:42Dw32 53:27:24 B17 14:07:46Dw33 54:35:18 B17 20.1 14:08:34Dw38 54:15:25 B17 19.9 14:01:18 55:26:49 B17 19.0 54:21:14 B17 18.9 1.8* B17 19.4 3.4 19.9 0.9 0.4* 25 1.5 30 1.4* 25 25 25 25 Name dwM1 13:43:07 58:13:40 M17 15.3 0.2 10 dwM2 13:55:11 51:54:29 M17 18.4 0.5 25 dwM3 14:08:41 56:55:38 M17 17.7 0.2 15 dwM4 14:12:11 56:08:31 M17 19.1 1.8 25 dwM5 14:16:59 57:54:39 M17 18.9 1.1 25 dwM6 14:47:00 58:34:04 M17 18.1 0.3 25 4.2. OBSERVATIONS AND ANALYSIS 86

0.005 (a) DF1

0.005 − 0.005 (b) DF2

0.005 − 0.005 (c) DF5

0.005 − 0.005 (d) dwD Flux (Jy)

0.005 − 0.005 (e) DGSAT1

0.005 − 0.005 (f) dwC

0.005 − 1000 0 1000 2000 3000 4000 5000 − VHelio(km/s)

Figure 4.2: GBT spectra of galaxies around M101 with no HI emission. Vertical dashed lines indicate channels that have had RFI removed in > 15% of scans (see 2.2). For the candidates that overlap the M101 HI disk, blue dash-dotted lines § indicate the velocity range of M101 HI emission. Milky Way emission is seen at velocities between -200 km/s and 150 km/s. 4.2. OBSERVATIONS AND ANALYSIS 87

0.005 (g) Dw3

0.005 − 0.005 (h) Dw4

0.005 − 0.005 (i) Dw5

0.005 − 0.005 (j) Dw6 Flux (Jy)

0.005 − 0.005 (k) Dw7

0.005 N5486 SDSS Gal. − 0.005 (l) Dw8

0.005 − 1000 0 1000 2000 3000 4000 5000 − VHelio(km/s) Figure 4.2 continued 4.2. OBSERVATIONS AND ANALYSIS 88

0.005 (m) Dw13

0.005 − 0.005 (n) Dw19

0.005 − 0.005 (o) Dw31

0.005 − 0.005 (p) Dw32 Flux (Jy)

0.005 − 0.005 (q) Dw33

0.005 − 0.005 (r) Dw38

0.005 − 1000 0 1000 2000 3000 4000 5000 − VHelio(km/s) Figure 4.2 continued 4.2. OBSERVATIONS AND ANALYSIS 89

0.005 (s) dwM2

0.005 −

0.005 (t) dwM4

0.005 −

0.005 (u) dwM5

0.005 SDSS Gal. Flux (Jy) −

0.005 (v) dwM6

0.005 − 1000 0 1000 2000 3000 4000 5000 − 0.005 (w) DF3

0.005 − 500 0 500 1000 − VHelio(km/s) Figure 4.2 continued 4.2. OBSERVATIONS AND ANALYSIS 90

55.5 Response level 10% 1% 55.0 0.1%

54.5 Declination (deg) 54.0

53.5

212.0 211.5 211.0 210.5 210.0 Right Ascension (deg)

Figure 4.3: Positions of satellite candidates with HI emission detected from the disk of M101. The hatched yellow ellipse approximately shows the extended HI disk mapped by Mihos et al.(2012). Red filled circles represent the objects with significant M101 HI emission in their spectra, decreasing in Dec.: Dw8, dwC, and Dw38, DF1. Blue filled circles show objects with less significant signs of HI emission from M101, decreasing in Dec.: Dw31, Dw13, and Dw32. The thin green (10%), blue (1%), and red (0.1%) open circles around these objects show the response level of the GBT beam at that position. 4.2. OBSERVATIONS AND ANALYSIS 91 Dw26 dwM3 (b) (d) s) 1000 1500 2000 2500 3000 10000 10500 11000 11500 12000 / 002 008 006 004 002 000 ...... 0007 0014 0007 0000 . 0 . . . 0 0 0 0 0 0 0 0 0 − − (km Helio V dwB dwM1 0 200 400 600 800 of scans. The solid green lines in panel a indicate Milky Way emission. 200 − 15% (a) > (c) 400 1000 1500 2000 2500 3000 − 004 001 016 012 008 004 000 020 005 003 001 ......

0 0 0 0 0 0 0 0 0 0 0

− − lx(Jy) Flux Figure 4.4: GBTwas spectra removed with in HI emission. The vertical dashed line in panel c indicates channels in which RFI 4.2. OBSERVATIONS AND ANALYSIS 92

Table 4.2: Upper limits for conﬁrmed satellites, satellite candidates, and background galaxies that are projected around M101 (d) (a) (b) (c) MHI Association log (MHI ) log (Lv) Name LV

M log (M ) log (L ) L DGSAT1 Can.* < 5.75 6.14 < 0.40 dwC Can. * < 5.52 5.59 < 0.84 dwD Can.* < 5.88 5.89 < 0.96 DF1 Conf.† < 5.71 6.09 < 0.41 DF2 Conf.† < 5.50 5.96 < 0.34 DF3 Conf.† < 5.68 6.54 < 0.14 DF5 BG† < (6.67 8.79) (7.31 9.41) < 0.23 − − Dw3 Can.* < 5.75 5.83 < 0.82 Dw4 Can.* < 5.56 5.63 < 0.83 Dw5 Can.* < 5.47 5.56 < 0.82 Dw6 BG† < (6.54 8.65) (6.58 8.70) < 0.90 − − Dw7 Can.* < 5.53 5.57 < 0.90 Dw8 Can.* < 5.69 5.80 < 0.76 Dw13 Can.* < 5.58 5.59 < 0.97 Dw19ˆ Can. < 5.69 5.59 < 1.25 Dw31 Can. < 5.90 6.00 < 0.79 Dw32ˆ Can. < 6.03 6.07 < 0.91 Dw33 Can. < 5.89 5.84 < 1.11 Dw38ˆ Can. < 5.99 5.64 < 2.24 dwM2 Can. < 6.11 6.27 < 0.68 dwM4 Can. < 5.92 6.00 < 0.83 dwM5 Can. < 5.97 6.05 < 0.84 dwM6 Can. < 6.32 6.36 < 0.90 ˆ - Objects with incomplete GBT HI observations * - Objects with forthcoming HST observations - Objects with existing HST observations (a):† Association with M101: Candidate satellite, Can.; Conﬁrmed satellite, Conf.; Background Galaxy, BG. (b): HI mass upper limit (c): V-band luminosity (d): HI-mass to V-band luminosity ratio 4.2. OBSERVATIONS AND ANALYSIS 93 (g)

23 82 22 . . . L V M HI L M )

(f) L 1 0 55 13 1 7 . . . 3 V,D 10 L × ˆ 4 0 4 5 . . 1 2 23 3 . 0 0 0 ± (e) ± ± ± 3 2 D . . 0 . ) (Mpc) ( 15 7 8 27

. . 2 27 0 26 156 0 (d) M ± 7 ± ± ± HI 7 . 5 10 . M 71 × . 2 6 4 0 6 88 4 6 . (c) . . . S N )( 1 4 1 5 11 4 4 (b) 16 4 − ,c ± ± ± ± 50 W ) (km s 12 81 4 69 2 58 1 37 1 − (a) ± ± ± ± sys Table 4.3: Properties of candidate satellites with HI detections V 195 (km s 1913 1904 10972 † - Object with existing HST observations dwB Dw26 Name dwM1 dwM3 ^ - Adopting the† M101 distance (a): Heliocentric systemic velocity (b): Velocity width at(c): 50% Peak of signal-to-noise proﬁle ratio peak,(d): corrected HI for mass redshift(e): and instrumental Distance broadening calculated using(f): Hubble’s V-band Law, luminosity unless calculated stated(g): at otherwise. HI the mass distance to in V-band (e) luminosity ratio 4.3. DISCUSSION 94

4.3 Discussion

In the previous section we presented the results of our observations along the lines- of-sight to 27 objects found in projection around M101. 4 of these galaxies have HI emission in their spectra while the remaining 23 have none. In this section, we discuss our ﬁndings and their implications in comparison to other work.

4.3.1 Satellites of M101

In Table 4.2, we separate our non-detections into three groups (col. 2 of Table 4.2): Candidates, Confirmed, and background galaxies. The second group includes galaxies that have been confirmed to be satellites of M101 through distance measurements using HST (Danieli et al., 2017). The third group includes two galaxies, DF5 and Dw6, that have published (Merritt et al., 2016) and preliminary (P. Bennet, private communication) HST data with no resolved stars and reside in the background. We will discuss the third group in the context of the next subsection ( 4.3.2) and the first § two in this one. The existing satellite population of M101 was increased with the addition of confirmed satellites DF1, DF2, and DF3. We find that these three galaxies have no HI emission along their lines of sight and are, therefore, gas-poor. These galaxies have projected separations (50 to 96 kpc) that place them well inside the virial radius of

M101 (rvir = 260 kpc; Merritt et al., 2014). Using the measured distances found by

+349 +278 Danieli et al.(2017), the 3D distances from M101 of these dwarfs 630−347 kpc, 161−36

+267 kpc, 490−241 kpc, also places them inside the virial radius within their uncertainties. The lack of gas in these three galaxies implies that they have potentially had their gas stripped away by M101. 4.3. DISCUSSION 95

Our observations conﬁrm a new galaxy to the M101 group, which we have labeled dwM1. Müller et al.(2017) classify dwM1 as likely being a BCD (Blue Compact Dwarf) due to its morphological properties: a high-surface brightness central region

MHI < 1 M surrounded by a low-surface brightness envelope. BCDs generally have LV L (Thuan and Martin, 1981) and HI properties of dwM1 are commensurate with BCDs with similar optical properties (see MK 475, Thuan et al., 1999).

We show the relative velocities, ∆V = √3 (Vsys Vsys,host), of M101 group mem- − bers as a function of their distance, r, from M101 in Figure 4.5. The factor of √3 statistically corrects from line-of-sight velocity to 3D velocity. We use the values found by Tikhonov et al.(2015) for the relative velocities and separations for the group members, as well as for the estimated dynamical mass of M101. dwM1’s sys-

−1 temic velocity, Vsys,dwM1 = 195 1 kms and separation place it just outside of the ± estimated escape velocity curves of M101. However, there are two factors that would allow dwM1 to be a group member. Firstly, the addition of more group members (DF1-DF3) could increase the estimated dynamical mass of M101. Secondly, the relative velocity calculation of dwM1 statistically corrects from the line-of-sight velocity to 3D velocities. If the line-of-sight velocity we measure is much larger than the tangential component, then we are over-correcting to a 3D velocity. Distance measurements using HST would prove to be useful in conﬁrming the association. 4.3. DISCUSSION 96

300 Vesc,M = 7.5e11 M ⊙ dwM1 200 M101 Group Members

100 )

0 km/s ( V ∆ 100 −

200 −

300 − 0 200 400 600 800 1000 r (kpc)

Figure 4.5: Relative velocities of M101 group members (red circles) and dwM1 (yellow star) as a function of separation from M101. The blue lines indicate the escape velocity 11 of a point mass with M = 7.5 10 M . ×

Accurate distances to nine satellite candidates with HI non-detections (marked with an * in Table 4.2) will be constrained with upcoming HST observations (PI: D.Sand). In the absence of distance estimates for the time being, we conclude this subsection by discussing the implications of assuming that they are at the distance of M101. We begin by comparing our sample with two larger samples of gas-rich dwarf galaxies. Huang et al.(2012) discovered their galaxies in the ALFALFA survey and 4.3. DISCUSSION 97

combined optical images along with GALEX UV measurements to estimate stellar masses. Bradford et al.(2015) take the complementary approach by first defining a sample of galaxies from existing optical catalogs and then obtaining follow-up HI measurements. They separate their sample into isolated and non-isolated galaxies and find no significant difference in the fraction of gas between them. We therefore consider just the isolated subsample as it is, generally, representative of the whole sample. Also, we only include galaxies with stellar mass estimates in our comparison, 74 from Huang et al.(2012) and 1700 from Bradford et al.(2015). We estimate ∼ the stellar masses of our sample galaxies using measured luminosities and assuming

Mstellar = 1 M LV L (e.g. Martin et al., 2008). The comparison of these three samples is shown in Figure 4.6. The ﬁrst thing to note is the vast range of stellar masses that are being compared, almost six orders of magnitude. The Huang et al.(2012) samples hovers in the mid- range of stellar and HI mass, whereas the sample presented here covers a narrow range in both. The black dashed line shows the linear ﬁt to the Bradford et al.(2015)

8.6 8.6 sample in two stellar mass ranges, Mstellar < 10 M and Mstellar 10 M . We ≥ extend their fit to lower masses to show where the Huang et al.(2012) sample and our sample lie relative to their low-mass trend. The Huang et al.(2012) galaxies loosely follow the Bradford et al.(2015) relation, with significant scatter in stellar mass. Some of the Huang et al.(2012) galaxies even fall into the gas-depleted region defined by Bradford et al.(2015). Our non-detections fall in a tight region between the fit to the Bradford et al.(2015) sample and the gas-depleted region (though the actual values could lie anywhere below the plotted points). Only the BCD and one of the confirmed M101 satellites fall into the gas-depleted region. It can also be seen 4.3. DISCUSSION 98

that gas-rich galaxies with similar stellar masses to those in our sample do exist. We must also consider the possibility that the galaxies without confirmed distance measurements may be members of the NGC 5485 group in the background. We show a red arrow in Figure 4.6 indicating the direction and magnitude of the shift a galaxy in our sample would take if placed at the background group distance instead of at the M101 distance. This then creates an interesting contrast to the other galaxies in that part of parameter space; we would add 18 gas-poor galaxies to a region that primarily consists of gas-rich galaxies. The implication is that the NGC 5485 group has stripped them of their gas. While environment may play a strong role in influencing the gas content of these dwarf galaxies, a detailed comparison of the individual galaxies in this region is beyond the scope of this work. We now compare our sample to the Satellites Around Galactic Analogs (SAGA) survey conducted by Geha et al.(2017). They characterized the satellite populations around Milky Way analogs by initially searching for satellite candidates within 300 kpc of their hosts in SDSS DR12 images. They then obtained follow-up spectroscopic Hα observations to measure redshifts and confirm the association with their hosts. They have initially presented results around 8 Milky Way analog galaxies and intend to expand this to 100 hosts with distances between 20 and 40 Mpc. Due to the large distances to these hosts, their search for satellite populations will be intrinsi- cally biased towards brighter objects. Their sample of galaxies have absolute V-band magnitudes in the range 10.8 < MV < 19.8. − − One of their key findings is that a vast majority of the satellites within the virial radii of these 8 hosts are star-forming. This is in stark contrast to that of the Milky Way and to our non-detections if the satellite candidates in our sample are associated 4.3. DISCUSSION 99

with M101. We show the comparison of absolute V-band magnitude as a function of projected distance for the Milky Way, the Geha et al.(2017) sample, and our sample in Figure 4.7. We can see the clear distinction between the three samples and the depth that they reach in absolute magnitude. The two galaxies in the Geha et al. (2017) sample that are not star-forming are shown as green circles instead of stars. The fainter of the two falls below the completeness limit of their sample. However, we include it here to show that it is similar to several of our galaxies and a few of the Milky Way’s. The two blue stars near the top-left of the figure are the LMC and the SMC which fit into the Geha et al.(2017) sample quite well. The stellar-HI mass relations in Bradford et al.(2015) would suggest that these brighter, more massive satellites in the Geha et al.(2017) sample have larger HI masses, as well. Of their 8 hosts, only 3 have four or more satellites that are star-forming. The remaining 5 have one to two star-forming satellites within their virial radii. This would suggest that the Milky Way is not particularly different from these other galaxies in terms of the number of bright satellites within their virial radii. By contrast, the M101 satellite candidates that we probe here are in a different luminosity regime from those probed by Geha et al.(2017). There is therefore no tension between their result that most of their galaxies are star-forming and our HI non-detections 4.3. DISCUSSION 100

9 ) ⊙

,M 8 HI M

log( 7

6 Bradford+ 15 Huang+ 12 This work 5 5 6 7 8 9 10 11 log(Mstellar,M ) ⊙

Figure 4.6: HI mass as a function of stellar mass for the samples of Bradford et al. (2015) (magenta), Huang et al.(2012) (green), and this work (red). Stars indicate objects with HI masses while circles with downward arrows show HI mass upper limits. The four galaxies from our sample that are conﬁrmed satellites of M101 are enclosed by a black square. The black dotted line shows the ﬁt to the Bradford et al. (2015) sample. The blue dash-dotted line shows the mass-dependent threshold below which galaxies are considered gas-depleted in the Bradford et al.(2015) sample. The red arrow illustrates how these objects would shift in this diagram if they were at the distance of the NGC5485 background group instead of at the M101 distance. 4.3. DISCUSSION 101

20 −

15 − V

M 10 −

Geha+ 17 5 − MW M101 candidate M101 conﬁrmed Gas-poor Gas-rich 0 102 103 Dproj(kpc)

Figure 4.7: Comparison of luminosity versus separation between the satellite populations of the Milky Way (blue), M101 satellite candidates (red), and those in the Geha et al.(2017) sample (green). The circles with boxes around them indicate the conﬁrmed satellites of M101. Circles indicate objects that are gas-poor and stars indicate those that are gas-rich. The vertical dashed line shows the virial radius of the Milky Way.

4.3.2 Galaxies in the Background

In our observations, we ﬁnd HI emission in three galaxies that places them in the background of M101. Two of them, dwB and dwM3, have velocities consistent with the NGC5485 background group while the other, Dw26, has a velocity that places 4.3. DISCUSSION 102

it much farther in the background. We list the properties of these three galaxies in Table 4.3. Two more galaxies, DF5 and Dw6, were not resolved into stars by HST and are not detected in HI. This places a minimum distance limit on these two galaxies of DHST = 17.5 Mpc (Merritt et al., 2016). Merritt et al.(2016) suggest that DF5, along with three other galaxies, are likely ultra-diﬀuse galaxies (UDG) associated with the background consisting of the two massive ellipticals NGC 5485 and NGC 5473. DF5 has a projected separation from

NGC 5485 of Dproj 350 kpc. Dw6 also lies in the background of M101 with a ∼ projected separation from NGC 5485 of Dproj 460 kpc. Its eﬀective radius at this ∼ distance, reff = 1.13 kpc, would fall below the UDG threshold (van Dokkum et al., 2015a). Should DF5 and Dw6 be members of this group, their HI non-detections are consistent with the possibility that they have been stripped of their gas by a hot intragroup medium. The extent of this group is made clear when we consider two galaxies in our sample with HI emission consistent with the velocity of the group. dwB and dwM3

−1 have similar line-of-sight velocities, Vsys,dwB = 1913 4 kms and Vsys,dwM3 = 1904 ± ± 2 kms−1. However, they have vastly diﬀerent projected separations from NGC 5485

Dproj,dwB 130 kpc and Dproj,dwM3 950 kpc. Despite the large projected separation ∼ ∼ of dwM3, its systemic velocity suggests that it is likely associated with NGC 5485 and this background group. The case is much more obvious for dwB due to its similar systemic velocity as well as its relatively small spatial separation. Our final confirmed background galaxy is Dw26. This is only galaxy in the Bennet et al.(2017) sample with significant UV emission seen in GALEX images. The UV 4.4. CONCLUSIONS 103

emission suggests that this galaxy is star forming or has undergone recent star formation which implies that a gas reservoir suﬃcient to fuel star formation is present. Our observations ﬁnd HI emission along the line-of-sight of Dw26 at a distance of

DDw26 156 Mpc, well behind M101 or NGC 5485. This shows that interlopers in ∼ the ﬁeld can also contaminate satellite galaxy samples.

4.4 Conclusions

We have shown the results of over 60 hours of GBT HI observations for 27 satellite candidates found around M101. We present the results of our observations showing that 1 galaxy (Dw26) lies far in the background, D 150 Mpc, 2 galaxies (dwB and ∼ dwM3) are likely associated with background group around NGC 5485, and 1 galaxy (dwM1) is likely associated with M101. The remainder of our observations were non- detections implying that these objects are gas-poor no matter their distance along the line-of-sight. 2 of these non-detections have published or preliminary (DF5 and Dw6, respectively) HST image analysis confirming that they lie in the background of M101. Of the remaining 21 galaxies, 3 have already been confirmed as M101 satellites in an independent study while nine others have upcoming HST observations to measure their distances. We assumed that our satellites with non-detections are associated with M101 for our discussion. We compared our sample of satellites around M101 to existing work on dwarf galaxies in isolation and around massive hosts. We find that our satellites probe a stellar and HI mass region that has not been thoroughly studied before. The dearth of HI in our sample is also in contrast to results from other studies that find satellites are either gas-rich or star-forming, suggesting that different mechanisms may be at play in this low-mass regime. While kinematic measurements 4.4. CONCLUSIONS 104

constrain the likelihood of a physical association between satellite candidate and host, accurate distances are required to determine if the satellites fall within the virial radius of the latter. Chapter 5

The Satellite Populations of Local Volume Galaxies

5.1 Galaxies in the Local Volume

Detailed studies of satellites around the Milky Way and M31 have revealed several obstacles to our understanding of galaxy formation and evolution; i.e. “the missing satellites” and “too big to fail” problems (see 1.2). To properly understand the satel- § lite populations in the Local Group, we need to compare them to those around other systems as we expect there to be halo-to-halo scatter (i.e cosmic scatter; Garrison- Kimmel et al., 2014). We discussed several optical imaging campaigns that reported discoveries of satellite candidates around the Milky Way-like galaxy M101 in the previous chapter. While studying the satellite population of one other galaxy is beneﬁcial, it does not en- able statistical satellite population studies around a variety of hosts. We expect that

105 5.1. GALAXIES IN THE LOCAL VOLUME 106

other Milky Way-like galaxies will reside in halos that are not exactly alike (Garrison- Kimmel et al., 2014). The number of subhalos and their distribution would then vary as well. This is another reason for studying the satellite populations of as many Milky Way-like galaxies as possible. In 4.3.1, we also discussed the SAGA (Satellite Around Galactic Analogs) survey § conducted by Geha et al.(2017). Using SDSS DR12 photometric imaging, they search for satellites galaxies within 300 kpc of 8 Milky Way-like hosts galaxies located at distances between 20 and 40 Mpc. To conﬁrm their ﬁndings in the SDSS data, they obtained spectroscopic observations and made use of existing spectroscopic surveys.

They ﬁnd 14 new satellites which satisfy their spectroscopic completeness limit, mr,0 < 20.75. This is in addition to 13 known satellites around these hosts. A key point to recall from our previous discussion of this work is that the majority of these satellites are brighter than those around the Milky Way, excluding the LMC and SMC. Low surface brightness (LSB) features were found projected near other galaxies in the Local Volume, most of which are Milky Way-like, in separate optical imaging campaigns by Karachentsev et al.(2015), hereafter K15, and Javanmardi et al.(2016), hereafter J16. The observations conducted by K15 were sensitive out to a distance of 10 Mpc, whereas, the J16 search probed slightly further to 16 Mpc. In addition ∼ ∼ to their discoveries we have already mentioned around M101, K15 (J16) ﬁnd 17 (14) LSBs/satellite candidates around 8 (5) hosts. K15 obtained further spectroscopic (Hα) observations for the two satellite candidates around the edge-on galaxy NGC 891. Neither object shows any signs of Hα emission despite the presence of faint UV emission through GALEX imaging. The 5.1. GALAXIES IN THE LOCAL VOLUME 107

Table 5.1: Properties of Local Volume Hosts in our sample 1 Name RA Dec Ref. D Nsat. (hh:mm:ss) (dd:mm:ss) (Mpc) NGC 672 01:47:53.2 27:03:01 K15 7.2 2 NGC 891 02:22:32.8 42:20:48 K15 10.0 1 NGC 1156 02:59:42.4 25:14:15 K15 7.8 2 NGC 3344 10:43:30.2 24:55:25 K15 9.8 1 NGC 4625 12:41:52.7 41:16:26 K15 7.9 1 NGC 2683 08:52:41.3 33:25:18 J16 10.5 4 NGC 4594 12:39:59.4 -11:37:23 J16 11.1 2 NGC 7814 00:031:4.9 16:08:44 J16 16.4 2 1Number of satellites around a host included in our sample brighter of these two objects were conﬁrmed as a satellite of NGC 891 through ab- sorption line measurements. Our goal is to follow up these satellite candidates and probe a population similar to the majority of the satellites that we see in the Local Group by searching for HI reservoirs. As explained in 3 and 4, a gas-rich satellite would be detected by § § our observations and provide a spectroscopic distance to the object which can be used to determine its association with its putative host. For HI non-detections, we set stringent upper limits on the gas richness of the object no matter its distance along the line of sight. For our sample, we select candidates from K15 and J16 with

6 mV < 19.5 mag which corresponds to LV & 10 L at the distances of their potential hosts. We do not include galaxies that have already been spectroscopically conﬁrmed. The details of the hosts in our sample are presented in Table 5.1. These observations of a broader sample of hosts should provide a clearer picture of the properties of faint satellites galaxy populations. 5.2. OBSERVATIONS AND ANALYSIS 108

5.2 Observations and Analysis

We conducted a total of 15 hours of observations for 15 satellite candidates with the ∼ GBT (PI: Spekkens, program AGBT16B-046) on multiple nights from August 2016 to January 2017. We use the instrumental setup described in 2.1.3 with VEGAS § in Mode 7. Recall that Mode 7 allows us to search for HI emission at recessional velocities ranging from 7000 kms−1 to 14000 kms−1. ∼ − ∼ We determine integration times for these targets using the same assumptions as for the M101 satellites (see 4.2). The targets in this sample span a large range in § brightnesses and as a result were observed with a variety of integration times. We list the details of these targets and their observations in Table 5.2. When targeting 7 of the satellite targets the GBT beam overlapped with the HI disk of the host galaxies. This occurred at different distances from the center of the GBT beam which results in different magnitudes of detection in our spectra (see 2.1). These § targets are indicated with a dagger in Table 5.2. NGC3344-dw1 was not observed for the required integration time and, therefore, we did not reach the required sensitivity to detect any potential HI emission along its line-of-sight. 14 of our satellite candidates show no sign of HI emission associated with them in their spectra. We show spectra for our non-detections in Figure 5.1 and list their upper limits in Table 5.3. The upper limits for the satellite candidates affected by the HI disk of their host are valid only outside of the contaminated velocity range. The spectrum for our only HI detection, NGC2683-DGSAT1, is shown in Figure 5.2 and the derived properties are shown in Table 5.4, following the methods described in 3.2.2. § 5.2. OBSERVATIONS AND ANALYSIS 109 ) d ( 38 84 60 77 57 21 58 43 26 48 30 03 65 56 57 V ...... ∆ 0 0 0 0 0 1 0 3 1 0 0 2 0 0 0 σ ) (mJy) ) 1 c − ( V ∆ (km s ) b ( -Apparent V-band magnitude, ) a ( - Velocity resolution of final spectrum, ) c ( ON+OFF Obs. Time ) a ( V m - GBT beam overlap with HI disk of Host, † 00:03:24.0 16:11:13.6 J15 18.2 0.83 25 00:03:06.8 16:18:33.3 J15 18.4 1.08 25 01:47:19.10 27:15:16.00 K15 18.8 1.18 25 03:00:18.20 25:14:56.00 K15 18.0 0.19 25 03:00:28.00 25:18:17.00 K15 18.4 0.76 25 10:42:44.00 25:01:30.00 K15 19.1 2.15 25 12:42:11.00 41:15:10.00 K15 17.9 0.62 25 (hh:mm:ss) (dd:mm:ss) (mag) (Hours) Table 5.2: Local Volume satellite candidates and their observations † † † , † † † ∗ † - RMS noise in spectrum at velocity resolution - Total observation time (including ON and OFF positions), ) ) b d Name RA Dec Ref. * - incomplete( observations, ( N4625-A N891-dwB 02:22:54.70 42:42:45.00 K15 18.0 0.93 25 N672-dwC 01:47:20.40 27:43:24.00 K15 17.7 0.92 25 N672-dwA N1156-dw1 N1156-dw2 N3344-dw1 N2683-DGSAT1 08:52:47.8 33:47:33.1 J15 14.6 0.1 5 N2683-DGSAT2 08:55:23.3 33:33:32.4 J15 16.2 0.15 25 N2683-DGSAT3 08:55:10.9 33:36:45.7 J15 18.9 2.27 25 N2683-DGSAT4 08:54:20.0 33:14:49.1 J15 19.1 2.16 25 N4594-DGSAT1 12:39:55.1 -11:44:38.4 J15 16.2 0.18 25 N4594-DGSAT3 12:39:32.8 -11:13:38.5 J15 18.0 0.56 25 N7814-DGSAT1 N7814-DGSAT2 5.2. OBSERVATIONS AND ANALYSIS 110

0.005 (a) N672-dwA

0.005 − 0.005 (b) N672-dwC

0.005 − 0.005 (d) N891-dwB

0.005 − 0.005 (c) N1156-dw1

0.005 − Flux (Jy) 0.005 (e) N1156-dw2

0.005 − 0.005 (f) N2683-DGSAT2

0.005 − 0.005 (g) N2683-DGSAT3

0.005 − 1000 0 1000 2000 3000 4000 5000 − VHelio(km/s)

Figure 5.1: GBT HI spectra of Local Volume satellite candidates with no HI emission attributable to the satellite candidate along the line of sight. Black dashed lines indicate channels that have had RFI removed in > 15% of scans. Blue dash-dotted lines show emission from the putative host galaxy. Emission from the Milky Way is between 150 kms−1 and 150 kms−1. − 5.2. OBSERVATIONS AND ANALYSIS 111

0.005 (h) N2683-DGSAT4

0.005 − 0.005 (i) N3344-dw1

0.005 − 0.005 (j) N4625-A

0.005 − 0.007 (k) N4594-DGSAT1

0.007

Flux (Jy) − 0.005 (l) N4594-DGSAT3

0.005 − 0.005 (m) N7814-DGSAT1

0.005 − 0.005 (n) N7814-DGSAT2

0.005 − 1000 0 1000 2000 3000 4000 5000 − VHelio(km/s) Figure 5.1 continued 5.2. OBSERVATIONS AND ANALYSIS 112

Table 5.3: Upper limits for satellite candidates with HI non-detections calculated at the distance of their putative host MHI log (MHI ) log (Lv) Name LV

M log (M ) log (L ) L N672-dwA † < 6.03 6.12 < 0.81 N672-dwC < 6.11 6.56 < 0.35 N891-dwB < 6.68 6.73 < 0.88 N1156-dw1 † < 6.15 6.51 < 0.43 N1156-dw2 † < 6.00 6.35 < 0.45 N3344-dw1 ∗,† < 6.55 6.27 < 1.90 N4625-A † < 5.95 6.56 < 0.24 N2683-DGSAT2 < 6.61 7.49 < 0.13 N2683-DGSAT3 < 6.20 6.41 < 0.61 N2683-DGSAT4 < 5.99 6.33 < 0.46 N4594-DGSAT1 < 6.82 7.54 < 0.19 N4594-DGSAT3 < 6.37 6.82 < 0.35 N7814-DGSAT1 † < 6.31 7.08 < 0.17 N7814-DGSAT2 † < 6.66 7.00 < 0.45 * - incomplete observations † - GBT beam overlap with HI disk of host 5.2. OBSERVATIONS AND ANALYSIS 113

V 44 HI . L L M M

L 8 36 0 V,D . 10 × 2 1 . 2 ± DL 5 . (Mpc) 8 10

. N2683-DGSAT1 1 M HI 7 ± M 10 0 . × s) 68 6 . S / N (km 1 1 31 ,c − Helio ± 50 V W km s 8 14 . 0 1 − ± sys 1 V . km s 463 Table 5.4: Properties of the satellite with HI emission 200 300 400 500 600 700

14 12 10 08 06 04 02 00 ......

0 0 0 0 0 0 0 0 lx(Jy) Flux Name Figure 5.2: GBT spectrum along the line of sight to NGC2683-DGSAT1. N2683-DGSAT1 5.3. DISCUSSION AND CONCLUSIONS 114

5.3 Discussion and Conclusions

The goal of this work was to follow up satellite candidates discovered around galaxies in the Local Volume, most of which are Milky Way-like. To accomplish this, we employed the same strategies as in the previous chapters of this thesis. We search for HI emission along the lines-of-sight to these satellite candidates, with a detection providing a spectroscopic distance and a non-detection constraining the gas content of the satellite. An exception to this is for 7 out of 15 targets HI emission from the host galaxy overlaps with GBT beam. We are essentially blind at the velocities corresponding to the host galaxy emission. Notwithstanding this, we can still set upper limits on all of our non-detections and conﬁrm that they are not gas-rich ﬁeld galaxies in the background. The following discussion assumes that our non-detections are associated with their host galaxies. All of our satellite candidates project well within the virial radius of a Milky Way mass halo, rvir 300 kpc. Our upper limits on of the HI masses of these satellites ∼ allow us to compare trends in these satellite populations to those we see in the Milky

MHI Way. In particular, we calculate LV and plot them as a function of projected distance from their potential host galaxy in Figure 5.3. Most of these galaxies have upper

MHI = 1 M limits that exclude LV L , with the exception of one which has incomplete

MHI observations. This same trend in LV is seen with the Milky Way satellites. The Hα non-detections, a tracer of star formation, in the satellite candidates around N891 (Karachentsev et al., 2015) are consistent with the lack of gas we ﬁnd in our observations.

−1 The systemic velocity of NGC2683-DGSAT1, VN2683−DG1 = 463 kms , and its

projected separation from NGC 2683, Dproj = 68 kpc, provide a strong case that it 5.3. DISCUSSION AND CONCLUSIONS 115

is a satellite of the latter. Taking a closer look at our gas-rich satellite galaxy we see it lies among the gas-poor objects in Figure 5.3. Its optical properties place it

8 at one extreme of our sample; with LV = 1.36 10 L and reff = 80.9” or 4.12 kpc × at the distance of NGC 2683. This makes it brighter and larger than most of the satellites and dwarf galaxies in the Local Group. In fact, it falls into the category of

2 ultra-diffuse galaxies, µg,0 > 24 mag arcsec and reff 1.5 kpc (van Dokkum et al., ≥ 2015a). Two satellites in the Local Group would also fall into this category: Milky Way dSph, Sagittarius (Majewski et al., 2003), and the M31 dSph, And XIX (Mc- Connachie et al., 2008). The former is an interacting satellite of the Milky Way and its extended morphology is likely due to tidal influences (e.g. Majewski et al., 2003; Martínez-Delgado et al., 2004). And XIX was recently determined to be more extended than previous estimates, reff 3 kpc (Martin et al., 2016). This discovery ∼ makes it quite different from the existing satellite population of the Local Group. With the confirmation of NGC2683-DGSAT1, some of the tension in the existence of And XIX may be alleviated. In their discovery paper, Javanmardi et al.(2016) report that DGSAT1 shows signs of tidal disruption which would explain its large

MHI reff . This might explain the lower than expected LV (see Figure 5.3) and coupled with its proximity suggests that it may be in the process of being stripped. We can compare the properties of these satellite candidates to those of other galaxies in a similar manner to that adopted for the M101 satellite candidates in the previous chapter (see 4.3). The upper limits on HI mass of the Local Volume satellite § candidates are shown as a function of stellar mass along with the Huang et al.(2012) and Bradford et al.(2015) samples in Figure 5.4. These satellites are more scattered in stellar mass when compared to the sample of satellites around M101. A larger 5.3. DISCUSSION AND CONCLUSIONS 116

Candidate Sats. 2.0 Gas-rich Incomplete Obs.

) 1.5 ⊙ /L ⊙ M ( V 1.0 /L HI M

0.5

101 102 Dproj(kpc)

Figure 5.3: HI mass to V-band luminosity ratio upper limits of satellite candidates as a function of projected distance from their putative host galaxy. Red circles with downward arrows indicate HI mass upper limits on satellite candidates with complete observations. The magenta triangle with the downward arrow shows the HI mass upper limit for NGC3344-dw1 using incomplete observations. The green star shows MHI = the gas-rich satellite, NGC2683-DGSAT1. The horizontal dashed line shows LV M 1 rvir = 300 kpc L and the vertical dash-dotted line shows . 5.3. DISCUSSION AND CONCLUSIONS 117

fraction of the satellites in this sample fall into the gas-depleted regime than those around M101. Some of our satellite candidates have HI mass upper limits similar to the measured masses of gas-rich dwarfs. Others have gas-rich counterparts with similar stellar masses. Our sample would add a new population of gas-poor objects to the regime of stellar and HI masses that have already been probed. Placing these objects at larger distances would push them up and to the right in Figure 5.4. If these galaxies were to lie in the background, they would be at odds with the results of Huang et al.(2012) and Bradford et al.(2015). We have presented the results from our observations of 15 satellite candidates found in projection around 8 Local Volume galaxies. All but one of the candidates show no HI emission likely to be associated with them in their spectra. Seven of the spectra with non-detections contain HI emission from their host galaxy. This makes it difficult to definitively determine the gas content of these satellites near the velocities of their hosts. However, we set upper limits on the HI mass of these satellites for the remainder of the velocity space in the spectra, thus confirming that they are not gas rich background field galaxies. The detection we report belongs to a satellite which falls into the UDG category, similar to And XIX around M31. This discovery can alleviate the tension created by such a large satellite around M31. Further observations (e.g. HST, Danieli et al., 2017) of these satellite candidates can definitively confirm whether they are associated with their putative hosts or lie in the background. 5.3. DISCUSSION AND CONCLUSIONS 118

9 ) ⊙ M )(

HI 8 M log( 7

6 Bradford+ 15 Huang+ 12 LV sats. 6 7 8 9 10 11 log(Mstellar)

Figure 5.4: HI mass as a function of stellar mass for samples of galaxies in Bradford et al.(2015) (magenta), Huang et al.(2012) (green), and this work (blue). Stars indicate objects with measured HI masses while circles with downward arrows show upper limits. The black dotted line shows the ﬁt to the Bradford et al.(2015) sample. The blue dash-dotted line shows the mass-dependent threshold below which galaxies are considered gas-depleted in the Bradford et al.(2015) sample. Chapter 6

Discussion

Throughout this work we have presented HI observations along the lines-of-sight to low-mass satellite galaxy candidates. Our goal was to use their HI content as a tracer of their association with their putative hosts. The spectra of gas-rich galaxies allow spectroscopic redshift estimates that can be used to confirm an association between the candidate and the putative host (e.g. 4.3). On the other hand, a non- § detection implies a gas-poor galaxy somewhere along the line of sight. It is possible that the galaxies lie close to their massive host and have had their gas removed by environmental processes (e.g. 1.4.2). It is also possible that there exists a population § of gas-poor field galaxies that has not yet been explored in the literature (e.g. 1.4.3). § Our total sample of 48 galaxies contains 6 galaxies with HI emission and 42 galaxies with no sign of HI emission. The implications of the detections and non-detections around their individual hosts (i.e. NGC 3109 in Chapter3, M101 in Chapter4, Local Volume galaxies in Chapter5) were discussed in the previous chapters. In this chapter, we aim to place these results into a broader context and, in particular, the implications of our non-detections for the gas content of nearby field dwarfs.

119 6.1. OUR NON-DETECTIONS AS SATELLITES 120

We separate the discussion into two main sections. Section 6.1 considers our non- detections as satellites of the nearest massive host and Section 6.2 considers them as background galaxies.

6.1 Our Non-Detections as Satellites

We ﬁrst treat our non-detections as satellites and place them at the distances of their hosts. This is consistent with the environmental dependence of the gas richness of satellite galaxies around the Milky Way and M31 ( 1.4.2), since the vast majority of § our candidates project within the virial radii of their putative hosts. In our comparison of the M101 satellite candidates and the SAGA survey sample (Geha et al., 2017), we identify a fairly clear separation in luminosity ( 4.3.1). This § eases the tension created by the large majority (26/27) of star-forming (i.e. gas-rich) SAGA satellites compared to our non-detections. We arrive at a similar conclusion if we expand our comparison to include our Local Volume sample and the one detection found near NGC 3113 and ESO 435 -G 039 (Figure 6.1). The galaxies in our Local Volume sample are slightly brighter than those around M101 and still fainter than the majority of the SAGA sample. The single HI detection in the Local Volume sample agrees well with the SAGA sample due to its high luminosity and presumably larger stellar mass. Our detection in the background of NGC 3109 ( 3.3) is represented § by the magenta star. If it is associated with either of the galaxies we identiﬁed as potential hosts, then it would fall at a similar luminosity as the majority of our total sample. The Milky Way satellites (McConnachie, 2012) are much more diverse in their luminosities in comparison to our sample and the SAGA sample. So far, our ability to probe to such faint luminosities is limited to the Local Group. 6.1. OUR NON-DETECTIONS AS SATELLITES 121

20 −

15 − V

M 10 −

Geha+ 17 Milky Way M101 5 Local Volume − N3113/ESO435-39 Conﬁrmed Gas-poor Gas-rich 102 103 Dproj(kpc)

Figure 6.1: Comparison of luminosity versus host separation between the satellite populations of the Milky Way (blue), M101 (red), our Local Volume Galaxy sample (cyan), and those in the Geha et al.(2017) sample (green). Our detection of Dw1 in the background of NGC 3109 is shown as a magenta star and Dproj deﬁned by its potential hosts. Circles indicate objects that are gas-poor and stars indicate those that are gas-rich. Symbols enclosed by boxes indicate conﬁrmed satellites. The vertical dashed line shows the virial radius of the Milky Way.

Due to scatter in the subhalo distributions of halos with similar masses (Garrison- Kimmel et al., 2014), we would expect diﬀerent halos to have diﬀerent satellite luminosity functions (see 1.1). Our Local Volume sample consists of satellite candidates § 6.1. OUR NON-DETECTIONS AS SATELLITES 122

around multiple hosts ( 5.1), whereas, the M101 sample ( 4.1) has many candidates § § around a single host. We follow the work of Danieli et al.(2017) by producing and comparing the luminosity functions for the Milky Way, M31, and our Local Volume hosts. This comparison is most useful for hosts with several satellite candidates, so we focus on NGC 2683 and M101. NGC 2683 has one confirmed satellite (DGSAT1) and three satellite candidates with HI non-detections. For completeness, we also include a faint LSB that we did not target in the luminosity function. Our M101 sample has 21 satellite candidates with HI non-detections and has one confirmed satellite (dwM1). We show the luminosity functions in Figure 6.2. We show two luminosity functions for M101: one includes our new detection (cyan squares) as well as those confirmed by Danieli et al.(2017). The other includes satellites and satellite candidates (red squares) and thus assumes that the non-detections are gas-poor satellites. The NGC 2683 luminosity function has been volume-corrected to account for the fraction of its virial volume that was not surveyed (Javanmardi et al., 2016). By and large, the luminosity functions are consistent with one another. The one exception is in the M101 total luminosity function where there is a drastically steeper slope at MV > 11 due − to the contribution of the faint satellite candidates in our sample. The emergence of the drastic steepening of the slope when all candidates are assumed to be satellites, as well as ongoing HST follow-up observations which imply that at least some candidates that were initially identified are in fact in the background (Merritt et al., 2016; P. Bennet, private communication), suggest that our HI non-detections of LSB features near M101 do not necessarily imply that they are satellites of the latter. We explore the possibility that a significant fraction of our non-detections are background galaxies in the next section. 6.1. OUR NON-DETECTIONS AS SATELLITES 123

102 Milky Way Andromeda M101 Total M101 Conﬁrmed NGC 2683 ) V 101 < M ( N

100 0 5 10 15 20 25 − − MV − − −

Figure 6.2: Cumulative luminosity functions for satellites around the Milky Way (blue triangles, McConnachie, 2012; Bechtol et al., 2015; Drlica-Wagner et al., 2015; Torrealba et al., 2016a), M31 (green, downward triangles, McConnachie, 2012), M101 conﬁrmed satellites (cyan squares), M101 candidates and conﬁrmed satellites (red squares), and NGC 2683 (magenta circles). We volume-correct the NGC 2683 sample. 6.2. OUR NON-DETECTIONS AS BACKGROUND GALAXIES 124

6.2 Our Non-detections as Background Galaxies

Our HI non-detections have two implications for the satellite candidates along their lines-of-sight: 1. they are likely gas-poor since they fall below our stringent sensitivity limits and 2. they can lie anywhere along the line-of-sight within our bandpass. If these galaxies do lie in the background, then they are gas-poor galaxies that may reside in the ﬁeld or a relatively isolated environment. As we start to place these galaxies at further distances in the background, we expect that they will start to probe higher stellar and HI masses in the dwarf galaxy regime. Therefore, there is a range of distances at which our non-detections are inconsistent with previous work on the gas scaling relations in ﬁeld galaxies (Huang et al., 2012; Bradford et al., 2015) Bradford et al.(2015) provide relations between HI and stellar mass for their sample of optically-selected isolated and non-isolated low-mass galaxies. To check our non-detections in the context of these relations, we estimate HI mass upper limits and stellar masses for the corresponding satellite candidates as a function of distance. Recall that HI mass scales as the square of the distance (Equation 1.3). We assume

Mstellar = 1 M LV L to estimate stellar masses (e.g. Martin et al., 2008). In Figure 6.3, we show the residuals (actual MHI MHI from relation) from the Bradford et al.(2015) − relation to the median HI upper limit and stellar mass of the galaxies in each of our subsamples. We also include the residuals for the ALFALFA dwarf galaxies (Huang et al., 2012) which probe lower stellar masses. Recall that, the Bradford et al.(2015) were selected optically, whereas, the Huang et al.(2012) were ﬁrst detected in HI. Our HI mass estimates are upper-limits and it follows that the residuals are as well, as indicated with downward arrows in Figure 6.3. Upper limits are calculated at 10 Mpc intervals out to the end of our bandpass 6.2. OUR NON-DETECTIONS AS BACKGROUND GALAXIES 125

coverage, D 200 Mpc. The NGC 3109 galaxies are fainter in apparent magnitude ∼ than those in the M101 sample. However, the lack of resolved stars in the preliminary analysis of HST data sets a lower limit on their distance that is larger than the M101 distance. Similarly, the Local Volume satellites are located at larger distances and are brighter than those in the M101 sample. This explains the higher starting stellar mass value for both samples.

At stellar masses below log(Mstellar) 8, corresponding to distances lower than ∼ 40 to 60 Mpc for our non-detections, all of our samples lie in a region that is not occupied by the ALFALFA dwarf galaxies or the Bradford et al.(2015) sample. This implies that there is a diversity in the properties in this mass regime should our galaxies be located at these distances. However, the diﬀerent sample masses in this case would imply that our non-detections and the gas-richness of the Bradford et al. (2015) sample are not inconsistent. On the other hand, placing these galaxies at distances greater than 40 to 60 Mpc begins to create some tension with the gas richness of ﬁeld dwarfs in other samples.

7 9 Geha et al.(2012) find that quiescent field dwarf galaxies 10 M < Mstellar < 10 M are scarce. Star formation in galaxies implies the existence of some form of a gas reservoir. By extension, their findings imply that field dwarfs are gas-rich consistent with the findings of Bradford et al.(2015). If we place our galaxies at distances of 15 Mpc for the Local Volume, 25 Mpc for the M101 sample, and 38 Mpc ∼ ∼ ∼ 7 for the NGC 3109 sample then they reach stellar masses Mstellar 10 M . All of ∼ our galaxies would fall into the mass regime probed by Geha et al.(2012), implying that our non-detections are inconsistent with their low fraction of quenched dwarfs. We note that if the Local Volume sample galaxies have, on average, distances greater 6.2. OUR NON-DETECTIONS AS BACKGROUND GALAXIES 126

9 than 150 Mpc, then their stellar masses would exceed Mstellar 10 M and therefore ∼ they would fall out of the range considered by Geha et al.(2012). The relative sizes of the samples being compared here should also be discussed. In the mass range where quiescent field dwarfs are rare, the Geha et al.(2012) sample contains 2500 galaxies in total. There are 105 galaxies at the low-mass end ∼ 7 9 (Mstellar 10 M ) and 900 galaxies at the high-mass end (Mstellar 10 M ). If ∼ ∼ ∼ we combined our total sample of 42 gas-poor galaxies with those in the lowest stellar mass bin, we would drastically increase the fraction of quiescent galaxies from 0 to 0.3. This assumes that all of our galaxies lie at the minimum distances listed above ∼ (15, 25, and 38 Mpc depending on the sample) The impact of adding our sample to the highest stellar mass bin (under the assumption that they all lie at 150 Mpc) would be small but not negligible. This is due to the large number of field galaxies in this bin as well as the fact that only one of our samples (Local Volume, 14 non-detections) reaches this mass range. Combining the galaxies in this mass range from the two samples, we increase the quiescent fraction in the highest stellar mass bin from 0 to 0.015. While the overall fraction of quiescent galaxies at high masses ∼ would not change with the addition of our sample to that of Geha et al.(2012), one still must explain why the relative fractions of quiescent, gas-poor systems within each individual sample are so different. 6.2. OUR NON-DETECTIONS AS BACKGROUND GALAXIES 127

Residual (dex) −

Local Volume 2 − N3109 M101 Bradford+ 15 Huang+ 12 3 − 5 6 7 8 9 10 11 log(Mstellar)(M ) ⊙

Figure 6.3: Residuals from the ﬁt between log (MHI ) and log (Mstellar) for the Brad- ford et al.(2015) sample. We show the median value for each of our samples as the assumed distance to the non-detections is increased. Starting at the coloured squares, the galaxy’s stellar mass and ﬁt residual upper limit are shown in 10 Mpc intervals. ∼ Also, included are the residuals for the Bradford et al.(2015), magenta stars, and Huang et al.(2012), green stars, samples.

The increase in the implied HI upper limits and stellar masses of our galaxies with distance is crucial to understanding how they compare with other populations of low-mass galaxies. However, as we invoke larger masses we also invoke larger physical sizes. We previously showed that our our sample probes a lower stellar mass 6.2. OUR NON-DETECTIONS AS BACKGROUND GALAXIES 128

6 7.5 range nearby (10 < Mstellar < 10 M , D < 40 Mpc) and higher stellar masses at

7.5 9 larger distances (10 < Mstellar < 10 M , 40 Mpc < D < 150 Mpc). We can perform

a similar estimate for how the luminosities (absolute V band magnitude,MV ) and − sizes (eﬀective radius,Reff ) as we place them at larger distances as in Figure 6.4. To calculate their luminosities and sizes, we use their apparent V-band magnitudes and angular eﬀective radii from their discovery papers. For the galaxies in the NGC

3109 sample, we assume an effective radius of reff = 9 arcseconds, visually estimated from Figure 3.1, due to a lack of accurate photometric analysis (D. Sand, private communication). We show the median value in implied luminosity and size for the galaxies in each of our samples by adopting distances in increments of 10 Mpc out to the edge of our bandpass (D 200 Mpc). We have truncated our range in effective ∼ radius at Reff = 7 kpc as galaxies with such large sizes for their corresponding masses have never been detected (see Figure 6.3, Mihos et al., 2015). For comparison at lower distances, we include some of the brighter and larger satellites in the Local Group in Figure 6.4 (blue circles and triangles), excluding the LMC and SMC. We also show ultra-diffuse galaxies (UDGs) found in the Coma cluster (grey crosses and diamonds, van Dokkum et al., 2015a,c; Kadowaki et al., 2017) and those found in relatively isolated environments (orange circles and stars, Dalcanton et al., 1997; Bellazzini et al., 2017; Papastergis et al., 2017). Recall that UDGs have similar luminosities and stellar masses to typical dwarf galaxies but are similar to Milky Way-type galaxies in size. When our samples are at their lowest

distances, they lie in a similar region of the MV Reff parameter space as satellites − in the Local Group. As the Local Volume and M101 samples are assumed to be at

larger distances (25 . DLV (Mpc) . 75 and 40 . DM101 (Mpc) . 120) we can see 6.2. OUR NON-DETECTIONS AS BACKGROUND GALAXIES 129

that they begin to overlap with the region occupied by UDGs. When the NGC 3109 sample is placed at distances between 20 and 60 Mpc, it covers a region in size that is similar to those of the UDGs presented here, however, they do not fall within a similar luminosity range. There are several on-going eﬀorts to search for UDGs in diﬀerent environments and some may reach fainter luminosities. At a distance of ∼ 60 Mpc, the NGC 3109 sample does have a similar size and luminosity to the tidally disturbed Sag dSph (McConnachie, 2012). The galaxies in our samples can only be placed out to a certain distance before becoming unreasonably large for their stellar masses. Looking at the right-most point for each sample in Figure 6.4, we can check Figure 6.3 to infer the corresponding

8 stellar masses. Our samples reach stellar masses Mstellar (4 6) 10 M and ∼ − × eﬀective radii around 6.5 to 6.9 kpc at distances between 90 and 190 Mpc. For ∼ ∼ comparison, Papastergis et al.(2017) obtain HI measurements of relatively isolated UDGs and estimate their stellar masses. The two orange circles in Figure 6.4 represent two of the more massive and gas-poor (HI non-detections) UDGs in their sample, with

8 8 Reff 4.0 kpc and 4.5 kpc and Mstellar 2.5 10 M and 4 10 M , respectively. ∼ ∼ × × There are two scenarios that may explain this. At present, we cannot rule out the possibility that galaxies with such large eﬀective radii exist, representing a seemingly new population of extremely diﬀuse galaxies. The other, more likely scenario, is they lie at lower distances and, therefore, fall into a similar mass and size regime of known galaxies like the ones mentioned above. 6.2. OUR NON-DETECTIONS AS BACKGROUND GALAXIES 130

18 −

16 −

14 V − M

12 − Local Volume M101 N3109 Local Group 10 Isolated UDGs − Coma UDGs 0 1 2 3 4 5 6 7 Reff (kpc)

Figure 6.4: Luminosity as a function of effective radius for sample galaxies if they are in the background. The initial median absolute V-band magnitudes and effective radii for each of our samples are shown as coloured squares. The luminosities and sizes implied by increasing the assumed distance to these systems are shown by circles, each separated by 10 Mpc. We include a few of the brighter, extended Milky Way ∼ satellites (blue circles, McConnachie, 2012); particularly, Sag dSph (blue, upside-down triangle) and And XIX (blue triangle). Also shown are UDGs from the Coma Cluster as grey crosses (van Dokkum et al., 2015a) and grey diamonds (spectroscopically confirmed, van Dokkum et al., 2015c; Kadowaki et al., 2017) and relatively isolated UDGs (Dalcanton et al., 1997; Bellazzini et al., 2017) as orange circles and stars (gas-rich, Papastergis et al., 2017). 6.3. CONSTRAINING DISTANCES TO NON-DETECTIONS 131

6.3 Constraining Distances to Non-Detections

In the previous sections, we considered our non-detections as satellites of the nearest massive hosts ( 6.1) and as isolated dwarf galaxies in the field ( 6.2). In the latter § § scenario, we found that placing these objects in the background relative to their putative hosts creates tension with existing studies of isolated dwarf galaxies and UDGs if the assumed distances, and therefore stellar masses and sizes, are sufficiently large. In this concluding section, we discuss the range of distances for our non- detections that avoids these tensions. If we assume that our non-detections are in the background and are not unusual in their properties relative to the broader population of field galaxies, then these distances represent the most likely values. We compared the luminosities and sizes of our samples to UDGs, which are very faint and extended galaxies, in Figure 6.4. To determine the corresponding constraints on the distances for our samples, we take the most extended (Reff 4.2 kpc) of the ∼ UDGs as our upper limit. Our galaxies in the NGC 3109, Local Volume, and M101 samples would need to be located at distances lower than approximately 60, 75, and 115 Mpc, respectively, to fall below this size threshold. At these distances, the gas-

poor galaxies in our samples would have stellar masses of 7.3 < log (Mstellar (M )) < 8.3, which is in tension with the gas-richness of dwarfs at these masses found by existing studies (Figure 6.3). We can place further constraints on the distances to the non-detections by con- sidering the work of Geha et al.(2012) and Bradford et al.(2015). Their samples suggest that isolated dwarf galaxies should be gas-rich in the stellar mass range of

7 9 10 < Mstellar < 10 M . In order to avoid any tension with either sample, the galaxies in our samples would have to lie at distances at which their stellar masses would 6.3. CONSTRAINING DISTANCES TO NON-DETECTIONS 132

7 be below Mstellar 10 M . The distances that corresponds to this threshold are ∼ 15, 25, and 38 Mpc for the Local Volume, M101, and NGC 3109 samples, ∼ ∼ ∼ respectively. With these constraints, the non-detections in the NGC 3109 sample are potentially located between 17.5 and 38 Mpc, where the lower limit is set by HST data (Merritt et al., 2016). A similar limit can be set for our M101 satellites that were not resolved into stars by HST (DF5 and Dw6); they are potentially located between 17.5 and 25 Mpc. The other satellites in the sample are potentially located between 7 and 25 Mpc. The distance ranges for the Local Volume sample are much more stringent than for the other samples. In other words, if the satellite candidates are not associated with their putative hosts, then they need to be located relatively nearby in order to be consistent with known scaling relations between luminosity, size, and gas content for ﬁeld dwarf galaxies. Chapter 7

Conclusions and Future Work

This thesis presented Green Bank Telescope (GBT) HI observations along the lines-of- sight to 48 low surface brightness satellite candidates identiﬁed in deep optical images near 10 putative hosts in the Local Volume. Using their HI content to constrain their association with these potential hosts, we aimed to study the satellite populations around these systems (Chapters1 and2). We considered three satellite candidate subsamples clustered around the dwarf NGC 3109 (Chapter3), the Milky Way-like galaxy M101 (Chapter4), and a variety of Local Volume galaxies targeted by small telescope networks (Chapter5). Five of the six ultra faint dwarf candidates around NGC 3109 are likely to be background objects from the lack of resolved stars in their HST images. The sixth candidate (Dw1) was also conﬁrmed to be in the background through our HI observa-

−1 tions, where we detect it at Vsys = 1066 km s (Figure 3.3). The ﬁve candidates that were not resolved into stars by HST were not detected in our HI observations (Figure 3.2). We computed 5σ upper limits on the HI mass along the line of sight to these candidates; which lie in the range of log (MHI (M )) < 6.3 to log (MHI (M )) < 8.7

133 CHAPTER 7. CONCLUSIONS AND FUTURE WORK 134

depending on the assumed distances between 17.5 Mpc (i.e. the lower limit implied by the lack of resolved stars in HST imaging) and 200 Mpc (i.e. the extent of the GBT bandpass). Our non-detections therefore imply that these objects are background, gas-poor dwarf galaxies. The largest sample of satellite candidates of a single host that we consider is around M101, with 27 LSBs in a highly asymmetric distribution skewed around a background group (Chapter4). Three of these LSBs (DF1-DF3) were independently confirmed as satellites of M101 (Danieli et al., 2017). We found no HI emission associated with these galaxies which is consistent with their optical properties. Two of the LSBs, DF5 and Dw6, were observed with HST and were not resolved into stars. This sets a lower distance limit of 17.5 Mpc for these objects. Merritt et al.(2016) suggest that DF5 is an ultra-diffuse galaxy (UDG) and member of the background group. Its lack of HI as determined by our observations (Figure 4.2) is consistent with this scenario, as its gas may have been stripped away; a similar case can be made for Dw6. We detect HI emission in the spectra of 4 LSBs: dwM1, Dw26, dwB, and dwM3 (Figure 4.4). dwM1 is the only galaxy in the M101 sample that we confirm as a new satellite. From its HI emission feature, we conclude that Dw26 is a massive background galaxy at a distance of 150 Mpc. This conclusion is supported by its ∼ bright UV emission (Bennet et al., 2017), which is unusual for a satellite galaxy. We find that the two other detections, dwB and dwM3, have velocities that are consistent with the background group centered around NGC 5485. The remaining 18 LSBs were not detected in our observations and we classify them as gas-poor. We compute 5σ upper limits on HI masses for all of our non-detections (Table 4.2). The LSBs in the Local Volume sample are, on average, brighter than those around CHAPTER 7. CONCLUSIONS AND FUTURE WORK 135

NGC 3109 and M101 (Chapter5). Our observations confirm a single satellite in this sample, NGC 2683-DGSAT1 (Figure 5.2). Its optical properties are consistent with UDGs or with a tidally disturbed satellite. We believe the latter is more likely because of its low gas fraction. These two properties would suggest that DGSAT1 is on the verge of being stripped by NGC 2683. We do not find any HI emission associated with the remaining 14 satellite candidates (Figure 5.1). However, we do detect emission from the HI disk of the putative host for 7 of our candidates. We calculate 5σ upper limits on the HI mass of our satellites outside of the velocity region contaminated by the parent HI disks (Table 5.3). Our HI non-detections do not provide concrete distances for satellite candidates along their lines of sight. They can be interpreted in two ways: either they are gas- poor satellites of their putative hosts or they are gas-poor dwarfs in the background. Assuming first that they are satellites, we made comparisons with the Geha et al. (2017) SAGA results, in which the majority of satellites (26/27) that they confirm around 8 Milky Way-like hosts are star forming. This is in contrast to the trends seen in the Local Group and in this thesis. However, the satellites that they discover are brighter and, therefore, are more massive than those in our sample (Figure 6.1). Again, under the assumption that are non-detections are satellites, we compared the satellite luminosity functions of the Milky Way, M31, M101, and NGC 2683 (Figure 6.2). We find that for the most part, they are broadly consistent. The exception is when all satellite candidates of M101 are included which produces a

drastically steeper slope at MV > 11 in the cumulative luminosity function. If all of − the candidates around M101 are in fact satellites, then the faint end of its luminosity function is much steeper than that of the Milky Way or M31. 7.1. FUTURE WORK 136

If we instead assume that the non-detections correspond to gas-poor background galaxies, we determined that they must lie in certain distance ranges in order to avoid tension with existing studies of the gas content of field dwarf galaxies ( 6.2). § Unless they stem from a fundamentally different population then that probed by existing studies, the LSB features near NGC 3109 have D > 17.5 Mpc set by HST observations and D < 38 Mpc set by the gas richness of field dwarfs. The satellite candidates around M101 likely lie between 7 Mpc (the distance of M101) and 25 Mpc, the limit beyond which their non-detections become inconsistent with the quenched field dwarf fraction and their sizes imply that they are ultra-diffuse. The upper limit is near the background group which also suggests that at least some of our non- detections may be members. Since they have the highest apparent brightnesses, the LSBs around Local Volume galaxies have the most strict distance limits in order to avoid tension with the known properties of field dwarfs. We determine that they must lie at D < 15 Mpc for consistency with the gas content and size-luminosity relations of local galaxies. The hosts of these galaxies have distances between 7.2 and 16.4 Mpc. The simplest scenario that explains our non-detections is therefore that they lie at the distance of their host.

7.1 Future Work

While this thesis may have presented the HI analysis for each satellite candidate and its implications for the satellite populations of nearby galaxies, there is still more to do. 7.1. FUTURE WORK 137

NGC 3109

In this work, we have set upper limits on the HI masses of LSB features around the dwarf galaxy NGC 3109 and placed a constraint on their distances. We await the full analysis of the HST data (by D. Sand and P. Bennet) to provide more detailed optical properties. These can be used to determine whether these galaxies are UDGs in the distance ranges implied by our non-detections.

M101

3 satellite candidates around M101 (Dw19, Dw33, and Dw42) have incomplete observations which will be completed in the near future (Program AGBT17B-235, PI: Karunakaran). 9 satellite candidates around M101 that we targeted in this thesis have forthcoming HST data which, like the previous sample, will provide us with more detailed optical properties. Whether or no HST resolves these objects into individual stars will further help constrain their distances, and therefore their physical properties given our HI non-detections.

Local Volume

One of the satellite candidates that we followed up around Local Volume galaxies has incomplete observations which need to be completed in the near future. We have submitted a GBT proposal (PI: Karunakaran) to expand our sample by targeting an additional 4 LSBs around NGC 7814, 2 around NGC 3625, and 1 around NGC 3669. 7.1. FUTURE WORK 138

Broader Goals

It would be incredibly useful to obtain distance information for our non-detections. While HST/JWST observations are one possibility, obtaining redshifts from absorp- tion/emission features using an 8m-class optical spectrometer is also feasible (van Dokkum et al., 2015c). We have calculated conservative 5σ upper limits for our spectra with non-detections. However, we conducted preliminary tests on the required signiﬁcance of a signal to be distinguished from an anomalous noise peak, and found that features with a S 2.5 N ∼ would be distinguished from the noise. As our follow-up data set gets larger, we could better characterize our ability to identify HI emission peaks corresponding to dwarf galaxies and determine a more strict HI mass limit implied by a non-detection. An extension to the one above goal would be to better understand low level noise in our data. This can be explored by performing a power spectrum analysis, as well as, characterizing the distrubtion of the noise in our spectra. As the number of satellite candidates detection around Local Volume galaxies grows, follow-up observations like the ones presented here will help constrain their association to their putative hosts as well as their physical properties. Eventually, we aim to study the statistical properties of satellite populations around host galaxies of diﬀerent masses in order to quantitatively compare them with predictions from simulations. Bibliography

R. G. Abraham and P. G. van Dokkum. Ultra-Low Surface Brightness Imaging with the Dragonﬂy Telephoto Array. Publications of the Astronomical Society of Paciﬁc, 126:55, January 2014. doi: 10.1086/674875.

J. K. Adelman-McCarthy, M. A. Agüeros, S. S. Allam, K. S. J. Anderson, S. F. Anderson, J. Annis, N. A. Bahcall, C. A. L. Bailer-Jones, I. K. Baldry, J. C. Barentine, T. C. Beers, V. Belokurov, A. Berlind, M. Bernardi, M. R. Blanton, J. J. Bochanski, W. N. Boroski, D. M. Bramich, H. J. Brewington, J. Brinchmann, J. Brinkmann, and SDSS Collaboration. The Fifth Data Release of the Sloan Digital Sky Survey. Astrophysical Journal Supplement Series, 172:634–644, October 2007. doi: 10.1086/518864.

N. C. Amorisco and A. Loeb. Ultradiﬀuse galaxies: the high-spin tail of the abundant dwarf galaxy population. Monthly Notices of the Royal Astronomical Society, 459: L51–L55, June 2016. doi: 10.1093/mnrasl/slw055.

H.C. Andrews, W.K. Pratt, and K. Caspari. Computer Techniques in Image Pro- cessing. Academic Press, 1970. URL https://books.google.ca/books?id= zOFRAAAAMAAJ.

139 BIBLIOGRAPHY 140

D. G. Barnes and W. J. G. de Blok. On the Neutral Gas Content and Environment of NGC 3109 and the Antlia Dwarf Galaxy. Astronomical Journal, 122:825–829, August 2001. doi: 10.1086/321170.

K. Bechtol, A. Drlica-Wagner, E. Balbinot, A. Pieres, J. D. Simon, B. Yanny, B. San- tiago, R. H. Wechsler, J. Frieman, A. R. Walker, P. Williams, E. Rozo, E. S. Rykoﬀ, A. Queiroz, E. Luque, A. Benoit-Lévy, D. Tucker, I. Sevilla, R. A. Gruendl, L. N. da Costa, A. Fausti Neto, M. A. G. Maia, and DES Collaboration. Eight New Milky Way Companions Discovered in First-year Dark Energy Survey Data. Astrophysical Journal, 807:50, July 2015. doi: 10.1088/0004-637X/807/1/50.

P. S. Behroozi, R. H. Wechsler, and C. Conroy. The Average Star Formation Histories of Galaxies in Dark Matter Halos from z = 0-8. Astrophysical Journal, 770:57, June 2013. doi: 10.1088/0004-637X/770/1/57.

E. F. Bell, C. T. Slater, and N. F. Martin. Andromeda XXIX: A New Dwarf Spheroidal Galaxy 200 kpc from Andromeda. Astrophysical Journal Letters, 742:L15, Novem- ber 2011. doi: 10.1088/2041-8205/742/1/L15.

M. Bellazzini, V. Belokurov, L. Magrini, F. Fraternali, V. Testa, G. Beccari, A. Marchetti, and R. Carini. Redshift, metallicity and size of two extended dwarf Irregular galaxies: a link between dwarf Irregulars and ultra diﬀuse galaxies? Monthly Notices of the Royal Astronomical Society, 467:3751–3758, May 2017. doi: 10.1093/mnras/stx236.

V. Belokurov, M. J. Irwin, S. E. Koposov, N. W. Evans, E. Gonzalez-Solares, N. Met- calfe, and T. Shanks. ATLAS lifts the Cup: discovery of a new Milky Way satellite BIBLIOGRAPHY 141

in Crater. Monthly Notices of the Royal Astronomical Society, 441:2124–2133, July 2014. doi: 10.1093/mnras/stu626.

P. Bennet, D. J. Sand, D. Crnojević, K. Spekkens, D. Zaritsky, and A. Karunakaran. Discovery of diﬀuse dwarf galaxy candidates around M101. ArXiv e-prints, arXiv:1710.01728, October 2017.

S. Bird, W. E. Harris, J. P. Blakeslee, and C. Flynn. The inner halo of M 87: a ﬁrst direct view of the red-giant population. Astronomy & Astrophysics, 524:A71, December 2010. doi: 10.1051/0004-6361/201014876.

M. R. Blanton, E. Kazin, D. Muna, B. A. Weaver, and A. Price-Whelan. Improved Background Subtraction for the Sloan Digital Sky Survey Images. Astronomical Journal, 142:31, July 2011. doi: 10.1088/0004-6256/142/1/31.

T. J. Bohuski, A. P. Fairall, and D. W. Weedman. The nature of Tololo emission-line galaxies. Astrophysical Journal, 221:776–787, May 1978. doi: 10.1086/156080.

G. D. Bothun, C. D. Impey, D. F. Malin, and J. R. Mould. Discovery of a huge low-surface-brightness galaxy - A protodisk galaxy at low redshift? Astronomical Journal, 94:23–29, July 1987. doi: 10.1086/114443.

G. D. Bothun, C. D. Impey, and D. F. Malin. Extremely low surface brightness galaxies in the Fornax Cluster - Properties, stability, and luminosity ﬂuctuations. Astrophysical Journal, 376:404–423, August 1991. doi: 10.1086/170290.

L. Bottinelli, N. Durand, P. Fouque, R. Garnier, L. Gouguenheim, M. Loulergue, G. Paturel, C. Petit, and P. Teerikorpi. Observational Data for the Kinematics BIBLIOGRAPHY 142

of the Local Universe - Part Two - Second Set of Radial Velocity Measurements. Astronomy & Astrophysics Supplement, 102:57, November 1993.

M. Boylan-Kolchin, J. S. Bullock, and M. Kaplinghat. Too big to fail? The puzzling darkness of massive Milky Way subhaloes. Monthly Notices of the Royal Astro- nomical Society, 415:L40–L44, July 2011. doi: 10.1111/j.1745-3933.2011.01074.x.

J. D. Bradford, M. C. Geha, and M. R. Blanton. A Study in Blue: The Baryon Content of Isolated Low-mass Galaxies. Astrophysical Journal, 809:146, August 2015. doi: 10.1088/0004-637X/809/2/146.

A. H. Broeils and M.-H. Rhee. Short 21-cm WSRT observations of spiral and irregular galaxies. HI properties. Astronomy & Astrophysics, 324:877–887, August 1997.

J. S. Bullock and M. Boylan-Kolchin. Small-Scale Challenges to the ΛCDM Paradigm. Annual Review of Astronomy and Astrophysics, 55:343–387, August 2017. doi: 10.1146/annurev-astro-091916-055313.

E. Carlesi, J. G. Sorce, Y. Hoﬀman, S. Gottlöber, G. Yepes, N. I. Libeskind, S. V. Pilipenko, A. Knebe, H. Courtois, R. B. Tully, and M. Steinmetz. Constrained Local UniversE Simulations: a Local Group factory. Monthly Notices of the Royal Astronomical Society, 458:900–911, May 2016. doi: 10.1093/mnras/stw357.

J. L. Carlin, D. J. Sand, P. Price, B. Willman, A. Karunakaran, K. Spekkens, E. F. Bell, J. P. Brodie, D. Crnojević, D. A. Forbes, J. Hargis, E. Kirby, R. Lupton, A. H. G. Peter, A. J. Romanowsky, and J. Strader. First Results from the MAD- CASH Survey: A Faint Dwarf Galaxy Companion to the Low-mass Spiral Galaxy BIBLIOGRAPHY 143

NGC 2403 at 3.2 Mpc. Astrophysical Journal Letters, 828:L5, September 2016. doi: 10.3847/2041-8205/828/1/L5.

K. C. Chambers, E. A. Magnier, N. Metcalfe, H. A. Flewelling, M. E. Huber, C. Z. Waters, L. Denneau, P. W. Draper, D. Farrow, D. P. Finkbeiner, C. Holmberg, J. Koppenhoefer, P. A. Price, R. P. Saglia, E. F. Schlaﬂy, S. J. Smartt, W. Sweeney, R. J. Wainscoat, W. S. Burgett, T. Grav, J. N. Heasley, K. W. Hodapp, R. Jedicke, N. Kaiser, R.-P. Kudritzki, G. A. Luppino, R. H. Lupton, D. G. Monet, J. S. Morgan, P. M. Onaka, C. W. Stubbs, J. L. Tonry, E. Banados, E. F. Bell, R. Bender, E. J. Bernard, M. T. Botticella, S. Casertano, S. Chastel, W.-P. Chen, X. Chen, S. Cole, N. Deacon, C. Frenk, A. Fitzsimmons, S. Gezari, C. Goessl, T. Goggia, B. Goldman, E. K. Grebel, N. C. Hambly, G. Hasinger, A. F. Heavens, T. M. Heckman, R. Henderson, T. Henning, M. Holman, U. Hopp, W.-H. Ip, S. Isani, C. D. Keyes, A. Koekemoer, R. Kotak, K. S. Long, J. R Lucey, M. Liu, N. F. Martin, B. McLean, E. Morganson, D. N. A. Murphy, M. A. Nieto-Santisteban, P. Norberg, J. A. Peacock, E. A. Pier, M. Postman, N. Primak, C. Rae, A. Rest, A. Riess, A. Riﬀeser, H. W. Rix, S. Roser, E. Schilbach, A. S. B. Schultz, D. Scolnic, A. Szalay, S. Seitz, B. Shiao, E. Small, K. W. Smith, D. Soderblom, A. N. Taylor, A. R. Thakar, J. Thiel, D. Thilker, Y. Urata, J. Valenti, F. Walter, S. P. Watters, S. Werner, R. White, W. M. Wood-Vasey, and R. Wyse. The Pan-STARRS1 Surveys. ArXiv e-prints, arXiv:1612.05560, December 2016.

K. Chiboucas, I. D. Karachentsev, and R. B. Tully. Discovery of New Dwarf Galaxies in the M81 Group. Astronomical Journal, 137:3009–3037, February 2009. doi: 10.1088/0004-6256/137/2/3009. BIBLIOGRAPHY 144

K. Chiboucas, B. A. Jacobs, R. B. Tully, and I. D. Karachentsev. Conﬁrmation of Faint Dwarf Galaxies in the M81 Group. Astronomical Journal, 146:126, November 2013. doi: 10.1088/0004-6256/146/5/126.

H. M. Courtois and R. B. Tully. Update on H I data collection from Green Bank, Parkes and Arecibo telescopes for the Cosmic Flows project. Monthly Notices of the Royal Astronomical Society, 447:1531–1534, February 2015. doi: 10.1093/mnras/ stu2405.

D. Crnojević, D. J. Sand, N. Caldwell, P. Guhathakurta, B. McLeod, A. Seth, J. D. Simon, J. Strader, and E. Toloba. Discovery of a Close Pair of Faint Dwarf Galaxies in the Halo of Centaurus A. Astrophysical Journal Letters, 795:L35, November 2014. doi: 10.1088/2041-8205/795/2/L35.

D. Crnojević, D. J. Sand, K. Spekkens, N. Caldwell, P. Guhathakurta, B. McLeod, A. Seth, J. D. Simon, J. Strader, and E. Toloba. The Extended Halo of Centaurus A: Uncovering Satellites, Streams, and Substructures. Astrophysical Journal, 823: 19, May 2016. doi: 10.3847/0004-637X/823/1/19.

J. J. Dalcanton, D. N. Spergel, J. E. Gunn, M. Schmidt, and D. P. Schneider. The Number Density of Low-Surface Brightness Galaxies with 23 < mu_0 < 25 V Mag/arcsecˆ2. Astronomical Journal, 114:635–654, August 1997. doi: 10.1086/ 118499.

S. Danieli, P. van Dokkum, A. Merritt, R. Abraham, J. Zhang, I. D. Karachentsev, and L. N. Makarova. The Dragonﬂy Nearby Galaxies Survey. III. The Luminosity Function of the M101 Group. Astrophysical Journal, 837:136, March 2017. doi: 10.3847/1538-4357/aa615b. BIBLIOGRAPHY 145

Dark Energy Survey Collaboration. The Dark Energy Survey: more than dark energy - an overview. Monthly Notices of the Royal Astronomical Society, 460:1270–1299, August 2016. doi: 10.1093/mnras/stw641.

J. I. Davies, S. Phillipps, and M. J. Disney. Low surface brightness galaxies in the cluster A1367. Monthly Notices of the Royal Astronomical Society, 239:703–714, August 1989. doi: 10.1093/mnras/239.3.703.

W. J. G. de Blok, F. Walter, E. Brinks, C. Trachternach, S.-H. Oh, and R. C. Kenni- cutt, Jr. High-Resolution Rotation Curves and Galaxy Mass Models from THINGS. Astronomical Journal, 136:2648-2719, December 2008. doi: 10.1088/0004-6256/ 136/6/2648.

A. Di Cintio, C. B. Brook, A. A. Dutton, A. V. Macciò, A. Obreja, and A. Dekel. NIHAO - XI. Formation of ultra-diﬀuse galaxies by outﬂows. Monthly Notices of the Royal Astronomical Society, 466:L1–L6, March 2017. doi: 10.1093/mnrasl/slw210.

J. Diemand, M. Kuhlen, and P. Madau. Formation and Evolution of Galaxy Dark Matter Halos and Their Substructure. Astrophysical Journal, 667:859–877, October 2007. doi: 10.1086/520573.

G. A. Dooley, A. H. G. Peter, J. L. Carlin, A. Frebel, K. Bechtol, and B. Willman. The predicted luminous satellite populations around SMC and LMC-mass galaxies - A missing satellite problem around the LMC? ArXiv e-prints, arXiv:1703.05321, March 2017.

M. T. Doyle, M. J. Drinkwater, D. J. Rohde, K. A. Pimbblet, M. Read, M. J. BIBLIOGRAPHY 146

Meyer, M. A. Zwaan, E. Ryan-Weber, J. Stevens, B. S. Koribalski, R. L. Web- ster, L. Staveley-Smith, D. G. Barnes, M. Howlett, V. A. Kilborn, M. Waugh, M. J. Pierce, R. Bhathal, W. J. G. de Blok, M. J. Disney, R. D. Ekers, K. C. Freeman, D. A. Garcia, B. K. Gibson, J. Harnett, P. A. Henning, H. Jerjen, M. J. Kesteven, P. M. Knezek, S. Mader, M. Marquarding, R. F. Minchin, J. O’Brien, T. Oosterloo, R. M. Price, M. E. Putman, S. D. Ryder, E. M. Sadler, I. M. Stewart, F. Stootman, and A. E. Wright. The HIPASS catalogue - III. Optical counterparts and isolated dark galaxies. Monthly Notices of the Royal Astronomical Society, 361: 34–44, July 2005. doi: 10.1111/j.1365-2966.2005.09159.x.

J. L. E. Dreyer. A New General Catalogue of Nebulæ and Clusters of Stars, being the Catalogue of the late Sir John F. W. Herschel, Bart, revised, corrected, and enlarged. Memoirs of the Royal Astronomical Society, 49:1, 1888.

A. Drlica-Wagner, K. Bechtol, E. S. Rykoﬀ, E. Luque, A. Queiroz, Y.-Y. Mao, R. H. Wechsler, J. D. Simon, B. Santiago, B. Yanny, and DES Collaboration. Eight Ultra-faint Galaxy Candidates Discovered in Year Two of the Dark Energy Survey. Astrophysical Journal, 813:109, November 2015. doi: 10.1088/0004-637X/813/2/ 109.

A. Drlica-Wagner, K. Bechtol, S. Allam, D. L. Tucker, R. A. Gruendl, M. D. Johnson, A. R. Walker, D. J. James, D. L. Nidever, K. A. G. Olsen, R. H. Wechsler, M. R. L. Cioni, B. C. Conn, K. Kuehn, T. S. Li, Y.-Y. Mao, N. F. Martin, E. Neilsen, N. E. D. Noel, A. Pieres, J. D. Simon, G. S. Stringfellow, R. P. van der Marel, and B. Yanny. An Ultra-faint Galaxy Candidate Discovered in Early Data from BIBLIOGRAPHY 147

the Magellanic Satellites Survey. Astrophysical Journal Letters, 833:L5, December 2016. doi: 10.3847/2041-8205/833/1/L5.

S. P. Fillingham, M. C. Cooper, A. B. Pace, M. Boylan-Kolchin, J. S. Bullock, S. Garrison-Kimmel, and C. Wheeler. Under pressure: quenching star formation in low-mass satellite galaxies via stripping. Monthly Notices of the Royal Astronomical Society, 463:1916–1928, December 2016. doi: 10.1093/mnras/stw2131.

S. Garrison-Kimmel, M. Boylan-Kolchin, J. S. Bullock, and K. Lee. ELVIS: Explor- ing the Local Volume in Simulations. Monthly Notices of the Royal Astronomical Society, 438:2578–2596, March 2014. doi: 10.1093/mnras/stt2377.

A. Gatto, F. Fraternali, J. I. Read, F. Marinacci, H. Lux, and S. Walch. Unveiling the corona of the Milky Way via ram-pressure stripping of dwarf satellites. Monthly Notices of the Royal Astronomical Society, 433:2749–2763, August 2013. doi: 10. 1093/mnras/stt896.

M. Geha, B. Willman, J. D. Simon, L. E. Strigari, E. N. Kirby, D. R. Law, and J. Strader. The Least-Luminous Galaxy: Spectroscopy of the Milky Way Satellite Segue 1. Astrophysical Journal, 692:1464–1475, February 2009. doi: 10.1088/ 0004-637X/692/2/1464.

M. Geha, M. R. Blanton, R. Yan, and J. L. Tinker. A Stellar Mass Threshold for Quenching of Field Galaxies. Astrophysical Journal, 757:85, September 2012. doi: 10.1088/0004-637X/757/1/85.

M. Geha, R. H. Wechsler, Y.-Y. Mao, E. J. Tollerud, B. Weiner, R. Bernstein, B. Hoyle, S. Marchi, P. J. Marshall, R. Munoz, and Y. Lu. The SAGA Survey: I. BIBLIOGRAPHY 148

Satellite Galaxy Populations Around Eight Milky Way Analogs. ArXiv e-prints, arXiv:1705.06743, May 2017.

R. Giovanelli, M. P. Haynes, B. R. Kent, P. Perillat, A. Saintonge, N. Brosch, B. Catinella, G. L. Hoﬀman, S. Stierwalt, K. Spekkens, M. S. Lerner, K. L. Mas- ters, E. Momjian, J. L. Rosenberg, C. M. Springob, A. Boselli, V. Charmandaris, J. K. Darling, J. Davies, D. Garcia Lambas, G. Gavazzi, C. Giovanardi, E. Hardy, L. K. Hunt, A. Iovino, I. D. Karachentsev, V. E. Karachentseva, R. A. Koop- mann, C. Marinoni, R. Minchin, E. Muller, M. Putman, C. Pantoja, J. J. Salzer, M. Scodeggio, E. Skillman, J. M. Solanes, C. Valotto, W. van Driel, and L. van Zee. The Arecibo Legacy Fast ALFA Survey. I. Science Goals, Survey Design, and Strat- egy. Astronomical Journal, 130:2598–2612, December 2005. doi: 10.1086/497431.

R. Giovanelli, M. P. Haynes, E. A. K. Adams, J. M. Cannon, K. L. Rhode, J. J. Salzer, E. D. Skillman, E. Z. Bernstein-Cooper, and K. B. W. McQuinn. AL- FALFA Discovery of the Nearby Gas-rich Dwarf Galaxy Leo P. I. H I Observations. Astronomical Journal, 146:15, July 2013. doi: 10.1088/0004-6256/146/1/15.

J. Grcevich and M. E. Putman. H I in Local Group Dwarf Galaxies and Stripping by the Galactic Halo. Astrophysical Journal, 696:385-395, May 2009. doi: 10.1088/ 0004-637X/696/1/385.

J. P. Greco, J. E. Greene, M. A. Strauss, L. A. MacArthur, X. Flowers, A. D. Gould- ing, S. Huang, J. H. Kim, Y. Komiyama, A. Leauthaud, L. Leisman, R. H. Lupton, C. Sifón, and S.-Y. Wang. Illuminating Low-Surface-Brightness Galaxies with the Hyper Suprime-Cam Survey. ArXiv e-prints, arXiv:1709.04474, September 2017.

B. F. Griﬀen, A. P. Ji, G. A. Dooley, F. A. Gómez, M. Vogelsberger, B. W. O’Shea, BIBLIOGRAPHY 149

and A. Frebel. The Caterpillar Project: A Large Suite of Milky Way Sized Halos. Astrophysical Journal, 818:10, February 2016. doi: 10.3847/0004-637X/818/1/10.

J. E. Gunn and J. R. Gott, III. On the Infall of Matter Into Clusters of Galaxies and Some Eﬀects on Their Evolution. Astrophysical Journal, 176:1, August 1972. doi: 10.1086/151605.

G. L. H. Harris, M. Rejkuba, and W. E. Harris. The Distance to NGC 5128 (Centaurus A). Publications of the Astronomical Society of Australia, 27:457–462, October 2010. doi: 10.1071/AS09061.

M. P. Haynes and R. Giovanelli. Neutral hydrogen in isolated galaxies. IV - Results for the Arecibo sample. Astronomical Journal, 89:758–800, June 1984. doi: 10. 1086/113573.

M. P. Haynes, R. Giovanelli, A. M. Martin, K. M. Hess, A. Saintonge, E. A. K. Adams, G. Hallenbeck, G. L. Hoﬀman, S. Huang, B. R. Kent, R. A. Koopmann, E. Pa- pastergis, S. Stierwalt, T. J. Balonek, D. W. Craig, S. J. U. Higdon, D. A. Kornre- ich, J. R. Miller, A. A. O’Donoghue, R. P. Olowin, J. L. Rosenberg, K. Spekkens, P. Troischt, and E. M. Wilcots. The Arecibo Legacy Fast ALFA Survey: The α.40 H I Source Catalog, Its Characteristics and Their Impact on the Derivation of the H I Mass Function. Astronomical Journal, 142:170, November 2011. doi: 10.1088/0004-6256/142/5/170.

E. Holmberg. A photographic photometry of extragalactic nebulae. Meddelanden fran Lunds Astronomiska Observatorium Serie II, 136:1, 1958.

D. Homma, M. Chiba, S. Okamoto, Y. Komiyama, M. Tanaka, M. Tanaka, M. N. BIBLIOGRAPHY 150

Ishigaki, M. Akiyama, N. Arimoto, J. A. Garmilla, R. H. Lupton, M. A. Strauss, H. Furusawa, S. Miyazaki, H. Murayama, A. J. Nishizawa, M. Takada, T. Usuda, and S.-Y. Wang. A New Milky Way Satellite Discovered in the Subaru/Hyper Suprime-Cam Survey. Astrophysical Journal, 832:21, November 2016. doi: 10. 3847/0004-637X/832/1/21.

D. Homma, M. Chiba, S. Okamoto, Y. Komiyama, M. Tanaka, M. Tanaka, M. N. Ishigaki, K. Hayashi, N. Arimoto, J. A. Garmilla, R. H. Lupton, M. A. Strauss, S. Miyazaki, S.-Y. Wang, and H. Murayama. Searches for New Milky Way Satellites from the First Two Years of Data of the Subaru/Hyper Suprime-Cam Survey: Discovery of Cetus˜III. ArXiv e-prints, arXiv:1704.05977, April 2017.

S. Huang, M. P. Haynes, R. Giovanelli, J. Brinchmann, S. Stierwalt, and S. G. Neﬀ. Gas, Stars, and Star Formation in ALFALFA Dwarf Galaxies. Astronomical Jour- nal, 143:133, June 2012. doi: 10.1088/0004-6256/143/6/133.

R. Ibata, N. F. Martin, M. Irwin, S. Chapman, A. M. N. Ferguson, G. F. Lewis, and A. W. McConnachie. The Haunted Halos of Andromeda and Triangulum: A Panorama of Galaxy Formation in Action. Astrophysical Journal, 671:1591–1623, December 2007. doi: 10.1086/522574.

R. A. Ibata, G. F. Lewis, A. R. Conn, M. J. Irwin, A. W. McConnachie, S. C. Chapman, M. L. Collins, M. Fardal, A. M. N. Ferguson, N. G. Ibata, A. D. Mackey, N. F. Martin, J. Navarro, R. M. Rich, D. Valls-Gabaud, and L. M. Widrow. A vast, thin plane of corotating dwarf galaxies orbiting the Andromeda galaxy. Nature, 493:62–65, January 2013. doi: 10.1038/nature11717. BIBLIOGRAPHY 151

C. Impey and G. Bothun. Malin 1 - A quiescent disk galaxy. Astrophysical Journal, 341:89–104, June 1989. doi: 10.1086/167474.

C. D. Impey, D. Sprayberry, M. J. Irwin, and G. D. Bothun. Low Surface Brightness Galaxies in the Local Universe. I. The Catalog. Astrophysical Journal Supplement Series, 105:209, August 1996. doi: 10.1086/192313.

M. J. Irwin, J. I. Davies, M. J. Disney, and S. Phillipps. Automated galaxy surface photometry. V - Detection of very low surface brightness galaxies. Monthly Notices of the Royal Astronomical Society, 245:289–304, July 1990.

M. J. Irwin, A. M. N. Ferguson, A. P. Huxor, N. R. Tanvir, R. A. Ibata, and G. F. Lewis. Andromeda XVII: A New Low-Luminosity Satellite of M31. Astrophysical Journal Letters, 676:L17, March 2008. doi: 10.1086/587100.

N. Jarosik, C. L. Bennett, J. Dunkley, B. Gold, M. R. Greason, M. Halpern, R. S. Hill, G. Hinshaw, A. Kogut, E. Komatsu, D. Larson, M. Limon, S. S. Meyer, M. R. Nolta, N. Odegard, L. Page, K. M. Smith, D. N. Spergel, G. S. Tucker, J. L. Weiland, E. Wollack, and E. L. Wright. Seven-year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Sky Maps, Systematic Errors, and Basic Results. Astrophysical Journal Supplement Series, 192:14, February 2011. doi: 10.1088/0067-0049/192/2/14.

B. Javanmardi, D. Martinez-Delgado, P. Kroupa, C. Henkel, K. Crawford, K. Teuwen, R. J. Gabany, M. Hanson, T. S. Chonis, and F. Neyer. DGSAT: Dwarf Galaxy Survey with Amateur Telescopes. I. Discovery of low surface brightness systems around nearby spiral galaxies. Astronomy & Astrophysics, 588:A89, April 2016. doi: 10.1051/0004-6361/201527745. BIBLIOGRAPHY 152

M. Jobin and C. Carignan. The dark side of NGC 3109. Astronomical Journal, 100: 648–662, September 1990. doi: 10.1086/115548.

J. Kadowaki, D. Zaritsky, and R. L. Donnerstein. Spectroscopy of Ultra-diﬀuse Galax- ies in the Coma Cluster. Astrophysical Journal Letters, 838:L21, April 2017. doi: 10.3847/2041-8213/aa653d.

N. Kaiser, H. Aussel, B. E. Burke, H. Boesgaard, K. Chambers, M. R. Chun, J. N. Heasley, K.-W. Hodapp, B. Hunt, R. Jedicke, D. Jewitt, R. Kudritzki, G. A. Lup- pino, M. Maberry, E. Magnier, D. G. Monet, P. M. Onaka, A. J. Pickles, P. H. H. Rhoads, T. Simon, A. Szalay, I. Szapudi, D. J. Tholen, J. L. Tonry, M. Waterson, and J. Wick. Pan-STARRS: A Large Synoptic Survey Telescope Array. In J. A. Tyson and S. Wolﬀ, editors, Survey and Other Telescope Technologies and Discov- eries, volume 4836 of Proceedings of the SPIE, pages 154–164, December 2002. doi: 10.1117/12.457365.

I. D. Karachentsev, D. Bautzmann, F. Neyer, R. Polzl, P. Riepe, T. Zilch, and B. Mat- tern. Three low surface brightness dwarfs discovered around NGC 4631. ArXiv e-prints, arXiv:1401.2719, January 2014.

I. D. Karachentsev, P. Riepe, T. Zilch, M. Blauensteiner, M. Elvov, P. Hochleitner, B. Hubl, G. Kerschhuber, S. Küppers, F. Neyer, R. Pölzl, P. Remmel, O. Schneider, R. Sparenberg, U. Trulson, G. Willems, and H. Ziegler. New low surface brightness dwarf galaxies detected around nearby spirals. Astrophysical Bulletin, 70:379–391, October 2015. doi: 10.1134/S199034131504001X.

D. Kim and H. Jerjen. Horologium II: A Second Ultra-faint Milky Way Satellite BIBLIOGRAPHY 153

in the Horologium Constellation. Astrophysical Journal Letters, 808:L39, August 2015. doi: 10.1088/2041-8205/808/2/L39.

A. Klypin, A. V. Kravtsov, O. Valenzuela, and F. Prada. Where Are the Missing Galactic Satellites? Astrophysical Journal, 522:82–92, September 1999. doi: 10. 1086/307643.

J. Koda, M. Yagi, H. Yamanoi, and Y. Komiyama. Approximately a Thousand Ultra- diﬀuse Galaxies in the Coma Cluster. Astrophysical Journal Letters, 807:L2, July 2015. doi: 10.1088/2041-8205/807/1/L2.

S. Koposov, V. Belokurov, N. W. Evans, P. C. Hewett, M. J. Irwin, G. Gilmore, D. B. Zucker, H.-W. Rix, M. Fellhauer, E. F. Bell, and E. V. Glushkova. The Luminosity Function of the Milky Way Satellites. Astrophysical Journal, 686:279-291, October 2008. doi: 10.1086/589911.

S. E. Koposov, V. Belokurov, G. Torrealba, and N. W. Evans. Beasts of the Southern Wild: Discovery of Nine Ultra Faint Satellites in the Vicinity of the Magellanic Clouds. Astrophysical Journal, 805:130, June 2015. doi: 10.1088/0004-637X/805/ 2/130.

B. S. Koribalski, L. Staveley-Smith, V. A. Kilborn, S. D. Ryder, R. C. Kraan- Korteweg, E. V. Ryan-Weber, R. D. Ekers, H. Jerjen, P. A. Henning, M. E. Putman, M. A. Zwaan, W. J. G. de Blok, M. R. Calabretta, M. J. Disney, R. F. Minchin, R. Bhathal, P. J. Boyce, M. J. Drinkwater, K. C. Freeman, B. K. Gibson, A. J. Green, R. F. Haynes, S. Juraszek, M. J. Kesteven, P. M. Knezek, S. Mader, M. Mar- quarding, M. Meyer, J. R. Mould, T. Oosterloo, J. O’Brien, R. M. Price, E. M. Sadler, A. Schröder, I. M. Stewart, F. Stootman, M. Waugh, B. E. Warren, R. L. BIBLIOGRAPHY 154

Webster, and A. E. Wright. The 1000 Brightest HIPASS Galaxies: H I Properties. Astronomical Journal, 128:16–46, July 2004a. doi: 10.1086/421744.

B. P. M. Laevens, N. F. Martin, B. Sesar, E. J. Bernard, H.-W. Rix, C. T. Slater, E. F. Bell, A. M. N. Ferguson, E. F. Schlaﬂy, W. S. Burgett, K. C. Chambers, L. Denneau, P. W. Draper, N. Kaiser, R.-P. Kudritzki, E. A. Magnier, N. Metcalfe, J. S. Morgan, P. A. Price, W. E. Sweeney, J. L. Tonry, R. J. Wainscoat, and C. Waters. A New Distant Milky Way Globular Cluster in the Pan-STARRS1 3π Survey. Astrophysical Journal Letters, 786:L3, May 2014. doi: 10.1088/2041-8205/ 786/1/L3.

B. P. M. Laevens, N. F. Martin, E. J. Bernard, E. F. Schlaﬂy, B. Sesar, H.-W. Rix, E. F. Bell, A. M. N. Ferguson, C. T. Slater, W. E. Sweeney, R. F. G. Wyse, A. P. Huxor, W. S. Burgett, K. C. Chambers, P. W. Draper, K. A. Hodapp, N. Kaiser, E. A. Magnier, N. Metcalfe, J. L. Tonry, R. J. Wainscoat, and C. Waters. Sagittarius II, Draco II and Laevens 3: Three New Milky Way Satellites Discovered BIBLIOGRAPHY 155

in the Pan-STARRS 1 3π Survey. Astrophysical Journal, 813:44, November 2015a. doi: 10.1088/0004-637X/813/1/44.

B. P. M. Laevens, N. F. Martin, R. A. Ibata, H.-W. Rix, E. J. Bernard, E. F. Bell, B. Sesar, A. M. N. Ferguson, E. F. Schlaﬂy, C. T. Slater, W. S. Burgett, K. C. Chambers, H. Flewelling, K. A. Hodapp, N. Kaiser, R.-P. Kudritzki, R. H. Lupton, E. A. Magnier, N. Metcalfe, J. S. Morgan, P. A. Price, J. L. Tonry, R. J. Wainscoat, and C. Waters. A New Faint Milky Way Satellite Discovered in the Pan-STARRS1 3π Survey. Astrophysical Journal Letters, 802:L18, April 2015b. doi: 10.1088/ 2041-8205/802/2/L18.

L. Leisman, M. P. Haynes, S. Janowiecki, G. Hallenbeck, G. Józsa, R. Giovanelli, E. A. K. Adams, D. Bernal Neira, J. M. Cannon, W. F. Janesh, K. L. Rhode, and J. J. Salzer. (Almost) Dark Galaxies in the ALFALFA Survey: Isolated H i-bearing Ultra-diﬀuse Galaxies. Astrophysical Journal, 842:133, June 2017. doi: 10.3847/1538-4357/aa7575.

E. Luque, A. Queiroz, B. Santiago, A. Pieres, E. Balbinot, K. Bechtol, A. Drlica- Wagner, A. F. Neto, L. N. da Costa, M. A. G. Maia, B. Yanny, T. Abbott, S. Allam, A. Benoit-Lévy, E. Bertin, D. Brooks, E. Buckley-Geer, D. L. Burke, A. C. Rosell, M. C. Kind, J. Carretero, C. E. Cunha, S. Desai, H. T. Diehl, J. P. Dietrich, T. F. Eiﬂer, D. A. Finley, B. Flaugher, P. Fosalba, J. Frieman, D. W. Gerdes, D. Gruen, G. Gutierrez, K. Honscheid, D. J. James, K. Kuehn, N. Kuropatkin, O. Lahav, T. S. Li, M. March, J. L. Marshall, P. Martini, R. Miquel, E. Neilsen, R. C. Nichol, B. Nord, R. Ogando, A. A. Plazas, A. K. Romer, A. Roodman, BIBLIOGRAPHY 156

E. Sanchez, V. Scarpine, M. Schubnell, I. Sevilla-Noarbe, R. C. Smith, M. Soares- Santos, F. Sobreira, E. Suchyta, M. E. C. Swanson, G. Tarle, J. Thaler, D. Tucker, A. R. Walker, and Y. Zhang. Digging deeper into the Southern skies: a compact Milky Way companion discovered in ﬁrst-year Dark Energy Survey data. Monthly Notices of the Royal Astronomical Society, 458:603–612, May 2016. doi: 10.1093/ mnras/stw302.

D. Lynden-Bell. Dwarf galaxies and globular clusters in high velocity hydrogen streams. Monthly Notices of the Royal Astronomical Society, 174:695–710, March 1976. doi: 10.1093/mnras/174.3.695.

P. Madau, S. Shen, and F. Governato. Dark Matter Heating and Early Core Formation in Dwarf Galaxies. Astrophysical Journal Letters, 789:L17, July 2014. doi: 10.1088/ 2041-8205/789/1/L17.

S. R. Majewski, M. F. Skrutskie, M. D. Weinberg, and J. C. Ostheimer. A Two Micron All Sky Survey View of the Sagittarius Dwarf Galaxy. I. Morphology of the Sagittarius Core and Tidal Arms. Astrophysical Journal, 599:1082–1115, December 2003. doi: 10.1086/379504.

D. Makarov, P. Prugniel, N. Terekhova, H. Courtois, and I. Vauglin. HyperLEDA. III. The catalogue of extragalactic distances. Astronomy & Astrophysics, 570:A13, October 2014. doi: 10.1051/0004-6361/201423496.

N. F. Martin, R. A. Ibata, M. J. Irwin, S. Chapman, G. F. Lewis, A. M. N. Ferguson, N. Tanvir, and A. W. McConnachie. Discovery and analysis of three faint dwarf galaxies and a globular cluster in the outer halo of the Andromeda galaxy. Monthly BIBLIOGRAPHY 157

Notices of the Royal Astronomical Society, 371:1983–1991, October 2006. doi: 10. 1111/j.1365-2966.2006.10823.x.

N. F. Martin, J. T. A. de Jong, and H.-W. Rix. A Comprehensive Maximum Like- lihood Analysis of the Structural Properties of Faint Milky Way Satellites. Astro- physical Journal, 684:1075-1092, September 2008. doi: 10.1086/590336.

N. F. Martin, A. W. McConnachie, M. Irwin, L. M. Widrow, A. M. N. Ferguson, R. A. Ibata, J. Dubinski, A. Babul, S. Chapman, M. Fardal, G. F. Lewis, J. Navarro, and R. M. Rich. PAndAS’ CUBS: Discovery of Two New Dwarf Galaxies in the Surroundings of the Andromeda and Triangulum Galaxies. Astrophysical Journal, 705:758–765, November 2009. doi: 10.1088/0004-637X/705/1/758.

N. F. Martin, R. A. Ibata, A. W. McConnachie, A. D. Mackey, A. M. N. Ferguson, M. J. Irwin, G. F. Lewis, and M. A. Fardal. The PAndAS View of the Andromeda Satellite System. I. A Bayesian Search for Dwarf Galaxies Using Spatial and Color- Magnitude Information. Astrophysical Journal, 776:80, October 2013a. doi: 10. 1088/0004-637X/776/2/80.

N. F. Martin, E. F. Schlaﬂy, C. T. Slater, E. J. Bernard, H.-W. Rix, E. F. Bell, A. M. N. Ferguson, D. P. Finkbeiner, B. P. M. Laevens, W. S. Burgett, K. C. Chambers, P. W. Draper, K. W. Hodapp, N. Kaiser, R.-P. Kudritzki, E. A. Magnier, N. Metcalfe, J. S. Morgan, P. A. Price, J. L. Tonry, R. J. Wainscoat, and C. Waters. Perseus I: A Distant Satellite Dwarf Galaxy of Andromeda. Astrophysical Journal Letters, 779:L10, December 2013b. doi: 10.1088/2041-8205/779/1/L10.

N. F. Martin, C. T. Slater, E. F. Schlaﬂy, E. Morganson, H.-W. Rix, E. F. Bell, BIBLIOGRAPHY 158

B. P. M. Laevens, E. J. Bernard, A. M. N. Ferguson, D. P. Finkbeiner, W. S. Bur- gett, K. C. Chambers, K. W. Hodapp, N. Kaiser, R.-P. Kudritzki, E. A. Magnier, J. S. Morgan, P. A. Price, J. L. Tonry, and R. J. Wainscoat. Lacerta I and Cas- siopeia III. Two Luminous and Distant Andromeda Satellite Dwarf Galaxies Found in the 3π Pan-STARRS1 Survey. Astrophysical Journal, 772:15, July 2013c. doi: 10.1088/0004-637X/772/1/15.

N. F. Martin, D. L. Nidever, G. Besla, K. Olsen, A. R. Walker, A. K. Vivas, R. A. Gruendl, C. C. Kaleida, R. R. Muñoz, R. D. Blum, A. Saha, B. C. Conn, E. F. Bell, Y.-H. Chu, M.-R. L. Cioni, T. J. L. de Boer, C. Gallart, S. Jin, A. Kunder, S. R. Majewski, D. Martinez-Delgado, A. Monachesi, M. Monelli, L. Monteagudo, N. E. D. Noël, E. W. Olszewski, G. S. Stringfellow, R. P. van der Marel, and D. Zaritsky. Hydra II: A Faint and Compact Milky Way Dwarf Galaxy Found in the Survey of the Magellanic Stellar History. Astrophysical Journal Letters, 804: L5, May 2015. doi: 10.1088/2041-8205/804/1/L5.

N. F. Martin, R. A. Ibata, G. F. Lewis, A. McConnachie, A. Babul, N. F. Bate, E. Bernard, S. C. Chapman, M. M. L. Collins, A. R. Conn, D. Crnojević, M. A. Fardal, A. M. N. Ferguson, M. Irwin, A. D. Mackey, B. McMonigal, J. F. Navarro, and R. M. Rich. The PAndAS View of the Andromeda Satellite System. II. Detailed Properties of 23 M31 Dwarf Spheroidal Galaxies. Astrophysical Journal, 833:167, December 2016. doi: 10.3847/1538-4357/833/2/167.

D. Martínez-Delgado, M. Á. Gómez-Flechoso, A. Aparicio, and R. Carrera. Trac- ing Out the Northern Tidal Stream of the Sagittarius Dwarf Spheroidal Galaxy. Astrophysical Journal, 601:242–259, January 2004. doi: 10.1086/380298. BIBLIOGRAPHY 159

D. Martínez-Delgado, J. Peñarrubia, R. J. Gabany, I. Trujillo, S. R. Majewski, and M. Pohlen. The Ghost of a Dwarf Galaxy: Fossils of the Hierarchical Formation of the Nearby Spiral Galaxy NGC 5907. Astrophysical Journal, 689:184-193, Decem- ber 2008. doi: 10.1086/592555.

D. Martínez-Delgado, R. J. Gabany, K. Crawford, S. Zibetti, S. R. Majewski, H.- W. Rix, J. Fliri, J. A. Carballo-Bello, D. C. Bardalez-Gagliuﬃ, J. Peñarrubia, T. S. Chonis, B. Madore, I. Trujillo, M. Schirmer, and D. A. McDavid. Stellar Tidal Streams in Spiral Galaxies of the Local Volume: A Pilot Survey with Modest Aperture Telescopes. Astronomical Journal, 140:962–967, October 2010. doi: 10. 1088/0004-6256/140/4/962.

D. Martínez-Delgado, A. J. Romanowsky, R. J. Gabany, F. Annibali, J. A. Arnold, J. Fliri, S. Zibetti, R. P. van der Marel, H.-W. Rix, T. S. Chonis, J. A. Carballo- Bello, A. Aloisi, A. V. Macciò, J. Gallego-Laborda, J. P. Brodie, and M. R. Mer- riﬁeld. Dwarfs Gobbling Dwarfs: A Stellar Tidal Stream around NGC 4449 and Hierarchical Galaxy Formation on Small Scales. Astrophysical Journal Letters, 748: L24, April 2012. doi: 10.1088/2041-8205/748/2/L24.

D. Martínez-Delgado, R. Läsker, M. Sharina, E. Toloba, J. Fliri, R. Beaton, D. Valls- Gabaud, I. D. Karachentsev, T. S. Chonis, E. K. Grebel, D. A. Forbes, A. J. Romanowsky, J. Gallego-Laborda, K. Teuwen, M. A. Gómez-Flechoso, J. Wang, P. Guhathakurta, S. Kaisin, and N. Ho. Discovery of an Ultra-diﬀuse Galaxy in the Pisces–Perseus Supercluster. Astronomical Journal, 151:96, April 2016. doi: 10.3847/0004-6256/151/4/96.

L. D. Matthews, J. S. Gallagher, III, and W. van Driel. The Extraordinary “Superthin” BIBLIOGRAPHY 160

Spiral Galaxy UGC 7321. I. Disk Color Gradients and Global Properties from Multiwavelength Observations. Astronomical Journal, 118:2751–2766, December 1999a. doi: 10.1086/301128.

L. D. Matthews, W. van Driel, and J. S. Gallagher. Properties of ‘Superthin’ Galaxies. ArXiv Astrophysics e-prints, November 1999b.

L. Mayer, C. Mastropietro, J. Wadsley, J. Stadel, and B. Moore. Simultane- ous ram pressure and tidal stripping; how dwarf spheroidals lost their gas. Monthly Notices of the Royal Astronomical Society, 369:1021–1038, July 2006. doi: 10.1111/j.1365-2966.2006.10403.x.

A. W. McConnachie. The Observed Properties of Dwarf Galaxies in and around the Local Group. Astronomical Journal, 144:4, July 2012. doi: 10.1088/0004-6256/ 144/1/4.

A. W. McConnachie, A. Huxor, N. F. Martin, M. J. Irwin, S. C. Chapman, G. Fahlman, A. M. N. Ferguson, R. A. Ibata, G. F. Lewis, H. Richer, and N. R. Tan- vir. A Trio of New Local Group Galaxies with Extreme Properties. Astrophysical Journal, 688:1009-1020, December 2008. doi: 10.1086/591313.

A. W. McConnachie, M. J. Irwin, R. A. Ibata, J. Dubinski, L. M. Widrow, N. F. Martin, P. Côté, A. L. Dotter, J. F. Navarro, A. M. N. Ferguson, T. H. Puzia, G. F. Lewis, A. Babul, P. Barmby, O. Bienaymé, S. C. Chapman, R. Cockcroft, M. L. M. Collins, M. A. Fardal, W. E. Harris, A. Huxor, A. D. Mackey, J. Peñarrubia, R. M. Rich, H. B. Richer, A. Siebert, N. Tanvir, D. Valls-Gabaud, and K. A. Venn. The remnants of galaxy formation from a panoramic survey of the region around M31. Nature, 461:66–69, September 2009a. doi: 10.1038/nature08327. BIBLIOGRAPHY 161

S. S. McGaugh, V. C. Rubin, and W. J. G. de Blok. High-Resolution Rotation Curves of Low Surface Brightness Galaxies. I. Data. Astronomical Journal, 122:2381–2395, November 2001. doi: 10.1086/323448.

K. B. W. McQuinn, E. D. Skillman, D. Berg, J. M. Cannon, J. J. Salzer, E. A. K. Adams, A. Dolphin, R. Giovanelli, M. P. Haynes, and K. L. Rhode. ALFALFA Discovery of the Nearby Gas-rich Dwarf Galaxy Leo P. IV. Distance Measurement from LBT Optical Imaging. Astronomical Journal, 146:145, December 2013. doi: 10.1088/0004-6256/146/6/145.

K. B. W. McQuinn, E. D. Skillman, A. Dolphin, J. M. Cannon, J. J. Salzer, K. L. Rhode, E. A. K. Adams, D. Berg, R. Giovanelli, L. Girardi, and M. P. Haynes. Leo P: An Unquenched Very Low-mass Galaxy. Astrophysical Journal, 812:158, October 2015. doi: 10.1088/0004-637X/812/2/158.

A. Merritt, P. van Dokkum, and R. Abraham. The Discovery of Seven Extremely Low Surface Brightness Galaxies in the Field of the Nearby Spiral Galaxy M101. Astrophysical Journal Letters, 787:L37, June 2014. doi: 10.1088/2041-8205/787/ 2/L37. BIBLIOGRAPHY 162

A. Merritt, P. van Dokkum, S. Danieli, R. Abraham, J. Zhang, I. D. Karachentsev, and L. N. Makarova. The Dragonﬂy Nearby Galaxies Survey. II. Ultra-Diﬀuse Galaxies near the Elliptical Galaxy NGC 5485. Astrophysical Journal, 833:168, December 2016. doi: 10.3847/1538-4357/833/2/168.

J. C. Mihos, K. M. Keating, K. Holley-Bockelmann, D. J. Pisano, and N. E. Kas- sim. The H I Environment of the M101 Group. Astrophysical Journal, 761:186, December 2012. doi: 10.1088/0004-637X/761/2/186.

J. C. Mihos, P. Harding, C. E. Spengler, C. S. Rudick, and J. J. Feldmeier. The Extended Optical Disk of M101. Astrophysical Journal, 762:82, January 2013. doi: 10.1088/0004-637X/762/2/82.

J. C. Mihos, P. R. Durrell, L. Ferrarese, J. J. Feldmeier, P. Côté, E. W. Peng, P. Hard- ing, C. Liu, S. Gwyn, and J.-C. Cuillandre. Galaxies at the Extremes: Ultra-diﬀuse Galaxies in the Virgo Cluster. Astrophysical Journal Letters, 809:L21, August 2015. doi: 10.1088/2041-8205/809/2/L21.

B. Moore, S. Ghigna, F. Governato, G. Lake, T. Quinn, J. Stadel, and P. Tozzi. Dark Matter Substructure within Galactic Halos. Astrophysical Journal Letters, 524: L19–L22, October 1999. doi: 10.1086/312287.

B. P. Moster, R. S. Somerville, J. A. Newman, and H.-W. Rix. A Cosmic Variance Cookbook. Astrophysical Journal, 731:113, April 2011. doi: 10.1088/0004-637X/ 731/2/113.

O. Müller, R. Scalera, B. Binggeli, and H. Jerjen. The M 101 group complex: new BIBLIOGRAPHY 163

dwarf galaxy candidates and spatial structure. Astronomy & Astrophysics, 602: A119, June 2017. doi: 10.1051/0004-6361/201730434.

K. O’Neil. Single-Dish Calibration Techniques at Radio Wavelengths. In S. Stan- imirovic, D. Altschuler, P. Goldsmith, and C. Salter, editors, Single-Dish Radio Astronomy: Techniques and Applications, volume 278 of Astronomical Society of the Paciﬁc Conference Series, pages 293–311, December 2002.

K. O’Neil, G. D. Bothun, and M. E. Cornell. A Wide Field CCD Survey for Low Surface Brightness Galaxies:I.Data Acquisition, Description, and Initial Results. Astronomical Journal, 113:1212, April 1997a. doi: 10.1086/118338.

K. O’Neil, G. D. Bothun, J. Schombert, M. E. Cornell, and C. D. Impey. A Wide Field CCD Survey for Low Surface Brightness Galaxies.II.Color Distributions, Stellar Populations, and Missing Baryons. Astronomical Journal, 114:2448, December 1997b. doi: 10.1086/118659.

K. O’Neil, G. D. Bothun, and C. D. Impey. Hubble Space Telescope WFPC2 Imaging of Three Low Surface Brightness Dwarf Elliptical Galaxies in the Virgo Cluster. Astronomical Journal, 118:1618–1634, October 1999. doi: 10.1086/301026.

E. Papastergis, R. Giovanelli, M. P. Haynes, and F. Shankar. Is there a “too big to fail” problem in the ﬁeld? Astronomy & Astrophysics, 574:A113, February 2015. doi: 10.1051/0004-6361/201424909.

E. Papastergis, E. A. K. Adams, and A. J. Romanowsky. The HI content of isolated ultra-diﬀuse galaxies: A sign of multiple formation mechanisms? Astronomy & Astrophysics, 601:L10, May 2017. doi: 10.1051/0004-6361/201730795. BIBLIOGRAPHY 164

M. S. Pawlowski, J. Pﬂamm-Altenburg, and P. Kroupa. The VPOS: a vast polar structure of satellite galaxies, globular clusters and streams around the Milky Way. Monthly Notices of the Royal Astronomical Society, 423:1109–1126, June 2012. doi: 10.1111/j.1365-2966.2012.20937.x.

Planck Collaboration, P. A. R. Ade, N. Aghanim, M. Arnaud, M. Ashdown, J. Au- mont, C. Baccigalupi, A. J. Banday, R. B. Barreiro, J. G. Bartlett, and et al. Planck 2015 results. XIII. Cosmological parameters. Astronomy & Astrophysics, 594:A13, September 2016. doi: 10.1051/0004-6361/201525830.

D. J. Radburn-Smith, R. S. de Jong, A. C. Seth, J. Bailin, E. F. Bell, T. M. Brown, J. S. Bullock, S. Courteau, J. J. Dalcanton, H. C. Ferguson, P. Goud- frooij, S. Holfeltz, B. W. Holwerda, C. Purcell, J. Sick, D. Streich, M. Vlajic, and D. B. Zucker. The GHOSTS Survey. I. Hubble Space Telescope Advanced Camera for Surveys Data. Astrophysical Journal Supplement Series, 195:18, August 2011. doi: 10.1088/0067-0049/195/2/18.

K. L. Rhode, J. J. Salzer, N. C. Haurberg, A. Van Sistine, M. D. Young, M. P. Haynes, R. Giovanelli, J. M. Cannon, E. D. Skillman, K. B. W. McQuinn, and E. A. K. Adams. ALFALFA Discovery of the Nearby Gas-rich Dwarf Galaxy Leo P. II. Optical Imaging Observations. Astronomical Journal, 145:149, June 2013. doi: 10.1088/0004-6256/145/6/149.

J. Román and I. Trujillo. Ultra-diﬀuse galaxies outside clusters: clues to their formation and evolution. Monthly Notices of the Royal Astronomical Society, 468: 4039–4047, July 2017. doi: 10.1093/mnras/stx694. BIBLIOGRAPHY 165

B. F. Roukema and B. A. Peterson. A survey for low surface brightness galaxies in the ﬁeld. Astronomy & Astrophysics Supplement, 109:511–521, March 1995.

D. J. Sand, D. Crnojević, J. Strader, E. Toloba, J. D. Simon, N. Caldwell, P. Guhathakurta, B. McLeod, and A. C. Seth. Discovery of a New Faint Dwarf Galaxy Associated with NGC 253. Astrophysical Journal Letters, 793:L7, Septem- ber 2014. doi: 10.1088/2041-8205/793/1/L7.

D. J. Sand, K. Spekkens, D. Crnojević, J. R. Hargis, B. Willman, J. Strader, and C. J. Grillmair. Antlia B: A Faint Dwarf Galaxy Member of the NGC 3109 Association. Astrophysical Journal Letters, 812:L13, October 2015. doi: 10.1088/2041-8205/ 812/1/L13.

T. Sawala, C. S. Frenk, A. Fattahi, J. F. Navarro, R. G. Bower, R. A. Crain, C. Dalla Vecchia, M. Furlong, J. C. Helly, A. Jenkins, K. A. Oman, M. Schaller, J. Schaye, T. Theuns, J. Trayford, and S. D. M. White. The APOSTLE simulations: solutions to the Local Group’s cosmic puzzles. Monthly Notices of the Royal Astronomical Society, 457:1931–1943, April 2016a. doi: 10.1093/mnras/stw145.

T. Sawala, C. S. Frenk, A. Fattahi, J. F. Navarro, T. Theuns, R. G. Bower, R. A. Crain, M. Furlong, A. Jenkins, M. Schaller, and J. Schaye. The chosen few: the low-mass haloes that host faint galaxies. Monthly Notices of the Royal Astronomical Society, 456:85–97, February 2016b. doi: 10.1093/mnras/stv2597.

J. Schaye, R. A. Crain, R. G. Bower, M. Furlong, M. Schaller, T. Theuns, C. Dalla Vecchia, C. S. Frenk, I. G. McCarthy, J. C. Helly, A. Jenkins, Y. M. Rosas- Guevara, S. D. M. White, M. Baes, C. M. Booth, P. Camps, J. F. Navarro, Y. Qu, A. Rahmati, T. Sawala, P. A. Thomas, and J. Trayford. The EAGLE BIBLIOGRAPHY 166

project: simulating the evolution and assembly of galaxies and their environments. Monthly Notices of the Royal Astronomical Society, 446:521–554, January 2015. doi: 10.1093/mnras/stu2058.

T. Shanks, N. Metcalfe, B. Chehade, J. R. Findlay, M. J. Irwin, E. Gonzalez-Solares, J. R. Lewis, A. K. Yoldas, R. G. Mann, M. A. Read, E. T. W. Sutorius, and S. Voutsinas. The VLT Survey Telescope ATLAS. Monthly Notices of the Royal Astronomical Society, 451:4238–4252, August 2015. doi: 10.1093/mnras/stv1130.

S. Shen, P. Madau, C. Conroy, F. Governato, and L. Mayer. The Baryon Cycle of Dwarf Galaxies: Dark, Bursty, Gas-rich Polluters. Astrophysical Journal, 792:99, September 2014. doi: 10.1088/0004-637X/792/2/99.

D. D. Shi, X. Z. Zheng, H. B. Zhao, Z. Z. Pan, B. Li, H. Zou, X. Zhou, K. Guo, F. X. An, and Y. B. Li. Deep Imaging of the HCG 95 Field. I. Ultra-diﬀuse Galaxies. Astrophysical Journal, 846:26, September 2017. doi: 10.3847/1538-4357/aa8327.

J. D. Simon and M. Geha. The Kinematics of the Ultra-faint Milky Way Satel- lites: Solving the Missing Satellite Problem. Astrophysical Journal, 670:313–331, November 2007. doi: 10.1086/521816.

C. M. Simpson, R. J. J. Grand, F. A. Gómez, F. Marinacci, R. Pakmor, V. Springel, D. J. R. Campbell, and C. S. Frenk. Quenching and ram pressure stripping of simulated Milky Way satellite galaxies. ArXiv e-prints, May 2017.

C. T. Slater, E. F. Bell, and N. F. Martin. Andromeda XXVIII: A Dwarf Galaxy More Than 350 kpc from Andromeda. Astrophysical Journal Letters, 742:L14, November 2011. doi: 10.1088/2041-8205/742/1/L14. BIBLIOGRAPHY 167

K. Spekkens, B. S. Mason, J. E. Aguirre, and B. Nhan. A Deep Search for Ex- tended Radio Continuum Emission from Dwarf Spheroidal Galaxies: Implica- tions for Particle Dark Matter. Astrophysical Journal, 773:61, August 2013. doi: 10.1088/0004-637X/773/1/61.

K. Spekkens, N. Urbancic, B. S. Mason, B. Willman, and J. E. Aguirre. The Dearth of Neutral Hydrogen in Galactic Dwarf Spheroidal Galaxies. Astrophysical Journal Letters, 795:L5, November 2014. doi: 10.1088/2041-8205/795/1/L5.

V. Springel, S. D. M. White, A. Jenkins, C. S. Frenk, N. Yoshida, L. Gao, J. Navarro, R. Thacker, D. Croton, J. Helly, J. A. Peacock, S. Cole, P. Thomas, H. Couchman, A. Evrard, J. Colberg, and F. Pearce. Simulations of the formation, evolution and clustering of galaxies and quasars. Nature, 435:629–636, June 2005. doi: 10.1038/nature03597.

V. Springel, J. Wang, M. Vogelsberger, A. Ludlow, A. Jenkins, A. Helmi, J. F. Navarro, C. S. Frenk, and S. D. M. White. The Aquarius Project: the subhaloes of galactic haloes. Monthly Notices of the Royal Astronomical Society, 391:1685–1711, December 2008. doi: 10.1111/j.1365-2966.2008.14066.x.

C. M. Springob, M. P. Haynes, R. Giovanelli, and B. R. Kent. A Digital Archive of H I 21 Centimeter Line Spectra of Optically Targeted Galaxies. Astrophysical Journal Supplement Series, 160:149–162, September 2005. doi: 10.1086/431550.

The Dark Energy Survey Collaboration. The Dark Energy Survey. ArXiv e-prints, arXiv:0510346, October 2005.

T. X. Thuan and G. E. Martin. Blue compact dwarf galaxies. I - Neutral hydrogen BIBLIOGRAPHY 168

observations of 115 galaxies. Astrophysical Journal, 247:823–848, August 1981. doi: 10.1086/159094.

T. X. Thuan, J.-M. Alimi, J. R. Gott, III, and S. E. Schneider. Northern dwarf and low surface brightness galaxies. IV - The large-scale space distribution. Astrophysical Journal, 370:25–48, March 1991. doi: 10.1086/169791.

T. X. Thuan, V. A. Lipovetsky, J.-M. Martin, and S. A. Pustilnik. HI observations of blue compact galaxies from the ﬁrst and second Byurakan surveys. Astronomy & Astrophysics Supplement, 139:1–24, October 1999. doi: 10.1051/aas:1999373.

N. A. Tikhonov, V. S. Lebedev, and O. A. Galazutdinova. M 101 group galaxies. Astronomy Letters, 41:239–251, June 2015. doi: 10.1134/S1063773715060080.

E. J. Tollerud, M. Boylan-Kolchin, and J. S. Bullock. M31 satellite masses compared to ΛCDM subhaloes. Monthly Notices of the Royal Astronomical Society, 440: 3511–3519, June 2014. doi: 10.1093/mnras/stu474.

E. Toloba, D. J. Sand, K. Spekkens, D. Crnojević, J. D. Simon, P. Guhathakurta, J. Strader, N. Caldwell, B. McLeod, and A. C. Seth. A Tidally Disrupting Dwarf Galaxy in the Halo of NGC 253. Astrophysical Journal Letters, 816:L5, January 2016. doi: 10.3847/2041-8205/816/1/L5.

G. Torrealba, S. E. Koposov, V. Belokurov, and M. Irwin. The feeble giant. Discov- ery of a large and diﬀuse Milky Way dwarf galaxy in the constellation of Crater. Monthly Notices of the Royal Astronomical Society, 459:2370–2378, July 2016a. doi: 10.1093/mnras/stw733. BIBLIOGRAPHY 169

G. Torrealba, S. E. Koposov, V. Belokurov, M. Irwin, M. Collins, M. Spencer, R. Ibata, M. Mateo, A. Bonaca, and P. Jethwa. At the survey limits: discovery of the Aquarius 2 dwarf galaxy in the VST ATLAS and the SDSS data. Monthly Notices of the Royal Astronomical Society, 463:712–722, November 2016b. doi: 10.1093/mnras/stw2051.

I. Trujillo, J. Roman, M. Filho, and J. Sánchez Almeida. The Nearest Ultra Diﬀuse Galaxy: UGC 2162. Astrophysical Journal, 836:191, February 2017. doi: 10.3847/ 1538-4357/aa5cbb.

R. B. Tully, L. Rizzi, A. E. Dolphin, I. D. Karachentsev, V. E. Karachentseva, D. I. Makarov, L. Makarova, S. Sakai, and E. J. Shaya. Associations of Dwarf Galaxies. Astronomical Journal, 132:729–748, August 2006. doi: 10.1086/505466.

R. B. Tully, E. J. Shaya, I. D. Karachentsev, H. M. Courtois, D. D. Kocevski, L. Rizzi, and A. Peel. Our Peculiar Motion Away from the Local Void. Astrophysical Journal, 676:184-205, March 2008. doi: 10.1086/527428.

R. F. J. van der Burg, H. Hoekstra, A. Muzzin, C. Sifon, M. Viola, M. N. Bremer, S. Brough, S. P. Driver, T. Erben, C. Heymans, H. Hildebrandt, B. W. Holwerda, D. Klaes, K. Kuijken, S. McGee, R. Nakajima, N. Napolitano, P. Norberg, E. N. Taylor, and E. Valentijn. The abundance of ultra-diﬀuse galaxies from groups to clusters: UDGs are relatively more common in more massive haloes. ArXiv e-prints, arXiv:1706.02704, June 2017.

P. G. van Dokkum, R. Abraham, and A. Merritt. First Results from the Dragon- ﬂy Telephoto Array: The Apparent Lack of a Stellar Halo in the Massive Spi- ral Galaxy M101. Astrophysical Journal Letters, 782:L24, February 2014. doi: BIBLIOGRAPHY 170

10.1088/2041-8205/782/2/L24.

P. G. van Dokkum, R. Abraham, A. Merritt, J. Zhang, M. Geha, and C. Conroy. Forty-seven Milky Way-sized, Extremely Diﬀuse Galaxies in the Coma Cluster. Astrophysical Journal Letters, 798:L45, January 2015a. doi: 10.1088/2041-8205/ 798/2/L45.

P. G. van Dokkum, A. J. Romanowsky, R. Abraham, J. P. Brodie, C. Conroy, M. Geha, A. Merritt, A. Villaume, and J. Zhang. Spectroscopic Conﬁrmation of the Existence of Large, Diﬀuse Galaxies in the Coma Cluster. Astrophysical Journal Letters, 804:L26, May 2015b. doi: 10.1088/2041-8205/804/1/L26.

M. Vogelsberger, S. Genel, V. Springel, P. Torrey, D. Sijacki, D. Xu, G. Snyder, D. Nelson, and L. Hernquist. Introducing the Illustris Project: simulating the coevolution of dark and visible matter in the Universe. Monthly Notices of the Royal Astronomical Society, 444:1518–1547, October 2014. doi: 10.1093/mnras/stu1536.

K. Voggel, M. Hilker, H. Baumgardt, M. L. M. Collins, E. K. Grebel, B. Husemann, T. Richtler, and M. J. Frank. Probing the boundary between star clusters and dwarf galaxies: A MUSE view on the dynamics of Crater/Laevens I. Monthly Notices of the Royal Astronomical Society, 460:3384–3397, August 2016. doi: 10.1093/mnras/ stw1132. BIBLIOGRAPHY 171

A. R. Wetzel, P. F. Hopkins, J.-h. Kim, C.-A. Faucher-Giguère, D. Kereš, and E. Quataert. Reconciling Dwarf Galaxies with ΛCDM Cosmology: Simulating a Realistic Population of Satellites around a Milky Way-mass Galaxy. Astrophysical Journal Letters, 827:L23, August 2016. doi: 10.3847/2041-8205/827/2/L23.

A. B. Whiting, M. J. Irwin, and G. K. T. Hau. A new galaxy in the local group: The antlia dwarf galaxy. Astronomical Journal, 114:996–1001, September 1997. doi: 10.1086/118530.

M. Yagi, J. Koda, Y. Komiyama, and H. Yamanoi. Catalog of Ultra-diﬀuse Galaxies in the Coma Clusters from Subaru Imaging Data. Astrophysical Journal Supplement Series, 225:11, July 2016. doi: 10.3847/0067-0049/225/1/11.

J. H. Yoon, K. V. Johnston, and D. W. Hogg. Clumpy Streams from Clumpy Halos: Detecting Missing Satellites with Cold Stellar Structures. Astrophysical Journal, 731:58, April 2011. doi: 10.1088/0004-637X/731/1/58.

D. B. Zucker, A. Y. Kniazev, E. F. Bell, D. Martínez-Delgado, E. K. Grebel, H.- W. Rix, C. M. Rockosi, J. A. Holtzman, R. A. M. Walterbos, J. Annis, D. G. York, Ž. Ivezić, J. Brinkmann, H. Brewington, M. Harvanek, G. Hennessy, S. J. Kleinman, J. Krzesinski, D. Long, P. R. Newman, A. Nitta, and S. A. Snedden. Andromeda IX: A New Dwarf Spheroidal Satellite of M31. Astrophysical Journal Letters, 612:L121–L124, September 2004. doi: 10.1086/424691.

D. B. Zucker, A. Y. Kniazev, D. Martínez-Delgado, E. F. Bell, H.-W. Rix, E. K. Grebel, J. A. Holtzman, R. A. M. Walterbos, C. M. Rockosi, D. G. York, J. C. Barentine, H. Brewington, J. Brinkmann, M. Harvanek, S. J. Kleinman, J. Krzesin- ski, D. Long, E. H. Neilsen, Jr., A. Nitta, and S. A. Snedden. Andromeda X, a New BIBLIOGRAPHY 172

Dwarf Spheroidal Satellite of M31: Photometry. Astrophysical Journal Letters, 659: L21–L24, April 2007. doi: 10.1086/516748.

F. Zwicky. On the Large Scale Distribution of Matter in the Universe. Physical Review, 61:489–503, April 1942. doi: 10.1103/PhysRev.61.489.